kernel-vserver-2.3.patch

   1 diff -NurpP --minimal linux-2.6.25.11/arch/alpha/Kconfig linux-2.6.25.11-vs2.3.0.34.14/arch/alpha/Kconfig
   2 --- linux-2.6.25.11/arch/alpha/Kconfig  2008-04-17 12:05:26.000000000 -0400
   3 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/alpha/Kconfig    2008-04-19 15:14:51.000000000 -0400
   4 @@ -671,6 +671,8 @@ config DUMMY_CONSOLE
   5         depends on VGA_HOSE
   6         default y
   7
   8 +source "kernel/vserver/Kconfig"
   9 +
  10  source "security/Kconfig"
  11
  12  source "crypto/Kconfig"
  13 diff -NurpP --minimal linux-2.6.25.11/arch/alpha/kernel/entry.S linux-2.6.25.11-vs2.3.0.34.14/arch/alpha/kernel/entry.S
  14 --- linux-2.6.25.11/arch/alpha/kernel/entry.S   2008-04-17 11:31:21.000000000 -0400
  15 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/alpha/kernel/entry.S     2008-04-21 11:09:01.000000000 -0400
  16 @@ -872,24 +872,15 @@ sys_getxgid:
  17         .globl  sys_getxpid
  18         .ent    sys_getxpid
  19  sys_getxpid:
  20 +       lda     $sp, -16($sp)
  21 +       stq     $26, 0($sp)
  22         .prologue 0
  23 -       ldq     $2, TI_TASK($8)
  24
  25 -       /* See linux/kernel/timer.c sys_getppid for discussion
  26 -          about this loop.  */
  27 -       ldq     $3, TASK_GROUP_LEADER($2)
  28 -       ldq     $4, TASK_REAL_PARENT($3)
  29 -       ldl     $0, TASK_TGID($2)
  30 -1:     ldl     $1, TASK_TGID($4)
  31 -#ifdef CONFIG_SMP
  32 -       mov     $4, $5
  33 -       mb
  34 -       ldq     $3, TASK_GROUP_LEADER($2)
  35 -       ldq     $4, TASK_REAL_PARENT($3)
  36 -       cmpeq   $4, $5, $5
  37 -       beq     $5, 1b
  38 -#endif
  39 -       stq     $1, 80($sp)
  40 +       lda     $16, 96($sp)
  41 +       jsr     $26, do_getxpid
  42 +       ldq     $26, 0($sp)
  43 +
  44 +       lda     $sp, 16($sp)
  45         ret
  46  .end sys_getxpid
  47
  48 diff -NurpP --minimal linux-2.6.25.11/arch/alpha/kernel/osf_sys.c linux-2.6.25.11-vs2.3.0.34.14/arch/alpha/kernel/osf_sys.c
  49 --- linux-2.6.25.11/arch/alpha/kernel/osf_sys.c 2008-07-15 22:05:16.000000000 -0400
  50 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/alpha/kernel/osf_sys.c   2008-05-21 14:30:40.000000000 -0400
  51 @@ -883,7 +883,7 @@ osf_gettimeofday(struct timeval32 __user
  52  {
  53         if (tv) {
  54                 struct timeval ktv;
  55 -               do_gettimeofday(&ktv);
  56 +               vx_gettimeofday(&ktv);
  57                 if (put_tv32(tv, &ktv))
  58                         return -EFAULT;
  59         }
  60 diff -NurpP --minimal linux-2.6.25.11/arch/alpha/kernel/ptrace.c linux-2.6.25.11-vs2.3.0.34.14/arch/alpha/kernel/ptrace.c
  61 --- linux-2.6.25.11/arch/alpha/kernel/ptrace.c  2008-04-17 11:31:21.000000000 -0400
  62 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/alpha/kernel/ptrace.c    2008-04-19 15:14:51.000000000 -0400
  63 @@ -15,6 +15,7 @@
  64  #include <linux/slab.h>
  65  #include <linux/security.h>
  66  #include <linux/signal.h>
  67 +#include <linux/vs_base.h>
  68
  69  #include <asm/uaccess.h>
  70  #include <asm/pgtable.h>
  71 diff -NurpP --minimal linux-2.6.25.11/arch/alpha/kernel/semaphore.c linux-2.6.25.11-vs2.3.0.34.14/arch/alpha/kernel/semaphore.c
  72 --- linux-2.6.25.11/arch/alpha/kernel/semaphore.c       2008-04-17 11:31:21.000000000 -0400
  73 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/alpha/kernel/semaphore.c 2008-04-19 15:14:51.000000000 -0400
  74 @@ -68,8 +68,8 @@ __down_failed(struct semaphore *sem)
  75         DECLARE_WAITQUEUE(wait, tsk);
  76
  77  #ifdef CONFIG_DEBUG_SEMAPHORE
  78 -       printk("%s(%d): down failed(%p)\n",
  79 -              tsk->comm, task_pid_nr(tsk), sem);
  80 +       printk("%s(%d:#%u): down failed(%p)\n",
  81 +              tsk->comm, task_pid_nr(tsk), tsk->xid, sem);
  82  #endif
  83
  84         tsk->state = TASK_UNINTERRUPTIBLE;
  85 @@ -97,8 +97,8 @@ __down_failed(struct semaphore *sem)
  86         wake_up(&sem->wait);
  87
  88  #ifdef CONFIG_DEBUG_SEMAPHORE
  89 -       printk("%s(%d): down acquired(%p)\n",
  90 -              tsk->comm, task_pid_nr(tsk), sem);
  91 +       printk("%s(%d:#%u): down acquired(%p)\n",
  92 +              tsk->comm, task_pid_nr(tsk), tsk->xid, sem);
  93  #endif
  94  }
  95
  96 @@ -110,8 +110,8 @@ __down_failed_interruptible(struct semap
  97         long ret = 0;
  98
  99  #ifdef CONFIG_DEBUG_SEMAPHORE
 100 -       printk("%s(%d): down failed(%p)\n",
 101 -              tsk->comm, task_pid_nr(tsk), sem);
 102 +       printk("%s(%d:#%u): down failed(%p)\n",
 103 +              tsk->comm, task_pid_nr(tsk), tsk->xid, sem);
 104  #endif
 105
 106         tsk->state = TASK_INTERRUPTIBLE;
 107 diff -NurpP --minimal linux-2.6.25.11/arch/alpha/kernel/systbls.S linux-2.6.25.11-vs2.3.0.34.14/arch/alpha/kernel/systbls.S
 108 --- linux-2.6.25.11/arch/alpha/kernel/systbls.S 2008-04-17 12:05:26.000000000 -0400
 109 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/alpha/kernel/systbls.S   2008-04-19 15:14:51.000000000 -0400
 110 @@ -446,7 +446,7 @@ sys_call_table:
 111         .quad sys_stat64                        /* 425 */
 112         .quad sys_lstat64
 113         .quad sys_fstat64
 114 -       .quad sys_ni_syscall                    /* sys_vserver */
 115 +       .quad sys_vserver                       /* sys_vserver */
 116         .quad sys_ni_syscall                    /* sys_mbind */
 117         .quad sys_ni_syscall                    /* sys_get_mempolicy */
 118         .quad sys_ni_syscall                    /* sys_set_mempolicy */
 119 diff -NurpP --minimal linux-2.6.25.11/arch/alpha/kernel/traps.c linux-2.6.25.11-vs2.3.0.34.14/arch/alpha/kernel/traps.c
 120 --- linux-2.6.25.11/arch/alpha/kernel/traps.c   2008-04-17 11:31:21.000000000 -0400
 121 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/alpha/kernel/traps.c     2008-04-19 15:14:51.000000000 -0400
 122 @@ -182,7 +182,8 @@ die_if_kernel(char * str, struct pt_regs
 123  #ifdef CONFIG_SMP
 124         printk("CPU %d ", hard_smp_processor_id());
 125  #endif
 126 -       printk("%s(%d): %s %ld\n", current->comm, task_pid_nr(current), str, err);
 127 +       printk("%s(%d[#%u]): %s %ld\n", current->comm,
 128 +               task_pid_nr(current), current->xid, str, err);
 129         dik_show_regs(regs, r9_15);
 130         add_taint(TAINT_DIE);
 131         dik_show_trace((unsigned long *)(regs+1));
 132 diff -NurpP --minimal linux-2.6.25.11/arch/alpha/mm/fault.c linux-2.6.25.11-vs2.3.0.34.14/arch/alpha/mm/fault.c
 133 --- linux-2.6.25.11/arch/alpha/mm/fault.c       2008-04-17 11:31:21.000000000 -0400
 134 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/alpha/mm/fault.c 2008-04-19 15:14:51.000000000 -0400
 135 @@ -193,8 +193,8 @@ do_page_fault(unsigned long address, uns
 136                 down_read(&mm->mmap_sem);
 137                 goto survive;
 138         }
 139 -       printk(KERN_ALERT "VM: killing process %s(%d)\n",
 140 -              current->comm, task_pid_nr(current));
 141 +       printk(KERN_ALERT "VM: killing process %s(%d:#%u)\n",
 142 +              current->comm, task_pid_nr(current), current->xid);
 143         if (!user_mode(regs))
 144                 goto no_context;
 145         do_group_exit(SIGKILL);
 146 diff -NurpP --minimal linux-2.6.25.11/arch/arm/Kconfig linux-2.6.25.11-vs2.3.0.34.14/arch/arm/Kconfig
 147 --- linux-2.6.25.11/arch/arm/Kconfig    2008-04-17 12:05:26.000000000 -0400
 148 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/arm/Kconfig      2008-04-19 15:14:51.000000000 -0400
 149 @@ -1180,6 +1180,8 @@ source "fs/Kconfig"
 150
 151  source "arch/arm/Kconfig.debug"
 152
 153 +source "kernel/vserver/Kconfig"
 154 +
 155  source "security/Kconfig"
 156
 157  source "crypto/Kconfig"
 158 diff -NurpP --minimal linux-2.6.25.11/arch/arm/kernel/calls.S linux-2.6.25.11-vs2.3.0.34.14/arch/arm/kernel/calls.S
 159 --- linux-2.6.25.11/arch/arm/kernel/calls.S     2008-04-17 12:05:26.000000000 -0400
 160 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/arm/kernel/calls.S       2008-04-19 15:14:51.000000000 -0400
 161 @@ -322,7 +322,7 @@
 162  /* 310 */      CALL(sys_request_key)
 163                 CALL(sys_keyctl)
 164                 CALL(ABI(sys_semtimedop, sys_oabi_semtimedop))
 165 -/* vserver */  CALL(sys_ni_syscall)
 166 +               CALL(sys_vserver)
 167                 CALL(sys_ioprio_set)
 168  /* 315 */      CALL(sys_ioprio_get)
 169                 CALL(sys_inotify_init)
 170 diff -NurpP --minimal linux-2.6.25.11/arch/arm/kernel/process.c linux-2.6.25.11-vs2.3.0.34.14/arch/arm/kernel/process.c
 171 --- linux-2.6.25.11/arch/arm/kernel/process.c   2008-04-17 12:05:26.000000000 -0400
 172 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/arm/kernel/process.c     2008-04-21 11:09:01.000000000 -0400
 173 @@ -264,7 +264,8 @@ void __show_regs(struct pt_regs *regs)
 174  void show_regs(struct pt_regs * regs)
 175  {
 176         printk("\n");
 177 -       printk("Pid: %d, comm: %20s\n", task_pid_nr(current), current->comm);
 178 +       printk("Pid: %d[#%u], comm: %20s\n",
 179 +               task_pid_nr(current), current->xid, current->comm);
 180         __show_regs(regs);
 181         __backtrace();
 182  }
 183 diff -NurpP --minimal linux-2.6.25.11/arch/arm/kernel/traps.c linux-2.6.25.11-vs2.3.0.34.14/arch/arm/kernel/traps.c
 184 --- linux-2.6.25.11/arch/arm/kernel/traps.c     2008-04-17 12:05:26.000000000 -0400
 185 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/arm/kernel/traps.c       2008-04-19 15:14:51.000000000 -0400
 186 @@ -214,8 +214,8 @@ static void __die(const char *str, int e
 187                str, err, ++die_counter);
 188         print_modules();
 189         __show_regs(regs);
 190 -       printk("Process %s (pid: %d, stack limit = 0x%p)\n",
 191 -               tsk->comm, task_pid_nr(tsk), thread + 1);
 192 +       printk("Process %s (pid: %d:#%u, stack limit = 0x%p)\n",
 193 +               tsk->comm, task_pid_nr(tsk), tsk->xid, thread + 1);
 194
 195         if (!user_mode(regs) || in_interrupt()) {
 196                 dump_mem("Stack: ", regs->ARM_sp,
 197 diff -NurpP --minimal linux-2.6.25.11/arch/arm/mm/fault.c linux-2.6.25.11-vs2.3.0.34.14/arch/arm/mm/fault.c
 198 --- linux-2.6.25.11/arch/arm/mm/fault.c 2008-04-17 12:05:27.000000000 -0400
 199 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/arm/mm/fault.c   2008-04-19 15:14:51.000000000 -0400
 200 @@ -292,7 +292,8 @@ do_page_fault(unsigned long addr, unsign
 201                  * happened to us that made us unable to handle
 202                  * the page fault gracefully.
 203                  */
 204 -               printk("VM: killing process %s\n", tsk->comm);
 205 +               printk("VM: killing process %s(%d:#%u)\n",
 206 +                       tsk->comm, task_pid_nr(tsk), tsk->xid);
 207                 do_group_exit(SIGKILL);
 208                 return 0;
 209         }
 210 diff -NurpP --minimal linux-2.6.25.11/arch/cris/Kconfig linux-2.6.25.11-vs2.3.0.34.14/arch/cris/Kconfig
 211 --- linux-2.6.25.11/arch/cris/Kconfig   2008-04-17 12:05:27.000000000 -0400
 212 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/cris/Kconfig     2008-04-19 15:14:51.000000000 -0400
 213 @@ -679,6 +679,8 @@ source "drivers/usb/Kconfig"
 214
 215  source "arch/cris/Kconfig.debug"
 216
 217 +source "kernel/vserver/Kconfig"
 218 +
 219  source "security/Kconfig"
 220
 221  source "crypto/Kconfig"
 222 diff -NurpP --minimal linux-2.6.25.11/arch/frv/kernel/kernel_thread.S linux-2.6.25.11-vs2.3.0.34.14/arch/frv/kernel/kernel_thread.S
 223 --- linux-2.6.25.11/arch/frv/kernel/kernel_thread.S     2007-02-04 13:44:54.000000000 -0500
 224 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/frv/kernel/kernel_thread.S       2008-04-21 11:09:01.000000000 -0400
 225 @@ -37,7 +37,7 @@ kernel_thread:
 226
 227         # start by forking the current process, but with shared VM
 228         setlos.p        #__NR_clone,gr7         ; syscall number
 229 -       ori             gr10,#CLONE_VM,gr8      ; first syscall arg     [clone_flags]
 230 +       ori             gr10,#CLONE_KT,gr8      ; first syscall arg     [clone_flags]
 231         sethi.p         #0xe4e4,gr9             ; second syscall arg    [newsp]
 232         setlo           #0xe4e4,gr9
 233         setlos.p        #0,gr10                 ; third syscall arg     [parent_tidptr]
 234 diff -NurpP --minimal linux-2.6.25.11/arch/h8300/Kconfig linux-2.6.25.11-vs2.3.0.34.14/arch/h8300/Kconfig
 235 --- linux-2.6.25.11/arch/h8300/Kconfig  2008-04-17 12:05:28.000000000 -0400
 236 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/h8300/Kconfig    2008-04-19 15:14:51.000000000 -0400
 237 @@ -233,6 +233,8 @@ source "fs/Kconfig"
 238
 239  source "arch/h8300/Kconfig.debug"
 240
 241 +source "kernel/vserver/Kconfig"
 242 +
 243  source "security/Kconfig"
 244
 245  source "crypto/Kconfig"
 246 diff -NurpP --minimal linux-2.6.25.11/arch/ia64/ia32/ia32_entry.S linux-2.6.25.11-vs2.3.0.34.14/arch/ia64/ia32/ia32_entry.S
 247 --- linux-2.6.25.11/arch/ia64/ia32/ia32_entry.S 2008-04-17 10:37:14.000000000 -0400
 248 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/ia64/ia32/ia32_entry.S   2008-04-19 15:14:51.000000000 -0400
 249 @@ -446,7 +446,7 @@ ia32_syscall_table:
 250         data8 sys_tgkill        /* 270 */
 251         data8 compat_sys_utimes
 252         data8 sys32_fadvise64_64
 253 -       data8 sys_ni_syscall
 254 +       data8 sys32_vserver
 255         data8 sys_ni_syscall
 256         data8 sys_ni_syscall    /* 275 */
 257         data8 sys_ni_syscall
 258 diff -NurpP --minimal linux-2.6.25.11/arch/ia64/ia32/sys_ia32.c linux-2.6.25.11-vs2.3.0.34.14/arch/ia64/ia32/sys_ia32.c
 259 --- linux-2.6.25.11/arch/ia64/ia32/sys_ia32.c   2008-04-17 12:05:28.000000000 -0400
 260 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/ia64/ia32/sys_ia32.c     2008-04-19 15:14:51.000000000 -0400
 261 @@ -1177,7 +1177,7 @@ sys32_gettimeofday (struct compat_timeva
 262  {
 263         if (tv) {
 264                 struct timeval ktv;
 265 -               do_gettimeofday(&ktv);
 266 +               vx_gettimeofday(&ktv);
 267                 if (put_tv32(tv, &ktv))
 268                         return -EFAULT;
 269         }
 270 diff -NurpP --minimal linux-2.6.25.11/arch/ia64/Kconfig linux-2.6.25.11-vs2.3.0.34.14/arch/ia64/Kconfig
 271 --- linux-2.6.25.11/arch/ia64/Kconfig   2008-04-17 12:05:28.000000000 -0400
 272 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/ia64/Kconfig     2008-04-19 15:14:51.000000000 -0400
 273 @@ -615,6 +615,8 @@ source "arch/ia64/hp/sim/Kconfig"
 274
 275  source "arch/ia64/Kconfig.debug"
 276
 277 +source "kernel/vserver/Kconfig"
 278 +
 279  source "security/Kconfig"
 280
 281  source "crypto/Kconfig"
 282 diff -NurpP --minimal linux-2.6.25.11/arch/ia64/kernel/entry.S linux-2.6.25.11-vs2.3.0.34.14/arch/ia64/kernel/entry.S
 283 --- linux-2.6.25.11/arch/ia64/kernel/entry.S    2008-04-17 12:05:28.000000000 -0400
 284 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/ia64/kernel/entry.S      2008-04-19 15:14:51.000000000 -0400
 285 @@ -1547,7 +1547,7 @@ sys_call_table:
 286         data8 sys_mq_notify
 287         data8 sys_mq_getsetattr
 288         data8 sys_kexec_load
 289 -       data8 sys_ni_syscall                    // reserved for vserver
 290 +       data8 sys_vserver
 291         data8 sys_waitid                        // 1270
 292         data8 sys_add_key
 293         data8 sys_request_key
 294 diff -NurpP --minimal linux-2.6.25.11/arch/ia64/kernel/perfmon.c linux-2.6.25.11-vs2.3.0.34.14/arch/ia64/kernel/perfmon.c
 295 --- linux-2.6.25.11/arch/ia64/kernel/perfmon.c  2008-04-17 12:05:28.000000000 -0400
 296 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/ia64/kernel/perfmon.c    2008-04-19 15:14:51.000000000 -0400
 297 @@ -40,6 +40,7 @@
 298  #include <linux/capability.h>
 299  #include <linux/rcupdate.h>
 300  #include <linux/completion.h>
 301 +#include <linux/vs_memory.h>
 302
 303  #include <asm/errno.h>
 304  #include <asm/intrinsics.h>
 305 @@ -2374,7 +2375,7 @@ pfm_smpl_buffer_alloc(struct task_struct
 306          */
 307         insert_vm_struct(mm, vma);
 308
 309 -       mm->total_vm  += size >> PAGE_SHIFT;
 310 +       vx_vmpages_add(mm, size >> PAGE_SHIFT);
 311         vm_stat_account(vma->vm_mm, vma->vm_flags, vma->vm_file,
 312                                                         vma_pages(vma));
 313         up_write(&task->mm->mmap_sem);
 314 diff -NurpP --minimal linux-2.6.25.11/arch/ia64/kernel/process.c linux-2.6.25.11-vs2.3.0.34.14/arch/ia64/kernel/process.c
 315 --- linux-2.6.25.11/arch/ia64/kernel/process.c  2008-04-17 12:05:28.000000000 -0400
 316 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/ia64/kernel/process.c    2008-04-21 11:09:01.000000000 -0400
 317 @@ -105,8 +105,8 @@ show_regs (struct pt_regs *regs)
 318         unsigned long ip = regs->cr_iip + ia64_psr(regs)->ri;
 319
 320         print_modules();
 321 -       printk("\nPid: %d, CPU %d, comm: %20s\n", task_pid_nr(current),
 322 -                       smp_processor_id(), current->comm);
 323 +       printk("\nPid: %d[#%u], CPU %d, comm: %20s\n", task_pid_nr(current),
 324 +                       current->xid, smp_processor_id(), current->comm);
 325         printk("psr : %016lx ifs : %016lx ip  : [<%016lx>]    %s (%s)\n",
 326                regs->cr_ipsr, regs->cr_ifs, ip, print_tainted(),
 327                init_utsname()->release);
 328 diff -NurpP --minimal linux-2.6.25.11/arch/ia64/kernel/ptrace.c linux-2.6.25.11-vs2.3.0.34.14/arch/ia64/kernel/ptrace.c
 329 --- linux-2.6.25.11/arch/ia64/kernel/ptrace.c   2008-04-17 12:05:28.000000000 -0400
 330 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/ia64/kernel/ptrace.c     2008-04-19 15:14:51.000000000 -0400
 331 @@ -17,6 +17,7 @@
 332  #include <linux/security.h>
 333  #include <linux/audit.h>
 334  #include <linux/signal.h>
 335 +#include <linux/vs_base.h>
 336
 337  #include <asm/pgtable.h>
 338  #include <asm/processor.h>
 339 diff -NurpP --minimal linux-2.6.25.11/arch/ia64/kernel/traps.c linux-2.6.25.11-vs2.3.0.34.14/arch/ia64/kernel/traps.c
 340 --- linux-2.6.25.11/arch/ia64/kernel/traps.c    2008-04-17 12:05:28.000000000 -0400
 341 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/ia64/kernel/traps.c      2008-04-21 10:33:04.000000000 -0400
 342 @@ -60,8 +60,9 @@ die (const char *str, struct pt_regs *re
 343         put_cpu();
 344
 345         if (++die.lock_owner_depth < 3) {
 346 -               printk("%s[%d]: %s %ld [%d]\n",
 347 -               current->comm, task_pid_nr(current), str, err, ++die_counter);
 348 +               printk("%s[%d[#%u]]: %s %ld [%d]\n",
 349 +                       current->comm, task_pid_nr(current), current->xid,
 350 +                       str, err, ++die_counter);
 351                 if (notify_die(DIE_OOPS, str, regs, err, 255, SIGSEGV)
 352                     != NOTIFY_STOP)
 353                         show_regs(regs);
 354 @@ -324,8 +325,9 @@ handle_fpu_swa (int fp_fault, struct pt_
 355                         if ((last.count & 15) < 5 && (ia64_fetchadd(1, &last.count, acq) & 15) < 5) {
 356                                 last.time = current_jiffies + 5 * HZ;
 357                                 printk(KERN_WARNING
 358 -                                       "%s(%d): floating-point assist fault at ip %016lx, isr %016lx\n",
 359 -                                       current->comm, task_pid_nr(current), regs->cr_iip + ia64_psr(regs)->ri, isr);
 360 +                                       "%s(%d[#%u]): floating-point assist fault at ip %016lx, isr %016lx\n",
 361 +                                       current->comm, task_pid_nr(current), current->xid,
 362 +                                       regs->cr_iip + ia64_psr(regs)->ri, isr);
 363                         }
 364                 }
 365         }
 366 diff -NurpP --minimal linux-2.6.25.11/arch/ia64/mm/fault.c linux-2.6.25.11-vs2.3.0.34.14/arch/ia64/mm/fault.c
 367 --- linux-2.6.25.11/arch/ia64/mm/fault.c        2008-04-17 12:05:28.000000000 -0400
 368 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/ia64/mm/fault.c  2008-04-19 15:14:52.000000000 -0400
 369 @@ -10,6 +10,7 @@
 370  #include <linux/interrupt.h>
 371  #include <linux/kprobes.h>
 372  #include <linux/kdebug.h>
 373 +#include <linux/vs_memory.h>
 374
 375  #include <asm/pgtable.h>
 376  #include <asm/processor.h>
 377 diff -NurpP --minimal linux-2.6.25.11/arch/m32r/kernel/traps.c linux-2.6.25.11-vs2.3.0.34.14/arch/m32r/kernel/traps.c
 378 --- linux-2.6.25.11/arch/m32r/kernel/traps.c    2008-04-17 11:31:23.000000000 -0400
 379 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/m32r/kernel/traps.c      2008-04-19 15:14:52.000000000 -0400
 380 @@ -195,8 +195,9 @@ static void show_registers(struct pt_reg
 381         } else {
 382                 printk("SPI: %08lx\n", sp);
 383         }
 384 -       printk("Process %s (pid: %d, process nr: %d, stackpage=%08lx)",
 385 -               current->comm, task_pid_nr(current), 0xffff & i, 4096+(unsigned long)current);
 386 +       printk("Process %s (pid: %d[#%u], process nr: %d, stackpage=%08lx)",
 387 +               current->comm, task_pid_nr(current), current->xid,
 388 +               0xffff & i, 4096+(unsigned long)current);
 389
 390         /*
 391          * When in-kernel, we also print out the stack and code at the
 392 diff -NurpP --minimal linux-2.6.25.11/arch/m68k/Kconfig linux-2.6.25.11-vs2.3.0.34.14/arch/m68k/Kconfig
 393 --- linux-2.6.25.11/arch/m68k/Kconfig   2008-04-17 12:05:28.000000000 -0400
 394 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/m68k/Kconfig     2008-04-19 15:14:52.000000000 -0400
 395 @@ -674,6 +674,8 @@ source "fs/Kconfig"
 396
 397  source "arch/m68k/Kconfig.debug"
 398
 399 +source "kernel/vserver/Kconfig"
 400 +
 401  source "security/Kconfig"
 402
 403  source "crypto/Kconfig"
 404 diff -NurpP --minimal linux-2.6.25.11/arch/m68k/kernel/ptrace.c linux-2.6.25.11-vs2.3.0.34.14/arch/m68k/kernel/ptrace.c
 405 --- linux-2.6.25.11/arch/m68k/kernel/ptrace.c   2008-04-17 11:31:23.000000000 -0400
 406 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/m68k/kernel/ptrace.c     2008-04-19 15:14:52.000000000 -0400
 407 @@ -18,6 +18,7 @@
 408  #include <linux/ptrace.h>
 409  #include <linux/user.h>
 410  #include <linux/signal.h>
 411 +#include <linux/vs_base.h>
 412
 413  #include <asm/uaccess.h>
 414  #include <asm/page.h>
 415 @@ -269,6 +270,8 @@ long arch_ptrace(struct task_struct *chi
 416                 ret = ptrace_request(child, request, addr, data);
 417                 break;
 418         }
 419 +       if (!vx_check(vx_task_xid(child), VS_WATCH_P | VS_IDENT))
 420 +               goto out_tsk;
 421
 422         return ret;
 423  out_eio:
 424 diff -NurpP --minimal linux-2.6.25.11/arch/m68k/kernel/traps.c linux-2.6.25.11-vs2.3.0.34.14/arch/m68k/kernel/traps.c
 425 --- linux-2.6.25.11/arch/m68k/kernel/traps.c    2008-04-17 12:05:28.000000000 -0400
 426 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/m68k/kernel/traps.c      2008-04-19 15:14:52.000000000 -0400
 427 @@ -898,8 +898,8 @@ void show_registers(struct pt_regs *regs
 428         printk("d4: %08lx    d5: %08lx    a0: %08lx    a1: %08lx\n",
 429                regs->d4, regs->d5, regs->a0, regs->a1);
 430
 431 -       printk("Process %s (pid: %d, task=%p)\n",
 432 -               current->comm, task_pid_nr(current), current);
 433 +       printk("Process %s (pid: %d[#%u], task=%p)\n",
 434 +               current->comm, task_pid_nr(current), current->xid, current);
 435         addr = (unsigned long)&fp->un;
 436         printk("Frame format=%X ", regs->format);
 437         switch (regs->format) {
 438 diff -NurpP --minimal linux-2.6.25.11/arch/m68knommu/Kconfig linux-2.6.25.11-vs2.3.0.34.14/arch/m68knommu/Kconfig
 439 --- linux-2.6.25.11/arch/m68knommu/Kconfig      2008-04-17 12:05:28.000000000 -0400
 440 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/m68knommu/Kconfig        2008-04-19 15:14:52.000000000 -0400
 441 @@ -722,6 +722,8 @@ source "fs/Kconfig"
 442
 443  source "arch/m68knommu/Kconfig.debug"
 444
 445 +source "kernel/vserver/Kconfig"
 446 +
 447  source "security/Kconfig"
 448
 449  source "crypto/Kconfig"
 450 diff -NurpP --minimal linux-2.6.25.11/arch/m68knommu/kernel/traps.c linux-2.6.25.11-vs2.3.0.34.14/arch/m68knommu/kernel/traps.c
 451 --- linux-2.6.25.11/arch/m68knommu/kernel/traps.c       2008-04-17 10:37:14.000000000 -0400
 452 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/m68knommu/kernel/traps.c 2008-04-19 15:14:52.000000000 -0400
 453 @@ -78,8 +78,9 @@ void die_if_kernel(char *str, struct pt_
 454         printk(KERN_EMERG "d4: %08lx    d5: %08lx    a0: %08lx    a1: %08lx\n",
 455                fp->d4, fp->d5, fp->a0, fp->a1);
 456
 457 -       printk(KERN_EMERG "Process %s (pid: %d, stackpage=%08lx)\n",
 458 -               current->comm, current->pid, PAGE_SIZE+(unsigned long)current);
 459 +       printk(KERN_EMERG "Process %s (pid: %d[#%u], stackpage=%08lx)\n",
 460 +               current->comm, task_pid_nr(current), current->xid,
 461 +               PAGE_SIZE+(unsigned long)current);
 462         show_stack(NULL, (unsigned long *)(fp + 1));
 463         add_taint(TAINT_DIE);
 464         do_exit(SIGSEGV);
 465 diff -NurpP --minimal linux-2.6.25.11/arch/mips/Kconfig linux-2.6.25.11-vs2.3.0.34.14/arch/mips/Kconfig
 466 --- linux-2.6.25.11/arch/mips/Kconfig   2008-04-17 12:05:28.000000000 -0400
 467 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/mips/Kconfig     2008-04-19 15:14:52.000000000 -0400
 468 @@ -2099,6 +2099,8 @@ source "fs/Kconfig"
 469
 470  source "arch/mips/Kconfig.debug"
 471
 472 +source "kernel/vserver/Kconfig"
 473 +
 474  source "security/Kconfig"
 475
 476  source "crypto/Kconfig"
 477 diff -NurpP --minimal linux-2.6.25.11/arch/mips/kernel/linux32.c linux-2.6.25.11-vs2.3.0.34.14/arch/mips/kernel/linux32.c
 478 --- linux-2.6.25.11/arch/mips/kernel/linux32.c  2008-04-17 12:05:29.000000000 -0400
 479 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/mips/kernel/linux32.c    2008-04-19 15:14:52.000000000 -0400
 480 @@ -209,7 +209,7 @@ sys32_gettimeofday(struct compat_timeval
 481  {
 482         if (tv) {
 483                 struct timeval ktv;
 484 -               do_gettimeofday(&ktv);
 485 +               vx_gettimeofday(&ktv);
 486                 if (put_tv32(tv, &ktv))
 487                         return -EFAULT;
 488         }
 489 diff -NurpP --minimal linux-2.6.25.11/arch/mips/kernel/ptrace.c linux-2.6.25.11-vs2.3.0.34.14/arch/mips/kernel/ptrace.c
 490 --- linux-2.6.25.11/arch/mips/kernel/ptrace.c   2008-04-17 11:31:23.000000000 -0400
 491 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/mips/kernel/ptrace.c     2008-04-19 15:14:52.000000000 -0400
 492 @@ -25,6 +25,7 @@
 493  #include <linux/security.h>
 494  #include <linux/audit.h>
 495  #include <linux/seccomp.h>
 496 +#include <linux/vs_base.h>
 497
 498  #include <asm/byteorder.h>
 499  #include <asm/cpu.h>
 500 @@ -171,6 +172,9 @@ long arch_ptrace(struct task_struct *chi
 501  {
 502         int ret;
 503
 504 +       if (!vx_check(vx_task_xid(child), VS_WATCH_P | VS_IDENT))
 505 +               goto out;
 506 +
 507         switch (request) {
 508         /* when I and D space are separate, these will need to be fixed. */
 509         case PTRACE_PEEKTEXT: /* read word at location addr. */
 510 diff -NurpP --minimal linux-2.6.25.11/arch/mips/kernel/scall32-o32.S linux-2.6.25.11-vs2.3.0.34.14/arch/mips/kernel/scall32-o32.S
 511 --- linux-2.6.25.11/arch/mips/kernel/scall32-o32.S      2008-04-17 12:05:29.000000000 -0400
 512 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/mips/kernel/scall32-o32.S        2008-04-19 15:14:52.000000000 -0400
 513 @@ -619,7 +619,7 @@ einval:     li      v0, -EINVAL
 514         sys     sys_mq_timedreceive     5
 515         sys     sys_mq_notify           2       /* 4275 */
 516         sys     sys_mq_getsetattr       3
 517 -       sys     sys_ni_syscall          0       /* sys_vserver */
 518 +       sys     sys_vserver             3
 519         sys     sys_waitid              5
 520         sys     sys_ni_syscall          0       /* available, was setaltroot */
 521         sys     sys_add_key             5       /* 4280 */
 522 diff -NurpP --minimal linux-2.6.25.11/arch/mips/kernel/scall64-64.S linux-2.6.25.11-vs2.3.0.34.14/arch/mips/kernel/scall64-64.S
 523 --- linux-2.6.25.11/arch/mips/kernel/scall64-64.S       2008-04-17 12:05:29.000000000 -0400
 524 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/mips/kernel/scall64-64.S 2008-04-19 15:14:52.000000000 -0400
 525 @@ -434,7 +434,7 @@ sys_call_table:
 526         PTR     sys_mq_timedreceive
 527         PTR     sys_mq_notify
 528         PTR     sys_mq_getsetattr               /* 5235 */
 529 -       PTR     sys_ni_syscall                  /* sys_vserver */
 530 +       PTR     sys_vserver
 531         PTR     sys_waitid
 532         PTR     sys_ni_syscall                  /* available, was setaltroot */
 533         PTR     sys_add_key
 534 diff -NurpP --minimal linux-2.6.25.11/arch/mips/kernel/scall64-n32.S linux-2.6.25.11-vs2.3.0.34.14/arch/mips/kernel/scall64-n32.S
 535 --- linux-2.6.25.11/arch/mips/kernel/scall64-n32.S      2008-04-17 12:05:29.000000000 -0400
 536 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/mips/kernel/scall64-n32.S        2008-04-19 15:14:52.000000000 -0400
 537 @@ -360,7 +360,7 @@ EXPORT(sysn32_call_table)
 538         PTR     compat_sys_mq_timedreceive
 539         PTR     compat_sys_mq_notify
 540         PTR     compat_sys_mq_getsetattr
 541 -       PTR     sys_ni_syscall                  /* 6240, sys_vserver */
 542 +       PTR     sys32_vserver                   /* 6240 */
 543         PTR     compat_sys_waitid
 544         PTR     sys_ni_syscall                  /* available, was setaltroot */
 545         PTR     sys_add_key
 546 diff -NurpP --minimal linux-2.6.25.11/arch/mips/kernel/scall64-o32.S linux-2.6.25.11-vs2.3.0.34.14/arch/mips/kernel/scall64-o32.S
 547 --- linux-2.6.25.11/arch/mips/kernel/scall64-o32.S      2008-04-17 12:05:29.000000000 -0400
 548 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/mips/kernel/scall64-o32.S        2008-04-19 15:14:52.000000000 -0400
 549 @@ -482,7 +482,7 @@ sys_call_table:
 550         PTR     compat_sys_mq_timedreceive
 551         PTR     compat_sys_mq_notify            /* 4275 */
 552         PTR     compat_sys_mq_getsetattr
 553 -       PTR     sys_ni_syscall                  /* sys_vserver */
 554 +       PTR     sys32_vserver
 555         PTR     sys32_waitid
 556         PTR     sys_ni_syscall                  /* available, was setaltroot */
 557         PTR     sys_add_key                     /* 4280 */
 558 diff -NurpP --minimal linux-2.6.25.11/arch/mips/kernel/traps.c linux-2.6.25.11-vs2.3.0.34.14/arch/mips/kernel/traps.c
 559 --- linux-2.6.25.11/arch/mips/kernel/traps.c    2008-04-17 12:05:29.000000000 -0400
 560 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/mips/kernel/traps.c      2008-04-19 15:14:52.000000000 -0400
 561 @@ -313,8 +313,9 @@ void show_registers(const struct pt_regs
 562  {
 563         __show_regs(regs);
 564         print_modules();
 565 -       printk("Process %s (pid: %d, threadinfo=%p, task=%p)\n",
 566 -               current->comm, task_pid_nr(current), current_thread_info(), current);
 567 +       printk("Process %s (pid: %d:#%u, threadinfo=%p, task=%p)\n",
 568 +               current->comm, task_pid_nr(current), current->xid,
 569 +               current_thread_info(), current);
 570         show_stacktrace(current, regs);
 571         show_code((unsigned int __user *) regs->cp0_epc);
 572         printk("\n");
 573 diff -NurpP --minimal linux-2.6.25.11/arch/mips/mm/fault.c linux-2.6.25.11-vs2.3.0.34.14/arch/mips/mm/fault.c
 574 --- linux-2.6.25.11/arch/mips/mm/fault.c        2008-04-17 11:31:24.000000000 -0400
 575 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/mips/mm/fault.c  2008-04-19 15:14:52.000000000 -0400
 576 @@ -178,7 +178,8 @@ out_of_memory:
 577                 down_read(&mm->mmap_sem);
 578                 goto survive;
 579         }
 580 -       printk("VM: killing process %s\n", tsk->comm);
 581 +       printk("VM: killing process %s(%d:#%u)\n",
 582 +               tsk->comm, tsk->pid, tsk->xid);
 583         if (user_mode(regs))
 584                 do_group_exit(SIGKILL);
 585         goto no_context;
 586 diff -NurpP --minimal linux-2.6.25.11/arch/parisc/Kconfig linux-2.6.25.11-vs2.3.0.34.14/arch/parisc/Kconfig
 587 --- linux-2.6.25.11/arch/parisc/Kconfig 2008-04-17 12:05:29.000000000 -0400
 588 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/parisc/Kconfig   2008-04-19 15:14:52.000000000 -0400
 589 @@ -278,6 +278,8 @@ source "fs/Kconfig"
 590
 591  source "arch/parisc/Kconfig.debug"
 592
 593 +source "kernel/vserver/Kconfig"
 594 +
 595  source "security/Kconfig"
 596
 597  source "crypto/Kconfig"
 598 diff -NurpP --minimal linux-2.6.25.11/arch/parisc/kernel/syscall_table.S linux-2.6.25.11-vs2.3.0.34.14/arch/parisc/kernel/syscall_table.S
 599 --- linux-2.6.25.11/arch/parisc/kernel/syscall_table.S  2008-04-17 12:05:29.000000000 -0400
 600 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/parisc/kernel/syscall_table.S    2008-04-19 15:14:52.000000000 -0400
 601 @@ -361,7 +361,7 @@
 602         ENTRY_COMP(mbind)               /* 260 */
 603         ENTRY_COMP(get_mempolicy)
 604         ENTRY_COMP(set_mempolicy)
 605 -       ENTRY_SAME(ni_syscall)  /* 263: reserved for vserver */
 606 +       ENTRY_DIFF(vserver)
 607         ENTRY_SAME(add_key)
 608         ENTRY_SAME(request_key)         /* 265 */
 609         ENTRY_SAME(keyctl)
 610 diff -NurpP --minimal linux-2.6.25.11/arch/parisc/kernel/sys_parisc32.c linux-2.6.25.11-vs2.3.0.34.14/arch/parisc/kernel/sys_parisc32.c
 611 --- linux-2.6.25.11/arch/parisc/kernel/sys_parisc32.c   2008-04-17 11:31:24.000000000 -0400
 612 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/parisc/kernel/sys_parisc32.c     2008-04-19 15:14:52.000000000 -0400
 613 @@ -204,11 +204,11 @@ static inline long get_ts32(struct times
 614  asmlinkage int
 615  sys32_gettimeofday(struct compat_timeval __user *tv, struct timezone __user *tz)
 616  {
 617 -    extern void do_gettimeofday(struct timeval *tv);
 618 +    extern void vx_gettimeofday(struct timeval *tv);
 619
 620      if (tv) {
 621             struct timeval ktv;
 622 -           do_gettimeofday(&ktv);
 623 +           vx_gettimeofday(&ktv);
 624             if (put_compat_timeval(tv, &ktv))
 625                     return -EFAULT;
 626      }
 627 diff -NurpP --minimal linux-2.6.25.11/arch/parisc/kernel/traps.c linux-2.6.25.11-vs2.3.0.34.14/arch/parisc/kernel/traps.c
 628 --- linux-2.6.25.11/arch/parisc/kernel/traps.c  2008-04-17 12:05:29.000000000 -0400
 629 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/parisc/kernel/traps.c    2008-04-19 15:14:52.000000000 -0400
 630 @@ -237,8 +237,9 @@ void die_if_kernel(char *str, struct pt_
 631                 if (err == 0)
 632                         return; /* STFU */
 633
 634 -               printk(KERN_CRIT "%s (pid %d): %s (code %ld) at " RFMT "\n",
 635 -                       current->comm, task_pid_nr(current), str, err, regs->iaoq[0]);
 636 +               printk(KERN_CRIT "%s (pid %d:#%u): %s (code %ld) at " RFMT "\n",
 637 +                       current->comm, task_pid_nr(current), current->xid,
 638 +                       str, err, regs->iaoq[0]);
 639  #ifdef PRINT_USER_FAULTS
 640                 /* XXX for debugging only */
 641                 show_regs(regs);
 642 @@ -270,8 +271,8 @@ KERN_CRIT "                     ||     |
 643                 pdc_console_restart();
 644
 645         if (err)
 646 -               printk(KERN_CRIT "%s (pid %d): %s (code %ld)\n",
 647 -                       current->comm, task_pid_nr(current), str, err);
 648 +               printk(KERN_CRIT "%s (pid %d:#%u): %s (code %ld)\n",
 649 +                       current->comm, task_pid_nr(current), current->xid, str, err);
 650
 651         /* Wot's wrong wif bein' racy? */
 652         if (current->thread.flags & PARISC_KERNEL_DEATH) {
 653 diff -NurpP --minimal linux-2.6.25.11/arch/parisc/mm/fault.c linux-2.6.25.11-vs2.3.0.34.14/arch/parisc/mm/fault.c
 654 --- linux-2.6.25.11/arch/parisc/mm/fault.c      2008-04-17 11:31:24.000000000 -0400
 655 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/parisc/mm/fault.c        2008-04-19 15:14:52.000000000 -0400
 656 @@ -210,8 +210,9 @@ bad_area:
 657
 658  #ifdef PRINT_USER_FAULTS
 659                 printk(KERN_DEBUG "\n");
 660 -               printk(KERN_DEBUG "do_page_fault() pid=%d command='%s' type=%lu address=0x%08lx\n",
 661 -                   task_pid_nr(tsk), tsk->comm, code, address);
 662 +               printk(KERN_DEBUG "do_page_fault() pid=%d:#%u "
 663 +                   "command='%s' type=%lu address=0x%08lx\n",
 664 +                   task_pid_nr(tsk), tsk->xid, tsk->comm, code, address);
 665                 if (vma) {
 666                         printk(KERN_DEBUG "vm_start = 0x%08lx, vm_end = 0x%08lx\n",
 667                                         vma->vm_start, vma->vm_end);
 668 @@ -261,7 +262,8 @@ no_context:
 669
 670    out_of_memory:
 671         up_read(&mm->mmap_sem);
 672 -       printk(KERN_CRIT "VM: killing process %s\n", current->comm);
 673 +       printk(KERN_CRIT "VM: killing process %s(%d:#%u)\n",
 674 +               current->comm, current->pid, current->xid);
 675         if (user_mode(regs))
 676                 do_group_exit(SIGKILL);
 677         goto no_context;
 678 diff -NurpP --minimal linux-2.6.25.11/arch/powerpc/Kconfig linux-2.6.25.11-vs2.3.0.34.14/arch/powerpc/Kconfig
 679 --- linux-2.6.25.11/arch/powerpc/Kconfig        2008-04-17 12:05:29.000000000 -0400
 680 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/powerpc/Kconfig  2008-04-19 15:14:52.000000000 -0400
 681 @@ -706,6 +706,8 @@ source "lib/Kconfig"
 682
 683  source "arch/powerpc/Kconfig.debug"
 684
 685 +source "kernel/vserver/Kconfig"
 686 +
 687  source "security/Kconfig"
 688
 689  config KEYS_COMPAT
 690 diff -NurpP --minimal linux-2.6.25.11/arch/powerpc/kernel/irq.c linux-2.6.25.11-vs2.3.0.34.14/arch/powerpc/kernel/irq.c
 691 --- linux-2.6.25.11/arch/powerpc/kernel/irq.c   2008-04-17 12:05:29.000000000 -0400
 692 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/powerpc/kernel/irq.c     2008-04-19 15:14:52.000000000 -0400
 693 @@ -53,6 +53,7 @@
 694  #include <linux/bootmem.h>
 695  #include <linux/pci.h>
 696  #include <linux/debugfs.h>
 697 +#include <linux/vs_context.h>
 698
 699  #include <asm/uaccess.h>
 700  #include <asm/system.h>
 701 diff -NurpP --minimal linux-2.6.25.11/arch/powerpc/kernel/process.c linux-2.6.25.11-vs2.3.0.34.14/arch/powerpc/kernel/process.c
 702 --- linux-2.6.25.11/arch/powerpc/kernel/process.c       2008-04-17 12:05:30.000000000 -0400
 703 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/powerpc/kernel/process.c 2008-04-19 15:14:52.000000000 -0400
 704 @@ -464,8 +464,9 @@ void show_regs(struct pt_regs * regs)
 705  #else
 706                 printk("DAR: "REG", DSISR: "REG"\n", regs->dar, regs->dsisr);
 707  #endif
 708 -       printk("TASK = %p[%d] '%s' THREAD: %p",
 709 -              current, task_pid_nr(current), current->comm, task_thread_info(current));
 710 +       printk("TASK = %p[%d,#%u] '%s' THREAD: %p",
 711 +              current, task_pid_nr(current), current->xid,
 712 +              current->comm, task_thread_info(current));
 713
 714  #ifdef CONFIG_SMP
 715         printk(" CPU: %d", raw_smp_processor_id());
 716 diff -NurpP --minimal linux-2.6.25.11/arch/powerpc/kernel/sys_ppc32.c linux-2.6.25.11-vs2.3.0.34.14/arch/powerpc/kernel/sys_ppc32.c
 717 --- linux-2.6.25.11/arch/powerpc/kernel/sys_ppc32.c     2008-04-17 11:31:24.000000000 -0400
 718 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/powerpc/kernel/sys_ppc32.c       2008-04-19 15:14:52.000000000 -0400
 719 @@ -205,7 +205,7 @@ asmlinkage long compat_sys_gettimeofday(
 720  {
 721         if (tv) {
 722                 struct timeval ktv;
 723 -               do_gettimeofday(&ktv);
 724 +               vx_gettimeofday(&ktv);
 725                 if (put_tv32(tv, &ktv))
 726                         return -EFAULT;
 727         }
 728 diff -NurpP --minimal linux-2.6.25.11/arch/powerpc/kernel/traps.c linux-2.6.25.11-vs2.3.0.34.14/arch/powerpc/kernel/traps.c
 729 --- linux-2.6.25.11/arch/powerpc/kernel/traps.c 2008-04-17 12:05:30.000000000 -0400
 730 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/powerpc/kernel/traps.c   2008-04-19 15:14:52.000000000 -0400
 731 @@ -941,8 +941,9 @@ void nonrecoverable_exception(struct pt_
 732
 733  void trace_syscall(struct pt_regs *regs)
 734  {
 735 -       printk("Task: %p(%d), PC: %08lX/%08lX, Syscall: %3ld, Result: %s%ld    %s\n",
 736 -              current, task_pid_nr(current), regs->nip, regs->link, regs->gpr[0],
 737 +       printk("Task: %p(%d[#%u]), PC: %08lX/%08lX, Syscall: %3ld, Result: %s%ld    %s\n",
 738 +              current, task_pid_nr(current), current->xid,
 739 +              regs->nip, regs->link, regs->gpr[0],
 740                regs->ccr&0x10000000?"Error=":"", regs->gpr[3], print_tainted());
 741  }
 742
 743 diff -NurpP --minimal linux-2.6.25.11/arch/powerpc/kernel/vdso.c linux-2.6.25.11-vs2.3.0.34.14/arch/powerpc/kernel/vdso.c
 744 --- linux-2.6.25.11/arch/powerpc/kernel/vdso.c  2008-07-15 22:05:16.000000000 -0400
 745 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/powerpc/kernel/vdso.c    2008-07-11 12:41:00.000000000 -0400
 746 @@ -21,6 +21,7 @@
 747  #include <linux/elf.h>
 748  #include <linux/security.h>
 749  #include <linux/bootmem.h>
 750 +#include <linux/vs_memory.h>
 751
 752  #include <asm/pgtable.h>
 753  #include <asm/system.h>
 754 diff -NurpP --minimal linux-2.6.25.11/arch/powerpc/mm/fault.c linux-2.6.25.11-vs2.3.0.34.14/arch/powerpc/mm/fault.c
 755 --- linux-2.6.25.11/arch/powerpc/mm/fault.c     2008-04-17 12:05:30.000000000 -0400
 756 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/powerpc/mm/fault.c       2008-04-19 15:14:52.000000000 -0400
 757 @@ -378,7 +378,8 @@ out_of_memory:
 758                 down_read(&mm->mmap_sem);
 759                 goto survive;
 760         }
 761 -       printk("VM: killing process %s\n", current->comm);
 762 +       printk("VM: killing process %s(%d:#%u)\n",
 763 +               current->comm, current->pid, current->xid);
 764         if (user_mode(regs))
 765                 do_group_exit(SIGKILL);
 766         return SIGKILL;
 767 diff -NurpP --minimal linux-2.6.25.11/arch/ppc/Kconfig linux-2.6.25.11-vs2.3.0.34.14/arch/ppc/Kconfig
 768 --- linux-2.6.25.11/arch/ppc/Kconfig    2008-04-17 12:05:30.000000000 -0400
 769 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/ppc/Kconfig      2008-04-19 15:14:52.000000000 -0400
 770 @@ -1261,6 +1261,8 @@ source "lib/Kconfig"
 771
 772  source "arch/ppc/Kconfig.debug"
 773
 774 +source "kernel/vserver/Kconfig"
 775 +
 776  source "security/Kconfig"
 777
 778  source "crypto/Kconfig"
 779 diff -NurpP --minimal linux-2.6.25.11/arch/ppc/kernel/traps.c linux-2.6.25.11-vs2.3.0.34.14/arch/ppc/kernel/traps.c
 780 --- linux-2.6.25.11/arch/ppc/kernel/traps.c     2008-04-17 12:05:30.000000000 -0400
 781 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/ppc/kernel/traps.c       2008-04-19 15:14:52.000000000 -0400
 782 @@ -669,8 +669,9 @@ void nonrecoverable_exception(struct pt_
 783
 784  void trace_syscall(struct pt_regs *regs)
 785  {
 786 -       printk("Task: %p(%d), PC: %08lX/%08lX, Syscall: %3ld, Result: %s%ld    %s\n",
 787 -              current, current->pid, regs->nip, regs->link, regs->gpr[0],
 788 +       printk("Task: %p(%d[#%u]), PC: %08lX/%08lX, Syscall: %3ld, Result: %s%ld    %s\n",
 789 +              current, current->pid, current->xid,
 790 +              regs->nip, regs->link, regs->gpr[0],
 791                regs->ccr&0x10000000?"Error=":"", regs->gpr[3], print_tainted());
 792  }
 793
 794 diff -NurpP --minimal linux-2.6.25.11/arch/ppc/mm/fault.c linux-2.6.25.11-vs2.3.0.34.14/arch/ppc/mm/fault.c
 795 --- linux-2.6.25.11/arch/ppc/mm/fault.c 2008-04-17 11:31:25.000000000 -0400
 796 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/ppc/mm/fault.c   2008-04-19 15:14:52.000000000 -0400
 797 @@ -295,7 +295,8 @@ out_of_memory:
 798                 down_read(&mm->mmap_sem);
 799                 goto survive;
 800         }
 801 -       printk("VM: killing process %s\n", current->comm);
 802 +       printk("VM: killing process %s(%d:#%u)\n",
 803 +               current->comm, current->pid, current->xid);
 804         if (user_mode(regs))
 805                 do_group_exit(SIGKILL);
 806         return SIGKILL;
 807 diff -NurpP --minimal linux-2.6.25.11/arch/s390/Kconfig linux-2.6.25.11-vs2.3.0.34.14/arch/s390/Kconfig
 808 --- linux-2.6.25.11/arch/s390/Kconfig   2008-04-17 12:05:30.000000000 -0400
 809 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/s390/Kconfig     2008-04-19 15:14:52.000000000 -0400
 810 @@ -544,6 +544,8 @@ source "fs/Kconfig"
 811
 812  source "arch/s390/Kconfig.debug"
 813
 814 +source "kernel/vserver/Kconfig"
 815 +
 816  source "security/Kconfig"
 817
 818  source "crypto/Kconfig"
 819 diff -NurpP --minimal linux-2.6.25.11/arch/s390/kernel/compat_linux.c linux-2.6.25.11-vs2.3.0.34.14/arch/s390/kernel/compat_linux.c
 820 --- linux-2.6.25.11/arch/s390/kernel/compat_linux.c     2008-04-17 11:31:25.000000000 -0400
 821 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/s390/kernel/compat_linux.c       2008-04-19 15:14:52.000000000 -0400
 822 @@ -567,7 +567,7 @@ asmlinkage long sys32_gettimeofday(struc
 823  {
 824         if (tv) {
 825                 struct timeval ktv;
 826 -               do_gettimeofday(&ktv);
 827 +               vx_gettimeofday(&ktv);
 828                 if (put_tv32(tv, &ktv))
 829                         return -EFAULT;
 830         }
 831 diff -NurpP --minimal linux-2.6.25.11/arch/s390/kernel/process.c linux-2.6.25.11-vs2.3.0.34.14/arch/s390/kernel/process.c
 832 --- linux-2.6.25.11/arch/s390/kernel/process.c  2008-04-17 12:05:30.000000000 -0400
 833 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/s390/kernel/process.c    2008-04-21 11:09:01.000000000 -0400
 834 @@ -194,9 +194,9 @@ void show_regs(struct pt_regs *regs)
 835                init_utsname()->release,
 836                (int)strcspn(init_utsname()->version, " "),
 837                init_utsname()->version);
 838 -       printk("Process %s (pid: %d, task: %p, ksp: %p)\n",
 839 -              current->comm, current->pid, current,
 840 -              (void *) current->thread.ksp);
 841 +       printk("Process %s (pid: %d[#%u], task: %p, ksp: %p)\n",
 842 +              current->comm, current->pid, current->xid,
 843 +              (void *) current, (void *) current->thread.ksp);
 844         show_registers(regs);
 845         /* Show stack backtrace if pt_regs is from kernel mode */
 846         if (!(regs->psw.mask & PSW_MASK_PSTATE))
 847 diff -NurpP --minimal linux-2.6.25.11/arch/s390/kernel/ptrace.c linux-2.6.25.11-vs2.3.0.34.14/arch/s390/kernel/ptrace.c
 848 --- linux-2.6.25.11/arch/s390/kernel/ptrace.c   2008-04-17 12:05:30.000000000 -0400
 849 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/s390/kernel/ptrace.c     2008-04-19 15:14:52.000000000 -0400
 850 @@ -33,6 +33,7 @@
 851  #include <linux/security.h>
 852  #include <linux/audit.h>
 853  #include <linux/signal.h>
 854 +#include <linux/vs_base.h>
 855
 856  #include <asm/segment.h>
 857  #include <asm/page.h>
 858 @@ -710,7 +711,13 @@ sys_ptrace(long request, long pid, long
 859                 goto out;
 860         }
 861
 862 +       if (!vx_check(vx_task_xid(child), VS_WATCH_P | VS_IDENT)) {
 863 +               ret = -EPERM;
 864 +               goto out_tsk;
 865 +       }
 866 +
 867         ret = do_ptrace(child, request, addr, data);
 868 +out_tsk:
 869         put_task_struct(child);
 870  out:
 871         unlock_kernel();
 872 diff -NurpP --minimal linux-2.6.25.11/arch/s390/kernel/syscalls.S linux-2.6.25.11-vs2.3.0.34.14/arch/s390/kernel/syscalls.S
 873 --- linux-2.6.25.11/arch/s390/kernel/syscalls.S 2008-04-17 12:05:30.000000000 -0400
 874 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/s390/kernel/syscalls.S   2008-04-19 15:14:52.000000000 -0400
 875 @@ -271,7 +271,7 @@ SYSCALL(sys_clock_settime,sys_clock_sett
 876  SYSCALL(sys_clock_gettime,sys_clock_gettime,sys32_clock_gettime_wrapper)       /* 260 */
 877  SYSCALL(sys_clock_getres,sys_clock_getres,sys32_clock_getres_wrapper)
 878  SYSCALL(sys_clock_nanosleep,sys_clock_nanosleep,sys32_clock_nanosleep_wrapper)
 879 -NI_SYSCALL                                                     /* reserved for vserver */
 880 +SYSCALL(sys_vserver,sys_vserver,sys32_vserver)
 881  SYSCALL(s390_fadvise64_64,sys_ni_syscall,sys32_fadvise64_64_wrapper)
 882  SYSCALL(sys_statfs64,sys_statfs64,compat_sys_statfs64_wrapper)
 883  SYSCALL(sys_fstatfs64,sys_fstatfs64,compat_sys_fstatfs64_wrapper)
 884 diff -NurpP --minimal linux-2.6.25.11/arch/s390/mm/fault.c linux-2.6.25.11-vs2.3.0.34.14/arch/s390/mm/fault.c
 885 --- linux-2.6.25.11/arch/s390/mm/fault.c        2008-04-17 12:05:30.000000000 -0400
 886 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/s390/mm/fault.c  2008-04-19 15:14:52.000000000 -0400
 887 @@ -217,7 +217,8 @@ static int do_out_of_memory(struct pt_re
 888                 down_read(&mm->mmap_sem);
 889                 return 1;
 890         }
 891 -       printk("VM: killing process %s\n", tsk->comm);
 892 +       printk("VM: killing process %s(%d:#%u)\n",
 893 +               tsk->comm, tsk->pid, tsk->xid);
 894         if (regs->psw.mask & PSW_MASK_PSTATE)
 895                 do_group_exit(SIGKILL);
 896         do_no_context(regs, error_code, address);
 897 diff -NurpP --minimal linux-2.6.25.11/arch/sh/Kconfig linux-2.6.25.11-vs2.3.0.34.14/arch/sh/Kconfig
 898 --- linux-2.6.25.11/arch/sh/Kconfig     2008-04-17 12:05:30.000000000 -0400
 899 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/sh/Kconfig       2008-04-19 15:14:52.000000000 -0400
 900 @@ -913,6 +913,8 @@ source "fs/Kconfig"
 901
 902  source "arch/sh/Kconfig.debug"
 903
 904 +source "kernel/vserver/Kconfig"
 905 +
 906  source "security/Kconfig"
 907
 908  source "crypto/Kconfig"
 909 diff -NurpP --minimal linux-2.6.25.11/arch/sh/kernel/irq.c linux-2.6.25.11-vs2.3.0.34.14/arch/sh/kernel/irq.c
 910 --- linux-2.6.25.11/arch/sh/kernel/irq.c        2008-04-17 12:05:30.000000000 -0400
 911 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/sh/kernel/irq.c  2008-04-19 15:14:52.000000000 -0400
 912 @@ -11,6 +11,7 @@
 913  #include <linux/module.h>
 914  #include <linux/kernel_stat.h>
 915  #include <linux/seq_file.h>
 916 +#include <linux/vs_context.h>
 917  #include <asm/processor.h>
 918  #include <asm/machvec.h>
 919  #include <asm/uaccess.h>
 920 diff -NurpP --minimal linux-2.6.25.11/arch/sh/kernel/vsyscall/vsyscall.c linux-2.6.25.11-vs2.3.0.34.14/arch/sh/kernel/vsyscall/vsyscall.c
 921 --- linux-2.6.25.11/arch/sh/kernel/vsyscall/vsyscall.c  2008-04-17 10:37:14.000000000 -0400
 922 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/sh/kernel/vsyscall/vsyscall.c    2008-04-19 15:14:52.000000000 -0400
 923 @@ -19,6 +19,7 @@
 924  #include <linux/elf.h>
 925  #include <linux/sched.h>
 926  #include <linux/err.h>
 927 +#include <linux/vs_memory.h>
 928
 929  /*
 930   * Should the kernel map a VDSO page into processes and pass its
 931 diff -NurpP --minimal linux-2.6.25.11/arch/sparc/Kconfig linux-2.6.25.11-vs2.3.0.34.14/arch/sparc/Kconfig
 932 --- linux-2.6.25.11/arch/sparc/Kconfig  2008-04-17 12:05:30.000000000 -0400
 933 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/sparc/Kconfig    2008-04-19 15:14:52.000000000 -0400
 934 @@ -330,6 +330,8 @@ source "fs/Kconfig"
 935
 936  source "arch/sparc/Kconfig.debug"
 937
 938 +source "kernel/vserver/Kconfig"
 939 +
 940  source "security/Kconfig"
 941
 942  source "crypto/Kconfig"
 943 diff -NurpP --minimal linux-2.6.25.11/arch/sparc/kernel/ptrace.c linux-2.6.25.11-vs2.3.0.34.14/arch/sparc/kernel/ptrace.c
 944 --- linux-2.6.25.11/arch/sparc/kernel/ptrace.c  2008-07-15 22:05:16.000000000 -0400
 945 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/sparc/kernel/ptrace.c    2008-07-12 19:54:52.000000000 -0400
 946 @@ -21,6 +21,7 @@
 947  #include <linux/signal.h>
 948  #include <linux/regset.h>
 949  #include <linux/elf.h>
 950 +#include <linux/vs_base.h>
 951
 952  #include <asm/pgtable.h>
 953  #include <asm/system.h>
 954 diff -NurpP --minimal linux-2.6.25.11/arch/sparc/kernel/systbls.S linux-2.6.25.11-vs2.3.0.34.14/arch/sparc/kernel/systbls.S
 955 --- linux-2.6.25.11/arch/sparc/kernel/systbls.S 2008-04-17 12:05:30.000000000 -0400
 956 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/sparc/kernel/systbls.S   2008-04-19 15:14:52.000000000 -0400
 957 @@ -70,7 +70,7 @@ sys_call_table:
 958  /*250*/        .long sparc_mremap, sys_sysctl, sys_getsid, sys_fdatasync, sys_nfsservctl
 959  /*255*/        .long sys_sync_file_range, sys_clock_settime, sys_clock_gettime, sys_clock_getres, sys_clock_nanosleep
 960  /*260*/        .long sys_sched_getaffinity, sys_sched_setaffinity, sys_timer_settime, sys_timer_gettime, sys_timer_getoverrun
 961 -/*265*/        .long sys_timer_delete, sys_timer_create, sys_nis_syscall, sys_io_setup, sys_io_destroy
 962 +/*265*/        .long sys_timer_delete, sys_timer_create, sys_vserver, sys_io_setup, sys_io_destroy
 963  /*270*/        .long sys_io_submit, sys_io_cancel, sys_io_getevents, sys_mq_open, sys_mq_unlink
 964  /*275*/        .long sys_mq_timedsend, sys_mq_timedreceive, sys_mq_notify, sys_mq_getsetattr, sys_waitid
 965  /*280*/        .long sys_tee, sys_add_key, sys_request_key, sys_keyctl, sys_openat
 966 diff -NurpP --minimal linux-2.6.25.11/arch/sparc/kernel/traps.c linux-2.6.25.11-vs2.3.0.34.14/arch/sparc/kernel/traps.c
 967 --- linux-2.6.25.11/arch/sparc/kernel/traps.c   2008-04-17 11:31:25.000000000 -0400
 968 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/sparc/kernel/traps.c     2008-04-19 15:14:52.000000000 -0400
 969 @@ -99,7 +99,8 @@ void die_if_kernel(char *str, struct pt_
 970  "              /_| \\__/ |_\\\n"
 971  "                 \\__U_/\n");
 972
 973 -       printk("%s(%d): %s [#%d]\n", current->comm, task_pid_nr(current), str, ++die_counter);
 974 +       printk("%s(%d[#%u]): %s [#%d]\n", current->comm,
 975 +               task_pid_nr(current), current->xid, str, ++die_counter);
 976         show_regs(regs);
 977         add_taint(TAINT_DIE);
 978
 979 diff -NurpP --minimal linux-2.6.25.11/arch/sparc/mm/fault.c linux-2.6.25.11-vs2.3.0.34.14/arch/sparc/mm/fault.c
 980 --- linux-2.6.25.11/arch/sparc/mm/fault.c       2008-04-17 11:31:25.000000000 -0400
 981 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/sparc/mm/fault.c 2008-04-19 15:14:52.000000000 -0400
 982 @@ -367,7 +367,8 @@ no_context:
 983   */
 984  out_of_memory:
 985         up_read(&mm->mmap_sem);
 986 -       printk("VM: killing process %s\n", tsk->comm);
 987 +       printk("VM: killing process %s(%d:#%u)\n",
 988 +               tsk->comm, tsk->pid, tsk->xid);
 989         if (from_user)
 990                 do_group_exit(SIGKILL);
 991         goto no_context;
 992 diff -NurpP --minimal linux-2.6.25.11/arch/sparc64/Kconfig linux-2.6.25.11-vs2.3.0.34.14/arch/sparc64/Kconfig
 993 --- linux-2.6.25.11/arch/sparc64/Kconfig        2008-04-17 12:05:30.000000000 -0400
 994 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/sparc64/Kconfig  2008-04-19 15:14:52.000000000 -0400
 995 @@ -471,6 +471,8 @@ source "fs/Kconfig"
 996
 997  source "arch/sparc64/Kconfig.debug"
 998
 999 +source "kernel/vserver/Kconfig"
1000 +
1001  source "security/Kconfig"
1002
1003  source "crypto/Kconfig"
1004 diff -NurpP --minimal linux-2.6.25.11/arch/sparc64/kernel/binfmt_aout32.c linux-2.6.25.11-vs2.3.0.34.14/arch/sparc64/kernel/binfmt_aout32.c
1005 --- linux-2.6.25.11/arch/sparc64/kernel/binfmt_aout32.c 2008-04-17 12:05:30.000000000 -0400
1006 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/sparc64/kernel/binfmt_aout32.c   2008-04-19 15:14:52.000000000 -0400
1007 @@ -27,6 +27,7 @@
1008  #include <linux/binfmts.h>
1009  #include <linux/personality.h>
1010  #include <linux/init.h>
1011 +#include <linux/vs_memory.h>
1012
1013  #include <asm/system.h>
1014  #include <asm/uaccess.h>
1015 diff -NurpP --minimal linux-2.6.25.11/arch/sparc64/kernel/ptrace.c linux-2.6.25.11-vs2.3.0.34.14/arch/sparc64/kernel/ptrace.c
1016 --- linux-2.6.25.11/arch/sparc64/kernel/ptrace.c        2008-07-15 22:05:16.000000000 -0400
1017 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/sparc64/kernel/ptrace.c  2008-07-12 19:51:45.000000000 -0400
1018 @@ -25,6 +25,7 @@
1019  #include <linux/regset.h>
1020  #include <linux/compat.h>
1021  #include <linux/elf.h>
1022 +#include <linux/vs_base.h>
1023
1024  #include <asm/asi.h>
1025  #include <asm/pgtable.h>
1026 diff -NurpP --minimal linux-2.6.25.11/arch/sparc64/kernel/sys_sparc32.c linux-2.6.25.11-vs2.3.0.34.14/arch/sparc64/kernel/sys_sparc32.c
1027 --- linux-2.6.25.11/arch/sparc64/kernel/sys_sparc32.c   2008-07-15 22:05:16.000000000 -0400
1028 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/sparc64/kernel/sys_sparc32.c     2008-05-21 14:30:40.000000000 -0400
1029 @@ -722,7 +722,7 @@ asmlinkage long sys32_gettimeofday(struc
1030  {
1031         if (tv) {
1032                 struct timeval ktv;
1033 -               do_gettimeofday(&ktv);
1034 +               vx_gettimeofday(&ktv);
1035                 if (put_tv32(tv, &ktv))
1036                         return -EFAULT;
1037         }
1038 diff -NurpP --minimal linux-2.6.25.11/arch/sparc64/kernel/systbls.S linux-2.6.25.11-vs2.3.0.34.14/arch/sparc64/kernel/systbls.S
1039 --- linux-2.6.25.11/arch/sparc64/kernel/systbls.S       2008-04-17 12:05:30.000000000 -0400
1040 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/sparc64/kernel/systbls.S 2008-04-19 15:14:52.000000000 -0400
1041 @@ -71,7 +71,7 @@ sys_call_table32:
1042  /*250*/        .word sys32_mremap, sys32_sysctl, sys32_getsid, sys_fdatasync, sys32_nfsservctl
1043         .word sys32_sync_file_range, compat_sys_clock_settime, compat_sys_clock_gettime, compat_sys_clock_getres, sys32_clock_nanosleep
1044  /*260*/        .word compat_sys_sched_getaffinity, compat_sys_sched_setaffinity, sys32_timer_settime, compat_sys_timer_gettime, sys_timer_getoverrun
1045 -       .word sys_timer_delete, compat_sys_timer_create, sys_ni_syscall, compat_sys_io_setup, sys_io_destroy
1046 +       .word sys_timer_delete, compat_sys_timer_create, sys32_vserver, compat_sys_io_setup, sys_io_destroy
1047  /*270*/        .word sys32_io_submit, sys_io_cancel, compat_sys_io_getevents, sys32_mq_open, sys_mq_unlink
1048         .word compat_sys_mq_timedsend, compat_sys_mq_timedreceive, compat_sys_mq_notify, compat_sys_mq_getsetattr, compat_sys_waitid
1049  /*280*/        .word sys32_tee, sys_add_key, sys_request_key, sys_keyctl, compat_sys_openat
1050 @@ -144,7 +144,7 @@ sys_call_table:
1051  /*250*/        .word sys64_mremap, sys_sysctl, sys_getsid, sys_fdatasync, sys_nfsservctl
1052         .word sys_sync_file_range, sys_clock_settime, sys_clock_gettime, sys_clock_getres, sys_clock_nanosleep
1053  /*260*/        .word sys_sched_getaffinity, sys_sched_setaffinity, sys_timer_settime, sys_timer_gettime, sys_timer_getoverrun
1054 -       .word sys_timer_delete, sys_timer_create, sys_ni_syscall, sys_io_setup, sys_io_destroy
1055 +       .word sys_timer_delete, sys_timer_create, sys_vserver, sys_io_setup, sys_io_destroy
1056  /*270*/        .word sys_io_submit, sys_io_cancel, sys_io_getevents, sys_mq_open, sys_mq_unlink
1057         .word sys_mq_timedsend, sys_mq_timedreceive, sys_mq_notify, sys_mq_getsetattr, sys_waitid
1058  /*280*/        .word sys_tee, sys_add_key, sys_request_key, sys_keyctl, sys_openat
1059 diff -NurpP --minimal linux-2.6.25.11/arch/sparc64/kernel/traps.c linux-2.6.25.11-vs2.3.0.34.14/arch/sparc64/kernel/traps.c
1060 --- linux-2.6.25.11/arch/sparc64/kernel/traps.c 2008-04-17 12:05:30.000000000 -0400
1061 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/sparc64/kernel/traps.c   2008-04-19 15:14:52.000000000 -0400
1062 @@ -2183,7 +2183,8 @@ void die_if_kernel(char *str, struct pt_
1063  "              /_| \\__/ |_\\\n"
1064  "                 \\__U_/\n");
1065
1066 -       printk("%s(%d): %s [#%d]\n", current->comm, task_pid_nr(current), str, ++die_counter);
1067 +       printk("%s(%d[#%u]): %s [#%d]\n", current->comm,
1068 +               task_pid_nr(current), current->xid, str, ++die_counter);
1069         notify_die(DIE_OOPS, str, regs, 0, 255, SIGSEGV);
1070         __asm__ __volatile__("flushw");
1071         __show_regs(regs);
1072 diff -NurpP --minimal linux-2.6.25.11/arch/sparc64/mm/fault.c linux-2.6.25.11-vs2.3.0.34.14/arch/sparc64/mm/fault.c
1073 --- linux-2.6.25.11/arch/sparc64/mm/fault.c     2008-04-17 12:05:30.000000000 -0400
1074 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/sparc64/mm/fault.c       2008-04-19 15:14:52.000000000 -0400
1075 @@ -453,7 +453,8 @@ handle_kernel_fault:
1076  out_of_memory:
1077         insn = get_fault_insn(regs, insn);
1078         up_read(&mm->mmap_sem);
1079 -       printk("VM: killing process %s\n", current->comm);
1080 +       printk("VM: killing process %s(%d:#%u)\n",
1081 +               current->comm, current->pid, current->xid);
1082         if (!(regs->tstate & TSTATE_PRIV))
1083                 do_group_exit(SIGKILL);
1084         goto handle_kernel_fault;
1085 diff -NurpP --minimal linux-2.6.25.11/arch/sparc64/solaris/fs.c linux-2.6.25.11-vs2.3.0.34.14/arch/sparc64/solaris/fs.c
1086 --- linux-2.6.25.11/arch/sparc64/solaris/fs.c   2008-04-17 12:05:30.000000000 -0400
1087 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/sparc64/solaris/fs.c     2008-04-19 15:14:52.000000000 -0400
1088 @@ -368,7 +368,7 @@ static int report_statvfs(struct vfsmoun
1089                 int j = strlen (p);
1090
1091                 if (j > 15) j = 15;
1092 -               if (IS_RDONLY(inode)) i = 1;
1093 +               if (IS_RDONLY(inode) || MNT_IS_RDONLY(mnt)) i = 1;
1094                 if (mnt->mnt_flags & MNT_NOSUID) i |= 2;
1095                 if (!sysv_valid_dev(inode->i_sb->s_dev))
1096                         return -EOVERFLOW;
1097 @@ -404,7 +404,7 @@ static int report_statvfs64(struct vfsmo
1098                 int j = strlen (p);
1099
1100                 if (j > 15) j = 15;
1101 -               if (IS_RDONLY(inode)) i = 1;
1102 +               if (IS_RDONLY(inode) || MNT_IS_RDONLY(mnt)) i = 1;
1103                 if (mnt->mnt_flags & MNT_NOSUID) i |= 2;
1104                 if (!sysv_valid_dev(inode->i_sb->s_dev))
1105                         return -EOVERFLOW;
1106 diff -NurpP --minimal linux-2.6.25.11/arch/um/Kconfig linux-2.6.25.11-vs2.3.0.34.14/arch/um/Kconfig
1107 --- linux-2.6.25.11/arch/um/Kconfig     2008-04-17 12:05:30.000000000 -0400
1108 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/um/Kconfig       2008-04-19 15:14:52.000000000 -0400
1109 @@ -245,6 +245,8 @@ source "drivers/connector/Kconfig"
1110
1111  source "fs/Kconfig"
1112
1113 +source "kernel/vserver/Kconfig"
1114 +
1115  source "security/Kconfig"
1116
1117  source "crypto/Kconfig"
1118 diff -NurpP --minimal linux-2.6.25.11/arch/um/kernel/trap.c linux-2.6.25.11-vs2.3.0.34.14/arch/um/kernel/trap.c
1119 --- linux-2.6.25.11/arch/um/kernel/trap.c       2008-04-17 12:05:30.000000000 -0400
1120 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/um/kernel/trap.c 2008-04-19 15:14:52.000000000 -0400
1121 @@ -215,7 +215,8 @@ unsigned long segv(struct faultinfo fi,
1122                 current->thread.arch.faultinfo = fi;
1123                 force_sig_info(SIGBUS, &si, current);
1124         } else if (err == -ENOMEM) {
1125 -               printk(KERN_INFO "VM: killing process %s\n", current->comm);
1126 +               printk(KERN_INFO "VM: killing process %s(%d:#%u)\n",
1127 +                       current->comm, task_pid_nr(current), current->xid);
1128                 do_exit(SIGKILL);
1129         } else {
1130                 BUG_ON(err != -EFAULT);
1131 diff -NurpP --minimal linux-2.6.25.11/arch/v850/Kconfig linux-2.6.25.11-vs2.3.0.34.14/arch/v850/Kconfig
1132 --- linux-2.6.25.11/arch/v850/Kconfig   2008-04-17 12:05:30.000000000 -0400
1133 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/v850/Kconfig     2008-04-19 15:14:52.000000000 -0400
1134 @@ -344,6 +344,8 @@ source "drivers/usb/Kconfig"
1135
1136  source "arch/v850/Kconfig.debug"
1137
1138 +source "kernel/vserver/Kconfig"
1139 +
1140  source "security/Kconfig"
1141
1142  source "crypto/Kconfig"
1143 diff -NurpP --minimal linux-2.6.25.11/arch/v850/kernel/ptrace.c linux-2.6.25.11-vs2.3.0.34.14/arch/v850/kernel/ptrace.c
1144 --- linux-2.6.25.11/arch/v850/kernel/ptrace.c   2008-04-17 10:37:14.000000000 -0400
1145 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/v850/kernel/ptrace.c     2008-04-19 15:14:52.000000000 -0400
1146 @@ -23,6 +23,7 @@
1147  #include <linux/sched.h>
1148  #include <linux/ptrace.h>
1149  #include <linux/signal.h>
1150 +#include <linux/vs_base.h>
1151
1152  #include <asm/errno.h>
1153  #include <asm/ptrace.h>
1154 @@ -116,6 +117,9 @@ long arch_ptrace(struct task_struct *chi
1155  {
1156         int rval;
1157
1158 +       if (!vx_check(vx_task_xid(child), VS_WATCH_P | VS_IDENT))
1159 +               goto out;
1160 +
1161         switch (request) {
1162                 unsigned long val;
1163
1164 diff -NurpP --minimal linux-2.6.25.11/arch/x86/ia32/ia32entry.S linux-2.6.25.11-vs2.3.0.34.14/arch/x86/ia32/ia32entry.S
1165 --- linux-2.6.25.11/arch/x86/ia32/ia32entry.S   2008-04-17 12:05:30.000000000 -0400
1166 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/x86/ia32/ia32entry.S     2008-04-22 20:07:49.000000000 -0400
1167 @@ -673,7 +673,7 @@ ia32_sys_call_table:
1168         .quad sys_tgkill                /* 270 */
1169         .quad compat_sys_utimes
1170         .quad sys32_fadvise64_64
1171 -       .quad quiet_ni_syscall  /* sys_vserver */
1172 +       .quad sys32_vserver
1173         .quad sys_mbind
1174         .quad compat_sys_get_mempolicy  /* 275 */
1175         .quad sys_set_mempolicy
1176 diff -NurpP --minimal linux-2.6.25.11/arch/x86/Kconfig linux-2.6.25.11-vs2.3.0.34.14/arch/x86/Kconfig
1177 --- linux-2.6.25.11/arch/x86/Kconfig    2008-07-15 22:05:16.000000000 -0400
1178 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/x86/Kconfig      2008-06-23 15:22:21.000000000 -0400
1179 @@ -1614,6 +1614,8 @@ source "fs/Kconfig"
1180
1181  source "arch/x86/Kconfig.debug"
1182
1183 +source "kernel/vserver/Kconfig"
1184 +
1185  source "security/Kconfig"
1186
1187  source "crypto/Kconfig"
1188 diff -NurpP --minimal linux-2.6.25.11/arch/x86/kernel/syscall_table_32.S linux-2.6.25.11-vs2.3.0.34.14/arch/x86/kernel/syscall_table_32.S
1189 --- linux-2.6.25.11/arch/x86/kernel/syscall_table_32.S  2008-04-17 12:05:30.000000000 -0400
1190 +++ linux-2.6.25.11-vs2.3.0.34.14/arch/x86/kernel/syscall_table_32.S    2008-04-19 15:14:52.000000000 -0400
1191 @@ -272,7 +272,7 @@ ENTRY(sys_call_table)
1192         .long sys_tgkill        /* 270 */
1193         .long sys_utimes
1194         .long sys_fadvise64_64
1195 -       .long sys_ni_syscall    /* sys_vserver */
1196 +       .long sys_vserver
1197         .long sys_mbind
1198         .long sys_get_mempolicy
1199         .long sys_set_mempolicy
1200 diff -NurpP --minimal linux-2.6.25.11/Documentation/vserver/debug.txt linux-2.6.25.11-vs2.3.0.34.14/Documentation/vserver/debug.txt
1201 --- linux-2.6.25.11/Documentation/vserver/debug.txt     1969-12-31 19:00:00.000000000 -0500
1202 +++ linux-2.6.25.11-vs2.3.0.34.14/Documentation/vserver/debug.txt       2008-04-19 15:14:51.000000000 -0400
1203 @@ -0,0 +1,154 @@
1204 +
1205 +debug_cvirt:
1206 +
1207 + 2   4 "vx_map_tgid: %p/%llx: %d -> %d"
1208 +       "vx_rmap_tgid: %p/%llx: %d -> %d"
1209 +
1210 +debug_dlim:
1211 +
1212 + 0   1 "ALLOC (%p,#%d)%c inode (%d)"
1213 +       "FREE  (%p,#%d)%c inode"
1214 + 1   2 "ALLOC (%p,#%d)%c %lld bytes (%d)"
1215 +       "FREE  (%p,#%d)%c %lld bytes"
1216 + 2   4 "ADJUST: %lld,%lld on %ld,%ld [mult=%d]"
1217 + 3   8 "ext3_has_free_blocks(%p): %lu<%lu+1, %c, %u!=%u r=%d"
1218 +       "ext3_has_free_blocks(%p): free=%lu, root=%lu"
1219 +       "rcu_free_dl_info(%p)"
1220 + 4  10 "alloc_dl_info(%p,%d) = %p"
1221 +       "dealloc_dl_info(%p)"
1222 +       "get_dl_info(%p[#%d.%d])"
1223 +       "put_dl_info(%p[#%d.%d])"
1224 + 5  20 "alloc_dl_info(%p,%d)*"
1225 + 6  40 "__hash_dl_info: %p[#%d]"
1226 +       "__unhash_dl_info: %p[#%d]"
1227 + 7  80 "locate_dl_info(%p,#%d) = %p"
1228 +
1229 +debug_misc:
1230 +
1231 + 0   1 "destroy_dqhash: %p [#0x%08x] c=%d"
1232 +       "new_dqhash: %p [#0x%08x]"
1233 +       "vroot[%d]_clr_dev: dev=%p[%lu,%d:%d]"
1234 +       "vroot[%d]_get_real_bdev: dev=%p[%lu,%d:%d]"
1235 +       "vroot[%d]_set_dev: dev=%p[%lu,%d:%d]"
1236 +       "vroot_get_real_bdev not set"
1237 + 1   2 "cow_break_link(»%s«)"
1238 +       "temp copy »%s«"
1239 + 2   4 "dentry_open(new): %p"
1240 +       "dentry_open(old): %p"
1241 +       "lookup_create(new): %p"
1242 +       "old path »%s«"
1243 +       "path_lookup(old): %d"
1244 +       "vfs_create(new): %d"
1245 +       "vfs_rename: %d"
1246 +       "vfs_sendfile: %d"
1247 + 3   8 "fput(new_file=%p[#%d])"
1248 +       "fput(old_file=%p[#%d])"
1249 + 4  10 "vx_info_kill(%p[#%d],%d,%d) = %d"
1250 +       "vx_info_kill(%p[#%d],%d,%d)*"
1251 + 5  20 "vs_reboot(%p[#%d],%d)"
1252 + 6  40 "dropping task %p[#%u,%u] for %p[#%u,%u]"
1253 +
1254 +debug_net:
1255 +
1256 + 2   4 "nx_addr_conflict(%p,%p) %d.%d,%d.%d"
1257 + 3   8 "inet_bind(%p) %d.%d.%d.%d, %d.%d.%d.%d, %d.%d.%d.%d"
1258 +       "inet_bind(%p)* %p,%p;%lx %d.%d.%d.%d"
1259 + 4  10 "ip_route_connect(%p) %p,%p;%lx"
1260 + 5  20 "__addr_in_socket(%p,%d.%d.%d.%d) %p:%d.%d.%d.%d %p;%lx"
1261 + 6  40 "sk,egf: %p [#%d] (from %d)"
1262 +       "sk,egn: %p [#%d] (from %d)"
1263 +       "sk,req: %p [#%d] (from %d)"
1264 +       "sk: %p [#%d] (from %d)"
1265 +       "tw: %p [#%d] (from %d)"
1266 + 7  80 "__sock_recvmsg: %p[%p,%p,%p;%d]:%d/%d"
1267 +       "__sock_sendmsg: %p[%p,%p,%p;%d]:%d/%d"
1268 +
1269 +debug_nid:
1270 +
1271 + 0   1 "__lookup_nx_info(#%u): %p[#%u]"
1272 +       "alloc_nx_info(%d) = %p"
1273 +       "create_nx_info(%d) (dynamic rejected)"
1274 +       "create_nx_info(%d) = %p (already there)"
1275 +       "create_nx_info(%d) = %p (new)"
1276 +       "dealloc_nx_info(%p)"
1277 + 1   2 "alloc_nx_info(%d)*"
1278 +       "create_nx_info(%d)*"
1279 + 2   4 "get_nx_info(%p[#%d.%d])"
1280 +       "put_nx_info(%p[#%d.%d])"
1281 + 3   8 "claim_nx_info(%p[#%d.%d.%d]) %p"
1282 +       "clr_nx_info(%p[#%d.%d])"
1283 +       "init_nx_info(%p[#%d.%d])"
1284 +       "release_nx_info(%p[#%d.%d.%d]) %p"
1285 +       "set_nx_info(%p[#%d.%d])"
1286 + 4  10 "__hash_nx_info: %p[#%d]"
1287 +       "__nx_dynamic_id: [#%d]"
1288 +       "__unhash_nx_info: %p[#%d.%d.%d]"
1289 + 5  20 "moved task %p into nxi:%p[#%d]"
1290 +       "nx_migrate_task(%p,%p[#%d.%d.%d])"
1291 +       "task_get_nx_info(%p)"
1292 + 6  40 "nx_clear_persistent(%p[#%d])"
1293 +
1294 +debug_quota:
1295 +
1296 + 0   1 "quota_sync_dqh(%p,%d) discard inode %p"
1297 + 1   2 "quota_sync_dqh(%p,%d)"
1298 +       "sync_dquots(%p,%d)"
1299 +       "sync_dquots_dqh(%p,%d)"
1300 + 3   8 "do_quotactl(%p,%d,cmd=%d,id=%d,%p)"
1301 +
1302 +debug_switch:
1303 +
1304 + 0   1 "vc: VCMD_%02d_%d[%d], %d,%p [%d,%d,%x,%x]"
1305 + 1   2 "vc: VCMD_%02d_%d[%d] = %08lx(%ld) [%d,%d]"
1306 + 4  10 "%s: (%s %s) returned %s with %d"
1307 +
1308 +debug_tag:
1309 +
1310 + 7  80 "dx_parse_tag(»%s«): %d:#%d"
1311 +       "dx_propagate_tag(%p[#%lu.%d]): %d,%d"
1312 +
1313 +debug_xid:
1314 +
1315 + 0   1 "__lookup_vx_info(#%u): %p[#%u]"
1316 +       "alloc_vx_info(%d) = %p"
1317 +       "alloc_vx_info(%d)*"
1318 +       "create_vx_info(%d) (dynamic rejected)"
1319 +       "create_vx_info(%d) = %p (already there)"
1320 +       "create_vx_info(%d) = %p (new)"
1321 +       "dealloc_vx_info(%p)"
1322 +       "loc_vx_info(%d) = %p (found)"
1323 +       "loc_vx_info(%d) = %p (new)"
1324 +       "loc_vx_info(%d) = %p (not available)"
1325 + 1   2 "create_vx_info(%d)*"
1326 +       "loc_vx_info(%d)*"
1327 + 2   4 "get_vx_info(%p[#%d.%d])"
1328 +       "put_vx_info(%p[#%d.%d])"
1329 + 3   8 "claim_vx_info(%p[#%d.%d.%d]) %p"
1330 +       "clr_vx_info(%p[#%d.%d])"
1331 +       "init_vx_info(%p[#%d.%d])"
1332 +       "release_vx_info(%p[#%d.%d.%d]) %p"
1333 +       "set_vx_info(%p[#%d.%d])"
1334 + 4  10 "__hash_vx_info: %p[#%d]"
1335 +       "__unhash_vx_info: %p[#%d.%d.%d]"
1336 +       "__vx_dynamic_id: [#%d]"
1337 + 5  20 "enter_vx_info(%p[#%d],%p) %p[#%d,%p]"
1338 +       "leave_vx_info(%p[#%d,%p]) %p[#%d,%p]"
1339 +       "moved task %p into vxi:%p[#%d]"
1340 +       "task_get_vx_info(%p)"
1341 +       "vx_migrate_task(%p,%p[#%d.%d])"
1342 + 6  40 "vx_clear_persistent(%p[#%d])"
1343 +       "vx_exit_init(%p[#%d],%p[#%d,%d,%d])"
1344 +       "vx_set_init(%p[#%d],%p[#%d,%d,%d])"
1345 +       "vx_set_persistent(%p[#%d])"
1346 +       "vx_set_reaper(%p[#%d],%p[#%d,%d])"
1347 + 7  80 "vx_child_reaper(%p[#%u,%u]) = %p[#%u,%u]"
1348 +
1349 +
1350 +debug_limit:
1351 +
1352 + n 2^n "vx_acc_cres[%5d,%s,%2d]: %5d%s"
1353 +       "vx_cres_avail[%5d,%s,%2d]: %5ld > %5d + %5d"
1354 +
1355 + m 2^m "vx_acc_page[%5d,%s,%2d]: %5d%s"
1356 +       "vx_acc_pages[%5d,%s,%2d]: %5d += %5d"
1357 +       "vx_pages_avail[%5d,%s,%2d]: %5ld > %5d + %5d"
1358 diff -NurpP --minimal linux-2.6.25.11/drivers/block/Kconfig linux-2.6.25.11-vs2.3.0.34.14/drivers/block/Kconfig
1359 --- linux-2.6.25.11/drivers/block/Kconfig       2008-04-17 12:05:31.000000000 -0400
1360 +++ linux-2.6.25.11-vs2.3.0.34.14/drivers/block/Kconfig 2008-04-19 15:14:52.000000000 -0400
1361 @@ -263,6 +263,13 @@ config BLK_DEV_CRYPTOLOOP
1362           instead, which can be configured to be on-disk compatible with the
1363           cryptoloop device.
1364
1365 +config BLK_DEV_VROOT
1366 +       tristate "Virtual Root device support"
1367 +       depends on QUOTACTL
1368 +       ---help---
1369 +         Saying Y here will allow you to use quota/fs ioctls on a shared
1370 +         partition within a virtual server without compromising security.
1371 +
1372  config BLK_DEV_NBD
1373         tristate "Network block device support"
1374         depends on NET
1375 diff -NurpP --minimal linux-2.6.25.11/drivers/block/loop.c linux-2.6.25.11-vs2.3.0.34.14/drivers/block/loop.c
1376 --- linux-2.6.25.11/drivers/block/loop.c        2008-04-17 12:05:32.000000000 -0400
1377 +++ linux-2.6.25.11-vs2.3.0.34.14/drivers/block/loop.c  2008-04-19 15:14:52.000000000 -0400
1378 @@ -76,6 +76,7 @@
1379  #include <linux/gfp.h>
1380  #include <linux/kthread.h>
1381  #include <linux/splice.h>
1382 +#include <linux/vs_context.h>
1383
1384  #include <asm/uaccess.h>
1385
1386 @@ -789,6 +790,7 @@ static int loop_set_fd(struct loop_devic
1387         lo->lo_blocksize = lo_blocksize;
1388         lo->lo_device = bdev;
1389         lo->lo_flags = lo_flags;
1390 +       lo->lo_xid = vx_current_xid();
1391         lo->lo_backing_file = file;
1392         lo->transfer = transfer_none;
1393         lo->ioctl = NULL;
1394 @@ -908,6 +910,7 @@ static int loop_clr_fd(struct loop_devic
1395         lo->lo_encrypt_key_size = 0;
1396         lo->lo_flags = 0;
1397         lo->lo_thread = NULL;
1398 +       lo->lo_xid = 0;
1399         memset(lo->lo_encrypt_key, 0, LO_KEY_SIZE);
1400         memset(lo->lo_crypt_name, 0, LO_NAME_SIZE);
1401         memset(lo->lo_file_name, 0, LO_NAME_SIZE);
1402 @@ -929,7 +932,7 @@ loop_set_status(struct loop_device *lo,
1403         struct loop_func_table *xfer;
1404
1405         if (lo->lo_encrypt_key_size && lo->lo_key_owner != current->uid &&
1406 -           !capable(CAP_SYS_ADMIN))
1407 +           !vx_capable(CAP_SYS_ADMIN, VXC_ADMIN_CLOOP))
1408                 return -EPERM;
1409         if (lo->lo_state != Lo_bound)
1410                 return -ENXIO;
1411 @@ -1013,7 +1016,8 @@ loop_get_status(struct loop_device *lo,
1412         memcpy(info->lo_crypt_name, lo->lo_crypt_name, LO_NAME_SIZE);
1413         info->lo_encrypt_type =
1414                 lo->lo_encryption ? lo->lo_encryption->number : 0;
1415 -       if (lo->lo_encrypt_key_size && capable(CAP_SYS_ADMIN)) {
1416 +       if (lo->lo_encrypt_key_size &&
1417 +               vx_capable(CAP_SYS_ADMIN, VXC_ADMIN_CLOOP)) {
1418                 info->lo_encrypt_key_size = lo->lo_encrypt_key_size;
1419                 memcpy(info->lo_encrypt_key, lo->lo_encrypt_key,
1420                        lo->lo_encrypt_key_size);
1421 @@ -1322,6 +1326,9 @@ static int lo_open(struct inode *inode,
1422  {
1423         struct loop_device *lo = inode->i_bdev->bd_disk->private_data;
1424
1425 +       if (!vx_check(lo->lo_xid, VS_IDENT|VS_HOSTID|VS_ADMIN_P))
1426 +               return -EACCES;
1427 +
1428         mutex_lock(&lo->lo_ctl_mutex);
1429         lo->lo_refcnt++;
1430         mutex_unlock(&lo->lo_ctl_mutex);
1431 diff -NurpP --minimal linux-2.6.25.11/drivers/block/Makefile linux-2.6.25.11-vs2.3.0.34.14/drivers/block/Makefile
1432 --- linux-2.6.25.11/drivers/block/Makefile      2008-04-17 12:05:31.000000000 -0400
1433 +++ linux-2.6.25.11-vs2.3.0.34.14/drivers/block/Makefile        2008-04-19 15:14:52.000000000 -0400
1434 @@ -29,5 +29,6 @@ obj-$(CONFIG_VIRTIO_BLK)      += virtio_blk.o
1435  obj-$(CONFIG_VIODASD)          += viodasd.o
1436  obj-$(CONFIG_BLK_DEV_SX8)      += sx8.o
1437  obj-$(CONFIG_BLK_DEV_UB)       += ub.o
1438 +obj-$(CONFIG_BLK_DEV_VROOT)    += vroot.o
1439
1440  obj-$(CONFIG_XEN_BLKDEV_FRONTEND)      += xen-blkfront.o
1441 diff -NurpP --minimal linux-2.6.25.11/drivers/block/vroot.c linux-2.6.25.11-vs2.3.0.34.14/drivers/block/vroot.c
1442 --- linux-2.6.25.11/drivers/block/vroot.c       1969-12-31 19:00:00.000000000 -0500
1443 +++ linux-2.6.25.11-vs2.3.0.34.14/drivers/block/vroot.c 2008-04-19 15:14:52.000000000 -0400
1444 @@ -0,0 +1,280 @@
1445 +/*
1446 + *  linux/drivers/block/vroot.c
1447 + *
1448 + *  written by Herbert Pötzl, 9/11/2002
1449 + *  ported to 2.6.10 by Herbert Pötzl, 30/12/2004
1450 + *
1451 + *  based on the loop.c code by Theodore Ts'o.
1452 + *
1453 + * Copyright (C) 2002-2007 by Herbert Pötzl.
1454 + * Redistribution of this file is permitted under the
1455 + * GNU General Public License.
1456 + *
1457 + */
1458 +
1459 +#include <linux/module.h>
1460 +#include <linux/moduleparam.h>
1461 +#include <linux/file.h>
1462 +#include <linux/major.h>
1463 +#include <linux/blkdev.h>
1464 +
1465 +#include <linux/vroot.h>
1466 +#include <linux/vs_context.h>
1467 +
1468 +
1469 +static int max_vroot = 8;
1470 +
1471 +static struct vroot_device *vroot_dev;
1472 +static struct gendisk **disks;
1473 +
1474 +
1475 +static int vroot_set_dev(
1476 +       struct vroot_device *vr,
1477 +       struct file *vr_file,
1478 +       struct block_device *bdev,
1479 +       unsigned int arg)
1480 +{
1481 +       struct block_device *real_bdev;
1482 +       struct file *file;
1483 +       struct inode *inode;
1484 +       int error;
1485 +
1486 +       error = -EBUSY;
1487 +       if (vr->vr_state != Vr_unbound)
1488 +               goto out;
1489 +
1490 +       error = -EBADF;
1491 +       file = fget(arg);
1492 +       if (!file)
1493 +               goto out;
1494 +
1495 +       error = -EINVAL;
1496 +       inode = file->f_dentry->d_inode;
1497 +
1498 +
1499 +       if (S_ISBLK(inode->i_mode)) {
1500 +               real_bdev = inode->i_bdev;
1501 +               vr->vr_device = real_bdev;
1502 +               __iget(real_bdev->bd_inode);
1503 +       } else
1504 +               goto out_fput;
1505 +
1506 +       vxdprintk(VXD_CBIT(misc, 0),
1507 +               "vroot[%d]_set_dev: dev=" VXF_DEV,
1508 +               vr->vr_number, VXD_DEV(real_bdev));
1509 +
1510 +       vr->vr_state = Vr_bound;
1511 +       error = 0;
1512 +
1513 + out_fput:
1514 +       fput(file);
1515 + out:
1516 +       return error;
1517 +}
1518 +
1519 +static int vroot_clr_dev(
1520 +       struct vroot_device *vr,
1521 +       struct file *vr_file,
1522 +       struct block_device *bdev)
1523 +{
1524 +       struct block_device *real_bdev;
1525 +
1526 +       if (vr->vr_state != Vr_bound)
1527 +               return -ENXIO;
1528 +       if (vr->vr_refcnt > 1)  /* we needed one fd for the ioctl */
1529 +               return -EBUSY;
1530 +
1531 +       real_bdev = vr->vr_device;
1532 +
1533 +       vxdprintk(VXD_CBIT(misc, 0),
1534 +               "vroot[%d]_clr_dev: dev=" VXF_DEV,
1535 +               vr->vr_number, VXD_DEV(real_bdev));
1536 +
1537 +       bdput(real_bdev);
1538 +       vr->vr_state = Vr_unbound;
1539 +       vr->vr_device = NULL;
1540 +       return 0;
1541 +}
1542 +
1543 +
1544 +static int vr_ioctl(struct inode *inode, struct file *file,
1545 +       unsigned int cmd, unsigned long arg)
1546 +{
1547 +       struct vroot_device *vr = inode->i_bdev->bd_disk->private_data;
1548 +       int err;
1549 +
1550 +       down(&vr->vr_ctl_mutex);
1551 +       switch (cmd) {
1552 +       case VROOT_SET_DEV:
1553 +               err = vroot_set_dev(vr, file, inode->i_bdev, arg);
1554 +               break;
1555 +       case VROOT_CLR_DEV:
1556 +               err = vroot_clr_dev(vr, file, inode->i_bdev);
1557 +               break;
1558 +       default:
1559 +               err = -EINVAL;
1560 +               break;
1561 +       }
1562 +       up(&vr->vr_ctl_mutex);
1563 +       return err;
1564 +}
1565 +
1566 +static int vr_open(struct inode *inode, struct file *file)
1567 +{
1568 +       struct vroot_device *vr = inode->i_bdev->bd_disk->private_data;
1569 +
1570 +       down(&vr->vr_ctl_mutex);
1571 +       vr->vr_refcnt++;
1572 +       up(&vr->vr_ctl_mutex);
1573 +       return 0;
1574 +}
1575 +
1576 +static int vr_release(struct inode *inode, struct file *file)
1577 +{
1578 +       struct vroot_device *vr = inode->i_bdev->bd_disk->private_data;
1579 +
1580 +       down(&vr->vr_ctl_mutex);
1581 +       --vr->vr_refcnt;
1582 +       up(&vr->vr_ctl_mutex);
1583 +       return 0;
1584 +}
1585 +
1586 +static struct block_device_operations vr_fops = {
1587 +       .owner =        THIS_MODULE,
1588 +       .open =         vr_open,
1589 +       .release =      vr_release,
1590 +       .ioctl =        vr_ioctl,
1591 +};
1592 +
1593 +struct block_device *__vroot_get_real_bdev(struct block_device *bdev)
1594 +{
1595 +       struct inode *inode = bdev->bd_inode;
1596 +       struct vroot_device *vr;
1597 +       struct block_device *real_bdev;
1598 +       int minor = iminor(inode);
1599 +
1600 +       vr = &vroot_dev[minor];
1601 +       real_bdev = vr->vr_device;
1602 +
1603 +       vxdprintk(VXD_CBIT(misc, 0),
1604 +               "vroot[%d]_get_real_bdev: dev=" VXF_DEV,
1605 +               vr->vr_number, VXD_DEV(real_bdev));
1606 +
1607 +       if (vr->vr_state != Vr_bound)
1608 +               return ERR_PTR(-ENXIO);
1609 +
1610 +       __iget(real_bdev->bd_inode);
1611 +       return real_bdev;
1612 +}
1613 +
1614 +/*
1615 + * And now the modules code and kernel interface.
1616 + */
1617 +
1618 +module_param(max_vroot, int, 0);
1619 +
1620 +MODULE_PARM_DESC(max_vroot, "Maximum number of vroot devices (1-256)");
1621 +MODULE_LICENSE("GPL");
1622 +MODULE_ALIAS_BLOCKDEV_MAJOR(VROOT_MAJOR);
1623 +
1624 +MODULE_AUTHOR ("Herbert Pötzl");
1625 +MODULE_DESCRIPTION ("Virtual Root Device Mapper");
1626 +
1627 +
1628 +int __init vroot_init(void)
1629 +{
1630 +       int err, i;
1631 +
1632 +       if (max_vroot < 1 || max_vroot > 256) {
1633 +               max_vroot = MAX_VROOT_DEFAULT;
1634 +               printk(KERN_WARNING "vroot: invalid max_vroot "
1635 +                       "(must be between 1 and 256), "
1636 +                       "using default (%d)\n", max_vroot);
1637 +       }
1638 +
1639 +       if (register_blkdev(VROOT_MAJOR, "vroot"))
1640 +               return -EIO;
1641 +
1642 +       err = -ENOMEM;
1643 +       vroot_dev = kmalloc(max_vroot * sizeof(struct vroot_device), GFP_KERNEL);
1644 +       if (!vroot_dev)
1645 +               goto out_mem1;
1646 +       memset(vroot_dev, 0, max_vroot * sizeof(struct vroot_device));
1647 +
1648 +       disks = kmalloc(max_vroot * sizeof(struct gendisk *), GFP_KERNEL);
1649 +       if (!disks)
1650 +               goto out_mem2;
1651 +
1652 +       for (i = 0; i < max_vroot; i++) {
1653 +               disks[i] = alloc_disk(1);
1654 +               if (!disks[i])
1655 +                       goto out_mem3;
1656 +       }
1657 +
1658 +       for (i = 0; i < max_vroot; i++) {
1659 +               struct vroot_device *vr = &vroot_dev[i];
1660 +               struct gendisk *disk = disks[i];
1661 +
1662 +               memset(vr, 0, sizeof(*vr));
1663 +               init_MUTEX(&vr->vr_ctl_mutex);
1664 +               vr->vr_number = i;
1665 +               disk->major = VROOT_MAJOR;
1666 +               disk->first_minor = i;
1667 +               disk->fops = &vr_fops;
1668 +               sprintf(disk->disk_name, "vroot%d", i);
1669 +               disk->private_data = vr;
1670 +       }
1671 +
1672 +       err = register_vroot_grb(&__vroot_get_real_bdev);
1673 +       if (err)
1674 +               goto out_mem3;
1675 +
1676 +       for (i = 0; i < max_vroot; i++)
1677 +               add_disk(disks[i]);
1678 +       printk(KERN_INFO "vroot: loaded (max %d devices)\n", max_vroot);
1679 +       return 0;
1680 +
1681 +out_mem3:
1682 +       while (i--)
1683 +               put_disk(disks[i]);
1684 +       kfree(disks);
1685 +out_mem2:
1686 +       kfree(vroot_dev);
1687 +out_mem1:
1688 +       unregister_blkdev(VROOT_MAJOR, "vroot");
1689 +       printk(KERN_ERR "vroot: ran out of memory\n");
1690 +       return err;
1691 +}
1692 +
1693 +void vroot_exit(void)
1694 +{
1695 +       int i;
1696 +
1697 +       if (unregister_vroot_grb(&__vroot_get_real_bdev))
1698 +               printk(KERN_WARNING "vroot: cannot unregister grb\n");
1699 +
1700 +       for (i = 0; i < max_vroot; i++) {
1701 +               del_gendisk(disks[i]);
1702 +               put_disk(disks[i]);
1703 +       }
1704 +       unregister_blkdev(VROOT_MAJOR, "vroot");
1705 +
1706 +       kfree(disks);
1707 +       kfree(vroot_dev);
1708 +}
1709 +
1710 +module_init(vroot_init);
1711 +module_exit(vroot_exit);
1712 +
1713 +#ifndef MODULE
1714 +
1715 +static int __init max_vroot_setup(char *str)
1716 +{
1717 +       max_vroot = simple_strtol(str, NULL, 0);
1718 +       return 1;
1719 +}
1720 +
1721 +__setup("max_vroot=", max_vroot_setup);
1722 +
1723 +#endif
1724 +
1725 diff -NurpP --minimal linux-2.6.25.11/drivers/char/sysrq.c linux-2.6.25.11-vs2.3.0.34.14/drivers/char/sysrq.c
1726 --- linux-2.6.25.11/drivers/char/sysrq.c        2008-04-17 11:31:27.000000000 -0400
1727 +++ linux-2.6.25.11-vs2.3.0.34.14/drivers/char/sysrq.c  2008-04-19 15:14:52.000000000 -0400
1728 @@ -37,6 +37,7 @@
1729  #include <linux/irq.h>
1730  #include <linux/hrtimer.h>
1731  #include <linux/oom.h>
1732 +#include <linux/vserver/debug.h>
1733
1734  #include <asm/ptrace.h>
1735  #include <asm/irq_regs.h>
1736 @@ -310,6 +311,21 @@ static struct sysrq_key_op sysrq_unrt_op
1737         .enable_mask    = SYSRQ_ENABLE_RTNICE,
1738  };
1739
1740 +
1741 +#ifdef CONFIG_VSERVER_DEBUG
1742 +static void sysrq_handle_vxinfo(int key, struct tty_struct *tty)
1743 +{
1744 +       dump_vx_info_inactive((key == 'x')?0:1);
1745 +}
1746 +
1747 +static struct sysrq_key_op sysrq_showvxinfo_op = {
1748 +       .handler        = sysrq_handle_vxinfo,
1749 +       .help_msg       = "conteXt",
1750 +       .action_msg     = "Show Context Info",
1751 +       .enable_mask    = SYSRQ_ENABLE_DUMP,
1752 +};
1753 +#endif
1754 +
1755  /* Key Operations table and lock */
1756  static DEFINE_SPINLOCK(sysrq_key_table_lock);
1757
1758 @@ -358,7 +374,11 @@ static struct sysrq_key_op *sysrq_key_ta
1759         /* x: May be registered on ppc/powerpc for xmon */
1760         NULL,                           /* x */
1761         NULL,                           /* y */
1762 -       NULL                            /* z */
1763 +#ifdef CONFIG_VSERVER_DEBUG
1764 +       &sysrq_showvxinfo_op,           /* z */
1765 +#else
1766 +       NULL,                           /* z */
1767 +#endif
1768  };
1769
1770  /* key2index calculation, -1 on invalid index */
1771 @@ -370,6 +390,8 @@ static int sysrq_key_table_key2index(int
1772                 retval = key - '0';
1773         else if ((key >= 'a') && (key <= 'z'))
1774                 retval = key + 10 - 'a';
1775 +       else if ((key >= 'A') && (key <= 'Z'))
1776 +               retval = key + 10 - 'A';
1777         else
1778                 retval = -1;
1779         return retval;
1780 diff -NurpP --minimal linux-2.6.25.11/drivers/char/tty_io.c linux-2.6.25.11-vs2.3.0.34.14/drivers/char/tty_io.c
1781 --- linux-2.6.25.11/drivers/char/tty_io.c       2008-04-17 12:05:32.000000000 -0400
1782 +++ linux-2.6.25.11-vs2.3.0.34.14/drivers/char/tty_io.c 2008-07-11 16:57:01.000000000 -0400
1783 @@ -105,6 +105,7 @@
1784
1785  #include <linux/kmod.h>
1786  #include <linux/nsproxy.h>
1787 +#include <linux/vs_pid.h>
1788
1789  #undef TTY_DEBUG_HANGUP
1790
1791 @@ -3142,6 +3143,7 @@ static int tiocspgrp(struct tty_struct *
1792                 return -ENOTTY;
1793         if (get_user(pgrp_nr, p))
1794                 return -EFAULT;
1795 +       pgrp_nr = vx_rmap_pid(pgrp_nr);
1796         if (pgrp_nr < 0)
1797                 return -EINVAL;
1798         rcu_read_lock();
1799 diff -NurpP --minimal linux-2.6.25.11/drivers/infiniband/hw/ipath/ipath_user_pages.c linux-2.6.25.11-vs2.3.0.34.14/drivers/infiniband/hw/ipath/ipath_user_pages.c
1800 --- linux-2.6.25.11/drivers/infiniband/hw/ipath/ipath_user_pages.c      2008-04-17 10:37:17.000000000 -0400
1801 +++ linux-2.6.25.11-vs2.3.0.34.14/drivers/infiniband/hw/ipath/ipath_user_pages.c        2008-04-19 15:14:52.000000000 -0400
1802 @@ -33,6 +33,7 @@
1803
1804  #include <linux/mm.h>
1805  #include <linux/device.h>
1806 +#include <linux/vs_memory.h>
1807
1808  #include "ipath_kernel.h"
1809
1810 @@ -61,7 +62,8 @@ static int __get_user_pages(unsigned lon
1811         lock_limit = current->signal->rlim[RLIMIT_MEMLOCK].rlim_cur >>
1812                 PAGE_SHIFT;
1813
1814 -       if (num_pages > lock_limit) {
1815 +       if (num_pages > lock_limit ||
1816 +               !vx_vmlocked_avail(current->mm, num_pages)) {
1817                 ret = -ENOMEM;
1818                 goto bail;
1819         }
1820 @@ -78,7 +80,7 @@ static int __get_user_pages(unsigned lon
1821                         goto bail_release;
1822         }
1823
1824 -       current->mm->locked_vm += num_pages;
1825 +       vx_vmlocked_add(current->mm, num_pages);
1826
1827         ret = 0;
1828         goto bail;
1829 @@ -177,7 +179,7 @@ void ipath_release_user_pages(struct pag
1830
1831         __ipath_release_user_pages(p, num_pages, 1);
1832
1833 -       current->mm->locked_vm -= num_pages;
1834 +       vx_vmlocked_sub(current->mm, num_pages);
1835
1836         up_write(&current->mm->mmap_sem);
1837  }
1838 @@ -194,7 +196,7 @@ static void user_pages_account(struct wo
1839                 container_of(_work, struct ipath_user_pages_work, work);
1840
1841         down_write(&work->mm->mmap_sem);
1842 -       work->mm->locked_vm -= work->num_pages;
1843 +       vx_vmlocked_sub(work->mm, work->num_pages);
1844         up_write(&work->mm->mmap_sem);
1845         mmput(work->mm);
1846         kfree(work);
1847 diff -NurpP --minimal linux-2.6.25.11/drivers/md/dm.c linux-2.6.25.11-vs2.3.0.34.14/drivers/md/dm.c
1848 --- linux-2.6.25.11/drivers/md/dm.c     2008-04-17 12:05:33.000000000 -0400
1849 +++ linux-2.6.25.11-vs2.3.0.34.14/drivers/md/dm.c       2008-04-19 15:14:52.000000000 -0400
1850 @@ -22,6 +22,7 @@
1851  #include <linux/hdreg.h>
1852  #include <linux/blktrace_api.h>
1853  #include <linux/smp_lock.h>
1854 +#include <linux/vs_base.h>
1855
1856  #define DM_MSG_PREFIX "core"
1857
1858 @@ -91,6 +92,7 @@ struct mapped_device {
1859         rwlock_t map_lock;
1860         atomic_t holders;
1861         atomic_t open_count;
1862 +       xid_t xid;
1863
1864         unsigned long flags;
1865
1866 @@ -250,6 +252,7 @@ static void __exit dm_exit(void)
1867  static int dm_blk_open(struct inode *inode, struct file *file)
1868  {
1869         struct mapped_device *md;
1870 +       int ret = -ENXIO;
1871
1872         spin_lock(&_minor_lock);
1873
1874 @@ -258,18 +261,19 @@ static int dm_blk_open(struct inode *ino
1875                 goto out;
1876
1877         if (test_bit(DMF_FREEING, &md->flags) ||
1878 -           test_bit(DMF_DELETING, &md->flags)) {
1879 -               md = NULL;
1880 +           test_bit(DMF_DELETING, &md->flags))
1881 +               goto out;
1882 +
1883 +       ret = -EACCES;
1884 +       if (!vx_check(md->xid, VS_IDENT|VS_HOSTID))
1885                 goto out;
1886 -       }
1887
1888         dm_get(md);
1889         atomic_inc(&md->open_count);
1890 -
1891 +       ret = 0;
1892  out:
1893         spin_unlock(&_minor_lock);
1894 -
1895 -       return md ? 0 : -ENXIO;
1896 +       return ret;
1897  }
1898
1899  static int dm_blk_close(struct inode *inode, struct file *file)
1900 @@ -465,6 +469,14 @@ int dm_set_geometry(struct mapped_device
1901         return 0;
1902  }
1903
1904 +/*
1905 + * Get the xid associated with a dm device
1906 + */
1907 +xid_t dm_get_xid(struct mapped_device *md)
1908 +{
1909 +       return md->xid;
1910 +}
1911 +
1912  /*-----------------------------------------------------------------
1913   * CRUD START:
1914   *   A more elegant soln is in the works that uses the queue
1915 @@ -1022,6 +1034,7 @@ static struct mapped_device *alloc_dev(i
1916         INIT_LIST_HEAD(&md->uevent_list);
1917         spin_lock_init(&md->uevent_lock);
1918
1919 +       md->xid = vx_current_xid();
1920         md->queue = blk_alloc_queue(GFP_KERNEL);
1921         if (!md->queue)
1922                 goto bad_queue;
1923 diff -NurpP --minimal linux-2.6.25.11/drivers/md/dm.h linux-2.6.25.11-vs2.3.0.34.14/drivers/md/dm.h
1924 --- linux-2.6.25.11/drivers/md/dm.h     2008-04-17 11:31:28.000000000 -0400
1925 +++ linux-2.6.25.11-vs2.3.0.34.14/drivers/md/dm.h       2008-04-19 15:14:52.000000000 -0400
1926 @@ -127,6 +127,8 @@ void dm_put_target_type(struct target_ty
1927  int dm_target_iterate(void (*iter_func)(struct target_type *tt,
1928                                         void *param), void *param);
1929
1930 +xid_t dm_get_xid(struct mapped_device *md);
1931 +
1932  /*-----------------------------------------------------------------
1933   * Useful inlines.
1934   *---------------------------------------------------------------*/
1935 diff -NurpP --minimal linux-2.6.25.11/drivers/md/dm-ioctl.c linux-2.6.25.11-vs2.3.0.34.14/drivers/md/dm-ioctl.c
1936 --- linux-2.6.25.11/drivers/md/dm-ioctl.c       2008-04-17 12:05:33.000000000 -0400
1937 +++ linux-2.6.25.11-vs2.3.0.34.14/drivers/md/dm-ioctl.c 2008-04-21 10:45:53.000000000 -0400
1938 @@ -16,6 +16,7 @@
1939  #include <linux/dm-ioctl.h>
1940  #include <linux/hdreg.h>
1941  #include <linux/compat.h>
1942 +#include <linux/vs_context.h>
1943
1944  #include <asm/uaccess.h>
1945
1946 @@ -101,7 +102,8 @@ static struct hash_cell *__get_name_cell
1947         unsigned int h = hash_str(str);
1948
1949         list_for_each_entry (hc, _name_buckets + h, name_list)
1950 -               if (!strcmp(hc->name, str)) {
1951 +               if (vx_check(dm_get_xid(hc->md), VS_WATCH_P | VS_IDENT) &&
1952 +                       !strcmp(hc->name, str)) {
1953                         dm_get(hc->md);
1954                         return hc;
1955                 }
1956 @@ -115,7 +117,8 @@ static struct hash_cell *__get_uuid_cell
1957         unsigned int h = hash_str(str);
1958
1959         list_for_each_entry (hc, _uuid_buckets + h, uuid_list)
1960 -               if (!strcmp(hc->uuid, str)) {
1961 +               if (vx_check(dm_get_xid(hc->md), VS_WATCH_P | VS_IDENT) &&
1962 +                       !strcmp(hc->uuid, str)) {
1963                         dm_get(hc->md);
1964                         return hc;
1965                 }
1966 @@ -352,6 +355,9 @@ typedef int (*ioctl_fn)(struct dm_ioctl
1967
1968  static int remove_all(struct dm_ioctl *param, size_t param_size)
1969  {
1970 +       if (!vx_check(0, VS_ADMIN))
1971 +               return -EPERM;
1972 +
1973         dm_hash_remove_all(1);
1974         param->data_size = 0;
1975         return 0;
1976 @@ -399,6 +405,8 @@ static int list_devices(struct dm_ioctl
1977          */
1978         for (i = 0; i < NUM_BUCKETS; i++) {
1979                 list_for_each_entry (hc, _name_buckets + i, name_list) {
1980 +                       if (!vx_check(dm_get_xid(hc->md), VS_WATCH_P | VS_IDENT))
1981 +                               continue;
1982                         needed += sizeof(struct dm_name_list);
1983                         needed += strlen(hc->name) + 1;
1984                         needed += ALIGN_MASK;
1985 @@ -422,6 +430,8 @@ static int list_devices(struct dm_ioctl
1986          */
1987         for (i = 0; i < NUM_BUCKETS; i++) {
1988                 list_for_each_entry (hc, _name_buckets + i, name_list) {
1989 +                       if (!vx_check(dm_get_xid(hc->md), VS_WATCH_P | VS_IDENT))
1990 +                               continue;
1991                         if (old_nl)
1992                                 old_nl->next = (uint32_t) ((void *) nl -
1993                                                            (void *) old_nl);
1994 @@ -612,10 +622,11 @@ static struct hash_cell *__find_device_h
1995         if (!md)
1996                 goto out;
1997
1998 -       mdptr = dm_get_mdptr(md);
1999 +       if (vx_check(dm_get_xid(md), VS_WATCH_P | VS_IDENT))
2000 +               mdptr = dm_get_mdptr(md);
2001 +
2002         if (!mdptr)
2003                 dm_put(md);
2004 -
2005  out:
2006         return mdptr;
2007  }
2008 @@ -1406,8 +1417,8 @@ static int ctl_ioctl(uint command, struc
2009         ioctl_fn fn = NULL;
2010         size_t param_size;
2011
2012 -       /* only root can play with this */
2013 -       if (!capable(CAP_SYS_ADMIN))
2014 +       /* only root and certain contexts can play with this */
2015 +       if (!vx_capable(CAP_SYS_ADMIN, VXC_ADMIN_MAPPER))
2016                 return -EACCES;
2017
2018         if (_IOC_TYPE(command) != DM_IOCTL)
2019 diff -NurpP --minimal linux-2.6.25.11/drivers/net/tun.c linux-2.6.25.11-vs2.3.0.34.14/drivers/net/tun.c
2020 --- linux-2.6.25.11/drivers/net/tun.c   2008-04-17 12:05:36.000000000 -0400
2021 +++ linux-2.6.25.11-vs2.3.0.34.14/drivers/net/tun.c     2008-04-19 15:39:04.000000000 -0400
2022 @@ -62,6 +62,7 @@
2023  #include <linux/if_ether.h>
2024  #include <linux/if_tun.h>
2025  #include <linux/crc32.h>
2026 +#include <linux/vs_network.h>
2027  #include <net/net_namespace.h>
2028
2029  #include <asm/system.h>
2030 @@ -86,6 +87,7 @@ struct tun_struct {
2031         int                     attached;
2032         uid_t                   owner;
2033         gid_t                   group;
2034 +       nid_t                   nid;
2035
2036         wait_queue_head_t       read_wait;
2037         struct sk_buff_head     readq;
2038 @@ -465,6 +467,7 @@ static void tun_setup(struct net_device
2039
2040         tun->owner = -1;
2041         tun->group = -1;
2042 +       tun->nid = current->nid;
2043
2044         dev->open = tun_net_open;
2045         dev->hard_start_xmit = tun_net_xmit;
2046 @@ -494,6 +497,9 @@ static int tun_set_iff(struct file *file
2047
2048         tun = tun_get_by_name(ifr->ifr_name);
2049         if (tun) {
2050 +               if (!nx_check(tun->nid, VS_IDENT | VS_HOSTID | VS_ADMIN_P))
2051 +                       return -EPERM;
2052 +
2053                 if (tun->attached)
2054                         return -EBUSY;
2055
2056 @@ -502,7 +508,7 @@ static int tun_set_iff(struct file *file
2057                       current->euid != tun->owner) ||
2058                      (tun->group != -1 &&
2059                       current->egid != tun->group)) &&
2060 -                    !capable(CAP_NET_ADMIN))
2061 +                    !cap_raised(current->cap_effective, CAP_NET_ADMIN))
2062                         return -EPERM;
2063         }
2064         else if (__dev_get_by_name(&init_net, ifr->ifr_name))
2065 @@ -513,7 +519,7 @@ static int tun_set_iff(struct file *file
2066
2067                 err = -EINVAL;
2068
2069 -               if (!capable(CAP_NET_ADMIN))
2070 +               if (!nx_capable(CAP_NET_ADMIN, NXC_TUN_CREATE))
2071                         return -EPERM;
2072
2073                 /* Set dev type */
2074 @@ -656,6 +662,16 @@ static int tun_chr_ioctl(struct inode *i
2075                 DBG(KERN_INFO "%s: group set to %d\n", tun->dev->name, tun->group);
2076                 break;
2077
2078 +       case TUNSETNID:
2079 +               if (!capable(CAP_CONTEXT))
2080 +                       return -EPERM;
2081 +
2082 +               /* Set nid owner of the device */
2083 +               tun->nid = (nid_t) arg;
2084 +
2085 +               DBG(KERN_INFO "%s: nid owner set to %u\n", tun->dev->name, tun->nid);
2086 +               break;
2087 +
2088         case TUNSETLINK:
2089                 /* Only allow setting the type when the interface is down */
2090                 if (tun->dev->flags & IFF_UP) {
2091 diff -NurpP --minimal linux-2.6.25.11/fs/attr.c linux-2.6.25.11-vs2.3.0.34.14/fs/attr.c
2092 --- linux-2.6.25.11/fs/attr.c   2008-04-17 11:31:35.000000000 -0400
2093 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/attr.c     2008-04-19 15:14:52.000000000 -0400
2094 @@ -14,6 +14,9 @@
2095  #include <linux/fcntl.h>
2096  #include <linux/quotaops.h>
2097  #include <linux/security.h>
2098 +#include <linux/proc_fs.h>
2099 +#include <linux/devpts_fs.h>
2100 +#include <linux/vs_base.h>
2101
2102  /* Taken over from the old code... */
2103
2104 @@ -55,6 +58,27 @@ int inode_change_ok(struct inode *inode,
2105                 if (!is_owner_or_cap(inode))
2106                         goto error;
2107         }
2108 +
2109 +       /* Check for evil vserver activity */
2110 +       if (vx_check(0, VS_ADMIN))
2111 +               goto fine;
2112 +
2113 +       if (IS_BARRIER(inode)) {
2114 +               vxwprintk_task(1, "messing with the barrier.");
2115 +               goto error;
2116 +       }
2117 +       switch (inode->i_sb->s_magic) {
2118 +               case PROC_SUPER_MAGIC:
2119 +                       /* maybe allow that in the future? */
2120 +                       vxwprintk_task(1, "messing with the procfs.");
2121 +                       goto error;
2122 +               case DEVPTS_SUPER_MAGIC:
2123 +                       /* devpts is xid tagged */
2124 +                       if (vx_check((xid_t)inode->i_tag, VS_IDENT))
2125 +                               goto fine;
2126 +                       vxwprintk_task(1, "messing with the devpts.");
2127 +                       goto error;
2128 +       }
2129  fine:
2130         retval = 0;
2131  error:
2132 @@ -78,6 +102,8 @@ int inode_setattr(struct inode * inode,
2133                 inode->i_uid = attr->ia_uid;
2134         if (ia_valid & ATTR_GID)
2135                 inode->i_gid = attr->ia_gid;
2136 +       if ((ia_valid & ATTR_TAG) && IS_TAGGED(inode))
2137 +               inode->i_tag = attr->ia_tag;
2138         if (ia_valid & ATTR_ATIME)
2139                 inode->i_atime = timespec_trunc(attr->ia_atime,
2140                                                 inode->i_sb->s_time_gran);
2141 @@ -167,7 +193,8 @@ int notify_change(struct dentry * dentry
2142                         error = security_inode_setattr(dentry, attr);
2143                 if (!error) {
2144                         if ((ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) ||
2145 -                           (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid))
2146 +                           (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid) ||
2147 +                           (ia_valid & ATTR_TAG && attr->ia_tag != inode->i_tag))
2148                                 error = DQUOT_TRANSFER(inode, attr) ? -EDQUOT : 0;
2149                         if (!error)
2150                                 error = inode_setattr(inode, attr);
2151 diff -NurpP --minimal linux-2.6.25.11/fs/binfmt_aout.c linux-2.6.25.11-vs2.3.0.34.14/fs/binfmt_aout.c
2152 --- linux-2.6.25.11/fs/binfmt_aout.c    2008-04-17 12:05:39.000000000 -0400
2153 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/binfmt_aout.c      2008-04-19 15:14:52.000000000 -0400
2154 @@ -24,6 +24,7 @@
2155  #include <linux/binfmts.h>
2156  #include <linux/personality.h>
2157  #include <linux/init.h>
2158 +#include <linux/vs_memory.h>
2159
2160  #include <asm/system.h>
2161  #include <asm/uaccess.h>
2162 diff -NurpP --minimal linux-2.6.25.11/fs/binfmt_elf.c linux-2.6.25.11-vs2.3.0.34.14/fs/binfmt_elf.c
2163 --- linux-2.6.25.11/fs/binfmt_elf.c     2008-04-17 12:05:39.000000000 -0400
2164 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/binfmt_elf.c       2008-04-19 15:14:52.000000000 -0400
2165 @@ -39,6 +39,7 @@
2166  #include <linux/random.h>
2167  #include <linux/elf.h>
2168  #include <linux/utsname.h>
2169 +#include <linux/vs_memory.h>
2170  #include <asm/uaccess.h>
2171  #include <asm/param.h>
2172  #include <asm/page.h>
2173 diff -NurpP --minimal linux-2.6.25.11/fs/binfmt_flat.c linux-2.6.25.11-vs2.3.0.34.14/fs/binfmt_flat.c
2174 --- linux-2.6.25.11/fs/binfmt_flat.c    2008-04-17 12:05:39.000000000 -0400
2175 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/binfmt_flat.c      2008-04-19 15:14:52.000000000 -0400
2176 @@ -35,6 +35,7 @@
2177  #include <linux/init.h>
2178  #include <linux/flat.h>
2179  #include <linux/syscalls.h>
2180 +#include <linux/vs_memory.h>
2181
2182  #include <asm/byteorder.h>
2183  #include <asm/system.h>
2184 diff -NurpP --minimal linux-2.6.25.11/fs/binfmt_som.c linux-2.6.25.11-vs2.3.0.34.14/fs/binfmt_som.c
2185 --- linux-2.6.25.11/fs/binfmt_som.c     2008-04-17 12:05:39.000000000 -0400
2186 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/binfmt_som.c       2008-04-19 15:14:52.000000000 -0400
2187 @@ -28,6 +28,7 @@
2188  #include <linux/shm.h>
2189  #include <linux/personality.h>
2190  #include <linux/init.h>
2191 +#include <linux/vs_memory.h>
2192
2193  #include <asm/uaccess.h>
2194  #include <asm/pgtable.h>
2195 diff -NurpP --minimal linux-2.6.25.11/fs/block_dev.c linux-2.6.25.11-vs2.3.0.34.14/fs/block_dev.c
2196 --- linux-2.6.25.11/fs/block_dev.c      2008-04-17 12:05:39.000000000 -0400
2197 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/block_dev.c        2008-04-19 15:14:52.000000000 -0400
2198 @@ -23,6 +23,7 @@
2199  #include <linux/uio.h>
2200  #include <linux/namei.h>
2201  #include <linux/log2.h>
2202 +#include <linux/vs_device.h>
2203  #include <asm/uaccess.h>
2204  #include "internal.h"
2205
2206 @@ -388,6 +389,7 @@ struct block_device *bdget(dev_t dev)
2207                 bdev->bd_invalidated = 0;
2208                 inode->i_mode = S_IFBLK;
2209                 inode->i_rdev = dev;
2210 +               inode->i_mdev = dev;
2211                 inode->i_bdev = bdev;
2212                 inode->i_data.a_ops = &def_blk_aops;
2213                 mapping_set_gfp_mask(&inode->i_data, GFP_USER);
2214 @@ -424,6 +426,11 @@ EXPORT_SYMBOL(bdput);
2215  static struct block_device *bd_acquire(struct inode *inode)
2216  {
2217         struct block_device *bdev;
2218 +       dev_t mdev;
2219 +
2220 +       if (!vs_map_blkdev(inode->i_rdev, &mdev, DATTR_OPEN))
2221 +               return NULL;
2222 +       inode->i_mdev = mdev;
2223
2224         spin_lock(&bdev_lock);
2225         bdev = inode->i_bdev;
2226 @@ -434,7 +441,7 @@ static struct block_device *bd_acquire(s
2227         }
2228         spin_unlock(&bdev_lock);
2229
2230 -       bdev = bdget(inode->i_rdev);
2231 +       bdev = bdget(mdev);
2232         if (bdev) {
2233                 spin_lock(&bdev_lock);
2234                 if (!inode->i_bdev) {
2235 diff -NurpP --minimal linux-2.6.25.11/fs/char_dev.c linux-2.6.25.11-vs2.3.0.34.14/fs/char_dev.c
2236 --- linux-2.6.25.11/fs/char_dev.c       2008-04-17 12:05:39.000000000 -0400
2237 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/char_dev.c 2008-04-29 18:44:50.000000000 -0400
2238 @@ -21,6 +21,8 @@
2239  #include <linux/cdev.h>
2240  #include <linux/mutex.h>
2241  #include <linux/backing-dev.h>
2242 +#include <linux/vs_context.h>
2243 +#include <linux/vs_device.h>
2244
2245  #ifdef CONFIG_KMOD
2246  #include <linux/kmod.h>
2247 @@ -362,14 +364,21 @@ static int chrdev_open(struct inode *ino
2248         struct cdev *p;
2249         struct cdev *new = NULL;
2250         int ret = 0;
2251 +       dev_t mdev;
2252 +
2253 +       if (!vs_map_chrdev(inode->i_rdev, &mdev, DATTR_OPEN))
2254 +               return -EPERM;
2255 +       inode->i_mdev = mdev;
2256
2257         spin_lock(&cdev_lock);
2258         p = inode->i_cdev;
2259         if (!p) {
2260                 struct kobject *kobj;
2261                 int idx;
2262 +
2263                 spin_unlock(&cdev_lock);
2264 -               kobj = kobj_lookup(cdev_map, inode->i_rdev, &idx);
2265 +
2266 +               kobj = kobj_lookup(cdev_map, mdev, &idx);
2267                 if (!kobj)
2268                         return -ENXIO;
2269                 new = container_of(kobj, struct cdev, kobj);
2270 diff -NurpP --minimal linux-2.6.25.11/fs/dcache.c linux-2.6.25.11-vs2.3.0.34.14/fs/dcache.c
2271 --- linux-2.6.25.11/fs/dcache.c 2008-04-17 12:05:39.000000000 -0400
2272 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/dcache.c   2008-04-19 17:06:15.000000000 -0400
2273 @@ -31,6 +31,7 @@
2274  #include <linux/seqlock.h>
2275  #include <linux/swap.h>
2276  #include <linux/bootmem.h>
2277 +#include <linux/vs_limit.h>
2278  #include "internal.h"
2279
2280
2281 @@ -184,6 +185,7 @@ void dput(struct dentry *dentry)
2282         if (!dentry)
2283                 return;
2284
2285 +       vx_dentry_dec(dentry);
2286  repeat:
2287         if (atomic_read(&dentry->d_count) == 1)
2288                 might_sleep();
2289 @@ -197,6 +199,8 @@ repeat:
2290                 return;
2291         }
2292
2293 +       vx_dentry_dec(dentry);
2294 +
2295         /*
2296          * AV: ->d_delete() is _NOT_ allowed to block now.
2297          */
2298 @@ -288,6 +292,7 @@ static inline struct dentry * __dget_loc
2299  {
2300         atomic_inc(&dentry->d_count);
2301         dentry_lru_remove(dentry);
2302 +       vx_dentry_inc(dentry);
2303         return dentry;
2304  }
2305
2306 @@ -885,6 +890,9 @@ struct dentry *d_alloc(struct dentry * p
2307         struct dentry *dentry;
2308         char *dname;
2309
2310 +       if (!vx_dentry_avail(1))
2311 +               return NULL;
2312 +
2313         dentry = kmem_cache_alloc(dentry_cache, GFP_KERNEL);
2314         if (!dentry)
2315                 return NULL;
2316 @@ -933,6 +941,7 @@ struct dentry *d_alloc(struct dentry * p
2317         if (parent)
2318                 list_add(&dentry->d_u.d_child, &parent->d_subdirs);
2319         dentry_stat.nr_dentry++;
2320 +       vx_dentry_inc(dentry);
2321         spin_unlock(&dcache_lock);
2322
2323         return dentry;
2324 @@ -1282,6 +1291,7 @@ struct dentry * __d_lookup(struct dentry
2325
2326                 if (!d_unhashed(dentry)) {
2327                         atomic_inc(&dentry->d_count);
2328 +                       vx_dentry_inc(dentry);
2329                         found = dentry;
2330                 }
2331                 spin_unlock(&dentry->d_lock);
2332 diff -NurpP --minimal linux-2.6.25.11/fs/devpts/inode.c linux-2.6.25.11-vs2.3.0.34.14/fs/devpts/inode.c
2333 --- linux-2.6.25.11/fs/devpts/inode.c   2008-04-17 12:05:39.000000000 -0400
2334 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/devpts/inode.c     2008-04-21 09:23:34.000000000 -0400
2335 @@ -17,15 +17,30 @@
2336  #include <linux/namei.h>
2337  #include <linux/mount.h>
2338  #include <linux/tty.h>
2339 +#include <linux/magic.h>
2340  #include <linux/devpts_fs.h>
2341  #include <linux/parser.h>
2342  #include <linux/fsnotify.h>
2343  #include <linux/seq_file.h>
2344 +#include <linux/vs_base.h>
2345
2346 -#define DEVPTS_SUPER_MAGIC 0x1cd1
2347
2348  #define DEVPTS_DEFAULT_MODE 0600
2349
2350 +static int devpts_permission(struct inode *inode, int mask, struct nameidata *nd)
2351 +{
2352 +       int ret = -EACCES;
2353 +
2354 +       /* devpts is xid tagged */
2355 +       if (vx_check((xid_t)inode->i_tag, VS_WATCH_P | VS_IDENT))
2356 +               ret = generic_permission(inode, mask, NULL);
2357 +       return ret;
2358 +}
2359 +
2360 +static struct inode_operations devpts_file_inode_operations = {
2361 +       .permission     = devpts_permission,
2362 +};
2363 +
2364  static struct vfsmount *devpts_mnt;
2365  static struct dentry *devpts_root;
2366
2367 @@ -106,6 +121,25 @@ static int devpts_show_options(struct se
2368         return 0;
2369  }
2370
2371 +static int devpts_filter(struct dentry *de)
2372 +{
2373 +       /* devpts is xid tagged */
2374 +       return vx_check((xid_t)de->d_inode->i_tag, VS_WATCH_P | VS_IDENT);
2375 +}
2376 +
2377 +static int devpts_readdir(struct file * filp, void * dirent, filldir_t filldir)
2378 +{
2379 +       return dcache_readdir_filter(filp, dirent, filldir, devpts_filter);
2380 +}
2381 +
2382 +static struct file_operations devpts_dir_operations = {
2383 +       .open           = dcache_dir_open,
2384 +       .release        = dcache_dir_close,
2385 +       .llseek         = dcache_dir_lseek,
2386 +       .read           = generic_read_dir,
2387 +       .readdir        = devpts_readdir,
2388 +};
2389 +
2390  static const struct super_operations devpts_sops = {
2391         .statfs         = simple_statfs,
2392         .remount_fs     = devpts_remount,
2393 @@ -132,8 +166,10 @@ devpts_fill_super(struct super_block *s,
2394         inode->i_uid = inode->i_gid = 0;
2395         inode->i_mode = S_IFDIR | S_IRUGO | S_IXUGO | S_IWUSR;
2396         inode->i_op = &simple_dir_inode_operations;
2397 -       inode->i_fop = &simple_dir_operations;
2398 +       inode->i_fop = &devpts_dir_operations;
2399         inode->i_nlink = 2;
2400 +       /* devpts is xid tagged */
2401 +       inode->i_tag = (tag_t)vx_current_xid();
2402
2403         devpts_root = s->s_root = d_alloc_root(inode);
2404         if (s->s_root)
2405 @@ -191,6 +227,9 @@ int devpts_pty_new(struct tty_struct *tt
2406         inode->i_gid = config.setgid ? config.gid : current->fsgid;
2407         inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
2408         init_special_inode(inode, S_IFCHR|config.mode, device);
2409 +       /* devpts is xid tagged */
2410 +       inode->i_tag = (tag_t)vx_current_xid();
2411 +       inode->i_op = &devpts_file_inode_operations;
2412         inode->i_private = tty;
2413
2414         dentry = get_node(number);
2415 diff -NurpP --minimal linux-2.6.25.11/fs/ecryptfs/inode.c linux-2.6.25.11-vs2.3.0.34.14/fs/ecryptfs/inode.c
2416 --- linux-2.6.25.11/fs/ecryptfs/inode.c 2008-07-15 22:05:17.000000000 -0400
2417 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ecryptfs/inode.c   2008-06-23 15:23:51.000000000 -0400
2418 @@ -394,7 +394,7 @@ static int ecryptfs_link(struct dentry *
2419         dget(lower_new_dentry);
2420         lower_dir_dentry = lock_parent(lower_new_dentry);
2421         rc = vfs_link(lower_old_dentry, lower_dir_dentry->d_inode,
2422 -                     lower_new_dentry);
2423 +                     lower_new_dentry, NULL);
2424         if (rc || !lower_new_dentry->d_inode)
2425                 goto out_lock;
2426         rc = ecryptfs_interpose(lower_new_dentry, new_dentry, dir->i_sb, 0);
2427 @@ -421,9 +421,9 @@ static int ecryptfs_unlink(struct inode
2428         struct dentry *lower_dentry = ecryptfs_dentry_to_lower(dentry);
2429         struct inode *lower_dir_inode = ecryptfs_inode_to_lower(dir);
2430         struct dentry *lower_dir_dentry;
2431 -
2432 +
2433         lower_dir_dentry = lock_parent(lower_dentry);
2434 -       rc = vfs_unlink(lower_dir_inode, lower_dentry);
2435 +       rc = vfs_unlink(lower_dir_inode, lower_dentry, NULL);
2436         if (rc) {
2437                 printk(KERN_ERR "Error in vfs_unlink; rc = [%d]\n", rc);
2438                 goto out_unlock;
2439 @@ -461,7 +461,7 @@ static int ecryptfs_symlink(struct inode
2440                 goto out_lock;
2441         }
2442         rc = vfs_symlink(lower_dir_dentry->d_inode, lower_dentry,
2443 -                        encoded_symname, mode);
2444 +                        encoded_symname, mode, NULL);
2445         kfree(encoded_symname);
2446         if (rc || !lower_dentry->d_inode)
2447                 goto out_lock;
2448 @@ -486,7 +486,7 @@ static int ecryptfs_mkdir(struct inode *
2449
2450         lower_dentry = ecryptfs_dentry_to_lower(dentry);
2451         lower_dir_dentry = lock_parent(lower_dentry);
2452 -       rc = vfs_mkdir(lower_dir_dentry->d_inode, lower_dentry, mode);
2453 +       rc = vfs_mkdir(lower_dir_dentry->d_inode, lower_dentry, mode, NULL);
2454         if (rc || !lower_dentry->d_inode)
2455                 goto out;
2456         rc = ecryptfs_interpose(lower_dentry, dentry, dir->i_sb, 0);
2457 @@ -512,7 +512,7 @@ static int ecryptfs_rmdir(struct inode *
2458         dget(dentry);
2459         lower_dir_dentry = lock_parent(lower_dentry);
2460         dget(lower_dentry);
2461 -       rc = vfs_rmdir(lower_dir_dentry->d_inode, lower_dentry);
2462 +       rc = vfs_rmdir(lower_dir_dentry->d_inode, lower_dentry, NULL);
2463         dput(lower_dentry);
2464         if (!rc)
2465                 d_delete(lower_dentry);
2466 @@ -534,7 +534,7 @@ ecryptfs_mknod(struct inode *dir, struct
2467
2468         lower_dentry = ecryptfs_dentry_to_lower(dentry);
2469         lower_dir_dentry = lock_parent(lower_dentry);
2470 -       rc = vfs_mknod(lower_dir_dentry->d_inode, lower_dentry, mode, dev);
2471 +       rc = vfs_mknod(lower_dir_dentry->d_inode, lower_dentry, mode, dev, NULL);
2472         if (rc || !lower_dentry->d_inode)
2473                 goto out;
2474         rc = ecryptfs_interpose(lower_dentry, dentry, dir->i_sb, 0);
2475 diff -NurpP --minimal linux-2.6.25.11/fs/exec.c linux-2.6.25.11-vs2.3.0.34.14/fs/exec.c
2476 --- linux-2.6.25.11/fs/exec.c   2008-04-17 12:05:40.000000000 -0400
2477 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/exec.c     2008-04-19 15:14:52.000000000 -0400
2478 @@ -249,7 +249,9 @@ static int __bprm_mm_init(struct linux_b
2479                 goto err;
2480         }
2481
2482 -       mm->stack_vm = mm->total_vm = 1;
2483 +       mm->total_vm = 0;
2484 +       vx_vmpages_inc(mm);
2485 +       mm->stack_vm = 1;
2486         up_write(&mm->mmap_sem);
2487
2488         bprm->p = vma->vm_end - sizeof(void *);
2489 @@ -1452,7 +1454,7 @@ static int format_corename(char *corenam
2490                         /* UNIX time of coredump */
2491                         case 't': {
2492                                 struct timeval tv;
2493 -                               do_gettimeofday(&tv);
2494 +                               vx_gettimeofday(&tv);
2495                                 rc = snprintf(out_ptr, out_end - out_ptr,
2496                                               "%lu", tv.tv_sec);
2497                                 if (rc > out_end - out_ptr)
2498 diff -NurpP --minimal linux-2.6.25.11/fs/ext2/balloc.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext2/balloc.c
2499 --- linux-2.6.25.11/fs/ext2/balloc.c    2008-04-17 12:05:40.000000000 -0400
2500 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext2/balloc.c      2008-04-19 15:14:52.000000000 -0400
2501 @@ -16,6 +16,8 @@
2502  #include <linux/sched.h>
2503  #include <linux/buffer_head.h>
2504  #include <linux/capability.h>
2505 +#include <linux/vs_dlimit.h>
2506 +#include <linux/vs_tag.h>
2507
2508  /*
2509   * balloc.c contains the blocks allocation and deallocation routines
2510 @@ -569,6 +571,7 @@ do_more:
2511         }
2512  error_return:
2513         brelse(bitmap_bh);
2514 +       DLIMIT_FREE_BLOCK(inode, freed);
2515         release_blocks(sb, freed);
2516         DQUOT_FREE_BLOCK(inode, freed);
2517  }
2518 @@ -701,7 +704,6 @@ ext2_try_to_allocate(struct super_block
2519                         start = 0;
2520                 end = EXT2_BLOCKS_PER_GROUP(sb);
2521         }
2522 -
2523         BUG_ON(start > EXT2_BLOCKS_PER_GROUP(sb));
2524
2525  repeat:
2526 @@ -1251,6 +1253,11 @@ ext2_fsblk_t ext2_new_blocks(struct inod
2527                 *errp = -EDQUOT;
2528                 return 0;
2529         }
2530 +       if (DLIMIT_ALLOC_BLOCK(inode, num)) {
2531 +               *errp = -ENOSPC;
2532 +               DQUOT_FREE_BLOCK(inode, num);
2533 +               return 0;
2534 +       }
2535
2536         sbi = EXT2_SB(sb);
2537         es = EXT2_SB(sb)->s_es;
2538 @@ -1403,6 +1410,7 @@ allocated:
2539
2540         *errp = 0;
2541         brelse(bitmap_bh);
2542 +       DLIMIT_FREE_BLOCK(inode, *count-num);
2543         DQUOT_FREE_BLOCK(inode, *count-num);
2544         *count = num;
2545         return ret_block;
2546 @@ -1413,8 +1421,10 @@ out:
2547         /*
2548          * Undo the block allocation
2549          */
2550 -       if (!performed_allocation)
2551 +       if (!performed_allocation) {
2552 +               DLIMIT_FREE_BLOCK(inode, *count);
2553                 DQUOT_FREE_BLOCK(inode, *count);
2554 +       }
2555         brelse(bitmap_bh);
2556         return 0;
2557  }
2558 diff -NurpP --minimal linux-2.6.25.11/fs/ext2/ext2.h linux-2.6.25.11-vs2.3.0.34.14/fs/ext2/ext2.h
2559 --- linux-2.6.25.11/fs/ext2/ext2.h      2008-04-17 12:05:40.000000000 -0400
2560 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext2/ext2.h        2008-04-19 15:14:52.000000000 -0400
2561 @@ -168,6 +168,7 @@ extern const struct file_operations ext2
2562  extern const struct address_space_operations ext2_aops;
2563  extern const struct address_space_operations ext2_aops_xip;
2564  extern const struct address_space_operations ext2_nobh_aops;
2565 +extern int ext2_sync_flags(struct inode *inode);
2566
2567  /* namei.c */
2568  extern const struct inode_operations ext2_dir_inode_operations;
2569 diff -NurpP --minimal linux-2.6.25.11/fs/ext2/file.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext2/file.c
2570 --- linux-2.6.25.11/fs/ext2/file.c      2008-04-17 12:05:40.000000000 -0400
2571 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext2/file.c        2008-04-19 15:14:52.000000000 -0400
2572 @@ -86,4 +86,5 @@ const struct inode_operations ext2_file_
2573  #endif
2574         .setattr        = ext2_setattr,
2575         .permission     = ext2_permission,
2576 +       .sync_flags     = ext2_sync_flags,
2577  };
2578 diff -NurpP --minimal linux-2.6.25.11/fs/ext2/ialloc.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext2/ialloc.c
2579 --- linux-2.6.25.11/fs/ext2/ialloc.c    2008-04-17 11:31:35.000000000 -0400
2580 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext2/ialloc.c      2008-07-08 04:31:24.000000000 -0400
2581 @@ -17,6 +17,8 @@
2582  #include <linux/backing-dev.h>
2583  #include <linux/buffer_head.h>
2584  #include <linux/random.h>
2585 +#include <linux/vs_dlimit.h>
2586 +#include <linux/vs_tag.h>
2587  #include "ext2.h"
2588  #include "xattr.h"
2589  #include "acl.h"
2590 @@ -125,6 +127,7 @@ void ext2_free_inode (struct inode * ino
2591                 ext2_xattr_delete_inode(inode);
2592                 DQUOT_FREE_INODE(inode);
2593                 DQUOT_DROP(inode);
2594 +               DLIMIT_FREE_INODE(inode);
2595         }
2596
2597         es = EXT2_SB(sb)->s_es;
2598 @@ -456,6 +459,11 @@ struct inode *ext2_new_inode(struct inod
2599         if (!inode)
2600                 return ERR_PTR(-ENOMEM);
2601
2602 +       inode->i_tag = dx_current_fstag(sb);
2603 +       if (DLIMIT_ALLOC_INODE(inode)) {
2604 +               err = -ENOSPC;
2605 +               goto fail_dlim;
2606 +       }
2607         ei = EXT2_I(inode);
2608         sbi = EXT2_SB(sb);
2609         es = sbi->s_es;
2610 @@ -569,7 +577,8 @@ got:
2611         inode->i_blocks = 0;
2612         inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME_SEC;
2613         memset(ei->i_data, 0, sizeof(ei->i_data));
2614 -       ei->i_flags = EXT2_I(dir)->i_flags & ~EXT2_BTREE_FL;
2615 +       ei->i_flags = EXT2_I(dir)->i_flags &
2616 +               ~(EXT2_BTREE_FL|EXT2_IXUNLINK_FL|EXT2_BARRIER_FL);
2617         if (S_ISLNK(mode))
2618                 ei->i_flags &= ~(EXT2_IMMUTABLE_FL|EXT2_APPEND_FL);
2619         /* dirsync is only applied to directories */
2620 @@ -614,12 +623,15 @@ fail_free_drop:
2621
2622  fail_drop:
2623         DQUOT_DROP(inode);
2624 +       DLIMIT_FREE_INODE(inode);
2625         inode->i_flags |= S_NOQUOTA;
2626         inode->i_nlink = 0;
2627         iput(inode);
2628         return ERR_PTR(err);
2629
2630  fail:
2631 +       DLIMIT_FREE_INODE(inode);
2632 +fail_dlim:
2633         make_bad_inode(inode);
2634         iput(inode);
2635         return ERR_PTR(err);
2636 diff -NurpP --minimal linux-2.6.25.11/fs/ext2/inode.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext2/inode.c
2637 --- linux-2.6.25.11/fs/ext2/inode.c     2008-04-17 12:05:40.000000000 -0400
2638 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext2/inode.c       2008-07-08 16:53:21.000000000 -0400
2639 @@ -31,6 +31,7 @@
2640  #include <linux/writeback.h>
2641  #include <linux/buffer_head.h>
2642  #include <linux/mpage.h>
2643 +#include <linux/vs_tag.h>
2644  #include "ext2.h"
2645  #include "acl.h"
2646  #include "xip.h"
2647 @@ -1011,7 +1012,7 @@ void ext2_truncate(struct inode *inode)
2648                 return;
2649         if (ext2_inode_is_fast_symlink(inode))
2650                 return;
2651 -       if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
2652 +       if (IS_APPEND(inode) || IS_IXORUNLINK(inode))
2653                 return;
2654
2655         blocksize = inode->i_sb->s_blocksize;
2656 @@ -1149,38 +1150,72 @@ void ext2_set_inode_flags(struct inode *
2657  {
2658         unsigned int flags = EXT2_I(inode)->i_flags;
2659
2660 -       inode->i_flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC);
2661 +       inode->i_flags &= ~(S_IMMUTABLE | S_IXUNLINK |
2662 +               S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
2663 +
2664 +
2665 +       if (flags & EXT2_IMMUTABLE_FL)
2666 +               inode->i_flags |= S_IMMUTABLE;
2667 +       if (flags & EXT2_IXUNLINK_FL)
2668 +               inode->i_flags |= S_IXUNLINK;
2669 +
2670         if (flags & EXT2_SYNC_FL)
2671                 inode->i_flags |= S_SYNC;
2672         if (flags & EXT2_APPEND_FL)
2673                 inode->i_flags |= S_APPEND;
2674 -       if (flags & EXT2_IMMUTABLE_FL)
2675 -               inode->i_flags |= S_IMMUTABLE;
2676         if (flags & EXT2_NOATIME_FL)
2677                 inode->i_flags |= S_NOATIME;
2678         if (flags & EXT2_DIRSYNC_FL)
2679                 inode->i_flags |= S_DIRSYNC;
2680 +
2681 +       inode->i_vflags &= ~(V_BARRIER | V_COW);
2682 +
2683 +       if (flags & EXT2_BARRIER_FL)
2684 +               inode->i_vflags |= V_BARRIER;
2685 +       if (flags & EXT2_COW_FL)
2686 +               inode->i_vflags |= V_COW;
2687  }
2688
2689  /* Propagate flags from i_flags to EXT2_I(inode)->i_flags */
2690  void ext2_get_inode_flags(struct ext2_inode_info *ei)
2691  {
2692         unsigned int flags = ei->vfs_inode.i_flags;
2693 +       unsigned int vflags = ei->vfs_inode.i_vflags;
2694 +
2695 +       ei->i_flags &= ~(EXT2_SYNC_FL | EXT2_APPEND_FL |
2696 +                       EXT2_IMMUTABLE_FL | EXT2_IXUNLINK_FL |
2697 +                       EXT2_NOATIME_FL | EXT2_DIRSYNC_FL |
2698 +                       EXT2_BARRIER_FL | EXT2_COW_FL);
2699 +
2700 +       if (flags & S_IMMUTABLE)
2701 +               ei->i_flags |= EXT2_IMMUTABLE_FL;
2702 +       if (flags & S_IXUNLINK)
2703 +               ei->i_flags |= EXT2_IXUNLINK_FL;
2704
2705 -       ei->i_flags &= ~(EXT2_SYNC_FL|EXT2_APPEND_FL|
2706 -                       EXT2_IMMUTABLE_FL|EXT2_NOATIME_FL|EXT2_DIRSYNC_FL);
2707         if (flags & S_SYNC)
2708                 ei->i_flags |= EXT2_SYNC_FL;
2709         if (flags & S_APPEND)
2710                 ei->i_flags |= EXT2_APPEND_FL;
2711 -       if (flags & S_IMMUTABLE)
2712 -               ei->i_flags |= EXT2_IMMUTABLE_FL;
2713         if (flags & S_NOATIME)
2714                 ei->i_flags |= EXT2_NOATIME_FL;
2715         if (flags & S_DIRSYNC)
2716                 ei->i_flags |= EXT2_DIRSYNC_FL;
2717 +
2718 +       if (vflags & V_BARRIER)
2719 +               ei->i_flags |= EXT2_BARRIER_FL;
2720 +       if (vflags & V_COW)
2721 +               ei->i_flags |= EXT2_COW_FL;
2722  }
2723
2724 +int ext2_sync_flags(struct inode *inode)
2725 +{
2726 +       ext2_get_inode_flags(EXT2_I(inode));
2727 +       inode->i_ctime = CURRENT_TIME;
2728 +       mark_inode_dirty(inode);
2729 +       return 0;
2730 +}
2731 +
2732 +
2733  struct inode *ext2_iget (struct super_block *sb, unsigned long ino)
2734  {
2735         struct ext2_inode_info *ei;
2736 @@ -1188,6 +1223,8 @@ struct inode *ext2_iget (struct super_bl
2737         struct ext2_inode *raw_inode;
2738         struct inode *inode;
2739         long ret = -EIO;
2740 +       uid_t uid;
2741 +       gid_t gid;
2742         int n;
2743
2744         inode = iget_locked(sb, ino);
2745 @@ -1210,12 +1247,17 @@ struct inode *ext2_iget (struct super_bl
2746         }
2747
2748         inode->i_mode = le16_to_cpu(raw_inode->i_mode);
2749 -       inode->i_uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
2750 -       inode->i_gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
2751 +       uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
2752 +       gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
2753         if (!(test_opt (inode->i_sb, NO_UID32))) {
2754 -               inode->i_uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
2755 -               inode->i_gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
2756 +               uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
2757 +               gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
2758         }
2759 +       inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
2760 +       inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
2761 +       inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid,
2762 +               le16_to_cpu(raw_inode->i_raw_tag));
2763 +
2764         inode->i_nlink = le16_to_cpu(raw_inode->i_links_count);
2765         inode->i_size = le32_to_cpu(raw_inode->i_size);
2766         inode->i_atime.tv_sec = (signed)le32_to_cpu(raw_inode->i_atime);
2767 @@ -1311,8 +1353,8 @@ static int ext2_update_inode(struct inod
2768         struct ext2_inode_info *ei = EXT2_I(inode);
2769         struct super_block *sb = inode->i_sb;
2770         ino_t ino = inode->i_ino;
2771 -       uid_t uid = inode->i_uid;
2772 -       gid_t gid = inode->i_gid;
2773 +       uid_t uid = TAGINO_UID(DX_TAG(inode), inode->i_uid, inode->i_tag);
2774 +       gid_t gid = TAGINO_GID(DX_TAG(inode), inode->i_gid, inode->i_tag);
2775         struct buffer_head * bh;
2776         struct ext2_inode * raw_inode = ext2_get_inode(sb, ino, &bh);
2777         int n;
2778 @@ -1348,6 +1390,9 @@ static int ext2_update_inode(struct inod
2779                 raw_inode->i_uid_high = 0;
2780                 raw_inode->i_gid_high = 0;
2781         }
2782 +#ifdef CONFIG_TAGGING_INTERN
2783 +       raw_inode->i_raw_tag = cpu_to_le16(inode->i_tag);
2784 +#endif
2785         raw_inode->i_links_count = cpu_to_le16(inode->i_nlink);
2786         raw_inode->i_size = cpu_to_le32(inode->i_size);
2787         raw_inode->i_atime = cpu_to_le32(inode->i_atime.tv_sec);
2788 @@ -1434,7 +1479,8 @@ int ext2_setattr(struct dentry *dentry,
2789         if (error)
2790                 return error;
2791         if ((iattr->ia_valid & ATTR_UID && iattr->ia_uid != inode->i_uid) ||
2792 -           (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid)) {
2793 +           (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid) ||
2794 +           (iattr->ia_valid & ATTR_TAG && iattr->ia_tag != inode->i_tag)) {
2795                 error = DQUOT_TRANSFER(inode, iattr) ? -EDQUOT : 0;
2796                 if (error)
2797                         return error;
2798 diff -NurpP --minimal linux-2.6.25.11/fs/ext2/ioctl.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext2/ioctl.c
2799 --- linux-2.6.25.11/fs/ext2/ioctl.c     2008-04-17 12:05:40.000000000 -0400
2800 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext2/ioctl.c       2008-07-11 02:48:03.000000000 -0400
2801 @@ -13,6 +13,7 @@
2802  #include <linux/sched.h>
2803  #include <linux/compat.h>
2804  #include <linux/smp_lock.h>
2805 +#include <linux/mount.h>
2806  #include <asm/current.h>
2807  #include <asm/uaccess.h>
2808
2809 @@ -34,7 +35,8 @@ long ext2_ioctl(struct file *filp, unsig
2810         case EXT2_IOC_SETFLAGS: {
2811                 unsigned int oldflags;
2812
2813 -               if (IS_RDONLY(inode))
2814 +               if (IS_RDONLY(inode) ||
2815 +                       (filp && MNT_IS_RDONLY(filp->f_vfsmnt)))
2816                         return -EROFS;
2817
2818                 if (!is_owner_or_cap(inode))
2819 @@ -46,6 +48,11 @@ long ext2_ioctl(struct file *filp, unsig
2820                 if (!S_ISDIR(inode->i_mode))
2821                         flags &= ~EXT2_DIRSYNC_FL;
2822
2823 +               if (IS_BARRIER(inode)) {
2824 +                       vxwprintk_task(1, "messing with the barrier.");
2825 +                       return -EACCES;
2826 +               }
2827 +
2828                 mutex_lock(&inode->i_mutex);
2829                 /* Is it quota file? Do not allow user to mess with it */
2830                 if (IS_NOQUOTA(inode)) {
2831 @@ -60,7 +67,9 @@ long ext2_ioctl(struct file *filp, unsig
2832                  *
2833                  * This test looks nicer. Thanks to Pauline Middelink
2834                  */
2835 -               if ((flags ^ oldflags) & (EXT2_APPEND_FL | EXT2_IMMUTABLE_FL)) {
2836 +               if ((oldflags & EXT2_IMMUTABLE_FL) ||
2837 +                       ((flags ^ oldflags) & (EXT2_APPEND_FL |
2838 +                       EXT2_IMMUTABLE_FL | EXT2_IXUNLINK_FL))) {
2839                         if (!capable(CAP_LINUX_IMMUTABLE)) {
2840                                 mutex_unlock(&inode->i_mutex);
2841                                 return -EPERM;
2842 @@ -82,7 +91,8 @@ long ext2_ioctl(struct file *filp, unsig
2843         case EXT2_IOC_SETVERSION:
2844                 if (!is_owner_or_cap(inode))
2845                         return -EPERM;
2846 -               if (IS_RDONLY(inode))
2847 +               if (IS_RDONLY(inode) ||
2848 +                       (filp && MNT_IS_RDONLY(filp->f_vfsmnt)))
2849                         return -EROFS;
2850                 if (get_user(inode->i_generation, (int __user *) arg))
2851                         return -EFAULT;
2852 diff -NurpP --minimal linux-2.6.25.11/fs/ext2/namei.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext2/namei.c
2853 --- linux-2.6.25.11/fs/ext2/namei.c     2008-04-17 12:05:40.000000000 -0400
2854 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext2/namei.c       2008-04-21 10:12:49.000000000 -0400
2855 @@ -31,6 +31,7 @@
2856   */
2857
2858  #include <linux/pagemap.h>
2859 +#include <linux/vs_tag.h>
2860  #include "ext2.h"
2861  #include "xattr.h"
2862  #include "acl.h"
2863 @@ -66,6 +67,7 @@ static struct dentry *ext2_lookup(struct
2864                 inode = ext2_iget(dir->i_sb, ino);
2865                 if (IS_ERR(inode))
2866                         return ERR_CAST(inode);
2867 +               dx_propagate_tag(nd, inode);
2868         }
2869         return d_splice_alias(inode, dentry);
2870  }
2871 @@ -391,6 +393,7 @@ const struct inode_operations ext2_dir_i
2872  #endif
2873         .setattr        = ext2_setattr,
2874         .permission     = ext2_permission,
2875 +       .sync_flags     = ext2_sync_flags,
2876  };
2877
2878  const struct inode_operations ext2_special_inode_operations = {
2879 @@ -402,4 +405,5 @@ const struct inode_operations ext2_speci
2880  #endif
2881         .setattr        = ext2_setattr,
2882         .permission     = ext2_permission,
2883 +       .sync_flags     = ext2_sync_flags,
2884  };
2885 diff -NurpP --minimal linux-2.6.25.11/fs/ext2/super.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext2/super.c
2886 --- linux-2.6.25.11/fs/ext2/super.c     2008-04-17 12:05:40.000000000 -0400
2887 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext2/super.c       2008-04-19 15:14:52.000000000 -0400
2888 @@ -390,7 +390,8 @@ enum {
2889         Opt_err_ro, Opt_nouid32, Opt_nocheck, Opt_debug,
2890         Opt_oldalloc, Opt_orlov, Opt_nobh, Opt_user_xattr, Opt_nouser_xattr,
2891         Opt_acl, Opt_noacl, Opt_xip, Opt_ignore, Opt_err, Opt_quota,
2892 -       Opt_usrquota, Opt_grpquota, Opt_reservation, Opt_noreservation
2893 +       Opt_usrquota, Opt_grpquota, Opt_reservation, Opt_noreservation,
2894 +       Opt_tag, Opt_notag, Opt_tagid
2895  };
2896
2897  static match_table_t tokens = {
2898 @@ -418,6 +419,9 @@ static match_table_t tokens = {
2899         {Opt_acl, "acl"},
2900         {Opt_noacl, "noacl"},
2901         {Opt_xip, "xip"},
2902 +       {Opt_tag, "tag"},
2903 +       {Opt_notag, "notag"},
2904 +       {Opt_tagid, "tagid=%u"},
2905         {Opt_grpquota, "grpquota"},
2906         {Opt_ignore, "noquota"},
2907         {Opt_quota, "quota"},
2908 @@ -488,6 +492,20 @@ static int parse_options (char * options
2909                 case Opt_nouid32:
2910                         set_opt (sbi->s_mount_opt, NO_UID32);
2911                         break;
2912 +#ifndef CONFIG_TAGGING_NONE
2913 +               case Opt_tag:
2914 +                       set_opt (sbi->s_mount_opt, TAGGED);
2915 +                       break;
2916 +               case Opt_notag:
2917 +                       clear_opt (sbi->s_mount_opt, TAGGED);
2918 +                       break;
2919 +#endif
2920 +#ifdef CONFIG_PROPAGATE
2921 +               case Opt_tagid:
2922 +                       /* use args[0] */
2923 +                       set_opt (sbi->s_mount_opt, TAGGED);
2924 +                       break;
2925 +#endif
2926                 case Opt_nocheck:
2927                         clear_opt (sbi->s_mount_opt, CHECK);
2928                         break;
2929 @@ -831,6 +849,8 @@ static int ext2_fill_super(struct super_
2930         if (!parse_options ((char *) data, sbi))
2931                 goto failed_mount;
2932
2933 +       if (EXT2_SB(sb)->s_mount_opt & EXT2_MOUNT_TAGGED)
2934 +               sb->s_flags |= MS_TAGGED;
2935         sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
2936                 ((EXT2_SB(sb)->s_mount_opt & EXT2_MOUNT_POSIX_ACL) ?
2937                  MS_POSIXACL : 0);
2938 @@ -1164,6 +1184,13 @@ static int ext2_remount (struct super_bl
2939                 goto restore_opts;
2940         }
2941
2942 +       if ((sbi->s_mount_opt & EXT2_MOUNT_TAGGED) &&
2943 +               !(sb->s_flags & MS_TAGGED)) {
2944 +               printk("EXT2-fs: %s: tagging not permitted on remount.\n",
2945 +                      sb->s_id);
2946 +               return -EINVAL;
2947 +       }
2948 +
2949         sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
2950                 ((sbi->s_mount_opt & EXT2_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0);
2951
2952 diff -NurpP --minimal linux-2.6.25.11/fs/ext2/symlink.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext2/symlink.c
2953 --- linux-2.6.25.11/fs/ext2/symlink.c   2008-04-17 10:32:27.000000000 -0400
2954 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext2/symlink.c     2008-04-19 15:14:52.000000000 -0400
2955 @@ -38,6 +38,7 @@ const struct inode_operations ext2_symli
2956         .listxattr      = ext2_listxattr,
2957         .removexattr    = generic_removexattr,
2958  #endif
2959 +       .sync_flags     = ext2_sync_flags,
2960  };
2961
2962  const struct inode_operations ext2_fast_symlink_inode_operations = {
2963 @@ -49,4 +50,5 @@ const struct inode_operations ext2_fast_
2964         .listxattr      = ext2_listxattr,
2965         .removexattr    = generic_removexattr,
2966  #endif
2967 +       .sync_flags     = ext2_sync_flags,
2968  };
2969 diff -NurpP --minimal linux-2.6.25.11/fs/ext2/xattr.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext2/xattr.c
2970 --- linux-2.6.25.11/fs/ext2/xattr.c     2008-04-17 12:05:40.000000000 -0400
2971 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext2/xattr.c       2008-04-19 15:14:52.000000000 -0400
2972 @@ -60,6 +60,7 @@
2973  #include <linux/mbcache.h>
2974  #include <linux/quotaops.h>
2975  #include <linux/rwsem.h>
2976 +#include <linux/vs_dlimit.h>
2977  #include "ext2.h"
2978  #include "xattr.h"
2979  #include "acl.h"
2980 @@ -641,8 +642,12 @@ ext2_xattr_set2(struct inode *inode, str
2981                                    the inode.  */
2982                                 ea_bdebug(new_bh, "reusing block");
2983
2984 +                               error = -ENOSPC;
2985 +                               if (DLIMIT_ALLOC_BLOCK(inode, 1))
2986 +                                       goto cleanup;
2987                                 error = -EDQUOT;
2988                                 if (DQUOT_ALLOC_BLOCK(inode, 1)) {
2989 +                                       DLIMIT_FREE_BLOCK(inode, 1);
2990                                         unlock_buffer(new_bh);
2991                                         goto cleanup;
2992                                 }
2993 @@ -735,6 +740,7 @@ ext2_xattr_set2(struct inode *inode, str
2994                                 le32_to_cpu(HDR(old_bh)->h_refcount) - 1);
2995                         if (ce)
2996                                 mb_cache_entry_release(ce);
2997 +                       DLIMIT_FREE_BLOCK(inode, 1);
2998                         DQUOT_FREE_BLOCK(inode, 1);
2999                         mark_buffer_dirty(old_bh);
3000                         ea_bdebug(old_bh, "refcount now=%d",
3001 @@ -799,6 +805,7 @@ ext2_xattr_delete_inode(struct inode *in
3002                 mark_buffer_dirty(bh);
3003                 if (IS_SYNC(inode))
3004                         sync_dirty_buffer(bh);
3005 +               DLIMIT_FREE_BLOCK(inode, 1);
3006                 DQUOT_FREE_BLOCK(inode, 1);
3007         }
3008         EXT2_I(inode)->i_file_acl = 0;
3009 diff -NurpP --minimal linux-2.6.25.11/fs/ext3/balloc.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext3/balloc.c
3010 --- linux-2.6.25.11/fs/ext3/balloc.c    2008-04-17 12:05:40.000000000 -0400
3011 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext3/balloc.c      2008-04-19 15:14:52.000000000 -0400
3012 @@ -19,6 +19,8 @@
3013  #include <linux/ext3_jbd.h>
3014  #include <linux/quotaops.h>
3015  #include <linux/buffer_head.h>
3016 +#include <linux/vs_dlimit.h>
3017 +#include <linux/vs_tag.h>
3018
3019  /*
3020   * balloc.c contains the blocks allocation and deallocation routines
3021 @@ -675,8 +677,10 @@ void ext3_free_blocks(handle_t *handle,
3022                 return;
3023         }
3024         ext3_free_blocks_sb(handle, sb, block, count, &dquot_freed_blocks);
3025 -       if (dquot_freed_blocks)
3026 +       if (dquot_freed_blocks) {
3027 +               DLIMIT_FREE_BLOCK(inode, dquot_freed_blocks);
3028                 DQUOT_FREE_BLOCK(inode, dquot_freed_blocks);
3029 +       }
3030         return;
3031  }
3032
3033 @@ -1415,18 +1419,33 @@ out:
3034   *
3035   * Check if filesystem has at least 1 free block available for allocation.
3036   */
3037 -static int ext3_has_free_blocks(struct ext3_sb_info *sbi)
3038 +static int ext3_has_free_blocks(struct super_block *sb)
3039  {
3040 -       ext3_fsblk_t free_blocks, root_blocks;
3041 +       struct ext3_sb_info *sbi = EXT3_SB(sb);
3042 +       unsigned long long free_blocks, root_blocks;
3043 +       int cond;
3044
3045         free_blocks = percpu_counter_read_positive(&sbi->s_freeblocks_counter);
3046         root_blocks = le32_to_cpu(sbi->s_es->s_r_blocks_count);
3047 -       if (free_blocks < root_blocks + 1 && !capable(CAP_SYS_RESOURCE) &&
3048 +
3049 +       vxdprintk(VXD_CBIT(dlim, 3),
3050 +               "ext3_has_free_blocks(%p): free=%llu, root=%llu",
3051 +               sb, free_blocks, root_blocks);
3052 +
3053 +       DLIMIT_ADJUST_BLOCK(sb, dx_current_tag(), &free_blocks, &root_blocks);
3054 +
3055 +       cond = (free_blocks < root_blocks + 1 &&
3056 +               !capable(CAP_SYS_RESOURCE) &&
3057                 sbi->s_resuid != current->fsuid &&
3058 -               (sbi->s_resgid == 0 || !in_group_p (sbi->s_resgid))) {
3059 -               return 0;
3060 -       }
3061 -       return 1;
3062 +               (sbi->s_resgid == 0 || !in_group_p (sbi->s_resgid)));
3063 +
3064 +       vxdprintk(VXD_CBIT(dlim, 3),
3065 +               "ext3_has_free_blocks(%p): %llu<%llu+1, %c, %u!=%u r=%d",
3066 +               sb, free_blocks, root_blocks,
3067 +               !capable(CAP_SYS_RESOURCE)?'1':'0',
3068 +               sbi->s_resuid, current->fsuid, cond?0:1);
3069 +
3070 +       return (cond ? 0 : 1);
3071  }
3072
3073  /**
3074 @@ -1443,7 +1462,7 @@ static int ext3_has_free_blocks(struct e
3075   */
3076  int ext3_should_retry_alloc(struct super_block *sb, int *retries)
3077  {
3078 -       if (!ext3_has_free_blocks(EXT3_SB(sb)) || (*retries)++ > 3)
3079 +       if (!ext3_has_free_blocks(sb) || (*retries)++ > 3)
3080                 return 0;
3081
3082         jbd_debug(1, "%s: retrying operation after ENOSPC\n", sb->s_id);
3083 @@ -1506,6 +1525,8 @@ ext3_fsblk_t ext3_new_blocks(handle_t *h
3084                 *errp = -EDQUOT;
3085                 return 0;
3086         }
3087 +       if (DLIMIT_ALLOC_BLOCK(inode, num))
3088 +           goto out_dlimit;
3089
3090         sbi = EXT3_SB(sb);
3091         es = EXT3_SB(sb)->s_es;
3092 @@ -1522,7 +1543,7 @@ ext3_fsblk_t ext3_new_blocks(handle_t *h
3093         if (block_i && ((windowsz = block_i->rsv_window_node.rsv_goal_size) > 0))
3094                 my_rsv = &block_i->rsv_window_node;
3095
3096 -       if (!ext3_has_free_blocks(sbi)) {
3097 +       if (!ext3_has_free_blocks(sb)) {
3098                 *errp = -ENOSPC;
3099                 goto out;
3100         }
3101 @@ -1710,12 +1731,16 @@ allocated:
3102         *errp = 0;
3103         brelse(bitmap_bh);
3104         DQUOT_FREE_BLOCK(inode, *count-num);
3105 +       DLIMIT_FREE_BLOCK(inode, *count-num);
3106         *count = num;
3107         return ret_block;
3108
3109  io_error:
3110         *errp = -EIO;
3111  out:
3112 +       if (!performed_allocation)
3113 +               DLIMIT_FREE_BLOCK(inode, *count);
3114 +out_dlimit:
3115         if (fatal) {
3116                 *errp = fatal;
3117                 ext3_std_error(sb, fatal);
3118 diff -NurpP --minimal linux-2.6.25.11/fs/ext3/file.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext3/file.c
3119 --- linux-2.6.25.11/fs/ext3/file.c      2008-04-17 10:37:23.000000000 -0400
3120 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext3/file.c        2008-04-19 15:14:52.000000000 -0400
3121 @@ -134,5 +134,6 @@ const struct inode_operations ext3_file_
3122         .removexattr    = generic_removexattr,
3123  #endif
3124         .permission     = ext3_permission,
3125 +       .sync_flags     = ext3_sync_flags,
3126  };
3127
3128 diff -NurpP --minimal linux-2.6.25.11/fs/ext3/ialloc.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext3/ialloc.c
3129 --- linux-2.6.25.11/fs/ext3/ialloc.c    2008-04-17 12:05:40.000000000 -0400
3130 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext3/ialloc.c      2008-07-08 04:31:24.000000000 -0400
3131 @@ -23,6 +23,8 @@
3132  #include <linux/buffer_head.h>
3133  #include <linux/random.h>
3134  #include <linux/bitops.h>
3135 +#include <linux/vs_dlimit.h>
3136 +#include <linux/vs_tag.h>
3137
3138  #include <asm/byteorder.h>
3139
3140 @@ -127,6 +129,7 @@ void ext3_free_inode (handle_t *handle,
3141         ext3_xattr_delete_inode(handle, inode);
3142         DQUOT_FREE_INODE(inode);
3143         DQUOT_DROP(inode);
3144 +       DLIMIT_FREE_INODE(inode);
3145
3146         is_directory = S_ISDIR(inode->i_mode);
3147
3148 @@ -440,6 +443,12 @@ struct inode *ext3_new_inode(handle_t *h
3149         inode = new_inode(sb);
3150         if (!inode)
3151                 return ERR_PTR(-ENOMEM);
3152 +
3153 +       inode->i_tag = dx_current_fstag(sb);
3154 +       if (DLIMIT_ALLOC_INODE(inode)) {
3155 +               err = -ENOSPC;
3156 +               goto out_dlimit;
3157 +       }
3158         ei = EXT3_I(inode);
3159
3160         sbi = EXT3_SB(sb);
3161 @@ -559,7 +568,8 @@ got:
3162         ei->i_dir_start_lookup = 0;
3163         ei->i_disksize = 0;
3164
3165 -       ei->i_flags = EXT3_I(dir)->i_flags & ~EXT3_INDEX_FL;
3166 +       ei->i_flags = EXT3_I(dir)->i_flags &
3167 +               ~(EXT3_INDEX_FL|EXT3_IXUNLINK_FL|EXT3_BARRIER_FL);
3168         if (S_ISLNK(mode))
3169                 ei->i_flags &= ~(EXT3_IMMUTABLE_FL|EXT3_APPEND_FL);
3170         /* dirsync only applies to directories */
3171 @@ -614,6 +624,8 @@ got:
3172  fail:
3173         ext3_std_error(sb, err);
3174  out:
3175 +       DLIMIT_FREE_INODE(inode);
3176 +out_dlimit:
3177         iput(inode);
3178         ret = ERR_PTR(err);
3179  really_out:
3180 @@ -625,6 +637,7 @@ fail_free_drop:
3181
3182  fail_drop:
3183         DQUOT_DROP(inode);
3184 +       DLIMIT_FREE_INODE(inode);
3185         inode->i_flags |= S_NOQUOTA;
3186         inode->i_nlink = 0;
3187         iput(inode);
3188 diff -NurpP --minimal linux-2.6.25.11/fs/ext3/inode.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext3/inode.c
3189 --- linux-2.6.25.11/fs/ext3/inode.c     2008-04-17 12:05:40.000000000 -0400
3190 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext3/inode.c       2008-07-08 16:53:39.000000000 -0400
3191 @@ -36,6 +36,7 @@
3192  #include <linux/mpage.h>
3193  #include <linux/uio.h>
3194  #include <linux/bio.h>
3195 +#include <linux/vs_tag.h>
3196  #include "xattr.h"
3197  #include "acl.h"
3198
3199 @@ -2300,7 +2301,7 @@ void ext3_truncate(struct inode *inode)
3200                 return;
3201         if (ext3_inode_is_fast_symlink(inode))
3202                 return;
3203 -       if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
3204 +       if (IS_APPEND(inode) || IS_IXORUNLINK(inode))
3205                 return;
3206
3207         /*
3208 @@ -2622,36 +2623,84 @@ void ext3_set_inode_flags(struct inode *
3209  {
3210         unsigned int flags = EXT3_I(inode)->i_flags;
3211
3212 -       inode->i_flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC);
3213 +       inode->i_flags &= ~(S_IMMUTABLE | S_IXUNLINK |
3214 +               S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
3215 +
3216 +       if (flags & EXT3_IMMUTABLE_FL)
3217 +               inode->i_flags |= S_IMMUTABLE;
3218 +       if (flags & EXT3_IXUNLINK_FL)
3219 +               inode->i_flags |= S_IXUNLINK;
3220 +
3221         if (flags & EXT3_SYNC_FL)
3222                 inode->i_flags |= S_SYNC;
3223         if (flags & EXT3_APPEND_FL)
3224                 inode->i_flags |= S_APPEND;
3225 -       if (flags & EXT3_IMMUTABLE_FL)
3226 -               inode->i_flags |= S_IMMUTABLE;
3227         if (flags & EXT3_NOATIME_FL)
3228                 inode->i_flags |= S_NOATIME;
3229         if (flags & EXT3_DIRSYNC_FL)
3230                 inode->i_flags |= S_DIRSYNC;
3231 +
3232 +       inode->i_vflags &= ~(V_BARRIER | V_COW);
3233 +
3234 +       if (flags & EXT3_BARRIER_FL)
3235 +               inode->i_vflags |= V_BARRIER;
3236 +       if (flags & EXT3_COW_FL)
3237 +               inode->i_vflags |= V_COW;
3238  }
3239
3240  /* Propagate flags from i_flags to EXT3_I(inode)->i_flags */
3241  void ext3_get_inode_flags(struct ext3_inode_info *ei)
3242  {
3243         unsigned int flags = ei->vfs_inode.i_flags;
3244 +       unsigned int vflags = ei->vfs_inode.i_vflags;
3245 +
3246 +       ei->i_flags &= ~(EXT3_SYNC_FL | EXT3_APPEND_FL |
3247 +                       EXT3_IMMUTABLE_FL | EXT3_IXUNLINK_FL |
3248 +                       EXT3_NOATIME_FL | EXT3_DIRSYNC_FL |
3249 +                       EXT3_BARRIER_FL | EXT3_COW_FL);
3250 +
3251 +       if (flags & S_IMMUTABLE)
3252 +               ei->i_flags |= EXT3_IMMUTABLE_FL;
3253 +       if (flags & S_IXUNLINK)
3254 +               ei->i_flags |= EXT3_IXUNLINK_FL;
3255
3256 -       ei->i_flags &= ~(EXT3_SYNC_FL|EXT3_APPEND_FL|
3257 -                       EXT3_IMMUTABLE_FL|EXT3_NOATIME_FL|EXT3_DIRSYNC_FL);
3258         if (flags & S_SYNC)
3259                 ei->i_flags |= EXT3_SYNC_FL;
3260         if (flags & S_APPEND)
3261                 ei->i_flags |= EXT3_APPEND_FL;
3262 -       if (flags & S_IMMUTABLE)
3263 -               ei->i_flags |= EXT3_IMMUTABLE_FL;
3264         if (flags & S_NOATIME)
3265                 ei->i_flags |= EXT3_NOATIME_FL;
3266         if (flags & S_DIRSYNC)
3267                 ei->i_flags |= EXT3_DIRSYNC_FL;
3268 +
3269 +       if (vflags & V_BARRIER)
3270 +               ei->i_flags |= EXT3_BARRIER_FL;
3271 +       if (vflags & V_COW)
3272 +               ei->i_flags |= EXT3_COW_FL;
3273 +}
3274 +
3275 +int ext3_sync_flags(struct inode *inode)
3276 +{
3277 +       struct ext3_iloc iloc;
3278 +       handle_t *handle;
3279 +       int err;
3280 +
3281 +       handle = ext3_journal_start(inode, 1);
3282 +       if (IS_ERR(handle))
3283 +               return PTR_ERR(handle);
3284 +       if (IS_SYNC(inode))
3285 +               handle->h_sync = 1;
3286 +       err = ext3_reserve_inode_write(handle, inode, &iloc);
3287 +       if (err)
3288 +               goto flags_err;
3289 +
3290 +       ext3_get_inode_flags(EXT3_I(inode));
3291 +       inode->i_ctime = CURRENT_TIME;
3292 +
3293 +       err = ext3_mark_iloc_dirty(handle, inode, &iloc);
3294 +flags_err:
3295 +       ext3_journal_stop(handle);
3296 +       return err;
3297  }
3298
3299  struct inode *ext3_iget(struct super_block *sb, unsigned long ino)
3300 @@ -2663,6 +2712,8 @@ struct inode *ext3_iget(struct super_blo
3301         struct inode *inode;
3302         long ret;
3303         int block;
3304 +       uid_t uid;
3305 +       gid_t gid;
3306
3307         inode = iget_locked(sb, ino);
3308         if (!inode)
3309 @@ -2683,12 +2734,17 @@ struct inode *ext3_iget(struct super_blo
3310         bh = iloc.bh;
3311         raw_inode = ext3_raw_inode(&iloc);
3312         inode->i_mode = le16_to_cpu(raw_inode->i_mode);
3313 -       inode->i_uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
3314 -       inode->i_gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
3315 +       uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
3316 +       gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
3317         if(!(test_opt (inode->i_sb, NO_UID32))) {
3318 -               inode->i_uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
3319 -               inode->i_gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
3320 +               uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
3321 +               gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
3322         }
3323 +       inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
3324 +       inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
3325 +       inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid,
3326 +               le16_to_cpu(raw_inode->i_raw_tag));
3327 +
3328         inode->i_nlink = le16_to_cpu(raw_inode->i_links_count);
3329         inode->i_size = le32_to_cpu(raw_inode->i_size);
3330         inode->i_atime.tv_sec = (signed)le32_to_cpu(raw_inode->i_atime);
3331 @@ -2817,6 +2873,8 @@ static int ext3_do_update_inode(handle_t
3332         struct ext3_inode *raw_inode = ext3_raw_inode(iloc);
3333         struct ext3_inode_info *ei = EXT3_I(inode);
3334         struct buffer_head *bh = iloc->bh;
3335 +       uid_t uid = TAGINO_UID(DX_TAG(inode), inode->i_uid, inode->i_tag);
3336 +       gid_t gid = TAGINO_GID(DX_TAG(inode), inode->i_gid, inode->i_tag);
3337         int err = 0, rc, block;
3338
3339         /* For fields not not tracking in the in-memory inode,
3340 @@ -2827,29 +2885,32 @@ static int ext3_do_update_inode(handle_t
3341         ext3_get_inode_flags(ei);
3342         raw_inode->i_mode = cpu_to_le16(inode->i_mode);
3343         if(!(test_opt(inode->i_sb, NO_UID32))) {
3344 -               raw_inode->i_uid_low = cpu_to_le16(low_16_bits(inode->i_uid));
3345 -               raw_inode->i_gid_low = cpu_to_le16(low_16_bits(inode->i_gid));
3346 +               raw_inode->i_uid_low = cpu_to_le16(low_16_bits(uid));
3347 +               raw_inode->i_gid_low = cpu_to_le16(low_16_bits(gid));
3348  /*
3349   * Fix up interoperability with old kernels. Otherwise, old inodes get
3350   * re-used with the upper 16 bits of the uid/gid intact
3351   */
3352                 if(!ei->i_dtime) {
3353                         raw_inode->i_uid_high =
3354 -                               cpu_to_le16(high_16_bits(inode->i_uid));
3355 +                               cpu_to_le16(high_16_bits(uid));
3356                         raw_inode->i_gid_high =
3357 -                               cpu_to_le16(high_16_bits(inode->i_gid));
3358 +                               cpu_to_le16(high_16_bits(gid));
3359                 } else {
3360                         raw_inode->i_uid_high = 0;
3361                         raw_inode->i_gid_high = 0;
3362                 }
3363         } else {
3364                 raw_inode->i_uid_low =
3365 -                       cpu_to_le16(fs_high2lowuid(inode->i_uid));
3366 +                       cpu_to_le16(fs_high2lowuid(uid));
3367                 raw_inode->i_gid_low =
3368 -                       cpu_to_le16(fs_high2lowgid(inode->i_gid));
3369 +                       cpu_to_le16(fs_high2lowgid(gid));
3370                 raw_inode->i_uid_high = 0;
3371                 raw_inode->i_gid_high = 0;
3372         }
3373 +#ifdef CONFIG_TAGGING_INTERN
3374 +       raw_inode->i_raw_tag = cpu_to_le16(inode->i_tag);
3375 +#endif
3376         raw_inode->i_links_count = cpu_to_le16(inode->i_nlink);
3377         raw_inode->i_size = cpu_to_le32(ei->i_disksize);
3378         raw_inode->i_atime = cpu_to_le32(inode->i_atime.tv_sec);
3379 @@ -3002,7 +3063,8 @@ int ext3_setattr(struct dentry *dentry,
3380                 return error;
3381
3382         if ((ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) ||
3383 -               (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) {
3384 +               (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid) ||
3385 +               (ia_valid & ATTR_TAG && attr->ia_tag != inode->i_tag)) {
3386                 handle_t *handle;
3387
3388                 /* (user+group)*(old+new) structure, inode write (sb,
3389 @@ -3024,6 +3086,8 @@ int ext3_setattr(struct dentry *dentry,
3390                         inode->i_uid = attr->ia_uid;
3391                 if (attr->ia_valid & ATTR_GID)
3392                         inode->i_gid = attr->ia_gid;
3393 +               if ((attr->ia_valid & ATTR_TAG) && IS_TAGGED(inode))
3394 +                       inode->i_tag = attr->ia_tag;
3395                 error = ext3_mark_inode_dirty(handle, inode);
3396                 ext3_journal_stop(handle);
3397         }
3398 diff -NurpP --minimal linux-2.6.25.11/fs/ext3/ioctl.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext3/ioctl.c
3399 --- linux-2.6.25.11/fs/ext3/ioctl.c     2008-04-17 11:31:35.000000000 -0400
3400 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext3/ioctl.c       2008-07-11 02:49:40.000000000 -0400
3401 @@ -8,6 +8,7 @@
3402   */
3403
3404  #include <linux/fs.h>
3405 +#include <linux/mount.h>
3406  #include <linux/jbd.h>
3407  #include <linux/capability.h>
3408  #include <linux/ext3_fs.h>
3409 @@ -15,6 +16,7 @@
3410  #include <linux/time.h>
3411  #include <linux/compat.h>
3412  #include <linux/smp_lock.h>
3413 +#include <linux/vs_tag.h>
3414  #include <asm/uaccess.h>
3415
3416  int ext3_ioctl (struct inode * inode, struct file * filp, unsigned int cmd,
3417 @@ -38,7 +40,8 @@ int ext3_ioctl (struct inode * inode, st
3418                 unsigned int oldflags;
3419                 unsigned int jflag;
3420
3421 -               if (IS_RDONLY(inode))
3422 +               if (IS_RDONLY(inode) ||
3423 +                       (filp && MNT_IS_RDONLY(filp->f_vfsmnt)))
3424                         return -EROFS;
3425
3426                 if (!is_owner_or_cap(inode))
3427 @@ -50,6 +53,11 @@ int ext3_ioctl (struct inode * inode, st
3428                 if (!S_ISDIR(inode->i_mode))
3429                         flags &= ~EXT3_DIRSYNC_FL;
3430
3431 +               if (IS_BARRIER(inode)) {
3432 +                       vxwprintk_task(1, "messing with the barrier.");
3433 +                       return -EACCES;
3434 +               }
3435 +
3436                 mutex_lock(&inode->i_mutex);
3437                 /* Is it quota file? Do not allow user to mess with it */
3438                 if (IS_NOQUOTA(inode)) {
3439 @@ -67,7 +75,9 @@ int ext3_ioctl (struct inode * inode, st
3440                  *
3441                  * This test looks nicer. Thanks to Pauline Middelink
3442                  */
3443 -               if ((flags ^ oldflags) & (EXT3_APPEND_FL | EXT3_IMMUTABLE_FL)) {
3444 +               if ((oldflags & EXT3_IMMUTABLE_FL) ||
3445 +                       ((flags ^ oldflags) & (EXT3_APPEND_FL |
3446 +                       EXT3_IMMUTABLE_FL | EXT3_IXUNLINK_FL))) {
3447                         if (!capable(CAP_LINUX_IMMUTABLE)) {
3448                                 mutex_unlock(&inode->i_mutex);
3449                                 return -EPERM;
3450 @@ -129,7 +139,8 @@ flags_err:
3451
3452                 if (!is_owner_or_cap(inode))
3453                         return -EPERM;
3454 -               if (IS_RDONLY(inode))
3455 +               if (IS_RDONLY(inode) ||
3456 +                       (filp && MNT_IS_RDONLY(filp->f_vfsmnt)))
3457                         return -EROFS;
3458                 if (get_user(generation, (int __user *) arg))
3459                         return -EFAULT;
3460 @@ -183,7 +194,8 @@ flags_err:
3461                 if (!test_opt(inode->i_sb, RESERVATION) ||!S_ISREG(inode->i_mode))
3462                         return -ENOTTY;
3463
3464 -               if (IS_RDONLY(inode))
3465 +               if (IS_RDONLY(inode) ||
3466 +                       (filp && MNT_IS_RDONLY(filp->f_vfsmnt)))
3467                         return -EROFS;
3468
3469                 if (!is_owner_or_cap(inode))
3470 @@ -218,7 +230,8 @@ flags_err:
3471                 if (!capable(CAP_SYS_RESOURCE))
3472                         return -EPERM;
3473
3474 -               if (IS_RDONLY(inode))
3475 +               if (IS_RDONLY(inode) ||
3476 +                       (filp && MNT_IS_RDONLY(filp->f_vfsmnt)))
3477                         return -EROFS;
3478
3479                 if (get_user(n_blocks_count, (__u32 __user *)arg))
3480 @@ -239,7 +252,8 @@ flags_err:
3481                 if (!capable(CAP_SYS_RESOURCE))
3482                         return -EPERM;
3483
3484 -               if (IS_RDONLY(inode))
3485 +               if (IS_RDONLY(inode) ||
3486 +                       (filp && MNT_IS_RDONLY(filp->f_vfsmnt)))
3487                         return -EROFS;
3488
3489                 if (copy_from_user(&input, (struct ext3_new_group_input __user *)arg,
3490 diff -NurpP --minimal linux-2.6.25.11/fs/ext3/namei.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext3/namei.c
3491 --- linux-2.6.25.11/fs/ext3/namei.c     2008-04-17 12:05:40.000000000 -0400
3492 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext3/namei.c       2008-04-19 15:14:52.000000000 -0400
3493 @@ -36,6 +36,7 @@
3494  #include <linux/quotaops.h>
3495  #include <linux/buffer_head.h>
3496  #include <linux/bio.h>
3497 +#include <linux/vs_tag.h>
3498
3499  #include "namei.h"
3500  #include "xattr.h"
3501 @@ -907,6 +908,7 @@ restart:
3502                                 if (bh)
3503                                         ll_rw_block(READ_META, 1, &bh);
3504                         }
3505 +               dx_propagate_tag(nd, inode);
3506                 }
3507                 if ((bh = bh_use[ra_ptr++]) == NULL)
3508                         goto next;
3509 @@ -2417,6 +2419,7 @@ const struct inode_operations ext3_dir_i
3510         .removexattr    = generic_removexattr,
3511  #endif
3512         .permission     = ext3_permission,
3513 +       .sync_flags     = ext3_sync_flags,
3514  };
3515
3516  const struct inode_operations ext3_special_inode_operations = {
3517 @@ -2428,4 +2431,5 @@ const struct inode_operations ext3_speci
3518         .removexattr    = generic_removexattr,
3519  #endif
3520         .permission     = ext3_permission,
3521 +       .sync_flags     = ext3_sync_flags,
3522  };
3523 diff -NurpP --minimal linux-2.6.25.11/fs/ext3/super.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext3/super.c
3524 --- linux-2.6.25.11/fs/ext3/super.c     2008-04-17 12:05:40.000000000 -0400
3525 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext3/super.c       2008-04-20 13:26:55.000000000 -0400
3526 @@ -756,7 +756,7 @@ enum {
3527         Opt_usrjquota, Opt_grpjquota, Opt_offusrjquota, Opt_offgrpjquota,
3528         Opt_jqfmt_vfsold, Opt_jqfmt_vfsv0, Opt_quota, Opt_noquota,
3529         Opt_ignore, Opt_barrier, Opt_err, Opt_resize, Opt_usrquota,
3530 -       Opt_grpquota
3531 +       Opt_grpquota, Opt_tag, Opt_notag, Opt_tagid
3532  };
3533
3534  static match_table_t tokens = {
3535 @@ -807,6 +807,9 @@ static match_table_t tokens = {
3536         {Opt_usrquota, "usrquota"},
3537         {Opt_barrier, "barrier=%u"},
3538         {Opt_resize, "resize"},
3539 +       {Opt_tag, "tag"},
3540 +       {Opt_notag, "notag"},
3541 +       {Opt_tagid, "tagid=%u"},
3542         {Opt_err, NULL},
3543  };
3544
3545 @@ -899,6 +902,20 @@ static int parse_options (char *options,
3546                 case Opt_nouid32:
3547                         set_opt (sbi->s_mount_opt, NO_UID32);
3548                         break;
3549 +#ifndef CONFIG_TAGGING_NONE
3550 +               case Opt_tag:
3551 +                       set_opt (sbi->s_mount_opt, TAGGED);
3552 +                       break;
3553 +               case Opt_notag:
3554 +                       clear_opt (sbi->s_mount_opt, TAGGED);
3555 +                       break;
3556 +#endif
3557 +#ifdef CONFIG_PROPAGATE
3558 +               case Opt_tagid:
3559 +                       /* use args[0] */
3560 +                       set_opt (sbi->s_mount_opt, TAGGED);
3561 +                       break;
3562 +#endif
3563                 case Opt_nocheck:
3564                         clear_opt (sbi->s_mount_opt, CHECK);
3565                         break;
3566 @@ -1591,6 +1608,9 @@ static int ext3_fill_super (struct super
3567                             NULL, 0))
3568                 goto failed_mount;
3569
3570 +       if (EXT3_SB(sb)->s_mount_opt & EXT3_MOUNT_TAGGED)
3571 +               sb->s_flags |= MS_TAGGED;
3572 +
3573         sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
3574                 ((sbi->s_mount_opt & EXT3_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0);
3575
3576 @@ -2428,6 +2448,12 @@ static int ext3_remount (struct super_bl
3577
3578         if (sbi->s_mount_opt & EXT3_MOUNT_ABORT)
3579                 ext3_abort(sb, __FUNCTION__, "Abort forced by user");
3580 +       if ((sbi->s_mount_opt & EXT3_MOUNT_TAGGED) &&
3581 +               !(sb->s_flags & MS_TAGGED)) {
3582 +               printk("EXT3-fs: %s: tagging not permitted on remount.\n",
3583 +                       sb->s_id);
3584 +               return -EINVAL;
3585 +       }
3586
3587         sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
3588                 ((sbi->s_mount_opt & EXT3_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0);
3589 diff -NurpP --minimal linux-2.6.25.11/fs/ext3/symlink.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext3/symlink.c
3590 --- linux-2.6.25.11/fs/ext3/symlink.c   2008-04-17 10:32:27.000000000 -0400
3591 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext3/symlink.c     2008-04-19 15:14:52.000000000 -0400
3592 @@ -40,6 +40,7 @@ const struct inode_operations ext3_symli
3593         .listxattr      = ext3_listxattr,
3594         .removexattr    = generic_removexattr,
3595  #endif
3596 +       .sync_flags     = ext3_sync_flags,
3597  };
3598
3599  const struct inode_operations ext3_fast_symlink_inode_operations = {
3600 @@ -51,4 +52,5 @@ const struct inode_operations ext3_fast_
3601         .listxattr      = ext3_listxattr,
3602         .removexattr    = generic_removexattr,
3603  #endif
3604 +       .sync_flags     = ext3_sync_flags,
3605  };
3606 diff -NurpP --minimal linux-2.6.25.11/fs/ext3/xattr.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext3/xattr.c
3607 --- linux-2.6.25.11/fs/ext3/xattr.c     2008-07-15 22:05:17.000000000 -0400
3608 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext3/xattr.c       2008-06-10 20:00:46.000000000 -0400
3609 @@ -58,6 +58,7 @@
3610  #include <linux/mbcache.h>
3611  #include <linux/quotaops.h>
3612  #include <linux/rwsem.h>
3613 +#include <linux/vs_dlimit.h>
3614  #include "xattr.h"
3615  #include "acl.h"
3616
3617 @@ -496,6 +497,7 @@ ext3_xattr_release_block(handle_t *handl
3618                 error = ext3_journal_dirty_metadata(handle, bh);
3619                 if (IS_SYNC(inode))
3620                         handle->h_sync = 1;
3621 +                       DLIMIT_FREE_BLOCK(inode, 1);
3622                 DQUOT_FREE_BLOCK(inode, 1);
3623                 ea_bdebug(bh, "refcount now=%d; releasing",
3624                           le32_to_cpu(BHDR(bh)->h_refcount));
3625 @@ -769,11 +771,14 @@ inserted:
3626                         if (new_bh == bs->bh)
3627                                 ea_bdebug(new_bh, "keeping");
3628                         else {
3629 +                               error = -ENOSPC;
3630 +                               if (DLIMIT_ALLOC_BLOCK(inode, 1))
3631 +                                       goto cleanup;
3632                                 /* The old block is released after updating
3633                                    the inode. */
3634                                 error = -EDQUOT;
3635                                 if (DQUOT_ALLOC_BLOCK(inode, 1))
3636 -                                       goto cleanup;
3637 +                                       goto cleanup_dlimit;
3638                                 error = ext3_journal_get_write_access(handle,
3639                                                                       new_bh);
3640                                 if (error)
3641 @@ -849,6 +854,8 @@ cleanup:
3642
3643  cleanup_dquot:
3644         DQUOT_FREE_BLOCK(inode, 1);
3645 +cleanup_dlimit:
3646 +       DLIMIT_FREE_BLOCK(inode, 1);
3647         goto cleanup;
3648
3649  bad_block:
3650 diff -NurpP --minimal linux-2.6.25.11/fs/ext4/balloc.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext4/balloc.c
3651 --- linux-2.6.25.11/fs/ext4/balloc.c    2008-04-17 12:05:40.000000000 -0400
3652 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext4/balloc.c      2008-04-21 10:23:33.000000000 -0400
3653 @@ -19,6 +19,8 @@
3654  #include <linux/ext4_jbd2.h>
3655  #include <linux/quotaops.h>
3656  #include <linux/buffer_head.h>
3657 +#include <linux/vs_dlimit.h>
3658 +#include <linux/vs_tag.h>
3659
3660  #include "group.h"
3661  /*
3662 @@ -810,8 +812,10 @@ void ext4_free_blocks(handle_t *handle,
3663         else
3664                 ext4_mb_free_blocks(handle, inode, block, count,
3665                                                 metadata, &dquot_freed_blocks);
3666 -       if (dquot_freed_blocks)
3667 +       if (dquot_freed_blocks) {
3668 +               DLIMIT_FREE_BLOCK(inode, dquot_freed_blocks);
3669                 DQUOT_FREE_BLOCK(inode, dquot_freed_blocks);
3670 +       }
3671         return;
3672  }
3673
3674 @@ -1551,18 +1555,33 @@ out:
3675   *
3676   * Check if filesystem has at least 1 free block available for allocation.
3677   */
3678 -static int ext4_has_free_blocks(struct ext4_sb_info *sbi)
3679 +static int ext4_has_free_blocks(struct super_block *sb)
3680  {
3681 +       struct ext4_sb_info *sbi = EXT4_SB(sb);
3682         ext4_fsblk_t free_blocks, root_blocks;
3683 +       int cond;
3684
3685         free_blocks = percpu_counter_read_positive(&sbi->s_freeblocks_counter);
3686         root_blocks = ext4_r_blocks_count(sbi->s_es);
3687 -       if (free_blocks < root_blocks + 1 && !capable(CAP_SYS_RESOURCE) &&
3688 +
3689 +       vxdprintk(VXD_CBIT(dlim, 3),
3690 +               "ext4_has_free_blocks(%p): free=%llu, root=%llu",
3691 +               sb, free_blocks, root_blocks);
3692 +
3693 +       DLIMIT_ADJUST_BLOCK(sb, dx_current_tag(), &free_blocks, &root_blocks);
3694 +
3695 +       cond = (free_blocks < root_blocks + 1 &&
3696 +               !capable(CAP_SYS_RESOURCE) &&
3697                 sbi->s_resuid != current->fsuid &&
3698 -               (sbi->s_resgid == 0 || !in_group_p (sbi->s_resgid))) {
3699 -               return 0;
3700 -       }
3701 -       return 1;
3702 +               (sbi->s_resgid == 0 || !in_group_p (sbi->s_resgid)));
3703 +
3704 +       vxdprintk(VXD_CBIT(dlim, 3),
3705 +               "ext4_has_free_blocks(%p): %llu<%llu+1, %c, %u!=%u r=%d",
3706 +               sb, free_blocks, root_blocks,
3707 +               !capable(CAP_SYS_RESOURCE)?'1':'0',
3708 +               sbi->s_resuid, current->fsuid, cond?0:1);
3709 +
3710 +       return (cond ? 0 : 1);
3711  }
3712
3713  /**
3714 @@ -1579,7 +1598,7 @@ static int ext4_has_free_blocks(struct e
3715   */
3716  int ext4_should_retry_alloc(struct super_block *sb, int *retries)
3717  {
3718 -       if (!ext4_has_free_blocks(EXT4_SB(sb)) || (*retries)++ > 3)
3719 +       if (!ext4_has_free_blocks(sb) || (*retries)++ > 3)
3720                 return 0;
3721
3722         jbd_debug(1, "%s: retrying operation after ENOSPC\n", sb->s_id);
3723 @@ -1639,6 +1658,8 @@ ext4_fsblk_t ext4_new_blocks_old(handle_
3724                 *errp = -EDQUOT;
3725                 return 0;
3726         }
3727 +       if (DLIMIT_ALLOC_BLOCK(inode, num))
3728 +           goto out_dlimit;
3729
3730         sbi = EXT4_SB(sb);
3731         es = EXT4_SB(sb)->s_es;
3732 @@ -1655,7 +1676,7 @@ ext4_fsblk_t ext4_new_blocks_old(handle_
3733         if (block_i && ((windowsz = block_i->rsv_window_node.rsv_goal_size) > 0))
3734                 my_rsv = &block_i->rsv_window_node;
3735
3736 -       if (!ext4_has_free_blocks(sbi)) {
3737 +       if (!ext4_has_free_blocks(sb)) {
3738                 *errp = -ENOSPC;
3739                 goto out;
3740         }
3741 @@ -1841,12 +1862,16 @@ allocated:
3742         *errp = 0;
3743         brelse(bitmap_bh);
3744         DQUOT_FREE_BLOCK(inode, *count-num);
3745 +       DLIMIT_FREE_BLOCK(inode, *count-num);
3746         *count = num;
3747         return ret_block;
3748
3749  io_error:
3750         *errp = -EIO;
3751  out:
3752 +       if (!performed_allocation)
3753 +               DLIMIT_FREE_BLOCK(inode, *count);
3754 +out_dlimit:
3755         if (fatal) {
3756                 *errp = fatal;
3757                 ext4_std_error(sb, fatal);
3758 diff -NurpP --minimal linux-2.6.25.11/fs/ext4/file.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext4/file.c
3759 --- linux-2.6.25.11/fs/ext4/file.c      2008-04-17 12:05:40.000000000 -0400
3760 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext4/file.c        2008-04-19 15:14:52.000000000 -0400
3761 @@ -152,5 +152,6 @@ const struct inode_operations ext4_file_
3762  #endif
3763         .permission     = ext4_permission,
3764         .fallocate      = ext4_fallocate,
3765 +       .sync_flags     = ext4_sync_flags,
3766  };
3767
3768 diff -NurpP --minimal linux-2.6.25.11/fs/ext4/ialloc.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext4/ialloc.c
3769 --- linux-2.6.25.11/fs/ext4/ialloc.c    2008-04-17 12:05:40.000000000 -0400
3770 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext4/ialloc.c      2008-07-08 04:31:24.000000000 -0400
3771 @@ -24,6 +24,8 @@
3772  #include <linux/random.h>
3773  #include <linux/bitops.h>
3774  #include <linux/blkdev.h>
3775 +#include <linux/vs_dlimit.h>
3776 +#include <linux/vs_tag.h>
3777  #include <asm/byteorder.h>
3778
3779  #include "xattr.h"
3780 @@ -186,6 +188,7 @@ void ext4_free_inode (handle_t *handle,
3781         ext4_xattr_delete_inode(handle, inode);
3782         DQUOT_FREE_INODE(inode);
3783         DQUOT_DROP(inode);
3784 +       DLIMIT_FREE_INODE(inode);
3785
3786         is_directory = S_ISDIR(inode->i_mode);
3787
3788 @@ -513,6 +516,12 @@ struct inode *ext4_new_inode(handle_t *h
3789         inode = new_inode(sb);
3790         if (!inode)
3791                 return ERR_PTR(-ENOMEM);
3792 +
3793 +       inode->i_tag = dx_current_fstag(sb);
3794 +       if (DLIMIT_ALLOC_INODE(inode)) {
3795 +               err = -ENOSPC;
3796 +               goto out_dlimit;
3797 +       }
3798         ei = EXT4_I(inode);
3799
3800         sbi = EXT4_SB(sb);
3801 @@ -707,7 +716,8 @@ got:
3802          * newly created directory and file only if -o extent mount option is
3803          * specified
3804          */
3805 -       ei->i_flags = EXT4_I(dir)->i_flags & ~(EXT4_INDEX_FL|EXT4_EXTENTS_FL);
3806 +       ei->i_flags = EXT4_I(dir)->i_flags &
3807 +               ~(EXT4_INDEX_FL|EXT4_EXTENTS_FL|EXT4_IXUNLINK_FL|EXT4_BARRIER_FL);
3808         if (S_ISLNK(mode))
3809                 ei->i_flags &= ~(EXT4_IMMUTABLE_FL|EXT4_APPEND_FL);
3810         /* dirsync only applies to directories */
3811 @@ -766,6 +776,8 @@ got:
3812  fail:
3813         ext4_std_error(sb, err);
3814  out:
3815 +       DLIMIT_FREE_INODE(inode);
3816 +out_dlimit:
3817         iput(inode);
3818         ret = ERR_PTR(err);
3819  really_out:
3820 @@ -777,6 +789,7 @@ fail_free_drop:
3821
3822  fail_drop:
3823         DQUOT_DROP(inode);
3824 +       DLIMIT_FREE_INODE(inode);
3825         inode->i_flags |= S_NOQUOTA;
3826         inode->i_nlink = 0;
3827         iput(inode);
3828 diff -NurpP --minimal linux-2.6.25.11/fs/ext4/inode.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext4/inode.c
3829 --- linux-2.6.25.11/fs/ext4/inode.c     2008-04-17 12:05:40.000000000 -0400
3830 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext4/inode.c       2008-07-11 03:12:05.000000000 -0400
3831 @@ -36,6 +36,7 @@
3832  #include <linux/mpage.h>
3833  #include <linux/uio.h>
3834  #include <linux/bio.h>
3835 +#include <linux/vs_tag.h>
3836  #include "xattr.h"
3837  #include "acl.h"
3838
3839 @@ -2342,7 +2343,7 @@ void ext4_truncate(struct inode *inode)
3840                 return;
3841         if (ext4_inode_is_fast_symlink(inode))
3842                 return;
3843 -       if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
3844 +       if (IS_APPEND(inode) || IS_IXORUNLINK(inode))
3845                 return;
3846
3847         /*
3848 @@ -2671,37 +2672,86 @@ void ext4_set_inode_flags(struct inode *
3849  {
3850         unsigned int flags = EXT4_I(inode)->i_flags;
3851
3852 -       inode->i_flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC);
3853 +       inode->i_flags &= ~(S_IMMUTABLE | S_IXUNLINK |
3854 +               S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
3855 +
3856 +       if (flags & EXT4_IMMUTABLE_FL)
3857 +               inode->i_flags |= S_IMMUTABLE;
3858 +       if (flags & EXT4_IXUNLINK_FL)
3859 +               inode->i_flags |= S_IXUNLINK;
3860 +
3861         if (flags & EXT4_SYNC_FL)
3862                 inode->i_flags |= S_SYNC;
3863         if (flags & EXT4_APPEND_FL)
3864                 inode->i_flags |= S_APPEND;
3865 -       if (flags & EXT4_IMMUTABLE_FL)
3866 -               inode->i_flags |= S_IMMUTABLE;
3867         if (flags & EXT4_NOATIME_FL)
3868                 inode->i_flags |= S_NOATIME;
3869         if (flags & EXT4_DIRSYNC_FL)
3870                 inode->i_flags |= S_DIRSYNC;
3871 +
3872 +       inode->i_vflags &= ~(V_BARRIER | V_COW);
3873 +
3874 +       if (flags & EXT4_BARRIER_FL)
3875 +               inode->i_vflags |= V_BARRIER;
3876 +       if (flags & EXT4_COW_FL)
3877 +               inode->i_vflags |= V_COW;
3878  }
3879
3880  /* Propagate flags from i_flags to EXT4_I(inode)->i_flags */
3881  void ext4_get_inode_flags(struct ext4_inode_info *ei)
3882  {
3883         unsigned int flags = ei->vfs_inode.i_flags;
3884 +       unsigned int vflags = ei->vfs_inode.i_vflags;
3885 +
3886 +       ei->i_flags &= ~(EXT4_SYNC_FL | EXT4_APPEND_FL |
3887 +                       EXT4_IMMUTABLE_FL | EXT4_IXUNLINK_FL |
3888 +                       EXT4_NOATIME_FL | EXT4_DIRSYNC_FL |
3889 +                       EXT4_BARRIER_FL | EXT4_COW_FL);
3890 +
3891 +       if (flags & S_IMMUTABLE)
3892 +               ei->i_flags |= EXT4_IMMUTABLE_FL;
3893 +       if (flags & S_IXUNLINK)
3894 +               ei->i_flags |= EXT4_IXUNLINK_FL;
3895
3896 -       ei->i_flags &= ~(EXT4_SYNC_FL|EXT4_APPEND_FL|
3897 -                       EXT4_IMMUTABLE_FL|EXT4_NOATIME_FL|EXT4_DIRSYNC_FL);
3898         if (flags & S_SYNC)
3899                 ei->i_flags |= EXT4_SYNC_FL;
3900         if (flags & S_APPEND)
3901                 ei->i_flags |= EXT4_APPEND_FL;
3902 -       if (flags & S_IMMUTABLE)
3903 -               ei->i_flags |= EXT4_IMMUTABLE_FL;
3904         if (flags & S_NOATIME)
3905                 ei->i_flags |= EXT4_NOATIME_FL;
3906         if (flags & S_DIRSYNC)
3907                 ei->i_flags |= EXT4_DIRSYNC_FL;
3908 +
3909 +       if (vflags & V_BARRIER)
3910 +               ei->i_flags |= EXT4_BARRIER_FL;
3911 +       if (vflags & V_COW)
3912 +               ei->i_flags |= EXT4_COW_FL;
3913  }
3914 +
3915 +int ext4_sync_flags(struct inode *inode)
3916 +{
3917 +       struct ext4_iloc iloc;
3918 +       handle_t *handle;
3919 +       int err;
3920 +
3921 +       handle = ext4_journal_start(inode, 1);
3922 +       if (IS_ERR(handle))
3923 +               return PTR_ERR(handle);
3924 +       if (IS_SYNC(inode))
3925 +               handle->h_sync = 1;
3926 +       err = ext4_reserve_inode_write(handle, inode, &iloc);
3927 +       if (err)
3928 +               goto flags_err;
3929 +
3930 +       ext4_get_inode_flags(EXT4_I(inode));
3931 +       inode->i_ctime = CURRENT_TIME;
3932 +
3933 +       err = ext4_mark_iloc_dirty(handle, inode, &iloc);
3934 +flags_err:
3935 +       ext4_journal_stop(handle);
3936 +       return err;
3937 +}
3938 +
3939  static blkcnt_t ext4_inode_blocks(struct ext4_inode *raw_inode,
3940                                         struct ext4_inode_info *ei)
3941  {
3942 @@ -2734,6 +2784,8 @@ struct inode *ext4_iget(struct super_blo
3943         struct inode *inode;
3944         long ret;
3945         int block;
3946 +       uid_t uid;
3947 +       gid_t gid;
3948
3949         inode = iget_locked(sb, ino);
3950         if (!inode)
3951 @@ -2754,12 +2806,17 @@ struct inode *ext4_iget(struct super_blo
3952         bh = iloc.bh;
3953         raw_inode = ext4_raw_inode(&iloc);
3954         inode->i_mode = le16_to_cpu(raw_inode->i_mode);
3955 -       inode->i_uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
3956 -       inode->i_gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
3957 +       uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
3958 +       gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
3959         if(!(test_opt (inode->i_sb, NO_UID32))) {
3960 -               inode->i_uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
3961 -               inode->i_gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
3962 +               uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
3963 +               gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
3964         }
3965 +       inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
3966 +       inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
3967 +       inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid,
3968 +               le16_to_cpu(raw_inode->i_raw_tag));
3969 +
3970         inode->i_nlink = le16_to_cpu(raw_inode->i_links_count);
3971
3972         ei->i_state = 0;
3973 @@ -2933,6 +2990,8 @@ static int ext4_do_update_inode(handle_t
3974         struct ext4_inode *raw_inode = ext4_raw_inode(iloc);
3975         struct ext4_inode_info *ei = EXT4_I(inode);
3976         struct buffer_head *bh = iloc->bh;
3977 +       uid_t uid = TAGINO_UID(DX_TAG(inode), inode->i_uid, inode->i_tag);
3978 +       gid_t gid = TAGINO_GID(DX_TAG(inode), inode->i_gid, inode->i_tag);
3979         int err = 0, rc, block;
3980
3981         /* For fields not not tracking in the in-memory inode,
3982 @@ -2943,29 +3002,32 @@ static int ext4_do_update_inode(handle_t
3983         ext4_get_inode_flags(ei);
3984         raw_inode->i_mode = cpu_to_le16(inode->i_mode);
3985         if(!(test_opt(inode->i_sb, NO_UID32))) {
3986 -               raw_inode->i_uid_low = cpu_to_le16(low_16_bits(inode->i_uid));
3987 -               raw_inode->i_gid_low = cpu_to_le16(low_16_bits(inode->i_gid));
3988 +               raw_inode->i_uid_low = cpu_to_le16(low_16_bits(uid));
3989 +               raw_inode->i_gid_low = cpu_to_le16(low_16_bits(gid));
3990  /*
3991   * Fix up interoperability with old kernels. Otherwise, old inodes get
3992   * re-used with the upper 16 bits of the uid/gid intact
3993   */
3994                 if(!ei->i_dtime) {
3995                         raw_inode->i_uid_high =
3996 -                               cpu_to_le16(high_16_bits(inode->i_uid));
3997 +                               cpu_to_le16(high_16_bits(uid));
3998                         raw_inode->i_gid_high =
3999 -                               cpu_to_le16(high_16_bits(inode->i_gid));
4000 +                               cpu_to_le16(high_16_bits(gid));
4001                 } else {
4002                         raw_inode->i_uid_high = 0;
4003                         raw_inode->i_gid_high = 0;
4004                 }
4005         } else {
4006                 raw_inode->i_uid_low =
4007 -                       cpu_to_le16(fs_high2lowuid(inode->i_uid));
4008 +                       cpu_to_le16(fs_high2lowuid(uid));
4009                 raw_inode->i_gid_low =
4010 -                       cpu_to_le16(fs_high2lowgid(inode->i_gid));
4011 +                       cpu_to_le16(fs_high2lowgid(gid));
4012                 raw_inode->i_uid_high = 0;
4013                 raw_inode->i_gid_high = 0;
4014         }
4015 +#ifdef CONFIG_TAGGING_INTERN
4016 +       raw_inode->i_raw_tag = cpu_to_le16(inode->i_tag);
4017 +#endif
4018         raw_inode->i_links_count = cpu_to_le16(inode->i_nlink);
4019
4020         EXT4_INODE_SET_XTIME(i_ctime, inode, raw_inode);
4021 @@ -3121,7 +3183,8 @@ int ext4_setattr(struct dentry *dentry,
4022                 return error;
4023
4024         if ((ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) ||
4025 -               (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) {
4026 +               (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid) ||
4027 +               (ia_valid & ATTR_TAG && attr->ia_tag != inode->i_tag)) {
4028                 handle_t *handle;
4029
4030                 /* (user+group)*(old+new) structure, inode write (sb,
4031 @@ -3143,6 +3206,8 @@ int ext4_setattr(struct dentry *dentry,
4032                         inode->i_uid = attr->ia_uid;
4033                 if (attr->ia_valid & ATTR_GID)
4034                         inode->i_gid = attr->ia_gid;
4035 +               if ((attr->ia_valid & ATTR_TAG) && IS_TAGGED(inode))
4036 +                       inode->i_tag = attr->ia_tag;
4037                 error = ext4_mark_inode_dirty(handle, inode);
4038                 ext4_journal_stop(handle);
4039         }
4040 diff -NurpP --minimal linux-2.6.25.11/fs/ext4/ioctl.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext4/ioctl.c
4041 --- linux-2.6.25.11/fs/ext4/ioctl.c     2008-04-17 12:05:40.000000000 -0400
4042 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext4/ioctl.c       2008-07-11 02:49:54.000000000 -0400
4043 @@ -8,6 +8,7 @@
4044   */
4045
4046  #include <linux/fs.h>
4047 +#include <linux/mount.h>
4048  #include <linux/jbd2.h>
4049  #include <linux/capability.h>
4050  #include <linux/ext4_fs.h>
4051 @@ -15,6 +16,7 @@
4052  #include <linux/time.h>
4053  #include <linux/compat.h>
4054  #include <linux/smp_lock.h>
4055 +#include <linux/vs_tag.h>
4056  #include <asm/uaccess.h>
4057
4058  int ext4_ioctl (struct inode * inode, struct file * filp, unsigned int cmd,
4059 @@ -38,7 +40,8 @@ int ext4_ioctl (struct inode * inode, st
4060                 unsigned int oldflags;
4061                 unsigned int jflag;
4062
4063 -               if (IS_RDONLY(inode))
4064 +               if (IS_RDONLY(inode) ||
4065 +                       (filp && MNT_IS_RDONLY(filp->f_vfsmnt)))
4066                         return -EROFS;
4067
4068                 if (!is_owner_or_cap(inode))
4069 @@ -50,6 +53,11 @@ int ext4_ioctl (struct inode * inode, st
4070                 if (!S_ISDIR(inode->i_mode))
4071                         flags &= ~EXT4_DIRSYNC_FL;
4072
4073 +               if (IS_BARRIER(inode)) {
4074 +                       vxwprintk_task(1, "messing with the barrier.");
4075 +                       return -EACCES;
4076 +               }
4077 +
4078                 mutex_lock(&inode->i_mutex);
4079                 /* Is it quota file? Do not allow user to mess with it */
4080                 if (IS_NOQUOTA(inode)) {
4081 @@ -67,7 +75,9 @@ int ext4_ioctl (struct inode * inode, st
4082                  *
4083                  * This test looks nicer. Thanks to Pauline Middelink
4084                  */
4085 -               if ((flags ^ oldflags) & (EXT4_APPEND_FL | EXT4_IMMUTABLE_FL)) {
4086 +               if ((oldflags & EXT4_IMMUTABLE_FL) ||
4087 +                       ((flags ^ oldflags) & (EXT4_APPEND_FL |
4088 +                       EXT4_IMMUTABLE_FL | EXT4_IXUNLINK_FL))) {
4089                         if (!capable(CAP_LINUX_IMMUTABLE)) {
4090                                 mutex_unlock(&inode->i_mutex);
4091                                 return -EPERM;
4092 @@ -129,7 +139,8 @@ flags_err:
4093
4094                 if (!is_owner_or_cap(inode))
4095                         return -EPERM;
4096 -               if (IS_RDONLY(inode))
4097 +               if (IS_RDONLY(inode) ||
4098 +                       (filp && MNT_IS_RDONLY(filp->f_vfsmnt)))
4099                         return -EROFS;
4100                 if (get_user(generation, (int __user *) arg))
4101                         return -EFAULT;
4102 @@ -183,7 +194,8 @@ flags_err:
4103                 if (!test_opt(inode->i_sb, RESERVATION) ||!S_ISREG(inode->i_mode))
4104                         return -ENOTTY;
4105
4106 -               if (IS_RDONLY(inode))
4107 +               if (IS_RDONLY(inode) ||
4108 +                       (filp && MNT_IS_RDONLY(filp->f_vfsmnt)))
4109                         return -EROFS;
4110
4111                 if (!is_owner_or_cap(inode))
4112 @@ -218,7 +230,8 @@ flags_err:
4113                 if (!capable(CAP_SYS_RESOURCE))
4114                         return -EPERM;
4115
4116 -               if (IS_RDONLY(inode))
4117 +               if (IS_RDONLY(inode) ||
4118 +                       (filp && MNT_IS_RDONLY(filp->f_vfsmnt)))
4119                         return -EROFS;
4120
4121                 if (get_user(n_blocks_count, (__u32 __user *)arg))
4122 @@ -239,7 +252,8 @@ flags_err:
4123                 if (!capable(CAP_SYS_RESOURCE))
4124                         return -EPERM;
4125
4126 -               if (IS_RDONLY(inode))
4127 +               if (IS_RDONLY(inode) ||
4128 +                       (filp && MNT_IS_RDONLY(filp->f_vfsmnt)))
4129                         return -EROFS;
4130
4131                 if (copy_from_user(&input, (struct ext4_new_group_input __user *)arg,
4132 diff -NurpP --minimal linux-2.6.25.11/fs/ext4/namei.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext4/namei.c
4133 --- linux-2.6.25.11/fs/ext4/namei.c     2008-04-17 12:05:40.000000000 -0400
4134 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext4/namei.c       2008-04-19 15:14:52.000000000 -0400
4135 @@ -36,6 +36,7 @@
4136  #include <linux/quotaops.h>
4137  #include <linux/buffer_head.h>
4138  #include <linux/bio.h>
4139 +#include <linux/vs_tag.h>
4140
4141  #include "namei.h"
4142  #include "xattr.h"
4143 @@ -908,6 +909,7 @@ restart:
4144                                 if (bh)
4145                                         ll_rw_block(READ_META, 1, &bh);
4146                         }
4147 +               dx_propagate_tag(nd, inode);
4148                 }
4149                 if ((bh = bh_use[ra_ptr++]) == NULL)
4150                         goto next;
4151 @@ -2448,6 +2450,7 @@ const struct inode_operations ext4_dir_i
4152         .removexattr    = generic_removexattr,
4153  #endif
4154         .permission     = ext4_permission,
4155 +       .sync_flags     = ext4_sync_flags,
4156  };
4157
4158  const struct inode_operations ext4_special_inode_operations = {
4159 @@ -2459,4 +2462,5 @@ const struct inode_operations ext4_speci
4160         .removexattr    = generic_removexattr,
4161  #endif
4162         .permission     = ext4_permission,
4163 +       .sync_flags     = ext4_sync_flags,
4164  };
4165 diff -NurpP --minimal linux-2.6.25.11/fs/ext4/super.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext4/super.c
4166 --- linux-2.6.25.11/fs/ext4/super.c     2008-04-17 12:05:40.000000000 -0400
4167 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext4/super.c       2008-04-21 10:31:22.000000000 -0400
4168 @@ -887,6 +887,7 @@ enum {
4169         Opt_ignore, Opt_barrier, Opt_err, Opt_resize, Opt_usrquota,
4170         Opt_grpquota, Opt_extents, Opt_noextents, Opt_i_version,
4171         Opt_mballoc, Opt_nomballoc, Opt_stripe,
4172 +       Opt_tag, Opt_notag, Opt_tagid
4173  };
4174
4175  static match_table_t tokens = {
4176 @@ -944,8 +945,11 @@ static match_table_t tokens = {
4177         {Opt_mballoc, "mballoc"},
4178         {Opt_nomballoc, "nomballoc"},
4179         {Opt_stripe, "stripe=%u"},
4180 -       {Opt_err, NULL},
4181         {Opt_resize, "resize"},
4182 +       {Opt_tag, "tag"},
4183 +       {Opt_notag, "notag"},
4184 +       {Opt_tagid, "tagid=%u"},
4185 +       {Opt_err, NULL},
4186  };
4187
4188  static ext4_fsblk_t get_sb_block(void **data)
4189 @@ -1037,6 +1041,20 @@ static int parse_options (char *options,
4190                 case Opt_nouid32:
4191                         set_opt (sbi->s_mount_opt, NO_UID32);
4192                         break;
4193 +#ifndef CONFIG_TAGGING_NONE
4194 +               case Opt_tag:
4195 +                       set_opt (sbi->s_mount_opt, TAGGED);
4196 +                       break;
4197 +               case Opt_notag:
4198 +                       clear_opt (sbi->s_mount_opt, TAGGED);
4199 +                       break;
4200 +#endif
4201 +#ifdef CONFIG_PROPAGATE
4202 +               case Opt_tagid:
4203 +                       /* use args[0] */
4204 +                       set_opt (sbi->s_mount_opt, TAGGED);
4205 +                       break;
4206 +#endif
4207                 case Opt_nocheck:
4208                         clear_opt (sbi->s_mount_opt, CHECK);
4209                         break;
4210 @@ -1909,6 +1927,9 @@ static int ext4_fill_super (struct super
4211                             NULL, 0))
4212                 goto failed_mount;
4213
4214 +       if (EXT4_SB(sb)->s_mount_opt & EXT4_MOUNT_TAGGED)
4215 +               sb->s_flags |= MS_TAGGED;
4216 +
4217         sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
4218                 ((sbi->s_mount_opt & EXT4_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0);
4219
4220 @@ -2828,6 +2849,12 @@ static int ext4_remount (struct super_bl
4221
4222         if (sbi->s_mount_opt & EXT4_MOUNT_ABORT)
4223                 ext4_abort(sb, __FUNCTION__, "Abort forced by user");
4224 +       if ((sbi->s_mount_opt & EXT4_MOUNT_TAGGED) &&
4225 +               !(sb->s_flags & MS_TAGGED)) {
4226 +               printk("EXT4-fs: %s: tagging not permitted on remount.\n",
4227 +                       sb->s_id);
4228 +               return -EINVAL;
4229 +       }
4230
4231         sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
4232                 ((sbi->s_mount_opt & EXT4_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0);
4233 diff -NurpP --minimal linux-2.6.25.11/fs/ext4/symlink.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext4/symlink.c
4234 --- linux-2.6.25.11/fs/ext4/symlink.c   2008-04-17 10:32:27.000000000 -0400
4235 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext4/symlink.c     2008-04-19 15:14:52.000000000 -0400
4236 @@ -40,6 +40,7 @@ const struct inode_operations ext4_symli
4237         .listxattr      = ext4_listxattr,
4238         .removexattr    = generic_removexattr,
4239  #endif
4240 +       .sync_flags     = ext4_sync_flags,
4241  };
4242
4243  const struct inode_operations ext4_fast_symlink_inode_operations = {
4244 @@ -51,4 +52,5 @@ const struct inode_operations ext4_fast_
4245         .listxattr      = ext4_listxattr,
4246         .removexattr    = generic_removexattr,
4247  #endif
4248 +       .sync_flags     = ext4_sync_flags,
4249  };
4250 diff -NurpP --minimal linux-2.6.25.11/fs/ext4/xattr.c linux-2.6.25.11-vs2.3.0.34.14/fs/ext4/xattr.c
4251 --- linux-2.6.25.11/fs/ext4/xattr.c     2008-07-15 22:05:17.000000000 -0400
4252 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ext4/xattr.c       2008-06-10 20:00:46.000000000 -0400
4253 @@ -58,6 +58,7 @@
4254  #include <linux/mbcache.h>
4255  #include <linux/quotaops.h>
4256  #include <linux/rwsem.h>
4257 +#include <linux/vs_dlimit.h>
4258  #include "xattr.h"
4259  #include "acl.h"
4260
4261 @@ -489,6 +490,7 @@ ext4_xattr_release_block(handle_t *handl
4262                 error = ext4_journal_dirty_metadata(handle, bh);
4263                 if (IS_SYNC(inode))
4264                         handle->h_sync = 1;
4265 +                       DLIMIT_FREE_BLOCK(inode, 1);
4266                 DQUOT_FREE_BLOCK(inode, 1);
4267                 ea_bdebug(bh, "refcount now=%d; releasing",
4268                           le32_to_cpu(BHDR(bh)->h_refcount));
4269 @@ -779,11 +781,14 @@ inserted:
4270                         if (new_bh == bs->bh)
4271                                 ea_bdebug(new_bh, "keeping");
4272                         else {
4273 +                               error = -ENOSPC;
4274 +                               if (DLIMIT_ALLOC_BLOCK(inode, 1))
4275 +                                       goto cleanup;
4276                                 /* The old block is released after updating
4277                                    the inode. */
4278                                 error = -EDQUOT;
4279                                 if (DQUOT_ALLOC_BLOCK(inode, 1))
4280 -                                       goto cleanup;
4281 +                                       goto cleanup_dlimit;
4282                                 error = ext4_journal_get_write_access(handle,
4283                                                                       new_bh);
4284                                 if (error)
4285 @@ -860,6 +865,8 @@ cleanup:
4286
4287  cleanup_dquot:
4288         DQUOT_FREE_BLOCK(inode, 1);
4289 +cleanup_dlimit:
4290 +       DLIMIT_FREE_BLOCK(inode, 1);
4291         goto cleanup;
4292
4293  bad_block:
4294 diff -NurpP --minimal linux-2.6.25.11/fs/fcntl.c linux-2.6.25.11-vs2.3.0.34.14/fs/fcntl.c
4295 --- linux-2.6.25.11/fs/fcntl.c  2008-04-17 12:05:40.000000000 -0400
4296 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/fcntl.c    2008-04-19 15:14:52.000000000 -0400
4297 @@ -19,6 +19,7 @@
4298  #include <linux/signal.h>
4299  #include <linux/rcupdate.h>
4300  #include <linux/pid_namespace.h>
4301 +#include <linux/vs_limit.h>
4302
4303  #include <asm/poll.h>
4304  #include <asm/siginfo.h>
4305 @@ -85,6 +86,8 @@ repeat:
4306         error = -EMFILE;
4307         if (newfd >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur)
4308                 goto out;
4309 +       if (!vx_files_avail(1))
4310 +               goto out;
4311
4312         error = expand_files(files, newfd);
4313         if (error < 0)
4314 @@ -128,6 +131,7 @@ static int dupfd(struct file *file, unsi
4315                 else
4316                         FD_CLR(fd, fdt->close_on_exec);
4317                 spin_unlock(&files->file_lock);
4318 +               vx_openfd_inc(fd);
4319                 fd_install(fd, file);
4320         } else {
4321                 spin_unlock(&files->file_lock);
4322 @@ -180,6 +184,9 @@ asmlinkage long sys_dup2(unsigned int ol
4323
4324         if (tofree)
4325                 filp_close(tofree, files);
4326 +       else
4327 +               vx_openfd_inc(newfd);   /* fd was unused */
4328 +
4329         err = newfd;
4330  out:
4331         return err;
4332 diff -NurpP --minimal linux-2.6.25.11/fs/file_table.c linux-2.6.25.11-vs2.3.0.34.14/fs/file_table.c
4333 --- linux-2.6.25.11/fs/file_table.c     2008-04-17 12:05:40.000000000 -0400
4334 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/file_table.c       2008-04-19 15:14:52.000000000 -0400
4335 @@ -20,6 +20,8 @@
4336  #include <linux/fsnotify.h>
4337  #include <linux/sysctl.h>
4338  #include <linux/percpu_counter.h>
4339 +#include <linux/vs_limit.h>
4340 +#include <linux/vs_context.h>
4341
4342  #include <asm/atomic.h>
4343
4344 @@ -124,6 +126,8 @@ struct file *get_empty_filp(void)
4345         f->f_gid = tsk->fsgid;
4346         eventpoll_init_file(f);
4347         /* f->f_version: 0 */
4348 +       f->f_xid = vx_current_xid();
4349 +       vx_files_inc(f);
4350         return f;
4351
4352  over:
4353 @@ -239,6 +243,8 @@ void __fput(struct file *file)
4354         if (file->f_mode & FMODE_WRITE)
4355                 put_write_access(inode);
4356         put_pid(file->f_owner.pid);
4357 +       vx_files_dec(file);
4358 +       file->f_xid = 0;
4359         file_kill(file);
4360         file->f_path.dentry = NULL;
4361         file->f_path.mnt = NULL;
4362 @@ -304,6 +310,8 @@ void put_filp(struct file *file)
4363  {
4364         if (atomic_dec_and_test(&file->f_count)) {
4365                 security_file_free(file);
4366 +               vx_files_dec(file);
4367 +               file->f_xid = 0;
4368                 file_kill(file);
4369                 file_free(file);
4370         }
4371 diff -NurpP --minimal linux-2.6.25.11/fs/hfsplus/ioctl.c linux-2.6.25.11-vs2.3.0.34.14/fs/hfsplus/ioctl.c
4372 --- linux-2.6.25.11/fs/hfsplus/ioctl.c  2008-04-17 10:37:23.000000000 -0400
4373 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/hfsplus/ioctl.c    2008-04-19 15:14:52.000000000 -0400
4374 @@ -16,6 +16,7 @@
4375  #include <linux/fs.h>
4376  #include <linux/sched.h>
4377  #include <linux/xattr.h>
4378 +#include <linux/mount.h>
4379  #include <asm/uaccess.h>
4380  #include "hfsplus_fs.h"
4381
4382 @@ -35,7 +36,8 @@ int hfsplus_ioctl(struct inode *inode, s
4383                         flags |= FS_NODUMP_FL; /* EXT2_NODUMP_FL */
4384                 return put_user(flags, (int __user *)arg);
4385         case HFSPLUS_IOC_EXT2_SETFLAGS: {
4386 -               if (IS_RDONLY(inode))
4387 +               if (IS_RDONLY(inode) ||
4388 +                       (filp && MNT_IS_RDONLY(filp->f_vfsmnt)))
4389                         return -EROFS;
4390
4391                 if (!is_owner_or_cap(inode))
4392 diff -NurpP --minimal linux-2.6.25.11/fs/inode.c linux-2.6.25.11-vs2.3.0.34.14/fs/inode.c
4393 --- linux-2.6.25.11/fs/inode.c  2008-04-17 12:05:40.000000000 -0400
4394 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/inode.c    2008-04-19 15:14:52.000000000 -0400
4395 @@ -124,6 +124,9 @@ static struct inode *alloc_inode(struct
4396                 struct address_space * const mapping = &inode->i_data;
4397
4398                 inode->i_sb = sb;
4399 +
4400 +               /* essential because of inode slab reuse */
4401 +               inode->i_tag = 0;
4402                 inode->i_blkbits = sb->s_blocksize_bits;
4403                 inode->i_flags = 0;
4404                 atomic_set(&inode->i_count, 1);
4405 @@ -142,6 +145,7 @@ static struct inode *alloc_inode(struct
4406                 inode->i_bdev = NULL;
4407                 inode->i_cdev = NULL;
4408                 inode->i_rdev = 0;
4409 +               inode->i_mdev = 0;
4410                 inode->dirtied_when = 0;
4411                 if (security_inode_alloc(inode)) {
4412                         if (inode->i_sb->s_op->destroy_inode)
4413 @@ -246,6 +250,8 @@ void __iget(struct inode * inode)
4414         inodes_stat.nr_unused--;
4415  }
4416
4417 +EXPORT_SYMBOL_GPL(__iget);
4418 +
4419  /**
4420   * clear_inode - clear an inode
4421   * @inode: inode to clear
4422 @@ -1434,9 +1440,11 @@ void init_special_inode(struct inode *in
4423         if (S_ISCHR(mode)) {
4424                 inode->i_fop = &def_chr_fops;
4425                 inode->i_rdev = rdev;
4426 +               inode->i_mdev = rdev;
4427         } else if (S_ISBLK(mode)) {
4428                 inode->i_fop = &def_blk_fops;
4429                 inode->i_rdev = rdev;
4430 +               inode->i_mdev = rdev;
4431         } else if (S_ISFIFO(mode))
4432                 inode->i_fop = &def_fifo_fops;
4433         else if (S_ISSOCK(mode))
4434 diff -NurpP --minimal linux-2.6.25.11/fs/ioctl.c linux-2.6.25.11-vs2.3.0.34.14/fs/ioctl.c
4435 --- linux-2.6.25.11/fs/ioctl.c  2008-04-17 12:05:40.000000000 -0400
4436 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ioctl.c    2008-04-21 09:25:46.000000000 -0400
4437 @@ -13,6 +13,9 @@
4438  #include <linux/security.h>
4439  #include <linux/module.h>
4440  #include <linux/uaccess.h>
4441 +#include <linux/proc_fs.h>
4442 +#include <linux/vserver/inode.h>
4443 +#include <linux/vs_tag.h>
4444
4445  #include <asm/ioctls.h>
4446
4447 diff -NurpP --minimal linux-2.6.25.11/fs/ioprio.c linux-2.6.25.11-vs2.3.0.34.14/fs/ioprio.c
4448 --- linux-2.6.25.11/fs/ioprio.c 2008-04-17 12:05:40.000000000 -0400
4449 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ioprio.c   2008-04-19 15:14:52.000000000 -0400
4450 @@ -26,6 +26,7 @@
4451  #include <linux/syscalls.h>
4452  #include <linux/security.h>
4453  #include <linux/pid_namespace.h>
4454 +#include <linux/vs_base.h>
4455
4456  static int set_task_ioprio(struct task_struct *task, int ioprio)
4457  {
4458 @@ -116,6 +117,8 @@ asmlinkage long sys_ioprio_set(int which
4459                         else
4460                                 pgrp = find_vpid(who);
4461                         do_each_pid_task(pgrp, PIDTYPE_PGID, p) {
4462 +                               if (!vx_check(p->xid, VS_ADMIN_P | VS_IDENT))
4463 +                                       continue;
4464                                 ret = set_task_ioprio(p, ioprio);
4465                                 if (ret)
4466                                         break;
4467 @@ -205,6 +208,8 @@ asmlinkage long sys_ioprio_get(int which
4468                         else
4469                                 pgrp = find_vpid(who);
4470                         do_each_pid_task(pgrp, PIDTYPE_PGID, p) {
4471 +                               if (!vx_check(p->xid, VS_ADMIN_P | VS_IDENT))
4472 +                                       continue;
4473                                 tmpio = get_task_ioprio(p);
4474                                 if (tmpio < 0)
4475                                         continue;
4476 diff -NurpP --minimal linux-2.6.25.11/fs/jfs/acl.c linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/acl.c
4477 --- linux-2.6.25.11/fs/jfs/acl.c        2007-02-04 13:44:54.000000000 -0500
4478 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/acl.c  2008-04-19 15:14:52.000000000 -0400
4479 @@ -232,7 +232,8 @@ int jfs_setattr(struct dentry *dentry, s
4480                 return rc;
4481
4482         if ((iattr->ia_valid & ATTR_UID && iattr->ia_uid != inode->i_uid) ||
4483 -           (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid)) {
4484 +           (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid) ||
4485 +           (iattr->ia_valid & ATTR_TAG && iattr->ia_tag != inode->i_tag)) {
4486                 if (DQUOT_TRANSFER(inode, iattr))
4487                         return -EDQUOT;
4488         }
4489 diff -NurpP --minimal linux-2.6.25.11/fs/jfs/file.c linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/file.c
4490 --- linux-2.6.25.11/fs/jfs/file.c       2008-04-17 12:05:40.000000000 -0400
4491 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/file.c 2008-04-19 15:14:52.000000000 -0400
4492 @@ -98,6 +98,7 @@ const struct inode_operations jfs_file_i
4493         .setattr        = jfs_setattr,
4494         .permission     = jfs_permission,
4495  #endif
4496 +       .sync_flags     = jfs_sync_flags,
4497  };
4498
4499  const struct file_operations jfs_file_operations = {
4500 diff -NurpP --minimal linux-2.6.25.11/fs/jfs/inode.c linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/inode.c
4501 --- linux-2.6.25.11/fs/jfs/inode.c      2008-04-17 12:05:40.000000000 -0400
4502 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/inode.c        2008-04-19 15:14:52.000000000 -0400
4503 @@ -22,6 +22,7 @@
4504  #include <linux/buffer_head.h>
4505  #include <linux/pagemap.h>
4506  #include <linux/quotaops.h>
4507 +#include <linux/vs_dlimit.h>
4508  #include "jfs_incore.h"
4509  #include "jfs_inode.h"
4510  #include "jfs_filsys.h"
4511 @@ -155,6 +156,7 @@ void jfs_delete_inode(struct inode *inod
4512                 DQUOT_INIT(inode);
4513                 DQUOT_FREE_INODE(inode);
4514                 DQUOT_DROP(inode);
4515 +               DLIMIT_FREE_INODE(inode);
4516         }
4517
4518         clear_inode(inode);
4519 diff -NurpP --minimal linux-2.6.25.11/fs/jfs/ioctl.c linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/ioctl.c
4520 --- linux-2.6.25.11/fs/jfs/ioctl.c      2008-04-17 12:05:40.000000000 -0400
4521 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/ioctl.c        2008-07-11 02:50:45.000000000 -0400
4522 @@ -10,6 +10,7 @@
4523  #include <linux/capability.h>
4524  #include <linux/time.h>
4525  #include <linux/sched.h>
4526 +#include <linux/mount.h>
4527  #include <asm/current.h>
4528  #include <asm/uaccess.h>
4529
4530 @@ -66,7 +67,8 @@ long jfs_ioctl(struct file *filp, unsign
4531         case JFS_IOC_SETFLAGS: {
4532                 unsigned int oldflags;
4533
4534 -               if (IS_RDONLY(inode))
4535 +               if (IS_RDONLY(inode) ||
4536 +                       (filp && MNT_IS_RDONLY(filp->f_vfsmnt)))
4537                         return -EROFS;
4538
4539                 if (!is_owner_or_cap(inode))
4540 @@ -79,6 +81,11 @@ long jfs_ioctl(struct file *filp, unsign
4541                 if (!S_ISDIR(inode->i_mode))
4542                         flags &= ~JFS_DIRSYNC_FL;
4543
4544 +               if (IS_BARRIER(inode)) {
4545 +                       vxwprintk_task(1, "messing with the barrier.");
4546 +                       return -EACCES;
4547 +               }
4548 +
4549                 /* Is it quota file? Do not allow user to mess with it */
4550                 if (IS_NOQUOTA(inode))
4551                         return -EPERM;
4552 @@ -94,8 +101,8 @@ long jfs_ioctl(struct file *filp, unsign
4553                  * the relevant capability.
4554                  */
4555                 if ((oldflags & JFS_IMMUTABLE_FL) ||
4556 -                       ((flags ^ oldflags) &
4557 -                       (JFS_APPEND_FL | JFS_IMMUTABLE_FL))) {
4558 +                       ((flags ^ oldflags) & (JFS_APPEND_FL |
4559 +                       JFS_IMMUTABLE_FL | JFS_IXUNLINK_FL))) {
4560                         if (!capable(CAP_LINUX_IMMUTABLE)) {
4561                                 mutex_unlock(&inode->i_mutex);
4562                                 return -EPERM;
4563 diff -NurpP --minimal linux-2.6.25.11/fs/jfs/jfs_dinode.h linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/jfs_dinode.h
4564 --- linux-2.6.25.11/fs/jfs/jfs_dinode.h 2008-04-17 12:05:40.000000000 -0400
4565 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/jfs_dinode.h   2008-07-08 13:52:44.000000000 -0400
4566 @@ -161,9 +161,13 @@ struct dinode {
4567
4568  #define JFS_APPEND_FL          0x01000000 /* writes to file may only append */
4569  #define JFS_IMMUTABLE_FL       0x02000000 /* Immutable file */
4570 +#define JFS_IXUNLINK_FL                0x04000000 /* Immutable invert on unlink */
4571
4572 -#define JFS_FL_USER_VISIBLE    0x03F80000
4573 -#define JFS_FL_USER_MODIFIABLE 0x03F80000
4574 +#define JFS_BARRIER_FL         0x10000000 /* Barrier for chroot() */
4575 +#define JFS_COW_FL             0x20000000 /* Copy on Write marker */
4576 +
4577 +#define JFS_FL_USER_VISIBLE    0x07F80000
4578 +#define JFS_FL_USER_MODIFIABLE 0x07F80000
4579  #define JFS_FL_INHERIT         0x03C80000
4580
4581  /* These are identical to EXT[23]_IOC_GETFLAGS/SETFLAGS */
4582 diff -NurpP --minimal linux-2.6.25.11/fs/jfs/jfs_dtree.c linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/jfs_dtree.c
4583 --- linux-2.6.25.11/fs/jfs/jfs_dtree.c  2008-04-17 12:05:40.000000000 -0400
4584 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/jfs_dtree.c    2008-04-19 15:14:52.000000000 -0400
4585 @@ -102,6 +102,7 @@
4586
4587  #include <linux/fs.h>
4588  #include <linux/quotaops.h>
4589 +#include <linux/vs_dlimit.h>
4590  #include "jfs_incore.h"
4591  #include "jfs_superblock.h"
4592  #include "jfs_filsys.h"
4593 @@ -383,10 +384,10 @@ static u32 add_index(tid_t tid, struct i
4594                  */
4595                 if (DQUOT_ALLOC_BLOCK(ip, sbi->nbperpage))
4596                         goto clean_up;
4597 -               if (dbAlloc(ip, 0, sbi->nbperpage, &xaddr)) {
4598 -                       DQUOT_FREE_BLOCK(ip, sbi->nbperpage);
4599 -                       goto clean_up;
4600 -               }
4601 +               if (DLIMIT_ALLOC_BLOCK(ip, sbi->nbperpage))
4602 +                       goto clean_up_dquot;
4603 +               if (dbAlloc(ip, 0, sbi->nbperpage, &xaddr))
4604 +                       goto clean_up_dlimit;
4605
4606                 /*
4607                  * Save the table, we're going to overwrite it with the
4608 @@ -480,6 +481,12 @@ static u32 add_index(tid_t tid, struct i
4609
4610         return index;
4611
4612 +      clean_up_dlimit:
4613 +       DLIMIT_FREE_BLOCK(ip, sbi->nbperpage);
4614 +
4615 +      clean_up_dquot:
4616 +       DQUOT_FREE_BLOCK(ip, sbi->nbperpage);
4617 +
4618        clean_up:
4619
4620         jfs_ip->next_index--;
4621 @@ -951,6 +958,7 @@ static int dtSplitUp(tid_t tid,
4622         struct tlock *tlck;
4623         struct lv *lv;
4624         int quota_allocation = 0;
4625 +       int dlimit_allocation = 0;
4626
4627         /* get split page */
4628         smp = split->mp;
4629 @@ -1033,6 +1041,12 @@ static int dtSplitUp(tid_t tid,
4630                 }
4631                 quota_allocation += n;
4632
4633 +               if (DLIMIT_ALLOC_BLOCK(ip, n)) {
4634 +                       rc = -ENOSPC;
4635 +                       goto extendOut;
4636 +               }
4637 +               dlimit_allocation += n;
4638 +
4639                 if ((rc = dbReAlloc(sbi->ipbmap, xaddr, (s64) xlen,
4640                                     (s64) n, &nxaddr)))
4641                         goto extendOut;
4642 @@ -1306,6 +1320,9 @@ static int dtSplitUp(tid_t tid,
4643        freeKeyName:
4644         kfree(key.name);
4645
4646 +       /* Rollback dlimit allocation */
4647 +       if (rc && dlimit_allocation)
4648 +               DLIMIT_FREE_BLOCK(ip, dlimit_allocation);
4649         /* Rollback quota allocation */
4650         if (rc && quota_allocation)
4651                 DQUOT_FREE_BLOCK(ip, quota_allocation);
4652 @@ -1373,6 +1390,12 @@ static int dtSplitPage(tid_t tid, struct
4653                 release_metapage(rmp);
4654                 return -EDQUOT;
4655         }
4656 +       /* Allocate blocks to dlimit. */
4657 +       if (DLIMIT_ALLOC_BLOCK(ip, lengthPXD(pxd))) {
4658 +               DQUOT_FREE_BLOCK(ip, lengthPXD(pxd));
4659 +               release_metapage(rmp);
4660 +               return -ENOSPC;
4661 +       }
4662
4663         jfs_info("dtSplitPage: ip:0x%p smp:0x%p rmp:0x%p", ip, smp, rmp);
4664
4665 @@ -1920,6 +1943,12 @@ static int dtSplitRoot(tid_t tid,
4666                 release_metapage(rmp);
4667                 return -EDQUOT;
4668         }
4669 +       /* Allocate blocks to dlimit. */
4670 +       if (DLIMIT_ALLOC_BLOCK(ip, lengthPXD(pxd))) {
4671 +               DQUOT_FREE_BLOCK(ip, lengthPXD(pxd));
4672 +               release_metapage(rmp);
4673 +               return -ENOSPC;
4674 +       }
4675
4676         BT_MARK_DIRTY(rmp, ip);
4677         /*
4678 @@ -2286,6 +2315,8 @@ static int dtDeleteUp(tid_t tid, struct
4679
4680         xlen = lengthPXD(&fp->header.self);
4681
4682 +       /* Free dlimit allocation. */
4683 +       DLIMIT_FREE_BLOCK(ip, xlen);
4684         /* Free quota allocation. */
4685         DQUOT_FREE_BLOCK(ip, xlen);
4686
4687 @@ -2362,6 +2393,8 @@ static int dtDeleteUp(tid_t tid, struct
4688
4689                                 xlen = lengthPXD(&p->header.self);
4690
4691 +                               /* Free dlimit allocation */
4692 +                               DLIMIT_FREE_BLOCK(ip, xlen);
4693                                 /* Free quota allocation */
4694                                 DQUOT_FREE_BLOCK(ip, xlen);
4695
4696 diff -NurpP --minimal linux-2.6.25.11/fs/jfs/jfs_extent.c linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/jfs_extent.c
4697 --- linux-2.6.25.11/fs/jfs/jfs_extent.c 2008-04-17 10:37:23.000000000 -0400
4698 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/jfs_extent.c   2008-04-19 15:14:52.000000000 -0400
4699 @@ -18,6 +18,7 @@
4700
4701  #include <linux/fs.h>
4702  #include <linux/quotaops.h>
4703 +#include <linux/vs_dlimit.h>
4704  #include "jfs_incore.h"
4705  #include "jfs_inode.h"
4706  #include "jfs_superblock.h"
4707 @@ -147,6 +148,14 @@ extAlloc(struct inode *ip, s64 xlen, s64
4708                 return -EDQUOT;
4709         }
4710
4711 +       /* Allocate blocks to dlimit. */
4712 +       if (DLIMIT_ALLOC_BLOCK(ip, nxlen)) {
4713 +               DQUOT_FREE_BLOCK(ip, nxlen);
4714 +               dbFree(ip, nxaddr, (s64) nxlen);
4715 +               mutex_unlock(&JFS_IP(ip)->commit_mutex);
4716 +               return -ENOSPC;
4717 +       }
4718 +
4719         /* determine the value of the extent flag */
4720         xflag = abnr ? XAD_NOTRECORDED : 0;
4721
4722 @@ -164,6 +173,7 @@ extAlloc(struct inode *ip, s64 xlen, s64
4723          */
4724         if (rc) {
4725                 dbFree(ip, nxaddr, nxlen);
4726 +               DLIMIT_FREE_BLOCK(ip, nxlen);
4727                 DQUOT_FREE_BLOCK(ip, nxlen);
4728                 mutex_unlock(&JFS_IP(ip)->commit_mutex);
4729                 return (rc);
4730 @@ -261,6 +271,13 @@ int extRealloc(struct inode *ip, s64 nxl
4731                 mutex_unlock(&JFS_IP(ip)->commit_mutex);
4732                 return -EDQUOT;
4733         }
4734 +       /* Allocate blocks to dlimit. */
4735 +       if (DLIMIT_ALLOC_BLOCK(ip, nxlen)) {
4736 +               DQUOT_FREE_BLOCK(ip, nxlen);
4737 +               dbFree(ip, nxaddr, (s64) nxlen);
4738 +               up(&JFS_IP(ip)->commit_sem);
4739 +               return -ENOSPC;
4740 +       }
4741
4742         delta = nxlen - xlen;
4743
4744 @@ -297,6 +314,7 @@ int extRealloc(struct inode *ip, s64 nxl
4745                 /* extend the extent */
4746                 if ((rc = xtExtend(0, ip, xoff + xlen, (int) nextend, 0))) {
4747                         dbFree(ip, xaddr + xlen, delta);
4748 +                       DLIMIT_FREE_BLOCK(ip, nxlen);
4749                         DQUOT_FREE_BLOCK(ip, nxlen);
4750                         goto exit;
4751                 }
4752 @@ -308,6 +326,7 @@ int extRealloc(struct inode *ip, s64 nxl
4753                  */
4754                 if ((rc = xtTailgate(0, ip, xoff, (int) ntail, nxaddr, 0))) {
4755                         dbFree(ip, nxaddr, nxlen);
4756 +                       DLIMIT_FREE_BLOCK(ip, nxlen);
4757                         DQUOT_FREE_BLOCK(ip, nxlen);
4758                         goto exit;
4759                 }
4760 diff -NurpP --minimal linux-2.6.25.11/fs/jfs/jfs_filsys.h linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/jfs_filsys.h
4761 --- linux-2.6.25.11/fs/jfs/jfs_filsys.h 2008-04-17 10:37:23.000000000 -0400
4762 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/jfs_filsys.h   2008-04-19 15:14:52.000000000 -0400
4763 @@ -263,6 +263,7 @@
4764  #define JFS_NAME_MAX   255
4765  #define JFS_PATH_MAX   BPSIZE
4766
4767 +#define JFS_TAGGED             0x00800000      /* Context Tagging */
4768
4769  /*
4770   *     file system state (superblock state)
4771 diff -NurpP --minimal linux-2.6.25.11/fs/jfs/jfs_imap.c linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/jfs_imap.c
4772 --- linux-2.6.25.11/fs/jfs/jfs_imap.c   2008-04-17 12:05:40.000000000 -0400
4773 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/jfs_imap.c     2008-04-19 15:14:52.000000000 -0400
4774 @@ -45,6 +45,7 @@
4775  #include <linux/buffer_head.h>
4776  #include <linux/pagemap.h>
4777  #include <linux/quotaops.h>
4778 +#include <linux/vs_tag.h>
4779
4780  #include "jfs_incore.h"
4781  #include "jfs_inode.h"
4782 @@ -3061,6 +3062,8 @@ static int copy_from_dinode(struct dinod
4783  {
4784         struct jfs_inode_info *jfs_ip = JFS_IP(ip);
4785         struct jfs_sb_info *sbi = JFS_SBI(ip->i_sb);
4786 +       uid_t uid;
4787 +       gid_t gid;
4788
4789         jfs_ip->fileset = le32_to_cpu(dip->di_fileset);
4790         jfs_ip->mode2 = le32_to_cpu(dip->di_mode);
4791 @@ -3081,14 +3084,18 @@ static int copy_from_dinode(struct dinod
4792         }
4793         ip->i_nlink = le32_to_cpu(dip->di_nlink);
4794
4795 -       jfs_ip->saved_uid = le32_to_cpu(dip->di_uid);
4796 +       uid = le32_to_cpu(dip->di_uid);
4797 +       gid = le32_to_cpu(dip->di_gid);
4798 +       ip->i_tag = INOTAG_TAG(DX_TAG(ip), uid, gid, 0);
4799 +
4800 +       jfs_ip->saved_uid = INOTAG_UID(DX_TAG(ip), uid, gid);
4801         if (sbi->uid == -1)
4802                 ip->i_uid = jfs_ip->saved_uid;
4803         else {
4804                 ip->i_uid = sbi->uid;
4805         }
4806
4807 -       jfs_ip->saved_gid = le32_to_cpu(dip->di_gid);
4808 +       jfs_ip->saved_gid = INOTAG_GID(DX_TAG(ip), uid, gid);
4809         if (sbi->gid == -1)
4810                 ip->i_gid = jfs_ip->saved_gid;
4811         else {
4812 @@ -3153,14 +3160,12 @@ static void copy_to_dinode(struct dinode
4813         dip->di_size = cpu_to_le64(ip->i_size);
4814         dip->di_nblocks = cpu_to_le64(PBLK2LBLK(ip->i_sb, ip->i_blocks));
4815         dip->di_nlink = cpu_to_le32(ip->i_nlink);
4816 -       if (sbi->uid == -1)
4817 -               dip->di_uid = cpu_to_le32(ip->i_uid);
4818 -       else
4819 -               dip->di_uid = cpu_to_le32(jfs_ip->saved_uid);
4820 -       if (sbi->gid == -1)
4821 -               dip->di_gid = cpu_to_le32(ip->i_gid);
4822 -       else
4823 -               dip->di_gid = cpu_to_le32(jfs_ip->saved_gid);
4824 +
4825 +       dip->di_uid = cpu_to_le32(TAGINO_UID(DX_TAG(ip),
4826 +               (sbi->uid == -1) ? ip->i_uid : jfs_ip->saved_uid, ip->i_tag));
4827 +       dip->di_gid = cpu_to_le32(TAGINO_GID(DX_TAG(ip),
4828 +               (sbi->gid == -1) ? ip->i_gid : jfs_ip->saved_gid, ip->i_tag));
4829 +
4830         jfs_get_inode_flags(jfs_ip);
4831         /*
4832          * mode2 is only needed for storing the higher order bits.
4833 diff -NurpP --minimal linux-2.6.25.11/fs/jfs/jfs_inode.c linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/jfs_inode.c
4834 --- linux-2.6.25.11/fs/jfs/jfs_inode.c  2008-04-17 10:33:02.000000000 -0400
4835 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/jfs_inode.c    2008-07-08 05:01:03.000000000 -0400
4836 @@ -18,6 +18,8 @@
4837
4838  #include <linux/fs.h>
4839  #include <linux/quotaops.h>
4840 +#include <linux/vs_dlimit.h>
4841 +#include <linux/vs_tag.h>
4842  #include "jfs_incore.h"
4843  #include "jfs_inode.h"
4844  #include "jfs_filsys.h"
4845 @@ -30,29 +32,46 @@ void jfs_set_inode_flags(struct inode *i
4846  {
4847         unsigned int flags = JFS_IP(inode)->mode2;
4848
4849 -       inode->i_flags &= ~(S_IMMUTABLE | S_APPEND |
4850 -               S_NOATIME | S_DIRSYNC | S_SYNC);
4851 +       inode->i_flags &= ~(S_IMMUTABLE | S_IXUNLINK |
4852 +               S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
4853
4854         if (flags & JFS_IMMUTABLE_FL)
4855                 inode->i_flags |= S_IMMUTABLE;
4856 +       if (flags & JFS_IXUNLINK_FL)
4857 +               inode->i_flags |= S_IXUNLINK;
4858 +
4859 +       if (flags & JFS_SYNC_FL)
4860 +               inode->i_flags |= S_SYNC;
4861         if (flags & JFS_APPEND_FL)
4862                 inode->i_flags |= S_APPEND;
4863         if (flags & JFS_NOATIME_FL)
4864                 inode->i_flags |= S_NOATIME;
4865         if (flags & JFS_DIRSYNC_FL)
4866                 inode->i_flags |= S_DIRSYNC;
4867 -       if (flags & JFS_SYNC_FL)
4868 -               inode->i_flags |= S_SYNC;
4869 +
4870 +       inode->i_vflags &= ~(V_BARRIER | V_COW);
4871 +
4872 +       if (flags & JFS_BARRIER_FL)
4873 +               inode->i_vflags |= V_BARRIER;
4874 +       if (flags & JFS_COW_FL)
4875 +               inode->i_vflags |= V_COW;
4876  }
4877
4878  void jfs_get_inode_flags(struct jfs_inode_info *jfs_ip)
4879  {
4880         unsigned int flags = jfs_ip->vfs_inode.i_flags;
4881 +       unsigned int vflags = jfs_ip->vfs_inode.i_vflags;
4882 +
4883 +       jfs_ip->mode2 &= ~(JFS_IMMUTABLE_FL | JFS_IXUNLINK_FL |
4884 +                          JFS_APPEND_FL | JFS_NOATIME_FL |
4885 +                          JFS_DIRSYNC_FL | JFS_SYNC_FL |
4886 +                          JFS_BARRIER_FL | JFS_COW_FL);
4887
4888 -       jfs_ip->mode2 &= ~(JFS_IMMUTABLE_FL | JFS_APPEND_FL | JFS_NOATIME_FL |
4889 -                          JFS_DIRSYNC_FL | JFS_SYNC_FL);
4890         if (flags & S_IMMUTABLE)
4891                 jfs_ip->mode2 |= JFS_IMMUTABLE_FL;
4892 +       if (flags & S_IXUNLINK)
4893 +               jfs_ip->mode2 |= JFS_IXUNLINK_FL;
4894 +
4895         if (flags & S_APPEND)
4896                 jfs_ip->mode2 |= JFS_APPEND_FL;
4897         if (flags & S_NOATIME)
4898 @@ -61,6 +80,19 @@ void jfs_get_inode_flags(struct jfs_inod
4899                 jfs_ip->mode2 |= JFS_DIRSYNC_FL;
4900         if (flags & S_SYNC)
4901                 jfs_ip->mode2 |= JFS_SYNC_FL;
4902 +
4903 +       if (vflags & V_BARRIER)
4904 +               jfs_ip->mode2 |= JFS_BARRIER_FL;
4905 +       if (vflags & V_COW)
4906 +               jfs_ip->mode2 |= JFS_COW_FL;
4907 +}
4908 +
4909 +int jfs_sync_flags(struct inode *inode)
4910 +{
4911 +       jfs_get_inode_flags(JFS_IP(inode));
4912 +       inode->i_ctime = CURRENT_TIME;
4913 +       mark_inode_dirty(inode);
4914 +       return 0;
4915  }
4916
4917  /*
4918 @@ -108,10 +140,17 @@ struct inode *ialloc(struct inode *paren
4919         jfs_inode->saved_uid = inode->i_uid;
4920         jfs_inode->saved_gid = inode->i_gid;
4921
4922 +       inode->i_tag = dx_current_fstag(sb);
4923 +       if (DLIMIT_ALLOC_INODE(inode)) {
4924 +               iput(inode);
4925 +               return ERR_PTR(-ENOSPC);
4926 +       }
4927 +
4928         /*
4929          * Allocate inode to quota.
4930          */
4931         if (DQUOT_ALLOC_INODE(inode)) {
4932 +               DLIMIT_FREE_INODE(inode);
4933                 DQUOT_DROP(inode);
4934                 inode->i_flags |= S_NOQUOTA;
4935                 inode->i_nlink = 0;
4936 diff -NurpP --minimal linux-2.6.25.11/fs/jfs/jfs_inode.h linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/jfs_inode.h
4937 --- linux-2.6.25.11/fs/jfs/jfs_inode.h  2008-04-17 12:05:40.000000000 -0400
4938 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/jfs_inode.h    2008-04-19 15:14:52.000000000 -0400
4939 @@ -39,6 +39,7 @@ extern struct dentry *jfs_fh_to_dentry(s
4940  extern struct dentry *jfs_fh_to_parent(struct super_block *sb, struct fid *fid,
4941         int fh_len, int fh_type);
4942  extern void jfs_set_inode_flags(struct inode *);
4943 +extern int jfs_sync_flags(struct inode *);
4944  extern int jfs_get_block(struct inode *, sector_t, struct buffer_head *, int);
4945
4946  extern const struct address_space_operations jfs_aops;
4947 diff -NurpP --minimal linux-2.6.25.11/fs/jfs/jfs_xtree.c linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/jfs_xtree.c
4948 --- linux-2.6.25.11/fs/jfs/jfs_xtree.c  2008-04-17 12:05:40.000000000 -0400
4949 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/jfs_xtree.c    2008-04-19 15:14:52.000000000 -0400
4950 @@ -21,6 +21,7 @@
4951
4952  #include <linux/fs.h>
4953  #include <linux/quotaops.h>
4954 +#include <linux/vs_dlimit.h>
4955  #include "jfs_incore.h"
4956  #include "jfs_filsys.h"
4957  #include "jfs_metapage.h"
4958 @@ -846,7 +847,12 @@ int xtInsert(tid_t tid,            /* transaction
4959                         hint = 0;
4960                 if ((rc = DQUOT_ALLOC_BLOCK(ip, xlen)))
4961                         goto out;
4962 +               if ((rc = DLIMIT_ALLOC_BLOCK(ip, xlen))) {
4963 +                       DQUOT_FREE_BLOCK(ip, xlen);
4964 +                       goto out;
4965 +               }
4966                 if ((rc = dbAlloc(ip, hint, (s64) xlen, &xaddr))) {
4967 +                       DLIMIT_FREE_BLOCK(ip, xlen);
4968                         DQUOT_FREE_BLOCK(ip, xlen);
4969                         goto out;
4970                 }
4971 @@ -876,6 +882,7 @@ int xtInsert(tid_t tid,             /* transaction
4972                         /* undo data extent allocation */
4973                         if (*xaddrp == 0) {
4974                                 dbFree(ip, xaddr, (s64) xlen);
4975 +                               DLIMIT_FREE_BLOCK(ip, xlen);
4976                                 DQUOT_FREE_BLOCK(ip, xlen);
4977                         }
4978                         return rc;
4979 @@ -1236,6 +1243,7 @@ xtSplitPage(tid_t tid, struct inode *ip,
4980         struct tlock *tlck;
4981         struct xtlock *sxtlck = NULL, *rxtlck = NULL;
4982         int quota_allocation = 0;
4983 +       int dlimit_allocation = 0;
4984
4985         smp = split->mp;
4986         sp = XT_PAGE(ip, smp);
4987 @@ -1255,6 +1263,13 @@ xtSplitPage(tid_t tid, struct inode *ip,
4988
4989         quota_allocation += lengthPXD(pxd);
4990
4991 +       /* Allocate blocks to dlimit. */
4992 +       if (DLIMIT_ALLOC_BLOCK(ip, lengthPXD(pxd))) {
4993 +              rc = -ENOSPC;
4994 +              goto clean_up;
4995 +       }
4996 +       dlimit_allocation += lengthPXD(pxd);
4997 +
4998         /*
4999          * allocate the new right page for the split
5000          */
5001 @@ -1456,6 +1471,9 @@ xtSplitPage(tid_t tid, struct inode *ip,
5002
5003        clean_up:
5004
5005 +       /* Rollback dlimit allocation. */
5006 +       if (dlimit_allocation)
5007 +               DLIMIT_FREE_BLOCK(ip, dlimit_allocation);
5008         /* Rollback quota allocation. */
5009         if (quota_allocation)
5010                 DQUOT_FREE_BLOCK(ip, quota_allocation);
5011 @@ -1519,6 +1537,12 @@ xtSplitRoot(tid_t tid,
5012                 release_metapage(rmp);
5013                 return -EDQUOT;
5014         }
5015 +       /* Allocate blocks to dlimit. */
5016 +       if (DLIMIT_ALLOC_BLOCK(ip, lengthPXD(pxd))) {
5017 +               DQUOT_FREE_BLOCK(ip, lengthPXD(pxd));
5018 +               release_metapage(rmp);
5019 +               return -ENOSPC;
5020 +       }
5021
5022         jfs_info("xtSplitRoot: ip:0x%p rmp:0x%p", ip, rmp);
5023
5024 @@ -3948,6 +3972,8 @@ s64 xtTruncate(tid_t tid, struct inode *
5025         else
5026                 ip->i_size = newsize;
5027
5028 +       /* update dlimit allocation to reflect freed blocks */
5029 +       DLIMIT_FREE_BLOCK(ip, nfreed);
5030         /* update quota allocation to reflect freed blocks */
5031         DQUOT_FREE_BLOCK(ip, nfreed);
5032
5033 diff -NurpP --minimal linux-2.6.25.11/fs/jfs/namei.c linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/namei.c
5034 --- linux-2.6.25.11/fs/jfs/namei.c      2008-04-17 12:05:40.000000000 -0400
5035 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/namei.c        2008-04-19 15:14:52.000000000 -0400
5036 @@ -21,6 +21,7 @@
5037  #include <linux/ctype.h>
5038  #include <linux/quotaops.h>
5039  #include <linux/exportfs.h>
5040 +#include <linux/vs_tag.h>
5041  #include "jfs_incore.h"
5042  #include "jfs_superblock.h"
5043  #include "jfs_inode.h"
5044 @@ -1468,6 +1469,7 @@ static struct dentry *jfs_lookup(struct
5045                 return ERR_CAST(ip);
5046         }
5047
5048 +       dx_propagate_tag(nd, ip);
5049         dentry = d_splice_alias(ip, dentry);
5050
5051         if (dentry && (JFS_SBI(dip->i_sb)->mntflag & JFS_OS2))
5052 @@ -1550,6 +1552,7 @@ const struct inode_operations jfs_dir_in
5053         .setattr        = jfs_setattr,
5054         .permission     = jfs_permission,
5055  #endif
5056 +       .sync_flags     = jfs_sync_flags,
5057  };
5058
5059  const struct file_operations jfs_dir_operations = {
5060 diff -NurpP --minimal linux-2.6.25.11/fs/jfs/super.c linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/super.c
5061 --- linux-2.6.25.11/fs/jfs/super.c      2008-04-17 12:05:40.000000000 -0400
5062 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/super.c        2008-04-19 15:14:52.000000000 -0400
5063 @@ -195,7 +195,8 @@ static void jfs_put_super(struct super_b
5064  enum {
5065         Opt_integrity, Opt_nointegrity, Opt_iocharset, Opt_resize,
5066         Opt_resize_nosize, Opt_errors, Opt_ignore, Opt_err, Opt_quota,
5067 -       Opt_usrquota, Opt_grpquota, Opt_uid, Opt_gid, Opt_umask
5068 +       Opt_usrquota, Opt_grpquota, Opt_uid, Opt_gid, Opt_umask,
5069 +       Opt_tag, Opt_notag, Opt_tagid
5070  };
5071
5072  static match_table_t tokens = {
5073 @@ -205,6 +206,10 @@ static match_table_t tokens = {
5074         {Opt_resize, "resize=%u"},
5075         {Opt_resize_nosize, "resize"},
5076         {Opt_errors, "errors=%s"},
5077 +       {Opt_tag, "tag"},
5078 +       {Opt_notag, "notag"},
5079 +       {Opt_tagid, "tagid=%u"},
5080 +       {Opt_tag, "tagxid"},
5081         {Opt_ignore, "noquota"},
5082         {Opt_ignore, "quota"},
5083         {Opt_usrquota, "usrquota"},
5084 @@ -339,6 +344,20 @@ static int parse_options(char *options,
5085                         }
5086                         break;
5087                 }
5088 +#ifndef CONFIG_TAGGING_NONE
5089 +               case Opt_tag:
5090 +                       *flag |= JFS_TAGGED;
5091 +                       break;
5092 +               case Opt_notag:
5093 +                       *flag &= JFS_TAGGED;
5094 +                       break;
5095 +#endif
5096 +#ifdef CONFIG_PROPAGATE
5097 +               case Opt_tagid:
5098 +                       /* use args[0] */
5099 +                       *flag |= JFS_TAGGED;
5100 +                       break;
5101 +#endif
5102                 default:
5103                         printk("jfs: Unrecognized mount option \"%s\" "
5104                                         " or missing value\n", p);
5105 @@ -369,6 +388,13 @@ static int jfs_remount(struct super_bloc
5106         if (!parse_options(data, sb, &newLVSize, &flag)) {
5107                 return -EINVAL;
5108         }
5109 +
5110 +       if ((flag & JFS_TAGGED) && !(sb->s_flags & MS_TAGGED)) {
5111 +               printk(KERN_ERR "JFS: %s: tagging not permitted on remount.\n",
5112 +                       sb->s_id);
5113 +               return -EINVAL;
5114 +       }
5115 +
5116         if (newLVSize) {
5117                 if (sb->s_flags & MS_RDONLY) {
5118                         printk(KERN_ERR
5119 @@ -440,6 +466,9 @@ static int jfs_fill_super(struct super_b
5120  #ifdef CONFIG_JFS_POSIX_ACL
5121         sb->s_flags |= MS_POSIXACL;
5122  #endif
5123 +       /* map mount option tagxid */
5124 +       if (sbi->flag & JFS_TAGGED)
5125 +               sb->s_flags |= MS_TAGGED;
5126
5127         if (newLVSize) {
5128                 printk(KERN_ERR "resize option for remount only\n");
5129 diff -NurpP --minimal linux-2.6.25.11/fs/jfs/xattr.c linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/xattr.c
5130 --- linux-2.6.25.11/fs/jfs/xattr.c      2008-04-17 10:37:23.000000000 -0400
5131 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/jfs/xattr.c        2008-04-19 15:14:52.000000000 -0400
5132 @@ -23,6 +23,7 @@
5133  #include <linux/posix_acl_xattr.h>
5134  #include <linux/quotaops.h>
5135  #include <linux/security.h>
5136 +#include <linux/vs_dlimit.h>
5137  #include "jfs_incore.h"
5138  #include "jfs_superblock.h"
5139  #include "jfs_dmap.h"
5140 @@ -263,9 +264,16 @@ static int ea_write(struct inode *ip, st
5141         if (DQUOT_ALLOC_BLOCK(ip, nblocks)) {
5142                 return -EDQUOT;
5143         }
5144 +       /* Allocate new blocks to dlimit. */
5145 +       if (DLIMIT_ALLOC_BLOCK(ip, nblocks)) {
5146 +               DQUOT_FREE_BLOCK(ip, nblocks);
5147 +               return -ENOSPC;
5148 +       }
5149
5150         rc = dbAlloc(ip, INOHINT(ip), nblocks, &blkno);
5151         if (rc) {
5152 +               /*Rollback dlimit allocation. */
5153 +               DLIMIT_FREE_BLOCK(ip, nblocks);
5154                 /*Rollback quota allocation. */
5155                 DQUOT_FREE_BLOCK(ip, nblocks);
5156                 return rc;
5157 @@ -332,6 +340,8 @@ static int ea_write(struct inode *ip, st
5158
5159        failed:
5160         /* Rollback quota allocation. */
5161 +       DLIMIT_FREE_BLOCK(ip, nblocks);
5162 +       /* Rollback quota allocation. */
5163         DQUOT_FREE_BLOCK(ip, nblocks);
5164
5165         dbFree(ip, blkno, nblocks);
5166 @@ -468,6 +478,7 @@ static int ea_get(struct inode *inode, s
5167         s64 blkno;
5168         int rc;
5169         int quota_allocation = 0;
5170 +       int dlimit_allocation = 0;
5171
5172         /* When fsck.jfs clears a bad ea, it doesn't clear the size */
5173         if (ji->ea.flag == 0)
5174 @@ -543,6 +554,12 @@ static int ea_get(struct inode *inode, s
5175
5176                 quota_allocation = blocks_needed;
5177
5178 +               /* Allocate new blocks to dlimit. */
5179 +               rc = -ENOSPC;
5180 +               if (DLIMIT_ALLOC_BLOCK(inode, blocks_needed))
5181 +                       goto clean_up;
5182 +               dlimit_allocation = blocks_needed;
5183 +
5184                 rc = dbAlloc(inode, INOHINT(inode), (s64) blocks_needed,
5185                              &blkno);
5186                 if (rc)
5187 @@ -600,6 +617,9 @@ static int ea_get(struct inode *inode, s
5188         return ea_size;
5189
5190        clean_up:
5191 +       /* Rollback dlimit allocation */
5192 +       if (dlimit_allocation)
5193 +               DLIMIT_FREE_BLOCK(inode, dlimit_allocation);
5194         /* Rollback quota allocation */
5195         if (quota_allocation)
5196                 DQUOT_FREE_BLOCK(inode, quota_allocation);
5197 @@ -676,8 +696,10 @@ static int ea_put(tid_t tid, struct inod
5198         }
5199
5200         /* If old blocks exist, they must be removed from quota allocation. */
5201 -       if (old_blocks)
5202 +       if (old_blocks) {
5203 +               DLIMIT_FREE_BLOCK(inode, old_blocks);
5204                 DQUOT_FREE_BLOCK(inode, old_blocks);
5205 +       }
5206
5207         inode->i_ctime = CURRENT_TIME;
5208
5209 diff -NurpP --minimal linux-2.6.25.11/fs/libfs.c linux-2.6.25.11-vs2.3.0.34.14/fs/libfs.c
5210 --- linux-2.6.25.11/fs/libfs.c  2008-04-17 12:05:40.000000000 -0400
5211 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/libfs.c    2008-04-19 15:14:52.000000000 -0400
5212 @@ -125,7 +125,8 @@ static inline unsigned char dt_type(stru
5213   * both impossible due to the lock on directory.
5214   */
5215
5216 -int dcache_readdir(struct file * filp, void * dirent, filldir_t filldir)
5217 +static inline int do_dcache_readdir_filter(struct file *filp,
5218 +       void *dirent, filldir_t filldir, int (*filter)(struct dentry *dentry))
5219  {
5220         struct dentry *dentry = filp->f_path.dentry;
5221         struct dentry *cursor = filp->private_data;
5222 @@ -158,6 +159,8 @@ int dcache_readdir(struct file * filp, v
5223                                 next = list_entry(p, struct dentry, d_u.d_child);
5224                                 if (d_unhashed(next) || !next->d_inode)
5225                                         continue;
5226 +                               if (filter && !filter(next))
5227 +                                       continue;
5228
5229                                 spin_unlock(&dcache_lock);
5230                                 if (filldir(dirent, next->d_name.name,
5231 @@ -176,6 +179,18 @@ int dcache_readdir(struct file * filp, v
5232         return 0;
5233  }
5234
5235 +int dcache_readdir(struct file *filp, void *dirent, filldir_t filldir)
5236 +{
5237 +       return do_dcache_readdir_filter(filp, dirent, filldir, NULL);
5238 +}
5239 +
5240 +int dcache_readdir_filter(struct file *filp, void *dirent, filldir_t filldir,
5241 +       int (*filter)(struct dentry *))
5242 +{
5243 +       return do_dcache_readdir_filter(filp, dirent, filldir, filter);
5244 +}
5245 +
5246 +
5247  ssize_t generic_read_dir(struct file *filp, char __user *buf, size_t siz, loff_t *ppos)
5248  {
5249         return -EISDIR;
5250 @@ -778,6 +793,7 @@ EXPORT_SYMBOL(dcache_dir_close);
5251  EXPORT_SYMBOL(dcache_dir_lseek);
5252  EXPORT_SYMBOL(dcache_dir_open);
5253  EXPORT_SYMBOL(dcache_readdir);
5254 +EXPORT_SYMBOL(dcache_readdir_filter);
5255  EXPORT_SYMBOL(generic_read_dir);
5256  EXPORT_SYMBOL(get_sb_pseudo);
5257  EXPORT_SYMBOL(simple_write_begin);
5258 diff -NurpP --minimal linux-2.6.25.11/fs/locks.c linux-2.6.25.11-vs2.3.0.34.14/fs/locks.c
5259 --- linux-2.6.25.11/fs/locks.c  2008-07-15 22:05:17.000000000 -0400
5260 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/locks.c    2008-05-21 14:30:41.000000000 -0400
5261 @@ -126,6 +126,8 @@
5262  #include <linux/time.h>
5263  #include <linux/rcupdate.h>
5264  #include <linux/pid_namespace.h>
5265 +#include <linux/vs_base.h>
5266 +#include <linux/vs_limit.h>
5267
5268  #include <asm/semaphore.h>
5269  #include <asm/uaccess.h>
5270 @@ -148,6 +150,8 @@ static struct kmem_cache *filelock_cache
5271  /* Allocate an empty lock structure. */
5272  static struct file_lock *locks_alloc_lock(void)
5273  {
5274 +       if (!vx_locks_avail(1))
5275 +               return NULL;
5276         return kmem_cache_alloc(filelock_cache, GFP_KERNEL);
5277  }
5278
5279 @@ -173,6 +177,7 @@ static void locks_free_lock(struct file_
5280         BUG_ON(!list_empty(&fl->fl_block));
5281         BUG_ON(!list_empty(&fl->fl_link));
5282
5283 +       vx_locks_dec(fl);
5284         locks_release_private(fl);
5285         kmem_cache_free(filelock_cache, fl);
5286  }
5287 @@ -193,6 +198,7 @@ void locks_init_lock(struct file_lock *f
5288         fl->fl_start = fl->fl_end = 0;
5289         fl->fl_ops = NULL;
5290         fl->fl_lmops = NULL;
5291 +       fl->fl_xid = -1;
5292  }
5293
5294  EXPORT_SYMBOL(locks_init_lock);
5295 @@ -246,6 +252,7 @@ void locks_copy_lock(struct file_lock *n
5296         new->fl_file = fl->fl_file;
5297         new->fl_ops = fl->fl_ops;
5298         new->fl_lmops = fl->fl_lmops;
5299 +       new->fl_xid = fl->fl_xid;
5300
5301         locks_copy_private(new, fl);
5302  }
5303 @@ -284,6 +291,11 @@ static int flock_make_lock(struct file *
5304         fl->fl_flags = FL_FLOCK;
5305         fl->fl_type = type;
5306         fl->fl_end = OFFSET_MAX;
5307 +
5308 +       vxd_assert(filp->f_xid == vx_current_xid(),
5309 +               "f_xid(%d) == current(%d)", filp->f_xid, vx_current_xid());
5310 +       fl->fl_xid = filp->f_xid;
5311 +       vx_locks_inc(fl);
5312
5313         *lock = fl;
5314         return 0;
5315 @@ -449,6 +461,7 @@ static int lease_init(struct file *filp,
5316
5317         fl->fl_owner = current->files;
5318         fl->fl_pid = current->tgid;
5319 +       fl->fl_xid = vx_current_xid();
5320
5321         fl->fl_file = filp;
5322         fl->fl_flags = FL_LEASE;
5323 @@ -468,6 +481,11 @@ static struct file_lock *lease_alloc(str
5324         if (fl == NULL)
5325                 return ERR_PTR(error);
5326
5327 +       fl->fl_xid = vx_current_xid();
5328 +       if (filp)
5329 +               vxd_assert(filp->f_xid == fl->fl_xid,
5330 +                       "f_xid(%d) == fl_xid(%d)", filp->f_xid, fl->fl_xid);
5331 +       vx_locks_inc(fl);
5332         error = lease_init(filp, type, fl);
5333         if (error) {
5334                 locks_free_lock(fl);
5335 @@ -774,6 +792,7 @@ static int flock_lock_file(struct file *
5336         if (found)
5337                 cond_resched();
5338
5339 +       new_fl->fl_xid = -1;
5340  find_conflict:
5341         for_each_lock(inode, before) {
5342                 struct file_lock *fl = *before;
5343 @@ -792,6 +811,7 @@ find_conflict:
5344                 goto out;
5345         locks_copy_lock(new_fl, request);
5346         locks_insert_lock(before, new_fl);
5347 +       vx_locks_inc(new_fl);
5348         new_fl = NULL;
5349         error = 0;
5350
5351 @@ -802,7 +822,8 @@ out:
5352         return error;
5353  }
5354
5355 -static int __posix_lock_file(struct inode *inode, struct file_lock *request, struct file_lock *conflock)
5356 +static int __posix_lock_file(struct inode *inode, struct file_lock *request,
5357 +       struct file_lock *conflock, xid_t xid)
5358  {
5359         struct file_lock *fl;
5360         struct file_lock *new_fl = NULL;
5361 @@ -812,6 +833,8 @@ static int __posix_lock_file(struct inod
5362         struct file_lock **before;
5363         int error, added = 0;
5364
5365 +       vxd_assert(xid == vx_current_xid(),
5366 +               "xid(%d) == current(%d)", xid, vx_current_xid());
5367         /*
5368          * We may need two file_lock structures for this operation,
5369          * so we get them in advance to avoid races.
5370 @@ -822,7 +845,11 @@ static int __posix_lock_file(struct inod
5371             (request->fl_type != F_UNLCK ||
5372              request->fl_start != 0 || request->fl_end != OFFSET_MAX)) {
5373                 new_fl = locks_alloc_lock();
5374 +               new_fl->fl_xid = xid;
5375 +               vx_locks_inc(new_fl);
5376                 new_fl2 = locks_alloc_lock();
5377 +               new_fl2->fl_xid = xid;
5378 +               vx_locks_inc(new_fl2);
5379         }
5380
5381         lock_kernel();
5382 @@ -1021,7 +1048,8 @@ static int __posix_lock_file(struct inod
5383  int posix_lock_file(struct file *filp, struct file_lock *fl,
5384                         struct file_lock *conflock)
5385  {
5386 -       return __posix_lock_file(filp->f_path.dentry->d_inode, fl, conflock);
5387 +       return __posix_lock_file(filp->f_path.dentry->d_inode,
5388 +               fl, conflock, filp->f_xid);
5389  }
5390  EXPORT_SYMBOL(posix_lock_file);
5391
5392 @@ -1111,7 +1139,7 @@ int locks_mandatory_area(int read_write,
5393         fl.fl_end = offset + count - 1;
5394
5395         for (;;) {
5396 -               error = __posix_lock_file(inode, &fl, NULL);
5397 +               error = __posix_lock_file(inode, &fl, NULL, filp->f_xid);
5398                 if (error != -EAGAIN)
5399                         break;
5400                 if (!(fl.fl_flags & FL_SLEEP))
5401 @@ -1425,6 +1453,7 @@ int generic_setlease(struct file *filp,
5402
5403         locks_copy_lock(new_fl, lease);
5404         locks_insert_lock(before, new_fl);
5405 +       vx_locks_inc(new_fl);
5406
5407         *flp = new_fl;
5408         return 0;
5409 @@ -1756,6 +1785,11 @@ int fcntl_setlk(unsigned int fd, struct
5410         if (file_lock == NULL)
5411                 return -ENOLCK;
5412
5413 +       vxd_assert(filp->f_xid == vx_current_xid(),
5414 +               "f_xid(%d) == current(%d)", filp->f_xid, vx_current_xid());
5415 +       file_lock->fl_xid = filp->f_xid;
5416 +       vx_locks_inc(file_lock);
5417 +
5418         /*
5419          * This might block, so we do it before checking the inode.
5420          */
5421 @@ -1893,6 +1927,11 @@ int fcntl_setlk64(unsigned int fd, struc
5422         if (file_lock == NULL)
5423                 return -ENOLCK;
5424
5425 +       vxd_assert(filp->f_xid == vx_current_xid(),
5426 +               "f_xid(%d) == current(%d)", filp->f_xid, vx_current_xid());
5427 +       file_lock->fl_xid = filp->f_xid;
5428 +       vx_locks_inc(file_lock);
5429 +
5430         /*
5431          * This might block, so we do it before checking the inode.
5432          */
5433 @@ -2176,8 +2215,11 @@ static int locks_show(struct seq_file *f
5434
5435         lock_get_status(f, fl, (long)f->private, "");
5436
5437 -       list_for_each_entry(bfl, &fl->fl_block, fl_block)
5438 +       list_for_each_entry(bfl, &fl->fl_block, fl_block) {
5439 +               if (!vx_check(fl->fl_xid, VS_WATCH_P | VS_IDENT))
5440 +                       continue;
5441                 lock_get_status(f, bfl, (long)f->private, " ->");
5442 +       }
5443
5444         f->private++;
5445         return 0;
5446 diff -NurpP --minimal linux-2.6.25.11/fs/namei.c linux-2.6.25.11-vs2.3.0.34.14/fs/namei.c
5447 --- linux-2.6.25.11/fs/namei.c  2008-04-17 12:05:40.000000000 -0400
5448 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/namei.c    2008-07-11 03:07:10.000000000 -0400
5449 @@ -30,6 +30,13 @@
5450  #include <linux/capability.h>
5451  #include <linux/file.h>
5452  #include <linux/fcntl.h>
5453 +#include <linux/proc_fs.h>
5454 +#include <linux/vserver/inode.h>
5455 +#include <linux/vs_base.h>
5456 +#include <linux/vs_tag.h>
5457 +#include <linux/vs_cowbl.h>
5458 +#include <linux/vs_device.h>
5459 +#include <linux/vs_context.h>
5460  #include <asm/namei.h>
5461  #include <asm/uaccess.h>
5462
5463 @@ -225,6 +232,28 @@ int generic_permission(struct inode *ino
5464         return -EACCES;
5465  }
5466
5467 +static inline int dx_barrier(struct inode *inode)
5468 +{
5469 +       if (IS_BARRIER(inode) && !vx_check(0, VS_ADMIN)) {
5470 +               vxwprintk_task(1, "did hit the barrier.");
5471 +               return 1;
5472 +       }
5473 +       return 0;
5474 +}
5475 +
5476 +static inline int dx_permission(struct inode *inode, int mask, struct nameidata *nd)
5477 +{
5478 +       if (dx_barrier(inode))
5479 +               return -EACCES;
5480 +       if (dx_notagcheck(nd) ||
5481 +           dx_check(inode->i_tag, DX_HOSTID|DX_ADMIN|DX_WATCH|DX_IDENT))
5482 +               return 0;
5483 +
5484 +       vxwprintk_task(1, "denied access to %p[#%d,%lu] »%s«.",
5485 +               inode, inode->i_tag, inode->i_ino, vxd_cond_path(nd));
5486 +       return -EACCES;
5487 +}
5488 +
5489  int permission(struct inode *inode, int mask, struct nameidata *nd)
5490  {
5491         int retval, submask;
5492 @@ -239,14 +268,14 @@ int permission(struct inode *inode, int
5493                 /*
5494                  * Nobody gets write access to a read-only fs.
5495                  */
5496 -               if (IS_RDONLY(inode) &&
5497 +               if ((IS_RDONLY(inode) || (nd && MNT_IS_RDONLY(nd->path.mnt))) &&
5498                     (S_ISREG(mode) || S_ISDIR(mode) || S_ISLNK(mode)))
5499                         return -EROFS;
5500
5501                 /*
5502                  * Nobody gets write access to an immutable file.
5503                  */
5504 -               if (IS_IMMUTABLE(inode))
5505 +               if (IS_IMMUTABLE(inode) && !IS_COW(inode))
5506                         return -EACCES;
5507         }
5508
5509 @@ -261,6 +290,11 @@ int permission(struct inode *inode, int
5510
5511         /* Ordinary permission routines do not understand MAY_APPEND. */
5512         submask = mask & ~MAY_APPEND;
5513 +       if ((inode->i_sb->s_magic != DEVPTS_SUPER_MAGIC) &&
5514 +               (inode->i_sb->s_magic != PROC_SUPER_MAGIC) &&
5515 +               (retval = dx_permission(inode, mask, nd)))
5516 +               return retval;
5517 +
5518         if (inode->i_op && inode->i_op->permission) {
5519                 retval = inode->i_op->permission(inode, submask, nd);
5520                 if (!retval) {
5521 @@ -459,6 +493,8 @@ static int exec_permission_lite(struct i
5522  {
5523         umode_t mode = inode->i_mode;
5524
5525 +       if (dx_barrier(inode))
5526 +               return -EACCES;
5527         if (inode->i_op && inode->i_op->permission)
5528                 return -EAGAIN;
5529
5530 @@ -789,7 +825,8 @@ static __always_inline void follow_dotdo
5531                 if (nd->path.dentry == fs->root.dentry &&
5532                     nd->path.mnt == fs->root.mnt) {
5533                          read_unlock(&fs->lock);
5534 -                       break;
5535 +                       /* for sane '/' avoid follow_mount() */
5536 +                       return;
5537                 }
5538                  read_unlock(&fs->lock);
5539                 spin_lock(&dcache_lock);
5540 @@ -826,16 +863,39 @@ static int do_lookup(struct nameidata *n
5541  {
5542         struct vfsmount *mnt = nd->path.mnt;
5543         struct dentry *dentry = __d_lookup(nd->path.dentry, name);
5544 +       struct inode *inode;
5545
5546         if (!dentry)
5547                 goto need_lookup;
5548         if (dentry->d_op && dentry->d_op->d_revalidate)
5549                 goto need_revalidate;
5550 +       inode = dentry->d_inode;
5551 +       if (!inode)
5552 +               goto done;
5553 +
5554 +       if (inode->i_sb->s_magic == PROC_SUPER_MAGIC) {
5555 +               struct proc_dir_entry *de = PDE(inode);
5556 +
5557 +               if (de && !vx_hide_check(0, de->vx_flags))
5558 +                       goto hidden;
5559 +       } else if (inode->i_sb->s_magic == DEVPTS_SUPER_MAGIC) {
5560 +               if (!vx_check((xid_t)inode->i_tag, VS_WATCH_P | VS_IDENT))
5561 +                       goto hidden;
5562 +       } else {
5563 +               if (!dx_notagcheck(nd) && !dx_check(inode->i_tag,
5564 +                       DX_WATCH | DX_ADMIN | DX_HOSTID | DX_IDENT))
5565 +                       goto hidden;
5566 +       }
5567  done:
5568         path->mnt = mnt;
5569         path->dentry = dentry;
5570         __follow_mount(path);
5571         return 0;
5572 +hidden:
5573 +       vxwprintk_task(1, "did lookup hidden %p[#%d,%lu] »%s«.",
5574 +               inode, inode->i_tag, inode->i_ino, vxd_path(path));
5575 +       dput(dentry);
5576 +       return -ENOENT;
5577
5578  need_lookup:
5579         dentry = real_lookup(nd->path.dentry, name, nd);
5580 @@ -1464,7 +1524,8 @@ static inline int check_sticky(struct in
5581   * 10. We don't allow removal of NFS sillyrenamed files; it's handled by
5582   *     nfs_async_unlink().
5583   */
5584 -static int may_delete(struct inode *dir,struct dentry *victim,int isdir)
5585 +static int may_delete(struct inode *dir, struct dentry *victim,
5586 +       int isdir, struct nameidata *nd)
5587  {
5588         int error;
5589
5590 @@ -1474,13 +1535,13 @@ static int may_delete(struct inode *dir,
5591         BUG_ON(victim->d_parent->d_inode != dir);
5592         audit_inode_child(victim->d_name.name, victim, dir);
5593
5594 -       error = permission(dir,MAY_WRITE | MAY_EXEC, NULL);
5595 +       error = permission(dir,MAY_WRITE | MAY_EXEC, nd);
5596         if (error)
5597                 return error;
5598         if (IS_APPEND(dir))
5599                 return -EPERM;
5600         if (check_sticky(dir, victim->d_inode)||IS_APPEND(victim->d_inode)||
5601 -           IS_IMMUTABLE(victim->d_inode))
5602 +               IS_IXORUNLINK(victim->d_inode))
5603                 return -EPERM;
5604         if (isdir) {
5605                 if (!S_ISDIR(victim->d_inode->i_mode))
5606 @@ -1626,6 +1687,14 @@ int may_open(struct nameidata *nd, int a
5607         } else if (IS_RDONLY(inode) && (acc_mode & MAY_WRITE))
5608                 return -EROFS;
5609
5610 +#ifdef CONFIG_VSERVER_COWBL
5611 +       if (IS_COW(inode) && (flag & FMODE_WRITE)) {
5612 +               if (IS_COW_LINK(inode))
5613 +                       return -EMLINK;
5614 +               inode->i_flags &= ~(S_IXUNLINK|S_IMMUTABLE);
5615 +               mark_inode_dirty(inode);
5616 +       }
5617 +#endif
5618         error = vfs_permission(nd, acc_mode);
5619         if (error)
5620                 return error;
5621 @@ -1717,6 +1786,11 @@ int open_namei(int dfd, const char *path
5622         struct dentry *dir;
5623         int count = 0;
5624
5625 +#ifdef CONFIG_VSERVER_COWBL
5626 +       int rflag = flag;
5627 +       int rmode = mode;
5628 +restart:
5629 +#endif
5630         acc_mode = ACC_MODE(flag);
5631
5632         /* O_TRUNC implies we need access checks for write permissions */
5633 @@ -1810,6 +1884,22 @@ do_last:
5634                 goto exit;
5635  ok:
5636         error = may_open(nd, acc_mode, flag);
5637 +#ifdef CONFIG_VSERVER_COWBL
5638 +       if (error == -EMLINK) {
5639 +               struct dentry *dentry;
5640 +               dentry = cow_break_link(pathname);
5641 +               if (IS_ERR(dentry)) {
5642 +                       error = PTR_ERR(dentry);
5643 +                       goto exit;
5644 +               }
5645 +               dput(dentry);
5646 +               release_open_intent(nd);
5647 +               path_put(&nd->path);
5648 +               flag = rflag;
5649 +               mode = rmode;
5650 +               goto restart;
5651 +       }
5652 +#endif
5653         if (error)
5654                 goto exit;
5655         return 0;
5656 @@ -1921,16 +2011,25 @@ fail:
5657  }
5658  EXPORT_SYMBOL_GPL(lookup_create);
5659
5660 -int vfs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t dev)
5661 +int vfs_mknod(struct inode *dir, struct dentry *dentry,
5662 +       int mode, dev_t dev, struct nameidata *nd)
5663  {
5664 -       int error = may_create(dir, dentry, NULL);
5665 +       int error = may_create(dir, dentry, nd);
5666
5667         if (error)
5668                 return error;
5669
5670 -       if ((S_ISCHR(mode) || S_ISBLK(mode)) && !capable(CAP_MKNOD))
5671 +       if (!(S_ISCHR(mode) || S_ISBLK(mode)))
5672 +               goto okay;
5673 +
5674 +       if (!capable(CAP_MKNOD))
5675                 return -EPERM;
5676
5677 +       if (S_ISCHR(mode) && !vs_chrdev_perm(dev, DATTR_CREATE))
5678 +               return -EPERM;
5679 +       if (S_ISBLK(mode) && !vs_blkdev_perm(dev, DATTR_CREATE))
5680 +               return -EPERM;
5681 +okay:
5682         if (!dir->i_op || !dir->i_op->mknod)
5683                 return -EPERM;
5684
5685 @@ -1973,11 +2072,12 @@ asmlinkage long sys_mknodat(int dfd, con
5686                         error = vfs_create(nd.path.dentry->d_inode,dentry,mode,&nd);
5687                         break;
5688                 case S_IFCHR: case S_IFBLK:
5689 -                       error = vfs_mknod(nd.path.dentry->d_inode,dentry,mode,
5690 -                                       new_decode_dev(dev));
5691 +                       error = vfs_mknod(nd.path.dentry->d_inode, dentry, mode,
5692 +                                       new_decode_dev(dev), &nd);
5693                         break;
5694                 case S_IFIFO: case S_IFSOCK:
5695 -                       error = vfs_mknod(nd.path.dentry->d_inode,dentry,mode,0);
5696 +                       error = vfs_mknod(nd.path.dentry->d_inode, dentry, mode,
5697 +                                       0, &nd);
5698                         break;
5699                 case S_IFDIR:
5700                         error = -EPERM;
5701 @@ -2000,9 +2100,10 @@ asmlinkage long sys_mknod(const char __u
5702         return sys_mknodat(AT_FDCWD, filename, mode, dev);
5703  }
5704
5705 -int vfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
5706 +int vfs_mkdir(struct inode *dir, struct dentry *dentry,
5707 +       int mode, struct nameidata *nd)
5708  {
5709 -       int error = may_create(dir, dentry, NULL);
5710 +       int error = may_create(dir, dentry, nd);
5711
5712         if (error)
5713                 return error;
5714 @@ -2044,7 +2145,7 @@ asmlinkage long sys_mkdirat(int dfd, con
5715
5716         if (!IS_POSIXACL(nd.path.dentry->d_inode))
5717                 mode &= ~current->fs->umask;
5718 -       error = vfs_mkdir(nd.path.dentry->d_inode, dentry, mode);
5719 +       error = vfs_mkdir(nd.path.dentry->d_inode, dentry, mode, &nd);
5720         dput(dentry);
5721  out_unlock:
5722         mutex_unlock(&nd.path.dentry->d_inode->i_mutex);
5723 @@ -2087,9 +2188,10 @@ void dentry_unhash(struct dentry *dentry
5724         spin_unlock(&dcache_lock);
5725  }
5726
5727 -int vfs_rmdir(struct inode *dir, struct dentry *dentry)
5728 +int vfs_rmdir(struct inode *dir, struct dentry *dentry,
5729 +       struct nameidata *nd)
5730  {
5731 -       int error = may_delete(dir, dentry, 1);
5732 +       int error = may_delete(dir, dentry, 1, nd);
5733
5734         if (error)
5735                 return error;
5736 @@ -2151,7 +2253,7 @@ static long do_rmdir(int dfd, const char
5737         error = PTR_ERR(dentry);
5738         if (IS_ERR(dentry))
5739                 goto exit2;
5740 -       error = vfs_rmdir(nd.path.dentry->d_inode, dentry);
5741 +       error = vfs_rmdir(nd.path.dentry->d_inode, dentry, &nd);
5742         dput(dentry);
5743  exit2:
5744         mutex_unlock(&nd.path.dentry->d_inode->i_mutex);
5745 @@ -2167,9 +2269,10 @@ asmlinkage long sys_rmdir(const char __u
5746         return do_rmdir(AT_FDCWD, pathname);
5747  }
5748
5749 -int vfs_unlink(struct inode *dir, struct dentry *dentry)
5750 +int vfs_unlink(struct inode *dir, struct dentry *dentry,
5751 +       struct nameidata *nd)
5752  {
5753 -       int error = may_delete(dir, dentry, 0);
5754 +       int error = may_delete(dir, dentry, 0, nd);
5755
5756         if (error)
5757                 return error;
5758 @@ -2232,7 +2335,7 @@ static long do_unlinkat(int dfd, const c
5759                 inode = dentry->d_inode;
5760                 if (inode)
5761                         atomic_inc(&inode->i_count);
5762 -               error = vfs_unlink(nd.path.dentry->d_inode, dentry);
5763 +               error = vfs_unlink(nd.path.dentry->d_inode, dentry, &nd);
5764         exit2:
5765                 dput(dentry);
5766         }
5767 @@ -2267,9 +2370,10 @@ asmlinkage long sys_unlink(const char __
5768         return do_unlinkat(AT_FDCWD, pathname);
5769  }
5770
5771 -int vfs_symlink(struct inode *dir, struct dentry *dentry, const char *oldname, int mode)
5772 +int vfs_symlink(struct inode *dir, struct dentry *dentry,
5773 +       const char *oldname, int mode, struct nameidata *nd)
5774  {
5775 -       int error = may_create(dir, dentry, NULL);
5776 +       int error = may_create(dir, dentry, nd);
5777
5778         if (error)
5779                 return error;
5780 @@ -2313,7 +2417,8 @@ asmlinkage long sys_symlinkat(const char
5781         if (IS_ERR(dentry))
5782                 goto out_unlock;
5783
5784 -       error = vfs_symlink(nd.path.dentry->d_inode, dentry, from, S_IALLUGO);
5785 +       error = vfs_symlink(nd.path.dentry->d_inode, dentry, from,
5786 +               S_IALLUGO, &nd);
5787         dput(dentry);
5788  out_unlock:
5789         mutex_unlock(&nd.path.dentry->d_inode->i_mutex);
5790 @@ -2330,7 +2435,8 @@ asmlinkage long sys_symlink(const char _
5791         return sys_symlinkat(oldname, AT_FDCWD, newname);
5792  }
5793
5794 -int vfs_link(struct dentry *old_dentry, struct inode *dir, struct dentry *new_dentry)
5795 +int vfs_link(struct dentry *old_dentry, struct inode *dir,
5796 +       struct dentry *new_dentry, struct nameidata *nd)
5797  {
5798         struct inode *inode = old_dentry->d_inode;
5799         int error;
5800 @@ -2338,7 +2444,7 @@ int vfs_link(struct dentry *old_dentry,
5801         if (!inode)
5802                 return -ENOENT;
5803
5804 -       error = may_create(dir, new_dentry, NULL);
5805 +       error = may_create(dir, new_dentry, nd);
5806         if (error)
5807                 return error;
5808
5809 @@ -2348,7 +2454,7 @@ int vfs_link(struct dentry *old_dentry,
5810         /*
5811          * A link to an append-only or immutable file cannot be created.
5812          */
5813 -       if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
5814 +       if (IS_APPEND(inode) || IS_IXORUNLINK(inode))
5815                 return -EPERM;
5816         if (!dir->i_op || !dir->i_op->link)
5817                 return -EPERM;
5818 @@ -2408,7 +2514,8 @@ asmlinkage long sys_linkat(int olddfd, c
5819         error = PTR_ERR(new_dentry);
5820         if (IS_ERR(new_dentry))
5821                 goto out_unlock;
5822 -       error = vfs_link(old_nd.path.dentry, nd.path.dentry->d_inode, new_dentry);
5823 +       error = vfs_link(old_nd.path.dentry, nd.path.dentry->d_inode,
5824 +               new_dentry, &nd);
5825         dput(new_dentry);
5826  out_unlock:
5827         mutex_unlock(&nd.path.dentry->d_inode->i_mutex);
5828 @@ -2540,14 +2647,14 @@ int vfs_rename(struct inode *old_dir, st
5829         if (old_dentry->d_inode == new_dentry->d_inode)
5830                 return 0;
5831
5832 -       error = may_delete(old_dir, old_dentry, is_dir);
5833 +       error = may_delete(old_dir, old_dentry, is_dir, NULL);
5834         if (error)
5835                 return error;
5836
5837         if (!new_dentry->d_inode)
5838                 error = may_create(new_dir, new_dentry, NULL);
5839         else
5840 -               error = may_delete(new_dir, new_dentry, is_dir);
5841 +               error = may_delete(new_dir, new_dentry, is_dir, NULL);
5842         if (error)
5843                 return error;
5844
5845 @@ -2625,6 +2732,9 @@ static int do_rename(int olddfd, const c
5846         error = -EINVAL;
5847         if (old_dentry == trap)
5848                 goto exit4;
5849 +       error = -EROFS;
5850 +       if (MNT_IS_RDONLY(newnd.path.mnt))
5851 +               goto exit4;
5852         new_dentry = lookup_hash(&newnd);
5853         error = PTR_ERR(new_dentry);
5854         if (IS_ERR(new_dentry))
5855 @@ -2718,6 +2828,214 @@ int vfs_follow_link(struct nameidata *nd
5856         return __vfs_follow_link(nd, link);
5857  }
5858
5859 +
5860 +#ifdef CONFIG_VSERVER_COWBL
5861 +
5862 +#include <linux/file.h>
5863 +
5864 +static inline
5865 +long do_cow_splice(struct file *in, struct file *out, size_t len)
5866 +{
5867 +       loff_t ppos = 0;
5868 +
5869 +       return do_splice_direct(in, &ppos, out, len, 0);
5870 +}
5871 +
5872 +struct dentry *cow_break_link(const char *pathname)
5873 +{
5874 +       int ret, mode, pathlen, redo = 0;
5875 +       struct nameidata old_nd, dir_nd;
5876 +       struct path old_path, new_path;
5877 +       struct dentry *dir, *res = NULL;
5878 +       struct file *old_file;
5879 +       struct file *new_file;
5880 +       char *to, *path, pad='\251';
5881 +       loff_t size;
5882 +
5883 +       vxdprintk(VXD_CBIT(misc, 1), "cow_break_link(»%s«)", pathname);
5884 +       path = kmalloc(PATH_MAX, GFP_KERNEL);
5885 +       ret = -ENOMEM;
5886 +       if (!path)
5887 +               goto out;
5888 +
5889 +       /* old_nd will have refs to dentry and mnt */
5890 +       ret = path_lookup(pathname, LOOKUP_FOLLOW, &old_nd);
5891 +       vxdprintk(VXD_CBIT(misc, 2), "path_lookup(old): %d", ret);
5892 +       if (ret < 0)
5893 +               goto out_free_path;
5894 +
5895 +       old_path = old_nd.path;
5896 +       mode = old_path.dentry->d_inode->i_mode;
5897 +
5898 +       to = d_path(&old_path, path, PATH_MAX-2);
5899 +       pathlen = strlen(to);
5900 +       vxdprintk(VXD_CBIT(misc, 2), "old path »%s« [»%.*s«:%d]", to,
5901 +               old_path.dentry->d_name.len, old_path.dentry->d_name.name,
5902 +               old_path.dentry->d_name.len);
5903 +
5904 +       to[pathlen + 1] = 0;
5905 +retry:
5906 +       to[pathlen] = pad--;
5907 +       ret = -EMLINK;
5908 +       if (pad <= '\240')
5909 +               goto out_rel_old;
5910 +
5911 +       vxdprintk(VXD_CBIT(misc, 1), "temp copy »%s«", to);
5912 +       /* dir_nd will have refs to dentry and mnt */
5913 +       ret = path_lookup(to,
5914 +               LOOKUP_PARENT | LOOKUP_OPEN | LOOKUP_CREATE, &dir_nd);
5915 +       vxdprintk(VXD_CBIT(misc, 2),
5916 +               "path_lookup(new): %d", ret);
5917 +       if (ret < 0)
5918 +               goto retry;
5919 +
5920 +       /* this puppy downs the inode mutex */
5921 +       new_path.dentry = lookup_create(&dir_nd, 0);
5922 +       vxdprintk(VXD_CBIT(misc, 2),
5923 +               "lookup_create(new): %p [»%.*s«:%d]", new_path.dentry,
5924 +               new_path.dentry->d_name.len, new_path.dentry->d_name.name,
5925 +               new_path.dentry->d_name.len);
5926 +       if (!new_path.dentry || IS_ERR(new_path.dentry)) {
5927 +               path_put(&dir_nd.path);
5928 +               goto retry;
5929 +       }
5930 +       dir = dir_nd.path.dentry;
5931 +
5932 +       ret = vfs_create(dir_nd.path.dentry->d_inode, new_path.dentry, mode, &dir_nd);
5933 +       vxdprintk(VXD_CBIT(misc, 2),
5934 +               "vfs_create(new): %d", ret);
5935 +       if (ret == -EEXIST) {
5936 +               mutex_unlock(&dir->d_inode->i_mutex);
5937 +               dput(new_path.dentry);
5938 +               path_put(&dir_nd.path);
5939 +               goto retry;
5940 +       }
5941 +       else if (ret < 0)
5942 +               goto out_unlock_new;
5943 +
5944 +       /* drop out early, ret passes ENOENT */
5945 +       ret = -ENOENT;
5946 +       if ((redo = d_unhashed(old_path.dentry)))
5947 +               goto out_unlock_new;
5948 +
5949 +       new_path.mnt = dir_nd.path.mnt;
5950 +       dget(old_path.dentry);
5951 +       mntget(old_path.mnt);
5952 +       /* this one cleans up the dentry/mnt in case of failure */
5953 +       old_file = dentry_open(old_path.dentry, old_path.mnt, O_RDONLY);
5954 +       vxdprintk(VXD_CBIT(misc, 2),
5955 +               "dentry_open(old): %p", old_file);
5956 +       if (!old_file || IS_ERR(old_file)) {
5957 +               res = IS_ERR(old_file) ? (void *) old_file : res;
5958 +               goto out_unlock_new;
5959 +       }
5960 +
5961 +       dget(new_path.dentry);
5962 +       mntget(new_path.mnt);
5963 +       /* this one cleans up the dentry/mnt in case of failure */
5964 +       new_file = dentry_open(new_path.dentry, new_path.mnt, O_WRONLY);
5965 +       vxdprintk(VXD_CBIT(misc, 2),
5966 +               "dentry_open(new): %p", new_file);
5967 +
5968 +       ret = IS_ERR(new_file) ? PTR_ERR(new_file) : -ENOENT;
5969 +       if (!new_file || IS_ERR(new_file))
5970 +               goto out_fput_old;
5971 +
5972 +       size = i_size_read(old_file->f_dentry->d_inode);
5973 +       ret = do_cow_splice(old_file, new_file, size);
5974 +       vxdprintk(VXD_CBIT(misc, 2), "do_splice_direct: %d", ret);
5975 +       if (ret < 0) {
5976 +               goto out_fput_both;
5977 +       } else if (ret < size) {
5978 +               ret = -ENOSPC;
5979 +               goto out_fput_both;
5980 +       } else {
5981 +               struct inode *old_inode = old_path.dentry->d_inode;
5982 +               struct inode *new_inode = new_path.dentry->d_inode;
5983 +               struct iattr attr = {
5984 +                       .ia_uid = old_inode->i_uid,
5985 +                       .ia_gid = old_inode->i_gid,
5986 +                       .ia_valid = ATTR_UID | ATTR_GID
5987 +                       };
5988 +
5989 +               ret = inode_setattr(new_inode, &attr);
5990 +               if (ret)
5991 +                       goto out_fput_both;
5992 +       }
5993 +
5994 +       mutex_lock(&old_path.dentry->d_inode->i_sb->s_vfs_rename_mutex);
5995 +
5996 +       /* drop out late */
5997 +       ret = -ENOENT;
5998 +       if ((redo = d_unhashed(old_path.dentry)))
5999 +               goto out_unlock;
6000 +
6001 +       vxdprintk(VXD_CBIT(misc, 2),
6002 +               "vfs_rename: [»%*s«:%d] -> [»%*s«:%d]",
6003 +               new_path.dentry->d_name.len, new_path.dentry->d_name.name,
6004 +               new_path.dentry->d_name.len,
6005 +               old_path.dentry->d_name.len, old_path.dentry->d_name.name,
6006 +               old_path.dentry->d_name.len);
6007 +       ret = vfs_rename(dir_nd.path.dentry->d_inode, new_path.dentry,
6008 +               old_nd.path.dentry->d_parent->d_inode, old_path.dentry);
6009 +       vxdprintk(VXD_CBIT(misc, 2), "vfs_rename: %d", ret);
6010 +       res = new_path.dentry;
6011 +
6012 +out_unlock:
6013 +       mutex_unlock(&old_path.dentry->d_inode->i_sb->s_vfs_rename_mutex);
6014 +
6015 +out_fput_both:
6016 +       vxdprintk(VXD_CBIT(misc, 3),
6017 +               "fput(new_file=%p[#%d])", new_file,
6018 +               atomic_read(&new_file->f_count));
6019 +       fput(new_file);
6020 +
6021 +out_fput_old:
6022 +       vxdprintk(VXD_CBIT(misc, 3),
6023 +               "fput(old_file=%p[#%d])", old_file,
6024 +               atomic_read(&old_file->f_count));
6025 +       fput(old_file);
6026 +
6027 +out_unlock_new:
6028 +       mutex_unlock(&dir->d_inode->i_mutex);
6029 +       if (!ret)
6030 +               goto out_redo;
6031 +
6032 +       /* error path cleanup */
6033 +       vfs_unlink(dir->d_inode, new_path.dentry, &dir_nd);
6034 +       dput(new_path.dentry);
6035 +
6036 +out_redo:
6037 +       if (!redo)
6038 +               goto out_rel_both;
6039 +       /* lookup dentry once again */
6040 +       path_put(&old_nd.path);
6041 +       ret = path_lookup(pathname, LOOKUP_FOLLOW, &old_nd);
6042 +       if (ret)
6043 +               goto out_rel_both;
6044 +
6045 +       new_path.dentry = old_nd.path.dentry;
6046 +       vxdprintk(VXD_CBIT(misc, 2),
6047 +               "path_lookup(redo): %p [»%.*s«:%d]", new_path.dentry,
6048 +               new_path.dentry->d_name.len, new_path.dentry->d_name.name,
6049 +               new_path.dentry->d_name.len);
6050 +       dget(new_path.dentry);
6051 +       res = new_path.dentry;
6052 +
6053 +out_rel_both:
6054 +       path_put(&dir_nd.path);
6055 +out_rel_old:
6056 +       path_put(&old_nd.path);
6057 +out_free_path:
6058 +       kfree(path);
6059 +out:
6060 +       if (ret)
6061 +               res = ERR_PTR(ret);
6062 +       return res;
6063 +}
6064 +
6065 +#endif
6066 +
6067  /* get the link contents into pagecache */
6068  static char *page_getlink(struct dentry * dentry, struct page **ppage)
6069  {
6070 diff -NurpP --minimal linux-2.6.25.11/fs/namespace.c linux-2.6.25.11-vs2.3.0.34.14/fs/namespace.c
6071 --- linux-2.6.25.11/fs/namespace.c      2008-04-17 12:05:40.000000000 -0400
6072 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/namespace.c        2008-04-21 17:20:53.000000000 -0400
6073 @@ -26,6 +26,11 @@
6074  #include <linux/mount.h>
6075  #include <linux/ramfs.h>
6076  #include <linux/log2.h>
6077 +#include <linux/vs_base.h>
6078 +#include <linux/vs_context.h>
6079 +#include <linux/vs_tag.h>
6080 +#include <linux/vserver/space.h>
6081 +#include <linux/vserver/global.h>
6082  #include <asm/uaccess.h>
6083  #include <asm/unistd.h>
6084  #include "pnode.h"
6085 @@ -244,6 +249,7 @@ static struct vfsmount *clone_mnt(struct
6086                 mnt->mnt_root = dget(root);
6087                 mnt->mnt_mountpoint = mnt->mnt_root;
6088                 mnt->mnt_parent = mnt;
6089 +               mnt->mnt_tag = old->mnt_tag;
6090
6091                 if (flag & CL_SLAVE) {
6092                         list_add(&mnt->mnt_slave, &old->mnt_slave_list);
6093 @@ -323,6 +329,31 @@ static inline void mangle(struct seq_fil
6094         seq_escape(m, s, " \t\n\\");
6095  }
6096
6097 +static int mnt_is_reachable(struct vfsmount *mnt)
6098 +{
6099 +       struct path root;
6100 +       struct dentry *point;
6101 +       int ret;
6102 +
6103 +       if (mnt == mnt->mnt_ns->root)
6104 +               return 1;
6105 +
6106 +       spin_lock(&vfsmount_lock);
6107 +       root = current->fs->root;
6108 +       point = root.dentry;
6109 +
6110 +       while ((mnt != mnt->mnt_parent) && (mnt != root.mnt)) {
6111 +               point = mnt->mnt_mountpoint;
6112 +               mnt = mnt->mnt_parent;
6113 +       }
6114 +
6115 +       ret = (mnt == root.mnt) && is_subdir(point, root.dentry);
6116 +
6117 +       spin_unlock(&vfsmount_lock);
6118 +
6119 +       return ret;
6120 +}
6121 +
6122  /*
6123   * Simple .show_options callback for filesystems which don't want to
6124   * implement more complex mount option showing.
6125 @@ -388,44 +419,61 @@ static int show_vfsmnt(struct seq_file *
6126         struct vfsmount *mnt = list_entry(v, struct vfsmount, mnt_list);
6127         int err = 0;
6128         static struct proc_fs_info {
6129 -               int flag;
6130 -               char *str;
6131 +               int s_flag;
6132 +               int mnt_flag;
6133 +               char *set_str;
6134 +               char *unset_str;
6135         } fs_info[] = {
6136 -               { MS_SYNCHRONOUS, ",sync" },
6137 -               { MS_DIRSYNC, ",dirsync" },
6138 -               { MS_MANDLOCK, ",mand" },
6139 -               { 0, NULL }
6140 -       };
6141 -       static struct proc_fs_info mnt_info[] = {
6142 -               { MNT_NOSUID, ",nosuid" },
6143 -               { MNT_NODEV, ",nodev" },
6144 -               { MNT_NOEXEC, ",noexec" },
6145 -               { MNT_NOATIME, ",noatime" },
6146 -               { MNT_NODIRATIME, ",nodiratime" },
6147 -               { MNT_RELATIME, ",relatime" },
6148 -               { 0, NULL }
6149 +               { MS_RDONLY, MNT_RDONLY, "ro", "rw" },
6150 +               { MS_SYNCHRONOUS, 0, ",sync", NULL },
6151 +               { MS_DIRSYNC, 0, ",dirsync", NULL },
6152 +               { MS_MANDLOCK, 0, ",mand", NULL },
6153 +               { MS_TAGGED, 0, ",tag", NULL },
6154 +               { MS_NOATIME, MNT_NOATIME, ",noatime", NULL },
6155 +               { MS_NODIRATIME, MNT_NODIRATIME, ",nodiratime", NULL },
6156 +               { MS_RELATIME, MNT_RELATIME, ",relatime", NULL },
6157 +               { 0, MNT_NOSUID, ",nosuid", NULL },
6158 +               { 0, MNT_NODEV, ",nodev", NULL },
6159 +               { 0, MNT_NOEXEC, ",noexec", NULL },
6160 +               { 0, 0, NULL, NULL }
6161         };
6162 -       struct proc_fs_info *fs_infop;
6163 -       struct path mnt_path = { .dentry = mnt->mnt_root, .mnt = mnt };
6164 +       struct proc_fs_info *p;
6165 +       unsigned long s_flags = mnt->mnt_sb->s_flags;
6166 +       int mnt_flags = mnt->mnt_flags;
6167
6168 -       mangle(m, mnt->mnt_devname ? mnt->mnt_devname : "none");
6169 -       seq_putc(m, ' ');
6170 -       seq_path(m, &mnt_path, " \t\n\\");
6171 -       seq_putc(m, ' ');
6172 -       mangle(m, mnt->mnt_sb->s_type->name);
6173 -       if (mnt->mnt_sb->s_subtype && mnt->mnt_sb->s_subtype[0]) {
6174 -               seq_putc(m, '.');
6175 -               mangle(m, mnt->mnt_sb->s_subtype);
6176 -       }
6177 -       seq_puts(m, mnt->mnt_sb->s_flags & MS_RDONLY ? " ro" : " rw");
6178 -       for (fs_infop = fs_info; fs_infop->flag; fs_infop++) {
6179 -               if (mnt->mnt_sb->s_flags & fs_infop->flag)
6180 -                       seq_puts(m, fs_infop->str);
6181 +       if (vx_flags(VXF_HIDE_MOUNT, 0))
6182 +               return 0;
6183 +       if (!mnt_is_reachable(mnt) && !vx_check(0, VS_WATCH_P))
6184 +               return 0;
6185 +
6186 +       if (!vx_check(0, VS_ADMIN|VS_WATCH) &&
6187 +               mnt == current->fs->root.mnt) {
6188 +               seq_puts(m, "/dev/root / ");
6189 +       } else {
6190 +               struct path mnt_path = { .dentry = mnt->mnt_root, .mnt = mnt };
6191 +               mangle(m, mnt->mnt_devname ? mnt->mnt_devname : "none");
6192 +               seq_putc(m, ' ');
6193 +               seq_path(m, &mnt_path, " \t\n\\");
6194 +               seq_putc(m, ' ');
6195 +
6196 +               if (mnt->mnt_sb->s_subtype && mnt->mnt_sb->s_subtype[0]) {
6197 +                       seq_putc(m, '.');
6198 +                       mangle(m, mnt->mnt_sb->s_subtype);
6199 +               }
6200         }
6201 -       for (fs_infop = mnt_info; fs_infop->flag; fs_infop++) {
6202 -               if (mnt->mnt_flags & fs_infop->flag)
6203 -                       seq_puts(m, fs_infop->str);
6204 +       mangle(m, mnt->mnt_sb->s_type->name);
6205 +       seq_putc(m, ' ');
6206 +       for (p = fs_info; (p->s_flag | p->mnt_flag) ; p++) {
6207 +               if ((s_flags & p->s_flag) || (mnt_flags & p->mnt_flag)) {
6208 +                       if (p->set_str)
6209 +                               seq_puts(m, p->set_str);
6210 +               } else {
6211 +                       if (p->unset_str)
6212 +                               seq_puts(m, p->unset_str);
6213 +               }
6214         }
6215 +       if (mnt->mnt_flags & MNT_TAGID)
6216 +               seq_printf(m, ",tag=%d", mnt->mnt_tag);
6217         if (mnt->mnt_sb->s_op->show_options)
6218                 err = mnt->mnt_sb->s_op->show_options(m, mnt);
6219         seq_puts(m, " 0 0\n");
6220 @@ -445,17 +493,27 @@ static int show_vfsstat(struct seq_file
6221         struct path mnt_path = { .dentry = mnt->mnt_root, .mnt = mnt };
6222         int err = 0;
6223
6224 -       /* device */
6225 -       if (mnt->mnt_devname) {
6226 -               seq_puts(m, "device ");
6227 -               mangle(m, mnt->mnt_devname);
6228 -       } else
6229 -               seq_puts(m, "no device");
6230 +       if (vx_flags(VXF_HIDE_MOUNT, 0))
6231 +               return 0;
6232 +       if (!mnt_is_reachable(mnt) && !vx_check(0, VS_WATCH_P))
6233 +               return 0;
6234
6235 -       /* mount point */
6236 -       seq_puts(m, " mounted on ");
6237 -       seq_path(m, &mnt_path, " \t\n\\");
6238 -       seq_putc(m, ' ');
6239 +       if (!vx_check(0, VS_ADMIN|VS_WATCH) &&
6240 +               mnt == current->fs->root.mnt) {
6241 +               seq_puts(m, "device /dev/root mounted on / ");
6242 +       } else {
6243 +               /* device */
6244 +               if (mnt->mnt_devname) {
6245 +                       seq_puts(m, "device ");
6246 +                       mangle(m, mnt->mnt_devname);
6247 +               } else
6248 +                       seq_puts(m, "no device");
6249 +
6250 +               /* mount point */
6251 +               seq_puts(m, " mounted on ");
6252 +               seq_path(m, &mnt_path, " \t\n\\");
6253 +               seq_putc(m, ' ');
6254 +       }
6255
6256         /* file system type */
6257         seq_puts(m, "with fstype ");
6258 @@ -693,7 +751,7 @@ asmlinkage long sys_umount(char __user *
6259                 goto dput_and_out;
6260
6261         retval = -EPERM;
6262 -       if (!capable(CAP_SYS_ADMIN))
6263 +       if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT))
6264                 goto dput_and_out;
6265
6266         retval = do_umount(nd.path.mnt, flags);
6267 @@ -719,7 +777,7 @@ asmlinkage long sys_oldumount(char __use
6268
6269  static int mount_is_safe(struct nameidata *nd)
6270  {
6271 -       if (capable(CAP_SYS_ADMIN))
6272 +       if (vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT))
6273                 return 0;
6274         return -EPERM;
6275  #ifdef notyet
6276 @@ -974,11 +1032,13 @@ static noinline int do_change_type(struc
6277   * noinline this do_mount helper to save do_mount stack space.
6278   */
6279  static noinline int do_loopback(struct nameidata *nd, char *old_name,
6280 -                               int recurse)
6281 +       tag_t tag, unsigned long flags, int mnt_flags)
6282  {
6283         struct nameidata old_nd;
6284         struct vfsmount *mnt = NULL;
6285         int err = mount_is_safe(nd);
6286 +       int recurse = flags & MS_REC;
6287 +
6288         if (err)
6289                 return err;
6290         if (!old_name || !*old_name)
6291 @@ -1004,6 +1064,12 @@ static noinline int do_loopback(struct n
6292         if (!mnt)
6293                 goto out;
6294
6295 +       mnt->mnt_flags = mnt_flags;
6296 +       if (flags & MS_TAGID) {
6297 +               mnt->mnt_tag = tag;
6298 +               mnt->mnt_flags |= MNT_TAGID;
6299 +       }
6300 +
6301         err = graft_tree(mnt, nd);
6302         if (err) {
6303                 LIST_HEAD(umount_list);
6304 @@ -1012,6 +1078,7 @@ static noinline int do_loopback(struct n
6305                 spin_unlock(&vfsmount_lock);
6306                 release_mounts(&umount_list);
6307         }
6308 +       mnt->mnt_flags = mnt_flags;
6309
6310  out:
6311         up_write(&namespace_sem);
6312 @@ -1026,12 +1093,12 @@ out:
6313   * noinline this do_mount helper to save do_mount stack space.
6314   */
6315  static noinline int do_remount(struct nameidata *nd, int flags, int mnt_flags,
6316 -                     void *data)
6317 +       void *data, xid_t xid)
6318  {
6319         int err;
6320         struct super_block *sb = nd->path.mnt->mnt_sb;
6321
6322 -       if (!capable(CAP_SYS_ADMIN))
6323 +       if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_REMOUNT))
6324                 return -EPERM;
6325
6326         if (!check_mnt(nd->path.mnt))
6327 @@ -1069,7 +1136,7 @@ static noinline int do_move_mount(struct
6328         struct path parent_path;
6329         struct vfsmount *p;
6330         int err = 0;
6331 -       if (!capable(CAP_SYS_ADMIN))
6332 +       if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT))
6333                 return -EPERM;
6334         if (!old_name || !*old_name)
6335                 return -EINVAL;
6336 @@ -1152,7 +1219,7 @@ static noinline int do_new_mount(struct
6337                 return -EINVAL;
6338
6339         /* we need capabilities... */
6340 -       if (!capable(CAP_SYS_ADMIN))
6341 +       if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT))
6342                 return -EPERM;
6343
6344         mnt = do_kern_mount(type, flags, name, data);
6345 @@ -1397,6 +1464,7 @@ long do_mount(char *dev_name, char *dir_
6346         struct nameidata nd;
6347         int retval = 0;
6348         int mnt_flags = 0;
6349 +       tag_t tag = 0;
6350
6351         /* Discard magic */
6352         if ((flags & MS_MGC_MSK) == MS_MGC_VAL)
6353 @@ -1412,7 +1480,17 @@ long do_mount(char *dev_name, char *dir_
6354         if (data_page)
6355                 ((char *)data_page)[PAGE_SIZE - 1] = 0;
6356
6357 +       retval = dx_parse_tag(data_page, &tag, 1);
6358 +       if (retval) {
6359 +               mnt_flags |= retval;
6360 +               /* FIXME: bind and re-mounts get the tag flag? */
6361 +               if (flags & (MS_BIND|MS_REMOUNT))
6362 +                       flags |= MS_TAGID;
6363 +       }
6364 +
6365         /* Separate the per-mountpoint flags */
6366 +       if (flags & MS_RDONLY)
6367 +               mnt_flags |= MNT_RDONLY;
6368         if (flags & MS_NOSUID)
6369                 mnt_flags |= MNT_NOSUID;
6370         if (flags & MS_NODEV)
6371 @@ -1426,6 +1504,8 @@ long do_mount(char *dev_name, char *dir_
6372         if (flags & MS_RELATIME)
6373                 mnt_flags |= MNT_RELATIME;
6374
6375 +       if (!capable(CAP_SYS_ADMIN))
6376 +               mnt_flags |= MNT_NODEV;
6377         flags &= ~(MS_NOSUID | MS_NOEXEC | MS_NODEV | MS_ACTIVE |
6378                    MS_NOATIME | MS_NODIRATIME | MS_RELATIME| MS_KERNMOUNT);
6379
6380 @@ -1440,9 +1520,9 @@ long do_mount(char *dev_name, char *dir_
6381
6382         if (flags & MS_REMOUNT)
6383                 retval = do_remount(&nd, flags & ~MS_REMOUNT, mnt_flags,
6384 -                                   data_page);
6385 +                                   data_page, tag);
6386         else if (flags & MS_BIND)
6387 -               retval = do_loopback(&nd, dev_name, flags & MS_REC);
6388 +               retval = do_loopback(&nd, dev_name, tag, flags, mnt_flags);
6389         else if (flags & (MS_SHARED | MS_PRIVATE | MS_SLAVE | MS_UNBINDABLE))
6390                 retval = do_change_type(&nd, flags);
6391         else if (flags & MS_MOVE)
6392 @@ -1515,6 +1595,7 @@ static struct mnt_namespace *dup_mnt_ns(
6393                 q = next_mnt(q, new_ns->root);
6394         }
6395         up_write(&namespace_sem);
6396 +       atomic_inc(&vs_global_mnt_ns);
6397
6398         if (rootmnt)
6399                 mntput(rootmnt);
6400 @@ -1850,5 +1931,6 @@ void __put_mnt_ns(struct mnt_namespace *
6401         spin_unlock(&vfsmount_lock);
6402         up_write(&namespace_sem);
6403         release_mounts(&umount_list);
6404 +       atomic_dec(&vs_global_mnt_ns);
6405         kfree(ns);
6406  }
6407 diff -NurpP --minimal linux-2.6.25.11/fs/nfs/client.c linux-2.6.25.11-vs2.3.0.34.14/fs/nfs/client.c
6408 --- linux-2.6.25.11/fs/nfs/client.c     2008-04-17 12:05:40.000000000 -0400
6409 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/nfs/client.c       2008-04-19 15:14:52.000000000 -0400
6410 @@ -589,6 +589,9 @@ static int nfs_init_server_rpcclient(str
6411         if (server->flags & NFS_MOUNT_SOFT)
6412                 server->client->cl_softrtry = 1;
6413
6414 +       server->client->cl_tag = 0;
6415 +       if (server->flags & NFS_MOUNT_TAGGED)
6416 +               server->client->cl_tag = 1;
6417         return 0;
6418  }
6419
6420 @@ -742,6 +745,10 @@ static void nfs_server_set_fsinfo(struct
6421                 server->acdirmin = server->acdirmax = 0;
6422         }
6423
6424 +       /* FIXME: needs fsinfo
6425 +       if (server->flags & NFS_MOUNT_TAGGED)
6426 +               sb->s_flags |= MS_TAGGED;       */
6427 +
6428         server->maxfilesize = fsinfo->maxfilesize;
6429
6430         /* We're airborne Set socket buffersize */
6431 diff -NurpP --minimal linux-2.6.25.11/fs/nfs/dir.c linux-2.6.25.11-vs2.3.0.34.14/fs/nfs/dir.c
6432 --- linux-2.6.25.11/fs/nfs/dir.c        2008-04-17 12:05:40.000000000 -0400
6433 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/nfs/dir.c  2008-04-21 16:52:03.000000000 -0400
6434 @@ -34,6 +34,7 @@
6435  #include <linux/namei.h>
6436  #include <linux/mount.h>
6437  #include <linux/sched.h>
6438 +#include <linux/vs_tag.h>
6439
6440  #include "nfs4_fs.h"
6441  #include "delegation.h"
6442 @@ -930,6 +931,7 @@ static struct dentry *nfs_lookup(struct
6443         if (IS_ERR(res))
6444                 goto out_unblock_sillyrename;
6445
6446 +       dx_propagate_tag(nd, inode);
6447  no_entry:
6448         res = d_materialise_unique(dentry, inode);
6449         if (res != NULL) {
6450 @@ -967,7 +969,8 @@ static int is_atomic_open(struct inode *
6451         if (nd->flags & LOOKUP_DIRECTORY)
6452                 return 0;
6453         /* Are we trying to write to a read only partition? */
6454 -       if (IS_RDONLY(dir) && (nd->intent.open.flags & (O_CREAT|O_TRUNC|FMODE_WRITE)))
6455 +       if ((IS_RDONLY(dir) || MNT_IS_RDONLY(nd->path.mnt)) &&
6456 +               (nd->intent.open.flags & (O_CREAT|O_TRUNC|FMODE_WRITE)))
6457                 return 0;
6458         return 1;
6459  }
6460 diff -NurpP --minimal linux-2.6.25.11/fs/nfs/inode.c linux-2.6.25.11-vs2.3.0.34.14/fs/nfs/inode.c
6461 --- linux-2.6.25.11/fs/nfs/inode.c      2008-04-17 12:05:40.000000000 -0400
6462 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/nfs/inode.c        2008-04-19 15:14:52.000000000 -0400
6463 @@ -37,6 +37,7 @@
6464  #include <linux/vfs.h>
6465  #include <linux/inet.h>
6466  #include <linux/nfs_xdr.h>
6467 +#include <linux/vs_tag.h>
6468
6469  #include <asm/system.h>
6470  #include <asm/uaccess.h>
6471 @@ -316,8 +317,10 @@ nfs_fhget(struct super_block *sb, struct
6472                         nfsi->change_attr = fattr->change_attr;
6473                 inode->i_size = nfs_size_to_loff_t(fattr->size);
6474                 inode->i_nlink = fattr->nlink;
6475 -               inode->i_uid = fattr->uid;
6476 -               inode->i_gid = fattr->gid;
6477 +               inode->i_uid = INOTAG_UID(DX_TAG(inode), fattr->uid, fattr->gid);
6478 +               inode->i_gid = INOTAG_GID(DX_TAG(inode), fattr->uid, fattr->gid);
6479 +               inode->i_tag = INOTAG_TAG(DX_TAG(inode), fattr->uid, fattr->gid, 0);
6480 +                                        /* maybe fattr->xid someday */
6481                 if (fattr->valid & (NFS_ATTR_FATTR_V3 | NFS_ATTR_FATTR_V4)) {
6482                         /*
6483                          * report the blocks in 512byte units
6484 @@ -410,6 +413,8 @@ void nfs_setattr_update_inode(struct ino
6485                         inode->i_uid = attr->ia_uid;
6486                 if ((attr->ia_valid & ATTR_GID) != 0)
6487                         inode->i_gid = attr->ia_gid;
6488 +               if ((attr->ia_valid & ATTR_TAG) && IS_TAGGED(inode))
6489 +                       inode->i_tag = attr->ia_tag;
6490                 spin_lock(&inode->i_lock);
6491                 NFS_I(inode)->cache_validity |= NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
6492                 spin_unlock(&inode->i_lock);
6493 @@ -841,6 +846,9 @@ static int nfs_check_inode_attributes(st
6494         struct nfs_inode *nfsi = NFS_I(inode);
6495         loff_t cur_size, new_isize;
6496         unsigned long invalid = 0;
6497 +       uid_t uid;
6498 +       gid_t gid;
6499 +       tag_t tag;
6500
6501
6502         /* Has the inode gone and changed behind our back? */
6503 @@ -865,10 +873,15 @@ static int nfs_check_inode_attributes(st
6504         if (cur_size != new_isize && nfsi->npages == 0)
6505                 invalid |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE;
6506
6507 +       uid = INOTAG_UID(DX_TAG(inode), fattr->uid, fattr->gid);
6508 +       gid = INOTAG_GID(DX_TAG(inode), fattr->uid, fattr->gid);
6509 +       tag = INOTAG_TAG(DX_TAG(inode), fattr->uid, fattr->gid, 0);
6510 +
6511         /* Have any file permissions changed? */
6512         if ((inode->i_mode & S_IALLUGO) != (fattr->mode & S_IALLUGO)
6513 -                       || inode->i_uid != fattr->uid
6514 -                       || inode->i_gid != fattr->gid)
6515 +                       || inode->i_uid != uid
6516 +                       || inode->i_gid != gid
6517 +                       || inode->i_tag != tag)
6518                 invalid |= NFS_INO_INVALID_ATTR | NFS_INO_INVALID_ACCESS | NFS_INO_INVALID_ACL;
6519
6520         /* Has the link count changed? */
6521 @@ -989,6 +1002,9 @@ static int nfs_update_inode(struct inode
6522         loff_t cur_isize, new_isize;
6523         unsigned long invalid = 0;
6524         unsigned long now = jiffies;
6525 +       uid_t uid;
6526 +       gid_t gid;
6527 +       tag_t tag;
6528
6529         dfprintk(VFS, "NFS: %s(%s/%ld ct=%d info=0x%x)\n",
6530                         __FUNCTION__, inode->i_sb->s_id, inode->i_ino,
6531 @@ -1062,15 +1078,21 @@ static int nfs_update_inode(struct inode
6532         memcpy(&inode->i_atime, &fattr->atime, sizeof(inode->i_atime));
6533         nfsi->change_attr = fattr->change_attr;
6534
6535 +       uid = INOTAG_UID(DX_TAG(inode), fattr->uid, fattr->gid);
6536 +       gid = INOTAG_GID(DX_TAG(inode), fattr->uid, fattr->gid);
6537 +       tag = INOTAG_TAG(DX_TAG(inode), fattr->uid, fattr->gid, 0);
6538 +
6539         if ((inode->i_mode & S_IALLUGO) != (fattr->mode & S_IALLUGO) ||
6540 -           inode->i_uid != fattr->uid ||
6541 -           inode->i_gid != fattr->gid)
6542 +           inode->i_uid != uid ||
6543 +           inode->i_gid != gid ||
6544 +           inode->i_tag != tag)
6545                 invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
6546
6547         inode->i_mode = fattr->mode;
6548         inode->i_nlink = fattr->nlink;
6549 -       inode->i_uid = fattr->uid;
6550 -       inode->i_gid = fattr->gid;
6551 +       inode->i_uid = uid;
6552 +       inode->i_gid = gid;
6553 +       inode->i_tag = tag;
6554
6555         if (fattr->valid & (NFS_ATTR_FATTR_V3 | NFS_ATTR_FATTR_V4)) {
6556                 /*
6557 diff -NurpP --minimal linux-2.6.25.11/fs/nfs/nfs3xdr.c linux-2.6.25.11-vs2.3.0.34.14/fs/nfs/nfs3xdr.c
6558 --- linux-2.6.25.11/fs/nfs/nfs3xdr.c    2008-04-17 12:05:40.000000000 -0400
6559 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/nfs/nfs3xdr.c      2008-04-19 15:14:52.000000000 -0400
6560 @@ -22,6 +22,7 @@
6561  #include <linux/nfs3.h>
6562  #include <linux/nfs_fs.h>
6563  #include <linux/nfsacl.h>
6564 +#include <linux/vs_tag.h>
6565  #include "internal.h"
6566
6567  #define NFSDBG_FACILITY                NFSDBG_XDR
6568 @@ -180,7 +181,7 @@ xdr_decode_fattr(__be32 *p, struct nfs_f
6569  }
6570
6571  static inline __be32 *
6572 -xdr_encode_sattr(__be32 *p, struct iattr *attr)
6573 +xdr_encode_sattr(__be32 *p, struct iattr *attr, int tag)
6574  {
6575         if (attr->ia_valid & ATTR_MODE) {
6576                 *p++ = xdr_one;
6577 @@ -188,15 +189,17 @@ xdr_encode_sattr(__be32 *p, struct iattr
6578         } else {
6579                 *p++ = xdr_zero;
6580         }
6581 -       if (attr->ia_valid & ATTR_UID) {
6582 +       if (attr->ia_valid & ATTR_UID ||
6583 +               (tag && (attr->ia_valid & ATTR_TAG))) {
6584                 *p++ = xdr_one;
6585 -               *p++ = htonl(attr->ia_uid);
6586 +               *p++ = htonl(TAGINO_UID(tag, attr->ia_uid, attr->ia_tag));
6587         } else {
6588                 *p++ = xdr_zero;
6589         }
6590 -       if (attr->ia_valid & ATTR_GID) {
6591 +       if (attr->ia_valid & ATTR_GID ||
6592 +               (tag && (attr->ia_valid & ATTR_TAG))) {
6593                 *p++ = xdr_one;
6594 -               *p++ = htonl(attr->ia_gid);
6595 +               *p++ = htonl(TAGINO_GID(tag, attr->ia_gid, attr->ia_tag));
6596         } else {
6597                 *p++ = xdr_zero;
6598         }
6599 @@ -281,7 +284,8 @@ static int
6600  nfs3_xdr_sattrargs(struct rpc_rqst *req, __be32 *p, struct nfs3_sattrargs *args)
6601  {
6602         p = xdr_encode_fhandle(p, args->fh);
6603 -       p = xdr_encode_sattr(p, args->sattr);
6604 +       p = xdr_encode_sattr(p, args->sattr,
6605 +               req->rq_task->tk_client->cl_tag);
6606         *p++ = htonl(args->guard);
6607         if (args->guard)
6608                 p = xdr_encode_time3(p, &args->guardtime);
6609 @@ -386,7 +390,8 @@ nfs3_xdr_createargs(struct rpc_rqst *req
6610                 *p++ = args->verifier[0];
6611                 *p++ = args->verifier[1];
6612         } else
6613 -               p = xdr_encode_sattr(p, args->sattr);
6614 +               p = xdr_encode_sattr(p, args->sattr,
6615 +                       req->rq_task->tk_client->cl_tag);
6616
6617         req->rq_slen = xdr_adjust_iovec(req->rq_svec, p);
6618         return 0;
6619 @@ -400,7 +405,8 @@ nfs3_xdr_mkdirargs(struct rpc_rqst *req,
6620  {
6621         p = xdr_encode_fhandle(p, args->fh);
6622         p = xdr_encode_array(p, args->name, args->len);
6623 -       p = xdr_encode_sattr(p, args->sattr);
6624 +       p = xdr_encode_sattr(p, args->sattr,
6625 +               req->rq_task->tk_client->cl_tag);
6626         req->rq_slen = xdr_adjust_iovec(req->rq_svec, p);
6627         return 0;
6628  }
6629 @@ -413,7 +419,8 @@ nfs3_xdr_symlinkargs(struct rpc_rqst *re
6630  {
6631         p = xdr_encode_fhandle(p, args->fromfh);
6632         p = xdr_encode_array(p, args->fromname, args->fromlen);
6633 -       p = xdr_encode_sattr(p, args->sattr);
6634 +       p = xdr_encode_sattr(p, args->sattr,
6635 +               req->rq_task->tk_client->cl_tag);
6636         *p++ = htonl(args->pathlen);
6637         req->rq_slen = xdr_adjust_iovec(req->rq_svec, p);
6638
6639 @@ -431,7 +438,8 @@ nfs3_xdr_mknodargs(struct rpc_rqst *req,
6640         p = xdr_encode_fhandle(p, args->fh);
6641         p = xdr_encode_array(p, args->name, args->len);
6642         *p++ = htonl(args->type);
6643 -       p = xdr_encode_sattr(p, args->sattr);
6644 +       p = xdr_encode_sattr(p, args->sattr,
6645 +               req->rq_task->tk_client->cl_tag);
6646         if (args->type == NF3CHR || args->type == NF3BLK) {
6647                 *p++ = htonl(MAJOR(args->rdev));
6648                 *p++ = htonl(MINOR(args->rdev));
6649 diff -NurpP --minimal linux-2.6.25.11/fs/nfs/nfsroot.c linux-2.6.25.11-vs2.3.0.34.14/fs/nfs/nfsroot.c
6650 --- linux-2.6.25.11/fs/nfs/nfsroot.c    2008-04-17 12:05:40.000000000 -0400
6651 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/nfs/nfsroot.c      2008-04-19 15:14:52.000000000 -0400
6652 @@ -119,12 +119,12 @@ static int mount_port __initdata = 0;             /
6653  enum {
6654         /* Options that take integer arguments */
6655         Opt_port, Opt_rsize, Opt_wsize, Opt_timeo, Opt_retrans, Opt_acregmin,
6656 -       Opt_acregmax, Opt_acdirmin, Opt_acdirmax,
6657 +       Opt_acregmax, Opt_acdirmin, Opt_acdirmax, Opt_tagid,
6658         /* Options that take no arguments */
6659         Opt_soft, Opt_hard, Opt_intr,
6660         Opt_nointr, Opt_posix, Opt_noposix, Opt_cto, Opt_nocto, Opt_ac,
6661         Opt_noac, Opt_lock, Opt_nolock, Opt_v2, Opt_v3, Opt_udp, Opt_tcp,
6662 -       Opt_acl, Opt_noacl,
6663 +       Opt_acl, Opt_noacl, Opt_tag, Opt_notag,
6664         /* Error token */
6665         Opt_err
6666  };
6667 @@ -161,6 +161,9 @@ static match_table_t __initdata tokens =
6668         {Opt_tcp, "tcp"},
6669         {Opt_acl, "acl"},
6670         {Opt_noacl, "noacl"},
6671 +       {Opt_tag, "tag"},
6672 +       {Opt_notag, "notag"},
6673 +       {Opt_tagid, "tagid=%u"},
6674         {Opt_err, NULL}
6675
6676  };
6677 @@ -272,6 +275,20 @@ static int __init root_nfs_parse(char *n
6678                         case Opt_noacl:
6679                                 nfs_data.flags |= NFS_MOUNT_NOACL;
6680                                 break;
6681 +#ifndef CONFIG_TAGGING_NONE
6682 +                       case Opt_tag:
6683 +                               nfs_data.flags |= NFS_MOUNT_TAGGED;
6684 +                               break;
6685 +                       case Opt_notag:
6686 +                               nfs_data.flags &= ~NFS_MOUNT_TAGGED;
6687 +                               break;
6688 +#endif
6689 +#ifdef CONFIG_PROPAGATE
6690 +                       case Opt_tagid:
6691 +                               /* use args[0] */
6692 +                               nfs_data.flags |= NFS_MOUNT_TAGGED;
6693 +                               break;
6694 +#endif
6695                         default:
6696                                 printk(KERN_WARNING "Root-NFS: unknown "
6697                                         "option: %s\n", p);
6698 diff -NurpP --minimal linux-2.6.25.11/fs/nfs/super.c linux-2.6.25.11-vs2.3.0.34.14/fs/nfs/super.c
6699 --- linux-2.6.25.11/fs/nfs/super.c      2008-04-17 12:05:40.000000000 -0400
6700 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/nfs/super.c        2008-04-19 15:14:52.000000000 -0400
6701 @@ -50,6 +50,7 @@
6702  #include <linux/nfs_xdr.h>
6703  #include <linux/magic.h>
6704  #include <linux/parser.h>
6705 +#include <linux/vs_tag.h>
6706
6707  #include <asm/system.h>
6708  #include <asm/uaccess.h>
6709 @@ -458,6 +459,7 @@ static void nfs_show_mount_options(struc
6710                 { NFS_MOUNT_NOACL, ",noacl", "" },
6711                 { NFS_MOUNT_NORDIRPLUS, ",nordirplus", "" },
6712                 { NFS_MOUNT_UNSHARED, ",nosharecache", ""},
6713 +               { NFS_MOUNT_TAGGED, ",tag", "" },
6714                 { 0, NULL, NULL }
6715         };
6716         const struct proc_nfs_info *nfs_infop;
6717 diff -NurpP --minimal linux-2.6.25.11/fs/nfsd/auth.c linux-2.6.25.11-vs2.3.0.34.14/fs/nfsd/auth.c
6718 --- linux-2.6.25.11/fs/nfsd/auth.c      2008-04-17 12:05:40.000000000 -0400
6719 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/nfsd/auth.c        2008-04-19 17:07:41.000000000 -0400
6720 @@ -10,6 +10,7 @@
6721  #include <linux/sunrpc/svcauth.h>
6722  #include <linux/nfsd/nfsd.h>
6723  #include <linux/nfsd/export.h>
6724 +#include <linux/vs_tag.h>
6725
6726  int nfsexp_flags(struct svc_rqst *rqstp, struct svc_export *exp)
6727  {
6728 @@ -54,19 +55,23 @@ int nfsd_setuser(struct svc_rqst *rqstp,
6729                 get_group_info(cred.cr_group_info);
6730
6731         if (cred.cr_uid != (uid_t) -1)
6732 -               current->fsuid = cred.cr_uid;
6733 +               current->fsuid = INOTAG_UID(DX_TAG_NFSD, cred.cr_uid, cred.cr_gid);
6734         else
6735                 current->fsuid = exp->ex_anon_uid;
6736         if (cred.cr_gid != (gid_t) -1)
6737 -               current->fsgid = cred.cr_gid;
6738 +               current->fsgid = INOTAG_GID(DX_TAG_NFSD, cred.cr_uid, cred.cr_gid);
6739         else
6740                 current->fsgid = exp->ex_anon_gid;
6741
6742 +       /* this desperately needs a tag :) */
6743 +       current->xid = (xid_t)INOTAG_TAG(DX_TAG_NFSD, cred.cr_uid, cred.cr_gid, 0);
6744 +
6745         if (!cred.cr_group_info)
6746                 return -ENOMEM;
6747         ret = set_current_groups(cred.cr_group_info);
6748         put_group_info(cred.cr_group_info);
6749 -       if ((cred.cr_uid)) {
6750 +
6751 +       if (INOTAG_UID(DX_TAG_NFSD, cred.cr_uid, cred.cr_gid)) {
6752                 current->cap_effective =
6753                         cap_drop_nfsd_set(current->cap_effective);
6754         } else {
6755 diff -NurpP --minimal linux-2.6.25.11/fs/nfsd/nfs3xdr.c linux-2.6.25.11-vs2.3.0.34.14/fs/nfsd/nfs3xdr.c
6756 --- linux-2.6.25.11/fs/nfsd/nfs3xdr.c   2008-04-17 12:05:40.000000000 -0400
6757 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/nfsd/nfs3xdr.c     2008-04-19 17:08:09.000000000 -0400
6758 @@ -21,6 +21,7 @@
6759  #include <linux/sunrpc/svc.h>
6760  #include <linux/nfsd/nfsd.h>
6761  #include <linux/nfsd/xdr3.h>
6762 +#include <linux/vs_tag.h>
6763  #include "auth.h"
6764
6765  #define NFSDDBG_FACILITY               NFSDDBG_XDR
6766 @@ -108,6 +109,8 @@ static __be32 *
6767  decode_sattr3(__be32 *p, struct iattr *iap)
6768  {
6769         u32     tmp;
6770 +       uid_t   uid = 0;
6771 +       gid_t   gid = 0;
6772
6773         iap->ia_valid = 0;
6774
6775 @@ -117,12 +120,15 @@ decode_sattr3(__be32 *p, struct iattr *i
6776         }
6777         if (*p++) {
6778                 iap->ia_valid |= ATTR_UID;
6779 -               iap->ia_uid = ntohl(*p++);
6780 +               uid = ntohl(*p++);
6781         }
6782         if (*p++) {
6783                 iap->ia_valid |= ATTR_GID;
6784 -               iap->ia_gid = ntohl(*p++);
6785 +               gid = ntohl(*p++);
6786         }
6787 +       iap->ia_uid = INOTAG_UID(DX_TAG_NFSD, uid, gid);
6788 +       iap->ia_gid = INOTAG_GID(DX_TAG_NFSD, uid, gid);
6789 +       iap->ia_tag = INOTAG_TAG(DX_TAG_NFSD, uid, gid, 0);
6790         if (*p++) {
6791                 u64     newsize;
6792
6793 @@ -178,8 +184,12 @@ encode_fattr3(struct svc_rqst *rqstp, __
6794         *p++ = htonl(nfs3_ftypes[(stat->mode & S_IFMT) >> 12]);
6795         *p++ = htonl((u32) stat->mode);
6796         *p++ = htonl((u32) stat->nlink);
6797 -       *p++ = htonl((u32) nfsd_ruid(rqstp, stat->uid));
6798 -       *p++ = htonl((u32) nfsd_rgid(rqstp, stat->gid));
6799 +       *p++ = htonl((u32) nfsd_ruid(rqstp,
6800 +               TAGINO_UID(0 /* FIXME: DX_TAG(dentry->d_inode) */,
6801 +               stat->uid, stat->tag)));
6802 +       *p++ = htonl((u32) nfsd_rgid(rqstp,
6803 +               TAGINO_GID(0 /* FIXME: DX_TAG(dentry->d_inode) */,
6804 +               stat->gid, stat->tag)));
6805         if (S_ISLNK(stat->mode) && stat->size > NFS3_MAXPATHLEN) {
6806                 p = xdr_encode_hyper(p, (u64) NFS3_MAXPATHLEN);
6807         } else {
6808 diff -NurpP --minimal linux-2.6.25.11/fs/nfsd/nfs4recover.c linux-2.6.25.11-vs2.3.0.34.14/fs/nfsd/nfs4recover.c
6809 --- linux-2.6.25.11/fs/nfsd/nfs4recover.c       2008-04-17 12:05:40.000000000 -0400
6810 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/nfsd/nfs4recover.c 2008-04-20 13:24:11.000000000 -0400
6811 @@ -154,7 +154,7 @@ nfsd4_create_clid_dir(struct nfs4_client
6812                 dprintk("NFSD: nfsd4_create_clid_dir: DIRECTORY EXISTS\n");
6813                 goto out_put;
6814         }
6815 -       status = vfs_mkdir(rec_dir.path.dentry->d_inode, dentry, S_IRWXU);
6816 +       status = vfs_mkdir(rec_dir.path.dentry->d_inode, dentry, S_IRWXU, NULL);
6817  out_put:
6818         dput(dentry);
6819  out_unlock:
6820 @@ -258,7 +258,7 @@ nfsd4_remove_clid_file(struct dentry *di
6821                 return -EINVAL;
6822         }
6823         mutex_lock_nested(&dir->d_inode->i_mutex, I_MUTEX_PARENT);
6824 -       status = vfs_unlink(dir->d_inode, dentry);
6825 +       status = vfs_unlink(dir->d_inode, dentry, NULL);
6826         mutex_unlock(&dir->d_inode->i_mutex);
6827         return status;
6828  }
6829 @@ -273,7 +273,7 @@ nfsd4_clear_clid_dir(struct dentry *dir,
6830          * a kernel from the future.... */
6831         nfsd4_list_rec_dir(dentry, nfsd4_remove_clid_file);
6832         mutex_lock_nested(&dir->d_inode->i_mutex, I_MUTEX_PARENT);
6833 -       status = vfs_rmdir(dir->d_inode, dentry);
6834 +       status = vfs_rmdir(dir->d_inode, dentry, NULL);
6835         mutex_unlock(&dir->d_inode->i_mutex);
6836         return status;
6837  }
6838 diff -NurpP --minimal linux-2.6.25.11/fs/nfsd/nfs4xdr.c linux-2.6.25.11-vs2.3.0.34.14/fs/nfsd/nfs4xdr.c
6839 --- linux-2.6.25.11/fs/nfsd/nfs4xdr.c   2008-04-17 12:05:40.000000000 -0400
6840 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/nfsd/nfs4xdr.c     2008-04-19 15:14:52.000000000 -0400
6841 @@ -58,6 +58,7 @@
6842  #include <linux/nfs4_acl.h>
6843  #include <linux/sunrpc/gss_api.h>
6844  #include <linux/sunrpc/svcauth_gss.h>
6845 +#include <linux/vs_tag.h>
6846
6847  #define NFSDDBG_FACILITY               NFSDDBG_XDR
6848
6849 @@ -1759,14 +1760,18 @@ out_acl:
6850                 WRITE32(stat.nlink);
6851         }
6852         if (bmval1 & FATTR4_WORD1_OWNER) {
6853 -               status = nfsd4_encode_user(rqstp, stat.uid, &p, &buflen);
6854 +               status = nfsd4_encode_user(rqstp,
6855 +                       TAGINO_UID(DX_TAG(dentry->d_inode),
6856 +                       stat.uid, stat.tag), &p, &buflen);
6857                 if (status == nfserr_resource)
6858                         goto out_resource;
6859                 if (status)
6860                         goto out;
6861         }
6862         if (bmval1 & FATTR4_WORD1_OWNER_GROUP) {
6863 -               status = nfsd4_encode_group(rqstp, stat.gid, &p, &buflen);
6864 +               status = nfsd4_encode_group(rqstp,
6865 +                       TAGINO_GID(DX_TAG(dentry->d_inode),
6866 +                       stat.gid, stat.tag), &p, &buflen);
6867                 if (status == nfserr_resource)
6868                         goto out_resource;
6869                 if (status)
6870 diff -NurpP --minimal linux-2.6.25.11/fs/nfsd/nfsxdr.c linux-2.6.25.11-vs2.3.0.34.14/fs/nfsd/nfsxdr.c
6871 --- linux-2.6.25.11/fs/nfsd/nfsxdr.c    2008-04-17 12:05:40.000000000 -0400
6872 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/nfsd/nfsxdr.c      2008-04-20 13:23:36.000000000 -0400
6873 @@ -15,6 +15,7 @@
6874  #include <linux/nfsd/nfsd.h>
6875  #include <linux/nfsd/xdr.h>
6876  #include <linux/mm.h>
6877 +#include <linux/vs_tag.h>
6878  #include "auth.h"
6879
6880  #define NFSDDBG_FACILITY               NFSDDBG_XDR
6881 @@ -98,6 +99,8 @@ static __be32 *
6882  decode_sattr(__be32 *p, struct iattr *iap)
6883  {
6884         u32     tmp, tmp1;
6885 +       uid_t   uid = 0;
6886 +       gid_t   gid = 0;
6887
6888         iap->ia_valid = 0;
6889
6890 @@ -111,12 +114,15 @@ decode_sattr(__be32 *p, struct iattr *ia
6891         }
6892         if ((tmp = ntohl(*p++)) != (u32)-1) {
6893                 iap->ia_valid |= ATTR_UID;
6894 -               iap->ia_uid = tmp;
6895 +               uid = tmp;
6896         }
6897         if ((tmp = ntohl(*p++)) != (u32)-1) {
6898                 iap->ia_valid |= ATTR_GID;
6899 -               iap->ia_gid = tmp;
6900 +               gid = tmp;
6901         }
6902 +       iap->ia_uid = INOTAG_UID(DX_TAG_NFSD, uid, gid);
6903 +       iap->ia_gid = INOTAG_GID(DX_TAG_NFSD, uid, gid);
6904 +       iap->ia_tag = INOTAG_TAG(DX_TAG_NFSD, uid, gid, 0);
6905         if ((tmp = ntohl(*p++)) != (u32)-1) {
6906                 iap->ia_valid |= ATTR_SIZE;
6907                 iap->ia_size = tmp;
6908 @@ -161,8 +167,10 @@ encode_fattr(struct svc_rqst *rqstp, __b
6909         *p++ = htonl(nfs_ftypes[type >> 12]);
6910         *p++ = htonl((u32) stat->mode);
6911         *p++ = htonl((u32) stat->nlink);
6912 -       *p++ = htonl((u32) nfsd_ruid(rqstp, stat->uid));
6913 -       *p++ = htonl((u32) nfsd_rgid(rqstp, stat->gid));
6914 +       *p++ = htonl((u32) nfsd_ruid(rqstp,
6915 +               TAGINO_UID(DX_TAG(dentry->d_inode), stat->uid, stat->tag)));
6916 +       *p++ = htonl((u32) nfsd_rgid(rqstp,
6917 +               TAGINO_GID(DX_TAG(dentry->d_inode), stat->gid, stat->tag)));
6918
6919         if (S_ISLNK(type) && stat->size > NFS_MAXPATHLEN) {
6920                 *p++ = htonl(NFS_MAXPATHLEN);
6921 diff -NurpP --minimal linux-2.6.25.11/fs/nfsd/vfs.c linux-2.6.25.11-vs2.3.0.34.14/fs/nfsd/vfs.c
6922 --- linux-2.6.25.11/fs/nfsd/vfs.c       2008-04-17 12:05:40.000000000 -0400
6923 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/nfsd/vfs.c 2008-04-21 17:24:34.000000000 -0400
6924 @@ -1258,13 +1258,13 @@ nfsd_create(struct svc_rqst *rqstp, stru
6925                 host_err = vfs_create(dirp, dchild, iap->ia_mode, NULL);
6926                 break;
6927         case S_IFDIR:
6928 -               host_err = vfs_mkdir(dirp, dchild, iap->ia_mode);
6929 +               host_err = vfs_mkdir(dirp, dchild, iap->ia_mode, NULL);
6930                 break;
6931         case S_IFCHR:
6932         case S_IFBLK:
6933         case S_IFIFO:
6934         case S_IFSOCK:
6935 -               host_err = vfs_mknod(dirp, dchild, iap->ia_mode, rdev);
6936 +               host_err = vfs_mknod(dirp, dchild, iap->ia_mode, rdev, NULL);
6937                 break;
6938         default:
6939                 printk("nfsd: bad file type %o in nfsd_create\n", type);
6940 @@ -1529,11 +1529,13 @@ nfsd_symlink(struct svc_rqst *rqstp, str
6941                 else {
6942                         strncpy(path_alloced, path, plen);
6943                         path_alloced[plen] = 0;
6944 -                       host_err = vfs_symlink(dentry->d_inode, dnew, path_alloced, mode);
6945 +                       host_err = vfs_symlink(dentry->d_inode, dnew,
6946 +                               path_alloced, mode, NULL);
6947                         kfree(path_alloced);
6948                 }
6949         } else
6950 -               host_err = vfs_symlink(dentry->d_inode, dnew, path, mode);
6951 +               host_err = vfs_symlink(dentry->d_inode, dnew,
6952 +                       path, mode, NULL);
6953
6954         if (!host_err) {
6955                 if (EX_ISSYNC(fhp->fh_export))
6956 @@ -1592,7 +1594,7 @@ nfsd_link(struct svc_rqst *rqstp, struct
6957         dold = tfhp->fh_dentry;
6958         dest = dold->d_inode;
6959
6960 -       host_err = vfs_link(dold, dirp, dnew);
6961 +       host_err = vfs_link(dold, dirp, dnew, NULL);
6962         if (!host_err) {
6963                 if (EX_ISSYNC(ffhp->fh_export)) {
6964                         err = nfserrno(nfsd_sync_dir(ddir));
6965 @@ -1757,9 +1759,9 @@ nfsd_unlink(struct svc_rqst *rqstp, stru
6966                         host_err = -EPERM;
6967                 } else
6968  #endif
6969 -               host_err = vfs_unlink(dirp, rdentry);
6970 +               host_err = vfs_unlink(dirp, rdentry, NULL);
6971         } else { /* It's RMDIR */
6972 -               host_err = vfs_rmdir(dirp, rdentry);
6973 +               host_err = vfs_rmdir(dirp, rdentry, NULL);
6974         }
6975
6976         dput(rdentry);
6977 @@ -1876,7 +1878,8 @@ nfsd_permission(struct svc_rqst *rqstp,
6978          */
6979         if (!(acc & MAY_LOCAL_ACCESS))
6980                 if (acc & (MAY_WRITE | MAY_SATTR | MAY_TRUNC)) {
6981 -                       if (exp_rdonly(rqstp, exp) || IS_RDONLY(inode))
6982 +                       if (exp_rdonly(rqstp, exp) || IS_RDONLY(inode)
6983 +                               || MNT_IS_RDONLY(exp->ex_path.mnt))
6984                                 return nfserr_rofs;
6985                         if (/* (acc & MAY_WRITE) && */ IS_IMMUTABLE(inode))
6986                                 return nfserr_perm;
6987 diff -NurpP --minimal linux-2.6.25.11/fs/ocfs2/dlm/dlmfs.c linux-2.6.25.11-vs2.3.0.34.14/fs/ocfs2/dlm/dlmfs.c
6988 --- linux-2.6.25.11/fs/ocfs2/dlm/dlmfs.c        2008-04-17 12:05:40.000000000 -0400
6989 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ocfs2/dlm/dlmfs.c  2008-04-19 15:14:52.000000000 -0400
6990 @@ -43,6 +43,7 @@
6991  #include <linux/init.h>
6992  #include <linux/string.h>
6993  #include <linux/backing-dev.h>
6994 +#include <linux/vs_tag.h>
6995
6996  #include <asm/uaccess.h>
6997
6998 @@ -342,6 +343,7 @@ static struct inode *dlmfs_get_root_inod
6999                 inode->i_mode = mode;
7000                 inode->i_uid = current->fsuid;
7001                 inode->i_gid = current->fsgid;
7002 +               inode->i_tag = dx_current_fstag(sb);
7003                 inode->i_blocks = 0;
7004                 inode->i_mapping->backing_dev_info = &dlmfs_backing_dev_info;
7005                 inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
7006 @@ -368,6 +370,7 @@ static struct inode *dlmfs_get_inode(str
7007         inode->i_mode = mode;
7008         inode->i_uid = current->fsuid;
7009         inode->i_gid = current->fsgid;
7010 +       inode->i_tag = dx_current_fstag(sb);
7011         inode->i_blocks = 0;
7012         inode->i_mapping->backing_dev_info = &dlmfs_backing_dev_info;
7013         inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
7014 diff -NurpP --minimal linux-2.6.25.11/fs/ocfs2/dlmglue.c linux-2.6.25.11-vs2.3.0.34.14/fs/ocfs2/dlmglue.c
7015 --- linux-2.6.25.11/fs/ocfs2/dlmglue.c  2008-04-17 12:05:40.000000000 -0400
7016 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ocfs2/dlmglue.c    2008-04-19 15:14:52.000000000 -0400
7017 @@ -1665,6 +1665,7 @@ static void __ocfs2_stuff_meta_lvb(struc
7018         lvb->lvb_iclusters = cpu_to_be32(oi->ip_clusters);
7019         lvb->lvb_iuid      = cpu_to_be32(inode->i_uid);
7020         lvb->lvb_igid      = cpu_to_be32(inode->i_gid);
7021 +       lvb->lvb_itag      = cpu_to_be16(inode->i_tag);
7022         lvb->lvb_imode     = cpu_to_be16(inode->i_mode);
7023         lvb->lvb_inlink    = cpu_to_be16(inode->i_nlink);
7024         lvb->lvb_iatime_packed  =
7025 @@ -1719,6 +1720,7 @@ static void ocfs2_refresh_inode_from_lvb
7026
7027         inode->i_uid     = be32_to_cpu(lvb->lvb_iuid);
7028         inode->i_gid     = be32_to_cpu(lvb->lvb_igid);
7029 +       inode->i_tag     = be16_to_cpu(lvb->lvb_itag);
7030         inode->i_mode    = be16_to_cpu(lvb->lvb_imode);
7031         inode->i_nlink   = be16_to_cpu(lvb->lvb_inlink);
7032         ocfs2_unpack_timespec(&inode->i_atime,
7033 diff -NurpP --minimal linux-2.6.25.11/fs/ocfs2/dlmglue.h linux-2.6.25.11-vs2.3.0.34.14/fs/ocfs2/dlmglue.h
7034 --- linux-2.6.25.11/fs/ocfs2/dlmglue.h  2008-04-17 12:05:40.000000000 -0400
7035 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ocfs2/dlmglue.h    2008-04-19 15:14:52.000000000 -0400
7036 @@ -46,7 +46,8 @@ struct ocfs2_meta_lvb {
7037         __be16       lvb_inlink;
7038         __be32       lvb_iattr;
7039         __be32       lvb_igeneration;
7040 -       __be32       lvb_reserved2;
7041 +       __be16       lvb_itag;
7042 +       __be16       lvb_reserved2;
7043  };
7044
7045  /* ocfs2_inode_lock_full() 'arg_flags' flags */
7046 diff -NurpP --minimal linux-2.6.25.11/fs/ocfs2/file.c linux-2.6.25.11-vs2.3.0.34.14/fs/ocfs2/file.c
7047 --- linux-2.6.25.11/fs/ocfs2/file.c     2008-04-17 12:05:40.000000000 -0400
7048 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ocfs2/file.c       2008-04-19 15:14:52.000000000 -0400
7049 @@ -1054,13 +1054,15 @@ int ocfs2_setattr(struct dentry *dentry,
7050                 mlog(0, "uid change: %d\n", attr->ia_uid);
7051         if (attr->ia_valid & ATTR_GID)
7052                 mlog(0, "gid change: %d\n", attr->ia_gid);
7053 +       if (attr->ia_valid & ATTR_TAG)
7054 +               mlog(0, "tag change: %d\n", attr->ia_tag);
7055         if (attr->ia_valid & ATTR_SIZE)
7056                 mlog(0, "size change...\n");
7057         if (attr->ia_valid & (ATTR_ATIME | ATTR_MTIME | ATTR_CTIME))
7058                 mlog(0, "time change...\n");
7059
7060  #define OCFS2_VALID_ATTRS (ATTR_ATIME | ATTR_MTIME | ATTR_CTIME | ATTR_SIZE \
7061 -                          | ATTR_GID | ATTR_UID | ATTR_MODE)
7062 +                          | ATTR_GID | ATTR_UID | ATTR_TAG | ATTR_MODE)
7063         if (!(attr->ia_valid & OCFS2_VALID_ATTRS)) {
7064                 mlog(0, "can't handle attrs: 0x%x\n", attr->ia_valid);
7065                 return 0;
7066 @@ -2229,6 +2231,7 @@ const struct inode_operations ocfs2_file
7067  const struct inode_operations ocfs2_special_file_iops = {
7068         .setattr        = ocfs2_setattr,
7069         .getattr        = ocfs2_getattr,
7070 +       .sync_flags     = ocfs2_sync_flags,
7071         .permission     = ocfs2_permission,
7072  };
7073
7074 diff -NurpP --minimal linux-2.6.25.11/fs/ocfs2/inode.c linux-2.6.25.11-vs2.3.0.34.14/fs/ocfs2/inode.c
7075 --- linux-2.6.25.11/fs/ocfs2/inode.c    2008-04-17 12:05:40.000000000 -0400
7076 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ocfs2/inode.c      2008-07-11 18:24:40.000000000 -0400
7077 @@ -28,6 +28,7 @@
7078  #include <linux/slab.h>
7079  #include <linux/highmem.h>
7080  #include <linux/pagemap.h>
7081 +#include <linux/vs_tag.h>
7082
7083  #include <asm/byteorder.h>
7084
7085 @@ -42,6 +43,7 @@
7086  #include "file.h"
7087  #include "heartbeat.h"
7088  #include "inode.h"
7089 +#include "ioctl.h"
7090  #include "journal.h"
7091  #include "namei.h"
7092  #include "suballoc.h"
7093 @@ -74,11 +76,13 @@ void ocfs2_set_inode_flags(struct inode
7094  {
7095         unsigned int flags = OCFS2_I(inode)->ip_attr;
7096
7097 -       inode->i_flags &= ~(S_IMMUTABLE |
7098 +       inode->i_flags &= ~(S_IMMUTABLE | S_IXUNLINK |
7099                 S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
7100
7101         if (flags & OCFS2_IMMUTABLE_FL)
7102                 inode->i_flags |= S_IMMUTABLE;
7103 +       if (flags & OCFS2_IXUNLINK_FL)
7104 +               inode->i_flags |= S_IXUNLINK;
7105
7106         if (flags & OCFS2_SYNC_FL)
7107                 inode->i_flags |= S_SYNC;
7108 @@ -88,25 +92,89 @@ void ocfs2_set_inode_flags(struct inode
7109                 inode->i_flags |= S_NOATIME;
7110         if (flags & OCFS2_DIRSYNC_FL)
7111                 inode->i_flags |= S_DIRSYNC;
7112 +
7113 +       inode->i_vflags &= ~(V_BARRIER | V_COW);
7114 +
7115 +       if (flags & OCFS2_BARRIER_FL)
7116 +               inode->i_vflags |= V_BARRIER;
7117 +       if (flags & OCFS2_COW_FL)
7118 +               inode->i_vflags |= V_COW;
7119  }
7120
7121  /* Propagate flags from i_flags to OCFS2_I(inode)->ip_attr */
7122  void ocfs2_get_inode_flags(struct ocfs2_inode_info *oi)
7123  {
7124         unsigned int flags = oi->vfs_inode.i_flags;
7125 +       unsigned int vflags = oi->vfs_inode.i_vflags;
7126 +
7127 +       oi->ip_attr &= ~(OCFS2_SYNC_FL | OCFS2_APPEND_FL |
7128 +                       OCFS2_IMMUTABLE_FL | OCFS2_IXUNLINK_FL |
7129 +                       OCFS2_NOATIME_FL | OCFS2_DIRSYNC_FL |
7130 +                       OCFS2_BARRIER_FL | OCFS2_COW_FL);
7131 +
7132 +       if (flags & S_IMMUTABLE)
7133 +               oi->ip_attr |= OCFS2_IMMUTABLE_FL;
7134 +       if (flags & S_IXUNLINK)
7135 +               oi->ip_attr |= OCFS2_IXUNLINK_FL;
7136
7137 -       oi->ip_attr &= ~(OCFS2_SYNC_FL|OCFS2_APPEND_FL|
7138 -                       OCFS2_IMMUTABLE_FL|OCFS2_NOATIME_FL|OCFS2_DIRSYNC_FL);
7139         if (flags & S_SYNC)
7140                 oi->ip_attr |= OCFS2_SYNC_FL;
7141         if (flags & S_APPEND)
7142                 oi->ip_attr |= OCFS2_APPEND_FL;
7143 -       if (flags & S_IMMUTABLE)
7144 -               oi->ip_attr |= OCFS2_IMMUTABLE_FL;
7145         if (flags & S_NOATIME)
7146                 oi->ip_attr |= OCFS2_NOATIME_FL;
7147         if (flags & S_DIRSYNC)
7148                 oi->ip_attr |= OCFS2_DIRSYNC_FL;
7149 +
7150 +       if (vflags & V_BARRIER)
7151 +               oi->ip_attr |= OCFS2_BARRIER_FL;
7152 +       if (vflags & V_COW)
7153 +               oi->ip_attr |= OCFS2_COW_FL;
7154 +}
7155 +
7156 +int ocfs2_sync_flags(struct inode *inode)
7157 +{
7158 +       struct ocfs2_inode_info *ocfs2_inode = OCFS2_I(inode);
7159 +       struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
7160 +       handle_t *handle = NULL;
7161 +       struct buffer_head *bh = NULL;
7162 +       int status;
7163 +
7164 +       mutex_lock(&inode->i_mutex);
7165 +
7166 +       status = ocfs2_inode_lock(inode, &bh, 1);
7167 +       if (status < 0) {
7168 +               mlog_errno(status);
7169 +               goto bail;
7170 +       }
7171 +
7172 +       status = -EROFS;
7173 +       if (IS_RDONLY(inode))
7174 +               goto bail_unlock;
7175 +
7176 +       handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS);
7177 +       if (IS_ERR(handle)) {
7178 +               status = PTR_ERR(handle);
7179 +               mlog_errno(status);
7180 +               goto bail_unlock;
7181 +       }
7182 +
7183 +       ocfs2_set_inode_flags(inode);
7184 +       status = ocfs2_mark_inode_dirty(handle, inode, bh);
7185 +       if (status < 0)
7186 +               mlog_errno(status);
7187 +
7188 +       ocfs2_commit_trans(osb, handle);
7189 +bail_unlock:
7190 +       ocfs2_inode_unlock(inode, 1);
7191 +bail:
7192 +       mutex_unlock(&inode->i_mutex);
7193 +
7194 +       if (bh)
7195 +               brelse(bh);
7196 +
7197 +       mlog_exit(status);
7198 +       return status;
7199  }
7200
7201  struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 blkno, unsigned flags,
7202 @@ -219,6 +287,8 @@ int ocfs2_populate_inode(struct inode *i
7203         struct super_block *sb;
7204         struct ocfs2_super *osb;
7205         int status = -EINVAL;
7206 +       uid_t uid;
7207 +       gid_t gid;
7208
7209         mlog_entry("(0x%p, size:%llu)\n", inode,
7210                    (unsigned long long)le64_to_cpu(fe->i_size));
7211 @@ -254,8 +324,12 @@ int ocfs2_populate_inode(struct inode *i
7212         inode->i_generation = le32_to_cpu(fe->i_generation);
7213         inode->i_rdev = huge_decode_dev(le64_to_cpu(fe->id1.dev1.i_rdev));
7214         inode->i_mode = le16_to_cpu(fe->i_mode);
7215 -       inode->i_uid = le32_to_cpu(fe->i_uid);
7216 -       inode->i_gid = le32_to_cpu(fe->i_gid);
7217 +       uid = le32_to_cpu(fe->i_uid);
7218 +       gid = le32_to_cpu(fe->i_gid);
7219 +       inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
7220 +       inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
7221 +       inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid,
7222 +               /* le16_to_cpu(raw_inode->i_raw_tag)i */ 0);
7223
7224         /* Fast symlinks will have i_size but no allocated clusters. */
7225         if (S_ISLNK(inode->i_mode) && !fe->i_clusters)
7226 @@ -1230,8 +1304,11 @@ int ocfs2_mark_inode_dirty(handle_t *han
7227
7228         fe->i_size = cpu_to_le64(i_size_read(inode));
7229         fe->i_links_count = cpu_to_le16(inode->i_nlink);
7230 -       fe->i_uid = cpu_to_le32(inode->i_uid);
7231 -       fe->i_gid = cpu_to_le32(inode->i_gid);
7232 +       fe->i_uid = cpu_to_le32(TAGINO_UID(DX_TAG(inode),
7233 +               inode->i_uid, inode->i_tag));
7234 +       fe->i_gid = cpu_to_le32(TAGINO_GID(DX_TAG(inode),
7235 +               inode->i_gid, inode->i_tag));
7236 +       /* i_tag = = cpu_to_le16(inode->i_tag); */
7237         fe->i_mode = cpu_to_le16(inode->i_mode);
7238         fe->i_atime = cpu_to_le64(inode->i_atime.tv_sec);
7239         fe->i_atime_nsec = cpu_to_le32(inode->i_atime.tv_nsec);
7240 @@ -1259,16 +1336,25 @@ leave:
7241  void ocfs2_refresh_inode(struct inode *inode,
7242                          struct ocfs2_dinode *fe)
7243  {
7244 +       uid_t uid;
7245 +       gid_t gid;
7246 +
7247         spin_lock(&OCFS2_I(inode)->ip_lock);
7248
7249         OCFS2_I(inode)->ip_clusters = le32_to_cpu(fe->i_clusters);
7250         OCFS2_I(inode)->ip_attr = le32_to_cpu(fe->i_attr);
7251 +       /* OCFS2_I(inode)->ip_flags &= ~OCFS2_FL_MASK;
7252 +          OCFS2_I(inode)->ip_flags |= le32_to_cpu(fe->i_flags) & OCFS2_FL_MASK; */
7253         OCFS2_I(inode)->ip_dyn_features = le16_to_cpu(fe->i_dyn_features);
7254         ocfs2_set_inode_flags(inode);
7255         i_size_write(inode, le64_to_cpu(fe->i_size));
7256         inode->i_nlink = le16_to_cpu(fe->i_links_count);
7257 -       inode->i_uid = le32_to_cpu(fe->i_uid);
7258 -       inode->i_gid = le32_to_cpu(fe->i_gid);
7259 +       uid = le32_to_cpu(fe->i_uid);
7260 +       gid = le32_to_cpu(fe->i_gid);
7261 +       inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
7262 +       inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
7263 +       inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid,
7264 +               /* le16_to_cpu(raw_inode->i_raw_tag)i */ 0);
7265         inode->i_mode = le16_to_cpu(fe->i_mode);
7266         if (S_ISLNK(inode->i_mode) && le32_to_cpu(fe->i_clusters) == 0)
7267                 inode->i_blocks = 0;
7268 diff -NurpP --minimal linux-2.6.25.11/fs/ocfs2/inode.h linux-2.6.25.11-vs2.3.0.34.14/fs/ocfs2/inode.h
7269 --- linux-2.6.25.11/fs/ocfs2/inode.h    2008-04-17 12:05:40.000000000 -0400
7270 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ocfs2/inode.h      2008-04-19 15:14:52.000000000 -0400
7271 @@ -143,6 +143,7 @@ int ocfs2_aio_write(struct file *file, s
7272
7273  void ocfs2_set_inode_flags(struct inode *inode);
7274  void ocfs2_get_inode_flags(struct ocfs2_inode_info *oi);
7275 +int ocfs2_sync_flags(struct inode *inode);
7276
7277  static inline blkcnt_t ocfs2_inode_sector_count(struct inode *inode)
7278  {
7279 diff -NurpP --minimal linux-2.6.25.11/fs/ocfs2/ioctl.c linux-2.6.25.11-vs2.3.0.34.14/fs/ocfs2/ioctl.c
7280 --- linux-2.6.25.11/fs/ocfs2/ioctl.c    2008-04-17 12:05:40.000000000 -0400
7281 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ocfs2/ioctl.c      2008-07-11 02:51:27.000000000 -0400
7282 @@ -41,7 +41,7 @@ static int ocfs2_get_inode_attr(struct i
7283         return status;
7284  }
7285
7286 -static int ocfs2_set_inode_attr(struct inode *inode, unsigned flags,
7287 +int ocfs2_set_inode_attr(struct inode *inode, unsigned flags,
7288                                 unsigned mask)
7289  {
7290         struct ocfs2_inode_info *ocfs2_inode = OCFS2_I(inode);
7291 @@ -70,6 +70,11 @@ static int ocfs2_set_inode_attr(struct i
7292         if (!S_ISDIR(inode->i_mode))
7293                 flags &= ~OCFS2_DIRSYNC_FL;
7294
7295 +       if (IS_BARRIER(inode)) {
7296 +               vxwprintk_task(1, "messing with the barrier.");
7297 +               goto bail_unlock;
7298 +       }
7299 +
7300         handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS);
7301         if (IS_ERR(handle)) {
7302                 status = PTR_ERR(handle);
7303 diff -NurpP --minimal linux-2.6.25.11/fs/ocfs2/ioctl.h linux-2.6.25.11-vs2.3.0.34.14/fs/ocfs2/ioctl.h
7304 --- linux-2.6.25.11/fs/ocfs2/ioctl.h    2008-04-17 10:33:02.000000000 -0400
7305 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ocfs2/ioctl.h      2008-04-19 15:14:52.000000000 -0400
7306 @@ -10,6 +10,9 @@
7307  #ifndef OCFS2_IOCTL_H
7308  #define OCFS2_IOCTL_H
7309
7310 +int ocfs2_set_inode_attr(struct inode *inode, unsigned flags,
7311 +                               unsigned mask);
7312 +
7313  int ocfs2_ioctl(struct inode * inode, struct file * filp,
7314         unsigned int cmd, unsigned long arg);
7315  long ocfs2_compat_ioctl(struct file *file, unsigned cmd, unsigned long arg);
7316 diff -NurpP --minimal linux-2.6.25.11/fs/ocfs2/namei.c linux-2.6.25.11-vs2.3.0.34.14/fs/ocfs2/namei.c
7317 --- linux-2.6.25.11/fs/ocfs2/namei.c    2008-04-17 12:05:40.000000000 -0400
7318 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ocfs2/namei.c      2008-04-19 15:14:52.000000000 -0400
7319 @@ -40,6 +40,7 @@
7320  #include <linux/types.h>
7321  #include <linux/slab.h>
7322  #include <linux/highmem.h>
7323 +#include <linux/vs_tag.h>
7324
7325  #define MLOG_MASK_PREFIX ML_NAMEI
7326  #include <cluster/masklog.h>
7327 @@ -366,6 +367,9 @@ static int ocfs2_mknod_locked(struct ocf
7328         u64 fe_blkno = 0;
7329         u16 suballoc_bit;
7330         struct inode *inode = NULL;
7331 +       uid_t uid;
7332 +       gid_t gid;
7333 +       tag_t tag;
7334
7335         mlog_entry("(0x%p, 0x%p, %d, %lu, '%.*s')\n", dir, dentry, mode,
7336                    (unsigned long)dev, dentry->d_name.len,
7337 @@ -425,13 +429,19 @@ static int ocfs2_mknod_locked(struct ocf
7338         fe->i_blkno = cpu_to_le64(fe_blkno);
7339         fe->i_suballoc_bit = cpu_to_le16(suballoc_bit);
7340         fe->i_suballoc_slot = cpu_to_le16(osb->slot_num);
7341 -       fe->i_uid = cpu_to_le32(current->fsuid);
7342 +
7343 +       tag = dx_current_fstag(osb->sb);
7344 +       uid = current->fsuid;
7345         if (dir->i_mode & S_ISGID) {
7346 -               fe->i_gid = cpu_to_le32(dir->i_gid);
7347 +               gid = dir->i_gid;
7348                 if (S_ISDIR(mode))
7349                         mode |= S_ISGID;
7350         } else
7351 -               fe->i_gid = cpu_to_le32(current->fsgid);
7352 +               gid = current->fsgid;
7353 +
7354 +       fe->i_uid = cpu_to_le32(TAGINO_UID(DX_TAG(inode), uid, tag));
7355 +       fe->i_gid = cpu_to_le32(TAGINO_GID(DX_TAG(inode), gid, tag));
7356 +       inode->i_tag = tag;
7357         fe->i_mode = cpu_to_le16(mode);
7358         if (S_ISCHR(mode) || S_ISBLK(mode))
7359                 fe->id1.dev1.i_rdev = cpu_to_le64(huge_encode_dev(dev));
7360 @@ -1917,5 +1927,6 @@ const struct inode_operations ocfs2_dir_
7361         .rename         = ocfs2_rename,
7362         .setattr        = ocfs2_setattr,
7363         .getattr        = ocfs2_getattr,
7364 +       .sync_flags     = ocfs2_sync_flags,
7365         .permission     = ocfs2_permission,
7366  };
7367 diff -NurpP --minimal linux-2.6.25.11/fs/ocfs2/ocfs2_fs.h linux-2.6.25.11-vs2.3.0.34.14/fs/ocfs2/ocfs2_fs.h
7368 --- linux-2.6.25.11/fs/ocfs2/ocfs2_fs.h 2008-04-17 12:05:40.000000000 -0400
7369 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ocfs2/ocfs2_fs.h   2008-07-08 13:51:49.000000000 -0400
7370 @@ -187,9 +187,14 @@
7371  #define OCFS2_NODUMP_FL                (0x00000040)    /* do not dump file */
7372  #define OCFS2_NOATIME_FL       (0x00000080)    /* do not update atime */
7373  #define OCFS2_DIRSYNC_FL       (0x00010000)    /* dirsync behaviour (directories only) */
7374 +#define OCFS2_IXUNLINK_FL      (0x00020000)    /* Immutable invert on unlink */
7375
7376 -#define OCFS2_FL_VISIBLE       (0x000100FF)    /* User visible flags */
7377 -#define OCFS2_FL_MODIFIABLE    (0x000100FF)    /* User modifiable flags */
7378 +#define OCFS2_BARRIER_FL       (0x01000000)    /* Barrier for chroot() */
7379 +#define OCFS2_COW_FL           (0x02000000)    /* Copy on Write marker */
7380 +
7381 +#define OCFS2_FL_VISIBLE       (0x000300FF)    /* User visible flags */
7382 +#define OCFS2_FL_MODIFIABLE    (0x000300FF)    /* User modifiable flags */
7383 +#define OCFS2_FL_MASK          (0x030300FF)
7384
7385  /*
7386   * Extent record flags (e_node.leaf.flags)
7387 diff -NurpP --minimal linux-2.6.25.11/fs/ocfs2/ocfs2.h linux-2.6.25.11-vs2.3.0.34.14/fs/ocfs2/ocfs2.h
7388 --- linux-2.6.25.11/fs/ocfs2/ocfs2.h    2008-04-17 12:05:40.000000000 -0400
7389 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ocfs2/ocfs2.h      2008-04-21 10:15:31.000000000 -0400
7390 @@ -171,6 +171,7 @@ enum ocfs2_mount_options
7391         OCFS2_MOUNT_ERRORS_PANIC = 1 << 3, /* Panic on errors */
7392         OCFS2_MOUNT_DATA_WRITEBACK = 1 << 4, /* No data ordering */
7393         OCFS2_MOUNT_LOCALFLOCKS = 1 << 5, /* No cluster aware user file locks */
7394 +       OCFS2_MOUNT_TAGGED = 1 << 8, /* use tagging */
7395  };
7396
7397  #define OCFS2_OSB_SOFT_RO      0x0001
7398 diff -NurpP --minimal linux-2.6.25.11/fs/ocfs2/super.c linux-2.6.25.11-vs2.3.0.34.14/fs/ocfs2/super.c
7399 --- linux-2.6.25.11/fs/ocfs2/super.c    2008-04-17 12:05:40.000000000 -0400
7400 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/ocfs2/super.c      2008-04-23 08:31:10.000000000 -0400
7401 @@ -154,6 +154,7 @@ enum {
7402         Opt_commit,
7403         Opt_localalloc,
7404         Opt_localflocks,
7405 +       Opt_tag, Opt_notag, Opt_tagid,
7406         Opt_err,
7407  };
7408
7409 @@ -172,6 +173,9 @@ static match_table_t tokens = {
7410         {Opt_commit, "commit=%u"},
7411         {Opt_localalloc, "localalloc=%d"},
7412         {Opt_localflocks, "localflocks"},
7413 +       {Opt_tag, "tag"},
7414 +       {Opt_notag, "notag"},
7415 +       {Opt_tagid, "tagid=%u"},
7416         {Opt_err, NULL}
7417  };
7418
7419 @@ -391,6 +395,13 @@ static int ocfs2_remount(struct super_bl
7420                 goto out;
7421         }
7422
7423 +       if ((parsed_options.mount_opt & OCFS2_MOUNT_TAGGED) &&
7424 +               !(sb->s_flags & MS_TAGGED)) {
7425 +               ret = -EINVAL;
7426 +               mlog(ML_ERROR, "Cannot change tagging on remount\n");
7427 +               goto out;
7428 +       }
7429 +
7430         if ((osb->s_mount_opt & OCFS2_MOUNT_HB_LOCAL) !=
7431             (parsed_options.mount_opt & OCFS2_MOUNT_HB_LOCAL)) {
7432                 ret = -EINVAL;
7433 @@ -691,6 +702,9 @@ static int ocfs2_fill_super(struct super
7434
7435         ocfs2_complete_mount_recovery(osb);
7436
7437 +       if (osb->s_mount_opt & OCFS2_MOUNT_TAGGED)
7438 +               sb->s_flags |= MS_TAGGED;
7439 +
7440         if (ocfs2_mount_local(osb))
7441                 snprintf(nodestr, sizeof(nodestr), "local");
7442         else
7443 @@ -864,6 +878,20 @@ static int ocfs2_parse_options(struct su
7444                         if (!is_remount)
7445                                 mopt->mount_opt |= OCFS2_MOUNT_LOCALFLOCKS;
7446                         break;
7447 +#ifndef CONFIG_TAGGING_NONE
7448 +               case Opt_tag:
7449 +                       mopt->mount_opt |= OCFS2_MOUNT_TAGGED;
7450 +                       break;
7451 +               case Opt_notag:
7452 +                       mopt->mount_opt &= ~OCFS2_MOUNT_TAGGED;
7453 +                       break;
7454 +#endif
7455 +#ifdef CONFIG_PROPAGATE
7456 +               case Opt_tagid:
7457 +                       /* use args[0] */
7458 +                       mopt->mount_opt |= OCFS2_MOUNT_TAGGED;
7459 +                       break;
7460 +#endif
7461                 default:
7462                         mlog(ML_ERROR,
7463                              "Unrecognized mount option \"%s\" "
7464 diff -NurpP --minimal linux-2.6.25.11/fs/open.c linux-2.6.25.11-vs2.3.0.34.14/fs/open.c
7465 --- linux-2.6.25.11/fs/open.c   2008-04-17 12:05:41.000000000 -0400
7466 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/open.c     2008-04-21 13:51:52.000000000 -0400
7467 @@ -27,22 +27,31 @@
7468  #include <linux/rcupdate.h>
7469  #include <linux/audit.h>
7470  #include <linux/falloc.h>
7471 +#include <linux/vs_base.h>
7472 +#include <linux/vs_limit.h>
7473 +#include <linux/vs_dlimit.h>
7474 +#include <linux/vs_tag.h>
7475 +#include <linux/vs_cowbl.h>
7476
7477  int vfs_statfs(struct dentry *dentry, struct kstatfs *buf)
7478  {
7479         int retval = -ENODEV;
7480
7481         if (dentry) {
7482 +               struct super_block *sb = dentry->d_sb;
7483 +
7484                 retval = -ENOSYS;
7485 -               if (dentry->d_sb->s_op->statfs) {
7486 +               if (sb->s_op->statfs) {
7487                         memset(buf, 0, sizeof(*buf));
7488                         retval = security_sb_statfs(dentry);
7489                         if (retval)
7490                                 return retval;
7491 -                       retval = dentry->d_sb->s_op->statfs(dentry, buf);
7492 +                       retval = sb->s_op->statfs(dentry, buf);
7493                         if (retval == 0 && buf->f_frsize == 0)
7494                                 buf->f_frsize = buf->f_bsize;
7495                 }
7496 +               if (!vx_check(0, VS_ADMIN|VS_WATCH))
7497 +                       vx_vsi_statfs(sb, buf);
7498         }
7499         return retval;
7500  }
7501 @@ -249,7 +258,7 @@ static long do_sys_truncate(const char _
7502                 goto dput_and_out;
7503
7504         error = -EROFS;
7505 -       if (IS_RDONLY(inode))
7506 +       if (IS_RDONLY(inode) || MNT_IS_RDONLY(nd.path.mnt))
7507                 goto dput_and_out;
7508
7509         error = -EPERM;
7510 @@ -458,7 +467,7 @@ asmlinkage long sys_faccessat(int dfd, c
7511            special_file(nd.path.dentry->d_inode->i_mode))
7512                 goto out_path_release;
7513
7514 -       if(IS_RDONLY(nd.path.dentry->d_inode))
7515 +       if(IS_RDONLY(nd.path.dentry->d_inode) || MNT_IS_RDONLY(nd.path.mnt))
7516                 res = -EROFS;
7517
7518  out_path_release:
7519 @@ -568,7 +577,7 @@ asmlinkage long sys_fchmod(unsigned int
7520         audit_inode(NULL, dentry);
7521
7522         err = -EROFS;
7523 -       if (IS_RDONLY(inode))
7524 +       if (IS_RDONLY(inode) || MNT_IS_RDONLY(file->f_vfsmnt))
7525                 goto out_putf;
7526         err = -EPERM;
7527         if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
7528 @@ -598,11 +607,11 @@ asmlinkage long sys_fchmodat(int dfd, co
7529         error = __user_walk_fd(dfd, filename, LOOKUP_FOLLOW, &nd);
7530         if (error)
7531                 goto out;
7532 -       inode = nd.path.dentry->d_inode;
7533 -
7534 -       error = -EROFS;
7535 -       if (IS_RDONLY(inode))
7536 +
7537 +       error = cow_check_and_break(&nd);
7538 +       if (error)
7539                 goto dput_and_out;
7540 +       inode = nd.path.dentry->d_inode;
7541
7542         error = -EPERM;
7543         if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
7544 @@ -627,7 +636,8 @@ asmlinkage long sys_chmod(const char __u
7545         return sys_fchmodat(AT_FDCWD, filename, mode);
7546  }
7547
7548 -static int chown_common(struct dentry * dentry, uid_t user, gid_t group)
7549 +static int chown_common(struct dentry *dentry, struct vfsmount *mnt,
7550 +       uid_t user, gid_t group)
7551  {
7552         struct inode * inode;
7553         int error;
7554 @@ -639,7 +649,7 @@ static int chown_common(struct dentry *
7555                 goto out;
7556         }
7557         error = -EROFS;
7558 -       if (IS_RDONLY(inode))
7559 +       if (IS_RDONLY(inode) || MNT_IS_RDONLY(mnt))
7560                 goto out;
7561         error = -EPERM;
7562         if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
7563 @@ -647,11 +657,11 @@ static int chown_common(struct dentry *
7564         newattrs.ia_valid =  ATTR_CTIME;
7565         if (user != (uid_t) -1) {
7566                 newattrs.ia_valid |= ATTR_UID;
7567 -               newattrs.ia_uid = user;
7568 +               newattrs.ia_uid = dx_map_uid(user);
7569         }
7570         if (group != (gid_t) -1) {
7571                 newattrs.ia_valid |= ATTR_GID;
7572 -               newattrs.ia_gid = group;
7573 +               newattrs.ia_gid = dx_map_gid(group);
7574         }
7575         if (!S_ISDIR(inode->i_mode))
7576                 newattrs.ia_valid |=
7577 @@ -671,7 +681,11 @@ asmlinkage long sys_chown(const char __u
7578         error = user_path_walk(filename, &nd);
7579         if (error)
7580                 goto out;
7581 -       error = chown_common(nd.path.dentry, user, group);
7582 +#ifdef CONFIG_VSERVER_COWBL
7583 +       error = cow_check_and_break(&nd);
7584 +       if (!error)
7585 +#endif
7586 +               error = chown_common(nd.path.dentry, nd.path.mnt, user, group);
7587         path_put(&nd.path);
7588  out:
7589         return error;
7590 @@ -691,7 +705,11 @@ asmlinkage long sys_fchownat(int dfd, co
7591         error = __user_walk_fd(dfd, filename, follow, &nd);
7592         if (error)
7593                 goto out;
7594 -       error = chown_common(nd.path.dentry, user, group);
7595 +#ifdef CONFIG_VSERVER_COWBL
7596 +       error = cow_check_and_break(&nd);
7597 +       if (!error)
7598 +#endif
7599 +               error = chown_common(nd.path.dentry, nd.path.mnt, user, group);
7600         path_put(&nd.path);
7601  out:
7602         return error;
7603 @@ -705,7 +723,11 @@ asmlinkage long sys_lchown(const char __
7604         error = user_path_walk_link(filename, &nd);
7605         if (error)
7606                 goto out;
7607 -       error = chown_common(nd.path.dentry, user, group);
7608 +#ifdef CONFIG_VSERVER_COWBL
7609 +       error = cow_check_and_break(&nd);
7610 +       if (!error)
7611 +#endif
7612 +               error = chown_common(nd.path.dentry, nd.path.mnt, user, group);
7613         path_put(&nd.path);
7614  out:
7615         return error;
7616 @@ -724,7 +746,7 @@ asmlinkage long sys_fchown(unsigned int
7617
7618         dentry = file->f_path.dentry;
7619         audit_inode(NULL, dentry);
7620 -       error = chown_common(dentry, user, group);
7621 +       error = chown_common(dentry, file->f_vfsmnt, user, group);
7622         fput(file);
7623  out:
7624         return error;
7625 @@ -971,6 +993,7 @@ repeat:
7626         else
7627                 FD_CLR(fd, fdt->close_on_exec);
7628         files->next_fd = fd + 1;
7629 +       vx_openfd_inc(fd);
7630  #if 1
7631         /* Sanity check */
7632         if (fdt->fd[fd] != NULL) {
7633 @@ -998,6 +1021,7 @@ static void __put_unused_fd(struct files
7634         __FD_CLR(fd, fdt->open_fds);
7635         if (fd < files->next_fd)
7636                 files->next_fd = fd;
7637 +       vx_openfd_dec(fd);
7638  }
7639
7640  void put_unused_fd(unsigned int fd)
7641 diff -NurpP --minimal linux-2.6.25.11/fs/proc/array.c linux-2.6.25.11-vs2.3.0.34.14/fs/proc/array.c
7642 --- linux-2.6.25.11/fs/proc/array.c     2008-07-15 22:05:17.000000000 -0400
7643 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/proc/array.c       2008-06-10 20:00:46.000000000 -0400
7644 @@ -79,6 +79,8 @@
7645  #include <linux/delayacct.h>
7646  #include <linux/seq_file.h>
7647  #include <linux/pid_namespace.h>
7648 +#include <linux/vs_context.h>
7649 +#include <linux/vs_network.h>
7650
7651  #include <asm/pgtable.h>
7652  #include <asm/processor.h>
7653 @@ -140,8 +142,9 @@ static const char *task_state_array[] =
7654         "D (disk sleep)",       /*  2 */
7655         "T (stopped)",          /*  4 */
7656         "T (tracing stop)",     /*  8 */
7657 -       "Z (zombie)",           /* 16 */
7658 -       "X (dead)"              /* 32 */
7659 +       "H (on hold)",          /* 16 */
7660 +       "Z (zombie)",           /* 32 */
7661 +       "X (dead)",             /* 64 */
7662  };
7663
7664  static inline const char *get_task_state(struct task_struct *tsk)
7665 @@ -162,6 +165,7 @@ static inline void task_state(struct seq
7666         struct group_info *group_info;
7667         int g;
7668         struct fdtable *fdt = NULL;
7669 +/* +   pid_t pid, ptgid, tppid, tgid; */
7670         pid_t ppid, tpid;
7671
7672         rcu_read_lock();
7673 @@ -169,6 +173,12 @@ static inline void task_state(struct seq
7674                 task_tgid_nr_ns(rcu_dereference(p->real_parent), ns) : 0;
7675         tpid = pid_alive(p) && p->ptrace ?
7676                 task_pid_nr_ns(rcu_dereference(p->parent), ns) : 0;
7677 +/*     tgid = vx_map_tgid(p->tgid);
7678 +       pid = vx_map_pid(p->pid);
7679 +       ptgid = vx_map_pid(pid_alive(p) ?
7680 +               rcu_dereference(p->real_parent)->tgid : 0);
7681 +       tppid = vx_map_pid(pid_alive(p) && p->ptrace ?
7682 +               rcu_dereference(p->parent)->pid : 0); */
7683         seq_printf(m,
7684                 "State:\t%s\n"
7685                 "Tgid:\t%d\n"
7686 @@ -280,7 +290,7 @@ static inline void task_sig(struct seq_f
7687  }
7688
7689  static void render_cap_t(struct seq_file *m, const char *header,
7690 -                       kernel_cap_t *a)
7691 +                       struct vx_info *vxi, kernel_cap_t *a)
7692  {
7693         unsigned __capi;
7694
7695 @@ -294,9 +304,9 @@ static void render_cap_t(struct seq_file
7696
7697  static inline void task_cap(struct seq_file *m, struct task_struct *p)
7698  {
7699 -       render_cap_t(m, "CapInh:\t", &p->cap_inheritable);
7700 -       render_cap_t(m, "CapPrm:\t", &p->cap_permitted);
7701 -       render_cap_t(m, "CapEff:\t", &p->cap_effective);
7702 +       render_cap_t(m, "CapInh:\t", p->vx_info, &p->cap_inheritable);
7703 +       render_cap_t(m, "CapPrm:\t", p->vx_info, &p->cap_permitted);
7704 +       render_cap_t(m, "CapEff:\t", p->vx_info, &p->cap_effective);
7705  }
7706
7707  static inline void task_context_switch_counts(struct seq_file *m,
7708 @@ -308,6 +318,45 @@ static inline void task_context_switch_c
7709                         p->nivcsw);
7710  }
7711
7712 +int proc_pid_nsproxy(struct seq_file *m, struct pid_namespace *ns,
7713 +                       struct pid *pid, struct task_struct *task)
7714 +{
7715 +       seq_printf(m,   "Proxy:\t%p(%c)\n"
7716 +                       "Count:\t%u\n"
7717 +                       "uts:\t%p(%c)\n"
7718 +                       "ipc:\t%p(%c)\n"
7719 +                       "mnt:\t%p(%c)\n"
7720 +                       "pid:\t%p(%c)\n"
7721 +                       "user:\t%p(%c)\n"
7722 +                       "net:\t%p(%c)\n",
7723 +                       task->nsproxy,
7724 +                       (task->nsproxy == init_task.nsproxy ? 'I' : '-'),
7725 +                       atomic_read(&task->nsproxy->count),
7726 +                       task->nsproxy->uts_ns,
7727 +                       (task->nsproxy->uts_ns == init_task.nsproxy->uts_ns ? 'I' : '-'),
7728 +                       task->nsproxy->ipc_ns,
7729 +                       (task->nsproxy->ipc_ns == init_task.nsproxy->ipc_ns ? 'I' : '-'),
7730 +                       task->nsproxy->mnt_ns,
7731 +                       (task->nsproxy->mnt_ns == init_task.nsproxy->mnt_ns ? 'I' : '-'),
7732 +                       task->nsproxy->pid_ns,
7733 +                       (task->nsproxy->pid_ns == init_task.nsproxy->pid_ns ? 'I' : '-'),
7734 +                       task->nsproxy->user_ns,
7735 +                       (task->nsproxy->user_ns == init_task.nsproxy->user_ns ? 'I' : '-'),
7736 +                       task->nsproxy->net_ns,
7737 +                       (task->nsproxy->net_ns == init_task.nsproxy->net_ns ? 'I' : '-'));
7738 +       return 0;
7739 +}
7740 +
7741 +void task_vs_id(struct seq_file *m, struct task_struct *task)
7742 +{
7743 +       if (task_vx_flags(task, VXF_HIDE_VINFO, 0))
7744 +               return;
7745 +
7746 +       seq_printf(m, "VxID: %d\n", vx_task_xid(task));
7747 +       seq_printf(m, "NxID: %d\n", nx_task_nid(task));
7748 +}
7749 +
7750 +
7751  int proc_pid_status(struct seq_file *m, struct pid_namespace *ns,
7752                         struct pid *pid, struct task_struct *task)
7753  {
7754 @@ -323,6 +372,7 @@ int proc_pid_status(struct seq_file *m,
7755         task_sig(m, task);
7756         task_cap(m, task);
7757         cpuset_task_status_allowed(m, task);
7758 +       task_vs_id(m, task);
7759  #if defined(CONFIG_S390)
7760         task_show_regs(m, task);
7761  #endif
7762 @@ -494,6 +544,17 @@ static int do_task_stat(struct seq_file
7763         /* convert nsec -> ticks */
7764         start_time = nsec_to_clock_t(start_time);
7765
7766 +       /* fixup start time for virt uptime */
7767 +       if (vx_flags(VXF_VIRT_UPTIME, 0)) {
7768 +               unsigned long long bias =
7769 +                       current->vx_info->cvirt.bias_clock;
7770 +
7771 +               if (start_time > bias)
7772 +                       start_time -= bias;
7773 +               else
7774 +                       start_time = 0;
7775 +       }
7776 +
7777         seq_printf(m, "%d (%s) %c %d %d %d %d %d %u %lu \
7778  %lu %lu %lu %lu %lu %ld %ld %ld %ld %d 0 %llu %lu %ld %lu %lu %lu %lu %lu \
7779  %lu %lu %lu %lu %lu %lu %lu %lu %d %d %u %u %llu %lu %ld\n",
7780 diff -NurpP --minimal linux-2.6.25.11/fs/proc/base.c linux-2.6.25.11-vs2.3.0.34.14/fs/proc/base.c
7781 --- linux-2.6.25.11/fs/proc/base.c      2008-07-15 22:05:17.000000000 -0400
7782 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/proc/base.c        2008-06-10 20:00:46.000000000 -0400
7783 @@ -76,6 +76,8 @@
7784  #include <linux/oom.h>
7785  #include <linux/elf.h>
7786  #include <linux/pid_namespace.h>
7787 +#include <linux/vs_context.h>
7788 +#include <linux/vs_network.h>
7789  #include "internal.h"
7790
7791  /* NOTE:
7792 @@ -1309,6 +1311,8 @@ static struct inode *proc_pid_make_inode
7793                 inode->i_uid = task->euid;
7794                 inode->i_gid = task->egid;
7795         }
7796 +       /* procfs is xid tagged */
7797 +       inode->i_tag = (tag_t)vx_task_xid(task);
7798         security_task_to_inode(task, inode);
7799
7800  out:
7801 @@ -1852,6 +1856,13 @@ static struct dentry *proc_pident_lookup
7802         if (!task)
7803                 goto out_no_task;
7804
7805 +       /* TODO: maybe we can come up with a generic approach? */
7806 +       if (task_vx_flags(task, VXF_HIDE_VINFO, 0) &&
7807 +               (dentry->d_name.len == 5) &&
7808 +               (!memcmp(dentry->d_name.name, "vinfo", 5) ||
7809 +               !memcmp(dentry->d_name.name, "ninfo", 5)))
7810 +               goto out;
7811 +
7812         /*
7813          * Yes, it does not scale. And it should not. Don't add
7814          * new entries into /proc/<tgid>/ without very good reasons.
7815 @@ -2239,7 +2250,7 @@ out_iput:
7816  static struct dentry *proc_base_lookup(struct inode *dir, struct dentry *dentry)
7817  {
7818         struct dentry *error;
7819 -       struct task_struct *task = get_proc_task(dir);
7820 +       struct task_struct *task = get_proc_task_real(dir);
7821         const struct pid_entry *p, *last;
7822
7823         error = ERR_PTR(-ENOENT);
7824 @@ -2304,6 +2315,9 @@ static int proc_pid_io_accounting(struct
7825  static const struct file_operations proc_task_operations;
7826  static const struct inode_operations proc_task_inode_operations;
7827
7828 +extern int proc_pid_vx_info(struct task_struct *, char *);
7829 +extern int proc_pid_nx_info(struct task_struct *, char *);
7830 +
7831  static const struct pid_entry tgid_base_stuff[] = {
7832         DIR("task",       S_IRUGO|S_IXUGO, task),
7833         DIR("fd",         S_IRUSR|S_IXUSR, fd),
7834 @@ -2354,6 +2368,8 @@ static const struct pid_entry tgid_base_
7835  #ifdef CONFIG_CGROUPS
7836         REG("cgroup",  S_IRUGO, cgroup),
7837  #endif
7838 +       INF("vinfo",      S_IRUGO, pid_vx_info),
7839 +       INF("ninfo",      S_IRUGO, pid_nx_info),
7840         INF("oom_score",  S_IRUGO, oom_score),
7841         REG("oom_adj",    S_IRUGO|S_IWUSR, oom_adjust),
7842  #ifdef CONFIG_AUDITSYSCALL
7843 @@ -2369,6 +2385,7 @@ static const struct pid_entry tgid_base_
7844  #ifdef CONFIG_TASK_IO_ACCOUNTING
7845         INF("io",       S_IRUGO, pid_io_accounting),
7846  #endif
7847 +       ONE("nsproxy",  S_IRUGO, pid_nsproxy),
7848  };
7849
7850  static int proc_tgid_base_readdir(struct file * filp,
7851 @@ -2565,7 +2582,7 @@ retry:
7852         iter.task = NULL;
7853         pid = find_ge_pid(iter.tgid, ns);
7854         if (pid) {
7855 -               iter.tgid = pid_nr_ns(pid, ns);
7856 +               iter.tgid = pid_unmapped_nr_ns(pid, ns);
7857                 iter.task = pid_task(pid, PIDTYPE_PID);
7858                 /* What we to know is if the pid we have find is the
7859                  * pid of a thread_group_leader.  Testing for task
7860 @@ -2595,7 +2612,7 @@ static int proc_pid_fill_cache(struct fi
7861         struct tgid_iter iter)
7862  {
7863         char name[PROC_NUMBUF];
7864 -       int len = snprintf(name, sizeof(name), "%d", iter.tgid);
7865 +       int len = snprintf(name, sizeof(name), "%d", vx_map_tgid(iter.tgid));
7866         return proc_fill_cache(filp, dirent, filldir, name, len,
7867                                 proc_pid_instantiate, iter.task, NULL);
7868  }
7869 @@ -2604,7 +2621,7 @@ static int proc_pid_fill_cache(struct fi
7870  int proc_pid_readdir(struct file * filp, void * dirent, filldir_t filldir)
7871  {
7872         unsigned int nr = filp->f_pos - FIRST_PROCESS_ENTRY;
7873 -       struct task_struct *reaper = get_proc_task(filp->f_path.dentry->d_inode);
7874 +       struct task_struct *reaper = get_proc_task_real(filp->f_path.dentry->d_inode);
7875         struct tgid_iter iter;
7876         struct pid_namespace *ns;
7877
7878 @@ -2624,6 +2641,8 @@ int proc_pid_readdir(struct file * filp,
7879              iter.task;
7880              iter.tgid += 1, iter = next_tgid(ns, iter)) {
7881                 filp->f_pos = iter.tgid + TGID_OFFSET;
7882 +               if (!vx_proc_task_visible(iter.task))
7883 +                       continue;
7884                 if (proc_pid_fill_cache(filp, dirent, filldir, iter) < 0) {
7885                         put_task_struct(iter.task);
7886                         goto out;
7887 @@ -2759,6 +2778,8 @@ static struct dentry *proc_task_lookup(s
7888         tid = name_to_int(dentry);
7889         if (tid == ~0U)
7890                 goto out;
7891 +       if (vx_current_initpid(tid))
7892 +               goto out;
7893
7894         ns = dentry->d_sb->s_fs_info;
7895         rcu_read_lock();
7896 diff -NurpP --minimal linux-2.6.25.11/fs/proc/generic.c linux-2.6.25.11-vs2.3.0.34.14/fs/proc/generic.c
7897 --- linux-2.6.25.11/fs/proc/generic.c   2008-04-17 12:05:41.000000000 -0400
7898 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/proc/generic.c     2008-04-20 14:23:26.000000000 -0400
7899 @@ -21,6 +21,7 @@
7900  #include <linux/bitops.h>
7901  #include <linux/spinlock.h>
7902  #include <linux/completion.h>
7903 +#include <linux/vserver/inode.h>
7904  #include <asm/uaccess.h>
7905
7906  #include "internal.h"
7907 @@ -389,6 +390,8 @@ struct dentry *proc_lookup_de(struct pro
7908                 for (de = de->subdir; de ; de = de->next) {
7909                         if (de->namelen != dentry->d_name.len)
7910                                 continue;
7911 +                       if (!vx_hide_check(0, de->vx_flags))
7912 +                               continue;
7913                         if (!memcmp(dentry->d_name.name, de->name, de->namelen)) {
7914                                 unsigned int ino;
7915
7916 @@ -397,6 +400,8 @@ struct dentry *proc_lookup_de(struct pro
7917                                 spin_unlock(&proc_subdir_lock);
7918                                 error = -EINVAL;
7919                                 inode = proc_get_inode(dir->i_sb, ino, de);
7920 +                               /* generic proc entries belong to the host */
7921 +                               inode->i_tag = 0;
7922                                 goto out_unlock;
7923                         }
7924                 }
7925 @@ -481,6 +486,8 @@ int proc_readdir_de(struct proc_dir_entr
7926
7927                                 /* filldir passes info to user space */
7928                                 de_get(de);
7929 +                               if (!vx_hide_check(0, de->vx_flags))
7930 +                                       goto skip;
7931                                 spin_unlock(&proc_subdir_lock);
7932                                 if (filldir(dirent, de->name, de->namelen, filp->f_pos,
7933                                             de->low_ino, de->mode >> 12) < 0) {
7934 @@ -488,6 +495,7 @@ int proc_readdir_de(struct proc_dir_entr
7935                                         goto out;
7936                                 }
7937                                 spin_lock(&proc_subdir_lock);
7938 +                       skip:
7939                                 filp->f_pos++;
7940                                 next = de->next;
7941                                 de_put(de);
7942 @@ -602,6 +610,7 @@ static struct proc_dir_entry *__proc_cre
7943         ent->nlink = nlink;
7944         atomic_set(&ent->count, 1);
7945         ent->pde_users = 0;
7946 +       ent->vx_flags = IATTR_PROC_DEFAULT;
7947         spin_lock_init(&ent->pde_unload_lock);
7948         ent->pde_unload_completion = NULL;
7949   out:
7950 @@ -624,7 +633,8 @@ struct proc_dir_entry *proc_symlink(cons
7951                                 kfree(ent->data);
7952                                 kfree(ent);
7953                                 ent = NULL;
7954 -                       }
7955 +                       } else
7956 +                               ent->vx_flags = IATTR_PROC_SYMLINK;
7957                 } else {
7958                         kfree(ent);
7959                         ent = NULL;
7960 diff -NurpP --minimal linux-2.6.25.11/fs/proc/inode.c linux-2.6.25.11-vs2.3.0.34.14/fs/proc/inode.c
7961 --- linux-2.6.25.11/fs/proc/inode.c     2008-04-17 12:05:41.000000000 -0400
7962 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/proc/inode.c       2008-04-19 15:14:52.000000000 -0400
7963 @@ -408,6 +408,8 @@ struct inode *proc_get_inode(struct supe
7964                                 inode->i_uid = de->uid;
7965                                 inode->i_gid = de->gid;
7966                         }
7967 +               if (de->vx_flags)
7968 +                       PROC_I(inode)->vx_flags = de->vx_flags;
7969                         if (de->size)
7970                                 inode->i_size = de->size;
7971                         if (de->nlink)
7972 diff -NurpP --minimal linux-2.6.25.11/fs/proc/internal.h linux-2.6.25.11-vs2.3.0.34.14/fs/proc/internal.h
7973 --- linux-2.6.25.11/fs/proc/internal.h  2008-04-17 12:05:41.000000000 -0400
7974 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/proc/internal.h    2008-05-21 14:08:19.000000000 -0400
7975 @@ -10,6 +10,7 @@
7976   */
7977
7978  #include <linux/proc_fs.h>
7979 +#include <linux/vs_pid.h>
7980
7981  #ifdef CONFIG_PROC_SYSCTL
7982  extern int proc_sys_init(void);
7983 @@ -57,6 +58,9 @@ extern int proc_pid_status(struct seq_fi
7984                                 struct pid *pid, struct task_struct *task);
7985  extern int proc_pid_statm(struct seq_file *m, struct pid_namespace *ns,
7986                                 struct pid *pid, struct task_struct *task);
7987 +extern int proc_pid_nsproxy(struct seq_file *m, struct pid_namespace *ns,
7988 +                               struct pid *pid, struct task_struct *task);
7989 +
7990  extern loff_t mem_lseek(struct file *file, loff_t offset, int orig);
7991
7992  extern const struct file_operations proc_maps_operations;
7993 @@ -76,11 +80,16 @@ static inline struct pid *proc_pid(struc
7994         return PROC_I(inode)->pid;
7995  }
7996
7997 -static inline struct task_struct *get_proc_task(struct inode *inode)
7998 +static inline struct task_struct *get_proc_task_real(struct inode *inode)
7999  {
8000         return get_pid_task(proc_pid(inode), PIDTYPE_PID);
8001  }
8002
8003 +static inline struct task_struct *get_proc_task(struct inode *inode)
8004 +{
8005 +       return vx_get_proc_task(inode, proc_pid(inode));
8006 +}
8007 +
8008  static inline int proc_fd(struct inode *inode)
8009  {
8010         return PROC_I(inode)->fd;
8011 diff -NurpP --minimal linux-2.6.25.11/fs/proc/proc_misc.c linux-2.6.25.11-vs2.3.0.34.14/fs/proc/proc_misc.c
8012 --- linux-2.6.25.11/fs/proc/proc_misc.c 2008-04-17 12:05:41.000000000 -0400
8013 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/proc/proc_misc.c   2008-04-23 11:56:24.000000000 -0400
8014 @@ -56,6 +56,8 @@
8015  #include <asm/div64.h>
8016  #include "internal.h"
8017
8018 +#include <linux/vs_cvirt.h>
8019 +
8020  #define LOAD_INT(x) ((x) >> FSHIFT)
8021  #define LOAD_FRAC(x) LOAD_INT(((x) & (FIXED_1-1)) * 100)
8022  /*
8023 @@ -83,22 +85,38 @@ static int proc_calc_metrics(char *page,
8024  static int loadavg_read_proc(char *page, char **start, off_t off,
8025                                  int count, int *eof, void *data)
8026  {
8027 +       unsigned long running;
8028 +       unsigned int threads;
8029         int a, b, c;
8030         int len;
8031         unsigned long seq;
8032
8033         do {
8034                 seq = read_seqbegin(&xtime_lock);
8035 -               a = avenrun[0] + (FIXED_1/200);
8036 -               b = avenrun[1] + (FIXED_1/200);
8037 -               c = avenrun[2] + (FIXED_1/200);
8038 +               if (vx_flags(VXF_VIRT_LOAD, 0)) {
8039 +                       struct vx_info *vxi = current->vx_info;
8040 +
8041 +                       a = vxi->cvirt.load[0] + (FIXED_1/200);
8042 +                       b = vxi->cvirt.load[1] + (FIXED_1/200);
8043 +                       c = vxi->cvirt.load[2] + (FIXED_1/200);
8044 +
8045 +                       running = atomic_read(&vxi->cvirt.nr_running);
8046 +                       threads = atomic_read(&vxi->cvirt.nr_threads);
8047 +               } else {
8048 +                       a = avenrun[0] + (FIXED_1/200);
8049 +                       b = avenrun[1] + (FIXED_1/200);
8050 +                       c = avenrun[2] + (FIXED_1/200);
8051 +
8052 +                       running = nr_running();
8053 +                       threads = nr_threads;
8054 +               }
8055         } while (read_seqretry(&xtime_lock, seq));
8056
8057         len = sprintf(page,"%d.%02d %d.%02d %d.%02d %ld/%d %d\n",
8058                 LOAD_INT(a), LOAD_FRAC(a),
8059                 LOAD_INT(b), LOAD_FRAC(b),
8060                 LOAD_INT(c), LOAD_FRAC(c),
8061 -               nr_running(), nr_threads,
8062 +               running, threads,
8063                 task_active_pid_ns(current)->last_pid);
8064         return proc_calc_metrics(page, start, off, count, eof, len);
8065  }
8066 @@ -114,6 +132,9 @@ static int uptime_read_proc(char *page,
8067         do_posix_clock_monotonic_gettime(&uptime);
8068         monotonic_to_bootbased(&uptime);
8069         cputime_to_timespec(idletime, &idle);
8070 +       if (vx_flags(VXF_VIRT_UPTIME, 0))
8071 +               vx_vsi_uptime(&uptime, &idle);
8072 +
8073         len = sprintf(page,"%lu.%02lu %lu.%02lu\n",
8074                         (unsigned long) uptime.tv_sec,
8075                         (uptime.tv_nsec / (NSEC_PER_SEC / 100)),
8076 @@ -145,7 +166,7 @@ static int meminfo_read_proc(char *page,
8077
8078         cached = global_page_state(NR_FILE_PAGES) -
8079                         total_swapcache_pages - i.bufferram;
8080 -       if (cached < 0)
8081 +       if (cached < 0 || vx_flags(VXF_VIRT_MEM, 0))
8082                 cached = 0;
8083
8084         get_vmalloc_info(&vmi);
8085 diff -NurpP --minimal linux-2.6.25.11/fs/proc/root.c linux-2.6.25.11-vs2.3.0.34.14/fs/proc/root.c
8086 --- linux-2.6.25.11/fs/proc/root.c      2008-04-17 12:05:41.000000000 -0400
8087 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/proc/root.c        2008-04-19 15:14:52.000000000 -0400
8088 @@ -23,6 +23,9 @@
8089  #include "internal.h"
8090
8091  struct proc_dir_entry *proc_bus, *proc_root_fs, *proc_root_driver;
8092 +struct proc_dir_entry *proc_virtual;
8093 +
8094 +extern void proc_vx_init(void);
8095
8096  static int proc_test_super(struct super_block *sb, void *data)
8097  {
8098 @@ -138,6 +141,7 @@ void __init proc_root_init(void)
8099         proc_device_tree_init();
8100  #endif
8101         proc_bus = proc_mkdir("bus", NULL);
8102 +       proc_vx_init();
8103         proc_sys_init();
8104  }
8105
8106 diff -NurpP --minimal linux-2.6.25.11/fs/quota.c linux-2.6.25.11-vs2.3.0.34.14/fs/quota.c
8107 --- linux-2.6.25.11/fs/quota.c  2008-04-17 12:05:41.000000000 -0400
8108 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/quota.c    2008-04-21 09:21:04.000000000 -0400
8109 @@ -18,6 +18,7 @@
8110  #include <linux/capability.h>
8111  #include <linux/quotaops.h>
8112  #include <linux/types.h>
8113 +#include <linux/vs_context.h>
8114
8115  /* Check validity of generic quotactl commands */
8116  static int generic_quotactl_valid(struct super_block *sb, int type, int cmd, qid_t id)
8117 @@ -82,11 +83,11 @@ static int generic_quotactl_valid(struct
8118         if (cmd == Q_GETQUOTA) {
8119                 if (((type == USRQUOTA && current->euid != id) ||
8120                      (type == GRPQUOTA && !in_egroup_p(id))) &&
8121 -                   !capable(CAP_SYS_ADMIN))
8122 +                   !vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL))
8123                         return -EPERM;
8124         }
8125         else if (cmd != Q_GETFMT && cmd != Q_SYNC && cmd != Q_GETINFO)
8126 -               if (!capable(CAP_SYS_ADMIN))
8127 +               if (!vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL))
8128                         return -EPERM;
8129
8130         return 0;
8131 @@ -133,10 +134,10 @@ static int xqm_quotactl_valid(struct sup
8132         if (cmd == Q_XGETQUOTA) {
8133                 if (((type == XQM_USRQUOTA && current->euid != id) ||
8134                      (type == XQM_GRPQUOTA && !in_egroup_p(id))) &&
8135 -                    !capable(CAP_SYS_ADMIN))
8136 +                    !vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL))
8137                         return -EPERM;
8138         } else if (cmd != Q_XGETQSTAT && cmd != Q_XQUOTASYNC) {
8139 -               if (!capable(CAP_SYS_ADMIN))
8140 +               if (!vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL))
8141                         return -EPERM;
8142         }
8143
8144 @@ -329,6 +330,46 @@ static int do_quotactl(struct super_bloc
8145         return 0;
8146  }
8147
8148 +#if defined(CONFIG_BLK_DEV_VROOT) || defined(CONFIG_BLK_DEV_VROOT_MODULE)
8149 +
8150 +#include <linux/vroot.h>
8151 +#include <linux/major.h>
8152 +#include <linux/module.h>
8153 +#include <linux/kallsyms.h>
8154 +#include <linux/vserver/debug.h>
8155 +
8156 +static vroot_grb_func *vroot_get_real_bdev = NULL;
8157 +
8158 +static spinlock_t vroot_grb_lock = SPIN_LOCK_UNLOCKED;
8159 +
8160 +int register_vroot_grb(vroot_grb_func *func) {
8161 +       int ret = -EBUSY;
8162 +
8163 +       spin_lock(&vroot_grb_lock);
8164 +       if (!vroot_get_real_bdev) {
8165 +               vroot_get_real_bdev = func;
8166 +               ret = 0;
8167 +       }
8168 +       spin_unlock(&vroot_grb_lock);
8169 +       return ret;
8170 +}
8171 +EXPORT_SYMBOL(register_vroot_grb);
8172 +
8173 +int unregister_vroot_grb(vroot_grb_func *func) {
8174 +       int ret = -EINVAL;
8175 +
8176 +       spin_lock(&vroot_grb_lock);
8177 +       if (vroot_get_real_bdev) {
8178 +               vroot_get_real_bdev = NULL;
8179 +               ret = 0;
8180 +       }
8181 +       spin_unlock(&vroot_grb_lock);
8182 +       return ret;
8183 +}
8184 +EXPORT_SYMBOL(unregister_vroot_grb);
8185 +
8186 +#endif
8187 +
8188  /*
8189   * look up a superblock on which quota ops will be performed
8190   * - use the name of a block device to find the superblock thereon
8191 @@ -346,6 +387,22 @@ static inline struct super_block *quotac
8192         putname(tmp);
8193         if (IS_ERR(bdev))
8194                 return ERR_CAST(bdev);
8195 +#if defined(CONFIG_BLK_DEV_VROOT) || defined(CONFIG_BLK_DEV_VROOT_MODULE)
8196 +       if (bdev && bdev->bd_inode &&
8197 +                       imajor(bdev->bd_inode) == VROOT_MAJOR) {
8198 +               struct block_device *bdnew = (void *)-EINVAL;
8199 +
8200 +               if (vroot_get_real_bdev)
8201 +                       bdnew = vroot_get_real_bdev(bdev);
8202 +               else
8203 +                       vxdprintk(VXD_CBIT(misc, 0),
8204 +                                       "vroot_get_real_bdev not set");
8205 +               bdput(bdev);
8206 +               if (IS_ERR(bdnew))
8207 +                       return ERR_PTR(PTR_ERR(bdnew));
8208 +               bdev = bdnew;
8209 +       }
8210 +#endif
8211         sb = get_super(bdev);
8212         bdput(bdev);
8213         if (!sb)
8214 diff -NurpP --minimal linux-2.6.25.11/fs/reiserfs/bitmap.c linux-2.6.25.11-vs2.3.0.34.14/fs/reiserfs/bitmap.c
8215 --- linux-2.6.25.11/fs/reiserfs/bitmap.c        2008-04-17 12:05:41.000000000 -0400
8216 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/reiserfs/bitmap.c  2008-04-19 15:14:52.000000000 -0400
8217 @@ -13,6 +13,7 @@
8218  #include <linux/reiserfs_fs_sb.h>
8219  #include <linux/reiserfs_fs_i.h>
8220  #include <linux/quotaops.h>
8221 +#include <linux/vs_dlimit.h>
8222
8223  #define PREALLOCATION_SIZE 9
8224
8225 @@ -429,8 +430,10 @@ static void _reiserfs_free_block(struct
8226         set_sb_free_blocks(rs, sb_free_blocks(rs) + 1);
8227
8228         journal_mark_dirty(th, s, sbh);
8229 -       if (for_unformatted)
8230 +       if (for_unformatted) {
8231 +               DLIMIT_FREE_BLOCK(inode, 1);
8232                 DQUOT_FREE_BLOCK_NODIRTY(inode, 1);
8233 +       }
8234  }
8235
8236  void reiserfs_free_block(struct reiserfs_transaction_handle *th,
8237 @@ -1045,6 +1048,7 @@ static inline int blocknrs_and_prealloc_
8238         b_blocknr_t finish = SB_BLOCK_COUNT(s) - 1;
8239         int passno = 0;
8240         int nr_allocated = 0;
8241 +       int blocks;
8242
8243         determine_prealloc_size(hint);
8244         if (!hint->formatted_node) {
8245 @@ -1054,19 +1058,30 @@ static inline int blocknrs_and_prealloc_
8246                                "reiserquota: allocating %d blocks id=%u",
8247                                amount_needed, hint->inode->i_uid);
8248  #endif
8249 -               quota_ret =
8250 -                   DQUOT_ALLOC_BLOCK_NODIRTY(hint->inode, amount_needed);
8251 -               if (quota_ret)  /* Quota exceeded? */
8252 +               quota_ret = DQUOT_ALLOC_BLOCK_NODIRTY(hint->inode,
8253 +                       amount_needed);
8254 +               if (quota_ret)
8255                         return QUOTA_EXCEEDED;
8256 +               if (DLIMIT_ALLOC_BLOCK(hint->inode, amount_needed)) {
8257 +                       DQUOT_FREE_BLOCK_NODIRTY(hint->inode,
8258 +                               amount_needed);
8259 +                       return NO_DISK_SPACE;
8260 +               }
8261 +
8262                 if (hint->preallocate && hint->prealloc_size) {
8263  #ifdef REISERQUOTA_DEBUG
8264                         reiserfs_debug(s, REISERFS_DEBUG_CODE,
8265                                        "reiserquota: allocating (prealloc) %d blocks id=%u",
8266                                        hint->prealloc_size, hint->inode->i_uid);
8267  #endif
8268 -                       quota_ret =
8269 -                           DQUOT_PREALLOC_BLOCK_NODIRTY(hint->inode,
8270 -                                                        hint->prealloc_size);
8271 +                       quota_ret = DQUOT_PREALLOC_BLOCK_NODIRTY(hint->inode,
8272 +                               hint->prealloc_size);
8273 +                       if (!quota_ret &&
8274 +                               DLIMIT_ALLOC_BLOCK(hint->inode, hint->prealloc_size)) {
8275 +                               DQUOT_FREE_BLOCK_NODIRTY(hint->inode,
8276 +                                       hint->prealloc_size);
8277 +                               quota_ret = 1;
8278 +                       }
8279                         if (quota_ret)
8280                                 hint->preallocate = hint->prealloc_size = 0;
8281                 }
8282 @@ -1098,7 +1113,10 @@ static inline int blocknrs_and_prealloc_
8283                                                nr_allocated,
8284                                                hint->inode->i_uid);
8285  #endif
8286 -                               DQUOT_FREE_BLOCK_NODIRTY(hint->inode, amount_needed + hint->prealloc_size - nr_allocated);      /* Free not allocated blocks */
8287 +                               /* Free not allocated blocks */
8288 +                               blocks = amount_needed + hint->prealloc_size - nr_allocated;
8289 +                               DLIMIT_FREE_BLOCK(hint->inode, blocks);
8290 +                               DQUOT_FREE_BLOCK_NODIRTY(hint->inode, blocks);
8291                         }
8292                         while (nr_allocated--)
8293                                 reiserfs_free_block(hint->th, hint->inode,
8294 @@ -1129,10 +1147,10 @@ static inline int blocknrs_and_prealloc_
8295                                REISERFS_I(hint->inode)->i_prealloc_count,
8296                                hint->inode->i_uid);
8297  #endif
8298 -               DQUOT_FREE_BLOCK_NODIRTY(hint->inode, amount_needed +
8299 -                                        hint->prealloc_size - nr_allocated -
8300 -                                        REISERFS_I(hint->inode)->
8301 -                                        i_prealloc_count);
8302 +               blocks = amount_needed + hint->prealloc_size - nr_allocated -
8303 +                       REISERFS_I(hint->inode)->i_prealloc_count;
8304 +               DLIMIT_FREE_BLOCK(hint->inode, blocks);
8305 +               DQUOT_FREE_BLOCK_NODIRTY(hint->inode, blocks);
8306         }
8307
8308         return CARRY_ON;
8309 diff -NurpP --minimal linux-2.6.25.11/fs/reiserfs/file.c linux-2.6.25.11-vs2.3.0.34.14/fs/reiserfs/file.c
8310 --- linux-2.6.25.11/fs/reiserfs/file.c  2008-04-17 11:31:38.000000000 -0400
8311 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/reiserfs/file.c    2008-04-19 15:14:52.000000000 -0400
8312 @@ -306,4 +306,5 @@ const struct inode_operations reiserfs_f
8313         .listxattr = reiserfs_listxattr,
8314         .removexattr = reiserfs_removexattr,
8315         .permission = reiserfs_permission,
8316 +       .sync_flags = reiserfs_sync_flags,
8317  };
8318 diff -NurpP --minimal linux-2.6.25.11/fs/reiserfs/inode.c linux-2.6.25.11-vs2.3.0.34.14/fs/reiserfs/inode.c
8319 --- linux-2.6.25.11/fs/reiserfs/inode.c 2008-04-17 12:05:41.000000000 -0400
8320 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/reiserfs/inode.c   2008-07-08 05:17:07.000000000 -0400
8321 @@ -18,6 +18,8 @@
8322  #include <linux/writeback.h>
8323  #include <linux/quotaops.h>
8324  #include <linux/swap.h>
8325 +#include <linux/vs_dlimit.h>
8326 +#include <linux/vs_tag.h>
8327
8328  int reiserfs_commit_write(struct file *f, struct page *page,
8329                           unsigned from, unsigned to);
8330 @@ -52,6 +54,7 @@ void reiserfs_delete_inode(struct inode
8331                  * stat data deletion */
8332                 if (!err)
8333                         DQUOT_FREE_INODE(inode);
8334 +               DLIMIT_FREE_INODE(inode);
8335
8336                 if (journal_end(&th, inode->i_sb, jbegin_count))
8337                         goto out;
8338 @@ -1114,6 +1117,8 @@ static void init_inode(struct inode *ino
8339         struct buffer_head *bh;
8340         struct item_head *ih;
8341         __u32 rdev;
8342 +       uid_t uid;
8343 +       gid_t gid;
8344         //int version = ITEM_VERSION_1;
8345
8346         bh = PATH_PLAST_BUFFER(path);
8347 @@ -1137,12 +1142,13 @@ static void init_inode(struct inode *ino
8348                     (struct stat_data_v1 *)B_I_PITEM(bh, ih);
8349                 unsigned long blocks;
8350
8351 +               uid = sd_v1_uid(sd);
8352 +               gid = sd_v1_gid(sd);
8353 +
8354                 set_inode_item_key_version(inode, KEY_FORMAT_3_5);
8355                 set_inode_sd_version(inode, STAT_DATA_V1);
8356                 inode->i_mode = sd_v1_mode(sd);
8357                 inode->i_nlink = sd_v1_nlink(sd);
8358 -               inode->i_uid = sd_v1_uid(sd);
8359 -               inode->i_gid = sd_v1_gid(sd);
8360                 inode->i_size = sd_v1_size(sd);
8361                 inode->i_atime.tv_sec = sd_v1_atime(sd);
8362                 inode->i_mtime.tv_sec = sd_v1_mtime(sd);
8363 @@ -1184,11 +1190,12 @@ static void init_inode(struct inode *ino
8364                 // (directories and symlinks)
8365                 struct stat_data *sd = (struct stat_data *)B_I_PITEM(bh, ih);
8366
8367 +               uid    = sd_v2_uid(sd);
8368 +               gid    = sd_v2_gid(sd);
8369 +
8370                 inode->i_mode = sd_v2_mode(sd);
8371                 inode->i_nlink = sd_v2_nlink(sd);
8372 -               inode->i_uid = sd_v2_uid(sd);
8373                 inode->i_size = sd_v2_size(sd);
8374 -               inode->i_gid = sd_v2_gid(sd);
8375                 inode->i_mtime.tv_sec = sd_v2_mtime(sd);
8376                 inode->i_atime.tv_sec = sd_v2_atime(sd);
8377                 inode->i_ctime.tv_sec = sd_v2_ctime(sd);
8378 @@ -1218,6 +1225,10 @@ static void init_inode(struct inode *ino
8379                 sd_attrs_to_i_attrs(sd_v2_attrs(sd), inode);
8380         }
8381
8382 +       inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
8383 +       inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
8384 +       inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid, 0);
8385 +
8386         pathrelse(path);
8387         if (S_ISREG(inode->i_mode)) {
8388                 inode->i_op = &reiserfs_file_inode_operations;
8389 @@ -1240,13 +1251,15 @@ static void init_inode(struct inode *ino
8390  static void inode2sd(void *sd, struct inode *inode, loff_t size)
8391  {
8392         struct stat_data *sd_v2 = (struct stat_data *)sd;
8393 +       uid_t uid = TAGINO_UID(DX_TAG(inode), inode->i_uid, inode->i_tag);
8394 +       gid_t gid = TAGINO_GID(DX_TAG(inode), inode->i_gid, inode->i_tag);
8395         __u16 flags;
8396
8397 +       set_sd_v2_uid(sd_v2, uid);
8398 +       set_sd_v2_gid(sd_v2, gid);
8399         set_sd_v2_mode(sd_v2, inode->i_mode);
8400         set_sd_v2_nlink(sd_v2, inode->i_nlink);
8401 -       set_sd_v2_uid(sd_v2, inode->i_uid);
8402         set_sd_v2_size(sd_v2, size);
8403 -       set_sd_v2_gid(sd_v2, inode->i_gid);
8404         set_sd_v2_mtime(sd_v2, inode->i_mtime.tv_sec);
8405         set_sd_v2_atime(sd_v2, inode->i_atime.tv_sec);
8406         set_sd_v2_ctime(sd_v2, inode->i_ctime.tv_sec);
8407 @@ -1769,6 +1782,10 @@ int reiserfs_new_inode(struct reiserfs_t
8408
8409         BUG_ON(!th->t_trans_id);
8410
8411 +       if (DLIMIT_ALLOC_INODE(inode)) {
8412 +               err = -ENOSPC;
8413 +               goto out_bad_dlimit;
8414 +       }
8415         if (DQUOT_ALLOC_INODE(inode)) {
8416                 err = -EDQUOT;
8417                 goto out_end_trans;
8418 @@ -1954,6 +1971,9 @@ int reiserfs_new_inode(struct reiserfs_t
8419         DQUOT_FREE_INODE(inode);
8420
8421        out_end_trans:
8422 +       DLIMIT_FREE_INODE(inode);
8423 +
8424 +      out_bad_dlimit:
8425         journal_end(th, th->t_super, th->t_blocks_allocated);
8426         /* Drop can be outside and it needs more credits so it's better to have it outside */
8427         DQUOT_DROP(inode);
8428 @@ -2840,14 +2860,19 @@ int reiserfs_commit_write(struct file *f
8429  void sd_attrs_to_i_attrs(__u16 sd_attrs, struct inode *inode)
8430  {
8431         if (reiserfs_attrs(inode->i_sb)) {
8432 -               if (sd_attrs & REISERFS_SYNC_FL)
8433 -                       inode->i_flags |= S_SYNC;
8434 -               else
8435 -                       inode->i_flags &= ~S_SYNC;
8436                 if (sd_attrs & REISERFS_IMMUTABLE_FL)
8437                         inode->i_flags |= S_IMMUTABLE;
8438                 else
8439                         inode->i_flags &= ~S_IMMUTABLE;
8440 +               if (sd_attrs & REISERFS_IXUNLINK_FL)
8441 +                       inode->i_flags |= S_IXUNLINK;
8442 +               else
8443 +                       inode->i_flags &= ~S_IXUNLINK;
8444 +
8445 +               if (sd_attrs & REISERFS_SYNC_FL)
8446 +                       inode->i_flags |= S_SYNC;
8447 +               else
8448 +                       inode->i_flags &= ~S_SYNC;
8449                 if (sd_attrs & REISERFS_APPEND_FL)
8450                         inode->i_flags |= S_APPEND;
8451                 else
8452 @@ -2860,6 +2885,15 @@ void sd_attrs_to_i_attrs(__u16 sd_attrs,
8453                         REISERFS_I(inode)->i_flags |= i_nopack_mask;
8454                 else
8455                         REISERFS_I(inode)->i_flags &= ~i_nopack_mask;
8456 +
8457 +               if (sd_attrs & REISERFS_BARRIER_FL)
8458 +                       inode->i_vflags |= V_BARRIER;
8459 +               else
8460 +                       inode->i_vflags &= ~V_BARRIER;
8461 +               if (sd_attrs & REISERFS_COW_FL)
8462 +                       inode->i_vflags |= V_COW;
8463 +               else
8464 +                       inode->i_vflags &= ~V_COW;
8465         }
8466  }
8467
8468 @@ -2870,6 +2904,11 @@ void i_attrs_to_sd_attrs(struct inode *i
8469                         *sd_attrs |= REISERFS_IMMUTABLE_FL;
8470                 else
8471                         *sd_attrs &= ~REISERFS_IMMUTABLE_FL;
8472 +               if (inode->i_flags & S_IXUNLINK)
8473 +                       *sd_attrs |= REISERFS_IXUNLINK_FL;
8474 +               else
8475 +                       *sd_attrs &= ~REISERFS_IXUNLINK_FL;
8476 +
8477                 if (inode->i_flags & S_SYNC)
8478                         *sd_attrs |= REISERFS_SYNC_FL;
8479                 else
8480 @@ -2882,6 +2921,15 @@ void i_attrs_to_sd_attrs(struct inode *i
8481                         *sd_attrs |= REISERFS_NOTAIL_FL;
8482                 else
8483                         *sd_attrs &= ~REISERFS_NOTAIL_FL;
8484 +
8485 +               if (inode->i_vflags & V_BARRIER)
8486 +                       *sd_attrs |= REISERFS_BARRIER_FL;
8487 +               else
8488 +                       *sd_attrs &= ~REISERFS_BARRIER_FL;
8489 +               if (inode->i_vflags & V_COW)
8490 +                       *sd_attrs |= REISERFS_COW_FL;
8491 +               else
8492 +                       *sd_attrs &= ~REISERFS_COW_FL;
8493         }
8494  }
8495
8496 @@ -3049,6 +3097,22 @@ static ssize_t reiserfs_direct_IO(int rw
8497                                   reiserfs_get_blocks_direct_io, NULL);
8498  }
8499
8500 +int reiserfs_sync_flags(struct inode *inode)
8501 +{
8502 +       u16 oldflags, newflags;
8503 +
8504 +       oldflags = REISERFS_I(inode)->i_attrs;
8505 +       newflags = oldflags;
8506 +       i_attrs_to_sd_attrs(inode, &newflags);
8507 +
8508 +       if (oldflags ^ newflags) {
8509 +               REISERFS_I(inode)->i_attrs = newflags;
8510 +               inode->i_ctime = CURRENT_TIME_SEC;
8511 +               mark_inode_dirty(inode);
8512 +       }
8513 +       return 0;
8514 +}
8515 +
8516  int reiserfs_setattr(struct dentry *dentry, struct iattr *attr)
8517  {
8518         struct inode *inode = dentry->d_inode;
8519 @@ -3102,9 +3166,11 @@ int reiserfs_setattr(struct dentry *dent
8520         }
8521
8522         error = inode_change_ok(inode, attr);
8523 +
8524         if (!error) {
8525                 if ((ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) ||
8526 -                   (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) {
8527 +                   (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid) ||
8528 +                   (ia_valid & ATTR_TAG && attr->ia_tag != inode->i_tag)) {
8529                         error = reiserfs_chown_xattrs(inode, attr);
8530
8531                         if (!error) {
8532 @@ -3134,6 +3200,9 @@ int reiserfs_setattr(struct dentry *dent
8533                                         inode->i_uid = attr->ia_uid;
8534                                 if (attr->ia_valid & ATTR_GID)
8535                                         inode->i_gid = attr->ia_gid;
8536 +                               if ((attr->ia_valid & ATTR_TAG) &&
8537 +                                       IS_TAGGED(inode))
8538 +                                       inode->i_tag = attr->ia_tag;
8539                                 mark_inode_dirty(inode);
8540                                 error =
8541                                     journal_end(&th, inode->i_sb, jbegin_count);
8542 diff -NurpP --minimal linux-2.6.25.11/fs/reiserfs/ioctl.c linux-2.6.25.11-vs2.3.0.34.14/fs/reiserfs/ioctl.c
8543 --- linux-2.6.25.11/fs/reiserfs/ioctl.c 2008-07-15 22:05:17.000000000 -0400
8544 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/reiserfs/ioctl.c   2008-07-11 02:52:05.000000000 -0400
8545 @@ -6,6 +6,7 @@
8546  #include <linux/fs.h>
8547  #include <linux/reiserfs_fs.h>
8548  #include <linux/time.h>
8549 +#include <linux/mount.h>
8550  #include <asm/uaccess.h>
8551  #include <linux/pagemap.h>
8552  #include <linux/smp_lock.h>
8553 @@ -22,7 +23,7 @@
8554  int reiserfs_ioctl(struct inode *inode, struct file *filp, unsigned int cmd,
8555                    unsigned long arg)
8556  {
8557 -       unsigned int flags;
8558 +       unsigned int flags, oldflags;
8559
8560         switch (cmd) {
8561         case REISERFS_IOC_UNPACK:
8562 @@ -41,12 +42,14 @@ int reiserfs_ioctl(struct inode *inode,
8563
8564                 flags = REISERFS_I(inode)->i_attrs;
8565                 i_attrs_to_sd_attrs(inode, (__u16 *) & flags);
8566 +               flags &= REISERFS_FL_USER_VISIBLE;
8567                 return put_user(flags, (int __user *)arg);
8568         case REISERFS_IOC_SETFLAGS:{
8569                         if (!reiserfs_attrs(inode->i_sb))
8570                                 return -ENOTTY;
8571
8572 -                       if (IS_RDONLY(inode))
8573 +                       if (IS_RDONLY(inode) ||
8574 +                               (filp && MNT_IS_RDONLY(filp->f_vfsmnt)))
8575                                 return -EROFS;
8576
8577                         if (!is_owner_or_cap(inode))
8578 @@ -55,6 +58,11 @@ int reiserfs_ioctl(struct inode *inode,
8579                         if (get_user(flags, (int __user *)arg))
8580                                 return -EFAULT;
8581
8582 +                       if (IS_BARRIER(inode)) {
8583 +                               vxwprintk_task(1, "messing with the barrier.");
8584 +                               return -EACCES;
8585 +                       }
8586 +
8587                         /* Is it quota file? Do not allow user to mess with it. */
8588                         if (IS_NOQUOTA(inode))
8589                                 return -EPERM;
8590 @@ -72,6 +80,10 @@ int reiserfs_ioctl(struct inode *inode,
8591                                 if (result)
8592                                         return result;
8593                         }
8594 +
8595 +                       oldflags = REISERFS_I(inode)->i_attrs;
8596 +                       flags = flags & REISERFS_FL_USER_MODIFIABLE;
8597 +                       flags |= oldflags & ~REISERFS_FL_USER_MODIFIABLE;
8598                         sd_attrs_to_i_attrs(flags, inode);
8599                         REISERFS_I(inode)->i_attrs = flags;
8600                         inode->i_ctime = CURRENT_TIME_SEC;
8601 @@ -83,7 +95,8 @@ int reiserfs_ioctl(struct inode *inode,
8602         case REISERFS_IOC_SETVERSION:
8603                 if (!is_owner_or_cap(inode))
8604                         return -EPERM;
8605 -               if (IS_RDONLY(inode))
8606 +               if (IS_RDONLY(inode) ||
8607 +                       (filp && MNT_IS_RDONLY(filp->f_vfsmnt)))
8608                         return -EROFS;
8609                 if (get_user(inode->i_generation, (int __user *)arg))
8610                         return -EFAULT;
8611 diff -NurpP --minimal linux-2.6.25.11/fs/reiserfs/namei.c linux-2.6.25.11-vs2.3.0.34.14/fs/reiserfs/namei.c
8612 --- linux-2.6.25.11/fs/reiserfs/namei.c 2008-04-17 12:05:41.000000000 -0400
8613 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/reiserfs/namei.c   2008-04-19 15:14:52.000000000 -0400
8614 @@ -17,6 +17,7 @@
8615  #include <linux/reiserfs_acl.h>
8616  #include <linux/reiserfs_xattr.h>
8617  #include <linux/quotaops.h>
8618 +#include <linux/vs_tag.h>
8619
8620  #define INC_DIR_INODE_NLINK(i) if (i->i_nlink != 1) { inc_nlink(i); if (i->i_nlink >= REISERFS_LINK_MAX) i->i_nlink=1; }
8621  #define DEC_DIR_INODE_NLINK(i) if (i->i_nlink != 1) drop_nlink(i);
8622 @@ -360,6 +361,7 @@ static struct dentry *reiserfs_lookup(st
8623                         reiserfs_write_unlock(dir->i_sb);
8624                         return ERR_PTR(-EACCES);
8625                 }
8626 +               dx_propagate_tag(nd, inode);
8627
8628                 /* Propogate the priv_object flag so we know we're in the priv tree */
8629                 if (is_reiserfs_priv_object(dir))
8630 @@ -595,6 +597,7 @@ static int new_inode_init(struct inode *
8631         } else {
8632                 inode->i_gid = current->fsgid;
8633         }
8634 +       inode->i_tag = dx_current_fstag(inode->i_sb);
8635         DQUOT_INIT(inode);
8636         return 0;
8637  }
8638 @@ -1541,6 +1544,7 @@ const struct inode_operations reiserfs_d
8639         .listxattr = reiserfs_listxattr,
8640         .removexattr = reiserfs_removexattr,
8641         .permission = reiserfs_permission,
8642 +       .sync_flags = reiserfs_sync_flags,
8643  };
8644
8645  /*
8646 @@ -1557,6 +1561,7 @@ const struct inode_operations reiserfs_s
8647         .listxattr = reiserfs_listxattr,
8648         .removexattr = reiserfs_removexattr,
8649         .permission = reiserfs_permission,
8650 +       .sync_flags = reiserfs_sync_flags,
8651
8652  };
8653
8654 @@ -1570,5 +1575,6 @@ const struct inode_operations reiserfs_s
8655         .listxattr = reiserfs_listxattr,
8656         .removexattr = reiserfs_removexattr,
8657         .permission = reiserfs_permission,
8658 +       .sync_flags = reiserfs_sync_flags,
8659
8660  };
8661 diff -NurpP --minimal linux-2.6.25.11/fs/reiserfs/stree.c linux-2.6.25.11-vs2.3.0.34.14/fs/reiserfs/stree.c
8662 --- linux-2.6.25.11/fs/reiserfs/stree.c 2008-04-17 11:31:38.000000000 -0400
8663 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/reiserfs/stree.c   2008-04-19 15:14:52.000000000 -0400
8664 @@ -55,6 +55,7 @@
8665  #include <linux/reiserfs_fs.h>
8666  #include <linux/buffer_head.h>
8667  #include <linux/quotaops.h>
8668 +#include <linux/vs_dlimit.h>
8669
8670  /* Does the buffer contain a disk block which is in the tree. */
8671  inline int B_IS_IN_TREE(const struct buffer_head *p_s_bh)
8672 @@ -1297,6 +1298,7 @@ int reiserfs_delete_item(struct reiserfs
8673                        "reiserquota delete_item(): freeing %u, id=%u type=%c",
8674                        quota_cut_bytes, p_s_inode->i_uid, head2type(&s_ih));
8675  #endif
8676 +       DLIMIT_FREE_SPACE(p_s_inode, quota_cut_bytes);
8677         DQUOT_FREE_SPACE_NODIRTY(p_s_inode, quota_cut_bytes);
8678
8679         /* Return deleted body length */
8680 @@ -1385,6 +1387,7 @@ void reiserfs_delete_solid_item(struct r
8681  #endif
8682                                 DQUOT_FREE_SPACE_NODIRTY(inode,
8683                                                          quota_cut_bytes);
8684 +                               DLIMIT_FREE_SPACE(inode, quota_cut_bytes);
8685                         }
8686                         break;
8687                 }
8688 @@ -1735,6 +1738,7 @@ int reiserfs_cut_from_item(struct reiser
8689                        "reiserquota cut_from_item(): freeing %u id=%u type=%c",
8690                        quota_cut_bytes, p_s_inode->i_uid, '?');
8691  #endif
8692 +       DLIMIT_FREE_SPACE(p_s_inode, quota_cut_bytes);
8693         DQUOT_FREE_SPACE_NODIRTY(p_s_inode, quota_cut_bytes);
8694         return n_ret_value;
8695  }
8696 @@ -1976,6 +1980,11 @@ int reiserfs_paste_into_item(struct reis
8697                 pathrelse(p_s_search_path);
8698                 return -EDQUOT;
8699         }
8700 +       if (DLIMIT_ALLOC_SPACE(inode, n_pasted_size)) {
8701 +               DQUOT_FREE_SPACE_NODIRTY(inode, n_pasted_size);
8702 +               pathrelse(p_s_search_path);
8703 +               return -ENOSPC;
8704 +       }
8705         init_tb_struct(th, &s_paste_balance, th->t_super, p_s_search_path,
8706                        n_pasted_size);
8707  #ifdef DISPLACE_NEW_PACKING_LOCALITIES
8708 @@ -2028,6 +2037,7 @@ int reiserfs_paste_into_item(struct reis
8709                        n_pasted_size, inode->i_uid,
8710                        key2type(&(p_s_key->on_disk_key)));
8711  #endif
8712 +       DLIMIT_FREE_SPACE(inode, n_pasted_size);
8713         DQUOT_FREE_SPACE_NODIRTY(inode, n_pasted_size);
8714         return retval;
8715  }
8716 @@ -2065,6 +2075,11 @@ int reiserfs_insert_item(struct reiserfs
8717                         pathrelse(p_s_path);
8718                         return -EDQUOT;
8719                 }
8720 +               if (DLIMIT_ALLOC_SPACE(inode, quota_bytes)) {
8721 +                       DQUOT_FREE_SPACE_NODIRTY(inode, quota_bytes);
8722 +                       pathrelse(p_s_path);
8723 +                       return -ENOSPC;
8724 +               }
8725         }
8726         init_tb_struct(th, &s_ins_balance, th->t_super, p_s_path,
8727                        IH_SIZE + ih_item_len(p_s_ih));
8728 @@ -2112,7 +2127,9 @@ int reiserfs_insert_item(struct reiserfs
8729                        "reiserquota insert_item(): freeing %u id=%u type=%c",
8730                        quota_bytes, inode->i_uid, head2type(p_s_ih));
8731  #endif
8732 -       if (inode)
8733 +       if (inode) {
8734 +               DLIMIT_FREE_SPACE(inode, quota_bytes);
8735                 DQUOT_FREE_SPACE_NODIRTY(inode, quota_bytes);
8736 +       }
8737         return retval;
8738  }
8739 diff -NurpP --minimal linux-2.6.25.11/fs/reiserfs/super.c linux-2.6.25.11-vs2.3.0.34.14/fs/reiserfs/super.c
8740 --- linux-2.6.25.11/fs/reiserfs/super.c 2008-07-15 22:05:17.000000000 -0400
8741 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/reiserfs/super.c   2008-05-21 14:30:41.000000000 -0400
8742 @@ -896,6 +896,14 @@ static int reiserfs_parse_options(struct
8743                 {"user_xattr",.setmask = 1 << REISERFS_UNSUPPORTED_OPT},
8744                 {"nouser_xattr",.clrmask = 1 << REISERFS_UNSUPPORTED_OPT},
8745  #endif
8746 +#ifndef CONFIG_TAGGING_NONE
8747 +               {"tagxid",.setmask = 1 << REISERFS_TAGGED},
8748 +               {"tag",.setmask = 1 << REISERFS_TAGGED},
8749 +               {"notag",.clrmask = 1 << REISERFS_TAGGED},
8750 +#endif
8751 +#ifdef CONFIG_PROPAGATE
8752 +               {"tag",.arg_required = 'T',.values = NULL},
8753 +#endif
8754  #ifdef CONFIG_REISERFS_FS_POSIX_ACL
8755                 {"acl",.setmask = 1 << REISERFS_POSIXACL},
8756                 {"noacl",.clrmask = 1 << REISERFS_POSIXACL},
8757 @@ -1159,6 +1167,12 @@ static int reiserfs_remount(struct super
8758                 goto out_err;
8759         }
8760
8761 +       if ((mount_options & (1 << REISERFS_TAGGED)) &&
8762 +               !(s->s_flags & MS_TAGGED)) {
8763 +               reiserfs_warning(s, "reiserfs: tagging not permitted on remount.");
8764 +               return -EINVAL;
8765 +       }
8766 +
8767         handle_attrs(s);
8768
8769         /* Add options that are safe here */
8770 @@ -1618,6 +1632,10 @@ static int reiserfs_fill_super(struct su
8771                 goto error;
8772         }
8773
8774 +       /* map mount option tagxid */
8775 +       if (REISERFS_SB(s)->s_mount_opt & (1 << REISERFS_TAGGED))
8776 +               s->s_flags |= MS_TAGGED;
8777 +
8778         rs = SB_DISK_SUPER_BLOCK(s);
8779         /* Let's do basic sanity check to verify that underlying device is not
8780            smaller than the filesystem. If the check fails then abort and scream,
8781 diff -NurpP --minimal linux-2.6.25.11/fs/reiserfs/xattr.c linux-2.6.25.11-vs2.3.0.34.14/fs/reiserfs/xattr.c
8782 --- linux-2.6.25.11/fs/reiserfs/xattr.c 2008-04-17 12:05:41.000000000 -0400
8783 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/reiserfs/xattr.c   2008-04-19 15:14:52.000000000 -0400
8784 @@ -35,6 +35,7 @@
8785  #include <linux/namei.h>
8786  #include <linux/errno.h>
8787  #include <linux/fs.h>
8788 +#include <linux/mount.h>
8789  #include <linux/file.h>
8790  #include <linux/pagemap.h>
8791  #include <linux/xattr.h>
8792 @@ -747,7 +748,7 @@ int reiserfs_delete_xattrs(struct inode
8793         if (dir->d_inode->i_nlink <= 2) {
8794                 root = get_xa_root(inode->i_sb, XATTR_REPLACE);
8795                 reiserfs_write_lock_xattrs(inode->i_sb);
8796 -               err = vfs_rmdir(root->d_inode, dir);
8797 +               err = vfs_rmdir(root->d_inode, dir, NULL);
8798                 reiserfs_write_unlock_xattrs(inode->i_sb);
8799                 dput(root);
8800         } else {
8801 diff -NurpP --minimal linux-2.6.25.11/fs/stat.c linux-2.6.25.11-vs2.3.0.34.14/fs/stat.c
8802 --- linux-2.6.25.11/fs/stat.c   2008-04-17 12:05:41.000000000 -0400
8803 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/stat.c     2008-04-19 15:14:52.000000000 -0400
8804 @@ -26,6 +26,7 @@ void generic_fillattr(struct inode *inod
8805         stat->nlink = inode->i_nlink;
8806         stat->uid = inode->i_uid;
8807         stat->gid = inode->i_gid;
8808 +       stat->tag = inode->i_tag;
8809         stat->rdev = inode->i_rdev;
8810         stat->atime = inode->i_atime;
8811         stat->mtime = inode->i_mtime;
8812 diff -NurpP --minimal linux-2.6.25.11/fs/super.c linux-2.6.25.11-vs2.3.0.34.14/fs/super.c
8813 --- linux-2.6.25.11/fs/super.c  2008-04-17 12:05:41.000000000 -0400
8814 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/super.c    2008-04-19 15:14:52.000000000 -0400
8815 @@ -37,6 +37,9 @@
8816  #include <linux/idr.h>
8817  #include <linux/kobject.h>
8818  #include <linux/mutex.h>
8819 +#include <linux/devpts_fs.h>
8820 +#include <linux/proc_fs.h>
8821 +#include <linux/vs_context.h>
8822  #include <asm/uaccess.h>
8823
8824
8825 @@ -859,12 +862,18 @@ struct vfsmount *
8826  vfs_kern_mount(struct file_system_type *type, int flags, const char *name, void *data)
8827  {
8828         struct vfsmount *mnt;
8829 +       struct super_block *sb;
8830         char *secdata = NULL;
8831         int error;
8832
8833         if (!type)
8834                 return ERR_PTR(-ENODEV);
8835
8836 +       error = -EPERM;
8837 +       if ((type->fs_flags & FS_BINARY_MOUNTDATA) &&
8838 +               !vx_capable(CAP_SYS_ADMIN, VXC_BINARY_MOUNT))
8839 +               goto out;
8840 +
8841         error = -ENOMEM;
8842         mnt = alloc_vfsmnt(name);
8843         if (!mnt)
8844 @@ -885,7 +894,14 @@ vfs_kern_mount(struct file_system_type *
8845                 goto out_free_secdata;
8846         BUG_ON(!mnt->mnt_sb);
8847
8848 -       error = security_sb_kern_mount(mnt->mnt_sb, secdata);
8849 +       sb = mnt->mnt_sb;
8850 +       error = -EPERM;
8851 +       if (!vx_capable(CAP_SYS_ADMIN, VXC_BINARY_MOUNT) && !sb->s_bdev &&
8852 +               (sb->s_magic != PROC_SUPER_MAGIC) &&
8853 +               (sb->s_magic != DEVPTS_SUPER_MAGIC))
8854 +               goto out_sb;
8855 +
8856 +       error = security_sb_kern_mount(sb, secdata);
8857         if (error)
8858                 goto out_sb;
8859
8860 diff -NurpP --minimal linux-2.6.25.11/fs/sysfs/mount.c linux-2.6.25.11-vs2.3.0.34.14/fs/sysfs/mount.c
8861 --- linux-2.6.25.11/fs/sysfs/mount.c    2008-04-17 11:31:38.000000000 -0400
8862 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/sysfs/mount.c      2008-04-19 15:14:52.000000000 -0400
8863 @@ -19,8 +19,6 @@
8864
8865  #include "sysfs.h"
8866
8867 -/* Random magic number */
8868 -#define SYSFS_MAGIC 0x62656572
8869
8870  static struct vfsmount *sysfs_mount;
8871  struct super_block * sysfs_sb = NULL;
8872 @@ -46,7 +44,7 @@ static int sysfs_fill_super(struct super
8873
8874         sb->s_blocksize = PAGE_CACHE_SIZE;
8875         sb->s_blocksize_bits = PAGE_CACHE_SHIFT;
8876 -       sb->s_magic = SYSFS_MAGIC;
8877 +       sb->s_magic = SYSFS_SUPER_MAGIC;
8878         sb->s_op = &sysfs_ops;
8879         sb->s_time_gran = 1;
8880         sysfs_sb = sb;
8881 diff -NurpP --minimal linux-2.6.25.11/fs/utimes.c linux-2.6.25.11-vs2.3.0.34.14/fs/utimes.c
8882 --- linux-2.6.25.11/fs/utimes.c 2008-07-15 22:05:17.000000000 -0400
8883 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/utimes.c   2008-05-21 14:30:41.000000000 -0400
8884 @@ -7,6 +7,8 @@
8885  #include <linux/stat.h>
8886  #include <linux/utime.h>
8887  #include <linux/syscalls.h>
8888 +#include <linux/mount.h>
8889 +#include <linux/vs_cowbl.h>
8890  #include <asm/uaccess.h>
8891  #include <asm/unistd.h>
8892
8893 @@ -89,6 +91,9 @@ long do_utimes(int dfd, char __user *fil
8894                 if (error)
8895                         goto out;
8896
8897 +               error = cow_check_and_break(&nd);
8898 +               if (error)
8899 +                       goto dput_and_out;
8900                 dentry = nd.path.dentry;
8901         }
8902
8903 diff -NurpP --minimal linux-2.6.25.11/fs/xattr.c linux-2.6.25.11-vs2.3.0.34.14/fs/xattr.c
8904 --- linux-2.6.25.11/fs/xattr.c  2008-04-17 12:05:41.000000000 -0400
8905 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/xattr.c    2008-04-21 13:52:50.000000000 -0400
8906 @@ -17,6 +17,7 @@
8907  #include <linux/module.h>
8908  #include <linux/fsnotify.h>
8909  #include <linux/audit.h>
8910 +#include <linux/mount.h>
8911  #include <asm/uaccess.h>
8912
8913
8914 @@ -220,7 +221,7 @@ EXPORT_SYMBOL_GPL(vfs_removexattr);
8915   */
8916  static long
8917  setxattr(struct dentry *d, char __user *name, void __user *value,
8918 -        size_t size, int flags)
8919 +        size_t size, int flags, struct vfsmount *mnt)
8920  {
8921         int error;
8922         void *kvalue = NULL;
8923 @@ -247,6 +248,9 @@ setxattr(struct dentry *d, char __user *
8924                 }
8925         }
8926
8927 +       if (MNT_IS_RDONLY(mnt))
8928 +               return -EROFS;
8929 +
8930         error = vfs_setxattr(d, kname, kvalue, size, flags);
8931         kfree(kvalue);
8932         return error;
8933 @@ -262,7 +266,7 @@ sys_setxattr(char __user *path, char __u
8934         error = user_path_walk(path, &nd);
8935         if (error)
8936                 return error;
8937 -       error = setxattr(nd.path.dentry, name, value, size, flags);
8938 +       error = setxattr(nd.path.dentry, name, value, size, flags, nd.path.mnt);
8939         path_put(&nd.path);
8940         return error;
8941  }
8942 @@ -277,7 +281,7 @@ sys_lsetxattr(char __user *path, char __
8943         error = user_path_walk_link(path, &nd);
8944         if (error)
8945                 return error;
8946 -       error = setxattr(nd.path.dentry, name, value, size, flags);
8947 +       error = setxattr(nd.path.dentry, name, value, size, flags, nd.path.mnt);
8948         path_put(&nd.path);
8949         return error;
8950  }
8951 @@ -295,7 +299,7 @@ sys_fsetxattr(int fd, char __user *name,
8952                 return error;
8953         dentry = f->f_path.dentry;
8954         audit_inode(NULL, dentry);
8955 -       error = setxattr(dentry, name, value, size, flags);
8956 +       error = setxattr(dentry, name, value, size, flags, f->f_vfsmnt);
8957         fput(f);
8958         return error;
8959  }
8960 @@ -459,7 +463,7 @@ sys_flistxattr(int fd, char __user *list
8961   * Extended attribute REMOVE operations
8962   */
8963  static long
8964 -removexattr(struct dentry *d, char __user *name)
8965 +removexattr(struct dentry *d, char __user *name, struct vfsmount *mnt)
8966  {
8967         int error;
8968         char kname[XATTR_NAME_MAX + 1];
8969 @@ -470,6 +474,9 @@ removexattr(struct dentry *d, char __use
8970         if (error < 0)
8971                 return error;
8972
8973 +       if (MNT_IS_RDONLY(mnt))
8974 +               return -EROFS;
8975 +
8976         return vfs_removexattr(d, kname);
8977  }
8978
8979 @@ -482,7 +489,7 @@ sys_removexattr(char __user *path, char
8980         error = user_path_walk(path, &nd);
8981         if (error)
8982                 return error;
8983 -       error = removexattr(nd.path.dentry, name);
8984 +       error = removexattr(nd.path.dentry, name, nd.path.mnt);
8985         path_put(&nd.path);
8986         return error;
8987  }
8988 @@ -496,7 +503,7 @@ sys_lremovexattr(char __user *path, char
8989         error = user_path_walk_link(path, &nd);
8990         if (error)
8991                 return error;
8992 -       error = removexattr(nd.path.dentry, name);
8993 +       error = removexattr(nd.path.dentry, name, nd.path.mnt);
8994         path_put(&nd.path);
8995         return error;
8996  }
8997 @@ -513,7 +520,7 @@ sys_fremovexattr(int fd, char __user *na
8998                 return error;
8999         dentry = f->f_path.dentry;
9000         audit_inode(NULL, dentry);
9001 -       error = removexattr(dentry, name);
9002 +       error = removexattr(dentry, name, f->f_vfsmnt);
9003         fput(f);
9004         return error;
9005  }
9006 diff -NurpP --minimal linux-2.6.25.11/fs/xfs/linux-2.6/xfs_ioctl.c linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/linux-2.6/xfs_ioctl.c
9007 --- linux-2.6.25.11/fs/xfs/linux-2.6/xfs_ioctl.c        2008-04-17 12:05:41.000000000 -0400
9008 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/linux-2.6/xfs_ioctl.c  2008-07-11 02:57:05.000000000 -0400
9009 @@ -770,9 +770,13 @@ xfs_ioctl(
9010                 return xfs_ioc_fsgetxattr(ip, 0, arg);
9011         case XFS_IOC_FSGETXATTRA:
9012                 return xfs_ioc_fsgetxattr(ip, 1, arg);
9013 -       case XFS_IOC_GETXFLAGS:
9014         case XFS_IOC_SETXFLAGS:
9015         case XFS_IOC_FSSETXATTR:
9016 +               if (IS_BARRIER(inode)) {
9017 +                       vxwprintk_task(1, "messing with the barrier.");
9018 +                       return -XFS_ERROR(EACCES);
9019 +               }
9020 +       case XFS_IOC_GETXFLAGS:
9021                 return xfs_ioc_xattr(ip, filp, cmd, arg);
9022
9023         case XFS_IOC_FSSETDM: {
9024 @@ -1125,12 +1129,16 @@ xfs_merge_ioc_xflags(
9025
9026  STATIC unsigned int
9027  xfs_di2lxflags(
9028 -       __uint16_t      di_flags)
9029 +       __uint16_t      di_flags,
9030 +       __uint16_t      di_vflags)
9031  {
9032         unsigned int    flags = 0;
9033
9034         if (di_flags & XFS_DIFLAG_IMMUTABLE)
9035                 flags |= FS_IMMUTABLE_FL;
9036 +       if (di_flags & XFS_DIFLAG_IXUNLINK)
9037 +               flags |= FS_IXUNLINK_FL;
9038 +
9039         if (di_flags & XFS_DIFLAG_APPEND)
9040                 flags |= FS_APPEND_FL;
9041         if (di_flags & XFS_DIFLAG_SYNC)
9042 @@ -1139,6 +1147,11 @@ xfs_di2lxflags(
9043                 flags |= FS_NOATIME_FL;
9044         if (di_flags & XFS_DIFLAG_NODUMP)
9045                 flags |= FS_NODUMP_FL;
9046 +
9047 +       if (di_vflags & XFS_DIVFLAG_BARRIER)
9048 +               flags |= FS_BARRIER_FL;
9049 +       if (di_vflags & XFS_DIVFLAG_COW)
9050 +               flags |= FS_COW_FL;
9051         return flags;
9052  }
9053
9054 @@ -1219,7 +1232,7 @@ xfs_ioc_xattr(
9055         }
9056
9057         case XFS_IOC_GETXFLAGS: {
9058 -               flags = xfs_di2lxflags(ip->i_d.di_flags);
9059 +               flags = xfs_di2lxflags(ip->i_d.di_flags, ip->i_d.di_vflags);
9060                 if (copy_to_user(arg, &flags, sizeof(flags)))
9061                         error = -EFAULT;
9062                 break;
9063 diff -NurpP --minimal linux-2.6.25.11/fs/xfs/linux-2.6/xfs_iops.c linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/linux-2.6/xfs_iops.c
9064 --- linux-2.6.25.11/fs/xfs/linux-2.6/xfs_iops.c 2008-04-17 12:05:41.000000000 -0400
9065 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/linux-2.6/xfs_iops.c   2008-07-10 23:30:31.000000000 -0400
9066 @@ -53,6 +53,7 @@
9067  #include <linux/namei.h>
9068  #include <linux/security.h>
9069  #include <linux/falloc.h>
9070 +#include <linux/vs_tag.h>
9071
9072  /*
9073   * Bring the atime in the XFS inode uptodate.
9074 @@ -391,6 +392,7 @@ xfs_vn_lookup(
9075                 d_add(dentry, NULL);
9076                 return NULL;
9077         }
9078 +       dx_propagate_tag(nd, vn_to_inode(cvp));
9079
9080         return d_splice_alias(vn_to_inode(cvp), dentry);
9081  }
9082 @@ -655,6 +657,10 @@ xfs_vn_setattr(
9083         int             flags = 0;
9084         int             error;
9085
9086 +       error = inode_change_ok(inode, attr);
9087 +       if (error)
9088 +               return error;
9089 +
9090         if (ia_valid & ATTR_UID) {
9091                 vattr.va_mask |= XFS_AT_UID;
9092                 vattr.va_uid = attr->ia_uid;
9093 @@ -663,6 +669,10 @@ xfs_vn_setattr(
9094                 vattr.va_mask |= XFS_AT_GID;
9095                 vattr.va_gid = attr->ia_gid;
9096         }
9097 +       if ((ia_valid & ATTR_TAG) && IS_TAGGED(inode)) {
9098 +               vattr.va_mask |= XFS_AT_TAG;
9099 +               vattr.va_tag = attr->ia_tag;
9100 +       }
9101         if (ia_valid & ATTR_SIZE) {
9102                 vattr.va_mask |= XFS_AT_SIZE;
9103                 vattr.va_size = attr->ia_size;
9104 @@ -708,6 +718,45 @@ xfs_vn_truncate(
9105  }
9106
9107  STATIC int
9108 +xfs_vn_sync_flags(struct inode *inode)
9109 +{
9110 +       unsigned int oldflags, newflags;
9111 +       int             flags = 0;
9112 +       int             error;
9113 +       bhv_vattr_t     vattr;
9114 +       bhv_vnode_t     *vp = vn_from_inode(inode);
9115 +
9116 +       memset(&vattr, 0, sizeof vattr);
9117 +
9118 +       vattr.va_mask = XFS_AT_XFLAGS;
9119 +       error = xfs_getattr(XFS_I(inode), &vattr, 0);
9120 +
9121 +       if (error)
9122 +               return error;
9123 +       oldflags = vattr.va_xflags;
9124 +       newflags = oldflags & ~(XFS_XFLAG_IMMUTABLE | XFS_XFLAG_IXUNLINK
9125 +               | XFS_XFLAG_BARRIER | XFS_XFLAG_COW);
9126 +
9127 +       if (IS_IMMUTABLE(inode))
9128 +               newflags |= XFS_XFLAG_IMMUTABLE;
9129 +       if (IS_IXUNLINK(inode))
9130 +               newflags |= XFS_XFLAG_IXUNLINK;
9131 +
9132 +       if (IS_BARRIER(inode))
9133 +               newflags |= XFS_XFLAG_BARRIER;
9134 +       if (IS_COW(inode))
9135 +               newflags |= XFS_XFLAG_COW;
9136 +
9137 +       if (oldflags ^ newflags) {
9138 +               vattr.va_xflags = newflags;
9139 +               vattr.va_mask |= XFS_AT_XFLAGS;
9140 +               error = xfs_setattr(XFS_I(inode), &vattr, flags, NULL);
9141 +       }
9142 +       vn_revalidate(vp);
9143 +       return error;
9144 +}
9145 +
9146 +STATIC int
9147  xfs_vn_setxattr(
9148         struct dentry   *dentry,
9149         const char      *name,
9150 @@ -862,6 +911,7 @@ const struct inode_operations xfs_inode_
9151         .listxattr              = xfs_vn_listxattr,
9152         .removexattr            = xfs_vn_removexattr,
9153         .fallocate              = xfs_vn_fallocate,
9154 +       .sync_flags             = xfs_vn_sync_flags,
9155  };
9156
9157  const struct inode_operations xfs_dir_inode_operations = {
9158 @@ -881,6 +931,7 @@ const struct inode_operations xfs_dir_in
9159         .getxattr               = xfs_vn_getxattr,
9160         .listxattr              = xfs_vn_listxattr,
9161         .removexattr            = xfs_vn_removexattr,
9162 +       .sync_flags             = xfs_vn_sync_flags,
9163  };
9164
9165  const struct inode_operations xfs_symlink_inode_operations = {
9166 @@ -894,4 +945,5 @@ const struct inode_operations xfs_symlin
9167         .getxattr               = xfs_vn_getxattr,
9168         .listxattr              = xfs_vn_listxattr,
9169         .removexattr            = xfs_vn_removexattr,
9170 +       .sync_flags             = xfs_vn_sync_flags,
9171  };
9172 diff -NurpP --minimal linux-2.6.25.11/fs/xfs/linux-2.6/xfs_linux.h linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/linux-2.6/xfs_linux.h
9173 --- linux-2.6.25.11/fs/xfs/linux-2.6/xfs_linux.h        2008-04-17 12:05:41.000000000 -0400
9174 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/linux-2.6/xfs_linux.h  2008-04-19 15:14:52.000000000 -0400
9175 @@ -128,6 +128,7 @@
9176  #define current_pid()          (current->pid)
9177  #define current_fsuid(cred)    (current->fsuid)
9178  #define current_fsgid(cred)    (current->fsgid)
9179 +#define current_fstag(cred,vp) (dx_current_fstag(vn_to_inode(vp)->i_sb))
9180  #define current_test_flags(f)  (current->flags & (f))
9181  #define current_set_flags_nested(sp, f)                \
9182                 (*(sp) = current->flags, current->flags |= (f))
9183 diff -NurpP --minimal linux-2.6.25.11/fs/xfs/linux-2.6/xfs_super.c linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/linux-2.6/xfs_super.c
9184 --- linux-2.6.25.11/fs/xfs/linux-2.6/xfs_super.c        2008-04-17 12:05:41.000000000 -0400
9185 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/linux-2.6/xfs_super.c  2008-07-11 00:48:43.000000000 -0400
9186 @@ -137,6 +137,9 @@ xfs_args_allocate(
9187  #define MNTOPT_DMAPI   "dmapi"         /* DMI enabled (DMAPI / XDSM) */
9188  #define MNTOPT_XDSM    "xdsm"          /* DMI enabled (DMAPI / XDSM) */
9189  #define MNTOPT_DMI     "dmi"           /* DMI enabled (DMAPI / XDSM) */
9190 +#define MNTOPT_TAGXID  "tagxid"        /* context tagging for inodes */
9191 +#define MNTOPT_TAGGED  "tag"           /* context tagging for inodes */
9192 +#define MNTOPT_NOTAGTAG        "notag"         /* do not use context tagging */
9193
9194  STATIC unsigned long
9195  suffix_strtoul(char *s, char **endp, unsigned int base)
9196 @@ -355,6 +358,19 @@ xfs_parseargs(
9197                 } else if (!strcmp(this_char, "irixsgid")) {
9198                         cmn_err(CE_WARN,
9199         "XFS: irixsgid is now a sysctl(2) variable, option is deprecated.");
9200 +#ifndef CONFIG_TAGGING_NONE
9201 +               } else if (!strcmp(this_char, MNTOPT_TAGGED)) {
9202 +                       args->flags2 |= XFSMNT2_TAGGED;
9203 +               } else if (!strcmp(this_char, MNTOPT_NOTAGTAG)) {
9204 +                       args->flags2 &= ~XFSMNT2_TAGGED;
9205 +               } else if (!strcmp(this_char, MNTOPT_TAGXID)) {
9206 +                       args->flags2 |= XFSMNT2_TAGGED;
9207 +#endif
9208 +#ifdef CONFIG_PROPAGATE
9209 +               } else if (!strcmp(this_char, MNTOPT_TAGGED)) {
9210 +                       /* use value */
9211 +                       args->flags2 |= XFSMNT2_TAGGED;
9212 +#endif
9213                 } else {
9214                         cmn_err(CE_WARN,
9215                                 "XFS: unknown mount option [%s].", this_char);
9216 @@ -591,6 +607,7 @@ xfs_revalidate_inode(
9217         inode->i_nlink  = ip->i_d.di_nlink;
9218         inode->i_uid    = ip->i_d.di_uid;
9219         inode->i_gid    = ip->i_d.di_gid;
9220 +       inode->i_tag    = ip->i_d.di_tag;
9221
9222         switch (inode->i_mode & S_IFMT) {
9223         case S_IFBLK:
9224 @@ -612,10 +629,16 @@ xfs_revalidate_inode(
9225         inode->i_mtime.tv_nsec  = ip->i_d.di_mtime.t_nsec;
9226         inode->i_ctime.tv_sec   = ip->i_d.di_ctime.t_sec;
9227         inode->i_ctime.tv_nsec  = ip->i_d.di_ctime.t_nsec;
9228 +
9229         if (ip->i_d.di_flags & XFS_DIFLAG_IMMUTABLE)
9230                 inode->i_flags |= S_IMMUTABLE;
9231         else
9232                 inode->i_flags &= ~S_IMMUTABLE;
9233 +       if (ip->i_d.di_flags & XFS_DIFLAG_IXUNLINK)
9234 +               inode->i_flags |= S_IXUNLINK;
9235 +       else
9236 +               inode->i_flags &= ~S_IXUNLINK;
9237 +
9238         if (ip->i_d.di_flags & XFS_DIFLAG_APPEND)
9239                 inode->i_flags |= S_APPEND;
9240         else
9241 @@ -628,6 +651,16 @@ xfs_revalidate_inode(
9242                 inode->i_flags |= S_NOATIME;
9243         else
9244                 inode->i_flags &= ~S_NOATIME;
9245 +
9246 +       if (ip->i_d.di_vflags & XFS_DIVFLAG_BARRIER)
9247 +               inode->i_vflags |= V_BARRIER;
9248 +       else
9249 +               inode->i_vflags &= ~V_BARRIER;
9250 +       if (ip->i_d.di_vflags & XFS_DIVFLAG_COW)
9251 +               inode->i_vflags |= V_COW;
9252 +       else
9253 +               inode->i_vflags &= ~V_COW;
9254 +
9255         xfs_iflags_clear(ip, XFS_IMODIFIED);
9256  }
9257
9258 @@ -1220,6 +1253,12 @@ xfs_fs_remount(
9259         int                     error;
9260
9261         error = xfs_parseargs(mp, options, args, 1);
9262 +       if ((args->flags2 & XFSMNT2_TAGGED) &&
9263 +               !(sb->s_flags & MS_TAGGED)) {
9264 +               printk("XFS: %s: tagging not permitted on remount.\n",
9265 +                       sb->s_id);
9266 +               error = EINVAL;
9267 +       }
9268         if (!error)
9269                 error = xfs_mntupdate(mp, flags, args);
9270         kmem_free(args, sizeof(*args));
9271 @@ -1336,6 +1375,9 @@ xfs_fs_fill_super(
9272         if (error)
9273                 goto fail_vfsop;
9274
9275 +       if (mp->m_flags & XFS_MOUNT_TAGGED)
9276 +               sb->s_flags |= MS_TAGGED;
9277 +
9278         sb->s_dirt = 1;
9279         sb->s_magic = XFS_SB_MAGIC;
9280         sb->s_blocksize = mp->m_sb.sb_blocksize;
9281 diff -NurpP --minimal linux-2.6.25.11/fs/xfs/linux-2.6/xfs_vnode.c linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/linux-2.6/xfs_vnode.c
9282 --- linux-2.6.25.11/fs/xfs/linux-2.6/xfs_vnode.c        2008-04-17 12:05:41.000000000 -0400
9283 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/linux-2.6/xfs_vnode.c  2008-07-08 16:54:48.000000000 -0400
9284 @@ -105,6 +105,7 @@ vn_revalidate(
9285         inode->i_mode       = ip->i_d.di_mode;
9286         inode->i_uid        = ip->i_d.di_uid;
9287         inode->i_gid        = ip->i_d.di_gid;
9288 +       inode->i_tag        = ip->i_d.di_tag;
9289         inode->i_mtime.tv_sec = ip->i_d.di_mtime.t_sec;
9290         inode->i_mtime.tv_nsec = ip->i_d.di_mtime.t_nsec;
9291         inode->i_ctime.tv_sec = ip->i_d.di_ctime.t_sec;
9292 @@ -115,6 +116,10 @@ vn_revalidate(
9293                 inode->i_flags |= S_IMMUTABLE;
9294         else
9295                 inode->i_flags &= ~S_IMMUTABLE;
9296 +       if (xflags & XFS_XFLAG_IXUNLINK)
9297 +               inode->i_flags |= S_IXUNLINK;
9298 +       else
9299 +               inode->i_flags &= ~S_IXUNLINK;
9300         if (xflags & XFS_XFLAG_APPEND)
9301                 inode->i_flags |= S_APPEND;
9302         else
9303 @@ -127,6 +132,15 @@ vn_revalidate(
9304                 inode->i_flags |= S_NOATIME;
9305         else
9306                 inode->i_flags &= ~S_NOATIME;
9307 +
9308 +       if (xflags & XFS_XFLAG_BARRIER)
9309 +               inode->i_vflags |= V_BARRIER;
9310 +       else
9311 +               inode->i_vflags &= ~V_BARRIER;
9312 +       if (xflags & XFS_XFLAG_COW)
9313 +               inode->i_vflags |= V_COW;
9314 +       else
9315 +               inode->i_vflags &= ~V_COW;
9316         xfs_iunlock(ip, XFS_ILOCK_SHARED);
9317
9318         xfs_iflags_clear(ip, XFS_IMODIFIED);
9319 diff -NurpP --minimal linux-2.6.25.11/fs/xfs/linux-2.6/xfs_vnode.h linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/linux-2.6/xfs_vnode.h
9320 --- linux-2.6.25.11/fs/xfs/linux-2.6/xfs_vnode.h        2008-04-17 12:05:41.000000000 -0400
9321 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/linux-2.6/xfs_vnode.h  2008-04-19 15:14:52.000000000 -0400
9322 @@ -99,6 +99,7 @@ typedef struct bhv_vattr {
9323         xfs_nlink_t     va_nlink;       /* number of references to file */
9324         uid_t           va_uid;         /* owner user id */
9325         gid_t           va_gid;         /* owner group id */
9326 +       tag_t           va_tag;         /* owner group id */
9327         xfs_ino_t       va_nodeid;      /* file id */
9328         xfs_off_t       va_size;        /* file size in bytes */
9329         u_long          va_blocksize;   /* blocksize preferred for i/o */
9330 @@ -147,13 +148,15 @@ typedef struct bhv_vattr {
9331  #define XFS_AT_PROJID          0x04000000
9332  #define XFS_AT_SIZE_NOPERM     0x08000000
9333  #define XFS_AT_GENCOUNT                0x10000000
9334 +#define XFS_AT_TAG             0x20000000
9335
9336  #define XFS_AT_ALL     (XFS_AT_TYPE|XFS_AT_MODE|XFS_AT_UID|XFS_AT_GID|\
9337                 XFS_AT_FSID|XFS_AT_NODEID|XFS_AT_NLINK|XFS_AT_SIZE|\
9338                 XFS_AT_ATIME|XFS_AT_MTIME|XFS_AT_CTIME|XFS_AT_RDEV|\
9339                 XFS_AT_BLKSIZE|XFS_AT_NBLOCKS|XFS_AT_VCODE|XFS_AT_MAC|\
9340                 XFS_AT_ACL|XFS_AT_CAP|XFS_AT_INF|XFS_AT_XFLAGS|XFS_AT_EXTSIZE|\
9341 -               XFS_AT_NEXTENTS|XFS_AT_ANEXTENTS|XFS_AT_PROJID|XFS_AT_GENCOUNT)
9342 +               XFS_AT_NEXTENTS|XFS_AT_ANEXTENTS|XFS_AT_PROJID|XFS_AT_GENCOUNT\
9343 +               XFS_AT_TAG)
9344
9345  #define XFS_AT_STAT    (XFS_AT_TYPE|XFS_AT_MODE|XFS_AT_UID|XFS_AT_GID|\
9346                 XFS_AT_FSID|XFS_AT_NODEID|XFS_AT_NLINK|XFS_AT_SIZE|\
9347 diff -NurpP --minimal linux-2.6.25.11/fs/xfs/quota/xfs_qm_syscalls.c linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/quota/xfs_qm_syscalls.c
9348 --- linux-2.6.25.11/fs/xfs/quota/xfs_qm_syscalls.c      2008-04-17 12:05:42.000000000 -0400
9349 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/quota/xfs_qm_syscalls.c        2008-04-19 15:14:52.000000000 -0400
9350 @@ -17,6 +17,7 @@
9351   */
9352
9353  #include <linux/capability.h>
9354 +#include <linux/vs_context.h>
9355
9356  #include "xfs.h"
9357  #include "xfs_fs.h"
9358 @@ -205,7 +206,7 @@ xfs_qm_scall_quotaoff(
9359         xfs_qoff_logitem_t      *qoffstart;
9360         int                     nculprits;
9361
9362 -       if (!force && !capable(CAP_SYS_ADMIN))
9363 +       if (!force && !vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL))
9364                 return XFS_ERROR(EPERM);
9365         /*
9366          * No file system can have quotas enabled on disk but not in core.
9367 @@ -374,7 +375,7 @@ xfs_qm_scall_trunc_qfiles(
9368         int             error;
9369         xfs_inode_t     *qip;
9370
9371 -       if (!capable(CAP_SYS_ADMIN))
9372 +       if (!vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL))
9373                 return XFS_ERROR(EPERM);
9374         error = 0;
9375         if (!xfs_sb_version_hasquota(&mp->m_sb) || flags == 0) {
9376 @@ -418,7 +419,7 @@ xfs_qm_scall_quotaon(
9377         uint            accflags;
9378         __int64_t       sbflags;
9379
9380 -       if (!capable(CAP_SYS_ADMIN))
9381 +       if (!vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL))
9382                 return XFS_ERROR(EPERM);
9383
9384         flags &= (XFS_ALL_QUOTA_ACCT | XFS_ALL_QUOTA_ENFD);
9385 @@ -587,7 +588,7 @@ xfs_qm_scall_setqlim(
9386         int                     error;
9387         xfs_qcnt_t              hard, soft;
9388
9389 -       if (!capable(CAP_SYS_ADMIN))
9390 +       if (!vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL))
9391                 return XFS_ERROR(EPERM);
9392
9393         if ((newlim->d_fieldmask &
9394 diff -NurpP --minimal linux-2.6.25.11/fs/xfs/xfs_clnt.h linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/xfs_clnt.h
9395 --- linux-2.6.25.11/fs/xfs/xfs_clnt.h   2008-04-17 12:05:42.000000000 -0400
9396 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/xfs_clnt.h     2008-07-11 01:28:55.000000000 -0400
9397 @@ -100,5 +100,6 @@ struct xfs_mount_args {
9398                                                  * I/O size in stat(2) */
9399  #define XFSMNT2_FILESTREAMS    0x00000002      /* enable the filestreams
9400                                                  * allocator */
9401 +#define XFSMNT2_TAGGED         0x10000000      /* context tagging */
9402
9403  #endif /* __XFS_CLNT_H__ */
9404 diff -NurpP --minimal linux-2.6.25.11/fs/xfs/xfs_dinode.h linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/xfs_dinode.h
9405 --- linux-2.6.25.11/fs/xfs/xfs_dinode.h 2008-04-17 12:05:42.000000000 -0400
9406 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/xfs_dinode.h   2008-07-11 00:30:17.000000000 -0400
9407 @@ -53,7 +53,9 @@ typedef struct xfs_dinode_core {
9408         __be32          di_gid;         /* owner's group id */
9409         __be32          di_nlink;       /* number of links to file */
9410         __be16          di_projid;      /* owner's project id */
9411 -       __u8            di_pad[8];      /* unused, zeroed space */
9412 +       __be16          di_tag;         /* context tagging */
9413 +       __be16          di_vflags;      /* vserver specific flags */
9414 +       __u8            di_pad[4];      /* unused, zeroed space */
9415         __be16          di_flushiter;   /* incremented on flush */
9416         xfs_timestamp_t di_atime;       /* time last accessed */
9417         xfs_timestamp_t di_mtime;       /* time last modified */
9418 @@ -136,7 +138,9 @@ typedef struct xfs_dinode
9419  #define        XFS_DI_NEXT_UNLINKED    0x1000000
9420  #define        XFS_DI_U                0x2000000
9421  #define        XFS_DI_A                0x4000000
9422 -#define        XFS_DI_NUM_BITS         27
9423 +#define        XFS_DI_TAG              0x8000000
9424 +#define        XFS_DI_VFLAGS           0x10000000
9425 +#define        XFS_DI_NUM_BITS         29
9426  #define        XFS_DI_ALL_BITS         ((1 << XFS_DI_NUM_BITS) - 1)
9427  #define        XFS_DI_CORE_BITS        (XFS_DI_ALL_BITS & ~(XFS_DI_U|XFS_DI_A))
9428
9429 @@ -223,6 +227,8 @@ typedef enum xfs_dinode_fmt
9430  #define XFS_DIFLAG_EXTSZINHERIT_BIT 12 /* inherit inode extent size */
9431  #define XFS_DIFLAG_NODEFRAG_BIT     13 /* do not reorganize/defragment */
9432  #define XFS_DIFLAG_FILESTREAM_BIT   14  /* use filestream allocator */
9433 +#define XFS_DIFLAG_IXUNLINK_BIT     15 /* Immutable inver on unlink */
9434 +
9435  #define XFS_DIFLAG_REALTIME      (1 << XFS_DIFLAG_REALTIME_BIT)
9436  #define XFS_DIFLAG_PREALLOC      (1 << XFS_DIFLAG_PREALLOC_BIT)
9437  #define XFS_DIFLAG_NEWRTBM       (1 << XFS_DIFLAG_NEWRTBM_BIT)
9438 @@ -238,6 +244,7 @@ typedef enum xfs_dinode_fmt
9439  #define XFS_DIFLAG_EXTSZINHERIT  (1 << XFS_DIFLAG_EXTSZINHERIT_BIT)
9440  #define XFS_DIFLAG_NODEFRAG      (1 << XFS_DIFLAG_NODEFRAG_BIT)
9441  #define XFS_DIFLAG_FILESTREAM    (1 << XFS_DIFLAG_FILESTREAM_BIT)
9442 +#define XFS_DIFLAG_IXUNLINK      (1 << XFS_DIFLAG_IXUNLINK_BIT)
9443
9444  #ifdef CONFIG_XFS_RT
9445  #define XFS_IS_REALTIME_INODE(ip) ((ip)->i_d.di_flags & XFS_DIFLAG_REALTIME)
9446 @@ -250,6 +257,10 @@ typedef enum xfs_dinode_fmt
9447          XFS_DIFLAG_IMMUTABLE | XFS_DIFLAG_APPEND | XFS_DIFLAG_SYNC | \
9448          XFS_DIFLAG_NOATIME | XFS_DIFLAG_NODUMP | XFS_DIFLAG_RTINHERIT | \
9449          XFS_DIFLAG_PROJINHERIT | XFS_DIFLAG_NOSYMLINKS | XFS_DIFLAG_EXTSIZE | \
9450 -        XFS_DIFLAG_EXTSZINHERIT | XFS_DIFLAG_NODEFRAG | XFS_DIFLAG_FILESTREAM)
9451 +        XFS_DIFLAG_EXTSZINHERIT | XFS_DIFLAG_NODEFRAG | XFS_DIFLAG_FILESTREAM | \
9452 +        XFS_DIFLAG_IXUNLINK)
9453 +
9454 +#define XFS_DIVFLAG_BARRIER    0x01
9455 +#define XFS_DIVFLAG_COW                0x02
9456
9457  #endif /* __XFS_DINODE_H__ */
9458 diff -NurpP --minimal linux-2.6.25.11/fs/xfs/xfs_fs.h linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/xfs_fs.h
9459 --- linux-2.6.25.11/fs/xfs/xfs_fs.h     2008-04-17 12:05:42.000000000 -0400
9460 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/xfs_fs.h       2008-07-08 16:11:02.000000000 -0400
9461 @@ -67,6 +67,9 @@ struct fsxattr {
9462  #define XFS_XFLAG_EXTSZINHERIT 0x00001000      /* inherit inode extent size */
9463  #define XFS_XFLAG_NODEFRAG     0x00002000      /* do not defragment */
9464  #define XFS_XFLAG_FILESTREAM   0x00004000      /* use filestream allocator */
9465 +#define XFS_XFLAG_IXUNLINK     0x00008000      /* immutable invert on unlink */
9466 +#define XFS_XFLAG_BARRIER      0x10000000      /* chroot() barrier */
9467 +#define XFS_XFLAG_COW          0x20000000      /* copy on write mark */
9468  #define XFS_XFLAG_HASATTR      0x80000000      /* no DIFLAG for this   */
9469
9470  /*
9471 @@ -296,7 +299,8 @@ typedef struct xfs_bstat {
9472         __s32           bs_extents;     /* number of extents            */
9473         __u32           bs_gen;         /* generation count             */
9474         __u16           bs_projid;      /* project id                   */
9475 -       unsigned char   bs_pad[14];     /* pad space, unused            */
9476 +       __u16           bs_tag;         /* context tagging              */
9477 +       unsigned char   bs_pad[12];     /* pad space, unused            */
9478         __u32           bs_dmevmask;    /* DMIG event mask              */
9479         __u16           bs_dmstate;     /* DMIG state info              */
9480         __u16           bs_aextents;    /* attribute number of extents  */
9481 diff -NurpP --minimal linux-2.6.25.11/fs/xfs/xfs_ialloc.c linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/xfs_ialloc.c
9482 --- linux-2.6.25.11/fs/xfs/xfs_ialloc.c 2008-04-17 12:05:42.000000000 -0400
9483 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/xfs_ialloc.c   2008-07-11 00:30:58.000000000 -0400
9484 @@ -84,6 +84,8 @@ xfs_ialloc_log_di(
9485                 offsetof(xfs_dinode_t, di_next_unlinked),
9486                 offsetof(xfs_dinode_t, di_u),
9487                 offsetof(xfs_dinode_t, di_a),
9488 +               offsetof(xfs_dinode_core_t, di_tag),
9489 +               offsetof(xfs_dinode_core_t, di_vflags),
9490                 sizeof(xfs_dinode_t)
9491         };
9492
9493 diff -NurpP --minimal linux-2.6.25.11/fs/xfs/xfs_inode.c linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/xfs_inode.c
9494 --- linux-2.6.25.11/fs/xfs/xfs_inode.c  2008-04-17 12:05:42.000000000 -0400
9495 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/xfs_inode.c    2008-07-11 01:49:20.000000000 -0400
9496 @@ -219,6 +219,7 @@ xfs_inotobp(
9497         return 0;
9498  }
9499
9500 +#include <linux/vs_tag.h>
9501
9502  /*
9503   * This routine is called to map an inode to the buffer containing
9504 @@ -714,15 +715,25 @@ xfs_iformat_btree(
9505  void
9506  xfs_dinode_from_disk(
9507         xfs_icdinode_t          *to,
9508 -       xfs_dinode_core_t       *from)
9509 +       xfs_dinode_core_t       *from,
9510 +       int tagged)
9511  {
9512 +       uint32_t uid, gid, tag;
9513 +
9514         to->di_magic = be16_to_cpu(from->di_magic);
9515         to->di_mode = be16_to_cpu(from->di_mode);
9516         to->di_version = from ->di_version;
9517         to->di_format = from->di_format;
9518         to->di_onlink = be16_to_cpu(from->di_onlink);
9519 -       to->di_uid = be32_to_cpu(from->di_uid);
9520 -       to->di_gid = be32_to_cpu(from->di_gid);
9521 +
9522 +       uid = be32_to_cpu(from->di_uid);
9523 +       gid = be32_to_cpu(from->di_gid);
9524 +       tag = be16_to_cpu(from->di_tag);
9525 +
9526 +       to->di_uid = INOTAG_UID(tagged, uid, gid);
9527 +       to->di_gid = INOTAG_GID(tagged, uid, gid);
9528 +       to->di_tag = INOTAG_TAG(tagged, uid, gid, tag);
9529 +
9530         to->di_nlink = be32_to_cpu(from->di_nlink);
9531         to->di_projid = be16_to_cpu(from->di_projid);
9532         memcpy(to->di_pad, from->di_pad, sizeof(to->di_pad));
9533 @@ -743,21 +754,26 @@ xfs_dinode_from_disk(
9534         to->di_dmevmask = be32_to_cpu(from->di_dmevmask);
9535         to->di_dmstate  = be16_to_cpu(from->di_dmstate);
9536         to->di_flags    = be16_to_cpu(from->di_flags);
9537 +       to->di_vflags   = be16_to_cpu(from->di_vflags);
9538         to->di_gen      = be32_to_cpu(from->di_gen);
9539  }
9540
9541  void
9542  xfs_dinode_to_disk(
9543         xfs_dinode_core_t       *to,
9544 -       xfs_icdinode_t          *from)
9545 +       xfs_icdinode_t          *from,
9546 +       int tagged)
9547  {
9548         to->di_magic = cpu_to_be16(from->di_magic);
9549         to->di_mode = cpu_to_be16(from->di_mode);
9550         to->di_version = from ->di_version;
9551         to->di_format = from->di_format;
9552         to->di_onlink = cpu_to_be16(from->di_onlink);
9553 -       to->di_uid = cpu_to_be32(from->di_uid);
9554 -       to->di_gid = cpu_to_be32(from->di_gid);
9555 +
9556 +       to->di_uid = cpu_to_be32(TAGINO_UID(tagged, from->di_uid, from->di_tag));
9557 +       to->di_gid = cpu_to_be32(TAGINO_GID(tagged, from->di_gid, from->di_tag));
9558 +       to->di_tag = cpu_to_be16(TAGINO_TAG(tagged, from->di_tag));
9559 +
9560         to->di_nlink = cpu_to_be32(from->di_nlink);
9561         to->di_projid = cpu_to_be16(from->di_projid);
9562         memcpy(to->di_pad, from->di_pad, sizeof(to->di_pad));
9563 @@ -778,12 +794,14 @@ xfs_dinode_to_disk(
9564         to->di_dmevmask = cpu_to_be32(from->di_dmevmask);
9565         to->di_dmstate = cpu_to_be16(from->di_dmstate);
9566         to->di_flags = cpu_to_be16(from->di_flags);
9567 +       to->di_vflags = cpu_to_be16(from->di_vflags);
9568         to->di_gen = cpu_to_be32(from->di_gen);
9569  }
9570
9571  STATIC uint
9572  _xfs_dic2xflags(
9573 -       __uint16_t              di_flags)
9574 +       __uint16_t              di_flags,
9575 +       __uint16_t              di_vflags)
9576  {
9577         uint                    flags = 0;
9578
9579 @@ -794,6 +812,8 @@ _xfs_dic2xflags(
9580                         flags |= XFS_XFLAG_PREALLOC;
9581                 if (di_flags & XFS_DIFLAG_IMMUTABLE)
9582                         flags |= XFS_XFLAG_IMMUTABLE;
9583 +               if (di_flags & XFS_DIFLAG_IXUNLINK)
9584 +                       flags |= XFS_XFLAG_IXUNLINK;
9585                 if (di_flags & XFS_DIFLAG_APPEND)
9586                         flags |= XFS_XFLAG_APPEND;
9587                 if (di_flags & XFS_DIFLAG_SYNC)
9588 @@ -817,7 +837,10 @@ _xfs_dic2xflags(
9589                 if (di_flags & XFS_DIFLAG_FILESTREAM)
9590                         flags |= XFS_XFLAG_FILESTREAM;
9591         }
9592 -
9593 +       if (di_vflags & XFS_DIVFLAG_BARRIER)
9594 +               flags |= XFS_XFLAG_BARRIER;
9595 +       if (di_vflags & XFS_DIVFLAG_COW)
9596 +               flags |= XFS_XFLAG_COW;
9597         return flags;
9598  }
9599
9600 @@ -827,7 +850,7 @@ xfs_ip2xflags(
9601  {
9602         xfs_icdinode_t          *dic = &ip->i_d;
9603
9604 -       return _xfs_dic2xflags(dic->di_flags) |
9605 +       return _xfs_dic2xflags(dic->di_flags, dic->di_vflags) |
9606                                 (XFS_IFORK_Q(ip) ? XFS_XFLAG_HASATTR : 0);
9607  }
9608
9609 @@ -837,7 +860,7 @@ xfs_dic2xflags(
9610  {
9611         xfs_dinode_core_t       *dic = &dip->di_core;
9612
9613 -       return _xfs_dic2xflags(be16_to_cpu(dic->di_flags)) |
9614 +       return _xfs_dic2xflags(be16_to_cpu(dic->di_flags), be16_to_cpu(dic->di_vflags)) |
9615                                 (XFS_DFORK_Q(dip) ? XFS_XFLAG_HASATTR : 0);
9616  }
9617
9618 @@ -932,7 +955,8 @@ xfs_iread(
9619          * Otherwise, just get the truly permanent information.
9620          */
9621         if (dip->di_core.di_mode) {
9622 -               xfs_dinode_from_disk(&ip->i_d, &dip->di_core);
9623 +               xfs_dinode_from_disk(&ip->i_d, &dip->di_core,
9624 +                       mp->m_flags & XFS_MOUNT_TAGGED);
9625                 error = xfs_iformat(ip, dip);
9626                 if (error)  {
9627                         kmem_zone_free(xfs_inode_zone, ip);
9628 @@ -1138,6 +1162,7 @@ xfs_ialloc(
9629         ASSERT(ip->i_d.di_nlink == nlink);
9630         ip->i_d.di_uid = current_fsuid(cr);
9631         ip->i_d.di_gid = current_fsgid(cr);
9632 +       ip->i_d.di_tag = current_fstag(cr, vp);
9633         ip->i_d.di_projid = prid;
9634         memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad));
9635
9636 @@ -1192,6 +1217,7 @@ xfs_ialloc(
9637         ip->i_d.di_dmevmask = 0;
9638         ip->i_d.di_dmstate = 0;
9639         ip->i_d.di_flags = 0;
9640 +       ip->i_d.di_vflags = 0;
9641         flags = XFS_ILOG_CORE;
9642         switch (mode & S_IFMT) {
9643         case S_IFIFO:
9644 @@ -1944,8 +1970,8 @@ xfs_iunlink(
9645         agi_ok =
9646                 be32_to_cpu(agi->agi_magicnum) == XFS_AGI_MAGIC &&
9647                 XFS_AGI_GOOD_VERSION(be32_to_cpu(agi->agi_versionnum));
9648 -       if (unlikely(XFS_TEST_ERROR(!agi_ok, mp, XFS_ERRTAG_IUNLINK,
9649 -                       XFS_RANDOM_IUNLINK))) {
9650 +       if (unlikely(XFS_TEST_ERROR(!agi_ok, mp, XFS_ERRTAG_IXUNLINK,
9651 +                       XFS_RANDOM_IXUNLINK))) {
9652                 XFS_CORRUPTION_ERROR("xfs_iunlink", XFS_ERRLEVEL_LOW, mp, agi);
9653                 xfs_trans_brelse(tp, agibp);
9654                 return XFS_ERROR(EFSCORRUPTED);
9655 @@ -2046,8 +2072,8 @@ xfs_iunlink_remove(
9656         agi_ok =
9657                 be32_to_cpu(agi->agi_magicnum) == XFS_AGI_MAGIC &&
9658                 XFS_AGI_GOOD_VERSION(be32_to_cpu(agi->agi_versionnum));
9659 -       if (unlikely(XFS_TEST_ERROR(!agi_ok, mp, XFS_ERRTAG_IUNLINK_REMOVE,
9660 -                       XFS_RANDOM_IUNLINK_REMOVE))) {
9661 +       if (unlikely(XFS_TEST_ERROR(!agi_ok, mp, XFS_ERRTAG_IXUNLINK_REMOVE,
9662 +                       XFS_RANDOM_IXUNLINK_REMOVE))) {
9663                 XFS_CORRUPTION_ERROR("xfs_iunlink_remove", XFS_ERRLEVEL_LOW,
9664                                      mp, agi);
9665                 xfs_trans_brelse(tp, agibp);
9666 @@ -2386,6 +2412,7 @@ xfs_ifree(
9667         }
9668         ip->i_d.di_mode = 0;            /* mark incore inode as free */
9669         ip->i_d.di_flags = 0;
9670 +       ip->i_d.di_vflags = 0;
9671         ip->i_d.di_dmevmask = 0;
9672         ip->i_d.di_forkoff = 0;         /* mark the attr fork not in use */
9673         ip->i_df.if_ext_max =
9674 @@ -3421,7 +3448,8 @@ xfs_iflush_int(
9675          * because if the inode is dirty at all the core must
9676          * be.
9677          */
9678 -       xfs_dinode_to_disk(&dip->di_core, &ip->i_d);
9679 +       xfs_dinode_to_disk(&dip->di_core, &ip->i_d,
9680 +               mp->m_flags & XFS_MOUNT_TAGGED);
9681
9682         /* Wrap, we never let the log put out DI_MAX_FLUSH */
9683         if (ip->i_d.di_flushiter == DI_MAX_FLUSH)
9684 diff -NurpP --minimal linux-2.6.25.11/fs/xfs/xfs_inode.h linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/xfs_inode.h
9685 --- linux-2.6.25.11/fs/xfs/xfs_inode.h  2008-04-17 12:05:42.000000000 -0400
9686 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/xfs_inode.h    2008-07-11 01:44:30.000000000 -0400
9687 @@ -187,7 +187,9 @@ typedef struct xfs_icdinode {
9688         __uint32_t      di_gid;         /* owner's group id */
9689         __uint32_t      di_nlink;       /* number of links to file */
9690         __uint16_t      di_projid;      /* owner's project id */
9691 -       __uint8_t       di_pad[8];      /* unused, zeroed space */
9692 +       __uint16_t      di_tag;         /* context tagging */
9693 +       __uint16_t      di_vflags;      /* vserver specific flags */
9694 +       __uint8_t       di_pad[4];      /* unused, zeroed space */
9695         __uint16_t      di_flushiter;   /* incremented on flush */
9696         xfs_ictimestamp_t di_atime;     /* time last accessed */
9697         xfs_ictimestamp_t di_mtime;     /* time last modified */
9698 @@ -523,9 +525,9 @@ int         xfs_ialloc(struct xfs_trans *, xfs_
9699                            xfs_nlink_t, xfs_dev_t, struct cred *, xfs_prid_t,
9700                            int, struct xfs_buf **, boolean_t *, xfs_inode_t **);
9701  void           xfs_dinode_from_disk(struct xfs_icdinode *,
9702 -                                    struct xfs_dinode_core *);
9703 +                                    struct xfs_dinode_core *, int tagged);
9704  void           xfs_dinode_to_disk(struct xfs_dinode_core *,
9705 -                                  struct xfs_icdinode *);
9706 +                                  struct xfs_icdinode *, int tagged);
9707
9708  uint           xfs_ip2xflags(struct xfs_inode *);
9709  uint           xfs_dic2xflags(struct xfs_dinode *);
9710 diff -NurpP --minimal linux-2.6.25.11/fs/xfs/xfs_itable.c linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/xfs_itable.c
9711 --- linux-2.6.25.11/fs/xfs/xfs_itable.c 2008-04-17 12:05:42.000000000 -0400
9712 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/xfs_itable.c   2008-04-19 15:14:52.000000000 -0400
9713 @@ -89,6 +89,7 @@ xfs_bulkstat_one_iget(
9714         buf->bs_mode = dic->di_mode;
9715         buf->bs_uid = dic->di_uid;
9716         buf->bs_gid = dic->di_gid;
9717 +       buf->bs_tag = dic->di_tag;
9718         buf->bs_size = dic->di_size;
9719         vn_atime_to_bstime(vp, &buf->bs_atime);
9720         buf->bs_mtime.tv_sec = dic->di_mtime.t_sec;
9721 diff -NurpP --minimal linux-2.6.25.11/fs/xfs/xfs_log_recover.c linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/xfs_log_recover.c
9722 --- linux-2.6.25.11/fs/xfs/xfs_log_recover.c    2008-04-17 12:05:43.000000000 -0400
9723 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/xfs_log_recover.c      2008-07-11 01:52:59.000000000 -0400
9724 @@ -2410,7 +2410,8 @@ xlog_recover_do_inode_trans(
9725
9726         /* The core is in in-core format */
9727         xfs_dinode_to_disk(&dip->di_core,
9728 -               (xfs_icdinode_t *)item->ri_buf[1].i_addr);
9729 +               (xfs_icdinode_t *)item->ri_buf[1].i_addr,
9730 +               mp->m_flags & XFS_MOUNT_TAGGED);
9731
9732         /* the rest is in on-disk format */
9733         if (item->ri_buf[1].i_len > sizeof(xfs_dinode_core_t)) {
9734 diff -NurpP --minimal linux-2.6.25.11/fs/xfs/xfs_mount.h linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/xfs_mount.h
9735 --- linux-2.6.25.11/fs/xfs/xfs_mount.h  2008-04-17 12:05:43.000000000 -0400
9736 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/xfs_mount.h    2008-04-19 15:14:52.000000000 -0400
9737 @@ -378,6 +378,7 @@ typedef struct xfs_mount {
9738  #define XFS_MOUNT_FILESTREAMS  (1ULL << 24)    /* enable the filestreams
9739                                                    allocator */
9740
9741 +#define XFS_MOUNT_TAGGED       (1ULL << 31)    /* context tagging */
9742
9743  /*
9744   * Default minimum read and write sizes.
9745 diff -NurpP --minimal linux-2.6.25.11/fs/xfs/xfs_vfsops.c linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/xfs_vfsops.c
9746 --- linux-2.6.25.11/fs/xfs/xfs_vfsops.c 2008-04-17 12:05:43.000000000 -0400
9747 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/xfs_vfsops.c   2008-04-26 09:52:49.000000000 -0400
9748 @@ -290,6 +290,8 @@ xfs_start_flags(
9749
9750         if (ap->flags2 & XFSMNT2_COMPAT_IOSIZE)
9751                 mp->m_flags |= XFS_MOUNT_COMPAT_IOSIZE;
9752 +       if (ap->flags2 & XFSMNT2_TAGGED)
9753 +               mp->m_flags |= XFS_MOUNT_TAGGED;
9754
9755         /*
9756          * no recovery flag requires a read-only mount
9757 @@ -402,7 +404,6 @@ xfs_finish_flags(
9758                 if (ap->flags & XFSMNT_PQUOTAENF)
9759                         mp->m_qflags |= XFS_OQUOTA_ENFD;
9760         }
9761 -
9762         return 0;
9763  }
9764
9765 diff -NurpP --minimal linux-2.6.25.11/fs/xfs/xfs_vnodeops.c linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/xfs_vnodeops.c
9766 --- linux-2.6.25.11/fs/xfs/xfs_vnodeops.c       2008-04-17 12:05:43.000000000 -0400
9767 +++ linux-2.6.25.11-vs2.3.0.34.14/fs/xfs/xfs_vnodeops.c 2008-07-08 16:08:14.000000000 -0400
9768 @@ -122,6 +122,7 @@ xfs_getattr(
9769         vap->va_mode = ip->i_d.di_mode;
9770         vap->va_uid = ip->i_d.di_uid;
9771         vap->va_gid = ip->i_d.di_gid;
9772 +       vap->va_tag = ip->i_d.di_tag;
9773         vap->va_projid = ip->i_d.di_projid;
9774
9775         /*
9776 @@ -221,6 +222,7 @@ xfs_setattr(
9777         uint                    commit_flags=0;
9778         uid_t                   uid=0, iuid=0;
9779         gid_t                   gid=0, igid=0;
9780 +       tag_t                   tag=0, itag=0;
9781         int                     timeflags = 0;
9782         xfs_prid_t              projid=0, iprojid=0;
9783         int                     mandlock_before, mandlock_after;
9784 @@ -272,6 +274,7 @@ xfs_setattr(
9785             (mask & (XFS_AT_UID|XFS_AT_GID|XFS_AT_PROJID))) {
9786                 uint    qflags = 0;
9787
9788 +               /* TODO: handle tagging? */
9789                 if ((mask & XFS_AT_UID) && XFS_IS_UQUOTA_ON(mp)) {
9790                         uid = vap->va_uid;
9791                         qflags |= XFS_QMOPT_UQUOTA;
9792 @@ -351,6 +354,8 @@ xfs_setattr(
9793         if (mask &
9794             (XFS_AT_MODE|XFS_AT_XFLAGS|XFS_AT_EXTSIZE|XFS_AT_UID|
9795              XFS_AT_GID|XFS_AT_PROJID)) {
9796 +               /* TODO: handle tagging? */
9797 +
9798                 /*
9799                  * CAP_FOWNER overrides the following restrictions:
9800                  *
9801 @@ -399,7 +404,7 @@ xfs_setattr(
9802          * and can change the group id only to a group of which he
9803          * or she is a member.
9804          */
9805 -       if (mask & (XFS_AT_UID|XFS_AT_GID|XFS_AT_PROJID)) {
9806 +       if (mask & (XFS_AT_UID|XFS_AT_GID|XFS_AT_TAG|XFS_AT_PROJID)) {
9807                 /*
9808                  * These IDs could have changed since we last looked at them.
9809                  * But, we're assured that if the ownership did change
9810 @@ -407,10 +412,12 @@ xfs_setattr(
9811                  * would have changed also.
9812                  */
9813                 iuid = ip->i_d.di_uid;
9814 -               iprojid = ip->i_d.di_projid;
9815                 igid = ip->i_d.di_gid;
9816 -               gid = (mask & XFS_AT_GID) ? vap->va_gid : igid;
9817 +               itag = ip->i_d.di_tag;
9818 +               iprojid = ip->i_d.di_projid;
9819                 uid = (mask & XFS_AT_UID) ? vap->va_uid : iuid;
9820 +               gid = (mask & XFS_AT_GID) ? vap->va_gid : igid;
9821 +               tag = (mask & XFS_AT_TAG) ? vap->va_tag : itag;
9822                 projid = (mask & XFS_AT_PROJID) ? (xfs_prid_t)vap->va_projid :
9823                          iprojid;
9824
9825 @@ -438,6 +445,7 @@ xfs_setattr(
9826                 if ((XFS_IS_UQUOTA_ON(mp) && iuid != uid) ||
9827                     (XFS_IS_PQUOTA_ON(mp) && iprojid != projid) ||
9828                     (XFS_IS_GQUOTA_ON(mp) && igid != gid)) {
9829 +                       /* TODO: handle tagging? */
9830                         ASSERT(tp);
9831                         code = XFS_QM_DQVOPCHOWNRESV(mp, tp, ip, udqp, gdqp,
9832                                                 capable(CAP_FOWNER) ?
9833 @@ -552,7 +560,7 @@ xfs_setattr(
9834                  */
9835                 if ((mask & XFS_AT_XFLAGS) &&
9836                     (ip->i_d.di_flags &
9837 -                               (XFS_DIFLAG_IMMUTABLE|XFS_DIFLAG_APPEND) ||
9838 +                               (XFS_DIFLAG_IMMUTABLE | XFS_DIFLAG_APPEND) ||
9839                      (vap->va_xflags &
9840                                 (XFS_XFLAG_IMMUTABLE | XFS_XFLAG_APPEND))) &&
9841                     !capable(CAP_LINUX_IMMUTABLE)) {
9842 @@ -686,7 +694,7 @@ xfs_setattr(
9843          * and can change the group id only to a group of which he
9844          * or she is a member.
9845          */
9846 -       if (mask & (XFS_AT_UID|XFS_AT_GID|XFS_AT_PROJID)) {
9847 +       if (mask & (XFS_AT_UID|XFS_AT_GID|XFS_AT_TAG|XFS_AT_PROJID)) {
9848                 /*
9849                  * CAP_FSETID overrides the following restrictions:
9850                  *
9851 @@ -702,6 +710,9 @@ xfs_setattr(
9852                  * Change the ownerships and register quota modifications
9853                  * in the transaction.
9854                  */
9855 +               if (itag != tag) {
9856 +                       ip->i_d.di_tag = tag;
9857 +               }
9858                 if (iuid != uid) {
9859                         if (XFS_IS_UQUOTA_ON(mp)) {
9860                                 ASSERT(mask & XFS_AT_UID);
9861 @@ -777,11 +788,14 @@ xfs_setattr(
9862                 }
9863                 if (mask & XFS_AT_XFLAGS) {
9864                         uint    di_flags;
9865 +                       uint    di_vflags = 0;
9866
9867                         /* can't set PREALLOC this way, just preserve it */
9868                         di_flags = (ip->i_d.di_flags & XFS_DIFLAG_PREALLOC);
9869                         if (vap->va_xflags & XFS_XFLAG_IMMUTABLE)
9870                                 di_flags |= XFS_DIFLAG_IMMUTABLE;
9871 +                       if (vap->va_xflags & XFS_XFLAG_IXUNLINK)
9872 +                               di_flags |= XFS_DIFLAG_IXUNLINK;
9873                         if (vap->va_xflags & XFS_XFLAG_APPEND)
9874                                 di_flags |= XFS_DIFLAG_APPEND;
9875                         if (vap->va_xflags & XFS_XFLAG_SYNC)
9876 @@ -810,6 +824,11 @@ xfs_setattr(
9877                                         di_flags |= XFS_DIFLAG_EXTSIZE;
9878                         }
9879                         ip->i_d.di_flags = di_flags;
9880 +                       if (vap->va_xflags & XFS_XFLAG_BARRIER)
9881 +                               di_vflags |= XFS_DIVFLAG_BARRIER;
9882 +                       if (vap->va_xflags & XFS_XFLAG_COW)
9883 +                               di_vflags |= XFS_DIVFLAG_COW;
9884 +                       ip->i_d.di_vflags = di_vflags;
9885                 }
9886                 xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
9887                 timeflags |= XFS_ICHGTIME_CHG;
9888 diff -NurpP --minimal linux-2.6.25.11/include/asm-arm/tlb.h linux-2.6.25.11-vs2.3.0.34.14/include/asm-arm/tlb.h
9889 --- linux-2.6.25.11/include/asm-arm/tlb.h       2008-04-17 12:05:43.000000000 -0400
9890 +++ linux-2.6.25.11-vs2.3.0.34.14/include/asm-arm/tlb.h 2008-04-19 15:14:52.000000000 -0400
9891 @@ -28,6 +28,7 @@
9892  #else /* !CONFIG_MMU */
9893
9894  #include <asm/pgalloc.h>
9895 +#include <linux/vs_memory.h>
9896
9897  /*
9898   * TLB handling.  This allows us to remove pages from the page
9899 diff -NurpP --minimal linux-2.6.25.11/include/asm-blackfin/unistd.h linux-2.6.25.11-vs2.3.0.34.14/include/asm-blackfin/unistd.h
9900 --- linux-2.6.25.11/include/asm-blackfin/unistd.h       2008-04-17 12:05:43.000000000 -0400
9901 +++ linux-2.6.25.11-vs2.3.0.34.14/include/asm-blackfin/unistd.h 2008-04-19 15:14:52.000000000 -0400
9902 @@ -278,7 +278,7 @@
9903  #define __NR_tgkill            271
9904  #define __NR_utimes            272
9905  #define __NR_fadvise64_64      273
9906 -                               /* 274 __NR_vserver */
9907 +#define __NR_vserver           274
9908                                 /* 275 __NR_mbind */
9909                                 /* 276 __NR_get_mempolicy */
9910                                 /* 277 __NR_set_mempolicy */
9911 diff -NurpP --minimal linux-2.6.25.11/include/asm-generic/tlb.h linux-2.6.25.11-vs2.3.0.34.14/include/asm-generic/tlb.h
9912 --- linux-2.6.25.11/include/asm-generic/tlb.h   2008-04-17 12:05:43.000000000 -0400
9913 +++ linux-2.6.25.11-vs2.3.0.34.14/include/asm-generic/tlb.h     2008-04-19 15:50:15.000000000 -0400
9914 @@ -14,6 +14,7 @@
9915  #define _ASM_GENERIC__TLB_H
9916
9917  #include <linux/swap.h>
9918 +#include <linux/vs_memory.h>
9919  #include <asm/pgalloc.h>
9920  #include <asm/tlbflush.h>
9921
9922 diff -NurpP --minimal linux-2.6.25.11/include/asm-ia64/tlb.h linux-2.6.25.11-vs2.3.0.34.14/include/asm-ia64/tlb.h
9923 --- linux-2.6.25.11/include/asm-ia64/tlb.h      2007-02-04 13:44:54.000000000 -0500
9924 +++ linux-2.6.25.11-vs2.3.0.34.14/include/asm-ia64/tlb.h        2008-04-19 15:14:52.000000000 -0400
9925 @@ -40,6 +40,7 @@
9926  #include <linux/mm.h>
9927  #include <linux/pagemap.h>
9928  #include <linux/swap.h>
9929 +#include <linux/vs_memory.h>
9930
9931  #include <asm/pgalloc.h>
9932  #include <asm/processor.h>
9933 diff -NurpP --minimal linux-2.6.25.11/include/asm-powerpc/systbl.h linux-2.6.25.11-vs2.3.0.34.14/include/asm-powerpc/systbl.h
9934 --- linux-2.6.25.11/include/asm-powerpc/systbl.h        2008-04-17 12:05:44.000000000 -0400
9935 +++ linux-2.6.25.11-vs2.3.0.34.14/include/asm-powerpc/systbl.h  2008-04-19 15:14:52.000000000 -0400
9936 @@ -260,7 +260,7 @@ COMPAT_SYS_SPU(fstatfs64)
9937  SYSX(sys_ni_syscall, ppc_fadvise64_64, ppc_fadvise64_64)
9938  PPC_SYS_SPU(rtas)
9939  OLDSYS(debug_setcontext)
9940 -SYSCALL(ni_syscall)
9941 +SYSX(sys_vserver, sys32_vserver, sys_vserver)
9942  COMPAT_SYS(migrate_pages)
9943  COMPAT_SYS(mbind)
9944  COMPAT_SYS(get_mempolicy)
9945 diff -NurpP --minimal linux-2.6.25.11/include/asm-powerpc/unistd.h linux-2.6.25.11-vs2.3.0.34.14/include/asm-powerpc/unistd.h
9946 --- linux-2.6.25.11/include/asm-powerpc/unistd.h        2008-04-17 12:05:44.000000000 -0400
9947 +++ linux-2.6.25.11-vs2.3.0.34.14/include/asm-powerpc/unistd.h  2008-04-19 15:14:52.000000000 -0400
9948 @@ -275,7 +275,7 @@
9949  #endif
9950  #define __NR_rtas              255
9951  #define __NR_sys_debug_setcontext 256
9952 -/* Number 257 is reserved for vserver */
9953 +#define __NR_vserver           257
9954  #define __NR_migrate_pages     258
9955  #define __NR_mbind             259
9956  #define __NR_get_mempolicy     260
9957 diff -NurpP --minimal linux-2.6.25.11/include/asm-s390/unistd.h linux-2.6.25.11-vs2.3.0.34.14/include/asm-s390/unistd.h
9958 --- linux-2.6.25.11/include/asm-s390/unistd.h   2008-04-17 12:05:44.000000000 -0400
9959 +++ linux-2.6.25.11-vs2.3.0.34.14/include/asm-s390/unistd.h     2008-04-19 15:14:52.000000000 -0400
9960 @@ -202,7 +202,7 @@
9961  #define __NR_clock_gettime     (__NR_timer_create+6)
9962  #define __NR_clock_getres      (__NR_timer_create+7)
9963  #define __NR_clock_nanosleep   (__NR_timer_create+8)
9964 -/* Number 263 is reserved for vserver */
9965 +#define __NR_vserver           263
9966  #define __NR_statfs64          265
9967  #define __NR_fstatfs64         266
9968  #define __NR_remap_file_pages  267
9969 diff -NurpP --minimal linux-2.6.25.11/include/asm-sparc/unistd.h linux-2.6.25.11-vs2.3.0.34.14/include/asm-sparc/unistd.h
9970 --- linux-2.6.25.11/include/asm-sparc/unistd.h  2008-04-17 12:05:44.000000000 -0400
9971 +++ linux-2.6.25.11-vs2.3.0.34.14/include/asm-sparc/unistd.h    2008-04-19 15:14:52.000000000 -0400
9972 @@ -282,7 +282,7 @@
9973  #define __NR_timer_getoverrun  264
9974  #define __NR_timer_delete      265
9975  #define __NR_timer_create      266
9976 -/* #define __NR_vserver                267 Reserved for VSERVER */
9977 +#define __NR_vserver           267
9978  #define __NR_io_setup          268
9979  #define __NR_io_destroy                269
9980  #define __NR_io_submit         270
9981 diff -NurpP --minimal linux-2.6.25.11/include/asm-sparc64/tlb.h linux-2.6.25.11-vs2.3.0.34.14/include/asm-sparc64/tlb.h
9982 --- linux-2.6.25.11/include/asm-sparc64/tlb.h   2008-04-17 12:05:44.000000000 -0400
9983 +++ linux-2.6.25.11-vs2.3.0.34.14/include/asm-sparc64/tlb.h     2008-04-19 15:14:52.000000000 -0400
9984 @@ -3,6 +3,7 @@
9985
9986  #include <linux/swap.h>
9987  #include <linux/pagemap.h>
9988 +#include <linux/vs_memory.h>
9989  #include <asm/pgalloc.h>
9990  #include <asm/tlbflush.h>
9991  #include <asm/mmu_context.h>
9992 diff -NurpP --minimal linux-2.6.25.11/include/asm-sparc64/unistd.h linux-2.6.25.11-vs2.3.0.34.14/include/asm-sparc64/unistd.h
9993 --- linux-2.6.25.11/include/asm-sparc64/unistd.h        2008-04-17 12:05:44.000000000 -0400
9994 +++ linux-2.6.25.11-vs2.3.0.34.14/include/asm-sparc64/unistd.h  2008-04-19 15:14:52.000000000 -0400
9995 @@ -284,7 +284,7 @@
9996  #define __NR_timer_getoverrun  264
9997  #define __NR_timer_delete      265
9998  #define __NR_timer_create      266
9999 -/* #define __NR_vserver                267 Reserved for VSERVER */
10000 +#define __NR_vserver           267
10001  #define __NR_io_setup          268
10002  #define __NR_io_destroy                269
10003  #define __NR_io_submit         270
10004 diff -NurpP --minimal linux-2.6.25.11/include/asm-x86/unistd_64.h linux-2.6.25.11-vs2.3.0.34.14/include/asm-x86/unistd_64.h
10005 --- linux-2.6.25.11/include/asm-x86/unistd_64.h 2008-04-17 12:05:44.000000000 -0400
10006 +++ linux-2.6.25.11-vs2.3.0.34.14/include/asm-x86/unistd_64.h   2008-04-19 15:14:52.000000000 -0400
10007 @@ -535,7 +535,7 @@ __SYSCALL(__NR_tgkill, sys_tgkill)
10008  #define __NR_utimes                            235
10009  __SYSCALL(__NR_utimes, sys_utimes)
10010  #define __NR_vserver                           236
10011 -__SYSCALL(__NR_vserver, sys_ni_syscall)
10012 +__SYSCALL(__NR_vserver, sys_vserver)
10013  #define __NR_mbind                             237
10014  __SYSCALL(__NR_mbind, sys_mbind)
10015  #define __NR_set_mempolicy                     238
10016 diff -NurpP --minimal linux-2.6.25.11/include/linux/capability.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/capability.h
10017 --- linux-2.6.25.11/include/linux/capability.h  2008-07-15 22:05:17.000000000 -0400
10018 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/capability.h    2008-06-10 20:00:46.000000000 -0400
10019 @@ -273,6 +273,7 @@ typedef struct kernel_cap_struct {
10020     arbitrary SCSI commands */
10021  /* Allow setting encryption key on loopback filesystem */
10022  /* Allow setting zone reclaim policy */
10023 +/* Allow the selection of a security context */
10024
10025  #define CAP_SYS_ADMIN        21
10026
10027 @@ -345,7 +346,13 @@ typedef struct kernel_cap_struct {
10028
10029  #define CAP_MAC_ADMIN        33
10030
10031 -#define CAP_LAST_CAP         CAP_MAC_ADMIN
10032 +/* Allow context manipulations */
10033 +/* Allow changing context info on files */
10034 +
10035 +#define CAP_CONTEXT         34
10036 +
10037 +
10038 +#define CAP_LAST_CAP         CAP_CONTEXT
10039
10040  #define cap_valid(x) ((x) >= 0 && (x) <= CAP_LAST_CAP)
10041
10042 diff -NurpP --minimal linux-2.6.25.11/include/linux/devpts_fs.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/devpts_fs.h
10043 --- linux-2.6.25.11/include/linux/devpts_fs.h   2007-02-04 13:44:54.000000000 -0500
10044 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/devpts_fs.h     2008-04-19 15:14:52.000000000 -0400
10045 @@ -30,5 +30,4 @@ static inline void devpts_pty_kill(int n
10046
10047  #endif
10048
10049 -
10050  #endif /* _LINUX_DEVPTS_FS_H */
10051 diff -NurpP --minimal linux-2.6.25.11/include/linux/ext2_fs.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/ext2_fs.h
10052 --- linux-2.6.25.11/include/linux/ext2_fs.h     2008-04-17 11:31:39.000000000 -0400
10053 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/ext2_fs.h       2008-07-08 13:53:46.000000000 -0400
10054 @@ -189,8 +189,12 @@ struct ext2_group_desc
10055  #define EXT2_NOTAIL_FL                 FS_NOTAIL_FL    /* file tail should not be merged */
10056  #define EXT2_DIRSYNC_FL                        FS_DIRSYNC_FL   /* dirsync behaviour (directories only) */
10057  #define EXT2_TOPDIR_FL                 FS_TOPDIR_FL    /* Top of directory hierarchies*/
10058 +#define EXT2_IXUNLINK_FL               FS_IXUNLINK_FL  /* Immutable invert on unlink */
10059  #define EXT2_RESERVED_FL               FS_RESERVED_FL  /* reserved for ext2 lib */
10060
10061 +#define EXT2_BARRIER_FL                        FS_BARRIER_FL   /* Barrier for chroot() */
10062 +#define EXT2_COW_FL                    FS_COW_FL       /* Copy on Write marker */
10063 +
10064  #define EXT2_FL_USER_VISIBLE           FS_FL_USER_VISIBLE      /* User visible flags */
10065  #define EXT2_FL_USER_MODIFIABLE                FS_FL_USER_MODIFIABLE   /* User modifiable flags */
10066
10067 @@ -247,7 +251,7 @@ struct ext2_inode {
10068                 struct {
10069                         __u8    l_i_frag;       /* Fragment number */
10070                         __u8    l_i_fsize;      /* Fragment size */
10071 -                       __u16   i_pad1;
10072 +                       __u16   l_i_tag;        /* Context Tag */
10073                         __le16  l_i_uid_high;   /* these 2 fields    */
10074                         __le16  l_i_gid_high;   /* were reserved2[0] */
10075                         __u32   l_i_reserved2;
10076 @@ -279,6 +283,7 @@ struct ext2_inode {
10077  #define i_gid_low      i_gid
10078  #define i_uid_high     osd2.linux2.l_i_uid_high
10079  #define i_gid_high     osd2.linux2.l_i_gid_high
10080 +#define i_raw_tag      osd2.linux2.l_i_tag
10081  #define i_reserved2    osd2.linux2.l_i_reserved2
10082  #endif
10083
10084 @@ -323,6 +328,7 @@ struct ext2_inode {
10085  #define EXT2_MOUNT_USRQUOTA            0x020000  /* user quota */
10086  #define EXT2_MOUNT_GRPQUOTA            0x040000  /* group quota */
10087  #define EXT2_MOUNT_RESERVATION         0x080000  /* Preallocation */
10088 +#define EXT2_MOUNT_TAGGED              (1<<24)   /* Enable Context Tags */
10089
10090
10091  #define clear_opt(o, opt)              o &= ~EXT2_MOUNT_##opt
10092 diff -NurpP --minimal linux-2.6.25.11/include/linux/ext3_fs.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/ext3_fs.h
10093 --- linux-2.6.25.11/include/linux/ext3_fs.h     2008-04-17 12:05:44.000000000 -0400
10094 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/ext3_fs.h       2008-07-08 14:08:49.000000000 -0400
10095 @@ -173,10 +173,14 @@ struct ext3_group_desc
10096  #define EXT3_NOTAIL_FL                 0x00008000 /* file tail should not be merged */
10097  #define EXT3_DIRSYNC_FL                        0x00010000 /* dirsync behaviour (directories only) */
10098  #define EXT3_TOPDIR_FL                 0x00020000 /* Top of directory hierarchies*/
10099 +#define EXT3_IXUNLINK_FL               0x01000000 /* Immutable invert on unlink */
10100  #define EXT3_RESERVED_FL               0x80000000 /* reserved for ext3 lib */
10101
10102 -#define EXT3_FL_USER_VISIBLE           0x0003DFFF /* User visible flags */
10103 -#define EXT3_FL_USER_MODIFIABLE                0x000380FF /* User modifiable flags */
10104 +#define EXT3_BARRIER_FL                        0x10000000 /* Barrier for chroot() */
10105 +#define EXT3_COW_FL                    0x20000000 /* Copy on Write marker */
10106 +
10107 +#define EXT3_FL_USER_VISIBLE           0x0103DFFF /* User visible flags */
10108 +#define EXT3_FL_USER_MODIFIABLE                0x010380FF /* User modifiable flags */
10109
10110  /*
10111   * Inode dynamic state flags
10112 @@ -292,7 +296,7 @@ struct ext3_inode {
10113                 struct {
10114                         __u8    l_i_frag;       /* Fragment number */
10115                         __u8    l_i_fsize;      /* Fragment size */
10116 -                       __u16   i_pad1;
10117 +                       __u16   l_i_tag;        /* Context Tag */
10118                         __le16  l_i_uid_high;   /* these 2 fields    */
10119                         __le16  l_i_gid_high;   /* were reserved2[0] */
10120                         __u32   l_i_reserved2;
10121 @@ -326,6 +330,7 @@ struct ext3_inode {
10122  #define i_gid_low      i_gid
10123  #define i_uid_high     osd2.linux2.l_i_uid_high
10124  #define i_gid_high     osd2.linux2.l_i_gid_high
10125 +#define i_raw_tag      osd2.linux2.l_i_tag
10126  #define i_reserved2    osd2.linux2.l_i_reserved2
10127
10128  #elif defined(__GNU__)
10129 @@ -380,6 +385,7 @@ struct ext3_inode {
10130  #define EXT3_MOUNT_QUOTA               0x80000 /* Some quota option set */
10131  #define EXT3_MOUNT_USRQUOTA            0x100000 /* "old" user quota */
10132  #define EXT3_MOUNT_GRPQUOTA            0x200000 /* "old" group quota */
10133 +#define EXT3_MOUNT_TAGGED              (1<<24) /* Enable Context Tags */
10134
10135  /* Compatibility, for having both ext2_fs.h and ext3_fs.h included at once */
10136  #ifndef _LINUX_EXT2_FS_H
10137 @@ -822,6 +828,7 @@ struct buffer_head * ext3_bread (handle_
10138  int ext3_get_blocks_handle(handle_t *handle, struct inode *inode,
10139         sector_t iblock, unsigned long maxblocks, struct buffer_head *bh_result,
10140         int create, int extend_disksize);
10141 +extern int ext3_sync_flags(struct inode *inode);
10142
10143  extern struct inode *ext3_iget(struct super_block *, unsigned long);
10144  extern int  ext3_write_inode (struct inode *, int);
10145 diff -NurpP --minimal linux-2.6.25.11/include/linux/ext4_fs.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/ext4_fs.h
10146 --- linux-2.6.25.11/include/linux/ext4_fs.h     2008-04-17 12:05:44.000000000 -0400
10147 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/ext4_fs.h       2008-07-08 14:09:54.000000000 -0400
10148 @@ -231,10 +231,14 @@ struct ext4_group_desc
10149  #define EXT4_TOPDIR_FL                 0x00020000 /* Top of directory hierarchies*/
10150  #define EXT4_HUGE_FILE_FL               0x00040000 /* Set to each huge file */
10151  #define EXT4_EXTENTS_FL                        0x00080000 /* Inode uses extents */
10152 +#define EXT4_IXUNLINK_FL               0x01000000 /* Immutable invert on unlink */
10153  #define EXT4_RESERVED_FL               0x80000000 /* reserved for ext4 lib */
10154
10155 -#define EXT4_FL_USER_VISIBLE           0x000BDFFF /* User visible flags */
10156 -#define EXT4_FL_USER_MODIFIABLE                0x000380FF /* User modifiable flags */
10157 +#define EXT4_BARRIER_FL                        0x10000000 /* Barrier for chroot() */
10158 +#define EXT4_COW_FL                    0x20000000 /* Copy on Write marker */
10159 +
10160 +#define EXT4_FL_USER_VISIBLE           0x010BDFFF /* User visible flags */
10161 +#define EXT4_FL_USER_MODIFIABLE                0x010380FF /* User modifiable flags */
10162
10163  /*
10164   * Inode dynamic state flags
10165 @@ -359,7 +363,8 @@ struct ext4_inode {
10166                         __le16  l_i_file_acl_high;
10167                         __le16  l_i_uid_high;   /* these 2 fields */
10168                         __le16  l_i_gid_high;   /* were reserved2[0] */
10169 -                       __u32   l_i_reserved2;
10170 +                       __u16   l_i_tag;        /* Context Tag */
10171 +                       __u16   l_i_reserved2;
10172                 } linux2;
10173                 struct {
10174                         __le16  h_i_reserved1;  /* Obsoleted fragment number/size which are removed in ext4 */
10175 @@ -465,6 +470,7 @@ do {                                                                               \
10176  #define i_gid_low      i_gid
10177  #define i_uid_high     osd2.linux2.l_i_uid_high
10178  #define i_gid_high     osd2.linux2.l_i_gid_high
10179 +#define i_raw_tag      osd2.linux2.l_i_tag
10180  #define i_reserved2    osd2.linux2.l_i_reserved2
10181
10182  #elif defined(__GNU__)
10183 @@ -528,6 +534,7 @@ do {                                                                               \
10184  #define EXT4_MOUNT_JOURNAL_ASYNC_COMMIT        0x1000000 /* Journal Async Commit */
10185  #define EXT4_MOUNT_I_VERSION            0x2000000 /* i_version support */
10186  #define EXT4_MOUNT_MBALLOC             0x4000000 /* Buddy allocation support */
10187 +#define EXT4_MOUNT_TAGGED              0x8000000 /* Enable Context Tags */
10188  /* Compatibility, for having both ext2_fs.h and ext4_fs.h included at once */
10189  #ifndef _LINUX_EXT2_FS_H
10190  #define clear_opt(o, opt)              o &= ~EXT4_MOUNT_##opt
10191 @@ -1030,6 +1037,7 @@ int ext4_get_blocks_handle(handle_t *han
10192                                 ext4_lblk_t iblock, unsigned long maxblocks,
10193                                 struct buffer_head *bh_result,
10194                                 int create, int extend_disksize);
10195 +extern int ext4_sync_flags(struct inode *inode);
10196
10197  extern struct inode *ext4_iget(struct super_block *, unsigned long);
10198  extern int  ext4_write_inode (struct inode *, int);
10199 diff -NurpP --minimal linux-2.6.25.11/include/linux/fs.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/fs.h
10200 --- linux-2.6.25.11/include/linux/fs.h  2008-04-17 12:05:44.000000000 -0400
10201 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/fs.h    2008-07-08 16:56:52.000000000 -0400
10202 @@ -125,6 +125,8 @@ extern int dir_notify_enable;
10203  #define MS_RELATIME    (1<<21) /* Update atime relative to mtime/ctime. */
10204  #define MS_KERNMOUNT   (1<<22) /* this is a kern_mount call */
10205  #define MS_I_VERSION   (1<<23) /* Update inode I_version field */
10206 +#define MS_TAGGED      (1<<24) /* use generic inode tagging */
10207 +#define MS_TAGID       (1<<25) /* use specific tag for this mount */
10208  #define MS_ACTIVE      (1<<30)
10209  #define MS_NOUSER      (1<<31)
10210
10211 @@ -151,6 +153,14 @@ extern int dir_notify_enable;
10212  #define S_NOCMTIME     128     /* Do not update file c/mtime */
10213  #define S_SWAPFILE     256     /* Do not truncate: swapon got its bmaps */
10214  #define S_PRIVATE      512     /* Inode is fs-internal */
10215 +#define S_IXUNLINK     1024    /* Immutable Invert on unlink */
10216 +
10217 +/* Linux-VServer related Inode flags */
10218 +
10219 +#define V_VALID                1
10220 +#define V_XATTR                2
10221 +#define V_BARRIER      4       /* Barrier for chroot() */
10222 +#define V_COW          8       /* Copy on Write */
10223
10224  /*
10225   * Note that nosuid etc flags are inode-specific: setting some file-system
10226 @@ -167,18 +177,21 @@ extern int dir_notify_enable;
10227   */
10228  #define __IS_FLG(inode,flg) ((inode)->i_sb->s_flags & (flg))
10229
10230 -#define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
10231 +#define IS_RDONLY(inode)       __IS_FLG(inode, MS_RDONLY)
10232  #define IS_SYNC(inode)         (__IS_FLG(inode, MS_SYNCHRONOUS) || \
10233                                         ((inode)->i_flags & S_SYNC))
10234  #define IS_DIRSYNC(inode)      (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
10235                                         ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
10236  #define IS_MANDLOCK(inode)     __IS_FLG(inode, MS_MANDLOCK)
10237 -#define IS_NOATIME(inode)   __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
10238 -#define IS_I_VERSION(inode)   __IS_FLG(inode, MS_I_VERSION)
10239 +#define IS_NOATIME(inode)      __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
10240 +#define IS_I_VERSION(inode)    __IS_FLG(inode, MS_I_VERSION)
10241 +#define IS_TAGGED(inode)       __IS_FLG(inode, MS_TAGGED)
10242
10243  #define IS_NOQUOTA(inode)      ((inode)->i_flags & S_NOQUOTA)
10244  #define IS_APPEND(inode)       ((inode)->i_flags & S_APPEND)
10245  #define IS_IMMUTABLE(inode)    ((inode)->i_flags & S_IMMUTABLE)
10246 +#define IS_IXUNLINK(inode)     ((inode)->i_flags & S_IXUNLINK)
10247 +#define IS_IXORUNLINK(inode)   ((IS_IXUNLINK(inode) ? S_IMMUTABLE : 0) ^ IS_IMMUTABLE(inode))
10248  #define IS_POSIXACL(inode)     __IS_FLG(inode, MS_POSIXACL)
10249
10250  #define IS_DEADDIR(inode)      ((inode)->i_flags & S_DEAD)
10251 @@ -186,6 +199,16 @@ extern int dir_notify_enable;
10252  #define IS_SWAPFILE(inode)     ((inode)->i_flags & S_SWAPFILE)
10253  #define IS_PRIVATE(inode)      ((inode)->i_flags & S_PRIVATE)
10254
10255 +#define IS_BARRIER(inode)      (S_ISDIR((inode)->i_mode) && ((inode)->i_vflags & V_BARRIER))
10256 +
10257 +#ifdef CONFIG_VSERVER_COWBL
10258 +#  define IS_COW(inode)                (IS_IXUNLINK(inode) && IS_IMMUTABLE(inode))
10259 +#  define IS_COW_LINK(inode)   (S_ISREG((inode)->i_mode) && ((inode)->i_nlink > 1))
10260 +#else
10261 +#  define IS_COW(inode)                (0)
10262 +#  define IS_COW_LINK(inode)   (0)
10263 +#endif
10264 +
10265  /* the read-only stuff doesn't really belong here, but any other place is
10266     probably as bad and I don't want to create yet another include file. */
10267
10268 @@ -259,11 +282,14 @@ extern int dir_notify_enable;
10269  #define FS_TOPDIR_FL                   0x00020000 /* Top of directory hierarchies*/
10270  #define FS_EXTENT_FL                   0x00080000 /* Extents */
10271  #define FS_DIRECTIO_FL                 0x00100000 /* Use direct i/o */
10272 +#define FS_IXUNLINK_FL                 0x01000000 /* Immutable invert on unlink */
10273  #define FS_RESERVED_FL                 0x80000000 /* reserved for ext2 lib */
10274
10275 -#define FS_FL_USER_VISIBLE             0x0003DFFF /* User visible flags */
10276 -#define FS_FL_USER_MODIFIABLE          0x000380FF /* User modifiable flags */
10277 +#define FS_BARRIER_FL                  0x10000000 /* Barrier for chroot() */
10278 +#define FS_COW_FL                      0x20000000 /* Copy on Write marker */
10279
10280 +#define FS_FL_USER_VISIBLE             0x0103DFFF /* User visible flags */
10281 +#define FS_FL_USER_MODIFIABLE          0x010380FF /* User modifiable flags */
10282
10283  #define SYNC_FILE_RANGE_WAIT_BEFORE    1
10284  #define SYNC_FILE_RANGE_WRITE          2
10285 @@ -334,6 +360,7 @@ typedef void (dio_iodone_t)(struct kiocb
10286  #define ATTR_FILE      8192
10287  #define ATTR_KILL_PRIV 16384
10288  #define ATTR_OPEN      32768   /* Truncating from open(O_TRUNC) */
10289 +#define ATTR_TAG       65536
10290
10291  /*
10292   * This is the Inode Attributes structure, used for notify_change().  It
10293 @@ -349,6 +376,7 @@ struct iattr {
10294         umode_t         ia_mode;
10295         uid_t           ia_uid;
10296         gid_t           ia_gid;
10297 +       tag_t           ia_tag;
10298         loff_t          ia_size;
10299         struct timespec ia_atime;
10300         struct timespec ia_mtime;
10301 @@ -362,6 +390,9 @@ struct iattr {
10302         struct file     *ia_file;
10303  };
10304
10305 +#define ATTR_FLAG_BARRIER      512     /* Barrier for chroot() */
10306 +#define ATTR_FLAG_IXUNLINK     1024    /* Immutable invert on unlink */
10307 +
10308  /*
10309   * Includes for diskquotas.
10310   */
10311 @@ -600,7 +631,9 @@ struct inode {
10312         unsigned int            i_nlink;
10313         uid_t                   i_uid;
10314         gid_t                   i_gid;
10315 +       tag_t                   i_tag;
10316         dev_t                   i_rdev;
10317 +       dev_t                   i_mdev;
10318         u64                     i_version;
10319         loff_t                  i_size;
10320  #ifdef __NEED_I_SIZE_ORDERED
10321 @@ -648,7 +681,8 @@ struct inode {
10322         unsigned long           i_state;
10323         unsigned long           dirtied_when;   /* jiffies of first dirtying */
10324
10325 -       unsigned int            i_flags;
10326 +       unsigned short          i_flags;
10327 +       unsigned short          i_vflags;
10328
10329         atomic_t                i_writecount;
10330  #ifdef CONFIG_SECURITY
10331 @@ -735,12 +769,12 @@ static inline void i_size_write(struct i
10332
10333  static inline unsigned iminor(const struct inode *inode)
10334  {
10335 -       return MINOR(inode->i_rdev);
10336 +       return MINOR(inode->i_mdev);
10337  }
10338
10339  static inline unsigned imajor(const struct inode *inode)
10340  {
10341 -       return MAJOR(inode->i_rdev);
10342 +       return MAJOR(inode->i_mdev);
10343  }
10344
10345  extern struct block_device *I_BDEV(struct inode *inode);
10346 @@ -795,6 +829,7 @@ struct file {
10347         loff_t                  f_pos;
10348         struct fown_struct      f_owner;
10349         unsigned int            f_uid, f_gid;
10350 +       xid_t                   f_xid;
10351         struct file_ra_state    f_ra;
10352
10353         u64                     f_version;
10354 @@ -879,6 +914,7 @@ struct file_lock {
10355         unsigned char fl_type;
10356         loff_t fl_start;
10357         loff_t fl_end;
10358 +       xid_t fl_xid;
10359
10360         struct fasync_struct *  fl_fasync; /* for lease break notifications */
10361         unsigned long fl_break_time;    /* for nonblocking lease breaks */
10362 @@ -1076,12 +1112,12 @@ extern void unlock_super(struct super_bl
10363   */
10364  extern int vfs_permission(struct nameidata *, int);
10365  extern int vfs_create(struct inode *, struct dentry *, int, struct nameidata *);
10366 -extern int vfs_mkdir(struct inode *, struct dentry *, int);
10367 -extern int vfs_mknod(struct inode *, struct dentry *, int, dev_t);
10368 -extern int vfs_symlink(struct inode *, struct dentry *, const char *, int);
10369 -extern int vfs_link(struct dentry *, struct inode *, struct dentry *);
10370 -extern int vfs_rmdir(struct inode *, struct dentry *);
10371 -extern int vfs_unlink(struct inode *, struct dentry *);
10372 +extern int vfs_mkdir(struct inode *, struct dentry *, int, struct nameidata *);
10373 +extern int vfs_mknod(struct inode *, struct dentry *, int, dev_t, struct nameidata *);
10374 +extern int vfs_symlink(struct inode *, struct dentry *, const char *, int, struct nameidata *);
10375 +extern int vfs_link(struct dentry *, struct inode *, struct dentry *, struct nameidata *);
10376 +extern int vfs_rmdir(struct inode *, struct dentry *, struct nameidata *);
10377 +extern int vfs_unlink(struct inode *, struct dentry *, struct nameidata *);
10378  extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
10379
10380  /*
10381 @@ -1223,6 +1259,7 @@ struct inode_operations {
10382         void (*truncate_range)(struct inode *, loff_t, loff_t);
10383         long (*fallocate)(struct inode *inode, int mode, loff_t offset,
10384                           loff_t len);
10385 +       int (*sync_flags) (struct inode *);
10386  };
10387
10388  struct seq_file;
10389 @@ -1238,6 +1275,7 @@ extern ssize_t vfs_readv(struct file *,
10390                 unsigned long, loff_t *);
10391  extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
10392                 unsigned long, loff_t *);
10393 +ssize_t vfs_sendfile(struct file *, struct file *, loff_t *, size_t, loff_t);
10394
10395  /*
10396   * NOTE: write_inode, delete_inode, clear_inode, put_inode can be called
10397 @@ -1934,6 +1972,7 @@ extern int dcache_dir_open(struct inode
10398  extern int dcache_dir_close(struct inode *, struct file *);
10399  extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
10400  extern int dcache_readdir(struct file *, void *, filldir_t);
10401 +extern int dcache_readdir_filter(struct file *, void *, filldir_t, int (*)(struct dentry *));
10402  extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
10403  extern int simple_statfs(struct dentry *, struct kstatfs *);
10404  extern int simple_link(struct dentry *, struct inode *, struct dentry *);
10405 diff -NurpP --minimal linux-2.6.25.11/include/linux/if_tun.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/if_tun.h
10406 --- linux-2.6.25.11/include/linux/if_tun.h      2008-04-17 12:05:44.000000000 -0400
10407 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/if_tun.h        2008-04-19 15:14:52.000000000 -0400
10408 @@ -42,6 +42,7 @@
10409  #define TUNSETOWNER   _IOW('T', 204, int)
10410  #define TUNSETLINK    _IOW('T', 205, int)
10411  #define TUNSETGROUP   _IOW('T', 206, int)
10412 +#define TUNSETNID     _IOW('T', 215, int)
10413
10414  /* TUNSETIFF ifr flags */
10415  #define IFF_TUN                0x0001
10416 diff -NurpP --minimal linux-2.6.25.11/include/linux/init_task.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/init_task.h
10417 --- linux-2.6.25.11/include/linux/init_task.h   2008-04-17 12:05:44.000000000 -0400
10418 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/init_task.h     2008-04-19 15:14:52.000000000 -0400
10419 @@ -196,6 +196,10 @@ extern struct group_info init_groups;
10420         INIT_IDS                                                        \
10421         INIT_TRACE_IRQFLAGS                                             \
10422         INIT_LOCKDEP                                                    \
10423 +       .xid            = 0,                                            \
10424 +       .vx_info        = NULL,                                         \
10425 +       .nid            = 0,                                            \
10426 +       .nx_info        = NULL,                                         \
10427  }
10428
10429
10430 diff -NurpP --minimal linux-2.6.25.11/include/linux/interrupt.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/interrupt.h
10431 --- linux-2.6.25.11/include/linux/interrupt.h   2008-04-17 12:05:44.000000000 -0400
10432 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/interrupt.h     2008-04-19 15:14:52.000000000 -0400
10433 @@ -8,8 +8,8 @@
10434  #include <linux/preempt.h>
10435  #include <linux/cpumask.h>
10436  #include <linux/irqreturn.h>
10437 -#include <linux/hardirq.h>
10438  #include <linux/sched.h>
10439 +#include <linux/hardirq.h>
10440  #include <linux/irqflags.h>
10441  #include <asm/atomic.h>
10442  #include <asm/ptrace.h>
10443 diff -NurpP --minimal linux-2.6.25.11/include/linux/ipc.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/ipc.h
10444 --- linux-2.6.25.11/include/linux/ipc.h 2008-04-17 12:05:44.000000000 -0400
10445 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/ipc.h   2008-04-19 15:14:52.000000000 -0400
10446 @@ -93,6 +93,7 @@ struct kern_ipc_perm
10447         key_t           key;
10448         uid_t           uid;
10449         gid_t           gid;
10450 +       xid_t           xid;
10451         uid_t           cuid;
10452         gid_t           cgid;
10453         mode_t          mode;
10454 diff -NurpP --minimal linux-2.6.25.11/include/linux/Kbuild linux-2.6.25.11-vs2.3.0.34.14/include/linux/Kbuild
10455 --- linux-2.6.25.11/include/linux/Kbuild        2008-04-17 12:05:44.000000000 -0400
10456 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/Kbuild  2008-04-19 15:14:52.000000000 -0400
10457 @@ -354,3 +354,6 @@ unifdef-y += xattr.h
10458  unifdef-y += xfrm.h
10459
10460  objhdr-y += version.h
10461 +
10462 +header-y += vserver/
10463 +
10464 diff -NurpP --minimal linux-2.6.25.11/include/linux/loop.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/loop.h
10465 --- linux-2.6.25.11/include/linux/loop.h        2008-04-17 12:05:44.000000000 -0400
10466 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/loop.h  2008-04-19 15:14:52.000000000 -0400
10467 @@ -45,6 +45,7 @@ struct loop_device {
10468         struct loop_func_table *lo_encryption;
10469         __u32           lo_init[2];
10470         uid_t           lo_key_owner;   /* Who set the key */
10471 +       xid_t           lo_xid;
10472         int             (*ioctl)(struct loop_device *, int cmd,
10473                                  unsigned long arg);
10474
10475 diff -NurpP --minimal linux-2.6.25.11/include/linux/magic.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/magic.h
10476 --- linux-2.6.25.11/include/linux/magic.h       2008-04-17 11:31:39.000000000 -0400
10477 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/magic.h 2008-04-19 15:14:52.000000000 -0400
10478 @@ -3,7 +3,7 @@
10479
10480  #define ADFS_SUPER_MAGIC       0xadf5
10481  #define AFFS_SUPER_MAGIC       0xadff
10482 -#define AFS_SUPER_MAGIC                0x5346414F
10483 +#define AFS_SUPER_MAGIC                0x5346414F
10484  #define AUTOFS_SUPER_MAGIC     0x0187
10485  #define CODA_SUPER_MAGIC       0x73757245
10486  #define EFS_SUPER_MAGIC                0x414A53
10487 @@ -26,6 +26,7 @@
10488  #define NFS_SUPER_MAGIC                0x6969
10489  #define OPENPROM_SUPER_MAGIC   0x9fa1
10490  #define PROC_SUPER_MAGIC       0x9fa0
10491 +#define DEVPTS_SUPER_MAGIC     0x1cd1
10492  #define QNX4_SUPER_MAGIC       0x002f          /* qnx4 fs detection */
10493
10494  #define REISERFS_SUPER_MAGIC   0x52654973      /* used by gcc */
10495 diff -NurpP --minimal linux-2.6.25.11/include/linux/major.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/major.h
10496 --- linux-2.6.25.11/include/linux/major.h       2008-04-17 10:37:24.000000000 -0400
10497 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/major.h 2008-04-19 15:14:52.000000000 -0400
10498 @@ -15,6 +15,7 @@
10499  #define HD_MAJOR               IDE0_MAJOR
10500  #define PTY_SLAVE_MAJOR                3
10501  #define TTY_MAJOR              4
10502 +#define VROOT_MAJOR            4
10503  #define TTYAUX_MAJOR           5
10504  #define LP_MAJOR               6
10505  #define VCS_MAJOR              7
10506 diff -NurpP --minimal linux-2.6.25.11/include/linux/mm_types.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/mm_types.h
10507 --- linux-2.6.25.11/include/linux/mm_types.h    2008-04-17 12:05:44.000000000 -0400
10508 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/mm_types.h      2008-04-19 15:14:52.000000000 -0400
10509 @@ -201,6 +201,7 @@ struct mm_struct {
10510
10511         /* Architecture-specific MM context */
10512         mm_context_t context;
10513 +       struct vx_info *mm_vx_info;
10514
10515         /* Swap token stuff */
10516         /*
10517 diff -NurpP --minimal linux-2.6.25.11/include/linux/mount.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/mount.h
10518 --- linux-2.6.25.11/include/linux/mount.h       2008-04-17 12:05:44.000000000 -0400
10519 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/mount.h 2008-04-19 15:45:19.000000000 -0400
10520 @@ -28,6 +28,9 @@ struct mnt_namespace;
10521  #define MNT_NOATIME    0x08
10522  #define MNT_NODIRATIME 0x10
10523  #define MNT_RELATIME   0x20
10524 +#define MNT_RDONLY     0x40
10525 +
10526 +#define MNT_IS_RDONLY(m)       ((m) && ((m)->mnt_flags & MNT_RDONLY))
10527
10528  #define MNT_SHRINKABLE 0x100
10529
10530 @@ -35,6 +38,10 @@ struct mnt_namespace;
10531  #define MNT_UNBINDABLE 0x2000  /* if the vfsmount is a unbindable mount */
10532  #define MNT_PNODE_MASK 0x3000  /* propagation flag mask */
10533
10534 +#define MNT_TAGID      0x10000
10535 +#define MNT_NOTAG      0x20000
10536 +#define MNT_NOTAGCHECK 0x40000
10537 +
10538  struct vfsmount {
10539         struct list_head mnt_hash;
10540         struct vfsmount *mnt_parent;    /* fs we are mounted on */
10541 @@ -62,6 +69,7 @@ struct vfsmount {
10542         int mnt_expiry_mark;            /* true if marked for expiry */
10543         int mnt_pinned;
10544         int mnt_ghosts;
10545 +       tag_t mnt_tag;                  /* tagging used for vfsmount */
10546  };
10547
10548  static inline struct vfsmount *mntget(struct vfsmount *mnt)
10549 diff -NurpP --minimal linux-2.6.25.11/include/linux/net.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/net.h
10550 --- linux-2.6.25.11/include/linux/net.h 2008-04-17 12:05:44.000000000 -0400
10551 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/net.h   2008-04-19 15:14:52.000000000 -0400
10552 @@ -65,6 +65,7 @@ typedef enum {
10553  #define SOCK_NOSPACE           2
10554  #define SOCK_PASSCRED          3
10555  #define SOCK_PASSSEC           4
10556 +#define SOCK_USER_SOCKET       5
10557
10558  #ifndef ARCH_HAS_SOCKET_TYPES
10559  /**
10560 diff -NurpP --minimal linux-2.6.25.11/include/linux/nfs_mount.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/nfs_mount.h
10561 --- linux-2.6.25.11/include/linux/nfs_mount.h   2008-04-17 12:05:44.000000000 -0400
10562 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/nfs_mount.h     2008-04-19 15:14:52.000000000 -0400
10563 @@ -63,6 +63,7 @@ struct nfs_mount_data {
10564  #define NFS_MOUNT_SECFLAVOUR   0x2000  /* 5 */
10565  #define NFS_MOUNT_NORDIRPLUS   0x4000  /* 5 */
10566  #define NFS_MOUNT_UNSHARED     0x8000  /* 5 */
10567 -#define NFS_MOUNT_FLAGMASK     0xFFFF
10568 +#define NFS_MOUNT_TAGGED       0x10000 /* context tagging */
10569 +#define NFS_MOUNT_FLAGMASK     0x1FFFF
10570
10571  #endif
10572 diff -NurpP --minimal linux-2.6.25.11/include/linux/nsproxy.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/nsproxy.h
10573 --- linux-2.6.25.11/include/linux/nsproxy.h     2008-04-17 11:31:39.000000000 -0400
10574 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/nsproxy.h       2008-04-19 15:14:52.000000000 -0400
10575 @@ -3,6 +3,7 @@
10576
10577  #include <linux/spinlock.h>
10578  #include <linux/sched.h>
10579 +#include <linux/vserver/debug.h>
10580
10581  struct mnt_namespace;
10582  struct uts_namespace;
10583 @@ -63,22 +64,33 @@ static inline struct nsproxy *task_nspro
10584  }
10585
10586  int copy_namespaces(unsigned long flags, struct task_struct *tsk);
10587 +struct nsproxy *copy_nsproxy(struct nsproxy *orig);
10588  void exit_task_namespaces(struct task_struct *tsk);
10589  void switch_task_namespaces(struct task_struct *tsk, struct nsproxy *new);
10590  void free_nsproxy(struct nsproxy *ns);
10591  int unshare_nsproxy_namespaces(unsigned long, struct nsproxy **,
10592         struct fs_struct *);
10593
10594 -static inline void put_nsproxy(struct nsproxy *ns)
10595 +#define        get_nsproxy(n)  __get_nsproxy(n, __FILE__, __LINE__)
10596 +
10597 +static inline void __get_nsproxy(struct nsproxy *ns,
10598 +       const char *_file, int _line)
10599  {
10600 -       if (atomic_dec_and_test(&ns->count)) {
10601 -               free_nsproxy(ns);
10602 -       }
10603 +       vxlprintk(VXD_CBIT(space, 0), "get_nsproxy(%p[%u])",
10604 +               ns, atomic_read(&ns->count), _file, _line);
10605 +       atomic_inc(&ns->count);
10606  }
10607
10608 -static inline void get_nsproxy(struct nsproxy *ns)
10609 +#define        put_nsproxy(n)  __put_nsproxy(n, __FILE__, __LINE__)
10610 +
10611 +static inline void __put_nsproxy(struct nsproxy *ns,
10612 +       const char *_file, int _line)
10613  {
10614 -       atomic_inc(&ns->count);
10615 +       vxlprintk(VXD_CBIT(space, 0), "put_nsproxy(%p[%u])",
10616 +               ns, atomic_read(&ns->count), _file, _line);
10617 +       if (atomic_dec_and_test(&ns->count)) {
10618 +               free_nsproxy(ns);
10619 +       }
10620  }
10621
10622  #ifdef CONFIG_CGROUP_NS
10623 diff -NurpP --minimal linux-2.6.25.11/include/linux/pid.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/pid.h
10624 --- linux-2.6.25.11/include/linux/pid.h 2008-04-17 12:05:44.000000000 -0400
10625 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/pid.h   2008-04-29 17:56:00.000000000 -0400
10626 @@ -8,7 +8,8 @@ enum pid_type
10627         PIDTYPE_PID,
10628         PIDTYPE_PGID,
10629         PIDTYPE_SID,
10630 -       PIDTYPE_MAX
10631 +       PIDTYPE_MAX,
10632 +       PIDTYPE_REALPID
10633  };
10634
10635  /*
10636 @@ -142,6 +143,7 @@ static inline pid_t pid_nr(struct pid *p
10637  }
10638
10639  pid_t pid_nr_ns(struct pid *pid, struct pid_namespace *ns);
10640 +pid_t pid_unmapped_nr_ns(struct pid *pid, struct pid_namespace *ns);
10641  pid_t pid_vnr(struct pid *pid);
10642
10643  #define do_each_pid_task(pid, type, task)                              \
10644 diff -NurpP --minimal linux-2.6.25.11/include/linux/proc_fs.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/proc_fs.h
10645 --- linux-2.6.25.11/include/linux/proc_fs.h     2008-04-17 12:05:44.000000000 -0400
10646 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/proc_fs.h       2008-04-19 15:44:54.000000000 -0400
10647 @@ -59,6 +59,7 @@ struct proc_dir_entry {
10648         nlink_t nlink;
10649         uid_t uid;
10650         gid_t gid;
10651 +       int vx_flags;
10652         loff_t size;
10653         const struct inode_operations *proc_iops;
10654         /*
10655 @@ -265,16 +266,23 @@ static inline void kclist_add(struct kco
10656  extern void kclist_add(struct kcore_list *, void *, size_t);
10657  #endif
10658
10659 +struct vx_info;
10660 +struct nx_info;
10661 +
10662  union proc_op {
10663         int (*proc_get_link)(struct inode *, struct path *);
10664         int (*proc_read)(struct task_struct *task, char *page);
10665         int (*proc_show)(struct seq_file *m,
10666                 struct pid_namespace *ns, struct pid *pid,
10667                 struct task_struct *task);
10668 +       int (*proc_vs_read)(char *page);
10669 +       int (*proc_vxi_read)(struct vx_info *vxi, char *page);
10670 +       int (*proc_nxi_read)(struct nx_info *nxi, char *page);
10671  };
10672
10673  struct proc_inode {
10674         struct pid *pid;
10675 +       int vx_flags;
10676         int fd;
10677         union proc_op op;
10678         struct proc_dir_entry *pde;
10679 diff -NurpP --minimal linux-2.6.25.11/include/linux/reiserfs_fs.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/reiserfs_fs.h
10680 --- linux-2.6.25.11/include/linux/reiserfs_fs.h 2008-07-15 22:05:17.000000000 -0400
10681 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/reiserfs_fs.h   2008-07-08 14:14:46.000000000 -0400
10682 @@ -837,6 +837,11 @@ struct stat_data_v1 {
10683  #define REISERFS_COMPR_FL     FS_COMPR_FL
10684  #define REISERFS_NOTAIL_FL    FS_NOTAIL_FL
10685
10686 +/* unfortunately reiserfs sdattr is only 16 bit */
10687 +#define REISERFS_IXUNLINK_FL  (FS_IXUNLINK_FL >> 16)
10688 +#define REISERFS_BARRIER_FL   (FS_BARRIER_FL >> 16)
10689 +#define REISERFS_COW_FL       (FS_COW_FL >> 16)
10690 +
10691  /* persistent flags that file inherits from the parent directory */
10692  #define REISERFS_INHERIT_MASK ( REISERFS_IMMUTABLE_FL |        \
10693                                 REISERFS_SYNC_FL |      \
10694 @@ -846,6 +851,9 @@ struct stat_data_v1 {
10695                                 REISERFS_COMPR_FL |     \
10696                                 REISERFS_NOTAIL_FL )
10697
10698 +#define REISERFS_FL_USER_VISIBLE       0x80FF
10699 +#define REISERFS_FL_USER_MODIFIABLE    0x80FF
10700 +
10701  /* Stat Data on disk (reiserfs version of UFS disk inode minus the
10702     address blocks) */
10703  struct stat_data {
10704 @@ -1911,6 +1919,7 @@ static inline void reiserfs_update_sd(st
10705  void sd_attrs_to_i_attrs(__u16 sd_attrs, struct inode *inode);
10706  void i_attrs_to_sd_attrs(struct inode *inode, __u16 * sd_attrs);
10707  int reiserfs_setattr(struct dentry *dentry, struct iattr *attr);
10708 +int reiserfs_sync_flags(struct inode *inode);
10709
10710  /* namei.c */
10711  void set_de_name_and_namelen(struct reiserfs_dir_entry *de);
10712 diff -NurpP --minimal linux-2.6.25.11/include/linux/reiserfs_fs_sb.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/reiserfs_fs_sb.h
10713 --- linux-2.6.25.11/include/linux/reiserfs_fs_sb.h      2008-04-17 12:05:44.000000000 -0400
10714 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/reiserfs_fs_sb.h        2008-04-19 15:14:52.000000000 -0400
10715 @@ -456,6 +456,7 @@ enum reiserfs_mount_options {
10716         REISERFS_POSIXACL,
10717         REISERFS_BARRIER_NONE,
10718         REISERFS_BARRIER_FLUSH,
10719 +       REISERFS_TAGGED,
10720
10721         /* Actions on error */
10722         REISERFS_ERROR_PANIC,
10723 diff -NurpP --minimal linux-2.6.25.11/include/linux/sched.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/sched.h
10724 --- linux-2.6.25.11/include/linux/sched.h       2008-04-17 12:05:44.000000000 -0400
10725 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/sched.h 2008-04-21 11:09:01.000000000 -0400
10726 @@ -73,7 +73,6 @@ struct sched_param {
10727  #include <linux/fs_struct.h>
10728  #include <linux/compiler.h>
10729  #include <linux/completion.h>
10730 -#include <linux/pid.h>
10731  #include <linux/percpu.h>
10732  #include <linux/topology.h>
10733  #include <linux/proportions.h>
10734 @@ -89,6 +88,7 @@ struct sched_param {
10735  #include <linux/task_io_accounting.h>
10736  #include <linux/kobject.h>
10737  #include <linux/latencytop.h>
10738 +#include <linux/pid.h>
10739
10740  #include <asm/processor.h>
10741
10742 @@ -355,25 +355,27 @@ extern void arch_unmap_area_topdown(stru
10743   * The mm counters are not protected by its page_table_lock,
10744   * so must be incremented atomically.
10745   */
10746 -#define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
10747 -#define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
10748 -#define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
10749 -#define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
10750 -#define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
10751 -
10752 +#define __set_mm_counter(mm, member, value) \
10753 +       atomic_long_set(&(mm)->_##member, value)
10754 +#define get_mm_counter(mm, member) \
10755 +       ((unsigned long)atomic_long_read(&(mm)->_##member))
10756  #else  /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
10757  /*
10758   * The mm counters are protected by its page_table_lock,
10759   * so can be incremented directly.
10760   */
10761 -#define set_mm_counter(mm, member, value) (mm)->_##member = (value)
10762 +#define __set_mm_counter(mm, member, value) (mm)->_##member = (value)
10763  #define get_mm_counter(mm, member) ((mm)->_##member)
10764 -#define add_mm_counter(mm, member, value) (mm)->_##member += (value)
10765 -#define inc_mm_counter(mm, member) (mm)->_##member++
10766 -#define dec_mm_counter(mm, member) (mm)->_##member--
10767
10768  #endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
10769
10770 +#define set_mm_counter(mm, member, value) \
10771 +       vx_ ## member ## pages_sub((mm), (get_mm_counter(mm, member) - value))
10772 +#define add_mm_counter(mm, member, value) \
10773 +       vx_ ## member ## pages_add((mm), (value))
10774 +#define inc_mm_counter(mm, member) vx_ ## member ## pages_inc((mm))
10775 +#define dec_mm_counter(mm, member) vx_ ## member ## pages_dec((mm))
10776 +
10777  #define get_mm_rss(mm)                                 \
10778         (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
10779  #define update_hiwater_rss(mm) do {                    \
10780 @@ -1162,6 +1164,14 @@ struct task_struct {
10781  #endif
10782         seccomp_t seccomp;
10783
10784 +/* vserver context data */
10785 +       struct vx_info *vx_info;
10786 +       struct nx_info *nx_info;
10787 +
10788 +       xid_t xid;
10789 +       nid_t nid;
10790 +       tag_t tag;
10791 +
10792  /* Thread group tracking */
10793         u32 parent_exec_id;
10794         u32 self_exec_id;
10795 @@ -1350,6 +1360,11 @@ struct pid_namespace;
10796   * see also pid_nr() etc in include/linux/pid.h
10797   */
10798
10799 +#include <linux/vserver/base.h>
10800 +#include <linux/vserver/context.h>
10801 +#include <linux/vserver/debug.h>
10802 +#include <linux/vserver/pid.h>
10803 +
10804  static inline pid_t task_pid_nr(struct task_struct *tsk)
10805  {
10806         return tsk->pid;
10807 @@ -1359,7 +1374,7 @@ pid_t task_pid_nr_ns(struct task_struct
10808
10809  static inline pid_t task_pid_vnr(struct task_struct *tsk)
10810  {
10811 -       return pid_vnr(task_pid(tsk));
10812 +       return vx_map_pid(pid_vnr(task_pid(tsk)));
10813  }
10814
10815
10816 @@ -1372,7 +1387,7 @@ pid_t task_tgid_nr_ns(struct task_struct
10817
10818  static inline pid_t task_tgid_vnr(struct task_struct *tsk)
10819  {
10820 -       return pid_vnr(task_tgid(tsk));
10821 +       return vx_map_tgid(pid_vnr(task_tgid(tsk)));
10822  }
10823
10824
10825 diff -NurpP --minimal linux-2.6.25.11/include/linux/shmem_fs.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/shmem_fs.h
10826 --- linux-2.6.25.11/include/linux/shmem_fs.h    2008-04-17 12:05:44.000000000 -0400
10827 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/shmem_fs.h      2008-04-19 15:14:52.000000000 -0400
10828 @@ -8,6 +8,9 @@
10829
10830  #define SHMEM_NR_DIRECT 16
10831
10832 +#define TMPFS_SUPER_MAGIC      0x01021994
10833 +
10834 +
10835  struct shmem_inode_info {
10836         spinlock_t              lock;
10837         unsigned long           flags;
10838 diff -NurpP --minimal linux-2.6.25.11/include/linux/stat.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/stat.h
10839 --- linux-2.6.25.11/include/linux/stat.h        2008-04-17 10:33:07.000000000 -0400
10840 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/stat.h  2008-04-19 15:14:52.000000000 -0400
10841 @@ -66,6 +66,7 @@ struct kstat {
10842         unsigned int    nlink;
10843         uid_t           uid;
10844         gid_t           gid;
10845 +       tag_t           tag;
10846         dev_t           rdev;
10847         loff_t          size;
10848         struct timespec  atime;
10849 diff -NurpP --minimal linux-2.6.25.11/include/linux/sunrpc/auth.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/sunrpc/auth.h
10850 --- linux-2.6.25.11/include/linux/sunrpc/auth.h 2008-04-17 10:37:24.000000000 -0400
10851 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/sunrpc/auth.h   2008-04-19 15:14:52.000000000 -0400
10852 @@ -25,6 +25,7 @@
10853  struct auth_cred {
10854         uid_t   uid;
10855         gid_t   gid;
10856 +       tag_t   tag;
10857         struct group_info *group_info;
10858  };
10859
10860 diff -NurpP --minimal linux-2.6.25.11/include/linux/sunrpc/clnt.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/sunrpc/clnt.h
10861 --- linux-2.6.25.11/include/linux/sunrpc/clnt.h 2008-04-17 12:05:44.000000000 -0400
10862 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/sunrpc/clnt.h   2008-04-19 15:14:52.000000000 -0400
10863 @@ -42,7 +42,8 @@ struct rpc_clnt {
10864
10865         unsigned int            cl_softrtry : 1,/* soft timeouts */
10866                                 cl_discrtry : 1,/* disconnect before retry */
10867 -                               cl_autobind : 1;/* use getport() */
10868 +                               cl_autobind : 1,/* use getport() */
10869 +                               cl_tag      : 1;/* context tagging */
10870
10871         struct rpc_rtt *        cl_rtt;         /* RTO estimator data */
10872         const struct rpc_timeout *cl_timeout;   /* Timeout strategy */
10873 diff -NurpP --minimal linux-2.6.25.11/include/linux/syscalls.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/syscalls.h
10874 --- linux-2.6.25.11/include/linux/syscalls.h    2008-04-17 12:05:44.000000000 -0400
10875 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/syscalls.h      2008-04-19 15:14:52.000000000 -0400
10876 @@ -294,6 +294,8 @@ asmlinkage long sys_symlink(const char _
10877  asmlinkage long sys_unlink(const char __user *pathname);
10878  asmlinkage long sys_rename(const char __user *oldname,
10879                                 const char __user *newname);
10880 +asmlinkage long sys_copyfile(const char __user *from, const char __user *to,
10881 +                               umode_t mode);
10882  asmlinkage long sys_chmod(const char __user *filename, mode_t mode);
10883  asmlinkage long sys_fchmod(unsigned int fd, mode_t mode);
10884
10885 diff -NurpP --minimal linux-2.6.25.11/include/linux/sysctl.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/sysctl.h
10886 --- linux-2.6.25.11/include/linux/sysctl.h      2008-04-17 12:05:44.000000000 -0400
10887 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/sysctl.h        2008-04-19 15:14:52.000000000 -0400
10888 @@ -70,6 +70,7 @@ enum
10889         CTL_ABI=9,              /* Binary emulation */
10890         CTL_CPU=10,             /* CPU stuff (speed scaling, etc) */
10891         CTL_ARLAN=254,          /* arlan wireless driver */
10892 +       CTL_VSERVER=4242,       /* Linux-VServer debug */
10893         CTL_S390DBF=5677,       /* s390 debug */
10894         CTL_SUNRPC=7249,        /* sunrpc debug */
10895         CTL_PM=9899,            /* frv power management */
10896 @@ -104,6 +105,7 @@ enum
10897
10898         KERN_PANIC=15,          /* int: panic timeout */
10899         KERN_REALROOTDEV=16,    /* real root device to mount after initrd */
10900 +       KERN_VSHELPER=17,       /* string: path to vshelper policy agent */
10901
10902         KERN_SPARC_REBOOT=21,   /* reboot command on Sparc */
10903         KERN_CTLALTDEL=22,      /* int: allow ctl-alt-del to reboot */
10904 diff -NurpP --minimal linux-2.6.25.11/include/linux/sysfs.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/sysfs.h
10905 --- linux-2.6.25.11/include/linux/sysfs.h       2008-04-17 12:05:44.000000000 -0400
10906 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/sysfs.h 2008-04-19 15:14:52.000000000 -0400
10907 @@ -17,6 +17,8 @@
10908  #include <linux/list.h>
10909  #include <asm/atomic.h>
10910
10911 +#define SYSFS_SUPER_MAGIC      0x62656572
10912 +
10913  struct kobject;
10914  struct module;
10915
10916 diff -NurpP --minimal linux-2.6.25.11/include/linux/time.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/time.h
10917 --- linux-2.6.25.11/include/linux/time.h        2008-04-17 12:05:44.000000000 -0400
10918 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/time.h  2008-04-19 15:14:52.000000000 -0400
10919 @@ -183,6 +183,9 @@ static inline void timespec_add_ns(struc
10920         }
10921         a->tv_nsec = ns;
10922  }
10923 +
10924 +#include <linux/vs_time.h>
10925 +
10926  #endif /* __KERNEL__ */
10927
10928  #define NFDBITS                        __NFDBITS
10929 diff -NurpP --minimal linux-2.6.25.11/include/linux/types.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/types.h
10930 --- linux-2.6.25.11/include/linux/types.h       2008-07-15 22:05:17.000000000 -0400
10931 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/types.h 2008-06-10 20:00:46.000000000 -0400
10932 @@ -36,6 +36,9 @@ typedef __kernel_uid32_t      uid_t;
10933  typedef __kernel_gid32_t       gid_t;
10934  typedef __kernel_uid16_t        uid16_t;
10935  typedef __kernel_gid16_t        gid16_t;
10936 +typedef unsigned int           xid_t;
10937 +typedef unsigned int           nid_t;
10938 +typedef unsigned int           tag_t;
10939
10940  typedef unsigned long          uintptr_t;
10941
10942 diff -NurpP --minimal linux-2.6.25.11/include/linux/vroot.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vroot.h
10943 --- linux-2.6.25.11/include/linux/vroot.h       1969-12-31 19:00:00.000000000 -0500
10944 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vroot.h 2008-04-19 15:14:52.000000000 -0400
10945 @@ -0,0 +1,51 @@
10946 +
10947 +/*
10948 + * include/linux/vroot.h
10949 + *
10950 + * written by Herbert Pötzl, 9/11/2002
10951 + * ported to 2.6 by Herbert Pötzl, 30/12/2004
10952 + *
10953 + * Copyright (C) 2002-2007 by Herbert Pötzl.
10954 + * Redistribution of this file is permitted under the
10955 + * GNU General Public License.
10956 + */
10957 +
10958 +#ifndef _LINUX_VROOT_H
10959 +#define _LINUX_VROOT_H
10960 +
10961 +
10962 +#ifdef __KERNEL__
10963 +
10964 +/* Possible states of device */
10965 +enum {
10966 +       Vr_unbound,
10967 +       Vr_bound,
10968 +};
10969 +
10970 +struct vroot_device {
10971 +       int             vr_number;
10972 +       int             vr_refcnt;
10973 +
10974 +       struct semaphore        vr_ctl_mutex;
10975 +       struct block_device    *vr_device;
10976 +       int                     vr_state;
10977 +};
10978 +
10979 +
10980 +typedef struct block_device *(vroot_grb_func)(struct block_device *);
10981 +
10982 +extern int register_vroot_grb(vroot_grb_func *);
10983 +extern int unregister_vroot_grb(vroot_grb_func *);
10984 +
10985 +#endif /* __KERNEL__ */
10986 +
10987 +#define MAX_VROOT_DEFAULT      8
10988 +
10989 +/*
10990 + * IOCTL commands --- we will commandeer 0x56 ('V')
10991 + */
10992 +
10993 +#define VROOT_SET_DEV          0x5600
10994 +#define VROOT_CLR_DEV          0x5601
10995 +
10996 +#endif /* _LINUX_VROOT_H */
10997 diff -NurpP --minimal linux-2.6.25.11/include/linux/vs_base.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_base.h
10998 --- linux-2.6.25.11/include/linux/vs_base.h     1969-12-31 19:00:00.000000000 -0500
10999 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_base.h       2008-04-19 15:14:52.000000000 -0400
11000 @@ -0,0 +1,10 @@
11001 +#ifndef _VS_BASE_H
11002 +#define _VS_BASE_H
11003 +
11004 +#include "vserver/base.h"
11005 +#include "vserver/check.h"
11006 +#include "vserver/debug.h"
11007 +
11008 +#else
11009 +#warning duplicate inclusion
11010 +#endif
11011 diff -NurpP --minimal linux-2.6.25.11/include/linux/vs_context.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_context.h
11012 --- linux-2.6.25.11/include/linux/vs_context.h  1969-12-31 19:00:00.000000000 -0500
11013 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_context.h    2008-04-29 18:42:09.000000000 -0400
11014 @@ -0,0 +1,227 @@
11015 +#ifndef _VS_CONTEXT_H
11016 +#define _VS_CONTEXT_H
11017 +
11018 +#include "vserver/base.h"
11019 +#include "vserver/check.h"
11020 +#include "vserver/context.h"
11021 +#include "vserver/history.h"
11022 +#include "vserver/debug.h"
11023 +
11024 +#include <linux/sched.h>
11025 +
11026 +
11027 +#define get_vx_info(i) __get_vx_info(i, __FILE__, __LINE__, __HERE__)
11028 +
11029 +static inline struct vx_info *__get_vx_info(struct vx_info *vxi,
11030 +       const char *_file, int _line, void *_here)
11031 +{
11032 +       if (!vxi)
11033 +               return NULL;
11034 +
11035 +       vxlprintk(VXD_CBIT(xid, 2), "get_vx_info(%p[#%d.%d])",
11036 +               vxi, vxi ? vxi->vx_id : 0,
11037 +               vxi ? atomic_read(&vxi->vx_usecnt) : 0,
11038 +               _file, _line);
11039 +       __vxh_get_vx_info(vxi, _here);
11040 +
11041 +       atomic_inc(&vxi->vx_usecnt);
11042 +       return vxi;
11043 +}
11044 +
11045 +
11046 +extern void free_vx_info(struct vx_info *);
11047 +
11048 +#define put_vx_info(i) __put_vx_info(i, __FILE__, __LINE__, __HERE__)
11049 +
11050 +static inline void __put_vx_info(struct vx_info *vxi,
11051 +       const char *_file, int _line, void *_here)
11052 +{
11053 +       if (!vxi)
11054 +               return;
11055 +
11056 +       vxlprintk(VXD_CBIT(xid, 2), "put_vx_info(%p[#%d.%d])",
11057 +               vxi, vxi ? vxi->vx_id : 0,
11058 +               vxi ? atomic_read(&vxi->vx_usecnt) : 0,
11059 +               _file, _line);
11060 +       __vxh_put_vx_info(vxi, _here);
11061 +
11062 +       if (atomic_dec_and_test(&vxi->vx_usecnt))
11063 +               free_vx_info(vxi);
11064 +}
11065 +
11066 +
11067 +#define init_vx_info(p, i) \
11068 +       __init_vx_info(p, i, __FILE__, __LINE__, __HERE__)
11069 +
11070 +static inline void __init_vx_info(struct vx_info **vxp, struct vx_info *vxi,
11071 +       const char *_file, int _line, void *_here)
11072 +{
11073 +       if (vxi) {
11074 +               vxlprintk(VXD_CBIT(xid, 3),
11075 +                       "init_vx_info(%p[#%d.%d])",
11076 +                       vxi, vxi ? vxi->vx_id : 0,
11077 +                       vxi ? atomic_read(&vxi->vx_usecnt) : 0,
11078 +                       _file, _line);
11079 +               __vxh_init_vx_info(vxi, vxp, _here);
11080 +
11081 +               atomic_inc(&vxi->vx_usecnt);
11082 +       }
11083 +       *vxp = vxi;
11084 +}
11085 +
11086 +
11087 +#define set_vx_info(p, i) \
11088 +       __set_vx_info(p, i, __FILE__, __LINE__, __HERE__)
11089 +
11090 +static inline void __set_vx_info(struct vx_info **vxp, struct vx_info *vxi,
11091 +       const char *_file, int _line, void *_here)
11092 +{
11093 +       struct vx_info *vxo;
11094 +
11095 +       if (!vxi)
11096 +               return;
11097 +
11098 +       vxlprintk(VXD_CBIT(xid, 3), "set_vx_info(%p[#%d.%d])",
11099 +               vxi, vxi ? vxi->vx_id : 0,
11100 +               vxi ? atomic_read(&vxi->vx_usecnt) : 0,
11101 +               _file, _line);
11102 +       __vxh_set_vx_info(vxi, vxp, _here);
11103 +
11104 +       atomic_inc(&vxi->vx_usecnt);
11105 +       vxo = xchg(vxp, vxi);
11106 +       BUG_ON(vxo);
11107 +}
11108 +
11109 +
11110 +#define clr_vx_info(p) __clr_vx_info(p, __FILE__, __LINE__, __HERE__)
11111 +
11112 +static inline void __clr_vx_info(struct vx_info **vxp,
11113 +       const char *_file, int _line, void *_here)
11114 +{
11115 +       struct vx_info *vxo;
11116 +
11117 +       vxo = xchg(vxp, NULL);
11118 +       if (!vxo)
11119 +               return;
11120 +
11121 +       vxlprintk(VXD_CBIT(xid, 3), "clr_vx_info(%p[#%d.%d])",
11122 +               vxo, vxo ? vxo->vx_id : 0,
11123 +               vxo ? atomic_read(&vxo->vx_usecnt) : 0,
11124 +               _file, _line);
11125 +       __vxh_clr_vx_info(vxo, vxp, _here);
11126 +
11127 +       if (atomic_dec_and_test(&vxo->vx_usecnt))
11128 +               free_vx_info(vxo);
11129 +}
11130 +
11131 +
11132 +#define claim_vx_info(v, p) \
11133 +       __claim_vx_info(v, p, __FILE__, __LINE__, __HERE__)
11134 +
11135 +static inline void __claim_vx_info(struct vx_info *vxi,
11136 +       struct task_struct *task,
11137 +       const char *_file, int _line, void *_here)
11138 +{
11139 +       vxlprintk(VXD_CBIT(xid, 3), "claim_vx_info(%p[#%d.%d.%d]) %p",
11140 +               vxi, vxi ? vxi->vx_id : 0,
11141 +               vxi ? atomic_read(&vxi->vx_usecnt) : 0,
11142 +               vxi ? atomic_read(&vxi->vx_tasks) : 0,
11143 +               task, _file, _line);
11144 +       __vxh_claim_vx_info(vxi, task, _here);
11145 +
11146 +       atomic_inc(&vxi->vx_tasks);
11147 +}
11148 +
11149 +
11150 +extern void unhash_vx_info(struct vx_info *);
11151 +
11152 +#define release_vx_info(v, p) \
11153 +       __release_vx_info(v, p, __FILE__, __LINE__, __HERE__)
11154 +
11155 +static inline void __release_vx_info(struct vx_info *vxi,
11156 +       struct task_struct *task,
11157 +       const char *_file, int _line, void *_here)
11158 +{
11159 +       vxlprintk(VXD_CBIT(xid, 3), "release_vx_info(%p[#%d.%d.%d]) %p",
11160 +               vxi, vxi ? vxi->vx_id : 0,
11161 +               vxi ? atomic_read(&vxi->vx_usecnt) : 0,
11162 +               vxi ? atomic_read(&vxi->vx_tasks) : 0,
11163 +               task, _file, _line);
11164 +       __vxh_release_vx_info(vxi, task, _here);
11165 +
11166 +       might_sleep();
11167 +
11168 +       if (atomic_dec_and_test(&vxi->vx_tasks))
11169 +               unhash_vx_info(vxi);
11170 +}
11171 +
11172 +
11173 +#define task_get_vx_info(p) \
11174 +       __task_get_vx_info(p, __FILE__, __LINE__, __HERE__)
11175 +
11176 +static inline struct vx_info *__task_get_vx_info(struct task_struct *p,
11177 +       const char *_file, int _line, void *_here)
11178 +{
11179 +       struct vx_info *vxi;
11180 +
11181 +       task_lock(p);
11182 +       vxlprintk(VXD_CBIT(xid, 5), "task_get_vx_info(%p)",
11183 +               p, _file, _line);
11184 +       vxi = __get_vx_info(p->vx_info, _file, _line, _here);
11185 +       task_unlock(p);
11186 +       return vxi;
11187 +}
11188 +
11189 +
11190 +static inline void __wakeup_vx_info(struct vx_info *vxi)
11191 +{
11192 +       if (waitqueue_active(&vxi->vx_wait))
11193 +               wake_up_interruptible(&vxi->vx_wait);
11194 +}
11195 +
11196 +
11197 +#define enter_vx_info(v, s) __enter_vx_info(v, s, __FILE__, __LINE__)
11198 +
11199 +static inline void __enter_vx_info(struct vx_info *vxi,
11200 +       struct vx_info_save *vxis, const char *_file, int _line)
11201 +{
11202 +       vxlprintk(VXD_CBIT(xid, 5), "enter_vx_info(%p[#%d],%p) %p[#%d,%p]",
11203 +               vxi, vxi ? vxi->vx_id : 0, vxis, current,
11204 +               current->xid, current->vx_info, _file, _line);
11205 +       vxis->vxi = xchg(&current->vx_info, vxi);
11206 +       vxis->xid = current->xid;
11207 +       current->xid = vxi ? vxi->vx_id : 0;
11208 +}
11209 +
11210 +#define leave_vx_info(s) __leave_vx_info(s, __FILE__, __LINE__)
11211 +
11212 +static inline void __leave_vx_info(struct vx_info_save *vxis,
11213 +       const char *_file, int _line)
11214 +{
11215 +       vxlprintk(VXD_CBIT(xid, 5), "leave_vx_info(%p[#%d,%p]) %p[#%d,%p]",
11216 +               vxis, vxis->xid, vxis->vxi, current,
11217 +               current->xid, current->vx_info, _file, _line);
11218 +       (void)xchg(&current->vx_info, vxis->vxi);
11219 +       current->xid = vxis->xid;
11220 +}
11221 +
11222 +
11223 +static inline void __enter_vx_admin(struct vx_info_save *vxis)
11224 +{
11225 +       vxis->vxi = xchg(&current->vx_info, NULL);
11226 +       vxis->xid = xchg(&current->xid, (xid_t)0);
11227 +}
11228 +
11229 +static inline void __leave_vx_admin(struct vx_info_save *vxis)
11230 +{
11231 +       (void)xchg(&current->xid, vxis->xid);
11232 +       (void)xchg(&current->vx_info, vxis->vxi);
11233 +}
11234 +
11235 +extern void exit_vx_info(struct task_struct *, int);
11236 +extern void exit_vx_info_early(struct task_struct *, int);
11237 +
11238 +
11239 +#else
11240 +#warning duplicate inclusion
11241 +#endif
11242 diff -NurpP --minimal linux-2.6.25.11/include/linux/vs_cowbl.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_cowbl.h
11243 --- linux-2.6.25.11/include/linux/vs_cowbl.h    1969-12-31 19:00:00.000000000 -0500
11244 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_cowbl.h      2008-07-08 04:31:24.000000000 -0400
11245 @@ -0,0 +1,44 @@
11246 +#ifndef _VS_COWBL_H
11247 +#define _VS_COWBL_H
11248 +
11249 +#include <linux/fs.h>
11250 +#include <linux/dcache.h>
11251 +#include <linux/namei.h>
11252 +
11253 +extern struct dentry *cow_break_link(const char *pathname);
11254 +
11255 +static inline int cow_check_and_break(struct nameidata *nd)
11256 +{
11257 +       struct inode *inode = nd->path.dentry->d_inode;
11258 +       int error = 0;
11259 +       if (IS_RDONLY(inode) || MNT_IS_RDONLY(nd->path.mnt))
11260 +               return -EROFS;
11261 +       if (IS_COW(inode)) {
11262 +               if (IS_COW_LINK(inode)) {
11263 +                       struct dentry *new_dentry, *old_dentry = nd->path.dentry;
11264 +                       char *path, *buf;
11265 +
11266 +                       buf = kmalloc(PATH_MAX, GFP_KERNEL);
11267 +                       if (!buf) {
11268 +                               return -ENOMEM;
11269 +                       }
11270 +                       path = d_path(&nd->path, buf, PATH_MAX);
11271 +                       new_dentry = cow_break_link(path);
11272 +                       kfree(buf);
11273 +                       if (!IS_ERR(new_dentry)) {
11274 +                               nd->path.dentry = new_dentry;
11275 +                               dput(old_dentry);
11276 +                       } else
11277 +                               error = PTR_ERR(new_dentry);
11278 +               } else {
11279 +                       inode->i_flags &= ~(S_IXUNLINK | S_IMMUTABLE);
11280 +                       inode->i_ctime = CURRENT_TIME;
11281 +                       mark_inode_dirty(inode);
11282 +               }
11283 +       }
11284 +       return error;
11285 +}
11286 +
11287 +#else
11288 +#warning duplicate inclusion
11289 +#endif
11290 diff -NurpP --minimal linux-2.6.25.11/include/linux/vs_cvirt.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_cvirt.h
11291 --- linux-2.6.25.11/include/linux/vs_cvirt.h    1969-12-31 19:00:00.000000000 -0500
11292 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_cvirt.h      2008-04-19 15:14:52.000000000 -0400
11293 @@ -0,0 +1,50 @@
11294 +#ifndef _VS_CVIRT_H
11295 +#define _VS_CVIRT_H
11296 +
11297 +#include "vserver/cvirt.h"
11298 +#include "vserver/context.h"
11299 +#include "vserver/base.h"
11300 +#include "vserver/check.h"
11301 +#include "vserver/debug.h"
11302 +
11303 +
11304 +static inline void vx_activate_task(struct task_struct *p)
11305 +{
11306 +       struct vx_info *vxi;
11307 +
11308 +       if ((vxi = p->vx_info)) {
11309 +               vx_update_load(vxi);
11310 +               atomic_inc(&vxi->cvirt.nr_running);
11311 +       }
11312 +}
11313 +
11314 +static inline void vx_deactivate_task(struct task_struct *p)
11315 +{
11316 +       struct vx_info *vxi;
11317 +
11318 +       if ((vxi = p->vx_info)) {
11319 +               vx_update_load(vxi);
11320 +               atomic_dec(&vxi->cvirt.nr_running);
11321 +       }
11322 +}
11323 +
11324 +static inline void vx_uninterruptible_inc(struct task_struct *p)
11325 +{
11326 +       struct vx_info *vxi;
11327 +
11328 +       if ((vxi = p->vx_info))
11329 +               atomic_inc(&vxi->cvirt.nr_uninterruptible);
11330 +}
11331 +
11332 +static inline void vx_uninterruptible_dec(struct task_struct *p)
11333 +{
11334 +       struct vx_info *vxi;
11335 +
11336 +       if ((vxi = p->vx_info))
11337 +               atomic_dec(&vxi->cvirt.nr_uninterruptible);
11338 +}
11339 +
11340 +
11341 +#else
11342 +#warning duplicate inclusion
11343 +#endif
11344 diff -NurpP --minimal linux-2.6.25.11/include/linux/vs_device.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_device.h
11345 --- linux-2.6.25.11/include/linux/vs_device.h   1969-12-31 19:00:00.000000000 -0500
11346 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_device.h     2008-04-19 15:14:52.000000000 -0400
11347 @@ -0,0 +1,45 @@
11348 +#ifndef _VS_DEVICE_H
11349 +#define _VS_DEVICE_H
11350 +
11351 +#include "vserver/base.h"
11352 +#include "vserver/device.h"
11353 +#include "vserver/debug.h"
11354 +
11355 +
11356 +#ifdef CONFIG_VSERVER_DEVICE
11357 +
11358 +int vs_map_device(struct vx_info *, dev_t, dev_t *, umode_t);
11359 +
11360 +#define vs_device_perm(v, d, m, p) \
11361 +       ((vs_map_device(current_vx_info(), d, NULL, m) & (p)) == (p))
11362 +
11363 +#else
11364 +
11365 +static inline
11366 +int vs_map_device(struct vx_info *vxi,
11367 +       dev_t device, dev_t *target, umode_t mode)
11368 +{
11369 +       if (target)
11370 +               *target = device;
11371 +       return ~0;
11372 +}
11373 +
11374 +#define vs_device_perm(v, d, m, p) ((p) == (p))
11375 +
11376 +#endif
11377 +
11378 +
11379 +#define vs_map_chrdev(d, t, p) \
11380 +       ((vs_map_device(current_vx_info(), d, t, S_IFCHR) & (p)) == (p))
11381 +#define vs_map_blkdev(d, t, p) \
11382 +       ((vs_map_device(current_vx_info(), d, t, S_IFBLK) & (p)) == (p))
11383 +
11384 +#define vs_chrdev_perm(d, p) \
11385 +       vs_device_perm(current_vx_info(), d, S_IFCHR, p)
11386 +#define vs_blkdev_perm(d, p) \
11387 +       vs_device_perm(current_vx_info(), d, S_IFBLK, p)
11388 +
11389 +
11390 +#else
11391 +#warning duplicate inclusion
11392 +#endif
11393 diff -NurpP --minimal linux-2.6.25.11/include/linux/vs_dlimit.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_dlimit.h
11394 --- linux-2.6.25.11/include/linux/vs_dlimit.h   1969-12-31 19:00:00.000000000 -0500
11395 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_dlimit.h     2008-04-19 15:14:52.000000000 -0400
11396 @@ -0,0 +1,211 @@
11397 +#ifndef _VS_DLIMIT_H
11398 +#define _VS_DLIMIT_H
11399 +
11400 +#include <linux/fs.h>
11401 +
11402 +#include "vserver/dlimit.h"
11403 +#include "vserver/base.h"
11404 +#include "vserver/debug.h"
11405 +
11406 +
11407 +#define get_dl_info(i) __get_dl_info(i, __FILE__, __LINE__)
11408 +
11409 +static inline struct dl_info *__get_dl_info(struct dl_info *dli,
11410 +       const char *_file, int _line)
11411 +{
11412 +       if (!dli)
11413 +               return NULL;
11414 +       vxlprintk(VXD_CBIT(dlim, 4), "get_dl_info(%p[#%d.%d])",
11415 +               dli, dli ? dli->dl_tag : 0,
11416 +               dli ? atomic_read(&dli->dl_usecnt) : 0,
11417 +               _file, _line);
11418 +       atomic_inc(&dli->dl_usecnt);
11419 +       return dli;
11420 +}
11421 +
11422 +
11423 +#define free_dl_info(i) \
11424 +       call_rcu(&(i)->dl_rcu, rcu_free_dl_info)
11425 +
11426 +#define put_dl_info(i) __put_dl_info(i, __FILE__, __LINE__)
11427 +
11428 +static inline void __put_dl_info(struct dl_info *dli,
11429 +       const char *_file, int _line)
11430 +{
11431 +       if (!dli)
11432 +               return;
11433 +       vxlprintk(VXD_CBIT(dlim, 4), "put_dl_info(%p[#%d.%d])",
11434 +               dli, dli ? dli->dl_tag : 0,
11435 +               dli ? atomic_read(&dli->dl_usecnt) : 0,
11436 +               _file, _line);
11437 +       if (atomic_dec_and_test(&dli->dl_usecnt))
11438 +               free_dl_info(dli);
11439 +}
11440 +
11441 +
11442 +#define __dlimit_char(d)       ((d) ? '*' : ' ')
11443 +
11444 +static inline int __dl_alloc_space(struct super_block *sb,
11445 +       tag_t tag, dlsize_t nr, const char *file, int line)
11446 +{
11447 +       struct dl_info *dli = NULL;
11448 +       int ret = 0;
11449 +
11450 +       if (nr == 0)
11451 +               goto out;
11452 +       dli = locate_dl_info(sb, tag);
11453 +       if (!dli)
11454 +               goto out;
11455 +
11456 +       spin_lock(&dli->dl_lock);
11457 +       ret = (dli->dl_space_used + nr > dli->dl_space_total);
11458 +       if (!ret)
11459 +               dli->dl_space_used += nr;
11460 +       spin_unlock(&dli->dl_lock);
11461 +       put_dl_info(dli);
11462 +out:
11463 +       vxlprintk(VXD_CBIT(dlim, 1),
11464 +               "ALLOC (%p,#%d)%c %lld bytes (%d)",
11465 +               sb, tag, __dlimit_char(dli), (long long)nr,
11466 +               ret, file, line);
11467 +       return ret;
11468 +}
11469 +
11470 +static inline void __dl_free_space(struct super_block *sb,
11471 +       tag_t tag, dlsize_t nr, const char *_file, int _line)
11472 +{
11473 +       struct dl_info *dli = NULL;
11474 +
11475 +       if (nr == 0)
11476 +               goto out;
11477 +       dli = locate_dl_info(sb, tag);
11478 +       if (!dli)
11479 +               goto out;
11480 +
11481 +       spin_lock(&dli->dl_lock);
11482 +       if (dli->dl_space_used > nr)
11483 +               dli->dl_space_used -= nr;
11484 +       else
11485 +               dli->dl_space_used = 0;
11486 +       spin_unlock(&dli->dl_lock);
11487 +       put_dl_info(dli);
11488 +out:
11489 +       vxlprintk(VXD_CBIT(dlim, 1),
11490 +               "FREE  (%p,#%d)%c %lld bytes",
11491 +               sb, tag, __dlimit_char(dli), (long long)nr,
11492 +               _file, _line);
11493 +}
11494 +
11495 +static inline int __dl_alloc_inode(struct super_block *sb,
11496 +       tag_t tag, const char *_file, int _line)
11497 +{
11498 +       struct dl_info *dli;
11499 +       int ret = 0;
11500 +
11501 +       dli = locate_dl_info(sb, tag);
11502 +       if (!dli)
11503 +               goto out;
11504 +
11505 +       spin_lock(&dli->dl_lock);
11506 +       ret = (dli->dl_inodes_used >= dli->dl_inodes_total);
11507 +       if (!ret)
11508 +               dli->dl_inodes_used++;
11509 +       spin_unlock(&dli->dl_lock);
11510 +       put_dl_info(dli);
11511 +out:
11512 +       vxlprintk(VXD_CBIT(dlim, 0),
11513 +               "ALLOC (%p,#%d)%c inode (%d)",
11514 +               sb, tag, __dlimit_char(dli), ret, _file, _line);
11515 +       return ret;
11516 +}
11517 +
11518 +static inline void __dl_free_inode(struct super_block *sb,
11519 +       tag_t tag, const char *_file, int _line)
11520 +{
11521 +       struct dl_info *dli;
11522 +
11523 +       dli = locate_dl_info(sb, tag);
11524 +       if (!dli)
11525 +               goto out;
11526 +
11527 +       spin_lock(&dli->dl_lock);
11528 +       if (dli->dl_inodes_used > 1)
11529 +               dli->dl_inodes_used--;
11530 +       else
11531 +               dli->dl_inodes_used = 0;
11532 +       spin_unlock(&dli->dl_lock);
11533 +       put_dl_info(dli);
11534 +out:
11535 +       vxlprintk(VXD_CBIT(dlim, 0),
11536 +               "FREE  (%p,#%d)%c inode",
11537 +               sb, tag, __dlimit_char(dli), _file, _line);
11538 +}
11539 +
11540 +static inline void __dl_adjust_block(struct super_block *sb, tag_t tag,
11541 +       unsigned long long *free_blocks, unsigned long long *root_blocks,
11542 +       const char *_file, int _line)
11543 +{
11544 +       struct dl_info *dli;
11545 +       uint64_t broot, bfree;
11546 +
11547 +       dli = locate_dl_info(sb, tag);
11548 +       if (!dli)
11549 +               return;
11550 +
11551 +       spin_lock(&dli->dl_lock);
11552 +       broot = (dli->dl_space_total -
11553 +               (dli->dl_space_total >> 10) * dli->dl_nrlmult)
11554 +               >> sb->s_blocksize_bits;
11555 +       bfree = (dli->dl_space_total - dli->dl_space_used)
11556 +                       >> sb->s_blocksize_bits;
11557 +       spin_unlock(&dli->dl_lock);
11558 +
11559 +       vxlprintk(VXD_CBIT(dlim, 2),
11560 +               "ADJUST: %lld,%lld on %lld,%lld [mult=%d]",
11561 +               (long long)bfree, (long long)broot,
11562 +               *free_blocks, *root_blocks, dli->dl_nrlmult,
11563 +               _file, _line);
11564 +       if (free_blocks) {
11565 +               if (*free_blocks > bfree)
11566 +                       *free_blocks = bfree;
11567 +       }
11568 +       if (root_blocks) {
11569 +               if (*root_blocks > broot)
11570 +                       *root_blocks = broot;
11571 +       }
11572 +       put_dl_info(dli);
11573 +}
11574 +
11575 +#define DLIMIT_ALLOC_SPACE(in, bytes) \
11576 +       __dl_alloc_space((in)->i_sb, (in)->i_tag, (dlsize_t)(bytes), \
11577 +               __FILE__, __LINE__ )
11578 +
11579 +#define DLIMIT_FREE_SPACE(in, bytes) \
11580 +       __dl_free_space((in)->i_sb, (in)->i_tag, (dlsize_t)(bytes), \
11581 +               __FILE__, __LINE__ )
11582 +
11583 +#define DLIMIT_ALLOC_BLOCK(in, nr) \
11584 +       __dl_alloc_space((in)->i_sb, (in)->i_tag, \
11585 +               ((dlsize_t)(nr)) << (in)->i_sb->s_blocksize_bits, \
11586 +               __FILE__, __LINE__ )
11587 +
11588 +#define DLIMIT_FREE_BLOCK(in, nr) \
11589 +       __dl_free_space((in)->i_sb, (in)->i_tag, \
11590 +               ((dlsize_t)(nr)) << (in)->i_sb->s_blocksize_bits, \
11591 +               __FILE__, __LINE__ )
11592 +
11593 +
11594 +#define DLIMIT_ALLOC_INODE(in) \
11595 +       __dl_alloc_inode((in)->i_sb, (in)->i_tag, __FILE__, __LINE__ )
11596 +
11597 +#define DLIMIT_FREE_INODE(in) \
11598 +       __dl_free_inode((in)->i_sb, (in)->i_tag, __FILE__, __LINE__ )
11599 +
11600 +
11601 +#define DLIMIT_ADJUST_BLOCK(sb, tag, fb, rb) \
11602 +       __dl_adjust_block(sb, tag, fb, rb, __FILE__, __LINE__ )
11603 +
11604 +
11605 +#else
11606 +#warning duplicate inclusion
11607 +#endif
11608 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/base.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/base.h
11609 --- linux-2.6.25.11/include/linux/vserver/base.h        1969-12-31 19:00:00.000000000 -0500
11610 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/base.h  2008-04-23 20:52:31.000000000 -0400
11611 @@ -0,0 +1,157 @@
11612 +#ifndef _VX_BASE_H
11613 +#define _VX_BASE_H
11614 +
11615 +
11616 +/* context state changes */
11617 +
11618 +enum {
11619 +       VSC_STARTUP = 1,
11620 +       VSC_SHUTDOWN,
11621 +
11622 +       VSC_NETUP,
11623 +       VSC_NETDOWN,
11624 +};
11625 +
11626 +
11627 +
11628 +#define vx_task_xid(t) ((t)->xid)
11629 +
11630 +#define vx_current_xid() vx_task_xid(current)
11631 +
11632 +#define current_vx_info() (current->vx_info)
11633 +
11634 +
11635 +#define nx_task_nid(t) ((t)->nid)
11636 +
11637 +#define nx_current_nid() nx_task_nid(current)
11638 +
11639 +#define current_nx_info() (current->nx_info)
11640 +
11641 +
11642 +/* generic flag merging */
11643 +
11644 +#define vs_check_flags(v, m, f)        (((v) & (m)) ^ (f))
11645 +
11646 +#define vs_mask_flags(v, f, m) (((v) & ~(m)) | ((f) & (m)))
11647 +
11648 +#define vs_mask_mask(v, f, m)  (((v) & ~(m)) | ((v) & (f) & (m)))
11649 +
11650 +#define vs_check_bit(v, n)     ((v) & (1LL << (n)))
11651 +
11652 +
11653 +/* context flags */
11654 +
11655 +#define __vx_flags(v)  ((v) ? (v)->vx_flags : 0)
11656 +
11657 +#define vx_current_flags()     __vx_flags(current->vx_info)
11658 +
11659 +#define vx_info_flags(v, m, f) \
11660 +       vs_check_flags(__vx_flags(v), m, f)
11661 +
11662 +#define task_vx_flags(t, m, f) \
11663 +       ((t) && vx_info_flags((t)->vx_info, m, f))
11664 +
11665 +#define vx_flags(m, f) vx_info_flags(current->vx_info, m, f)
11666 +
11667 +
11668 +/* context caps */
11669 +
11670 +#define __vx_ccaps(v)  ((v) ? (v)->vx_ccaps : 0)
11671 +
11672 +#define vx_current_ccaps()     __vx_ccaps(current->vx_info)
11673 +
11674 +#define vx_info_ccaps(v, c)    (__vx_ccaps(v) & (c))
11675 +
11676 +#define vx_ccaps(c)    vx_info_ccaps(current->vx_info, (c))
11677 +
11678 +
11679 +
11680 +/* network flags */
11681 +
11682 +#define __nx_flags(n)  ((n) ? (n)->nx_flags : 0)
11683 +
11684 +#define nx_current_flags()     __nx_flags(current->nx_info)
11685 +
11686 +#define nx_info_flags(n, m, f) \
11687 +       vs_check_flags(__nx_flags(n), m, f)
11688 +
11689 +#define task_nx_flags(t, m, f) \
11690 +       ((t) && nx_info_flags((t)->nx_info, m, f))
11691 +
11692 +#define nx_flags(m, f) nx_info_flags(current->nx_info, m, f)
11693 +
11694 +
11695 +/* network caps */
11696 +
11697 +#define __nx_ncaps(n)  ((n) ? (n)->nx_ncaps : 0)
11698 +
11699 +#define nx_current_ncaps()     __nx_ncaps(current->nx_info)
11700 +
11701 +#define nx_info_ncaps(n, c)    (__nx_ncaps(n) & (c))
11702 +
11703 +#define nx_ncaps(c)    nx_info_ncaps(current->nx_info, c)
11704 +
11705 +
11706 +/* context mask capabilities */
11707 +
11708 +#define __vx_mcaps(v)  ((v) ? (v)->vx_ccaps >> 32UL : ~0 )
11709 +
11710 +#define vx_info_mcaps(v, c)    (__vx_mcaps(v) & (c))
11711 +
11712 +#define vx_mcaps(c)    vx_info_mcaps(current->vx_info, c)
11713 +
11714 +
11715 +/* context bcap mask */
11716 +
11717 +#define __vx_bcaps(v)          ((v)->vx_bcaps)
11718 +
11719 +#define vx_current_bcaps()     __vx_bcaps(current->vx_info)
11720 +
11721 +
11722 +/* mask given bcaps */
11723 +
11724 +#define vx_info_mbcaps(v, c)   ((v) ? cap_intersect(__vx_bcaps(v), c) : c)
11725 +
11726 +#define vx_mbcaps(c)           vx_info_mbcaps(current->vx_info, c)
11727 +
11728 +
11729 +/* masked cap_bset */
11730 +
11731 +#define vx_info_cap_bset(v)    vx_info_mbcaps(v, current->cap_bset)
11732 +
11733 +#define vx_current_cap_bset()  vx_info_cap_bset(current->vx_info)
11734 +
11735 +#if 0
11736 +#define vx_info_mbcap(v, b) \
11737 +       (!vx_info_flags(v, VXF_STATE_SETUP, 0) ? \
11738 +       vx_info_bcaps(v, b) : (b))
11739 +
11740 +#define task_vx_mbcap(t, b) \
11741 +       vx_info_mbcap((t)->vx_info, (t)->b)
11742 +
11743 +#define vx_mbcap(b)    task_vx_mbcap(current, b)
11744 +#endif
11745 +
11746 +#define vx_cap_raised(v, c, f) cap_raised(vx_info_mbcaps(v, c), f)
11747 +
11748 +#define vx_capable(b, c) (capable(b) || \
11749 +       (cap_raised(current->cap_effective, b) && vx_ccaps(c)))
11750 +
11751 +#define nx_capable(b, c) (capable(b) || \
11752 +       (cap_raised(current->cap_effective, b) && nx_ncaps(c)))
11753 +
11754 +#define vx_current_initpid(n) \
11755 +       (current->vx_info && \
11756 +       (current->vx_info->vx_initpid == (n)))
11757 +
11758 +
11759 +#define __vx_state(v)  ((v) ? ((v)->vx_state) : 0)
11760 +
11761 +#define vx_info_state(v, m)    (__vx_state(v) & (m))
11762 +
11763 +
11764 +#define __nx_state(n)  ((n) ? ((n)->nx_state) : 0)
11765 +
11766 +#define nx_info_state(n, m)    (__nx_state(n) & (m))
11767 +
11768 +#endif
11769 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/cacct_cmd.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/cacct_cmd.h
11770 --- linux-2.6.25.11/include/linux/vserver/cacct_cmd.h   1969-12-31 19:00:00.000000000 -0500
11771 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/cacct_cmd.h     2008-04-19 15:14:52.000000000 -0400
11772 @@ -0,0 +1,23 @@
11773 +#ifndef _VX_CACCT_CMD_H
11774 +#define _VX_CACCT_CMD_H
11775 +
11776 +
11777 +/* virtual host info name commands */
11778 +
11779 +#define VCMD_sock_stat         VC_CMD(VSTAT, 5, 0)
11780 +
11781 +struct vcmd_sock_stat_v0 {
11782 +       uint32_t field;
11783 +       uint32_t count[3];
11784 +       uint64_t total[3];
11785 +};
11786 +
11787 +
11788 +#ifdef __KERNEL__
11789 +
11790 +#include <linux/compiler.h>
11791 +
11792 +extern int vc_sock_stat(struct vx_info *, void __user *);
11793 +
11794 +#endif /* __KERNEL__ */
11795 +#endif /* _VX_CACCT_CMD_H */
11796 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/cacct_def.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/cacct_def.h
11797 --- linux-2.6.25.11/include/linux/vserver/cacct_def.h   1969-12-31 19:00:00.000000000 -0500
11798 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/cacct_def.h     2008-04-19 15:14:52.000000000 -0400
11799 @@ -0,0 +1,43 @@
11800 +#ifndef _VX_CACCT_DEF_H
11801 +#define _VX_CACCT_DEF_H
11802 +
11803 +#include <asm/atomic.h>
11804 +#include <linux/vserver/cacct.h>
11805 +
11806 +
11807 +struct _vx_sock_acc {
11808 +       atomic_long_t count;
11809 +       atomic_long_t total;
11810 +};
11811 +
11812 +/* context sub struct */
11813 +
11814 +struct _vx_cacct {
11815 +       struct _vx_sock_acc sock[VXA_SOCK_SIZE][3];
11816 +       atomic_t slab[8];
11817 +       atomic_t page[6][8];
11818 +};
11819 +
11820 +#ifdef CONFIG_VSERVER_DEBUG
11821 +
11822 +static inline void __dump_vx_cacct(struct _vx_cacct *cacct)
11823 +{
11824 +       int i, j;
11825 +
11826 +       printk("\t_vx_cacct:");
11827 +       for (i = 0; i < 6; i++) {
11828 +               struct _vx_sock_acc *ptr = cacct->sock[i];
11829 +
11830 +               printk("\t [%d] =", i);
11831 +               for (j = 0; j < 3; j++) {
11832 +                       printk(" [%d] = %8lu, %8lu", j,
11833 +                               atomic_long_read(&ptr[j].count),
11834 +                               atomic_long_read(&ptr[j].total));
11835 +               }
11836 +               printk("\n");
11837 +       }
11838 +}
11839 +
11840 +#endif
11841 +
11842 +#endif /* _VX_CACCT_DEF_H */
11843 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/cacct.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/cacct.h
11844 --- linux-2.6.25.11/include/linux/vserver/cacct.h       1969-12-31 19:00:00.000000000 -0500
11845 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/cacct.h 2008-04-19 15:14:52.000000000 -0400
11846 @@ -0,0 +1,15 @@
11847 +#ifndef _VX_CACCT_H
11848 +#define _VX_CACCT_H
11849 +
11850 +
11851 +enum sock_acc_field {
11852 +       VXA_SOCK_UNSPEC = 0,
11853 +       VXA_SOCK_UNIX,
11854 +       VXA_SOCK_INET,
11855 +       VXA_SOCK_INET6,
11856 +       VXA_SOCK_PACKET,
11857 +       VXA_SOCK_OTHER,
11858 +       VXA_SOCK_SIZE   /* array size */
11859 +};
11860 +
11861 +#endif /* _VX_CACCT_H */
11862 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/cacct_int.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/cacct_int.h
11863 --- linux-2.6.25.11/include/linux/vserver/cacct_int.h   1969-12-31 19:00:00.000000000 -0500
11864 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/cacct_int.h     2008-04-19 15:14:52.000000000 -0400
11865 @@ -0,0 +1,21 @@
11866 +#ifndef _VX_CACCT_INT_H
11867 +#define _VX_CACCT_INT_H
11868 +
11869 +
11870 +#ifdef __KERNEL__
11871 +
11872 +static inline
11873 +unsigned long vx_sock_count(struct _vx_cacct *cacct, int type, int pos)
11874 +{
11875 +       return atomic_long_read(&cacct->sock[type][pos].count);
11876 +}
11877 +
11878 +
11879 +static inline
11880 +unsigned long vx_sock_total(struct _vx_cacct *cacct, int type, int pos)
11881 +{
11882 +       return atomic_long_read(&cacct->sock[type][pos].total);
11883 +}
11884 +
11885 +#endif /* __KERNEL__ */
11886 +#endif /* _VX_CACCT_INT_H */
11887 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/check.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/check.h
11888 --- linux-2.6.25.11/include/linux/vserver/check.h       1969-12-31 19:00:00.000000000 -0500
11889 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/check.h 2008-04-19 15:14:52.000000000 -0400
11890 @@ -0,0 +1,89 @@
11891 +#ifndef _VS_CHECK_H
11892 +#define _VS_CHECK_H
11893 +
11894 +
11895 +#define MAX_S_CONTEXT  65535   /* Arbitrary limit */
11896 +
11897 +#ifdef CONFIG_VSERVER_DYNAMIC_IDS
11898 +#define MIN_D_CONTEXT  49152   /* dynamic contexts start here */
11899 +#else
11900 +#define MIN_D_CONTEXT  65536
11901 +#endif
11902 +
11903 +/* check conditions */
11904 +
11905 +#define VS_ADMIN       0x0001
11906 +#define VS_WATCH       0x0002
11907 +#define VS_HIDE                0x0004
11908 +#define VS_HOSTID      0x0008
11909 +
11910 +#define VS_IDENT       0x0010
11911 +#define VS_EQUIV       0x0020
11912 +#define VS_PARENT      0x0040
11913 +#define VS_CHILD       0x0080
11914 +
11915 +#define VS_ARG_MASK    0x00F0
11916 +
11917 +#define VS_DYNAMIC     0x0100
11918 +#define VS_STATIC      0x0200
11919 +
11920 +#define VS_ATR_MASK    0x0F00
11921 +
11922 +#ifdef CONFIG_VSERVER_PRIVACY
11923 +#define VS_ADMIN_P     (0)
11924 +#define VS_WATCH_P     (0)
11925 +#else
11926 +#define VS_ADMIN_P     VS_ADMIN
11927 +#define VS_WATCH_P     VS_WATCH
11928 +#endif
11929 +
11930 +#define VS_HARDIRQ     0x1000
11931 +#define VS_SOFTIRQ     0x2000
11932 +#define VS_IRQ         0x4000
11933 +
11934 +#define VS_IRQ_MASK    0xF000
11935 +
11936 +#include <linux/hardirq.h>
11937 +
11938 +/*
11939 + * check current context for ADMIN/WATCH and
11940 + * optionally against supplied argument
11941 + */
11942 +static inline int __vs_check(int cid, int id, unsigned int mode)
11943 +{
11944 +       if (mode & VS_ARG_MASK) {
11945 +               if ((mode & VS_IDENT) && (id == cid))
11946 +                       return 1;
11947 +       }
11948 +       if (mode & VS_ATR_MASK) {
11949 +               if ((mode & VS_DYNAMIC) &&
11950 +                       (id >= MIN_D_CONTEXT) &&
11951 +                       (id <= MAX_S_CONTEXT))
11952 +                       return 1;
11953 +               if ((mode & VS_STATIC) &&
11954 +                       (id > 1) && (id < MIN_D_CONTEXT))
11955 +                       return 1;
11956 +       }
11957 +       if (mode & VS_IRQ_MASK) {
11958 +               if ((mode & VS_IRQ) && unlikely(in_interrupt()))
11959 +                       return 1;
11960 +               if ((mode & VS_HARDIRQ) && unlikely(in_irq()))
11961 +                       return 1;
11962 +               if ((mode & VS_SOFTIRQ) && unlikely(in_softirq()))
11963 +                       return 1;
11964 +       }
11965 +       return (((mode & VS_ADMIN) && (cid == 0)) ||
11966 +               ((mode & VS_WATCH) && (cid == 1)) ||
11967 +               ((mode & VS_HOSTID) && (id == 0)));
11968 +}
11969 +
11970 +#define vx_check(c, m) __vs_check(vx_current_xid(), c, (m) | VS_IRQ)
11971 +
11972 +#define vx_weak_check(c, m)    ((m) ? vx_check(c, m) : 1)
11973 +
11974 +
11975 +#define nx_check(c, m) __vs_check(nx_current_nid(), c, m)
11976 +
11977 +#define nx_weak_check(c, m)    ((m) ? nx_check(c, m) : 1)
11978 +
11979 +#endif
11980 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/context_cmd.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/context_cmd.h
11981 --- linux-2.6.25.11/include/linux/vserver/context_cmd.h 1969-12-31 19:00:00.000000000 -0500
11982 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/context_cmd.h   2008-04-19 15:14:52.000000000 -0400
11983 @@ -0,0 +1,128 @@
11984 +#ifndef _VX_CONTEXT_CMD_H
11985 +#define _VX_CONTEXT_CMD_H
11986 +
11987 +
11988 +/* vinfo commands */
11989 +
11990 +#define VCMD_task_xid          VC_CMD(VINFO, 1, 0)
11991 +
11992 +#ifdef __KERNEL__
11993 +extern int vc_task_xid(uint32_t);
11994 +
11995 +#endif /* __KERNEL__ */
11996 +
11997 +#define VCMD_vx_info           VC_CMD(VINFO, 5, 0)
11998 +
11999 +struct vcmd_vx_info_v0 {
12000 +       uint32_t xid;
12001 +       uint32_t initpid;
12002 +       /* more to come */
12003 +};
12004 +
12005 +#ifdef __KERNEL__
12006 +extern int vc_vx_info(struct vx_info *, void __user *);
12007 +
12008 +#endif /* __KERNEL__ */
12009 +
12010 +#define VCMD_ctx_stat          VC_CMD(VSTAT, 0, 0)
12011 +
12012 +struct vcmd_ctx_stat_v0 {
12013 +       uint32_t usecnt;
12014 +       uint32_t tasks;
12015 +       /* more to come */
12016 +};
12017 +
12018 +#ifdef __KERNEL__
12019 +extern int vc_ctx_stat(struct vx_info *, void __user *);
12020 +
12021 +#endif /* __KERNEL__ */
12022 +
12023 +/* context commands */
12024 +
12025 +#define VCMD_ctx_create_v0     VC_CMD(VPROC, 1, 0)
12026 +#define VCMD_ctx_create                VC_CMD(VPROC, 1, 1)
12027 +
12028 +struct vcmd_ctx_create {
12029 +       uint64_t flagword;
12030 +};
12031 +
12032 +#define VCMD_ctx_migrate_v0    VC_CMD(PROCMIG, 1, 0)
12033 +#define VCMD_ctx_migrate       VC_CMD(PROCMIG, 1, 1)
12034 +
12035 +struct vcmd_ctx_migrate {
12036 +       uint64_t flagword;
12037 +};
12038 +
12039 +#ifdef __KERNEL__
12040 +extern int vc_ctx_create(uint32_t, void __user *);
12041 +extern int vc_ctx_migrate(struct vx_info *, void __user *);
12042 +
12043 +#endif /* __KERNEL__ */
12044 +
12045 +
12046 +/* flag commands */
12047 +
12048 +#define VCMD_get_cflags                VC_CMD(FLAGS, 1, 0)
12049 +#define VCMD_set_cflags                VC_CMD(FLAGS, 2, 0)
12050 +
12051 +struct vcmd_ctx_flags_v0 {
12052 +       uint64_t flagword;
12053 +       uint64_t mask;
12054 +};
12055 +
12056 +#ifdef __KERNEL__
12057 +extern int vc_get_cflags(struct vx_info *, void __user *);
12058 +extern int vc_set_cflags(struct vx_info *, void __user *);
12059 +
12060 +#endif /* __KERNEL__ */
12061 +
12062 +
12063 +/* context caps commands */
12064 +
12065 +#define VCMD_get_ccaps         VC_CMD(FLAGS, 3, 1)
12066 +#define VCMD_set_ccaps         VC_CMD(FLAGS, 4, 1)
12067 +
12068 +struct vcmd_ctx_caps_v1 {
12069 +       uint64_t ccaps;
12070 +       uint64_t cmask;
12071 +};
12072 +
12073 +#ifdef __KERNEL__
12074 +extern int vc_get_ccaps(struct vx_info *, void __user *);
12075 +extern int vc_set_ccaps(struct vx_info *, void __user *);
12076 +
12077 +#endif /* __KERNEL__ */
12078 +
12079 +
12080 +/* bcaps commands */
12081 +
12082 +#define VCMD_get_bcaps         VC_CMD(FLAGS, 9, 0)
12083 +#define VCMD_set_bcaps         VC_CMD(FLAGS, 10, 0)
12084 +
12085 +struct vcmd_bcaps {
12086 +       uint64_t bcaps;
12087 +       uint64_t bmask;
12088 +};
12089 +
12090 +#ifdef __KERNEL__
12091 +extern int vc_get_bcaps(struct vx_info *, void __user *);
12092 +extern int vc_set_bcaps(struct vx_info *, void __user *);
12093 +
12094 +#endif /* __KERNEL__ */
12095 +
12096 +
12097 +/* OOM badness */
12098 +
12099 +#define VCMD_get_badness       VC_CMD(MEMCTRL, 5, 0)
12100 +#define VCMD_set_badness       VC_CMD(MEMCTRL, 6, 0)
12101 +
12102 +struct vcmd_badness_v0 {
12103 +       int64_t bias;
12104 +};
12105 +
12106 +#ifdef __KERNEL__
12107 +extern int vc_get_badness(struct vx_info *, void __user *);
12108 +extern int vc_set_badness(struct vx_info *, void __user *);
12109 +
12110 +#endif /* __KERNEL__ */
12111 +#endif /* _VX_CONTEXT_CMD_H */
12112 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/context.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/context.h
12113 --- linux-2.6.25.11/include/linux/vserver/context.h     1969-12-31 19:00:00.000000000 -0500
12114 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/context.h       2008-04-22 15:14:28.000000000 -0400
12115 @@ -0,0 +1,176 @@
12116 +#ifndef _VX_CONTEXT_H
12117 +#define _VX_CONTEXT_H
12118 +
12119 +#include <linux/types.h>
12120 +#include <linux/capability.h>
12121 +
12122 +
12123 +/* context flags */
12124 +
12125 +#define VXF_INFO_SCHED         0x00000002
12126 +#define VXF_INFO_NPROC         0x00000004
12127 +#define VXF_INFO_PRIVATE       0x00000008
12128 +
12129 +#define VXF_INFO_INIT          0x00000010
12130 +#define VXF_INFO_HIDE          0x00000020
12131 +#define VXF_INFO_ULIMIT                0x00000040
12132 +#define VXF_INFO_NSPACE                0x00000080
12133 +
12134 +#define VXF_SCHED_HARD         0x00000100
12135 +#define VXF_SCHED_PRIO         0x00000200
12136 +#define VXF_SCHED_PAUSE                0x00000400
12137 +
12138 +#define VXF_VIRT_MEM           0x00010000
12139 +#define VXF_VIRT_UPTIME                0x00020000
12140 +#define VXF_VIRT_CPU           0x00040000
12141 +#define VXF_VIRT_LOAD          0x00080000
12142 +#define VXF_VIRT_TIME          0x00100000
12143 +
12144 +#define VXF_HIDE_MOUNT         0x01000000
12145 +/* was VXF_HIDE_NETIF          0x02000000 */
12146 +#define VXF_HIDE_VINFO         0x04000000
12147 +
12148 +#define VXF_STATE_SETUP                (1ULL << 32)
12149 +#define VXF_STATE_INIT         (1ULL << 33)
12150 +#define VXF_STATE_ADMIN                (1ULL << 34)
12151 +
12152 +#define VXF_SC_HELPER          (1ULL << 36)
12153 +#define VXF_REBOOT_KILL                (1ULL << 37)
12154 +#define VXF_PERSISTENT         (1ULL << 38)
12155 +
12156 +#define VXF_FORK_RSS           (1ULL << 48)
12157 +#define VXF_PROLIFIC           (1ULL << 49)
12158 +
12159 +#define VXF_IGNEG_NICE         (1ULL << 52)
12160 +
12161 +#define VXF_ONE_TIME           (0x0007ULL << 32)
12162 +
12163 +#define VXF_INIT_SET           (VXF_STATE_SETUP | VXF_STATE_INIT | VXF_STATE_ADMIN)
12164 +
12165 +
12166 +/* context migration */
12167 +
12168 +#define VXM_SET_INIT           0x00000001
12169 +#define VXM_SET_REAPER         0x00000002
12170 +
12171 +/* context caps */
12172 +
12173 +#define VXC_CAP_MASK           0x00000000
12174 +
12175 +#define VXC_SET_UTSNAME                0x00000001
12176 +#define VXC_SET_RLIMIT         0x00000002
12177 +
12178 +/* was VXC_RAW_ICMP            0x00000100 */
12179 +#define VXC_SYSLOG             0x00001000
12180 +
12181 +#define VXC_SECURE_MOUNT       0x00010000
12182 +#define VXC_SECURE_REMOUNT     0x00020000
12183 +#define VXC_BINARY_MOUNT       0x00040000
12184 +
12185 +#define VXC_QUOTA_CTL          0x00100000
12186 +#define VXC_ADMIN_MAPPER       0x00200000
12187 +#define VXC_ADMIN_CLOOP                0x00400000
12188 +
12189 +#define VXC_KTHREAD            0x01000000
12190 +
12191 +
12192 +#ifdef __KERNEL__
12193 +
12194 +#include <linux/list.h>
12195 +#include <linux/spinlock.h>
12196 +#include <linux/rcupdate.h>
12197 +
12198 +#include "limit_def.h"
12199 +#include "sched_def.h"
12200 +#include "cvirt_def.h"
12201 +#include "cacct_def.h"
12202 +#include "device_def.h"
12203 +
12204 +struct _vx_info_pc {
12205 +       struct _vx_sched_pc sched_pc;
12206 +       struct _vx_cvirt_pc cvirt_pc;
12207 +};
12208 +
12209 +struct vx_info {
12210 +       struct hlist_node vx_hlist;             /* linked list of contexts */
12211 +       xid_t vx_id;                            /* context id */
12212 +       atomic_t vx_usecnt;                     /* usage count */
12213 +       atomic_t vx_tasks;                      /* tasks count */
12214 +       struct vx_info *vx_parent;              /* parent context */
12215 +       int vx_state;                           /* context state */
12216 +
12217 +       unsigned long vx_nsmask;                /* assignment mask */
12218 +       struct nsproxy *vx_nsproxy;             /* private namespace */
12219 +       struct fs_struct *vx_fs;                /* private namespace fs */
12220 +
12221 +       uint64_t vx_flags;                      /* context flags */
12222 +       uint64_t vx_ccaps;                      /* context caps (vserver) */
12223 +       kernel_cap_t vx_bcaps;                  /* bounding caps (system) */
12224 +       // kernel_cap_t vx_cap_bset;            /* the guest's bset */
12225 +
12226 +       struct task_struct *vx_reaper;          /* guest reaper process */
12227 +       pid_t vx_initpid;                       /* PID of guest init */
12228 +       int64_t vx_badness_bias;                /* OOM points bias */
12229 +
12230 +       struct _vx_limit limit;                 /* vserver limits */
12231 +       struct _vx_sched sched;                 /* vserver scheduler */
12232 +       struct _vx_cvirt cvirt;                 /* virtual/bias stuff */
12233 +       struct _vx_cacct cacct;                 /* context accounting */
12234 +
12235 +       struct _vx_device dmap;                 /* default device map targets */
12236 +
12237 +#ifndef CONFIG_SMP
12238 +       struct _vx_info_pc info_pc;             /* per cpu data */
12239 +#else
12240 +       struct _vx_info_pc *ptr_pc;             /* per cpu array */
12241 +#endif
12242 +
12243 +       wait_queue_head_t vx_wait;              /* context exit waitqueue */
12244 +       int reboot_cmd;                         /* last sys_reboot() cmd */
12245 +       int exit_code;                          /* last process exit code */
12246 +
12247 +       char vx_name[65];                       /* vserver name */
12248 +};
12249 +
12250 +#ifndef CONFIG_SMP
12251 +#define        vx_ptr_pc(vxi)          (&(vxi)->info_pc)
12252 +#define        vx_per_cpu(vxi, v, id)  vx_ptr_pc(vxi)->v
12253 +#else
12254 +#define        vx_ptr_pc(vxi)          ((vxi)->ptr_pc)
12255 +#define        vx_per_cpu(vxi, v, id)  per_cpu_ptr(vx_ptr_pc(vxi), id)->v
12256 +#endif
12257 +
12258 +#define        vx_cpu(vxi, v)          vx_per_cpu(vxi, v, smp_processor_id())
12259 +
12260 +
12261 +struct vx_info_save {
12262 +       struct vx_info *vxi;
12263 +       xid_t xid;
12264 +};
12265 +
12266 +
12267 +/* status flags */
12268 +
12269 +#define VXS_HASHED     0x0001
12270 +#define VXS_PAUSED     0x0010
12271 +#define VXS_SHUTDOWN   0x0100
12272 +#define VXS_HELPER     0x1000
12273 +#define VXS_RELEASED   0x8000
12274 +
12275 +
12276 +extern void claim_vx_info(struct vx_info *, struct task_struct *);
12277 +extern void release_vx_info(struct vx_info *, struct task_struct *);
12278 +
12279 +extern struct vx_info *lookup_vx_info(int);
12280 +extern struct vx_info *lookup_or_create_vx_info(int);
12281 +
12282 +extern int get_xid_list(int, unsigned int *, int);
12283 +extern int xid_is_hashed(xid_t);
12284 +
12285 +extern int vx_migrate_task(struct task_struct *, struct vx_info *, int);
12286 +
12287 +extern long vs_state_change(struct vx_info *, unsigned int);
12288 +
12289 +
12290 +#endif /* __KERNEL__ */
12291 +#endif /* _VX_CONTEXT_H */
12292 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/cvirt_cmd.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/cvirt_cmd.h
12293 --- linux-2.6.25.11/include/linux/vserver/cvirt_cmd.h   1969-12-31 19:00:00.000000000 -0500
12294 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/cvirt_cmd.h     2008-04-19 15:14:52.000000000 -0400
12295 @@ -0,0 +1,53 @@
12296 +#ifndef _VX_CVIRT_CMD_H
12297 +#define _VX_CVIRT_CMD_H
12298 +
12299 +
12300 +/* virtual host info name commands */
12301 +
12302 +#define VCMD_set_vhi_name      VC_CMD(VHOST, 1, 0)
12303 +#define VCMD_get_vhi_name      VC_CMD(VHOST, 2, 0)
12304 +
12305 +struct vcmd_vhi_name_v0 {
12306 +       uint32_t field;
12307 +       char name[65];
12308 +};
12309 +
12310 +
12311 +enum vhi_name_field {
12312 +       VHIN_CONTEXT = 0,
12313 +       VHIN_SYSNAME,
12314 +       VHIN_NODENAME,
12315 +       VHIN_RELEASE,
12316 +       VHIN_VERSION,
12317 +       VHIN_MACHINE,
12318 +       VHIN_DOMAINNAME,
12319 +};
12320 +
12321 +
12322 +#ifdef __KERNEL__
12323 +
12324 +#include <linux/compiler.h>
12325 +
12326 +extern int vc_set_vhi_name(struct vx_info *, void __user *);
12327 +extern int vc_get_vhi_name(struct vx_info *, void __user *);
12328 +
12329 +#endif /* __KERNEL__ */
12330 +
12331 +#define VCMD_virt_stat         VC_CMD(VSTAT, 3, 0)
12332 +
12333 +struct vcmd_virt_stat_v0 {
12334 +       uint64_t offset;
12335 +       uint64_t uptime;
12336 +       uint32_t nr_threads;
12337 +       uint32_t nr_running;
12338 +       uint32_t nr_uninterruptible;
12339 +       uint32_t nr_onhold;
12340 +       uint32_t nr_forks;
12341 +       uint32_t load[3];
12342 +};
12343 +
12344 +#ifdef __KERNEL__
12345 +extern int vc_virt_stat(struct vx_info *, void __user *);
12346 +
12347 +#endif /* __KERNEL__ */
12348 +#endif /* _VX_CVIRT_CMD_H */
12349 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/cvirt_def.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/cvirt_def.h
12350 --- linux-2.6.25.11/include/linux/vserver/cvirt_def.h   1969-12-31 19:00:00.000000000 -0500
12351 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/cvirt_def.h     2008-04-19 15:14:52.000000000 -0400
12352 @@ -0,0 +1,80 @@
12353 +#ifndef _VX_CVIRT_DEF_H
12354 +#define _VX_CVIRT_DEF_H
12355 +
12356 +#include <linux/jiffies.h>
12357 +#include <linux/spinlock.h>
12358 +#include <linux/wait.h>
12359 +#include <linux/time.h>
12360 +#include <asm/atomic.h>
12361 +
12362 +
12363 +struct _vx_usage_stat {
12364 +       uint64_t user;
12365 +       uint64_t nice;
12366 +       uint64_t system;
12367 +       uint64_t softirq;
12368 +       uint64_t irq;
12369 +       uint64_t idle;
12370 +       uint64_t iowait;
12371 +};
12372 +
12373 +struct _vx_syslog {
12374 +       wait_queue_head_t log_wait;
12375 +       spinlock_t logbuf_lock;         /* lock for the log buffer */
12376 +
12377 +       unsigned long log_start;        /* next char to be read by syslog() */
12378 +       unsigned long con_start;        /* next char to be sent to consoles */
12379 +       unsigned long log_end;  /* most-recently-written-char + 1 */
12380 +       unsigned long logged_chars;     /* #chars since last read+clear operation */
12381 +
12382 +       char log_buf[1024];
12383 +};
12384 +
12385 +
12386 +/* context sub struct */
12387 +
12388 +struct _vx_cvirt {
12389 +       atomic_t nr_threads;            /* number of current threads */
12390 +       atomic_t nr_running;            /* number of running threads */
12391 +       atomic_t nr_uninterruptible;    /* number of uninterruptible threads */
12392 +
12393 +       atomic_t nr_onhold;             /* processes on hold */
12394 +       uint32_t onhold_last;           /* jiffies when put on hold */
12395 +
12396 +       struct timeval bias_tv;         /* time offset to the host */
12397 +       struct timespec bias_idle;
12398 +       struct timespec bias_uptime;    /* context creation point */
12399 +       uint64_t bias_clock;            /* offset in clock_t */
12400 +
12401 +       spinlock_t load_lock;           /* lock for the load averages */
12402 +       atomic_t load_updates;          /* nr of load updates done so far */
12403 +       uint32_t load_last;             /* last time load was calculated */
12404 +       uint32_t load[3];               /* load averages 1,5,15 */
12405 +
12406 +       atomic_t total_forks;           /* number of forks so far */
12407 +
12408 +       struct _vx_syslog syslog;
12409 +};
12410 +
12411 +struct _vx_cvirt_pc {
12412 +       struct _vx_usage_stat cpustat;
12413 +};
12414 +
12415 +
12416 +#ifdef CONFIG_VSERVER_DEBUG
12417 +
12418 +static inline void __dump_vx_cvirt(struct _vx_cvirt *cvirt)
12419 +{
12420 +       printk("\t_vx_cvirt:\n");
12421 +       printk("\t threads: %4d, %4d, %4d, %4d\n",
12422 +               atomic_read(&cvirt->nr_threads),
12423 +               atomic_read(&cvirt->nr_running),
12424 +               atomic_read(&cvirt->nr_uninterruptible),
12425 +               atomic_read(&cvirt->nr_onhold));
12426 +       /* add rest here */
12427 +       printk("\t total_forks = %d\n", atomic_read(&cvirt->total_forks));
12428 +}
12429 +
12430 +#endif
12431 +
12432 +#endif /* _VX_CVIRT_DEF_H */
12433 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/cvirt.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/cvirt.h
12434 --- linux-2.6.25.11/include/linux/vserver/cvirt.h       1969-12-31 19:00:00.000000000 -0500
12435 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/cvirt.h 2008-04-19 15:14:52.000000000 -0400
12436 @@ -0,0 +1,20 @@
12437 +#ifndef _VX_CVIRT_H
12438 +#define _VX_CVIRT_H
12439 +
12440 +
12441 +#ifdef __KERNEL__
12442 +
12443 +struct timespec;
12444 +
12445 +void vx_vsi_uptime(struct timespec *, struct timespec *);
12446 +
12447 +
12448 +struct vx_info;
12449 +
12450 +void vx_update_load(struct vx_info *);
12451 +
12452 +
12453 +int vx_do_syslog(int, char __user *, int);
12454 +
12455 +#endif /* __KERNEL__ */
12456 +#endif /* _VX_CVIRT_H */
12457 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/debug_cmd.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/debug_cmd.h
12458 --- linux-2.6.25.11/include/linux/vserver/debug_cmd.h   1969-12-31 19:00:00.000000000 -0500
12459 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/debug_cmd.h     2008-04-19 15:14:52.000000000 -0400
12460 @@ -0,0 +1,58 @@
12461 +#ifndef _VX_DEBUG_CMD_H
12462 +#define _VX_DEBUG_CMD_H
12463 +
12464 +
12465 +/* debug commands */
12466 +
12467 +#define VCMD_dump_history      VC_CMD(DEBUG, 1, 0)
12468 +
12469 +#define VCMD_read_history      VC_CMD(DEBUG, 5, 0)
12470 +#define VCMD_read_monitor      VC_CMD(DEBUG, 6, 0)
12471 +
12472 +struct  vcmd_read_history_v0 {
12473 +       uint32_t index;
12474 +       uint32_t count;
12475 +       char __user *data;
12476 +};
12477 +
12478 +struct  vcmd_read_monitor_v0 {
12479 +       uint32_t index;
12480 +       uint32_t count;
12481 +       char __user *data;
12482 +};
12483 +
12484 +
12485 +#ifdef __KERNEL__
12486 +
12487 +#ifdef CONFIG_COMPAT
12488 +
12489 +#include <asm/compat.h>
12490 +
12491 +struct vcmd_read_history_v0_x32 {
12492 +       uint32_t index;
12493 +       uint32_t count;
12494 +       compat_uptr_t data_ptr;
12495 +};
12496 +
12497 +struct vcmd_read_monitor_v0_x32 {
12498 +       uint32_t index;
12499 +       uint32_t count;
12500 +       compat_uptr_t data_ptr;
12501 +};
12502 +
12503 +#endif  /* CONFIG_COMPAT */
12504 +
12505 +extern int vc_dump_history(uint32_t);
12506 +
12507 +extern int vc_read_history(uint32_t, void __user *);
12508 +extern int vc_read_monitor(uint32_t, void __user *);
12509 +
12510 +#ifdef CONFIG_COMPAT
12511 +
12512 +extern int vc_read_history_x32(uint32_t, void __user *);
12513 +extern int vc_read_monitor_x32(uint32_t, void __user *);
12514 +
12515 +#endif  /* CONFIG_COMPAT */
12516 +
12517 +#endif /* __KERNEL__ */
12518 +#endif /* _VX_DEBUG_CMD_H */
12519 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/debug.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/debug.h
12520 --- linux-2.6.25.11/include/linux/vserver/debug.h       1969-12-31 19:00:00.000000000 -0500
12521 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/debug.h 2008-04-21 13:55:14.000000000 -0400
12522 @@ -0,0 +1,130 @@
12523 +#ifndef _VX_DEBUG_H
12524 +#define _VX_DEBUG_H
12525 +
12526 +
12527 +#define VXD_CBIT(n, m) (vx_debug_ ## n & (1 << (m)))
12528 +#define VXD_CMIN(n, m) (vx_debug_ ## n > (m))
12529 +#define VXD_MASK(n, m) (vx_debug_ ## n & (m))
12530 +
12531 +#define VXD_DEV(d)     (d), (d)->bd_inode->i_ino,              \
12532 +                       imajor((d)->bd_inode), iminor((d)->bd_inode)
12533 +#define VXF_DEV                "%p[%lu,%d:%d]"
12534 +
12535 +
12536 +#define __FUNC__       __func__
12537 +
12538 +
12539 +#define vxd_path(p)                                            \
12540 +       ({ static char _buffer[PATH_MAX];                       \
12541 +          d_path(p, _buffer, sizeof(_buffer)); })
12542 +
12543 +#define vxd_cond_path(n)                                       \
12544 +       ((n) ? vxd_path(&(n)->path) : "<null>" )
12545 +
12546 +
12547 +#ifdef CONFIG_VSERVER_DEBUG
12548 +
12549 +extern unsigned int vx_debug_switch;
12550 +extern unsigned int vx_debug_xid;
12551 +extern unsigned int vx_debug_nid;
12552 +extern unsigned int vx_debug_tag;
12553 +extern unsigned int vx_debug_net;
12554 +extern unsigned int vx_debug_limit;
12555 +extern unsigned int vx_debug_cres;
12556 +extern unsigned int vx_debug_dlim;
12557 +extern unsigned int vx_debug_quota;
12558 +extern unsigned int vx_debug_cvirt;
12559 +extern unsigned int vx_debug_space;
12560 +extern unsigned int vx_debug_misc;
12561 +
12562 +
12563 +#define VX_LOGLEVEL    "vxD: "
12564 +#define VX_PROC_FMT    "%p: "
12565 +#define VX_PROCESS     current
12566 +
12567 +#define vxdprintk(c, f, x...)                                  \
12568 +       do {                                                    \
12569 +               if (c)                                          \
12570 +                       printk(VX_LOGLEVEL VX_PROC_FMT f "\n",  \
12571 +                               VX_PROCESS , ##x);              \
12572 +       } while (0)
12573 +
12574 +#define vxlprintk(c, f, x...)                                  \
12575 +       do {                                                    \
12576 +               if (c)                                          \
12577 +                       printk(VX_LOGLEVEL f " @%s:%d\n", x);   \
12578 +       } while (0)
12579 +
12580 +#define vxfprintk(c, f, x...)                                  \
12581 +       do {                                                    \
12582 +               if (c)                                          \
12583 +                       printk(VX_LOGLEVEL f " %s@%s:%d\n", x); \
12584 +       } while (0)
12585 +
12586 +
12587 +struct vx_info;
12588 +
12589 +void dump_vx_info(struct vx_info *, int);
12590 +void dump_vx_info_inactive(int);
12591 +
12592 +#else  /* CONFIG_VSERVER_DEBUG */
12593 +
12594 +#define vx_debug_switch 0
12595 +#define vx_debug_xid   0
12596 +#define vx_debug_nid   0
12597 +#define vx_debug_tag   0
12598 +#define vx_debug_net   0
12599 +#define vx_debug_limit 0
12600 +#define vx_debug_cres  0
12601 +#define vx_debug_dlim  0
12602 +#define vx_debug_cvirt 0
12603 +
12604 +#define vxdprintk(x...) do { } while (0)
12605 +#define vxlprintk(x...) do { } while (0)
12606 +#define vxfprintk(x...) do { } while (0)
12607 +
12608 +#endif /* CONFIG_VSERVER_DEBUG */
12609 +
12610 +
12611 +#ifdef CONFIG_VSERVER_WARN
12612 +
12613 +#define VX_WARNLEVEL   KERN_WARNING "vxW: "
12614 +#define VX_WARN_TASK   "[»%s«,%u:#%u|%u|%u] "
12615 +#define VX_WARN_XID    "[xid #%u] "
12616 +#define VX_WARN_NID    "[nid #%u] "
12617 +#define VX_WARN_TAG    "[tag #%u] "
12618 +
12619 +#define vxwprintk(c, f, x...)                                  \
12620 +       do {                                                    \
12621 +               if (c)                                          \
12622 +                       printk(VX_WARNLEVEL f "\n", ##x);       \
12623 +       } while (0)
12624 +
12625 +#else  /* CONFIG_VSERVER_WARN */
12626 +
12627 +#define vxwprintk(x...) do { } while (0)
12628 +
12629 +#endif /* CONFIG_VSERVER_WARN */
12630 +
12631 +#define vxwprintk_task(c, f, x...)                             \
12632 +       vxwprintk(c, VX_WARN_TASK f,                            \
12633 +               current->comm, current->pid,                    \
12634 +               current->xid, current->nid, current->tag, ##x)
12635 +#define vxwprintk_xid(c, f, x...)                              \
12636 +       vxwprintk(c, VX_WARN_XID f, current->xid, x)
12637 +#define vxwprintk_nid(c, f, x...)                              \
12638 +       vxwprintk(c, VX_WARN_NID f, current->nid, x)
12639 +#define vxwprintk_tag(c, f, x...)                              \
12640 +       vxwprintk(c, VX_WARN_TAG f, current->tag, x)
12641 +
12642 +#ifdef CONFIG_VSERVER_DEBUG
12643 +#define vxd_assert_lock(l)     assert_spin_locked(l)
12644 +#define vxd_assert(c, f, x...) vxlprintk(!(c), \
12645 +       "assertion [" f "] failed.", ##x, __FILE__, __LINE__)
12646 +#else
12647 +#define vxd_assert_lock(l)     do { } while (0)
12648 +#define vxd_assert(c, f, x...) do { } while (0)
12649 +#endif
12650 +
12651 +
12652 +#endif /* _VX_DEBUG_H */
12653 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/device_cmd.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/device_cmd.h
12654 --- linux-2.6.25.11/include/linux/vserver/device_cmd.h  1969-12-31 19:00:00.000000000 -0500
12655 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/device_cmd.h    2008-04-19 15:14:52.000000000 -0400
12656 @@ -0,0 +1,44 @@
12657 +#ifndef _VX_DEVICE_CMD_H
12658 +#define _VX_DEVICE_CMD_H
12659 +
12660 +
12661 +/*  device vserver commands */
12662 +
12663 +#define VCMD_set_mapping       VC_CMD(DEVICE, 1, 0)
12664 +#define VCMD_unset_mapping     VC_CMD(DEVICE, 2, 0)
12665 +
12666 +struct vcmd_set_mapping_v0 {
12667 +       const char __user *device;
12668 +       const char __user *target;
12669 +       uint32_t flags;
12670 +};
12671 +
12672 +
12673 +#ifdef __KERNEL__
12674 +
12675 +#ifdef CONFIG_COMPAT
12676 +
12677 +#include <asm/compat.h>
12678 +
12679 +struct vcmd_set_mapping_v0_x32 {
12680 +       compat_uptr_t device_ptr;
12681 +       compat_uptr_t target_ptr;
12682 +       uint32_t flags;
12683 +};
12684 +
12685 +#endif /* CONFIG_COMPAT */
12686 +
12687 +#include <linux/compiler.h>
12688 +
12689 +extern int vc_set_mapping(struct vx_info *, void __user *);
12690 +extern int vc_unset_mapping(struct vx_info *, void __user *);
12691 +
12692 +#ifdef CONFIG_COMPAT
12693 +
12694 +extern int vc_set_mapping_x32(struct vx_info *, void __user *);
12695 +extern int vc_unset_mapping_x32(struct vx_info *, void __user *);
12696 +
12697 +#endif /* CONFIG_COMPAT */
12698 +
12699 +#endif /* __KERNEL__ */
12700 +#endif /* _VX_DEVICE_CMD_H */
12701 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/device_def.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/device_def.h
12702 --- linux-2.6.25.11/include/linux/vserver/device_def.h  1969-12-31 19:00:00.000000000 -0500
12703 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/device_def.h    2008-04-19 15:14:52.000000000 -0400
12704 @@ -0,0 +1,17 @@
12705 +#ifndef _VX_DEVICE_DEF_H
12706 +#define _VX_DEVICE_DEF_H
12707 +
12708 +#include <linux/types.h>
12709 +
12710 +struct vx_dmap_target {
12711 +       dev_t target;
12712 +       uint32_t flags;
12713 +};
12714 +
12715 +struct _vx_device {
12716 +#ifdef CONFIG_VSERVER_DEVICE
12717 +       struct vx_dmap_target targets[2];
12718 +#endif
12719 +};
12720 +
12721 +#endif /* _VX_DEVICE_DEF_H */
12722 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/device.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/device.h
12723 --- linux-2.6.25.11/include/linux/vserver/device.h      1969-12-31 19:00:00.000000000 -0500
12724 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/device.h        2008-04-19 15:14:52.000000000 -0400
12725 @@ -0,0 +1,15 @@
12726 +#ifndef _VX_DEVICE_H
12727 +#define _VX_DEVICE_H
12728 +
12729 +
12730 +#define DATTR_CREATE   0x00000001
12731 +#define DATTR_OPEN     0x00000002
12732 +
12733 +#define DATTR_REMAP    0x00000010
12734 +
12735 +#define DATTR_MASK     0x00000013
12736 +
12737 +
12738 +#else  /* _VX_DEVICE_H */
12739 +#warning duplicate inclusion
12740 +#endif /* _VX_DEVICE_H */
12741 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/dlimit_cmd.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/dlimit_cmd.h
12742 --- linux-2.6.25.11/include/linux/vserver/dlimit_cmd.h  1969-12-31 19:00:00.000000000 -0500
12743 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/dlimit_cmd.h    2008-04-19 15:14:52.000000000 -0400
12744 @@ -0,0 +1,74 @@
12745 +#ifndef _VX_DLIMIT_CMD_H
12746 +#define _VX_DLIMIT_CMD_H
12747 +
12748 +
12749 +/*  dlimit vserver commands */
12750 +
12751 +#define VCMD_add_dlimit                VC_CMD(DLIMIT, 1, 0)
12752 +#define VCMD_rem_dlimit                VC_CMD(DLIMIT, 2, 0)
12753 +
12754 +#define VCMD_set_dlimit                VC_CMD(DLIMIT, 5, 0)
12755 +#define VCMD_get_dlimit                VC_CMD(DLIMIT, 6, 0)
12756 +
12757 +struct vcmd_ctx_dlimit_base_v0 {
12758 +       const char __user *name;
12759 +       uint32_t flags;
12760 +};
12761 +
12762 +struct vcmd_ctx_dlimit_v0 {
12763 +       const char __user *name;
12764 +       uint32_t space_used;                    /* used space in kbytes */
12765 +       uint32_t space_total;                   /* maximum space in kbytes */
12766 +       uint32_t inodes_used;                   /* used inodes */
12767 +       uint32_t inodes_total;                  /* maximum inodes */
12768 +       uint32_t reserved;                      /* reserved for root in % */
12769 +       uint32_t flags;
12770 +};
12771 +
12772 +#define CDLIM_UNSET            ((uint32_t)0UL)
12773 +#define CDLIM_INFINITY         ((uint32_t)~0UL)
12774 +#define CDLIM_KEEP             ((uint32_t)~1UL)
12775 +
12776 +#ifdef __KERNEL__
12777 +
12778 +#ifdef CONFIG_COMPAT
12779 +
12780 +#include <asm/compat.h>
12781 +
12782 +struct vcmd_ctx_dlimit_base_v0_x32 {
12783 +       compat_uptr_t name_ptr;
12784 +       uint32_t flags;
12785 +};
12786 +
12787 +struct vcmd_ctx_dlimit_v0_x32 {
12788 +       compat_uptr_t name_ptr;
12789 +       uint32_t space_used;                    /* used space in kbytes */
12790 +       uint32_t space_total;                   /* maximum space in kbytes */
12791 +       uint32_t inodes_used;                   /* used inodes */
12792 +       uint32_t inodes_total;                  /* maximum inodes */
12793 +       uint32_t reserved;                      /* reserved for root in % */
12794 +       uint32_t flags;
12795 +};
12796 +
12797 +#endif /* CONFIG_COMPAT */
12798 +
12799 +#include <linux/compiler.h>
12800 +
12801 +extern int vc_add_dlimit(uint32_t, void __user *);
12802 +extern int vc_rem_dlimit(uint32_t, void __user *);
12803 +
12804 +extern int vc_set_dlimit(uint32_t, void __user *);
12805 +extern int vc_get_dlimit(uint32_t, void __user *);
12806 +
12807 +#ifdef CONFIG_COMPAT
12808 +
12809 +extern int vc_add_dlimit_x32(uint32_t, void __user *);
12810 +extern int vc_rem_dlimit_x32(uint32_t, void __user *);
12811 +
12812 +extern int vc_set_dlimit_x32(uint32_t, void __user *);
12813 +extern int vc_get_dlimit_x32(uint32_t, void __user *);
12814 +
12815 +#endif /* CONFIG_COMPAT */
12816 +
12817 +#endif /* __KERNEL__ */
12818 +#endif /* _VX_DLIMIT_CMD_H */
12819 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/dlimit.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/dlimit.h
12820 --- linux-2.6.25.11/include/linux/vserver/dlimit.h      1969-12-31 19:00:00.000000000 -0500
12821 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/dlimit.h        2008-04-19 15:14:52.000000000 -0400
12822 @@ -0,0 +1,54 @@
12823 +#ifndef _VX_DLIMIT_H
12824 +#define _VX_DLIMIT_H
12825 +
12826 +#include "switch.h"
12827 +
12828 +
12829 +#ifdef __KERNEL__
12830 +
12831 +/*      keep in sync with CDLIM_INFINITY       */
12832 +
12833 +#define DLIM_INFINITY          (~0ULL)
12834 +
12835 +#include <linux/spinlock.h>
12836 +#include <linux/rcupdate.h>
12837 +
12838 +struct super_block;
12839 +
12840 +struct dl_info {
12841 +       struct hlist_node dl_hlist;             /* linked list of contexts */
12842 +       struct rcu_head dl_rcu;                 /* the rcu head */
12843 +       tag_t dl_tag;                           /* context tag */
12844 +       atomic_t dl_usecnt;                     /* usage count */
12845 +       atomic_t dl_refcnt;                     /* reference count */
12846 +
12847 +       struct super_block *dl_sb;              /* associated superblock */
12848 +
12849 +       spinlock_t dl_lock;                     /* protect the values */
12850 +
12851 +       unsigned long long dl_space_used;       /* used space in bytes */
12852 +       unsigned long long dl_space_total;      /* maximum space in bytes */
12853 +       unsigned long dl_inodes_used;           /* used inodes */
12854 +       unsigned long dl_inodes_total;          /* maximum inodes */
12855 +
12856 +       unsigned int dl_nrlmult;                /* non root limit mult */
12857 +};
12858 +
12859 +struct rcu_head;
12860 +
12861 +extern void rcu_free_dl_info(struct rcu_head *);
12862 +extern void unhash_dl_info(struct dl_info *);
12863 +
12864 +extern struct dl_info *locate_dl_info(struct super_block *, tag_t);
12865 +
12866 +
12867 +struct kstatfs;
12868 +
12869 +extern void vx_vsi_statfs(struct super_block *, struct kstatfs *);
12870 +
12871 +typedef uint64_t dlsize_t;
12872 +
12873 +#endif /* __KERNEL__ */
12874 +#else  /* _VX_DLIMIT_H */
12875 +#warning duplicate inclusion
12876 +#endif /* _VX_DLIMIT_H */
12877 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/global.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/global.h
12878 --- linux-2.6.25.11/include/linux/vserver/global.h      1969-12-31 19:00:00.000000000 -0500
12879 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/global.h        2008-04-19 15:14:52.000000000 -0400
12880 @@ -0,0 +1,20 @@
12881 +#ifndef _VX_GLOBAL_H
12882 +#define _VX_GLOBAL_H
12883 +
12884 +
12885 +extern atomic_t vx_global_ctotal;
12886 +extern atomic_t vx_global_cactive;
12887 +
12888 +extern atomic_t nx_global_ctotal;
12889 +extern atomic_t nx_global_cactive;
12890 +
12891 +extern atomic_t vs_global_nsproxy;
12892 +extern atomic_t vs_global_fs;
12893 +extern atomic_t vs_global_mnt_ns;
12894 +extern atomic_t vs_global_uts_ns;
12895 +extern atomic_t vs_global_ipc_ns;
12896 +extern atomic_t vs_global_user_ns;
12897 +extern atomic_t vs_global_pid_ns;
12898 +
12899 +
12900 +#endif /* _VX_GLOBAL_H */
12901 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/history.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/history.h
12902 --- linux-2.6.25.11/include/linux/vserver/history.h     1969-12-31 19:00:00.000000000 -0500
12903 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/history.h       2008-04-19 15:14:52.000000000 -0400
12904 @@ -0,0 +1,197 @@
12905 +#ifndef _VX_HISTORY_H
12906 +#define _VX_HISTORY_H
12907 +
12908 +
12909 +enum {
12910 +       VXH_UNUSED = 0,
12911 +       VXH_THROW_OOPS = 1,
12912 +
12913 +       VXH_GET_VX_INFO,
12914 +       VXH_PUT_VX_INFO,
12915 +       VXH_INIT_VX_INFO,
12916 +       VXH_SET_VX_INFO,
12917 +       VXH_CLR_VX_INFO,
12918 +       VXH_CLAIM_VX_INFO,
12919 +       VXH_RELEASE_VX_INFO,
12920 +       VXH_ALLOC_VX_INFO,
12921 +       VXH_DEALLOC_VX_INFO,
12922 +       VXH_HASH_VX_INFO,
12923 +       VXH_UNHASH_VX_INFO,
12924 +       VXH_LOC_VX_INFO,
12925 +       VXH_LOOKUP_VX_INFO,
12926 +       VXH_CREATE_VX_INFO,
12927 +};
12928 +
12929 +struct _vxhe_vxi {
12930 +       struct vx_info *ptr;
12931 +       unsigned xid;
12932 +       unsigned usecnt;
12933 +       unsigned tasks;
12934 +};
12935 +
12936 +struct _vxhe_set_clr {
12937 +       void *data;
12938 +};
12939 +
12940 +struct _vxhe_loc_lookup {
12941 +       unsigned arg;
12942 +};
12943 +
12944 +struct _vx_hist_entry {
12945 +       void *loc;
12946 +       unsigned short seq;
12947 +       unsigned short type;
12948 +       struct _vxhe_vxi vxi;
12949 +       union {
12950 +               struct _vxhe_set_clr sc;
12951 +               struct _vxhe_loc_lookup ll;
12952 +       };
12953 +};
12954 +
12955 +#ifdef CONFIG_VSERVER_HISTORY
12956 +
12957 +extern unsigned volatile int vxh_active;
12958 +
12959 +struct _vx_hist_entry *vxh_advance(void *loc);
12960 +
12961 +
12962 +static inline
12963 +void   __vxh_copy_vxi(struct _vx_hist_entry *entry, struct vx_info *vxi)
12964 +{
12965 +       entry->vxi.ptr = vxi;
12966 +       if (vxi) {
12967 +               entry->vxi.usecnt = atomic_read(&vxi->vx_usecnt);
12968 +               entry->vxi.tasks = atomic_read(&vxi->vx_tasks);
12969 +               entry->vxi.xid = vxi->vx_id;
12970 +       }
12971 +}
12972 +
12973 +
12974 +#define        __HERE__ current_text_addr()
12975 +
12976 +#define __VXH_BODY(__type, __data, __here)     \
12977 +       struct _vx_hist_entry *entry;           \
12978 +                                               \
12979 +       preempt_disable();                      \
12980 +       entry = vxh_advance(__here);            \
12981 +       __data;                                 \
12982 +       entry->type = __type;                   \
12983 +       preempt_enable();
12984 +
12985 +
12986 +       /* pass vxi only */
12987 +
12988 +#define __VXH_SMPL                             \
12989 +       __vxh_copy_vxi(entry, vxi)
12990 +
12991 +static inline
12992 +void   __vxh_smpl(struct vx_info *vxi, int __type, void *__here)
12993 +{
12994 +       __VXH_BODY(__type, __VXH_SMPL, __here)
12995 +}
12996 +
12997 +       /* pass vxi and data (void *) */
12998 +
12999 +#define __VXH_DATA                             \
13000 +       __vxh_copy_vxi(entry, vxi);             \
13001 +       entry->sc.data = data
13002 +
13003 +static inline
13004 +void   __vxh_data(struct vx_info *vxi, void *data,
13005 +                       int __type, void *__here)
13006 +{
13007 +       __VXH_BODY(__type, __VXH_DATA, __here)
13008 +}
13009 +
13010 +       /* pass vxi and arg (long) */
13011 +
13012 +#define __VXH_LONG                             \
13013 +       __vxh_copy_vxi(entry, vxi);             \
13014 +       entry->ll.arg = arg
13015 +
13016 +static inline
13017 +void   __vxh_long(struct vx_info *vxi, long arg,
13018 +                       int __type, void *__here)
13019 +{
13020 +       __VXH_BODY(__type, __VXH_LONG, __here)
13021 +}
13022 +
13023 +
13024 +static inline
13025 +void   __vxh_throw_oops(void *__here)
13026 +{
13027 +       __VXH_BODY(VXH_THROW_OOPS, {}, __here);
13028 +       /* prevent further acquisition */
13029 +       vxh_active = 0;
13030 +}
13031 +
13032 +
13033 +#define vxh_throw_oops()       __vxh_throw_oops(__HERE__);
13034 +
13035 +#define __vxh_get_vx_info(v, h)        __vxh_smpl(v, VXH_GET_VX_INFO, h);
13036 +#define __vxh_put_vx_info(v, h)        __vxh_smpl(v, VXH_PUT_VX_INFO, h);
13037 +
13038 +#define __vxh_init_vx_info(v, d, h) \
13039 +       __vxh_data(v, d, VXH_INIT_VX_INFO, h);
13040 +#define __vxh_set_vx_info(v, d, h) \
13041 +       __vxh_data(v, d, VXH_SET_VX_INFO, h);
13042 +#define __vxh_clr_vx_info(v, d, h) \
13043 +       __vxh_data(v, d, VXH_CLR_VX_INFO, h);
13044 +
13045 +#define __vxh_claim_vx_info(v, d, h) \
13046 +       __vxh_data(v, d, VXH_CLAIM_VX_INFO, h);
13047 +#define __vxh_release_vx_info(v, d, h) \
13048 +       __vxh_data(v, d, VXH_RELEASE_VX_INFO, h);
13049 +
13050 +#define vxh_alloc_vx_info(v) \
13051 +       __vxh_smpl(v, VXH_ALLOC_VX_INFO, __HERE__);
13052 +#define vxh_dealloc_vx_info(v) \
13053 +       __vxh_smpl(v, VXH_DEALLOC_VX_INFO, __HERE__);
13054 +
13055 +#define vxh_hash_vx_info(v) \
13056 +       __vxh_smpl(v, VXH_HASH_VX_INFO, __HERE__);
13057 +#define vxh_unhash_vx_info(v) \
13058 +       __vxh_smpl(v, VXH_UNHASH_VX_INFO, __HERE__);
13059 +
13060 +#define vxh_loc_vx_info(v, l) \
13061 +       __vxh_long(v, l, VXH_LOC_VX_INFO, __HERE__);
13062 +#define vxh_lookup_vx_info(v, l) \
13063 +       __vxh_long(v, l, VXH_LOOKUP_VX_INFO, __HERE__);
13064 +#define vxh_create_vx_info(v, l) \
13065 +       __vxh_long(v, l, VXH_CREATE_VX_INFO, __HERE__);
13066 +
13067 +extern void vxh_dump_history(void);
13068 +
13069 +
13070 +#else  /* CONFIG_VSERVER_HISTORY */
13071 +
13072 +#define        __HERE__        0
13073 +
13074 +#define vxh_throw_oops()               do { } while (0)
13075 +
13076 +#define __vxh_get_vx_info(v, h)                do { } while (0)
13077 +#define __vxh_put_vx_info(v, h)                do { } while (0)
13078 +
13079 +#define __vxh_init_vx_info(v, d, h)    do { } while (0)
13080 +#define __vxh_set_vx_info(v, d, h)     do { } while (0)
13081 +#define __vxh_clr_vx_info(v, d, h)     do { } while (0)
13082 +
13083 +#define __vxh_claim_vx_info(v, d, h)   do { } while (0)
13084 +#define __vxh_release_vx_info(v, d, h) do { } while (0)
13085 +
13086 +#define vxh_alloc_vx_info(v)           do { } while (0)
13087 +#define vxh_dealloc_vx_info(v)         do { } while (0)
13088 +
13089 +#define vxh_hash_vx_info(v)            do { } while (0)
13090 +#define vxh_unhash_vx_info(v)          do { } while (0)
13091 +
13092 +#define vxh_loc_vx_info(v, l)          do { } while (0)
13093 +#define vxh_lookup_vx_info(v, l)       do { } while (0)
13094 +#define vxh_create_vx_info(v, l)       do { } while (0)
13095 +
13096 +#define vxh_dump_history()             do { } while (0)
13097 +
13098 +
13099 +#endif /* CONFIG_VSERVER_HISTORY */
13100 +
13101 +#endif /* _VX_HISTORY_H */
13102 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/inode_cmd.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/inode_cmd.h
13103 --- linux-2.6.25.11/include/linux/vserver/inode_cmd.h   1969-12-31 19:00:00.000000000 -0500
13104 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/inode_cmd.h     2008-04-19 15:14:52.000000000 -0400
13105 @@ -0,0 +1,59 @@
13106 +#ifndef _VX_INODE_CMD_H
13107 +#define _VX_INODE_CMD_H
13108 +
13109 +
13110 +/*  inode vserver commands */
13111 +
13112 +#define VCMD_get_iattr         VC_CMD(INODE, 1, 1)
13113 +#define VCMD_set_iattr         VC_CMD(INODE, 2, 1)
13114 +
13115 +#define VCMD_fget_iattr                VC_CMD(INODE, 3, 0)
13116 +#define VCMD_fset_iattr                VC_CMD(INODE, 4, 0)
13117 +
13118 +struct vcmd_ctx_iattr_v1 {
13119 +       const char __user *name;
13120 +       uint32_t tag;
13121 +       uint32_t flags;
13122 +       uint32_t mask;
13123 +};
13124 +
13125 +struct vcmd_ctx_fiattr_v0 {
13126 +       uint32_t tag;
13127 +       uint32_t flags;
13128 +       uint32_t mask;
13129 +};
13130 +
13131 +
13132 +#ifdef __KERNEL__
13133 +
13134 +
13135 +#ifdef CONFIG_COMPAT
13136 +
13137 +#include <asm/compat.h>
13138 +
13139 +struct vcmd_ctx_iattr_v1_x32 {
13140 +       compat_uptr_t name_ptr;
13141 +       uint32_t tag;
13142 +       uint32_t flags;
13143 +       uint32_t mask;
13144 +};
13145 +
13146 +#endif /* CONFIG_COMPAT */
13147 +
13148 +#include <linux/compiler.h>
13149 +
13150 +extern int vc_get_iattr(void __user *);
13151 +extern int vc_set_iattr(void __user *);
13152 +
13153 +extern int vc_fget_iattr(uint32_t, void __user *);
13154 +extern int vc_fset_iattr(uint32_t, void __user *);
13155 +
13156 +#ifdef CONFIG_COMPAT
13157 +
13158 +extern int vc_get_iattr_x32(void __user *);
13159 +extern int vc_set_iattr_x32(void __user *);
13160 +
13161 +#endif /* CONFIG_COMPAT */
13162 +
13163 +#endif /* __KERNEL__ */
13164 +#endif /* _VX_INODE_CMD_H */
13165 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/inode.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/inode.h
13166 --- linux-2.6.25.11/include/linux/vserver/inode.h       1969-12-31 19:00:00.000000000 -0500
13167 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/inode.h 2008-07-08 04:31:24.000000000 -0400
13168 @@ -0,0 +1,38 @@
13169 +#ifndef _VX_INODE_H
13170 +#define _VX_INODE_H
13171 +
13172 +
13173 +#define IATTR_TAG      0x01000000
13174 +
13175 +#define IATTR_ADMIN    0x00000001
13176 +#define IATTR_WATCH    0x00000002
13177 +#define IATTR_HIDE     0x00000004
13178 +#define IATTR_FLAGS    0x00000007
13179 +
13180 +#define IATTR_BARRIER  0x00010000
13181 +#define IATTR_IXUNLINK 0x00020000
13182 +#define IATTR_IMMUTABLE 0x00040000
13183 +
13184 +#ifdef __KERNEL__
13185 +
13186 +
13187 +#ifdef CONFIG_VSERVER_PROC_SECURE
13188 +#define IATTR_PROC_DEFAULT     ( IATTR_ADMIN | IATTR_HIDE )
13189 +#define IATTR_PROC_SYMLINK     ( IATTR_ADMIN )
13190 +#else
13191 +#define IATTR_PROC_DEFAULT     ( IATTR_ADMIN )
13192 +#define IATTR_PROC_SYMLINK     ( IATTR_ADMIN )
13193 +#endif
13194 +
13195 +#define vx_hide_check(c, m)    (((m) & IATTR_HIDE) ? vx_check(c, m) : 1)
13196 +
13197 +#endif /* __KERNEL__ */
13198 +
13199 +/* inode ioctls */
13200 +
13201 +#define FIOC_GETXFLG   _IOR('x', 5, long)
13202 +#define FIOC_SETXFLG   _IOW('x', 6, long)
13203 +
13204 +#else  /* _VX_INODE_H */
13205 +#warning duplicate inclusion
13206 +#endif /* _VX_INODE_H */
13207 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/Kbuild linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/Kbuild
13208 --- linux-2.6.25.11/include/linux/vserver/Kbuild        1969-12-31 19:00:00.000000000 -0500
13209 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/Kbuild  2008-04-19 15:14:52.000000000 -0400
13210 @@ -0,0 +1,8 @@
13211 +
13212 +unifdef-y += context_cmd.h network_cmd.h space_cmd.h \
13213 +       cacct_cmd.h cvirt_cmd.h limit_cmd.h dlimit_cmd.h \
13214 +       inode_cmd.h tag_cmd.h sched_cmd.h signal_cmd.h \
13215 +       debug_cmd.h device_cmd.h
13216 +
13217 +unifdef-y += switch.h network.h monitor.h inode.h device.h
13218 +
13219 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/limit_cmd.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/limit_cmd.h
13220 --- linux-2.6.25.11/include/linux/vserver/limit_cmd.h   1969-12-31 19:00:00.000000000 -0500
13221 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/limit_cmd.h     2008-04-19 15:14:52.000000000 -0400
13222 @@ -0,0 +1,69 @@
13223 +#ifndef _VX_LIMIT_CMD_H
13224 +#define _VX_LIMIT_CMD_H
13225 +
13226 +
13227 +/*  rlimit vserver commands */
13228 +
13229 +#define VCMD_get_rlimit                VC_CMD(RLIMIT, 1, 0)
13230 +#define VCMD_set_rlimit                VC_CMD(RLIMIT, 2, 0)
13231 +#define VCMD_get_rlimit_mask   VC_CMD(RLIMIT, 3, 0)
13232 +#define VCMD_reset_minmax      VC_CMD(RLIMIT, 9, 0)
13233 +
13234 +struct vcmd_ctx_rlimit_v0 {
13235 +       uint32_t id;
13236 +       uint64_t minimum;
13237 +       uint64_t softlimit;
13238 +       uint64_t maximum;
13239 +};
13240 +
13241 +struct vcmd_ctx_rlimit_mask_v0 {
13242 +       uint32_t minimum;
13243 +       uint32_t softlimit;
13244 +       uint32_t maximum;
13245 +};
13246 +
13247 +#define VCMD_rlimit_stat       VC_CMD(VSTAT, 1, 0)
13248 +
13249 +struct vcmd_rlimit_stat_v0 {
13250 +       uint32_t id;
13251 +       uint32_t hits;
13252 +       uint64_t value;
13253 +       uint64_t minimum;
13254 +       uint64_t maximum;
13255 +};
13256 +
13257 +#define CRLIM_UNSET            (0ULL)
13258 +#define CRLIM_INFINITY         (~0ULL)
13259 +#define CRLIM_KEEP             (~1ULL)
13260 +
13261 +#ifdef __KERNEL__
13262 +
13263 +#ifdef CONFIG_IA32_EMULATION
13264 +
13265 +struct vcmd_ctx_rlimit_v0_x32 {
13266 +       uint32_t id;
13267 +       uint64_t minimum;
13268 +       uint64_t softlimit;
13269 +       uint64_t maximum;
13270 +} __attribute__ ((aligned (4)));
13271 +
13272 +#endif /* CONFIG_IA32_EMULATION */
13273 +
13274 +#include <linux/compiler.h>
13275 +
13276 +extern int vc_get_rlimit_mask(uint32_t, void __user *);
13277 +extern int vc_get_rlimit(struct vx_info *, void __user *);
13278 +extern int vc_set_rlimit(struct vx_info *, void __user *);
13279 +extern int vc_reset_minmax(struct vx_info *, void __user *);
13280 +
13281 +extern int vc_rlimit_stat(struct vx_info *, void __user *);
13282 +
13283 +#ifdef CONFIG_IA32_EMULATION
13284 +
13285 +extern int vc_get_rlimit_x32(struct vx_info *, void __user *);
13286 +extern int vc_set_rlimit_x32(struct vx_info *, void __user *);
13287 +
13288 +#endif /* CONFIG_IA32_EMULATION */
13289 +
13290 +#endif /* __KERNEL__ */
13291 +#endif /* _VX_LIMIT_CMD_H */
13292 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/limit_def.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/limit_def.h
13293 --- linux-2.6.25.11/include/linux/vserver/limit_def.h   1969-12-31 19:00:00.000000000 -0500
13294 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/limit_def.h     2008-04-19 15:14:52.000000000 -0400
13295 @@ -0,0 +1,47 @@
13296 +#ifndef _VX_LIMIT_DEF_H
13297 +#define _VX_LIMIT_DEF_H
13298 +
13299 +#include <asm/atomic.h>
13300 +#include <asm/resource.h>
13301 +
13302 +#include "limit.h"
13303 +
13304 +
13305 +struct _vx_res_limit {
13306 +       rlim_t soft;            /* Context soft limit */
13307 +       rlim_t hard;            /* Context hard limit */
13308 +
13309 +       rlim_atomic_t rcur;     /* Current value */
13310 +       rlim_t rmin;            /* Context minimum */
13311 +       rlim_t rmax;            /* Context maximum */
13312 +
13313 +       atomic_t lhit;          /* Limit hits */
13314 +};
13315 +
13316 +/* context sub struct */
13317 +
13318 +struct _vx_limit {
13319 +       struct _vx_res_limit res[NUM_LIMITS];
13320 +};
13321 +
13322 +#ifdef CONFIG_VSERVER_DEBUG
13323 +
13324 +static inline void __dump_vx_limit(struct _vx_limit *limit)
13325 +{
13326 +       int i;
13327 +
13328 +       printk("\t_vx_limit:");
13329 +       for (i = 0; i < NUM_LIMITS; i++) {
13330 +               printk("\t [%2d] = %8lu %8lu/%8lu, %8ld/%8ld, %8d\n",
13331 +                       i, (unsigned long)__rlim_get(limit, i),
13332 +                       (unsigned long)__rlim_rmin(limit, i),
13333 +                       (unsigned long)__rlim_rmax(limit, i),
13334 +                       (long)__rlim_soft(limit, i),
13335 +                       (long)__rlim_hard(limit, i),
13336 +                       atomic_read(&__rlim_lhit(limit, i)));
13337 +       }
13338 +}
13339 +
13340 +#endif
13341 +
13342 +#endif /* _VX_LIMIT_DEF_H */
13343 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/limit.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/limit.h
13344 --- linux-2.6.25.11/include/linux/vserver/limit.h       1969-12-31 19:00:00.000000000 -0500
13345 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/limit.h 2008-04-19 15:14:52.000000000 -0400
13346 @@ -0,0 +1,70 @@
13347 +#ifndef _VX_LIMIT_H
13348 +#define _VX_LIMIT_H
13349 +
13350 +#define VLIMIT_NSOCK   16
13351 +#define VLIMIT_OPENFD  17
13352 +#define VLIMIT_ANON    18
13353 +#define VLIMIT_SHMEM   19
13354 +#define VLIMIT_SEMARY  20
13355 +#define VLIMIT_NSEMS   21
13356 +#define VLIMIT_DENTRY  22
13357 +#define VLIMIT_MAPPED  23
13358 +
13359 +
13360 +#ifdef __KERNEL__
13361 +
13362 +#define        VLIM_NOCHECK    ((1L << VLIMIT_DENTRY) | (1L << RLIMIT_RSS))
13363 +
13364 +/*     keep in sync with CRLIM_INFINITY */
13365 +
13366 +#define        VLIM_INFINITY   (~0ULL)
13367 +
13368 +#include <asm/atomic.h>
13369 +#include <asm/resource.h>
13370 +
13371 +#ifndef RLIM_INFINITY
13372 +#warning RLIM_INFINITY is undefined
13373 +#endif
13374 +
13375 +#define __rlim_val(l, r, v)    ((l)->res[r].v)
13376 +
13377 +#define __rlim_soft(l, r)      __rlim_val(l, r, soft)
13378 +#define __rlim_hard(l, r)      __rlim_val(l, r, hard)
13379 +
13380 +#define __rlim_rcur(l, r)      __rlim_val(l, r, rcur)
13381 +#define __rlim_rmin(l, r)      __rlim_val(l, r, rmin)
13382 +#define __rlim_rmax(l, r)      __rlim_val(l, r, rmax)
13383 +
13384 +#define __rlim_lhit(l, r)      __rlim_val(l, r, lhit)
13385 +#define __rlim_hit(l, r)       atomic_inc(&__rlim_lhit(l, r))
13386 +
13387 +typedef atomic_long_t rlim_atomic_t;
13388 +typedef unsigned long rlim_t;
13389 +
13390 +#define __rlim_get(l, r)       atomic_long_read(&__rlim_rcur(l, r))
13391 +#define __rlim_set(l, r, v)    atomic_long_set(&__rlim_rcur(l, r), v)
13392 +#define __rlim_inc(l, r)       atomic_long_inc(&__rlim_rcur(l, r))
13393 +#define __rlim_dec(l, r)       atomic_long_dec(&__rlim_rcur(l, r))
13394 +#define __rlim_add(l, r, v)    atomic_long_add(v, &__rlim_rcur(l, r))
13395 +#define __rlim_sub(l, r, v)    atomic_long_sub(v, &__rlim_rcur(l, r))
13396 +
13397 +
13398 +#if    (RLIM_INFINITY == VLIM_INFINITY)
13399 +#define        VX_VLIM(r) ((long long)(long)(r))
13400 +#define        VX_RLIM(v) ((rlim_t)(v))
13401 +#else
13402 +#define        VX_VLIM(r) (((r) == RLIM_INFINITY) \
13403 +               ? VLIM_INFINITY : (long long)(r))
13404 +#define        VX_RLIM(v) (((v) == VLIM_INFINITY) \
13405 +               ? RLIM_INFINITY : (rlim_t)(v))
13406 +#endif
13407 +
13408 +struct sysinfo;
13409 +
13410 +void vx_vsi_meminfo(struct sysinfo *);
13411 +void vx_vsi_swapinfo(struct sysinfo *);
13412 +
13413 +#define NUM_LIMITS     24
13414 +
13415 +#endif /* __KERNEL__ */
13416 +#endif /* _VX_LIMIT_H */
13417 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/limit_int.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/limit_int.h
13418 --- linux-2.6.25.11/include/linux/vserver/limit_int.h   1969-12-31 19:00:00.000000000 -0500
13419 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/limit_int.h     2008-04-19 15:14:52.000000000 -0400
13420 @@ -0,0 +1,198 @@
13421 +#ifndef _VX_LIMIT_INT_H
13422 +#define _VX_LIMIT_INT_H
13423 +
13424 +#include "context.h"
13425 +
13426 +#ifdef __KERNEL__
13427 +
13428 +#define VXD_RCRES_COND(r)      VXD_CBIT(cres, r)
13429 +#define VXD_RLIMIT_COND(r)     VXD_CBIT(limit, r)
13430 +
13431 +extern const char *vlimit_name[NUM_LIMITS];
13432 +
13433 +static inline void __vx_acc_cres(struct vx_info *vxi,
13434 +       int res, int dir, void *_data, char *_file, int _line)
13435 +{
13436 +       if (VXD_RCRES_COND(res))
13437 +               vxlprintk(1, "vx_acc_cres[%5d,%s,%2d]: %5ld%s (%p)",
13438 +                       (vxi ? vxi->vx_id : -1), vlimit_name[res], res,
13439 +                       (vxi ? (long)__rlim_get(&vxi->limit, res) : 0),
13440 +                       (dir > 0) ? "++" : "--", _data, _file, _line);
13441 +       if (!vxi)
13442 +               return;
13443 +
13444 +       if (dir > 0)
13445 +               __rlim_inc(&vxi->limit, res);
13446 +       else
13447 +               __rlim_dec(&vxi->limit, res);
13448 +}
13449 +
13450 +static inline void __vx_add_cres(struct vx_info *vxi,
13451 +       int res, int amount, void *_data, char *_file, int _line)
13452 +{
13453 +       if (VXD_RCRES_COND(res))
13454 +               vxlprintk(1, "vx_add_cres[%5d,%s,%2d]: %5ld += %5d (%p)",
13455 +                       (vxi ? vxi->vx_id : -1), vlimit_name[res], res,
13456 +                       (vxi ? (long)__rlim_get(&vxi->limit, res) : 0),
13457 +                       amount, _data, _file, _line);
13458 +       if (amount == 0)
13459 +               return;
13460 +       if (!vxi)
13461 +               return;
13462 +       __rlim_add(&vxi->limit, res, amount);
13463 +}
13464 +
13465 +static inline
13466 +int __vx_cres_adjust_max(struct _vx_limit *limit, int res, rlim_t value)
13467 +{
13468 +       int cond = (value > __rlim_rmax(limit, res));
13469 +
13470 +       if (cond)
13471 +               __rlim_rmax(limit, res) = value;
13472 +       return cond;
13473 +}
13474 +
13475 +static inline
13476 +int __vx_cres_adjust_min(struct _vx_limit *limit, int res, rlim_t value)
13477 +{
13478 +       int cond = (value < __rlim_rmin(limit, res));
13479 +
13480 +       if (cond)
13481 +               __rlim_rmin(limit, res) = value;
13482 +       return cond;
13483 +}
13484 +
13485 +static inline
13486 +void __vx_cres_fixup(struct _vx_limit *limit, int res, rlim_t value)
13487 +{
13488 +       if (!__vx_cres_adjust_max(limit, res, value))
13489 +               __vx_cres_adjust_min(limit, res, value);
13490 +}
13491 +
13492 +
13493 +/*     return values:
13494 +        +1 ... no limit hit
13495 +        -1 ... over soft limit
13496 +         0 ... over hard limit         */
13497 +
13498 +static inline int __vx_cres_avail(struct vx_info *vxi,
13499 +       int res, int num, char *_file, int _line)
13500 +{
13501 +       struct _vx_limit *limit;
13502 +       rlim_t value;
13503 +
13504 +       if (VXD_RLIMIT_COND(res))
13505 +               vxlprintk(1, "vx_cres_avail[%5d,%s,%2d]: %5ld/%5ld > %5ld + %5d",
13506 +                       (vxi ? vxi->vx_id : -1), vlimit_name[res], res,
13507 +                       (vxi ? (long)__rlim_soft(&vxi->limit, res) : -1),
13508 +                       (vxi ? (long)__rlim_hard(&vxi->limit, res) : -1),
13509 +                       (vxi ? (long)__rlim_get(&vxi->limit, res) : 0),
13510 +                       num, _file, _line);
13511 +       if (!vxi)
13512 +               return 1;
13513 +
13514 +       limit = &vxi->limit;
13515 +       value = __rlim_get(limit, res);
13516 +
13517 +       if (!__vx_cres_adjust_max(limit, res, value))
13518 +               __vx_cres_adjust_min(limit, res, value);
13519 +
13520 +       if (num == 0)
13521 +               return 1;
13522 +
13523 +       if (__rlim_soft(limit, res) == RLIM_INFINITY)
13524 +               return -1;
13525 +       if (value + num <= __rlim_soft(limit, res))
13526 +               return -1;
13527 +
13528 +       if (__rlim_hard(limit, res) == RLIM_INFINITY)
13529 +               return 1;
13530 +       if (value + num <= __rlim_hard(limit, res))
13531 +               return 1;
13532 +
13533 +       __rlim_hit(limit, res);
13534 +       return 0;
13535 +}
13536 +
13537 +
13538 +static const int VLA_RSS[] = { RLIMIT_RSS, VLIMIT_ANON, VLIMIT_MAPPED, 0 };
13539 +
13540 +static inline
13541 +rlim_t __vx_cres_array_sum(struct _vx_limit *limit, const int *array)
13542 +{
13543 +       rlim_t value, sum = 0;
13544 +       int res;
13545 +
13546 +       while ((res = *array++)) {
13547 +               value = __rlim_get(limit, res);
13548 +               __vx_cres_fixup(limit, res, value);
13549 +               sum += value;
13550 +       }
13551 +       return sum;
13552 +}
13553 +
13554 +static inline
13555 +rlim_t __vx_cres_array_fixup(struct _vx_limit *limit, const int *array)
13556 +{
13557 +       rlim_t value = __vx_cres_array_sum(limit, array + 1);
13558 +       int res = *array;
13559 +
13560 +       if (value == __rlim_get(limit, res))
13561 +               return value;
13562 +
13563 +       __rlim_set(limit, res, value);
13564 +       /* now adjust min/max */
13565 +       if (!__vx_cres_adjust_max(limit, res, value))
13566 +               __vx_cres_adjust_min(limit, res, value);
13567 +
13568 +       return value;
13569 +}
13570 +
13571 +static inline int __vx_cres_array_avail(struct vx_info *vxi,
13572 +       const int *array, int num, char *_file, int _line)
13573 +{
13574 +       struct _vx_limit *limit;
13575 +       rlim_t value = 0;
13576 +       int res;
13577 +
13578 +       if (num == 0)
13579 +               return 1;
13580 +       if (!vxi)
13581 +               return 1;
13582 +
13583 +       limit = &vxi->limit;
13584 +       res = *array;
13585 +       value = __vx_cres_array_sum(limit, array + 1);
13586 +
13587 +       __rlim_set(limit, res, value);
13588 +       __vx_cres_fixup(limit, res, value);
13589 +
13590 +       return __vx_cres_avail(vxi, res, num, _file, _line);
13591 +}
13592 +
13593 +
13594 +static inline void vx_limit_fixup(struct _vx_limit *limit, int id)
13595 +{
13596 +       rlim_t value;
13597 +       int res;
13598 +
13599 +       /* complex resources first */
13600 +       if ((id < 0) || (id == RLIMIT_RSS))
13601 +               __vx_cres_array_fixup(limit, VLA_RSS);
13602 +
13603 +       for (res = 0; res < NUM_LIMITS; res++) {
13604 +               if ((id > 0) && (res != id))
13605 +                       continue;
13606 +
13607 +               value = __rlim_get(limit, res);
13608 +               __vx_cres_fixup(limit, res, value);
13609 +
13610 +               /* not supposed to happen, maybe warn? */
13611 +               if (__rlim_rmax(limit, res) > __rlim_hard(limit, res))
13612 +                       __rlim_rmax(limit, res) = __rlim_hard(limit, res);
13613 +       }
13614 +}
13615 +
13616 +
13617 +#endif /* __KERNEL__ */
13618 +#endif /* _VX_LIMIT_INT_H */
13619 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/monitor.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/monitor.h
13620 --- linux-2.6.25.11/include/linux/vserver/monitor.h     1969-12-31 19:00:00.000000000 -0500
13621 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/monitor.h       2008-04-19 15:14:52.000000000 -0400
13622 @@ -0,0 +1,96 @@
13623 +#ifndef _VX_MONITOR_H
13624 +#define _VX_MONITOR_H
13625 +
13626 +#include <linux/types.h>
13627 +
13628 +enum {
13629 +       VXM_UNUSED = 0,
13630 +
13631 +       VXM_SYNC = 0x10,
13632 +
13633 +       VXM_UPDATE = 0x20,
13634 +       VXM_UPDATE_1,
13635 +       VXM_UPDATE_2,
13636 +
13637 +       VXM_RQINFO_1 = 0x24,
13638 +       VXM_RQINFO_2,
13639 +
13640 +       VXM_ACTIVATE = 0x40,
13641 +       VXM_DEACTIVATE,
13642 +       VXM_IDLE,
13643 +
13644 +       VXM_HOLD = 0x44,
13645 +       VXM_UNHOLD,
13646 +
13647 +       VXM_MIGRATE = 0x48,
13648 +       VXM_RESCHED,
13649 +
13650 +       /* all other bits are flags */
13651 +       VXM_SCHED = 0x80,
13652 +};
13653 +
13654 +struct _vxm_update_1 {
13655 +       uint32_t tokens_max;
13656 +       uint32_t fill_rate;
13657 +       uint32_t interval;
13658 +};
13659 +
13660 +struct _vxm_update_2 {
13661 +       uint32_t tokens_min;
13662 +       uint32_t fill_rate;
13663 +       uint32_t interval;
13664 +};
13665 +
13666 +struct _vxm_rqinfo_1 {
13667 +       uint16_t running;
13668 +       uint16_t onhold;
13669 +       uint16_t iowait;
13670 +       uint16_t uintr;
13671 +       uint32_t idle_tokens;
13672 +};
13673 +
13674 +struct _vxm_rqinfo_2 {
13675 +       uint32_t norm_time;
13676 +       uint32_t idle_time;
13677 +       uint32_t idle_skip;
13678 +};
13679 +
13680 +struct _vxm_sched {
13681 +       uint32_t tokens;
13682 +       uint32_t norm_time;
13683 +       uint32_t idle_time;
13684 +};
13685 +
13686 +struct _vxm_task {
13687 +       uint16_t pid;
13688 +       uint16_t state;
13689 +};
13690 +
13691 +struct _vxm_event {
13692 +       uint32_t jif;
13693 +       union {
13694 +               uint32_t seq;
13695 +               uint32_t sec;
13696 +       };
13697 +       union {
13698 +               uint32_t tokens;
13699 +               uint32_t nsec;
13700 +               struct _vxm_task tsk;
13701 +       };
13702 +};
13703 +
13704 +struct _vx_mon_entry {
13705 +       uint16_t type;
13706 +       uint16_t xid;
13707 +       union {
13708 +               struct _vxm_event ev;
13709 +               struct _vxm_sched sd;
13710 +               struct _vxm_update_1 u1;
13711 +               struct _vxm_update_2 u2;
13712 +               struct _vxm_rqinfo_1 q1;
13713 +               struct _vxm_rqinfo_2 q2;
13714 +       };
13715 +};
13716 +
13717 +
13718 +#endif /* _VX_MONITOR_H */
13719 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/network_cmd.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/network_cmd.h
13720 --- linux-2.6.25.11/include/linux/vserver/network_cmd.h 1969-12-31 19:00:00.000000000 -0500
13721 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/network_cmd.h   2008-04-19 15:14:52.000000000 -0400
13722 @@ -0,0 +1,150 @@
13723 +#ifndef _VX_NETWORK_CMD_H
13724 +#define _VX_NETWORK_CMD_H
13725 +
13726 +
13727 +/* vinfo commands */
13728 +
13729 +#define VCMD_task_nid          VC_CMD(VINFO, 2, 0)
13730 +
13731 +#ifdef __KERNEL__
13732 +extern int vc_task_nid(uint32_t);
13733 +
13734 +#endif /* __KERNEL__ */
13735 +
13736 +#define VCMD_nx_info           VC_CMD(VINFO, 6, 0)
13737 +
13738 +struct vcmd_nx_info_v0 {
13739 +       uint32_t nid;
13740 +       /* more to come */
13741 +};
13742 +
13743 +#ifdef __KERNEL__
13744 +extern int vc_nx_info(struct nx_info *, void __user *);
13745 +
13746 +#endif /* __KERNEL__ */
13747 +
13748 +#include <linux/in.h>
13749 +#include <linux/in6.h>
13750 +
13751 +#define VCMD_net_create_v0     VC_CMD(VNET, 1, 0)
13752 +#define VCMD_net_create                VC_CMD(VNET, 1, 1)
13753 +
13754 +struct  vcmd_net_create {
13755 +       uint64_t flagword;
13756 +};
13757 +
13758 +#define VCMD_net_migrate       VC_CMD(NETMIG, 1, 0)
13759 +
13760 +#define VCMD_net_add           VC_CMD(NETALT, 1, 0)
13761 +#define VCMD_net_remove                VC_CMD(NETALT, 2, 0)
13762 +
13763 +struct vcmd_net_addr_v0 {
13764 +       uint16_t type;
13765 +       uint16_t count;
13766 +       struct in_addr ip[4];
13767 +       struct in_addr mask[4];
13768 +};
13769 +
13770 +#define VCMD_net_add_ipv4      VC_CMD(NETALT, 1, 1)
13771 +#define VCMD_net_remove_ipv4   VC_CMD(NETALT, 2, 1)
13772 +
13773 +struct vcmd_net_addr_ipv4_v1 {
13774 +       uint16_t type;
13775 +       uint16_t flags;
13776 +       struct in_addr ip;
13777 +       struct in_addr mask;
13778 +};
13779 +
13780 +#define VCMD_net_add_ipv6      VC_CMD(NETALT, 3, 1)
13781 +#define VCMD_net_remove_ipv6   VC_CMD(NETALT, 4, 1)
13782 +
13783 +struct vcmd_net_addr_ipv6_v1 {
13784 +       uint16_t type;
13785 +       uint16_t flags;
13786 +       uint32_t prefix;
13787 +       struct in6_addr ip;
13788 +       struct in6_addr mask;
13789 +};
13790 +
13791 +#define VCMD_add_match_ipv4    VC_CMD(NETALT, 5, 0)
13792 +#define VCMD_get_match_ipv4    VC_CMD(NETALT, 6, 0)
13793 +
13794 +struct vcmd_match_ipv4_v0 {
13795 +       uint16_t type;
13796 +       uint16_t flags;
13797 +       uint16_t parent;
13798 +       uint16_t prefix;
13799 +       struct in_addr ip;
13800 +       struct in_addr ip2;
13801 +       struct in_addr mask;
13802 +};
13803 +
13804 +#define VCMD_add_match_ipv6    VC_CMD(NETALT, 7, 0)
13805 +#define VCMD_get_match_ipv6    VC_CMD(NETALT, 8, 0)
13806 +
13807 +struct vcmd_match_ipv6_v0 {
13808 +       uint16_t type;
13809 +       uint16_t flags;
13810 +       uint16_t parent;
13811 +       uint16_t prefix;
13812 +       struct in6_addr ip;
13813 +       struct in6_addr ip2;
13814 +       struct in6_addr mask;
13815 +};
13816 +
13817 +
13818 +#ifdef __KERNEL__
13819 +extern int vc_net_create(uint32_t, void __user *);
13820 +extern int vc_net_migrate(struct nx_info *, void __user *);
13821 +
13822 +extern int vc_net_add(struct nx_info *, void __user *);
13823 +extern int vc_net_remove(struct nx_info *, void __user *);
13824 +
13825 +extern int vc_net_add_ipv4(struct nx_info *, void __user *);
13826 +extern int vc_net_remove_ipv4(struct nx_info *, void __user *);
13827 +
13828 +extern int vc_net_add_ipv6(struct nx_info *, void __user *);
13829 +extern int vc_net_remove_ipv6(struct nx_info *, void __user *);
13830 +
13831 +extern int vc_add_match_ipv4(struct nx_info *, void __user *);
13832 +extern int vc_get_match_ipv4(struct nx_info *, void __user *);
13833 +
13834 +extern int vc_add_match_ipv6(struct nx_info *, void __user *);
13835 +extern int vc_get_match_ipv6(struct nx_info *, void __user *);
13836 +
13837 +#endif /* __KERNEL__ */
13838 +
13839 +
13840 +/* flag commands */
13841 +
13842 +#define VCMD_get_nflags                VC_CMD(FLAGS, 5, 0)
13843 +#define VCMD_set_nflags                VC_CMD(FLAGS, 6, 0)
13844 +
13845 +struct vcmd_net_flags_v0 {
13846 +       uint64_t flagword;
13847 +       uint64_t mask;
13848 +};
13849 +
13850 +#ifdef __KERNEL__
13851 +extern int vc_get_nflags(struct nx_info *, void __user *);
13852 +extern int vc_set_nflags(struct nx_info *, void __user *);
13853 +
13854 +#endif /* __KERNEL__ */
13855 +
13856 +
13857 +/* network caps commands */
13858 +
13859 +#define VCMD_get_ncaps         VC_CMD(FLAGS, 7, 0)
13860 +#define VCMD_set_ncaps         VC_CMD(FLAGS, 8, 0)
13861 +
13862 +struct vcmd_net_caps_v0 {
13863 +       uint64_t ncaps;
13864 +       uint64_t cmask;
13865 +};
13866 +
13867 +#ifdef __KERNEL__
13868 +extern int vc_get_ncaps(struct nx_info *, void __user *);
13869 +extern int vc_set_ncaps(struct nx_info *, void __user *);
13870 +
13871 +#endif /* __KERNEL__ */
13872 +#endif /* _VX_CONTEXT_CMD_H */
13873 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/network.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/network.h
13874 --- linux-2.6.25.11/include/linux/vserver/network.h     1969-12-31 19:00:00.000000000 -0500
13875 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/network.h       2008-04-19 15:14:52.000000000 -0400
13876 @@ -0,0 +1,146 @@
13877 +#ifndef _VX_NETWORK_H
13878 +#define _VX_NETWORK_H
13879 +
13880 +#include <linux/types.h>
13881 +
13882 +
13883 +#define MAX_N_CONTEXT  65535   /* Arbitrary limit */
13884 +
13885 +
13886 +/* network flags */
13887 +
13888 +#define NXF_INFO_PRIVATE       0x00000008
13889 +
13890 +#define NXF_SINGLE_IP          0x00000100
13891 +#define NXF_LBACK_REMAP                0x00000200
13892 +#define NXF_LBACK_ALLOW                0x00000400
13893 +
13894 +#define NXF_HIDE_NETIF         0x02000000
13895 +#define NXF_HIDE_LBACK         0x04000000
13896 +
13897 +#define NXF_STATE_SETUP                (1ULL << 32)
13898 +#define NXF_STATE_ADMIN                (1ULL << 34)
13899 +
13900 +#define NXF_SC_HELPER          (1ULL << 36)
13901 +#define NXF_PERSISTENT         (1ULL << 38)
13902 +
13903 +#define NXF_ONE_TIME           (0x0005ULL << 32)
13904 +
13905 +
13906 +#define        NXF_INIT_SET            (__nxf_init_set())
13907 +
13908 +static inline uint64_t __nxf_init_set(void) {
13909 +       return    NXF_STATE_ADMIN
13910 +#ifdef CONFIG_VSERVER_AUTO_LBACK
13911 +               | NXF_LBACK_REMAP
13912 +               | NXF_HIDE_LBACK
13913 +#endif
13914 +#ifdef CONFIG_VSERVER_AUTO_SINGLE
13915 +               | NXF_SINGLE_IP
13916 +#endif
13917 +               | NXF_HIDE_NETIF;
13918 +}
13919 +
13920 +
13921 +/* network caps */
13922 +
13923 +#define NXC_TUN_CREATE         0x00000001
13924 +
13925 +#define NXC_RAW_ICMP           0x00000100
13926 +
13927 +
13928 +/* address types */
13929 +
13930 +#define NXA_TYPE_IPV4          0x0001
13931 +#define NXA_TYPE_IPV6          0x0002
13932 +
13933 +#define NXA_TYPE_NONE          0x0000
13934 +#define NXA_TYPE_ANY           0x00FF
13935 +
13936 +#define NXA_TYPE_ADDR          0x0010
13937 +#define NXA_TYPE_MASK          0x0020
13938 +#define NXA_TYPE_RANGE         0x0040
13939 +
13940 +#define NXA_MASK_ALL           (NXA_TYPE_ADDR | NXA_TYPE_MASK | NXA_TYPE_RANGE)
13941 +
13942 +#define NXA_MOD_BCAST          0x0100
13943 +#define NXA_MOD_LBACK          0x0200
13944 +
13945 +#define NXA_LOOPBACK           0x1000
13946 +
13947 +#define NXA_MASK_BIND          (NXA_MASK_ALL | NXA_MOD_BCAST | NXA_MOD_LBACK)
13948 +#define NXA_MASK_SHOW          (NXA_MASK_ALL | NXA_LOOPBACK)
13949 +
13950 +#ifdef __KERNEL__
13951 +
13952 +#include <linux/list.h>
13953 +#include <linux/spinlock.h>
13954 +#include <linux/rcupdate.h>
13955 +#include <linux/in.h>
13956 +#include <linux/in6.h>
13957 +#include <asm/atomic.h>
13958 +
13959 +struct nx_addr_v4 {
13960 +       struct nx_addr_v4 *next;
13961 +       struct in_addr ip[2];
13962 +       struct in_addr mask;
13963 +       uint16_t type;
13964 +       uint16_t flags;
13965 +};
13966 +
13967 +struct nx_addr_v6 {
13968 +       struct nx_addr_v6 *next;
13969 +       struct in6_addr ip;
13970 +       struct in6_addr mask;
13971 +       uint32_t prefix;
13972 +       uint16_t type;
13973 +       uint16_t flags;
13974 +};
13975 +
13976 +struct nx_info {
13977 +       struct hlist_node nx_hlist;     /* linked list of nxinfos */
13978 +       nid_t nx_id;                    /* vnet id */
13979 +       atomic_t nx_usecnt;             /* usage count */
13980 +       atomic_t nx_tasks;              /* tasks count */
13981 +       int nx_state;                   /* context state */
13982 +
13983 +       uint64_t nx_flags;              /* network flag word */
13984 +       uint64_t nx_ncaps;              /* network capabilities */
13985 +
13986 +       struct in_addr v4_lback;        /* Loopback address */
13987 +       struct in_addr v4_bcast;        /* Broadcast address */
13988 +       struct nx_addr_v4 v4;           /* First/Single ipv4 address */
13989 +#ifdef CONFIG_IPV6
13990 +       struct nx_addr_v6 v6;           /* First/Single ipv6 address */
13991 +#endif
13992 +       char nx_name[65];               /* network context name */
13993 +};
13994 +
13995 +
13996 +/* status flags */
13997 +
13998 +#define NXS_HASHED      0x0001
13999 +#define NXS_SHUTDOWN    0x0100
14000 +#define NXS_RELEASED    0x8000
14001 +
14002 +extern struct nx_info *lookup_nx_info(int);
14003 +
14004 +extern int get_nid_list(int, unsigned int *, int);
14005 +extern int nid_is_hashed(nid_t);
14006 +
14007 +extern int nx_migrate_task(struct task_struct *, struct nx_info *);
14008 +
14009 +extern long vs_net_change(struct nx_info *, unsigned int);
14010 +
14011 +struct sock;
14012 +
14013 +
14014 +#define NX_IPV4(n)     ((n)->v4.type != NXA_TYPE_NONE)
14015 +#ifdef  CONFIG_IPV6
14016 +#define NX_IPV6(n)     ((n)->v6.type != NXA_TYPE_NONE)
14017 +#else
14018 +#define NX_IPV6(n)     (0)
14019 +#endif
14020 +
14021 +#endif /* __KERNEL__ */
14022 +#endif /* _VX_NETWORK_H */
14023 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/percpu.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/percpu.h
14024 --- linux-2.6.25.11/include/linux/vserver/percpu.h      1969-12-31 19:00:00.000000000 -0500
14025 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/percpu.h        2008-04-19 15:14:52.000000000 -0400
14026 @@ -0,0 +1,14 @@
14027 +#ifndef _VX_PERCPU_H
14028 +#define _VX_PERCPU_H
14029 +
14030 +#include "cvirt_def.h"
14031 +#include "sched_def.h"
14032 +
14033 +struct _vx_percpu {
14034 +       struct _vx_cvirt_pc cvirt;
14035 +       struct _vx_sched_pc sched;
14036 +};
14037 +
14038 +#define        PERCPU_PERCTX   (sizeof(struct _vx_percpu))
14039 +
14040 +#endif /* _VX_PERCPU_H */
14041 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/pid.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/pid.h
14042 --- linux-2.6.25.11/include/linux/vserver/pid.h 1969-12-31 19:00:00.000000000 -0500
14043 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/pid.h   2008-04-19 15:14:52.000000000 -0400
14044 @@ -0,0 +1,51 @@
14045 +#ifndef _VSERVER_PID_H
14046 +#define _VSERVER_PID_H
14047 +
14048 +/* pid faking stuff */
14049 +
14050 +#define vx_info_map_pid(v, p) \
14051 +       __vx_info_map_pid((v), (p), __FUNC__, __FILE__, __LINE__)
14052 +#define vx_info_map_tgid(v,p)  vx_info_map_pid(v,p)
14053 +#define vx_map_pid(p) vx_info_map_pid(current->vx_info, p)
14054 +#define vx_map_tgid(p) vx_map_pid(p)
14055 +
14056 +static inline int __vx_info_map_pid(struct vx_info *vxi, int pid,
14057 +       const char *func, const char *file, int line)
14058 +{
14059 +       if (vx_info_flags(vxi, VXF_INFO_INIT, 0)) {
14060 +               vxfprintk(VXD_CBIT(cvirt, 2),
14061 +                       "vx_map_tgid: %p/%llx: %d -> %d",
14062 +                       vxi, (long long)vxi->vx_flags, pid,
14063 +                       (pid && pid == vxi->vx_initpid) ? 1 : pid,
14064 +                       func, file, line);
14065 +               if (pid == 0)
14066 +                       return 0;
14067 +               if (pid == vxi->vx_initpid)
14068 +                       return 1;
14069 +       }
14070 +       return pid;
14071 +}
14072 +
14073 +#define vx_info_rmap_pid(v, p) \
14074 +       __vx_info_rmap_pid((v), (p), __FUNC__, __FILE__, __LINE__)
14075 +#define vx_rmap_pid(p) vx_info_rmap_pid(current->vx_info, p)
14076 +#define vx_rmap_tgid(p) vx_rmap_pid(p)
14077 +
14078 +static inline int __vx_info_rmap_pid(struct vx_info *vxi, int pid,
14079 +       const char *func, const char *file, int line)
14080 +{
14081 +       if (vx_info_flags(vxi, VXF_INFO_INIT, 0)) {
14082 +               vxfprintk(VXD_CBIT(cvirt, 2),
14083 +                       "vx_rmap_tgid: %p/%llx: %d -> %d",
14084 +                       vxi, (long long)vxi->vx_flags, pid,
14085 +                       (pid == 1) ? vxi->vx_initpid : pid,
14086 +                       func, file, line);
14087 +               if ((pid == 1) && vxi->vx_initpid)
14088 +                       return vxi->vx_initpid;
14089 +               if (pid == vxi->vx_initpid)
14090 +                       return ~0U;
14091 +       }
14092 +       return pid;
14093 +}
14094 +
14095 +#endif
14096 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/sched_cmd.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/sched_cmd.h
14097 --- linux-2.6.25.11/include/linux/vserver/sched_cmd.h   1969-12-31 19:00:00.000000000 -0500
14098 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/sched_cmd.h     2008-04-19 15:14:52.000000000 -0400
14099 @@ -0,0 +1,108 @@
14100 +#ifndef _VX_SCHED_CMD_H
14101 +#define _VX_SCHED_CMD_H
14102 +
14103 +
14104 +/*  sched vserver commands */
14105 +
14106 +#define VCMD_set_sched_v2      VC_CMD(SCHED, 1, 2)
14107 +#define VCMD_set_sched_v3      VC_CMD(SCHED, 1, 3)
14108 +#define VCMD_set_sched_v4      VC_CMD(SCHED, 1, 4)
14109 +
14110 +struct vcmd_set_sched_v2 {
14111 +       int32_t fill_rate;
14112 +       int32_t interval;
14113 +       int32_t tokens;
14114 +       int32_t tokens_min;
14115 +       int32_t tokens_max;
14116 +       uint64_t cpu_mask;
14117 +};
14118 +
14119 +struct vcmd_set_sched_v3 {
14120 +       uint32_t set_mask;
14121 +       int32_t fill_rate;
14122 +       int32_t interval;
14123 +       int32_t tokens;
14124 +       int32_t tokens_min;
14125 +       int32_t tokens_max;
14126 +       int32_t priority_bias;
14127 +};
14128 +
14129 +struct vcmd_set_sched_v4 {
14130 +       uint32_t set_mask;
14131 +       int32_t fill_rate;
14132 +       int32_t interval;
14133 +       int32_t tokens;
14134 +       int32_t tokens_min;
14135 +       int32_t tokens_max;
14136 +       int32_t prio_bias;
14137 +       int32_t cpu_id;
14138 +       int32_t bucket_id;
14139 +};
14140 +
14141 +#define VCMD_set_sched         VC_CMD(SCHED, 1, 5)
14142 +#define VCMD_get_sched         VC_CMD(SCHED, 2, 5)
14143 +
14144 +struct vcmd_sched_v5 {
14145 +       uint32_t mask;
14146 +       int32_t cpu_id;
14147 +       int32_t bucket_id;
14148 +       int32_t fill_rate[2];
14149 +       int32_t interval[2];
14150 +       int32_t tokens;
14151 +       int32_t tokens_min;
14152 +       int32_t tokens_max;
14153 +       int32_t prio_bias;
14154 +};
14155 +
14156 +#define VXSM_FILL_RATE         0x0001
14157 +#define VXSM_INTERVAL          0x0002
14158 +#define VXSM_FILL_RATE2                0x0004
14159 +#define VXSM_INTERVAL2         0x0008
14160 +#define VXSM_TOKENS            0x0010
14161 +#define VXSM_TOKENS_MIN                0x0020
14162 +#define VXSM_TOKENS_MAX                0x0040
14163 +#define VXSM_PRIO_BIAS         0x0100
14164 +
14165 +#define VXSM_IDLE_TIME         0x0200
14166 +#define VXSM_FORCE             0x0400
14167 +
14168 +#define        VXSM_V3_MASK            0x0173
14169 +#define        VXSM_SET_MASK           0x01FF
14170 +
14171 +#define VXSM_CPU_ID            0x1000
14172 +#define VXSM_BUCKET_ID         0x2000
14173 +
14174 +#define VXSM_MSEC              0x4000
14175 +
14176 +#define SCHED_KEEP             (-2)    /* only for v2 */
14177 +
14178 +#ifdef __KERNEL__
14179 +
14180 +#include <linux/compiler.h>
14181 +
14182 +extern int vc_set_sched_v2(struct vx_info *, void __user *);
14183 +extern int vc_set_sched_v3(struct vx_info *, void __user *);
14184 +extern int vc_set_sched_v4(struct vx_info *, void __user *);
14185 +extern int vc_set_sched(struct vx_info *, void __user *);
14186 +extern int vc_get_sched(struct vx_info *, void __user *);
14187 +
14188 +#endif /* __KERNEL__ */
14189 +
14190 +#define VCMD_sched_info                VC_CMD(SCHED, 3, 0)
14191 +
14192 +struct vcmd_sched_info {
14193 +       int32_t cpu_id;
14194 +       int32_t bucket_id;
14195 +       uint64_t user_msec;
14196 +       uint64_t sys_msec;
14197 +       uint64_t hold_msec;
14198 +       uint32_t token_usec;
14199 +       int32_t vavavoom;
14200 +};
14201 +
14202 +#ifdef __KERNEL__
14203 +
14204 +extern int vc_sched_info(struct vx_info *, void __user *);
14205 +
14206 +#endif /* __KERNEL__ */
14207 +#endif /* _VX_SCHED_CMD_H */
14208 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/sched_def.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/sched_def.h
14209 --- linux-2.6.25.11/include/linux/vserver/sched_def.h   1969-12-31 19:00:00.000000000 -0500
14210 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/sched_def.h     2008-04-19 15:14:52.000000000 -0400
14211 @@ -0,0 +1,68 @@
14212 +#ifndef _VX_SCHED_DEF_H
14213 +#define _VX_SCHED_DEF_H
14214 +
14215 +#include <linux/spinlock.h>
14216 +#include <linux/jiffies.h>
14217 +#include <linux/cpumask.h>
14218 +#include <asm/atomic.h>
14219 +#include <asm/param.h>
14220 +
14221 +
14222 +/* context sub struct */
14223 +
14224 +struct _vx_sched {
14225 +       spinlock_t tokens_lock;         /* lock for token bucket */
14226 +
14227 +       int tokens;                     /* number of CPU tokens */
14228 +       int fill_rate[2];               /* Fill rate: add X tokens... */
14229 +       int interval[2];                /* Divisor:   per Y jiffies   */
14230 +       int tokens_min;                 /* Limit:     minimum for unhold */
14231 +       int tokens_max;                 /* Limit:     no more than N tokens */
14232 +
14233 +       int prio_bias;                  /* bias offset for priority */
14234 +
14235 +       unsigned update_mask;           /* which features should be updated */
14236 +       cpumask_t update;               /* CPUs which should update */
14237 +};
14238 +
14239 +struct _vx_sched_pc {
14240 +       int tokens;                     /* number of CPU tokens */
14241 +       int flags;                      /* bucket flags */
14242 +
14243 +       int fill_rate[2];               /* Fill rate: add X tokens... */
14244 +       int interval[2];                /* Divisor:   per Y jiffies   */
14245 +       int tokens_min;                 /* Limit:     minimum for unhold */
14246 +       int tokens_max;                 /* Limit:     no more than N tokens */
14247 +
14248 +       int prio_bias;                  /* bias offset for priority */
14249 +       int vavavoom;                   /* last calculated vavavoom */
14250 +
14251 +       unsigned long norm_time;        /* last time accounted */
14252 +       unsigned long idle_time;        /* non linear time for fair sched */
14253 +       unsigned long token_time;       /* token time for accounting */
14254 +       unsigned long onhold;           /* jiffies when put on hold */
14255 +
14256 +       uint64_t user_ticks;            /* token tick events */
14257 +       uint64_t sys_ticks;             /* token tick events */
14258 +       uint64_t hold_ticks;            /* token ticks paused */
14259 +};
14260 +
14261 +
14262 +#define VXSF_ONHOLD    0x0001
14263 +#define VXSF_IDLE_TIME 0x0100
14264 +
14265 +#ifdef CONFIG_VSERVER_DEBUG
14266 +
14267 +static inline void __dump_vx_sched(struct _vx_sched *sched)
14268 +{
14269 +       printk("\t_vx_sched:\n");
14270 +       printk("\t tokens: %4d/%4d, %4d/%4d, %4d, %4d\n",
14271 +               sched->fill_rate[0], sched->interval[0],
14272 +               sched->fill_rate[1], sched->interval[1],
14273 +               sched->tokens_min, sched->tokens_max);
14274 +       printk("\t priority = %4d\n", sched->prio_bias);
14275 +}
14276 +
14277 +#endif
14278 +
14279 +#endif /* _VX_SCHED_DEF_H */
14280 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/sched.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/sched.h
14281 --- linux-2.6.25.11/include/linux/vserver/sched.h       1969-12-31 19:00:00.000000000 -0500
14282 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/sched.h 2008-04-19 15:14:52.000000000 -0400
14283 @@ -0,0 +1,26 @@
14284 +#ifndef _VX_SCHED_H
14285 +#define _VX_SCHED_H
14286 +
14287 +
14288 +#ifdef __KERNEL__
14289 +
14290 +struct timespec;
14291 +
14292 +void vx_vsi_uptime(struct timespec *, struct timespec *);
14293 +
14294 +
14295 +struct vx_info;
14296 +
14297 +void vx_update_load(struct vx_info *);
14298 +
14299 +
14300 +int vx_tokens_recalc(struct _vx_sched_pc *,
14301 +       unsigned long *, unsigned long *, int [2]);
14302 +
14303 +void vx_update_sched_param(struct _vx_sched *sched,
14304 +       struct _vx_sched_pc *sched_pc);
14305 +
14306 +#endif /* __KERNEL__ */
14307 +#else  /* _VX_SCHED_H */
14308 +#warning duplicate inclusion
14309 +#endif /* _VX_SCHED_H */
14310 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/signal_cmd.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/signal_cmd.h
14311 --- linux-2.6.25.11/include/linux/vserver/signal_cmd.h  1969-12-31 19:00:00.000000000 -0500
14312 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/signal_cmd.h    2008-04-19 15:14:52.000000000 -0400
14313 @@ -0,0 +1,43 @@
14314 +#ifndef _VX_SIGNAL_CMD_H
14315 +#define _VX_SIGNAL_CMD_H
14316 +
14317 +
14318 +/*  signalling vserver commands */
14319 +
14320 +#define VCMD_ctx_kill          VC_CMD(PROCTRL, 1, 0)
14321 +#define VCMD_wait_exit         VC_CMD(EVENT, 99, 0)
14322 +
14323 +struct vcmd_ctx_kill_v0 {
14324 +       int32_t pid;
14325 +       int32_t sig;
14326 +};
14327 +
14328 +struct vcmd_wait_exit_v0 {
14329 +       int32_t reboot_cmd;
14330 +       int32_t exit_code;
14331 +};
14332 +
14333 +#ifdef __KERNEL__
14334 +
14335 +extern int vc_ctx_kill(struct vx_info *, void __user *);
14336 +extern int vc_wait_exit(struct vx_info *, void __user *);
14337 +
14338 +#endif /* __KERNEL__ */
14339 +
14340 +/*  process alteration commands */
14341 +
14342 +#define VCMD_get_pflags                VC_CMD(PROCALT, 5, 0)
14343 +#define VCMD_set_pflags                VC_CMD(PROCALT, 6, 0)
14344 +
14345 +struct vcmd_pflags_v0 {
14346 +       uint32_t flagword;
14347 +       uint32_t mask;
14348 +};
14349 +
14350 +#ifdef __KERNEL__
14351 +
14352 +extern int vc_get_pflags(uint32_t pid, void __user *);
14353 +extern int vc_set_pflags(uint32_t pid, void __user *);
14354 +
14355 +#endif /* __KERNEL__ */
14356 +#endif /* _VX_SIGNAL_CMD_H */
14357 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/signal.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/signal.h
14358 --- linux-2.6.25.11/include/linux/vserver/signal.h      1969-12-31 19:00:00.000000000 -0500
14359 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/signal.h        2008-04-19 15:14:52.000000000 -0400
14360 @@ -0,0 +1,14 @@
14361 +#ifndef _VX_SIGNAL_H
14362 +#define _VX_SIGNAL_H
14363 +
14364 +
14365 +#ifdef __KERNEL__
14366 +
14367 +struct vx_info;
14368 +
14369 +int vx_info_kill(struct vx_info *, int, int);
14370 +
14371 +#endif /* __KERNEL__ */
14372 +#else  /* _VX_SIGNAL_H */
14373 +#warning duplicate inclusion
14374 +#endif /* _VX_SIGNAL_H */
14375 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/space_cmd.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/space_cmd.h
14376 --- linux-2.6.25.11/include/linux/vserver/space_cmd.h   1969-12-31 19:00:00.000000000 -0500
14377 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/space_cmd.h     2008-07-16 00:21:29.000000000 -0400
14378 @@ -0,0 +1,29 @@
14379 +#ifndef _VX_SPACE_CMD_H
14380 +#define _VX_SPACE_CMD_H
14381 +
14382 +
14383 +#define VCMD_enter_space_v0    VC_CMD(PROCALT, 1, 0)
14384 +#define VCMD_enter_space       VC_CMD(PROCALT, 1, 1)
14385 +
14386 +#define VCMD_set_space_v0      VC_CMD(PROCALT, 3, 0)
14387 +#define VCMD_set_space         VC_CMD(PROCALT, 3, 1)
14388 +
14389 +#define VCMD_get_space_mask_v0 VC_CMD(PROCALT, 4, 0)
14390 +
14391 +#define VCMD_get_space_mask    VC_CMD(VSPACE, 0, 1)
14392 +#define VCMD_get_space_default VC_CMD(VSPACE, 1, 0)
14393 +
14394 +
14395 +struct vcmd_space_mask {
14396 +       uint64_t mask;
14397 +};
14398 +
14399 +
14400 +#ifdef __KERNEL__
14401 +
14402 +extern int vc_enter_space(struct vx_info *, void __user *);
14403 +extern int vc_set_space(struct vx_info *, void __user *);
14404 +extern int vc_get_space_mask(void __user *, int);
14405 +
14406 +#endif /* __KERNEL__ */
14407 +#endif /* _VX_SPACE_CMD_H */
14408 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/space.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/space.h
14409 --- linux-2.6.25.11/include/linux/vserver/space.h       1969-12-31 19:00:00.000000000 -0500
14410 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/space.h 2008-04-19 15:14:52.000000000 -0400
14411 @@ -0,0 +1,13 @@
14412 +#ifndef _VX_SPACE_H
14413 +#define _VX_SPACE_H
14414 +
14415 +
14416 +#include <linux/types.h>
14417 +
14418 +struct vx_info;
14419 +
14420 +int vx_set_space(struct vx_info *vxi, unsigned long mask);
14421 +
14422 +#else  /* _VX_SPACE_H */
14423 +#warning duplicate inclusion
14424 +#endif /* _VX_SPACE_H */
14425 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/switch.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/switch.h
14426 --- linux-2.6.25.11/include/linux/vserver/switch.h      1969-12-31 19:00:00.000000000 -0500
14427 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/switch.h        2008-07-15 22:47:32.000000000 -0400
14428 @@ -0,0 +1,100 @@
14429 +#ifndef _VX_SWITCH_H
14430 +#define _VX_SWITCH_H
14431 +
14432 +#include <linux/types.h>
14433 +
14434 +
14435 +#define VC_CATEGORY(c)         (((c) >> 24) & 0x3F)
14436 +#define VC_COMMAND(c)          (((c) >> 16) & 0xFF)
14437 +#define VC_VERSION(c)          ((c) & 0xFFF)
14438 +
14439 +#define VC_CMD(c, i, v)                ((((VC_CAT_ ## c) & 0x3F) << 24) \
14440 +                               | (((i) & 0xFF) << 16) | ((v) & 0xFFF))
14441 +
14442 +/*
14443 +
14444 +  Syscall Matrix V2.8
14445 +
14446 +        |VERSION|CREATE |MODIFY |MIGRATE|CONTROL|EXPERIM| |SPECIAL|SPECIAL|
14447 +        |STATS  |DESTROY|ALTER  |CHANGE |LIMIT  |TEST   | |       |       |
14448 +        |INFO   |SETUP  |       |MOVE   |       |       | |       |       |
14449 +  -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
14450 +  SYSTEM |VERSION|VSETUP |VHOST  |       |       |       | |DEVICE |       |
14451 +  HOST   |     00|     01|     02|     03|     04|     05| |     06|     07|
14452 +  -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
14453 +  CPU    |       |VPROC  |PROCALT|PROCMIG|PROCTRL|       | |SCHED. |       |
14454 +  PROCESS|     08|     09|     10|     11|     12|     13| |     14|     15|
14455 +  -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
14456 +  MEMORY |       |       |       |       |MEMCTRL|       | |SWAP   |       |
14457 +        |     16|     17|     18|     19|     20|     21| |     22|     23|
14458 +  -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
14459 +  NETWORK|       |VNET   |NETALT |NETMIG |NETCTL |       | |SERIAL |       |
14460 +        |     24|     25|     26|     27|     28|     29| |     30|     31|
14461 +  -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
14462 +  DISK   |       |       |       |TAGMIG |DLIMIT |       | |INODE  |       |
14463 +  VFS    |     32|     33|     34|     35|     36|     37| |     38|     39|
14464 +  -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
14465 +  OTHER  |VSTAT  |       |       |       |       |       | |VINFO  |       |
14466 +        |     40|     41|     42|     43|     44|     45| |     46|     47|
14467 +  =======+=======+=======+=======+=======+=======+=======+ +=======+=======+
14468 +  SPECIAL|EVENT  |       |       |       |FLAGS  |       | |VSPACE |       |
14469 +        |     48|     49|     50|     51|     52|     53| |     54|     55|
14470 +  -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
14471 +  SPECIAL|DEBUG  |       |       |       |RLIMIT |SYSCALL| |       |COMPAT |
14472 +        |     56|     57|     58|     59|     60|TEST 61| |     62|     63|
14473 +  -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
14474 +
14475 +*/
14476 +
14477 +#define VC_CAT_VERSION         0
14478 +
14479 +#define VC_CAT_VSETUP          1
14480 +#define VC_CAT_VHOST           2
14481 +
14482 +#define VC_CAT_DEVICE          6
14483 +
14484 +#define VC_CAT_VPROC           9
14485 +#define VC_CAT_PROCALT         10
14486 +#define VC_CAT_PROCMIG         11
14487 +#define VC_CAT_PROCTRL         12
14488 +
14489 +#define VC_CAT_SCHED           14
14490 +#define VC_CAT_MEMCTRL         20
14491 +
14492 +#define VC_CAT_VNET            25
14493 +#define VC_CAT_NETALT          26
14494 +#define VC_CAT_NETMIG          27
14495 +#define VC_CAT_NETCTRL         28
14496 +
14497 +#define VC_CAT_TAGMIG          35
14498 +#define VC_CAT_DLIMIT          36
14499 +#define VC_CAT_INODE           38
14500 +
14501 +#define VC_CAT_VSTAT           40
14502 +#define VC_CAT_VINFO           46
14503 +#define VC_CAT_EVENT           48
14504 +
14505 +#define VC_CAT_FLAGS           52
14506 +#define VC_CAT_VSPACE          54
14507 +#define VC_CAT_DEBUG           56
14508 +#define VC_CAT_RLIMIT          60
14509 +
14510 +#define VC_CAT_SYSTEST         61
14511 +#define VC_CAT_COMPAT          63
14512 +
14513 +/*  query version */
14514 +
14515 +#define VCMD_get_version       VC_CMD(VERSION, 0, 0)
14516 +#define VCMD_get_vci           VC_CMD(VERSION, 1, 0)
14517 +
14518 +
14519 +#ifdef __KERNEL__
14520 +
14521 +#include <linux/errno.h>
14522 +
14523 +
14524 +#else  /* __KERNEL__ */
14525 +#define __user
14526 +#endif /* __KERNEL__ */
14527 +
14528 +#endif /* _VX_SWITCH_H */
14529 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/tag_cmd.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/tag_cmd.h
14530 --- linux-2.6.25.11/include/linux/vserver/tag_cmd.h     1969-12-31 19:00:00.000000000 -0500
14531 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/tag_cmd.h       2008-04-19 15:14:52.000000000 -0400
14532 @@ -0,0 +1,22 @@
14533 +#ifndef _VX_TAG_CMD_H
14534 +#define _VX_TAG_CMD_H
14535 +
14536 +
14537 +/* vinfo commands */
14538 +
14539 +#define VCMD_task_tag          VC_CMD(VINFO, 3, 0)
14540 +
14541 +#ifdef __KERNEL__
14542 +extern int vc_task_tag(uint32_t);
14543 +
14544 +#endif /* __KERNEL__ */
14545 +
14546 +/* context commands */
14547 +
14548 +#define VCMD_tag_migrate       VC_CMD(TAGMIG, 1, 0)
14549 +
14550 +#ifdef __KERNEL__
14551 +extern int vc_tag_migrate(uint32_t);
14552 +
14553 +#endif /* __KERNEL__ */
14554 +#endif /* _VX_TAG_CMD_H */
14555 diff -NurpP --minimal linux-2.6.25.11/include/linux/vserver/tag.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/tag.h
14556 --- linux-2.6.25.11/include/linux/vserver/tag.h 1969-12-31 19:00:00.000000000 -0500
14557 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vserver/tag.h   2008-04-21 13:53:47.000000000 -0400
14558 @@ -0,0 +1,143 @@
14559 +#ifndef _DX_TAG_H
14560 +#define _DX_TAG_H
14561 +
14562 +#include <linux/types.h>
14563 +
14564 +
14565 +#define DX_TAG(in)     (IS_TAGGED(in))
14566 +
14567 +
14568 +#ifdef CONFIG_DX_TAG_NFSD
14569 +#define DX_TAG_NFSD    1
14570 +#else
14571 +#define DX_TAG_NFSD    0
14572 +#endif
14573 +
14574 +
14575 +#ifdef CONFIG_TAGGING_NONE
14576 +
14577 +#define MAX_UID                0xFFFFFFFF
14578 +#define MAX_GID                0xFFFFFFFF
14579 +
14580 +#define INOTAG_TAG(cond, uid, gid, tag)        (0)
14581 +
14582 +#define TAGINO_UID(cond, uid, tag)     (uid)
14583 +#define TAGINO_GID(cond, gid, tag)     (gid)
14584 +
14585 +#endif
14586 +
14587 +
14588 +#ifdef CONFIG_TAGGING_GID16
14589 +
14590 +#define MAX_UID                0xFFFFFFFF
14591 +#define MAX_GID                0x0000FFFF
14592 +
14593 +#define INOTAG_TAG(cond, uid, gid, tag)        \
14594 +       ((cond) ? (((gid) >> 16) & 0xFFFF) : 0)
14595 +
14596 +#define TAGINO_UID(cond, uid, tag)     (uid)
14597 +#define TAGINO_GID(cond, gid, tag)     \
14598 +       ((cond) ? (((gid) & 0xFFFF) | ((tag) << 16)) : (gid))
14599 +
14600 +#endif
14601 +
14602 +
14603 +#ifdef CONFIG_TAGGING_ID24
14604 +
14605 +#define MAX_UID                0x00FFFFFF
14606 +#define MAX_GID                0x00FFFFFF
14607 +
14608 +#define INOTAG_TAG(cond, uid, gid, tag)        \
14609 +       ((cond) ? ((((uid) >> 16) & 0xFF00) | (((gid) >> 24) & 0xFF)) : 0)
14610 +
14611 +#define TAGINO_UID(cond, uid, tag)     \
14612 +       ((cond) ? (((uid) & 0xFFFFFF) | (((tag) & 0xFF00) << 16)) : (uid))
14613 +#define TAGINO_GID(cond, gid, tag)     \
14614 +       ((cond) ? (((gid) & 0xFFFFFF) | (((tag) & 0x00FF) << 24)) : (gid))
14615 +
14616 +#endif
14617 +
14618 +
14619 +#ifdef CONFIG_TAGGING_UID16
14620 +
14621 +#define MAX_UID                0x0000FFFF
14622 +#define MAX_GID                0xFFFFFFFF
14623 +
14624 +#define INOTAG_TAG(cond, uid, gid, tag)        \
14625 +       ((cond) ? (((uid) >> 16) & 0xFFFF) : 0)
14626 +
14627 +#define TAGINO_UID(cond, uid, tag)     \
14628 +       ((cond) ? (((uid) & 0xFFFF) | ((tag) << 16)) : (uid))
14629 +#define TAGINO_GID(cond, gid, tag)     (gid)
14630 +
14631 +#endif
14632 +
14633 +
14634 +#ifdef CONFIG_TAGGING_INTERN
14635 +
14636 +#define MAX_UID                0xFFFFFFFF
14637 +#define MAX_GID                0xFFFFFFFF
14638 +
14639 +#define INOTAG_TAG(cond, uid, gid, tag)        \
14640 +       ((cond) ? (tag) : 0)
14641 +
14642 +#define TAGINO_UID(cond, uid, tag)     (uid)
14643 +#define TAGINO_GID(cond, gid, tag)     (gid)
14644 +
14645 +#endif
14646 +
14647 +
14648 +#ifndef CONFIG_TAGGING_NONE
14649 +#define dx_current_fstag(sb)   \
14650 +       ((sb)->s_flags & MS_TAGGED ? dx_current_tag() : 0)
14651 +#else
14652 +#define dx_current_fstag(sb)   (0)
14653 +#endif
14654 +
14655 +#ifndef CONFIG_TAGGING_INTERN
14656 +#define TAGINO_TAG(cond, tag)  (0)
14657 +#else
14658 +#define TAGINO_TAG(cond, tag)  ((cond) ? (tag) : 0)
14659 +#endif
14660 +
14661 +#define INOTAG_UID(cond, uid, gid)     \
14662 +       ((cond) ? ((uid) & MAX_UID) : (uid))
14663 +#define INOTAG_GID(cond, uid, gid)     \
14664 +       ((cond) ? ((gid) & MAX_GID) : (gid))
14665 +
14666 +
14667 +static inline uid_t dx_map_uid(uid_t uid)
14668 +{
14669 +       if ((uid > MAX_UID) && (uid != -1))
14670 +               uid = -2;
14671 +       return (uid & MAX_UID);
14672 +}
14673 +
14674 +static inline gid_t dx_map_gid(gid_t gid)
14675 +{
14676 +       if ((gid > MAX_GID) && (gid != -1))
14677 +               gid = -2;
14678 +       return (gid & MAX_GID);
14679 +}
14680 +
14681 +struct peer_tag {
14682 +       int32_t xid;
14683 +       int32_t nid;
14684 +};
14685 +
14686 +#define dx_notagcheck(nd) ((nd) && (nd)->path.mnt && \
14687 +       ((nd)->path.mnt->mnt_flags & MNT_NOTAGCHECK))
14688 +
14689 +int dx_parse_tag(char *string, tag_t *tag, int remove);
14690 +
14691 +#ifdef CONFIG_PROPAGATE
14692 +
14693 +void __dx_propagate_tag(struct nameidata *nd, struct inode *inode);
14694 +
14695 +#define dx_propagate_tag(n, i) __dx_propagate_tag(n, i)
14696 +
14697 +#else
14698 +#define dx_propagate_tag(n, i) do { } while (0)
14699 +#endif
14700 +
14701 +#endif /* _DX_TAG_H */
14702 diff -NurpP --minimal linux-2.6.25.11/include/linux/vs_inet6.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_inet6.h
14703 --- linux-2.6.25.11/include/linux/vs_inet6.h    1969-12-31 19:00:00.000000000 -0500
14704 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_inet6.h      2008-04-19 15:14:52.000000000 -0400
14705 @@ -0,0 +1,229 @@
14706 +#ifndef _VS_INET6_H
14707 +#define _VS_INET6_H
14708 +
14709 +#include "vserver/base.h"
14710 +#include "vserver/network.h"
14711 +#include "vserver/debug.h"
14712 +
14713 +#include <net/ipv6.h>
14714 +
14715 +#define NXAV6(a)       NIP6((a)->ip), NIP6((a)->mask), (a)->prefix, (a)->type
14716 +#define NXAV6_FMT      "[" NIP6_FMT "/" NIP6_FMT "/%d:%04x]"
14717 +
14718 +
14719 +#ifdef CONFIG_IPV6
14720 +
14721 +static inline
14722 +int v6_addr_match(struct nx_addr_v6 *nxa,
14723 +       const struct in6_addr *addr, uint16_t mask)
14724 +{
14725 +       switch (nxa->type & mask) {
14726 +       case NXA_TYPE_MASK:
14727 +               return ipv6_masked_addr_cmp(&nxa->ip, &nxa->mask, addr);
14728 +       case NXA_TYPE_ADDR:
14729 +               return ipv6_addr_equal(&nxa->ip, addr);
14730 +       case NXA_TYPE_ANY:
14731 +               return 1;
14732 +       default:
14733 +               return 0;
14734 +       }
14735 +}
14736 +
14737 +static inline
14738 +int v6_addr_in_nx_info(struct nx_info *nxi,
14739 +       const struct in6_addr *addr, uint16_t mask)
14740 +{
14741 +       struct nx_addr_v6 *nxa;
14742 +
14743 +       if (!nxi)
14744 +               return 1;
14745 +       for (nxa = &nxi->v6; nxa; nxa = nxa->next)
14746 +               if (v6_addr_match(nxa, addr, mask))
14747 +                       return 1;
14748 +       return 0;
14749 +}
14750 +
14751 +static inline
14752 +int v6_nx_addr_match(struct nx_addr_v6 *nxa, struct nx_addr_v6 *addr, uint16_t mask)
14753 +{
14754 +       /* FIXME: needs full range checks */
14755 +       return v6_addr_match(nxa, &addr->ip, mask);
14756 +}
14757 +
14758 +static inline
14759 +int v6_nx_addr_in_nx_info(struct nx_info *nxi, struct nx_addr_v6 *nxa, uint16_t mask)
14760 +{
14761 +       struct nx_addr_v6 *ptr;
14762 +
14763 +       for (ptr = &nxi->v6; ptr; ptr = ptr->next)
14764 +               if (v6_nx_addr_match(ptr, nxa, mask))
14765 +                       return 1;
14766 +       return 0;
14767 +}
14768 +
14769 +
14770 +/*
14771 + *     Check if a given address matches for a socket
14772 + *
14773 + *     nxi:            the socket's nx_info if any
14774 + *     addr:           to be verified address
14775 + */
14776 +static inline
14777 +int v6_sock_addr_match (
14778 +       struct nx_info *nxi,
14779 +       struct inet_sock *inet,
14780 +       struct in6_addr *addr)
14781 +{
14782 +       struct sock *sk = &inet->sk;
14783 +       struct in6_addr *saddr = inet6_rcv_saddr(sk);
14784 +
14785 +       if (!ipv6_addr_any(addr) &&
14786 +               ipv6_addr_equal(saddr, addr))
14787 +               return 1;
14788 +       if (ipv6_addr_any(saddr))
14789 +               return v6_addr_in_nx_info(nxi, addr, -1);
14790 +       return 0;
14791 +}
14792 +
14793 +/*
14794 + *     check if address is covered by socket
14795 + *
14796 + *     sk:     the socket to check against
14797 + *     addr:   the address in question (must be != 0)
14798 + */
14799 +
14800 +static inline
14801 +int __v6_addr_match_socket(const struct sock *sk, struct nx_addr_v6 *nxa)
14802 +{
14803 +       struct nx_info *nxi = sk->sk_nx_info;
14804 +       struct in6_addr *saddr = inet6_rcv_saddr(sk);
14805 +
14806 +       vxdprintk(VXD_CBIT(net, 5),
14807 +               "__v6_addr_in_socket(%p," NXAV6_FMT ") %p:" NIP6_FMT " %p;%lx",
14808 +               sk, NXAV6(nxa), nxi, NIP6(*saddr), sk->sk_socket,
14809 +               (sk->sk_socket?sk->sk_socket->flags:0));
14810 +
14811 +       if (!ipv6_addr_any(saddr)) {    /* direct address match */
14812 +               return v6_addr_match(nxa, saddr, -1);
14813 +       } else if (nxi) {               /* match against nx_info */
14814 +               return v6_nx_addr_in_nx_info(nxi, nxa, -1);
14815 +       } else {                        /* unrestricted any socket */
14816 +               return 1;
14817 +       }
14818 +}
14819 +
14820 +
14821 +/* inet related checks and helpers */
14822 +
14823 +
14824 +struct in_ifaddr;
14825 +struct net_device;
14826 +struct sock;
14827 +
14828 +
14829 +#include <linux/netdevice.h>
14830 +#include <linux/inetdevice.h>
14831 +#include <net/inet_timewait_sock.h>
14832 +
14833 +
14834 +int dev_in_nx_info(struct net_device *, struct nx_info *);
14835 +int v6_dev_in_nx_info(struct net_device *, struct nx_info *);
14836 +int nx_v6_addr_conflict(struct nx_info *, struct nx_info *);
14837 +
14838 +
14839 +
14840 +static inline
14841 +int v6_ifa_in_nx_info(struct inet6_ifaddr *ifa, struct nx_info *nxi)
14842 +{
14843 +       if (!nxi)
14844 +               return 1;
14845 +       if (!ifa)
14846 +               return 0;
14847 +       return v6_addr_in_nx_info(nxi, &ifa->addr, -1);
14848 +}
14849 +
14850 +static inline
14851 +int nx_v6_ifa_visible(struct nx_info *nxi, struct inet6_ifaddr *ifa)
14852 +{
14853 +       if (!nx_info_flags(nxi, NXF_HIDE_NETIF, 0))
14854 +               return 1;
14855 +       if (v6_ifa_in_nx_info(ifa, nxi))
14856 +               return 1;
14857 +       return 0;
14858 +}
14859 +
14860 +
14861 +struct nx_v6_sock_addr {
14862 +       struct in6_addr saddr;  /* Address used for validation */
14863 +       struct in6_addr baddr;  /* Address used for socket bind */
14864 +};
14865 +
14866 +static inline
14867 +int v6_map_sock_addr(struct inet_sock *inet, struct sockaddr_in6 *addr,
14868 +       struct nx_v6_sock_addr *nsa)
14869 +{
14870 +       // struct sock *sk = &inet->sk;
14871 +       // struct nx_info *nxi = sk->sk_nx_info;
14872 +       struct in6_addr saddr = addr->sin6_addr;
14873 +       struct in6_addr baddr = saddr;
14874 +
14875 +       nsa->saddr = saddr;
14876 +       nsa->baddr = baddr;
14877 +       return 0;
14878 +}
14879 +
14880 +static inline
14881 +void v6_set_sock_addr(struct inet_sock *inet, struct nx_v6_sock_addr *nsa)
14882 +{
14883 +       // struct sock *sk = &inet->sk;
14884 +       // struct in6_addr *saddr = inet6_rcv_saddr(sk);
14885 +
14886 +       // *saddr = nsa->baddr;
14887 +       // inet->saddr = nsa->baddr;
14888 +}
14889 +
14890 +static inline
14891 +int nx_info_has_v6(struct nx_info *nxi)
14892 +{
14893 +       if (!nxi)
14894 +               return 1;
14895 +       if (NX_IPV6(nxi))
14896 +               return 1;
14897 +       return 0;
14898 +}
14899 +
14900 +#else /* CONFIG_IPV6 */
14901 +
14902 +static inline
14903 +int nx_v6_dev_visible(struct nx_info *n, struct net_device *d)
14904 +{
14905 +       return 1;
14906 +}
14907 +
14908 +
14909 +static inline
14910 +int nx_v6_addr_conflict(struct nx_info *n, uint32_t a, const struct sock *s)
14911 +{
14912 +       return 1;
14913 +}
14914 +
14915 +static inline
14916 +int v6_ifa_in_nx_info(struct in_ifaddr *a, struct nx_info *n)
14917 +{
14918 +       return 1;
14919 +}
14920 +
14921 +static inline
14922 +int nx_info_has_v6(struct nx_info *nxi)
14923 +{
14924 +       return 0;
14925 +}
14926 +
14927 +#endif /* CONFIG_IPV6 */
14928 +
14929 +#define current_nx_info_has_v6() \
14930 +       nx_info_has_v6(current_nx_info())
14931 +
14932 +#else
14933 +#warning duplicate inclusion
14934 +#endif
14935 diff -NurpP --minimal linux-2.6.25.11/include/linux/vs_inet.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_inet.h
14936 --- linux-2.6.25.11/include/linux/vs_inet.h     1969-12-31 19:00:00.000000000 -0500
14937 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_inet.h       2008-04-19 15:14:52.000000000 -0400
14938 @@ -0,0 +1,342 @@
14939 +#ifndef _VS_INET_H
14940 +#define _VS_INET_H
14941 +
14942 +#include "vserver/base.h"
14943 +#include "vserver/network.h"
14944 +#include "vserver/debug.h"
14945 +
14946 +#define IPI_LOOPBACK   htonl(INADDR_LOOPBACK)
14947 +
14948 +#define NXAV4(a)       NIPQUAD((a)->ip[0]), NIPQUAD((a)->ip[1]), \
14949 +                       NIPQUAD((a)->mask), (a)->type
14950 +#define NXAV4_FMT      "[" NIPQUAD_FMT "-" NIPQUAD_FMT "/" NIPQUAD_FMT ":%04x]"
14951 +
14952 +
14953 +static inline
14954 +int v4_addr_match(struct nx_addr_v4 *nxa, __be32 addr, uint16_t tmask)
14955 +{
14956 +       __be32 ip = nxa->ip[0].s_addr;
14957 +       __be32 mask = nxa->mask.s_addr;
14958 +       __be32 bcast = ip | ~mask;
14959 +       int ret = 0;
14960 +
14961 +       switch (nxa->type & tmask) {
14962 +       case NXA_TYPE_MASK:
14963 +               ret = (ip == (addr & mask));
14964 +               break;
14965 +       case NXA_TYPE_ADDR:
14966 +               ret = 3;
14967 +               if (addr == ip)
14968 +                       break;
14969 +               /* fall through to broadcast */
14970 +       case NXA_MOD_BCAST:
14971 +               ret = ((tmask & NXA_MOD_BCAST) && (addr == bcast));
14972 +               break;
14973 +       case NXA_TYPE_RANGE:
14974 +               ret = ((nxa->ip[0].s_addr <= addr) &&
14975 +                       (nxa->ip[1].s_addr > addr));
14976 +               break;
14977 +       case NXA_TYPE_ANY:
14978 +               ret = 2;
14979 +               break;
14980 +       }
14981 +
14982 +       vxdprintk(VXD_CBIT(net, 0),
14983 +               "v4_addr_match(%p" NXAV4_FMT "," NIPQUAD_FMT ",%04x) = %d",
14984 +               nxa, NXAV4(nxa), NIPQUAD(addr), tmask, ret);
14985 +       return ret;
14986 +}
14987 +
14988 +static inline
14989 +int v4_addr_in_nx_info(struct nx_info *nxi, __be32 addr, uint16_t tmask)
14990 +{
14991 +       struct nx_addr_v4 *nxa;
14992 +       int ret = 1;
14993 +
14994 +       if (!nxi)
14995 +               goto out;
14996 +
14997 +       ret = 2;
14998 +       /* allow 127.0.0.1 when remapping lback */
14999 +       if ((tmask & NXA_LOOPBACK) &&
15000 +               (addr == IPI_LOOPBACK) &&
15001 +               nx_info_flags(nxi, NXF_LBACK_REMAP, 0))
15002 +               goto out;
15003 +       ret = 3;
15004 +       /* check for lback address */
15005 +       if ((tmask & NXA_MOD_LBACK) &&
15006 +               (nxi->v4_lback.s_addr == addr))
15007 +               goto out;
15008 +       ret = 4;
15009 +       /* check for broadcast address */
15010 +       if ((tmask & NXA_MOD_BCAST) &&
15011 +               (nxi->v4_bcast.s_addr == addr))
15012 +               goto out;
15013 +       ret = 5;
15014 +       /* check for v4 addresses */
15015 +       for (nxa = &nxi->v4; nxa; nxa = nxa->next)
15016 +               if (v4_addr_match(nxa, addr, tmask))
15017 +                       goto out;
15018 +       ret = 0;
15019 +out:
15020 +       vxdprintk(VXD_CBIT(net, 0),
15021 +               "v4_addr_in_nx_info(%p[#%u]," NIPQUAD_FMT ",%04x) = %d",
15022 +               nxi, nxi ? nxi->nx_id : 0, NIPQUAD(addr), tmask, ret);
15023 +       return ret;
15024 +}
15025 +
15026 +static inline
15027 +int v4_nx_addr_match(struct nx_addr_v4 *nxa, struct nx_addr_v4 *addr, uint16_t mask)
15028 +{
15029 +       /* FIXME: needs full range checks */
15030 +       return v4_addr_match(nxa, addr->ip[0].s_addr, mask);
15031 +}
15032 +
15033 +static inline
15034 +int v4_nx_addr_in_nx_info(struct nx_info *nxi, struct nx_addr_v4 *nxa, uint16_t mask)
15035 +{
15036 +       struct nx_addr_v4 *ptr;
15037 +
15038 +       for (ptr = &nxi->v4; ptr; ptr = ptr->next)
15039 +               if (v4_nx_addr_match(ptr, nxa, mask))
15040 +                       return 1;
15041 +       return 0;
15042 +}
15043 +
15044 +#include <net/inet_sock.h>
15045 +
15046 +/*
15047 + *     Check if a given address matches for a socket
15048 + *
15049 + *     nxi:            the socket's nx_info if any
15050 + *     addr:           to be verified address
15051 + */
15052 +static inline
15053 +int v4_sock_addr_match (
15054 +       struct nx_info *nxi,
15055 +       struct inet_sock *inet,
15056 +       __be32 addr)
15057 +{
15058 +       __be32 saddr = inet->rcv_saddr;
15059 +       __be32 bcast = nxi ? nxi->v4_bcast.s_addr : INADDR_BROADCAST;
15060 +
15061 +       if (addr && (saddr == addr || bcast == addr))
15062 +               return 1;
15063 +       if (!saddr)
15064 +               return v4_addr_in_nx_info(nxi, addr, NXA_MASK_BIND);
15065 +       return 0;
15066 +}
15067 +
15068 +
15069 +/* inet related checks and helpers */
15070 +
15071 +
15072 +struct in_ifaddr;
15073 +struct net_device;
15074 +struct sock;
15075 +
15076 +#ifdef CONFIG_INET
15077 +
15078 +#include <linux/netdevice.h>
15079 +#include <linux/inetdevice.h>
15080 +#include <net/inet_sock.h>
15081 +#include <net/inet_timewait_sock.h>
15082 +
15083 +
15084 +int dev_in_nx_info(struct net_device *, struct nx_info *);
15085 +int v4_dev_in_nx_info(struct net_device *, struct nx_info *);
15086 +int nx_v4_addr_conflict(struct nx_info *, struct nx_info *);
15087 +
15088 +
15089 +/*
15090 + *     check if address is covered by socket
15091 + *
15092 + *     sk:     the socket to check against
15093 + *     addr:   the address in question (must be != 0)
15094 + */
15095 +
15096 +static inline
15097 +int __v4_addr_match_socket(const struct sock *sk, struct nx_addr_v4 *nxa)
15098 +{
15099 +       struct nx_info *nxi = sk->sk_nx_info;
15100 +       __be32 saddr = inet_rcv_saddr(sk);
15101 +
15102 +       vxdprintk(VXD_CBIT(net, 5),
15103 +               "__v4_addr_in_socket(%p," NXAV4_FMT ") %p:" NIPQUAD_FMT " %p;%lx",
15104 +               sk, NXAV4(nxa), nxi, NIPQUAD(saddr), sk->sk_socket,
15105 +               (sk->sk_socket?sk->sk_socket->flags:0));
15106 +
15107 +       if (saddr) {            /* direct address match */
15108 +               return v4_addr_match(nxa, saddr, -1);
15109 +       } else if (nxi) {       /* match against nx_info */
15110 +               return v4_nx_addr_in_nx_info(nxi, nxa, -1);
15111 +       } else {                /* unrestricted any socket */
15112 +               return 1;
15113 +       }
15114 +}
15115 +
15116 +
15117 +
15118 +static inline
15119 +int nx_dev_visible(struct nx_info *nxi, struct net_device *dev)
15120 +{
15121 +       vxdprintk(VXD_CBIT(net, 1), "nx_dev_visible(%p[#%u],%p »%s«) %d",
15122 +               nxi, nxi ? nxi->nx_id : 0, dev, dev->name,
15123 +               nxi ? dev_in_nx_info(dev, nxi) : 0);
15124 +
15125 +       if (!nx_info_flags(nxi, NXF_HIDE_NETIF, 0))
15126 +               return 1;
15127 +       if (dev_in_nx_info(dev, nxi))
15128 +               return 1;
15129 +       return 0;
15130 +}
15131 +
15132 +
15133 +static inline
15134 +int v4_ifa_in_nx_info(struct in_ifaddr *ifa, struct nx_info *nxi)
15135 +{
15136 +       if (!nxi)
15137 +               return 1;
15138 +       if (!ifa)
15139 +               return 0;
15140 +       return v4_addr_in_nx_info(nxi, ifa->ifa_local, NXA_MASK_SHOW);
15141 +}
15142 +
15143 +static inline
15144 +int nx_v4_ifa_visible(struct nx_info *nxi, struct in_ifaddr *ifa)
15145 +{
15146 +       vxdprintk(VXD_CBIT(net, 1), "nx_v4_ifa_visible(%p[#%u],%p) %d",
15147 +               nxi, nxi ? nxi->nx_id : 0, ifa,
15148 +               nxi ? v4_ifa_in_nx_info(ifa, nxi) : 0);
15149 +
15150 +       if (!nx_info_flags(nxi, NXF_HIDE_NETIF, 0))
15151 +               return 1;
15152 +       if (v4_ifa_in_nx_info(ifa, nxi))
15153 +               return 1;
15154 +       return 0;
15155 +}
15156 +
15157 +
15158 +struct nx_v4_sock_addr {
15159 +       __be32 saddr;   /* Address used for validation */
15160 +       __be32 baddr;   /* Address used for socket bind */
15161 +};
15162 +
15163 +static inline
15164 +int v4_map_sock_addr(struct inet_sock *inet, struct sockaddr_in *addr,
15165 +       struct nx_v4_sock_addr *nsa)
15166 +{
15167 +       struct sock *sk = &inet->sk;
15168 +       struct nx_info *nxi = sk->sk_nx_info;
15169 +       __be32 saddr = addr->sin_addr.s_addr;
15170 +       __be32 baddr = saddr;
15171 +
15172 +       vxdprintk(VXD_CBIT(net, 3),
15173 +               "inet_bind(%p)* %p,%p;%lx " NIPQUAD_FMT,
15174 +               sk, sk->sk_nx_info, sk->sk_socket,
15175 +               (sk->sk_socket ? sk->sk_socket->flags : 0),
15176 +               NIPQUAD(saddr));
15177 +
15178 +       if (nxi) {
15179 +               if (saddr == INADDR_ANY) {
15180 +                       if (nx_info_flags(nxi, NXF_SINGLE_IP, 0))
15181 +                               baddr = nxi->v4.ip[0].s_addr;
15182 +               } else if (saddr == IPI_LOOPBACK) {
15183 +                       if (nx_info_flags(nxi, NXF_LBACK_REMAP, 0))
15184 +                               baddr = nxi->v4_lback.s_addr;
15185 +               } else {        /* normal address bind */
15186 +                       if (!v4_addr_in_nx_info(nxi, saddr, NXA_MASK_BIND))
15187 +                               return -EADDRNOTAVAIL;
15188 +               }
15189 +       }
15190 +
15191 +       vxdprintk(VXD_CBIT(net, 3),
15192 +               "inet_bind(%p) " NIPQUAD_FMT ", " NIPQUAD_FMT,
15193 +               sk, NIPQUAD(saddr), NIPQUAD(baddr));
15194 +
15195 +       nsa->saddr = saddr;
15196 +       nsa->baddr = baddr;
15197 +       return 0;
15198 +}
15199 +
15200 +static inline
15201 +void v4_set_sock_addr(struct inet_sock *inet, struct nx_v4_sock_addr *nsa)
15202 +{
15203 +       inet->saddr = nsa->baddr;
15204 +       inet->rcv_saddr = nsa->baddr;
15205 +}
15206 +
15207 +
15208 +/*
15209 + *      helper to simplify inet_lookup_listener
15210 + *
15211 + *      nxi:   the socket's nx_info if any
15212 + *      addr:  to be verified address
15213 + *      saddr: socket address
15214 + */
15215 +static inline int v4_inet_addr_match (
15216 +       struct nx_info *nxi,
15217 +       __be32 addr,
15218 +       __be32 saddr)
15219 +{
15220 +       if (addr && (saddr == addr))
15221 +               return 1;
15222 +       if (!saddr)
15223 +               return nxi ? v4_addr_in_nx_info(nxi, addr, NXA_MASK_BIND) : 1;
15224 +       return 0;
15225 +}
15226 +
15227 +static inline __be32 nx_map_sock_lback(struct nx_info *nxi, __be32 addr)
15228 +{
15229 +       if (nx_info_flags(nxi, NXF_HIDE_LBACK, 0) &&
15230 +               (addr == nxi->v4_lback.s_addr))
15231 +               return IPI_LOOPBACK;
15232 +       return addr;
15233 +}
15234 +
15235 +static inline
15236 +int nx_info_has_v4(struct nx_info *nxi)
15237 +{
15238 +       if (!nxi)
15239 +               return 1;
15240 +       if (NX_IPV4(nxi))
15241 +               return 1;
15242 +       if (nx_info_flags(nxi, NXF_LBACK_REMAP, 0))
15243 +               return 1;
15244 +       return 0;
15245 +}
15246 +
15247 +#else /* CONFIG_INET */
15248 +
15249 +static inline
15250 +int nx_dev_visible(struct nx_info *n, struct net_device *d)
15251 +{
15252 +       return 1;
15253 +}
15254 +
15255 +static inline
15256 +int nx_v4_addr_conflict(struct nx_info *n, uint32_t a, const struct sock *s)
15257 +{
15258 +       return 1;
15259 +}
15260 +
15261 +static inline
15262 +int v4_ifa_in_nx_info(struct in_ifaddr *a, struct nx_info *n)
15263 +{
15264 +       return 1;
15265 +}
15266 +
15267 +static inline
15268 +int nx_info_has_v4(struct nx_info *nxi)
15269 +{
15270 +       return 0;
15271 +}
15272 +
15273 +#endif /* CONFIG_INET */
15274 +
15275 +#define current_nx_info_has_v4() \
15276 +       nx_info_has_v4(current_nx_info())
15277 +
15278 +#else
15279 +// #warning duplicate inclusion
15280 +#endif
15281 diff -NurpP --minimal linux-2.6.25.11/include/linux/vs_limit.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_limit.h
15282 --- linux-2.6.25.11/include/linux/vs_limit.h    1969-12-31 19:00:00.000000000 -0500
15283 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_limit.h      2008-04-19 15:14:52.000000000 -0400
15284 @@ -0,0 +1,140 @@
15285 +#ifndef _VS_LIMIT_H
15286 +#define _VS_LIMIT_H
15287 +
15288 +#include "vserver/limit.h"
15289 +#include "vserver/base.h"
15290 +#include "vserver/context.h"
15291 +#include "vserver/debug.h"
15292 +#include "vserver/context.h"
15293 +#include "vserver/limit_int.h"
15294 +
15295 +
15296 +#define vx_acc_cres(v, d, p, r) \
15297 +       __vx_acc_cres(v, r, d, p, __FILE__, __LINE__)
15298 +
15299 +#define vx_acc_cres_cond(x, d, p, r) \
15300 +       __vx_acc_cres(((x) == vx_current_xid()) ? current->vx_info : 0, \
15301 +       r, d, p, __FILE__, __LINE__)
15302 +
15303 +
15304 +#define vx_add_cres(v, a, p, r) \
15305 +       __vx_add_cres(v, r, a, p, __FILE__, __LINE__)
15306 +#define vx_sub_cres(v, a, p, r)                vx_add_cres(v, -(a), p, r)
15307 +
15308 +#define vx_add_cres_cond(x, a, p, r) \
15309 +       __vx_add_cres(((x) == vx_current_xid()) ? current->vx_info : 0, \
15310 +       r, a, p, __FILE__, __LINE__)
15311 +#define vx_sub_cres_cond(x, a, p, r)   vx_add_cres_cond(x, -(a), p, r)
15312 +
15313 +
15314 +/* process and file limits */
15315 +
15316 +#define vx_nproc_inc(p) \
15317 +       vx_acc_cres((p)->vx_info, 1, p, RLIMIT_NPROC)
15318 +
15319 +#define vx_nproc_dec(p) \
15320 +       vx_acc_cres((p)->vx_info,-1, p, RLIMIT_NPROC)
15321 +
15322 +#define vx_files_inc(f) \
15323 +       vx_acc_cres_cond((f)->f_xid, 1, f, RLIMIT_NOFILE)
15324 +
15325 +#define vx_files_dec(f) \
15326 +       vx_acc_cres_cond((f)->f_xid,-1, f, RLIMIT_NOFILE)
15327 +
15328 +#define vx_locks_inc(l) \
15329 +       vx_acc_cres_cond((l)->fl_xid, 1, l, RLIMIT_LOCKS)
15330 +
15331 +#define vx_locks_dec(l) \
15332 +       vx_acc_cres_cond((l)->fl_xid,-1, l, RLIMIT_LOCKS)
15333 +
15334 +#define vx_openfd_inc(f) \
15335 +       vx_acc_cres(current->vx_info, 1, (void *)(long)(f), VLIMIT_OPENFD)
15336 +
15337 +#define vx_openfd_dec(f) \
15338 +       vx_acc_cres(current->vx_info,-1, (void *)(long)(f), VLIMIT_OPENFD)
15339 +
15340 +
15341 +#define vx_cres_avail(v, n, r) \
15342 +       __vx_cres_avail(v, r, n, __FILE__, __LINE__)
15343 +
15344 +
15345 +#define vx_nproc_avail(n) \
15346 +       vx_cres_avail(current->vx_info, n, RLIMIT_NPROC)
15347 +
15348 +#define vx_files_avail(n) \
15349 +       vx_cres_avail(current->vx_info, n, RLIMIT_NOFILE)
15350 +
15351 +#define vx_locks_avail(n) \
15352 +       vx_cres_avail(current->vx_info, n, RLIMIT_LOCKS)
15353 +
15354 +#define vx_openfd_avail(n) \
15355 +       vx_cres_avail(current->vx_info, n, VLIMIT_OPENFD)
15356 +
15357 +
15358 +/* dentry limits */
15359 +
15360 +#define vx_dentry_inc(d) do {                                          \
15361 +       if (atomic_read(&d->d_count) == 1)                              \
15362 +               vx_acc_cres(current->vx_info, 1, d, VLIMIT_DENTRY);     \
15363 +       } while (0)
15364 +
15365 +#define vx_dentry_dec(d) do {                                          \
15366 +       if (atomic_read(&d->d_count) == 0)                              \
15367 +               vx_acc_cres(current->vx_info,-1, d, VLIMIT_DENTRY);     \
15368 +       } while (0)
15369 +
15370 +#define vx_dentry_avail(n) \
15371 +       vx_cres_avail(current->vx_info, n, VLIMIT_DENTRY)
15372 +
15373 +
15374 +/* socket limits */
15375 +
15376 +#define vx_sock_inc(s) \
15377 +       vx_acc_cres((s)->sk_vx_info, 1, s, VLIMIT_NSOCK)
15378 +
15379 +#define vx_sock_dec(s) \
15380 +       vx_acc_cres((s)->sk_vx_info,-1, s, VLIMIT_NSOCK)
15381 +
15382 +#define vx_sock_avail(n) \
15383 +       vx_cres_avail(current->vx_info, n, VLIMIT_NSOCK)
15384 +
15385 +
15386 +/* ipc resource limits */
15387 +
15388 +#define vx_ipcmsg_add(v, u, a) \
15389 +       vx_add_cres(v, a, u, RLIMIT_MSGQUEUE)
15390 +
15391 +#define vx_ipcmsg_sub(v, u, a) \
15392 +       vx_sub_cres(v, a, u, RLIMIT_MSGQUEUE)
15393 +
15394 +#define vx_ipcmsg_avail(v, a) \
15395 +       vx_cres_avail(v, a, RLIMIT_MSGQUEUE)
15396 +
15397 +
15398 +#define vx_ipcshm_add(v, k, a) \
15399 +       vx_add_cres(v, a, (void *)(long)(k), VLIMIT_SHMEM)
15400 +
15401 +#define vx_ipcshm_sub(v, k, a) \
15402 +       vx_sub_cres(v, a, (void *)(long)(k), VLIMIT_SHMEM)
15403 +
15404 +#define vx_ipcshm_avail(v, a) \
15405 +       vx_cres_avail(v, a, VLIMIT_SHMEM)
15406 +
15407 +
15408 +#define vx_semary_inc(a) \
15409 +       vx_acc_cres(current->vx_info, 1, a, VLIMIT_SEMARY)
15410 +
15411 +#define vx_semary_dec(a) \
15412 +       vx_acc_cres(current->vx_info, -1, a, VLIMIT_SEMARY)
15413 +
15414 +
15415 +#define vx_nsems_add(a,n) \
15416 +       vx_add_cres(current->vx_info, n, a, VLIMIT_NSEMS)
15417 +
15418 +#define vx_nsems_sub(a,n) \
15419 +       vx_sub_cres(current->vx_info, n, a, VLIMIT_NSEMS)
15420 +
15421 +
15422 +#else
15423 +#warning duplicate inclusion
15424 +#endif
15425 diff -NurpP --minimal linux-2.6.25.11/include/linux/vs_memory.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_memory.h
15426 --- linux-2.6.25.11/include/linux/vs_memory.h   1969-12-31 19:00:00.000000000 -0500
15427 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_memory.h     2008-04-19 15:14:52.000000000 -0400
15428 @@ -0,0 +1,159 @@
15429 +#ifndef _VS_MEMORY_H
15430 +#define _VS_MEMORY_H
15431 +
15432 +#include "vserver/limit.h"
15433 +#include "vserver/base.h"
15434 +#include "vserver/context.h"
15435 +#include "vserver/debug.h"
15436 +#include "vserver/context.h"
15437 +#include "vserver/limit_int.h"
15438 +
15439 +
15440 +#define __acc_add_long(a, v)   (*(v) += (a))
15441 +#define __acc_inc_long(v)      (++*(v))
15442 +#define __acc_dec_long(v)      (--*(v))
15443 +
15444 +#if    NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS
15445 +#define __acc_add_atomic(a, v) atomic_long_add(a, v)
15446 +#define __acc_inc_atomic(v)    atomic_long_inc(v)
15447 +#define __acc_dec_atomic(v)    atomic_long_dec(v)
15448 +#else  /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
15449 +#define __acc_add_atomic(a, v) __acc_add_long(a, v)
15450 +#define __acc_inc_atomic(v)    __acc_inc_long(v)
15451 +#define __acc_dec_atomic(v)    __acc_dec_long(v)
15452 +#endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
15453 +
15454 +
15455 +#define vx_acc_page(m, d, v, r) do {                                   \
15456 +       if ((d) > 0)                                                    \
15457 +               __acc_inc_long(&(m)->v);                                \
15458 +       else                                                            \
15459 +               __acc_dec_long(&(m)->v);                                \
15460 +       __vx_acc_cres(m->mm_vx_info, r, d, m, __FILE__, __LINE__);      \
15461 +} while (0)
15462 +
15463 +#define vx_acc_page_atomic(m, d, v, r) do {                            \
15464 +       if ((d) > 0)                                                    \
15465 +               __acc_inc_atomic(&(m)->v);                              \
15466 +       else                                                            \
15467 +               __acc_dec_atomic(&(m)->v);                              \
15468 +       __vx_acc_cres(m->mm_vx_info, r, d, m, __FILE__, __LINE__);      \
15469 +} while (0)
15470 +
15471 +
15472 +#define vx_acc_pages(m, p, v, r) do {                                  \
15473 +       unsigned long __p = (p);                                        \
15474 +       __acc_add_long(__p, &(m)->v);                                   \
15475 +       __vx_add_cres(m->mm_vx_info, r, __p, m, __FILE__, __LINE__);    \
15476 +} while (0)
15477 +
15478 +#define vx_acc_pages_atomic(m, p, v, r) do {                           \
15479 +       unsigned long __p = (p);                                        \
15480 +       __acc_add_atomic(__p, &(m)->v);                                 \
15481 +       __vx_add_cres(m->mm_vx_info, r, __p, m, __FILE__, __LINE__);    \
15482 +} while (0)
15483 +
15484 +
15485 +
15486 +#define vx_acc_vmpage(m, d) \
15487 +       vx_acc_page(m, d, total_vm,  RLIMIT_AS)
15488 +#define vx_acc_vmlpage(m, d) \
15489 +       vx_acc_page(m, d, locked_vm, RLIMIT_MEMLOCK)
15490 +#define vx_acc_file_rsspage(m, d) \
15491 +       vx_acc_page_atomic(m, d, _file_rss, VLIMIT_MAPPED)
15492 +#define vx_acc_anon_rsspage(m, d) \
15493 +       vx_acc_page_atomic(m, d, _anon_rss, VLIMIT_ANON)
15494 +
15495 +#define vx_acc_vmpages(m, p) \
15496 +       vx_acc_pages(m, p, total_vm,  RLIMIT_AS)
15497 +#define vx_acc_vmlpages(m, p) \
15498 +       vx_acc_pages(m, p, locked_vm, RLIMIT_MEMLOCK)
15499 +#define vx_acc_file_rsspages(m, p) \
15500 +       vx_acc_pages_atomic(m, p, _file_rss, VLIMIT_MAPPED)
15501 +#define vx_acc_anon_rsspages(m, p) \
15502 +       vx_acc_pages_atomic(m, p, _anon_rss, VLIMIT_ANON)
15503 +
15504 +#define vx_pages_add(s, r, p)  __vx_add_cres(s, r, p, 0, __FILE__, __LINE__)
15505 +#define vx_pages_sub(s, r, p)  vx_pages_add(s, r, -(p))
15506 +
15507 +#define vx_vmpages_inc(m)              vx_acc_vmpage(m, 1)
15508 +#define vx_vmpages_dec(m)              vx_acc_vmpage(m, -1)
15509 +#define vx_vmpages_add(m, p)           vx_acc_vmpages(m, p)
15510 +#define vx_vmpages_sub(m, p)           vx_acc_vmpages(m, -(p))
15511 +
15512 +#define vx_vmlocked_inc(m)             vx_acc_vmlpage(m, 1)
15513 +#define vx_vmlocked_dec(m)             vx_acc_vmlpage(m, -1)
15514 +#define vx_vmlocked_add(m, p)          vx_acc_vmlpages(m, p)
15515 +#define vx_vmlocked_sub(m, p)          vx_acc_vmlpages(m, -(p))
15516 +
15517 +#define vx_file_rsspages_inc(m)                vx_acc_file_rsspage(m, 1)
15518 +#define vx_file_rsspages_dec(m)                vx_acc_file_rsspage(m, -1)
15519 +#define vx_file_rsspages_add(m, p)     vx_acc_file_rsspages(m, p)
15520 +#define vx_file_rsspages_sub(m, p)     vx_acc_file_rsspages(m, -(p))
15521 +
15522 +#define vx_anon_rsspages_inc(m)                vx_acc_anon_rsspage(m, 1)
15523 +#define vx_anon_rsspages_dec(m)                vx_acc_anon_rsspage(m, -1)
15524 +#define vx_anon_rsspages_add(m, p)     vx_acc_anon_rsspages(m, p)
15525 +#define vx_anon_rsspages_sub(m, p)     vx_acc_anon_rsspages(m, -(p))
15526 +
15527 +
15528 +#define vx_pages_avail(m, p, r) \
15529 +       __vx_cres_avail((m)->mm_vx_info, r, p, __FILE__, __LINE__)
15530 +
15531 +#define vx_vmpages_avail(m, p) vx_pages_avail(m, p, RLIMIT_AS)
15532 +#define vx_vmlocked_avail(m, p)        vx_pages_avail(m, p, RLIMIT_MEMLOCK)
15533 +#define vx_anon_avail(m, p)    vx_pages_avail(m, p, VLIMIT_ANON)
15534 +#define vx_mapped_avail(m, p)  vx_pages_avail(m, p, VLIMIT_MAPPED)
15535 +
15536 +#define vx_rss_avail(m, p) \
15537 +       __vx_cres_array_avail((m)->mm_vx_info, VLA_RSS, p, __FILE__, __LINE__)
15538 +
15539 +
15540 +enum {
15541 +       VXPT_UNKNOWN = 0,
15542 +       VXPT_ANON,
15543 +       VXPT_NONE,
15544 +       VXPT_FILE,
15545 +       VXPT_SWAP,
15546 +       VXPT_WRITE
15547 +};
15548 +
15549 +#if 0
15550 +#define        vx_page_fault(mm, vma, type, ret)
15551 +#else
15552 +
15553 +static inline
15554 +void __vx_page_fault(struct mm_struct *mm,
15555 +       struct vm_area_struct *vma, int type, int ret)
15556 +{
15557 +       struct vx_info *vxi = mm->mm_vx_info;
15558 +       int what;
15559 +/*
15560 +       static char *page_type[6] =
15561 +               { "UNKNOWN", "ANON", "NONE", "FILE", "SWAP", "WRITE" };
15562 +       static char *page_what[4] =
15563 +               { "FAULT_OOM", "FAULT_SIGBUS", "FAULT_MINOR", "FAULT_MAJOR" };
15564 +*/
15565 +
15566 +       if (!vxi)
15567 +               return;
15568 +
15569 +       what = (ret & 0x3);
15570 +
15571 +/*     printk("[%d] page[%d][%d] %2x %s %s\n", vxi->vx_id,
15572 +               type, what, ret, page_type[type], page_what[what]);
15573 +*/
15574 +       if (ret & VM_FAULT_WRITE)
15575 +               what |= 0x4;
15576 +       atomic_inc(&vxi->cacct.page[type][what]);
15577 +}
15578 +
15579 +#define        vx_page_fault(mm, vma, type, ret)       __vx_page_fault(mm, vma, type, ret)
15580 +#endif
15581 +
15582 +
15583 +extern unsigned long vx_badness(struct task_struct *task, struct mm_struct *mm);
15584 +
15585 +#else
15586 +#warning duplicate inclusion
15587 +#endif
15588 diff -NurpP --minimal linux-2.6.25.11/include/linux/vs_network.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_network.h
15589 --- linux-2.6.25.11/include/linux/vs_network.h  1969-12-31 19:00:00.000000000 -0500
15590 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_network.h    2008-04-29 18:42:49.000000000 -0400
15591 @@ -0,0 +1,169 @@
15592 +#ifndef _NX_VS_NETWORK_H
15593 +#define _NX_VS_NETWORK_H
15594 +
15595 +#include "vserver/context.h"
15596 +#include "vserver/network.h"
15597 +#include "vserver/base.h"
15598 +#include "vserver/check.h"
15599 +#include "vserver/debug.h"
15600 +
15601 +#include <linux/sched.h>
15602 +
15603 +
15604 +#define get_nx_info(i) __get_nx_info(i, __FILE__, __LINE__)
15605 +
15606 +static inline struct nx_info *__get_nx_info(struct nx_info *nxi,
15607 +       const char *_file, int _line)
15608 +{
15609 +       if (!nxi)
15610 +               return NULL;
15611 +
15612 +       vxlprintk(VXD_CBIT(nid, 2), "get_nx_info(%p[#%d.%d])",
15613 +               nxi, nxi ? nxi->nx_id : 0,
15614 +               nxi ? atomic_read(&nxi->nx_usecnt) : 0,
15615 +               _file, _line);
15616 +
15617 +       atomic_inc(&nxi->nx_usecnt);
15618 +       return nxi;
15619 +}
15620 +
15621 +
15622 +extern void free_nx_info(struct nx_info *);
15623 +
15624 +#define put_nx_info(i) __put_nx_info(i, __FILE__, __LINE__)
15625 +
15626 +static inline void __put_nx_info(struct nx_info *nxi, const char *_file, int _line)
15627 +{
15628 +       if (!nxi)
15629 +               return;
15630 +
15631 +       vxlprintk(VXD_CBIT(nid, 2), "put_nx_info(%p[#%d.%d])",
15632 +               nxi, nxi ? nxi->nx_id : 0,
15633 +               nxi ? atomic_read(&nxi->nx_usecnt) : 0,
15634 +               _file, _line);
15635 +
15636 +       if (atomic_dec_and_test(&nxi->nx_usecnt))
15637 +               free_nx_info(nxi);
15638 +}
15639 +
15640 +
15641 +#define init_nx_info(p, i) __init_nx_info(p, i, __FILE__, __LINE__)
15642 +
15643 +static inline void __init_nx_info(struct nx_info **nxp, struct nx_info *nxi,
15644 +               const char *_file, int _line)
15645 +{
15646 +       if (nxi) {
15647 +               vxlprintk(VXD_CBIT(nid, 3),
15648 +                       "init_nx_info(%p[#%d.%d])",
15649 +                       nxi, nxi ? nxi->nx_id : 0,
15650 +                       nxi ? atomic_read(&nxi->nx_usecnt) : 0,
15651 +                       _file, _line);
15652 +
15653 +               atomic_inc(&nxi->nx_usecnt);
15654 +       }
15655 +       *nxp = nxi;
15656 +}
15657 +
15658 +
15659 +#define set_nx_info(p, i) __set_nx_info(p, i, __FILE__, __LINE__)
15660 +
15661 +static inline void __set_nx_info(struct nx_info **nxp, struct nx_info *nxi,
15662 +       const char *_file, int _line)
15663 +{
15664 +       struct nx_info *nxo;
15665 +
15666 +       if (!nxi)
15667 +               return;
15668 +
15669 +       vxlprintk(VXD_CBIT(nid, 3), "set_nx_info(%p[#%d.%d])",
15670 +               nxi, nxi ? nxi->nx_id : 0,
15671 +               nxi ? atomic_read(&nxi->nx_usecnt) : 0,
15672 +               _file, _line);
15673 +
15674 +       atomic_inc(&nxi->nx_usecnt);
15675 +       nxo = xchg(nxp, nxi);
15676 +       BUG_ON(nxo);
15677 +}
15678 +
15679 +#define clr_nx_info(p) __clr_nx_info(p, __FILE__, __LINE__)
15680 +
15681 +static inline void __clr_nx_info(struct nx_info **nxp,
15682 +       const char *_file, int _line)
15683 +{
15684 +       struct nx_info *nxo;
15685 +
15686 +       nxo = xchg(nxp, NULL);
15687 +       if (!nxo)
15688 +               return;
15689 +
15690 +       vxlprintk(VXD_CBIT(nid, 3), "clr_nx_info(%p[#%d.%d])",
15691 +               nxo, nxo ? nxo->nx_id : 0,
15692 +               nxo ? atomic_read(&nxo->nx_usecnt) : 0,
15693 +               _file, _line);
15694 +
15695 +       if (atomic_dec_and_test(&nxo->nx_usecnt))
15696 +               free_nx_info(nxo);
15697 +}
15698 +
15699 +
15700 +#define claim_nx_info(v, p) __claim_nx_info(v, p, __FILE__, __LINE__)
15701 +
15702 +static inline void __claim_nx_info(struct nx_info *nxi,
15703 +       struct task_struct *task, const char *_file, int _line)
15704 +{
15705 +       vxlprintk(VXD_CBIT(nid, 3), "claim_nx_info(%p[#%d.%d.%d]) %p",
15706 +               nxi, nxi ? nxi->nx_id : 0,
15707 +               nxi?atomic_read(&nxi->nx_usecnt):0,
15708 +               nxi?atomic_read(&nxi->nx_tasks):0,
15709 +               task, _file, _line);
15710 +
15711 +       atomic_inc(&nxi->nx_tasks);
15712 +}
15713 +
15714 +
15715 +extern void unhash_nx_info(struct nx_info *);
15716 +
15717 +#define release_nx_info(v, p) __release_nx_info(v, p, __FILE__, __LINE__)
15718 +
15719 +static inline void __release_nx_info(struct nx_info *nxi,
15720 +       struct task_struct *task, const char *_file, int _line)
15721 +{
15722 +       vxlprintk(VXD_CBIT(nid, 3), "release_nx_info(%p[#%d.%d.%d]) %p",
15723 +               nxi, nxi ? nxi->nx_id : 0,
15724 +               nxi ? atomic_read(&nxi->nx_usecnt) : 0,
15725 +               nxi ? atomic_read(&nxi->nx_tasks) : 0,
15726 +               task, _file, _line);
15727 +
15728 +       might_sleep();
15729 +
15730 +       if (atomic_dec_and_test(&nxi->nx_tasks))
15731 +               unhash_nx_info(nxi);
15732 +}
15733 +
15734 +
15735 +#define task_get_nx_info(i)    __task_get_nx_info(i, __FILE__, __LINE__)
15736 +
15737 +static __inline__ struct nx_info *__task_get_nx_info(struct task_struct *p,
15738 +       const char *_file, int _line)
15739 +{
15740 +       struct nx_info *nxi;
15741 +
15742 +       task_lock(p);
15743 +       vxlprintk(VXD_CBIT(nid, 5), "task_get_nx_info(%p)",
15744 +               p, _file, _line);
15745 +       nxi = __get_nx_info(p->nx_info, _file, _line);
15746 +       task_unlock(p);
15747 +       return nxi;
15748 +}
15749 +
15750 +
15751 +static inline void exit_nx_info(struct task_struct *p)
15752 +{
15753 +       if (p->nx_info)
15754 +               release_nx_info(p->nx_info, p);
15755 +}
15756 +
15757 +
15758 +#else
15759 +#warning duplicate inclusion
15760 +#endif
15761 diff -NurpP --minimal linux-2.6.25.11/include/linux/vs_pid.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_pid.h
15762 --- linux-2.6.25.11/include/linux/vs_pid.h      1969-12-31 19:00:00.000000000 -0500
15763 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_pid.h        2008-04-19 15:14:52.000000000 -0400
15764 @@ -0,0 +1,95 @@
15765 +#ifndef _VS_PID_H
15766 +#define _VS_PID_H
15767 +
15768 +#include "vserver/base.h"
15769 +#include "vserver/check.h"
15770 +#include "vserver/context.h"
15771 +#include "vserver/debug.h"
15772 +#include "vserver/pid.h"
15773 +#include <linux/pid_namespace.h>
15774 +
15775 +
15776 +#define VXF_FAKE_INIT  (VXF_INFO_INIT | VXF_STATE_INIT)
15777 +
15778 +static inline
15779 +int vx_proc_task_visible(struct task_struct *task)
15780 +{
15781 +       if ((task->pid == 1) &&
15782 +               !vx_flags(VXF_FAKE_INIT, VXF_FAKE_INIT))
15783 +               /* show a blend through init */
15784 +               goto visible;
15785 +       if (vx_check(vx_task_xid(task), VS_WATCH | VS_IDENT))
15786 +               goto visible;
15787 +       return 0;
15788 +visible:
15789 +       return 1;
15790 +}
15791 +
15792 +#define find_task_by_real_pid find_task_by_pid
15793 +
15794 +#if 0
15795 +
15796 +static inline
15797 +struct task_struct *vx_find_proc_task_by_pid(int pid)
15798 +{
15799 +       struct task_struct *task = find_task_by_real_pid(pid);
15800 +
15801 +       if (task && !vx_proc_task_visible(task)) {
15802 +               vxdprintk(VXD_CBIT(misc, 6),
15803 +                       "dropping task (find) %p[#%u,%u] for %p[#%u,%u]",
15804 +                       task, task->xid, task->pid,
15805 +                       current, current->xid, current->pid);
15806 +               task = NULL;
15807 +       }
15808 +       return task;
15809 +}
15810 +
15811 +#endif
15812 +
15813 +static inline
15814 +struct task_struct *vx_get_proc_task(struct inode *inode, struct pid *pid)
15815 +{
15816 +       struct task_struct *task = get_pid_task(pid, PIDTYPE_PID);
15817 +
15818 +       if (task && !vx_proc_task_visible(task)) {
15819 +               vxdprintk(VXD_CBIT(misc, 6),
15820 +                       "dropping task (get) %p[#%u,%u] for %p[#%u,%u]",
15821 +                       task, task->xid, task->pid,
15822 +                       current, current->xid, current->pid);
15823 +               put_task_struct(task);
15824 +               task = NULL;
15825 +       }
15826 +       return task;
15827 +}
15828 +
15829 +#if 0
15830 +
15831 +static inline
15832 +struct task_struct *vx_child_reaper(struct task_struct *p)
15833 +{
15834 +       struct vx_info *vxi = p->vx_info;
15835 +       struct task_struct *reaper = child_reaper(p);
15836 +
15837 +       if (!vxi)
15838 +               goto out;
15839 +
15840 +       BUG_ON(!p->vx_info->vx_reaper);
15841 +
15842 +       /* child reaper for the guest reaper */
15843 +       if (vxi->vx_reaper == p)
15844 +               goto out;
15845 +
15846 +       reaper = vxi->vx_reaper;
15847 +out:
15848 +       vxdprintk(VXD_CBIT(xid, 7),
15849 +               "vx_child_reaper(%p[#%u,%u]) = %p[#%u,%u]",
15850 +               p, p->xid, p->pid, reaper, reaper->xid, reaper->pid);
15851 +       return reaper;
15852 +}
15853 +
15854 +#endif
15855 +
15856 +
15857 +#else
15858 +#warning duplicate inclusion
15859 +#endif
15860 diff -NurpP --minimal linux-2.6.25.11/include/linux/vs_sched.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_sched.h
15861 --- linux-2.6.25.11/include/linux/vs_sched.h    1969-12-31 19:00:00.000000000 -0500
15862 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_sched.h      2008-04-19 15:14:52.000000000 -0400
15863 @@ -0,0 +1,110 @@
15864 +#ifndef _VS_SCHED_H
15865 +#define _VS_SCHED_H
15866 +
15867 +#include "vserver/base.h"
15868 +#include "vserver/context.h"
15869 +#include "vserver/sched.h"
15870 +
15871 +
15872 +#define VAVAVOOM_RATIO          50
15873 +
15874 +#define MAX_PRIO_BIAS           20
15875 +#define MIN_PRIO_BIAS          -20
15876 +
15877 +
15878 +#ifdef CONFIG_VSERVER_HARDCPU
15879 +
15880 +/*
15881 + * effective_prio - return the priority that is based on the static
15882 + * priority but is modified by bonuses/penalties.
15883 + *
15884 + * We scale the actual sleep average [0 .... MAX_SLEEP_AVG]
15885 + * into a -4 ... 0 ... +4 bonus/penalty range.
15886 + *
15887 + * Additionally, we scale another amount based on the number of
15888 + * CPU tokens currently held by the context, if the process is
15889 + * part of a context (and the appropriate SCHED flag is set).
15890 + * This ranges from -5 ... 0 ... +15, quadratically.
15891 + *
15892 + * So, the total bonus is -9 .. 0 .. +19
15893 + * We use ~50% of the full 0...39 priority range so that:
15894 + *
15895 + * 1) nice +19 interactive tasks do not preempt nice 0 CPU hogs.
15896 + * 2) nice -20 CPU hogs do not get preempted by nice 0 tasks.
15897 + *    unless that context is far exceeding its CPU allocation.
15898 + *
15899 + * Both properties are important to certain workloads.
15900 + */
15901 +static inline
15902 +int vx_effective_vavavoom(struct _vx_sched_pc *sched_pc, int max_prio)
15903 +{
15904 +       int vavavoom, max;
15905 +
15906 +       /* lots of tokens = lots of vavavoom
15907 +        *      no tokens = no vavavoom      */
15908 +       if ((vavavoom = sched_pc->tokens) >= 0) {
15909 +               max = sched_pc->tokens_max;
15910 +               vavavoom = max - vavavoom;
15911 +               max = max * max;
15912 +               vavavoom = max_prio * VAVAVOOM_RATIO / 100
15913 +                       * (vavavoom*vavavoom - (max >> 2)) / max;
15914 +               return vavavoom;
15915 +       }
15916 +       return 0;
15917 +}
15918 +
15919 +
15920 +static inline
15921 +int vx_adjust_prio(struct task_struct *p, int prio, int max_user)
15922 +{
15923 +       struct vx_info *vxi = p->vx_info;
15924 +       struct _vx_sched_pc *sched_pc;
15925 +
15926 +       if (!vxi)
15927 +               return prio;
15928 +
15929 +       sched_pc = &vx_cpu(vxi, sched_pc);
15930 +       if (vx_info_flags(vxi, VXF_SCHED_PRIO, 0)) {
15931 +               int vavavoom = vx_effective_vavavoom(sched_pc, max_user);
15932 +
15933 +               sched_pc->vavavoom = vavavoom;
15934 +               prio += vavavoom;
15935 +       }
15936 +       prio += sched_pc->prio_bias;
15937 +       return prio;
15938 +}
15939 +
15940 +#else /* !CONFIG_VSERVER_HARDCPU */
15941 +
15942 +static inline
15943 +int vx_adjust_prio(struct task_struct *p, int prio, int max_user)
15944 +{
15945 +       struct vx_info *vxi = p->vx_info;
15946 +
15947 +       if (vxi)
15948 +               prio += vx_cpu(vxi, sched_pc).prio_bias;
15949 +       return prio;
15950 +}
15951 +
15952 +#endif /* CONFIG_VSERVER_HARDCPU */
15953 +
15954 +
15955 +static inline void vx_account_user(struct vx_info *vxi,
15956 +       cputime_t cputime, int nice)
15957 +{
15958 +       if (!vxi)
15959 +               return;
15960 +       vx_cpu(vxi, sched_pc).user_ticks += cputime;
15961 +}
15962 +
15963 +static inline void vx_account_system(struct vx_info *vxi,
15964 +       cputime_t cputime, int idle)
15965 +{
15966 +       if (!vxi)
15967 +               return;
15968 +       vx_cpu(vxi, sched_pc).sys_ticks += cputime;
15969 +}
15970 +
15971 +#else
15972 +#warning duplicate inclusion
15973 +#endif
15974 diff -NurpP --minimal linux-2.6.25.11/include/linux/vs_socket.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_socket.h
15975 --- linux-2.6.25.11/include/linux/vs_socket.h   1969-12-31 19:00:00.000000000 -0500
15976 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_socket.h     2008-04-23 14:32:00.000000000 -0400
15977 @@ -0,0 +1,67 @@
15978 +#ifndef _VS_SOCKET_H
15979 +#define _VS_SOCKET_H
15980 +
15981 +#include "vserver/debug.h"
15982 +#include "vserver/base.h"
15983 +#include "vserver/cacct.h"
15984 +#include "vserver/context.h"
15985 +#include "vserver/tag.h"
15986 +
15987 +
15988 +/* socket accounting */
15989 +
15990 +#include <linux/socket.h>
15991 +
15992 +static inline int vx_sock_type(int family)
15993 +{
15994 +       switch (family) {
15995 +       case PF_UNSPEC:
15996 +               return VXA_SOCK_UNSPEC;
15997 +       case PF_UNIX:
15998 +               return VXA_SOCK_UNIX;
15999 +       case PF_INET:
16000 +               return VXA_SOCK_INET;
16001 +       case PF_INET6:
16002 +               return VXA_SOCK_INET6;
16003 +       case PF_PACKET:
16004 +               return VXA_SOCK_PACKET;
16005 +       default:
16006 +               return VXA_SOCK_OTHER;
16007 +       }
16008 +}
16009 +
16010 +#define vx_acc_sock(v, f, p, s) \
16011 +       __vx_acc_sock(v, f, p, s, __FILE__, __LINE__)
16012 +
16013 +static inline void __vx_acc_sock(struct vx_info *vxi,
16014 +       int family, int pos, int size, char *file, int line)
16015 +{
16016 +       if (vxi) {
16017 +               int type = vx_sock_type(family);
16018 +
16019 +               atomic_long_inc(&vxi->cacct.sock[type][pos].count);
16020 +               atomic_long_add(size, &vxi->cacct.sock[type][pos].total);
16021 +       }
16022 +}
16023 +
16024 +#define vx_sock_recv(sk, s) \
16025 +       vx_acc_sock((sk)->sk_vx_info, (sk)->sk_family, 0, s)
16026 +#define vx_sock_send(sk, s) \
16027 +       vx_acc_sock((sk)->sk_vx_info, (sk)->sk_family, 1, s)
16028 +#define vx_sock_fail(sk, s) \
16029 +       vx_acc_sock((sk)->sk_vx_info, (sk)->sk_family, 2, s)
16030 +
16031 +
16032 +#define sock_vx_init(s) do {           \
16033 +       (s)->sk_xid = 0;                \
16034 +       (s)->sk_vx_info = NULL;         \
16035 +       } while (0)
16036 +
16037 +#define sock_nx_init(s) do {           \
16038 +       (s)->sk_nid = 0;                \
16039 +       (s)->sk_nx_info = NULL;         \
16040 +       } while (0)
16041 +
16042 +#else
16043 +#warning duplicate inclusion
16044 +#endif
16045 diff -NurpP --minimal linux-2.6.25.11/include/linux/vs_tag.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_tag.h
16046 --- linux-2.6.25.11/include/linux/vs_tag.h      1969-12-31 19:00:00.000000000 -0500
16047 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_tag.h        2008-04-19 15:14:52.000000000 -0400
16048 @@ -0,0 +1,43 @@
16049 +#ifndef _VS_TAG_H
16050 +#define _VS_TAG_H
16051 +
16052 +#include <linux/vserver/tag.h>
16053 +
16054 +/* check conditions */
16055 +
16056 +#define DX_ADMIN       0x0001
16057 +#define DX_WATCH       0x0002
16058 +#define DX_HOSTID      0x0008
16059 +
16060 +#define DX_IDENT       0x0010
16061 +
16062 +#define DX_ARG_MASK    0x0010
16063 +
16064 +
16065 +#define dx_task_tag(t) ((t)->tag)
16066 +
16067 +#define dx_current_tag() dx_task_tag(current)
16068 +
16069 +#define dx_check(c, m) __dx_check(dx_current_tag(), c, m)
16070 +
16071 +#define dx_weak_check(c, m)    ((m) ? dx_check(c, m) : 1)
16072 +
16073 +
16074 +/*
16075 + * check current context for ADMIN/WATCH and
16076 + * optionally against supplied argument
16077 + */
16078 +static inline int __dx_check(tag_t cid, tag_t id, unsigned int mode)
16079 +{
16080 +       if (mode & DX_ARG_MASK) {
16081 +               if ((mode & DX_IDENT) && (id == cid))
16082 +                       return 1;
16083 +       }
16084 +       return (((mode & DX_ADMIN) && (cid == 0)) ||
16085 +               ((mode & DX_WATCH) && (cid == 1)) ||
16086 +               ((mode & DX_HOSTID) && (id == 0)));
16087 +}
16088 +
16089 +#else
16090 +#warning duplicate inclusion
16091 +#endif
16092 diff -NurpP --minimal linux-2.6.25.11/include/linux/vs_time.h linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_time.h
16093 --- linux-2.6.25.11/include/linux/vs_time.h     1969-12-31 19:00:00.000000000 -0500
16094 +++ linux-2.6.25.11-vs2.3.0.34.14/include/linux/vs_time.h       2008-04-19 15:14:52.000000000 -0400
16095 @@ -0,0 +1,19 @@
16096 +#ifndef _VS_TIME_H
16097 +#define _VS_TIME_H
16098 +
16099 +
16100 +/* time faking stuff */
16101 +
16102 +#ifdef CONFIG_VSERVER_VTIME
16103 +
16104 +extern void vx_gettimeofday(struct timeval *tv);
16105 +extern int vx_settimeofday(struct timespec *ts);
16106 +
16107 +#else
16108 +#define        vx_gettimeofday(t)      do_gettimeofday(t)
16109 +#define        vx_settimeofday(t)      do_settimeofday(t)
16110 +#endif
16111 +
16112 +#else
16113 +#warning duplicate inclusion
16114 +#endif
16115 diff -NurpP --minimal linux-2.6.25.11/include/net/addrconf.h linux-2.6.25.11-vs2.3.0.34.14/include/net/addrconf.h
16116 --- linux-2.6.25.11/include/net/addrconf.h      2008-04-17 12:05:44.000000000 -0400
16117 +++ linux-2.6.25.11-vs2.3.0.34.14/include/net/addrconf.h        2008-04-19 15:14:52.000000000 -0400
16118 @@ -75,10 +75,12 @@ extern struct inet6_ifaddr      *ipv6_ge
16119
16120  extern int                     ipv6_get_saddr(struct dst_entry *dst,
16121                                                struct in6_addr *daddr,
16122 -                                              struct in6_addr *saddr);
16123 +                                              struct in6_addr *saddr,
16124 +                                              struct nx_info *nxi);
16125  extern int                     ipv6_dev_get_saddr(struct net_device *dev,
16126                                                struct in6_addr *daddr,
16127 -                                              struct in6_addr *saddr);
16128 +                                              struct in6_addr *saddr,
16129 +                                              struct nx_info *nxi);
16130  extern int                     ipv6_get_lladdr(struct net_device *dev,
16131                                                 struct in6_addr *addr,
16132                                                 unsigned char banned_flags);
16133 diff -NurpP --minimal linux-2.6.25.11/include/net/af_unix.h linux-2.6.25.11-vs2.3.0.34.14/include/net/af_unix.h
16134 --- linux-2.6.25.11/include/net/af_unix.h       2008-04-17 12:05:44.000000000 -0400
16135 +++ linux-2.6.25.11-vs2.3.0.34.14/include/net/af_unix.h 2008-04-19 15:14:52.000000000 -0400
16136 @@ -4,6 +4,7 @@
16137  #include <linux/socket.h>
16138  #include <linux/un.h>
16139  #include <linux/mutex.h>
16140 +#include <linux/vs_base.h>
16141  #include <net/sock.h>
16142
16143  extern void unix_inflight(struct file *fp);
16144 diff -NurpP --minimal linux-2.6.25.11/include/net/inet_sock.h linux-2.6.25.11-vs2.3.0.34.14/include/net/inet_sock.h
16145 --- linux-2.6.25.11/include/net/inet_sock.h     2008-04-17 12:05:44.000000000 -0400
16146 +++ linux-2.6.25.11-vs2.3.0.34.14/include/net/inet_sock.h       2008-04-19 15:14:52.000000000 -0400
16147 @@ -24,7 +24,7 @@
16148  #include <net/flow.h>
16149  #include <net/sock.h>
16150  #include <net/request_sock.h>
16151 -#include <net/route.h>
16152 +// #include <net/route.h>
16153
16154  /** struct ip_options - IP Options
16155   *
16156 @@ -193,9 +193,4 @@ static inline int inet_sk_ehashfn(const
16157  }
16158
16159
16160 -static inline int inet_iif(const struct sk_buff *skb)
16161 -{
16162 -       return ((struct rtable *)skb->dst)->rt_iif;
16163 -}
16164 -
16165  #endif /* _INET_SOCK_H */
16166 diff -NurpP --minimal linux-2.6.25.11/include/net/inet_timewait_sock.h linux-2.6.25.11-vs2.3.0.34.14/include/net/inet_timewait_sock.h
16167 --- linux-2.6.25.11/include/net/inet_timewait_sock.h    2008-04-17 12:05:44.000000000 -0400
16168 +++ linux-2.6.25.11-vs2.3.0.34.14/include/net/inet_timewait_sock.h      2008-04-19 15:25:34.000000000 -0400
16169 @@ -15,15 +15,14 @@
16170  #ifndef _INET_TIMEWAIT_SOCK_
16171  #define _INET_TIMEWAIT_SOCK_
16172
16173 +// #include <net/inet_sock.h>
16174 +#include <net/sock.h>
16175
16176  #include <linux/list.h>
16177  #include <linux/module.h>
16178  #include <linux/timer.h>
16179  #include <linux/types.h>
16180  #include <linux/workqueue.h>
16181 -
16182 -#include <net/inet_sock.h>
16183 -#include <net/sock.h>
16184  #include <net/tcp_states.h>
16185  #include <net/timewait_sock.h>
16186
16187 @@ -116,6 +115,10 @@ struct inet_timewait_sock {
16188  #define tw_hash                        __tw_common.skc_hash
16189  #define tw_prot                        __tw_common.skc_prot
16190  #define tw_net                 __tw_common.skc_net
16191 +#define tw_xid                 __tw_common.skc_xid
16192 +#define tw_vx_info             __tw_common.skc_vx_info
16193 +#define tw_nid                 __tw_common.skc_nid
16194 +#define tw_nx_info             __tw_common.skc_nx_info
16195         int                     tw_timeout;
16196         volatile unsigned char  tw_substate;
16197         /* 3 bits hole, try to pack */
16198 diff -NurpP --minimal linux-2.6.25.11/include/net/route.h linux-2.6.25.11-vs2.3.0.34.14/include/net/route.h
16199 --- linux-2.6.25.11/include/net/route.h 2008-04-17 12:05:44.000000000 -0400
16200 +++ linux-2.6.25.11-vs2.3.0.34.14/include/net/route.h   2008-04-21 12:39:35.000000000 -0400
16201 @@ -34,7 +34,7 @@
16202  #include <linux/ip.h>
16203  #include <linux/cache.h>
16204  #include <linux/security.h>
16205 -#include <net/sock.h>
16206 +#include <linux/in.h>
16207
16208  #ifndef __KERNEL__
16209  #warning This file is not supposed to be used outside of kernel.
16210 @@ -86,6 +86,11 @@ struct ip_rt_acct
16211         __u32   i_packets;
16212  };
16213
16214 +static inline int inet_iif(const struct sk_buff *skb)
16215 +{
16216 +       return ((struct rtable *)skb->dst)->rt_iif;
16217 +}
16218 +
16219  struct rt_cache_stat
16220  {
16221          unsigned int in_hit;
16222 @@ -136,6 +141,9 @@ static inline void ip_rt_put(struct rtab
16223                 dst_release(&rt->u.dst);
16224  }
16225
16226 +#include <linux/vs_base.h>
16227 +#include <linux/vs_inet.h>
16228 +
16229  #define IPTOS_RT_MASK  (IPTOS_TOS_MASK & ~3)
16230
16231  extern const __u8 ip_tos2prio[16];
16232 @@ -145,6 +153,9 @@ static inline char rt_tos2priority(u8 to
16233         return ip_tos2prio[IPTOS_TOS(tos)>>1];
16234  }
16235
16236 +extern int ip_v4_find_src(struct net *net, struct nx_info *,
16237 +       struct rtable **, struct flowi *);
16238 +
16239  static inline int ip_route_connect(struct rtable **rp, __be32 dst,
16240                                    __be32 src, u32 tos, int oif, u8 protocol,
16241                                    __be16 sport, __be16 dport, struct sock *sk,
16242 @@ -162,7 +173,21 @@ static inline int ip_route_connect(struc
16243
16244         int err;
16245         struct net *net = sk->sk_net;
16246 -       if (!dst || !src) {
16247 +       struct nx_info *nx_info = current->nx_info;
16248 +
16249 +       if (sk)
16250 +               nx_info = sk->sk_nx_info;
16251 +
16252 +       vxdprintk(VXD_CBIT(net, 4),
16253 +               "ip_route_connect(%p) %p,%p;%lx",
16254 +               sk, nx_info, sk->sk_socket,
16255 +               (sk->sk_socket?sk->sk_socket->flags:0));
16256 +
16257 +       err = ip_v4_find_src(net, nx_info, rp, &fl);
16258 +       if (err)
16259 +               return err;
16260 +
16261 +       if (!fl.fl4_dst || !fl.fl4_src) {
16262                 err = __ip_route_output_key(net, rp, &fl);
16263                 if (err)
16264                         return err;
16265 diff -NurpP --minimal linux-2.6.25.11/include/net/sock.h linux-2.6.25.11-vs2.3.0.34.14/include/net/sock.h
16266 --- linux-2.6.25.11/include/net/sock.h  2008-04-17 12:05:44.000000000 -0400
16267 +++ linux-2.6.25.11-vs2.3.0.34.14/include/net/sock.h    2008-04-19 15:14:52.000000000 -0400
16268 @@ -123,6 +123,10 @@ struct sock_common {
16269         unsigned int            skc_hash;
16270         struct proto            *skc_prot;
16271         struct net              *skc_net;
16272 +       xid_t                   skc_xid;
16273 +       struct vx_info          *skc_vx_info;
16274 +       nid_t                   skc_nid;
16275 +       struct nx_info          *skc_nx_info;
16276  };
16277
16278  /**
16279 @@ -205,6 +209,10 @@ struct sock {
16280  #define sk_hash                        __sk_common.skc_hash
16281  #define sk_prot                        __sk_common.skc_prot
16282  #define sk_net                 __sk_common.skc_net
16283 +#define sk_xid                 __sk_common.skc_xid
16284 +#define sk_vx_info             __sk_common.skc_vx_info
16285 +#define sk_nid                 __sk_common.skc_nid
16286 +#define sk_nx_info             __sk_common.skc_nx_info
16287         unsigned char           sk_shutdown : 2,
16288                                 sk_no_check : 2,
16289                                 sk_userlocks : 4;
16290 diff -NurpP --minimal linux-2.6.25.11/init/main.c linux-2.6.25.11-vs2.3.0.34.14/init/main.c
16291 --- linux-2.6.25.11/init/main.c 2008-04-17 12:05:44.000000000 -0400
16292 +++ linux-2.6.25.11-vs2.3.0.34.14/init/main.c   2008-04-21 10:46:10.000000000 -0400
16293 @@ -58,6 +58,7 @@
16294  #include <linux/kthread.h>
16295  #include <linux/sched.h>
16296  #include <linux/signal.h>
16297 +#include <linux/vserver/percpu.h>
16298
16299  #include <asm/io.h>
16300  #include <asm/bugs.h>
16301 @@ -370,12 +371,14 @@ EXPORT_SYMBOL(__per_cpu_offset);
16302
16303  static void __init setup_per_cpu_areas(void)
16304  {
16305 -       unsigned long size, i;
16306 +       unsigned long size, vspc, i;
16307         char *ptr;
16308         unsigned long nr_possible_cpus = num_possible_cpus();
16309
16310 +       vspc = PERCPU_PERCTX * CONFIG_VSERVER_CONTEXTS;
16311 +
16312         /* Copy section for each CPU (we discard the original) */
16313 -       size = ALIGN(PERCPU_ENOUGH_ROOM, PAGE_SIZE);
16314 +       size = ALIGN(PERCPU_ENOUGH_ROOM + vspc, PAGE_SIZE);
16315         ptr = alloc_bootmem_pages(size * nr_possible_cpus);
16316
16317         for_each_possible_cpu(i) {
16318 diff -NurpP --minimal linux-2.6.25.11/ipc/mqueue.c linux-2.6.25.11-vs2.3.0.34.14/ipc/mqueue.c
16319 --- linux-2.6.25.11/ipc/mqueue.c        2008-04-17 12:05:44.000000000 -0400
16320 +++ linux-2.6.25.11-vs2.3.0.34.14/ipc/mqueue.c  2008-04-19 15:14:52.000000000 -0400
16321 @@ -31,6 +31,8 @@
16322  #include <linux/mutex.h>
16323  #include <linux/nsproxy.h>
16324  #include <linux/pid.h>
16325 +#include <linux/vs_context.h>
16326 +#include <linux/vs_limit.h>
16327
16328  #include <net/sock.h>
16329  #include "util.h"
16330 @@ -71,6 +73,7 @@ struct mqueue_inode_info {
16331         struct sigevent notify;
16332         struct pid* notify_owner;
16333         struct user_struct *user;       /* user who created, for accounting */
16334 +       struct vx_info *vxi;
16335         struct sock *notify_sock;
16336         struct sk_buff *notify_cookie;
16337
16338 @@ -119,6 +122,7 @@ static struct inode *mqueue_get_inode(st
16339                         struct mqueue_inode_info *info;
16340                         struct task_struct *p = current;
16341                         struct user_struct *u = p->user;
16342 +                       struct vx_info *vxi = p->vx_info;
16343                         unsigned long mq_bytes, mq_msg_tblsz;
16344
16345                         inode->i_fop = &mqueue_file_operations;
16346 @@ -133,6 +137,7 @@ static struct inode *mqueue_get_inode(st
16347                         info->notify_owner = NULL;
16348                         info->qsize = 0;
16349                         info->user = NULL;      /* set when all is ok */
16350 +                       info->vxi = NULL;
16351                         memset(&info->attr, 0, sizeof(info->attr));
16352                         info->attr.mq_maxmsg = DFLT_MSGMAX;
16353                         info->attr.mq_msgsize = DFLT_MSGSIZEMAX;
16354 @@ -147,22 +152,26 @@ static struct inode *mqueue_get_inode(st
16355                         spin_lock(&mq_lock);
16356                         if (u->mq_bytes + mq_bytes < u->mq_bytes ||
16357                             u->mq_bytes + mq_bytes >
16358 -                           p->signal->rlim[RLIMIT_MSGQUEUE].rlim_cur) {
16359 +                           p->signal->rlim[RLIMIT_MSGQUEUE].rlim_cur ||
16360 +                           !vx_ipcmsg_avail(vxi, mq_bytes)) {
16361                                 spin_unlock(&mq_lock);
16362                                 goto out_inode;
16363                         }
16364                         u->mq_bytes += mq_bytes;
16365 +                       vx_ipcmsg_add(vxi, u, mq_bytes);
16366                         spin_unlock(&mq_lock);
16367
16368                         info->messages = kmalloc(mq_msg_tblsz, GFP_KERNEL);
16369                         if (!info->messages) {
16370                                 spin_lock(&mq_lock);
16371                                 u->mq_bytes -= mq_bytes;
16372 +                               vx_ipcmsg_sub(vxi, u, mq_bytes);
16373                                 spin_unlock(&mq_lock);
16374                                 goto out_inode;
16375                         }
16376                         /* all is ok */
16377                         info->user = get_uid(u);
16378 +                       info->vxi = get_vx_info(vxi);
16379                 } else if (S_ISDIR(mode)) {
16380                         inc_nlink(inode);
16381                         /* Some things misbehave if size == 0 on a directory */
16382 @@ -253,10 +262,14 @@ static void mqueue_delete_inode(struct i
16383                    (info->attr.mq_maxmsg * info->attr.mq_msgsize));
16384         user = info->user;
16385         if (user) {
16386 +               struct vx_info *vxi = info->vxi;
16387 +
16388                 spin_lock(&mq_lock);
16389                 user->mq_bytes -= mq_bytes;
16390 +               vx_ipcmsg_sub(vxi, user, mq_bytes);
16391                 queues_count--;
16392                 spin_unlock(&mq_lock);
16393 +               put_vx_info(vxi);
16394                 free_uid(user);
16395         }
16396  }
16397 @@ -743,7 +756,7 @@ asmlinkage long sys_mq_unlink(const char
16398         if (inode)
16399                 atomic_inc(&inode->i_count);
16400
16401 -       err = vfs_unlink(dentry->d_parent->d_inode, dentry);
16402 +       err = vfs_unlink(dentry->d_parent->d_inode, dentry, NULL);
16403  out_err:
16404         dput(dentry);
16405
16406 diff -NurpP --minimal linux-2.6.25.11/ipc/msg.c linux-2.6.25.11-vs2.3.0.34.14/ipc/msg.c
16407 --- linux-2.6.25.11/ipc/msg.c   2008-04-17 12:05:44.000000000 -0400
16408 +++ linux-2.6.25.11-vs2.3.0.34.14/ipc/msg.c     2008-04-21 10:41:47.000000000 -0400
16409 @@ -37,6 +37,7 @@
16410  #include <linux/rwsem.h>
16411  #include <linux/nsproxy.h>
16412  #include <linux/ipc_namespace.h>
16413 +#include <linux/vs_base.h>
16414
16415  #include <asm/current.h>
16416  #include <asm/uaccess.h>
16417 @@ -168,6 +169,7 @@ static int newque(struct ipc_namespace *
16418
16419         msq->q_perm.mode = msgflg & S_IRWXUGO;
16420         msq->q_perm.key = key;
16421 +       msq->q_perm.xid = vx_current_xid();
16422
16423         msq->q_perm.security = NULL;
16424         retval = security_msg_queue_alloc(msq);
16425 diff -NurpP --minimal linux-2.6.25.11/ipc/namespace.c linux-2.6.25.11-vs2.3.0.34.14/ipc/namespace.c
16426 --- linux-2.6.25.11/ipc/namespace.c     2008-04-17 12:05:44.000000000 -0400
16427 +++ linux-2.6.25.11-vs2.3.0.34.14/ipc/namespace.c       2008-04-21 10:44:58.000000000 -0400
16428 @@ -9,6 +9,8 @@
16429  #include <linux/rcupdate.h>
16430  #include <linux/nsproxy.h>
16431  #include <linux/slab.h>
16432 +#include <linux/vs_base.h>
16433 +#include <linux/vserver/global.h>
16434
16435  #include "util.h"
16436
16437 @@ -25,6 +27,7 @@ static struct ipc_namespace *clone_ipc_n
16438         shm_init_ns(ns);
16439
16440         kref_init(&ns->kref);
16441 +       atomic_inc(&vs_global_ipc_ns);
16442         return ns;
16443  }
16444
16445 @@ -82,5 +85,6 @@ void free_ipc_ns(struct kref *kref)
16446         sem_exit_ns(ns);
16447         msg_exit_ns(ns);
16448         shm_exit_ns(ns);
16449 +       atomic_dec(&vs_global_ipc_ns);
16450         kfree(ns);
16451  }
16452 diff -NurpP --minimal linux-2.6.25.11/ipc/sem.c linux-2.6.25.11-vs2.3.0.34.14/ipc/sem.c
16453 --- linux-2.6.25.11/ipc/sem.c   2008-04-17 12:05:44.000000000 -0400
16454 +++ linux-2.6.25.11-vs2.3.0.34.14/ipc/sem.c     2008-04-21 10:45:22.000000000 -0400
16455 @@ -83,6 +83,8 @@
16456  #include <linux/rwsem.h>
16457  #include <linux/nsproxy.h>
16458  #include <linux/ipc_namespace.h>
16459 +#include <linux/vs_base.h>
16460 +#include <linux/vs_limit.h>
16461
16462  #include <asm/uaccess.h>
16463  #include "util.h"
16464 @@ -252,6 +254,7 @@ static int newary(struct ipc_namespace *
16465
16466         sma->sem_perm.mode = (semflg & S_IRWXUGO);
16467         sma->sem_perm.key = key;
16468 +       sma->sem_perm.xid = vx_current_xid();
16469
16470         sma->sem_perm.security = NULL;
16471         retval = security_sem_alloc(sma);
16472 @@ -267,6 +270,9 @@ static int newary(struct ipc_namespace *
16473                 return id;
16474         }
16475         ns->used_sems += nsems;
16476 +       /* FIXME: obsoleted? */
16477 +       vx_semary_inc(sma);
16478 +       vx_nsems_add(sma, nsems);
16479
16480         sma->sem_perm.id = sem_buildid(id, sma->sem_perm.seq);
16481         sma->sem_base = (struct sem *) &sma[1];
16482 diff -NurpP --minimal linux-2.6.25.11/ipc/shm.c linux-2.6.25.11-vs2.3.0.34.14/ipc/shm.c
16483 --- linux-2.6.25.11/ipc/shm.c   2008-04-17 12:05:44.000000000 -0400
16484 +++ linux-2.6.25.11-vs2.3.0.34.14/ipc/shm.c     2008-04-21 10:45:38.000000000 -0400
16485 @@ -39,6 +39,8 @@
16486  #include <linux/nsproxy.h>
16487  #include <linux/mount.h>
16488  #include <linux/ipc_namespace.h>
16489 +#include <linux/vs_context.h>
16490 +#include <linux/vs_limit.h>
16491
16492  #include <asm/uaccess.h>
16493
16494 @@ -202,7 +204,12 @@ static void shm_open(struct vm_area_stru
16495   */
16496  static void shm_destroy(struct ipc_namespace *ns, struct shmid_kernel *shp)
16497  {
16498 -       ns->shm_tot -= (shp->shm_segsz + PAGE_SIZE - 1) >> PAGE_SHIFT;
16499 +       struct vx_info *vxi = lookup_vx_info(shp->shm_perm.xid);
16500 +       int numpages = (shp->shm_segsz + PAGE_SIZE - 1) >> PAGE_SHIFT;
16501 +
16502 +       vx_ipcshm_sub(vxi, shp, numpages);
16503 +       ns->shm_tot -= numpages;
16504 +
16505         shm_rmid(ns, shp);
16506         shm_unlock(shp);
16507         if (!is_file_hugepages(shp->shm_file))
16508 @@ -212,6 +219,7 @@ static void shm_destroy(struct ipc_names
16509                                                 shp->mlock_user);
16510         fput (shp->shm_file);
16511         security_shm_free(shp);
16512 +       put_vx_info(vxi);
16513         ipc_rcu_putref(shp);
16514  }
16515
16516 @@ -383,11 +391,15 @@ static int newseg(struct ipc_namespace *
16517         if (ns->shm_tot + numpages > ns->shm_ctlall)
16518                 return -ENOSPC;
16519
16520 +       if (!vx_ipcshm_avail(current->vx_info, numpages))
16521 +               return -ENOSPC;
16522 +
16523         shp = ipc_rcu_alloc(sizeof(*shp));
16524         if (!shp)
16525                 return -ENOMEM;
16526
16527         shp->shm_perm.key = key;
16528 +       shp->shm_perm.xid = vx_current_xid();
16529         shp->shm_perm.mode = (shmflg & S_IRWXUGO);
16530         shp->mlock_user = NULL;
16531
16532 @@ -441,6 +453,7 @@ static int newseg(struct ipc_namespace *
16533         ns->shm_tot += numpages;
16534         error = shp->shm_perm.id;
16535         shm_unlock(shp);
16536 +       vx_ipcshm_add(current->vx_info, key, numpages);
16537         return error;
16538
16539  no_id:
16540 diff -NurpP --minimal linux-2.6.25.11/kernel/capability.c linux-2.6.25.11-vs2.3.0.34.14/kernel/capability.c
16541 --- linux-2.6.25.11/kernel/capability.c 2008-07-15 22:05:17.000000000 -0400
16542 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/capability.c   2008-06-10 20:00:46.000000000 -0400
16543 @@ -13,6 +13,7 @@
16544  #include <linux/security.h>
16545  #include <linux/syscalls.h>
16546  #include <linux/pid_namespace.h>
16547 +#include <linux/vs_context.h>
16548  #include <asm/uaccess.h>
16549
16550  /*
16551 @@ -220,6 +221,8 @@ static inline int cap_set_pg(int pgrp_nr
16552
16553         pgrp = find_vpid(pgrp_nr);
16554         do_each_pid_task(pgrp, PIDTYPE_PGID, g) {
16555 +               if (!vx_check(g->xid, VS_ADMIN_P | VS_IDENT))
16556 +                       continue;
16557                 target = g;
16558                 while_each_thread(g, target) {
16559                         if (!security_capset_check(target, effective,
16560 @@ -370,8 +373,12 @@ int __capable(struct task_struct *t, int
16561         return 0;
16562  }
16563
16564 +#include <linux/vserver/base.h>
16565  int capable(int cap)
16566  {
16567 +       /* here for now so we don't require task locking */
16568 +       if (vs_check_bit(VXC_CAP_MASK, cap) && !vx_mcaps(1L << cap))
16569 +               return 0;
16570         return __capable(current, cap);
16571  }
16572  EXPORT_SYMBOL(capable);
16573 diff -NurpP --minimal linux-2.6.25.11/kernel/cgroup.c linux-2.6.25.11-vs2.3.0.34.14/kernel/cgroup.c
16574 --- linux-2.6.25.11/kernel/cgroup.c     2008-07-15 22:05:17.000000000 -0400
16575 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/cgroup.c       2008-06-10 20:00:46.000000000 -0400
16576 @@ -2833,7 +2833,7 @@ int cgroup_clone(struct task_struct *tsk
16577         }
16578
16579         /* Create the cgroup directory, which also creates the cgroup */
16580 -       ret = vfs_mkdir(inode, dentry, S_IFDIR | 0755);
16581 +       ret = vfs_mkdir(inode, dentry, S_IFDIR | 0755, NULL);
16582         child = __d_cgrp(dentry);
16583         dput(dentry);
16584         if (ret) {
16585 diff -NurpP --minimal linux-2.6.25.11/kernel/compat.c linux-2.6.25.11-vs2.3.0.34.14/kernel/compat.c
16586 --- linux-2.6.25.11/kernel/compat.c     2008-04-17 12:05:44.000000000 -0400
16587 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/compat.c       2008-04-19 15:14:52.000000000 -0400
16588 @@ -846,7 +846,7 @@ asmlinkage long compat_sys_time(compat_t
16589         compat_time_t i;
16590         struct timeval tv;
16591
16592 -       do_gettimeofday(&tv);
16593 +       vx_gettimeofday(&tv);
16594         i = tv.tv_sec;
16595
16596         if (tloc) {
16597 @@ -870,7 +870,7 @@ asmlinkage long compat_sys_stime(compat_
16598         if (err)
16599                 return err;
16600
16601 -       do_settimeofday(&tv);
16602 +       vx_settimeofday(&tv);
16603         return 0;
16604  }
16605
16606 diff -NurpP --minimal linux-2.6.25.11/kernel/exit.c linux-2.6.25.11-vs2.3.0.34.14/kernel/exit.c
16607 --- linux-2.6.25.11/kernel/exit.c       2008-04-17 12:05:44.000000000 -0400
16608 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/exit.c 2008-04-19 15:14:52.000000000 -0400
16609 @@ -44,6 +44,11 @@
16610  #include <linux/resource.h>
16611  #include <linux/blkdev.h>
16612  #include <linux/task_io_accounting_ops.h>
16613 +#include <linux/vs_limit.h>
16614 +#include <linux/vs_context.h>
16615 +#include <linux/vs_network.h>
16616 +#include <linux/vs_pid.h>
16617 +#include <linux/vserver/global.h>
16618
16619  #include <asm/uaccess.h>
16620  #include <asm/unistd.h>
16621 @@ -468,9 +473,11 @@ static void close_files(struct files_str
16622                                         filp_close(file, files);
16623                                         cond_resched();
16624                                 }
16625 +                               vx_openfd_dec(i);
16626                         }
16627                         i++;
16628                         set >>= 1;
16629 +                       cond_resched();
16630                 }
16631         }
16632  }
16633 @@ -1014,6 +1021,10 @@ NORET_TYPE void do_exit(long code)
16634         if (tsk->splice_pipe)
16635                 __free_pipe_info(tsk->splice_pipe);
16636
16637 +       /* needs to stay after exit_notify() */
16638 +       exit_vx_info(tsk, code);
16639 +       exit_nx_info(tsk);
16640 +
16641         preempt_disable();
16642         /* causes final put_task_struct in finish_task_switch(). */
16643         tsk->state = TASK_DEAD;
16644 diff -NurpP --minimal linux-2.6.25.11/kernel/fork.c linux-2.6.25.11-vs2.3.0.34.14/kernel/fork.c
16645 --- linux-2.6.25.11/kernel/fork.c       2008-04-17 12:05:44.000000000 -0400
16646 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/fork.c 2008-04-21 11:09:01.000000000 -0400
16647 @@ -53,6 +53,11 @@
16648  #include <linux/tty.h>
16649  #include <linux/proc_fs.h>
16650  #include <linux/blkdev.h>
16651 +#include <linux/vs_context.h>
16652 +#include <linux/vs_network.h>
16653 +#include <linux/vs_limit.h>
16654 +#include <linux/vs_memory.h>
16655 +#include <linux/vserver/global.h>
16656
16657  #include <asm/pgtable.h>
16658  #include <asm/pgalloc.h>
16659 @@ -113,6 +118,8 @@ void free_task(struct task_struct *tsk)
16660         prop_local_destroy_single(&tsk->dirties);
16661         free_thread_info(tsk->stack);
16662         rt_mutex_debug_task_free(tsk);
16663 +       clr_vx_info(&tsk->vx_info);
16664 +       clr_nx_info(&tsk->nx_info);
16665         free_task_struct(tsk);
16666  }
16667  EXPORT_SYMBOL(free_task);
16668 @@ -229,6 +236,8 @@ static int dup_mmap(struct mm_struct *mm
16669         mm->free_area_cache = oldmm->mmap_base;
16670         mm->cached_hole_size = ~0UL;
16671         mm->map_count = 0;
16672 +       __set_mm_counter(mm, file_rss, 0);
16673 +       __set_mm_counter(mm, anon_rss, 0);
16674         cpus_clear(mm->cpu_vm_mask);
16675         mm->mm_rb = RB_ROOT;
16676         rb_link = &mm->mm_rb.rb_node;
16677 @@ -240,7 +249,7 @@ static int dup_mmap(struct mm_struct *mm
16678
16679                 if (mpnt->vm_flags & VM_DONTCOPY) {
16680                         long pages = vma_pages(mpnt);
16681 -                       mm->total_vm -= pages;
16682 +                       vx_vmpages_sub(mm, pages);
16683                         vm_stat_account(mm, mpnt->vm_flags, mpnt->vm_file,
16684                                                                 -pages);
16685                         continue;
16686 @@ -351,8 +360,8 @@ static struct mm_struct * mm_init(struct
16687                                   : MMF_DUMP_FILTER_DEFAULT;
16688         mm->core_waiters = 0;
16689         mm->nr_ptes = 0;
16690 -       set_mm_counter(mm, file_rss, 0);
16691 -       set_mm_counter(mm, anon_rss, 0);
16692 +       __set_mm_counter(mm, file_rss, 0);
16693 +       __set_mm_counter(mm, anon_rss, 0);
16694         spin_lock_init(&mm->page_table_lock);
16695         rwlock_init(&mm->ioctx_list_lock);
16696         mm->ioctx_list = NULL;
16697 @@ -362,6 +371,7 @@ static struct mm_struct * mm_init(struct
16698
16699         if (likely(!mm_alloc_pgd(mm))) {
16700                 mm->def_flags = 0;
16701 +               set_vx_info(&mm->mm_vx_info, p->vx_info);
16702                 return mm;
16703         }
16704
16705 @@ -395,6 +405,7 @@ void __mmdrop(struct mm_struct *mm)
16706         BUG_ON(mm == &init_mm);
16707         mm_free_pgd(mm);
16708         destroy_context(mm);
16709 +       clr_vx_info(&mm->mm_vx_info);
16710         free_mm(mm);
16711  }
16712  EXPORT_SYMBOL_GPL(__mmdrop);
16713 @@ -511,6 +522,7 @@ static struct mm_struct *dup_mm(struct t
16714                 goto fail_nomem;
16715
16716         memcpy(mm, oldmm, sizeof(*mm));
16717 +       mm->mm_vx_info = NULL;
16718
16719         /* Initializing for Swap token stuff */
16720         mm->token_priority = 0;
16721 @@ -542,6 +554,7 @@ fail_nocontext:
16722          * If init_new_context() failed, we cannot use mmput() to free the mm
16723          * because it calls destroy_context()
16724          */
16725 +       clr_vx_info(&mm->mm_vx_info);
16726         mm_free_pgd(mm);
16727         free_mm(mm);
16728         return NULL;
16729 @@ -612,6 +625,7 @@ static struct fs_struct *__copy_fs_struc
16730                         fs->altroot.dentry = NULL;
16731                 }
16732                 read_unlock(&old->lock);
16733 +               atomic_inc(&vs_global_fs);
16734         }
16735         return fs;
16736  }
16737 @@ -730,6 +744,8 @@ static struct files_struct *dup_fd(struc
16738                 struct file *f = *old_fds++;
16739                 if (f) {
16740                         get_file(f);
16741 +                       /* TODO: sum it first for check and performance */
16742 +                       vx_openfd_inc(open_files - i);
16743                 } else {
16744                         /*
16745                          * The fd may be claimed in the fd bitmap but not yet
16746 @@ -1011,6 +1027,8 @@ static struct task_struct *copy_process(
16747         int retval;
16748         struct task_struct *p;
16749         int cgroup_callbacks_done = 0;
16750 +       struct vx_info *vxi;
16751 +       struct nx_info *nxi;
16752
16753         if ((clone_flags & (CLONE_NEWNS|CLONE_FS)) == (CLONE_NEWNS|CLONE_FS))
16754                 return ERR_PTR(-EINVAL);
16755 @@ -1045,12 +1063,28 @@ static struct task_struct *copy_process(
16756         DEBUG_LOCKS_WARN_ON(!p->hardirqs_enabled);
16757         DEBUG_LOCKS_WARN_ON(!p->softirqs_enabled);
16758  #endif
16759 +       init_vx_info(&p->vx_info, current->vx_info);
16760 +       init_nx_info(&p->nx_info, current->nx_info);
16761 +
16762 +       /* check vserver memory */
16763 +       if (p->mm && !(clone_flags & CLONE_VM)) {
16764 +               if (vx_vmpages_avail(p->mm, p->mm->total_vm))
16765 +                       vx_pages_add(p->vx_info, RLIMIT_AS, p->mm->total_vm);
16766 +               else
16767 +                       goto bad_fork_free;
16768 +       }
16769 +       if (p->mm && vx_flags(VXF_FORK_RSS, 0)) {
16770 +               if (!vx_rss_avail(p->mm, get_mm_counter(p->mm, file_rss)))
16771 +                       goto bad_fork_cleanup_vm;
16772 +       }
16773         retval = -EAGAIN;
16774 +       if (!vx_nproc_avail(1))
16775 +               goto bad_fork_cleanup_vm;
16776         if (atomic_read(&p->user->processes) >=
16777                         p->signal->rlim[RLIMIT_NPROC].rlim_cur) {
16778                 if (!capable(CAP_SYS_ADMIN) && !capable(CAP_SYS_RESOURCE) &&
16779                     p->user != current->nsproxy->user_ns->root_user)
16780 -                       goto bad_fork_free;
16781 +                       goto bad_fork_cleanup_vm;
16782         }
16783
16784         atomic_inc(&p->user->__count);
16785 @@ -1352,6 +1386,18 @@ static struct task_struct *copy_process(
16786
16787         total_forks++;
16788         spin_unlock(&current->sighand->siglock);
16789 +
16790 +       /* p is copy of current */
16791 +       vxi = p->vx_info;
16792 +       if (vxi) {
16793 +               claim_vx_info(vxi, p);
16794 +               atomic_inc(&vxi->cvirt.nr_threads);
16795 +               atomic_inc(&vxi->cvirt.total_forks);
16796 +               vx_nproc_inc(p);
16797 +       }
16798 +       nxi = p->nx_info;
16799 +       if (nxi)
16800 +               claim_nx_info(nxi, p);
16801         write_unlock_irq(&tasklist_lock);
16802         proc_fork_connector(p);
16803         cgroup_post_fork(p);
16804 @@ -1398,6 +1444,9 @@ bad_fork_cleanup_count:
16805         put_group_info(p->group_info);
16806         atomic_dec(&p->user->processes);
16807         free_uid(p->user);
16808 +bad_fork_cleanup_vm:
16809 +       if (p->mm && !(clone_flags & CLONE_VM))
16810 +               vx_pages_sub(p->vx_info, RLIMIT_AS, p->mm->total_vm);
16811  bad_fork_free:
16812         free_task(p);
16813  fork_out:
16814 diff -NurpP --minimal linux-2.6.25.11/kernel/kthread.c linux-2.6.25.11-vs2.3.0.34.14/kernel/kthread.c
16815 --- linux-2.6.25.11/kernel/kthread.c    2008-04-17 12:05:44.000000000 -0400
16816 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/kthread.c      2008-04-19 15:14:52.000000000 -0400
16817 @@ -13,6 +13,7 @@
16818  #include <linux/file.h>
16819  #include <linux/module.h>
16820  #include <linux/mutex.h>
16821 +#include <linux/vs_pid.h>
16822  #include <asm/semaphore.h>
16823
16824  #define KTHREAD_NICE_LEVEL (-5)
16825 @@ -99,7 +100,7 @@ static void create_kthread(struct kthrea
16826                 struct sched_param param = { .sched_priority = 0 };
16827                 wait_for_completion(&create->started);
16828                 read_lock(&tasklist_lock);
16829 -               create->result = find_task_by_pid(pid);
16830 +               create->result = find_task_by_real_pid(pid);
16831                 read_unlock(&tasklist_lock);
16832                 /*
16833                  * root may have changed our (kthreadd's) priority or CPU mask.
16834 diff -NurpP --minimal linux-2.6.25.11/kernel/Makefile linux-2.6.25.11-vs2.3.0.34.14/kernel/Makefile
16835 --- linux-2.6.25.11/kernel/Makefile     2008-04-17 12:05:44.000000000 -0400
16836 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/Makefile       2008-04-19 15:14:52.000000000 -0400
16837 @@ -11,6 +11,8 @@ obj-y     = sched.o fork.o exec_domain.o
16838             hrtimer.o rwsem.o nsproxy.o srcu.o \
16839             notifier.o ksysfs.o pm_qos_params.o
16840
16841 +obj-y += vserver/
16842 +
16843  obj-$(CONFIG_SYSCTL) += sysctl_check.o
16844  obj-$(CONFIG_STACKTRACE) += stacktrace.o
16845  obj-y += time/
16846 diff -NurpP --minimal linux-2.6.25.11/kernel/nsproxy.c linux-2.6.25.11-vs2.3.0.34.14/kernel/nsproxy.c
16847 --- linux-2.6.25.11/kernel/nsproxy.c    2008-04-17 12:05:44.000000000 -0400
16848 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/nsproxy.c      2008-07-16 00:10:51.000000000 -0400
16849 @@ -20,6 +20,8 @@
16850  #include <linux/mnt_namespace.h>
16851  #include <linux/utsname.h>
16852  #include <linux/pid_namespace.h>
16853 +#include <linux/vserver/global.h>
16854 +#include <linux/vserver/debug.h>
16855  #include <net/net_namespace.h>
16856  #include <linux/ipc_namespace.h>
16857
16858 @@ -38,6 +40,9 @@ static inline struct nsproxy *clone_nspr
16859         if (ns) {
16860                 memcpy(ns, orig, sizeof(struct nsproxy));
16861                 atomic_set(&ns->count, 1);
16862 +               vxdprintk(VXD_CBIT(space, 2), "clone_nsproxy(%p[%u] = %p[1]",
16863 +                       orig, atomic_read(&orig->count), ns);
16864 +               atomic_inc(&vs_global_nsproxy);
16865         }
16866         return ns;
16867  }
16868 @@ -47,47 +52,51 @@ static inline struct nsproxy *clone_nspr
16869   * Return the newly created nsproxy.  Do not attach this to the task,
16870   * leave it to the caller to do proper locking and attach it to task.
16871   */
16872 -static struct nsproxy *create_new_namespaces(unsigned long flags,
16873 -                       struct task_struct *tsk, struct fs_struct *new_fs)
16874 +static struct nsproxy *unshare_namespaces(unsigned long flags,
16875 +                       struct nsproxy *orig, struct fs_struct *new_fs)
16876  {
16877         struct nsproxy *new_nsp;
16878         int err;
16879
16880 -       new_nsp = clone_nsproxy(tsk->nsproxy);
16881 +       vxdprintk(VXD_CBIT(space, 4),
16882 +               "unshare_namespaces(0x%08lx,%p,%p)",
16883 +               flags, orig, new_fs);
16884 +
16885 +       new_nsp = clone_nsproxy(orig);
16886         if (!new_nsp)
16887                 return ERR_PTR(-ENOMEM);
16888
16889 -       new_nsp->mnt_ns = copy_mnt_ns(flags, tsk->nsproxy->mnt_ns, new_fs);
16890 +       new_nsp->mnt_ns = copy_mnt_ns(flags, orig->mnt_ns, new_fs);
16891         if (IS_ERR(new_nsp->mnt_ns)) {
16892                 err = PTR_ERR(new_nsp->mnt_ns);
16893                 goto out_ns;
16894         }
16895
16896 -       new_nsp->uts_ns = copy_utsname(flags, tsk->nsproxy->uts_ns);
16897 +       new_nsp->uts_ns = copy_utsname(flags, orig->uts_ns);
16898         if (IS_ERR(new_nsp->uts_ns)) {
16899                 err = PTR_ERR(new_nsp->uts_ns);
16900                 goto out_uts;
16901         }
16902
16903 -       new_nsp->ipc_ns = copy_ipcs(flags, tsk->nsproxy->ipc_ns);
16904 +       new_nsp->ipc_ns = copy_ipcs(flags, orig->ipc_ns);
16905         if (IS_ERR(new_nsp->ipc_ns)) {
16906                 err = PTR_ERR(new_nsp->ipc_ns);
16907                 goto out_ipc;
16908         }
16909
16910 -       new_nsp->pid_ns = copy_pid_ns(flags, task_active_pid_ns(tsk));
16911 +       new_nsp->pid_ns = copy_pid_ns(flags, orig->pid_ns);
16912         if (IS_ERR(new_nsp->pid_ns)) {
16913                 err = PTR_ERR(new_nsp->pid_ns);
16914                 goto out_pid;
16915         }
16916
16917 -       new_nsp->user_ns = copy_user_ns(flags, tsk->nsproxy->user_ns);
16918 +       new_nsp->user_ns = copy_user_ns(flags, orig->user_ns);
16919         if (IS_ERR(new_nsp->user_ns)) {
16920                 err = PTR_ERR(new_nsp->user_ns);
16921                 goto out_user;
16922         }
16923
16924 -       new_nsp->net_ns = copy_net_ns(flags, tsk->nsproxy->net_ns);
16925 +       new_nsp->net_ns = copy_net_ns(flags, orig->net_ns);
16926         if (IS_ERR(new_nsp->net_ns)) {
16927                 err = PTR_ERR(new_nsp->net_ns);
16928                 goto out_net;
16929 @@ -115,6 +124,37 @@ out_ns:
16930         return ERR_PTR(err);
16931  }
16932
16933 +static struct nsproxy *create_new_namespaces(int flags, struct task_struct *tsk,
16934 +                       struct fs_struct *new_fs)
16935 +{
16936 +       return unshare_namespaces(flags, tsk->nsproxy, new_fs);
16937 +}
16938 +
16939 +/*
16940 + * copies the nsproxy, setting refcount to 1, and grabbing a
16941 + * reference to all contained namespaces.
16942 + */
16943 +struct nsproxy *copy_nsproxy(struct nsproxy *orig)
16944 +{
16945 +       struct nsproxy *ns = clone_nsproxy(orig);
16946 +
16947 +       if (ns) {
16948 +               if (ns->mnt_ns)
16949 +                       get_mnt_ns(ns->mnt_ns);
16950 +               if (ns->uts_ns)
16951 +                       get_uts_ns(ns->uts_ns);
16952 +               if (ns->ipc_ns)
16953 +                       get_ipc_ns(ns->ipc_ns);
16954 +               if (ns->pid_ns)
16955 +                       get_pid_ns(ns->pid_ns);
16956 +               if (ns->user_ns)
16957 +                       get_user_ns(ns->user_ns);
16958 +               if (ns->net_ns)
16959 +                       get_net(ns->net_ns);
16960 +       }
16961 +       return ns;
16962 +}
16963 +
16964  /*
16965   * called from clone.  This now handles copy for nsproxy and all
16966   * namespaces therein.
16967 @@ -122,9 +162,12 @@ out_ns:
16968  int copy_namespaces(unsigned long flags, struct task_struct *tsk)
16969  {
16970         struct nsproxy *old_ns = tsk->nsproxy;
16971 -       struct nsproxy *new_ns;
16972 +       struct nsproxy *new_ns = NULL;
16973         int err = 0;
16974
16975 +       vxdprintk(VXD_CBIT(space, 7), "copy_namespaces(0x%08lx,%p[%p])",
16976 +               flags, tsk, old_ns);
16977 +
16978         if (!old_ns)
16979                 return 0;
16980
16981 @@ -155,6 +198,9 @@ int copy_namespaces(unsigned long flags,
16982
16983  out:
16984         put_nsproxy(old_ns);
16985 +       vxdprintk(VXD_CBIT(space, 3),
16986 +               "copy_namespaces(0x%08lx,%p[%p]) = %d [%p]",
16987 +               flags, tsk, old_ns, err, new_ns);
16988         return err;
16989  }
16990
16991 @@ -170,7 +216,9 @@ void free_nsproxy(struct nsproxy *ns)
16992                 put_pid_ns(ns->pid_ns);
16993         if (ns->user_ns)
16994                 put_user_ns(ns->user_ns);
16995 -       put_net(ns->net_ns);
16996 +       if (ns->net_ns)
16997 +               put_net(ns->net_ns);
16998 +       atomic_dec(&vs_global_nsproxy);
16999         kmem_cache_free(nsproxy_cachep, ns);
17000  }
17001
17002 @@ -183,6 +231,10 @@ int unshare_nsproxy_namespaces(unsigned
17003  {
17004         int err = 0;
17005
17006 +       vxdprintk(VXD_CBIT(space, 4),
17007 +               "unshare_nsproxy_namespaces(0x%08lx,[%p])",
17008 +               unshare_flags, current->nsproxy);
17009 +
17010         if (!(unshare_flags & (CLONE_NEWNS | CLONE_NEWUTS | CLONE_NEWIPC |
17011                                CLONE_NEWUSER | CLONE_NEWNET)))
17012                 return 0;
17013 diff -NurpP --minimal linux-2.6.25.11/kernel/pid.c linux-2.6.25.11-vs2.3.0.34.14/kernel/pid.c
17014 --- linux-2.6.25.11/kernel/pid.c        2008-04-17 12:05:44.000000000 -0400
17015 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/pid.c  2008-04-29 17:56:00.000000000 -0400
17016 @@ -35,6 +35,8 @@
17017  #include <linux/pid_namespace.h>
17018  #include <linux/init_task.h>
17019  #include <linux/syscalls.h>
17020 +#include <linux/vs_pid.h>
17021 +#include <linux/vserver/global.h>
17022
17023  #define pid_hashfn(nr, ns)     \
17024         hash_long((unsigned long)nr + (unsigned long)ns, pidhash_shift)
17025 @@ -303,7 +305,7 @@ EXPORT_SYMBOL_GPL(find_pid_ns);
17026
17027  struct pid *find_vpid(int nr)
17028  {
17029 -       return find_pid_ns(nr, current->nsproxy->pid_ns);
17030 +       return find_pid_ns(vx_rmap_pid(nr), current->nsproxy->pid_ns);
17031  }
17032  EXPORT_SYMBOL_GPL(find_vpid);
17033
17034 @@ -359,6 +361,9 @@ void transfer_pid(struct task_struct *ol
17035  struct task_struct *pid_task(struct pid *pid, enum pid_type type)
17036  {
17037         struct task_struct *result = NULL;
17038 +
17039 +       if (type == PIDTYPE_REALPID)
17040 +               type = PIDTYPE_PID;
17041         if (pid) {
17042                 struct hlist_node *first;
17043                 first = rcu_dereference(pid->tasks[type].first);
17044 @@ -388,14 +393,14 @@ EXPORT_SYMBOL(find_task_by_pid);
17045
17046  struct task_struct *find_task_by_vpid(pid_t vnr)
17047  {
17048 -       return find_task_by_pid_type_ns(PIDTYPE_PID, vnr,
17049 +       return find_task_by_pid_type_ns(PIDTYPE_PID, vx_rmap_pid(vnr),
17050                         current->nsproxy->pid_ns);
17051  }
17052  EXPORT_SYMBOL(find_task_by_vpid);
17053
17054  struct task_struct *find_task_by_pid_ns(pid_t nr, struct pid_namespace *ns)
17055  {
17056 -       return find_task_by_pid_type_ns(PIDTYPE_PID, nr, ns);
17057 +       return find_task_by_pid_type_ns(PIDTYPE_PID, vx_rmap_pid(nr), ns);
17058  }
17059  EXPORT_SYMBOL(find_task_by_pid_ns);
17060
17061 @@ -430,7 +435,7 @@ struct pid *find_get_pid(pid_t nr)
17062         return pid;
17063  }
17064
17065 -pid_t pid_nr_ns(struct pid *pid, struct pid_namespace *ns)
17066 +pid_t pid_unmapped_nr_ns(struct pid *pid, struct pid_namespace *ns)
17067  {
17068         struct upid *upid;
17069         pid_t nr = 0;
17070 @@ -443,6 +448,11 @@ pid_t pid_nr_ns(struct pid *pid, struct
17071         return nr;
17072  }
17073
17074 +pid_t pid_nr_ns(struct pid *pid, struct pid_namespace *ns)
17075 +{
17076 +       return vx_map_pid(pid_unmapped_nr_ns(pid, ns));
17077 +}
17078 +
17079  pid_t pid_vnr(struct pid *pid)
17080  {
17081         return pid_nr_ns(pid, current->nsproxy->pid_ns);
17082 diff -NurpP --minimal linux-2.6.25.11/kernel/pid_namespace.c linux-2.6.25.11-vs2.3.0.34.14/kernel/pid_namespace.c
17083 --- linux-2.6.25.11/kernel/pid_namespace.c      2008-04-17 12:05:44.000000000 -0400
17084 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/pid_namespace.c        2008-04-23 11:52:08.000000000 -0400
17085 @@ -12,6 +12,7 @@
17086  #include <linux/pid_namespace.h>
17087  #include <linux/syscalls.h>
17088  #include <linux/err.h>
17089 +#include <linux/vserver/global.h>
17090
17091  #define BITS_PER_PAGE          (PAGE_SIZE*8)
17092
17093 @@ -84,6 +85,7 @@ static struct pid_namespace *create_pid_
17094                 goto out_free_map;
17095
17096         kref_init(&ns->kref);
17097 +       atomic_inc(&vs_global_pid_ns);
17098         ns->last_pid = 0;
17099         ns->child_reaper = NULL;
17100         ns->level = level;
17101 @@ -112,6 +114,7 @@ static void destroy_pid_namespace(struct
17102
17103         for (i = 0; i < PIDMAP_ENTRIES; i++)
17104                 kfree(ns->pidmap[i].page);
17105 +       atomic_dec(&vs_global_pid_ns);
17106         kmem_cache_free(pid_ns_cachep, ns);
17107  }
17108
17109 diff -NurpP --minimal linux-2.6.25.11/kernel/posix-timers.c linux-2.6.25.11-vs2.3.0.34.14/kernel/posix-timers.c
17110 --- linux-2.6.25.11/kernel/posix-timers.c       2008-04-17 12:05:44.000000000 -0400
17111 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/posix-timers.c 2008-04-19 15:14:52.000000000 -0400
17112 @@ -47,6 +47,7 @@
17113  #include <linux/wait.h>
17114  #include <linux/workqueue.h>
17115  #include <linux/module.h>
17116 +#include <linux/vs_context.h>
17117
17118  /*
17119   * Management arrays for POSIX timers.  Timers are kept in slab memory
17120 @@ -299,6 +300,12 @@ void do_schedule_next_timer(struct sigin
17121
17122  int posix_timer_event(struct k_itimer *timr,int si_private)
17123  {
17124 +       struct vx_info_save vxis;
17125 +       struct vx_info *vxi;
17126 +       int ret;
17127 +
17128 +       vxi = task_get_vx_info(timr->it_process);
17129 +       enter_vx_info(vxi, &vxis);
17130         memset(&timr->sigq->info, 0, sizeof(siginfo_t));
17131         timr->sigq->info.si_sys_private = si_private;
17132         /* Send signal to the process that owns this timer.*/
17133 @@ -311,11 +318,11 @@ int posix_timer_event(struct k_itimer *t
17134
17135         if (timr->it_sigev_notify & SIGEV_THREAD_ID) {
17136                 struct task_struct *leader;
17137 -               int ret = send_sigqueue(timr->it_sigev_signo, timr->sigq,
17138 -                                       timr->it_process);
17139
17140 +               ret = send_sigqueue(timr->it_sigev_signo, timr->sigq,
17141 +                                   timr->it_process);
17142                 if (likely(ret >= 0))
17143 -                       return ret;
17144 +                       goto out;
17145
17146                 timr->it_sigev_notify = SIGEV_SIGNAL;
17147                 leader = timr->it_process->group_leader;
17148 @@ -323,8 +330,12 @@ int posix_timer_event(struct k_itimer *t
17149                 timr->it_process = leader;
17150         }
17151
17152 -       return send_group_sigqueue(timr->it_sigev_signo, timr->sigq,
17153 +       ret = send_group_sigqueue(timr->it_sigev_signo, timr->sigq,
17154                                    timr->it_process);
17155 +out:
17156 +       leave_vx_info(&vxis);
17157 +       put_vx_info(vxi);
17158 +       return ret;
17159  }
17160  EXPORT_SYMBOL_GPL(posix_timer_event);
17161
17162 diff -NurpP --minimal linux-2.6.25.11/kernel/printk.c linux-2.6.25.11-vs2.3.0.34.14/kernel/printk.c
17163 --- linux-2.6.25.11/kernel/printk.c     2008-04-17 12:05:44.000000000 -0400
17164 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/printk.c       2008-04-21 10:59:28.000000000 -0400
17165 @@ -32,6 +32,7 @@
17166  #include <linux/security.h>
17167  #include <linux/bootmem.h>
17168  #include <linux/syscalls.h>
17169 +#include <linux/vs_cvirt.h>
17170
17171  #include <asm/uaccess.h>
17172
17173 @@ -297,18 +298,13 @@ int do_syslog(int type, char __user *buf
17174         unsigned i, j, limit, count;
17175         int do_clear = 0;
17176         char c;
17177 -       int error = 0;
17178 +       int error;
17179
17180         error = security_syslog(type);
17181         if (error)
17182                 return error;
17183
17184 -       switch (type) {
17185 -       case 0:         /* Close log */
17186 -               break;
17187 -       case 1:         /* Open log */
17188 -               break;
17189 -       case 2:         /* Read from log */
17190 +       if ((type >= 2) && (type <= 4)) {
17191                 error = -EINVAL;
17192                 if (!buf || len < 0)
17193                         goto out;
17194 @@ -319,6 +315,16 @@ int do_syslog(int type, char __user *buf
17195                         error = -EFAULT;
17196                         goto out;
17197                 }
17198 +       }
17199 +       if (!vx_check(0, VS_ADMIN|VS_WATCH))
17200 +               return vx_do_syslog(type, buf, len);
17201 +
17202 +       switch (type) {
17203 +       case 0:         /* Close log */
17204 +               break;
17205 +       case 1:         /* Open log */
17206 +               break;
17207 +       case 2:         /* Read from log */
17208                 error = wait_event_interruptible(log_wait,
17209                                                         (log_start - log_end));
17210                 if (error)
17211 @@ -343,16 +349,6 @@ int do_syslog(int type, char __user *buf
17212                 do_clear = 1;
17213                 /* FALL THRU */
17214         case 3:         /* Read last kernel messages */
17215 -               error = -EINVAL;
17216 -               if (!buf || len < 0)
17217 -                       goto out;
17218 -               error = 0;
17219 -               if (!len)
17220 -                       goto out;
17221 -               if (!access_ok(VERIFY_WRITE, buf, len)) {
17222 -                       error = -EFAULT;
17223 -                       goto out;
17224 -               }
17225                 count = len;
17226                 if (count > log_buf_len)
17227                         count = log_buf_len;
17228 diff -NurpP --minimal linux-2.6.25.11/kernel/ptrace.c linux-2.6.25.11-vs2.3.0.34.14/kernel/ptrace.c
17229 --- linux-2.6.25.11/kernel/ptrace.c     2008-04-17 12:05:44.000000000 -0400
17230 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/ptrace.c       2008-04-21 10:50:28.000000000 -0400
17231 @@ -21,6 +21,7 @@
17232  #include <linux/audit.h>
17233  #include <linux/pid_namespace.h>
17234  #include <linux/syscalls.h>
17235 +#include <linux/vs_context.h>
17236
17237  #include <asm/pgtable.h>
17238  #include <asm/uaccess.h>
17239 @@ -147,6 +148,11 @@ int __ptrace_may_attach(struct task_stru
17240                 dumpable = get_dumpable(task->mm);
17241         if (!dumpable && !capable(CAP_SYS_PTRACE))
17242                 return -EPERM;
17243 +       if (!vx_check(task->xid, VS_ADMIN_P|VS_IDENT))
17244 +               return -EPERM;
17245 +       if (!vx_check(task->xid, VS_IDENT) &&
17246 +               !task_vx_flags(task, VXF_STATE_ADMIN, 0))
17247 +               return -EACCES;
17248
17249         return security_ptrace(current, task);
17250  }
17251 @@ -562,6 +568,10 @@ asmlinkage long sys_ptrace(long request,
17252                 goto out;
17253         }
17254
17255 +       ret = -EPERM;
17256 +       if (!vx_check(vx_task_xid(child), VS_WATCH_P | VS_IDENT))
17257 +               goto out_put_task_struct;
17258 +
17259         if (request == PTRACE_ATTACH) {
17260                 ret = ptrace_attach(child);
17261                 /*
17262 diff -NurpP --minimal linux-2.6.25.11/kernel/sched.c linux-2.6.25.11-vs2.3.0.34.14/kernel/sched.c
17263 --- linux-2.6.25.11/kernel/sched.c      2008-07-15 22:05:18.000000000 -0400
17264 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/sched.c        2008-07-11 12:41:01.000000000 -0400
17265 @@ -66,6 +66,8 @@
17266  #include <linux/unistd.h>
17267  #include <linux/pagemap.h>
17268  #include <linux/hrtimer.h>
17269 +#include <linux/vs_sched.h>
17270 +#include <linux/vs_cvirt.h>
17271
17272  #include <asm/tlb.h>
17273  #include <asm/irq_regs.h>
17274 @@ -375,6 +377,16 @@ struct root_domain {
17275  static struct root_domain def_root_domain;
17276
17277  #endif
17278 +       unsigned long norm_time;
17279 +       unsigned long idle_time;
17280 +#ifdef CONFIG_VSERVER_IDLETIME
17281 +       int idle_skip;
17282 +#endif
17283 +#ifdef CONFIG_VSERVER_HARDCPU
17284 +       struct list_head hold_queue;
17285 +       unsigned long nr_onhold;
17286 +       int idle_tokens;
17287 +#endif
17288
17289  /*
17290   * This is the main, per-CPU runqueue data structure.
17291 @@ -1366,6 +1378,7 @@ static void set_load_weight(struct task_
17292
17293  static void enqueue_task(struct rq *rq, struct task_struct *p, int wakeup)
17294  {
17295 +       // BUG_ON(p->state & TASK_ONHOLD);
17296         sched_info_queued(p);
17297         p->sched_class->enqueue_task(rq, p, wakeup);
17298         p->se.on_rq = 1;
17299 @@ -1556,6 +1569,9 @@ struct migration_req {
17300         struct completion done;
17301  };
17302
17303 +#include "sched_mon.h"
17304 +
17305 +
17306  /*
17307   * The task's runqueue lock must be held.
17308   * Returns true if you have to wait for migration thread.
17309 @@ -1565,6 +1581,7 @@ migrate_task(struct task_struct *p, int
17310  {
17311         struct rq *rq = task_rq(p);
17312
17313 +       vxm_migrate_task(p, rq, dest_cpu);
17314         /*
17315          * If the task is not on a runqueue (and not running), then
17316          * it is sufficient to simply update the task's cpu field.
17317 @@ -1926,6 +1943,12 @@ static int try_to_wake_up(struct task_st
17318                 /* might preempt at this point */
17319                 rq = task_rq_lock(p, &flags);
17320                 old_state = p->state;
17321 +
17322 +       /* we need to unhold suspended tasks
17323 +       if (old_state & TASK_ONHOLD) {
17324 +               vx_unhold_task(p, rq);
17325 +               old_state = p->state;
17326 +       } */
17327                 if (!(old_state & state))
17328                         goto out;
17329                 if (p->se.on_rq)
17330 @@ -3697,13 +3720,16 @@ unsigned long long task_sched_runtime(st
17331  void account_user_time(struct task_struct *p, cputime_t cputime)
17332  {
17333         struct cpu_usage_stat *cpustat = &kstat_this_cpu.cpustat;
17334 +       struct vx_info *vxi = p->vx_info;  /* p is _always_ current */
17335         cputime64_t tmp;
17336 +       int nice = (TASK_NICE(p) > 0);
17337
17338         p->utime = cputime_add(p->utime, cputime);
17339 +       vx_account_user(vxi, cputime, nice);
17340
17341         /* Add user time to cpustat. */
17342         tmp = cputime_to_cputime64(cputime);
17343 -       if (TASK_NICE(p) > 0)
17344 +       if (nice)
17345                 cpustat->nice = cputime64_add(cpustat->nice, tmp);
17346         else
17347                 cpustat->user = cputime64_add(cpustat->user, tmp);
17348 @@ -3748,6 +3774,7 @@ void account_system_time(struct task_str
17349                          cputime_t cputime)
17350  {
17351         struct cpu_usage_stat *cpustat = &kstat_this_cpu.cpustat;
17352 +       struct vx_info *vxi = p->vx_info;  /* p is _always_ current */
17353         struct rq *rq = this_rq();
17354         cputime64_t tmp;
17355
17356 @@ -3755,6 +3782,7 @@ void account_system_time(struct task_str
17357                 return account_guest_time(p, cputime);
17358
17359         p->stime = cputime_add(p->stime, cputime);
17360 +       vx_account_system(vxi, cputime, (p == rq->idle));
17361
17362         /* Add system time to cpustat. */
17363         tmp = cputime_to_cputime64(cputime);
17364 @@ -4500,7 +4528,7 @@ asmlinkage long sys_nice(int increment)
17365                 nice = 19;
17366
17367         if (increment < 0 && !can_nice(current, nice))
17368 -               return -EPERM;
17369 +               return vx_flags(VXF_IGNEG_NICE, 0) ? 0 : -EPERM;
17370
17371         retval = security_task_setnice(current, nice);
17372         if (retval)
17373 diff -NurpP --minimal linux-2.6.25.11/kernel/sched_fair.c linux-2.6.25.11-vs2.3.0.34.14/kernel/sched_fair.c
17374 --- linux-2.6.25.11/kernel/sched_fair.c 2008-04-17 12:05:44.000000000 -0400
17375 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/sched_fair.c   2008-04-19 15:14:52.000000000 -0400
17376 @@ -537,6 +537,8 @@ enqueue_entity(struct cfs_rq *cfs_rq, st
17377         check_spread(cfs_rq, se);
17378         if (se != cfs_rq->curr)
17379                 __enqueue_entity(cfs_rq, se);
17380 +       if (entity_is_task(se))
17381 +               vx_activate_task(task_of(se));
17382         account_entity_enqueue(cfs_rq, se);
17383  }
17384
17385 @@ -580,6 +582,8 @@ dequeue_entity(struct cfs_rq *cfs_rq, st
17386
17387         if (se != cfs_rq->curr)
17388                 __dequeue_entity(cfs_rq, se);
17389 +       if (entity_is_task(se))
17390 +               vx_deactivate_task(task_of(se));
17391         account_entity_dequeue(cfs_rq, se);
17392  }
17393
17394 diff -NurpP --minimal linux-2.6.25.11/kernel/sched_hard.h linux-2.6.25.11-vs2.3.0.34.14/kernel/sched_hard.h
17395 --- linux-2.6.25.11/kernel/sched_hard.h 1969-12-31 19:00:00.000000000 -0500
17396 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/sched_hard.h   2008-04-19 15:14:52.000000000 -0400
17397 @@ -0,0 +1,324 @@
17398 +
17399 +#ifdef CONFIG_VSERVER_IDLELIMIT
17400 +
17401 +/*
17402 + * vx_idle_resched - reschedule after maxidle
17403 + */
17404 +static inline
17405 +void vx_idle_resched(struct rq *rq)
17406 +{
17407 +       /* maybe have a better criterion for paused */
17408 +       if (!--rq->idle_tokens && !list_empty(&rq->hold_queue))
17409 +               set_need_resched();
17410 +}
17411 +
17412 +#else /* !CONFIG_VSERVER_IDLELIMIT */
17413 +
17414 +#define vx_idle_resched(rq)
17415 +
17416 +#endif /* CONFIG_VSERVER_IDLELIMIT */
17417 +
17418 +
17419 +
17420 +#ifdef CONFIG_VSERVER_IDLETIME
17421 +
17422 +#define vx_set_rq_min_skip(rq, min)            \
17423 +       (rq)->idle_skip = (min)
17424 +
17425 +#define vx_save_min_skip(ret, min, val)                \
17426 +       __vx_save_min_skip(ret, min, val)
17427 +
17428 +static inline
17429 +void __vx_save_min_skip(int ret, int *min, int val)
17430 +{
17431 +       if (ret > -2)
17432 +               return;
17433 +       if ((*min > val) || !*min)
17434 +               *min = val;
17435 +}
17436 +
17437 +static inline
17438 +int vx_try_skip(struct rq *rq, int cpu)
17439 +{
17440 +       /* artificially advance time */
17441 +       if (rq->idle_skip > 0) {
17442 +               vxdprintk(list_empty(&rq->hold_queue),
17443 +                       "hold queue empty on cpu %d", cpu);
17444 +               rq->idle_time += rq->idle_skip;
17445 +               vxm_idle_skip(rq, cpu);
17446 +               return 1;
17447 +       }
17448 +       return 0;
17449 +}
17450 +
17451 +#else /* !CONFIG_VSERVER_IDLETIME */
17452 +
17453 +#define vx_set_rq_min_skip(rq, min)            \
17454 +       ({ int dummy = (min); dummy; })
17455 +
17456 +#define vx_save_min_skip(ret, min, val)
17457 +
17458 +static inline
17459 +int vx_try_skip(struct rq *rq, int cpu)
17460 +{
17461 +       return 0;
17462 +}
17463 +
17464 +#endif /* CONFIG_VSERVER_IDLETIME */
17465 +
17466 +
17467 +
17468 +#ifdef CONFIG_VSERVER_HARDCPU
17469 +
17470 +#define vx_set_rq_max_idle(rq, max)            \
17471 +       (rq)->idle_tokens = (max)
17472 +
17473 +#define vx_save_max_idle(ret, min, val)                \
17474 +       __vx_save_max_idle(ret, min, val)
17475 +
17476 +static inline
17477 +void __vx_save_max_idle(int ret, int *min, int val)
17478 +{
17479 +       if (*min > val)
17480 +               *min = val;
17481 +}
17482 +
17483 +
17484 +/*
17485 + * vx_hold_task - put a task on the hold queue
17486 + */
17487 +static inline
17488 +void vx_hold_task(struct task_struct *p, struct rq *rq)
17489 +{
17490 +       __deactivate_task(p, rq);
17491 +       p->state |= TASK_ONHOLD;
17492 +       /* a new one on hold */
17493 +       rq->nr_onhold++;
17494 +       vxm_hold_task(p, rq);
17495 +       list_add_tail(&p->run_list, &rq->hold_queue);
17496 +}
17497 +
17498 +/*
17499 + * vx_unhold_task - put a task back to the runqueue
17500 + */
17501 +static inline
17502 +void vx_unhold_task(struct task_struct *p, struct rq *rq)
17503 +{
17504 +       list_del(&p->run_list);
17505 +       /* one less waiting */
17506 +       rq->nr_onhold--;
17507 +       p->state &= ~TASK_ONHOLD;
17508 +       enqueue_task(p, rq->expired);
17509 +       inc_nr_running(p, rq);
17510 +       vxm_unhold_task(p, rq);
17511 +
17512 +       if (p->static_prio < rq->best_expired_prio)
17513 +               rq->best_expired_prio = p->static_prio;
17514 +}
17515 +
17516 +unsigned long nr_onhold(void)
17517 +{
17518 +       unsigned long i, sum = 0;
17519 +
17520 +       for_each_online_cpu(i)
17521 +               sum += cpu_rq(i)->nr_onhold;
17522 +
17523 +       return sum;
17524 +}
17525 +
17526 +
17527 +
17528 +static inline
17529 +int __vx_tokens_avail(struct _vx_sched_pc *sched_pc)
17530 +{
17531 +       return sched_pc->tokens;
17532 +}
17533 +
17534 +static inline
17535 +void __vx_consume_token(struct _vx_sched_pc *sched_pc)
17536 +{
17537 +       sched_pc->tokens--;
17538 +}
17539 +
17540 +static inline
17541 +int vx_need_resched(struct task_struct *p, int slice, int cpu)
17542 +{
17543 +       struct vx_info *vxi = p->vx_info;
17544 +
17545 +       if (vx_info_flags(vxi, VXF_SCHED_HARD|VXF_SCHED_PRIO, 0)) {
17546 +               struct _vx_sched_pc *sched_pc =
17547 +                       &vx_per_cpu(vxi, sched_pc, cpu);
17548 +               int tokens;
17549 +
17550 +               /* maybe we can simplify that to decrement
17551 +                  the token counter unconditional? */
17552 +
17553 +               if ((tokens = __vx_tokens_avail(sched_pc)) > 0)
17554 +                       __vx_consume_token(sched_pc);
17555 +
17556 +               /* for tokens > 0, one token was consumed */
17557 +               if (tokens < 2)
17558 +                       slice = 0;
17559 +       }
17560 +       vxm_need_resched(p, slice, cpu);
17561 +       return (slice == 0);
17562 +}
17563 +
17564 +
17565 +#define vx_set_rq_time(rq, time) do {  \
17566 +       rq->norm_time = time;           \
17567 +} while (0)
17568 +
17569 +
17570 +static inline
17571 +void vx_try_unhold(struct rq *rq, int cpu)
17572 +{
17573 +       struct vx_info *vxi = NULL;
17574 +       struct list_head *l, *n;
17575 +       int maxidle = HZ;
17576 +       int minskip = 0;
17577 +
17578 +       /* nothing to do? what about pause? */
17579 +       if (list_empty(&rq->hold_queue))
17580 +               return;
17581 +
17582 +       list_for_each_safe(l, n, &rq->hold_queue) {
17583 +               int ret, delta_min[2];
17584 +               struct _vx_sched_pc *sched_pc;
17585 +               struct task_struct *p;
17586 +
17587 +               p = list_entry(l, struct task_struct, run_list);
17588 +               /* don't bother with same context */
17589 +               if (vxi == p->vx_info)
17590 +                       continue;
17591 +
17592 +               vxi = p->vx_info;
17593 +               /* ignore paused contexts */
17594 +               if (vx_info_flags(vxi, VXF_SCHED_PAUSE, 0))
17595 +                       continue;
17596 +
17597 +               sched_pc = &vx_per_cpu(vxi, sched_pc, cpu);
17598 +
17599 +               /* recalc tokens */
17600 +               vxm_sched_info(sched_pc, vxi, cpu);
17601 +               ret = vx_tokens_recalc(sched_pc,
17602 +                       &rq->norm_time, &rq->idle_time, delta_min);
17603 +               vxm_tokens_recalc(sched_pc, rq, vxi, cpu);
17604 +
17605 +               if (ret > 0) {
17606 +                       /* we found a runable context */
17607 +                       vx_unhold_task(p, rq);
17608 +                       break;
17609 +               }
17610 +               vx_save_max_idle(ret, &maxidle, delta_min[0]);
17611 +               vx_save_min_skip(ret, &minskip, delta_min[1]);
17612 +       }
17613 +       vx_set_rq_max_idle(rq, maxidle);
17614 +       vx_set_rq_min_skip(rq, minskip);
17615 +       vxm_rq_max_min(rq, cpu);
17616 +}
17617 +
17618 +
17619 +static inline
17620 +int vx_schedule(struct task_struct *next, struct rq *rq, int cpu)
17621 +{
17622 +       struct vx_info *vxi = next->vx_info;
17623 +       struct _vx_sched_pc *sched_pc;
17624 +       int delta_min[2];
17625 +       int flags, ret;
17626 +
17627 +       if (!vxi)
17628 +               return 1;
17629 +
17630 +       flags = vxi->vx_flags;
17631 +
17632 +       if (unlikely(vs_check_flags(flags, VXF_SCHED_PAUSE, 0)))
17633 +               goto put_on_hold;
17634 +       if (!vs_check_flags(flags, VXF_SCHED_HARD | VXF_SCHED_PRIO, 0))
17635 +               return 1;
17636 +
17637 +       sched_pc = &vx_per_cpu(vxi, sched_pc, cpu);
17638 +#ifdef CONFIG_SMP
17639 +       /* update scheduler params */
17640 +       if (cpu_isset(cpu, vxi->sched.update)) {
17641 +               vx_update_sched_param(&vxi->sched, sched_pc);
17642 +               vxm_update_sched(sched_pc, vxi, cpu);
17643 +               cpu_clear(cpu, vxi->sched.update);
17644 +       }
17645 +#endif
17646 +       vxm_sched_info(sched_pc, vxi, cpu);
17647 +       ret  = vx_tokens_recalc(sched_pc,
17648 +               &rq->norm_time, &rq->idle_time, delta_min);
17649 +       vxm_tokens_recalc(sched_pc, rq, vxi, cpu);
17650 +
17651 +       if (!vs_check_flags(flags, VXF_SCHED_HARD, 0))
17652 +               return 1;
17653 +
17654 +       if (unlikely(ret < 0)) {
17655 +               vx_save_max_idle(ret, &rq->idle_tokens, delta_min[0]);
17656 +               vx_save_min_skip(ret, &rq->idle_skip, delta_min[1]);
17657 +               vxm_rq_max_min(rq, cpu);
17658 +       put_on_hold:
17659 +               vx_hold_task(next, rq);
17660 +               return 0;
17661 +       }
17662 +       return 1;
17663 +}
17664 +
17665 +
17666 +#else /* CONFIG_VSERVER_HARDCPU */
17667 +
17668 +static inline
17669 +void vx_hold_task(struct task_struct *p, struct rq *rq)
17670 +{
17671 +       return;
17672 +}
17673 +
17674 +static inline
17675 +void vx_unhold_task(struct task_struct *p, struct rq *rq)
17676 +{
17677 +       return;
17678 +}
17679 +
17680 +unsigned long nr_onhold(void)
17681 +{
17682 +       return 0;
17683 +}
17684 +
17685 +
17686 +static inline
17687 +int vx_need_resched(struct task_struct *p, int slice, int cpu)
17688 +{
17689 +       return (slice == 0);
17690 +}
17691 +
17692 +
17693 +#define vx_set_rq_time(rq, time)
17694 +
17695 +static inline
17696 +void vx_try_unhold(struct rq *rq, int cpu)
17697 +{
17698 +       return;
17699 +}
17700 +
17701 +static inline
17702 +int vx_schedule(struct task_struct *next, struct rq *rq, int cpu)
17703 +{
17704 +       struct vx_info *vxi = next->vx_info;
17705 +       struct _vx_sched_pc *sched_pc;
17706 +       int delta_min[2];
17707 +       int ret;
17708 +
17709 +       if (!vx_info_flags(vxi, VXF_SCHED_PRIO, 0))
17710 +               return 1;
17711 +
17712 +       sched_pc = &vx_per_cpu(vxi, sched_pc, cpu);
17713 +       vxm_sched_info(sched_pc, vxi, cpu);
17714 +       ret  = vx_tokens_recalc(sched_pc,
17715 +               &rq->norm_time, &rq->idle_time, delta_min);
17716 +       vxm_tokens_recalc(sched_pc, rq, vxi, cpu);
17717 +       return 1;
17718 +}
17719 +
17720 +#endif /* CONFIG_VSERVER_HARDCPU */
17721 +
17722 diff -NurpP --minimal linux-2.6.25.11/kernel/sched_mon.h linux-2.6.25.11-vs2.3.0.34.14/kernel/sched_mon.h
17723 --- linux-2.6.25.11/kernel/sched_mon.h  1969-12-31 19:00:00.000000000 -0500
17724 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/sched_mon.h    2008-04-19 15:14:52.000000000 -0400
17725 @@ -0,0 +1,200 @@
17726 +
17727 +#include <linux/vserver/monitor.h>
17728 +
17729 +#ifdef  CONFIG_VSERVER_MONITOR
17730 +
17731 +#ifdef CONFIG_VSERVER_HARDCPU
17732 +#define HARDCPU(x) (x)
17733 +#else
17734 +#define HARDCPU(x) (0)
17735 +#endif
17736 +
17737 +#ifdef CONFIG_VSERVER_IDLETIME
17738 +#define IDLETIME(x) (x)
17739 +#else
17740 +#define IDLETIME(x) (0)
17741 +#endif
17742 +
17743 +struct _vx_mon_entry *vxm_advance(int cpu);
17744 +
17745 +
17746 +static inline
17747 +void   __vxm_basic(struct _vx_mon_entry *entry, xid_t xid, int type)
17748 +{
17749 +       entry->type = type;
17750 +       entry->xid = xid;
17751 +}
17752 +
17753 +static inline
17754 +void   __vxm_sync(int cpu)
17755 +{
17756 +       struct _vx_mon_entry *entry = vxm_advance(cpu);
17757 +
17758 +       __vxm_basic(entry, 0, VXM_SYNC);
17759 +       entry->ev.sec = xtime.tv_sec;
17760 +       entry->ev.nsec = xtime.tv_nsec;
17761 +}
17762 +
17763 +static inline
17764 +void   __vxm_task(struct task_struct *p, int type)
17765 +{
17766 +       struct _vx_mon_entry *entry = vxm_advance(task_cpu(p));
17767 +
17768 +       __vxm_basic(entry, p->xid, type);
17769 +       entry->ev.tsk.pid = p->pid;
17770 +       entry->ev.tsk.state = p->state;
17771 +}
17772 +
17773 +static inline
17774 +void   __vxm_sched(struct _vx_sched_pc *s, struct vx_info *vxi, int cpu)
17775 +{
17776 +       struct _vx_mon_entry *entry = vxm_advance(cpu);
17777 +
17778 +       __vxm_basic(entry, vxi->vx_id, (VXM_SCHED | s->flags));
17779 +       entry->sd.tokens = s->tokens;
17780 +       entry->sd.norm_time = s->norm_time;
17781 +       entry->sd.idle_time = s->idle_time;
17782 +}
17783 +
17784 +static inline
17785 +void   __vxm_rqinfo1(struct rq *q, int cpu)
17786 +{
17787 +       struct _vx_mon_entry *entry = vxm_advance(cpu);
17788 +
17789 +       entry->type = VXM_RQINFO_1;
17790 +       entry->xid = ((unsigned long)q >> 16) & 0xffff;
17791 +       entry->q1.running = q->nr_running;
17792 +       entry->q1.onhold = HARDCPU(q->nr_onhold);
17793 +       entry->q1.iowait = atomic_read(&q->nr_iowait);
17794 +       entry->q1.uintr = q->nr_uninterruptible;
17795 +       entry->q1.idle_tokens = IDLETIME(q->idle_tokens);
17796 +}
17797 +
17798 +static inline
17799 +void   __vxm_rqinfo2(struct rq *q, int cpu)
17800 +{
17801 +       struct _vx_mon_entry *entry = vxm_advance(cpu);
17802 +
17803 +       entry->type = VXM_RQINFO_2;
17804 +       entry->xid = (unsigned long)q & 0xffff;
17805 +       entry->q2.norm_time = q->norm_time;
17806 +       entry->q2.idle_time = q->idle_time;
17807 +       entry->q2.idle_skip = IDLETIME(q->idle_skip);
17808 +}
17809 +
17810 +static inline
17811 +void   __vxm_update(struct _vx_sched_pc *s, struct vx_info *vxi, int cpu)
17812 +{
17813 +       struct _vx_mon_entry *entry = vxm_advance(cpu);
17814 +
17815 +       __vxm_basic(entry, vxi->vx_id, VXM_UPDATE);
17816 +       entry->ev.tokens = s->tokens;
17817 +}
17818 +
17819 +static inline
17820 +void   __vxm_update1(struct _vx_sched_pc *s, struct vx_info *vxi, int cpu)
17821 +{
17822 +       struct _vx_mon_entry *entry = vxm_advance(cpu);
17823 +
17824 +       __vxm_basic(entry, vxi->vx_id, VXM_UPDATE_1);
17825 +       entry->u1.tokens_max = s->tokens_max;
17826 +       entry->u1.fill_rate = s->fill_rate[0];
17827 +       entry->u1.interval = s->interval[0];
17828 +}
17829 +
17830 +static inline
17831 +void   __vxm_update2(struct _vx_sched_pc *s, struct vx_info *vxi, int cpu)
17832 +{
17833 +       struct _vx_mon_entry *entry = vxm_advance(cpu);
17834 +
17835 +       __vxm_basic(entry, vxi->vx_id, VXM_UPDATE_2);
17836 +       entry->u2.tokens_min = s->tokens_min;
17837 +       entry->u2.fill_rate = s->fill_rate[1];
17838 +       entry->u2.interval = s->interval[1];
17839 +}
17840 +
17841 +
17842 +#define        vxm_activate_task(p,q)          __vxm_task(p, VXM_ACTIVATE)
17843 +#define        vxm_activate_idle(p,q)          __vxm_task(p, VXM_IDLE)
17844 +#define        vxm_deactivate_task(p,q)        __vxm_task(p, VXM_DEACTIVATE)
17845 +#define        vxm_hold_task(p,q)              __vxm_task(p, VXM_HOLD)
17846 +#define        vxm_unhold_task(p,q)            __vxm_task(p, VXM_UNHOLD)
17847 +
17848 +static inline
17849 +void   vxm_migrate_task(struct task_struct *p, struct rq *rq, int dest)
17850 +{
17851 +       __vxm_task(p, VXM_MIGRATE);
17852 +       __vxm_rqinfo1(rq, task_cpu(p));
17853 +       __vxm_rqinfo2(rq, task_cpu(p));
17854 +}
17855 +
17856 +static inline
17857 +void   vxm_idle_skip(struct rq *rq, int cpu)
17858 +{
17859 +       __vxm_rqinfo1(rq, cpu);
17860 +       __vxm_rqinfo2(rq, cpu);
17861 +}
17862 +
17863 +static inline
17864 +void   vxm_need_resched(struct task_struct *p, int slice, int cpu)
17865 +{
17866 +       if (slice)
17867 +               return;
17868 +
17869 +       __vxm_task(p, VXM_RESCHED);
17870 +}
17871 +
17872 +static inline
17873 +void   vxm_sync(unsigned long now, int cpu)
17874 +{
17875 +       if (!CONFIG_VSERVER_MONITOR_SYNC ||
17876 +               (now % CONFIG_VSERVER_MONITOR_SYNC))
17877 +               return;
17878 +
17879 +       __vxm_sync(cpu);
17880 +}
17881 +
17882 +#define        vxm_sched_info(s,v,c)           __vxm_sched(s,v,c)
17883 +
17884 +static inline
17885 +void   vxm_tokens_recalc(struct _vx_sched_pc *s, struct rq *rq,
17886 +       struct vx_info *vxi, int cpu)
17887 +{
17888 +       __vxm_sched(s, vxi, cpu);
17889 +       __vxm_rqinfo2(rq, cpu);
17890 +}
17891 +
17892 +static inline
17893 +void   vxm_update_sched(struct _vx_sched_pc *s, struct vx_info *vxi, int cpu)
17894 +{
17895 +       __vxm_sched(s, vxi, cpu);
17896 +       __vxm_update(s, vxi, cpu);
17897 +       __vxm_update1(s, vxi, cpu);
17898 +       __vxm_update2(s, vxi, cpu);
17899 +}
17900 +
17901 +static inline
17902 +void   vxm_rq_max_min(struct rq *rq, int cpu)
17903 +{
17904 +       __vxm_rqinfo1(rq, cpu);
17905 +       __vxm_rqinfo2(rq, cpu);
17906 +}
17907 +
17908 +#else  /* CONFIG_VSERVER_MONITOR */
17909 +
17910 +#define        vxm_activate_task(t,q)          do { } while (0)
17911 +#define        vxm_activate_idle(t,q)          do { } while (0)
17912 +#define        vxm_deactivate_task(t,q)        do { } while (0)
17913 +#define        vxm_hold_task(t,q)              do { } while (0)
17914 +#define        vxm_unhold_task(t,q)            do { } while (0)
17915 +#define        vxm_migrate_task(t,q,d)         do { } while (0)
17916 +#define        vxm_idle_skip(q,c)              do { } while (0)
17917 +#define        vxm_need_resched(t,s,c)         do { } while (0)
17918 +#define        vxm_sync(s,c)                   do { } while (0)
17919 +#define        vxm_sched_info(s,v,c)           do { } while (0)
17920 +#define        vxm_tokens_recalc(s,q,v,c)      do { } while (0)
17921 +#define        vxm_update_sched(s,v,c)         do { } while (0)
17922 +#define        vxm_rq_max_min(q,c)             do { } while (0)
17923 +
17924 +#endif /* CONFIG_VSERVER_MONITOR */
17925 +
17926 diff -NurpP --minimal linux-2.6.25.11/kernel/signal.c linux-2.6.25.11-vs2.3.0.34.14/kernel/signal.c
17927 --- linux-2.6.25.11/kernel/signal.c     2008-04-17 12:05:44.000000000 -0400
17928 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/signal.c       2008-07-11 12:42:23.000000000 -0400
17929 @@ -26,6 +26,8 @@
17930  #include <linux/freezer.h>
17931  #include <linux/pid_namespace.h>
17932  #include <linux/nsproxy.h>
17933 +#include <linux/vs_context.h>
17934 +#include <linux/vs_pid.h>
17935
17936  #include <asm/param.h>
17937  #include <asm/uaccess.h>
17938 @@ -530,6 +532,14 @@ static int check_kill_permission(int sig
17939         if (!valid_signal(sig))
17940                 return error;
17941
17942 +       if ((info != SEND_SIG_NOINFO) &&
17943 +               (is_si_special(info) || !SI_FROMUSER(info)))
17944 +               goto skip;
17945 +
17946 +       vxdprintk(VXD_CBIT(misc, 7),
17947 +               "check_kill_permission(%d,%p,%p[#%u,%u])",
17948 +               sig, info, t, vx_task_xid(t), t->pid);
17949 +
17950         if (info == SEND_SIG_NOINFO || (!is_si_special(info) && SI_FROMUSER(info))) {
17951                 error = audit_signal_info(sig, t); /* Let audit system see the signal */
17952                 if (error)
17953 @@ -543,6 +553,18 @@ static int check_kill_permission(int sig
17954                 return error;
17955         }
17956
17957 +       error = -EPERM;
17958 +       if (t->pid == 1 && current->xid)
17959 +               return error;
17960 +
17961 +       error = -ESRCH;
17962 +       if (!vx_check(vx_task_xid(t), VS_WATCH_P | VS_IDENT)) {
17963 +               vxdprintk(current->xid || VXD_CBIT(misc, 7),
17964 +                       "signal %d[%p] xid mismatch %p[#%u,%u] xid=#%u",
17965 +                       sig, info, t, vx_task_xid(t), t->pid, current->xid);
17966 +               return error;
17967 +       }
17968 +skip:
17969         return security_task_kill(t, info, sig, 0);
17970  }
17971
17972 @@ -1048,7 +1070,7 @@ int kill_pid_info(int sig, struct siginf
17973
17974  retry:
17975         p = pid_task(pid, PIDTYPE_PID);
17976 -       if (p) {
17977 +       if (p && vx_check(vx_task_xid(p), VS_WATCH | VS_IDENT)) {
17978                 error = group_send_sig_info(sig, info, p);
17979                 if (unlikely(error == -ESRCH))
17980                         /*
17981 @@ -1088,7 +1110,7 @@ int kill_pid_info_as_uid(int sig, struct
17982
17983         read_lock(&tasklist_lock);
17984         p = pid_task(pid, PIDTYPE_PID);
17985 -       if (!p) {
17986 +       if (!p || !vx_check(vx_task_xid(p), VS_IDENT)) {
17987                 ret = -ESRCH;
17988                 goto out_unlock;
17989         }
17990 @@ -1140,7 +1162,9 @@ static int kill_something_info(int sig,
17991                 struct task_struct * p;
17992
17993                 for_each_process(p) {
17994 -                       if (p->pid > 1 && !same_thread_group(p, current)) {
17995 +                       if (vx_check(vx_task_xid(p), VS_ADMIN|VS_IDENT) &&
17996 +                               p->pid > 1 && !same_thread_group(p, current) &&
17997 +                               !vx_current_initpid(p->pid)) {
17998                                 int err = group_send_sig_info(sig, info, p);
17999                                 ++count;
18000                                 if (err != -EPERM)
18001 @@ -1842,6 +1866,11 @@ relock:
18002                 if (is_global_init(current))
18003                         continue;
18004
18005 +               /* virtual init is protected against user signals */
18006 +               if ((info->si_code == SI_USER) &&
18007 +                       vx_current_initpid(current->pid))
18008 +                       continue;
18009 +
18010                 if (sig_kernel_stop(signr)) {
18011                         /*
18012                          * The default action is to stop all threads in
18013 diff -NurpP --minimal linux-2.6.25.11/kernel/softirq.c linux-2.6.25.11-vs2.3.0.34.14/kernel/softirq.c
18014 --- linux-2.6.25.11/kernel/softirq.c    2008-04-17 12:05:44.000000000 -0400
18015 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/softirq.c      2008-04-19 15:14:52.000000000 -0400
18016 @@ -21,6 +21,7 @@
18017  #include <linux/rcupdate.h>
18018  #include <linux/smp.h>
18019  #include <linux/tick.h>
18020 +#include <linux/vs_context.h>
18021
18022  #include <asm/irq.h>
18023  /*
18024 diff -NurpP --minimal linux-2.6.25.11/kernel/sys.c linux-2.6.25.11-vs2.3.0.34.14/kernel/sys.c
18025 --- linux-2.6.25.11/kernel/sys.c        2008-04-17 12:05:44.000000000 -0400
18026 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/sys.c  2008-04-19 15:14:52.000000000 -0400
18027 @@ -38,6 +38,7 @@
18028  #include <linux/syscalls.h>
18029  #include <linux/kprobes.h>
18030  #include <linux/user_namespace.h>
18031 +#include <linux/vs_pid.h>
18032
18033  #include <asm/uaccess.h>
18034  #include <asm/io.h>
18035 @@ -116,7 +117,10 @@ static int set_one_prio(struct task_stru
18036                 goto out;
18037         }
18038         if (niceval < task_nice(p) && !can_nice(p, niceval)) {
18039 -               error = -EACCES;
18040 +               if (vx_flags(VXF_IGNEG_NICE, 0))
18041 +                       error = 0;
18042 +               else
18043 +                       error = -EACCES;
18044                 goto out;
18045         }
18046         no_nice = security_task_setnice(p, niceval);
18047 @@ -164,6 +168,8 @@ asmlinkage long sys_setpriority(int whic
18048                         else
18049                                 pgrp = task_pgrp(current);
18050                         do_each_pid_task(pgrp, PIDTYPE_PGID, p) {
18051 +                               if (!vx_check(p->xid, VS_ADMIN_P | VS_IDENT))
18052 +                                       continue;
18053                                 error = set_one_prio(p, niceval, error);
18054                         } while_each_pid_task(pgrp, PIDTYPE_PGID, p);
18055                         break;
18056 @@ -224,6 +230,8 @@ asmlinkage long sys_getpriority(int whic
18057                         else
18058                                 pgrp = task_pgrp(current);
18059                         do_each_pid_task(pgrp, PIDTYPE_PGID, p) {
18060 +                               if (!vx_check(p->xid, VS_ADMIN_P | VS_IDENT))
18061 +                                       continue;
18062                                 niceval = 20 - task_nice(p);
18063                                 if (niceval > retval)
18064                                         retval = niceval;
18065 @@ -353,6 +361,9 @@ void kernel_power_off(void)
18066         machine_power_off();
18067  }
18068  EXPORT_SYMBOL_GPL(kernel_power_off);
18069 +
18070 +long vs_reboot(unsigned int, void __user *);
18071 +
18072  /*
18073   * Reboot system call: for obvious reasons only root may call it,
18074   * and even root needs to set up some magic numbers in the registers
18075 @@ -383,6 +394,9 @@ asmlinkage long sys_reboot(int magic1, i
18076         if ((cmd == LINUX_REBOOT_CMD_POWER_OFF) && !pm_power_off)
18077                 cmd = LINUX_REBOOT_CMD_HALT;
18078
18079 +       if (!vx_check(0, VS_ADMIN|VS_WATCH))
18080 +               return vs_reboot(cmd, arg);
18081 +
18082         lock_kernel();
18083         switch (cmd) {
18084         case LINUX_REBOOT_CMD_RESTART:
18085 @@ -1343,7 +1357,7 @@ asmlinkage long sys_sethostname(char __u
18086         int errno;
18087         char tmp[__NEW_UTS_LEN];
18088
18089 -       if (!capable(CAP_SYS_ADMIN))
18090 +       if (!vx_capable(CAP_SYS_ADMIN, VXC_SET_UTSNAME))
18091                 return -EPERM;
18092         if (len < 0 || len > __NEW_UTS_LEN)
18093                 return -EINVAL;
18094 @@ -1388,7 +1402,7 @@ asmlinkage long sys_setdomainname(char _
18095         int errno;
18096         char tmp[__NEW_UTS_LEN];
18097
18098 -       if (!capable(CAP_SYS_ADMIN))
18099 +       if (!vx_capable(CAP_SYS_ADMIN, VXC_SET_UTSNAME))
18100                 return -EPERM;
18101         if (len < 0 || len > __NEW_UTS_LEN)
18102                 return -EINVAL;
18103 @@ -1455,7 +1469,7 @@ asmlinkage long sys_setrlimit(unsigned i
18104                 return -EINVAL;
18105         old_rlim = current->signal->rlim + resource;
18106         if ((new_rlim.rlim_max > old_rlim->rlim_max) &&
18107 -           !capable(CAP_SYS_RESOURCE))
18108 +           !vx_capable(CAP_SYS_RESOURCE, VXC_SET_RLIMIT))
18109                 return -EPERM;
18110         if (resource == RLIMIT_NOFILE && new_rlim.rlim_max > sysctl_nr_open)
18111                 return -EPERM;
18112 diff -NurpP --minimal linux-2.6.25.11/kernel/sysctl.c linux-2.6.25.11-vs2.3.0.34.14/kernel/sysctl.c
18113 --- linux-2.6.25.11/kernel/sysctl.c     2008-04-17 12:05:44.000000000 -0400
18114 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/sysctl.c       2008-04-19 15:14:52.000000000 -0400
18115 @@ -107,6 +107,7 @@ static int ngroups_max = NGROUPS_MAX;
18116  #ifdef CONFIG_KMOD
18117  extern char modprobe_path[];
18118  #endif
18119 +extern char vshelper_path[];
18120  #ifdef CONFIG_CHR_DEV_SG
18121  extern int sg_big_buff;
18122  #endif
18123 @@ -492,6 +493,15 @@ static struct ctl_table kern_table[] = {
18124                 .strategy       = &sysctl_string,
18125         },
18126  #endif
18127 +       {
18128 +               .ctl_name       = KERN_VSHELPER,
18129 +               .procname       = "vshelper",
18130 +               .data           = &vshelper_path,
18131 +               .maxlen         = 256,
18132 +               .mode           = 0644,
18133 +               .proc_handler   = &proc_dostring,
18134 +               .strategy       = &sysctl_string,
18135 +       },
18136  #ifdef CONFIG_CHR_DEV_SG
18137         {
18138                 .ctl_name       = KERN_SG_BIG_BUFF,
18139 diff -NurpP --minimal linux-2.6.25.11/kernel/sysctl_check.c linux-2.6.25.11-vs2.3.0.34.14/kernel/sysctl_check.c
18140 --- linux-2.6.25.11/kernel/sysctl_check.c       2008-04-17 12:05:44.000000000 -0400
18141 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/sysctl_check.c 2008-04-21 13:42:56.000000000 -0400
18142 @@ -39,6 +39,7 @@ static const struct trans_ctl_table tran
18143
18144         { KERN_PANIC,                   "panic" },
18145         { KERN_REALROOTDEV,             "real-root-dev" },
18146 +       { KERN_VSHELPER,                "vshelper", },
18147
18148         { KERN_SPARC_REBOOT,            "reboot-cmd" },
18149         { KERN_CTLALTDEL,               "ctrl-alt-del" },
18150 @@ -1217,6 +1218,22 @@ static const struct trans_ctl_table tran
18151         {}
18152  };
18153
18154 +static struct trans_ctl_table trans_vserver_table[] = {
18155 +       { 1,    "debug_switch" },
18156 +       { 2,    "debug_xid" },
18157 +       { 3,    "debug_nid" },
18158 +       { 4,    "debug_tag" },
18159 +       { 5,    "debug_net" },
18160 +       { 6,    "debug_limit" },
18161 +       { 7,    "debug_cres" },
18162 +       { 8,    "debug_dlim" },
18163 +       { 9,    "debug_quota" },
18164 +       { 10,   "debug_cvirt" },
18165 +       { 11,   "debug_space" },
18166 +       { 12,   "debug_misc" },
18167 +       {}
18168 +};
18169 +
18170  static const struct trans_ctl_table trans_root_table[] = {
18171         { CTL_KERN,     "kernel",       trans_kern_table },
18172         { CTL_VM,       "vm",           trans_vm_table },
18173 @@ -1233,6 +1250,7 @@ static const struct trans_ctl_table tran
18174         { CTL_SUNRPC,   "sunrpc",       trans_sunrpc_table },
18175         { CTL_PM,       "pm",           trans_pm_table },
18176         { CTL_FRV,      "frv",          trans_frv_table },
18177 +       { CTL_VSERVER,  "vserver",      trans_vserver_table },
18178         {}
18179  };
18180
18181 diff -NurpP --minimal linux-2.6.25.11/kernel/time.c linux-2.6.25.11-vs2.3.0.34.14/kernel/time.c
18182 --- linux-2.6.25.11/kernel/time.c       2008-04-17 12:05:44.000000000 -0400
18183 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/time.c 2008-04-19 15:14:52.000000000 -0400
18184 @@ -60,6 +60,7 @@ EXPORT_SYMBOL(sys_tz);
18185  asmlinkage long sys_time(time_t __user * tloc)
18186  {
18187         time_t i = get_seconds();
18188 +/*     FIXME: do_gettimeofday(&tv) -> vx_gettimeofday(&tv) */
18189
18190         if (tloc) {
18191                 if (put_user(i,tloc))
18192 @@ -89,7 +90,7 @@ asmlinkage long sys_stime(time_t __user
18193         if (err)
18194                 return err;
18195
18196 -       do_settimeofday(&tv);
18197 +       vx_settimeofday(&tv);
18198         return 0;
18199  }
18200
18201 @@ -100,7 +101,7 @@ asmlinkage long sys_gettimeofday(struct
18202  {
18203         if (likely(tv != NULL)) {
18204                 struct timeval ktv;
18205 -               do_gettimeofday(&ktv);
18206 +               vx_gettimeofday(&ktv);
18207                 if (copy_to_user(tv, &ktv, sizeof(ktv)))
18208                         return -EFAULT;
18209         }
18210 @@ -175,7 +176,7 @@ int do_sys_settimeofday(struct timespec
18211                 /* SMP safe, again the code in arch/foo/time.c should
18212                  * globally block out interrupts when it runs.
18213                  */
18214 -               return do_settimeofday(tv);
18215 +               return vx_settimeofday(tv);
18216         }
18217         return 0;
18218  }
18219 @@ -307,7 +308,7 @@ void getnstimeofday(struct timespec *tv)
18220  {
18221         struct timeval x;
18222
18223 -       do_gettimeofday(&x);
18224 +       vx_gettimeofday(&x);
18225         tv->tv_sec = x.tv_sec;
18226         tv->tv_nsec = x.tv_usec * NSEC_PER_USEC;
18227  }
18228 diff -NurpP --minimal linux-2.6.25.11/kernel/timer.c linux-2.6.25.11-vs2.3.0.34.14/kernel/timer.c
18229 --- linux-2.6.25.11/kernel/timer.c      2008-04-17 12:05:44.000000000 -0400
18230 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/timer.c        2008-04-19 15:14:52.000000000 -0400
18231 @@ -37,6 +37,10 @@
18232  #include <linux/delay.h>
18233  #include <linux/tick.h>
18234  #include <linux/kallsyms.h>
18235 +#include <linux/vs_base.h>
18236 +#include <linux/vs_cvirt.h>
18237 +#include <linux/vs_pid.h>
18238 +#include <linux/vserver/sched.h>
18239
18240  #include <asm/uaccess.h>
18241  #include <asm/unistd.h>
18242 @@ -955,12 +959,6 @@ asmlinkage unsigned long sys_alarm(unsig
18243
18244  #endif
18245
18246 -#ifndef __alpha__
18247 -
18248 -/*
18249 - * The Alpha uses getxpid, getxuid, and getxgid instead.  Maybe this
18250 - * should be moved into arch/i386 instead?
18251 - */
18252
18253  /**
18254   * sys_getpid - return the thread group id of the current process
18255 @@ -989,10 +987,23 @@ asmlinkage long sys_getppid(void)
18256         rcu_read_lock();
18257         pid = task_tgid_vnr(current->real_parent);
18258         rcu_read_unlock();
18259 +       return vx_map_pid(pid);
18260 +}
18261
18262 -       return pid;
18263 +#ifdef __alpha__
18264 +
18265 +/*
18266 + * The Alpha uses getxpid, getxuid, and getxgid instead.
18267 + */
18268 +
18269 +asmlinkage long do_getxpid(long *ppid)
18270 +{
18271 +       *ppid = sys_getppid();
18272 +       return sys_getpid();
18273  }
18274
18275 +#else /* _alpha_ */
18276 +
18277  asmlinkage long sys_getuid(void)
18278  {
18279         /* Only we change this so SMP safe */
18280 @@ -1160,6 +1171,8 @@ int do_sysinfo(struct sysinfo *info)
18281                         tp.tv_nsec = tp.tv_nsec - NSEC_PER_SEC;
18282                         tp.tv_sec++;
18283                 }
18284 +               if (vx_flags(VXF_VIRT_UPTIME, 0))
18285 +                       vx_vsi_uptime(&tp, NULL);
18286                 info->uptime = tp.tv_sec + (tp.tv_nsec ? 1 : 0);
18287
18288                 info->loads[0] = avenrun[0] << (SI_LOAD_SHIFT - FSHIFT);
18289 diff -NurpP --minimal linux-2.6.25.11/kernel/user.c linux-2.6.25.11-vs2.3.0.34.14/kernel/user.c
18290 --- linux-2.6.25.11/kernel/user.c       2008-04-17 12:05:44.000000000 -0400
18291 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/user.c 2008-04-23 16:24:56.000000000 -0400
18292 @@ -219,14 +219,15 @@ static struct kobj_type uids_ktype = {
18293  };
18294
18295  /* create /sys/kernel/uids/<uid>/cpu_share file for this user */
18296 -static int uids_user_create(struct user_struct *up)
18297 +static int uids_user_create(struct user_namespace *ns, struct user_struct *up)
18298  {
18299         struct kobject *kobj = &up->kobj;
18300         int error;
18301
18302         memset(kobj, 0, sizeof(struct kobject));
18303         kobj->kset = uids_kset;
18304 -       error = kobject_init_and_add(kobj, &uids_ktype, NULL, "%d", up->uid);
18305 +       error = kobject_init_and_add(kobj, &uids_ktype, NULL,
18306 +               "%p:%d", ns, up->uid);
18307         if (error) {
18308                 kobject_put(kobj);
18309                 goto done;
18310 @@ -248,7 +249,7 @@ int __init uids_sysfs_init(void)
18311         if (!uids_kset)
18312                 return -ENOMEM;
18313
18314 -       return uids_user_create(&root_user);
18315 +       return uids_user_create(NULL, &root_user);
18316  }
18317
18318  /* work function to remove sysfs directory for a user and free up
18319 @@ -308,7 +309,8 @@ static inline void free_user(struct user
18320  #else  /* CONFIG_USER_SCHED && CONFIG_SYSFS */
18321
18322  int uids_sysfs_init(void) { return 0; }
18323 -static inline int uids_user_create(struct user_struct *up) { return 0; }
18324 +static inline int uids_user_create(struct user_namespace *ns,
18325 +       struct user_struct *up) { return 0; }
18326  static inline void uids_mutex_lock(void) { }
18327  static inline void uids_mutex_unlock(void) { }
18328
18329 @@ -399,7 +401,7 @@ struct user_struct * alloc_uid(struct us
18330                 if (sched_create_user(new) < 0)
18331                         goto out_put_keys;
18332
18333 -               if (uids_user_create(new))
18334 +               if (uids_user_create(ns, new))
18335                         goto out_destoy_sched;
18336
18337                 /*
18338 diff -NurpP --minimal linux-2.6.25.11/kernel/user_namespace.c linux-2.6.25.11-vs2.3.0.34.14/kernel/user_namespace.c
18339 --- linux-2.6.25.11/kernel/user_namespace.c     2008-04-17 12:05:44.000000000 -0400
18340 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/user_namespace.c       2008-04-19 15:14:52.000000000 -0400
18341 @@ -9,6 +9,7 @@
18342  #include <linux/version.h>
18343  #include <linux/nsproxy.h>
18344  #include <linux/user_namespace.h>
18345 +#include <linux/vserver/global.h>
18346
18347  /*
18348   * Clone a new ns copying an original user ns, setting refcount to 1
18349 @@ -26,6 +27,7 @@ static struct user_namespace *clone_user
18350                 return ERR_PTR(-ENOMEM);
18351
18352         kref_init(&ns->kref);
18353 +       atomic_inc(&vs_global_user_ns);
18354
18355         for (n = 0; n < UIDHASH_SZ; ++n)
18356                 INIT_HLIST_HEAD(ns->uidhash_table + n);
18357 @@ -71,5 +73,6 @@ void free_user_ns(struct kref *kref)
18358
18359         ns = container_of(kref, struct user_namespace, kref);
18360         release_uids(ns);
18361 +       atomic_dec(&vs_global_user_ns);
18362         kfree(ns);
18363  }
18364 diff -NurpP --minimal linux-2.6.25.11/kernel/utsname.c linux-2.6.25.11-vs2.3.0.34.14/kernel/utsname.c
18365 --- linux-2.6.25.11/kernel/utsname.c    2008-04-17 10:37:25.000000000 -0400
18366 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/utsname.c      2008-04-19 15:14:52.000000000 -0400
18367 @@ -14,6 +14,7 @@
18368  #include <linux/utsname.h>
18369  #include <linux/version.h>
18370  #include <linux/err.h>
18371 +#include <linux/vserver/global.h>
18372
18373  /*
18374   * Clone a new ns copying an original utsname, setting refcount to 1
18375 @@ -32,6 +33,7 @@ static struct uts_namespace *clone_uts_n
18376         memcpy(&ns->name, &old_ns->name, sizeof(ns->name));
18377         up_read(&uts_sem);
18378         kref_init(&ns->kref);
18379 +       atomic_inc(&vs_global_uts_ns);
18380         return ns;
18381  }
18382
18383 @@ -62,5 +64,6 @@ void free_uts_ns(struct kref *kref)
18384         struct uts_namespace *ns;
18385
18386         ns = container_of(kref, struct uts_namespace, kref);
18387 +       atomic_dec(&vs_global_uts_ns);
18388         kfree(ns);
18389  }
18390 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/cacct.c linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/cacct.c
18391 --- linux-2.6.25.11/kernel/vserver/cacct.c      1969-12-31 19:00:00.000000000 -0500
18392 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/cacct.c        2008-04-19 15:14:52.000000000 -0400
18393 @@ -0,0 +1,42 @@
18394 +/*
18395 + *  linux/kernel/vserver/cacct.c
18396 + *
18397 + *  Virtual Server: Context Accounting
18398 + *
18399 + *  Copyright (C) 2006-2007 Herbert Pötzl
18400 + *
18401 + *  V0.01  added accounting stats
18402 + *
18403 + */
18404 +
18405 +#include <linux/types.h>
18406 +#include <linux/vs_context.h>
18407 +#include <linux/vserver/cacct_cmd.h>
18408 +#include <linux/vserver/cacct_int.h>
18409 +
18410 +#include <asm/errno.h>
18411 +#include <asm/uaccess.h>
18412 +
18413 +
18414 +int vc_sock_stat(struct vx_info *vxi, void __user *data)
18415 +{
18416 +       struct vcmd_sock_stat_v0 vc_data;
18417 +       int j, field;
18418 +
18419 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18420 +               return -EFAULT;
18421 +
18422 +       field = vc_data.field;
18423 +       if ((field < 0) || (field >= VXA_SOCK_SIZE))
18424 +               return -EINVAL;
18425 +
18426 +       for (j = 0; j < 3; j++) {
18427 +               vc_data.count[j] = vx_sock_count(&vxi->cacct, field, j);
18428 +               vc_data.total[j] = vx_sock_total(&vxi->cacct, field, j);
18429 +       }
18430 +
18431 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
18432 +               return -EFAULT;
18433 +       return 0;
18434 +}
18435 +
18436 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/cacct_init.h linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/cacct_init.h
18437 --- linux-2.6.25.11/kernel/vserver/cacct_init.h 1969-12-31 19:00:00.000000000 -0500
18438 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/cacct_init.h   2008-04-19 15:14:52.000000000 -0400
18439 @@ -0,0 +1,25 @@
18440 +
18441 +
18442 +static inline void vx_info_init_cacct(struct _vx_cacct *cacct)
18443 +{
18444 +       int i, j;
18445 +
18446 +
18447 +       for (i = 0; i < VXA_SOCK_SIZE; i++) {
18448 +               for (j = 0; j < 3; j++) {
18449 +                       atomic_set(&cacct->sock[i][j].count, 0);
18450 +                       atomic_set(&cacct->sock[i][j].total, 0);
18451 +               }
18452 +       }
18453 +       for (i = 0; i < 8; i++)
18454 +               atomic_set(&cacct->slab[i], 0);
18455 +       for (i = 0; i < 5; i++)
18456 +               for (j = 0; j < 4; j++)
18457 +                       atomic_set(&cacct->page[i][j], 0);
18458 +}
18459 +
18460 +static inline void vx_info_exit_cacct(struct _vx_cacct *cacct)
18461 +{
18462 +       return;
18463 +}
18464 +
18465 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/cacct_proc.h linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/cacct_proc.h
18466 --- linux-2.6.25.11/kernel/vserver/cacct_proc.h 1969-12-31 19:00:00.000000000 -0500
18467 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/cacct_proc.h   2008-04-19 15:14:52.000000000 -0400
18468 @@ -0,0 +1,53 @@
18469 +#ifndef _VX_CACCT_PROC_H
18470 +#define _VX_CACCT_PROC_H
18471 +
18472 +#include <linux/vserver/cacct_int.h>
18473 +
18474 +
18475 +#define VX_SOCKA_TOP   \
18476 +       "Type\t    recv #/bytes\t\t   send #/bytes\t\t    fail #/bytes\n"
18477 +
18478 +static inline int vx_info_proc_cacct(struct _vx_cacct *cacct, char *buffer)
18479 +{
18480 +       int i, j, length = 0;
18481 +       static char *type[VXA_SOCK_SIZE] = {
18482 +               "UNSPEC", "UNIX", "INET", "INET6", "PACKET", "OTHER"
18483 +       };
18484 +
18485 +       length += sprintf(buffer + length, VX_SOCKA_TOP);
18486 +       for (i = 0; i < VXA_SOCK_SIZE; i++) {
18487 +               length += sprintf(buffer + length, "%s:", type[i]);
18488 +               for (j = 0; j < 3; j++) {
18489 +                       length += sprintf(buffer + length,
18490 +                               "\t%10lu/%-10lu",
18491 +                               vx_sock_count(cacct, i, j),
18492 +                               vx_sock_total(cacct, i, j));
18493 +               }
18494 +               buffer[length++] = '\n';
18495 +       }
18496 +
18497 +       length += sprintf(buffer + length, "\n");
18498 +       length += sprintf(buffer + length,
18499 +               "slab:\t %8u %8u %8u %8u\n",
18500 +               atomic_read(&cacct->slab[1]),
18501 +               atomic_read(&cacct->slab[4]),
18502 +               atomic_read(&cacct->slab[0]),
18503 +               atomic_read(&cacct->slab[2]));
18504 +
18505 +       length += sprintf(buffer + length, "\n");
18506 +       for (i = 0; i < 5; i++) {
18507 +               length += sprintf(buffer + length,
18508 +                       "page[%d]: %8u %8u %8u %8u\t %8u %8u %8u %8u\n", i,
18509 +                       atomic_read(&cacct->page[i][0]),
18510 +                       atomic_read(&cacct->page[i][1]),
18511 +                       atomic_read(&cacct->page[i][2]),
18512 +                       atomic_read(&cacct->page[i][3]),
18513 +                       atomic_read(&cacct->page[i][4]),
18514 +                       atomic_read(&cacct->page[i][5]),
18515 +                       atomic_read(&cacct->page[i][6]),
18516 +                       atomic_read(&cacct->page[i][7]));
18517 +       }
18518 +       return length;
18519 +}
18520 +
18521 +#endif /* _VX_CACCT_PROC_H */
18522 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/context.c linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/context.c
18523 --- linux-2.6.25.11/kernel/vserver/context.c    1969-12-31 19:00:00.000000000 -0500
18524 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/context.c      2008-07-16 00:35:54.000000000 -0400
18525 @@ -0,0 +1,1009 @@
18526 +/*
18527 + *  linux/kernel/vserver/context.c
18528 + *
18529 + *  Virtual Server: Context Support
18530 + *
18531 + *  Copyright (C) 2003-2007  Herbert Pötzl
18532 + *
18533 + *  V0.01  context helper
18534 + *  V0.02  vx_ctx_kill syscall command
18535 + *  V0.03  replaced context_info calls
18536 + *  V0.04  redesign of struct (de)alloc
18537 + *  V0.05  rlimit basic implementation
18538 + *  V0.06  task_xid and info commands
18539 + *  V0.07  context flags and caps
18540 + *  V0.08  switch to RCU based hash
18541 + *  V0.09  revert to non RCU for now
18542 + *  V0.10  and back to working RCU hash
18543 + *  V0.11  and back to locking again
18544 + *  V0.12  referenced context store
18545 + *  V0.13  separate per cpu data
18546 + *  V0.14  changed vcmds to vxi arg
18547 + *  V0.15  added context stat
18548 + *  V0.16  have __create claim() the vxi
18549 + *  V0.17  removed older and legacy stuff
18550 + *
18551 + */
18552 +
18553 +#include <linux/slab.h>
18554 +#include <linux/types.h>
18555 +#include <linux/security.h>
18556 +#include <linux/pid_namespace.h>
18557 +
18558 +#include <linux/vserver/context.h>
18559 +#include <linux/vserver/network.h>
18560 +#include <linux/vserver/debug.h>
18561 +#include <linux/vserver/limit.h>
18562 +#include <linux/vserver/limit_int.h>
18563 +#include <linux/vserver/space.h>
18564 +
18565 +#include <linux/vs_context.h>
18566 +#include <linux/vs_limit.h>
18567 +#include <linux/vs_pid.h>
18568 +#include <linux/vserver/context_cmd.h>
18569 +
18570 +#include "cvirt_init.h"
18571 +#include "cacct_init.h"
18572 +#include "limit_init.h"
18573 +#include "sched_init.h"
18574 +
18575 +
18576 +atomic_t vx_global_ctotal      = ATOMIC_INIT(0);
18577 +atomic_t vx_global_cactive     = ATOMIC_INIT(0);
18578 +
18579 +
18580 +/*     now inactive context structures */
18581 +
18582 +static struct hlist_head vx_info_inactive = HLIST_HEAD_INIT;
18583 +
18584 +static spinlock_t vx_info_inactive_lock = SPIN_LOCK_UNLOCKED;
18585 +
18586 +
18587 +/*     __alloc_vx_info()
18588 +
18589 +       * allocate an initialized vx_info struct
18590 +       * doesn't make it visible (hash)                        */
18591 +
18592 +static struct vx_info *__alloc_vx_info(xid_t xid)
18593 +{
18594 +       struct vx_info *new = NULL;
18595 +       int cpu;
18596 +
18597 +       vxdprintk(VXD_CBIT(xid, 0), "alloc_vx_info(%d)*", xid);
18598 +
18599 +       /* would this benefit from a slab cache? */
18600 +       new = kmalloc(sizeof(struct vx_info), GFP_KERNEL);
18601 +       if (!new)
18602 +               return 0;
18603 +
18604 +       memset(new, 0, sizeof(struct vx_info));
18605 +#ifdef CONFIG_SMP
18606 +       new->ptr_pc = alloc_percpu(struct _vx_info_pc);
18607 +       if (!new->ptr_pc)
18608 +               goto error;
18609 +#endif
18610 +       new->vx_id = xid;
18611 +       INIT_HLIST_NODE(&new->vx_hlist);
18612 +       atomic_set(&new->vx_usecnt, 0);
18613 +       atomic_set(&new->vx_tasks, 0);
18614 +       new->vx_parent = NULL;
18615 +       new->vx_state = 0;
18616 +       init_waitqueue_head(&new->vx_wait);
18617 +
18618 +       /* prepare reaper */
18619 +       get_task_struct(init_pid_ns.child_reaper);
18620 +       new->vx_reaper = init_pid_ns.child_reaper;
18621 +       new->vx_badness_bias = 0;
18622 +
18623 +       /* rest of init goes here */
18624 +       vx_info_init_limit(&new->limit);
18625 +       vx_info_init_sched(&new->sched);
18626 +       vx_info_init_cvirt(&new->cvirt);
18627 +       vx_info_init_cacct(&new->cacct);
18628 +
18629 +       /* per cpu data structures */
18630 +       for_each_possible_cpu(cpu) {
18631 +               vx_info_init_sched_pc(
18632 +                       &vx_per_cpu(new, sched_pc, cpu), cpu);
18633 +               vx_info_init_cvirt_pc(
18634 +                       &vx_per_cpu(new, cvirt_pc, cpu), cpu);
18635 +       }
18636 +
18637 +       new->vx_flags = VXF_INIT_SET;
18638 +       cap_set_init_eff(new->vx_bcaps);
18639 +       new->vx_ccaps = 0;
18640 +       // new->vx_cap_bset = current->cap_bset;
18641 +
18642 +       new->reboot_cmd = 0;
18643 +       new->exit_code = 0;
18644 +
18645 +       vxdprintk(VXD_CBIT(xid, 0),
18646 +               "alloc_vx_info(%d) = %p", xid, new);
18647 +       vxh_alloc_vx_info(new);
18648 +       atomic_inc(&vx_global_ctotal);
18649 +       return new;
18650 +#ifdef CONFIG_SMP
18651 +error:
18652 +       kfree(new);
18653 +       return 0;
18654 +#endif
18655 +}
18656 +
18657 +/*     __dealloc_vx_info()
18658 +
18659 +       * final disposal of vx_info                             */
18660 +
18661 +static void __dealloc_vx_info(struct vx_info *vxi)
18662 +{
18663 +       struct vx_info_save vxis;
18664 +       int cpu;
18665 +
18666 +       vxdprintk(VXD_CBIT(xid, 0),
18667 +               "dealloc_vx_info(%p)", vxi);
18668 +       vxh_dealloc_vx_info(vxi);
18669 +
18670 +#ifdef CONFIG_VSERVER_WARN
18671 +       enter_vx_info(vxi, &vxis);
18672 +       vx_info_exit_limit(&vxi->limit);
18673 +       vx_info_exit_sched(&vxi->sched);
18674 +       vx_info_exit_cvirt(&vxi->cvirt);
18675 +       vx_info_exit_cacct(&vxi->cacct);
18676 +
18677 +       for_each_possible_cpu(cpu) {
18678 +               vx_info_exit_sched_pc(
18679 +                       &vx_per_cpu(vxi, sched_pc, cpu), cpu);
18680 +               vx_info_exit_cvirt_pc(
18681 +                       &vx_per_cpu(vxi, cvirt_pc, cpu), cpu);
18682 +       }
18683 +       leave_vx_info(&vxis);
18684 +#endif
18685 +
18686 +       vxi->vx_id = -1;
18687 +       vxi->vx_state |= VXS_RELEASED;
18688 +
18689 +#ifdef CONFIG_SMP
18690 +       free_percpu(vxi->ptr_pc);
18691 +#endif
18692 +       kfree(vxi);
18693 +       atomic_dec(&vx_global_ctotal);
18694 +}
18695 +
18696 +static void __shutdown_vx_info(struct vx_info *vxi)
18697 +{
18698 +       struct nsproxy *nsproxy;
18699 +       struct fs_struct *fs;
18700 +
18701 +       might_sleep();
18702 +
18703 +       vxi->vx_state |= VXS_SHUTDOWN;
18704 +       vs_state_change(vxi, VSC_SHUTDOWN);
18705 +
18706 +       nsproxy = xchg(&vxi->vx_nsproxy, NULL);
18707 +       fs = xchg(&vxi->vx_fs, NULL);
18708 +
18709 +       if (nsproxy)
18710 +               put_nsproxy(nsproxy);
18711 +       if (fs)
18712 +               put_fs_struct(fs);
18713 +}
18714 +
18715 +/* exported stuff */
18716 +
18717 +void free_vx_info(struct vx_info *vxi)
18718 +{
18719 +       unsigned long flags;
18720 +
18721 +       /* check for reference counts first */
18722 +       BUG_ON(atomic_read(&vxi->vx_usecnt));
18723 +       BUG_ON(atomic_read(&vxi->vx_tasks));
18724 +
18725 +       /* context must not be hashed */
18726 +       BUG_ON(vx_info_state(vxi, VXS_HASHED));
18727 +
18728 +       /* context shutdown is mandatory */
18729 +       BUG_ON(!vx_info_state(vxi, VXS_SHUTDOWN));
18730 +
18731 +       BUG_ON(vxi->vx_nsproxy);
18732 +       BUG_ON(vxi->vx_fs);
18733 +
18734 +       spin_lock_irqsave(&vx_info_inactive_lock, flags);
18735 +       hlist_del(&vxi->vx_hlist);
18736 +       spin_unlock_irqrestore(&vx_info_inactive_lock, flags);
18737 +
18738 +       __dealloc_vx_info(vxi);
18739 +}
18740 +
18741 +
18742 +/*     hash table for vx_info hash */
18743 +
18744 +#define VX_HASH_SIZE   13
18745 +
18746 +static struct hlist_head vx_info_hash[VX_HASH_SIZE] =
18747 +       { [0 ... VX_HASH_SIZE-1] = HLIST_HEAD_INIT };
18748 +
18749 +static spinlock_t vx_info_hash_lock = SPIN_LOCK_UNLOCKED;
18750 +
18751 +
18752 +static inline unsigned int __hashval(xid_t xid)
18753 +{
18754 +       return (xid % VX_HASH_SIZE);
18755 +}
18756 +
18757 +
18758 +
18759 +/*     __hash_vx_info()
18760 +
18761 +       * add the vxi to the global hash table
18762 +       * requires the hash_lock to be held                     */
18763 +
18764 +static inline void __hash_vx_info(struct vx_info *vxi)
18765 +{
18766 +       struct hlist_head *head;
18767 +
18768 +       vxd_assert_lock(&vx_info_hash_lock);
18769 +       vxdprintk(VXD_CBIT(xid, 4),
18770 +               "__hash_vx_info: %p[#%d]", vxi, vxi->vx_id);
18771 +       vxh_hash_vx_info(vxi);
18772 +
18773 +       /* context must not be hashed */
18774 +       BUG_ON(vx_info_state(vxi, VXS_HASHED));
18775 +
18776 +       vxi->vx_state |= VXS_HASHED;
18777 +       head = &vx_info_hash[__hashval(vxi->vx_id)];
18778 +       hlist_add_head(&vxi->vx_hlist, head);
18779 +       atomic_inc(&vx_global_cactive);
18780 +}
18781 +
18782 +/*     __unhash_vx_info()
18783 +
18784 +       * remove the vxi from the global hash table
18785 +       * requires the hash_lock to be held                     */
18786 +
18787 +static inline void __unhash_vx_info(struct vx_info *vxi)
18788 +{
18789 +       unsigned long flags;
18790 +
18791 +       vxd_assert_lock(&vx_info_hash_lock);
18792 +       vxdprintk(VXD_CBIT(xid, 4),
18793 +               "__unhash_vx_info: %p[#%d.%d.%d]", vxi, vxi->vx_id,
18794 +               atomic_read(&vxi->vx_usecnt), atomic_read(&vxi->vx_tasks));
18795 +       vxh_unhash_vx_info(vxi);
18796 +
18797 +       /* context must be hashed */
18798 +       BUG_ON(!vx_info_state(vxi, VXS_HASHED));
18799 +       /* but without tasks */
18800 +       BUG_ON(atomic_read(&vxi->vx_tasks));
18801 +
18802 +       vxi->vx_state &= ~VXS_HASHED;
18803 +       hlist_del_init(&vxi->vx_hlist);
18804 +       spin_lock_irqsave(&vx_info_inactive_lock, flags);
18805 +       hlist_add_head(&vxi->vx_hlist, &vx_info_inactive);
18806 +       spin_unlock_irqrestore(&vx_info_inactive_lock, flags);
18807 +       atomic_dec(&vx_global_cactive);
18808 +}
18809 +
18810 +
18811 +/*     __lookup_vx_info()
18812 +
18813 +       * requires the hash_lock to be held
18814 +       * doesn't increment the vx_refcnt                       */
18815 +
18816 +static inline struct vx_info *__lookup_vx_info(xid_t xid)
18817 +{
18818 +       struct hlist_head *head = &vx_info_hash[__hashval(xid)];
18819 +       struct hlist_node *pos;
18820 +       struct vx_info *vxi;
18821 +
18822 +       vxd_assert_lock(&vx_info_hash_lock);
18823 +       hlist_for_each(pos, head) {
18824 +               vxi = hlist_entry(pos, struct vx_info, vx_hlist);
18825 +
18826 +               if (vxi->vx_id == xid)
18827 +                       goto found;
18828 +       }
18829 +       vxi = NULL;
18830 +found:
18831 +       vxdprintk(VXD_CBIT(xid, 0),
18832 +               "__lookup_vx_info(#%u): %p[#%u]",
18833 +               xid, vxi, vxi ? vxi->vx_id : 0);
18834 +       vxh_lookup_vx_info(vxi, xid);
18835 +       return vxi;
18836 +}
18837 +
18838 +
18839 +/*     __create_vx_info()
18840 +
18841 +       * create the requested context
18842 +       * get(), claim() and hash it                            */
18843 +
18844 +static struct vx_info *__create_vx_info(int id)
18845 +{
18846 +       struct vx_info *new, *vxi = NULL;
18847 +
18848 +       vxdprintk(VXD_CBIT(xid, 1), "create_vx_info(%d)*", id);
18849 +
18850 +       if (!(new = __alloc_vx_info(id)))
18851 +               return ERR_PTR(-ENOMEM);
18852 +
18853 +       /* required to make dynamic xids unique */
18854 +       spin_lock(&vx_info_hash_lock);
18855 +
18856 +       /* static context requested */
18857 +       if ((vxi = __lookup_vx_info(id))) {
18858 +               vxdprintk(VXD_CBIT(xid, 0),
18859 +                       "create_vx_info(%d) = %p (already there)", id, vxi);
18860 +               if (vx_info_flags(vxi, VXF_STATE_SETUP, 0))
18861 +                       vxi = ERR_PTR(-EBUSY);
18862 +               else
18863 +                       vxi = ERR_PTR(-EEXIST);
18864 +               goto out_unlock;
18865 +       }
18866 +       /* new context */
18867 +       vxdprintk(VXD_CBIT(xid, 0),
18868 +               "create_vx_info(%d) = %p (new)", id, new);
18869 +       claim_vx_info(new, NULL);
18870 +       __hash_vx_info(get_vx_info(new));
18871 +       vxi = new, new = NULL;
18872 +
18873 +out_unlock:
18874 +       spin_unlock(&vx_info_hash_lock);
18875 +       vxh_create_vx_info(IS_ERR(vxi) ? NULL : vxi, id);
18876 +       if (new)
18877 +               __dealloc_vx_info(new);
18878 +       return vxi;
18879 +}
18880 +
18881 +
18882 +/*     exported stuff                                          */
18883 +
18884 +
18885 +void unhash_vx_info(struct vx_info *vxi)
18886 +{
18887 +       __shutdown_vx_info(vxi);
18888 +       spin_lock(&vx_info_hash_lock);
18889 +       __unhash_vx_info(vxi);
18890 +       spin_unlock(&vx_info_hash_lock);
18891 +       __wakeup_vx_info(vxi);
18892 +}
18893 +
18894 +
18895 +/*     lookup_vx_info()
18896 +
18897 +       * search for a vx_info and get() it
18898 +       * negative id means current                             */
18899 +
18900 +struct vx_info *lookup_vx_info(int id)
18901 +{
18902 +       struct vx_info *vxi = NULL;
18903 +
18904 +       if (id < 0) {
18905 +               vxi = get_vx_info(current->vx_info);
18906 +       } else if (id > 1) {
18907 +               spin_lock(&vx_info_hash_lock);
18908 +               vxi = get_vx_info(__lookup_vx_info(id));
18909 +               spin_unlock(&vx_info_hash_lock);
18910 +       }
18911 +       return vxi;
18912 +}
18913 +
18914 +/*     xid_is_hashed()
18915 +
18916 +       * verify that xid is still hashed                       */
18917 +
18918 +int xid_is_hashed(xid_t xid)
18919 +{
18920 +       int hashed;
18921 +
18922 +       spin_lock(&vx_info_hash_lock);
18923 +       hashed = (__lookup_vx_info(xid) != NULL);
18924 +       spin_unlock(&vx_info_hash_lock);
18925 +       return hashed;
18926 +}
18927 +
18928 +#ifdef CONFIG_PROC_FS
18929 +
18930 +/*     get_xid_list()
18931 +
18932 +       * get a subset of hashed xids for proc
18933 +       * assumes size is at least one                          */
18934 +
18935 +int get_xid_list(int index, unsigned int *xids, int size)
18936 +{
18937 +       int hindex, nr_xids = 0;
18938 +
18939 +       /* only show current and children */
18940 +       if (!vx_check(0, VS_ADMIN | VS_WATCH)) {
18941 +               if (index > 0)
18942 +                       return 0;
18943 +               xids[nr_xids] = vx_current_xid();
18944 +               return 1;
18945 +       }
18946 +
18947 +       for (hindex = 0; hindex < VX_HASH_SIZE; hindex++) {
18948 +               struct hlist_head *head = &vx_info_hash[hindex];
18949 +               struct hlist_node *pos;
18950 +
18951 +               spin_lock(&vx_info_hash_lock);
18952 +               hlist_for_each(pos, head) {
18953 +                       struct vx_info *vxi;
18954 +
18955 +                       if (--index > 0)
18956 +                               continue;
18957 +
18958 +                       vxi = hlist_entry(pos, struct vx_info, vx_hlist);
18959 +                       xids[nr_xids] = vxi->vx_id;
18960 +                       if (++nr_xids >= size) {
18961 +                               spin_unlock(&vx_info_hash_lock);
18962 +                               goto out;
18963 +                       }
18964 +               }
18965 +               /* keep the lock time short */
18966 +               spin_unlock(&vx_info_hash_lock);
18967 +       }
18968 +out:
18969 +       return nr_xids;
18970 +}
18971 +#endif
18972 +
18973 +#ifdef CONFIG_VSERVER_DEBUG
18974 +
18975 +void   dump_vx_info_inactive(int level)
18976 +{
18977 +       struct hlist_node *entry, *next;
18978 +
18979 +       hlist_for_each_safe(entry, next, &vx_info_inactive) {
18980 +               struct vx_info *vxi =
18981 +                       list_entry(entry, struct vx_info, vx_hlist);
18982 +
18983 +               dump_vx_info(vxi, level);
18984 +       }
18985 +}
18986 +
18987 +#endif
18988 +
18989 +#if 0
18990 +int vx_migrate_user(struct task_struct *p, struct vx_info *vxi)
18991 +{
18992 +       struct user_struct *new_user, *old_user;
18993 +
18994 +       if (!p || !vxi)
18995 +               BUG();
18996 +
18997 +       if (vx_info_flags(vxi, VXF_INFO_PRIVATE, 0))
18998 +               return -EACCES;
18999 +
19000 +       new_user = alloc_uid(vxi->vx_id, p->uid);
19001 +       if (!new_user)
19002 +               return -ENOMEM;
19003 +
19004 +       old_user = p->user;
19005 +       if (new_user != old_user) {
19006 +               atomic_inc(&new_user->processes);
19007 +               atomic_dec(&old_user->processes);
19008 +               p->user = new_user;
19009 +       }
19010 +       free_uid(old_user);
19011 +       return 0;
19012 +}
19013 +#endif
19014 +
19015 +#if 0
19016 +void vx_mask_cap_bset(struct vx_info *vxi, struct task_struct *p)
19017 +{
19018 +       // p->cap_effective &= vxi->vx_cap_bset;
19019 +       p->cap_effective =
19020 +               cap_intersect(p->cap_effective, vxi->cap_bset);
19021 +       // p->cap_inheritable &= vxi->vx_cap_bset;
19022 +       p->cap_inheritable =
19023 +               cap_intersect(p->cap_inheritable, vxi->cap_bset);
19024 +       // p->cap_permitted &= vxi->vx_cap_bset;
19025 +       p->cap_permitted =
19026 +               cap_intersect(p->cap_permitted, vxi->cap_bset);
19027 +}
19028 +#endif
19029 +
19030 +
19031 +#include <linux/file.h>
19032 +
19033 +static int vx_openfd_task(struct task_struct *tsk)
19034 +{
19035 +       struct files_struct *files = tsk->files;
19036 +       struct fdtable *fdt;
19037 +       const unsigned long *bptr;
19038 +       int count, total;
19039 +
19040 +       /* no rcu_read_lock() because of spin_lock() */
19041 +       spin_lock(&files->file_lock);
19042 +       fdt = files_fdtable(files);
19043 +       bptr = fdt->open_fds->fds_bits;
19044 +       count = fdt->max_fds / (sizeof(unsigned long) * 8);
19045 +       for (total = 0; count > 0; count--) {
19046 +               if (*bptr)
19047 +                       total += hweight_long(*bptr);
19048 +               bptr++;
19049 +       }
19050 +       spin_unlock(&files->file_lock);
19051 +       return total;
19052 +}
19053 +
19054 +
19055 +/*     for *space compatibility */
19056 +
19057 +asmlinkage long sys_unshare(unsigned long);
19058 +
19059 +/*
19060 + *     migrate task to new context
19061 + *     gets vxi, puts old_vxi on change
19062 + *     optionally unshares namespaces (hack)
19063 + */
19064 +
19065 +int vx_migrate_task(struct task_struct *p, struct vx_info *vxi, int unshare)
19066 +{
19067 +       struct vx_info *old_vxi;
19068 +       int ret = 0;
19069 +
19070 +       if (!p || !vxi)
19071 +               BUG();
19072 +
19073 +       vxdprintk(VXD_CBIT(xid, 5),
19074 +               "vx_migrate_task(%p,%p[#%d.%d])", p, vxi,
19075 +               vxi->vx_id, atomic_read(&vxi->vx_usecnt));
19076 +
19077 +       if (vx_info_flags(vxi, VXF_INFO_PRIVATE, 0) &&
19078 +               !vx_info_flags(vxi, VXF_STATE_SETUP, 0))
19079 +               return -EACCES;
19080 +
19081 +       if (vx_info_state(vxi, VXS_SHUTDOWN))
19082 +               return -EFAULT;
19083 +
19084 +       old_vxi = task_get_vx_info(p);
19085 +       if (old_vxi == vxi)
19086 +               goto out;
19087 +
19088 +//     if (!(ret = vx_migrate_user(p, vxi))) {
19089 +       {
19090 +               int openfd;
19091 +
19092 +               task_lock(p);
19093 +               openfd = vx_openfd_task(p);
19094 +
19095 +               if (old_vxi) {
19096 +                       atomic_dec(&old_vxi->cvirt.nr_threads);
19097 +                       atomic_dec(&old_vxi->cvirt.nr_running);
19098 +                       __rlim_dec(&old_vxi->limit, RLIMIT_NPROC);
19099 +                       /* FIXME: what about the struct files here? */
19100 +                       __rlim_sub(&old_vxi->limit, VLIMIT_OPENFD, openfd);
19101 +                       /* account for the executable */
19102 +                       __rlim_dec(&old_vxi->limit, VLIMIT_DENTRY);
19103 +               }
19104 +               atomic_inc(&vxi->cvirt.nr_threads);
19105 +               atomic_inc(&vxi->cvirt.nr_running);
19106 +               __rlim_inc(&vxi->limit, RLIMIT_NPROC);
19107 +               /* FIXME: what about the struct files here? */
19108 +               __rlim_add(&vxi->limit, VLIMIT_OPENFD, openfd);
19109 +               /* account for the executable */
19110 +               __rlim_inc(&vxi->limit, VLIMIT_DENTRY);
19111 +
19112 +               if (old_vxi) {
19113 +                       release_vx_info(old_vxi, p);
19114 +                       clr_vx_info(&p->vx_info);
19115 +               }
19116 +               claim_vx_info(vxi, p);
19117 +               set_vx_info(&p->vx_info, vxi);
19118 +               p->xid = vxi->vx_id;
19119 +
19120 +               vxdprintk(VXD_CBIT(xid, 5),
19121 +                       "moved task %p into vxi:%p[#%d]",
19122 +                       p, vxi, vxi->vx_id);
19123 +
19124 +               // vx_mask_cap_bset(vxi, p);
19125 +               task_unlock(p);
19126 +
19127 +               /* hack for *spaces to provide compatibility */
19128 +               if (unshare) {
19129 +                       struct nsproxy *old_nsp, *new_nsp;
19130 +
19131 +                       ret = unshare_nsproxy_namespaces(
19132 +                               CLONE_NEWUTS | CLONE_NEWIPC | CLONE_NEWUSER,
19133 +                               &new_nsp, NULL);
19134 +                       if (ret)
19135 +                               goto out;
19136 +
19137 +                       old_nsp = xchg(&p->nsproxy, new_nsp);
19138 +                       vx_set_space(vxi, CLONE_NEWUTS | CLONE_NEWIPC | CLONE_NEWUSER);
19139 +                       put_nsproxy(old_nsp);
19140 +               }
19141 +       }
19142 +out:
19143 +       put_vx_info(old_vxi);
19144 +       return ret;
19145 +}
19146 +
19147 +int vx_set_reaper(struct vx_info *vxi, struct task_struct *p)
19148 +{
19149 +       struct task_struct *old_reaper;
19150 +
19151 +       if (!vxi)
19152 +               return -EINVAL;
19153 +
19154 +       vxdprintk(VXD_CBIT(xid, 6),
19155 +               "vx_set_reaper(%p[#%d],%p[#%d,%d])",
19156 +               vxi, vxi->vx_id, p, p->xid, p->pid);
19157 +
19158 +       old_reaper = vxi->vx_reaper;
19159 +       if (old_reaper == p)
19160 +               return 0;
19161 +
19162 +       /* set new child reaper */
19163 +       get_task_struct(p);
19164 +       vxi->vx_reaper = p;
19165 +       put_task_struct(old_reaper);
19166 +       return 0;
19167 +}
19168 +
19169 +int vx_set_init(struct vx_info *vxi, struct task_struct *p)
19170 +{
19171 +       if (!vxi)
19172 +               return -EINVAL;
19173 +
19174 +       vxdprintk(VXD_CBIT(xid, 6),
19175 +               "vx_set_init(%p[#%d],%p[#%d,%d,%d])",
19176 +               vxi, vxi->vx_id, p, p->xid, p->pid, p->tgid);
19177 +
19178 +       vxi->vx_flags &= ~VXF_STATE_INIT;
19179 +       vxi->vx_initpid = p->tgid;
19180 +       return 0;
19181 +}
19182 +
19183 +void vx_exit_init(struct vx_info *vxi, struct task_struct *p, int code)
19184 +{
19185 +       vxdprintk(VXD_CBIT(xid, 6),
19186 +               "vx_exit_init(%p[#%d],%p[#%d,%d,%d])",
19187 +               vxi, vxi->vx_id, p, p->xid, p->pid, p->tgid);
19188 +
19189 +       vxi->exit_code = code;
19190 +       vxi->vx_initpid = 0;
19191 +}
19192 +
19193 +
19194 +void vx_set_persistent(struct vx_info *vxi)
19195 +{
19196 +       vxdprintk(VXD_CBIT(xid, 6),
19197 +               "vx_set_persistent(%p[#%d])", vxi, vxi->vx_id);
19198 +
19199 +       get_vx_info(vxi);
19200 +       claim_vx_info(vxi, NULL);
19201 +}
19202 +
19203 +void vx_clear_persistent(struct vx_info *vxi)
19204 +{
19205 +       vxdprintk(VXD_CBIT(xid, 6),
19206 +               "vx_clear_persistent(%p[#%d])", vxi, vxi->vx_id);
19207 +
19208 +       release_vx_info(vxi, NULL);
19209 +       put_vx_info(vxi);
19210 +}
19211 +
19212 +void vx_update_persistent(struct vx_info *vxi)
19213 +{
19214 +       if (vx_info_flags(vxi, VXF_PERSISTENT, 0))
19215 +               vx_set_persistent(vxi);
19216 +       else
19217 +               vx_clear_persistent(vxi);
19218 +}
19219 +
19220 +
19221 +/*     task must be current or locked          */
19222 +
19223 +void   exit_vx_info(struct task_struct *p, int code)
19224 +{
19225 +       struct vx_info *vxi = p->vx_info;
19226 +
19227 +       if (vxi) {
19228 +               atomic_dec(&vxi->cvirt.nr_threads);
19229 +               vx_nproc_dec(p);
19230 +
19231 +               vxi->exit_code = code;
19232 +               release_vx_info(vxi, p);
19233 +       }
19234 +}
19235 +
19236 +void   exit_vx_info_early(struct task_struct *p, int code)
19237 +{
19238 +       struct vx_info *vxi = p->vx_info;
19239 +
19240 +       if (vxi) {
19241 +               if (vxi->vx_initpid == p->tgid)
19242 +                       vx_exit_init(vxi, p, code);
19243 +               if (vxi->vx_reaper == p)
19244 +                       vx_set_reaper(vxi, init_pid_ns.child_reaper);
19245 +       }
19246 +}
19247 +
19248 +
19249 +/* vserver syscall commands below here */
19250 +
19251 +/* taks xid and vx_info functions */
19252 +
19253 +#include <asm/uaccess.h>
19254 +
19255 +
19256 +int vc_task_xid(uint32_t id)
19257 +{
19258 +       xid_t xid;
19259 +
19260 +       if (id) {
19261 +               struct task_struct *tsk;
19262 +
19263 +               read_lock(&tasklist_lock);
19264 +               tsk = find_task_by_real_pid(id);
19265 +               xid = (tsk) ? tsk->xid : -ESRCH;
19266 +               read_unlock(&tasklist_lock);
19267 +       } else
19268 +               xid = vx_current_xid();
19269 +       return xid;
19270 +}
19271 +
19272 +
19273 +int vc_vx_info(struct vx_info *vxi, void __user *data)
19274 +{
19275 +       struct vcmd_vx_info_v0 vc_data;
19276 +
19277 +       vc_data.xid = vxi->vx_id;
19278 +       vc_data.initpid = vxi->vx_initpid;
19279 +
19280 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19281 +               return -EFAULT;
19282 +       return 0;
19283 +}
19284 +
19285 +
19286 +int vc_ctx_stat(struct vx_info *vxi, void __user *data)
19287 +{
19288 +       struct vcmd_ctx_stat_v0 vc_data;
19289 +
19290 +       vc_data.usecnt = atomic_read(&vxi->vx_usecnt);
19291 +       vc_data.tasks = atomic_read(&vxi->vx_tasks);
19292 +
19293 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19294 +               return -EFAULT;
19295 +       return 0;
19296 +}
19297 +
19298 +
19299 +/* context functions */
19300 +
19301 +int vc_ctx_create(uint32_t xid, void __user *data)
19302 +{
19303 +       struct vcmd_ctx_create vc_data = { .flagword = VXF_INIT_SET };
19304 +       struct vx_info *new_vxi;
19305 +       int ret;
19306 +
19307 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
19308 +               return -EFAULT;
19309 +
19310 +       if ((xid > MAX_S_CONTEXT) || (xid < 2))
19311 +               return -EINVAL;
19312 +
19313 +       new_vxi = __create_vx_info(xid);
19314 +       if (IS_ERR(new_vxi))
19315 +               return PTR_ERR(new_vxi);
19316 +
19317 +       /* initial flags */
19318 +       new_vxi->vx_flags = vc_data.flagword;
19319 +
19320 +       ret = -ENOEXEC;
19321 +       if (vs_state_change(new_vxi, VSC_STARTUP))
19322 +               goto out;
19323 +
19324 +       ret = vx_migrate_task(current, new_vxi, (!data));
19325 +       if (ret)
19326 +               goto out;
19327 +
19328 +       /* return context id on success */
19329 +       ret = new_vxi->vx_id;
19330 +
19331 +       /* get a reference for persistent contexts */
19332 +       if ((vc_data.flagword & VXF_PERSISTENT))
19333 +               vx_set_persistent(new_vxi);
19334 +out:
19335 +       release_vx_info(new_vxi, NULL);
19336 +       put_vx_info(new_vxi);
19337 +       return ret;
19338 +}
19339 +
19340 +
19341 +int vc_ctx_migrate(struct vx_info *vxi, void __user *data)
19342 +{
19343 +       struct vcmd_ctx_migrate vc_data = { .flagword = 0 };
19344 +       int ret;
19345 +
19346 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
19347 +               return -EFAULT;
19348 +
19349 +       ret = vx_migrate_task(current, vxi, 0);
19350 +       if (ret)
19351 +               return ret;
19352 +       if (vc_data.flagword & VXM_SET_INIT)
19353 +               ret = vx_set_init(vxi, current);
19354 +       if (ret)
19355 +               return ret;
19356 +       if (vc_data.flagword & VXM_SET_REAPER)
19357 +               ret = vx_set_reaper(vxi, current);
19358 +       return ret;
19359 +}
19360 +
19361 +
19362 +int vc_get_cflags(struct vx_info *vxi, void __user *data)
19363 +{
19364 +       struct vcmd_ctx_flags_v0 vc_data;
19365 +
19366 +       vc_data.flagword = vxi->vx_flags;
19367 +
19368 +       /* special STATE flag handling */
19369 +       vc_data.mask = vs_mask_flags(~0ULL, vxi->vx_flags, VXF_ONE_TIME);
19370 +
19371 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19372 +               return -EFAULT;
19373 +       return 0;
19374 +}
19375 +
19376 +int vc_set_cflags(struct vx_info *vxi, void __user *data)
19377 +{
19378 +       struct vcmd_ctx_flags_v0 vc_data;
19379 +       uint64_t mask, trigger;
19380 +
19381 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19382 +               return -EFAULT;
19383 +
19384 +       /* special STATE flag handling */
19385 +       mask = vs_mask_mask(vc_data.mask, vxi->vx_flags, VXF_ONE_TIME);
19386 +       trigger = (mask & vxi->vx_flags) ^ (mask & vc_data.flagword);
19387 +
19388 +       if (vxi == current->vx_info) {
19389 +               /* if (trigger & VXF_STATE_SETUP)
19390 +                       vx_mask_cap_bset(vxi, current); */
19391 +               if (trigger & VXF_STATE_INIT) {
19392 +                       int ret;
19393 +
19394 +                       ret = vx_set_init(vxi, current);
19395 +                       if (ret)
19396 +                               return ret;
19397 +                       ret = vx_set_reaper(vxi, current);
19398 +                       if (ret)
19399 +                               return ret;
19400 +               }
19401 +       }
19402 +
19403 +       vxi->vx_flags = vs_mask_flags(vxi->vx_flags,
19404 +               vc_data.flagword, mask);
19405 +       if (trigger & VXF_PERSISTENT)
19406 +               vx_update_persistent(vxi);
19407 +
19408 +       return 0;
19409 +}
19410 +
19411 +
19412 +static inline uint64_t caps_from_cap_t(kernel_cap_t c)
19413 +{
19414 +       uint64_t v = c.cap[0] | ((uint64_t)c.cap[1] << 32);
19415 +
19416 +       // printk("caps_from_cap_t(%08x:%08x) = %016llx\n", c.cap[1], c.cap[0], v);
19417 +       return v;
19418 +}
19419 +
19420 +static inline kernel_cap_t cap_t_from_caps(uint64_t v)
19421 +{
19422 +       kernel_cap_t c = __cap_empty_set;
19423 +
19424 +       c.cap[0] = v & 0xFFFFFFFF;
19425 +       c.cap[1] = (v >> 32) & 0xFFFFFFFF;
19426 +
19427 +       // printk("cap_t_from_caps(%016llx) = %08x:%08x\n", v, c.cap[1], c.cap[0]);
19428 +       return c;
19429 +}
19430 +
19431 +
19432 +static int do_get_caps(struct vx_info *vxi, uint64_t *bcaps, uint64_t *ccaps)
19433 +{
19434 +       if (bcaps)
19435 +               *bcaps = caps_from_cap_t(vxi->vx_bcaps);
19436 +       if (ccaps)
19437 +               *ccaps = vxi->vx_ccaps;
19438 +
19439 +       return 0;
19440 +}
19441 +
19442 +int vc_get_ccaps(struct vx_info *vxi, void __user *data)
19443 +{
19444 +       struct vcmd_ctx_caps_v1 vc_data;
19445 +       int ret;
19446 +
19447 +       ret = do_get_caps(vxi, NULL, &vc_data.ccaps);
19448 +       if (ret)
19449 +               return ret;
19450 +       vc_data.cmask = ~0ULL;
19451 +
19452 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19453 +               return -EFAULT;
19454 +       return 0;
19455 +}
19456 +
19457 +static int do_set_caps(struct vx_info *vxi,
19458 +       uint64_t bcaps, uint64_t bmask, uint64_t ccaps, uint64_t cmask)
19459 +{
19460 +       uint64_t bcold = caps_from_cap_t(vxi->vx_bcaps);
19461 +
19462 +#if 0
19463 +       printk("do_set_caps(%16llx, %16llx, %16llx, %16llx)\n",
19464 +               bcaps, bmask, ccaps, cmask);
19465 +#endif
19466 +       vxi->vx_bcaps = cap_t_from_caps(
19467 +               vs_mask_flags(bcold, bcaps, bmask));
19468 +       vxi->vx_ccaps = vs_mask_flags(vxi->vx_ccaps, ccaps, cmask);
19469 +
19470 +       return 0;
19471 +}
19472 +
19473 +int vc_set_ccaps(struct vx_info *vxi, void __user *data)
19474 +{
19475 +       struct vcmd_ctx_caps_v1 vc_data;
19476 +
19477 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19478 +               return -EFAULT;
19479 +
19480 +       return do_set_caps(vxi, 0, 0, vc_data.ccaps, vc_data.cmask);
19481 +}
19482 +
19483 +int vc_get_bcaps(struct vx_info *vxi, void __user *data)
19484 +{
19485 +       struct vcmd_bcaps vc_data;
19486 +       int ret;
19487 +
19488 +       ret = do_get_caps(vxi, &vc_data.bcaps, NULL);
19489 +       if (ret)
19490 +               return ret;
19491 +       vc_data.bmask = ~0ULL;
19492 +
19493 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19494 +               return -EFAULT;
19495 +       return 0;
19496 +}
19497 +
19498 +int vc_set_bcaps(struct vx_info *vxi, void __user *data)
19499 +{
19500 +       struct vcmd_bcaps vc_data;
19501 +
19502 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19503 +               return -EFAULT;
19504 +
19505 +       return do_set_caps(vxi, vc_data.bcaps, vc_data.bmask, 0, 0);
19506 +}
19507 +
19508 +
19509 +int vc_get_badness(struct vx_info *vxi, void __user *data)
19510 +{
19511 +       struct vcmd_badness_v0 vc_data;
19512 +
19513 +       vc_data.bias = vxi->vx_badness_bias;
19514 +
19515 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19516 +               return -EFAULT;
19517 +       return 0;
19518 +}
19519 +
19520 +int vc_set_badness(struct vx_info *vxi, void __user *data)
19521 +{
19522 +       struct vcmd_badness_v0 vc_data;
19523 +
19524 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19525 +               return -EFAULT;
19526 +
19527 +       vxi->vx_badness_bias = vc_data.bias;
19528 +       return 0;
19529 +}
19530 +
19531 +#include <linux/module.h>
19532 +
19533 +EXPORT_SYMBOL_GPL(free_vx_info);
19534 +
19535 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/cvirt.c linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/cvirt.c
19536 --- linux-2.6.25.11/kernel/vserver/cvirt.c      1969-12-31 19:00:00.000000000 -0500
19537 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/cvirt.c        2008-04-19 15:14:52.000000000 -0400
19538 @@ -0,0 +1,301 @@
19539 +/*
19540 + *  linux/kernel/vserver/cvirt.c
19541 + *
19542 + *  Virtual Server: Context Virtualization
19543 + *
19544 + *  Copyright (C) 2004-2007  Herbert Pötzl
19545 + *
19546 + *  V0.01  broken out from limit.c
19547 + *  V0.02  added utsname stuff
19548 + *  V0.03  changed vcmds to vxi arg
19549 + *
19550 + */
19551 +
19552 +#include <linux/types.h>
19553 +#include <linux/utsname.h>
19554 +#include <linux/vs_cvirt.h>
19555 +#include <linux/vserver/switch.h>
19556 +#include <linux/vserver/cvirt_cmd.h>
19557 +
19558 +#include <asm/uaccess.h>
19559 +
19560 +
19561 +void vx_vsi_uptime(struct timespec *uptime, struct timespec *idle)
19562 +{
19563 +       struct vx_info *vxi = current->vx_info;
19564 +
19565 +       set_normalized_timespec(uptime,
19566 +               uptime->tv_sec - vxi->cvirt.bias_uptime.tv_sec,
19567 +               uptime->tv_nsec - vxi->cvirt.bias_uptime.tv_nsec);
19568 +       if (!idle)
19569 +               return;
19570 +       set_normalized_timespec(idle,
19571 +               idle->tv_sec - vxi->cvirt.bias_idle.tv_sec,
19572 +               idle->tv_nsec - vxi->cvirt.bias_idle.tv_nsec);
19573 +       return;
19574 +}
19575 +
19576 +uint64_t vx_idle_jiffies(void)
19577 +{
19578 +       return init_task.utime + init_task.stime;
19579 +}
19580 +
19581 +
19582 +
19583 +static inline uint32_t __update_loadavg(uint32_t load,
19584 +       int wsize, int delta, int n)
19585 +{
19586 +       unsigned long long calc, prev;
19587 +
19588 +       /* just set it to n */
19589 +       if (unlikely(delta >= wsize))
19590 +               return (n << FSHIFT);
19591 +
19592 +       calc = delta * n;
19593 +       calc <<= FSHIFT;
19594 +       prev = (wsize - delta);
19595 +       prev *= load;
19596 +       calc += prev;
19597 +       do_div(calc, wsize);
19598 +       return calc;
19599 +}
19600 +
19601 +
19602 +void vx_update_load(struct vx_info *vxi)
19603 +{
19604 +       uint32_t now, last, delta;
19605 +       unsigned int nr_running, nr_uninterruptible;
19606 +       unsigned int total;
19607 +       unsigned long flags;
19608 +
19609 +       spin_lock_irqsave(&vxi->cvirt.load_lock, flags);
19610 +
19611 +       now = jiffies;
19612 +       last = vxi->cvirt.load_last;
19613 +       delta = now - last;
19614 +
19615 +       if (delta < 5*HZ)
19616 +               goto out;
19617 +
19618 +       nr_running = atomic_read(&vxi->cvirt.nr_running);
19619 +       nr_uninterruptible = atomic_read(&vxi->cvirt.nr_uninterruptible);
19620 +       total = nr_running + nr_uninterruptible;
19621 +
19622 +       vxi->cvirt.load[0] = __update_loadavg(vxi->cvirt.load[0],
19623 +               60*HZ, delta, total);
19624 +       vxi->cvirt.load[1] = __update_loadavg(vxi->cvirt.load[1],
19625 +               5*60*HZ, delta, total);
19626 +       vxi->cvirt.load[2] = __update_loadavg(vxi->cvirt.load[2],
19627 +               15*60*HZ, delta, total);
19628 +
19629 +       vxi->cvirt.load_last = now;
19630 +out:
19631 +       atomic_inc(&vxi->cvirt.load_updates);
19632 +       spin_unlock_irqrestore(&vxi->cvirt.load_lock, flags);
19633 +}
19634 +
19635 +
19636 +/*
19637 + * Commands to do_syslog:
19638 + *
19639 + *      0 -- Close the log.  Currently a NOP.
19640 + *      1 -- Open the log. Currently a NOP.
19641 + *      2 -- Read from the log.
19642 + *      3 -- Read all messages remaining in the ring buffer.
19643 + *      4 -- Read and clear all messages remaining in the ring buffer
19644 + *      5 -- Clear ring buffer.
19645 + *      6 -- Disable printk's to console
19646 + *      7 -- Enable printk's to console
19647 + *      8 -- Set level of messages printed to console
19648 + *      9 -- Return number of unread characters in the log buffer
19649 + *     10 -- Return size of the log buffer
19650 + */
19651 +int vx_do_syslog(int type, char __user *buf, int len)
19652 +{
19653 +       int error = 0;
19654 +       int do_clear = 0;
19655 +       struct vx_info *vxi = current->vx_info;
19656 +       struct _vx_syslog *log;
19657 +
19658 +       if (!vxi)
19659 +               return -EINVAL;
19660 +       log = &vxi->cvirt.syslog;
19661 +
19662 +       switch (type) {
19663 +       case 0:         /* Close log */
19664 +       case 1:         /* Open log */
19665 +               break;
19666 +       case 2:         /* Read from log */
19667 +               error = wait_event_interruptible(log->log_wait,
19668 +                       (log->log_start - log->log_end));
19669 +               if (error)
19670 +                       break;
19671 +               spin_lock_irq(&log->logbuf_lock);
19672 +               spin_unlock_irq(&log->logbuf_lock);
19673 +               break;
19674 +       case 4:         /* Read/clear last kernel messages */
19675 +               do_clear = 1;
19676 +               /* fall through */
19677 +       case 3:         /* Read last kernel messages */
19678 +               return 0;
19679 +
19680 +       case 5:         /* Clear ring buffer */
19681 +               return 0;
19682 +
19683 +       case 6:         /* Disable logging to console */
19684 +       case 7:         /* Enable logging to console */
19685 +       case 8:         /* Set level of messages printed to console */
19686 +               break;
19687 +
19688 +       case 9:         /* Number of chars in the log buffer */
19689 +               return 0;
19690 +       case 10:        /* Size of the log buffer */
19691 +               return 0;
19692 +       default:
19693 +               error = -EINVAL;
19694 +               break;
19695 +       }
19696 +       return error;
19697 +}
19698 +
19699 +
19700 +/* virtual host info names */
19701 +
19702 +static char *vx_vhi_name(struct vx_info *vxi, int id)
19703 +{
19704 +       struct nsproxy *nsproxy;
19705 +       struct uts_namespace *uts;
19706 +
19707 +
19708 +       if (id == VHIN_CONTEXT)
19709 +               return vxi->vx_name;
19710 +
19711 +       nsproxy = vxi->vx_nsproxy;
19712 +       if (!nsproxy)
19713 +               return NULL;
19714 +
19715 +       uts = nsproxy->uts_ns;
19716 +       if (!uts)
19717 +               return NULL;
19718 +
19719 +       switch (id) {
19720 +       case VHIN_SYSNAME:
19721 +               return uts->name.sysname;
19722 +       case VHIN_NODENAME:
19723 +               return uts->name.nodename;
19724 +       case VHIN_RELEASE:
19725 +               return uts->name.release;
19726 +       case VHIN_VERSION:
19727 +               return uts->name.version;
19728 +       case VHIN_MACHINE:
19729 +               return uts->name.machine;
19730 +       case VHIN_DOMAINNAME:
19731 +               return uts->name.domainname;
19732 +       default:
19733 +               return NULL;
19734 +       }
19735 +       return NULL;
19736 +}
19737 +
19738 +int vc_set_vhi_name(struct vx_info *vxi, void __user *data)
19739 +{
19740 +       struct vcmd_vhi_name_v0 vc_data;
19741 +       char *name;
19742 +
19743 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19744 +               return -EFAULT;
19745 +
19746 +       name = vx_vhi_name(vxi, vc_data.field);
19747 +       if (!name)
19748 +               return -EINVAL;
19749 +
19750 +       memcpy(name, vc_data.name, 65);
19751 +       return 0;
19752 +}
19753 +
19754 +int vc_get_vhi_name(struct vx_info *vxi, void __user *data)
19755 +{
19756 +       struct vcmd_vhi_name_v0 vc_data;
19757 +       char *name;
19758 +
19759 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19760 +               return -EFAULT;
19761 +
19762 +       name = vx_vhi_name(vxi, vc_data.field);
19763 +       if (!name)
19764 +               return -EINVAL;
19765 +
19766 +       memcpy(vc_data.name, name, 65);
19767 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19768 +               return -EFAULT;
19769 +       return 0;
19770 +}
19771 +
19772 +
19773 +int vc_virt_stat(struct vx_info *vxi, void __user *data)
19774 +{
19775 +       struct vcmd_virt_stat_v0 vc_data;
19776 +       struct _vx_cvirt *cvirt = &vxi->cvirt;
19777 +       struct timespec uptime;
19778 +
19779 +       do_posix_clock_monotonic_gettime(&uptime);
19780 +       set_normalized_timespec(&uptime,
19781 +               uptime.tv_sec - cvirt->bias_uptime.tv_sec,
19782 +               uptime.tv_nsec - cvirt->bias_uptime.tv_nsec);
19783 +
19784 +       vc_data.offset = timeval_to_ns(&cvirt->bias_tv);
19785 +       vc_data.uptime = timespec_to_ns(&uptime);
19786 +       vc_data.nr_threads = atomic_read(&cvirt->nr_threads);
19787 +       vc_data.nr_running = atomic_read(&cvirt->nr_running);
19788 +       vc_data.nr_uninterruptible = atomic_read(&cvirt->nr_uninterruptible);
19789 +       vc_data.nr_onhold = atomic_read(&cvirt->nr_onhold);
19790 +       vc_data.nr_forks = atomic_read(&cvirt->total_forks);
19791 +       vc_data.load[0] = cvirt->load[0];
19792 +       vc_data.load[1] = cvirt->load[1];
19793 +       vc_data.load[2] = cvirt->load[2];
19794 +
19795 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19796 +               return -EFAULT;
19797 +       return 0;
19798 +}
19799 +
19800 +
19801 +#ifdef CONFIG_VSERVER_VTIME
19802 +
19803 +/* virtualized time base */
19804 +
19805 +void vx_gettimeofday(struct timeval *tv)
19806 +{
19807 +       do_gettimeofday(tv);
19808 +       if (!vx_flags(VXF_VIRT_TIME, 0))
19809 +               return;
19810 +
19811 +       tv->tv_sec += current->vx_info->cvirt.bias_tv.tv_sec;
19812 +       tv->tv_usec += current->vx_info->cvirt.bias_tv.tv_usec;
19813 +
19814 +       if (tv->tv_usec >= USEC_PER_SEC) {
19815 +               tv->tv_sec++;
19816 +               tv->tv_usec -= USEC_PER_SEC;
19817 +       } else if (tv->tv_usec < 0) {
19818 +               tv->tv_sec--;
19819 +               tv->tv_usec += USEC_PER_SEC;
19820 +       }
19821 +}
19822 +
19823 +int vx_settimeofday(struct timespec *ts)
19824 +{
19825 +       struct timeval tv;
19826 +
19827 +       if (!vx_flags(VXF_VIRT_TIME, 0))
19828 +               return do_settimeofday(ts);
19829 +
19830 +       do_gettimeofday(&tv);
19831 +       current->vx_info->cvirt.bias_tv.tv_sec =
19832 +               ts->tv_sec - tv.tv_sec;
19833 +       current->vx_info->cvirt.bias_tv.tv_usec =
19834 +               (ts->tv_nsec/NSEC_PER_USEC) - tv.tv_usec;
19835 +       return 0;
19836 +}
19837 +
19838 +#endif
19839 +
19840 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/cvirt_init.h linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/cvirt_init.h
19841 --- linux-2.6.25.11/kernel/vserver/cvirt_init.h 1969-12-31 19:00:00.000000000 -0500
19842 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/cvirt_init.h   2008-04-19 15:14:52.000000000 -0400
19843 @@ -0,0 +1,69 @@
19844 +
19845 +
19846 +extern uint64_t vx_idle_jiffies(void);
19847 +
19848 +static inline void vx_info_init_cvirt(struct _vx_cvirt *cvirt)
19849 +{
19850 +       uint64_t idle_jiffies = vx_idle_jiffies();
19851 +       uint64_t nsuptime;
19852 +
19853 +       do_posix_clock_monotonic_gettime(&cvirt->bias_uptime);
19854 +       nsuptime = (unsigned long long)cvirt->bias_uptime.tv_sec
19855 +               * NSEC_PER_SEC + cvirt->bias_uptime.tv_nsec;
19856 +       cvirt->bias_clock = nsec_to_clock_t(nsuptime);
19857 +       cvirt->bias_tv.tv_sec = 0;
19858 +       cvirt->bias_tv.tv_usec = 0;
19859 +
19860 +       jiffies_to_timespec(idle_jiffies, &cvirt->bias_idle);
19861 +       atomic_set(&cvirt->nr_threads, 0);
19862 +       atomic_set(&cvirt->nr_running, 0);
19863 +       atomic_set(&cvirt->nr_uninterruptible, 0);
19864 +       atomic_set(&cvirt->nr_onhold, 0);
19865 +
19866 +       spin_lock_init(&cvirt->load_lock);
19867 +       cvirt->load_last = jiffies;
19868 +       atomic_set(&cvirt->load_updates, 0);
19869 +       cvirt->load[0] = 0;
19870 +       cvirt->load[1] = 0;
19871 +       cvirt->load[2] = 0;
19872 +       atomic_set(&cvirt->total_forks, 0);
19873 +
19874 +       spin_lock_init(&cvirt->syslog.logbuf_lock);
19875 +       init_waitqueue_head(&cvirt->syslog.log_wait);
19876 +       cvirt->syslog.log_start = 0;
19877 +       cvirt->syslog.log_end = 0;
19878 +       cvirt->syslog.con_start = 0;
19879 +       cvirt->syslog.logged_chars = 0;
19880 +}
19881 +
19882 +static inline
19883 +void vx_info_init_cvirt_pc(struct _vx_cvirt_pc *cvirt_pc, int cpu)
19884 +{
19885 +       // cvirt_pc->cpustat = { 0 };
19886 +}
19887 +
19888 +static inline void vx_info_exit_cvirt(struct _vx_cvirt *cvirt)
19889 +{
19890 +       int value;
19891 +
19892 +       vxwprintk_xid((value = atomic_read(&cvirt->nr_threads)),
19893 +               "!!! cvirt: %p[nr_threads] = %d on exit.",
19894 +               cvirt, value);
19895 +       vxwprintk_xid((value = atomic_read(&cvirt->nr_running)),
19896 +               "!!! cvirt: %p[nr_running] = %d on exit.",
19897 +               cvirt, value);
19898 +       vxwprintk_xid((value = atomic_read(&cvirt->nr_uninterruptible)),
19899 +               "!!! cvirt: %p[nr_uninterruptible] = %d on exit.",
19900 +               cvirt, value);
19901 +       vxwprintk_xid((value = atomic_read(&cvirt->nr_onhold)),
19902 +               "!!! cvirt: %p[nr_onhold] = %d on exit.",
19903 +               cvirt, value);
19904 +       return;
19905 +}
19906 +
19907 +static inline
19908 +void vx_info_exit_cvirt_pc(struct _vx_cvirt_pc *cvirt_pc, int cpu)
19909 +{
19910 +       return;
19911 +}
19912 +
19913 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/cvirt_proc.h linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/cvirt_proc.h
19914 --- linux-2.6.25.11/kernel/vserver/cvirt_proc.h 1969-12-31 19:00:00.000000000 -0500
19915 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/cvirt_proc.h   2008-04-21 13:01:29.000000000 -0400
19916 @@ -0,0 +1,135 @@
19917 +#ifndef _VX_CVIRT_PROC_H
19918 +#define _VX_CVIRT_PROC_H
19919 +
19920 +#include <linux/nsproxy.h>
19921 +#include <linux/mnt_namespace.h>
19922 +#include <linux/ipc_namespace.h>
19923 +#include <linux/utsname.h>
19924 +#include <linux/ipc.h>
19925 +
19926 +
19927 +static inline
19928 +int vx_info_proc_nsproxy(struct nsproxy *nsproxy, char *buffer)
19929 +{
19930 +       struct mnt_namespace *ns;
19931 +       struct uts_namespace *uts;
19932 +       struct ipc_namespace *ipc;
19933 +       struct path path;
19934 +       char *pstr, *root;
19935 +       int length = 0;
19936 +
19937 +       if (!nsproxy)
19938 +               goto out;
19939 +
19940 +       length += sprintf(buffer + length,
19941 +               "NSProxy:\t%p [%p,%p,%p]\n",
19942 +               nsproxy, nsproxy->mnt_ns,
19943 +               nsproxy->uts_ns, nsproxy->ipc_ns);
19944 +
19945 +       ns = nsproxy->mnt_ns;
19946 +       if (!ns)
19947 +               goto skip_ns;
19948 +
19949 +       pstr = kmalloc(PATH_MAX, GFP_KERNEL);
19950 +       if (!pstr)
19951 +               goto skip_ns;
19952 +
19953 +       path.mnt = ns->root;
19954 +       path.dentry = ns->root->mnt_root;
19955 +       root = d_path(&path, pstr, PATH_MAX - 2);
19956 +       length += sprintf(buffer + length,
19957 +               "Namespace:\t%p [#%u]\n"
19958 +               "RootPath:\t%s\n",
19959 +               ns, atomic_read(&ns->count),
19960 +               root);
19961 +       kfree(pstr);
19962 +skip_ns:
19963 +
19964 +       uts = nsproxy->uts_ns;
19965 +       if (!uts)
19966 +               goto skip_uts;
19967 +
19968 +       length += sprintf(buffer + length,
19969 +               "SysName:\t%.*s\n"
19970 +               "NodeName:\t%.*s\n"
19971 +               "Release:\t%.*s\n"
19972 +               "Version:\t%.*s\n"
19973 +               "Machine:\t%.*s\n"
19974 +               "DomainName:\t%.*s\n",
19975 +               __NEW_UTS_LEN, uts->name.sysname,
19976 +               __NEW_UTS_LEN, uts->name.nodename,
19977 +               __NEW_UTS_LEN, uts->name.release,
19978 +               __NEW_UTS_LEN, uts->name.version,
19979 +               __NEW_UTS_LEN, uts->name.machine,
19980 +               __NEW_UTS_LEN, uts->name.domainname);
19981 +skip_uts:
19982 +
19983 +       ipc = nsproxy->ipc_ns;
19984 +       if (!ipc)
19985 +               goto skip_ipc;
19986 +
19987 +       length += sprintf(buffer + length,
19988 +               "SEMS:\t\t%d %d %d %d  %d\n"
19989 +               "MSG:\t\t%d %d %d\n"
19990 +               "SHM:\t\t%lu %lu  %d %d\n",
19991 +               ipc->sem_ctls[0], ipc->sem_ctls[1],
19992 +               ipc->sem_ctls[2], ipc->sem_ctls[3],
19993 +               ipc->used_sems,
19994 +               ipc->msg_ctlmax, ipc->msg_ctlmnb, ipc->msg_ctlmni,
19995 +               (unsigned long)ipc->shm_ctlmax,
19996 +               (unsigned long)ipc->shm_ctlall,
19997 +               ipc->shm_ctlmni, ipc->shm_tot);
19998 +skip_ipc:
19999 +out:
20000 +       return length;
20001 +}
20002 +
20003 +
20004 +#include <linux/sched.h>
20005 +
20006 +#define LOAD_INT(x) ((x) >> FSHIFT)
20007 +#define LOAD_FRAC(x) LOAD_INT(((x) & (FIXED_1 - 1)) * 100)
20008 +
20009 +static inline
20010 +int vx_info_proc_cvirt(struct _vx_cvirt *cvirt, char *buffer)
20011 +{
20012 +       int length = 0;
20013 +       int a, b, c;
20014 +
20015 +       length += sprintf(buffer + length,
20016 +               "BiasUptime:\t%lu.%02lu\n",
20017 +               (unsigned long)cvirt->bias_uptime.tv_sec,
20018 +               (cvirt->bias_uptime.tv_nsec / (NSEC_PER_SEC / 100)));
20019 +
20020 +       a = cvirt->load[0] + (FIXED_1 / 200);
20021 +       b = cvirt->load[1] + (FIXED_1 / 200);
20022 +       c = cvirt->load[2] + (FIXED_1 / 200);
20023 +       length += sprintf(buffer + length,
20024 +               "nr_threads:\t%d\n"
20025 +               "nr_running:\t%d\n"
20026 +               "nr_unintr:\t%d\n"
20027 +               "nr_onhold:\t%d\n"
20028 +               "load_updates:\t%d\n"
20029 +               "loadavg:\t%d.%02d %d.%02d %d.%02d\n"
20030 +               "total_forks:\t%d\n",
20031 +               atomic_read(&cvirt->nr_threads),
20032 +               atomic_read(&cvirt->nr_running),
20033 +               atomic_read(&cvirt->nr_uninterruptible),
20034 +               atomic_read(&cvirt->nr_onhold),
20035 +               atomic_read(&cvirt->load_updates),
20036 +               LOAD_INT(a), LOAD_FRAC(a),
20037 +               LOAD_INT(b), LOAD_FRAC(b),
20038 +               LOAD_INT(c), LOAD_FRAC(c),
20039 +               atomic_read(&cvirt->total_forks));
20040 +       return length;
20041 +}
20042 +
20043 +static inline
20044 +int vx_info_proc_cvirt_pc(struct _vx_cvirt_pc *cvirt_pc,
20045 +       char *buffer, int cpu)
20046 +{
20047 +       int length = 0;
20048 +       return length;
20049 +}
20050 +
20051 +#endif /* _VX_CVIRT_PROC_H */
20052 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/debug.c linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/debug.c
20053 --- linux-2.6.25.11/kernel/vserver/debug.c      1969-12-31 19:00:00.000000000 -0500
20054 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/debug.c        2008-04-19 15:14:52.000000000 -0400
20055 @@ -0,0 +1,32 @@
20056 +/*
20057 + *  kernel/vserver/debug.c
20058 + *
20059 + *  Copyright (C) 2005-2007 Herbert Pötzl
20060 + *
20061 + *  V0.01  vx_info dump support
20062 + *
20063 + */
20064 +
20065 +#include <linux/module.h>
20066 +
20067 +#include <linux/vserver/context.h>
20068 +
20069 +
20070 +void   dump_vx_info(struct vx_info *vxi, int level)
20071 +{
20072 +       printk("vx_info %p[#%d, %d.%d, %4x]\n", vxi, vxi->vx_id,
20073 +               atomic_read(&vxi->vx_usecnt),
20074 +               atomic_read(&vxi->vx_tasks),
20075 +               vxi->vx_state);
20076 +       if (level > 0) {
20077 +               __dump_vx_limit(&vxi->limit);
20078 +               __dump_vx_sched(&vxi->sched);
20079 +               __dump_vx_cvirt(&vxi->cvirt);
20080 +               __dump_vx_cacct(&vxi->cacct);
20081 +       }
20082 +       printk("---\n");
20083 +}
20084 +
20085 +
20086 +EXPORT_SYMBOL_GPL(dump_vx_info);
20087 +
20088 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/device.c linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/device.c
20089 --- linux-2.6.25.11/kernel/vserver/device.c     1969-12-31 19:00:00.000000000 -0500
20090 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/device.c       2008-04-21 12:35:24.000000000 -0400
20091 @@ -0,0 +1,443 @@
20092 +/*
20093 + *  linux/kernel/vserver/device.c
20094 + *
20095 + *  Linux-VServer: Device Support
20096 + *
20097 + *  Copyright (C) 2006  Herbert Pötzl
20098 + *  Copyright (C) 2007  Daniel Hokka Zakrisson
20099 + *
20100 + *  V0.01  device mapping basics
20101 + *  V0.02  added defaults
20102 + *
20103 + */
20104 +
20105 +#include <linux/slab.h>
20106 +#include <linux/rcupdate.h>
20107 +#include <linux/fs.h>
20108 +#include <linux/namei.h>
20109 +#include <linux/hash.h>
20110 +
20111 +#include <asm/errno.h>
20112 +#include <asm/uaccess.h>
20113 +#include <linux/vserver/base.h>
20114 +#include <linux/vserver/debug.h>
20115 +#include <linux/vserver/context.h>
20116 +#include <linux/vserver/device.h>
20117 +#include <linux/vserver/device_cmd.h>
20118 +
20119 +
20120 +#define DMAP_HASH_BITS 4
20121 +
20122 +
20123 +struct vs_mapping {
20124 +       union {
20125 +               struct hlist_node hlist;
20126 +               struct list_head list;
20127 +       } u;
20128 +#define dm_hlist       u.hlist
20129 +#define dm_list                u.list
20130 +       xid_t xid;
20131 +       dev_t device;
20132 +       struct vx_dmap_target target;
20133 +};
20134 +
20135 +
20136 +static struct hlist_head dmap_main_hash[1 << DMAP_HASH_BITS];
20137 +
20138 +static spinlock_t dmap_main_hash_lock = SPIN_LOCK_UNLOCKED;
20139 +
20140 +static struct vx_dmap_target dmap_defaults[2] = {
20141 +       { .flags = DATTR_OPEN },
20142 +       { .flags = DATTR_OPEN },
20143 +};
20144 +
20145 +
20146 +struct kmem_cache *dmap_cachep __read_mostly;
20147 +
20148 +int __init dmap_cache_init(void)
20149 +{
20150 +       dmap_cachep = kmem_cache_create("dmap_cache",
20151 +               sizeof(struct vs_mapping), 0,
20152 +               SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
20153 +       return 0;
20154 +}
20155 +
20156 +__initcall(dmap_cache_init);
20157 +
20158 +
20159 +static inline unsigned int __hashval(dev_t dev, int bits)
20160 +{
20161 +       return hash_long((unsigned long)dev, bits);
20162 +}
20163 +
20164 +
20165 +/*     __hash_mapping()
20166 + *     add the mapping to the hash table
20167 + */
20168 +static inline void __hash_mapping(struct vx_info *vxi, struct vs_mapping *vdm)
20169 +{
20170 +       spinlock_t *hash_lock = &dmap_main_hash_lock;
20171 +       struct hlist_head *head, *hash = dmap_main_hash;
20172 +       int device = vdm->device;
20173 +
20174 +       spin_lock(hash_lock);
20175 +       vxdprintk(VXD_CBIT(misc, 8), "__hash_mapping: %p[#%d] %08x:%08x",
20176 +               vxi, vxi ? vxi->vx_id : 0, device, vdm->target.target);
20177 +
20178 +       head = &hash[__hashval(device, DMAP_HASH_BITS)];
20179 +       hlist_add_head(&vdm->dm_hlist, head);
20180 +       spin_unlock(hash_lock);
20181 +}
20182 +
20183 +
20184 +static inline int __mode_to_default(umode_t mode)
20185 +{
20186 +       switch (mode) {
20187 +       case S_IFBLK:
20188 +               return 0;
20189 +       case S_IFCHR:
20190 +               return 1;
20191 +       default:
20192 +               BUG();
20193 +       }
20194 +}
20195 +
20196 +
20197 +/*     __set_default()
20198 + *     set a default
20199 + */
20200 +static inline void __set_default(struct vx_info *vxi, umode_t mode,
20201 +       struct vx_dmap_target *vdmt)
20202 +{
20203 +       spinlock_t *hash_lock = &dmap_main_hash_lock;
20204 +       spin_lock(hash_lock);
20205 +
20206 +       if (vxi)
20207 +               vxi->dmap.targets[__mode_to_default(mode)] = *vdmt;
20208 +       else
20209 +               dmap_defaults[__mode_to_default(mode)] = *vdmt;
20210 +
20211 +
20212 +       spin_unlock(hash_lock);
20213 +
20214 +       vxdprintk(VXD_CBIT(misc, 8), "__set_default: %p[#%u] %08x %04x",
20215 +                 vxi, vxi ? vxi->vx_id : 0, vdmt->target, vdmt->flags);
20216 +}
20217 +
20218 +
20219 +/*     __remove_default()
20220 + *     remove a default
20221 + */
20222 +static inline int __remove_default(struct vx_info *vxi, umode_t mode)
20223 +{
20224 +       spinlock_t *hash_lock = &dmap_main_hash_lock;
20225 +       spin_lock(hash_lock);
20226 +
20227 +       if (vxi)
20228 +               vxi->dmap.targets[__mode_to_default(mode)].flags = 0;
20229 +       else    /* remove == reset */
20230 +               dmap_defaults[__mode_to_default(mode)].flags = DATTR_OPEN | mode;
20231 +
20232 +       spin_unlock(hash_lock);
20233 +       return 0;
20234 +}
20235 +
20236 +
20237 +/*     __find_mapping()
20238 + *     find a mapping in the hash table
20239 + *
20240 + *     caller must hold hash_lock
20241 + */
20242 +static inline int __find_mapping(xid_t xid, dev_t device, umode_t mode,
20243 +       struct vs_mapping **local, struct vs_mapping **global)
20244 +{
20245 +       struct hlist_head *hash = dmap_main_hash;
20246 +       struct hlist_head *head = &hash[__hashval(device, DMAP_HASH_BITS)];
20247 +       struct hlist_node *pos;
20248 +       struct vs_mapping *vdm;
20249 +
20250 +       *local = NULL;
20251 +       if (global)
20252 +               *global = NULL;
20253 +
20254 +       hlist_for_each(pos, head) {
20255 +               vdm = hlist_entry(pos, struct vs_mapping, dm_hlist);
20256 +
20257 +               if ((vdm->device == device) &&
20258 +                       !((vdm->target.flags ^ mode) & S_IFMT)) {
20259 +                       if (vdm->xid == xid) {
20260 +                               *local = vdm;
20261 +                               return 1;
20262 +                       } else if (global && vdm->xid == 0)
20263 +                               *global = vdm;
20264 +               }
20265 +       }
20266 +
20267 +       if (global && *global)
20268 +               return 0;
20269 +       else
20270 +               return -ENOENT;
20271 +}
20272 +
20273 +
20274 +/*     __lookup_mapping()
20275 + *     find a mapping and store the result in target and flags
20276 + */
20277 +static inline int __lookup_mapping(struct vx_info *vxi,
20278 +       dev_t device, dev_t *target, int *flags, umode_t mode)
20279 +{
20280 +       spinlock_t *hash_lock = &dmap_main_hash_lock;
20281 +       struct vs_mapping *vdm, *global;
20282 +       struct vx_dmap_target *vdmt;
20283 +       int ret = 0;
20284 +       xid_t xid = vxi->vx_id;
20285 +       int index;
20286 +
20287 +       spin_lock(hash_lock);
20288 +       if (__find_mapping(xid, device, mode, &vdm, &global) > 0) {
20289 +               ret = 1;
20290 +               vdmt = &vdm->target;
20291 +               goto found;
20292 +       }
20293 +
20294 +       index = __mode_to_default(mode);
20295 +       if (vxi && vxi->dmap.targets[index].flags) {
20296 +               ret = 2;
20297 +               vdmt = &vxi->dmap.targets[index];
20298 +       } else if (global) {
20299 +               ret = 3;
20300 +               vdmt = &global->target;
20301 +               goto found;
20302 +       } else {
20303 +               ret = 4;
20304 +               vdmt = &dmap_defaults[index];
20305 +       }
20306 +
20307 +found:
20308 +       if (target && (vdmt->flags & DATTR_REMAP))
20309 +               *target = vdmt->target;
20310 +       else if (target)
20311 +               *target = device;
20312 +       if (flags)
20313 +               *flags = vdmt->flags;
20314 +
20315 +       spin_unlock(hash_lock);
20316 +
20317 +       return ret;
20318 +}
20319 +
20320 +
20321 +/*     __remove_mapping()
20322 + *     remove a mapping from the hash table
20323 + */
20324 +static inline int __remove_mapping(struct vx_info *vxi, dev_t device,
20325 +       umode_t mode)
20326 +{
20327 +       spinlock_t *hash_lock = &dmap_main_hash_lock;
20328 +       struct vs_mapping *vdm = NULL;
20329 +       int ret = 0;
20330 +
20331 +       spin_lock(hash_lock);
20332 +
20333 +       ret = __find_mapping((vxi ? vxi->vx_id : 0), device, mode, &vdm,
20334 +               NULL);
20335 +       vxdprintk(VXD_CBIT(misc, 8), "__remove_mapping: %p[#%d] %08x %04x",
20336 +               vxi, vxi ? vxi->vx_id : 0, device, mode);
20337 +       if (ret < 0)
20338 +               goto out;
20339 +       hlist_del(&vdm->dm_hlist);
20340 +
20341 +out:
20342 +       spin_unlock(hash_lock);
20343 +       if (vdm)
20344 +               kmem_cache_free(dmap_cachep, vdm);
20345 +       return ret;
20346 +}
20347 +
20348 +
20349 +
20350 +int vs_map_device(struct vx_info *vxi,
20351 +       dev_t device, dev_t *target, umode_t mode)
20352 +{
20353 +       int ret, flags = DATTR_MASK;
20354 +
20355 +       if (!vxi) {
20356 +               if (target)
20357 +                       *target = device;
20358 +               goto out;
20359 +       }
20360 +       ret = __lookup_mapping(vxi, device, target, &flags, mode);
20361 +       vxdprintk(VXD_CBIT(misc, 8), "vs_map_device: %08x target: %08x flags: %04x mode: %04x mapped=%d",
20362 +               device, target ? *target : 0, flags, mode, ret);
20363 +out:
20364 +       return (flags & DATTR_MASK);
20365 +}
20366 +
20367 +
20368 +
20369 +static int do_set_mapping(struct vx_info *vxi,
20370 +       dev_t device, dev_t target, int flags, umode_t mode)
20371 +{
20372 +       if (device) {
20373 +               struct vs_mapping *new;
20374 +
20375 +               new = kmem_cache_alloc(dmap_cachep, GFP_KERNEL);
20376 +               if (!new)
20377 +                       return -ENOMEM;
20378 +
20379 +               INIT_HLIST_NODE(&new->dm_hlist);
20380 +               new->device = device;
20381 +               new->target.target = target;
20382 +               new->target.flags = flags | mode;
20383 +               new->xid = (vxi ? vxi->vx_id : 0);
20384 +
20385 +               vxdprintk(VXD_CBIT(misc, 8), "do_set_mapping: %08x target: %08x flags: %04x", device, target, flags);
20386 +               __hash_mapping(vxi, new);
20387 +       } else {
20388 +               struct vx_dmap_target new = {
20389 +                       .target = target,
20390 +                       .flags = flags | mode,
20391 +               };
20392 +               __set_default(vxi, mode, &new);
20393 +       }
20394 +       return 0;
20395 +}
20396 +
20397 +
20398 +static int do_unset_mapping(struct vx_info *vxi,
20399 +       dev_t device, dev_t target, int flags, umode_t mode)
20400 +{
20401 +       int ret = -EINVAL;
20402 +
20403 +       if (device) {
20404 +               ret = __remove_mapping(vxi, device, mode);
20405 +               if (ret < 0)
20406 +                       goto out;
20407 +       } else {
20408 +               ret = __remove_default(vxi, mode);
20409 +               if (ret < 0)
20410 +                       goto out;
20411 +       }
20412 +
20413 +out:
20414 +       return ret;
20415 +}
20416 +
20417 +
20418 +static inline int __user_device(const char __user *name, dev_t *dev,
20419 +       umode_t *mode)
20420 +{
20421 +       struct nameidata nd;
20422 +       int ret;
20423 +
20424 +       if (!name) {
20425 +               *dev = 0;
20426 +               return 0;
20427 +       }
20428 +       ret = user_path_walk_link(name, &nd);
20429 +       if (ret)
20430 +               return ret;
20431 +       if (nd.path.dentry->d_inode) {
20432 +               *dev = nd.path.dentry->d_inode->i_rdev;
20433 +               *mode = nd.path.dentry->d_inode->i_mode;
20434 +       }
20435 +       path_put(&nd.path);
20436 +       return 0;
20437 +}
20438 +
20439 +static inline int __mapping_mode(dev_t device, dev_t target,
20440 +       umode_t device_mode, umode_t target_mode, umode_t *mode)
20441 +{
20442 +       if (device)
20443 +               *mode = device_mode & S_IFMT;
20444 +       else if (target)
20445 +               *mode = target_mode & S_IFMT;
20446 +       else
20447 +               return -EINVAL;
20448 +
20449 +       /* if both given, device and target mode have to match */
20450 +       if (device && target &&
20451 +               ((device_mode ^ target_mode) & S_IFMT))
20452 +               return -EINVAL;
20453 +       return 0;
20454 +}
20455 +
20456 +
20457 +static inline int do_mapping(struct vx_info *vxi, const char __user *device_path,
20458 +       const char __user *target_path, int flags, int set)
20459 +{
20460 +       dev_t device = ~0, target = ~0;
20461 +       umode_t device_mode = 0, target_mode = 0, mode;
20462 +       int ret;
20463 +
20464 +       ret = __user_device(device_path, &device, &device_mode);
20465 +       if (ret)
20466 +               return ret;
20467 +       ret = __user_device(target_path, &target, &target_mode);
20468 +       if (ret)
20469 +               return ret;
20470 +
20471 +       ret = __mapping_mode(device, target,
20472 +               device_mode, target_mode, &mode);
20473 +       if (ret)
20474 +               return ret;
20475 +
20476 +       if (set)
20477 +               return do_set_mapping(vxi, device, target,
20478 +                       flags, mode);
20479 +       else
20480 +               return do_unset_mapping(vxi, device, target,
20481 +                       flags, mode);
20482 +}
20483 +
20484 +
20485 +int vc_set_mapping(struct vx_info *vxi, void __user *data)
20486 +{
20487 +       struct vcmd_set_mapping_v0 vc_data;
20488 +
20489 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20490 +               return -EFAULT;
20491 +
20492 +       return do_mapping(vxi, vc_data.device, vc_data.target,
20493 +               vc_data.flags, 1);
20494 +}
20495 +
20496 +int vc_unset_mapping(struct vx_info *vxi, void __user *data)
20497 +{
20498 +       struct vcmd_set_mapping_v0 vc_data;
20499 +
20500 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20501 +               return -EFAULT;
20502 +
20503 +       return do_mapping(vxi, vc_data.device, vc_data.target,
20504 +               vc_data.flags, 0);
20505 +}
20506 +
20507 +
20508 +#ifdef CONFIG_COMPAT
20509 +
20510 +int vc_set_mapping_x32(struct vx_info *vxi, void __user *data)
20511 +{
20512 +       struct vcmd_set_mapping_v0_x32 vc_data;
20513 +
20514 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20515 +               return -EFAULT;
20516 +
20517 +       return do_mapping(vxi, compat_ptr(vc_data.device_ptr),
20518 +               compat_ptr(vc_data.target_ptr), vc_data.flags, 1);
20519 +}
20520 +
20521 +int vc_unset_mapping_x32(struct vx_info *vxi, void __user *data)
20522 +{
20523 +       struct vcmd_set_mapping_v0_x32 vc_data;
20524 +
20525 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20526 +               return -EFAULT;
20527 +
20528 +       return do_mapping(vxi, compat_ptr(vc_data.device_ptr),
20529 +               compat_ptr(vc_data.target_ptr), vc_data.flags, 0);
20530 +}
20531 +
20532 +#endif /* CONFIG_COMPAT */
20533 +
20534 +
20535 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/dlimit.c linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/dlimit.c
20536 --- linux-2.6.25.11/kernel/vserver/dlimit.c     1969-12-31 19:00:00.000000000 -0500
20537 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/dlimit.c       2008-04-21 12:36:09.000000000 -0400
20538 @@ -0,0 +1,521 @@
20539 +/*
20540 + *  linux/kernel/vserver/dlimit.c
20541 + *
20542 + *  Virtual Server: Context Disk Limits
20543 + *
20544 + *  Copyright (C) 2004-2007  Herbert Pötzl
20545 + *
20546 + *  V0.01  initial version
20547 + *  V0.02  compat32 splitup
20548 + *
20549 + */
20550 +
20551 +#include <linux/statfs.h>
20552 +#include <linux/sched.h>
20553 +#include <linux/vs_tag.h>
20554 +#include <linux/vs_dlimit.h>
20555 +#include <linux/vserver/dlimit_cmd.h>
20556 +
20557 +#include <asm/uaccess.h>
20558 +
20559 +/*     __alloc_dl_info()
20560 +
20561 +       * allocate an initialized dl_info struct
20562 +       * doesn't make it visible (hash)                        */
20563 +
20564 +static struct dl_info *__alloc_dl_info(struct super_block *sb, tag_t tag)
20565 +{
20566 +       struct dl_info *new = NULL;
20567 +
20568 +       vxdprintk(VXD_CBIT(dlim, 5),
20569 +               "alloc_dl_info(%p,%d)*", sb, tag);
20570 +
20571 +       /* would this benefit from a slab cache? */
20572 +       new = kmalloc(sizeof(struct dl_info), GFP_KERNEL);
20573 +       if (!new)
20574 +               return 0;
20575 +
20576 +       memset(new, 0, sizeof(struct dl_info));
20577 +       new->dl_tag = tag;
20578 +       new->dl_sb = sb;
20579 +       INIT_RCU_HEAD(&new->dl_rcu);
20580 +       INIT_HLIST_NODE(&new->dl_hlist);
20581 +       spin_lock_init(&new->dl_lock);
20582 +       atomic_set(&new->dl_refcnt, 0);
20583 +       atomic_set(&new->dl_usecnt, 0);
20584 +
20585 +       /* rest of init goes here */
20586 +
20587 +       vxdprintk(VXD_CBIT(dlim, 4),
20588 +               "alloc_dl_info(%p,%d) = %p", sb, tag, new);
20589 +       return new;
20590 +}
20591 +
20592 +/*     __dealloc_dl_info()
20593 +
20594 +       * final disposal of dl_info                             */
20595 +
20596 +static void __dealloc_dl_info(struct dl_info *dli)
20597 +{
20598 +       vxdprintk(VXD_CBIT(dlim, 4),
20599 +               "dealloc_dl_info(%p)", dli);
20600 +
20601 +       dli->dl_hlist.next = LIST_POISON1;
20602 +       dli->dl_tag = -1;
20603 +       dli->dl_sb = 0;
20604 +
20605 +       BUG_ON(atomic_read(&dli->dl_usecnt));
20606 +       BUG_ON(atomic_read(&dli->dl_refcnt));
20607 +
20608 +       kfree(dli);
20609 +}
20610 +
20611 +
20612 +/*     hash table for dl_info hash */
20613 +
20614 +#define DL_HASH_SIZE   13
20615 +
20616 +struct hlist_head dl_info_hash[DL_HASH_SIZE];
20617 +
20618 +static spinlock_t dl_info_hash_lock = SPIN_LOCK_UNLOCKED;
20619 +
20620 +
20621 +static inline unsigned int __hashval(struct super_block *sb, tag_t tag)
20622 +{
20623 +       return ((tag ^ (unsigned long)sb) % DL_HASH_SIZE);
20624 +}
20625 +
20626 +
20627 +
20628 +/*     __hash_dl_info()
20629 +
20630 +       * add the dli to the global hash table
20631 +       * requires the hash_lock to be held                     */
20632 +
20633 +static inline void __hash_dl_info(struct dl_info *dli)
20634 +{
20635 +       struct hlist_head *head;
20636 +
20637 +       vxdprintk(VXD_CBIT(dlim, 6),
20638 +               "__hash_dl_info: %p[#%d]", dli, dli->dl_tag);
20639 +       get_dl_info(dli);
20640 +       head = &dl_info_hash[__hashval(dli->dl_sb, dli->dl_tag)];
20641 +       hlist_add_head_rcu(&dli->dl_hlist, head);
20642 +}
20643 +
20644 +/*     __unhash_dl_info()
20645 +
20646 +       * remove the dli from the global hash table
20647 +       * requires the hash_lock to be held                     */
20648 +
20649 +static inline void __unhash_dl_info(struct dl_info *dli)
20650 +{
20651 +       vxdprintk(VXD_CBIT(dlim, 6),
20652 +               "__unhash_dl_info: %p[#%d]", dli, dli->dl_tag);
20653 +       hlist_del_rcu(&dli->dl_hlist);
20654 +       put_dl_info(dli);
20655 +}
20656 +
20657 +
20658 +/*     __lookup_dl_info()
20659 +
20660 +       * requires the rcu_read_lock()
20661 +       * doesn't increment the dl_refcnt                       */
20662 +
20663 +static inline struct dl_info *__lookup_dl_info(struct super_block *sb, tag_t tag)
20664 +{
20665 +       struct hlist_head *head = &dl_info_hash[__hashval(sb, tag)];
20666 +       struct hlist_node *pos;
20667 +       struct dl_info *dli;
20668 +
20669 +       hlist_for_each_entry_rcu(dli, pos, head, dl_hlist) {
20670 +
20671 +               if (dli->dl_tag == tag && dli->dl_sb == sb) {
20672 +                       return dli;
20673 +               }
20674 +       }
20675 +       return NULL;
20676 +}
20677 +
20678 +
20679 +struct dl_info *locate_dl_info(struct super_block *sb, tag_t tag)
20680 +{
20681 +       struct dl_info *dli;
20682 +
20683 +       rcu_read_lock();
20684 +       dli = get_dl_info(__lookup_dl_info(sb, tag));
20685 +       vxdprintk(VXD_CBIT(dlim, 7),
20686 +               "locate_dl_info(%p,#%d) = %p", sb, tag, dli);
20687 +       rcu_read_unlock();
20688 +       return dli;
20689 +}
20690 +
20691 +void rcu_free_dl_info(struct rcu_head *head)
20692 +{
20693 +       struct dl_info *dli = container_of(head, struct dl_info, dl_rcu);
20694 +       int usecnt, refcnt;
20695 +
20696 +       BUG_ON(!dli || !head);
20697 +
20698 +       usecnt = atomic_read(&dli->dl_usecnt);
20699 +       BUG_ON(usecnt < 0);
20700 +
20701 +       refcnt = atomic_read(&dli->dl_refcnt);
20702 +       BUG_ON(refcnt < 0);
20703 +
20704 +       vxdprintk(VXD_CBIT(dlim, 3),
20705 +               "rcu_free_dl_info(%p)", dli);
20706 +       if (!usecnt)
20707 +               __dealloc_dl_info(dli);
20708 +       else
20709 +               printk("!!! rcu didn't free\n");
20710 +}
20711 +
20712 +
20713 +
20714 +
20715 +static int do_addrem_dlimit(uint32_t id, const char __user *name,
20716 +       uint32_t flags, int add)
20717 +{
20718 +       struct nameidata nd;
20719 +       int ret;
20720 +
20721 +       ret = user_path_walk_link(name, &nd);
20722 +       if (!ret) {
20723 +               struct super_block *sb;
20724 +               struct dl_info *dli;
20725 +
20726 +               ret = -EINVAL;
20727 +               if (!nd.path.dentry->d_inode)
20728 +                       goto out_release;
20729 +               if (!(sb = nd.path.dentry->d_inode->i_sb))
20730 +                       goto out_release;
20731 +
20732 +               if (add) {
20733 +                       dli = __alloc_dl_info(sb, id);
20734 +                       spin_lock(&dl_info_hash_lock);
20735 +
20736 +                       ret = -EEXIST;
20737 +                       if (__lookup_dl_info(sb, id))
20738 +                               goto out_unlock;
20739 +                       __hash_dl_info(dli);
20740 +                       dli = NULL;
20741 +               } else {
20742 +                       spin_lock(&dl_info_hash_lock);
20743 +                       dli = __lookup_dl_info(sb, id);
20744 +
20745 +                       ret = -ESRCH;
20746 +                       if (!dli)
20747 +                               goto out_unlock;
20748 +                       __unhash_dl_info(dli);
20749 +               }
20750 +               ret = 0;
20751 +       out_unlock:
20752 +               spin_unlock(&dl_info_hash_lock);
20753 +               if (add && dli)
20754 +                       __dealloc_dl_info(dli);
20755 +       out_release:
20756 +               path_put(&nd.path);
20757 +       }
20758 +       return ret;
20759 +}
20760 +
20761 +int vc_add_dlimit(uint32_t id, void __user *data)
20762 +{
20763 +       struct vcmd_ctx_dlimit_base_v0 vc_data;
20764 +
20765 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20766 +               return -EFAULT;
20767 +
20768 +       return do_addrem_dlimit(id, vc_data.name, vc_data.flags, 1);
20769 +}
20770 +
20771 +int vc_rem_dlimit(uint32_t id, void __user *data)
20772 +{
20773 +       struct vcmd_ctx_dlimit_base_v0 vc_data;
20774 +
20775 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20776 +               return -EFAULT;
20777 +
20778 +       return do_addrem_dlimit(id, vc_data.name, vc_data.flags, 0);
20779 +}
20780 +
20781 +#ifdef CONFIG_COMPAT
20782 +
20783 +int vc_add_dlimit_x32(uint32_t id, void __user *data)
20784 +{
20785 +       struct vcmd_ctx_dlimit_base_v0_x32 vc_data;
20786 +
20787 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20788 +               return -EFAULT;
20789 +
20790 +       return do_addrem_dlimit(id,
20791 +               compat_ptr(vc_data.name_ptr), vc_data.flags, 1);
20792 +}
20793 +
20794 +int vc_rem_dlimit_x32(uint32_t id, void __user *data)
20795 +{
20796 +       struct vcmd_ctx_dlimit_base_v0_x32 vc_data;
20797 +
20798 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20799 +               return -EFAULT;
20800 +
20801 +       return do_addrem_dlimit(id,
20802 +               compat_ptr(vc_data.name_ptr), vc_data.flags, 0);
20803 +}
20804 +
20805 +#endif /* CONFIG_COMPAT */
20806 +
20807 +
20808 +static inline
20809 +int do_set_dlimit(uint32_t id, const char __user *name,
20810 +       uint32_t space_used, uint32_t space_total,
20811 +       uint32_t inodes_used, uint32_t inodes_total,
20812 +       uint32_t reserved, uint32_t flags)
20813 +{
20814 +       struct nameidata nd;
20815 +       int ret;
20816 +
20817 +       ret = user_path_walk_link(name, &nd);
20818 +       if (!ret) {
20819 +               struct super_block *sb;
20820 +               struct dl_info *dli;
20821 +
20822 +               ret = -EINVAL;
20823 +               if (!nd.path.dentry->d_inode)
20824 +                       goto out_release;
20825 +               if (!(sb = nd.path.dentry->d_inode->i_sb))
20826 +                       goto out_release;
20827 +               if ((reserved != CDLIM_KEEP &&
20828 +                       reserved > 100) ||
20829 +                       (inodes_used != CDLIM_KEEP &&
20830 +                       inodes_used > inodes_total) ||
20831 +                       (space_used != CDLIM_KEEP &&
20832 +                       space_used > space_total))
20833 +                       goto out_release;
20834 +
20835 +               ret = -ESRCH;
20836 +               dli = locate_dl_info(sb, id);
20837 +               if (!dli)
20838 +                       goto out_release;
20839 +
20840 +               spin_lock(&dli->dl_lock);
20841 +
20842 +               if (inodes_used != CDLIM_KEEP)
20843 +                       dli->dl_inodes_used = inodes_used;
20844 +               if (inodes_total != CDLIM_KEEP)
20845 +                       dli->dl_inodes_total = inodes_total;
20846 +               if (space_used != CDLIM_KEEP) {
20847 +                       dli->dl_space_used = space_used;
20848 +                       dli->dl_space_used <<= 10;
20849 +               }
20850 +               if (space_total == CDLIM_INFINITY)
20851 +                       dli->dl_space_total = DLIM_INFINITY;
20852 +               else if (space_total != CDLIM_KEEP) {
20853 +                       dli->dl_space_total = space_total;
20854 +                       dli->dl_space_total <<= 10;
20855 +               }
20856 +               if (reserved != CDLIM_KEEP)
20857 +                       dli->dl_nrlmult = (1 << 10) * (100 - reserved) / 100;
20858 +
20859 +               spin_unlock(&dli->dl_lock);
20860 +
20861 +               put_dl_info(dli);
20862 +               ret = 0;
20863 +
20864 +       out_release:
20865 +               path_put(&nd.path);
20866 +       }
20867 +       return ret;
20868 +}
20869 +
20870 +int vc_set_dlimit(uint32_t id, void __user *data)
20871 +{
20872 +       struct vcmd_ctx_dlimit_v0 vc_data;
20873 +
20874 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20875 +               return -EFAULT;
20876 +
20877 +       return do_set_dlimit(id, vc_data.name,
20878 +               vc_data.space_used, vc_data.space_total,
20879 +               vc_data.inodes_used, vc_data.inodes_total,
20880 +               vc_data.reserved, vc_data.flags);
20881 +}
20882 +
20883 +#ifdef CONFIG_COMPAT
20884 +
20885 +int vc_set_dlimit_x32(uint32_t id, void __user *data)
20886 +{
20887 +       struct vcmd_ctx_dlimit_v0_x32 vc_data;
20888 +
20889 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20890 +               return -EFAULT;
20891 +
20892 +       return do_set_dlimit(id, compat_ptr(vc_data.name_ptr),
20893 +               vc_data.space_used, vc_data.space_total,
20894 +               vc_data.inodes_used, vc_data.inodes_total,
20895 +               vc_data.reserved, vc_data.flags);
20896 +}
20897 +
20898 +#endif /* CONFIG_COMPAT */
20899 +
20900 +
20901 +static inline
20902 +int do_get_dlimit(uint32_t id, const char __user *name,
20903 +       uint32_t *space_used, uint32_t *space_total,
20904 +       uint32_t *inodes_used, uint32_t *inodes_total,
20905 +       uint32_t *reserved, uint32_t *flags)
20906 +{
20907 +       struct nameidata nd;
20908 +       int ret;
20909 +
20910 +       ret = user_path_walk_link(name, &nd);
20911 +       if (!ret) {
20912 +               struct super_block *sb;
20913 +               struct dl_info *dli;
20914 +
20915 +               ret = -EINVAL;
20916 +               if (!nd.path.dentry->d_inode)
20917 +                       goto out_release;
20918 +               if (!(sb = nd.path.dentry->d_inode->i_sb))
20919 +                       goto out_release;
20920 +
20921 +               ret = -ESRCH;
20922 +               dli = locate_dl_info(sb, id);
20923 +               if (!dli)
20924 +                       goto out_release;
20925 +
20926 +               spin_lock(&dli->dl_lock);
20927 +               *inodes_used = dli->dl_inodes_used;
20928 +               *inodes_total = dli->dl_inodes_total;
20929 +               *space_used = dli->dl_space_used >> 10;
20930 +               if (dli->dl_space_total == DLIM_INFINITY)
20931 +                       *space_total = CDLIM_INFINITY;
20932 +               else
20933 +                       *space_total = dli->dl_space_total >> 10;
20934 +
20935 +               *reserved = 100 - ((dli->dl_nrlmult * 100 + 512) >> 10);
20936 +               spin_unlock(&dli->dl_lock);
20937 +
20938 +               put_dl_info(dli);
20939 +               ret = -EFAULT;
20940 +
20941 +               ret = 0;
20942 +       out_release:
20943 +               path_put(&nd.path);
20944 +       }
20945 +       return ret;
20946 +}
20947 +
20948 +
20949 +int vc_get_dlimit(uint32_t id, void __user *data)
20950 +{
20951 +       struct vcmd_ctx_dlimit_v0 vc_data;
20952 +       int ret;
20953 +
20954 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20955 +               return -EFAULT;
20956 +
20957 +       ret = do_get_dlimit(id, vc_data.name,
20958 +               &vc_data.space_used, &vc_data.space_total,
20959 +               &vc_data.inodes_used, &vc_data.inodes_total,
20960 +               &vc_data.reserved, &vc_data.flags);
20961 +       if (ret)
20962 +               return ret;
20963 +
20964 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
20965 +               return -EFAULT;
20966 +       return 0;
20967 +}
20968 +
20969 +#ifdef CONFIG_COMPAT
20970 +
20971 +int vc_get_dlimit_x32(uint32_t id, void __user *data)
20972 +{
20973 +       struct vcmd_ctx_dlimit_v0_x32 vc_data;
20974 +       int ret;
20975 +
20976 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20977 +               return -EFAULT;
20978 +
20979 +       ret = do_get_dlimit(id, compat_ptr(vc_data.name_ptr),
20980 +               &vc_data.space_used, &vc_data.space_total,
20981 +               &vc_data.inodes_used, &vc_data.inodes_total,
20982 +               &vc_data.reserved, &vc_data.flags);
20983 +       if (ret)
20984 +               return ret;
20985 +
20986 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
20987 +               return -EFAULT;
20988 +       return 0;
20989 +}
20990 +
20991 +#endif /* CONFIG_COMPAT */
20992 +
20993 +
20994 +void vx_vsi_statfs(struct super_block *sb, struct kstatfs *buf)
20995 +{
20996 +       struct dl_info *dli;
20997 +       __u64 blimit, bfree, bavail;
20998 +       __u32 ifree;
20999 +
21000 +       dli = locate_dl_info(sb, dx_current_tag());
21001 +       if (!dli)
21002 +               return;
21003 +
21004 +       spin_lock(&dli->dl_lock);
21005 +       if (dli->dl_inodes_total == (unsigned long)DLIM_INFINITY)
21006 +               goto no_ilim;
21007 +
21008 +       /* reduce max inodes available to limit */
21009 +       if (buf->f_files > dli->dl_inodes_total)
21010 +               buf->f_files = dli->dl_inodes_total;
21011 +
21012 +       ifree = dli->dl_inodes_total - dli->dl_inodes_used;
21013 +       /* reduce free inodes to min */
21014 +       if (ifree < buf->f_ffree)
21015 +               buf->f_ffree = ifree;
21016 +
21017 +no_ilim:
21018 +       if (dli->dl_space_total == DLIM_INFINITY)
21019 +               goto no_blim;
21020 +
21021 +       blimit = dli->dl_space_total >> sb->s_blocksize_bits;
21022 +
21023 +       if (dli->dl_space_total < dli->dl_space_used)
21024 +               bfree = 0;
21025 +       else
21026 +               bfree = (dli->dl_space_total - dli->dl_space_used)
21027 +                       >> sb->s_blocksize_bits;
21028 +
21029 +       bavail = ((dli->dl_space_total >> 10) * dli->dl_nrlmult);
21030 +       if (bavail < dli->dl_space_used)
21031 +               bavail = 0;
21032 +       else
21033 +               bavail = (bavail - dli->dl_space_used)
21034 +                       >> sb->s_blocksize_bits;
21035 +
21036 +       /* reduce max space available to limit */
21037 +       if (buf->f_blocks > blimit)
21038 +               buf->f_blocks = blimit;
21039 +
21040 +       /* reduce free space to min */
21041 +       if (bfree < buf->f_bfree)
21042 +               buf->f_bfree = bfree;
21043 +
21044 +       /* reduce avail space to min */
21045 +       if (bavail < buf->f_bavail)
21046 +               buf->f_bavail = bavail;
21047 +
21048 +no_blim:
21049 +       spin_unlock(&dli->dl_lock);
21050 +       put_dl_info(dli);
21051 +
21052 +       return;
21053 +}
21054 +
21055 +#include <linux/module.h>
21056 +
21057 +EXPORT_SYMBOL_GPL(locate_dl_info);
21058 +EXPORT_SYMBOL_GPL(rcu_free_dl_info);
21059 +
21060 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/helper.c linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/helper.c
21061 --- linux-2.6.25.11/kernel/vserver/helper.c     1969-12-31 19:00:00.000000000 -0500
21062 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/helper.c       2008-04-19 15:14:52.000000000 -0400
21063 @@ -0,0 +1,199 @@
21064 +/*
21065 + *  linux/kernel/vserver/helper.c
21066 + *
21067 + *  Virtual Context Support
21068 + *
21069 + *  Copyright (C) 2004-2007  Herbert Pötzl
21070 + *
21071 + *  V0.01  basic helper
21072 + *
21073 + */
21074 +
21075 +#include <linux/kmod.h>
21076 +#include <linux/reboot.h>
21077 +#include <linux/vs_context.h>
21078 +#include <linux/vs_network.h>
21079 +#include <linux/vserver/signal.h>
21080 +
21081 +
21082 +char vshelper_path[255] = "/sbin/vshelper";
21083 +
21084 +
21085 +static int do_vshelper(char *name, char *argv[], char *envp[], int sync)
21086 +{
21087 +       int ret;
21088 +
21089 +       if ((ret = call_usermodehelper(name, argv, envp, sync))) {
21090 +               printk( KERN_WARNING
21091 +                       "%s: (%s %s) returned %s with %d\n",
21092 +                       name, argv[1], argv[2],
21093 +                       sync ? "sync" : "async", ret);
21094 +       }
21095 +       vxdprintk(VXD_CBIT(switch, 4),
21096 +               "%s: (%s %s) returned %s with %d",
21097 +               name, argv[1], argv[2], sync ? "sync" : "async", ret);
21098 +       return ret;
21099 +}
21100 +
21101 +/*
21102 + *      vshelper path is set via /proc/sys
21103 + *      invoked by vserver sys_reboot(), with
21104 + *      the following arguments
21105 + *
21106 + *      argv [0] = vshelper_path;
21107 + *      argv [1] = action: "restart", "halt", "poweroff", ...
21108 + *      argv [2] = context identifier
21109 + *
21110 + *      envp [*] = type-specific parameters
21111 + */
21112 +
21113 +long vs_reboot_helper(struct vx_info *vxi, int cmd, void __user *arg)
21114 +{
21115 +       char id_buf[8], cmd_buf[16];
21116 +       char uid_buf[16], pid_buf[16];
21117 +       int ret;
21118 +
21119 +       char *argv[] = {vshelper_path, NULL, id_buf, 0};
21120 +       char *envp[] = {"HOME=/", "TERM=linux",
21121 +                       "PATH=/sbin:/usr/sbin:/bin:/usr/bin",
21122 +                       uid_buf, pid_buf, cmd_buf, 0};
21123 +
21124 +       if (vx_info_state(vxi, VXS_HELPER))
21125 +               return -EAGAIN;
21126 +       vxi->vx_state |= VXS_HELPER;
21127 +
21128 +       snprintf(id_buf, sizeof(id_buf)-1, "%d", vxi->vx_id);
21129 +
21130 +       snprintf(cmd_buf, sizeof(cmd_buf)-1, "VS_CMD=%08x", cmd);
21131 +       snprintf(uid_buf, sizeof(uid_buf)-1, "VS_UID=%d", current->uid);
21132 +       snprintf(pid_buf, sizeof(pid_buf)-1, "VS_PID=%d", current->pid);
21133 +
21134 +       switch (cmd) {
21135 +       case LINUX_REBOOT_CMD_RESTART:
21136 +               argv[1] = "restart";
21137 +               break;
21138 +
21139 +       case LINUX_REBOOT_CMD_HALT:
21140 +               argv[1] = "halt";
21141 +               break;
21142 +
21143 +       case LINUX_REBOOT_CMD_POWER_OFF:
21144 +               argv[1] = "poweroff";
21145 +               break;
21146 +
21147 +       case LINUX_REBOOT_CMD_SW_SUSPEND:
21148 +               argv[1] = "swsusp";
21149 +               break;
21150 +
21151 +       default:
21152 +               vxi->vx_state &= ~VXS_HELPER;
21153 +               return 0;
21154 +       }
21155 +
21156 +       ret = do_vshelper(vshelper_path, argv, envp, 0);
21157 +       vxi->vx_state &= ~VXS_HELPER;
21158 +       __wakeup_vx_info(vxi);
21159 +       return (ret) ? -EPERM : 0;
21160 +}
21161 +
21162 +
21163 +long vs_reboot(unsigned int cmd, void __user *arg)
21164 +{
21165 +       struct vx_info *vxi = current->vx_info;
21166 +       long ret = 0;
21167 +
21168 +       vxdprintk(VXD_CBIT(misc, 5),
21169 +               "vs_reboot(%p[#%d],%d)",
21170 +               vxi, vxi ? vxi->vx_id : 0, cmd);
21171 +
21172 +       ret = vs_reboot_helper(vxi, cmd, arg);
21173 +       if (ret)
21174 +               return ret;
21175 +
21176 +       vxi->reboot_cmd = cmd;
21177 +       if (vx_info_flags(vxi, VXF_REBOOT_KILL, 0)) {
21178 +               switch (cmd) {
21179 +               case LINUX_REBOOT_CMD_RESTART:
21180 +               case LINUX_REBOOT_CMD_HALT:
21181 +               case LINUX_REBOOT_CMD_POWER_OFF:
21182 +                       vx_info_kill(vxi, 0, SIGKILL);
21183 +                       vx_info_kill(vxi, 1, SIGKILL);
21184 +               default:
21185 +                       break;
21186 +               }
21187 +       }
21188 +       return 0;
21189 +}
21190 +
21191 +
21192 +/*
21193 + *      argv [0] = vshelper_path;
21194 + *      argv [1] = action: "startup", "shutdown"
21195 + *      argv [2] = context identifier
21196 + *
21197 + *      envp [*] = type-specific parameters
21198 + */
21199 +
21200 +long vs_state_change(struct vx_info *vxi, unsigned int cmd)
21201 +{
21202 +       char id_buf[8], cmd_buf[16];
21203 +       char *argv[] = {vshelper_path, NULL, id_buf, 0};
21204 +       char *envp[] = {"HOME=/", "TERM=linux",
21205 +                       "PATH=/sbin:/usr/sbin:/bin:/usr/bin", cmd_buf, 0};
21206 +
21207 +       if (!vx_info_flags(vxi, VXF_SC_HELPER, 0))
21208 +               return 0;
21209 +
21210 +       snprintf(id_buf, sizeof(id_buf)-1, "%d", vxi->vx_id);
21211 +       snprintf(cmd_buf, sizeof(cmd_buf)-1, "VS_CMD=%08x", cmd);
21212 +
21213 +       switch (cmd) {
21214 +       case VSC_STARTUP:
21215 +               argv[1] = "startup";
21216 +               break;
21217 +       case VSC_SHUTDOWN:
21218 +               argv[1] = "shutdown";
21219 +               break;
21220 +       default:
21221 +               return 0;
21222 +       }
21223 +
21224 +       return do_vshelper(vshelper_path, argv, envp, 1);
21225 +}
21226 +
21227 +
21228 +/*
21229 + *      argv [0] = vshelper_path;
21230 + *      argv [1] = action: "netup", "netdown"
21231 + *      argv [2] = context identifier
21232 + *
21233 + *      envp [*] = type-specific parameters
21234 + */
21235 +
21236 +long vs_net_change(struct nx_info *nxi, unsigned int cmd)
21237 +{
21238 +       char id_buf[8], cmd_buf[16];
21239 +       char *argv[] = {vshelper_path, NULL, id_buf, 0};
21240 +       char *envp[] = {"HOME=/", "TERM=linux",
21241 +                       "PATH=/sbin:/usr/sbin:/bin:/usr/bin", cmd_buf, 0};
21242 +
21243 +       if (!nx_info_flags(nxi, NXF_SC_HELPER, 0))
21244 +               return 0;
21245 +
21246 +       snprintf(id_buf, sizeof(id_buf)-1, "%d", nxi->nx_id);
21247 +       snprintf(cmd_buf, sizeof(cmd_buf)-1, "VS_CMD=%08x", cmd);
21248 +
21249 +       switch (cmd) {
21250 +       case VSC_NETUP:
21251 +               argv[1] = "netup";
21252 +               break;
21253 +       case VSC_NETDOWN:
21254 +               argv[1] = "netdown";
21255 +               break;
21256 +       default:
21257 +               return 0;
21258 +       }
21259 +
21260 +       return do_vshelper(vshelper_path, argv, envp, 1);
21261 +}
21262 +
21263 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/history.c linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/history.c
21264 --- linux-2.6.25.11/kernel/vserver/history.c    1969-12-31 19:00:00.000000000 -0500
21265 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/history.c      2008-04-19 15:14:52.000000000 -0400
21266 @@ -0,0 +1,258 @@
21267 +/*
21268 + *  kernel/vserver/history.c
21269 + *
21270 + *  Virtual Context History Backtrace
21271 + *
21272 + *  Copyright (C) 2004-2007  Herbert Pötzl
21273 + *
21274 + *  V0.01  basic structure
21275 + *  V0.02  hash/unhash and trace
21276 + *  V0.03  preemption fixes
21277 + *
21278 + */
21279 +
21280 +#include <linux/module.h>
21281 +#include <asm/uaccess.h>
21282 +
21283 +#include <linux/vserver/context.h>
21284 +#include <linux/vserver/debug.h>
21285 +#include <linux/vserver/debug_cmd.h>
21286 +#include <linux/vserver/history.h>
21287 +
21288 +
21289 +#ifdef CONFIG_VSERVER_HISTORY
21290 +#define VXH_SIZE       CONFIG_VSERVER_HISTORY_SIZE
21291 +#else
21292 +#define VXH_SIZE       64
21293 +#endif
21294 +
21295 +struct _vx_history {
21296 +       unsigned int counter;
21297 +
21298 +       struct _vx_hist_entry entry[VXH_SIZE + 1];
21299 +};
21300 +
21301 +
21302 +DEFINE_PER_CPU(struct _vx_history, vx_history_buffer);
21303 +
21304 +unsigned volatile int vxh_active = 1;
21305 +
21306 +static atomic_t sequence = ATOMIC_INIT(0);
21307 +
21308 +
21309 +/*     vxh_advance()
21310 +
21311 +       * requires disabled preemption                          */
21312 +
21313 +struct _vx_hist_entry *vxh_advance(void *loc)
21314 +{
21315 +       unsigned int cpu = smp_processor_id();
21316 +       struct _vx_history *hist = &per_cpu(vx_history_buffer, cpu);
21317 +       struct _vx_hist_entry *entry;
21318 +       unsigned int index;
21319 +
21320 +       index = vxh_active ? (hist->counter++ % VXH_SIZE) : VXH_SIZE;
21321 +       entry = &hist->entry[index];
21322 +
21323 +       entry->seq = atomic_inc_return(&sequence);
21324 +       entry->loc = loc;
21325 +       return entry;
21326 +}
21327 +
21328 +EXPORT_SYMBOL_GPL(vxh_advance);
21329 +
21330 +
21331 +#define VXH_LOC_FMTS   "(#%04x,*%d):%p"
21332 +
21333 +#define VXH_LOC_ARGS(e)        (e)->seq, cpu, (e)->loc
21334 +
21335 +
21336 +#define VXH_VXI_FMTS   "%p[#%d,%d.%d]"
21337 +
21338 +#define VXH_VXI_ARGS(e)        (e)->vxi.ptr,                           \
21339 +                       (e)->vxi.ptr ? (e)->vxi.xid : 0,        \
21340 +                       (e)->vxi.ptr ? (e)->vxi.usecnt : 0,     \
21341 +                       (e)->vxi.ptr ? (e)->vxi.tasks : 0
21342 +
21343 +void   vxh_dump_entry(struct _vx_hist_entry *e, unsigned cpu)
21344 +{
21345 +       switch (e->type) {
21346 +       case VXH_THROW_OOPS:
21347 +               printk( VXH_LOC_FMTS " oops \n", VXH_LOC_ARGS(e));
21348 +               break;
21349 +
21350 +       case VXH_GET_VX_INFO:
21351 +       case VXH_PUT_VX_INFO:
21352 +               printk( VXH_LOC_FMTS " %s_vx_info " VXH_VXI_FMTS "\n",
21353 +                       VXH_LOC_ARGS(e),
21354 +                       (e->type == VXH_GET_VX_INFO) ? "get" : "put",
21355 +                       VXH_VXI_ARGS(e));
21356 +               break;
21357 +
21358 +       case VXH_INIT_VX_INFO:
21359 +       case VXH_SET_VX_INFO:
21360 +       case VXH_CLR_VX_INFO:
21361 +               printk( VXH_LOC_FMTS " %s_vx_info " VXH_VXI_FMTS " @%p\n",
21362 +                       VXH_LOC_ARGS(e),
21363 +                       (e->type == VXH_INIT_VX_INFO) ? "init" :
21364 +                       ((e->type == VXH_SET_VX_INFO) ? "set" : "clr"),
21365 +                       VXH_VXI_ARGS(e), e->sc.data);
21366 +               break;
21367 +
21368 +       case VXH_CLAIM_VX_INFO:
21369 +       case VXH_RELEASE_VX_INFO:
21370 +               printk( VXH_LOC_FMTS " %s_vx_info " VXH_VXI_FMTS " @%p\n",
21371 +                       VXH_LOC_ARGS(e),
21372 +                       (e->type == VXH_CLAIM_VX_INFO) ? "claim" : "release",
21373 +                       VXH_VXI_ARGS(e), e->sc.data);
21374 +               break;
21375 +
21376 +       case VXH_ALLOC_VX_INFO:
21377 +       case VXH_DEALLOC_VX_INFO:
21378 +               printk( VXH_LOC_FMTS " %s_vx_info " VXH_VXI_FMTS "\n",
21379 +                       VXH_LOC_ARGS(e),
21380 +                       (e->type == VXH_ALLOC_VX_INFO) ? "alloc" : "dealloc",
21381 +                       VXH_VXI_ARGS(e));
21382 +               break;
21383 +
21384 +       case VXH_HASH_VX_INFO:
21385 +       case VXH_UNHASH_VX_INFO:
21386 +               printk( VXH_LOC_FMTS " __%s_vx_info " VXH_VXI_FMTS "\n",
21387 +                       VXH_LOC_ARGS(e),
21388 +                       (e->type == VXH_HASH_VX_INFO) ? "hash" : "unhash",
21389 +                       VXH_VXI_ARGS(e));
21390 +               break;
21391 +
21392 +       case VXH_LOC_VX_INFO:
21393 +       case VXH_LOOKUP_VX_INFO:
21394 +       case VXH_CREATE_VX_INFO:
21395 +               printk( VXH_LOC_FMTS " __%s_vx_info [#%d] -> " VXH_VXI_FMTS "\n",
21396 +                       VXH_LOC_ARGS(e),
21397 +                       (e->type == VXH_CREATE_VX_INFO) ? "create" :
21398 +                       ((e->type == VXH_LOC_VX_INFO) ? "loc" : "lookup"),
21399 +                       e->ll.arg, VXH_VXI_ARGS(e));
21400 +               break;
21401 +       }
21402 +}
21403 +
21404 +static void __vxh_dump_history(void)
21405 +{
21406 +       unsigned int i, cpu;
21407 +
21408 +       printk("History:\tSEQ: %8x\tNR_CPUS: %d\n",
21409 +               atomic_read(&sequence), NR_CPUS);
21410 +
21411 +       for (i = 0; i < VXH_SIZE; i++) {
21412 +               for_each_online_cpu(cpu) {
21413 +                       struct _vx_history *hist =
21414 +                               &per_cpu(vx_history_buffer, cpu);
21415 +                       unsigned int index = (hist->counter - i) % VXH_SIZE;
21416 +                       struct _vx_hist_entry *entry = &hist->entry[index];
21417 +
21418 +                       vxh_dump_entry(entry, cpu);
21419 +               }
21420 +       }
21421 +}
21422 +
21423 +void   vxh_dump_history(void)
21424 +{
21425 +       vxh_active = 0;
21426 +#ifdef CONFIG_SMP
21427 +       local_irq_enable();
21428 +       smp_send_stop();
21429 +       local_irq_disable();
21430 +#endif
21431 +       __vxh_dump_history();
21432 +}
21433 +
21434 +
21435 +/* vserver syscall commands below here */
21436 +
21437 +
21438 +int vc_dump_history(uint32_t id)
21439 +{
21440 +       vxh_active = 0;
21441 +       __vxh_dump_history();
21442 +       vxh_active = 1;
21443 +
21444 +       return 0;
21445 +}
21446 +
21447 +
21448 +int do_read_history(struct __user _vx_hist_entry *data,
21449 +       int cpu, uint32_t *index, uint32_t *count)
21450 +{
21451 +       int pos, ret = 0;
21452 +       struct _vx_history *hist = &per_cpu(vx_history_buffer, cpu);
21453 +       int end = hist->counter;
21454 +       int start = end - VXH_SIZE + 2;
21455 +       int idx = *index;
21456 +
21457 +       /* special case: get current pos */
21458 +       if (!*count) {
21459 +               *index = end;
21460 +               return 0;
21461 +       }
21462 +
21463 +       /* have we lost some data? */
21464 +       if (idx < start)
21465 +               idx = start;
21466 +
21467 +       for (pos = 0; (pos < *count) && (idx < end); pos++, idx++) {
21468 +               struct _vx_hist_entry *entry =
21469 +                       &hist->entry[idx % VXH_SIZE];
21470 +
21471 +               /* send entry to userspace */
21472 +               ret = copy_to_user(&data[pos], entry, sizeof(*entry));
21473 +               if (ret)
21474 +                       break;
21475 +       }
21476 +       /* save new index and count */
21477 +       *index = idx;
21478 +       *count = pos;
21479 +       return ret ? ret : (*index < end);
21480 +}
21481 +
21482 +int vc_read_history(uint32_t id, void __user *data)
21483 +{
21484 +       struct vcmd_read_history_v0 vc_data;
21485 +       int ret;
21486 +
21487 +       if (id >= NR_CPUS)
21488 +               return -EINVAL;
21489 +
21490 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
21491 +               return -EFAULT;
21492 +
21493 +       ret = do_read_history((struct __user _vx_hist_entry *)vc_data.data,
21494 +               id, &vc_data.index, &vc_data.count);
21495 +
21496 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
21497 +               return -EFAULT;
21498 +       return ret;
21499 +}
21500 +
21501 +#ifdef CONFIG_COMPAT
21502 +
21503 +int vc_read_history_x32(uint32_t id, void __user *data)
21504 +{
21505 +       struct vcmd_read_history_v0_x32 vc_data;
21506 +       int ret;
21507 +
21508 +       if (id >= NR_CPUS)
21509 +               return -EINVAL;
21510 +
21511 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
21512 +               return -EFAULT;
21513 +
21514 +       ret = do_read_history((struct __user _vx_hist_entry *)
21515 +               compat_ptr(vc_data.data_ptr),
21516 +               id, &vc_data.index, &vc_data.count);
21517 +
21518 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
21519 +               return -EFAULT;
21520 +       return ret;
21521 +}
21522 +
21523 +#endif /* CONFIG_COMPAT */
21524 +
21525 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/inet.c linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/inet.c
21526 --- linux-2.6.25.11/kernel/vserver/inet.c       1969-12-31 19:00:00.000000000 -0500
21527 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/inet.c 2008-05-29 18:56:59.000000000 -0400
21528 @@ -0,0 +1,225 @@
21529 +
21530 +#include <linux/in.h>
21531 +#include <linux/inetdevice.h>
21532 +#include <linux/vs_inet.h>
21533 +#include <linux/vs_inet6.h>
21534 +#include <linux/vserver/debug.h>
21535 +#include <net/route.h>
21536 +#include <net/addrconf.h>
21537 +
21538 +
21539 +int nx_v4_addr_conflict(struct nx_info *nxi1, struct nx_info *nxi2)
21540 +{
21541 +       int ret = 0;
21542 +
21543 +       if (!nxi1 || !nxi2 || nxi1 == nxi2)
21544 +               ret = 1;
21545 +       else {
21546 +               struct nx_addr_v4 *ptr;
21547 +
21548 +               for (ptr = &nxi1->v4; ptr; ptr = ptr->next) {
21549 +                       if (v4_nx_addr_in_nx_info(nxi2, ptr, -1)) {
21550 +                               ret = 1;
21551 +                               break;
21552 +                       }
21553 +               }
21554 +       }
21555 +
21556 +       vxdprintk(VXD_CBIT(net, 2),
21557 +               "nx_v4_addr_conflict(%p,%p): %d",
21558 +               nxi1, nxi2, ret);
21559 +
21560 +       return ret;
21561 +}
21562 +
21563 +
21564 +#ifdef CONFIG_IPV6
21565 +
21566 +int nx_v6_addr_conflict(struct nx_info *nxi1, struct nx_info *nxi2)
21567 +{
21568 +       int ret = 0;
21569 +
21570 +       if (!nxi1 || !nxi2 || nxi1 == nxi2)
21571 +               ret = 1;
21572 +       else {
21573 +               struct nx_addr_v6 *ptr;
21574 +
21575 +               for (ptr = &nxi1->v6; ptr; ptr = ptr->next) {
21576 +                       if (v6_nx_addr_in_nx_info(nxi2, ptr, -1)) {
21577 +                               ret = 1;
21578 +                               break;
21579 +                       }
21580 +               }
21581 +       }
21582 +
21583 +       vxdprintk(VXD_CBIT(net, 2),
21584 +               "nx_v6_addr_conflict(%p,%p): %d",
21585 +               nxi1, nxi2, ret);
21586 +
21587 +       return ret;
21588 +}
21589 +
21590 +#endif
21591 +
21592 +int v4_dev_in_nx_info(struct net_device *dev, struct nx_info *nxi)
21593 +{
21594 +       struct in_device *in_dev;
21595 +       struct in_ifaddr **ifap;
21596 +       struct in_ifaddr *ifa;
21597 +       int ret = 0;
21598 +
21599 +       if (!dev)
21600 +               goto out;
21601 +       in_dev = in_dev_get(dev);
21602 +       if (!in_dev)
21603 +               goto out;
21604 +
21605 +       for (ifap = &in_dev->ifa_list; (ifa = *ifap) != NULL;
21606 +               ifap = &ifa->ifa_next) {
21607 +               if (v4_addr_in_nx_info(nxi, ifa->ifa_local, NXA_MASK_SHOW)) {
21608 +                       ret = 1;
21609 +                       break;
21610 +               }
21611 +       }
21612 +       in_dev_put(in_dev);
21613 +out:
21614 +       return ret;
21615 +}
21616 +
21617 +
21618 +#ifdef CONFIG_IPV6
21619 +
21620 +int v6_dev_in_nx_info(struct net_device *dev, struct nx_info *nxi)
21621 +{
21622 +       struct inet6_dev *in_dev;
21623 +       struct inet6_ifaddr **ifap;
21624 +       struct inet6_ifaddr *ifa;
21625 +       int ret = 0;
21626 +
21627 +       if (!dev)
21628 +               goto out;
21629 +       in_dev = in6_dev_get(dev);
21630 +       if (!in_dev)
21631 +               goto out;
21632 +
21633 +       for (ifap = &in_dev->addr_list; (ifa = *ifap) != NULL;
21634 +               ifap = &ifa->if_next) {
21635 +               if (v6_addr_in_nx_info(nxi, &ifa->addr, -1)) {
21636 +                       ret = 1;
21637 +                       break;
21638 +               }
21639 +       }
21640 +       in6_dev_put(in_dev);
21641 +out:
21642 +       return ret;
21643 +}
21644 +
21645 +#endif
21646 +
21647 +int dev_in_nx_info(struct net_device *dev, struct nx_info *nxi)
21648 +{
21649 +       int ret = 1;
21650 +
21651 +       if (!nxi)
21652 +               goto out;
21653 +       if (nxi->v4.type && v4_dev_in_nx_info(dev, nxi))
21654 +               goto out;
21655 +#ifdef CONFIG_IPV6
21656 +       ret = 2;
21657 +       if (nxi->v6.type && v6_dev_in_nx_info(dev, nxi))
21658 +               goto out;
21659 +#endif
21660 +       ret = 0;
21661 +out:
21662 +       vxdprintk(VXD_CBIT(net, 3),
21663 +               "dev_in_nx_info(%p,%p[#%d]) = %d",
21664 +               dev, nxi, nxi ? nxi->nx_id : 0, ret);
21665 +       return ret;
21666 +}
21667 +
21668 +int ip_v4_find_src(struct net *net, struct nx_info *nxi,
21669 +       struct rtable **rp, struct flowi *fl)
21670 +{
21671 +       if (!nxi)
21672 +               return 0;
21673 +
21674 +       /* FIXME: handle lback only case */
21675 +       if (!NX_IPV4(nxi))
21676 +               return -EPERM;
21677 +
21678 +       vxdprintk(VXD_CBIT(net, 4),
21679 +               "ip_v4_find_src(%p[#%u]) " NIPQUAD_FMT " -> " NIPQUAD_FMT,
21680 +               nxi, nxi ? nxi->nx_id : 0,
21681 +               NIPQUAD(fl->fl4_src), NIPQUAD(fl->fl4_dst));
21682 +
21683 +       /* single IP is unconditional */
21684 +       if (nx_info_flags(nxi, NXF_SINGLE_IP, 0) &&
21685 +               (fl->fl4_src == INADDR_ANY))
21686 +               fl->fl4_src = nxi->v4.ip[0].s_addr;
21687 +
21688 +       if (fl->fl4_src == INADDR_ANY) {
21689 +               struct nx_addr_v4 *ptr;
21690 +               __be32 found = 0;
21691 +               int err;
21692 +
21693 +               err = __ip_route_output_key(net, rp, fl);
21694 +               if (!err) {
21695 +                       found = (*rp)->rt_src;
21696 +                       ip_rt_put(*rp);
21697 +                       vxdprintk(VXD_CBIT(net, 4),
21698 +                               "ip_v4_find_src(%p[#%u]) rok[%u]: " NIPQUAD_FMT,
21699 +                               nxi, nxi ? nxi->nx_id : 0, fl->oif, NIPQUAD(found));
21700 +                       if (v4_addr_in_nx_info(nxi, found, NXA_MASK_BIND))
21701 +                               goto found;
21702 +               }
21703 +
21704 +               for (ptr = &nxi->v4; ptr; ptr = ptr->next) {
21705 +                       __be32 primary = ptr->ip[0].s_addr;
21706 +                       __be32 mask = ptr->mask.s_addr;
21707 +                       __be32 neta = primary & mask;
21708 +
21709 +                       vxdprintk(VXD_CBIT(net, 4), "ip_v4_find_src(%p[#%u]) chk: "
21710 +                               NIPQUAD_FMT "/" NIPQUAD_FMT "/" NIPQUAD_FMT,
21711 +                               nxi, nxi ? nxi->nx_id : 0, NIPQUAD(primary),
21712 +                               NIPQUAD(mask), NIPQUAD(neta));
21713 +                       if ((found & mask) != neta)
21714 +                               continue;
21715 +
21716 +                       fl->fl4_src = primary;
21717 +                       err = __ip_route_output_key(net, rp, fl);
21718 +                       vxdprintk(VXD_CBIT(net, 4),
21719 +                               "ip_v4_find_src(%p[#%u]) rok[%u]: " NIPQUAD_FMT,
21720 +                               nxi, nxi ? nxi->nx_id : 0, fl->oif, NIPQUAD(primary));
21721 +                       if (!err) {
21722 +                               found = (*rp)->rt_src;
21723 +                               ip_rt_put(*rp);
21724 +                               if (found == primary)
21725 +                                       goto found;
21726 +                       }
21727 +               }
21728 +               /* still no source ip? */
21729 +               found = ipv4_is_loopback(fl->fl4_dst)
21730 +                       ? IPI_LOOPBACK : nxi->v4.ip[0].s_addr;
21731 +       found:
21732 +               /* assign src ip to flow */
21733 +               fl->fl4_src = found;
21734 +
21735 +       } else {
21736 +               if (!v4_addr_in_nx_info(nxi, fl->fl4_src, NXA_MASK_BIND))
21737 +                       return -EPERM;
21738 +       }
21739 +
21740 +       if (nx_info_flags(nxi, NXF_LBACK_REMAP, 0)) {
21741 +               if (ipv4_is_loopback(fl->fl4_dst))
21742 +                       fl->fl4_dst = nxi->v4_lback.s_addr;
21743 +               if (ipv4_is_loopback(fl->fl4_src))
21744 +                       fl->fl4_src = nxi->v4_lback.s_addr;
21745 +       } else if (ipv4_is_loopback(fl->fl4_dst) &&
21746 +               !nx_info_flags(nxi, NXF_LBACK_ALLOW, 0))
21747 +               return -EPERM;
21748 +
21749 +       return 0;
21750 +}
21751 +
21752 +EXPORT_SYMBOL_GPL(ip_v4_find_src);
21753 +
21754 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/init.c linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/init.c
21755 --- linux-2.6.25.11/kernel/vserver/init.c       1969-12-31 19:00:00.000000000 -0500
21756 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/init.c 2008-04-19 15:14:52.000000000 -0400
21757 @@ -0,0 +1,45 @@
21758 +/*
21759 + *  linux/kernel/init.c
21760 + *
21761 + *  Virtual Server Init
21762 + *
21763 + *  Copyright (C) 2004-2007  Herbert Pötzl
21764 + *
21765 + *  V0.01  basic structure
21766 + *
21767 + */
21768 +
21769 +#include <linux/init.h>
21770 +
21771 +int    vserver_register_sysctl(void);
21772 +void   vserver_unregister_sysctl(void);
21773 +
21774 +
21775 +static int __init init_vserver(void)
21776 +{
21777 +       int ret = 0;
21778 +
21779 +#ifdef CONFIG_VSERVER_DEBUG
21780 +       vserver_register_sysctl();
21781 +#endif
21782 +       return ret;
21783 +}
21784 +
21785 +
21786 +static void __exit exit_vserver(void)
21787 +{
21788 +
21789 +#ifdef CONFIG_VSERVER_DEBUG
21790 +       vserver_unregister_sysctl();
21791 +#endif
21792 +       return;
21793 +}
21794 +
21795 +/* FIXME: GFP_ZONETYPES gone
21796 +long vx_slab[GFP_ZONETYPES]; */
21797 +long vx_area;
21798 +
21799 +
21800 +module_init(init_vserver);
21801 +module_exit(exit_vserver);
21802 +
21803 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/inode.c linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/inode.c
21804 --- linux-2.6.25.11/kernel/vserver/inode.c      1969-12-31 19:00:00.000000000 -0500
21805 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/inode.c        2008-07-08 13:41:38.000000000 -0400
21806 @@ -0,0 +1,409 @@
21807 +/*
21808 + *  linux/kernel/vserver/inode.c
21809 + *
21810 + *  Virtual Server: File System Support
21811 + *
21812 + *  Copyright (C) 2004-2007  Herbert Pötzl
21813 + *
21814 + *  V0.01  separated from vcontext V0.05
21815 + *  V0.02  moved to tag (instead of xid)
21816 + *
21817 + */
21818 +
21819 +#include <linux/tty.h>
21820 +#include <linux/proc_fs.h>
21821 +#include <linux/devpts_fs.h>
21822 +#include <linux/fs.h>
21823 +#include <linux/file.h>
21824 +#include <linux/mount.h>
21825 +#include <linux/parser.h>
21826 +#include <linux/vserver/inode.h>
21827 +#include <linux/vserver/inode_cmd.h>
21828 +#include <linux/vs_base.h>
21829 +#include <linux/vs_tag.h>
21830 +
21831 +#include <asm/uaccess.h>
21832 +
21833 +
21834 +static int __vc_get_iattr(struct inode *in, uint32_t *tag, uint32_t *flags, uint32_t *mask)
21835 +{
21836 +       struct proc_dir_entry *entry;
21837 +
21838 +       if (!in || !in->i_sb)
21839 +               return -ESRCH;
21840 +
21841 +       *flags = IATTR_TAG
21842 +               | (IS_BARRIER(in) ? IATTR_BARRIER : 0)
21843 +               | (IS_IXUNLINK(in) ? IATTR_IXUNLINK : 0)
21844 +               | (IS_IMMUTABLE(in) ? IATTR_IMMUTABLE : 0);
21845 +       *mask = IATTR_IXUNLINK | IATTR_IMMUTABLE;
21846 +
21847 +       if (S_ISDIR(in->i_mode))
21848 +               *mask |= IATTR_BARRIER;
21849 +
21850 +       if (IS_TAGGED(in)) {
21851 +               *tag = in->i_tag;
21852 +               *mask |= IATTR_TAG;
21853 +       }
21854 +
21855 +       switch (in->i_sb->s_magic) {
21856 +       case PROC_SUPER_MAGIC:
21857 +               entry = PROC_I(in)->pde;
21858 +
21859 +               /* check for specific inodes? */
21860 +               if (entry)
21861 +                       *mask |= IATTR_FLAGS;
21862 +               if (entry)
21863 +                       *flags |= (entry->vx_flags & IATTR_FLAGS);
21864 +               else
21865 +                       *flags |= (PROC_I(in)->vx_flags & IATTR_FLAGS);
21866 +               break;
21867 +
21868 +       case DEVPTS_SUPER_MAGIC:
21869 +               *tag = in->i_tag;
21870 +               *mask |= IATTR_TAG;
21871 +               break;
21872 +
21873 +       default:
21874 +               break;
21875 +       }
21876 +       return 0;
21877 +}
21878 +
21879 +int vc_get_iattr(void __user *data)
21880 +{
21881 +       struct nameidata nd;
21882 +       struct vcmd_ctx_iattr_v1 vc_data = { .tag = -1 };
21883 +       int ret;
21884 +
21885 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
21886 +               return -EFAULT;
21887 +
21888 +       ret = user_path_walk_link(vc_data.name, &nd);
21889 +       if (!ret) {
21890 +               ret = __vc_get_iattr(nd.path.dentry->d_inode,
21891 +                       &vc_data.tag, &vc_data.flags, &vc_data.mask);
21892 +               path_put(&nd.path);
21893 +       }
21894 +       if (ret)
21895 +               return ret;
21896 +
21897 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
21898 +               ret = -EFAULT;
21899 +       return ret;
21900 +}
21901 +
21902 +#ifdef CONFIG_COMPAT
21903 +
21904 +int vc_get_iattr_x32(void __user *data)
21905 +{
21906 +       struct nameidata nd;
21907 +       struct vcmd_ctx_iattr_v1_x32 vc_data = { .tag = -1 };
21908 +       int ret;
21909 +
21910 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
21911 +               return -EFAULT;
21912 +
21913 +       ret = user_path_walk_link(compat_ptr(vc_data.name_ptr), &nd);
21914 +       if (!ret) {
21915 +               ret = __vc_get_iattr(nd.path.dentry->d_inode,
21916 +                       &vc_data.tag, &vc_data.flags, &vc_data.mask);
21917 +               path_put(&nd.path);
21918 +       }
21919 +       if (ret)
21920 +               return ret;
21921 +
21922 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
21923 +               ret = -EFAULT;
21924 +       return ret;
21925 +}
21926 +
21927 +#endif /* CONFIG_COMPAT */
21928 +
21929 +
21930 +int vc_fget_iattr(uint32_t fd, void __user *data)
21931 +{
21932 +       struct file *filp;
21933 +       struct vcmd_ctx_fiattr_v0 vc_data = { .tag = -1 };
21934 +       int ret;
21935 +
21936 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
21937 +               return -EFAULT;
21938 +
21939 +       filp = fget(fd);
21940 +       if (!filp || !filp->f_dentry || !filp->f_dentry->d_inode)
21941 +               return -EBADF;
21942 +
21943 +       ret = __vc_get_iattr(filp->f_dentry->d_inode,
21944 +               &vc_data.tag, &vc_data.flags, &vc_data.mask);
21945 +
21946 +       fput(filp);
21947 +
21948 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
21949 +               ret = -EFAULT;
21950 +       return ret;
21951 +}
21952 +
21953 +
21954 +static int __vc_set_iattr(struct dentry *de, uint32_t *tag, uint32_t *flags, uint32_t *mask)
21955 +{
21956 +       struct inode *in = de->d_inode;
21957 +       int error = 0, is_proc = 0, has_tag = 0;
21958 +       struct iattr attr = { 0 };
21959 +
21960 +       if (!in || !in->i_sb)
21961 +               return -ESRCH;
21962 +
21963 +       is_proc = (in->i_sb->s_magic == PROC_SUPER_MAGIC);
21964 +       if ((*mask & IATTR_FLAGS) && !is_proc)
21965 +               return -EINVAL;
21966 +
21967 +       has_tag = IS_TAGGED(in) ||
21968 +               (in->i_sb->s_magic == DEVPTS_SUPER_MAGIC);
21969 +       if ((*mask & IATTR_TAG) && !has_tag)
21970 +               return -EINVAL;
21971 +
21972 +       mutex_lock(&in->i_mutex);
21973 +       if (*mask & IATTR_TAG) {
21974 +               attr.ia_tag = *tag;
21975 +               attr.ia_valid |= ATTR_TAG;
21976 +       }
21977 +
21978 +       if (*mask & IATTR_FLAGS) {
21979 +               struct proc_dir_entry *entry = PROC_I(in)->pde;
21980 +               unsigned int iflags = PROC_I(in)->vx_flags;
21981 +
21982 +               iflags = (iflags & ~(*mask & IATTR_FLAGS))
21983 +                       | (*flags & IATTR_FLAGS);
21984 +               PROC_I(in)->vx_flags = iflags;
21985 +               if (entry)
21986 +                       entry->vx_flags = iflags;
21987 +       }
21988 +
21989 +       if (*mask & (IATTR_BARRIER | IATTR_IXUNLINK | IATTR_IMMUTABLE)) {
21990 +               if (*mask & IATTR_IMMUTABLE) {
21991 +                       if (*flags & IATTR_IMMUTABLE)
21992 +                               in->i_flags |= S_IMMUTABLE;
21993 +                       else
21994 +                               in->i_flags &= ~S_IMMUTABLE;
21995 +               }
21996 +               if (*mask & IATTR_IXUNLINK) {
21997 +                       if (*flags & IATTR_IXUNLINK)
21998 +                               in->i_flags |= S_IXUNLINK;
21999 +                       else
22000 +                               in->i_flags &= ~S_IXUNLINK;
22001 +               }
22002 +               if (S_ISDIR(in->i_mode) && (*mask & IATTR_BARRIER)) {
22003 +                       if (*flags & IATTR_BARRIER)
22004 +                               in->i_vflags |= V_BARRIER;
22005 +                       else
22006 +                               in->i_vflags &= ~V_BARRIER;
22007 +               }
22008 +               if (in->i_op && in->i_op->sync_flags) {
22009 +                       error = in->i_op->sync_flags(in);
22010 +                       if (error)
22011 +                               goto out;
22012 +               }
22013 +       }
22014 +
22015 +       if (attr.ia_valid) {
22016 +               if (in->i_op && in->i_op->setattr)
22017 +                       error = in->i_op->setattr(de, &attr);
22018 +               else {
22019 +                       error = inode_change_ok(in, &attr);
22020 +                       if (!error)
22021 +                               error = inode_setattr(in, &attr);
22022 +               }
22023 +       }
22024 +
22025 +out:
22026 +       mutex_unlock(&in->i_mutex);
22027 +       return error;
22028 +}
22029 +
22030 +int vc_set_iattr(void __user *data)
22031 +{
22032 +       struct nameidata nd;
22033 +       struct vcmd_ctx_iattr_v1 vc_data;
22034 +       int ret;
22035 +
22036 +       if (!capable(CAP_LINUX_IMMUTABLE))
22037 +               return -EPERM;
22038 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
22039 +               return -EFAULT;
22040 +
22041 +       ret = user_path_walk_link(vc_data.name, &nd);
22042 +       if (!ret) {
22043 +               ret = __vc_set_iattr(nd.path.dentry,
22044 +                       &vc_data.tag, &vc_data.flags, &vc_data.mask);
22045 +               path_put(&nd.path);
22046 +       }
22047 +
22048 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
22049 +               ret = -EFAULT;
22050 +       return ret;
22051 +}
22052 +
22053 +#ifdef CONFIG_COMPAT
22054 +
22055 +int vc_set_iattr_x32(void __user *data)
22056 +{
22057 +       struct nameidata nd;
22058 +       struct vcmd_ctx_iattr_v1_x32 vc_data;
22059 +       int ret;
22060 +
22061 +       if (!capable(CAP_LINUX_IMMUTABLE))
22062 +               return -EPERM;
22063 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
22064 +               return -EFAULT;
22065 +
22066 +       ret = user_path_walk_link(compat_ptr(vc_data.name_ptr), &nd);
22067 +       if (!ret) {
22068 +               ret = __vc_set_iattr(nd.path.dentry,
22069 +                       &vc_data.tag, &vc_data.flags, &vc_data.mask);
22070 +               path_put(&nd.path);
22071 +       }
22072 +
22073 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
22074 +               ret = -EFAULT;
22075 +       return ret;
22076 +}
22077 +
22078 +#endif /* CONFIG_COMPAT */
22079 +
22080 +int vc_fset_iattr(uint32_t fd, void __user *data)
22081 +{
22082 +       struct file *filp;
22083 +       struct vcmd_ctx_fiattr_v0 vc_data;
22084 +       int ret;
22085 +
22086 +       if (!capable(CAP_LINUX_IMMUTABLE))
22087 +               return -EPERM;
22088 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
22089 +               return -EFAULT;
22090 +
22091 +       filp = fget(fd);
22092 +       if (!filp || !filp->f_dentry || !filp->f_dentry->d_inode)
22093 +               return -EBADF;
22094 +
22095 +       ret = __vc_set_iattr(filp->f_dentry, &vc_data.tag,
22096 +               &vc_data.flags, &vc_data.mask);
22097 +
22098 +       fput(filp);
22099 +
22100 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
22101 +               return -EFAULT;
22102 +       return ret;
22103 +}
22104 +
22105 +
22106 +enum { Opt_notagcheck, Opt_tag, Opt_notag, Opt_tagid, Opt_err };
22107 +
22108 +static match_table_t tokens = {
22109 +       {Opt_notagcheck, "notagcheck"},
22110 +#ifdef CONFIG_PROPAGATE
22111 +       {Opt_notag, "notag"},
22112 +       {Opt_tag, "tag"},
22113 +       {Opt_tagid, "tagid=%u"},
22114 +#endif
22115 +       {Opt_err, NULL}
22116 +};
22117 +
22118 +
22119 +static void __dx_parse_remove(char *string, char *opt)
22120 +{
22121 +       char *p = strstr(string, opt);
22122 +       char *q = p;
22123 +
22124 +       if (p) {
22125 +               while (*q != '\0' && *q != ',')
22126 +                       q++;
22127 +               while (*q)
22128 +                       *p++ = *q++;
22129 +               while (*p)
22130 +                       *p++ = '\0';
22131 +       }
22132 +}
22133 +
22134 +static inline
22135 +int __dx_parse_tag(char *string, tag_t *tag, int remove)
22136 +{
22137 +       substring_t args[MAX_OPT_ARGS];
22138 +       int token, option = 0;
22139 +
22140 +       if (!string)
22141 +               return 0;
22142 +
22143 +       token = match_token(string, tokens, args);
22144 +
22145 +       vxdprintk(VXD_CBIT(tag, 7),
22146 +               "dx_parse_tag(»%s«): %d:#%d",
22147 +               string, token, option);
22148 +
22149 +       switch (token) {
22150 +       case Opt_tag:
22151 +               if (tag)
22152 +                       *tag = 0;
22153 +               if (remove)
22154 +                       __dx_parse_remove(string, "tag");
22155 +               return MNT_TAGID;
22156 +       case Opt_notag:
22157 +               if (remove)
22158 +                       __dx_parse_remove(string, "notag");
22159 +               return MNT_NOTAG;
22160 +       case Opt_notagcheck:
22161 +               if (remove)
22162 +                       __dx_parse_remove(string, "notagcheck");
22163 +               return MNT_NOTAGCHECK;
22164 +       case Opt_tagid:
22165 +               if (tag && !match_int(args, &option))
22166 +                       *tag = option;
22167 +               if (remove)
22168 +                       __dx_parse_remove(string, "tagid");
22169 +               return MNT_TAGID;
22170 +       }
22171 +       return 0;
22172 +}
22173 +
22174 +int dx_parse_tag(char *string, tag_t *tag, int remove)
22175 +{
22176 +       int retval, flags = 0;
22177 +
22178 +       while ((retval = __dx_parse_tag(string, tag, remove)))
22179 +               flags |= retval;
22180 +       return flags;
22181 +}
22182 +
22183 +#ifdef CONFIG_PROPAGATE
22184 +
22185 +void __dx_propagate_tag(struct nameidata *nd, struct inode *inode)
22186 +{
22187 +       tag_t new_tag = 0;
22188 +       struct vfsmount *mnt;
22189 +       int propagate;
22190 +
22191 +       if (!nd)
22192 +               return;
22193 +       mnt = nd->path.mnt;
22194 +       if (!mnt)
22195 +               return;
22196 +
22197 +       propagate = (mnt->mnt_flags & MNT_TAGID);
22198 +       if (propagate)
22199 +               new_tag = mnt->mnt_tag;
22200 +
22201 +       vxdprintk(VXD_CBIT(tag, 7),
22202 +               "dx_propagate_tag(%p[#%lu.%d]): %d,%d",
22203 +               inode, inode->i_ino, inode->i_tag,
22204 +               new_tag, (propagate) ? 1 : 0);
22205 +
22206 +       if (propagate)
22207 +               inode->i_tag = new_tag;
22208 +}
22209 +
22210 +#include <linux/module.h>
22211 +
22212 +EXPORT_SYMBOL_GPL(__dx_propagate_tag);
22213 +
22214 +#endif /* CONFIG_PROPAGATE */
22215 +
22216 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/Kconfig linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/Kconfig
22217 --- linux-2.6.25.11/kernel/vserver/Kconfig      1969-12-31 19:00:00.000000000 -0500
22218 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/Kconfig        2008-07-15 22:46:23.000000000 -0400
22219 @@ -0,0 +1,251 @@
22220 +#
22221 +# Linux VServer configuration
22222 +#
22223 +
22224 +menu "Linux VServer"
22225 +
22226 +config VSERVER_AUTO_LBACK
22227 +       bool    "Automatically Assign Loopback IP"
22228 +       default y
22229 +       help
22230 +         Automatically assign a guest specific loopback
22231 +         IP and add it to the kernel network stack on
22232 +         startup.
22233 +
22234 +config VSERVER_AUTO_SINGLE
22235 +       bool    "Automatic Single IP Special Casing"
22236 +       depends on EXPERIMENTAL
22237 +       default y
22238 +       help
22239 +         This allows network contexts with a single IP to
22240 +         automatically remap 0.0.0.0 bindings to that IP,
22241 +         avoiding further network checks and improving
22242 +         performance.
22243 +
22244 +         (note: such guests do not allow to change the ip
22245 +          on the fly and do not show loopback addresses)
22246 +
22247 +config VSERVER_COWBL
22248 +       bool    "Enable COW Immutable Link Breaking"
22249 +       default y
22250 +       help
22251 +         This enables the COW (Copy-On-Write) link break code.
22252 +         It allows you to treat unified files like normal files
22253 +         when writing to them (which will implicitely break the
22254 +         link and create a copy of the unified file)
22255 +
22256 +config VSERVER_VTIME
22257 +       bool    "Enable Virtualized Guest Time"
22258 +       depends on EXPERIMENTAL
22259 +       default n
22260 +       help
22261 +         This enables per guest time offsets to allow for
22262 +         adjusting the system clock individually per guest.
22263 +         this adds some overhead to the time functions and
22264 +         therefore should not be enabled without good reason.
22265 +
22266 +config VSERVER_DEVICE
22267 +       bool    "Enable Guest Device Mapping"
22268 +       depends on EXPERIMENTAL
22269 +       default n
22270 +       help
22271 +         This enables generic device remapping.
22272 +
22273 +config VSERVER_PROC_SECURE
22274 +       bool    "Enable Proc Security"
22275 +       depends on PROC_FS
22276 +       default y
22277 +       help
22278 +         This configures ProcFS security to initially hide
22279 +         non-process entries for all contexts except the main and
22280 +         spectator context (i.e. for all guests), which is a secure
22281 +         default.
22282 +
22283 +         (note: on 1.2x the entries were visible by default)
22284 +
22285 +config VSERVER_HARDCPU
22286 +       bool    "Enable Hard CPU Limits"
22287 +       default y
22288 +       help
22289 +         Activate the Hard CPU Limits
22290 +
22291 +         This will compile in code that allows the Token Bucket
22292 +         Scheduler to put processes on hold when a context's
22293 +         tokens are depleted (provided that its per-context
22294 +         sched_hard flag is set).
22295 +
22296 +         Processes belonging to that context will not be able
22297 +         to consume CPU resources again until a per-context
22298 +         configured minimum of tokens has been reached.
22299 +
22300 +config VSERVER_IDLETIME
22301 +       bool    "Avoid idle CPUs by skipping Time"
22302 +       depends on VSERVER_HARDCPU
22303 +       default y
22304 +       help
22305 +         This option allows the scheduler to artificially
22306 +         advance time (per cpu) when otherwise the idle
22307 +         task would be scheduled, thus keeping the cpu
22308 +         busy and sharing the available resources among
22309 +         certain contexts.
22310 +
22311 +config VSERVER_IDLELIMIT
22312 +       bool    "Limit the IDLE task"
22313 +       depends on VSERVER_HARDCPU
22314 +       default n
22315 +       help
22316 +         Limit the idle slices, so the the next context
22317 +         will be scheduled as soon as possible.
22318 +
22319 +         This might improve interactivity and latency, but
22320 +         will also marginally increase scheduling overhead.
22321 +
22322 +choice
22323 +       prompt  "Persistent Inode Tagging"
22324 +       default TAGGING_ID24
22325 +       help
22326 +         This adds persistent context information to filesystems
22327 +         mounted with the tagxid option. Tagging is a requirement
22328 +         for per-context disk limits and per-context quota.
22329 +
22330 +
22331 +config TAGGING_NONE
22332 +       bool    "Disabled"
22333 +       help
22334 +         do not store per-context information in inodes.
22335 +
22336 +config TAGGING_UID16
22337 +       bool    "UID16/GID32"
22338 +       help
22339 +         reduces UID to 16 bit, but leaves GID at 32 bit.
22340 +
22341 +config TAGGING_GID16
22342 +       bool    "UID32/GID16"
22343 +       help
22344 +         reduces GID to 16 bit, but leaves UID at 32 bit.
22345 +
22346 +config TAGGING_ID24
22347 +       bool    "UID24/GID24"
22348 +       help
22349 +         uses the upper 8bit from UID and GID for XID tagging
22350 +         which leaves 24bit for UID/GID each, which should be
22351 +         more than sufficient for normal use.
22352 +
22353 +config TAGGING_INTERN
22354 +       bool    "UID32/GID32"
22355 +       help
22356 +         this uses otherwise reserved inode fields in the on
22357 +         disk representation, which limits the use to a few
22358 +         filesystems (currently ext2 and ext3)
22359 +
22360 +endchoice
22361 +
22362 +config TAG_NFSD
22363 +       bool    "Tag NFSD User Auth and Files"
22364 +       default n
22365 +       help
22366 +         Enable this if you do want the in-kernel NFS
22367 +         Server to use the tagging specified above.
22368 +         (will require patched clients too)
22369 +
22370 +config VSERVER_PRIVACY
22371 +       bool    "Honor Privacy Aspects of Guests"
22372 +       default n
22373 +       help
22374 +         When enabled, most context checks will disallow
22375 +         access to structures assigned to a specific context,
22376 +         like ptys or loop devices.
22377 +
22378 +config VSERVER_CONTEXTS
22379 +       int     "Maximum number of Contexts (1-65533)"  if EMBEDDED
22380 +       range 1 65533
22381 +       default "768"   if 64BIT
22382 +       default "256"
22383 +       help
22384 +         This setting will optimize certain data structures
22385 +         and memory allocations according to the expected
22386 +         maximum.
22387 +
22388 +         note: this is not a strict upper limit.
22389 +
22390 +config VSERVER_WARN
22391 +       bool    "VServer Warnings"
22392 +       default y
22393 +       help
22394 +         This enables various runtime warnings, which will
22395 +         notify about potential manipulation attempts or
22396 +         resource shortage. It is generally considered to
22397 +         be a good idea to have that enabled.
22398 +
22399 +config VSERVER_DEBUG
22400 +       bool    "VServer Debugging Code"
22401 +       default n
22402 +       help
22403 +         Set this to yes if you want to be able to activate
22404 +         debugging output at runtime. It adds a very small
22405 +         overhead to all vserver related functions and
22406 +         increases the kernel size by about 20k.
22407 +
22408 +config VSERVER_HISTORY
22409 +       bool    "VServer History Tracing"
22410 +       depends on VSERVER_DEBUG
22411 +       default n
22412 +       help
22413 +         Set this to yes if you want to record the history of
22414 +         linux-vserver activities, so they can be replayed in
22415 +         the event of a kernel panic or oops.
22416 +
22417 +config VSERVER_HISTORY_SIZE
22418 +       int     "Per-CPU History Size (32-65536)"
22419 +       depends on VSERVER_HISTORY
22420 +       range 32 65536
22421 +       default 64
22422 +       help
22423 +         This allows you to specify the number of entries in
22424 +         the per-CPU history buffer.
22425 +
22426 +config VSERVER_MONITOR
22427 +       bool    "VServer Scheduling Monitor"
22428 +       depends on VSERVER_DISABLED
22429 +       default n
22430 +       help
22431 +         Set this to yes if you want to record the scheduling
22432 +         decisions, so that they can be relayed to userspace
22433 +         for detailed analysis.
22434 +
22435 +config VSERVER_MONITOR_SIZE
22436 +       int     "Per-CPU Monitor Queue Size (32-65536)"
22437 +       depends on VSERVER_MONITOR
22438 +       range 32 65536
22439 +       default 1024
22440 +       help
22441 +         This allows you to specify the number of entries in
22442 +         the per-CPU scheduling monitor buffer.
22443 +
22444 +config VSERVER_MONITOR_SYNC
22445 +       int     "Per-CPU Monitor Sync Interval (0-65536)"
22446 +       depends on VSERVER_MONITOR
22447 +       range 0 65536
22448 +       default 256
22449 +       help
22450 +         This allows you to specify the interval in ticks
22451 +         when a time sync entry is inserted.
22452 +
22453 +endmenu
22454 +
22455 +
22456 +config VSERVER
22457 +       bool
22458 +       default y
22459 +       select NAMESPACES
22460 +       select UTS_NS
22461 +       select IPC_NS
22462 +       select USER_NS
22463 +       select SYSVIPC
22464 +
22465 +config VSERVER_SECURITY
22466 +       bool
22467 +       depends on SECURITY
22468 +       default y
22469 +       select SECURITY_CAPABILITIES
22470 +
22471 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/limit.c linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/limit.c
22472 --- linux-2.6.25.11/kernel/vserver/limit.c      1969-12-31 19:00:00.000000000 -0500
22473 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/limit.c        2008-04-19 15:14:52.000000000 -0400
22474 @@ -0,0 +1,319 @@
22475 +/*
22476 + *  linux/kernel/vserver/limit.c
22477 + *
22478 + *  Virtual Server: Context Limits
22479 + *
22480 + *  Copyright (C) 2004-2007  Herbert Pötzl
22481 + *
22482 + *  V0.01  broken out from vcontext V0.05
22483 + *  V0.02  changed vcmds to vxi arg
22484 + *
22485 + */
22486 +
22487 +#include <linux/sched.h>
22488 +#include <linux/module.h>
22489 +#include <linux/vs_limit.h>
22490 +#include <linux/vserver/limit.h>
22491 +#include <linux/vserver/limit_cmd.h>
22492 +
22493 +#include <asm/uaccess.h>
22494 +
22495 +
22496 +const char *vlimit_name[NUM_LIMITS] = {
22497 +       [RLIMIT_CPU]            = "CPU",
22498 +       [RLIMIT_RSS]            = "RSS",
22499 +       [RLIMIT_NPROC]          = "NPROC",
22500 +       [RLIMIT_NOFILE]         = "NOFILE",
22501 +       [RLIMIT_MEMLOCK]        = "VML",
22502 +       [RLIMIT_AS]             = "VM",
22503 +       [RLIMIT_LOCKS]          = "LOCKS",
22504 +       [RLIMIT_SIGPENDING]     = "SIGP",
22505 +       [RLIMIT_MSGQUEUE]       = "MSGQ",
22506 +
22507 +       [VLIMIT_NSOCK]          = "NSOCK",
22508 +       [VLIMIT_OPENFD]         = "OPENFD",
22509 +       [VLIMIT_ANON]           = "ANON",
22510 +       [VLIMIT_SHMEM]          = "SHMEM",
22511 +       [VLIMIT_DENTRY]         = "DENTRY",
22512 +};
22513 +
22514 +EXPORT_SYMBOL_GPL(vlimit_name);
22515 +
22516 +#define MASK_ENTRY(x)  (1 << (x))
22517 +
22518 +const struct vcmd_ctx_rlimit_mask_v0 vlimit_mask = {
22519 +               /* minimum */
22520 +       0
22521 +       ,       /* softlimit */
22522 +       MASK_ENTRY( RLIMIT_RSS          ) |
22523 +       MASK_ENTRY( VLIMIT_ANON         ) |
22524 +       0
22525 +       ,       /* maximum */
22526 +       MASK_ENTRY( RLIMIT_RSS          ) |
22527 +       MASK_ENTRY( RLIMIT_NPROC        ) |
22528 +       MASK_ENTRY( RLIMIT_NOFILE       ) |
22529 +       MASK_ENTRY( RLIMIT_MEMLOCK      ) |
22530 +       MASK_ENTRY( RLIMIT_AS           ) |
22531 +       MASK_ENTRY( RLIMIT_LOCKS        ) |
22532 +       MASK_ENTRY( RLIMIT_MSGQUEUE     ) |
22533 +
22534 +       MASK_ENTRY( VLIMIT_NSOCK        ) |
22535 +       MASK_ENTRY( VLIMIT_OPENFD       ) |
22536 +       MASK_ENTRY( VLIMIT_ANON         ) |
22537 +       MASK_ENTRY( VLIMIT_SHMEM        ) |
22538 +       MASK_ENTRY( VLIMIT_DENTRY       ) |
22539 +       0
22540 +};
22541 +               /* accounting only */
22542 +uint32_t account_mask =
22543 +       MASK_ENTRY( VLIMIT_SEMARY       ) |
22544 +       MASK_ENTRY( VLIMIT_NSEMS        ) |
22545 +       MASK_ENTRY( VLIMIT_MAPPED       ) |
22546 +       0;
22547 +
22548 +
22549 +static int is_valid_vlimit(int id)
22550 +{
22551 +       uint32_t mask = vlimit_mask.minimum |
22552 +               vlimit_mask.softlimit | vlimit_mask.maximum;
22553 +       return mask & (1 << id);
22554 +}
22555 +
22556 +static int is_accounted_vlimit(int id)
22557 +{
22558 +       if (is_valid_vlimit(id))
22559 +               return 1;
22560 +       return account_mask & (1 << id);
22561 +}
22562 +
22563 +
22564 +static inline uint64_t vc_get_soft(struct vx_info *vxi, int id)
22565 +{
22566 +       rlim_t limit = __rlim_soft(&vxi->limit, id);
22567 +       return VX_VLIM(limit);
22568 +}
22569 +
22570 +static inline uint64_t vc_get_hard(struct vx_info *vxi, int id)
22571 +{
22572 +       rlim_t limit = __rlim_hard(&vxi->limit, id);
22573 +       return VX_VLIM(limit);
22574 +}
22575 +
22576 +static int do_get_rlimit(struct vx_info *vxi, uint32_t id,
22577 +       uint64_t *minimum, uint64_t *softlimit, uint64_t *maximum)
22578 +{
22579 +       if (!is_valid_vlimit(id))
22580 +               return -EINVAL;
22581 +
22582 +       if (minimum)
22583 +               *minimum = CRLIM_UNSET;
22584 +       if (softlimit)
22585 +               *softlimit = vc_get_soft(vxi, id);
22586 +       if (maximum)
22587 +               *maximum = vc_get_hard(vxi, id);
22588 +       return 0;
22589 +}
22590 +
22591 +int vc_get_rlimit(struct vx_info *vxi, void __user *data)
22592 +{
22593 +       struct vcmd_ctx_rlimit_v0 vc_data;
22594 +       int ret;
22595 +
22596 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
22597 +               return -EFAULT;
22598 +
22599 +       ret = do_get_rlimit(vxi, vc_data.id,
22600 +               &vc_data.minimum, &vc_data.softlimit, &vc_data.maximum);
22601 +       if (ret)
22602 +               return ret;
22603 +
22604 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
22605 +               return -EFAULT;
22606 +       return 0;
22607 +}
22608 +
22609 +static int do_set_rlimit(struct vx_info *vxi, uint32_t id,
22610 +       uint64_t minimum, uint64_t softlimit, uint64_t maximum)
22611 +{
22612 +       if (!is_valid_vlimit(id))
22613 +               return -EINVAL;
22614 +
22615 +       if (maximum != CRLIM_KEEP)
22616 +               __rlim_hard(&vxi->limit, id) = VX_RLIM(maximum);
22617 +       if (softlimit != CRLIM_KEEP)
22618 +               __rlim_soft(&vxi->limit, id) = VX_RLIM(softlimit);
22619 +
22620 +       /* clamp soft limit */
22621 +       if (__rlim_soft(&vxi->limit, id) > __rlim_hard(&vxi->limit, id))
22622 +               __rlim_soft(&vxi->limit, id) = __rlim_hard(&vxi->limit, id);
22623 +
22624 +       return 0;
22625 +}
22626 +
22627 +int vc_set_rlimit(struct vx_info *vxi, void __user *data)
22628 +{
22629 +       struct vcmd_ctx_rlimit_v0 vc_data;
22630 +
22631 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
22632 +               return -EFAULT;
22633 +
22634 +       return do_set_rlimit(vxi, vc_data.id,
22635 +               vc_data.minimum, vc_data.softlimit, vc_data.maximum);
22636 +}
22637 +
22638 +#ifdef CONFIG_IA32_EMULATION
22639 +
22640 +int vc_set_rlimit_x32(struct vx_info *vxi, void __user *data)
22641 +{
22642 +       struct vcmd_ctx_rlimit_v0_x32 vc_data;
22643 +
22644 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
22645 +               return -EFAULT;
22646 +
22647 +       return do_set_rlimit(vxi, vc_data.id,
22648 +               vc_data.minimum, vc_data.softlimit, vc_data.maximum);
22649 +}
22650 +
22651 +int vc_get_rlimit_x32(struct vx_info *vxi, void __user *data)
22652 +{
22653 +       struct vcmd_ctx_rlimit_v0_x32 vc_data;
22654 +       int ret;
22655 +
22656 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
22657 +               return -EFAULT;
22658 +
22659 +       ret = do_get_rlimit(vxi, vc_data.id,
22660 +               &vc_data.minimum, &vc_data.softlimit, &vc_data.maximum);
22661 +       if (ret)
22662 +               return ret;
22663 +
22664 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
22665 +               return -EFAULT;
22666 +       return 0;
22667 +}
22668 +
22669 +#endif /* CONFIG_IA32_EMULATION */
22670 +
22671 +
22672 +int vc_get_rlimit_mask(uint32_t id, void __user *data)
22673 +{
22674 +       if (copy_to_user(data, &vlimit_mask, sizeof(vlimit_mask)))
22675 +               return -EFAULT;
22676 +       return 0;
22677 +}
22678 +
22679 +
22680 +static inline void vx_reset_minmax(struct _vx_limit *limit)
22681 +{
22682 +       rlim_t value;
22683 +       int lim;
22684 +
22685 +       for (lim = 0; lim < NUM_LIMITS; lim++) {
22686 +               value = __rlim_get(limit, lim);
22687 +               __rlim_rmax(limit, lim) = value;
22688 +               __rlim_rmin(limit, lim) = value;
22689 +       }
22690 +}
22691 +
22692 +
22693 +int vc_reset_minmax(struct vx_info *vxi, void __user *data)
22694 +{
22695 +       vx_reset_minmax(&vxi->limit);
22696 +       return 0;
22697 +}
22698 +
22699 +
22700 +int vc_rlimit_stat(struct vx_info *vxi, void __user *data)
22701 +{
22702 +       struct vcmd_rlimit_stat_v0 vc_data;
22703 +       struct _vx_limit *limit = &vxi->limit;
22704 +       int id;
22705 +
22706 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
22707 +               return -EFAULT;
22708 +
22709 +       id = vc_data.id;
22710 +       if (!is_accounted_vlimit(id))
22711 +               return -EINVAL;
22712 +
22713 +       vx_limit_fixup(limit, id);
22714 +       vc_data.hits = atomic_read(&__rlim_lhit(limit, id));
22715 +       vc_data.value = __rlim_get(limit, id);
22716 +       vc_data.minimum = __rlim_rmin(limit, id);
22717 +       vc_data.maximum = __rlim_rmax(limit, id);
22718 +
22719 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
22720 +               return -EFAULT;
22721 +       return 0;
22722 +}
22723 +
22724 +
22725 +void vx_vsi_meminfo(struct sysinfo *val)
22726 +{
22727 +       struct vx_info *vxi = current->vx_info;
22728 +       unsigned long totalram, freeram;
22729 +       rlim_t v;
22730 +
22731 +       /* we blindly accept the max */
22732 +       v = __rlim_soft(&vxi->limit, RLIMIT_RSS);
22733 +       totalram = (v != RLIM_INFINITY) ? v : val->totalram;
22734 +
22735 +       /* total minus used equals free */
22736 +       v = __vx_cres_array_fixup(&vxi->limit, VLA_RSS);
22737 +       freeram = (v < totalram) ? totalram - v : 0;
22738 +
22739 +       val->totalram = totalram;
22740 +       val->freeram = freeram;
22741 +       val->bufferram = 0;
22742 +       val->totalhigh = 0;
22743 +       val->freehigh = 0;
22744 +       return;
22745 +}
22746 +
22747 +void vx_vsi_swapinfo(struct sysinfo *val)
22748 +{
22749 +       struct vx_info *vxi = current->vx_info;
22750 +       unsigned long totalswap, freeswap;
22751 +       rlim_t v, w;
22752 +
22753 +       v = __rlim_soft(&vxi->limit, RLIMIT_RSS);
22754 +       if (v == RLIM_INFINITY) {
22755 +               val->freeswap = val->totalswap;
22756 +               return;
22757 +       }
22758 +
22759 +       /* we blindly accept the max */
22760 +       w = __rlim_hard(&vxi->limit, RLIMIT_RSS);
22761 +       totalswap = (w != RLIM_INFINITY) ? (w - v) : val->totalswap;
22762 +
22763 +       /* currently 'used' swap */
22764 +       w = __vx_cres_array_fixup(&vxi->limit, VLA_RSS);
22765 +       w -= (w > v) ? v : w;
22766 +
22767 +       /* total minus used equals free */
22768 +       freeswap = (w < totalswap) ? totalswap - w : 0;
22769 +
22770 +       val->totalswap = totalswap;
22771 +       val->freeswap = freeswap;
22772 +       return;
22773 +}
22774 +
22775 +
22776 +unsigned long vx_badness(struct task_struct *task, struct mm_struct *mm)
22777 +{
22778 +       struct vx_info *vxi = mm->mm_vx_info;
22779 +       unsigned long points;
22780 +       rlim_t v, w;
22781 +
22782 +       if (!vxi)
22783 +               return 0;
22784 +
22785 +       points = vxi->vx_badness_bias;
22786 +
22787 +       v = __vx_cres_array_fixup(&vxi->limit, VLA_RSS);
22788 +       w = __rlim_soft(&vxi->limit, RLIMIT_RSS);
22789 +       points += (v > w) ? (v - w) : 0;
22790 +
22791 +       return points;
22792 +}
22793 +
22794 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/limit_init.h linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/limit_init.h
22795 --- linux-2.6.25.11/kernel/vserver/limit_init.h 1969-12-31 19:00:00.000000000 -0500
22796 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/limit_init.h   2008-07-16 00:34:58.000000000 -0400
22797 @@ -0,0 +1,31 @@
22798 +
22799 +
22800 +static inline void vx_info_init_limit(struct _vx_limit *limit)
22801 +{
22802 +       int lim;
22803 +
22804 +       for (lim = 0; lim < NUM_LIMITS; lim++) {
22805 +               __rlim_soft(limit, lim) = RLIM_INFINITY;
22806 +               __rlim_hard(limit, lim) = RLIM_INFINITY;
22807 +               __rlim_set(limit, lim, 0);
22808 +               atomic_set(&__rlim_lhit(limit, lim), 0);
22809 +               __rlim_rmin(limit, lim) = 0;
22810 +               __rlim_rmax(limit, lim) = 0;
22811 +       }
22812 +}
22813 +
22814 +static inline void vx_info_exit_limit(struct _vx_limit *limit)
22815 +{
22816 +       rlim_t value;
22817 +       int lim;
22818 +
22819 +       for (lim = 0; lim < NUM_LIMITS; lim++) {
22820 +               if ((1 << lim) & VLIM_NOCHECK)
22821 +                       continue;
22822 +               value = __rlim_get(limit, lim);
22823 +               vxwprintk_xid(value,
22824 +                       "!!! limit: %p[%s,%d] = %ld on exit.",
22825 +                       limit, vlimit_name[lim], lim, (long)value);
22826 +       }
22827 +}
22828 +
22829 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/limit_proc.h linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/limit_proc.h
22830 --- linux-2.6.25.11/kernel/vserver/limit_proc.h 1969-12-31 19:00:00.000000000 -0500
22831 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/limit_proc.h   2008-04-19 15:14:52.000000000 -0400
22832 @@ -0,0 +1,57 @@
22833 +#ifndef _VX_LIMIT_PROC_H
22834 +#define _VX_LIMIT_PROC_H
22835 +
22836 +#include <linux/vserver/limit_int.h>
22837 +
22838 +
22839 +#define VX_LIMIT_FMT   ":\t%8ld\t%8ld/%8ld\t%8lld/%8lld\t%6d\n"
22840 +#define VX_LIMIT_TOP   \
22841 +       "Limit\t current\t     min/max\t\t    soft/hard\t\thits\n"
22842 +
22843 +#define VX_LIMIT_ARG(r)                                \
22844 +       (unsigned long)__rlim_get(limit, r),    \
22845 +       (unsigned long)__rlim_rmin(limit, r),   \
22846 +       (unsigned long)__rlim_rmax(limit, r),   \
22847 +       VX_VLIM(__rlim_soft(limit, r)),         \
22848 +       VX_VLIM(__rlim_hard(limit, r)),         \
22849 +       atomic_read(&__rlim_lhit(limit, r))
22850 +
22851 +static inline int vx_info_proc_limit(struct _vx_limit *limit, char *buffer)
22852 +{
22853 +       vx_limit_fixup(limit, -1);
22854 +       return sprintf(buffer, VX_LIMIT_TOP
22855 +               "PROC"  VX_LIMIT_FMT
22856 +               "VM"    VX_LIMIT_FMT
22857 +               "VML"   VX_LIMIT_FMT
22858 +               "RSS"   VX_LIMIT_FMT
22859 +               "ANON"  VX_LIMIT_FMT
22860 +               "RMAP"  VX_LIMIT_FMT
22861 +               "FILES" VX_LIMIT_FMT
22862 +               "OFD"   VX_LIMIT_FMT
22863 +               "LOCKS" VX_LIMIT_FMT
22864 +               "SOCK"  VX_LIMIT_FMT
22865 +               "MSGQ"  VX_LIMIT_FMT
22866 +               "SHM"   VX_LIMIT_FMT
22867 +               "SEMA"  VX_LIMIT_FMT
22868 +               "SEMS"  VX_LIMIT_FMT
22869 +               "DENT"  VX_LIMIT_FMT,
22870 +               VX_LIMIT_ARG(RLIMIT_NPROC),
22871 +               VX_LIMIT_ARG(RLIMIT_AS),
22872 +               VX_LIMIT_ARG(RLIMIT_MEMLOCK),
22873 +               VX_LIMIT_ARG(RLIMIT_RSS),
22874 +               VX_LIMIT_ARG(VLIMIT_ANON),
22875 +               VX_LIMIT_ARG(VLIMIT_MAPPED),
22876 +               VX_LIMIT_ARG(RLIMIT_NOFILE),
22877 +               VX_LIMIT_ARG(VLIMIT_OPENFD),
22878 +               VX_LIMIT_ARG(RLIMIT_LOCKS),
22879 +               VX_LIMIT_ARG(VLIMIT_NSOCK),
22880 +               VX_LIMIT_ARG(RLIMIT_MSGQUEUE),
22881 +               VX_LIMIT_ARG(VLIMIT_SHMEM),
22882 +               VX_LIMIT_ARG(VLIMIT_SEMARY),
22883 +               VX_LIMIT_ARG(VLIMIT_NSEMS),
22884 +               VX_LIMIT_ARG(VLIMIT_DENTRY));
22885 +}
22886 +
22887 +#endif /* _VX_LIMIT_PROC_H */
22888 +
22889 +
22890 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/Makefile linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/Makefile
22891 --- linux-2.6.25.11/kernel/vserver/Makefile     1969-12-31 19:00:00.000000000 -0500
22892 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/Makefile       2008-07-11 03:07:10.000000000 -0400
22893 @@ -0,0 +1,18 @@
22894 +#
22895 +# Makefile for the Linux vserver routines.
22896 +#
22897 +
22898 +
22899 +obj-y          += vserver.o
22900 +
22901 +vserver-y      := switch.o context.o space.o sched.o network.o inode.o \
22902 +                  limit.o cvirt.o cacct.o signal.o helper.o init.o \
22903 +                  dlimit.o tag.o
22904 +
22905 +vserver-$(CONFIG_INET) += inet.o
22906 +vserver-$(CONFIG_PROC_FS) += proc.o
22907 +vserver-$(CONFIG_VSERVER_DEBUG) += sysctl.o debug.o
22908 +vserver-$(CONFIG_VSERVER_HISTORY) += history.o
22909 +vserver-$(CONFIG_VSERVER_MONITOR) += monitor.o
22910 +vserver-$(CONFIG_VSERVER_DEVICE) += device.o
22911 +
22912 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/monitor.c linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/monitor.c
22913 --- linux-2.6.25.11/kernel/vserver/monitor.c    1969-12-31 19:00:00.000000000 -0500
22914 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/monitor.c      2008-04-19 15:14:52.000000000 -0400
22915 @@ -0,0 +1,138 @@
22916 +/*
22917 + *  kernel/vserver/monitor.c
22918 + *
22919 + *  Virtual Context Scheduler Monitor
22920 + *
22921 + *  Copyright (C) 2006-2007 Herbert Pötzl
22922 + *
22923 + *  V0.01  basic design
22924 + *
22925 + */
22926 +
22927 +#include <linux/module.h>
22928 +#include <linux/jiffies.h>
22929 +#include <asm/uaccess.h>
22930 +#include <asm/atomic.h>
22931 +
22932 +#include <linux/vserver/monitor.h>
22933 +#include <linux/vserver/debug_cmd.h>
22934 +
22935 +
22936 +#ifdef CONFIG_VSERVER_MONITOR
22937 +#define VXM_SIZE       CONFIG_VSERVER_MONITOR_SIZE
22938 +#else
22939 +#define VXM_SIZE       64
22940 +#endif
22941 +
22942 +struct _vx_monitor {
22943 +       unsigned int counter;
22944 +
22945 +       struct _vx_mon_entry entry[VXM_SIZE+1];
22946 +};
22947 +
22948 +
22949 +DEFINE_PER_CPU(struct _vx_monitor, vx_monitor_buffer);
22950 +
22951 +unsigned volatile int vxm_active = 1;
22952 +
22953 +static atomic_t sequence = ATOMIC_INIT(0);
22954 +
22955 +
22956 +/*     vxm_advance()
22957 +
22958 +       * requires disabled preemption                          */
22959 +
22960 +struct _vx_mon_entry *vxm_advance(int cpu)
22961 +{
22962 +       struct _vx_monitor *mon = &per_cpu(vx_monitor_buffer, cpu);
22963 +       struct _vx_mon_entry *entry;
22964 +       unsigned int index;
22965 +
22966 +       index = vxm_active ? (mon->counter++ % VXM_SIZE) : VXM_SIZE;
22967 +       entry = &mon->entry[index];
22968 +
22969 +       entry->ev.seq = atomic_inc_return(&sequence);
22970 +       entry->ev.jif = jiffies;
22971 +       return entry;
22972 +}
22973 +
22974 +EXPORT_SYMBOL_GPL(vxm_advance);
22975 +
22976 +
22977 +int do_read_monitor(struct __user _vx_mon_entry *data,
22978 +       int cpu, uint32_t *index, uint32_t *count)
22979 +{
22980 +       int pos, ret = 0;
22981 +       struct _vx_monitor *mon = &per_cpu(vx_monitor_buffer, cpu);
22982 +       int end = mon->counter;
22983 +       int start = end - VXM_SIZE + 2;
22984 +       int idx = *index;
22985 +
22986 +       /* special case: get current pos */
22987 +       if (!*count) {
22988 +               *index = end;
22989 +               return 0;
22990 +       }
22991 +
22992 +       /* have we lost some data? */
22993 +       if (idx < start)
22994 +               idx = start;
22995 +
22996 +       for (pos = 0; (pos < *count) && (idx < end); pos++, idx++) {
22997 +               struct _vx_mon_entry *entry =
22998 +                       &mon->entry[idx % VXM_SIZE];
22999 +
23000 +               /* send entry to userspace */
23001 +               ret = copy_to_user(&data[pos], entry, sizeof(*entry));
23002 +               if (ret)
23003 +                       break;
23004 +       }
23005 +       /* save new index and count */
23006 +       *index = idx;
23007 +       *count = pos;
23008 +       return ret ? ret : (*index < end);
23009 +}
23010 +
23011 +int vc_read_monitor(uint32_t id, void __user *data)
23012 +{
23013 +       struct vcmd_read_monitor_v0 vc_data;
23014 +       int ret;
23015 +
23016 +       if (id >= NR_CPUS)
23017 +               return -EINVAL;
23018 +
23019 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
23020 +               return -EFAULT;
23021 +
23022 +       ret = do_read_monitor((struct __user _vx_mon_entry *)vc_data.data,
23023 +               id, &vc_data.index, &vc_data.count);
23024 +
23025 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
23026 +               return -EFAULT;
23027 +       return ret;
23028 +}
23029 +
23030 +#ifdef CONFIG_COMPAT
23031 +
23032 +int vc_read_monitor_x32(uint32_t id, void __user *data)
23033 +{
23034 +       struct vcmd_read_monitor_v0_x32 vc_data;
23035 +       int ret;
23036 +
23037 +       if (id >= NR_CPUS)
23038 +               return -EINVAL;
23039 +
23040 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
23041 +               return -EFAULT;
23042 +
23043 +       ret = do_read_monitor((struct __user _vx_mon_entry *)
23044 +               compat_ptr(vc_data.data_ptr),
23045 +               id, &vc_data.index, &vc_data.count);
23046 +
23047 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
23048 +               return -EFAULT;
23049 +       return ret;
23050 +}
23051 +
23052 +#endif /* CONFIG_COMPAT */
23053 +
23054 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/network.c linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/network.c
23055 --- linux-2.6.25.11/kernel/vserver/network.c    1969-12-31 19:00:00.000000000 -0500
23056 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/network.c      2008-04-19 15:14:52.000000000 -0400
23057 @@ -0,0 +1,864 @@
23058 +/*
23059 + *  linux/kernel/vserver/network.c
23060 + *
23061 + *  Virtual Server: Network Support
23062 + *
23063 + *  Copyright (C) 2003-2007  Herbert Pötzl
23064 + *
23065 + *  V0.01  broken out from vcontext V0.05
23066 + *  V0.02  cleaned up implementation
23067 + *  V0.03  added equiv nx commands
23068 + *  V0.04  switch to RCU based hash
23069 + *  V0.05  and back to locking again
23070 + *  V0.06  changed vcmds to nxi arg
23071 + *  V0.07  have __create claim() the nxi
23072 + *
23073 + */
23074 +
23075 +#include <linux/err.h>
23076 +#include <linux/slab.h>
23077 +#include <linux/rcupdate.h>
23078 +
23079 +#include <linux/vs_network.h>
23080 +#include <linux/vs_pid.h>
23081 +#include <linux/vserver/network_cmd.h>
23082 +
23083 +
23084 +atomic_t nx_global_ctotal      = ATOMIC_INIT(0);
23085 +atomic_t nx_global_cactive     = ATOMIC_INIT(0);
23086 +
23087 +static struct kmem_cache *nx_addr_v4_cachep = NULL;
23088 +static struct kmem_cache *nx_addr_v6_cachep = NULL;
23089 +
23090 +
23091 +static int __init init_network(void)
23092 +{
23093 +       nx_addr_v4_cachep = kmem_cache_create("nx_v4_addr_cache",
23094 +               sizeof(struct nx_addr_v4), 0,
23095 +               SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
23096 +       nx_addr_v6_cachep = kmem_cache_create("nx_v6_addr_cache",
23097 +               sizeof(struct nx_addr_v6), 0,
23098 +               SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
23099 +       return 0;
23100 +}
23101 +
23102 +
23103 +/*     __alloc_nx_addr_v4()                                    */
23104 +
23105 +static inline struct nx_addr_v4 *__alloc_nx_addr_v4(void)
23106 +{
23107 +       struct nx_addr_v4 *nxa = kmem_cache_alloc(
23108 +               nx_addr_v4_cachep, GFP_KERNEL);
23109 +
23110 +       if (!IS_ERR(nxa))
23111 +               memset(nxa, 0, sizeof(*nxa));
23112 +       return nxa;
23113 +}
23114 +
23115 +/*     __dealloc_nx_addr_v4()                                  */
23116 +
23117 +static inline void __dealloc_nx_addr_v4(struct nx_addr_v4 *nxa)
23118 +{
23119 +       kmem_cache_free(nx_addr_v4_cachep, nxa);
23120 +}
23121 +
23122 +/*     __dealloc_nx_addr_v4_all()                              */
23123 +
23124 +static inline void __dealloc_nx_addr_v4_all(struct nx_addr_v4 *nxa)
23125 +{
23126 +       while (nxa) {
23127 +               struct nx_addr_v4 *next = nxa->next;
23128 +
23129 +               __dealloc_nx_addr_v4(nxa);
23130 +               nxa = next;
23131 +       }
23132 +}
23133 +
23134 +
23135 +#ifdef CONFIG_IPV6
23136 +
23137 +/*     __alloc_nx_addr_v6()                                    */
23138 +
23139 +static inline struct nx_addr_v6 *__alloc_nx_addr_v6(void)
23140 +{
23141 +       struct nx_addr_v6 *nxa = kmem_cache_alloc(
23142 +               nx_addr_v6_cachep, GFP_KERNEL);
23143 +
23144 +       if (!IS_ERR(nxa))
23145 +               memset(nxa, 0, sizeof(*nxa));
23146 +       return nxa;
23147 +}
23148 +
23149 +/*     __dealloc_nx_addr_v6()                                  */
23150 +
23151 +static inline void __dealloc_nx_addr_v6(struct nx_addr_v6 *nxa)
23152 +{
23153 +       kmem_cache_free(nx_addr_v6_cachep, nxa);
23154 +}
23155 +
23156 +/*     __dealloc_nx_addr_v6_all()                              */
23157 +
23158 +static inline void __dealloc_nx_addr_v6_all(struct nx_addr_v6 *nxa)
23159 +{
23160 +       while (nxa) {
23161 +               struct nx_addr_v6 *next = nxa->next;
23162 +
23163 +               __dealloc_nx_addr_v6(nxa);
23164 +               nxa = next;
23165 +       }
23166 +}
23167 +
23168 +#endif /* CONFIG_IPV6 */
23169 +
23170 +/*     __alloc_nx_info()
23171 +
23172 +       * allocate an initialized nx_info struct
23173 +       * doesn't make it visible (hash)                        */
23174 +
23175 +static struct nx_info *__alloc_nx_info(nid_t nid)
23176 +{
23177 +       struct nx_info *new = NULL;
23178 +
23179 +       vxdprintk(VXD_CBIT(nid, 1), "alloc_nx_info(%d)*", nid);
23180 +
23181 +       /* would this benefit from a slab cache? */
23182 +       new = kmalloc(sizeof(struct nx_info), GFP_KERNEL);
23183 +       if (!new)
23184 +               return 0;
23185 +
23186 +       memset(new, 0, sizeof(struct nx_info));
23187 +       new->nx_id = nid;
23188 +       INIT_HLIST_NODE(&new->nx_hlist);
23189 +       atomic_set(&new->nx_usecnt, 0);
23190 +       atomic_set(&new->nx_tasks, 0);
23191 +       new->nx_state = 0;
23192 +
23193 +       new->nx_flags = NXF_INIT_SET;
23194 +
23195 +       /* rest of init goes here */
23196 +
23197 +       new->v4_lback.s_addr = htonl(INADDR_LOOPBACK);
23198 +       new->v4_bcast.s_addr = htonl(INADDR_BROADCAST);
23199 +
23200 +       vxdprintk(VXD_CBIT(nid, 0),
23201 +               "alloc_nx_info(%d) = %p", nid, new);
23202 +       atomic_inc(&nx_global_ctotal);
23203 +       return new;
23204 +}
23205 +
23206 +/*     __dealloc_nx_info()
23207 +
23208 +       * final disposal of nx_info                             */
23209 +
23210 +static void __dealloc_nx_info(struct nx_info *nxi)
23211 +{
23212 +       vxdprintk(VXD_CBIT(nid, 0),
23213 +               "dealloc_nx_info(%p)", nxi);
23214 +
23215 +       nxi->nx_hlist.next = LIST_POISON1;
23216 +       nxi->nx_id = -1;
23217 +
23218 +       BUG_ON(atomic_read(&nxi->nx_usecnt));
23219 +       BUG_ON(atomic_read(&nxi->nx_tasks));
23220 +
23221 +       __dealloc_nx_addr_v4_all(nxi->v4.next);
23222 +
23223 +       nxi->nx_state |= NXS_RELEASED;
23224 +       kfree(nxi);
23225 +       atomic_dec(&nx_global_ctotal);
23226 +}
23227 +
23228 +static void __shutdown_nx_info(struct nx_info *nxi)
23229 +{
23230 +       nxi->nx_state |= NXS_SHUTDOWN;
23231 +       vs_net_change(nxi, VSC_NETDOWN);
23232 +}
23233 +
23234 +/*     exported stuff                                          */
23235 +
23236 +void free_nx_info(struct nx_info *nxi)
23237 +{
23238 +       /* context shutdown is mandatory */
23239 +       BUG_ON(nxi->nx_state != NXS_SHUTDOWN);
23240 +
23241 +       /* context must not be hashed */
23242 +       BUG_ON(nxi->nx_state & NXS_HASHED);
23243 +
23244 +       BUG_ON(atomic_read(&nxi->nx_usecnt));
23245 +       BUG_ON(atomic_read(&nxi->nx_tasks));
23246 +
23247 +       __dealloc_nx_info(nxi);
23248 +}
23249 +
23250 +
23251 +void __nx_set_lback(struct nx_info *nxi)
23252 +{
23253 +       int nid = nxi->nx_id;
23254 +       __be32 lback = htonl(INADDR_LOOPBACK ^ ((nid & 0xFFFF) << 8));
23255 +
23256 +       nxi->v4_lback.s_addr = lback;
23257 +}
23258 +
23259 +extern int __nx_inet_add_lback(__be32 addr);
23260 +extern int __nx_inet_del_lback(__be32 addr);
23261 +
23262 +
23263 +/*     hash table for nx_info hash */
23264 +
23265 +#define NX_HASH_SIZE   13
23266 +
23267 +struct hlist_head nx_info_hash[NX_HASH_SIZE];
23268 +
23269 +static spinlock_t nx_info_hash_lock = SPIN_LOCK_UNLOCKED;
23270 +
23271 +
23272 +static inline unsigned int __hashval(nid_t nid)
23273 +{
23274 +       return (nid % NX_HASH_SIZE);
23275 +}
23276 +
23277 +
23278 +
23279 +/*     __hash_nx_info()
23280 +
23281 +       * add the nxi to the global hash table
23282 +       * requires the hash_lock to be held                     */
23283 +
23284 +static inline void __hash_nx_info(struct nx_info *nxi)
23285 +{
23286 +       struct hlist_head *head;
23287 +
23288 +       vxd_assert_lock(&nx_info_hash_lock);
23289 +       vxdprintk(VXD_CBIT(nid, 4),
23290 +               "__hash_nx_info: %p[#%d]", nxi, nxi->nx_id);
23291 +
23292 +       /* context must not be hashed */
23293 +       BUG_ON(nx_info_state(nxi, NXS_HASHED));
23294 +
23295 +       nxi->nx_state |= NXS_HASHED;
23296 +       head = &nx_info_hash[__hashval(nxi->nx_id)];
23297 +       hlist_add_head(&nxi->nx_hlist, head);
23298 +       atomic_inc(&nx_global_cactive);
23299 +}
23300 +
23301 +/*     __unhash_nx_info()
23302 +
23303 +       * remove the nxi from the global hash table
23304 +       * requires the hash_lock to be held                     */
23305 +
23306 +static inline void __unhash_nx_info(struct nx_info *nxi)
23307 +{
23308 +       vxd_assert_lock(&nx_info_hash_lock);
23309 +       vxdprintk(VXD_CBIT(nid, 4),
23310 +               "__unhash_nx_info: %p[#%d.%d.%d]", nxi, nxi->nx_id,
23311 +               atomic_read(&nxi->nx_usecnt), atomic_read(&nxi->nx_tasks));
23312 +
23313 +       /* context must be hashed */
23314 +       BUG_ON(!nx_info_state(nxi, NXS_HASHED));
23315 +       /* but without tasks */
23316 +       BUG_ON(atomic_read(&nxi->nx_tasks));
23317 +
23318 +       nxi->nx_state &= ~NXS_HASHED;
23319 +       hlist_del(&nxi->nx_hlist);
23320 +       atomic_dec(&nx_global_cactive);
23321 +}
23322 +
23323 +
23324 +/*     __lookup_nx_info()
23325 +
23326 +       * requires the hash_lock to be held
23327 +       * doesn't increment the nx_refcnt                       */
23328 +
23329 +static inline struct nx_info *__lookup_nx_info(nid_t nid)
23330 +{
23331 +       struct hlist_head *head = &nx_info_hash[__hashval(nid)];
23332 +       struct hlist_node *pos;
23333 +       struct nx_info *nxi;
23334 +
23335 +       vxd_assert_lock(&nx_info_hash_lock);
23336 +       hlist_for_each(pos, head) {
23337 +               nxi = hlist_entry(pos, struct nx_info, nx_hlist);
23338 +
23339 +               if (nxi->nx_id == nid)
23340 +                       goto found;
23341 +       }
23342 +       nxi = NULL;
23343 +found:
23344 +       vxdprintk(VXD_CBIT(nid, 0),
23345 +               "__lookup_nx_info(#%u): %p[#%u]",
23346 +               nid, nxi, nxi ? nxi->nx_id : 0);
23347 +       return nxi;
23348 +}
23349 +
23350 +
23351 +/*     __create_nx_info()
23352 +
23353 +       * create the requested context
23354 +       * get(), claim() and hash it                            */
23355 +
23356 +static struct nx_info *__create_nx_info(int id)
23357 +{
23358 +       struct nx_info *new, *nxi = NULL;
23359 +
23360 +       vxdprintk(VXD_CBIT(nid, 1), "create_nx_info(%d)*", id);
23361 +
23362 +       if (!(new = __alloc_nx_info(id)))
23363 +               return ERR_PTR(-ENOMEM);
23364 +
23365 +       /* required to make dynamic xids unique */
23366 +       spin_lock(&nx_info_hash_lock);
23367 +
23368 +       /* static context requested */
23369 +       if ((nxi = __lookup_nx_info(id))) {
23370 +               vxdprintk(VXD_CBIT(nid, 0),
23371 +                       "create_nx_info(%d) = %p (already there)", id, nxi);
23372 +               if (nx_info_flags(nxi, NXF_STATE_SETUP, 0))
23373 +                       nxi = ERR_PTR(-EBUSY);
23374 +               else
23375 +                       nxi = ERR_PTR(-EEXIST);
23376 +               goto out_unlock;
23377 +       }
23378 +       /* new context */
23379 +       vxdprintk(VXD_CBIT(nid, 0),
23380 +               "create_nx_info(%d) = %p (new)", id, new);
23381 +       claim_nx_info(new, NULL);
23382 +       __nx_set_lback(new);
23383 +       __hash_nx_info(get_nx_info(new));
23384 +       nxi = new, new = NULL;
23385 +
23386 +out_unlock:
23387 +       spin_unlock(&nx_info_hash_lock);
23388 +       if (new)
23389 +               __dealloc_nx_info(new);
23390 +       return nxi;
23391 +}
23392 +
23393 +
23394 +
23395 +/*     exported stuff                                          */
23396 +
23397 +
23398 +void unhash_nx_info(struct nx_info *nxi)
23399 +{
23400 +       __shutdown_nx_info(nxi);
23401 +       spin_lock(&nx_info_hash_lock);
23402 +       __unhash_nx_info(nxi);
23403 +       spin_unlock(&nx_info_hash_lock);
23404 +}
23405 +
23406 +/*     lookup_nx_info()
23407 +
23408 +       * search for a nx_info and get() it
23409 +       * negative id means current                             */
23410 +
23411 +struct nx_info *lookup_nx_info(int id)
23412 +{
23413 +       struct nx_info *nxi = NULL;
23414 +
23415 +       if (id < 0) {
23416 +               nxi = get_nx_info(current->nx_info);
23417 +       } else if (id > 1) {
23418 +               spin_lock(&nx_info_hash_lock);
23419 +               nxi = get_nx_info(__lookup_nx_info(id));
23420 +               spin_unlock(&nx_info_hash_lock);
23421 +       }
23422 +       return nxi;
23423 +}
23424 +
23425 +/*     nid_is_hashed()
23426 +
23427 +       * verify that nid is still hashed                       */
23428 +
23429 +int nid_is_hashed(nid_t nid)
23430 +{
23431 +       int hashed;
23432 +
23433 +       spin_lock(&nx_info_hash_lock);
23434 +       hashed = (__lookup_nx_info(nid) != NULL);
23435 +       spin_unlock(&nx_info_hash_lock);
23436 +       return hashed;
23437 +}
23438 +
23439 +
23440 +#ifdef CONFIG_PROC_FS
23441 +
23442 +/*     get_nid_list()
23443 +
23444 +       * get a subset of hashed nids for proc
23445 +       * assumes size is at least one                          */
23446 +
23447 +int get_nid_list(int index, unsigned int *nids, int size)
23448 +{
23449 +       int hindex, nr_nids = 0;
23450 +
23451 +       /* only show current and children */
23452 +       if (!nx_check(0, VS_ADMIN | VS_WATCH)) {
23453 +               if (index > 0)
23454 +                       return 0;
23455 +               nids[nr_nids] = nx_current_nid();
23456 +               return 1;
23457 +       }
23458 +
23459 +       for (hindex = 0; hindex < NX_HASH_SIZE; hindex++) {
23460 +               struct hlist_head *head = &nx_info_hash[hindex];
23461 +               struct hlist_node *pos;
23462 +
23463 +               spin_lock(&nx_info_hash_lock);
23464 +               hlist_for_each(pos, head) {
23465 +                       struct nx_info *nxi;
23466 +
23467 +                       if (--index > 0)
23468 +                               continue;
23469 +
23470 +                       nxi = hlist_entry(pos, struct nx_info, nx_hlist);
23471 +                       nids[nr_nids] = nxi->nx_id;
23472 +                       if (++nr_nids >= size) {
23473 +                               spin_unlock(&nx_info_hash_lock);
23474 +                               goto out;
23475 +                       }
23476 +               }
23477 +               /* keep the lock time short */
23478 +               spin_unlock(&nx_info_hash_lock);
23479 +       }
23480 +out:
23481 +       return nr_nids;
23482 +}
23483 +#endif
23484 +
23485 +
23486 +/*
23487 + *     migrate task to new network
23488 + *     gets nxi, puts old_nxi on change
23489 + */
23490 +
23491 +int nx_migrate_task(struct task_struct *p, struct nx_info *nxi)
23492 +{
23493 +       struct nx_info *old_nxi;
23494 +       int ret = 0;
23495 +
23496 +       if (!p || !nxi)
23497 +               BUG();
23498 +
23499 +       vxdprintk(VXD_CBIT(nid, 5),
23500 +               "nx_migrate_task(%p,%p[#%d.%d.%d])",
23501 +               p, nxi, nxi->nx_id,
23502 +               atomic_read(&nxi->nx_usecnt),
23503 +               atomic_read(&nxi->nx_tasks));
23504 +
23505 +       if (nx_info_flags(nxi, NXF_INFO_PRIVATE, 0) &&
23506 +               !nx_info_flags(nxi, NXF_STATE_SETUP, 0))
23507 +               return -EACCES;
23508 +
23509 +       if (nx_info_state(nxi, NXS_SHUTDOWN))
23510 +               return -EFAULT;
23511 +
23512 +       /* maybe disallow this completely? */
23513 +       old_nxi = task_get_nx_info(p);
23514 +       if (old_nxi == nxi)
23515 +               goto out;
23516 +
23517 +       task_lock(p);
23518 +       if (old_nxi)
23519 +               clr_nx_info(&p->nx_info);
23520 +       claim_nx_info(nxi, p);
23521 +       set_nx_info(&p->nx_info, nxi);
23522 +       p->nid = nxi->nx_id;
23523 +       task_unlock(p);
23524 +
23525 +       vxdprintk(VXD_CBIT(nid, 5),
23526 +               "moved task %p into nxi:%p[#%d]",
23527 +               p, nxi, nxi->nx_id);
23528 +
23529 +       if (old_nxi)
23530 +               release_nx_info(old_nxi, p);
23531 +       ret = 0;
23532 +out:
23533 +       put_nx_info(old_nxi);
23534 +       return ret;
23535 +}
23536 +
23537 +
23538 +void nx_set_persistent(struct nx_info *nxi)
23539 +{
23540 +       vxdprintk(VXD_CBIT(nid, 6),
23541 +               "nx_set_persistent(%p[#%d])", nxi, nxi->nx_id);
23542 +
23543 +       get_nx_info(nxi);
23544 +       claim_nx_info(nxi, NULL);
23545 +}
23546 +
23547 +void nx_clear_persistent(struct nx_info *nxi)
23548 +{
23549 +       vxdprintk(VXD_CBIT(nid, 6),
23550 +               "nx_clear_persistent(%p[#%d])", nxi, nxi->nx_id);
23551 +
23552 +       release_nx_info(nxi, NULL);
23553 +       put_nx_info(nxi);
23554 +}
23555 +
23556 +void nx_update_persistent(struct nx_info *nxi)
23557 +{
23558 +       if (nx_info_flags(nxi, NXF_PERSISTENT, 0))
23559 +               nx_set_persistent(nxi);
23560 +       else
23561 +               nx_clear_persistent(nxi);
23562 +}
23563 +
23564 +/* vserver syscall commands below here */
23565 +
23566 +/* taks nid and nx_info functions */
23567 +
23568 +#include <asm/uaccess.h>
23569 +
23570 +
23571 +int vc_task_nid(uint32_t id)
23572 +{
23573 +       nid_t nid;
23574 +
23575 +       if (id) {
23576 +               struct task_struct *tsk;
23577 +
23578 +               read_lock(&tasklist_lock);
23579 +               tsk = find_task_by_real_pid(id);
23580 +               nid = (tsk) ? tsk->nid : -ESRCH;
23581 +               read_unlock(&tasklist_lock);
23582 +       } else
23583 +               nid = nx_current_nid();
23584 +       return nid;
23585 +}
23586 +
23587 +
23588 +int vc_nx_info(struct nx_info *nxi, void __user *data)
23589 +{
23590 +       struct vcmd_nx_info_v0 vc_data;
23591 +
23592 +       vc_data.nid = nxi->nx_id;
23593 +
23594 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
23595 +               return -EFAULT;
23596 +       return 0;
23597 +}
23598 +
23599 +
23600 +/* network functions */
23601 +
23602 +int vc_net_create(uint32_t nid, void __user *data)
23603 +{
23604 +       struct vcmd_net_create vc_data = { .flagword = NXF_INIT_SET };
23605 +       struct nx_info *new_nxi;
23606 +       int ret;
23607 +
23608 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
23609 +               return -EFAULT;
23610 +
23611 +       if ((nid > MAX_S_CONTEXT) || (nid < 2))
23612 +               return -EINVAL;
23613 +
23614 +       new_nxi = __create_nx_info(nid);
23615 +       if (IS_ERR(new_nxi))
23616 +               return PTR_ERR(new_nxi);
23617 +
23618 +       /* initial flags */
23619 +       new_nxi->nx_flags = vc_data.flagword;
23620 +
23621 +       ret = -ENOEXEC;
23622 +       if (vs_net_change(new_nxi, VSC_NETUP))
23623 +               goto out;
23624 +
23625 +       ret = nx_migrate_task(current, new_nxi);
23626 +       if (ret)
23627 +               goto out;
23628 +
23629 +       /* return context id on success */
23630 +       ret = new_nxi->nx_id;
23631 +
23632 +       /* get a reference for persistent contexts */
23633 +       if ((vc_data.flagword & NXF_PERSISTENT))
23634 +               nx_set_persistent(new_nxi);
23635 +out:
23636 +       release_nx_info(new_nxi, NULL);
23637 +       put_nx_info(new_nxi);
23638 +       return ret;
23639 +}
23640 +
23641 +
23642 +int vc_net_migrate(struct nx_info *nxi, void __user *data)
23643 +{
23644 +       return nx_migrate_task(current, nxi);
23645 +}
23646 +
23647 +
23648 +
23649 +int do_add_v4_addr(struct nx_info *nxi, __be32 ip, __be32 ip2, __be32 mask,
23650 +       uint16_t type, uint16_t flags)
23651 +{
23652 +       struct nx_addr_v4 *nxa = &nxi->v4;
23653 +
23654 +       if (NX_IPV4(nxi)) {
23655 +               /* locate last entry */
23656 +               for (; nxa->next; nxa = nxa->next);
23657 +               nxa->next = __alloc_nx_addr_v4();
23658 +               nxa = nxa->next;
23659 +
23660 +               if (IS_ERR(nxa))
23661 +                       return PTR_ERR(nxa);
23662 +       }
23663 +
23664 +       if (nxi->v4.next)
23665 +               /* remove single ip for ip list */
23666 +               nxi->nx_flags &= ~NXF_SINGLE_IP;
23667 +
23668 +       nxa->ip[0].s_addr = ip;
23669 +       nxa->ip[1].s_addr = ip2;
23670 +       nxa->mask.s_addr = mask;
23671 +       nxa->type = type;
23672 +       nxa->flags = flags;
23673 +       return 0;
23674 +}
23675 +
23676 +
23677 +int vc_net_add(struct nx_info *nxi, void __user *data)
23678 +{
23679 +       struct vcmd_net_addr_v0 vc_data;
23680 +       int index, ret = 0;
23681 +
23682 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
23683 +               return -EFAULT;
23684 +
23685 +       switch (vc_data.type) {
23686 +       case NXA_TYPE_IPV4:
23687 +               if ((vc_data.count < 1) || (vc_data.count > 4))
23688 +                       return -EINVAL;
23689 +
23690 +               index = 0;
23691 +               while (index < vc_data.count) {
23692 +                       ret = do_add_v4_addr(nxi, vc_data.ip[index].s_addr, 0,
23693 +                               vc_data.mask[index].s_addr, NXA_TYPE_ADDR, 0);
23694 +                       if (ret)
23695 +                               return ret;
23696 +                       index++;
23697 +               }
23698 +               ret = index;
23699 +               break;
23700 +
23701 +       case NXA_TYPE_IPV4|NXA_MOD_BCAST:
23702 +               nxi->v4_bcast = vc_data.ip[0];
23703 +               ret = 1;
23704 +               break;
23705 +
23706 +       case NXA_TYPE_IPV4|NXA_MOD_LBACK:
23707 +               nxi->v4_lback = vc_data.ip[0];
23708 +               ret = 1;
23709 +               break;
23710 +
23711 +       default:
23712 +               ret = -EINVAL;
23713 +               break;
23714 +       }
23715 +       return ret;
23716 +}
23717 +
23718 +int vc_net_remove(struct nx_info *nxi, void __user *data)
23719 +{
23720 +       struct vcmd_net_addr_v0 vc_data;
23721 +
23722 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
23723 +               return -EFAULT;
23724 +
23725 +       switch (vc_data.type) {
23726 +       case NXA_TYPE_ANY:
23727 +               __dealloc_nx_addr_v4_all(xchg(&nxi->v4.next, NULL));
23728 +               memset(&nxi->v4, 0, sizeof(nxi->v4));
23729 +               break;
23730 +
23731 +       default:
23732 +               return -EINVAL;
23733 +       }
23734 +       return 0;
23735 +}
23736 +
23737 +
23738 +int vc_net_add_ipv4(struct nx_info *nxi, void __user *data)
23739 +{
23740 +       struct vcmd_net_addr_ipv4_v1 vc_data;
23741 +
23742 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
23743 +               return -EFAULT;
23744 +
23745 +       switch (vc_data.type) {
23746 +       case NXA_TYPE_ADDR:
23747 +       case NXA_TYPE_RANGE:
23748 +       case NXA_TYPE_MASK:
23749 +               return do_add_v4_addr(nxi, vc_data.ip.s_addr, 0,
23750 +                       vc_data.mask.s_addr, vc_data.type, vc_data.flags);
23751 +
23752 +       case NXA_TYPE_ADDR | NXA_MOD_BCAST:
23753 +               nxi->v4_bcast = vc_data.ip;
23754 +               break;
23755 +
23756 +       case NXA_TYPE_ADDR | NXA_MOD_LBACK:
23757 +               nxi->v4_lback = vc_data.ip;
23758 +               break;
23759 +
23760 +       default:
23761 +               return -EINVAL;
23762 +       }
23763 +       return 0;
23764 +}
23765 +
23766 +int vc_net_remove_ipv4(struct nx_info *nxi, void __user *data)
23767 +{
23768 +       struct vcmd_net_addr_ipv4_v1 vc_data;
23769 +
23770 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
23771 +               return -EFAULT;
23772 +
23773 +       switch (vc_data.type) {
23774 +/*     case NXA_TYPE_ADDR:
23775 +               break;          */
23776 +
23777 +       case NXA_TYPE_ANY:
23778 +               __dealloc_nx_addr_v4_all(xchg(&nxi->v4.next, NULL));
23779 +               memset(&nxi->v4, 0, sizeof(nxi->v4));
23780 +               break;
23781 +
23782 +       default:
23783 +               return -EINVAL;
23784 +       }
23785 +       return 0;
23786 +}
23787 +
23788 +
23789 +#ifdef CONFIG_IPV6
23790 +
23791 +int do_add_v6_addr(struct nx_info *nxi,
23792 +       struct in6_addr *ip, struct in6_addr *mask,
23793 +       uint32_t prefix, uint16_t type, uint16_t flags)
23794 +{
23795 +       struct nx_addr_v6 *nxa = &nxi->v6;
23796 +
23797 +       if (NX_IPV6(nxi)) {
23798 +               /* locate last entry */
23799 +               for (; nxa->next; nxa = nxa->next);
23800 +               nxa->next = __alloc_nx_addr_v6();
23801 +               nxa = nxa->next;
23802 +
23803 +               if (IS_ERR(nxa))
23804 +                       return PTR_ERR(nxa);
23805 +       }
23806 +
23807 +       nxa->ip = *ip;
23808 +       nxa->mask = *mask;
23809 +       nxa->prefix = prefix;
23810 +       nxa->type = type;
23811 +       nxa->flags = flags;
23812 +       return 0;
23813 +}
23814 +
23815 +
23816 +int vc_net_add_ipv6(struct nx_info *nxi, void __user *data)
23817 +{
23818 +       struct vcmd_net_addr_ipv6_v1 vc_data;
23819 +
23820 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
23821 +               return -EFAULT;
23822 +
23823 +       switch (vc_data.type) {
23824 +       case NXA_TYPE_ADDR:
23825 +       case NXA_TYPE_MASK:
23826 +               return do_add_v6_addr(nxi, &vc_data.ip, &vc_data.mask,
23827 +                       vc_data.prefix, vc_data.type, vc_data.flags);
23828 +       default:
23829 +               return -EINVAL;
23830 +       }
23831 +       return 0;
23832 +}
23833 +
23834 +int vc_net_remove_ipv6(struct nx_info *nxi, void __user *data)
23835 +{
23836 +       struct vcmd_net_addr_ipv6_v1 vc_data;
23837 +
23838 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
23839 +               return -EFAULT;
23840 +
23841 +       switch (vc_data.type) {
23842 +       case NXA_TYPE_ANY:
23843 +               __dealloc_nx_addr_v6_all(xchg(&nxi->v6.next, NULL));
23844 +               memset(&nxi->v6, 0, sizeof(nxi->v6));
23845 +               break;
23846 +
23847 +       default:
23848 +               return -EINVAL;
23849 +       }
23850 +       return 0;
23851 +}
23852 +
23853 +#endif /* CONFIG_IPV6 */
23854 +
23855 +
23856 +int vc_get_nflags(struct nx_info *nxi, void __user *data)
23857 +{
23858 +       struct vcmd_net_flags_v0 vc_data;
23859 +
23860 +       vc_data.flagword = nxi->nx_flags;
23861 +
23862 +       /* special STATE flag handling */
23863 +       vc_data.mask = vs_mask_flags(~0ULL, nxi->nx_flags, NXF_ONE_TIME);
23864 +
23865 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
23866 +               return -EFAULT;
23867 +       return 0;
23868 +}
23869 +
23870 +int vc_set_nflags(struct nx_info *nxi, void __user *data)
23871 +{
23872 +       struct vcmd_net_flags_v0 vc_data;
23873 +       uint64_t mask, trigger;
23874 +
23875 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
23876 +               return -EFAULT;
23877 +
23878 +       /* special STATE flag handling */
23879 +       mask = vs_mask_mask(vc_data.mask, nxi->nx_flags, NXF_ONE_TIME);
23880 +       trigger = (mask & nxi->nx_flags) ^ (mask & vc_data.flagword);
23881 +
23882 +       nxi->nx_flags = vs_mask_flags(nxi->nx_flags,
23883 +               vc_data.flagword, mask);
23884 +       if (trigger & NXF_PERSISTENT)
23885 +               nx_update_persistent(nxi);
23886 +
23887 +       return 0;
23888 +}
23889 +
23890 +int vc_get_ncaps(struct nx_info *nxi, void __user *data)
23891 +{
23892 +       struct vcmd_net_caps_v0 vc_data;
23893 +
23894 +       vc_data.ncaps = nxi->nx_ncaps;
23895 +       vc_data.cmask = ~0ULL;
23896 +
23897 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
23898 +               return -EFAULT;
23899 +       return 0;
23900 +}
23901 +
23902 +int vc_set_ncaps(struct nx_info *nxi, void __user *data)
23903 +{
23904 +       struct vcmd_net_caps_v0 vc_data;
23905 +
23906 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
23907 +               return -EFAULT;
23908 +
23909 +       nxi->nx_ncaps = vs_mask_flags(nxi->nx_ncaps,
23910 +               vc_data.ncaps, vc_data.cmask);
23911 +       return 0;
23912 +}
23913 +
23914 +
23915 +#include <linux/module.h>
23916 +
23917 +module_init(init_network);
23918 +
23919 +EXPORT_SYMBOL_GPL(free_nx_info);
23920 +EXPORT_SYMBOL_GPL(unhash_nx_info);
23921 +
23922 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/proc.c linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/proc.c
23923 --- linux-2.6.25.11/kernel/vserver/proc.c       1969-12-31 19:00:00.000000000 -0500
23924 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/proc.c 2008-06-10 20:05:58.000000000 -0400
23925 @@ -0,0 +1,1086 @@
23926 +/*
23927 + *  linux/kernel/vserver/proc.c
23928 + *
23929 + *  Virtual Context Support
23930 + *
23931 + *  Copyright (C) 2003-2007  Herbert Pötzl
23932 + *
23933 + *  V0.01  basic structure
23934 + *  V0.02  adaptation vs1.3.0
23935 + *  V0.03  proc permissions
23936 + *  V0.04  locking/generic
23937 + *  V0.05  next generation procfs
23938 + *  V0.06  inode validation
23939 + *  V0.07  generic rewrite vid
23940 + *  V0.08  remove inode type
23941 + *
23942 + */
23943 +
23944 +#include <linux/proc_fs.h>
23945 +#include <asm/unistd.h>
23946 +
23947 +#include <linux/vs_context.h>
23948 +#include <linux/vs_network.h>
23949 +#include <linux/vs_cvirt.h>
23950 +
23951 +#include <linux/in.h>
23952 +#include <linux/inetdevice.h>
23953 +#include <linux/vs_inet.h>
23954 +#include <linux/vs_inet6.h>
23955 +
23956 +#include <linux/vserver/global.h>
23957 +
23958 +#include "cvirt_proc.h"
23959 +#include "cacct_proc.h"
23960 +#include "limit_proc.h"
23961 +#include "sched_proc.h"
23962 +#include "vci_config.h"
23963 +
23964 +
23965 +static inline char *print_cap_t(char *buffer, kernel_cap_t *c)
23966 +{
23967 +       unsigned __capi;
23968 +
23969 +       CAP_FOR_EACH_U32(__capi) {
23970 +               buffer += sprintf(buffer, "%08x",
23971 +                       c->cap[(_KERNEL_CAPABILITY_U32S-1) - __capi]);
23972 +       }
23973 +       return buffer;
23974 +}
23975 +
23976 +
23977 +static struct proc_dir_entry *proc_virtual;
23978 +
23979 +static struct proc_dir_entry *proc_virtnet;
23980 +
23981 +
23982 +/* first the actual feeds */
23983 +
23984 +
23985 +static int proc_vci(char *buffer)
23986 +{
23987 +       return sprintf(buffer,
23988 +               "VCIVersion:\t%04x:%04x\n"
23989 +               "VCISyscall:\t%d\n"
23990 +               "VCIKernel:\t%08x\n",
23991 +               VCI_VERSION >> 16,
23992 +               VCI_VERSION & 0xFFFF,
23993 +               __NR_vserver,
23994 +               vci_kernel_config());
23995 +}
23996 +
23997 +static int proc_virtual_info(char *buffer)
23998 +{
23999 +       return proc_vci(buffer);
24000 +}
24001 +
24002 +static int proc_virtual_status(char *buffer)
24003 +{
24004 +       return sprintf(buffer,
24005 +               "#CTotal:\t%d\n"
24006 +               "#CActive:\t%d\n"
24007 +               "#NSProxy:\t%d\t%d %d %d %d %d %d\n",
24008 +               atomic_read(&vx_global_ctotal),
24009 +               atomic_read(&vx_global_cactive),
24010 +               atomic_read(&vs_global_nsproxy),
24011 +               atomic_read(&vs_global_fs),
24012 +               atomic_read(&vs_global_mnt_ns),
24013 +               atomic_read(&vs_global_uts_ns),
24014 +               atomic_read(&vs_global_ipc_ns),
24015 +               atomic_read(&vs_global_user_ns),
24016 +               atomic_read(&vs_global_pid_ns));
24017 +}
24018 +
24019 +
24020 +int proc_vxi_info(struct vx_info *vxi, char *buffer)
24021 +{
24022 +       int length;
24023 +
24024 +       length = sprintf(buffer,
24025 +               "ID:\t%d\n"
24026 +               "Info:\t%p\n"
24027 +               "Init:\t%d\n"
24028 +               "OOM:\t%lld\n",
24029 +               vxi->vx_id,
24030 +               vxi,
24031 +               vxi->vx_initpid,
24032 +               vxi->vx_badness_bias);
24033 +       return length;
24034 +}
24035 +
24036 +int proc_vxi_status(struct vx_info *vxi, char *buffer)
24037 +{
24038 +       char *orig = buffer;
24039 +
24040 +       buffer += sprintf(buffer,
24041 +               "UseCnt:\t%d\n"
24042 +               "Tasks:\t%d\n"
24043 +               "Flags:\t%016llx\n",
24044 +               atomic_read(&vxi->vx_usecnt),
24045 +               atomic_read(&vxi->vx_tasks),
24046 +               (unsigned long long)vxi->vx_flags);
24047 +
24048 +       buffer += sprintf(buffer, "BCaps:\t");
24049 +       buffer = print_cap_t(buffer, &vxi->vx_bcaps);
24050 +       buffer += sprintf(buffer, "\n");
24051 +
24052 +       buffer += sprintf(buffer,
24053 +               "CCaps:\t%016llx\n"
24054 +               "Spaces:\t%08lx\n",
24055 +               (unsigned long long)vxi->vx_ccaps,
24056 +               vxi->vx_nsmask);
24057 +       return buffer - orig;
24058 +}
24059 +
24060 +int proc_vxi_limit(struct vx_info *vxi, char *buffer)
24061 +{
24062 +       return vx_info_proc_limit(&vxi->limit, buffer);
24063 +}
24064 +
24065 +int proc_vxi_sched(struct vx_info *vxi, char *buffer)
24066 +{
24067 +       int cpu, length;
24068 +
24069 +       length = vx_info_proc_sched(&vxi->sched, buffer);
24070 +       for_each_online_cpu(cpu) {
24071 +               length += vx_info_proc_sched_pc(
24072 +                       &vx_per_cpu(vxi, sched_pc, cpu),
24073 +                       buffer + length, cpu);
24074 +       }
24075 +       return length;
24076 +}
24077 +
24078 +int proc_vxi_nsproxy(struct vx_info *vxi, char *buffer)
24079 +{
24080 +       return vx_info_proc_nsproxy(vxi->vx_nsproxy, buffer);
24081 +}
24082 +
24083 +int proc_vxi_cvirt(struct vx_info *vxi, char *buffer)
24084 +{
24085 +       int cpu, length;
24086 +
24087 +       vx_update_load(vxi);
24088 +       length = vx_info_proc_cvirt(&vxi->cvirt, buffer);
24089 +       for_each_online_cpu(cpu) {
24090 +               length += vx_info_proc_cvirt_pc(
24091 +                       &vx_per_cpu(vxi, cvirt_pc, cpu),
24092 +                       buffer + length, cpu);
24093 +       }
24094 +       return length;
24095 +}
24096 +
24097 +int proc_vxi_cacct(struct vx_info *vxi, char *buffer)
24098 +{
24099 +       return vx_info_proc_cacct(&vxi->cacct, buffer);
24100 +}
24101 +
24102 +
24103 +static int proc_virtnet_info(char *buffer)
24104 +{
24105 +       return proc_vci(buffer);
24106 +}
24107 +
24108 +static int proc_virtnet_status(char *buffer)
24109 +{
24110 +       return sprintf(buffer,
24111 +               "#CTotal:\t%d\n"
24112 +               "#CActive:\t%d\n",
24113 +               atomic_read(&nx_global_ctotal),
24114 +               atomic_read(&nx_global_cactive));
24115 +}
24116 +
24117 +int proc_nxi_info(struct nx_info *nxi, char *buffer)
24118 +{
24119 +       struct nx_addr_v4 *v4a;
24120 +#ifdef CONFIG_IPV6
24121 +       struct nx_addr_v6 *v6a;
24122 +#endif
24123 +       int length, i;
24124 +
24125 +       length = sprintf(buffer,
24126 +               "ID:\t%d\n"
24127 +               "Info:\t%p\n"
24128 +               "Bcast:\t" NIPQUAD_FMT "\n"
24129 +               "Lback:\t" NIPQUAD_FMT "\n",
24130 +               nxi->nx_id,
24131 +               nxi,
24132 +               NIPQUAD(nxi->v4_bcast.s_addr),
24133 +               NIPQUAD(nxi->v4_lback.s_addr));
24134 +
24135 +       if (!NX_IPV4(nxi))
24136 +               goto skip_v4;
24137 +       for (i = 0, v4a = &nxi->v4; v4a; i++, v4a = v4a->next)
24138 +               length += sprintf(buffer + length, "%d:\t" NXAV4_FMT "\n",
24139 +                       i, NXAV4(v4a));
24140 +skip_v4:
24141 +#ifdef CONFIG_IPV6
24142 +       if (!NX_IPV6(nxi))
24143 +               goto skip_v6;
24144 +       for (i = 0, v6a = &nxi->v6; v6a; i++, v6a = v6a->next)
24145 +               length += sprintf(buffer + length, "%d:\t" NXAV6_FMT "\n",
24146 +                       i, NXAV6(v6a));
24147 +skip_v6:
24148 +#endif
24149 +       return length;
24150 +}
24151 +
24152 +int proc_nxi_status(struct nx_info *nxi, char *buffer)
24153 +{
24154 +       int length;
24155 +
24156 +       length = sprintf(buffer,
24157 +               "UseCnt:\t%d\n"
24158 +               "Tasks:\t%d\n"
24159 +               "Flags:\t%016llx\n"
24160 +               "NCaps:\t%016llx\n",
24161 +               atomic_read(&nxi->nx_usecnt),
24162 +               atomic_read(&nxi->nx_tasks),
24163 +               (unsigned long long)nxi->nx_flags,
24164 +               (unsigned long long)nxi->nx_ncaps);
24165 +       return length;
24166 +}
24167 +
24168 +
24169 +
24170 +/* here the inode helpers */
24171 +
24172 +struct vs_entry {
24173 +       int len;
24174 +       char *name;
24175 +       mode_t mode;
24176 +       struct inode_operations *iop;
24177 +       struct file_operations *fop;
24178 +       union proc_op op;
24179 +};
24180 +
24181 +static struct inode *vs_proc_make_inode(struct super_block *sb, struct vs_entry *p)
24182 +{
24183 +       struct inode *inode = new_inode(sb);
24184 +
24185 +       if (!inode)
24186 +               goto out;
24187 +
24188 +       inode->i_mode = p->mode;
24189 +       if (p->iop)
24190 +               inode->i_op = p->iop;
24191 +       if (p->fop)
24192 +               inode->i_fop = p->fop;
24193 +
24194 +       inode->i_nlink = (p->mode & S_IFDIR) ? 2 : 1;
24195 +       inode->i_flags |= S_IMMUTABLE;
24196 +
24197 +       inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
24198 +
24199 +       inode->i_uid = 0;
24200 +       inode->i_gid = 0;
24201 +       inode->i_tag = 0;
24202 +out:
24203 +       return inode;
24204 +}
24205 +
24206 +static struct dentry *vs_proc_instantiate(struct inode *dir,
24207 +       struct dentry *dentry, int id, void *ptr)
24208 +{
24209 +       struct vs_entry *p = ptr;
24210 +       struct inode *inode = vs_proc_make_inode(dir->i_sb, p);
24211 +       struct dentry *error = ERR_PTR(-EINVAL);
24212 +
24213 +       if (!inode)
24214 +               goto out;
24215 +
24216 +       PROC_I(inode)->op = p->op;
24217 +       PROC_I(inode)->fd = id;
24218 +       d_add(dentry, inode);
24219 +       error = NULL;
24220 +out:
24221 +       return error;
24222 +}
24223 +
24224 +/* Lookups */
24225 +
24226 +typedef struct dentry *instantiate_t(struct inode *, struct dentry *, int, void *);
24227 +
24228 +/*
24229 + * Fill a directory entry.
24230 + *
24231 + * If possible create the dcache entry and derive our inode number and
24232 + * file type from dcache entry.
24233 + *
24234 + * Since all of the proc inode numbers are dynamically generated, the inode
24235 + * numbers do not exist until the inode is cache.  This means creating the
24236 + * the dcache entry in readdir is necessary to keep the inode numbers
24237 + * reported by readdir in sync with the inode numbers reported
24238 + * by stat.
24239 + */
24240 +static int proc_fill_cache(struct file *filp, void *dirent, filldir_t filldir,
24241 +       char *name, int len, instantiate_t instantiate, int id, void *ptr)
24242 +{
24243 +       struct dentry *child, *dir = filp->f_dentry;
24244 +       struct inode *inode;
24245 +       struct qstr qname;
24246 +       ino_t ino = 0;
24247 +       unsigned type = DT_UNKNOWN;
24248 +
24249 +       qname.name = name;
24250 +       qname.len  = len;
24251 +       qname.hash = full_name_hash(name, len);
24252 +
24253 +       child = d_lookup(dir, &qname);
24254 +       if (!child) {
24255 +               struct dentry *new;
24256 +               new = d_alloc(dir, &qname);
24257 +               if (new) {
24258 +                       child = instantiate(dir->d_inode, new, id, ptr);
24259 +                       if (child)
24260 +                               dput(new);
24261 +                       else
24262 +                               child = new;
24263 +               }
24264 +       }
24265 +       if (!child || IS_ERR(child) || !child->d_inode)
24266 +               goto end_instantiate;
24267 +       inode = child->d_inode;
24268 +       if (inode) {
24269 +               ino = inode->i_ino;
24270 +               type = inode->i_mode >> 12;
24271 +       }
24272 +       dput(child);
24273 +end_instantiate:
24274 +       if (!ino)
24275 +               ino = find_inode_number(dir, &qname);
24276 +       if (!ino)
24277 +               ino = 1;
24278 +       return filldir(dirent, name, len, filp->f_pos, ino, type);
24279 +}
24280 +
24281 +
24282 +
24283 +/* get and revalidate vx_info/xid */
24284 +
24285 +static inline
24286 +struct vx_info *get_proc_vx_info(struct inode *inode)
24287 +{
24288 +       return lookup_vx_info(PROC_I(inode)->fd);
24289 +}
24290 +
24291 +static int proc_xid_revalidate(struct dentry *dentry, struct nameidata *nd)
24292 +{
24293 +       struct inode *inode = dentry->d_inode;
24294 +       xid_t xid = PROC_I(inode)->fd;
24295 +
24296 +       if (!xid || xid_is_hashed(xid))
24297 +               return 1;
24298 +       d_drop(dentry);
24299 +       return 0;
24300 +}
24301 +
24302 +
24303 +/* get and revalidate nx_info/nid */
24304 +
24305 +static int proc_nid_revalidate(struct dentry *dentry, struct nameidata *nd)
24306 +{
24307 +       struct inode *inode = dentry->d_inode;
24308 +       nid_t nid = PROC_I(inode)->fd;
24309 +
24310 +       if (!nid || nid_is_hashed(nid))
24311 +               return 1;
24312 +       d_drop(dentry);
24313 +       return 0;
24314 +}
24315 +
24316 +
24317 +
24318 +#define PROC_BLOCK_SIZE (PAGE_SIZE - 1024)
24319 +
24320 +static ssize_t proc_vs_info_read(struct file *file, char __user *buf,
24321 +                         size_t count, loff_t *ppos)
24322 +{
24323 +       struct inode *inode = file->f_dentry->d_inode;
24324 +       unsigned long page;
24325 +       ssize_t length = 0;
24326 +
24327 +       if (count > PROC_BLOCK_SIZE)
24328 +               count = PROC_BLOCK_SIZE;
24329 +
24330 +       /* fade that out as soon as stable */
24331 +       WARN_ON(PROC_I(inode)->fd);
24332 +
24333 +       if (!(page = __get_free_page(GFP_KERNEL)))
24334 +               return -ENOMEM;
24335 +
24336 +       BUG_ON(!PROC_I(inode)->op.proc_vs_read);
24337 +       length = PROC_I(inode)->op.proc_vs_read((char *)page);
24338 +
24339 +       if (length >= 0)
24340 +               length = simple_read_from_buffer(buf, count, ppos,
24341 +                       (char *)page, length);
24342 +
24343 +       free_page(page);
24344 +       return length;
24345 +}
24346 +
24347 +static ssize_t proc_vx_info_read(struct file *file, char __user *buf,
24348 +                         size_t count, loff_t *ppos)
24349 +{
24350 +       struct inode *inode = file->f_dentry->d_inode;
24351 +       struct vx_info *vxi = NULL;
24352 +       xid_t xid = PROC_I(inode)->fd;
24353 +       unsigned long page;
24354 +       ssize_t length = 0;
24355 +
24356 +       if (count > PROC_BLOCK_SIZE)
24357 +               count = PROC_BLOCK_SIZE;
24358 +
24359 +       /* fade that out as soon as stable */
24360 +       WARN_ON(!xid);
24361 +       vxi = lookup_vx_info(xid);
24362 +       if (!vxi)
24363 +               goto out;
24364 +
24365 +       length = -ENOMEM;
24366 +       if (!(page = __get_free_page(GFP_KERNEL)))
24367 +               goto out_put;
24368 +
24369 +       BUG_ON(!PROC_I(inode)->op.proc_vxi_read);
24370 +       length = PROC_I(inode)->op.proc_vxi_read(vxi, (char *)page);
24371 +
24372 +       if (length >= 0)
24373 +               length = simple_read_from_buffer(buf, count, ppos,
24374 +                       (char *)page, length);
24375 +
24376 +       free_page(page);
24377 +out_put:
24378 +       put_vx_info(vxi);
24379 +out:
24380 +       return length;
24381 +}
24382 +
24383 +static ssize_t proc_nx_info_read(struct file *file, char __user *buf,
24384 +                         size_t count, loff_t *ppos)
24385 +{
24386 +       struct inode *inode = file->f_dentry->d_inode;
24387 +       struct nx_info *nxi = NULL;
24388 +       nid_t nid = PROC_I(inode)->fd;
24389 +       unsigned long page;
24390 +       ssize_t length = 0;
24391 +
24392 +       if (count > PROC_BLOCK_SIZE)
24393 +               count = PROC_BLOCK_SIZE;
24394 +
24395 +       /* fade that out as soon as stable */
24396 +       WARN_ON(!nid);
24397 +       nxi = lookup_nx_info(nid);
24398 +       if (!nxi)
24399 +               goto out;
24400 +
24401 +       length = -ENOMEM;
24402 +       if (!(page = __get_free_page(GFP_KERNEL)))
24403 +               goto out_put;
24404 +
24405 +       BUG_ON(!PROC_I(inode)->op.proc_nxi_read);
24406 +       length = PROC_I(inode)->op.proc_nxi_read(nxi, (char *)page);
24407 +
24408 +       if (length >= 0)
24409 +               length = simple_read_from_buffer(buf, count, ppos,
24410 +                       (char *)page, length);
24411 +
24412 +       free_page(page);
24413 +out_put:
24414 +       put_nx_info(nxi);
24415 +out:
24416 +       return length;
24417 +}
24418 +
24419 +
24420 +
24421 +/* here comes the lower level */
24422 +
24423 +
24424 +#define NOD(NAME, MODE, IOP, FOP, OP) {        \
24425 +       .len  = sizeof(NAME) - 1,       \
24426 +       .name = (NAME),                 \
24427 +       .mode = MODE,                   \
24428 +       .iop  = IOP,                    \
24429 +       .fop  = FOP,                    \
24430 +       .op   = OP,                     \
24431 +}
24432 +
24433 +
24434 +#define DIR(NAME, MODE, OTYPE)                         \
24435 +       NOD(NAME, (S_IFDIR | (MODE)),                   \
24436 +               &proc_ ## OTYPE ## _inode_operations,   \
24437 +               &proc_ ## OTYPE ## _file_operations, { } )
24438 +
24439 +#define INF(NAME, MODE, OTYPE)                         \
24440 +       NOD(NAME, (S_IFREG | (MODE)), NULL,             \
24441 +               &proc_vs_info_file_operations,          \
24442 +               { .proc_vs_read = &proc_##OTYPE } )
24443 +
24444 +#define VINF(NAME, MODE, OTYPE)                                \
24445 +       NOD(NAME, (S_IFREG | (MODE)), NULL,             \
24446 +               &proc_vx_info_file_operations,          \
24447 +               { .proc_vxi_read = &proc_##OTYPE } )
24448 +
24449 +#define NINF(NAME, MODE, OTYPE)                                \
24450 +       NOD(NAME, (S_IFREG | (MODE)), NULL,             \
24451 +               &proc_nx_info_file_operations,          \
24452 +               { .proc_nxi_read = &proc_##OTYPE } )
24453 +
24454 +
24455 +static struct file_operations proc_vs_info_file_operations = {
24456 +       .read =         proc_vs_info_read,
24457 +};
24458 +
24459 +static struct file_operations proc_vx_info_file_operations = {
24460 +       .read =         proc_vx_info_read,
24461 +};
24462 +
24463 +static struct dentry_operations proc_xid_dentry_operations = {
24464 +       .d_revalidate = proc_xid_revalidate,
24465 +};
24466 +
24467 +static struct vs_entry vx_base_stuff[] = {
24468 +       VINF("info",    S_IRUGO, vxi_info),
24469 +       VINF("status",  S_IRUGO, vxi_status),
24470 +       VINF("limit",   S_IRUGO, vxi_limit),
24471 +       VINF("sched",   S_IRUGO, vxi_sched),
24472 +       VINF("nsproxy", S_IRUGO, vxi_nsproxy),
24473 +       VINF("cvirt",   S_IRUGO, vxi_cvirt),
24474 +       VINF("cacct",   S_IRUGO, vxi_cacct),
24475 +       {}
24476 +};
24477 +
24478 +
24479 +
24480 +
24481 +static struct dentry *proc_xid_instantiate(struct inode *dir,
24482 +       struct dentry *dentry, int id, void *ptr)
24483 +{
24484 +       dentry->d_op = &proc_xid_dentry_operations;
24485 +       return vs_proc_instantiate(dir, dentry, id, ptr);
24486 +}
24487 +
24488 +static struct dentry *proc_xid_lookup(struct inode *dir,
24489 +       struct dentry *dentry, struct nameidata *nd)
24490 +{
24491 +       struct vs_entry *p = vx_base_stuff;
24492 +       struct dentry *error = ERR_PTR(-ENOENT);
24493 +
24494 +       for (; p->name; p++) {
24495 +               if (p->len != dentry->d_name.len)
24496 +                       continue;
24497 +               if (!memcmp(dentry->d_name.name, p->name, p->len))
24498 +                       break;
24499 +       }
24500 +       if (!p->name)
24501 +               goto out;
24502 +
24503 +       error = proc_xid_instantiate(dir, dentry, PROC_I(dir)->fd, p);
24504 +out:
24505 +       return error;
24506 +}
24507 +
24508 +static int proc_xid_readdir(struct file *filp,
24509 +       void *dirent, filldir_t filldir)
24510 +{
24511 +       struct dentry *dentry = filp->f_dentry;
24512 +       struct inode *inode = dentry->d_inode;
24513 +       struct vs_entry *p = vx_base_stuff;
24514 +       int size = sizeof(vx_base_stuff) / sizeof(struct vs_entry);
24515 +       int pos, index;
24516 +       u64 ino;
24517 +
24518 +       pos = filp->f_pos;
24519 +       switch (pos) {
24520 +       case 0:
24521 +               ino = inode->i_ino;
24522 +               if (filldir(dirent, ".", 1, pos, ino, DT_DIR) < 0)
24523 +                       goto out;
24524 +               pos++;
24525 +               /* fall through */
24526 +       case 1:
24527 +               ino = parent_ino(dentry);
24528 +               if (filldir(dirent, "..", 2, pos, ino, DT_DIR) < 0)
24529 +                       goto out;
24530 +               pos++;
24531 +               /* fall through */
24532 +       default:
24533 +               index = pos - 2;
24534 +               if (index >= size)
24535 +                       goto out;
24536 +               for (p += index; p->name; p++) {
24537 +                       if (proc_fill_cache(filp, dirent, filldir, p->name, p->len,
24538 +                               vs_proc_instantiate, PROC_I(inode)->fd, p))
24539 +                               goto out;
24540 +                       pos++;
24541 +               }
24542 +       }
24543 +out:
24544 +       filp->f_pos = pos;
24545 +       return 1;
24546 +}
24547 +
24548 +
24549 +
24550 +static struct file_operations proc_nx_info_file_operations = {
24551 +       .read =         proc_nx_info_read,
24552 +};
24553 +
24554 +static struct dentry_operations proc_nid_dentry_operations = {
24555 +       .d_revalidate = proc_nid_revalidate,
24556 +};
24557 +
24558 +static struct vs_entry nx_base_stuff[] = {
24559 +       NINF("info",    S_IRUGO, nxi_info),
24560 +       NINF("status",  S_IRUGO, nxi_status),
24561 +       {}
24562 +};
24563 +
24564 +
24565 +static struct dentry *proc_nid_instantiate(struct inode *dir,
24566 +       struct dentry *dentry, int id, void *ptr)
24567 +{
24568 +       dentry->d_op = &proc_nid_dentry_operations;
24569 +       return vs_proc_instantiate(dir, dentry, id, ptr);
24570 +}
24571 +
24572 +static struct dentry *proc_nid_lookup(struct inode *dir,
24573 +       struct dentry *dentry, struct nameidata *nd)
24574 +{
24575 +       struct vs_entry *p = nx_base_stuff;
24576 +       struct dentry *error = ERR_PTR(-ENOENT);
24577 +
24578 +       for (; p->name; p++) {
24579 +               if (p->len != dentry->d_name.len)
24580 +                       continue;
24581 +               if (!memcmp(dentry->d_name.name, p->name, p->len))
24582 +                       break;
24583 +       }
24584 +       if (!p->name)
24585 +               goto out;
24586 +
24587 +       error = proc_nid_instantiate(dir, dentry, PROC_I(dir)->fd, p);
24588 +out:
24589 +       return error;
24590 +}
24591 +
24592 +static int proc_nid_readdir(struct file *filp,
24593 +       void *dirent, filldir_t filldir)
24594 +{
24595 +       struct dentry *dentry = filp->f_dentry;
24596 +       struct inode *inode = dentry->d_inode;
24597 +       struct vs_entry *p = nx_base_stuff;
24598 +       int size = sizeof(nx_base_stuff) / sizeof(struct vs_entry);
24599 +       int pos, index;
24600 +       u64 ino;
24601 +
24602 +       pos = filp->f_pos;
24603 +       switch (pos) {
24604 +       case 0:
24605 +               ino = inode->i_ino;
24606 +               if (filldir(dirent, ".", 1, pos, ino, DT_DIR) < 0)
24607 +                       goto out;
24608 +               pos++;
24609 +               /* fall through */
24610 +       case 1:
24611 +               ino = parent_ino(dentry);
24612 +               if (filldir(dirent, "..", 2, pos, ino, DT_DIR) < 0)
24613 +                       goto out;
24614 +               pos++;
24615 +               /* fall through */
24616 +       default:
24617 +               index = pos - 2;
24618 +               if (index >= size)
24619 +                       goto out;
24620 +               for (p += index; p->name; p++) {
24621 +                       if (proc_fill_cache(filp, dirent, filldir, p->name, p->len,
24622 +                               vs_proc_instantiate, PROC_I(inode)->fd, p))
24623 +                               goto out;
24624 +                       pos++;
24625 +               }
24626 +       }
24627 +out:
24628 +       filp->f_pos = pos;
24629 +       return 1;
24630 +}
24631 +
24632 +
24633 +#define MAX_MULBY10    ((~0U - 9) / 10)
24634 +
24635 +static inline int atovid(const char *str, int len)
24636 +{
24637 +       int vid, c;
24638 +
24639 +       vid = 0;
24640 +       while (len-- > 0) {
24641 +               c = *str - '0';
24642 +               str++;
24643 +               if (c > 9)
24644 +                       return -1;
24645 +               if (vid >= MAX_MULBY10)
24646 +                       return -1;
24647 +               vid *= 10;
24648 +               vid += c;
24649 +               if (!vid)
24650 +                       return -1;
24651 +       }
24652 +       return vid;
24653 +}
24654 +
24655 +/* now the upper level (virtual) */
24656 +
24657 +
24658 +static struct file_operations proc_xid_file_operations = {
24659 +       .read =         generic_read_dir,
24660 +       .readdir =      proc_xid_readdir,
24661 +};
24662 +
24663 +static struct inode_operations proc_xid_inode_operations = {
24664 +       .lookup =       proc_xid_lookup,
24665 +};
24666 +
24667 +static struct vs_entry vx_virtual_stuff[] = {
24668 +       INF("info",     S_IRUGO, virtual_info),
24669 +       INF("status",   S_IRUGO, virtual_status),
24670 +       DIR(NULL,       S_IRUGO | S_IXUGO, xid),
24671 +};
24672 +
24673 +
24674 +static struct dentry *proc_virtual_lookup(struct inode *dir,
24675 +       struct dentry *dentry, struct nameidata *nd)
24676 +{
24677 +       struct vs_entry *p = vx_virtual_stuff;
24678 +       struct dentry *error = ERR_PTR(-ENOENT);
24679 +       int id = 0;
24680 +
24681 +       for (; p->name; p++) {
24682 +               if (p->len != dentry->d_name.len)
24683 +                       continue;
24684 +               if (!memcmp(dentry->d_name.name, p->name, p->len))
24685 +                       break;
24686 +       }
24687 +       if (p->name)
24688 +               goto instantiate;
24689 +
24690 +       id = atovid(dentry->d_name.name, dentry->d_name.len);
24691 +       if ((id < 0) || !xid_is_hashed(id))
24692 +               goto out;
24693 +
24694 +instantiate:
24695 +       error = proc_xid_instantiate(dir, dentry, id, p);
24696 +out:
24697 +       return error;
24698 +}
24699 +
24700 +static struct file_operations proc_nid_file_operations = {
24701 +       .read =         generic_read_dir,
24702 +       .readdir =      proc_nid_readdir,
24703 +};
24704 +
24705 +static struct inode_operations proc_nid_inode_operations = {
24706 +       .lookup =       proc_nid_lookup,
24707 +};
24708 +
24709 +static struct vs_entry nx_virtnet_stuff[] = {
24710 +       INF("info",     S_IRUGO, virtnet_info),
24711 +       INF("status",   S_IRUGO, virtnet_status),
24712 +       DIR(NULL,       S_IRUGO | S_IXUGO, nid),
24713 +};
24714 +
24715 +
24716 +static struct dentry *proc_virtnet_lookup(struct inode *dir,
24717 +       struct dentry *dentry, struct nameidata *nd)
24718 +{
24719 +       struct vs_entry *p = nx_virtnet_stuff;
24720 +       struct dentry *error = ERR_PTR(-ENOENT);
24721 +       int id = 0;
24722 +
24723 +       for (; p->name; p++) {
24724 +               if (p->len != dentry->d_name.len)
24725 +                       continue;
24726 +               if (!memcmp(dentry->d_name.name, p->name, p->len))
24727 +                       break;
24728 +       }
24729 +       if (p->name)
24730 +               goto instantiate;
24731 +
24732 +       id = atovid(dentry->d_name.name, dentry->d_name.len);
24733 +       if ((id < 0) || !nid_is_hashed(id))
24734 +               goto out;
24735 +
24736 +instantiate:
24737 +       error = proc_nid_instantiate(dir, dentry, id, p);
24738 +out:
24739 +       return error;
24740 +}
24741 +
24742 +
24743 +#define PROC_MAXVIDS 32
24744 +
24745 +int proc_virtual_readdir(struct file *filp,
24746 +       void *dirent, filldir_t filldir)
24747 +{
24748 +       struct dentry *dentry = filp->f_dentry;
24749 +       struct inode *inode = dentry->d_inode;
24750 +       struct vs_entry *p = vx_virtual_stuff;
24751 +       int size = sizeof(vx_virtual_stuff) / sizeof(struct vs_entry);
24752 +       int pos, index;
24753 +       unsigned int xid_array[PROC_MAXVIDS];
24754 +       char buf[PROC_NUMBUF];
24755 +       unsigned int nr_xids, i;
24756 +       u64 ino;
24757 +
24758 +       pos = filp->f_pos;
24759 +       switch (pos) {
24760 +       case 0:
24761 +               ino = inode->i_ino;
24762 +               if (filldir(dirent, ".", 1, pos, ino, DT_DIR) < 0)
24763 +                       goto out;
24764 +               pos++;
24765 +               /* fall through */
24766 +       case 1:
24767 +               ino = parent_ino(dentry);
24768 +               if (filldir(dirent, "..", 2, pos, ino, DT_DIR) < 0)
24769 +                       goto out;
24770 +               pos++;
24771 +               /* fall through */
24772 +       default:
24773 +               index = pos - 2;
24774 +               if (index >= size)
24775 +                       goto entries;
24776 +               for (p += index; p->name; p++) {
24777 +                       if (proc_fill_cache(filp, dirent, filldir, p->name, p->len,
24778 +                               vs_proc_instantiate, 0, p))
24779 +                               goto out;
24780 +                       pos++;
24781 +               }
24782 +       entries:
24783 +               index = pos - size;
24784 +               p = &vx_virtual_stuff[size - 1];
24785 +               nr_xids = get_xid_list(index, xid_array, PROC_MAXVIDS);
24786 +               for (i = 0; i < nr_xids; i++) {
24787 +                       int n, xid = xid_array[i];
24788 +                       unsigned int j = PROC_NUMBUF;
24789 +
24790 +                       n = xid;
24791 +                       do
24792 +                               buf[--j] = '0' + (n % 10);
24793 +                       while (n /= 10);
24794 +
24795 +                       if (proc_fill_cache(filp, dirent, filldir,
24796 +                               buf + j, PROC_NUMBUF - j,
24797 +                               vs_proc_instantiate, xid, p))
24798 +                               goto out;
24799 +                       pos++;
24800 +               }
24801 +       }
24802 +out:
24803 +       filp->f_pos = pos;
24804 +       return 0;
24805 +}
24806 +
24807 +static int proc_virtual_getattr(struct vfsmount *mnt,
24808 +       struct dentry *dentry, struct kstat *stat)
24809 +{
24810 +       struct inode *inode = dentry->d_inode;
24811 +
24812 +       generic_fillattr(inode, stat);
24813 +       stat->nlink = 2 + atomic_read(&vx_global_cactive);
24814 +       return 0;
24815 +}
24816 +
24817 +static struct file_operations proc_virtual_dir_operations = {
24818 +       .read =         generic_read_dir,
24819 +       .readdir =      proc_virtual_readdir,
24820 +};
24821 +
24822 +static struct inode_operations proc_virtual_dir_inode_operations = {
24823 +       .getattr =      proc_virtual_getattr,
24824 +       .lookup =       proc_virtual_lookup,
24825 +};
24826 +
24827 +
24828 +
24829 +
24830 +
24831 +int proc_virtnet_readdir(struct file *filp,
24832 +       void *dirent, filldir_t filldir)
24833 +{
24834 +       struct dentry *dentry = filp->f_dentry;
24835 +       struct inode *inode = dentry->d_inode;
24836 +       struct vs_entry *p = nx_virtnet_stuff;
24837 +       int size = sizeof(nx_virtnet_stuff) / sizeof(struct vs_entry);
24838 +       int pos, index;
24839 +       unsigned int nid_array[PROC_MAXVIDS];
24840 +       char buf[PROC_NUMBUF];
24841 +       unsigned int nr_nids, i;
24842 +       u64 ino;
24843 +
24844 +       pos = filp->f_pos;
24845 +       switch (pos) {
24846 +       case 0:
24847 +               ino = inode->i_ino;
24848 +               if (filldir(dirent, ".", 1, pos, ino, DT_DIR) < 0)
24849 +                       goto out;
24850 +               pos++;
24851 +               /* fall through */
24852 +       case 1:
24853 +               ino = parent_ino(dentry);
24854 +               if (filldir(dirent, "..", 2, pos, ino, DT_DIR) < 0)
24855 +                       goto out;
24856 +               pos++;
24857 +               /* fall through */
24858 +       default:
24859 +               index = pos - 2;
24860 +               if (index >= size)
24861 +                       goto entries;
24862 +               for (p += index; p->name; p++) {
24863 +                       if (proc_fill_cache(filp, dirent, filldir, p->name, p->len,
24864 +                               vs_proc_instantiate, 0, p))
24865 +                               goto out;
24866 +                       pos++;
24867 +               }
24868 +       entries:
24869 +               index = pos - size;
24870 +               p = &nx_virtnet_stuff[size - 1];
24871 +               nr_nids = get_nid_list(index, nid_array, PROC_MAXVIDS);
24872 +               for (i = 0; i < nr_nids; i++) {
24873 +                       int n, nid = nid_array[i];
24874 +                       unsigned int j = PROC_NUMBUF;
24875 +
24876 +                       n = nid;
24877 +                       do
24878 +                               buf[--j] = '0' + (n % 10);
24879 +                       while (n /= 10);
24880 +
24881 +                       if (proc_fill_cache(filp, dirent, filldir,
24882 +                               buf + j, PROC_NUMBUF - j,
24883 +                               vs_proc_instantiate, nid, p))
24884 +                               goto out;
24885 +                       pos++;
24886 +               }
24887 +       }
24888 +out:
24889 +       filp->f_pos = pos;
24890 +       return 0;
24891 +}
24892 +
24893 +static int proc_virtnet_getattr(struct vfsmount *mnt,
24894 +       struct dentry *dentry, struct kstat *stat)
24895 +{
24896 +       struct inode *inode = dentry->d_inode;
24897 +
24898 +       generic_fillattr(inode, stat);
24899 +       stat->nlink = 2 + atomic_read(&nx_global_cactive);
24900 +       return 0;
24901 +}
24902 +
24903 +static struct file_operations proc_virtnet_dir_operations = {
24904 +       .read =         generic_read_dir,
24905 +       .readdir =      proc_virtnet_readdir,
24906 +};
24907 +
24908 +static struct inode_operations proc_virtnet_dir_inode_operations = {
24909 +       .getattr =      proc_virtnet_getattr,
24910 +       .lookup =       proc_virtnet_lookup,
24911 +};
24912 +
24913 +
24914 +
24915 +void proc_vx_init(void)
24916 +{
24917 +       struct proc_dir_entry *ent;
24918 +
24919 +       ent = proc_mkdir("virtual", 0);
24920 +       if (ent) {
24921 +               ent->proc_fops = &proc_virtual_dir_operations;
24922 +               ent->proc_iops = &proc_virtual_dir_inode_operations;
24923 +       }
24924 +       proc_virtual = ent;
24925 +
24926 +       ent = proc_mkdir("virtnet", 0);
24927 +       if (ent) {
24928 +               ent->proc_fops = &proc_virtnet_dir_operations;
24929 +               ent->proc_iops = &proc_virtnet_dir_inode_operations;
24930 +       }
24931 +       proc_virtnet = ent;
24932 +}
24933 +
24934 +
24935 +
24936 +
24937 +/* per pid info */
24938 +
24939 +
24940 +int proc_pid_vx_info(struct task_struct *p, char *buffer)
24941 +{
24942 +       struct vx_info *vxi;
24943 +       char *orig = buffer;
24944 +
24945 +       buffer += sprintf(buffer, "XID:\t%d\n", vx_task_xid(p));
24946 +
24947 +       vxi = task_get_vx_info(p);
24948 +       if (!vxi)
24949 +               goto out;
24950 +
24951 +       buffer += sprintf(buffer, "BCaps:\t");
24952 +       buffer = print_cap_t(buffer, &vxi->vx_bcaps);
24953 +       buffer += sprintf(buffer, "\n");
24954 +       buffer += sprintf(buffer, "CCaps:\t%016llx\n",
24955 +               (unsigned long long)vxi->vx_ccaps);
24956 +       buffer += sprintf(buffer, "CFlags:\t%016llx\n",
24957 +               (unsigned long long)vxi->vx_flags);
24958 +       buffer += sprintf(buffer, "CIPid:\t%d\n", vxi->vx_initpid);
24959 +
24960 +       put_vx_info(vxi);
24961 +out:
24962 +       return buffer - orig;
24963 +}
24964 +
24965 +
24966 +int proc_pid_nx_info(struct task_struct *p, char *buffer)
24967 +{
24968 +       struct nx_info *nxi;
24969 +       struct nx_addr_v4 *v4a;
24970 +#ifdef CONFIG_IPV6
24971 +       struct nx_addr_v6 *v6a;
24972 +#endif
24973 +       char *orig = buffer;
24974 +       int i;
24975 +
24976 +       buffer += sprintf(buffer, "NID:\t%d\n", nx_task_nid(p));
24977 +
24978 +       nxi = task_get_nx_info(p);
24979 +       if (!nxi)
24980 +               goto out;
24981 +
24982 +       buffer += sprintf(buffer, "NCaps:\t%016llx\n",
24983 +               (unsigned long long)nxi->nx_ncaps);
24984 +       buffer += sprintf(buffer, "NFlags:\t%016llx\n",
24985 +               (unsigned long long)nxi->nx_flags);
24986 +
24987 +       buffer += sprintf(buffer,
24988 +               "V4Root[bcast]:\t" NIPQUAD_FMT "\n",
24989 +               NIPQUAD(nxi->v4_bcast.s_addr));
24990 +       buffer += sprintf (buffer,
24991 +               "V4Root[lback]:\t" NIPQUAD_FMT "\n",
24992 +               NIPQUAD(nxi->v4_lback.s_addr));
24993 +       if (!NX_IPV4(nxi))
24994 +               goto skip_v4;
24995 +       for (i = 0, v4a = &nxi->v4; v4a; i++, v4a = v4a->next)
24996 +               buffer += sprintf(buffer, "V4Root[%d]:\t" NXAV4_FMT "\n",
24997 +                       i, NXAV4(v4a));
24998 +skip_v4:
24999 +#ifdef CONFIG_IPV6
25000 +       if (!NX_IPV6(nxi))
25001 +               goto skip_v6;
25002 +       for (i = 0, v6a = &nxi->v6; v6a; i++, v6a = v6a->next)
25003 +               buffer += sprintf(buffer, "V6Root[%d]:\t" NXAV6_FMT "\n",
25004 +                       i, NXAV6(v6a));
25005 +skip_v6:
25006 +#endif
25007 +       put_nx_info(nxi);
25008 +out:
25009 +       return buffer - orig;
25010 +}
25011 +
25012 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/sched.c linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/sched.c
25013 --- linux-2.6.25.11/kernel/vserver/sched.c      1969-12-31 19:00:00.000000000 -0500
25014 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/sched.c        2008-04-29 18:40:09.000000000 -0400
25015 @@ -0,0 +1,413 @@
25016 +/*
25017 + *  linux/kernel/vserver/sched.c
25018 + *
25019 + *  Virtual Server: Scheduler Support
25020 + *
25021 + *  Copyright (C) 2004-2007  Herbert Pötzl
25022 + *
25023 + *  V0.01  adapted Sam Vilains version to 2.6.3
25024 + *  V0.02  removed legacy interface
25025 + *  V0.03  changed vcmds to vxi arg
25026 + *  V0.04  removed older and legacy interfaces
25027 + *
25028 + */
25029 +
25030 +#include <linux/vs_context.h>
25031 +#include <linux/vs_sched.h>
25032 +#include <linux/vserver/sched_cmd.h>
25033 +
25034 +#include <asm/uaccess.h>
25035 +
25036 +
25037 +#define vxd_check_range(val, min, max) do {            \
25038 +       vxlprintk((val < min) || (val > max),           \
25039 +               "check_range(%ld,%ld,%ld)",             \
25040 +               (long)val, (long)min, (long)max,        \
25041 +               __FILE__, __LINE__);                    \
25042 +       } while (0)
25043 +
25044 +
25045 +void vx_update_sched_param(struct _vx_sched *sched,
25046 +       struct _vx_sched_pc *sched_pc)
25047 +{
25048 +       unsigned int set_mask = sched->update_mask;
25049 +
25050 +       if (set_mask & VXSM_FILL_RATE)
25051 +               sched_pc->fill_rate[0] = sched->fill_rate[0];
25052 +       if (set_mask & VXSM_INTERVAL)
25053 +               sched_pc->interval[0] = sched->interval[0];
25054 +       if (set_mask & VXSM_FILL_RATE2)
25055 +               sched_pc->fill_rate[1] = sched->fill_rate[1];
25056 +       if (set_mask & VXSM_INTERVAL2)
25057 +               sched_pc->interval[1] = sched->interval[1];
25058 +       if (set_mask & VXSM_TOKENS)
25059 +               sched_pc->tokens = sched->tokens;
25060 +       if (set_mask & VXSM_TOKENS_MIN)
25061 +               sched_pc->tokens_min = sched->tokens_min;
25062 +       if (set_mask & VXSM_TOKENS_MAX)
25063 +               sched_pc->tokens_max = sched->tokens_max;
25064 +       if (set_mask & VXSM_PRIO_BIAS)
25065 +               sched_pc->prio_bias = sched->prio_bias;
25066 +
25067 +       if (set_mask & VXSM_IDLE_TIME)
25068 +               sched_pc->flags |= VXSF_IDLE_TIME;
25069 +       else
25070 +               sched_pc->flags &= ~VXSF_IDLE_TIME;
25071 +
25072 +       /* reset time */
25073 +       sched_pc->norm_time = jiffies;
25074 +}
25075 +
25076 +
25077 +/*
25078 + * recalculate the context's scheduling tokens
25079 + *
25080 + * ret > 0 : number of tokens available
25081 + * ret < 0 : on hold, check delta_min[]
25082 + *          -1 only jiffies
25083 + *          -2 also idle time
25084 + *
25085 + */
25086 +int vx_tokens_recalc(struct _vx_sched_pc *sched_pc,
25087 +       unsigned long *norm_time, unsigned long *idle_time, int delta_min[2])
25088 +{
25089 +       long delta;
25090 +       long tokens = 0;
25091 +       int flags = sched_pc->flags;
25092 +
25093 +       /* how much time did pass? */
25094 +       delta = *norm_time - sched_pc->norm_time;
25095 +       vxd_check_range(delta, 0, INT_MAX);
25096 +
25097 +       if (delta >= sched_pc->interval[0]) {
25098 +               long tokens, integral;
25099 +
25100 +               /* calc integral token part */
25101 +               tokens = delta / sched_pc->interval[0];
25102 +               integral = tokens * sched_pc->interval[0];
25103 +               tokens *= sched_pc->fill_rate[0];
25104 +#ifdef CONFIG_VSERVER_HARDCPU
25105 +               delta_min[0] = delta - integral;
25106 +               vxd_check_range(delta_min[0], 0, sched_pc->interval[0]);
25107 +#endif
25108 +               /* advance time */
25109 +               sched_pc->norm_time += delta;
25110 +
25111 +               /* add tokens */
25112 +               sched_pc->tokens += tokens;
25113 +               sched_pc->token_time += tokens;
25114 +       } else
25115 +               delta_min[0] = delta;
25116 +
25117 +#ifdef CONFIG_VSERVER_IDLETIME
25118 +       if (!(flags & VXSF_IDLE_TIME))
25119 +               goto skip_idle;
25120 +
25121 +       /* how much was the idle skip? */
25122 +       delta = *idle_time - sched_pc->idle_time;
25123 +       vxd_check_range(delta, 0, INT_MAX);
25124 +
25125 +       if (delta >= sched_pc->interval[1]) {
25126 +               long tokens, integral;
25127 +
25128 +               /* calc fair share token part */
25129 +               tokens = delta / sched_pc->interval[1];
25130 +               integral = tokens * sched_pc->interval[1];
25131 +               tokens *= sched_pc->fill_rate[1];
25132 +               delta_min[1] = delta - integral;
25133 +               vxd_check_range(delta_min[1], 0, sched_pc->interval[1]);
25134 +
25135 +               /* advance idle time */
25136 +               sched_pc->idle_time += integral;
25137 +
25138 +               /* add tokens */
25139 +               sched_pc->tokens += tokens;
25140 +               sched_pc->token_time += tokens;
25141 +       } else
25142 +               delta_min[1] = delta;
25143 +skip_idle:
25144 +#endif
25145 +
25146 +       /* clip at maximum */
25147 +       if (sched_pc->tokens > sched_pc->tokens_max)
25148 +               sched_pc->tokens = sched_pc->tokens_max;
25149 +       tokens = sched_pc->tokens;
25150 +
25151 +       if ((flags & VXSF_ONHOLD)) {
25152 +               /* can we unhold? */
25153 +               if (tokens >= sched_pc->tokens_min) {
25154 +                       flags &= ~VXSF_ONHOLD;
25155 +                       sched_pc->hold_ticks +=
25156 +                               *norm_time - sched_pc->onhold;
25157 +               } else
25158 +                       goto on_hold;
25159 +       } else {
25160 +               /* put on hold? */
25161 +               if (tokens <= 0) {
25162 +                       flags |= VXSF_ONHOLD;
25163 +                       sched_pc->onhold = *norm_time;
25164 +                       goto on_hold;
25165 +               }
25166 +       }
25167 +       sched_pc->flags = flags;
25168 +       return tokens;
25169 +
25170 +on_hold:
25171 +       tokens = sched_pc->tokens_min - tokens;
25172 +       sched_pc->flags = flags;
25173 +       BUG_ON(tokens < 0);
25174 +
25175 +#ifdef CONFIG_VSERVER_HARDCPU
25176 +       /* next interval? */
25177 +       if (!sched_pc->fill_rate[0])
25178 +               delta_min[0] = HZ;
25179 +       else if (tokens > sched_pc->fill_rate[0])
25180 +               delta_min[0] += sched_pc->interval[0] *
25181 +                       tokens / sched_pc->fill_rate[0];
25182 +       else
25183 +               delta_min[0] = sched_pc->interval[0] - delta_min[0];
25184 +       vxd_check_range(delta_min[0], 0, INT_MAX);
25185 +
25186 +#ifdef CONFIG_VSERVER_IDLETIME
25187 +       if (!(flags & VXSF_IDLE_TIME))
25188 +               return -1;
25189 +
25190 +       /* next interval? */
25191 +       if (!sched_pc->fill_rate[1])
25192 +               delta_min[1] = HZ;
25193 +       else if (tokens > sched_pc->fill_rate[1])
25194 +               delta_min[1] += sched_pc->interval[1] *
25195 +                       tokens / sched_pc->fill_rate[1];
25196 +       else
25197 +               delta_min[1] = sched_pc->interval[1] - delta_min[1];
25198 +       vxd_check_range(delta_min[1], 0, INT_MAX);
25199 +
25200 +       return -2;
25201 +#else
25202 +       return -1;
25203 +#endif /* CONFIG_VSERVER_IDLETIME */
25204 +#else
25205 +       return 0;
25206 +#endif /* CONFIG_VSERVER_HARDCPU */
25207 +}
25208 +
25209 +static inline unsigned long msec_to_ticks(unsigned long msec)
25210 +{
25211 +       return msecs_to_jiffies(msec);
25212 +}
25213 +
25214 +static inline unsigned long ticks_to_msec(unsigned long ticks)
25215 +{
25216 +       return jiffies_to_msecs(ticks);
25217 +}
25218 +
25219 +static inline unsigned long ticks_to_usec(unsigned long ticks)
25220 +{
25221 +       return jiffies_to_usecs(ticks);
25222 +}
25223 +
25224 +
25225 +static int do_set_sched(struct vx_info *vxi, struct vcmd_sched_v5 *data)
25226 +{
25227 +       unsigned int set_mask = data->mask;
25228 +       unsigned int update_mask;
25229 +       int i, cpu;
25230 +
25231 +       /* Sanity check data values */
25232 +       if (data->tokens_max <= 0)
25233 +               data->tokens_max = HZ;
25234 +       if (data->tokens_min < 0)
25235 +               data->tokens_min = HZ / 3;
25236 +       if (data->tokens_min >= data->tokens_max)
25237 +               data->tokens_min = data->tokens_max;
25238 +
25239 +       if (data->prio_bias > MAX_PRIO_BIAS)
25240 +               data->prio_bias = MAX_PRIO_BIAS;
25241 +       if (data->prio_bias < MIN_PRIO_BIAS)
25242 +               data->prio_bias = MIN_PRIO_BIAS;
25243 +
25244 +       spin_lock(&vxi->sched.tokens_lock);
25245 +
25246 +       /* sync up on delayed updates */
25247 +       for_each_cpu_mask(cpu, vxi->sched.update)
25248 +               vx_update_sched_param(&vxi->sched,
25249 +                       &vx_per_cpu(vxi, sched_pc, cpu));
25250 +
25251 +       if (set_mask & VXSM_FILL_RATE)
25252 +               vxi->sched.fill_rate[0] = data->fill_rate[0];
25253 +       if (set_mask & VXSM_FILL_RATE2)
25254 +               vxi->sched.fill_rate[1] = data->fill_rate[1];
25255 +       if (set_mask & VXSM_INTERVAL)
25256 +               vxi->sched.interval[0] = (set_mask & VXSM_MSEC) ?
25257 +                       msec_to_ticks(data->interval[0]) : data->interval[0];
25258 +       if (set_mask & VXSM_INTERVAL2)
25259 +               vxi->sched.interval[1] = (set_mask & VXSM_MSEC) ?
25260 +                       msec_to_ticks(data->interval[1]) : data->interval[1];
25261 +       if (set_mask & VXSM_TOKENS)
25262 +               vxi->sched.tokens = data->tokens;
25263 +       if (set_mask & VXSM_TOKENS_MIN)
25264 +               vxi->sched.tokens_min = data->tokens_min;
25265 +       if (set_mask & VXSM_TOKENS_MAX)
25266 +               vxi->sched.tokens_max = data->tokens_max;
25267 +       if (set_mask & VXSM_PRIO_BIAS)
25268 +               vxi->sched.prio_bias = data->prio_bias;
25269 +
25270 +       /* Sanity check rate/interval */
25271 +       for (i = 0; i < 2; i++) {
25272 +               if (data->fill_rate[i] < 0)
25273 +                       data->fill_rate[i] = 0;
25274 +               if (data->interval[i] <= 0)
25275 +                       data->interval[i] = HZ;
25276 +       }
25277 +
25278 +       update_mask = vxi->sched.update_mask & VXSM_SET_MASK;
25279 +       update_mask |= (set_mask & (VXSM_SET_MASK | VXSM_IDLE_TIME));
25280 +       vxi->sched.update_mask = update_mask;
25281 +
25282 +#ifdef CONFIG_SMP
25283 +       rmb();
25284 +       if (set_mask & VXSM_CPU_ID) {
25285 +               vxi->sched.update = cpumask_of_cpu(data->cpu_id);
25286 +               cpus_and(vxi->sched.update, cpu_online_map,
25287 +                       vxi->sched.update);
25288 +       } else
25289 +               vxi->sched.update = cpu_online_map;
25290 +
25291 +       /* forced reload? */
25292 +       if (set_mask & VXSM_FORCE) {
25293 +               for_each_cpu_mask(cpu, vxi->sched.update)
25294 +                       vx_update_sched_param(&vxi->sched,
25295 +                               &vx_per_cpu(vxi, sched_pc, cpu));
25296 +               vxi->sched.update = CPU_MASK_NONE;
25297 +       }
25298 +#else
25299 +       /* on UP we update immediately */
25300 +       vx_update_sched_param(&vxi->sched,
25301 +               &vx_per_cpu(vxi, sched_pc, 0));
25302 +#endif
25303 +
25304 +       spin_unlock(&vxi->sched.tokens_lock);
25305 +       return 0;
25306 +}
25307 +
25308 +
25309 +#define COPY_IDS(C) C(cpu_id); C(bucket_id)
25310 +#define COPY_PRI(C) C(prio_bias)
25311 +#define COPY_TOK(C) C(tokens); C(tokens_min); C(tokens_max)
25312 +#define COPY_FRI(C) C(fill_rate[0]); C(interval[0]);   \
25313 +                   C(fill_rate[1]); C(interval[1]);
25314 +
25315 +#define COPY_VALUE(name) vc_data.name = data->name
25316 +
25317 +static int do_set_sched_v4(struct vx_info *vxi, struct vcmd_set_sched_v4 *data)
25318 +{
25319 +       struct vcmd_sched_v5 vc_data;
25320 +
25321 +       vc_data.mask = data->set_mask;
25322 +       COPY_IDS(COPY_VALUE);
25323 +       COPY_PRI(COPY_VALUE);
25324 +       COPY_TOK(COPY_VALUE);
25325 +       vc_data.fill_rate[0] = vc_data.fill_rate[1] = data->fill_rate;
25326 +       vc_data.interval[0] = vc_data.interval[1] = data->interval;
25327 +       return do_set_sched(vxi, &vc_data);
25328 +}
25329 +
25330 +int vc_set_sched_v4(struct vx_info *vxi, void __user *data)
25331 +{
25332 +       struct vcmd_set_sched_v4 vc_data;
25333 +
25334 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
25335 +               return -EFAULT;
25336 +
25337 +       return do_set_sched_v4(vxi, &vc_data);
25338 +}
25339 +
25340 +       /* latest interface is v5 */
25341 +
25342 +int vc_set_sched(struct vx_info *vxi, void __user *data)
25343 +{
25344 +       struct vcmd_sched_v5 vc_data;
25345 +
25346 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
25347 +               return -EFAULT;
25348 +
25349 +       return do_set_sched(vxi, &vc_data);
25350 +}
25351 +
25352 +
25353 +#define COPY_PRI(C) C(prio_bias)
25354 +#define COPY_TOK(C) C(tokens); C(tokens_min); C(tokens_max)
25355 +#define COPY_FRI(C) C(fill_rate[0]); C(interval[0]);    \
25356 +                   C(fill_rate[1]); C(interval[1]);
25357 +
25358 +#define COPY_VALUE(name) vc_data.name = data->name
25359 +
25360 +
25361 +int vc_get_sched(struct vx_info *vxi, void __user *data)
25362 +{
25363 +       struct vcmd_sched_v5 vc_data;
25364 +
25365 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
25366 +               return -EFAULT;
25367 +
25368 +       if (vc_data.mask & VXSM_CPU_ID) {
25369 +               int cpu = vc_data.cpu_id;
25370 +               struct _vx_sched_pc *data;
25371 +
25372 +               if (!cpu_possible(cpu))
25373 +                       return -EINVAL;
25374 +
25375 +               data = &vx_per_cpu(vxi, sched_pc, cpu);
25376 +               COPY_TOK(COPY_VALUE);
25377 +               COPY_PRI(COPY_VALUE);
25378 +               COPY_FRI(COPY_VALUE);
25379 +
25380 +               if (data->flags & VXSF_IDLE_TIME)
25381 +                       vc_data.mask |= VXSM_IDLE_TIME;
25382 +       } else {
25383 +               struct _vx_sched *data = &vxi->sched;
25384 +
25385 +               COPY_TOK(COPY_VALUE);
25386 +               COPY_PRI(COPY_VALUE);
25387 +               COPY_FRI(COPY_VALUE);
25388 +       }
25389 +
25390 +       if (vc_data.mask & VXSM_MSEC) {
25391 +               vc_data.interval[0] = ticks_to_msec(vc_data.interval[0]);
25392 +               vc_data.interval[1] = ticks_to_msec(vc_data.interval[1]);
25393 +       }
25394 +
25395 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
25396 +               return -EFAULT;
25397 +       return 0;
25398 +}
25399 +
25400 +
25401 +int vc_sched_info(struct vx_info *vxi, void __user *data)
25402 +{
25403 +       struct vcmd_sched_info vc_data;
25404 +       int cpu;
25405 +
25406 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
25407 +               return -EFAULT;
25408 +
25409 +       cpu = vc_data.cpu_id;
25410 +       if (!cpu_possible(cpu))
25411 +               return -EINVAL;
25412 +
25413 +       if (vxi) {
25414 +               struct _vx_sched_pc *sched_pc =
25415 +                       &vx_per_cpu(vxi, sched_pc, cpu);
25416 +
25417 +               vc_data.user_msec = ticks_to_msec(sched_pc->user_ticks);
25418 +               vc_data.sys_msec = ticks_to_msec(sched_pc->sys_ticks);
25419 +               vc_data.hold_msec = ticks_to_msec(sched_pc->hold_ticks);
25420 +               vc_data.vavavoom = sched_pc->vavavoom;
25421 +       }
25422 +       vc_data.token_usec = ticks_to_usec(1);
25423 +
25424 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
25425 +               return -EFAULT;
25426 +       return 0;
25427 +}
25428 +
25429 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/sched_init.h linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/sched_init.h
25430 --- linux-2.6.25.11/kernel/vserver/sched_init.h 1969-12-31 19:00:00.000000000 -0500
25431 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/sched_init.h   2008-04-19 15:14:52.000000000 -0400
25432 @@ -0,0 +1,50 @@
25433 +
25434 +static inline void vx_info_init_sched(struct _vx_sched *sched)
25435 +{
25436 +       static struct lock_class_key tokens_lock_key;
25437 +
25438 +       /* scheduling; hard code starting values as constants */
25439 +       sched->fill_rate[0]     = 1;
25440 +       sched->interval[0]      = 4;
25441 +       sched->fill_rate[1]     = 1;
25442 +       sched->interval[1]      = 8;
25443 +       sched->tokens           = HZ >> 2;
25444 +       sched->tokens_min       = HZ >> 4;
25445 +       sched->tokens_max       = HZ >> 1;
25446 +       sched->tokens_lock      = SPIN_LOCK_UNLOCKED;
25447 +       sched->prio_bias        = 0;
25448 +
25449 +       lockdep_set_class(&sched->tokens_lock, &tokens_lock_key);
25450 +}
25451 +
25452 +static inline
25453 +void vx_info_init_sched_pc(struct _vx_sched_pc *sched_pc, int cpu)
25454 +{
25455 +       sched_pc->fill_rate[0]  = 1;
25456 +       sched_pc->interval[0]   = 4;
25457 +       sched_pc->fill_rate[1]  = 1;
25458 +       sched_pc->interval[1]   = 8;
25459 +       sched_pc->tokens        = HZ >> 2;
25460 +       sched_pc->tokens_min    = HZ >> 4;
25461 +       sched_pc->tokens_max    = HZ >> 1;
25462 +       sched_pc->prio_bias     = 0;
25463 +       sched_pc->vavavoom      = 0;
25464 +       sched_pc->token_time    = 0;
25465 +       sched_pc->idle_time     = 0;
25466 +       sched_pc->norm_time     = jiffies;
25467 +
25468 +       sched_pc->user_ticks = 0;
25469 +       sched_pc->sys_ticks = 0;
25470 +       sched_pc->hold_ticks = 0;
25471 +}
25472 +
25473 +static inline void vx_info_exit_sched(struct _vx_sched *sched)
25474 +{
25475 +       return;
25476 +}
25477 +
25478 +static inline
25479 +void vx_info_exit_sched_pc(struct _vx_sched_pc *sched_pc, int cpu)
25480 +{
25481 +       return;
25482 +}
25483 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/sched_proc.h linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/sched_proc.h
25484 --- linux-2.6.25.11/kernel/vserver/sched_proc.h 1969-12-31 19:00:00.000000000 -0500
25485 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/sched_proc.h   2008-04-19 15:14:52.000000000 -0400
25486 @@ -0,0 +1,57 @@
25487 +#ifndef _VX_SCHED_PROC_H
25488 +#define _VX_SCHED_PROC_H
25489 +
25490 +
25491 +static inline
25492 +int vx_info_proc_sched(struct _vx_sched *sched, char *buffer)
25493 +{
25494 +       int length = 0;
25495 +
25496 +       length += sprintf(buffer,
25497 +               "FillRate:\t%8d,%d\n"
25498 +               "Interval:\t%8d,%d\n"
25499 +               "TokensMin:\t%8d\n"
25500 +               "TokensMax:\t%8d\n"
25501 +               "PrioBias:\t%8d\n",
25502 +               sched->fill_rate[0],
25503 +               sched->fill_rate[1],
25504 +               sched->interval[0],
25505 +               sched->interval[1],
25506 +               sched->tokens_min,
25507 +               sched->tokens_max,
25508 +               sched->prio_bias);
25509 +       return length;
25510 +}
25511 +
25512 +static inline
25513 +int vx_info_proc_sched_pc(struct _vx_sched_pc *sched_pc,
25514 +       char *buffer, int cpu)
25515 +{
25516 +       int length = 0;
25517 +
25518 +       length += sprintf(buffer + length,
25519 +               "cpu %d: %lld %lld %lld %ld %ld", cpu,
25520 +               (unsigned long long)sched_pc->user_ticks,
25521 +               (unsigned long long)sched_pc->sys_ticks,
25522 +               (unsigned long long)sched_pc->hold_ticks,
25523 +               sched_pc->token_time,
25524 +               sched_pc->idle_time);
25525 +       length += sprintf(buffer + length,
25526 +               " %c%c %d %d %d %d/%d %d/%d",
25527 +               (sched_pc->flags & VXSF_ONHOLD) ? 'H' : 'R',
25528 +               (sched_pc->flags & VXSF_IDLE_TIME) ? 'I' : '-',
25529 +               sched_pc->tokens,
25530 +               sched_pc->tokens_min,
25531 +               sched_pc->tokens_max,
25532 +               sched_pc->fill_rate[0],
25533 +               sched_pc->interval[0],
25534 +               sched_pc->fill_rate[1],
25535 +               sched_pc->interval[1]);
25536 +       length += sprintf(buffer + length,
25537 +               " %d %d\n",
25538 +               sched_pc->prio_bias,
25539 +               sched_pc->vavavoom);
25540 +       return length;
25541 +}
25542 +
25543 +#endif /* _VX_SCHED_PROC_H */
25544 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/signal.c linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/signal.c
25545 --- linux-2.6.25.11/kernel/vserver/signal.c     1969-12-31 19:00:00.000000000 -0500
25546 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/signal.c       2008-04-19 15:14:52.000000000 -0400
25547 @@ -0,0 +1,132 @@
25548 +/*
25549 + *  linux/kernel/vserver/signal.c
25550 + *
25551 + *  Virtual Server: Signal Support
25552 + *
25553 + *  Copyright (C) 2003-2007  Herbert Pötzl
25554 + *
25555 + *  V0.01  broken out from vcontext V0.05
25556 + *  V0.02  changed vcmds to vxi arg
25557 + *  V0.03  adjusted siginfo for kill
25558 + *
25559 + */
25560 +
25561 +#include <asm/uaccess.h>
25562 +
25563 +#include <linux/vs_context.h>
25564 +#include <linux/vs_pid.h>
25565 +#include <linux/vserver/signal_cmd.h>
25566 +
25567 +
25568 +int vx_info_kill(struct vx_info *vxi, int pid, int sig)
25569 +{
25570 +       int retval, count = 0;
25571 +       struct task_struct *p;
25572 +       struct siginfo *sip = SEND_SIG_PRIV;
25573 +
25574 +       retval = -ESRCH;
25575 +       vxdprintk(VXD_CBIT(misc, 4),
25576 +               "vx_info_kill(%p[#%d],%d,%d)*",
25577 +               vxi, vxi->vx_id, pid, sig);
25578 +       read_lock(&tasklist_lock);
25579 +       switch (pid) {
25580 +       case  0:
25581 +       case -1:
25582 +               for_each_process(p) {
25583 +                       int err = 0;
25584 +
25585 +                       if (vx_task_xid(p) != vxi->vx_id || p->pid <= 1 ||
25586 +                               (pid && vxi->vx_initpid == p->pid))
25587 +                               continue;
25588 +
25589 +                       err = group_send_sig_info(sig, sip, p);
25590 +                       ++count;
25591 +                       if (err != -EPERM)
25592 +                               retval = err;
25593 +               }
25594 +               break;
25595 +
25596 +       case 1:
25597 +               if (vxi->vx_initpid) {
25598 +                       pid = vxi->vx_initpid;
25599 +                       /* for now, only SIGINT to private init ... */
25600 +                       if (!vx_info_flags(vxi, VXF_STATE_ADMIN, 0) &&
25601 +                               /* ... as long as there are tasks left */
25602 +                               (atomic_read(&vxi->vx_tasks) > 1))
25603 +                               sig = SIGINT;
25604 +               }
25605 +               /* fallthrough */
25606 +       default:
25607 +               p = find_task_by_real_pid(pid);
25608 +               if (p) {
25609 +                       if (vx_task_xid(p) == vxi->vx_id)
25610 +                               retval = group_send_sig_info(sig, sip, p);
25611 +               }
25612 +               break;
25613 +       }
25614 +       read_unlock(&tasklist_lock);
25615 +       vxdprintk(VXD_CBIT(misc, 4),
25616 +               "vx_info_kill(%p[#%d],%d,%d,%ld) = %d",
25617 +               vxi, vxi->vx_id, pid, sig, (long)sip, retval);
25618 +       return retval;
25619 +}
25620 +
25621 +int vc_ctx_kill(struct vx_info *vxi, void __user *data)
25622 +{
25623 +       struct vcmd_ctx_kill_v0 vc_data;
25624 +
25625 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
25626 +               return -EFAULT;
25627 +
25628 +       /* special check to allow guest shutdown */
25629 +       if (!vx_info_flags(vxi, VXF_STATE_ADMIN, 0) &&
25630 +               /* forbid killall pid=0 when init is present */
25631 +               (((vc_data.pid < 1) && vxi->vx_initpid) ||
25632 +               (vc_data.pid > 1)))
25633 +               return -EACCES;
25634 +
25635 +       return vx_info_kill(vxi, vc_data.pid, vc_data.sig);
25636 +}
25637 +
25638 +
25639 +static int __wait_exit(struct vx_info *vxi)
25640 +{
25641 +       DECLARE_WAITQUEUE(wait, current);
25642 +       int ret = 0;
25643 +
25644 +       add_wait_queue(&vxi->vx_wait, &wait);
25645 +       set_current_state(TASK_INTERRUPTIBLE);
25646 +
25647 +wait:
25648 +       if (vx_info_state(vxi,
25649 +               VXS_SHUTDOWN | VXS_HASHED | VXS_HELPER) == VXS_SHUTDOWN)
25650 +               goto out;
25651 +       if (signal_pending(current)) {
25652 +               ret = -ERESTARTSYS;
25653 +               goto out;
25654 +       }
25655 +       schedule();
25656 +       goto wait;
25657 +
25658 +out:
25659 +       set_current_state(TASK_RUNNING);
25660 +       remove_wait_queue(&vxi->vx_wait, &wait);
25661 +       return ret;
25662 +}
25663 +
25664 +
25665 +
25666 +int vc_wait_exit(struct vx_info *vxi, void __user *data)
25667 +{
25668 +       struct vcmd_wait_exit_v0 vc_data;
25669 +       int ret;
25670 +
25671 +       ret = __wait_exit(vxi);
25672 +       vc_data.reboot_cmd = vxi->reboot_cmd;
25673 +       vc_data.exit_code = vxi->exit_code;
25674 +
25675 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
25676 +               ret = -EFAULT;
25677 +       return ret;
25678 +}
25679 +
25680 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/space.c linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/space.c
25681 --- linux-2.6.25.11/kernel/vserver/space.c      1969-12-31 19:00:00.000000000 -0500
25682 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/space.c        2008-07-16 00:21:04.000000000 -0400
25683 @@ -0,0 +1,372 @@
25684 +/*
25685 + *  linux/kernel/vserver/space.c
25686 + *
25687 + *  Virtual Server: Context Space Support
25688 + *
25689 + *  Copyright (C) 2003-2007  Herbert Pötzl
25690 + *
25691 + *  V0.01  broken out from context.c 0.07
25692 + *  V0.02  added task locking for namespace
25693 + *  V0.03  broken out vx_enter_namespace
25694 + *  V0.04  added *space support and commands
25695 + *
25696 + */
25697 +
25698 +#include <linux/utsname.h>
25699 +#include <linux/nsproxy.h>
25700 +#include <linux/err.h>
25701 +#include <asm/uaccess.h>
25702 +
25703 +#include <linux/vs_context.h>
25704 +#include <linux/vserver/space.h>
25705 +#include <linux/vserver/space_cmd.h>
25706 +
25707 +atomic_t vs_global_nsproxy     = ATOMIC_INIT(0);
25708 +atomic_t vs_global_fs          = ATOMIC_INIT(0);
25709 +atomic_t vs_global_mnt_ns      = ATOMIC_INIT(0);
25710 +atomic_t vs_global_uts_ns      = ATOMIC_INIT(0);
25711 +atomic_t vs_global_ipc_ns      = ATOMIC_INIT(0);
25712 +atomic_t vs_global_user_ns     = ATOMIC_INIT(0);
25713 +atomic_t vs_global_pid_ns      = ATOMIC_INIT(0);
25714 +
25715 +
25716 +/* namespace functions */
25717 +
25718 +#include <linux/mnt_namespace.h>
25719 +#include <linux/user_namespace.h>
25720 +#include <linux/pid_namespace.h>
25721 +#include <linux/ipc_namespace.h>
25722 +#include <net/net_namespace.h>
25723 +
25724 +
25725 +static const struct vcmd_space_mask space_mask_v0 = {
25726 +       .mask = CLONE_FS |
25727 +               CLONE_NEWNS |
25728 +               CLONE_NEWUTS |
25729 +               CLONE_NEWIPC |
25730 +               CLONE_NEWUSER |
25731 +               0
25732 +};
25733 +
25734 +static const struct vcmd_space_mask space_mask = {
25735 +       .mask = CLONE_FS |
25736 +               CLONE_NEWNS |
25737 +               CLONE_NEWUTS |
25738 +               CLONE_NEWIPC |
25739 +               CLONE_NEWUSER |
25740 +#ifdef CONFIG_PID_NS
25741 +               CLONE_NEWPID |
25742 +#endif
25743 +#ifdef CONFIG_NET_NS
25744 +               CLONE_NEWNET |
25745 +#endif
25746 +               0
25747 +};
25748 +
25749 +static const struct vcmd_space_mask default_space_mask = {
25750 +       .mask = CLONE_FS |
25751 +               CLONE_NEWNS |
25752 +               CLONE_NEWUTS |
25753 +               CLONE_NEWIPC |
25754 +               CLONE_NEWUSER |
25755 +#ifdef CONFIG_PID_NS
25756 +//             CLONE_NEWPID |
25757 +#endif
25758 +               0
25759 +};
25760 +
25761 +/*
25762 + *     build a new nsproxy mix
25763 + *      assumes that both proxies are 'const'
25764 + *     does not touch nsproxy refcounts
25765 + *     will hold a reference on the result.
25766 + */
25767 +
25768 +struct nsproxy *vs_mix_nsproxy(struct nsproxy *old_nsproxy,
25769 +       struct nsproxy *new_nsproxy, unsigned long mask)
25770 +{
25771 +       struct mnt_namespace *old_ns;
25772 +       struct uts_namespace *old_uts;
25773 +       struct ipc_namespace *old_ipc;
25774 +       struct user_namespace *old_user;
25775 +#ifdef CONFIG_PID_NS
25776 +       struct pid_namespace *old_pid;
25777 +#endif
25778 +#ifdef CONFIG_NET_NS
25779 +       struct net *old_net;
25780 +#endif
25781 +       struct nsproxy *nsproxy;
25782 +
25783 +       nsproxy = copy_nsproxy(old_nsproxy);
25784 +       if (!nsproxy)
25785 +               goto out;
25786 +
25787 +       if (mask & CLONE_NEWNS) {
25788 +               old_ns = nsproxy->mnt_ns;
25789 +               nsproxy->mnt_ns = new_nsproxy->mnt_ns;
25790 +               if (nsproxy->mnt_ns)
25791 +                       get_mnt_ns(nsproxy->mnt_ns);
25792 +       } else
25793 +               old_ns = NULL;
25794 +
25795 +       if (mask & CLONE_NEWUTS) {
25796 +               old_uts = nsproxy->uts_ns;
25797 +               nsproxy->uts_ns = new_nsproxy->uts_ns;
25798 +               if (nsproxy->uts_ns)
25799 +                       get_uts_ns(nsproxy->uts_ns);
25800 +       } else
25801 +               old_uts = NULL;
25802 +
25803 +       if (mask & CLONE_NEWIPC) {
25804 +               old_ipc = nsproxy->ipc_ns;
25805 +               nsproxy->ipc_ns = new_nsproxy->ipc_ns;
25806 +               if (nsproxy->ipc_ns)
25807 +                       get_ipc_ns(nsproxy->ipc_ns);
25808 +       } else
25809 +               old_ipc = NULL;
25810 +
25811 +       if (mask & CLONE_NEWUSER) {
25812 +               old_user = nsproxy->user_ns;
25813 +               nsproxy->user_ns = new_nsproxy->user_ns;
25814 +               if (nsproxy->user_ns)
25815 +                       get_user_ns(nsproxy->user_ns);
25816 +       } else
25817 +               old_user = NULL;
25818 +
25819 +#ifdef CONFIG_PID_NS
25820 +       if (mask & CLONE_NEWPID) {
25821 +               old_pid = nsproxy->pid_ns;
25822 +               nsproxy->pid_ns = new_nsproxy->pid_ns;
25823 +               if (nsproxy->pid_ns)
25824 +                       get_pid_ns(nsproxy->pid_ns);
25825 +       } else
25826 +               old_pid = NULL;
25827 +#endif
25828 +#ifdef CONFIG_NET_NS
25829 +       if (mask & CLONE_NEWNET) {
25830 +               old_net = nsproxy->net_ns;
25831 +               nsproxy->net_ns = new_nsproxy->net_ns;
25832 +               if (nsproxy->net_ns)
25833 +                       get_net(nsproxy->net_ns);
25834 +       } else
25835 +               old_net = NULL;
25836 +#endif
25837 +       if (old_ns)
25838 +               put_mnt_ns(old_ns);
25839 +       if (old_uts)
25840 +               put_uts_ns(old_uts);
25841 +       if (old_ipc)
25842 +               put_ipc_ns(old_ipc);
25843 +       if (old_user)
25844 +               put_user_ns(old_user);
25845 +#ifdef CONFIG_PID_NS
25846 +       if (old_pid)
25847 +               put_pid_ns(old_pid);
25848 +#endif
25849 +#ifdef CONFIG_NET_NS
25850 +       if (old_net)
25851 +               put_net(old_net);
25852 +#endif
25853 +out:
25854 +       return nsproxy;
25855 +}
25856 +
25857 +
25858 +/*
25859 + *     merge two nsproxy structs into a new one.
25860 + *     will hold a reference on the result.
25861 + */
25862 +
25863 +static inline
25864 +struct nsproxy *__vs_merge_nsproxy(struct nsproxy *old,
25865 +       struct nsproxy *proxy, unsigned long mask)
25866 +{
25867 +       struct nsproxy null_proxy = { .mnt_ns = NULL };
25868 +
25869 +       if (!proxy)
25870 +               return NULL;
25871 +
25872 +       if (mask) {
25873 +               /* vs_mix_nsproxy returns with reference */
25874 +               return vs_mix_nsproxy(old ? old : &null_proxy,
25875 +                       proxy, mask);
25876 +       }
25877 +       get_nsproxy(proxy);
25878 +       return proxy;
25879 +}
25880 +
25881 +/*
25882 + *     merge two fs structs into a new one.
25883 + *     will take a reference on the result.
25884 + */
25885 +
25886 +static inline
25887 +struct fs_struct *__vs_merge_fs(struct fs_struct *old,
25888 +       struct fs_struct *fs, unsigned long mask)
25889 +{
25890 +       if (!(mask & CLONE_FS)) {
25891 +               if (old)
25892 +                       atomic_inc(&old->count);
25893 +               return old;
25894 +       }
25895 +
25896 +       if (!fs)
25897 +               return NULL;
25898 +
25899 +       return copy_fs_struct(fs);
25900 +}
25901 +
25902 +
25903 +int vx_enter_space(struct vx_info *vxi, unsigned long mask)
25904 +{
25905 +       struct nsproxy *proxy, *proxy_cur, *proxy_new;
25906 +       struct fs_struct *fs, *fs_cur, *fs_new;
25907 +       int ret;
25908 +
25909 +       vxdprintk(VXD_CBIT(space, 8), "vx_enter_space(%p[#%u],0x%08lx)",
25910 +               vxi, vxi->vx_id, mask);
25911 +
25912 +       if (vx_info_flags(vxi, VXF_INFO_PRIVATE, 0))
25913 +               return -EACCES;
25914 +
25915 +       if (!mask)
25916 +               mask = vxi->vx_nsmask;
25917 +
25918 +       if ((mask & vxi->vx_nsmask) != mask)
25919 +               return -EINVAL;
25920 +
25921 +       proxy = vxi->vx_nsproxy;
25922 +       fs = vxi->vx_fs;
25923 +
25924 +       task_lock(current);
25925 +       fs_cur = current->fs;
25926 +       atomic_inc(&fs_cur->count);
25927 +       proxy_cur = current->nsproxy;
25928 +       get_nsproxy(proxy_cur);
25929 +       task_unlock(current);
25930 +
25931 +       fs_new = __vs_merge_fs(fs_cur, fs, mask);
25932 +       if (IS_ERR(fs_new)) {
25933 +               ret = PTR_ERR(fs_new);
25934 +               goto out_put;
25935 +       }
25936 +
25937 +       proxy_new = __vs_merge_nsproxy(proxy_cur, proxy, mask);
25938 +       if (IS_ERR(proxy_new)) {
25939 +               ret = PTR_ERR(proxy_new);
25940 +               goto out_put_fs;
25941 +       }
25942 +
25943 +       fs_new = xchg(&current->fs, fs_new);
25944 +       proxy_new = xchg(&current->nsproxy, proxy_new);
25945 +       ret = 0;
25946 +
25947 +       if (proxy_new)
25948 +               put_nsproxy(proxy_new);
25949 +out_put_fs:
25950 +       if (fs_new)
25951 +               put_fs_struct(fs_new);
25952 +out_put:
25953 +       if (proxy_cur)
25954 +               put_nsproxy(proxy_cur);
25955 +       if (fs_cur)
25956 +               put_fs_struct(fs_cur);
25957 +       return ret;
25958 +}
25959 +
25960 +
25961 +int vx_set_space(struct vx_info *vxi, unsigned long mask)
25962 +{
25963 +       struct nsproxy *proxy_vxi, *proxy_cur, *proxy_new;
25964 +       struct fs_struct *fs_vxi, *fs_cur, *fs_new;
25965 +       int ret;
25966 +
25967 +       vxdprintk(VXD_CBIT(space, 8), "vx_set_space(%p[#%u],0x%08lx)",
25968 +               vxi, vxi->vx_id, mask);
25969 +#if 0
25970 +       if (!mask)
25971 +               mask = default_space_mask.mask;
25972 +#endif
25973 +       if ((mask & space_mask.mask) != mask)
25974 +               return -EINVAL;
25975 +
25976 +       proxy_vxi = vxi->vx_nsproxy;
25977 +       fs_vxi = vxi->vx_fs;
25978 +
25979 +       task_lock(current);
25980 +       fs_cur = current->fs;
25981 +       atomic_inc(&fs_cur->count);
25982 +       proxy_cur = current->nsproxy;
25983 +       get_nsproxy(proxy_cur);
25984 +       task_unlock(current);
25985 +
25986 +       fs_new = __vs_merge_fs(fs_vxi, fs_cur, mask);
25987 +       if (IS_ERR(fs_new)) {
25988 +               ret = PTR_ERR(fs_new);
25989 +               goto out_put;
25990 +       }
25991 +
25992 +       proxy_new = __vs_merge_nsproxy(proxy_vxi, proxy_cur, mask);
25993 +       if (IS_ERR(proxy_new)) {
25994 +               ret = PTR_ERR(proxy_new);
25995 +               goto out_put_fs;
25996 +       }
25997 +
25998 +       fs_new = xchg(&vxi->vx_fs, fs_new);
25999 +       proxy_new = xchg(&vxi->vx_nsproxy, proxy_new);
26000 +       vxi->vx_nsmask |= mask;
26001 +       ret = 0;
26002 +
26003 +       if (proxy_new)
26004 +               put_nsproxy(proxy_new);
26005 +out_put_fs:
26006 +       if (fs_new)
26007 +               put_fs_struct(fs_new);
26008 +out_put:
26009 +       if (proxy_cur)
26010 +               put_nsproxy(proxy_cur);
26011 +       if (fs_cur)
26012 +               put_fs_struct(fs_cur);
26013 +       return ret;
26014 +}
26015 +
26016 +
26017 +int vc_enter_space(struct vx_info *vxi, void __user *data)
26018 +{
26019 +       struct vcmd_space_mask vc_data = { .mask = 0 };
26020 +
26021 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
26022 +               return -EFAULT;
26023 +
26024 +       return vx_enter_space(vxi, vc_data.mask);
26025 +}
26026 +
26027 +int vc_set_space(struct vx_info *vxi, void __user *data)
26028 +{
26029 +       struct vcmd_space_mask vc_data = { .mask = 0 };
26030 +
26031 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
26032 +               return -EFAULT;
26033 +
26034 +       return vx_set_space(vxi, vc_data.mask);
26035 +}
26036 +
26037 +int vc_get_space_mask(void __user *data, int type)
26038 +{
26039 +       const struct vcmd_space_mask *mask;
26040 +
26041 +       if (type == 0)
26042 +               mask = &space_mask_v0;
26043 +       else if (type == 1)
26044 +               mask = &space_mask;
26045 +       else
26046 +               mask = &default_space_mask;
26047 +
26048 +       vxdprintk(VXD_CBIT(space, 10),
26049 +               "vc_get_space_mask(%d) = %08llx", type, mask->mask);
26050 +
26051 +       if (copy_to_user(data, mask, sizeof(struct vcmd_space_mask)))
26052 +               return -EFAULT;
26053 +       return 0;
26054 +}
26055 +
26056 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/switch.c linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/switch.c
26057 --- linux-2.6.25.11/kernel/vserver/switch.c     1969-12-31 19:00:00.000000000 -0500
26058 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/switch.c       2008-07-16 00:21:47.000000000 -0400
26059 @@ -0,0 +1,537 @@
26060 +/*
26061 + *  linux/kernel/vserver/switch.c
26062 + *
26063 + *  Virtual Server: Syscall Switch
26064 + *
26065 + *  Copyright (C) 2003-2007  Herbert Pötzl
26066 + *
26067 + *  V0.01  syscall switch
26068 + *  V0.02  added signal to context
26069 + *  V0.03  added rlimit functions
26070 + *  V0.04  added iattr, task/xid functions
26071 + *  V0.05  added debug/history stuff
26072 + *  V0.06  added compat32 layer
26073 + *  V0.07  vcmd args and perms
26074 + *  V0.08  added status commands
26075 + *  V0.09  added tag commands
26076 + *  V0.10  added oom bias
26077 + *  V0.11  added device commands
26078 + *
26079 + */
26080 +
26081 +#include <linux/vs_context.h>
26082 +#include <linux/vs_network.h>
26083 +#include <linux/vserver/switch.h>
26084 +
26085 +#include "vci_config.h"
26086 +
26087 +
26088 +static inline
26089 +int vc_get_version(uint32_t id)
26090 +{
26091 +       return VCI_VERSION;
26092 +}
26093 +
26094 +static inline
26095 +int vc_get_vci(uint32_t id)
26096 +{
26097 +       return vci_kernel_config();
26098 +}
26099 +
26100 +#include <linux/vserver/context_cmd.h>
26101 +#include <linux/vserver/cvirt_cmd.h>
26102 +#include <linux/vserver/cacct_cmd.h>
26103 +#include <linux/vserver/limit_cmd.h>
26104 +#include <linux/vserver/network_cmd.h>
26105 +#include <linux/vserver/sched_cmd.h>
26106 +#include <linux/vserver/debug_cmd.h>
26107 +#include <linux/vserver/inode_cmd.h>
26108 +#include <linux/vserver/dlimit_cmd.h>
26109 +#include <linux/vserver/signal_cmd.h>
26110 +#include <linux/vserver/space_cmd.h>
26111 +#include <linux/vserver/tag_cmd.h>
26112 +#include <linux/vserver/device_cmd.h>
26113 +
26114 +#include <linux/vserver/inode.h>
26115 +#include <linux/vserver/dlimit.h>
26116 +
26117 +
26118 +#ifdef CONFIG_COMPAT
26119 +#define __COMPAT(name, id, data, compat)       \
26120 +       (compat) ? name ## _x32(id, data) : name(id, data)
26121 +#define __COMPAT_NO_ID(name, data, compat)     \
26122 +       (compat) ? name ## _x32(data) : name(data)
26123 +#else
26124 +#define __COMPAT(name, id, data, compat)       \
26125 +       name(id, data)
26126 +#define __COMPAT_NO_ID(name, data, compat)     \
26127 +       name(data)
26128 +#endif
26129 +
26130 +
26131 +static inline
26132 +long do_vcmd(uint32_t cmd, uint32_t id,
26133 +       struct vx_info *vxi, struct nx_info *nxi,
26134 +       void __user *data, int compat)
26135 +{
26136 +       switch (cmd) {
26137 +
26138 +       case VCMD_get_version:
26139 +               return vc_get_version(id);
26140 +       case VCMD_get_vci:
26141 +               return vc_get_vci(id);
26142 +
26143 +       case VCMD_task_xid:
26144 +               return vc_task_xid(id);
26145 +       case VCMD_vx_info:
26146 +               return vc_vx_info(vxi, data);
26147 +
26148 +       case VCMD_task_nid:
26149 +               return vc_task_nid(id);
26150 +       case VCMD_nx_info:
26151 +               return vc_nx_info(nxi, data);
26152 +
26153 +       case VCMD_task_tag:
26154 +               return vc_task_tag(id);
26155 +
26156 +       /* this is version 1 */
26157 +       case VCMD_set_space:
26158 +               return vc_set_space(vxi, data);
26159 +
26160 +       case VCMD_get_space_mask_v0:
26161 +               return vc_get_space_mask(data, 0);
26162 +       /* this is version 1 */
26163 +       case VCMD_get_space_mask:
26164 +               return vc_get_space_mask(data, 1);
26165 +
26166 +       case VCMD_get_space_default:
26167 +               return vc_get_space_mask(data, -1);
26168 +
26169 +#ifdef CONFIG_IA32_EMULATION
26170 +       case VCMD_get_rlimit:
26171 +               return __COMPAT(vc_get_rlimit, vxi, data, compat);
26172 +       case VCMD_set_rlimit:
26173 +               return __COMPAT(vc_set_rlimit, vxi, data, compat);
26174 +#else
26175 +       case VCMD_get_rlimit:
26176 +               return vc_get_rlimit(vxi, data);
26177 +       case VCMD_set_rlimit:
26178 +               return vc_set_rlimit(vxi, data);
26179 +#endif
26180 +       case VCMD_get_rlimit_mask:
26181 +               return vc_get_rlimit_mask(id, data);
26182 +       case VCMD_reset_minmax:
26183 +               return vc_reset_minmax(vxi, data);
26184 +
26185 +       case VCMD_get_vhi_name:
26186 +               return vc_get_vhi_name(vxi, data);
26187 +       case VCMD_set_vhi_name:
26188 +               return vc_set_vhi_name(vxi, data);
26189 +
26190 +       case VCMD_ctx_stat:
26191 +               return vc_ctx_stat(vxi, data);
26192 +       case VCMD_virt_stat:
26193 +               return vc_virt_stat(vxi, data);
26194 +       case VCMD_sock_stat:
26195 +               return vc_sock_stat(vxi, data);
26196 +       case VCMD_rlimit_stat:
26197 +               return vc_rlimit_stat(vxi, data);
26198 +
26199 +       case VCMD_set_cflags:
26200 +               return vc_set_cflags(vxi, data);
26201 +       case VCMD_get_cflags:
26202 +               return vc_get_cflags(vxi, data);
26203 +
26204 +       /* this is version 1 */
26205 +       case VCMD_set_ccaps:
26206 +               return vc_set_ccaps(vxi, data);
26207 +       /* this is version 1 */
26208 +       case VCMD_get_ccaps:
26209 +               return vc_get_ccaps(vxi, data);
26210 +       case VCMD_set_bcaps:
26211 +               return vc_set_bcaps(vxi, data);
26212 +       case VCMD_get_bcaps:
26213 +               return vc_get_bcaps(vxi, data);
26214 +
26215 +       case VCMD_set_badness:
26216 +               return vc_set_badness(vxi, data);
26217 +       case VCMD_get_badness:
26218 +               return vc_get_badness(vxi, data);
26219 +
26220 +       case VCMD_set_nflags:
26221 +               return vc_set_nflags(nxi, data);
26222 +       case VCMD_get_nflags:
26223 +               return vc_get_nflags(nxi, data);
26224 +
26225 +       case VCMD_set_ncaps:
26226 +               return vc_set_ncaps(nxi, data);
26227 +       case VCMD_get_ncaps:
26228 +               return vc_get_ncaps(nxi, data);
26229 +
26230 +       case VCMD_set_sched_v4:
26231 +               return vc_set_sched_v4(vxi, data);
26232 +       /* this is version 5 */
26233 +       case VCMD_set_sched:
26234 +               return vc_set_sched(vxi, data);
26235 +       case VCMD_get_sched:
26236 +               return vc_get_sched(vxi, data);
26237 +       case VCMD_sched_info:
26238 +               return vc_sched_info(vxi, data);
26239 +
26240 +       case VCMD_add_dlimit:
26241 +               return __COMPAT(vc_add_dlimit, id, data, compat);
26242 +       case VCMD_rem_dlimit:
26243 +               return __COMPAT(vc_rem_dlimit, id, data, compat);
26244 +       case VCMD_set_dlimit:
26245 +               return __COMPAT(vc_set_dlimit, id, data, compat);
26246 +       case VCMD_get_dlimit:
26247 +               return __COMPAT(vc_get_dlimit, id, data, compat);
26248 +
26249 +       case VCMD_ctx_kill:
26250 +               return vc_ctx_kill(vxi, data);
26251 +
26252 +       case VCMD_wait_exit:
26253 +               return vc_wait_exit(vxi, data);
26254 +
26255 +       case VCMD_get_iattr:
26256 +               return __COMPAT_NO_ID(vc_get_iattr, data, compat);
26257 +       case VCMD_set_iattr:
26258 +               return __COMPAT_NO_ID(vc_set_iattr, data, compat);
26259 +
26260 +       case VCMD_fget_iattr:
26261 +               return vc_fget_iattr(id, data);
26262 +       case VCMD_fset_iattr:
26263 +               return vc_fset_iattr(id, data);
26264 +
26265 +       case VCMD_enter_space_v0:
26266 +               return vc_enter_space(vxi, NULL);
26267 +       /* this is version 1 */
26268 +       case VCMD_enter_space:
26269 +               return vc_enter_space(vxi, data);
26270 +
26271 +       case VCMD_ctx_create_v0:
26272 +               return vc_ctx_create(id, NULL);
26273 +       case VCMD_ctx_create:
26274 +               return vc_ctx_create(id, data);
26275 +       case VCMD_ctx_migrate_v0:
26276 +               return vc_ctx_migrate(vxi, NULL);
26277 +       case VCMD_ctx_migrate:
26278 +               return vc_ctx_migrate(vxi, data);
26279 +
26280 +       case VCMD_net_create_v0:
26281 +               return vc_net_create(id, NULL);
26282 +       case VCMD_net_create:
26283 +               return vc_net_create(id, data);
26284 +       case VCMD_net_migrate:
26285 +               return vc_net_migrate(nxi, data);
26286 +
26287 +       case VCMD_tag_migrate:
26288 +               return vc_tag_migrate(id);
26289 +
26290 +       case VCMD_net_add:
26291 +               return vc_net_add(nxi, data);
26292 +       case VCMD_net_remove:
26293 +               return vc_net_remove(nxi, data);
26294 +
26295 +       case VCMD_net_add_ipv4:
26296 +               return vc_net_add_ipv4(nxi, data);
26297 +       case VCMD_net_remove_ipv4:
26298 +               return vc_net_remove_ipv4(nxi, data);
26299 +#ifdef CONFIG_IPV6
26300 +       case VCMD_net_add_ipv6:
26301 +               return vc_net_add_ipv6(nxi, data);
26302 +       case VCMD_net_remove_ipv6:
26303 +               return vc_net_remove_ipv6(nxi, data);
26304 +#endif
26305 +/*     case VCMD_add_match_ipv4:
26306 +               return vc_add_match_ipv4(nxi, data);
26307 +       case VCMD_get_match_ipv4:
26308 +               return vc_get_match_ipv4(nxi, data);
26309 +#ifdef CONFIG_IPV6
26310 +       case VCMD_add_match_ipv6:
26311 +               return vc_add_match_ipv6(nxi, data);
26312 +       case VCMD_get_match_ipv6:
26313 +               return vc_get_match_ipv6(nxi, data);
26314 +#endif */
26315 +
26316 +#ifdef CONFIG_VSERVER_DEVICE
26317 +       case VCMD_set_mapping:
26318 +               return __COMPAT(vc_set_mapping, vxi, data, compat);
26319 +       case VCMD_unset_mapping:
26320 +               return __COMPAT(vc_unset_mapping, vxi, data, compat);
26321 +#endif
26322 +#ifdef CONFIG_VSERVER_HISTORY
26323 +       case VCMD_dump_history:
26324 +               return vc_dump_history(id);
26325 +       case VCMD_read_history:
26326 +               return __COMPAT(vc_read_history, id, data, compat);
26327 +#endif
26328 +#ifdef CONFIG_VSERVER_MONITOR
26329 +       case VCMD_read_monitor:
26330 +               return __COMPAT(vc_read_monitor, id, data, compat);
26331 +#endif
26332 +       default:
26333 +               vxwprintk_task(1, "unimplemented VCMD_%02d_%d[%d]",
26334 +                       VC_CATEGORY(cmd), VC_COMMAND(cmd), VC_VERSION(cmd));
26335 +       }
26336 +       return -ENOSYS;
26337 +}
26338 +
26339 +
26340 +#define        __VCMD(vcmd, _perm, _args, _flags)              \
26341 +       case VCMD_ ## vcmd: perm = _perm;               \
26342 +               args = _args; flags = _flags; break
26343 +
26344 +
26345 +#define VCA_NONE       0x00
26346 +#define VCA_VXI                0x01
26347 +#define VCA_NXI                0x02
26348 +
26349 +#define VCF_NONE       0x00
26350 +#define VCF_INFO       0x01
26351 +#define VCF_ADMIN      0x02
26352 +#define VCF_ARES       0x06    /* includes admin */
26353 +#define VCF_SETUP      0x08
26354 +
26355 +#define VCF_ZIDOK      0x10    /* zero id okay */
26356 +
26357 +
26358 +static inline
26359 +long do_vserver(uint32_t cmd, uint32_t id, void __user *data, int compat)
26360 +{
26361 +       long ret;
26362 +       int permit = -1, state = 0;
26363 +       int perm = -1, args = 0, flags = 0;
26364 +       struct vx_info *vxi = NULL;
26365 +       struct nx_info *nxi = NULL;
26366 +
26367 +       switch (cmd) {
26368 +       /* unpriviledged commands */
26369 +       __VCMD(get_version,      0, VCA_NONE,   0);
26370 +       __VCMD(get_vci,          0, VCA_NONE,   0);
26371 +       __VCMD(get_rlimit_mask,  0, VCA_NONE,   0);
26372 +       __VCMD(get_space_mask_v0,0, VCA_NONE,   0);
26373 +       __VCMD(get_space_mask,   0, VCA_NONE,   0);
26374 +       __VCMD(get_space_default,0, VCA_NONE,   0);
26375 +
26376 +       /* info commands */
26377 +       __VCMD(task_xid,         2, VCA_NONE,   0);
26378 +       __VCMD(reset_minmax,     2, VCA_VXI,    0);
26379 +       __VCMD(vx_info,          3, VCA_VXI,    VCF_INFO);
26380 +       __VCMD(get_bcaps,        3, VCA_VXI,    VCF_INFO);
26381 +       __VCMD(get_ccaps,        3, VCA_VXI,    VCF_INFO);
26382 +       __VCMD(get_cflags,       3, VCA_VXI,    VCF_INFO);
26383 +       __VCMD(get_badness,      3, VCA_VXI,    VCF_INFO);
26384 +       __VCMD(get_vhi_name,     3, VCA_VXI,    VCF_INFO);
26385 +       __VCMD(get_rlimit,       3, VCA_VXI,    VCF_INFO);
26386 +
26387 +       __VCMD(ctx_stat,         3, VCA_VXI,    VCF_INFO);
26388 +       __VCMD(virt_stat,        3, VCA_VXI,    VCF_INFO);
26389 +       __VCMD(sock_stat,        3, VCA_VXI,    VCF_INFO);
26390 +       __VCMD(rlimit_stat,      3, VCA_VXI,    VCF_INFO);
26391 +
26392 +       __VCMD(task_nid,         2, VCA_NONE,   0);
26393 +       __VCMD(nx_info,          3, VCA_NXI,    VCF_INFO);
26394 +       __VCMD(get_ncaps,        3, VCA_NXI,    VCF_INFO);
26395 +       __VCMD(get_nflags,       3, VCA_NXI,    VCF_INFO);
26396 +
26397 +       __VCMD(task_tag,         2, VCA_NONE,   0);
26398 +
26399 +       __VCMD(get_iattr,        2, VCA_NONE,   0);
26400 +       __VCMD(fget_iattr,       2, VCA_NONE,   0);
26401 +       __VCMD(get_dlimit,       3, VCA_NONE,   VCF_INFO);
26402 +       __VCMD(get_sched,        3, VCA_VXI,    VCF_INFO);
26403 +       __VCMD(sched_info,       3, VCA_VXI,    VCF_INFO | VCF_ZIDOK);
26404 +
26405 +       /* lower admin commands */
26406 +       __VCMD(wait_exit,        4, VCA_VXI,    VCF_INFO);
26407 +       __VCMD(ctx_create_v0,    5, VCA_NONE,   0);
26408 +       __VCMD(ctx_create,       5, VCA_NONE,   0);
26409 +       __VCMD(ctx_migrate_v0,   5, VCA_VXI,    VCF_ADMIN);
26410 +       __VCMD(ctx_migrate,      5, VCA_VXI,    VCF_ADMIN);
26411 +       __VCMD(enter_space_v0,   5, VCA_VXI,    VCF_ADMIN);
26412 +       __VCMD(enter_space,      5, VCA_VXI,    VCF_ADMIN);
26413 +
26414 +       __VCMD(net_create_v0,    5, VCA_NONE,   0);
26415 +       __VCMD(net_create,       5, VCA_NONE,   0);
26416 +       __VCMD(net_migrate,      5, VCA_NXI,    VCF_ADMIN);
26417 +
26418 +       __VCMD(tag_migrate,      5, VCA_NONE,   VCF_ADMIN);
26419 +
26420 +       /* higher admin commands */
26421 +       __VCMD(ctx_kill,         6, VCA_VXI,    VCF_ARES);
26422 +       __VCMD(set_space,        7, VCA_VXI,    VCF_ARES | VCF_SETUP);
26423 +
26424 +       __VCMD(set_ccaps,        7, VCA_VXI,    VCF_ARES | VCF_SETUP);
26425 +       __VCMD(set_bcaps,        7, VCA_VXI,    VCF_ARES | VCF_SETUP);
26426 +       __VCMD(set_cflags,       7, VCA_VXI,    VCF_ARES | VCF_SETUP);
26427 +       __VCMD(set_badness,      7, VCA_VXI,    VCF_ARES | VCF_SETUP);
26428 +
26429 +       __VCMD(set_vhi_name,     7, VCA_VXI,    VCF_ARES | VCF_SETUP);
26430 +       __VCMD(set_rlimit,       7, VCA_VXI,    VCF_ARES | VCF_SETUP);
26431 +       __VCMD(set_sched,        7, VCA_VXI,    VCF_ARES | VCF_SETUP);
26432 +       __VCMD(set_sched_v4,     7, VCA_VXI,    VCF_ARES | VCF_SETUP);
26433 +
26434 +       __VCMD(set_ncaps,        7, VCA_NXI,    VCF_ARES | VCF_SETUP);
26435 +       __VCMD(set_nflags,       7, VCA_NXI,    VCF_ARES | VCF_SETUP);
26436 +       __VCMD(net_add,          8, VCA_NXI,    VCF_ARES | VCF_SETUP);
26437 +       __VCMD(net_remove,       8, VCA_NXI,    VCF_ARES | VCF_SETUP);
26438 +       __VCMD(net_add_ipv4,     8, VCA_NXI,    VCF_ARES | VCF_SETUP);
26439 +       __VCMD(net_remove_ipv4,  8, VCA_NXI,    VCF_ARES | VCF_SETUP);
26440 +#ifdef CONFIG_IPV6
26441 +       __VCMD(net_add_ipv6,     8, VCA_NXI,    VCF_ARES | VCF_SETUP);
26442 +       __VCMD(net_remove_ipv6,  8, VCA_NXI,    VCF_ARES | VCF_SETUP);
26443 +#endif
26444 +       __VCMD(set_iattr,        7, VCA_NONE,   0);
26445 +       __VCMD(fset_iattr,       7, VCA_NONE,   0);
26446 +       __VCMD(set_dlimit,       7, VCA_NONE,   VCF_ARES);
26447 +       __VCMD(add_dlimit,       8, VCA_NONE,   VCF_ARES);
26448 +       __VCMD(rem_dlimit,       8, VCA_NONE,   VCF_ARES);
26449 +
26450 +#ifdef CONFIG_VSERVER_DEVICE
26451 +       __VCMD(set_mapping,      8, VCA_VXI,    VCF_ARES|VCF_ZIDOK);
26452 +       __VCMD(unset_mapping,    8, VCA_VXI,    VCF_ARES|VCF_ZIDOK);
26453 +#endif
26454 +       /* debug level admin commands */
26455 +#ifdef CONFIG_VSERVER_HISTORY
26456 +       __VCMD(dump_history,     9, VCA_NONE,   0);
26457 +       __VCMD(read_history,     9, VCA_NONE,   0);
26458 +#endif
26459 +#ifdef CONFIG_VSERVER_MONITOR
26460 +       __VCMD(read_monitor,     9, VCA_NONE,   0);
26461 +#endif
26462 +
26463 +       default:
26464 +               perm = -1;
26465 +       }
26466 +
26467 +       vxdprintk(VXD_CBIT(switch, 0),
26468 +               "vc: VCMD_%02d_%d[%d], %d,%p [%d,%d,%x,%x]",
26469 +               VC_CATEGORY(cmd), VC_COMMAND(cmd),
26470 +               VC_VERSION(cmd), id, data, compat,
26471 +               perm, args, flags);
26472 +
26473 +       ret = -ENOSYS;
26474 +       if (perm < 0)
26475 +               goto out;
26476 +
26477 +       state = 1;
26478 +       if (!capable(CAP_CONTEXT))
26479 +               goto out;
26480 +
26481 +       state = 2;
26482 +       /* moved here from the individual commands */
26483 +       ret = -EPERM;
26484 +       if ((perm > 1) && !capable(CAP_SYS_ADMIN))
26485 +               goto out;
26486 +
26487 +       state = 3;
26488 +       /* vcmd involves resource management  */
26489 +       ret = -EPERM;
26490 +       if ((flags & VCF_ARES) && !capable(CAP_SYS_RESOURCE))
26491 +               goto out;
26492 +
26493 +       state = 4;
26494 +       /* various legacy exceptions */
26495 +       switch (cmd) {
26496 +       /* will go away when spectator is a cap */
26497 +       case VCMD_ctx_migrate_v0:
26498 +       case VCMD_ctx_migrate:
26499 +               if (id == 1) {
26500 +                       current->xid = 1;
26501 +                       ret = 1;
26502 +                       goto out;
26503 +               }
26504 +               break;
26505 +
26506 +       /* will go away when spectator is a cap */
26507 +       case VCMD_net_migrate:
26508 +               if (id == 1) {
26509 +                       current->nid = 1;
26510 +                       ret = 1;
26511 +                       goto out;
26512 +               }
26513 +               break;
26514 +       }
26515 +
26516 +       /* vcmds are fine by default */
26517 +       permit = 1;
26518 +
26519 +       /* admin type vcmds require admin ... */
26520 +       if (flags & VCF_ADMIN)
26521 +               permit = vx_check(0, VS_ADMIN) ? 1 : 0;
26522 +
26523 +       /* ... but setup type vcmds override that */
26524 +       if (!permit && (flags & VCF_SETUP))
26525 +               permit = vx_flags(VXF_STATE_SETUP, 0) ? 2 : 0;
26526 +
26527 +       state = 5;
26528 +       ret = -EPERM;
26529 +       if (!permit)
26530 +               goto out;
26531 +
26532 +       state = 6;
26533 +       if (!id && (flags & VCF_ZIDOK))
26534 +               goto skip_id;
26535 +
26536 +       ret = -ESRCH;
26537 +       if (args & VCA_VXI) {
26538 +               vxi = lookup_vx_info(id);
26539 +               if (!vxi)
26540 +                       goto out;
26541 +
26542 +               if ((flags & VCF_ADMIN) &&
26543 +                       /* special case kill for shutdown */
26544 +                       (cmd != VCMD_ctx_kill) &&
26545 +                       /* can context be administrated? */
26546 +                       !vx_info_flags(vxi, VXF_STATE_ADMIN, 0)) {
26547 +                       ret = -EACCES;
26548 +                       goto out_vxi;
26549 +               }
26550 +       }
26551 +       state = 7;
26552 +       if (args & VCA_NXI) {
26553 +               nxi = lookup_nx_info(id);
26554 +               if (!nxi)
26555 +                       goto out_vxi;
26556 +
26557 +               if ((flags & VCF_ADMIN) &&
26558 +                       /* can context be administrated? */
26559 +                       !nx_info_flags(nxi, NXF_STATE_ADMIN, 0)) {
26560 +                       ret = -EACCES;
26561 +                       goto out_nxi;
26562 +               }
26563 +       }
26564 +skip_id:
26565 +       state = 8;
26566 +       ret = do_vcmd(cmd, id, vxi, nxi, data, compat);
26567 +
26568 +out_nxi:
26569 +       if ((args & VCA_NXI) && nxi)
26570 +               put_nx_info(nxi);
26571 +out_vxi:
26572 +       if ((args & VCA_VXI) && vxi)
26573 +               put_vx_info(vxi);
26574 +out:
26575 +       vxdprintk(VXD_CBIT(switch, 1),
26576 +               "vc: VCMD_%02d_%d[%d] = %08lx(%ld) [%d,%d]",
26577 +               VC_CATEGORY(cmd), VC_COMMAND(cmd),
26578 +               VC_VERSION(cmd), ret, ret, state, permit);
26579 +       return ret;
26580 +}
26581 +
26582 +asmlinkage long
26583 +sys_vserver(uint32_t cmd, uint32_t id, void __user *data)
26584 +{
26585 +       return do_vserver(cmd, id, data, 0);
26586 +}
26587 +
26588 +#ifdef CONFIG_COMPAT
26589 +
26590 +asmlinkage long
26591 +sys32_vserver(uint32_t cmd, uint32_t id, void __user *data)
26592 +{
26593 +       return do_vserver(cmd, id, data, 1);
26594 +}
26595 +
26596 +#endif /* CONFIG_COMPAT */
26597 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/sysctl.c linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/sysctl.c
26598 --- linux-2.6.25.11/kernel/vserver/sysctl.c     1969-12-31 19:00:00.000000000 -0500
26599 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/sysctl.c       2008-04-19 15:14:52.000000000 -0400
26600 @@ -0,0 +1,244 @@
26601 +/*
26602 + *  kernel/vserver/sysctl.c
26603 + *
26604 + *  Virtual Context Support
26605 + *
26606 + *  Copyright (C) 2004-2007  Herbert Pötzl
26607 + *
26608 + *  V0.01  basic structure
26609 + *
26610 + */
26611 +
26612 +#include <linux/module.h>
26613 +#include <linux/ctype.h>
26614 +#include <linux/sysctl.h>
26615 +#include <linux/parser.h>
26616 +#include <asm/uaccess.h>
26617 +
26618 +
26619 +enum {
26620 +       CTL_DEBUG_ERROR         = 0,
26621 +       CTL_DEBUG_SWITCH        = 1,
26622 +       CTL_DEBUG_XID,
26623 +       CTL_DEBUG_NID,
26624 +       CTL_DEBUG_TAG,
26625 +       CTL_DEBUG_NET,
26626 +       CTL_DEBUG_LIMIT,
26627 +       CTL_DEBUG_CRES,
26628 +       CTL_DEBUG_DLIM,
26629 +       CTL_DEBUG_QUOTA,
26630 +       CTL_DEBUG_CVIRT,
26631 +       CTL_DEBUG_SPACE,
26632 +       CTL_DEBUG_MISC,
26633 +};
26634 +
26635 +
26636 +unsigned int vx_debug_switch   = 0;
26637 +unsigned int vx_debug_xid      = 0;
26638 +unsigned int vx_debug_nid      = 0;
26639 +unsigned int vx_debug_tag      = 0;
26640 +unsigned int vx_debug_net      = 0;
26641 +unsigned int vx_debug_limit    = 0;
26642 +unsigned int vx_debug_cres     = 0;
26643 +unsigned int vx_debug_dlim     = 0;
26644 +unsigned int vx_debug_quota    = 0;
26645 +unsigned int vx_debug_cvirt    = 0;
26646 +unsigned int vx_debug_space    = 0;
26647 +unsigned int vx_debug_misc     = 0;
26648 +
26649 +
26650 +static struct ctl_table_header *vserver_table_header;
26651 +static ctl_table vserver_root_table[];
26652 +
26653 +
26654 +void vserver_register_sysctl(void)
26655 +{
26656 +       if (!vserver_table_header) {
26657 +               vserver_table_header = register_sysctl_table(vserver_root_table);
26658 +       }
26659 +
26660 +}
26661 +
26662 +void vserver_unregister_sysctl(void)
26663 +{
26664 +       if (vserver_table_header) {
26665 +               unregister_sysctl_table(vserver_table_header);
26666 +               vserver_table_header = NULL;
26667 +       }
26668 +}
26669 +
26670 +
26671 +static int proc_dodebug(ctl_table *table, int write,
26672 +       struct file *filp, void __user *buffer, size_t *lenp, loff_t *ppos)
26673 +{
26674 +       char            tmpbuf[20], *p, c;
26675 +       unsigned int    value;
26676 +       size_t          left, len;
26677 +
26678 +       if ((*ppos && !write) || !*lenp) {
26679 +               *lenp = 0;
26680 +               return 0;
26681 +       }
26682 +
26683 +       left = *lenp;
26684 +
26685 +       if (write) {
26686 +               if (!access_ok(VERIFY_READ, buffer, left))
26687 +                       return -EFAULT;
26688 +               p = (char *)buffer;
26689 +               while (left && __get_user(c, p) >= 0 && isspace(c))
26690 +                       left--, p++;
26691 +               if (!left)
26692 +                       goto done;
26693 +
26694 +               if (left > sizeof(tmpbuf) - 1)
26695 +                       return -EINVAL;
26696 +               if (copy_from_user(tmpbuf, p, left))
26697 +                       return -EFAULT;
26698 +               tmpbuf[left] = '\0';
26699 +
26700 +               for (p = tmpbuf, value = 0; '0' <= *p && *p <= '9'; p++, left--)
26701 +                       value = 10 * value + (*p - '0');
26702 +               if (*p && !isspace(*p))
26703 +                       return -EINVAL;
26704 +               while (left && isspace(*p))
26705 +                       left--, p++;
26706 +               *(unsigned int *)table->data = value;
26707 +       } else {
26708 +               if (!access_ok(VERIFY_WRITE, buffer, left))
26709 +                       return -EFAULT;
26710 +               len = sprintf(tmpbuf, "%d", *(unsigned int *)table->data);
26711 +               if (len > left)
26712 +                       len = left;
26713 +               if (__copy_to_user(buffer, tmpbuf, len))
26714 +                       return -EFAULT;
26715 +               if ((left -= len) > 0) {
26716 +                       if (put_user('\n', (char *)buffer + len))
26717 +                               return -EFAULT;
26718 +                       left--;
26719 +               }
26720 +       }
26721 +
26722 +done:
26723 +       *lenp -= left;
26724 +       *ppos += *lenp;
26725 +       return 0;
26726 +}
26727 +
26728 +static int zero;
26729 +
26730 +#define        CTL_ENTRY(ctl, name)                            \
26731 +       {                                               \
26732 +               .ctl_name       = ctl,                  \
26733 +               .procname       = #name,                \
26734 +               .data           = &vx_ ## name,         \
26735 +               .maxlen         = sizeof(int),          \
26736 +               .mode           = 0644,                 \
26737 +               .proc_handler   = &proc_dodebug,        \
26738 +               .strategy       = &sysctl_intvec,       \
26739 +               .extra1         = &zero,                \
26740 +       }
26741 +
26742 +static ctl_table vserver_debug_table[] = {
26743 +       CTL_ENTRY(CTL_DEBUG_SWITCH,     debug_switch),
26744 +       CTL_ENTRY(CTL_DEBUG_XID,        debug_xid),
26745 +       CTL_ENTRY(CTL_DEBUG_NID,        debug_nid),
26746 +       CTL_ENTRY(CTL_DEBUG_TAG,        debug_tag),
26747 +       CTL_ENTRY(CTL_DEBUG_NET,        debug_net),
26748 +       CTL_ENTRY(CTL_DEBUG_LIMIT,      debug_limit),
26749 +       CTL_ENTRY(CTL_DEBUG_CRES,       debug_cres),
26750 +       CTL_ENTRY(CTL_DEBUG_DLIM,       debug_dlim),
26751 +       CTL_ENTRY(CTL_DEBUG_QUOTA,      debug_quota),
26752 +       CTL_ENTRY(CTL_DEBUG_CVIRT,      debug_cvirt),
26753 +       CTL_ENTRY(CTL_DEBUG_SPACE,      debug_space),
26754 +       CTL_ENTRY(CTL_DEBUG_MISC,       debug_misc),
26755 +       { .ctl_name = 0 }
26756 +};
26757 +
26758 +static ctl_table vserver_root_table[] = {
26759 +       {
26760 +               .ctl_name       = CTL_VSERVER,
26761 +               .procname       = "vserver",
26762 +               .mode           = 0555,
26763 +               .child          = vserver_debug_table
26764 +       },
26765 +       { .ctl_name = 0 }
26766 +};
26767 +
26768 +
26769 +static match_table_t tokens = {
26770 +       { CTL_DEBUG_SWITCH,     "switch=%x"     },
26771 +       { CTL_DEBUG_XID,        "xid=%x"        },
26772 +       { CTL_DEBUG_NID,        "nid=%x"        },
26773 +       { CTL_DEBUG_TAG,        "tag=%x"        },
26774 +       { CTL_DEBUG_NET,        "net=%x"        },
26775 +       { CTL_DEBUG_LIMIT,      "limit=%x"      },
26776 +       { CTL_DEBUG_CRES,       "cres=%x"       },
26777 +       { CTL_DEBUG_DLIM,       "dlim=%x"       },
26778 +       { CTL_DEBUG_QUOTA,      "quota=%x"      },
26779 +       { CTL_DEBUG_CVIRT,      "cvirt=%x"      },
26780 +       { CTL_DEBUG_SPACE,      "space=%x"      },
26781 +       { CTL_DEBUG_MISC,       "misc=%x"       },
26782 +       { CTL_DEBUG_ERROR,      NULL            }
26783 +};
26784 +
26785 +#define        HANDLE_CASE(id, name, val)                              \
26786 +       case CTL_DEBUG_ ## id:                                  \
26787 +               vx_debug_ ## name = val;                        \
26788 +               printk("vs_debug_" #name "=0x%x\n", val);       \
26789 +               break
26790 +
26791 +
26792 +static int __init vs_debug_setup(char *str)
26793 +{
26794 +       char *p;
26795 +       int token;
26796 +
26797 +       printk("vs_debug_setup(%s)\n", str);
26798 +       while ((p = strsep(&str, ",")) != NULL) {
26799 +               substring_t args[MAX_OPT_ARGS];
26800 +               unsigned int value;
26801 +
26802 +               if (!*p)
26803 +                       continue;
26804 +
26805 +               token = match_token(p, tokens, args);
26806 +               value = (token > 0) ? simple_strtoul(args[0].from, NULL, 0) : 0;
26807 +
26808 +               switch (token) {
26809 +               HANDLE_CASE(SWITCH, switch, value);
26810 +               HANDLE_CASE(XID,    xid,    value);
26811 +               HANDLE_CASE(NID,    nid,    value);
26812 +               HANDLE_CASE(TAG,    tag,    value);
26813 +               HANDLE_CASE(NET,    net,    value);
26814 +               HANDLE_CASE(LIMIT,  limit,  value);
26815 +               HANDLE_CASE(CRES,   cres,   value);
26816 +               HANDLE_CASE(DLIM,   dlim,   value);
26817 +               HANDLE_CASE(QUOTA,  quota,  value);
26818 +               HANDLE_CASE(CVIRT,  cvirt,  value);
26819 +               HANDLE_CASE(SPACE,  space,  value);
26820 +               HANDLE_CASE(MISC,   misc,   value);
26821 +               default:
26822 +                       return -EINVAL;
26823 +                       break;
26824 +               }
26825 +       }
26826 +       return 1;
26827 +}
26828 +
26829 +__setup("vsdebug=", vs_debug_setup);
26830 +
26831 +
26832 +
26833 +EXPORT_SYMBOL_GPL(vx_debug_switch);
26834 +EXPORT_SYMBOL_GPL(vx_debug_xid);
26835 +EXPORT_SYMBOL_GPL(vx_debug_nid);
26836 +EXPORT_SYMBOL_GPL(vx_debug_net);
26837 +EXPORT_SYMBOL_GPL(vx_debug_limit);
26838 +EXPORT_SYMBOL_GPL(vx_debug_cres);
26839 +EXPORT_SYMBOL_GPL(vx_debug_dlim);
26840 +EXPORT_SYMBOL_GPL(vx_debug_quota);
26841 +EXPORT_SYMBOL_GPL(vx_debug_cvirt);
26842 +EXPORT_SYMBOL_GPL(vx_debug_space);
26843 +EXPORT_SYMBOL_GPL(vx_debug_misc);
26844 +
26845 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/tag.c linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/tag.c
26846 --- linux-2.6.25.11/kernel/vserver/tag.c        1969-12-31 19:00:00.000000000 -0500
26847 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/tag.c  2008-04-19 15:14:52.000000000 -0400
26848 @@ -0,0 +1,63 @@
26849 +/*
26850 + *  linux/kernel/vserver/tag.c
26851 + *
26852 + *  Virtual Server: Shallow Tag Space
26853 + *
26854 + *  Copyright (C) 2007  Herbert Pötzl
26855 + *
26856 + *  V0.01  basic implementation
26857 + *
26858 + */
26859 +
26860 +#include <linux/sched.h>
26861 +#include <linux/vserver/debug.h>
26862 +#include <linux/vs_pid.h>
26863 +#include <linux/vs_tag.h>
26864 +
26865 +#include <linux/vserver/tag_cmd.h>
26866 +
26867 +
26868 +int dx_migrate_task(struct task_struct *p, tag_t tag)
26869 +{
26870 +       if (!p)
26871 +               BUG();
26872 +
26873 +       vxdprintk(VXD_CBIT(tag, 5),
26874 +               "dx_migrate_task(%p[#%d],#%d)", p, p->tag, tag);
26875 +
26876 +       task_lock(p);
26877 +       p->tag = tag;
26878 +       task_unlock(p);
26879 +
26880 +       vxdprintk(VXD_CBIT(tag, 5),
26881 +               "moved task %p into [#%d]", p, tag);
26882 +       return 0;
26883 +}
26884 +
26885 +/* vserver syscall commands below here */
26886 +
26887 +/* taks xid and vx_info functions */
26888 +
26889 +
26890 +int vc_task_tag(uint32_t id)
26891 +{
26892 +       tag_t tag;
26893 +
26894 +       if (id) {
26895 +               struct task_struct *tsk;
26896 +               read_lock(&tasklist_lock);
26897 +               tsk = find_task_by_real_pid(id);
26898 +               tag = (tsk) ? tsk->tag : -ESRCH;
26899 +               read_unlock(&tasklist_lock);
26900 +       } else
26901 +               tag = dx_current_tag();
26902 +       return tag;
26903 +}
26904 +
26905 +
26906 +int vc_tag_migrate(uint32_t tag)
26907 +{
26908 +       return dx_migrate_task(current, tag & 0xFFFF);
26909 +}
26910 +
26911 +
26912 diff -NurpP --minimal linux-2.6.25.11/kernel/vserver/vci_config.h linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/vci_config.h
26913 --- linux-2.6.25.11/kernel/vserver/vci_config.h 1969-12-31 19:00:00.000000000 -0500
26914 +++ linux-2.6.25.11-vs2.3.0.34.14/kernel/vserver/vci_config.h   2008-07-15 22:48:08.000000000 -0400
26915 @@ -0,0 +1,81 @@
26916 +
26917 +/*  interface version */
26918 +
26919 +#define VCI_VERSION            0x00020303
26920 +
26921 +
26922 +enum {
26923 +       VCI_KCBIT_NO_DYNAMIC = 0,
26924 +
26925 +       VCI_KCBIT_PROC_SECURE = 4,
26926 +       VCI_KCBIT_HARDCPU = 5,
26927 +       VCI_KCBIT_IDLELIMIT = 6,
26928 +       VCI_KCBIT_IDLETIME = 7,
26929 +
26930 +       VCI_KCBIT_COWBL = 8,
26931 +       VCI_KCBIT_FULLCOWBL = 9,
26932 +       VCI_KCBIT_SPACES = 10,
26933 +       VCI_KCBIT_NETV2 = 11,
26934 +
26935 +       VCI_KCBIT_DEBUG = 16,
26936 +       VCI_KCBIT_HISTORY = 20,
26937 +       VCI_KCBIT_TAGGED = 24,
26938 +       VCI_KCBIT_PPTAG = 28,
26939 +
26940 +       VCI_KCBIT_MORE = 31,
26941 +};
26942 +
26943 +
26944 +static inline uint32_t vci_kernel_config(void)
26945 +{
26946 +       return
26947 +       (1 << VCI_KCBIT_NO_DYNAMIC) |
26948 +
26949 +       /* configured features */
26950 +#ifdef CONFIG_VSERVER_PROC_SECURE
26951 +       (1 << VCI_KCBIT_PROC_SECURE) |
26952 +#endif
26953 +#ifdef CONFIG_VSERVER_HARDCPU
26954 +       (1 << VCI_KCBIT_HARDCPU) |
26955 +#endif
26956 +#ifdef CONFIG_VSERVER_IDLELIMIT
26957 +       (1 << VCI_KCBIT_IDLELIMIT) |
26958 +#endif
26959 +#ifdef CONFIG_VSERVER_IDLETIME
26960 +       (1 << VCI_KCBIT_IDLETIME) |
26961 +#endif
26962 +#ifdef CONFIG_VSERVER_COWBL
26963 +       (1 << VCI_KCBIT_COWBL) |
26964 +       (1 << VCI_KCBIT_FULLCOWBL) |
26965 +#endif
26966 +       (1 << VCI_KCBIT_SPACES) |
26967 +       (1 << VCI_KCBIT_NETV2) |
26968 +
26969 +       /* debug options */
26970 +#ifdef CONFIG_VSERVER_DEBUG
26971 +       (1 << VCI_KCBIT_DEBUG) |
26972 +#endif
26973 +#ifdef CONFIG_VSERVER_HISTORY
26974 +       (1 << VCI_KCBIT_HISTORY) |
26975 +#endif
26976 +
26977 +       /* inode context tagging */
26978 +#if    defined(CONFIG_TAGGING_NONE)
26979 +       (0 << VCI_KCBIT_TAGGED) |
26980 +#elif  defined(CONFIG_TAGGING_UID16)
26981 +       (1 << VCI_KCBIT_TAGGED) |
26982 +#elif  defined(CONFIG_TAGGING_GID16)
26983 +       (2 << VCI_KCBIT_TAGGED) |
26984 +#elif  defined(CONFIG_TAGGING_ID24)
26985 +       (3 << VCI_KCBIT_TAGGED) |
26986 +#elif  defined(CONFIG_TAGGING_INTERN)
26987 +       (4 << VCI_KCBIT_TAGGED) |
26988 +#elif  defined(CONFIG_TAGGING_RUNTIME)
26989 +       (5 << VCI_KCBIT_TAGGED) |
26990 +#else
26991 +       (7 << VCI_KCBIT_TAGGED) |
26992 +#endif
26993 +       (1 << VCI_KCBIT_PPTAG) |
26994 +       0;
26995 +}
26996 +
26997 diff -NurpP --minimal linux-2.6.25.11/Makefile linux-2.6.25.11-vs2.3.0.34.14/Makefile
26998 --- linux-2.6.25.11/Makefile    2008-07-15 22:05:16.000000000 -0400
26999 +++ linux-2.6.25.11-vs2.3.0.34.14/Makefile      2008-07-16 00:41:15.000000000 -0400
27000 @@ -1,7 +1,7 @@
27001  VERSION = 2
27002  PATCHLEVEL = 6
27003  SUBLEVEL = 25
27004 -EXTRAVERSION = .14
27005 +EXTRAVERSION = .14-vs2.3.0.34.14
27006  NAME = Funky Weasel is Jiggy wit it
27007
27008  # *DOCUMENTATION*
27009 diff -NurpP --minimal linux-2.6.25.11/mm/filemap_xip.c linux-2.6.25.11-vs2.3.0.34.14/mm/filemap_xip.c
27010 --- linux-2.6.25.11/mm/filemap_xip.c    2008-04-17 12:05:44.000000000 -0400
27011 +++ linux-2.6.25.11-vs2.3.0.34.14/mm/filemap_xip.c      2008-04-19 15:14:52.000000000 -0400
27012 @@ -14,6 +14,7 @@
27013  #include <linux/uio.h>
27014  #include <linux/rmap.h>
27015  #include <linux/sched.h>
27016 +#include <linux/vs_memory.h>
27017  #include <asm/tlbflush.h>
27018
27019  /*
27020 diff -NurpP --minimal linux-2.6.25.11/mm/fremap.c linux-2.6.25.11-vs2.3.0.34.14/mm/fremap.c
27021 --- linux-2.6.25.11/mm/fremap.c 2008-04-17 12:05:44.000000000 -0400
27022 +++ linux-2.6.25.11-vs2.3.0.34.14/mm/fremap.c   2008-04-19 15:14:52.000000000 -0400
27023 @@ -15,6 +15,7 @@
27024  #include <linux/rmap.h>
27025  #include <linux/module.h>
27026  #include <linux/syscalls.h>
27027 +#include <linux/vs_memory.h>
27028
27029  #include <asm/mmu_context.h>
27030  #include <asm/cacheflush.h>
27031 diff -NurpP --minimal linux-2.6.25.11/mm/hugetlb.c linux-2.6.25.11-vs2.3.0.34.14/mm/hugetlb.c
27032 --- linux-2.6.25.11/mm/hugetlb.c        2008-04-17 12:05:44.000000000 -0400
27033 +++ linux-2.6.25.11-vs2.3.0.34.14/mm/hugetlb.c  2008-04-19 15:14:52.000000000 -0400
27034 @@ -19,6 +19,7 @@
27035  #include <asm/pgtable.h>
27036
27037  #include <linux/hugetlb.h>
27038 +#include <linux/vs_memory.h>
27039  #include "internal.h"
27040
27041  const unsigned long hugetlb_zero = 0, hugetlb_infinity = ~0UL;
27042 diff -NurpP --minimal linux-2.6.25.11/mm/memory.c linux-2.6.25.11-vs2.3.0.34.14/mm/memory.c
27043 --- linux-2.6.25.11/mm/memory.c 2008-07-15 22:05:18.000000000 -0400
27044 +++ linux-2.6.25.11-vs2.3.0.34.14/mm/memory.c   2008-07-11 12:41:01.000000000 -0400
27045 @@ -505,6 +505,9 @@ static int copy_pte_range(struct mm_stru
27046         int progress = 0;
27047         int rss[2];
27048
27049 +       if (!vx_rss_avail(dst_mm, ((end - addr)/PAGE_SIZE + 1)))
27050 +               return -ENOMEM;
27051 +
27052  again:
27053         rss[1] = rss[0] = 0;
27054         dst_pte = pte_alloc_map_lock(dst_mm, dst_pmd, addr, &dst_ptl);
27055 @@ -2085,6 +2088,11 @@ static int do_swap_page(struct mm_struct
27056                 goto out;
27057         }
27058
27059 +       if (!vx_rss_avail(mm, 1)) {
27060 +               ret = VM_FAULT_OOM;
27061 +               goto out;
27062 +       }
27063 +
27064         mark_page_accessed(page);
27065         lock_page(page);
27066         delayacct_clear_flag(DELAYACCT_PF_SWAPIN);
27067 @@ -2156,6 +2164,8 @@ static int do_anonymous_page(struct mm_s
27068         /* Allocate our own private page. */
27069         pte_unmap(page_table);
27070
27071 +       if (!vx_rss_avail(mm, 1))
27072 +               goto oom;
27073         if (unlikely(anon_vma_prepare(vma)))
27074                 goto oom;
27075         page = alloc_zeroed_user_highpage_movable(vma, address);
27076 @@ -2480,6 +2490,7 @@ static inline int handle_pte_fault(struc
27077  {
27078         pte_t entry;
27079         spinlock_t *ptl;
27080 +       int ret = 0, type = VXPT_UNKNOWN;
27081
27082         entry = *pte;
27083         if (!pte_present(entry)) {
27084 @@ -2507,9 +2518,12 @@ static inline int handle_pte_fault(struc
27085         if (unlikely(!pte_same(*pte, entry)))
27086                 goto unlock;
27087         if (write_access) {
27088 -               if (!pte_write(entry))
27089 -                       return do_wp_page(mm, vma, address,
27090 +               if (!pte_write(entry)) {
27091 +                       ret = do_wp_page(mm, vma, address,
27092                                         pte, pmd, ptl, entry);
27093 +                       type = VXPT_WRITE;
27094 +                       goto out;
27095 +               }
27096                 entry = pte_mkdirty(entry);
27097         }
27098         entry = pte_mkyoung(entry);
27099 @@ -2527,7 +2541,10 @@ static inline int handle_pte_fault(struc
27100         }
27101  unlock:
27102         pte_unmap_unlock(pte, ptl);
27103 -       return 0;
27104 +       ret = 0;
27105 +out:
27106 +       vx_page_fault(mm, vma, type, ret);
27107 +       return ret;
27108  }
27109
27110  /*
27111 diff -NurpP --minimal linux-2.6.25.11/mm/mlock.c linux-2.6.25.11-vs2.3.0.34.14/mm/mlock.c
27112 --- linux-2.6.25.11/mm/mlock.c  2008-04-17 10:37:25.000000000 -0400
27113 +++ linux-2.6.25.11-vs2.3.0.34.14/mm/mlock.c    2008-04-19 15:14:52.000000000 -0400
27114 @@ -12,6 +12,7 @@
27115  #include <linux/syscalls.h>
27116  #include <linux/sched.h>
27117  #include <linux/module.h>
27118 +#include <linux/vs_memory.h>
27119
27120  int can_do_mlock(void)
27121  {
27122 @@ -76,7 +77,7 @@ success:
27123                         ret = make_pages_present(start, end);
27124         }
27125
27126 -       mm->locked_vm -= pages;
27127 +       vx_vmlocked_sub(mm, pages);
27128  out:
27129         if (ret == -ENOMEM)
27130                 ret = -EAGAIN;
27131 @@ -134,7 +135,7 @@ static int do_mlock(unsigned long start,
27132
27133  asmlinkage long sys_mlock(unsigned long start, size_t len)
27134  {
27135 -       unsigned long locked;
27136 +       unsigned long locked, grow;
27137         unsigned long lock_limit;
27138         int error = -ENOMEM;
27139
27140 @@ -145,8 +146,10 @@ asmlinkage long sys_mlock(unsigned long
27141         len = PAGE_ALIGN(len + (start & ~PAGE_MASK));
27142         start &= PAGE_MASK;
27143
27144 -       locked = len >> PAGE_SHIFT;
27145 -       locked += current->mm->locked_vm;
27146 +       grow = len >> PAGE_SHIFT;
27147 +       if (!vx_vmlocked_avail(current->mm, grow))
27148 +               goto out;
27149 +       locked = current->mm->locked_vm + grow;
27150
27151         lock_limit = current->signal->rlim[RLIMIT_MEMLOCK].rlim_cur;
27152         lock_limit >>= PAGE_SHIFT;
27153 @@ -154,6 +157,7 @@ asmlinkage long sys_mlock(unsigned long
27154         /* check against resource limits */
27155         if ((locked <= lock_limit) || capable(CAP_IPC_LOCK))
27156                 error = do_mlock(start, len, 1);
27157 +out:
27158         up_write(&current->mm->mmap_sem);
27159         return error;
27160  }
27161 @@ -213,6 +217,8 @@ asmlinkage long sys_mlockall(int flags)
27162         lock_limit >>= PAGE_SHIFT;
27163
27164         ret = -ENOMEM;
27165 +       if (!vx_vmlocked_avail(current->mm, current->mm->total_vm))
27166 +               goto out;
27167         if (!(flags & MCL_CURRENT) || (current->mm->total_vm <= lock_limit) ||
27168             capable(CAP_IPC_LOCK))
27169                 ret = do_mlockall(flags);
27170 diff -NurpP --minimal linux-2.6.25.11/mm/mmap.c linux-2.6.25.11-vs2.3.0.34.14/mm/mmap.c
27171 --- linux-2.6.25.11/mm/mmap.c   2008-07-15 22:05:18.000000000 -0400
27172 +++ linux-2.6.25.11-vs2.3.0.34.14/mm/mmap.c     2008-06-10 20:00:46.000000000 -0400
27173 @@ -1203,10 +1203,10 @@ munmap_back:
27174                 kmem_cache_free(vm_area_cachep, vma);
27175         }
27176  out:
27177 -       mm->total_vm += len >> PAGE_SHIFT;
27178 +       vx_vmpages_add(mm, len >> PAGE_SHIFT);
27179         vm_stat_account(mm, vm_flags, file, len >> PAGE_SHIFT);
27180         if (vm_flags & VM_LOCKED) {
27181 -               mm->locked_vm += len >> PAGE_SHIFT;
27182 +               vx_vmlocked_add(mm, len >> PAGE_SHIFT);
27183                 make_pages_present(addr, addr + len);
27184         }
27185         if ((flags & MAP_POPULATE) && !(flags & MAP_NONBLOCK))
27186 @@ -1555,9 +1555,9 @@ static int acct_stack_growth(struct vm_a
27187                 return -ENOMEM;
27188
27189         /* Ok, everything looks good - let it rip */
27190 -       mm->total_vm += grow;
27191 +       vx_vmpages_add(mm, grow);
27192         if (vma->vm_flags & VM_LOCKED)
27193 -               mm->locked_vm += grow;
27194 +               vx_vmlocked_add(mm, grow);
27195         vm_stat_account(mm, vma->vm_flags, vma->vm_file, grow);
27196         return 0;
27197  }
27198 @@ -1728,9 +1728,9 @@ static void remove_vma_list(struct mm_st
27199         do {
27200                 long nrpages = vma_pages(vma);
27201
27202 -               mm->total_vm -= nrpages;
27203 +               vx_vmpages_sub(mm, nrpages);
27204                 if (vma->vm_flags & VM_LOCKED)
27205 -                       mm->locked_vm -= nrpages;
27206 +                       vx_vmlocked_sub(mm, nrpages);
27207                 vm_stat_account(mm, vma->vm_flags, vma->vm_file, -nrpages);
27208                 vma = remove_vma(vma);
27209         } while (vma);
27210 @@ -1973,6 +1973,8 @@ unsigned long do_brk(unsigned long addr,
27211                 lock_limit >>= PAGE_SHIFT;
27212                 if (locked > lock_limit && !capable(CAP_IPC_LOCK))
27213                         return -EAGAIN;
27214 +               if (!vx_vmlocked_avail(mm, len >> PAGE_SHIFT))
27215 +                       return -ENOMEM;
27216         }
27217
27218         /*
27219 @@ -1999,7 +2001,8 @@ unsigned long do_brk(unsigned long addr,
27220         if (mm->map_count > sysctl_max_map_count)
27221                 return -ENOMEM;
27222
27223 -       if (security_vm_enough_memory(len >> PAGE_SHIFT))
27224 +       if (security_vm_enough_memory(len >> PAGE_SHIFT) ||
27225 +               !vx_vmpages_avail(mm, len >> PAGE_SHIFT))
27226                 return -ENOMEM;
27227
27228         /* Can we just expand an old private anonymous mapping? */
27229 @@ -2024,9 +2027,9 @@ unsigned long do_brk(unsigned long addr,
27230         vma->vm_page_prot = vm_get_page_prot(flags);
27231         vma_link(mm, vma, prev, rb_link, rb_parent);
27232  out:
27233 -       mm->total_vm += len >> PAGE_SHIFT;
27234 +       vx_vmpages_add(mm, len >> PAGE_SHIFT);
27235         if (flags & VM_LOCKED) {
27236 -               mm->locked_vm += len >> PAGE_SHIFT;
27237 +               vx_vmlocked_add(mm, len >> PAGE_SHIFT);
27238                 make_pages_present(addr, addr + len);
27239         }
27240         return addr;
27241 @@ -2055,6 +2058,11 @@ void exit_mmap(struct mm_struct *mm)
27242         free_pgtables(&tlb, vma, FIRST_USER_ADDRESS, 0);
27243         tlb_finish_mmu(tlb, 0, end);
27244
27245 +       set_mm_counter(mm, file_rss, 0);
27246 +       set_mm_counter(mm, anon_rss, 0);
27247 +       vx_vmpages_sub(mm, mm->total_vm);
27248 +       vx_vmlocked_sub(mm, mm->locked_vm);
27249 +
27250         /*
27251          * Walk the list again, actually closing and freeing it,
27252          * with preemption enabled, without holding any MM locks.
27253 @@ -2094,7 +2102,8 @@ int insert_vm_struct(struct mm_struct *
27254         if (__vma && __vma->vm_start < vma->vm_end)
27255                 return -ENOMEM;
27256         if ((vma->vm_flags & VM_ACCOUNT) &&
27257 -            security_vm_enough_memory_mm(mm, vma_pages(vma)))
27258 +               (security_vm_enough_memory_mm(mm, vma_pages(vma)) ||
27259 +               !vx_vmpages_avail(mm, vma_pages(vma))))
27260                 return -ENOMEM;
27261         vma_link(mm, vma, prev, rb_link, rb_parent);
27262         return 0;
27263 @@ -2167,6 +2176,8 @@ int may_expand_vm(struct mm_struct *mm,
27264
27265         if (cur + npages > lim)
27266                 return 0;
27267 +       if (!vx_vmpages_avail(mm, npages))
27268 +               return 0;
27269         return 1;
27270  }
27271
27272 @@ -2244,7 +2255,6 @@ int install_special_mapping(struct mm_st
27273                 return -ENOMEM;
27274         }
27275
27276 -       mm->total_vm += len >> PAGE_SHIFT;
27277 -
27278 +       vx_vmpages_add(mm, len >> PAGE_SHIFT);
27279         return 0;
27280  }
27281 diff -NurpP --minimal linux-2.6.25.11/mm/mremap.c linux-2.6.25.11-vs2.3.0.34.14/mm/mremap.c
27282 --- linux-2.6.25.11/mm/mremap.c 2008-04-17 11:31:40.000000000 -0400
27283 +++ linux-2.6.25.11-vs2.3.0.34.14/mm/mremap.c   2008-04-19 15:14:52.000000000 -0400
27284 @@ -18,6 +18,7 @@
27285  #include <linux/highmem.h>
27286  #include <linux/security.h>
27287  #include <linux/syscalls.h>
27288 +#include <linux/vs_memory.h>
27289
27290  #include <asm/uaccess.h>
27291  #include <asm/cacheflush.h>
27292 @@ -212,7 +213,7 @@ static unsigned long move_vma(struct vm_
27293          * If this were a serious issue, we'd add a flag to do_munmap().
27294          */
27295         hiwater_vm = mm->hiwater_vm;
27296 -       mm->total_vm += new_len >> PAGE_SHIFT;
27297 +       vx_vmpages_add(mm, new_len >> PAGE_SHIFT);
27298         vm_stat_account(mm, vma->vm_flags, vma->vm_file, new_len>>PAGE_SHIFT);
27299
27300         if (do_munmap(mm, old_addr, old_len) < 0) {
27301 @@ -230,7 +231,7 @@ static unsigned long move_vma(struct vm_
27302         }
27303
27304         if (vm_flags & VM_LOCKED) {
27305 -               mm->locked_vm += new_len >> PAGE_SHIFT;
27306 +               vx_vmlocked_add(mm, new_len >> PAGE_SHIFT);
27307                 if (new_len > old_len)
27308                         make_pages_present(new_addr + old_len,
27309                                            new_addr + new_len);
27310 @@ -341,6 +342,9 @@ unsigned long do_mremap(unsigned long ad
27311                 ret = -EAGAIN;
27312                 if (locked > lock_limit && !capable(CAP_IPC_LOCK))
27313                         goto out;
27314 +               if (!vx_vmlocked_avail(current->mm,
27315 +                       (new_len - old_len) >> PAGE_SHIFT))
27316 +                       goto out;
27317         }
27318         if (!may_expand_vm(mm, (new_len - old_len) >> PAGE_SHIFT)) {
27319                 ret = -ENOMEM;
27320 @@ -369,10 +373,10 @@ unsigned long do_mremap(unsigned long ad
27321                         vma_adjust(vma, vma->vm_start,
27322                                 addr + new_len, vma->vm_pgoff, NULL);
27323
27324 -                       mm->total_vm += pages;
27325 +                       vx_vmpages_add(mm, pages);
27326                         vm_stat_account(mm, vma->vm_flags, vma->vm_file, pages);
27327                         if (vma->vm_flags & VM_LOCKED) {
27328 -                               mm->locked_vm += pages;
27329 +                               vx_vmlocked_add(mm, pages);
27330                                 make_pages_present(addr + old_len,
27331                                                    addr + new_len);
27332                         }
27333 diff -NurpP --minimal linux-2.6.25.11/mm/nommu.c linux-2.6.25.11-vs2.3.0.34.14/mm/nommu.c
27334 --- linux-2.6.25.11/mm/nommu.c  2008-04-17 12:05:44.000000000 -0400
27335 +++ linux-2.6.25.11-vs2.3.0.34.14/mm/nommu.c    2008-04-19 15:14:52.000000000 -0400
27336 @@ -991,7 +991,7 @@ unsigned long do_mmap_pgoff(struct file
27337         realalloc += kobjsize(vma);
27338         askedalloc += sizeof(*vma);
27339
27340 -       current->mm->total_vm += len >> PAGE_SHIFT;
27341 +       vx_vmpages_add(current->mm, len >> PAGE_SHIFT);
27342
27343         add_nommu_vma(vma);
27344
27345 @@ -1117,7 +1117,7 @@ int do_munmap(struct mm_struct *mm, unsi
27346         kfree(vml);
27347
27348         update_hiwater_vm(mm);
27349 -       mm->total_vm -= len >> PAGE_SHIFT;
27350 +       vx_vmpages_sub(mm, len >> PAGE_SHIFT);
27351
27352  #ifdef DEBUG
27353         show_process_blocks();
27354 @@ -1150,7 +1150,7 @@ void exit_mmap(struct mm_struct * mm)
27355                 printk("Exit_mmap:\n");
27356  #endif
27357
27358 -               mm->total_vm = 0;
27359 +               vx_vmpages_sub(mm, mm->total_vm);
27360
27361                 while ((tmp = mm->context.vmlist)) {
27362                         mm->context.vmlist = tmp->next;
27363 diff -NurpP --minimal linux-2.6.25.11/mm/oom_kill.c linux-2.6.25.11-vs2.3.0.34.14/mm/oom_kill.c
27364 --- linux-2.6.25.11/mm/oom_kill.c       2008-04-17 12:05:44.000000000 -0400
27365 +++ linux-2.6.25.11-vs2.3.0.34.14/mm/oom_kill.c 2008-04-19 16:55:20.000000000 -0400
27366 @@ -26,6 +26,7 @@
27367  #include <linux/module.h>
27368  #include <linux/notifier.h>
27369  #include <linux/memcontrol.h>
27370 +#include <linux/vs_memory.h>
27371
27372  int sysctl_panic_on_oom;
27373  int sysctl_oom_kill_allocating_task;
27374 @@ -73,6 +74,12 @@ unsigned long badness(struct task_struct
27375         points = mm->total_vm;
27376
27377         /*
27378 +        * add points for context badness
27379 +        */
27380 +
27381 +       points += vx_badness(p, mm);
27382 +
27383 +       /*
27384          * After this unlock we can no longer dereference local variable `mm'
27385          */
27386         task_unlock(p);
27387 @@ -162,8 +169,8 @@ unsigned long badness(struct task_struct
27388         }
27389
27390  #ifdef DEBUG
27391 -       printk(KERN_DEBUG "OOMkill: task %d (%s) got %lu points\n",
27392 -       p->pid, p->comm, points);
27393 +       printk(KERN_DEBUG "OOMkill: task %d:#%u (%s) got %d points\n",
27394 +               task_pid_nr(p), p->xid, p->comm, points);
27395  #endif
27396         return points;
27397  }
27398 @@ -322,8 +329,8 @@ static void __oom_kill_task(struct task_
27399         }
27400
27401         if (verbose)
27402 -               printk(KERN_ERR "Killed process %d (%s)\n",
27403 -                               task_pid_nr(p), p->comm);
27404 +               printk(KERN_ERR "Killed process %d:#%u (%s)\n",
27405 +                               task_pid_nr(p), p->xid, p->comm);
27406
27407         /*
27408          * We give our sacrificial lamb high priority and access to
27409 @@ -403,8 +410,8 @@ static int oom_kill_process(struct task_
27410                 return 0;
27411         }
27412
27413 -       printk(KERN_ERR "%s: kill process %d (%s) score %li or a child\n",
27414 -                                       message, task_pid_nr(p), p->comm, points);
27415 +       printk(KERN_ERR "%s: kill process %d:#%u (%s) score %li or a child\n",
27416 +                               message, task_pid_nr(p), p->xid, p->comm, points);
27417
27418         /* Try to kill a child first */
27419         list_for_each_entry(c, &p->children, sibling) {
27420 diff -NurpP --minimal linux-2.6.25.11/mm/page_alloc.c linux-2.6.25.11-vs2.3.0.34.14/mm/page_alloc.c
27421 --- linux-2.6.25.11/mm/page_alloc.c     2008-07-15 22:05:18.000000000 -0400
27422 +++ linux-2.6.25.11-vs2.3.0.34.14/mm/page_alloc.c       2008-06-10 20:00:46.000000000 -0400
27423 @@ -45,6 +45,8 @@
27424  #include <linux/fault-inject.h>
27425  #include <linux/page-isolation.h>
27426  #include <linux/memcontrol.h>
27427 +#include <linux/vs_base.h>
27428 +#include <linux/vs_limit.h>
27429
27430  #include <asm/tlbflush.h>
27431  #include <asm/div64.h>
27432 @@ -1762,6 +1764,9 @@ void si_meminfo(struct sysinfo *val)
27433         val->totalhigh = totalhigh_pages;
27434         val->freehigh = nr_free_highpages();
27435         val->mem_unit = PAGE_SIZE;
27436 +
27437 +       if (vx_flags(VXF_VIRT_MEM, 0))
27438 +               vx_vsi_meminfo(val);
27439  }
27440
27441  EXPORT_SYMBOL(si_meminfo);
27442 @@ -1782,6 +1787,9 @@ void si_meminfo_node(struct sysinfo *val
27443         val->freehigh = 0;
27444  #endif
27445         val->mem_unit = PAGE_SIZE;
27446 +
27447 +       if (vx_flags(VXF_VIRT_MEM, 0))
27448 +               vx_vsi_meminfo(val);
27449  }
27450  #endif
27451
27452 diff -NurpP --minimal linux-2.6.25.11/mm/rmap.c linux-2.6.25.11-vs2.3.0.34.14/mm/rmap.c
27453 --- linux-2.6.25.11/mm/rmap.c   2008-04-17 12:05:44.000000000 -0400
27454 +++ linux-2.6.25.11-vs2.3.0.34.14/mm/rmap.c     2008-04-19 16:53:36.000000000 -0400
27455 @@ -49,6 +49,7 @@
27456  #include <linux/module.h>
27457  #include <linux/kallsyms.h>
27458  #include <linux/memcontrol.h>
27459 +#include <linux/vs_memory.h>
27460
27461  #include <asm/tlbflush.h>
27462
27463 diff -NurpP --minimal linux-2.6.25.11/mm/shmem.c linux-2.6.25.11-vs2.3.0.34.14/mm/shmem.c
27464 --- linux-2.6.25.11/mm/shmem.c  2008-04-17 12:05:44.000000000 -0400
27465 +++ linux-2.6.25.11-vs2.3.0.34.14/mm/shmem.c    2008-04-19 15:14:52.000000000 -0400
27466 @@ -56,7 +56,6 @@
27467  #include <asm/pgtable.h>
27468
27469  /* This magic number is used in glibc for posix shared memory */
27470 -#define TMPFS_MAGIC    0x01021994
27471
27472  #define ENTRIES_PER_PAGE (PAGE_CACHE_SIZE/sizeof(unsigned long))
27473  #define ENTRIES_PER_PAGEPAGE (ENTRIES_PER_PAGE*ENTRIES_PER_PAGE)
27474 @@ -1773,7 +1772,7 @@ static int shmem_statfs(struct dentry *d
27475  {
27476         struct shmem_sb_info *sbinfo = SHMEM_SB(dentry->d_sb);
27477
27478 -       buf->f_type = TMPFS_MAGIC;
27479 +       buf->f_type = TMPFS_SUPER_MAGIC;
27480         buf->f_bsize = PAGE_CACHE_SIZE;
27481         buf->f_namelen = NAME_MAX;
27482         spin_lock(&sbinfo->stat_lock);
27483 @@ -2341,7 +2340,7 @@ static int shmem_fill_super(struct super
27484         sb->s_maxbytes = SHMEM_MAX_BYTES;
27485         sb->s_blocksize = PAGE_CACHE_SIZE;
27486         sb->s_blocksize_bits = PAGE_CACHE_SHIFT;
27487 -       sb->s_magic = TMPFS_MAGIC;
27488 +       sb->s_magic = TMPFS_SUPER_MAGIC;
27489         sb->s_op = &shmem_ops;
27490         sb->s_time_gran = 1;
27491  #ifdef CONFIG_TMPFS_POSIX_ACL
27492 diff -NurpP --minimal linux-2.6.25.11/mm/slab.c linux-2.6.25.11-vs2.3.0.34.14/mm/slab.c
27493 --- linux-2.6.25.11/mm/slab.c   2008-04-17 12:05:44.000000000 -0400
27494 +++ linux-2.6.25.11-vs2.3.0.34.14/mm/slab.c     2008-04-19 15:14:52.000000000 -0400
27495 @@ -509,6 +509,8 @@ struct kmem_cache {
27496  #define STATS_INC_FREEMISS(x)  do { } while (0)
27497  #endif
27498
27499 +#include "slab_vs.h"
27500 +
27501  #if DEBUG
27502
27503  /*
27504 @@ -3344,6 +3346,7 @@ retry:
27505
27506         obj = slab_get_obj(cachep, slabp, nodeid);
27507         check_slabp(cachep, slabp);
27508 +       vx_slab_alloc(cachep, flags);
27509         l3->free_objects--;
27510         /* move slabp to correct slabp list: */
27511         list_del(&slabp->list);
27512 @@ -3416,6 +3419,7 @@ __cache_alloc_node(struct kmem_cache *ca
27513         /* ___cache_alloc_node can fall back to other nodes */
27514         ptr = ____cache_alloc_node(cachep, flags, nodeid);
27515    out:
27516 +       vx_slab_alloc(cachep, flags);
27517         local_irq_restore(save_flags);
27518         ptr = cache_alloc_debugcheck_after(cachep, flags, ptr, caller);
27519
27520 @@ -3587,6 +3591,7 @@ static inline void __cache_free(struct k
27521
27522         check_irq_off();
27523         objp = cache_free_debugcheck(cachep, objp, __builtin_return_address(0));
27524 +       vx_slab_free(cachep);
27525
27526         /*
27527          * Skip calling cache_free_alien() when the platform is not numa.
27528 diff -NurpP --minimal linux-2.6.25.11/mm/slab_vs.h linux-2.6.25.11-vs2.3.0.34.14/mm/slab_vs.h
27529 --- linux-2.6.25.11/mm/slab_vs.h        1969-12-31 19:00:00.000000000 -0500
27530 +++ linux-2.6.25.11-vs2.3.0.34.14/mm/slab_vs.h  2008-04-19 15:14:52.000000000 -0400
27531 @@ -0,0 +1,27 @@
27532 +
27533 +#include <linux/vserver/context.h>
27534 +
27535 +#include <linux/vs_context.h>
27536 +
27537 +static inline
27538 +void vx_slab_alloc(struct kmem_cache *cachep, gfp_t flags)
27539 +{
27540 +       int what = gfp_zone(cachep->gfpflags);
27541 +
27542 +       if (!current->vx_info)
27543 +               return;
27544 +
27545 +       atomic_add(cachep->buffer_size, &current->vx_info->cacct.slab[what]);
27546 +}
27547 +
27548 +static inline
27549 +void vx_slab_free(struct kmem_cache *cachep)
27550 +{
27551 +       int what = gfp_zone(cachep->gfpflags);
27552 +
27553 +       if (!current->vx_info)
27554 +               return;
27555 +
27556 +       atomic_sub(cachep->buffer_size, &current->vx_info->cacct.slab[what]);
27557 +}
27558 +
27559 diff -NurpP --minimal linux-2.6.25.11/mm/swapfile.c linux-2.6.25.11-vs2.3.0.34.14/mm/swapfile.c
27560 --- linux-2.6.25.11/mm/swapfile.c       2008-04-17 12:05:44.000000000 -0400
27561 +++ linux-2.6.25.11-vs2.3.0.34.14/mm/swapfile.c 2008-04-19 15:14:52.000000000 -0400
27562 @@ -32,6 +32,8 @@
27563  #include <asm/pgtable.h>
27564  #include <asm/tlbflush.h>
27565  #include <linux/swapops.h>
27566 +#include <linux/vs_base.h>
27567 +#include <linux/vs_memory.h>
27568
27569  DEFINE_SPINLOCK(swap_lock);
27570  unsigned int nr_swapfiles;
27571 @@ -1743,6 +1745,8 @@ void si_swapinfo(struct sysinfo *val)
27572         val->freeswap = nr_swap_pages + nr_to_be_unused;
27573         val->totalswap = total_swap_pages + nr_to_be_unused;
27574         spin_unlock(&swap_lock);
27575 +       if (vx_flags(VXF_VIRT_MEM, 0))
27576 +               vx_vsi_swapinfo(val);
27577  }
27578
27579  /*
27580 diff -NurpP --minimal linux-2.6.25.11/net/core/dev.c linux-2.6.25.11-vs2.3.0.34.14/net/core/dev.c
27581 --- linux-2.6.25.11/net/core/dev.c      2008-07-15 22:05:18.000000000 -0400
27582 +++ linux-2.6.25.11-vs2.3.0.34.14/net/core/dev.c        2008-06-23 15:22:21.000000000 -0400
27583 @@ -119,6 +119,7 @@
27584  #include <linux/err.h>
27585  #include <linux/ctype.h>
27586  #include <linux/if_arp.h>
27587 +#include <linux/vs_inet.h>
27588
27589  #include "net-sysfs.h"
27590
27591 @@ -2336,6 +2337,8 @@ static int dev_ifconf(struct net *net, c
27592
27593         total = 0;
27594         for_each_netdev(net, dev) {
27595 +               if (!nx_dev_visible(current->nx_info, dev))
27596 +                       continue;
27597                 for (i = 0; i < NPROTO; i++) {
27598                         if (gifconf_list[i]) {
27599                                 int done;
27600 @@ -2404,6 +2407,9 @@ static void dev_seq_printf_stats(struct
27601  {
27602         struct net_device_stats *stats = dev->get_stats(dev);
27603
27604 +       if (!nx_dev_visible(current->nx_info, dev))
27605 +               return;
27606 +
27607         seq_printf(seq, "%6s:%8lu %7lu %4lu %4lu %4lu %5lu %10lu %9lu "
27608                    "%8lu %7lu %4lu %4lu %4lu %5lu %7lu %10lu\n",
27609                    dev->name, stats->rx_bytes, stats->rx_packets,
27610 diff -NurpP --minimal linux-2.6.25.11/net/core/rtnetlink.c linux-2.6.25.11-vs2.3.0.34.14/net/core/rtnetlink.c
27611 --- linux-2.6.25.11/net/core/rtnetlink.c        2008-07-15 22:05:18.000000000 -0400
27612 +++ linux-2.6.25.11-vs2.3.0.34.14/net/core/rtnetlink.c  2008-05-21 14:30:41.000000000 -0400
27613 @@ -674,6 +674,8 @@ static int rtnl_dump_ifinfo(struct sk_bu
27614
27615         idx = 0;
27616         for_each_netdev(net, dev) {
27617 +               if (!nx_dev_visible(skb->sk->sk_nx_info, dev))
27618 +                       continue;
27619                 if (idx < s_idx)
27620                         goto cont;
27621                 if (rtnl_fill_ifinfo(skb, dev, RTM_NEWLINK,
27622 @@ -1207,6 +1209,9 @@ void rtmsg_ifinfo(int type, struct net_d
27623         struct sk_buff *skb;
27624         int err = -ENOBUFS;
27625
27626 +       if (!nx_dev_visible(current->nx_info, dev))
27627 +               return;
27628 +
27629         skb = nlmsg_new(if_nlmsg_size(dev), GFP_KERNEL);
27630         if (skb == NULL)
27631                 goto errout;
27632 diff -NurpP --minimal linux-2.6.25.11/net/core/sock.c linux-2.6.25.11-vs2.3.0.34.14/net/core/sock.c
27633 --- linux-2.6.25.11/net/core/sock.c     2008-04-17 12:05:44.000000000 -0400
27634 +++ linux-2.6.25.11-vs2.3.0.34.14/net/core/sock.c       2008-04-23 14:31:31.000000000 -0400
27635 @@ -126,6 +126,10 @@
27636  #include <linux/ipsec.h>
27637
27638  #include <linux/filter.h>
27639 +#include <linux/vs_socket.h>
27640 +#include <linux/vs_limit.h>
27641 +#include <linux/vs_context.h>
27642 +#include <linux/vs_network.h>
27643
27644  #ifdef CONFIG_INET
27645  #include <net/tcp.h>
27646 @@ -907,6 +911,8 @@ static struct sock *sk_prot_alloc(struct
27647                 if (!try_module_get(prot->owner))
27648                         goto out_free_sec;
27649         }
27650 +               sock_vx_init(sk);
27651 +               sock_nx_init(sk);
27652
27653         return sk;
27654
27655 @@ -984,6 +990,11 @@ void sk_free(struct sock *sk)
27656                        __FUNCTION__, atomic_read(&sk->sk_omem_alloc));
27657
27658         put_net(sk->sk_net);
27659 +       vx_sock_dec(sk);
27660 +       clr_vx_info(&sk->sk_vx_info);
27661 +       sk->sk_xid = -1;
27662 +       clr_nx_info(&sk->sk_nx_info);
27663 +       sk->sk_nid = -1;
27664         sk_prot_free(sk->sk_prot_creator, sk);
27665  }
27666
27667 @@ -999,6 +1010,8 @@ struct sock *sk_clone(const struct sock
27668
27669                 /* SANITY */
27670                 get_net(newsk->sk_net);
27671 +               sock_vx_init(newsk);
27672 +               sock_nx_init(newsk);
27673                 sk_node_init(&newsk->sk_node);
27674                 sock_lock_init(newsk);
27675                 bh_lock_sock(newsk);
27676 @@ -1045,6 +1058,12 @@ struct sock *sk_clone(const struct sock
27677                 newsk->sk_priority = 0;
27678                 atomic_set(&newsk->sk_refcnt, 2);
27679
27680 +               set_vx_info(&newsk->sk_vx_info, sk->sk_vx_info);
27681 +               newsk->sk_xid = sk->sk_xid;
27682 +               vx_sock_inc(newsk);
27683 +               set_nx_info(&newsk->sk_nx_info, sk->sk_nx_info);
27684 +               newsk->sk_nid = sk->sk_nid;
27685 +
27686                 /*
27687                  * Increment the counter in the same struct proto as the master
27688                  * sock (sk_refcnt_debug_inc uses newsk->sk_prot->socks, that
27689 @@ -1727,6 +1746,11 @@ void sock_init_data(struct socket *sock,
27690
27691         sk->sk_stamp = ktime_set(-1L, 0);
27692
27693 +       set_vx_info(&sk->sk_vx_info, current->vx_info);
27694 +       sk->sk_xid = vx_current_xid();
27695 +       vx_sock_inc(sk);
27696 +       set_nx_info(&sk->sk_nx_info, current->nx_info);
27697 +       sk->sk_nid = nx_current_nid();
27698         atomic_set(&sk->sk_refcnt, 1);
27699         atomic_set(&sk->sk_drops, 0);
27700  }
27701 diff -NurpP --minimal linux-2.6.25.11/net/ipv4/af_inet.c linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/af_inet.c
27702 --- linux-2.6.25.11/net/ipv4/af_inet.c  2008-04-17 12:05:44.000000000 -0400
27703 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/af_inet.c    2008-04-19 16:08:42.000000000 -0400
27704 @@ -115,6 +115,7 @@
27705  #ifdef CONFIG_IP_MROUTE
27706  #include <linux/mroute.h>
27707  #endif
27708 +#include <linux/vs_limit.h>
27709
27710  DEFINE_SNMP_STAT(struct linux_mib, net_statistics) __read_mostly;
27711
27712 @@ -317,9 +318,12 @@ lookup_protocol:
27713         }
27714
27715         err = -EPERM;
27716 +       if ((protocol == IPPROTO_ICMP) &&
27717 +               nx_capable(answer->capability, NXC_RAW_ICMP))
27718 +               goto override;
27719         if (answer->capability > 0 && !capable(answer->capability))
27720                 goto out_rcu_unlock;
27721 -
27722 +override:
27723         sock->ops = answer->ops;
27724         answer_prot = answer->prot;
27725         answer_no_check = answer->no_check;
27726 @@ -433,6 +437,7 @@ int inet_bind(struct socket *sock, struc
27727         struct sockaddr_in *addr = (struct sockaddr_in *)uaddr;
27728         struct sock *sk = sock->sk;
27729         struct inet_sock *inet = inet_sk(sk);
27730 +       struct nx_v4_sock_addr nsa;
27731         unsigned short snum;
27732         int chk_addr_ret;
27733         int err;
27734 @@ -446,7 +451,11 @@ int inet_bind(struct socket *sock, struc
27735         if (addr_len < sizeof(struct sockaddr_in))
27736                 goto out;
27737
27738 -       chk_addr_ret = inet_addr_type(&init_net, addr->sin_addr.s_addr);
27739 +       err = v4_map_sock_addr(inet, addr, &nsa);
27740 +       if (err)
27741 +               goto out;
27742 +
27743 +       chk_addr_ret = inet_addr_type(&init_net, nsa.saddr);
27744
27745         /* Not specified by any standard per-se, however it breaks too
27746          * many applications when removed.  It is unfortunate since
27747 @@ -458,7 +467,7 @@ int inet_bind(struct socket *sock, struc
27748         err = -EADDRNOTAVAIL;
27749         if (!sysctl_ip_nonlocal_bind &&
27750             !inet->freebind &&
27751 -           addr->sin_addr.s_addr != htonl(INADDR_ANY) &&
27752 +           nsa.saddr != htonl(INADDR_ANY) &&
27753             chk_addr_ret != RTN_LOCAL &&
27754             chk_addr_ret != RTN_MULTICAST &&
27755             chk_addr_ret != RTN_BROADCAST)
27756 @@ -483,7 +492,7 @@ int inet_bind(struct socket *sock, struc
27757         if (sk->sk_state != TCP_CLOSE || inet->num)
27758                 goto out_release_sock;
27759
27760 -       inet->rcv_saddr = inet->saddr = addr->sin_addr.s_addr;
27761 +       v4_set_sock_addr(inet, &nsa);
27762         if (chk_addr_ret == RTN_MULTICAST || chk_addr_ret == RTN_BROADCAST)
27763                 inet->saddr = 0;  /* Use device */
27764
27765 @@ -676,11 +685,13 @@ int inet_getname(struct socket *sock, st
27766                      peer == 1))
27767                         return -ENOTCONN;
27768                 sin->sin_port = inet->dport;
27769 -               sin->sin_addr.s_addr = inet->daddr;
27770 +               sin->sin_addr.s_addr =
27771 +                       nx_map_sock_lback(sk->sk_nx_info, inet->daddr);
27772         } else {
27773                 __be32 addr = inet->rcv_saddr;
27774                 if (!addr)
27775                         addr = inet->saddr;
27776 +               addr = nx_map_sock_lback(sk->sk_nx_info, addr);
27777                 sin->sin_port = inet->sport;
27778                 sin->sin_addr.s_addr = addr;
27779         }
27780 diff -NurpP --minimal linux-2.6.25.11/net/ipv4/devinet.c linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/devinet.c
27781 --- linux-2.6.25.11/net/ipv4/devinet.c  2008-04-17 12:05:44.000000000 -0400
27782 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/devinet.c    2008-04-19 15:14:52.000000000 -0400
27783 @@ -421,6 +421,7 @@ struct in_device *inetdev_by_index(struc
27784         return in_dev;
27785  }
27786
27787 +
27788  /* Called only from RTNL semaphored context. No locks. */
27789
27790  struct in_ifaddr *inet_ifa_byprefix(struct in_device *in_dev, __be32 prefix,
27791 @@ -672,6 +673,8 @@ int devinet_ioctl(unsigned int cmd, void
27792                 *colon = ':';
27793
27794         if ((in_dev = __in_dev_get_rtnl(dev)) != NULL) {
27795 +               struct nx_info *nxi = current->nx_info;
27796 +
27797                 if (tryaddrmatch) {
27798                         /* Matthias Andree */
27799                         /* compare label and address (4.4BSD style) */
27800 @@ -680,6 +683,8 @@ int devinet_ioctl(unsigned int cmd, void
27801                            This is checked above. */
27802                         for (ifap = &in_dev->ifa_list; (ifa = *ifap) != NULL;
27803                              ifap = &ifa->ifa_next) {
27804 +                               if (!nx_v4_ifa_visible(nxi, ifa))
27805 +                                       continue;
27806                                 if (!strcmp(ifr.ifr_name, ifa->ifa_label) &&
27807                                     sin_orig.sin_addr.s_addr ==
27808                                                         ifa->ifa_address) {
27809 @@ -692,9 +697,12 @@ int devinet_ioctl(unsigned int cmd, void
27810                    comparing just the label */
27811                 if (!ifa) {
27812                         for (ifap = &in_dev->ifa_list; (ifa = *ifap) != NULL;
27813 -                            ifap = &ifa->ifa_next)
27814 +                            ifap = &ifa->ifa_next) {
27815 +                               if (!nx_v4_ifa_visible(nxi, ifa))
27816 +                                       continue;
27817                                 if (!strcmp(ifr.ifr_name, ifa->ifa_label))
27818                                         break;
27819 +                       }
27820                 }
27821         }
27822
27823 @@ -846,6 +854,8 @@ static int inet_gifconf(struct net_devic
27824                 goto out;
27825
27826         for (; ifa; ifa = ifa->ifa_next) {
27827 +               if (!nx_v4_ifa_visible(current->nx_info, ifa))
27828 +                       continue;
27829                 if (!buf) {
27830                         done += sizeof(ifr);
27831                         continue;
27832 @@ -1171,6 +1181,7 @@ static int inet_dump_ifaddr(struct sk_bu
27833         struct net_device *dev;
27834         struct in_device *in_dev;
27835         struct in_ifaddr *ifa;
27836 +       struct sock *sk = skb->sk;
27837         int s_ip_idx, s_idx = cb->args[0];
27838
27839         if (net != &init_net)
27840 @@ -1188,6 +1199,8 @@ static int inet_dump_ifaddr(struct sk_bu
27841
27842                 for (ifa = in_dev->ifa_list, ip_idx = 0; ifa;
27843                      ifa = ifa->ifa_next, ip_idx++) {
27844 +                       if (sk && !nx_v4_ifa_visible(sk->sk_nx_info, ifa))
27845 +                               continue;
27846                         if (ip_idx < s_ip_idx)
27847                                 continue;
27848                         if (inet_fill_ifaddr(skb, ifa, NETLINK_CB(cb->skb).pid,
27849 diff -NurpP --minimal linux-2.6.25.11/net/ipv4/fib_hash.c linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/fib_hash.c
27850 --- linux-2.6.25.11/net/ipv4/fib_hash.c 2008-04-17 12:05:44.000000000 -0400
27851 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/fib_hash.c   2008-04-19 15:14:52.000000000 -0400
27852 @@ -1025,7 +1025,7 @@ static int fib_seq_show(struct seq_file
27853         prefix  = f->fn_key;
27854         mask    = FZ_MASK(iter->zone);
27855         flags   = fib_flag_trans(fa->fa_type, mask, fi);
27856 -       if (fi)
27857 +       if (fi && nx_dev_visible(current->nx_info, fi->fib_dev))
27858                 snprintf(bf, sizeof(bf),
27859                          "%s\t%08X\t%08X\t%04X\t%d\t%u\t%d\t%08X\t%d\t%u\t%u",
27860                          fi->fib_dev ? fi->fib_dev->name : "*", prefix,
27861 diff -NurpP --minimal linux-2.6.25.11/net/ipv4/inet_connection_sock.c linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/inet_connection_sock.c
27862 --- linux-2.6.25.11/net/ipv4/inet_connection_sock.c     2008-04-17 12:05:44.000000000 -0400
27863 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/inet_connection_sock.c       2008-04-19 15:14:52.000000000 -0400
27864 @@ -47,10 +47,40 @@ void inet_get_local_port_range(int *low,
27865  }
27866  EXPORT_SYMBOL(inet_get_local_port_range);
27867
27868 +int ipv4_rcv_saddr_equal(const struct sock *sk1, const struct sock *sk2)
27869 +{
27870 +       __be32  sk1_rcv_saddr = inet_rcv_saddr(sk1),
27871 +               sk2_rcv_saddr = inet_rcv_saddr(sk2);
27872 +
27873 +       if (inet_v6_ipv6only(sk2))
27874 +               return 0;
27875 +
27876 +       if (sk1_rcv_saddr &&
27877 +           sk2_rcv_saddr &&
27878 +           sk1_rcv_saddr == sk2_rcv_saddr)
27879 +               return 1;
27880 +
27881 +       if (sk1_rcv_saddr &&
27882 +           !sk2_rcv_saddr &&
27883 +           v4_addr_in_nx_info(sk2->sk_nx_info, sk1_rcv_saddr, NXA_MASK_BIND))
27884 +               return 1;
27885 +
27886 +       if (sk2_rcv_saddr &&
27887 +           !sk1_rcv_saddr &&
27888 +           v4_addr_in_nx_info(sk1->sk_nx_info, sk2_rcv_saddr, NXA_MASK_BIND))
27889 +               return 1;
27890 +
27891 +       if (!sk1_rcv_saddr &&
27892 +           !sk2_rcv_saddr &&
27893 +           nx_v4_addr_conflict(sk1->sk_nx_info, sk2->sk_nx_info))
27894 +               return 1;
27895 +
27896 +       return 0;
27897 +}
27898 +
27899  int inet_csk_bind_conflict(const struct sock *sk,
27900                            const struct inet_bind_bucket *tb)
27901  {
27902 -       const __be32 sk_rcv_saddr = inet_rcv_saddr(sk);
27903         struct sock *sk2;
27904         struct hlist_node *node;
27905         int reuse = sk->sk_reuse;
27906 @@ -63,9 +93,7 @@ int inet_csk_bind_conflict(const struct
27907                      sk->sk_bound_dev_if == sk2->sk_bound_dev_if)) {
27908                         if (!reuse || !sk2->sk_reuse ||
27909                             sk2->sk_state == TCP_LISTEN) {
27910 -                               const __be32 sk2_rcv_saddr = inet_rcv_saddr(sk2);
27911 -                               if (!sk2_rcv_saddr || !sk_rcv_saddr ||
27912 -                                   sk2_rcv_saddr == sk_rcv_saddr)
27913 +                               if (ipv4_rcv_saddr_equal(sk, sk2))
27914                                         break;
27915                         }
27916                 }
27917 diff -NurpP --minimal linux-2.6.25.11/net/ipv4/inet_diag.c linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/inet_diag.c
27918 --- linux-2.6.25.11/net/ipv4/inet_diag.c        2008-04-17 12:05:44.000000000 -0400
27919 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/inet_diag.c  2008-04-19 16:05:51.000000000 -0400
27920 @@ -34,6 +34,8 @@
27921  #include <linux/stddef.h>
27922
27923  #include <linux/inet_diag.h>
27924 +#include <linux/vs_network.h>
27925 +#include <linux/vs_inet.h>
27926
27927  static const struct inet_diag_handler **inet_diag_table;
27928
27929 @@ -122,8 +124,8 @@ static int inet_csk_diag_fill(struct soc
27930
27931         r->id.idiag_sport = inet->sport;
27932         r->id.idiag_dport = inet->dport;
27933 -       r->id.idiag_src[0] = inet->rcv_saddr;
27934 -       r->id.idiag_dst[0] = inet->daddr;
27935 +       r->id.idiag_src[0] = nx_map_sock_lback(sk->sk_nx_info, inet->rcv_saddr);
27936 +       r->id.idiag_dst[0] = nx_map_sock_lback(sk->sk_nx_info, inet->daddr);
27937
27938  #if defined(CONFIG_IPV6) || defined (CONFIG_IPV6_MODULE)
27939         if (r->idiag_family == AF_INET6) {
27940 @@ -210,8 +212,8 @@ static int inet_twsk_diag_fill(struct in
27941         r->id.idiag_cookie[1] = (u32)(((unsigned long)tw >> 31) >> 1);
27942         r->id.idiag_sport     = tw->tw_sport;
27943         r->id.idiag_dport     = tw->tw_dport;
27944 -       r->id.idiag_src[0]    = tw->tw_rcv_saddr;
27945 -       r->id.idiag_dst[0]    = tw->tw_daddr;
27946 +       r->id.idiag_src[0]    = nx_map_sock_lback(tw->tw_nx_info, tw->tw_rcv_saddr);
27947 +       r->id.idiag_dst[0]    = nx_map_sock_lback(tw->tw_nx_info, tw->tw_daddr);
27948         r->idiag_state        = tw->tw_substate;
27949         r->idiag_timer        = 3;
27950         r->idiag_expires      = DIV_ROUND_UP(tmo * 1000, HZ);
27951 @@ -268,6 +270,7 @@ static int inet_diag_get_exact(struct sk
27952         err = -EINVAL;
27953
27954         if (req->idiag_family == AF_INET) {
27955 +               /* TODO: lback */
27956                 sk = inet_lookup(&init_net, hashinfo, req->id.idiag_dst[0],
27957                                  req->id.idiag_dport, req->id.idiag_src[0],
27958                                  req->id.idiag_sport, req->id.idiag_if);
27959 @@ -510,6 +513,7 @@ static int inet_csk_diag_dump(struct soc
27960                 } else
27961  #endif
27962                 {
27963 +                       /* TODO: lback */
27964                         entry.saddr = &inet->rcv_saddr;
27965                         entry.daddr = &inet->daddr;
27966                 }
27967 @@ -546,6 +550,7 @@ static int inet_twsk_diag_dump(struct in
27968                 } else
27969  #endif
27970                 {
27971 +                       /* TODO: lback */
27972                         entry.saddr = &tw->tw_rcv_saddr;
27973                         entry.daddr = &tw->tw_daddr;
27974                 }
27975 @@ -592,8 +597,8 @@ static int inet_diag_fill_req(struct sk_
27976
27977         r->id.idiag_sport = inet->sport;
27978         r->id.idiag_dport = ireq->rmt_port;
27979 -       r->id.idiag_src[0] = ireq->loc_addr;
27980 -       r->id.idiag_dst[0] = ireq->rmt_addr;
27981 +       r->id.idiag_src[0] = nx_map_sock_lback(sk->sk_nx_info, ireq->loc_addr);
27982 +       r->id.idiag_dst[0] = nx_map_sock_lback(sk->sk_nx_info, ireq->rmt_addr);
27983         r->idiag_expires = jiffies_to_msecs(tmo);
27984         r->idiag_rqueue = 0;
27985         r->idiag_wqueue = 0;
27986 @@ -663,6 +668,7 @@ static int inet_diag_dump_reqs(struct sk
27987                                 continue;
27988
27989                         if (bc) {
27990 +                               /* TODO: lback */
27991                                 entry.saddr =
27992  #if defined(CONFIG_IPV6) || defined (CONFIG_IPV6_MODULE)
27993                                         (entry.family == AF_INET6) ?
27994 @@ -731,6 +737,8 @@ static int inet_diag_dump(struct sk_buff
27995                         sk_for_each(sk, node, &hashinfo->listening_hash[i]) {
27996                                 struct inet_sock *inet = inet_sk(sk);
27997
27998 +                               if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
27999 +                                       continue;
28000                                 if (num < s_num) {
28001                                         num++;
28002                                         continue;
28003 @@ -792,6 +800,8 @@ skip_listen_ht:
28004                 sk_for_each(sk, node, &head->chain) {
28005                         struct inet_sock *inet = inet_sk(sk);
28006
28007 +                       if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
28008 +                               continue;
28009                         if (num < s_num)
28010                                 goto next_normal;
28011                         if (!(r->idiag_states & (1 << sk->sk_state)))
28012 @@ -816,6 +826,8 @@ next_normal:
28013                         inet_twsk_for_each(tw, node,
28014                                     &head->twchain) {
28015
28016 +                               if (!nx_check(tw->tw_nid, VS_WATCH_P | VS_IDENT))
28017 +                                       continue;
28018                                 if (num < s_num)
28019                                         goto next_dying;
28020                                 if (r->id.idiag_sport != tw->tw_sport &&
28021 diff -NurpP --minimal linux-2.6.25.11/net/ipv4/inet_hashtables.c linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/inet_hashtables.c
28022 --- linux-2.6.25.11/net/ipv4/inet_hashtables.c  2008-04-17 12:05:44.000000000 -0400
28023 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/inet_hashtables.c    2008-04-19 15:14:52.000000000 -0400
28024 @@ -21,6 +21,7 @@
28025
28026  #include <net/inet_connection_sock.h>
28027  #include <net/inet_hashtables.h>
28028 +#include <net/route.h>
28029  #include <net/ip.h>
28030
28031  /*
28032 @@ -144,11 +145,10 @@ static struct sock *inet_lookup_listener
28033                         const __be32 rcv_saddr = inet->rcv_saddr;
28034                         int score = sk->sk_family == PF_INET ? 1 : 0;
28035
28036 -                       if (rcv_saddr) {
28037 -                               if (rcv_saddr != daddr)
28038 -                                       continue;
28039 +                       if (v4_inet_addr_match(sk->sk_nx_info, daddr, rcv_saddr))
28040                                 score += 2;
28041 -                       }
28042 +                       else
28043 +                               continue;
28044                         if (sk->sk_bound_dev_if) {
28045                                 if (sk->sk_bound_dev_if != dif)
28046                                         continue;
28047 @@ -180,7 +180,7 @@ struct sock *__inet_lookup_listener(stru
28048                 const struct inet_sock *inet = inet_sk((sk = __sk_head(head)));
28049
28050                 if (inet->num == hnum && !sk->sk_node.next &&
28051 -                   (!inet->rcv_saddr || inet->rcv_saddr == daddr) &&
28052 +                   v4_inet_addr_match(sk->sk_nx_info, daddr, inet->rcv_saddr) &&
28053                     (sk->sk_family == PF_INET || !ipv6_only_sock(sk)) &&
28054                     !sk->sk_bound_dev_if && sk->sk_net == net)
28055                         goto sherry_cache;
28056 diff -NurpP --minimal linux-2.6.25.11/net/ipv4/netfilter/nf_nat_helper.c linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/netfilter/nf_nat_helper.c
28057 --- linux-2.6.25.11/net/ipv4/netfilter/nf_nat_helper.c  2008-04-17 12:05:45.000000000 -0400
28058 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/netfilter/nf_nat_helper.c    2008-04-19 15:14:52.000000000 -0400
28059 @@ -18,6 +18,7 @@
28060  #include <net/tcp.h>
28061
28062  #include <linux/netfilter_ipv4.h>
28063 +#include <net/route.h>
28064  #include <net/netfilter/nf_conntrack.h>
28065  #include <net/netfilter/nf_conntrack_helper.h>
28066  #include <net/netfilter/nf_conntrack_ecache.h>
28067 diff -NurpP --minimal linux-2.6.25.11/net/ipv4/netfilter.c linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/netfilter.c
28068 --- linux-2.6.25.11/net/ipv4/netfilter.c        2008-04-17 12:05:45.000000000 -0400
28069 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/netfilter.c  2008-04-19 15:14:52.000000000 -0400
28070 @@ -4,7 +4,7 @@
28071  #include <linux/netfilter_ipv4.h>
28072  #include <linux/ip.h>
28073  #include <linux/skbuff.h>
28074 -#include <net/route.h>
28075 +// #include <net/route.h>
28076  #include <net/xfrm.h>
28077  #include <net/ip.h>
28078  #include <net/netfilter/nf_queue.h>
28079 diff -NurpP --minimal linux-2.6.25.11/net/ipv4/raw.c linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/raw.c
28080 --- linux-2.6.25.11/net/ipv4/raw.c      2008-04-17 12:05:45.000000000 -0400
28081 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/raw.c        2008-04-29 20:28:52.000000000 -0400
28082 @@ -126,7 +126,7 @@ static struct sock *__raw_v4_lookup(stru
28083
28084                 if (sk->sk_net == net && inet->num == num               &&
28085                     !(inet->daddr && inet->daddr != raddr)              &&
28086 -                   !(inet->rcv_saddr && inet->rcv_saddr != laddr)      &&
28087 +                   v4_sock_addr_match(sk->sk_nx_info, inet, laddr)     &&
28088                     !(sk->sk_bound_dev_if && sk->sk_bound_dev_if != dif))
28089                         goto found; /* gotcha */
28090         }
28091 @@ -382,6 +382,12 @@ static int raw_send_hdrinc(struct sock *
28092                 icmp_out_count(((struct icmphdr *)
28093                         skb_transport_header(skb))->type);
28094
28095 +       err = -EPERM;
28096 +       if (!nx_check(0, VS_ADMIN) && !capable(CAP_NET_RAW) &&
28097 +               sk->sk_nx_info &&
28098 +               !v4_addr_in_nx_info(sk->sk_nx_info, iph->saddr, NXA_MASK_BIND))
28099 +               goto error_free;
28100 +
28101         err = NF_HOOK(PF_INET, NF_INET_LOCAL_OUT, skb, NULL, rt->u.dst.dev,
28102                       dst_output);
28103         if (err > 0)
28104 @@ -393,6 +399,7 @@ out:
28105
28106  error_fault:
28107         err = -EFAULT;
28108 +error_free:
28109         kfree_skb(skb);
28110  error:
28111         IP_INC_STATS(IPSTATS_MIB_OUTDISCARDS);
28112 @@ -560,6 +567,13 @@ static int raw_sendmsg(struct kiocb *ioc
28113                 }
28114
28115                 security_sk_classify_flow(sk, &fl);
28116 +               if (sk->sk_nx_info) {
28117 +                       err = ip_v4_find_src(sk->sk_net,
28118 +                               sk->sk_nx_info, &rt, &fl);
28119 +
28120 +                       if (err)
28121 +                               goto done;
28122 +               }
28123                 err = ip_route_output_flow(&init_net, &rt, &fl, sk, 1);
28124         }
28125         if (err)
28126 @@ -622,17 +636,19 @@ static int raw_bind(struct sock *sk, str
28127  {
28128         struct inet_sock *inet = inet_sk(sk);
28129         struct sockaddr_in *addr = (struct sockaddr_in *) uaddr;
28130 +       struct nx_v4_sock_addr nsa = { 0 };
28131         int ret = -EINVAL;
28132         int chk_addr_ret;
28133
28134         if (sk->sk_state != TCP_CLOSE || addr_len < sizeof(struct sockaddr_in))
28135                 goto out;
28136 -       chk_addr_ret = inet_addr_type(sk->sk_net, addr->sin_addr.s_addr);
28137 +       v4_map_sock_addr(inet, addr, &nsa);
28138 +       chk_addr_ret = inet_addr_type(sk->sk_net, nsa.saddr);
28139         ret = -EADDRNOTAVAIL;
28140 -       if (addr->sin_addr.s_addr && chk_addr_ret != RTN_LOCAL &&
28141 +       if (nsa.saddr && chk_addr_ret != RTN_LOCAL &&
28142             chk_addr_ret != RTN_MULTICAST && chk_addr_ret != RTN_BROADCAST)
28143                 goto out;
28144 -       inet->rcv_saddr = inet->saddr = addr->sin_addr.s_addr;
28145 +       v4_set_sock_addr(inet, &nsa);
28146         if (chk_addr_ret == RTN_MULTICAST || chk_addr_ret == RTN_BROADCAST)
28147                 inet->saddr = 0;  /* Use device */
28148         sk_dst_reset(sk);
28149 @@ -684,7 +700,8 @@ static int raw_recvmsg(struct kiocb *ioc
28150         /* Copy the address. */
28151         if (sin) {
28152                 sin->sin_family = AF_INET;
28153 -               sin->sin_addr.s_addr = ip_hdr(skb)->saddr;
28154 +               sin->sin_addr.s_addr =
28155 +                       nx_map_sock_lback(sk->sk_nx_info, ip_hdr(skb)->saddr);
28156                 sin->sin_port = 0;
28157                 memset(&sin->sin_zero, 0, sizeof(sin->sin_zero));
28158         }
28159 @@ -862,7 +879,8 @@ static struct sock *raw_get_first(struct
28160                 struct hlist_node *node;
28161
28162                 sk_for_each(sk, node, &state->h->ht[state->bucket])
28163 -                       if (sk->sk_net == state->p.net)
28164 +                       if ((sk->sk_net == state->p.net) &&
28165 +                               nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
28166                                 goto found;
28167         }
28168         sk = NULL;
28169 @@ -878,7 +896,8 @@ static struct sock *raw_get_next(struct
28170                 sk = sk_next(sk);
28171  try_again:
28172                 ;
28173 -       } while (sk && sk->sk_net != state->p.net);
28174 +       } while (sk && ((sk->sk_net != state->p.net) ||
28175 +               !nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT)));
28176
28177         if (!sk && ++state->bucket < RAW_HTABLE_SIZE) {
28178                 sk = sk_head(&state->h->ht[state->bucket]);
28179 @@ -937,7 +956,10 @@ static void raw_sock_seq_show(struct seq
28180
28181         seq_printf(seq, "%4d: %08X:%04X %08X:%04X"
28182                 " %02X %08X:%08X %02X:%08lX %08X %5d %8d %lu %d %p %d\n",
28183 -               i, src, srcp, dest, destp, sp->sk_state,
28184 +               i,
28185 +               nx_map_sock_lback(current_nx_info(), src), srcp,
28186 +               nx_map_sock_lback(current_nx_info(), dest), destp,
28187 +               sp->sk_state,
28188                 atomic_read(&sp->sk_wmem_alloc),
28189                 atomic_read(&sp->sk_rmem_alloc),
28190                 0, 0L, 0, sock_i_uid(sp), 0, sock_i_ino(sp),
28191 diff -NurpP --minimal linux-2.6.25.11/net/ipv4/syncookies.c linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/syncookies.c
28192 --- linux-2.6.25.11/net/ipv4/syncookies.c       2008-04-17 12:05:45.000000000 -0400
28193 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/syncookies.c 2008-04-19 15:14:52.000000000 -0400
28194 @@ -20,6 +20,7 @@
28195  #include <linux/cryptohash.h>
28196  #include <linux/kernel.h>
28197  #include <net/tcp.h>
28198 +#include <net/route.h>
28199
28200  extern int sysctl_tcp_syncookies;
28201
28202 diff -NurpP --minimal linux-2.6.25.11/net/ipv4/tcp.c linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/tcp.c
28203 --- linux-2.6.25.11/net/ipv4/tcp.c      2008-04-17 12:05:45.000000000 -0400
28204 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/tcp.c        2008-04-19 15:14:52.000000000 -0400
28205 @@ -263,6 +263,7 @@
28206  #include <linux/cache.h>
28207  #include <linux/err.h>
28208  #include <linux/crypto.h>
28209 +#include <linux/in.h>
28210
28211  #include <net/icmp.h>
28212  #include <net/tcp.h>
28213 diff -NurpP --minimal linux-2.6.25.11/net/ipv4/tcp_ipv4.c linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/tcp_ipv4.c
28214 --- linux-2.6.25.11/net/ipv4/tcp_ipv4.c 2008-04-17 12:05:45.000000000 -0400
28215 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/tcp_ipv4.c   2008-04-19 15:14:52.000000000 -0400
28216 @@ -1965,6 +1965,12 @@ static void *listening_get_next(struct s
28217                 req = req->dl_next;
28218                 while (1) {
28219                         while (req) {
28220 +                               vxdprintk(VXD_CBIT(net, 6),
28221 +                                       "sk,req: %p [#%d] (from %d)", req->sk,
28222 +                                       (req->sk)?req->sk->sk_nid:0, nx_current_nid());
28223 +                               if (req->sk &&
28224 +                                       !nx_check(req->sk->sk_nid, VS_WATCH_P | VS_IDENT))
28225 +                                       continue;
28226                                 if (req->rsk_ops->family == st->family) {
28227                                         cur = req;
28228                                         goto out;
28229 @@ -1989,6 +1995,10 @@ get_req:
28230         }
28231  get_sk:
28232         sk_for_each_from(sk, node) {
28233 +               vxdprintk(VXD_CBIT(net, 6), "sk: %p [#%d] (from %d)",
28234 +                       sk, sk->sk_nid, nx_current_nid());
28235 +               if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
28236 +                       continue;
28237                 if (sk->sk_family == st->family) {
28238                         cur = sk;
28239                         goto out;
28240 @@ -2038,18 +2048,26 @@ static void *established_get_first(struc
28241
28242                 read_lock_bh(lock);
28243                 sk_for_each(sk, node, &tcp_hashinfo.ehash[st->bucket].chain) {
28244 -                       if (sk->sk_family != st->family) {
28245 +                       vxdprintk(VXD_CBIT(net, 6),
28246 +                               "sk,egf: %p [#%d] (from %d)",
28247 +                               sk, sk->sk_nid, nx_current_nid());
28248 +                       if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
28249 +                               continue;
28250 +                       if (sk->sk_family != st->family)
28251                                 continue;
28252 -                       }
28253                         rc = sk;
28254                         goto out;
28255                 }
28256                 st->state = TCP_SEQ_STATE_TIME_WAIT;
28257                 inet_twsk_for_each(tw, node,
28258                                    &tcp_hashinfo.ehash[st->bucket].twchain) {
28259 -                       if (tw->tw_family != st->family) {
28260 +                       vxdprintk(VXD_CBIT(net, 6),
28261 +                               "tw: %p [#%d] (from %d)",
28262 +                               tw, tw->tw_nid, nx_current_nid());
28263 +                       if (!nx_check(tw->tw_nid, VS_WATCH_P | VS_IDENT))
28264 +                               continue;
28265 +                       if (tw->tw_family != st->family)
28266                                 continue;
28267 -                       }
28268                         rc = tw;
28269                         goto out;
28270                 }
28271 @@ -2073,7 +2091,8 @@ static void *established_get_next(struct
28272                 tw = cur;
28273                 tw = tw_next(tw);
28274  get_tw:
28275 -               while (tw && tw->tw_family != st->family) {
28276 +               while (tw && (tw->tw_family != st->family ||
28277 +                       !nx_check(tw->tw_nid, VS_WATCH_P | VS_IDENT))) {
28278                         tw = tw_next(tw);
28279                 }
28280                 if (tw) {
28281 @@ -2094,6 +2113,11 @@ get_tw:
28282                 sk = sk_next(sk);
28283
28284         sk_for_each_from(sk, node) {
28285 +               vxdprintk(VXD_CBIT(net, 6),
28286 +                       "sk,egn: %p [#%d] (from %d)",
28287 +                       sk, sk->sk_nid, nx_current_nid());
28288 +               if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
28289 +                       continue;
28290                 if (sk->sk_family == st->family)
28291                         goto found;
28292         }
28293 @@ -2266,9 +2290,9 @@ static void get_openreq4(struct sock *sk
28294         sprintf(tmpbuf, "%4d: %08X:%04X %08X:%04X"
28295                 " %02X %08X:%08X %02X:%08lX %08X %5d %8d %u %d %p",
28296                 i,
28297 -               ireq->loc_addr,
28298 +               nx_map_sock_lback(current_nx_info(), ireq->loc_addr),
28299                 ntohs(inet_sk(sk)->sport),
28300 -               ireq->rmt_addr,
28301 +               nx_map_sock_lback(current_nx_info(), ireq->rmt_addr),
28302                 ntohs(ireq->rmt_port),
28303                 TCP_SYN_RECV,
28304                 0, 0, /* could print option size, but that is af dependent. */
28305 @@ -2310,7 +2334,10 @@ static void get_tcp4_sock(struct sock *s
28306
28307         sprintf(tmpbuf, "%4d: %08X:%04X %08X:%04X %02X %08X:%08X %02X:%08lX "
28308                         "%08X %5d %8d %lu %d %p %u %u %u %u %d",
28309 -               i, src, srcp, dest, destp, sk->sk_state,
28310 +               i,
28311 +               nx_map_sock_lback(current_nx_info(), src), srcp,
28312 +               nx_map_sock_lback(current_nx_info(), dest), destp,
28313 +               sk->sk_state,
28314                 tp->write_seq - tp->snd_una,
28315                 sk->sk_state == TCP_LISTEN ? sk->sk_ack_backlog :
28316                                              (tp->rcv_nxt - tp->copied_seq),
28317 @@ -2345,7 +2372,10 @@ static void get_timewait4_sock(struct in
28318
28319         sprintf(tmpbuf, "%4d: %08X:%04X %08X:%04X"
28320                 " %02X %08X:%08X %02X:%08lX %08X %5d %8d %d %d %p",
28321 -               i, src, srcp, dest, destp, tw->tw_substate, 0, 0,
28322 +               i,
28323 +               nx_map_sock_lback(current_nx_info(), src), srcp,
28324 +               nx_map_sock_lback(current_nx_info(), dest), destp,
28325 +               tw->tw_substate, 0, 0,
28326                 3, jiffies_to_clock_t(ttd), 0, 0, 0, 0,
28327                 atomic_read(&tw->tw_refcnt), tw);
28328  }
28329 diff -NurpP --minimal linux-2.6.25.11/net/ipv4/tcp_minisocks.c linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/tcp_minisocks.c
28330 --- linux-2.6.25.11/net/ipv4/tcp_minisocks.c    2008-04-17 11:31:40.000000000 -0400
28331 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/tcp_minisocks.c      2008-04-19 15:14:52.000000000 -0400
28332 @@ -28,6 +28,10 @@
28333  #include <net/inet_common.h>
28334  #include <net/xfrm.h>
28335
28336 +#include <linux/vs_limit.h>
28337 +#include <linux/vs_socket.h>
28338 +#include <linux/vs_context.h>
28339 +
28340  #ifdef CONFIG_SYSCTL
28341  #define SYNC_INIT 0 /* let the user enable it */
28342  #else
28343 @@ -293,6 +297,11 @@ void tcp_time_wait(struct sock *sk, int
28344                 tcptw->tw_ts_recent     = tp->rx_opt.ts_recent;
28345                 tcptw->tw_ts_recent_stamp = tp->rx_opt.ts_recent_stamp;
28346
28347 +               tw->tw_xid              = sk->sk_xid;
28348 +               tw->tw_vx_info          = NULL;
28349 +               tw->tw_nid              = sk->sk_nid;
28350 +               tw->tw_nx_info          = NULL;
28351 +
28352  #if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
28353                 if (tw->tw_family == PF_INET6) {
28354                         struct ipv6_pinfo *np = inet6_sk(sk);
28355 diff -NurpP --minimal linux-2.6.25.11/net/ipv4/udp.c linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/udp.c
28356 --- linux-2.6.25.11/net/ipv4/udp.c      2008-04-17 12:05:45.000000000 -0400
28357 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv4/udp.c        2008-04-21 12:41:01.000000000 -0400
28358 @@ -246,14 +246,7 @@ int udp_get_port(struct sock *sk, unsign
28359         return  __udp_lib_get_port(sk, snum, udp_hash, scmp);
28360  }
28361
28362 -int ipv4_rcv_saddr_equal(const struct sock *sk1, const struct sock *sk2)
28363 -{
28364 -       struct inet_sock *inet1 = inet_sk(sk1), *inet2 = inet_sk(sk2);
28365 -
28366 -       return  ( !ipv6_only_sock(sk2)  &&
28367 -                 (!inet1->rcv_saddr || !inet2->rcv_saddr ||
28368 -                  inet1->rcv_saddr == inet2->rcv_saddr      ));
28369 -}
28370 +extern int ipv4_rcv_saddr_equal(const struct sock *sk1, const struct sock *sk2);
28371
28372  static inline int udp_v4_get_port(struct sock *sk, unsigned short snum)
28373  {
28374 @@ -273,16 +266,23 @@ static struct sock *__udp4_lib_lookup(st
28375         int badness = -1;
28376
28377         read_lock(&udp_hash_lock);
28378 +
28379         sk_for_each(sk, node, &udptable[hnum & (UDP_HTABLE_SIZE - 1)]) {
28380                 struct inet_sock *inet = inet_sk(sk);
28381
28382                 if (sk->sk_net == net && sk->sk_hash == hnum &&
28383                                 !ipv6_only_sock(sk)) {
28384                         int score = (sk->sk_family == PF_INET ? 1 : 0);
28385 +
28386                         if (inet->rcv_saddr) {
28387                                 if (inet->rcv_saddr != daddr)
28388                                         continue;
28389                                 score+=2;
28390 +                       } else {
28391 +                               /* block non nx_info ips */
28392 +                               if (!v4_addr_in_nx_info(sk->sk_nx_info,
28393 +                                       daddr, NXA_MASK_BIND))
28394 +                                       continue;
28395                         }
28396                         if (inet->daddr) {
28397                                 if (inet->daddr != saddr)
28398 @@ -308,6 +308,7 @@ static struct sock *__udp4_lib_lookup(st
28399                         }
28400                 }
28401         }
28402 +
28403         if (result)
28404                 sock_hold(result);
28405         read_unlock(&udp_hash_lock);
28406 @@ -329,7 +330,7 @@ static inline struct sock *udp_v4_mcast_
28407                 if (s->sk_hash != hnum                                  ||
28408                     (inet->daddr && inet->daddr != rmt_addr)            ||
28409                     (inet->dport != rmt_port && inet->dport)            ||
28410 -                   (inet->rcv_saddr && inet->rcv_saddr != loc_addr)    ||
28411 +                   !v4_sock_addr_match(sk->sk_nx_info, inet, loc_addr) ||
28412                     ipv6_only_sock(s)                                   ||
28413                     (s->sk_bound_dev_if && s->sk_bound_dev_if != dif))
28414                         continue;
28415 @@ -662,7 +663,15 @@ int udp_sendmsg(struct kiocb *iocb, stru
28416                                     .uli_u = { .ports =
28417                                                { .sport = inet->sport,
28418                                                  .dport = dport } } };
28419 +               struct nx_info *nxi = sk->sk_nx_info;
28420 +               struct net *net = sk->sk_net;
28421 +
28422                 security_sk_classify_flow(sk, &fl);
28423 +
28424 +               err = ip_v4_find_src(net, nxi, &rt, &fl);
28425 +               if (err)
28426 +                       goto out;
28427 +
28428                 err = ip_route_output_flow(&init_net, &rt, &fl, sk, 1);
28429                 if (err) {
28430                         if (err == -ENETUNREACH)
28431 @@ -905,7 +914,8 @@ try_again:
28432         {
28433                 sin->sin_family = AF_INET;
28434                 sin->sin_port = udp_hdr(skb)->source;
28435 -               sin->sin_addr.s_addr = ip_hdr(skb)->saddr;
28436 +               sin->sin_addr.s_addr = nx_map_sock_lback(
28437 +                       skb->sk->sk_nx_info, ip_hdr(skb)->saddr);
28438                 memset(sin->sin_zero, 0, sizeof(sin->sin_zero));
28439         }
28440         if (inet->cmsg_flags)
28441 @@ -1516,7 +1526,8 @@ static struct sock *udp_get_first(struct
28442         for (state->bucket = 0; state->bucket < UDP_HTABLE_SIZE; ++state->bucket) {
28443                 struct hlist_node *node;
28444                 sk_for_each(sk, node, state->hashtable + state->bucket) {
28445 -                       if (sk->sk_family == state->family)
28446 +                       if (sk->sk_family == state->family &&
28447 +                               nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
28448                                 goto found;
28449                 }
28450         }
28451 @@ -1533,7 +1544,8 @@ static struct sock *udp_get_next(struct
28452                 sk = sk_next(sk);
28453  try_again:
28454                 ;
28455 -       } while (sk && sk->sk_family != state->family);
28456 +       } while (sk && (sk->sk_family != state->family ||
28457 +               !nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT)));
28458
28459         if (!sk && ++state->bucket < UDP_HTABLE_SIZE) {
28460                 sk = sk_head(state->hashtable + state->bucket);
28461 @@ -1648,7 +1660,10 @@ static void udp4_format_sock(struct sock
28462
28463         sprintf(tmpbuf, "%4d: %08X:%04X %08X:%04X"
28464                 " %02X %08X:%08X %02X:%08lX %08X %5d %8d %lu %d %p",
28465 -               bucket, src, srcp, dest, destp, sp->sk_state,
28466 +               bucket,
28467 +               nx_map_sock_lback(current_nx_info(), src), srcp,
28468 +               nx_map_sock_lback(current_nx_info(), dest), destp,
28469 +               sp->sk_state,
28470                 atomic_read(&sp->sk_wmem_alloc),
28471                 atomic_read(&sp->sk_rmem_alloc),
28472                 0, 0L, 0, sock_i_uid(sp), 0, sock_i_ino(sp),
28473 diff -NurpP --minimal linux-2.6.25.11/net/ipv6/addrconf.c linux-2.6.25.11-vs2.3.0.34.14/net/ipv6/addrconf.c
28474 --- linux-2.6.25.11/net/ipv6/addrconf.c 2008-04-17 12:05:45.000000000 -0400
28475 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv6/addrconf.c   2008-06-23 16:09:17.000000000 -0400
28476 @@ -87,6 +87,8 @@
28477
28478  #include <linux/proc_fs.h>
28479  #include <linux/seq_file.h>
28480 +#include <linux/vs_network.h>
28481 +#include <linux/vs_inet6.h>
28482
28483  /* Set to 3 to get tracing... */
28484  #define ACONF_DEBUG 2
28485 @@ -918,7 +920,8 @@ static inline int ipv6_saddr_preferred(i
28486  }
28487
28488  int ipv6_dev_get_saddr(struct net_device *daddr_dev,
28489 -                      struct in6_addr *daddr, struct in6_addr *saddr)
28490 +                      struct in6_addr *daddr, struct in6_addr *saddr,
28491 +                      struct nx_info *nxi)
28492  {
28493         struct ipv6_saddr_score hiscore;
28494         struct inet6_ifaddr *ifa_result = NULL;
28495 @@ -964,6 +967,10 @@ int ipv6_dev_get_saddr(struct net_device
28496
28497                         score.addr_type = __ipv6_addr_type(&ifa->addr);
28498
28499 +                       /* Use only addresses assigned to the context */
28500 +                       if (!v6_ifa_in_nx_info(ifa, nxi))
28501 +                               continue;
28502 +
28503                         /* Rule 0:
28504                          * - Tentative Address (RFC2462 section 5.4)
28505                          *  - A tentative address is not considered
28506 @@ -1181,9 +1188,10 @@ record_it:
28507
28508
28509  int ipv6_get_saddr(struct dst_entry *dst,
28510 -                  struct in6_addr *daddr, struct in6_addr *saddr)
28511 +                  struct in6_addr *daddr, struct in6_addr *saddr,
28512 +                  struct nx_info *nxi)
28513  {
28514 -       return ipv6_dev_get_saddr(dst ? ip6_dst_idev(dst)->dev : NULL, daddr, saddr);
28515 +       return ipv6_dev_get_saddr(dst ? ip6_dst_idev(dst)->dev : NULL, daddr, saddr, nxi);
28516  }
28517
28518  EXPORT_SYMBOL(ipv6_get_saddr);
28519 @@ -1287,35 +1295,46 @@ struct inet6_ifaddr *ipv6_get_ifaddr(str
28520         return ifp;
28521  }
28522
28523 +extern int ipv4_rcv_saddr_equal(const struct sock *sk1, const struct sock *sk2);
28524 +
28525  int ipv6_rcv_saddr_equal(const struct sock *sk, const struct sock *sk2)
28526  {
28527         const struct in6_addr *sk_rcv_saddr6 = &inet6_sk(sk)->rcv_saddr;
28528         const struct in6_addr *sk2_rcv_saddr6 = inet6_rcv_saddr(sk2);
28529 -       __be32 sk_rcv_saddr = inet_sk(sk)->rcv_saddr;
28530         __be32 sk2_rcv_saddr = inet_rcv_saddr(sk2);
28531         int sk_ipv6only = ipv6_only_sock(sk);
28532         int sk2_ipv6only = inet_v6_ipv6only(sk2);
28533         int addr_type = ipv6_addr_type(sk_rcv_saddr6);
28534         int addr_type2 = sk2_rcv_saddr6 ? ipv6_addr_type(sk2_rcv_saddr6) : IPV6_ADDR_MAPPED;
28535
28536 -       if (!sk2_rcv_saddr && !sk_ipv6only)
28537 +       /* FIXME: needs handling for v4 ANY */
28538 +       if (!sk2_rcv_saddr && !sk_ipv6only && !sk2->sk_nx_info)
28539                 return 1;
28540
28541         if (addr_type2 == IPV6_ADDR_ANY &&
28542 -           !(sk2_ipv6only && addr_type == IPV6_ADDR_MAPPED))
28543 +           !(sk2_ipv6only && addr_type == IPV6_ADDR_MAPPED) &&
28544 +           v6_addr_in_nx_info(sk2->sk_nx_info, sk_rcv_saddr6, -1))
28545                 return 1;
28546
28547         if (addr_type == IPV6_ADDR_ANY &&
28548 -           !(sk_ipv6only && addr_type2 == IPV6_ADDR_MAPPED))
28549 +           !(sk_ipv6only && addr_type2 == IPV6_ADDR_MAPPED) &&
28550 +           (sk2_rcv_saddr6 && v6_addr_in_nx_info(sk->sk_nx_info, sk2_rcv_saddr6, -1)))
28551 +               return 1;
28552 +
28553 +       if (addr_type == IPV6_ADDR_ANY &&
28554 +           addr_type2 == IPV6_ADDR_ANY &&
28555 +           nx_v6_addr_conflict(sk->sk_nx_info, sk2->sk_nx_info))
28556                 return 1;
28557
28558         if (sk2_rcv_saddr6 &&
28559 +           addr_type != IPV6_ADDR_ANY &&
28560 +           addr_type != IPV6_ADDR_ANY &&
28561             ipv6_addr_equal(sk_rcv_saddr6, sk2_rcv_saddr6))
28562                 return 1;
28563
28564         if (addr_type == IPV6_ADDR_MAPPED &&
28565             !sk2_ipv6only &&
28566 -           (!sk2_rcv_saddr || !sk_rcv_saddr || sk_rcv_saddr == sk2_rcv_saddr))
28567 +           ipv4_rcv_saddr_equal(sk, sk2))
28568                 return 1;
28569
28570         return 0;
28571 @@ -2846,7 +2865,10 @@ static void if6_seq_stop(struct seq_file
28572  static int if6_seq_show(struct seq_file *seq, void *v)
28573  {
28574         struct inet6_ifaddr *ifp = (struct inet6_ifaddr *)v;
28575 -       seq_printf(seq,
28576 +
28577 +       if (nx_check(0, VS_ADMIN|VS_WATCH) ||
28578 +           v6_addr_in_nx_info(current_nx_info(), &ifp->addr, -1))
28579 +               seq_printf(seq,
28580                    NIP6_SEQFMT " %02x %02x %02x %02x %8s\n",
28581                    NIP6(ifp->addr),
28582                    ifp->idev->dev->ifindex,
28583 @@ -3337,6 +3359,11 @@ static int inet6_dump_addr(struct sk_buf
28584         struct inet6_ifaddr *ifa;
28585         struct ifmcaddr6 *ifmca;
28586         struct ifacaddr6 *ifaca;
28587 +       struct nx_info *nxi = skb->sk ? skb->sk->sk_nx_info : NULL;
28588 +
28589 +       /* disable ipv6 on non v6 guests */
28590 +       if (nxi && !nx_info_has_v6(nxi))
28591 +               return skb->len;
28592
28593         s_idx = cb->args[0];
28594         s_ip_idx = ip_idx = cb->args[1];
28595 @@ -3358,6 +3385,8 @@ static int inet6_dump_addr(struct sk_buf
28596                              ifa = ifa->if_next, ip_idx++) {
28597                                 if (ip_idx < s_ip_idx)
28598                                         continue;
28599 +                               if (!v6_addr_in_nx_info(nxi, &ifa->addr, -1))
28600 +                                       continue;
28601                                 err = inet6_fill_ifaddr(skb, ifa,
28602                                                         NETLINK_CB(cb->skb).pid,
28603                                                         cb->nlh->nlmsg_seq,
28604 @@ -3371,6 +3400,8 @@ static int inet6_dump_addr(struct sk_buf
28605                              ifmca = ifmca->next, ip_idx++) {
28606                                 if (ip_idx < s_ip_idx)
28607                                         continue;
28608 +                               if (!v6_addr_in_nx_info(nxi, &ifmca->mca_addr, -1))
28609 +                                       continue;
28610                                 err = inet6_fill_ifmcaddr(skb, ifmca,
28611                                                           NETLINK_CB(cb->skb).pid,
28612                                                           cb->nlh->nlmsg_seq,
28613 @@ -3384,6 +3415,8 @@ static int inet6_dump_addr(struct sk_buf
28614                              ifaca = ifaca->aca_next, ip_idx++) {
28615                                 if (ip_idx < s_ip_idx)
28616                                         continue;
28617 +                               if (!v6_addr_in_nx_info(nxi, &ifaca->aca_addr, -1))
28618 +                                       continue;
28619                                 err = inet6_fill_ifacaddr(skb, ifaca,
28620                                                           NETLINK_CB(cb->skb).pid,
28621                                                           cb->nlh->nlmsg_seq,
28622 @@ -3678,6 +3711,11 @@ static int inet6_dump_ifinfo(struct sk_b
28623         int s_idx = cb->args[0];
28624         struct net_device *dev;
28625         struct inet6_dev *idev;
28626 +       struct nx_info *nxi = skb->sk ? skb->sk->sk_nx_info : NULL;
28627 +
28628 +       /* FIXME: maybe disable ipv6 on non v6 guests?
28629 +       if (skb->sk && skb->sk->sk_vx_info)
28630 +               return skb->len; */
28631
28632         if (net != &init_net)
28633                 return 0;
28634 @@ -3687,6 +3725,8 @@ static int inet6_dump_ifinfo(struct sk_b
28635         for_each_netdev(&init_net, dev) {
28636                 if (idx < s_idx)
28637                         goto cont;
28638 +               if (!v6_dev_in_nx_info(dev, nxi))
28639 +                       goto cont;
28640                 if ((idev = in6_dev_get(dev)) == NULL)
28641                         goto cont;
28642                 err = inet6_fill_ifinfo(skb, idev, NETLINK_CB(cb->skb).pid,
28643 diff -NurpP --minimal linux-2.6.25.11/net/ipv6/af_inet6.c linux-2.6.25.11-vs2.3.0.34.14/net/ipv6/af_inet6.c
28644 --- linux-2.6.25.11/net/ipv6/af_inet6.c 2008-04-17 12:05:45.000000000 -0400
28645 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv6/af_inet6.c   2008-04-19 15:14:52.000000000 -0400
28646 @@ -43,6 +43,8 @@
28647  #include <linux/netdevice.h>
28648  #include <linux/icmpv6.h>
28649  #include <linux/netfilter_ipv6.h>
28650 +#include <linux/vs_inet.h>
28651 +#include <linux/vs_inet6.h>
28652
28653  #include <net/ip.h>
28654  #include <net/ipv6.h>
28655 @@ -51,6 +53,7 @@
28656  #include <net/tcp.h>
28657  #include <net/ipip.h>
28658  #include <net/protocol.h>
28659 +#include <net/route.h>
28660  #include <net/inet_common.h>
28661  #include <net/transp_v6.h>
28662  #include <net/ip6_route.h>
28663 @@ -148,9 +151,12 @@ lookup_protocol:
28664         }
28665
28666         err = -EPERM;
28667 +       if ((protocol == IPPROTO_ICMPV6) &&
28668 +               nx_capable(answer->capability, NXC_RAW_ICMP))
28669 +               goto override;
28670         if (answer->capability > 0 && !capable(answer->capability))
28671                 goto out_rcu_unlock;
28672 -
28673 +override:
28674         sock->ops = answer->ops;
28675         answer_prot = answer->prot;
28676         answer_no_check = answer->no_check;
28677 @@ -248,6 +254,7 @@ int inet6_bind(struct socket *sock, stru
28678         struct sock *sk = sock->sk;
28679         struct inet_sock *inet = inet_sk(sk);
28680         struct ipv6_pinfo *np = inet6_sk(sk);
28681 +       struct nx_v6_sock_addr nsa;
28682         __be32 v4addr = 0;
28683         unsigned short snum;
28684         int addr_type = 0;
28685 @@ -259,6 +266,11 @@ int inet6_bind(struct socket *sock, stru
28686
28687         if (addr_len < SIN6_LEN_RFC2133)
28688                 return -EINVAL;
28689 +
28690 +       err = v6_map_sock_addr(inet, addr, &nsa);
28691 +       if (err)
28692 +               return err;
28693 +
28694         addr_type = ipv6_addr_type(&addr->sin6_addr);
28695         if ((addr_type & IPV6_ADDR_MULTICAST) && sock->type == SOCK_STREAM)
28696                 return -EINVAL;
28697 @@ -282,6 +294,10 @@ int inet6_bind(struct socket *sock, stru
28698                         err = -EADDRNOTAVAIL;
28699                         goto out;
28700                 }
28701 +               if (!v4_addr_in_nx_info(sk->sk_nx_info, v4addr, NXA_MASK_BIND)) {
28702 +                       err = -EADDRNOTAVAIL;
28703 +                       goto out;
28704 +               }
28705         } else {
28706                 if (addr_type != IPV6_ADDR_ANY) {
28707                         struct net_device *dev = NULL;
28708 @@ -307,6 +323,11 @@ int inet6_bind(struct socket *sock, stru
28709                                 }
28710                         }
28711
28712 +                       if (!v6_addr_in_nx_info(sk->sk_nx_info, &addr->sin6_addr, -1)) {
28713 +                               err = -EADDRNOTAVAIL;
28714 +                               goto out;
28715 +                       }
28716 +
28717                         /* ipv4 addr of the socket is invalid.  Only the
28718                          * unspecified and mapped address have a v4 equivalent.
28719                          */
28720 @@ -325,6 +346,8 @@ int inet6_bind(struct socket *sock, stru
28721                 }
28722         }
28723
28724 +       v6_set_sock_addr(inet, &nsa);
28725 +
28726         inet->rcv_saddr = v4addr;
28727         inet->saddr = v4addr;
28728
28729 @@ -419,9 +442,11 @@ int inet6_getname(struct socket *sock, s
28730                         return -ENOTCONN;
28731                 sin->sin6_port = inet->dport;
28732                 ipv6_addr_copy(&sin->sin6_addr, &np->daddr);
28733 +               /* FIXME: remap lback? */
28734                 if (np->sndflow)
28735                         sin->sin6_flowinfo = np->flow_label;
28736         } else {
28737 +               /* FIXME: remap lback? */
28738                 if (ipv6_addr_any(&np->rcv_saddr))
28739                         ipv6_addr_copy(&sin->sin6_addr, &np->saddr);
28740                 else
28741 diff -NurpP --minimal linux-2.6.25.11/net/ipv6/fib6_rules.c linux-2.6.25.11-vs2.3.0.34.14/net/ipv6/fib6_rules.c
28742 --- linux-2.6.25.11/net/ipv6/fib6_rules.c       2008-04-17 12:05:45.000000000 -0400
28743 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv6/fib6_rules.c 2008-04-19 15:14:52.000000000 -0400
28744 @@ -86,7 +86,7 @@ static int fib6_rule_action(struct fib_r
28745                     r->src.plen && !(flags & RT6_LOOKUP_F_HAS_SADDR)) {
28746                         struct in6_addr saddr;
28747                         if (ipv6_get_saddr(&rt->u.dst, &flp->fl6_dst,
28748 -                                          &saddr))
28749 +                                          &saddr, NULL))
28750                                 goto again;
28751                         if (!ipv6_prefix_equal(&saddr, &r->src.addr,
28752                                                r->src.plen))
28753 diff -NurpP --minimal linux-2.6.25.11/net/ipv6/inet6_hashtables.c linux-2.6.25.11-vs2.3.0.34.14/net/ipv6/inet6_hashtables.c
28754 --- linux-2.6.25.11/net/ipv6/inet6_hashtables.c 2008-04-17 12:05:45.000000000 -0400
28755 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv6/inet6_hashtables.c   2008-04-19 15:14:52.000000000 -0400
28756 @@ -16,6 +16,7 @@
28757
28758  #include <linux/module.h>
28759  #include <linux/random.h>
28760 +#include <linux/vs_inet6.h>
28761
28762  #include <net/inet_connection_sock.h>
28763  #include <net/inet_hashtables.h>
28764 @@ -114,6 +115,9 @@ struct sock *inet6_lookup_listener(struc
28765                                 if (!ipv6_addr_equal(&np->rcv_saddr, daddr))
28766                                         continue;
28767                                 score++;
28768 +                       } else {
28769 +                               if (!v6_addr_in_nx_info(sk->sk_nx_info, daddr, -1))
28770 +                                       continue;
28771                         }
28772                         if (sk->sk_bound_dev_if) {
28773                                 if (sk->sk_bound_dev_if != dif)
28774 diff -NurpP --minimal linux-2.6.25.11/net/ipv6/ip6_output.c linux-2.6.25.11-vs2.3.0.34.14/net/ipv6/ip6_output.c
28775 --- linux-2.6.25.11/net/ipv6/ip6_output.c       2008-04-17 12:05:45.000000000 -0400
28776 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv6/ip6_output.c 2008-04-19 15:14:52.000000000 -0400
28777 @@ -920,7 +920,7 @@ static int ip6_dst_lookup_tail(struct so
28778                 goto out_err_release;
28779
28780         if (ipv6_addr_any(&fl->fl6_src)) {
28781 -               err = ipv6_get_saddr(*dst, &fl->fl6_dst, &fl->fl6_src);
28782 +               err = ipv6_get_saddr(*dst, &fl->fl6_dst, &fl->fl6_src, sk->sk_nx_info);
28783                 if (err)
28784                         goto out_err_release;
28785         }
28786 diff -NurpP --minimal linux-2.6.25.11/net/ipv6/Kconfig linux-2.6.25.11-vs2.3.0.34.14/net/ipv6/Kconfig
28787 --- linux-2.6.25.11/net/ipv6/Kconfig    2008-04-17 12:05:45.000000000 -0400
28788 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv6/Kconfig      2008-04-19 15:14:52.000000000 -0400
28789 @@ -4,8 +4,8 @@
28790
28791  #   IPv6 as module will cause a CRASH if you try to unload it
28792  config IPV6
28793 -       tristate "The IPv6 protocol"
28794 -       default m
28795 +       bool "The IPv6 protocol"
28796 +       default n
28797         ---help---
28798           This is complemental support for the IP version 6.
28799           You will still be able to do traditional IPv4 networking as well.
28800 diff -NurpP --minimal linux-2.6.25.11/net/ipv6/ndisc.c linux-2.6.25.11-vs2.3.0.34.14/net/ipv6/ndisc.c
28801 --- linux-2.6.25.11/net/ipv6/ndisc.c    2008-04-17 12:05:45.000000000 -0400
28802 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv6/ndisc.c      2008-04-19 15:14:52.000000000 -0400
28803 @@ -563,7 +563,7 @@ static void ndisc_send_na(struct net_dev
28804                         override = 0;
28805                 in6_ifa_put(ifp);
28806         } else {
28807 -               if (ipv6_dev_get_saddr(dev, daddr, &tmpaddr))
28808 +               if (ipv6_dev_get_saddr(dev, daddr, &tmpaddr, NULL))
28809                         return;
28810                 src_addr = &tmpaddr;
28811         }
28812 diff -NurpP --minimal linux-2.6.25.11/net/ipv6/route.c linux-2.6.25.11-vs2.3.0.34.14/net/ipv6/route.c
28813 --- linux-2.6.25.11/net/ipv6/route.c    2008-07-15 22:05:18.000000000 -0400
28814 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv6/route.c      2008-06-23 15:22:21.000000000 -0400
28815 @@ -2122,7 +2122,7 @@ static int rt6_fill_node(struct sk_buff
28816                 NLA_PUT_U32(skb, RTA_IIF, iif);
28817         else if (dst) {
28818                 struct in6_addr saddr_buf;
28819 -               if (ipv6_get_saddr(&rt->u.dst, dst, &saddr_buf) == 0)
28820 +               if (ipv6_get_saddr(&rt->u.dst, dst, &saddr_buf, (skb->sk ? skb->sk->sk_nx_info : NULL)) == 0)
28821                         NLA_PUT(skb, RTA_PREFSRC, 16, &saddr_buf);
28822         }
28823
28824 diff -NurpP --minimal linux-2.6.25.11/net/ipv6/tcp_ipv6.c linux-2.6.25.11-vs2.3.0.34.14/net/ipv6/tcp_ipv6.c
28825 --- linux-2.6.25.11/net/ipv6/tcp_ipv6.c 2008-04-17 12:05:45.000000000 -0400
28826 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv6/tcp_ipv6.c   2008-04-19 15:14:52.000000000 -0400
28827 @@ -68,6 +68,7 @@
28828
28829  #include <linux/crypto.h>
28830  #include <linux/scatterlist.h>
28831 +#include <linux/vs_inet6.h>
28832
28833  /* Socket used for sending RSTs and ACKs */
28834  static struct socket *tcp6_socket;
28835 @@ -154,8 +155,15 @@ static int tcp_v6_connect(struct sock *s
28836          *      connect() to INADDR_ANY means loopback (BSD'ism).
28837          */
28838
28839 -       if(ipv6_addr_any(&usin->sin6_addr))
28840 -               usin->sin6_addr.s6_addr[15] = 0x1;
28841 +       if(ipv6_addr_any(&usin->sin6_addr)) {
28842 +               struct nx_info *nxi =  sk->sk_nx_info;
28843 +
28844 +               if (nxi && nx_info_has_v6(nxi))
28845 +                       /* FIXME: remap lback? */
28846 +                       usin->sin6_addr = nxi->v6.ip;
28847 +               else
28848 +                       usin->sin6_addr.s6_addr[15] = 0x1;
28849 +       }
28850
28851         addr_type = ipv6_addr_type(&usin->sin6_addr);
28852
28853 diff -NurpP --minimal linux-2.6.25.11/net/ipv6/udp.c linux-2.6.25.11-vs2.3.0.34.14/net/ipv6/udp.c
28854 --- linux-2.6.25.11/net/ipv6/udp.c      2008-04-17 12:05:45.000000000 -0400
28855 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv6/udp.c        2008-04-19 15:14:52.000000000 -0400
28856 @@ -49,6 +49,7 @@
28857
28858  #include <linux/proc_fs.h>
28859  #include <linux/seq_file.h>
28860 +#include <linux/vs_inet6.h>
28861  #include "udp_impl.h"
28862
28863  static inline int udp_v6_get_port(struct sock *sk, unsigned short snum)
28864 @@ -83,6 +84,10 @@ static struct sock *__udp6_lib_lookup(st
28865                                 if (!ipv6_addr_equal(&np->rcv_saddr, daddr))
28866                                         continue;
28867                                 score++;
28868 +                       } else {
28869 +                               /* block non nx_info ips */
28870 +                               if (!v6_addr_in_nx_info(sk->sk_nx_info, daddr, -1))
28871 +                                       continue;
28872                         }
28873                         if (!ipv6_addr_any(&np->daddr)) {
28874                                 if (!ipv6_addr_equal(&np->daddr, saddr))
28875 diff -NurpP --minimal linux-2.6.25.11/net/ipv6/xfrm6_policy.c linux-2.6.25.11-vs2.3.0.34.14/net/ipv6/xfrm6_policy.c
28876 --- linux-2.6.25.11/net/ipv6/xfrm6_policy.c     2008-04-17 12:05:45.000000000 -0400
28877 +++ linux-2.6.25.11-vs2.3.0.34.14/net/ipv6/xfrm6_policy.c       2008-04-19 16:13:21.000000000 -0400
28878 @@ -58,7 +58,7 @@ static int xfrm6_get_saddr(xfrm_address_
28879                 return -EHOSTUNREACH;
28880
28881         ipv6_get_saddr(dst, (struct in6_addr *)&daddr->a6,
28882 -                      (struct in6_addr *)&saddr->a6);
28883 +                      (struct in6_addr *)&saddr->a6, NULL);
28884         dst_release(dst);
28885         return 0;
28886  }
28887 diff -NurpP --minimal linux-2.6.25.11/net/netlink/af_netlink.c linux-2.6.25.11-vs2.3.0.34.14/net/netlink/af_netlink.c
28888 --- linux-2.6.25.11/net/netlink/af_netlink.c    2008-04-17 12:05:45.000000000 -0400
28889 +++ linux-2.6.25.11-vs2.3.0.34.14/net/netlink/af_netlink.c      2008-07-16 00:02:34.000000000 -0400
28890 @@ -56,6 +56,9 @@
28891  #include <linux/audit.h>
28892  #include <linux/selinux.h>
28893  #include <linux/mutex.h>
28894 +#include <linux/vs_context.h>
28895 +#include <linux/vs_network.h>
28896 +#include <linux/vs_limit.h>
28897
28898  #include <net/net_namespace.h>
28899  #include <net/sock.h>
28900 @@ -1761,6 +1764,8 @@ static struct sock *netlink_seq_socket_i
28901                         sk_for_each(s, node, &hash->table[j]) {
28902                                 if (iter->p.net != s->sk_net)
28903                                         continue;
28904 +                               if (!nx_check(s->sk_nid, VS_WATCH_P | VS_IDENT))
28905 +                                       continue;
28906                                 if (off == pos) {
28907                                         iter->link = i;
28908                                         iter->hash_idx = j;
28909 @@ -1795,7 +1800,8 @@ static void *netlink_seq_next(struct seq
28910         s = v;
28911         do {
28912                 s = sk_next(s);
28913 -       } while (s && (iter->p.net != s->sk_net));
28914 +       } while (s && (iter->p.net != s->sk_net ||
28915 +               !nx_check(s->sk_nid, VS_WATCH_P | VS_IDENT)));
28916         if (s)
28917                 return s;
28918
28919 @@ -1807,7 +1813,8 @@ static void *netlink_seq_next(struct seq
28920
28921                 for (; j <= hash->mask; j++) {
28922                         s = sk_head(&hash->table[j]);
28923 -                       while (s && (iter->p.net != s->sk_net))
28924 +                       while (s && (iter->p.net != s->sk_net ||
28925 +                               !nx_check(s->sk_nid, VS_WATCH_P | VS_IDENT)))
28926                                 s = sk_next(s);
28927                         if (s) {
28928                                 iter->link = i;
28929 diff -NurpP --minimal linux-2.6.25.11/net/sctp/ipv6.c linux-2.6.25.11-vs2.3.0.34.14/net/sctp/ipv6.c
28930 --- linux-2.6.25.11/net/sctp/ipv6.c     2008-04-17 12:05:45.000000000 -0400
28931 +++ linux-2.6.25.11-vs2.3.0.34.14/net/sctp/ipv6.c       2008-04-19 15:14:52.000000000 -0400
28932 @@ -316,7 +316,7 @@ static void sctp_v6_get_saddr(struct sct
28933                           __FUNCTION__, asoc, dst, NIP6(daddr->v6.sin6_addr));
28934
28935         if (!asoc) {
28936 -               ipv6_get_saddr(dst, &daddr->v6.sin6_addr,&saddr->v6.sin6_addr);
28937 +               ipv6_get_saddr(dst, &daddr->v6.sin6_addr,&saddr->v6.sin6_addr, asoc->base.sk->sk_nx_info);
28938                 SCTP_DEBUG_PRINTK("saddr from ipv6_get_saddr: " NIP6_FMT "\n",
28939                                   NIP6(saddr->v6.sin6_addr));
28940                 return;
28941 diff -NurpP --minimal linux-2.6.25.11/net/socket.c linux-2.6.25.11-vs2.3.0.34.14/net/socket.c
28942 --- linux-2.6.25.11/net/socket.c        2008-04-17 12:05:45.000000000 -0400
28943 +++ linux-2.6.25.11-vs2.3.0.34.14/net/socket.c  2008-04-19 15:14:52.000000000 -0400
28944 @@ -93,6 +93,10 @@
28945
28946  #include <net/sock.h>
28947  #include <linux/netfilter.h>
28948 +#include <linux/vs_base.h>
28949 +#include <linux/vs_socket.h>
28950 +#include <linux/vs_inet.h>
28951 +#include <linux/vs_inet6.h>
28952
28953  static int sock_no_open(struct inode *irrelevant, struct file *dontcare);
28954  static ssize_t sock_aio_read(struct kiocb *iocb, const struct iovec *iov,
28955 @@ -547,7 +551,7 @@ static inline int __sock_sendmsg(struct
28956                                  struct msghdr *msg, size_t size)
28957  {
28958         struct sock_iocb *si = kiocb_to_siocb(iocb);
28959 -       int err;
28960 +       int err, len;
28961
28962         si->sock = sock;
28963         si->scm = NULL;
28964 @@ -558,7 +562,22 @@ static inline int __sock_sendmsg(struct
28965         if (err)
28966                 return err;
28967
28968 -       return sock->ops->sendmsg(iocb, sock, msg, size);
28969 +       len = sock->ops->sendmsg(iocb, sock, msg, size);
28970 +       if (sock->sk) {
28971 +               if (len == size)
28972 +                       vx_sock_send(sock->sk, size);
28973 +               else
28974 +                       vx_sock_fail(sock->sk, size);
28975 +       }
28976 +       vxdprintk(VXD_CBIT(net, 7),
28977 +               "__sock_sendmsg: %p[%p,%p,%p;%d/%d]:%d/%d",
28978 +               sock, sock->sk,
28979 +               (sock->sk)?sock->sk->sk_nx_info:0,
28980 +               (sock->sk)?sock->sk->sk_vx_info:0,
28981 +               (sock->sk)?sock->sk->sk_xid:0,
28982 +               (sock->sk)?sock->sk->sk_nid:0,
28983 +               (unsigned int)size, len);
28984 +       return len;
28985  }
28986
28987  int sock_sendmsg(struct socket *sock, struct msghdr *msg, size_t size)
28988 @@ -627,7 +646,7 @@ EXPORT_SYMBOL_GPL(__sock_recv_timestamp)
28989  static inline int __sock_recvmsg(struct kiocb *iocb, struct socket *sock,
28990                                  struct msghdr *msg, size_t size, int flags)
28991  {
28992 -       int err;
28993 +       int err, len;
28994         struct sock_iocb *si = kiocb_to_siocb(iocb);
28995
28996         si->sock = sock;
28997 @@ -640,7 +659,18 @@ static inline int __sock_recvmsg(struct
28998         if (err)
28999                 return err;
29000
29001 -       return sock->ops->recvmsg(iocb, sock, msg, size, flags);
29002 +       len = sock->ops->recvmsg(iocb, sock, msg, size, flags);
29003 +       if ((len >= 0) && sock->sk)
29004 +               vx_sock_recv(sock->sk, len);
29005 +       vxdprintk(VXD_CBIT(net, 7),
29006 +               "__sock_recvmsg: %p[%p,%p,%p;%d/%d]:%d/%d",
29007 +               sock, sock->sk,
29008 +               (sock->sk)?sock->sk->sk_nx_info:0,
29009 +               (sock->sk)?sock->sk->sk_vx_info:0,
29010 +               (sock->sk)?sock->sk->sk_xid:0,
29011 +               (sock->sk)?sock->sk->sk_nid:0,
29012 +               (unsigned int)size, len);
29013 +       return len;
29014  }
29015
29016  int sock_recvmsg(struct socket *sock, struct msghdr *msg,
29017 @@ -1105,6 +1135,13 @@ static int __sock_create(struct net *net
29018         if (type < 0 || type >= SOCK_MAX)
29019                 return -EINVAL;
29020
29021 +       if (!nx_check(0, VS_ADMIN)) {
29022 +               if (family == PF_INET && !current_nx_info_has_v4())
29023 +                       return -EAFNOSUPPORT;
29024 +               if (family == PF_INET6 && !current_nx_info_has_v6())
29025 +                       return -EAFNOSUPPORT;
29026 +       }
29027 +
29028         /* Compatibility.
29029
29030            This uglymoron is moved from INET layer to here to avoid
29031 @@ -1222,6 +1259,7 @@ asmlinkage long sys_socket(int family, i
29032         if (retval < 0)
29033                 goto out;
29034
29035 +       set_bit(SOCK_USER_SOCKET, &sock->flags);
29036         retval = sock_map_fd(sock);
29037         if (retval < 0)
29038                 goto out_release;
29039 @@ -1254,10 +1292,12 @@ asmlinkage long sys_socketpair(int famil
29040         err = sock_create(family, type, protocol, &sock1);
29041         if (err < 0)
29042                 goto out;
29043 +       set_bit(SOCK_USER_SOCKET, &sock1->flags);
29044
29045         err = sock_create(family, type, protocol, &sock2);
29046         if (err < 0)
29047                 goto out_release_1;
29048 +       set_bit(SOCK_USER_SOCKET, &sock2->flags);
29049
29050         err = sock1->ops->socketpair(sock1, sock2);
29051         if (err < 0)
29052 diff -NurpP --minimal linux-2.6.25.11/net/sunrpc/auth.c linux-2.6.25.11-vs2.3.0.34.14/net/sunrpc/auth.c
29053 --- linux-2.6.25.11/net/sunrpc/auth.c   2008-04-17 12:05:45.000000000 -0400
29054 +++ linux-2.6.25.11-vs2.3.0.34.14/net/sunrpc/auth.c     2008-04-19 15:14:52.000000000 -0400
29055 @@ -13,6 +13,7 @@
29056  #include <linux/errno.h>
29057  #include <linux/sunrpc/clnt.h>
29058  #include <linux/spinlock.h>
29059 +#include <linux/vs_tag.h>
29060
29061  #ifdef RPC_DEBUG
29062  # define RPCDBG_FACILITY       RPCDBG_AUTH
29063 @@ -345,6 +346,7 @@ rpcauth_lookupcred(struct rpc_auth *auth
29064         struct auth_cred acred = {
29065                 .uid = current->fsuid,
29066                 .gid = current->fsgid,
29067 +               .tag = dx_current_tag(),
29068                 .group_info = current->group_info,
29069         };
29070         struct rpc_cred *ret;
29071 @@ -382,6 +384,7 @@ rpcauth_bindcred(struct rpc_task *task)
29072         struct auth_cred acred = {
29073                 .uid = current->fsuid,
29074                 .gid = current->fsgid,
29075 +               .tag = dx_current_tag(),
29076                 .group_info = current->group_info,
29077         };
29078         struct rpc_cred *ret;
29079 diff -NurpP --minimal linux-2.6.25.11/net/sunrpc/auth_unix.c linux-2.6.25.11-vs2.3.0.34.14/net/sunrpc/auth_unix.c
29080 --- linux-2.6.25.11/net/sunrpc/auth_unix.c      2008-04-17 10:37:27.000000000 -0400
29081 +++ linux-2.6.25.11-vs2.3.0.34.14/net/sunrpc/auth_unix.c        2008-04-19 15:14:52.000000000 -0400
29082 @@ -11,12 +11,14 @@
29083  #include <linux/module.h>
29084  #include <linux/sunrpc/clnt.h>
29085  #include <linux/sunrpc/auth.h>
29086 +#include <linux/vs_tag.h>
29087
29088  #define NFS_NGROUPS    16
29089
29090  struct unx_cred {
29091         struct rpc_cred         uc_base;
29092         gid_t                   uc_gid;
29093 +       tag_t                   uc_tag;
29094         gid_t                   uc_gids[NFS_NGROUPS];
29095  };
29096  #define uc_uid                 uc_base.cr_uid
29097 @@ -73,6 +75,7 @@ unx_create_cred(struct rpc_auth *auth, s
29098         if (flags & RPCAUTH_LOOKUP_ROOTCREDS) {
29099                 cred->uc_uid = 0;
29100                 cred->uc_gid = 0;
29101 +               cred->uc_tag = dx_current_tag();
29102                 cred->uc_gids[0] = NOGROUP;
29103         } else {
29104                 int groups = acred->group_info->ngroups;
29105 @@ -80,6 +83,7 @@ unx_create_cred(struct rpc_auth *auth, s
29106                         groups = NFS_NGROUPS;
29107
29108                 cred->uc_gid = acred->gid;
29109 +               cred->uc_tag = acred->tag;
29110                 for (i = 0; i < groups; i++)
29111                         cred->uc_gids[i] = GROUP_AT(acred->group_info, i);
29112                 if (i < NFS_NGROUPS)
29113 @@ -124,7 +128,8 @@ unx_match(struct auth_cred *acred, struc
29114                 int groups;
29115
29116                 if (cred->uc_uid != acred->uid
29117 -                || cred->uc_gid != acred->gid)
29118 +                || cred->uc_gid != acred->gid
29119 +                || cred->uc_tag != acred->tag)
29120                         return 0;
29121
29122                 groups = acred->group_info->ngroups;
29123 @@ -150,7 +155,7 @@ unx_marshal(struct rpc_task *task, __be3
29124         struct rpc_clnt *clnt = task->tk_client;
29125         struct unx_cred *cred = container_of(task->tk_msg.rpc_cred, struct unx_cred, uc_base);
29126         __be32          *base, *hold;
29127 -       int             i;
29128 +       int             i, tag;
29129
29130         *p++ = htonl(RPC_AUTH_UNIX);
29131         base = p++;
29132 @@ -160,9 +165,12 @@ unx_marshal(struct rpc_task *task, __be3
29133          * Copy the UTS nodename captured when the client was created.
29134          */
29135         p = xdr_encode_array(p, clnt->cl_nodename, clnt->cl_nodelen);
29136 +       tag = task->tk_client->cl_tag;
29137
29138 -       *p++ = htonl((u32) cred->uc_uid);
29139 -       *p++ = htonl((u32) cred->uc_gid);
29140 +       *p++ = htonl((u32) TAGINO_UID(tag,
29141 +               cred->uc_uid, cred->uc_tag));
29142 +       *p++ = htonl((u32) TAGINO_GID(tag,
29143 +               cred->uc_gid, cred->uc_tag));
29144         hold = p++;
29145         for (i = 0; i < 16 && cred->uc_gids[i] != (gid_t) NOGROUP; i++)
29146                 *p++ = htonl((u32) cred->uc_gids[i]);
29147 diff -NurpP --minimal linux-2.6.25.11/net/sunrpc/clnt.c linux-2.6.25.11-vs2.3.0.34.14/net/sunrpc/clnt.c
29148 --- linux-2.6.25.11/net/sunrpc/clnt.c   2008-04-17 12:05:45.000000000 -0400
29149 +++ linux-2.6.25.11-vs2.3.0.34.14/net/sunrpc/clnt.c     2008-04-19 16:09:36.000000000 -0400
29150 @@ -31,6 +31,7 @@
29151  #include <linux/utsname.h>
29152  #include <linux/workqueue.h>
29153  #include <linux/in6.h>
29154 +#include <linux/vs_cvirt.h>
29155
29156  #include <linux/sunrpc/clnt.h>
29157  #include <linux/sunrpc/rpc_pipe_fs.h>
29158 @@ -324,7 +325,9 @@ struct rpc_clnt *rpc_create(struct rpc_c
29159                 clnt->cl_autobind = 1;
29160         if (args->flags & RPC_CLNT_CREATE_DISCRTRY)
29161                 clnt->cl_discrtry = 1;
29162 -
29163 +       /* TODO: handle RPC_CLNT_CREATE_TAGGED
29164 +       if (args->flags & RPC_CLNT_CREATE_TAGGED)
29165 +               clnt->cl_tag = 1; */
29166         return clnt;
29167  }
29168  EXPORT_SYMBOL_GPL(rpc_create);
29169 diff -NurpP --minimal linux-2.6.25.11/net/unix/af_unix.c linux-2.6.25.11-vs2.3.0.34.14/net/unix/af_unix.c
29170 --- linux-2.6.25.11/net/unix/af_unix.c  2008-04-17 12:05:45.000000000 -0400
29171 +++ linux-2.6.25.11-vs2.3.0.34.14/net/unix/af_unix.c    2008-07-16 00:03:11.000000000 -0400
29172 @@ -116,6 +116,8 @@
29173  #include <linux/mount.h>
29174  #include <net/checksum.h>
29175  #include <linux/security.h>
29176 +#include <linux/vs_context.h>
29177 +#include <linux/vs_limit.h>
29178
29179  static struct hlist_head unix_socket_table[UNIX_HASH_SIZE + 1];
29180  static DEFINE_SPINLOCK(unix_table_lock);
29181 @@ -255,6 +257,8 @@ static struct sock *__unix_find_socket_b
29182                 if (s->sk_net != net)
29183                         continue;
29184
29185 +               if (!nx_check(s->sk_nid, VS_WATCH_P | VS_IDENT))
29186 +                       continue;
29187                 if (u->addr->len == len &&
29188                     !memcmp(u->addr->name, sunname, len))
29189                         goto found;
29190 @@ -819,7 +823,7 @@ static int unix_bind(struct socket *sock
29191                  */
29192                 mode = S_IFSOCK |
29193                        (SOCK_INODE(sock)->i_mode & ~current->fs->umask);
29194 -               err = vfs_mknod(nd.path.dentry->d_inode, dentry, mode, 0);
29195 +               err = vfs_mknod(nd.path.dentry->d_inode, dentry, mode, 0, NULL);
29196                 if (err)
29197                         goto out_mknod_dput;
29198                 mutex_unlock(&nd.path.dentry->d_inode->i_mutex);
29199 @@ -2024,6 +2028,8 @@ static struct sock *unix_seq_idx(struct
29200         for (s = first_unix_socket(&iter->i); s; s = next_unix_socket(&iter->i, s)) {
29201                 if (s->sk_net != iter->p.net)
29202                         continue;
29203 +               if (!nx_check(s->sk_nid, VS_WATCH_P | VS_IDENT))
29204 +                       continue;
29205                 if (off == pos)
29206                         return s;
29207                 ++off;
29208 @@ -2050,7 +2056,8 @@ static void *unix_seq_next(struct seq_fi
29209                 sk = first_unix_socket(&iter->i);
29210         else
29211                 sk = next_unix_socket(&iter->i, sk);
29212 -       while (sk && (sk->sk_net != iter->p.net))
29213 +       while (sk && (sk->sk_net != iter->p.net ||
29214 +               !nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT)))
29215                 sk = next_unix_socket(&iter->i, sk);
29216         return sk;
29217  }
29218 diff -NurpP --minimal linux-2.6.25.11/net/x25/af_x25.c linux-2.6.25.11-vs2.3.0.34.14/net/x25/af_x25.c
29219 --- linux-2.6.25.11/net/x25/af_x25.c    2008-04-17 12:05:45.000000000 -0400
29220 +++ linux-2.6.25.11-vs2.3.0.34.14/net/x25/af_x25.c      2008-04-19 15:14:52.000000000 -0400
29221 @@ -506,7 +506,10 @@ static int x25_create(struct net *net, s
29222
29223         x25 = x25_sk(sk);
29224
29225 -       sock_init_data(sock, sk);
29226 +       sk->sk_socket = sock;
29227 +       sk->sk_type = sock->type;
29228 +       sk->sk_sleep = &sock->wait;
29229 +       sock->sk = sk;
29230
29231         x25_init_timers(sk);
29232
29233 diff -NurpP --minimal linux-2.6.25.11/scripts/checksyscalls.sh linux-2.6.25.11-vs2.3.0.34.14/scripts/checksyscalls.sh
29234 --- linux-2.6.25.11/scripts/checksyscalls.sh    2008-04-17 11:31:42.000000000 -0400
29235 +++ linux-2.6.25.11-vs2.3.0.34.14/scripts/checksyscalls.sh      2008-04-19 15:14:52.000000000 -0400
29236 @@ -108,7 +108,6 @@ cat << EOF
29237  #define __IGNORE_afs_syscall
29238  #define __IGNORE_getpmsg
29239  #define __IGNORE_putpmsg
29240 -#define __IGNORE_vserver
29241  EOF
29242  }
29243
29244 diff -NurpP --minimal linux-2.6.25.11/security/commoncap.c linux-2.6.25.11-vs2.3.0.34.14/security/commoncap.c
29245 --- linux-2.6.25.11/security/commoncap.c        2008-04-17 12:05:46.000000000 -0400
29246 +++ linux-2.6.25.11-vs2.3.0.34.14/security/commoncap.c  2008-04-23 22:22:54.000000000 -0400
29247 @@ -24,6 +24,7 @@
29248  #include <linux/hugetlb.h>
29249  #include <linux/mount.h>
29250  #include <linux/sched.h>
29251 +#include <linux/vs_context.h>
29252
29253  /* Global security state */
29254
29255 @@ -32,7 +33,7 @@ EXPORT_SYMBOL(securebits);
29256
29257  int cap_netlink_send(struct sock *sk, struct sk_buff *skb)
29258  {
29259 -       NETLINK_CB(skb).eff_cap = current->cap_effective;
29260 +       NETLINK_CB(skb).eff_cap = vx_mbcaps(current->cap_effective);
29261         return 0;
29262  }
29263
29264 @@ -53,9 +54,24 @@ EXPORT_SYMBOL(cap_netlink_recv);
29265   */
29266  int cap_capable (struct task_struct *tsk, int cap)
29267  {
29268 +       struct vx_info *vxi = tsk->vx_info;
29269 +
29270 +#if 0
29271 +       printk("cap_capable() VXF_STATE_SETUP = %llx, raised = %x, eff = %08x:%08x\n",
29272 +               vx_info_flags(vxi, VXF_STATE_SETUP, 0),
29273 +               cap_raised(tsk->cap_effective, cap),
29274 +               tsk->cap_effective.cap[1], tsk->cap_effective.cap[0]);
29275 +#endif
29276 +
29277 +       /* special case SETUP */
29278 +       if (vx_info_flags(vxi, VXF_STATE_SETUP, 0) &&
29279 +               cap_raised(tsk->cap_effective, cap))
29280 +               return 0;
29281 +
29282         /* Derived from include/linux/sched.h:capable. */
29283 -       if (cap_raised(tsk->cap_effective, cap))
29284 +       if (vx_cap_raised(vxi, tsk->cap_effective, cap))
29285                 return 0;
29286 +
29287         return -EPERM;
29288  }
29289
29290 @@ -583,7 +599,8 @@ void cap_task_reparent_to_init (struct t
29291
29292  int cap_syslog (int type)
29293  {
29294 -       if ((type != 3 && type != 10) && !capable(CAP_SYS_ADMIN))
29295 +       if ((type != 3 && type != 10) &&
29296 +               !vx_capable(CAP_SYS_ADMIN, VXC_SYSLOG))
29297                 return -EPERM;
29298         return 0;
29299  }
29300 diff -NurpP --minimal linux-2.6.25.11/security/dummy.c linux-2.6.25.11-vs2.3.0.34.14/security/dummy.c
29301 --- linux-2.6.25.11/security/dummy.c    2008-04-17 12:05:46.000000000 -0400
29302 +++ linux-2.6.25.11-vs2.3.0.34.14/security/dummy.c      2008-04-23 20:28:54.000000000 -0400
29303 @@ -27,6 +27,7 @@
29304  #include <linux/hugetlb.h>
29305  #include <linux/ptrace.h>
29306  #include <linux/file.h>
29307 +#include <linux/vs_context.h>
29308
29309  static int dummy_ptrace (struct task_struct *parent, struct task_struct *child)
29310  {
29311 @@ -714,7 +715,7 @@ static int dummy_sem_semop (struct sem_a
29312
29313  static int dummy_netlink_send (struct sock *sk, struct sk_buff *skb)
29314  {
29315 -       NETLINK_CB(skb).eff_cap = current->cap_effective;
29316 +       NETLINK_CB(skb).eff_cap = vx_mbcaps(current->cap_effective);
29317         return 0;
29318  }
29319
29320 diff -NurpP --minimal linux-2.6.25.11/security/selinux/hooks.c linux-2.6.25.11-vs2.3.0.34.14/security/selinux/hooks.c
29321 --- linux-2.6.25.11/security/selinux/hooks.c    2008-07-15 22:05:18.000000000 -0400
29322 +++ linux-2.6.25.11-vs2.3.0.34.14/security/selinux/hooks.c      2008-05-21 14:30:41.000000000 -0400
29323 @@ -64,7 +64,6 @@
29324  #include <linux/dccp.h>
29325  #include <linux/quota.h>
29326  #include <linux/un.h>          /* for Unix socket types */
29327 -#include <net/af_unix.h>       /* for Unix socket types */
29328  #include <linux/parser.h>
29329  #include <linux/nfs_mount.h>
29330  #include <net/ipv6.h>