kernel-vserver-2.3.patch

   1 diff -NurpP --minimal linux-3.0-rc5/Documentation/vserver/debug.txt linux-3.0-rc5-vs2.3.1-pre3/Documentation/vserver/debug.txt
   2 --- linux-3.0-rc5/Documentation/vserver/debug.txt       1970-01-01 01:00:00.000000000 +0100
   3 +++ linux-3.0-rc5-vs2.3.1-pre3/Documentation/vserver/debug.txt  2011-06-10 22:11:24.000000000 +0200
   4 @@ -0,0 +1,154 @@
   5 +
   6 +debug_cvirt:
   7 +
   8 + 2   4 "vx_map_tgid: %p/%llx: %d -> %d"
   9 +       "vx_rmap_tgid: %p/%llx: %d -> %d"
  10 +
  11 +debug_dlim:
  12 +
  13 + 0   1 "ALLOC (%p,#%d)%c inode (%d)"
  14 +       "FREE  (%p,#%d)%c inode"
  15 + 1   2 "ALLOC (%p,#%d)%c %lld bytes (%d)"
  16 +       "FREE  (%p,#%d)%c %lld bytes"
  17 + 2   4 "ADJUST: %lld,%lld on %ld,%ld [mult=%d]"
  18 + 3   8 "ext3_has_free_blocks(%p): %lu<%lu+1, %c, %u!=%u r=%d"
  19 +       "ext3_has_free_blocks(%p): free=%lu, root=%lu"
  20 +       "rcu_free_dl_info(%p)"
  21 + 4  10 "alloc_dl_info(%p,%d) = %p"
  22 +       "dealloc_dl_info(%p)"
  23 +       "get_dl_info(%p[#%d.%d])"
  24 +       "put_dl_info(%p[#%d.%d])"
  25 + 5  20 "alloc_dl_info(%p,%d)*"
  26 + 6  40 "__hash_dl_info: %p[#%d]"
  27 +       "__unhash_dl_info: %p[#%d]"
  28 + 7  80 "locate_dl_info(%p,#%d) = %p"
  29 +
  30 +debug_misc:
  31 +
  32 + 0   1 "destroy_dqhash: %p [#0x%08x] c=%d"
  33 +       "new_dqhash: %p [#0x%08x]"
  34 +       "vroot[%d]_clr_dev: dev=%p[%lu,%d:%d]"
  35 +       "vroot[%d]_get_real_bdev: dev=%p[%lu,%d:%d]"
  36 +       "vroot[%d]_set_dev: dev=%p[%lu,%d:%d]"
  37 +       "vroot_get_real_bdev not set"
  38 + 1   2 "cow_break_link(»%s«)"
  39 +       "temp copy »%s«"
  40 + 2   4 "dentry_open(new): %p"
  41 +       "dentry_open(old): %p"
  42 +       "lookup_create(new): %p"
  43 +       "old path »%s«"
  44 +       "path_lookup(old): %d"
  45 +       "vfs_create(new): %d"
  46 +       "vfs_rename: %d"
  47 +       "vfs_sendfile: %d"
  48 + 3   8 "fput(new_file=%p[#%d])"
  49 +       "fput(old_file=%p[#%d])"
  50 + 4  10 "vx_info_kill(%p[#%d],%d,%d) = %d"
  51 +       "vx_info_kill(%p[#%d],%d,%d)*"
  52 + 5  20 "vs_reboot(%p[#%d],%d)"
  53 + 6  40 "dropping task %p[#%u,%u] for %p[#%u,%u]"
  54 +
  55 +debug_net:
  56 +
  57 + 2   4 "nx_addr_conflict(%p,%p) %d.%d,%d.%d"
  58 + 3   8 "inet_bind(%p) %d.%d.%d.%d, %d.%d.%d.%d, %d.%d.%d.%d"
  59 +       "inet_bind(%p)* %p,%p;%lx %d.%d.%d.%d"
  60 + 4  10 "ip_route_connect(%p) %p,%p;%lx"
  61 + 5  20 "__addr_in_socket(%p,%d.%d.%d.%d) %p:%d.%d.%d.%d %p;%lx"
  62 + 6  40 "sk,egf: %p [#%d] (from %d)"
  63 +       "sk,egn: %p [#%d] (from %d)"
  64 +       "sk,req: %p [#%d] (from %d)"
  65 +       "sk: %p [#%d] (from %d)"
  66 +       "tw: %p [#%d] (from %d)"
  67 + 7  80 "__sock_recvmsg: %p[%p,%p,%p;%d]:%d/%d"
  68 +       "__sock_sendmsg: %p[%p,%p,%p;%d]:%d/%d"
  69 +
  70 +debug_nid:
  71 +
  72 + 0   1 "__lookup_nx_info(#%u): %p[#%u]"
  73 +       "alloc_nx_info(%d) = %p"
  74 +       "create_nx_info(%d) (dynamic rejected)"
  75 +       "create_nx_info(%d) = %p (already there)"
  76 +       "create_nx_info(%d) = %p (new)"
  77 +       "dealloc_nx_info(%p)"
  78 + 1   2 "alloc_nx_info(%d)*"
  79 +       "create_nx_info(%d)*"
  80 + 2   4 "get_nx_info(%p[#%d.%d])"
  81 +       "put_nx_info(%p[#%d.%d])"
  82 + 3   8 "claim_nx_info(%p[#%d.%d.%d]) %p"
  83 +       "clr_nx_info(%p[#%d.%d])"
  84 +       "init_nx_info(%p[#%d.%d])"
  85 +       "release_nx_info(%p[#%d.%d.%d]) %p"
  86 +       "set_nx_info(%p[#%d.%d])"
  87 + 4  10 "__hash_nx_info: %p[#%d]"
  88 +       "__nx_dynamic_id: [#%d]"
  89 +       "__unhash_nx_info: %p[#%d.%d.%d]"
  90 + 5  20 "moved task %p into nxi:%p[#%d]"
  91 +       "nx_migrate_task(%p,%p[#%d.%d.%d])"
  92 +       "task_get_nx_info(%p)"
  93 + 6  40 "nx_clear_persistent(%p[#%d])"
  94 +
  95 +debug_quota:
  96 +
  97 + 0   1 "quota_sync_dqh(%p,%d) discard inode %p"
  98 + 1   2 "quota_sync_dqh(%p,%d)"
  99 +       "sync_dquots(%p,%d)"
 100 +       "sync_dquots_dqh(%p,%d)"
 101 + 3   8 "do_quotactl(%p,%d,cmd=%d,id=%d,%p)"
 102 +
 103 +debug_switch:
 104 +
 105 + 0   1 "vc: VCMD_%02d_%d[%d], %d,%p [%d,%d,%x,%x]"
 106 + 1   2 "vc: VCMD_%02d_%d[%d] = %08lx(%ld) [%d,%d]"
 107 + 4  10 "%s: (%s %s) returned %s with %d"
 108 +
 109 +debug_tag:
 110 +
 111 + 7  80 "dx_parse_tag(»%s«): %d:#%d"
 112 +       "dx_propagate_tag(%p[#%lu.%d]): %d,%d"
 113 +
 114 +debug_xid:
 115 +
 116 + 0   1 "__lookup_vx_info(#%u): %p[#%u]"
 117 +       "alloc_vx_info(%d) = %p"
 118 +       "alloc_vx_info(%d)*"
 119 +       "create_vx_info(%d) (dynamic rejected)"
 120 +       "create_vx_info(%d) = %p (already there)"
 121 +       "create_vx_info(%d) = %p (new)"
 122 +       "dealloc_vx_info(%p)"
 123 +       "loc_vx_info(%d) = %p (found)"
 124 +       "loc_vx_info(%d) = %p (new)"
 125 +       "loc_vx_info(%d) = %p (not available)"
 126 + 1   2 "create_vx_info(%d)*"
 127 +       "loc_vx_info(%d)*"
 128 + 2   4 "get_vx_info(%p[#%d.%d])"
 129 +       "put_vx_info(%p[#%d.%d])"
 130 + 3   8 "claim_vx_info(%p[#%d.%d.%d]) %p"
 131 +       "clr_vx_info(%p[#%d.%d])"
 132 +       "init_vx_info(%p[#%d.%d])"
 133 +       "release_vx_info(%p[#%d.%d.%d]) %p"
 134 +       "set_vx_info(%p[#%d.%d])"
 135 + 4  10 "__hash_vx_info: %p[#%d]"
 136 +       "__unhash_vx_info: %p[#%d.%d.%d]"
 137 +       "__vx_dynamic_id: [#%d]"
 138 + 5  20 "enter_vx_info(%p[#%d],%p) %p[#%d,%p]"
 139 +       "leave_vx_info(%p[#%d,%p]) %p[#%d,%p]"
 140 +       "moved task %p into vxi:%p[#%d]"
 141 +       "task_get_vx_info(%p)"
 142 +       "vx_migrate_task(%p,%p[#%d.%d])"
 143 + 6  40 "vx_clear_persistent(%p[#%d])"
 144 +       "vx_exit_init(%p[#%d],%p[#%d,%d,%d])"
 145 +       "vx_set_init(%p[#%d],%p[#%d,%d,%d])"
 146 +       "vx_set_persistent(%p[#%d])"
 147 +       "vx_set_reaper(%p[#%d],%p[#%d,%d])"
 148 + 7  80 "vx_child_reaper(%p[#%u,%u]) = %p[#%u,%u]"
 149 +
 150 +
 151 +debug_limit:
 152 +
 153 + n 2^n "vx_acc_cres[%5d,%s,%2d]: %5d%s"
 154 +       "vx_cres_avail[%5d,%s,%2d]: %5ld > %5d + %5d"
 155 +
 156 + m 2^m "vx_acc_page[%5d,%s,%2d]: %5d%s"
 157 +       "vx_acc_pages[%5d,%s,%2d]: %5d += %5d"
 158 +       "vx_pages_avail[%5d,%s,%2d]: %5ld > %5d + %5d"
 159 diff -NurpP --minimal linux-3.0-rc5/arch/alpha/Kconfig linux-3.0-rc5-vs2.3.1-pre3/arch/alpha/Kconfig
 160 --- linux-3.0-rc5/arch/alpha/Kconfig    2011-07-01 11:13:32.000000000 +0200
 161 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/alpha/Kconfig       2011-06-10 22:11:24.000000000 +0200
 162 @@ -668,6 +668,8 @@ config DUMMY_CONSOLE
 163         depends on VGA_HOSE
 164         default y
 165
 166 +source "kernel/vserver/Kconfig"
 167 +
 168  source "security/Kconfig"
 169
 170  source "crypto/Kconfig"
 171 diff -NurpP --minimal linux-3.0-rc5/arch/alpha/kernel/entry.S linux-3.0-rc5-vs2.3.1-pre3/arch/alpha/kernel/entry.S
 172 --- linux-3.0-rc5/arch/alpha/kernel/entry.S     2010-10-21 13:06:45.000000000 +0200
 173 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/alpha/kernel/entry.S        2011-06-10 22:11:24.000000000 +0200
 174 @@ -860,24 +860,15 @@ sys_getxgid:
 175         .globl  sys_getxpid
 176         .ent    sys_getxpid
 177  sys_getxpid:
 178 +       lda     $sp, -16($sp)
 179 +       stq     $26, 0($sp)
 180         .prologue 0
 181 -       ldq     $2, TI_TASK($8)
 182
 183 -       /* See linux/kernel/timer.c sys_getppid for discussion
 184 -          about this loop.  */
 185 -       ldq     $3, TASK_GROUP_LEADER($2)
 186 -       ldq     $4, TASK_REAL_PARENT($3)
 187 -       ldl     $0, TASK_TGID($2)
 188 -1:     ldl     $1, TASK_TGID($4)
 189 -#ifdef CONFIG_SMP
 190 -       mov     $4, $5
 191 -       mb
 192 -       ldq     $3, TASK_GROUP_LEADER($2)
 193 -       ldq     $4, TASK_REAL_PARENT($3)
 194 -       cmpeq   $4, $5, $5
 195 -       beq     $5, 1b
 196 -#endif
 197 -       stq     $1, 80($sp)
 198 +       lda     $16, 96($sp)
 199 +       jsr     $26, do_getxpid
 200 +       ldq     $26, 0($sp)
 201 +
 202 +       lda     $sp, 16($sp)
 203         ret
 204  .end sys_getxpid
 205
 206 diff -NurpP --minimal linux-3.0-rc5/arch/alpha/kernel/ptrace.c linux-3.0-rc5-vs2.3.1-pre3/arch/alpha/kernel/ptrace.c
 207 --- linux-3.0-rc5/arch/alpha/kernel/ptrace.c    2011-01-05 21:48:40.000000000 +0100
 208 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/alpha/kernel/ptrace.c       2011-06-10 22:11:24.000000000 +0200
 209 @@ -13,6 +13,7 @@
 210  #include <linux/user.h>
 211  #include <linux/security.h>
 212  #include <linux/signal.h>
 213 +#include <linux/vs_base.h>
 214
 215  #include <asm/uaccess.h>
 216  #include <asm/pgtable.h>
 217 diff -NurpP --minimal linux-3.0-rc5/arch/alpha/kernel/systbls.S linux-3.0-rc5-vs2.3.1-pre3/arch/alpha/kernel/systbls.S
 218 --- linux-3.0-rc5/arch/alpha/kernel/systbls.S   2011-07-01 11:13:32.000000000 +0200
 219 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/alpha/kernel/systbls.S      2011-06-10 22:11:24.000000000 +0200
 220 @@ -446,7 +446,7 @@ sys_call_table:
 221         .quad sys_stat64                        /* 425 */
 222         .quad sys_lstat64
 223         .quad sys_fstat64
 224 -       .quad sys_ni_syscall                    /* sys_vserver */
 225 +       .quad sys_vserver                       /* sys_vserver */
 226         .quad sys_ni_syscall                    /* sys_mbind */
 227         .quad sys_ni_syscall                    /* sys_get_mempolicy */
 228         .quad sys_ni_syscall                    /* sys_set_mempolicy */
 229 diff -NurpP --minimal linux-3.0-rc5/arch/alpha/kernel/traps.c linux-3.0-rc5-vs2.3.1-pre3/arch/alpha/kernel/traps.c
 230 --- linux-3.0-rc5/arch/alpha/kernel/traps.c     2010-10-21 13:06:46.000000000 +0200
 231 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/alpha/kernel/traps.c        2011-06-10 22:11:24.000000000 +0200
 232 @@ -183,7 +183,8 @@ die_if_kernel(char * str, struct pt_regs
 233  #ifdef CONFIG_SMP
 234         printk("CPU %d ", hard_smp_processor_id());
 235  #endif
 236 -       printk("%s(%d): %s %ld\n", current->comm, task_pid_nr(current), str, err);
 237 +       printk("%s(%d[#%u]): %s %ld\n", current->comm,
 238 +               task_pid_nr(current), current->xid, str, err);
 239         dik_show_regs(regs, r9_15);
 240         add_taint(TAINT_DIE);
 241         dik_show_trace((unsigned long *)(regs+1));
 242 diff -NurpP --minimal linux-3.0-rc5/arch/arm/Kconfig linux-3.0-rc5-vs2.3.1-pre3/arch/arm/Kconfig
 243 --- linux-3.0-rc5/arch/arm/Kconfig      2011-07-01 11:13:32.000000000 +0200
 244 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/arm/Kconfig 2011-06-10 22:11:24.000000000 +0200
 245 @@ -2035,6 +2035,8 @@ source "fs/Kconfig"
 246
 247  source "arch/arm/Kconfig.debug"
 248
 249 +source "kernel/vserver/Kconfig"
 250 +
 251  source "security/Kconfig"
 252
 253  source "crypto/Kconfig"
 254 diff -NurpP --minimal linux-3.0-rc5/arch/arm/kernel/calls.S linux-3.0-rc5-vs2.3.1-pre3/arch/arm/kernel/calls.S
 255 --- linux-3.0-rc5/arch/arm/kernel/calls.S       2011-07-01 11:13:32.000000000 +0200
 256 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/arm/kernel/calls.S  2011-06-10 22:11:24.000000000 +0200
 257 @@ -322,7 +322,7 @@
 258  /* 310 */      CALL(sys_request_key)
 259                 CALL(sys_keyctl)
 260                 CALL(ABI(sys_semtimedop, sys_oabi_semtimedop))
 261 -/* vserver */  CALL(sys_ni_syscall)
 262 +               CALL(sys_vserver)
 263                 CALL(sys_ioprio_set)
 264  /* 315 */      CALL(sys_ioprio_get)
 265                 CALL(sys_inotify_init)
 266 diff -NurpP --minimal linux-3.0-rc5/arch/arm/kernel/process.c linux-3.0-rc5-vs2.3.1-pre3/arch/arm/kernel/process.c
 267 --- linux-3.0-rc5/arch/arm/kernel/process.c     2011-05-22 16:16:47.000000000 +0200
 268 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/arm/kernel/process.c        2011-06-10 22:11:24.000000000 +0200
 269 @@ -315,7 +315,8 @@ void __show_regs(struct pt_regs *regs)
 270  void show_regs(struct pt_regs * regs)
 271  {
 272         printk("\n");
 273 -       printk("Pid: %d, comm: %20s\n", task_pid_nr(current), current->comm);
 274 +       printk("Pid: %d[#%u], comm: %20s\n",
 275 +               task_pid_nr(current), current->xid, current->comm);
 276         __show_regs(regs);
 277         __backtrace();
 278  }
 279 diff -NurpP --minimal linux-3.0-rc5/arch/arm/kernel/traps.c linux-3.0-rc5-vs2.3.1-pre3/arch/arm/kernel/traps.c
 280 --- linux-3.0-rc5/arch/arm/kernel/traps.c       2011-07-01 11:13:32.000000000 +0200
 281 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/arm/kernel/traps.c  2011-06-22 12:39:12.000000000 +0200
 282 @@ -242,8 +242,8 @@ static int __die(const char *str, int er
 283
 284         print_modules();
 285         __show_regs(regs);
 286 -       printk(KERN_EMERG "Process %.*s (pid: %d, stack limit = 0x%p)\n",
 287 -               TASK_COMM_LEN, tsk->comm, task_pid_nr(tsk), thread + 1);
 288 +       printk(KERN_EMERG "Process %.*s (pid: %d:#%u, stack limit = 0x%p)\n",
 289 +               TASK_COMM_LEN, tsk->comm, task_pid_nr(tsk), tsk->xid, thread + 1);
 290
 291         if (!user_mode(regs) || in_interrupt()) {
 292                 dump_mem(KERN_EMERG, "Stack: ", regs->ARM_sp,
 293 diff -NurpP --minimal linux-3.0-rc5/arch/cris/Kconfig linux-3.0-rc5-vs2.3.1-pre3/arch/cris/Kconfig
 294 --- linux-3.0-rc5/arch/cris/Kconfig     2011-07-01 11:13:37.000000000 +0200
 295 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/cris/Kconfig        2011-06-10 22:11:24.000000000 +0200
 296 @@ -678,6 +678,8 @@ source "drivers/staging/Kconfig"
 297
 298  source "arch/cris/Kconfig.debug"
 299
 300 +source "kernel/vserver/Kconfig"
 301 +
 302  source "security/Kconfig"
 303
 304  source "crypto/Kconfig"
 305 diff -NurpP --minimal linux-3.0-rc5/arch/frv/kernel/kernel_thread.S linux-3.0-rc5-vs2.3.1-pre3/arch/frv/kernel/kernel_thread.S
 306 --- linux-3.0-rc5/arch/frv/kernel/kernel_thread.S       2008-12-25 00:26:37.000000000 +0100
 307 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/frv/kernel/kernel_thread.S  2011-06-10 22:11:24.000000000 +0200
 308 @@ -37,7 +37,7 @@ kernel_thread:
 309
 310         # start by forking the current process, but with shared VM
 311         setlos.p        #__NR_clone,gr7         ; syscall number
 312 -       ori             gr10,#CLONE_VM,gr8      ; first syscall arg     [clone_flags]
 313 +       ori             gr10,#CLONE_KT,gr8      ; first syscall arg     [clone_flags]
 314         sethi.p         #0xe4e4,gr9             ; second syscall arg    [newsp]
 315         setlo           #0xe4e4,gr9
 316         setlos.p        #0,gr10                 ; third syscall arg     [parent_tidptr]
 317 diff -NurpP --minimal linux-3.0-rc5/arch/h8300/Kconfig linux-3.0-rc5-vs2.3.1-pre3/arch/h8300/Kconfig
 318 --- linux-3.0-rc5/arch/h8300/Kconfig    2011-07-01 11:13:37.000000000 +0200
 319 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/h8300/Kconfig       2011-06-10 22:11:24.000000000 +0200
 320 @@ -213,6 +213,8 @@ source "fs/Kconfig"
 321
 322  source "arch/h8300/Kconfig.debug"
 323
 324 +source "kernel/vserver/Kconfig"
 325 +
 326  source "security/Kconfig"
 327
 328  source "crypto/Kconfig"
 329 diff -NurpP --minimal linux-3.0-rc5/arch/ia64/Kconfig linux-3.0-rc5-vs2.3.1-pre3/arch/ia64/Kconfig
 330 --- linux-3.0-rc5/arch/ia64/Kconfig     2011-07-01 11:13:37.000000000 +0200
 331 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/ia64/Kconfig        2011-06-10 22:11:24.000000000 +0200
 332 @@ -671,6 +671,8 @@ source "fs/Kconfig"
 333
 334  source "arch/ia64/Kconfig.debug"
 335
 336 +source "kernel/vserver/Kconfig"
 337 +
 338  source "security/Kconfig"
 339
 340  source "crypto/Kconfig"
 341 diff -NurpP --minimal linux-3.0-rc5/arch/ia64/include/asm/tlb.h linux-3.0-rc5-vs2.3.1-pre3/arch/ia64/include/asm/tlb.h
 342 --- linux-3.0-rc5/arch/ia64/include/asm/tlb.h   2011-07-01 11:13:37.000000000 +0200
 343 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/ia64/include/asm/tlb.h      2011-06-10 22:11:24.000000000 +0200
 344 @@ -40,6 +40,7 @@
 345  #include <linux/mm.h>
 346  #include <linux/pagemap.h>
 347  #include <linux/swap.h>
 348 +#include <linux/vs_memory.h>
 349
 350  #include <asm/pgalloc.h>
 351  #include <asm/processor.h>
 352 diff -NurpP --minimal linux-3.0-rc5/arch/ia64/kernel/entry.S linux-3.0-rc5-vs2.3.1-pre3/arch/ia64/kernel/entry.S
 353 --- linux-3.0-rc5/arch/ia64/kernel/entry.S      2011-07-01 11:13:38.000000000 +0200
 354 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/ia64/kernel/entry.S 2011-06-10 22:11:24.000000000 +0200
 355 @@ -1714,7 +1714,7 @@ sys_call_table:
 356         data8 sys_mq_notify
 357         data8 sys_mq_getsetattr
 358         data8 sys_kexec_load
 359 -       data8 sys_ni_syscall                    // reserved for vserver
 360 +       data8 sys_vserver
 361         data8 sys_waitid                        // 1270
 362         data8 sys_add_key
 363         data8 sys_request_key
 364 diff -NurpP --minimal linux-3.0-rc5/arch/ia64/kernel/perfmon.c linux-3.0-rc5-vs2.3.1-pre3/arch/ia64/kernel/perfmon.c
 365 --- linux-3.0-rc5/arch/ia64/kernel/perfmon.c    2011-03-15 18:06:39.000000000 +0100
 366 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/ia64/kernel/perfmon.c       2011-06-10 22:11:24.000000000 +0200
 367 @@ -42,6 +42,7 @@
 368  #include <linux/completion.h>
 369  #include <linux/tracehook.h>
 370  #include <linux/slab.h>
 371 +#include <linux/vs_memory.h>
 372
 373  #include <asm/errno.h>
 374  #include <asm/intrinsics.h>
 375 diff -NurpP --minimal linux-3.0-rc5/arch/ia64/kernel/process.c linux-3.0-rc5-vs2.3.1-pre3/arch/ia64/kernel/process.c
 376 --- linux-3.0-rc5/arch/ia64/kernel/process.c    2011-03-15 18:06:39.000000000 +0100
 377 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/ia64/kernel/process.c       2011-06-10 22:11:24.000000000 +0200
 378 @@ -109,8 +109,8 @@ show_regs (struct pt_regs *regs)
 379         unsigned long ip = regs->cr_iip + ia64_psr(regs)->ri;
 380
 381         print_modules();
 382 -       printk("\nPid: %d, CPU %d, comm: %20s\n", task_pid_nr(current),
 383 -                       smp_processor_id(), current->comm);
 384 +       printk("\nPid: %d[#%u], CPU %d, comm: %20s\n", task_pid_nr(current),
 385 +                       current->xid, smp_processor_id(), current->comm);
 386         printk("psr : %016lx ifs : %016lx ip  : [<%016lx>]    %s (%s)\n",
 387                regs->cr_ipsr, regs->cr_ifs, ip, print_tainted(),
 388                init_utsname()->release);
 389 diff -NurpP --minimal linux-3.0-rc5/arch/ia64/kernel/ptrace.c linux-3.0-rc5-vs2.3.1-pre3/arch/ia64/kernel/ptrace.c
 390 --- linux-3.0-rc5/arch/ia64/kernel/ptrace.c     2011-01-05 21:48:59.000000000 +0100
 391 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/ia64/kernel/ptrace.c        2011-06-10 22:11:24.000000000 +0200
 392 @@ -21,6 +21,7 @@
 393  #include <linux/regset.h>
 394  #include <linux/elf.h>
 395  #include <linux/tracehook.h>
 396 +#include <linux/vs_base.h>
 397
 398  #include <asm/pgtable.h>
 399  #include <asm/processor.h>
 400 diff -NurpP --minimal linux-3.0-rc5/arch/ia64/kernel/traps.c linux-3.0-rc5-vs2.3.1-pre3/arch/ia64/kernel/traps.c
 401 --- linux-3.0-rc5/arch/ia64/kernel/traps.c      2010-07-07 18:31:01.000000000 +0200
 402 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/ia64/kernel/traps.c 2011-06-10 22:11:24.000000000 +0200
 403 @@ -59,8 +59,9 @@ die (const char *str, struct pt_regs *re
 404         put_cpu();
 405
 406         if (++die.lock_owner_depth < 3) {
 407 -               printk("%s[%d]: %s %ld [%d]\n",
 408 -               current->comm, task_pid_nr(current), str, err, ++die_counter);
 409 +               printk("%s[%d[#%u]]: %s %ld [%d]\n",
 410 +                       current->comm, task_pid_nr(current), current->xid,
 411 +                       str, err, ++die_counter);
 412                 if (notify_die(DIE_OOPS, str, regs, err, 255, SIGSEGV)
 413                     != NOTIFY_STOP)
 414                         show_regs(regs);
 415 @@ -323,8 +324,9 @@ handle_fpu_swa (int fp_fault, struct pt_
 416                         if ((last.count & 15) < 5 && (ia64_fetchadd(1, &last.count, acq) & 15) < 5) {
 417                                 last.time = current_jiffies + 5 * HZ;
 418                                 printk(KERN_WARNING
 419 -                                       "%s(%d): floating-point assist fault at ip %016lx, isr %016lx\n",
 420 -                                       current->comm, task_pid_nr(current), regs->cr_iip + ia64_psr(regs)->ri, isr);
 421 +                                       "%s(%d[#%u]): floating-point assist fault at ip %016lx, isr %016lx\n",
 422 +                                       current->comm, task_pid_nr(current), current->xid,
 423 +                                       regs->cr_iip + ia64_psr(regs)->ri, isr);
 424                         }
 425                 }
 426         }
 427 diff -NurpP --minimal linux-3.0-rc5/arch/ia64/mm/fault.c linux-3.0-rc5-vs2.3.1-pre3/arch/ia64/mm/fault.c
 428 --- linux-3.0-rc5/arch/ia64/mm/fault.c  2011-07-01 11:13:38.000000000 +0200
 429 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/ia64/mm/fault.c     2011-06-10 22:28:23.000000000 +0200
 430 @@ -11,6 +11,7 @@
 431  #include <linux/kprobes.h>
 432  #include <linux/kdebug.h>
 433  #include <linux/prefetch.h>
 434 +#include <linux/vs_memory.h>
 435
 436  #include <asm/pgtable.h>
 437  #include <asm/processor.h>
 438 diff -NurpP --minimal linux-3.0-rc5/arch/m32r/kernel/traps.c linux-3.0-rc5-vs2.3.1-pre3/arch/m32r/kernel/traps.c
 439 --- linux-3.0-rc5/arch/m32r/kernel/traps.c      2009-12-03 20:01:57.000000000 +0100
 440 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/m32r/kernel/traps.c 2011-06-10 22:11:24.000000000 +0200
 441 @@ -196,8 +196,9 @@ static void show_registers(struct pt_reg
 442         } else {
 443                 printk("SPI: %08lx\n", sp);
 444         }
 445 -       printk("Process %s (pid: %d, process nr: %d, stackpage=%08lx)",
 446 -               current->comm, task_pid_nr(current), 0xffff & i, 4096+(unsigned long)current);
 447 +       printk("Process %s (pid: %d[#%u], process nr: %d, stackpage=%08lx)",
 448 +               current->comm, task_pid_nr(current), current->xid,
 449 +               0xffff & i, 4096+(unsigned long)current);
 450
 451         /*
 452          * When in-kernel, we also print out the stack and code at the
 453 diff -NurpP --minimal linux-3.0-rc5/arch/m68k/Kconfig linux-3.0-rc5-vs2.3.1-pre3/arch/m68k/Kconfig
 454 --- linux-3.0-rc5/arch/m68k/Kconfig     2011-07-01 11:13:38.000000000 +0200
 455 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/m68k/Kconfig        2011-06-10 22:11:24.000000000 +0200
 456 @@ -241,6 +241,8 @@ source "fs/Kconfig"
 457
 458  source "arch/m68k/Kconfig.debug"
 459
 460 +source "kernel/vserver/Kconfig"
 461 +
 462  source "security/Kconfig"
 463
 464  source "crypto/Kconfig"
 465 diff -NurpP --minimal linux-3.0-rc5/arch/mips/Kconfig linux-3.0-rc5-vs2.3.1-pre3/arch/mips/Kconfig
 466 --- linux-3.0-rc5/arch/mips/Kconfig     2011-07-01 11:13:38.000000000 +0200
 467 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/mips/Kconfig        2011-06-10 22:11:24.000000000 +0200
 468 @@ -2485,6 +2485,8 @@ 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-3.0-rc5/arch/mips/kernel/ptrace.c linux-3.0-rc5-vs2.3.1-pre3/arch/mips/kernel/ptrace.c
 478 --- linux-3.0-rc5/arch/mips/kernel/ptrace.c     2011-07-01 11:13:39.000000000 +0200
 479 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/mips/kernel/ptrace.c        2011-06-10 22:11:24.000000000 +0200
 480 @@ -25,6 +25,7 @@
 481  #include <linux/security.h>
 482  #include <linux/audit.h>
 483  #include <linux/seccomp.h>
 484 +#include <linux/vs_base.h>
 485
 486  #include <asm/byteorder.h>
 487  #include <asm/cpu.h>
 488 @@ -263,6 +264,9 @@ long arch_ptrace(struct task_struct *chi
 489         void __user *datavp = (void __user *) data;
 490         unsigned long __user *datalp = (void __user *) data;
 491
 492 +       if (!vx_check(vx_task_xid(child), VS_WATCH_P | VS_IDENT))
 493 +               goto out;
 494 +
 495         switch (request) {
 496         /* when I and D space are separate, these will need to be fixed. */
 497         case PTRACE_PEEKTEXT: /* read word at location addr. */
 498 diff -NurpP --minimal linux-3.0-rc5/arch/mips/kernel/scall32-o32.S linux-3.0-rc5-vs2.3.1-pre3/arch/mips/kernel/scall32-o32.S
 499 --- linux-3.0-rc5/arch/mips/kernel/scall32-o32.S        2011-07-01 11:13:39.000000000 +0200
 500 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/mips/kernel/scall32-o32.S   2011-06-10 22:11:24.000000000 +0200
 501 @@ -523,7 +523,7 @@ einval:     li      v0, -ENOSYS
 502         sys     sys_mq_timedreceive     5
 503         sys     sys_mq_notify           2       /* 4275 */
 504         sys     sys_mq_getsetattr       3
 505 -       sys     sys_ni_syscall          0       /* sys_vserver */
 506 +       sys     sys_vserver             3
 507         sys     sys_waitid              5
 508         sys     sys_ni_syscall          0       /* available, was setaltroot */
 509         sys     sys_add_key             5       /* 4280 */
 510 diff -NurpP --minimal linux-3.0-rc5/arch/mips/kernel/scall64-64.S linux-3.0-rc5-vs2.3.1-pre3/arch/mips/kernel/scall64-64.S
 511 --- linux-3.0-rc5/arch/mips/kernel/scall64-64.S 2011-07-01 11:13:39.000000000 +0200
 512 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/mips/kernel/scall64-64.S    2011-06-10 22:11:24.000000000 +0200
 513 @@ -362,7 +362,7 @@ sys_call_table:
 514         PTR     sys_mq_timedreceive
 515         PTR     sys_mq_notify
 516         PTR     sys_mq_getsetattr               /* 5235 */
 517 -       PTR     sys_ni_syscall                  /* sys_vserver */
 518 +       PTR     sys_vserver
 519         PTR     sys_waitid
 520         PTR     sys_ni_syscall                  /* available, was setaltroot */
 521         PTR     sys_add_key
 522 diff -NurpP --minimal linux-3.0-rc5/arch/mips/kernel/scall64-n32.S linux-3.0-rc5-vs2.3.1-pre3/arch/mips/kernel/scall64-n32.S
 523 --- linux-3.0-rc5/arch/mips/kernel/scall64-n32.S        2011-07-01 11:13:39.000000000 +0200
 524 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/mips/kernel/scall64-n32.S   2011-06-10 22:11:24.000000000 +0200
 525 @@ -361,7 +361,7 @@ EXPORT(sysn32_call_table)
 526         PTR     compat_sys_mq_timedreceive
 527         PTR     compat_sys_mq_notify
 528         PTR     compat_sys_mq_getsetattr
 529 -       PTR     sys_ni_syscall                  /* 6240, sys_vserver */
 530 +       PTR     sys32_vserver                   /* 6240 */
 531         PTR     compat_sys_waitid
 532         PTR     sys_ni_syscall                  /* available, was setaltroot */
 533         PTR     sys_add_key
 534 diff -NurpP --minimal linux-3.0-rc5/arch/mips/kernel/scall64-o32.S linux-3.0-rc5-vs2.3.1-pre3/arch/mips/kernel/scall64-o32.S
 535 --- linux-3.0-rc5/arch/mips/kernel/scall64-o32.S        2011-07-01 11:13:39.000000000 +0200
 536 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/mips/kernel/scall64-o32.S   2011-06-10 22:11:24.000000000 +0200
 537 @@ -480,7 +480,7 @@ sys_call_table:
 538         PTR     compat_sys_mq_timedreceive
 539         PTR     compat_sys_mq_notify            /* 4275 */
 540         PTR     compat_sys_mq_getsetattr
 541 -       PTR     sys_ni_syscall                  /* sys_vserver */
 542 +       PTR     sys32_vserver
 543         PTR     sys_32_waitid
 544         PTR     sys_ni_syscall                  /* available, was setaltroot */
 545         PTR     sys_add_key                     /* 4280 */
 546 diff -NurpP --minimal linux-3.0-rc5/arch/mips/kernel/traps.c linux-3.0-rc5-vs2.3.1-pre3/arch/mips/kernel/traps.c
 547 --- linux-3.0-rc5/arch/mips/kernel/traps.c      2011-05-22 16:17:00.000000000 +0200
 548 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/mips/kernel/traps.c 2011-06-10 22:11:24.000000000 +0200
 549 @@ -343,9 +343,10 @@ void show_registers(struct pt_regs *regs
 550
 551         __show_regs(regs);
 552         print_modules();
 553 -       printk("Process %s (pid: %d, threadinfo=%p, task=%p, tls=%0*lx)\n",
 554 -              current->comm, current->pid, current_thread_info(), current,
 555 -             field, current_thread_info()->tp_value);
 556 +       printk("Process %s (pid: %d:#%u, threadinfo=%p, task=%p, tls=%0*lx)\n",
 557 +               current->comm, task_pid_nr(current), current->xid,
 558 +               current_thread_info(), current,
 559 +               field, current_thread_info()->tp_value);
 560         if (cpu_has_userlocal) {
 561                 unsigned long tls;
 562
 563 diff -NurpP --minimal linux-3.0-rc5/arch/parisc/Kconfig linux-3.0-rc5-vs2.3.1-pre3/arch/parisc/Kconfig
 564 --- linux-3.0-rc5/arch/parisc/Kconfig   2011-07-01 11:13:42.000000000 +0200
 565 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/parisc/Kconfig      2011-06-10 22:11:24.000000000 +0200
 566 @@ -279,6 +279,8 @@ source "fs/Kconfig"
 567
 568  source "arch/parisc/Kconfig.debug"
 569
 570 +source "kernel/vserver/Kconfig"
 571 +
 572  source "security/Kconfig"
 573
 574  source "crypto/Kconfig"
 575 diff -NurpP --minimal linux-3.0-rc5/arch/parisc/kernel/syscall_table.S linux-3.0-rc5-vs2.3.1-pre3/arch/parisc/kernel/syscall_table.S
 576 --- linux-3.0-rc5/arch/parisc/kernel/syscall_table.S    2011-07-01 11:13:43.000000000 +0200
 577 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/parisc/kernel/syscall_table.S       2011-06-10 22:11:24.000000000 +0200
 578 @@ -361,7 +361,7 @@
 579         ENTRY_COMP(mbind)               /* 260 */
 580         ENTRY_COMP(get_mempolicy)
 581         ENTRY_COMP(set_mempolicy)
 582 -       ENTRY_SAME(ni_syscall)  /* 263: reserved for vserver */
 583 +       ENTRY_DIFF(vserver)
 584         ENTRY_SAME(add_key)
 585         ENTRY_SAME(request_key)         /* 265 */
 586         ENTRY_SAME(keyctl)
 587 diff -NurpP --minimal linux-3.0-rc5/arch/parisc/kernel/traps.c linux-3.0-rc5-vs2.3.1-pre3/arch/parisc/kernel/traps.c
 588 --- linux-3.0-rc5/arch/parisc/kernel/traps.c    2009-09-10 15:25:40.000000000 +0200
 589 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/parisc/kernel/traps.c       2011-06-10 22:11:24.000000000 +0200
 590 @@ -236,8 +236,9 @@ void die_if_kernel(char *str, struct pt_
 591                 if (err == 0)
 592                         return; /* STFU */
 593
 594 -               printk(KERN_CRIT "%s (pid %d): %s (code %ld) at " RFMT "\n",
 595 -                       current->comm, task_pid_nr(current), str, err, regs->iaoq[0]);
 596 +               printk(KERN_CRIT "%s (pid %d:#%u): %s (code %ld) at " RFMT "\n",
 597 +                       current->comm, task_pid_nr(current), current->xid,
 598 +                       str, err, regs->iaoq[0]);
 599  #ifdef PRINT_USER_FAULTS
 600                 /* XXX for debugging only */
 601                 show_regs(regs);
 602 @@ -270,8 +271,8 @@ void die_if_kernel(char *str, struct pt_
 603                 pdc_console_restart();
 604
 605         if (err)
 606 -               printk(KERN_CRIT "%s (pid %d): %s (code %ld)\n",
 607 -                       current->comm, task_pid_nr(current), str, err);
 608 +               printk(KERN_CRIT "%s (pid %d:#%u): %s (code %ld)\n",
 609 +                       current->comm, task_pid_nr(current), current->xid, str, err);
 610
 611         /* Wot's wrong wif bein' racy? */
 612         if (current->thread.flags & PARISC_KERNEL_DEATH) {
 613 diff -NurpP --minimal linux-3.0-rc5/arch/parisc/mm/fault.c linux-3.0-rc5-vs2.3.1-pre3/arch/parisc/mm/fault.c
 614 --- linux-3.0-rc5/arch/parisc/mm/fault.c        2010-08-02 16:52:06.000000000 +0200
 615 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/parisc/mm/fault.c   2011-06-10 22:11:24.000000000 +0200
 616 @@ -237,8 +237,9 @@ bad_area:
 617
 618  #ifdef PRINT_USER_FAULTS
 619                 printk(KERN_DEBUG "\n");
 620 -               printk(KERN_DEBUG "do_page_fault() pid=%d command='%s' type=%lu address=0x%08lx\n",
 621 -                   task_pid_nr(tsk), tsk->comm, code, address);
 622 +               printk(KERN_DEBUG "do_page_fault() pid=%d:#%u "
 623 +                   "command='%s' type=%lu address=0x%08lx\n",
 624 +                   task_pid_nr(tsk), tsk->xid, tsk->comm, code, address);
 625                 if (vma) {
 626                         printk(KERN_DEBUG "vm_start = 0x%08lx, vm_end = 0x%08lx\n",
 627                                         vma->vm_start, vma->vm_end);
 628 diff -NurpP --minimal linux-3.0-rc5/arch/powerpc/Kconfig linux-3.0-rc5-vs2.3.1-pre3/arch/powerpc/Kconfig
 629 --- linux-3.0-rc5/arch/powerpc/Kconfig  2011-07-01 11:13:43.000000000 +0200
 630 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/powerpc/Kconfig     2011-06-10 22:11:24.000000000 +0200
 631 @@ -978,6 +978,8 @@ source "lib/Kconfig"
 632
 633  source "arch/powerpc/Kconfig.debug"
 634
 635 +source "kernel/vserver/Kconfig"
 636 +
 637  source "security/Kconfig"
 638
 639  config KEYS_COMPAT
 640 diff -NurpP --minimal linux-3.0-rc5/arch/powerpc/include/asm/unistd.h linux-3.0-rc5-vs2.3.1-pre3/arch/powerpc/include/asm/unistd.h
 641 --- linux-3.0-rc5/arch/powerpc/include/asm/unistd.h     2011-07-01 11:13:44.000000000 +0200
 642 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/powerpc/include/asm/unistd.h        2011-06-10 22:11:24.000000000 +0200
 643 @@ -275,7 +275,7 @@
 644  #endif
 645  #define __NR_rtas              255
 646  #define __NR_sys_debug_setcontext 256
 647 -/* Number 257 is reserved for vserver */
 648 +#define __NR_vserver           257
 649  #define __NR_migrate_pages     258
 650  #define __NR_mbind             259
 651  #define __NR_get_mempolicy     260
 652 diff -NurpP --minimal linux-3.0-rc5/arch/powerpc/kernel/process.c linux-3.0-rc5-vs2.3.1-pre3/arch/powerpc/kernel/process.c
 653 --- linux-3.0-rc5/arch/powerpc/kernel/process.c 2011-07-01 11:13:44.000000000 +0200
 654 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/powerpc/kernel/process.c    2011-06-10 22:11:24.000000000 +0200
 655 @@ -656,8 +656,9 @@ void show_regs(struct pt_regs * regs)
 656  #else
 657                 printk("DAR: "REG", DSISR: %08lx\n", regs->dar, regs->dsisr);
 658  #endif
 659 -       printk("TASK = %p[%d] '%s' THREAD: %p",
 660 -              current, task_pid_nr(current), current->comm, task_thread_info(current));
 661 +       printk("TASK = %p[%d,#%u] '%s' THREAD: %p",
 662 +              current, task_pid_nr(current), current->xid,
 663 +              current->comm, task_thread_info(current));
 664
 665  #ifdef CONFIG_SMP
 666         printk(" CPU: %d", raw_smp_processor_id());
 667 diff -NurpP --minimal linux-3.0-rc5/arch/powerpc/kernel/traps.c linux-3.0-rc5-vs2.3.1-pre3/arch/powerpc/kernel/traps.c
 668 --- linux-3.0-rc5/arch/powerpc/kernel/traps.c   2011-07-01 11:13:44.000000000 +0200
 669 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/powerpc/kernel/traps.c      2011-06-10 22:11:24.000000000 +0200
 670 @@ -1075,8 +1075,9 @@ void nonrecoverable_exception(struct pt_
 671
 672  void trace_syscall(struct pt_regs *regs)
 673  {
 674 -       printk("Task: %p(%d), PC: %08lX/%08lX, Syscall: %3ld, Result: %s%ld    %s\n",
 675 -              current, task_pid_nr(current), regs->nip, regs->link, regs->gpr[0],
 676 +       printk("Task: %p(%d[#%u]), PC: %08lX/%08lX, Syscall: %3ld, Result: %s%ld    %s\n",
 677 +              current, task_pid_nr(current), current->xid,
 678 +              regs->nip, regs->link, regs->gpr[0],
 679                regs->ccr&0x10000000?"Error=":"", regs->gpr[3], print_tainted());
 680  }
 681
 682 diff -NurpP --minimal linux-3.0-rc5/arch/powerpc/kernel/vdso.c linux-3.0-rc5-vs2.3.1-pre3/arch/powerpc/kernel/vdso.c
 683 --- linux-3.0-rc5/arch/powerpc/kernel/vdso.c    2011-05-22 16:17:02.000000000 +0200
 684 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/powerpc/kernel/vdso.c       2011-06-10 22:11:24.000000000 +0200
 685 @@ -23,6 +23,7 @@
 686  #include <linux/security.h>
 687  #include <linux/bootmem.h>
 688  #include <linux/memblock.h>
 689 +#include <linux/vs_memory.h>
 690
 691  #include <asm/pgtable.h>
 692  #include <asm/system.h>
 693 diff -NurpP --minimal linux-3.0-rc5/arch/s390/Kconfig linux-3.0-rc5-vs2.3.1-pre3/arch/s390/Kconfig
 694 --- linux-3.0-rc5/arch/s390/Kconfig     2011-07-01 11:13:45.000000000 +0200
 695 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/s390/Kconfig        2011-07-01 11:35:34.000000000 +0200
 696 @@ -628,6 +628,8 @@ source "fs/Kconfig"
 697
 698  source "arch/s390/Kconfig.debug"
 699
 700 +source "kernel/vserver/Kconfig"
 701 +
 702  source "security/Kconfig"
 703
 704  source "crypto/Kconfig"
 705 diff -NurpP --minimal linux-3.0-rc5/arch/s390/include/asm/tlb.h linux-3.0-rc5-vs2.3.1-pre3/arch/s390/include/asm/tlb.h
 706 --- linux-3.0-rc5/arch/s390/include/asm/tlb.h   2011-07-01 11:13:45.000000000 +0200
 707 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/s390/include/asm/tlb.h      2011-06-15 02:40:14.000000000 +0200
 708 @@ -24,6 +24,8 @@
 709  #include <linux/mm.h>
 710  #include <linux/pagemap.h>
 711  #include <linux/swap.h>
 712 +#include <linux/vs_memory.h>
 713 +
 714  #include <asm/processor.h>
 715  #include <asm/pgalloc.h>
 716  #include <asm/tlbflush.h>
 717 diff -NurpP --minimal linux-3.0-rc5/arch/s390/include/asm/unistd.h linux-3.0-rc5-vs2.3.1-pre3/arch/s390/include/asm/unistd.h
 718 --- linux-3.0-rc5/arch/s390/include/asm/unistd.h        2011-07-01 11:13:45.000000000 +0200
 719 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/s390/include/asm/unistd.h   2011-06-10 22:11:24.000000000 +0200
 720 @@ -202,7 +202,7 @@
 721  #define __NR_clock_gettime     (__NR_timer_create+6)
 722  #define __NR_clock_getres      (__NR_timer_create+7)
 723  #define __NR_clock_nanosleep   (__NR_timer_create+8)
 724 -/* Number 263 is reserved for vserver */
 725 +#define __NR_vserver           263
 726  #define __NR_statfs64          265
 727  #define __NR_fstatfs64         266
 728  #define __NR_remap_file_pages  267
 729 diff -NurpP --minimal linux-3.0-rc5/arch/s390/kernel/ptrace.c linux-3.0-rc5-vs2.3.1-pre3/arch/s390/kernel/ptrace.c
 730 --- linux-3.0-rc5/arch/s390/kernel/ptrace.c     2011-03-15 18:06:45.000000000 +0100
 731 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/s390/kernel/ptrace.c        2011-06-10 22:11:24.000000000 +0200
 732 @@ -20,6 +20,7 @@
 733  #include <linux/regset.h>
 734  #include <linux/tracehook.h>
 735  #include <linux/seccomp.h>
 736 +#include <linux/vs_base.h>
 737  #include <trace/syscall.h>
 738  #include <asm/compat.h>
 739  #include <asm/segment.h>
 740 diff -NurpP --minimal linux-3.0-rc5/arch/s390/kernel/syscalls.S linux-3.0-rc5-vs2.3.1-pre3/arch/s390/kernel/syscalls.S
 741 --- linux-3.0-rc5/arch/s390/kernel/syscalls.S   2011-07-01 11:13:45.000000000 +0200
 742 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/s390/kernel/syscalls.S      2011-06-10 22:11:24.000000000 +0200
 743 @@ -271,7 +271,7 @@ SYSCALL(sys_clock_settime,sys_clock_sett
 744  SYSCALL(sys_clock_gettime,sys_clock_gettime,sys32_clock_gettime_wrapper)       /* 260 */
 745  SYSCALL(sys_clock_getres,sys_clock_getres,sys32_clock_getres_wrapper)
 746  SYSCALL(sys_clock_nanosleep,sys_clock_nanosleep,sys32_clock_nanosleep_wrapper)
 747 -NI_SYSCALL                                                     /* reserved for vserver */
 748 +SYSCALL(sys_vserver,sys_vserver,sys32_vserver)
 749  SYSCALL(sys_s390_fadvise64_64,sys_ni_syscall,sys32_fadvise64_64_wrapper)
 750  SYSCALL(sys_statfs64,sys_statfs64,compat_sys_statfs64_wrapper)
 751  SYSCALL(sys_fstatfs64,sys_fstatfs64,compat_sys_fstatfs64_wrapper)
 752 diff -NurpP --minimal linux-3.0-rc5/arch/sh/Kconfig linux-3.0-rc5-vs2.3.1-pre3/arch/sh/Kconfig
 753 --- linux-3.0-rc5/arch/sh/Kconfig       2011-07-01 11:13:45.000000000 +0200
 754 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/sh/Kconfig  2011-06-10 22:11:24.000000000 +0200
 755 @@ -888,6 +888,8 @@ source "fs/Kconfig"
 756
 757  source "arch/sh/Kconfig.debug"
 758
 759 +source "kernel/vserver/Kconfig"
 760 +
 761  source "security/Kconfig"
 762
 763  source "crypto/Kconfig"
 764 diff -NurpP --minimal linux-3.0-rc5/arch/sh/kernel/irq.c linux-3.0-rc5-vs2.3.1-pre3/arch/sh/kernel/irq.c
 765 --- linux-3.0-rc5/arch/sh/kernel/irq.c  2011-05-22 16:17:07.000000000 +0200
 766 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/sh/kernel/irq.c     2011-06-10 22:11:24.000000000 +0200
 767 @@ -13,6 +13,7 @@
 768  #include <linux/seq_file.h>
 769  #include <linux/ftrace.h>
 770  #include <linux/delay.h>
 771 +// #include <linux/vs_context.h>
 772  #include <linux/ratelimit.h>
 773  #include <asm/processor.h>
 774  #include <asm/machvec.h>
 775 diff -NurpP --minimal linux-3.0-rc5/arch/sh/kernel/vsyscall/vsyscall.c linux-3.0-rc5-vs2.3.1-pre3/arch/sh/kernel/vsyscall/vsyscall.c
 776 --- linux-3.0-rc5/arch/sh/kernel/vsyscall/vsyscall.c    2011-05-22 16:17:07.000000000 +0200
 777 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/sh/kernel/vsyscall/vsyscall.c       2011-06-10 22:11:24.000000000 +0200
 778 @@ -18,6 +18,7 @@
 779  #include <linux/elf.h>
 780  #include <linux/sched.h>
 781  #include <linux/err.h>
 782 +#include <linux/vs_memory.h>
 783
 784  /*
 785   * Should the kernel map a VDSO page into processes and pass its
 786 diff -NurpP --minimal linux-3.0-rc5/arch/sparc/Kconfig linux-3.0-rc5-vs2.3.1-pre3/arch/sparc/Kconfig
 787 --- linux-3.0-rc5/arch/sparc/Kconfig    2011-07-01 11:13:46.000000000 +0200
 788 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/sparc/Kconfig       2011-06-15 02:40:14.000000000 +0200
 789 @@ -601,6 +601,8 @@ source "fs/Kconfig"
 790
 791  source "arch/sparc/Kconfig.debug"
 792
 793 +source "kernel/vserver/Kconfig"
 794 +
 795  source "security/Kconfig"
 796
 797  source "crypto/Kconfig"
 798 diff -NurpP --minimal linux-3.0-rc5/arch/sparc/include/asm/tlb_64.h linux-3.0-rc5-vs2.3.1-pre3/arch/sparc/include/asm/tlb_64.h
 799 --- linux-3.0-rc5/arch/sparc/include/asm/tlb_64.h       2011-07-01 11:13:46.000000000 +0200
 800 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/sparc/include/asm/tlb_64.h  2011-06-10 22:11:24.000000000 +0200
 801 @@ -3,6 +3,7 @@
 802
 803  #include <linux/swap.h>
 804  #include <linux/pagemap.h>
 805 +#include <linux/vs_memory.h>
 806  #include <asm/pgalloc.h>
 807  #include <asm/tlbflush.h>
 808  #include <asm/mmu_context.h>
 809 diff -NurpP --minimal linux-3.0-rc5/arch/sparc/include/asm/unistd.h linux-3.0-rc5-vs2.3.1-pre3/arch/sparc/include/asm/unistd.h
 810 --- linux-3.0-rc5/arch/sparc/include/asm/unistd.h       2011-07-01 11:13:46.000000000 +0200
 811 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/sparc/include/asm/unistd.h  2011-06-10 22:11:24.000000000 +0200
 812 @@ -335,7 +335,7 @@
 813  #define __NR_timer_getoverrun  264
 814  #define __NR_timer_delete      265
 815  #define __NR_timer_create      266
 816 -/* #define __NR_vserver                267 Reserved for VSERVER */
 817 +#define __NR_vserver           267
 818  #define __NR_io_setup          268
 819  #define __NR_io_destroy                269
 820  #define __NR_io_submit         270
 821 diff -NurpP --minimal linux-3.0-rc5/arch/sparc/kernel/systbls_32.S linux-3.0-rc5-vs2.3.1-pre3/arch/sparc/kernel/systbls_32.S
 822 --- linux-3.0-rc5/arch/sparc/kernel/systbls_32.S        2011-07-01 11:13:46.000000000 +0200
 823 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/sparc/kernel/systbls_32.S   2011-06-10 22:11:24.000000000 +0200
 824 @@ -70,7 +70,7 @@ sys_call_table:
 825  /*250*/        .long sys_mremap, sys_sysctl, sys_getsid, sys_fdatasync, sys_nfsservctl
 826  /*255*/        .long sys_sync_file_range, sys_clock_settime, sys_clock_gettime, sys_clock_getres, sys_clock_nanosleep
 827  /*260*/        .long sys_sched_getaffinity, sys_sched_setaffinity, sys_timer_settime, sys_timer_gettime, sys_timer_getoverrun
 828 -/*265*/        .long sys_timer_delete, sys_timer_create, sys_nis_syscall, sys_io_setup, sys_io_destroy
 829 +/*265*/        .long sys_timer_delete, sys_timer_create, sys_vserver, sys_io_setup, sys_io_destroy
 830  /*270*/        .long sys_io_submit, sys_io_cancel, sys_io_getevents, sys_mq_open, sys_mq_unlink
 831  /*275*/        .long sys_mq_timedsend, sys_mq_timedreceive, sys_mq_notify, sys_mq_getsetattr, sys_waitid
 832  /*280*/        .long sys_tee, sys_add_key, sys_request_key, sys_keyctl, sys_openat
 833 diff -NurpP --minimal linux-3.0-rc5/arch/sparc/kernel/systbls_64.S linux-3.0-rc5-vs2.3.1-pre3/arch/sparc/kernel/systbls_64.S
 834 --- linux-3.0-rc5/arch/sparc/kernel/systbls_64.S        2011-07-01 11:13:46.000000000 +0200
 835 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/sparc/kernel/systbls_64.S   2011-06-10 22:11:24.000000000 +0200
 836 @@ -71,7 +71,7 @@ sys_call_table32:
 837  /*250*/        .word sys_mremap, compat_sys_sysctl, sys32_getsid, sys_fdatasync, sys32_nfsservctl
 838         .word sys32_sync_file_range, compat_sys_clock_settime, compat_sys_clock_gettime, compat_sys_clock_getres, sys32_clock_nanosleep
 839  /*260*/        .word compat_sys_sched_getaffinity, compat_sys_sched_setaffinity, sys32_timer_settime, compat_sys_timer_gettime, sys_timer_getoverrun
 840 -       .word sys_timer_delete, compat_sys_timer_create, sys_ni_syscall, compat_sys_io_setup, sys_io_destroy
 841 +       .word sys_timer_delete, compat_sys_timer_create, sys32_vserver, compat_sys_io_setup, sys_io_destroy
 842  /*270*/        .word sys32_io_submit, sys_io_cancel, compat_sys_io_getevents, sys32_mq_open, sys_mq_unlink
 843         .word compat_sys_mq_timedsend, compat_sys_mq_timedreceive, compat_sys_mq_notify, compat_sys_mq_getsetattr, compat_sys_waitid
 844  /*280*/        .word sys32_tee, sys_add_key, sys_request_key, sys_keyctl, compat_sys_openat
 845 @@ -148,7 +148,7 @@ sys_call_table:
 846  /*250*/        .word sys_64_mremap, sys_sysctl, sys_getsid, sys_fdatasync, sys_nfsservctl
 847         .word sys_sync_file_range, sys_clock_settime, sys_clock_gettime, sys_clock_getres, sys_clock_nanosleep
 848  /*260*/        .word sys_sched_getaffinity, sys_sched_setaffinity, sys_timer_settime, sys_timer_gettime, sys_timer_getoverrun
 849 -       .word sys_timer_delete, sys_timer_create, sys_ni_syscall, sys_io_setup, sys_io_destroy
 850 +       .word sys_timer_delete, sys_timer_create, sys_vserver, sys_io_setup, sys_io_destroy
 851  /*270*/        .word sys_io_submit, sys_io_cancel, sys_io_getevents, sys_mq_open, sys_mq_unlink
 852         .word sys_mq_timedsend, sys_mq_timedreceive, sys_mq_notify, sys_mq_getsetattr, sys_waitid
 853  /*280*/        .word sys_tee, sys_add_key, sys_request_key, sys_keyctl, sys_openat
 854 diff -NurpP --minimal linux-3.0-rc5/arch/um/Kconfig.rest linux-3.0-rc5-vs2.3.1-pre3/arch/um/Kconfig.rest
 855 --- linux-3.0-rc5/arch/um/Kconfig.rest  2009-06-11 17:12:19.000000000 +0200
 856 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/um/Kconfig.rest     2011-06-10 22:11:24.000000000 +0200
 857 @@ -18,6 +18,8 @@ source "drivers/connector/Kconfig"
 858
 859  source "fs/Kconfig"
 860
 861 +source "kernel/vserver/Kconfig"
 862 +
 863  source "security/Kconfig"
 864
 865  source "crypto/Kconfig"
 866 diff -NurpP --minimal linux-3.0-rc5/arch/um/include/asm/tlb.h linux-3.0-rc5-vs2.3.1-pre3/arch/um/include/asm/tlb.h
 867 --- linux-3.0-rc5/arch/um/include/asm/tlb.h     2011-07-01 11:13:46.000000000 +0200
 868 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/um/include/asm/tlb.h        2011-06-10 22:11:24.000000000 +0200
 869 @@ -3,6 +3,7 @@
 870
 871  #include <linux/pagemap.h>
 872  #include <linux/swap.h>
 873 +#include <linux/vs_memory.h>
 874  #include <asm/percpu.h>
 875  #include <asm/pgalloc.h>
 876  #include <asm/tlbflush.h>
 877 diff -NurpP --minimal linux-3.0-rc5/arch/um/include/shared/kern_constants.h linux-3.0-rc5-vs2.3.1-pre3/arch/um/include/shared/kern_constants.h
 878 --- linux-3.0-rc5/arch/um/include/shared/kern_constants.h       1970-01-01 01:00:00.000000000 +0100
 879 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/um/include/shared/kern_constants.h  2011-06-10 22:11:24.000000000 +0200
 880 @@ -0,0 +1 @@
 881 +#include "../../../../include/generated/asm-offsets.h"
 882 diff -NurpP --minimal linux-3.0-rc5/arch/um/include/shared/user_constants.h linux-3.0-rc5-vs2.3.1-pre3/arch/um/include/shared/user_constants.h
 883 --- linux-3.0-rc5/arch/um/include/shared/user_constants.h       1970-01-01 01:00:00.000000000 +0100
 884 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/um/include/shared/user_constants.h  2011-06-10 22:11:24.000000000 +0200
 885 @@ -0,0 +1,40 @@
 886 +/*
 887 + * DO NOT MODIFY.
 888 + *
 889 + * This file was generated by arch/um/Makefile
 890 + *
 891 + */
 892 +
 893 +#define HOST_SC_CR2 176 /* offsetof(struct sigcontext, cr2)    # */
 894 +#define HOST_SC_ERR 152 /* offsetof(struct sigcontext, err)    # */
 895 +#define HOST_SC_TRAPNO 160 /* offsetof(struct sigcontext, trapno)      # */
 896 +#define HOST_FP_SIZE 64 /* sizeof(struct _fpstate) / sizeof(unsigned long)     # */
 897 +#define HOST_RBX 5 /* RBX      # */
 898 +#define HOST_RCX 11 /* RCX     # */
 899 +#define HOST_RDI 14 /* RDI     # */
 900 +#define HOST_RSI 13 /* RSI     # */
 901 +#define HOST_RDX 12 /* RDX     # */
 902 +#define HOST_RBP 4 /* RBP      # */
 903 +#define HOST_RAX 10 /* RAX     # */
 904 +#define HOST_R8 9 /* R8        # */
 905 +#define HOST_R9 8 /* R9        # */
 906 +#define HOST_R10 7 /* R10      # */
 907 +#define HOST_R11 6 /* R11      # */
 908 +#define HOST_R12 3 /* R12      # */
 909 +#define HOST_R13 2 /* R13      # */
 910 +#define HOST_R14 1 /* R14      # */
 911 +#define HOST_R15 0 /* R15      # */
 912 +#define HOST_ORIG_RAX 15 /* ORIG_RAX   # */
 913 +#define HOST_CS 17 /* CS       # */
 914 +#define HOST_SS 20 /* SS       # */
 915 +#define HOST_EFLAGS 18 /* EFLAGS       # */
 916 +#define HOST_IP 16 /* RIP      # */
 917 +#define HOST_SP 19 /* RSP      # */
 918 +#define UM_FRAME_SIZE 216 /* sizeof(struct user_regs_struct)   # */
 919 +#define UM_POLLIN 1 /* POLLIN  # */
 920 +#define UM_POLLPRI 2 /* POLLPRI        # */
 921 +#define UM_POLLOUT 4 /* POLLOUT        # */
 922 +#define UM_PROT_READ 1 /* PROT_READ    # */
 923 +#define UM_PROT_WRITE 2 /* PROT_WRITE  # */
 924 +#define UM_PROT_EXEC 4 /* PROT_EXEC    # */
 925 +
 926 diff -NurpP --minimal linux-3.0-rc5/arch/x86/Kconfig linux-3.0-rc5-vs2.3.1-pre3/arch/x86/Kconfig
 927 --- linux-3.0-rc5/arch/x86/Kconfig      2011-07-01 11:13:47.000000000 +0200
 928 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/x86/Kconfig 2011-06-10 22:11:24.000000000 +0200
 929 @@ -2159,6 +2159,8 @@ source "fs/Kconfig"
 930
 931  source "arch/x86/Kconfig.debug"
 932
 933 +source "kernel/vserver/Kconfig"
 934 +
 935  source "security/Kconfig"
 936
 937  source "crypto/Kconfig"
 938 diff -NurpP --minimal linux-3.0-rc5/arch/x86/ia32/ia32entry.S linux-3.0-rc5-vs2.3.1-pre3/arch/x86/ia32/ia32entry.S
 939 --- linux-3.0-rc5/arch/x86/ia32/ia32entry.S     2011-07-01 11:13:47.000000000 +0200
 940 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/x86/ia32/ia32entry.S        2011-06-10 22:11:24.000000000 +0200
 941 @@ -776,7 +776,7 @@ ia32_sys_call_table:
 942         .quad sys_tgkill                /* 270 */
 943         .quad compat_sys_utimes
 944         .quad sys32_fadvise64_64
 945 -       .quad quiet_ni_syscall  /* sys_vserver */
 946 +       .quad sys32_vserver
 947         .quad sys_mbind
 948         .quad compat_sys_get_mempolicy  /* 275 */
 949         .quad sys_set_mempolicy
 950 diff -NurpP --minimal linux-3.0-rc5/arch/x86/include/asm/unistd_64.h linux-3.0-rc5-vs2.3.1-pre3/arch/x86/include/asm/unistd_64.h
 951 --- linux-3.0-rc5/arch/x86/include/asm/unistd_64.h      2011-07-01 11:13:47.000000000 +0200
 952 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/x86/include/asm/unistd_64.h 2011-06-10 22:11:24.000000000 +0200
 953 @@ -535,7 +535,7 @@ __SYSCALL(__NR_tgkill, sys_tgkill)
 954  #define __NR_utimes                            235
 955  __SYSCALL(__NR_utimes, sys_utimes)
 956  #define __NR_vserver                           236
 957 -__SYSCALL(__NR_vserver, sys_ni_syscall)
 958 +__SYSCALL(__NR_vserver, sys_vserver)
 959  #define __NR_mbind                             237
 960  __SYSCALL(__NR_mbind, sys_mbind)
 961  #define __NR_set_mempolicy                     238
 962 diff -NurpP --minimal linux-3.0-rc5/arch/x86/kernel/syscall_table_32.S linux-3.0-rc5-vs2.3.1-pre3/arch/x86/kernel/syscall_table_32.S
 963 --- linux-3.0-rc5/arch/x86/kernel/syscall_table_32.S    2011-07-01 11:13:48.000000000 +0200
 964 +++ linux-3.0-rc5-vs2.3.1-pre3/arch/x86/kernel/syscall_table_32.S       2011-06-10 22:11:24.000000000 +0200
 965 @@ -272,7 +272,7 @@ ENTRY(sys_call_table)
 966         .long sys_tgkill        /* 270 */
 967         .long sys_utimes
 968         .long sys_fadvise64_64
 969 -       .long sys_ni_syscall    /* sys_vserver */
 970 +       .long sys_vserver
 971         .long sys_mbind
 972         .long sys_get_mempolicy
 973         .long sys_set_mempolicy
 974 diff -NurpP --minimal linux-3.0-rc5/block/genhd.c linux-3.0-rc5-vs2.3.1-pre3/block/genhd.c
 975 --- linux-3.0-rc5/block/genhd.c 2011-07-01 11:13:48.000000000 +0200
 976 +++ linux-3.0-rc5-vs2.3.1-pre3/block/genhd.c    2011-07-01 11:35:34.000000000 +0200
 977 @@ -1154,17 +1154,17 @@ static int diskstats_show(struct seq_fil
 978                 cpu = part_stat_lock();
 979                 part_round_stats(cpu, hd);
 980                 part_stat_unlock();
 981 -               seq_printf(seqf, "%4d %7d %s %lu %lu %llu "
 982 -                          "%u %lu %lu %llu %u %u %u %u\n",
 983 +               seq_printf(seqf, "%4d %7d %s %lu %lu %lu "
 984 +                          "%u %lu %lu %lu %u %u %u %u\n",
 985                            MAJOR(part_devt(hd)), MINOR(part_devt(hd)),
 986                            disk_name(gp, hd->partno, buf),
 987                            part_stat_read(hd, ios[READ]),
 988                            part_stat_read(hd, merges[READ]),
 989 -                          (unsigned long long)part_stat_read(hd, sectors[READ]),
 990 +                          part_stat_read(hd, sectors[READ]),
 991                            jiffies_to_msecs(part_stat_read(hd, ticks[READ])),
 992                            part_stat_read(hd, ios[WRITE]),
 993                            part_stat_read(hd, merges[WRITE]),
 994 -                          (unsigned long long)part_stat_read(hd, sectors[WRITE]),
 995 +                          part_stat_read(hd, sectors[WRITE]),
 996                            jiffies_to_msecs(part_stat_read(hd, ticks[WRITE])),
 997                            part_in_flight(hd),
 998                            jiffies_to_msecs(part_stat_read(hd, io_ticks)),
 999 diff -NurpP --minimal linux-3.0-rc5/drivers/block/Kconfig linux-3.0-rc5-vs2.3.1-pre3/drivers/block/Kconfig
1000 --- linux-3.0-rc5/drivers/block/Kconfig 2011-07-01 11:13:50.000000000 +0200
1001 +++ linux-3.0-rc5-vs2.3.1-pre3/drivers/block/Kconfig    2011-06-10 22:11:24.000000000 +0200
1002 @@ -273,6 +273,13 @@ config BLK_DEV_CRYPTOLOOP
1003
1004  source "drivers/block/drbd/Kconfig"
1005
1006 +config BLK_DEV_VROOT
1007 +       tristate "Virtual Root device support"
1008 +       depends on QUOTACTL
1009 +       ---help---
1010 +         Saying Y here will allow you to use quota/fs ioctls on a shared
1011 +         partition within a virtual server without compromising security.
1012 +
1013  config BLK_DEV_NBD
1014         tristate "Network block device support"
1015         depends on NET
1016 diff -NurpP --minimal linux-3.0-rc5/drivers/block/Makefile linux-3.0-rc5-vs2.3.1-pre3/drivers/block/Makefile
1017 --- linux-3.0-rc5/drivers/block/Makefile        2011-07-01 11:13:50.000000000 +0200
1018 +++ linux-3.0-rc5-vs2.3.1-pre3/drivers/block/Makefile   2011-06-10 22:11:24.000000000 +0200
1019 @@ -34,6 +34,7 @@ obj-$(CONFIG_VIODASD)         += viodasd.o
1020  obj-$(CONFIG_BLK_DEV_SX8)      += sx8.o
1021  obj-$(CONFIG_BLK_DEV_UB)       += ub.o
1022  obj-$(CONFIG_BLK_DEV_HD)       += hd.o
1023 +obj-$(CONFIG_BLK_DEV_VROOT)    += vroot.o
1024
1025  obj-$(CONFIG_XEN_BLKDEV_FRONTEND)      += xen-blkfront.o
1026  obj-$(CONFIG_XEN_BLKDEV_BACKEND)       += xen-blkback/
1027 diff -NurpP --minimal linux-3.0-rc5/drivers/block/loop.c linux-3.0-rc5-vs2.3.1-pre3/drivers/block/loop.c
1028 --- linux-3.0-rc5/drivers/block/loop.c  2011-07-01 11:13:50.000000000 +0200
1029 +++ linux-3.0-rc5-vs2.3.1-pre3/drivers/block/loop.c     2011-06-10 22:11:24.000000000 +0200
1030 @@ -75,6 +75,7 @@
1031  #include <linux/kthread.h>
1032  #include <linux/splice.h>
1033  #include <linux/sysfs.h>
1034 +#include <linux/vs_context.h>
1035
1036  #include <asm/uaccess.h>
1037
1038 @@ -891,6 +892,7 @@ static int loop_set_fd(struct loop_devic
1039         lo->lo_blocksize = lo_blocksize;
1040         lo->lo_device = bdev;
1041         lo->lo_flags = lo_flags;
1042 +       lo->lo_xid = vx_current_xid();
1043         lo->lo_backing_file = file;
1044         lo->transfer = transfer_none;
1045         lo->ioctl = NULL;
1046 @@ -1019,6 +1021,7 @@ static int loop_clr_fd(struct loop_devic
1047         lo->lo_encrypt_key_size = 0;
1048         lo->lo_flags = 0;
1049         lo->lo_thread = NULL;
1050 +       lo->lo_xid = 0;
1051         memset(lo->lo_encrypt_key, 0, LO_KEY_SIZE);
1052         memset(lo->lo_crypt_name, 0, LO_NAME_SIZE);
1053         memset(lo->lo_file_name, 0, LO_NAME_SIZE);
1054 @@ -1057,7 +1060,7 @@ loop_set_status(struct loop_device *lo,
1055
1056         if (lo->lo_encrypt_key_size &&
1057             lo->lo_key_owner != uid &&
1058 -           !capable(CAP_SYS_ADMIN))
1059 +           !vx_capable(CAP_SYS_ADMIN, VXC_ADMIN_CLOOP))
1060                 return -EPERM;
1061         if (lo->lo_state != Lo_bound)
1062                 return -ENXIO;
1063 @@ -1141,7 +1144,8 @@ loop_get_status(struct loop_device *lo,
1064         memcpy(info->lo_crypt_name, lo->lo_crypt_name, LO_NAME_SIZE);
1065         info->lo_encrypt_type =
1066                 lo->lo_encryption ? lo->lo_encryption->number : 0;
1067 -       if (lo->lo_encrypt_key_size && capable(CAP_SYS_ADMIN)) {
1068 +       if (lo->lo_encrypt_key_size &&
1069 +               vx_capable(CAP_SYS_ADMIN, VXC_ADMIN_CLOOP)) {
1070                 info->lo_encrypt_key_size = lo->lo_encrypt_key_size;
1071                 memcpy(info->lo_encrypt_key, lo->lo_encrypt_key,
1072                        lo->lo_encrypt_key_size);
1073 @@ -1487,6 +1491,9 @@ static int lo_open(struct block_device *
1074  {
1075         struct loop_device *lo = bdev->bd_disk->private_data;
1076
1077 +       if (!vx_check(lo->lo_xid, VS_IDENT|VS_HOSTID|VS_ADMIN_P))
1078 +               return -EACCES;
1079 +
1080         mutex_lock(&lo->lo_ctl_mutex);
1081         lo->lo_refcnt++;
1082         mutex_unlock(&lo->lo_ctl_mutex);
1083 diff -NurpP --minimal linux-3.0-rc5/drivers/block/vroot.c linux-3.0-rc5-vs2.3.1-pre3/drivers/block/vroot.c
1084 --- linux-3.0-rc5/drivers/block/vroot.c 1970-01-01 01:00:00.000000000 +0100
1085 +++ linux-3.0-rc5-vs2.3.1-pre3/drivers/block/vroot.c    2011-06-10 22:11:24.000000000 +0200
1086 @@ -0,0 +1,292 @@
1087 +/*
1088 + *  linux/drivers/block/vroot.c
1089 + *
1090 + *  written by Herbert Pötzl, 9/11/2002
1091 + *  ported to 2.6.10 by Herbert Pötzl, 30/12/2004
1092 + *
1093 + *  based on the loop.c code by Theodore Ts'o.
1094 + *
1095 + * Copyright (C) 2002-2007 by Herbert Pötzl.
1096 + * Redistribution of this file is permitted under the
1097 + * GNU General Public License.
1098 + *
1099 + */
1100 +
1101 +#include <linux/module.h>
1102 +#include <linux/moduleparam.h>
1103 +#include <linux/file.h>
1104 +#include <linux/major.h>
1105 +#include <linux/blkdev.h>
1106 +#include <linux/slab.h>
1107 +
1108 +#include <linux/vroot.h>
1109 +#include <linux/vs_context.h>
1110 +
1111 +
1112 +static int max_vroot = 8;
1113 +
1114 +static struct vroot_device *vroot_dev;
1115 +static struct gendisk **disks;
1116 +
1117 +
1118 +static int vroot_set_dev(
1119 +       struct vroot_device *vr,
1120 +       struct block_device *bdev,
1121 +       unsigned int arg)
1122 +{
1123 +       struct block_device *real_bdev;
1124 +       struct file *file;
1125 +       struct inode *inode;
1126 +       int error;
1127 +
1128 +       error = -EBUSY;
1129 +       if (vr->vr_state != Vr_unbound)
1130 +               goto out;
1131 +
1132 +       error = -EBADF;
1133 +       file = fget(arg);
1134 +       if (!file)
1135 +               goto out;
1136 +
1137 +       error = -EINVAL;
1138 +       inode = file->f_dentry->d_inode;
1139 +
1140 +
1141 +       if (S_ISBLK(inode->i_mode)) {
1142 +               real_bdev = inode->i_bdev;
1143 +               vr->vr_device = real_bdev;
1144 +               __iget(real_bdev->bd_inode);
1145 +       } else
1146 +               goto out_fput;
1147 +
1148 +       vxdprintk(VXD_CBIT(misc, 0),
1149 +               "vroot[%d]_set_dev: dev=" VXF_DEV,
1150 +               vr->vr_number, VXD_DEV(real_bdev));
1151 +
1152 +       vr->vr_state = Vr_bound;
1153 +       error = 0;
1154 +
1155 + out_fput:
1156 +       fput(file);
1157 + out:
1158 +       return error;
1159 +}
1160 +
1161 +static int vroot_clr_dev(
1162 +       struct vroot_device *vr,
1163 +       struct block_device *bdev)
1164 +{
1165 +       struct block_device *real_bdev;
1166 +
1167 +       if (vr->vr_state != Vr_bound)
1168 +               return -ENXIO;
1169 +       if (vr->vr_refcnt > 1)  /* we needed one fd for the ioctl */
1170 +               return -EBUSY;
1171 +
1172 +       real_bdev = vr->vr_device;
1173 +
1174 +       vxdprintk(VXD_CBIT(misc, 0),
1175 +               "vroot[%d]_clr_dev: dev=" VXF_DEV,
1176 +               vr->vr_number, VXD_DEV(real_bdev));
1177 +
1178 +       bdput(real_bdev);
1179 +       vr->vr_state = Vr_unbound;
1180 +       vr->vr_device = NULL;
1181 +       return 0;
1182 +}
1183 +
1184 +
1185 +static int vr_ioctl(struct block_device *bdev, fmode_t mode,
1186 +       unsigned int cmd, unsigned long arg)
1187 +{
1188 +       struct vroot_device *vr = bdev->bd_disk->private_data;
1189 +       int err;
1190 +
1191 +       down(&vr->vr_ctl_mutex);
1192 +       switch (cmd) {
1193 +       case VROOT_SET_DEV:
1194 +               err = vroot_set_dev(vr, bdev, arg);
1195 +               break;
1196 +       case VROOT_CLR_DEV:
1197 +               err = vroot_clr_dev(vr, bdev);
1198 +               break;
1199 +       default:
1200 +               err = -EINVAL;
1201 +               break;
1202 +       }
1203 +       up(&vr->vr_ctl_mutex);
1204 +       return err;
1205 +}
1206 +
1207 +static int vr_open(struct block_device *bdev, fmode_t mode)
1208 +{
1209 +       struct vroot_device *vr = bdev->bd_disk->private_data;
1210 +
1211 +       down(&vr->vr_ctl_mutex);
1212 +       vr->vr_refcnt++;
1213 +       up(&vr->vr_ctl_mutex);
1214 +       return 0;
1215 +}
1216 +
1217 +static int vr_release(struct gendisk *disk, fmode_t mode)
1218 +{
1219 +       struct vroot_device *vr = disk->private_data;
1220 +
1221 +       down(&vr->vr_ctl_mutex);
1222 +       --vr->vr_refcnt;
1223 +       up(&vr->vr_ctl_mutex);
1224 +       return 0;
1225 +}
1226 +
1227 +static struct block_device_operations vr_fops = {
1228 +       .owner =        THIS_MODULE,
1229 +       .open =         vr_open,
1230 +       .release =      vr_release,
1231 +       .ioctl =        vr_ioctl,
1232 +};
1233 +
1234 +static int vroot_make_request(struct request_queue *q, struct bio *bio)
1235 +{
1236 +       printk("vroot_make_request %p, %p\n", q, bio);
1237 +       bio_io_error(bio);
1238 +       return 0;
1239 +}
1240 +
1241 +struct block_device *__vroot_get_real_bdev(struct block_device *bdev)
1242 +{
1243 +       struct inode *inode = bdev->bd_inode;
1244 +       struct vroot_device *vr;
1245 +       struct block_device *real_bdev;
1246 +       int minor = iminor(inode);
1247 +
1248 +       vr = &vroot_dev[minor];
1249 +       real_bdev = vr->vr_device;
1250 +
1251 +       vxdprintk(VXD_CBIT(misc, 0),
1252 +               "vroot[%d]_get_real_bdev: dev=" VXF_DEV,
1253 +               vr->vr_number, VXD_DEV(real_bdev));
1254 +
1255 +       if (vr->vr_state != Vr_bound)
1256 +               return ERR_PTR(-ENXIO);
1257 +
1258 +       __iget(real_bdev->bd_inode);
1259 +       return real_bdev;
1260 +}
1261 +
1262 +
1263 +
1264 +/*
1265 + * And now the modules code and kernel interface.
1266 + */
1267 +
1268 +module_param(max_vroot, int, 0);
1269 +
1270 +MODULE_PARM_DESC(max_vroot, "Maximum number of vroot devices (1-256)");
1271 +MODULE_LICENSE("GPL");
1272 +MODULE_ALIAS_BLOCKDEV_MAJOR(VROOT_MAJOR);
1273 +
1274 +MODULE_AUTHOR ("Herbert Pötzl");
1275 +MODULE_DESCRIPTION ("Virtual Root Device Mapper");
1276 +
1277 +
1278 +int __init vroot_init(void)
1279 +{
1280 +       int err, i;
1281 +
1282 +       if (max_vroot < 1 || max_vroot > 256) {
1283 +               max_vroot = MAX_VROOT_DEFAULT;
1284 +               printk(KERN_WARNING "vroot: invalid max_vroot "
1285 +                       "(must be between 1 and 256), "
1286 +                       "using default (%d)\n", max_vroot);
1287 +       }
1288 +
1289 +       if (register_blkdev(VROOT_MAJOR, "vroot"))
1290 +               return -EIO;
1291 +
1292 +       err = -ENOMEM;
1293 +       vroot_dev = kmalloc(max_vroot * sizeof(struct vroot_device), GFP_KERNEL);
1294 +       if (!vroot_dev)
1295 +               goto out_mem1;
1296 +       memset(vroot_dev, 0, max_vroot * sizeof(struct vroot_device));
1297 +
1298 +       disks = kmalloc(max_vroot * sizeof(struct gendisk *), GFP_KERNEL);
1299 +       if (!disks)
1300 +               goto out_mem2;
1301 +
1302 +       for (i = 0; i < max_vroot; i++) {
1303 +               disks[i] = alloc_disk(1);
1304 +               if (!disks[i])
1305 +                       goto out_mem3;
1306 +               disks[i]->queue = blk_alloc_queue(GFP_KERNEL);
1307 +               if (!disks[i]->queue)
1308 +                       goto out_mem3;
1309 +               blk_queue_make_request(disks[i]->queue, vroot_make_request);
1310 +       }
1311 +
1312 +       for (i = 0; i < max_vroot; i++) {
1313 +               struct vroot_device *vr = &vroot_dev[i];
1314 +               struct gendisk *disk = disks[i];
1315 +
1316 +               memset(vr, 0, sizeof(*vr));
1317 +               sema_init(&vr->vr_ctl_mutex, 1);
1318 +               vr->vr_number = i;
1319 +               disk->major = VROOT_MAJOR;
1320 +               disk->first_minor = i;
1321 +               disk->fops = &vr_fops;
1322 +               sprintf(disk->disk_name, "vroot%d", i);
1323 +               disk->private_data = vr;
1324 +       }
1325 +
1326 +       err = register_vroot_grb(&__vroot_get_real_bdev);
1327 +       if (err)
1328 +               goto out_mem3;
1329 +
1330 +       for (i = 0; i < max_vroot; i++)
1331 +               add_disk(disks[i]);
1332 +       printk(KERN_INFO "vroot: loaded (max %d devices)\n", max_vroot);
1333 +       return 0;
1334 +
1335 +out_mem3:
1336 +       while (i--)
1337 +               put_disk(disks[i]);
1338 +       kfree(disks);
1339 +out_mem2:
1340 +       kfree(vroot_dev);
1341 +out_mem1:
1342 +       unregister_blkdev(VROOT_MAJOR, "vroot");
1343 +       printk(KERN_ERR "vroot: ran out of memory\n");
1344 +       return err;
1345 +}
1346 +
1347 +void vroot_exit(void)
1348 +{
1349 +       int i;
1350 +
1351 +       if (unregister_vroot_grb(&__vroot_get_real_bdev))
1352 +               printk(KERN_WARNING "vroot: cannot unregister grb\n");
1353 +
1354 +       for (i = 0; i < max_vroot; i++) {
1355 +               del_gendisk(disks[i]);
1356 +               put_disk(disks[i]);
1357 +       }
1358 +       unregister_blkdev(VROOT_MAJOR, "vroot");
1359 +
1360 +       kfree(disks);
1361 +       kfree(vroot_dev);
1362 +}
1363 +
1364 +module_init(vroot_init);
1365 +module_exit(vroot_exit);
1366 +
1367 +#ifndef MODULE
1368 +
1369 +static int __init max_vroot_setup(char *str)
1370 +{
1371 +       max_vroot = simple_strtol(str, NULL, 0);
1372 +       return 1;
1373 +}
1374 +
1375 +__setup("max_vroot=", max_vroot_setup);
1376 +
1377 +#endif
1378 +
1379 diff -NurpP --minimal linux-3.0-rc5/drivers/gpu/drm/radeon/r100_reg_safe.h linux-3.0-rc5-vs2.3.1-pre3/drivers/gpu/drm/radeon/r100_reg_safe.h
1380 --- linux-3.0-rc5/drivers/gpu/drm/radeon/r100_reg_safe.h        1970-01-01 01:00:00.000000000 +0100
1381 +++ linux-3.0-rc5-vs2.3.1-pre3/drivers/gpu/drm/radeon/r100_reg_safe.h   2011-06-10 22:11:24.000000000 +0200
1382 @@ -0,0 +1,28 @@
1383 +static const unsigned r100_reg_safe_bm[102] = {
1384 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1385 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1386 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1387 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1388 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1389 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1390 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1391 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1392 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1393 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1394 +       0x17FF1FFF, 0xFFFFFFFC, 0xFFFFFFFF, 0xFF30FFBF,
1395 +       0xFFFFFFF8, 0xC3E6FFFF, 0xFFFFF6DF, 0xFFFFFFFF,
1396 +       0xFFFFFFCF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1397 +       0xFFFFFF9F, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1398 +       0x38E7FE1F, 0xFFC3FF8E, 0x7FF8FFFF, 0xFFFF803C,
1399 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1400 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFEFFFF, 0xFFFFFFFF,
1401 +       0x00000000, 0xFFFFFFFD, 0xFFFFFFFF, 0xFFFFFFFF,
1402 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1403 +       0xFFFFFFFF, 0xFFFCFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1404 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1405 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1406 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1407 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1408 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1409 +       0xFFFFFFFF, 0xFFFFFFEF,
1410 +};
1411 diff -NurpP --minimal linux-3.0-rc5/drivers/gpu/drm/radeon/r200_reg_safe.h linux-3.0-rc5-vs2.3.1-pre3/drivers/gpu/drm/radeon/r200_reg_safe.h
1412 --- linux-3.0-rc5/drivers/gpu/drm/radeon/r200_reg_safe.h        1970-01-01 01:00:00.000000000 +0100
1413 +++ linux-3.0-rc5-vs2.3.1-pre3/drivers/gpu/drm/radeon/r200_reg_safe.h   2011-06-10 22:11:24.000000000 +0200
1414 @@ -0,0 +1,28 @@
1415 +static const unsigned r200_reg_safe_bm[102] = {
1416 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1417 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1418 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1419 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1420 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1421 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1422 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1423 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1424 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1425 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1426 +       0x17FF1FFF, 0xFFFFFFFC, 0xFFFFFFFF, 0xFF30FFBF,
1427 +       0xFFFFFFF8, 0xC3E6FFFF, 0xFFFFF6DF, 0xFFFFFFFF,
1428 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1429 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1430 +       0xFFE7FE1F, 0xF003FFFF, 0x7EFFFFFF, 0xFFFF803C,
1431 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1432 +       0xFFFFFFFF, 0xFFFFEFCE, 0xFFFEFFFF, 0xFFFFFFFE,
1433 +       0x020E0FF0, 0xFFCC83FD, 0xFFFFFFFF, 0xFFFFFFFF,
1434 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1435 +       0xFFFBFFFF, 0xEFFCFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1436 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1437 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1438 +       0xDFDFDFDF, 0x3FFDDFDF, 0xFFFFFFFF, 0xFFFFFF7F,
1439 +       0xFFFFFFFF, 0x00FFFFFF, 0x00000000, 0x00000000,
1440 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1441 +       0xFFFFFE3F, 0xFFFFFFEF,
1442 +};
1443 diff -NurpP --minimal linux-3.0-rc5/drivers/gpu/drm/radeon/r300_reg_safe.h linux-3.0-rc5-vs2.3.1-pre3/drivers/gpu/drm/radeon/r300_reg_safe.h
1444 --- linux-3.0-rc5/drivers/gpu/drm/radeon/r300_reg_safe.h        1970-01-01 01:00:00.000000000 +0100
1445 +++ linux-3.0-rc5-vs2.3.1-pre3/drivers/gpu/drm/radeon/r300_reg_safe.h   2011-06-10 22:11:24.000000000 +0200
1446 @@ -0,0 +1,42 @@
1447 +static const unsigned r300_reg_safe_bm[159] = {
1448 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1449 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1450 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1451 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1452 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1453 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1454 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1455 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1456 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1457 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1458 +       0x17FF1FFF, 0xFFFFFFFC, 0xFFFFFFFF, 0xFF30FFBF,
1459 +       0xFFFFFFF8, 0xC3E6FFFF, 0xFFFFF6DF, 0xFFFFFFFF,
1460 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1461 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1462 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFF03F,
1463 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1464 +       0xFFFFFFFF, 0xFFFFEFCE, 0xF00EBFFF, 0x007C0000,
1465 +       0xF0000078, 0xFF000009, 0xFFFFFFFF, 0xFFFFFFFF,
1466 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1467 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1468 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1469 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1470 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1471 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1472 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1473 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1474 +       0xFFFFF7FF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1475 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1476 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1477 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1478 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1479 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1480 +       0xFFFFFC78, 0xFFFFFFFF, 0xFFFFFFFE, 0xFFFFFFFF,
1481 +       0x38FF8F50, 0xFFF88082, 0xF000000C, 0xFAE009FF,
1482 +       0x0000FFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0x00000000,
1483 +       0x00000000, 0x0000C100, 0x00000000, 0x00000000,
1484 +       0x00000000, 0x00000000, 0x00000000, 0x00000000,
1485 +       0x00000000, 0xFFFF0000, 0xFFFFFFFF, 0xFF80FFFF,
1486 +       0x00000000, 0x00000000, 0x00000000, 0x00000000,
1487 +       0x0003FC01, 0xFFFFFCF8, 0xFF800B19,
1488 +};
1489 diff -NurpP --minimal linux-3.0-rc5/drivers/gpu/drm/radeon/r420_reg_safe.h linux-3.0-rc5-vs2.3.1-pre3/drivers/gpu/drm/radeon/r420_reg_safe.h
1490 --- linux-3.0-rc5/drivers/gpu/drm/radeon/r420_reg_safe.h        1970-01-01 01:00:00.000000000 +0100
1491 +++ linux-3.0-rc5-vs2.3.1-pre3/drivers/gpu/drm/radeon/r420_reg_safe.h   2011-06-10 22:11:24.000000000 +0200
1492 @@ -0,0 +1,42 @@
1493 +static const unsigned r420_reg_safe_bm[159] = {
1494 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1495 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1496 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1497 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1498 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1499 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1500 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1501 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1502 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1503 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1504 +       0x17FF1FFF, 0xFFFFFFFC, 0xFFFFFFFF, 0xFF30FFBF,
1505 +       0xFFFFFFF8, 0xC3E6FFFF, 0xFFFFF6DF, 0xFFFFFFFF,
1506 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1507 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1508 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFF03F,
1509 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1510 +       0xFFFFFFFF, 0xFFFFEFCE, 0xF00EBFFF, 0x007C0000,
1511 +       0xF0000078, 0xFF000009, 0xFFFFFFFF, 0xFFFFFFFF,
1512 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1513 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1514 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1515 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1516 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1517 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1518 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1519 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1520 +       0xFFFFF7FF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1521 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1522 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1523 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1524 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1525 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1526 +       0xFFFFFC78, 0xFFFFFFFF, 0xFFFFFFFE, 0xFFFFFFFF,
1527 +       0x38FF8F50, 0xFFF88082, 0xF000000C, 0xFAE009FF,
1528 +       0x0000FFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0x00000000,
1529 +       0x00000000, 0x00000100, 0x00000000, 0x00000000,
1530 +       0x00000000, 0x00000000, 0x00000000, 0x00000000,
1531 +       0x00000000, 0x00000000, 0x00000000, 0xFF800000,
1532 +       0x00000000, 0x00000000, 0x00000000, 0x00000000,
1533 +       0x0003FC01, 0xFFFFFCF8, 0xFF800B19,
1534 +};
1535 diff -NurpP --minimal linux-3.0-rc5/drivers/gpu/drm/radeon/rn50_reg_safe.h linux-3.0-rc5-vs2.3.1-pre3/drivers/gpu/drm/radeon/rn50_reg_safe.h
1536 --- linux-3.0-rc5/drivers/gpu/drm/radeon/rn50_reg_safe.h        1970-01-01 01:00:00.000000000 +0100
1537 +++ linux-3.0-rc5-vs2.3.1-pre3/drivers/gpu/drm/radeon/rn50_reg_safe.h   2011-06-10 22:11:24.000000000 +0200
1538 @@ -0,0 +1,28 @@
1539 +static const unsigned rn50_reg_safe_bm[102] = {
1540 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1541 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1542 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1543 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1544 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1545 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1546 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1547 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1548 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1549 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1550 +       0x17FF1FFF, 0xFFFFFFFC, 0xFFFFFFFF, 0xFF30FFBF,
1551 +       0xFFFFFFF8, 0xC3E6FFFF, 0xFFFFF6DF, 0xFFFFFFFF,
1552 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1553 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1554 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1555 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1556 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1557 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1558 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1559 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1560 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1561 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1562 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1563 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1564 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1565 +       0xFFFFFFFF, 0xFFFFFFFF,
1566 +};
1567 diff -NurpP --minimal linux-3.0-rc5/drivers/gpu/drm/radeon/rs600_reg_safe.h linux-3.0-rc5-vs2.3.1-pre3/drivers/gpu/drm/radeon/rs600_reg_safe.h
1568 --- linux-3.0-rc5/drivers/gpu/drm/radeon/rs600_reg_safe.h       1970-01-01 01:00:00.000000000 +0100
1569 +++ linux-3.0-rc5-vs2.3.1-pre3/drivers/gpu/drm/radeon/rs600_reg_safe.h  2011-06-10 22:11:24.000000000 +0200
1570 @@ -0,0 +1,57 @@
1571 +static const unsigned rs600_reg_safe_bm[219] = {
1572 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1573 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1574 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1575 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1576 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1577 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1578 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1579 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1580 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1581 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1582 +       0x17FF1FFF, 0xFFFFFFFC, 0xFFFFFFFF, 0xFF30FFBF,
1583 +       0xFFFFFFF8, 0xC3E6FFFF, 0xFFFFF6DF, 0xFFFFFFFF,
1584 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1585 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1586 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFF03F,
1587 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1588 +       0xFFFFFFFF, 0xFFFFEFCE, 0xF00EBFFF, 0x007C0000,
1589 +       0xF0000078, 0xFF000009, 0xFFFFFFFF, 0xFFFFFFFF,
1590 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1591 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1592 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1593 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1594 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1595 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1596 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1597 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1598 +       0xFFFFF7FF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1599 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1600 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1601 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1602 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1603 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1604 +       0xFFFFFC78, 0xFFFFFFFF, 0xFFFFFFFE, 0xFFFFFFFF,
1605 +       0x38FF8F50, 0xFFF88082, 0xF000000C, 0xFAE009FF,
1606 +       0x0000FFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0x00000000,
1607 +       0x00000000, 0x00000100, 0x00000000, 0x00000000,
1608 +       0x00000000, 0x00000000, 0x00000000, 0x00000000,
1609 +       0x00000000, 0x00000000, 0x00000000, 0xFF800000,
1610 +       0x00000000, 0x00000000, 0x00000000, 0x00000000,
1611 +       0x0003FC01, 0xFFFFFCF8, 0xFF800B19, 0xFFFFFFFF,
1612 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1613 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1614 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1615 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1616 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1617 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1618 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1619 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1620 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1621 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1622 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1623 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1624 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1625 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1626 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1627 +};
1628 diff -NurpP --minimal linux-3.0-rc5/drivers/gpu/drm/radeon/rv515_reg_safe.h linux-3.0-rc5-vs2.3.1-pre3/drivers/gpu/drm/radeon/rv515_reg_safe.h
1629 --- linux-3.0-rc5/drivers/gpu/drm/radeon/rv515_reg_safe.h       1970-01-01 01:00:00.000000000 +0100
1630 +++ linux-3.0-rc5-vs2.3.1-pre3/drivers/gpu/drm/radeon/rv515_reg_safe.h  2011-06-10 22:11:24.000000000 +0200
1631 @@ -0,0 +1,57 @@
1632 +static const unsigned rv515_reg_safe_bm[219] = {
1633 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1634 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1635 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1636 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1637 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1638 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1639 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1640 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1641 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1642 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1643 +       0x17FF1FFF, 0xFFFFFFFC, 0xFFFFFFFF, 0xFF30FFBF,
1644 +       0xFFFFFFF8, 0xC3E6FFFF, 0xFFFFF6DF, 0xFFFFFFFF,
1645 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1646 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1647 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFF03F,
1648 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1649 +       0xFFFFFFFF, 0xFFFFEFCE, 0xF00EBFFF, 0x007C0000,
1650 +       0xF0000038, 0xFF000009, 0xFFFFFFFF, 0xFFFFFFFF,
1651 +       0xFFFFFFFF, 0xFFFFFFFF, 0x00000000, 0xFFFFFFFF,
1652 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1653 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1654 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1655 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1656 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1657 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1658 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1659 +       0xFFFFF7FF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1660 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1661 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1662 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1663 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1664 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1665 +       0x1FFFF878, 0xFFFFE000, 0xFFFFFE1E, 0xFFFFFFFF,
1666 +       0x388F8F50, 0xFFF88082, 0xFF0000FC, 0xFAE009FF,
1667 +       0x0000FFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0x00000000,
1668 +       0xFFFF8CFC, 0xFFFFC1FF, 0xFFFFFFFF, 0xFFFFFFFF,
1669 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1670 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFF80FFFF,
1671 +       0x00000000, 0x00000000, 0x00000000, 0x00000000,
1672 +       0x0003FC01, 0x3FFFFCF8, 0xFF800B19, 0xFFDFFFFF,
1673 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1674 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1675 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1676 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1677 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1678 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1679 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1680 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1681 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1682 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1683 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1684 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1685 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1686 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1687 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1688 +};
1689 diff -NurpP --minimal linux-3.0-rc5/drivers/infiniband/core/addr.c linux-3.0-rc5-vs2.3.1-pre3/drivers/infiniband/core/addr.c
1690 --- linux-3.0-rc5/drivers/infiniband/core/addr.c        2011-07-01 11:13:52.000000000 +0200
1691 +++ linux-3.0-rc5-vs2.3.1-pre3/drivers/infiniband/core/addr.c   2011-06-16 14:16:51.000000000 +0200
1692 @@ -252,7 +252,7 @@ static int addr6_resolve(struct sockaddr
1693
1694         if (ipv6_addr_any(&fl6.saddr)) {
1695                 ret = ipv6_dev_get_saddr(&init_net, ip6_dst_idev(dst)->dev,
1696 -                                        &fl6.daddr, 0, &fl6.saddr);
1697 +                                        &fl6.daddr, 0, &fl6.saddr, NULL);
1698                 if (ret)
1699                         goto put;
1700
1701 diff -NurpP --minimal linux-3.0-rc5/drivers/infiniband/hw/ipath/ipath_user_pages.c linux-3.0-rc5-vs2.3.1-pre3/drivers/infiniband/hw/ipath/ipath_user_pages.c
1702 --- linux-3.0-rc5/drivers/infiniband/hw/ipath/ipath_user_pages.c        2011-05-22 16:17:16.000000000 +0200
1703 +++ linux-3.0-rc5-vs2.3.1-pre3/drivers/infiniband/hw/ipath/ipath_user_pages.c   2011-06-10 22:11:24.000000000 +0200
1704 @@ -35,6 +35,7 @@
1705  #include <linux/device.h>
1706  #include <linux/slab.h>
1707  #include <linux/sched.h>
1708 +#include <linux/vs_memory.h>
1709
1710  #include "ipath_kernel.h"
1711
1712 diff -NurpP --minimal linux-3.0-rc5/drivers/md/dm-ioctl.c linux-3.0-rc5-vs2.3.1-pre3/drivers/md/dm-ioctl.c
1713 --- linux-3.0-rc5/drivers/md/dm-ioctl.c 2011-05-22 16:17:18.000000000 +0200
1714 +++ linux-3.0-rc5-vs2.3.1-pre3/drivers/md/dm-ioctl.c    2011-06-10 22:11:24.000000000 +0200
1715 @@ -16,6 +16,7 @@
1716  #include <linux/dm-ioctl.h>
1717  #include <linux/hdreg.h>
1718  #include <linux/compat.h>
1719 +#include <linux/vs_context.h>
1720
1721  #include <asm/uaccess.h>
1722
1723 @@ -106,7 +107,8 @@ static struct hash_cell *__get_name_cell
1724         unsigned int h = hash_str(str);
1725
1726         list_for_each_entry (hc, _name_buckets + h, name_list)
1727 -               if (!strcmp(hc->name, str)) {
1728 +               if (vx_check(dm_get_xid(hc->md), VS_WATCH_P | VS_IDENT) &&
1729 +                       !strcmp(hc->name, str)) {
1730                         dm_get(hc->md);
1731                         return hc;
1732                 }
1733 @@ -120,7 +122,8 @@ static struct hash_cell *__get_uuid_cell
1734         unsigned int h = hash_str(str);
1735
1736         list_for_each_entry (hc, _uuid_buckets + h, uuid_list)
1737 -               if (!strcmp(hc->uuid, str)) {
1738 +               if (vx_check(dm_get_xid(hc->md), VS_WATCH_P | VS_IDENT) &&
1739 +                       !strcmp(hc->uuid, str)) {
1740                         dm_get(hc->md);
1741                         return hc;
1742                 }
1743 @@ -427,6 +430,9 @@ typedef int (*ioctl_fn)(struct dm_ioctl
1744
1745  static int remove_all(struct dm_ioctl *param, size_t param_size)
1746  {
1747 +       if (!vx_check(0, VS_ADMIN))
1748 +               return -EPERM;
1749 +
1750         dm_hash_remove_all(1);
1751         param->data_size = 0;
1752         return 0;
1753 @@ -474,6 +480,8 @@ static int list_devices(struct dm_ioctl
1754          */
1755         for (i = 0; i < NUM_BUCKETS; i++) {
1756                 list_for_each_entry (hc, _name_buckets + i, name_list) {
1757 +                       if (!vx_check(dm_get_xid(hc->md), VS_WATCH_P | VS_IDENT))
1758 +                               continue;
1759                         needed += sizeof(struct dm_name_list);
1760                         needed += strlen(hc->name) + 1;
1761                         needed += ALIGN_MASK;
1762 @@ -497,6 +505,8 @@ static int list_devices(struct dm_ioctl
1763          */
1764         for (i = 0; i < NUM_BUCKETS; i++) {
1765                 list_for_each_entry (hc, _name_buckets + i, name_list) {
1766 +                       if (!vx_check(dm_get_xid(hc->md), VS_WATCH_P | VS_IDENT))
1767 +                               continue;
1768                         if (old_nl)
1769                                 old_nl->next = (uint32_t) ((void *) nl -
1770                                                            (void *) old_nl);
1771 @@ -731,10 +741,11 @@ static struct hash_cell *__find_device_h
1772         if (!md)
1773                 goto out;
1774
1775 -       mdptr = dm_get_mdptr(md);
1776 +       if (vx_check(dm_get_xid(md), VS_WATCH_P | VS_IDENT))
1777 +               mdptr = dm_get_mdptr(md);
1778 +
1779         if (!mdptr)
1780                 dm_put(md);
1781 -
1782  out:
1783         return mdptr;
1784  }
1785 @@ -1577,8 +1588,8 @@ static int ctl_ioctl(uint command, struc
1786         ioctl_fn fn = NULL;
1787         size_t input_param_size;
1788
1789 -       /* only root can play with this */
1790 -       if (!capable(CAP_SYS_ADMIN))
1791 +       /* only root and certain contexts can play with this */
1792 +       if (!vx_capable(CAP_SYS_ADMIN, VXC_ADMIN_MAPPER))
1793                 return -EACCES;
1794
1795         if (_IOC_TYPE(command) != DM_IOCTL)
1796 diff -NurpP --minimal linux-3.0-rc5/drivers/md/dm.c linux-3.0-rc5-vs2.3.1-pre3/drivers/md/dm.c
1797 --- linux-3.0-rc5/drivers/md/dm.c       2011-05-22 16:17:18.000000000 +0200
1798 +++ linux-3.0-rc5-vs2.3.1-pre3/drivers/md/dm.c  2011-06-10 22:11:24.000000000 +0200
1799 @@ -20,6 +20,7 @@
1800  #include <linux/idr.h>
1801  #include <linux/hdreg.h>
1802  #include <linux/delay.h>
1803 +#include <linux/vs_base.h>
1804
1805  #include <trace/events/block.h>
1806
1807 @@ -119,6 +120,7 @@ struct mapped_device {
1808         rwlock_t map_lock;
1809         atomic_t holders;
1810         atomic_t open_count;
1811 +       xid_t xid;
1812
1813         unsigned long flags;
1814
1815 @@ -326,6 +328,7 @@ int dm_deleting_md(struct mapped_device
1816  static int dm_blk_open(struct block_device *bdev, fmode_t mode)
1817  {
1818         struct mapped_device *md;
1819 +       int ret = -ENXIO;
1820
1821         spin_lock(&_minor_lock);
1822
1823 @@ -334,18 +337,19 @@ static int dm_blk_open(struct block_devi
1824                 goto out;
1825
1826         if (test_bit(DMF_FREEING, &md->flags) ||
1827 -           dm_deleting_md(md)) {
1828 -               md = NULL;
1829 +           dm_deleting_md(md))
1830 +               goto out;
1831 +
1832 +       ret = -EACCES;
1833 +       if (!vx_check(md->xid, VS_IDENT|VS_HOSTID))
1834                 goto out;
1835 -       }
1836
1837         dm_get(md);
1838         atomic_inc(&md->open_count);
1839 -
1840 +       ret = 0;
1841  out:
1842         spin_unlock(&_minor_lock);
1843 -
1844 -       return md ? 0 : -ENXIO;
1845 +       return ret;
1846  }
1847
1848  static int dm_blk_close(struct gendisk *disk, fmode_t mode)
1849 @@ -566,6 +570,14 @@ int dm_set_geometry(struct mapped_device
1850         return 0;
1851  }
1852
1853 +/*
1854 + * Get the xid associated with a dm device
1855 + */
1856 +xid_t dm_get_xid(struct mapped_device *md)
1857 +{
1858 +       return md->xid;
1859 +}
1860 +
1861  /*-----------------------------------------------------------------
1862   * CRUD START:
1863   *   A more elegant soln is in the works that uses the queue
1864 @@ -1841,6 +1853,7 @@ static struct mapped_device *alloc_dev(i
1865         INIT_LIST_HEAD(&md->uevent_list);
1866         spin_lock_init(&md->uevent_lock);
1867
1868 +       md->xid = vx_current_xid();
1869         md->queue = blk_alloc_queue(GFP_KERNEL);
1870         if (!md->queue)
1871                 goto bad_queue;
1872 diff -NurpP --minimal linux-3.0-rc5/drivers/md/dm.h linux-3.0-rc5-vs2.3.1-pre3/drivers/md/dm.h
1873 --- linux-3.0-rc5/drivers/md/dm.h       2011-05-22 16:17:18.000000000 +0200
1874 +++ linux-3.0-rc5-vs2.3.1-pre3/drivers/md/dm.h  2011-06-10 22:11:24.000000000 +0200
1875 @@ -41,6 +41,8 @@ struct dm_dev_internal {
1876  struct dm_table;
1877  struct dm_md_mempools;
1878
1879 +xid_t dm_get_xid(struct mapped_device *md);
1880 +
1881  /*-----------------------------------------------------------------
1882   * Internal table functions.
1883   *---------------------------------------------------------------*/
1884 diff -NurpP --minimal linux-3.0-rc5/drivers/net/tun.c linux-3.0-rc5-vs2.3.1-pre3/drivers/net/tun.c
1885 --- linux-3.0-rc5/drivers/net/tun.c     2011-07-01 11:14:01.000000000 +0200
1886 +++ linux-3.0-rc5-vs2.3.1-pre3/drivers/net/tun.c        2011-06-22 12:39:14.000000000 +0200
1887 @@ -64,6 +64,7 @@
1888  #include <linux/nsproxy.h>
1889  #include <linux/virtio_net.h>
1890  #include <linux/rcupdate.h>
1891 +#include <linux/vs_network.h>
1892  #include <net/net_namespace.h>
1893  #include <net/netns/generic.h>
1894  #include <net/rtnetlink.h>
1895 @@ -121,6 +122,7 @@ struct tun_struct {
1896         unsigned int            flags;
1897         uid_t                   owner;
1898         gid_t                   group;
1899 +       nid_t                   nid;
1900
1901         struct net_device       *dev;
1902         u32                     set_features;
1903 @@ -904,6 +906,7 @@ static void tun_setup(struct net_device
1904
1905         tun->owner = -1;
1906         tun->group = -1;
1907 +       tun->nid = current->nid;
1908
1909         dev->ethtool_ops = &tun_ethtool_ops;
1910         dev->destructor = tun_free_netdev;
1911 @@ -1054,7 +1057,7 @@ static int tun_set_iff(struct net *net,
1912
1913                 if (((tun->owner != -1 && cred->euid != tun->owner) ||
1914                      (tun->group != -1 && !in_egroup_p(tun->group))) &&
1915 -                   !capable(CAP_NET_ADMIN))
1916 +               !cap_raised(current_cap(), CAP_NET_ADMIN))
1917                         return -EPERM;
1918                 err = security_tun_dev_attach(tun->socket.sk);
1919                 if (err < 0)
1920 @@ -1068,7 +1071,7 @@ static int tun_set_iff(struct net *net,
1921                 char *name;
1922                 unsigned long flags = 0;
1923
1924 -               if (!capable(CAP_NET_ADMIN))
1925 +               if (!nx_capable(CAP_NET_ADMIN, NXC_TUN_CREATE))
1926                         return -EPERM;
1927                 err = security_tun_dev_create();
1928                 if (err < 0)
1929 @@ -1136,6 +1139,9 @@ static int tun_set_iff(struct net *net,
1930
1931                 sk->sk_destruct = tun_sock_destruct;
1932
1933 +               if (!nx_check(tun->nid, VS_IDENT | VS_HOSTID | VS_ADMIN_P))
1934 +                       return -EPERM;
1935 +
1936                 err = tun_attach(tun, file);
1937                 if (err < 0)
1938                         goto failed;
1939 @@ -1317,6 +1323,16 @@ static long __tun_chr_ioctl(struct file
1940                 tun_debug(KERN_INFO, tun, "group set to %d\n", tun->group);
1941                 break;
1942
1943 +       case TUNSETNID:
1944 +               if (!capable(CAP_CONTEXT))
1945 +                       return -EPERM;
1946 +
1947 +               /* Set nid owner of the device */
1948 +               tun->nid = (nid_t) arg;
1949 +
1950 +               tun_debug(KERN_INFO, tun, "nid owner set to %u\n", tun->nid);
1951 +               break;
1952 +
1953         case TUNSETLINK:
1954                 /* Only allow setting the type when the interface is down */
1955                 if (tun->dev->flags & IFF_UP) {
1956 diff -NurpP --minimal linux-3.0-rc5/drivers/tty/sysrq.c linux-3.0-rc5-vs2.3.1-pre3/drivers/tty/sysrq.c
1957 --- linux-3.0-rc5/drivers/tty/sysrq.c   2011-05-22 16:17:44.000000000 +0200
1958 +++ linux-3.0-rc5-vs2.3.1-pre3/drivers/tty/sysrq.c      2011-06-10 22:11:24.000000000 +0200
1959 @@ -41,6 +41,7 @@
1960  #include <linux/oom.h>
1961  #include <linux/slab.h>
1962  #include <linux/input.h>
1963 +#include <linux/vserver/debug.h>
1964
1965  #include <asm/ptrace.h>
1966  #include <asm/irq_regs.h>
1967 @@ -395,6 +396,21 @@ static struct sysrq_key_op sysrq_unrt_op
1968         .enable_mask    = SYSRQ_ENABLE_RTNICE,
1969  };
1970
1971 +
1972 +#ifdef CONFIG_VSERVER_DEBUG
1973 +static void sysrq_handle_vxinfo(int key)
1974 +{
1975 +       dump_vx_info_inactive((key == 'x') ? 0 : 1);
1976 +}
1977 +
1978 +static struct sysrq_key_op sysrq_showvxinfo_op = {
1979 +       .handler        = sysrq_handle_vxinfo,
1980 +       .help_msg       = "conteXt",
1981 +       .action_msg     = "Show Context Info",
1982 +       .enable_mask    = SYSRQ_ENABLE_DUMP,
1983 +};
1984 +#endif
1985 +
1986  /* Key Operations table and lock */
1987  static DEFINE_SPINLOCK(sysrq_key_table_lock);
1988
1989 @@ -449,7 +465,11 @@ static struct sysrq_key_op *sysrq_key_ta
1990         NULL,                           /* v */
1991         &sysrq_showstate_blocked_op,    /* w */
1992         /* x: May be registered on ppc/powerpc for xmon */
1993 +#ifdef CONFIG_VSERVER_DEBUG
1994 +       &sysrq_showvxinfo_op,           /* x */
1995 +#else
1996         NULL,                           /* x */
1997 +#endif
1998         /* y: May be registered on sparc64 for global register dump */
1999         NULL,                           /* y */
2000         &sysrq_ftrace_dump_op,          /* z */
2001 @@ -464,6 +484,8 @@ static int sysrq_key_table_key2index(int
2002                 retval = key - '0';
2003         else if ((key >= 'a') && (key <= 'z'))
2004                 retval = key + 10 - 'a';
2005 +       else if ((key >= 'A') && (key <= 'Z'))
2006 +               retval = key + 10 - 'A';
2007         else
2008                 retval = -1;
2009         return retval;
2010 diff -NurpP --minimal linux-3.0-rc5/drivers/tty/tty_io.c linux-3.0-rc5-vs2.3.1-pre3/drivers/tty/tty_io.c
2011 --- linux-3.0-rc5/drivers/tty/tty_io.c  2011-07-01 11:14:14.000000000 +0200
2012 +++ linux-3.0-rc5-vs2.3.1-pre3/drivers/tty/tty_io.c     2011-06-10 22:11:24.000000000 +0200
2013 @@ -104,6 +104,7 @@
2014
2015  #include <linux/kmod.h>
2016  #include <linux/nsproxy.h>
2017 +#include <linux/vs_pid.h>
2018
2019  #undef TTY_DEBUG_HANGUP
2020
2021 @@ -2057,7 +2058,8 @@ static int tiocsti(struct tty_struct *tt
2022         char ch, mbz = 0;
2023         struct tty_ldisc *ld;
2024
2025 -       if ((current->signal->tty != tty) && !capable(CAP_SYS_ADMIN))
2026 +       if (((current->signal->tty != tty) &&
2027 +               !vx_capable(CAP_SYS_ADMIN, VXC_TIOCSTI)))
2028                 return -EPERM;
2029         if (get_user(ch, p))
2030                 return -EFAULT;
2031 @@ -2345,6 +2347,7 @@ static int tiocspgrp(struct tty_struct *
2032                 return -ENOTTY;
2033         if (get_user(pgrp_nr, p))
2034                 return -EFAULT;
2035 +       pgrp_nr = vx_rmap_pid(pgrp_nr);
2036         if (pgrp_nr < 0)
2037                 return -EINVAL;
2038         rcu_read_lock();
2039 diff -NurpP --minimal linux-3.0-rc5/fs/attr.c linux-3.0-rc5-vs2.3.1-pre3/fs/attr.c
2040 --- linux-3.0-rc5/fs/attr.c     2011-07-01 11:14:16.000000000 +0200
2041 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/attr.c        2011-06-10 22:11:24.000000000 +0200
2042 @@ -13,6 +13,9 @@
2043  #include <linux/fsnotify.h>
2044  #include <linux/fcntl.h>
2045  #include <linux/security.h>
2046 +#include <linux/proc_fs.h>
2047 +#include <linux/devpts_fs.h>
2048 +#include <linux/vs_tag.h>
2049
2050  /**
2051   * inode_change_ok - check if attribute changes to an inode are allowed
2052 @@ -73,6 +76,10 @@ int inode_change_ok(const struct inode *
2053                         return -EPERM;
2054         }
2055
2056 +       /* check for inode tag permission */
2057 +       if (dx_permission(inode, MAY_WRITE))
2058 +               return -EACCES;
2059 +
2060         return 0;
2061  }
2062  EXPORT_SYMBOL(inode_change_ok);
2063 @@ -143,6 +150,8 @@ void setattr_copy(struct inode *inode, c
2064                 inode->i_uid = attr->ia_uid;
2065         if (ia_valid & ATTR_GID)
2066                 inode->i_gid = attr->ia_gid;
2067 +       if ((ia_valid & ATTR_TAG) && IS_TAGGED(inode))
2068 +               inode->i_tag = attr->ia_tag;
2069         if (ia_valid & ATTR_ATIME)
2070                 inode->i_atime = timespec_trunc(attr->ia_atime,
2071                                                 inode->i_sb->s_time_gran);
2072 @@ -170,7 +179,8 @@ int notify_change(struct dentry * dentry
2073         struct timespec now;
2074         unsigned int ia_valid = attr->ia_valid;
2075
2076 -       if (ia_valid & (ATTR_MODE | ATTR_UID | ATTR_GID | ATTR_TIMES_SET)) {
2077 +       if (ia_valid & (ATTR_MODE | ATTR_UID | ATTR_GID |
2078 +               ATTR_TAG | ATTR_TIMES_SET)) {
2079                 if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
2080                         return -EPERM;
2081         }
2082 diff -NurpP --minimal linux-3.0-rc5/fs/binfmt_aout.c linux-3.0-rc5-vs2.3.1-pre3/fs/binfmt_aout.c
2083 --- linux-3.0-rc5/fs/binfmt_aout.c      2010-10-21 13:07:47.000000000 +0200
2084 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/binfmt_aout.c 2011-06-10 22:11:24.000000000 +0200
2085 @@ -25,6 +25,7 @@
2086  #include <linux/init.h>
2087  #include <linux/coredump.h>
2088  #include <linux/slab.h>
2089 +#include <linux/vs_memory.h>
2090
2091  #include <asm/system.h>
2092  #include <asm/uaccess.h>
2093 diff -NurpP --minimal linux-3.0-rc5/fs/binfmt_elf.c linux-3.0-rc5-vs2.3.1-pre3/fs/binfmt_elf.c
2094 --- linux-3.0-rc5/fs/binfmt_elf.c       2011-05-22 16:17:48.000000000 +0200
2095 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/binfmt_elf.c  2011-06-10 22:11:24.000000000 +0200
2096 @@ -32,6 +32,7 @@
2097  #include <linux/elf.h>
2098  #include <linux/utsname.h>
2099  #include <linux/coredump.h>
2100 +#include <linux/vs_memory.h>
2101  #include <asm/uaccess.h>
2102  #include <asm/param.h>
2103  #include <asm/page.h>
2104 diff -NurpP --minimal linux-3.0-rc5/fs/binfmt_flat.c linux-3.0-rc5-vs2.3.1-pre3/fs/binfmt_flat.c
2105 --- linux-3.0-rc5/fs/binfmt_flat.c      2011-07-01 11:14:16.000000000 +0200
2106 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/binfmt_flat.c 2011-06-10 22:11:24.000000000 +0200
2107 @@ -35,6 +35,7 @@
2108  #include <linux/init.h>
2109  #include <linux/flat.h>
2110  #include <linux/syscalls.h>
2111 +#include <linux/vs_memory.h>
2112
2113  #include <asm/byteorder.h>
2114  #include <asm/system.h>
2115 diff -NurpP --minimal linux-3.0-rc5/fs/binfmt_som.c linux-3.0-rc5-vs2.3.1-pre3/fs/binfmt_som.c
2116 --- linux-3.0-rc5/fs/binfmt_som.c       2010-02-25 11:52:04.000000000 +0100
2117 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/binfmt_som.c  2011-06-10 22:11:24.000000000 +0200
2118 @@ -28,6 +28,7 @@
2119  #include <linux/shm.h>
2120  #include <linux/personality.h>
2121  #include <linux/init.h>
2122 +#include <linux/vs_memory.h>
2123
2124  #include <asm/uaccess.h>
2125  #include <asm/pgtable.h>
2126 diff -NurpP --minimal linux-3.0-rc5/fs/block_dev.c linux-3.0-rc5-vs2.3.1-pre3/fs/block_dev.c
2127 --- linux-3.0-rc5/fs/block_dev.c        2011-07-01 11:14:16.000000000 +0200
2128 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/block_dev.c   2011-07-01 11:35:34.000000000 +0200
2129 @@ -25,6 +25,7 @@
2130  #include <linux/namei.h>
2131  #include <linux/log2.h>
2132  #include <linux/kmemleak.h>
2133 +#include <linux/vs_device.h>
2134  #include <asm/uaccess.h>
2135  #include "internal.h"
2136
2137 @@ -553,6 +554,7 @@ struct block_device *bdget(dev_t dev)
2138                 bdev->bd_invalidated = 0;
2139                 inode->i_mode = S_IFBLK;
2140                 inode->i_rdev = dev;
2141 +               inode->i_mdev = dev;
2142                 inode->i_bdev = bdev;
2143                 inode->i_data.a_ops = &def_blk_aops;
2144                 mapping_set_gfp_mask(&inode->i_data, GFP_USER);
2145 @@ -599,6 +601,11 @@ EXPORT_SYMBOL(bdput);
2146  static struct block_device *bd_acquire(struct inode *inode)
2147  {
2148         struct block_device *bdev;
2149 +       dev_t mdev;
2150 +
2151 +       if (!vs_map_blkdev(inode->i_rdev, &mdev, DATTR_OPEN))
2152 +               return NULL;
2153 +       inode->i_mdev = mdev;
2154
2155         spin_lock(&bdev_lock);
2156         bdev = inode->i_bdev;
2157 @@ -609,7 +616,7 @@ static struct block_device *bd_acquire(s
2158         }
2159         spin_unlock(&bdev_lock);
2160
2161 -       bdev = bdget(inode->i_rdev);
2162 +       bdev = bdget(mdev);
2163         if (bdev) {
2164                 spin_lock(&bdev_lock);
2165                 if (!inode->i_bdev) {
2166 diff -NurpP --minimal linux-3.0-rc5/fs/btrfs/ctree.h linux-3.0-rc5-vs2.3.1-pre3/fs/btrfs/ctree.h
2167 --- linux-3.0-rc5/fs/btrfs/ctree.h      2011-07-01 11:14:16.000000000 +0200
2168 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/btrfs/ctree.h 2011-07-01 11:35:34.000000000 +0200
2169 @@ -600,11 +600,14 @@ struct btrfs_inode_item {
2170         /* modification sequence number for NFS */
2171         __le64 sequence;
2172
2173 +       __le16 tag;
2174         /*
2175          * a little future expansion, for more than this we can
2176          * just grow the inode item and version it
2177          */
2178 -       __le64 reserved[4];
2179 +       __le16 reserved16;
2180 +       __le32 reserved32;
2181 +       __le64 reserved[3];
2182         struct btrfs_timespec atime;
2183         struct btrfs_timespec ctime;
2184         struct btrfs_timespec mtime;
2185 @@ -1354,6 +1357,8 @@ struct btrfs_ioctl_defrag_range_args {
2186  #define BTRFS_MOUNT_AUTO_DEFRAG                (1 << 16)
2187  #define BTRFS_MOUNT_INODE_MAP_CACHE    (1 << 17)
2188
2189 +#define BTRFS_MOUNT_TAGGED             (1 << 24)
2190 +
2191  #define btrfs_clear_opt(o, opt)                ((o) &= ~BTRFS_MOUNT_##opt)
2192  #define btrfs_set_opt(o, opt)          ((o) |= BTRFS_MOUNT_##opt)
2193  #define btrfs_test_opt(root, opt)      ((root)->fs_info->mount_opt & \
2194 @@ -1563,6 +1568,7 @@ BTRFS_SETGET_FUNCS(inode_block_group, st
2195  BTRFS_SETGET_FUNCS(inode_nlink, struct btrfs_inode_item, nlink, 32);
2196  BTRFS_SETGET_FUNCS(inode_uid, struct btrfs_inode_item, uid, 32);
2197  BTRFS_SETGET_FUNCS(inode_gid, struct btrfs_inode_item, gid, 32);
2198 +BTRFS_SETGET_FUNCS(inode_tag, struct btrfs_inode_item, tag, 16);
2199  BTRFS_SETGET_FUNCS(inode_mode, struct btrfs_inode_item, mode, 32);
2200  BTRFS_SETGET_FUNCS(inode_rdev, struct btrfs_inode_item, rdev, 64);
2201  BTRFS_SETGET_FUNCS(inode_flags, struct btrfs_inode_item, flags, 64);
2202 @@ -1616,6 +1622,10 @@ BTRFS_SETGET_FUNCS(extent_flags, struct
2203
2204  BTRFS_SETGET_FUNCS(extent_refs_v0, struct btrfs_extent_item_v0, refs, 32);
2205
2206 +#define BTRFS_INODE_IXUNLINK           (1 << 24)
2207 +#define BTRFS_INODE_BARRIER            (1 << 25)
2208 +#define BTRFS_INODE_COW                        (1 << 26)
2209 +
2210
2211  BTRFS_SETGET_FUNCS(tree_block_level, struct btrfs_tree_block_info, level, 8);
2212
2213 @@ -2590,6 +2600,7 @@ extern const struct dentry_operations bt
2214  long btrfs_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
2215  void btrfs_update_iflags(struct inode *inode);
2216  void btrfs_inherit_iflags(struct inode *inode, struct inode *dir);
2217 +int btrfs_sync_flags(struct inode *inode, int, int);
2218  int btrfs_defrag_file(struct inode *inode, struct file *file,
2219                       struct btrfs_ioctl_defrag_range_args *range,
2220                       u64 newer_than, unsigned long max_pages);
2221 diff -NurpP --minimal linux-3.0-rc5/fs/btrfs/disk-io.c linux-3.0-rc5-vs2.3.1-pre3/fs/btrfs/disk-io.c
2222 --- linux-3.0-rc5/fs/btrfs/disk-io.c    2011-07-01 11:14:16.000000000 +0200
2223 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/btrfs/disk-io.c       2011-06-22 12:39:15.000000000 +0200
2224 @@ -1765,6 +1765,9 @@ struct btrfs_root *open_ctree(struct sup
2225                 goto fail_alloc;
2226         }
2227
2228 +       if (btrfs_test_opt(tree_root, TAGGED))
2229 +               sb->s_flags |= MS_TAGGED;
2230 +
2231         features = btrfs_super_incompat_flags(disk_super) &
2232                 ~BTRFS_FEATURE_INCOMPAT_SUPP;
2233         if (features) {
2234 diff -NurpP --minimal linux-3.0-rc5/fs/btrfs/inode.c linux-3.0-rc5-vs2.3.1-pre3/fs/btrfs/inode.c
2235 --- linux-3.0-rc5/fs/btrfs/inode.c      2011-07-01 11:14:17.000000000 +0200
2236 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/btrfs/inode.c 2011-07-01 11:35:34.000000000 +0200
2237 @@ -38,6 +38,7 @@
2238  #include <linux/falloc.h>
2239  #include <linux/slab.h>
2240  #include <linux/ratelimit.h>
2241 +#include <linux/vs_tag.h>
2242  #include "compat.h"
2243  #include "ctree.h"
2244  #include "disk-io.h"
2245 @@ -2508,6 +2509,8 @@ static void btrfs_read_locked_inode(stru
2246         struct btrfs_key location;
2247         int maybe_acls;
2248         u32 rdev;
2249 +       uid_t uid;
2250 +       gid_t gid;
2251         int ret;
2252         bool filled = false;
2253
2254 @@ -2540,8 +2543,13 @@ static void btrfs_read_locked_inode(stru
2255
2256         inode->i_mode = btrfs_inode_mode(leaf, inode_item);
2257         inode->i_nlink = btrfs_inode_nlink(leaf, inode_item);
2258 -       inode->i_uid = btrfs_inode_uid(leaf, inode_item);
2259 -       inode->i_gid = btrfs_inode_gid(leaf, inode_item);
2260 +
2261 +       uid = btrfs_inode_uid(leaf, inode_item);
2262 +       gid = btrfs_inode_gid(leaf, inode_item);
2263 +       inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
2264 +       inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
2265 +       inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid,
2266 +               btrfs_inode_tag(leaf, inode_item));
2267         btrfs_i_size_write(inode, btrfs_inode_size(leaf, inode_item));
2268
2269         tspec = btrfs_inode_atime(inode_item);
2270 @@ -2624,6 +2632,9 @@ static void fill_inode_item(struct btrfs
2271                             struct btrfs_inode_item *item,
2272                             struct inode *inode)
2273  {
2274 +       uid_t uid = TAGINO_UID(DX_TAG(inode), inode->i_uid, inode->i_tag);
2275 +       gid_t gid = TAGINO_GID(DX_TAG(inode), inode->i_gid, inode->i_tag);
2276 +
2277         if (!leaf->map_token)
2278                 map_private_extent_buffer(leaf, (unsigned long)item,
2279                                           sizeof(struct btrfs_inode_item),
2280 @@ -2631,8 +2642,11 @@ static void fill_inode_item(struct btrfs
2281                                           &leaf->map_start, &leaf->map_len,
2282                                           KM_USER1);
2283
2284 -       btrfs_set_inode_uid(leaf, item, inode->i_uid);
2285 -       btrfs_set_inode_gid(leaf, item, inode->i_gid);
2286 +       btrfs_set_inode_uid(leaf, item, uid);
2287 +       btrfs_set_inode_gid(leaf, item, gid);
2288 +#ifdef CONFIG_TAGGING_INTERN
2289 +       btrfs_set_inode_tag(leaf, item, inode->i_tag);
2290 +#endif
2291         btrfs_set_inode_size(leaf, item, BTRFS_I(inode)->disk_i_size);
2292         btrfs_set_inode_mode(leaf, item, inode->i_mode);
2293         btrfs_set_inode_nlink(leaf, item, inode->i_nlink);
2294 @@ -7357,6 +7371,7 @@ static const struct inode_operations btr
2295         .listxattr      = btrfs_listxattr,
2296         .removexattr    = btrfs_removexattr,
2297         .permission     = btrfs_permission,
2298 +       .sync_flags     = btrfs_sync_flags,
2299  };
2300  static const struct inode_operations btrfs_dir_ro_inode_operations = {
2301         .lookup         = btrfs_lookup,
2302 @@ -7429,6 +7444,7 @@ static const struct inode_operations btr
2303         .removexattr    = btrfs_removexattr,
2304         .permission     = btrfs_permission,
2305         .fiemap         = btrfs_fiemap,
2306 +       .sync_flags     = btrfs_sync_flags,
2307  };
2308  static const struct inode_operations btrfs_special_inode_operations = {
2309         .getattr        = btrfs_getattr,
2310 diff -NurpP --minimal linux-3.0-rc5/fs/btrfs/ioctl.c linux-3.0-rc5-vs2.3.1-pre3/fs/btrfs/ioctl.c
2311 --- linux-3.0-rc5/fs/btrfs/ioctl.c      2011-07-01 11:14:17.000000000 +0200
2312 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/btrfs/ioctl.c 2011-06-22 12:39:15.000000000 +0200
2313 @@ -70,10 +70,13 @@ static unsigned int btrfs_flags_to_ioctl
2314  {
2315         unsigned int iflags = 0;
2316
2317 -       if (flags & BTRFS_INODE_SYNC)
2318 -               iflags |= FS_SYNC_FL;
2319         if (flags & BTRFS_INODE_IMMUTABLE)
2320                 iflags |= FS_IMMUTABLE_FL;
2321 +       if (flags & BTRFS_INODE_IXUNLINK)
2322 +               iflags |= FS_IXUNLINK_FL;
2323 +
2324 +       if (flags & BTRFS_INODE_SYNC)
2325 +               iflags |= FS_SYNC_FL;
2326         if (flags & BTRFS_INODE_APPEND)
2327                 iflags |= FS_APPEND_FL;
2328         if (flags & BTRFS_INODE_NODUMP)
2329 @@ -90,28 +93,78 @@ static unsigned int btrfs_flags_to_ioctl
2330         else if (flags & BTRFS_INODE_NOCOMPRESS)
2331                 iflags |= FS_NOCOMP_FL;
2332
2333 +       if (flags & BTRFS_INODE_BARRIER)
2334 +               iflags |= FS_BARRIER_FL;
2335 +       if (flags & BTRFS_INODE_COW)
2336 +               iflags |= FS_COW_FL;
2337         return iflags;
2338  }
2339
2340  /*
2341 - * Update inode->i_flags based on the btrfs internal flags.
2342 + * Update inode->i_(v)flags based on the btrfs internal flags.
2343   */
2344  void btrfs_update_iflags(struct inode *inode)
2345  {
2346         struct btrfs_inode *ip = BTRFS_I(inode);
2347
2348 -       inode->i_flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC);
2349 +       inode->i_flags &= ~(S_IMMUTABLE | S_IXUNLINK |
2350 +               S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
2351
2352 -       if (ip->flags & BTRFS_INODE_SYNC)
2353 -               inode->i_flags |= S_SYNC;
2354         if (ip->flags & BTRFS_INODE_IMMUTABLE)
2355                 inode->i_flags |= S_IMMUTABLE;
2356 +       if (ip->flags & BTRFS_INODE_IXUNLINK)
2357 +               inode->i_flags |= S_IXUNLINK;
2358 +
2359 +       if (ip->flags & BTRFS_INODE_SYNC)
2360 +               inode->i_flags |= S_SYNC;
2361         if (ip->flags & BTRFS_INODE_APPEND)
2362                 inode->i_flags |= S_APPEND;
2363         if (ip->flags & BTRFS_INODE_NOATIME)
2364                 inode->i_flags |= S_NOATIME;
2365         if (ip->flags & BTRFS_INODE_DIRSYNC)
2366                 inode->i_flags |= S_DIRSYNC;
2367 +
2368 +       inode->i_vflags &= ~(V_BARRIER | V_COW);
2369 +
2370 +       if (ip->flags & BTRFS_INODE_BARRIER)
2371 +               inode->i_vflags |= V_BARRIER;
2372 +       if (ip->flags & BTRFS_INODE_COW)
2373 +               inode->i_vflags |= V_COW;
2374 +}
2375 +
2376 +/*
2377 + * Update btrfs internal flags from inode->i_(v)flags.
2378 + */
2379 +void btrfs_update_flags(struct inode *inode)
2380 +{
2381 +       struct btrfs_inode *ip = BTRFS_I(inode);
2382 +
2383 +       unsigned int flags = inode->i_flags;
2384 +       unsigned int vflags = inode->i_vflags;
2385 +
2386 +       ip->flags &= ~(BTRFS_INODE_SYNC | BTRFS_INODE_APPEND |
2387 +                       BTRFS_INODE_IMMUTABLE | BTRFS_INODE_IXUNLINK |
2388 +                       BTRFS_INODE_NOATIME | BTRFS_INODE_DIRSYNC |
2389 +                       BTRFS_INODE_BARRIER | BTRFS_INODE_COW);
2390 +
2391 +       if (flags & S_IMMUTABLE)
2392 +               ip->flags |= BTRFS_INODE_IMMUTABLE;
2393 +       if (flags & S_IXUNLINK)
2394 +               ip->flags |= BTRFS_INODE_IXUNLINK;
2395 +
2396 +       if (flags & S_SYNC)
2397 +               ip->flags |= BTRFS_INODE_SYNC;
2398 +       if (flags & S_APPEND)
2399 +               ip->flags |= BTRFS_INODE_APPEND;
2400 +       if (flags & S_NOATIME)
2401 +               ip->flags |= BTRFS_INODE_NOATIME;
2402 +       if (flags & S_DIRSYNC)
2403 +               ip->flags |= BTRFS_INODE_DIRSYNC;
2404 +
2405 +       if (vflags & V_BARRIER)
2406 +               ip->flags |= BTRFS_INODE_BARRIER;
2407 +       if (vflags & V_COW)
2408 +               ip->flags |= BTRFS_INODE_COW;
2409  }
2410
2411  /*
2412 @@ -129,7 +182,7 @@ void btrfs_inherit_iflags(struct inode *
2413         flags = BTRFS_I(dir)->flags;
2414
2415         if (S_ISREG(inode->i_mode))
2416 -               flags &= ~BTRFS_INODE_DIRSYNC;
2417 +               flags &= ~(BTRFS_INODE_DIRSYNC | BTRFS_INODE_BARRIER);
2418         else if (!S_ISDIR(inode->i_mode))
2419                 flags &= (BTRFS_INODE_NODUMP | BTRFS_INODE_NOATIME);
2420
2421 @@ -137,6 +190,30 @@ void btrfs_inherit_iflags(struct inode *
2422         btrfs_update_iflags(inode);
2423  }
2424
2425 +int btrfs_sync_flags(struct inode *inode, int flags, int vflags)
2426 +{
2427 +       struct btrfs_inode *ip = BTRFS_I(inode);
2428 +       struct btrfs_root *root = ip->root;
2429 +       struct btrfs_trans_handle *trans;
2430 +       int ret;
2431 +
2432 +       trans = btrfs_join_transaction(root);
2433 +       BUG_ON(!trans);
2434 +
2435 +       inode->i_flags = flags;
2436 +       inode->i_vflags = vflags;
2437 +       btrfs_update_flags(inode);
2438 +
2439 +       ret = btrfs_update_inode(trans, root, inode);
2440 +       BUG_ON(ret);
2441 +
2442 +       btrfs_update_iflags(inode);
2443 +       inode->i_ctime = CURRENT_TIME;
2444 +       btrfs_end_transaction(trans, root);
2445 +
2446 +       return 0;
2447 +}
2448 +
2449  static int btrfs_ioctl_getflags(struct file *file, void __user *arg)
2450  {
2451         struct btrfs_inode *ip = BTRFS_I(file->f_path.dentry->d_inode);
2452 @@ -188,7 +265,8 @@ static int btrfs_ioctl_setflags(struct f
2453
2454         flags = btrfs_mask_flags(inode->i_mode, flags);
2455         oldflags = btrfs_flags_to_ioctl(ip->flags);
2456 -       if ((flags ^ oldflags) & (FS_APPEND_FL | FS_IMMUTABLE_FL)) {
2457 +       if ((flags ^ oldflags) & (FS_APPEND_FL |
2458 +               FS_IMMUTABLE_FL | FS_IXUNLINK_FL)) {
2459                 if (!capable(CAP_LINUX_IMMUTABLE)) {
2460                         ret = -EPERM;
2461                         goto out_unlock;
2462 @@ -199,14 +277,19 @@ static int btrfs_ioctl_setflags(struct f
2463         if (ret)
2464                 goto out_unlock;
2465
2466 -       if (flags & FS_SYNC_FL)
2467 -               ip->flags |= BTRFS_INODE_SYNC;
2468 -       else
2469 -               ip->flags &= ~BTRFS_INODE_SYNC;
2470         if (flags & FS_IMMUTABLE_FL)
2471                 ip->flags |= BTRFS_INODE_IMMUTABLE;
2472         else
2473                 ip->flags &= ~BTRFS_INODE_IMMUTABLE;
2474 +       if (flags & FS_IXUNLINK_FL)
2475 +               ip->flags |= BTRFS_INODE_IXUNLINK;
2476 +       else
2477 +               ip->flags &= ~BTRFS_INODE_IXUNLINK;
2478 +
2479 +       if (flags & FS_SYNC_FL)
2480 +               ip->flags |= BTRFS_INODE_SYNC;
2481 +       else
2482 +               ip->flags &= ~BTRFS_INODE_SYNC;
2483         if (flags & FS_APPEND_FL)
2484                 ip->flags |= BTRFS_INODE_APPEND;
2485         else
2486 diff -NurpP --minimal linux-3.0-rc5/fs/btrfs/super.c linux-3.0-rc5-vs2.3.1-pre3/fs/btrfs/super.c
2487 --- linux-3.0-rc5/fs/btrfs/super.c      2011-07-01 11:14:17.000000000 +0200
2488 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/btrfs/super.c 2011-06-15 02:40:14.000000000 +0200
2489 @@ -162,7 +162,7 @@ enum {
2490         Opt_notreelog, Opt_ratio, Opt_flushoncommit, Opt_discard,
2491         Opt_space_cache, Opt_clear_cache, Opt_user_subvol_rm_allowed,
2492         Opt_enospc_debug, Opt_subvolrootid, Opt_defrag,
2493 -       Opt_inode_cache, Opt_err,
2494 +       Opt_inode_cache, Opt_tag, Opt_notag, Opt_tagid, Opt_err,
2495  };
2496
2497  static match_table_t tokens = {
2498 @@ -195,6 +195,9 @@ static match_table_t tokens = {
2499         {Opt_subvolrootid, "subvolrootid=%d"},
2500         {Opt_defrag, "autodefrag"},
2501         {Opt_inode_cache, "inode_cache"},
2502 +       {Opt_tag, "tag"},
2503 +       {Opt_notag, "notag"},
2504 +       {Opt_tagid, "tagid=%u"},
2505         {Opt_err, NULL},
2506  };
2507
2508 @@ -381,6 +384,22 @@ int btrfs_parse_options(struct btrfs_roo
2509                         printk(KERN_INFO "btrfs: enabling auto defrag");
2510                         btrfs_set_opt(info->mount_opt, AUTO_DEFRAG);
2511                         break;
2512 +#ifndef CONFIG_TAGGING_NONE
2513 +               case Opt_tag:
2514 +                       printk(KERN_INFO "btrfs: use tagging\n");
2515 +                       btrfs_set_opt(info->mount_opt, TAGGED);
2516 +                       break;
2517 +               case Opt_notag:
2518 +                       printk(KERN_INFO "btrfs: disabled tagging\n");
2519 +                       btrfs_clear_opt(info->mount_opt, TAGGED);
2520 +                       break;
2521 +#endif
2522 +#ifdef CONFIG_PROPAGATE
2523 +               case Opt_tagid:
2524 +                       /* use args[0] */
2525 +                       btrfs_set_opt(info->mount_opt, TAGGED);
2526 +                       break;
2527 +#endif
2528                 case Opt_err:
2529                         printk(KERN_INFO "btrfs: unrecognized mount option "
2530                                "'%s'\n", p);
2531 @@ -901,6 +920,12 @@ static int btrfs_remount(struct super_bl
2532         if (ret)
2533                 return -EINVAL;
2534
2535 +       if (btrfs_test_opt(root, TAGGED) && !(sb->s_flags & MS_TAGGED)) {
2536 +               printk("btrfs: %s: tagging not permitted on remount.\n",
2537 +                       sb->s_id);
2538 +               return -EINVAL;
2539 +       }
2540 +
2541         if ((*flags & MS_RDONLY) == (sb->s_flags & MS_RDONLY))
2542                 return 0;
2543
2544 diff -NurpP --minimal linux-3.0-rc5/fs/char_dev.c linux-3.0-rc5-vs2.3.1-pre3/fs/char_dev.c
2545 --- linux-3.0-rc5/fs/char_dev.c 2011-03-15 18:07:31.000000000 +0100
2546 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/char_dev.c    2011-06-10 22:11:24.000000000 +0200
2547 @@ -21,6 +21,8 @@
2548  #include <linux/mutex.h>
2549  #include <linux/backing-dev.h>
2550  #include <linux/tty.h>
2551 +#include <linux/vs_context.h>
2552 +#include <linux/vs_device.h>
2553
2554  #include "internal.h"
2555
2556 @@ -371,14 +373,21 @@ static int chrdev_open(struct inode *ino
2557         struct cdev *p;
2558         struct cdev *new = NULL;
2559         int ret = 0;
2560 +       dev_t mdev;
2561 +
2562 +       if (!vs_map_chrdev(inode->i_rdev, &mdev, DATTR_OPEN))
2563 +               return -EPERM;
2564 +       inode->i_mdev = mdev;
2565
2566         spin_lock(&cdev_lock);
2567         p = inode->i_cdev;
2568         if (!p) {
2569                 struct kobject *kobj;
2570                 int idx;
2571 +
2572                 spin_unlock(&cdev_lock);
2573 -               kobj = kobj_lookup(cdev_map, inode->i_rdev, &idx);
2574 +
2575 +               kobj = kobj_lookup(cdev_map, mdev, &idx);
2576                 if (!kobj)
2577                         return -ENXIO;
2578                 new = container_of(kobj, struct cdev, kobj);
2579 diff -NurpP --minimal linux-3.0-rc5/fs/dcache.c linux-3.0-rc5-vs2.3.1-pre3/fs/dcache.c
2580 --- linux-3.0-rc5/fs/dcache.c   2011-07-01 11:14:17.000000000 +0200
2581 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/dcache.c      2011-06-10 22:35:26.000000000 +0200
2582 @@ -36,6 +36,7 @@
2583  #include <linux/bit_spinlock.h>
2584  #include <linux/rculist_bl.h>
2585  #include <linux/prefetch.h>
2586 +#include <linux/vs_limit.h>
2587  #include "internal.h"
2588
2589  /*
2590 @@ -479,6 +480,8 @@ int d_invalidate(struct dentry * dentry)
2591                 spin_lock(&dentry->d_lock);
2592         }
2593
2594 +       vx_dentry_dec(dentry);
2595 +
2596         /*
2597          * Somebody else still using it?
2598          *
2599 @@ -506,6 +509,7 @@ EXPORT_SYMBOL(d_invalidate);
2600  static inline void __dget_dlock(struct dentry *dentry)
2601  {
2602         dentry->d_count++;
2603 +       vx_dentry_inc(dentry);
2604  }
2605
2606  static inline void __dget(struct dentry *dentry)
2607 @@ -1266,6 +1270,9 @@ struct dentry *d_alloc(struct dentry * p
2608         struct dentry *dentry;
2609         char *dname;
2610
2611 +       if (!vx_dentry_avail(1))
2612 +               return NULL;
2613 +
2614         dentry = kmem_cache_alloc(dentry_cache, GFP_KERNEL);
2615         if (!dentry)
2616                 return NULL;
2617 @@ -1288,6 +1295,7 @@ struct dentry *d_alloc(struct dentry * p
2618
2619         dentry->d_count = 1;
2620         dentry->d_flags = 0;
2621 +       vx_dentry_inc(dentry);
2622         spin_lock_init(&dentry->d_lock);
2623         seqcount_init(&dentry->d_seq);
2624         dentry->d_inode = NULL;
2625 @@ -1947,6 +1955,7 @@ struct dentry *__d_lookup(struct dentry
2626                 }
2627
2628                 dentry->d_count++;
2629 +               vx_dentry_inc(dentry);
2630                 found = dentry;
2631                 spin_unlock(&dentry->d_lock);
2632                 break;
2633 diff -NurpP --minimal linux-3.0-rc5/fs/devpts/inode.c linux-3.0-rc5-vs2.3.1-pre3/fs/devpts/inode.c
2634 --- linux-3.0-rc5/fs/devpts/inode.c     2011-05-22 16:17:50.000000000 +0200
2635 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/devpts/inode.c        2011-06-10 22:11:24.000000000 +0200
2636 @@ -25,6 +25,7 @@
2637  #include <linux/parser.h>
2638  #include <linux/fsnotify.h>
2639  #include <linux/seq_file.h>
2640 +#include <linux/vs_base.h>
2641
2642  #define DEVPTS_DEFAULT_MODE 0600
2643  /*
2644 @@ -36,6 +37,20 @@
2645  #define DEVPTS_DEFAULT_PTMX_MODE 0000
2646  #define PTMX_MINOR     2
2647
2648 +static int devpts_permission(struct inode *inode, int mask, unsigned int flags)
2649 +{
2650 +       int ret = -EACCES;
2651 +
2652 +       /* devpts is xid tagged */
2653 +       if (vx_check((xid_t)inode->i_tag, VS_WATCH_P | VS_IDENT))
2654 +               ret = generic_permission(inode, mask, flags, NULL);
2655 +       return ret;
2656 +}
2657 +
2658 +static struct inode_operations devpts_file_inode_operations = {
2659 +       .permission     = devpts_permission,
2660 +};
2661 +
2662  extern int pty_limit;                  /* Config limit on Unix98 ptys */
2663  static DEFINE_MUTEX(allocated_ptys_lock);
2664
2665 @@ -263,6 +278,34 @@ static int devpts_show_options(struct se
2666         return 0;
2667  }
2668
2669 +static int devpts_filter(struct dentry *de)
2670 +{
2671 +       xid_t xid = 0;
2672 +
2673 +       /* devpts is xid tagged */
2674 +       if (de && de->d_inode)
2675 +               xid = (xid_t)de->d_inode->i_tag;
2676 +#ifdef CONFIG_VSERVER_WARN_DEVPTS
2677 +       else
2678 +               vxwprintk_task(1, "devpts " VS_Q("%.*s") " without inode.",
2679 +                       de->d_name.len, de->d_name.name);
2680 +#endif
2681 +       return vx_check(xid, VS_WATCH_P | VS_IDENT);
2682 +}
2683 +
2684 +static int devpts_readdir(struct file * filp, void * dirent, filldir_t filldir)
2685 +{
2686 +       return dcache_readdir_filter(filp, dirent, filldir, devpts_filter);
2687 +}
2688 +
2689 +static struct file_operations devpts_dir_operations = {
2690 +       .open           = dcache_dir_open,
2691 +       .release        = dcache_dir_close,
2692 +       .llseek         = dcache_dir_lseek,
2693 +       .read           = generic_read_dir,
2694 +       .readdir        = devpts_readdir,
2695 +};
2696 +
2697  static const struct super_operations devpts_sops = {
2698         .statfs         = simple_statfs,
2699         .remount_fs     = devpts_remount,
2700 @@ -302,12 +345,15 @@ devpts_fill_super(struct super_block *s,
2701         inode = new_inode(s);
2702         if (!inode)
2703                 goto free_fsi;
2704 +
2705         inode->i_ino = 1;
2706         inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
2707         inode->i_mode = S_IFDIR | S_IRUGO | S_IXUGO | S_IWUSR;
2708         inode->i_op = &simple_dir_inode_operations;
2709 -       inode->i_fop = &simple_dir_operations;
2710 +       inode->i_fop = &devpts_dir_operations;
2711         inode->i_nlink = 2;
2712 +       /* devpts is xid tagged */
2713 +       inode->i_tag = (tag_t)vx_current_xid();
2714
2715         s->s_root = d_alloc_root(inode);
2716         if (s->s_root)
2717 @@ -494,6 +540,9 @@ int devpts_pty_new(struct inode *ptmx_in
2718         inode->i_gid = opts->setgid ? opts->gid : current_fsgid();
2719         inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
2720         init_special_inode(inode, S_IFCHR|opts->mode, device);
2721 +       /* devpts is xid tagged */
2722 +       inode->i_tag = (tag_t)vx_current_xid();
2723 +       inode->i_op = &devpts_file_inode_operations;
2724         inode->i_private = tty;
2725         tty->driver_data = inode;
2726
2727 diff -NurpP --minimal linux-3.0-rc5/fs/ext2/balloc.c linux-3.0-rc5-vs2.3.1-pre3/fs/ext2/balloc.c
2728 --- linux-3.0-rc5/fs/ext2/balloc.c      2011-05-22 16:17:51.000000000 +0200
2729 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ext2/balloc.c 2011-06-10 22:11:24.000000000 +0200
2730 @@ -701,7 +701,6 @@ ext2_try_to_allocate(struct super_block
2731                         start = 0;
2732                 end = EXT2_BLOCKS_PER_GROUP(sb);
2733         }
2734 -
2735         BUG_ON(start > EXT2_BLOCKS_PER_GROUP(sb));
2736
2737  repeat:
2738 diff -NurpP --minimal linux-3.0-rc5/fs/ext2/ext2.h linux-3.0-rc5-vs2.3.1-pre3/fs/ext2/ext2.h
2739 --- linux-3.0-rc5/fs/ext2/ext2.h        2011-05-22 16:17:51.000000000 +0200
2740 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ext2/ext2.h   2011-06-10 22:11:24.000000000 +0200
2741 @@ -126,6 +126,7 @@ extern void ext2_set_inode_flags(struct
2742  extern void ext2_get_inode_flags(struct ext2_inode_info *);
2743  extern int ext2_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
2744                        u64 start, u64 len);
2745 +extern int ext2_sync_flags(struct inode *, int, int);
2746
2747  /* ioctl.c */
2748  extern long ext2_ioctl(struct file *, unsigned int, unsigned long);
2749 diff -NurpP --minimal linux-3.0-rc5/fs/ext2/file.c linux-3.0-rc5-vs2.3.1-pre3/fs/ext2/file.c
2750 --- linux-3.0-rc5/fs/ext2/file.c        2010-08-02 16:52:48.000000000 +0200
2751 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ext2/file.c   2011-06-10 22:11:24.000000000 +0200
2752 @@ -104,4 +104,5 @@ const struct inode_operations ext2_file_
2753         .setattr        = ext2_setattr,
2754         .check_acl      = ext2_check_acl,
2755         .fiemap         = ext2_fiemap,
2756 +       .sync_flags     = ext2_sync_flags,
2757  };
2758 diff -NurpP --minimal linux-3.0-rc5/fs/ext2/ialloc.c linux-3.0-rc5-vs2.3.1-pre3/fs/ext2/ialloc.c
2759 --- linux-3.0-rc5/fs/ext2/ialloc.c      2011-05-22 16:17:51.000000000 +0200
2760 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ext2/ialloc.c 2011-06-10 22:11:24.000000000 +0200
2761 @@ -17,6 +17,7 @@
2762  #include <linux/backing-dev.h>
2763  #include <linux/buffer_head.h>
2764  #include <linux/random.h>
2765 +#include <linux/vs_tag.h>
2766  #include "ext2.h"
2767  #include "xattr.h"
2768  #include "acl.h"
2769 @@ -549,6 +550,7 @@ got:
2770                 inode->i_mode = mode;
2771                 inode->i_uid = current_fsuid();
2772                 inode->i_gid = dir->i_gid;
2773 +               inode->i_tag = dx_current_fstag(sb);
2774         } else
2775                 inode_init_owner(inode, dir, mode);
2776
2777 diff -NurpP --minimal linux-3.0-rc5/fs/ext2/inode.c linux-3.0-rc5-vs2.3.1-pre3/fs/ext2/inode.c
2778 --- linux-3.0-rc5/fs/ext2/inode.c       2011-05-22 16:17:51.000000000 +0200
2779 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ext2/inode.c  2011-06-10 22:11:24.000000000 +0200
2780 @@ -32,6 +32,7 @@
2781  #include <linux/mpage.h>
2782  #include <linux/fiemap.h>
2783  #include <linux/namei.h>
2784 +#include <linux/vs_tag.h>
2785  #include "ext2.h"
2786  #include "acl.h"
2787  #include "xip.h"
2788 @@ -1167,7 +1168,7 @@ static void ext2_truncate_blocks(struct
2789                 return;
2790         if (ext2_inode_is_fast_symlink(inode))
2791                 return;
2792 -       if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
2793 +       if (IS_APPEND(inode) || IS_IXORUNLINK(inode))
2794                 return;
2795         __ext2_truncate_blocks(inode, offset);
2796  }
2797 @@ -1256,36 +1257,61 @@ void ext2_set_inode_flags(struct inode *
2798  {
2799         unsigned int flags = EXT2_I(inode)->i_flags;
2800
2801 -       inode->i_flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC);
2802 +       inode->i_flags &= ~(S_IMMUTABLE | S_IXUNLINK |
2803 +               S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
2804 +
2805 +
2806 +       if (flags & EXT2_IMMUTABLE_FL)
2807 +               inode->i_flags |= S_IMMUTABLE;
2808 +       if (flags & EXT2_IXUNLINK_FL)
2809 +               inode->i_flags |= S_IXUNLINK;
2810 +
2811         if (flags & EXT2_SYNC_FL)
2812                 inode->i_flags |= S_SYNC;
2813         if (flags & EXT2_APPEND_FL)
2814                 inode->i_flags |= S_APPEND;
2815 -       if (flags & EXT2_IMMUTABLE_FL)
2816 -               inode->i_flags |= S_IMMUTABLE;
2817         if (flags & EXT2_NOATIME_FL)
2818                 inode->i_flags |= S_NOATIME;
2819         if (flags & EXT2_DIRSYNC_FL)
2820                 inode->i_flags |= S_DIRSYNC;
2821 +
2822 +       inode->i_vflags &= ~(V_BARRIER | V_COW);
2823 +
2824 +       if (flags & EXT2_BARRIER_FL)
2825 +               inode->i_vflags |= V_BARRIER;
2826 +       if (flags & EXT2_COW_FL)
2827 +               inode->i_vflags |= V_COW;
2828  }
2829
2830  /* Propagate flags from i_flags to EXT2_I(inode)->i_flags */
2831  void ext2_get_inode_flags(struct ext2_inode_info *ei)
2832  {
2833         unsigned int flags = ei->vfs_inode.i_flags;
2834 +       unsigned int vflags = ei->vfs_inode.i_vflags;
2835 +
2836 +       ei->i_flags &= ~(EXT2_SYNC_FL | EXT2_APPEND_FL |
2837 +                       EXT2_IMMUTABLE_FL | EXT2_IXUNLINK_FL |
2838 +                       EXT2_NOATIME_FL | EXT2_DIRSYNC_FL |
2839 +                       EXT2_BARRIER_FL | EXT2_COW_FL);
2840 +
2841 +       if (flags & S_IMMUTABLE)
2842 +               ei->i_flags |= EXT2_IMMUTABLE_FL;
2843 +       if (flags & S_IXUNLINK)
2844 +               ei->i_flags |= EXT2_IXUNLINK_FL;
2845
2846 -       ei->i_flags &= ~(EXT2_SYNC_FL|EXT2_APPEND_FL|
2847 -                       EXT2_IMMUTABLE_FL|EXT2_NOATIME_FL|EXT2_DIRSYNC_FL);
2848         if (flags & S_SYNC)
2849                 ei->i_flags |= EXT2_SYNC_FL;
2850         if (flags & S_APPEND)
2851                 ei->i_flags |= EXT2_APPEND_FL;
2852 -       if (flags & S_IMMUTABLE)
2853 -               ei->i_flags |= EXT2_IMMUTABLE_FL;
2854         if (flags & S_NOATIME)
2855                 ei->i_flags |= EXT2_NOATIME_FL;
2856         if (flags & S_DIRSYNC)
2857                 ei->i_flags |= EXT2_DIRSYNC_FL;
2858 +
2859 +       if (vflags & V_BARRIER)
2860 +               ei->i_flags |= EXT2_BARRIER_FL;
2861 +       if (vflags & V_COW)
2862 +               ei->i_flags |= EXT2_COW_FL;
2863  }
2864
2865  struct inode *ext2_iget (struct super_block *sb, unsigned long ino)
2866 @@ -1295,6 +1321,8 @@ struct inode *ext2_iget (struct super_bl
2867         struct ext2_inode *raw_inode;
2868         struct inode *inode;
2869         long ret = -EIO;
2870 +       uid_t uid;
2871 +       gid_t gid;
2872         int n;
2873
2874         inode = iget_locked(sb, ino);
2875 @@ -1313,12 +1341,17 @@ struct inode *ext2_iget (struct super_bl
2876         }
2877
2878         inode->i_mode = le16_to_cpu(raw_inode->i_mode);
2879 -       inode->i_uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
2880 -       inode->i_gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
2881 +       uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
2882 +       gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
2883         if (!(test_opt (inode->i_sb, NO_UID32))) {
2884 -               inode->i_uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
2885 -               inode->i_gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
2886 +               uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
2887 +               gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
2888         }
2889 +       inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
2890 +       inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
2891 +       inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid,
2892 +               le16_to_cpu(raw_inode->i_raw_tag));
2893 +
2894         inode->i_nlink = le16_to_cpu(raw_inode->i_links_count);
2895         inode->i_size = le32_to_cpu(raw_inode->i_size);
2896         inode->i_atime.tv_sec = (signed)le32_to_cpu(raw_inode->i_atime);
2897 @@ -1416,8 +1449,8 @@ static int __ext2_write_inode(struct ino
2898         struct ext2_inode_info *ei = EXT2_I(inode);
2899         struct super_block *sb = inode->i_sb;
2900         ino_t ino = inode->i_ino;
2901 -       uid_t uid = inode->i_uid;
2902 -       gid_t gid = inode->i_gid;
2903 +       uid_t uid = TAGINO_UID(DX_TAG(inode), inode->i_uid, inode->i_tag);
2904 +       gid_t gid = TAGINO_GID(DX_TAG(inode), inode->i_gid, inode->i_tag);
2905         struct buffer_head * bh;
2906         struct ext2_inode * raw_inode = ext2_get_inode(sb, ino, &bh);
2907         int n;
2908 @@ -1453,6 +1486,9 @@ static int __ext2_write_inode(struct ino
2909                 raw_inode->i_uid_high = 0;
2910                 raw_inode->i_gid_high = 0;
2911         }
2912 +#ifdef CONFIG_TAGGING_INTERN
2913 +       raw_inode->i_raw_tag = cpu_to_le16(inode->i_tag);
2914 +#endif
2915         raw_inode->i_links_count = cpu_to_le16(inode->i_nlink);
2916         raw_inode->i_size = cpu_to_le32(inode->i_size);
2917         raw_inode->i_atime = cpu_to_le32(inode->i_atime.tv_sec);
2918 @@ -1533,7 +1569,8 @@ int ext2_setattr(struct dentry *dentry,
2919         if (is_quota_modification(inode, iattr))
2920                 dquot_initialize(inode);
2921         if ((iattr->ia_valid & ATTR_UID && iattr->ia_uid != inode->i_uid) ||
2922 -           (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid)) {
2923 +           (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid) ||
2924 +           (iattr->ia_valid & ATTR_TAG && iattr->ia_tag != inode->i_tag)) {
2925                 error = dquot_transfer(inode, iattr);
2926                 if (error)
2927                         return error;
2928 diff -NurpP --minimal linux-3.0-rc5/fs/ext2/ioctl.c linux-3.0-rc5-vs2.3.1-pre3/fs/ext2/ioctl.c
2929 --- linux-3.0-rc5/fs/ext2/ioctl.c       2011-05-22 16:17:51.000000000 +0200
2930 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ext2/ioctl.c  2011-06-10 22:11:24.000000000 +0200
2931 @@ -17,6 +17,16 @@
2932  #include <asm/uaccess.h>
2933
2934
2935 +int ext2_sync_flags(struct inode *inode, int flags, int vflags)
2936 +{
2937 +       inode->i_flags = flags;
2938 +       inode->i_vflags = vflags;
2939 +       ext2_get_inode_flags(EXT2_I(inode));
2940 +       inode->i_ctime = CURRENT_TIME_SEC;
2941 +       mark_inode_dirty(inode);
2942 +       return 0;
2943 +}
2944 +
2945  long ext2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
2946  {
2947         struct inode *inode = filp->f_dentry->d_inode;
2948 @@ -51,6 +61,11 @@ long ext2_ioctl(struct file *filp, unsig
2949
2950                 flags = ext2_mask_flags(inode->i_mode, flags);
2951
2952 +               if (IS_BARRIER(inode)) {
2953 +                       vxwprintk_task(1, "messing with the barrier.");
2954 +                       return -EACCES;
2955 +               }
2956 +
2957                 mutex_lock(&inode->i_mutex);
2958                 /* Is it quota file? Do not allow user to mess with it */
2959                 if (IS_NOQUOTA(inode)) {
2960 @@ -66,7 +81,9 @@ long ext2_ioctl(struct file *filp, unsig
2961                  *
2962                  * This test looks nicer. Thanks to Pauline Middelink
2963                  */
2964 -               if ((flags ^ oldflags) & (EXT2_APPEND_FL | EXT2_IMMUTABLE_FL)) {
2965 +               if ((oldflags & EXT2_IMMUTABLE_FL) ||
2966 +                       ((flags ^ oldflags) & (EXT2_APPEND_FL |
2967 +                       EXT2_IMMUTABLE_FL | EXT2_IXUNLINK_FL))) {
2968                         if (!capable(CAP_LINUX_IMMUTABLE)) {
2969                                 mutex_unlock(&inode->i_mutex);
2970                                 ret = -EPERM;
2971 @@ -74,7 +91,7 @@ long ext2_ioctl(struct file *filp, unsig
2972                         }
2973                 }
2974
2975 -               flags = flags & EXT2_FL_USER_MODIFIABLE;
2976 +               flags &= EXT2_FL_USER_MODIFIABLE;
2977                 flags |= oldflags & ~EXT2_FL_USER_MODIFIABLE;
2978                 ei->i_flags = flags;
2979                 mutex_unlock(&inode->i_mutex);
2980 diff -NurpP --minimal linux-3.0-rc5/fs/ext2/namei.c linux-3.0-rc5-vs2.3.1-pre3/fs/ext2/namei.c
2981 --- linux-3.0-rc5/fs/ext2/namei.c       2011-05-22 16:17:51.000000000 +0200
2982 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ext2/namei.c  2011-06-10 22:11:24.000000000 +0200
2983 @@ -32,6 +32,7 @@
2984
2985  #include <linux/pagemap.h>
2986  #include <linux/quotaops.h>
2987 +#include <linux/vs_tag.h>
2988  #include "ext2.h"
2989  #include "xattr.h"
2990  #include "acl.h"
2991 @@ -75,6 +76,7 @@ static struct dentry *ext2_lookup(struct
2992                                 return ERR_PTR(-EIO);
2993                         } else {
2994                                 return ERR_CAST(inode);
2995 +               dx_propagate_tag(nd, inode);
2996                         }
2997                 }
2998         }
2999 @@ -413,6 +415,7 @@ const struct inode_operations ext2_dir_i
3000  #endif
3001         .setattr        = ext2_setattr,
3002         .check_acl      = ext2_check_acl,
3003 +       .sync_flags     = ext2_sync_flags,
3004  };
3005
3006  const struct inode_operations ext2_special_inode_operations = {
3007 diff -NurpP --minimal linux-3.0-rc5/fs/ext2/super.c linux-3.0-rc5-vs2.3.1-pre3/fs/ext2/super.c
3008 --- linux-3.0-rc5/fs/ext2/super.c       2011-07-01 11:14:18.000000000 +0200
3009 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ext2/super.c  2011-06-10 22:11:24.000000000 +0200
3010 @@ -394,7 +394,8 @@ enum {
3011         Opt_err_ro, Opt_nouid32, Opt_nocheck, Opt_debug,
3012         Opt_oldalloc, Opt_orlov, Opt_nobh, Opt_user_xattr, Opt_nouser_xattr,
3013         Opt_acl, Opt_noacl, Opt_xip, Opt_ignore, Opt_err, Opt_quota,
3014 -       Opt_usrquota, Opt_grpquota, Opt_reservation, Opt_noreservation
3015 +       Opt_usrquota, Opt_grpquota, Opt_reservation, Opt_noreservation,
3016 +       Opt_tag, Opt_notag, Opt_tagid
3017  };
3018
3019  static const match_table_t tokens = {
3020 @@ -422,6 +423,9 @@ static const match_table_t tokens = {
3021         {Opt_acl, "acl"},
3022         {Opt_noacl, "noacl"},
3023         {Opt_xip, "xip"},
3024 +       {Opt_tag, "tag"},
3025 +       {Opt_notag, "notag"},
3026 +       {Opt_tagid, "tagid=%u"},
3027         {Opt_grpquota, "grpquota"},
3028         {Opt_ignore, "noquota"},
3029         {Opt_quota, "quota"},
3030 @@ -492,6 +496,20 @@ static int parse_options(char *options,
3031                 case Opt_nouid32:
3032                         set_opt (sbi->s_mount_opt, NO_UID32);
3033                         break;
3034 +#ifndef CONFIG_TAGGING_NONE
3035 +               case Opt_tag:
3036 +                       set_opt (sbi->s_mount_opt, TAGGED);
3037 +                       break;
3038 +               case Opt_notag:
3039 +                       clear_opt (sbi->s_mount_opt, TAGGED);
3040 +                       break;
3041 +#endif
3042 +#ifdef CONFIG_PROPAGATE
3043 +               case Opt_tagid:
3044 +                       /* use args[0] */
3045 +                       set_opt (sbi->s_mount_opt, TAGGED);
3046 +                       break;
3047 +#endif
3048                 case Opt_nocheck:
3049                         clear_opt (sbi->s_mount_opt, CHECK);
3050                         break;
3051 @@ -850,6 +868,8 @@ static int ext2_fill_super(struct super_
3052         if (!parse_options((char *) data, sb))
3053                 goto failed_mount;
3054
3055 +       if (EXT2_SB(sb)->s_mount_opt & EXT2_MOUNT_TAGGED)
3056 +               sb->s_flags |= MS_TAGGED;
3057         sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
3058                 ((EXT2_SB(sb)->s_mount_opt & EXT2_MOUNT_POSIX_ACL) ?
3059                  MS_POSIXACL : 0);
3060 @@ -1224,6 +1244,14 @@ static int ext2_remount (struct super_bl
3061                 goto restore_opts;
3062         }
3063
3064 +       if ((sbi->s_mount_opt & EXT2_MOUNT_TAGGED) &&
3065 +               !(sb->s_flags & MS_TAGGED)) {
3066 +               printk("EXT2-fs: %s: tagging not permitted on remount.\n",
3067 +                      sb->s_id);
3068 +               err = -EINVAL;
3069 +               goto restore_opts;
3070 +       }
3071 +
3072         sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
3073                 ((sbi->s_mount_opt & EXT2_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0);
3074
3075 diff -NurpP --minimal linux-3.0-rc5/fs/ext3/file.c linux-3.0-rc5-vs2.3.1-pre3/fs/ext3/file.c
3076 --- linux-3.0-rc5/fs/ext3/file.c        2010-07-07 18:31:51.000000000 +0200
3077 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ext3/file.c   2011-06-10 22:11:24.000000000 +0200
3078 @@ -81,5 +81,6 @@ const struct inode_operations ext3_file_
3079  #endif
3080         .check_acl      = ext3_check_acl,
3081         .fiemap         = ext3_fiemap,
3082 +       .sync_flags     = ext3_sync_flags,
3083  };
3084
3085 diff -NurpP --minimal linux-3.0-rc5/fs/ext3/ialloc.c linux-3.0-rc5-vs2.3.1-pre3/fs/ext3/ialloc.c
3086 --- linux-3.0-rc5/fs/ext3/ialloc.c      2011-05-22 16:17:52.000000000 +0200
3087 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ext3/ialloc.c 2011-06-10 22:11:24.000000000 +0200
3088 @@ -23,6 +23,7 @@
3089  #include <linux/buffer_head.h>
3090  #include <linux/random.h>
3091  #include <linux/bitops.h>
3092 +#include <linux/vs_tag.h>
3093
3094  #include <asm/byteorder.h>
3095
3096 @@ -532,6 +533,7 @@ got:
3097                 inode->i_mode = mode;
3098                 inode->i_uid = current_fsuid();
3099                 inode->i_gid = dir->i_gid;
3100 +               inode->i_tag = dx_current_fstag(sb);
3101         } else
3102                 inode_init_owner(inode, dir, mode);
3103
3104 diff -NurpP --minimal linux-3.0-rc5/fs/ext3/inode.c linux-3.0-rc5-vs2.3.1-pre3/fs/ext3/inode.c
3105 --- linux-3.0-rc5/fs/ext3/inode.c       2011-07-01 11:14:18.000000000 +0200
3106 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ext3/inode.c  2011-06-10 22:11:24.000000000 +0200
3107 @@ -38,6 +38,7 @@
3108  #include <linux/bio.h>
3109  #include <linux/fiemap.h>
3110  #include <linux/namei.h>
3111 +#include <linux/vs_tag.h>
3112  #include "xattr.h"
3113  #include "acl.h"
3114
3115 @@ -2391,7 +2392,7 @@ static void ext3_free_branches(handle_t
3116
3117  int ext3_can_truncate(struct inode *inode)
3118  {
3119 -       if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
3120 +       if (IS_APPEND(inode) || IS_IXORUNLINK(inode))
3121                 return 0;
3122         if (S_ISREG(inode->i_mode))
3123                 return 1;
3124 @@ -2775,36 +2776,60 @@ void ext3_set_inode_flags(struct inode *
3125  {
3126         unsigned int flags = EXT3_I(inode)->i_flags;
3127
3128 -       inode->i_flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC);
3129 +       inode->i_flags &= ~(S_IMMUTABLE | S_IXUNLINK |
3130 +               S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
3131 +
3132 +       if (flags & EXT3_IMMUTABLE_FL)
3133 +               inode->i_flags |= S_IMMUTABLE;
3134 +       if (flags & EXT3_IXUNLINK_FL)
3135 +               inode->i_flags |= S_IXUNLINK;
3136 +
3137         if (flags & EXT3_SYNC_FL)
3138                 inode->i_flags |= S_SYNC;
3139         if (flags & EXT3_APPEND_FL)
3140                 inode->i_flags |= S_APPEND;
3141 -       if (flags & EXT3_IMMUTABLE_FL)
3142 -               inode->i_flags |= S_IMMUTABLE;
3143         if (flags & EXT3_NOATIME_FL)
3144                 inode->i_flags |= S_NOATIME;
3145         if (flags & EXT3_DIRSYNC_FL)
3146                 inode->i_flags |= S_DIRSYNC;
3147 +
3148 +       inode->i_vflags &= ~(V_BARRIER | V_COW);
3149 +
3150 +       if (flags & EXT3_BARRIER_FL)
3151 +               inode->i_vflags |= V_BARRIER;
3152 +       if (flags & EXT3_COW_FL)
3153 +               inode->i_vflags |= V_COW;
3154  }
3155
3156  /* Propagate flags from i_flags to EXT3_I(inode)->i_flags */
3157  void ext3_get_inode_flags(struct ext3_inode_info *ei)
3158  {
3159         unsigned int flags = ei->vfs_inode.i_flags;
3160 +       unsigned int vflags = ei->vfs_inode.i_vflags;
3161 +
3162 +       ei->i_flags &= ~(EXT3_SYNC_FL | EXT3_APPEND_FL |
3163 +                       EXT3_IMMUTABLE_FL | EXT3_IXUNLINK_FL |
3164 +                       EXT3_NOATIME_FL | EXT3_DIRSYNC_FL |
3165 +                       EXT3_BARRIER_FL | EXT3_COW_FL);
3166 +
3167 +       if (flags & S_IMMUTABLE)
3168 +               ei->i_flags |= EXT3_IMMUTABLE_FL;
3169 +       if (flags & S_IXUNLINK)
3170 +               ei->i_flags |= EXT3_IXUNLINK_FL;
3171
3172 -       ei->i_flags &= ~(EXT3_SYNC_FL|EXT3_APPEND_FL|
3173 -                       EXT3_IMMUTABLE_FL|EXT3_NOATIME_FL|EXT3_DIRSYNC_FL);
3174         if (flags & S_SYNC)
3175                 ei->i_flags |= EXT3_SYNC_FL;
3176         if (flags & S_APPEND)
3177                 ei->i_flags |= EXT3_APPEND_FL;
3178 -       if (flags & S_IMMUTABLE)
3179 -               ei->i_flags |= EXT3_IMMUTABLE_FL;
3180         if (flags & S_NOATIME)
3181                 ei->i_flags |= EXT3_NOATIME_FL;
3182         if (flags & S_DIRSYNC)
3183                 ei->i_flags |= EXT3_DIRSYNC_FL;
3184 +
3185 +       if (vflags & V_BARRIER)
3186 +               ei->i_flags |= EXT3_BARRIER_FL;
3187 +       if (vflags & V_COW)
3188 +               ei->i_flags |= EXT3_COW_FL;
3189  }
3190
3191  struct inode *ext3_iget(struct super_block *sb, unsigned long ino)
3192 @@ -2818,6 +2843,8 @@ struct inode *ext3_iget(struct super_blo
3193         transaction_t *transaction;
3194         long ret;
3195         int block;
3196 +       uid_t uid;
3197 +       gid_t gid;
3198
3199         inode = iget_locked(sb, ino);
3200         if (!inode)
3201 @@ -2834,12 +2861,17 @@ struct inode *ext3_iget(struct super_blo
3202         bh = iloc.bh;
3203         raw_inode = ext3_raw_inode(&iloc);
3204         inode->i_mode = le16_to_cpu(raw_inode->i_mode);
3205 -       inode->i_uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
3206 -       inode->i_gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
3207 +       uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
3208 +       gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
3209         if(!(test_opt (inode->i_sb, NO_UID32))) {
3210 -               inode->i_uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
3211 -               inode->i_gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
3212 +               uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
3213 +               gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
3214         }
3215 +       inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
3216 +       inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
3217 +       inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid,
3218 +               le16_to_cpu(raw_inode->i_raw_tag));
3219 +
3220         inode->i_nlink = le16_to_cpu(raw_inode->i_links_count);
3221         inode->i_size = le32_to_cpu(raw_inode->i_size);
3222         inode->i_atime.tv_sec = (signed)le32_to_cpu(raw_inode->i_atime);
3223 @@ -2994,6 +3026,8 @@ static int ext3_do_update_inode(handle_t
3224         struct ext3_inode *raw_inode = ext3_raw_inode(iloc);
3225         struct ext3_inode_info *ei = EXT3_I(inode);
3226         struct buffer_head *bh = iloc->bh;
3227 +       uid_t uid = TAGINO_UID(DX_TAG(inode), inode->i_uid, inode->i_tag);
3228 +       gid_t gid = TAGINO_GID(DX_TAG(inode), inode->i_gid, inode->i_tag);
3229         int err = 0, rc, block;
3230
3231  again:
3232 @@ -3008,29 +3042,32 @@ again:
3233         ext3_get_inode_flags(ei);
3234         raw_inode->i_mode = cpu_to_le16(inode->i_mode);
3235         if(!(test_opt(inode->i_sb, NO_UID32))) {
3236 -               raw_inode->i_uid_low = cpu_to_le16(low_16_bits(inode->i_uid));
3237 -               raw_inode->i_gid_low = cpu_to_le16(low_16_bits(inode->i_gid));
3238 +               raw_inode->i_uid_low = cpu_to_le16(low_16_bits(uid));
3239 +               raw_inode->i_gid_low = cpu_to_le16(low_16_bits(gid));
3240  /*
3241   * Fix up interoperability with old kernels. Otherwise, old inodes get
3242   * re-used with the upper 16 bits of the uid/gid intact
3243   */
3244                 if(!ei->i_dtime) {
3245                         raw_inode->i_uid_high =
3246 -                               cpu_to_le16(high_16_bits(inode->i_uid));
3247 +                               cpu_to_le16(high_16_bits(uid));
3248                         raw_inode->i_gid_high =
3249 -                               cpu_to_le16(high_16_bits(inode->i_gid));
3250 +                               cpu_to_le16(high_16_bits(gid));
3251                 } else {
3252                         raw_inode->i_uid_high = 0;
3253                         raw_inode->i_gid_high = 0;
3254                 }
3255         } else {
3256                 raw_inode->i_uid_low =
3257 -                       cpu_to_le16(fs_high2lowuid(inode->i_uid));
3258 +                       cpu_to_le16(fs_high2lowuid(uid));
3259                 raw_inode->i_gid_low =
3260 -                       cpu_to_le16(fs_high2lowgid(inode->i_gid));
3261 +                       cpu_to_le16(fs_high2lowgid(gid));
3262                 raw_inode->i_uid_high = 0;
3263                 raw_inode->i_gid_high = 0;
3264         }
3265 +#ifdef CONFIG_TAGGING_INTERN
3266 +       raw_inode->i_raw_tag = cpu_to_le16(inode->i_tag);
3267 +#endif
3268         raw_inode->i_links_count = cpu_to_le16(inode->i_nlink);
3269         raw_inode->i_size = cpu_to_le32(ei->i_disksize);
3270         raw_inode->i_atime = cpu_to_le32(inode->i_atime.tv_sec);
3271 @@ -3190,7 +3227,8 @@ int ext3_setattr(struct dentry *dentry,
3272         if (is_quota_modification(inode, attr))
3273                 dquot_initialize(inode);
3274         if ((ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) ||
3275 -               (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) {
3276 +               (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid) ||
3277 +               (ia_valid & ATTR_TAG && attr->ia_tag != inode->i_tag)) {
3278                 handle_t *handle;
3279
3280                 /* (user+group)*(old+new) structure, inode write (sb,
3281 @@ -3212,6 +3250,8 @@ int ext3_setattr(struct dentry *dentry,
3282                         inode->i_uid = attr->ia_uid;
3283                 if (attr->ia_valid & ATTR_GID)
3284                         inode->i_gid = attr->ia_gid;
3285 +               if ((attr->ia_valid & ATTR_TAG) && IS_TAGGED(inode))
3286 +                       inode->i_tag = attr->ia_tag;
3287                 error = ext3_mark_inode_dirty(handle, inode);
3288                 ext3_journal_stop(handle);
3289         }
3290 diff -NurpP --minimal linux-3.0-rc5/fs/ext3/ioctl.c linux-3.0-rc5-vs2.3.1-pre3/fs/ext3/ioctl.c
3291 --- linux-3.0-rc5/fs/ext3/ioctl.c       2011-05-22 16:17:52.000000000 +0200
3292 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ext3/ioctl.c  2011-06-10 22:11:24.000000000 +0200
3293 @@ -8,6 +8,7 @@
3294   */
3295
3296  #include <linux/fs.h>
3297 +#include <linux/mount.h>
3298  #include <linux/jbd.h>
3299  #include <linux/capability.h>
3300  #include <linux/ext3_fs.h>
3301 @@ -17,6 +18,34 @@
3302  #include <linux/compat.h>
3303  #include <asm/uaccess.h>
3304
3305 +
3306 +int ext3_sync_flags(struct inode *inode, int flags, int vflags)
3307 +{
3308 +       handle_t *handle = NULL;
3309 +       struct ext3_iloc iloc;
3310 +       int err;
3311 +
3312 +       handle = ext3_journal_start(inode, 1);
3313 +       if (IS_ERR(handle))
3314 +               return PTR_ERR(handle);
3315 +
3316 +       if (IS_SYNC(inode))
3317 +               handle->h_sync = 1;
3318 +       err = ext3_reserve_inode_write(handle, inode, &iloc);
3319 +       if (err)
3320 +               goto flags_err;
3321 +
3322 +       inode->i_flags = flags;
3323 +       inode->i_vflags = vflags;
3324 +       ext3_get_inode_flags(EXT3_I(inode));
3325 +       inode->i_ctime = CURRENT_TIME_SEC;
3326 +
3327 +       err = ext3_mark_iloc_dirty(handle, inode, &iloc);
3328 +flags_err:
3329 +       ext3_journal_stop(handle);
3330 +       return err;
3331 +}
3332 +
3333  long ext3_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
3334  {
3335         struct inode *inode = filp->f_dentry->d_inode;
3336 @@ -50,6 +79,11 @@ long ext3_ioctl(struct file *filp, unsig
3337
3338                 flags = ext3_mask_flags(inode->i_mode, flags);
3339
3340 +               if (IS_BARRIER(inode)) {
3341 +                       vxwprintk_task(1, "messing with the barrier.");
3342 +                       return -EACCES;
3343 +               }
3344 +
3345                 mutex_lock(&inode->i_mutex);
3346
3347                 /* Is it quota file? Do not allow user to mess with it */
3348 @@ -68,7 +102,9 @@ long ext3_ioctl(struct file *filp, unsig
3349                  *
3350                  * This test looks nicer. Thanks to Pauline Middelink
3351                  */
3352 -               if ((flags ^ oldflags) & (EXT3_APPEND_FL | EXT3_IMMUTABLE_FL)) {
3353 +               if ((oldflags & EXT3_IMMUTABLE_FL) ||
3354 +                       ((flags ^ oldflags) & (EXT3_APPEND_FL |
3355 +                       EXT3_IMMUTABLE_FL | EXT3_IXUNLINK_FL))) {
3356                         if (!capable(CAP_LINUX_IMMUTABLE))
3357                                 goto flags_out;
3358                 }
3359 @@ -93,7 +129,7 @@ long ext3_ioctl(struct file *filp, unsig
3360                 if (err)
3361                         goto flags_err;
3362
3363 -               flags = flags & EXT3_FL_USER_MODIFIABLE;
3364 +               flags &= EXT3_FL_USER_MODIFIABLE;
3365                 flags |= oldflags & ~EXT3_FL_USER_MODIFIABLE;
3366                 ei->i_flags = flags;
3367
3368 diff -NurpP --minimal linux-3.0-rc5/fs/ext3/namei.c linux-3.0-rc5-vs2.3.1-pre3/fs/ext3/namei.c
3369 --- linux-3.0-rc5/fs/ext3/namei.c       2011-07-01 11:14:18.000000000 +0200
3370 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ext3/namei.c  2011-06-10 22:11:24.000000000 +0200
3371 @@ -36,6 +36,7 @@
3372  #include <linux/quotaops.h>
3373  #include <linux/buffer_head.h>
3374  #include <linux/bio.h>
3375 +#include <linux/vs_tag.h>
3376
3377  #include "namei.h"
3378  #include "xattr.h"
3379 @@ -923,6 +924,7 @@ restart:
3380                                 if (bh)
3381                                         ll_rw_block(READ_META, 1, &bh);
3382                         }
3383 +               dx_propagate_tag(nd, inode);
3384                 }
3385                 if ((bh = bh_use[ra_ptr++]) == NULL)
3386                         goto next;
3387 @@ -2534,6 +2536,7 @@ const struct inode_operations ext3_dir_i
3388         .removexattr    = generic_removexattr,
3389  #endif
3390         .check_acl      = ext3_check_acl,
3391 +       .sync_flags     = ext3_sync_flags,
3392  };
3393
3394  const struct inode_operations ext3_special_inode_operations = {
3395 diff -NurpP --minimal linux-3.0-rc5/fs/ext3/super.c linux-3.0-rc5-vs2.3.1-pre3/fs/ext3/super.c
3396 --- linux-3.0-rc5/fs/ext3/super.c       2011-07-01 11:14:18.000000000 +0200
3397 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ext3/super.c  2011-06-10 22:11:24.000000000 +0200
3398 @@ -821,7 +821,8 @@ enum {
3399         Opt_usrjquota, Opt_grpjquota, Opt_offusrjquota, Opt_offgrpjquota,
3400         Opt_jqfmt_vfsold, Opt_jqfmt_vfsv0, Opt_jqfmt_vfsv1, Opt_quota,
3401         Opt_noquota, Opt_ignore, Opt_barrier, Opt_nobarrier, Opt_err,
3402 -       Opt_resize, Opt_usrquota, Opt_grpquota
3403 +       Opt_resize, Opt_usrquota, Opt_grpquota,
3404 +       Opt_tag, Opt_notag, Opt_tagid
3405  };
3406
3407  static const match_table_t tokens = {
3408 @@ -878,6 +879,9 @@ static const match_table_t tokens = {
3409         {Opt_barrier, "barrier"},
3410         {Opt_nobarrier, "nobarrier"},
3411         {Opt_resize, "resize"},
3412 +       {Opt_tag, "tag"},
3413 +       {Opt_notag, "notag"},
3414 +       {Opt_tagid, "tagid=%u"},
3415         {Opt_err, NULL},
3416  };
3417
3418 @@ -1030,6 +1034,20 @@ static int parse_options (char *options,
3419                 case Opt_nouid32:
3420                         set_opt (sbi->s_mount_opt, NO_UID32);
3421                         break;
3422 +#ifndef CONFIG_TAGGING_NONE
3423 +               case Opt_tag:
3424 +                       set_opt (sbi->s_mount_opt, TAGGED);
3425 +                       break;
3426 +               case Opt_notag:
3427 +                       clear_opt (sbi->s_mount_opt, TAGGED);
3428 +                       break;
3429 +#endif
3430 +#ifdef CONFIG_PROPAGATE
3431 +               case Opt_tagid:
3432 +                       /* use args[0] */
3433 +                       set_opt (sbi->s_mount_opt, TAGGED);
3434 +                       break;
3435 +#endif
3436                 case Opt_nocheck:
3437                         clear_opt (sbi->s_mount_opt, CHECK);
3438                         break;
3439 @@ -1724,6 +1742,9 @@ static int ext3_fill_super (struct super
3440                             NULL, 0))
3441                 goto failed_mount;
3442
3443 +       if (EXT3_SB(sb)->s_mount_opt & EXT3_MOUNT_TAGGED)
3444 +               sb->s_flags |= MS_TAGGED;
3445 +
3446         sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
3447                 (test_opt(sb, POSIX_ACL) ? MS_POSIXACL : 0);
3448
3449 @@ -2604,6 +2625,14 @@ static int ext3_remount (struct super_bl
3450         if (test_opt(sb, ABORT))
3451                 ext3_abort(sb, __func__, "Abort forced by user");
3452
3453 +       if ((sbi->s_mount_opt & EXT3_MOUNT_TAGGED) &&
3454 +               !(sb->s_flags & MS_TAGGED)) {
3455 +               printk("EXT3-fs: %s: tagging not permitted on remount.\n",
3456 +                       sb->s_id);
3457 +               err = -EINVAL;
3458 +               goto restore_opts;
3459 +       }
3460 +
3461         sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
3462                 (test_opt(sb, POSIX_ACL) ? MS_POSIXACL : 0);
3463
3464 diff -NurpP --minimal linux-3.0-rc5/fs/ext4/ext4.h linux-3.0-rc5-vs2.3.1-pre3/fs/ext4/ext4.h
3465 --- linux-3.0-rc5/fs/ext4/ext4.h        2011-07-01 11:14:18.000000000 +0200
3466 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ext4/ext4.h   2011-06-10 22:11:24.000000000 +0200
3467 @@ -350,8 +350,12 @@ struct flex_groups {
3468  #define EXT4_EXTENTS_FL                        0x00080000 /* Inode uses extents */
3469  #define EXT4_EA_INODE_FL               0x00200000 /* Inode used for large EA */
3470  #define EXT4_EOFBLOCKS_FL              0x00400000 /* Blocks allocated beyond EOF */
3471 +#define EXT4_IXUNLINK_FL               0x08000000 /* Immutable invert on unlink */
3472  #define EXT4_RESERVED_FL               0x80000000 /* reserved for ext4 lib */
3473
3474 +#define EXT4_BARRIER_FL                        0x04000000 /* Barrier for chroot() */
3475 +#define EXT4_COW_FL                    0x20000000 /* Copy on Write marker */
3476 +
3477  #define EXT4_FL_USER_VISIBLE           0x004BDFFF /* User visible flags */
3478  #define EXT4_FL_USER_MODIFIABLE                0x004B80FF /* User modifiable flags */
3479
3480 @@ -608,7 +612,8 @@ struct ext4_inode {
3481                         __le16  l_i_file_acl_high;
3482                         __le16  l_i_uid_high;   /* these 2 fields */
3483                         __le16  l_i_gid_high;   /* were reserved2[0] */
3484 -                       __u32   l_i_reserved2;
3485 +                       __le16  l_i_tag;        /* Context Tag */
3486 +                       __u16   l_i_reserved2;
3487                 } linux2;
3488                 struct {
3489                         __le16  h_i_reserved1;  /* Obsoleted fragment number/size which are removed in ext4 */
3490 @@ -726,6 +731,7 @@ do {                                                                               \
3491  #define i_gid_low      i_gid
3492  #define i_uid_high     osd2.linux2.l_i_uid_high
3493  #define i_gid_high     osd2.linux2.l_i_gid_high
3494 +#define i_raw_tag      osd2.linux2.l_i_tag
3495  #define i_reserved2    osd2.linux2.l_i_reserved2
3496
3497  #elif defined(__GNU__)
3498 @@ -902,6 +908,7 @@ struct ext4_inode_info {
3499  #define EXT4_MOUNT_POSIX_ACL           0x08000 /* POSIX Access Control Lists */
3500  #define EXT4_MOUNT_NO_AUTO_DA_ALLOC    0x10000 /* No auto delalloc mapping */
3501  #define EXT4_MOUNT_BARRIER             0x20000 /* Use block barriers */
3502 +#define EXT4_MOUNT_TAGGED              0x40000 /* Enable Context Tags */
3503  #define EXT4_MOUNT_QUOTA               0x80000 /* Some quota option set */
3504  #define EXT4_MOUNT_USRQUOTA            0x100000 /* "old" user quota */
3505  #define EXT4_MOUNT_GRPQUOTA            0x200000 /* "old" group quota */
3506 @@ -2172,6 +2179,7 @@ extern int ext4_map_blocks(handle_t *han
3507                            struct ext4_map_blocks *map, int flags);
3508  extern int ext4_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
3509                         __u64 start, __u64 len);
3510 +extern int ext4_sync_flags(struct inode *, int, int);
3511  /* move_extent.c */
3512  extern int ext4_move_extents(struct file *o_filp, struct file *d_filp,
3513                              __u64 start_orig, __u64 start_donor,
3514 diff -NurpP --minimal linux-3.0-rc5/fs/ext4/file.c linux-3.0-rc5-vs2.3.1-pre3/fs/ext4/file.c
3515 --- linux-3.0-rc5/fs/ext4/file.c        2011-07-01 11:14:18.000000000 +0200
3516 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ext4/file.c   2011-06-10 22:11:24.000000000 +0200
3517 @@ -282,5 +282,6 @@ const struct inode_operations ext4_file_
3518  #endif
3519         .check_acl      = ext4_check_acl,
3520         .fiemap         = ext4_fiemap,
3521 +       .sync_flags     = ext4_sync_flags,
3522  };
3523
3524 diff -NurpP --minimal linux-3.0-rc5/fs/ext4/ialloc.c linux-3.0-rc5-vs2.3.1-pre3/fs/ext4/ialloc.c
3525 --- linux-3.0-rc5/fs/ext4/ialloc.c      2011-05-22 16:17:52.000000000 +0200
3526 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ext4/ialloc.c 2011-06-10 22:11:24.000000000 +0200
3527 @@ -22,6 +22,7 @@
3528  #include <linux/random.h>
3529  #include <linux/bitops.h>
3530  #include <linux/blkdev.h>
3531 +#include <linux/vs_tag.h>
3532  #include <asm/byteorder.h>
3533
3534  #include "ext4.h"
3535 @@ -992,6 +993,7 @@ got:
3536                 inode->i_mode = mode;
3537                 inode->i_uid = current_fsuid();
3538                 inode->i_gid = dir->i_gid;
3539 +               inode->i_tag = dx_current_fstag(sb);
3540         } else
3541                 inode_init_owner(inode, dir, mode);
3542
3543 diff -NurpP --minimal linux-3.0-rc5/fs/ext4/inode.c linux-3.0-rc5-vs2.3.1-pre3/fs/ext4/inode.c
3544 --- linux-3.0-rc5/fs/ext4/inode.c       2011-07-01 11:14:18.000000000 +0200
3545 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ext4/inode.c  2011-07-01 11:35:34.000000000 +0200
3546 @@ -42,6 +42,7 @@
3547  #include <linux/printk.h>
3548  #include <linux/slab.h>
3549  #include <linux/ratelimit.h>
3550 +#include <linux/vs_tag.h>
3551
3552  #include "ext4_jbd2.h"
3553  #include "xattr.h"
3554 @@ -4797,41 +4798,64 @@ void ext4_set_inode_flags(struct inode *
3555  {
3556         unsigned int flags = EXT4_I(inode)->i_flags;
3557
3558 -       inode->i_flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC);
3559 +       inode->i_flags &= ~(S_IMMUTABLE | S_IXUNLINK |
3560 +               S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
3561 +
3562 +       if (flags & EXT4_IMMUTABLE_FL)
3563 +               inode->i_flags |= S_IMMUTABLE;
3564 +       if (flags & EXT4_IXUNLINK_FL)
3565 +               inode->i_flags |= S_IXUNLINK;
3566 +
3567         if (flags & EXT4_SYNC_FL)
3568                 inode->i_flags |= S_SYNC;
3569         if (flags & EXT4_APPEND_FL)
3570                 inode->i_flags |= S_APPEND;
3571 -       if (flags & EXT4_IMMUTABLE_FL)
3572 -               inode->i_flags |= S_IMMUTABLE;
3573         if (flags & EXT4_NOATIME_FL)
3574                 inode->i_flags |= S_NOATIME;
3575         if (flags & EXT4_DIRSYNC_FL)
3576                 inode->i_flags |= S_DIRSYNC;
3577 +
3578 +       inode->i_vflags &= ~(V_BARRIER | V_COW);
3579 +
3580 +       if (flags & EXT4_BARRIER_FL)
3581 +               inode->i_vflags |= V_BARRIER;
3582 +       if (flags & EXT4_COW_FL)
3583 +               inode->i_vflags |= V_COW;
3584  }
3585
3586  /* Propagate flags from i_flags to EXT4_I(inode)->i_flags */
3587  void ext4_get_inode_flags(struct ext4_inode_info *ei)
3588  {
3589 -       unsigned int vfs_fl;
3590 +       unsigned int vfs_fl, vfs_vf;
3591         unsigned long old_fl, new_fl;
3592
3593         do {
3594                 vfs_fl = ei->vfs_inode.i_flags;
3595 +               vfs_vf = ei->vfs_inode.i_vflags;
3596                 old_fl = ei->i_flags;
3597                 new_fl = old_fl & ~(EXT4_SYNC_FL|EXT4_APPEND_FL|
3598                                 EXT4_IMMUTABLE_FL|EXT4_NOATIME_FL|
3599 -                               EXT4_DIRSYNC_FL);
3600 +                               EXT4_DIRSYNC_FL|EXT4_BARRIER_FL|
3601 +                               EXT4_COW_FL);
3602 +
3603 +               if (vfs_fl & S_IMMUTABLE)
3604 +                       new_fl |= EXT4_IMMUTABLE_FL;
3605 +               if (vfs_fl & S_IXUNLINK)
3606 +                       new_fl |= EXT4_IXUNLINK_FL;
3607 +
3608                 if (vfs_fl & S_SYNC)
3609                         new_fl |= EXT4_SYNC_FL;
3610                 if (vfs_fl & S_APPEND)
3611                         new_fl |= EXT4_APPEND_FL;
3612 -               if (vfs_fl & S_IMMUTABLE)
3613 -                       new_fl |= EXT4_IMMUTABLE_FL;
3614                 if (vfs_fl & S_NOATIME)
3615                         new_fl |= EXT4_NOATIME_FL;
3616                 if (vfs_fl & S_DIRSYNC)
3617                         new_fl |= EXT4_DIRSYNC_FL;
3618 +
3619 +               if (vfs_vf & V_BARRIER)
3620 +                       new_fl |= EXT4_BARRIER_FL;
3621 +               if (vfs_vf & V_COW)
3622 +                       new_fl |= EXT4_COW_FL;
3623         } while (cmpxchg(&ei->i_flags, old_fl, new_fl) != old_fl);
3624  }
3625
3626 @@ -4867,6 +4891,8 @@ struct inode *ext4_iget(struct super_blo
3627         journal_t *journal = EXT4_SB(sb)->s_journal;
3628         long ret;
3629         int block;
3630 +       uid_t uid;
3631 +       gid_t gid;
3632
3633         inode = iget_locked(sb, ino);
3634         if (!inode)
3635 @@ -4882,12 +4908,16 @@ struct inode *ext4_iget(struct super_blo
3636                 goto bad_inode;
3637         raw_inode = ext4_raw_inode(&iloc);
3638         inode->i_mode = le16_to_cpu(raw_inode->i_mode);
3639 -       inode->i_uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
3640 -       inode->i_gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
3641 +       uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
3642 +       gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
3643         if (!(test_opt(inode->i_sb, NO_UID32))) {
3644 -               inode->i_uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
3645 -               inode->i_gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
3646 +               uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
3647 +               gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
3648         }
3649 +       inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
3650 +       inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
3651 +       inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid,
3652 +               le16_to_cpu(raw_inode->i_raw_tag));
3653         inode->i_nlink = le16_to_cpu(raw_inode->i_links_count);
3654
3655         ext4_clear_state_flags(ei);     /* Only relevant on 32-bit archs */
3656 @@ -5106,6 +5136,8 @@ static int ext4_do_update_inode(handle_t
3657         struct ext4_inode *raw_inode = ext4_raw_inode(iloc);
3658         struct ext4_inode_info *ei = EXT4_I(inode);
3659         struct buffer_head *bh = iloc->bh;
3660 +       uid_t uid = TAGINO_UID(DX_TAG(inode), inode->i_uid, inode->i_tag);
3661 +       gid_t gid = TAGINO_GID(DX_TAG(inode), inode->i_gid, inode->i_tag);
3662         int err = 0, rc, block;
3663
3664         /* For fields not not tracking in the in-memory inode,
3665 @@ -5116,29 +5148,32 @@ static int ext4_do_update_inode(handle_t
3666         ext4_get_inode_flags(ei);
3667         raw_inode->i_mode = cpu_to_le16(inode->i_mode);
3668         if (!(test_opt(inode->i_sb, NO_UID32))) {
3669 -               raw_inode->i_uid_low = cpu_to_le16(low_16_bits(inode->i_uid));
3670 -               raw_inode->i_gid_low = cpu_to_le16(low_16_bits(inode->i_gid));
3671 +               raw_inode->i_uid_low = cpu_to_le16(low_16_bits(uid));
3672 +               raw_inode->i_gid_low = cpu_to_le16(low_16_bits(gid));
3673  /*
3674   * Fix up interoperability with old kernels. Otherwise, old inodes get
3675   * re-used with the upper 16 bits of the uid/gid intact
3676   */
3677                 if (!ei->i_dtime) {
3678                         raw_inode->i_uid_high =
3679 -                               cpu_to_le16(high_16_bits(inode->i_uid));
3680 +                               cpu_to_le16(high_16_bits(uid));
3681                         raw_inode->i_gid_high =
3682 -                               cpu_to_le16(high_16_bits(inode->i_gid));
3683 +                               cpu_to_le16(high_16_bits(gid));
3684                 } else {
3685                         raw_inode->i_uid_high = 0;
3686                         raw_inode->i_gid_high = 0;
3687                 }
3688         } else {
3689                 raw_inode->i_uid_low =
3690 -                       cpu_to_le16(fs_high2lowuid(inode->i_uid));
3691 +                       cpu_to_le16(fs_high2lowuid(uid));
3692                 raw_inode->i_gid_low =
3693 -                       cpu_to_le16(fs_high2lowgid(inode->i_gid));
3694 +                       cpu_to_le16(fs_high2lowgid(gid));
3695                 raw_inode->i_uid_high = 0;
3696                 raw_inode->i_gid_high = 0;
3697         }
3698 +#ifdef CONFIG_TAGGING_INTERN
3699 +       raw_inode->i_raw_tag = cpu_to_le16(inode->i_tag);
3700 +#endif
3701         raw_inode->i_links_count = cpu_to_le16(inode->i_nlink);
3702
3703         EXT4_INODE_SET_XTIME(i_ctime, inode, raw_inode);
3704 @@ -5324,7 +5359,8 @@ int ext4_setattr(struct dentry *dentry,
3705         if (is_quota_modification(inode, attr))
3706                 dquot_initialize(inode);
3707         if ((ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) ||
3708 -               (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) {
3709 +               (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid) ||
3710 +               (ia_valid & ATTR_TAG && attr->ia_tag != inode->i_tag)) {
3711                 handle_t *handle;
3712
3713                 /* (user+group)*(old+new) structure, inode write (sb,
3714 @@ -5346,6 +5382,8 @@ int ext4_setattr(struct dentry *dentry,
3715                         inode->i_uid = attr->ia_uid;
3716                 if (attr->ia_valid & ATTR_GID)
3717                         inode->i_gid = attr->ia_gid;
3718 +               if ((attr->ia_valid & ATTR_TAG) && IS_TAGGED(inode))
3719 +                       inode->i_tag = attr->ia_tag;
3720                 error = ext4_mark_inode_dirty(handle, inode);
3721                 ext4_journal_stop(handle);
3722         }
3723 diff -NurpP --minimal linux-3.0-rc5/fs/ext4/ioctl.c linux-3.0-rc5-vs2.3.1-pre3/fs/ext4/ioctl.c
3724 --- linux-3.0-rc5/fs/ext4/ioctl.c       2011-05-22 16:17:52.000000000 +0200
3725 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ext4/ioctl.c  2011-06-10 22:11:24.000000000 +0200
3726 @@ -14,10 +14,39 @@
3727  #include <linux/compat.h>
3728  #include <linux/mount.h>
3729  #include <linux/file.h>
3730 +#include <linux/vs_tag.h>
3731  #include <asm/uaccess.h>
3732  #include "ext4_jbd2.h"
3733  #include "ext4.h"
3734
3735 +
3736 +int ext4_sync_flags(struct inode *inode, int flags, int vflags)
3737 +{
3738 +       handle_t *handle = NULL;
3739 +       struct ext4_iloc iloc;
3740 +       int err;
3741 +
3742 +       handle = ext4_journal_start(inode, 1);
3743 +       if (IS_ERR(handle))
3744 +               return PTR_ERR(handle);
3745 +
3746 +       if (IS_SYNC(inode))
3747 +               ext4_handle_sync(handle);
3748 +       err = ext4_reserve_inode_write(handle, inode, &iloc);
3749 +       if (err)
3750 +               goto flags_err;
3751 +
3752 +       inode->i_flags = flags;
3753 +       inode->i_vflags = vflags;
3754 +       ext4_get_inode_flags(EXT4_I(inode));
3755 +       inode->i_ctime = ext4_current_time(inode);
3756 +
3757 +       err = ext4_mark_iloc_dirty(handle, inode, &iloc);
3758 +flags_err:
3759 +       ext4_journal_stop(handle);
3760 +       return err;
3761 +}
3762 +
3763  long ext4_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
3764  {
3765         struct inode *inode = filp->f_dentry->d_inode;
3766 @@ -50,6 +79,11 @@ long ext4_ioctl(struct file *filp, unsig
3767
3768                 flags = ext4_mask_flags(inode->i_mode, flags);
3769
3770 +               if (IS_BARRIER(inode)) {
3771 +                       vxwprintk_task(1, "messing with the barrier.");
3772 +                       return -EACCES;
3773 +               }
3774 +
3775                 err = -EPERM;
3776                 mutex_lock(&inode->i_mutex);
3777                 /* Is it quota file? Do not allow user to mess with it */
3778 @@ -67,7 +101,9 @@ long ext4_ioctl(struct file *filp, unsig
3779                  *
3780                  * This test looks nicer. Thanks to Pauline Middelink
3781                  */
3782 -               if ((flags ^ oldflags) & (EXT4_APPEND_FL | EXT4_IMMUTABLE_FL)) {
3783 +               if ((oldflags & EXT4_IMMUTABLE_FL) ||
3784 +                       ((flags ^ oldflags) & (EXT4_APPEND_FL |
3785 +                       EXT4_IMMUTABLE_FL | EXT4_IXUNLINK_FL))) {
3786                         if (!capable(CAP_LINUX_IMMUTABLE))
3787                                 goto flags_out;
3788                 }
3789 diff -NurpP --minimal linux-3.0-rc5/fs/ext4/namei.c linux-3.0-rc5-vs2.3.1-pre3/fs/ext4/namei.c
3790 --- linux-3.0-rc5/fs/ext4/namei.c       2011-07-01 11:14:18.000000000 +0200
3791 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ext4/namei.c  2011-06-10 22:11:24.000000000 +0200
3792 @@ -34,6 +34,7 @@
3793  #include <linux/quotaops.h>
3794  #include <linux/buffer_head.h>
3795  #include <linux/bio.h>
3796 +#include <linux/vs_tag.h>
3797  #include "ext4.h"
3798  #include "ext4_jbd2.h"
3799
3800 @@ -924,6 +925,7 @@ restart:
3801                                 if (bh)
3802                                         ll_rw_block(READ_META, 1, &bh);
3803                         }
3804 +               dx_propagate_tag(nd, inode);
3805                 }
3806                 if ((bh = bh_use[ra_ptr++]) == NULL)
3807                         goto next;
3808 @@ -2596,6 +2598,7 @@ const struct inode_operations ext4_dir_i
3809  #endif
3810         .check_acl      = ext4_check_acl,
3811         .fiemap         = ext4_fiemap,
3812 +       .sync_flags     = ext4_sync_flags,
3813  };
3814
3815  const struct inode_operations ext4_special_inode_operations = {
3816 diff -NurpP --minimal linux-3.0-rc5/fs/ext4/super.c linux-3.0-rc5-vs2.3.1-pre3/fs/ext4/super.c
3817 --- linux-3.0-rc5/fs/ext4/super.c       2011-07-01 11:14:18.000000000 +0200
3818 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ext4/super.c  2011-07-01 11:35:34.000000000 +0200
3819 @@ -1294,6 +1294,7 @@ enum {
3820         Opt_dioread_nolock, Opt_dioread_lock,
3821         Opt_discard, Opt_nodiscard,
3822         Opt_init_inode_table, Opt_noinit_inode_table,
3823 +       Opt_tag, Opt_notag, Opt_tagid
3824  };
3825
3826  static const match_table_t tokens = {
3827 @@ -1369,6 +1370,9 @@ static const match_table_t tokens = {
3828         {Opt_init_inode_table, "init_itable=%u"},
3829         {Opt_init_inode_table, "init_itable"},
3830         {Opt_noinit_inode_table, "noinit_itable"},
3831 +       {Opt_tag, "tag"},
3832 +       {Opt_notag, "notag"},
3833 +       {Opt_tagid, "tagid=%u"},
3834         {Opt_err, NULL},
3835  };
3836
3837 @@ -1537,6 +1541,20 @@ static int parse_options(char *options,
3838                 case Opt_nouid32:
3839                         set_opt(sb, NO_UID32);
3840                         break;
3841 +#ifndef CONFIG_TAGGING_NONE
3842 +               case Opt_tag:
3843 +                       set_opt(sb, TAGGED);
3844 +                       break;
3845 +               case Opt_notag:
3846 +                       clear_opt(sb, TAGGED);
3847 +                       break;
3848 +#endif
3849 +#ifdef CONFIG_PROPAGATE
3850 +               case Opt_tagid:
3851 +                       /* use args[0] */
3852 +                       set_opt(sb, TAGGED);
3853 +                       break;
3854 +#endif
3855                 case Opt_debug:
3856                         set_opt(sb, DEBUG);
3857                         break;
3858 @@ -3194,6 +3212,9 @@ static int ext4_fill_super(struct super_
3859                            &journal_ioprio, NULL, 0))
3860                 goto failed_mount;
3861
3862 +       if (EXT4_SB(sb)->s_mount_opt & EXT4_MOUNT_TAGGED)
3863 +               sb->s_flags |= MS_TAGGED;
3864 +
3865         sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
3866                 (test_opt(sb, POSIX_ACL) ? MS_POSIXACL : 0);
3867
3868 @@ -4324,6 +4345,14 @@ static int ext4_remount(struct super_blo
3869         if (sbi->s_mount_flags & EXT4_MF_FS_ABORTED)
3870                 ext4_abort(sb, "Abort forced by user");
3871
3872 +       if ((sbi->s_mount_opt & EXT4_MOUNT_TAGGED) &&
3873 +               !(sb->s_flags & MS_TAGGED)) {
3874 +               printk("EXT4-fs: %s: tagging not permitted on remount.\n",
3875 +                       sb->s_id);
3876 +               err = -EINVAL;
3877 +               goto restore_opts;
3878 +       }
3879 +
3880         sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
3881                 (test_opt(sb, POSIX_ACL) ? MS_POSIXACL : 0);
3882
3883 diff -NurpP --minimal linux-3.0-rc5/fs/fcntl.c linux-3.0-rc5-vs2.3.1-pre3/fs/fcntl.c
3884 --- linux-3.0-rc5/fs/fcntl.c    2011-05-22 16:17:52.000000000 +0200
3885 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/fcntl.c       2011-06-10 22:11:24.000000000 +0200
3886 @@ -20,6 +20,7 @@
3887  #include <linux/signal.h>
3888  #include <linux/rcupdate.h>
3889  #include <linux/pid_namespace.h>
3890 +#include <linux/vs_limit.h>
3891
3892  #include <asm/poll.h>
3893  #include <asm/siginfo.h>
3894 @@ -103,6 +104,8 @@ SYSCALL_DEFINE3(dup3, unsigned int, oldf
3895
3896         if (tofree)
3897                 filp_close(tofree, files);
3898 +       else
3899 +               vx_openfd_inc(newfd);   /* fd was unused */
3900
3901         return newfd;
3902
3903 @@ -447,6 +450,8 @@ SYSCALL_DEFINE3(fcntl, unsigned int, fd,
3904         filp = fget_raw(fd);
3905         if (!filp)
3906                 goto out;
3907 +       if (!vx_files_avail(1))
3908 +               goto out;
3909
3910         if (unlikely(filp->f_mode & FMODE_PATH)) {
3911                 if (!check_fcntl_cmd(cmd)) {
3912 diff -NurpP --minimal linux-3.0-rc5/fs/file.c linux-3.0-rc5-vs2.3.1-pre3/fs/file.c
3913 --- linux-3.0-rc5/fs/file.c     2011-05-22 16:17:52.000000000 +0200
3914 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/file.c        2011-06-10 22:11:24.000000000 +0200
3915 @@ -21,6 +21,7 @@
3916  #include <linux/spinlock.h>
3917  #include <linux/rcupdate.h>
3918  #include <linux/workqueue.h>
3919 +#include <linux/vs_limit.h>
3920
3921  struct fdtable_defer {
3922         spinlock_t lock;
3923 @@ -359,6 +360,8 @@ struct files_struct *dup_fd(struct files
3924                 struct file *f = *old_fds++;
3925                 if (f) {
3926                         get_file(f);
3927 +                       /* TODO: sum it first for check and performance */
3928 +                       vx_openfd_inc(open_files - i);
3929                 } else {
3930                         /*
3931                          * The fd may be claimed in the fd bitmap but not yet
3932 @@ -466,6 +469,7 @@ repeat:
3933         else
3934                 FD_CLR(fd, fdt->close_on_exec);
3935         error = fd;
3936 +       vx_openfd_inc(fd);
3937  #if 1
3938         /* Sanity check */
3939         if (rcu_dereference_raw(fdt->fd[fd]) != NULL) {
3940 diff -NurpP --minimal linux-3.0-rc5/fs/file_table.c linux-3.0-rc5-vs2.3.1-pre3/fs/file_table.c
3941 --- linux-3.0-rc5/fs/file_table.c       2011-05-22 16:17:52.000000000 +0200
3942 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/file_table.c  2011-06-10 22:11:24.000000000 +0200
3943 @@ -24,6 +24,8 @@
3944  #include <linux/percpu_counter.h>
3945  #include <linux/percpu.h>
3946  #include <linux/ima.h>
3947 +#include <linux/vs_limit.h>
3948 +#include <linux/vs_context.h>
3949
3950  #include <asm/atomic.h>
3951
3952 @@ -135,6 +137,8 @@ struct file *get_empty_filp(void)
3953         spin_lock_init(&f->f_lock);
3954         eventpoll_init_file(f);
3955         /* f->f_version: 0 */
3956 +       f->f_xid = vx_current_xid();
3957 +       vx_files_inc(f);
3958         return f;
3959
3960  over:
3961 @@ -253,6 +257,8 @@ static void __fput(struct file *file)
3962         }
3963         fops_put(file->f_op);
3964         put_pid(file->f_owner.pid);
3965 +       vx_files_dec(file);
3966 +       file->f_xid = 0;
3967         file_sb_list_del(file);
3968         if ((file->f_mode & (FMODE_READ | FMODE_WRITE)) == FMODE_READ)
3969                 i_readcount_dec(inode);
3970 @@ -383,6 +389,8 @@ void put_filp(struct file *file)
3971  {
3972         if (atomic_long_dec_and_test(&file->f_count)) {
3973                 security_file_free(file);
3974 +               vx_files_dec(file);
3975 +               file->f_xid = 0;
3976                 file_sb_list_del(file);
3977                 file_free(file);
3978         }
3979 diff -NurpP --minimal linux-3.0-rc5/fs/fs_struct.c linux-3.0-rc5-vs2.3.1-pre3/fs/fs_struct.c
3980 --- linux-3.0-rc5/fs/fs_struct.c        2011-03-15 18:07:31.000000000 +0100
3981 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/fs_struct.c   2011-06-10 22:11:24.000000000 +0200
3982 @@ -4,6 +4,7 @@
3983  #include <linux/path.h>
3984  #include <linux/slab.h>
3985  #include <linux/fs_struct.h>
3986 +#include <linux/vserver/global.h>
3987  #include "internal.h"
3988
3989  static inline void path_get_longterm(struct path *path)
3990 @@ -96,6 +97,7 @@ void free_fs_struct(struct fs_struct *fs
3991  {
3992         path_put_longterm(&fs->root);
3993         path_put_longterm(&fs->pwd);
3994 +       atomic_dec(&vs_global_fs);
3995         kmem_cache_free(fs_cachep, fs);
3996  }
3997
3998 @@ -135,6 +137,7 @@ struct fs_struct *copy_fs_struct(struct
3999                 fs->pwd = old->pwd;
4000                 path_get_longterm(&fs->pwd);
4001                 spin_unlock(&old->lock);
4002 +               atomic_inc(&vs_global_fs);
4003         }
4004         return fs;
4005  }
4006 diff -NurpP --minimal linux-3.0-rc5/fs/gfs2/file.c linux-3.0-rc5-vs2.3.1-pre3/fs/gfs2/file.c
4007 --- linux-3.0-rc5/fs/gfs2/file.c        2011-07-01 11:14:18.000000000 +0200
4008 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/gfs2/file.c   2011-06-10 22:11:24.000000000 +0200
4009 @@ -134,6 +134,9 @@ static const u32 fsflags_to_gfs2[32] = {
4010         [7] = GFS2_DIF_NOATIME,
4011         [12] = GFS2_DIF_EXHASH,
4012         [14] = GFS2_DIF_INHERIT_JDATA,
4013 +       [27] = GFS2_DIF_IXUNLINK,
4014 +       [26] = GFS2_DIF_BARRIER,
4015 +       [29] = GFS2_DIF_COW,
4016  };
4017
4018  static const u32 gfs2_to_fsflags[32] = {
4019 @@ -143,6 +146,9 @@ static const u32 gfs2_to_fsflags[32] = {
4020         [gfs2fl_NoAtime] = FS_NOATIME_FL,
4021         [gfs2fl_ExHash] = FS_INDEX_FL,
4022         [gfs2fl_InheritJdata] = FS_JOURNAL_DATA_FL,
4023 +       [gfs2fl_IXUnlink] = FS_IXUNLINK_FL,
4024 +       [gfs2fl_Barrier] = FS_BARRIER_FL,
4025 +       [gfs2fl_Cow] = FS_COW_FL,
4026  };
4027
4028  static int gfs2_get_flags(struct file *filp, u32 __user *ptr)
4029 @@ -173,10 +179,16 @@ void gfs2_set_inode_flags(struct inode *
4030  {
4031         struct gfs2_inode *ip = GFS2_I(inode);
4032         unsigned int flags = inode->i_flags;
4033 +       unsigned int vflags = inode->i_vflags;
4034 +
4035 +       flags &= ~(S_IMMUTABLE | S_IXUNLINK |
4036 +               S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
4037
4038 -       flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC);
4039         if (ip->i_diskflags & GFS2_DIF_IMMUTABLE)
4040                 flags |= S_IMMUTABLE;
4041 +       if (ip->i_diskflags & GFS2_DIF_IXUNLINK)
4042 +               flags |= S_IXUNLINK;
4043 +
4044         if (ip->i_diskflags & GFS2_DIF_APPENDONLY)
4045                 flags |= S_APPEND;
4046         if (ip->i_diskflags & GFS2_DIF_NOATIME)
4047 @@ -184,6 +196,43 @@ void gfs2_set_inode_flags(struct inode *
4048         if (ip->i_diskflags & GFS2_DIF_SYNC)
4049                 flags |= S_SYNC;
4050         inode->i_flags = flags;
4051 +
4052 +       vflags &= ~(V_BARRIER | V_COW);
4053 +
4054 +       if (ip->i_diskflags & GFS2_DIF_BARRIER)
4055 +               vflags |= V_BARRIER;
4056 +       if (ip->i_diskflags & GFS2_DIF_COW)
4057 +               vflags |= V_COW;
4058 +       inode->i_vflags = vflags;
4059 +}
4060 +
4061 +void gfs2_get_inode_flags(struct inode *inode)
4062 +{
4063 +       struct gfs2_inode *ip = GFS2_I(inode);
4064 +       unsigned int flags = inode->i_flags;
4065 +       unsigned int vflags = inode->i_vflags;
4066 +
4067 +       ip->i_diskflags &= ~(GFS2_DIF_APPENDONLY |
4068 +                       GFS2_DIF_NOATIME | GFS2_DIF_SYNC |
4069 +                       GFS2_DIF_IMMUTABLE | GFS2_DIF_IXUNLINK |
4070 +                       GFS2_DIF_BARRIER | GFS2_DIF_COW);
4071 +
4072 +       if (flags & S_IMMUTABLE)
4073 +               ip->i_diskflags |= GFS2_DIF_IMMUTABLE;
4074 +       if (flags & S_IXUNLINK)
4075 +               ip->i_diskflags |= GFS2_DIF_IXUNLINK;
4076 +
4077 +       if (flags & S_APPEND)
4078 +               ip->i_diskflags |= GFS2_DIF_APPENDONLY;
4079 +       if (flags & S_NOATIME)
4080 +               ip->i_diskflags |= GFS2_DIF_NOATIME;
4081 +       if (flags & S_SYNC)
4082 +               ip->i_diskflags |= GFS2_DIF_SYNC;
4083 +
4084 +       if (vflags & V_BARRIER)
4085 +               ip->i_diskflags |= GFS2_DIF_BARRIER;
4086 +       if (vflags & V_COW)
4087 +               ip->i_diskflags |= GFS2_DIF_COW;
4088  }
4089
4090  /* Flags that can be set by user space */
4091 @@ -295,6 +344,37 @@ static int gfs2_set_flags(struct file *f
4092         return do_gfs2_set_flags(filp, gfsflags, ~GFS2_DIF_JDATA);
4093  }
4094
4095 +int gfs2_sync_flags(struct inode *inode, int flags, int vflags)
4096 +{
4097 +       struct gfs2_inode *ip = GFS2_I(inode);
4098 +       struct gfs2_sbd *sdp = GFS2_SB(inode);
4099 +       struct buffer_head *bh;
4100 +       struct gfs2_holder gh;
4101 +       int error;
4102 +
4103 +       error = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
4104 +       if (error)
4105 +               return error;
4106 +       error = gfs2_trans_begin(sdp, RES_DINODE, 0);
4107 +       if (error)
4108 +               goto out;
4109 +       error = gfs2_meta_inode_buffer(ip, &bh);
4110 +       if (error)
4111 +               goto out_trans_end;
4112 +       gfs2_trans_add_bh(ip->i_gl, bh, 1);
4113 +       inode->i_flags = flags;
4114 +       inode->i_vflags = vflags;
4115 +       gfs2_get_inode_flags(inode);
4116 +       gfs2_dinode_out(ip, bh->b_data);
4117 +       brelse(bh);
4118 +       gfs2_set_aops(inode);
4119 +out_trans_end:
4120 +       gfs2_trans_end(sdp);
4121 +out:
4122 +       gfs2_glock_dq_uninit(&gh);
4123 +       return error;
4124 +}
4125 +
4126  static long gfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
4127  {
4128         switch(cmd) {
4129 diff -NurpP --minimal linux-3.0-rc5/fs/gfs2/inode.h linux-3.0-rc5-vs2.3.1-pre3/fs/gfs2/inode.h
4130 --- linux-3.0-rc5/fs/gfs2/inode.h       2011-07-01 11:14:18.000000000 +0200
4131 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/gfs2/inode.h  2011-06-10 22:11:24.000000000 +0200
4132 @@ -120,6 +120,7 @@ extern const struct file_operations gfs2
4133  extern const struct file_operations gfs2_dir_fops_nolock;
4134
4135  extern void gfs2_set_inode_flags(struct inode *inode);
4136 +extern int gfs2_sync_flags(struct inode *inode, int flags, int vflags);
4137
4138  #ifdef CONFIG_GFS2_FS_LOCKING_DLM
4139  extern const struct file_operations gfs2_file_fops;
4140 diff -NurpP --minimal linux-3.0-rc5/fs/inode.c linux-3.0-rc5-vs2.3.1-pre3/fs/inode.c
4141 --- linux-3.0-rc5/fs/inode.c    2011-07-01 11:14:18.000000000 +0200
4142 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/inode.c       2011-07-01 11:35:34.000000000 +0200
4143 @@ -26,6 +26,7 @@
4144  #include <linux/ima.h>
4145  #include <linux/cred.h>
4146  #include <linux/buffer_head.h> /* for inode_has_buffers */
4147 +#include <linux/vs_tag.h>
4148  #include "internal.h"
4149
4150  /*
4151 @@ -146,6 +147,9 @@ int inode_init_always(struct super_block
4152         struct address_space *const mapping = &inode->i_data;
4153
4154         inode->i_sb = sb;
4155 +
4156 +       /* essential because of inode slab reuse */
4157 +       inode->i_tag = 0;
4158         inode->i_blkbits = sb->s_blocksize_bits;
4159         inode->i_flags = 0;
4160         atomic_set(&inode->i_count, 1);
4161 @@ -166,6 +170,7 @@ int inode_init_always(struct super_block
4162         inode->i_bdev = NULL;
4163         inode->i_cdev = NULL;
4164         inode->i_rdev = 0;
4165 +       inode->i_mdev = 0;
4166         inode->dirtied_when = 0;
4167
4168         if (security_inode_alloc(inode))
4169 @@ -404,6 +409,8 @@ void __insert_inode_hash(struct inode *i
4170  }
4171  EXPORT_SYMBOL(__insert_inode_hash);
4172
4173 +EXPORT_SYMBOL_GPL(__iget);
4174 +
4175  /**
4176   *     remove_inode_hash - remove an inode from the hash
4177   *     @inode: inode to unhash
4178 @@ -1643,9 +1650,11 @@ void init_special_inode(struct inode *in
4179         if (S_ISCHR(mode)) {
4180                 inode->i_fop = &def_chr_fops;
4181                 inode->i_rdev = rdev;
4182 +               inode->i_mdev = rdev;
4183         } else if (S_ISBLK(mode)) {
4184                 inode->i_fop = &def_blk_fops;
4185                 inode->i_rdev = rdev;
4186 +               inode->i_mdev = rdev;
4187         } else if (S_ISFIFO(mode))
4188                 inode->i_fop = &def_fifo_fops;
4189         else if (S_ISSOCK(mode))
4190 @@ -1674,6 +1683,7 @@ void inode_init_owner(struct inode *inod
4191         } else
4192                 inode->i_gid = current_fsgid();
4193         inode->i_mode = mode;
4194 +       inode->i_tag = dx_current_fstag(inode->i_sb);
4195  }
4196  EXPORT_SYMBOL(inode_init_owner);
4197
4198 diff -NurpP --minimal linux-3.0-rc5/fs/ioctl.c linux-3.0-rc5-vs2.3.1-pre3/fs/ioctl.c
4199 --- linux-3.0-rc5/fs/ioctl.c    2011-05-22 16:17:52.000000000 +0200
4200 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ioctl.c       2011-06-10 22:11:24.000000000 +0200
4201 @@ -15,6 +15,9 @@
4202  #include <linux/writeback.h>
4203  #include <linux/buffer_head.h>
4204  #include <linux/falloc.h>
4205 +#include <linux/proc_fs.h>
4206 +#include <linux/vserver/inode.h>
4207 +#include <linux/vs_tag.h>
4208
4209  #include <asm/ioctls.h>
4210
4211 diff -NurpP --minimal linux-3.0-rc5/fs/ioprio.c linux-3.0-rc5-vs2.3.1-pre3/fs/ioprio.c
4212 --- linux-3.0-rc5/fs/ioprio.c   2011-01-05 21:50:24.000000000 +0100
4213 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ioprio.c      2011-06-10 22:11:24.000000000 +0200
4214 @@ -27,6 +27,7 @@
4215  #include <linux/syscalls.h>
4216  #include <linux/security.h>
4217  #include <linux/pid_namespace.h>
4218 +#include <linux/vs_base.h>
4219
4220  int set_task_ioprio(struct task_struct *task, int ioprio)
4221  {
4222 @@ -119,6 +120,8 @@ SYSCALL_DEFINE3(ioprio_set, int, which,
4223                         else
4224                                 pgrp = find_vpid(who);
4225                         do_each_pid_thread(pgrp, PIDTYPE_PGID, p) {
4226 +                               if (!vx_check(p->xid, VS_ADMIN_P | VS_IDENT))
4227 +                                       continue;
4228                                 ret = set_task_ioprio(p, ioprio);
4229                                 if (ret)
4230                                         break;
4231 @@ -208,6 +211,8 @@ SYSCALL_DEFINE2(ioprio_get, int, which,
4232                         else
4233                                 pgrp = find_vpid(who);
4234                         do_each_pid_thread(pgrp, PIDTYPE_PGID, p) {
4235 +                               if (!vx_check(p->xid, VS_ADMIN_P | VS_IDENT))
4236 +                                       continue;
4237                                 tmpio = get_task_ioprio(p);
4238                                 if (tmpio < 0)
4239                                         continue;
4240 diff -NurpP --minimal linux-3.0-rc5/fs/jfs/file.c linux-3.0-rc5-vs2.3.1-pre3/fs/jfs/file.c
4241 --- linux-3.0-rc5/fs/jfs/file.c 2011-07-01 11:14:18.000000000 +0200
4242 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/jfs/file.c    2011-07-01 11:35:34.000000000 +0200
4243 @@ -102,7 +102,8 @@ int jfs_setattr(struct dentry *dentry, s
4244         if (is_quota_modification(inode, iattr))
4245                 dquot_initialize(inode);
4246         if ((iattr->ia_valid & ATTR_UID && iattr->ia_uid != inode->i_uid) ||
4247 -           (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid)) {
4248 +           (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid) ||
4249 +           (iattr->ia_valid & ATTR_TAG && iattr->ia_tag != inode->i_tag)) {
4250                 rc = dquot_transfer(inode, iattr);
4251                 if (rc)
4252                         return rc;
4253 @@ -133,6 +134,7 @@ const struct inode_operations jfs_file_i
4254  #ifdef CONFIG_JFS_POSIX_ACL
4255         .check_acl      = jfs_check_acl,
4256  #endif
4257 +       .sync_flags     = jfs_sync_flags,
4258  };
4259
4260  const struct file_operations jfs_file_operations = {
4261 diff -NurpP --minimal linux-3.0-rc5/fs/jfs/ioctl.c linux-3.0-rc5-vs2.3.1-pre3/fs/jfs/ioctl.c
4262 --- linux-3.0-rc5/fs/jfs/ioctl.c        2011-05-22 16:17:52.000000000 +0200
4263 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/jfs/ioctl.c   2011-06-10 22:11:24.000000000 +0200
4264 @@ -11,6 +11,7 @@
4265  #include <linux/mount.h>
4266  #include <linux/time.h>
4267  #include <linux/sched.h>
4268 +#include <linux/mount.h>
4269  #include <asm/current.h>
4270  #include <asm/uaccess.h>
4271
4272 @@ -52,6 +53,16 @@ static long jfs_map_ext2(unsigned long f
4273  }
4274
4275
4276 +int jfs_sync_flags(struct inode *inode, int flags, int vflags)
4277 +{
4278 +       inode->i_flags = flags;
4279 +       inode->i_vflags = vflags;
4280 +       jfs_get_inode_flags(JFS_IP(inode));
4281 +       inode->i_ctime = CURRENT_TIME_SEC;
4282 +       mark_inode_dirty(inode);
4283 +       return 0;
4284 +}
4285 +
4286  long jfs_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
4287  {
4288         struct inode *inode = filp->f_dentry->d_inode;
4289 @@ -85,6 +96,11 @@ long jfs_ioctl(struct file *filp, unsign
4290                 if (!S_ISDIR(inode->i_mode))
4291                         flags &= ~JFS_DIRSYNC_FL;
4292
4293 +               if (IS_BARRIER(inode)) {
4294 +                       vxwprintk_task(1, "messing with the barrier.");
4295 +                       return -EACCES;
4296 +               }
4297 +
4298                 /* Is it quota file? Do not allow user to mess with it */
4299                 if (IS_NOQUOTA(inode)) {
4300                         err = -EPERM;
4301 @@ -102,8 +118,8 @@ long jfs_ioctl(struct file *filp, unsign
4302                  * the relevant capability.
4303                  */
4304                 if ((oldflags & JFS_IMMUTABLE_FL) ||
4305 -                       ((flags ^ oldflags) &
4306 -                       (JFS_APPEND_FL | JFS_IMMUTABLE_FL))) {
4307 +                       ((flags ^ oldflags) & (JFS_APPEND_FL |
4308 +                       JFS_IMMUTABLE_FL | JFS_IXUNLINK_FL))) {
4309                         if (!capable(CAP_LINUX_IMMUTABLE)) {
4310                                 mutex_unlock(&inode->i_mutex);
4311                                 err = -EPERM;
4312 @@ -111,7 +127,7 @@ long jfs_ioctl(struct file *filp, unsign
4313                         }
4314                 }
4315
4316 -               flags = flags & JFS_FL_USER_MODIFIABLE;
4317 +               flags &= JFS_FL_USER_MODIFIABLE;
4318                 flags |= oldflags & ~JFS_FL_USER_MODIFIABLE;
4319                 jfs_inode->mode2 = flags;
4320
4321 diff -NurpP --minimal linux-3.0-rc5/fs/jfs/jfs_dinode.h linux-3.0-rc5-vs2.3.1-pre3/fs/jfs/jfs_dinode.h
4322 --- linux-3.0-rc5/fs/jfs/jfs_dinode.h   2008-12-25 00:26:37.000000000 +0100
4323 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/jfs/jfs_dinode.h      2011-06-10 22:11:24.000000000 +0200
4324 @@ -161,9 +161,13 @@ struct dinode {
4325
4326  #define JFS_APPEND_FL          0x01000000 /* writes to file may only append */
4327  #define JFS_IMMUTABLE_FL       0x02000000 /* Immutable file */
4328 +#define JFS_IXUNLINK_FL                0x08000000 /* Immutable invert on unlink */
4329
4330 -#define JFS_FL_USER_VISIBLE    0x03F80000
4331 -#define JFS_FL_USER_MODIFIABLE 0x03F80000
4332 +#define JFS_BARRIER_FL         0x04000000 /* Barrier for chroot() */
4333 +#define JFS_COW_FL             0x20000000 /* Copy on Write marker */
4334 +
4335 +#define JFS_FL_USER_VISIBLE    0x07F80000
4336 +#define JFS_FL_USER_MODIFIABLE 0x07F80000
4337  #define JFS_FL_INHERIT         0x03C80000
4338
4339  /* These are identical to EXT[23]_IOC_GETFLAGS/SETFLAGS */
4340 diff -NurpP --minimal linux-3.0-rc5/fs/jfs/jfs_filsys.h linux-3.0-rc5-vs2.3.1-pre3/fs/jfs/jfs_filsys.h
4341 --- linux-3.0-rc5/fs/jfs/jfs_filsys.h   2008-12-25 00:26:37.000000000 +0100
4342 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/jfs/jfs_filsys.h      2011-06-10 22:11:24.000000000 +0200
4343 @@ -263,6 +263,7 @@
4344  #define JFS_NAME_MAX   255
4345  #define JFS_PATH_MAX   BPSIZE
4346
4347 +#define JFS_TAGGED             0x00800000      /* Context Tagging */
4348
4349  /*
4350   *     file system state (superblock state)
4351 diff -NurpP --minimal linux-3.0-rc5/fs/jfs/jfs_imap.c linux-3.0-rc5-vs2.3.1-pre3/fs/jfs/jfs_imap.c
4352 --- linux-3.0-rc5/fs/jfs/jfs_imap.c     2011-07-01 11:14:18.000000000 +0200
4353 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/jfs/jfs_imap.c        2011-07-01 11:35:34.000000000 +0200
4354 @@ -46,6 +46,7 @@
4355  #include <linux/pagemap.h>
4356  #include <linux/quotaops.h>
4357  #include <linux/slab.h>
4358 +#include <linux/vs_tag.h>
4359
4360  #include "jfs_incore.h"
4361  #include "jfs_inode.h"
4362 @@ -3058,6 +3059,8 @@ static int copy_from_dinode(struct dinod
4363  {
4364         struct jfs_inode_info *jfs_ip = JFS_IP(ip);
4365         struct jfs_sb_info *sbi = JFS_SBI(ip->i_sb);
4366 +       uid_t uid;
4367 +       gid_t gid;
4368
4369         jfs_ip->fileset = le32_to_cpu(dip->di_fileset);
4370         jfs_ip->mode2 = le32_to_cpu(dip->di_mode);
4371 @@ -3078,14 +3081,18 @@ static int copy_from_dinode(struct dinod
4372         }
4373         ip->i_nlink = le32_to_cpu(dip->di_nlink);
4374
4375 -       jfs_ip->saved_uid = le32_to_cpu(dip->di_uid);
4376 +       uid = le32_to_cpu(dip->di_uid);
4377 +       gid = le32_to_cpu(dip->di_gid);
4378 +       ip->i_tag = INOTAG_TAG(DX_TAG(ip), uid, gid, 0);
4379 +
4380 +       jfs_ip->saved_uid = INOTAG_UID(DX_TAG(ip), uid, gid);
4381         if (sbi->uid == -1)
4382                 ip->i_uid = jfs_ip->saved_uid;
4383         else {
4384                 ip->i_uid = sbi->uid;
4385         }
4386
4387 -       jfs_ip->saved_gid = le32_to_cpu(dip->di_gid);
4388 +       jfs_ip->saved_gid = INOTAG_GID(DX_TAG(ip), uid, gid);
4389         if (sbi->gid == -1)
4390                 ip->i_gid = jfs_ip->saved_gid;
4391         else {
4392 @@ -3150,14 +3157,12 @@ static void copy_to_dinode(struct dinode
4393         dip->di_size = cpu_to_le64(ip->i_size);
4394         dip->di_nblocks = cpu_to_le64(PBLK2LBLK(ip->i_sb, ip->i_blocks));
4395         dip->di_nlink = cpu_to_le32(ip->i_nlink);
4396 -       if (sbi->uid == -1)
4397 -               dip->di_uid = cpu_to_le32(ip->i_uid);
4398 -       else
4399 -               dip->di_uid = cpu_to_le32(jfs_ip->saved_uid);
4400 -       if (sbi->gid == -1)
4401 -               dip->di_gid = cpu_to_le32(ip->i_gid);
4402 -       else
4403 -               dip->di_gid = cpu_to_le32(jfs_ip->saved_gid);
4404 +
4405 +       dip->di_uid = cpu_to_le32(TAGINO_UID(DX_TAG(ip),
4406 +               (sbi->uid == -1) ? ip->i_uid : jfs_ip->saved_uid, ip->i_tag));
4407 +       dip->di_gid = cpu_to_le32(TAGINO_GID(DX_TAG(ip),
4408 +               (sbi->gid == -1) ? ip->i_gid : jfs_ip->saved_gid, ip->i_tag));
4409 +
4410         jfs_get_inode_flags(jfs_ip);
4411         /*
4412          * mode2 is only needed for storing the higher order bits.
4413 diff -NurpP --minimal linux-3.0-rc5/fs/jfs/jfs_inode.c linux-3.0-rc5-vs2.3.1-pre3/fs/jfs/jfs_inode.c
4414 --- linux-3.0-rc5/fs/jfs/jfs_inode.c    2010-08-02 16:52:49.000000000 +0200
4415 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/jfs/jfs_inode.c       2011-06-10 22:11:24.000000000 +0200
4416 @@ -18,6 +18,7 @@
4417
4418  #include <linux/fs.h>
4419  #include <linux/quotaops.h>
4420 +#include <linux/vs_tag.h>
4421  #include "jfs_incore.h"
4422  #include "jfs_inode.h"
4423  #include "jfs_filsys.h"
4424 @@ -30,29 +31,46 @@ void jfs_set_inode_flags(struct inode *i
4425  {
4426         unsigned int flags = JFS_IP(inode)->mode2;
4427
4428 -       inode->i_flags &= ~(S_IMMUTABLE | S_APPEND |
4429 -               S_NOATIME | S_DIRSYNC | S_SYNC);
4430 +       inode->i_flags &= ~(S_IMMUTABLE | S_IXUNLINK |
4431 +               S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
4432
4433         if (flags & JFS_IMMUTABLE_FL)
4434                 inode->i_flags |= S_IMMUTABLE;
4435 +       if (flags & JFS_IXUNLINK_FL)
4436 +               inode->i_flags |= S_IXUNLINK;
4437 +
4438 +       if (flags & JFS_SYNC_FL)
4439 +               inode->i_flags |= S_SYNC;
4440         if (flags & JFS_APPEND_FL)
4441                 inode->i_flags |= S_APPEND;
4442         if (flags & JFS_NOATIME_FL)
4443                 inode->i_flags |= S_NOATIME;
4444         if (flags & JFS_DIRSYNC_FL)
4445                 inode->i_flags |= S_DIRSYNC;
4446 -       if (flags & JFS_SYNC_FL)
4447 -               inode->i_flags |= S_SYNC;
4448 +
4449 +       inode->i_vflags &= ~(V_BARRIER | V_COW);
4450 +
4451 +       if (flags & JFS_BARRIER_FL)
4452 +               inode->i_vflags |= V_BARRIER;
4453 +       if (flags & JFS_COW_FL)
4454 +               inode->i_vflags |= V_COW;
4455  }
4456
4457  void jfs_get_inode_flags(struct jfs_inode_info *jfs_ip)
4458  {
4459         unsigned int flags = jfs_ip->vfs_inode.i_flags;
4460 +       unsigned int vflags = jfs_ip->vfs_inode.i_vflags;
4461 +
4462 +       jfs_ip->mode2 &= ~(JFS_IMMUTABLE_FL | JFS_IXUNLINK_FL |
4463 +                          JFS_APPEND_FL | JFS_NOATIME_FL |
4464 +                          JFS_DIRSYNC_FL | JFS_SYNC_FL |
4465 +                          JFS_BARRIER_FL | JFS_COW_FL);
4466
4467 -       jfs_ip->mode2 &= ~(JFS_IMMUTABLE_FL | JFS_APPEND_FL | JFS_NOATIME_FL |
4468 -                          JFS_DIRSYNC_FL | JFS_SYNC_FL);
4469         if (flags & S_IMMUTABLE)
4470                 jfs_ip->mode2 |= JFS_IMMUTABLE_FL;
4471 +       if (flags & S_IXUNLINK)
4472 +               jfs_ip->mode2 |= JFS_IXUNLINK_FL;
4473 +
4474         if (flags & S_APPEND)
4475                 jfs_ip->mode2 |= JFS_APPEND_FL;
4476         if (flags & S_NOATIME)
4477 @@ -61,6 +79,11 @@ void jfs_get_inode_flags(struct jfs_inod
4478                 jfs_ip->mode2 |= JFS_DIRSYNC_FL;
4479         if (flags & S_SYNC)
4480                 jfs_ip->mode2 |= JFS_SYNC_FL;
4481 +
4482 +       if (vflags & V_BARRIER)
4483 +               jfs_ip->mode2 |= JFS_BARRIER_FL;
4484 +       if (vflags & V_COW)
4485 +               jfs_ip->mode2 |= JFS_COW_FL;
4486  }
4487
4488  /*
4489 diff -NurpP --minimal linux-3.0-rc5/fs/jfs/jfs_inode.h linux-3.0-rc5-vs2.3.1-pre3/fs/jfs/jfs_inode.h
4490 --- linux-3.0-rc5/fs/jfs/jfs_inode.h    2011-07-01 11:14:18.000000000 +0200
4491 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/jfs/jfs_inode.h       2011-06-10 22:11:24.000000000 +0200
4492 @@ -39,6 +39,7 @@ extern struct dentry *jfs_fh_to_dentry(s
4493  extern struct dentry *jfs_fh_to_parent(struct super_block *sb, struct fid *fid,
4494         int fh_len, int fh_type);
4495  extern void jfs_set_inode_flags(struct inode *);
4496 +extern int jfs_sync_flags(struct inode *, int, int);
4497  extern int jfs_get_block(struct inode *, sector_t, struct buffer_head *, int);
4498  extern int jfs_setattr(struct dentry *, struct iattr *);
4499
4500 diff -NurpP --minimal linux-3.0-rc5/fs/jfs/namei.c linux-3.0-rc5-vs2.3.1-pre3/fs/jfs/namei.c
4501 --- linux-3.0-rc5/fs/jfs/namei.c        2011-05-22 16:17:53.000000000 +0200
4502 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/jfs/namei.c   2011-06-10 22:11:24.000000000 +0200
4503 @@ -22,6 +22,7 @@
4504  #include <linux/ctype.h>
4505  #include <linux/quotaops.h>
4506  #include <linux/exportfs.h>
4507 +#include <linux/vs_tag.h>
4508  #include "jfs_incore.h"
4509  #include "jfs_superblock.h"
4510  #include "jfs_inode.h"
4511 @@ -1486,6 +1487,7 @@ static struct dentry *jfs_lookup(struct
4512                 return ERR_CAST(ip);
4513         }
4514
4515 +       dx_propagate_tag(nd, ip);
4516         return d_splice_alias(ip, dentry);
4517  }
4518
4519 @@ -1550,6 +1552,7 @@ const struct inode_operations jfs_dir_in
4520  #ifdef CONFIG_JFS_POSIX_ACL
4521         .check_acl      = jfs_check_acl,
4522  #endif
4523 +       .sync_flags     = jfs_sync_flags,
4524  };
4525
4526  const struct file_operations jfs_dir_operations = {
4527 diff -NurpP --minimal linux-3.0-rc5/fs/jfs/super.c linux-3.0-rc5-vs2.3.1-pre3/fs/jfs/super.c
4528 --- linux-3.0-rc5/fs/jfs/super.c        2011-05-22 16:17:53.000000000 +0200
4529 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/jfs/super.c   2011-06-10 22:11:24.000000000 +0200
4530 @@ -198,7 +198,8 @@ static void jfs_put_super(struct super_b
4531  enum {
4532         Opt_integrity, Opt_nointegrity, Opt_iocharset, Opt_resize,
4533         Opt_resize_nosize, Opt_errors, Opt_ignore, Opt_err, Opt_quota,
4534 -       Opt_usrquota, Opt_grpquota, Opt_uid, Opt_gid, Opt_umask
4535 +       Opt_usrquota, Opt_grpquota, Opt_uid, Opt_gid, Opt_umask,
4536 +       Opt_tag, Opt_notag, Opt_tagid
4537  };
4538
4539  static const match_table_t tokens = {
4540 @@ -208,6 +209,10 @@ static const match_table_t tokens = {
4541         {Opt_resize, "resize=%u"},
4542         {Opt_resize_nosize, "resize"},
4543         {Opt_errors, "errors=%s"},
4544 +       {Opt_tag, "tag"},
4545 +       {Opt_notag, "notag"},
4546 +       {Opt_tagid, "tagid=%u"},
4547 +       {Opt_tag, "tagxid"},
4548         {Opt_ignore, "noquota"},
4549         {Opt_ignore, "quota"},
4550         {Opt_usrquota, "usrquota"},
4551 @@ -342,6 +347,20 @@ static int parse_options(char *options,
4552                         }
4553                         break;
4554                 }
4555 +#ifndef CONFIG_TAGGING_NONE
4556 +               case Opt_tag:
4557 +                       *flag |= JFS_TAGGED;
4558 +                       break;
4559 +               case Opt_notag:
4560 +                       *flag &= JFS_TAGGED;
4561 +                       break;
4562 +#endif
4563 +#ifdef CONFIG_PROPAGATE
4564 +               case Opt_tagid:
4565 +                       /* use args[0] */
4566 +                       *flag |= JFS_TAGGED;
4567 +                       break;
4568 +#endif
4569                 default:
4570                         printk("jfs: Unrecognized mount option \"%s\" "
4571                                         " or missing value\n", p);
4572 @@ -373,6 +392,12 @@ static int jfs_remount(struct super_bloc
4573                 return -EINVAL;
4574         }
4575
4576 +       if ((flag & JFS_TAGGED) && !(sb->s_flags & MS_TAGGED)) {
4577 +               printk(KERN_ERR "JFS: %s: tagging not permitted on remount.\n",
4578 +                       sb->s_id);
4579 +               return -EINVAL;
4580 +       }
4581 +
4582         if (newLVSize) {
4583                 if (sb->s_flags & MS_RDONLY) {
4584                         printk(KERN_ERR
4585 @@ -455,6 +480,9 @@ static int jfs_fill_super(struct super_b
4586  #ifdef CONFIG_JFS_POSIX_ACL
4587         sb->s_flags |= MS_POSIXACL;
4588  #endif
4589 +       /* map mount option tagxid */
4590 +       if (sbi->flag & JFS_TAGGED)
4591 +               sb->s_flags |= MS_TAGGED;
4592
4593         if (newLVSize) {
4594                 printk(KERN_ERR "resize option for remount only\n");
4595 diff -NurpP --minimal linux-3.0-rc5/fs/libfs.c linux-3.0-rc5-vs2.3.1-pre3/fs/libfs.c
4596 --- linux-3.0-rc5/fs/libfs.c    2011-03-15 18:07:32.000000000 +0100
4597 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/libfs.c       2011-06-10 22:11:24.000000000 +0200
4598 @@ -133,7 +133,8 @@ static inline unsigned char dt_type(stru
4599   * both impossible due to the lock on directory.
4600   */
4601
4602 -int dcache_readdir(struct file * filp, void * dirent, filldir_t filldir)
4603 +static inline int do_dcache_readdir_filter(struct file *filp,
4604 +       void *dirent, filldir_t filldir, int (*filter)(struct dentry *dentry))
4605  {
4606         struct dentry *dentry = filp->f_path.dentry;
4607         struct dentry *cursor = filp->private_data;
4608 @@ -164,6 +165,8 @@ int dcache_readdir(struct file * filp, v
4609                         for (p=q->next; p != &dentry->d_subdirs; p=p->next) {
4610                                 struct dentry *next;
4611                                 next = list_entry(p, struct dentry, d_u.d_child);
4612 +                               if (filter && !filter(next))
4613 +                                       continue;
4614                                 spin_lock_nested(&next->d_lock, DENTRY_D_LOCK_NESTED);
4615                                 if (!simple_positive(next)) {
4616                                         spin_unlock(&next->d_lock);
4617 @@ -190,6 +193,17 @@ int dcache_readdir(struct file * filp, v
4618         return 0;
4619  }
4620
4621 +int dcache_readdir(struct file *filp, void *dirent, filldir_t filldir)
4622 +{
4623 +       return do_dcache_readdir_filter(filp, dirent, filldir, NULL);
4624 +}
4625 +
4626 +int dcache_readdir_filter(struct file *filp, void *dirent, filldir_t filldir,
4627 +       int (*filter)(struct dentry *))
4628 +{
4629 +       return do_dcache_readdir_filter(filp, dirent, filldir, filter);
4630 +}
4631 +
4632  ssize_t generic_read_dir(struct file *filp, char __user *buf, size_t siz, loff_t *ppos)
4633  {
4634         return -EISDIR;
4635 @@ -965,6 +979,7 @@ EXPORT_SYMBOL(dcache_dir_close);
4636  EXPORT_SYMBOL(dcache_dir_lseek);
4637  EXPORT_SYMBOL(dcache_dir_open);
4638  EXPORT_SYMBOL(dcache_readdir);
4639 +EXPORT_SYMBOL(dcache_readdir_filter);
4640  EXPORT_SYMBOL(generic_read_dir);
4641  EXPORT_SYMBOL(mount_pseudo);
4642  EXPORT_SYMBOL(simple_write_begin);
4643 diff -NurpP --minimal linux-3.0-rc5/fs/locks.c linux-3.0-rc5-vs2.3.1-pre3/fs/locks.c
4644 --- linux-3.0-rc5/fs/locks.c    2011-05-22 16:17:53.000000000 +0200
4645 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/locks.c       2011-06-10 22:11:24.000000000 +0200
4646 @@ -126,6 +126,8 @@
4647  #include <linux/time.h>
4648  #include <linux/rcupdate.h>
4649  #include <linux/pid_namespace.h>
4650 +#include <linux/vs_base.h>
4651 +#include <linux/vs_limit.h>
4652
4653  #include <asm/uaccess.h>
4654
4655 @@ -163,6 +165,8 @@ static struct kmem_cache *filelock_cache
4656  /* Allocate an empty lock structure. */
4657  struct file_lock *locks_alloc_lock(void)
4658  {
4659 +       if (!vx_locks_avail(1))
4660 +               return NULL;
4661         return kmem_cache_alloc(filelock_cache, GFP_KERNEL);
4662  }
4663  EXPORT_SYMBOL_GPL(locks_alloc_lock);
4664 @@ -190,6 +194,7 @@ void locks_free_lock(struct file_lock *f
4665         BUG_ON(!list_empty(&fl->fl_block));
4666         BUG_ON(!list_empty(&fl->fl_link));
4667
4668 +       vx_locks_dec(fl);
4669         locks_release_private(fl);
4670         kmem_cache_free(filelock_cache, fl);
4671  }
4672 @@ -211,6 +216,7 @@ void locks_init_lock(struct file_lock *f
4673         fl->fl_start = fl->fl_end = 0;
4674         fl->fl_ops = NULL;
4675         fl->fl_lmops = NULL;
4676 +       fl->fl_xid = -1;
4677  }
4678
4679  EXPORT_SYMBOL(locks_init_lock);
4680 @@ -262,6 +268,7 @@ void locks_copy_lock(struct file_lock *n
4681         new->fl_file = fl->fl_file;
4682         new->fl_ops = fl->fl_ops;
4683         new->fl_lmops = fl->fl_lmops;
4684 +       new->fl_xid = fl->fl_xid;
4685
4686         locks_copy_private(new, fl);
4687  }
4688 @@ -300,6 +307,11 @@ static int flock_make_lock(struct file *
4689         fl->fl_flags = FL_FLOCK;
4690         fl->fl_type = type;
4691         fl->fl_end = OFFSET_MAX;
4692 +
4693 +       vxd_assert(filp->f_xid == vx_current_xid(),
4694 +               "f_xid(%d) == current(%d)", filp->f_xid, vx_current_xid());
4695 +       fl->fl_xid = filp->f_xid;
4696 +       vx_locks_inc(fl);
4697
4698         *lock = fl;
4699         return 0;
4700 @@ -449,6 +461,7 @@ static int lease_init(struct file *filp,
4701
4702         fl->fl_owner = current->files;
4703         fl->fl_pid = current->tgid;
4704 +       fl->fl_xid = vx_current_xid();
4705
4706         fl->fl_file = filp;
4707         fl->fl_flags = FL_LEASE;
4708 @@ -468,6 +481,11 @@ static struct file_lock *lease_alloc(str
4709         if (fl == NULL)
4710                 return ERR_PTR(error);
4711
4712 +       fl->fl_xid = vx_current_xid();
4713 +       if (filp)
4714 +               vxd_assert(filp->f_xid == fl->fl_xid,
4715 +                       "f_xid(%d) == fl_xid(%d)", filp->f_xid, fl->fl_xid);
4716 +       vx_locks_inc(fl);
4717         error = lease_init(filp, type, fl);
4718         if (error) {
4719                 locks_free_lock(fl);
4720 @@ -769,6 +787,7 @@ static int flock_lock_file(struct file *
4721                 lock_flocks();
4722         }
4723
4724 +       new_fl->fl_xid = -1;
4725  find_conflict:
4726         for_each_lock(inode, before) {
4727                 struct file_lock *fl = *before;
4728 @@ -789,6 +808,7 @@ find_conflict:
4729                 goto out;
4730         locks_copy_lock(new_fl, request);
4731         locks_insert_lock(before, new_fl);
4732 +       vx_locks_inc(new_fl);
4733         new_fl = NULL;
4734         error = 0;
4735
4736 @@ -799,7 +819,8 @@ out:
4737         return error;
4738  }
4739
4740 -static int __posix_lock_file(struct inode *inode, struct file_lock *request, struct file_lock *conflock)
4741 +static int __posix_lock_file(struct inode *inode, struct file_lock *request,
4742 +       struct file_lock *conflock, xid_t xid)
4743  {
4744         struct file_lock *fl;
4745         struct file_lock *new_fl = NULL;
4746 @@ -809,6 +830,8 @@ static int __posix_lock_file(struct inod
4747         struct file_lock **before;
4748         int error, added = 0;
4749
4750 +       vxd_assert(xid == vx_current_xid(),
4751 +               "xid(%d) == current(%d)", xid, vx_current_xid());
4752         /*
4753          * We may need two file_lock structures for this operation,
4754          * so we get them in advance to avoid races.
4755 @@ -819,7 +842,11 @@ static int __posix_lock_file(struct inod
4756             (request->fl_type != F_UNLCK ||
4757              request->fl_start != 0 || request->fl_end != OFFSET_MAX)) {
4758                 new_fl = locks_alloc_lock();
4759 +               new_fl->fl_xid = xid;
4760 +               vx_locks_inc(new_fl);
4761                 new_fl2 = locks_alloc_lock();
4762 +               new_fl2->fl_xid = xid;
4763 +               vx_locks_inc(new_fl2);
4764         }
4765
4766         lock_flocks();
4767 @@ -1018,7 +1045,8 @@ static int __posix_lock_file(struct inod
4768  int posix_lock_file(struct file *filp, struct file_lock *fl,
4769                         struct file_lock *conflock)
4770  {
4771 -       return __posix_lock_file(filp->f_path.dentry->d_inode, fl, conflock);
4772 +       return __posix_lock_file(filp->f_path.dentry->d_inode,
4773 +               fl, conflock, filp->f_xid);
4774  }
4775  EXPORT_SYMBOL(posix_lock_file);
4776
4777 @@ -1108,7 +1136,7 @@ int locks_mandatory_area(int read_write,
4778         fl.fl_end = offset + count - 1;
4779
4780         for (;;) {
4781 -               error = __posix_lock_file(inode, &fl, NULL);
4782 +               error = __posix_lock_file(inode, &fl, NULL, filp->f_xid);
4783                 if (error != FILE_LOCK_DEFERRED)
4784                         break;
4785                 error = wait_event_interruptible(fl.fl_wait, !fl.fl_next);
4786 @@ -1421,6 +1449,7 @@ int generic_setlease(struct file *filp,
4787                 goto out;
4788
4789         locks_insert_lock(before, lease);
4790 +       vx_locks_inc(lease);
4791         return 0;
4792
4793  out:
4794 @@ -1805,6 +1834,11 @@ int fcntl_setlk(unsigned int fd, struct
4795         if (file_lock == NULL)
4796                 return -ENOLCK;
4797
4798 +       vxd_assert(filp->f_xid == vx_current_xid(),
4799 +               "f_xid(%d) == current(%d)", filp->f_xid, vx_current_xid());
4800 +       file_lock->fl_xid = filp->f_xid;
4801 +       vx_locks_inc(file_lock);
4802 +
4803         /*
4804          * This might block, so we do it before checking the inode.
4805          */
4806 @@ -1923,6 +1957,11 @@ int fcntl_setlk64(unsigned int fd, struc
4807         if (file_lock == NULL)
4808                 return -ENOLCK;
4809
4810 +       vxd_assert(filp->f_xid == vx_current_xid(),
4811 +               "f_xid(%d) == current(%d)", filp->f_xid, vx_current_xid());
4812 +       file_lock->fl_xid = filp->f_xid;
4813 +       vx_locks_inc(file_lock);
4814 +
4815         /*
4816          * This might block, so we do it before checking the inode.
4817          */
4818 @@ -2188,8 +2227,11 @@ static int locks_show(struct seq_file *f
4819
4820         lock_get_status(f, fl, *((loff_t *)f->private), "");
4821
4822 -       list_for_each_entry(bfl, &fl->fl_block, fl_block)
4823 +       list_for_each_entry(bfl, &fl->fl_block, fl_block) {
4824 +               if (!vx_check(fl->fl_xid, VS_WATCH_P | VS_IDENT))
4825 +                       continue;
4826                 lock_get_status(f, bfl, *((loff_t *)f->private), " ->");
4827 +       }
4828
4829         return 0;
4830  }
4831 diff -NurpP --minimal linux-3.0-rc5/fs/namei.c linux-3.0-rc5-vs2.3.1-pre3/fs/namei.c
4832 --- linux-3.0-rc5/fs/namei.c    2011-07-01 11:14:19.000000000 +0200
4833 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/namei.c       2011-06-22 12:39:15.000000000 +0200
4834 @@ -32,6 +32,14 @@
4835  #include <linux/fcntl.h>
4836  #include <linux/device_cgroup.h>
4837  #include <linux/fs_struct.h>
4838 +#include <linux/proc_fs.h>
4839 +#include <linux/vserver/inode.h>
4840 +#include <linux/vs_base.h>
4841 +#include <linux/vs_tag.h>
4842 +#include <linux/vs_cowbl.h>
4843 +#include <linux/vs_device.h>
4844 +#include <linux/vs_context.h>
4845 +#include <linux/pid_namespace.h>
4846  #include <asm/uaccess.h>
4847
4848  #include "internal.h"
4849 @@ -173,6 +181,84 @@ void putname(const char *name)
4850  EXPORT_SYMBOL(putname);
4851  #endif
4852
4853 +static inline int dx_barrier(const struct inode *inode)
4854 +{
4855 +       if (IS_BARRIER(inode) && !vx_check(0, VS_ADMIN | VS_WATCH)) {
4856 +               vxwprintk_task(1, "did hit the barrier.");
4857 +               return 1;
4858 +       }
4859 +       return 0;
4860 +}
4861 +
4862 +static int __dx_permission(const struct inode *inode, int mask)
4863 +{
4864 +       if (dx_barrier(inode))
4865 +               return -EACCES;
4866 +
4867 +       if (inode->i_sb->s_magic == DEVPTS_SUPER_MAGIC) {
4868 +               /* devpts is xid tagged */
4869 +               if (S_ISDIR(inode->i_mode) ||
4870 +                   vx_check((xid_t)inode->i_tag, VS_IDENT | VS_WATCH_P))
4871 +                       return 0;
4872 +
4873 +               /* just pretend we didn't find anything */
4874 +               return -ENOENT;
4875 +       }
4876 +       else if (inode->i_sb->s_magic == PROC_SUPER_MAGIC) {
4877 +               struct proc_dir_entry *de = PDE(inode);
4878 +
4879 +               if (de && !vx_hide_check(0, de->vx_flags))
4880 +                       goto out;
4881 +
4882 +               if ((mask & (MAY_WRITE | MAY_APPEND))) {
4883 +                       struct pid *pid;
4884 +                       struct task_struct *tsk;
4885 +
4886 +                       if (vx_check(0, VS_ADMIN | VS_WATCH_P) ||
4887 +                           vx_flags(VXF_STATE_SETUP, 0))
4888 +                               return 0;
4889 +
4890 +                       pid = PROC_I(inode)->pid;
4891 +                       if (!pid)
4892 +                               goto out;
4893 +
4894 +                       tsk = pid_task(pid, PIDTYPE_PID);
4895 +                       vxdprintk(VXD_CBIT(tag, 0), "accessing %p[#%u]",
4896 +                                 tsk, (tsk ? vx_task_xid(tsk) : 0));
4897 +                       if (tsk && vx_check(vx_task_xid(tsk), VS_IDENT | VS_WATCH_P))
4898 +                               return 0;
4899 +               }
4900 +               else {
4901 +                       /* FIXME: Should we block some entries here? */
4902 +                       return 0;
4903 +               }
4904 +       }
4905 +       else {
4906 +               if (dx_notagcheck(inode->i_sb) ||
4907 +                   dx_check(inode->i_tag, DX_HOSTID | DX_ADMIN | DX_WATCH |
4908 +                            DX_IDENT))
4909 +                       return 0;
4910 +       }
4911 +
4912 +out:
4913 +       return -EACCES;
4914 +}
4915 +
4916 +int dx_permission(const struct inode *inode, int mask)
4917 +{
4918 +       int ret = __dx_permission(inode, mask);
4919 +       if (unlikely(ret)) {
4920 +#ifndef        CONFIG_VSERVER_WARN_DEVPTS
4921 +               if (inode->i_sb->s_magic != DEVPTS_SUPER_MAGIC)
4922 +#endif
4923 +                   vxwprintk_task(1,
4924 +                       "denied [0x%x] access to inode %s:%p[#%d,%lu]",
4925 +                       mask, inode->i_sb->s_id, inode, inode->i_tag,
4926 +                       inode->i_ino);
4927 +       }
4928 +       return ret;
4929 +}
4930 +
4931  /*
4932   * This does basic POSIX ACL permission checking
4933   */
4934 @@ -283,10 +369,14 @@ int inode_permission(struct inode *inode
4935                 /*
4936                  * Nobody gets write access to an immutable file.
4937                  */
4938 -               if (IS_IMMUTABLE(inode))
4939 +               if (IS_IMMUTABLE(inode) && !IS_COW(inode))
4940                         return -EACCES;
4941         }
4942
4943 +       retval = dx_permission(inode, mask);
4944 +       if (retval)
4945 +               return retval;
4946 +
4947         if (inode->i_op->permission)
4948                 retval = inode->i_op->permission(inode, mask, 0);
4949         else
4950 @@ -580,6 +670,9 @@ static inline int exec_permission(struct
4951         int ret;
4952         struct user_namespace *ns = inode_userns(inode);
4953
4954 +       if (dx_barrier(inode))
4955 +               return -EACCES;
4956 +
4957         if (inode->i_op->permission) {
4958                 ret = inode->i_op->permission(inode, MAY_EXEC, flags);
4959         } else {
4960 @@ -1082,7 +1175,8 @@ static void follow_dotdot(struct nameida
4961
4962                 if (nd->path.dentry == nd->root.dentry &&
4963                     nd->path.mnt == nd->root.mnt) {
4964 -                       break;
4965 +                       /* for sane '/' avoid follow_mount() */
4966 +                       return;
4967                 }
4968                 if (nd->path.dentry != nd->path.mnt->mnt_root) {
4969                         /* rare case of legitimate dget_parent()... */
4970 @@ -1832,7 +1926,7 @@ static int may_delete(struct inode *dir,
4971         if (IS_APPEND(dir))
4972                 return -EPERM;
4973         if (check_sticky(dir, victim->d_inode)||IS_APPEND(victim->d_inode)||
4974 -           IS_IMMUTABLE(victim->d_inode) || IS_SWAPFILE(victim->d_inode))
4975 +               IS_IXORUNLINK(victim->d_inode) || IS_SWAPFILE(victim->d_inode))
4976                 return -EPERM;
4977         if (isdir) {
4978                 if (!S_ISDIR(victim->d_inode->i_mode))
4979 @@ -1959,6 +2053,15 @@ static int may_open(struct path *path, i
4980                 break;
4981         }
4982
4983 +#ifdef CONFIG_VSERVER_COWBL
4984 +       if (IS_COW(inode) &&
4985 +               ((flag & O_ACCMODE) != O_RDONLY)) {
4986 +               if (IS_COW_LINK(inode))
4987 +                       return -EMLINK;
4988 +               inode->i_flags &= ~(S_IXUNLINK|S_IMMUTABLE);
4989 +               mark_inode_dirty(inode);
4990 +       }
4991 +#endif
4992         error = inode_permission(inode, acc_mode);
4993         if (error)
4994                 return error;
4995 @@ -2852,7 +2955,7 @@ int vfs_link(struct dentry *old_dentry,
4996         /*
4997          * A link to an append-only or immutable file cannot be created.
4998          */
4999 -       if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
5000 +       if (IS_APPEND(inode) || IS_IXORUNLINK(inode))
5001                 return -EPERM;
5002         if (!dir->i_op->link)
5003                 return -EPERM;
5004 @@ -3238,6 +3341,222 @@ int vfs_follow_link(struct nameidata *nd
5005         return __vfs_follow_link(nd, link);
5006  }
5007
5008 +
5009 +#ifdef CONFIG_VSERVER_COWBL
5010 +
5011 +#include <linux/file.h>
5012 +
5013 +static inline
5014 +long do_cow_splice(struct file *in, struct file *out, size_t len)
5015 +{
5016 +       loff_t ppos = 0;
5017 +
5018 +       return do_splice_direct(in, &ppos, out, len, 0);
5019 +}
5020 +
5021 +struct dentry *cow_break_link(const char *pathname)
5022 +{
5023 +       int ret, mode, pathlen, redo = 0;
5024 +       struct nameidata old_nd, dir_nd;
5025 +       struct path old_path, new_path;
5026 +       struct dentry *dir, *res = NULL;
5027 +       struct file *old_file;
5028 +       struct file *new_file;
5029 +       char *to, *path, pad='\251';
5030 +       loff_t size;
5031 +
5032 +       vxdprintk(VXD_CBIT(misc, 1),
5033 +               "cow_break_link(" VS_Q("%s") ")", pathname);
5034 +       path = kmalloc(PATH_MAX, GFP_KERNEL);
5035 +       ret = -ENOMEM;
5036 +       if (!path)
5037 +               goto out;
5038 +
5039 +       /* old_nd will have refs to dentry and mnt */
5040 +       ret = do_path_lookup(AT_FDCWD, pathname, LOOKUP_FOLLOW, &old_nd);
5041 +       vxdprintk(VXD_CBIT(misc, 2), "do_path_lookup(old): %d", ret);
5042 +       if (ret < 0)
5043 +               goto out_free_path;
5044 +
5045 +       old_path = old_nd.path;
5046 +       mode = old_path.dentry->d_inode->i_mode;
5047 +
5048 +       to = d_path(&old_path, path, PATH_MAX-2);
5049 +       pathlen = strlen(to);
5050 +       vxdprintk(VXD_CBIT(misc, 2),
5051 +               "old path " VS_Q("%s") " [" VS_Q("%.*s") ":%d]", to,
5052 +               old_path.dentry->d_name.len, old_path.dentry->d_name.name,
5053 +               old_path.dentry->d_name.len);
5054 +
5055 +       to[pathlen + 1] = 0;
5056 +retry:
5057 +       to[pathlen] = pad--;
5058 +       ret = -EMLINK;
5059 +       if (pad <= '\240')
5060 +               goto out_rel_old;
5061 +
5062 +       vxdprintk(VXD_CBIT(misc, 1), "temp copy " VS_Q("%s"), to);
5063 +       /* dir_nd will have refs to dentry and mnt */
5064 +       ret = do_path_lookup(AT_FDCWD, to,
5065 +               LOOKUP_PARENT | LOOKUP_OPEN | LOOKUP_CREATE, &dir_nd);
5066 +       vxdprintk(VXD_CBIT(misc, 2),
5067 +               "do_path_lookup(new): %d", ret);
5068 +       if (ret < 0)
5069 +               goto retry;
5070 +
5071 +       /* this puppy downs the inode mutex */
5072 +       new_path.dentry = lookup_create(&dir_nd, 0);
5073 +       if (!new_path.dentry || IS_ERR(new_path.dentry)) {
5074 +               vxdprintk(VXD_CBIT(misc, 2),
5075 +                       "lookup_create(new): %p", new_path.dentry);
5076 +               mutex_unlock(&dir_nd.path.dentry->d_inode->i_mutex);
5077 +               path_put(&dir_nd.path);
5078 +               goto retry;
5079 +       }
5080 +       vxdprintk(VXD_CBIT(misc, 2),
5081 +               "lookup_create(new): %p [" VS_Q("%.*s") ":%d]",
5082 +               new_path.dentry,
5083 +               new_path.dentry->d_name.len, new_path.dentry->d_name.name,
5084 +               new_path.dentry->d_name.len);
5085 +       dir = dir_nd.path.dentry;
5086 +
5087 +       ret = vfs_create(dir_nd.path.dentry->d_inode, new_path.dentry, mode, &dir_nd);
5088 +       vxdprintk(VXD_CBIT(misc, 2),
5089 +               "vfs_create(new): %d", ret);
5090 +       if (ret == -EEXIST) {
5091 +               mutex_unlock(&dir->d_inode->i_mutex);
5092 +               dput(new_path.dentry);
5093 +               path_put(&dir_nd.path);
5094 +               goto retry;
5095 +       }
5096 +       else if (ret < 0)
5097 +               goto out_unlock_new;
5098 +
5099 +       /* drop out early, ret passes ENOENT */
5100 +       ret = -ENOENT;
5101 +       if ((redo = d_unhashed(old_path.dentry)))
5102 +               goto out_unlock_new;
5103 +
5104 +       new_path.mnt = dir_nd.path.mnt;
5105 +       dget(old_path.dentry);
5106 +       mntget(old_path.mnt);
5107 +       /* this one cleans up the dentry/mnt in case of failure */
5108 +       old_file = dentry_open(old_path.dentry, old_path.mnt,
5109 +               O_RDONLY, current_cred());
5110 +       vxdprintk(VXD_CBIT(misc, 2),
5111 +               "dentry_open(old): %p", old_file);
5112 +       if (!old_file || IS_ERR(old_file)) {
5113 +               res = IS_ERR(old_file) ? (void *) old_file : res;
5114 +               goto out_unlock_new;
5115 +       }
5116 +
5117 +       dget(new_path.dentry);
5118 +       mntget(new_path.mnt);
5119 +       /* this one cleans up the dentry/mnt in case of failure */
5120 +       new_file = dentry_open(new_path.dentry, new_path.mnt,
5121 +               O_WRONLY, current_cred());
5122 +       vxdprintk(VXD_CBIT(misc, 2),
5123 +               "dentry_open(new): %p", new_file);
5124 +
5125 +       ret = IS_ERR(new_file) ? PTR_ERR(new_file) : -ENOENT;
5126 +       if (!new_file || IS_ERR(new_file))
5127 +               goto out_fput_old;
5128 +
5129 +       size = i_size_read(old_file->f_dentry->d_inode);
5130 +       ret = do_cow_splice(old_file, new_file, size);
5131 +       vxdprintk(VXD_CBIT(misc, 2), "do_splice_direct: %d", ret);
5132 +       if (ret < 0) {
5133 +               goto out_fput_both;
5134 +       } else if (ret < size) {
5135 +               ret = -ENOSPC;
5136 +               goto out_fput_both;
5137 +       } else {
5138 +               struct inode *old_inode = old_path.dentry->d_inode;
5139 +               struct inode *new_inode = new_path.dentry->d_inode;
5140 +               struct iattr attr = {
5141 +                       .ia_uid = old_inode->i_uid,
5142 +                       .ia_gid = old_inode->i_gid,
5143 +                       .ia_valid = ATTR_UID | ATTR_GID
5144 +                       };
5145 +
5146 +               setattr_copy(new_inode, &attr);
5147 +               mark_inode_dirty(new_inode);
5148 +       }
5149 +
5150 +       mutex_lock(&old_path.dentry->d_inode->i_sb->s_vfs_rename_mutex);
5151 +
5152 +       /* drop out late */
5153 +       ret = -ENOENT;
5154 +       if ((redo = d_unhashed(old_path.dentry)))
5155 +               goto out_unlock;
5156 +
5157 +       vxdprintk(VXD_CBIT(misc, 2),
5158 +               "vfs_rename: [" VS_Q("%*s") ":%d] -> [" VS_Q("%*s") ":%d]",
5159 +               new_path.dentry->d_name.len, new_path.dentry->d_name.name,
5160 +               new_path.dentry->d_name.len,
5161 +               old_path.dentry->d_name.len, old_path.dentry->d_name.name,
5162 +               old_path.dentry->d_name.len);
5163 +       ret = vfs_rename(dir_nd.path.dentry->d_inode, new_path.dentry,
5164 +               old_nd.path.dentry->d_parent->d_inode, old_path.dentry);
5165 +       vxdprintk(VXD_CBIT(misc, 2), "vfs_rename: %d", ret);
5166 +       res = new_path.dentry;
5167 +
5168 +out_unlock:
5169 +       mutex_unlock(&old_path.dentry->d_inode->i_sb->s_vfs_rename_mutex);
5170 +
5171 +out_fput_both:
5172 +       vxdprintk(VXD_CBIT(misc, 3),
5173 +               "fput(new_file=%p[#%ld])", new_file,
5174 +               atomic_long_read(&new_file->f_count));
5175 +       fput(new_file);
5176 +
5177 +out_fput_old:
5178 +       vxdprintk(VXD_CBIT(misc, 3),
5179 +               "fput(old_file=%p[#%ld])", old_file,
5180 +               atomic_long_read(&old_file->f_count));
5181 +       fput(old_file);
5182 +
5183 +out_unlock_new:
5184 +       mutex_unlock(&dir->d_inode->i_mutex);
5185 +       if (!ret)
5186 +               goto out_redo;
5187 +
5188 +       /* error path cleanup */
5189 +       vfs_unlink(dir->d_inode, new_path.dentry);
5190 +       dput(new_path.dentry);
5191 +
5192 +out_redo:
5193 +       if (!redo)
5194 +               goto out_rel_both;
5195 +       /* lookup dentry once again */
5196 +       path_put(&old_nd.path);
5197 +       ret = do_path_lookup(AT_FDCWD, pathname, LOOKUP_FOLLOW, &old_nd);
5198 +       if (ret)
5199 +               goto out_rel_both;
5200 +
5201 +       new_path.dentry = old_nd.path.dentry;
5202 +       vxdprintk(VXD_CBIT(misc, 2),
5203 +               "do_path_lookup(redo): %p [" VS_Q("%.*s") ":%d]",
5204 +               new_path.dentry,
5205 +               new_path.dentry->d_name.len, new_path.dentry->d_name.name,
5206 +               new_path.dentry->d_name.len);
5207 +       dget(new_path.dentry);
5208 +       res = new_path.dentry;
5209 +
5210 +out_rel_both:
5211 +       path_put(&dir_nd.path);
5212 +out_rel_old:
5213 +       path_put(&old_nd.path);
5214 +out_free_path:
5215 +       kfree(path);
5216 +out:
5217 +       if (ret)
5218 +               res = ERR_PTR(ret);
5219 +       return res;
5220 +}
5221 +
5222 +#endif
5223 +
5224  /* get the link contents into pagecache */
5225  static char *page_getlink(struct dentry * dentry, struct page **ppage)
5226  {
5227 diff -NurpP --minimal linux-3.0-rc5/fs/namespace.c linux-3.0-rc5-vs2.3.1-pre3/fs/namespace.c
5228 --- linux-3.0-rc5/fs/namespace.c        2011-07-01 11:14:19.000000000 +0200
5229 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/namespace.c   2011-06-13 18:26:48.000000000 +0200
5230 @@ -31,6 +31,11 @@
5231  #include <linux/idr.h>
5232  #include <linux/fs_struct.h>
5233  #include <linux/fsnotify.h>
5234 +#include <linux/vs_base.h>
5235 +#include <linux/vs_context.h>
5236 +#include <linux/vs_tag.h>
5237 +#include <linux/vserver/space.h>
5238 +#include <linux/vserver/global.h>
5239  #include <asm/uaccess.h>
5240  #include <asm/unistd.h>
5241  #include "pnode.h"
5242 @@ -724,6 +729,7 @@ static struct vfsmount *clone_mnt(struct
5243                 mnt->mnt_root = dget(root);
5244                 mnt->mnt_mountpoint = mnt->mnt_root;
5245                 mnt->mnt_parent = mnt;
5246 +               mnt->mnt_tag = old->mnt_tag;
5247
5248                 if (flag & CL_SLAVE) {
5249                         list_add(&mnt->mnt_slave, &old->mnt_slave_list);
5250 @@ -852,6 +858,31 @@ static inline void mangle(struct seq_fil
5251         seq_escape(m, s, " \t\n\\");
5252  }
5253
5254 +static int mnt_is_reachable(struct vfsmount *mnt)
5255 +{
5256 +       struct path root;
5257 +       struct dentry *point;
5258 +       int ret;
5259 +
5260 +       if (mnt == mnt->mnt_ns->root)
5261 +               return 1;
5262 +
5263 +       br_read_lock(vfsmount_lock);
5264 +       root = current->fs->root;
5265 +       point = root.dentry;
5266 +
5267 +       while ((mnt != mnt->mnt_parent) && (mnt != root.mnt)) {
5268 +               point = mnt->mnt_mountpoint;
5269 +               mnt = mnt->mnt_parent;
5270 +       }
5271 +
5272 +       ret = (mnt == root.mnt) && is_subdir(point, root.dentry);
5273 +
5274 +       br_read_unlock(vfsmount_lock);
5275 +
5276 +       return ret;
5277 +}
5278 +
5279  /*
5280   * Simple .show_options callback for filesystems which don't want to
5281   * implement more complex mount option showing.
5282 @@ -954,6 +985,8 @@ static int show_sb_opts(struct seq_file
5283                 { MS_SYNCHRONOUS, ",sync" },
5284                 { MS_DIRSYNC, ",dirsync" },
5285                 { MS_MANDLOCK, ",mand" },
5286 +               { MS_TAGGED, ",tag" },
5287 +               { MS_NOTAGCHECK, ",notagcheck" },
5288                 { 0, NULL }
5289         };
5290         const struct proc_fs_info *fs_infop;
5291 @@ -1000,16 +1033,26 @@ static int show_vfsmnt(struct seq_file *
5292         int err = 0;
5293         struct path mnt_path = { .dentry = mnt->mnt_root, .mnt = mnt };
5294
5295 -       if (mnt->mnt_sb->s_op->show_devname) {
5296 -               err = mnt->mnt_sb->s_op->show_devname(m, mnt);
5297 -               if (err)
5298 -                       goto out;
5299 +       if (vx_flags(VXF_HIDE_MOUNT, 0))
5300 +               return SEQ_SKIP;
5301 +       if (!mnt_is_reachable(mnt) && !vx_check(0, VS_WATCH_P))
5302 +               return SEQ_SKIP;
5303 +
5304 +       if (!vx_check(0, VS_ADMIN|VS_WATCH) &&
5305 +               mnt == current->fs->root.mnt) {
5306 +               seq_puts(m, "/dev/root / ");
5307         } else {
5308 -               mangle(m, mnt->mnt_devname ? mnt->mnt_devname : "none");
5309 +               if (mnt->mnt_sb->s_op->show_devname) {
5310 +                       err = mnt->mnt_sb->s_op->show_devname(m, mnt);
5311 +                       if (err)
5312 +                               goto out;
5313 +               } else {
5314 +                       mangle(m, mnt->mnt_devname ? mnt->mnt_devname : "none");
5315 +               }
5316 +               seq_putc(m, ' ');
5317 +               seq_path(m, &mnt_path, " \t\n\\");
5318 +               seq_putc(m, ' ');
5319         }
5320 -       seq_putc(m, ' ');
5321 -       seq_path(m, &mnt_path, " \t\n\\");
5322 -       seq_putc(m, ' ');
5323         show_type(m, mnt->mnt_sb);
5324         seq_puts(m, __mnt_is_readonly(mnt) ? " ro" : " rw");
5325         err = show_sb_opts(m, mnt->mnt_sb);
5326 @@ -1039,6 +1082,11 @@ static int show_mountinfo(struct seq_fil
5327         struct path root = p->root;
5328         int err = 0;
5329
5330 +       if (vx_flags(VXF_HIDE_MOUNT, 0))
5331 +               return SEQ_SKIP;
5332 +       if (!mnt_is_reachable(mnt) && !vx_check(0, VS_WATCH_P))
5333 +               return SEQ_SKIP;
5334 +
5335         seq_printf(m, "%i %i %u:%u ", mnt->mnt_id, mnt->mnt_parent->mnt_id,
5336                    MAJOR(sb->s_dev), MINOR(sb->s_dev));
5337         if (sb->s_op->show_path)
5338 @@ -1107,21 +1155,31 @@ static int show_vfsstat(struct seq_file
5339         struct path mnt_path = { .dentry = mnt->mnt_root, .mnt = mnt };
5340         int err = 0;
5341
5342 -       /* device */
5343 -       if (mnt->mnt_sb->s_op->show_devname) {
5344 -               err = mnt->mnt_sb->s_op->show_devname(m, mnt);
5345 +       if (vx_flags(VXF_HIDE_MOUNT, 0))
5346 +               return SEQ_SKIP;
5347 +       if (!mnt_is_reachable(mnt) && !vx_check(0, VS_WATCH_P))
5348 +               return SEQ_SKIP;
5349 +
5350 +       if (!vx_check(0, VS_ADMIN|VS_WATCH) &&
5351 +               mnt == current->fs->root.mnt) {
5352 +               seq_puts(m, "device /dev/root mounted on / ");
5353         } else {
5354 -               if (mnt->mnt_devname) {
5355 -                       seq_puts(m, "device ");
5356 -                       mangle(m, mnt->mnt_devname);
5357 -               } else
5358 -                       seq_puts(m, "no device");
5359 -       }
5360 +               /* device */
5361 +               if (mnt->mnt_sb->s_op->show_devname) {
5362 +                       err = mnt->mnt_sb->s_op->show_devname(m, mnt);
5363 +               } else {
5364 +                       if (mnt->mnt_devname) {
5365 +                               seq_puts(m, "device ");
5366 +                               mangle(m, mnt->mnt_devname);
5367 +                       } else
5368 +                               seq_puts(m, "no device");
5369 +               }
5370
5371 -       /* mount point */
5372 -       seq_puts(m, " mounted on ");
5373 -       seq_path(m, &mnt_path, " \t\n\\");
5374 -       seq_putc(m, ' ');
5375 +               /* mount point */
5376 +               seq_puts(m, " mounted on ");
5377 +               seq_path(m, &mnt_path, " \t\n\\");
5378 +               seq_putc(m, ' ');
5379 +       }
5380
5381         /* file system type */
5382         seq_puts(m, "with fstype ");
5383 @@ -1380,7 +1438,7 @@ SYSCALL_DEFINE2(umount, char __user *, n
5384                 goto dput_and_out;
5385
5386         retval = -EPERM;
5387 -       if (!capable(CAP_SYS_ADMIN))
5388 +       if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT))
5389                 goto dput_and_out;
5390
5391         retval = do_umount(path.mnt, flags);
5392 @@ -1406,7 +1464,7 @@ SYSCALL_DEFINE1(oldumount, char __user *
5393
5394  static int mount_is_safe(struct path *path)
5395  {
5396 -       if (capable(CAP_SYS_ADMIN))
5397 +       if (vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT))
5398                 return 0;
5399         return -EPERM;
5400  #ifdef notyet
5401 @@ -1716,7 +1774,7 @@ static int do_change_type(struct path *p
5402         int type;
5403         int err = 0;
5404
5405 -       if (!capable(CAP_SYS_ADMIN))
5406 +       if (!vx_capable(CAP_SYS_ADMIN, VXC_NAMESPACE))
5407                 return -EPERM;
5408
5409         if (path->dentry != path->mnt->mnt_root)
5410 @@ -1732,6 +1790,7 @@ static int do_change_type(struct path *p
5411                 if (err)
5412                         goto out_unlock;
5413         }
5414 +       // mnt->mnt_flags = mnt_flags;
5415
5416         br_write_lock(vfsmount_lock);
5417         for (m = mnt; m; m = (recurse ? next_mnt(m, mnt) : NULL))
5418 @@ -1747,12 +1806,14 @@ static int do_change_type(struct path *p
5419   * do loopback mount.
5420   */
5421  static int do_loopback(struct path *path, char *old_name,
5422 -                               int recurse)
5423 +       tag_t tag, unsigned long flags, int mnt_flags)
5424  {
5425         LIST_HEAD(umount_list);
5426         struct path old_path;
5427         struct vfsmount *mnt = NULL;
5428         int err = mount_is_safe(path);
5429 +       int recurse = flags & MS_REC;
5430 +
5431         if (err)
5432                 return err;
5433         if (!old_name || !*old_name)
5434 @@ -1818,12 +1879,12 @@ static int change_mount_flags(struct vfs
5435   * on it - tough luck.
5436   */
5437  static int do_remount(struct path *path, int flags, int mnt_flags,
5438 -                     void *data)
5439 +       void *data, xid_t xid)
5440  {
5441         int err;
5442         struct super_block *sb = path->mnt->mnt_sb;
5443
5444 -       if (!capable(CAP_SYS_ADMIN))
5445 +       if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_REMOUNT))
5446                 return -EPERM;
5447
5448         if (!check_mnt(path->mnt))
5449 @@ -1871,7 +1932,7 @@ static int do_move_mount(struct path *pa
5450         struct path old_path, parent_path;
5451         struct vfsmount *p;
5452         int err = 0;
5453 -       if (!capable(CAP_SYS_ADMIN))
5454 +       if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT))
5455                 return -EPERM;
5456         if (!old_name || !*old_name)
5457                 return -EINVAL;
5458 @@ -2022,7 +2083,7 @@ static int do_new_mount(struct path *pat
5459                 return -EINVAL;
5460
5461         /* we need capabilities... */
5462 -       if (!capable(CAP_SYS_ADMIN))
5463 +       if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT))
5464                 return -EPERM;
5465
5466         mnt = do_kern_mount(type, flags, name, data);
5467 @@ -2291,6 +2352,7 @@ long do_mount(char *dev_name, char *dir_
5468         struct path path;
5469         int retval = 0;
5470         int mnt_flags = 0;
5471 +       tag_t tag = 0;
5472
5473         /* Discard magic */
5474         if ((flags & MS_MGC_MSK) == MS_MGC_VAL)
5475 @@ -2318,6 +2380,12 @@ long do_mount(char *dev_name, char *dir_
5476         if (!(flags & MS_NOATIME))
5477                 mnt_flags |= MNT_RELATIME;
5478
5479 +       if (dx_parse_tag(data_page, &tag, 1, &mnt_flags, &flags)) {
5480 +               /* FIXME: bind and re-mounts get the tag flag? */
5481 +               if (flags & (MS_BIND|MS_REMOUNT))
5482 +                       flags |= MS_TAGID;
5483 +       }
5484 +
5485         /* Separate the per-mountpoint flags */
5486         if (flags & MS_NOSUID)
5487                 mnt_flags |= MNT_NOSUID;
5488 @@ -2334,15 +2402,17 @@ long do_mount(char *dev_name, char *dir_
5489         if (flags & MS_RDONLY)
5490                 mnt_flags |= MNT_READONLY;
5491
5492 +       if (!capable(CAP_SYS_ADMIN))
5493 +               mnt_flags |= MNT_NODEV;
5494         flags &= ~(MS_NOSUID | MS_NOEXEC | MS_NODEV | MS_ACTIVE | MS_BORN |
5495                    MS_NOATIME | MS_NODIRATIME | MS_RELATIME| MS_KERNMOUNT |
5496                    MS_STRICTATIME);
5497
5498         if (flags & MS_REMOUNT)
5499                 retval = do_remount(&path, flags & ~MS_REMOUNT, mnt_flags,
5500 -                                   data_page);
5501 +                                   data_page, tag);
5502         else if (flags & MS_BIND)
5503 -               retval = do_loopback(&path, dev_name, flags & MS_REC);
5504 +               retval = do_loopback(&path, dev_name, tag, flags, mnt_flags);
5505         else if (flags & (MS_SHARED | MS_PRIVATE | MS_SLAVE | MS_UNBINDABLE))
5506                 retval = do_change_type(&path, flags);
5507         else if (flags & MS_MOVE)
5508 @@ -2442,6 +2512,7 @@ static struct mnt_namespace *dup_mnt_ns(
5509                 q = next_mnt(q, new_ns->root);
5510         }
5511         up_write(&namespace_sem);
5512 +       atomic_inc(&vs_global_mnt_ns);
5513
5514         if (rootmnt)
5515                 mntput(rootmnt);
5516 @@ -2581,9 +2652,10 @@ SYSCALL_DEFINE2(pivot_root, const char _
5517                 goto out3;
5518
5519         error = -EINVAL;
5520 -       if (IS_MNT_SHARED(old.mnt) ||
5521 -               IS_MNT_SHARED(new.mnt->mnt_parent) ||
5522 -               IS_MNT_SHARED(root.mnt->mnt_parent))
5523 +       if ((IS_MNT_SHARED(old.mnt) ||
5524 +               IS_MNT_SHARED(new.mnt->mnt_parent) ||
5525 +               IS_MNT_SHARED(root.mnt->mnt_parent)) &&
5526 +               !vx_flags(VXF_STATE_SETUP, 0))
5527                 goto out4;
5528         if (!check_mnt(root.mnt) || !check_mnt(new.mnt))
5529                 goto out4;
5530 @@ -2715,6 +2787,7 @@ void put_mnt_ns(struct mnt_namespace *ns
5531         br_write_unlock(vfsmount_lock);
5532         up_write(&namespace_sem);
5533         release_mounts(&umount_list);
5534 +       atomic_dec(&vs_global_mnt_ns);
5535         kfree(ns);
5536  }
5537  EXPORT_SYMBOL(put_mnt_ns);
5538 diff -NurpP --minimal linux-3.0-rc5/fs/nfs/client.c linux-3.0-rc5-vs2.3.1-pre3/fs/nfs/client.c
5539 --- linux-3.0-rc5/fs/nfs/client.c       2011-07-01 11:14:19.000000000 +0200
5540 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/nfs/client.c  2011-06-10 22:11:24.000000000 +0200
5541 @@ -780,6 +780,9 @@ static int nfs_init_server_rpcclient(str
5542         if (server->flags & NFS_MOUNT_SOFT)
5543                 server->client->cl_softrtry = 1;
5544
5545 +       server->client->cl_tag = 0;
5546 +       if (server->flags & NFS_MOUNT_TAGGED)
5547 +               server->client->cl_tag = 1;
5548         return 0;
5549  }
5550
5551 @@ -951,6 +954,10 @@ static void nfs_server_set_fsinfo(struct
5552                 server->acdirmin = server->acdirmax = 0;
5553         }
5554
5555 +       /* FIXME: needs fsinfo
5556 +       if (server->flags & NFS_MOUNT_TAGGED)
5557 +               sb->s_flags |= MS_TAGGED;       */
5558 +
5559         server->maxfilesize = fsinfo->maxfilesize;
5560
5561         server->time_delta = fsinfo->time_delta;
5562 diff -NurpP --minimal linux-3.0-rc5/fs/nfs/dir.c linux-3.0-rc5-vs2.3.1-pre3/fs/nfs/dir.c
5563 --- linux-3.0-rc5/fs/nfs/dir.c  2011-07-01 11:14:19.000000000 +0200
5564 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/nfs/dir.c     2011-06-10 22:11:24.000000000 +0200
5565 @@ -35,6 +35,7 @@
5566  #include <linux/sched.h>
5567  #include <linux/kmemleak.h>
5568  #include <linux/xattr.h>
5569 +#include <linux/vs_tag.h>
5570
5571  #include "delegation.h"
5572  #include "iostat.h"
5573 @@ -1298,6 +1299,7 @@ static struct dentry *nfs_lookup(struct
5574         if (IS_ERR(res))
5575                 goto out_unblock_sillyrename;
5576
5577 +       dx_propagate_tag(nd, inode);
5578  no_entry:
5579         res = d_materialise_unique(dentry, inode);
5580         if (res != NULL) {
5581 diff -NurpP --minimal linux-3.0-rc5/fs/nfs/inode.c linux-3.0-rc5-vs2.3.1-pre3/fs/nfs/inode.c
5582 --- linux-3.0-rc5/fs/nfs/inode.c        2011-07-01 11:14:19.000000000 +0200
5583 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/nfs/inode.c   2011-07-01 11:35:34.000000000 +0200
5584 @@ -38,6 +38,7 @@
5585  #include <linux/nfs_xdr.h>
5586  #include <linux/slab.h>
5587  #include <linux/compat.h>
5588 +#include <linux/vs_tag.h>
5589
5590  #include <asm/system.h>
5591  #include <asm/uaccess.h>
5592 @@ -273,6 +274,8 @@ nfs_fhget(struct super_block *sb, struct
5593         if (inode->i_state & I_NEW) {
5594                 struct nfs_inode *nfsi = NFS_I(inode);
5595                 unsigned long now = jiffies;
5596 +               uid_t uid;
5597 +               gid_t gid;
5598
5599                 /* We set i_ino for the few things that still rely on it,
5600                  * such as stat(2) */
5601 @@ -321,8 +324,8 @@ nfs_fhget(struct super_block *sb, struct
5602                 nfsi->change_attr = 0;
5603                 inode->i_size = 0;
5604                 inode->i_nlink = 0;
5605 -               inode->i_uid = -2;
5606 -               inode->i_gid = -2;
5607 +               uid = -2;
5608 +               gid = -2;
5609                 inode->i_blocks = 0;
5610                 memset(nfsi->cookieverf, 0, sizeof(nfsi->cookieverf));
5611
5612 @@ -359,13 +362,13 @@ nfs_fhget(struct super_block *sb, struct
5613                 else if (nfs_server_capable(inode, NFS_CAP_NLINK))
5614                         nfsi->cache_validity |= NFS_INO_INVALID_ATTR;
5615                 if (fattr->valid & NFS_ATTR_FATTR_OWNER)
5616 -                       inode->i_uid = fattr->uid;
5617 +                       uid = fattr->uid;
5618                 else if (nfs_server_capable(inode, NFS_CAP_OWNER))
5619                         nfsi->cache_validity |= NFS_INO_INVALID_ATTR
5620                                 | NFS_INO_INVALID_ACCESS
5621                                 | NFS_INO_INVALID_ACL;
5622                 if (fattr->valid & NFS_ATTR_FATTR_GROUP)
5623 -                       inode->i_gid = fattr->gid;
5624 +                       gid = fattr->gid;
5625                 else if (nfs_server_capable(inode, NFS_CAP_OWNER_GROUP))
5626                         nfsi->cache_validity |= NFS_INO_INVALID_ATTR
5627                                 | NFS_INO_INVALID_ACCESS
5628 @@ -378,6 +381,11 @@ nfs_fhget(struct super_block *sb, struct
5629                          */
5630                         inode->i_blocks = nfs_calc_block_size(fattr->du.nfs3.used);
5631                 }
5632 +               inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
5633 +               inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
5634 +               inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid, 0);
5635 +                               /* maybe fattr->xid someday */
5636 +
5637                 nfsi->attrtimeo = NFS_MINATTRTIMEO(inode);
5638                 nfsi->attrtimeo_timestamp = now;
5639                 nfsi->access_cache = RB_ROOT;
5640 @@ -494,6 +502,8 @@ void nfs_setattr_update_inode(struct ino
5641                         inode->i_uid = attr->ia_uid;
5642                 if ((attr->ia_valid & ATTR_GID) != 0)
5643                         inode->i_gid = attr->ia_gid;
5644 +               if ((attr->ia_valid & ATTR_TAG) && IS_TAGGED(inode))
5645 +                       inode->i_tag = attr->ia_tag;
5646                 NFS_I(inode)->cache_validity |= NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
5647                 spin_unlock(&inode->i_lock);
5648         }
5649 @@ -941,6 +951,9 @@ static int nfs_check_inode_attributes(st
5650         struct nfs_inode *nfsi = NFS_I(inode);
5651         loff_t cur_size, new_isize;
5652         unsigned long invalid = 0;
5653 +       uid_t uid;
5654 +       gid_t gid;
5655 +       tag_t tag;
5656
5657
5658         /* Has the inode gone and changed behind our back? */
5659 @@ -964,13 +977,18 @@ static int nfs_check_inode_attributes(st
5660                         invalid |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE;
5661         }
5662
5663 +       uid = INOTAG_UID(DX_TAG(inode), fattr->uid, fattr->gid);
5664 +       gid = INOTAG_GID(DX_TAG(inode), fattr->uid, fattr->gid);
5665 +       tag = INOTAG_TAG(DX_TAG(inode), fattr->uid, fattr->gid, 0);
5666 +
5667         /* Have any file permissions changed? */
5668         if ((fattr->valid & NFS_ATTR_FATTR_MODE) && (inode->i_mode & S_IALLUGO) != (fattr->mode & S_IALLUGO))
5669                 invalid |= NFS_INO_INVALID_ATTR | NFS_INO_INVALID_ACCESS | NFS_INO_INVALID_ACL;
5670 -       if ((fattr->valid & NFS_ATTR_FATTR_OWNER) && inode->i_uid != fattr->uid)
5671 +       if ((fattr->valid & NFS_ATTR_FATTR_OWNER) && uid != fattr->uid)
5672                 invalid |= NFS_INO_INVALID_ATTR | NFS_INO_INVALID_ACCESS | NFS_INO_INVALID_ACL;
5673 -       if ((fattr->valid & NFS_ATTR_FATTR_GROUP) && inode->i_gid != fattr->gid)
5674 +       if ((fattr->valid & NFS_ATTR_FATTR_GROUP) && gid != fattr->gid)
5675                 invalid |= NFS_INO_INVALID_ATTR | NFS_INO_INVALID_ACCESS | NFS_INO_INVALID_ACL;
5676 +               /* maybe check for tag too? */
5677
5678         /* Has the link count changed? */
5679         if ((fattr->valid & NFS_ATTR_FATTR_NLINK) && inode->i_nlink != fattr->nlink)
5680 @@ -1205,6 +1223,9 @@ static int nfs_update_inode(struct inode
5681         unsigned long invalid = 0;
5682         unsigned long now = jiffies;
5683         unsigned long save_cache_validity;
5684 +       uid_t uid;
5685 +       gid_t gid;
5686 +       tag_t tag;
5687
5688         dfprintk(VFS, "NFS: %s(%s/%ld ct=%d info=0x%x)\n",
5689                         __func__, inode->i_sb->s_id, inode->i_ino,
5690 @@ -1312,6 +1333,9 @@ static int nfs_update_inode(struct inode
5691                                 | NFS_INO_REVAL_PAGECACHE
5692                                 | NFS_INO_REVAL_FORCED);
5693
5694 +       uid = INOTAG_UID(DX_TAG(inode), fattr->uid, fattr->gid);
5695 +       gid = INOTAG_GID(DX_TAG(inode), fattr->uid, fattr->gid);
5696 +       tag = INOTAG_TAG(DX_TAG(inode), fattr->uid, fattr->gid, 0);
5697
5698         if (fattr->valid & NFS_ATTR_FATTR_ATIME)
5699                 memcpy(&inode->i_atime, &fattr->atime, sizeof(inode->i_atime));
5700 @@ -1333,9 +1357,9 @@ static int nfs_update_inode(struct inode
5701                                 | NFS_INO_REVAL_FORCED);
5702
5703         if (fattr->valid & NFS_ATTR_FATTR_OWNER) {
5704 -               if (inode->i_uid != fattr->uid) {
5705 +               if (uid != fattr->uid) {
5706                         invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
5707 -                       inode->i_uid = fattr->uid;
5708 +                       uid = fattr->uid;
5709                 }
5710         } else if (server->caps & NFS_CAP_OWNER)
5711                 invalid |= save_cache_validity & (NFS_INO_INVALID_ATTR
5712 @@ -1344,9 +1368,9 @@ static int nfs_update_inode(struct inode
5713                                 | NFS_INO_REVAL_FORCED);
5714
5715         if (fattr->valid & NFS_ATTR_FATTR_GROUP) {
5716 -               if (inode->i_gid != fattr->gid) {
5717 +               if (gid != fattr->gid) {
5718                         invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
5719 -                       inode->i_gid = fattr->gid;
5720 +                       gid = fattr->gid;
5721                 }
5722         } else if (server->caps & NFS_CAP_OWNER_GROUP)
5723                 invalid |= save_cache_validity & (NFS_INO_INVALID_ATTR
5724 @@ -1354,6 +1378,10 @@ static int nfs_update_inode(struct inode
5725                                 | NFS_INO_INVALID_ACL
5726                                 | NFS_INO_REVAL_FORCED);
5727
5728 +       inode->i_uid = uid;
5729 +       inode->i_gid = gid;
5730 +       inode->i_tag = tag;
5731 +
5732         if (fattr->valid & NFS_ATTR_FATTR_NLINK) {
5733                 if (inode->i_nlink != fattr->nlink) {
5734                         invalid |= NFS_INO_INVALID_ATTR;
5735 diff -NurpP --minimal linux-3.0-rc5/fs/nfs/nfs3xdr.c linux-3.0-rc5-vs2.3.1-pre3/fs/nfs/nfs3xdr.c
5736 --- linux-3.0-rc5/fs/nfs/nfs3xdr.c      2011-03-15 18:07:32.000000000 +0100
5737 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/nfs/nfs3xdr.c 2011-06-10 22:11:24.000000000 +0200
5738 @@ -20,6 +20,7 @@
5739  #include <linux/nfs3.h>
5740  #include <linux/nfs_fs.h>
5741  #include <linux/nfsacl.h>
5742 +#include <linux/vs_tag.h>
5743  #include "internal.h"
5744
5745  #define NFSDBG_FACILITY                NFSDBG_XDR
5746 @@ -562,7 +563,8 @@ static __be32 *xdr_decode_nfstime3(__be3
5747   *             set_mtime       mtime;
5748   *     };
5749   */
5750 -static void encode_sattr3(struct xdr_stream *xdr, const struct iattr *attr)
5751 +static void encode_sattr3(struct xdr_stream *xdr,
5752 +       const struct iattr *attr, int tag)
5753  {
5754         u32 nbytes;
5755         __be32 *p;
5756 @@ -594,15 +596,19 @@ static void encode_sattr3(struct xdr_str
5757         } else
5758                 *p++ = xdr_zero;
5759
5760 -       if (attr->ia_valid & ATTR_UID) {
5761 +       if (attr->ia_valid & ATTR_UID ||
5762 +               (tag && (attr->ia_valid & ATTR_TAG))) {
5763                 *p++ = xdr_one;
5764 -               *p++ = cpu_to_be32(attr->ia_uid);
5765 +               *p++ = cpu_to_be32(TAGINO_UID(tag,
5766 +                       attr->ia_uid, attr->ia_tag));
5767         } else
5768                 *p++ = xdr_zero;
5769
5770 -       if (attr->ia_valid & ATTR_GID) {
5771 +       if (attr->ia_valid & ATTR_GID ||
5772 +               (tag && (attr->ia_valid & ATTR_TAG))) {
5773                 *p++ = xdr_one;
5774 -               *p++ = cpu_to_be32(attr->ia_gid);
5775 +               *p++ = cpu_to_be32(TAGINO_GID(tag,
5776 +                       attr->ia_gid, attr->ia_tag));
5777         } else
5778                 *p++ = xdr_zero;
5779
5780 @@ -878,7 +884,7 @@ static void nfs3_xdr_enc_setattr3args(st
5781                                       const struct nfs3_sattrargs *args)
5782  {
5783         encode_nfs_fh3(xdr, args->fh);
5784 -       encode_sattr3(xdr, args->sattr);
5785 +       encode_sattr3(xdr, args->sattr, req->rq_task->tk_client->cl_tag);
5786         encode_sattrguard3(xdr, args);
5787  }
5788
5789 @@ -1028,13 +1034,13 @@ static void nfs3_xdr_enc_write3args(stru
5790   *     };
5791   */
5792  static void encode_createhow3(struct xdr_stream *xdr,
5793 -                             const struct nfs3_createargs *args)
5794 +       const struct nfs3_createargs *args, int tag)
5795  {
5796         encode_uint32(xdr, args->createmode);
5797         switch (args->createmode) {
5798         case NFS3_CREATE_UNCHECKED:
5799         case NFS3_CREATE_GUARDED:
5800 -               encode_sattr3(xdr, args->sattr);
5801 +               encode_sattr3(xdr, args->sattr, tag);
5802                 break;
5803         case NFS3_CREATE_EXCLUSIVE:
5804                 encode_createverf3(xdr, args->verifier);
5805 @@ -1049,7 +1055,7 @@ static void nfs3_xdr_enc_create3args(str
5806                                      const struct nfs3_createargs *args)
5807  {
5808         encode_diropargs3(xdr, args->fh, args->name, args->len);
5809 -       encode_createhow3(xdr, args);
5810 +       encode_createhow3(xdr, args, req->rq_task->tk_client->cl_tag);
5811  }
5812
5813  /*
5814 @@ -1065,7 +1071,7 @@ static void nfs3_xdr_enc_mkdir3args(stru
5815                                     const struct nfs3_mkdirargs *args)
5816  {
5817         encode_diropargs3(xdr, args->fh, args->name, args->len);
5818 -       encode_sattr3(xdr, args->sattr);
5819 +       encode_sattr3(xdr, args->sattr, req->rq_task->tk_client->cl_tag);
5820  }
5821
5822  /*
5823 @@ -1082,9 +1088,9 @@ static void nfs3_xdr_enc_mkdir3args(stru
5824   *     };
5825   */
5826  static void encode_symlinkdata3(struct xdr_stream *xdr,
5827 -                               const struct nfs3_symlinkargs *args)
5828 +       const struct nfs3_symlinkargs *args, int tag)
5829  {
5830 -       encode_sattr3(xdr, args->sattr);
5831 +       encode_sattr3(xdr, args->sattr, tag);
5832         encode_nfspath3(xdr, args->pages, args->pathlen);
5833  }
5834
5835 @@ -1093,7 +1099,7 @@ static void nfs3_xdr_enc_symlink3args(st
5836                                       const struct nfs3_symlinkargs *args)
5837  {
5838         encode_diropargs3(xdr, args->fromfh, args->fromname, args->fromlen);
5839 -       encode_symlinkdata3(xdr, args);
5840 +       encode_symlinkdata3(xdr, args, req->rq_task->tk_client->cl_tag);
5841  }
5842
5843  /*
5844 @@ -1121,24 +1127,24 @@ static void nfs3_xdr_enc_symlink3args(st
5845   *     };
5846   */
5847  static void encode_devicedata3(struct xdr_stream *xdr,
5848 -                              const struct nfs3_mknodargs *args)
5849 +       const struct nfs3_mknodargs *args, int tag)
5850  {
5851 -       encode_sattr3(xdr, args->sattr);
5852 +       encode_sattr3(xdr, args->sattr, tag);
5853         encode_specdata3(xdr, args->rdev);
5854  }
5855
5856  static void encode_mknoddata3(struct xdr_stream *xdr,
5857 -                             const struct nfs3_mknodargs *args)
5858 +       const struct nfs3_mknodargs *args, int tag)
5859  {
5860         encode_ftype3(xdr, args->type);
5861         switch (args->type) {
5862         case NF3CHR:
5863         case NF3BLK:
5864 -               encode_devicedata3(xdr, args);
5865 +               encode_devicedata3(xdr, args, tag);
5866                 break;
5867         case NF3SOCK:
5868         case NF3FIFO:
5869 -               encode_sattr3(xdr, args->sattr);
5870 +               encode_sattr3(xdr, args->sattr, tag);
5871                 break;
5872         case NF3REG:
5873         case NF3DIR:
5874 @@ -1153,7 +1159,7 @@ static void nfs3_xdr_enc_mknod3args(stru
5875                                     const struct nfs3_mknodargs *args)
5876  {
5877         encode_diropargs3(xdr, args->fh, args->name, args->len);
5878 -       encode_mknoddata3(xdr, args);
5879 +       encode_mknoddata3(xdr, args, req->rq_task->tk_client->cl_tag);
5880  }
5881
5882  /*
5883 diff -NurpP --minimal linux-3.0-rc5/fs/nfs/super.c linux-3.0-rc5-vs2.3.1-pre3/fs/nfs/super.c
5884 --- linux-3.0-rc5/fs/nfs/super.c        2011-07-01 11:14:19.000000000 +0200
5885 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/nfs/super.c   2011-06-10 22:11:24.000000000 +0200
5886 @@ -53,6 +53,7 @@
5887  #include <linux/nfs_xdr.h>
5888  #include <linux/magic.h>
5889  #include <linux/parser.h>
5890 +#include <linux/vs_tag.h>
5891
5892  #include <asm/system.h>
5893  #include <asm/uaccess.h>
5894 @@ -87,6 +88,7 @@ enum {
5895         Opt_sharecache, Opt_nosharecache,
5896         Opt_resvport, Opt_noresvport,
5897         Opt_fscache, Opt_nofscache,
5898 +       Opt_tag, Opt_notag,
5899
5900         /* Mount options that take integer arguments */
5901         Opt_port,
5902 @@ -100,6 +102,7 @@ enum {
5903         Opt_mountvers,
5904         Opt_nfsvers,
5905         Opt_minorversion,
5906 +       Opt_tagid,
5907
5908         /* Mount options that take string arguments */
5909         Opt_sec, Opt_proto, Opt_mountproto, Opt_mounthost,
5910 @@ -180,6 +183,10 @@ static const match_table_t nfs_mount_opt
5911         { Opt_fscache_uniq, "fsc=%s" },
5912         { Opt_local_lock, "local_lock=%s" },
5913
5914 +       { Opt_tag, "tag" },
5915 +       { Opt_notag, "notag" },
5916 +       { Opt_tagid, "tagid=%u" },
5917 +
5918         { Opt_err, NULL }
5919  };
5920
5921 @@ -650,6 +657,7 @@ static void nfs_show_mount_options(struc
5922                 { NFS_MOUNT_NORDIRPLUS, ",nordirplus", "" },
5923                 { NFS_MOUNT_UNSHARED, ",nosharecache", "" },
5924                 { NFS_MOUNT_NORESVPORT, ",noresvport", "" },
5925 +               { NFS_MOUNT_TAGGED, ",tag", "" },
5926                 { 0, NULL, NULL }
5927         };
5928         const struct proc_nfs_info *nfs_infop;
5929 @@ -1198,6 +1206,14 @@ static int nfs_parse_mount_options(char
5930                         kfree(mnt->fscache_uniq);
5931                         mnt->fscache_uniq = NULL;
5932                         break;
5933 +#ifndef CONFIG_TAGGING_NONE
5934 +               case Opt_tag:
5935 +                       mnt->flags |= NFS_MOUNT_TAGGED;
5936 +                       break;
5937 +               case Opt_notag:
5938 +                       mnt->flags &= ~NFS_MOUNT_TAGGED;
5939 +                       break;
5940 +#endif
5941
5942                 /*
5943                  * options that take numeric values
5944 @@ -1304,6 +1320,12 @@ static int nfs_parse_mount_options(char
5945                                 goto out_invalid_value;
5946                         mnt->minorversion = option;
5947                         break;
5948 +#ifdef CONFIG_PROPAGATE
5949 +               case Opt_tagid:
5950 +                       /* use args[0] */
5951 +                       nfs_data.flags |= NFS_MOUNT_TAGGED;
5952 +                       break;
5953 +#endif
5954
5955                 /*
5956                  * options that take text values
5957 diff -NurpP --minimal linux-3.0-rc5/fs/nfsd/auth.c linux-3.0-rc5-vs2.3.1-pre3/fs/nfsd/auth.c
5958 --- linux-3.0-rc5/fs/nfsd/auth.c        2010-02-25 11:52:05.000000000 +0100
5959 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/nfsd/auth.c   2011-06-10 22:11:24.000000000 +0200
5960 @@ -1,6 +1,7 @@
5961  /* Copyright (C) 1995, 1996 Olaf Kirch <okir@monad.swb.de> */
5962
5963  #include <linux/sched.h>
5964 +#include <linux/vs_tag.h>
5965  #include "nfsd.h"
5966  #include "auth.h"
5967
5968 @@ -36,6 +37,9 @@ int nfsd_setuser(struct svc_rqst *rqstp,
5969
5970         new->fsuid = rqstp->rq_cred.cr_uid;
5971         new->fsgid = rqstp->rq_cred.cr_gid;
5972 +       /* FIXME: this desperately needs a tag :)
5973 +       new->xid = (xid_t)INOTAG_TAG(DX_TAG_NFSD, cred.cr_uid, cred.cr_gid, 0);
5974 +                       */
5975
5976         rqgi = rqstp->rq_cred.cr_group_info;
5977
5978 diff -NurpP --minimal linux-3.0-rc5/fs/nfsd/nfs3xdr.c linux-3.0-rc5-vs2.3.1-pre3/fs/nfsd/nfs3xdr.c
5979 --- linux-3.0-rc5/fs/nfsd/nfs3xdr.c     2011-07-01 11:14:19.000000000 +0200
5980 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/nfsd/nfs3xdr.c        2011-06-10 22:11:24.000000000 +0200
5981 @@ -7,6 +7,7 @@
5982   */
5983
5984  #include <linux/namei.h>
5985 +#include <linux/vs_tag.h>
5986  #include "xdr3.h"
5987  #include "auth.h"
5988
5989 @@ -95,6 +96,8 @@ static __be32 *
5990  decode_sattr3(__be32 *p, struct iattr *iap)
5991  {
5992         u32     tmp;
5993 +       uid_t   uid = 0;
5994 +       gid_t   gid = 0;
5995
5996         iap->ia_valid = 0;
5997
5998 @@ -104,12 +107,15 @@ decode_sattr3(__be32 *p, struct iattr *i
5999         }
6000         if (*p++) {
6001                 iap->ia_valid |= ATTR_UID;
6002 -               iap->ia_uid = ntohl(*p++);
6003 +               uid = ntohl(*p++);
6004         }
6005         if (*p++) {
6006                 iap->ia_valid |= ATTR_GID;
6007 -               iap->ia_gid = ntohl(*p++);
6008 +               gid = ntohl(*p++);
6009         }
6010 +       iap->ia_uid = INOTAG_UID(DX_TAG_NFSD, uid, gid);
6011 +       iap->ia_gid = INOTAG_GID(DX_TAG_NFSD, uid, gid);
6012 +       iap->ia_tag = INOTAG_TAG(DX_TAG_NFSD, uid, gid, 0);
6013         if (*p++) {
6014                 u64     newsize;
6015
6016 @@ -165,8 +171,12 @@ encode_fattr3(struct svc_rqst *rqstp, __
6017         *p++ = htonl(nfs3_ftypes[(stat->mode & S_IFMT) >> 12]);
6018         *p++ = htonl((u32) stat->mode);
6019         *p++ = htonl((u32) stat->nlink);
6020 -       *p++ = htonl((u32) nfsd_ruid(rqstp, stat->uid));
6021 -       *p++ = htonl((u32) nfsd_rgid(rqstp, stat->gid));
6022 +       *p++ = htonl((u32) nfsd_ruid(rqstp,
6023 +               TAGINO_UID(0 /* FIXME: DX_TAG(dentry->d_inode) */,
6024 +               stat->uid, stat->tag)));
6025 +       *p++ = htonl((u32) nfsd_rgid(rqstp,
6026 +               TAGINO_GID(0 /* FIXME: DX_TAG(dentry->d_inode) */,
6027 +               stat->gid, stat->tag)));
6028         if (S_ISLNK(stat->mode) && stat->size > NFS3_MAXPATHLEN) {
6029                 p = xdr_encode_hyper(p, (u64) NFS3_MAXPATHLEN);
6030         } else {
6031 diff -NurpP --minimal linux-3.0-rc5/fs/nfsd/nfs4xdr.c linux-3.0-rc5-vs2.3.1-pre3/fs/nfsd/nfs4xdr.c
6032 --- linux-3.0-rc5/fs/nfsd/nfs4xdr.c     2011-07-01 11:14:19.000000000 +0200
6033 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/nfsd/nfs4xdr.c        2011-06-10 22:11:24.000000000 +0200
6034 @@ -45,6 +45,7 @@
6035  #include <linux/statfs.h>
6036  #include <linux/utsname.h>
6037  #include <linux/sunrpc/svcauth_gss.h>
6038 +#include <linux/vs_tag.h>
6039
6040  #include "idmap.h"
6041  #include "acl.h"
6042 @@ -2090,14 +2091,18 @@ out_acl:
6043                 WRITE32(stat.nlink);
6044         }
6045         if (bmval1 & FATTR4_WORD1_OWNER) {
6046 -               status = nfsd4_encode_user(rqstp, stat.uid, &p, &buflen);
6047 +               status = nfsd4_encode_user(rqstp,
6048 +                       TAGINO_UID(DX_TAG(dentry->d_inode),
6049 +                       stat.uid, stat.tag), &p, &buflen);
6050                 if (status == nfserr_resource)
6051                         goto out_resource;
6052                 if (status)
6053                         goto out;
6054         }
6055         if (bmval1 & FATTR4_WORD1_OWNER_GROUP) {
6056 -               status = nfsd4_encode_group(rqstp, stat.gid, &p, &buflen);
6057 +               status = nfsd4_encode_group(rqstp,
6058 +                       TAGINO_GID(DX_TAG(dentry->d_inode),
6059 +                       stat.gid, stat.tag), &p, &buflen);
6060                 if (status == nfserr_resource)
6061                         goto out_resource;
6062                 if (status)
6063 diff -NurpP --minimal linux-3.0-rc5/fs/nfsd/nfsxdr.c linux-3.0-rc5-vs2.3.1-pre3/fs/nfsd/nfsxdr.c
6064 --- linux-3.0-rc5/fs/nfsd/nfsxdr.c      2011-05-22 16:17:53.000000000 +0200
6065 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/nfsd/nfsxdr.c 2011-06-10 22:11:24.000000000 +0200
6066 @@ -6,6 +6,7 @@
6067
6068  #include "xdr.h"
6069  #include "auth.h"
6070 +#include <linux/vs_tag.h>
6071
6072  #define NFSDDBG_FACILITY               NFSDDBG_XDR
6073
6074 @@ -88,6 +89,8 @@ static __be32 *
6075  decode_sattr(__be32 *p, struct iattr *iap)
6076  {
6077         u32     tmp, tmp1;
6078 +       uid_t   uid = 0;
6079 +       gid_t   gid = 0;
6080
6081         iap->ia_valid = 0;
6082
6083 @@ -101,12 +104,15 @@ decode_sattr(__be32 *p, struct iattr *ia
6084         }
6085         if ((tmp = ntohl(*p++)) != (u32)-1) {
6086                 iap->ia_valid |= ATTR_UID;
6087 -               iap->ia_uid = tmp;
6088 +               uid = tmp;
6089         }
6090         if ((tmp = ntohl(*p++)) != (u32)-1) {
6091                 iap->ia_valid |= ATTR_GID;
6092 -               iap->ia_gid = tmp;
6093 +               gid = tmp;
6094         }
6095 +       iap->ia_uid = INOTAG_UID(DX_TAG_NFSD, uid, gid);
6096 +       iap->ia_gid = INOTAG_GID(DX_TAG_NFSD, uid, gid);
6097 +       iap->ia_tag = INOTAG_TAG(DX_TAG_NFSD, uid, gid, 0);
6098         if ((tmp = ntohl(*p++)) != (u32)-1) {
6099                 iap->ia_valid |= ATTR_SIZE;
6100                 iap->ia_size = tmp;
6101 @@ -151,8 +157,10 @@ encode_fattr(struct svc_rqst *rqstp, __b
6102         *p++ = htonl(nfs_ftypes[type >> 12]);
6103         *p++ = htonl((u32) stat->mode);
6104         *p++ = htonl((u32) stat->nlink);
6105 -       *p++ = htonl((u32) nfsd_ruid(rqstp, stat->uid));
6106 -       *p++ = htonl((u32) nfsd_rgid(rqstp, stat->gid));
6107 +       *p++ = htonl((u32) nfsd_ruid(rqstp,
6108 +               TAGINO_UID(DX_TAG(dentry->d_inode), stat->uid, stat->tag)));
6109 +       *p++ = htonl((u32) nfsd_rgid(rqstp,
6110 +               TAGINO_GID(DX_TAG(dentry->d_inode), stat->gid, stat->tag)));
6111
6112         if (S_ISLNK(type) && stat->size > NFS_MAXPATHLEN) {
6113                 *p++ = htonl(NFS_MAXPATHLEN);
6114 diff -NurpP --minimal linux-3.0-rc5/fs/ocfs2/dlmglue.c linux-3.0-rc5-vs2.3.1-pre3/fs/ocfs2/dlmglue.c
6115 --- linux-3.0-rc5/fs/ocfs2/dlmglue.c    2011-05-22 16:17:53.000000000 +0200
6116 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ocfs2/dlmglue.c       2011-06-10 22:11:24.000000000 +0200
6117 @@ -2041,6 +2041,7 @@ static void __ocfs2_stuff_meta_lvb(struc
6118         lvb->lvb_iclusters = cpu_to_be32(oi->ip_clusters);
6119         lvb->lvb_iuid      = cpu_to_be32(inode->i_uid);
6120         lvb->lvb_igid      = cpu_to_be32(inode->i_gid);
6121 +       lvb->lvb_itag      = cpu_to_be16(inode->i_tag);
6122         lvb->lvb_imode     = cpu_to_be16(inode->i_mode);
6123         lvb->lvb_inlink    = cpu_to_be16(inode->i_nlink);
6124         lvb->lvb_iatime_packed  =
6125 @@ -2091,6 +2092,7 @@ static void ocfs2_refresh_inode_from_lvb
6126
6127         inode->i_uid     = be32_to_cpu(lvb->lvb_iuid);
6128         inode->i_gid     = be32_to_cpu(lvb->lvb_igid);
6129 +       inode->i_tag     = be16_to_cpu(lvb->lvb_itag);
6130         inode->i_mode    = be16_to_cpu(lvb->lvb_imode);
6131         inode->i_nlink   = be16_to_cpu(lvb->lvb_inlink);
6132         ocfs2_unpack_timespec(&inode->i_atime,
6133 diff -NurpP --minimal linux-3.0-rc5/fs/ocfs2/dlmglue.h linux-3.0-rc5-vs2.3.1-pre3/fs/ocfs2/dlmglue.h
6134 --- linux-3.0-rc5/fs/ocfs2/dlmglue.h    2010-10-21 13:07:50.000000000 +0200
6135 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ocfs2/dlmglue.h       2011-06-10 22:11:24.000000000 +0200
6136 @@ -46,7 +46,8 @@ struct ocfs2_meta_lvb {
6137         __be16       lvb_inlink;
6138         __be32       lvb_iattr;
6139         __be32       lvb_igeneration;
6140 -       __be32       lvb_reserved2;
6141 +       __be16       lvb_itag;
6142 +       __be16       lvb_reserved2;
6143  };
6144
6145  #define OCFS2_QINFO_LVB_VERSION 1
6146 diff -NurpP --minimal linux-3.0-rc5/fs/ocfs2/file.c linux-3.0-rc5-vs2.3.1-pre3/fs/ocfs2/file.c
6147 --- linux-3.0-rc5/fs/ocfs2/file.c       2011-07-01 11:14:20.000000000 +0200
6148 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ocfs2/file.c  2011-06-10 22:43:33.000000000 +0200
6149 @@ -1111,7 +1111,7 @@ int ocfs2_setattr(struct dentry *dentry,
6150                 attr->ia_valid &= ~ATTR_SIZE;
6151
6152  #define OCFS2_VALID_ATTRS (ATTR_ATIME | ATTR_MTIME | ATTR_CTIME | ATTR_SIZE \
6153 -                          | ATTR_GID | ATTR_UID | ATTR_MODE)
6154 +                          | ATTR_GID | ATTR_UID | ATTR_TAG | ATTR_MODE)
6155         if (!(attr->ia_valid & OCFS2_VALID_ATTRS))
6156                 return 0;
6157
6158 diff -NurpP --minimal linux-3.0-rc5/fs/ocfs2/inode.c linux-3.0-rc5-vs2.3.1-pre3/fs/ocfs2/inode.c
6159 --- linux-3.0-rc5/fs/ocfs2/inode.c      2011-05-22 16:17:53.000000000 +0200
6160 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ocfs2/inode.c 2011-06-10 22:11:24.000000000 +0200
6161 @@ -28,6 +28,7 @@
6162  #include <linux/highmem.h>
6163  #include <linux/pagemap.h>
6164  #include <linux/quotaops.h>
6165 +#include <linux/vs_tag.h>
6166
6167  #include <asm/byteorder.h>
6168
6169 @@ -78,11 +79,13 @@ void ocfs2_set_inode_flags(struct inode
6170  {
6171         unsigned int flags = OCFS2_I(inode)->ip_attr;
6172
6173 -       inode->i_flags &= ~(S_IMMUTABLE |
6174 +       inode->i_flags &= ~(S_IMMUTABLE | S_IXUNLINK |
6175                 S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
6176
6177         if (flags & OCFS2_IMMUTABLE_FL)
6178                 inode->i_flags |= S_IMMUTABLE;
6179 +       if (flags & OCFS2_IXUNLINK_FL)
6180 +               inode->i_flags |= S_IXUNLINK;
6181
6182         if (flags & OCFS2_SYNC_FL)
6183                 inode->i_flags |= S_SYNC;
6184 @@ -92,25 +95,44 @@ void ocfs2_set_inode_flags(struct inode
6185                 inode->i_flags |= S_NOATIME;
6186         if (flags & OCFS2_DIRSYNC_FL)
6187                 inode->i_flags |= S_DIRSYNC;
6188 +
6189 +       inode->i_vflags &= ~(V_BARRIER | V_COW);
6190 +
6191 +       if (flags & OCFS2_BARRIER_FL)
6192 +               inode->i_vflags |= V_BARRIER;
6193 +       if (flags & OCFS2_COW_FL)
6194 +               inode->i_vflags |= V_COW;
6195  }
6196
6197  /* Propagate flags from i_flags to OCFS2_I(inode)->ip_attr */
6198  void ocfs2_get_inode_flags(struct ocfs2_inode_info *oi)
6199  {
6200         unsigned int flags = oi->vfs_inode.i_flags;
6201 +       unsigned int vflags = oi->vfs_inode.i_vflags;
6202 +
6203 +       oi->ip_attr &= ~(OCFS2_SYNC_FL | OCFS2_APPEND_FL |
6204 +                       OCFS2_IMMUTABLE_FL | OCFS2_IXUNLINK_FL |
6205 +                       OCFS2_NOATIME_FL | OCFS2_DIRSYNC_FL |
6206 +                       OCFS2_BARRIER_FL | OCFS2_COW_FL);
6207 +
6208 +       if (flags & S_IMMUTABLE)
6209 +               oi->ip_attr |= OCFS2_IMMUTABLE_FL;
6210 +       if (flags & S_IXUNLINK)
6211 +               oi->ip_attr |= OCFS2_IXUNLINK_FL;
6212
6213 -       oi->ip_attr &= ~(OCFS2_SYNC_FL|OCFS2_APPEND_FL|
6214 -                       OCFS2_IMMUTABLE_FL|OCFS2_NOATIME_FL|OCFS2_DIRSYNC_FL);
6215         if (flags & S_SYNC)
6216                 oi->ip_attr |= OCFS2_SYNC_FL;
6217         if (flags & S_APPEND)
6218                 oi->ip_attr |= OCFS2_APPEND_FL;
6219 -       if (flags & S_IMMUTABLE)
6220 -               oi->ip_attr |= OCFS2_IMMUTABLE_FL;
6221         if (flags & S_NOATIME)
6222                 oi->ip_attr |= OCFS2_NOATIME_FL;
6223         if (flags & S_DIRSYNC)
6224                 oi->ip_attr |= OCFS2_DIRSYNC_FL;
6225 +
6226 +       if (vflags & V_BARRIER)
6227 +               oi->ip_attr |= OCFS2_BARRIER_FL;
6228 +       if (vflags & V_COW)
6229 +               oi->ip_attr |= OCFS2_COW_FL;
6230  }
6231
6232  struct inode *ocfs2_ilookup(struct super_block *sb, u64 blkno)
6233 @@ -241,6 +263,8 @@ void ocfs2_populate_inode(struct inode *
6234         struct super_block *sb;
6235         struct ocfs2_super *osb;
6236         int use_plocks = 1;
6237 +       uid_t uid;
6238 +       gid_t gid;
6239
6240         sb = inode->i_sb;
6241         osb = OCFS2_SB(sb);
6242 @@ -269,8 +293,12 @@ void ocfs2_populate_inode(struct inode *
6243         inode->i_generation = le32_to_cpu(fe->i_generation);
6244         inode->i_rdev = huge_decode_dev(le64_to_cpu(fe->id1.dev1.i_rdev));
6245         inode->i_mode = le16_to_cpu(fe->i_mode);
6246 -       inode->i_uid = le32_to_cpu(fe->i_uid);
6247 -       inode->i_gid = le32_to_cpu(fe->i_gid);
6248 +       uid = le32_to_cpu(fe->i_uid);
6249 +       gid = le32_to_cpu(fe->i_gid);
6250 +       inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
6251 +       inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
6252 +       inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid,
6253 +               /* le16_to_cpu(raw_inode->i_raw_tag)i */ 0);
6254
6255         /* Fast symlinks will have i_size but no allocated clusters. */
6256         if (S_ISLNK(inode->i_mode) && !fe->i_clusters)
6257 diff -NurpP --minimal linux-3.0-rc5/fs/ocfs2/inode.h linux-3.0-rc5-vs2.3.1-pre3/fs/ocfs2/inode.h
6258 --- linux-3.0-rc5/fs/ocfs2/inode.h      2011-01-05 21:50:26.000000000 +0100
6259 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ocfs2/inode.h 2011-06-10 22:11:24.000000000 +0200
6260 @@ -151,6 +151,7 @@ struct buffer_head *ocfs2_bread(struct i
6261
6262  void ocfs2_set_inode_flags(struct inode *inode);
6263  void ocfs2_get_inode_flags(struct ocfs2_inode_info *oi);
6264 +int ocfs2_sync_flags(struct inode *inode, int, int);
6265
6266  static inline blkcnt_t ocfs2_inode_sector_count(struct inode *inode)
6267  {
6268 diff -NurpP --minimal linux-3.0-rc5/fs/ocfs2/ioctl.c linux-3.0-rc5-vs2.3.1-pre3/fs/ocfs2/ioctl.c
6269 --- linux-3.0-rc5/fs/ocfs2/ioctl.c      2011-07-01 11:14:20.000000000 +0200
6270 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ocfs2/ioctl.c 2011-06-10 22:11:24.000000000 +0200
6271 @@ -78,7 +78,41 @@ static int ocfs2_get_inode_attr(struct i
6272         return status;
6273  }
6274
6275 -static int ocfs2_set_inode_attr(struct inode *inode, unsigned flags,
6276 +int ocfs2_sync_flags(struct inode *inode, int flags, int vflags)
6277 +{
6278 +       struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
6279 +       struct buffer_head *bh = NULL;
6280 +       handle_t *handle = NULL;
6281 +       int status;
6282 +
6283 +       status = ocfs2_inode_lock(inode, &bh, 1);
6284 +       if (status < 0) {
6285 +               mlog_errno(status);
6286 +               return status;
6287 +       }
6288 +       handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS);
6289 +       if (IS_ERR(handle)) {
6290 +               status = PTR_ERR(handle);
6291 +               mlog_errno(status);
6292 +               goto bail_unlock;
6293 +       }
6294 +
6295 +       inode->i_flags = flags;
6296 +       inode->i_vflags = vflags;
6297 +       ocfs2_get_inode_flags(OCFS2_I(inode));
6298 +
6299 +       status = ocfs2_mark_inode_dirty(handle, inode, bh);
6300 +       if (status < 0)
6301 +               mlog_errno(status);
6302 +
6303 +       ocfs2_commit_trans(osb, handle);
6304 +bail_unlock:
6305 +       ocfs2_inode_unlock(inode, 1);
6306 +       brelse(bh);
6307 +       return status;
6308 +}
6309 +
6310 +int ocfs2_set_inode_attr(struct inode *inode, unsigned flags,
6311                                 unsigned mask)
6312  {
6313         struct ocfs2_inode_info *ocfs2_inode = OCFS2_I(inode);
6314 @@ -103,6 +137,11 @@ static int ocfs2_set_inode_attr(struct i
6315         if (!S_ISDIR(inode->i_mode))
6316                 flags &= ~OCFS2_DIRSYNC_FL;
6317
6318 +       if (IS_BARRIER(inode)) {
6319 +               vxwprintk_task(1, "messing with the barrier.");
6320 +               goto bail_unlock;
6321 +       }
6322 +
6323         handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS);
6324         if (IS_ERR(handle)) {
6325                 status = PTR_ERR(handle);
6326 @@ -880,6 +919,7 @@ bail:
6327         return status;
6328  }
6329
6330 +
6331  long ocfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
6332  {
6333         struct inode *inode = filp->f_path.dentry->d_inode;
6334 diff -NurpP --minimal linux-3.0-rc5/fs/ocfs2/namei.c linux-3.0-rc5-vs2.3.1-pre3/fs/ocfs2/namei.c
6335 --- linux-3.0-rc5/fs/ocfs2/namei.c      2011-05-22 16:17:53.000000000 +0200
6336 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ocfs2/namei.c 2011-06-10 22:11:24.000000000 +0200
6337 @@ -41,6 +41,7 @@
6338  #include <linux/slab.h>
6339  #include <linux/highmem.h>
6340  #include <linux/quotaops.h>
6341 +#include <linux/vs_tag.h>
6342
6343  #include <cluster/masklog.h>
6344
6345 @@ -477,6 +478,7 @@ static int __ocfs2_mknod_locked(struct i
6346         struct ocfs2_dinode *fe = NULL;
6347         struct ocfs2_extent_list *fel;
6348         u16 feat;
6349 +       tag_t tag;
6350
6351         *new_fe_bh = NULL;
6352
6353 @@ -514,8 +516,11 @@ static int __ocfs2_mknod_locked(struct i
6354         fe->i_suballoc_loc = cpu_to_le64(suballoc_loc);
6355         fe->i_suballoc_bit = cpu_to_le16(suballoc_bit);
6356         fe->i_suballoc_slot = cpu_to_le16(inode_ac->ac_alloc_slot);
6357 -       fe->i_uid = cpu_to_le32(inode->i_uid);
6358 -       fe->i_gid = cpu_to_le32(inode->i_gid);
6359 +
6360 +       tag = dx_current_fstag(osb->sb);
6361 +       fe->i_uid = cpu_to_le32(TAGINO_UID(DX_TAG(inode), inode->i_uid, tag));
6362 +       fe->i_gid = cpu_to_le32(TAGINO_GID(DX_TAG(inode), inode->i_gid, tag));
6363 +       inode->i_tag = tag;
6364         fe->i_mode = cpu_to_le16(inode->i_mode);
6365         if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode))
6366                 fe->id1.dev1.i_rdev = cpu_to_le64(huge_encode_dev(dev));
6367 diff -NurpP --minimal linux-3.0-rc5/fs/ocfs2/ocfs2.h linux-3.0-rc5-vs2.3.1-pre3/fs/ocfs2/ocfs2.h
6368 --- linux-3.0-rc5/fs/ocfs2/ocfs2.h      2011-05-22 16:17:53.000000000 +0200
6369 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ocfs2/ocfs2.h 2011-06-10 22:11:24.000000000 +0200
6370 @@ -272,6 +272,7 @@ enum ocfs2_mount_options
6371                                                      writes */
6372         OCFS2_MOUNT_HB_NONE = 1 << 13, /* No heartbeat */
6373         OCFS2_MOUNT_HB_GLOBAL = 1 << 14, /* Global heartbeat */
6374 +       OCFS2_MOUNT_TAGGED = 1 << 15, /* use tagging */
6375  };
6376
6377  #define OCFS2_OSB_SOFT_RO                      0x0001
6378 diff -NurpP --minimal linux-3.0-rc5/fs/ocfs2/ocfs2_fs.h linux-3.0-rc5-vs2.3.1-pre3/fs/ocfs2/ocfs2_fs.h
6379 --- linux-3.0-rc5/fs/ocfs2/ocfs2_fs.h   2011-05-22 16:17:53.000000000 +0200
6380 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ocfs2/ocfs2_fs.h      2011-06-10 22:11:24.000000000 +0200
6381 @@ -266,6 +266,11 @@
6382  #define OCFS2_TOPDIR_FL                        FS_TOPDIR_FL    /* Top of directory hierarchies*/
6383  #define OCFS2_RESERVED_FL              FS_RESERVED_FL  /* reserved for ext2 lib */
6384
6385 +#define OCFS2_IXUNLINK_FL              FS_IXUNLINK_FL  /* Immutable invert on unlink */
6386 +
6387 +#define OCFS2_BARRIER_FL               FS_BARRIER_FL   /* Barrier for chroot() */
6388 +#define OCFS2_COW_FL                   FS_COW_FL       /* Copy on Write marker */
6389 +
6390  #define OCFS2_FL_VISIBLE               FS_FL_USER_VISIBLE      /* User visible flags */
6391  #define OCFS2_FL_MODIFIABLE            FS_FL_USER_MODIFIABLE   /* User modifiable flags */
6392
6393 diff -NurpP --minimal linux-3.0-rc5/fs/ocfs2/super.c linux-3.0-rc5-vs2.3.1-pre3/fs/ocfs2/super.c
6394 --- linux-3.0-rc5/fs/ocfs2/super.c      2011-07-01 11:14:20.000000000 +0200
6395 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/ocfs2/super.c 2011-06-15 02:40:14.000000000 +0200
6396 @@ -184,6 +184,7 @@ enum {
6397         Opt_coherency_full,
6398         Opt_resv_level,
6399         Opt_dir_resv_level,
6400 +       Opt_tag, Opt_notag, Opt_tagid,
6401         Opt_err,
6402  };
6403
6404 @@ -215,6 +216,9 @@ static const match_table_t tokens = {
6405         {Opt_coherency_full, "coherency=full"},
6406         {Opt_resv_level, "resv_level=%u"},
6407         {Opt_dir_resv_level, "dir_resv_level=%u"},
6408 +       {Opt_tag, "tag"},
6409 +       {Opt_notag, "notag"},
6410 +       {Opt_tagid, "tagid=%u"},
6411         {Opt_err, NULL}
6412  };
6413
6414 @@ -662,6 +666,13 @@ static int ocfs2_remount(struct super_bl
6415                 goto out;
6416         }
6417
6418 +       if ((osb->s_mount_opt & OCFS2_MOUNT_TAGGED) !=
6419 +           (parsed_options.mount_opt & OCFS2_MOUNT_TAGGED)) {
6420 +               ret = -EINVAL;
6421 +               mlog(ML_ERROR, "Cannot change tagging on remount\n");
6422 +               goto out;
6423 +       }
6424 +
6425         /* We're going to/from readonly mode. */
6426         if ((*flags & MS_RDONLY) != (sb->s_flags & MS_RDONLY)) {
6427                 /* Disable quota accounting before remounting RO */
6428 @@ -1177,6 +1188,9 @@ static int ocfs2_fill_super(struct super
6429
6430         ocfs2_complete_mount_recovery(osb);
6431
6432 +       if (osb->s_mount_opt & OCFS2_MOUNT_TAGGED)
6433 +               sb->s_flags |= MS_TAGGED;
6434 +
6435         if (ocfs2_mount_local(osb))
6436                 snprintf(nodestr, sizeof(nodestr), "local");
6437         else
6438 @@ -1506,6 +1520,20 @@ static int ocfs2_parse_options(struct su
6439                             option < OCFS2_MAX_RESV_LEVEL)
6440                                 mopt->dir_resv_level = option;
6441                         break;
6442 +#ifndef CONFIG_TAGGING_NONE
6443 +               case Opt_tag:
6444 +                       mopt->mount_opt |= OCFS2_MOUNT_TAGGED;
6445 +                       break;
6446 +               case Opt_notag:
6447 +                       mopt->mount_opt &= ~OCFS2_MOUNT_TAGGED;
6448 +                       break;
6449 +#endif
6450 +#ifdef CONFIG_PROPAGATE
6451 +               case Opt_tagid:
6452 +                       /* use args[0] */
6453 +                       mopt->mount_opt |= OCFS2_MOUNT_TAGGED;
6454 +                       break;
6455 +#endif
6456                 default:
6457                         mlog(ML_ERROR,
6458                              "Unrecognized mount option \"%s\" "
6459 diff -NurpP --minimal linux-3.0-rc5/fs/open.c linux-3.0-rc5-vs2.3.1-pre3/fs/open.c
6460 --- linux-3.0-rc5/fs/open.c     2011-05-22 16:17:53.000000000 +0200
6461 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/open.c        2011-06-10 22:11:24.000000000 +0200
6462 @@ -30,6 +30,11 @@
6463  #include <linux/fs_struct.h>
6464  #include <linux/ima.h>
6465  #include <linux/dnotify.h>
6466 +#include <linux/vs_base.h>
6467 +#include <linux/vs_limit.h>
6468 +#include <linux/vs_tag.h>
6469 +#include <linux/vs_cowbl.h>
6470 +#include <linux/vserver/dlimit.h>
6471
6472  #include "internal.h"
6473
6474 @@ -494,6 +499,12 @@ SYSCALL_DEFINE3(fchmodat, int, dfd, cons
6475         error = user_path_at(dfd, filename, LOOKUP_FOLLOW, &path);
6476         if (error)
6477                 goto out;
6478 +
6479 +#ifdef CONFIG_VSERVER_COWBL
6480 +       error = cow_check_and_break(&path);
6481 +       if (error)
6482 +               goto dput_and_out;
6483 +#endif
6484         inode = path.dentry->d_inode;
6485
6486         error = mnt_want_write(path.mnt);
6487 @@ -531,11 +542,11 @@ static int chown_common(struct path *pat
6488         newattrs.ia_valid =  ATTR_CTIME;
6489         if (user != (uid_t) -1) {
6490                 newattrs.ia_valid |= ATTR_UID;
6491 -               newattrs.ia_uid = user;
6492 +               newattrs.ia_uid = dx_map_uid(user);
6493         }
6494         if (group != (gid_t) -1) {
6495                 newattrs.ia_valid |= ATTR_GID;
6496 -               newattrs.ia_gid = group;
6497 +               newattrs.ia_gid = dx_map_gid(group);
6498         }
6499         if (!S_ISDIR(inode->i_mode))
6500                 newattrs.ia_valid |=
6501 @@ -560,6 +571,10 @@ SYSCALL_DEFINE3(chown, const char __user
6502         error = mnt_want_write(path.mnt);
6503         if (error)
6504                 goto out_release;
6505 +#ifdef CONFIG_VSERVER_COWBL
6506 +       error = cow_check_and_break(&path);
6507 +       if (!error)
6508 +#endif
6509         error = chown_common(&path, user, group);
6510         mnt_drop_write(path.mnt);
6511  out_release:
6512 @@ -587,6 +602,10 @@ SYSCALL_DEFINE5(fchownat, int, dfd, cons
6513         error = mnt_want_write(path.mnt);
6514         if (error)
6515                 goto out_release;
6516 +#ifdef CONFIG_VSERVER_COWBL
6517 +       error = cow_check_and_break(&path);
6518 +       if (!error)
6519 +#endif
6520         error = chown_common(&path, user, group);
6521         mnt_drop_write(path.mnt);
6522  out_release:
6523 @@ -606,6 +625,10 @@ SYSCALL_DEFINE3(lchown, const char __use
6524         error = mnt_want_write(path.mnt);
6525         if (error)
6526                 goto out_release;
6527 +#ifdef CONFIG_VSERVER_COWBL
6528 +       error = cow_check_and_break(&path);
6529 +       if (!error)
6530 +#endif
6531         error = chown_common(&path, user, group);
6532         mnt_drop_write(path.mnt);
6533  out_release:
6534 @@ -857,6 +880,7 @@ static void __put_unused_fd(struct files
6535         __FD_CLR(fd, fdt->open_fds);
6536         if (fd < files->next_fd)
6537                 files->next_fd = fd;
6538 +       vx_openfd_dec(fd);
6539  }
6540
6541  void put_unused_fd(unsigned int fd)
6542 diff -NurpP --minimal linux-3.0-rc5/fs/proc/array.c linux-3.0-rc5-vs2.3.1-pre3/fs/proc/array.c
6543 --- linux-3.0-rc5/fs/proc/array.c       2011-07-01 11:14:20.000000000 +0200
6544 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/proc/array.c  2011-06-10 22:11:24.000000000 +0200
6545 @@ -81,6 +81,8 @@
6546  #include <linux/pid_namespace.h>
6547  #include <linux/ptrace.h>
6548  #include <linux/tracehook.h>
6549 +#include <linux/vs_context.h>
6550 +#include <linux/vs_network.h>
6551
6552  #include <asm/pgtable.h>
6553  #include <asm/processor.h>
6554 @@ -170,6 +172,9 @@ static inline void task_state(struct seq
6555         rcu_read_lock();
6556         ppid = pid_alive(p) ?
6557                 task_tgid_nr_ns(rcu_dereference(p->real_parent), ns) : 0;
6558 +       if (unlikely(vx_current_initpid(p->pid)))
6559 +               ppid = 0;
6560 +
6561         tpid = 0;
6562         if (pid_alive(p)) {
6563                 struct task_struct *tracer = tracehook_tracer_task(p);
6564 @@ -287,7 +292,7 @@ static inline void task_sig(struct seq_f
6565  }
6566
6567  static void render_cap_t(struct seq_file *m, const char *header,
6568 -                       kernel_cap_t *a)
6569 +                       struct vx_info *vxi, kernel_cap_t *a)
6570  {
6571         unsigned __capi;
6572
6573 @@ -312,10 +317,11 @@ static inline void task_cap(struct seq_f
6574         cap_bset        = cred->cap_bset;
6575         rcu_read_unlock();
6576
6577 -       render_cap_t(m, "CapInh:\t", &cap_inheritable);
6578 -       render_cap_t(m, "CapPrm:\t", &cap_permitted);
6579 -       render_cap_t(m, "CapEff:\t", &cap_effective);
6580 -       render_cap_t(m, "CapBnd:\t", &cap_bset);
6581 +       /* FIXME: maybe move the p->vx_info masking to __task_cred() ? */
6582 +       render_cap_t(m, "CapInh:\t", p->vx_info, &cap_inheritable);
6583 +       render_cap_t(m, "CapPrm:\t", p->vx_info, &cap_permitted);
6584 +       render_cap_t(m, "CapEff:\t", p->vx_info, &cap_effective);
6585 +       render_cap_t(m, "CapBnd:\t", p->vx_info, &cap_bset);
6586  }
6587
6588  static inline void task_context_switch_counts(struct seq_file *m,
6589 @@ -337,6 +343,42 @@ static void task_cpus_allowed(struct seq
6590         seq_putc(m, '\n');
6591  }
6592
6593 +int proc_pid_nsproxy(struct seq_file *m, struct pid_namespace *ns,
6594 +                       struct pid *pid, struct task_struct *task)
6595 +{
6596 +       seq_printf(m,   "Proxy:\t%p(%c)\n"
6597 +                       "Count:\t%u\n"
6598 +                       "uts:\t%p(%c)\n"
6599 +                       "ipc:\t%p(%c)\n"
6600 +                       "mnt:\t%p(%c)\n"
6601 +                       "pid:\t%p(%c)\n"
6602 +                       "net:\t%p(%c)\n",
6603 +                       task->nsproxy,
6604 +                       (task->nsproxy == init_task.nsproxy ? 'I' : '-'),
6605 +                       atomic_read(&task->nsproxy->count),
6606 +                       task->nsproxy->uts_ns,
6607 +                       (task->nsproxy->uts_ns == init_task.nsproxy->uts_ns ? 'I' : '-'),
6608 +                       task->nsproxy->ipc_ns,
6609 +                       (task->nsproxy->ipc_ns == init_task.nsproxy->ipc_ns ? 'I' : '-'),
6610 +                       task->nsproxy->mnt_ns,
6611 +                       (task->nsproxy->mnt_ns == init_task.nsproxy->mnt_ns ? 'I' : '-'),
6612 +                       task->nsproxy->pid_ns,
6613 +                       (task->nsproxy->pid_ns == init_task.nsproxy->pid_ns ? 'I' : '-'),
6614 +                       task->nsproxy->net_ns,
6615 +                       (task->nsproxy->net_ns == init_task.nsproxy->net_ns ? 'I' : '-'));
6616 +       return 0;
6617 +}
6618 +
6619 +void task_vs_id(struct seq_file *m, struct task_struct *task)
6620 +{
6621 +       if (task_vx_flags(task, VXF_HIDE_VINFO, 0))
6622 +               return;
6623 +
6624 +       seq_printf(m, "VxID: %d\n", vx_task_xid(task));
6625 +       seq_printf(m, "NxID: %d\n", nx_task_nid(task));
6626 +}
6627 +
6628 +
6629  int proc_pid_status(struct seq_file *m, struct pid_namespace *ns,
6630                         struct pid *pid, struct task_struct *task)
6631  {
6632 @@ -353,6 +395,7 @@ int proc_pid_status(struct seq_file *m,
6633         task_cap(m, task);
6634         task_cpus_allowed(m, task);
6635         cpuset_task_status_allowed(m, task);
6636 +       task_vs_id(m, task);
6637         task_context_switch_counts(m, task);
6638         return 0;
6639  }
6640 @@ -462,6 +505,17 @@ static int do_task_stat(struct seq_file
6641         /* convert nsec -> ticks */
6642         start_time = nsec_to_clock_t(start_time);
6643
6644 +       /* fixup start time for virt uptime */
6645 +       if (vx_flags(VXF_VIRT_UPTIME, 0)) {
6646 +               unsigned long long bias =
6647 +                       current->vx_info->cvirt.bias_clock;
6648 +
6649 +               if (start_time > bias)
6650 +                       start_time -= bias;
6651 +               else
6652 +                       start_time = 0;
6653 +       }
6654 +
6655         seq_printf(m, "%d (%s) %c %d %d %d %d %d %u %lu \
6656  %lu %lu %lu %lu %lu %ld %ld %ld %ld %d 0 %llu %lu %ld %lu %lu %lu %lu %lu \
6657  %lu %lu %lu %lu %lu %lu %lu %lu %d %d %u %u %llu %lu %ld\n",
6658 diff -NurpP --minimal linux-3.0-rc5/fs/proc/base.c linux-3.0-rc5-vs2.3.1-pre3/fs/proc/base.c
6659 --- linux-3.0-rc5/fs/proc/base.c        2011-07-01 11:14:20.000000000 +0200
6660 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/proc/base.c   2011-06-22 12:39:15.000000000 +0200
6661 @@ -83,6 +83,8 @@
6662  #include <linux/pid_namespace.h>
6663  #include <linux/fs_struct.h>
6664  #include <linux/slab.h>
6665 +#include <linux/vs_context.h>
6666 +#include <linux/vs_network.h>
6667  #ifdef CONFIG_HARDWALL
6668  #include <asm/hardwall.h>
6669  #endif
6670 @@ -1102,11 +1104,16 @@ static ssize_t oom_adjust_write(struct f
6671                 goto err_task_lock;
6672         }
6673
6674 -       if (oom_adjust < task->signal->oom_adj && !capable(CAP_SYS_RESOURCE)) {
6675 +       if (oom_adjust < task->signal->oom_adj &&
6676 +               !vx_capable(CAP_SYS_RESOURCE, VXC_OOM_ADJUST)) {
6677                 err = -EACCES;
6678                 goto err_sighand;
6679         }
6680
6681 +       /* prevent guest processes from circumventing the oom killer */
6682 +       if (vx_current_xid() && (oom_adjust == OOM_DISABLE))
6683 +               oom_adjust = OOM_ADJUST_MIN;
6684 +
6685         if (oom_adjust != task->signal->oom_adj) {
6686                 if (oom_adjust == OOM_DISABLE)
6687                         atomic_inc(&task->mm->oom_disable_count);
6688 @@ -1275,7 +1282,7 @@ static ssize_t proc_loginuid_write(struc
6689         ssize_t length;
6690         uid_t loginuid;
6691
6692 -       if (!capable(CAP_AUDIT_CONTROL))
6693 +       if (!vx_capable(CAP_AUDIT_CONTROL, VXC_AUDIT_CONTROL))
6694                 return -EPERM;
6695
6696         rcu_read_lock();
6697 @@ -1722,6 +1729,8 @@ struct inode *proc_pid_make_inode(struct
6698                 inode->i_gid = cred->egid;
6699                 rcu_read_unlock();
6700         }
6701 +       /* procfs is xid tagged */
6702 +       inode->i_tag = (tag_t)vx_task_xid(task);
6703         security_task_to_inode(task, inode);
6704
6705  out:
6706 @@ -1758,6 +1767,8 @@ int pid_getattr(struct vfsmount *mnt, st
6707
6708  /* dentry stuff */
6709
6710 +static unsigned name_to_int(struct dentry *dentry);
6711 +
6712  /*
6713   *     Exceptional case: normally we are not allowed to unhash a busy
6714   * directory. In this case, however, we can do it - no aliasing problems
6715 @@ -1786,6 +1797,12 @@ int pid_revalidate(struct dentry *dentry
6716         task = get_proc_task(inode);
6717
6718         if (task) {
6719 +               unsigned pid = name_to_int(dentry);
6720 +
6721 +               if (pid != ~0U && pid != vx_map_pid(task->pid)) {
6722 +                       put_task_struct(task);
6723 +                       goto drop;
6724 +               }
6725                 if ((inode->i_mode == (S_IFDIR|S_IRUGO|S_IXUGO)) ||
6726                     task_dumpable(task)) {
6727                         rcu_read_lock();
6728 @@ -1802,6 +1819,7 @@ int pid_revalidate(struct dentry *dentry
6729                 put_task_struct(task);
6730                 return 1;
6731         }
6732 +drop:
6733         d_drop(dentry);
6734         return 0;
6735  }
6736 @@ -2283,6 +2301,13 @@ static struct dentry *proc_pident_lookup
6737         if (!task)
6738                 goto out_no_task;
6739
6740 +       /* TODO: maybe we can come up with a generic approach? */
6741 +       if (task_vx_flags(task, VXF_HIDE_VINFO, 0) &&
6742 +               (dentry->d_name.len == 5) &&
6743 +               (!memcmp(dentry->d_name.name, "vinfo", 5) ||
6744 +               !memcmp(dentry->d_name.name, "ninfo", 5)))
6745 +               goto out;
6746 +
6747         /*
6748          * Yes, it does not scale. And it should not. Don't add
6749          * new entries into /proc/<tgid>/ without very good reasons.
6750 @@ -2668,7 +2693,7 @@ out_iput:
6751  static struct dentry *proc_base_lookup(struct inode *dir, struct dentry *dentry)
6752  {
6753         struct dentry *error;
6754 -       struct task_struct *task = get_proc_task(dir);
6755 +       struct task_struct *task = get_proc_task_real(dir);
6756         const struct pid_entry *p, *last;
6757
6758         error = ERR_PTR(-ENOENT);
6759 @@ -2762,6 +2787,9 @@ static int proc_pid_personality(struct s
6760  static const struct file_operations proc_task_operations;
6761  static const struct inode_operations proc_task_inode_operations;
6762
6763 +extern int proc_pid_vx_info(struct task_struct *, char *);
6764 +extern int proc_pid_nx_info(struct task_struct *, char *);
6765 +
6766  static const struct pid_entry tgid_base_stuff[] = {
6767         DIR("task",       S_IRUGO|S_IXUGO, proc_task_inode_operations, proc_task_operations),
6768         DIR("fd",         S_IRUSR|S_IXUSR, proc_fd_inode_operations, proc_fd_operations),
6769 @@ -2825,6 +2853,8 @@ static const struct pid_entry tgid_base_
6770  #ifdef CONFIG_CGROUPS
6771         REG("cgroup",  S_IRUGO, proc_cgroup_operations),
6772  #endif
6773 +       INF("vinfo",      S_IRUGO, proc_pid_vx_info),
6774 +       INF("ninfo",      S_IRUGO, proc_pid_nx_info),
6775         INF("oom_score",  S_IRUGO, proc_oom_score),
6776         REG("oom_adj",    S_IRUGO|S_IWUSR, proc_oom_adjust_operations),
6777         REG("oom_score_adj", S_IRUGO|S_IWUSR, proc_oom_score_adj_operations),
6778 @@ -2844,6 +2874,7 @@ static const struct pid_entry tgid_base_
6779  #ifdef CONFIG_HARDWALL
6780         INF("hardwall",   S_IRUGO, proc_pid_hardwall),
6781  #endif
6782 +       ONE("nsproxy",  S_IRUGO, proc_pid_nsproxy),
6783  };
6784
6785  static int proc_tgid_base_readdir(struct file * filp,
6786 @@ -3036,7 +3067,7 @@ retry:
6787         iter.task = NULL;
6788         pid = find_ge_pid(iter.tgid, ns);
6789         if (pid) {
6790 -               iter.tgid = pid_nr_ns(pid, ns);
6791 +               iter.tgid = pid_unmapped_nr_ns(pid, ns);
6792                 iter.task = pid_task(pid, PIDTYPE_PID);
6793                 /* What we to know is if the pid we have find is the
6794                  * pid of a thread_group_leader.  Testing for task
6795 @@ -3066,7 +3097,7 @@ static int proc_pid_fill_cache(struct fi
6796         struct tgid_iter iter)
6797  {
6798         char name[PROC_NUMBUF];
6799 -       int len = snprintf(name, sizeof(name), "%d", iter.tgid);
6800 +       int len = snprintf(name, sizeof(name), "%d", vx_map_tgid(iter.tgid));
6801         return proc_fill_cache(filp, dirent, filldir, name, len,
6802                                 proc_pid_instantiate, iter.task, NULL);
6803  }
6804 @@ -3083,7 +3114,7 @@ int proc_pid_readdir(struct file * filp,
6805                 goto out_no_task;
6806         nr = filp->f_pos - FIRST_PROCESS_ENTRY;
6807
6808 -       reaper = get_proc_task(filp->f_path.dentry->d_inode);
6809 +       reaper = get_proc_task_real(filp->f_path.dentry->d_inode);
6810         if (!reaper)
6811                 goto out_no_task;
6812
6813 @@ -3100,6 +3131,8 @@ int proc_pid_readdir(struct file * filp,
6814              iter.task;
6815              iter.tgid += 1, iter = next_tgid(ns, iter)) {
6816                 filp->f_pos = iter.tgid + TGID_OFFSET;
6817 +               if (!vx_proc_task_visible(iter.task))
6818 +                       continue;
6819                 if (proc_pid_fill_cache(filp, dirent, filldir, iter) < 0) {
6820                         put_task_struct(iter.task);
6821                         goto out;
6822 @@ -3253,6 +3286,8 @@ static struct dentry *proc_task_lookup(s
6823         tid = name_to_int(dentry);
6824         if (tid == ~0U)
6825                 goto out;
6826 +       if (vx_current_initpid(tid))
6827 +               goto out;
6828
6829         ns = dentry->d_sb->s_fs_info;
6830         rcu_read_lock();
6831 diff -NurpP --minimal linux-3.0-rc5/fs/proc/generic.c linux-3.0-rc5-vs2.3.1-pre3/fs/proc/generic.c
6832 --- linux-3.0-rc5/fs/proc/generic.c     2011-07-01 11:14:20.000000000 +0200
6833 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/proc/generic.c        2011-06-10 22:11:24.000000000 +0200
6834 @@ -22,6 +22,7 @@
6835  #include <linux/bitops.h>
6836  #include <linux/spinlock.h>
6837  #include <linux/completion.h>
6838 +#include <linux/vserver/inode.h>
6839  #include <asm/uaccess.h>
6840
6841  #include "internal.h"
6842 @@ -424,11 +425,15 @@ struct dentry *proc_lookup_de(struct pro
6843         for (de = de->subdir; de ; de = de->next) {
6844                 if (de->namelen != dentry->d_name.len)
6845                         continue;
6846 +               if (!vx_hide_check(0, de->vx_flags))
6847 +                       continue;
6848                 if (!memcmp(dentry->d_name.name, de->name, de->namelen)) {
6849                         pde_get(de);
6850                         spin_unlock(&proc_subdir_lock);
6851                         error = -EINVAL;
6852                         inode = proc_get_inode(dir->i_sb, de);
6853 +                       /* generic proc entries belong to the host */
6854 +                       inode->i_tag = 0;
6855                         goto out_unlock;
6856                 }
6857         }
6858 @@ -506,6 +511,8 @@ int proc_readdir_de(struct proc_dir_entr
6859
6860                                 /* filldir passes info to user space */
6861                                 pde_get(de);
6862 +                               if (!vx_hide_check(0, de->vx_flags))
6863 +                                       goto skip;
6864                                 spin_unlock(&proc_subdir_lock);
6865                                 if (filldir(dirent, de->name, de->namelen, filp->f_pos,
6866                                             de->low_ino, de->mode >> 12) < 0) {
6867 @@ -513,6 +520,7 @@ int proc_readdir_de(struct proc_dir_entr
6868                                         goto out;
6869                                 }
6870                                 spin_lock(&proc_subdir_lock);
6871 +                       skip:
6872                                 filp->f_pos++;
6873                                 next = de->next;
6874                                 pde_put(de);
6875 @@ -627,6 +635,7 @@ static struct proc_dir_entry *__proc_cre
6876         ent->nlink = nlink;
6877         atomic_set(&ent->count, 1);
6878         ent->pde_users = 0;
6879 +       ent->vx_flags = IATTR_PROC_DEFAULT;
6880         spin_lock_init(&ent->pde_unload_lock);
6881         ent->pde_unload_completion = NULL;
6882         INIT_LIST_HEAD(&ent->pde_openers);
6883 @@ -650,7 +659,8 @@ struct proc_dir_entry *proc_symlink(cons
6884                                 kfree(ent->data);
6885                                 kfree(ent);
6886                                 ent = NULL;
6887 -                       }
6888 +                       } else
6889 +                               ent->vx_flags = IATTR_PROC_SYMLINK;
6890                 } else {
6891                         kfree(ent);
6892                         ent = NULL;
6893 diff -NurpP --minimal linux-3.0-rc5/fs/proc/inode.c linux-3.0-rc5-vs2.3.1-pre3/fs/proc/inode.c
6894 --- linux-3.0-rc5/fs/proc/inode.c       2011-07-01 11:14:20.000000000 +0200
6895 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/proc/inode.c  2011-06-10 22:11:24.000000000 +0200
6896 @@ -442,6 +442,8 @@ struct inode *proc_get_inode(struct supe
6897                         inode->i_uid = de->uid;
6898                         inode->i_gid = de->gid;
6899                 }
6900 +               if (de->vx_flags)
6901 +                       PROC_I(inode)->vx_flags = de->vx_flags;
6902                 if (de->size)
6903                         inode->i_size = de->size;
6904                 if (de->nlink)
6905 diff -NurpP --minimal linux-3.0-rc5/fs/proc/internal.h linux-3.0-rc5-vs2.3.1-pre3/fs/proc/internal.h
6906 --- linux-3.0-rc5/fs/proc/internal.h    2011-07-01 11:14:20.000000000 +0200
6907 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/proc/internal.h       2011-06-10 22:11:24.000000000 +0200
6908 @@ -10,6 +10,7 @@
6909   */
6910
6911  #include <linux/proc_fs.h>
6912 +#include <linux/vs_pid.h>
6913
6914  extern struct proc_dir_entry proc_root;
6915  #ifdef CONFIG_PROC_SYSCTL
6916 @@ -51,6 +52,9 @@ extern int proc_pid_status(struct seq_fi
6917                                 struct pid *pid, struct task_struct *task);
6918  extern int proc_pid_statm(struct seq_file *m, struct pid_namespace *ns,
6919                                 struct pid *pid, struct task_struct *task);
6920 +extern int proc_pid_nsproxy(struct seq_file *m, struct pid_namespace *ns,
6921 +                               struct pid *pid, struct task_struct *task);
6922 +
6923  extern loff_t mem_lseek(struct file *file, loff_t offset, int orig);
6924
6925  extern const struct file_operations proc_maps_operations;
6926 @@ -76,11 +80,16 @@ static inline struct pid *proc_pid(struc
6927         return PROC_I(inode)->pid;
6928  }
6929
6930 -static inline struct task_struct *get_proc_task(struct inode *inode)
6931 +static inline struct task_struct *get_proc_task_real(struct inode *inode)
6932  {
6933         return get_pid_task(proc_pid(inode), PIDTYPE_PID);
6934  }
6935
6936 +static inline struct task_struct *get_proc_task(struct inode *inode)
6937 +{
6938 +       return vx_get_proc_task(inode, proc_pid(inode));
6939 +}
6940 +
6941  static inline int proc_fd(struct inode *inode)
6942  {
6943         return PROC_I(inode)->fd;
6944 diff -NurpP --minimal linux-3.0-rc5/fs/proc/loadavg.c linux-3.0-rc5-vs2.3.1-pre3/fs/proc/loadavg.c
6945 --- linux-3.0-rc5/fs/proc/loadavg.c     2009-09-10 15:26:23.000000000 +0200
6946 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/proc/loadavg.c        2011-06-10 22:11:24.000000000 +0200
6947 @@ -12,15 +12,27 @@
6948
6949  static int loadavg_proc_show(struct seq_file *m, void *v)
6950  {
6951 +       unsigned long running;
6952 +       unsigned int threads;
6953         unsigned long avnrun[3];
6954
6955         get_avenrun(avnrun, FIXED_1/200, 0);
6956
6957 +       if (vx_flags(VXF_VIRT_LOAD, 0)) {
6958 +               struct vx_info *vxi = current_vx_info();
6959 +
6960 +               running = atomic_read(&vxi->cvirt.nr_running);
6961 +               threads = atomic_read(&vxi->cvirt.nr_threads);
6962 +       } else {
6963 +               running = nr_running();
6964 +               threads = nr_threads;
6965 +       }
6966 +
6967         seq_printf(m, "%lu.%02lu %lu.%02lu %lu.%02lu %ld/%d %d\n",
6968                 LOAD_INT(avnrun[0]), LOAD_FRAC(avnrun[0]),
6969                 LOAD_INT(avnrun[1]), LOAD_FRAC(avnrun[1]),
6970                 LOAD_INT(avnrun[2]), LOAD_FRAC(avnrun[2]),
6971 -               nr_running(), nr_threads,
6972 +               running, threads,
6973                 task_active_pid_ns(current)->last_pid);
6974         return 0;
6975  }
6976 diff -NurpP --minimal linux-3.0-rc5/fs/proc/meminfo.c linux-3.0-rc5-vs2.3.1-pre3/fs/proc/meminfo.c
6977 --- linux-3.0-rc5/fs/proc/meminfo.c     2011-03-15 18:07:33.000000000 +0100
6978 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/proc/meminfo.c        2011-06-10 22:11:24.000000000 +0200
6979 @@ -39,7 +39,8 @@ static int meminfo_proc_show(struct seq_
6980         allowed = ((totalram_pages - hugetlb_total_pages())
6981                 * sysctl_overcommit_ratio / 100) + total_swap_pages;
6982
6983 -       cached = global_page_state(NR_FILE_PAGES) -
6984 +       cached = vx_flags(VXF_VIRT_MEM, 0) ?
6985 +               vx_vsi_cached(&i) : global_page_state(NR_FILE_PAGES) -
6986                         total_swapcache_pages - i.bufferram;
6987         if (cached < 0)
6988                 cached = 0;
6989 diff -NurpP --minimal linux-3.0-rc5/fs/proc/root.c linux-3.0-rc5-vs2.3.1-pre3/fs/proc/root.c
6990 --- linux-3.0-rc5/fs/proc/root.c        2011-07-01 11:14:20.000000000 +0200
6991 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/proc/root.c   2011-06-22 12:39:15.000000000 +0200
6992 @@ -18,9 +18,14 @@
6993  #include <linux/bitops.h>
6994  #include <linux/mount.h>
6995  #include <linux/pid_namespace.h>
6996 +#include <linux/vserver/inode.h>
6997
6998  #include "internal.h"
6999
7000 +struct proc_dir_entry *proc_virtual;
7001 +
7002 +extern void proc_vx_init(void);
7003 +
7004  static int proc_test_super(struct super_block *sb, void *data)
7005  {
7006         return sb->s_fs_info == data;
7007 @@ -125,6 +130,7 @@ void __init proc_root_init(void)
7008  #endif
7009         proc_mkdir("bus", NULL);
7010         proc_sys_init();
7011 +       proc_vx_init();
7012  }
7013
7014  static int proc_root_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat
7015 @@ -193,6 +199,7 @@ struct proc_dir_entry proc_root = {
7016         .proc_iops      = &proc_root_inode_operations,
7017         .proc_fops      = &proc_root_operations,
7018         .parent         = &proc_root,
7019 +       .vx_flags       = IATTR_ADMIN | IATTR_WATCH,
7020  };
7021
7022  int pid_ns_prepare_proc(struct pid_namespace *ns)
7023 diff -NurpP --minimal linux-3.0-rc5/fs/proc/uptime.c linux-3.0-rc5-vs2.3.1-pre3/fs/proc/uptime.c
7024 --- linux-3.0-rc5/fs/proc/uptime.c      2009-12-03 20:02:53.000000000 +0100
7025 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/proc/uptime.c 2011-06-10 22:11:24.000000000 +0200
7026 @@ -4,22 +4,22 @@
7027  #include <linux/sched.h>
7028  #include <linux/seq_file.h>
7029  #include <linux/time.h>
7030 -#include <linux/kernel_stat.h>
7031 +#include <linux/vserver/cvirt.h>
7032  #include <asm/cputime.h>
7033
7034  static int uptime_proc_show(struct seq_file *m, void *v)
7035  {
7036         struct timespec uptime;
7037         struct timespec idle;
7038 -       int i;
7039 -       cputime_t idletime = cputime_zero;
7040 -
7041 -       for_each_possible_cpu(i)
7042 -               idletime = cputime64_add(idletime, kstat_cpu(i).cpustat.idle);
7043 +       cputime_t idletime = cputime_add(init_task.utime, init_task.stime);
7044
7045         do_posix_clock_monotonic_gettime(&uptime);
7046         monotonic_to_bootbased(&uptime);
7047         cputime_to_timespec(idletime, &idle);
7048 +
7049 +       if (vx_flags(VXF_VIRT_UPTIME, 0))
7050 +               vx_vsi_uptime(&uptime, &idle);
7051 +
7052         seq_printf(m, "%lu.%02lu %lu.%02lu\n",
7053                         (unsigned long) uptime.tv_sec,
7054                         (uptime.tv_nsec / (NSEC_PER_SEC / 100)),
7055 diff -NurpP --minimal linux-3.0-rc5/fs/quota/dquot.c linux-3.0-rc5-vs2.3.1-pre3/fs/quota/dquot.c
7056 --- linux-3.0-rc5/fs/quota/dquot.c      2011-07-01 11:14:20.000000000 +0200
7057 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/quota/dquot.c 2011-06-10 22:11:24.000000000 +0200
7058 @@ -1548,6 +1548,9 @@ int __dquot_alloc_space(struct inode *in
7059         int reserve = flags & DQUOT_SPACE_RESERVE;
7060         int nofail = flags & DQUOT_SPACE_NOFAIL;
7061
7062 +       if ((ret = dl_alloc_space(inode, number)))
7063 +               return ret;
7064 +
7065         /*
7066          * First test before acquiring mutex - solves deadlocks when we
7067          * re-enter the quota code and are already holding the mutex
7068 @@ -1602,6 +1605,9 @@ int dquot_alloc_inode(const struct inode
7069         int cnt, ret = 0;
7070         char warntype[MAXQUOTAS];
7071
7072 +       if ((ret = dl_alloc_inode(inode)))
7073 +               return ret;
7074 +
7075         /* First test before acquiring mutex - solves deadlocks when we
7076           * re-enter the quota code and are already holding the mutex */
7077         if (!dquot_active(inode))
7078 @@ -1672,6 +1678,8 @@ void __dquot_free_space(struct inode *in
7079         char warntype[MAXQUOTAS];
7080         int reserve = flags & DQUOT_SPACE_RESERVE;
7081
7082 +       dl_free_space(inode, number);
7083 +
7084         /* First test before acquiring mutex - solves deadlocks when we
7085           * re-enter the quota code and are already holding the mutex */
7086         if (!dquot_active(inode)) {
7087 @@ -1710,6 +1718,8 @@ void dquot_free_inode(const struct inode
7088         unsigned int cnt;
7089         char warntype[MAXQUOTAS];
7090
7091 +       dl_free_inode(inode);
7092 +
7093         /* First test before acquiring mutex - solves deadlocks when we
7094           * re-enter the quota code and are already holding the mutex */
7095         if (!dquot_active(inode))
7096 diff -NurpP --minimal linux-3.0-rc5/fs/quota/quota.c linux-3.0-rc5-vs2.3.1-pre3/fs/quota/quota.c
7097 --- linux-3.0-rc5/fs/quota/quota.c      2011-03-15 18:07:34.000000000 +0100
7098 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/quota/quota.c 2011-06-13 18:19:47.000000000 +0200
7099 @@ -8,6 +8,7 @@
7100  #include <linux/fs.h>
7101  #include <linux/namei.h>
7102  #include <linux/slab.h>
7103 +#include <linux/vs_context.h>
7104  #include <asm/current.h>
7105  #include <asm/uaccess.h>
7106  #include <linux/kernel.h>
7107 @@ -38,7 +39,7 @@ static int check_quotactl_permission(str
7108                         break;
7109                 /*FALLTHROUGH*/
7110         default:
7111 -               if (!capable(CAP_SYS_ADMIN))
7112 +               if (!vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL))
7113                         return -EPERM;
7114         }
7115
7116 @@ -293,6 +294,46 @@ static int do_quotactl(struct super_bloc
7117         }
7118  }
7119
7120 +#if defined(CONFIG_BLK_DEV_VROOT) || defined(CONFIG_BLK_DEV_VROOT_MODULE)
7121 +
7122 +#include <linux/vroot.h>
7123 +#include <linux/major.h>
7124 +#include <linux/module.h>
7125 +#include <linux/kallsyms.h>
7126 +#include <linux/vserver/debug.h>
7127 +
7128 +static vroot_grb_func *vroot_get_real_bdev = NULL;
7129 +
7130 +static DEFINE_SPINLOCK(vroot_grb_lock);
7131 +
7132 +int register_vroot_grb(vroot_grb_func *func) {
7133 +       int ret = -EBUSY;
7134 +
7135 +       spin_lock(&vroot_grb_lock);
7136 +       if (!vroot_get_real_bdev) {
7137 +               vroot_get_real_bdev = func;
7138 +               ret = 0;
7139 +       }
7140 +       spin_unlock(&vroot_grb_lock);
7141 +       return ret;
7142 +}
7143 +EXPORT_SYMBOL(register_vroot_grb);
7144 +
7145 +int unregister_vroot_grb(vroot_grb_func *func) {
7146 +       int ret = -EINVAL;
7147 +
7148 +       spin_lock(&vroot_grb_lock);
7149 +       if (vroot_get_real_bdev) {
7150 +               vroot_get_real_bdev = NULL;
7151 +               ret = 0;
7152 +       }
7153 +       spin_unlock(&vroot_grb_lock);
7154 +       return ret;
7155 +}
7156 +EXPORT_SYMBOL(unregister_vroot_grb);
7157 +
7158 +#endif
7159 +
7160  /*
7161   * look up a superblock on which quota ops will be performed
7162   * - use the name of a block device to find the superblock thereon
7163 @@ -310,6 +351,22 @@ static struct super_block *quotactl_bloc
7164         putname(tmp);
7165         if (IS_ERR(bdev))
7166                 return ERR_CAST(bdev);
7167 +#if defined(CONFIG_BLK_DEV_VROOT) || defined(CONFIG_BLK_DEV_VROOT_MODULE)
7168 +       if (bdev && bdev->bd_inode &&
7169 +                       imajor(bdev->bd_inode) == VROOT_MAJOR) {
7170 +               struct block_device *bdnew = (void *)-EINVAL;
7171 +
7172 +               if (vroot_get_real_bdev)
7173 +                       bdnew = vroot_get_real_bdev(bdev);
7174 +               else
7175 +                       vxdprintk(VXD_CBIT(misc, 0),
7176 +                                       "vroot_get_real_bdev not set");
7177 +               bdput(bdev);
7178 +               if (IS_ERR(bdnew))
7179 +                       return ERR_PTR(PTR_ERR(bdnew));
7180 +               bdev = bdnew;
7181 +       }
7182 +#endif
7183         sb = get_super(bdev);
7184         bdput(bdev);
7185         if (!sb)
7186 diff -NurpP --minimal linux-3.0-rc5/fs/reiserfs/file.c linux-3.0-rc5-vs2.3.1-pre3/fs/reiserfs/file.c
7187 --- linux-3.0-rc5/fs/reiserfs/file.c    2011-01-05 21:50:26.000000000 +0100
7188 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/reiserfs/file.c       2011-06-10 22:11:24.000000000 +0200
7189 @@ -312,4 +312,5 @@ const struct inode_operations reiserfs_f
7190         .listxattr = reiserfs_listxattr,
7191         .removexattr = reiserfs_removexattr,
7192         .permission = reiserfs_permission,
7193 +       .sync_flags = reiserfs_sync_flags,
7194  };
7195 diff -NurpP --minimal linux-3.0-rc5/fs/reiserfs/inode.c linux-3.0-rc5-vs2.3.1-pre3/fs/reiserfs/inode.c
7196 --- linux-3.0-rc5/fs/reiserfs/inode.c   2011-05-22 16:17:53.000000000 +0200
7197 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/reiserfs/inode.c      2011-06-10 22:11:24.000000000 +0200
7198 @@ -18,6 +18,7 @@
7199  #include <linux/writeback.h>
7200  #include <linux/quotaops.h>
7201  #include <linux/swap.h>
7202 +#include <linux/vs_tag.h>
7203
7204  int reiserfs_commit_write(struct file *f, struct page *page,
7205                           unsigned from, unsigned to);
7206 @@ -1131,6 +1132,8 @@ static void init_inode(struct inode *ino
7207         struct buffer_head *bh;
7208         struct item_head *ih;
7209         __u32 rdev;
7210 +       uid_t uid;
7211 +       gid_t gid;
7212         //int version = ITEM_VERSION_1;
7213
7214         bh = PATH_PLAST_BUFFER(path);
7215 @@ -1151,12 +1154,13 @@ static void init_inode(struct inode *ino
7216                     (struct stat_data_v1 *)B_I_PITEM(bh, ih);
7217                 unsigned long blocks;
7218
7219 +               uid = sd_v1_uid(sd);
7220 +               gid = sd_v1_gid(sd);
7221 +
7222                 set_inode_item_key_version(inode, KEY_FORMAT_3_5);
7223                 set_inode_sd_version(inode, STAT_DATA_V1);
7224                 inode->i_mode = sd_v1_mode(sd);
7225                 inode->i_nlink = sd_v1_nlink(sd);
7226 -               inode->i_uid = sd_v1_uid(sd);
7227 -               inode->i_gid = sd_v1_gid(sd);
7228                 inode->i_size = sd_v1_size(sd);
7229                 inode->i_atime.tv_sec = sd_v1_atime(sd);
7230                 inode->i_mtime.tv_sec = sd_v1_mtime(sd);
7231 @@ -1198,11 +1202,12 @@ static void init_inode(struct inode *ino
7232                 // (directories and symlinks)
7233                 struct stat_data *sd = (struct stat_data *)B_I_PITEM(bh, ih);
7234
7235 +               uid    = sd_v2_uid(sd);
7236 +               gid    = sd_v2_gid(sd);
7237 +
7238                 inode->i_mode = sd_v2_mode(sd);
7239                 inode->i_nlink = sd_v2_nlink(sd);
7240 -               inode->i_uid = sd_v2_uid(sd);
7241                 inode->i_size = sd_v2_size(sd);
7242 -               inode->i_gid = sd_v2_gid(sd);
7243                 inode->i_mtime.tv_sec = sd_v2_mtime(sd);
7244                 inode->i_atime.tv_sec = sd_v2_atime(sd);
7245                 inode->i_ctime.tv_sec = sd_v2_ctime(sd);
7246 @@ -1232,6 +1237,10 @@ static void init_inode(struct inode *ino
7247                 sd_attrs_to_i_attrs(sd_v2_attrs(sd), inode);
7248         }
7249
7250 +       inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
7251 +       inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
7252 +       inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid, 0);
7253 +
7254         pathrelse(path);
7255         if (S_ISREG(inode->i_mode)) {
7256                 inode->i_op = &reiserfs_file_inode_operations;
7257 @@ -1254,13 +1263,15 @@ static void init_inode(struct inode *ino
7258  static void inode2sd(void *sd, struct inode *inode, loff_t size)
7259  {
7260         struct stat_data *sd_v2 = (struct stat_data *)sd;
7261 +       uid_t uid = TAGINO_UID(DX_TAG(inode), inode->i_uid, inode->i_tag);
7262 +       gid_t gid = TAGINO_GID(DX_TAG(inode), inode->i_gid, inode->i_tag);
7263         __u16 flags;
7264
7265 +       set_sd_v2_uid(sd_v2, uid);
7266 +       set_sd_v2_gid(sd_v2, gid);
7267         set_sd_v2_mode(sd_v2, inode->i_mode);
7268         set_sd_v2_nlink(sd_v2, inode->i_nlink);
7269 -       set_sd_v2_uid(sd_v2, inode->i_uid);
7270         set_sd_v2_size(sd_v2, size);
7271 -       set_sd_v2_gid(sd_v2, inode->i_gid);
7272         set_sd_v2_mtime(sd_v2, inode->i_mtime.tv_sec);
7273         set_sd_v2_atime(sd_v2, inode->i_atime.tv_sec);
7274         set_sd_v2_ctime(sd_v2, inode->i_ctime.tv_sec);
7275 @@ -2863,14 +2874,19 @@ int reiserfs_commit_write(struct file *f
7276  void sd_attrs_to_i_attrs(__u16 sd_attrs, struct inode *inode)
7277  {
7278         if (reiserfs_attrs(inode->i_sb)) {
7279 -               if (sd_attrs & REISERFS_SYNC_FL)
7280 -                       inode->i_flags |= S_SYNC;
7281 -               else
7282 -                       inode->i_flags &= ~S_SYNC;
7283                 if (sd_attrs & REISERFS_IMMUTABLE_FL)
7284                         inode->i_flags |= S_IMMUTABLE;
7285                 else
7286                         inode->i_flags &= ~S_IMMUTABLE;
7287 +               if (sd_attrs & REISERFS_IXUNLINK_FL)
7288 +                       inode->i_flags |= S_IXUNLINK;
7289 +               else
7290 +                       inode->i_flags &= ~S_IXUNLINK;
7291 +
7292 +               if (sd_attrs & REISERFS_SYNC_FL)
7293 +                       inode->i_flags |= S_SYNC;
7294 +               else
7295 +                       inode->i_flags &= ~S_SYNC;
7296                 if (sd_attrs & REISERFS_APPEND_FL)
7297                         inode->i_flags |= S_APPEND;
7298                 else
7299 @@ -2883,6 +2899,15 @@ void sd_attrs_to_i_attrs(__u16 sd_attrs,
7300                         REISERFS_I(inode)->i_flags |= i_nopack_mask;
7301                 else
7302                         REISERFS_I(inode)->i_flags &= ~i_nopack_mask;
7303 +
7304 +               if (sd_attrs & REISERFS_BARRIER_FL)
7305 +                       inode->i_vflags |= V_BARRIER;
7306 +               else
7307 +                       inode->i_vflags &= ~V_BARRIER;
7308 +               if (sd_attrs & REISERFS_COW_FL)
7309 +                       inode->i_vflags |= V_COW;
7310 +               else
7311 +                       inode->i_vflags &= ~V_COW;
7312         }
7313  }
7314
7315 @@ -2893,6 +2918,11 @@ void i_attrs_to_sd_attrs(struct inode *i
7316                         *sd_attrs |= REISERFS_IMMUTABLE_FL;
7317                 else
7318                         *sd_attrs &= ~REISERFS_IMMUTABLE_FL;
7319 +               if (inode->i_flags & S_IXUNLINK)
7320 +                       *sd_attrs |= REISERFS_IXUNLINK_FL;
7321 +               else
7322 +                       *sd_attrs &= ~REISERFS_IXUNLINK_FL;
7323 +
7324                 if (inode->i_flags & S_SYNC)
7325                         *sd_attrs |= REISERFS_SYNC_FL;
7326                 else
7327 @@ -2905,6 +2935,15 @@ void i_attrs_to_sd_attrs(struct inode *i
7328                         *sd_attrs |= REISERFS_NOTAIL_FL;
7329                 else
7330                         *sd_attrs &= ~REISERFS_NOTAIL_FL;
7331 +
7332 +               if (inode->i_vflags & V_BARRIER)
7333 +                       *sd_attrs |= REISERFS_BARRIER_FL;
7334 +               else
7335 +                       *sd_attrs &= ~REISERFS_BARRIER_FL;
7336 +               if (inode->i_vflags & V_COW)
7337 +                       *sd_attrs |= REISERFS_COW_FL;
7338 +               else
7339 +                       *sd_attrs &= ~REISERFS_COW_FL;
7340         }
7341  }
7342
7343 @@ -3148,7 +3187,8 @@ int reiserfs_setattr(struct dentry *dent
7344         }
7345
7346         if ((ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) ||
7347 -           (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) {
7348 +           (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid) ||
7349 +           (ia_valid & ATTR_TAG && attr->ia_tag != inode->i_tag)) {
7350                 struct reiserfs_transaction_handle th;
7351                 int jbegin_count =
7352                     2 *
7353 @@ -3177,6 +3217,9 @@ int reiserfs_setattr(struct dentry *dent
7354                         inode->i_uid = attr->ia_uid;
7355                 if (attr->ia_valid & ATTR_GID)
7356                         inode->i_gid = attr->ia_gid;
7357 +                               if ((attr->ia_valid & ATTR_TAG) &&
7358 +                                       IS_TAGGED(inode))
7359 +                                       inode->i_tag = attr->ia_tag;
7360                 mark_inode_dirty(inode);
7361                 error = journal_end(&th, inode->i_sb, jbegin_count);
7362                 if (error)
7363 diff -NurpP --minimal linux-3.0-rc5/fs/reiserfs/ioctl.c linux-3.0-rc5-vs2.3.1-pre3/fs/reiserfs/ioctl.c
7364 --- linux-3.0-rc5/fs/reiserfs/ioctl.c   2011-05-22 16:17:53.000000000 +0200
7365 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/reiserfs/ioctl.c      2011-06-10 22:11:24.000000000 +0200
7366 @@ -11,6 +11,21 @@
7367  #include <linux/pagemap.h>
7368  #include <linux/compat.h>
7369
7370 +
7371 +int reiserfs_sync_flags(struct inode *inode, int flags, int vflags)
7372 +{
7373 +       __u16 sd_attrs = 0;
7374 +
7375 +       inode->i_flags = flags;
7376 +       inode->i_vflags = vflags;
7377 +
7378 +       i_attrs_to_sd_attrs(inode, &sd_attrs);
7379 +       REISERFS_I(inode)->i_attrs = sd_attrs;
7380 +       inode->i_ctime = CURRENT_TIME_SEC;
7381 +       mark_inode_dirty(inode);
7382 +       return 0;
7383 +}
7384 +
7385  /*
7386   * reiserfs_ioctl - handler for ioctl for inode
7387   * supported commands:
7388 @@ -22,7 +37,7 @@
7389  long reiserfs_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
7390  {
7391         struct inode *inode = filp->f_path.dentry->d_inode;
7392 -       unsigned int flags;
7393 +       unsigned int flags, oldflags;
7394         int err = 0;
7395
7396         reiserfs_write_lock(inode->i_sb);
7397 @@ -47,6 +62,7 @@ long reiserfs_ioctl(struct file *filp, u
7398
7399                 flags = REISERFS_I(inode)->i_attrs;
7400                 i_attrs_to_sd_attrs(inode, (__u16 *) & flags);
7401 +               flags &= REISERFS_FL_USER_VISIBLE;
7402                 err = put_user(flags, (int __user *)arg);
7403                 break;
7404         case REISERFS_IOC_SETFLAGS:{
7405 @@ -67,6 +83,10 @@ long reiserfs_ioctl(struct file *filp, u
7406                                 err = -EFAULT;
7407                                 goto setflags_out;
7408                         }
7409 +                       if (IS_BARRIER(inode)) {
7410 +                               vxwprintk_task(1, "messing with the barrier.");
7411 +                               return -EACCES;
7412 +                       }
7413                         /*
7414                          * Is it quota file? Do not allow user to mess with it
7415                          */
7416 @@ -91,6 +111,10 @@ long reiserfs_ioctl(struct file *filp, u
7417                                         goto setflags_out;
7418                                 }
7419                         }
7420 +
7421 +                       oldflags = REISERFS_I(inode)->i_attrs;
7422 +                       flags &= REISERFS_FL_USER_MODIFIABLE;
7423 +                       flags |= oldflags & ~REISERFS_FL_USER_MODIFIABLE;
7424                         sd_attrs_to_i_attrs(flags, inode);
7425                         REISERFS_I(inode)->i_attrs = flags;
7426                         inode->i_ctime = CURRENT_TIME_SEC;
7427 diff -NurpP --minimal linux-3.0-rc5/fs/reiserfs/namei.c linux-3.0-rc5-vs2.3.1-pre3/fs/reiserfs/namei.c
7428 --- linux-3.0-rc5/fs/reiserfs/namei.c   2011-05-22 16:17:53.000000000 +0200
7429 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/reiserfs/namei.c      2011-06-10 22:11:24.000000000 +0200
7430 @@ -18,6 +18,7 @@
7431  #include <linux/reiserfs_acl.h>
7432  #include <linux/reiserfs_xattr.h>
7433  #include <linux/quotaops.h>
7434 +#include <linux/vs_tag.h>
7435
7436  #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; }
7437  #define DEC_DIR_INODE_NLINK(i) if (i->i_nlink != 1) drop_nlink(i);
7438 @@ -362,6 +363,7 @@ static struct dentry *reiserfs_lookup(st
7439         if (retval == IO_ERROR) {
7440                 return ERR_PTR(-EIO);
7441         }
7442 +               dx_propagate_tag(nd, inode);
7443
7444         return d_splice_alias(inode, dentry);
7445  }
7446 @@ -1529,6 +1531,7 @@ const struct inode_operations reiserfs_d
7447         .listxattr = reiserfs_listxattr,
7448         .removexattr = reiserfs_removexattr,
7449         .permission = reiserfs_permission,
7450 +       .sync_flags = reiserfs_sync_flags,
7451  };
7452
7453  /*
7454 diff -NurpP --minimal linux-3.0-rc5/fs/reiserfs/super.c linux-3.0-rc5-vs2.3.1-pre3/fs/reiserfs/super.c
7455 --- linux-3.0-rc5/fs/reiserfs/super.c   2011-07-01 11:14:20.000000000 +0200
7456 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/reiserfs/super.c      2011-06-10 22:11:24.000000000 +0200
7457 @@ -899,6 +899,14 @@ static int reiserfs_parse_options(struct
7458                 {"user_xattr",.setmask = 1 << REISERFS_UNSUPPORTED_OPT},
7459                 {"nouser_xattr",.clrmask = 1 << REISERFS_UNSUPPORTED_OPT},
7460  #endif
7461 +#ifndef CONFIG_TAGGING_NONE
7462 +               {"tagxid",.setmask = 1 << REISERFS_TAGGED},
7463 +               {"tag",.setmask = 1 << REISERFS_TAGGED},
7464 +               {"notag",.clrmask = 1 << REISERFS_TAGGED},
7465 +#endif
7466 +#ifdef CONFIG_PROPAGATE
7467 +               {"tag",.arg_required = 'T',.values = NULL},
7468 +#endif
7469  #ifdef CONFIG_REISERFS_FS_POSIX_ACL
7470                 {"acl",.setmask = 1 << REISERFS_POSIXACL},
7471                 {"noacl",.clrmask = 1 << REISERFS_POSIXACL},
7472 @@ -1208,6 +1216,14 @@ static int reiserfs_remount(struct super
7473         handle_quota_files(s, qf_names, &qfmt);
7474  #endif
7475
7476 +       if ((mount_options & (1 << REISERFS_TAGGED)) &&
7477 +               !(s->s_flags & MS_TAGGED)) {
7478 +               reiserfs_warning(s, "super-vs01",
7479 +                       "reiserfs: tagging not permitted on remount.");
7480 +               err = -EINVAL;
7481 +               goto out_err;
7482 +       }
7483 +
7484         handle_attrs(s);
7485
7486         /* Add options that are safe here */
7487 @@ -1690,6 +1706,10 @@ static int reiserfs_fill_super(struct su
7488                 goto error;
7489         }
7490
7491 +       /* map mount option tagxid */
7492 +       if (REISERFS_SB(s)->s_mount_opt & (1 << REISERFS_TAGGED))
7493 +               s->s_flags |= MS_TAGGED;
7494 +
7495         rs = SB_DISK_SUPER_BLOCK(s);
7496         /* Let's do basic sanity check to verify that underlying device is not
7497            smaller than the filesystem. If the check fails then abort and scream,
7498 diff -NurpP --minimal linux-3.0-rc5/fs/reiserfs/xattr.c linux-3.0-rc5-vs2.3.1-pre3/fs/reiserfs/xattr.c
7499 --- linux-3.0-rc5/fs/reiserfs/xattr.c   2011-07-01 11:14:20.000000000 +0200
7500 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/reiserfs/xattr.c      2011-06-22 12:39:15.000000000 +0200
7501 @@ -40,6 +40,7 @@
7502  #include <linux/errno.h>
7503  #include <linux/gfp.h>
7504  #include <linux/fs.h>
7505 +#include <linux/mount.h>
7506  #include <linux/file.h>
7507  #include <linux/pagemap.h>
7508  #include <linux/xattr.h>
7509 diff -NurpP --minimal linux-3.0-rc5/fs/stat.c linux-3.0-rc5-vs2.3.1-pre3/fs/stat.c
7510 --- linux-3.0-rc5/fs/stat.c     2011-05-22 16:17:54.000000000 +0200
7511 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/stat.c        2011-06-10 22:11:24.000000000 +0200
7512 @@ -26,6 +26,7 @@ void generic_fillattr(struct inode *inod
7513         stat->nlink = inode->i_nlink;
7514         stat->uid = inode->i_uid;
7515         stat->gid = inode->i_gid;
7516 +       stat->tag = inode->i_tag;
7517         stat->rdev = inode->i_rdev;
7518         stat->atime = inode->i_atime;
7519         stat->mtime = inode->i_mtime;
7520 diff -NurpP --minimal linux-3.0-rc5/fs/statfs.c linux-3.0-rc5-vs2.3.1-pre3/fs/statfs.c
7521 --- linux-3.0-rc5/fs/statfs.c   2011-05-22 16:17:54.000000000 +0200
7522 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/statfs.c      2011-06-10 22:11:24.000000000 +0200
7523 @@ -7,6 +7,8 @@
7524  #include <linux/statfs.h>
7525  #include <linux/security.h>
7526  #include <linux/uaccess.h>
7527 +#include <linux/vs_base.h>
7528 +#include <linux/vs_dlimit.h>
7529
7530  static int flags_by_mnt(int mnt_flags)
7531  {
7532 @@ -59,6 +61,8 @@ int statfs_by_dentry(struct dentry *dent
7533         retval = dentry->d_sb->s_op->statfs(dentry, buf);
7534         if (retval == 0 && buf->f_frsize == 0)
7535                 buf->f_frsize = buf->f_bsize;
7536 +       if (!vx_check(0, VS_ADMIN|VS_WATCH))
7537 +               vx_vsi_statfs(dentry->d_sb, buf);
7538         return retval;
7539  }
7540
7541 diff -NurpP --minimal linux-3.0-rc5/fs/sysfs/mount.c linux-3.0-rc5-vs2.3.1-pre3/fs/sysfs/mount.c
7542 --- linux-3.0-rc5/fs/sysfs/mount.c      2011-07-01 11:14:20.000000000 +0200
7543 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/sysfs/mount.c 2011-06-22 12:39:15.000000000 +0200
7544 @@ -47,7 +47,7 @@ static int sysfs_fill_super(struct super
7545
7546         sb->s_blocksize = PAGE_CACHE_SIZE;
7547         sb->s_blocksize_bits = PAGE_CACHE_SHIFT;
7548 -       sb->s_magic = SYSFS_MAGIC;
7549 +       sb->s_magic = SYSFS_SUPER_MAGIC;
7550         sb->s_op = &sysfs_ops;
7551         sb->s_time_gran = 1;
7552
7553 diff -NurpP --minimal linux-3.0-rc5/fs/utimes.c linux-3.0-rc5-vs2.3.1-pre3/fs/utimes.c
7554 --- linux-3.0-rc5/fs/utimes.c   2011-05-22 16:17:54.000000000 +0200
7555 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/utimes.c      2011-06-10 22:11:24.000000000 +0200
7556 @@ -8,6 +8,8 @@
7557  #include <linux/stat.h>
7558  #include <linux/utime.h>
7559  #include <linux/syscalls.h>
7560 +#include <linux/mount.h>
7561 +#include <linux/vs_cowbl.h>
7562  #include <asm/uaccess.h>
7563  #include <asm/unistd.h>
7564
7565 @@ -52,12 +54,18 @@ static int utimes_common(struct path *pa
7566  {
7567         int error;
7568         struct iattr newattrs;
7569 -       struct inode *inode = path->dentry->d_inode;
7570 +       struct inode *inode;
7571
7572         error = mnt_want_write(path->mnt);
7573         if (error)
7574                 goto out;
7575
7576 +       error = cow_check_and_break(path);
7577 +       if (error)
7578 +               goto mnt_drop_write_and_out;
7579 +
7580 +       inode = path->dentry->d_inode;
7581 +
7582         if (times && times[0].tv_nsec == UTIME_NOW &&
7583                      times[1].tv_nsec == UTIME_NOW)
7584                 times = NULL;
7585 diff -NurpP --minimal linux-3.0-rc5/fs/xattr.c linux-3.0-rc5-vs2.3.1-pre3/fs/xattr.c
7586 --- linux-3.0-rc5/fs/xattr.c    2011-07-01 11:14:20.000000000 +0200
7587 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/xattr.c       2011-06-10 23:10:19.000000000 +0200
7588 @@ -18,6 +18,7 @@
7589  #include <linux/module.h>
7590  #include <linux/fsnotify.h>
7591  #include <linux/audit.h>
7592 +#include <linux/mount.h>
7593  #include <asm/uaccess.h>
7594
7595
7596 @@ -49,7 +50,7 @@ xattr_permission(struct inode *inode, co
7597          * The trusted.* namespace can only be accessed by privileged users.
7598          */
7599         if (!strncmp(name, XATTR_TRUSTED_PREFIX, XATTR_TRUSTED_PREFIX_LEN)) {
7600 -               if (!capable(CAP_SYS_ADMIN))
7601 +               if (!vx_capable(CAP_SYS_ADMIN, VXC_FS_TRUSTED))
7602                         return (mask & MAY_WRITE) ? -EPERM : -ENODATA;
7603                 return 0;
7604         }
7605 diff -NurpP --minimal linux-3.0-rc5/fs/xfs/linux-2.6/xfs_ioctl.c linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/linux-2.6/xfs_ioctl.c
7606 --- linux-3.0-rc5/fs/xfs/linux-2.6/xfs_ioctl.c  2011-05-22 16:17:54.000000000 +0200
7607 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/linux-2.6/xfs_ioctl.c     2011-06-10 22:11:24.000000000 +0200
7608 @@ -28,7 +28,7 @@
7609  #include "xfs_bmap_btree.h"
7610  #include "xfs_dinode.h"
7611  #include "xfs_inode.h"
7612 -#include "xfs_ioctl.h"
7613 +// #include "xfs_ioctl.h"
7614  #include "xfs_rtalloc.h"
7615  #include "xfs_itable.h"
7616  #include "xfs_error.h"
7617 @@ -748,6 +748,10 @@ xfs_merge_ioc_xflags(
7618                 xflags |= XFS_XFLAG_IMMUTABLE;
7619         else
7620                 xflags &= ~XFS_XFLAG_IMMUTABLE;
7621 +       if (flags & FS_IXUNLINK_FL)
7622 +               xflags |= XFS_XFLAG_IXUNLINK;
7623 +       else
7624 +               xflags &= ~XFS_XFLAG_IXUNLINK;
7625         if (flags & FS_APPEND_FL)
7626                 xflags |= XFS_XFLAG_APPEND;
7627         else
7628 @@ -776,6 +780,8 @@ xfs_di2lxflags(
7629
7630         if (di_flags & XFS_DIFLAG_IMMUTABLE)
7631                 flags |= FS_IMMUTABLE_FL;
7632 +       if (di_flags & XFS_DIFLAG_IXUNLINK)
7633 +               flags |= FS_IXUNLINK_FL;
7634         if (di_flags & XFS_DIFLAG_APPEND)
7635                 flags |= FS_APPEND_FL;
7636         if (di_flags & XFS_DIFLAG_SYNC)
7637 @@ -836,6 +842,8 @@ xfs_set_diflags(
7638         di_flags = (ip->i_d.di_flags & XFS_DIFLAG_PREALLOC);
7639         if (xflags & XFS_XFLAG_IMMUTABLE)
7640                 di_flags |= XFS_DIFLAG_IMMUTABLE;
7641 +       if (xflags & XFS_XFLAG_IXUNLINK)
7642 +               di_flags |= XFS_DIFLAG_IXUNLINK;
7643         if (xflags & XFS_XFLAG_APPEND)
7644                 di_flags |= XFS_DIFLAG_APPEND;
7645         if (xflags & XFS_XFLAG_SYNC)
7646 @@ -878,6 +886,10 @@ xfs_diflags_to_linux(
7647                 inode->i_flags |= S_IMMUTABLE;
7648         else
7649                 inode->i_flags &= ~S_IMMUTABLE;
7650 +       if (xflags & XFS_XFLAG_IXUNLINK)
7651 +               inode->i_flags |= S_IXUNLINK;
7652 +       else
7653 +               inode->i_flags &= ~S_IXUNLINK;
7654         if (xflags & XFS_XFLAG_APPEND)
7655                 inode->i_flags |= S_APPEND;
7656         else
7657 @@ -1370,10 +1382,18 @@ xfs_file_ioctl(
7658         case XFS_IOC_FSGETXATTRA:
7659                 return xfs_ioc_fsgetxattr(ip, 1, arg);
7660         case XFS_IOC_FSSETXATTR:
7661 +               if (IS_BARRIER(inode)) {
7662 +                       vxwprintk_task(1, "messing with the barrier.");
7663 +                       return -XFS_ERROR(EACCES);
7664 +               }
7665                 return xfs_ioc_fssetxattr(ip, filp, arg);
7666         case XFS_IOC_GETXFLAGS:
7667                 return xfs_ioc_getxflags(ip, arg);
7668         case XFS_IOC_SETXFLAGS:
7669 +               if (IS_BARRIER(inode)) {
7670 +                       vxwprintk_task(1, "messing with the barrier.");
7671 +                       return -XFS_ERROR(EACCES);
7672 +               }
7673                 return xfs_ioc_setxflags(ip, filp, arg);
7674
7675         case XFS_IOC_FSSETDM: {
7676 diff -NurpP --minimal linux-3.0-rc5/fs/xfs/linux-2.6/xfs_ioctl.h linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/linux-2.6/xfs_ioctl.h
7677 --- linux-3.0-rc5/fs/xfs/linux-2.6/xfs_ioctl.h  2010-07-07 18:31:54.000000000 +0200
7678 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/linux-2.6/xfs_ioctl.h     2011-06-10 22:11:24.000000000 +0200
7679 @@ -70,6 +70,12 @@ xfs_handle_to_dentry(
7680         void __user             *uhandle,
7681         u32                     hlen);
7682
7683 +extern int
7684 +xfs_sync_flags(
7685 +       struct inode            *inode,
7686 +       int                     flags,
7687 +       int                     vflags);
7688 +
7689  extern long
7690  xfs_file_ioctl(
7691         struct file             *filp,
7692 diff -NurpP --minimal linux-3.0-rc5/fs/xfs/linux-2.6/xfs_iops.c linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/linux-2.6/xfs_iops.c
7693 --- linux-3.0-rc5/fs/xfs/linux-2.6/xfs_iops.c   2011-07-01 11:14:20.000000000 +0200
7694 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/linux-2.6/xfs_iops.c      2011-06-22 12:39:15.000000000 +0200
7695 @@ -30,6 +30,7 @@
7696  #include "xfs_bmap_btree.h"
7697  #include "xfs_dinode.h"
7698  #include "xfs_inode.h"
7699 +#include "xfs_ioctl.h"
7700  #include "xfs_bmap.h"
7701  #include "xfs_rtalloc.h"
7702  #include "xfs_error.h"
7703 @@ -48,6 +49,7 @@
7704  #include <linux/security.h>
7705  #include <linux/fiemap.h>
7706  #include <linux/slab.h>
7707 +#include <linux/vs_tag.h>
7708
7709  /*
7710   * Bring the timestamps in the XFS inode uptodate.
7711 @@ -464,6 +466,7 @@ xfs_vn_getattr(
7712         stat->nlink = ip->i_d.di_nlink;
7713         stat->uid = ip->i_d.di_uid;
7714         stat->gid = ip->i_d.di_gid;
7715 +       stat->tag = ip->i_d.di_tag;
7716         stat->ino = ip->i_ino;
7717         stat->atime = inode->i_atime;
7718         stat->mtime = inode->i_mtime;
7719 @@ -599,6 +602,7 @@ static const struct inode_operations xfs
7720         .removexattr            = generic_removexattr,
7721         .listxattr              = xfs_vn_listxattr,
7722         .fiemap                 = xfs_vn_fiemap,
7723 +       .sync_flags             = xfs_sync_flags,
7724  };
7725
7726  static const struct inode_operations xfs_dir_inode_operations = {
7727 @@ -624,6 +628,7 @@ static const struct inode_operations xfs
7728         .getxattr               = generic_getxattr,
7729         .removexattr            = generic_removexattr,
7730         .listxattr              = xfs_vn_listxattr,
7731 +       .sync_flags             = xfs_sync_flags,
7732  };
7733
7734  static const struct inode_operations xfs_dir_ci_inode_operations = {
7735 @@ -673,6 +678,10 @@ xfs_diflags_to_iflags(
7736                 inode->i_flags |= S_IMMUTABLE;
7737         else
7738                 inode->i_flags &= ~S_IMMUTABLE;
7739 +       if (ip->i_d.di_flags & XFS_DIFLAG_IXUNLINK)
7740 +               inode->i_flags |= S_IXUNLINK;
7741 +       else
7742 +               inode->i_flags &= ~S_IXUNLINK;
7743         if (ip->i_d.di_flags & XFS_DIFLAG_APPEND)
7744                 inode->i_flags |= S_APPEND;
7745         else
7746 @@ -685,6 +694,15 @@ xfs_diflags_to_iflags(
7747                 inode->i_flags |= S_NOATIME;
7748         else
7749                 inode->i_flags &= ~S_NOATIME;
7750 +
7751 +       if (ip->i_d.di_vflags & XFS_DIVFLAG_BARRIER)
7752 +               inode->i_vflags |= V_BARRIER;
7753 +       else
7754 +               inode->i_vflags &= ~V_BARRIER;
7755 +       if (ip->i_d.di_vflags & XFS_DIVFLAG_COW)
7756 +               inode->i_vflags |= V_COW;
7757 +       else
7758 +               inode->i_vflags &= ~V_COW;
7759  }
7760
7761  /*
7762 @@ -716,6 +734,7 @@ xfs_setup_inode(
7763         inode->i_nlink  = ip->i_d.di_nlink;
7764         inode->i_uid    = ip->i_d.di_uid;
7765         inode->i_gid    = ip->i_d.di_gid;
7766 +       inode->i_tag    = ip->i_d.di_tag;
7767
7768         switch (inode->i_mode & S_IFMT) {
7769         case S_IFBLK:
7770 diff -NurpP --minimal linux-3.0-rc5/fs/xfs/linux-2.6/xfs_linux.h linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/linux-2.6/xfs_linux.h
7771 --- linux-3.0-rc5/fs/xfs/linux-2.6/xfs_linux.h  2011-07-01 11:14:20.000000000 +0200
7772 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/linux-2.6/xfs_linux.h     2011-06-10 22:11:24.000000000 +0200
7773 @@ -115,6 +115,7 @@
7774
7775  #define current_cpu()          (raw_smp_processor_id())
7776  #define current_pid()          (current->pid)
7777 +#define current_fstag(vp)      (dx_current_fstag((vp)->i_sb))
7778  #define current_test_flags(f)  (current->flags & (f))
7779  #define current_set_flags_nested(sp, f)                \
7780                 (*(sp) = current->flags, current->flags |= (f))
7781 diff -NurpP --minimal linux-3.0-rc5/fs/xfs/linux-2.6/xfs_super.c linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/linux-2.6/xfs_super.c
7782 --- linux-3.0-rc5/fs/xfs/linux-2.6/xfs_super.c  2011-07-01 11:14:20.000000000 +0200
7783 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/linux-2.6/xfs_super.c     2011-06-22 12:39:15.000000000 +0200
7784 @@ -114,6 +114,9 @@ mempool_t *xfs_ioend_pool;
7785  #define MNTOPT_NODELAYLOG  "nodelaylog"        /* Delayed logging disabled */
7786  #define MNTOPT_DISCARD    "discard"    /* Discard unused blocks */
7787  #define MNTOPT_NODISCARD   "nodiscard" /* Do not discard unused blocks */
7788 +#define MNTOPT_TAGXID  "tagxid"        /* context tagging for inodes */
7789 +#define MNTOPT_TAGGED  "tag"           /* context tagging for inodes */
7790 +#define MNTOPT_NOTAGTAG        "notag"         /* do not use context tagging */
7791
7792  /*
7793   * Table driven mount option parser.
7794 @@ -122,10 +125,14 @@ mempool_t *xfs_ioend_pool;
7795   * in the future, too.
7796   */
7797  enum {
7798 +       Opt_tag, Opt_notag,
7799         Opt_barrier, Opt_nobarrier, Opt_err
7800  };
7801
7802  static const match_table_t tokens = {
7803 +       {Opt_tag, "tagxid"},
7804 +       {Opt_tag, "tag"},
7805 +       {Opt_notag, "notag"},
7806         {Opt_barrier, "barrier"},
7807         {Opt_nobarrier, "nobarrier"},
7808         {Opt_err, NULL}
7809 @@ -373,6 +380,19 @@ xfs_parseargs(
7810                 } else if (!strcmp(this_char, "irixsgid")) {
7811                         xfs_warn(mp,
7812         "irixsgid is now a sysctl(2) variable, option is deprecated.");
7813 +#ifndef CONFIG_TAGGING_NONE
7814 +               } else if (!strcmp(this_char, MNTOPT_TAGGED)) {
7815 +                       mp->m_flags |= XFS_MOUNT_TAGGED;
7816 +               } else if (!strcmp(this_char, MNTOPT_NOTAGTAG)) {
7817 +                       mp->m_flags &= ~XFS_MOUNT_TAGGED;
7818 +               } else if (!strcmp(this_char, MNTOPT_TAGXID)) {
7819 +                       mp->m_flags |= XFS_MOUNT_TAGGED;
7820 +#endif
7821 +#ifdef CONFIG_PROPAGATE
7822 +               } else if (!strcmp(this_char, MNTOPT_TAGGED)) {
7823 +                       /* use value */
7824 +                       mp->m_flags |= XFS_MOUNT_TAGGED;
7825 +#endif
7826                 } else {
7827                         xfs_warn(mp, "unknown mount option [%s].", this_char);
7828                         return EINVAL;
7829 @@ -1182,6 +1202,16 @@ xfs_fs_remount(
7830                 case Opt_nobarrier:
7831                         mp->m_flags &= ~XFS_MOUNT_BARRIER;
7832                         break;
7833 +               case Opt_tag:
7834 +                       if (!(sb->s_flags & MS_TAGGED)) {
7835 +                               printk(KERN_INFO
7836 +                                       "XFS: %s: tagging not permitted on remount.\n",
7837 +                                       sb->s_id);
7838 +                               return -EINVAL;
7839 +                       }
7840 +                       break;
7841 +               case Opt_notag:
7842 +                       break;
7843                 default:
7844                         /*
7845                          * Logically we would return an error here to prevent
7846 @@ -1397,6 +1427,9 @@ xfs_fs_fill_super(
7847         if (error)
7848                 goto out_free_sb;
7849
7850 +       if (mp->m_flags & XFS_MOUNT_TAGGED)
7851 +               sb->s_flags |= MS_TAGGED;
7852 +
7853         /*
7854          * we must configure the block size in the superblock before we run the
7855          * full mount process as the mount process can lookup and cache inodes.
7856 diff -NurpP --minimal linux-3.0-rc5/fs/xfs/xfs_dinode.h linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/xfs_dinode.h
7857 --- linux-3.0-rc5/fs/xfs/xfs_dinode.h   2011-01-05 21:50:28.000000000 +0100
7858 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/xfs_dinode.h      2011-06-10 22:11:24.000000000 +0200
7859 @@ -51,7 +51,9 @@ typedef struct xfs_dinode {
7860         __be32          di_nlink;       /* number of links to file */
7861         __be16          di_projid_lo;   /* lower part of owner's project id */
7862         __be16          di_projid_hi;   /* higher part owner's project id */
7863 -       __u8            di_pad[6];      /* unused, zeroed space */
7864 +       __u8            di_pad[2];      /* unused, zeroed space */
7865 +       __be16          di_tag;         /* context tagging */
7866 +       __be16          di_vflags;      /* vserver specific flags */
7867         __be16          di_flushiter;   /* incremented on flush */
7868         xfs_timestamp_t di_atime;       /* time last accessed */
7869         xfs_timestamp_t di_mtime;       /* time last modified */
7870 @@ -184,6 +186,8 @@ static inline void xfs_dinode_put_rdev(s
7871  #define XFS_DIFLAG_EXTSZINHERIT_BIT 12 /* inherit inode extent size */
7872  #define XFS_DIFLAG_NODEFRAG_BIT     13 /* do not reorganize/defragment */
7873  #define XFS_DIFLAG_FILESTREAM_BIT   14  /* use filestream allocator */
7874 +#define XFS_DIFLAG_IXUNLINK_BIT     15 /* Immutable inver on unlink */
7875 +
7876  #define XFS_DIFLAG_REALTIME      (1 << XFS_DIFLAG_REALTIME_BIT)
7877  #define XFS_DIFLAG_PREALLOC      (1 << XFS_DIFLAG_PREALLOC_BIT)
7878  #define XFS_DIFLAG_NEWRTBM       (1 << XFS_DIFLAG_NEWRTBM_BIT)
7879 @@ -199,6 +203,7 @@ static inline void xfs_dinode_put_rdev(s
7880  #define XFS_DIFLAG_EXTSZINHERIT  (1 << XFS_DIFLAG_EXTSZINHERIT_BIT)
7881  #define XFS_DIFLAG_NODEFRAG      (1 << XFS_DIFLAG_NODEFRAG_BIT)
7882  #define XFS_DIFLAG_FILESTREAM    (1 << XFS_DIFLAG_FILESTREAM_BIT)
7883 +#define XFS_DIFLAG_IXUNLINK      (1 << XFS_DIFLAG_IXUNLINK_BIT)
7884
7885  #ifdef CONFIG_XFS_RT
7886  #define XFS_IS_REALTIME_INODE(ip) ((ip)->i_d.di_flags & XFS_DIFLAG_REALTIME)
7887 @@ -211,6 +216,10 @@ static inline void xfs_dinode_put_rdev(s
7888          XFS_DIFLAG_IMMUTABLE | XFS_DIFLAG_APPEND | XFS_DIFLAG_SYNC | \
7889          XFS_DIFLAG_NOATIME | XFS_DIFLAG_NODUMP | XFS_DIFLAG_RTINHERIT | \
7890          XFS_DIFLAG_PROJINHERIT | XFS_DIFLAG_NOSYMLINKS | XFS_DIFLAG_EXTSIZE | \
7891 -        XFS_DIFLAG_EXTSZINHERIT | XFS_DIFLAG_NODEFRAG | XFS_DIFLAG_FILESTREAM)
7892 +        XFS_DIFLAG_EXTSZINHERIT | XFS_DIFLAG_NODEFRAG | XFS_DIFLAG_FILESTREAM | \
7893 +        XFS_DIFLAG_IXUNLINK)
7894 +
7895 +#define XFS_DIVFLAG_BARRIER    0x01
7896 +#define XFS_DIVFLAG_COW                0x02
7897
7898  #endif /* __XFS_DINODE_H__ */
7899 diff -NurpP --minimal linux-3.0-rc5/fs/xfs/xfs_fs.h linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/xfs_fs.h
7900 --- linux-3.0-rc5/fs/xfs/xfs_fs.h       2011-01-05 21:50:28.000000000 +0100
7901 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/xfs_fs.h  2011-06-10 22:11:24.000000000 +0200
7902 @@ -67,6 +67,9 @@ struct fsxattr {
7903  #define XFS_XFLAG_EXTSZINHERIT 0x00001000      /* inherit inode extent size */
7904  #define XFS_XFLAG_NODEFRAG     0x00002000      /* do not defragment */
7905  #define XFS_XFLAG_FILESTREAM   0x00004000      /* use filestream allocator */
7906 +#define XFS_XFLAG_IXUNLINK     0x00008000      /* immutable invert on unlink */
7907 +#define XFS_XFLAG_BARRIER      0x10000000      /* chroot() barrier */
7908 +#define XFS_XFLAG_COW          0x20000000      /* copy on write mark */
7909  #define XFS_XFLAG_HASATTR      0x80000000      /* no DIFLAG for this   */
7910
7911  /*
7912 @@ -297,7 +300,8 @@ typedef struct xfs_bstat {
7913  #define        bs_projid       bs_projid_lo    /* (previously just bs_projid)  */
7914         __u16           bs_forkoff;     /* inode fork offset in bytes   */
7915         __u16           bs_projid_hi;   /* higher part of project id    */
7916 -       unsigned char   bs_pad[10];     /* pad space, unused            */
7917 +       unsigned char   bs_pad[8];      /* pad space, unused            */
7918 +       __u16           bs_tag;         /* context tagging              */
7919         __u32           bs_dmevmask;    /* DMIG event mask              */
7920         __u16           bs_dmstate;     /* DMIG state info              */
7921         __u16           bs_aextents;    /* attribute number of extents  */
7922 diff -NurpP --minimal linux-3.0-rc5/fs/xfs/xfs_ialloc.c linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/xfs_ialloc.c
7923 --- linux-3.0-rc5/fs/xfs/xfs_ialloc.c   2011-05-22 16:17:54.000000000 +0200
7924 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/xfs_ialloc.c      2011-06-10 22:11:24.000000000 +0200
7925 @@ -37,7 +37,6 @@
7926  #include "xfs_error.h"
7927  #include "xfs_bmap.h"
7928
7929 -
7930  /*
7931   * Allocation group level functions.
7932   */
7933 diff -NurpP --minimal linux-3.0-rc5/fs/xfs/xfs_inode.c linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/xfs_inode.c
7934 --- linux-3.0-rc5/fs/xfs/xfs_inode.c    2011-07-01 11:14:20.000000000 +0200
7935 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/xfs_inode.c       2011-06-10 22:11:24.000000000 +0200
7936 @@ -243,6 +243,7 @@ xfs_inotobp(
7937         return 0;
7938  }
7939
7940 +#include <linux/vs_tag.h>
7941
7942  /*
7943   * This routine is called to map an inode to the buffer containing
7944 @@ -641,15 +642,25 @@ xfs_iformat_btree(
7945  STATIC void
7946  xfs_dinode_from_disk(
7947         xfs_icdinode_t          *to,
7948 -       xfs_dinode_t            *from)
7949 +       xfs_dinode_t            *from,
7950 +       int tagged)
7951  {
7952 +       uint32_t uid, gid, tag;
7953 +
7954         to->di_magic = be16_to_cpu(from->di_magic);
7955         to->di_mode = be16_to_cpu(from->di_mode);
7956         to->di_version = from ->di_version;
7957         to->di_format = from->di_format;
7958         to->di_onlink = be16_to_cpu(from->di_onlink);
7959 -       to->di_uid = be32_to_cpu(from->di_uid);
7960 -       to->di_gid = be32_to_cpu(from->di_gid);
7961 +
7962 +       uid = be32_to_cpu(from->di_uid);
7963 +       gid = be32_to_cpu(from->di_gid);
7964 +       tag = be16_to_cpu(from->di_tag);
7965 +
7966 +       to->di_uid = INOTAG_UID(tagged, uid, gid);
7967 +       to->di_gid = INOTAG_GID(tagged, uid, gid);
7968 +       to->di_tag = INOTAG_TAG(tagged, uid, gid, tag);
7969 +
7970         to->di_nlink = be32_to_cpu(from->di_nlink);
7971         to->di_projid_lo = be16_to_cpu(from->di_projid_lo);
7972         to->di_projid_hi = be16_to_cpu(from->di_projid_hi);
7973 @@ -671,21 +682,26 @@ xfs_dinode_from_disk(
7974         to->di_dmevmask = be32_to_cpu(from->di_dmevmask);
7975         to->di_dmstate  = be16_to_cpu(from->di_dmstate);
7976         to->di_flags    = be16_to_cpu(from->di_flags);
7977 +       to->di_vflags   = be16_to_cpu(from->di_vflags);
7978         to->di_gen      = be32_to_cpu(from->di_gen);
7979  }
7980
7981  void
7982  xfs_dinode_to_disk(
7983         xfs_dinode_t            *to,
7984 -       xfs_icdinode_t          *from)
7985 +       xfs_icdinode_t          *from,
7986 +       int tagged)
7987  {
7988         to->di_magic = cpu_to_be16(from->di_magic);
7989         to->di_mode = cpu_to_be16(from->di_mode);
7990         to->di_version = from ->di_version;
7991         to->di_format = from->di_format;
7992         to->di_onlink = cpu_to_be16(from->di_onlink);
7993 -       to->di_uid = cpu_to_be32(from->di_uid);
7994 -       to->di_gid = cpu_to_be32(from->di_gid);
7995 +
7996 +       to->di_uid = cpu_to_be32(TAGINO_UID(tagged, from->di_uid, from->di_tag));
7997 +       to->di_gid = cpu_to_be32(TAGINO_GID(tagged, from->di_gid, from->di_tag));
7998 +       to->di_tag = cpu_to_be16(TAGINO_TAG(tagged, from->di_tag));
7999 +
8000         to->di_nlink = cpu_to_be32(from->di_nlink);
8001         to->di_projid_lo = cpu_to_be16(from->di_projid_lo);
8002         to->di_projid_hi = cpu_to_be16(from->di_projid_hi);
8003 @@ -707,12 +723,14 @@ xfs_dinode_to_disk(
8004         to->di_dmevmask = cpu_to_be32(from->di_dmevmask);
8005         to->di_dmstate = cpu_to_be16(from->di_dmstate);
8006         to->di_flags = cpu_to_be16(from->di_flags);
8007 +       to->di_vflags = cpu_to_be16(from->di_vflags);
8008         to->di_gen = cpu_to_be32(from->di_gen);
8009  }
8010
8011  STATIC uint
8012  _xfs_dic2xflags(
8013 -       __uint16_t              di_flags)
8014 +       __uint16_t              di_flags,
8015 +       __uint16_t              di_vflags)
8016  {
8017         uint                    flags = 0;
8018
8019 @@ -723,6 +741,8 @@ _xfs_dic2xflags(
8020                         flags |= XFS_XFLAG_PREALLOC;
8021                 if (di_flags & XFS_DIFLAG_IMMUTABLE)
8022                         flags |= XFS_XFLAG_IMMUTABLE;
8023 +               if (di_flags & XFS_DIFLAG_IXUNLINK)
8024 +                       flags |= XFS_XFLAG_IXUNLINK;
8025                 if (di_flags & XFS_DIFLAG_APPEND)
8026                         flags |= XFS_XFLAG_APPEND;
8027                 if (di_flags & XFS_DIFLAG_SYNC)
8028 @@ -747,6 +767,10 @@ _xfs_dic2xflags(
8029                         flags |= XFS_XFLAG_FILESTREAM;
8030         }
8031
8032 +       if (di_vflags & XFS_DIVFLAG_BARRIER)
8033 +               flags |= FS_BARRIER_FL;
8034 +       if (di_vflags & XFS_DIVFLAG_COW)
8035 +               flags |= FS_COW_FL;
8036         return flags;
8037  }
8038
8039 @@ -756,7 +780,7 @@ xfs_ip2xflags(
8040  {
8041         xfs_icdinode_t          *dic = &ip->i_d;
8042
8043 -       return _xfs_dic2xflags(dic->di_flags) |
8044 +       return _xfs_dic2xflags(dic->di_flags, dic->di_vflags) |
8045                                 (XFS_IFORK_Q(ip) ? XFS_XFLAG_HASATTR : 0);
8046  }
8047
8048 @@ -764,7 +788,8 @@ uint
8049  xfs_dic2xflags(
8050         xfs_dinode_t            *dip)
8051  {
8052 -       return _xfs_dic2xflags(be16_to_cpu(dip->di_flags)) |
8053 +       return _xfs_dic2xflags(be16_to_cpu(dip->di_flags),
8054 +                               be16_to_cpu(dip->di_vflags)) |
8055                                 (XFS_DFORK_Q(dip) ? XFS_XFLAG_HASATTR : 0);
8056  }
8057
8058 @@ -797,7 +822,6 @@ xfs_iread(
8059         if (error)
8060                 return error;
8061         dip = (xfs_dinode_t *)xfs_buf_offset(bp, ip->i_imap.im_boffset);
8062 -
8063         /*
8064          * If we got something that isn't an inode it means someone
8065          * (nfs or dmi) has a stale handle.
8066 @@ -820,7 +844,8 @@ xfs_iread(
8067          * Otherwise, just get the truly permanent information.
8068          */
8069         if (dip->di_mode) {
8070 -               xfs_dinode_from_disk(&ip->i_d, dip);
8071 +               xfs_dinode_from_disk(&ip->i_d, dip,
8072 +                       mp->m_flags & XFS_MOUNT_TAGGED);
8073                 error = xfs_iformat(ip, dip);
8074                 if (error)  {
8075  #ifdef DEBUG
8076 @@ -1015,6 +1040,7 @@ xfs_ialloc(
8077         ASSERT(ip->i_d.di_nlink == nlink);
8078         ip->i_d.di_uid = current_fsuid();
8079         ip->i_d.di_gid = current_fsgid();
8080 +       ip->i_d.di_tag = current_fstag(&ip->i_vnode);
8081         xfs_set_projid(ip, prid);
8082         memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad));
8083
8084 @@ -1075,6 +1101,7 @@ xfs_ialloc(
8085         ip->i_d.di_dmevmask = 0;
8086         ip->i_d.di_dmstate = 0;
8087         ip->i_d.di_flags = 0;
8088 +       ip->i_d.di_vflags = 0;
8089         flags = XFS_ILOG_CORE;
8090         switch (mode & S_IFMT) {
8091         case S_IFIFO:
8092 @@ -2108,6 +2135,7 @@ xfs_ifree(
8093         }
8094         ip->i_d.di_mode = 0;            /* mark incore inode as free */
8095         ip->i_d.di_flags = 0;
8096 +       ip->i_d.di_vflags = 0;
8097         ip->i_d.di_dmevmask = 0;
8098         ip->i_d.di_forkoff = 0;         /* mark the attr fork not in use */
8099         ip->i_df.if_ext_max =
8100 @@ -2987,7 +3015,8 @@ xfs_iflush_int(
8101          * because if the inode is dirty at all the core must
8102          * be.
8103          */
8104 -       xfs_dinode_to_disk(dip, &ip->i_d);
8105 +       xfs_dinode_to_disk(dip, &ip->i_d,
8106 +               mp->m_flags & XFS_MOUNT_TAGGED);
8107
8108         /* Wrap, we never let the log put out DI_MAX_FLUSH */
8109         if (ip->i_d.di_flushiter == DI_MAX_FLUSH)
8110 diff -NurpP --minimal linux-3.0-rc5/fs/xfs/xfs_inode.h linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/xfs_inode.h
8111 --- linux-3.0-rc5/fs/xfs/xfs_inode.h    2011-07-01 11:14:20.000000000 +0200
8112 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/xfs_inode.h       2011-07-01 11:35:35.000000000 +0200
8113 @@ -135,7 +135,9 @@ typedef struct xfs_icdinode {
8114         __uint32_t      di_nlink;       /* number of links to file */
8115         __uint16_t      di_projid_lo;   /* lower part of owner's project id */
8116         __uint16_t      di_projid_hi;   /* higher part of owner's project id */
8117 -       __uint8_t       di_pad[6];      /* unused, zeroed space */
8118 +       __uint8_t       di_pad[2];      /* unused, zeroed space */
8119 +       __uint16_t      di_tag;         /* context tagging */
8120 +       __uint16_t      di_vflags;      /* vserver specific flags */
8121         __uint16_t      di_flushiter;   /* incremented on flush */
8122         xfs_ictimestamp_t di_atime;     /* time last accessed */
8123         xfs_ictimestamp_t di_mtime;     /* time last modified */
8124 @@ -546,7 +548,7 @@ int         xfs_itobp(struct xfs_mount *, struc
8125  int            xfs_iread(struct xfs_mount *, struct xfs_trans *,
8126                           struct xfs_inode *, uint);
8127  void           xfs_dinode_to_disk(struct xfs_dinode *,
8128 -                                  struct xfs_icdinode *);
8129 +                                  struct xfs_icdinode *, int);
8130  void           xfs_idestroy_fork(struct xfs_inode *, int);
8131  void           xfs_idata_realloc(struct xfs_inode *, int, int);
8132  void           xfs_iroot_realloc(struct xfs_inode *, int, int);
8133 diff -NurpP --minimal linux-3.0-rc5/fs/xfs/xfs_itable.c linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/xfs_itable.c
8134 --- linux-3.0-rc5/fs/xfs/xfs_itable.c   2011-05-22 16:17:54.000000000 +0200
8135 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/xfs_itable.c      2011-06-10 22:11:24.000000000 +0200
8136 @@ -98,6 +98,7 @@ xfs_bulkstat_one_int(
8137         buf->bs_mode = dic->di_mode;
8138         buf->bs_uid = dic->di_uid;
8139         buf->bs_gid = dic->di_gid;
8140 +       buf->bs_tag = dic->di_tag;
8141         buf->bs_size = dic->di_size;
8142
8143         /*
8144 diff -NurpP --minimal linux-3.0-rc5/fs/xfs/xfs_log_recover.c linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/xfs_log_recover.c
8145 --- linux-3.0-rc5/fs/xfs/xfs_log_recover.c      2011-07-01 11:14:20.000000000 +0200
8146 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/xfs_log_recover.c 2011-06-10 22:11:24.000000000 +0200
8147 @@ -2343,7 +2343,8 @@ xlog_recover_inode_pass2(
8148         }
8149
8150         /* The core is in in-core format */
8151 -       xfs_dinode_to_disk(dip, item->ri_buf[1].i_addr);
8152 +       xfs_dinode_to_disk(dip, item->ri_buf[1].i_addr,
8153 +               mp->m_flags & XFS_MOUNT_TAGGED);
8154
8155         /* the rest is in on-disk format */
8156         if (item->ri_buf[1].i_len > sizeof(struct xfs_icdinode)) {
8157 diff -NurpP --minimal linux-3.0-rc5/fs/xfs/xfs_mount.h linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/xfs_mount.h
8158 --- linux-3.0-rc5/fs/xfs/xfs_mount.h    2011-07-01 11:14:20.000000000 +0200
8159 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/xfs_mount.h       2011-06-10 22:11:24.000000000 +0200
8160 @@ -249,6 +249,7 @@ typedef struct xfs_mount {
8161                                                    allocator */
8162  #define XFS_MOUNT_NOATTR2      (1ULL << 25)    /* disable use of attr2 format */
8163
8164 +#define XFS_MOUNT_TAGGED       (1ULL << 31)    /* context tagging */
8165
8166  /*
8167   * Default minimum read and write sizes.
8168 diff -NurpP --minimal linux-3.0-rc5/fs/xfs/xfs_vnodeops.c linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/xfs_vnodeops.c
8169 --- linux-3.0-rc5/fs/xfs/xfs_vnodeops.c 2011-07-01 11:14:20.000000000 +0200
8170 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/xfs_vnodeops.c    2011-07-01 11:35:35.000000000 +0200
8171 @@ -50,6 +50,78 @@
8172  #include "xfs_vnodeops.h"
8173  #include "xfs_trace.h"
8174
8175 +
8176 +STATIC void
8177 +xfs_get_inode_flags(
8178 +       xfs_inode_t     *ip)
8179 +{
8180 +       struct inode    *inode = VFS_I(ip);
8181 +       unsigned int    flags = inode->i_flags;
8182 +       unsigned int    vflags = inode->i_vflags;
8183 +
8184 +       if (flags & S_IMMUTABLE)
8185 +               ip->i_d.di_flags |= XFS_DIFLAG_IMMUTABLE;
8186 +       else
8187 +               ip->i_d.di_flags &= ~XFS_DIFLAG_IMMUTABLE;
8188 +       if (flags & S_IXUNLINK)
8189 +               ip->i_d.di_flags |= XFS_DIFLAG_IXUNLINK;
8190 +       else
8191 +               ip->i_d.di_flags &= ~XFS_DIFLAG_IXUNLINK;
8192 +
8193 +       if (vflags & V_BARRIER)
8194 +               ip->i_d.di_vflags |= XFS_DIVFLAG_BARRIER;
8195 +       else
8196 +               ip->i_d.di_vflags &= ~XFS_DIVFLAG_BARRIER;
8197 +       if (vflags & V_COW)
8198 +               ip->i_d.di_vflags |= XFS_DIVFLAG_COW;
8199 +       else
8200 +               ip->i_d.di_vflags &= ~XFS_DIVFLAG_COW;
8201 +}
8202 +
8203 +int
8204 +xfs_sync_flags(
8205 +       struct inode            *inode,
8206 +       int                     flags,
8207 +       int                     vflags)
8208 +{
8209 +       struct xfs_inode        *ip = XFS_I(inode);
8210 +       struct xfs_mount        *mp = ip->i_mount;
8211 +       struct xfs_trans        *tp;
8212 +       unsigned int            lock_flags = 0;
8213 +       int                     code;
8214 +
8215 +       tp = xfs_trans_alloc(mp, XFS_TRANS_SETATTR_NOT_SIZE);
8216 +       code = xfs_trans_reserve(tp, 0, XFS_ICHANGE_LOG_RES(mp), 0, 0, 0);
8217 +       if (code)
8218 +               goto error_out;
8219 +
8220 +       xfs_ilock(ip, XFS_ILOCK_EXCL);
8221 +
8222 +       xfs_trans_ijoin(tp, ip);
8223 +
8224 +       inode->i_flags = flags;
8225 +       inode->i_vflags = vflags;
8226 +       xfs_get_inode_flags(ip);
8227 +
8228 +       xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
8229 +       xfs_trans_ichgtime(tp, ip, XFS_ICHGTIME_CHG);
8230 +
8231 +       XFS_STATS_INC(xs_ig_attrchg);
8232 +
8233 +       if (mp->m_flags & XFS_MOUNT_WSYNC)
8234 +               xfs_trans_set_sync(tp);
8235 +       code = xfs_trans_commit(tp, 0);
8236 +       xfs_iunlock(ip, XFS_ILOCK_EXCL);
8237 +       return code;
8238 +
8239 +error_out:
8240 +       xfs_trans_cancel(tp, 0);
8241 +       if (lock_flags)
8242 +               xfs_iunlock(ip, XFS_ILOCK_EXCL);
8243 +       return code;
8244 +}
8245 +
8246 +
8247  int
8248  xfs_setattr(
8249         struct xfs_inode        *ip,
8250 @@ -65,6 +137,7 @@ xfs_setattr(
8251         uint                    commit_flags=0;
8252         uid_t                   uid=0, iuid=0;
8253         gid_t                   gid=0, igid=0;
8254 +       tag_t                   tag=0, itag=0;
8255         struct xfs_dquot        *udqp, *gdqp, *olddquot1, *olddquot2;
8256         int                     need_iolock = 1;
8257
8258 @@ -147,7 +220,7 @@ xfs_setattr(
8259         /*
8260          * Change file ownership.  Must be the owner or privileged.
8261          */
8262 -       if (mask & (ATTR_UID|ATTR_GID)) {
8263 +       if (mask & (ATTR_UID|ATTR_GID|ATTR_TAG)) {
8264                 /*
8265                  * These IDs could have changed since we last looked at them.
8266                  * But, we're assured that if the ownership did change
8267 @@ -156,8 +229,10 @@ xfs_setattr(
8268                  */
8269                 iuid = ip->i_d.di_uid;
8270                 igid = ip->i_d.di_gid;
8271 +               itag = ip->i_d.di_tag;
8272                 gid = (mask & ATTR_GID) ? iattr->ia_gid : igid;
8273                 uid = (mask & ATTR_UID) ? iattr->ia_uid : iuid;
8274 +               tag = (mask & ATTR_TAG) ? iattr->ia_tag : itag;
8275
8276                 /*
8277                  * Do a quota reservation only if uid/gid is actually
8278 @@ -165,7 +240,8 @@ xfs_setattr(
8279                  */
8280                 if (XFS_IS_QUOTA_RUNNING(mp) &&
8281                     ((XFS_IS_UQUOTA_ON(mp) && iuid != uid) ||
8282 -                    (XFS_IS_GQUOTA_ON(mp) && igid != gid))) {
8283 +                    (XFS_IS_GQUOTA_ON(mp) && igid != gid) ||
8284 +                    (XFS_IS_GQUOTA_ON(mp) && itag != tag))) {
8285                         ASSERT(tp);
8286                         code = xfs_qm_vop_chown_reserve(tp, ip, udqp, gdqp,
8287                                                 capable(CAP_FOWNER) ?
8288 @@ -329,7 +405,7 @@ xfs_setattr(
8289         /*
8290          * Change file ownership.  Must be the owner or privileged.
8291          */
8292 -       if (mask & (ATTR_UID|ATTR_GID)) {
8293 +       if (mask & (ATTR_UID|ATTR_GID|ATTR_TAG)) {
8294                 /*
8295                  * CAP_FSETID overrides the following restrictions:
8296                  *
8297 @@ -345,6 +421,10 @@ xfs_setattr(
8298                  * Change the ownerships and register quota modifications
8299                  * in the transaction.
8300                  */
8301 +               if (itag != tag) {
8302 +                       ip->i_d.di_tag = tag;
8303 +                       inode->i_tag = tag;
8304 +               }
8305                 if (iuid != uid) {
8306                         if (XFS_IS_QUOTA_RUNNING(mp) && XFS_IS_UQUOTA_ON(mp)) {
8307                                 ASSERT(mask & ATTR_UID);
8308 diff -NurpP --minimal linux-3.0-rc5/fs/xfs/xfs_vnodeops.h linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/xfs_vnodeops.h
8309 --- linux-3.0-rc5/fs/xfs/xfs_vnodeops.h 2011-05-22 16:17:54.000000000 +0200
8310 +++ linux-3.0-rc5-vs2.3.1-pre3/fs/xfs/xfs_vnodeops.h    2011-06-10 22:11:24.000000000 +0200
8311 @@ -13,6 +13,7 @@ struct xfs_inode;
8312  struct xfs_iomap;
8313
8314
8315 +int xfs_sync_xflags(struct xfs_inode *ip);
8316  int xfs_setattr(struct xfs_inode *ip, struct iattr *vap, int flags);
8317  #define        XFS_ATTR_DMI            0x01    /* invocation from a DMI function */
8318  #define        XFS_ATTR_NONBLOCK       0x02    /* return EAGAIN if operation would block */
8319 diff -NurpP --minimal linux-3.0-rc5/include/asm-generic/tlb.h linux-3.0-rc5-vs2.3.1-pre3/include/asm-generic/tlb.h
8320 --- linux-3.0-rc5/include/asm-generic/tlb.h     2011-07-01 11:14:20.000000000 +0200
8321 +++ linux-3.0-rc5-vs2.3.1-pre3/include/asm-generic/tlb.h        2011-06-10 22:11:24.000000000 +0200
8322 @@ -16,6 +16,7 @@
8323  #define _ASM_GENERIC__TLB_H
8324
8325  #include <linux/swap.h>
8326 +#include <linux/vs_memory.h>
8327  #include <asm/pgalloc.h>
8328  #include <asm/tlbflush.h>
8329
8330 diff -NurpP --minimal linux-3.0-rc5/include/linux/Kbuild linux-3.0-rc5-vs2.3.1-pre3/include/linux/Kbuild
8331 --- linux-3.0-rc5/include/linux/Kbuild  2011-07-01 11:14:20.000000000 +0200
8332 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/Kbuild     2011-06-10 22:11:24.000000000 +0200
8333 @@ -17,6 +17,7 @@ header-y += netfilter_bridge/
8334  header-y += netfilter_ipv4/
8335  header-y += netfilter_ipv6/
8336  header-y += usb/
8337 +header-y += vserver/
8338  header-y += wimax/
8339
8340  objhdr-y += version.h
8341 diff -NurpP --minimal linux-3.0-rc5/include/linux/capability.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/capability.h
8342 --- linux-3.0-rc5/include/linux/capability.h    2011-07-01 11:14:21.000000000 +0200
8343 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/capability.h       2011-06-10 22:11:24.000000000 +0200
8344 @@ -279,6 +279,7 @@ struct cpu_vfs_cap_data {
8345     arbitrary SCSI commands */
8346  /* Allow setting encryption key on loopback filesystem */
8347  /* Allow setting zone reclaim policy */
8348 +/* Allow the selection of a security context */
8349
8350  #define CAP_SYS_ADMIN        21
8351
8352 @@ -362,7 +363,12 @@ struct cpu_vfs_cap_data {
8353
8354  #define CAP_LAST_CAP         CAP_WAKE_ALARM
8355
8356 -#define cap_valid(x) ((x) >= 0 && (x) <= CAP_LAST_CAP)
8357 +/* Allow context manipulations */
8358 +/* Allow changing context info on files */
8359 +
8360 +#define CAP_CONTEXT         63
8361 +
8362 +#define cap_valid(x) ((x) >= 0 && ((x) <= CAP_LAST_CAP || (x) == CAP_CONTEXT))
8363
8364  /*
8365   * Bit location of each capability (used by user-space library and kernel)
8366 diff -NurpP --minimal linux-3.0-rc5/include/linux/cred.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/cred.h
8367 --- linux-3.0-rc5/include/linux/cred.h  2011-07-01 11:14:21.000000000 +0200
8368 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/cred.h     2011-06-10 22:11:24.000000000 +0200
8369 @@ -156,6 +156,7 @@ extern void exit_creds(struct task_struc
8370  extern int copy_creds(struct task_struct *, unsigned long);
8371  extern const struct cred *get_task_cred(struct task_struct *);
8372  extern struct cred *cred_alloc_blank(void);
8373 +extern struct cred *__prepare_creds(const struct cred *);
8374  extern struct cred *prepare_creds(void);
8375  extern struct cred *prepare_exec_creds(void);
8376  extern int commit_creds(struct cred *);
8377 @@ -209,6 +210,31 @@ static inline void validate_process_cred
8378  }
8379  #endif
8380
8381 +static inline void set_cred_subscribers(struct cred *cred, int n)
8382 +{
8383 +#ifdef CONFIG_DEBUG_CREDENTIALS
8384 +       atomic_set(&cred->subscribers, n);
8385 +#endif
8386 +}
8387 +
8388 +static inline int read_cred_subscribers(const struct cred *cred)
8389 +{
8390 +#ifdef CONFIG_DEBUG_CREDENTIALS
8391 +       return atomic_read(&cred->subscribers);
8392 +#else
8393 +       return 0;
8394 +#endif
8395 +}
8396 +
8397 +static inline void alter_cred_subscribers(const struct cred *_cred, int n)
8398 +{
8399 +#ifdef CONFIG_DEBUG_CREDENTIALS
8400 +       struct cred *cred = (struct cred *) _cred;
8401 +
8402 +       atomic_add(n, &cred->subscribers);
8403 +#endif
8404 +}
8405 +
8406  /**
8407   * get_new_cred - Get a reference on a new set of credentials
8408   * @cred: The new credentials to reference
8409 diff -NurpP --minimal linux-3.0-rc5/include/linux/devpts_fs.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/devpts_fs.h
8410 --- linux-3.0-rc5/include/linux/devpts_fs.h     2008-12-25 00:26:37.000000000 +0100
8411 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/devpts_fs.h        2011-06-10 22:11:24.000000000 +0200
8412 @@ -45,5 +45,4 @@ static inline void devpts_pty_kill(struc
8413
8414  #endif
8415
8416 -
8417  #endif /* _LINUX_DEVPTS_FS_H */
8418 diff -NurpP --minimal linux-3.0-rc5/include/linux/ext2_fs.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/ext2_fs.h
8419 --- linux-3.0-rc5/include/linux/ext2_fs.h       2010-02-25 11:52:07.000000000 +0100
8420 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/ext2_fs.h  2011-06-10 22:11:24.000000000 +0200
8421 @@ -189,8 +189,12 @@ struct ext2_group_desc
8422  #define EXT2_NOTAIL_FL                 FS_NOTAIL_FL    /* file tail should not be merged */
8423  #define EXT2_DIRSYNC_FL                        FS_DIRSYNC_FL   /* dirsync behaviour (directories only) */
8424  #define EXT2_TOPDIR_FL                 FS_TOPDIR_FL    /* Top of directory hierarchies*/
8425 +#define EXT2_IXUNLINK_FL               FS_IXUNLINK_FL  /* Immutable invert on unlink */
8426  #define EXT2_RESERVED_FL               FS_RESERVED_FL  /* reserved for ext2 lib */
8427
8428 +#define EXT2_BARRIER_FL                        FS_BARRIER_FL   /* Barrier for chroot() */
8429 +#define EXT2_COW_FL                    FS_COW_FL       /* Copy on Write marker */
8430 +
8431  #define EXT2_FL_USER_VISIBLE           FS_FL_USER_VISIBLE      /* User visible flags */
8432  #define EXT2_FL_USER_MODIFIABLE                FS_FL_USER_MODIFIABLE   /* User modifiable flags */
8433
8434 @@ -274,7 +278,8 @@ struct ext2_inode {
8435                         __u16   i_pad1;
8436                         __le16  l_i_uid_high;   /* these 2 fields    */
8437                         __le16  l_i_gid_high;   /* were reserved2[0] */
8438 -                       __u32   l_i_reserved2;
8439 +                       __le16  l_i_tag;        /* Context Tag */
8440 +                       __u16   l_i_reserved2;
8441                 } linux2;
8442                 struct {
8443                         __u8    h_i_frag;       /* Fragment number */
8444 @@ -303,6 +308,7 @@ struct ext2_inode {
8445  #define i_gid_low      i_gid
8446  #define i_uid_high     osd2.linux2.l_i_uid_high
8447  #define i_gid_high     osd2.linux2.l_i_gid_high
8448 +#define i_raw_tag      osd2.linux2.l_i_tag
8449  #define i_reserved2    osd2.linux2.l_i_reserved2
8450  #endif
8451
8452 @@ -347,6 +353,7 @@ struct ext2_inode {
8453  #define EXT2_MOUNT_USRQUOTA            0x020000  /* user quota */
8454  #define EXT2_MOUNT_GRPQUOTA            0x040000  /* group quota */
8455  #define EXT2_MOUNT_RESERVATION         0x080000  /* Preallocation */
8456 +#define EXT2_MOUNT_TAGGED              (1<<24)   /* Enable Context Tags */
8457
8458
8459  #define clear_opt(o, opt)              o &= ~EXT2_MOUNT_##opt
8460 diff -NurpP --minimal linux-3.0-rc5/include/linux/ext3_fs.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/ext3_fs.h
8461 --- linux-3.0-rc5/include/linux/ext3_fs.h       2011-07-01 11:14:21.000000000 +0200
8462 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/ext3_fs.h  2011-06-10 22:11:24.000000000 +0200
8463 @@ -173,10 +173,14 @@ struct ext3_group_desc
8464  #define EXT3_NOTAIL_FL                 0x00008000 /* file tail should not be merged */
8465  #define EXT3_DIRSYNC_FL                        0x00010000 /* dirsync behaviour (directories only) */
8466  #define EXT3_TOPDIR_FL                 0x00020000 /* Top of directory hierarchies*/
8467 +#define EXT3_IXUNLINK_FL               0x08000000 /* Immutable invert on unlink */
8468  #define EXT3_RESERVED_FL               0x80000000 /* reserved for ext3 lib */
8469
8470 -#define EXT3_FL_USER_VISIBLE           0x0003DFFF /* User visible flags */
8471 -#define EXT3_FL_USER_MODIFIABLE                0x000380FF /* User modifiable flags */
8472 +#define EXT3_BARRIER_FL                        0x04000000 /* Barrier for chroot() */
8473 +#define EXT3_COW_FL                    0x20000000 /* Copy on Write marker */
8474 +
8475 +#define EXT3_FL_USER_VISIBLE           0x0103DFFF /* User visible flags */
8476 +#define EXT3_FL_USER_MODIFIABLE                0x010380FF /* User modifiable flags */
8477
8478  /* Flags that should be inherited by new inodes from their parent. */
8479  #define EXT3_FL_INHERITED (EXT3_SECRM_FL | EXT3_UNRM_FL | EXT3_COMPR_FL |\
8480 @@ -312,7 +316,8 @@ struct ext3_inode {
8481                         __u16   i_pad1;
8482                         __le16  l_i_uid_high;   /* these 2 fields    */
8483                         __le16  l_i_gid_high;   /* were reserved2[0] */
8484 -                       __u32   l_i_reserved2;
8485 +                       __le16  l_i_tag;        /* Context Tag */
8486 +                       __u16   l_i_reserved2;
8487                 } linux2;
8488                 struct {
8489                         __u8    h_i_frag;       /* Fragment number */
8490 @@ -343,6 +348,7 @@ struct ext3_inode {
8491  #define i_gid_low      i_gid
8492  #define i_uid_high     osd2.linux2.l_i_uid_high
8493  #define i_gid_high     osd2.linux2.l_i_gid_high
8494 +#define i_raw_tag      osd2.linux2.l_i_tag
8495  #define i_reserved2    osd2.linux2.l_i_reserved2
8496
8497  #elif defined(__GNU__)
8498 @@ -405,6 +411,7 @@ struct ext3_inode {
8499  #define EXT3_MOUNT_GRPQUOTA            0x200000 /* "old" group quota */
8500  #define EXT3_MOUNT_DATA_ERR_ABORT      0x400000 /* Abort on file data write
8501                                                   * error in ordered mode */
8502 +#define EXT3_MOUNT_TAGGED              (1<<24) /* Enable Context Tags */
8503
8504  /* Compatibility, for having both ext2_fs.h and ext3_fs.h included at once */
8505  #ifndef _LINUX_EXT2_FS_H
8506 @@ -919,6 +926,7 @@ extern void ext3_get_inode_flags(struct
8507  extern void ext3_set_aops(struct inode *inode);
8508  extern int ext3_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
8509                        u64 start, u64 len);
8510 +extern int ext3_sync_flags(struct inode *, int, int);
8511
8512  /* ioctl.c */
8513  extern long ext3_ioctl(struct file *, unsigned int, unsigned long);
8514 diff -NurpP --minimal linux-3.0-rc5/include/linux/fs.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/fs.h
8515 --- linux-3.0-rc5/include/linux/fs.h    2011-07-01 11:14:21.000000000 +0200
8516 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/fs.h       2011-07-01 11:35:35.000000000 +0200
8517 @@ -208,6 +208,9 @@ struct inodes_stat_t {
8518  #define MS_KERNMOUNT   (1<<22) /* this is a kern_mount call */
8519  #define MS_I_VERSION   (1<<23) /* Update inode I_version field */
8520  #define MS_STRICTATIME (1<<24) /* Always perform atime updates */
8521 +#define MS_TAGGED      (1<<25) /* use generic inode tagging */
8522 +#define MS_TAGID       (1<<26) /* use specific tag for this mount */
8523 +#define MS_NOTAGCHECK  (1<<27) /* don't check tags */
8524  #define MS_NOSEC       (1<<28)
8525  #define MS_BORN                (1<<29)
8526  #define MS_ACTIVE      (1<<30)
8527 @@ -239,6 +242,14 @@ struct inodes_stat_t {
8528  #define S_IMA          1024    /* Inode has an associated IMA struct */
8529  #define S_AUTOMOUNT    2048    /* Automount/referral quasi-directory */
8530  #define S_NOSEC                4096    /* no suid or xattr security attributes */
8531 +#define S_IXUNLINK     16384   /* Immutable Invert on unlink */
8532 +
8533 +/* Linux-VServer related Inode flags */
8534 +
8535 +#define V_VALID                1
8536 +#define V_XATTR                2
8537 +#define V_BARRIER      4       /* Barrier for chroot() */
8538 +#define V_COW          8       /* Copy on Write */
8539
8540  /*
8541   * Note that nosuid etc flags are inode-specific: setting some file-system
8542 @@ -261,12 +272,15 @@ struct inodes_stat_t {
8543  #define IS_DIRSYNC(inode)      (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
8544                                         ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
8545  #define IS_MANDLOCK(inode)     __IS_FLG(inode, MS_MANDLOCK)
8546 -#define IS_NOATIME(inode)   __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
8547 -#define IS_I_VERSION(inode)   __IS_FLG(inode, MS_I_VERSION)
8548 +#define IS_NOATIME(inode)      __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
8549 +#define IS_I_VERSION(inode)    __IS_FLG(inode, MS_I_VERSION)
8550 +#define IS_TAGGED(inode)       __IS_FLG(inode, MS_TAGGED)
8551
8552  #define IS_NOQUOTA(inode)      ((inode)->i_flags & S_NOQUOTA)
8553  #define IS_APPEND(inode)       ((inode)->i_flags & S_APPEND)
8554  #define IS_IMMUTABLE(inode)    ((inode)->i_flags & S_IMMUTABLE)
8555 +#define IS_IXUNLINK(inode)     ((inode)->i_flags & S_IXUNLINK)
8556 +#define IS_IXORUNLINK(inode)   ((IS_IXUNLINK(inode) ? S_IMMUTABLE : 0) ^ IS_IMMUTABLE(inode))
8557  #define IS_POSIXACL(inode)     __IS_FLG(inode, MS_POSIXACL)
8558
8559  #define IS_DEADDIR(inode)      ((inode)->i_flags & S_DEAD)
8560 @@ -277,6 +291,16 @@ struct inodes_stat_t {
8561  #define IS_AUTOMOUNT(inode)    ((inode)->i_flags & S_AUTOMOUNT)
8562  #define IS_NOSEC(inode)                ((inode)->i_flags & S_NOSEC)
8563
8564 +#define IS_BARRIER(inode)      (S_ISDIR((inode)->i_mode) && ((inode)->i_vflags & V_BARRIER))
8565 +
8566 +#ifdef CONFIG_VSERVER_COWBL
8567 +#  define IS_COW(inode)                (IS_IXUNLINK(inode) && IS_IMMUTABLE(inode))
8568 +#  define IS_COW_LINK(inode)   (S_ISREG((inode)->i_mode) && ((inode)->i_nlink > 1))
8569 +#else
8570 +#  define IS_COW(inode)                (0)
8571 +#  define IS_COW_LINK(inode)   (0)
8572 +#endif
8573 +
8574  /* the read-only stuff doesn't really belong here, but any other place is
8575     probably as bad and I don't want to create yet another include file. */
8576
8577 @@ -362,11 +386,14 @@ struct inodes_stat_t {
8578  #define FS_EXTENT_FL                   0x00080000 /* Extents */
8579  #define FS_DIRECTIO_FL                 0x00100000 /* Use direct i/o */
8580  #define FS_NOCOW_FL                    0x00800000 /* Do not cow file */
8581 +#define FS_IXUNLINK_FL                 0x08000000 /* Immutable invert on unlink */
8582  #define FS_RESERVED_FL                 0x80000000 /* reserved for ext2 lib */
8583
8584 -#define FS_FL_USER_VISIBLE             0x0003DFFF /* User visible flags */
8585 -#define FS_FL_USER_MODIFIABLE          0x000380FF /* User modifiable flags */
8586 -
8587 +#define FS_BARRIER_FL                  0x04000000 /* Barrier for chroot() */
8588 +#define FS_COW_FL                      0x20000000 /* Copy on Write marker */
8589 +
8590 +#define FS_FL_USER_VISIBLE             0x0103DFFF /* User visible flags */
8591 +#define FS_FL_USER_MODIFIABLE          0x010380FF /* User modifiable flags */
8592
8593  #define SYNC_FILE_RANGE_WAIT_BEFORE    1
8594  #define SYNC_FILE_RANGE_WRITE          2
8595 @@ -447,6 +474,7 @@ typedef void (dio_iodone_t)(struct kiocb
8596  #define ATTR_KILL_PRIV (1 << 14)
8597  #define ATTR_OPEN      (1 << 15) /* Truncating from open(O_TRUNC) */
8598  #define ATTR_TIMES_SET (1 << 16)
8599 +#define ATTR_TAG       (1 << 17)
8600
8601  /*
8602   * This is the Inode Attributes structure, used for notify_change().  It
8603 @@ -462,6 +490,7 @@ struct iattr {
8604         umode_t         ia_mode;
8605         uid_t           ia_uid;
8606         gid_t           ia_gid;
8607 +       tag_t           ia_tag;
8608         loff_t          ia_size;
8609         struct timespec ia_atime;
8610         struct timespec ia_mtime;
8611 @@ -475,6 +504,9 @@ struct iattr {
8612         struct file     *ia_file;
8613  };
8614
8615 +#define ATTR_FLAG_BARRIER      512     /* Barrier for chroot() */
8616 +#define ATTR_FLAG_IXUNLINK     1024    /* Immutable invert on unlink */
8617 +
8618  /*
8619   * Includes for diskquotas.
8620   */
8621 @@ -740,11 +772,13 @@ struct inode {
8622         umode_t                 i_mode;
8623         uid_t                   i_uid;
8624         gid_t                   i_gid;
8625 +       tag_t                   i_tag;
8626         const struct inode_operations   *i_op;
8627         struct super_block      *i_sb;
8628
8629         spinlock_t              i_lock; /* i_blocks, i_bytes, maybe i_size */
8630 -       unsigned int            i_flags;
8631 +       unsigned short          i_flags;
8632 +       unsigned short          i_vflags;
8633         unsigned long           i_state;
8634  #ifdef CONFIG_SECURITY
8635         void                    *i_security;
8636 @@ -766,6 +800,7 @@ struct inode {
8637         atomic_t                i_count;
8638         unsigned int            i_nlink;
8639         dev_t                   i_rdev;
8640 +       dev_t                   i_mdev;
8641         unsigned int            i_blkbits;
8642         u64                     i_version;
8643         loff_t                  i_size;
8644 @@ -890,12 +925,12 @@ static inline void i_size_write(struct i
8645
8646  static inline unsigned iminor(const struct inode *inode)
8647  {
8648 -       return MINOR(inode->i_rdev);
8649 +       return MINOR(inode->i_mdev);
8650  }
8651
8652  static inline unsigned imajor(const struct inode *inode)
8653  {
8654 -       return MAJOR(inode->i_rdev);
8655 +       return MAJOR(inode->i_mdev);
8656  }
8657
8658  extern struct block_device *I_BDEV(struct inode *inode);
8659 @@ -957,6 +992,7 @@ struct file {
8660         loff_t                  f_pos;
8661         struct fown_struct      f_owner;
8662         const struct cred       *f_cred;
8663 +       xid_t                   f_xid;
8664         struct file_ra_state    f_ra;
8665
8666         u64                     f_version;
8667 @@ -1101,6 +1137,7 @@ struct file_lock {
8668         struct file *fl_file;
8669         loff_t fl_start;
8670         loff_t fl_end;
8671 +       xid_t fl_xid;
8672
8673         struct fasync_struct *  fl_fasync; /* for lease break notifications */
8674         unsigned long fl_break_time;    /* for nonblocking lease breaks */
8675 @@ -1600,6 +1637,7 @@ struct inode_operations {
8676         ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
8677         ssize_t (*listxattr) (struct dentry *, char *, size_t);
8678         int (*removexattr) (struct dentry *, const char *);
8679 +       int (*sync_flags) (struct inode *, int, int);
8680         void (*truncate_range)(struct inode *, loff_t, loff_t);
8681         int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start,
8682                       u64 len);
8683 @@ -1618,6 +1656,7 @@ extern ssize_t vfs_readv(struct file *,
8684                 unsigned long, loff_t *);
8685  extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
8686                 unsigned long, loff_t *);
8687 +ssize_t vfs_sendfile(struct file *, struct file *, loff_t *, size_t, loff_t);
8688
8689  struct super_operations {
8690         struct inode *(*alloc_inode)(struct super_block *sb);
8691 @@ -2437,6 +2476,7 @@ extern int dcache_dir_open(struct inode
8692  extern int dcache_dir_close(struct inode *, struct file *);
8693  extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
8694  extern int dcache_readdir(struct file *, void *, filldir_t);
8695 +extern int dcache_readdir_filter(struct file *, void *, filldir_t, int (*)(struct dentry *));
8696  extern int simple_setattr(struct dentry *, struct iattr *);
8697  extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
8698  extern int simple_statfs(struct dentry *, struct kstatfs *);
8699 diff -NurpP --minimal linux-3.0-rc5/include/linux/gfs2_ondisk.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/gfs2_ondisk.h
8700 --- linux-3.0-rc5/include/linux/gfs2_ondisk.h   2010-07-07 18:31:55.000000000 +0200
8701 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/gfs2_ondisk.h      2011-06-10 22:11:24.000000000 +0200
8702 @@ -211,6 +211,9 @@ enum {
8703         gfs2fl_NoAtime          = 7,
8704         gfs2fl_Sync             = 8,
8705         gfs2fl_System           = 9,
8706 +       gfs2fl_IXUnlink         = 16,
8707 +       gfs2fl_Barrier          = 17,
8708 +       gfs2fl_Cow              = 18,
8709         gfs2fl_TruncInProg      = 29,
8710         gfs2fl_InheritDirectio  = 30,
8711         gfs2fl_InheritJdata     = 31,
8712 @@ -227,6 +230,9 @@ enum {
8713  #define GFS2_DIF_NOATIME               0x00000080
8714  #define GFS2_DIF_SYNC                  0x00000100
8715  #define GFS2_DIF_SYSTEM                        0x00000200 /* New in gfs2 */
8716 +#define GFS2_DIF_IXUNLINK              0x00010000
8717 +#define GFS2_DIF_BARRIER               0x00020000
8718 +#define GFS2_DIF_COW                   0x00040000
8719  #define GFS2_DIF_TRUNC_IN_PROG         0x20000000 /* New in gfs2 */
8720  #define GFS2_DIF_INHERIT_DIRECTIO      0x40000000
8721  #define GFS2_DIF_INHERIT_JDATA         0x80000000
8722 diff -NurpP --minimal linux-3.0-rc5/include/linux/if_tun.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/if_tun.h
8723 --- linux-3.0-rc5/include/linux/if_tun.h        2010-08-02 16:52:54.000000000 +0200
8724 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/if_tun.h   2011-06-10 22:11:24.000000000 +0200
8725 @@ -53,6 +53,7 @@
8726  #define TUNDETACHFILTER _IOW('T', 214, struct sock_fprog)
8727  #define TUNGETVNETHDRSZ _IOR('T', 215, int)
8728  #define TUNSETVNETHDRSZ _IOW('T', 216, int)
8729 +#define TUNSETNID     _IOW('T', 217, int)
8730
8731  /* TUNSETIFF ifr flags */
8732  #define IFF_TUN                0x0001
8733 diff -NurpP --minimal linux-3.0-rc5/include/linux/init_task.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/init_task.h
8734 --- linux-3.0-rc5/include/linux/init_task.h     2011-07-01 11:14:21.000000000 +0200
8735 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/init_task.h        2011-06-10 22:11:24.000000000 +0200
8736 @@ -193,6 +193,10 @@ extern struct cred init_cred;
8737         INIT_FTRACE_GRAPH                                               \
8738         INIT_TRACE_RECURSION                                            \
8739         INIT_TASK_RCU_PREEMPT(tsk)                                      \
8740 +       .xid            = 0,                                            \
8741 +       .vx_info        = NULL,                                         \
8742 +       .nid            = 0,                                            \
8743 +       .nx_info        = NULL,                                         \
8744  }
8745
8746
8747 diff -NurpP --minimal linux-3.0-rc5/include/linux/ipc.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/ipc.h
8748 --- linux-3.0-rc5/include/linux/ipc.h   2009-12-03 20:02:55.000000000 +0100
8749 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/ipc.h      2011-06-10 22:11:24.000000000 +0200
8750 @@ -91,6 +91,7 @@ struct kern_ipc_perm
8751         key_t           key;
8752         uid_t           uid;
8753         gid_t           gid;
8754 +       xid_t           xid;
8755         uid_t           cuid;
8756         gid_t           cgid;
8757         mode_t          mode;
8758 diff -NurpP --minimal linux-3.0-rc5/include/linux/ipc_namespace.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/ipc_namespace.h
8759 --- linux-3.0-rc5/include/linux/ipc_namespace.h 2011-05-22 16:17:55.000000000 +0200
8760 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/ipc_namespace.h    2011-06-13 14:09:44.000000000 +0200
8761 @@ -94,7 +94,8 @@ static inline int mq_init_ns(struct ipc_
8762
8763  #if defined(CONFIG_IPC_NS)
8764  extern struct ipc_namespace *copy_ipcs(unsigned long flags,
8765 -                                      struct task_struct *tsk);
8766 +                                      struct ipc_namespace *old_ns,
8767 +                                      struct user_namespace *user_ns);
8768  static inline struct ipc_namespace *get_ipc_ns(struct ipc_namespace *ns)
8769  {
8770         if (ns)
8771 @@ -105,12 +106,13 @@ static inline struct ipc_namespace *get_
8772  extern void put_ipc_ns(struct ipc_namespace *ns);
8773  #else
8774  static inline struct ipc_namespace *copy_ipcs(unsigned long flags,
8775 -                                             struct task_struct *tsk)
8776 +                                             struct ipc_namespace *old_ns,
8777 +                                             struct user_namespace *user_ns)
8778  {
8779         if (flags & CLONE_NEWIPC)
8780                 return ERR_PTR(-EINVAL);
8781
8782 -       return tsk->nsproxy->ipc_ns;
8783 +       return old_ns;
8784  }
8785
8786  static inline struct ipc_namespace *get_ipc_ns(struct ipc_namespace *ns)
8787 diff -NurpP --minimal linux-3.0-rc5/include/linux/loop.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/loop.h
8788 --- linux-3.0-rc5/include/linux/loop.h  2009-09-10 15:26:25.000000000 +0200
8789 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/loop.h     2011-06-10 22:11:24.000000000 +0200
8790 @@ -45,6 +45,7 @@ struct loop_device {
8791         struct loop_func_table *lo_encryption;
8792         __u32           lo_init[2];
8793         uid_t           lo_key_owner;   /* Who set the key */
8794 +       xid_t           lo_xid;
8795         int             (*ioctl)(struct loop_device *, int cmd,
8796                                  unsigned long arg);
8797
8798 diff -NurpP --minimal linux-3.0-rc5/include/linux/magic.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/magic.h
8799 --- linux-3.0-rc5/include/linux/magic.h 2011-05-22 16:17:55.000000000 +0200
8800 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/magic.h    2011-06-10 22:11:24.000000000 +0200
8801 @@ -3,7 +3,7 @@
8802
8803  #define ADFS_SUPER_MAGIC       0xadf5
8804  #define AFFS_SUPER_MAGIC       0xadff
8805 -#define AFS_SUPER_MAGIC                0x5346414F
8806 +#define AFS_SUPER_MAGIC                0x5346414F
8807  #define AUTOFS_SUPER_MAGIC     0x0187
8808  #define CODA_SUPER_MAGIC       0x73757245
8809  #define CRAMFS_MAGIC           0x28cd3d45      /* some random number */
8810 @@ -41,6 +41,7 @@
8811  #define NFS_SUPER_MAGIC                0x6969
8812  #define OPENPROM_SUPER_MAGIC   0x9fa1
8813  #define PROC_SUPER_MAGIC       0x9fa0
8814 +#define DEVPTS_SUPER_MAGIC     0x1cd1
8815  #define QNX4_SUPER_MAGIC       0x002f          /* qnx4 fs detection */
8816
8817  #define REISERFS_SUPER_MAGIC   0x52654973      /* used by gcc */
8818 diff -NurpP --minimal linux-3.0-rc5/include/linux/major.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/major.h
8819 --- linux-3.0-rc5/include/linux/major.h 2009-09-10 15:26:25.000000000 +0200
8820 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/major.h    2011-06-10 22:11:24.000000000 +0200
8821 @@ -15,6 +15,7 @@
8822  #define HD_MAJOR               IDE0_MAJOR
8823  #define PTY_SLAVE_MAJOR                3
8824  #define TTY_MAJOR              4
8825 +#define VROOT_MAJOR            4
8826  #define TTYAUX_MAJOR           5
8827  #define LP_MAJOR               6
8828  #define VCS_MAJOR              7
8829 diff -NurpP --minimal linux-3.0-rc5/include/linux/memcontrol.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/memcontrol.h
8830 --- linux-3.0-rc5/include/linux/memcontrol.h    2011-07-01 11:14:21.000000000 +0200
8831 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/memcontrol.h       2011-06-22 12:39:15.000000000 +0200
8832 @@ -86,6 +86,13 @@ extern struct mem_cgroup *try_get_mem_cg
8833  extern struct mem_cgroup *mem_cgroup_from_task(struct task_struct *p);
8834  extern struct mem_cgroup *try_get_mem_cgroup_from_mm(struct mm_struct *mm);
8835
8836 +extern u64 mem_cgroup_res_read_u64(struct mem_cgroup *mem, int member);
8837 +extern u64 mem_cgroup_memsw_read_u64(struct mem_cgroup *mem, int member);
8838 +
8839 +extern s64 mem_cgroup_stat_read_cache(struct mem_cgroup *mem);
8840 +extern s64 mem_cgroup_stat_read_anon(struct mem_cgroup *mem);
8841 +extern s64 mem_cgroup_stat_read_mapped(struct mem_cgroup *mem);
8842 +
8843  static inline
8844  int mm_match_cgroup(const struct mm_struct *mm, const struct mem_cgroup *cgroup)
8845  {
8846 diff -NurpP --minimal linux-3.0-rc5/include/linux/mm_types.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/mm_types.h
8847 --- linux-3.0-rc5/include/linux/mm_types.h      2011-07-01 11:14:21.000000000 +0200
8848 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/mm_types.h 2011-06-10 22:11:24.000000000 +0200
8849 @@ -268,6 +268,7 @@ struct mm_struct {
8850
8851         /* Architecture-specific MM context */
8852         mm_context_t context;
8853 +       struct vx_info *mm_vx_info;
8854
8855         /* Swap token stuff */
8856         /*
8857 diff -NurpP --minimal linux-3.0-rc5/include/linux/mmzone.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/mmzone.h
8858 --- linux-3.0-rc5/include/linux/mmzone.h        2011-07-01 11:14:21.000000000 +0200
8859 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/mmzone.h   2011-07-01 11:35:35.000000000 +0200
8860 @@ -654,6 +654,13 @@ typedef struct pglist_data {
8861         __pgdat->node_start_pfn + __pgdat->node_spanned_pages;\
8862  })
8863
8864 +#define node_start_pfn(nid)    (NODE_DATA(nid)->node_start_pfn)
8865 +
8866 +#define node_end_pfn(nid) ({\
8867 +       pg_data_t *__pgdat = NODE_DATA(nid);\
8868 +       __pgdat->node_start_pfn + __pgdat->node_spanned_pages;\
8869 +})
8870 +
8871  #include <linux/memory_hotplug.h>
8872
8873  extern struct mutex zonelists_mutex;
8874 diff -NurpP --minimal linux-3.0-rc5/include/linux/mount.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/mount.h
8875 --- linux-3.0-rc5/include/linux/mount.h 2011-03-15 18:07:39.000000000 +0100
8876 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/mount.h    2011-06-10 22:11:24.000000000 +0200
8877 @@ -52,6 +52,9 @@ struct mnt_pcp {
8878         int mnt_writers;
8879  };
8880
8881 +#define MNT_TAGID      0x10000
8882 +#define MNT_NOTAG      0x20000
8883 +
8884  struct vfsmount {
8885         struct list_head mnt_hash;
8886         struct vfsmount *mnt_parent;    /* fs we are mounted on */
8887 @@ -86,6 +89,7 @@ struct vfsmount {
8888         int mnt_expiry_mark;            /* true if marked for expiry */
8889         int mnt_pinned;
8890         int mnt_ghosts;
8891 +       tag_t mnt_tag;                  /* tagging used for vfsmount */
8892  };
8893
8894  struct file; /* forward dec */
8895 diff -NurpP --minimal linux-3.0-rc5/include/linux/net.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/net.h
8896 --- linux-3.0-rc5/include/linux/net.h   2011-07-01 11:14:21.000000000 +0200
8897 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/net.h      2011-06-10 22:11:24.000000000 +0200
8898 @@ -72,6 +72,7 @@ struct net;
8899  #define SOCK_NOSPACE           2
8900  #define SOCK_PASSCRED          3
8901  #define SOCK_PASSSEC           4
8902 +#define SOCK_USER_SOCKET       5
8903
8904  #ifndef ARCH_HAS_SOCKET_TYPES
8905  /**
8906 diff -NurpP --minimal linux-3.0-rc5/include/linux/nfs_mount.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/nfs_mount.h
8907 --- linux-3.0-rc5/include/linux/nfs_mount.h     2011-01-05 21:50:31.000000000 +0100
8908 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/nfs_mount.h        2011-06-10 22:11:24.000000000 +0200
8909 @@ -63,7 +63,8 @@ struct nfs_mount_data {
8910  #define NFS_MOUNT_SECFLAVOUR   0x2000  /* 5 */
8911  #define NFS_MOUNT_NORDIRPLUS   0x4000  /* 5 */
8912  #define NFS_MOUNT_UNSHARED     0x8000  /* 5 */
8913 -#define NFS_MOUNT_FLAGMASK     0xFFFF
8914 +#define NFS_MOUNT_TAGGED       0x10000 /* context tagging */
8915 +#define NFS_MOUNT_FLAGMASK     0x1FFFF
8916
8917  /* The following are for internal use only */
8918  #define NFS_MOUNT_LOOKUP_CACHE_NONEG   0x10000
8919 diff -NurpP --minimal linux-3.0-rc5/include/linux/nsproxy.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/nsproxy.h
8920 --- linux-3.0-rc5/include/linux/nsproxy.h       2011-07-01 11:14:21.000000000 +0200
8921 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/nsproxy.h  2011-06-10 22:11:24.000000000 +0200
8922 @@ -3,6 +3,7 @@
8923
8924  #include <linux/spinlock.h>
8925  #include <linux/sched.h>
8926 +#include <linux/vserver/debug.h>
8927
8928  struct mnt_namespace;
8929  struct uts_namespace;
8930 @@ -63,22 +64,33 @@ static inline struct nsproxy *task_nspro
8931  }
8932
8933  int copy_namespaces(unsigned long flags, struct task_struct *tsk);
8934 +struct nsproxy *copy_nsproxy(struct nsproxy *orig);
8935  void exit_task_namespaces(struct task_struct *tsk);
8936  void switch_task_namespaces(struct task_struct *tsk, struct nsproxy *new);
8937  void free_nsproxy(struct nsproxy *ns);
8938  int unshare_nsproxy_namespaces(unsigned long, struct nsproxy **,
8939         struct fs_struct *);
8940
8941 -static inline void put_nsproxy(struct nsproxy *ns)
8942 +#define        get_nsproxy(n)  __get_nsproxy(n, __FILE__, __LINE__)
8943 +
8944 +static inline void __get_nsproxy(struct nsproxy *ns,
8945 +       const char *_file, int _line)
8946  {
8947 -       if (atomic_dec_and_test(&ns->count)) {
8948 -               free_nsproxy(ns);
8949 -       }
8950 +       vxlprintk(VXD_CBIT(space, 0), "get_nsproxy(%p[%u])",
8951 +               ns, atomic_read(&ns->count), _file, _line);
8952 +       atomic_inc(&ns->count);
8953  }
8954
8955 -static inline void get_nsproxy(struct nsproxy *ns)
8956 +#define        put_nsproxy(n)  __put_nsproxy(n, __FILE__, __LINE__)
8957 +
8958 +static inline void __put_nsproxy(struct nsproxy *ns,
8959 +       const char *_file, int _line)
8960  {
8961 -       atomic_inc(&ns->count);
8962 +       vxlprintk(VXD_CBIT(space, 0), "put_nsproxy(%p[%u])",
8963 +               ns, atomic_read(&ns->count), _file, _line);
8964 +       if (atomic_dec_and_test(&ns->count)) {
8965 +               free_nsproxy(ns);
8966 +       }
8967  }
8968
8969  #endif
8970 diff -NurpP --minimal linux-3.0-rc5/include/linux/pid.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/pid.h
8971 --- linux-3.0-rc5/include/linux/pid.h   2011-07-01 11:14:21.000000000 +0200
8972 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/pid.h      2011-06-10 22:11:24.000000000 +0200
8973 @@ -8,7 +8,8 @@ enum pid_type
8974         PIDTYPE_PID,
8975         PIDTYPE_PGID,
8976         PIDTYPE_SID,
8977 -       PIDTYPE_MAX
8978 +       PIDTYPE_MAX,
8979 +       PIDTYPE_REALPID
8980  };
8981
8982  /*
8983 @@ -171,6 +172,7 @@ static inline pid_t pid_nr(struct pid *p
8984  }
8985
8986  pid_t pid_nr_ns(struct pid *pid, struct pid_namespace *ns);
8987 +pid_t pid_unmapped_nr_ns(struct pid *pid, struct pid_namespace *ns);
8988  pid_t pid_vnr(struct pid *pid);
8989
8990  #define do_each_pid_task(pid, type, task)                              \
8991 diff -NurpP --minimal linux-3.0-rc5/include/linux/proc_fs.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/proc_fs.h
8992 --- linux-3.0-rc5/include/linux/proc_fs.h       2011-07-01 11:14:21.000000000 +0200
8993 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/proc_fs.h  2011-06-10 22:11:24.000000000 +0200
8994 @@ -56,6 +56,7 @@ struct proc_dir_entry {
8995         nlink_t nlink;
8996         uid_t uid;
8997         gid_t gid;
8998 +       int vx_flags;
8999         loff_t size;
9000         const struct inode_operations *proc_iops;
9001         /*
9002 @@ -252,12 +253,18 @@ extern const struct proc_ns_operations n
9003  extern const struct proc_ns_operations utsns_operations;
9004  extern const struct proc_ns_operations ipcns_operations;
9005
9006 +struct vx_info;
9007 +struct nx_info;
9008 +
9009  union proc_op {
9010         int (*proc_get_link)(struct inode *, struct path *);
9011         int (*proc_read)(struct task_struct *task, char *page);
9012         int (*proc_show)(struct seq_file *m,
9013                 struct pid_namespace *ns, struct pid *pid,
9014                 struct task_struct *task);
9015 +       int (*proc_vs_read)(char *page);
9016 +       int (*proc_vxi_read)(struct vx_info *vxi, char *page);
9017 +       int (*proc_nxi_read)(struct nx_info *nxi, char *page);
9018  };
9019
9020  struct ctl_table_header;
9021 @@ -265,6 +272,7 @@ struct ctl_table;
9022
9023  struct proc_inode {
9024         struct pid *pid;
9025 +       int vx_flags;
9026         int fd;
9027         union proc_op op;
9028         struct proc_dir_entry *pde;
9029 diff -NurpP --minimal linux-3.0-rc5/include/linux/quotaops.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/quotaops.h
9030 --- linux-3.0-rc5/include/linux/quotaops.h      2011-05-22 16:17:57.000000000 +0200
9031 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/quotaops.h 2011-06-10 22:11:24.000000000 +0200
9032 @@ -8,6 +8,7 @@
9033  #define _LINUX_QUOTAOPS_
9034
9035  #include <linux/fs.h>
9036 +#include <linux/vs_dlimit.h>
9037
9038  #define DQUOT_SPACE_WARN       0x1
9039  #define DQUOT_SPACE_RESERVE    0x2
9040 @@ -204,11 +205,12 @@ static inline void dquot_drop(struct ino
9041
9042  static inline int dquot_alloc_inode(const struct inode *inode)
9043  {
9044 -       return 0;
9045 +       return dl_alloc_inode(inode);
9046  }
9047
9048  static inline void dquot_free_inode(const struct inode *inode)
9049  {
9050 +       dl_free_inode(inode);
9051  }
9052
9053  static inline int dquot_transfer(struct inode *inode, struct iattr *iattr)
9054 @@ -219,6 +221,10 @@ static inline int dquot_transfer(struct
9055  static inline int __dquot_alloc_space(struct inode *inode, qsize_t number,
9056                 int flags)
9057  {
9058 +       int ret = 0;
9059 +
9060 +       if ((ret = dl_alloc_space(inode, number)))
9061 +               return ret;
9062         if (!(flags & DQUOT_SPACE_RESERVE))
9063                 inode_add_bytes(inode, number);
9064         return 0;
9065 @@ -229,6 +235,7 @@ static inline void __dquot_free_space(st
9066  {
9067         if (!(flags & DQUOT_SPACE_RESERVE))
9068                 inode_sub_bytes(inode, number);
9069 +       dl_free_space(inode, number);
9070  }
9071
9072  static inline int dquot_claim_space_nodirty(struct inode *inode, qsize_t number)
9073 diff -NurpP --minimal linux-3.0-rc5/include/linux/reboot.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/reboot.h
9074 --- linux-3.0-rc5/include/linux/reboot.h        2010-07-07 18:31:56.000000000 +0200
9075 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/reboot.h   2011-06-10 22:11:24.000000000 +0200
9076 @@ -33,6 +33,7 @@
9077  #define        LINUX_REBOOT_CMD_RESTART2       0xA1B2C3D4
9078  #define        LINUX_REBOOT_CMD_SW_SUSPEND     0xD000FCE2
9079  #define        LINUX_REBOOT_CMD_KEXEC          0x45584543
9080 +#define        LINUX_REBOOT_CMD_OOM            0xDEADBEEF
9081
9082
9083  #ifdef __KERNEL__
9084 diff -NurpP --minimal linux-3.0-rc5/include/linux/reiserfs_fs.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/reiserfs_fs.h
9085 --- linux-3.0-rc5/include/linux/reiserfs_fs.h   2011-05-22 16:17:58.000000000 +0200
9086 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/reiserfs_fs.h      2011-06-10 22:11:24.000000000 +0200
9087 @@ -976,6 +976,11 @@ struct stat_data_v1 {
9088  #define REISERFS_COMPR_FL     FS_COMPR_FL
9089  #define REISERFS_NOTAIL_FL    FS_NOTAIL_FL
9090
9091 +/* unfortunately reiserfs sdattr is only 16 bit */
9092 +#define REISERFS_IXUNLINK_FL  (FS_IXUNLINK_FL >> 16)
9093 +#define REISERFS_BARRIER_FL   (FS_BARRIER_FL >> 16)
9094 +#define REISERFS_COW_FL       (FS_COW_FL >> 16)
9095 +
9096  /* persistent flags that file inherits from the parent directory */
9097  #define REISERFS_INHERIT_MASK ( REISERFS_IMMUTABLE_FL |        \
9098                                 REISERFS_SYNC_FL |      \
9099 @@ -985,6 +990,9 @@ struct stat_data_v1 {
9100                                 REISERFS_COMPR_FL |     \
9101                                 REISERFS_NOTAIL_FL )
9102
9103 +#define REISERFS_FL_USER_VISIBLE       0x80FF
9104 +#define REISERFS_FL_USER_MODIFIABLE    0x80FF
9105 +
9106  /* Stat Data on disk (reiserfs version of UFS disk inode minus the
9107     address blocks) */
9108  struct stat_data {
9109 @@ -2073,6 +2081,7 @@ static inline void reiserfs_update_sd(st
9110  void sd_attrs_to_i_attrs(__u16 sd_attrs, struct inode *inode);
9111  void i_attrs_to_sd_attrs(struct inode *inode, __u16 * sd_attrs);
9112  int reiserfs_setattr(struct dentry *dentry, struct iattr *attr);
9113 +int reiserfs_sync_flags(struct inode *inode, int, int);
9114
9115  int __reiserfs_write_begin(struct page *page, unsigned from, unsigned len);
9116
9117 diff -NurpP --minimal linux-3.0-rc5/include/linux/reiserfs_fs_sb.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/reiserfs_fs_sb.h
9118 --- linux-3.0-rc5/include/linux/reiserfs_fs_sb.h        2010-02-25 11:52:07.000000000 +0100
9119 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/reiserfs_fs_sb.h   2011-06-10 22:11:24.000000000 +0200
9120 @@ -476,6 +476,7 @@ enum reiserfs_mount_options {
9121         REISERFS_EXPOSE_PRIVROOT,
9122         REISERFS_BARRIER_NONE,
9123         REISERFS_BARRIER_FLUSH,
9124 +       REISERFS_TAGGED,
9125
9126         /* Actions on error */
9127         REISERFS_ERROR_PANIC,
9128 diff -NurpP --minimal linux-3.0-rc5/include/linux/sched.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/sched.h
9129 --- linux-3.0-rc5/include/linux/sched.h 2011-07-01 11:14:21.000000000 +0200
9130 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/sched.h    2011-06-15 02:40:14.000000000 +0200
9131 @@ -1397,6 +1397,14 @@ struct task_struct {
9132  #endif
9133         seccomp_t seccomp;
9134
9135 +/* vserver context data */
9136 +       struct vx_info *vx_info;
9137 +       struct nx_info *nx_info;
9138 +
9139 +       xid_t xid;
9140 +       nid_t nid;
9141 +       tag_t tag;
9142 +
9143  /* Thread group tracking */
9144         u32 parent_exec_id;
9145         u32 self_exec_id;
9146 @@ -1640,6 +1648,11 @@ struct pid_namespace;
9147  pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type,
9148                         struct pid_namespace *ns);
9149
9150 +#include <linux/vserver/base.h>
9151 +#include <linux/vserver/context.h>
9152 +#include <linux/vserver/debug.h>
9153 +#include <linux/vserver/pid.h>
9154 +
9155  static inline pid_t task_pid_nr(struct task_struct *tsk)
9156  {
9157         return tsk->pid;
9158 @@ -1653,7 +1666,8 @@ static inline pid_t task_pid_nr_ns(struc
9159
9160  static inline pid_t task_pid_vnr(struct task_struct *tsk)
9161  {
9162 -       return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
9163 +       // return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
9164 +       return vx_map_pid(__task_pid_nr_ns(tsk, PIDTYPE_PID, NULL));
9165  }
9166
9167
9168 @@ -1666,7 +1680,7 @@ pid_t task_tgid_nr_ns(struct task_struct
9169
9170  static inline pid_t task_tgid_vnr(struct task_struct *tsk)
9171  {
9172 -       return pid_vnr(task_tgid(tsk));
9173 +       return vx_map_tgid(pid_vnr(task_tgid(tsk)));
9174  }
9175
9176
9177 diff -NurpP --minimal linux-3.0-rc5/include/linux/shmem_fs.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/shmem_fs.h
9178 --- linux-3.0-rc5/include/linux/shmem_fs.h      2011-07-01 11:14:21.000000000 +0200
9179 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/shmem_fs.h 2011-07-01 11:35:35.000000000 +0200
9180 @@ -12,6 +12,9 @@
9181
9182  #define SHMEM_SYMLINK_INLINE_LEN (SHMEM_NR_DIRECT * sizeof(swp_entry_t))
9183
9184 +#define TMPFS_SUPER_MAGIC      0x01021994
9185 +
9186 +
9187  struct shmem_inode_info {
9188         spinlock_t              lock;
9189         unsigned long           flags;
9190 diff -NurpP --minimal linux-3.0-rc5/include/linux/stat.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/stat.h
9191 --- linux-3.0-rc5/include/linux/stat.h  2008-12-25 00:26:37.000000000 +0100
9192 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/stat.h     2011-06-10 22:11:24.000000000 +0200
9193 @@ -66,6 +66,7 @@ struct kstat {
9194         unsigned int    nlink;
9195         uid_t           uid;
9196         gid_t           gid;
9197 +       tag_t           tag;
9198         dev_t           rdev;
9199         loff_t          size;
9200         struct timespec  atime;
9201 diff -NurpP --minimal linux-3.0-rc5/include/linux/sunrpc/auth.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/sunrpc/auth.h
9202 --- linux-3.0-rc5/include/linux/sunrpc/auth.h   2011-03-15 18:07:39.000000000 +0100
9203 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/sunrpc/auth.h      2011-06-10 22:11:24.000000000 +0200
9204 @@ -25,6 +25,7 @@
9205  struct auth_cred {
9206         uid_t   uid;
9207         gid_t   gid;
9208 +       tag_t   tag;
9209         struct group_info *group_info;
9210         unsigned char machine_cred : 1;
9211  };
9212 diff -NurpP --minimal linux-3.0-rc5/include/linux/sunrpc/clnt.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/sunrpc/clnt.h
9213 --- linux-3.0-rc5/include/linux/sunrpc/clnt.h   2011-05-22 16:17:58.000000000 +0200
9214 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/sunrpc/clnt.h      2011-06-10 22:11:24.000000000 +0200
9215 @@ -49,7 +49,8 @@ struct rpc_clnt {
9216         unsigned int            cl_softrtry : 1,/* soft timeouts */
9217                                 cl_discrtry : 1,/* disconnect before retry */
9218                                 cl_autobind : 1,/* use getport() */
9219 -                               cl_chatty   : 1;/* be verbose */
9220 +                               cl_chatty   : 1,/* be verbose */
9221 +                               cl_tag      : 1;/* context tagging */
9222
9223         struct rpc_rtt *        cl_rtt;         /* RTO estimator data */
9224         const struct rpc_timeout *cl_timeout;   /* Timeout strategy */
9225 diff -NurpP --minimal linux-3.0-rc5/include/linux/syscalls.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/syscalls.h
9226 --- linux-3.0-rc5/include/linux/syscalls.h      2011-07-01 11:14:22.000000000 +0200
9227 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/syscalls.h 2011-06-10 22:11:24.000000000 +0200
9228 @@ -483,6 +483,8 @@ asmlinkage long sys_symlink(const char _
9229  asmlinkage long sys_unlink(const char __user *pathname);
9230  asmlinkage long sys_rename(const char __user *oldname,
9231                                 const char __user *newname);
9232 +asmlinkage long sys_copyfile(const char __user *from, const char __user *to,
9233 +                               umode_t mode);
9234  asmlinkage long sys_chmod(const char __user *filename, mode_t mode);
9235  asmlinkage long sys_fchmod(unsigned int fd, mode_t mode);
9236
9237 diff -NurpP --minimal linux-3.0-rc5/include/linux/sysctl.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/sysctl.h
9238 --- linux-3.0-rc5/include/linux/sysctl.h        2011-03-15 18:07:40.000000000 +0100
9239 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/sysctl.h   2011-06-10 22:11:24.000000000 +0200
9240 @@ -60,6 +60,7 @@ enum
9241         CTL_ABI=9,              /* Binary emulation */
9242         CTL_CPU=10,             /* CPU stuff (speed scaling, etc) */
9243         CTL_ARLAN=254,          /* arlan wireless driver */
9244 +       CTL_VSERVER=4242,       /* Linux-VServer debug */
9245         CTL_S390DBF=5677,       /* s390 debug */
9246         CTL_SUNRPC=7249,        /* sunrpc debug */
9247         CTL_PM=9899,            /* frv power management */
9248 @@ -94,6 +95,7 @@ enum
9249
9250         KERN_PANIC=15,          /* int: panic timeout */
9251         KERN_REALROOTDEV=16,    /* real root device to mount after initrd */
9252 +       KERN_VSHELPER=17,       /* string: path to vshelper policy agent */
9253
9254         KERN_SPARC_REBOOT=21,   /* reboot command on Sparc */
9255         KERN_CTLALTDEL=22,      /* int: allow ctl-alt-del to reboot */
9256 diff -NurpP --minimal linux-3.0-rc5/include/linux/sysfs.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/sysfs.h
9257 --- linux-3.0-rc5/include/linux/sysfs.h 2011-07-01 11:14:22.000000000 +0200
9258 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/sysfs.h    2011-06-22 12:39:15.000000000 +0200
9259 @@ -19,6 +19,8 @@
9260  #include <linux/kobject_ns.h>
9261  #include <asm/atomic.h>
9262
9263 +#define SYSFS_SUPER_MAGIC      0x62656572
9264 +
9265  struct kobject;
9266  struct module;
9267  enum kobj_ns_type;
9268 diff -NurpP --minimal linux-3.0-rc5/include/linux/time.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/time.h
9269 --- linux-3.0-rc5/include/linux/time.h  2011-07-01 11:14:22.000000000 +0200
9270 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/time.h     2011-06-10 22:11:24.000000000 +0200
9271 @@ -256,6 +256,9 @@ static __always_inline void timespec_add
9272         a->tv_sec += __iter_div_u64_rem(a->tv_nsec + ns, NSEC_PER_SEC, &ns);
9273         a->tv_nsec = ns;
9274  }
9275 +
9276 +#include <linux/vs_time.h>
9277 +
9278  #endif /* __KERNEL__ */
9279
9280  #define NFDBITS                        __NFDBITS
9281 diff -NurpP --minimal linux-3.0-rc5/include/linux/types.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/types.h
9282 --- linux-3.0-rc5/include/linux/types.h 2011-05-22 16:17:58.000000000 +0200
9283 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/types.h    2011-06-10 22:11:24.000000000 +0200
9284 @@ -40,6 +40,9 @@ typedef __kernel_uid32_t      uid_t;
9285  typedef __kernel_gid32_t       gid_t;
9286  typedef __kernel_uid16_t        uid16_t;
9287  typedef __kernel_gid16_t        gid16_t;
9288 +typedef unsigned int           xid_t;
9289 +typedef unsigned int           nid_t;
9290 +typedef unsigned int           tag_t;
9291
9292  typedef unsigned long          uintptr_t;
9293
9294 diff -NurpP --minimal linux-3.0-rc5/include/linux/utsname.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/utsname.h
9295 --- linux-3.0-rc5/include/linux/utsname.h       2011-05-22 16:17:58.000000000 +0200
9296 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/utsname.h  2011-06-13 14:36:48.000000000 +0200
9297 @@ -54,7 +54,8 @@ static inline void get_uts_ns(struct uts
9298  }
9299
9300  extern struct uts_namespace *copy_utsname(unsigned long flags,
9301 -                                         struct task_struct *tsk);
9302 +                                         struct uts_namespace *old_ns,
9303 +                                         struct user_namespace *user_ns);
9304  extern void free_uts_ns(struct kref *kref);
9305
9306  static inline void put_uts_ns(struct uts_namespace *ns)
9307 @@ -71,12 +72,13 @@ static inline void put_uts_ns(struct uts
9308  }
9309
9310  static inline struct uts_namespace *copy_utsname(unsigned long flags,
9311 -                                                struct task_struct *tsk)
9312 +                                                struct uts_namespace *old_ns,
9313 +                                                struct user_namespace *user_ns)
9314  {
9315         if (flags & CLONE_NEWUTS)
9316                 return ERR_PTR(-EINVAL);
9317
9318 -       return tsk->nsproxy->uts_ns;
9319 +       return old_ns;
9320  }
9321  #endif
9322
9323 diff -NurpP --minimal linux-3.0-rc5/include/linux/vroot.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vroot.h
9324 --- linux-3.0-rc5/include/linux/vroot.h 1970-01-01 01:00:00.000000000 +0100
9325 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vroot.h    2011-06-10 22:11:24.000000000 +0200
9326 @@ -0,0 +1,51 @@
9327 +
9328 +/*
9329 + * include/linux/vroot.h
9330 + *
9331 + * written by Herbert Pötzl, 9/11/2002
9332 + * ported to 2.6 by Herbert Pötzl, 30/12/2004
9333 + *
9334 + * Copyright (C) 2002-2007 by Herbert Pötzl.
9335 + * Redistribution of this file is permitted under the
9336 + * GNU General Public License.
9337 + */
9338 +
9339 +#ifndef _LINUX_VROOT_H
9340 +#define _LINUX_VROOT_H
9341 +
9342 +
9343 +#ifdef __KERNEL__
9344 +
9345 +/* Possible states of device */
9346 +enum {
9347 +       Vr_unbound,
9348 +       Vr_bound,
9349 +};
9350 +
9351 +struct vroot_device {
9352 +       int             vr_number;
9353 +       int             vr_refcnt;
9354 +
9355 +       struct semaphore        vr_ctl_mutex;
9356 +       struct block_device    *vr_device;
9357 +       int                     vr_state;
9358 +};
9359 +
9360 +
9361 +typedef struct block_device *(vroot_grb_func)(struct block_device *);
9362 +
9363 +extern int register_vroot_grb(vroot_grb_func *);
9364 +extern int unregister_vroot_grb(vroot_grb_func *);
9365 +
9366 +#endif /* __KERNEL__ */
9367 +
9368 +#define MAX_VROOT_DEFAULT      8
9369 +
9370 +/*
9371 + * IOCTL commands --- we will commandeer 0x56 ('V')
9372 + */
9373 +
9374 +#define VROOT_SET_DEV          0x5600
9375 +#define VROOT_CLR_DEV          0x5601
9376 +
9377 +#endif /* _LINUX_VROOT_H */
9378 diff -NurpP --minimal linux-3.0-rc5/include/linux/vs_base.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_base.h
9379 --- linux-3.0-rc5/include/linux/vs_base.h       1970-01-01 01:00:00.000000000 +0100
9380 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_base.h  2011-06-10 22:11:24.000000000 +0200
9381 @@ -0,0 +1,10 @@
9382 +#ifndef _VS_BASE_H
9383 +#define _VS_BASE_H
9384 +
9385 +#include "vserver/base.h"
9386 +#include "vserver/check.h"
9387 +#include "vserver/debug.h"
9388 +
9389 +#else
9390 +#warning duplicate inclusion
9391 +#endif
9392 diff -NurpP --minimal linux-3.0-rc5/include/linux/vs_context.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_context.h
9393 --- linux-3.0-rc5/include/linux/vs_context.h    1970-01-01 01:00:00.000000000 +0100
9394 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_context.h       2011-06-10 22:11:24.000000000 +0200
9395 @@ -0,0 +1,242 @@
9396 +#ifndef _VS_CONTEXT_H
9397 +#define _VS_CONTEXT_H
9398 +
9399 +#include "vserver/base.h"
9400 +#include "vserver/check.h"
9401 +#include "vserver/context.h"
9402 +#include "vserver/history.h"
9403 +#include "vserver/debug.h"
9404 +
9405 +#include <linux/sched.h>
9406 +
9407 +
9408 +#define get_vx_info(i) __get_vx_info(i, __FILE__, __LINE__, __HERE__)
9409 +
9410 +static inline struct vx_info *__get_vx_info(struct vx_info *vxi,
9411 +       const char *_file, int _line, void *_here)
9412 +{
9413 +       if (!vxi)
9414 +               return NULL;
9415 +
9416 +       vxlprintk(VXD_CBIT(xid, 2), "get_vx_info(%p[#%d.%d])",
9417 +               vxi, vxi ? vxi->vx_id : 0,
9418 +               vxi ? atomic_read(&vxi->vx_usecnt) : 0,
9419 +               _file, _line);
9420 +       __vxh_get_vx_info(vxi, _here);
9421 +
9422 +       atomic_inc(&vxi->vx_usecnt);
9423 +       return vxi;
9424 +}
9425 +
9426 +
9427 +extern void free_vx_info(struct vx_info *);
9428 +
9429 +#define put_vx_info(i) __put_vx_info(i, __FILE__, __LINE__, __HERE__)
9430 +
9431 +static inline void __put_vx_info(struct vx_info *vxi,
9432 +       const char *_file, int _line, void *_here)
9433 +{
9434 +       if (!vxi)
9435 +               return;
9436 +
9437 +       vxlprintk(VXD_CBIT(xid, 2), "put_vx_info(%p[#%d.%d])",
9438 +               vxi, vxi ? vxi->vx_id : 0,
9439 +               vxi ? atomic_read(&vxi->vx_usecnt) : 0,
9440 +               _file, _line);
9441 +       __vxh_put_vx_info(vxi, _here);
9442 +
9443 +       if (atomic_dec_and_test(&vxi->vx_usecnt))
9444 +               free_vx_info(vxi);
9445 +}
9446 +
9447 +
9448 +#define init_vx_info(p, i) \
9449 +       __init_vx_info(p, i, __FILE__, __LINE__, __HERE__)
9450 +
9451 +static inline void __init_vx_info(struct vx_info **vxp, struct vx_info *vxi,
9452 +       const char *_file, int _line, void *_here)
9453 +{
9454 +       if (vxi) {
9455 +               vxlprintk(VXD_CBIT(xid, 3),
9456 +                       "init_vx_info(%p[#%d.%d])",
9457 +                       vxi, vxi ? vxi->vx_id : 0,
9458 +                       vxi ? atomic_read(&vxi->vx_usecnt) : 0,
9459 +                       _file, _line);
9460 +               __vxh_init_vx_info(vxi, vxp, _here);
9461 +
9462 +               atomic_inc(&vxi->vx_usecnt);
9463 +       }
9464 +       *vxp = vxi;
9465 +}
9466 +
9467 +
9468 +#define set_vx_info(p, i) \
9469 +       __set_vx_info(p, i, __FILE__, __LINE__, __HERE__)
9470 +
9471 +static inline void __set_vx_info(struct vx_info **vxp, struct vx_info *vxi,
9472 +       const char *_file, int _line, void *_here)
9473 +{
9474 +       struct vx_info *vxo;
9475 +
9476 +       if (!vxi)
9477 +               return;
9478 +
9479 +       vxlprintk(VXD_CBIT(xid, 3), "set_vx_info(%p[#%d.%d])",
9480 +               vxi, vxi ? vxi->vx_id : 0,
9481 +               vxi ? atomic_read(&vxi->vx_usecnt) : 0,
9482 +               _file, _line);
9483 +       __vxh_set_vx_info(vxi, vxp, _here);
9484 +
9485 +       atomic_inc(&vxi->vx_usecnt);
9486 +       vxo = xchg(vxp, vxi);
9487 +       BUG_ON(vxo);
9488 +}
9489 +
9490 +
9491 +#define clr_vx_info(p) __clr_vx_info(p, __FILE__, __LINE__, __HERE__)
9492 +
9493 +static inline void __clr_vx_info(struct vx_info **vxp,
9494 +       const char *_file, int _line, void *_here)
9495 +{
9496 +       struct vx_info *vxo;
9497 +
9498 +       vxo = xchg(vxp, NULL);
9499 +       if (!vxo)
9500 +               return;
9501 +
9502 +       vxlprintk(VXD_CBIT(xid, 3), "clr_vx_info(%p[#%d.%d])",
9503 +               vxo, vxo ? vxo->vx_id : 0,
9504 +               vxo ? atomic_read(&vxo->vx_usecnt) : 0,
9505 +               _file, _line);
9506 +       __vxh_clr_vx_info(vxo, vxp, _here);
9507 +
9508 +       if (atomic_dec_and_test(&vxo->vx_usecnt))
9509 +               free_vx_info(vxo);
9510 +}
9511 +
9512 +
9513 +#define claim_vx_info(v, p) \
9514 +       __claim_vx_info(v, p, __FILE__, __LINE__, __HERE__)
9515 +
9516 +static inline void __claim_vx_info(struct vx_info *vxi,
9517 +       struct task_struct *task,
9518 +       const char *_file, int _line, void *_here)
9519 +{
9520 +       vxlprintk(VXD_CBIT(xid, 3), "claim_vx_info(%p[#%d.%d.%d]) %p",
9521 +               vxi, vxi ? vxi->vx_id : 0,
9522 +               vxi ? atomic_read(&vxi->vx_usecnt) : 0,
9523 +               vxi ? atomic_read(&vxi->vx_tasks) : 0,
9524 +               task, _file, _line);
9525 +       __vxh_claim_vx_info(vxi, task, _here);
9526 +
9527 +       atomic_inc(&vxi->vx_tasks);
9528 +}
9529 +
9530 +
9531 +extern void unhash_vx_info(struct vx_info *);
9532 +
9533 +#define release_vx_info(v, p) \
9534 +       __release_vx_info(v, p, __FILE__, __LINE__, __HERE__)
9535 +
9536 +static inline void __release_vx_info(struct vx_info *vxi,
9537 +       struct task_struct *task,
9538 +       const char *_file, int _line, void *_here)
9539 +{
9540 +       vxlprintk(VXD_CBIT(xid, 3), "release_vx_info(%p[#%d.%d.%d]) %p",
9541 +               vxi, vxi ? vxi->vx_id : 0,
9542 +               vxi ? atomic_read(&vxi->vx_usecnt) : 0,
9543 +               vxi ? atomic_read(&vxi->vx_tasks) : 0,
9544 +               task, _file, _line);
9545 +       __vxh_release_vx_info(vxi, task, _here);
9546 +
9547 +       might_sleep();
9548 +
9549 +       if (atomic_dec_and_test(&vxi->vx_tasks))
9550 +               unhash_vx_info(vxi);
9551 +}
9552 +
9553 +
9554 +#define task_get_vx_info(p) \
9555 +       __task_get_vx_info(p, __FILE__, __LINE__, __HERE__)
9556 +
9557 +static inline struct vx_info *__task_get_vx_info(struct task_struct *p,
9558 +       const char *_file, int _line, void *_here)
9559 +{
9560 +       struct vx_info *vxi;
9561 +
9562 +       task_lock(p);
9563 +       vxlprintk(VXD_CBIT(xid, 5), "task_get_vx_info(%p)",
9564 +               p, _file, _line);
9565 +       vxi = __get_vx_info(p->vx_info, _file, _line, _here);
9566 +       task_unlock(p);
9567 +       return vxi;
9568 +}
9569 +
9570 +
9571 +static inline void __wakeup_vx_info(struct vx_info *vxi)
9572 +{
9573 +       if (waitqueue_active(&vxi->vx_wait))
9574 +               wake_up_interruptible(&vxi->vx_wait);
9575 +}
9576 +
9577 +
9578 +#define enter_vx_info(v, s) __enter_vx_info(v, s, __FILE__, __LINE__)
9579 +
9580 +static inline void __enter_vx_info(struct vx_info *vxi,
9581 +       struct vx_info_save *vxis, const char *_file, int _line)
9582 +{
9583 +       vxlprintk(VXD_CBIT(xid, 5), "enter_vx_info(%p[#%d],%p) %p[#%d,%p]",
9584 +               vxi, vxi ? vxi->vx_id : 0, vxis, current,
9585 +               current->xid, current->vx_info, _file, _line);
9586 +       vxis->vxi = xchg(&current->vx_info, vxi);
9587 +       vxis->xid = current->xid;
9588 +       current->xid = vxi ? vxi->vx_id : 0;
9589 +}
9590 +
9591 +#define leave_vx_info(s) __leave_vx_info(s, __FILE__, __LINE__)
9592 +
9593 +static inline void __leave_vx_info(struct vx_info_save *vxis,
9594 +       const char *_file, int _line)
9595 +{
9596 +       vxlprintk(VXD_CBIT(xid, 5), "leave_vx_info(%p[#%d,%p]) %p[#%d,%p]",
9597 +               vxis, vxis->xid, vxis->vxi, current,
9598 +               current->xid, current->vx_info, _file, _line);
9599 +       (void)xchg(&current->vx_info, vxis->vxi);
9600 +       current->xid = vxis->xid;
9601 +}
9602 +
9603 +
9604 +static inline void __enter_vx_admin(struct vx_info_save *vxis)
9605 +{
9606 +       vxis->vxi = xchg(&current->vx_info, NULL);
9607 +       vxis->xid = xchg(&current->xid, (xid_t)0);
9608 +}
9609 +
9610 +static inline void __leave_vx_admin(struct vx_info_save *vxis)
9611 +{
9612 +       (void)xchg(&current->xid, vxis->xid);
9613 +       (void)xchg(&current->vx_info, vxis->vxi);
9614 +}
9615 +
9616 +#define task_is_init(p) \
9617 +       __task_is_init(p, __FILE__, __LINE__, __HERE__)
9618 +
9619 +static inline int __task_is_init(struct task_struct *p,
9620 +       const char *_file, int _line, void *_here)
9621 +{
9622 +       int is_init = is_global_init(p);
9623 +
9624 +       task_lock(p);
9625 +       if (p->vx_info)
9626 +               is_init = p->vx_info->vx_initpid == p->pid;
9627 +       task_unlock(p);
9628 +       return is_init;
9629 +}
9630 +
9631 +extern void exit_vx_info(struct task_struct *, int);
9632 +extern void exit_vx_info_early(struct task_struct *, int);
9633 +
9634 +
9635 +#else
9636 +#warning duplicate inclusion
9637 +#endif
9638 diff -NurpP --minimal linux-3.0-rc5/include/linux/vs_cowbl.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_cowbl.h
9639 --- linux-3.0-rc5/include/linux/vs_cowbl.h      1970-01-01 01:00:00.000000000 +0100
9640 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_cowbl.h 2011-06-10 22:11:24.000000000 +0200
9641 @@ -0,0 +1,48 @@
9642 +#ifndef _VS_COWBL_H
9643 +#define _VS_COWBL_H
9644 +
9645 +#include <linux/fs.h>
9646 +#include <linux/dcache.h>
9647 +#include <linux/namei.h>
9648 +#include <linux/slab.h>
9649 +
9650 +extern struct dentry *cow_break_link(const char *pathname);
9651 +
9652 +static inline int cow_check_and_break(struct path *path)
9653 +{
9654 +       struct inode *inode = path->dentry->d_inode;
9655 +       int error = 0;
9656 +
9657 +       /* do we need this check? */
9658 +       if (IS_RDONLY(inode))
9659 +               return -EROFS;
9660 +
9661 +       if (IS_COW(inode)) {
9662 +               if (IS_COW_LINK(inode)) {
9663 +                       struct dentry *new_dentry, *old_dentry = path->dentry;
9664 +                       char *pp, *buf;
9665 +
9666 +                       buf = kmalloc(PATH_MAX, GFP_KERNEL);
9667 +                       if (!buf) {
9668 +                               return -ENOMEM;
9669 +                       }
9670 +                       pp = d_path(path, buf, PATH_MAX);
9671 +                       new_dentry = cow_break_link(pp);
9672 +                       kfree(buf);
9673 +                       if (!IS_ERR(new_dentry)) {
9674 +                               path->dentry = new_dentry;
9675 +                               dput(old_dentry);
9676 +                       } else
9677 +                               error = PTR_ERR(new_dentry);
9678 +               } else {
9679 +                       inode->i_flags &= ~(S_IXUNLINK | S_IMMUTABLE);
9680 +                       inode->i_ctime = CURRENT_TIME;
9681 +                       mark_inode_dirty(inode);
9682 +               }
9683 +       }
9684 +       return error;
9685 +}
9686 +
9687 +#else
9688 +#warning duplicate inclusion
9689 +#endif
9690 diff -NurpP --minimal linux-3.0-rc5/include/linux/vs_cvirt.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_cvirt.h
9691 --- linux-3.0-rc5/include/linux/vs_cvirt.h      1970-01-01 01:00:00.000000000 +0100
9692 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_cvirt.h 2011-06-10 22:11:24.000000000 +0200
9693 @@ -0,0 +1,50 @@
9694 +#ifndef _VS_CVIRT_H
9695 +#define _VS_CVIRT_H
9696 +
9697 +#include "vserver/cvirt.h"
9698 +#include "vserver/context.h"
9699 +#include "vserver/base.h"
9700 +#include "vserver/check.h"
9701 +#include "vserver/debug.h"
9702 +
9703 +
9704 +static inline void vx_activate_task(struct task_struct *p)
9705 +{
9706 +       struct vx_info *vxi;
9707 +
9708 +       if ((vxi = p->vx_info)) {
9709 +               vx_update_load(vxi);
9710 +               atomic_inc(&vxi->cvirt.nr_running);
9711 +       }
9712 +}
9713 +
9714 +static inline void vx_deactivate_task(struct task_struct *p)
9715 +{
9716 +       struct vx_info *vxi;
9717 +
9718 +       if ((vxi = p->vx_info)) {
9719 +               vx_update_load(vxi);
9720 +               atomic_dec(&vxi->cvirt.nr_running);
9721 +       }
9722 +}
9723 +
9724 +static inline void vx_uninterruptible_inc(struct task_struct *p)
9725 +{
9726 +       struct vx_info *vxi;
9727 +
9728 +       if ((vxi = p->vx_info))
9729 +               atomic_inc(&vxi->cvirt.nr_uninterruptible);
9730 +}
9731 +
9732 +static inline void vx_uninterruptible_dec(struct task_struct *p)
9733 +{
9734 +       struct vx_info *vxi;
9735 +
9736 +       if ((vxi = p->vx_info))
9737 +               atomic_dec(&vxi->cvirt.nr_uninterruptible);
9738 +}
9739 +
9740 +
9741 +#else
9742 +#warning duplicate inclusion
9743 +#endif
9744 diff -NurpP --minimal linux-3.0-rc5/include/linux/vs_device.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_device.h
9745 --- linux-3.0-rc5/include/linux/vs_device.h     1970-01-01 01:00:00.000000000 +0100
9746 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_device.h        2011-06-10 22:11:24.000000000 +0200
9747 @@ -0,0 +1,45 @@
9748 +#ifndef _VS_DEVICE_H
9749 +#define _VS_DEVICE_H
9750 +
9751 +#include "vserver/base.h"
9752 +#include "vserver/device.h"
9753 +#include "vserver/debug.h"
9754 +
9755 +
9756 +#ifdef CONFIG_VSERVER_DEVICE
9757 +
9758 +int vs_map_device(struct vx_info *, dev_t, dev_t *, umode_t);
9759 +
9760 +#define vs_device_perm(v, d, m, p) \
9761 +       ((vs_map_device(current_vx_info(), d, NULL, m) & (p)) == (p))
9762 +
9763 +#else
9764 +
9765 +static inline
9766 +int vs_map_device(struct vx_info *vxi,
9767 +       dev_t device, dev_t *target, umode_t mode)
9768 +{
9769 +       if (target)
9770 +               *target = device;
9771 +       return ~0;
9772 +}
9773 +
9774 +#define vs_device_perm(v, d, m, p) ((p) == (p))
9775 +
9776 +#endif
9777 +
9778 +
9779 +#define vs_map_chrdev(d, t, p) \
9780 +       ((vs_map_device(current_vx_info(), d, t, S_IFCHR) & (p)) == (p))
9781 +#define vs_map_blkdev(d, t, p) \
9782 +       ((vs_map_device(current_vx_info(), d, t, S_IFBLK) & (p)) == (p))
9783 +
9784 +#define vs_chrdev_perm(d, p) \
9785 +       vs_device_perm(current_vx_info(), d, S_IFCHR, p)
9786 +#define vs_blkdev_perm(d, p) \
9787 +       vs_device_perm(current_vx_info(), d, S_IFBLK, p)
9788 +
9789 +
9790 +#else
9791 +#warning duplicate inclusion
9792 +#endif
9793 diff -NurpP --minimal linux-3.0-rc5/include/linux/vs_dlimit.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_dlimit.h
9794 --- linux-3.0-rc5/include/linux/vs_dlimit.h     1970-01-01 01:00:00.000000000 +0100
9795 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_dlimit.h        2011-06-10 22:11:24.000000000 +0200
9796 @@ -0,0 +1,215 @@
9797 +#ifndef _VS_DLIMIT_H
9798 +#define _VS_DLIMIT_H
9799 +
9800 +#include <linux/fs.h>
9801 +
9802 +#include "vserver/dlimit.h"
9803 +#include "vserver/base.h"
9804 +#include "vserver/debug.h"
9805 +
9806 +
9807 +#define get_dl_info(i) __get_dl_info(i, __FILE__, __LINE__)
9808 +
9809 +static inline struct dl_info *__get_dl_info(struct dl_info *dli,
9810 +       const char *_file, int _line)
9811 +{
9812 +       if (!dli)
9813 +               return NULL;
9814 +       vxlprintk(VXD_CBIT(dlim, 4), "get_dl_info(%p[#%d.%d])",
9815 +               dli, dli ? dli->dl_tag : 0,
9816 +               dli ? atomic_read(&dli->dl_usecnt) : 0,
9817 +               _file, _line);
9818 +       atomic_inc(&dli->dl_usecnt);
9819 +       return dli;
9820 +}
9821 +
9822 +
9823 +#define free_dl_info(i) \
9824 +       call_rcu(&(i)->dl_rcu, rcu_free_dl_info)
9825 +
9826 +#define put_dl_info(i) __put_dl_info(i, __FILE__, __LINE__)
9827 +
9828 +static inline void __put_dl_info(struct dl_info *dli,
9829 +       const char *_file, int _line)
9830 +{
9831 +       if (!dli)
9832 +               return;
9833 +       vxlprintk(VXD_CBIT(dlim, 4), "put_dl_info(%p[#%d.%d])",
9834 +               dli, dli ? dli->dl_tag : 0,
9835 +               dli ? atomic_read(&dli->dl_usecnt) : 0,
9836 +               _file, _line);
9837 +       if (atomic_dec_and_test(&dli->dl_usecnt))
9838 +               free_dl_info(dli);
9839 +}
9840 +
9841 +
9842 +#define __dlimit_char(d)       ((d) ? '*' : ' ')
9843 +
9844 +static inline int __dl_alloc_space(struct super_block *sb,
9845 +       tag_t tag, dlsize_t nr, const char *file, int line)
9846 +{
9847 +       struct dl_info *dli = NULL;
9848 +       int ret = 0;
9849 +
9850 +       if (nr == 0)
9851 +               goto out;
9852 +       dli = locate_dl_info(sb, tag);
9853 +       if (!dli)
9854 +               goto out;
9855 +
9856 +       spin_lock(&dli->dl_lock);
9857 +       ret = (dli->dl_space_used + nr > dli->dl_space_total);
9858 +       if (!ret)
9859 +               dli->dl_space_used += nr;
9860 +       spin_unlock(&dli->dl_lock);
9861 +       put_dl_info(dli);
9862 +out:
9863 +       vxlprintk(VXD_CBIT(dlim, 1),
9864 +               "ALLOC (%p,#%d)%c %lld bytes (%d)",
9865 +               sb, tag, __dlimit_char(dli), (long long)nr,
9866 +               ret, file, line);
9867 +       return ret ? -ENOSPC : 0;
9868 +}
9869 +
9870 +static inline void __dl_free_space(struct super_block *sb,
9871 +       tag_t tag, dlsize_t nr, const char *_file, int _line)
9872 +{
9873 +       struct dl_info *dli = NULL;
9874 +
9875 +       if (nr == 0)
9876 +               goto out;
9877 +       dli = locate_dl_info(sb, tag);
9878 +       if (!dli)
9879 +               goto out;
9880 +
9881 +       spin_lock(&dli->dl_lock);
9882 +       if (dli->dl_space_used > nr)
9883 +               dli->dl_space_used -= nr;
9884 +       else
9885 +               dli->dl_space_used = 0;
9886 +       spin_unlock(&dli->dl_lock);
9887 +       put_dl_info(dli);
9888 +out:
9889 +       vxlprintk(VXD_CBIT(dlim, 1),
9890 +               "FREE  (%p,#%d)%c %lld bytes",
9891 +               sb, tag, __dlimit_char(dli), (long long)nr,
9892 +               _file, _line);
9893 +}
9894 +
9895 +static inline int __dl_alloc_inode(struct super_block *sb,
9896 +       tag_t tag, const char *_file, int _line)
9897 +{
9898 +       struct dl_info *dli;
9899 +       int ret = 0;
9900 +
9901 +       dli = locate_dl_info(sb, tag);
9902 +       if (!dli)
9903 +               goto out;
9904 +
9905 +       spin_lock(&dli->dl_lock);
9906 +       dli->dl_inodes_used++;
9907 +       ret = (dli->dl_inodes_used > dli->dl_inodes_total);
9908 +       spin_unlock(&dli->dl_lock);
9909 +       put_dl_info(dli);
9910 +out:
9911 +       vxlprintk(VXD_CBIT(dlim, 0),
9912 +               "ALLOC (%p,#%d)%c inode (%d)",
9913 +               sb, tag, __dlimit_char(dli), ret, _file, _line);
9914 +       return ret ? -ENOSPC : 0;
9915 +}
9916 +
9917 +static inline void __dl_free_inode(struct super_block *sb,
9918 +       tag_t tag, const char *_file, int _line)
9919 +{
9920 +       struct dl_info *dli;
9921 +
9922 +       dli = locate_dl_info(sb, tag);
9923 +       if (!dli)
9924 +               goto out;
9925 +
9926 +       spin_lock(&dli->dl_lock);
9927 +       if (dli->dl_inodes_used > 1)
9928 +               dli->dl_inodes_used--;
9929 +       else
9930 +               dli->dl_inodes_used = 0;
9931 +       spin_unlock(&dli->dl_lock);
9932 +       put_dl_info(dli);
9933 +out:
9934 +       vxlprintk(VXD_CBIT(dlim, 0),
9935 +               "FREE  (%p,#%d)%c inode",
9936 +               sb, tag, __dlimit_char(dli), _file, _line);
9937 +}
9938 +
9939 +static inline void __dl_adjust_block(struct super_block *sb, tag_t tag,
9940 +       unsigned long long *free_blocks, unsigned long long *root_blocks,
9941 +       const char *_file, int _line)
9942 +{
9943 +       struct dl_info *dli;
9944 +       uint64_t broot, bfree;
9945 +
9946 +       dli = locate_dl_info(sb, tag);
9947 +       if (!dli)
9948 +               return;
9949 +
9950 +       spin_lock(&dli->dl_lock);
9951 +       broot = (dli->dl_space_total -
9952 +               (dli->dl_space_total >> 10) * dli->dl_nrlmult)
9953 +               >> sb->s_blocksize_bits;
9954 +       bfree = (dli->dl_space_total - dli->dl_space_used)
9955 +                       >> sb->s_blocksize_bits;
9956 +       spin_unlock(&dli->dl_lock);
9957 +
9958 +       vxlprintk(VXD_CBIT(dlim, 2),
9959 +               "ADJUST: %lld,%lld on %lld,%lld [mult=%d]",
9960 +               (long long)bfree, (long long)broot,
9961 +               *free_blocks, *root_blocks, dli->dl_nrlmult,
9962 +               _file, _line);
9963 +       if (free_blocks) {
9964 +               if (*free_blocks > bfree)
9965 +                       *free_blocks = bfree;
9966 +       }
9967 +       if (root_blocks) {
9968 +               if (*root_blocks > broot)
9969 +                       *root_blocks = broot;
9970 +       }
9971 +       put_dl_info(dli);
9972 +}
9973 +
9974 +#define dl_prealloc_space(in, bytes) \
9975 +       __dl_alloc_space((in)->i_sb, (in)->i_tag, (dlsize_t)(bytes), \
9976 +               __FILE__, __LINE__ )
9977 +
9978 +#define dl_alloc_space(in, bytes) \
9979 +       __dl_alloc_space((in)->i_sb, (in)->i_tag, (dlsize_t)(bytes), \
9980 +               __FILE__, __LINE__ )
9981 +
9982 +#define dl_reserve_space(in, bytes) \
9983 +       __dl_alloc_space((in)->i_sb, (in)->i_tag, (dlsize_t)(bytes), \
9984 +               __FILE__, __LINE__ )
9985 +
9986 +#define dl_claim_space(in, bytes) (0)
9987 +
9988 +#define dl_release_space(in, bytes) \
9989 +       __dl_free_space((in)->i_sb, (in)->i_tag, (dlsize_t)(bytes), \
9990 +               __FILE__, __LINE__ )
9991 +
9992 +#define dl_free_space(in, bytes) \
9993 +       __dl_free_space((in)->i_sb, (in)->i_tag, (dlsize_t)(bytes), \
9994 +               __FILE__, __LINE__ )
9995 +
9996 +
9997 +
9998 +#define dl_alloc_inode(in) \
9999 +       __dl_alloc_inode((in)->i_sb, (in)->i_tag, __FILE__, __LINE__ )
10000 +
10001 +#define dl_free_inode(in) \
10002 +       __dl_free_inode((in)->i_sb, (in)->i_tag, __FILE__, __LINE__ )
10003 +
10004 +
10005 +#define dl_adjust_block(sb, tag, fb, rb) \
10006 +       __dl_adjust_block(sb, tag, fb, rb, __FILE__, __LINE__ )
10007 +
10008 +
10009 +#else
10010 +#warning duplicate inclusion
10011 +#endif
10012 diff -NurpP --minimal linux-3.0-rc5/include/linux/vs_inet.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_inet.h
10013 --- linux-3.0-rc5/include/linux/vs_inet.h       1970-01-01 01:00:00.000000000 +0100
10014 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_inet.h  2011-06-10 22:11:24.000000000 +0200
10015 @@ -0,0 +1,351 @@
10016 +#ifndef _VS_INET_H
10017 +#define _VS_INET_H
10018 +
10019 +#include "vserver/base.h"
10020 +#include "vserver/network.h"
10021 +#include "vserver/debug.h"
10022 +
10023 +#define IPI_LOOPBACK   htonl(INADDR_LOOPBACK)
10024 +
10025 +#define NXAV4(a)       NIPQUAD((a)->ip[0]), NIPQUAD((a)->ip[1]), \
10026 +                       NIPQUAD((a)->mask), (a)->type
10027 +#define NXAV4_FMT      "[" NIPQUAD_FMT "-" NIPQUAD_FMT "/" NIPQUAD_FMT ":%04x]"
10028 +
10029 +#define NIPQUAD(addr) \
10030 +       ((unsigned char *)&addr)[0], \
10031 +       ((unsigned char *)&addr)[1], \
10032 +       ((unsigned char *)&addr)[2], \
10033 +       ((unsigned char *)&addr)[3]
10034 +
10035 +#define NIPQUAD_FMT "%u.%u.%u.%u"
10036 +
10037 +
10038 +static inline
10039 +int v4_addr_match(struct nx_addr_v4 *nxa, __be32 addr, uint16_t tmask)
10040 +{
10041 +       __be32 ip = nxa->ip[0].s_addr;
10042 +       __be32 mask = nxa->mask.s_addr;
10043 +       __be32 bcast = ip | ~mask;
10044 +       int ret = 0;
10045 +
10046 +       switch (nxa->type & tmask) {
10047 +       case NXA_TYPE_MASK:
10048 +               ret = (ip == (addr & mask));
10049 +               break;
10050 +       case NXA_TYPE_ADDR:
10051 +               ret = 3;
10052 +               if (addr == ip)
10053 +                       break;
10054 +               /* fall through to broadcast */
10055 +       case NXA_MOD_BCAST:
10056 +               ret = ((tmask & NXA_MOD_BCAST) && (addr == bcast));
10057 +               break;
10058 +       case NXA_TYPE_RANGE:
10059 +               ret = ((nxa->ip[0].s_addr <= addr) &&
10060 +                       (nxa->ip[1].s_addr > addr));
10061 +               break;
10062 +       case NXA_TYPE_ANY:
10063 +               ret = 2;
10064 +               break;
10065 +       }
10066 +
10067 +       vxdprintk(VXD_CBIT(net, 0),
10068 +               "v4_addr_match(%p" NXAV4_FMT "," NIPQUAD_FMT ",%04x) = %d",
10069 +               nxa, NXAV4(nxa), NIPQUAD(addr), tmask, ret);
10070 +       return ret;
10071 +}
10072 +
10073 +static inline
10074 +int v4_addr_in_nx_info(struct nx_info *nxi, __be32 addr, uint16_t tmask)
10075 +{
10076 +       struct nx_addr_v4 *nxa;
10077 +       int ret = 1;
10078 +
10079 +       if (!nxi)
10080 +               goto out;
10081 +
10082 +       ret = 2;
10083 +       /* allow 127.0.0.1 when remapping lback */
10084 +       if ((tmask & NXA_LOOPBACK) &&
10085 +               (addr == IPI_LOOPBACK) &&
10086 +               nx_info_flags(nxi, NXF_LBACK_REMAP, 0))
10087 +               goto out;
10088 +       ret = 3;
10089 +       /* check for lback address */
10090 +       if ((tmask & NXA_MOD_LBACK) &&
10091 +               (nxi->v4_lback.s_addr == addr))
10092 +               goto out;
10093 +       ret = 4;
10094 +       /* check for broadcast address */
10095 +       if ((tmask & NXA_MOD_BCAST) &&
10096 +               (nxi->v4_bcast.s_addr == addr))
10097 +               goto out;
10098 +       ret = 5;
10099 +       /* check for v4 addresses */
10100 +       for (nxa = &nxi->v4; nxa; nxa = nxa->next)
10101 +               if (v4_addr_match(nxa, addr, tmask))
10102 +                       goto out;
10103 +       ret = 0;
10104 +out:
10105 +       vxdprintk(VXD_CBIT(net, 0),
10106 +               "v4_addr_in_nx_info(%p[#%u]," NIPQUAD_FMT ",%04x) = %d",
10107 +               nxi, nxi ? nxi->nx_id : 0, NIPQUAD(addr), tmask, ret);
10108 +       return ret;
10109 +}
10110 +
10111 +static inline
10112 +int v4_nx_addr_match(struct nx_addr_v4 *nxa, struct nx_addr_v4 *addr, uint16_t mask)
10113 +{
10114 +       /* FIXME: needs full range checks */
10115 +       return v4_addr_match(nxa, addr->ip[0].s_addr, mask);
10116 +}
10117 +
10118 +static inline
10119 +int v4_nx_addr_in_nx_info(struct nx_info *nxi, struct nx_addr_v4 *nxa, uint16_t mask)
10120 +{
10121 +       struct nx_addr_v4 *ptr;
10122 +
10123 +       for (ptr = &nxi->v4; ptr; ptr = ptr->next)
10124 +               if (v4_nx_addr_match(ptr, nxa, mask))
10125 +                       return 1;
10126 +       return 0;
10127 +}
10128 +
10129 +#include <net/inet_sock.h>
10130 +
10131 +/*
10132 + *     Check if a given address matches for a socket
10133 + *
10134 + *     nxi:            the socket's nx_info if any
10135 + *     addr:           to be verified address
10136 + */
10137 +static inline
10138 +int v4_sock_addr_match (
10139 +       struct nx_info *nxi,
10140 +       struct inet_sock *inet,
10141 +       __be32 addr)
10142 +{
10143 +       __be32 saddr = inet->inet_rcv_saddr;
10144 +       __be32 bcast = nxi ? nxi->v4_bcast.s_addr : INADDR_BROADCAST;
10145 +
10146 +       if (addr && (saddr == addr || bcast == addr))
10147 +               return 1;
10148 +       if (!saddr)
10149 +               return v4_addr_in_nx_info(nxi, addr, NXA_MASK_BIND);
10150 +       return 0;
10151 +}
10152 +
10153 +
10154 +/* inet related checks and helpers */
10155 +
10156 +
10157 +struct in_ifaddr;
10158 +struct net_device;
10159 +struct sock;
10160 +
10161 +#ifdef CONFIG_INET
10162 +
10163 +#include <linux/netdevice.h>
10164 +#include <linux/inetdevice.h>
10165 +#include <net/inet_sock.h>
10166 +#include <net/inet_timewait_sock.h>
10167 +
10168 +
10169 +int dev_in_nx_info(struct net_device *, struct nx_info *);
10170 +int v4_dev_in_nx_info(struct net_device *, struct nx_info *);
10171 +int nx_v4_addr_conflict(struct nx_info *, struct nx_info *);
10172 +
10173 +
10174 +/*
10175 + *     check if address is covered by socket
10176 + *
10177 + *     sk:     the socket to check against
10178 + *     addr:   the address in question (must be != 0)
10179 + */
10180 +
10181 +static inline
10182 +int __v4_addr_match_socket(const struct sock *sk, struct nx_addr_v4 *nxa)
10183 +{
10184 +       struct nx_info *nxi = sk->sk_nx_info;
10185 +       __be32 saddr = sk_rcv_saddr(sk);
10186 +
10187 +       vxdprintk(VXD_CBIT(net, 5),
10188 +               "__v4_addr_in_socket(%p," NXAV4_FMT ") %p:" NIPQUAD_FMT " %p;%lx",
10189 +               sk, NXAV4(nxa), nxi, NIPQUAD(saddr), sk->sk_socket,
10190 +               (sk->sk_socket?sk->sk_socket->flags:0));
10191 +
10192 +       if (saddr) {            /* direct address match */
10193 +               return v4_addr_match(nxa, saddr, -1);
10194 +       } else if (nxi) {       /* match against nx_info */
10195 +               return v4_nx_addr_in_nx_info(nxi, nxa, -1);
10196 +       } else {                /* unrestricted any socket */
10197 +               return 1;
10198 +       }
10199 +}
10200 +
10201 +
10202 +
10203 +static inline
10204 +int nx_dev_visible(struct nx_info *nxi, struct net_device *dev)
10205 +{
10206 +       vxdprintk(VXD_CBIT(net, 1),
10207 +               "nx_dev_visible(%p[#%u],%p " VS_Q("%s") ") %d",
10208 +               nxi, nxi ? nxi->nx_id : 0, dev, dev->name,
10209 +               nxi ? dev_in_nx_info(dev, nxi) : 0);
10210 +
10211 +       if (!nx_info_flags(nxi, NXF_HIDE_NETIF, 0))
10212 +               return 1;
10213 +       if (dev_in_nx_info(dev, nxi))
10214 +               return 1;
10215 +       return 0;
10216 +}
10217 +
10218 +
10219 +static inline
10220 +int v4_ifa_in_nx_info(struct in_ifaddr *ifa, struct nx_info *nxi)
10221 +{
10222 +       if (!nxi)
10223 +               return 1;
10224 +       if (!ifa)
10225 +               return 0;
10226 +       return v4_addr_in_nx_info(nxi, ifa->ifa_local, NXA_MASK_SHOW);
10227 +}
10228 +
10229 +static inline
10230 +int nx_v4_ifa_visible(struct nx_info *nxi, struct in_ifaddr *ifa)
10231 +{
10232 +       vxdprintk(VXD_CBIT(net, 1), "nx_v4_ifa_visible(%p[#%u],%p) %d",
10233 +               nxi, nxi ? nxi->nx_id : 0, ifa,
10234 +               nxi ? v4_ifa_in_nx_info(ifa, nxi) : 0);
10235 +
10236 +       if (!nx_info_flags(nxi, NXF_HIDE_NETIF, 0))
10237 +               return 1;
10238 +       if (v4_ifa_in_nx_info(ifa, nxi))
10239 +               return 1;
10240 +       return 0;
10241 +}
10242 +
10243 +
10244 +struct nx_v4_sock_addr {
10245 +       __be32 saddr;   /* Address used for validation */
10246 +       __be32 baddr;   /* Address used for socket bind */
10247 +};
10248 +
10249 +static inline
10250 +int v4_map_sock_addr(struct inet_sock *inet, struct sockaddr_in *addr,
10251 +       struct nx_v4_sock_addr *nsa)
10252 +{
10253 +       struct sock *sk = &inet->sk;
10254 +       struct nx_info *nxi = sk->sk_nx_info;
10255 +       __be32 saddr = addr->sin_addr.s_addr;
10256 +       __be32 baddr = saddr;
10257 +
10258 +       vxdprintk(VXD_CBIT(net, 3),
10259 +               "inet_bind(%p)* %p,%p;%lx " NIPQUAD_FMT,
10260 +               sk, sk->sk_nx_info, sk->sk_socket,
10261 +               (sk->sk_socket ? sk->sk_socket->flags : 0),
10262 +               NIPQUAD(saddr));
10263 +
10264 +       if (nxi) {
10265 +               if (saddr == INADDR_ANY) {
10266 +                       if (nx_info_flags(nxi, NXF_SINGLE_IP, 0))
10267 +                               baddr = nxi->v4.ip[0].s_addr;
10268 +               } else if (saddr == IPI_LOOPBACK) {
10269 +                       if (nx_info_flags(nxi, NXF_LBACK_REMAP, 0))
10270 +                               baddr = nxi->v4_lback.s_addr;
10271 +               } else {        /* normal address bind */
10272 +                       if (!v4_addr_in_nx_info(nxi, saddr, NXA_MASK_BIND))
10273 +                               return -EADDRNOTAVAIL;
10274 +               }
10275 +       }
10276 +
10277 +       vxdprintk(VXD_CBIT(net, 3),
10278 +               "inet_bind(%p) " NIPQUAD_FMT ", " NIPQUAD_FMT,
10279 +               sk, NIPQUAD(saddr), NIPQUAD(baddr));
10280 +
10281 +       nsa->saddr = saddr;
10282 +       nsa->baddr = baddr;
10283 +       return 0;
10284 +}
10285 +
10286 +static inline
10287 +void v4_set_sock_addr(struct inet_sock *inet, struct nx_v4_sock_addr *nsa)
10288 +{
10289 +       inet->inet_saddr = nsa->baddr;
10290 +       inet->inet_rcv_saddr = nsa->baddr;
10291 +}
10292 +
10293 +
10294 +/*
10295 + *      helper to simplify inet_lookup_listener
10296 + *
10297 + *      nxi:   the socket's nx_info if any
10298 + *      addr:  to be verified address
10299 + *      saddr: socket address
10300 + */
10301 +static inline int v4_inet_addr_match (
10302 +       struct nx_info *nxi,
10303 +       __be32 addr,
10304 +       __be32 saddr)
10305 +{
10306 +       if (addr && (saddr == addr))
10307 +               return 1;
10308 +       if (!saddr)
10309 +               return nxi ? v4_addr_in_nx_info(nxi, addr, NXA_MASK_BIND) : 1;
10310 +       return 0;
10311 +}
10312 +
10313 +static inline __be32 nx_map_sock_lback(struct nx_info *nxi, __be32 addr)
10314 +{
10315 +       if (nx_info_flags(nxi, NXF_HIDE_LBACK, 0) &&
10316 +               (addr == nxi->v4_lback.s_addr))
10317 +               return IPI_LOOPBACK;
10318 +       return addr;
10319 +}
10320 +
10321 +static inline
10322 +int nx_info_has_v4(struct nx_info *nxi)
10323 +{
10324 +       if (!nxi)
10325 +               return 1;
10326 +       if (NX_IPV4(nxi))
10327 +               return 1;
10328 +       if (nx_info_flags(nxi, NXF_LBACK_REMAP, 0))
10329 +               return 1;
10330 +       return 0;
10331 +}
10332 +
10333 +#else /* CONFIG_INET */
10334 +
10335 +static inline
10336 +int nx_dev_visible(struct nx_info *n, struct net_device *d)
10337 +{
10338 +       return 1;
10339 +}
10340 +
10341 +static inline
10342 +int nx_v4_addr_conflict(struct nx_info *n, uint32_t a, const struct sock *s)
10343 +{
10344 +       return 1;
10345 +}
10346 +
10347 +static inline
10348 +int v4_ifa_in_nx_info(struct in_ifaddr *a, struct nx_info *n)
10349 +{
10350 +       return 1;
10351 +}
10352 +
10353 +static inline
10354 +int nx_info_has_v4(struct nx_info *nxi)
10355 +{
10356 +       return 0;
10357 +}
10358 +
10359 +#endif /* CONFIG_INET */
10360 +
10361 +#define current_nx_info_has_v4() \
10362 +       nx_info_has_v4(current_nx_info())
10363 +
10364 +#else
10365 +// #warning duplicate inclusion
10366 +#endif
10367 diff -NurpP --minimal linux-3.0-rc5/include/linux/vs_inet6.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_inet6.h
10368 --- linux-3.0-rc5/include/linux/vs_inet6.h      1970-01-01 01:00:00.000000000 +0100
10369 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_inet6.h 2011-06-10 22:11:24.000000000 +0200
10370 @@ -0,0 +1,246 @@
10371 +#ifndef _VS_INET6_H
10372 +#define _VS_INET6_H
10373 +
10374 +#include "vserver/base.h"
10375 +#include "vserver/network.h"
10376 +#include "vserver/debug.h"
10377 +
10378 +#include <net/ipv6.h>
10379 +
10380 +#define NXAV6(a)       &(a)->ip, &(a)->mask, (a)->prefix, (a)->type
10381 +#define NXAV6_FMT      "[%pI6/%pI6/%d:%04x]"
10382 +
10383 +
10384 +#ifdef CONFIG_IPV6
10385 +
10386 +static inline
10387 +int v6_addr_match(struct nx_addr_v6 *nxa,
10388 +       const struct in6_addr *addr, uint16_t mask)
10389 +{
10390 +       int ret = 0;
10391 +
10392 +       switch (nxa->type & mask) {
10393 +       case NXA_TYPE_MASK:
10394 +               ret = ipv6_masked_addr_cmp(&nxa->ip, &nxa->mask, addr);
10395 +               break;
10396 +       case NXA_TYPE_ADDR:
10397 +               ret = ipv6_addr_equal(&nxa->ip, addr);
10398 +               break;
10399 +       case NXA_TYPE_ANY:
10400 +               ret = 1;
10401 +               break;
10402 +       }
10403 +       vxdprintk(VXD_CBIT(net, 0),
10404 +               "v6_addr_match(%p" NXAV6_FMT ",%pI6,%04x) = %d",
10405 +               nxa, NXAV6(nxa), addr, mask, ret);
10406 +       return ret;
10407 +}
10408 +
10409 +static inline
10410 +int v6_addr_in_nx_info(struct nx_info *nxi,
10411 +       const struct in6_addr *addr, uint16_t mask)
10412 +{
10413 +       struct nx_addr_v6 *nxa;
10414 +       int ret = 1;
10415 +
10416 +       if (!nxi)
10417 +               goto out;
10418 +       for (nxa = &nxi->v6; nxa; nxa = nxa->next)
10419 +               if (v6_addr_match(nxa, addr, mask))
10420 +                       goto out;
10421 +       ret = 0;
10422 +out:
10423 +       vxdprintk(VXD_CBIT(net, 0),
10424 +               "v6_addr_in_nx_info(%p[#%u],%pI6,%04x) = %d",
10425 +               nxi, nxi ? nxi->nx_id : 0, addr, mask, ret);
10426 +       return ret;
10427 +}
10428 +
10429 +static inline
10430 +int v6_nx_addr_match(struct nx_addr_v6 *nxa, struct nx_addr_v6 *addr, uint16_t mask)
10431 +{
10432 +       /* FIXME: needs full range checks */
10433 +       return v6_addr_match(nxa, &addr->ip, mask);
10434 +}
10435 +
10436 +static inline
10437 +int v6_nx_addr_in_nx_info(struct nx_info *nxi, struct nx_addr_v6 *nxa, uint16_t mask)
10438 +{
10439 +       struct nx_addr_v6 *ptr;
10440 +
10441 +       for (ptr = &nxi->v6; ptr; ptr = ptr->next)
10442 +               if (v6_nx_addr_match(ptr, nxa, mask))
10443 +                       return 1;
10444 +       return 0;
10445 +}
10446 +
10447 +
10448 +/*
10449 + *     Check if a given address matches for a socket
10450 + *
10451 + *     nxi:            the socket's nx_info if any
10452 + *     addr:           to be verified address
10453 + */
10454 +static inline
10455 +int v6_sock_addr_match (
10456 +       struct nx_info *nxi,
10457 +       struct inet_sock *inet,
10458 +       struct in6_addr *addr)
10459 +{
10460 +       struct sock *sk = &inet->sk;
10461 +       struct in6_addr *saddr = inet6_rcv_saddr(sk);
10462 +
10463 +       if (!ipv6_addr_any(addr) &&
10464 +               ipv6_addr_equal(saddr, addr))
10465 +               return 1;
10466 +       if (ipv6_addr_any(saddr))
10467 +               return v6_addr_in_nx_info(nxi, addr, -1);
10468 +       return 0;
10469 +}
10470 +
10471 +/*
10472 + *     check if address is covered by socket
10473 + *
10474 + *     sk:     the socket to check against
10475 + *     addr:   the address in question (must be != 0)
10476 + */
10477 +
10478 +static inline
10479 +int __v6_addr_match_socket(const struct sock *sk, struct nx_addr_v6 *nxa)
10480 +{
10481 +       struct nx_info *nxi = sk->sk_nx_info;
10482 +       struct in6_addr *saddr = inet6_rcv_saddr(sk);
10483 +
10484 +       vxdprintk(VXD_CBIT(net, 5),
10485 +               "__v6_addr_in_socket(%p," NXAV6_FMT ") %p:%pI6 %p;%lx",
10486 +               sk, NXAV6(nxa), nxi, saddr, sk->sk_socket,
10487 +               (sk->sk_socket?sk->sk_socket->flags:0));
10488 +
10489 +       if (!ipv6_addr_any(saddr)) {    /* direct address match */
10490 +               return v6_addr_match(nxa, saddr, -1);
10491 +       } else if (nxi) {               /* match against nx_info */
10492 +               return v6_nx_addr_in_nx_info(nxi, nxa, -1);
10493 +       } else {                        /* unrestricted any socket */
10494 +               return 1;
10495 +       }
10496 +}
10497 +
10498 +
10499 +/* inet related checks and helpers */
10500 +
10501 +
10502 +struct in_ifaddr;
10503 +struct net_device;
10504 +struct sock;
10505 +
10506 +
10507 +#include <linux/netdevice.h>
10508 +#include <linux/inetdevice.h>
10509 +#include <net/inet_timewait_sock.h>
10510 +
10511 +
10512 +int dev_in_nx_info(struct net_device *, struct nx_info *);
10513 +int v6_dev_in_nx_info(struct net_device *, struct nx_info *);
10514 +int nx_v6_addr_conflict(struct nx_info *, struct nx_info *);
10515 +
10516 +
10517 +
10518 +static inline
10519 +int v6_ifa_in_nx_info(struct inet6_ifaddr *ifa, struct nx_info *nxi)
10520 +{
10521 +       if (!nxi)
10522 +               return 1;
10523 +       if (!ifa)
10524 +               return 0;
10525 +       return v6_addr_in_nx_info(nxi, &ifa->addr, -1);
10526 +}
10527 +
10528 +static inline
10529 +int nx_v6_ifa_visible(struct nx_info *nxi, struct inet6_ifaddr *ifa)
10530 +{
10531 +       vxdprintk(VXD_CBIT(net, 1), "nx_v6_ifa_visible(%p[#%u],%p) %d",
10532 +               nxi, nxi ? nxi->nx_id : 0, ifa,
10533 +               nxi ? v6_ifa_in_nx_info(ifa, nxi) : 0);
10534 +
10535 +       if (!nx_info_flags(nxi, NXF_HIDE_NETIF, 0))
10536 +               return 1;
10537 +       if (v6_ifa_in_nx_info(ifa, nxi))
10538 +               return 1;
10539 +       return 0;
10540 +}
10541 +
10542 +
10543 +struct nx_v6_sock_addr {
10544 +       struct in6_addr saddr;  /* Address used for validation */
10545 +       struct in6_addr baddr;  /* Address used for socket bind */
10546 +};
10547 +
10548 +static inline
10549 +int v6_map_sock_addr(struct inet_sock *inet, struct sockaddr_in6 *addr,
10550 +       struct nx_v6_sock_addr *nsa)
10551 +{
10552 +       // struct sock *sk = &inet->sk;
10553 +       // struct nx_info *nxi = sk->sk_nx_info;
10554 +       struct in6_addr saddr = addr->sin6_addr;
10555 +       struct in6_addr baddr = saddr;
10556 +
10557 +       nsa->saddr = saddr;
10558 +       nsa->baddr = baddr;
10559 +       return 0;
10560 +}
10561 +
10562 +static inline
10563 +void v6_set_sock_addr(struct inet_sock *inet, struct nx_v6_sock_addr *nsa)
10564 +{
10565 +       // struct sock *sk = &inet->sk;
10566 +       // struct in6_addr *saddr = inet6_rcv_saddr(sk);
10567 +
10568 +       // *saddr = nsa->baddr;
10569 +       // inet->inet_saddr = nsa->baddr;
10570 +}
10571 +
10572 +static inline
10573 +int nx_info_has_v6(struct nx_info *nxi)
10574 +{
10575 +       if (!nxi)
10576 +               return 1;
10577 +       if (NX_IPV6(nxi))
10578 +               return 1;
10579 +       return 0;
10580 +}
10581 +
10582 +#else /* CONFIG_IPV6 */
10583 +
10584 +static inline
10585 +int nx_v6_dev_visible(struct nx_info *n, struct net_device *d)
10586 +{
10587 +       return 1;
10588 +}
10589 +
10590 +
10591 +static inline
10592 +int nx_v6_addr_conflict(struct nx_info *n, uint32_t a, const struct sock *s)
10593 +{
10594 +       return 1;
10595 +}
10596 +
10597 +static inline
10598 +int v6_ifa_in_nx_info(struct in_ifaddr *a, struct nx_info *n)
10599 +{
10600 +       return 1;
10601 +}
10602 +
10603 +static inline
10604 +int nx_info_has_v6(struct nx_info *nxi)
10605 +{
10606 +       return 0;
10607 +}
10608 +
10609 +#endif /* CONFIG_IPV6 */
10610 +
10611 +#define current_nx_info_has_v6() \
10612 +       nx_info_has_v6(current_nx_info())
10613 +
10614 +#else
10615 +#warning duplicate inclusion
10616 +#endif
10617 diff -NurpP --minimal linux-3.0-rc5/include/linux/vs_limit.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_limit.h
10618 --- linux-3.0-rc5/include/linux/vs_limit.h      1970-01-01 01:00:00.000000000 +0100
10619 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_limit.h 2011-06-10 22:11:24.000000000 +0200
10620 @@ -0,0 +1,140 @@
10621 +#ifndef _VS_LIMIT_H
10622 +#define _VS_LIMIT_H
10623 +
10624 +#include "vserver/limit.h"
10625 +#include "vserver/base.h"
10626 +#include "vserver/context.h"
10627 +#include "vserver/debug.h"
10628 +#include "vserver/context.h"
10629 +#include "vserver/limit_int.h"
10630 +
10631 +
10632 +#define vx_acc_cres(v, d, p, r) \
10633 +       __vx_acc_cres(v, r, d, p, __FILE__, __LINE__)
10634 +
10635 +#define vx_acc_cres_cond(x, d, p, r) \
10636 +       __vx_acc_cres(((x) == vx_current_xid()) ? current_vx_info() : 0, \
10637 +       r, d, p, __FILE__, __LINE__)
10638 +
10639 +
10640 +#define vx_add_cres(v, a, p, r) \
10641 +       __vx_add_cres(v, r, a, p, __FILE__, __LINE__)
10642 +#define vx_sub_cres(v, a, p, r)                vx_add_cres(v, -(a), p, r)
10643 +
10644 +#define vx_add_cres_cond(x, a, p, r) \
10645 +       __vx_add_cres(((x) == vx_current_xid()) ? current_vx_info() : 0, \
10646 +       r, a, p, __FILE__, __LINE__)
10647 +#define vx_sub_cres_cond(x, a, p, r)   vx_add_cres_cond(x, -(a), p, r)
10648 +
10649 +
10650 +/* process and file limits */
10651 +
10652 +#define vx_nproc_inc(p) \
10653 +       vx_acc_cres((p)->vx_info, 1, p, RLIMIT_NPROC)
10654 +
10655 +#define vx_nproc_dec(p) \
10656 +       vx_acc_cres((p)->vx_info,-1, p, RLIMIT_NPROC)
10657 +
10658 +#define vx_files_inc(f) \
10659 +       vx_acc_cres_cond((f)->f_xid, 1, f, RLIMIT_NOFILE)
10660 +
10661 +#define vx_files_dec(f) \
10662 +       vx_acc_cres_cond((f)->f_xid,-1, f, RLIMIT_NOFILE)
10663 +
10664 +#define vx_locks_inc(l) \
10665 +       vx_acc_cres_cond((l)->fl_xid, 1, l, RLIMIT_LOCKS)
10666 +
10667 +#define vx_locks_dec(l) \
10668 +       vx_acc_cres_cond((l)->fl_xid,-1, l, RLIMIT_LOCKS)
10669 +
10670 +#define vx_openfd_inc(f) \
10671 +       vx_acc_cres(current_vx_info(), 1, (void *)(long)(f), VLIMIT_OPENFD)
10672 +
10673 +#define vx_openfd_dec(f) \
10674 +       vx_acc_cres(current_vx_info(),-1, (void *)(long)(f), VLIMIT_OPENFD)
10675 +
10676 +
10677 +#define vx_cres_avail(v, n, r) \
10678 +       __vx_cres_avail(v, r, n, __FILE__, __LINE__)
10679 +
10680 +
10681 +#define vx_nproc_avail(n) \
10682 +       vx_cres_avail(current_vx_info(), n, RLIMIT_NPROC)
10683 +
10684 +#define vx_files_avail(n) \
10685 +       vx_cres_avail(current_vx_info(), n, RLIMIT_NOFILE)
10686 +
10687 +#define vx_locks_avail(n) \
10688 +       vx_cres_avail(current_vx_info(), n, RLIMIT_LOCKS)
10689 +
10690 +#define vx_openfd_avail(n) \
10691 +       vx_cres_avail(current_vx_info(), n, VLIMIT_OPENFD)
10692 +
10693 +
10694 +/* dentry limits */
10695 +
10696 +#define vx_dentry_inc(d) do {                                          \
10697 +       if ((d)->d_count == 1)                                          \
10698 +               vx_acc_cres(current_vx_info(), 1, d, VLIMIT_DENTRY);    \
10699 +       } while (0)
10700 +
10701 +#define vx_dentry_dec(d) do {                                          \
10702 +       if ((d)->d_count == 0)                                          \
10703 +               vx_acc_cres(current_vx_info(),-1, d, VLIMIT_DENTRY);    \
10704 +       } while (0)
10705 +
10706 +#define vx_dentry_avail(n) \
10707 +       vx_cres_avail(current_vx_info(), n, VLIMIT_DENTRY)
10708 +
10709 +
10710 +/* socket limits */
10711 +
10712 +#define vx_sock_inc(s) \
10713 +       vx_acc_cres((s)->sk_vx_info, 1, s, VLIMIT_NSOCK)
10714 +
10715 +#define vx_sock_dec(s) \
10716 +       vx_acc_cres((s)->sk_vx_info,-1, s, VLIMIT_NSOCK)
10717 +
10718 +#define vx_sock_avail(n) \
10719 +       vx_cres_avail(current_vx_info(), n, VLIMIT_NSOCK)
10720 +
10721 +
10722 +/* ipc resource limits */
10723 +
10724 +#define vx_ipcmsg_add(v, u, a) \
10725 +       vx_add_cres(v, a, u, RLIMIT_MSGQUEUE)
10726 +
10727 +#define vx_ipcmsg_sub(v, u, a) \
10728 +       vx_sub_cres(v, a, u, RLIMIT_MSGQUEUE)
10729 +
10730 +#define vx_ipcmsg_avail(v, a) \
10731 +       vx_cres_avail(v, a, RLIMIT_MSGQUEUE)
10732 +
10733 +
10734 +#define vx_ipcshm_add(v, k, a) \
10735 +       vx_add_cres(v, a, (void *)(long)(k), VLIMIT_SHMEM)
10736 +
10737 +#define vx_ipcshm_sub(v, k, a) \
10738 +       vx_sub_cres(v, a, (void *)(long)(k), VLIMIT_SHMEM)
10739 +
10740 +#define vx_ipcshm_avail(v, a) \
10741 +       vx_cres_avail(v, a, VLIMIT_SHMEM)
10742 +
10743 +
10744 +#define vx_semary_inc(a) \
10745 +       vx_acc_cres(current_vx_info(), 1, a, VLIMIT_SEMARY)
10746 +
10747 +#define vx_semary_dec(a) \
10748 +       vx_acc_cres(current_vx_info(), -1, a, VLIMIT_SEMARY)
10749 +
10750 +
10751 +#define vx_nsems_add(a,n) \
10752 +       vx_add_cres(current_vx_info(), n, a, VLIMIT_NSEMS)
10753 +
10754 +#define vx_nsems_sub(a,n) \
10755 +       vx_sub_cres(current_vx_info(), n, a, VLIMIT_NSEMS)
10756 +
10757 +
10758 +#else
10759 +#warning duplicate inclusion
10760 +#endif
10761 diff -NurpP --minimal linux-3.0-rc5/include/linux/vs_memory.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_memory.h
10762 --- linux-3.0-rc5/include/linux/vs_memory.h     1970-01-01 01:00:00.000000000 +0100
10763 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_memory.h        2011-06-10 22:11:24.000000000 +0200
10764 @@ -0,0 +1,58 @@
10765 +#ifndef _VS_MEMORY_H
10766 +#define _VS_MEMORY_H
10767 +
10768 +#include "vserver/limit.h"
10769 +#include "vserver/base.h"
10770 +#include "vserver/context.h"
10771 +#include "vserver/debug.h"
10772 +#include "vserver/context.h"
10773 +#include "vserver/limit_int.h"
10774 +
10775 +enum {
10776 +       VXPT_UNKNOWN = 0,
10777 +       VXPT_ANON,
10778 +       VXPT_NONE,
10779 +       VXPT_FILE,
10780 +       VXPT_SWAP,
10781 +       VXPT_WRITE
10782 +};
10783 +
10784 +#if 0
10785 +#define        vx_page_fault(mm, vma, type, ret)
10786 +#else
10787 +
10788 +static inline
10789 +void __vx_page_fault(struct mm_struct *mm,
10790 +       struct vm_area_struct *vma, int type, int ret)
10791 +{
10792 +       struct vx_info *vxi = mm->mm_vx_info;
10793 +       int what;
10794 +/*
10795 +       static char *page_type[6] =
10796 +               { "UNKNOWN", "ANON", "NONE", "FILE", "SWAP", "WRITE" };
10797 +       static char *page_what[4] =
10798 +               { "FAULT_OOM", "FAULT_SIGBUS", "FAULT_MINOR", "FAULT_MAJOR" };
10799 +*/
10800 +
10801 +       if (!vxi)
10802 +               return;
10803 +
10804 +       what = (ret & 0x3);
10805 +
10806 +/*     printk("[%d] page[%d][%d] %2x %s %s\n", vxi->vx_id,
10807 +               type, what, ret, page_type[type], page_what[what]);
10808 +*/
10809 +       if (ret & VM_FAULT_WRITE)
10810 +               what |= 0x4;
10811 +       atomic_inc(&vxi->cacct.page[type][what]);
10812 +}
10813 +
10814 +#define        vx_page_fault(mm, vma, type, ret)       __vx_page_fault(mm, vma, type, ret)
10815 +#endif
10816 +
10817 +
10818 +extern unsigned long vx_badness(struct task_struct *task, struct mm_struct *mm);
10819 +
10820 +#else
10821 +#warning duplicate inclusion
10822 +#endif
10823 diff -NurpP --minimal linux-3.0-rc5/include/linux/vs_network.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_network.h
10824 --- linux-3.0-rc5/include/linux/vs_network.h    1970-01-01 01:00:00.000000000 +0100
10825 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_network.h       2011-06-10 22:11:24.000000000 +0200
10826 @@ -0,0 +1,169 @@
10827 +#ifndef _NX_VS_NETWORK_H
10828 +#define _NX_VS_NETWORK_H
10829 +
10830 +#include "vserver/context.h"
10831 +#include "vserver/network.h"
10832 +#include "vserver/base.h"
10833 +#include "vserver/check.h"
10834 +#include "vserver/debug.h"
10835 +
10836 +#include <linux/sched.h>
10837 +
10838 +
10839 +#define get_nx_info(i) __get_nx_info(i, __FILE__, __LINE__)
10840 +
10841 +static inline struct nx_info *__get_nx_info(struct nx_info *nxi,
10842 +       const char *_file, int _line)
10843 +{
10844 +       if (!nxi)
10845 +               return NULL;
10846 +
10847 +       vxlprintk(VXD_CBIT(nid, 2), "get_nx_info(%p[#%d.%d])",
10848 +               nxi, nxi ? nxi->nx_id : 0,
10849 +               nxi ? atomic_read(&nxi->nx_usecnt) : 0,
10850 +               _file, _line);
10851 +
10852 +       atomic_inc(&nxi->nx_usecnt);
10853 +       return nxi;
10854 +}
10855 +
10856 +
10857 +extern void free_nx_info(struct nx_info *);
10858 +
10859 +#define put_nx_info(i) __put_nx_info(i, __FILE__, __LINE__)
10860 +
10861 +static inline void __put_nx_info(struct nx_info *nxi, const char *_file, int _line)
10862 +{
10863 +       if (!nxi)
10864 +               return;
10865 +
10866 +       vxlprintk(VXD_CBIT(nid, 2), "put_nx_info(%p[#%d.%d])",
10867 +               nxi, nxi ? nxi->nx_id : 0,
10868 +               nxi ? atomic_read(&nxi->nx_usecnt) : 0,
10869 +               _file, _line);
10870 +
10871 +       if (atomic_dec_and_test(&nxi->nx_usecnt))
10872 +               free_nx_info(nxi);
10873 +}
10874 +
10875 +
10876 +#define init_nx_info(p, i) __init_nx_info(p, i, __FILE__, __LINE__)
10877 +
10878 +static inline void __init_nx_info(struct nx_info **nxp, struct nx_info *nxi,
10879 +               const char *_file, int _line)
10880 +{
10881 +       if (nxi) {
10882 +               vxlprintk(VXD_CBIT(nid, 3),
10883 +                       "init_nx_info(%p[#%d.%d])",
10884 +                       nxi, nxi ? nxi->nx_id : 0,
10885 +                       nxi ? atomic_read(&nxi->nx_usecnt) : 0,
10886 +                       _file, _line);
10887 +
10888 +               atomic_inc(&nxi->nx_usecnt);
10889 +       }
10890 +       *nxp = nxi;
10891 +}
10892 +
10893 +
10894 +#define set_nx_info(p, i) __set_nx_info(p, i, __FILE__, __LINE__)
10895 +
10896 +static inline void __set_nx_info(struct nx_info **nxp, struct nx_info *nxi,
10897 +       const char *_file, int _line)
10898 +{
10899 +       struct nx_info *nxo;
10900 +
10901 +       if (!nxi)
10902 +               return;
10903 +
10904 +       vxlprintk(VXD_CBIT(nid, 3), "set_nx_info(%p[#%d.%d])",
10905 +               nxi, nxi ? nxi->nx_id : 0,
10906 +               nxi ? atomic_read(&nxi->nx_usecnt) : 0,
10907 +               _file, _line);
10908 +
10909 +       atomic_inc(&nxi->nx_usecnt);
10910 +       nxo = xchg(nxp, nxi);
10911 +       BUG_ON(nxo);
10912 +}
10913 +
10914 +#define clr_nx_info(p) __clr_nx_info(p, __FILE__, __LINE__)
10915 +
10916 +static inline void __clr_nx_info(struct nx_info **nxp,
10917 +       const char *_file, int _line)
10918 +{
10919 +       struct nx_info *nxo;
10920 +
10921 +       nxo = xchg(nxp, NULL);
10922 +       if (!nxo)
10923 +               return;
10924 +
10925 +       vxlprintk(VXD_CBIT(nid, 3), "clr_nx_info(%p[#%d.%d])",
10926 +               nxo, nxo ? nxo->nx_id : 0,
10927 +               nxo ? atomic_read(&nxo->nx_usecnt) : 0,
10928 +               _file, _line);
10929 +
10930 +       if (atomic_dec_and_test(&nxo->nx_usecnt))
10931 +               free_nx_info(nxo);
10932 +}
10933 +
10934 +
10935 +#define claim_nx_info(v, p) __claim_nx_info(v, p, __FILE__, __LINE__)
10936 +
10937 +static inline void __claim_nx_info(struct nx_info *nxi,
10938 +       struct task_struct *task, const char *_file, int _line)
10939 +{
10940 +       vxlprintk(VXD_CBIT(nid, 3), "claim_nx_info(%p[#%d.%d.%d]) %p",
10941 +               nxi, nxi ? nxi->nx_id : 0,
10942 +               nxi?atomic_read(&nxi->nx_usecnt):0,
10943 +               nxi?atomic_read(&nxi->nx_tasks):0,
10944 +               task, _file, _line);
10945 +
10946 +       atomic_inc(&nxi->nx_tasks);
10947 +}
10948 +
10949 +
10950 +extern void unhash_nx_info(struct nx_info *);
10951 +
10952 +#define release_nx_info(v, p) __release_nx_info(v, p, __FILE__, __LINE__)
10953 +
10954 +static inline void __release_nx_info(struct nx_info *nxi,
10955 +       struct task_struct *task, const char *_file, int _line)
10956 +{
10957 +       vxlprintk(VXD_CBIT(nid, 3), "release_nx_info(%p[#%d.%d.%d]) %p",
10958 +               nxi, nxi ? nxi->nx_id : 0,
10959 +               nxi ? atomic_read(&nxi->nx_usecnt) : 0,
10960 +               nxi ? atomic_read(&nxi->nx_tasks) : 0,
10961 +               task, _file, _line);
10962 +
10963 +       might_sleep();
10964 +
10965 +       if (atomic_dec_and_test(&nxi->nx_tasks))
10966 +               unhash_nx_info(nxi);
10967 +}
10968 +
10969 +
10970 +#define task_get_nx_info(i)    __task_get_nx_info(i, __FILE__, __LINE__)
10971 +
10972 +static __inline__ struct nx_info *__task_get_nx_info(struct task_struct *p,
10973 +       const char *_file, int _line)
10974 +{
10975 +       struct nx_info *nxi;
10976 +
10977 +       task_lock(p);
10978 +       vxlprintk(VXD_CBIT(nid, 5), "task_get_nx_info(%p)",
10979 +               p, _file, _line);
10980 +       nxi = __get_nx_info(p->nx_info, _file, _line);
10981 +       task_unlock(p);
10982 +       return nxi;
10983 +}
10984 +
10985 +
10986 +static inline void exit_nx_info(struct task_struct *p)
10987 +{
10988 +       if (p->nx_info)
10989 +               release_nx_info(p->nx_info, p);
10990 +}
10991 +
10992 +
10993 +#else
10994 +#warning duplicate inclusion
10995 +#endif
10996 diff -NurpP --minimal linux-3.0-rc5/include/linux/vs_pid.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_pid.h
10997 --- linux-3.0-rc5/include/linux/vs_pid.h        1970-01-01 01:00:00.000000000 +0100
10998 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_pid.h   2011-06-10 22:11:24.000000000 +0200
10999 @@ -0,0 +1,50 @@
11000 +#ifndef _VS_PID_H
11001 +#define _VS_PID_H
11002 +
11003 +#include "vserver/base.h"
11004 +#include "vserver/check.h"
11005 +#include "vserver/context.h"
11006 +#include "vserver/debug.h"
11007 +#include "vserver/pid.h"
11008 +#include <linux/pid_namespace.h>
11009 +
11010 +
11011 +#define VXF_FAKE_INIT  (VXF_INFO_INIT | VXF_STATE_INIT)
11012 +
11013 +static inline
11014 +int vx_proc_task_visible(struct task_struct *task)
11015 +{
11016 +       if ((task->pid == 1) &&
11017 +               !vx_flags(VXF_FAKE_INIT, VXF_FAKE_INIT))
11018 +               /* show a blend through init */
11019 +               goto visible;
11020 +       if (vx_check(vx_task_xid(task), VS_WATCH | VS_IDENT))
11021 +               goto visible;
11022 +       return 0;
11023 +visible:
11024 +       return 1;
11025 +}
11026 +
11027 +#define find_task_by_real_pid(pid) find_task_by_pid_ns(pid, &init_pid_ns)
11028 +
11029 +
11030 +static inline
11031 +struct task_struct *vx_get_proc_task(struct inode *inode, struct pid *pid)
11032 +{
11033 +       struct task_struct *task = get_pid_task(pid, PIDTYPE_PID);
11034 +
11035 +       if (task && !vx_proc_task_visible(task)) {
11036 +               vxdprintk(VXD_CBIT(misc, 6),
11037 +                       "dropping task (get) %p[#%u,%u] for %p[#%u,%u]",
11038 +                       task, task->xid, task->pid,
11039 +                       current, current->xid, current->pid);
11040 +               put_task_struct(task);
11041 +               task = NULL;
11042 +       }
11043 +       return task;
11044 +}
11045 +
11046 +
11047 +#else
11048 +#warning duplicate inclusion
11049 +#endif
11050 diff -NurpP --minimal linux-3.0-rc5/include/linux/vs_sched.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_sched.h
11051 --- linux-3.0-rc5/include/linux/vs_sched.h      1970-01-01 01:00:00.000000000 +0100
11052 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_sched.h 2011-06-10 22:11:24.000000000 +0200
11053 @@ -0,0 +1,40 @@
11054 +#ifndef _VS_SCHED_H
11055 +#define _VS_SCHED_H
11056 +
11057 +#include "vserver/base.h"
11058 +#include "vserver/context.h"
11059 +#include "vserver/sched.h"
11060 +
11061 +
11062 +#define MAX_PRIO_BIAS           20
11063 +#define MIN_PRIO_BIAS          -20
11064 +
11065 +static inline
11066 +int vx_adjust_prio(struct task_struct *p, int prio, int max_user)
11067 +{
11068 +       struct vx_info *vxi = p->vx_info;
11069 +
11070 +       if (vxi)
11071 +               prio += vx_cpu(vxi, sched_pc).prio_bias;
11072 +       return prio;
11073 +}
11074 +
11075 +static inline void vx_account_user(struct vx_info *vxi,
11076 +       cputime_t cputime, int nice)
11077 +{
11078 +       if (!vxi)
11079 +               return;
11080 +       vx_cpu(vxi, sched_pc).user_ticks += cputime;
11081 +}
11082 +
11083 +static inline void vx_account_system(struct vx_info *vxi,
11084 +       cputime_t cputime, int idle)
11085 +{
11086 +       if (!vxi)
11087 +               return;
11088 +       vx_cpu(vxi, sched_pc).sys_ticks += cputime;
11089 +}
11090 +
11091 +#else
11092 +#warning duplicate inclusion
11093 +#endif
11094 diff -NurpP --minimal linux-3.0-rc5/include/linux/vs_socket.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_socket.h
11095 --- linux-3.0-rc5/include/linux/vs_socket.h     1970-01-01 01:00:00.000000000 +0100
11096 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_socket.h        2011-06-10 22:11:24.000000000 +0200
11097 @@ -0,0 +1,67 @@
11098 +#ifndef _VS_SOCKET_H
11099 +#define _VS_SOCKET_H
11100 +
11101 +#include "vserver/debug.h"
11102 +#include "vserver/base.h"
11103 +#include "vserver/cacct.h"
11104 +#include "vserver/context.h"
11105 +#include "vserver/tag.h"
11106 +
11107 +
11108 +/* socket accounting */
11109 +
11110 +#include <linux/socket.h>
11111 +
11112 +static inline int vx_sock_type(int family)
11113 +{
11114 +       switch (family) {
11115 +       case PF_UNSPEC:
11116 +               return VXA_SOCK_UNSPEC;
11117 +       case PF_UNIX:
11118 +               return VXA_SOCK_UNIX;
11119 +       case PF_INET:
11120 +               return VXA_SOCK_INET;
11121 +       case PF_INET6:
11122 +               return VXA_SOCK_INET6;
11123 +       case PF_PACKET:
11124 +               return VXA_SOCK_PACKET;
11125 +       default:
11126 +               return VXA_SOCK_OTHER;
11127 +       }
11128 +}
11129 +
11130 +#define vx_acc_sock(v, f, p, s) \
11131 +       __vx_acc_sock(v, f, p, s, __FILE__, __LINE__)
11132 +
11133 +static inline void __vx_acc_sock(struct vx_info *vxi,
11134 +       int family, int pos, int size, char *file, int line)
11135 +{
11136 +       if (vxi) {
11137 +               int type = vx_sock_type(family);
11138 +
11139 +               atomic_long_inc(&vxi->cacct.sock[type][pos].count);
11140 +               atomic_long_add(size, &vxi->cacct.sock[type][pos].total);
11141 +       }
11142 +}
11143 +
11144 +#define vx_sock_recv(sk, s) \
11145 +       vx_acc_sock((sk)->sk_vx_info, (sk)->sk_family, 0, s)
11146 +#define vx_sock_send(sk, s) \
11147 +       vx_acc_sock((sk)->sk_vx_info, (sk)->sk_family, 1, s)
11148 +#define vx_sock_fail(sk, s) \
11149 +       vx_acc_sock((sk)->sk_vx_info, (sk)->sk_family, 2, s)
11150 +
11151 +
11152 +#define sock_vx_init(s) do {           \
11153 +       (s)->sk_xid = 0;                \
11154 +       (s)->sk_vx_info = NULL;         \
11155 +       } while (0)
11156 +
11157 +#define sock_nx_init(s) do {           \
11158 +       (s)->sk_nid = 0;                \
11159 +       (s)->sk_nx_info = NULL;         \
11160 +       } while (0)
11161 +
11162 +#else
11163 +#warning duplicate inclusion
11164 +#endif
11165 diff -NurpP --minimal linux-3.0-rc5/include/linux/vs_tag.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_tag.h
11166 --- linux-3.0-rc5/include/linux/vs_tag.h        1970-01-01 01:00:00.000000000 +0100
11167 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_tag.h   2011-06-10 22:11:24.000000000 +0200
11168 @@ -0,0 +1,47 @@
11169 +#ifndef _VS_TAG_H
11170 +#define _VS_TAG_H
11171 +
11172 +#include <linux/vserver/tag.h>
11173 +
11174 +/* check conditions */
11175 +
11176 +#define DX_ADMIN       0x0001
11177 +#define DX_WATCH       0x0002
11178 +#define DX_HOSTID      0x0008
11179 +
11180 +#define DX_IDENT       0x0010
11181 +
11182 +#define DX_ARG_MASK    0x0010
11183 +
11184 +
11185 +#define dx_task_tag(t) ((t)->tag)
11186 +
11187 +#define dx_current_tag() dx_task_tag(current)
11188 +
11189 +#define dx_check(c, m) __dx_check(dx_current_tag(), c, m)
11190 +
11191 +#define dx_weak_check(c, m)    ((m) ? dx_check(c, m) : 1)
11192 +
11193 +
11194 +/*
11195 + * check current context for ADMIN/WATCH and
11196 + * optionally against supplied argument
11197 + */
11198 +static inline int __dx_check(tag_t cid, tag_t id, unsigned int mode)
11199 +{
11200 +       if (mode & DX_ARG_MASK) {
11201 +               if ((mode & DX_IDENT) && (id == cid))
11202 +                       return 1;
11203 +       }
11204 +       return (((mode & DX_ADMIN) && (cid == 0)) ||
11205 +               ((mode & DX_WATCH) && (cid == 1)) ||
11206 +               ((mode & DX_HOSTID) && (id == 0)));
11207 +}
11208 +
11209 +struct inode;
11210 +int dx_permission(const struct inode *inode, int mask);
11211 +
11212 +
11213 +#else
11214 +#warning duplicate inclusion
11215 +#endif
11216 diff -NurpP --minimal linux-3.0-rc5/include/linux/vs_time.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_time.h
11217 --- linux-3.0-rc5/include/linux/vs_time.h       1970-01-01 01:00:00.000000000 +0100
11218 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vs_time.h  2011-06-13 14:57:45.000000000 +0200
11219 @@ -0,0 +1,19 @@
11220 +#ifndef _VS_TIME_H
11221 +#define _VS_TIME_H
11222 +
11223 +
11224 +/* time faking stuff */
11225 +
11226 +#ifdef CONFIG_VSERVER_VTIME
11227 +
11228 +extern void vx_adjust_timespec(struct timespec *ts);
11229 +extern int vx_settimeofday(const struct timespec *ts);
11230 +
11231 +#else
11232 +#define        vx_adjust_timespec(t)   do { } while (0)
11233 +#define        vx_settimeofday(t)      do_settimeofday(t)
11234 +#endif
11235 +
11236 +#else
11237 +#warning duplicate inclusion
11238 +#endif
11239 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/Kbuild linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/Kbuild
11240 --- linux-3.0-rc5/include/linux/vserver/Kbuild  1970-01-01 01:00:00.000000000 +0100
11241 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/Kbuild     2011-06-10 22:11:24.000000000 +0200
11242 @@ -0,0 +1,8 @@
11243 +
11244 +unifdef-y += context_cmd.h network_cmd.h space_cmd.h \
11245 +       cacct_cmd.h cvirt_cmd.h limit_cmd.h dlimit_cmd.h \
11246 +       inode_cmd.h tag_cmd.h sched_cmd.h signal_cmd.h \
11247 +       debug_cmd.h device_cmd.h
11248 +
11249 +unifdef-y += switch.h network.h monitor.h inode.h device.h
11250 +
11251 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/base.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/base.h
11252 --- linux-3.0-rc5/include/linux/vserver/base.h  1970-01-01 01:00:00.000000000 +0100
11253 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/base.h     2011-06-13 14:09:44.000000000 +0200
11254 @@ -0,0 +1,173 @@
11255 +#ifndef _VX_BASE_H
11256 +#define _VX_BASE_H
11257 +
11258 +
11259 +/* context state changes */
11260 +
11261 +enum {
11262 +       VSC_STARTUP = 1,
11263 +       VSC_SHUTDOWN,
11264 +
11265 +       VSC_NETUP,
11266 +       VSC_NETDOWN,
11267 +};
11268 +
11269 +
11270 +
11271 +#define vx_task_xid(t) ((t)->xid)
11272 +
11273 +#define vx_current_xid() vx_task_xid(current)
11274 +
11275 +#define current_vx_info() (current->vx_info)
11276 +
11277 +
11278 +#define nx_task_nid(t) ((t)->nid)
11279 +
11280 +#define nx_current_nid() nx_task_nid(current)
11281 +
11282 +#define current_nx_info() (current->nx_info)
11283 +
11284 +
11285 +/* generic flag merging */
11286 +
11287 +#define vs_check_flags(v, m, f)        (((v) & (m)) ^ (f))
11288 +
11289 +#define vs_mask_flags(v, f, m) (((v) & ~(m)) | ((f) & (m)))
11290 +
11291 +#define vs_mask_mask(v, f, m)  (((v) & ~(m)) | ((v) & (f) & (m)))
11292 +
11293 +#define vs_check_bit(v, n)     ((v) & (1LL << (n)))
11294 +
11295 +
11296 +/* context flags */
11297 +
11298 +#define __vx_flags(v)  ((v) ? (v)->vx_flags : 0)
11299 +
11300 +#define vx_current_flags()     __vx_flags(current_vx_info())
11301 +
11302 +#define vx_info_flags(v, m, f) \
11303 +       vs_check_flags(__vx_flags(v), m, f)
11304 +
11305 +#define task_vx_flags(t, m, f) \
11306 +       ((t) && vx_info_flags((t)->vx_info, m, f))
11307 +
11308 +#define vx_flags(m, f) vx_info_flags(current_vx_info(), m, f)
11309 +
11310 +
11311 +/* context caps */
11312 +
11313 +#define __vx_ccaps(v)  ((v) ? (v)->vx_ccaps : 0)
11314 +
11315 +#define vx_current_ccaps()     __vx_ccaps(current_vx_info())
11316 +
11317 +#define vx_info_ccaps(v, c)    (__vx_ccaps(v) & (c))
11318 +
11319 +#define vx_ccaps(c)    vx_info_ccaps(current_vx_info(), (c))
11320 +
11321 +
11322 +
11323 +/* network flags */
11324 +
11325 +#define __nx_flags(n)  ((n) ? (n)->nx_flags : 0)
11326 +
11327 +#define nx_current_flags()     __nx_flags(current_nx_info())
11328 +
11329 +#define nx_info_flags(n, m, f) \
11330 +       vs_check_flags(__nx_flags(n), m, f)
11331 +
11332 +#define task_nx_flags(t, m, f) \
11333 +       ((t) && nx_info_flags((t)->nx_info, m, f))
11334 +
11335 +#define nx_flags(m, f) nx_info_flags(current_nx_info(), m, f)
11336 +
11337 +
11338 +/* network caps */
11339 +
11340 +#define __nx_ncaps(n)  ((n) ? (n)->nx_ncaps : 0)
11341 +
11342 +#define nx_current_ncaps()     __nx_ncaps(current_nx_info())
11343 +
11344 +#define nx_info_ncaps(n, c)    (__nx_ncaps(n) & (c))
11345 +
11346 +#define nx_ncaps(c)    nx_info_ncaps(current_nx_info(), c)
11347 +
11348 +
11349 +/* context mask capabilities */
11350 +
11351 +#define __vx_mcaps(v)  ((v) ? (v)->vx_ccaps >> 32UL : ~0 )
11352 +
11353 +#define vx_info_mcaps(v, c)    (__vx_mcaps(v) & (c))
11354 +
11355 +#define vx_mcaps(c)    vx_info_mcaps(current_vx_info(), c)
11356 +
11357 +
11358 +/* context bcap mask */
11359 +
11360 +#define __vx_bcaps(v)          ((v)->vx_bcaps)
11361 +
11362 +#define vx_current_bcaps()     __vx_bcaps(current_vx_info())
11363 +
11364 +
11365 +/* mask given bcaps */
11366 +
11367 +#define vx_info_mbcaps(v, c)   ((v) ? cap_intersect(__vx_bcaps(v), c) : c)
11368 +
11369 +#define vx_mbcaps(c)           vx_info_mbcaps(current_vx_info(), c)
11370 +
11371 +
11372 +/* masked cap_bset */
11373 +
11374 +#define vx_info_cap_bset(v)    vx_info_mbcaps(v, current->cap_bset)
11375 +
11376 +#define vx_current_cap_bset()  vx_info_cap_bset(current_vx_info())
11377 +
11378 +#if 0
11379 +#define vx_info_mbcap(v, b) \
11380 +       (!vx_info_flags(v, VXF_STATE_SETUP, 0) ? \
11381 +       vx_info_bcaps(v, b) : (b))
11382 +
11383 +#define task_vx_mbcap(t, b) \
11384 +       vx_info_mbcap((t)->vx_info, (t)->b)
11385 +
11386 +#define vx_mbcap(b)    task_vx_mbcap(current, b)
11387 +#endif
11388 +
11389 +#define vx_cap_raised(v, c, f) cap_raised(vx_info_mbcaps(v, c), f)
11390 +
11391 +#define vx_capable(b, c) (capable(b) || \
11392 +       (cap_raised(current_cap(), b) && vx_ccaps(c)))
11393 +
11394 +#define vx_ns_capable(n, b, c) (ns_capable(n, b) || \
11395 +       (cap_raised(current_cap(), b) && vx_ccaps(c)))
11396 +
11397 +#define nx_capable(b, c) (capable(b) || \
11398 +       (cap_raised(current_cap(), b) && nx_ncaps(c)))
11399 +
11400 +#define vx_task_initpid(t, n) \
11401 +       ((t)->vx_info && \
11402 +       ((t)->vx_info->vx_initpid == (n)))
11403 +
11404 +#define vx_current_initpid(n)  vx_task_initpid(current, n)
11405 +
11406 +
11407 +/* context unshare mask */
11408 +
11409 +#define __vx_umask(v)          ((v)->vx_umask)
11410 +
11411 +#define vx_current_umask()     __vx_umask(current_vx_info())
11412 +
11413 +#define vx_can_unshare(b, f) (capable(b) || \
11414 +       (cap_raised(current_cap(), b) && \
11415 +       !((f) & ~vx_current_umask())))
11416 +
11417 +
11418 +#define __vx_state(v)  ((v) ? ((v)->vx_state) : 0)
11419 +
11420 +#define vx_info_state(v, m)    (__vx_state(v) & (m))
11421 +
11422 +
11423 +#define __nx_state(n)  ((n) ? ((n)->nx_state) : 0)
11424 +
11425 +#define nx_info_state(n, m)    (__nx_state(n) & (m))
11426 +
11427 +#endif
11428 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/cacct.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/cacct.h
11429 --- linux-3.0-rc5/include/linux/vserver/cacct.h 1970-01-01 01:00:00.000000000 +0100
11430 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/cacct.h    2011-06-10 22:11:24.000000000 +0200
11431 @@ -0,0 +1,15 @@
11432 +#ifndef _VX_CACCT_H
11433 +#define _VX_CACCT_H
11434 +
11435 +
11436 +enum sock_acc_field {
11437 +       VXA_SOCK_UNSPEC = 0,
11438 +       VXA_SOCK_UNIX,
11439 +       VXA_SOCK_INET,
11440 +       VXA_SOCK_INET6,
11441 +       VXA_SOCK_PACKET,
11442 +       VXA_SOCK_OTHER,
11443 +       VXA_SOCK_SIZE   /* array size */
11444 +};
11445 +
11446 +#endif /* _VX_CACCT_H */
11447 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/cacct_cmd.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/cacct_cmd.h
11448 --- linux-3.0-rc5/include/linux/vserver/cacct_cmd.h     1970-01-01 01:00:00.000000000 +0100
11449 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/cacct_cmd.h        2011-06-10 22:11:24.000000000 +0200
11450 @@ -0,0 +1,23 @@
11451 +#ifndef _VX_CACCT_CMD_H
11452 +#define _VX_CACCT_CMD_H
11453 +
11454 +
11455 +/* virtual host info name commands */
11456 +
11457 +#define VCMD_sock_stat         VC_CMD(VSTAT, 5, 0)
11458 +
11459 +struct vcmd_sock_stat_v0 {
11460 +       uint32_t field;
11461 +       uint32_t count[3];
11462 +       uint64_t total[3];
11463 +};
11464 +
11465 +
11466 +#ifdef __KERNEL__
11467 +
11468 +#include <linux/compiler.h>
11469 +
11470 +extern int vc_sock_stat(struct vx_info *, void __user *);
11471 +
11472 +#endif /* __KERNEL__ */
11473 +#endif /* _VX_CACCT_CMD_H */
11474 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/cacct_def.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/cacct_def.h
11475 --- linux-3.0-rc5/include/linux/vserver/cacct_def.h     1970-01-01 01:00:00.000000000 +0100
11476 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/cacct_def.h        2011-06-10 22:11:24.000000000 +0200
11477 @@ -0,0 +1,43 @@
11478 +#ifndef _VX_CACCT_DEF_H
11479 +#define _VX_CACCT_DEF_H
11480 +
11481 +#include <asm/atomic.h>
11482 +#include <linux/vserver/cacct.h>
11483 +
11484 +
11485 +struct _vx_sock_acc {
11486 +       atomic_long_t count;
11487 +       atomic_long_t total;
11488 +};
11489 +
11490 +/* context sub struct */
11491 +
11492 +struct _vx_cacct {
11493 +       struct _vx_sock_acc sock[VXA_SOCK_SIZE][3];
11494 +       atomic_t slab[8];
11495 +       atomic_t page[6][8];
11496 +};
11497 +
11498 +#ifdef CONFIG_VSERVER_DEBUG
11499 +
11500 +static inline void __dump_vx_cacct(struct _vx_cacct *cacct)
11501 +{
11502 +       int i, j;
11503 +
11504 +       printk("\t_vx_cacct:");
11505 +       for (i = 0; i < 6; i++) {
11506 +               struct _vx_sock_acc *ptr = cacct->sock[i];
11507 +
11508 +               printk("\t [%d] =", i);
11509 +               for (j = 0; j < 3; j++) {
11510 +                       printk(" [%d] = %8lu, %8lu", j,
11511 +                               atomic_long_read(&ptr[j].count),
11512 +                               atomic_long_read(&ptr[j].total));
11513 +               }
11514 +               printk("\n");
11515 +       }
11516 +}
11517 +
11518 +#endif
11519 +
11520 +#endif /* _VX_CACCT_DEF_H */
11521 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/cacct_int.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/cacct_int.h
11522 --- linux-3.0-rc5/include/linux/vserver/cacct_int.h     1970-01-01 01:00:00.000000000 +0100
11523 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/cacct_int.h        2011-06-10 22:11:24.000000000 +0200
11524 @@ -0,0 +1,21 @@
11525 +#ifndef _VX_CACCT_INT_H
11526 +#define _VX_CACCT_INT_H
11527 +
11528 +
11529 +#ifdef __KERNEL__
11530 +
11531 +static inline
11532 +unsigned long vx_sock_count(struct _vx_cacct *cacct, int type, int pos)
11533 +{
11534 +       return atomic_long_read(&cacct->sock[type][pos].count);
11535 +}
11536 +
11537 +
11538 +static inline
11539 +unsigned long vx_sock_total(struct _vx_cacct *cacct, int type, int pos)
11540 +{
11541 +       return atomic_long_read(&cacct->sock[type][pos].total);
11542 +}
11543 +
11544 +#endif /* __KERNEL__ */
11545 +#endif /* _VX_CACCT_INT_H */
11546 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/check.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/check.h
11547 --- linux-3.0-rc5/include/linux/vserver/check.h 1970-01-01 01:00:00.000000000 +0100
11548 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/check.h    2011-06-10 22:11:24.000000000 +0200
11549 @@ -0,0 +1,89 @@
11550 +#ifndef _VS_CHECK_H
11551 +#define _VS_CHECK_H
11552 +
11553 +
11554 +#define MAX_S_CONTEXT  65535   /* Arbitrary limit */
11555 +
11556 +#ifdef CONFIG_VSERVER_DYNAMIC_IDS
11557 +#define MIN_D_CONTEXT  49152   /* dynamic contexts start here */
11558 +#else
11559 +#define MIN_D_CONTEXT  65536
11560 +#endif
11561 +
11562 +/* check conditions */
11563 +
11564 +#define VS_ADMIN       0x0001
11565 +#define VS_WATCH       0x0002
11566 +#define VS_HIDE                0x0004
11567 +#define VS_HOSTID      0x0008
11568 +
11569 +#define VS_IDENT       0x0010
11570 +#define VS_EQUIV       0x0020
11571 +#define VS_PARENT      0x0040
11572 +#define VS_CHILD       0x0080
11573 +
11574 +#define VS_ARG_MASK    0x00F0
11575 +
11576 +#define VS_DYNAMIC     0x0100
11577 +#define VS_STATIC      0x0200
11578 +
11579 +#define VS_ATR_MASK    0x0F00
11580 +
11581 +#ifdef CONFIG_VSERVER_PRIVACY
11582 +#define VS_ADMIN_P     (0)
11583 +#define VS_WATCH_P     (0)
11584 +#else
11585 +#define VS_ADMIN_P     VS_ADMIN
11586 +#define VS_WATCH_P     VS_WATCH
11587 +#endif
11588 +
11589 +#define VS_HARDIRQ     0x1000
11590 +#define VS_SOFTIRQ     0x2000
11591 +#define VS_IRQ         0x4000
11592 +
11593 +#define VS_IRQ_MASK    0xF000
11594 +
11595 +#include <linux/hardirq.h>
11596 +
11597 +/*
11598 + * check current context for ADMIN/WATCH and
11599 + * optionally against supplied argument
11600 + */
11601 +static inline int __vs_check(int cid, int id, unsigned int mode)
11602 +{
11603 +       if (mode & VS_ARG_MASK) {
11604 +               if ((mode & VS_IDENT) && (id == cid))
11605 +                       return 1;
11606 +       }
11607 +       if (mode & VS_ATR_MASK) {
11608 +               if ((mode & VS_DYNAMIC) &&
11609 +                       (id >= MIN_D_CONTEXT) &&
11610 +                       (id <= MAX_S_CONTEXT))
11611 +                       return 1;
11612 +               if ((mode & VS_STATIC) &&
11613 +                       (id > 1) && (id < MIN_D_CONTEXT))
11614 +                       return 1;
11615 +       }
11616 +       if (mode & VS_IRQ_MASK) {
11617 +               if ((mode & VS_IRQ) && unlikely(in_interrupt()))
11618 +                       return 1;
11619 +               if ((mode & VS_HARDIRQ) && unlikely(in_irq()))
11620 +                       return 1;
11621 +               if ((mode & VS_SOFTIRQ) && unlikely(in_softirq()))
11622 +                       return 1;
11623 +       }
11624 +       return (((mode & VS_ADMIN) && (cid == 0)) ||
11625 +               ((mode & VS_WATCH) && (cid == 1)) ||
11626 +               ((mode & VS_HOSTID) && (id == 0)));
11627 +}
11628 +
11629 +#define vx_check(c, m) __vs_check(vx_current_xid(), c, (m) | VS_IRQ)
11630 +
11631 +#define vx_weak_check(c, m)    ((m) ? vx_check(c, m) : 1)
11632 +
11633 +
11634 +#define nx_check(c, m) __vs_check(nx_current_nid(), c, m)
11635 +
11636 +#define nx_weak_check(c, m)    ((m) ? nx_check(c, m) : 1)
11637 +
11638 +#endif
11639 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/context.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/context.h
11640 --- linux-3.0-rc5/include/linux/vserver/context.h       1970-01-01 01:00:00.000000000 +0100
11641 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/context.h  2011-06-10 22:11:24.000000000 +0200
11642 @@ -0,0 +1,189 @@
11643 +#ifndef _VX_CONTEXT_H
11644 +#define _VX_CONTEXT_H
11645 +
11646 +#include <linux/types.h>
11647 +#include <linux/capability.h>
11648 +
11649 +
11650 +/* context flags */
11651 +
11652 +#define VXF_INFO_SCHED         0x00000002
11653 +#define VXF_INFO_NPROC         0x00000004
11654 +#define VXF_INFO_PRIVATE       0x00000008
11655 +
11656 +#define VXF_INFO_INIT          0x00000010
11657 +#define VXF_INFO_HIDE          0x00000020
11658 +#define VXF_INFO_ULIMIT                0x00000040
11659 +#define VXF_INFO_NSPACE                0x00000080
11660 +
11661 +#define VXF_SCHED_HARD         0x00000100
11662 +#define VXF_SCHED_PRIO         0x00000200
11663 +#define VXF_SCHED_PAUSE                0x00000400
11664 +
11665 +#define VXF_VIRT_MEM           0x00010000
11666 +#define VXF_VIRT_UPTIME                0x00020000
11667 +#define VXF_VIRT_CPU           0x00040000
11668 +#define VXF_VIRT_LOAD          0x00080000
11669 +#define VXF_VIRT_TIME          0x00100000
11670 +
11671 +#define VXF_HIDE_MOUNT         0x01000000
11672 +/* was VXF_HIDE_NETIF          0x02000000 */
11673 +#define VXF_HIDE_VINFO         0x04000000
11674 +
11675 +#define VXF_STATE_SETUP                (1ULL << 32)
11676 +#define VXF_STATE_INIT         (1ULL << 33)
11677 +#define VXF_STATE_ADMIN                (1ULL << 34)
11678 +
11679 +#define VXF_SC_HELPER          (1ULL << 36)
11680 +#define VXF_REBOOT_KILL                (1ULL << 37)
11681 +#define VXF_PERSISTENT         (1ULL << 38)
11682 +
11683 +#define VXF_FORK_RSS           (1ULL << 48)
11684 +#define VXF_PROLIFIC           (1ULL << 49)
11685 +
11686 +#define VXF_IGNEG_NICE         (1ULL << 52)
11687 +
11688 +#define VXF_ONE_TIME           (0x0007ULL << 32)
11689 +
11690 +#define VXF_INIT_SET           (VXF_STATE_SETUP | VXF_STATE_INIT | VXF_STATE_ADMIN)
11691 +
11692 +
11693 +/* context migration */
11694 +
11695 +#define VXM_SET_INIT           0x00000001
11696 +#define VXM_SET_REAPER         0x00000002
11697 +
11698 +/* context caps */
11699 +
11700 +#define VXC_CAP_MASK           0x00000000
11701 +
11702 +#define VXC_SET_UTSNAME                0x00000001
11703 +#define VXC_SET_RLIMIT         0x00000002
11704 +#define VXC_FS_SECURITY                0x00000004
11705 +#define VXC_FS_TRUSTED         0x00000008
11706 +#define VXC_TIOCSTI            0x00000010
11707 +
11708 +/* was VXC_RAW_ICMP            0x00000100 */
11709 +#define VXC_SYSLOG             0x00001000
11710 +#define VXC_OOM_ADJUST         0x00002000
11711 +#define VXC_AUDIT_CONTROL      0x00004000
11712 +
11713 +#define VXC_SECURE_MOUNT       0x00010000
11714 +#define VXC_SECURE_REMOUNT     0x00020000
11715 +#define VXC_BINARY_MOUNT       0x00040000
11716 +
11717 +#define VXC_QUOTA_CTL          0x00100000
11718 +#define VXC_ADMIN_MAPPER       0x00200000
11719 +#define VXC_ADMIN_CLOOP                0x00400000
11720 +
11721 +#define VXC_KTHREAD            0x01000000
11722 +#define VXC_NAMESPACE          0x02000000
11723 +
11724 +
11725 +#ifdef __KERNEL__
11726 +
11727 +#include <linux/list.h>
11728 +#include <linux/spinlock.h>
11729 +#include <linux/rcupdate.h>
11730 +
11731 +#include "limit_def.h"
11732 +#include "sched_def.h"
11733 +#include "cvirt_def.h"
11734 +#include "cacct_def.h"
11735 +#include "device_def.h"
11736 +
11737 +#define VX_SPACES      2
11738 +
11739 +struct _vx_info_pc {
11740 +       struct _vx_sched_pc sched_pc;
11741 +       struct _vx_cvirt_pc cvirt_pc;
11742 +};
11743 +
11744 +struct _vx_space {
11745 +       unsigned long vx_nsmask;                /* assignment mask */
11746 +       struct nsproxy *vx_nsproxy;             /* private namespaces */
11747 +       struct fs_struct *vx_fs;                /* private namespace fs */
11748 +       const struct cred *vx_cred;             /* task credentials */
11749 +};
11750 +
11751 +struct vx_info {
11752 +       struct hlist_node vx_hlist;             /* linked list of contexts */
11753 +       xid_t vx_id;                            /* context id */
11754 +       atomic_t vx_usecnt;                     /* usage count */
11755 +       atomic_t vx_tasks;                      /* tasks count */
11756 +       struct vx_info *vx_parent;              /* parent context */
11757 +       int vx_state;                           /* context state */
11758 +
11759 +       struct _vx_space space[VX_SPACES];      /* namespace store */
11760 +
11761 +       uint64_t vx_flags;                      /* context flags */
11762 +       uint64_t vx_ccaps;                      /* context caps (vserver) */
11763 +       uint64_t vx_umask;                      /* unshare mask (guest) */
11764 +       kernel_cap_t vx_bcaps;                  /* bounding caps (system) */
11765 +
11766 +       struct task_struct *vx_reaper;          /* guest reaper process */
11767 +       pid_t vx_initpid;                       /* PID of guest init */
11768 +       int64_t vx_badness_bias;                /* OOM points bias */
11769 +
11770 +       struct _vx_limit limit;                 /* vserver limits */
11771 +       struct _vx_sched sched;                 /* vserver scheduler */
11772 +       struct _vx_cvirt cvirt;                 /* virtual/bias stuff */
11773 +       struct _vx_cacct cacct;                 /* context accounting */
11774 +
11775 +       struct _vx_device dmap;                 /* default device map targets */
11776 +
11777 +#ifndef CONFIG_SMP
11778 +       struct _vx_info_pc info_pc;             /* per cpu data */
11779 +#else
11780 +       struct _vx_info_pc *ptr_pc;             /* per cpu array */
11781 +#endif
11782 +
11783 +       wait_queue_head_t vx_wait;              /* context exit waitqueue */
11784 +       int reboot_cmd;                         /* last sys_reboot() cmd */
11785 +       int exit_code;                          /* last process exit code */
11786 +
11787 +       char vx_name[65];                       /* vserver name */
11788 +};
11789 +
11790 +#ifndef CONFIG_SMP
11791 +#define        vx_ptr_pc(vxi)          (&(vxi)->info_pc)
11792 +#define        vx_per_cpu(vxi, v, id)  vx_ptr_pc(vxi)->v
11793 +#else
11794 +#define        vx_ptr_pc(vxi)          ((vxi)->ptr_pc)
11795 +#define        vx_per_cpu(vxi, v, id)  per_cpu_ptr(vx_ptr_pc(vxi), id)->v
11796 +#endif
11797 +
11798 +#define        vx_cpu(vxi, v)          vx_per_cpu(vxi, v, smp_processor_id())
11799 +
11800 +
11801 +struct vx_info_save {
11802 +       struct vx_info *vxi;
11803 +       xid_t xid;
11804 +};
11805 +
11806 +
11807 +/* status flags */
11808 +
11809 +#define VXS_HASHED     0x0001
11810 +#define VXS_PAUSED     0x0010
11811 +#define VXS_SHUTDOWN   0x0100
11812 +#define VXS_HELPER     0x1000
11813 +#define VXS_RELEASED   0x8000
11814 +
11815 +
11816 +extern void claim_vx_info(struct vx_info *, struct task_struct *);
11817 +extern void release_vx_info(struct vx_info *, struct task_struct *);
11818 +
11819 +extern struct vx_info *lookup_vx_info(int);
11820 +extern struct vx_info *lookup_or_create_vx_info(int);
11821 +
11822 +extern int get_xid_list(int, unsigned int *, int);
11823 +extern int xid_is_hashed(xid_t);
11824 +
11825 +extern int vx_migrate_task(struct task_struct *, struct vx_info *, int);
11826 +
11827 +extern long vs_state_change(struct vx_info *, unsigned int);
11828 +
11829 +
11830 +#endif /* __KERNEL__ */
11831 +#endif /* _VX_CONTEXT_H */
11832 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/context_cmd.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/context_cmd.h
11833 --- linux-3.0-rc5/include/linux/vserver/context_cmd.h   1970-01-01 01:00:00.000000000 +0100
11834 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/context_cmd.h      2011-06-10 22:11:24.000000000 +0200
11835 @@ -0,0 +1,145 @@
11836 +#ifndef _VX_CONTEXT_CMD_H
11837 +#define _VX_CONTEXT_CMD_H
11838 +
11839 +
11840 +/* vinfo commands */
11841 +
11842 +#define VCMD_task_xid          VC_CMD(VINFO, 1, 0)
11843 +
11844 +#ifdef __KERNEL__
11845 +extern int vc_task_xid(uint32_t);
11846 +
11847 +#endif /* __KERNEL__ */
11848 +
11849 +#define VCMD_vx_info           VC_CMD(VINFO, 5, 0)
11850 +
11851 +struct vcmd_vx_info_v0 {
11852 +       uint32_t xid;
11853 +       uint32_t initpid;
11854 +       /* more to come */
11855 +};
11856 +
11857 +#ifdef __KERNEL__
11858 +extern int vc_vx_info(struct vx_info *, void __user *);
11859 +
11860 +#endif /* __KERNEL__ */
11861 +
11862 +#define VCMD_ctx_stat          VC_CMD(VSTAT, 0, 0)
11863 +
11864 +struct vcmd_ctx_stat_v0 {
11865 +       uint32_t usecnt;
11866 +       uint32_t tasks;
11867 +       /* more to come */
11868 +};
11869 +
11870 +#ifdef __KERNEL__
11871 +extern int vc_ctx_stat(struct vx_info *, void __user *);
11872 +
11873 +#endif /* __KERNEL__ */
11874 +
11875 +/* context commands */
11876 +
11877 +#define VCMD_ctx_create_v0     VC_CMD(VPROC, 1, 0)
11878 +#define VCMD_ctx_create                VC_CMD(VPROC, 1, 1)
11879 +
11880 +struct vcmd_ctx_create {
11881 +       uint64_t flagword;
11882 +};
11883 +
11884 +#define VCMD_ctx_migrate_v0    VC_CMD(PROCMIG, 1, 0)
11885 +#define VCMD_ctx_migrate       VC_CMD(PROCMIG, 1, 1)
11886 +
11887 +struct vcmd_ctx_migrate {
11888 +       uint64_t flagword;
11889 +};
11890 +
11891 +#ifdef __KERNEL__
11892 +extern int vc_ctx_create(uint32_t, void __user *);
11893 +extern int vc_ctx_migrate(struct vx_info *, void __user *);
11894 +
11895 +#endif /* __KERNEL__ */
11896 +
11897 +
11898 +/* flag commands */
11899 +
11900 +#define VCMD_get_cflags                VC_CMD(FLAGS, 1, 0)
11901 +#define VCMD_set_cflags                VC_CMD(FLAGS, 2, 0)
11902 +
11903 +struct vcmd_ctx_flags_v0 {
11904 +       uint64_t flagword;
11905 +       uint64_t mask;
11906 +};
11907 +
11908 +#ifdef __KERNEL__
11909 +extern int vc_get_cflags(struct vx_info *, void __user *);
11910 +extern int vc_set_cflags(struct vx_info *, void __user *);
11911 +
11912 +#endif /* __KERNEL__ */
11913 +
11914 +
11915 +/* context caps commands */
11916 +
11917 +#define VCMD_get_ccaps         VC_CMD(FLAGS, 3, 1)
11918 +#define VCMD_set_ccaps         VC_CMD(FLAGS, 4, 1)
11919 +
11920 +struct vcmd_ctx_caps_v1 {
11921 +       uint64_t ccaps;
11922 +       uint64_t cmask;
11923 +};
11924 +
11925 +#ifdef __KERNEL__
11926 +extern int vc_get_ccaps(struct vx_info *, void __user *);
11927 +extern int vc_set_ccaps(struct vx_info *, void __user *);
11928 +
11929 +#endif /* __KERNEL__ */
11930 +
11931 +
11932 +/* bcaps commands */
11933 +
11934 +#define VCMD_get_bcaps         VC_CMD(FLAGS, 9, 0)
11935 +#define VCMD_set_bcaps         VC_CMD(FLAGS, 10, 0)
11936 +
11937 +struct vcmd_bcaps {
11938 +       uint64_t bcaps;
11939 +       uint64_t bmask;
11940 +};
11941 +
11942 +#ifdef __KERNEL__
11943 +extern int vc_get_bcaps(struct vx_info *, void __user *);
11944 +extern int vc_set_bcaps(struct vx_info *, void __user *);
11945 +
11946 +#endif /* __KERNEL__ */
11947 +
11948 +
11949 +/* umask commands */
11950 +
11951 +#define VCMD_get_umask         VC_CMD(FLAGS, 13, 0)
11952 +#define VCMD_set_umask         VC_CMD(FLAGS, 14, 0)
11953 +
11954 +struct vcmd_umask {
11955 +       uint64_t umask;
11956 +       uint64_t mask;
11957 +};
11958 +
11959 +#ifdef __KERNEL__
11960 +extern int vc_get_umask(struct vx_info *, void __user *);
11961 +extern int vc_set_umask(struct vx_info *, void __user *);
11962 +
11963 +#endif /* __KERNEL__ */
11964 +
11965 +
11966 +/* OOM badness */
11967 +
11968 +#define VCMD_get_badness       VC_CMD(MEMCTRL, 5, 0)
11969 +#define VCMD_set_badness       VC_CMD(MEMCTRL, 6, 0)
11970 +
11971 +struct vcmd_badness_v0 {
11972 +       int64_t bias;
11973 +};
11974 +
11975 +#ifdef __KERNEL__
11976 +extern int vc_get_badness(struct vx_info *, void __user *);
11977 +extern int vc_set_badness(struct vx_info *, void __user *);
11978 +
11979 +#endif /* __KERNEL__ */
11980 +#endif /* _VX_CONTEXT_CMD_H */
11981 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/cvirt.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/cvirt.h
11982 --- linux-3.0-rc5/include/linux/vserver/cvirt.h 1970-01-01 01:00:00.000000000 +0100
11983 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/cvirt.h    2011-06-10 22:11:24.000000000 +0200
11984 @@ -0,0 +1,20 @@
11985 +#ifndef _VX_CVIRT_H
11986 +#define _VX_CVIRT_H
11987 +
11988 +
11989 +#ifdef __KERNEL__
11990 +
11991 +struct timespec;
11992 +
11993 +void vx_vsi_uptime(struct timespec *, struct timespec *);
11994 +
11995 +
11996 +struct vx_info;
11997 +
11998 +void vx_update_load(struct vx_info *);
11999 +
12000 +
12001 +int vx_do_syslog(int, char __user *, int);
12002 +
12003 +#endif /* __KERNEL__ */
12004 +#endif /* _VX_CVIRT_H */
12005 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/cvirt_cmd.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/cvirt_cmd.h
12006 --- linux-3.0-rc5/include/linux/vserver/cvirt_cmd.h     1970-01-01 01:00:00.000000000 +0100
12007 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/cvirt_cmd.h        2011-06-10 22:11:24.000000000 +0200
12008 @@ -0,0 +1,53 @@
12009 +#ifndef _VX_CVIRT_CMD_H
12010 +#define _VX_CVIRT_CMD_H
12011 +
12012 +
12013 +/* virtual host info name commands */
12014 +
12015 +#define VCMD_set_vhi_name      VC_CMD(VHOST, 1, 0)
12016 +#define VCMD_get_vhi_name      VC_CMD(VHOST, 2, 0)
12017 +
12018 +struct vcmd_vhi_name_v0 {
12019 +       uint32_t field;
12020 +       char name[65];
12021 +};
12022 +
12023 +
12024 +enum vhi_name_field {
12025 +       VHIN_CONTEXT = 0,
12026 +       VHIN_SYSNAME,
12027 +       VHIN_NODENAME,
12028 +       VHIN_RELEASE,
12029 +       VHIN_VERSION,
12030 +       VHIN_MACHINE,
12031 +       VHIN_DOMAINNAME,
12032 +};
12033 +
12034 +
12035 +#ifdef __KERNEL__
12036 +
12037 +#include <linux/compiler.h>
12038 +
12039 +extern int vc_set_vhi_name(struct vx_info *, void __user *);
12040 +extern int vc_get_vhi_name(struct vx_info *, void __user *);
12041 +
12042 +#endif /* __KERNEL__ */
12043 +
12044 +#define VCMD_virt_stat         VC_CMD(VSTAT, 3, 0)
12045 +
12046 +struct vcmd_virt_stat_v0 {
12047 +       uint64_t offset;
12048 +       uint64_t uptime;
12049 +       uint32_t nr_threads;
12050 +       uint32_t nr_running;
12051 +       uint32_t nr_uninterruptible;
12052 +       uint32_t nr_onhold;
12053 +       uint32_t nr_forks;
12054 +       uint32_t load[3];
12055 +};
12056 +
12057 +#ifdef __KERNEL__
12058 +extern int vc_virt_stat(struct vx_info *, void __user *);
12059 +
12060 +#endif /* __KERNEL__ */
12061 +#endif /* _VX_CVIRT_CMD_H */
12062 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/cvirt_def.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/cvirt_def.h
12063 --- linux-3.0-rc5/include/linux/vserver/cvirt_def.h     1970-01-01 01:00:00.000000000 +0100
12064 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/cvirt_def.h        2011-06-10 22:11:24.000000000 +0200
12065 @@ -0,0 +1,80 @@
12066 +#ifndef _VX_CVIRT_DEF_H
12067 +#define _VX_CVIRT_DEF_H
12068 +
12069 +#include <linux/jiffies.h>
12070 +#include <linux/spinlock.h>
12071 +#include <linux/wait.h>
12072 +#include <linux/time.h>
12073 +#include <asm/atomic.h>
12074 +
12075 +
12076 +struct _vx_usage_stat {
12077 +       uint64_t user;
12078 +       uint64_t nice;
12079 +       uint64_t system;
12080 +       uint64_t softirq;
12081 +       uint64_t irq;
12082 +       uint64_t idle;
12083 +       uint64_t iowait;
12084 +};
12085 +
12086 +struct _vx_syslog {
12087 +       wait_queue_head_t log_wait;
12088 +       spinlock_t logbuf_lock;         /* lock for the log buffer */
12089 +
12090 +       unsigned long log_start;        /* next char to be read by syslog() */
12091 +       unsigned long con_start;        /* next char to be sent to consoles */
12092 +       unsigned long log_end;  /* most-recently-written-char + 1 */
12093 +       unsigned long logged_chars;     /* #chars since last read+clear operation */
12094 +
12095 +       char log_buf[1024];
12096 +};
12097 +
12098 +
12099 +/* context sub struct */
12100 +
12101 +struct _vx_cvirt {
12102 +       atomic_t nr_threads;            /* number of current threads */
12103 +       atomic_t nr_running;            /* number of running threads */
12104 +       atomic_t nr_uninterruptible;    /* number of uninterruptible threads */
12105 +
12106 +       atomic_t nr_onhold;             /* processes on hold */
12107 +       uint32_t onhold_last;           /* jiffies when put on hold */
12108 +
12109 +       struct timespec bias_ts;        /* time offset to the host */
12110 +       struct timespec bias_idle;
12111 +       struct timespec bias_uptime;    /* context creation point */
12112 +       uint64_t bias_clock;            /* offset in clock_t */
12113 +
12114 +       spinlock_t load_lock;           /* lock for the load averages */
12115 +       atomic_t load_updates;          /* nr of load updates done so far */
12116 +       uint32_t load_last;             /* last time load was calculated */
12117 +       uint32_t load[3];               /* load averages 1,5,15 */
12118 +
12119 +       atomic_t total_forks;           /* number of forks so far */
12120 +
12121 +       struct _vx_syslog syslog;
12122 +};
12123 +
12124 +struct _vx_cvirt_pc {
12125 +       struct _vx_usage_stat cpustat;
12126 +};
12127 +
12128 +
12129 +#ifdef CONFIG_VSERVER_DEBUG
12130 +
12131 +static inline void __dump_vx_cvirt(struct _vx_cvirt *cvirt)
12132 +{
12133 +       printk("\t_vx_cvirt:\n");
12134 +       printk("\t threads: %4d, %4d, %4d, %4d\n",
12135 +               atomic_read(&cvirt->nr_threads),
12136 +               atomic_read(&cvirt->nr_running),
12137 +               atomic_read(&cvirt->nr_uninterruptible),
12138 +               atomic_read(&cvirt->nr_onhold));
12139 +       /* add rest here */
12140 +       printk("\t total_forks = %d\n", atomic_read(&cvirt->total_forks));
12141 +}
12142 +
12143 +#endif
12144 +
12145 +#endif /* _VX_CVIRT_DEF_H */
12146 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/debug.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/debug.h
12147 --- linux-3.0-rc5/include/linux/vserver/debug.h 1970-01-01 01:00:00.000000000 +0100
12148 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/debug.h    2011-06-10 22:11:24.000000000 +0200
12149 @@ -0,0 +1,140 @@
12150 +#ifndef _VX_DEBUG_H
12151 +#define _VX_DEBUG_H
12152 +
12153 +
12154 +#define VXD_CBIT(n, m) (vx_debug_ ## n & (1 << (m)))
12155 +#define VXD_CMIN(n, m) (vx_debug_ ## n > (m))
12156 +#define VXD_MASK(n, m) (vx_debug_ ## n & (m))
12157 +
12158 +#define VXD_DEV(d)     (d), (d)->bd_inode->i_ino,              \
12159 +                       imajor((d)->bd_inode), iminor((d)->bd_inode)
12160 +#define VXF_DEV                "%p[%lu,%d:%d]"
12161 +
12162 +#if    defined(CONFIG_QUOTES_UTF8)
12163 +#define        VS_Q_LQM        "\xc2\xbb"
12164 +#define        VS_Q_RQM        "\xc2\xab"
12165 +#elif  defined(CONFIG_QUOTES_ASCII)
12166 +#define        VS_Q_LQM        "\x27"
12167 +#define        VS_Q_RQM        "\x27"
12168 +#else
12169 +#define        VS_Q_LQM        "\xbb"
12170 +#define        VS_Q_RQM        "\xab"
12171 +#endif
12172 +
12173 +#define        VS_Q(f)         VS_Q_LQM f VS_Q_RQM
12174 +
12175 +
12176 +#define vxd_path(p)                                            \
12177 +       ({ static char _buffer[PATH_MAX];                       \
12178 +          d_path(p, _buffer, sizeof(_buffer)); })
12179 +
12180 +#define vxd_cond_path(n)                                       \
12181 +       ((n) ? vxd_path(&(n)->path) : "<null>" )
12182 +
12183 +
12184 +#ifdef CONFIG_VSERVER_DEBUG
12185 +
12186 +extern unsigned int vx_debug_switch;
12187 +extern unsigned int vx_debug_xid;
12188 +extern unsigned int vx_debug_nid;
12189 +extern unsigned int vx_debug_tag;
12190 +extern unsigned int vx_debug_net;
12191 +extern unsigned int vx_debug_limit;
12192 +extern unsigned int vx_debug_cres;
12193 +extern unsigned int vx_debug_dlim;
12194 +extern unsigned int vx_debug_quota;
12195 +extern unsigned int vx_debug_cvirt;
12196 +extern unsigned int vx_debug_space;
12197 +extern unsigned int vx_debug_misc;
12198 +
12199 +
12200 +#define VX_LOGLEVEL    "vxD: "
12201 +#define VX_PROC_FMT    "%p: "
12202 +#define VX_PROCESS     current
12203 +
12204 +#define vxdprintk(c, f, x...)                                  \
12205 +       do {                                                    \
12206 +               if (c)                                          \
12207 +                       printk(VX_LOGLEVEL VX_PROC_FMT f "\n",  \
12208 +                               VX_PROCESS , ##x);              \
12209 +       } while (0)
12210 +
12211 +#define vxlprintk(c, f, x...)                                  \
12212 +       do {                                                    \
12213 +               if (c)                                          \
12214 +                       printk(VX_LOGLEVEL f " @%s:%d\n", x);   \
12215 +       } while (0)
12216 +
12217 +#define vxfprintk(c, f, x...)                                  \
12218 +       do {                                                    \
12219 +               if (c)                                          \
12220 +                       printk(VX_LOGLEVEL f " %s@%s:%d\n", x); \
12221 +       } while (0)
12222 +
12223 +
12224 +struct vx_info;
12225 +
12226 +void dump_vx_info(struct vx_info *, int);
12227 +void dump_vx_info_inactive(int);
12228 +
12229 +#else  /* CONFIG_VSERVER_DEBUG */
12230 +
12231 +#define vx_debug_switch 0
12232 +#define vx_debug_xid   0
12233 +#define vx_debug_nid   0
12234 +#define vx_debug_tag   0
12235 +#define vx_debug_net   0
12236 +#define vx_debug_limit 0
12237 +#define vx_debug_cres  0
12238 +#define vx_debug_dlim  0
12239 +#define vx_debug_cvirt 0
12240 +
12241 +#define vxdprintk(x...) do { } while (0)
12242 +#define vxlprintk(x...) do { } while (0)
12243 +#define vxfprintk(x...) do { } while (0)
12244 +
12245 +#endif /* CONFIG_VSERVER_DEBUG */
12246 +
12247 +
12248 +#ifdef CONFIG_VSERVER_WARN
12249 +
12250 +#define VX_WARNLEVEL   KERN_WARNING "vxW: "
12251 +#define VX_WARN_TASK   "[" VS_Q("%s") ",%u:#%u|%u|%u] "
12252 +#define VX_WARN_XID    "[xid #%u] "
12253 +#define VX_WARN_NID    "[nid #%u] "
12254 +#define VX_WARN_TAG    "[tag #%u] "
12255 +
12256 +#define vxwprintk(c, f, x...)                                  \
12257 +       do {                                                    \
12258 +               if (c)                                          \
12259 +                       printk(VX_WARNLEVEL f "\n", ##x);       \
12260 +       } while (0)
12261 +
12262 +#else  /* CONFIG_VSERVER_WARN */
12263 +
12264 +#define vxwprintk(x...) do { } while (0)
12265 +
12266 +#endif /* CONFIG_VSERVER_WARN */
12267 +
12268 +#define vxwprintk_task(c, f, x...)                             \
12269 +       vxwprintk(c, VX_WARN_TASK f,                            \
12270 +               current->comm, current->pid,                    \
12271 +               current->xid, current->nid, current->tag, ##x)
12272 +#define vxwprintk_xid(c, f, x...)                              \
12273 +       vxwprintk(c, VX_WARN_XID f, current->xid, x)
12274 +#define vxwprintk_nid(c, f, x...)                              \
12275 +       vxwprintk(c, VX_WARN_NID f, current->nid, x)
12276 +#define vxwprintk_tag(c, f, x...)                              \
12277 +       vxwprintk(c, VX_WARN_TAG f, current->tag, x)
12278 +
12279 +#ifdef CONFIG_VSERVER_DEBUG
12280 +#define vxd_assert_lock(l)     assert_spin_locked(l)
12281 +#define vxd_assert(c, f, x...) vxlprintk(!(c), \
12282 +       "assertion [" f "] failed.", ##x, __FILE__, __LINE__)
12283 +#else
12284 +#define vxd_assert_lock(l)     do { } while (0)
12285 +#define vxd_assert(c, f, x...) do { } while (0)
12286 +#endif
12287 +
12288 +
12289 +#endif /* _VX_DEBUG_H */
12290 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/debug_cmd.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/debug_cmd.h
12291 --- linux-3.0-rc5/include/linux/vserver/debug_cmd.h     1970-01-01 01:00:00.000000000 +0100
12292 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/debug_cmd.h        2011-06-10 22:11:24.000000000 +0200
12293 @@ -0,0 +1,58 @@
12294 +#ifndef _VX_DEBUG_CMD_H
12295 +#define _VX_DEBUG_CMD_H
12296 +
12297 +
12298 +/* debug commands */
12299 +
12300 +#define VCMD_dump_history      VC_CMD(DEBUG, 1, 0)
12301 +
12302 +#define VCMD_read_history      VC_CMD(DEBUG, 5, 0)
12303 +#define VCMD_read_monitor      VC_CMD(DEBUG, 6, 0)
12304 +
12305 +struct  vcmd_read_history_v0 {
12306 +       uint32_t index;
12307 +       uint32_t count;
12308 +       char __user *data;
12309 +};
12310 +
12311 +struct  vcmd_read_monitor_v0 {
12312 +       uint32_t index;
12313 +       uint32_t count;
12314 +       char __user *data;
12315 +};
12316 +
12317 +
12318 +#ifdef __KERNEL__
12319 +
12320 +#ifdef CONFIG_COMPAT
12321 +
12322 +#include <asm/compat.h>
12323 +
12324 +struct vcmd_read_history_v0_x32 {
12325 +       uint32_t index;
12326 +       uint32_t count;
12327 +       compat_uptr_t data_ptr;
12328 +};
12329 +
12330 +struct vcmd_read_monitor_v0_x32 {
12331 +       uint32_t index;
12332 +       uint32_t count;
12333 +       compat_uptr_t data_ptr;
12334 +};
12335 +
12336 +#endif  /* CONFIG_COMPAT */
12337 +
12338 +extern int vc_dump_history(uint32_t);
12339 +
12340 +extern int vc_read_history(uint32_t, void __user *);
12341 +extern int vc_read_monitor(uint32_t, void __user *);
12342 +
12343 +#ifdef CONFIG_COMPAT
12344 +
12345 +extern int vc_read_history_x32(uint32_t, void __user *);
12346 +extern int vc_read_monitor_x32(uint32_t, void __user *);
12347 +
12348 +#endif  /* CONFIG_COMPAT */
12349 +
12350 +#endif /* __KERNEL__ */
12351 +#endif /* _VX_DEBUG_CMD_H */
12352 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/device.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/device.h
12353 --- linux-3.0-rc5/include/linux/vserver/device.h        1970-01-01 01:00:00.000000000 +0100
12354 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/device.h   2011-06-10 22:11:24.000000000 +0200
12355 @@ -0,0 +1,15 @@
12356 +#ifndef _VX_DEVICE_H
12357 +#define _VX_DEVICE_H
12358 +
12359 +
12360 +#define DATTR_CREATE   0x00000001
12361 +#define DATTR_OPEN     0x00000002
12362 +
12363 +#define DATTR_REMAP    0x00000010
12364 +
12365 +#define DATTR_MASK     0x00000013
12366 +
12367 +
12368 +#else  /* _VX_DEVICE_H */
12369 +#warning duplicate inclusion
12370 +#endif /* _VX_DEVICE_H */
12371 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/device_cmd.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/device_cmd.h
12372 --- linux-3.0-rc5/include/linux/vserver/device_cmd.h    1970-01-01 01:00:00.000000000 +0100
12373 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/device_cmd.h       2011-06-10 22:11:24.000000000 +0200
12374 @@ -0,0 +1,44 @@
12375 +#ifndef _VX_DEVICE_CMD_H
12376 +#define _VX_DEVICE_CMD_H
12377 +
12378 +
12379 +/*  device vserver commands */
12380 +
12381 +#define VCMD_set_mapping       VC_CMD(DEVICE, 1, 0)
12382 +#define VCMD_unset_mapping     VC_CMD(DEVICE, 2, 0)
12383 +
12384 +struct vcmd_set_mapping_v0 {
12385 +       const char __user *device;
12386 +       const char __user *target;
12387 +       uint32_t flags;
12388 +};
12389 +
12390 +
12391 +#ifdef __KERNEL__
12392 +
12393 +#ifdef CONFIG_COMPAT
12394 +
12395 +#include <asm/compat.h>
12396 +
12397 +struct vcmd_set_mapping_v0_x32 {
12398 +       compat_uptr_t device_ptr;
12399 +       compat_uptr_t target_ptr;
12400 +       uint32_t flags;
12401 +};
12402 +
12403 +#endif /* CONFIG_COMPAT */
12404 +
12405 +#include <linux/compiler.h>
12406 +
12407 +extern int vc_set_mapping(struct vx_info *, void __user *);
12408 +extern int vc_unset_mapping(struct vx_info *, void __user *);
12409 +
12410 +#ifdef CONFIG_COMPAT
12411 +
12412 +extern int vc_set_mapping_x32(struct vx_info *, void __user *);
12413 +extern int vc_unset_mapping_x32(struct vx_info *, void __user *);
12414 +
12415 +#endif /* CONFIG_COMPAT */
12416 +
12417 +#endif /* __KERNEL__ */
12418 +#endif /* _VX_DEVICE_CMD_H */
12419 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/device_def.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/device_def.h
12420 --- linux-3.0-rc5/include/linux/vserver/device_def.h    1970-01-01 01:00:00.000000000 +0100
12421 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/device_def.h       2011-06-10 22:11:24.000000000 +0200
12422 @@ -0,0 +1,17 @@
12423 +#ifndef _VX_DEVICE_DEF_H
12424 +#define _VX_DEVICE_DEF_H
12425 +
12426 +#include <linux/types.h>
12427 +
12428 +struct vx_dmap_target {
12429 +       dev_t target;
12430 +       uint32_t flags;
12431 +};
12432 +
12433 +struct _vx_device {
12434 +#ifdef CONFIG_VSERVER_DEVICE
12435 +       struct vx_dmap_target targets[2];
12436 +#endif
12437 +};
12438 +
12439 +#endif /* _VX_DEVICE_DEF_H */
12440 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/dlimit.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/dlimit.h
12441 --- linux-3.0-rc5/include/linux/vserver/dlimit.h        1970-01-01 01:00:00.000000000 +0100
12442 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/dlimit.h   2011-06-10 22:11:24.000000000 +0200
12443 @@ -0,0 +1,54 @@
12444 +#ifndef _VX_DLIMIT_H
12445 +#define _VX_DLIMIT_H
12446 +
12447 +#include "switch.h"
12448 +
12449 +
12450 +#ifdef __KERNEL__
12451 +
12452 +/*      keep in sync with CDLIM_INFINITY       */
12453 +
12454 +#define DLIM_INFINITY          (~0ULL)
12455 +
12456 +#include <linux/spinlock.h>
12457 +#include <linux/rcupdate.h>
12458 +
12459 +struct super_block;
12460 +
12461 +struct dl_info {
12462 +       struct hlist_node dl_hlist;             /* linked list of contexts */
12463 +       struct rcu_head dl_rcu;                 /* the rcu head */
12464 +       tag_t dl_tag;                           /* context tag */
12465 +       atomic_t dl_usecnt;                     /* usage count */
12466 +       atomic_t dl_refcnt;                     /* reference count */
12467 +
12468 +       struct super_block *dl_sb;              /* associated superblock */
12469 +
12470 +       spinlock_t dl_lock;                     /* protect the values */
12471 +
12472 +       unsigned long long dl_space_used;       /* used space in bytes */
12473 +       unsigned long long dl_space_total;      /* maximum space in bytes */
12474 +       unsigned long dl_inodes_used;           /* used inodes */
12475 +       unsigned long dl_inodes_total;          /* maximum inodes */
12476 +
12477 +       unsigned int dl_nrlmult;                /* non root limit mult */
12478 +};
12479 +
12480 +struct rcu_head;
12481 +
12482 +extern void rcu_free_dl_info(struct rcu_head *);
12483 +extern void unhash_dl_info(struct dl_info *);
12484 +
12485 +extern struct dl_info *locate_dl_info(struct super_block *, tag_t);
12486 +
12487 +
12488 +struct kstatfs;
12489 +
12490 +extern void vx_vsi_statfs(struct super_block *, struct kstatfs *);
12491 +
12492 +typedef uint64_t dlsize_t;
12493 +
12494 +#endif /* __KERNEL__ */
12495 +#else  /* _VX_DLIMIT_H */
12496 +#warning duplicate inclusion
12497 +#endif /* _VX_DLIMIT_H */
12498 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/dlimit_cmd.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/dlimit_cmd.h
12499 --- linux-3.0-rc5/include/linux/vserver/dlimit_cmd.h    1970-01-01 01:00:00.000000000 +0100
12500 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/dlimit_cmd.h       2011-06-10 22:11:24.000000000 +0200
12501 @@ -0,0 +1,109 @@
12502 +#ifndef _VX_DLIMIT_CMD_H
12503 +#define _VX_DLIMIT_CMD_H
12504 +
12505 +
12506 +/*  dlimit vserver commands */
12507 +
12508 +#define VCMD_add_dlimit                VC_CMD(DLIMIT, 1, 0)
12509 +#define VCMD_rem_dlimit                VC_CMD(DLIMIT, 2, 0)
12510 +
12511 +#define VCMD_set_dlimit                VC_CMD(DLIMIT, 5, 0)
12512 +#define VCMD_get_dlimit                VC_CMD(DLIMIT, 6, 0)
12513 +
12514 +struct vcmd_ctx_dlimit_base_v0 {
12515 +       const char __user *name;
12516 +       uint32_t flags;
12517 +};
12518 +
12519 +struct vcmd_ctx_dlimit_v0 {
12520 +       const char __user *name;
12521 +       uint32_t space_used;                    /* used space in kbytes */
12522 +       uint32_t space_total;                   /* maximum space in kbytes */
12523 +       uint32_t inodes_used;                   /* used inodes */
12524 +       uint32_t inodes_total;                  /* maximum inodes */
12525 +       uint32_t reserved;                      /* reserved for root in % */
12526 +       uint32_t flags;
12527 +};
12528 +
12529 +#define CDLIM_UNSET            ((uint32_t)0UL)
12530 +#define CDLIM_INFINITY         ((uint32_t)~0UL)
12531 +#define CDLIM_KEEP             ((uint32_t)~1UL)
12532 +
12533 +#define DLIME_UNIT     0
12534 +#define DLIME_KILO     1
12535 +#define DLIME_MEGA     2
12536 +#define DLIME_GIGA     3
12537 +
12538 +#define DLIMF_SHIFT    0x10
12539 +
12540 +#define DLIMS_USED     0
12541 +#define DLIMS_TOTAL    2
12542 +
12543 +static inline
12544 +uint64_t dlimit_space_32to64(uint32_t val, uint32_t flags, int shift)
12545 +{
12546 +       int exp = (flags & DLIMF_SHIFT) ?
12547 +               (flags >> shift) & DLIME_GIGA : DLIME_KILO;
12548 +       return ((uint64_t)val) << (10 * exp);
12549 +}
12550 +
12551 +static inline
12552 +uint32_t dlimit_space_64to32(uint64_t val, uint32_t *flags, int shift)
12553 +{
12554 +       int exp = 0;
12555 +
12556 +       if (*flags & DLIMF_SHIFT) {
12557 +               while (val > (1LL << 32) && (exp < 3)) {
12558 +                       val >>= 10;
12559 +                       exp++;
12560 +               }
12561 +               *flags &= ~(DLIME_GIGA << shift);
12562 +               *flags |= exp << shift;
12563 +       } else
12564 +               val >>= 10;
12565 +       return val;
12566 +}
12567 +
12568 +#ifdef __KERNEL__
12569 +
12570 +#ifdef CONFIG_COMPAT
12571 +
12572 +#include <asm/compat.h>
12573 +
12574 +struct vcmd_ctx_dlimit_base_v0_x32 {
12575 +       compat_uptr_t name_ptr;
12576 +       uint32_t flags;
12577 +};
12578 +
12579 +struct vcmd_ctx_dlimit_v0_x32 {
12580 +       compat_uptr_t name_ptr;
12581 +       uint32_t space_used;                    /* used space in kbytes */
12582 +       uint32_t space_total;                   /* maximum space in kbytes */
12583 +       uint32_t inodes_used;                   /* used inodes */
12584 +       uint32_t inodes_total;                  /* maximum inodes */
12585 +       uint32_t reserved;                      /* reserved for root in % */
12586 +       uint32_t flags;
12587 +};
12588 +
12589 +#endif /* CONFIG_COMPAT */
12590 +
12591 +#include <linux/compiler.h>
12592 +
12593 +extern int vc_add_dlimit(uint32_t, void __user *);
12594 +extern int vc_rem_dlimit(uint32_t, void __user *);
12595 +
12596 +extern int vc_set_dlimit(uint32_t, void __user *);
12597 +extern int vc_get_dlimit(uint32_t, void __user *);
12598 +
12599 +#ifdef CONFIG_COMPAT
12600 +
12601 +extern int vc_add_dlimit_x32(uint32_t, void __user *);
12602 +extern int vc_rem_dlimit_x32(uint32_t, void __user *);
12603 +
12604 +extern int vc_set_dlimit_x32(uint32_t, void __user *);
12605 +extern int vc_get_dlimit_x32(uint32_t, void __user *);
12606 +
12607 +#endif /* CONFIG_COMPAT */
12608 +
12609 +#endif /* __KERNEL__ */
12610 +#endif /* _VX_DLIMIT_CMD_H */
12611 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/global.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/global.h
12612 --- linux-3.0-rc5/include/linux/vserver/global.h        1970-01-01 01:00:00.000000000 +0100
12613 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/global.h   2011-06-10 22:11:24.000000000 +0200
12614 @@ -0,0 +1,19 @@
12615 +#ifndef _VX_GLOBAL_H
12616 +#define _VX_GLOBAL_H
12617 +
12618 +
12619 +extern atomic_t vx_global_ctotal;
12620 +extern atomic_t vx_global_cactive;
12621 +
12622 +extern atomic_t nx_global_ctotal;
12623 +extern atomic_t nx_global_cactive;
12624 +
12625 +extern atomic_t vs_global_nsproxy;
12626 +extern atomic_t vs_global_fs;
12627 +extern atomic_t vs_global_mnt_ns;
12628 +extern atomic_t vs_global_uts_ns;
12629 +extern atomic_t vs_global_user_ns;
12630 +extern atomic_t vs_global_pid_ns;
12631 +
12632 +
12633 +#endif /* _VX_GLOBAL_H */
12634 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/history.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/history.h
12635 --- linux-3.0-rc5/include/linux/vserver/history.h       1970-01-01 01:00:00.000000000 +0100
12636 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/history.h  2011-06-10 22:11:24.000000000 +0200
12637 @@ -0,0 +1,197 @@
12638 +#ifndef _VX_HISTORY_H
12639 +#define _VX_HISTORY_H
12640 +
12641 +
12642 +enum {
12643 +       VXH_UNUSED = 0,
12644 +       VXH_THROW_OOPS = 1,
12645 +
12646 +       VXH_GET_VX_INFO,
12647 +       VXH_PUT_VX_INFO,
12648 +       VXH_INIT_VX_INFO,
12649 +       VXH_SET_VX_INFO,
12650 +       VXH_CLR_VX_INFO,
12651 +       VXH_CLAIM_VX_INFO,
12652 +       VXH_RELEASE_VX_INFO,
12653 +       VXH_ALLOC_VX_INFO,
12654 +       VXH_DEALLOC_VX_INFO,
12655 +       VXH_HASH_VX_INFO,
12656 +       VXH_UNHASH_VX_INFO,
12657 +       VXH_LOC_VX_INFO,
12658 +       VXH_LOOKUP_VX_INFO,
12659 +       VXH_CREATE_VX_INFO,
12660 +};
12661 +
12662 +struct _vxhe_vxi {
12663 +       struct vx_info *ptr;
12664 +       unsigned xid;
12665 +       unsigned usecnt;
12666 +       unsigned tasks;
12667 +};
12668 +
12669 +struct _vxhe_set_clr {
12670 +       void *data;
12671 +};
12672 +
12673 +struct _vxhe_loc_lookup {
12674 +       unsigned arg;
12675 +};
12676 +
12677 +struct _vx_hist_entry {
12678 +       void *loc;
12679 +       unsigned short seq;
12680 +       unsigned short type;
12681 +       struct _vxhe_vxi vxi;
12682 +       union {
12683 +               struct _vxhe_set_clr sc;
12684 +               struct _vxhe_loc_lookup ll;
12685 +       };
12686 +};
12687 +
12688 +#ifdef CONFIG_VSERVER_HISTORY
12689 +
12690 +extern unsigned volatile int vxh_active;
12691 +
12692 +struct _vx_hist_entry *vxh_advance(void *loc);
12693 +
12694 +
12695 +static inline
12696 +void   __vxh_copy_vxi(struct _vx_hist_entry *entry, struct vx_info *vxi)
12697 +{
12698 +       entry->vxi.ptr = vxi;
12699 +       if (vxi) {
12700 +               entry->vxi.usecnt = atomic_read(&vxi->vx_usecnt);
12701 +               entry->vxi.tasks = atomic_read(&vxi->vx_tasks);
12702 +               entry->vxi.xid = vxi->vx_id;
12703 +       }
12704 +}
12705 +
12706 +
12707 +#define        __HERE__ current_text_addr()
12708 +
12709 +#define __VXH_BODY(__type, __data, __here)     \
12710 +       struct _vx_hist_entry *entry;           \
12711 +                                               \
12712 +       preempt_disable();                      \
12713 +       entry = vxh_advance(__here);            \
12714 +       __data;                                 \
12715 +       entry->type = __type;                   \
12716 +       preempt_enable();
12717 +
12718 +
12719 +       /* pass vxi only */
12720 +
12721 +#define __VXH_SMPL                             \
12722 +       __vxh_copy_vxi(entry, vxi)
12723 +
12724 +static inline
12725 +void   __vxh_smpl(struct vx_info *vxi, int __type, void *__here)
12726 +{
12727 +       __VXH_BODY(__type, __VXH_SMPL, __here)
12728 +}
12729 +
12730 +       /* pass vxi and data (void *) */
12731 +
12732 +#define __VXH_DATA                             \
12733 +       __vxh_copy_vxi(entry, vxi);             \
12734 +       entry->sc.data = data
12735 +
12736 +static inline
12737 +void   __vxh_data(struct vx_info *vxi, void *data,
12738 +                       int __type, void *__here)
12739 +{
12740 +       __VXH_BODY(__type, __VXH_DATA, __here)
12741 +}
12742 +
12743 +       /* pass vxi and arg (long) */
12744 +
12745 +#define __VXH_LONG                             \
12746 +       __vxh_copy_vxi(entry, vxi);             \
12747 +       entry->ll.arg = arg
12748 +
12749 +static inline
12750 +void   __vxh_long(struct vx_info *vxi, long arg,
12751 +                       int __type, void *__here)
12752 +{
12753 +       __VXH_BODY(__type, __VXH_LONG, __here)
12754 +}
12755 +
12756 +
12757 +static inline
12758 +void   __vxh_throw_oops(void *__here)
12759 +{
12760 +       __VXH_BODY(VXH_THROW_OOPS, {}, __here);
12761 +       /* prevent further acquisition */
12762 +       vxh_active = 0;
12763 +}
12764 +
12765 +
12766 +#define vxh_throw_oops()       __vxh_throw_oops(__HERE__);
12767 +
12768 +#define __vxh_get_vx_info(v, h)        __vxh_smpl(v, VXH_GET_VX_INFO, h);
12769 +#define __vxh_put_vx_info(v, h)        __vxh_smpl(v, VXH_PUT_VX_INFO, h);
12770 +
12771 +#define __vxh_init_vx_info(v, d, h) \
12772 +       __vxh_data(v, d, VXH_INIT_VX_INFO, h);
12773 +#define __vxh_set_vx_info(v, d, h) \
12774 +       __vxh_data(v, d, VXH_SET_VX_INFO, h);
12775 +#define __vxh_clr_vx_info(v, d, h) \
12776 +       __vxh_data(v, d, VXH_CLR_VX_INFO, h);
12777 +
12778 +#define __vxh_claim_vx_info(v, d, h) \
12779 +       __vxh_data(v, d, VXH_CLAIM_VX_INFO, h);
12780 +#define __vxh_release_vx_info(v, d, h) \
12781 +       __vxh_data(v, d, VXH_RELEASE_VX_INFO, h);
12782 +
12783 +#define vxh_alloc_vx_info(v) \
12784 +       __vxh_smpl(v, VXH_ALLOC_VX_INFO, __HERE__);
12785 +#define vxh_dealloc_vx_info(v) \
12786 +       __vxh_smpl(v, VXH_DEALLOC_VX_INFO, __HERE__);
12787 +
12788 +#define vxh_hash_vx_info(v) \
12789 +       __vxh_smpl(v, VXH_HASH_VX_INFO, __HERE__);
12790 +#define vxh_unhash_vx_info(v) \
12791 +       __vxh_smpl(v, VXH_UNHASH_VX_INFO, __HERE__);
12792 +
12793 +#define vxh_loc_vx_info(v, l) \
12794 +       __vxh_long(v, l, VXH_LOC_VX_INFO, __HERE__);
12795 +#define vxh_lookup_vx_info(v, l) \
12796 +       __vxh_long(v, l, VXH_LOOKUP_VX_INFO, __HERE__);
12797 +#define vxh_create_vx_info(v, l) \
12798 +       __vxh_long(v, l, VXH_CREATE_VX_INFO, __HERE__);
12799 +
12800 +extern void vxh_dump_history(void);
12801 +
12802 +
12803 +#else  /* CONFIG_VSERVER_HISTORY */
12804 +
12805 +#define        __HERE__        0
12806 +
12807 +#define vxh_throw_oops()               do { } while (0)
12808 +
12809 +#define __vxh_get_vx_info(v, h)                do { } while (0)
12810 +#define __vxh_put_vx_info(v, h)                do { } while (0)
12811 +
12812 +#define __vxh_init_vx_info(v, d, h)    do { } while (0)
12813 +#define __vxh_set_vx_info(v, d, h)     do { } while (0)
12814 +#define __vxh_clr_vx_info(v, d, h)     do { } while (0)
12815 +
12816 +#define __vxh_claim_vx_info(v, d, h)   do { } while (0)
12817 +#define __vxh_release_vx_info(v, d, h) do { } while (0)
12818 +
12819 +#define vxh_alloc_vx_info(v)           do { } while (0)
12820 +#define vxh_dealloc_vx_info(v)         do { } while (0)
12821 +
12822 +#define vxh_hash_vx_info(v)            do { } while (0)
12823 +#define vxh_unhash_vx_info(v)          do { } while (0)
12824 +
12825 +#define vxh_loc_vx_info(v, l)          do { } while (0)
12826 +#define vxh_lookup_vx_info(v, l)       do { } while (0)
12827 +#define vxh_create_vx_info(v, l)       do { } while (0)
12828 +
12829 +#define vxh_dump_history()             do { } while (0)
12830 +
12831 +
12832 +#endif /* CONFIG_VSERVER_HISTORY */
12833 +
12834 +#endif /* _VX_HISTORY_H */
12835 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/inode.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/inode.h
12836 --- linux-3.0-rc5/include/linux/vserver/inode.h 1970-01-01 01:00:00.000000000 +0100
12837 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/inode.h    2011-06-10 22:11:24.000000000 +0200
12838 @@ -0,0 +1,39 @@
12839 +#ifndef _VX_INODE_H
12840 +#define _VX_INODE_H
12841 +
12842 +
12843 +#define IATTR_TAG      0x01000000
12844 +
12845 +#define IATTR_ADMIN    0x00000001
12846 +#define IATTR_WATCH    0x00000002
12847 +#define IATTR_HIDE     0x00000004
12848 +#define IATTR_FLAGS    0x00000007
12849 +
12850 +#define IATTR_BARRIER  0x00010000
12851 +#define IATTR_IXUNLINK 0x00020000
12852 +#define IATTR_IMMUTABLE 0x00040000
12853 +#define IATTR_COW      0x00080000
12854 +
12855 +#ifdef __KERNEL__
12856 +
12857 +
12858 +#ifdef CONFIG_VSERVER_PROC_SECURE
12859 +#define IATTR_PROC_DEFAULT     ( IATTR_ADMIN | IATTR_HIDE )
12860 +#define IATTR_PROC_SYMLINK     ( IATTR_ADMIN )
12861 +#else
12862 +#define IATTR_PROC_DEFAULT     ( IATTR_ADMIN )
12863 +#define IATTR_PROC_SYMLINK     ( IATTR_ADMIN )
12864 +#endif
12865 +
12866 +#define vx_hide_check(c, m)    (((m) & IATTR_HIDE) ? vx_check(c, m) : 1)
12867 +
12868 +#endif /* __KERNEL__ */
12869 +
12870 +/* inode ioctls */
12871 +
12872 +#define FIOC_GETXFLG   _IOR('x', 5, long)
12873 +#define FIOC_SETXFLG   _IOW('x', 6, long)
12874 +
12875 +#else  /* _VX_INODE_H */
12876 +#warning duplicate inclusion
12877 +#endif /* _VX_INODE_H */
12878 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/inode_cmd.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/inode_cmd.h
12879 --- linux-3.0-rc5/include/linux/vserver/inode_cmd.h     1970-01-01 01:00:00.000000000 +0100
12880 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/inode_cmd.h        2011-06-10 22:11:24.000000000 +0200
12881 @@ -0,0 +1,59 @@
12882 +#ifndef _VX_INODE_CMD_H
12883 +#define _VX_INODE_CMD_H
12884 +
12885 +
12886 +/*  inode vserver commands */
12887 +
12888 +#define VCMD_get_iattr         VC_CMD(INODE, 1, 1)
12889 +#define VCMD_set_iattr         VC_CMD(INODE, 2, 1)
12890 +
12891 +#define VCMD_fget_iattr                VC_CMD(INODE, 3, 0)
12892 +#define VCMD_fset_iattr                VC_CMD(INODE, 4, 0)
12893 +
12894 +struct vcmd_ctx_iattr_v1 {
12895 +       const char __user *name;
12896 +       uint32_t tag;
12897 +       uint32_t flags;
12898 +       uint32_t mask;
12899 +};
12900 +
12901 +struct vcmd_ctx_fiattr_v0 {
12902 +       uint32_t tag;
12903 +       uint32_t flags;
12904 +       uint32_t mask;
12905 +};
12906 +
12907 +
12908 +#ifdef __KERNEL__
12909 +
12910 +
12911 +#ifdef CONFIG_COMPAT
12912 +
12913 +#include <asm/compat.h>
12914 +
12915 +struct vcmd_ctx_iattr_v1_x32 {
12916 +       compat_uptr_t name_ptr;
12917 +       uint32_t tag;
12918 +       uint32_t flags;
12919 +       uint32_t mask;
12920 +};
12921 +
12922 +#endif /* CONFIG_COMPAT */
12923 +
12924 +#include <linux/compiler.h>
12925 +
12926 +extern int vc_get_iattr(void __user *);
12927 +extern int vc_set_iattr(void __user *);
12928 +
12929 +extern int vc_fget_iattr(uint32_t, void __user *);
12930 +extern int vc_fset_iattr(uint32_t, void __user *);
12931 +
12932 +#ifdef CONFIG_COMPAT
12933 +
12934 +extern int vc_get_iattr_x32(void __user *);
12935 +extern int vc_set_iattr_x32(void __user *);
12936 +
12937 +#endif /* CONFIG_COMPAT */
12938 +
12939 +#endif /* __KERNEL__ */
12940 +#endif /* _VX_INODE_CMD_H */
12941 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/limit.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/limit.h
12942 --- linux-3.0-rc5/include/linux/vserver/limit.h 1970-01-01 01:00:00.000000000 +0100
12943 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/limit.h    2011-06-10 22:11:24.000000000 +0200
12944 @@ -0,0 +1,71 @@
12945 +#ifndef _VX_LIMIT_H
12946 +#define _VX_LIMIT_H
12947 +
12948 +#define VLIMIT_NSOCK   16
12949 +#define VLIMIT_OPENFD  17
12950 +#define VLIMIT_ANON    18
12951 +#define VLIMIT_SHMEM   19
12952 +#define VLIMIT_SEMARY  20
12953 +#define VLIMIT_NSEMS   21
12954 +#define VLIMIT_DENTRY  22
12955 +#define VLIMIT_MAPPED  23
12956 +
12957 +
12958 +#ifdef __KERNEL__
12959 +
12960 +#define        VLIM_NOCHECK    ((1L << VLIMIT_DENTRY) | (1L << RLIMIT_RSS))
12961 +
12962 +/*     keep in sync with CRLIM_INFINITY */
12963 +
12964 +#define        VLIM_INFINITY   (~0ULL)
12965 +
12966 +#include <asm/atomic.h>
12967 +#include <asm/resource.h>
12968 +
12969 +#ifndef RLIM_INFINITY
12970 +#warning RLIM_INFINITY is undefined
12971 +#endif
12972 +
12973 +#define __rlim_val(l, r, v)    ((l)->res[r].v)
12974 +
12975 +#define __rlim_soft(l, r)      __rlim_val(l, r, soft)
12976 +#define __rlim_hard(l, r)      __rlim_val(l, r, hard)
12977 +
12978 +#define __rlim_rcur(l, r)      __rlim_val(l, r, rcur)
12979 +#define __rlim_rmin(l, r)      __rlim_val(l, r, rmin)
12980 +#define __rlim_rmax(l, r)      __rlim_val(l, r, rmax)
12981 +
12982 +#define __rlim_lhit(l, r)      __rlim_val(l, r, lhit)
12983 +#define __rlim_hit(l, r)       atomic_inc(&__rlim_lhit(l, r))
12984 +
12985 +typedef atomic_long_t rlim_atomic_t;
12986 +typedef unsigned long rlim_t;
12987 +
12988 +#define __rlim_get(l, r)       atomic_long_read(&__rlim_rcur(l, r))
12989 +#define __rlim_set(l, r, v)    atomic_long_set(&__rlim_rcur(l, r), v)
12990 +#define __rlim_inc(l, r)       atomic_long_inc(&__rlim_rcur(l, r))
12991 +#define __rlim_dec(l, r)       atomic_long_dec(&__rlim_rcur(l, r))
12992 +#define __rlim_add(l, r, v)    atomic_long_add(v, &__rlim_rcur(l, r))
12993 +#define __rlim_sub(l, r, v)    atomic_long_sub(v, &__rlim_rcur(l, r))
12994 +
12995 +
12996 +#if    (RLIM_INFINITY == VLIM_INFINITY)
12997 +#define        VX_VLIM(r) ((long long)(long)(r))
12998 +#define        VX_RLIM(v) ((rlim_t)(v))
12999 +#else
13000 +#define        VX_VLIM(r) (((r) == RLIM_INFINITY) \
13001 +               ? VLIM_INFINITY : (long long)(r))
13002 +#define        VX_RLIM(v) (((v) == VLIM_INFINITY) \
13003 +               ? RLIM_INFINITY : (rlim_t)(v))
13004 +#endif
13005 +
13006 +struct sysinfo;
13007 +
13008 +void vx_vsi_meminfo(struct sysinfo *);
13009 +void vx_vsi_swapinfo(struct sysinfo *);
13010 +long vx_vsi_cached(struct sysinfo *);
13011 +
13012 +#define NUM_LIMITS     24
13013 +
13014 +#endif /* __KERNEL__ */
13015 +#endif /* _VX_LIMIT_H */
13016 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/limit_cmd.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/limit_cmd.h
13017 --- linux-3.0-rc5/include/linux/vserver/limit_cmd.h     1970-01-01 01:00:00.000000000 +0100
13018 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/limit_cmd.h        2011-06-10 22:11:24.000000000 +0200
13019 @@ -0,0 +1,71 @@
13020 +#ifndef _VX_LIMIT_CMD_H
13021 +#define _VX_LIMIT_CMD_H
13022 +
13023 +
13024 +/*  rlimit vserver commands */
13025 +
13026 +#define VCMD_get_rlimit                VC_CMD(RLIMIT, 1, 0)
13027 +#define VCMD_set_rlimit                VC_CMD(RLIMIT, 2, 0)
13028 +#define VCMD_get_rlimit_mask   VC_CMD(RLIMIT, 3, 0)
13029 +#define VCMD_reset_hits                VC_CMD(RLIMIT, 7, 0)
13030 +#define VCMD_reset_minmax      VC_CMD(RLIMIT, 9, 0)
13031 +
13032 +struct vcmd_ctx_rlimit_v0 {
13033 +       uint32_t id;
13034 +       uint64_t minimum;
13035 +       uint64_t softlimit;
13036 +       uint64_t maximum;
13037 +};
13038 +
13039 +struct vcmd_ctx_rlimit_mask_v0 {
13040 +       uint32_t minimum;
13041 +       uint32_t softlimit;
13042 +       uint32_t maximum;
13043 +};
13044 +
13045 +#define VCMD_rlimit_stat       VC_CMD(VSTAT, 1, 0)
13046 +
13047 +struct vcmd_rlimit_stat_v0 {
13048 +       uint32_t id;
13049 +       uint32_t hits;
13050 +       uint64_t value;
13051 +       uint64_t minimum;
13052 +       uint64_t maximum;
13053 +};
13054 +
13055 +#define CRLIM_UNSET            (0ULL)
13056 +#define CRLIM_INFINITY         (~0ULL)
13057 +#define CRLIM_KEEP             (~1ULL)
13058 +
13059 +#ifdef __KERNEL__
13060 +
13061 +#ifdef CONFIG_IA32_EMULATION
13062 +
13063 +struct vcmd_ctx_rlimit_v0_x32 {
13064 +       uint32_t id;
13065 +       uint64_t minimum;
13066 +       uint64_t softlimit;
13067 +       uint64_t maximum;
13068 +} __attribute__ ((packed));
13069 +
13070 +#endif /* CONFIG_IA32_EMULATION */
13071 +
13072 +#include <linux/compiler.h>
13073 +
13074 +extern int vc_get_rlimit_mask(uint32_t, void __user *);
13075 +extern int vc_get_rlimit(struct vx_info *, void __user *);
13076 +extern int vc_set_rlimit(struct vx_info *, void __user *);
13077 +extern int vc_reset_hits(struct vx_info *, void __user *);
13078 +extern int vc_reset_minmax(struct vx_info *, void __user *);
13079 +
13080 +extern int vc_rlimit_stat(struct vx_info *, void __user *);
13081 +
13082 +#ifdef CONFIG_IA32_EMULATION
13083 +
13084 +extern int vc_get_rlimit_x32(struct vx_info *, void __user *);
13085 +extern int vc_set_rlimit_x32(struct vx_info *, void __user *);
13086 +
13087 +#endif /* CONFIG_IA32_EMULATION */
13088 +
13089 +#endif /* __KERNEL__ */
13090 +#endif /* _VX_LIMIT_CMD_H */
13091 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/limit_def.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/limit_def.h
13092 --- linux-3.0-rc5/include/linux/vserver/limit_def.h     1970-01-01 01:00:00.000000000 +0100
13093 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/limit_def.h        2011-06-10 22:11:24.000000000 +0200
13094 @@ -0,0 +1,47 @@
13095 +#ifndef _VX_LIMIT_DEF_H
13096 +#define _VX_LIMIT_DEF_H
13097 +
13098 +#include <asm/atomic.h>
13099 +#include <asm/resource.h>
13100 +
13101 +#include "limit.h"
13102 +
13103 +
13104 +struct _vx_res_limit {
13105 +       rlim_t soft;            /* Context soft limit */
13106 +       rlim_t hard;            /* Context hard limit */
13107 +
13108 +       rlim_atomic_t rcur;     /* Current value */
13109 +       rlim_t rmin;            /* Context minimum */
13110 +       rlim_t rmax;            /* Context maximum */
13111 +
13112 +       atomic_t lhit;          /* Limit hits */
13113 +};
13114 +
13115 +/* context sub struct */
13116 +
13117 +struct _vx_limit {
13118 +       struct _vx_res_limit res[NUM_LIMITS];
13119 +};
13120 +
13121 +#ifdef CONFIG_VSERVER_DEBUG
13122 +
13123 +static inline void __dump_vx_limit(struct _vx_limit *limit)
13124 +{
13125 +       int i;
13126 +
13127 +       printk("\t_vx_limit:");
13128 +       for (i = 0; i < NUM_LIMITS; i++) {
13129 +               printk("\t [%2d] = %8lu %8lu/%8lu, %8ld/%8ld, %8d\n",
13130 +                       i, (unsigned long)__rlim_get(limit, i),
13131 +                       (unsigned long)__rlim_rmin(limit, i),
13132 +                       (unsigned long)__rlim_rmax(limit, i),
13133 +                       (long)__rlim_soft(limit, i),
13134 +                       (long)__rlim_hard(limit, i),
13135 +                       atomic_read(&__rlim_lhit(limit, i)));
13136 +       }
13137 +}
13138 +
13139 +#endif
13140 +
13141 +#endif /* _VX_LIMIT_DEF_H */
13142 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/limit_int.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/limit_int.h
13143 --- linux-3.0-rc5/include/linux/vserver/limit_int.h     1970-01-01 01:00:00.000000000 +0100
13144 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/limit_int.h        2011-06-10 22:11:24.000000000 +0200
13145 @@ -0,0 +1,198 @@
13146 +#ifndef _VX_LIMIT_INT_H
13147 +#define _VX_LIMIT_INT_H
13148 +
13149 +#include "context.h"
13150 +
13151 +#ifdef __KERNEL__
13152 +
13153 +#define VXD_RCRES_COND(r)      VXD_CBIT(cres, r)
13154 +#define VXD_RLIMIT_COND(r)     VXD_CBIT(limit, r)
13155 +
13156 +extern const char *vlimit_name[NUM_LIMITS];
13157 +
13158 +static inline void __vx_acc_cres(struct vx_info *vxi,
13159 +       int res, int dir, void *_data, char *_file, int _line)
13160 +{
13161 +       if (VXD_RCRES_COND(res))
13162 +               vxlprintk(1, "vx_acc_cres[%5d,%s,%2d]: %5ld%s (%p)",
13163 +                       (vxi ? vxi->vx_id : -1), vlimit_name[res], res,
13164 +                       (vxi ? (long)__rlim_get(&vxi->limit, res) : 0),
13165 +                       (dir > 0) ? "++" : "--", _data, _file, _line);
13166 +       if (!vxi)
13167 +               return;
13168 +
13169 +       if (dir > 0)
13170 +               __rlim_inc(&vxi->limit, res);
13171 +       else
13172 +               __rlim_dec(&vxi->limit, res);
13173 +}
13174 +
13175 +static inline void __vx_add_cres(struct vx_info *vxi,
13176 +       int res, int amount, void *_data, char *_file, int _line)
13177 +{
13178 +       if (VXD_RCRES_COND(res))
13179 +               vxlprintk(1, "vx_add_cres[%5d,%s,%2d]: %5ld += %5d (%p)",
13180 +                       (vxi ? vxi->vx_id : -1), vlimit_name[res], res,
13181 +                       (vxi ? (long)__rlim_get(&vxi->limit, res) : 0),
13182 +                       amount, _data, _file, _line);
13183 +       if (amount == 0)
13184 +               return;
13185 +       if (!vxi)
13186 +               return;
13187 +       __rlim_add(&vxi->limit, res, amount);
13188 +}
13189 +
13190 +static inline
13191 +int __vx_cres_adjust_max(struct _vx_limit *limit, int res, rlim_t value)
13192 +{
13193 +       int cond = (value > __rlim_rmax(limit, res));
13194 +
13195 +       if (cond)
13196 +               __rlim_rmax(limit, res) = value;
13197 +       return cond;
13198 +}
13199 +
13200 +static inline
13201 +int __vx_cres_adjust_min(struct _vx_limit *limit, int res, rlim_t value)
13202 +{
13203 +       int cond = (value < __rlim_rmin(limit, res));
13204 +
13205 +       if (cond)
13206 +               __rlim_rmin(limit, res) = value;
13207 +       return cond;
13208 +}
13209 +
13210 +static inline
13211 +void __vx_cres_fixup(struct _vx_limit *limit, int res, rlim_t value)
13212 +{
13213 +       if (!__vx_cres_adjust_max(limit, res, value))
13214 +               __vx_cres_adjust_min(limit, res, value);
13215 +}
13216 +
13217 +
13218 +/*     return values:
13219 +        +1 ... no limit hit
13220 +        -1 ... over soft limit
13221 +         0 ... over hard limit         */
13222 +
13223 +static inline int __vx_cres_avail(struct vx_info *vxi,
13224 +       int res, int num, char *_file, int _line)
13225 +{
13226 +       struct _vx_limit *limit;
13227 +       rlim_t value;
13228 +
13229 +       if (VXD_RLIMIT_COND(res))
13230 +               vxlprintk(1, "vx_cres_avail[%5d,%s,%2d]: %5ld/%5ld > %5ld + %5d",
13231 +                       (vxi ? vxi->vx_id : -1), vlimit_name[res], res,
13232 +                       (vxi ? (long)__rlim_soft(&vxi->limit, res) : -1),
13233 +                       (vxi ? (long)__rlim_hard(&vxi->limit, res) : -1),
13234 +                       (vxi ? (long)__rlim_get(&vxi->limit, res) : 0),
13235 +                       num, _file, _line);
13236 +       if (!vxi)
13237 +               return 1;
13238 +
13239 +       limit = &vxi->limit;
13240 +       value = __rlim_get(limit, res);
13241 +
13242 +       if (!__vx_cres_adjust_max(limit, res, value))
13243 +               __vx_cres_adjust_min(limit, res, value);
13244 +
13245 +       if (num == 0)
13246 +               return 1;
13247 +
13248 +       if (__rlim_soft(limit, res) == RLIM_INFINITY)
13249 +               return -1;
13250 +       if (value + num <= __rlim_soft(limit, res))
13251 +               return -1;
13252 +
13253 +       if (__rlim_hard(limit, res) == RLIM_INFINITY)
13254 +               return 1;
13255 +       if (value + num <= __rlim_hard(limit, res))
13256 +               return 1;
13257 +
13258 +       __rlim_hit(limit, res);
13259 +       return 0;
13260 +}
13261 +
13262 +
13263 +static const int VLA_RSS[] = { RLIMIT_RSS, VLIMIT_ANON, VLIMIT_MAPPED, 0 };
13264 +
13265 +static inline
13266 +rlim_t __vx_cres_array_sum(struct _vx_limit *limit, const int *array)
13267 +{
13268 +       rlim_t value, sum = 0;
13269 +       int res;
13270 +
13271 +       while ((res = *array++)) {
13272 +               value = __rlim_get(limit, res);
13273 +               __vx_cres_fixup(limit, res, value);
13274 +               sum += value;
13275 +       }
13276 +       return sum;
13277 +}
13278 +
13279 +static inline
13280 +rlim_t __vx_cres_array_fixup(struct _vx_limit *limit, const int *array)
13281 +{
13282 +       rlim_t value = __vx_cres_array_sum(limit, array + 1);
13283 +       int res = *array;
13284 +
13285 +       if (value == __rlim_get(limit, res))
13286 +               return value;
13287 +
13288 +       __rlim_set(limit, res, value);
13289 +       /* now adjust min/max */
13290 +       if (!__vx_cres_adjust_max(limit, res, value))
13291 +               __vx_cres_adjust_min(limit, res, value);
13292 +
13293 +       return value;
13294 +}
13295 +
13296 +static inline int __vx_cres_array_avail(struct vx_info *vxi,
13297 +       const int *array, int num, char *_file, int _line)
13298 +{
13299 +       struct _vx_limit *limit;
13300 +       rlim_t value = 0;
13301 +       int res;
13302 +
13303 +       if (num == 0)
13304 +               return 1;
13305 +       if (!vxi)
13306 +               return 1;
13307 +
13308 +       limit = &vxi->limit;
13309 +       res = *array;
13310 +       value = __vx_cres_array_sum(limit, array + 1);
13311 +
13312 +       __rlim_set(limit, res, value);
13313 +       __vx_cres_fixup(limit, res, value);
13314 +
13315 +       return __vx_cres_avail(vxi, res, num, _file, _line);
13316 +}
13317 +
13318 +
13319 +static inline void vx_limit_fixup(struct _vx_limit *limit, int id)
13320 +{
13321 +       rlim_t value;
13322 +       int res;
13323 +
13324 +       /* complex resources first */
13325 +       if ((id < 0) || (id == RLIMIT_RSS))
13326 +               __vx_cres_array_fixup(limit, VLA_RSS);
13327 +
13328 +       for (res = 0; res < NUM_LIMITS; res++) {
13329 +               if ((id > 0) && (res != id))
13330 +                       continue;
13331 +
13332 +               value = __rlim_get(limit, res);
13333 +               __vx_cres_fixup(limit, res, value);
13334 +
13335 +               /* not supposed to happen, maybe warn? */
13336 +               if (__rlim_rmax(limit, res) > __rlim_hard(limit, res))
13337 +                       __rlim_rmax(limit, res) = __rlim_hard(limit, res);
13338 +       }
13339 +}
13340 +
13341 +
13342 +#endif /* __KERNEL__ */
13343 +#endif /* _VX_LIMIT_INT_H */
13344 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/monitor.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/monitor.h
13345 --- linux-3.0-rc5/include/linux/vserver/monitor.h       1970-01-01 01:00:00.000000000 +0100
13346 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/monitor.h  2011-06-10 22:11:24.000000000 +0200
13347 @@ -0,0 +1,96 @@
13348 +#ifndef _VX_MONITOR_H
13349 +#define _VX_MONITOR_H
13350 +
13351 +#include <linux/types.h>
13352 +
13353 +enum {
13354 +       VXM_UNUSED = 0,
13355 +
13356 +       VXM_SYNC = 0x10,
13357 +
13358 +       VXM_UPDATE = 0x20,
13359 +       VXM_UPDATE_1,
13360 +       VXM_UPDATE_2,
13361 +
13362 +       VXM_RQINFO_1 = 0x24,
13363 +       VXM_RQINFO_2,
13364 +
13365 +       VXM_ACTIVATE = 0x40,
13366 +       VXM_DEACTIVATE,
13367 +       VXM_IDLE,
13368 +
13369 +       VXM_HOLD = 0x44,
13370 +       VXM_UNHOLD,
13371 +
13372 +       VXM_MIGRATE = 0x48,
13373 +       VXM_RESCHED,
13374 +
13375 +       /* all other bits are flags */
13376 +       VXM_SCHED = 0x80,
13377 +};
13378 +
13379 +struct _vxm_update_1 {
13380 +       uint32_t tokens_max;
13381 +       uint32_t fill_rate;
13382 +       uint32_t interval;
13383 +};
13384 +
13385 +struct _vxm_update_2 {
13386 +       uint32_t tokens_min;
13387 +       uint32_t fill_rate;
13388 +       uint32_t interval;
13389 +};
13390 +
13391 +struct _vxm_rqinfo_1 {
13392 +       uint16_t running;
13393 +       uint16_t onhold;
13394 +       uint16_t iowait;
13395 +       uint16_t uintr;
13396 +       uint32_t idle_tokens;
13397 +};
13398 +
13399 +struct _vxm_rqinfo_2 {
13400 +       uint32_t norm_time;
13401 +       uint32_t idle_time;
13402 +       uint32_t idle_skip;
13403 +};
13404 +
13405 +struct _vxm_sched {
13406 +       uint32_t tokens;
13407 +       uint32_t norm_time;
13408 +       uint32_t idle_time;
13409 +};
13410 +
13411 +struct _vxm_task {
13412 +       uint16_t pid;
13413 +       uint16_t state;
13414 +};
13415 +
13416 +struct _vxm_event {
13417 +       uint32_t jif;
13418 +       union {
13419 +               uint32_t seq;
13420 +               uint32_t sec;
13421 +       };
13422 +       union {
13423 +               uint32_t tokens;
13424 +               uint32_t nsec;
13425 +               struct _vxm_task tsk;
13426 +       };
13427 +};
13428 +
13429 +struct _vx_mon_entry {
13430 +       uint16_t type;
13431 +       uint16_t xid;
13432 +       union {
13433 +               struct _vxm_event ev;
13434 +               struct _vxm_sched sd;
13435 +               struct _vxm_update_1 u1;
13436 +               struct _vxm_update_2 u2;
13437 +               struct _vxm_rqinfo_1 q1;
13438 +               struct _vxm_rqinfo_2 q2;
13439 +       };
13440 +};
13441 +
13442 +
13443 +#endif /* _VX_MONITOR_H */
13444 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/network.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/network.h
13445 --- linux-3.0-rc5/include/linux/vserver/network.h       1970-01-01 01:00:00.000000000 +0100
13446 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/network.h  2011-06-10 22:11:24.000000000 +0200
13447 @@ -0,0 +1,146 @@
13448 +#ifndef _VX_NETWORK_H
13449 +#define _VX_NETWORK_H
13450 +
13451 +#include <linux/types.h>
13452 +
13453 +
13454 +#define MAX_N_CONTEXT  65535   /* Arbitrary limit */
13455 +
13456 +
13457 +/* network flags */
13458 +
13459 +#define NXF_INFO_PRIVATE       0x00000008
13460 +
13461 +#define NXF_SINGLE_IP          0x00000100
13462 +#define NXF_LBACK_REMAP                0x00000200
13463 +#define NXF_LBACK_ALLOW                0x00000400
13464 +
13465 +#define NXF_HIDE_NETIF         0x02000000
13466 +#define NXF_HIDE_LBACK         0x04000000
13467 +
13468 +#define NXF_STATE_SETUP                (1ULL << 32)
13469 +#define NXF_STATE_ADMIN                (1ULL << 34)
13470 +
13471 +#define NXF_SC_HELPER          (1ULL << 36)
13472 +#define NXF_PERSISTENT         (1ULL << 38)
13473 +
13474 +#define NXF_ONE_TIME           (0x0005ULL << 32)
13475 +
13476 +
13477 +#define        NXF_INIT_SET            (__nxf_init_set())
13478 +
13479 +static inline uint64_t __nxf_init_set(void) {
13480 +       return    NXF_STATE_ADMIN
13481 +#ifdef CONFIG_VSERVER_AUTO_LBACK
13482 +               | NXF_LBACK_REMAP
13483 +               | NXF_HIDE_LBACK
13484 +#endif
13485 +#ifdef CONFIG_VSERVER_AUTO_SINGLE
13486 +               | NXF_SINGLE_IP
13487 +#endif
13488 +               | NXF_HIDE_NETIF;
13489 +}
13490 +
13491 +
13492 +/* network caps */
13493 +
13494 +#define NXC_TUN_CREATE         0x00000001
13495 +
13496 +#define NXC_RAW_ICMP           0x00000100
13497 +
13498 +
13499 +/* address types */
13500 +
13501 +#define NXA_TYPE_IPV4          0x0001
13502 +#define NXA_TYPE_IPV6          0x0002
13503 +
13504 +#define NXA_TYPE_NONE          0x0000
13505 +#define NXA_TYPE_ANY           0x00FF
13506 +
13507 +#define NXA_TYPE_ADDR          0x0010
13508 +#define NXA_TYPE_MASK          0x0020
13509 +#define NXA_TYPE_RANGE         0x0040
13510 +
13511 +#define NXA_MASK_ALL           (NXA_TYPE_ADDR | NXA_TYPE_MASK | NXA_TYPE_RANGE)
13512 +
13513 +#define NXA_MOD_BCAST          0x0100
13514 +#define NXA_MOD_LBACK          0x0200
13515 +
13516 +#define NXA_LOOPBACK           0x1000
13517 +
13518 +#define NXA_MASK_BIND          (NXA_MASK_ALL | NXA_MOD_BCAST | NXA_MOD_LBACK)
13519 +#define NXA_MASK_SHOW          (NXA_MASK_ALL | NXA_LOOPBACK)
13520 +
13521 +#ifdef __KERNEL__
13522 +
13523 +#include <linux/list.h>
13524 +#include <linux/spinlock.h>
13525 +#include <linux/rcupdate.h>
13526 +#include <linux/in.h>
13527 +#include <linux/in6.h>
13528 +#include <asm/atomic.h>
13529 +
13530 +struct nx_addr_v4 {
13531 +       struct nx_addr_v4 *next;
13532 +       struct in_addr ip[2];
13533 +       struct in_addr mask;
13534 +       uint16_t type;
13535 +       uint16_t flags;
13536 +};
13537 +
13538 +struct nx_addr_v6 {
13539 +       struct nx_addr_v6 *next;
13540 +       struct in6_addr ip;
13541 +       struct in6_addr mask;
13542 +       uint32_t prefix;
13543 +       uint16_t type;
13544 +       uint16_t flags;
13545 +};
13546 +
13547 +struct nx_info {
13548 +       struct hlist_node nx_hlist;     /* linked list of nxinfos */
13549 +       nid_t nx_id;                    /* vnet id */
13550 +       atomic_t nx_usecnt;             /* usage count */
13551 +       atomic_t nx_tasks;              /* tasks count */
13552 +       int nx_state;                   /* context state */
13553 +
13554 +       uint64_t nx_flags;              /* network flag word */
13555 +       uint64_t nx_ncaps;              /* network capabilities */
13556 +
13557 +       struct in_addr v4_lback;        /* Loopback address */
13558 +       struct in_addr v4_bcast;        /* Broadcast address */
13559 +       struct nx_addr_v4 v4;           /* First/Single ipv4 address */
13560 +#ifdef CONFIG_IPV6
13561 +       struct nx_addr_v6 v6;           /* First/Single ipv6 address */
13562 +#endif
13563 +       char nx_name[65];               /* network context name */
13564 +};
13565 +
13566 +
13567 +/* status flags */
13568 +
13569 +#define NXS_HASHED      0x0001
13570 +#define NXS_SHUTDOWN    0x0100
13571 +#define NXS_RELEASED    0x8000
13572 +
13573 +extern struct nx_info *lookup_nx_info(int);
13574 +
13575 +extern int get_nid_list(int, unsigned int *, int);
13576 +extern int nid_is_hashed(nid_t);
13577 +
13578 +extern int nx_migrate_task(struct task_struct *, struct nx_info *);
13579 +
13580 +extern long vs_net_change(struct nx_info *, unsigned int);
13581 +
13582 +struct sock;
13583 +
13584 +
13585 +#define NX_IPV4(n)     ((n)->v4.type != NXA_TYPE_NONE)
13586 +#ifdef  CONFIG_IPV6
13587 +#define NX_IPV6(n)     ((n)->v6.type != NXA_TYPE_NONE)
13588 +#else
13589 +#define NX_IPV6(n)     (0)
13590 +#endif
13591 +
13592 +#endif /* __KERNEL__ */
13593 +#endif /* _VX_NETWORK_H */
13594 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/network_cmd.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/network_cmd.h
13595 --- linux-3.0-rc5/include/linux/vserver/network_cmd.h   1970-01-01 01:00:00.000000000 +0100
13596 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/network_cmd.h      2011-06-10 22:11:24.000000000 +0200
13597 @@ -0,0 +1,164 @@
13598 +#ifndef _VX_NETWORK_CMD_H
13599 +#define _VX_NETWORK_CMD_H
13600 +
13601 +
13602 +/* vinfo commands */
13603 +
13604 +#define VCMD_task_nid          VC_CMD(VINFO, 2, 0)
13605 +
13606 +#ifdef __KERNEL__
13607 +extern int vc_task_nid(uint32_t);
13608 +
13609 +#endif /* __KERNEL__ */
13610 +
13611 +#define VCMD_nx_info           VC_CMD(VINFO, 6, 0)
13612 +
13613 +struct vcmd_nx_info_v0 {
13614 +       uint32_t nid;
13615 +       /* more to come */
13616 +};
13617 +
13618 +#ifdef __KERNEL__
13619 +extern int vc_nx_info(struct nx_info *, void __user *);
13620 +
13621 +#endif /* __KERNEL__ */
13622 +
13623 +#include <linux/in.h>
13624 +#include <linux/in6.h>
13625 +
13626 +#define VCMD_net_create_v0     VC_CMD(VNET, 1, 0)
13627 +#define VCMD_net_create                VC_CMD(VNET, 1, 1)
13628 +
13629 +struct  vcmd_net_create {
13630 +       uint64_t flagword;
13631 +};
13632 +
13633 +#define VCMD_net_migrate       VC_CMD(NETMIG, 1, 0)
13634 +
13635 +#define VCMD_net_add           VC_CMD(NETALT, 1, 0)
13636 +#define VCMD_net_remove                VC_CMD(NETALT, 2, 0)
13637 +
13638 +struct vcmd_net_addr_v0 {
13639 +       uint16_t type;
13640 +       uint16_t count;
13641 +       struct in_addr ip[4];
13642 +       struct in_addr mask[4];
13643 +};
13644 +
13645 +#define VCMD_net_add_ipv4_v1   VC_CMD(NETALT, 1, 1)
13646 +#define VCMD_net_rem_ipv4_v1   VC_CMD(NETALT, 2, 1)
13647 +
13648 +struct vcmd_net_addr_ipv4_v1 {
13649 +       uint16_t type;
13650 +       uint16_t flags;
13651 +       struct in_addr ip;
13652 +       struct in_addr mask;
13653 +};
13654 +
13655 +#define VCMD_net_add_ipv4      VC_CMD(NETALT, 1, 2)
13656 +#define VCMD_net_rem_ipv4      VC_CMD(NETALT, 2, 2)
13657 +
13658 +struct vcmd_net_addr_ipv4_v2 {
13659 +       uint16_t type;
13660 +       uint16_t flags;
13661 +       struct in_addr ip;
13662 +       struct in_addr ip2;
13663 +       struct in_addr mask;
13664 +};
13665 +
13666 +#define VCMD_net_add_ipv6      VC_CMD(NETALT, 3, 1)
13667 +#define VCMD_net_remove_ipv6   VC_CMD(NETALT, 4, 1)
13668 +
13669 +struct vcmd_net_addr_ipv6_v1 {
13670 +       uint16_t type;
13671 +       uint16_t flags;
13672 +       uint32_t prefix;
13673 +       struct in6_addr ip;
13674 +       struct in6_addr mask;
13675 +};
13676 +
13677 +#define VCMD_add_match_ipv4    VC_CMD(NETALT, 5, 0)
13678 +#define VCMD_get_match_ipv4    VC_CMD(NETALT, 6, 0)
13679 +
13680 +struct vcmd_match_ipv4_v0 {
13681 +       uint16_t type;
13682 +       uint16_t flags;
13683 +       uint16_t parent;
13684 +       uint16_t prefix;
13685 +       struct in_addr ip;
13686 +       struct in_addr ip2;
13687 +       struct in_addr mask;
13688 +};
13689 +
13690 +#define VCMD_add_match_ipv6    VC_CMD(NETALT, 7, 0)
13691 +#define VCMD_get_match_ipv6    VC_CMD(NETALT, 8, 0)
13692 +
13693 +struct vcmd_match_ipv6_v0 {
13694 +       uint16_t type;
13695 +       uint16_t flags;
13696 +       uint16_t parent;
13697 +       uint16_t prefix;
13698 +       struct in6_addr ip;
13699 +       struct in6_addr ip2;
13700 +       struct in6_addr mask;
13701 +};
13702 +
13703 +
13704 +#ifdef __KERNEL__
13705 +extern int vc_net_create(uint32_t, void __user *);
13706 +extern int vc_net_migrate(struct nx_info *, void __user *);
13707 +
13708 +extern int vc_net_add(struct nx_info *, void __user *);
13709 +extern int vc_net_remove(struct nx_info *, void __user *);
13710 +
13711 +extern int vc_net_add_ipv4_v1(struct nx_info *, void __user *);
13712 +extern int vc_net_add_ipv4(struct nx_info *, void __user *);
13713 +
13714 +extern int vc_net_rem_ipv4_v1(struct nx_info *, void __user *);
13715 +extern int vc_net_rem_ipv4(struct nx_info *, void __user *);
13716 +
13717 +extern int vc_net_add_ipv6(struct nx_info *, void __user *);
13718 +extern int vc_net_remove_ipv6(struct nx_info *, void __user *);
13719 +
13720 +extern int vc_add_match_ipv4(struct nx_info *, void __user *);
13721 +extern int vc_get_match_ipv4(struct nx_info *, void __user *);
13722 +
13723 +extern int vc_add_match_ipv6(struct nx_info *, void __user *);
13724 +extern int vc_get_match_ipv6(struct nx_info *, void __user *);
13725 +
13726 +#endif /* __KERNEL__ */
13727 +
13728 +
13729 +/* flag commands */
13730 +
13731 +#define VCMD_get_nflags                VC_CMD(FLAGS, 5, 0)
13732 +#define VCMD_set_nflags                VC_CMD(FLAGS, 6, 0)
13733 +
13734 +struct vcmd_net_flags_v0 {
13735 +       uint64_t flagword;
13736 +       uint64_t mask;
13737 +};
13738 +
13739 +#ifdef __KERNEL__
13740 +extern int vc_get_nflags(struct nx_info *, void __user *);
13741 +extern int vc_set_nflags(struct nx_info *, void __user *);
13742 +
13743 +#endif /* __KERNEL__ */
13744 +
13745 +
13746 +/* network caps commands */
13747 +
13748 +#define VCMD_get_ncaps         VC_CMD(FLAGS, 7, 0)
13749 +#define VCMD_set_ncaps         VC_CMD(FLAGS, 8, 0)
13750 +
13751 +struct vcmd_net_caps_v0 {
13752 +       uint64_t ncaps;
13753 +       uint64_t cmask;
13754 +};
13755 +
13756 +#ifdef __KERNEL__
13757 +extern int vc_get_ncaps(struct nx_info *, void __user *);
13758 +extern int vc_set_ncaps(struct nx_info *, void __user *);
13759 +
13760 +#endif /* __KERNEL__ */
13761 +#endif /* _VX_CONTEXT_CMD_H */
13762 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/percpu.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/percpu.h
13763 --- linux-3.0-rc5/include/linux/vserver/percpu.h        1970-01-01 01:00:00.000000000 +0100
13764 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/percpu.h   2011-06-10 22:11:24.000000000 +0200
13765 @@ -0,0 +1,14 @@
13766 +#ifndef _VX_PERCPU_H
13767 +#define _VX_PERCPU_H
13768 +
13769 +#include "cvirt_def.h"
13770 +#include "sched_def.h"
13771 +
13772 +struct _vx_percpu {
13773 +       struct _vx_cvirt_pc cvirt;
13774 +       struct _vx_sched_pc sched;
13775 +};
13776 +
13777 +#define        PERCPU_PERCTX   (sizeof(struct _vx_percpu))
13778 +
13779 +#endif /* _VX_PERCPU_H */
13780 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/pid.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/pid.h
13781 --- linux-3.0-rc5/include/linux/vserver/pid.h   1970-01-01 01:00:00.000000000 +0100
13782 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/pid.h      2011-06-10 22:11:24.000000000 +0200
13783 @@ -0,0 +1,51 @@
13784 +#ifndef _VSERVER_PID_H
13785 +#define _VSERVER_PID_H
13786 +
13787 +/* pid faking stuff */
13788 +
13789 +#define vx_info_map_pid(v, p) \
13790 +       __vx_info_map_pid((v), (p), __func__, __FILE__, __LINE__)
13791 +#define vx_info_map_tgid(v,p)  vx_info_map_pid(v,p)
13792 +#define vx_map_pid(p) vx_info_map_pid(current_vx_info(), p)
13793 +#define vx_map_tgid(p) vx_map_pid(p)
13794 +
13795 +static inline int __vx_info_map_pid(struct vx_info *vxi, int pid,
13796 +       const char *func, const char *file, int line)
13797 +{
13798 +       if (vx_info_flags(vxi, VXF_INFO_INIT, 0)) {
13799 +               vxfprintk(VXD_CBIT(cvirt, 2),
13800 +                       "vx_map_tgid: %p/%llx: %d -> %d",
13801 +                       vxi, (long long)vxi->vx_flags, pid,
13802 +                       (pid && pid == vxi->vx_initpid) ? 1 : pid,
13803 +                       func, file, line);
13804 +               if (pid == 0)
13805 +                       return 0;
13806 +               if (pid == vxi->vx_initpid)
13807 +                       return 1;
13808 +       }
13809 +       return pid;
13810 +}
13811 +
13812 +#define vx_info_rmap_pid(v, p) \
13813 +       __vx_info_rmap_pid((v), (p), __func__, __FILE__, __LINE__)
13814 +#define vx_rmap_pid(p) vx_info_rmap_pid(current_vx_info(), p)
13815 +#define vx_rmap_tgid(p) vx_rmap_pid(p)
13816 +
13817 +static inline int __vx_info_rmap_pid(struct vx_info *vxi, int pid,
13818 +       const char *func, const char *file, int line)
13819 +{
13820 +       if (vx_info_flags(vxi, VXF_INFO_INIT, 0)) {
13821 +               vxfprintk(VXD_CBIT(cvirt, 2),
13822 +                       "vx_rmap_tgid: %p/%llx: %d -> %d",
13823 +                       vxi, (long long)vxi->vx_flags, pid,
13824 +                       (pid == 1) ? vxi->vx_initpid : pid,
13825 +                       func, file, line);
13826 +               if ((pid == 1) && vxi->vx_initpid)
13827 +                       return vxi->vx_initpid;
13828 +               if (pid == vxi->vx_initpid)
13829 +                       return ~0U;
13830 +       }
13831 +       return pid;
13832 +}
13833 +
13834 +#endif
13835 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/sched.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/sched.h
13836 --- linux-3.0-rc5/include/linux/vserver/sched.h 1970-01-01 01:00:00.000000000 +0100
13837 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/sched.h    2011-06-10 22:11:24.000000000 +0200
13838 @@ -0,0 +1,23 @@
13839 +#ifndef _VX_SCHED_H
13840 +#define _VX_SCHED_H
13841 +
13842 +
13843 +#ifdef __KERNEL__
13844 +
13845 +struct timespec;
13846 +
13847 +void vx_vsi_uptime(struct timespec *, struct timespec *);
13848 +
13849 +
13850 +struct vx_info;
13851 +
13852 +void vx_update_load(struct vx_info *);
13853 +
13854 +
13855 +void vx_update_sched_param(struct _vx_sched *sched,
13856 +       struct _vx_sched_pc *sched_pc);
13857 +
13858 +#endif /* __KERNEL__ */
13859 +#else  /* _VX_SCHED_H */
13860 +#warning duplicate inclusion
13861 +#endif /* _VX_SCHED_H */
13862 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/sched_cmd.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/sched_cmd.h
13863 --- linux-3.0-rc5/include/linux/vserver/sched_cmd.h     1970-01-01 01:00:00.000000000 +0100
13864 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/sched_cmd.h        2011-06-10 22:11:24.000000000 +0200
13865 @@ -0,0 +1,21 @@
13866 +#ifndef _VX_SCHED_CMD_H
13867 +#define _VX_SCHED_CMD_H
13868 +
13869 +
13870 +struct vcmd_prio_bias {
13871 +       int32_t cpu_id;
13872 +       int32_t prio_bias;
13873 +};
13874 +
13875 +#define VCMD_set_prio_bias     VC_CMD(SCHED, 4, 0)
13876 +#define VCMD_get_prio_bias     VC_CMD(SCHED, 5, 0)
13877 +
13878 +#ifdef __KERNEL__
13879 +
13880 +#include <linux/compiler.h>
13881 +
13882 +extern int vc_set_prio_bias(struct vx_info *, void __user *);
13883 +extern int vc_get_prio_bias(struct vx_info *, void __user *);
13884 +
13885 +#endif /* __KERNEL__ */
13886 +#endif /* _VX_SCHED_CMD_H */
13887 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/sched_def.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/sched_def.h
13888 --- linux-3.0-rc5/include/linux/vserver/sched_def.h     1970-01-01 01:00:00.000000000 +0100
13889 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/sched_def.h        2011-06-10 22:11:24.000000000 +0200
13890 @@ -0,0 +1,38 @@
13891 +#ifndef _VX_SCHED_DEF_H
13892 +#define _VX_SCHED_DEF_H
13893 +
13894 +#include <linux/spinlock.h>
13895 +#include <linux/jiffies.h>
13896 +#include <linux/cpumask.h>
13897 +#include <asm/atomic.h>
13898 +#include <asm/param.h>
13899 +
13900 +
13901 +/* context sub struct */
13902 +
13903 +struct _vx_sched {
13904 +       int prio_bias;                  /* bias offset for priority */
13905 +
13906 +       cpumask_t update;               /* CPUs which should update */
13907 +};
13908 +
13909 +struct _vx_sched_pc {
13910 +       int prio_bias;                  /* bias offset for priority */
13911 +
13912 +       uint64_t user_ticks;            /* token tick events */
13913 +       uint64_t sys_ticks;             /* token tick events */
13914 +       uint64_t hold_ticks;            /* token ticks paused */
13915 +};
13916 +
13917 +
13918 +#ifdef CONFIG_VSERVER_DEBUG
13919 +
13920 +static inline void __dump_vx_sched(struct _vx_sched *sched)
13921 +{
13922 +       printk("\t_vx_sched:\n");
13923 +       printk("\t priority = %4d\n", sched->prio_bias);
13924 +}
13925 +
13926 +#endif
13927 +
13928 +#endif /* _VX_SCHED_DEF_H */
13929 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/signal.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/signal.h
13930 --- linux-3.0-rc5/include/linux/vserver/signal.h        1970-01-01 01:00:00.000000000 +0100
13931 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/signal.h   2011-06-10 22:11:24.000000000 +0200
13932 @@ -0,0 +1,14 @@
13933 +#ifndef _VX_SIGNAL_H
13934 +#define _VX_SIGNAL_H
13935 +
13936 +
13937 +#ifdef __KERNEL__
13938 +
13939 +struct vx_info;
13940 +
13941 +int vx_info_kill(struct vx_info *, int, int);
13942 +
13943 +#endif /* __KERNEL__ */
13944 +#else  /* _VX_SIGNAL_H */
13945 +#warning duplicate inclusion
13946 +#endif /* _VX_SIGNAL_H */
13947 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/signal_cmd.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/signal_cmd.h
13948 --- linux-3.0-rc5/include/linux/vserver/signal_cmd.h    1970-01-01 01:00:00.000000000 +0100
13949 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/signal_cmd.h       2011-06-10 22:11:24.000000000 +0200
13950 @@ -0,0 +1,43 @@
13951 +#ifndef _VX_SIGNAL_CMD_H
13952 +#define _VX_SIGNAL_CMD_H
13953 +
13954 +
13955 +/*  signalling vserver commands */
13956 +
13957 +#define VCMD_ctx_kill          VC_CMD(PROCTRL, 1, 0)
13958 +#define VCMD_wait_exit         VC_CMD(EVENT, 99, 0)
13959 +
13960 +struct vcmd_ctx_kill_v0 {
13961 +       int32_t pid;
13962 +       int32_t sig;
13963 +};
13964 +
13965 +struct vcmd_wait_exit_v0 {
13966 +       int32_t reboot_cmd;
13967 +       int32_t exit_code;
13968 +};
13969 +
13970 +#ifdef __KERNEL__
13971 +
13972 +extern int vc_ctx_kill(struct vx_info *, void __user *);
13973 +extern int vc_wait_exit(struct vx_info *, void __user *);
13974 +
13975 +#endif /* __KERNEL__ */
13976 +
13977 +/*  process alteration commands */
13978 +
13979 +#define VCMD_get_pflags                VC_CMD(PROCALT, 5, 0)
13980 +#define VCMD_set_pflags                VC_CMD(PROCALT, 6, 0)
13981 +
13982 +struct vcmd_pflags_v0 {
13983 +       uint32_t flagword;
13984 +       uint32_t mask;
13985 +};
13986 +
13987 +#ifdef __KERNEL__
13988 +
13989 +extern int vc_get_pflags(uint32_t pid, void __user *);
13990 +extern int vc_set_pflags(uint32_t pid, void __user *);
13991 +
13992 +#endif /* __KERNEL__ */
13993 +#endif /* _VX_SIGNAL_CMD_H */
13994 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/space.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/space.h
13995 --- linux-3.0-rc5/include/linux/vserver/space.h 1970-01-01 01:00:00.000000000 +0100
13996 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/space.h    2011-06-10 22:11:24.000000000 +0200
13997 @@ -0,0 +1,12 @@
13998 +#ifndef _VX_SPACE_H
13999 +#define _VX_SPACE_H
14000 +
14001 +#include <linux/types.h>
14002 +
14003 +struct vx_info;
14004 +
14005 +int vx_set_space(struct vx_info *vxi, unsigned long mask, unsigned index);
14006 +
14007 +#else  /* _VX_SPACE_H */
14008 +#warning duplicate inclusion
14009 +#endif /* _VX_SPACE_H */
14010 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/space_cmd.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/space_cmd.h
14011 --- linux-3.0-rc5/include/linux/vserver/space_cmd.h     1970-01-01 01:00:00.000000000 +0100
14012 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/space_cmd.h        2011-06-10 22:11:24.000000000 +0200
14013 @@ -0,0 +1,38 @@
14014 +#ifndef _VX_SPACE_CMD_H
14015 +#define _VX_SPACE_CMD_H
14016 +
14017 +
14018 +#define VCMD_enter_space_v0    VC_CMD(PROCALT, 1, 0)
14019 +#define VCMD_enter_space_v1    VC_CMD(PROCALT, 1, 1)
14020 +#define VCMD_enter_space       VC_CMD(PROCALT, 1, 2)
14021 +
14022 +#define VCMD_set_space_v0      VC_CMD(PROCALT, 3, 0)
14023 +#define VCMD_set_space_v1      VC_CMD(PROCALT, 3, 1)
14024 +#define VCMD_set_space         VC_CMD(PROCALT, 3, 2)
14025 +
14026 +#define VCMD_get_space_mask_v0 VC_CMD(PROCALT, 4, 0)
14027 +
14028 +#define VCMD_get_space_mask    VC_CMD(VSPACE, 0, 1)
14029 +#define VCMD_get_space_default VC_CMD(VSPACE, 1, 0)
14030 +
14031 +
14032 +struct vcmd_space_mask_v1 {
14033 +       uint64_t mask;
14034 +};
14035 +
14036 +struct vcmd_space_mask_v2 {
14037 +       uint64_t mask;
14038 +       uint32_t index;
14039 +};
14040 +
14041 +
14042 +#ifdef __KERNEL__
14043 +
14044 +extern int vc_enter_space_v1(struct vx_info *, void __user *);
14045 +extern int vc_set_space_v1(struct vx_info *, void __user *);
14046 +extern int vc_enter_space(struct vx_info *, void __user *);
14047 +extern int vc_set_space(struct vx_info *, void __user *);
14048 +extern int vc_get_space_mask(void __user *, int);
14049 +
14050 +#endif /* __KERNEL__ */
14051 +#endif /* _VX_SPACE_CMD_H */
14052 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/switch.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/switch.h
14053 --- linux-3.0-rc5/include/linux/vserver/switch.h        1970-01-01 01:00:00.000000000 +0100
14054 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/switch.h   2011-06-10 22:11:24.000000000 +0200
14055 @@ -0,0 +1,98 @@
14056 +#ifndef _VX_SWITCH_H
14057 +#define _VX_SWITCH_H
14058 +
14059 +#include <linux/types.h>
14060 +
14061 +
14062 +#define VC_CATEGORY(c)         (((c) >> 24) & 0x3F)
14063 +#define VC_COMMAND(c)          (((c) >> 16) & 0xFF)
14064 +#define VC_VERSION(c)          ((c) & 0xFFF)
14065 +
14066 +#define VC_CMD(c, i, v)                ((((VC_CAT_ ## c) & 0x3F) << 24) \
14067 +                               | (((i) & 0xFF) << 16) | ((v) & 0xFFF))
14068 +
14069 +/*
14070 +
14071 +  Syscall Matrix V2.8
14072 +
14073 +        |VERSION|CREATE |MODIFY |MIGRATE|CONTROL|EXPERIM| |SPECIAL|SPECIAL|
14074 +        |STATS  |DESTROY|ALTER  |CHANGE |LIMIT  |TEST   | |       |       |
14075 +        |INFO   |SETUP  |       |MOVE   |       |       | |       |       |
14076 +  -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
14077 +  SYSTEM |VERSION|VSETUP |VHOST  |       |       |       | |DEVICE |       |
14078 +  HOST   |     00|     01|     02|     03|     04|     05| |     06|     07|
14079 +  -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
14080 +  CPU    |       |VPROC  |PROCALT|PROCMIG|PROCTRL|       | |SCHED. |       |
14081 +  PROCESS|     08|     09|     10|     11|     12|     13| |     14|     15|
14082 +  -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
14083 +  MEMORY |       |       |       |       |MEMCTRL|       | |SWAP   |       |
14084 +        |     16|     17|     18|     19|     20|     21| |     22|     23|
14085 +  -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
14086 +  NETWORK|       |VNET   |NETALT |NETMIG |NETCTL |       | |SERIAL |       |
14087 +        |     24|     25|     26|     27|     28|     29| |     30|     31|
14088 +  -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
14089 +  DISK   |       |       |       |TAGMIG |DLIMIT |       | |INODE  |       |
14090 +  VFS    |     32|     33|     34|     35|     36|     37| |     38|     39|
14091 +  -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
14092 +  OTHER  |VSTAT  |       |       |       |       |       | |VINFO  |       |
14093 +        |     40|     41|     42|     43|     44|     45| |     46|     47|
14094 +  =======+=======+=======+=======+=======+=======+=======+ +=======+=======+
14095 +  SPECIAL|EVENT  |       |       |       |FLAGS  |       | |VSPACE |       |
14096 +        |     48|     49|     50|     51|     52|     53| |     54|     55|
14097 +  -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
14098 +  SPECIAL|DEBUG  |       |       |       |RLIMIT |SYSCALL| |       |COMPAT |
14099 +        |     56|     57|     58|     59|     60|TEST 61| |     62|     63|
14100 +  -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
14101 +
14102 +*/
14103 +
14104 +#define VC_CAT_VERSION         0
14105 +
14106 +#define VC_CAT_VSETUP          1
14107 +#define VC_CAT_VHOST           2
14108 +
14109 +#define VC_CAT_DEVICE          6
14110 +
14111 +#define VC_CAT_VPROC           9
14112 +#define VC_CAT_PROCALT         10
14113 +#define VC_CAT_PROCMIG         11
14114 +#define VC_CAT_PROCTRL         12
14115 +
14116 +#define VC_CAT_SCHED           14
14117 +#define VC_CAT_MEMCTRL         20
14118 +
14119 +#define VC_CAT_VNET            25
14120 +#define VC_CAT_NETALT          26
14121 +#define VC_CAT_NETMIG          27
14122 +#define VC_CAT_NETCTRL         28
14123 +
14124 +#define VC_CAT_TAGMIG          35
14125 +#define VC_CAT_DLIMIT          36
14126 +#define VC_CAT_INODE           38
14127 +
14128 +#define VC_CAT_VSTAT           40
14129 +#define VC_CAT_VINFO           46
14130 +#define VC_CAT_EVENT           48
14131 +
14132 +#define VC_CAT_FLAGS           52
14133 +#define VC_CAT_VSPACE          54
14134 +#define VC_CAT_DEBUG           56
14135 +#define VC_CAT_RLIMIT          60
14136 +
14137 +#define VC_CAT_SYSTEST         61
14138 +#define VC_CAT_COMPAT          63
14139 +
14140 +/*  query version */
14141 +
14142 +#define VCMD_get_version       VC_CMD(VERSION, 0, 0)
14143 +#define VCMD_get_vci           VC_CMD(VERSION, 1, 0)
14144 +
14145 +
14146 +#ifdef __KERNEL__
14147 +
14148 +#include <linux/errno.h>
14149 +
14150 +#endif /* __KERNEL__ */
14151 +
14152 +#endif /* _VX_SWITCH_H */
14153 +
14154 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/tag.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/tag.h
14155 --- linux-3.0-rc5/include/linux/vserver/tag.h   1970-01-01 01:00:00.000000000 +0100
14156 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/tag.h      2011-06-10 22:11:24.000000000 +0200
14157 @@ -0,0 +1,143 @@
14158 +#ifndef _DX_TAG_H
14159 +#define _DX_TAG_H
14160 +
14161 +#include <linux/types.h>
14162 +
14163 +
14164 +#define DX_TAG(in)     (IS_TAGGED(in))
14165 +
14166 +
14167 +#ifdef CONFIG_TAG_NFSD
14168 +#define DX_TAG_NFSD    1
14169 +#else
14170 +#define DX_TAG_NFSD    0
14171 +#endif
14172 +
14173 +
14174 +#ifdef CONFIG_TAGGING_NONE
14175 +
14176 +#define MAX_UID                0xFFFFFFFF
14177 +#define MAX_GID                0xFFFFFFFF
14178 +
14179 +#define INOTAG_TAG(cond, uid, gid, tag)        (0)
14180 +
14181 +#define TAGINO_UID(cond, uid, tag)     (uid)
14182 +#define TAGINO_GID(cond, gid, tag)     (gid)
14183 +
14184 +#endif
14185 +
14186 +
14187 +#ifdef CONFIG_TAGGING_GID16
14188 +
14189 +#define MAX_UID                0xFFFFFFFF
14190 +#define MAX_GID                0x0000FFFF
14191 +
14192 +#define INOTAG_TAG(cond, uid, gid, tag)        \
14193 +       ((cond) ? (((gid) >> 16) & 0xFFFF) : 0)
14194 +
14195 +#define TAGINO_UID(cond, uid, tag)     (uid)
14196 +#define TAGINO_GID(cond, gid, tag)     \
14197 +       ((cond) ? (((gid) & 0xFFFF) | ((tag) << 16)) : (gid))
14198 +
14199 +#endif
14200 +
14201 +
14202 +#ifdef CONFIG_TAGGING_ID24
14203 +
14204 +#define MAX_UID                0x00FFFFFF
14205 +#define MAX_GID                0x00FFFFFF
14206 +
14207 +#define INOTAG_TAG(cond, uid, gid, tag)        \
14208 +       ((cond) ? ((((uid) >> 16) & 0xFF00) | (((gid) >> 24) & 0xFF)) : 0)
14209 +
14210 +#define TAGINO_UID(cond, uid, tag)     \
14211 +       ((cond) ? (((uid) & 0xFFFFFF) | (((tag) & 0xFF00) << 16)) : (uid))
14212 +#define TAGINO_GID(cond, gid, tag)     \
14213 +       ((cond) ? (((gid) & 0xFFFFFF) | (((tag) & 0x00FF) << 24)) : (gid))
14214 +
14215 +#endif
14216 +
14217 +
14218 +#ifdef CONFIG_TAGGING_UID16
14219 +
14220 +#define MAX_UID                0x0000FFFF
14221 +#define MAX_GID                0xFFFFFFFF
14222 +
14223 +#define INOTAG_TAG(cond, uid, gid, tag)        \
14224 +       ((cond) ? (((uid) >> 16) & 0xFFFF) : 0)
14225 +
14226 +#define TAGINO_UID(cond, uid, tag)     \
14227 +       ((cond) ? (((uid) & 0xFFFF) | ((tag) << 16)) : (uid))
14228 +#define TAGINO_GID(cond, gid, tag)     (gid)
14229 +
14230 +#endif
14231 +
14232 +
14233 +#ifdef CONFIG_TAGGING_INTERN
14234 +
14235 +#define MAX_UID                0xFFFFFFFF
14236 +#define MAX_GID                0xFFFFFFFF
14237 +
14238 +#define INOTAG_TAG(cond, uid, gid, tag)        \
14239 +       ((cond) ? (tag) : 0)
14240 +
14241 +#define TAGINO_UID(cond, uid, tag)     (uid)
14242 +#define TAGINO_GID(cond, gid, tag)     (gid)
14243 +
14244 +#endif
14245 +
14246 +
14247 +#ifndef CONFIG_TAGGING_NONE
14248 +#define dx_current_fstag(sb)   \
14249 +       ((sb)->s_flags & MS_TAGGED ? dx_current_tag() : 0)
14250 +#else
14251 +#define dx_current_fstag(sb)   (0)
14252 +#endif
14253 +
14254 +#ifndef CONFIG_TAGGING_INTERN
14255 +#define TAGINO_TAG(cond, tag)  (0)
14256 +#else
14257 +#define TAGINO_TAG(cond, tag)  ((cond) ? (tag) : 0)
14258 +#endif
14259 +
14260 +#define INOTAG_UID(cond, uid, gid)     \
14261 +       ((cond) ? ((uid) & MAX_UID) : (uid))
14262 +#define INOTAG_GID(cond, uid, gid)     \
14263 +       ((cond) ? ((gid) & MAX_GID) : (gid))
14264 +
14265 +
14266 +static inline uid_t dx_map_uid(uid_t uid)
14267 +{
14268 +       if ((uid > MAX_UID) && (uid != -1))
14269 +               uid = -2;
14270 +       return (uid & MAX_UID);
14271 +}
14272 +
14273 +static inline gid_t dx_map_gid(gid_t gid)
14274 +{
14275 +       if ((gid > MAX_GID) && (gid != -1))
14276 +               gid = -2;
14277 +       return (gid & MAX_GID);
14278 +}
14279 +
14280 +struct peer_tag {
14281 +       int32_t xid;
14282 +       int32_t nid;
14283 +};
14284 +
14285 +#define dx_notagcheck(sb) ((sb) && ((sb)->s_flags & MS_NOTAGCHECK))
14286 +
14287 +int dx_parse_tag(char *string, tag_t *tag, int remove, int *mnt_flags,
14288 +                unsigned long *flags);
14289 +
14290 +#ifdef CONFIG_PROPAGATE
14291 +
14292 +void __dx_propagate_tag(struct nameidata *nd, struct inode *inode);
14293 +
14294 +#define dx_propagate_tag(n, i) __dx_propagate_tag(n, i)
14295 +
14296 +#else
14297 +#define dx_propagate_tag(n, i) do { } while (0)
14298 +#endif
14299 +
14300 +#endif /* _DX_TAG_H */
14301 diff -NurpP --minimal linux-3.0-rc5/include/linux/vserver/tag_cmd.h linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/tag_cmd.h
14302 --- linux-3.0-rc5/include/linux/vserver/tag_cmd.h       1970-01-01 01:00:00.000000000 +0100
14303 +++ linux-3.0-rc5-vs2.3.1-pre3/include/linux/vserver/tag_cmd.h  2011-06-10 22:11:24.000000000 +0200
14304 @@ -0,0 +1,22 @@
14305 +#ifndef _VX_TAG_CMD_H
14306 +#define _VX_TAG_CMD_H
14307 +
14308 +
14309 +/* vinfo commands */
14310 +
14311 +#define VCMD_task_tag          VC_CMD(VINFO, 3, 0)
14312 +
14313 +#ifdef __KERNEL__
14314 +extern int vc_task_tag(uint32_t);
14315 +
14316 +#endif /* __KERNEL__ */
14317 +
14318 +/* context commands */
14319 +
14320 +#define VCMD_tag_migrate       VC_CMD(TAGMIG, 1, 0)
14321 +
14322 +#ifdef __KERNEL__
14323 +extern int vc_tag_migrate(uint32_t);
14324 +
14325 +#endif /* __KERNEL__ */
14326 +#endif /* _VX_TAG_CMD_H */
14327 diff -NurpP --minimal linux-3.0-rc5/include/net/addrconf.h linux-3.0-rc5-vs2.3.1-pre3/include/net/addrconf.h
14328 --- linux-3.0-rc5/include/net/addrconf.h        2011-07-01 11:14:22.000000000 +0200
14329 +++ linux-3.0-rc5-vs2.3.1-pre3/include/net/addrconf.h   2011-06-10 22:11:24.000000000 +0200
14330 @@ -80,7 +80,8 @@ extern int                    ipv6_dev_get_saddr(struct n
14331                                                struct net_device *dev,
14332                                                const struct in6_addr *daddr,
14333                                                unsigned int srcprefs,
14334 -                                              struct in6_addr *saddr);
14335 +                                              struct in6_addr *saddr,
14336 +                                              struct nx_info *nxi);
14337  extern int                     ipv6_get_lladdr(struct net_device *dev,
14338                                                 struct in6_addr *addr,
14339                                                 unsigned char banned_flags);
14340 diff -NurpP --minimal linux-3.0-rc5/include/net/af_unix.h linux-3.0-rc5-vs2.3.1-pre3/include/net/af_unix.h
14341 --- linux-3.0-rc5/include/net/af_unix.h 2011-07-01 11:14:22.000000000 +0200
14342 +++ linux-3.0-rc5-vs2.3.1-pre3/include/net/af_unix.h    2011-06-10 22:11:24.000000000 +0200
14343 @@ -4,6 +4,7 @@
14344  #include <linux/socket.h>
14345  #include <linux/un.h>
14346  #include <linux/mutex.h>
14347 +#include <linux/vs_base.h>
14348  #include <net/sock.h>
14349
14350  extern void unix_inflight(struct file *fp);
14351 diff -NurpP --minimal linux-3.0-rc5/include/net/inet_timewait_sock.h linux-3.0-rc5-vs2.3.1-pre3/include/net/inet_timewait_sock.h
14352 --- linux-3.0-rc5/include/net/inet_timewait_sock.h      2011-03-15 18:07:40.000000000 +0100
14353 +++ linux-3.0-rc5-vs2.3.1-pre3/include/net/inet_timewait_sock.h 2011-06-10 22:11:24.000000000 +0200
14354 @@ -113,6 +113,10 @@ struct inet_timewait_sock {
14355  #define tw_net                 __tw_common.skc_net
14356  #define tw_daddr               __tw_common.skc_daddr
14357  #define tw_rcv_saddr           __tw_common.skc_rcv_saddr
14358 +#define tw_xid                 __tw_common.skc_xid
14359 +#define tw_vx_info             __tw_common.skc_vx_info
14360 +#define tw_nid                 __tw_common.skc_nid
14361 +#define tw_nx_info             __tw_common.skc_nx_info
14362         int                     tw_timeout;
14363         volatile unsigned char  tw_substate;
14364         unsigned char           tw_rcv_wscale;
14365 diff -NurpP --minimal linux-3.0-rc5/include/net/ip6_route.h linux-3.0-rc5-vs2.3.1-pre3/include/net/ip6_route.h
14366 --- linux-3.0-rc5/include/net/ip6_route.h       2011-07-01 11:14:22.000000000 +0200
14367 +++ linux-3.0-rc5-vs2.3.1-pre3/include/net/ip6_route.h  2011-06-16 14:16:51.000000000 +0200
14368 @@ -86,7 +86,8 @@ extern int                    ip6_route_get_saddr(struct
14369                                                     struct rt6_info *rt,
14370                                                     const struct in6_addr *daddr,
14371                                                     unsigned int prefs,
14372 -                                                   struct in6_addr *saddr);
14373 +                                                   struct in6_addr *saddr,
14374 +                                                   struct nx_info *nxi);
14375
14376  extern struct rt6_info         *rt6_lookup(struct net *net,
14377                                             const struct in6_addr *daddr,
14378 diff -NurpP --minimal linux-3.0-rc5/include/net/route.h linux-3.0-rc5-vs2.3.1-pre3/include/net/route.h
14379 --- linux-3.0-rc5/include/net/route.h   2011-07-01 11:14:22.000000000 +0200
14380 +++ linux-3.0-rc5-vs2.3.1-pre3/include/net/route.h      2011-06-13 18:20:44.000000000 +0200
14381 @@ -202,6 +202,9 @@ static inline void ip_rt_put(struct rtab
14382                 dst_release(&rt->dst);
14383  }
14384
14385 +#include <linux/vs_base.h>
14386 +#include <linux/vs_inet.h>
14387 +
14388  #define IPTOS_RT_MASK  (IPTOS_TOS_MASK & ~3)
14389
14390  extern const __u8 ip_tos2prio[16];
14391 @@ -253,6 +256,9 @@ static inline void ip_route_connect_init
14392                            protocol, flow_flags, dst, src, dport, sport);
14393  }
14394
14395 +extern struct rtable *ip_v4_find_src(struct net *net, struct nx_info *,
14396 +       struct flowi4 *);
14397 +
14398  static inline struct rtable *ip_route_connect(struct flowi4 *fl4,
14399                                               __be32 dst, __be32 src, u32 tos,
14400                                               int oif, u8 protocol,
14401 @@ -261,11 +267,24 @@ static inline struct rtable *ip_route_co
14402  {
14403         struct net *net = sock_net(sk);
14404         struct rtable *rt;
14405 +       struct nx_info *nx_info = current_nx_info();
14406
14407         ip_route_connect_init(fl4, dst, src, tos, oif, protocol,
14408                               sport, dport, sk, can_sleep);
14409
14410 -       if (!dst || !src) {
14411 +       if (sk)
14412 +               nx_info = sk->sk_nx_info;
14413 +
14414 +       vxdprintk(VXD_CBIT(net, 4),
14415 +               "ip_route_connect(%p) %p,%p;%lx",
14416 +               sk, nx_info, sk->sk_socket,
14417 +               (sk->sk_socket?sk->sk_socket->flags:0));
14418 +
14419 +       rt = ip_v4_find_src(net, nx_info, fl4);
14420 +       if (IS_ERR(rt))
14421 +               return rt;
14422 +
14423 +       if (!fl4->daddr || !fl4->saddr) {
14424                 rt = __ip_route_output_key(net, fl4);
14425                 if (IS_ERR(rt))
14426                         return rt;
14427 diff -NurpP --minimal linux-3.0-rc5/include/net/sock.h linux-3.0-rc5-vs2.3.1-pre3/include/net/sock.h
14428 --- linux-3.0-rc5/include/net/sock.h    2011-07-01 11:14:22.000000000 +0200
14429 +++ linux-3.0-rc5-vs2.3.1-pre3/include/net/sock.h       2011-07-01 11:35:35.000000000 +0200
14430 @@ -149,6 +149,10 @@ struct sock_common {
14431  #ifdef CONFIG_NET_NS
14432         struct net              *skc_net;
14433  #endif
14434 +       xid_t                   skc_xid;
14435 +       struct vx_info          *skc_vx_info;
14436 +       nid_t                   skc_nid;
14437 +       struct nx_info          *skc_nx_info;
14438         /*
14439          * fields between dontcopy_begin/dontcopy_end
14440          * are not copied in sock_copy()
14441 @@ -256,6 +260,10 @@ struct sock {
14442  #define sk_bind_node           __sk_common.skc_bind_node
14443  #define sk_prot                        __sk_common.skc_prot
14444  #define sk_net                 __sk_common.skc_net
14445 +#define sk_xid                 __sk_common.skc_xid
14446 +#define sk_vx_info             __sk_common.skc_vx_info
14447 +#define sk_nid                 __sk_common.skc_nid
14448 +#define sk_nx_info             __sk_common.skc_nx_info
14449         socket_lock_t           sk_lock;
14450         struct sk_buff_head     sk_receive_queue;
14451         /*
14452 diff -NurpP --minimal linux-3.0-rc5/init/Kconfig linux-3.0-rc5-vs2.3.1-pre3/init/Kconfig
14453 --- linux-3.0-rc5/init/Kconfig  2011-07-01 11:14:22.000000000 +0200
14454 +++ linux-3.0-rc5-vs2.3.1-pre3/init/Kconfig     2011-06-22 12:39:15.000000000 +0200
14455 @@ -574,6 +574,7 @@ config HAVE_UNSTABLE_SCHED_CLOCK
14456  menuconfig CGROUPS
14457         boolean "Control Group support"
14458         depends on EVENTFD
14459 +       default y
14460         help
14461           This option adds support for grouping sets of processes together, for
14462           use with process control subsystems such as Cpusets, CFS, memory
14463 diff -NurpP --minimal linux-3.0-rc5/init/main.c linux-3.0-rc5-vs2.3.1-pre3/init/main.c
14464 --- linux-3.0-rc5/init/main.c   2011-07-01 11:14:22.000000000 +0200
14465 +++ linux-3.0-rc5-vs2.3.1-pre3/init/main.c      2011-06-22 12:39:15.000000000 +0200
14466 @@ -68,6 +68,7 @@
14467  #include <linux/shmem_fs.h>
14468  #include <linux/slab.h>
14469  #include <linux/perf_event.h>
14470 +#include <linux/vserver/percpu.h>
14471
14472  #include <asm/io.h>
14473  #include <asm/bugs.h>
14474 diff -NurpP --minimal linux-3.0-rc5/ipc/mqueue.c linux-3.0-rc5-vs2.3.1-pre3/ipc/mqueue.c
14475 --- linux-3.0-rc5/ipc/mqueue.c  2011-03-15 18:07:41.000000000 +0100
14476 +++ linux-3.0-rc5-vs2.3.1-pre3/ipc/mqueue.c     2011-06-10 22:11:24.000000000 +0200
14477 @@ -33,6 +33,8 @@
14478  #include <linux/pid.h>
14479  #include <linux/ipc_namespace.h>
14480  #include <linux/slab.h>
14481 +#include <linux/vs_context.h>
14482 +#include <linux/vs_limit.h>
14483
14484  #include <net/sock.h>
14485  #include "util.h"
14486 @@ -66,6 +68,7 @@ struct mqueue_inode_info {
14487         struct sigevent notify;
14488         struct pid* notify_owner;
14489         struct user_struct *user;       /* user who created, for accounting */
14490 +       struct vx_info *vxi;
14491         struct sock *notify_sock;
14492         struct sk_buff *notify_cookie;
14493
14494 @@ -126,6 +129,7 @@ static struct inode *mqueue_get_inode(st
14495                 if (S_ISREG(mode)) {
14496                         struct mqueue_inode_info *info;
14497                         struct task_struct *p = current;
14498 +                       struct vx_info *vxi = p->vx_info;
14499                         unsigned long mq_bytes, mq_msg_tblsz;
14500
14501                         inode->i_fop = &mqueue_file_operations;
14502 @@ -139,6 +143,7 @@ static struct inode *mqueue_get_inode(st
14503                         info->notify_owner = NULL;
14504                         info->qsize = 0;
14505                         info->user = NULL;      /* set when all is ok */
14506 +                       info->vxi = NULL;
14507                         memset(&info->attr, 0, sizeof(info->attr));
14508                         info->attr.mq_maxmsg = ipc_ns->mq_msg_max;
14509                         info->attr.mq_msgsize = ipc_ns->mq_msgsize_max;
14510 @@ -157,16 +162,19 @@ static struct inode *mqueue_get_inode(st
14511                         spin_lock(&mq_lock);
14512                         if (u->mq_bytes + mq_bytes < u->mq_bytes ||
14513                             u->mq_bytes + mq_bytes >
14514 -                           task_rlimit(p, RLIMIT_MSGQUEUE)) {
14515 +                           task_rlimit(p, RLIMIT_MSGQUEUE) ||
14516 +                           !vx_ipcmsg_avail(vxi, mq_bytes)) {
14517                                 spin_unlock(&mq_lock);
14518                                 /* mqueue_evict_inode() releases info->messages */
14519                                 goto out_inode;
14520                         }
14521                         u->mq_bytes += mq_bytes;
14522 +                       vx_ipcmsg_add(vxi, u, mq_bytes);
14523                         spin_unlock(&mq_lock);
14524
14525                         /* all is ok */
14526                         info->user = get_uid(u);
14527 +                       info->vxi = get_vx_info(vxi);
14528                 } else if (S_ISDIR(mode)) {
14529                         inc_nlink(inode);
14530                         /* Some things misbehave if size == 0 on a directory */
14531 @@ -275,8 +283,11 @@ static void mqueue_evict_inode(struct in
14532             + info->attr.mq_msgsize);
14533         user = info->user;
14534         if (user) {
14535 +               struct vx_info *vxi = info->vxi;
14536 +
14537                 spin_lock(&mq_lock);
14538                 user->mq_bytes -= mq_bytes;
14539 +               vx_ipcmsg_sub(vxi, user, mq_bytes);
14540                 /*
14541                  * get_ns_from_inode() ensures that the
14542                  * (ipc_ns = sb->s_fs_info) is either a valid ipc_ns
14543 @@ -286,6 +297,7 @@ static void mqueue_evict_inode(struct in
14544                 if (ipc_ns)
14545                         ipc_ns->mq_queues_count--;
14546                 spin_unlock(&mq_lock);
14547 +               put_vx_info(vxi);
14548                 free_uid(user);
14549         }
14550         if (ipc_ns)
14551 diff -NurpP --minimal linux-3.0-rc5/ipc/msg.c linux-3.0-rc5-vs2.3.1-pre3/ipc/msg.c
14552 --- linux-3.0-rc5/ipc/msg.c     2011-05-22 16:17:59.000000000 +0200
14553 +++ linux-3.0-rc5-vs2.3.1-pre3/ipc/msg.c        2011-06-10 22:11:24.000000000 +0200
14554 @@ -37,6 +37,7 @@
14555  #include <linux/rwsem.h>
14556  #include <linux/nsproxy.h>
14557  #include <linux/ipc_namespace.h>
14558 +#include <linux/vs_base.h>
14559
14560  #include <asm/current.h>
14561  #include <asm/uaccess.h>
14562 @@ -190,6 +191,7 @@ static int newque(struct ipc_namespace *
14563
14564         msq->q_perm.mode = msgflg & S_IRWXUGO;
14565         msq->q_perm.key = key;
14566 +       msq->q_perm.xid = vx_current_xid();
14567
14568         msq->q_perm.security = NULL;
14569         retval = security_msg_queue_alloc(msq);
14570 diff -NurpP --minimal linux-3.0-rc5/ipc/namespace.c linux-3.0-rc5-vs2.3.1-pre3/ipc/namespace.c
14571 --- linux-3.0-rc5/ipc/namespace.c       2011-07-01 11:14:22.000000000 +0200
14572 +++ linux-3.0-rc5-vs2.3.1-pre3/ipc/namespace.c  2011-06-13 14:09:44.000000000 +0200
14573 @@ -13,11 +13,12 @@
14574  #include <linux/mount.h>
14575  #include <linux/user_namespace.h>
14576  #include <linux/proc_fs.h>
14577 +#include <linux/vs_base.h>
14578 +#include <linux/vserver/global.h>
14579
14580  #include "util.h"
14581
14582 -static struct ipc_namespace *create_ipc_ns(struct task_struct *tsk,
14583 -                                          struct ipc_namespace *old_ns)
14584 +static struct ipc_namespace *create_ipc_ns(struct user_namespace *user_ns)
14585  {
14586         struct ipc_namespace *ns;
14587         int err;
14588 @@ -46,19 +47,18 @@ static struct ipc_namespace *create_ipc_
14589         ipcns_notify(IPCNS_CREATED);
14590         register_ipcns_notifier(ns);
14591
14592 -       ns->user_ns = get_user_ns(task_cred_xxx(tsk, user)->user_ns);
14593 +       ns->user_ns = get_user_ns(user_ns);
14594
14595         return ns;
14596  }
14597
14598  struct ipc_namespace *copy_ipcs(unsigned long flags,
14599 -                               struct task_struct *tsk)
14600 +                               struct ipc_namespace *old_ns,
14601 +                               struct user_namespace *user_ns)
14602  {
14603 -       struct ipc_namespace *ns = tsk->nsproxy->ipc_ns;
14604 -
14605         if (!(flags & CLONE_NEWIPC))
14606 -               return get_ipc_ns(ns);
14607 -       return create_ipc_ns(tsk, ns);
14608 +               return get_ipc_ns(old_ns);
14609 +       return create_ipc_ns(user_ns);
14610  }
14611
14612  /*
14613 diff -NurpP --minimal linux-3.0-rc5/ipc/sem.c linux-3.0-rc5-vs2.3.1-pre3/ipc/sem.c
14614 --- linux-3.0-rc5/ipc/sem.c     2011-05-22 16:17:59.000000000 +0200
14615 +++ linux-3.0-rc5-vs2.3.1-pre3/ipc/sem.c        2011-06-10 22:11:24.000000000 +0200
14616 @@ -86,6 +86,8 @@
14617  #include <linux/rwsem.h>
14618  #include <linux/nsproxy.h>
14619  #include <linux/ipc_namespace.h>
14620 +#include <linux/vs_base.h>
14621 +#include <linux/vs_limit.h>
14622
14623  #include <asm/uaccess.h>
14624  #include "util.h"
14625 @@ -260,6 +262,7 @@ static int newary(struct ipc_namespace *
14626
14627         sma->sem_perm.mode = (semflg & S_IRWXUGO);
14628         sma->sem_perm.key = key;
14629 +       sma->sem_perm.xid = vx_current_xid();
14630
14631         sma->sem_perm.security = NULL;
14632         retval = security_sem_alloc(sma);
14633 @@ -275,6 +278,9 @@ static int newary(struct ipc_namespace *
14634                 return id;
14635         }
14636         ns->used_sems += nsems;
14637 +       /* FIXME: obsoleted? */
14638 +       vx_semary_inc(sma);
14639 +       vx_nsems_add(sma, nsems);
14640
14641         sma->sem_base = (struct sem *) &sma[1];
14642
14643 @@ -730,6 +736,9 @@ static void freeary(struct ipc_namespace
14644
14645         wake_up_sem_queue_do(&tasks);
14646         ns->used_sems -= sma->sem_nsems;
14647 +       /* FIXME: obsoleted? */
14648 +       vx_nsems_sub(sma, sma->sem_nsems);
14649 +       vx_semary_dec(sma);
14650         security_sem_free(sma);
14651         ipc_rcu_putref(sma);
14652  }
14653 diff -NurpP --minimal linux-3.0-rc5/ipc/shm.c linux-3.0-rc5-vs2.3.1-pre3/ipc/shm.c
14654 --- linux-3.0-rc5/ipc/shm.c     2011-07-01 11:14:22.000000000 +0200
14655 +++ linux-3.0-rc5-vs2.3.1-pre3/ipc/shm.c        2011-06-10 22:11:24.000000000 +0200
14656 @@ -39,6 +39,8 @@
14657  #include <linux/nsproxy.h>
14658  #include <linux/mount.h>
14659  #include <linux/ipc_namespace.h>
14660 +#include <linux/vs_context.h>
14661 +#include <linux/vs_limit.h>
14662
14663  #include <asm/uaccess.h>
14664
14665 @@ -173,7 +175,12 @@ static void shm_open(struct vm_area_stru
14666   */
14667  static void shm_destroy(struct ipc_namespace *ns, struct shmid_kernel *shp)
14668  {
14669 -       ns->shm_tot -= (shp->shm_segsz + PAGE_SIZE - 1) >> PAGE_SHIFT;
14670 +       struct vx_info *vxi = lookup_vx_info(shp->shm_perm.xid);
14671 +       int numpages = (shp->shm_segsz + PAGE_SIZE - 1) >> PAGE_SHIFT;
14672 +
14673 +       vx_ipcshm_sub(vxi, shp, numpages);
14674 +       ns->shm_tot -= numpages;
14675 +
14676         shm_rmid(ns, shp);
14677         shm_unlock(shp);
14678         if (!is_file_hugepages(shp->shm_file))
14679 @@ -183,6 +190,7 @@ static void shm_destroy(struct ipc_names
14680                                                 shp->mlock_user);
14681         fput (shp->shm_file);
14682         security_shm_free(shp);
14683 +       put_vx_info(vxi);
14684         ipc_rcu_putref(shp);
14685  }
14686
14687 @@ -355,11 +363,15 @@ static int newseg(struct ipc_namespace *
14688         if (ns->shm_tot + numpages > ns->shm_ctlall)
14689                 return -ENOSPC;
14690
14691 +       if (!vx_ipcshm_avail(current_vx_info(), numpages))
14692 +               return -ENOSPC;
14693 +
14694         shp = ipc_rcu_alloc(sizeof(*shp));
14695         if (!shp)
14696                 return -ENOMEM;
14697
14698         shp->shm_perm.key = key;
14699 +       shp->shm_perm.xid = vx_current_xid();
14700         shp->shm_perm.mode = (shmflg & S_IRWXUGO);
14701         shp->mlock_user = NULL;
14702
14703 @@ -413,6 +425,7 @@ static int newseg(struct ipc_namespace *
14704         ns->shm_tot += numpages;
14705         error = shp->shm_perm.id;
14706         shm_unlock(shp);
14707 +       vx_ipcshm_add(current_vx_info(), key, numpages);
14708         return error;
14709
14710  no_id:
14711 diff -NurpP --minimal linux-3.0-rc5/kernel/Makefile linux-3.0-rc5-vs2.3.1-pre3/kernel/Makefile
14712 --- linux-3.0-rc5/kernel/Makefile       2011-07-01 11:14:22.000000000 +0200
14713 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/Makefile  2011-06-10 22:11:24.000000000 +0200
14714 @@ -24,6 +24,7 @@ CFLAGS_REMOVE_sched_clock.o = -pg
14715  CFLAGS_REMOVE_irq_work.o = -pg
14716  endif
14717
14718 +obj-y += vserver/
14719  obj-$(CONFIG_FREEZER) += freezer.o
14720  obj-$(CONFIG_PROFILING) += profile.o
14721  obj-$(CONFIG_SYSCTL_SYSCALL_CHECK) += sysctl_check.o
14722 diff -NurpP --minimal linux-3.0-rc5/kernel/capability.c linux-3.0-rc5-vs2.3.1-pre3/kernel/capability.c
14723 --- linux-3.0-rc5/kernel/capability.c   2011-07-01 11:14:22.000000000 +0200
14724 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/capability.c      2011-06-10 22:21:49.000000000 +0200
14725 @@ -15,6 +15,7 @@
14726  #include <linux/syscalls.h>
14727  #include <linux/pid_namespace.h>
14728  #include <linux/user_namespace.h>
14729 +#include <linux/vs_context.h>
14730  #include <asm/uaccess.h>
14731
14732  /*
14733 @@ -116,6 +117,7 @@ static int cap_validate_magic(cap_user_h
14734         return 0;
14735  }
14736
14737 +
14738  /*
14739   * The only thing that can change the capabilities of the current
14740   * process is the current process. As such, we can't be in this code
14741 @@ -340,6 +342,8 @@ bool has_capability_noaudit(struct task_
14742         return (ret == 0);
14743  }
14744
14745 +#include <linux/vserver/base.h>
14746 +
14747  /**
14748   * capable - Determine if the current task has a superior capability in effect
14749   * @cap: The capability to be tested for
14750 @@ -369,6 +373,9 @@ EXPORT_SYMBOL(capable);
14751   */
14752  bool ns_capable(struct user_namespace *ns, int cap)
14753  {
14754 +       /* here for now so we don't require task locking */
14755 +       if (vs_check_bit(VXC_CAP_MASK, cap) && !vx_mcaps(1L << cap))
14756 +               return 0;
14757         if (unlikely(!cap_valid(cap))) {
14758                 printk(KERN_CRIT "capable() called with invalid cap=%u\n", cap);
14759                 BUG();
14760 diff -NurpP --minimal linux-3.0-rc5/kernel/compat.c linux-3.0-rc5-vs2.3.1-pre3/kernel/compat.c
14761 --- linux-3.0-rc5/kernel/compat.c       2011-07-01 11:14:22.000000000 +0200
14762 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/compat.c  2011-06-10 22:11:24.000000000 +0200
14763 @@ -970,7 +970,7 @@ asmlinkage long compat_sys_stime(compat_
14764         if (err)
14765                 return err;
14766
14767 -       do_settimeofday(&tv);
14768 +       vx_settimeofday(&tv);
14769         return 0;
14770  }
14771
14772 diff -NurpP --minimal linux-3.0-rc5/kernel/cred.c linux-3.0-rc5-vs2.3.1-pre3/kernel/cred.c
14773 --- linux-3.0-rc5/kernel/cred.c 2011-07-01 11:14:22.000000000 +0200
14774 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/cred.c    2011-06-10 22:11:24.000000000 +0200
14775 @@ -61,31 +61,6 @@ struct cred init_cred = {
14776  #endif
14777  };
14778
14779 -static inline void set_cred_subscribers(struct cred *cred, int n)
14780 -{
14781 -#ifdef CONFIG_DEBUG_CREDENTIALS
14782 -       atomic_set(&cred->subscribers, n);
14783 -#endif
14784 -}
14785 -
14786 -static inline int read_cred_subscribers(const struct cred *cred)
14787 -{
14788 -#ifdef CONFIG_DEBUG_CREDENTIALS
14789 -       return atomic_read(&cred->subscribers);
14790 -#else
14791 -       return 0;
14792 -#endif
14793 -}
14794 -
14795 -static inline void alter_cred_subscribers(const struct cred *_cred, int n)
14796 -{
14797 -#ifdef CONFIG_DEBUG_CREDENTIALS
14798 -       struct cred *cred = (struct cred *) _cred;
14799 -
14800 -       atomic_add(n, &cred->subscribers);
14801 -#endif
14802 -}
14803 -
14804  /*
14805   * Dispose of the shared task group credentials
14806   */
14807 @@ -281,21 +256,16 @@ error:
14808   *
14809   * Call commit_creds() or abort_creds() to clean up.
14810   */
14811 -struct cred *prepare_creds(void)
14812 +struct cred *__prepare_creds(const struct cred *old)
14813  {
14814 -       struct task_struct *task = current;
14815 -       const struct cred *old;
14816         struct cred *new;
14817
14818 -       validate_process_creds();
14819 -
14820         new = kmem_cache_alloc(cred_jar, GFP_KERNEL);
14821         if (!new)
14822                 return NULL;
14823
14824         kdebug("prepare_creds() alloc %p", new);
14825
14826 -       old = task->cred;
14827         memcpy(new, old, sizeof(struct cred));
14828
14829         atomic_set(&new->usage, 1);
14830 @@ -322,6 +292,13 @@ error:
14831         abort_creds(new);
14832         return NULL;
14833  }
14834 +
14835 +struct cred *prepare_creds(void)
14836 +{
14837 +       validate_process_creds();
14838 +
14839 +       return __prepare_creds(current->cred);
14840 +}
14841  EXPORT_SYMBOL(prepare_creds);
14842
14843  /*
14844 diff -NurpP --minimal linux-3.0-rc5/kernel/exit.c linux-3.0-rc5-vs2.3.1-pre3/kernel/exit.c
14845 --- linux-3.0-rc5/kernel/exit.c 2011-07-01 11:14:22.000000000 +0200
14846 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/exit.c    2011-06-22 12:39:15.000000000 +0200
14847 @@ -48,6 +48,10 @@
14848  #include <linux/fs_struct.h>
14849  #include <linux/init_task.h>
14850  #include <linux/perf_event.h>
14851 +#include <linux/vs_limit.h>
14852 +#include <linux/vs_context.h>
14853 +#include <linux/vs_network.h>
14854 +#include <linux/vs_pid.h>
14855  #include <trace/events/sched.h>
14856  #include <linux/hw_breakpoint.h>
14857  #include <linux/oom.h>
14858 @@ -494,9 +498,11 @@ static void close_files(struct files_str
14859                                         filp_close(file, files);
14860                                         cond_resched();
14861                                 }
14862 +                               vx_openfd_dec(i);
14863                         }
14864                         i++;
14865                         set >>= 1;
14866 +                       cond_resched();
14867                 }
14868         }
14869  }
14870 @@ -1047,11 +1053,16 @@ NORET_TYPE void do_exit(long code)
14871
14872         validate_creds_for_do_exit(tsk);
14873
14874 +       /* needs to stay after exit_notify() */
14875 +       exit_vx_info(tsk, code);
14876 +       exit_nx_info(tsk);
14877 +
14878         preempt_disable();
14879         exit_rcu();
14880         /* causes final put_task_struct in finish_task_switch(). */
14881         tsk->state = TASK_DEAD;
14882         schedule();
14883 +       printk("bad task: %p [%lx]\n", current, current->state);
14884         BUG();
14885         /* Avoid "noreturn function does return".  */
14886         for (;;)
14887 diff -NurpP --minimal linux-3.0-rc5/kernel/fork.c linux-3.0-rc5-vs2.3.1-pre3/kernel/fork.c
14888 --- linux-3.0-rc5/kernel/fork.c 2011-07-01 11:14:22.000000000 +0200
14889 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/fork.c    2011-06-10 22:21:02.000000000 +0200
14890 @@ -67,6 +67,10 @@
14891  #include <linux/user-return-notifier.h>
14892  #include <linux/oom.h>
14893  #include <linux/khugepaged.h>
14894 +#include <linux/vs_context.h>
14895 +#include <linux/vs_network.h>
14896 +#include <linux/vs_limit.h>
14897 +#include <linux/vs_memory.h>
14898
14899  #include <asm/pgtable.h>
14900  #include <asm/pgalloc.h>
14901 @@ -167,6 +171,8 @@ void free_task(struct task_struct *tsk)
14902         account_kernel_stack(tsk->stack, -1);
14903         free_thread_info(tsk->stack);
14904         rt_mutex_debug_task_free(tsk);
14905 +       clr_vx_info(&tsk->vx_info);
14906 +       clr_nx_info(&tsk->nx_info);
14907         ftrace_graph_exit_task(tsk);
14908         free_task_struct(tsk);
14909  }
14910 @@ -505,6 +511,7 @@ static struct mm_struct * mm_init(struct
14911         if (likely(!mm_alloc_pgd(mm))) {
14912                 mm->def_flags = 0;
14913                 mmu_notifier_mm_init(mm);
14914 +               set_vx_info(&mm->mm_vx_info, p->vx_info);
14915                 return mm;
14916         }
14917
14918 @@ -542,6 +549,7 @@ void __mmdrop(struct mm_struct *mm)
14919  #ifdef CONFIG_TRANSPARENT_HUGEPAGE
14920         VM_BUG_ON(mm->pmd_huge_pte);
14921  #endif
14922 +       clr_vx_info(&mm->mm_vx_info);
14923         free_mm(mm);
14924  }
14925  EXPORT_SYMBOL_GPL(__mmdrop);
14926 @@ -729,6 +737,7 @@ struct mm_struct *dup_mm(struct task_str
14927                 goto fail_nomem;
14928
14929         memcpy(mm, oldmm, sizeof(*mm));
14930 +       mm->mm_vx_info = NULL;
14931         mm_init_cpumask(mm);
14932
14933         /* Initializing for Swap token stuff */
14934 @@ -772,6 +781,7 @@ fail_nocontext:
14935          * If init_new_context() failed, we cannot use mmput() to free the mm
14936          * because it calls destroy_context()
14937          */
14938 +       clr_vx_info(&mm->mm_vx_info);
14939         mm_free_pgd(mm);
14940         free_mm(mm);
14941         return NULL;
14942 @@ -1057,6 +1067,8 @@ static struct task_struct *copy_process(
14943         int retval;
14944         struct task_struct *p;
14945         int cgroup_callbacks_done = 0;
14946 +       struct vx_info *vxi;
14947 +       struct nx_info *nxi;
14948
14949         if ((clone_flags & (CLONE_NEWNS|CLONE_FS)) == (CLONE_NEWNS|CLONE_FS))
14950                 return ERR_PTR(-EINVAL);
14951 @@ -1103,7 +1115,12 @@ static struct task_struct *copy_process(
14952         DEBUG_LOCKS_WARN_ON(!p->hardirqs_enabled);
14953         DEBUG_LOCKS_WARN_ON(!p->softirqs_enabled);
14954  #endif
14955 +       init_vx_info(&p->vx_info, current_vx_info());
14956 +       init_nx_info(&p->nx_info, current_nx_info());
14957 +
14958         retval = -EAGAIN;
14959 +       if (!vx_nproc_avail(1))
14960 +               goto bad_fork_free;
14961         if (atomic_read(&p->real_cred->user->processes) >=
14962                         task_rlimit(p, RLIMIT_NPROC)) {
14963                 if (!capable(CAP_SYS_ADMIN) && !capable(CAP_SYS_RESOURCE) &&
14964 @@ -1360,6 +1377,18 @@ static struct task_struct *copy_process(
14965
14966         total_forks++;
14967         spin_unlock(&current->sighand->siglock);
14968 +
14969 +       /* p is copy of current */
14970 +       vxi = p->vx_info;
14971 +       if (vxi) {
14972 +               claim_vx_info(vxi, p);
14973 +               atomic_inc(&vxi->cvirt.nr_threads);
14974 +               atomic_inc(&vxi->cvirt.total_forks);
14975 +               vx_nproc_inc(p);
14976 +       }
14977 +       nxi = p->nx_info;
14978 +       if (nxi)
14979 +               claim_nx_info(nxi, p);
14980         write_unlock_irq(&tasklist_lock);
14981         proc_fork_connector(p);
14982         cgroup_post_fork(p);
14983 diff -NurpP --minimal linux-3.0-rc5/kernel/kthread.c linux-3.0-rc5-vs2.3.1-pre3/kernel/kthread.c
14984 --- linux-3.0-rc5/kernel/kthread.c      2011-07-01 11:14:22.000000000 +0200
14985 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/kthread.c 2011-06-10 22:11:24.000000000 +0200
14986 @@ -16,6 +16,7 @@
14987  #include <linux/mutex.h>
14988  #include <linux/slab.h>
14989  #include <linux/freezer.h>
14990 +#include <linux/vs_pid.h>
14991  #include <trace/events/sched.h>
14992
14993  static DEFINE_SPINLOCK(kthread_create_lock);
14994 diff -NurpP --minimal linux-3.0-rc5/kernel/nsproxy.c linux-3.0-rc5-vs2.3.1-pre3/kernel/nsproxy.c
14995 --- linux-3.0-rc5/kernel/nsproxy.c      2011-07-01 11:14:23.000000000 +0200
14996 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/nsproxy.c 2011-06-13 14:09:44.000000000 +0200
14997 @@ -20,6 +20,8 @@
14998  #include <linux/mnt_namespace.h>
14999  #include <linux/utsname.h>
15000  #include <linux/pid_namespace.h>
15001 +#include <linux/vserver/global.h>
15002 +#include <linux/vserver/debug.h>
15003  #include <net/net_namespace.h>
15004  #include <linux/ipc_namespace.h>
15005  #include <linux/proc_fs.h>
15006 @@ -46,8 +48,11 @@ static inline struct nsproxy *create_nsp
15007         struct nsproxy *nsproxy;
15008
15009         nsproxy = kmem_cache_alloc(nsproxy_cachep, GFP_KERNEL);
15010 -       if (nsproxy)
15011 +       if (nsproxy) {
15012                 atomic_set(&nsproxy->count, 1);
15013 +               atomic_inc(&vs_global_nsproxy);
15014 +       }
15015 +       vxdprintk(VXD_CBIT(space, 2), "create_nsproxy = %p[1]", nsproxy);
15016         return nsproxy;
15017  }
15018
15019 @@ -56,8 +61,11 @@ static inline struct nsproxy *create_nsp
15020   * Return the newly created nsproxy.  Do not attach this to the task,
15021   * leave it to the caller to do proper locking and attach it to task.
15022   */
15023 -static struct nsproxy *create_new_namespaces(unsigned long flags,
15024 -                       struct task_struct *tsk, struct fs_struct *new_fs)
15025 +static struct nsproxy *unshare_namespaces(unsigned long flags,
15026 +                       struct nsproxy *orig,
15027 +                       struct fs_struct *new_fs,
15028 +                       struct user_namespace *new_user,
15029 +                       struct pid_namespace *new_pid)
15030  {
15031         struct nsproxy *new_nsp;
15032         int err;
15033 @@ -66,31 +74,31 @@ static struct nsproxy *create_new_namesp
15034         if (!new_nsp)
15035                 return ERR_PTR(-ENOMEM);
15036
15037 -       new_nsp->mnt_ns = copy_mnt_ns(flags, tsk->nsproxy->mnt_ns, new_fs);
15038 +       new_nsp->mnt_ns = copy_mnt_ns(flags, orig->mnt_ns, new_fs);
15039         if (IS_ERR(new_nsp->mnt_ns)) {
15040                 err = PTR_ERR(new_nsp->mnt_ns);
15041                 goto out_ns;
15042         }
15043
15044 -       new_nsp->uts_ns = copy_utsname(flags, tsk);
15045 +       new_nsp->uts_ns = copy_utsname(flags, orig->uts_ns, new_user);
15046         if (IS_ERR(new_nsp->uts_ns)) {
15047                 err = PTR_ERR(new_nsp->uts_ns);
15048                 goto out_uts;
15049         }
15050
15051 -       new_nsp->ipc_ns = copy_ipcs(flags, tsk);
15052 +       new_nsp->ipc_ns = copy_ipcs(flags, orig->ipc_ns, new_user);
15053         if (IS_ERR(new_nsp->ipc_ns)) {
15054                 err = PTR_ERR(new_nsp->ipc_ns);
15055                 goto out_ipc;
15056         }
15057
15058 -       new_nsp->pid_ns = copy_pid_ns(flags, task_active_pid_ns(tsk));
15059 +       new_nsp->pid_ns = copy_pid_ns(flags, new_pid);
15060         if (IS_ERR(new_nsp->pid_ns)) {
15061                 err = PTR_ERR(new_nsp->pid_ns);
15062                 goto out_pid;
15063         }
15064
15065 -       new_nsp->net_ns = copy_net_ns(flags, tsk->nsproxy->net_ns);
15066 +       new_nsp->net_ns = copy_net_ns(flags, orig->net_ns);
15067         if (IS_ERR(new_nsp->net_ns)) {
15068                 err = PTR_ERR(new_nsp->net_ns);
15069                 goto out_net;
15070 @@ -115,6 +123,40 @@ out_ns:
15071         return ERR_PTR(err);
15072  }
15073
15074 +static struct nsproxy *create_new_namespaces(unsigned long flags,
15075 +                       struct task_struct *tsk, struct fs_struct *new_fs)
15076 +{
15077 +       return unshare_namespaces(flags, tsk->nsproxy,
15078 +               new_fs, task_cred_xxx(tsk, user)->user_ns,
15079 +               task_active_pid_ns(tsk));
15080 +}
15081 +
15082 +/*
15083 + * copies the nsproxy, setting refcount to 1, and grabbing a
15084 + * reference to all contained namespaces.
15085 + */
15086 +struct nsproxy *copy_nsproxy(struct nsproxy *orig)
15087 +{
15088 +       struct nsproxy *ns = create_nsproxy();
15089 +
15090 +       if (ns) {
15091 +               memcpy(ns, orig, sizeof(struct nsproxy));
15092 +               atomic_set(&ns->count, 1);
15093 +
15094 +               if (ns->mnt_ns)
15095 +                       get_mnt_ns(ns->mnt_ns);
15096 +               if (ns->uts_ns)
15097 +                       get_uts_ns(ns->uts_ns);
15098 +               if (ns->ipc_ns)
15099 +                       get_ipc_ns(ns->ipc_ns);
15100 +               if (ns->pid_ns)
15101 +                       get_pid_ns(ns->pid_ns);
15102 +               if (ns->net_ns)
15103 +                       get_net(ns->net_ns);
15104 +       }
15105 +       return ns;
15106 +}
15107 +
15108  /*
15109   * called from clone.  This now handles copy for nsproxy and all
15110   * namespaces therein.
15111 @@ -122,9 +164,12 @@ out_ns:
15112  int copy_namespaces(unsigned long flags, struct task_struct *tsk)
15113  {
15114         struct nsproxy *old_ns = tsk->nsproxy;
15115 -       struct nsproxy *new_ns;
15116 +       struct nsproxy *new_ns = NULL;
15117         int err = 0;
15118
15119 +       vxdprintk(VXD_CBIT(space, 7), "copy_namespaces(0x%08lx,%p[%p])",
15120 +               flags, tsk, old_ns);
15121 +
15122         if (!old_ns)
15123                 return 0;
15124
15125 @@ -134,7 +179,7 @@ int copy_namespaces(unsigned long flags,
15126                                 CLONE_NEWPID | CLONE_NEWNET)))
15127                 return 0;
15128
15129 -       if (!capable(CAP_SYS_ADMIN)) {
15130 +       if (!vx_can_unshare(CAP_SYS_ADMIN, flags)) {
15131                 err = -EPERM;
15132                 goto out;
15133         }
15134 @@ -161,6 +206,9 @@ int copy_namespaces(unsigned long flags,
15135
15136  out:
15137         put_nsproxy(old_ns);
15138 +       vxdprintk(VXD_CBIT(space, 3),
15139 +               "copy_namespaces(0x%08lx,%p[%p]) = %d [%p]",
15140 +               flags, tsk, old_ns, err, new_ns);
15141         return err;
15142  }
15143
15144 @@ -174,7 +222,9 @@ void free_nsproxy(struct nsproxy *ns)
15145                 put_ipc_ns(ns->ipc_ns);
15146         if (ns->pid_ns)
15147                 put_pid_ns(ns->pid_ns);
15148 -       put_net(ns->net_ns);
15149 +       if (ns->net_ns)
15150 +               put_net(ns->net_ns);
15151 +       atomic_dec(&vs_global_nsproxy);
15152         kmem_cache_free(nsproxy_cachep, ns);
15153  }
15154
15155 @@ -187,11 +237,15 @@ int unshare_nsproxy_namespaces(unsigned
15156  {
15157         int err = 0;
15158
15159 +       vxdprintk(VXD_CBIT(space, 4),
15160 +               "unshare_nsproxy_namespaces(0x%08lx,[%p])",
15161 +               unshare_flags, current->nsproxy);
15162 +
15163         if (!(unshare_flags & (CLONE_NEWNS | CLONE_NEWUTS | CLONE_NEWIPC |
15164                                CLONE_NEWNET)))
15165                 return 0;
15166
15167 -       if (!capable(CAP_SYS_ADMIN))
15168 +       if (!vx_can_unshare(CAP_SYS_ADMIN, unshare_flags))
15169                 return -EPERM;
15170
15171         *new_nsp = create_new_namespaces(unshare_flags, current,
15172 diff -NurpP --minimal linux-3.0-rc5/kernel/pid.c linux-3.0-rc5-vs2.3.1-pre3/kernel/pid.c
15173 --- linux-3.0-rc5/kernel/pid.c  2011-05-22 16:17:59.000000000 +0200
15174 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/pid.c     2011-06-10 22:11:24.000000000 +0200
15175 @@ -36,6 +36,7 @@
15176  #include <linux/pid_namespace.h>
15177  #include <linux/init_task.h>
15178  #include <linux/syscalls.h>
15179 +#include <linux/vs_pid.h>
15180
15181  #define pid_hashfn(nr, ns)     \
15182         hash_long((unsigned long)nr + (unsigned long)ns, pidhash_shift)
15183 @@ -342,7 +343,7 @@ EXPORT_SYMBOL_GPL(find_pid_ns);
15184
15185  struct pid *find_vpid(int nr)
15186  {
15187 -       return find_pid_ns(nr, current->nsproxy->pid_ns);
15188 +       return find_pid_ns(vx_rmap_pid(nr), current->nsproxy->pid_ns);
15189  }
15190  EXPORT_SYMBOL_GPL(find_vpid);
15191
15192 @@ -402,6 +403,9 @@ void transfer_pid(struct task_struct *ol
15193  struct task_struct *pid_task(struct pid *pid, enum pid_type type)
15194  {
15195         struct task_struct *result = NULL;
15196 +
15197 +       if (type == PIDTYPE_REALPID)
15198 +               type = PIDTYPE_PID;
15199         if (pid) {
15200                 struct hlist_node *first;
15201                 first = rcu_dereference_check(hlist_first_rcu(&pid->tasks[type]),
15202 @@ -420,7 +424,7 @@ EXPORT_SYMBOL(pid_task);
15203  struct task_struct *find_task_by_pid_ns(pid_t nr, struct pid_namespace *ns)
15204  {
15205         rcu_lockdep_assert(rcu_read_lock_held());
15206 -       return pid_task(find_pid_ns(nr, ns), PIDTYPE_PID);
15207 +       return pid_task(find_pid_ns(vx_rmap_pid(nr), ns), PIDTYPE_PID);
15208  }
15209
15210  struct task_struct *find_task_by_vpid(pid_t vnr)
15211 @@ -464,7 +468,7 @@ struct pid *find_get_pid(pid_t nr)
15212  }
15213  EXPORT_SYMBOL_GPL(find_get_pid);
15214
15215 -pid_t pid_nr_ns(struct pid *pid, struct pid_namespace *ns)
15216 +pid_t pid_unmapped_nr_ns(struct pid *pid, struct pid_namespace *ns)
15217  {
15218         struct upid *upid;
15219         pid_t nr = 0;
15220 @@ -477,6 +481,11 @@ pid_t pid_nr_ns(struct pid *pid, struct
15221         return nr;
15222  }
15223
15224 +pid_t pid_nr_ns(struct pid *pid, struct pid_namespace *ns)
15225 +{
15226 +       return vx_map_pid(pid_unmapped_nr_ns(pid, ns));
15227 +}
15228 +
15229  pid_t pid_vnr(struct pid *pid)
15230  {
15231         return pid_nr_ns(pid, current->nsproxy->pid_ns);
15232 diff -NurpP --minimal linux-3.0-rc5/kernel/pid_namespace.c linux-3.0-rc5-vs2.3.1-pre3/kernel/pid_namespace.c
15233 --- linux-3.0-rc5/kernel/pid_namespace.c        2011-05-22 16:17:59.000000000 +0200
15234 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/pid_namespace.c   2011-06-10 22:17:45.000000000 +0200
15235 @@ -15,6 +15,7 @@
15236  #include <linux/acct.h>
15237  #include <linux/slab.h>
15238  #include <linux/proc_fs.h>
15239 +#include <linux/vserver/global.h>
15240
15241  #define BITS_PER_PAGE          (PAGE_SIZE*8)
15242
15243 @@ -88,6 +89,7 @@ static struct pid_namespace *create_pid_
15244                 goto out_free_map;
15245
15246         kref_init(&ns->kref);
15247 +       atomic_inc(&vs_global_pid_ns);
15248         ns->level = level;
15249         ns->parent = get_pid_ns(parent_pid_ns);
15250
15251 @@ -119,6 +121,7 @@ static void destroy_pid_namespace(struct
15252
15253         for (i = 0; i < PIDMAP_ENTRIES; i++)
15254                 kfree(ns->pidmap[i].page);
15255 +       atomic_dec(&vs_global_pid_ns);
15256         kmem_cache_free(pid_ns_cachep, ns);
15257  }
15258
15259 diff -NurpP --minimal linux-3.0-rc5/kernel/posix-timers.c linux-3.0-rc5-vs2.3.1-pre3/kernel/posix-timers.c
15260 --- linux-3.0-rc5/kernel/posix-timers.c 2011-07-01 11:14:23.000000000 +0200
15261 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/posix-timers.c    2011-06-10 22:11:24.000000000 +0200
15262 @@ -47,6 +47,7 @@
15263  #include <linux/wait.h>
15264  #include <linux/workqueue.h>
15265  #include <linux/module.h>
15266 +#include <linux/vs_context.h>
15267
15268  /*
15269   * Management arrays for POSIX timers.  Timers are kept in slab memory
15270 @@ -340,6 +341,7 @@ int posix_timer_event(struct k_itimer *t
15271  {
15272         struct task_struct *task;
15273         int shared, ret = -1;
15274 +
15275         /*
15276          * FIXME: if ->sigq is queued we can race with
15277          * dequeue_signal()->do_schedule_next_timer().
15278 @@ -356,10 +358,18 @@ int posix_timer_event(struct k_itimer *t
15279         rcu_read_lock();
15280         task = pid_task(timr->it_pid, PIDTYPE_PID);
15281         if (task) {
15282 +               struct vx_info_save vxis;
15283 +               struct vx_info *vxi;
15284 +
15285 +               vxi = get_vx_info(task->vx_info);
15286 +               enter_vx_info(vxi, &vxis);
15287                 shared = !(timr->it_sigev_notify & SIGEV_THREAD_ID);
15288                 ret = send_sigqueue(timr->sigq, task, shared);
15289 +               leave_vx_info(&vxis);
15290 +               put_vx_info(vxi);
15291         }
15292         rcu_read_unlock();
15293 +
15294         /* If we failed to send the signal the timer stops. */
15295         return ret > 0;
15296  }
15297 diff -NurpP --minimal linux-3.0-rc5/kernel/printk.c linux-3.0-rc5-vs2.3.1-pre3/kernel/printk.c
15298 --- linux-3.0-rc5/kernel/printk.c       2011-07-01 11:14:23.000000000 +0200
15299 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/printk.c  2011-06-10 22:11:24.000000000 +0200
15300 @@ -41,6 +41,7 @@
15301  #include <linux/cpu.h>
15302  #include <linux/notifier.h>
15303  #include <linux/rculist.h>
15304 +#include <linux/vs_cvirt.h>
15305
15306  #include <asm/uaccess.h>
15307
15308 @@ -314,7 +315,7 @@ static int check_syslog_permissions(int
15309                 return 0;
15310
15311         if (syslog_action_restricted(type)) {
15312 -               if (capable(CAP_SYSLOG))
15313 +               if (vx_capable(CAP_SYSLOG, VXC_SYSLOG))
15314                         return 0;
15315                 /* For historical reasons, accept CAP_SYS_ADMIN too, with a warning */
15316                 if (capable(CAP_SYS_ADMIN)) {
15317 @@ -342,12 +343,9 @@ int do_syslog(int type, char __user *buf
15318         if (error)
15319                 return error;
15320
15321 -       switch (type) {
15322 -       case SYSLOG_ACTION_CLOSE:       /* Close log */
15323 -               break;
15324 -       case SYSLOG_ACTION_OPEN:        /* Open log */
15325 -               break;
15326 -       case SYSLOG_ACTION_READ:        /* Read from log */
15327 +       if ((type == SYSLOG_ACTION_READ) ||
15328 +           (type == SYSLOG_ACTION_READ_ALL) ||
15329 +           (type == SYSLOG_ACTION_READ_CLEAR)) {
15330                 error = -EINVAL;
15331                 if (!buf || len < 0)
15332                         goto out;
15333 @@ -358,6 +356,16 @@ int do_syslog(int type, char __user *buf
15334                         error = -EFAULT;
15335                         goto out;
15336                 }
15337 +       }
15338 +       if (!vx_check(0, VS_ADMIN|VS_WATCH))
15339 +               return vx_do_syslog(type, buf, len);
15340 +
15341 +       switch (type) {
15342 +       case SYSLOG_ACTION_CLOSE:       /* Close log */
15343 +               break;
15344 +       case SYSLOG_ACTION_OPEN:        /* Open log */
15345 +               break;
15346 +       case SYSLOG_ACTION_READ:        /* Read from log */
15347                 error = wait_event_interruptible(log_wait,
15348                                                         (log_start - log_end));
15349                 if (error)
15350 @@ -384,16 +392,6 @@ int do_syslog(int type, char __user *buf
15351                 /* FALL THRU */
15352         /* Read last kernel messages */
15353         case SYSLOG_ACTION_READ_ALL:
15354 -               error = -EINVAL;
15355 -               if (!buf || len < 0)
15356 -                       goto out;
15357 -               error = 0;
15358 -               if (!len)
15359 -                       goto out;
15360 -               if (!access_ok(VERIFY_WRITE, buf, len)) {
15361 -                       error = -EFAULT;
15362 -                       goto out;
15363 -               }
15364                 count = len;
15365                 if (count > log_buf_len)
15366                         count = log_buf_len;
15367 diff -NurpP --minimal linux-3.0-rc5/kernel/ptrace.c linux-3.0-rc5-vs2.3.1-pre3/kernel/ptrace.c
15368 --- linux-3.0-rc5/kernel/ptrace.c       2011-07-01 11:14:23.000000000 +0200
15369 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/ptrace.c  2011-06-10 22:11:24.000000000 +0200
15370 @@ -22,6 +22,7 @@
15371  #include <linux/syscalls.h>
15372  #include <linux/uaccess.h>
15373  #include <linux/regset.h>
15374 +#include <linux/vs_context.h>
15375  #include <linux/hw_breakpoint.h>
15376
15377
15378 @@ -169,6 +170,11 @@ ok:
15379                 dumpable = get_dumpable(task->mm);
15380         if (!dumpable && !task_ns_capable(task, CAP_SYS_PTRACE))
15381                 return -EPERM;
15382 +       if (!vx_check(task->xid, VS_ADMIN_P|VS_WATCH_P|VS_IDENT))
15383 +               return -EPERM;
15384 +       if (!vx_check(task->xid, VS_IDENT) &&
15385 +               !task_vx_flags(task, VXF_STATE_ADMIN, 0))
15386 +               return -EACCES;
15387
15388         return security_ptrace_access_check(task, mode);
15389  }
15390 diff -NurpP --minimal linux-3.0-rc5/kernel/sched.c linux-3.0-rc5-vs2.3.1-pre3/kernel/sched.c
15391 --- linux-3.0-rc5/kernel/sched.c        2011-07-01 11:14:23.000000000 +0200
15392 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/sched.c   2011-06-15 02:40:14.000000000 +0200
15393 @@ -71,6 +71,8 @@
15394  #include <linux/ctype.h>
15395  #include <linux/ftrace.h>
15396  #include <linux/slab.h>
15397 +#include <linux/vs_sched.h>
15398 +#include <linux/vs_cvirt.h>
15399
15400  #include <asm/tlb.h>
15401  #include <asm/irq_regs.h>
15402 @@ -3429,9 +3431,17 @@ static void calc_global_nohz(unsigned lo
15403   */
15404  void get_avenrun(unsigned long *loads, unsigned long offset, int shift)
15405  {
15406 -       loads[0] = (avenrun[0] + offset) << shift;
15407 -       loads[1] = (avenrun[1] + offset) << shift;
15408 -       loads[2] = (avenrun[2] + offset) << shift;
15409 +       if (vx_flags(VXF_VIRT_LOAD, 0)) {
15410 +               struct vx_info *vxi = current_vx_info();
15411 +
15412 +               loads[0] = (vxi->cvirt.load[0] + offset) << shift;
15413 +               loads[1] = (vxi->cvirt.load[1] + offset) << shift;
15414 +               loads[2] = (vxi->cvirt.load[2] + offset) << shift;
15415 +       } else {
15416 +               loads[0] = (avenrun[0] + offset) << shift;
15417 +               loads[1] = (avenrun[1] + offset) << shift;
15418 +               loads[2] = (avenrun[2] + offset) << shift;
15419 +       }
15420  }
15421
15422  /*
15423 @@ -3714,16 +3724,19 @@ void account_user_time(struct task_struc
15424                        cputime_t cputime_scaled)
15425  {
15426         struct cpu_usage_stat *cpustat = &kstat_this_cpu.cpustat;
15427 +       struct vx_info *vxi = p->vx_info;  /* p is _always_ current */
15428         cputime64_t tmp;
15429 +       int nice = (TASK_NICE(p) > 0);
15430
15431         /* Add user time to process. */
15432         p->utime = cputime_add(p->utime, cputime);
15433         p->utimescaled = cputime_add(p->utimescaled, cputime_scaled);
15434 +       vx_account_user(vxi, cputime, nice);
15435         account_group_user_time(p, cputime);
15436
15437         /* Add user time to cpustat. */
15438         tmp = cputime_to_cputime64(cputime);
15439 -       if (TASK_NICE(p) > 0)
15440 +       if (nice)
15441                 cpustat->nice = cputime64_add(cpustat->nice, tmp);
15442         else
15443                 cpustat->user = cputime64_add(cpustat->user, tmp);
15444 @@ -3775,10 +3788,12 @@ void __account_system_time(struct task_s
15445                         cputime_t cputime_scaled, cputime64_t *target_cputime64)
15446  {
15447         cputime64_t tmp = cputime_to_cputime64(cputime);
15448 +       struct vx_info *vxi = p->vx_info;  /* p is _always_ current */
15449
15450         /* Add system time to process. */
15451         p->stime = cputime_add(p->stime, cputime);
15452         p->stimescaled = cputime_add(p->stimescaled, cputime_scaled);
15453 +       vx_account_system(vxi, cputime, 0 /* do we have idle time? */);
15454         account_group_system_time(p, cputime);
15455
15456         /* Add system time to cpustat. */
15457 @@ -4936,7 +4951,7 @@ SYSCALL_DEFINE1(nice, int, increment)
15458                 nice = 19;
15459
15460         if (increment < 0 && !can_nice(current, nice))
15461 -               return -EPERM;
15462 +               return vx_flags(VXF_IGNEG_NICE, 0) ? 0 : -EPERM;
15463
15464         retval = security_task_setnice(current, nice);
15465         if (retval)
15466 diff -NurpP --minimal linux-3.0-rc5/kernel/sched_fair.c linux-3.0-rc5-vs2.3.1-pre3/kernel/sched_fair.c
15467 --- linux-3.0-rc5/kernel/sched_fair.c   2011-07-01 11:14:23.000000000 +0200
15468 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/sched_fair.c      2011-06-10 22:11:24.000000000 +0200
15469 @@ -998,6 +998,8 @@ enqueue_entity(struct cfs_rq *cfs_rq, st
15470                 __enqueue_entity(cfs_rq, se);
15471         se->on_rq = 1;
15472
15473 +       if (entity_is_task(se))
15474 +               vx_activate_task(task_of(se));
15475         if (cfs_rq->nr_running == 1)
15476                 list_add_leaf_cfs_rq(cfs_rq);
15477  }
15478 @@ -1074,6 +1076,8 @@ dequeue_entity(struct cfs_rq *cfs_rq, st
15479         if (se != cfs_rq->curr)
15480                 __dequeue_entity(cfs_rq, se);
15481         se->on_rq = 0;
15482 +       if (entity_is_task(se))
15483 +               vx_deactivate_task(task_of(se));
15484         update_cfs_load(cfs_rq, 0);
15485         account_entity_dequeue(cfs_rq, se);
15486
15487 diff -NurpP --minimal linux-3.0-rc5/kernel/signal.c linux-3.0-rc5-vs2.3.1-pre3/kernel/signal.c
15488 --- linux-3.0-rc5/kernel/signal.c       2011-07-01 11:14:23.000000000 +0200
15489 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/signal.c  2011-06-22 12:39:16.000000000 +0200
15490 @@ -28,6 +28,8 @@
15491  #include <linux/freezer.h>
15492  #include <linux/pid_namespace.h>
15493  #include <linux/nsproxy.h>
15494 +#include <linux/vs_context.h>
15495 +#include <linux/vs_pid.h>
15496  #define CREATE_TRACE_POINTS
15497  #include <trace/events/signal.h>
15498
15499 @@ -744,9 +746,18 @@ static int check_kill_permission(int sig
15500         struct pid *sid;
15501         int error;
15502
15503 +       vxdprintk(VXD_CBIT(misc, 7),
15504 +               "check_kill_permission(%d,%p,%p[#%u,%u])",
15505 +               sig, info, t, vx_task_xid(t), t->pid);
15506 +
15507         if (!valid_signal(sig))
15508                 return -EINVAL;
15509
15510 +/*     FIXME: needed? if so, why?
15511 +       if ((info != SEND_SIG_NOINFO) &&
15512 +               (is_si_special(info) || !si_fromuser(info)))
15513 +               goto skip;      */
15514 +
15515         if (!si_fromuser(info))
15516                 return 0;
15517
15518 @@ -770,6 +781,20 @@ static int check_kill_permission(int sig
15519                 }
15520         }
15521
15522 +       error = -EPERM;
15523 +       if (t->pid == 1 && current->xid)
15524 +               return error;
15525 +
15526 +       error = -ESRCH;
15527 +       /* FIXME: we shouldn't return ESRCH ever, to avoid
15528 +                 loops, maybe ENOENT or EACCES? */
15529 +       if (!vx_check(vx_task_xid(t), VS_WATCH_P | VS_IDENT)) {
15530 +               vxdprintk(current->xid || VXD_CBIT(misc, 7),
15531 +                       "signal %d[%p] xid mismatch %p[#%u,%u] xid=#%u",
15532 +                       sig, info, t, vx_task_xid(t), t->pid, current->xid);
15533 +               return error;
15534 +       }
15535 +/* skip: */
15536         return security_task_kill(t, info, sig, 0);
15537  }
15538
15539 @@ -1239,7 +1264,7 @@ int kill_pid_info(int sig, struct siginf
15540         rcu_read_lock();
15541  retry:
15542         p = pid_task(pid, PIDTYPE_PID);
15543 -       if (p) {
15544 +       if (p && vx_check(vx_task_xid(p), VS_IDENT)) {
15545                 error = group_send_sig_info(sig, info, p);
15546                 if (unlikely(error == -ESRCH))
15547                         /*
15548 @@ -1278,7 +1303,7 @@ int kill_pid_info_as_uid(int sig, struct
15549
15550         rcu_read_lock();
15551         p = pid_task(pid, PIDTYPE_PID);
15552 -       if (!p) {
15553 +       if (!p || !vx_check(vx_task_xid(p), VS_IDENT)) {
15554                 ret = -ESRCH;
15555                 goto out_unlock;
15556         }
15557 @@ -1333,8 +1358,10 @@ static int kill_something_info(int sig,
15558                 struct task_struct * p;
15559
15560                 for_each_process(p) {
15561 -                       if (task_pid_vnr(p) > 1 &&
15562 -                                       !same_thread_group(p, current)) {
15563 +                       if (vx_check(vx_task_xid(p), VS_ADMIN|VS_IDENT) &&
15564 +                               task_pid_vnr(p) > 1 &&
15565 +                               !same_thread_group(p, current) &&
15566 +                               !vx_current_initpid(p->pid)) {
15567                                 int err = group_send_sig_info(sig, info, p);
15568                                 ++count;
15569                                 if (err != -EPERM)
15570 @@ -2133,6 +2160,11 @@ relock:
15571                                 !sig_kernel_only(signr))
15572                         continue;
15573
15574 +               /* virtual init is protected against user signals */
15575 +               if ((info->si_code == SI_USER) &&
15576 +                       vx_current_initpid(current->pid))
15577 +                       continue;
15578 +
15579                 if (sig_kernel_stop(signr)) {
15580                         /*
15581                          * The default action is to stop all threads in
15582 diff -NurpP --minimal linux-3.0-rc5/kernel/softirq.c linux-3.0-rc5-vs2.3.1-pre3/kernel/softirq.c
15583 --- linux-3.0-rc5/kernel/softirq.c      2011-07-01 11:14:23.000000000 +0200
15584 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/softirq.c 2011-06-22 12:39:16.000000000 +0200
15585 @@ -24,6 +24,7 @@
15586  #include <linux/ftrace.h>
15587  #include <linux/smp.h>
15588  #include <linux/tick.h>
15589 +#include <linux/vs_context.h>
15590
15591  #define CREATE_TRACE_POINTS
15592  #include <trace/events/irq.h>
15593 diff -NurpP --minimal linux-3.0-rc5/kernel/sys.c linux-3.0-rc5-vs2.3.1-pre3/kernel/sys.c
15594 --- linux-3.0-rc5/kernel/sys.c  2011-07-01 11:14:23.000000000 +0200
15595 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/sys.c     2011-06-10 22:11:24.000000000 +0200
15596 @@ -43,6 +43,7 @@
15597  #include <linux/syscalls.h>
15598  #include <linux/kprobes.h>
15599  #include <linux/user_namespace.h>
15600 +#include <linux/vs_pid.h>
15601
15602  #include <linux/kmsg_dump.h>
15603
15604 @@ -151,7 +152,10 @@ static int set_one_prio(struct task_stru
15605                 goto out;
15606         }
15607         if (niceval < task_nice(p) && !can_nice(p, niceval)) {
15608 -               error = -EACCES;
15609 +               if (vx_flags(VXF_IGNEG_NICE, 0))
15610 +                       error = 0;
15611 +               else
15612 +                       error = -EACCES;
15613                 goto out;
15614         }
15615         no_nice = security_task_setnice(p, niceval);
15616 @@ -201,6 +205,8 @@ SYSCALL_DEFINE3(setpriority, int, which,
15617                         else
15618                                 pgrp = task_pgrp(current);
15619                         do_each_pid_thread(pgrp, PIDTYPE_PGID, p) {
15620 +                               if (!vx_check(p->xid, VS_ADMIN_P | VS_IDENT))
15621 +                                       continue;
15622                                 error = set_one_prio(p, niceval, error);
15623                         } while_each_pid_thread(pgrp, PIDTYPE_PGID, p);
15624                         break;
15625 @@ -264,6 +270,8 @@ SYSCALL_DEFINE2(getpriority, int, which,
15626                         else
15627                                 pgrp = task_pgrp(current);
15628                         do_each_pid_thread(pgrp, PIDTYPE_PGID, p) {
15629 +                               if (!vx_check(p->xid, VS_ADMIN_P | VS_IDENT))
15630 +                                       continue;
15631                                 niceval = 20 - task_nice(p);
15632                                 if (niceval > retval)
15633                                         retval = niceval;
15634 @@ -383,6 +391,8 @@ EXPORT_SYMBOL_GPL(kernel_power_off);
15635
15636  static DEFINE_MUTEX(reboot_mutex);
15637
15638 +long vs_reboot(unsigned int, void __user *);
15639 +
15640  /*
15641   * Reboot system call: for obvious reasons only root may call it,
15642   * and even root needs to set up some magic numbers in the registers
15643 @@ -415,6 +425,9 @@ SYSCALL_DEFINE4(reboot, int, magic1, int
15644         if ((cmd == LINUX_REBOOT_CMD_POWER_OFF) && !pm_power_off)
15645                 cmd = LINUX_REBOOT_CMD_HALT;
15646
15647 +       if (!vx_check(0, VS_ADMIN|VS_WATCH))
15648 +               return vs_reboot(cmd, arg);
15649 +
15650         mutex_lock(&reboot_mutex);
15651         switch (cmd) {
15652         case LINUX_REBOOT_CMD_RESTART:
15653 @@ -1365,7 +1378,7 @@ int do_prlimit(struct task_struct *tsk,
15654                 /* Keep the capable check against init_user_ns until
15655                    cgroups can contain all limits */
15656                 if (new_rlim->rlim_max > rlim->rlim_max &&
15657 -                               !capable(CAP_SYS_RESOURCE))
15658 +                       !vx_capable(CAP_SYS_RESOURCE, VXC_SET_RLIMIT))
15659                         retval = -EPERM;
15660                 if (!retval)
15661                         retval = security_task_setrlimit(tsk->group_leader,
15662 diff -NurpP --minimal linux-3.0-rc5/kernel/sysctl.c linux-3.0-rc5-vs2.3.1-pre3/kernel/sysctl.c
15663 --- linux-3.0-rc5/kernel/sysctl.c       2011-07-01 11:14:23.000000000 +0200
15664 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/sysctl.c  2011-06-15 02:40:14.000000000 +0200
15665 @@ -75,6 +75,7 @@
15666  #if defined(CONFIG_PROVE_LOCKING) || defined(CONFIG_LOCK_STAT)
15667  #include <linux/lockdep.h>
15668  #endif
15669 +extern char vshelper_path[];
15670  #ifdef CONFIG_CHR_DEV_SG
15671  #include <scsi/sg.h>
15672  #endif
15673 @@ -568,6 +569,13 @@ static struct ctl_table kern_table[] = {
15674                 .proc_handler   = proc_dostring,
15675         },
15676  #endif
15677 +       {
15678 +               .procname       = "vshelper",
15679 +               .data           = &vshelper_path,
15680 +               .maxlen         = 256,
15681 +               .mode           = 0644,
15682 +               .proc_handler   = &proc_dostring,
15683 +       },
15684  #ifdef CONFIG_CHR_DEV_SG
15685         {
15686                 .procname       = "sg-big-buff",
15687 diff -NurpP --minimal linux-3.0-rc5/kernel/sysctl_binary.c linux-3.0-rc5-vs2.3.1-pre3/kernel/sysctl_binary.c
15688 --- linux-3.0-rc5/kernel/sysctl_binary.c        2011-05-22 16:17:59.000000000 +0200
15689 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/sysctl_binary.c   2011-06-10 22:11:24.000000000 +0200
15690 @@ -73,6 +73,7 @@ static const struct bin_table bin_kern_t
15691
15692         { CTL_INT,      KERN_PANIC,                     "panic" },
15693         { CTL_INT,      KERN_REALROOTDEV,               "real-root-dev" },
15694 +       { CTL_STR,      KERN_VSHELPER,                  "vshelper" },
15695
15696         { CTL_STR,      KERN_SPARC_REBOOT,              "reboot-cmd" },
15697         { CTL_INT,      KERN_CTLALTDEL,                 "ctrl-alt-del" },
15698 diff -NurpP --minimal linux-3.0-rc5/kernel/time/timekeeping.c linux-3.0-rc5-vs2.3.1-pre3/kernel/time/timekeeping.c
15699 --- linux-3.0-rc5/kernel/time/timekeeping.c     2011-07-01 11:14:23.000000000 +0200
15700 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/time/timekeeping.c        2011-06-10 22:11:24.000000000 +0200
15701 @@ -233,6 +233,7 @@ void getnstimeofday(struct timespec *ts)
15702         } while (read_seqretry(&xtime_lock, seq));
15703
15704         timespec_add_ns(ts, nsecs);
15705 +       vx_adjust_timespec(ts);
15706  }
15707
15708  EXPORT_SYMBOL(getnstimeofday);
15709 diff -NurpP --minimal linux-3.0-rc5/kernel/time.c linux-3.0-rc5-vs2.3.1-pre3/kernel/time.c
15710 --- linux-3.0-rc5/kernel/time.c 2011-05-22 16:17:59.000000000 +0200
15711 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/time.c    2011-06-10 22:11:24.000000000 +0200
15712 @@ -92,7 +92,7 @@ SYSCALL_DEFINE1(stime, time_t __user *,
15713         if (err)
15714                 return err;
15715
15716 -       do_settimeofday(&tv);
15717 +       vx_settimeofday(&tv);
15718         return 0;
15719  }
15720
15721 @@ -177,7 +177,7 @@ int do_sys_settimeofday(const struct tim
15722                 /* SMP safe, again the code in arch/foo/time.c should
15723                  * globally block out interrupts when it runs.
15724                  */
15725 -               return do_settimeofday(tv);
15726 +               return vx_settimeofday(tv);
15727         }
15728         return 0;
15729  }
15730 diff -NurpP --minimal linux-3.0-rc5/kernel/timer.c linux-3.0-rc5-vs2.3.1-pre3/kernel/timer.c
15731 --- linux-3.0-rc5/kernel/timer.c        2011-07-01 11:14:23.000000000 +0200
15732 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/timer.c   2011-06-15 02:40:14.000000000 +0200
15733 @@ -40,6 +40,10 @@
15734  #include <linux/irq_work.h>
15735  #include <linux/sched.h>
15736  #include <linux/slab.h>
15737 +#include <linux/vs_base.h>
15738 +#include <linux/vs_cvirt.h>
15739 +#include <linux/vs_pid.h>
15740 +#include <linux/vserver/sched.h>
15741
15742  #include <asm/uaccess.h>
15743  #include <asm/unistd.h>
15744 @@ -1336,12 +1340,6 @@ SYSCALL_DEFINE1(alarm, unsigned int, sec
15745
15746  #endif
15747
15748 -#ifndef __alpha__
15749 -
15750 -/*
15751 - * The Alpha uses getxpid, getxuid, and getxgid instead.  Maybe this
15752 - * should be moved into arch/i386 instead?
15753 - */
15754
15755  /**
15756   * sys_getpid - return the thread group id of the current process
15757 @@ -1370,10 +1368,23 @@ SYSCALL_DEFINE0(getppid)
15758         rcu_read_lock();
15759         pid = task_tgid_vnr(current->real_parent);
15760         rcu_read_unlock();
15761 +       return vx_map_pid(pid);
15762 +}
15763
15764 -       return pid;
15765 +#ifdef __alpha__
15766 +
15767 +/*
15768 + * The Alpha uses getxpid, getxuid, and getxgid instead.
15769 + */
15770 +
15771 +asmlinkage long do_getxpid(long *ppid)
15772 +{
15773 +       *ppid = sys_getppid();
15774 +       return sys_getpid();
15775  }
15776
15777 +#else /* _alpha_ */
15778 +
15779  SYSCALL_DEFINE0(getuid)
15780  {
15781         /* Only we change this so SMP safe */
15782 diff -NurpP --minimal linux-3.0-rc5/kernel/user_namespace.c linux-3.0-rc5-vs2.3.1-pre3/kernel/user_namespace.c
15783 --- linux-3.0-rc5/kernel/user_namespace.c       2011-03-15 18:07:42.000000000 +0100
15784 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/user_namespace.c  2011-06-10 22:11:24.000000000 +0200
15785 @@ -11,6 +11,7 @@
15786  #include <linux/user_namespace.h>
15787  #include <linux/highuid.h>
15788  #include <linux/cred.h>
15789 +#include <linux/vserver/global.h>
15790
15791  static struct kmem_cache *user_ns_cachep __read_mostly;
15792
15793 @@ -33,6 +34,7 @@ int create_user_ns(struct cred *new)
15794                 return -ENOMEM;
15795
15796         kref_init(&ns->kref);
15797 +       atomic_inc(&vs_global_user_ns);
15798
15799         for (n = 0; n < UIDHASH_SZ; ++n)
15800                 INIT_HLIST_HEAD(ns->uidhash_table + n);
15801 @@ -81,6 +83,8 @@ void free_user_ns(struct kref *kref)
15802         struct user_namespace *ns =
15803                 container_of(kref, struct user_namespace, kref);
15804
15805 +       /* FIXME: maybe move into destroyer? */
15806 +       atomic_dec(&vs_global_user_ns);
15807         INIT_WORK(&ns->destroyer, free_user_ns_work);
15808         schedule_work(&ns->destroyer);
15809  }
15810 diff -NurpP --minimal linux-3.0-rc5/kernel/utsname.c linux-3.0-rc5-vs2.3.1-pre3/kernel/utsname.c
15811 --- linux-3.0-rc5/kernel/utsname.c      2011-07-01 11:14:23.000000000 +0200
15812 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/utsname.c 2011-06-13 14:09:44.000000000 +0200
15813 @@ -16,14 +16,17 @@
15814  #include <linux/slab.h>
15815  #include <linux/user_namespace.h>
15816  #include <linux/proc_fs.h>
15817 +#include <linux/vserver/global.h>
15818
15819  static struct uts_namespace *create_uts_ns(void)
15820  {
15821         struct uts_namespace *uts_ns;
15822
15823         uts_ns = kmalloc(sizeof(struct uts_namespace), GFP_KERNEL);
15824 -       if (uts_ns)
15825 +       if (uts_ns) {
15826                 kref_init(&uts_ns->kref);
15827 +               atomic_inc(&vs_global_uts_ns);
15828 +       }
15829         return uts_ns;
15830  }
15831
15832 @@ -32,8 +35,8 @@ static struct uts_namespace *create_uts_
15833   * @old_ns: namespace to clone
15834   * Return NULL on error (failure to kmalloc), new ns otherwise
15835   */
15836 -static struct uts_namespace *clone_uts_ns(struct task_struct *tsk,
15837 -                                         struct uts_namespace *old_ns)
15838 +static struct uts_namespace *clone_uts_ns(struct uts_namespace *old_ns,
15839 +                                         struct user_namespace *old_user)
15840  {
15841         struct uts_namespace *ns;
15842
15843 @@ -43,7 +46,7 @@ static struct uts_namespace *clone_uts_n
15844
15845         down_read(&uts_sem);
15846         memcpy(&ns->name, &old_ns->name, sizeof(ns->name));
15847 -       ns->user_ns = get_user_ns(task_cred_xxx(tsk, user)->user_ns);
15848 +       ns->user_ns = get_user_ns(old_user);
15849         up_read(&uts_sem);
15850         return ns;
15851  }
15852 @@ -55,9 +58,9 @@ static struct uts_namespace *clone_uts_n
15853   * versa.
15854   */
15855  struct uts_namespace *copy_utsname(unsigned long flags,
15856 -                                  struct task_struct *tsk)
15857 +                                  struct uts_namespace *old_ns,
15858 +                                  struct user_namespace *user_ns)
15859  {
15860 -       struct uts_namespace *old_ns = tsk->nsproxy->uts_ns;
15861         struct uts_namespace *new_ns;
15862
15863         BUG_ON(!old_ns);
15864 @@ -66,7 +69,7 @@ struct uts_namespace *copy_utsname(unsig
15865         if (!(flags & CLONE_NEWUTS))
15866                 return old_ns;
15867
15868 -       new_ns = clone_uts_ns(tsk, old_ns);
15869 +       new_ns = clone_uts_ns(old_ns, user_ns);
15870
15871         put_uts_ns(old_ns);
15872         return new_ns;
15873 @@ -78,6 +81,7 @@ void free_uts_ns(struct kref *kref)
15874
15875         ns = container_of(kref, struct uts_namespace, kref);
15876         put_user_ns(ns->user_ns);
15877 +       atomic_dec(&vs_global_uts_ns);
15878         kfree(ns);
15879  }
15880
15881 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/Kconfig linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/Kconfig
15882 --- linux-3.0-rc5/kernel/vserver/Kconfig        1970-01-01 01:00:00.000000000 +0100
15883 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/Kconfig   2011-06-10 22:11:24.000000000 +0200
15884 @@ -0,0 +1,228 @@
15885 +#
15886 +# Linux VServer configuration
15887 +#
15888 +
15889 +menu "Linux VServer"
15890 +
15891 +config VSERVER_AUTO_LBACK
15892 +       bool    "Automatically Assign Loopback IP"
15893 +       default y
15894 +       help
15895 +         Automatically assign a guest specific loopback
15896 +         IP and add it to the kernel network stack on
15897 +         startup.
15898 +
15899 +config VSERVER_AUTO_SINGLE
15900 +       bool    "Automatic Single IP Special Casing"
15901 +       depends on EXPERIMENTAL
15902 +       default y
15903 +       help
15904 +         This allows network contexts with a single IP to
15905 +         automatically remap 0.0.0.0 bindings to that IP,
15906 +         avoiding further network checks and improving
15907 +         performance.
15908 +
15909 +         (note: such guests do not allow to change the ip
15910 +          on the fly and do not show loopback addresses)
15911 +
15912 +config VSERVER_COWBL
15913 +       bool    "Enable COW Immutable Link Breaking"
15914 +       default y
15915 +       help
15916 +         This enables the COW (Copy-On-Write) link break code.
15917 +         It allows you to treat unified files like normal files
15918 +         when writing to them (which will implicitely break the
15919 +         link and create a copy of the unified file)
15920 +
15921 +config VSERVER_VTIME
15922 +       bool    "Enable Virtualized Guest Time"
15923 +       depends on EXPERIMENTAL
15924 +       default n
15925 +       help
15926 +         This enables per guest time offsets to allow for
15927 +         adjusting the system clock individually per guest.
15928 +         this adds some overhead to the time functions and
15929 +         therefore should not be enabled without good reason.
15930 +
15931 +config VSERVER_DEVICE
15932 +       bool    "Enable Guest Device Mapping"
15933 +       depends on EXPERIMENTAL
15934 +       default n
15935 +       help
15936 +         This enables generic device remapping.
15937 +
15938 +config VSERVER_PROC_SECURE
15939 +       bool    "Enable Proc Security"
15940 +       depends on PROC_FS
15941 +       default y
15942 +       help
15943 +         This configures ProcFS security to initially hide
15944 +         non-process entries for all contexts except the main and
15945 +         spectator context (i.e. for all guests), which is a secure
15946 +         default.
15947 +
15948 +         (note: on 1.2x the entries were visible by default)
15949 +
15950 +choice
15951 +       prompt  "Persistent Inode Tagging"
15952 +       default TAGGING_ID24
15953 +       help
15954 +         This adds persistent context information to filesystems
15955 +         mounted with the tagxid option. Tagging is a requirement
15956 +         for per-context disk limits and per-context quota.
15957 +
15958 +
15959 +config TAGGING_NONE
15960 +       bool    "Disabled"
15961 +       help
15962 +         do not store per-context information in inodes.
15963 +
15964 +config TAGGING_UID16
15965 +       bool    "UID16/GID32"
15966 +       help
15967 +         reduces UID to 16 bit, but leaves GID at 32 bit.
15968 +
15969 +config TAGGING_GID16
15970 +       bool    "UID32/GID16"
15971 +       help
15972 +         reduces GID to 16 bit, but leaves UID at 32 bit.
15973 +
15974 +config TAGGING_ID24
15975 +       bool    "UID24/GID24"
15976 +       help
15977 +         uses the upper 8bit from UID and GID for XID tagging
15978 +         which leaves 24bit for UID/GID each, which should be
15979 +         more than sufficient for normal use.
15980 +
15981 +config TAGGING_INTERN
15982 +       bool    "UID32/GID32"
15983 +       help
15984 +         this uses otherwise reserved inode fields in the on
15985 +         disk representation, which limits the use to a few
15986 +         filesystems (currently ext2 and ext3)
15987 +
15988 +endchoice
15989 +
15990 +config TAG_NFSD
15991 +       bool    "Tag NFSD User Auth and Files"
15992 +       default n
15993 +       help
15994 +         Enable this if you do want the in-kernel NFS
15995 +         Server to use the tagging specified above.
15996 +         (will require patched clients too)
15997 +
15998 +config VSERVER_PRIVACY
15999 +       bool    "Honor Privacy Aspects of Guests"
16000 +       default n
16001 +       help
16002 +         When enabled, most context checks will disallow
16003 +         access to structures assigned to a specific context,
16004 +         like ptys or loop devices.
16005 +
16006 +config VSERVER_CONTEXTS
16007 +       int     "Maximum number of Contexts (1-65533)"  if EMBEDDED
16008 +       range 1 65533
16009 +       default "768"   if 64BIT
16010 +       default "256"
16011 +       help
16012 +         This setting will optimize certain data structures
16013 +         and memory allocations according to the expected
16014 +         maximum.
16015 +
16016 +         note: this is not a strict upper limit.
16017 +
16018 +config VSERVER_WARN
16019 +       bool    "VServer Warnings"
16020 +       default y
16021 +       help
16022 +         This enables various runtime warnings, which will
16023 +         notify about potential manipulation attempts or
16024 +         resource shortage. It is generally considered to
16025 +         be a good idea to have that enabled.
16026 +
16027 +config VSERVER_WARN_DEVPTS
16028 +       bool    "VServer DevPTS Warnings"
16029 +       depends on VSERVER_WARN
16030 +       default y
16031 +       help
16032 +         This enables DevPTS related warnings, issued when a
16033 +         process inside a context tries to lookup or access
16034 +         a dynamic pts from the host or a different context.
16035 +
16036 +config VSERVER_DEBUG
16037 +       bool    "VServer Debugging Code"
16038 +       default n
16039 +       help
16040 +         Set this to yes if you want to be able to activate
16041 +         debugging output at runtime. It adds a very small
16042 +         overhead to all vserver related functions and
16043 +         increases the kernel size by about 20k.
16044 +
16045 +config VSERVER_HISTORY
16046 +       bool    "VServer History Tracing"
16047 +       depends on VSERVER_DEBUG
16048 +       default n
16049 +       help
16050 +         Set this to yes if you want to record the history of
16051 +         linux-vserver activities, so they can be replayed in
16052 +         the event of a kernel panic or oops.
16053 +
16054 +config VSERVER_HISTORY_SIZE
16055 +       int     "Per-CPU History Size (32-65536)"
16056 +       depends on VSERVER_HISTORY
16057 +       range 32 65536
16058 +       default 64
16059 +       help
16060 +         This allows you to specify the number of entries in
16061 +         the per-CPU history buffer.
16062 +
16063 +config VSERVER_LEGACY_MEM
16064 +       bool    "Legacy Memory Limits"
16065 +       default n
16066 +       help
16067 +         This provides fake memory limits to keep
16068 +         older tools happy in the face of memory
16069 +         cgroups
16070 +
16071 +choice
16072 +       prompt  "Quotes used in debug and warn messages"
16073 +       default QUOTES_ISO8859
16074 +
16075 +config QUOTES_ISO8859
16076 +       bool    "Extended ASCII (ISO 8859) angle quotes"
16077 +       help
16078 +         This uses the extended ASCII characters \xbb
16079 +         and \xab for quoting file and process names.
16080 +
16081 +config QUOTES_UTF8
16082 +       bool    "UTF-8 angle quotes"
16083 +       help
16084 +         This uses the the UTF-8 sequences for angle
16085 +         quotes to quote file and process names.
16086 +
16087 +config QUOTES_ASCII
16088 +       bool    "ASCII single quotes"
16089 +       help
16090 +         This uses the ASCII single quote character
16091 +         (\x27) to quote file and process names.
16092 +
16093 +endchoice
16094 +
16095 +endmenu
16096 +
16097 +
16098 +config VSERVER
16099 +       bool
16100 +       default y
16101 +       select NAMESPACES
16102 +       select UTS_NS
16103 +       select IPC_NS
16104 +       select USER_NS
16105 +       select SYSVIPC
16106 +
16107 +config VSERVER_SECURITY
16108 +       bool
16109 +       depends on SECURITY
16110 +       default y
16111 +       select SECURITY_CAPABILITIES
16112 +
16113 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/Makefile linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/Makefile
16114 --- linux-3.0-rc5/kernel/vserver/Makefile       1970-01-01 01:00:00.000000000 +0100
16115 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/Makefile  2011-06-10 22:11:24.000000000 +0200
16116 @@ -0,0 +1,18 @@
16117 +#
16118 +# Makefile for the Linux vserver routines.
16119 +#
16120 +
16121 +
16122 +obj-y          += vserver.o
16123 +
16124 +vserver-y      := switch.o context.o space.o sched.o network.o inode.o \
16125 +                  limit.o cvirt.o cacct.o signal.o helper.o init.o \
16126 +                  dlimit.o tag.o
16127 +
16128 +vserver-$(CONFIG_INET) += inet.o
16129 +vserver-$(CONFIG_PROC_FS) += proc.o
16130 +vserver-$(CONFIG_VSERVER_DEBUG) += sysctl.o debug.o
16131 +vserver-$(CONFIG_VSERVER_HISTORY) += history.o
16132 +vserver-$(CONFIG_VSERVER_MONITOR) += monitor.o
16133 +vserver-$(CONFIG_VSERVER_DEVICE) += device.o
16134 +
16135 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/cacct.c linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/cacct.c
16136 --- linux-3.0-rc5/kernel/vserver/cacct.c        1970-01-01 01:00:00.000000000 +0100
16137 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/cacct.c   2011-06-10 22:11:24.000000000 +0200
16138 @@ -0,0 +1,42 @@
16139 +/*
16140 + *  linux/kernel/vserver/cacct.c
16141 + *
16142 + *  Virtual Server: Context Accounting
16143 + *
16144 + *  Copyright (C) 2006-2007 Herbert Pötzl
16145 + *
16146 + *  V0.01  added accounting stats
16147 + *
16148 + */
16149 +
16150 +#include <linux/types.h>
16151 +#include <linux/vs_context.h>
16152 +#include <linux/vserver/cacct_cmd.h>
16153 +#include <linux/vserver/cacct_int.h>
16154 +
16155 +#include <asm/errno.h>
16156 +#include <asm/uaccess.h>
16157 +
16158 +
16159 +int vc_sock_stat(struct vx_info *vxi, void __user *data)
16160 +{
16161 +       struct vcmd_sock_stat_v0 vc_data;
16162 +       int j, field;
16163 +
16164 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
16165 +               return -EFAULT;
16166 +
16167 +       field = vc_data.field;
16168 +       if ((field < 0) || (field >= VXA_SOCK_SIZE))
16169 +               return -EINVAL;
16170 +
16171 +       for (j = 0; j < 3; j++) {
16172 +               vc_data.count[j] = vx_sock_count(&vxi->cacct, field, j);
16173 +               vc_data.total[j] = vx_sock_total(&vxi->cacct, field, j);
16174 +       }
16175 +
16176 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
16177 +               return -EFAULT;
16178 +       return 0;
16179 +}
16180 +
16181 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/cacct_init.h linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/cacct_init.h
16182 --- linux-3.0-rc5/kernel/vserver/cacct_init.h   1970-01-01 01:00:00.000000000 +0100
16183 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/cacct_init.h      2011-06-10 22:11:24.000000000 +0200
16184 @@ -0,0 +1,25 @@
16185 +
16186 +
16187 +static inline void vx_info_init_cacct(struct _vx_cacct *cacct)
16188 +{
16189 +       int i, j;
16190 +
16191 +
16192 +       for (i = 0; i < VXA_SOCK_SIZE; i++) {
16193 +               for (j = 0; j < 3; j++) {
16194 +                       atomic_long_set(&cacct->sock[i][j].count, 0);
16195 +                       atomic_long_set(&cacct->sock[i][j].total, 0);
16196 +               }
16197 +       }
16198 +       for (i = 0; i < 8; i++)
16199 +               atomic_set(&cacct->slab[i], 0);
16200 +       for (i = 0; i < 5; i++)
16201 +               for (j = 0; j < 4; j++)
16202 +                       atomic_set(&cacct->page[i][j], 0);
16203 +}
16204 +
16205 +static inline void vx_info_exit_cacct(struct _vx_cacct *cacct)
16206 +{
16207 +       return;
16208 +}
16209 +
16210 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/cacct_proc.h linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/cacct_proc.h
16211 --- linux-3.0-rc5/kernel/vserver/cacct_proc.h   1970-01-01 01:00:00.000000000 +0100
16212 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/cacct_proc.h      2011-06-10 22:11:24.000000000 +0200
16213 @@ -0,0 +1,53 @@
16214 +#ifndef _VX_CACCT_PROC_H
16215 +#define _VX_CACCT_PROC_H
16216 +
16217 +#include <linux/vserver/cacct_int.h>
16218 +
16219 +
16220 +#define VX_SOCKA_TOP   \
16221 +       "Type\t    recv #/bytes\t\t   send #/bytes\t\t    fail #/bytes\n"
16222 +
16223 +static inline int vx_info_proc_cacct(struct _vx_cacct *cacct, char *buffer)
16224 +{
16225 +       int i, j, length = 0;
16226 +       static char *type[VXA_SOCK_SIZE] = {
16227 +               "UNSPEC", "UNIX", "INET", "INET6", "PACKET", "OTHER"
16228 +       };
16229 +
16230 +       length += sprintf(buffer + length, VX_SOCKA_TOP);
16231 +       for (i = 0; i < VXA_SOCK_SIZE; i++) {
16232 +               length += sprintf(buffer + length, "%s:", type[i]);
16233 +               for (j = 0; j < 3; j++) {
16234 +                       length += sprintf(buffer + length,
16235 +                               "\t%10lu/%-10lu",
16236 +                               vx_sock_count(cacct, i, j),
16237 +                               vx_sock_total(cacct, i, j));
16238 +               }
16239 +               buffer[length++] = '\n';
16240 +       }
16241 +
16242 +       length += sprintf(buffer + length, "\n");
16243 +       length += sprintf(buffer + length,
16244 +               "slab:\t %8u %8u %8u %8u\n",
16245 +               atomic_read(&cacct->slab[1]),
16246 +               atomic_read(&cacct->slab[4]),
16247 +               atomic_read(&cacct->slab[0]),
16248 +               atomic_read(&cacct->slab[2]));
16249 +
16250 +       length += sprintf(buffer + length, "\n");
16251 +       for (i = 0; i < 5; i++) {
16252 +               length += sprintf(buffer + length,
16253 +                       "page[%d]: %8u %8u %8u %8u\t %8u %8u %8u %8u\n", i,
16254 +                       atomic_read(&cacct->page[i][0]),
16255 +                       atomic_read(&cacct->page[i][1]),
16256 +                       atomic_read(&cacct->page[i][2]),
16257 +                       atomic_read(&cacct->page[i][3]),
16258 +                       atomic_read(&cacct->page[i][4]),
16259 +                       atomic_read(&cacct->page[i][5]),
16260 +                       atomic_read(&cacct->page[i][6]),
16261 +                       atomic_read(&cacct->page[i][7]));
16262 +       }
16263 +       return length;
16264 +}
16265 +
16266 +#endif /* _VX_CACCT_PROC_H */
16267 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/context.c linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/context.c
16268 --- linux-3.0-rc5/kernel/vserver/context.c      1970-01-01 01:00:00.000000000 +0100
16269 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/context.c 2011-06-16 14:16:51.000000000 +0200
16270 @@ -0,0 +1,1080 @@
16271 +/*
16272 + *  linux/kernel/vserver/context.c
16273 + *
16274 + *  Virtual Server: Context Support
16275 + *
16276 + *  Copyright (C) 2003-2010  Herbert Pötzl
16277 + *
16278 + *  V0.01  context helper
16279 + *  V0.02  vx_ctx_kill syscall command
16280 + *  V0.03  replaced context_info calls
16281 + *  V0.04  redesign of struct (de)alloc
16282 + *  V0.05  rlimit basic implementation
16283 + *  V0.06  task_xid and info commands
16284 + *  V0.07  context flags and caps
16285 + *  V0.08  switch to RCU based hash
16286 + *  V0.09  revert to non RCU for now
16287 + *  V0.10  and back to working RCU hash
16288 + *  V0.11  and back to locking again
16289 + *  V0.12  referenced context store
16290 + *  V0.13  separate per cpu data
16291 + *  V0.14  changed vcmds to vxi arg
16292 + *  V0.15  added context stat
16293 + *  V0.16  have __create claim() the vxi
16294 + *  V0.17  removed older and legacy stuff
16295 + *  V0.18  added user credentials
16296 + *
16297 + */
16298 +
16299 +#include <linux/slab.h>
16300 +#include <linux/types.h>
16301 +#include <linux/security.h>
16302 +#include <linux/pid_namespace.h>
16303 +#include <linux/capability.h>
16304 +
16305 +#include <linux/vserver/context.h>
16306 +#include <linux/vserver/network.h>
16307 +#include <linux/vserver/debug.h>
16308 +#include <linux/vserver/limit.h>
16309 +#include <linux/vserver/limit_int.h>
16310 +#include <linux/vserver/space.h>
16311 +#include <linux/init_task.h>
16312 +#include <linux/fs_struct.h>
16313 +#include <linux/cred.h>
16314 +
16315 +#include <linux/vs_context.h>
16316 +#include <linux/vs_limit.h>
16317 +#include <linux/vs_pid.h>
16318 +#include <linux/vserver/context_cmd.h>
16319 +
16320 +#include "cvirt_init.h"
16321 +#include "cacct_init.h"
16322 +#include "limit_init.h"
16323 +#include "sched_init.h"
16324 +
16325 +
16326 +atomic_t vx_global_ctotal      = ATOMIC_INIT(0);
16327 +atomic_t vx_global_cactive     = ATOMIC_INIT(0);
16328 +
16329 +
16330 +/*     now inactive context structures */
16331 +
16332 +static struct hlist_head vx_info_inactive = HLIST_HEAD_INIT;
16333 +
16334 +static DEFINE_SPINLOCK(vx_info_inactive_lock);
16335 +
16336 +
16337 +/*     __alloc_vx_info()
16338 +
16339 +       * allocate an initialized vx_info struct
16340 +       * doesn't make it visible (hash)                        */
16341 +
16342 +static struct vx_info *__alloc_vx_info(xid_t xid)
16343 +{
16344 +       struct vx_info *new = NULL;
16345 +       int cpu, index;
16346 +
16347 +       vxdprintk(VXD_CBIT(xid, 0), "alloc_vx_info(%d)*", xid);
16348 +
16349 +       /* would this benefit from a slab cache? */
16350 +       new = kmalloc(sizeof(struct vx_info), GFP_KERNEL);
16351 +       if (!new)
16352 +               return 0;
16353 +
16354 +       memset(new, 0, sizeof(struct vx_info));
16355 +#ifdef CONFIG_SMP
16356 +       new->ptr_pc = alloc_percpu(struct _vx_info_pc);
16357 +       if (!new->ptr_pc)
16358 +               goto error;
16359 +#endif
16360 +       new->vx_id = xid;
16361 +       INIT_HLIST_NODE(&new->vx_hlist);
16362 +       atomic_set(&new->vx_usecnt, 0);
16363 +       atomic_set(&new->vx_tasks, 0);
16364 +       new->vx_parent = NULL;
16365 +       new->vx_state = 0;
16366 +       init_waitqueue_head(&new->vx_wait);
16367 +
16368 +       /* prepare reaper */
16369 +       get_task_struct(init_pid_ns.child_reaper);
16370 +       new->vx_reaper = init_pid_ns.child_reaper;
16371 +       new->vx_badness_bias = 0;
16372 +
16373 +       /* rest of init goes here */
16374 +       vx_info_init_limit(&new->limit);
16375 +       vx_info_init_sched(&new->sched);
16376 +       vx_info_init_cvirt(&new->cvirt);
16377 +       vx_info_init_cacct(&new->cacct);
16378 +
16379 +       /* per cpu data structures */
16380 +       for_each_possible_cpu(cpu) {
16381 +               vx_info_init_sched_pc(
16382 +                       &vx_per_cpu(new, sched_pc, cpu), cpu);
16383 +               vx_info_init_cvirt_pc(
16384 +                       &vx_per_cpu(new, cvirt_pc, cpu), cpu);
16385 +       }
16386 +
16387 +       new->vx_flags = VXF_INIT_SET;
16388 +       new->vx_bcaps = CAP_FULL_SET;
16389 +       new->vx_ccaps = 0;
16390 +       new->vx_umask = 0;
16391 +
16392 +       new->reboot_cmd = 0;
16393 +       new->exit_code = 0;
16394 +
16395 +       // preconfig spaces
16396 +       for (index = 0; index < VX_SPACES; index++) {
16397 +               struct _vx_space *space = &new->space[index];
16398 +
16399 +               // filesystem
16400 +               spin_lock(&init_fs.lock);
16401 +               init_fs.users++;
16402 +               spin_unlock(&init_fs.lock);
16403 +               space->vx_fs = &init_fs;
16404 +
16405 +               /* FIXME: do we want defaults? */
16406 +               // space->vx_real_cred = 0;
16407 +               // space->vx_cred = 0;
16408 +       }
16409 +
16410 +
16411 +       vxdprintk(VXD_CBIT(xid, 0),
16412 +               "alloc_vx_info(%d) = %p", xid, new);
16413 +       vxh_alloc_vx_info(new);
16414 +       atomic_inc(&vx_global_ctotal);
16415 +       return new;
16416 +#ifdef CONFIG_SMP
16417 +error:
16418 +       kfree(new);
16419 +       return 0;
16420 +#endif
16421 +}
16422 +
16423 +/*     __dealloc_vx_info()
16424 +
16425 +       * final disposal of vx_info                             */
16426 +
16427 +static void __dealloc_vx_info(struct vx_info *vxi)
16428 +{
16429 +#ifdef CONFIG_VSERVER_WARN
16430 +       struct vx_info_save vxis;
16431 +       int cpu;
16432 +#endif
16433 +       vxdprintk(VXD_CBIT(xid, 0),
16434 +               "dealloc_vx_info(%p)", vxi);
16435 +       vxh_dealloc_vx_info(vxi);
16436 +
16437 +#ifdef CONFIG_VSERVER_WARN
16438 +       enter_vx_info(vxi, &vxis);
16439 +       vx_info_exit_limit(&vxi->limit);
16440 +       vx_info_exit_sched(&vxi->sched);
16441 +       vx_info_exit_cvirt(&vxi->cvirt);
16442 +       vx_info_exit_cacct(&vxi->cacct);
16443 +
16444 +       for_each_possible_cpu(cpu) {
16445 +               vx_info_exit_sched_pc(
16446 +                       &vx_per_cpu(vxi, sched_pc, cpu), cpu);
16447 +               vx_info_exit_cvirt_pc(
16448 +                       &vx_per_cpu(vxi, cvirt_pc, cpu), cpu);
16449 +       }
16450 +       leave_vx_info(&vxis);
16451 +#endif
16452 +
16453 +       vxi->vx_id = -1;
16454 +       vxi->vx_state |= VXS_RELEASED;
16455 +
16456 +#ifdef CONFIG_SMP
16457 +       free_percpu(vxi->ptr_pc);
16458 +#endif
16459 +       kfree(vxi);
16460 +       atomic_dec(&vx_global_ctotal);
16461 +}
16462 +
16463 +static void __shutdown_vx_info(struct vx_info *vxi)
16464 +{
16465 +       struct nsproxy *nsproxy;
16466 +       struct fs_struct *fs;
16467 +       struct cred *cred;
16468 +       int index, kill;
16469 +
16470 +       might_sleep();
16471 +
16472 +       vxi->vx_state |= VXS_SHUTDOWN;
16473 +       vs_state_change(vxi, VSC_SHUTDOWN);
16474 +
16475 +       for (index = 0; index < VX_SPACES; index++) {
16476 +               struct _vx_space *space = &vxi->space[index];
16477 +
16478 +               nsproxy = xchg(&space->vx_nsproxy, NULL);
16479 +               if (nsproxy)
16480 +                       put_nsproxy(nsproxy);
16481 +
16482 +               fs = xchg(&space->vx_fs, NULL);
16483 +               spin_lock(&fs->lock);
16484 +               kill = !--fs->users;
16485 +               spin_unlock(&fs->lock);
16486 +               if (kill)
16487 +                       free_fs_struct(fs);
16488 +
16489 +               cred = (struct cred *)xchg(&space->vx_cred, NULL);
16490 +               if (cred)
16491 +                       abort_creds(cred);
16492 +       }
16493 +}
16494 +
16495 +/* exported stuff */
16496 +
16497 +void free_vx_info(struct vx_info *vxi)
16498 +{
16499 +       unsigned long flags;
16500 +       unsigned index;
16501 +
16502 +       /* check for reference counts first */
16503 +       BUG_ON(atomic_read(&vxi->vx_usecnt));
16504 +       BUG_ON(atomic_read(&vxi->vx_tasks));
16505 +
16506 +       /* context must not be hashed */
16507 +       BUG_ON(vx_info_state(vxi, VXS_HASHED));
16508 +
16509 +       /* context shutdown is mandatory */
16510 +       BUG_ON(!vx_info_state(vxi, VXS_SHUTDOWN));
16511 +
16512 +       /* spaces check */
16513 +       for (index = 0; index < VX_SPACES; index++) {
16514 +               struct _vx_space *space = &vxi->space[index];
16515 +
16516 +               BUG_ON(space->vx_nsproxy);
16517 +               BUG_ON(space->vx_fs);
16518 +               // BUG_ON(space->vx_real_cred);
16519 +               // BUG_ON(space->vx_cred);
16520 +       }
16521 +
16522 +       spin_lock_irqsave(&vx_info_inactive_lock, flags);
16523 +       hlist_del(&vxi->vx_hlist);
16524 +       spin_unlock_irqrestore(&vx_info_inactive_lock, flags);
16525 +
16526 +       __dealloc_vx_info(vxi);
16527 +}
16528 +
16529 +
16530 +/*     hash table for vx_info hash */
16531 +
16532 +#define VX_HASH_SIZE   13
16533 +
16534 +static struct hlist_head vx_info_hash[VX_HASH_SIZE] =
16535 +       { [0 ... VX_HASH_SIZE-1] = HLIST_HEAD_INIT };
16536 +
16537 +static DEFINE_SPINLOCK(vx_info_hash_lock);
16538 +
16539 +
16540 +static inline unsigned int __hashval(xid_t xid)
16541 +{
16542 +       return (xid % VX_HASH_SIZE);
16543 +}
16544 +
16545 +
16546 +
16547 +/*     __hash_vx_info()
16548 +
16549 +       * add the vxi to the global hash table
16550 +       * requires the hash_lock to be held                     */
16551 +
16552 +static inline void __hash_vx_info(struct vx_info *vxi)
16553 +{
16554 +       struct hlist_head *head;
16555 +
16556 +       vxd_assert_lock(&vx_info_hash_lock);
16557 +       vxdprintk(VXD_CBIT(xid, 4),
16558 +               "__hash_vx_info: %p[#%d]", vxi, vxi->vx_id);
16559 +       vxh_hash_vx_info(vxi);
16560 +
16561 +       /* context must not be hashed */
16562 +       BUG_ON(vx_info_state(vxi, VXS_HASHED));
16563 +
16564 +       vxi->vx_state |= VXS_HASHED;
16565 +       head = &vx_info_hash[__hashval(vxi->vx_id)];
16566 +       hlist_add_head(&vxi->vx_hlist, head);
16567 +       atomic_inc(&vx_global_cactive);
16568 +}
16569 +
16570 +/*     __unhash_vx_info()
16571 +
16572 +       * remove the vxi from the global hash table
16573 +       * requires the hash_lock to be held                     */
16574 +
16575 +static inline void __unhash_vx_info(struct vx_info *vxi)
16576 +{
16577 +       unsigned long flags;
16578 +
16579 +       vxd_assert_lock(&vx_info_hash_lock);
16580 +       vxdprintk(VXD_CBIT(xid, 4),
16581 +               "__unhash_vx_info: %p[#%d.%d.%d]", vxi, vxi->vx_id,
16582 +               atomic_read(&vxi->vx_usecnt), atomic_read(&vxi->vx_tasks));
16583 +       vxh_unhash_vx_info(vxi);
16584 +
16585 +       /* context must be hashed */
16586 +       BUG_ON(!vx_info_state(vxi, VXS_HASHED));
16587 +       /* but without tasks */
16588 +       BUG_ON(atomic_read(&vxi->vx_tasks));
16589 +
16590 +       vxi->vx_state &= ~VXS_HASHED;
16591 +       hlist_del_init(&vxi->vx_hlist);
16592 +       spin_lock_irqsave(&vx_info_inactive_lock, flags);
16593 +       hlist_add_head(&vxi->vx_hlist, &vx_info_inactive);
16594 +       spin_unlock_irqrestore(&vx_info_inactive_lock, flags);
16595 +       atomic_dec(&vx_global_cactive);
16596 +}
16597 +
16598 +
16599 +/*     __lookup_vx_info()
16600 +
16601 +       * requires the hash_lock to be held
16602 +       * doesn't increment the vx_refcnt                       */
16603 +
16604 +static inline struct vx_info *__lookup_vx_info(xid_t xid)
16605 +{
16606 +       struct hlist_head *head = &vx_info_hash[__hashval(xid)];
16607 +       struct hlist_node *pos;
16608 +       struct vx_info *vxi;
16609 +
16610 +       vxd_assert_lock(&vx_info_hash_lock);
16611 +       hlist_for_each(pos, head) {
16612 +               vxi = hlist_entry(pos, struct vx_info, vx_hlist);
16613 +
16614 +               if (vxi->vx_id == xid)
16615 +                       goto found;
16616 +       }
16617 +       vxi = NULL;
16618 +found:
16619 +       vxdprintk(VXD_CBIT(xid, 0),
16620 +               "__lookup_vx_info(#%u): %p[#%u]",
16621 +               xid, vxi, vxi ? vxi->vx_id : 0);
16622 +       vxh_lookup_vx_info(vxi, xid);
16623 +       return vxi;
16624 +}
16625 +
16626 +
16627 +/*     __create_vx_info()
16628 +
16629 +       * create the requested context
16630 +       * get(), claim() and hash it                            */
16631 +
16632 +static struct vx_info *__create_vx_info(int id)
16633 +{
16634 +       struct vx_info *new, *vxi = NULL;
16635 +
16636 +       vxdprintk(VXD_CBIT(xid, 1), "create_vx_info(%d)*", id);
16637 +
16638 +       if (!(new = __alloc_vx_info(id)))
16639 +               return ERR_PTR(-ENOMEM);
16640 +
16641 +       /* required to make dynamic xids unique */
16642 +       spin_lock(&vx_info_hash_lock);
16643 +
16644 +       /* static context requested */
16645 +       if ((vxi = __lookup_vx_info(id))) {
16646 +               vxdprintk(VXD_CBIT(xid, 0),
16647 +                       "create_vx_info(%d) = %p (already there)", id, vxi);
16648 +               if (vx_info_flags(vxi, VXF_STATE_SETUP, 0))
16649 +                       vxi = ERR_PTR(-EBUSY);
16650 +               else
16651 +                       vxi = ERR_PTR(-EEXIST);
16652 +               goto out_unlock;
16653 +       }
16654 +       /* new context */
16655 +       vxdprintk(VXD_CBIT(xid, 0),
16656 +               "create_vx_info(%d) = %p (new)", id, new);
16657 +       claim_vx_info(new, NULL);
16658 +       __hash_vx_info(get_vx_info(new));
16659 +       vxi = new, new = NULL;
16660 +
16661 +out_unlock:
16662 +       spin_unlock(&vx_info_hash_lock);
16663 +       vxh_create_vx_info(IS_ERR(vxi) ? NULL : vxi, id);
16664 +       if (new)
16665 +               __dealloc_vx_info(new);
16666 +       return vxi;
16667 +}
16668 +
16669 +
16670 +/*     exported stuff                                          */
16671 +
16672 +
16673 +void unhash_vx_info(struct vx_info *vxi)
16674 +{
16675 +       spin_lock(&vx_info_hash_lock);
16676 +       __unhash_vx_info(vxi);
16677 +       spin_unlock(&vx_info_hash_lock);
16678 +       __shutdown_vx_info(vxi);
16679 +       __wakeup_vx_info(vxi);
16680 +}
16681 +
16682 +
16683 +/*     lookup_vx_info()
16684 +
16685 +       * search for a vx_info and get() it
16686 +       * negative id means current                             */
16687 +
16688 +struct vx_info *lookup_vx_info(int id)
16689 +{
16690 +       struct vx_info *vxi = NULL;
16691 +
16692 +       if (id < 0) {
16693 +               vxi = get_vx_info(current_vx_info());
16694 +       } else if (id > 1) {
16695 +               spin_lock(&vx_info_hash_lock);
16696 +               vxi = get_vx_info(__lookup_vx_info(id));
16697 +               spin_unlock(&vx_info_hash_lock);
16698 +       }
16699 +       return vxi;
16700 +}
16701 +
16702 +/*     xid_is_hashed()
16703 +
16704 +       * verify that xid is still hashed                       */
16705 +
16706 +int xid_is_hashed(xid_t xid)
16707 +{
16708 +       int hashed;
16709 +
16710 +       spin_lock(&vx_info_hash_lock);
16711 +       hashed = (__lookup_vx_info(xid) != NULL);
16712 +       spin_unlock(&vx_info_hash_lock);
16713 +       return hashed;
16714 +}
16715 +
16716 +#ifdef CONFIG_PROC_FS
16717 +
16718 +/*     get_xid_list()
16719 +
16720 +       * get a subset of hashed xids for proc
16721 +       * assumes size is at least one                          */
16722 +
16723 +int get_xid_list(int index, unsigned int *xids, int size)
16724 +{
16725 +       int hindex, nr_xids = 0;
16726 +
16727 +       /* only show current and children */
16728 +       if (!vx_check(0, VS_ADMIN | VS_WATCH)) {
16729 +               if (index > 0)
16730 +                       return 0;
16731 +               xids[nr_xids] = vx_current_xid();
16732 +               return 1;
16733 +       }
16734 +
16735 +       for (hindex = 0; hindex < VX_HASH_SIZE; hindex++) {
16736 +               struct hlist_head *head = &vx_info_hash[hindex];
16737 +               struct hlist_node *pos;
16738 +
16739 +               spin_lock(&vx_info_hash_lock);
16740 +               hlist_for_each(pos, head) {
16741 +                       struct vx_info *vxi;
16742 +
16743 +                       if (--index > 0)
16744 +                               continue;
16745 +
16746 +                       vxi = hlist_entry(pos, struct vx_info, vx_hlist);
16747 +                       xids[nr_xids] = vxi->vx_id;
16748 +                       if (++nr_xids >= size) {
16749 +                               spin_unlock(&vx_info_hash_lock);
16750 +                               goto out;
16751 +                       }
16752 +               }
16753 +               /* keep the lock time short */
16754 +               spin_unlock(&vx_info_hash_lock);
16755 +       }
16756 +out:
16757 +       return nr_xids;
16758 +}
16759 +#endif
16760 +
16761 +#ifdef CONFIG_VSERVER_DEBUG
16762 +
16763 +void   dump_vx_info_inactive(int level)
16764 +{
16765 +       struct hlist_node *entry, *next;
16766 +
16767 +       hlist_for_each_safe(entry, next, &vx_info_inactive) {
16768 +               struct vx_info *vxi =
16769 +                       list_entry(entry, struct vx_info, vx_hlist);
16770 +
16771 +               dump_vx_info(vxi, level);
16772 +       }
16773 +}
16774 +
16775 +#endif
16776 +
16777 +#if 0
16778 +int vx_migrate_user(struct task_struct *p, struct vx_info *vxi)
16779 +{
16780 +       struct user_struct *new_user, *old_user;
16781 +
16782 +       if (!p || !vxi)
16783 +               BUG();
16784 +
16785 +       if (vx_info_flags(vxi, VXF_INFO_PRIVATE, 0))
16786 +               return -EACCES;
16787 +
16788 +       new_user = alloc_uid(vxi->vx_id, p->uid);
16789 +       if (!new_user)
16790 +               return -ENOMEM;
16791 +
16792 +       old_user = p->user;
16793 +       if (new_user != old_user) {
16794 +               atomic_inc(&new_user->processes);
16795 +               atomic_dec(&old_user->processes);
16796 +               p->user = new_user;
16797 +       }
16798 +       free_uid(old_user);
16799 +       return 0;
16800 +}
16801 +#endif
16802 +
16803 +#if 0
16804 +void vx_mask_cap_bset(struct vx_info *vxi, struct task_struct *p)
16805 +{
16806 +       // p->cap_effective &= vxi->vx_cap_bset;
16807 +       p->cap_effective =
16808 +               cap_intersect(p->cap_effective, vxi->cap_bset);
16809 +       // p->cap_inheritable &= vxi->vx_cap_bset;
16810 +       p->cap_inheritable =
16811 +               cap_intersect(p->cap_inheritable, vxi->cap_bset);
16812 +       // p->cap_permitted &= vxi->vx_cap_bset;
16813 +       p->cap_permitted =
16814 +               cap_intersect(p->cap_permitted, vxi->cap_bset);
16815 +}
16816 +#endif
16817 +
16818 +
16819 +#include <linux/file.h>
16820 +#include <linux/fdtable.h>
16821 +
16822 +static int vx_openfd_task(struct task_struct *tsk)
16823 +{
16824 +       struct files_struct *files = tsk->files;
16825 +       struct fdtable *fdt;
16826 +       const unsigned long *bptr;
16827 +       int count, total;
16828 +
16829 +       /* no rcu_read_lock() because of spin_lock() */
16830 +       spin_lock(&files->file_lock);
16831 +       fdt = files_fdtable(files);
16832 +       bptr = fdt->open_fds->fds_bits;
16833 +       count = fdt->max_fds / (sizeof(unsigned long) * 8);
16834 +       for (total = 0; count > 0; count--) {
16835 +               if (*bptr)
16836 +                       total += hweight_long(*bptr);
16837 +               bptr++;
16838 +       }
16839 +       spin_unlock(&files->file_lock);
16840 +       return total;
16841 +}
16842 +
16843 +
16844 +/*     for *space compatibility */
16845 +
16846 +asmlinkage long sys_unshare(unsigned long);
16847 +
16848 +/*
16849 + *     migrate task to new context
16850 + *     gets vxi, puts old_vxi on change
16851 + *     optionally unshares namespaces (hack)
16852 + */
16853 +
16854 +int vx_migrate_task(struct task_struct *p, struct vx_info *vxi, int unshare)
16855 +{
16856 +       struct vx_info *old_vxi;
16857 +       int ret = 0;
16858 +
16859 +       if (!p || !vxi)
16860 +               BUG();
16861 +
16862 +       vxdprintk(VXD_CBIT(xid, 5),
16863 +               "vx_migrate_task(%p,%p[#%d.%d])", p, vxi,
16864 +               vxi->vx_id, atomic_read(&vxi->vx_usecnt));
16865 +
16866 +       if (vx_info_flags(vxi, VXF_INFO_PRIVATE, 0) &&
16867 +               !vx_info_flags(vxi, VXF_STATE_SETUP, 0))
16868 +               return -EACCES;
16869 +
16870 +       if (vx_info_state(vxi, VXS_SHUTDOWN))
16871 +               return -EFAULT;
16872 +
16873 +       old_vxi = task_get_vx_info(p);
16874 +       if (old_vxi == vxi)
16875 +               goto out;
16876 +
16877 +//     if (!(ret = vx_migrate_user(p, vxi))) {
16878 +       {
16879 +               int openfd;
16880 +
16881 +               task_lock(p);
16882 +               openfd = vx_openfd_task(p);
16883 +
16884 +               if (old_vxi) {
16885 +                       atomic_dec(&old_vxi->cvirt.nr_threads);
16886 +                       atomic_dec(&old_vxi->cvirt.nr_running);
16887 +                       __rlim_dec(&old_vxi->limit, RLIMIT_NPROC);
16888 +                       /* FIXME: what about the struct files here? */
16889 +                       __rlim_sub(&old_vxi->limit, VLIMIT_OPENFD, openfd);
16890 +                       /* account for the executable */
16891 +                       __rlim_dec(&old_vxi->limit, VLIMIT_DENTRY);
16892 +               }
16893 +               atomic_inc(&vxi->cvirt.nr_threads);
16894 +               atomic_inc(&vxi->cvirt.nr_running);
16895 +               __rlim_inc(&vxi->limit, RLIMIT_NPROC);
16896 +               /* FIXME: what about the struct files here? */
16897 +               __rlim_add(&vxi->limit, VLIMIT_OPENFD, openfd);
16898 +               /* account for the executable */
16899 +               __rlim_inc(&vxi->limit, VLIMIT_DENTRY);
16900 +
16901 +               if (old_vxi) {
16902 +                       release_vx_info(old_vxi, p);
16903 +                       clr_vx_info(&p->vx_info);
16904 +               }
16905 +               claim_vx_info(vxi, p);
16906 +               set_vx_info(&p->vx_info, vxi);
16907 +               p->xid = vxi->vx_id;
16908 +
16909 +               vxdprintk(VXD_CBIT(xid, 5),
16910 +                       "moved task %p into vxi:%p[#%d]",
16911 +                       p, vxi, vxi->vx_id);
16912 +
16913 +               // vx_mask_cap_bset(vxi, p);
16914 +               task_unlock(p);
16915 +
16916 +               /* hack for *spaces to provide compatibility */
16917 +               if (unshare) {
16918 +                       struct nsproxy *old_nsp, *new_nsp;
16919 +
16920 +                       ret = unshare_nsproxy_namespaces(
16921 +                               CLONE_NEWUTS | CLONE_NEWIPC | CLONE_NEWUSER,
16922 +                               &new_nsp, NULL);
16923 +                       if (ret)
16924 +                               goto out;
16925 +
16926 +                       old_nsp = xchg(&p->nsproxy, new_nsp);
16927 +                       vx_set_space(vxi,
16928 +                               CLONE_NEWUTS | CLONE_NEWIPC | CLONE_NEWUSER, 0);
16929 +                       put_nsproxy(old_nsp);
16930 +               }
16931 +       }
16932 +out:
16933 +       put_vx_info(old_vxi);
16934 +       return ret;
16935 +}
16936 +
16937 +int vx_set_reaper(struct vx_info *vxi, struct task_struct *p)
16938 +{
16939 +       struct task_struct *old_reaper;
16940 +
16941 +       if (!vxi)
16942 +               return -EINVAL;
16943 +
16944 +       vxdprintk(VXD_CBIT(xid, 6),
16945 +               "vx_set_reaper(%p[#%d],%p[#%d,%d])",
16946 +               vxi, vxi->vx_id, p, p->xid, p->pid);
16947 +
16948 +       old_reaper = vxi->vx_reaper;
16949 +       if (old_reaper == p)
16950 +               return 0;
16951 +
16952 +       /* set new child reaper */
16953 +       get_task_struct(p);
16954 +       vxi->vx_reaper = p;
16955 +       put_task_struct(old_reaper);
16956 +       return 0;
16957 +}
16958 +
16959 +int vx_set_init(struct vx_info *vxi, struct task_struct *p)
16960 +{
16961 +       if (!vxi)
16962 +               return -EINVAL;
16963 +
16964 +       vxdprintk(VXD_CBIT(xid, 6),
16965 +               "vx_set_init(%p[#%d],%p[#%d,%d,%d])",
16966 +               vxi, vxi->vx_id, p, p->xid, p->pid, p->tgid);
16967 +
16968 +       vxi->vx_flags &= ~VXF_STATE_INIT;
16969 +       // vxi->vx_initpid = p->tgid;
16970 +       vxi->vx_initpid = p->pid;
16971 +       return 0;
16972 +}
16973 +
16974 +void vx_exit_init(struct vx_info *vxi, struct task_struct *p, int code)
16975 +{
16976 +       vxdprintk(VXD_CBIT(xid, 6),
16977 +               "vx_exit_init(%p[#%d],%p[#%d,%d,%d])",
16978 +               vxi, vxi->vx_id, p, p->xid, p->pid, p->tgid);
16979 +
16980 +       vxi->exit_code = code;
16981 +       vxi->vx_initpid = 0;
16982 +}
16983 +
16984 +
16985 +void vx_set_persistent(struct vx_info *vxi)
16986 +{
16987 +       vxdprintk(VXD_CBIT(xid, 6),
16988 +               "vx_set_persistent(%p[#%d])", vxi, vxi->vx_id);
16989 +
16990 +       get_vx_info(vxi);
16991 +       claim_vx_info(vxi, NULL);
16992 +}
16993 +
16994 +void vx_clear_persistent(struct vx_info *vxi)
16995 +{
16996 +       vxdprintk(VXD_CBIT(xid, 6),
16997 +               "vx_clear_persistent(%p[#%d])", vxi, vxi->vx_id);
16998 +
16999 +       release_vx_info(vxi, NULL);
17000 +       put_vx_info(vxi);
17001 +}
17002 +
17003 +void vx_update_persistent(struct vx_info *vxi)
17004 +{
17005 +       if (vx_info_flags(vxi, VXF_PERSISTENT, 0))
17006 +               vx_set_persistent(vxi);
17007 +       else
17008 +               vx_clear_persistent(vxi);
17009 +}
17010 +
17011 +
17012 +/*     task must be current or locked          */
17013 +
17014 +void   exit_vx_info(struct task_struct *p, int code)
17015 +{
17016 +       struct vx_info *vxi = p->vx_info;
17017 +
17018 +       if (vxi) {
17019 +               atomic_dec(&vxi->cvirt.nr_threads);
17020 +               vx_nproc_dec(p);
17021 +
17022 +               vxi->exit_code = code;
17023 +               release_vx_info(vxi, p);
17024 +       }
17025 +}
17026 +
17027 +void   exit_vx_info_early(struct task_struct *p, int code)
17028 +{
17029 +       struct vx_info *vxi = p->vx_info;
17030 +
17031 +       if (vxi) {
17032 +               if (vxi->vx_initpid == p->pid)
17033 +                       vx_exit_init(vxi, p, code);
17034 +               if (vxi->vx_reaper == p)
17035 +                       vx_set_reaper(vxi, init_pid_ns.child_reaper);
17036 +       }
17037 +}
17038 +
17039 +
17040 +/* vserver syscall commands below here */
17041 +
17042 +/* taks xid and vx_info functions */
17043 +
17044 +#include <asm/uaccess.h>
17045 +
17046 +
17047 +int vc_task_xid(uint32_t id)
17048 +{
17049 +       xid_t xid;
17050 +
17051 +       if (id) {
17052 +               struct task_struct *tsk;
17053 +
17054 +               rcu_read_lock();
17055 +               tsk = find_task_by_real_pid(id);
17056 +               xid = (tsk) ? tsk->xid : -ESRCH;
17057 +               rcu_read_unlock();
17058 +       } else
17059 +               xid = vx_current_xid();
17060 +       return xid;
17061 +}
17062 +
17063 +
17064 +int vc_vx_info(struct vx_info *vxi, void __user *data)
17065 +{
17066 +       struct vcmd_vx_info_v0 vc_data;
17067 +
17068 +       vc_data.xid = vxi->vx_id;
17069 +       vc_data.initpid = vxi->vx_initpid;
17070 +
17071 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
17072 +               return -EFAULT;
17073 +       return 0;
17074 +}
17075 +
17076 +
17077 +int vc_ctx_stat(struct vx_info *vxi, void __user *data)
17078 +{
17079 +       struct vcmd_ctx_stat_v0 vc_data;
17080 +
17081 +       vc_data.usecnt = atomic_read(&vxi->vx_usecnt);
17082 +       vc_data.tasks = atomic_read(&vxi->vx_tasks);
17083 +
17084 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
17085 +               return -EFAULT;
17086 +       return 0;
17087 +}
17088 +
17089 +
17090 +/* context functions */
17091 +
17092 +int vc_ctx_create(uint32_t xid, void __user *data)
17093 +{
17094 +       struct vcmd_ctx_create vc_data = { .flagword = VXF_INIT_SET };
17095 +       struct vx_info *new_vxi;
17096 +       int ret;
17097 +
17098 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
17099 +               return -EFAULT;
17100 +
17101 +       if ((xid > MAX_S_CONTEXT) || (xid < 2))
17102 +               return -EINVAL;
17103 +
17104 +       new_vxi = __create_vx_info(xid);
17105 +       if (IS_ERR(new_vxi))
17106 +               return PTR_ERR(new_vxi);
17107 +
17108 +       /* initial flags */
17109 +       new_vxi->vx_flags = vc_data.flagword;
17110 +
17111 +       ret = -ENOEXEC;
17112 +       if (vs_state_change(new_vxi, VSC_STARTUP))
17113 +               goto out;
17114 +
17115 +       ret = vx_migrate_task(current, new_vxi, (!data));
17116 +       if (ret)
17117 +               goto out;
17118 +
17119 +       /* return context id on success */
17120 +       ret = new_vxi->vx_id;
17121 +
17122 +       /* get a reference for persistent contexts */
17123 +       if ((vc_data.flagword & VXF_PERSISTENT))
17124 +               vx_set_persistent(new_vxi);
17125 +out:
17126 +       release_vx_info(new_vxi, NULL);
17127 +       put_vx_info(new_vxi);
17128 +       return ret;
17129 +}
17130 +
17131 +
17132 +int vc_ctx_migrate(struct vx_info *vxi, void __user *data)
17133 +{
17134 +       struct vcmd_ctx_migrate vc_data = { .flagword = 0 };
17135 +       int ret;
17136 +
17137 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
17138 +               return -EFAULT;
17139 +
17140 +       ret = vx_migrate_task(current, vxi, 0);
17141 +       if (ret)
17142 +               return ret;
17143 +       if (vc_data.flagword & VXM_SET_INIT)
17144 +               ret = vx_set_init(vxi, current);
17145 +       if (ret)
17146 +               return ret;
17147 +       if (vc_data.flagword & VXM_SET_REAPER)
17148 +               ret = vx_set_reaper(vxi, current);
17149 +       return ret;
17150 +}
17151 +
17152 +
17153 +int vc_get_cflags(struct vx_info *vxi, void __user *data)
17154 +{
17155 +       struct vcmd_ctx_flags_v0 vc_data;
17156 +
17157 +       vc_data.flagword = vxi->vx_flags;
17158 +
17159 +       /* special STATE flag handling */
17160 +       vc_data.mask = vs_mask_flags(~0ULL, vxi->vx_flags, VXF_ONE_TIME);
17161 +
17162 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
17163 +               return -EFAULT;
17164 +       return 0;
17165 +}
17166 +
17167 +int vc_set_cflags(struct vx_info *vxi, void __user *data)
17168 +{
17169 +       struct vcmd_ctx_flags_v0 vc_data;
17170 +       uint64_t mask, trigger;
17171 +
17172 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
17173 +               return -EFAULT;
17174 +
17175 +       /* special STATE flag handling */
17176 +       mask = vs_mask_mask(vc_data.mask, vxi->vx_flags, VXF_ONE_TIME);
17177 +       trigger = (mask & vxi->vx_flags) ^ (mask & vc_data.flagword);
17178 +
17179 +       if (vxi == current_vx_info()) {
17180 +               /* if (trigger & VXF_STATE_SETUP)
17181 +                       vx_mask_cap_bset(vxi, current); */
17182 +               if (trigger & VXF_STATE_INIT) {
17183 +                       int ret;
17184 +
17185 +                       ret = vx_set_init(vxi, current);
17186 +                       if (ret)
17187 +                               return ret;
17188 +                       ret = vx_set_reaper(vxi, current);
17189 +                       if (ret)
17190 +                               return ret;
17191 +               }
17192 +       }
17193 +
17194 +       vxi->vx_flags = vs_mask_flags(vxi->vx_flags,
17195 +               vc_data.flagword, mask);
17196 +       if (trigger & VXF_PERSISTENT)
17197 +               vx_update_persistent(vxi);
17198 +
17199 +       return 0;
17200 +}
17201 +
17202 +
17203 +static inline uint64_t caps_from_cap_t(kernel_cap_t c)
17204 +{
17205 +       uint64_t v = c.cap[0] | ((uint64_t)c.cap[1] << 32);
17206 +
17207 +       // printk("caps_from_cap_t(%08x:%08x) = %016llx\n", c.cap[1], c.cap[0], v);
17208 +       return v;
17209 +}
17210 +
17211 +static inline kernel_cap_t cap_t_from_caps(uint64_t v)
17212 +{
17213 +       kernel_cap_t c = __cap_empty_set;
17214 +
17215 +       c.cap[0] = v & 0xFFFFFFFF;
17216 +       c.cap[1] = (v >> 32) & 0xFFFFFFFF;
17217 +
17218 +       // printk("cap_t_from_caps(%016llx) = %08x:%08x\n", v, c.cap[1], c.cap[0]);
17219 +       return c;
17220 +}
17221 +
17222 +
17223 +static int do_get_caps(struct vx_info *vxi, uint64_t *bcaps, uint64_t *ccaps)
17224 +{
17225 +       if (bcaps)
17226 +               *bcaps = caps_from_cap_t(vxi->vx_bcaps);
17227 +       if (ccaps)
17228 +               *ccaps = vxi->vx_ccaps;
17229 +
17230 +       return 0;
17231 +}
17232 +
17233 +int vc_get_ccaps(struct vx_info *vxi, void __user *data)
17234 +{
17235 +       struct vcmd_ctx_caps_v1 vc_data;
17236 +       int ret;
17237 +
17238 +       ret = do_get_caps(vxi, NULL, &vc_data.ccaps);
17239 +       if (ret)
17240 +               return ret;
17241 +       vc_data.cmask = ~0ULL;
17242 +
17243 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
17244 +               return -EFAULT;
17245 +       return 0;
17246 +}
17247 +
17248 +static int do_set_caps(struct vx_info *vxi,
17249 +       uint64_t bcaps, uint64_t bmask, uint64_t ccaps, uint64_t cmask)
17250 +{
17251 +       uint64_t bcold = caps_from_cap_t(vxi->vx_bcaps);
17252 +
17253 +#if 0
17254 +       printk("do_set_caps(%16llx, %16llx, %16llx, %16llx)\n",
17255 +               bcaps, bmask, ccaps, cmask);
17256 +#endif
17257 +       vxi->vx_bcaps = cap_t_from_caps(
17258 +               vs_mask_flags(bcold, bcaps, bmask));
17259 +       vxi->vx_ccaps = vs_mask_flags(vxi->vx_ccaps, ccaps, cmask);
17260 +
17261 +       return 0;
17262 +}
17263 +
17264 +int vc_set_ccaps(struct vx_info *vxi, void __user *data)
17265 +{
17266 +       struct vcmd_ctx_caps_v1 vc_data;
17267 +
17268 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
17269 +               return -EFAULT;
17270 +
17271 +       return do_set_caps(vxi, 0, 0, vc_data.ccaps, vc_data.cmask);
17272 +}
17273 +
17274 +int vc_get_bcaps(struct vx_info *vxi, void __user *data)
17275 +{
17276 +       struct vcmd_bcaps vc_data;
17277 +       int ret;
17278 +
17279 +       ret = do_get_caps(vxi, &vc_data.bcaps, NULL);
17280 +       if (ret)
17281 +               return ret;
17282 +       vc_data.bmask = ~0ULL;
17283 +
17284 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
17285 +               return -EFAULT;
17286 +       return 0;
17287 +}
17288 +
17289 +int vc_set_bcaps(struct vx_info *vxi, void __user *data)
17290 +{
17291 +       struct vcmd_bcaps vc_data;
17292 +
17293 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
17294 +               return -EFAULT;
17295 +
17296 +       return do_set_caps(vxi, vc_data.bcaps, vc_data.bmask, 0, 0);
17297 +}
17298 +
17299 +
17300 +int vc_get_umask(struct vx_info *vxi, void __user *data)
17301 +{
17302 +       struct vcmd_umask vc_data;
17303 +
17304 +       vc_data.umask = vxi->vx_umask;
17305 +       vc_data.mask = ~0ULL;
17306 +
17307 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
17308 +               return -EFAULT;
17309 +       return 0;
17310 +}
17311 +
17312 +int vc_set_umask(struct vx_info *vxi, void __user *data)
17313 +{
17314 +       struct vcmd_umask vc_data;
17315 +
17316 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
17317 +               return -EFAULT;
17318 +
17319 +       vxi->vx_umask = vs_mask_flags(vxi->vx_umask,
17320 +               vc_data.umask, vc_data.mask);
17321 +       return 0;
17322 +}
17323 +
17324 +
17325 +int vc_get_badness(struct vx_info *vxi, void __user *data)
17326 +{
17327 +       struct vcmd_badness_v0 vc_data;
17328 +
17329 +       vc_data.bias = vxi->vx_badness_bias;
17330 +
17331 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
17332 +               return -EFAULT;
17333 +       return 0;
17334 +}
17335 +
17336 +int vc_set_badness(struct vx_info *vxi, void __user *data)
17337 +{
17338 +       struct vcmd_badness_v0 vc_data;
17339 +
17340 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
17341 +               return -EFAULT;
17342 +
17343 +       vxi->vx_badness_bias = vc_data.bias;
17344 +       return 0;
17345 +}
17346 +
17347 +#include <linux/module.h>
17348 +
17349 +EXPORT_SYMBOL_GPL(free_vx_info);
17350 +
17351 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/cvirt.c linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/cvirt.c
17352 --- linux-3.0-rc5/kernel/vserver/cvirt.c        1970-01-01 01:00:00.000000000 +0100
17353 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/cvirt.c   2011-06-13 14:57:26.000000000 +0200
17354 @@ -0,0 +1,303 @@
17355 +/*
17356 + *  linux/kernel/vserver/cvirt.c
17357 + *
17358 + *  Virtual Server: Context Virtualization
17359 + *
17360 + *  Copyright (C) 2004-2007  Herbert Pötzl
17361 + *
17362 + *  V0.01  broken out from limit.c
17363 + *  V0.02  added utsname stuff
17364 + *  V0.03  changed vcmds to vxi arg
17365 + *
17366 + */
17367 +
17368 +#include <linux/types.h>
17369 +#include <linux/utsname.h>
17370 +#include <linux/vs_cvirt.h>
17371 +#include <linux/vserver/switch.h>
17372 +#include <linux/vserver/cvirt_cmd.h>
17373 +
17374 +#include <asm/uaccess.h>
17375 +
17376 +
17377 +void vx_vsi_uptime(struct timespec *uptime, struct timespec *idle)
17378 +{
17379 +       struct vx_info *vxi = current_vx_info();
17380 +
17381 +       set_normalized_timespec(uptime,
17382 +               uptime->tv_sec - vxi->cvirt.bias_uptime.tv_sec,
17383 +               uptime->tv_nsec - vxi->cvirt.bias_uptime.tv_nsec);
17384 +       if (!idle)
17385 +               return;
17386 +       set_normalized_timespec(idle,
17387 +               idle->tv_sec - vxi->cvirt.bias_idle.tv_sec,
17388 +               idle->tv_nsec - vxi->cvirt.bias_idle.tv_nsec);
17389 +       return;
17390 +}
17391 +
17392 +uint64_t vx_idle_jiffies(void)
17393 +{
17394 +       return init_task.utime + init_task.stime;
17395 +}
17396 +
17397 +
17398 +
17399 +static inline uint32_t __update_loadavg(uint32_t load,
17400 +       int wsize, int delta, int n)
17401 +{
17402 +       unsigned long long calc, prev;
17403 +
17404 +       /* just set it to n */
17405 +       if (unlikely(delta >= wsize))
17406 +               return (n << FSHIFT);
17407 +
17408 +       calc = delta * n;
17409 +       calc <<= FSHIFT;
17410 +       prev = (wsize - delta);
17411 +       prev *= load;
17412 +       calc += prev;
17413 +       do_div(calc, wsize);
17414 +       return calc;
17415 +}
17416 +
17417 +
17418 +void vx_update_load(struct vx_info *vxi)
17419 +{
17420 +       uint32_t now, last, delta;
17421 +       unsigned int nr_running, nr_uninterruptible;
17422 +       unsigned int total;
17423 +       unsigned long flags;
17424 +
17425 +       spin_lock_irqsave(&vxi->cvirt.load_lock, flags);
17426 +
17427 +       now = jiffies;
17428 +       last = vxi->cvirt.load_last;
17429 +       delta = now - last;
17430 +
17431 +       if (delta < 5*HZ)
17432 +               goto out;
17433 +
17434 +       nr_running = atomic_read(&vxi->cvirt.nr_running);
17435 +       nr_uninterruptible = atomic_read(&vxi->cvirt.nr_uninterruptible);
17436 +       total = nr_running + nr_uninterruptible;
17437 +
17438 +       vxi->cvirt.load[0] = __update_loadavg(vxi->cvirt.load[0],
17439 +               60*HZ, delta, total);
17440 +       vxi->cvirt.load[1] = __update_loadavg(vxi->cvirt.load[1],
17441 +               5*60*HZ, delta, total);
17442 +       vxi->cvirt.load[2] = __update_loadavg(vxi->cvirt.load[2],
17443 +               15*60*HZ, delta, total);
17444 +
17445 +       vxi->cvirt.load_last = now;
17446 +out:
17447 +       atomic_inc(&vxi->cvirt.load_updates);
17448 +       spin_unlock_irqrestore(&vxi->cvirt.load_lock, flags);
17449 +}
17450 +
17451 +
17452 +/*
17453 + * Commands to do_syslog:
17454 + *
17455 + *      0 -- Close the log.  Currently a NOP.
17456 + *      1 -- Open the log. Currently a NOP.
17457 + *      2 -- Read from the log.
17458 + *      3 -- Read all messages remaining in the ring buffer.
17459 + *      4 -- Read and clear all messages remaining in the ring buffer
17460 + *      5 -- Clear ring buffer.
17461 + *      6 -- Disable printk's to console
17462 + *      7 -- Enable printk's to console
17463 + *      8 -- Set level of messages printed to console
17464 + *      9 -- Return number of unread characters in the log buffer
17465 + *     10 -- Return size of the log buffer
17466 + */
17467 +int vx_do_syslog(int type, char __user *buf, int len)
17468 +{
17469 +       int error = 0;
17470 +       int do_clear = 0;
17471 +       struct vx_info *vxi = current_vx_info();
17472 +       struct _vx_syslog *log;
17473 +
17474 +       if (!vxi)
17475 +               return -EINVAL;
17476 +       log = &vxi->cvirt.syslog;
17477 +
17478 +       switch (type) {
17479 +       case 0:         /* Close log */
17480 +       case 1:         /* Open log */
17481 +               break;
17482 +       case 2:         /* Read from log */
17483 +               error = wait_event_interruptible(log->log_wait,
17484 +                       (log->log_start - log->log_end));
17485 +               if (error)
17486 +                       break;
17487 +               spin_lock_irq(&log->logbuf_lock);
17488 +               spin_unlock_irq(&log->logbuf_lock);
17489 +               break;
17490 +       case 4:         /* Read/clear last kernel messages */
17491 +               do_clear = 1;
17492 +               /* fall through */
17493 +       case 3:         /* Read last kernel messages */
17494 +               return 0;
17495 +
17496 +       case 5:         /* Clear ring buffer */
17497 +               return 0;
17498 +
17499 +       case 6:         /* Disable logging to console */
17500 +       case 7:         /* Enable logging to console */
17501 +       case 8:         /* Set level of messages printed to console */
17502 +               break;
17503 +
17504 +       case 9:         /* Number of chars in the log buffer */
17505 +               return 0;
17506 +       case 10:        /* Size of the log buffer */
17507 +               return 0;
17508 +       default:
17509 +               error = -EINVAL;
17510 +               break;
17511 +       }
17512 +       return error;
17513 +}
17514 +
17515 +
17516 +/* virtual host info names */
17517 +
17518 +static char *vx_vhi_name(struct vx_info *vxi, int id)
17519 +{
17520 +       struct nsproxy *nsproxy;
17521 +       struct uts_namespace *uts;
17522 +
17523 +       if (id == VHIN_CONTEXT)
17524 +               return vxi->vx_name;
17525 +
17526 +       nsproxy = vxi->space[0].vx_nsproxy;
17527 +       if (!nsproxy)
17528 +               return NULL;
17529 +
17530 +       uts = nsproxy->uts_ns;
17531 +       if (!uts)
17532 +               return NULL;
17533 +
17534 +       switch (id) {
17535 +       case VHIN_SYSNAME:
17536 +               return uts->name.sysname;
17537 +       case VHIN_NODENAME:
17538 +               return uts->name.nodename;
17539 +       case VHIN_RELEASE:
17540 +               return uts->name.release;
17541 +       case VHIN_VERSION:
17542 +               return uts->name.version;
17543 +       case VHIN_MACHINE:
17544 +               return uts->name.machine;
17545 +       case VHIN_DOMAINNAME:
17546 +               return uts->name.domainname;
17547 +       default:
17548 +               return NULL;
17549 +       }
17550 +       return NULL;
17551 +}
17552 +
17553 +int vc_set_vhi_name(struct vx_info *vxi, void __user *data)
17554 +{
17555 +       struct vcmd_vhi_name_v0 vc_data;
17556 +       char *name;
17557 +
17558 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
17559 +               return -EFAULT;
17560 +
17561 +       name = vx_vhi_name(vxi, vc_data.field);
17562 +       if (!name)
17563 +               return -EINVAL;
17564 +
17565 +       memcpy(name, vc_data.name, 65);
17566 +       return 0;
17567 +}
17568 +
17569 +int vc_get_vhi_name(struct vx_info *vxi, void __user *data)
17570 +{
17571 +       struct vcmd_vhi_name_v0 vc_data;
17572 +       char *name;
17573 +
17574 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
17575 +               return -EFAULT;
17576 +
17577 +       name = vx_vhi_name(vxi, vc_data.field);
17578 +       if (!name)
17579 +               return -EINVAL;
17580 +
17581 +       memcpy(vc_data.name, name, 65);
17582 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
17583 +               return -EFAULT;
17584 +       return 0;
17585 +}
17586 +
17587 +
17588 +int vc_virt_stat(struct vx_info *vxi, void __user *data)
17589 +{
17590 +       struct vcmd_virt_stat_v0 vc_data;
17591 +       struct _vx_cvirt *cvirt = &vxi->cvirt;
17592 +       struct timespec uptime;
17593 +
17594 +       do_posix_clock_monotonic_gettime(&uptime);
17595 +       set_normalized_timespec(&uptime,
17596 +               uptime.tv_sec - cvirt->bias_uptime.tv_sec,
17597 +               uptime.tv_nsec - cvirt->bias_uptime.tv_nsec);
17598 +
17599 +       vc_data.offset = timespec_to_ns(&cvirt->bias_ts);
17600 +       vc_data.uptime = timespec_to_ns(&uptime);
17601 +       vc_data.nr_threads = atomic_read(&cvirt->nr_threads);
17602 +       vc_data.nr_running = atomic_read(&cvirt->nr_running);
17603 +       vc_data.nr_uninterruptible = atomic_read(&cvirt->nr_uninterruptible);
17604 +       vc_data.nr_onhold = atomic_read(&cvirt->nr_onhold);
17605 +       vc_data.nr_forks = atomic_read(&cvirt->total_forks);
17606 +       vc_data.load[0] = cvirt->load[0];
17607 +       vc_data.load[1] = cvirt->load[1];
17608 +       vc_data.load[2] = cvirt->load[2];
17609 +
17610 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
17611 +               return -EFAULT;
17612 +       return 0;
17613 +}
17614 +
17615 +
17616 +#ifdef CONFIG_VSERVER_VTIME
17617 +
17618 +/* virtualized time base */
17619 +
17620 +void vx_adjust_timespec(struct timespec *ts)
17621 +{
17622 +       struct vx_info *vxi;
17623 +
17624 +       if (!vx_flags(VXF_VIRT_TIME, 0))
17625 +               return;
17626 +
17627 +       vxi = current_vx_info();
17628 +       ts->tv_sec += vxi->cvirt.bias_ts.tv_sec;
17629 +       ts->tv_nsec += vxi->cvirt.bias_ts.tv_nsec;
17630 +
17631 +       if (ts->tv_nsec >= NSEC_PER_SEC) {
17632 +               ts->tv_sec++;
17633 +               ts->tv_nsec -= NSEC_PER_SEC;
17634 +       } else if (ts->tv_nsec < 0) {
17635 +               ts->tv_sec--;
17636 +               ts->tv_nsec += NSEC_PER_SEC;
17637 +       }
17638 +}
17639 +
17640 +int vx_settimeofday(const struct timespec *ts)
17641 +{
17642 +       struct timespec ats, delta;
17643 +       struct vx_info *vxi;
17644 +
17645 +       if (!vx_flags(VXF_VIRT_TIME, 0))
17646 +               return do_settimeofday(ts);
17647 +
17648 +       getnstimeofday(&ats);
17649 +       delta = timespec_sub(*ts, ats);
17650 +
17651 +       vxi = current_vx_info();
17652 +       vxi->cvirt.bias_ts = timespec_add(vxi->cvirt.bias_ts, delta);
17653 +       return 0;
17654 +}
17655 +
17656 +#endif
17657 +
17658 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/cvirt_init.h linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/cvirt_init.h
17659 --- linux-3.0-rc5/kernel/vserver/cvirt_init.h   1970-01-01 01:00:00.000000000 +0100
17660 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/cvirt_init.h      2011-06-10 22:11:24.000000000 +0200
17661 @@ -0,0 +1,70 @@
17662 +
17663 +
17664 +extern uint64_t vx_idle_jiffies(void);
17665 +
17666 +static inline void vx_info_init_cvirt(struct _vx_cvirt *cvirt)
17667 +{
17668 +       uint64_t idle_jiffies = vx_idle_jiffies();
17669 +       uint64_t nsuptime;
17670 +
17671 +       do_posix_clock_monotonic_gettime(&cvirt->bias_uptime);
17672 +       nsuptime = (unsigned long long)cvirt->bias_uptime.tv_sec
17673 +               * NSEC_PER_SEC + cvirt->bias_uptime.tv_nsec;
17674 +       cvirt->bias_clock = nsec_to_clock_t(nsuptime);
17675 +       cvirt->bias_ts.tv_sec = 0;
17676 +       cvirt->bias_ts.tv_nsec = 0;
17677 +
17678 +       jiffies_to_timespec(idle_jiffies, &cvirt->bias_idle);
17679 +       atomic_set(&cvirt->nr_threads, 0);
17680 +       atomic_set(&cvirt->nr_running, 0);
17681 +       atomic_set(&cvirt->nr_uninterruptible, 0);
17682 +       atomic_set(&cvirt->nr_onhold, 0);
17683 +
17684 +       spin_lock_init(&cvirt->load_lock);
17685 +       cvirt->load_last = jiffies;
17686 +       atomic_set(&cvirt->load_updates, 0);
17687 +       cvirt->load[0] = 0;
17688 +       cvirt->load[1] = 0;
17689 +       cvirt->load[2] = 0;
17690 +       atomic_set(&cvirt->total_forks, 0);
17691 +
17692 +       spin_lock_init(&cvirt->syslog.logbuf_lock);
17693 +       init_waitqueue_head(&cvirt->syslog.log_wait);
17694 +       cvirt->syslog.log_start = 0;
17695 +       cvirt->syslog.log_end = 0;
17696 +       cvirt->syslog.con_start = 0;
17697 +       cvirt->syslog.logged_chars = 0;
17698 +}
17699 +
17700 +static inline
17701 +void vx_info_init_cvirt_pc(struct _vx_cvirt_pc *cvirt_pc, int cpu)
17702 +{
17703 +       // cvirt_pc->cpustat = { 0 };
17704 +}
17705 +
17706 +static inline void vx_info_exit_cvirt(struct _vx_cvirt *cvirt)
17707 +{
17708 +#ifdef CONFIG_VSERVER_WARN
17709 +       int value;
17710 +#endif
17711 +       vxwprintk_xid((value = atomic_read(&cvirt->nr_threads)),
17712 +               "!!! cvirt: %p[nr_threads] = %d on exit.",
17713 +               cvirt, value);
17714 +       vxwprintk_xid((value = atomic_read(&cvirt->nr_running)),
17715 +               "!!! cvirt: %p[nr_running] = %d on exit.",
17716 +               cvirt, value);
17717 +       vxwprintk_xid((value = atomic_read(&cvirt->nr_uninterruptible)),
17718 +               "!!! cvirt: %p[nr_uninterruptible] = %d on exit.",
17719 +               cvirt, value);
17720 +       vxwprintk_xid((value = atomic_read(&cvirt->nr_onhold)),
17721 +               "!!! cvirt: %p[nr_onhold] = %d on exit.",
17722 +               cvirt, value);
17723 +       return;
17724 +}
17725 +
17726 +static inline
17727 +void vx_info_exit_cvirt_pc(struct _vx_cvirt_pc *cvirt_pc, int cpu)
17728 +{
17729 +       return;
17730 +}
17731 +
17732 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/cvirt_proc.h linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/cvirt_proc.h
17733 --- linux-3.0-rc5/kernel/vserver/cvirt_proc.h   1970-01-01 01:00:00.000000000 +0100
17734 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/cvirt_proc.h      2011-06-10 22:11:24.000000000 +0200
17735 @@ -0,0 +1,135 @@
17736 +#ifndef _VX_CVIRT_PROC_H
17737 +#define _VX_CVIRT_PROC_H
17738 +
17739 +#include <linux/nsproxy.h>
17740 +#include <linux/mnt_namespace.h>
17741 +#include <linux/ipc_namespace.h>
17742 +#include <linux/utsname.h>
17743 +#include <linux/ipc.h>
17744 +
17745 +
17746 +static inline
17747 +int vx_info_proc_nsproxy(struct nsproxy *nsproxy, char *buffer)
17748 +{
17749 +       struct mnt_namespace *ns;
17750 +       struct uts_namespace *uts;
17751 +       struct ipc_namespace *ipc;
17752 +       struct path path;
17753 +       char *pstr, *root;
17754 +       int length = 0;
17755 +
17756 +       if (!nsproxy)
17757 +               goto out;
17758 +
17759 +       length += sprintf(buffer + length,
17760 +               "NSProxy:\t%p [%p,%p,%p]\n",
17761 +               nsproxy, nsproxy->mnt_ns,
17762 +               nsproxy->uts_ns, nsproxy->ipc_ns);
17763 +
17764 +       ns = nsproxy->mnt_ns;
17765 +       if (!ns)
17766 +               goto skip_ns;
17767 +
17768 +       pstr = kmalloc(PATH_MAX, GFP_KERNEL);
17769 +       if (!pstr)
17770 +               goto skip_ns;
17771 +
17772 +       path.mnt = ns->root;
17773 +       path.dentry = ns->root->mnt_root;
17774 +       root = d_path(&path, pstr, PATH_MAX - 2);
17775 +       length += sprintf(buffer + length,
17776 +               "Namespace:\t%p [#%u]\n"
17777 +               "RootPath:\t%s\n",
17778 +               ns, atomic_read(&ns->count),
17779 +               root);
17780 +       kfree(pstr);
17781 +skip_ns:
17782 +
17783 +       uts = nsproxy->uts_ns;
17784 +       if (!uts)
17785 +               goto skip_uts;
17786 +
17787 +       length += sprintf(buffer + length,
17788 +               "SysName:\t%.*s\n"
17789 +               "NodeName:\t%.*s\n"
17790 +               "Release:\t%.*s\n"
17791 +               "Version:\t%.*s\n"
17792 +               "Machine:\t%.*s\n"
17793 +               "DomainName:\t%.*s\n",
17794 +               __NEW_UTS_LEN, uts->name.sysname,
17795 +               __NEW_UTS_LEN, uts->name.nodename,
17796 +               __NEW_UTS_LEN, uts->name.release,
17797 +               __NEW_UTS_LEN, uts->name.version,
17798 +               __NEW_UTS_LEN, uts->name.machine,
17799 +               __NEW_UTS_LEN, uts->name.domainname);
17800 +skip_uts:
17801 +
17802 +       ipc = nsproxy->ipc_ns;
17803 +       if (!ipc)
17804 +               goto skip_ipc;
17805 +
17806 +       length += sprintf(buffer + length,
17807 +               "SEMS:\t\t%d %d %d %d  %d\n"
17808 +               "MSG:\t\t%d %d %d\n"
17809 +               "SHM:\t\t%lu %lu  %d %d\n",
17810 +               ipc->sem_ctls[0], ipc->sem_ctls[1],
17811 +               ipc->sem_ctls[2], ipc->sem_ctls[3],
17812 +               ipc->used_sems,
17813 +               ipc->msg_ctlmax, ipc->msg_ctlmnb, ipc->msg_ctlmni,
17814 +               (unsigned long)ipc->shm_ctlmax,
17815 +               (unsigned long)ipc->shm_ctlall,
17816 +               ipc->shm_ctlmni, ipc->shm_tot);
17817 +skip_ipc:
17818 +out:
17819 +       return length;
17820 +}
17821 +
17822 +
17823 +#include <linux/sched.h>
17824 +
17825 +#define LOAD_INT(x) ((x) >> FSHIFT)
17826 +#define LOAD_FRAC(x) LOAD_INT(((x) & (FIXED_1 - 1)) * 100)
17827 +
17828 +static inline
17829 +int vx_info_proc_cvirt(struct _vx_cvirt *cvirt, char *buffer)
17830 +{
17831 +       int length = 0;
17832 +       int a, b, c;
17833 +
17834 +       length += sprintf(buffer + length,
17835 +               "BiasUptime:\t%lu.%02lu\n",
17836 +               (unsigned long)cvirt->bias_uptime.tv_sec,
17837 +               (cvirt->bias_uptime.tv_nsec / (NSEC_PER_SEC / 100)));
17838 +
17839 +       a = cvirt->load[0] + (FIXED_1 / 200);
17840 +       b = cvirt->load[1] + (FIXED_1 / 200);
17841 +       c = cvirt->load[2] + (FIXED_1 / 200);
17842 +       length += sprintf(buffer + length,
17843 +               "nr_threads:\t%d\n"
17844 +               "nr_running:\t%d\n"
17845 +               "nr_unintr:\t%d\n"
17846 +               "nr_onhold:\t%d\n"
17847 +               "load_updates:\t%d\n"
17848 +               "loadavg:\t%d.%02d %d.%02d %d.%02d\n"
17849 +               "total_forks:\t%d\n",
17850 +               atomic_read(&cvirt->nr_threads),
17851 +               atomic_read(&cvirt->nr_running),
17852 +               atomic_read(&cvirt->nr_uninterruptible),
17853 +               atomic_read(&cvirt->nr_onhold),
17854 +               atomic_read(&cvirt->load_updates),
17855 +               LOAD_INT(a), LOAD_FRAC(a),
17856 +               LOAD_INT(b), LOAD_FRAC(b),
17857 +               LOAD_INT(c), LOAD_FRAC(c),
17858 +               atomic_read(&cvirt->total_forks));
17859 +       return length;
17860 +}
17861 +
17862 +static inline
17863 +int vx_info_proc_cvirt_pc(struct _vx_cvirt_pc *cvirt_pc,
17864 +       char *buffer, int cpu)
17865 +{
17866 +       int length = 0;
17867 +       return length;
17868 +}
17869 +
17870 +#endif /* _VX_CVIRT_PROC_H */
17871 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/debug.c linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/debug.c
17872 --- linux-3.0-rc5/kernel/vserver/debug.c        1970-01-01 01:00:00.000000000 +0100
17873 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/debug.c   2011-06-10 22:11:24.000000000 +0200
17874 @@ -0,0 +1,32 @@
17875 +/*
17876 + *  kernel/vserver/debug.c
17877 + *
17878 + *  Copyright (C) 2005-2007 Herbert Pötzl
17879 + *
17880 + *  V0.01  vx_info dump support
17881 + *
17882 + */
17883 +
17884 +#include <linux/module.h>
17885 +
17886 +#include <linux/vserver/context.h>
17887 +
17888 +
17889 +void   dump_vx_info(struct vx_info *vxi, int level)
17890 +{
17891 +       printk("vx_info %p[#%d, %d.%d, %4x]\n", vxi, vxi->vx_id,
17892 +               atomic_read(&vxi->vx_usecnt),
17893 +               atomic_read(&vxi->vx_tasks),
17894 +               vxi->vx_state);
17895 +       if (level > 0) {
17896 +               __dump_vx_limit(&vxi->limit);
17897 +               __dump_vx_sched(&vxi->sched);
17898 +               __dump_vx_cvirt(&vxi->cvirt);
17899 +               __dump_vx_cacct(&vxi->cacct);
17900 +       }
17901 +       printk("---\n");
17902 +}
17903 +
17904 +
17905 +EXPORT_SYMBOL_GPL(dump_vx_info);
17906 +
17907 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/device.c linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/device.c
17908 --- linux-3.0-rc5/kernel/vserver/device.c       1970-01-01 01:00:00.000000000 +0100
17909 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/device.c  2011-06-10 23:20:56.000000000 +0200
17910 @@ -0,0 +1,443 @@
17911 +/*
17912 + *  linux/kernel/vserver/device.c
17913 + *
17914 + *  Linux-VServer: Device Support
17915 + *
17916 + *  Copyright (C) 2006  Herbert Pötzl
17917 + *  Copyright (C) 2007  Daniel Hokka Zakrisson
17918 + *
17919 + *  V0.01  device mapping basics
17920 + *  V0.02  added defaults
17921 + *
17922 + */
17923 +
17924 +#include <linux/slab.h>
17925 +#include <linux/rcupdate.h>
17926 +#include <linux/fs.h>
17927 +#include <linux/namei.h>
17928 +#include <linux/hash.h>
17929 +
17930 +#include <asm/errno.h>
17931 +#include <asm/uaccess.h>
17932 +#include <linux/vserver/base.h>
17933 +#include <linux/vserver/debug.h>
17934 +#include <linux/vserver/context.h>
17935 +#include <linux/vserver/device.h>
17936 +#include <linux/vserver/device_cmd.h>
17937 +
17938 +
17939 +#define DMAP_HASH_BITS 4
17940 +
17941 +
17942 +struct vs_mapping {
17943 +       union {
17944 +               struct hlist_node hlist;
17945 +               struct list_head list;
17946 +       } u;
17947 +#define dm_hlist       u.hlist
17948 +#define dm_list                u.list
17949 +       xid_t xid;
17950 +       dev_t device;
17951 +       struct vx_dmap_target target;
17952 +};
17953 +
17954 +
17955 +static struct hlist_head dmap_main_hash[1 << DMAP_HASH_BITS];
17956 +
17957 +static DEFINE_SPINLOCK(dmap_main_hash_lock);
17958 +
17959 +static struct vx_dmap_target dmap_defaults[2] = {
17960 +       { .flags = DATTR_OPEN },
17961 +       { .flags = DATTR_OPEN },
17962 +};
17963 +
17964 +
17965 +struct kmem_cache *dmap_cachep __read_mostly;
17966 +
17967 +int __init dmap_cache_init(void)
17968 +{
17969 +       dmap_cachep = kmem_cache_create("dmap_cache",
17970 +               sizeof(struct vs_mapping), 0,
17971 +               SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
17972 +       return 0;
17973 +}
17974 +
17975 +__initcall(dmap_cache_init);
17976 +
17977 +
17978 +static inline unsigned int __hashval(dev_t dev, int bits)
17979 +{
17980 +       return hash_long((unsigned long)dev, bits);
17981 +}
17982 +
17983 +
17984 +/*     __hash_mapping()
17985 + *     add the mapping to the hash table
17986 + */
17987 +static inline void __hash_mapping(struct vx_info *vxi, struct vs_mapping *vdm)
17988 +{
17989 +       spinlock_t *hash_lock = &dmap_main_hash_lock;
17990 +       struct hlist_head *head, *hash = dmap_main_hash;
17991 +       int device = vdm->device;
17992 +
17993 +       spin_lock(hash_lock);
17994 +       vxdprintk(VXD_CBIT(misc, 8), "__hash_mapping: %p[#%d] %08x:%08x",
17995 +               vxi, vxi ? vxi->vx_id : 0, device, vdm->target.target);
17996 +
17997 +       head = &hash[__hashval(device, DMAP_HASH_BITS)];
17998 +       hlist_add_head(&vdm->dm_hlist, head);
17999 +       spin_unlock(hash_lock);
18000 +}
18001 +
18002 +
18003 +static inline int __mode_to_default(umode_t mode)
18004 +{
18005 +       switch (mode) {
18006 +       case S_IFBLK:
18007 +               return 0;
18008 +       case S_IFCHR:
18009 +               return 1;
18010 +       default:
18011 +               BUG();
18012 +       }
18013 +}
18014 +
18015 +
18016 +/*     __set_default()
18017 + *     set a default
18018 + */
18019 +static inline void __set_default(struct vx_info *vxi, umode_t mode,
18020 +       struct vx_dmap_target *vdmt)
18021 +{
18022 +       spinlock_t *hash_lock = &dmap_main_hash_lock;
18023 +       spin_lock(hash_lock);
18024 +
18025 +       if (vxi)
18026 +               vxi->dmap.targets[__mode_to_default(mode)] = *vdmt;
18027 +       else
18028 +               dmap_defaults[__mode_to_default(mode)] = *vdmt;
18029 +
18030 +
18031 +       spin_unlock(hash_lock);
18032 +
18033 +       vxdprintk(VXD_CBIT(misc, 8), "__set_default: %p[#%u] %08x %04x",
18034 +                 vxi, vxi ? vxi->vx_id : 0, vdmt->target, vdmt->flags);
18035 +}
18036 +
18037 +
18038 +/*     __remove_default()
18039 + *     remove a default
18040 + */
18041 +static inline int __remove_default(struct vx_info *vxi, umode_t mode)
18042 +{
18043 +       spinlock_t *hash_lock = &dmap_main_hash_lock;
18044 +       spin_lock(hash_lock);
18045 +
18046 +       if (vxi)
18047 +               vxi->dmap.targets[__mode_to_default(mode)].flags = 0;
18048 +       else    /* remove == reset */
18049 +               dmap_defaults[__mode_to_default(mode)].flags = DATTR_OPEN | mode;
18050 +
18051 +       spin_unlock(hash_lock);
18052 +       return 0;
18053 +}
18054 +
18055 +
18056 +/*     __find_mapping()
18057 + *     find a mapping in the hash table
18058 + *
18059 + *     caller must hold hash_lock
18060 + */
18061 +static inline int __find_mapping(xid_t xid, dev_t device, umode_t mode,
18062 +       struct vs_mapping **local, struct vs_mapping **global)
18063 +{
18064 +       struct hlist_head *hash = dmap_main_hash;
18065 +       struct hlist_head *head = &hash[__hashval(device, DMAP_HASH_BITS)];
18066 +       struct hlist_node *pos;
18067 +       struct vs_mapping *vdm;
18068 +
18069 +       *local = NULL;
18070 +       if (global)
18071 +               *global = NULL;
18072 +
18073 +       hlist_for_each(pos, head) {
18074 +               vdm = hlist_entry(pos, struct vs_mapping, dm_hlist);
18075 +
18076 +               if ((vdm->device == device) &&
18077 +                       !((vdm->target.flags ^ mode) & S_IFMT)) {
18078 +                       if (vdm->xid == xid) {
18079 +                               *local = vdm;
18080 +                               return 1;
18081 +                       } else if (global && vdm->xid == 0)
18082 +                               *global = vdm;
18083 +               }
18084 +       }
18085 +
18086 +       if (global && *global)
18087 +               return 0;
18088 +       else
18089 +               return -ENOENT;
18090 +}
18091 +
18092 +
18093 +/*     __lookup_mapping()
18094 + *     find a mapping and store the result in target and flags
18095 + */
18096 +static inline int __lookup_mapping(struct vx_info *vxi,
18097 +       dev_t device, dev_t *target, int *flags, umode_t mode)
18098 +{
18099 +       spinlock_t *hash_lock = &dmap_main_hash_lock;
18100 +       struct vs_mapping *vdm, *global;
18101 +       struct vx_dmap_target *vdmt;
18102 +       int ret = 0;
18103 +       xid_t xid = vxi->vx_id;
18104 +       int index;
18105 +
18106 +       spin_lock(hash_lock);
18107 +       if (__find_mapping(xid, device, mode, &vdm, &global) > 0) {
18108 +               ret = 1;
18109 +               vdmt = &vdm->target;
18110 +               goto found;
18111 +       }
18112 +
18113 +       index = __mode_to_default(mode);
18114 +       if (vxi && vxi->dmap.targets[index].flags) {
18115 +               ret = 2;
18116 +               vdmt = &vxi->dmap.targets[index];
18117 +       } else if (global) {
18118 +               ret = 3;
18119 +               vdmt = &global->target;
18120 +               goto found;
18121 +       } else {
18122 +               ret = 4;
18123 +               vdmt = &dmap_defaults[index];
18124 +       }
18125 +
18126 +found:
18127 +       if (target && (vdmt->flags & DATTR_REMAP))
18128 +               *target = vdmt->target;
18129 +       else if (target)
18130 +               *target = device;
18131 +       if (flags)
18132 +               *flags = vdmt->flags;
18133 +
18134 +       spin_unlock(hash_lock);
18135 +
18136 +       return ret;
18137 +}
18138 +
18139 +
18140 +/*     __remove_mapping()
18141 + *     remove a mapping from the hash table
18142 + */
18143 +static inline int __remove_mapping(struct vx_info *vxi, dev_t device,
18144 +       umode_t mode)
18145 +{
18146 +       spinlock_t *hash_lock = &dmap_main_hash_lock;
18147 +       struct vs_mapping *vdm = NULL;
18148 +       int ret = 0;
18149 +
18150 +       spin_lock(hash_lock);
18151 +
18152 +       ret = __find_mapping((vxi ? vxi->vx_id : 0), device, mode, &vdm,
18153 +               NULL);
18154 +       vxdprintk(VXD_CBIT(misc, 8), "__remove_mapping: %p[#%d] %08x %04x",
18155 +               vxi, vxi ? vxi->vx_id : 0, device, mode);
18156 +       if (ret < 0)
18157 +               goto out;
18158 +       hlist_del(&vdm->dm_hlist);
18159 +
18160 +out:
18161 +       spin_unlock(hash_lock);
18162 +       if (vdm)
18163 +               kmem_cache_free(dmap_cachep, vdm);
18164 +       return ret;
18165 +}
18166 +
18167 +
18168 +
18169 +int vs_map_device(struct vx_info *vxi,
18170 +       dev_t device, dev_t *target, umode_t mode)
18171 +{
18172 +       int ret, flags = DATTR_MASK;
18173 +
18174 +       if (!vxi) {
18175 +               if (target)
18176 +                       *target = device;
18177 +               goto out;
18178 +       }
18179 +       ret = __lookup_mapping(vxi, device, target, &flags, mode);
18180 +       vxdprintk(VXD_CBIT(misc, 8), "vs_map_device: %08x target: %08x flags: %04x mode: %04x mapped=%d",
18181 +               device, target ? *target : 0, flags, mode, ret);
18182 +out:
18183 +       return (flags & DATTR_MASK);
18184 +}
18185 +
18186 +
18187 +
18188 +static int do_set_mapping(struct vx_info *vxi,
18189 +       dev_t device, dev_t target, int flags, umode_t mode)
18190 +{
18191 +       if (device) {
18192 +               struct vs_mapping *new;
18193 +
18194 +               new = kmem_cache_alloc(dmap_cachep, GFP_KERNEL);
18195 +               if (!new)
18196 +                       return -ENOMEM;
18197 +
18198 +               INIT_HLIST_NODE(&new->dm_hlist);
18199 +               new->device = device;
18200 +               new->target.target = target;
18201 +               new->target.flags = flags | mode;
18202 +               new->xid = (vxi ? vxi->vx_id : 0);
18203 +
18204 +               vxdprintk(VXD_CBIT(misc, 8), "do_set_mapping: %08x target: %08x flags: %04x", device, target, flags);
18205 +               __hash_mapping(vxi, new);
18206 +       } else {
18207 +               struct vx_dmap_target new = {
18208 +                       .target = target,
18209 +                       .flags = flags | mode,
18210 +               };
18211 +               __set_default(vxi, mode, &new);
18212 +       }
18213 +       return 0;
18214 +}
18215 +
18216 +
18217 +static int do_unset_mapping(struct vx_info *vxi,
18218 +       dev_t device, dev_t target, int flags, umode_t mode)
18219 +{
18220 +       int ret = -EINVAL;
18221 +
18222 +       if (device) {
18223 +               ret = __remove_mapping(vxi, device, mode);
18224 +               if (ret < 0)
18225 +                       goto out;
18226 +       } else {
18227 +               ret = __remove_default(vxi, mode);
18228 +               if (ret < 0)
18229 +                       goto out;
18230 +       }
18231 +
18232 +out:
18233 +       return ret;
18234 +}
18235 +
18236 +
18237 +static inline int __user_device(const char __user *name, dev_t *dev,
18238 +       umode_t *mode)
18239 +{
18240 +       struct nameidata nd;
18241 +       int ret;
18242 +
18243 +       if (!name) {
18244 +               *dev = 0;
18245 +               return 0;
18246 +       }
18247 +       ret = user_lpath(name, &nd.path);
18248 +       if (ret)
18249 +               return ret;
18250 +       if (nd.path.dentry->d_inode) {
18251 +               *dev = nd.path.dentry->d_inode->i_rdev;
18252 +               *mode = nd.path.dentry->d_inode->i_mode;
18253 +       }
18254 +       path_put(&nd.path);
18255 +       return 0;
18256 +}
18257 +
18258 +static inline int __mapping_mode(dev_t device, dev_t target,
18259 +       umode_t device_mode, umode_t target_mode, umode_t *mode)
18260 +{
18261 +       if (device)
18262 +               *mode = device_mode & S_IFMT;
18263 +       else if (target)
18264 +               *mode = target_mode & S_IFMT;
18265 +       else
18266 +               return -EINVAL;
18267 +
18268 +       /* if both given, device and target mode have to match */
18269 +       if (device && target &&
18270 +               ((device_mode ^ target_mode) & S_IFMT))
18271 +               return -EINVAL;
18272 +       return 0;
18273 +}
18274 +
18275 +
18276 +static inline int do_mapping(struct vx_info *vxi, const char __user *device_path,
18277 +       const char __user *target_path, int flags, int set)
18278 +{
18279 +       dev_t device = ~0, target = ~0;
18280 +       umode_t device_mode = 0, target_mode = 0, mode;
18281 +       int ret;
18282 +
18283 +       ret = __user_device(device_path, &device, &device_mode);
18284 +       if (ret)
18285 +               return ret;
18286 +       ret = __user_device(target_path, &target, &target_mode);
18287 +       if (ret)
18288 +               return ret;
18289 +
18290 +       ret = __mapping_mode(device, target,
18291 +               device_mode, target_mode, &mode);
18292 +       if (ret)
18293 +               return ret;
18294 +
18295 +       if (set)
18296 +               return do_set_mapping(vxi, device, target,
18297 +                       flags, mode);
18298 +       else
18299 +               return do_unset_mapping(vxi, device, target,
18300 +                       flags, mode);
18301 +}
18302 +
18303 +
18304 +int vc_set_mapping(struct vx_info *vxi, void __user *data)
18305 +{
18306 +       struct vcmd_set_mapping_v0 vc_data;
18307 +
18308 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18309 +               return -EFAULT;
18310 +
18311 +       return do_mapping(vxi, vc_data.device, vc_data.target,
18312 +               vc_data.flags, 1);
18313 +}
18314 +
18315 +int vc_unset_mapping(struct vx_info *vxi, void __user *data)
18316 +{
18317 +       struct vcmd_set_mapping_v0 vc_data;
18318 +
18319 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18320 +               return -EFAULT;
18321 +
18322 +       return do_mapping(vxi, vc_data.device, vc_data.target,
18323 +               vc_data.flags, 0);
18324 +}
18325 +
18326 +
18327 +#ifdef CONFIG_COMPAT
18328 +
18329 +int vc_set_mapping_x32(struct vx_info *vxi, void __user *data)
18330 +{
18331 +       struct vcmd_set_mapping_v0_x32 vc_data;
18332 +
18333 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18334 +               return -EFAULT;
18335 +
18336 +       return do_mapping(vxi, compat_ptr(vc_data.device_ptr),
18337 +               compat_ptr(vc_data.target_ptr), vc_data.flags, 1);
18338 +}
18339 +
18340 +int vc_unset_mapping_x32(struct vx_info *vxi, void __user *data)
18341 +{
18342 +       struct vcmd_set_mapping_v0_x32 vc_data;
18343 +
18344 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18345 +               return -EFAULT;
18346 +
18347 +       return do_mapping(vxi, compat_ptr(vc_data.device_ptr),
18348 +               compat_ptr(vc_data.target_ptr), vc_data.flags, 0);
18349 +}
18350 +
18351 +#endif /* CONFIG_COMPAT */
18352 +
18353 +
18354 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/dlimit.c linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/dlimit.c
18355 --- linux-3.0-rc5/kernel/vserver/dlimit.c       1970-01-01 01:00:00.000000000 +0100
18356 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/dlimit.c  2011-06-10 23:20:56.000000000 +0200
18357 @@ -0,0 +1,531 @@
18358 +/*
18359 + *  linux/kernel/vserver/dlimit.c
18360 + *
18361 + *  Virtual Server: Context Disk Limits
18362 + *
18363 + *  Copyright (C) 2004-2009  Herbert Pötzl
18364 + *
18365 + *  V0.01  initial version
18366 + *  V0.02  compat32 splitup
18367 + *  V0.03  extended interface
18368 + *
18369 + */
18370 +
18371 +#include <linux/statfs.h>
18372 +#include <linux/sched.h>
18373 +#include <linux/namei.h>
18374 +#include <linux/vs_tag.h>
18375 +#include <linux/vs_dlimit.h>
18376 +#include <linux/vserver/dlimit_cmd.h>
18377 +#include <linux/slab.h>
18378 +// #include <linux/gfp.h>
18379 +
18380 +#include <asm/uaccess.h>
18381 +
18382 +/*     __alloc_dl_info()
18383 +
18384 +       * allocate an initialized dl_info struct
18385 +       * doesn't make it visible (hash)                        */
18386 +
18387 +static struct dl_info *__alloc_dl_info(struct super_block *sb, tag_t tag)
18388 +{
18389 +       struct dl_info *new = NULL;
18390 +
18391 +       vxdprintk(VXD_CBIT(dlim, 5),
18392 +               "alloc_dl_info(%p,%d)*", sb, tag);
18393 +
18394 +       /* would this benefit from a slab cache? */
18395 +       new = kmalloc(sizeof(struct dl_info), GFP_KERNEL);
18396 +       if (!new)
18397 +               return 0;
18398 +
18399 +       memset(new, 0, sizeof(struct dl_info));
18400 +       new->dl_tag = tag;
18401 +       new->dl_sb = sb;
18402 +       // INIT_RCU_HEAD(&new->dl_rcu);
18403 +       INIT_HLIST_NODE(&new->dl_hlist);
18404 +       spin_lock_init(&new->dl_lock);
18405 +       atomic_set(&new->dl_refcnt, 0);
18406 +       atomic_set(&new->dl_usecnt, 0);
18407 +
18408 +       /* rest of init goes here */
18409 +
18410 +       vxdprintk(VXD_CBIT(dlim, 4),
18411 +               "alloc_dl_info(%p,%d) = %p", sb, tag, new);
18412 +       return new;
18413 +}
18414 +
18415 +/*     __dealloc_dl_info()
18416 +
18417 +       * final disposal of dl_info                             */
18418 +
18419 +static void __dealloc_dl_info(struct dl_info *dli)
18420 +{
18421 +       vxdprintk(VXD_CBIT(dlim, 4),
18422 +               "dealloc_dl_info(%p)", dli);
18423 +
18424 +       dli->dl_hlist.next = LIST_POISON1;
18425 +       dli->dl_tag = -1;
18426 +       dli->dl_sb = 0;
18427 +
18428 +       BUG_ON(atomic_read(&dli->dl_usecnt));
18429 +       BUG_ON(atomic_read(&dli->dl_refcnt));
18430 +
18431 +       kfree(dli);
18432 +}
18433 +
18434 +
18435 +/*     hash table for dl_info hash */
18436 +
18437 +#define DL_HASH_SIZE   13
18438 +
18439 +struct hlist_head dl_info_hash[DL_HASH_SIZE];
18440 +
18441 +static DEFINE_SPINLOCK(dl_info_hash_lock);
18442 +
18443 +
18444 +static inline unsigned int __hashval(struct super_block *sb, tag_t tag)
18445 +{
18446 +       return ((tag ^ (unsigned long)sb) % DL_HASH_SIZE);
18447 +}
18448 +
18449 +
18450 +
18451 +/*     __hash_dl_info()
18452 +
18453 +       * add the dli to the global hash table
18454 +       * requires the hash_lock to be held                     */
18455 +
18456 +static inline void __hash_dl_info(struct dl_info *dli)
18457 +{
18458 +       struct hlist_head *head;
18459 +
18460 +       vxdprintk(VXD_CBIT(dlim, 6),
18461 +               "__hash_dl_info: %p[#%d]", dli, dli->dl_tag);
18462 +       get_dl_info(dli);
18463 +       head = &dl_info_hash[__hashval(dli->dl_sb, dli->dl_tag)];
18464 +       hlist_add_head_rcu(&dli->dl_hlist, head);
18465 +}
18466 +
18467 +/*     __unhash_dl_info()
18468 +
18469 +       * remove the dli from the global hash table
18470 +       * requires the hash_lock to be held                     */
18471 +
18472 +static inline void __unhash_dl_info(struct dl_info *dli)
18473 +{
18474 +       vxdprintk(VXD_CBIT(dlim, 6),
18475 +               "__unhash_dl_info: %p[#%d]", dli, dli->dl_tag);
18476 +       hlist_del_rcu(&dli->dl_hlist);
18477 +       put_dl_info(dli);
18478 +}
18479 +
18480 +
18481 +/*     __lookup_dl_info()
18482 +
18483 +       * requires the rcu_read_lock()
18484 +       * doesn't increment the dl_refcnt                       */
18485 +
18486 +static inline struct dl_info *__lookup_dl_info(struct super_block *sb, tag_t tag)
18487 +{
18488 +       struct hlist_head *head = &dl_info_hash[__hashval(sb, tag)];
18489 +       struct hlist_node *pos;
18490 +       struct dl_info *dli;
18491 +
18492 +       hlist_for_each_entry_rcu(dli, pos, head, dl_hlist) {
18493 +
18494 +               if (dli->dl_tag == tag && dli->dl_sb == sb) {
18495 +                       return dli;
18496 +               }
18497 +       }
18498 +       return NULL;
18499 +}
18500 +
18501 +
18502 +struct dl_info *locate_dl_info(struct super_block *sb, tag_t tag)
18503 +{
18504 +       struct dl_info *dli;
18505 +
18506 +       rcu_read_lock();
18507 +       dli = get_dl_info(__lookup_dl_info(sb, tag));
18508 +       vxdprintk(VXD_CBIT(dlim, 7),
18509 +               "locate_dl_info(%p,#%d) = %p", sb, tag, dli);
18510 +       rcu_read_unlock();
18511 +       return dli;
18512 +}
18513 +
18514 +void rcu_free_dl_info(struct rcu_head *head)
18515 +{
18516 +       struct dl_info *dli = container_of(head, struct dl_info, dl_rcu);
18517 +       int usecnt, refcnt;
18518 +
18519 +       BUG_ON(!dli || !head);
18520 +
18521 +       usecnt = atomic_read(&dli->dl_usecnt);
18522 +       BUG_ON(usecnt < 0);
18523 +
18524 +       refcnt = atomic_read(&dli->dl_refcnt);
18525 +       BUG_ON(refcnt < 0);
18526 +
18527 +       vxdprintk(VXD_CBIT(dlim, 3),
18528 +               "rcu_free_dl_info(%p)", dli);
18529 +       if (!usecnt)
18530 +               __dealloc_dl_info(dli);
18531 +       else
18532 +               printk("!!! rcu didn't free\n");
18533 +}
18534 +
18535 +
18536 +
18537 +
18538 +static int do_addrem_dlimit(uint32_t id, const char __user *name,
18539 +       uint32_t flags, int add)
18540 +{
18541 +       struct path path;
18542 +       int ret;
18543 +
18544 +       ret = user_lpath(name, &path);
18545 +       if (!ret) {
18546 +               struct super_block *sb;
18547 +               struct dl_info *dli;
18548 +
18549 +               ret = -EINVAL;
18550 +               if (!path.dentry->d_inode)
18551 +                       goto out_release;
18552 +               if (!(sb = path.dentry->d_inode->i_sb))
18553 +                       goto out_release;
18554 +
18555 +               if (add) {
18556 +                       dli = __alloc_dl_info(sb, id);
18557 +                       spin_lock(&dl_info_hash_lock);
18558 +
18559 +                       ret = -EEXIST;
18560 +                       if (__lookup_dl_info(sb, id))
18561 +                               goto out_unlock;
18562 +                       __hash_dl_info(dli);
18563 +                       dli = NULL;
18564 +               } else {
18565 +                       spin_lock(&dl_info_hash_lock);
18566 +                       dli = __lookup_dl_info(sb, id);
18567 +
18568 +                       ret = -ESRCH;
18569 +                       if (!dli)
18570 +                               goto out_unlock;
18571 +                       __unhash_dl_info(dli);
18572 +               }
18573 +               ret = 0;
18574 +       out_unlock:
18575 +               spin_unlock(&dl_info_hash_lock);
18576 +               if (add && dli)
18577 +                       __dealloc_dl_info(dli);
18578 +       out_release:
18579 +               path_put(&path);
18580 +       }
18581 +       return ret;
18582 +}
18583 +
18584 +int vc_add_dlimit(uint32_t id, void __user *data)
18585 +{
18586 +       struct vcmd_ctx_dlimit_base_v0 vc_data;
18587 +
18588 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18589 +               return -EFAULT;
18590 +
18591 +       return do_addrem_dlimit(id, vc_data.name, vc_data.flags, 1);
18592 +}
18593 +
18594 +int vc_rem_dlimit(uint32_t id, void __user *data)
18595 +{
18596 +       struct vcmd_ctx_dlimit_base_v0 vc_data;
18597 +
18598 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18599 +               return -EFAULT;
18600 +
18601 +       return do_addrem_dlimit(id, vc_data.name, vc_data.flags, 0);
18602 +}
18603 +
18604 +#ifdef CONFIG_COMPAT
18605 +
18606 +int vc_add_dlimit_x32(uint32_t id, void __user *data)
18607 +{
18608 +       struct vcmd_ctx_dlimit_base_v0_x32 vc_data;
18609 +
18610 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18611 +               return -EFAULT;
18612 +
18613 +       return do_addrem_dlimit(id,
18614 +               compat_ptr(vc_data.name_ptr), vc_data.flags, 1);
18615 +}
18616 +
18617 +int vc_rem_dlimit_x32(uint32_t id, void __user *data)
18618 +{
18619 +       struct vcmd_ctx_dlimit_base_v0_x32 vc_data;
18620 +
18621 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18622 +               return -EFAULT;
18623 +
18624 +       return do_addrem_dlimit(id,
18625 +               compat_ptr(vc_data.name_ptr), vc_data.flags, 0);
18626 +}
18627 +
18628 +#endif /* CONFIG_COMPAT */
18629 +
18630 +
18631 +static inline
18632 +int do_set_dlimit(uint32_t id, const char __user *name,
18633 +       uint32_t space_used, uint32_t space_total,
18634 +       uint32_t inodes_used, uint32_t inodes_total,
18635 +       uint32_t reserved, uint32_t flags)
18636 +{
18637 +       struct path path;
18638 +       int ret;
18639 +
18640 +       ret = user_lpath(name, &path);
18641 +       if (!ret) {
18642 +               struct super_block *sb;
18643 +               struct dl_info *dli;
18644 +
18645 +               ret = -EINVAL;
18646 +               if (!path.dentry->d_inode)
18647 +                       goto out_release;
18648 +               if (!(sb = path.dentry->d_inode->i_sb))
18649 +                       goto out_release;
18650 +
18651 +               /* sanity checks */
18652 +               if ((reserved != CDLIM_KEEP &&
18653 +                       reserved > 100) ||
18654 +                       (inodes_used != CDLIM_KEEP &&
18655 +                       inodes_used > inodes_total) ||
18656 +                       (space_used != CDLIM_KEEP &&
18657 +                       space_used > space_total))
18658 +                       goto out_release;
18659 +
18660 +               ret = -ESRCH;
18661 +               dli = locate_dl_info(sb, id);
18662 +               if (!dli)
18663 +                       goto out_release;
18664 +
18665 +               spin_lock(&dli->dl_lock);
18666 +
18667 +               if (inodes_used != CDLIM_KEEP)
18668 +                       dli->dl_inodes_used = inodes_used;
18669 +               if (inodes_total != CDLIM_KEEP)
18670 +                       dli->dl_inodes_total = inodes_total;
18671 +               if (space_used != CDLIM_KEEP)
18672 +                       dli->dl_space_used = dlimit_space_32to64(
18673 +                               space_used, flags, DLIMS_USED);
18674 +
18675 +               if (space_total == CDLIM_INFINITY)
18676 +                       dli->dl_space_total = DLIM_INFINITY;
18677 +               else if (space_total != CDLIM_KEEP)
18678 +                       dli->dl_space_total = dlimit_space_32to64(
18679 +                               space_total, flags, DLIMS_TOTAL);
18680 +
18681 +               if (reserved != CDLIM_KEEP)
18682 +                       dli->dl_nrlmult = (1 << 10) * (100 - reserved) / 100;
18683 +
18684 +               spin_unlock(&dli->dl_lock);
18685 +
18686 +               put_dl_info(dli);
18687 +               ret = 0;
18688 +
18689 +       out_release:
18690 +               path_put(&path);
18691 +       }
18692 +       return ret;
18693 +}
18694 +
18695 +int vc_set_dlimit(uint32_t id, void __user *data)
18696 +{
18697 +       struct vcmd_ctx_dlimit_v0 vc_data;
18698 +
18699 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18700 +               return -EFAULT;
18701 +
18702 +       return do_set_dlimit(id, vc_data.name,
18703 +               vc_data.space_used, vc_data.space_total,
18704 +               vc_data.inodes_used, vc_data.inodes_total,
18705 +               vc_data.reserved, vc_data.flags);
18706 +}
18707 +
18708 +#ifdef CONFIG_COMPAT
18709 +
18710 +int vc_set_dlimit_x32(uint32_t id, void __user *data)
18711 +{
18712 +       struct vcmd_ctx_dlimit_v0_x32 vc_data;
18713 +
18714 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18715 +               return -EFAULT;
18716 +
18717 +       return do_set_dlimit(id, compat_ptr(vc_data.name_ptr),
18718 +               vc_data.space_used, vc_data.space_total,
18719 +               vc_data.inodes_used, vc_data.inodes_total,
18720 +               vc_data.reserved, vc_data.flags);
18721 +}
18722 +
18723 +#endif /* CONFIG_COMPAT */
18724 +
18725 +
18726 +static inline
18727 +int do_get_dlimit(uint32_t id, const char __user *name,
18728 +       uint32_t *space_used, uint32_t *space_total,
18729 +       uint32_t *inodes_used, uint32_t *inodes_total,
18730 +       uint32_t *reserved, uint32_t *flags)
18731 +{
18732 +       struct path path;
18733 +       int ret;
18734 +
18735 +       ret = user_lpath(name, &path);
18736 +       if (!ret) {
18737 +               struct super_block *sb;
18738 +               struct dl_info *dli;
18739 +
18740 +               ret = -EINVAL;
18741 +               if (!path.dentry->d_inode)
18742 +                       goto out_release;
18743 +               if (!(sb = path.dentry->d_inode->i_sb))
18744 +                       goto out_release;
18745 +
18746 +               ret = -ESRCH;
18747 +               dli = locate_dl_info(sb, id);
18748 +               if (!dli)
18749 +                       goto out_release;
18750 +
18751 +               spin_lock(&dli->dl_lock);
18752 +               *inodes_used = dli->dl_inodes_used;
18753 +               *inodes_total = dli->dl_inodes_total;
18754 +
18755 +               *space_used = dlimit_space_64to32(
18756 +                       dli->dl_space_used, flags, DLIMS_USED);
18757 +
18758 +               if (dli->dl_space_total == DLIM_INFINITY)
18759 +                       *space_total = CDLIM_INFINITY;
18760 +               else
18761 +                       *space_total = dlimit_space_64to32(
18762 +                               dli->dl_space_total, flags, DLIMS_TOTAL);
18763 +
18764 +               *reserved = 100 - ((dli->dl_nrlmult * 100 + 512) >> 10);
18765 +               spin_unlock(&dli->dl_lock);
18766 +
18767 +               put_dl_info(dli);
18768 +               ret = -EFAULT;
18769 +
18770 +               ret = 0;
18771 +       out_release:
18772 +               path_put(&path);
18773 +       }
18774 +       return ret;
18775 +}
18776 +
18777 +
18778 +int vc_get_dlimit(uint32_t id, void __user *data)
18779 +{
18780 +       struct vcmd_ctx_dlimit_v0 vc_data;
18781 +       int ret;
18782 +
18783 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18784 +               return -EFAULT;
18785 +
18786 +       ret = do_get_dlimit(id, vc_data.name,
18787 +               &vc_data.space_used, &vc_data.space_total,
18788 +               &vc_data.inodes_used, &vc_data.inodes_total,
18789 +               &vc_data.reserved, &vc_data.flags);
18790 +       if (ret)
18791 +               return ret;
18792 +
18793 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
18794 +               return -EFAULT;
18795 +       return 0;
18796 +}
18797 +
18798 +#ifdef CONFIG_COMPAT
18799 +
18800 +int vc_get_dlimit_x32(uint32_t id, void __user *data)
18801 +{
18802 +       struct vcmd_ctx_dlimit_v0_x32 vc_data;
18803 +       int ret;
18804 +
18805 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18806 +               return -EFAULT;
18807 +
18808 +       ret = do_get_dlimit(id, compat_ptr(vc_data.name_ptr),
18809 +               &vc_data.space_used, &vc_data.space_total,
18810 +               &vc_data.inodes_used, &vc_data.inodes_total,
18811 +               &vc_data.reserved, &vc_data.flags);
18812 +       if (ret)
18813 +               return ret;
18814 +
18815 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
18816 +               return -EFAULT;
18817 +       return 0;
18818 +}
18819 +
18820 +#endif /* CONFIG_COMPAT */
18821 +
18822 +
18823 +void vx_vsi_statfs(struct super_block *sb, struct kstatfs *buf)
18824 +{
18825 +       struct dl_info *dli;
18826 +       __u64 blimit, bfree, bavail;
18827 +       __u32 ifree;
18828 +
18829 +       dli = locate_dl_info(sb, dx_current_tag());
18830 +       if (!dli)
18831 +               return;
18832 +
18833 +       spin_lock(&dli->dl_lock);
18834 +       if (dli->dl_inodes_total == (unsigned long)DLIM_INFINITY)
18835 +               goto no_ilim;
18836 +
18837 +       /* reduce max inodes available to limit */
18838 +       if (buf->f_files > dli->dl_inodes_total)
18839 +               buf->f_files = dli->dl_inodes_total;
18840 +
18841 +       ifree = dli->dl_inodes_total - dli->dl_inodes_used;
18842 +       /* reduce free inodes to min */
18843 +       if (ifree < buf->f_ffree)
18844 +               buf->f_ffree = ifree;
18845 +
18846 +no_ilim:
18847 +       if (dli->dl_space_total == DLIM_INFINITY)
18848 +               goto no_blim;
18849 +
18850 +       blimit = dli->dl_space_total >> sb->s_blocksize_bits;
18851 +
18852 +       if (dli->dl_space_total < dli->dl_space_used)
18853 +               bfree = 0;
18854 +       else
18855 +               bfree = (dli->dl_space_total - dli->dl_space_used)
18856 +                       >> sb->s_blocksize_bits;
18857 +
18858 +       bavail = ((dli->dl_space_total >> 10) * dli->dl_nrlmult);
18859 +       if (bavail < dli->dl_space_used)
18860 +               bavail = 0;
18861 +       else
18862 +               bavail = (bavail - dli->dl_space_used)
18863 +                       >> sb->s_blocksize_bits;
18864 +
18865 +       /* reduce max space available to limit */
18866 +       if (buf->f_blocks > blimit)
18867 +               buf->f_blocks = blimit;
18868 +
18869 +       /* reduce free space to min */
18870 +       if (bfree < buf->f_bfree)
18871 +               buf->f_bfree = bfree;
18872 +
18873 +       /* reduce avail space to min */
18874 +       if (bavail < buf->f_bavail)
18875 +               buf->f_bavail = bavail;
18876 +
18877 +no_blim:
18878 +       spin_unlock(&dli->dl_lock);
18879 +       put_dl_info(dli);
18880 +
18881 +       return;
18882 +}
18883 +
18884 +#include <linux/module.h>
18885 +
18886 +EXPORT_SYMBOL_GPL(locate_dl_info);
18887 +EXPORT_SYMBOL_GPL(rcu_free_dl_info);
18888 +
18889 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/helper.c linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/helper.c
18890 --- linux-3.0-rc5/kernel/vserver/helper.c       1970-01-01 01:00:00.000000000 +0100
18891 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/helper.c  2011-06-10 22:11:24.000000000 +0200
18892 @@ -0,0 +1,223 @@
18893 +/*
18894 + *  linux/kernel/vserver/helper.c
18895 + *
18896 + *  Virtual Context Support
18897 + *
18898 + *  Copyright (C) 2004-2007  Herbert Pötzl
18899 + *
18900 + *  V0.01  basic helper
18901 + *
18902 + */
18903 +
18904 +#include <linux/kmod.h>
18905 +#include <linux/reboot.h>
18906 +#include <linux/vs_context.h>
18907 +#include <linux/vs_network.h>
18908 +#include <linux/vserver/signal.h>
18909 +
18910 +
18911 +char vshelper_path[255] = "/sbin/vshelper";
18912 +
18913 +
18914 +static int do_vshelper(char *name, char *argv[], char *envp[], int sync)
18915 +{
18916 +       int ret;
18917 +
18918 +       if ((ret = call_usermodehelper(name, argv, envp, sync))) {
18919 +               printk( KERN_WARNING
18920 +                       "%s: (%s %s) returned %s with %d\n",
18921 +                       name, argv[1], argv[2],
18922 +                       sync ? "sync" : "async", ret);
18923 +       }
18924 +       vxdprintk(VXD_CBIT(switch, 4),
18925 +               "%s: (%s %s) returned %s with %d",
18926 +               name, argv[1], argv[2], sync ? "sync" : "async", ret);
18927 +       return ret;
18928 +}
18929 +
18930 +/*
18931 + *      vshelper path is set via /proc/sys
18932 + *      invoked by vserver sys_reboot(), with
18933 + *      the following arguments
18934 + *
18935 + *      argv [0] = vshelper_path;
18936 + *      argv [1] = action: "restart", "halt", "poweroff", ...
18937 + *      argv [2] = context identifier
18938 + *
18939 + *      envp [*] = type-specific parameters
18940 + */
18941 +
18942 +long vs_reboot_helper(struct vx_info *vxi, int cmd, void __user *arg)
18943 +{
18944 +       char id_buf[8], cmd_buf[16];
18945 +       char uid_buf[16], pid_buf[16];
18946 +       int ret;
18947 +
18948 +       char *argv[] = {vshelper_path, NULL, id_buf, 0};
18949 +       char *envp[] = {"HOME=/", "TERM=linux",
18950 +                       "PATH=/sbin:/usr/sbin:/bin:/usr/bin",
18951 +                       uid_buf, pid_buf, cmd_buf, 0};
18952 +
18953 +       if (vx_info_state(vxi, VXS_HELPER))
18954 +               return -EAGAIN;
18955 +       vxi->vx_state |= VXS_HELPER;
18956 +
18957 +       snprintf(id_buf, sizeof(id_buf)-1, "%d", vxi->vx_id);
18958 +
18959 +       snprintf(cmd_buf, sizeof(cmd_buf)-1, "VS_CMD=%08x", cmd);
18960 +       snprintf(uid_buf, sizeof(uid_buf)-1, "VS_UID=%d", current_uid());
18961 +       snprintf(pid_buf, sizeof(pid_buf)-1, "VS_PID=%d", current->pid);
18962 +
18963 +       switch (cmd) {
18964 +       case LINUX_REBOOT_CMD_RESTART:
18965 +               argv[1] = "restart";
18966 +               break;
18967 +
18968 +       case LINUX_REBOOT_CMD_HALT:
18969 +               argv[1] = "halt";
18970 +               break;
18971 +
18972 +       case LINUX_REBOOT_CMD_POWER_OFF:
18973 +               argv[1] = "poweroff";
18974 +               break;
18975 +
18976 +       case LINUX_REBOOT_CMD_SW_SUSPEND:
18977 +               argv[1] = "swsusp";
18978 +               break;
18979 +
18980 +       case LINUX_REBOOT_CMD_OOM:
18981 +               argv[1] = "oom";
18982 +               break;
18983 +
18984 +       default:
18985 +               vxi->vx_state &= ~VXS_HELPER;
18986 +               return 0;
18987 +       }
18988 +
18989 +       ret = do_vshelper(vshelper_path, argv, envp, 0);
18990 +       vxi->vx_state &= ~VXS_HELPER;
18991 +       __wakeup_vx_info(vxi);
18992 +       return (ret) ? -EPERM : 0;
18993 +}
18994 +
18995 +
18996 +long vs_reboot(unsigned int cmd, void __user *arg)
18997 +{
18998 +       struct vx_info *vxi = current_vx_info();
18999 +       long ret = 0;
19000 +
19001 +       vxdprintk(VXD_CBIT(misc, 5),
19002 +               "vs_reboot(%p[#%d],%u)",
19003 +               vxi, vxi ? vxi->vx_id : 0, cmd);
19004 +
19005 +       ret = vs_reboot_helper(vxi, cmd, arg);
19006 +       if (ret)
19007 +               return ret;
19008 +
19009 +       vxi->reboot_cmd = cmd;
19010 +       if (vx_info_flags(vxi, VXF_REBOOT_KILL, 0)) {
19011 +               switch (cmd) {
19012 +               case LINUX_REBOOT_CMD_RESTART:
19013 +               case LINUX_REBOOT_CMD_HALT:
19014 +               case LINUX_REBOOT_CMD_POWER_OFF:
19015 +                       vx_info_kill(vxi, 0, SIGKILL);
19016 +                       vx_info_kill(vxi, 1, SIGKILL);
19017 +               default:
19018 +                       break;
19019 +               }
19020 +       }
19021 +       return 0;
19022 +}
19023 +
19024 +long vs_oom_action(unsigned int cmd)
19025 +{
19026 +       struct vx_info *vxi = current_vx_info();
19027 +       long ret = 0;
19028 +
19029 +       vxdprintk(VXD_CBIT(misc, 5),
19030 +               "vs_oom_action(%p[#%d],%u)",
19031 +               vxi, vxi ? vxi->vx_id : 0, cmd);
19032 +
19033 +       ret = vs_reboot_helper(vxi, cmd, NULL);
19034 +       if (ret)
19035 +               return ret;
19036 +
19037 +       vxi->reboot_cmd = cmd;
19038 +       if (vx_info_flags(vxi, VXF_REBOOT_KILL, 0)) {
19039 +               vx_info_kill(vxi, 0, SIGKILL);
19040 +               vx_info_kill(vxi, 1, SIGKILL);
19041 +       }
19042 +       return 0;
19043 +}
19044 +
19045 +/*
19046 + *      argv [0] = vshelper_path;
19047 + *      argv [1] = action: "startup", "shutdown"
19048 + *      argv [2] = context identifier
19049 + *
19050 + *      envp [*] = type-specific parameters
19051 + */
19052 +
19053 +long vs_state_change(struct vx_info *vxi, unsigned int cmd)
19054 +{
19055 +       char id_buf[8], cmd_buf[16];
19056 +       char *argv[] = {vshelper_path, NULL, id_buf, 0};
19057 +       char *envp[] = {"HOME=/", "TERM=linux",
19058 +                       "PATH=/sbin:/usr/sbin:/bin:/usr/bin", cmd_buf, 0};
19059 +
19060 +       if (!vx_info_flags(vxi, VXF_SC_HELPER, 0))
19061 +               return 0;
19062 +
19063 +       snprintf(id_buf, sizeof(id_buf)-1, "%d", vxi->vx_id);
19064 +       snprintf(cmd_buf, sizeof(cmd_buf)-1, "VS_CMD=%08x", cmd);
19065 +
19066 +       switch (cmd) {
19067 +       case VSC_STARTUP:
19068 +               argv[1] = "startup";
19069 +               break;
19070 +       case VSC_SHUTDOWN:
19071 +               argv[1] = "shutdown";
19072 +               break;
19073 +       default:
19074 +               return 0;
19075 +       }
19076 +
19077 +       return do_vshelper(vshelper_path, argv, envp, 1);
19078 +}
19079 +
19080 +
19081 +/*
19082 + *      argv [0] = vshelper_path;
19083 + *      argv [1] = action: "netup", "netdown"
19084 + *      argv [2] = context identifier
19085 + *
19086 + *      envp [*] = type-specific parameters
19087 + */
19088 +
19089 +long vs_net_change(struct nx_info *nxi, unsigned int cmd)
19090 +{
19091 +       char id_buf[8], cmd_buf[16];
19092 +       char *argv[] = {vshelper_path, NULL, id_buf, 0};
19093 +       char *envp[] = {"HOME=/", "TERM=linux",
19094 +                       "PATH=/sbin:/usr/sbin:/bin:/usr/bin", cmd_buf, 0};
19095 +
19096 +       if (!nx_info_flags(nxi, NXF_SC_HELPER, 0))
19097 +               return 0;
19098 +
19099 +       snprintf(id_buf, sizeof(id_buf)-1, "%d", nxi->nx_id);
19100 +       snprintf(cmd_buf, sizeof(cmd_buf)-1, "VS_CMD=%08x", cmd);
19101 +
19102 +       switch (cmd) {
19103 +       case VSC_NETUP:
19104 +               argv[1] = "netup";
19105 +               break;
19106 +       case VSC_NETDOWN:
19107 +               argv[1] = "netdown";
19108 +               break;
19109 +       default:
19110 +               return 0;
19111 +       }
19112 +
19113 +       return do_vshelper(vshelper_path, argv, envp, 1);
19114 +}
19115 +
19116 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/history.c linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/history.c
19117 --- linux-3.0-rc5/kernel/vserver/history.c      1970-01-01 01:00:00.000000000 +0100
19118 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/history.c 2011-06-10 22:11:24.000000000 +0200
19119 @@ -0,0 +1,258 @@
19120 +/*
19121 + *  kernel/vserver/history.c
19122 + *
19123 + *  Virtual Context History Backtrace
19124 + *
19125 + *  Copyright (C) 2004-2007  Herbert Pötzl
19126 + *
19127 + *  V0.01  basic structure
19128 + *  V0.02  hash/unhash and trace
19129 + *  V0.03  preemption fixes
19130 + *
19131 + */
19132 +
19133 +#include <linux/module.h>
19134 +#include <asm/uaccess.h>
19135 +
19136 +#include <linux/vserver/context.h>
19137 +#include <linux/vserver/debug.h>
19138 +#include <linux/vserver/debug_cmd.h>
19139 +#include <linux/vserver/history.h>
19140 +
19141 +
19142 +#ifdef CONFIG_VSERVER_HISTORY
19143 +#define VXH_SIZE       CONFIG_VSERVER_HISTORY_SIZE
19144 +#else
19145 +#define VXH_SIZE       64
19146 +#endif
19147 +
19148 +struct _vx_history {
19149 +       unsigned int counter;
19150 +
19151 +       struct _vx_hist_entry entry[VXH_SIZE + 1];
19152 +};
19153 +
19154 +
19155 +DEFINE_PER_CPU(struct _vx_history, vx_history_buffer);
19156 +
19157 +unsigned volatile int vxh_active = 1;
19158 +
19159 +static atomic_t sequence = ATOMIC_INIT(0);
19160 +
19161 +
19162 +/*     vxh_advance()
19163 +
19164 +       * requires disabled preemption                          */
19165 +
19166 +struct _vx_hist_entry *vxh_advance(void *loc)
19167 +{
19168 +       unsigned int cpu = smp_processor_id();
19169 +       struct _vx_history *hist = &per_cpu(vx_history_buffer, cpu);
19170 +       struct _vx_hist_entry *entry;
19171 +       unsigned int index;
19172 +
19173 +       index = vxh_active ? (hist->counter++ % VXH_SIZE) : VXH_SIZE;
19174 +       entry = &hist->entry[index];
19175 +
19176 +       entry->seq = atomic_inc_return(&sequence);
19177 +       entry->loc = loc;
19178 +       return entry;
19179 +}
19180 +
19181 +EXPORT_SYMBOL_GPL(vxh_advance);
19182 +
19183 +
19184 +#define VXH_LOC_FMTS   "(#%04x,*%d):%p"
19185 +
19186 +#define VXH_LOC_ARGS(e)        (e)->seq, cpu, (e)->loc
19187 +
19188 +
19189 +#define VXH_VXI_FMTS   "%p[#%d,%d.%d]"
19190 +
19191 +#define VXH_VXI_ARGS(e)        (e)->vxi.ptr,                           \
19192 +                       (e)->vxi.ptr ? (e)->vxi.xid : 0,        \
19193 +                       (e)->vxi.ptr ? (e)->vxi.usecnt : 0,     \
19194 +                       (e)->vxi.ptr ? (e)->vxi.tasks : 0
19195 +
19196 +void   vxh_dump_entry(struct _vx_hist_entry *e, unsigned cpu)
19197 +{
19198 +       switch (e->type) {
19199 +       case VXH_THROW_OOPS:
19200 +               printk( VXH_LOC_FMTS " oops \n", VXH_LOC_ARGS(e));
19201 +               break;
19202 +
19203 +       case VXH_GET_VX_INFO:
19204 +       case VXH_PUT_VX_INFO:
19205 +               printk( VXH_LOC_FMTS " %s_vx_info " VXH_VXI_FMTS "\n",
19206 +                       VXH_LOC_ARGS(e),
19207 +                       (e->type == VXH_GET_VX_INFO) ? "get" : "put",
19208 +                       VXH_VXI_ARGS(e));
19209 +               break;
19210 +
19211 +       case VXH_INIT_VX_INFO:
19212 +       case VXH_SET_VX_INFO:
19213 +       case VXH_CLR_VX_INFO:
19214 +               printk( VXH_LOC_FMTS " %s_vx_info " VXH_VXI_FMTS " @%p\n",
19215 +                       VXH_LOC_ARGS(e),
19216 +                       (e->type == VXH_INIT_VX_INFO) ? "init" :
19217 +                       ((e->type == VXH_SET_VX_INFO) ? "set" : "clr"),
19218 +                       VXH_VXI_ARGS(e), e->sc.data);
19219 +               break;
19220 +
19221 +       case VXH_CLAIM_VX_INFO:
19222 +       case VXH_RELEASE_VX_INFO:
19223 +               printk( VXH_LOC_FMTS " %s_vx_info " VXH_VXI_FMTS " @%p\n",
19224 +                       VXH_LOC_ARGS(e),
19225 +                       (e->type == VXH_CLAIM_VX_INFO) ? "claim" : "release",
19226 +                       VXH_VXI_ARGS(e), e->sc.data);
19227 +               break;
19228 +
19229 +       case VXH_ALLOC_VX_INFO:
19230 +       case VXH_DEALLOC_VX_INFO:
19231 +               printk( VXH_LOC_FMTS " %s_vx_info " VXH_VXI_FMTS "\n",
19232 +                       VXH_LOC_ARGS(e),
19233 +                       (e->type == VXH_ALLOC_VX_INFO) ? "alloc" : "dealloc",
19234 +                       VXH_VXI_ARGS(e));
19235 +               break;
19236 +
19237 +       case VXH_HASH_VX_INFO:
19238 +       case VXH_UNHASH_VX_INFO:
19239 +               printk( VXH_LOC_FMTS " __%s_vx_info " VXH_VXI_FMTS "\n",
19240 +                       VXH_LOC_ARGS(e),
19241 +                       (e->type == VXH_HASH_VX_INFO) ? "hash" : "unhash",
19242 +                       VXH_VXI_ARGS(e));
19243 +               break;
19244 +
19245 +       case VXH_LOC_VX_INFO:
19246 +       case VXH_LOOKUP_VX_INFO:
19247 +       case VXH_CREATE_VX_INFO:
19248 +               printk( VXH_LOC_FMTS " __%s_vx_info [#%d] -> " VXH_VXI_FMTS "\n",
19249 +                       VXH_LOC_ARGS(e),
19250 +                       (e->type == VXH_CREATE_VX_INFO) ? "create" :
19251 +                       ((e->type == VXH_LOC_VX_INFO) ? "loc" : "lookup"),
19252 +                       e->ll.arg, VXH_VXI_ARGS(e));
19253 +               break;
19254 +       }
19255 +}
19256 +
19257 +static void __vxh_dump_history(void)
19258 +{
19259 +       unsigned int i, cpu;
19260 +
19261 +       printk("History:\tSEQ: %8x\tNR_CPUS: %d\n",
19262 +               atomic_read(&sequence), NR_CPUS);
19263 +
19264 +       for (i = 0; i < VXH_SIZE; i++) {
19265 +               for_each_online_cpu(cpu) {
19266 +                       struct _vx_history *hist =
19267 +                               &per_cpu(vx_history_buffer, cpu);
19268 +                       unsigned int index = (hist->counter - i) % VXH_SIZE;
19269 +                       struct _vx_hist_entry *entry = &hist->entry[index];
19270 +
19271 +                       vxh_dump_entry(entry, cpu);
19272 +               }
19273 +       }
19274 +}
19275 +
19276 +void   vxh_dump_history(void)
19277 +{
19278 +       vxh_active = 0;
19279 +#ifdef CONFIG_SMP
19280 +       local_irq_enable();
19281 +       smp_send_stop();
19282 +       local_irq_disable();
19283 +#endif
19284 +       __vxh_dump_history();
19285 +}
19286 +
19287 +
19288 +/* vserver syscall commands below here */
19289 +
19290 +
19291 +int vc_dump_history(uint32_t id)
19292 +{
19293 +       vxh_active = 0;
19294 +       __vxh_dump_history();
19295 +       vxh_active = 1;
19296 +
19297 +       return 0;
19298 +}
19299 +
19300 +
19301 +int do_read_history(struct __user _vx_hist_entry *data,
19302 +       int cpu, uint32_t *index, uint32_t *count)
19303 +{
19304 +       int pos, ret = 0;
19305 +       struct _vx_history *hist = &per_cpu(vx_history_buffer, cpu);
19306 +       int end = hist->counter;
19307 +       int start = end - VXH_SIZE + 2;
19308 +       int idx = *index;
19309 +
19310 +       /* special case: get current pos */
19311 +       if (!*count) {
19312 +               *index = end;
19313 +               return 0;
19314 +       }
19315 +
19316 +       /* have we lost some data? */
19317 +       if (idx < start)
19318 +               idx = start;
19319 +
19320 +       for (pos = 0; (pos < *count) && (idx < end); pos++, idx++) {
19321 +               struct _vx_hist_entry *entry =
19322 +                       &hist->entry[idx % VXH_SIZE];
19323 +
19324 +               /* send entry to userspace */
19325 +               ret = copy_to_user(&data[pos], entry, sizeof(*entry));
19326 +               if (ret)
19327 +                       break;
19328 +       }
19329 +       /* save new index and count */
19330 +       *index = idx;
19331 +       *count = pos;
19332 +       return ret ? ret : (*index < end);
19333 +}
19334 +
19335 +int vc_read_history(uint32_t id, void __user *data)
19336 +{
19337 +       struct vcmd_read_history_v0 vc_data;
19338 +       int ret;
19339 +
19340 +       if (id >= NR_CPUS)
19341 +               return -EINVAL;
19342 +
19343 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19344 +               return -EFAULT;
19345 +
19346 +       ret = do_read_history((struct __user _vx_hist_entry *)vc_data.data,
19347 +               id, &vc_data.index, &vc_data.count);
19348 +
19349 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19350 +               return -EFAULT;
19351 +       return ret;
19352 +}
19353 +
19354 +#ifdef CONFIG_COMPAT
19355 +
19356 +int vc_read_history_x32(uint32_t id, void __user *data)
19357 +{
19358 +       struct vcmd_read_history_v0_x32 vc_data;
19359 +       int ret;
19360 +
19361 +       if (id >= NR_CPUS)
19362 +               return -EINVAL;
19363 +
19364 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19365 +               return -EFAULT;
19366 +
19367 +       ret = do_read_history((struct __user _vx_hist_entry *)
19368 +               compat_ptr(vc_data.data_ptr),
19369 +               id, &vc_data.index, &vc_data.count);
19370 +
19371 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19372 +               return -EFAULT;
19373 +       return ret;
19374 +}
19375 +
19376 +#endif /* CONFIG_COMPAT */
19377 +
19378 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/inet.c linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/inet.c
19379 --- linux-3.0-rc5/kernel/vserver/inet.c 1970-01-01 01:00:00.000000000 +0100
19380 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/inet.c    2011-06-13 18:20:44.000000000 +0200
19381 @@ -0,0 +1,225 @@
19382 +
19383 +#include <linux/in.h>
19384 +#include <linux/inetdevice.h>
19385 +#include <linux/vs_inet.h>
19386 +#include <linux/vs_inet6.h>
19387 +#include <linux/vserver/debug.h>
19388 +#include <net/route.h>
19389 +#include <net/addrconf.h>
19390 +
19391 +
19392 +int nx_v4_addr_conflict(struct nx_info *nxi1, struct nx_info *nxi2)
19393 +{
19394 +       int ret = 0;
19395 +
19396 +       if (!nxi1 || !nxi2 || nxi1 == nxi2)
19397 +               ret = 1;
19398 +       else {
19399 +               struct nx_addr_v4 *ptr;
19400 +
19401 +               for (ptr = &nxi1->v4; ptr; ptr = ptr->next) {
19402 +                       if (v4_nx_addr_in_nx_info(nxi2, ptr, -1)) {
19403 +                               ret = 1;
19404 +                               break;
19405 +                       }
19406 +               }
19407 +       }
19408 +
19409 +       vxdprintk(VXD_CBIT(net, 2),
19410 +               "nx_v4_addr_conflict(%p,%p): %d",
19411 +               nxi1, nxi2, ret);
19412 +
19413 +       return ret;
19414 +}
19415 +
19416 +
19417 +#ifdef CONFIG_IPV6
19418 +
19419 +int nx_v6_addr_conflict(struct nx_info *nxi1, struct nx_info *nxi2)
19420 +{
19421 +       int ret = 0;
19422 +
19423 +       if (!nxi1 || !nxi2 || nxi1 == nxi2)
19424 +               ret = 1;
19425 +       else {
19426 +               struct nx_addr_v6 *ptr;
19427 +
19428 +               for (ptr = &nxi1->v6; ptr; ptr = ptr->next) {
19429 +                       if (v6_nx_addr_in_nx_info(nxi2, ptr, -1)) {
19430 +                               ret = 1;
19431 +                               break;
19432 +                       }
19433 +               }
19434 +       }
19435 +
19436 +       vxdprintk(VXD_CBIT(net, 2),
19437 +               "nx_v6_addr_conflict(%p,%p): %d",
19438 +               nxi1, nxi2, ret);
19439 +
19440 +       return ret;
19441 +}
19442 +
19443 +#endif
19444 +
19445 +int v4_dev_in_nx_info(struct net_device *dev, struct nx_info *nxi)
19446 +{
19447 +       struct in_device *in_dev;
19448 +       struct in_ifaddr **ifap;
19449 +       struct in_ifaddr *ifa;
19450 +       int ret = 0;
19451 +
19452 +       if (!dev)
19453 +               goto out;
19454 +       in_dev = in_dev_get(dev);
19455 +       if (!in_dev)
19456 +               goto out;
19457 +
19458 +       for (ifap = &in_dev->ifa_list; (ifa = *ifap) != NULL;
19459 +               ifap = &ifa->ifa_next) {
19460 +               if (v4_addr_in_nx_info(nxi, ifa->ifa_local, NXA_MASK_SHOW)) {
19461 +                       ret = 1;
19462 +                       break;
19463 +               }
19464 +       }
19465 +       in_dev_put(in_dev);
19466 +out:
19467 +       return ret;
19468 +}
19469 +
19470 +
19471 +#ifdef CONFIG_IPV6
19472 +
19473 +int v6_dev_in_nx_info(struct net_device *dev, struct nx_info *nxi)
19474 +{
19475 +       struct inet6_dev *in_dev;
19476 +       struct inet6_ifaddr *ifa;
19477 +       int ret = 0;
19478 +
19479 +       if (!dev)
19480 +               goto out;
19481 +       in_dev = in6_dev_get(dev);
19482 +       if (!in_dev)
19483 +               goto out;
19484 +
19485 +       // for (ifap = &in_dev->addr_list; (ifa = *ifap) != NULL;
19486 +       list_for_each_entry(ifa, &in_dev->addr_list, if_list) {
19487 +               if (v6_addr_in_nx_info(nxi, &ifa->addr, -1)) {
19488 +                       ret = 1;
19489 +                       break;
19490 +               }
19491 +       }
19492 +       in6_dev_put(in_dev);
19493 +out:
19494 +       return ret;
19495 +}
19496 +
19497 +#endif
19498 +
19499 +int dev_in_nx_info(struct net_device *dev, struct nx_info *nxi)
19500 +{
19501 +       int ret = 1;
19502 +
19503 +       if (!nxi)
19504 +               goto out;
19505 +       if (nxi->v4.type && v4_dev_in_nx_info(dev, nxi))
19506 +               goto out;
19507 +#ifdef CONFIG_IPV6
19508 +       ret = 2;
19509 +       if (nxi->v6.type && v6_dev_in_nx_info(dev, nxi))
19510 +               goto out;
19511 +#endif
19512 +       ret = 0;
19513 +out:
19514 +       vxdprintk(VXD_CBIT(net, 3),
19515 +               "dev_in_nx_info(%p,%p[#%d]) = %d",
19516 +               dev, nxi, nxi ? nxi->nx_id : 0, ret);
19517 +       return ret;
19518 +}
19519 +
19520 +struct rtable *ip_v4_find_src(struct net *net, struct nx_info *nxi,
19521 +       struct flowi4 *fl4)
19522 +{
19523 +       struct rtable *rt;
19524 +
19525 +       if (!nxi)
19526 +               return NULL;
19527 +
19528 +       /* FIXME: handle lback only case */
19529 +       if (!NX_IPV4(nxi))
19530 +               return ERR_PTR(EPERM);
19531 +
19532 +       vxdprintk(VXD_CBIT(net, 4),
19533 +               "ip_v4_find_src(%p[#%u]) " NIPQUAD_FMT " -> " NIPQUAD_FMT,
19534 +               nxi, nxi ? nxi->nx_id : 0,
19535 +               NIPQUAD(fl4->saddr), NIPQUAD(fl4->daddr));
19536 +
19537 +       /* single IP is unconditional */
19538 +       if (nx_info_flags(nxi, NXF_SINGLE_IP, 0) &&
19539 +               (fl4->saddr == INADDR_ANY))
19540 +               fl4->saddr = nxi->v4.ip[0].s_addr;
19541 +
19542 +       if (fl4->saddr == INADDR_ANY) {
19543 +               struct nx_addr_v4 *ptr;
19544 +               __be32 found = 0;
19545 +
19546 +               rt = __ip_route_output_key(net, fl4);
19547 +               if (!IS_ERR(rt)) {
19548 +                       found = fl4->saddr;
19549 +                       ip_rt_put(rt);
19550 +                       vxdprintk(VXD_CBIT(net, 4),
19551 +                               "ip_v4_find_src(%p[#%u]) rok[%u]: " NIPQUAD_FMT,
19552 +                               nxi, nxi ? nxi->nx_id : 0, fl4->flowi4_oif, NIPQUAD(found));
19553 +                       if (v4_addr_in_nx_info(nxi, found, NXA_MASK_BIND))
19554 +                               goto found;
19555 +               }
19556 +
19557 +               for (ptr = &nxi->v4; ptr; ptr = ptr->next) {
19558 +                       __be32 primary = ptr->ip[0].s_addr;
19559 +                       __be32 mask = ptr->mask.s_addr;
19560 +                       __be32 neta = primary & mask;
19561 +
19562 +                       vxdprintk(VXD_CBIT(net, 4), "ip_v4_find_src(%p[#%u]) chk: "
19563 +                               NIPQUAD_FMT "/" NIPQUAD_FMT "/" NIPQUAD_FMT,
19564 +                               nxi, nxi ? nxi->nx_id : 0, NIPQUAD(primary),
19565 +                               NIPQUAD(mask), NIPQUAD(neta));
19566 +                       if ((found & mask) != neta)
19567 +                               continue;
19568 +
19569 +                       fl4->saddr = primary;
19570 +                       rt = __ip_route_output_key(net, fl4);
19571 +                       vxdprintk(VXD_CBIT(net, 4),
19572 +                               "ip_v4_find_src(%p[#%u]) rok[%u]: " NIPQUAD_FMT,
19573 +                               nxi, nxi ? nxi->nx_id : 0, fl4->flowi4_oif, NIPQUAD(primary));
19574 +                       if (!IS_ERR(rt)) {
19575 +                               found = fl4->saddr;
19576 +                               ip_rt_put(rt);
19577 +                               if (found == primary)
19578 +                                       goto found;
19579 +                       }
19580 +               }
19581 +               /* still no source ip? */
19582 +               found = ipv4_is_loopback(fl4->daddr)
19583 +                       ? IPI_LOOPBACK : nxi->v4.ip[0].s_addr;
19584 +       found:
19585 +               /* assign src ip to flow */
19586 +               fl4->saddr = found;
19587 +
19588 +       } else {
19589 +               if (!v4_addr_in_nx_info(nxi, fl4->saddr, NXA_MASK_BIND))
19590 +                       return ERR_PTR(EPERM);
19591 +       }
19592 +
19593 +       if (nx_info_flags(nxi, NXF_LBACK_REMAP, 0)) {
19594 +               if (ipv4_is_loopback(fl4->daddr))
19595 +                       fl4->daddr = nxi->v4_lback.s_addr;
19596 +               if (ipv4_is_loopback(fl4->saddr))
19597 +                       fl4->saddr = nxi->v4_lback.s_addr;
19598 +       } else if (ipv4_is_loopback(fl4->daddr) &&
19599 +               !nx_info_flags(nxi, NXF_LBACK_ALLOW, 0))
19600 +               return ERR_PTR(EPERM);
19601 +
19602 +       return NULL;
19603 +}
19604 +
19605 +EXPORT_SYMBOL_GPL(ip_v4_find_src);
19606 +
19607 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/init.c linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/init.c
19608 --- linux-3.0-rc5/kernel/vserver/init.c 1970-01-01 01:00:00.000000000 +0100
19609 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/init.c    2011-06-10 22:11:24.000000000 +0200
19610 @@ -0,0 +1,45 @@
19611 +/*
19612 + *  linux/kernel/init.c
19613 + *
19614 + *  Virtual Server Init
19615 + *
19616 + *  Copyright (C) 2004-2007  Herbert Pötzl
19617 + *
19618 + *  V0.01  basic structure
19619 + *
19620 + */
19621 +
19622 +#include <linux/init.h>
19623 +
19624 +int    vserver_register_sysctl(void);
19625 +void   vserver_unregister_sysctl(void);
19626 +
19627 +
19628 +static int __init init_vserver(void)
19629 +{
19630 +       int ret = 0;
19631 +
19632 +#ifdef CONFIG_VSERVER_DEBUG
19633 +       vserver_register_sysctl();
19634 +#endif
19635 +       return ret;
19636 +}
19637 +
19638 +
19639 +static void __exit exit_vserver(void)
19640 +{
19641 +
19642 +#ifdef CONFIG_VSERVER_DEBUG
19643 +       vserver_unregister_sysctl();
19644 +#endif
19645 +       return;
19646 +}
19647 +
19648 +/* FIXME: GFP_ZONETYPES gone
19649 +long vx_slab[GFP_ZONETYPES]; */
19650 +long vx_area;
19651 +
19652 +
19653 +module_init(init_vserver);
19654 +module_exit(exit_vserver);
19655 +
19656 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/inode.c linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/inode.c
19657 --- linux-3.0-rc5/kernel/vserver/inode.c        1970-01-01 01:00:00.000000000 +0100
19658 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/inode.c   2011-06-10 22:11:24.000000000 +0200
19659 @@ -0,0 +1,437 @@
19660 +/*
19661 + *  linux/kernel/vserver/inode.c
19662 + *
19663 + *  Virtual Server: File System Support
19664 + *
19665 + *  Copyright (C) 2004-2007  Herbert Pötzl
19666 + *
19667 + *  V0.01  separated from vcontext V0.05
19668 + *  V0.02  moved to tag (instead of xid)
19669 + *
19670 + */
19671 +
19672 +#include <linux/tty.h>
19673 +#include <linux/proc_fs.h>
19674 +#include <linux/devpts_fs.h>
19675 +#include <linux/fs.h>
19676 +#include <linux/file.h>
19677 +#include <linux/mount.h>
19678 +#include <linux/parser.h>
19679 +#include <linux/namei.h>
19680 +#include <linux/vserver/inode.h>
19681 +#include <linux/vserver/inode_cmd.h>
19682 +#include <linux/vs_base.h>
19683 +#include <linux/vs_tag.h>
19684 +
19685 +#include <asm/uaccess.h>
19686 +
19687 +
19688 +static int __vc_get_iattr(struct inode *in, uint32_t *tag, uint32_t *flags, uint32_t *mask)
19689 +{
19690 +       struct proc_dir_entry *entry;
19691 +
19692 +       if (!in || !in->i_sb)
19693 +               return -ESRCH;
19694 +
19695 +       *flags = IATTR_TAG
19696 +               | (IS_IMMUTABLE(in) ? IATTR_IMMUTABLE : 0)
19697 +               | (IS_IXUNLINK(in) ? IATTR_IXUNLINK : 0)
19698 +               | (IS_BARRIER(in) ? IATTR_BARRIER : 0)
19699 +               | (IS_COW(in) ? IATTR_COW : 0);
19700 +       *mask = IATTR_IXUNLINK | IATTR_IMMUTABLE | IATTR_COW;
19701 +
19702 +       if (S_ISDIR(in->i_mode))
19703 +               *mask |= IATTR_BARRIER;
19704 +
19705 +       if (IS_TAGGED(in)) {
19706 +               *tag = in->i_tag;
19707 +               *mask |= IATTR_TAG;
19708 +       }
19709 +
19710 +       switch (in->i_sb->s_magic) {
19711 +       case PROC_SUPER_MAGIC:
19712 +               entry = PROC_I(in)->pde;
19713 +
19714 +               /* check for specific inodes? */
19715 +               if (entry)
19716 +                       *mask |= IATTR_FLAGS;
19717 +               if (entry)
19718 +                       *flags |= (entry->vx_flags & IATTR_FLAGS);
19719 +               else
19720 +                       *flags |= (PROC_I(in)->vx_flags & IATTR_FLAGS);
19721 +               break;
19722 +
19723 +       case DEVPTS_SUPER_MAGIC:
19724 +               *tag = in->i_tag;
19725 +               *mask |= IATTR_TAG;
19726 +               break;
19727 +
19728 +       default:
19729 +               break;
19730 +       }
19731 +       return 0;
19732 +}
19733 +
19734 +int vc_get_iattr(void __user *data)
19735 +{
19736 +       struct path path;
19737 +       struct vcmd_ctx_iattr_v1 vc_data = { .tag = -1 };
19738 +       int ret;
19739 +
19740 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19741 +               return -EFAULT;
19742 +
19743 +       ret = user_lpath(vc_data.name, &path);
19744 +       if (!ret) {
19745 +               ret = __vc_get_iattr(path.dentry->d_inode,
19746 +                       &vc_data.tag, &vc_data.flags, &vc_data.mask);
19747 +               path_put(&path);
19748 +       }
19749 +       if (ret)
19750 +               return ret;
19751 +
19752 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19753 +               ret = -EFAULT;
19754 +       return ret;
19755 +}
19756 +
19757 +#ifdef CONFIG_COMPAT
19758 +
19759 +int vc_get_iattr_x32(void __user *data)
19760 +{
19761 +       struct path path;
19762 +       struct vcmd_ctx_iattr_v1_x32 vc_data = { .tag = -1 };
19763 +       int ret;
19764 +
19765 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19766 +               return -EFAULT;
19767 +
19768 +       ret = user_lpath(compat_ptr(vc_data.name_ptr), &path);
19769 +       if (!ret) {
19770 +               ret = __vc_get_iattr(path.dentry->d_inode,
19771 +                       &vc_data.tag, &vc_data.flags, &vc_data.mask);
19772 +               path_put(&path);
19773 +       }
19774 +       if (ret)
19775 +               return ret;
19776 +
19777 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19778 +               ret = -EFAULT;
19779 +       return ret;
19780 +}
19781 +
19782 +#endif /* CONFIG_COMPAT */
19783 +
19784 +
19785 +int vc_fget_iattr(uint32_t fd, void __user *data)
19786 +{
19787 +       struct file *filp;
19788 +       struct vcmd_ctx_fiattr_v0 vc_data = { .tag = -1 };
19789 +       int ret;
19790 +
19791 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19792 +               return -EFAULT;
19793 +
19794 +       filp = fget(fd);
19795 +       if (!filp || !filp->f_dentry || !filp->f_dentry->d_inode)
19796 +               return -EBADF;
19797 +
19798 +       ret = __vc_get_iattr(filp->f_dentry->d_inode,
19799 +               &vc_data.tag, &vc_data.flags, &vc_data.mask);
19800 +
19801 +       fput(filp);
19802 +
19803 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19804 +               ret = -EFAULT;
19805 +       return ret;
19806 +}
19807 +
19808 +
19809 +static int __vc_set_iattr(struct dentry *de, uint32_t *tag, uint32_t *flags, uint32_t *mask)
19810 +{
19811 +       struct inode *in = de->d_inode;
19812 +       int error = 0, is_proc = 0, has_tag = 0;
19813 +       struct iattr attr = { 0 };
19814 +
19815 +       if (!in || !in->i_sb)
19816 +               return -ESRCH;
19817 +
19818 +       is_proc = (in->i_sb->s_magic == PROC_SUPER_MAGIC);
19819 +       if ((*mask & IATTR_FLAGS) && !is_proc)
19820 +               return -EINVAL;
19821 +
19822 +       has_tag = IS_TAGGED(in) ||
19823 +               (in->i_sb->s_magic == DEVPTS_SUPER_MAGIC);
19824 +       if ((*mask & IATTR_TAG) && !has_tag)
19825 +               return -EINVAL;
19826 +
19827 +       mutex_lock(&in->i_mutex);
19828 +       if (*mask & IATTR_TAG) {
19829 +               attr.ia_tag = *tag;
19830 +               attr.ia_valid |= ATTR_TAG;
19831 +       }
19832 +
19833 +       if (*mask & IATTR_FLAGS) {
19834 +               struct proc_dir_entry *entry = PROC_I(in)->pde;
19835 +               unsigned int iflags = PROC_I(in)->vx_flags;
19836 +
19837 +               iflags = (iflags & ~(*mask & IATTR_FLAGS))
19838 +                       | (*flags & IATTR_FLAGS);
19839 +               PROC_I(in)->vx_flags = iflags;
19840 +               if (entry)
19841 +                       entry->vx_flags = iflags;
19842 +       }
19843 +
19844 +       if (*mask & (IATTR_IMMUTABLE | IATTR_IXUNLINK |
19845 +               IATTR_BARRIER | IATTR_COW)) {
19846 +               int iflags = in->i_flags;
19847 +               int vflags = in->i_vflags;
19848 +
19849 +               if (*mask & IATTR_IMMUTABLE) {
19850 +                       if (*flags & IATTR_IMMUTABLE)
19851 +                               iflags |= S_IMMUTABLE;
19852 +                       else
19853 +                               iflags &= ~S_IMMUTABLE;
19854 +               }
19855 +               if (*mask & IATTR_IXUNLINK) {
19856 +                       if (*flags & IATTR_IXUNLINK)
19857 +                               iflags |= S_IXUNLINK;
19858 +                       else
19859 +                               iflags &= ~S_IXUNLINK;
19860 +               }
19861 +               if (S_ISDIR(in->i_mode) && (*mask & IATTR_BARRIER)) {
19862 +                       if (*flags & IATTR_BARRIER)
19863 +                               vflags |= V_BARRIER;
19864 +                       else
19865 +                               vflags &= ~V_BARRIER;
19866 +               }
19867 +               if (S_ISREG(in->i_mode) && (*mask & IATTR_COW)) {
19868 +                       if (*flags & IATTR_COW)
19869 +                               vflags |= V_COW;
19870 +                       else
19871 +                               vflags &= ~V_COW;
19872 +               }
19873 +               if (in->i_op && in->i_op->sync_flags) {
19874 +                       error = in->i_op->sync_flags(in, iflags, vflags);
19875 +                       if (error)
19876 +                               goto out;
19877 +               }
19878 +       }
19879 +
19880 +       if (attr.ia_valid) {
19881 +               if (in->i_op && in->i_op->setattr)
19882 +                       error = in->i_op->setattr(de, &attr);
19883 +               else {
19884 +                       error = inode_change_ok(in, &attr);
19885 +                       if (!error) {
19886 +                               setattr_copy(in, &attr);
19887 +                               mark_inode_dirty(in);
19888 +                       }
19889 +               }
19890 +       }
19891 +
19892 +out:
19893 +       mutex_unlock(&in->i_mutex);
19894 +       return error;
19895 +}
19896 +
19897 +int vc_set_iattr(void __user *data)
19898 +{
19899 +       struct path path;
19900 +       struct vcmd_ctx_iattr_v1 vc_data;
19901 +       int ret;
19902 +
19903 +       if (!capable(CAP_LINUX_IMMUTABLE))
19904 +               return -EPERM;
19905 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19906 +               return -EFAULT;
19907 +
19908 +       ret = user_lpath(vc_data.name, &path);
19909 +       if (!ret) {
19910 +               ret = __vc_set_iattr(path.dentry,
19911 +                       &vc_data.tag, &vc_data.flags, &vc_data.mask);
19912 +               path_put(&path);
19913 +       }
19914 +
19915 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19916 +               ret = -EFAULT;
19917 +       return ret;
19918 +}
19919 +
19920 +#ifdef CONFIG_COMPAT
19921 +
19922 +int vc_set_iattr_x32(void __user *data)
19923 +{
19924 +       struct path path;
19925 +       struct vcmd_ctx_iattr_v1_x32 vc_data;
19926 +       int ret;
19927 +
19928 +       if (!capable(CAP_LINUX_IMMUTABLE))
19929 +               return -EPERM;
19930 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19931 +               return -EFAULT;
19932 +
19933 +       ret = user_lpath(compat_ptr(vc_data.name_ptr), &path);
19934 +       if (!ret) {
19935 +               ret = __vc_set_iattr(path.dentry,
19936 +                       &vc_data.tag, &vc_data.flags, &vc_data.mask);
19937 +               path_put(&path);
19938 +       }
19939 +
19940 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19941 +               ret = -EFAULT;
19942 +       return ret;
19943 +}
19944 +
19945 +#endif /* CONFIG_COMPAT */
19946 +
19947 +int vc_fset_iattr(uint32_t fd, void __user *data)
19948 +{
19949 +       struct file *filp;
19950 +       struct vcmd_ctx_fiattr_v0 vc_data;
19951 +       int ret;
19952 +
19953 +       if (!capable(CAP_LINUX_IMMUTABLE))
19954 +               return -EPERM;
19955 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19956 +               return -EFAULT;
19957 +
19958 +       filp = fget(fd);
19959 +       if (!filp || !filp->f_dentry || !filp->f_dentry->d_inode)
19960 +               return -EBADF;
19961 +
19962 +       ret = __vc_set_iattr(filp->f_dentry, &vc_data.tag,
19963 +               &vc_data.flags, &vc_data.mask);
19964 +
19965 +       fput(filp);
19966 +
19967 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19968 +               return -EFAULT;
19969 +       return ret;
19970 +}
19971 +
19972 +
19973 +enum { Opt_notagcheck, Opt_tag, Opt_notag, Opt_tagid, Opt_err };
19974 +
19975 +static match_table_t tokens = {
19976 +       {Opt_notagcheck, "notagcheck"},
19977 +#ifdef CONFIG_PROPAGATE
19978 +       {Opt_notag, "notag"},
19979 +       {Opt_tag, "tag"},
19980 +       {Opt_tagid, "tagid=%u"},
19981 +#endif
19982 +       {Opt_err, NULL}
19983 +};
19984 +
19985 +
19986 +static void __dx_parse_remove(char *string, char *opt)
19987 +{
19988 +       char *p = strstr(string, opt);
19989 +       char *q = p;
19990 +
19991 +       if (p) {
19992 +               while (*q != '\0' && *q != ',')
19993 +                       q++;
19994 +               while (*q)
19995 +                       *p++ = *q++;
19996 +               while (*p)
19997 +                       *p++ = '\0';
19998 +       }
19999 +}
20000 +
20001 +int dx_parse_tag(char *string, tag_t *tag, int remove, int *mnt_flags,
20002 +                unsigned long *flags)
20003 +{
20004 +       int set = 0;
20005 +       substring_t args[MAX_OPT_ARGS];
20006 +       int token;
20007 +       char *s, *p, *opts;
20008 +#if defined(CONFIG_PROPAGATE) || defined(CONFIG_VSERVER_WARN)
20009 +       int option = 0;
20010 +#endif
20011 +
20012 +       if (!string)
20013 +               return 0;
20014 +       s = kstrdup(string, GFP_KERNEL | GFP_ATOMIC);
20015 +       if (!s)
20016 +               return 0;
20017 +
20018 +       opts = s;
20019 +       while ((p = strsep(&opts, ",")) != NULL) {
20020 +               token = match_token(p, tokens, args);
20021 +
20022 +               switch (token) {
20023 +#ifdef CONFIG_PROPAGATE
20024 +               case Opt_tag:
20025 +                       if (tag)
20026 +                               *tag = 0;
20027 +                       if (remove)
20028 +                               __dx_parse_remove(s, "tag");
20029 +                       *mnt_flags |= MNT_TAGID;
20030 +                       set |= MNT_TAGID;
20031 +                       break;
20032 +               case Opt_notag:
20033 +                       if (remove)
20034 +                               __dx_parse_remove(s, "notag");
20035 +                       *mnt_flags |= MNT_NOTAG;
20036 +                       set |= MNT_NOTAG;
20037 +                       break;
20038 +               case Opt_tagid:
20039 +                       if (tag && !match_int(args, &option))
20040 +                               *tag = option;
20041 +                       if (remove)
20042 +                               __dx_parse_remove(s, "tagid");
20043 +                       *mnt_flags |= MNT_TAGID;
20044 +                       set |= MNT_TAGID;
20045 +                       break;
20046 +#endif
20047 +               case Opt_notagcheck:
20048 +                       if (remove)
20049 +                               __dx_parse_remove(s, "notagcheck");
20050 +                       *flags |= MS_NOTAGCHECK;
20051 +                       set |= MS_NOTAGCHECK;
20052 +                       break;
20053 +               }
20054 +               vxdprintk(VXD_CBIT(tag, 7),
20055 +                       "dx_parse_tag(" VS_Q("%s") "): %d:#%d",
20056 +                       p, token, option);
20057 +       }
20058 +       if (set)
20059 +               strcpy(string, s);
20060 +       kfree(s);
20061 +       return set;
20062 +}
20063 +
20064 +#ifdef CONFIG_PROPAGATE
20065 +
20066 +void __dx_propagate_tag(struct nameidata *nd, struct inode *inode)
20067 +{
20068 +       tag_t new_tag = 0;
20069 +       struct vfsmount *mnt;
20070 +       int propagate;
20071 +
20072 +       if (!nd)
20073 +               return;
20074 +       mnt = nd->path.mnt;
20075 +       if (!mnt)
20076 +               return;
20077 +
20078 +       propagate = (mnt->mnt_flags & MNT_TAGID);
20079 +       if (propagate)
20080 +               new_tag = mnt->mnt_tag;
20081 +
20082 +       vxdprintk(VXD_CBIT(tag, 7),
20083 +               "dx_propagate_tag(%p[#%lu.%d]): %d,%d",
20084 +               inode, inode->i_ino, inode->i_tag,
20085 +               new_tag, (propagate) ? 1 : 0);
20086 +
20087 +       if (propagate)
20088 +               inode->i_tag = new_tag;
20089 +}
20090 +
20091 +#include <linux/module.h>
20092 +
20093 +EXPORT_SYMBOL_GPL(__dx_propagate_tag);
20094 +
20095 +#endif /* CONFIG_PROPAGATE */
20096 +
20097 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/limit.c linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/limit.c
20098 --- linux-3.0-rc5/kernel/vserver/limit.c        1970-01-01 01:00:00.000000000 +0100
20099 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/limit.c   2011-06-10 22:11:24.000000000 +0200
20100 @@ -0,0 +1,354 @@
20101 +/*
20102 + *  linux/kernel/vserver/limit.c
20103 + *
20104 + *  Virtual Server: Context Limits
20105 + *
20106 + *  Copyright (C) 2004-2010  Herbert Pötzl
20107 + *
20108 + *  V0.01  broken out from vcontext V0.05
20109 + *  V0.02  changed vcmds to vxi arg
20110 + *  V0.03  added memory cgroup support
20111 + *
20112 + */
20113 +
20114 +#include <linux/sched.h>
20115 +#include <linux/module.h>
20116 +#include <linux/memcontrol.h>
20117 +#include <linux/res_counter.h>
20118 +#include <linux/vs_limit.h>
20119 +#include <linux/vserver/limit.h>
20120 +#include <linux/vserver/limit_cmd.h>
20121 +
20122 +#include <asm/uaccess.h>
20123 +
20124 +
20125 +const char *vlimit_name[NUM_LIMITS] = {
20126 +#ifdef CONFIG_VSERVER_LEGACY_MEM
20127 +       [RLIMIT_RSS]            = "RSS",
20128 +       [RLIMIT_AS]             = "VM",
20129 +#endif /* CONFIG_VSERVER_LEGACY_MEM */
20130 +       [RLIMIT_CPU]            = "CPU",
20131 +       [RLIMIT_NPROC]          = "NPROC",
20132 +       [RLIMIT_NOFILE]         = "NOFILE",
20133 +       [RLIMIT_LOCKS]          = "LOCKS",
20134 +       [RLIMIT_SIGPENDING]     = "SIGP",
20135 +       [RLIMIT_MSGQUEUE]       = "MSGQ",
20136 +
20137 +       [VLIMIT_NSOCK]          = "NSOCK",
20138 +       [VLIMIT_OPENFD]         = "OPENFD",
20139 +       [VLIMIT_SHMEM]          = "SHMEM",
20140 +       [VLIMIT_DENTRY]         = "DENTRY",
20141 +};
20142 +
20143 +EXPORT_SYMBOL_GPL(vlimit_name);
20144 +
20145 +#define MASK_ENTRY(x)  (1 << (x))
20146 +
20147 +const struct vcmd_ctx_rlimit_mask_v0 vlimit_mask = {
20148 +               /* minimum */
20149 +       0
20150 +       ,       /* softlimit */
20151 +#ifdef CONFIG_VSERVER_LEGACY_MEM
20152 +       MASK_ENTRY( RLIMIT_RSS          ) |
20153 +#endif /* CONFIG_VSERVER_LEGACY_MEM */
20154 +       0
20155 +       ,       /* maximum */
20156 +#ifdef CONFIG_VSERVER_LEGACY_MEM
20157 +       MASK_ENTRY( RLIMIT_RSS          ) |
20158 +       MASK_ENTRY( RLIMIT_AS           ) |
20159 +#endif /* CONFIG_VSERVER_LEGACY_MEM */
20160 +       MASK_ENTRY( RLIMIT_NPROC        ) |
20161 +       MASK_ENTRY( RLIMIT_NOFILE       ) |
20162 +       MASK_ENTRY( RLIMIT_LOCKS        ) |
20163 +       MASK_ENTRY( RLIMIT_MSGQUEUE     ) |
20164 +
20165 +       MASK_ENTRY( VLIMIT_NSOCK        ) |
20166 +       MASK_ENTRY( VLIMIT_OPENFD       ) |
20167 +       MASK_ENTRY( VLIMIT_SHMEM        ) |
20168 +       MASK_ENTRY( VLIMIT_DENTRY       ) |
20169 +       0
20170 +};
20171 +               /* accounting only */
20172 +uint32_t account_mask =
20173 +       MASK_ENTRY( VLIMIT_SEMARY       ) |
20174 +       MASK_ENTRY( VLIMIT_NSEMS        ) |
20175 +       MASK_ENTRY( VLIMIT_MAPPED       ) |
20176 +       0;
20177 +
20178 +
20179 +static int is_valid_vlimit(int id)
20180 +{
20181 +       uint32_t mask = vlimit_mask.minimum |
20182 +               vlimit_mask.softlimit | vlimit_mask.maximum;
20183 +       return mask & (1 << id);
20184 +}
20185 +
20186 +static int is_accounted_vlimit(int id)
20187 +{
20188 +       if (is_valid_vlimit(id))
20189 +               return 1;
20190 +       return account_mask & (1 << id);
20191 +}
20192 +
20193 +
20194 +static inline uint64_t vc_get_soft(struct vx_info *vxi, int id)
20195 +{
20196 +       rlim_t limit = __rlim_soft(&vxi->limit, id);
20197 +       return VX_VLIM(limit);
20198 +}
20199 +
20200 +static inline uint64_t vc_get_hard(struct vx_info *vxi, int id)
20201 +{
20202 +       rlim_t limit = __rlim_hard(&vxi->limit, id);
20203 +       return VX_VLIM(limit);
20204 +}
20205 +
20206 +static int do_get_rlimit(struct vx_info *vxi, uint32_t id,
20207 +       uint64_t *minimum, uint64_t *softlimit, uint64_t *maximum)
20208 +{
20209 +       if (!is_valid_vlimit(id))
20210 +               return -EINVAL;
20211 +
20212 +       if (minimum)
20213 +               *minimum = CRLIM_UNSET;
20214 +       if (softlimit)
20215 +               *softlimit = vc_get_soft(vxi, id);
20216 +       if (maximum)
20217 +               *maximum = vc_get_hard(vxi, id);
20218 +       return 0;
20219 +}
20220 +
20221 +int vc_get_rlimit(struct vx_info *vxi, void __user *data)
20222 +{
20223 +       struct vcmd_ctx_rlimit_v0 vc_data;
20224 +       int ret;
20225 +
20226 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20227 +               return -EFAULT;
20228 +
20229 +       ret = do_get_rlimit(vxi, vc_data.id,
20230 +               &vc_data.minimum, &vc_data.softlimit, &vc_data.maximum);
20231 +       if (ret)
20232 +               return ret;
20233 +
20234 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
20235 +               return -EFAULT;
20236 +       return 0;
20237 +}
20238 +
20239 +static int do_set_rlimit(struct vx_info *vxi, uint32_t id,
20240 +       uint64_t minimum, uint64_t softlimit, uint64_t maximum)
20241 +{
20242 +       if (!is_valid_vlimit(id))
20243 +               return -EINVAL;
20244 +
20245 +       if (maximum != CRLIM_KEEP)
20246 +               __rlim_hard(&vxi->limit, id) = VX_RLIM(maximum);
20247 +       if (softlimit != CRLIM_KEEP)
20248 +               __rlim_soft(&vxi->limit, id) = VX_RLIM(softlimit);
20249 +
20250 +       /* clamp soft limit */
20251 +       if (__rlim_soft(&vxi->limit, id) > __rlim_hard(&vxi->limit, id))
20252 +               __rlim_soft(&vxi->limit, id) = __rlim_hard(&vxi->limit, id);
20253 +
20254 +       return 0;
20255 +}
20256 +
20257 +int vc_set_rlimit(struct vx_info *vxi, void __user *data)
20258 +{
20259 +       struct vcmd_ctx_rlimit_v0 vc_data;
20260 +
20261 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20262 +               return -EFAULT;
20263 +
20264 +       return do_set_rlimit(vxi, vc_data.id,
20265 +               vc_data.minimum, vc_data.softlimit, vc_data.maximum);
20266 +}
20267 +
20268 +#ifdef CONFIG_IA32_EMULATION
20269 +
20270 +int vc_set_rlimit_x32(struct vx_info *vxi, void __user *data)
20271 +{
20272 +       struct vcmd_ctx_rlimit_v0_x32 vc_data;
20273 +
20274 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20275 +               return -EFAULT;
20276 +
20277 +       return do_set_rlimit(vxi, vc_data.id,
20278 +               vc_data.minimum, vc_data.softlimit, vc_data.maximum);
20279 +}
20280 +
20281 +int vc_get_rlimit_x32(struct vx_info *vxi, void __user *data)
20282 +{
20283 +       struct vcmd_ctx_rlimit_v0_x32 vc_data;
20284 +       int ret;
20285 +
20286 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20287 +               return -EFAULT;
20288 +
20289 +       ret = do_get_rlimit(vxi, vc_data.id,
20290 +               &vc_data.minimum, &vc_data.softlimit, &vc_data.maximum);
20291 +       if (ret)
20292 +               return ret;
20293 +
20294 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
20295 +               return -EFAULT;
20296 +       return 0;
20297 +}
20298 +
20299 +#endif /* CONFIG_IA32_EMULATION */
20300 +
20301 +
20302 +int vc_get_rlimit_mask(uint32_t id, void __user *data)
20303 +{
20304 +       if (copy_to_user(data, &vlimit_mask, sizeof(vlimit_mask)))
20305 +               return -EFAULT;
20306 +       return 0;
20307 +}
20308 +
20309 +
20310 +static inline void vx_reset_hits(struct _vx_limit *limit)
20311 +{
20312 +       int lim;
20313 +
20314 +       for (lim = 0; lim < NUM_LIMITS; lim++) {
20315 +               atomic_set(&__rlim_lhit(limit, lim), 0);
20316 +       }
20317 +}
20318 +
20319 +int vc_reset_hits(struct vx_info *vxi, void __user *data)
20320 +{
20321 +       vx_reset_hits(&vxi->limit);
20322 +       return 0;
20323 +}
20324 +
20325 +static inline void vx_reset_minmax(struct _vx_limit *limit)
20326 +{
20327 +       rlim_t value;
20328 +       int lim;
20329 +
20330 +       for (lim = 0; lim < NUM_LIMITS; lim++) {
20331 +               value = __rlim_get(limit, lim);
20332 +               __rlim_rmax(limit, lim) = value;
20333 +               __rlim_rmin(limit, lim) = value;
20334 +       }
20335 +}
20336 +
20337 +int vc_reset_minmax(struct vx_info *vxi, void __user *data)
20338 +{
20339 +       vx_reset_minmax(&vxi->limit);
20340 +       return 0;
20341 +}
20342 +
20343 +
20344 +int vc_rlimit_stat(struct vx_info *vxi, void __user *data)
20345 +{
20346 +       struct vcmd_rlimit_stat_v0 vc_data;
20347 +       struct _vx_limit *limit = &vxi->limit;
20348 +       int id;
20349 +
20350 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20351 +               return -EFAULT;
20352 +
20353 +       id = vc_data.id;
20354 +       if (!is_accounted_vlimit(id))
20355 +               return -EINVAL;
20356 +
20357 +       vx_limit_fixup(limit, id);
20358 +       vc_data.hits = atomic_read(&__rlim_lhit(limit, id));
20359 +       vc_data.value = __rlim_get(limit, id);
20360 +       vc_data.minimum = __rlim_rmin(limit, id);
20361 +       vc_data.maximum = __rlim_rmax(limit, id);
20362 +
20363 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
20364 +               return -EFAULT;
20365 +       return 0;
20366 +}
20367 +
20368 +
20369 +void vx_vsi_meminfo(struct sysinfo *val)
20370 +{
20371 +#ifdef CONFIG_CGROUP_MEM_RES_CTLR
20372 +       struct mem_cgroup *mcg = mem_cgroup_from_task(current);
20373 +       u64 res_limit, res_usage;
20374 +
20375 +       if (!mcg)
20376 +               return;
20377 +
20378 +       res_limit = mem_cgroup_res_read_u64(mcg, RES_LIMIT);
20379 +       res_usage = mem_cgroup_res_read_u64(mcg, RES_USAGE);
20380 +
20381 +       if (res_limit != RESOURCE_MAX)
20382 +               val->totalram = (res_limit >> PAGE_SHIFT);
20383 +       val->freeram = val->totalram - (res_usage >> PAGE_SHIFT);
20384 +       val->bufferram = 0;
20385 +       val->totalhigh = 0;
20386 +       val->freehigh = 0;
20387 +#endif /* CONFIG_CGROUP_MEM_RES_CTLR */
20388 +       return;
20389 +}
20390 +
20391 +void vx_vsi_swapinfo(struct sysinfo *val)
20392 +{
20393 +#ifdef CONFIG_CGROUP_MEM_RES_CTLR
20394 +#ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP
20395 +       struct mem_cgroup *mcg = mem_cgroup_from_task(current);
20396 +       u64 res_limit, res_usage, memsw_limit, memsw_usage;
20397 +       s64 swap_limit, swap_usage;
20398 +
20399 +       if (!mcg)
20400 +               return;
20401 +
20402 +       res_limit = mem_cgroup_res_read_u64(mcg, RES_LIMIT);
20403 +       res_usage = mem_cgroup_res_read_u64(mcg, RES_USAGE);
20404 +       memsw_limit = mem_cgroup_memsw_read_u64(mcg, RES_LIMIT);
20405 +       memsw_usage = mem_cgroup_memsw_read_u64(mcg, RES_USAGE);
20406 +
20407 +       if (res_limit == RESOURCE_MAX)
20408 +               return;
20409 +
20410 +       swap_limit = memsw_limit - res_limit;
20411 +       if (memsw_limit != RESOURCE_MAX)
20412 +               val->totalswap = swap_limit >> PAGE_SHIFT;
20413 +
20414 +       swap_usage = memsw_usage - res_usage;
20415 +       val->freeswap = (swap_usage < swap_limit) ?
20416 +               val->totalswap - (swap_usage >> PAGE_SHIFT) : 0;
20417 +#else  /* !CONFIG_CGROUP_MEM_RES_CTLR_SWAP */
20418 +       val->totalswap = 0;
20419 +       val->freeswap = 0;
20420 +#endif /* !CONFIG_CGROUP_MEM_RES_CTLR_SWAP */
20421 +#endif /* CONFIG_CGROUP_MEM_RES_CTLR */
20422 +       return;
20423 +}
20424 +
20425 +long vx_vsi_cached(struct sysinfo *val)
20426 +{
20427 +#ifdef CONFIG_CGROUP_MEM_RES_CTLR
20428 +       struct mem_cgroup *mcg = mem_cgroup_from_task(current);
20429 +
20430 +       return mem_cgroup_stat_read_cache(mcg);
20431 +#else
20432 +       return 0;
20433 +#endif
20434 +}
20435 +
20436 +
20437 +unsigned long vx_badness(struct task_struct *task, struct mm_struct *mm)
20438 +{
20439 +       struct vx_info *vxi = mm->mm_vx_info;
20440 +       unsigned long points;
20441 +       rlim_t v, w;
20442 +
20443 +       if (!vxi)
20444 +               return 0;
20445 +
20446 +       points = vxi->vx_badness_bias;
20447 +
20448 +       v = __vx_cres_array_fixup(&vxi->limit, VLA_RSS);
20449 +       w = __rlim_soft(&vxi->limit, RLIMIT_RSS);
20450 +       points += (v > w) ? (v - w) : 0;
20451 +
20452 +       return points;
20453 +}
20454 +
20455 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/limit_init.h linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/limit_init.h
20456 --- linux-3.0-rc5/kernel/vserver/limit_init.h   1970-01-01 01:00:00.000000000 +0100
20457 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/limit_init.h      2011-06-10 22:11:24.000000000 +0200
20458 @@ -0,0 +1,31 @@
20459 +
20460 +
20461 +static inline void vx_info_init_limit(struct _vx_limit *limit)
20462 +{
20463 +       int lim;
20464 +
20465 +       for (lim = 0; lim < NUM_LIMITS; lim++) {
20466 +               __rlim_soft(limit, lim) = RLIM_INFINITY;
20467 +               __rlim_hard(limit, lim) = RLIM_INFINITY;
20468 +               __rlim_set(limit, lim, 0);
20469 +               atomic_set(&__rlim_lhit(limit, lim), 0);
20470 +               __rlim_rmin(limit, lim) = 0;
20471 +               __rlim_rmax(limit, lim) = 0;
20472 +       }
20473 +}
20474 +
20475 +static inline void vx_info_exit_limit(struct _vx_limit *limit)
20476 +{
20477 +       rlim_t value;
20478 +       int lim;
20479 +
20480 +       for (lim = 0; lim < NUM_LIMITS; lim++) {
20481 +               if ((1 << lim) & VLIM_NOCHECK)
20482 +                       continue;
20483 +               value = __rlim_get(limit, lim);
20484 +               vxwprintk_xid(value,
20485 +                       "!!! limit: %p[%s,%d] = %ld on exit.",
20486 +                       limit, vlimit_name[lim], lim, (long)value);
20487 +       }
20488 +}
20489 +
20490 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/limit_proc.h linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/limit_proc.h
20491 --- linux-3.0-rc5/kernel/vserver/limit_proc.h   1970-01-01 01:00:00.000000000 +0100
20492 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/limit_proc.h      2011-06-10 22:11:24.000000000 +0200
20493 @@ -0,0 +1,57 @@
20494 +#ifndef _VX_LIMIT_PROC_H
20495 +#define _VX_LIMIT_PROC_H
20496 +
20497 +#include <linux/vserver/limit_int.h>
20498 +
20499 +
20500 +#define VX_LIMIT_FMT   ":\t%8ld\t%8ld/%8ld\t%8lld/%8lld\t%6d\n"
20501 +#define VX_LIMIT_TOP   \
20502 +       "Limit\t current\t     min/max\t\t    soft/hard\t\thits\n"
20503 +
20504 +#define VX_LIMIT_ARG(r)                                \
20505 +       (unsigned long)__rlim_get(limit, r),    \
20506 +       (unsigned long)__rlim_rmin(limit, r),   \
20507 +       (unsigned long)__rlim_rmax(limit, r),   \
20508 +       VX_VLIM(__rlim_soft(limit, r)),         \
20509 +       VX_VLIM(__rlim_hard(limit, r)),         \
20510 +       atomic_read(&__rlim_lhit(limit, r))
20511 +
20512 +static inline int vx_info_proc_limit(struct _vx_limit *limit, char *buffer)
20513 +{
20514 +       vx_limit_fixup(limit, -1);
20515 +       return sprintf(buffer, VX_LIMIT_TOP
20516 +               "PROC"  VX_LIMIT_FMT
20517 +               "VM"    VX_LIMIT_FMT
20518 +               "VML"   VX_LIMIT_FMT
20519 +               "RSS"   VX_LIMIT_FMT
20520 +               "ANON"  VX_LIMIT_FMT
20521 +               "RMAP"  VX_LIMIT_FMT
20522 +               "FILES" VX_LIMIT_FMT
20523 +               "OFD"   VX_LIMIT_FMT
20524 +               "LOCKS" VX_LIMIT_FMT
20525 +               "SOCK"  VX_LIMIT_FMT
20526 +               "MSGQ"  VX_LIMIT_FMT
20527 +               "SHM"   VX_LIMIT_FMT
20528 +               "SEMA"  VX_LIMIT_FMT
20529 +               "SEMS"  VX_LIMIT_FMT
20530 +               "DENT"  VX_LIMIT_FMT,
20531 +               VX_LIMIT_ARG(RLIMIT_NPROC),
20532 +               VX_LIMIT_ARG(RLIMIT_AS),
20533 +               VX_LIMIT_ARG(RLIMIT_MEMLOCK),
20534 +               VX_LIMIT_ARG(RLIMIT_RSS),
20535 +               VX_LIMIT_ARG(VLIMIT_ANON),
20536 +               VX_LIMIT_ARG(VLIMIT_MAPPED),
20537 +               VX_LIMIT_ARG(RLIMIT_NOFILE),
20538 +               VX_LIMIT_ARG(VLIMIT_OPENFD),
20539 +               VX_LIMIT_ARG(RLIMIT_LOCKS),
20540 +               VX_LIMIT_ARG(VLIMIT_NSOCK),
20541 +               VX_LIMIT_ARG(RLIMIT_MSGQUEUE),
20542 +               VX_LIMIT_ARG(VLIMIT_SHMEM),
20543 +               VX_LIMIT_ARG(VLIMIT_SEMARY),
20544 +               VX_LIMIT_ARG(VLIMIT_NSEMS),
20545 +               VX_LIMIT_ARG(VLIMIT_DENTRY));
20546 +}
20547 +
20548 +#endif /* _VX_LIMIT_PROC_H */
20549 +
20550 +
20551 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/network.c linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/network.c
20552 --- linux-3.0-rc5/kernel/vserver/network.c      1970-01-01 01:00:00.000000000 +0100
20553 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/network.c 2011-06-10 23:20:56.000000000 +0200
20554 @@ -0,0 +1,912 @@
20555 +/*
20556 + *  linux/kernel/vserver/network.c
20557 + *
20558 + *  Virtual Server: Network Support
20559 + *
20560 + *  Copyright (C) 2003-2007  Herbert Pötzl
20561 + *
20562 + *  V0.01  broken out from vcontext V0.05
20563 + *  V0.02  cleaned up implementation
20564 + *  V0.03  added equiv nx commands
20565 + *  V0.04  switch to RCU based hash
20566 + *  V0.05  and back to locking again
20567 + *  V0.06  changed vcmds to nxi arg
20568 + *  V0.07  have __create claim() the nxi
20569 + *
20570 + */
20571 +
20572 +#include <linux/err.h>
20573 +#include <linux/slab.h>
20574 +#include <linux/rcupdate.h>
20575 +
20576 +#include <linux/vs_network.h>
20577 +#include <linux/vs_pid.h>
20578 +#include <linux/vserver/network_cmd.h>
20579 +
20580 +
20581 +atomic_t nx_global_ctotal      = ATOMIC_INIT(0);
20582 +atomic_t nx_global_cactive     = ATOMIC_INIT(0);
20583 +
20584 +static struct kmem_cache *nx_addr_v4_cachep = NULL;
20585 +static struct kmem_cache *nx_addr_v6_cachep = NULL;
20586 +
20587 +
20588 +static int __init init_network(void)
20589 +{
20590 +       nx_addr_v4_cachep = kmem_cache_create("nx_v4_addr_cache",
20591 +               sizeof(struct nx_addr_v4), 0,
20592 +               SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
20593 +       nx_addr_v6_cachep = kmem_cache_create("nx_v6_addr_cache",
20594 +               sizeof(struct nx_addr_v6), 0,
20595 +               SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
20596 +       return 0;
20597 +}
20598 +
20599 +
20600 +/*     __alloc_nx_addr_v4()                                    */
20601 +
20602 +static inline struct nx_addr_v4 *__alloc_nx_addr_v4(void)
20603 +{
20604 +       struct nx_addr_v4 *nxa = kmem_cache_alloc(
20605 +               nx_addr_v4_cachep, GFP_KERNEL);
20606 +
20607 +       if (!IS_ERR(nxa))
20608 +               memset(nxa, 0, sizeof(*nxa));
20609 +       return nxa;
20610 +}
20611 +
20612 +/*     __dealloc_nx_addr_v4()                                  */
20613 +
20614 +static inline void __dealloc_nx_addr_v4(struct nx_addr_v4 *nxa)
20615 +{
20616 +       kmem_cache_free(nx_addr_v4_cachep, nxa);
20617 +}
20618 +
20619 +/*     __dealloc_nx_addr_v4_all()                              */
20620 +
20621 +static inline void __dealloc_nx_addr_v4_all(struct nx_addr_v4 *nxa)
20622 +{
20623 +       while (nxa) {
20624 +               struct nx_addr_v4 *next = nxa->next;
20625 +
20626 +               __dealloc_nx_addr_v4(nxa);
20627 +               nxa = next;
20628 +       }
20629 +}
20630 +
20631 +
20632 +#ifdef CONFIG_IPV6
20633 +
20634 +/*     __alloc_nx_addr_v6()                                    */
20635 +
20636 +static inline struct nx_addr_v6 *__alloc_nx_addr_v6(void)
20637 +{
20638 +       struct nx_addr_v6 *nxa = kmem_cache_alloc(
20639 +               nx_addr_v6_cachep, GFP_KERNEL);
20640 +
20641 +       if (!IS_ERR(nxa))
20642 +               memset(nxa, 0, sizeof(*nxa));
20643 +       return nxa;
20644 +}
20645 +
20646 +/*     __dealloc_nx_addr_v6()                                  */
20647 +
20648 +static inline void __dealloc_nx_addr_v6(struct nx_addr_v6 *nxa)
20649 +{
20650 +       kmem_cache_free(nx_addr_v6_cachep, nxa);
20651 +}
20652 +
20653 +/*     __dealloc_nx_addr_v6_all()                              */
20654 +
20655 +static inline void __dealloc_nx_addr_v6_all(struct nx_addr_v6 *nxa)
20656 +{
20657 +       while (nxa) {
20658 +               struct nx_addr_v6 *next = nxa->next;
20659 +
20660 +               __dealloc_nx_addr_v6(nxa);
20661 +               nxa = next;
20662 +       }
20663 +}
20664 +
20665 +#endif /* CONFIG_IPV6 */
20666 +
20667 +/*     __alloc_nx_info()
20668 +
20669 +       * allocate an initialized nx_info struct
20670 +       * doesn't make it visible (hash)                        */
20671 +
20672 +static struct nx_info *__alloc_nx_info(nid_t nid)
20673 +{
20674 +       struct nx_info *new = NULL;
20675 +
20676 +       vxdprintk(VXD_CBIT(nid, 1), "alloc_nx_info(%d)*", nid);
20677 +
20678 +       /* would this benefit from a slab cache? */
20679 +       new = kmalloc(sizeof(struct nx_info), GFP_KERNEL);
20680 +       if (!new)
20681 +               return 0;
20682 +
20683 +       memset(new, 0, sizeof(struct nx_info));
20684 +       new->nx_id = nid;
20685 +       INIT_HLIST_NODE(&new->nx_hlist);
20686 +       atomic_set(&new->nx_usecnt, 0);
20687 +       atomic_set(&new->nx_tasks, 0);
20688 +       new->nx_state = 0;
20689 +
20690 +       new->nx_flags = NXF_INIT_SET;
20691 +
20692 +       /* rest of init goes here */
20693 +
20694 +       new->v4_lback.s_addr = htonl(INADDR_LOOPBACK);
20695 +       new->v4_bcast.s_addr = htonl(INADDR_BROADCAST);
20696 +
20697 +       vxdprintk(VXD_CBIT(nid, 0),
20698 +               "alloc_nx_info(%d) = %p", nid, new);
20699 +       atomic_inc(&nx_global_ctotal);
20700 +       return new;
20701 +}
20702 +
20703 +/*     __dealloc_nx_info()
20704 +
20705 +       * final disposal of nx_info                             */
20706 +
20707 +static void __dealloc_nx_info(struct nx_info *nxi)
20708 +{
20709 +       vxdprintk(VXD_CBIT(nid, 0),
20710 +               "dealloc_nx_info(%p)", nxi);
20711 +
20712 +       nxi->nx_hlist.next = LIST_POISON1;
20713 +       nxi->nx_id = -1;
20714 +
20715 +       BUG_ON(atomic_read(&nxi->nx_usecnt));
20716 +       BUG_ON(atomic_read(&nxi->nx_tasks));
20717 +
20718 +       __dealloc_nx_addr_v4_all(nxi->v4.next);
20719 +
20720 +       nxi->nx_state |= NXS_RELEASED;
20721 +       kfree(nxi);
20722 +       atomic_dec(&nx_global_ctotal);
20723 +}
20724 +
20725 +static void __shutdown_nx_info(struct nx_info *nxi)
20726 +{
20727 +       nxi->nx_state |= NXS_SHUTDOWN;
20728 +       vs_net_change(nxi, VSC_NETDOWN);
20729 +}
20730 +
20731 +/*     exported stuff                                          */
20732 +
20733 +void free_nx_info(struct nx_info *nxi)
20734 +{
20735 +       /* context shutdown is mandatory */
20736 +       BUG_ON(nxi->nx_state != NXS_SHUTDOWN);
20737 +
20738 +       /* context must not be hashed */
20739 +       BUG_ON(nxi->nx_state & NXS_HASHED);
20740 +
20741 +       BUG_ON(atomic_read(&nxi->nx_usecnt));
20742 +       BUG_ON(atomic_read(&nxi->nx_tasks));
20743 +
20744 +       __dealloc_nx_info(nxi);
20745 +}
20746 +
20747 +
20748 +void __nx_set_lback(struct nx_info *nxi)
20749 +{
20750 +       int nid = nxi->nx_id;
20751 +       __be32 lback = htonl(INADDR_LOOPBACK ^ ((nid & 0xFFFF) << 8));
20752 +
20753 +       nxi->v4_lback.s_addr = lback;
20754 +}
20755 +
20756 +extern int __nx_inet_add_lback(__be32 addr);
20757 +extern int __nx_inet_del_lback(__be32 addr);
20758 +
20759 +
20760 +/*     hash table for nx_info hash */
20761 +
20762 +#define NX_HASH_SIZE   13
20763 +
20764 +struct hlist_head nx_info_hash[NX_HASH_SIZE];
20765 +
20766 +static DEFINE_SPINLOCK(nx_info_hash_lock);
20767 +
20768 +
20769 +static inline unsigned int __hashval(nid_t nid)
20770 +{
20771 +       return (nid % NX_HASH_SIZE);
20772 +}
20773 +
20774 +
20775 +
20776 +/*     __hash_nx_info()
20777 +
20778 +       * add the nxi to the global hash table
20779 +       * requires the hash_lock to be held                     */
20780 +
20781 +static inline void __hash_nx_info(struct nx_info *nxi)
20782 +{
20783 +       struct hlist_head *head;
20784 +
20785 +       vxd_assert_lock(&nx_info_hash_lock);
20786 +       vxdprintk(VXD_CBIT(nid, 4),
20787 +               "__hash_nx_info: %p[#%d]", nxi, nxi->nx_id);
20788 +
20789 +       /* context must not be hashed */
20790 +       BUG_ON(nx_info_state(nxi, NXS_HASHED));
20791 +
20792 +       nxi->nx_state |= NXS_HASHED;
20793 +       head = &nx_info_hash[__hashval(nxi->nx_id)];
20794 +       hlist_add_head(&nxi->nx_hlist, head);
20795 +       atomic_inc(&nx_global_cactive);
20796 +}
20797 +
20798 +/*     __unhash_nx_info()
20799 +
20800 +       * remove the nxi from the global hash table
20801 +       * requires the hash_lock to be held                     */
20802 +
20803 +static inline void __unhash_nx_info(struct nx_info *nxi)
20804 +{
20805 +       vxd_assert_lock(&nx_info_hash_lock);
20806 +       vxdprintk(VXD_CBIT(nid, 4),
20807 +               "__unhash_nx_info: %p[#%d.%d.%d]", nxi, nxi->nx_id,
20808 +               atomic_read(&nxi->nx_usecnt), atomic_read(&nxi->nx_tasks));
20809 +
20810 +       /* context must be hashed */
20811 +       BUG_ON(!nx_info_state(nxi, NXS_HASHED));
20812 +       /* but without tasks */
20813 +       BUG_ON(atomic_read(&nxi->nx_tasks));
20814 +
20815 +       nxi->nx_state &= ~NXS_HASHED;
20816 +       hlist_del(&nxi->nx_hlist);
20817 +       atomic_dec(&nx_global_cactive);
20818 +}
20819 +
20820 +
20821 +/*     __lookup_nx_info()
20822 +
20823 +       * requires the hash_lock to be held
20824 +       * doesn't increment the nx_refcnt                       */
20825 +
20826 +static inline struct nx_info *__lookup_nx_info(nid_t nid)
20827 +{
20828 +       struct hlist_head *head = &nx_info_hash[__hashval(nid)];
20829 +       struct hlist_node *pos;
20830 +       struct nx_info *nxi;
20831 +
20832 +       vxd_assert_lock(&nx_info_hash_lock);
20833 +       hlist_for_each(pos, head) {
20834 +               nxi = hlist_entry(pos, struct nx_info, nx_hlist);
20835 +
20836 +               if (nxi->nx_id == nid)
20837 +                       goto found;
20838 +       }
20839 +       nxi = NULL;
20840 +found:
20841 +       vxdprintk(VXD_CBIT(nid, 0),
20842 +               "__lookup_nx_info(#%u): %p[#%u]",
20843 +               nid, nxi, nxi ? nxi->nx_id : 0);
20844 +       return nxi;
20845 +}
20846 +
20847 +
20848 +/*     __create_nx_info()
20849 +
20850 +       * create the requested context
20851 +       * get(), claim() and hash it                            */
20852 +
20853 +static struct nx_info *__create_nx_info(int id)
20854 +{
20855 +       struct nx_info *new, *nxi = NULL;
20856 +
20857 +       vxdprintk(VXD_CBIT(nid, 1), "create_nx_info(%d)*", id);
20858 +
20859 +       if (!(new = __alloc_nx_info(id)))
20860 +               return ERR_PTR(-ENOMEM);
20861 +
20862 +       /* required to make dynamic xids unique */
20863 +       spin_lock(&nx_info_hash_lock);
20864 +
20865 +       /* static context requested */
20866 +       if ((nxi = __lookup_nx_info(id))) {
20867 +               vxdprintk(VXD_CBIT(nid, 0),
20868 +                       "create_nx_info(%d) = %p (already there)", id, nxi);
20869 +               if (nx_info_flags(nxi, NXF_STATE_SETUP, 0))
20870 +                       nxi = ERR_PTR(-EBUSY);
20871 +               else
20872 +                       nxi = ERR_PTR(-EEXIST);
20873 +               goto out_unlock;
20874 +       }
20875 +       /* new context */
20876 +       vxdprintk(VXD_CBIT(nid, 0),
20877 +               "create_nx_info(%d) = %p (new)", id, new);
20878 +       claim_nx_info(new, NULL);
20879 +       __nx_set_lback(new);
20880 +       __hash_nx_info(get_nx_info(new));
20881 +       nxi = new, new = NULL;
20882 +
20883 +out_unlock:
20884 +       spin_unlock(&nx_info_hash_lock);
20885 +       if (new)
20886 +               __dealloc_nx_info(new);
20887 +       return nxi;
20888 +}
20889 +
20890 +
20891 +
20892 +/*     exported stuff                                          */
20893 +
20894 +
20895 +void unhash_nx_info(struct nx_info *nxi)
20896 +{
20897 +       __shutdown_nx_info(nxi);
20898 +       spin_lock(&nx_info_hash_lock);
20899 +       __unhash_nx_info(nxi);
20900 +       spin_unlock(&nx_info_hash_lock);
20901 +}
20902 +
20903 +/*     lookup_nx_info()
20904 +
20905 +       * search for a nx_info and get() it
20906 +       * negative id means current                             */
20907 +
20908 +struct nx_info *lookup_nx_info(int id)
20909 +{
20910 +       struct nx_info *nxi = NULL;
20911 +
20912 +       if (id < 0) {
20913 +               nxi = get_nx_info(current_nx_info());
20914 +       } else if (id > 1) {
20915 +               spin_lock(&nx_info_hash_lock);
20916 +               nxi = get_nx_info(__lookup_nx_info(id));
20917 +               spin_unlock(&nx_info_hash_lock);
20918 +       }
20919 +       return nxi;
20920 +}
20921 +
20922 +/*     nid_is_hashed()
20923 +
20924 +       * verify that nid is still hashed                       */
20925 +
20926 +int nid_is_hashed(nid_t nid)
20927 +{
20928 +       int hashed;
20929 +
20930 +       spin_lock(&nx_info_hash_lock);
20931 +       hashed = (__lookup_nx_info(nid) != NULL);
20932 +       spin_unlock(&nx_info_hash_lock);
20933 +       return hashed;
20934 +}
20935 +
20936 +
20937 +#ifdef CONFIG_PROC_FS
20938 +
20939 +/*     get_nid_list()
20940 +
20941 +       * get a subset of hashed nids for proc
20942 +       * assumes size is at least one                          */
20943 +
20944 +int get_nid_list(int index, unsigned int *nids, int size)
20945 +{
20946 +       int hindex, nr_nids = 0;
20947 +
20948 +       /* only show current and children */
20949 +       if (!nx_check(0, VS_ADMIN | VS_WATCH)) {
20950 +               if (index > 0)
20951 +                       return 0;
20952 +               nids[nr_nids] = nx_current_nid();
20953 +               return 1;
20954 +       }
20955 +
20956 +       for (hindex = 0; hindex < NX_HASH_SIZE; hindex++) {
20957 +               struct hlist_head *head = &nx_info_hash[hindex];
20958 +               struct hlist_node *pos;
20959 +
20960 +               spin_lock(&nx_info_hash_lock);
20961 +               hlist_for_each(pos, head) {
20962 +                       struct nx_info *nxi;
20963 +
20964 +                       if (--index > 0)
20965 +                               continue;
20966 +
20967 +                       nxi = hlist_entry(pos, struct nx_info, nx_hlist);
20968 +                       nids[nr_nids] = nxi->nx_id;
20969 +                       if (++nr_nids >= size) {
20970 +                               spin_unlock(&nx_info_hash_lock);
20971 +                               goto out;
20972 +                       }
20973 +               }
20974 +               /* keep the lock time short */
20975 +               spin_unlock(&nx_info_hash_lock);
20976 +       }
20977 +out:
20978 +       return nr_nids;
20979 +}
20980 +#endif
20981 +
20982 +
20983 +/*
20984 + *     migrate task to new network
20985 + *     gets nxi, puts old_nxi on change
20986 + */
20987 +
20988 +int nx_migrate_task(struct task_struct *p, struct nx_info *nxi)
20989 +{
20990 +       struct nx_info *old_nxi;
20991 +       int ret = 0;
20992 +
20993 +       if (!p || !nxi)
20994 +               BUG();
20995 +
20996 +       vxdprintk(VXD_CBIT(nid, 5),
20997 +               "nx_migrate_task(%p,%p[#%d.%d.%d])",
20998 +               p, nxi, nxi->nx_id,
20999 +               atomic_read(&nxi->nx_usecnt),
21000 +               atomic_read(&nxi->nx_tasks));
21001 +
21002 +       if (nx_info_flags(nxi, NXF_INFO_PRIVATE, 0) &&
21003 +               !nx_info_flags(nxi, NXF_STATE_SETUP, 0))
21004 +               return -EACCES;
21005 +
21006 +       if (nx_info_state(nxi, NXS_SHUTDOWN))
21007 +               return -EFAULT;
21008 +
21009 +       /* maybe disallow this completely? */
21010 +       old_nxi = task_get_nx_info(p);
21011 +       if (old_nxi == nxi)
21012 +               goto out;
21013 +
21014 +       task_lock(p);
21015 +       if (old_nxi)
21016 +               clr_nx_info(&p->nx_info);
21017 +       claim_nx_info(nxi, p);
21018 +       set_nx_info(&p->nx_info, nxi);
21019 +       p->nid = nxi->nx_id;
21020 +       task_unlock(p);
21021 +
21022 +       vxdprintk(VXD_CBIT(nid, 5),
21023 +               "moved task %p into nxi:%p[#%d]",
21024 +               p, nxi, nxi->nx_id);
21025 +
21026 +       if (old_nxi)
21027 +               release_nx_info(old_nxi, p);
21028 +       ret = 0;
21029 +out:
21030 +       put_nx_info(old_nxi);
21031 +       return ret;
21032 +}
21033 +
21034 +
21035 +void nx_set_persistent(struct nx_info *nxi)
21036 +{
21037 +       vxdprintk(VXD_CBIT(nid, 6),
21038 +               "nx_set_persistent(%p[#%d])", nxi, nxi->nx_id);
21039 +
21040 +       get_nx_info(nxi);
21041 +       claim_nx_info(nxi, NULL);
21042 +}
21043 +
21044 +void nx_clear_persistent(struct nx_info *nxi)
21045 +{
21046 +       vxdprintk(VXD_CBIT(nid, 6),
21047 +               "nx_clear_persistent(%p[#%d])", nxi, nxi->nx_id);
21048 +
21049 +       release_nx_info(nxi, NULL);
21050 +       put_nx_info(nxi);
21051 +}
21052 +
21053 +void nx_update_persistent(struct nx_info *nxi)
21054 +{
21055 +       if (nx_info_flags(nxi, NXF_PERSISTENT, 0))
21056 +               nx_set_persistent(nxi);
21057 +       else
21058 +               nx_clear_persistent(nxi);
21059 +}
21060 +
21061 +/* vserver syscall commands below here */
21062 +
21063 +/* taks nid and nx_info functions */
21064 +
21065 +#include <asm/uaccess.h>
21066 +
21067 +
21068 +int vc_task_nid(uint32_t id)
21069 +{
21070 +       nid_t nid;
21071 +
21072 +       if (id) {
21073 +               struct task_struct *tsk;
21074 +
21075 +               rcu_read_lock();
21076 +               tsk = find_task_by_real_pid(id);
21077 +               nid = (tsk) ? tsk->nid : -ESRCH;
21078 +               rcu_read_unlock();
21079 +       } else
21080 +               nid = nx_current_nid();
21081 +       return nid;
21082 +}
21083 +
21084 +
21085 +int vc_nx_info(struct nx_info *nxi, void __user *data)
21086 +{
21087 +       struct vcmd_nx_info_v0 vc_data;
21088 +
21089 +       vc_data.nid = nxi->nx_id;
21090 +
21091 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
21092 +               return -EFAULT;
21093 +       return 0;
21094 +}
21095 +
21096 +
21097 +/* network functions */
21098 +
21099 +int vc_net_create(uint32_t nid, void __user *data)
21100 +{
21101 +       struct vcmd_net_create vc_data = { .flagword = NXF_INIT_SET };
21102 +       struct nx_info *new_nxi;
21103 +       int ret;
21104 +
21105 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
21106 +               return -EFAULT;
21107 +
21108 +       if ((nid > MAX_S_CONTEXT) || (nid < 2))
21109 +               return -EINVAL;
21110 +
21111 +       new_nxi = __create_nx_info(nid);
21112 +       if (IS_ERR(new_nxi))
21113 +               return PTR_ERR(new_nxi);
21114 +
21115 +       /* initial flags */
21116 +       new_nxi->nx_flags = vc_data.flagword;
21117 +
21118 +       ret = -ENOEXEC;
21119 +       if (vs_net_change(new_nxi, VSC_NETUP))
21120 +               goto out;
21121 +
21122 +       ret = nx_migrate_task(current, new_nxi);
21123 +       if (ret)
21124 +               goto out;
21125 +
21126 +       /* return context id on success */
21127 +       ret = new_nxi->nx_id;
21128 +
21129 +       /* get a reference for persistent contexts */
21130 +       if ((vc_data.flagword & NXF_PERSISTENT))
21131 +               nx_set_persistent(new_nxi);
21132 +out:
21133 +       release_nx_info(new_nxi, NULL);
21134 +       put_nx_info(new_nxi);
21135 +       return ret;
21136 +}
21137 +
21138 +
21139 +int vc_net_migrate(struct nx_info *nxi, void __user *data)
21140 +{
21141 +       return nx_migrate_task(current, nxi);
21142 +}
21143 +
21144 +
21145 +
21146 +int do_add_v4_addr(struct nx_info *nxi, __be32 ip, __be32 ip2, __be32 mask,
21147 +       uint16_t type, uint16_t flags)
21148 +{
21149 +       struct nx_addr_v4 *nxa = &nxi->v4;
21150 +
21151 +       if (NX_IPV4(nxi)) {
21152 +               /* locate last entry */
21153 +               for (; nxa->next; nxa = nxa->next);
21154 +               nxa->next = __alloc_nx_addr_v4();
21155 +               nxa = nxa->next;
21156 +
21157 +               if (IS_ERR(nxa))
21158 +                       return PTR_ERR(nxa);
21159 +       }
21160 +
21161 +       if (nxi->v4.next)
21162 +               /* remove single ip for ip list */
21163 +               nxi->nx_flags &= ~NXF_SINGLE_IP;
21164 +
21165 +       nxa->ip[0].s_addr = ip;
21166 +       nxa->ip[1].s_addr = ip2;
21167 +       nxa->mask.s_addr = mask;
21168 +       nxa->type = type;
21169 +       nxa->flags = flags;
21170 +       return 0;
21171 +}
21172 +
21173 +int do_remove_v4_addr(struct nx_info *nxi, __be32 ip, __be32 ip2, __be32 mask,
21174 +       uint16_t type, uint16_t flags)
21175 +{
21176 +       struct nx_addr_v4 *nxa = &nxi->v4;
21177 +
21178 +       switch (type) {
21179 +/*     case NXA_TYPE_ADDR:
21180 +               break;          */
21181 +
21182 +       case NXA_TYPE_ANY:
21183 +               __dealloc_nx_addr_v4_all(xchg(&nxa->next, NULL));
21184 +               memset(nxa, 0, sizeof(*nxa));
21185 +               break;
21186 +
21187 +       default:
21188 +               return -EINVAL;
21189 +       }
21190 +       return 0;
21191 +}
21192 +
21193 +
21194 +int vc_net_add(struct nx_info *nxi, void __user *data)
21195 +{
21196 +       struct vcmd_net_addr_v0 vc_data;
21197 +       int index, ret = 0;
21198 +
21199 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
21200 +               return -EFAULT;
21201 +
21202 +       switch (vc_data.type) {
21203 +       case NXA_TYPE_IPV4:
21204 +               if ((vc_data.count < 1) || (vc_data.count > 4))
21205 +                       return -EINVAL;
21206 +
21207 +               index = 0;
21208 +               while (index < vc_data.count) {
21209 +                       ret = do_add_v4_addr(nxi, vc_data.ip[index].s_addr, 0,
21210 +                               vc_data.mask[index].s_addr, NXA_TYPE_ADDR, 0);
21211 +                       if (ret)
21212 +                               return ret;
21213 +                       index++;
21214 +               }
21215 +               ret = index;
21216 +               break;
21217 +
21218 +       case NXA_TYPE_IPV4|NXA_MOD_BCAST:
21219 +               nxi->v4_bcast = vc_data.ip[0];
21220 +               ret = 1;
21221 +               break;
21222 +
21223 +       case NXA_TYPE_IPV4|NXA_MOD_LBACK:
21224 +               nxi->v4_lback = vc_data.ip[0];
21225 +               ret = 1;
21226 +               break;
21227 +
21228 +       default:
21229 +               ret = -EINVAL;
21230 +               break;
21231 +       }
21232 +       return ret;
21233 +}
21234 +
21235 +int vc_net_remove(struct nx_info *nxi, void __user *data)
21236 +{
21237 +       struct vcmd_net_addr_v0 vc_data;
21238 +
21239 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
21240 +               return -EFAULT;
21241 +
21242 +       switch (vc_data.type) {
21243 +       case NXA_TYPE_ANY:
21244 +               __dealloc_nx_addr_v4_all(xchg(&nxi->v4.next, NULL));
21245 +               memset(&nxi->v4, 0, sizeof(nxi->v4));
21246 +               break;
21247 +
21248 +       default:
21249 +               return -EINVAL;
21250 +       }
21251 +       return 0;
21252 +}
21253 +
21254 +
21255 +int vc_net_add_ipv4_v1(struct nx_info *nxi, void __user *data)
21256 +{
21257 +       struct vcmd_net_addr_ipv4_v1 vc_data;
21258 +
21259 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
21260 +               return -EFAULT;
21261 +
21262 +       switch (vc_data.type) {
21263 +       case NXA_TYPE_ADDR:
21264 +       case NXA_TYPE_MASK:
21265 +               return do_add_v4_addr(nxi, vc_data.ip.s_addr, 0,
21266 +                       vc_data.mask.s_addr, vc_data.type, vc_data.flags);
21267 +
21268 +       case NXA_TYPE_ADDR | NXA_MOD_BCAST:
21269 +               nxi->v4_bcast = vc_data.ip;
21270 +               break;
21271 +
21272 +       case NXA_TYPE_ADDR | NXA_MOD_LBACK:
21273 +               nxi->v4_lback = vc_data.ip;
21274 +               break;
21275 +
21276 +       default:
21277 +               return -EINVAL;
21278 +       }
21279 +       return 0;
21280 +}
21281 +
21282 +int vc_net_add_ipv4(struct nx_info *nxi, void __user *data)
21283 +{
21284 +       struct vcmd_net_addr_ipv4_v2 vc_data;
21285 +
21286 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
21287 +               return -EFAULT;
21288 +
21289 +       switch (vc_data.type) {
21290 +       case NXA_TYPE_ADDR:
21291 +       case NXA_TYPE_MASK:
21292 +       case NXA_TYPE_RANGE:
21293 +               return do_add_v4_addr(nxi, vc_data.ip.s_addr, vc_data.ip2.s_addr,
21294 +                       vc_data.mask.s_addr, vc_data.type, vc_data.flags);
21295 +
21296 +       case NXA_TYPE_ADDR | NXA_MOD_BCAST:
21297 +               nxi->v4_bcast = vc_data.ip;
21298 +               break;
21299 +
21300 +       case NXA_TYPE_ADDR | NXA_MOD_LBACK:
21301 +               nxi->v4_lback = vc_data.ip;
21302 +               break;
21303 +
21304 +       default:
21305 +               return -EINVAL;
21306 +       }
21307 +       return 0;
21308 +}
21309 +
21310 +int vc_net_rem_ipv4_v1(struct nx_info *nxi, void __user *data)
21311 +{
21312 +       struct vcmd_net_addr_ipv4_v1 vc_data;
21313 +
21314 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
21315 +               return -EFAULT;
21316 +
21317 +       return do_remove_v4_addr(nxi, vc_data.ip.s_addr, 0,
21318 +               vc_data.mask.s_addr, vc_data.type, vc_data.flags);
21319 +}
21320 +
21321 +int vc_net_rem_ipv4(struct nx_info *nxi, void __user *data)
21322 +{
21323 +       struct vcmd_net_addr_ipv4_v2 vc_data;
21324 +
21325 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
21326 +               return -EFAULT;
21327 +
21328 +       return do_remove_v4_addr(nxi, vc_data.ip.s_addr, vc_data.ip2.s_addr,
21329 +               vc_data.mask.s_addr, vc_data.type, vc_data.flags);
21330 +}
21331 +
21332 +#ifdef CONFIG_IPV6
21333 +
21334 +int do_add_v6_addr(struct nx_info *nxi,
21335 +       struct in6_addr *ip, struct in6_addr *mask,
21336 +       uint32_t prefix, uint16_t type, uint16_t flags)
21337 +{
21338 +       struct nx_addr_v6 *nxa = &nxi->v6;
21339 +
21340 +       if (NX_IPV6(nxi)) {
21341 +               /* locate last entry */
21342 +               for (; nxa->next; nxa = nxa->next);
21343 +               nxa->next = __alloc_nx_addr_v6();
21344 +               nxa = nxa->next;
21345 +
21346 +               if (IS_ERR(nxa))
21347 +                       return PTR_ERR(nxa);
21348 +       }
21349 +
21350 +       nxa->ip = *ip;
21351 +       nxa->mask = *mask;
21352 +       nxa->prefix = prefix;
21353 +       nxa->type = type;
21354 +       nxa->flags = flags;
21355 +       return 0;
21356 +}
21357 +
21358 +
21359 +int vc_net_add_ipv6(struct nx_info *nxi, void __user *data)
21360 +{
21361 +       struct vcmd_net_addr_ipv6_v1 vc_data;
21362 +
21363 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
21364 +               return -EFAULT;
21365 +
21366 +       switch (vc_data.type) {
21367 +       case NXA_TYPE_ADDR:
21368 +               memset(&vc_data.mask, ~0, sizeof(vc_data.mask));
21369 +               /* fallthrough */
21370 +       case NXA_TYPE_MASK:
21371 +               return do_add_v6_addr(nxi, &vc_data.ip, &vc_data.mask,
21372 +                       vc_data.prefix, vc_data.type, vc_data.flags);
21373 +       default:
21374 +               return -EINVAL;
21375 +       }
21376 +       return 0;
21377 +}
21378 +
21379 +int vc_net_remove_ipv6(struct nx_info *nxi, void __user *data)
21380 +{
21381 +       struct vcmd_net_addr_ipv6_v1 vc_data;
21382 +
21383 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
21384 +               return -EFAULT;
21385 +
21386 +       switch (vc_data.type) {
21387 +       case NXA_TYPE_ANY:
21388 +               __dealloc_nx_addr_v6_all(xchg(&nxi->v6.next, NULL));
21389 +               memset(&nxi->v6, 0, sizeof(nxi->v6));
21390 +               break;
21391 +
21392 +       default:
21393 +               return -EINVAL;
21394 +       }
21395 +       return 0;
21396 +}
21397 +
21398 +#endif /* CONFIG_IPV6 */
21399 +
21400 +
21401 +int vc_get_nflags(struct nx_info *nxi, void __user *data)
21402 +{
21403 +       struct vcmd_net_flags_v0 vc_data;
21404 +
21405 +       vc_data.flagword = nxi->nx_flags;
21406 +
21407 +       /* special STATE flag handling */
21408 +       vc_data.mask = vs_mask_flags(~0ULL, nxi->nx_flags, NXF_ONE_TIME);
21409 +
21410 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
21411 +               return -EFAULT;
21412 +       return 0;
21413 +}
21414 +
21415 +int vc_set_nflags(struct nx_info *nxi, void __user *data)
21416 +{
21417 +       struct vcmd_net_flags_v0 vc_data;
21418 +       uint64_t mask, trigger;
21419 +
21420 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
21421 +               return -EFAULT;
21422 +
21423 +       /* special STATE flag handling */
21424 +       mask = vs_mask_mask(vc_data.mask, nxi->nx_flags, NXF_ONE_TIME);
21425 +       trigger = (mask & nxi->nx_flags) ^ (mask & vc_data.flagword);
21426 +
21427 +       nxi->nx_flags = vs_mask_flags(nxi->nx_flags,
21428 +               vc_data.flagword, mask);
21429 +       if (trigger & NXF_PERSISTENT)
21430 +               nx_update_persistent(nxi);
21431 +
21432 +       return 0;
21433 +}
21434 +
21435 +int vc_get_ncaps(struct nx_info *nxi, void __user *data)
21436 +{
21437 +       struct vcmd_net_caps_v0 vc_data;
21438 +
21439 +       vc_data.ncaps = nxi->nx_ncaps;
21440 +       vc_data.cmask = ~0ULL;
21441 +
21442 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
21443 +               return -EFAULT;
21444 +       return 0;
21445 +}
21446 +
21447 +int vc_set_ncaps(struct nx_info *nxi, void __user *data)
21448 +{
21449 +       struct vcmd_net_caps_v0 vc_data;
21450 +
21451 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
21452 +               return -EFAULT;
21453 +
21454 +       nxi->nx_ncaps = vs_mask_flags(nxi->nx_ncaps,
21455 +               vc_data.ncaps, vc_data.cmask);
21456 +       return 0;
21457 +}
21458 +
21459 +
21460 +#include <linux/module.h>
21461 +
21462 +module_init(init_network);
21463 +
21464 +EXPORT_SYMBOL_GPL(free_nx_info);
21465 +EXPORT_SYMBOL_GPL(unhash_nx_info);
21466 +
21467 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/proc.c linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/proc.c
21468 --- linux-3.0-rc5/kernel/vserver/proc.c 1970-01-01 01:00:00.000000000 +0100
21469 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/proc.c    2011-06-10 22:11:24.000000000 +0200
21470 @@ -0,0 +1,1100 @@
21471 +/*
21472 + *  linux/kernel/vserver/proc.c
21473 + *
21474 + *  Virtual Context Support
21475 + *
21476 + *  Copyright (C) 2003-2007  Herbert Pötzl
21477 + *
21478 + *  V0.01  basic structure
21479 + *  V0.02  adaptation vs1.3.0
21480 + *  V0.03  proc permissions
21481 + *  V0.04  locking/generic
21482 + *  V0.05  next generation procfs
21483 + *  V0.06  inode validation
21484 + *  V0.07  generic rewrite vid
21485 + *  V0.08  remove inode type
21486 + *
21487 + */
21488 +
21489 +#include <linux/proc_fs.h>
21490 +#include <linux/fs_struct.h>
21491 +#include <linux/mount.h>
21492 +#include <asm/unistd.h>
21493 +
21494 +#include <linux/vs_context.h>
21495 +#include <linux/vs_network.h>
21496 +#include <linux/vs_cvirt.h>
21497 +
21498 +#include <linux/in.h>
21499 +#include <linux/inetdevice.h>
21500 +#include <linux/vs_inet.h>
21501 +#include <linux/vs_inet6.h>
21502 +
21503 +#include <linux/vserver/global.h>
21504 +
21505 +#include "cvirt_proc.h"
21506 +#include "cacct_proc.h"
21507 +#include "limit_proc.h"
21508 +#include "sched_proc.h"
21509 +#include "vci_config.h"
21510 +
21511 +
21512 +static inline char *print_cap_t(char *buffer, kernel_cap_t *c)
21513 +{
21514 +       unsigned __capi;
21515 +
21516 +       CAP_FOR_EACH_U32(__capi) {
21517 +               buffer += sprintf(buffer, "%08x",
21518 +                       c->cap[(_KERNEL_CAPABILITY_U32S-1) - __capi]);
21519 +       }
21520 +       return buffer;
21521 +}
21522 +
21523 +
21524 +static struct proc_dir_entry *proc_virtual;
21525 +
21526 +static struct proc_dir_entry *proc_virtnet;
21527 +
21528 +
21529 +/* first the actual feeds */
21530 +
21531 +
21532 +static int proc_vci(char *buffer)
21533 +{
21534 +       return sprintf(buffer,
21535 +               "VCIVersion:\t%04x:%04x\n"
21536 +               "VCISyscall:\t%d\n"
21537 +               "VCIKernel:\t%08x\n",
21538 +               VCI_VERSION >> 16,
21539 +               VCI_VERSION & 0xFFFF,
21540 +               __NR_vserver,
21541 +               vci_kernel_config());
21542 +}
21543 +
21544 +static int proc_virtual_info(char *buffer)
21545 +{
21546 +       return proc_vci(buffer);
21547 +}
21548 +
21549 +static int proc_virtual_status(char *buffer)
21550 +{
21551 +       return sprintf(buffer,
21552 +               "#CTotal:\t%d\n"
21553 +               "#CActive:\t%d\n"
21554 +               "#NSProxy:\t%d\t%d %d %d %d %d %d\n"
21555 +               "#InitTask:\t%d\t%d %d\n",
21556 +               atomic_read(&vx_global_ctotal),
21557 +               atomic_read(&vx_global_cactive),
21558 +               atomic_read(&vs_global_nsproxy),
21559 +               atomic_read(&vs_global_fs),
21560 +               atomic_read(&vs_global_mnt_ns),
21561 +               atomic_read(&vs_global_uts_ns),
21562 +               atomic_read(&nr_ipc_ns),
21563 +               atomic_read(&vs_global_user_ns),
21564 +               atomic_read(&vs_global_pid_ns),
21565 +               atomic_read(&init_task.usage),
21566 +               atomic_read(&init_task.nsproxy->count),
21567 +               init_task.fs->users);
21568 +}
21569 +
21570 +
21571 +int proc_vxi_info(struct vx_info *vxi, char *buffer)
21572 +{
21573 +       int length;
21574 +
21575 +       length = sprintf(buffer,
21576 +               "ID:\t%d\n"
21577 +               "Info:\t%p\n"
21578 +               "Init:\t%d\n"
21579 +               "OOM:\t%lld\n",
21580 +               vxi->vx_id,
21581 +               vxi,
21582 +               vxi->vx_initpid,
21583 +               vxi->vx_badness_bias);
21584 +       return length;
21585 +}
21586 +
21587 +int proc_vxi_status(struct vx_info *vxi, char *buffer)
21588 +{
21589 +       char *orig = buffer;
21590 +
21591 +       buffer += sprintf(buffer,
21592 +               "UseCnt:\t%d\n"
21593 +               "Tasks:\t%d\n"
21594 +               "Flags:\t%016llx\n",
21595 +               atomic_read(&vxi->vx_usecnt),
21596 +               atomic_read(&vxi->vx_tasks),
21597 +               (unsigned long long)vxi->vx_flags);
21598 +
21599 +       buffer += sprintf(buffer, "BCaps:\t");
21600 +       buffer = print_cap_t(buffer, &vxi->vx_bcaps);
21601 +       buffer += sprintf(buffer, "\n");
21602 +
21603 +       buffer += sprintf(buffer,
21604 +               "CCaps:\t%016llx\n"
21605 +               "Umask:\t%16llx\n"
21606 +               "Spaces:\t%08lx %08lx\n",
21607 +               (unsigned long long)vxi->vx_ccaps,
21608 +               (unsigned long long)vxi->vx_umask,
21609 +               vxi->space[0].vx_nsmask, vxi->space[1].vx_nsmask);
21610 +       return buffer - orig;
21611 +}
21612 +
21613 +int proc_vxi_limit(struct vx_info *vxi, char *buffer)
21614 +{
21615 +       return vx_info_proc_limit(&vxi->limit, buffer);
21616 +}
21617 +
21618 +int proc_vxi_sched(struct vx_info *vxi, char *buffer)
21619 +{
21620 +       int cpu, length;
21621 +
21622 +       length = vx_info_proc_sched(&vxi->sched, buffer);
21623 +       for_each_online_cpu(cpu) {
21624 +               length += vx_info_proc_sched_pc(
21625 +                       &vx_per_cpu(vxi, sched_pc, cpu),
21626 +                       buffer + length, cpu);
21627 +       }
21628 +       return length;
21629 +}
21630 +
21631 +int proc_vxi_nsproxy0(struct vx_info *vxi, char *buffer)
21632 +{
21633 +       return vx_info_proc_nsproxy(vxi->space[0].vx_nsproxy, buffer);
21634 +}
21635 +
21636 +int proc_vxi_nsproxy1(struct vx_info *vxi, char *buffer)
21637 +{
21638 +       return vx_info_proc_nsproxy(vxi->space[1].vx_nsproxy, buffer);
21639 +}
21640 +
21641 +int proc_vxi_cvirt(struct vx_info *vxi, char *buffer)
21642 +{
21643 +       int cpu, length;
21644 +
21645 +       vx_update_load(vxi);
21646 +       length = vx_info_proc_cvirt(&vxi->cvirt, buffer);
21647 +       for_each_online_cpu(cpu) {
21648 +               length += vx_info_proc_cvirt_pc(
21649 +                       &vx_per_cpu(vxi, cvirt_pc, cpu),
21650 +                       buffer + length, cpu);
21651 +       }
21652 +       return length;
21653 +}
21654 +
21655 +int proc_vxi_cacct(struct vx_info *vxi, char *buffer)
21656 +{
21657 +       return vx_info_proc_cacct(&vxi->cacct, buffer);
21658 +}
21659 +
21660 +
21661 +static int proc_virtnet_info(char *buffer)
21662 +{
21663 +       return proc_vci(buffer);
21664 +}
21665 +
21666 +static int proc_virtnet_status(char *buffer)
21667 +{
21668 +       return sprintf(buffer,
21669 +               "#CTotal:\t%d\n"
21670 +               "#CActive:\t%d\n",
21671 +               atomic_read(&nx_global_ctotal),
21672 +               atomic_read(&nx_global_cactive));
21673 +}
21674 +
21675 +int proc_nxi_info(struct nx_info *nxi, char *buffer)
21676 +{
21677 +       struct nx_addr_v4 *v4a;
21678 +#ifdef CONFIG_IPV6
21679 +       struct nx_addr_v6 *v6a;
21680 +#endif
21681 +       int length, i;
21682 +
21683 +       length = sprintf(buffer,
21684 +               "ID:\t%d\n"
21685 +               "Info:\t%p\n"
21686 +               "Bcast:\t" NIPQUAD_FMT "\n"
21687 +               "Lback:\t" NIPQUAD_FMT "\n",
21688 +               nxi->nx_id,
21689 +               nxi,
21690 +               NIPQUAD(nxi->v4_bcast.s_addr),
21691 +               NIPQUAD(nxi->v4_lback.s_addr));
21692 +
21693 +       if (!NX_IPV4(nxi))
21694 +               goto skip_v4;
21695 +       for (i = 0, v4a = &nxi->v4; v4a; i++, v4a = v4a->next)
21696 +               length += sprintf(buffer + length, "%d:\t" NXAV4_FMT "\n",
21697 +                       i, NXAV4(v4a));
21698 +skip_v4:
21699 +#ifdef CONFIG_IPV6
21700 +       if (!NX_IPV6(nxi))
21701 +               goto skip_v6;
21702 +       for (i = 0, v6a = &nxi->v6; v6a; i++, v6a = v6a->next)
21703 +               length += sprintf(buffer + length, "%d:\t" NXAV6_FMT "\n",
21704 +                       i, NXAV6(v6a));
21705 +skip_v6:
21706 +#endif
21707 +       return length;
21708 +}
21709 +
21710 +int proc_nxi_status(struct nx_info *nxi, char *buffer)
21711 +{
21712 +       int length;
21713 +
21714 +       length = sprintf(buffer,
21715 +               "UseCnt:\t%d\n"
21716 +               "Tasks:\t%d\n"
21717 +               "Flags:\t%016llx\n"
21718 +               "NCaps:\t%016llx\n",
21719 +               atomic_read(&nxi->nx_usecnt),
21720 +               atomic_read(&nxi->nx_tasks),
21721 +               (unsigned long long)nxi->nx_flags,
21722 +               (unsigned long long)nxi->nx_ncaps);
21723 +       return length;
21724 +}
21725 +
21726 +
21727 +
21728 +/* here the inode helpers */
21729 +
21730 +struct vs_entry {
21731 +       int len;
21732 +       char *name;
21733 +       mode_t mode;
21734 +       struct inode_operations *iop;
21735 +       struct file_operations *fop;
21736 +       union proc_op op;
21737 +};
21738 +
21739 +static struct inode *vs_proc_make_inode(struct super_block *sb, struct vs_entry *p)
21740 +{
21741 +       struct inode *inode = new_inode(sb);
21742 +
21743 +       if (!inode)
21744 +               goto out;
21745 +
21746 +       inode->i_mode = p->mode;
21747 +       if (p->iop)
21748 +               inode->i_op = p->iop;
21749 +       if (p->fop)
21750 +               inode->i_fop = p->fop;
21751 +
21752 +       inode->i_nlink = (p->mode & S_IFDIR) ? 2 : 1;
21753 +       inode->i_flags |= S_IMMUTABLE;
21754 +
21755 +       inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
21756 +
21757 +       inode->i_uid = 0;
21758 +       inode->i_gid = 0;
21759 +       inode->i_tag = 0;
21760 +out:
21761 +       return inode;
21762 +}
21763 +
21764 +static struct dentry *vs_proc_instantiate(struct inode *dir,
21765 +       struct dentry *dentry, int id, void *ptr)
21766 +{
21767 +       struct vs_entry *p = ptr;
21768 +       struct inode *inode = vs_proc_make_inode(dir->i_sb, p);
21769 +       struct dentry *error = ERR_PTR(-EINVAL);
21770 +
21771 +       if (!inode)
21772 +               goto out;
21773 +
21774 +       PROC_I(inode)->op = p->op;
21775 +       PROC_I(inode)->fd = id;
21776 +       d_add(dentry, inode);
21777 +       error = NULL;
21778 +out:
21779 +       return error;
21780 +}
21781 +
21782 +/* Lookups */
21783 +
21784 +typedef struct dentry *instantiate_t(struct inode *, struct dentry *, int, void *);
21785 +
21786 +/*
21787 + * Fill a directory entry.
21788 + *
21789 + * If possible create the dcache entry and derive our inode number and
21790 + * file type from dcache entry.
21791 + *
21792 + * Since all of the proc inode numbers are dynamically generated, the inode
21793 + * numbers do not exist until the inode is cache.  This means creating the
21794 + * the dcache entry in readdir is necessary to keep the inode numbers
21795 + * reported by readdir in sync with the inode numbers reported
21796 + * by stat.
21797 + */
21798 +static int proc_fill_cache(struct file *filp, void *dirent, filldir_t filldir,
21799 +       char *name, int len, instantiate_t instantiate, int id, void *ptr)
21800 +{
21801 +       struct dentry *child, *dir = filp->f_dentry;
21802 +       struct inode *inode;
21803 +       struct qstr qname;
21804 +       ino_t ino = 0;
21805 +       unsigned type = DT_UNKNOWN;
21806 +
21807 +       qname.name = name;
21808 +       qname.len  = len;
21809 +       qname.hash = full_name_hash(name, len);
21810 +
21811 +       child = d_lookup(dir, &qname);
21812 +       if (!child) {
21813 +               struct dentry *new;
21814 +               new = d_alloc(dir, &qname);
21815 +               if (new) {
21816 +                       child = instantiate(dir->d_inode, new, id, ptr);
21817 +                       if (child)
21818 +                               dput(new);
21819 +                       else
21820 +                               child = new;
21821 +               }
21822 +       }
21823 +       if (!child || IS_ERR(child) || !child->d_inode)
21824 +               goto end_instantiate;
21825 +       inode = child->d_inode;
21826 +       if (inode) {
21827 +               ino = inode->i_ino;
21828 +               type = inode->i_mode >> 12;
21829 +       }
21830 +       dput(child);
21831 +end_instantiate:
21832 +       if (!ino)
21833 +               ino = find_inode_number(dir, &qname);
21834 +       if (!ino)
21835 +               ino = 1;
21836 +       return filldir(dirent, name, len, filp->f_pos, ino, type);
21837 +}
21838 +
21839 +
21840 +
21841 +/* get and revalidate vx_info/xid */
21842 +
21843 +static inline
21844 +struct vx_info *get_proc_vx_info(struct inode *inode)
21845 +{
21846 +       return lookup_vx_info(PROC_I(inode)->fd);
21847 +}
21848 +
21849 +static int proc_xid_revalidate(struct dentry *dentry, struct nameidata *nd)
21850 +{
21851 +       struct inode *inode = dentry->d_inode;
21852 +       xid_t xid = PROC_I(inode)->fd;
21853 +
21854 +       if (!xid || xid_is_hashed(xid))
21855 +               return 1;
21856 +       d_drop(dentry);
21857 +       return 0;
21858 +}
21859 +
21860 +
21861 +/* get and revalidate nx_info/nid */
21862 +
21863 +static int proc_nid_revalidate(struct dentry *dentry, struct nameidata *nd)
21864 +{
21865 +       struct inode *inode = dentry->d_inode;
21866 +       nid_t nid = PROC_I(inode)->fd;
21867 +
21868 +       if (!nid || nid_is_hashed(nid))
21869 +               return 1;
21870 +       d_drop(dentry);
21871 +       return 0;
21872 +}
21873 +
21874 +
21875 +
21876 +#define PROC_BLOCK_SIZE (PAGE_SIZE - 1024)
21877 +
21878 +static ssize_t proc_vs_info_read(struct file *file, char __user *buf,
21879 +                         size_t count, loff_t *ppos)
21880 +{
21881 +       struct inode *inode = file->f_dentry->d_inode;
21882 +       unsigned long page;
21883 +       ssize_t length = 0;
21884 +
21885 +       if (count > PROC_BLOCK_SIZE)
21886 +               count = PROC_BLOCK_SIZE;
21887 +
21888 +       /* fade that out as soon as stable */
21889 +       WARN_ON(PROC_I(inode)->fd);
21890 +
21891 +       if (!(page = __get_free_page(GFP_KERNEL)))
21892 +               return -ENOMEM;
21893 +
21894 +       BUG_ON(!PROC_I(inode)->op.proc_vs_read);
21895 +       length = PROC_I(inode)->op.proc_vs_read((char *)page);
21896 +
21897 +       if (length >= 0)
21898 +               length = simple_read_from_buffer(buf, count, ppos,
21899 +                       (char *)page, length);
21900 +
21901 +       free_page(page);
21902 +       return length;
21903 +}
21904 +
21905 +static ssize_t proc_vx_info_read(struct file *file, char __user *buf,
21906 +                         size_t count, loff_t *ppos)
21907 +{
21908 +       struct inode *inode = file->f_dentry->d_inode;
21909 +       struct vx_info *vxi = NULL;
21910 +       xid_t xid = PROC_I(inode)->fd;
21911 +       unsigned long page;
21912 +       ssize_t length = 0;
21913 +
21914 +       if (count > PROC_BLOCK_SIZE)
21915 +               count = PROC_BLOCK_SIZE;
21916 +
21917 +       /* fade that out as soon as stable */
21918 +       WARN_ON(!xid);
21919 +       vxi = lookup_vx_info(xid);
21920 +       if (!vxi)
21921 +               goto out;
21922 +
21923 +       length = -ENOMEM;
21924 +       if (!(page = __get_free_page(GFP_KERNEL)))
21925 +               goto out_put;
21926 +
21927 +       BUG_ON(!PROC_I(inode)->op.proc_vxi_read);
21928 +       length = PROC_I(inode)->op.proc_vxi_read(vxi, (char *)page);
21929 +
21930 +       if (length >= 0)
21931 +               length = simple_read_from_buffer(buf, count, ppos,
21932 +                       (char *)page, length);
21933 +
21934 +       free_page(page);
21935 +out_put:
21936 +       put_vx_info(vxi);
21937 +out:
21938 +       return length;
21939 +}
21940 +
21941 +static ssize_t proc_nx_info_read(struct file *file, char __user *buf,
21942 +                         size_t count, loff_t *ppos)
21943 +{
21944 +       struct inode *inode = file->f_dentry->d_inode;
21945 +       struct nx_info *nxi = NULL;
21946 +       nid_t nid = PROC_I(inode)->fd;
21947 +       unsigned long page;
21948 +       ssize_t length = 0;
21949 +
21950 +       if (count > PROC_BLOCK_SIZE)
21951 +               count = PROC_BLOCK_SIZE;
21952 +
21953 +       /* fade that out as soon as stable */
21954 +       WARN_ON(!nid);
21955 +       nxi = lookup_nx_info(nid);
21956 +       if (!nxi)
21957 +               goto out;
21958 +
21959 +       length = -ENOMEM;
21960 +       if (!(page = __get_free_page(GFP_KERNEL)))
21961 +               goto out_put;
21962 +
21963 +       BUG_ON(!PROC_I(inode)->op.proc_nxi_read);
21964 +       length = PROC_I(inode)->op.proc_nxi_read(nxi, (char *)page);
21965 +
21966 +       if (length >= 0)
21967 +               length = simple_read_from_buffer(buf, count, ppos,
21968 +                       (char *)page, length);
21969 +
21970 +       free_page(page);
21971 +out_put:
21972 +       put_nx_info(nxi);
21973 +out:
21974 +       return length;
21975 +}
21976 +
21977 +
21978 +
21979 +/* here comes the lower level */
21980 +
21981 +
21982 +#define NOD(NAME, MODE, IOP, FOP, OP) {        \
21983 +       .len  = sizeof(NAME) - 1,       \
21984 +       .name = (NAME),                 \
21985 +       .mode = MODE,                   \
21986 +       .iop  = IOP,                    \
21987 +       .fop  = FOP,                    \
21988 +       .op   = OP,                     \
21989 +}
21990 +
21991 +
21992 +#define DIR(NAME, MODE, OTYPE)                         \
21993 +       NOD(NAME, (S_IFDIR | (MODE)),                   \
21994 +               &proc_ ## OTYPE ## _inode_operations,   \
21995 +               &proc_ ## OTYPE ## _file_operations, { } )
21996 +
21997 +#define INF(NAME, MODE, OTYPE)                         \
21998 +       NOD(NAME, (S_IFREG | (MODE)), NULL,             \
21999 +               &proc_vs_info_file_operations,          \
22000 +               { .proc_vs_read = &proc_##OTYPE } )
22001 +
22002 +#define VINF(NAME, MODE, OTYPE)                                \
22003 +       NOD(NAME, (S_IFREG | (MODE)), NULL,             \
22004 +               &proc_vx_info_file_operations,          \
22005 +               { .proc_vxi_read = &proc_##OTYPE } )
22006 +
22007 +#define NINF(NAME, MODE, OTYPE)                                \
22008 +       NOD(NAME, (S_IFREG | (MODE)), NULL,             \
22009 +               &proc_nx_info_file_operations,          \
22010 +               { .proc_nxi_read = &proc_##OTYPE } )
22011 +
22012 +
22013 +static struct file_operations proc_vs_info_file_operations = {
22014 +       .read =         proc_vs_info_read,
22015 +};
22016 +
22017 +static struct file_operations proc_vx_info_file_operations = {
22018 +       .read =         proc_vx_info_read,
22019 +};
22020 +
22021 +static struct dentry_operations proc_xid_dentry_operations = {
22022 +       .d_revalidate = proc_xid_revalidate,
22023 +};
22024 +
22025 +static struct vs_entry vx_base_stuff[] = {
22026 +       VINF("info",    S_IRUGO, vxi_info),
22027 +       VINF("status",  S_IRUGO, vxi_status),
22028 +       VINF("limit",   S_IRUGO, vxi_limit),
22029 +       VINF("sched",   S_IRUGO, vxi_sched),
22030 +       VINF("nsproxy", S_IRUGO, vxi_nsproxy0),
22031 +       VINF("nsproxy1",S_IRUGO, vxi_nsproxy1),
22032 +       VINF("cvirt",   S_IRUGO, vxi_cvirt),
22033 +       VINF("cacct",   S_IRUGO, vxi_cacct),
22034 +       {}
22035 +};
22036 +
22037 +
22038 +
22039 +
22040 +static struct dentry *proc_xid_instantiate(struct inode *dir,
22041 +       struct dentry *dentry, int id, void *ptr)
22042 +{
22043 +       dentry->d_op = &proc_xid_dentry_operations;
22044 +       return vs_proc_instantiate(dir, dentry, id, ptr);
22045 +}
22046 +
22047 +static struct dentry *proc_xid_lookup(struct inode *dir,
22048 +       struct dentry *dentry, struct nameidata *nd)
22049 +{
22050 +       struct vs_entry *p = vx_base_stuff;
22051 +       struct dentry *error = ERR_PTR(-ENOENT);
22052 +
22053 +       for (; p->name; p++) {
22054 +               if (p->len != dentry->d_name.len)
22055 +                       continue;
22056 +               if (!memcmp(dentry->d_name.name, p->name, p->len))
22057 +                       break;
22058 +       }
22059 +       if (!p->name)
22060 +               goto out;
22061 +
22062 +       error = proc_xid_instantiate(dir, dentry, PROC_I(dir)->fd, p);
22063 +out:
22064 +       return error;
22065 +}
22066 +
22067 +static int proc_xid_readdir(struct file *filp,
22068 +       void *dirent, filldir_t filldir)
22069 +{
22070 +       struct dentry *dentry = filp->f_dentry;
22071 +       struct inode *inode = dentry->d_inode;
22072 +       struct vs_entry *p = vx_base_stuff;
22073 +       int size = sizeof(vx_base_stuff) / sizeof(struct vs_entry);
22074 +       int pos, index;
22075 +       u64 ino;
22076 +
22077 +       pos = filp->f_pos;
22078 +       switch (pos) {
22079 +       case 0:
22080 +               ino = inode->i_ino;
22081 +               if (filldir(dirent, ".", 1, pos, ino, DT_DIR) < 0)
22082 +                       goto out;
22083 +               pos++;
22084 +               /* fall through */
22085 +       case 1:
22086 +               ino = parent_ino(dentry);
22087 +               if (filldir(dirent, "..", 2, pos, ino, DT_DIR) < 0)
22088 +                       goto out;
22089 +               pos++;
22090 +               /* fall through */
22091 +       default:
22092 +               index = pos - 2;
22093 +               if (index >= size)
22094 +                       goto out;
22095 +               for (p += index; p->name; p++) {
22096 +                       if (proc_fill_cache(filp, dirent, filldir, p->name, p->len,
22097 +                               vs_proc_instantiate, PROC_I(inode)->fd, p))
22098 +                               goto out;
22099 +                       pos++;
22100 +               }
22101 +       }
22102 +out:
22103 +       filp->f_pos = pos;
22104 +       return 1;
22105 +}
22106 +
22107 +
22108 +
22109 +static struct file_operations proc_nx_info_file_operations = {
22110 +       .read =         proc_nx_info_read,
22111 +};
22112 +
22113 +static struct dentry_operations proc_nid_dentry_operations = {
22114 +       .d_revalidate = proc_nid_revalidate,
22115 +};
22116 +
22117 +static struct vs_entry nx_base_stuff[] = {
22118 +       NINF("info",    S_IRUGO, nxi_info),
22119 +       NINF("status",  S_IRUGO, nxi_status),
22120 +       {}
22121 +};
22122 +
22123 +
22124 +static struct dentry *proc_nid_instantiate(struct inode *dir,
22125 +       struct dentry *dentry, int id, void *ptr)
22126 +{
22127 +       dentry->d_op = &proc_nid_dentry_operations;
22128 +       return vs_proc_instantiate(dir, dentry, id, ptr);
22129 +}
22130 +
22131 +static struct dentry *proc_nid_lookup(struct inode *dir,
22132 +       struct dentry *dentry, struct nameidata *nd)
22133 +{
22134 +       struct vs_entry *p = nx_base_stuff;
22135 +       struct dentry *error = ERR_PTR(-ENOENT);
22136 +
22137 +       for (; p->name; p++) {
22138 +               if (p->len != dentry->d_name.len)
22139 +                       continue;
22140 +               if (!memcmp(dentry->d_name.name, p->name, p->len))
22141 +                       break;
22142 +       }
22143 +       if (!p->name)
22144 +               goto out;
22145 +
22146 +       error = proc_nid_instantiate(dir, dentry, PROC_I(dir)->fd, p);
22147 +out:
22148 +       return error;
22149 +}
22150 +
22151 +static int proc_nid_readdir(struct file *filp,
22152 +       void *dirent, filldir_t filldir)
22153 +{
22154 +       struct dentry *dentry = filp->f_dentry;
22155 +       struct inode *inode = dentry->d_inode;
22156 +       struct vs_entry *p = nx_base_stuff;
22157 +       int size = sizeof(nx_base_stuff) / sizeof(struct vs_entry);
22158 +       int pos, index;
22159 +       u64 ino;
22160 +
22161 +       pos = filp->f_pos;
22162 +       switch (pos) {
22163 +       case 0:
22164 +               ino = inode->i_ino;
22165 +               if (filldir(dirent, ".", 1, pos, ino, DT_DIR) < 0)
22166 +                       goto out;
22167 +               pos++;
22168 +               /* fall through */
22169 +       case 1:
22170 +               ino = parent_ino(dentry);
22171 +               if (filldir(dirent, "..", 2, pos, ino, DT_DIR) < 0)
22172 +                       goto out;
22173 +               pos++;
22174 +               /* fall through */
22175 +       default:
22176 +               index = pos - 2;
22177 +               if (index >= size)
22178 +                       goto out;
22179 +               for (p += index; p->name; p++) {
22180 +                       if (proc_fill_cache(filp, dirent, filldir, p->name, p->len,
22181 +                               vs_proc_instantiate, PROC_I(inode)->fd, p))
22182 +                               goto out;
22183 +                       pos++;
22184 +               }
22185 +       }
22186 +out:
22187 +       filp->f_pos = pos;
22188 +       return 1;
22189 +}
22190 +
22191 +
22192 +#define MAX_MULBY10    ((~0U - 9) / 10)
22193 +
22194 +static inline int atovid(const char *str, int len)
22195 +{
22196 +       int vid, c;
22197 +
22198 +       vid = 0;
22199 +       while (len-- > 0) {
22200 +               c = *str - '0';
22201 +               str++;
22202 +               if (c > 9)
22203 +                       return -1;
22204 +               if (vid >= MAX_MULBY10)
22205 +                       return -1;
22206 +               vid *= 10;
22207 +               vid += c;
22208 +               if (!vid)
22209 +                       return -1;
22210 +       }
22211 +       return vid;
22212 +}
22213 +
22214 +/* now the upper level (virtual) */
22215 +
22216 +
22217 +static struct file_operations proc_xid_file_operations = {
22218 +       .read =         generic_read_dir,
22219 +       .readdir =      proc_xid_readdir,
22220 +};
22221 +
22222 +static struct inode_operations proc_xid_inode_operations = {
22223 +       .lookup =       proc_xid_lookup,
22224 +};
22225 +
22226 +static struct vs_entry vx_virtual_stuff[] = {
22227 +       INF("info",     S_IRUGO, virtual_info),
22228 +       INF("status",   S_IRUGO, virtual_status),
22229 +       DIR(NULL,       S_IRUGO | S_IXUGO, xid),
22230 +};
22231 +
22232 +
22233 +static struct dentry *proc_virtual_lookup(struct inode *dir,
22234 +       struct dentry *dentry, struct nameidata *nd)
22235 +{
22236 +       struct vs_entry *p = vx_virtual_stuff;
22237 +       struct dentry *error = ERR_PTR(-ENOENT);
22238 +       int id = 0;
22239 +
22240 +       for (; p->name; p++) {
22241 +               if (p->len != dentry->d_name.len)
22242 +                       continue;
22243 +               if (!memcmp(dentry->d_name.name, p->name, p->len))
22244 +                       break;
22245 +       }
22246 +       if (p->name)
22247 +               goto instantiate;
22248 +
22249 +       id = atovid(dentry->d_name.name, dentry->d_name.len);
22250 +       if ((id < 0) || !xid_is_hashed(id))
22251 +               goto out;
22252 +
22253 +instantiate:
22254 +       error = proc_xid_instantiate(dir, dentry, id, p);
22255 +out:
22256 +       return error;
22257 +}
22258 +
22259 +static struct file_operations proc_nid_file_operations = {
22260 +       .read =         generic_read_dir,
22261 +       .readdir =      proc_nid_readdir,
22262 +};
22263 +
22264 +static struct inode_operations proc_nid_inode_operations = {
22265 +       .lookup =       proc_nid_lookup,
22266 +};
22267 +
22268 +static struct vs_entry nx_virtnet_stuff[] = {
22269 +       INF("info",     S_IRUGO, virtnet_info),
22270 +       INF("status",   S_IRUGO, virtnet_status),
22271 +       DIR(NULL,       S_IRUGO | S_IXUGO, nid),
22272 +};
22273 +
22274 +
22275 +static struct dentry *proc_virtnet_lookup(struct inode *dir,
22276 +       struct dentry *dentry, struct nameidata *nd)
22277 +{
22278 +       struct vs_entry *p = nx_virtnet_stuff;
22279 +       struct dentry *error = ERR_PTR(-ENOENT);
22280 +       int id = 0;
22281 +
22282 +       for (; p->name; p++) {
22283 +               if (p->len != dentry->d_name.len)
22284 +                       continue;
22285 +               if (!memcmp(dentry->d_name.name, p->name, p->len))
22286 +                       break;
22287 +       }
22288 +       if (p->name)
22289 +               goto instantiate;
22290 +
22291 +       id = atovid(dentry->d_name.name, dentry->d_name.len);
22292 +       if ((id < 0) || !nid_is_hashed(id))
22293 +               goto out;
22294 +
22295 +instantiate:
22296 +       error = proc_nid_instantiate(dir, dentry, id, p);
22297 +out:
22298 +       return error;
22299 +}
22300 +
22301 +
22302 +#define PROC_MAXVIDS 32
22303 +
22304 +int proc_virtual_readdir(struct file *filp,
22305 +       void *dirent, filldir_t filldir)
22306 +{
22307 +       struct dentry *dentry = filp->f_dentry;
22308 +       struct inode *inode = dentry->d_inode;
22309 +       struct vs_entry *p = vx_virtual_stuff;
22310 +       int size = sizeof(vx_virtual_stuff) / sizeof(struct vs_entry);
22311 +       int pos, index;
22312 +       unsigned int xid_array[PROC_MAXVIDS];
22313 +       char buf[PROC_NUMBUF];
22314 +       unsigned int nr_xids, i;
22315 +       u64 ino;
22316 +
22317 +       pos = filp->f_pos;
22318 +       switch (pos) {
22319 +       case 0:
22320 +               ino = inode->i_ino;
22321 +               if (filldir(dirent, ".", 1, pos, ino, DT_DIR) < 0)
22322 +                       goto out;
22323 +               pos++;
22324 +               /* fall through */
22325 +       case 1:
22326 +               ino = parent_ino(dentry);
22327 +               if (filldir(dirent, "..", 2, pos, ino, DT_DIR) < 0)
22328 +                       goto out;
22329 +               pos++;
22330 +               /* fall through */
22331 +       default:
22332 +               index = pos - 2;
22333 +               if (index >= size)
22334 +                       goto entries;
22335 +               for (p += index; p->name; p++) {
22336 +                       if (proc_fill_cache(filp, dirent, filldir, p->name, p->len,
22337 +                               vs_proc_instantiate, 0, p))
22338 +                               goto out;
22339 +                       pos++;
22340 +               }
22341 +       entries:
22342 +               index = pos - size;
22343 +               p = &vx_virtual_stuff[size - 1];
22344 +               nr_xids = get_xid_list(index, xid_array, PROC_MAXVIDS);
22345 +               for (i = 0; i < nr_xids; i++) {
22346 +                       int n, xid = xid_array[i];
22347 +                       unsigned int j = PROC_NUMBUF;
22348 +
22349 +                       n = xid;
22350 +                       do
22351 +                               buf[--j] = '0' + (n % 10);
22352 +                       while (n /= 10);
22353 +
22354 +                       if (proc_fill_cache(filp, dirent, filldir,
22355 +                               buf + j, PROC_NUMBUF - j,
22356 +                               vs_proc_instantiate, xid, p))
22357 +                               goto out;
22358 +                       pos++;
22359 +               }
22360 +       }
22361 +out:
22362 +       filp->f_pos = pos;
22363 +       return 0;
22364 +}
22365 +
22366 +static int proc_virtual_getattr(struct vfsmount *mnt,
22367 +       struct dentry *dentry, struct kstat *stat)
22368 +{
22369 +       struct inode *inode = dentry->d_inode;
22370 +
22371 +       generic_fillattr(inode, stat);
22372 +       stat->nlink = 2 + atomic_read(&vx_global_cactive);
22373 +       return 0;
22374 +}
22375 +
22376 +static struct file_operations proc_virtual_dir_operations = {
22377 +       .read =         generic_read_dir,
22378 +       .readdir =      proc_virtual_readdir,
22379 +};
22380 +
22381 +static struct inode_operations proc_virtual_dir_inode_operations = {
22382 +       .getattr =      proc_virtual_getattr,
22383 +       .lookup =       proc_virtual_lookup,
22384 +};
22385 +
22386 +
22387 +
22388 +
22389 +
22390 +int proc_virtnet_readdir(struct file *filp,
22391 +       void *dirent, filldir_t filldir)
22392 +{
22393 +       struct dentry *dentry = filp->f_dentry;
22394 +       struct inode *inode = dentry->d_inode;
22395 +       struct vs_entry *p = nx_virtnet_stuff;
22396 +       int size = sizeof(nx_virtnet_stuff) / sizeof(struct vs_entry);
22397 +       int pos, index;
22398 +       unsigned int nid_array[PROC_MAXVIDS];
22399 +       char buf[PROC_NUMBUF];
22400 +       unsigned int nr_nids, i;
22401 +       u64 ino;
22402 +
22403 +       pos = filp->f_pos;
22404 +       switch (pos) {
22405 +       case 0:
22406 +               ino = inode->i_ino;
22407 +               if (filldir(dirent, ".", 1, pos, ino, DT_DIR) < 0)
22408 +                       goto out;
22409 +               pos++;
22410 +               /* fall through */
22411 +       case 1:
22412 +               ino = parent_ino(dentry);
22413 +               if (filldir(dirent, "..", 2, pos, ino, DT_DIR) < 0)
22414 +                       goto out;
22415 +               pos++;
22416 +               /* fall through */
22417 +       default:
22418 +               index = pos - 2;
22419 +               if (index >= size)
22420 +                       goto entries;
22421 +               for (p += index; p->name; p++) {
22422 +                       if (proc_fill_cache(filp, dirent, filldir, p->name, p->len,
22423 +                               vs_proc_instantiate, 0, p))
22424 +                               goto out;
22425 +                       pos++;
22426 +               }
22427 +       entries:
22428 +               index = pos - size;
22429 +               p = &nx_virtnet_stuff[size - 1];
22430 +               nr_nids = get_nid_list(index, nid_array, PROC_MAXVIDS);
22431 +               for (i = 0; i < nr_nids; i++) {
22432 +                       int n, nid = nid_array[i];
22433 +                       unsigned int j = PROC_NUMBUF;
22434 +
22435 +                       n = nid;
22436 +                       do
22437 +                               buf[--j] = '0' + (n % 10);
22438 +                       while (n /= 10);
22439 +
22440 +                       if (proc_fill_cache(filp, dirent, filldir,
22441 +                               buf + j, PROC_NUMBUF - j,
22442 +                               vs_proc_instantiate, nid, p))
22443 +                               goto out;
22444 +                       pos++;
22445 +               }
22446 +       }
22447 +out:
22448 +       filp->f_pos = pos;
22449 +       return 0;
22450 +}
22451 +
22452 +static int proc_virtnet_getattr(struct vfsmount *mnt,
22453 +       struct dentry *dentry, struct kstat *stat)
22454 +{
22455 +       struct inode *inode = dentry->d_inode;
22456 +
22457 +       generic_fillattr(inode, stat);
22458 +       stat->nlink = 2 + atomic_read(&nx_global_cactive);
22459 +       return 0;
22460 +}
22461 +
22462 +static struct file_operations proc_virtnet_dir_operations = {
22463 +       .read =         generic_read_dir,
22464 +       .readdir =      proc_virtnet_readdir,
22465 +};
22466 +
22467 +static struct inode_operations proc_virtnet_dir_inode_operations = {
22468 +       .getattr =      proc_virtnet_getattr,
22469 +       .lookup =       proc_virtnet_lookup,
22470 +};
22471 +
22472 +
22473 +
22474 +void proc_vx_init(void)
22475 +{
22476 +       struct proc_dir_entry *ent;
22477 +
22478 +       ent = proc_mkdir("virtual", 0);
22479 +       if (ent) {
22480 +               ent->proc_fops = &proc_virtual_dir_operations;
22481 +               ent->proc_iops = &proc_virtual_dir_inode_operations;
22482 +       }
22483 +       proc_virtual = ent;
22484 +
22485 +       ent = proc_mkdir("virtnet", 0);
22486 +       if (ent) {
22487 +               ent->proc_fops = &proc_virtnet_dir_operations;
22488 +               ent->proc_iops = &proc_virtnet_dir_inode_operations;
22489 +       }
22490 +       proc_virtnet = ent;
22491 +}
22492 +
22493 +
22494 +
22495 +
22496 +/* per pid info */
22497 +
22498 +
22499 +int proc_pid_vx_info(struct task_struct *p, char *buffer)
22500 +{
22501 +       struct vx_info *vxi;
22502 +       char *orig = buffer;
22503 +
22504 +       buffer += sprintf(buffer, "XID:\t%d\n", vx_task_xid(p));
22505 +
22506 +       vxi = task_get_vx_info(p);
22507 +       if (!vxi)
22508 +               goto out;
22509 +
22510 +       buffer += sprintf(buffer, "BCaps:\t");
22511 +       buffer = print_cap_t(buffer, &vxi->vx_bcaps);
22512 +       buffer += sprintf(buffer, "\n");
22513 +       buffer += sprintf(buffer, "CCaps:\t%016llx\n",
22514 +               (unsigned long long)vxi->vx_ccaps);
22515 +       buffer += sprintf(buffer, "CFlags:\t%016llx\n",
22516 +               (unsigned long long)vxi->vx_flags);
22517 +       buffer += sprintf(buffer, "CIPid:\t%d\n", vxi->vx_initpid);
22518 +
22519 +       put_vx_info(vxi);
22520 +out:
22521 +       return buffer - orig;
22522 +}
22523 +
22524 +
22525 +int proc_pid_nx_info(struct task_struct *p, char *buffer)
22526 +{
22527 +       struct nx_info *nxi;
22528 +       struct nx_addr_v4 *v4a;
22529 +#ifdef CONFIG_IPV6
22530 +       struct nx_addr_v6 *v6a;
22531 +#endif
22532 +       char *orig = buffer;
22533 +       int i;
22534 +
22535 +       buffer += sprintf(buffer, "NID:\t%d\n", nx_task_nid(p));
22536 +
22537 +       nxi = task_get_nx_info(p);
22538 +       if (!nxi)
22539 +               goto out;
22540 +
22541 +       buffer += sprintf(buffer, "NCaps:\t%016llx\n",
22542 +               (unsigned long long)nxi->nx_ncaps);
22543 +       buffer += sprintf(buffer, "NFlags:\t%016llx\n",
22544 +               (unsigned long long)nxi->nx_flags);
22545 +
22546 +       buffer += sprintf(buffer,
22547 +               "V4Root[bcast]:\t" NIPQUAD_FMT "\n",
22548 +               NIPQUAD(nxi->v4_bcast.s_addr));
22549 +       buffer += sprintf (buffer,
22550 +               "V4Root[lback]:\t" NIPQUAD_FMT "\n",
22551 +               NIPQUAD(nxi->v4_lback.s_addr));
22552 +       if (!NX_IPV4(nxi))
22553 +               goto skip_v4;
22554 +       for (i = 0, v4a = &nxi->v4; v4a; i++, v4a = v4a->next)
22555 +               buffer += sprintf(buffer, "V4Root[%d]:\t" NXAV4_FMT "\n",
22556 +                       i, NXAV4(v4a));
22557 +skip_v4:
22558 +#ifdef CONFIG_IPV6
22559 +       if (!NX_IPV6(nxi))
22560 +               goto skip_v6;
22561 +       for (i = 0, v6a = &nxi->v6; v6a; i++, v6a = v6a->next)
22562 +               buffer += sprintf(buffer, "V6Root[%d]:\t" NXAV6_FMT "\n",
22563 +                       i, NXAV6(v6a));
22564 +skip_v6:
22565 +#endif
22566 +       put_nx_info(nxi);
22567 +out:
22568 +       return buffer - orig;
22569 +}
22570 +
22571 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/sched.c linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/sched.c
22572 --- linux-3.0-rc5/kernel/vserver/sched.c        1970-01-01 01:00:00.000000000 +0100
22573 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/sched.c   2011-06-10 22:11:24.000000000 +0200
22574 @@ -0,0 +1,82 @@
22575 +/*
22576 + *  linux/kernel/vserver/sched.c
22577 + *
22578 + *  Virtual Server: Scheduler Support
22579 + *
22580 + *  Copyright (C) 2004-2010  Herbert Pötzl
22581 + *
22582 + *  V0.01  adapted Sam Vilains version to 2.6.3
22583 + *  V0.02  removed legacy interface
22584 + *  V0.03  changed vcmds to vxi arg
22585 + *  V0.04  removed older and legacy interfaces
22586 + *  V0.05  removed scheduler code/commands
22587 + *
22588 + */
22589 +
22590 +#include <linux/vs_context.h>
22591 +#include <linux/vs_sched.h>
22592 +#include <linux/vserver/sched_cmd.h>
22593 +
22594 +#include <asm/uaccess.h>
22595 +
22596 +
22597 +void vx_update_sched_param(struct _vx_sched *sched,
22598 +       struct _vx_sched_pc *sched_pc)
22599 +{
22600 +       sched_pc->prio_bias = sched->prio_bias;
22601 +}
22602 +
22603 +static int do_set_prio_bias(struct vx_info *vxi, struct vcmd_prio_bias *data)
22604 +{
22605 +       int cpu;
22606 +
22607 +       if (data->prio_bias > MAX_PRIO_BIAS)
22608 +               data->prio_bias = MAX_PRIO_BIAS;
22609 +       if (data->prio_bias < MIN_PRIO_BIAS)
22610 +               data->prio_bias = MIN_PRIO_BIAS;
22611 +
22612 +       if (data->cpu_id != ~0) {
22613 +               vxi->sched.update = cpumask_of_cpu(data->cpu_id);
22614 +               cpus_and(vxi->sched.update, cpu_online_map,
22615 +                       vxi->sched.update);
22616 +       } else
22617 +               vxi->sched.update = cpu_online_map;
22618 +
22619 +       for_each_cpu_mask(cpu, vxi->sched.update)
22620 +               vx_update_sched_param(&vxi->sched,
22621 +                       &vx_per_cpu(vxi, sched_pc, cpu));
22622 +       return 0;
22623 +}
22624 +
22625 +int vc_set_prio_bias(struct vx_info *vxi, void __user *data)
22626 +{
22627 +       struct vcmd_prio_bias vc_data;
22628 +
22629 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
22630 +               return -EFAULT;
22631 +
22632 +       return do_set_prio_bias(vxi, &vc_data);
22633 +}
22634 +
22635 +int vc_get_prio_bias(struct vx_info *vxi, void __user *data)
22636 +{
22637 +       struct vcmd_prio_bias vc_data;
22638 +       struct _vx_sched_pc *pcd;
22639 +       int cpu;
22640 +
22641 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
22642 +               return -EFAULT;
22643 +
22644 +       cpu = vc_data.cpu_id;
22645 +
22646 +       if (!cpu_possible(cpu))
22647 +               return -EINVAL;
22648 +
22649 +       pcd = &vx_per_cpu(vxi, sched_pc, cpu);
22650 +       vc_data.prio_bias = pcd->prio_bias;
22651 +
22652 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
22653 +               return -EFAULT;
22654 +       return 0;
22655 +}
22656 +
22657 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/sched_init.h linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/sched_init.h
22658 --- linux-3.0-rc5/kernel/vserver/sched_init.h   1970-01-01 01:00:00.000000000 +0100
22659 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/sched_init.h      2011-06-10 22:11:24.000000000 +0200
22660 @@ -0,0 +1,27 @@
22661 +
22662 +static inline void vx_info_init_sched(struct _vx_sched *sched)
22663 +{
22664 +       /* scheduling; hard code starting values as constants */
22665 +       sched->prio_bias = 0;
22666 +}
22667 +
22668 +static inline
22669 +void vx_info_init_sched_pc(struct _vx_sched_pc *sched_pc, int cpu)
22670 +{
22671 +       sched_pc->prio_bias = 0;
22672 +
22673 +       sched_pc->user_ticks = 0;
22674 +       sched_pc->sys_ticks = 0;
22675 +       sched_pc->hold_ticks = 0;
22676 +}
22677 +
22678 +static inline void vx_info_exit_sched(struct _vx_sched *sched)
22679 +{
22680 +       return;
22681 +}
22682 +
22683 +static inline
22684 +void vx_info_exit_sched_pc(struct _vx_sched_pc *sched_pc, int cpu)
22685 +{
22686 +       return;
22687 +}
22688 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/sched_proc.h linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/sched_proc.h
22689 --- linux-3.0-rc5/kernel/vserver/sched_proc.h   1970-01-01 01:00:00.000000000 +0100
22690 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/sched_proc.h      2011-06-10 22:11:24.000000000 +0200
22691 @@ -0,0 +1,32 @@
22692 +#ifndef _VX_SCHED_PROC_H
22693 +#define _VX_SCHED_PROC_H
22694 +
22695 +
22696 +static inline
22697 +int vx_info_proc_sched(struct _vx_sched *sched, char *buffer)
22698 +{
22699 +       int length = 0;
22700 +
22701 +       length += sprintf(buffer,
22702 +               "PrioBias:\t%8d\n",
22703 +               sched->prio_bias);
22704 +       return length;
22705 +}
22706 +
22707 +static inline
22708 +int vx_info_proc_sched_pc(struct _vx_sched_pc *sched_pc,
22709 +       char *buffer, int cpu)
22710 +{
22711 +       int length = 0;
22712 +
22713 +       length += sprintf(buffer + length,
22714 +               "cpu %d: %lld %lld %lld", cpu,
22715 +               (unsigned long long)sched_pc->user_ticks,
22716 +               (unsigned long long)sched_pc->sys_ticks,
22717 +               (unsigned long long)sched_pc->hold_ticks);
22718 +       length += sprintf(buffer + length,
22719 +               " %d\n", sched_pc->prio_bias);
22720 +       return length;
22721 +}
22722 +
22723 +#endif /* _VX_SCHED_PROC_H */
22724 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/signal.c linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/signal.c
22725 --- linux-3.0-rc5/kernel/vserver/signal.c       1970-01-01 01:00:00.000000000 +0100
22726 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/signal.c  2011-06-10 22:11:24.000000000 +0200
22727 @@ -0,0 +1,134 @@
22728 +/*
22729 + *  linux/kernel/vserver/signal.c
22730 + *
22731 + *  Virtual Server: Signal Support
22732 + *
22733 + *  Copyright (C) 2003-2007  Herbert Pötzl
22734 + *
22735 + *  V0.01  broken out from vcontext V0.05
22736 + *  V0.02  changed vcmds to vxi arg
22737 + *  V0.03  adjusted siginfo for kill
22738 + *
22739 + */
22740 +
22741 +#include <asm/uaccess.h>
22742 +
22743 +#include <linux/vs_context.h>
22744 +#include <linux/vs_pid.h>
22745 +#include <linux/vserver/signal_cmd.h>
22746 +
22747 +
22748 +int vx_info_kill(struct vx_info *vxi, int pid, int sig)
22749 +{
22750 +       int retval, count = 0;
22751 +       struct task_struct *p;
22752 +       struct siginfo *sip = SEND_SIG_PRIV;
22753 +
22754 +       retval = -ESRCH;
22755 +       vxdprintk(VXD_CBIT(misc, 4),
22756 +               "vx_info_kill(%p[#%d],%d,%d)*",
22757 +               vxi, vxi->vx_id, pid, sig);
22758 +       read_lock(&tasklist_lock);
22759 +       switch (pid) {
22760 +       case  0:
22761 +       case -1:
22762 +               for_each_process(p) {
22763 +                       int err = 0;
22764 +
22765 +                       if (vx_task_xid(p) != vxi->vx_id || p->pid <= 1 ||
22766 +                               (pid && vxi->vx_initpid == p->pid))
22767 +                               continue;
22768 +
22769 +                       err = group_send_sig_info(sig, sip, p);
22770 +                       ++count;
22771 +                       if (err != -EPERM)
22772 +                               retval = err;
22773 +               }
22774 +               break;
22775 +
22776 +       case 1:
22777 +               if (vxi->vx_initpid) {
22778 +                       pid = vxi->vx_initpid;
22779 +                       /* for now, only SIGINT to private init ... */
22780 +                       if (!vx_info_flags(vxi, VXF_STATE_ADMIN, 0) &&
22781 +                               /* ... as long as there are tasks left */
22782 +                               (atomic_read(&vxi->vx_tasks) > 1))
22783 +                               sig = SIGINT;
22784 +               }
22785 +               /* fallthrough */
22786 +       default:
22787 +               rcu_read_lock();
22788 +               p = find_task_by_real_pid(pid);
22789 +               rcu_read_unlock();
22790 +               if (p) {
22791 +                       if (vx_task_xid(p) == vxi->vx_id)
22792 +                               retval = group_send_sig_info(sig, sip, p);
22793 +               }
22794 +               break;
22795 +       }
22796 +       read_unlock(&tasklist_lock);
22797 +       vxdprintk(VXD_CBIT(misc, 4),
22798 +               "vx_info_kill(%p[#%d],%d,%d,%ld) = %d",
22799 +               vxi, vxi->vx_id, pid, sig, (long)sip, retval);
22800 +       return retval;
22801 +}
22802 +
22803 +int vc_ctx_kill(struct vx_info *vxi, void __user *data)
22804 +{
22805 +       struct vcmd_ctx_kill_v0 vc_data;
22806 +
22807 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
22808 +               return -EFAULT;
22809 +
22810 +       /* special check to allow guest shutdown */
22811 +       if (!vx_info_flags(vxi, VXF_STATE_ADMIN, 0) &&
22812 +               /* forbid killall pid=0 when init is present */
22813 +               (((vc_data.pid < 1) && vxi->vx_initpid) ||
22814 +               (vc_data.pid > 1)))
22815 +               return -EACCES;
22816 +
22817 +       return vx_info_kill(vxi, vc_data.pid, vc_data.sig);
22818 +}
22819 +
22820 +
22821 +static int __wait_exit(struct vx_info *vxi)
22822 +{
22823 +       DECLARE_WAITQUEUE(wait, current);
22824 +       int ret = 0;
22825 +
22826 +       add_wait_queue(&vxi->vx_wait, &wait);
22827 +       set_current_state(TASK_INTERRUPTIBLE);
22828 +
22829 +wait:
22830 +       if (vx_info_state(vxi,
22831 +               VXS_SHUTDOWN | VXS_HASHED | VXS_HELPER) == VXS_SHUTDOWN)
22832 +               goto out;
22833 +       if (signal_pending(current)) {
22834 +               ret = -ERESTARTSYS;
22835 +               goto out;
22836 +       }
22837 +       schedule();
22838 +       goto wait;
22839 +
22840 +out:
22841 +       set_current_state(TASK_RUNNING);
22842 +       remove_wait_queue(&vxi->vx_wait, &wait);
22843 +       return ret;
22844 +}
22845 +
22846 +
22847 +
22848 +int vc_wait_exit(struct vx_info *vxi, void __user *data)
22849 +{
22850 +       struct vcmd_wait_exit_v0 vc_data;
22851 +       int ret;
22852 +
22853 +       ret = __wait_exit(vxi);
22854 +       vc_data.reboot_cmd = vxi->reboot_cmd;
22855 +       vc_data.exit_code = vxi->exit_code;
22856 +
22857 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
22858 +               ret = -EFAULT;
22859 +       return ret;
22860 +}
22861 +
22862 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/space.c linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/space.c
22863 --- linux-3.0-rc5/kernel/vserver/space.c        1970-01-01 01:00:00.000000000 +0100
22864 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/space.c   2011-06-10 22:11:24.000000000 +0200
22865 @@ -0,0 +1,417 @@
22866 +/*
22867 + *  linux/kernel/vserver/space.c
22868 + *
22869 + *  Virtual Server: Context Space Support
22870 + *
22871 + *  Copyright (C) 2003-2010  Herbert Pötzl
22872 + *
22873 + *  V0.01  broken out from context.c 0.07
22874 + *  V0.02  added task locking for namespace
22875 + *  V0.03  broken out vx_enter_namespace
22876 + *  V0.04  added *space support and commands
22877 + *  V0.05  added credential support
22878 + *
22879 + */
22880 +
22881 +#include <linux/utsname.h>
22882 +#include <linux/nsproxy.h>
22883 +#include <linux/err.h>
22884 +#include <linux/fs_struct.h>
22885 +#include <linux/cred.h>
22886 +#include <asm/uaccess.h>
22887 +
22888 +#include <linux/vs_context.h>
22889 +#include <linux/vserver/space.h>
22890 +#include <linux/vserver/space_cmd.h>
22891 +
22892 +atomic_t vs_global_nsproxy     = ATOMIC_INIT(0);
22893 +atomic_t vs_global_fs          = ATOMIC_INIT(0);
22894 +atomic_t vs_global_mnt_ns      = ATOMIC_INIT(0);
22895 +atomic_t vs_global_uts_ns      = ATOMIC_INIT(0);
22896 +atomic_t vs_global_user_ns     = ATOMIC_INIT(0);
22897 +atomic_t vs_global_pid_ns      = ATOMIC_INIT(0);
22898 +
22899 +
22900 +/* namespace functions */
22901 +
22902 +#include <linux/mnt_namespace.h>
22903 +#include <linux/user_namespace.h>
22904 +#include <linux/pid_namespace.h>
22905 +#include <linux/ipc_namespace.h>
22906 +#include <net/net_namespace.h>
22907 +
22908 +
22909 +static const struct vcmd_space_mask_v1 space_mask_v0 = {
22910 +       .mask = CLONE_FS |
22911 +               CLONE_NEWNS |
22912 +               CLONE_NEWUTS |
22913 +               CLONE_NEWIPC |
22914 +               CLONE_NEWUSER |
22915 +               0
22916 +};
22917 +
22918 +static const struct vcmd_space_mask_v1 space_mask = {
22919 +       .mask = CLONE_FS |
22920 +               CLONE_NEWNS |
22921 +               CLONE_NEWUTS |
22922 +               CLONE_NEWIPC |
22923 +               CLONE_NEWUSER |
22924 +#ifdef CONFIG_PID_NS
22925 +               CLONE_NEWPID |
22926 +#endif
22927 +#ifdef CONFIG_NET_NS
22928 +               CLONE_NEWNET |
22929 +#endif
22930 +               0
22931 +};
22932 +
22933 +static const struct vcmd_space_mask_v1 default_space_mask = {
22934 +       .mask = CLONE_FS |
22935 +               CLONE_NEWNS |
22936 +               CLONE_NEWUTS |
22937 +               CLONE_NEWIPC |
22938 +               CLONE_NEWUSER |
22939 +#ifdef CONFIG_PID_NS
22940 +//             CLONE_NEWPID |
22941 +#endif
22942 +               0
22943 +};
22944 +
22945 +/*
22946 + *     build a new nsproxy mix
22947 + *      assumes that both proxies are 'const'
22948 + *     does not touch nsproxy refcounts
22949 + *     will hold a reference on the result.
22950 + */
22951 +
22952 +struct nsproxy *vs_mix_nsproxy(struct nsproxy *old_nsproxy,
22953 +       struct nsproxy *new_nsproxy, unsigned long mask)
22954 +{
22955 +       struct mnt_namespace *old_ns;
22956 +       struct uts_namespace *old_uts;
22957 +       struct ipc_namespace *old_ipc;
22958 +#ifdef CONFIG_PID_NS
22959 +       struct pid_namespace *old_pid;
22960 +#endif
22961 +#ifdef CONFIG_NET_NS
22962 +       struct net *old_net;
22963 +#endif
22964 +       struct nsproxy *nsproxy;
22965 +
22966 +       nsproxy = copy_nsproxy(old_nsproxy);
22967 +       if (!nsproxy)
22968 +               goto out;
22969 +
22970 +       if (mask & CLONE_NEWNS) {
22971 +               old_ns = nsproxy->mnt_ns;
22972 +               nsproxy->mnt_ns = new_nsproxy->mnt_ns;
22973 +               if (nsproxy->mnt_ns)
22974 +                       get_mnt_ns(nsproxy->mnt_ns);
22975 +       } else
22976 +               old_ns = NULL;
22977 +
22978 +       if (mask & CLONE_NEWUTS) {
22979 +               old_uts = nsproxy->uts_ns;
22980 +               nsproxy->uts_ns = new_nsproxy->uts_ns;
22981 +               if (nsproxy->uts_ns)
22982 +                       get_uts_ns(nsproxy->uts_ns);
22983 +       } else
22984 +               old_uts = NULL;
22985 +
22986 +       if (mask & CLONE_NEWIPC) {
22987 +               old_ipc = nsproxy->ipc_ns;
22988 +               nsproxy->ipc_ns = new_nsproxy->ipc_ns;
22989 +               if (nsproxy->ipc_ns)
22990 +                       get_ipc_ns(nsproxy->ipc_ns);
22991 +       } else
22992 +               old_ipc = NULL;
22993 +
22994 +#ifdef CONFIG_PID_NS
22995 +       if (mask & CLONE_NEWPID) {
22996 +               old_pid = nsproxy->pid_ns;
22997 +               nsproxy->pid_ns = new_nsproxy->pid_ns;
22998 +               if (nsproxy->pid_ns)
22999 +                       get_pid_ns(nsproxy->pid_ns);
23000 +       } else
23001 +               old_pid = NULL;
23002 +#endif
23003 +#ifdef CONFIG_NET_NS
23004 +       if (mask & CLONE_NEWNET) {
23005 +               old_net = nsproxy->net_ns;
23006 +               nsproxy->net_ns = new_nsproxy->net_ns;
23007 +               if (nsproxy->net_ns)
23008 +                       get_net(nsproxy->net_ns);
23009 +       } else
23010 +               old_net = NULL;
23011 +#endif
23012 +       if (old_ns)
23013 +               put_mnt_ns(old_ns);
23014 +       if (old_uts)
23015 +               put_uts_ns(old_uts);
23016 +       if (old_ipc)
23017 +               put_ipc_ns(old_ipc);
23018 +#ifdef CONFIG_PID_NS
23019 +       if (old_pid)
23020 +               put_pid_ns(old_pid);
23021 +#endif
23022 +#ifdef CONFIG_NET_NS
23023 +       if (old_net)
23024 +               put_net(old_net);
23025 +#endif
23026 +out:
23027 +       return nsproxy;
23028 +}
23029 +
23030 +
23031 +/*
23032 + *     merge two nsproxy structs into a new one.
23033 + *     will hold a reference on the result.
23034 + */
23035 +
23036 +static inline
23037 +struct nsproxy *__vs_merge_nsproxy(struct nsproxy *old,
23038 +       struct nsproxy *proxy, unsigned long mask)
23039 +{
23040 +       struct nsproxy null_proxy = { .mnt_ns = NULL };
23041 +
23042 +       if (!proxy)
23043 +               return NULL;
23044 +
23045 +       if (mask) {
23046 +               /* vs_mix_nsproxy returns with reference */
23047 +               return vs_mix_nsproxy(old ? old : &null_proxy,
23048 +                       proxy, mask);
23049 +       }
23050 +       get_nsproxy(proxy);
23051 +       return proxy;
23052 +}
23053 +
23054 +
23055 +int vx_enter_space(struct vx_info *vxi, unsigned long mask, unsigned index)
23056 +{
23057 +       struct nsproxy *proxy, *proxy_cur, *proxy_new;
23058 +       struct fs_struct *fs_cur, *fs = NULL;
23059 +       struct _vx_space *space;
23060 +       int ret, kill = 0;
23061 +
23062 +       vxdprintk(VXD_CBIT(space, 8), "vx_enter_space(%p[#%u],0x%08lx,%d)",
23063 +               vxi, vxi->vx_id, mask, index);
23064 +
23065 +       if (vx_info_flags(vxi, VXF_INFO_PRIVATE, 0))
23066 +               return -EACCES;
23067 +
23068 +       if (index >= VX_SPACES)
23069 +               return -EINVAL;
23070 +
23071 +       space = &vxi->space[index];
23072 +
23073 +       if (!mask)
23074 +               mask = space->vx_nsmask;
23075 +
23076 +       if ((mask & space->vx_nsmask) != mask)
23077 +               return -EINVAL;
23078 +
23079 +       if (mask & CLONE_FS) {
23080 +               fs = copy_fs_struct(space->vx_fs);
23081 +               if (!fs)
23082 +                       return -ENOMEM;
23083 +       }
23084 +       proxy = space->vx_nsproxy;
23085 +
23086 +       vxdprintk(VXD_CBIT(space, 9),
23087 +               "vx_enter_space(%p[#%u],0x%08lx,%d) -> (%p,%p)",
23088 +               vxi, vxi->vx_id, mask, index, proxy, fs);
23089 +
23090 +       task_lock(current);
23091 +       fs_cur = current->fs;
23092 +
23093 +       if (mask & CLONE_FS) {
23094 +               spin_lock(&fs_cur->lock);
23095 +               current->fs = fs;
23096 +               kill = !--fs_cur->users;
23097 +               spin_unlock(&fs_cur->lock);
23098 +       }
23099 +
23100 +       proxy_cur = current->nsproxy;
23101 +       get_nsproxy(proxy_cur);
23102 +       task_unlock(current);
23103 +
23104 +       if (kill)
23105 +               free_fs_struct(fs_cur);
23106 +
23107 +       proxy_new = __vs_merge_nsproxy(proxy_cur, proxy, mask);
23108 +       if (IS_ERR(proxy_new)) {
23109 +               ret = PTR_ERR(proxy_new);
23110 +               goto out_put;
23111 +       }
23112 +
23113 +       proxy_new = xchg(&current->nsproxy, proxy_new);
23114 +
23115 +       if (mask & CLONE_NEWUSER) {
23116 +               struct cred *cred;
23117 +
23118 +               vxdprintk(VXD_CBIT(space, 10),
23119 +                       "vx_enter_space(%p[#%u],%p) cred (%p,%p)",
23120 +                       vxi, vxi->vx_id, space->vx_cred,
23121 +                       current->real_cred, current->cred);
23122 +
23123 +               if (space->vx_cred) {
23124 +                       cred = __prepare_creds(space->vx_cred);
23125 +                       if (cred)
23126 +                               commit_creds(cred);
23127 +               }
23128 +       }
23129 +
23130 +       ret = 0;
23131 +
23132 +       if (proxy_new)
23133 +               put_nsproxy(proxy_new);
23134 +out_put:
23135 +       if (proxy_cur)
23136 +               put_nsproxy(proxy_cur);
23137 +       return ret;
23138 +}
23139 +
23140 +
23141 +int vx_set_space(struct vx_info *vxi, unsigned long mask, unsigned index)
23142 +{
23143 +       struct nsproxy *proxy_vxi, *proxy_cur, *proxy_new;
23144 +       struct fs_struct *fs_vxi, *fs;
23145 +       struct _vx_space *space;
23146 +       int ret, kill = 0;
23147 +
23148 +       vxdprintk(VXD_CBIT(space, 8), "vx_set_space(%p[#%u],0x%08lx,%d)",
23149 +               vxi, vxi->vx_id, mask, index);
23150 +
23151 +       if ((mask & space_mask.mask) != mask)
23152 +               return -EINVAL;
23153 +
23154 +       if (index >= VX_SPACES)
23155 +               return -EINVAL;
23156 +
23157 +       space = &vxi->space[index];
23158 +
23159 +       proxy_vxi = space->vx_nsproxy;
23160 +       fs_vxi = space->vx_fs;
23161 +
23162 +       if (mask & CLONE_FS) {
23163 +               fs = copy_fs_struct(current->fs);
23164 +               if (!fs)
23165 +                       return -ENOMEM;
23166 +       }
23167 +
23168 +       task_lock(current);
23169 +
23170 +       if (mask & CLONE_FS) {
23171 +               spin_lock(&fs_vxi->lock);
23172 +               space->vx_fs = fs;
23173 +               kill = !--fs_vxi->users;
23174 +               spin_unlock(&fs_vxi->lock);
23175 +       }
23176 +
23177 +       proxy_cur = current->nsproxy;
23178 +       get_nsproxy(proxy_cur);
23179 +       task_unlock(current);
23180 +
23181 +       if (kill)
23182 +               free_fs_struct(fs_vxi);
23183 +
23184 +       proxy_new = __vs_merge_nsproxy(proxy_vxi, proxy_cur, mask);
23185 +       if (IS_ERR(proxy_new)) {
23186 +               ret = PTR_ERR(proxy_new);
23187 +               goto out_put;
23188 +       }
23189 +
23190 +       proxy_new = xchg(&space->vx_nsproxy, proxy_new);
23191 +       space->vx_nsmask |= mask;
23192 +
23193 +       if (mask & CLONE_NEWUSER) {
23194 +               struct cred *cred;
23195 +
23196 +               vxdprintk(VXD_CBIT(space, 10),
23197 +                       "vx_set_space(%p[#%u],%p) cred (%p,%p)",
23198 +                       vxi, vxi->vx_id, space->vx_cred,
23199 +                       current->real_cred, current->cred);
23200 +
23201 +               cred = prepare_creds();
23202 +               cred = (struct cred *)xchg(&space->vx_cred, cred);
23203 +               if (cred)
23204 +                       abort_creds(cred);
23205 +       }
23206 +
23207 +       ret = 0;
23208 +
23209 +       if (proxy_new)
23210 +               put_nsproxy(proxy_new);
23211 +out_put:
23212 +       if (proxy_cur)
23213 +               put_nsproxy(proxy_cur);
23214 +       return ret;
23215 +}
23216 +
23217 +
23218 +int vc_enter_space_v1(struct vx_info *vxi, void __user *data)
23219 +{
23220 +       struct vcmd_space_mask_v1 vc_data = { .mask = 0 };
23221 +
23222 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
23223 +               return -EFAULT;
23224 +
23225 +       return vx_enter_space(vxi, vc_data.mask, 0);
23226 +}
23227 +
23228 +int vc_enter_space(struct vx_info *vxi, void __user *data)
23229 +{
23230 +       struct vcmd_space_mask_v2 vc_data = { .mask = 0 };
23231 +
23232 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
23233 +               return -EFAULT;
23234 +
23235 +       if (vc_data.index >= VX_SPACES)
23236 +               return -EINVAL;
23237 +
23238 +       return vx_enter_space(vxi, vc_data.mask, vc_data.index);
23239 +}
23240 +
23241 +int vc_set_space_v1(struct vx_info *vxi, void __user *data)
23242 +{
23243 +       struct vcmd_space_mask_v1 vc_data = { .mask = 0 };
23244 +
23245 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
23246 +               return -EFAULT;
23247 +
23248 +       return vx_set_space(vxi, vc_data.mask, 0);
23249 +}
23250 +
23251 +int vc_set_space(struct vx_info *vxi, void __user *data)
23252 +{
23253 +       struct vcmd_space_mask_v2 vc_data = { .mask = 0 };
23254 +
23255 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
23256 +               return -EFAULT;
23257 +
23258 +       if (vc_data.index >= VX_SPACES)
23259 +               return -EINVAL;
23260 +
23261 +       return vx_set_space(vxi, vc_data.mask, vc_data.index);
23262 +}
23263 +
23264 +int vc_get_space_mask(void __user *data, int type)
23265 +{
23266 +       const struct vcmd_space_mask_v1 *mask;
23267 +
23268 +       if (type == 0)
23269 +               mask = &space_mask_v0;
23270 +       else if (type == 1)
23271 +               mask = &space_mask;
23272 +       else
23273 +               mask = &default_space_mask;
23274 +
23275 +       vxdprintk(VXD_CBIT(space, 10),
23276 +               "vc_get_space_mask(%d) = %08llx", type, mask->mask);
23277 +
23278 +       if (copy_to_user(data, mask, sizeof(*mask)))
23279 +               return -EFAULT;
23280 +       return 0;
23281 +}
23282 +
23283 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/switch.c linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/switch.c
23284 --- linux-3.0-rc5/kernel/vserver/switch.c       1970-01-01 01:00:00.000000000 +0100
23285 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/switch.c  2011-06-10 22:11:24.000000000 +0200
23286 @@ -0,0 +1,548 @@
23287 +/*
23288 + *  linux/kernel/vserver/switch.c
23289 + *
23290 + *  Virtual Server: Syscall Switch
23291 + *
23292 + *  Copyright (C) 2003-2007  Herbert Pötzl
23293 + *
23294 + *  V0.01  syscall switch
23295 + *  V0.02  added signal to context
23296 + *  V0.03  added rlimit functions
23297 + *  V0.04  added iattr, task/xid functions
23298 + *  V0.05  added debug/history stuff
23299 + *  V0.06  added compat32 layer
23300 + *  V0.07  vcmd args and perms
23301 + *  V0.08  added status commands
23302 + *  V0.09  added tag commands
23303 + *  V0.10  added oom bias
23304 + *  V0.11  added device commands
23305 + *
23306 + */
23307 +
23308 +#include <linux/vs_context.h>
23309 +#include <linux/vs_network.h>
23310 +#include <linux/vserver/switch.h>
23311 +
23312 +#include "vci_config.h"
23313 +
23314 +
23315 +static inline
23316 +int vc_get_version(uint32_t id)
23317 +{
23318 +       return VCI_VERSION;
23319 +}
23320 +
23321 +static inline
23322 +int vc_get_vci(uint32_t id)
23323 +{
23324 +       return vci_kernel_config();
23325 +}
23326 +
23327 +#include <linux/vserver/context_cmd.h>
23328 +#include <linux/vserver/cvirt_cmd.h>
23329 +#include <linux/vserver/cacct_cmd.h>
23330 +#include <linux/vserver/limit_cmd.h>
23331 +#include <linux/vserver/network_cmd.h>
23332 +#include <linux/vserver/sched_cmd.h>
23333 +#include <linux/vserver/debug_cmd.h>
23334 +#include <linux/vserver/inode_cmd.h>
23335 +#include <linux/vserver/dlimit_cmd.h>
23336 +#include <linux/vserver/signal_cmd.h>
23337 +#include <linux/vserver/space_cmd.h>
23338 +#include <linux/vserver/tag_cmd.h>
23339 +#include <linux/vserver/device_cmd.h>
23340 +
23341 +#include <linux/vserver/inode.h>
23342 +#include <linux/vserver/dlimit.h>
23343 +
23344 +
23345 +#ifdef CONFIG_COMPAT
23346 +#define __COMPAT(name, id, data, compat)       \
23347 +       (compat) ? name ## _x32(id, data) : name(id, data)
23348 +#define __COMPAT_NO_ID(name, data, compat)     \
23349 +       (compat) ? name ## _x32(data) : name(data)
23350 +#else
23351 +#define __COMPAT(name, id, data, compat)       \
23352 +       name(id, data)
23353 +#define __COMPAT_NO_ID(name, data, compat)     \
23354 +       name(data)
23355 +#endif
23356 +
23357 +
23358 +static inline
23359 +long do_vcmd(uint32_t cmd, uint32_t id,
23360 +       struct vx_info *vxi, struct nx_info *nxi,
23361 +       void __user *data, int compat)
23362 +{
23363 +       switch (cmd) {
23364 +
23365 +       case VCMD_get_version:
23366 +               return vc_get_version(id);
23367 +       case VCMD_get_vci:
23368 +               return vc_get_vci(id);
23369 +
23370 +       case VCMD_task_xid:
23371 +               return vc_task_xid(id);
23372 +       case VCMD_vx_info:
23373 +               return vc_vx_info(vxi, data);
23374 +
23375 +       case VCMD_task_nid:
23376 +               return vc_task_nid(id);
23377 +       case VCMD_nx_info:
23378 +               return vc_nx_info(nxi, data);
23379 +
23380 +       case VCMD_task_tag:
23381 +               return vc_task_tag(id);
23382 +
23383 +       case VCMD_set_space_v1:
23384 +               return vc_set_space_v1(vxi, data);
23385 +       /* this is version 2 */
23386 +       case VCMD_set_space:
23387 +               return vc_set_space(vxi, data);
23388 +
23389 +       case VCMD_get_space_mask_v0:
23390 +               return vc_get_space_mask(data, 0);
23391 +       /* this is version 1 */
23392 +       case VCMD_get_space_mask:
23393 +               return vc_get_space_mask(data, 1);
23394 +
23395 +       case VCMD_get_space_default:
23396 +               return vc_get_space_mask(data, -1);
23397 +
23398 +       case VCMD_set_umask:
23399 +               return vc_set_umask(vxi, data);
23400 +
23401 +       case VCMD_get_umask:
23402 +               return vc_get_umask(vxi, data);
23403 +
23404 +#ifdef CONFIG_IA32_EMULATION
23405 +       case VCMD_get_rlimit:
23406 +               return __COMPAT(vc_get_rlimit, vxi, data, compat);
23407 +       case VCMD_set_rlimit:
23408 +               return __COMPAT(vc_set_rlimit, vxi, data, compat);
23409 +#else
23410 +       case VCMD_get_rlimit:
23411 +               return vc_get_rlimit(vxi, data);
23412 +       case VCMD_set_rlimit:
23413 +               return vc_set_rlimit(vxi, data);
23414 +#endif
23415 +       case VCMD_get_rlimit_mask:
23416 +               return vc_get_rlimit_mask(id, data);
23417 +       case VCMD_reset_hits:
23418 +               return vc_reset_hits(vxi, data);
23419 +       case VCMD_reset_minmax:
23420 +               return vc_reset_minmax(vxi, data);
23421 +
23422 +       case VCMD_get_vhi_name:
23423 +               return vc_get_vhi_name(vxi, data);
23424 +       case VCMD_set_vhi_name:
23425 +               return vc_set_vhi_name(vxi, data);
23426 +
23427 +       case VCMD_ctx_stat:
23428 +               return vc_ctx_stat(vxi, data);
23429 +       case VCMD_virt_stat:
23430 +               return vc_virt_stat(vxi, data);
23431 +       case VCMD_sock_stat:
23432 +               return vc_sock_stat(vxi, data);
23433 +       case VCMD_rlimit_stat:
23434 +               return vc_rlimit_stat(vxi, data);
23435 +
23436 +       case VCMD_set_cflags:
23437 +               return vc_set_cflags(vxi, data);
23438 +       case VCMD_get_cflags:
23439 +               return vc_get_cflags(vxi, data);
23440 +
23441 +       /* this is version 1 */
23442 +       case VCMD_set_ccaps:
23443 +               return vc_set_ccaps(vxi, data);
23444 +       /* this is version 1 */
23445 +       case VCMD_get_ccaps:
23446 +               return vc_get_ccaps(vxi, data);
23447 +       case VCMD_set_bcaps:
23448 +               return vc_set_bcaps(vxi, data);
23449 +       case VCMD_get_bcaps:
23450 +               return vc_get_bcaps(vxi, data);
23451 +
23452 +       case VCMD_set_badness:
23453 +               return vc_set_badness(vxi, data);
23454 +       case VCMD_get_badness:
23455 +               return vc_get_badness(vxi, data);
23456 +
23457 +       case VCMD_set_nflags:
23458 +               return vc_set_nflags(nxi, data);
23459 +       case VCMD_get_nflags:
23460 +               return vc_get_nflags(nxi, data);
23461 +
23462 +       case VCMD_set_ncaps:
23463 +               return vc_set_ncaps(nxi, data);
23464 +       case VCMD_get_ncaps:
23465 +               return vc_get_ncaps(nxi, data);
23466 +
23467 +       case VCMD_set_prio_bias:
23468 +               return vc_set_prio_bias(vxi, data);
23469 +       case VCMD_get_prio_bias:
23470 +               return vc_get_prio_bias(vxi, data);
23471 +       case VCMD_add_dlimit:
23472 +               return __COMPAT(vc_add_dlimit, id, data, compat);
23473 +       case VCMD_rem_dlimit:
23474 +               return __COMPAT(vc_rem_dlimit, id, data, compat);
23475 +       case VCMD_set_dlimit:
23476 +               return __COMPAT(vc_set_dlimit, id, data, compat);
23477 +       case VCMD_get_dlimit:
23478 +               return __COMPAT(vc_get_dlimit, id, data, compat);
23479 +
23480 +       case VCMD_ctx_kill:
23481 +               return vc_ctx_kill(vxi, data);
23482 +
23483 +       case VCMD_wait_exit:
23484 +               return vc_wait_exit(vxi, data);
23485 +
23486 +       case VCMD_get_iattr:
23487 +               return __COMPAT_NO_ID(vc_get_iattr, data, compat);
23488 +       case VCMD_set_iattr:
23489 +               return __COMPAT_NO_ID(vc_set_iattr, data, compat);
23490 +
23491 +       case VCMD_fget_iattr:
23492 +               return vc_fget_iattr(id, data);
23493 +       case VCMD_fset_iattr:
23494 +               return vc_fset_iattr(id, data);
23495 +
23496 +       case VCMD_enter_space_v0:
23497 +               return vc_enter_space_v1(vxi, NULL);
23498 +       case VCMD_enter_space_v1:
23499 +               return vc_enter_space_v1(vxi, data);
23500 +       /* this is version 2 */
23501 +       case VCMD_enter_space:
23502 +               return vc_enter_space(vxi, data);
23503 +
23504 +       case VCMD_ctx_create_v0:
23505 +               return vc_ctx_create(id, NULL);
23506 +       case VCMD_ctx_create:
23507 +               return vc_ctx_create(id, data);
23508 +       case VCMD_ctx_migrate_v0:
23509 +               return vc_ctx_migrate(vxi, NULL);
23510 +       case VCMD_ctx_migrate:
23511 +               return vc_ctx_migrate(vxi, data);
23512 +
23513 +       case VCMD_net_create_v0:
23514 +               return vc_net_create(id, NULL);
23515 +       case VCMD_net_create:
23516 +               return vc_net_create(id, data);
23517 +       case VCMD_net_migrate:
23518 +               return vc_net_migrate(nxi, data);
23519 +
23520 +       case VCMD_tag_migrate:
23521 +               return vc_tag_migrate(id);
23522 +
23523 +       case VCMD_net_add:
23524 +               return vc_net_add(nxi, data);
23525 +       case VCMD_net_remove:
23526 +               return vc_net_remove(nxi, data);
23527 +
23528 +       case VCMD_net_add_ipv4_v1:
23529 +               return vc_net_add_ipv4_v1(nxi, data);
23530 +       /* this is version 2 */
23531 +       case VCMD_net_add_ipv4:
23532 +               return vc_net_add_ipv4(nxi, data);
23533 +
23534 +       case VCMD_net_rem_ipv4_v1:
23535 +               return vc_net_rem_ipv4_v1(nxi, data);
23536 +       /* this is version 2 */
23537 +       case VCMD_net_rem_ipv4:
23538 +               return vc_net_rem_ipv4(nxi, data);
23539 +#ifdef CONFIG_IPV6
23540 +       case VCMD_net_add_ipv6:
23541 +               return vc_net_add_ipv6(nxi, data);
23542 +       case VCMD_net_remove_ipv6:
23543 +               return vc_net_remove_ipv6(nxi, data);
23544 +#endif
23545 +/*     case VCMD_add_match_ipv4:
23546 +               return vc_add_match_ipv4(nxi, data);
23547 +       case VCMD_get_match_ipv4:
23548 +               return vc_get_match_ipv4(nxi, data);
23549 +#ifdef CONFIG_IPV6
23550 +       case VCMD_add_match_ipv6:
23551 +               return vc_add_match_ipv6(nxi, data);
23552 +       case VCMD_get_match_ipv6:
23553 +               return vc_get_match_ipv6(nxi, data);
23554 +#endif */
23555 +
23556 +#ifdef CONFIG_VSERVER_DEVICE
23557 +       case VCMD_set_mapping:
23558 +               return __COMPAT(vc_set_mapping, vxi, data, compat);
23559 +       case VCMD_unset_mapping:
23560 +               return __COMPAT(vc_unset_mapping, vxi, data, compat);
23561 +#endif
23562 +#ifdef CONFIG_VSERVER_HISTORY
23563 +       case VCMD_dump_history:
23564 +               return vc_dump_history(id);
23565 +       case VCMD_read_history:
23566 +               return __COMPAT(vc_read_history, id, data, compat);
23567 +#endif
23568 +       default:
23569 +               vxwprintk_task(1, "unimplemented VCMD_%02d_%d[%d]",
23570 +                       VC_CATEGORY(cmd), VC_COMMAND(cmd), VC_VERSION(cmd));
23571 +       }
23572 +       return -ENOSYS;
23573 +}
23574 +
23575 +
23576 +#define        __VCMD(vcmd, _perm, _args, _flags)              \
23577 +       case VCMD_ ## vcmd: perm = _perm;               \
23578 +               args = _args; flags = _flags; break
23579 +
23580 +
23581 +#define VCA_NONE       0x00
23582 +#define VCA_VXI                0x01
23583 +#define VCA_NXI                0x02
23584 +
23585 +#define VCF_NONE       0x00
23586 +#define VCF_INFO       0x01
23587 +#define VCF_ADMIN      0x02
23588 +#define VCF_ARES       0x06    /* includes admin */
23589 +#define VCF_SETUP      0x08
23590 +
23591 +#define VCF_ZIDOK      0x10    /* zero id okay */
23592 +
23593 +
23594 +static inline
23595 +long do_vserver(uint32_t cmd, uint32_t id, void __user *data, int compat)
23596 +{
23597 +       long ret;
23598 +       int permit = -1, state = 0;
23599 +       int perm = -1, args = 0, flags = 0;
23600 +       struct vx_info *vxi = NULL;
23601 +       struct nx_info *nxi = NULL;
23602 +
23603 +       switch (cmd) {
23604 +       /* unpriviledged commands */
23605 +       __VCMD(get_version,      0, VCA_NONE,   0);
23606 +       __VCMD(get_vci,          0, VCA_NONE,   0);
23607 +       __VCMD(get_rlimit_mask,  0, VCA_NONE,   0);
23608 +       __VCMD(get_space_mask_v0,0, VCA_NONE,   0);
23609 +       __VCMD(get_space_mask,   0, VCA_NONE,   0);
23610 +       __VCMD(get_space_default,0, VCA_NONE,   0);
23611 +
23612 +       /* info commands */
23613 +       __VCMD(task_xid,         2, VCA_NONE,   0);
23614 +       __VCMD(reset_hits,       2, VCA_VXI,    0);
23615 +       __VCMD(reset_minmax,     2, VCA_VXI,    0);
23616 +       __VCMD(vx_info,          3, VCA_VXI,    VCF_INFO);
23617 +       __VCMD(get_bcaps,        3, VCA_VXI,    VCF_INFO);
23618 +       __VCMD(get_ccaps,        3, VCA_VXI,    VCF_INFO);
23619 +       __VCMD(get_cflags,       3, VCA_VXI,    VCF_INFO);
23620 +       __VCMD(get_umask,        3, VCA_VXI,    VCF_INFO);
23621 +       __VCMD(get_badness,      3, VCA_VXI,    VCF_INFO);
23622 +       __VCMD(get_vhi_name,     3, VCA_VXI,    VCF_INFO);
23623 +       __VCMD(get_rlimit,       3, VCA_VXI,    VCF_INFO);
23624 +
23625 +       __VCMD(ctx_stat,         3, VCA_VXI,    VCF_INFO);
23626 +       __VCMD(virt_stat,        3, VCA_VXI,    VCF_INFO);
23627 +       __VCMD(sock_stat,        3, VCA_VXI,    VCF_INFO);
23628 +       __VCMD(rlimit_stat,      3, VCA_VXI,    VCF_INFO);
23629 +
23630 +       __VCMD(task_nid,         2, VCA_NONE,   0);
23631 +       __VCMD(nx_info,          3, VCA_NXI,    VCF_INFO);
23632 +       __VCMD(get_ncaps,        3, VCA_NXI,    VCF_INFO);
23633 +       __VCMD(get_nflags,       3, VCA_NXI,    VCF_INFO);
23634 +
23635 +       __VCMD(task_tag,         2, VCA_NONE,   0);
23636 +
23637 +       __VCMD(get_iattr,        2, VCA_NONE,   0);
23638 +       __VCMD(fget_iattr,       2, VCA_NONE,   0);
23639 +       __VCMD(get_dlimit,       3, VCA_NONE,   VCF_INFO);
23640 +       __VCMD(get_prio_bias,    3, VCA_VXI,    VCF_INFO);
23641 +
23642 +       /* lower admin commands */
23643 +       __VCMD(wait_exit,        4, VCA_VXI,    VCF_INFO);
23644 +       __VCMD(ctx_create_v0,    5, VCA_NONE,   0);
23645 +       __VCMD(ctx_create,       5, VCA_NONE,   0);
23646 +       __VCMD(ctx_migrate_v0,   5, VCA_VXI,    VCF_ADMIN);
23647 +       __VCMD(ctx_migrate,      5, VCA_VXI,    VCF_ADMIN);
23648 +       __VCMD(enter_space_v0,   5, VCA_VXI,    VCF_ADMIN);
23649 +       __VCMD(enter_space_v1,   5, VCA_VXI,    VCF_ADMIN);
23650 +       __VCMD(enter_space,      5, VCA_VXI,    VCF_ADMIN);
23651 +
23652 +       __VCMD(net_create_v0,    5, VCA_NONE,   0);
23653 +       __VCMD(net_create,       5, VCA_NONE,   0);
23654 +       __VCMD(net_migrate,      5, VCA_NXI,    VCF_ADMIN);
23655 +
23656 +       __VCMD(tag_migrate,      5, VCA_NONE,   VCF_ADMIN);
23657 +
23658 +       /* higher admin commands */
23659 +       __VCMD(ctx_kill,         6, VCA_VXI,    VCF_ARES);
23660 +       __VCMD(set_space_v1,     7, VCA_VXI,    VCF_ARES | VCF_SETUP);
23661 +       __VCMD(set_space,        7, VCA_VXI,    VCF_ARES | VCF_SETUP);
23662 +
23663 +       __VCMD(set_ccaps,        7, VCA_VXI,    VCF_ARES | VCF_SETUP);
23664 +       __VCMD(set_bcaps,        7, VCA_VXI,    VCF_ARES | VCF_SETUP);
23665 +       __VCMD(set_cflags,       7, VCA_VXI,    VCF_ARES | VCF_SETUP);
23666 +       __VCMD(set_umask,        7, VCA_VXI,    VCF_ARES | VCF_SETUP);
23667 +       __VCMD(set_badness,      7, VCA_VXI,    VCF_ARES | VCF_SETUP);
23668 +
23669 +       __VCMD(set_vhi_name,     7, VCA_VXI,    VCF_ARES | VCF_SETUP);
23670 +       __VCMD(set_rlimit,       7, VCA_VXI,    VCF_ARES | VCF_SETUP);
23671 +       __VCMD(set_prio_bias,    7, VCA_VXI,    VCF_ARES | VCF_SETUP);
23672 +
23673 +       __VCMD(set_ncaps,        7, VCA_NXI,    VCF_ARES | VCF_SETUP);
23674 +       __VCMD(set_nflags,       7, VCA_NXI,    VCF_ARES | VCF_SETUP);
23675 +       __VCMD(net_add,          8, VCA_NXI,    VCF_ARES | VCF_SETUP);
23676 +       __VCMD(net_remove,       8, VCA_NXI,    VCF_ARES | VCF_SETUP);
23677 +       __VCMD(net_add_ipv4_v1,  8, VCA_NXI,    VCF_ARES | VCF_SETUP);
23678 +       __VCMD(net_rem_ipv4_v1,  8, VCA_NXI,    VCF_ARES | VCF_SETUP);
23679 +       __VCMD(net_add_ipv4,     8, VCA_NXI,    VCF_ARES | VCF_SETUP);
23680 +       __VCMD(net_rem_ipv4,     8, VCA_NXI,    VCF_ARES | VCF_SETUP);
23681 +#ifdef CONFIG_IPV6
23682 +       __VCMD(net_add_ipv6,     8, VCA_NXI,    VCF_ARES | VCF_SETUP);
23683 +       __VCMD(net_remove_ipv6,  8, VCA_NXI,    VCF_ARES | VCF_SETUP);
23684 +#endif
23685 +       __VCMD(set_iattr,        7, VCA_NONE,   0);
23686 +       __VCMD(fset_iattr,       7, VCA_NONE,   0);
23687 +       __VCMD(set_dlimit,       7, VCA_NONE,   VCF_ARES);
23688 +       __VCMD(add_dlimit,       8, VCA_NONE,   VCF_ARES);
23689 +       __VCMD(rem_dlimit,       8, VCA_NONE,   VCF_ARES);
23690 +
23691 +#ifdef CONFIG_VSERVER_DEVICE
23692 +       __VCMD(set_mapping,      8, VCA_VXI,    VCF_ARES|VCF_ZIDOK);
23693 +       __VCMD(unset_mapping,    8, VCA_VXI,    VCF_ARES|VCF_ZIDOK);
23694 +#endif
23695 +       /* debug level admin commands */
23696 +#ifdef CONFIG_VSERVER_HISTORY
23697 +       __VCMD(dump_history,     9, VCA_NONE,   0);
23698 +       __VCMD(read_history,     9, VCA_NONE,   0);
23699 +#endif
23700 +
23701 +       default:
23702 +               perm = -1;
23703 +       }
23704 +
23705 +       vxdprintk(VXD_CBIT(switch, 0),
23706 +               "vc: VCMD_%02d_%d[%d], %d,%p [%d,%d,%x,%x]",
23707 +               VC_CATEGORY(cmd), VC_COMMAND(cmd),
23708 +               VC_VERSION(cmd), id, data, compat,
23709 +               perm, args, flags);
23710 +
23711 +       ret = -ENOSYS;
23712 +       if (perm < 0)
23713 +               goto out;
23714 +
23715 +       state = 1;
23716 +       if (!capable(CAP_CONTEXT))
23717 +               goto out;
23718 +
23719 +       state = 2;
23720 +       /* moved here from the individual commands */
23721 +       ret = -EPERM;
23722 +       if ((perm > 1) && !capable(CAP_SYS_ADMIN))
23723 +               goto out;
23724 +
23725 +       state = 3;
23726 +       /* vcmd involves resource management  */
23727 +       ret = -EPERM;
23728 +       if ((flags & VCF_ARES) && !capable(CAP_SYS_RESOURCE))
23729 +               goto out;
23730 +
23731 +       state = 4;
23732 +       /* various legacy exceptions */
23733 +       switch (cmd) {
23734 +       /* will go away when spectator is a cap */
23735 +       case VCMD_ctx_migrate_v0:
23736 +       case VCMD_ctx_migrate:
23737 +               if (id == 1) {
23738 +                       current->xid = 1;
23739 +                       ret = 1;
23740 +                       goto out;
23741 +               }
23742 +               break;
23743 +
23744 +       /* will go away when spectator is a cap */
23745 +       case VCMD_net_migrate:
23746 +               if (id == 1) {
23747 +                       current->nid = 1;
23748 +                       ret = 1;
23749 +                       goto out;
23750 +               }
23751 +               break;
23752 +       }
23753 +
23754 +       /* vcmds are fine by default */
23755 +       permit = 1;
23756 +
23757 +       /* admin type vcmds require admin ... */
23758 +       if (flags & VCF_ADMIN)
23759 +               permit = vx_check(0, VS_ADMIN) ? 1 : 0;
23760 +
23761 +       /* ... but setup type vcmds override that */
23762 +       if (!permit && (flags & VCF_SETUP))
23763 +               permit = vx_flags(VXF_STATE_SETUP, 0) ? 2 : 0;
23764 +
23765 +       state = 5;
23766 +       ret = -EPERM;
23767 +       if (!permit)
23768 +               goto out;
23769 +
23770 +       state = 6;
23771 +       if (!id && (flags & VCF_ZIDOK))
23772 +               goto skip_id;
23773 +
23774 +       ret = -ESRCH;
23775 +       if (args & VCA_VXI) {
23776 +               vxi = lookup_vx_info(id);
23777 +               if (!vxi)
23778 +                       goto out;
23779 +
23780 +               if ((flags & VCF_ADMIN) &&
23781 +                       /* special case kill for shutdown */
23782 +                       (cmd != VCMD_ctx_kill) &&
23783 +                       /* can context be administrated? */
23784 +                       !vx_info_flags(vxi, VXF_STATE_ADMIN, 0)) {
23785 +                       ret = -EACCES;
23786 +                       goto out_vxi;
23787 +               }
23788 +       }
23789 +       state = 7;
23790 +       if (args & VCA_NXI) {
23791 +               nxi = lookup_nx_info(id);
23792 +               if (!nxi)
23793 +                       goto out_vxi;
23794 +
23795 +               if ((flags & VCF_ADMIN) &&
23796 +                       /* can context be administrated? */
23797 +                       !nx_info_flags(nxi, NXF_STATE_ADMIN, 0)) {
23798 +                       ret = -EACCES;
23799 +                       goto out_nxi;
23800 +               }
23801 +       }
23802 +skip_id:
23803 +       state = 8;
23804 +       ret = do_vcmd(cmd, id, vxi, nxi, data, compat);
23805 +
23806 +out_nxi:
23807 +       if ((args & VCA_NXI) && nxi)
23808 +               put_nx_info(nxi);
23809 +out_vxi:
23810 +       if ((args & VCA_VXI) && vxi)
23811 +               put_vx_info(vxi);
23812 +out:
23813 +       vxdprintk(VXD_CBIT(switch, 1),
23814 +               "vc: VCMD_%02d_%d[%d] = %08lx(%ld) [%d,%d]",
23815 +               VC_CATEGORY(cmd), VC_COMMAND(cmd),
23816 +               VC_VERSION(cmd), ret, ret, state, permit);
23817 +       return ret;
23818 +}
23819 +
23820 +asmlinkage long
23821 +sys_vserver(uint32_t cmd, uint32_t id, void __user *data)
23822 +{
23823 +       return do_vserver(cmd, id, data, 0);
23824 +}
23825 +
23826 +#ifdef CONFIG_COMPAT
23827 +
23828 +asmlinkage long
23829 +sys32_vserver(uint32_t cmd, uint32_t id, void __user *data)
23830 +{
23831 +       return do_vserver(cmd, id, data, 1);
23832 +}
23833 +
23834 +#endif /* CONFIG_COMPAT */
23835 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/sysctl.c linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/sysctl.c
23836 --- linux-3.0-rc5/kernel/vserver/sysctl.c       1970-01-01 01:00:00.000000000 +0100
23837 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/sysctl.c  2011-06-10 22:11:24.000000000 +0200
23838 @@ -0,0 +1,241 @@
23839 +/*
23840 + *  kernel/vserver/sysctl.c
23841 + *
23842 + *  Virtual Context Support
23843 + *
23844 + *  Copyright (C) 2004-2007  Herbert Pötzl
23845 + *
23846 + *  V0.01  basic structure
23847 + *
23848 + */
23849 +
23850 +#include <linux/module.h>
23851 +#include <linux/ctype.h>
23852 +#include <linux/sysctl.h>
23853 +#include <linux/parser.h>
23854 +#include <asm/uaccess.h>
23855 +
23856 +enum {
23857 +       CTL_DEBUG_ERROR         = 0,
23858 +       CTL_DEBUG_SWITCH        = 1,
23859 +       CTL_DEBUG_XID,
23860 +       CTL_DEBUG_NID,
23861 +       CTL_DEBUG_TAG,
23862 +       CTL_DEBUG_NET,
23863 +       CTL_DEBUG_LIMIT,
23864 +       CTL_DEBUG_CRES,
23865 +       CTL_DEBUG_DLIM,
23866 +       CTL_DEBUG_QUOTA,
23867 +       CTL_DEBUG_CVIRT,
23868 +       CTL_DEBUG_SPACE,
23869 +       CTL_DEBUG_MISC,
23870 +};
23871 +
23872 +
23873 +unsigned int vx_debug_switch   = 0;
23874 +unsigned int vx_debug_xid      = 0;
23875 +unsigned int vx_debug_nid      = 0;
23876 +unsigned int vx_debug_tag      = 0;
23877 +unsigned int vx_debug_net      = 0;
23878 +unsigned int vx_debug_limit    = 0;
23879 +unsigned int vx_debug_cres     = 0;
23880 +unsigned int vx_debug_dlim     = 0;
23881 +unsigned int vx_debug_quota    = 0;
23882 +unsigned int vx_debug_cvirt    = 0;
23883 +unsigned int vx_debug_space    = 0;
23884 +unsigned int vx_debug_misc     = 0;
23885 +
23886 +
23887 +static struct ctl_table_header *vserver_table_header;
23888 +static ctl_table vserver_root_table[];
23889 +
23890 +
23891 +void vserver_register_sysctl(void)
23892 +{
23893 +       if (!vserver_table_header) {
23894 +               vserver_table_header = register_sysctl_table(vserver_root_table);
23895 +       }
23896 +
23897 +}
23898 +
23899 +void vserver_unregister_sysctl(void)
23900 +{
23901 +       if (vserver_table_header) {
23902 +               unregister_sysctl_table(vserver_table_header);
23903 +               vserver_table_header = NULL;
23904 +       }
23905 +}
23906 +
23907 +
23908 +static int proc_dodebug(ctl_table *table, int write,
23909 +       void __user *buffer, size_t *lenp, loff_t *ppos)
23910 +{
23911 +       char            tmpbuf[20], *p, c;
23912 +       unsigned int    value;
23913 +       size_t          left, len;
23914 +
23915 +       if ((*ppos && !write) || !*lenp) {
23916 +               *lenp = 0;
23917 +               return 0;
23918 +       }
23919 +
23920 +       left = *lenp;
23921 +
23922 +       if (write) {
23923 +               if (!access_ok(VERIFY_READ, buffer, left))
23924 +                       return -EFAULT;
23925 +               p = (char *)buffer;
23926 +               while (left && __get_user(c, p) >= 0 && isspace(c))
23927 +                       left--, p++;
23928 +               if (!left)
23929 +                       goto done;
23930 +
23931 +               if (left > sizeof(tmpbuf) - 1)
23932 +                       return -EINVAL;
23933 +               if (copy_from_user(tmpbuf, p, left))
23934 +                       return -EFAULT;
23935 +               tmpbuf[left] = '\0';
23936 +
23937 +               for (p = tmpbuf, value = 0; '0' <= *p && *p <= '9'; p++, left--)
23938 +                       value = 10 * value + (*p - '0');
23939 +               if (*p && !isspace(*p))
23940 +                       return -EINVAL;
23941 +               while (left && isspace(*p))
23942 +                       left--, p++;
23943 +               *(unsigned int *)table->data = value;
23944 +       } else {
23945 +               if (!access_ok(VERIFY_WRITE, buffer, left))
23946 +                       return -EFAULT;
23947 +               len = sprintf(tmpbuf, "%d", *(unsigned int *)table->data);
23948 +               if (len > left)
23949 +                       len = left;
23950 +               if (__copy_to_user(buffer, tmpbuf, len))
23951 +                       return -EFAULT;
23952 +               if ((left -= len) > 0) {
23953 +                       if (put_user('\n', (char *)buffer + len))
23954 +                               return -EFAULT;
23955 +                       left--;
23956 +               }
23957 +       }
23958 +
23959 +done:
23960 +       *lenp -= left;
23961 +       *ppos += *lenp;
23962 +       return 0;
23963 +}
23964 +
23965 +static int zero;
23966 +
23967 +#define        CTL_ENTRY(ctl, name)                            \
23968 +       {                                               \
23969 +               .procname       = #name,                \
23970 +               .data           = &vx_ ## name,         \
23971 +               .maxlen         = sizeof(int),          \
23972 +               .mode           = 0644,                 \
23973 +               .proc_handler   = &proc_dodebug,        \
23974 +               .extra1         = &zero,                \
23975 +               .extra2         = &zero,                \
23976 +       }
23977 +
23978 +static ctl_table vserver_debug_table[] = {
23979 +       CTL_ENTRY(CTL_DEBUG_SWITCH,     debug_switch),
23980 +       CTL_ENTRY(CTL_DEBUG_XID,        debug_xid),
23981 +       CTL_ENTRY(CTL_DEBUG_NID,        debug_nid),
23982 +       CTL_ENTRY(CTL_DEBUG_TAG,        debug_tag),
23983 +       CTL_ENTRY(CTL_DEBUG_NET,        debug_net),
23984 +       CTL_ENTRY(CTL_DEBUG_LIMIT,      debug_limit),
23985 +       CTL_ENTRY(CTL_DEBUG_CRES,       debug_cres),
23986 +       CTL_ENTRY(CTL_DEBUG_DLIM,       debug_dlim),
23987 +       CTL_ENTRY(CTL_DEBUG_QUOTA,      debug_quota),
23988 +       CTL_ENTRY(CTL_DEBUG_CVIRT,      debug_cvirt),
23989 +       CTL_ENTRY(CTL_DEBUG_SPACE,      debug_space),
23990 +       CTL_ENTRY(CTL_DEBUG_MISC,       debug_misc),
23991 +       { 0 }
23992 +};
23993 +
23994 +static ctl_table vserver_root_table[] = {
23995 +       {
23996 +               .procname       = "vserver",
23997 +               .mode           = 0555,
23998 +               .child          = vserver_debug_table
23999 +       },
24000 +       { 0 }
24001 +};
24002 +
24003 +
24004 +static match_table_t tokens = {
24005 +       { CTL_DEBUG_SWITCH,     "switch=%x"     },
24006 +       { CTL_DEBUG_XID,        "xid=%x"        },
24007 +       { CTL_DEBUG_NID,        "nid=%x"        },
24008 +       { CTL_DEBUG_TAG,        "tag=%x"        },
24009 +       { CTL_DEBUG_NET,        "net=%x"        },
24010 +       { CTL_DEBUG_LIMIT,      "limit=%x"      },
24011 +       { CTL_DEBUG_CRES,       "cres=%x"       },
24012 +       { CTL_DEBUG_DLIM,       "dlim=%x"       },
24013 +       { CTL_DEBUG_QUOTA,      "quota=%x"      },
24014 +       { CTL_DEBUG_CVIRT,      "cvirt=%x"      },
24015 +       { CTL_DEBUG_SPACE,      "space=%x"      },
24016 +       { CTL_DEBUG_MISC,       "misc=%x"       },
24017 +       { CTL_DEBUG_ERROR,      NULL            }
24018 +};
24019 +
24020 +#define        HANDLE_CASE(id, name, val)                              \
24021 +       case CTL_DEBUG_ ## id:                                  \
24022 +               vx_debug_ ## name = val;                        \
24023 +               printk("vs_debug_" #name "=0x%x\n", val);       \
24024 +               break
24025 +
24026 +
24027 +static int __init vs_debug_setup(char *str)
24028 +{
24029 +       char *p;
24030 +       int token;
24031 +
24032 +       printk("vs_debug_setup(%s)\n", str);
24033 +       while ((p = strsep(&str, ",")) != NULL) {
24034 +               substring_t args[MAX_OPT_ARGS];
24035 +               unsigned int value;
24036 +
24037 +               if (!*p)
24038 +                       continue;
24039 +
24040 +               token = match_token(p, tokens, args);
24041 +               value = (token > 0) ? simple_strtoul(args[0].from, NULL, 0) : 0;
24042 +
24043 +               switch (token) {
24044 +               HANDLE_CASE(SWITCH, switch, value);
24045 +               HANDLE_CASE(XID,    xid,    value);
24046 +               HANDLE_CASE(NID,    nid,    value);
24047 +               HANDLE_CASE(TAG,    tag,    value);
24048 +               HANDLE_CASE(NET,    net,    value);
24049 +               HANDLE_CASE(LIMIT,  limit,  value);
24050 +               HANDLE_CASE(CRES,   cres,   value);
24051 +               HANDLE_CASE(DLIM,   dlim,   value);
24052 +               HANDLE_CASE(QUOTA,  quota,  value);
24053 +               HANDLE_CASE(CVIRT,  cvirt,  value);
24054 +               HANDLE_CASE(SPACE,  space,  value);
24055 +               HANDLE_CASE(MISC,   misc,   value);
24056 +               default:
24057 +                       return -EINVAL;
24058 +                       break;
24059 +               }
24060 +       }
24061 +       return 1;
24062 +}
24063 +
24064 +__setup("vsdebug=", vs_debug_setup);
24065 +
24066 +
24067 +
24068 +EXPORT_SYMBOL_GPL(vx_debug_switch);
24069 +EXPORT_SYMBOL_GPL(vx_debug_xid);
24070 +EXPORT_SYMBOL_GPL(vx_debug_nid);
24071 +EXPORT_SYMBOL_GPL(vx_debug_net);
24072 +EXPORT_SYMBOL_GPL(vx_debug_limit);
24073 +EXPORT_SYMBOL_GPL(vx_debug_cres);
24074 +EXPORT_SYMBOL_GPL(vx_debug_dlim);
24075 +EXPORT_SYMBOL_GPL(vx_debug_quota);
24076 +EXPORT_SYMBOL_GPL(vx_debug_cvirt);
24077 +EXPORT_SYMBOL_GPL(vx_debug_space);
24078 +EXPORT_SYMBOL_GPL(vx_debug_misc);
24079 +
24080 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/tag.c linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/tag.c
24081 --- linux-3.0-rc5/kernel/vserver/tag.c  1970-01-01 01:00:00.000000000 +0100
24082 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/tag.c     2011-06-10 22:11:24.000000000 +0200
24083 @@ -0,0 +1,63 @@
24084 +/*
24085 + *  linux/kernel/vserver/tag.c
24086 + *
24087 + *  Virtual Server: Shallow Tag Space
24088 + *
24089 + *  Copyright (C) 2007  Herbert Pötzl
24090 + *
24091 + *  V0.01  basic implementation
24092 + *
24093 + */
24094 +
24095 +#include <linux/sched.h>
24096 +#include <linux/vserver/debug.h>
24097 +#include <linux/vs_pid.h>
24098 +#include <linux/vs_tag.h>
24099 +
24100 +#include <linux/vserver/tag_cmd.h>
24101 +
24102 +
24103 +int dx_migrate_task(struct task_struct *p, tag_t tag)
24104 +{
24105 +       if (!p)
24106 +               BUG();
24107 +
24108 +       vxdprintk(VXD_CBIT(tag, 5),
24109 +               "dx_migrate_task(%p[#%d],#%d)", p, p->tag, tag);
24110 +
24111 +       task_lock(p);
24112 +       p->tag = tag;
24113 +       task_unlock(p);
24114 +
24115 +       vxdprintk(VXD_CBIT(tag, 5),
24116 +               "moved task %p into [#%d]", p, tag);
24117 +       return 0;
24118 +}
24119 +
24120 +/* vserver syscall commands below here */
24121 +
24122 +/* taks xid and vx_info functions */
24123 +
24124 +
24125 +int vc_task_tag(uint32_t id)
24126 +{
24127 +       tag_t tag;
24128 +
24129 +       if (id) {
24130 +               struct task_struct *tsk;
24131 +               rcu_read_lock();
24132 +               tsk = find_task_by_real_pid(id);
24133 +               tag = (tsk) ? tsk->tag : -ESRCH;
24134 +               rcu_read_unlock();
24135 +       } else
24136 +               tag = dx_current_tag();
24137 +       return tag;
24138 +}
24139 +
24140 +
24141 +int vc_tag_migrate(uint32_t tag)
24142 +{
24143 +       return dx_migrate_task(current, tag & 0xFFFF);
24144 +}
24145 +
24146 +
24147 diff -NurpP --minimal linux-3.0-rc5/kernel/vserver/vci_config.h linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/vci_config.h
24148 --- linux-3.0-rc5/kernel/vserver/vci_config.h   1970-01-01 01:00:00.000000000 +0100
24149 +++ linux-3.0-rc5-vs2.3.1-pre3/kernel/vserver/vci_config.h      2011-06-10 22:11:24.000000000 +0200
24150 @@ -0,0 +1,76 @@
24151 +
24152 +/*  interface version */
24153 +
24154 +#define VCI_VERSION            0x00020308
24155 +
24156 +
24157 +enum {
24158 +       VCI_KCBIT_NO_DYNAMIC = 0,
24159 +
24160 +       VCI_KCBIT_PROC_SECURE = 4,
24161 +       /* VCI_KCBIT_HARDCPU = 5, */
24162 +       /* VCI_KCBIT_IDLELIMIT = 6, */
24163 +       /* VCI_KCBIT_IDLETIME = 7, */
24164 +
24165 +       VCI_KCBIT_COWBL = 8,
24166 +       VCI_KCBIT_FULLCOWBL = 9,
24167 +       VCI_KCBIT_SPACES = 10,
24168 +       VCI_KCBIT_NETV2 = 11,
24169 +       VCI_KCBIT_MEMCG = 12,
24170 +
24171 +       VCI_KCBIT_DEBUG = 16,
24172 +       VCI_KCBIT_HISTORY = 20,
24173 +       VCI_KCBIT_TAGGED = 24,
24174 +       VCI_KCBIT_PPTAG = 28,
24175 +
24176 +       VCI_KCBIT_MORE = 31,
24177 +};
24178 +
24179 +
24180 +static inline uint32_t vci_kernel_config(void)
24181 +{
24182 +       return
24183 +       (1 << VCI_KCBIT_NO_DYNAMIC) |
24184 +
24185 +       /* configured features */
24186 +#ifdef CONFIG_VSERVER_PROC_SECURE
24187 +       (1 << VCI_KCBIT_PROC_SECURE) |
24188 +#endif
24189 +#ifdef CONFIG_VSERVER_COWBL
24190 +       (1 << VCI_KCBIT_COWBL) |
24191 +       (1 << VCI_KCBIT_FULLCOWBL) |
24192 +#endif
24193 +       (1 << VCI_KCBIT_SPACES) |
24194 +       (1 << VCI_KCBIT_NETV2) |
24195 +#ifdef CONFIG_CGROUP_MEM_RES_CTLR
24196 +       (1 << VCI_KCBIT_MEMCG) |
24197 +#endif
24198 +
24199 +       /* debug options */
24200 +#ifdef CONFIG_VSERVER_DEBUG
24201 +       (1 << VCI_KCBIT_DEBUG) |
24202 +#endif
24203 +#ifdef CONFIG_VSERVER_HISTORY
24204 +       (1 << VCI_KCBIT_HISTORY) |
24205 +#endif
24206 +
24207 +       /* inode context tagging */
24208 +#if    defined(CONFIG_TAGGING_NONE)
24209 +       (0 << VCI_KCBIT_TAGGED) |
24210 +#elif  defined(CONFIG_TAGGING_UID16)
24211 +       (1 << VCI_KCBIT_TAGGED) |
24212 +#elif  defined(CONFIG_TAGGING_GID16)
24213 +       (2 << VCI_KCBIT_TAGGED) |
24214 +#elif  defined(CONFIG_TAGGING_ID24)
24215 +       (3 << VCI_KCBIT_TAGGED) |
24216 +#elif  defined(CONFIG_TAGGING_INTERN)
24217 +       (4 << VCI_KCBIT_TAGGED) |
24218 +#elif  defined(CONFIG_TAGGING_RUNTIME)
24219 +       (5 << VCI_KCBIT_TAGGED) |
24220 +#else
24221 +       (7 << VCI_KCBIT_TAGGED) |
24222 +#endif
24223 +       (1 << VCI_KCBIT_PPTAG) |
24224 +       0;
24225 +}
24226 +
24227 diff -NurpP --minimal linux-3.0-rc5/mm/filemap_xip.c linux-3.0-rc5-vs2.3.1-pre3/mm/filemap_xip.c
24228 --- linux-3.0-rc5/mm/filemap_xip.c      2011-07-01 11:14:23.000000000 +0200
24229 +++ linux-3.0-rc5-vs2.3.1-pre3/mm/filemap_xip.c 2011-06-10 22:11:24.000000000 +0200
24230 @@ -18,6 +18,7 @@
24231  #include <linux/seqlock.h>
24232  #include <linux/mutex.h>
24233  #include <linux/gfp.h>
24234 +#include <linux/vs_memory.h>
24235  #include <asm/tlbflush.h>
24236  #include <asm/io.h>
24237
24238 diff -NurpP --minimal linux-3.0-rc5/mm/fremap.c linux-3.0-rc5-vs2.3.1-pre3/mm/fremap.c
24239 --- linux-3.0-rc5/mm/fremap.c   2011-07-01 11:14:23.000000000 +0200
24240 +++ linux-3.0-rc5-vs2.3.1-pre3/mm/fremap.c      2011-06-10 22:11:24.000000000 +0200
24241 @@ -16,6 +16,7 @@
24242  #include <linux/module.h>
24243  #include <linux/syscalls.h>
24244  #include <linux/mmu_notifier.h>
24245 +#include <linux/vs_memory.h>
24246
24247  #include <asm/mmu_context.h>
24248  #include <asm/cacheflush.h>
24249 diff -NurpP --minimal linux-3.0-rc5/mm/hugetlb.c linux-3.0-rc5-vs2.3.1-pre3/mm/hugetlb.c
24250 --- linux-3.0-rc5/mm/hugetlb.c  2011-07-01 11:14:23.000000000 +0200
24251 +++ linux-3.0-rc5-vs2.3.1-pre3/mm/hugetlb.c     2011-06-22 12:39:16.000000000 +0200
24252 @@ -28,6 +28,7 @@
24253
24254  #include <linux/hugetlb.h>
24255  #include <linux/node.h>
24256 +#include <linux/vs_memory.h>
24257  #include "internal.h"
24258
24259  const unsigned long hugetlb_zero = 0, hugetlb_infinity = ~0UL;
24260 diff -NurpP --minimal linux-3.0-rc5/mm/memcontrol.c linux-3.0-rc5-vs2.3.1-pre3/mm/memcontrol.c
24261 --- linux-3.0-rc5/mm/memcontrol.c       2011-07-01 11:14:23.000000000 +0200
24262 +++ linux-3.0-rc5-vs2.3.1-pre3/mm/memcontrol.c  2011-07-01 11:35:35.000000000 +0200
24263 @@ -736,6 +736,31 @@ struct mem_cgroup *mem_cgroup_from_task(
24264                                 struct mem_cgroup, css);
24265  }
24266
24267 +u64 mem_cgroup_res_read_u64(struct mem_cgroup *mem, int member)
24268 +{
24269 +       return res_counter_read_u64(&mem->res, member);
24270 +}
24271 +
24272 +u64 mem_cgroup_memsw_read_u64(struct mem_cgroup *mem, int member)
24273 +{
24274 +       return res_counter_read_u64(&mem->memsw, member);
24275 +}
24276 +
24277 +s64 mem_cgroup_stat_read_cache(struct mem_cgroup *mem)
24278 +{
24279 +       return mem_cgroup_read_stat(mem, MEM_CGROUP_STAT_CACHE);
24280 +}
24281 +
24282 +s64 mem_cgroup_stat_read_anon(struct mem_cgroup *mem)
24283 +{
24284 +       return mem_cgroup_read_stat(mem, MEM_CGROUP_STAT_RSS);
24285 +}
24286 +
24287 +s64 mem_cgroup_stat_read_mapped(struct mem_cgroup *mem)
24288 +{
24289 +       return mem_cgroup_read_stat(mem, MEM_CGROUP_STAT_FILE_MAPPED);
24290 +}
24291 +
24292  struct mem_cgroup *try_get_mem_cgroup_from_mm(struct mm_struct *mm)
24293  {
24294         struct mem_cgroup *mem = NULL;
24295 diff -NurpP --minimal linux-3.0-rc5/mm/memory.c linux-3.0-rc5-vs2.3.1-pre3/mm/memory.c
24296 --- linux-3.0-rc5/mm/memory.c   2011-07-01 11:14:23.000000000 +0200
24297 +++ linux-3.0-rc5-vs2.3.1-pre3/mm/memory.c      2011-07-01 11:35:35.000000000 +0200
24298 @@ -3331,6 +3331,7 @@ int handle_pte_fault(struct mm_struct *m
24299  {
24300         pte_t entry;
24301         spinlock_t *ptl;
24302 +       int ret = 0, type = VXPT_UNKNOWN;
24303
24304         entry = *pte;
24305         if (!pte_present(entry)) {
24306 @@ -3355,9 +3356,12 @@ int handle_pte_fault(struct mm_struct *m
24307         if (unlikely(!pte_same(*pte, entry)))
24308                 goto unlock;
24309         if (flags & FAULT_FLAG_WRITE) {
24310 -               if (!pte_write(entry))
24311 -                       return do_wp_page(mm, vma, address,
24312 +               if (!pte_write(entry)) {
24313 +                       ret = do_wp_page(mm, vma, address,
24314                                         pte, pmd, ptl, entry);
24315 +                       type = VXPT_WRITE;
24316 +                       goto out;
24317 +               }
24318                 entry = pte_mkdirty(entry);
24319         }
24320         entry = pte_mkyoung(entry);
24321 @@ -3375,7 +3379,10 @@ int handle_pte_fault(struct mm_struct *m
24322         }
24323  unlock:
24324         pte_unmap_unlock(pte, ptl);
24325 -       return 0;
24326 +       ret = 0;
24327 +out:
24328 +       vx_page_fault(mm, vma, type, ret);
24329 +       return ret;
24330  }
24331
24332  /*
24333 diff -NurpP --minimal linux-3.0-rc5/mm/mremap.c linux-3.0-rc5-vs2.3.1-pre3/mm/mremap.c
24334 --- linux-3.0-rc5/mm/mremap.c   2011-07-01 11:14:23.000000000 +0200
24335 +++ linux-3.0-rc5-vs2.3.1-pre3/mm/mremap.c      2011-06-10 22:11:24.000000000 +0200
24336 @@ -19,6 +19,7 @@
24337  #include <linux/security.h>
24338  #include <linux/syscalls.h>
24339  #include <linux/mmu_notifier.h>
24340 +#include <linux/vs_memory.h>
24341
24342  #include <asm/uaccess.h>
24343  #include <asm/cacheflush.h>
24344 diff -NurpP --minimal linux-3.0-rc5/mm/oom_kill.c linux-3.0-rc5-vs2.3.1-pre3/mm/oom_kill.c
24345 --- linux-3.0-rc5/mm/oom_kill.c 2011-07-01 11:14:23.000000000 +0200
24346 +++ linux-3.0-rc5-vs2.3.1-pre3/mm/oom_kill.c    2011-06-10 22:11:24.000000000 +0200
24347 @@ -32,6 +32,9 @@
24348  #include <linux/mempolicy.h>
24349  #include <linux/security.h>
24350  #include <linux/ptrace.h>
24351 +#include <linux/reboot.h>
24352 +#include <linux/vs_memory.h>
24353 +#include <linux/vs_context.h>
24354
24355  int sysctl_panic_on_oom;
24356  int sysctl_oom_kill_allocating_task;
24357 @@ -134,11 +137,18 @@ struct task_struct *find_lock_task_mm(st
24358  static bool oom_unkillable_task(struct task_struct *p,
24359                 const struct mem_cgroup *mem, const nodemask_t *nodemask)
24360  {
24361 -       if (is_global_init(p))
24362 +       unsigned xid = vx_current_xid();
24363 +
24364 +       /* skip the init task, global and per guest */
24365 +       if (task_is_init(p))
24366                 return true;
24367         if (p->flags & PF_KTHREAD)
24368                 return true;
24369
24370 +       /* skip other guest and host processes if oom in guest */
24371 +       if (xid && vx_task_xid(p) != xid)
24372 +               return true;
24373 +
24374         /* When mem_cgroup_out_of_memory() and p is not member of the group */
24375         if (mem && !task_in_mem_cgroup(p, mem))
24376                 return true;
24377 @@ -214,6 +224,18 @@ unsigned int oom_badness(struct task_str
24378         points += p->signal->oom_score_adj;
24379
24380         /*
24381 +        * add points for context badness and
24382 +        * reduce badness for processes belonging to
24383 +        * a different context
24384 +        */
24385 +
24386 +       points += vx_badness(p, p->mm);
24387 +
24388 +       if ((vx_current_xid() > 1) &&
24389 +               vx_current_xid() != vx_task_xid(p))
24390 +               points /= 16;
24391 +
24392 +       /*
24393          * Never return 0 for an eligible task that may be killed since it's
24394          * possible that no single user task uses more than 0.1% of memory and
24395          * no single admin tasks uses more than 3.0%.
24396 @@ -427,8 +449,8 @@ static int oom_kill_task(struct task_str
24397         /* mm cannot be safely dereferenced after task_unlock(p) */
24398         mm = p->mm;
24399
24400 -       pr_err("Killed process %d (%s) total-vm:%lukB, anon-rss:%lukB, file-rss:%lukB\n",
24401 -               task_pid_nr(p), p->comm, K(p->mm->total_vm),
24402 +       pr_err("Killed process %d:#%u (%s) total-vm:%lukB, anon-rss:%lukB, file-rss:%lukB\n",
24403 +               task_pid_nr(p), p->xid, p->comm, K(p->mm->total_vm),
24404                 K(get_mm_counter(p->mm, MM_ANONPAGES)),
24405                 K(get_mm_counter(p->mm, MM_FILEPAGES)));
24406         task_unlock(p);
24407 @@ -482,8 +504,8 @@ static int oom_kill_process(struct task_
24408         }
24409
24410         task_lock(p);
24411 -       pr_err("%s: Kill process %d (%s) score %d or sacrifice child\n",
24412 -               message, task_pid_nr(p), p->comm, points);
24413 +       pr_err("%s: Kill process %d:#%u (%s) score %d or sacrifice child\n",
24414 +               message, task_pid_nr(p), p->xid, p->comm, points);
24415         task_unlock(p);
24416
24417         /*
24418 @@ -584,6 +606,8 @@ int unregister_oom_notifier(struct notif
24419  }
24420  EXPORT_SYMBOL_GPL(unregister_oom_notifier);
24421
24422 +long vs_oom_action(unsigned int);
24423 +
24424  /*
24425   * Try to acquire the OOM killer lock for the zones in zonelist.  Returns zero
24426   * if a parallel OOM killing is already taking place that includes a zone in
24427 @@ -742,7 +766,12 @@ retry:
24428         if (!p) {
24429                 dump_header(NULL, gfp_mask, order, NULL, mpol_mask);
24430                 read_unlock(&tasklist_lock);
24431 -               panic("Out of memory and no killable processes...\n");
24432 +
24433 +               /* avoid panic for guest OOM */
24434 +               if (current->xid)
24435 +                       vs_oom_action(LINUX_REBOOT_CMD_OOM);
24436 +               else
24437 +                       panic("Out of memory and no killable processes...\n");
24438         }
24439
24440         if (oom_kill_process(p, gfp_mask, order, points, totalpages, NULL,
24441 diff -NurpP --minimal linux-3.0-rc5/mm/page_alloc.c linux-3.0-rc5-vs2.3.1-pre3/mm/page_alloc.c
24442 --- linux-3.0-rc5/mm/page_alloc.c       2011-07-01 11:14:23.000000000 +0200
24443 +++ linux-3.0-rc5-vs2.3.1-pre3/mm/page_alloc.c  2011-06-10 22:24:12.000000000 +0200
24444 @@ -57,6 +57,8 @@
24445  #include <linux/ftrace_event.h>
24446  #include <linux/memcontrol.h>
24447  #include <linux/prefetch.h>
24448 +#include <linux/vs_base.h>
24449 +#include <linux/vs_limit.h>
24450
24451  #include <asm/tlbflush.h>
24452  #include <asm/div64.h>
24453 @@ -2470,6 +2472,9 @@ void si_meminfo(struct sysinfo *val)
24454         val->totalhigh = totalhigh_pages;
24455         val->freehigh = nr_free_highpages();
24456         val->mem_unit = PAGE_SIZE;
24457 +
24458 +       if (vx_flags(VXF_VIRT_MEM, 0))
24459 +               vx_vsi_meminfo(val);
24460  }
24461
24462  EXPORT_SYMBOL(si_meminfo);
24463 @@ -2490,6 +2495,9 @@ void si_meminfo_node(struct sysinfo *val
24464         val->freehigh = 0;
24465  #endif
24466         val->mem_unit = PAGE_SIZE;
24467 +
24468 +       if (vx_flags(VXF_VIRT_MEM, 0))
24469 +               vx_vsi_meminfo(val);
24470  }
24471  #endif
24472
24473 diff -NurpP --minimal linux-3.0-rc5/mm/pgtable-generic.c linux-3.0-rc5-vs2.3.1-pre3/mm/pgtable-generic.c
24474 --- linux-3.0-rc5/mm/pgtable-generic.c  2011-03-15 18:07:42.000000000 +0100
24475 +++ linux-3.0-rc5-vs2.3.1-pre3/mm/pgtable-generic.c     2011-06-10 22:11:24.000000000 +0200
24476 @@ -6,6 +6,8 @@
24477   *  Copyright (C) 2010  Linus Torvalds
24478   */
24479
24480 +#include <linux/mm.h>
24481 +
24482  #include <linux/pagemap.h>
24483  #include <asm/tlb.h>
24484  #include <asm-generic/pgtable.h>
24485 diff -NurpP --minimal linux-3.0-rc5/mm/rmap.c linux-3.0-rc5-vs2.3.1-pre3/mm/rmap.c
24486 --- linux-3.0-rc5/mm/rmap.c     2011-07-01 11:14:23.000000000 +0200
24487 +++ linux-3.0-rc5-vs2.3.1-pre3/mm/rmap.c        2011-07-01 11:35:35.000000000 +0200
24488 @@ -57,6 +57,7 @@
24489  #include <linux/mmu_notifier.h>
24490  #include <linux/migrate.h>
24491  #include <linux/hugetlb.h>
24492 +#include <linux/vs_memory.h>
24493
24494  #include <asm/tlbflush.h>
24495
24496 diff -NurpP --minimal linux-3.0-rc5/mm/shmem.c linux-3.0-rc5-vs2.3.1-pre3/mm/shmem.c
24497 --- linux-3.0-rc5/mm/shmem.c    2011-07-01 11:14:23.000000000 +0200
24498 +++ linux-3.0-rc5-vs2.3.1-pre3/mm/shmem.c       2011-07-01 11:35:35.000000000 +0200
24499 @@ -1850,7 +1850,7 @@ static int shmem_statfs(struct dentry *d
24500  {
24501         struct shmem_sb_info *sbinfo = SHMEM_SB(dentry->d_sb);
24502
24503 -       buf->f_type = TMPFS_MAGIC;
24504 +       buf->f_type = TMPFS_SUPER_MAGIC;
24505         buf->f_bsize = PAGE_CACHE_SIZE;
24506         buf->f_namelen = NAME_MAX;
24507         if (sbinfo->max_blocks) {
24508 @@ -2605,7 +2605,7 @@ int shmem_fill_super(struct super_block
24509         sb->s_maxbytes = SHMEM_MAX_BYTES;
24510         sb->s_blocksize = PAGE_CACHE_SIZE;
24511         sb->s_blocksize_bits = PAGE_CACHE_SHIFT;
24512 -       sb->s_magic = TMPFS_MAGIC;
24513 +       sb->s_magic = TMPFS_SUPER_MAGIC;
24514         sb->s_op = &shmem_ops;
24515         sb->s_time_gran = 1;
24516  #ifdef CONFIG_TMPFS_XATTR
24517 diff -NurpP --minimal linux-3.0-rc5/mm/slab.c linux-3.0-rc5-vs2.3.1-pre3/mm/slab.c
24518 --- linux-3.0-rc5/mm/slab.c     2011-07-01 11:14:23.000000000 +0200
24519 +++ linux-3.0-rc5-vs2.3.1-pre3/mm/slab.c        2011-06-15 02:41:23.000000000 +0200
24520 @@ -411,6 +411,8 @@ static void kmem_list3_init(struct kmem_
24521  #define STATS_INC_FREEMISS(x)  do { } while (0)
24522  #endif
24523
24524 +#include "slab_vs.h"
24525 +
24526  #if DEBUG
24527
24528  /*
24529 @@ -3348,6 +3350,7 @@ retry:
24530
24531         obj = slab_get_obj(cachep, slabp, nodeid);
24532         check_slabp(cachep, slabp);
24533 +       vx_slab_alloc(cachep, flags);
24534         l3->free_objects--;
24535         /* move slabp to correct slabp list: */
24536         list_del(&slabp->list);
24537 @@ -3425,6 +3428,7 @@ __cache_alloc_node(struct kmem_cache *ca
24538         /* ___cache_alloc_node can fall back to other nodes */
24539         ptr = ____cache_alloc_node(cachep, flags, nodeid);
24540    out:
24541 +       vx_slab_alloc(cachep, flags);
24542         local_irq_restore(save_flags);
24543         ptr = cache_alloc_debugcheck_after(cachep, flags, ptr, caller);
24544         kmemleak_alloc_recursive(ptr, obj_size(cachep), 1, cachep->flags,
24545 @@ -3612,6 +3616,7 @@ static inline void __cache_free(struct k
24546         check_irq_off();
24547         kmemleak_free_recursive(objp, cachep->flags);
24548         objp = cache_free_debugcheck(cachep, objp, caller);
24549 +       vx_slab_free(cachep);
24550
24551         kmemcheck_slab_free(cachep, objp, obj_size(cachep));
24552
24553 diff -NurpP --minimal linux-3.0-rc5/mm/slab_vs.h linux-3.0-rc5-vs2.3.1-pre3/mm/slab_vs.h
24554 --- linux-3.0-rc5/mm/slab_vs.h  1970-01-01 01:00:00.000000000 +0100
24555 +++ linux-3.0-rc5-vs2.3.1-pre3/mm/slab_vs.h     2011-06-10 22:11:24.000000000 +0200
24556 @@ -0,0 +1,29 @@
24557 +
24558 +#include <linux/vserver/context.h>
24559 +
24560 +#include <linux/vs_context.h>
24561 +
24562 +static inline
24563 +void vx_slab_alloc(struct kmem_cache *cachep, gfp_t flags)
24564 +{
24565 +       int what = gfp_zone(cachep->gfpflags);
24566 +       struct vx_info *vxi = current_vx_info();
24567 +
24568 +       if (!vxi)
24569 +               return;
24570 +
24571 +       atomic_add(cachep->buffer_size, &vxi->cacct.slab[what]);
24572 +}
24573 +
24574 +static inline
24575 +void vx_slab_free(struct kmem_cache *cachep)
24576 +{
24577 +       int what = gfp_zone(cachep->gfpflags);
24578 +       struct vx_info *vxi = current_vx_info();
24579 +
24580 +       if (!vxi)
24581 +               return;
24582 +
24583 +       atomic_sub(cachep->buffer_size, &vxi->cacct.slab[what]);
24584 +}
24585 +
24586 diff -NurpP --minimal linux-3.0-rc5/mm/swapfile.c linux-3.0-rc5-vs2.3.1-pre3/mm/swapfile.c
24587 --- linux-3.0-rc5/mm/swapfile.c 2011-07-01 11:14:23.000000000 +0200
24588 +++ linux-3.0-rc5-vs2.3.1-pre3/mm/swapfile.c    2011-07-01 11:35:35.000000000 +0200
24589 @@ -37,6 +37,8 @@
24590  #include <asm/tlbflush.h>
24591  #include <linux/swapops.h>
24592  #include <linux/page_cgroup.h>
24593 +#include <linux/vs_base.h>
24594 +#include <linux/vs_memory.h>
24595
24596  static bool swap_count_continued(struct swap_info_struct *, pgoff_t,
24597                                  unsigned char);
24598 @@ -1759,6 +1761,16 @@ static int swap_show(struct seq_file *sw
24599
24600         if (si == SEQ_START_TOKEN) {
24601                 seq_puts(swap,"Filename\t\t\t\tType\t\tSize\tUsed\tPriority\n");
24602 +               if (vx_flags(VXF_VIRT_MEM, 0)) {
24603 +                       struct sysinfo si;
24604 +
24605 +                       vx_vsi_swapinfo(&si);
24606 +                       if (si.totalswap < (1 << 10))
24607 +                               return 0;
24608 +                       seq_printf(swap, "%s\t\t\t\t\t%s\t%lu\t%lu\t%d\n",
24609 +                               "hdv0", "partition", si.totalswap >> 10,
24610 +                               (si.totalswap - si.freeswap) >> 10, -1);
24611 +               }
24612                 return 0;
24613         }
24614
24615 @@ -2186,6 +2198,8 @@ void si_swapinfo(struct sysinfo *val)
24616         val->freeswap = nr_swap_pages + nr_to_be_unused;
24617         val->totalswap = total_swap_pages + nr_to_be_unused;
24618         spin_unlock(&swap_lock);
24619 +       if (vx_flags(VXF_VIRT_MEM, 0))
24620 +               vx_vsi_swapinfo(val);
24621  }
24622
24623  /*
24624 diff -NurpP --minimal linux-3.0-rc5/net/bridge/br_multicast.c linux-3.0-rc5-vs2.3.1-pre3/net/bridge/br_multicast.c
24625 --- linux-3.0-rc5/net/bridge/br_multicast.c     2011-07-01 11:14:26.000000000 +0200
24626 +++ linux-3.0-rc5-vs2.3.1-pre3/net/bridge/br_multicast.c        2011-06-22 12:39:16.000000000 +0200
24627 @@ -447,7 +447,7 @@ static struct sk_buff *br_ip6_multicast_
24628         ip6h->hop_limit = 1;
24629         ipv6_addr_set(&ip6h->daddr, htonl(0xff020000), 0, 0, htonl(1));
24630         ipv6_dev_get_saddr(dev_net(br->dev), br->dev, &ip6h->daddr, 0,
24631 -                          &ip6h->saddr);
24632 +                          &ip6h->saddr, NULL);
24633         ipv6_eth_mc_map(&ip6h->daddr, eth->h_dest);
24634
24635         hopopt = (u8 *)(ip6h + 1);
24636 diff -NurpP --minimal linux-3.0-rc5/net/core/dev.c linux-3.0-rc5-vs2.3.1-pre3/net/core/dev.c
24637 --- linux-3.0-rc5/net/core/dev.c        2011-07-01 11:14:26.000000000 +0200
24638 +++ linux-3.0-rc5-vs2.3.1-pre3/net/core/dev.c   2011-06-15 02:40:14.000000000 +0200
24639 @@ -127,6 +127,7 @@
24640  #include <linux/in.h>
24641  #include <linux/jhash.h>
24642  #include <linux/random.h>
24643 +#include <linux/vs_inet.h>
24644  #include <trace/events/napi.h>
24645  #include <trace/events/net.h>
24646  #include <trace/events/skb.h>
24647 @@ -609,7 +610,8 @@ struct net_device *__dev_get_by_name(str
24648         struct hlist_head *head = dev_name_hash(net, name);
24649
24650         hlist_for_each_entry(dev, p, head, name_hlist)
24651 -               if (!strncmp(dev->name, name, IFNAMSIZ))
24652 +               if (!strncmp(dev->name, name, IFNAMSIZ) &&
24653 +                   nx_dev_visible(current_nx_info(), dev))
24654                         return dev;
24655
24656         return NULL;
24657 @@ -635,7 +637,8 @@ struct net_device *dev_get_by_name_rcu(s
24658         struct hlist_head *head = dev_name_hash(net, name);
24659
24660         hlist_for_each_entry_rcu(dev, p, head, name_hlist)
24661 -               if (!strncmp(dev->name, name, IFNAMSIZ))
24662 +               if (!strncmp(dev->name, name, IFNAMSIZ) &&
24663 +                   nx_dev_visible(current_nx_info(), dev))
24664                         return dev;
24665
24666         return NULL;
24667 @@ -686,7 +689,8 @@ struct net_device *__dev_get_by_index(st
24668         struct hlist_head *head = dev_index_hash(net, ifindex);
24669
24670         hlist_for_each_entry(dev, p, head, index_hlist)
24671 -               if (dev->ifindex == ifindex)
24672 +               if ((dev->ifindex == ifindex) &&
24673 +                   nx_dev_visible(current_nx_info(), dev))
24674                         return dev;
24675
24676         return NULL;
24677 @@ -711,7 +715,8 @@ struct net_device *dev_get_by_index_rcu(
24678         struct hlist_head *head = dev_index_hash(net, ifindex);
24679
24680         hlist_for_each_entry_rcu(dev, p, head, index_hlist)
24681 -               if (dev->ifindex == ifindex)
24682 +               if ((dev->ifindex == ifindex) &&
24683 +                   nx_dev_visible(current_nx_info(), dev))
24684                         return dev;
24685
24686         return NULL;
24687 @@ -764,7 +769,8 @@ struct net_device *dev_getbyhwaddr_rcu(s
24688
24689         for_each_netdev_rcu(net, dev)
24690                 if (dev->type == type &&
24691 -                   !memcmp(dev->dev_addr, ha, dev->addr_len))
24692 +                   !memcmp(dev->dev_addr, ha, dev->addr_len) &&
24693 +                   nx_dev_visible(current_nx_info(), dev))
24694                         return dev;
24695
24696         return NULL;
24697 @@ -776,9 +782,11 @@ struct net_device *__dev_getfirstbyhwtyp
24698         struct net_device *dev;
24699
24700         ASSERT_RTNL();
24701 -       for_each_netdev(net, dev)
24702 -               if (dev->type == type)
24703 +       for_each_netdev(net, dev) {
24704 +               if ((dev->type == type) &&
24705 +                   nx_dev_visible(current_nx_info(), dev))
24706                         return dev;
24707 +       }
24708
24709         return NULL;
24710  }
24711 @@ -896,6 +904,8 @@ static int __dev_alloc_name(struct net *
24712                                 continue;
24713                         if (i < 0 || i >= max_netdevices)
24714                                 continue;
24715 +                       if (!nx_dev_visible(current_nx_info(), d))
24716 +                               continue;
24717
24718                         /*  avoid cases where sscanf is not exact inverse of printf */
24719                         snprintf(buf, IFNAMSIZ, name, i);
24720 @@ -3937,6 +3947,8 @@ static int dev_ifconf(struct net *net, c
24721
24722         total = 0;
24723         for_each_netdev(net, dev) {
24724 +               if (!nx_dev_visible(current_nx_info(), dev))
24725 +                       continue;
24726                 for (i = 0; i < NPROTO; i++) {
24727                         if (gifconf_list[i]) {
24728                                 int done;
24729 @@ -4011,6 +4023,10 @@ static void dev_seq_printf_stats(struct
24730         struct rtnl_link_stats64 temp;
24731         const struct rtnl_link_stats64 *stats = dev_get_stats(dev, &temp);
24732
24733 +       /* device visible inside network context? */
24734 +       if (!nx_dev_visible(current_nx_info(), dev))
24735 +               return;
24736 +
24737         seq_printf(seq, "%6s: %7llu %7llu %4llu %4llu %4llu %5llu %10llu %9llu "
24738                    "%8llu %7llu %4llu %4llu %4llu %5llu %7llu %10llu\n",
24739                    dev->name, stats->rx_bytes, stats->rx_packets,
24740 diff -NurpP --minimal linux-3.0-rc5/net/core/rtnetlink.c linux-3.0-rc5-vs2.3.1-pre3/net/core/rtnetlink.c
24741 --- linux-3.0-rc5/net/core/rtnetlink.c  2011-07-01 11:14:26.000000000 +0200
24742 +++ linux-3.0-rc5-vs2.3.1-pre3/net/core/rtnetlink.c     2011-06-10 22:11:24.000000000 +0200
24743 @@ -1015,6 +1015,8 @@ static int rtnl_dump_ifinfo(struct sk_bu
24744                 hlist_for_each_entry_rcu(dev, node, head, index_hlist) {
24745                         if (idx < s_idx)
24746                                 goto cont;
24747 +                       if (!nx_dev_visible(skb->sk->sk_nx_info, dev))
24748 +                               continue;
24749                         if (rtnl_fill_ifinfo(skb, dev, RTM_NEWLINK,
24750                                              NETLINK_CB(cb->skb).pid,
24751                                              cb->nlh->nlmsg_seq, 0,
24752 @@ -1848,6 +1850,9 @@ void rtmsg_ifinfo(int type, struct net_d
24753         struct sk_buff *skb;
24754         int err = -ENOBUFS;
24755
24756 +       if (!nx_dev_visible(current_nx_info(), dev))
24757 +               return;
24758 +
24759         skb = nlmsg_new(if_nlmsg_size(dev), GFP_KERNEL);
24760         if (skb == NULL)
24761                 goto errout;
24762 diff -NurpP --minimal linux-3.0-rc5/net/core/sock.c linux-3.0-rc5-vs2.3.1-pre3/net/core/sock.c
24763 --- linux-3.0-rc5/net/core/sock.c       2011-05-22 16:18:00.000000000 +0200
24764 +++ linux-3.0-rc5-vs2.3.1-pre3/net/core/sock.c  2011-06-10 22:11:24.000000000 +0200
24765 @@ -127,6 +127,10 @@
24766  #include <net/cls_cgroup.h>
24767
24768  #include <linux/filter.h>
24769 +#include <linux/vs_socket.h>
24770 +#include <linux/vs_limit.h>
24771 +#include <linux/vs_context.h>
24772 +#include <linux/vs_network.h>
24773
24774  #ifdef CONFIG_INET
24775  #include <net/tcp.h>
24776 @@ -1070,6 +1074,8 @@ static struct sock *sk_prot_alloc(struct
24777                         goto out_free_sec;
24778                 sk_tx_queue_clear(sk);
24779         }
24780 +               sock_vx_init(sk);
24781 +               sock_nx_init(sk);
24782
24783         return sk;
24784
24785 @@ -1169,6 +1175,11 @@ static void __sk_free(struct sock *sk)
24786                 put_cred(sk->sk_peer_cred);
24787         put_pid(sk->sk_peer_pid);
24788         put_net(sock_net(sk));
24789 +       vx_sock_dec(sk);
24790 +       clr_vx_info(&sk->sk_vx_info);
24791 +       sk->sk_xid = -1;
24792 +       clr_nx_info(&sk->sk_nx_info);
24793 +       sk->sk_nid = -1;
24794         sk_prot_free(sk->sk_prot_creator, sk);
24795  }
24796
24797 @@ -1216,6 +1227,8 @@ struct sock *sk_clone(const struct sock
24798
24799                 /* SANITY */
24800                 get_net(sock_net(newsk));
24801 +               sock_vx_init(newsk);
24802 +               sock_nx_init(newsk);
24803                 sk_node_init(&newsk->sk_node);
24804                 sock_lock_init(newsk);
24805                 bh_lock_sock(newsk);
24806 @@ -1271,6 +1284,12 @@ struct sock *sk_clone(const struct sock
24807                 smp_wmb();
24808                 atomic_set(&newsk->sk_refcnt, 2);
24809
24810 +               set_vx_info(&newsk->sk_vx_info, sk->sk_vx_info);
24811 +               newsk->sk_xid = sk->sk_xid;
24812 +               vx_sock_inc(newsk);
24813 +               set_nx_info(&newsk->sk_nx_info, sk->sk_nx_info);
24814 +               newsk->sk_nid = sk->sk_nid;
24815 +
24816                 /*
24817                  * Increment the counter in the same struct proto as the master
24818                  * sock (sk_refcnt_debug_inc uses newsk->sk_prot->socks, that
24819 @@ -2017,6 +2036,12 @@ void sock_init_data(struct socket *sock,
24820
24821         sk->sk_stamp = ktime_set(-1L, 0);
24822
24823 +       set_vx_info(&sk->sk_vx_info, current_vx_info());
24824 +       sk->sk_xid = vx_current_xid();
24825 +       vx_sock_inc(sk);
24826 +       set_nx_info(&sk->sk_nx_info, current_nx_info());
24827 +       sk->sk_nid = nx_current_nid();
24828 +
24829         /*
24830          * Before updating sk_refcnt, we must commit prior changes to memory
24831          * (Documentation/RCU/rculist_nulls.txt for details)
24832 diff -NurpP --minimal linux-3.0-rc5/net/ipv4/af_inet.c linux-3.0-rc5-vs2.3.1-pre3/net/ipv4/af_inet.c
24833 --- linux-3.0-rc5/net/ipv4/af_inet.c    2011-07-01 11:14:26.000000000 +0200
24834 +++ linux-3.0-rc5-vs2.3.1-pre3/net/ipv4/af_inet.c       2011-06-22 12:39:16.000000000 +0200
24835 @@ -117,6 +117,7 @@
24836  #ifdef CONFIG_IP_MROUTE
24837  #include <linux/mroute.h>
24838  #endif
24839 +#include <linux/vs_limit.h>
24840
24841
24842  /* The inetsw table contains everything that inet_create needs to
24843 @@ -326,9 +327,13 @@ lookup_protocol:
24844         }
24845
24846         err = -EPERM;
24847 +       if ((protocol == IPPROTO_ICMP) &&
24848 +               nx_capable(CAP_NET_RAW, NXC_RAW_ICMP))
24849 +               goto override;
24850 +
24851         if (sock->type == SOCK_RAW && !kern && !capable(CAP_NET_RAW))
24852                 goto out_rcu_unlock;
24853 -
24854 +override:
24855         err = -EAFNOSUPPORT;
24856         if (!inet_netns_ok(net, protocol))
24857                 goto out_rcu_unlock;
24858 @@ -452,6 +457,7 @@ int inet_bind(struct socket *sock, struc
24859         struct sockaddr_in *addr = (struct sockaddr_in *)uaddr;
24860         struct sock *sk = sock->sk;
24861         struct inet_sock *inet = inet_sk(sk);
24862 +       struct nx_v4_sock_addr nsa;
24863         unsigned short snum;
24864         int chk_addr_ret;
24865         int err;
24866 @@ -468,7 +474,11 @@ int inet_bind(struct socket *sock, struc
24867         if (addr->sin_family != AF_INET)
24868                 goto out;
24869
24870 -       chk_addr_ret = inet_addr_type(sock_net(sk), addr->sin_addr.s_addr);
24871 +       err = v4_map_sock_addr(inet, addr, &nsa);
24872 +       if (err)
24873 +               goto out;
24874 +
24875 +       chk_addr_ret = inet_addr_type(sock_net(sk), nsa.saddr);
24876
24877         /* Not specified by any standard per-se, however it breaks too
24878          * many applications when removed.  It is unfortunate since
24879 @@ -480,7 +490,7 @@ int inet_bind(struct socket *sock, struc
24880         err = -EADDRNOTAVAIL;
24881         if (!sysctl_ip_nonlocal_bind &&
24882             !(inet->freebind || inet->transparent) &&
24883 -           addr->sin_addr.s_addr != htonl(INADDR_ANY) &&
24884 +           nsa.saddr != htonl(INADDR_ANY) &&
24885             chk_addr_ret != RTN_LOCAL &&
24886             chk_addr_ret != RTN_MULTICAST &&
24887             chk_addr_ret != RTN_BROADCAST)
24888 @@ -505,7 +515,7 @@ int inet_bind(struct socket *sock, struc
24889         if (sk->sk_state != TCP_CLOSE || inet->inet_num)
24890                 goto out_release_sock;
24891
24892 -       inet->inet_rcv_saddr = inet->inet_saddr = addr->sin_addr.s_addr;
24893 +       v4_set_sock_addr(inet, &nsa);
24894         if (chk_addr_ret == RTN_MULTICAST || chk_addr_ret == RTN_BROADCAST)
24895                 inet->inet_saddr = 0;  /* Use device */
24896
24897 @@ -708,11 +718,13 @@ int inet_getname(struct socket *sock, st
24898                      peer == 1))
24899                         return -ENOTCONN;
24900                 sin->sin_port = inet->inet_dport;
24901 -               sin->sin_addr.s_addr = inet->inet_daddr;
24902 +               sin->sin_addr.s_addr =
24903 +                       nx_map_sock_lback(sk->sk_nx_info, inet->inet_daddr);
24904         } else {
24905                 __be32 addr = inet->inet_rcv_saddr;
24906                 if (!addr)
24907                         addr = inet->inet_saddr;
24908 +               addr = nx_map_sock_lback(sk->sk_nx_info, addr);
24909                 sin->sin_port = inet->inet_sport;
24910                 sin->sin_addr.s_addr = addr;
24911         }
24912 diff -NurpP --minimal linux-3.0-rc5/net/ipv4/devinet.c linux-3.0-rc5-vs2.3.1-pre3/net/ipv4/devinet.c
24913 --- linux-3.0-rc5/net/ipv4/devinet.c    2011-07-01 11:14:26.000000000 +0200
24914 +++ linux-3.0-rc5-vs2.3.1-pre3/net/ipv4/devinet.c       2011-06-10 22:11:24.000000000 +0200
24915 @@ -518,6 +518,7 @@ struct in_device *inetdev_by_index(struc
24916  }
24917  EXPORT_SYMBOL(inetdev_by_index);
24918
24919 +
24920  /* Called only from RTNL semaphored context. No locks. */
24921
24922  struct in_ifaddr *inet_ifa_byprefix(struct in_device *in_dev, __be32 prefix,
24923 @@ -759,6 +760,8 @@ int devinet_ioctl(struct net *net, unsig
24924
24925         in_dev = __in_dev_get_rtnl(dev);
24926         if (in_dev) {
24927 +               struct nx_info *nxi = current_nx_info();
24928 +
24929                 if (tryaddrmatch) {
24930                         /* Matthias Andree */
24931                         /* compare label and address (4.4BSD style) */
24932 @@ -767,6 +770,8 @@ int devinet_ioctl(struct net *net, unsig
24933                            This is checked above. */
24934                         for (ifap = &in_dev->ifa_list; (ifa = *ifap) != NULL;
24935                              ifap = &ifa->ifa_next) {
24936 +                               if (!nx_v4_ifa_visible(nxi, ifa))
24937 +                                       continue;
24938                                 if (!strcmp(ifr.ifr_name, ifa->ifa_label) &&
24939                                     sin_orig.sin_addr.s_addr ==
24940                                                         ifa->ifa_local) {
24941 @@ -779,9 +784,12 @@ int devinet_ioctl(struct net *net, unsig
24942                    comparing just the label */
24943                 if (!ifa) {
24944                         for (ifap = &in_dev->ifa_list; (ifa = *ifap) != NULL;
24945 -                            ifap = &ifa->ifa_next)
24946 +                            ifap = &ifa->ifa_next) {
24947 +                               if (!nx_v4_ifa_visible(nxi, ifa))
24948 +                                       continue;
24949                                 if (!strcmp(ifr.ifr_name, ifa->ifa_label))
24950                                         break;
24951 +                       }
24952                 }
24953         }
24954
24955 @@ -934,6 +942,8 @@ static int inet_gifconf(struct net_devic
24956                 goto out;
24957
24958         for (ifa = in_dev->ifa_list; ifa; ifa = ifa->ifa_next) {
24959 +               if (!nx_v4_ifa_visible(current_nx_info(), ifa))
24960 +                       continue;
24961                 if (!buf) {
24962                         done += sizeof(ifr);
24963                         continue;
24964 @@ -1294,6 +1304,7 @@ static int inet_dump_ifaddr(struct sk_bu
24965         struct net_device *dev;
24966         struct in_device *in_dev;
24967         struct in_ifaddr *ifa;
24968 +       struct sock *sk = skb->sk;
24969         struct hlist_head *head;
24970         struct hlist_node *node;
24971
24972 @@ -1316,6 +1327,8 @@ static int inet_dump_ifaddr(struct sk_bu
24973
24974                         for (ifa = in_dev->ifa_list, ip_idx = 0; ifa;
24975                              ifa = ifa->ifa_next, ip_idx++) {
24976 +                       if (sk && !nx_v4_ifa_visible(sk->sk_nx_info, ifa))
24977 +                               continue;
24978                                 if (ip_idx < s_ip_idx)
24979                                         continue;
24980                                 if (inet_fill_ifaddr(skb, ifa,
24981 diff -NurpP --minimal linux-3.0-rc5/net/ipv4/inet_connection_sock.c linux-3.0-rc5-vs2.3.1-pre3/net/ipv4/inet_connection_sock.c
24982 --- linux-3.0-rc5/net/ipv4/inet_connection_sock.c       2011-07-01 11:14:26.000000000 +0200
24983 +++ linux-3.0-rc5-vs2.3.1-pre3/net/ipv4/inet_connection_sock.c  2011-06-10 22:11:24.000000000 +0200
24984 @@ -52,6 +52,37 @@ void inet_get_local_port_range(int *low,
24985  }
24986  EXPORT_SYMBOL(inet_get_local_port_range);
24987
24988 +int ipv4_rcv_saddr_equal(const struct sock *sk1, const struct sock *sk2)
24989 +{
24990 +       __be32  sk1_rcv_saddr = sk_rcv_saddr(sk1),
24991 +               sk2_rcv_saddr = sk_rcv_saddr(sk2);
24992 +
24993 +       if (inet_v6_ipv6only(sk2))
24994 +               return 0;
24995 +
24996 +       if (sk1_rcv_saddr &&
24997 +           sk2_rcv_saddr &&
24998 +           sk1_rcv_saddr == sk2_rcv_saddr)
24999 +               return 1;
25000 +
25001 +       if (sk1_rcv_saddr &&
25002 +           !sk2_rcv_saddr &&
25003 +           v4_addr_in_nx_info(sk2->sk_nx_info, sk1_rcv_saddr, NXA_MASK_BIND))
25004 +               return 1;
25005 +
25006 +       if (sk2_rcv_saddr &&
25007 +           !sk1_rcv_saddr &&
25008 +           v4_addr_in_nx_info(sk1->sk_nx_info, sk2_rcv_saddr, NXA_MASK_BIND))
25009 +               return 1;
25010 +
25011 +       if (!sk1_rcv_saddr &&
25012 +           !sk2_rcv_saddr &&
25013 +           nx_v4_addr_conflict(sk1->sk_nx_info, sk2->sk_nx_info))
25014 +               return 1;
25015 +
25016 +       return 0;
25017 +}
25018 +
25019  int inet_csk_bind_conflict(const struct sock *sk,
25020                            const struct inet_bind_bucket *tb)
25021  {
25022 @@ -74,9 +105,7 @@ int inet_csk_bind_conflict(const struct
25023                      sk->sk_bound_dev_if == sk2->sk_bound_dev_if)) {
25024                         if (!reuse || !sk2->sk_reuse ||
25025                             sk2->sk_state == TCP_LISTEN) {
25026 -                               const __be32 sk2_rcv_saddr = sk_rcv_saddr(sk2);
25027 -                               if (!sk2_rcv_saddr || !sk_rcv_saddr(sk) ||
25028 -                                   sk2_rcv_saddr == sk_rcv_saddr(sk))
25029 +                               if (ipv4_rcv_saddr_equal(sk, sk2))
25030                                         break;
25031                         }
25032                 }
25033 diff -NurpP --minimal linux-3.0-rc5/net/ipv4/inet_diag.c linux-3.0-rc5-vs2.3.1-pre3/net/ipv4/inet_diag.c
25034 --- linux-3.0-rc5/net/ipv4/inet_diag.c  2011-07-01 11:14:26.000000000 +0200
25035 +++ linux-3.0-rc5-vs2.3.1-pre3/net/ipv4/inet_diag.c     2011-06-22 12:39:16.000000000 +0200
25036 @@ -33,6 +33,8 @@
25037  #include <linux/stddef.h>
25038
25039  #include <linux/inet_diag.h>
25040 +#include <linux/vs_network.h>
25041 +#include <linux/vs_inet.h>
25042
25043  static const struct inet_diag_handler **inet_diag_table;
25044
25045 @@ -119,8 +121,10 @@ static int inet_csk_diag_fill(struct soc
25046
25047         r->id.idiag_sport = inet->inet_sport;
25048         r->id.idiag_dport = inet->inet_dport;
25049 -       r->id.idiag_src[0] = inet->inet_rcv_saddr;
25050 -       r->id.idiag_dst[0] = inet->inet_daddr;
25051 +       r->id.idiag_src[0] = nx_map_sock_lback(sk->sk_nx_info,
25052 +               inet->inet_rcv_saddr);
25053 +       r->id.idiag_dst[0] = nx_map_sock_lback(sk->sk_nx_info,
25054 +               inet->inet_daddr);
25055
25056  #if defined(CONFIG_IPV6) || defined (CONFIG_IPV6_MODULE)
25057         if (r->idiag_family == AF_INET6) {
25058 @@ -205,8 +209,8 @@ static int inet_twsk_diag_fill(struct in
25059         r->id.idiag_cookie[1] = (u32)(((unsigned long)tw >> 31) >> 1);
25060         r->id.idiag_sport     = tw->tw_sport;
25061         r->id.idiag_dport     = tw->tw_dport;
25062 -       r->id.idiag_src[0]    = tw->tw_rcv_saddr;
25063 -       r->id.idiag_dst[0]    = tw->tw_daddr;
25064 +       r->id.idiag_src[0]    = nx_map_sock_lback(tw->tw_nx_info, tw->tw_rcv_saddr);
25065 +       r->id.idiag_dst[0]    = nx_map_sock_lback(tw->tw_nx_info, tw->tw_daddr);
25066         r->idiag_state        = tw->tw_substate;
25067         r->idiag_timer        = 3;
25068         r->idiag_expires      = DIV_ROUND_UP(tmo * 1000, HZ);
25069 @@ -263,6 +267,7 @@ static int inet_diag_get_exact(struct sk
25070         err = -EINVAL;
25071
25072         if (req->idiag_family == AF_INET) {
25073 +               /* TODO: lback */
25074                 sk = inet_lookup(&init_net, hashinfo, req->id.idiag_dst[0],
25075                                  req->id.idiag_dport, req->id.idiag_src[0],
25076                                  req->id.idiag_sport, req->id.idiag_if);
25077 @@ -505,6 +510,7 @@ static int inet_csk_diag_dump(struct soc
25078                 } else
25079  #endif
25080                 {
25081 +                       /* TODO: lback */
25082                         entry.saddr = &inet->inet_rcv_saddr;
25083                         entry.daddr = &inet->inet_daddr;
25084                 }
25085 @@ -543,6 +549,7 @@ static int inet_twsk_diag_dump(struct in
25086                 } else
25087  #endif
25088                 {
25089 +                       /* TODO: lback */
25090                         entry.saddr = &tw->tw_rcv_saddr;
25091                         entry.daddr = &tw->tw_daddr;
25092                 }
25093 @@ -589,8 +596,8 @@ static int inet_diag_fill_req(struct sk_
25094
25095         r->id.idiag_sport = inet->inet_sport;
25096         r->id.idiag_dport = ireq->rmt_port;
25097 -       r->id.idiag_src[0] = ireq->loc_addr;
25098 -       r->id.idiag_dst[0] = ireq->rmt_addr;
25099 +       r->id.idiag_src[0] = nx_map_sock_lback(sk->sk_nx_info, ireq->loc_addr);
25100 +       r->id.idiag_dst[0] = nx_map_sock_lback(sk->sk_nx_info, ireq->rmt_addr);
25101         r->idiag_expires = jiffies_to_msecs(tmo);
25102         r->idiag_rqueue = 0;
25103         r->idiag_wqueue = 0;
25104 @@ -661,6 +668,7 @@ static int inet_diag_dump_reqs(struct sk
25105                                 continue;
25106
25107                         if (bc) {
25108 +                               /* TODO: lback */
25109                                 entry.saddr =
25110  #if defined(CONFIG_IPV6) || defined (CONFIG_IPV6_MODULE)
25111                                         (entry.family == AF_INET6) ?
25112 @@ -731,6 +739,8 @@ static int inet_diag_dump(struct sk_buff
25113                         sk_nulls_for_each(sk, node, &ilb->head) {
25114                                 struct inet_sock *inet = inet_sk(sk);
25115
25116 +                               if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
25117 +                                       continue;
25118                                 if (num < s_num) {
25119                                         num++;
25120                                         continue;
25121 @@ -797,6 +807,8 @@ skip_listen_ht:
25122                 sk_nulls_for_each(sk, node, &head->chain) {
25123                         struct inet_sock *inet = inet_sk(sk);
25124
25125 +                       if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
25126 +                               continue;
25127                         if (num < s_num)
25128                                 goto next_normal;
25129                         if (!(r->idiag_states & (1 << sk->sk_state)))
25130 @@ -821,6 +833,8 @@ next_normal:
25131                         inet_twsk_for_each(tw, node,
25132                                     &head->twchain) {
25133
25134 +                               if (!nx_check(tw->tw_nid, VS_WATCH_P | VS_IDENT))
25135 +                                       continue;
25136                                 if (num < s_num)
25137                                         goto next_dying;
25138                                 if (r->id.idiag_sport != tw->tw_sport &&
25139 diff -NurpP --minimal linux-3.0-rc5/net/ipv4/inet_hashtables.c linux-3.0-rc5-vs2.3.1-pre3/net/ipv4/inet_hashtables.c
25140 --- linux-3.0-rc5/net/ipv4/inet_hashtables.c    2011-01-05 21:50:42.000000000 +0100
25141 +++ linux-3.0-rc5-vs2.3.1-pre3/net/ipv4/inet_hashtables.c       2011-06-10 22:11:24.000000000 +0200
25142 @@ -21,6 +21,7 @@
25143
25144  #include <net/inet_connection_sock.h>
25145  #include <net/inet_hashtables.h>
25146 +#include <net/route.h>
25147  #include <net/ip.h>
25148
25149  /*
25150 @@ -155,6 +156,11 @@ static inline int compute_score(struct s
25151                         if (rcv_saddr != daddr)
25152                                 return -1;
25153                         score += 2;
25154 +               } else {
25155 +                       /* block non nx_info ips */
25156 +                       if (!v4_addr_in_nx_info(sk->sk_nx_info,
25157 +                               daddr, NXA_MASK_BIND))
25158 +                               return -1;
25159                 }
25160                 if (sk->sk_bound_dev_if) {
25161                         if (sk->sk_bound_dev_if != dif)
25162 @@ -172,7 +178,6 @@ static inline int compute_score(struct s
25163   * wildcarded during the search since they can never be otherwise.
25164   */
25165
25166 -
25167  struct sock *__inet_lookup_listener(struct net *net,
25168                                     struct inet_hashinfo *hashinfo,
25169                                     const __be32 daddr, const unsigned short hnum,
25170 @@ -195,6 +200,7 @@ begin:
25171                         hiscore = score;
25172                 }
25173         }
25174 +
25175         /*
25176          * if the nulls value we got at the end of this lookup is
25177          * not the expected one, we must restart lookup.
25178 diff -NurpP --minimal linux-3.0-rc5/net/ipv4/netfilter/nf_nat_helper.c linux-3.0-rc5-vs2.3.1-pre3/net/ipv4/netfilter/nf_nat_helper.c
25179 --- linux-3.0-rc5/net/ipv4/netfilter/nf_nat_helper.c    2011-07-01 11:14:26.000000000 +0200
25180 +++ linux-3.0-rc5-vs2.3.1-pre3/net/ipv4/netfilter/nf_nat_helper.c       2011-06-15 02:40:14.000000000 +0200
25181 @@ -20,6 +20,7 @@
25182  #include <net/route.h>
25183
25184  #include <linux/netfilter_ipv4.h>
25185 +#include <net/route.h>
25186  #include <net/netfilter/nf_conntrack.h>
25187  #include <net/netfilter/nf_conntrack_helper.h>
25188  #include <net/netfilter/nf_conntrack_ecache.h>
25189 diff -NurpP --minimal linux-3.0-rc5/net/ipv4/netfilter.c linux-3.0-rc5-vs2.3.1-pre3/net/ipv4/netfilter.c
25190 --- linux-3.0-rc5/net/ipv4/netfilter.c  2011-05-22 16:18:00.000000000 +0200
25191 +++ linux-3.0-rc5-vs2.3.1-pre3/net/ipv4/netfilter.c     2011-06-10 22:11:24.000000000 +0200
25192 @@ -5,7 +5,7 @@
25193  #include <linux/ip.h>
25194  #include <linux/skbuff.h>
25195  #include <linux/gfp.h>
25196 -#include <net/route.h>
25197 +// #include <net/route.h>
25198  #include <net/xfrm.h>
25199  #include <net/ip.h>
25200  #include <net/netfilter/nf_queue.h>
25201 diff -NurpP --minimal linux-3.0-rc5/net/ipv4/raw.c linux-3.0-rc5-vs2.3.1-pre3/net/ipv4/raw.c
25202 --- linux-3.0-rc5/net/ipv4/raw.c        2011-07-01 11:14:26.000000000 +0200
25203 +++ linux-3.0-rc5-vs2.3.1-pre3/net/ipv4/raw.c   2011-06-10 22:11:24.000000000 +0200
25204 @@ -117,7 +117,7 @@ static struct sock *__raw_v4_lookup(stru
25205
25206                 if (net_eq(sock_net(sk), net) && inet->inet_num == num  &&
25207                     !(inet->inet_daddr && inet->inet_daddr != raddr)    &&
25208 -                   !(inet->inet_rcv_saddr && inet->inet_rcv_saddr != laddr) &&
25209 +                   v4_sock_addr_match(sk->sk_nx_info, inet, laddr)     &&
25210                     !(sk->sk_bound_dev_if && sk->sk_bound_dev_if != dif))
25211                         goto found; /* gotcha */
25212         }
25213 @@ -385,6 +385,12 @@ static int raw_send_hdrinc(struct sock *
25214                 icmp_out_count(net, ((struct icmphdr *)
25215                         skb_transport_header(skb))->type);
25216
25217 +       err = -EPERM;
25218 +       if (!nx_check(0, VS_ADMIN) && !capable(CAP_NET_RAW) &&
25219 +               sk->sk_nx_info &&
25220 +               !v4_addr_in_nx_info(sk->sk_nx_info, iph->saddr, NXA_MASK_BIND))
25221 +               goto error_free;
25222 +
25223         err = NF_HOOK(NFPROTO_IPV4, NF_INET_LOCAL_OUT, skb, NULL,
25224                       rt->dst.dev, dst_output);
25225         if (err > 0)
25226 @@ -647,17 +653,19 @@ static int raw_bind(struct sock *sk, str
25227  {
25228         struct inet_sock *inet = inet_sk(sk);
25229         struct sockaddr_in *addr = (struct sockaddr_in *) uaddr;
25230 +       struct nx_v4_sock_addr nsa = { 0 };
25231         int ret = -EINVAL;
25232         int chk_addr_ret;
25233
25234         if (sk->sk_state != TCP_CLOSE || addr_len < sizeof(struct sockaddr_in))
25235                 goto out;
25236 -       chk_addr_ret = inet_addr_type(sock_net(sk), addr->sin_addr.s_addr);
25237 +       v4_map_sock_addr(inet, addr, &nsa);
25238 +       chk_addr_ret = inet_addr_type(sock_net(sk), nsa.saddr);
25239         ret = -EADDRNOTAVAIL;
25240 -       if (addr->sin_addr.s_addr && chk_addr_ret != RTN_LOCAL &&
25241 +       if (nsa.saddr && chk_addr_ret != RTN_LOCAL &&
25242             chk_addr_ret != RTN_MULTICAST && chk_addr_ret != RTN_BROADCAST)
25243                 goto out;
25244 -       inet->inet_rcv_saddr = inet->inet_saddr = addr->sin_addr.s_addr;
25245 +       v4_set_sock_addr(inet, &nsa);
25246         if (chk_addr_ret == RTN_MULTICAST || chk_addr_ret == RTN_BROADCAST)
25247                 inet->inet_saddr = 0;  /* Use device */
25248         sk_dst_reset(sk);
25249 @@ -709,7 +717,8 @@ static int raw_recvmsg(struct kiocb *ioc
25250         /* Copy the address. */
25251         if (sin) {
25252                 sin->sin_family = AF_INET;
25253 -               sin->sin_addr.s_addr = ip_hdr(skb)->saddr;
25254 +               sin->sin_addr.s_addr =
25255 +                       nx_map_sock_lback(sk->sk_nx_info, ip_hdr(skb)->saddr);
25256                 sin->sin_port = 0;
25257                 memset(&sin->sin_zero, 0, sizeof(sin->sin_zero));
25258         }
25259 @@ -905,7 +914,8 @@ static struct sock *raw_get_first(struct
25260                 struct hlist_node *node;
25261
25262                 sk_for_each(sk, node, &state->h->ht[state->bucket])
25263 -                       if (sock_net(sk) == seq_file_net(seq))
25264 +                       if ((sock_net(sk) == seq_file_net(seq)) &&
25265 +                               nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
25266                                 goto found;
25267         }
25268         sk = NULL;
25269 @@ -921,7 +931,8 @@ static struct sock *raw_get_next(struct
25270                 sk = sk_next(sk);
25271  try_again:
25272                 ;
25273 -       } while (sk && sock_net(sk) != seq_file_net(seq));
25274 +       } while (sk && ((sock_net(sk) != seq_file_net(seq)) ||
25275 +               !nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT)));
25276
25277         if (!sk && ++state->bucket < RAW_HTABLE_SIZE) {
25278                 sk = sk_head(&state->h->ht[state->bucket]);
25279 diff -NurpP --minimal linux-3.0-rc5/net/ipv4/tcp.c linux-3.0-rc5-vs2.3.1-pre3/net/ipv4/tcp.c
25280 --- linux-3.0-rc5/net/ipv4/tcp.c        2011-07-01 11:14:26.000000000 +0200
25281 +++ linux-3.0-rc5-vs2.3.1-pre3/net/ipv4/tcp.c   2011-06-10 22:11:24.000000000 +0200
25282 @@ -266,6 +266,7 @@
25283  #include <linux/crypto.h>
25284  #include <linux/time.h>
25285  #include <linux/slab.h>
25286 +#include <linux/in.h>
25287
25288  #include <net/icmp.h>
25289  #include <net/tcp.h>
25290 diff -NurpP --minimal linux-3.0-rc5/net/ipv4/tcp_ipv4.c linux-3.0-rc5-vs2.3.1-pre3/net/ipv4/tcp_ipv4.c
25291 --- linux-3.0-rc5/net/ipv4/tcp_ipv4.c   2011-07-01 11:14:26.000000000 +0200
25292 +++ linux-3.0-rc5-vs2.3.1-pre3/net/ipv4/tcp_ipv4.c      2011-06-22 12:39:16.000000000 +0200
25293 @@ -2002,6 +2002,12 @@ static void *listening_get_next(struct s
25294                 req = req->dl_next;
25295                 while (1) {
25296                         while (req) {
25297 +                               vxdprintk(VXD_CBIT(net, 6),
25298 +                                       "sk,req: %p [#%d] (from %d)", req->sk,
25299 +                                       (req->sk)?req->sk->sk_nid:0, nx_current_nid());
25300 +                               if (req->sk &&
25301 +                                       !nx_check(req->sk->sk_nid, VS_WATCH_P | VS_IDENT))
25302 +                                       continue;
25303                                 if (req->rsk_ops->family == st->family) {
25304                                         cur = req;
25305                                         goto out;
25306 @@ -2026,6 +2032,10 @@ get_req:
25307         }
25308  get_sk:
25309         sk_nulls_for_each_from(sk, node) {
25310 +               vxdprintk(VXD_CBIT(net, 6), "sk: %p [#%d] (from %d)",
25311 +                       sk, sk->sk_nid, nx_current_nid());
25312 +               if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
25313 +                       continue;
25314                 if (!net_eq(sock_net(sk), net))
25315                         continue;
25316                 if (sk->sk_family == st->family) {
25317 @@ -2102,6 +2112,11 @@ static void *established_get_first(struc
25318
25319                 spin_lock_bh(lock);
25320                 sk_nulls_for_each(sk, node, &tcp_hashinfo.ehash[st->bucket].chain) {
25321 +                       vxdprintk(VXD_CBIT(net, 6),
25322 +                               "sk,egf: %p [#%d] (from %d)",
25323 +                               sk, sk->sk_nid, nx_current_nid());
25324 +                       if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
25325 +                               continue;
25326                         if (sk->sk_family != st->family ||
25327                             !net_eq(sock_net(sk), net)) {
25328                                 continue;
25329 @@ -2112,6 +2127,11 @@ static void *established_get_first(struc
25330                 st->state = TCP_SEQ_STATE_TIME_WAIT;
25331                 inet_twsk_for_each(tw, node,
25332                                    &tcp_hashinfo.ehash[st->bucket].twchain) {
25333 +                       vxdprintk(VXD_CBIT(net, 6),
25334 +                               "tw: %p [#%d] (from %d)",
25335 +                               tw, tw->tw_nid, nx_current_nid());
25336 +                       if (!nx_check(tw->tw_nid, VS_WATCH_P | VS_IDENT))
25337 +                               continue;
25338                         if (tw->tw_family != st->family ||
25339                             !net_eq(twsk_net(tw), net)) {
25340                                 continue;
25341 @@ -2141,7 +2161,9 @@ static void *established_get_next(struct
25342                 tw = cur;
25343                 tw = tw_next(tw);
25344  get_tw:
25345 -               while (tw && (tw->tw_family != st->family || !net_eq(twsk_net(tw), net))) {
25346 +               while (tw && (tw->tw_family != st->family ||
25347 +                       !net_eq(twsk_net(tw), net) ||
25348 +                       !nx_check(tw->tw_nid, VS_WATCH_P | VS_IDENT))) {
25349                         tw = tw_next(tw);
25350                 }
25351                 if (tw) {
25352 @@ -2165,6 +2187,11 @@ get_tw:
25353                 sk = sk_nulls_next(sk);
25354
25355         sk_nulls_for_each_from(sk, node) {
25356 +               vxdprintk(VXD_CBIT(net, 6),
25357 +                       "sk,egn: %p [#%d] (from %d)",
25358 +                       sk, sk->sk_nid, nx_current_nid());
25359 +               if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
25360 +                       continue;
25361                 if (sk->sk_family == st->family && net_eq(sock_net(sk), net))
25362                         goto found;
25363         }
25364 @@ -2374,9 +2401,9 @@ static void get_openreq4(struct sock *sk
25365         seq_printf(f, "%4d: %08X:%04X %08X:%04X"
25366                 " %02X %08X:%08X %02X:%08lX %08X %5d %8d %u %d %pK%n",
25367                 i,
25368 -               ireq->loc_addr,
25369 +               nx_map_sock_lback(current_nx_info(), ireq->loc_addr),
25370                 ntohs(inet_sk(sk)->inet_sport),
25371 -               ireq->rmt_addr,
25372 +               nx_map_sock_lback(current_nx_info(), ireq->rmt_addr),
25373                 ntohs(ireq->rmt_port),
25374                 TCP_SYN_RECV,
25375                 0, 0, /* could print option size, but that is af dependent. */
25376 diff -NurpP --minimal linux-3.0-rc5/net/ipv4/tcp_minisocks.c linux-3.0-rc5-vs2.3.1-pre3/net/ipv4/tcp_minisocks.c
25377 --- linux-3.0-rc5/net/ipv4/tcp_minisocks.c      2011-03-15 18:07:45.000000000 +0100
25378 +++ linux-3.0-rc5-vs2.3.1-pre3/net/ipv4/tcp_minisocks.c 2011-06-10 22:11:24.000000000 +0200
25379 @@ -23,6 +23,9 @@
25380  #include <linux/slab.h>
25381  #include <linux/sysctl.h>
25382  #include <linux/workqueue.h>
25383 +#include <linux/vs_limit.h>
25384 +#include <linux/vs_socket.h>
25385 +#include <linux/vs_context.h>
25386  #include <net/tcp.h>
25387  #include <net/inet_common.h>
25388  #include <net/xfrm.h>
25389 @@ -335,6 +338,11 @@ void tcp_time_wait(struct sock *sk, int
25390                 tcptw->tw_ts_recent     = tp->rx_opt.ts_recent;
25391                 tcptw->tw_ts_recent_stamp = tp->rx_opt.ts_recent_stamp;
25392
25393 +               tw->tw_xid              = sk->sk_xid;
25394 +               tw->tw_vx_info          = NULL;
25395 +               tw->tw_nid              = sk->sk_nid;
25396 +               tw->tw_nx_info          = NULL;
25397 +
25398  #if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
25399                 if (tw->tw_family == PF_INET6) {
25400                         struct ipv6_pinfo *np = inet6_sk(sk);
25401 diff -NurpP --minimal linux-3.0-rc5/net/ipv4/udp.c linux-3.0-rc5-vs2.3.1-pre3/net/ipv4/udp.c
25402 --- linux-3.0-rc5/net/ipv4/udp.c        2011-07-01 11:14:26.000000000 +0200
25403 +++ linux-3.0-rc5-vs2.3.1-pre3/net/ipv4/udp.c   2011-06-10 22:11:24.000000000 +0200
25404 @@ -296,14 +296,7 @@ fail:
25405  }
25406  EXPORT_SYMBOL(udp_lib_get_port);
25407
25408 -static int ipv4_rcv_saddr_equal(const struct sock *sk1, const struct sock *sk2)
25409 -{
25410 -       struct inet_sock *inet1 = inet_sk(sk1), *inet2 = inet_sk(sk2);
25411 -
25412 -       return  (!ipv6_only_sock(sk2)  &&
25413 -                (!inet1->inet_rcv_saddr || !inet2->inet_rcv_saddr ||
25414 -                  inet1->inet_rcv_saddr == inet2->inet_rcv_saddr));
25415 -}
25416 +extern int ipv4_rcv_saddr_equal(const struct sock *, const struct sock *);
25417
25418  static unsigned int udp4_portaddr_hash(struct net *net, __be32 saddr,
25419                                        unsigned int port)
25420 @@ -338,6 +331,11 @@ static inline int compute_score(struct s
25421                         if (inet->inet_rcv_saddr != daddr)
25422                                 return -1;
25423                         score += 2;
25424 +               } else {
25425 +                       /* block non nx_info ips */
25426 +                       if (!v4_addr_in_nx_info(sk->sk_nx_info,
25427 +                               daddr, NXA_MASK_BIND))
25428 +                               return -1;
25429                 }
25430                 if (inet->inet_daddr) {
25431                         if (inet->inet_daddr != saddr)
25432 @@ -441,6 +439,7 @@ exact_match:
25433         return result;
25434  }
25435
25436 +
25437  /* UDP is nearly always wildcards out the wazoo, it makes no sense to try
25438   * harder than this. -DaveM
25439   */
25440 @@ -486,6 +485,11 @@ begin:
25441         sk_nulls_for_each_rcu(sk, node, &hslot->head) {
25442                 score = compute_score(sk, net, saddr, hnum, sport,
25443                                       daddr, dport, dif);
25444 +               /* FIXME: disabled?
25445 +               if (score == 9) {
25446 +                       result = sk;
25447 +                       break;
25448 +               } else */
25449                 if (score > badness) {
25450                         result = sk;
25451                         badness = score;
25452 @@ -499,6 +503,7 @@ begin:
25453         if (get_nulls_value(node) != slot)
25454                 goto begin;
25455
25456 +
25457         if (result) {
25458                 if (unlikely(!atomic_inc_not_zero_hint(&result->sk_refcnt, 2)))
25459                         result = NULL;
25460 @@ -508,6 +513,7 @@ begin:
25461                         goto begin;
25462                 }
25463         }
25464 +
25465         rcu_read_unlock();
25466         return result;
25467  }
25468 @@ -550,8 +556,7 @@ static inline struct sock *udp_v4_mcast_
25469                     udp_sk(s)->udp_port_hash != hnum ||
25470                     (inet->inet_daddr && inet->inet_daddr != rmt_addr) ||
25471                     (inet->inet_dport != rmt_port && inet->inet_dport) ||
25472 -                   (inet->inet_rcv_saddr &&
25473 -                    inet->inet_rcv_saddr != loc_addr) ||
25474 +                   !v4_sock_addr_match(sk->sk_nx_info, inet, loc_addr) ||
25475                     ipv6_only_sock(s) ||
25476                     (s->sk_bound_dev_if && s->sk_bound_dev_if != dif))
25477                         continue;
25478 @@ -1226,7 +1231,8 @@ try_again:
25479         if (sin) {
25480                 sin->sin_family = AF_INET;
25481                 sin->sin_port = udp_hdr(skb)->source;
25482 -               sin->sin_addr.s_addr = ip_hdr(skb)->saddr;
25483 +               sin->sin_addr.s_addr = nx_map_sock_lback(
25484 +                       skb->sk->sk_nx_info, ip_hdr(skb)->saddr);
25485                 memset(sin->sin_zero, 0, sizeof(sin->sin_zero));
25486         }
25487         if (inet->cmsg_flags)
25488 @@ -1967,6 +1973,8 @@ static struct sock *udp_get_first(struct
25489                 sk_nulls_for_each(sk, node, &hslot->head) {
25490                         if (!net_eq(sock_net(sk), net))
25491                                 continue;
25492 +                       if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
25493 +                               continue;
25494                         if (sk->sk_family == state->family)
25495                                 goto found;
25496                 }
25497 @@ -1984,7 +1992,9 @@ static struct sock *udp_get_next(struct
25498
25499         do {
25500                 sk = sk_nulls_next(sk);
25501 -       } while (sk && (!net_eq(sock_net(sk), net) || sk->sk_family != state->family));
25502 +       } while (sk && (!net_eq(sock_net(sk), net) ||
25503 +               sk->sk_family != state->family ||
25504 +               !nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT)));
25505
25506         if (!sk) {
25507                 if (state->bucket <= state->udp_table->mask)
25508 diff -NurpP --minimal linux-3.0-rc5/net/ipv6/Kconfig linux-3.0-rc5-vs2.3.1-pre3/net/ipv6/Kconfig
25509 --- linux-3.0-rc5/net/ipv6/Kconfig      2010-08-02 16:52:59.000000000 +0200
25510 +++ linux-3.0-rc5-vs2.3.1-pre3/net/ipv6/Kconfig 2011-06-10 22:11:24.000000000 +0200
25511 @@ -4,8 +4,8 @@
25512
25513  #   IPv6 as module will cause a CRASH if you try to unload it
25514  menuconfig IPV6
25515 -       tristate "The IPv6 protocol"
25516 -       default m
25517 +       bool "The IPv6 protocol"
25518 +       default n
25519         ---help---
25520           This is complemental support for the IP version 6.
25521           You will still be able to do traditional IPv4 networking as well.
25522 diff -NurpP --minimal linux-3.0-rc5/net/ipv6/addrconf.c linux-3.0-rc5-vs2.3.1-pre3/net/ipv6/addrconf.c
25523 --- linux-3.0-rc5/net/ipv6/addrconf.c   2011-07-01 11:14:26.000000000 +0200
25524 +++ linux-3.0-rc5-vs2.3.1-pre3/net/ipv6/addrconf.c      2011-06-10 22:11:24.000000000 +0200
25525 @@ -87,6 +87,8 @@
25526
25527  #include <linux/proc_fs.h>
25528  #include <linux/seq_file.h>
25529 +#include <linux/vs_network.h>
25530 +#include <linux/vs_inet6.h>
25531
25532  /* Set to 3 to get tracing... */
25533  #define ACONF_DEBUG 2
25534 @@ -1108,7 +1110,7 @@ out:
25535
25536  int ipv6_dev_get_saddr(struct net *net, struct net_device *dst_dev,
25537                        const struct in6_addr *daddr, unsigned int prefs,
25538 -                      struct in6_addr *saddr)
25539 +                      struct in6_addr *saddr, struct nx_info *nxi)
25540  {
25541         struct ipv6_saddr_score scores[2],
25542                                 *score = &scores[0], *hiscore = &scores[1];
25543 @@ -1180,6 +1182,8 @@ int ipv6_dev_get_saddr(struct net *net,
25544                                                dev->name);
25545                                 continue;
25546                         }
25547 +                       if (!v6_addr_in_nx_info(nxi, &score->ifa->addr, -1))
25548 +                               continue;
25549
25550                         score->rule = -1;
25551                         bitmap_zero(score->scorebits, IPV6_SADDR_RULE_MAX);
25552 @@ -3048,7 +3052,10 @@ static void if6_seq_stop(struct seq_file
25553  static int if6_seq_show(struct seq_file *seq, void *v)
25554  {
25555         struct inet6_ifaddr *ifp = (struct inet6_ifaddr *)v;
25556 -       seq_printf(seq, "%pi6 %02x %02x %02x %02x %8s\n",
25557 +
25558 +       if (nx_check(0, VS_ADMIN|VS_WATCH) ||
25559 +           v6_addr_in_nx_info(current_nx_info(), &ifp->addr, -1))
25560 +               seq_printf(seq, "%pi6 %02x %02x %02x %02x %8s\n",
25561                    &ifp->addr,
25562                    ifp->idev->dev->ifindex,
25563                    ifp->prefix_len,
25564 @@ -3554,6 +3561,11 @@ static int in6_dump_addrs(struct inet6_d
25565         struct ifacaddr6 *ifaca;
25566         int err = 1;
25567         int ip_idx = *p_ip_idx;
25568 +       struct nx_info *nxi = skb->sk ? skb->sk->sk_nx_info : NULL;
25569 +
25570 +       /* disable ipv6 on non v6 guests */
25571 +       if (nxi && !nx_info_has_v6(nxi))
25572 +               return skb->len;
25573
25574         read_lock_bh(&idev->lock);
25575         switch (type) {
25576 @@ -3564,6 +3576,8 @@ static int in6_dump_addrs(struct inet6_d
25577                 list_for_each_entry(ifa, &idev->addr_list, if_list) {
25578                         if (++ip_idx < s_ip_idx)
25579                                 continue;
25580 +                               if (!v6_addr_in_nx_info(nxi, &ifa->addr, -1))
25581 +                                       continue;
25582                         err = inet6_fill_ifaddr(skb, ifa,
25583                                                 NETLINK_CB(cb->skb).pid,
25584                                                 cb->nlh->nlmsg_seq,
25585 @@ -3580,6 +3594,8 @@ static int in6_dump_addrs(struct inet6_d
25586                      ifmca = ifmca->next, ip_idx++) {
25587                         if (ip_idx < s_ip_idx)
25588                                 continue;
25589 +                               if (!v6_addr_in_nx_info(nxi, &ifmca->mca_addr, -1))
25590 +                                       continue;
25591                         err = inet6_fill_ifmcaddr(skb, ifmca,
25592                                                   NETLINK_CB(cb->skb).pid,
25593                                                   cb->nlh->nlmsg_seq,
25594 @@ -3595,6 +3611,8 @@ static int in6_dump_addrs(struct inet6_d
25595                      ifaca = ifaca->aca_next, ip_idx++) {
25596                         if (ip_idx < s_ip_idx)
25597                                 continue;
25598 +                               if (!v6_addr_in_nx_info(nxi, &ifaca->aca_addr, -1))
25599 +                                       continue;
25600                         err = inet6_fill_ifacaddr(skb, ifaca,
25601                                                   NETLINK_CB(cb->skb).pid,
25602                                                   cb->nlh->nlmsg_seq,
25603 @@ -3980,6 +3998,11 @@ static int inet6_dump_ifinfo(struct sk_b
25604         struct inet6_dev *idev;
25605         struct hlist_head *head;
25606         struct hlist_node *node;
25607 +       struct nx_info *nxi = skb->sk ? skb->sk->sk_nx_info : NULL;
25608 +
25609 +       /* FIXME: maybe disable ipv6 on non v6 guests?
25610 +       if (skb->sk && skb->sk->sk_vx_info)
25611 +               return skb->len; */
25612
25613         s_h = cb->args[0];
25614         s_idx = cb->args[1];
25615 @@ -3991,6 +4014,8 @@ static int inet6_dump_ifinfo(struct sk_b
25616                 hlist_for_each_entry_rcu(dev, node, head, index_hlist) {
25617                         if (idx < s_idx)
25618                                 goto cont;
25619 +                       if (!v6_dev_in_nx_info(dev, nxi))
25620 +                               goto cont;
25621                         idev = __in6_dev_get(dev);
25622                         if (!idev)
25623                                 goto cont;
25624 diff -NurpP --minimal linux-3.0-rc5/net/ipv6/af_inet6.c linux-3.0-rc5-vs2.3.1-pre3/net/ipv6/af_inet6.c
25625 --- linux-3.0-rc5/net/ipv6/af_inet6.c   2011-07-01 11:14:26.000000000 +0200
25626 +++ linux-3.0-rc5-vs2.3.1-pre3/net/ipv6/af_inet6.c      2011-06-15 02:44:10.000000000 +0200
25627 @@ -42,6 +42,8 @@
25628  #include <linux/netdevice.h>
25629  #include <linux/icmpv6.h>
25630  #include <linux/netfilter_ipv6.h>
25631 +#include <linux/vs_inet.h>
25632 +#include <linux/vs_inet6.h>
25633
25634  #include <net/ip.h>
25635  #include <net/ipv6.h>
25636 @@ -160,9 +162,12 @@ lookup_protocol:
25637         }
25638
25639         err = -EPERM;
25640 +       if ((protocol == IPPROTO_ICMPV6) &&
25641 +               nx_capable(CAP_NET_RAW, NXC_RAW_ICMP))
25642 +               goto override;
25643         if (sock->type == SOCK_RAW && !kern && !capable(CAP_NET_RAW))
25644                 goto out_rcu_unlock;
25645 -
25646 +override:
25647         sock->ops = answer->ops;
25648         answer_prot = answer->prot;
25649         answer_no_check = answer->no_check;
25650 @@ -261,6 +266,7 @@ int inet6_bind(struct socket *sock, stru
25651         struct inet_sock *inet = inet_sk(sk);
25652         struct ipv6_pinfo *np = inet6_sk(sk);
25653         struct net *net = sock_net(sk);
25654 +       struct nx_v6_sock_addr nsa;
25655         __be32 v4addr = 0;
25656         unsigned short snum;
25657         int addr_type = 0;
25658 @@ -276,6 +282,10 @@ int inet6_bind(struct socket *sock, stru
25659         if (addr->sin6_family != AF_INET6)
25660                 return -EINVAL;
25661
25662 +       err = v6_map_sock_addr(inet, addr, &nsa);
25663 +       if (err)
25664 +               return err;
25665 +
25666         addr_type = ipv6_addr_type(&addr->sin6_addr);
25667         if ((addr_type & IPV6_ADDR_MULTICAST) && sock->type == SOCK_STREAM)
25668                 return -EINVAL;
25669 @@ -307,6 +317,7 @@ int inet6_bind(struct socket *sock, stru
25670                 /* Reproduce AF_INET checks to make the bindings consistent */
25671                 v4addr = addr->sin6_addr.s6_addr32[3];
25672                 chk_addr_ret = inet_addr_type(net, v4addr);
25673 +
25674                 if (!sysctl_ip_nonlocal_bind &&
25675                     !(inet->freebind || inet->transparent) &&
25676                     v4addr != htonl(INADDR_ANY) &&
25677 @@ -316,6 +327,10 @@ int inet6_bind(struct socket *sock, stru
25678                         err = -EADDRNOTAVAIL;
25679                         goto out;
25680                 }
25681 +               if (!v4_addr_in_nx_info(sk->sk_nx_info, v4addr, NXA_MASK_BIND)) {
25682 +                       err = -EADDRNOTAVAIL;
25683 +                       goto out;
25684 +               }
25685         } else {
25686                 if (addr_type != IPV6_ADDR_ANY) {
25687                         struct net_device *dev = NULL;
25688 @@ -342,6 +357,11 @@ int inet6_bind(struct socket *sock, stru
25689                                 }
25690                         }
25691
25692 +                       if (!v6_addr_in_nx_info(sk->sk_nx_info, &addr->sin6_addr, -1)) {
25693 +                               err = -EADDRNOTAVAIL;
25694 +                               goto out;
25695 +                       }
25696 +
25697                         /* ipv4 addr of the socket is invalid.  Only the
25698                          * unspecified and mapped address have a v4 equivalent.
25699                          */
25700 @@ -358,6 +378,9 @@ int inet6_bind(struct socket *sock, stru
25701                 }
25702         }
25703
25704 +       /* what's that for? */
25705 +       v6_set_sock_addr(inet, &nsa);
25706 +
25707         inet->inet_rcv_saddr = v4addr;
25708         inet->inet_saddr = v4addr;
25709
25710 @@ -459,9 +482,11 @@ int inet6_getname(struct socket *sock, s
25711                         return -ENOTCONN;
25712                 sin->sin6_port = inet->inet_dport;
25713                 ipv6_addr_copy(&sin->sin6_addr, &np->daddr);
25714 +               /* FIXME: remap lback? */
25715                 if (np->sndflow)
25716                         sin->sin6_flowinfo = np->flow_label;
25717         } else {
25718 +               /* FIXME: remap lback? */
25719                 if (ipv6_addr_any(&np->rcv_saddr))
25720                         ipv6_addr_copy(&sin->sin6_addr, &np->saddr);
25721                 else
25722 diff -NurpP --minimal linux-3.0-rc5/net/ipv6/fib6_rules.c linux-3.0-rc5-vs2.3.1-pre3/net/ipv6/fib6_rules.c
25723 --- linux-3.0-rc5/net/ipv6/fib6_rules.c 2011-05-22 16:18:00.000000000 +0200
25724 +++ linux-3.0-rc5-vs2.3.1-pre3/net/ipv6/fib6_rules.c    2011-06-10 22:11:24.000000000 +0200
25725 @@ -90,7 +90,7 @@ static int fib6_rule_action(struct fib_r
25726                                                ip6_dst_idev(&rt->dst)->dev,
25727                                                &flp6->daddr,
25728                                                rt6_flags2srcprefs(flags),
25729 -                                              &saddr))
25730 +                                              &saddr, NULL))
25731                                 goto again;
25732                         if (!ipv6_prefix_equal(&saddr, &r->src.addr,
25733                                                r->src.plen))
25734 diff -NurpP --minimal linux-3.0-rc5/net/ipv6/inet6_hashtables.c linux-3.0-rc5-vs2.3.1-pre3/net/ipv6/inet6_hashtables.c
25735 --- linux-3.0-rc5/net/ipv6/inet6_hashtables.c   2011-05-22 16:18:00.000000000 +0200
25736 +++ linux-3.0-rc5-vs2.3.1-pre3/net/ipv6/inet6_hashtables.c      2011-06-10 22:11:24.000000000 +0200
25737 @@ -16,6 +16,7 @@
25738
25739  #include <linux/module.h>
25740  #include <linux/random.h>
25741 +#include <linux/vs_inet6.h>
25742
25743  #include <net/inet_connection_sock.h>
25744  #include <net/inet_hashtables.h>
25745 @@ -82,7 +83,6 @@ struct sock *__inet6_lookup_established(
25746         unsigned int slot = hash & hashinfo->ehash_mask;
25747         struct inet_ehash_bucket *head = &hashinfo->ehash[slot];
25748
25749 -
25750         rcu_read_lock();
25751  begin:
25752         sk_nulls_for_each_rcu(sk, node, &head->chain) {
25753 @@ -94,7 +94,7 @@ begin:
25754                                 sock_put(sk);
25755                                 goto begin;
25756                         }
25757 -               goto out;
25758 +                       goto out;
25759                 }
25760         }
25761         if (get_nulls_value(node) != slot)
25762 @@ -140,6 +140,9 @@ static inline int compute_score(struct s
25763                         if (!ipv6_addr_equal(&np->rcv_saddr, daddr))
25764                                 return -1;
25765                         score++;
25766 +               } else {
25767 +                       if (!v6_addr_in_nx_info(sk->sk_nx_info, daddr, -1))
25768 +                               return -1;
25769                 }
25770                 if (sk->sk_bound_dev_if) {
25771                         if (sk->sk_bound_dev_if != dif)
25772 diff -NurpP --minimal linux-3.0-rc5/net/ipv6/ip6_output.c linux-3.0-rc5-vs2.3.1-pre3/net/ipv6/ip6_output.c
25773 --- linux-3.0-rc5/net/ipv6/ip6_output.c 2011-07-01 11:14:26.000000000 +0200
25774 +++ linux-3.0-rc5-vs2.3.1-pre3/net/ipv6/ip6_output.c    2011-06-16 14:16:51.000000000 +0200
25775 @@ -933,7 +933,8 @@ static int ip6_dst_lookup_tail(struct so
25776                 struct rt6_info *rt = (struct rt6_info *) *dst;
25777                 err = ip6_route_get_saddr(net, rt, &fl6->daddr,
25778                                           sk ? inet6_sk(sk)->srcprefs : 0,
25779 -                                         &fl6->saddr);
25780 +                                         &fl6->saddr,
25781 +                                         sk ? sk->sk_nx_info : NULL);
25782                 if (err)
25783                         goto out_err_release;
25784         }
25785 diff -NurpP --minimal linux-3.0-rc5/net/ipv6/ndisc.c linux-3.0-rc5-vs2.3.1-pre3/net/ipv6/ndisc.c
25786 --- linux-3.0-rc5/net/ipv6/ndisc.c      2011-07-01 11:14:26.000000000 +0200
25787 +++ linux-3.0-rc5-vs2.3.1-pre3/net/ipv6/ndisc.c 2011-06-10 22:11:24.000000000 +0200
25788 @@ -597,7 +597,7 @@ static void ndisc_send_na(struct net_dev
25789         } else {
25790                 if (ipv6_dev_get_saddr(dev_net(dev), dev, daddr,
25791                                        inet6_sk(dev_net(dev)->ipv6.ndisc_sk)->srcprefs,
25792 -                                      &tmpaddr))
25793 +                                      &tmpaddr, NULL))
25794                         return;
25795                 src_addr = &tmpaddr;
25796         }
25797 diff -NurpP --minimal linux-3.0-rc5/net/ipv6/raw.c linux-3.0-rc5-vs2.3.1-pre3/net/ipv6/raw.c
25798 --- linux-3.0-rc5/net/ipv6/raw.c        2011-07-01 11:14:26.000000000 +0200
25799 +++ linux-3.0-rc5-vs2.3.1-pre3/net/ipv6/raw.c   2011-06-10 22:11:24.000000000 +0200
25800 @@ -30,6 +30,7 @@
25801  #include <linux/icmpv6.h>
25802  #include <linux/netfilter.h>
25803  #include <linux/netfilter_ipv6.h>
25804 +#include <linux/vs_inet6.h>
25805  #include <linux/skbuff.h>
25806  #include <linux/compat.h>
25807  #include <asm/uaccess.h>
25808 @@ -284,6 +285,13 @@ static int rawv6_bind(struct sock *sk, s
25809                                 goto out_unlock;
25810                 }
25811
25812 +               if (!v6_addr_in_nx_info(sk->sk_nx_info, &addr->sin6_addr, -1)) {
25813 +                       err = -EADDRNOTAVAIL;
25814 +                       if (dev)
25815 +                               dev_put(dev);
25816 +                       goto out;
25817 +               }
25818 +
25819                 /* ipv4 addr of the socket is invalid.  Only the
25820                  * unspecified and mapped address have a v4 equivalent.
25821                  */
25822 diff -NurpP --minimal linux-3.0-rc5/net/ipv6/route.c linux-3.0-rc5-vs2.3.1-pre3/net/ipv6/route.c
25823 --- linux-3.0-rc5/net/ipv6/route.c      2011-07-01 11:14:26.000000000 +0200
25824 +++ linux-3.0-rc5-vs2.3.1-pre3/net/ipv6/route.c 2011-06-16 14:16:51.000000000 +0200
25825 @@ -54,6 +54,7 @@
25826  #include <net/xfrm.h>
25827  #include <net/netevent.h>
25828  #include <net/netlink.h>
25829 +#include <linux/vs_inet6.h>
25830
25831  #include <asm/uaccess.h>
25832
25833 @@ -2057,15 +2058,17 @@ int ip6_route_get_saddr(struct net *net,
25834                         struct rt6_info *rt,
25835                         const struct in6_addr *daddr,
25836                         unsigned int prefs,
25837 -                       struct in6_addr *saddr)
25838 +                       struct in6_addr *saddr,
25839 +                       struct nx_info *nxi)
25840  {
25841         struct inet6_dev *idev = ip6_dst_idev((struct dst_entry*)rt);
25842         int err = 0;
25843 -       if (rt->rt6i_prefsrc.plen)
25844 +       if (rt->rt6i_prefsrc.plen && (!nxi ||
25845 +           v6_addr_in_nx_info(nxi, &rt->rt6i_prefsrc.addr, NXA_TYPE_ADDR)))
25846                 ipv6_addr_copy(saddr, &rt->rt6i_prefsrc.addr);
25847         else
25848                 err = ipv6_dev_get_saddr(net, idev ? idev->dev : NULL,
25849 -                                        daddr, prefs, saddr);
25850 +                                        daddr, prefs, saddr, nxi);
25851         return err;
25852  }
25853
25854 @@ -2394,7 +2397,8 @@ static int rt6_fill_node(struct net *net
25855                         NLA_PUT_U32(skb, RTA_IIF, iif);
25856         } else if (dst) {
25857                 struct in6_addr saddr_buf;
25858 -               if (ip6_route_get_saddr(net, rt, dst, 0, &saddr_buf) == 0)
25859 +               if (ip6_route_get_saddr(net, rt, dst, 0, &saddr_buf,
25860 +                                       (skb->sk ? skb->sk->sk_nx_info : NULL)) == 0)
25861                         NLA_PUT(skb, RTA_PREFSRC, 16, &saddr_buf);
25862         }
25863
25864 diff -NurpP --minimal linux-3.0-rc5/net/ipv6/tcp_ipv6.c linux-3.0-rc5-vs2.3.1-pre3/net/ipv6/tcp_ipv6.c
25865 --- linux-3.0-rc5/net/ipv6/tcp_ipv6.c   2011-07-01 11:14:26.000000000 +0200
25866 +++ linux-3.0-rc5-vs2.3.1-pre3/net/ipv6/tcp_ipv6.c      2011-06-22 12:39:16.000000000 +0200
25867 @@ -69,6 +69,7 @@
25868
25869  #include <linux/crypto.h>
25870  #include <linux/scatterlist.h>
25871 +#include <linux/vs_inet6.h>
25872
25873  static void    tcp_v6_send_reset(struct sock *sk, struct sk_buff *skb);
25874  static void    tcp_v6_reqsk_send_ack(struct sock *sk, struct sk_buff *skb,
25875 @@ -161,8 +162,15 @@ static int tcp_v6_connect(struct sock *s
25876          *      connect() to INADDR_ANY means loopback (BSD'ism).
25877          */
25878
25879 -       if(ipv6_addr_any(&usin->sin6_addr))
25880 -               usin->sin6_addr.s6_addr[15] = 0x1;
25881 +       if(ipv6_addr_any(&usin->sin6_addr)) {
25882 +               struct nx_info *nxi =  sk->sk_nx_info;
25883 +
25884 +               if (nxi && nx_info_has_v6(nxi))
25885 +                       /* FIXME: remap lback? */
25886 +                       usin->sin6_addr = nxi->v6.ip;
25887 +               else
25888 +                       usin->sin6_addr.s6_addr[15] = 0x1;
25889 +       }
25890
25891         addr_type = ipv6_addr_type(&usin->sin6_addr);
25892
25893 diff -NurpP --minimal linux-3.0-rc5/net/ipv6/udp.c linux-3.0-rc5-vs2.3.1-pre3/net/ipv6/udp.c
25894 --- linux-3.0-rc5/net/ipv6/udp.c        2011-07-01 11:14:26.000000000 +0200
25895 +++ linux-3.0-rc5-vs2.3.1-pre3/net/ipv6/udp.c   2011-06-10 22:11:24.000000000 +0200
25896 @@ -45,41 +45,67 @@
25897  #include <net/tcp_states.h>
25898  #include <net/ip6_checksum.h>
25899  #include <net/xfrm.h>
25900 +#include <linux/vs_inet6.h>
25901
25902  #include <linux/proc_fs.h>
25903  #include <linux/seq_file.h>
25904  #include "udp_impl.h"
25905
25906 -int ipv6_rcv_saddr_equal(const struct sock *sk, const struct sock *sk2)
25907 +int ipv6_rcv_saddr_equal(const struct sock *sk1, const struct sock *sk2)
25908  {
25909 -       const struct in6_addr *sk_rcv_saddr6 = &inet6_sk(sk)->rcv_saddr;
25910 +       const struct in6_addr *sk1_rcv_saddr6 = &inet6_sk(sk1)->rcv_saddr;
25911         const struct in6_addr *sk2_rcv_saddr6 = inet6_rcv_saddr(sk2);
25912 -       __be32 sk1_rcv_saddr = sk_rcv_saddr(sk);
25913 +       __be32 sk1_rcv_saddr = sk_rcv_saddr(sk1);
25914         __be32 sk2_rcv_saddr = sk_rcv_saddr(sk2);
25915 -       int sk_ipv6only = ipv6_only_sock(sk);
25916 +       int sk1_ipv6only = ipv6_only_sock(sk1);
25917         int sk2_ipv6only = inet_v6_ipv6only(sk2);
25918 -       int addr_type = ipv6_addr_type(sk_rcv_saddr6);
25919 +       int addr_type = ipv6_addr_type(sk1_rcv_saddr6);
25920         int addr_type2 = sk2_rcv_saddr6 ? ipv6_addr_type(sk2_rcv_saddr6) : IPV6_ADDR_MAPPED;
25921
25922         /* if both are mapped, treat as IPv4 */
25923 -       if (addr_type == IPV6_ADDR_MAPPED && addr_type2 == IPV6_ADDR_MAPPED)
25924 -               return (!sk2_ipv6only &&
25925 +       if (addr_type == IPV6_ADDR_MAPPED && addr_type2 == IPV6_ADDR_MAPPED) {
25926 +               if (!sk2_ipv6only &&
25927                         (!sk1_rcv_saddr || !sk2_rcv_saddr ||
25928 -                         sk1_rcv_saddr == sk2_rcv_saddr));
25929 +                         sk1_rcv_saddr == sk2_rcv_saddr))
25930 +                       goto vs_v4;
25931 +               else
25932 +                       return 0;
25933 +       }
25934
25935         if (addr_type2 == IPV6_ADDR_ANY &&
25936             !(sk2_ipv6only && addr_type == IPV6_ADDR_MAPPED))
25937 -               return 1;
25938 +               goto vs;
25939
25940         if (addr_type == IPV6_ADDR_ANY &&
25941 -           !(sk_ipv6only && addr_type2 == IPV6_ADDR_MAPPED))
25942 -               return 1;
25943 +           !(sk1_ipv6only && addr_type2 == IPV6_ADDR_MAPPED))
25944 +               goto vs;
25945
25946         if (sk2_rcv_saddr6 &&
25947 -           ipv6_addr_equal(sk_rcv_saddr6, sk2_rcv_saddr6))
25948 -               return 1;
25949 +           ipv6_addr_equal(sk1_rcv_saddr6, sk2_rcv_saddr6))
25950 +               goto vs;
25951
25952         return 0;
25953 +
25954 +vs_v4:
25955 +       if (!sk1_rcv_saddr && !sk2_rcv_saddr)
25956 +               return nx_v4_addr_conflict(sk1->sk_nx_info, sk2->sk_nx_info);
25957 +       if (!sk2_rcv_saddr)
25958 +               return v4_addr_in_nx_info(sk1->sk_nx_info, sk2_rcv_saddr, -1);
25959 +       if (!sk1_rcv_saddr)
25960 +               return v4_addr_in_nx_info(sk2->sk_nx_info, sk1_rcv_saddr, -1);
25961 +       return 1;
25962 +vs:
25963 +       if (addr_type2 == IPV6_ADDR_ANY && addr_type == IPV6_ADDR_ANY)
25964 +               return nx_v6_addr_conflict(sk1->sk_nx_info, sk2->sk_nx_info);
25965 +       else if (addr_type2 == IPV6_ADDR_ANY)
25966 +               return v6_addr_in_nx_info(sk2->sk_nx_info, sk1_rcv_saddr6, -1);
25967 +       else if (addr_type == IPV6_ADDR_ANY) {
25968 +               if (addr_type2 == IPV6_ADDR_MAPPED)
25969 +                       return nx_v4_addr_conflict(sk1->sk_nx_info, sk2->sk_nx_info);
25970 +               else
25971 +                       return v6_addr_in_nx_info(sk1->sk_nx_info, sk2_rcv_saddr6, -1);
25972 +       }
25973 +       return 1;
25974  }
25975
25976  static unsigned int udp6_portaddr_hash(struct net *net,
25977 @@ -143,6 +169,10 @@ static inline int compute_score(struct s
25978                         if (!ipv6_addr_equal(&np->rcv_saddr, daddr))
25979                                 return -1;
25980                         score++;
25981 +               } else {
25982 +                       /* block non nx_info ips */
25983 +                       if (!v6_addr_in_nx_info(sk->sk_nx_info, daddr, -1))
25984 +                               return -1;
25985                 }
25986                 if (!ipv6_addr_any(&np->daddr)) {
25987                         if (!ipv6_addr_equal(&np->daddr, saddr))
25988 diff -NurpP --minimal linux-3.0-rc5/net/ipv6/xfrm6_policy.c linux-3.0-rc5-vs2.3.1-pre3/net/ipv6/xfrm6_policy.c
25989 --- linux-3.0-rc5/net/ipv6/xfrm6_policy.c       2011-07-01 11:14:26.000000000 +0200
25990 +++ linux-3.0-rc5-vs2.3.1-pre3/net/ipv6/xfrm6_policy.c  2011-06-10 22:11:24.000000000 +0200
25991 @@ -63,7 +63,7 @@ static int xfrm6_get_saddr(struct net *n
25992         dev = ip6_dst_idev(dst)->dev;
25993         ipv6_dev_get_saddr(dev_net(dev), dev,
25994                            (struct in6_addr *)&daddr->a6, 0,
25995 -                          (struct in6_addr *)&saddr->a6);
25996 +                          (struct in6_addr *)&saddr->a6, NULL);
25997         dst_release(dst);
25998         return 0;
25999  }
26000 diff -NurpP --minimal linux-3.0-rc5/netfilter/ipvs/ip_vs_xmit.c linux-3.0-rc5-vs2.3.1-pre3/net/netfilter/ipvs/ip_vs_xmit.c
26001 --- linux-3.0-rc5/net/netfilter/ipvs/ip_vs_xmit.c
26002 +++ linux-3.0-rc5-vs2.3.1-pre3/net/netfilter/ipvs/ip_vs_xmit.c
26003 @@ -213,7 +213,7 @@ __ip_vs_route_output_v6(struct net *net,
26004                 return dst;
26005         if (ipv6_addr_any(&fl6.saddr) &&
26006             ipv6_dev_get_saddr(net, ip6_dst_idev(dst)->dev,
26007 -                              &fl6.daddr, 0, &fl6.saddr) < 0)
26008 +                              &fl6.daddr, 0, &fl6.saddr, NULL) < 0)
26009                 goto out_err;
26010         if (do_xfrm) {
26011                 dst = xfrm_lookup(net, dst, flowi6_to_flowi(&fl6), NULL, 0);
26012 diff -NurpP --minimal linux-3.0-rc5/net/netlink/af_netlink.c linux-3.0-rc5-vs2.3.1-pre3/net/netlink/af_netlink.c
26013 --- linux-3.0-rc5/net/netlink/af_netlink.c      2011-07-01 11:14:27.000000000 +0200
26014 +++ linux-3.0-rc5-vs2.3.1-pre3/net/netlink/af_netlink.c 2011-06-10 22:11:24.000000000 +0200
26015 @@ -55,6 +55,9 @@
26016  #include <linux/types.h>
26017  #include <linux/audit.h>
26018  #include <linux/mutex.h>
26019 +#include <linux/vs_context.h>
26020 +#include <linux/vs_network.h>
26021 +#include <linux/vs_limit.h>
26022
26023  #include <net/net_namespace.h>
26024  #include <net/sock.h>
26025 @@ -1907,6 +1910,8 @@ static struct sock *netlink_seq_socket_i
26026                         sk_for_each(s, node, &hash->table[j]) {
26027                                 if (sock_net(s) != seq_file_net(seq))
26028                                         continue;
26029 +                               if (!nx_check(s->sk_nid, VS_WATCH_P | VS_IDENT))
26030 +                                       continue;
26031                                 if (off == pos) {
26032                                         iter->link = i;
26033                                         iter->hash_idx = j;
26034 @@ -1941,7 +1946,8 @@ static void *netlink_seq_next(struct seq
26035         s = v;
26036         do {
26037                 s = sk_next(s);
26038 -       } while (s && sock_net(s) != seq_file_net(seq));
26039 +       } while (s && (sock_net(s) != seq_file_net(seq) ||
26040 +               !nx_check(s->sk_nid, VS_WATCH_P | VS_IDENT)));
26041         if (s)
26042                 return s;
26043
26044 @@ -1953,7 +1959,8 @@ static void *netlink_seq_next(struct seq
26045
26046                 for (; j <= hash->mask; j++) {
26047                         s = sk_head(&hash->table[j]);
26048 -                       while (s && sock_net(s) != seq_file_net(seq))
26049 +                       while (s && (sock_net(s) != seq_file_net(seq) ||
26050 +                               !nx_check(s->sk_nid, VS_WATCH_P | VS_IDENT)))
26051                                 s = sk_next(s);
26052                         if (s) {
26053                                 iter->link = i;
26054 diff -NurpP --minimal linux-3.0-rc5/net/socket.c linux-3.0-rc5-vs2.3.1-pre3/net/socket.c
26055 --- linux-3.0-rc5/net/socket.c  2011-07-01 11:14:27.000000000 +0200
26056 +++ linux-3.0-rc5-vs2.3.1-pre3/net/socket.c     2011-06-16 14:16:51.000000000 +0200
26057 @@ -98,6 +98,10 @@
26058
26059  #include <net/sock.h>
26060  #include <linux/netfilter.h>
26061 +#include <linux/vs_base.h>
26062 +#include <linux/vs_socket.h>
26063 +#include <linux/vs_inet.h>
26064 +#include <linux/vs_inet6.h>
26065
26066  #include <linux/if_tun.h>
26067  #include <linux/ipv6_route.h>
26068 @@ -546,6 +550,7 @@ static inline int __sock_sendmsg_nosec(s
26069                                        struct msghdr *msg, size_t size)
26070  {
26071         struct sock_iocb *si = kiocb_to_siocb(iocb);
26072 +       size_t len;
26073
26074         sock_update_classid(sock->sk);
26075
26076 @@ -554,7 +559,22 @@ static inline int __sock_sendmsg_nosec(s
26077         si->msg = msg;
26078         si->size = size;
26079
26080 -       return sock->ops->sendmsg(iocb, sock, msg, size);
26081 +       len = sock->ops->sendmsg(iocb, sock, msg, size);
26082 +       if (sock->sk) {
26083 +               if (len == size)
26084 +                       vx_sock_send(sock->sk, size);
26085 +               else
26086 +                       vx_sock_fail(sock->sk, size);
26087 +       }
26088 +       vxdprintk(VXD_CBIT(net, 7),
26089 +               "__sock_sendmsg: %p[%p,%p,%p;%d/%d]:%d/%d",
26090 +               sock, sock->sk,
26091 +               (sock->sk)?sock->sk->sk_nx_info:0,
26092 +               (sock->sk)?sock->sk->sk_vx_info:0,
26093 +               (sock->sk)?sock->sk->sk_xid:0,
26094 +               (sock->sk)?sock->sk->sk_nid:0,
26095 +               (unsigned int)size, len);
26096 +       return len;
26097  }
26098
26099  static inline int __sock_sendmsg(struct kiocb *iocb, struct socket *sock,
26100 @@ -694,6 +714,7 @@ static inline int __sock_recvmsg_nosec(s
26101                                        struct msghdr *msg, size_t size, int flags)
26102  {
26103         struct sock_iocb *si = kiocb_to_siocb(iocb);
26104 +       int len;
26105
26106         sock_update_classid(sock->sk);
26107
26108 @@ -703,7 +724,18 @@ static inline int __sock_recvmsg_nosec(s
26109         si->size = size;
26110         si->flags = flags;
26111
26112 -       return sock->ops->recvmsg(iocb, sock, msg, size, flags);
26113 +       len = sock->ops->recvmsg(iocb, sock, msg, size, flags);
26114 +       if ((len >= 0) && sock->sk)
26115 +               vx_sock_recv(sock->sk, len);
26116 +       vxdprintk(VXD_CBIT(net, 7),
26117 +               "__sock_recvmsg: %p[%p,%p,%p;%d/%d]:%d/%d",
26118 +               sock, sock->sk,
26119 +               (sock->sk)?sock->sk->sk_nx_info:0,
26120 +               (sock->sk)?sock->sk->sk_vx_info:0,
26121 +               (sock->sk)?sock->sk->sk_xid:0,
26122 +               (sock->sk)?sock->sk->sk_nid:0,
26123 +               (unsigned int)size, len);
26124 +       return len;
26125  }
26126
26127  static inline int __sock_recvmsg(struct kiocb *iocb, struct socket *sock,
26128 @@ -1188,6 +1220,13 @@ int __sock_create(struct net *net, int f
26129         if (type < 0 || type >= SOCK_MAX)
26130                 return -EINVAL;
26131
26132 +       if (!nx_check(0, VS_ADMIN)) {
26133 +               if (family == PF_INET && !current_nx_info_has_v4())
26134 +                       return -EAFNOSUPPORT;
26135 +               if (family == PF_INET6 && !current_nx_info_has_v6())
26136 +                       return -EAFNOSUPPORT;
26137 +       }
26138 +
26139         /* Compatibility.
26140
26141            This uglymoron is moved from INET layer to here to avoid
26142 @@ -1323,6 +1362,7 @@ SYSCALL_DEFINE3(socket, int, family, int
26143         if (retval < 0)
26144                 goto out;
26145
26146 +       set_bit(SOCK_USER_SOCKET, &sock->flags);
26147         retval = sock_map_fd(sock, flags & (O_CLOEXEC | O_NONBLOCK));
26148         if (retval < 0)
26149                 goto out_release;
26150 @@ -1364,10 +1404,12 @@ SYSCALL_DEFINE4(socketpair, int, family,
26151         err = sock_create(family, type, protocol, &sock1);
26152         if (err < 0)
26153                 goto out;
26154 +       set_bit(SOCK_USER_SOCKET, &sock1->flags);
26155
26156         err = sock_create(family, type, protocol, &sock2);
26157         if (err < 0)
26158                 goto out_release_1;
26159 +       set_bit(SOCK_USER_SOCKET, &sock2->flags);
26160
26161         err = sock1->ops->socketpair(sock1, sock2);
26162         if (err < 0)
26163 diff -NurpP --minimal linux-3.0-rc5/net/sunrpc/auth.c linux-3.0-rc5-vs2.3.1-pre3/net/sunrpc/auth.c
26164 --- linux-3.0-rc5/net/sunrpc/auth.c     2011-07-01 11:14:27.000000000 +0200
26165 +++ linux-3.0-rc5-vs2.3.1-pre3/net/sunrpc/auth.c        2011-06-10 22:11:24.000000000 +0200
26166 @@ -14,6 +14,7 @@
26167  #include <linux/hash.h>
26168  #include <linux/sunrpc/clnt.h>
26169  #include <linux/spinlock.h>
26170 +#include <linux/vs_tag.h>
26171
26172  #ifdef RPC_DEBUG
26173  # define RPCDBG_FACILITY       RPCDBG_AUTH
26174 @@ -427,6 +428,7 @@ rpcauth_lookupcred(struct rpc_auth *auth
26175         memset(&acred, 0, sizeof(acred));
26176         acred.uid = cred->fsuid;
26177         acred.gid = cred->fsgid;
26178 +       acred.tag = dx_current_tag();
26179         acred.group_info = get_group_info(((struct cred *)cred)->group_info);
26180
26181         ret = auth->au_ops->lookup_cred(auth, &acred, flags);
26182 @@ -467,6 +469,7 @@ rpcauth_bind_root_cred(struct rpc_task *
26183         struct auth_cred acred = {
26184                 .uid = 0,
26185                 .gid = 0,
26186 +               .tag = dx_current_tag(),
26187         };
26188
26189         dprintk("RPC: %5u looking up %s cred\n",
26190 diff -NurpP --minimal linux-3.0-rc5/net/sunrpc/auth_unix.c linux-3.0-rc5-vs2.3.1-pre3/net/sunrpc/auth_unix.c
26191 --- linux-3.0-rc5/net/sunrpc/auth_unix.c        2010-10-21 13:08:01.000000000 +0200
26192 +++ linux-3.0-rc5-vs2.3.1-pre3/net/sunrpc/auth_unix.c   2011-06-10 22:11:24.000000000 +0200
26193 @@ -12,12 +12,14 @@
26194  #include <linux/module.h>
26195  #include <linux/sunrpc/clnt.h>
26196  #include <linux/sunrpc/auth.h>
26197 +#include <linux/vs_tag.h>
26198
26199  #define NFS_NGROUPS    16
26200
26201  struct unx_cred {
26202         struct rpc_cred         uc_base;
26203         gid_t                   uc_gid;
26204 +       tag_t                   uc_tag;
26205         gid_t                   uc_gids[NFS_NGROUPS];
26206  };
26207  #define uc_uid                 uc_base.cr_uid
26208 @@ -78,6 +80,7 @@ unx_create_cred(struct rpc_auth *auth, s
26209                 groups = NFS_NGROUPS;
26210
26211         cred->uc_gid = acred->gid;
26212 +       cred->uc_tag = acred->tag;
26213         for (i = 0; i < groups; i++)
26214                 cred->uc_gids[i] = GROUP_AT(acred->group_info, i);
26215         if (i < NFS_NGROUPS)
26216 @@ -119,7 +122,9 @@ unx_match(struct auth_cred *acred, struc
26217         unsigned int i;
26218
26219
26220 -       if (cred->uc_uid != acred->uid || cred->uc_gid != acred->gid)
26221 +       if (cred->uc_uid != acred->uid ||
26222 +               cred->uc_gid != acred->gid ||
26223 +               cred->uc_tag != acred->tag)
26224                 return 0;
26225
26226         if (acred->group_info != NULL)
26227 @@ -142,7 +147,7 @@ unx_marshal(struct rpc_task *task, __be3
26228         struct rpc_clnt *clnt = task->tk_client;
26229         struct unx_cred *cred = container_of(task->tk_rqstp->rq_cred, struct unx_cred, uc_base);
26230         __be32          *base, *hold;
26231 -       int             i;
26232 +       int             i, tag;
26233
26234         *p++ = htonl(RPC_AUTH_UNIX);
26235         base = p++;
26236 @@ -152,9 +157,12 @@ unx_marshal(struct rpc_task *task, __be3
26237          * Copy the UTS nodename captured when the client was created.
26238          */
26239         p = xdr_encode_array(p, clnt->cl_nodename, clnt->cl_nodelen);
26240 +       tag = task->tk_client->cl_tag;
26241
26242 -       *p++ = htonl((u32) cred->uc_uid);
26243 -       *p++ = htonl((u32) cred->uc_gid);
26244 +       *p++ = htonl((u32) TAGINO_UID(tag,
26245 +               cred->uc_uid, cred->uc_tag));
26246 +       *p++ = htonl((u32) TAGINO_GID(tag,
26247 +               cred->uc_gid, cred->uc_tag));
26248         hold = p++;
26249         for (i = 0; i < 16 && cred->uc_gids[i] != (gid_t) NOGROUP; i++)
26250                 *p++ = htonl((u32) cred->uc_gids[i]);
26251 diff -NurpP --minimal linux-3.0-rc5/net/sunrpc/clnt.c linux-3.0-rc5-vs2.3.1-pre3/net/sunrpc/clnt.c
26252 --- linux-3.0-rc5/net/sunrpc/clnt.c     2011-07-01 11:14:27.000000000 +0200
26253 +++ linux-3.0-rc5-vs2.3.1-pre3/net/sunrpc/clnt.c        2011-07-01 11:35:35.000000000 +0200
26254 @@ -31,6 +31,7 @@
26255  #include <linux/in.h>
26256  #include <linux/in6.h>
26257  #include <linux/un.h>
26258 +#include <linux/vs_cvirt.h>
26259
26260  #include <linux/sunrpc/clnt.h>
26261  #include <linux/sunrpc/rpc_pipe_fs.h>
26262 @@ -362,6 +363,9 @@ struct rpc_clnt *rpc_create(struct rpc_c
26263         if (!(args->flags & RPC_CLNT_CREATE_QUIET))
26264                 clnt->cl_chatty = 1;
26265
26266 +       /* TODO: handle RPC_CLNT_CREATE_TAGGED
26267 +       if (args->flags & RPC_CLNT_CREATE_TAGGED)
26268 +               clnt->cl_tag = 1; */
26269         return clnt;
26270  }
26271  EXPORT_SYMBOL_GPL(rpc_create);
26272 diff -NurpP --minimal linux-3.0-rc5/net/unix/af_unix.c linux-3.0-rc5-vs2.3.1-pre3/net/unix/af_unix.c
26273 --- linux-3.0-rc5/net/unix/af_unix.c    2011-07-01 11:14:27.000000000 +0200
26274 +++ linux-3.0-rc5-vs2.3.1-pre3/net/unix/af_unix.c       2011-06-10 22:11:24.000000000 +0200
26275 @@ -114,6 +114,8 @@
26276  #include <linux/mount.h>
26277  #include <net/checksum.h>
26278  #include <linux/security.h>
26279 +#include <linux/vs_context.h>
26280 +#include <linux/vs_limit.h>
26281
26282  static struct hlist_head unix_socket_table[UNIX_HASH_SIZE + 1];
26283  static DEFINE_SPINLOCK(unix_table_lock);
26284 @@ -258,6 +260,8 @@ static struct sock *__unix_find_socket_b
26285                 if (!net_eq(sock_net(s), net))
26286                         continue;
26287
26288 +               if (!nx_check(s->sk_nid, VS_WATCH_P | VS_IDENT))
26289 +                       continue;
26290                 if (u->addr->len == len &&
26291                     !memcmp(u->addr->name, sunname, len))
26292                         goto found;
26293 @@ -2208,6 +2212,8 @@ static struct sock *unix_seq_idx(struct
26294         for (s = first_unix_socket(&iter->i); s; s = next_unix_socket(&iter->i, s)) {
26295                 if (sock_net(s) != seq_file_net(seq))
26296                         continue;
26297 +               if (!nx_check(s->sk_nid, VS_WATCH_P | VS_IDENT))
26298 +                       continue;
26299                 if (off == pos)
26300                         return s;
26301                 ++off;
26302 @@ -2232,7 +2238,8 @@ static void *unix_seq_next(struct seq_fi
26303                 sk = first_unix_socket(&iter->i);
26304         else
26305                 sk = next_unix_socket(&iter->i, sk);
26306 -       while (sk && (sock_net(sk) != seq_file_net(seq)))
26307 +       while (sk && (sock_net(sk) != seq_file_net(seq) ||
26308 +               !nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT)))
26309                 sk = next_unix_socket(&iter->i, sk);
26310         return sk;
26311  }
26312 diff -NurpP --minimal linux-3.0-rc5/scripts/checksyscalls.sh linux-3.0-rc5-vs2.3.1-pre3/scripts/checksyscalls.sh
26313 --- linux-3.0-rc5/scripts/checksyscalls.sh      2011-03-15 18:07:46.000000000 +0100
26314 +++ linux-3.0-rc5-vs2.3.1-pre3/scripts/checksyscalls.sh 2011-06-10 22:11:24.000000000 +0200
26315 @@ -193,7 +193,6 @@ cat << EOF
26316  #define __IGNORE_afs_syscall
26317  #define __IGNORE_getpmsg
26318  #define __IGNORE_putpmsg
26319 -#define __IGNORE_vserver
26320  EOF
26321  }
26322
26323 diff -NurpP --minimal linux-3.0-rc5/security/commoncap.c linux-3.0-rc5-vs2.3.1-pre3/security/commoncap.c
26324 --- linux-3.0-rc5/security/commoncap.c  2011-07-01 11:14:27.000000000 +0200
26325 +++ linux-3.0-rc5-vs2.3.1-pre3/security/commoncap.c     2011-06-16 14:16:51.000000000 +0200
26326 @@ -62,6 +62,7 @@ int cap_netlink_recv(struct sk_buff *skb
26327                 return -EPERM;
26328         return 0;
26329  }
26330 +
26331  EXPORT_SYMBOL(cap_netlink_recv);
26332
26333  /**
26334 @@ -83,14 +84,20 @@ EXPORT_SYMBOL(cap_netlink_recv);
26335  int cap_capable(struct task_struct *tsk, const struct cred *cred,
26336                 struct user_namespace *targ_ns, int cap, int audit)
26337  {
26338 +       struct vx_info *vxi = tsk->vx_info;
26339 +
26340         for (;;) {
26341                 /* The creator of the user namespace has all caps. */
26342                 if (targ_ns != &init_user_ns && targ_ns->creator == cred->user)
26343                         return 0;
26344
26345                 /* Do we have the necessary capabilities? */
26346 -               if (targ_ns == cred->user->user_ns)
26347 -                       return cap_raised(cred->cap_effective, cap) ? 0 : -EPERM;
26348 +               if (targ_ns == cred->user->user_ns) {
26349 +                       if (vx_info_flags(vxi, VXF_STATE_SETUP, 0) &&
26350 +                           cap_raised(cred->cap_effective, cap))
26351 +                               return 0;
26352 +                       return vx_cap_raised(vxi, cred->cap_effective, cap) ? 0 : -EPERM;
26353 +               }
26354
26355                 /* Have we tried all of the parent namespaces? */
26356                 if (targ_ns == &init_user_ns)
26357 @@ -611,7 +618,7 @@ int cap_inode_setxattr(struct dentry *de
26358
26359         if (!strncmp(name, XATTR_SECURITY_PREFIX,
26360                      sizeof(XATTR_SECURITY_PREFIX) - 1) &&
26361 -           !capable(CAP_SYS_ADMIN))
26362 +               !vx_capable(CAP_SYS_ADMIN, VXC_FS_SECURITY))
26363                 return -EPERM;
26364         return 0;
26365  }
26366 @@ -637,7 +644,7 @@ int cap_inode_removexattr(struct dentry
26367
26368         if (!strncmp(name, XATTR_SECURITY_PREFIX,
26369                      sizeof(XATTR_SECURITY_PREFIX) - 1) &&
26370 -           !capable(CAP_SYS_ADMIN))
26371 +               !vx_capable(CAP_SYS_ADMIN, VXC_FS_SECURITY))
26372                 return -EPERM;
26373         return 0;
26374  }
26375 @@ -971,3 +978,4 @@ int cap_file_mmap(struct file *file, uns
26376         }
26377         return ret;
26378  }
26379 +
26380 diff -NurpP --minimal linux-3.0-rc5/security/selinux/hooks.c linux-3.0-rc5-vs2.3.1-pre3/security/selinux/hooks.c
26381 --- linux-3.0-rc5/security/selinux/hooks.c      2011-07-01 11:14:28.000000000 +0200
26382 +++ linux-3.0-rc5-vs2.3.1-pre3/security/selinux/hooks.c 2011-06-15 02:40:14.000000000 +0200
26383 @@ -67,7 +67,6 @@
26384  #include <linux/dccp.h>
26385  #include <linux/quota.h>
26386  #include <linux/un.h>          /* for Unix socket types */
26387 -#include <net/af_unix.h>       /* for Unix socket types */
26388  #include <linux/parser.h>
26389  #include <linux/nfs_mount.h>
26390  #include <net/ipv6.h>