1 diff -NurpP --minimal linux-2.6.36/arch/alpha/Kconfig linux-2.6.36-vs2.3.0.36.33/arch/alpha/Kconfig
2 --- linux-2.6.36/arch/alpha/Kconfig 2010-10-21 13:06:45.000000000 +0200
3 +++ linux-2.6.36-vs2.3.0.36.33/arch/alpha/Kconfig 2010-10-21 13:09:36.000000000 +0200
4 @@ -677,6 +677,8 @@ config DUMMY_CONSOLE
8 +source "kernel/vserver/Kconfig"
10 source "security/Kconfig"
12 source "crypto/Kconfig"
13 diff -NurpP --minimal linux-2.6.36/arch/alpha/kernel/entry.S linux-2.6.36-vs2.3.0.36.33/arch/alpha/kernel/entry.S
14 --- linux-2.6.36/arch/alpha/kernel/entry.S 2010-10-21 13:06:45.000000000 +0200
15 +++ linux-2.6.36-vs2.3.0.36.33/arch/alpha/kernel/entry.S 2010-10-21 13:09:36.000000000 +0200
16 @@ -860,24 +860,15 @@ sys_getxgid:
25 - /* See linux/kernel/timer.c sys_getppid for discussion
27 - ldq $3, TASK_GROUP_LEADER($2)
28 - ldq $4, TASK_REAL_PARENT($3)
29 - ldl $0, TASK_TGID($2)
30 -1: ldl $1, TASK_TGID($4)
34 - ldq $3, TASK_GROUP_LEADER($2)
35 - ldq $4, TASK_REAL_PARENT($3)
48 diff -NurpP --minimal linux-2.6.36/arch/alpha/kernel/ptrace.c linux-2.6.36-vs2.3.0.36.33/arch/alpha/kernel/ptrace.c
49 --- linux-2.6.36/arch/alpha/kernel/ptrace.c 2010-07-07 18:30:51.000000000 +0200
50 +++ linux-2.6.36-vs2.3.0.36.33/arch/alpha/kernel/ptrace.c 2010-10-21 13:09:36.000000000 +0200
52 #include <linux/user.h>
53 #include <linux/security.h>
54 #include <linux/signal.h>
55 +#include <linux/vs_base.h>
57 #include <asm/uaccess.h>
58 #include <asm/pgtable.h>
59 diff -NurpP --minimal linux-2.6.36/arch/alpha/kernel/systbls.S linux-2.6.36-vs2.3.0.36.33/arch/alpha/kernel/systbls.S
60 --- linux-2.6.36/arch/alpha/kernel/systbls.S 2010-10-21 13:06:46.000000000 +0200
61 +++ linux-2.6.36-vs2.3.0.36.33/arch/alpha/kernel/systbls.S 2010-10-21 13:09:36.000000000 +0200
62 @@ -446,7 +446,7 @@ sys_call_table:
63 .quad sys_stat64 /* 425 */
66 - .quad sys_ni_syscall /* sys_vserver */
67 + .quad sys_vserver /* sys_vserver */
68 .quad sys_ni_syscall /* sys_mbind */
69 .quad sys_ni_syscall /* sys_get_mempolicy */
70 .quad sys_ni_syscall /* sys_set_mempolicy */
71 diff -NurpP --minimal linux-2.6.36/arch/alpha/kernel/traps.c linux-2.6.36-vs2.3.0.36.33/arch/alpha/kernel/traps.c
72 --- linux-2.6.36/arch/alpha/kernel/traps.c 2010-10-21 13:06:46.000000000 +0200
73 +++ linux-2.6.36-vs2.3.0.36.33/arch/alpha/kernel/traps.c 2010-10-21 13:09:36.000000000 +0200
74 @@ -183,7 +183,8 @@ die_if_kernel(char * str, struct pt_regs
76 printk("CPU %d ", hard_smp_processor_id());
78 - printk("%s(%d): %s %ld\n", current->comm, task_pid_nr(current), str, err);
79 + printk("%s(%d[#%u]): %s %ld\n", current->comm,
80 + task_pid_nr(current), current->xid, str, err);
81 dik_show_regs(regs, r9_15);
83 dik_show_trace((unsigned long *)(regs+1));
84 diff -NurpP --minimal linux-2.6.36/arch/arm/include/asm/tlb.h linux-2.6.36-vs2.3.0.36.33/arch/arm/include/asm/tlb.h
85 --- linux-2.6.36/arch/arm/include/asm/tlb.h 2009-09-10 15:25:15.000000000 +0200
86 +++ linux-2.6.36-vs2.3.0.36.33/arch/arm/include/asm/tlb.h 2010-10-21 13:09:36.000000000 +0200
89 #else /* !CONFIG_MMU */
91 +#include <linux/vs_memory.h>
92 #include <asm/pgalloc.h>
95 diff -NurpP --minimal linux-2.6.36/arch/arm/Kconfig linux-2.6.36-vs2.3.0.36.33/arch/arm/Kconfig
96 --- linux-2.6.36/arch/arm/Kconfig 2010-10-21 13:06:46.000000000 +0200
97 +++ linux-2.6.36-vs2.3.0.36.33/arch/arm/Kconfig 2010-10-21 13:09:36.000000000 +0200
98 @@ -1803,6 +1803,8 @@ source "fs/Kconfig"
100 source "arch/arm/Kconfig.debug"
102 +source "kernel/vserver/Kconfig"
104 source "security/Kconfig"
106 source "crypto/Kconfig"
107 diff -NurpP --minimal linux-2.6.36/arch/arm/kernel/calls.S linux-2.6.36-vs2.3.0.36.33/arch/arm/kernel/calls.S
108 --- linux-2.6.36/arch/arm/kernel/calls.S 2010-10-21 13:06:46.000000000 +0200
109 +++ linux-2.6.36-vs2.3.0.36.33/arch/arm/kernel/calls.S 2010-10-21 13:09:36.000000000 +0200
111 /* 310 */ CALL(sys_request_key)
113 CALL(ABI(sys_semtimedop, sys_oabi_semtimedop))
114 -/* vserver */ CALL(sys_ni_syscall)
117 /* 315 */ CALL(sys_ioprio_get)
118 CALL(sys_inotify_init)
119 diff -NurpP --minimal linux-2.6.36/arch/arm/kernel/process.c linux-2.6.36-vs2.3.0.36.33/arch/arm/kernel/process.c
120 --- linux-2.6.36/arch/arm/kernel/process.c 2010-10-21 13:06:46.000000000 +0200
121 +++ linux-2.6.36-vs2.3.0.36.33/arch/arm/kernel/process.c 2010-10-21 13:09:36.000000000 +0200
122 @@ -295,7 +295,8 @@ void __show_regs(struct pt_regs *regs)
123 void show_regs(struct pt_regs * regs)
126 - printk("Pid: %d, comm: %20s\n", task_pid_nr(current), current->comm);
127 + printk("Pid: %d[#%u], comm: %20s\n",
128 + task_pid_nr(current), current->xid, current->comm);
132 diff -NurpP --minimal linux-2.6.36/arch/arm/kernel/traps.c linux-2.6.36-vs2.3.0.36.33/arch/arm/kernel/traps.c
133 --- linux-2.6.36/arch/arm/kernel/traps.c 2010-10-21 13:06:46.000000000 +0200
134 +++ linux-2.6.36-vs2.3.0.36.33/arch/arm/kernel/traps.c 2010-10-21 13:09:36.000000000 +0200
135 @@ -244,8 +244,8 @@ static int __die(const char *str, int er
139 - printk(KERN_EMERG "Process %.*s (pid: %d, stack limit = 0x%p)\n",
140 - TASK_COMM_LEN, tsk->comm, task_pid_nr(tsk), thread + 1);
141 + printk(KERN_EMERG "Process %.*s (pid: %d:#%u, stack limit = 0x%p)\n",
142 + TASK_COMM_LEN, tsk->comm, task_pid_nr(tsk), tsk->xid, thread + 1);
144 if (!user_mode(regs) || in_interrupt()) {
145 dump_mem(KERN_EMERG, "Stack: ", regs->ARM_sp,
146 diff -NurpP --minimal linux-2.6.36/arch/cris/Kconfig linux-2.6.36-vs2.3.0.36.33/arch/cris/Kconfig
147 --- linux-2.6.36/arch/cris/Kconfig 2010-10-21 13:06:51.000000000 +0200
148 +++ linux-2.6.36-vs2.3.0.36.33/arch/cris/Kconfig 2010-10-21 13:09:36.000000000 +0200
149 @@ -694,6 +694,8 @@ source "drivers/staging/Kconfig"
151 source "arch/cris/Kconfig.debug"
153 +source "kernel/vserver/Kconfig"
155 source "security/Kconfig"
157 source "crypto/Kconfig"
158 diff -NurpP --minimal linux-2.6.36/arch/frv/kernel/kernel_thread.S linux-2.6.36-vs2.3.0.36.33/arch/frv/kernel/kernel_thread.S
159 --- linux-2.6.36/arch/frv/kernel/kernel_thread.S 2008-12-25 00:26:37.000000000 +0100
160 +++ linux-2.6.36-vs2.3.0.36.33/arch/frv/kernel/kernel_thread.S 2010-10-21 13:09:36.000000000 +0200
161 @@ -37,7 +37,7 @@ kernel_thread:
163 # start by forking the current process, but with shared VM
164 setlos.p #__NR_clone,gr7 ; syscall number
165 - ori gr10,#CLONE_VM,gr8 ; first syscall arg [clone_flags]
166 + ori gr10,#CLONE_KT,gr8 ; first syscall arg [clone_flags]
167 sethi.p #0xe4e4,gr9 ; second syscall arg [newsp]
169 setlos.p #0,gr10 ; third syscall arg [parent_tidptr]
170 diff -NurpP --minimal linux-2.6.36/arch/h8300/Kconfig linux-2.6.36-vs2.3.0.36.33/arch/h8300/Kconfig
171 --- linux-2.6.36/arch/h8300/Kconfig 2010-10-21 13:06:51.000000000 +0200
172 +++ linux-2.6.36-vs2.3.0.36.33/arch/h8300/Kconfig 2010-10-21 13:09:36.000000000 +0200
173 @@ -226,6 +226,8 @@ source "fs/Kconfig"
175 source "arch/h8300/Kconfig.debug"
177 +source "kernel/vserver/Kconfig"
179 source "security/Kconfig"
181 source "crypto/Kconfig"
182 diff -NurpP --minimal linux-2.6.36/arch/ia64/include/asm/tlb.h linux-2.6.36-vs2.3.0.36.33/arch/ia64/include/asm/tlb.h
183 --- linux-2.6.36/arch/ia64/include/asm/tlb.h 2010-02-25 11:51:26.000000000 +0100
184 +++ linux-2.6.36-vs2.3.0.36.33/arch/ia64/include/asm/tlb.h 2010-10-21 13:09:36.000000000 +0200
186 #include <linux/mm.h>
187 #include <linux/pagemap.h>
188 #include <linux/swap.h>
189 +#include <linux/vs_memory.h>
191 #include <asm/pgalloc.h>
192 #include <asm/processor.h>
193 diff -NurpP --minimal linux-2.6.36/arch/ia64/Kconfig linux-2.6.36-vs2.3.0.36.33/arch/ia64/Kconfig
194 --- linux-2.6.36/arch/ia64/Kconfig 2010-10-21 13:06:52.000000000 +0200
195 +++ linux-2.6.36-vs2.3.0.36.33/arch/ia64/Kconfig 2010-10-21 13:09:36.000000000 +0200
196 @@ -671,6 +671,8 @@ source "fs/Kconfig"
198 source "arch/ia64/Kconfig.debug"
200 +source "kernel/vserver/Kconfig"
202 source "security/Kconfig"
204 source "crypto/Kconfig"
205 diff -NurpP --minimal linux-2.6.36/arch/ia64/kernel/entry.S linux-2.6.36-vs2.3.0.36.33/arch/ia64/kernel/entry.S
206 --- linux-2.6.36/arch/ia64/kernel/entry.S 2010-10-21 13:06:52.000000000 +0200
207 +++ linux-2.6.36-vs2.3.0.36.33/arch/ia64/kernel/entry.S 2010-10-21 13:09:36.000000000 +0200
208 @@ -1714,7 +1714,7 @@ sys_call_table:
210 data8 sys_mq_getsetattr
212 - data8 sys_ni_syscall // reserved for vserver
214 data8 sys_waitid // 1270
216 data8 sys_request_key
217 diff -NurpP --minimal linux-2.6.36/arch/ia64/kernel/perfmon.c linux-2.6.36-vs2.3.0.36.33/arch/ia64/kernel/perfmon.c
218 --- linux-2.6.36/arch/ia64/kernel/perfmon.c 2010-10-21 13:06:52.000000000 +0200
219 +++ linux-2.6.36-vs2.3.0.36.33/arch/ia64/kernel/perfmon.c 2010-10-21 13:09:36.000000000 +0200
221 #include <linux/completion.h>
222 #include <linux/tracehook.h>
223 #include <linux/slab.h>
224 +#include <linux/vs_memory.h>
226 #include <asm/errno.h>
227 #include <asm/intrinsics.h>
228 diff -NurpP --minimal linux-2.6.36/arch/ia64/kernel/process.c linux-2.6.36-vs2.3.0.36.33/arch/ia64/kernel/process.c
229 --- linux-2.6.36/arch/ia64/kernel/process.c 2010-10-21 13:06:52.000000000 +0200
230 +++ linux-2.6.36-vs2.3.0.36.33/arch/ia64/kernel/process.c 2010-10-21 13:09:36.000000000 +0200
231 @@ -113,8 +113,8 @@ show_regs (struct pt_regs *regs)
232 unsigned long ip = regs->cr_iip + ia64_psr(regs)->ri;
235 - printk("\nPid: %d, CPU %d, comm: %20s\n", task_pid_nr(current),
236 - smp_processor_id(), current->comm);
237 + printk("\nPid: %d[#%u], CPU %d, comm: %20s\n", task_pid_nr(current),
238 + current->xid, smp_processor_id(), current->comm);
239 printk("psr : %016lx ifs : %016lx ip : [<%016lx>] %s (%s)\n",
240 regs->cr_ipsr, regs->cr_ifs, ip, print_tainted(),
241 init_utsname()->release);
242 diff -NurpP --minimal linux-2.6.36/arch/ia64/kernel/ptrace.c linux-2.6.36-vs2.3.0.36.33/arch/ia64/kernel/ptrace.c
243 --- linux-2.6.36/arch/ia64/kernel/ptrace.c 2010-08-02 16:52:04.000000000 +0200
244 +++ linux-2.6.36-vs2.3.0.36.33/arch/ia64/kernel/ptrace.c 2010-10-21 13:09:36.000000000 +0200
246 #include <linux/regset.h>
247 #include <linux/elf.h>
248 #include <linux/tracehook.h>
249 +#include <linux/vs_base.h>
251 #include <asm/pgtable.h>
252 #include <asm/processor.h>
253 diff -NurpP --minimal linux-2.6.36/arch/ia64/kernel/traps.c linux-2.6.36-vs2.3.0.36.33/arch/ia64/kernel/traps.c
254 --- linux-2.6.36/arch/ia64/kernel/traps.c 2010-07-07 18:31:01.000000000 +0200
255 +++ linux-2.6.36-vs2.3.0.36.33/arch/ia64/kernel/traps.c 2010-10-21 13:09:36.000000000 +0200
256 @@ -59,8 +59,9 @@ die (const char *str, struct pt_regs *re
259 if (++die.lock_owner_depth < 3) {
260 - printk("%s[%d]: %s %ld [%d]\n",
261 - current->comm, task_pid_nr(current), str, err, ++die_counter);
262 + printk("%s[%d[#%u]]: %s %ld [%d]\n",
263 + current->comm, task_pid_nr(current), current->xid,
264 + str, err, ++die_counter);
265 if (notify_die(DIE_OOPS, str, regs, err, 255, SIGSEGV)
268 @@ -323,8 +324,9 @@ handle_fpu_swa (int fp_fault, struct pt_
269 if ((last.count & 15) < 5 && (ia64_fetchadd(1, &last.count, acq) & 15) < 5) {
270 last.time = current_jiffies + 5 * HZ;
272 - "%s(%d): floating-point assist fault at ip %016lx, isr %016lx\n",
273 - current->comm, task_pid_nr(current), regs->cr_iip + ia64_psr(regs)->ri, isr);
274 + "%s(%d[#%u]): floating-point assist fault at ip %016lx, isr %016lx\n",
275 + current->comm, task_pid_nr(current), current->xid,
276 + regs->cr_iip + ia64_psr(regs)->ri, isr);
280 diff -NurpP --minimal linux-2.6.36/arch/ia64/mm/fault.c linux-2.6.36-vs2.3.0.36.33/arch/ia64/mm/fault.c
281 --- linux-2.6.36/arch/ia64/mm/fault.c 2010-08-02 16:52:04.000000000 +0200
282 +++ linux-2.6.36-vs2.3.0.36.33/arch/ia64/mm/fault.c 2010-10-21 13:09:36.000000000 +0200
284 #include <linux/interrupt.h>
285 #include <linux/kprobes.h>
286 #include <linux/kdebug.h>
287 +#include <linux/vs_memory.h>
289 #include <asm/pgtable.h>
290 #include <asm/processor.h>
291 diff -NurpP --minimal linux-2.6.36/arch/m32r/kernel/traps.c linux-2.6.36-vs2.3.0.36.33/arch/m32r/kernel/traps.c
292 --- linux-2.6.36/arch/m32r/kernel/traps.c 2009-12-03 20:01:57.000000000 +0100
293 +++ linux-2.6.36-vs2.3.0.36.33/arch/m32r/kernel/traps.c 2010-10-21 13:09:36.000000000 +0200
294 @@ -196,8 +196,9 @@ static void show_registers(struct pt_reg
296 printk("SPI: %08lx\n", sp);
298 - printk("Process %s (pid: %d, process nr: %d, stackpage=%08lx)",
299 - current->comm, task_pid_nr(current), 0xffff & i, 4096+(unsigned long)current);
300 + printk("Process %s (pid: %d[#%u], process nr: %d, stackpage=%08lx)",
301 + current->comm, task_pid_nr(current), current->xid,
302 + 0xffff & i, 4096+(unsigned long)current);
305 * When in-kernel, we also print out the stack and code at the
306 diff -NurpP --minimal linux-2.6.36/arch/m68k/Kconfig linux-2.6.36-vs2.3.0.36.33/arch/m68k/Kconfig
307 --- linux-2.6.36/arch/m68k/Kconfig 2010-10-21 13:06:52.000000000 +0200
308 +++ linux-2.6.36-vs2.3.0.36.33/arch/m68k/Kconfig 2010-10-21 13:09:36.000000000 +0200
309 @@ -616,6 +616,8 @@ source "fs/Kconfig"
311 source "arch/m68k/Kconfig.debug"
313 +source "kernel/vserver/Kconfig"
315 source "security/Kconfig"
317 source "crypto/Kconfig"
318 diff -NurpP --minimal linux-2.6.36/arch/m68k/kernel/ptrace.c linux-2.6.36-vs2.3.0.36.33/arch/m68k/kernel/ptrace.c
319 --- linux-2.6.36/arch/m68k/kernel/ptrace.c 2010-07-07 18:31:02.000000000 +0200
320 +++ linux-2.6.36-vs2.3.0.36.33/arch/m68k/kernel/ptrace.c 2010-10-21 13:09:36.000000000 +0200
322 #include <linux/ptrace.h>
323 #include <linux/user.h>
324 #include <linux/signal.h>
325 +#include <linux/vs_base.h>
327 #include <asm/uaccess.h>
328 #include <asm/page.h>
329 @@ -254,6 +255,8 @@ long arch_ptrace(struct task_struct *chi
330 ret = ptrace_request(child, request, addr, data);
333 + if (!vx_check(vx_task_xid(child), VS_WATCH_P | VS_IDENT))
338 diff -NurpP --minimal linux-2.6.36/arch/m68k/kernel/traps.c linux-2.6.36-vs2.3.0.36.33/arch/m68k/kernel/traps.c
339 --- linux-2.6.36/arch/m68k/kernel/traps.c 2010-08-02 16:52:04.000000000 +0200
340 +++ linux-2.6.36-vs2.3.0.36.33/arch/m68k/kernel/traps.c 2010-10-21 13:09:36.000000000 +0200
341 @@ -906,8 +906,8 @@ void show_registers(struct pt_regs *regs
342 printk("d4: %08lx d5: %08lx a0: %08lx a1: %08lx\n",
343 regs->d4, regs->d5, regs->a0, regs->a1);
345 - printk("Process %s (pid: %d, task=%p)\n",
346 - current->comm, task_pid_nr(current), current);
347 + printk("Process %s (pid: %d[#%u], task=%p)\n",
348 + current->comm, task_pid_nr(current), current->xid, current);
349 addr = (unsigned long)&fp->un;
350 printk("Frame format=%X ", regs->format);
351 switch (regs->format) {
352 diff -NurpP --minimal linux-2.6.36/arch/m68knommu/Kconfig linux-2.6.36-vs2.3.0.36.33/arch/m68knommu/Kconfig
353 --- linux-2.6.36/arch/m68knommu/Kconfig 2010-10-21 13:06:53.000000000 +0200
354 +++ linux-2.6.36-vs2.3.0.36.33/arch/m68knommu/Kconfig 2010-10-21 13:09:36.000000000 +0200
355 @@ -730,6 +730,8 @@ source "fs/Kconfig"
357 source "arch/m68knommu/Kconfig.debug"
359 +source "kernel/vserver/Kconfig"
361 source "security/Kconfig"
363 source "crypto/Kconfig"
364 diff -NurpP --minimal linux-2.6.36/arch/m68knommu/kernel/traps.c linux-2.6.36-vs2.3.0.36.33/arch/m68knommu/kernel/traps.c
365 --- linux-2.6.36/arch/m68knommu/kernel/traps.c 2009-09-10 15:25:23.000000000 +0200
366 +++ linux-2.6.36-vs2.3.0.36.33/arch/m68knommu/kernel/traps.c 2010-10-21 13:09:36.000000000 +0200
367 @@ -78,8 +78,9 @@ void die_if_kernel(char *str, struct pt_
368 printk(KERN_EMERG "d4: %08lx d5: %08lx a0: %08lx a1: %08lx\n",
369 fp->d4, fp->d5, fp->a0, fp->a1);
371 - printk(KERN_EMERG "Process %s (pid: %d, stackpage=%08lx)\n",
372 - current->comm, current->pid, PAGE_SIZE+(unsigned long)current);
373 + printk(KERN_EMERG "Process %s (pid: %d[#%u], stackpage=%08lx)\n",
374 + current->comm, task_pid_nr(current), current->xid,
375 + PAGE_SIZE+(unsigned long)current);
376 show_stack(NULL, (unsigned long *)(fp + 1));
377 add_taint(TAINT_DIE);
379 diff -NurpP --minimal linux-2.6.36/arch/mips/Kconfig linux-2.6.36-vs2.3.0.36.33/arch/mips/Kconfig
380 --- linux-2.6.36/arch/mips/Kconfig 2010-10-21 13:06:53.000000000 +0200
381 +++ linux-2.6.36-vs2.3.0.36.33/arch/mips/Kconfig 2010-10-21 13:09:36.000000000 +0200
382 @@ -2298,6 +2298,8 @@ source "fs/Kconfig"
384 source "arch/mips/Kconfig.debug"
386 +source "kernel/vserver/Kconfig"
388 source "security/Kconfig"
390 source "crypto/Kconfig"
391 diff -NurpP --minimal linux-2.6.36/arch/mips/kernel/ptrace.c linux-2.6.36-vs2.3.0.36.33/arch/mips/kernel/ptrace.c
392 --- linux-2.6.36/arch/mips/kernel/ptrace.c 2010-10-21 13:06:54.000000000 +0200
393 +++ linux-2.6.36-vs2.3.0.36.33/arch/mips/kernel/ptrace.c 2010-10-21 13:09:36.000000000 +0200
395 #include <linux/security.h>
396 #include <linux/audit.h>
397 #include <linux/seccomp.h>
398 +#include <linux/vs_base.h>
400 #include <asm/byteorder.h>
402 @@ -259,6 +260,9 @@ long arch_ptrace(struct task_struct *chi
406 + if (!vx_check(vx_task_xid(child), VS_WATCH_P | VS_IDENT))
410 /* when I and D space are separate, these will need to be fixed. */
411 case PTRACE_PEEKTEXT: /* read word at location addr. */
412 diff -NurpP --minimal linux-2.6.36/arch/mips/kernel/scall32-o32.S linux-2.6.36-vs2.3.0.36.33/arch/mips/kernel/scall32-o32.S
413 --- linux-2.6.36/arch/mips/kernel/scall32-o32.S 2010-10-21 13:06:54.000000000 +0200
414 +++ linux-2.6.36-vs2.3.0.36.33/arch/mips/kernel/scall32-o32.S 2010-10-21 13:09:36.000000000 +0200
415 @@ -524,7 +524,7 @@ einval: li v0, -ENOSYS
416 sys sys_mq_timedreceive 5
417 sys sys_mq_notify 2 /* 4275 */
418 sys sys_mq_getsetattr 3
419 - sys sys_ni_syscall 0 /* sys_vserver */
422 sys sys_ni_syscall 0 /* available, was setaltroot */
423 sys sys_add_key 5 /* 4280 */
424 diff -NurpP --minimal linux-2.6.36/arch/mips/kernel/scall64-64.S linux-2.6.36-vs2.3.0.36.33/arch/mips/kernel/scall64-64.S
425 --- linux-2.6.36/arch/mips/kernel/scall64-64.S 2010-10-21 13:06:54.000000000 +0200
426 +++ linux-2.6.36-vs2.3.0.36.33/arch/mips/kernel/scall64-64.S 2010-10-21 13:09:36.000000000 +0200
427 @@ -363,7 +363,7 @@ sys_call_table:
428 PTR sys_mq_timedreceive
430 PTR sys_mq_getsetattr /* 5235 */
431 - PTR sys_ni_syscall /* sys_vserver */
434 PTR sys_ni_syscall /* available, was setaltroot */
436 diff -NurpP --minimal linux-2.6.36/arch/mips/kernel/scall64-n32.S linux-2.6.36-vs2.3.0.36.33/arch/mips/kernel/scall64-n32.S
437 --- linux-2.6.36/arch/mips/kernel/scall64-n32.S 2010-10-21 13:06:54.000000000 +0200
438 +++ linux-2.6.36-vs2.3.0.36.33/arch/mips/kernel/scall64-n32.S 2010-10-21 13:09:36.000000000 +0200
439 @@ -362,7 +362,7 @@ EXPORT(sysn32_call_table)
440 PTR compat_sys_mq_timedreceive
441 PTR compat_sys_mq_notify
442 PTR compat_sys_mq_getsetattr
443 - PTR sys_ni_syscall /* 6240, sys_vserver */
444 + PTR sys32_vserver /* 6240 */
445 PTR compat_sys_waitid
446 PTR sys_ni_syscall /* available, was setaltroot */
448 diff -NurpP --minimal linux-2.6.36/arch/mips/kernel/scall64-o32.S linux-2.6.36-vs2.3.0.36.33/arch/mips/kernel/scall64-o32.S
449 --- linux-2.6.36/arch/mips/kernel/scall64-o32.S 2010-10-21 13:06:54.000000000 +0200
450 +++ linux-2.6.36-vs2.3.0.36.33/arch/mips/kernel/scall64-o32.S 2010-10-21 13:09:36.000000000 +0200
451 @@ -481,7 +481,7 @@ sys_call_table:
452 PTR compat_sys_mq_timedreceive
453 PTR compat_sys_mq_notify /* 4275 */
454 PTR compat_sys_mq_getsetattr
455 - PTR sys_ni_syscall /* sys_vserver */
458 PTR sys_ni_syscall /* available, was setaltroot */
459 PTR sys_add_key /* 4280 */
460 diff -NurpP --minimal linux-2.6.36/arch/mips/kernel/traps.c linux-2.6.36-vs2.3.0.36.33/arch/mips/kernel/traps.c
461 --- linux-2.6.36/arch/mips/kernel/traps.c 2010-10-21 13:06:54.000000000 +0200
462 +++ linux-2.6.36-vs2.3.0.36.33/arch/mips/kernel/traps.c 2010-10-21 13:09:36.000000000 +0200
463 @@ -341,9 +341,10 @@ void show_registers(struct pt_regs *regs
467 - printk("Process %s (pid: %d, threadinfo=%p, task=%p, tls=%0*lx)\n",
468 - current->comm, current->pid, current_thread_info(), current,
469 - field, current_thread_info()->tp_value);
470 + printk("Process %s (pid: %d:#%u, threadinfo=%p, task=%p, tls=%0*lx)\n",
471 + current->comm, task_pid_nr(current), current->xid,
472 + current_thread_info(), current,
473 + field, current_thread_info()->tp_value);
474 if (cpu_has_userlocal) {
477 diff -NurpP --minimal linux-2.6.36/arch/parisc/Kconfig linux-2.6.36-vs2.3.0.36.33/arch/parisc/Kconfig
478 --- linux-2.6.36/arch/parisc/Kconfig 2010-10-21 13:06:55.000000000 +0200
479 +++ linux-2.6.36-vs2.3.0.36.33/arch/parisc/Kconfig 2010-10-21 13:09:36.000000000 +0200
480 @@ -295,6 +295,8 @@ source "fs/Kconfig"
482 source "arch/parisc/Kconfig.debug"
484 +source "kernel/vserver/Kconfig"
486 source "security/Kconfig"
488 source "crypto/Kconfig"
489 diff -NurpP --minimal linux-2.6.36/arch/parisc/kernel/syscall_table.S linux-2.6.36-vs2.3.0.36.33/arch/parisc/kernel/syscall_table.S
490 --- linux-2.6.36/arch/parisc/kernel/syscall_table.S 2010-07-07 18:31:04.000000000 +0200
491 +++ linux-2.6.36-vs2.3.0.36.33/arch/parisc/kernel/syscall_table.S 2010-10-21 13:09:36.000000000 +0200
493 ENTRY_COMP(mbind) /* 260 */
494 ENTRY_COMP(get_mempolicy)
495 ENTRY_COMP(set_mempolicy)
496 - ENTRY_SAME(ni_syscall) /* 263: reserved for vserver */
497 + ENTRY_DIFF(vserver)
499 ENTRY_SAME(request_key) /* 265 */
501 diff -NurpP --minimal linux-2.6.36/arch/parisc/kernel/traps.c linux-2.6.36-vs2.3.0.36.33/arch/parisc/kernel/traps.c
502 --- linux-2.6.36/arch/parisc/kernel/traps.c 2009-09-10 15:25:40.000000000 +0200
503 +++ linux-2.6.36-vs2.3.0.36.33/arch/parisc/kernel/traps.c 2010-10-21 13:09:36.000000000 +0200
504 @@ -236,8 +236,9 @@ void die_if_kernel(char *str, struct pt_
508 - printk(KERN_CRIT "%s (pid %d): %s (code %ld) at " RFMT "\n",
509 - current->comm, task_pid_nr(current), str, err, regs->iaoq[0]);
510 + printk(KERN_CRIT "%s (pid %d:#%u): %s (code %ld) at " RFMT "\n",
511 + current->comm, task_pid_nr(current), current->xid,
512 + str, err, regs->iaoq[0]);
513 #ifdef PRINT_USER_FAULTS
514 /* XXX for debugging only */
516 @@ -270,8 +271,8 @@ void die_if_kernel(char *str, struct pt_
517 pdc_console_restart();
520 - printk(KERN_CRIT "%s (pid %d): %s (code %ld)\n",
521 - current->comm, task_pid_nr(current), str, err);
522 + printk(KERN_CRIT "%s (pid %d:#%u): %s (code %ld)\n",
523 + current->comm, task_pid_nr(current), current->xid, str, err);
525 /* Wot's wrong wif bein' racy? */
526 if (current->thread.flags & PARISC_KERNEL_DEATH) {
527 diff -NurpP --minimal linux-2.6.36/arch/parisc/mm/fault.c linux-2.6.36-vs2.3.0.36.33/arch/parisc/mm/fault.c
528 --- linux-2.6.36/arch/parisc/mm/fault.c 2010-08-02 16:52:06.000000000 +0200
529 +++ linux-2.6.36-vs2.3.0.36.33/arch/parisc/mm/fault.c 2010-10-21 13:09:36.000000000 +0200
530 @@ -237,8 +237,9 @@ bad_area:
532 #ifdef PRINT_USER_FAULTS
533 printk(KERN_DEBUG "\n");
534 - printk(KERN_DEBUG "do_page_fault() pid=%d command='%s' type=%lu address=0x%08lx\n",
535 - task_pid_nr(tsk), tsk->comm, code, address);
536 + printk(KERN_DEBUG "do_page_fault() pid=%d:#%u "
537 + "command='%s' type=%lu address=0x%08lx\n",
538 + task_pid_nr(tsk), tsk->xid, tsk->comm, code, address);
540 printk(KERN_DEBUG "vm_start = 0x%08lx, vm_end = 0x%08lx\n",
541 vma->vm_start, vma->vm_end);
542 diff -NurpP --minimal linux-2.6.36/arch/powerpc/include/asm/unistd.h linux-2.6.36-vs2.3.0.36.33/arch/powerpc/include/asm/unistd.h
543 --- linux-2.6.36/arch/powerpc/include/asm/unistd.h 2010-10-21 13:06:56.000000000 +0200
544 +++ linux-2.6.36-vs2.3.0.36.33/arch/powerpc/include/asm/unistd.h 2010-10-21 13:09:36.000000000 +0200
547 #define __NR_rtas 255
548 #define __NR_sys_debug_setcontext 256
549 -/* Number 257 is reserved for vserver */
550 +#define __NR_vserver 257
551 #define __NR_migrate_pages 258
552 #define __NR_mbind 259
553 #define __NR_get_mempolicy 260
554 diff -NurpP --minimal linux-2.6.36/arch/powerpc/Kconfig linux-2.6.36-vs2.3.0.36.33/arch/powerpc/Kconfig
555 --- linux-2.6.36/arch/powerpc/Kconfig 2010-10-21 13:06:55.000000000 +0200
556 +++ linux-2.6.36-vs2.3.0.36.33/arch/powerpc/Kconfig 2010-10-21 13:09:36.000000000 +0200
557 @@ -974,6 +974,8 @@ source "lib/Kconfig"
559 source "arch/powerpc/Kconfig.debug"
561 +source "kernel/vserver/Kconfig"
563 source "security/Kconfig"
566 diff -NurpP --minimal linux-2.6.36/arch/powerpc/kernel/process.c linux-2.6.36-vs2.3.0.36.33/arch/powerpc/kernel/process.c
567 --- linux-2.6.36/arch/powerpc/kernel/process.c 2010-10-21 13:06:56.000000000 +0200
568 +++ linux-2.6.36-vs2.3.0.36.33/arch/powerpc/kernel/process.c 2010-10-21 13:09:36.000000000 +0200
569 @@ -634,8 +634,9 @@ void show_regs(struct pt_regs * regs)
571 printk("DAR: "REG", DSISR: "REG"\n", regs->dar, regs->dsisr);
573 - printk("TASK = %p[%d] '%s' THREAD: %p",
574 - current, task_pid_nr(current), current->comm, task_thread_info(current));
575 + printk("TASK = %p[%d,#%u] '%s' THREAD: %p",
576 + current, task_pid_nr(current), current->xid,
577 + current->comm, task_thread_info(current));
580 printk(" CPU: %d", raw_smp_processor_id());
581 diff -NurpP --minimal linux-2.6.36/arch/powerpc/kernel/traps.c linux-2.6.36-vs2.3.0.36.33/arch/powerpc/kernel/traps.c
582 --- linux-2.6.36/arch/powerpc/kernel/traps.c 2010-10-21 13:06:57.000000000 +0200
583 +++ linux-2.6.36-vs2.3.0.36.33/arch/powerpc/kernel/traps.c 2010-10-21 13:09:36.000000000 +0200
584 @@ -1048,8 +1048,9 @@ void nonrecoverable_exception(struct pt_
586 void trace_syscall(struct pt_regs *regs)
588 - printk("Task: %p(%d), PC: %08lX/%08lX, Syscall: %3ld, Result: %s%ld %s\n",
589 - current, task_pid_nr(current), regs->nip, regs->link, regs->gpr[0],
590 + printk("Task: %p(%d[#%u]), PC: %08lX/%08lX, Syscall: %3ld, Result: %s%ld %s\n",
591 + current, task_pid_nr(current), current->xid,
592 + regs->nip, regs->link, regs->gpr[0],
593 regs->ccr&0x10000000?"Error=":"", regs->gpr[3], print_tainted());
596 diff -NurpP --minimal linux-2.6.36/arch/powerpc/kernel/vdso.c linux-2.6.36-vs2.3.0.36.33/arch/powerpc/kernel/vdso.c
597 --- linux-2.6.36/arch/powerpc/kernel/vdso.c 2010-08-02 16:52:07.000000000 +0200
598 +++ linux-2.6.36-vs2.3.0.36.33/arch/powerpc/kernel/vdso.c 2010-10-21 13:09:36.000000000 +0200
600 #include <linux/security.h>
601 #include <linux/bootmem.h>
602 #include <linux/memblock.h>
603 +#include <linux/vs_memory.h>
605 #include <asm/pgtable.h>
606 #include <asm/system.h>
607 diff -NurpP --minimal linux-2.6.36/arch/s390/include/asm/tlb.h linux-2.6.36-vs2.3.0.36.33/arch/s390/include/asm/tlb.h
608 --- linux-2.6.36/arch/s390/include/asm/tlb.h 2010-10-21 13:06:58.000000000 +0200
609 +++ linux-2.6.36-vs2.3.0.36.33/arch/s390/include/asm/tlb.h 2010-10-21 13:09:36.000000000 +0200
612 #include <linux/mm.h>
613 #include <linux/swap.h>
614 +#include <linux/vs_memory.h>
616 #include <asm/processor.h>
617 #include <asm/pgalloc.h>
619 diff -NurpP --minimal linux-2.6.36/arch/s390/include/asm/unistd.h linux-2.6.36-vs2.3.0.36.33/arch/s390/include/asm/unistd.h
620 --- linux-2.6.36/arch/s390/include/asm/unistd.h 2010-10-21 13:06:58.000000000 +0200
621 +++ linux-2.6.36-vs2.3.0.36.33/arch/s390/include/asm/unistd.h 2010-10-21 13:09:36.000000000 +0200
623 #define __NR_clock_gettime (__NR_timer_create+6)
624 #define __NR_clock_getres (__NR_timer_create+7)
625 #define __NR_clock_nanosleep (__NR_timer_create+8)
626 -/* Number 263 is reserved for vserver */
627 +#define __NR_vserver 263
628 #define __NR_statfs64 265
629 #define __NR_fstatfs64 266
630 #define __NR_remap_file_pages 267
631 diff -NurpP --minimal linux-2.6.36/arch/s390/Kconfig linux-2.6.36-vs2.3.0.36.33/arch/s390/Kconfig
632 --- linux-2.6.36/arch/s390/Kconfig 2010-10-21 13:06:57.000000000 +0200
633 +++ linux-2.6.36-vs2.3.0.36.33/arch/s390/Kconfig 2010-10-21 13:09:36.000000000 +0200
634 @@ -620,6 +620,8 @@ source "fs/Kconfig"
636 source "arch/s390/Kconfig.debug"
638 +source "kernel/vserver/Kconfig"
640 source "security/Kconfig"
642 source "crypto/Kconfig"
643 diff -NurpP --minimal linux-2.6.36/arch/s390/kernel/ptrace.c linux-2.6.36-vs2.3.0.36.33/arch/s390/kernel/ptrace.c
644 --- linux-2.6.36/arch/s390/kernel/ptrace.c 2010-08-02 16:52:09.000000000 +0200
645 +++ linux-2.6.36-vs2.3.0.36.33/arch/s390/kernel/ptrace.c 2010-10-21 13:09:36.000000000 +0200
647 #include <linux/regset.h>
648 #include <linux/tracehook.h>
649 #include <linux/seccomp.h>
650 +#include <linux/vs_base.h>
651 #include <trace/syscall.h>
652 #include <asm/compat.h>
653 #include <asm/segment.h>
654 diff -NurpP --minimal linux-2.6.36/arch/s390/kernel/syscalls.S linux-2.6.36-vs2.3.0.36.33/arch/s390/kernel/syscalls.S
655 --- linux-2.6.36/arch/s390/kernel/syscalls.S 2010-10-21 13:06:58.000000000 +0200
656 +++ linux-2.6.36-vs2.3.0.36.33/arch/s390/kernel/syscalls.S 2010-10-21 13:09:36.000000000 +0200
657 @@ -271,7 +271,7 @@ SYSCALL(sys_clock_settime,sys_clock_sett
658 SYSCALL(sys_clock_gettime,sys_clock_gettime,sys32_clock_gettime_wrapper) /* 260 */
659 SYSCALL(sys_clock_getres,sys_clock_getres,sys32_clock_getres_wrapper)
660 SYSCALL(sys_clock_nanosleep,sys_clock_nanosleep,sys32_clock_nanosleep_wrapper)
661 -NI_SYSCALL /* reserved for vserver */
662 +SYSCALL(sys_vserver,sys_vserver,sys32_vserver)
663 SYSCALL(sys_s390_fadvise64_64,sys_ni_syscall,sys32_fadvise64_64_wrapper)
664 SYSCALL(sys_statfs64,sys_statfs64,compat_sys_statfs64_wrapper)
665 SYSCALL(sys_fstatfs64,sys_fstatfs64,compat_sys_fstatfs64_wrapper)
666 diff -NurpP --minimal linux-2.6.36/arch/sh/Kconfig linux-2.6.36-vs2.3.0.36.33/arch/sh/Kconfig
667 --- linux-2.6.36/arch/sh/Kconfig 2010-10-21 13:06:58.000000000 +0200
668 +++ linux-2.6.36-vs2.3.0.36.33/arch/sh/Kconfig 2010-10-21 13:09:36.000000000 +0200
669 @@ -878,6 +878,8 @@ source "fs/Kconfig"
671 source "arch/sh/Kconfig.debug"
673 +source "kernel/vserver/Kconfig"
675 source "security/Kconfig"
677 source "crypto/Kconfig"
678 diff -NurpP --minimal linux-2.6.36/arch/sh/kernel/irq.c linux-2.6.36-vs2.3.0.36.33/arch/sh/kernel/irq.c
679 --- linux-2.6.36/arch/sh/kernel/irq.c 2010-08-02 16:52:10.000000000 +0200
680 +++ linux-2.6.36-vs2.3.0.36.33/arch/sh/kernel/irq.c 2010-10-21 13:09:36.000000000 +0200
682 #include <linux/seq_file.h>
683 #include <linux/ftrace.h>
684 #include <linux/delay.h>
685 +// #include <linux/vs_context.h>
686 #include <asm/processor.h>
687 #include <asm/machvec.h>
688 #include <asm/uaccess.h>
689 diff -NurpP --minimal linux-2.6.36/arch/sh/kernel/vsyscall/vsyscall.c linux-2.6.36-vs2.3.0.36.33/arch/sh/kernel/vsyscall/vsyscall.c
690 --- linux-2.6.36/arch/sh/kernel/vsyscall/vsyscall.c 2010-07-07 18:31:10.000000000 +0200
691 +++ linux-2.6.36-vs2.3.0.36.33/arch/sh/kernel/vsyscall/vsyscall.c 2010-10-21 13:09:36.000000000 +0200
693 #include <linux/elf.h>
694 #include <linux/sched.h>
695 #include <linux/err.h>
696 +#include <linux/vs_memory.h>
699 * Should the kernel map a VDSO page into processes and pass its
700 diff -NurpP --minimal linux-2.6.36/arch/sparc/include/asm/tlb_64.h linux-2.6.36-vs2.3.0.36.33/arch/sparc/include/asm/tlb_64.h
701 --- linux-2.6.36/arch/sparc/include/asm/tlb_64.h 2009-09-10 15:25:45.000000000 +0200
702 +++ linux-2.6.36-vs2.3.0.36.33/arch/sparc/include/asm/tlb_64.h 2010-10-21 13:09:36.000000000 +0200
705 #include <linux/swap.h>
706 #include <linux/pagemap.h>
707 +#include <linux/vs_memory.h>
708 #include <asm/pgalloc.h>
709 #include <asm/tlbflush.h>
710 #include <asm/mmu_context.h>
711 diff -NurpP --minimal linux-2.6.36/arch/sparc/include/asm/unistd.h linux-2.6.36-vs2.3.0.36.33/arch/sparc/include/asm/unistd.h
712 --- linux-2.6.36/arch/sparc/include/asm/unistd.h 2010-10-21 13:06:58.000000000 +0200
713 +++ linux-2.6.36-vs2.3.0.36.33/arch/sparc/include/asm/unistd.h 2010-10-21 13:09:36.000000000 +0200
715 #define __NR_timer_getoverrun 264
716 #define __NR_timer_delete 265
717 #define __NR_timer_create 266
718 -/* #define __NR_vserver 267 Reserved for VSERVER */
719 +#define __NR_vserver 267
720 #define __NR_io_setup 268
721 #define __NR_io_destroy 269
722 #define __NR_io_submit 270
723 diff -NurpP --minimal linux-2.6.36/arch/sparc/Kconfig linux-2.6.36-vs2.3.0.36.33/arch/sparc/Kconfig
724 --- linux-2.6.36/arch/sparc/Kconfig 2010-10-21 13:06:58.000000000 +0200
725 +++ linux-2.6.36-vs2.3.0.36.33/arch/sparc/Kconfig 2010-10-21 13:09:36.000000000 +0200
726 @@ -565,6 +565,8 @@ source "fs/Kconfig"
728 source "arch/sparc/Kconfig.debug"
730 +source "kernel/vserver/Kconfig"
732 source "security/Kconfig"
734 source "crypto/Kconfig"
735 diff -NurpP --minimal linux-2.6.36/arch/sparc/kernel/systbls_32.S linux-2.6.36-vs2.3.0.36.33/arch/sparc/kernel/systbls_32.S
736 --- linux-2.6.36/arch/sparc/kernel/systbls_32.S 2010-10-21 13:06:59.000000000 +0200
737 +++ linux-2.6.36-vs2.3.0.36.33/arch/sparc/kernel/systbls_32.S 2010-10-21 13:09:36.000000000 +0200
738 @@ -70,7 +70,7 @@ sys_call_table:
739 /*250*/ .long sys_mremap, sys_sysctl, sys_getsid, sys_fdatasync, sys_nfsservctl
740 /*255*/ .long sys_sync_file_range, sys_clock_settime, sys_clock_gettime, sys_clock_getres, sys_clock_nanosleep
741 /*260*/ .long sys_sched_getaffinity, sys_sched_setaffinity, sys_timer_settime, sys_timer_gettime, sys_timer_getoverrun
742 -/*265*/ .long sys_timer_delete, sys_timer_create, sys_nis_syscall, sys_io_setup, sys_io_destroy
743 +/*265*/ .long sys_timer_delete, sys_timer_create, sys_vserver, sys_io_setup, sys_io_destroy
744 /*270*/ .long sys_io_submit, sys_io_cancel, sys_io_getevents, sys_mq_open, sys_mq_unlink
745 /*275*/ .long sys_mq_timedsend, sys_mq_timedreceive, sys_mq_notify, sys_mq_getsetattr, sys_waitid
746 /*280*/ .long sys_tee, sys_add_key, sys_request_key, sys_keyctl, sys_openat
747 diff -NurpP --minimal linux-2.6.36/arch/sparc/kernel/systbls_64.S linux-2.6.36-vs2.3.0.36.33/arch/sparc/kernel/systbls_64.S
748 --- linux-2.6.36/arch/sparc/kernel/systbls_64.S 2010-10-21 13:06:59.000000000 +0200
749 +++ linux-2.6.36-vs2.3.0.36.33/arch/sparc/kernel/systbls_64.S 2010-10-21 13:09:36.000000000 +0200
750 @@ -71,7 +71,7 @@ sys_call_table32:
751 /*250*/ .word sys_mremap, compat_sys_sysctl, sys32_getsid, sys_fdatasync, sys32_nfsservctl
752 .word sys32_sync_file_range, compat_sys_clock_settime, compat_sys_clock_gettime, compat_sys_clock_getres, sys32_clock_nanosleep
753 /*260*/ .word compat_sys_sched_getaffinity, compat_sys_sched_setaffinity, sys32_timer_settime, compat_sys_timer_gettime, sys_timer_getoverrun
754 - .word sys_timer_delete, compat_sys_timer_create, sys_ni_syscall, compat_sys_io_setup, sys_io_destroy
755 + .word sys_timer_delete, compat_sys_timer_create, sys32_vserver, compat_sys_io_setup, sys_io_destroy
756 /*270*/ .word sys32_io_submit, sys_io_cancel, compat_sys_io_getevents, sys32_mq_open, sys_mq_unlink
757 .word compat_sys_mq_timedsend, compat_sys_mq_timedreceive, compat_sys_mq_notify, compat_sys_mq_getsetattr, compat_sys_waitid
758 /*280*/ .word sys32_tee, sys_add_key, sys_request_key, sys_keyctl, compat_sys_openat
759 @@ -147,7 +147,7 @@ sys_call_table:
760 /*250*/ .word sys_64_mremap, sys_sysctl, sys_getsid, sys_fdatasync, sys_nfsservctl
761 .word sys_sync_file_range, sys_clock_settime, sys_clock_gettime, sys_clock_getres, sys_clock_nanosleep
762 /*260*/ .word sys_sched_getaffinity, sys_sched_setaffinity, sys_timer_settime, sys_timer_gettime, sys_timer_getoverrun
763 - .word sys_timer_delete, sys_timer_create, sys_ni_syscall, sys_io_setup, sys_io_destroy
764 + .word sys_timer_delete, sys_timer_create, sys_vserver, sys_io_setup, sys_io_destroy
765 /*270*/ .word sys_io_submit, sys_io_cancel, sys_io_getevents, sys_mq_open, sys_mq_unlink
766 .word sys_mq_timedsend, sys_mq_timedreceive, sys_mq_notify, sys_mq_getsetattr, sys_waitid
767 /*280*/ .word sys_tee, sys_add_key, sys_request_key, sys_keyctl, sys_openat
768 diff -NurpP --minimal linux-2.6.36/arch/um/include/asm/tlb.h linux-2.6.36-vs2.3.0.36.33/arch/um/include/asm/tlb.h
769 --- linux-2.6.36/arch/um/include/asm/tlb.h 2009-09-10 15:25:46.000000000 +0200
770 +++ linux-2.6.36-vs2.3.0.36.33/arch/um/include/asm/tlb.h 2010-10-21 13:09:36.000000000 +0200
773 #include <linux/pagemap.h>
774 #include <linux/swap.h>
775 +#include <linux/vs_memory.h>
776 #include <asm/percpu.h>
777 #include <asm/pgalloc.h>
778 #include <asm/tlbflush.h>
779 diff -NurpP --minimal linux-2.6.36/arch/um/include/shared/kern_constants.h linux-2.6.36-vs2.3.0.36.33/arch/um/include/shared/kern_constants.h
780 --- linux-2.6.36/arch/um/include/shared/kern_constants.h 1970-01-01 01:00:00.000000000 +0100
781 +++ linux-2.6.36-vs2.3.0.36.33/arch/um/include/shared/kern_constants.h 2010-10-21 13:09:36.000000000 +0200
783 +#include "../../../../include/generated/asm-offsets.h"
784 diff -NurpP --minimal linux-2.6.36/arch/um/include/shared/user_constants.h linux-2.6.36-vs2.3.0.36.33/arch/um/include/shared/user_constants.h
785 --- linux-2.6.36/arch/um/include/shared/user_constants.h 1970-01-01 01:00:00.000000000 +0100
786 +++ linux-2.6.36-vs2.3.0.36.33/arch/um/include/shared/user_constants.h 2010-10-21 13:09:36.000000000 +0200
791 + * This file was generated by arch/um/Makefile
795 +#define HOST_SC_CR2 176 /* offsetof(struct sigcontext, cr2) # */
796 +#define HOST_SC_ERR 152 /* offsetof(struct sigcontext, err) # */
797 +#define HOST_SC_TRAPNO 160 /* offsetof(struct sigcontext, trapno) # */
798 +#define HOST_FP_SIZE 64 /* sizeof(struct _fpstate) / sizeof(unsigned long) # */
799 +#define HOST_RBX 5 /* RBX # */
800 +#define HOST_RCX 11 /* RCX # */
801 +#define HOST_RDI 14 /* RDI # */
802 +#define HOST_RSI 13 /* RSI # */
803 +#define HOST_RDX 12 /* RDX # */
804 +#define HOST_RBP 4 /* RBP # */
805 +#define HOST_RAX 10 /* RAX # */
806 +#define HOST_R8 9 /* R8 # */
807 +#define HOST_R9 8 /* R9 # */
808 +#define HOST_R10 7 /* R10 # */
809 +#define HOST_R11 6 /* R11 # */
810 +#define HOST_R12 3 /* R12 # */
811 +#define HOST_R13 2 /* R13 # */
812 +#define HOST_R14 1 /* R14 # */
813 +#define HOST_R15 0 /* R15 # */
814 +#define HOST_ORIG_RAX 15 /* ORIG_RAX # */
815 +#define HOST_CS 17 /* CS # */
816 +#define HOST_SS 20 /* SS # */
817 +#define HOST_EFLAGS 18 /* EFLAGS # */
818 +#define HOST_IP 16 /* RIP # */
819 +#define HOST_SP 19 /* RSP # */
820 +#define UM_FRAME_SIZE 216 /* sizeof(struct user_regs_struct) # */
821 +#define UM_POLLIN 1 /* POLLIN # */
822 +#define UM_POLLPRI 2 /* POLLPRI # */
823 +#define UM_POLLOUT 4 /* POLLOUT # */
824 +#define UM_PROT_READ 1 /* PROT_READ # */
825 +#define UM_PROT_WRITE 2 /* PROT_WRITE # */
826 +#define UM_PROT_EXEC 4 /* PROT_EXEC # */
828 diff -NurpP --minimal linux-2.6.36/arch/um/Kconfig.rest linux-2.6.36-vs2.3.0.36.33/arch/um/Kconfig.rest
829 --- linux-2.6.36/arch/um/Kconfig.rest 2009-06-11 17:12:19.000000000 +0200
830 +++ linux-2.6.36-vs2.3.0.36.33/arch/um/Kconfig.rest 2010-10-21 13:09:36.000000000 +0200
831 @@ -18,6 +18,8 @@ source "drivers/connector/Kconfig"
835 +source "kernel/vserver/Kconfig"
837 source "security/Kconfig"
839 source "crypto/Kconfig"
840 diff -NurpP --minimal linux-2.6.36/arch/x86/ia32/ia32entry.S linux-2.6.36-vs2.3.0.36.33/arch/x86/ia32/ia32entry.S
841 --- linux-2.6.36/arch/x86/ia32/ia32entry.S 2010-10-21 13:06:59.000000000 +0200
842 +++ linux-2.6.36-vs2.3.0.36.33/arch/x86/ia32/ia32entry.S 2010-10-21 13:09:36.000000000 +0200
843 @@ -783,7 +783,7 @@ ia32_sys_call_table:
844 .quad sys_tgkill /* 270 */
845 .quad compat_sys_utimes
846 .quad sys32_fadvise64_64
847 - .quad quiet_ni_syscall /* sys_vserver */
848 + .quad sys32_vserver
850 .quad compat_sys_get_mempolicy /* 275 */
851 .quad sys_set_mempolicy
852 diff -NurpP --minimal linux-2.6.36/arch/x86/include/asm/unistd_64.h linux-2.6.36-vs2.3.0.36.33/arch/x86/include/asm/unistd_64.h
853 --- linux-2.6.36/arch/x86/include/asm/unistd_64.h 2010-10-21 13:06:59.000000000 +0200
854 +++ linux-2.6.36-vs2.3.0.36.33/arch/x86/include/asm/unistd_64.h 2010-10-21 13:09:36.000000000 +0200
855 @@ -535,7 +535,7 @@ __SYSCALL(__NR_tgkill, sys_tgkill)
856 #define __NR_utimes 235
857 __SYSCALL(__NR_utimes, sys_utimes)
858 #define __NR_vserver 236
859 -__SYSCALL(__NR_vserver, sys_ni_syscall)
860 +__SYSCALL(__NR_vserver, sys_vserver)
861 #define __NR_mbind 237
862 __SYSCALL(__NR_mbind, sys_mbind)
863 #define __NR_set_mempolicy 238
864 diff -NurpP --minimal linux-2.6.36/arch/x86/Kconfig linux-2.6.36-vs2.3.0.36.33/arch/x86/Kconfig
865 --- linux-2.6.36/arch/x86/Kconfig 2010-10-21 13:06:59.000000000 +0200
866 +++ linux-2.6.36-vs2.3.0.36.33/arch/x86/Kconfig 2010-10-21 13:09:36.000000000 +0200
867 @@ -2135,6 +2135,8 @@ source "fs/Kconfig"
869 source "arch/x86/Kconfig.debug"
871 +source "kernel/vserver/Kconfig"
873 source "security/Kconfig"
875 source "crypto/Kconfig"
876 diff -NurpP --minimal linux-2.6.36/arch/x86/kernel/syscall_table_32.S linux-2.6.36-vs2.3.0.36.33/arch/x86/kernel/syscall_table_32.S
877 --- linux-2.6.36/arch/x86/kernel/syscall_table_32.S 2010-10-21 13:07:00.000000000 +0200
878 +++ linux-2.6.36-vs2.3.0.36.33/arch/x86/kernel/syscall_table_32.S 2010-10-21 13:09:36.000000000 +0200
879 @@ -272,7 +272,7 @@ ENTRY(sys_call_table)
880 .long sys_tgkill /* 270 */
882 .long sys_fadvise64_64
883 - .long sys_ni_syscall /* sys_vserver */
886 .long sys_get_mempolicy
887 .long sys_set_mempolicy
888 diff -NurpP --minimal linux-2.6.36/Documentation/vserver/debug.txt linux-2.6.36-vs2.3.0.36.33/Documentation/vserver/debug.txt
889 --- linux-2.6.36/Documentation/vserver/debug.txt 1970-01-01 01:00:00.000000000 +0100
890 +++ linux-2.6.36-vs2.3.0.36.33/Documentation/vserver/debug.txt 2010-10-21 13:09:36.000000000 +0200
895 + 2 4 "vx_map_tgid: %p/%llx: %d -> %d"
896 + "vx_rmap_tgid: %p/%llx: %d -> %d"
900 + 0 1 "ALLOC (%p,#%d)%c inode (%d)"
901 + "FREE (%p,#%d)%c inode"
902 + 1 2 "ALLOC (%p,#%d)%c %lld bytes (%d)"
903 + "FREE (%p,#%d)%c %lld bytes"
904 + 2 4 "ADJUST: %lld,%lld on %ld,%ld [mult=%d]"
905 + 3 8 "ext3_has_free_blocks(%p): %lu<%lu+1, %c, %u!=%u r=%d"
906 + "ext3_has_free_blocks(%p): free=%lu, root=%lu"
907 + "rcu_free_dl_info(%p)"
908 + 4 10 "alloc_dl_info(%p,%d) = %p"
909 + "dealloc_dl_info(%p)"
910 + "get_dl_info(%p[#%d.%d])"
911 + "put_dl_info(%p[#%d.%d])"
912 + 5 20 "alloc_dl_info(%p,%d)*"
913 + 6 40 "__hash_dl_info: %p[#%d]"
914 + "__unhash_dl_info: %p[#%d]"
915 + 7 80 "locate_dl_info(%p,#%d) = %p"
919 + 0 1 "destroy_dqhash: %p [#0x%08x] c=%d"
920 + "new_dqhash: %p [#0x%08x]"
921 + "vroot[%d]_clr_dev: dev=%p[%lu,%d:%d]"
922 + "vroot[%d]_get_real_bdev: dev=%p[%lu,%d:%d]"
923 + "vroot[%d]_set_dev: dev=%p[%lu,%d:%d]"
924 + "vroot_get_real_bdev not set"
925 + 1 2 "cow_break_link(»%s«)"
927 + 2 4 "dentry_open(new): %p"
928 + "dentry_open(old): %p"
929 + "lookup_create(new): %p"
931 + "path_lookup(old): %d"
932 + "vfs_create(new): %d"
935 + 3 8 "fput(new_file=%p[#%d])"
936 + "fput(old_file=%p[#%d])"
937 + 4 10 "vx_info_kill(%p[#%d],%d,%d) = %d"
938 + "vx_info_kill(%p[#%d],%d,%d)*"
939 + 5 20 "vs_reboot(%p[#%d],%d)"
940 + 6 40 "dropping task %p[#%u,%u] for %p[#%u,%u]"
944 + 2 4 "nx_addr_conflict(%p,%p) %d.%d,%d.%d"
945 + 3 8 "inet_bind(%p) %d.%d.%d.%d, %d.%d.%d.%d, %d.%d.%d.%d"
946 + "inet_bind(%p)* %p,%p;%lx %d.%d.%d.%d"
947 + 4 10 "ip_route_connect(%p) %p,%p;%lx"
948 + 5 20 "__addr_in_socket(%p,%d.%d.%d.%d) %p:%d.%d.%d.%d %p;%lx"
949 + 6 40 "sk,egf: %p [#%d] (from %d)"
950 + "sk,egn: %p [#%d] (from %d)"
951 + "sk,req: %p [#%d] (from %d)"
952 + "sk: %p [#%d] (from %d)"
953 + "tw: %p [#%d] (from %d)"
954 + 7 80 "__sock_recvmsg: %p[%p,%p,%p;%d]:%d/%d"
955 + "__sock_sendmsg: %p[%p,%p,%p;%d]:%d/%d"
959 + 0 1 "__lookup_nx_info(#%u): %p[#%u]"
960 + "alloc_nx_info(%d) = %p"
961 + "create_nx_info(%d) (dynamic rejected)"
962 + "create_nx_info(%d) = %p (already there)"
963 + "create_nx_info(%d) = %p (new)"
964 + "dealloc_nx_info(%p)"
965 + 1 2 "alloc_nx_info(%d)*"
966 + "create_nx_info(%d)*"
967 + 2 4 "get_nx_info(%p[#%d.%d])"
968 + "put_nx_info(%p[#%d.%d])"
969 + 3 8 "claim_nx_info(%p[#%d.%d.%d]) %p"
970 + "clr_nx_info(%p[#%d.%d])"
971 + "init_nx_info(%p[#%d.%d])"
972 + "release_nx_info(%p[#%d.%d.%d]) %p"
973 + "set_nx_info(%p[#%d.%d])"
974 + 4 10 "__hash_nx_info: %p[#%d]"
975 + "__nx_dynamic_id: [#%d]"
976 + "__unhash_nx_info: %p[#%d.%d.%d]"
977 + 5 20 "moved task %p into nxi:%p[#%d]"
978 + "nx_migrate_task(%p,%p[#%d.%d.%d])"
979 + "task_get_nx_info(%p)"
980 + 6 40 "nx_clear_persistent(%p[#%d])"
984 + 0 1 "quota_sync_dqh(%p,%d) discard inode %p"
985 + 1 2 "quota_sync_dqh(%p,%d)"
986 + "sync_dquots(%p,%d)"
987 + "sync_dquots_dqh(%p,%d)"
988 + 3 8 "do_quotactl(%p,%d,cmd=%d,id=%d,%p)"
992 + 0 1 "vc: VCMD_%02d_%d[%d], %d,%p [%d,%d,%x,%x]"
993 + 1 2 "vc: VCMD_%02d_%d[%d] = %08lx(%ld) [%d,%d]"
994 + 4 10 "%s: (%s %s) returned %s with %d"
998 + 7 80 "dx_parse_tag(»%s«): %d:#%d"
999 + "dx_propagate_tag(%p[#%lu.%d]): %d,%d"
1003 + 0 1 "__lookup_vx_info(#%u): %p[#%u]"
1004 + "alloc_vx_info(%d) = %p"
1005 + "alloc_vx_info(%d)*"
1006 + "create_vx_info(%d) (dynamic rejected)"
1007 + "create_vx_info(%d) = %p (already there)"
1008 + "create_vx_info(%d) = %p (new)"
1009 + "dealloc_vx_info(%p)"
1010 + "loc_vx_info(%d) = %p (found)"
1011 + "loc_vx_info(%d) = %p (new)"
1012 + "loc_vx_info(%d) = %p (not available)"
1013 + 1 2 "create_vx_info(%d)*"
1014 + "loc_vx_info(%d)*"
1015 + 2 4 "get_vx_info(%p[#%d.%d])"
1016 + "put_vx_info(%p[#%d.%d])"
1017 + 3 8 "claim_vx_info(%p[#%d.%d.%d]) %p"
1018 + "clr_vx_info(%p[#%d.%d])"
1019 + "init_vx_info(%p[#%d.%d])"
1020 + "release_vx_info(%p[#%d.%d.%d]) %p"
1021 + "set_vx_info(%p[#%d.%d])"
1022 + 4 10 "__hash_vx_info: %p[#%d]"
1023 + "__unhash_vx_info: %p[#%d.%d.%d]"
1024 + "__vx_dynamic_id: [#%d]"
1025 + 5 20 "enter_vx_info(%p[#%d],%p) %p[#%d,%p]"
1026 + "leave_vx_info(%p[#%d,%p]) %p[#%d,%p]"
1027 + "moved task %p into vxi:%p[#%d]"
1028 + "task_get_vx_info(%p)"
1029 + "vx_migrate_task(%p,%p[#%d.%d])"
1030 + 6 40 "vx_clear_persistent(%p[#%d])"
1031 + "vx_exit_init(%p[#%d],%p[#%d,%d,%d])"
1032 + "vx_set_init(%p[#%d],%p[#%d,%d,%d])"
1033 + "vx_set_persistent(%p[#%d])"
1034 + "vx_set_reaper(%p[#%d],%p[#%d,%d])"
1035 + 7 80 "vx_child_reaper(%p[#%u,%u]) = %p[#%u,%u]"
1040 + n 2^n "vx_acc_cres[%5d,%s,%2d]: %5d%s"
1041 + "vx_cres_avail[%5d,%s,%2d]: %5ld > %5d + %5d"
1043 + m 2^m "vx_acc_page[%5d,%s,%2d]: %5d%s"
1044 + "vx_acc_pages[%5d,%s,%2d]: %5d += %5d"
1045 + "vx_pages_avail[%5d,%s,%2d]: %5ld > %5d + %5d"
1046 diff -NurpP --minimal linux-2.6.36/drivers/block/Kconfig linux-2.6.36-vs2.3.0.36.33/drivers/block/Kconfig
1047 --- linux-2.6.36/drivers/block/Kconfig 2010-08-02 16:52:14.000000000 +0200
1048 +++ linux-2.6.36-vs2.3.0.36.33/drivers/block/Kconfig 2010-10-21 13:09:36.000000000 +0200
1049 @@ -273,6 +273,13 @@ config BLK_DEV_CRYPTOLOOP
1051 source "drivers/block/drbd/Kconfig"
1053 +config BLK_DEV_VROOT
1054 + tristate "Virtual Root device support"
1055 + depends on QUOTACTL
1057 + Saying Y here will allow you to use quota/fs ioctls on a shared
1058 + partition within a virtual server without compromising security.
1061 tristate "Network block device support"
1063 diff -NurpP --minimal linux-2.6.36/drivers/block/loop.c linux-2.6.36-vs2.3.0.36.33/drivers/block/loop.c
1064 --- linux-2.6.36/drivers/block/loop.c 2010-10-21 13:07:02.000000000 +0200
1065 +++ linux-2.6.36-vs2.3.0.36.33/drivers/block/loop.c 2010-10-21 13:11:25.000000000 +0200
1067 #include <linux/highmem.h>
1068 #include <linux/kthread.h>
1069 #include <linux/splice.h>
1070 +#include <linux/vs_context.h>
1072 #include <asm/uaccess.h>
1074 @@ -814,6 +815,7 @@ static int loop_set_fd(struct loop_devic
1075 lo->lo_blocksize = lo_blocksize;
1076 lo->lo_device = bdev;
1077 lo->lo_flags = lo_flags;
1078 + lo->lo_xid = vx_current_xid();
1079 lo->lo_backing_file = file;
1080 lo->transfer = transfer_none;
1082 @@ -942,6 +944,7 @@ static int loop_clr_fd(struct loop_devic
1083 lo->lo_encrypt_key_size = 0;
1085 lo->lo_thread = NULL;
1087 memset(lo->lo_encrypt_key, 0, LO_KEY_SIZE);
1088 memset(lo->lo_crypt_name, 0, LO_NAME_SIZE);
1089 memset(lo->lo_file_name, 0, LO_NAME_SIZE);
1090 @@ -979,7 +982,7 @@ loop_set_status(struct loop_device *lo,
1092 if (lo->lo_encrypt_key_size &&
1093 lo->lo_key_owner != uid &&
1094 - !capable(CAP_SYS_ADMIN))
1095 + !vx_capable(CAP_SYS_ADMIN, VXC_ADMIN_CLOOP))
1097 if (lo->lo_state != Lo_bound)
1099 @@ -1063,7 +1066,8 @@ loop_get_status(struct loop_device *lo,
1100 memcpy(info->lo_crypt_name, lo->lo_crypt_name, LO_NAME_SIZE);
1101 info->lo_encrypt_type =
1102 lo->lo_encryption ? lo->lo_encryption->number : 0;
1103 - if (lo->lo_encrypt_key_size && capable(CAP_SYS_ADMIN)) {
1104 + if (lo->lo_encrypt_key_size &&
1105 + vx_capable(CAP_SYS_ADMIN, VXC_ADMIN_CLOOP)) {
1106 info->lo_encrypt_key_size = lo->lo_encrypt_key_size;
1107 memcpy(info->lo_encrypt_key, lo->lo_encrypt_key,
1108 lo->lo_encrypt_key_size);
1109 @@ -1409,6 +1413,9 @@ static int lo_open(struct block_device *
1111 struct loop_device *lo = bdev->bd_disk->private_data;
1113 + if (!vx_check(lo->lo_xid, VS_IDENT|VS_HOSTID|VS_ADMIN_P))
1117 mutex_lock(&lo->lo_ctl_mutex);
1119 diff -NurpP --minimal linux-2.6.36/drivers/block/Makefile linux-2.6.36-vs2.3.0.36.33/drivers/block/Makefile
1120 --- linux-2.6.36/drivers/block/Makefile 2010-02-25 11:51:36.000000000 +0100
1121 +++ linux-2.6.36-vs2.3.0.36.33/drivers/block/Makefile 2010-10-21 13:09:36.000000000 +0200
1122 @@ -34,6 +34,7 @@ obj-$(CONFIG_VIODASD) += viodasd.o
1123 obj-$(CONFIG_BLK_DEV_SX8) += sx8.o
1124 obj-$(CONFIG_BLK_DEV_UB) += ub.o
1125 obj-$(CONFIG_BLK_DEV_HD) += hd.o
1126 +obj-$(CONFIG_BLK_DEV_VROOT) += vroot.o
1128 obj-$(CONFIG_XEN_BLKDEV_FRONTEND) += xen-blkfront.o
1129 obj-$(CONFIG_BLK_DEV_DRBD) += drbd/
1130 diff -NurpP --minimal linux-2.6.36/drivers/block/vroot.c linux-2.6.36-vs2.3.0.36.33/drivers/block/vroot.c
1131 --- linux-2.6.36/drivers/block/vroot.c 1970-01-01 01:00:00.000000000 +0100
1132 +++ linux-2.6.36-vs2.3.0.36.33/drivers/block/vroot.c 2010-10-21 13:09:36.000000000 +0200
1135 + * linux/drivers/block/vroot.c
1137 + * written by Herbert Pötzl, 9/11/2002
1138 + * ported to 2.6.10 by Herbert Pötzl, 30/12/2004
1140 + * based on the loop.c code by Theodore Ts'o.
1142 + * Copyright (C) 2002-2007 by Herbert Pötzl.
1143 + * Redistribution of this file is permitted under the
1144 + * GNU General Public License.
1148 +#include <linux/module.h>
1149 +#include <linux/moduleparam.h>
1150 +#include <linux/file.h>
1151 +#include <linux/major.h>
1152 +#include <linux/blkdev.h>
1153 +#include <linux/slab.h>
1155 +#include <linux/vroot.h>
1156 +#include <linux/vs_context.h>
1159 +static int max_vroot = 8;
1161 +static struct vroot_device *vroot_dev;
1162 +static struct gendisk **disks;
1165 +static int vroot_set_dev(
1166 + struct vroot_device *vr,
1167 + struct block_device *bdev,
1170 + struct block_device *real_bdev;
1171 + struct file *file;
1172 + struct inode *inode;
1176 + if (vr->vr_state != Vr_unbound)
1185 + inode = file->f_dentry->d_inode;
1188 + if (S_ISBLK(inode->i_mode)) {
1189 + real_bdev = inode->i_bdev;
1190 + vr->vr_device = real_bdev;
1191 + __iget(real_bdev->bd_inode);
1195 + vxdprintk(VXD_CBIT(misc, 0),
1196 + "vroot[%d]_set_dev: dev=" VXF_DEV,
1197 + vr->vr_number, VXD_DEV(real_bdev));
1199 + vr->vr_state = Vr_bound;
1208 +static int vroot_clr_dev(
1209 + struct vroot_device *vr,
1210 + struct block_device *bdev)
1212 + struct block_device *real_bdev;
1214 + if (vr->vr_state != Vr_bound)
1216 + if (vr->vr_refcnt > 1) /* we needed one fd for the ioctl */
1219 + real_bdev = vr->vr_device;
1221 + vxdprintk(VXD_CBIT(misc, 0),
1222 + "vroot[%d]_clr_dev: dev=" VXF_DEV,
1223 + vr->vr_number, VXD_DEV(real_bdev));
1226 + vr->vr_state = Vr_unbound;
1227 + vr->vr_device = NULL;
1232 +static int vr_ioctl(struct block_device *bdev, fmode_t mode,
1233 + unsigned int cmd, unsigned long arg)
1235 + struct vroot_device *vr = bdev->bd_disk->private_data;
1238 + down(&vr->vr_ctl_mutex);
1240 + case VROOT_SET_DEV:
1241 + err = vroot_set_dev(vr, bdev, arg);
1243 + case VROOT_CLR_DEV:
1244 + err = vroot_clr_dev(vr, bdev);
1250 + up(&vr->vr_ctl_mutex);
1254 +static int vr_open(struct block_device *bdev, fmode_t mode)
1256 + struct vroot_device *vr = bdev->bd_disk->private_data;
1258 + down(&vr->vr_ctl_mutex);
1260 + up(&vr->vr_ctl_mutex);
1264 +static int vr_release(struct gendisk *disk, fmode_t mode)
1266 + struct vroot_device *vr = disk->private_data;
1268 + down(&vr->vr_ctl_mutex);
1270 + up(&vr->vr_ctl_mutex);
1274 +static struct block_device_operations vr_fops = {
1275 + .owner = THIS_MODULE,
1277 + .release = vr_release,
1278 + .ioctl = vr_ioctl,
1281 +struct block_device *__vroot_get_real_bdev(struct block_device *bdev)
1283 + struct inode *inode = bdev->bd_inode;
1284 + struct vroot_device *vr;
1285 + struct block_device *real_bdev;
1286 + int minor = iminor(inode);
1288 + vr = &vroot_dev[minor];
1289 + real_bdev = vr->vr_device;
1291 + vxdprintk(VXD_CBIT(misc, 0),
1292 + "vroot[%d]_get_real_bdev: dev=" VXF_DEV,
1293 + vr->vr_number, VXD_DEV(real_bdev));
1295 + if (vr->vr_state != Vr_bound)
1296 + return ERR_PTR(-ENXIO);
1298 + __iget(real_bdev->bd_inode);
1303 + * And now the modules code and kernel interface.
1306 +module_param(max_vroot, int, 0);
1308 +MODULE_PARM_DESC(max_vroot, "Maximum number of vroot devices (1-256)");
1309 +MODULE_LICENSE("GPL");
1310 +MODULE_ALIAS_BLOCKDEV_MAJOR(VROOT_MAJOR);
1312 +MODULE_AUTHOR ("Herbert Pötzl");
1313 +MODULE_DESCRIPTION ("Virtual Root Device Mapper");
1316 +int __init vroot_init(void)
1320 + if (max_vroot < 1 || max_vroot > 256) {
1321 + max_vroot = MAX_VROOT_DEFAULT;
1322 + printk(KERN_WARNING "vroot: invalid max_vroot "
1323 + "(must be between 1 and 256), "
1324 + "using default (%d)\n", max_vroot);
1327 + if (register_blkdev(VROOT_MAJOR, "vroot"))
1331 + vroot_dev = kmalloc(max_vroot * sizeof(struct vroot_device), GFP_KERNEL);
1334 + memset(vroot_dev, 0, max_vroot * sizeof(struct vroot_device));
1336 + disks = kmalloc(max_vroot * sizeof(struct gendisk *), GFP_KERNEL);
1340 + for (i = 0; i < max_vroot; i++) {
1341 + disks[i] = alloc_disk(1);
1344 + disks[i]->queue = blk_alloc_queue(GFP_KERNEL);
1345 + if (!disks[i]->queue)
1349 + for (i = 0; i < max_vroot; i++) {
1350 + struct vroot_device *vr = &vroot_dev[i];
1351 + struct gendisk *disk = disks[i];
1353 + memset(vr, 0, sizeof(*vr));
1354 + init_MUTEX(&vr->vr_ctl_mutex);
1355 + vr->vr_number = i;
1356 + disk->major = VROOT_MAJOR;
1357 + disk->first_minor = i;
1358 + disk->fops = &vr_fops;
1359 + sprintf(disk->disk_name, "vroot%d", i);
1360 + disk->private_data = vr;
1363 + err = register_vroot_grb(&__vroot_get_real_bdev);
1367 + for (i = 0; i < max_vroot; i++)
1368 + add_disk(disks[i]);
1369 + printk(KERN_INFO "vroot: loaded (max %d devices)\n", max_vroot);
1374 + put_disk(disks[i]);
1379 + unregister_blkdev(VROOT_MAJOR, "vroot");
1380 + printk(KERN_ERR "vroot: ran out of memory\n");
1384 +void vroot_exit(void)
1388 + if (unregister_vroot_grb(&__vroot_get_real_bdev))
1389 + printk(KERN_WARNING "vroot: cannot unregister grb\n");
1391 + for (i = 0; i < max_vroot; i++) {
1392 + del_gendisk(disks[i]);
1393 + put_disk(disks[i]);
1395 + unregister_blkdev(VROOT_MAJOR, "vroot");
1401 +module_init(vroot_init);
1402 +module_exit(vroot_exit);
1406 +static int __init max_vroot_setup(char *str)
1408 + max_vroot = simple_strtol(str, NULL, 0);
1412 +__setup("max_vroot=", max_vroot_setup);
1416 diff -NurpP --minimal linux-2.6.36/drivers/char/sysrq.c linux-2.6.36-vs2.3.0.36.33/drivers/char/sysrq.c
1417 --- linux-2.6.36/drivers/char/sysrq.c 2010-10-21 13:07:02.000000000 +0200
1418 +++ linux-2.6.36-vs2.3.0.36.33/drivers/char/sysrq.c 2010-10-21 13:09:36.000000000 +0200
1420 #include <linux/oom.h>
1421 #include <linux/slab.h>
1422 #include <linux/input.h>
1423 +#include <linux/vserver/debug.h>
1425 #include <asm/ptrace.h>
1426 #include <asm/irq_regs.h>
1427 @@ -395,6 +396,21 @@ static struct sysrq_key_op sysrq_unrt_op
1428 .enable_mask = SYSRQ_ENABLE_RTNICE,
1432 +#ifdef CONFIG_VSERVER_DEBUG
1433 +static void sysrq_handle_vxinfo(int key, struct tty_struct *tty)
1435 + dump_vx_info_inactive((key == 'x')?0:1);
1438 +static struct sysrq_key_op sysrq_showvxinfo_op = {
1439 + .handler = sysrq_handle_vxinfo,
1440 + .help_msg = "conteXt",
1441 + .action_msg = "Show Context Info",
1442 + .enable_mask = SYSRQ_ENABLE_DUMP,
1446 /* Key Operations table and lock */
1447 static DEFINE_SPINLOCK(sysrq_key_table_lock);
1449 @@ -449,7 +465,11 @@ static struct sysrq_key_op *sysrq_key_ta
1451 &sysrq_showstate_blocked_op, /* w */
1452 /* x: May be registered on ppc/powerpc for xmon */
1453 +#ifdef CONFIG_VSERVER_DEBUG
1454 + &sysrq_showvxinfo_op, /* x */
1458 /* y: May be registered on sparc64 for global register dump */
1460 &sysrq_ftrace_dump_op, /* z */
1461 @@ -464,6 +484,8 @@ static int sysrq_key_table_key2index(int
1463 else if ((key >= 'a') && (key <= 'z'))
1464 retval = key + 10 - 'a';
1465 + else if ((key >= 'A') && (key <= 'Z'))
1466 + retval = key + 10 - 'A';
1470 diff -NurpP --minimal linux-2.6.36/drivers/char/tty_io.c linux-2.6.36-vs2.3.0.36.33/drivers/char/tty_io.c
1471 --- linux-2.6.36/drivers/char/tty_io.c 2010-10-21 13:07:02.000000000 +0200
1472 +++ linux-2.6.36-vs2.3.0.36.33/drivers/char/tty_io.c 2010-10-21 13:09:36.000000000 +0200
1475 #include <linux/kmod.h>
1476 #include <linux/nsproxy.h>
1477 +#include <linux/vs_pid.h>
1479 #undef TTY_DEBUG_HANGUP
1481 @@ -2038,7 +2039,8 @@ static int tiocsti(struct tty_struct *tt
1483 struct tty_ldisc *ld;
1485 - if ((current->signal->tty != tty) && !capable(CAP_SYS_ADMIN))
1486 + if (((current->signal->tty != tty) &&
1487 + !vx_capable(CAP_SYS_ADMIN, VXC_TIOCSTI)))
1489 if (get_user(ch, p))
1491 @@ -2326,6 +2328,7 @@ static int tiocspgrp(struct tty_struct *
1493 if (get_user(pgrp_nr, p))
1495 + pgrp_nr = vx_rmap_pid(pgrp_nr);
1499 diff -NurpP --minimal linux-2.6.36/drivers/gpu/drm/radeon/r100_reg_safe.h linux-2.6.36-vs2.3.0.36.33/drivers/gpu/drm/radeon/r100_reg_safe.h
1500 --- linux-2.6.36/drivers/gpu/drm/radeon/r100_reg_safe.h 1970-01-01 01:00:00.000000000 +0100
1501 +++ linux-2.6.36-vs2.3.0.36.33/drivers/gpu/drm/radeon/r100_reg_safe.h 2010-10-21 13:09:36.000000000 +0200
1503 +static const unsigned r100_reg_safe_bm[102] = {
1504 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1505 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1506 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1507 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1508 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1509 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1510 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1511 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1512 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1513 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1514 + 0x17FF1FFF, 0xFFFFFFFC, 0xFFFFFFFF, 0xFF30FFBF,
1515 + 0xFFFFFFF8, 0xC3E6FFFF, 0xFFFFF6DF, 0xFFFFFFFF,
1516 + 0xFFFFFFCF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1517 + 0xFFFFFF9F, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1518 + 0x38E7FE1F, 0xFFC3FF8E, 0x7FF8FFFF, 0xFFFF803C,
1519 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1520 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFEFFFF, 0xFFFFFFFF,
1521 + 0x00000000, 0xFFFFFFFD, 0xFFFFFFFF, 0xFFFFFFFF,
1522 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1523 + 0xFFFFFFFF, 0xFFFCFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1524 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1525 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1526 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1527 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1528 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1529 + 0xFFFFFFFF, 0xFFFFFFEF,
1531 diff -NurpP --minimal linux-2.6.36/drivers/gpu/drm/radeon/r200_reg_safe.h linux-2.6.36-vs2.3.0.36.33/drivers/gpu/drm/radeon/r200_reg_safe.h
1532 --- linux-2.6.36/drivers/gpu/drm/radeon/r200_reg_safe.h 1970-01-01 01:00:00.000000000 +0100
1533 +++ linux-2.6.36-vs2.3.0.36.33/drivers/gpu/drm/radeon/r200_reg_safe.h 2010-10-21 13:09:36.000000000 +0200
1535 +static const unsigned r200_reg_safe_bm[102] = {
1536 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1537 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1538 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1539 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
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 + 0x17FF1FFF, 0xFFFFFFFC, 0xFFFFFFFF, 0xFF30FFBF,
1547 + 0xFFFFFFF8, 0xC3E6FFFF, 0xFFFFF6DF, 0xFFFFFFFF,
1548 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1549 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1550 + 0xFFE7FE1F, 0xF003FFFF, 0x7EFFFFFF, 0xFFFF803C,
1551 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1552 + 0xFFFFFFFF, 0xFFFFEFCE, 0xFFFEFFFF, 0xFFFFFFFE,
1553 + 0x020E0FF0, 0xFFCC83FD, 0xFFFFFFFF, 0xFFFFFFFF,
1554 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1555 + 0xFFFBFFFF, 0xEFFCFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1556 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1557 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1558 + 0xDFDFDFDF, 0x3FFDDFDF, 0xFFFFFFFF, 0xFFFFFF7F,
1559 + 0xFFFFFFFF, 0x00FFFFFF, 0x00000000, 0x00000000,
1560 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1561 + 0xFFFFFE3F, 0xFFFFFFEF,
1563 diff -NurpP --minimal linux-2.6.36/drivers/gpu/drm/radeon/r300_reg_safe.h linux-2.6.36-vs2.3.0.36.33/drivers/gpu/drm/radeon/r300_reg_safe.h
1564 --- linux-2.6.36/drivers/gpu/drm/radeon/r300_reg_safe.h 1970-01-01 01:00:00.000000000 +0100
1565 +++ linux-2.6.36-vs2.3.0.36.33/drivers/gpu/drm/radeon/r300_reg_safe.h 2010-10-21 13:09:36.000000000 +0200
1567 +static const unsigned r300_reg_safe_bm[159] = {
1568 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1569 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1570 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1571 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
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 + 0x17FF1FFF, 0xFFFFFFFC, 0xFFFFFFFF, 0xFF30FFBF,
1579 + 0xFFFFFFF8, 0xC3E6FFFF, 0xFFFFF6DF, 0xFFFFFFFF,
1580 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1581 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1582 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFF03F,
1583 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1584 + 0xFFFFFFFF, 0xFFFFEFCE, 0xF00EBFFF, 0x007C0000,
1585 + 0xF0000078, 0xFF000009, 0xFFFFFFFF, 0xFFFFFFFF,
1586 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1587 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1588 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1589 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1590 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1591 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1592 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1593 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1594 + 0xFFFFF7FF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1595 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1596 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1597 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1598 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1599 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1600 + 0xFFFFFC78, 0xFFFFFFFF, 0xFFFFFFFE, 0xFFFFFFFF,
1601 + 0x38FF8F50, 0xFFF88082, 0xF000000C, 0xFAE009FF,
1602 + 0x0000FFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0x00000000,
1603 + 0x00000000, 0x0000C100, 0x00000000, 0x00000000,
1604 + 0x00000000, 0x00000000, 0x00000000, 0x00000000,
1605 + 0x00000000, 0xFFFF0000, 0xFFFFFFFF, 0xFF80FFFF,
1606 + 0x00000000, 0x00000000, 0x00000000, 0x00000000,
1607 + 0x0003FC01, 0xFFFFFCF8, 0xFF800B19,
1609 diff -NurpP --minimal linux-2.6.36/drivers/gpu/drm/radeon/r420_reg_safe.h linux-2.6.36-vs2.3.0.36.33/drivers/gpu/drm/radeon/r420_reg_safe.h
1610 --- linux-2.6.36/drivers/gpu/drm/radeon/r420_reg_safe.h 1970-01-01 01:00:00.000000000 +0100
1611 +++ linux-2.6.36-vs2.3.0.36.33/drivers/gpu/drm/radeon/r420_reg_safe.h 2010-10-21 13:09:36.000000000 +0200
1613 +static const unsigned r420_reg_safe_bm[159] = {
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 + 0x17FF1FFF, 0xFFFFFFFC, 0xFFFFFFFF, 0xFF30FFBF,
1625 + 0xFFFFFFF8, 0xC3E6FFFF, 0xFFFFF6DF, 0xFFFFFFFF,
1626 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1627 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1628 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFF03F,
1629 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1630 + 0xFFFFFFFF, 0xFFFFEFCE, 0xF00EBFFF, 0x007C0000,
1631 + 0xF0000078, 0xFF000009, 0xFFFFFFFF, 0xFFFFFFFF,
1632 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
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 + 0xFFFFF7FF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1641 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1642 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1643 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1644 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1645 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1646 + 0xFFFFFC78, 0xFFFFFFFF, 0xFFFFFFFE, 0xFFFFFFFF,
1647 + 0x38FF8F50, 0xFFF88082, 0xF000000C, 0xFAE009FF,
1648 + 0x0000FFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0x00000000,
1649 + 0x00000000, 0x00000100, 0x00000000, 0x00000000,
1650 + 0x00000000, 0x00000000, 0x00000000, 0x00000000,
1651 + 0x00000000, 0x00000000, 0x00000000, 0xFF800000,
1652 + 0x00000000, 0x00000000, 0x00000000, 0x00000000,
1653 + 0x0003FC01, 0xFFFFFCF8, 0xFF800B19,
1655 diff -NurpP --minimal linux-2.6.36/drivers/gpu/drm/radeon/rn50_reg_safe.h linux-2.6.36-vs2.3.0.36.33/drivers/gpu/drm/radeon/rn50_reg_safe.h
1656 --- linux-2.6.36/drivers/gpu/drm/radeon/rn50_reg_safe.h 1970-01-01 01:00:00.000000000 +0100
1657 +++ linux-2.6.36-vs2.3.0.36.33/drivers/gpu/drm/radeon/rn50_reg_safe.h 2010-10-21 13:09:36.000000000 +0200
1659 +static const unsigned rn50_reg_safe_bm[102] = {
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 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1666 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1667 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1668 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1669 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1670 + 0x17FF1FFF, 0xFFFFFFFC, 0xFFFFFFFF, 0xFF30FFBF,
1671 + 0xFFFFFFF8, 0xC3E6FFFF, 0xFFFFF6DF, 0xFFFFFFFF,
1672 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
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,
1687 diff -NurpP --minimal linux-2.6.36/drivers/gpu/drm/radeon/rs600_reg_safe.h linux-2.6.36-vs2.3.0.36.33/drivers/gpu/drm/radeon/rs600_reg_safe.h
1688 --- linux-2.6.36/drivers/gpu/drm/radeon/rs600_reg_safe.h 1970-01-01 01:00:00.000000000 +0100
1689 +++ linux-2.6.36-vs2.3.0.36.33/drivers/gpu/drm/radeon/rs600_reg_safe.h 2010-10-21 13:09:36.000000000 +0200
1691 +static const unsigned rs600_reg_safe_bm[219] = {
1692 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1693 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1694 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1695 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1696 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1697 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1698 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1699 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1700 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1701 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1702 + 0x17FF1FFF, 0xFFFFFFFC, 0xFFFFFFFF, 0xFF30FFBF,
1703 + 0xFFFFFFF8, 0xC3E6FFFF, 0xFFFFF6DF, 0xFFFFFFFF,
1704 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1705 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1706 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFF03F,
1707 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1708 + 0xFFFFFFFF, 0xFFFFEFCE, 0xF00EBFFF, 0x007C0000,
1709 + 0xF0000078, 0xFF000009, 0xFFFFFFFF, 0xFFFFFFFF,
1710 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1711 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1712 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1713 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1714 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1715 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1716 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1717 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1718 + 0xFFFFF7FF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1719 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1720 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1721 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1722 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1723 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1724 + 0xFFFFFC78, 0xFFFFFFFF, 0xFFFFFFFE, 0xFFFFFFFF,
1725 + 0x38FF8F50, 0xFFF88082, 0xF000000C, 0xFAE009FF,
1726 + 0x0000FFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0x00000000,
1727 + 0x00000000, 0x00000100, 0x00000000, 0x00000000,
1728 + 0x00000000, 0x00000000, 0x00000000, 0x00000000,
1729 + 0x00000000, 0x00000000, 0x00000000, 0xFF800000,
1730 + 0x00000000, 0x00000000, 0x00000000, 0x00000000,
1731 + 0x0003FC01, 0xFFFFFCF8, 0xFF800B19, 0xFFFFFFFF,
1732 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1733 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1734 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1735 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1736 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1737 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1738 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1739 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1740 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1741 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1742 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1743 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1744 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1745 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1746 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1748 diff -NurpP --minimal linux-2.6.36/drivers/gpu/drm/radeon/rv515_reg_safe.h linux-2.6.36-vs2.3.0.36.33/drivers/gpu/drm/radeon/rv515_reg_safe.h
1749 --- linux-2.6.36/drivers/gpu/drm/radeon/rv515_reg_safe.h 1970-01-01 01:00:00.000000000 +0100
1750 +++ linux-2.6.36-vs2.3.0.36.33/drivers/gpu/drm/radeon/rv515_reg_safe.h 2010-10-21 13:09:36.000000000 +0200
1752 +static const unsigned rv515_reg_safe_bm[219] = {
1753 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1754 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1755 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1756 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1757 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1758 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1759 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1760 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1761 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1762 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1763 + 0x17FF1FFF, 0xFFFFFFFC, 0xFFFFFFFF, 0xFF30FFBF,
1764 + 0xFFFFFFF8, 0xC3E6FFFF, 0xFFFFF6DF, 0xFFFFFFFF,
1765 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1766 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1767 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFF03F,
1768 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1769 + 0xFFFFFFFF, 0xFFFFEFCE, 0xF00EBFFF, 0x007C0000,
1770 + 0xF0000038, 0xFF000009, 0xFFFFFFFF, 0xFFFFFFFF,
1771 + 0xFFFFFFFF, 0xFFFFFFFF, 0x00000000, 0xFFFFFFFF,
1772 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1773 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1774 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1775 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1776 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1777 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1778 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1779 + 0xFFFFF7FF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1780 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1781 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1782 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1783 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1784 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1785 + 0x1FFFF878, 0xFFFFE000, 0xFFFFFE1E, 0xFFFFFFFF,
1786 + 0x388F8F50, 0xFFF88082, 0xFF0000FC, 0xFAE009FF,
1787 + 0x0000FFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0x00000000,
1788 + 0xFFFF8CFC, 0xFFFFC1FF, 0xFFFFFFFF, 0xFFFFFFFF,
1789 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1790 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFF80FFFF,
1791 + 0x00000000, 0x00000000, 0x00000000, 0x00000000,
1792 + 0x0003FC01, 0x3FFFFCF8, 0xFF800B19, 0xFFDFFFFF,
1793 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1794 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1795 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1796 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1797 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1798 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1799 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1800 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1801 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1802 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1803 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1804 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1805 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1806 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1807 + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
1809 diff -NurpP --minimal linux-2.6.36/drivers/infiniband/core/addr.c linux-2.6.36-vs2.3.0.36.33/drivers/infiniband/core/addr.c
1810 --- linux-2.6.36/drivers/infiniband/core/addr.c 2010-10-21 13:07:05.000000000 +0200
1811 +++ linux-2.6.36-vs2.3.0.36.33/drivers/infiniband/core/addr.c 2010-10-21 13:09:36.000000000 +0200
1812 @@ -252,7 +252,7 @@ static int addr6_resolve(struct sockaddr
1814 if (ipv6_addr_any(&fl.fl6_src)) {
1815 ret = ipv6_dev_get_saddr(&init_net, ip6_dst_idev(dst)->dev,
1816 - &fl.fl6_dst, 0, &fl.fl6_src);
1817 + &fl.fl6_dst, 0, &fl.fl6_src, NULL);
1821 diff -NurpP --minimal linux-2.6.36/drivers/infiniband/hw/ipath/ipath_user_pages.c linux-2.6.36-vs2.3.0.36.33/drivers/infiniband/hw/ipath/ipath_user_pages.c
1822 --- linux-2.6.36/drivers/infiniband/hw/ipath/ipath_user_pages.c 2010-07-07 18:31:19.000000000 +0200
1823 +++ linux-2.6.36-vs2.3.0.36.33/drivers/infiniband/hw/ipath/ipath_user_pages.c 2010-10-21 13:09:36.000000000 +0200
1825 #include <linux/device.h>
1826 #include <linux/slab.h>
1827 #include <linux/sched.h>
1828 +#include <linux/vs_memory.h>
1830 #include "ipath_kernel.h"
1832 diff -NurpP --minimal linux-2.6.36/drivers/md/dm.c linux-2.6.36-vs2.3.0.36.33/drivers/md/dm.c
1833 --- linux-2.6.36/drivers/md/dm.c 2010-10-21 13:07:07.000000000 +0200
1834 +++ linux-2.6.36-vs2.3.0.36.33/drivers/md/dm.c 2010-10-21 15:19:30.000000000 +0200
1836 #include <linux/idr.h>
1837 #include <linux/hdreg.h>
1838 #include <linux/delay.h>
1839 +#include <linux/vs_base.h>
1841 #include <trace/events/block.h>
1843 @@ -121,6 +122,7 @@ struct mapped_device {
1846 atomic_t open_count;
1849 unsigned long flags;
1851 @@ -343,6 +345,7 @@ int dm_deleting_md(struct mapped_device
1852 static int dm_blk_open(struct block_device *bdev, fmode_t mode)
1854 struct mapped_device *md;
1858 spin_lock(&_minor_lock);
1859 @@ -352,19 +355,20 @@ static int dm_blk_open(struct block_devi
1862 if (test_bit(DMF_FREEING, &md->flags) ||
1863 - dm_deleting_md(md)) {
1865 + dm_deleting_md(md))
1869 + if (!vx_check(md->xid, VS_IDENT|VS_HOSTID))
1874 atomic_inc(&md->open_count);
1878 spin_unlock(&_minor_lock);
1881 - return md ? 0 : -ENXIO;
1885 static int dm_blk_close(struct gendisk *disk, fmode_t mode)
1886 @@ -586,6 +590,14 @@ int dm_set_geometry(struct mapped_device
1891 + * Get the xid associated with a dm device
1893 +xid_t dm_get_xid(struct mapped_device *md)
1898 /*-----------------------------------------------------------------
1900 * A more elegant soln is in the works that uses the queue
1901 @@ -1981,6 +1993,7 @@ static struct mapped_device *alloc_dev(i
1902 INIT_LIST_HEAD(&md->uevent_list);
1903 spin_lock_init(&md->uevent_lock);
1905 + md->xid = vx_current_xid();
1906 md->queue = blk_alloc_queue(GFP_KERNEL);
1909 diff -NurpP --minimal linux-2.6.36/drivers/md/dm.h linux-2.6.36-vs2.3.0.36.33/drivers/md/dm.h
1910 --- linux-2.6.36/drivers/md/dm.h 2010-10-21 13:07:07.000000000 +0200
1911 +++ linux-2.6.36-vs2.3.0.36.33/drivers/md/dm.h 2010-10-21 13:09:36.000000000 +0200
1912 @@ -41,6 +41,8 @@ struct dm_dev_internal {
1914 struct dm_md_mempools;
1916 +xid_t dm_get_xid(struct mapped_device *md);
1918 /*-----------------------------------------------------------------
1919 * Internal table functions.
1920 *---------------------------------------------------------------*/
1921 diff -NurpP --minimal linux-2.6.36/drivers/md/dm-ioctl.c linux-2.6.36-vs2.3.0.36.33/drivers/md/dm-ioctl.c
1922 --- linux-2.6.36/drivers/md/dm-ioctl.c 2010-10-21 13:07:07.000000000 +0200
1923 +++ linux-2.6.36-vs2.3.0.36.33/drivers/md/dm-ioctl.c 2010-10-21 13:09:36.000000000 +0200
1925 #include <linux/dm-ioctl.h>
1926 #include <linux/hdreg.h>
1927 #include <linux/compat.h>
1928 +#include <linux/vs_context.h>
1930 #include <asm/uaccess.h>
1932 @@ -106,7 +107,8 @@ static struct hash_cell *__get_name_cell
1933 unsigned int h = hash_str(str);
1935 list_for_each_entry (hc, _name_buckets + h, name_list)
1936 - if (!strcmp(hc->name, str)) {
1937 + if (vx_check(dm_get_xid(hc->md), VS_WATCH_P | VS_IDENT) &&
1938 + !strcmp(hc->name, str)) {
1942 @@ -120,7 +122,8 @@ static struct hash_cell *__get_uuid_cell
1943 unsigned int h = hash_str(str);
1945 list_for_each_entry (hc, _uuid_buckets + h, uuid_list)
1946 - if (!strcmp(hc->uuid, str)) {
1947 + if (vx_check(dm_get_xid(hc->md), VS_WATCH_P | VS_IDENT) &&
1948 + !strcmp(hc->uuid, str)) {
1952 @@ -377,6 +380,9 @@ typedef int (*ioctl_fn)(struct dm_ioctl
1954 static int remove_all(struct dm_ioctl *param, size_t param_size)
1956 + if (!vx_check(0, VS_ADMIN))
1959 dm_hash_remove_all(1);
1960 param->data_size = 0;
1962 @@ -424,6 +430,8 @@ static int list_devices(struct dm_ioctl
1964 for (i = 0; i < NUM_BUCKETS; i++) {
1965 list_for_each_entry (hc, _name_buckets + i, name_list) {
1966 + if (!vx_check(dm_get_xid(hc->md), VS_WATCH_P | VS_IDENT))
1968 needed += sizeof(struct dm_name_list);
1969 needed += strlen(hc->name) + 1;
1970 needed += ALIGN_MASK;
1971 @@ -447,6 +455,8 @@ static int list_devices(struct dm_ioctl
1973 for (i = 0; i < NUM_BUCKETS; i++) {
1974 list_for_each_entry (hc, _name_buckets + i, name_list) {
1975 + if (!vx_check(dm_get_xid(hc->md), VS_WATCH_P | VS_IDENT))
1978 old_nl->next = (uint32_t) ((void *) nl -
1980 @@ -681,10 +691,11 @@ static struct hash_cell *__find_device_h
1984 - mdptr = dm_get_mdptr(md);
1985 + if (vx_check(dm_get_xid(md), VS_WATCH_P | VS_IDENT))
1986 + mdptr = dm_get_mdptr(md);
1994 @@ -1513,8 +1524,8 @@ static int ctl_ioctl(uint command, struc
1998 - /* only root can play with this */
1999 - if (!capable(CAP_SYS_ADMIN))
2000 + /* only root and certain contexts can play with this */
2001 + if (!vx_capable(CAP_SYS_ADMIN, VXC_ADMIN_MAPPER))
2004 if (_IOC_TYPE(command) != DM_IOCTL)
2005 diff -NurpP --minimal linux-2.6.36/drivers/net/tun.c linux-2.6.36-vs2.3.0.36.33/drivers/net/tun.c
2006 --- linux-2.6.36/drivers/net/tun.c 2010-10-21 13:07:15.000000000 +0200
2007 +++ linux-2.6.36-vs2.3.0.36.33/drivers/net/tun.c 2010-10-21 13:09:36.000000000 +0200
2009 #include <linux/nsproxy.h>
2010 #include <linux/virtio_net.h>
2011 #include <linux/rcupdate.h>
2012 +#include <linux/vs_network.h>
2013 #include <net/net_namespace.h>
2014 #include <net/netns/generic.h>
2015 #include <net/rtnetlink.h>
2016 @@ -103,6 +104,7 @@ struct tun_struct {
2022 struct net_device *dev;
2023 struct fasync_struct *fasync;
2024 @@ -853,6 +855,7 @@ static void tun_setup(struct net_device
2028 + tun->nid = current->nid;
2030 dev->ethtool_ops = &tun_ethtool_ops;
2031 dev->destructor = tun_free_netdev;
2032 @@ -1003,7 +1006,7 @@ static int tun_set_iff(struct net *net,
2034 if (((tun->owner != -1 && cred->euid != tun->owner) ||
2035 (tun->group != -1 && !in_egroup_p(tun->group))) &&
2036 - !capable(CAP_NET_ADMIN))
2037 + !cap_raised(current_cap(), CAP_NET_ADMIN))
2039 err = security_tun_dev_attach(tun->socket.sk);
2041 @@ -1017,7 +1020,7 @@ static int tun_set_iff(struct net *net,
2043 unsigned long flags = 0;
2045 - if (!capable(CAP_NET_ADMIN))
2046 + if (!nx_capable(CAP_NET_ADMIN, NXC_TUN_CREATE))
2048 err = security_tun_dev_create();
2050 @@ -1087,6 +1090,9 @@ static int tun_set_iff(struct net *net,
2052 sk->sk_destruct = tun_sock_destruct;
2054 + if (!nx_check(tun->nid, VS_IDENT | VS_HOSTID | VS_ADMIN_P))
2057 err = tun_attach(tun, file);
2060 @@ -1278,6 +1284,16 @@ static long __tun_chr_ioctl(struct file
2061 DBG(KERN_INFO "%s: group set to %d\n", tun->dev->name, tun->group);
2065 + if (!capable(CAP_CONTEXT))
2068 + /* Set nid owner of the device */
2069 + tun->nid = (nid_t) arg;
2071 + DBG(KERN_INFO "%s: nid owner set to %u\n", tun->dev->name, tun->nid);
2075 /* Only allow setting the type when the interface is down */
2076 if (tun->dev->flags & IFF_UP) {
2077 diff -NurpP --minimal linux-2.6.36/fs/attr.c linux-2.6.36-vs2.3.0.36.33/fs/attr.c
2078 --- linux-2.6.36/fs/attr.c 2010-10-21 13:07:47.000000000 +0200
2079 +++ linux-2.6.36-vs2.3.0.36.33/fs/attr.c 2010-10-21 14:04:40.000000000 +0200
2081 #include <linux/fsnotify.h>
2082 #include <linux/fcntl.h>
2083 #include <linux/security.h>
2084 +#include <linux/proc_fs.h>
2085 +#include <linux/devpts_fs.h>
2086 +#include <linux/vs_tag.h>
2089 * inode_change_ok - check if attribute changes to an inode are allowed
2090 @@ -73,6 +76,10 @@ int inode_change_ok(const struct inode *
2094 + /* check for inode tag permission */
2095 + if (dx_permission(inode, MAY_WRITE))
2100 EXPORT_SYMBOL(inode_change_ok);
2101 @@ -143,6 +150,8 @@ void setattr_copy(struct inode *inode, c
2102 inode->i_uid = attr->ia_uid;
2103 if (ia_valid & ATTR_GID)
2104 inode->i_gid = attr->ia_gid;
2105 + if ((ia_valid & ATTR_TAG) && IS_TAGGED(inode))
2106 + inode->i_tag = attr->ia_tag;
2107 if (ia_valid & ATTR_ATIME)
2108 inode->i_atime = timespec_trunc(attr->ia_atime,
2109 inode->i_sb->s_time_gran);
2110 @@ -170,7 +179,8 @@ int notify_change(struct dentry * dentry
2111 struct timespec now;
2112 unsigned int ia_valid = attr->ia_valid;
2114 - if (ia_valid & (ATTR_MODE | ATTR_UID | ATTR_GID | ATTR_TIMES_SET)) {
2115 + if (ia_valid & (ATTR_MODE | ATTR_UID | ATTR_GID |
2116 + ATTR_TAG | ATTR_TIMES_SET)) {
2117 if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
2120 diff -NurpP --minimal linux-2.6.36/fs/binfmt_aout.c linux-2.6.36-vs2.3.0.36.33/fs/binfmt_aout.c
2121 --- linux-2.6.36/fs/binfmt_aout.c 2010-10-21 13:07:47.000000000 +0200
2122 +++ linux-2.6.36-vs2.3.0.36.33/fs/binfmt_aout.c 2010-10-21 13:09:36.000000000 +0200
2124 #include <linux/init.h>
2125 #include <linux/coredump.h>
2126 #include <linux/slab.h>
2127 +#include <linux/vs_memory.h>
2129 #include <asm/system.h>
2130 #include <asm/uaccess.h>
2131 diff -NurpP --minimal linux-2.6.36/fs/binfmt_elf.c linux-2.6.36-vs2.3.0.36.33/fs/binfmt_elf.c
2132 --- linux-2.6.36/fs/binfmt_elf.c 2010-07-07 18:31:50.000000000 +0200
2133 +++ linux-2.6.36-vs2.3.0.36.33/fs/binfmt_elf.c 2010-10-21 13:09:36.000000000 +0200
2135 #include <linux/elf.h>
2136 #include <linux/utsname.h>
2137 #include <linux/coredump.h>
2138 +#include <linux/vs_memory.h>
2139 #include <asm/uaccess.h>
2140 #include <asm/param.h>
2141 #include <asm/page.h>
2142 diff -NurpP --minimal linux-2.6.36/fs/binfmt_flat.c linux-2.6.36-vs2.3.0.36.33/fs/binfmt_flat.c
2143 --- linux-2.6.36/fs/binfmt_flat.c 2010-08-02 16:52:47.000000000 +0200
2144 +++ linux-2.6.36-vs2.3.0.36.33/fs/binfmt_flat.c 2010-10-21 13:09:36.000000000 +0200
2146 #include <linux/init.h>
2147 #include <linux/flat.h>
2148 #include <linux/syscalls.h>
2149 +#include <linux/vs_memory.h>
2151 #include <asm/byteorder.h>
2152 #include <asm/system.h>
2153 diff -NurpP --minimal linux-2.6.36/fs/binfmt_som.c linux-2.6.36-vs2.3.0.36.33/fs/binfmt_som.c
2154 --- linux-2.6.36/fs/binfmt_som.c 2010-02-25 11:52:04.000000000 +0100
2155 +++ linux-2.6.36-vs2.3.0.36.33/fs/binfmt_som.c 2010-10-21 13:09:36.000000000 +0200
2157 #include <linux/shm.h>
2158 #include <linux/personality.h>
2159 #include <linux/init.h>
2160 +#include <linux/vs_memory.h>
2162 #include <asm/uaccess.h>
2163 #include <asm/pgtable.h>
2164 diff -NurpP --minimal linux-2.6.36/fs/block_dev.c linux-2.6.36-vs2.3.0.36.33/fs/block_dev.c
2165 --- linux-2.6.36/fs/block_dev.c 2010-10-21 13:07:47.000000000 +0200
2166 +++ linux-2.6.36-vs2.3.0.36.33/fs/block_dev.c 2010-10-21 13:09:36.000000000 +0200
2168 #include <linux/namei.h>
2169 #include <linux/log2.h>
2170 #include <linux/kmemleak.h>
2171 +#include <linux/vs_device.h>
2172 #include <asm/uaccess.h>
2173 #include "internal.h"
2175 @@ -530,6 +531,7 @@ struct block_device *bdget(dev_t dev)
2176 bdev->bd_invalidated = 0;
2177 inode->i_mode = S_IFBLK;
2178 inode->i_rdev = dev;
2179 + inode->i_mdev = dev;
2180 inode->i_bdev = bdev;
2181 inode->i_data.a_ops = &def_blk_aops;
2182 mapping_set_gfp_mask(&inode->i_data, GFP_USER);
2183 @@ -576,6 +578,11 @@ EXPORT_SYMBOL(bdput);
2184 static struct block_device *bd_acquire(struct inode *inode)
2186 struct block_device *bdev;
2189 + if (!vs_map_blkdev(inode->i_rdev, &mdev, DATTR_OPEN))
2191 + inode->i_mdev = mdev;
2193 spin_lock(&bdev_lock);
2194 bdev = inode->i_bdev;
2195 @@ -586,7 +593,7 @@ static struct block_device *bd_acquire(s
2197 spin_unlock(&bdev_lock);
2199 - bdev = bdget(inode->i_rdev);
2200 + bdev = bdget(mdev);
2202 spin_lock(&bdev_lock);
2203 if (!inode->i_bdev) {
2204 diff -NurpP --minimal linux-2.6.36/fs/btrfs/ctree.h linux-2.6.36-vs2.3.0.36.33/fs/btrfs/ctree.h
2205 --- linux-2.6.36/fs/btrfs/ctree.h 2010-10-21 13:07:48.000000000 +0200
2206 +++ linux-2.6.36-vs2.3.0.36.33/fs/btrfs/ctree.h 2010-10-21 13:09:36.000000000 +0200
2207 @@ -551,11 +551,14 @@ struct btrfs_inode_item {
2208 /* modification sequence number for NFS */
2213 * a little future expansion, for more than this we can
2214 * just grow the inode item and version it
2216 - __le64 reserved[4];
2217 + __le16 reserved16;
2218 + __le32 reserved32;
2219 + __le64 reserved[3];
2220 struct btrfs_timespec atime;
2221 struct btrfs_timespec ctime;
2222 struct btrfs_timespec mtime;
2223 @@ -1193,6 +1196,8 @@ struct btrfs_root {
2224 #define BTRFS_MOUNT_DISCARD (1 << 10)
2225 #define BTRFS_MOUNT_FORCE_COMPRESS (1 << 11)
2227 +#define BTRFS_MOUNT_TAGGED (1 << 24)
2229 #define btrfs_clear_opt(o, opt) ((o) &= ~BTRFS_MOUNT_##opt)
2230 #define btrfs_set_opt(o, opt) ((o) |= BTRFS_MOUNT_##opt)
2231 #define btrfs_test_opt(root, opt) ((root)->fs_info->mount_opt & \
2232 @@ -1413,6 +1418,7 @@ BTRFS_SETGET_FUNCS(inode_block_group, st
2233 BTRFS_SETGET_FUNCS(inode_nlink, struct btrfs_inode_item, nlink, 32);
2234 BTRFS_SETGET_FUNCS(inode_uid, struct btrfs_inode_item, uid, 32);
2235 BTRFS_SETGET_FUNCS(inode_gid, struct btrfs_inode_item, gid, 32);
2236 +BTRFS_SETGET_FUNCS(inode_tag, struct btrfs_inode_item, tag, 16);
2237 BTRFS_SETGET_FUNCS(inode_mode, struct btrfs_inode_item, mode, 32);
2238 BTRFS_SETGET_FUNCS(inode_rdev, struct btrfs_inode_item, rdev, 64);
2239 BTRFS_SETGET_FUNCS(inode_flags, struct btrfs_inode_item, flags, 64);
2240 @@ -1474,6 +1480,10 @@ BTRFS_SETGET_FUNCS(extent_flags, struct
2242 BTRFS_SETGET_FUNCS(extent_refs_v0, struct btrfs_extent_item_v0, refs, 32);
2244 +#define BTRFS_INODE_IXUNLINK (1 << 24)
2245 +#define BTRFS_INODE_BARRIER (1 << 25)
2246 +#define BTRFS_INODE_COW (1 << 26)
2249 BTRFS_SETGET_FUNCS(tree_block_level, struct btrfs_tree_block_info, level, 8);
2251 @@ -2432,6 +2442,7 @@ extern const struct dentry_operations bt
2252 long btrfs_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
2253 void btrfs_update_iflags(struct inode *inode);
2254 void btrfs_inherit_iflags(struct inode *inode, struct inode *dir);
2255 +int btrfs_sync_flags(struct inode *inode, int, int);
2258 int btrfs_sync_file(struct file *file, int datasync);
2259 diff -NurpP --minimal linux-2.6.36/fs/btrfs/disk-io.c linux-2.6.36-vs2.3.0.36.33/fs/btrfs/disk-io.c
2260 --- linux-2.6.36/fs/btrfs/disk-io.c 2010-10-21 13:07:48.000000000 +0200
2261 +++ linux-2.6.36-vs2.3.0.36.33/fs/btrfs/disk-io.c 2010-10-21 13:09:36.000000000 +0200
2262 @@ -1707,6 +1707,9 @@ struct btrfs_root *open_ctree(struct sup
2266 + if (btrfs_test_opt(tree_root, TAGGED))
2267 + sb->s_flags |= MS_TAGGED;
2269 features = btrfs_super_incompat_flags(disk_super) &
2270 ~BTRFS_FEATURE_INCOMPAT_SUPP;
2272 diff -NurpP --minimal linux-2.6.36/fs/btrfs/inode.c linux-2.6.36-vs2.3.0.36.33/fs/btrfs/inode.c
2273 --- linux-2.6.36/fs/btrfs/inode.c 2010-10-21 13:07:48.000000000 +0200
2274 +++ linux-2.6.36-vs2.3.0.36.33/fs/btrfs/inode.c 2010-10-21 13:09:36.000000000 +0200
2276 #include <linux/posix_acl.h>
2277 #include <linux/falloc.h>
2278 #include <linux/slab.h>
2279 +#include <linux/vs_tag.h>
2283 #include "disk-io.h"
2284 @@ -2415,6 +2417,8 @@ static void btrfs_read_locked_inode(stru
2286 u64 alloc_group_block;
2292 path = btrfs_alloc_path();
2293 @@ -2431,8 +2435,13 @@ static void btrfs_read_locked_inode(stru
2295 inode->i_mode = btrfs_inode_mode(leaf, inode_item);
2296 inode->i_nlink = btrfs_inode_nlink(leaf, inode_item);
2297 - inode->i_uid = btrfs_inode_uid(leaf, inode_item);
2298 - inode->i_gid = btrfs_inode_gid(leaf, inode_item);
2300 + uid = btrfs_inode_uid(leaf, inode_item);
2301 + gid = btrfs_inode_gid(leaf, inode_item);
2302 + inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
2303 + inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
2304 + inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid,
2305 + btrfs_inode_tag(leaf, inode_item));
2306 btrfs_i_size_write(inode, btrfs_inode_size(leaf, inode_item));
2308 tspec = btrfs_inode_atime(inode_item);
2309 @@ -2514,8 +2523,15 @@ static void fill_inode_item(struct btrfs
2310 struct btrfs_inode_item *item,
2311 struct inode *inode)
2313 - btrfs_set_inode_uid(leaf, item, inode->i_uid);
2314 - btrfs_set_inode_gid(leaf, item, inode->i_gid);
2315 + uid_t uid = TAGINO_UID(DX_TAG(inode), inode->i_uid, inode->i_tag);
2316 + gid_t gid = TAGINO_GID(DX_TAG(inode), inode->i_gid, inode->i_tag);
2318 + btrfs_set_inode_uid(leaf, item, uid);
2319 + btrfs_set_inode_gid(leaf, item, gid);
2320 +#ifdef CONFIG_TAGGING_INTERN
2321 + btrfs_set_inode_tag(leaf, item, inode->i_tag);
2324 btrfs_set_inode_size(leaf, item, BTRFS_I(inode)->disk_i_size);
2325 btrfs_set_inode_mode(leaf, item, inode->i_mode);
2326 btrfs_set_inode_nlink(leaf, item, inode->i_nlink);
2327 @@ -6946,6 +6962,7 @@ static const struct inode_operations btr
2328 .listxattr = btrfs_listxattr,
2329 .removexattr = btrfs_removexattr,
2330 .permission = btrfs_permission,
2331 + .sync_flags = btrfs_sync_flags,
2333 static const struct inode_operations btrfs_dir_ro_inode_operations = {
2334 .lookup = btrfs_lookup,
2335 @@ -7021,6 +7038,7 @@ static const struct inode_operations btr
2336 .permission = btrfs_permission,
2337 .fallocate = btrfs_fallocate,
2338 .fiemap = btrfs_fiemap,
2339 + .sync_flags = btrfs_sync_flags,
2341 static const struct inode_operations btrfs_special_inode_operations = {
2342 .getattr = btrfs_getattr,
2343 diff -NurpP --minimal linux-2.6.36/fs/btrfs/ioctl.c linux-2.6.36-vs2.3.0.36.33/fs/btrfs/ioctl.c
2344 --- linux-2.6.36/fs/btrfs/ioctl.c 2010-08-02 16:52:47.000000000 +0200
2345 +++ linux-2.6.36-vs2.3.0.36.33/fs/btrfs/ioctl.c 2010-10-21 13:09:36.000000000 +0200
2346 @@ -68,10 +68,13 @@ static unsigned int btrfs_flags_to_ioctl
2348 unsigned int iflags = 0;
2350 - if (flags & BTRFS_INODE_SYNC)
2351 - iflags |= FS_SYNC_FL;
2352 if (flags & BTRFS_INODE_IMMUTABLE)
2353 iflags |= FS_IMMUTABLE_FL;
2354 + if (flags & BTRFS_INODE_IXUNLINK)
2355 + iflags |= FS_IXUNLINK_FL;
2357 + if (flags & BTRFS_INODE_SYNC)
2358 + iflags |= FS_SYNC_FL;
2359 if (flags & BTRFS_INODE_APPEND)
2360 iflags |= FS_APPEND_FL;
2361 if (flags & BTRFS_INODE_NODUMP)
2362 @@ -81,28 +84,78 @@ static unsigned int btrfs_flags_to_ioctl
2363 if (flags & BTRFS_INODE_DIRSYNC)
2364 iflags |= FS_DIRSYNC_FL;
2366 + if (flags & BTRFS_INODE_BARRIER)
2367 + iflags |= FS_BARRIER_FL;
2368 + if (flags & BTRFS_INODE_COW)
2369 + iflags |= FS_COW_FL;
2374 - * Update inode->i_flags based on the btrfs internal flags.
2375 + * Update inode->i_(v)flags based on the btrfs internal flags.
2377 void btrfs_update_iflags(struct inode *inode)
2379 struct btrfs_inode *ip = BTRFS_I(inode);
2381 - inode->i_flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC);
2382 + inode->i_flags &= ~(S_IMMUTABLE | S_IXUNLINK |
2383 + S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
2385 - if (ip->flags & BTRFS_INODE_SYNC)
2386 - inode->i_flags |= S_SYNC;
2387 if (ip->flags & BTRFS_INODE_IMMUTABLE)
2388 inode->i_flags |= S_IMMUTABLE;
2389 + if (ip->flags & BTRFS_INODE_IXUNLINK)
2390 + inode->i_flags |= S_IXUNLINK;
2392 + if (ip->flags & BTRFS_INODE_SYNC)
2393 + inode->i_flags |= S_SYNC;
2394 if (ip->flags & BTRFS_INODE_APPEND)
2395 inode->i_flags |= S_APPEND;
2396 if (ip->flags & BTRFS_INODE_NOATIME)
2397 inode->i_flags |= S_NOATIME;
2398 if (ip->flags & BTRFS_INODE_DIRSYNC)
2399 inode->i_flags |= S_DIRSYNC;
2401 + inode->i_vflags &= ~(V_BARRIER | V_COW);
2403 + if (ip->flags & BTRFS_INODE_BARRIER)
2404 + inode->i_vflags |= V_BARRIER;
2405 + if (ip->flags & BTRFS_INODE_COW)
2406 + inode->i_vflags |= V_COW;
2410 + * Update btrfs internal flags from inode->i_(v)flags.
2412 +void btrfs_update_flags(struct inode *inode)
2414 + struct btrfs_inode *ip = BTRFS_I(inode);
2416 + unsigned int flags = inode->i_flags;
2417 + unsigned int vflags = inode->i_vflags;
2419 + ip->flags &= ~(BTRFS_INODE_SYNC | BTRFS_INODE_APPEND |
2420 + BTRFS_INODE_IMMUTABLE | BTRFS_INODE_IXUNLINK |
2421 + BTRFS_INODE_NOATIME | BTRFS_INODE_DIRSYNC |
2422 + BTRFS_INODE_BARRIER | BTRFS_INODE_COW);
2424 + if (flags & S_IMMUTABLE)
2425 + ip->flags |= BTRFS_INODE_IMMUTABLE;
2426 + if (flags & S_IXUNLINK)
2427 + ip->flags |= BTRFS_INODE_IXUNLINK;
2429 + if (flags & S_SYNC)
2430 + ip->flags |= BTRFS_INODE_SYNC;
2431 + if (flags & S_APPEND)
2432 + ip->flags |= BTRFS_INODE_APPEND;
2433 + if (flags & S_NOATIME)
2434 + ip->flags |= BTRFS_INODE_NOATIME;
2435 + if (flags & S_DIRSYNC)
2436 + ip->flags |= BTRFS_INODE_DIRSYNC;
2438 + if (vflags & V_BARRIER)
2439 + ip->flags |= BTRFS_INODE_BARRIER;
2440 + if (vflags & V_COW)
2441 + ip->flags |= BTRFS_INODE_COW;
2445 @@ -120,7 +173,7 @@ void btrfs_inherit_iflags(struct inode *
2446 flags = BTRFS_I(dir)->flags;
2448 if (S_ISREG(inode->i_mode))
2449 - flags &= ~BTRFS_INODE_DIRSYNC;
2450 + flags &= ~(BTRFS_INODE_DIRSYNC | BTRFS_INODE_BARRIER);
2451 else if (!S_ISDIR(inode->i_mode))
2452 flags &= (BTRFS_INODE_NODUMP | BTRFS_INODE_NOATIME);
2454 @@ -128,6 +181,30 @@ void btrfs_inherit_iflags(struct inode *
2455 btrfs_update_iflags(inode);
2458 +int btrfs_sync_flags(struct inode *inode, int flags, int vflags)
2460 + struct btrfs_inode *ip = BTRFS_I(inode);
2461 + struct btrfs_root *root = ip->root;
2462 + struct btrfs_trans_handle *trans;
2465 + trans = btrfs_join_transaction(root, 1);
2468 + inode->i_flags = flags;
2469 + inode->i_vflags = vflags;
2470 + btrfs_update_flags(inode);
2472 + ret = btrfs_update_inode(trans, root, inode);
2475 + btrfs_update_iflags(inode);
2476 + inode->i_ctime = CURRENT_TIME;
2477 + btrfs_end_transaction(trans, root);
2482 static int btrfs_ioctl_getflags(struct file *file, void __user *arg)
2484 struct btrfs_inode *ip = BTRFS_I(file->f_path.dentry->d_inode);
2485 @@ -150,6 +227,7 @@ static int btrfs_ioctl_setflags(struct f
2486 if (copy_from_user(&flags, arg, sizeof(flags)))
2489 + /* maybe add FS_IXUNLINK_FL ? */
2490 if (flags & ~(FS_IMMUTABLE_FL | FS_APPEND_FL | \
2491 FS_NOATIME_FL | FS_NODUMP_FL | \
2492 FS_SYNC_FL | FS_DIRSYNC_FL))
2493 @@ -162,7 +240,8 @@ static int btrfs_ioctl_setflags(struct f
2495 flags = btrfs_mask_flags(inode->i_mode, flags);
2496 oldflags = btrfs_flags_to_ioctl(ip->flags);
2497 - if ((flags ^ oldflags) & (FS_APPEND_FL | FS_IMMUTABLE_FL)) {
2498 + if ((flags ^ oldflags) & (FS_APPEND_FL |
2499 + FS_IMMUTABLE_FL | FS_IXUNLINK_FL)) {
2500 if (!capable(CAP_LINUX_IMMUTABLE)) {
2503 @@ -173,14 +252,19 @@ static int btrfs_ioctl_setflags(struct f
2507 - if (flags & FS_SYNC_FL)
2508 - ip->flags |= BTRFS_INODE_SYNC;
2510 - ip->flags &= ~BTRFS_INODE_SYNC;
2511 if (flags & FS_IMMUTABLE_FL)
2512 ip->flags |= BTRFS_INODE_IMMUTABLE;
2514 ip->flags &= ~BTRFS_INODE_IMMUTABLE;
2515 + if (flags & FS_IXUNLINK_FL)
2516 + ip->flags |= BTRFS_INODE_IXUNLINK;
2518 + ip->flags &= ~BTRFS_INODE_IXUNLINK;
2520 + if (flags & FS_SYNC_FL)
2521 + ip->flags |= BTRFS_INODE_SYNC;
2523 + ip->flags &= ~BTRFS_INODE_SYNC;
2524 if (flags & FS_APPEND_FL)
2525 ip->flags |= BTRFS_INODE_APPEND;
2527 diff -NurpP --minimal linux-2.6.36/fs/btrfs/super.c linux-2.6.36-vs2.3.0.36.33/fs/btrfs/super.c
2528 --- linux-2.6.36/fs/btrfs/super.c 2010-10-21 13:07:48.000000000 +0200
2529 +++ linux-2.6.36-vs2.3.0.36.33/fs/btrfs/super.c 2010-10-21 13:09:36.000000000 +0200
2530 @@ -68,7 +68,7 @@ enum {
2531 Opt_nodatacow, Opt_max_inline, Opt_alloc_start, Opt_nobarrier, Opt_ssd,
2532 Opt_nossd, Opt_ssd_spread, Opt_thread_pool, Opt_noacl, Opt_compress,
2533 Opt_compress_force, Opt_notreelog, Opt_ratio, Opt_flushoncommit,
2534 - Opt_discard, Opt_err,
2535 + Opt_tag, Opt_notag, Opt_tagid, Opt_discard, Opt_err,
2538 static match_table_t tokens = {
2539 @@ -92,6 +92,9 @@ static match_table_t tokens = {
2540 {Opt_flushoncommit, "flushoncommit"},
2541 {Opt_ratio, "metadata_ratio=%d"},
2542 {Opt_discard, "discard"},
2544 + {Opt_notag, "notag"},
2545 + {Opt_tagid, "tagid=%u"},
2549 @@ -235,6 +238,22 @@ int btrfs_parse_options(struct btrfs_roo
2551 btrfs_set_opt(info->mount_opt, DISCARD);
2553 +#ifndef CONFIG_TAGGING_NONE
2555 + printk(KERN_INFO "btrfs: use tagging\n");
2556 + btrfs_set_opt(info->mount_opt, TAGGED);
2559 + printk(KERN_INFO "btrfs: disabled tagging\n");
2560 + btrfs_clear_opt(info->mount_opt, TAGGED);
2563 +#ifdef CONFIG_PROPAGATE
2566 + btrfs_set_opt(info->mount_opt, TAGGED);
2570 printk(KERN_INFO "btrfs: unrecognized mount option "
2572 @@ -681,6 +700,12 @@ static int btrfs_remount(struct super_bl
2576 + if (btrfs_test_opt(root, TAGGED) && !(sb->s_flags & MS_TAGGED)) {
2577 + printk("btrfs: %s: tagging not permitted on remount.\n",
2582 if ((*flags & MS_RDONLY) == (sb->s_flags & MS_RDONLY))
2585 diff -NurpP --minimal linux-2.6.36/fs/char_dev.c linux-2.6.36-vs2.3.0.36.33/fs/char_dev.c
2586 --- linux-2.6.36/fs/char_dev.c 2010-10-21 13:07:48.000000000 +0200
2587 +++ linux-2.6.36-vs2.3.0.36.33/fs/char_dev.c 2010-10-21 13:09:36.000000000 +0200
2589 #include <linux/mutex.h>
2590 #include <linux/backing-dev.h>
2591 #include <linux/tty.h>
2592 +#include <linux/vs_context.h>
2593 +#include <linux/vs_device.h>
2595 #include "internal.h"
2597 @@ -371,14 +373,21 @@ static int chrdev_open(struct inode *ino
2599 struct cdev *new = NULL;
2603 + if (!vs_map_chrdev(inode->i_rdev, &mdev, DATTR_OPEN))
2605 + inode->i_mdev = mdev;
2607 spin_lock(&cdev_lock);
2610 struct kobject *kobj;
2613 spin_unlock(&cdev_lock);
2614 - kobj = kobj_lookup(cdev_map, inode->i_rdev, &idx);
2616 + kobj = kobj_lookup(cdev_map, mdev, &idx);
2619 new = container_of(kobj, struct cdev, kobj);
2620 diff -NurpP --minimal linux-2.6.36/fs/dcache.c linux-2.6.36-vs2.3.0.36.33/fs/dcache.c
2621 --- linux-2.6.36/fs/dcache.c 2010-10-21 13:07:48.000000000 +0200
2622 +++ linux-2.6.36-vs2.3.0.36.33/fs/dcache.c 2010-10-21 13:09:36.000000000 +0200
2624 #include <linux/bootmem.h>
2625 #include <linux/fs_struct.h>
2626 #include <linux/hardirq.h>
2627 +#include <linux/vs_limit.h>
2628 #include "internal.h"
2630 int sysctl_vfs_cache_pressure __read_mostly = 100;
2631 @@ -230,6 +231,8 @@ repeat:
2635 + vx_dentry_dec(dentry);
2638 * AV: ->d_delete() is _NOT_ allowed to block now.
2640 @@ -323,6 +326,7 @@ static inline struct dentry * __dget_loc
2642 atomic_inc(&dentry->d_count);
2643 dentry_lru_del_init(dentry);
2644 + vx_dentry_inc(dentry);
2648 @@ -928,6 +932,9 @@ struct dentry *d_alloc(struct dentry * p
2649 struct dentry *dentry;
2652 + if (!vx_dentry_avail(1))
2655 dentry = kmem_cache_alloc(dentry_cache, GFP_KERNEL);
2658 @@ -973,6 +980,7 @@ struct dentry *d_alloc(struct dentry * p
2660 list_add(&dentry->d_u.d_child, &parent->d_subdirs);
2661 dentry_stat.nr_dentry++;
2662 + vx_dentry_inc(dentry);
2663 spin_unlock(&dcache_lock);
2666 @@ -1432,6 +1440,7 @@ struct dentry * __d_lookup(struct dentry
2669 atomic_inc(&dentry->d_count);
2670 + vx_dentry_inc(dentry);
2672 spin_unlock(&dentry->d_lock);
2674 diff -NurpP --minimal linux-2.6.36/fs/devpts/inode.c linux-2.6.36-vs2.3.0.36.33/fs/devpts/inode.c
2675 --- linux-2.6.36/fs/devpts/inode.c 2010-08-02 16:52:48.000000000 +0200
2676 +++ linux-2.6.36-vs2.3.0.36.33/fs/devpts/inode.c 2010-10-21 13:09:36.000000000 +0200
2678 #include <linux/parser.h>
2679 #include <linux/fsnotify.h>
2680 #include <linux/seq_file.h>
2681 +#include <linux/vs_base.h>
2683 #define DEVPTS_DEFAULT_MODE 0600
2686 #define DEVPTS_DEFAULT_PTMX_MODE 0000
2687 #define PTMX_MINOR 2
2689 +static int devpts_permission(struct inode *inode, int mask)
2691 + int ret = -EACCES;
2693 + /* devpts is xid tagged */
2694 + if (vx_check((xid_t)inode->i_tag, VS_WATCH_P | VS_IDENT))
2695 + ret = generic_permission(inode, mask, NULL);
2699 +static struct inode_operations devpts_file_inode_operations = {
2700 + .permission = devpts_permission,
2703 extern int pty_limit; /* Config limit on Unix98 ptys */
2704 static DEFINE_MUTEX(allocated_ptys_lock);
2706 @@ -263,6 +278,25 @@ static int devpts_show_options(struct se
2710 +static int devpts_filter(struct dentry *de)
2712 + /* devpts is xid tagged */
2713 + return vx_check((xid_t)de->d_inode->i_tag, VS_WATCH_P | VS_IDENT);
2716 +static int devpts_readdir(struct file * filp, void * dirent, filldir_t filldir)
2718 + return dcache_readdir_filter(filp, dirent, filldir, devpts_filter);
2721 +static struct file_operations devpts_dir_operations = {
2722 + .open = dcache_dir_open,
2723 + .release = dcache_dir_close,
2724 + .llseek = dcache_dir_lseek,
2725 + .read = generic_read_dir,
2726 + .readdir = devpts_readdir,
2729 static const struct super_operations devpts_sops = {
2730 .statfs = simple_statfs,
2731 .remount_fs = devpts_remount,
2732 @@ -302,12 +336,15 @@ devpts_fill_super(struct super_block *s,
2733 inode = new_inode(s);
2738 inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
2739 inode->i_mode = S_IFDIR | S_IRUGO | S_IXUGO | S_IWUSR;
2740 inode->i_op = &simple_dir_inode_operations;
2741 - inode->i_fop = &simple_dir_operations;
2742 + inode->i_fop = &devpts_dir_operations;
2744 + /* devpts is xid tagged */
2745 + inode->i_tag = (tag_t)vx_current_xid();
2747 s->s_root = d_alloc_root(inode);
2749 @@ -495,6 +532,9 @@ int devpts_pty_new(struct inode *ptmx_in
2750 inode->i_gid = opts->setgid ? opts->gid : current_fsgid();
2751 inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
2752 init_special_inode(inode, S_IFCHR|opts->mode, device);
2753 + /* devpts is xid tagged */
2754 + inode->i_tag = (tag_t)vx_current_xid();
2755 + inode->i_op = &devpts_file_inode_operations;
2756 inode->i_private = tty;
2757 tty->driver_data = inode;
2759 diff -NurpP --minimal linux-2.6.36/fs/ext2/balloc.c linux-2.6.36-vs2.3.0.36.33/fs/ext2/balloc.c
2760 --- linux-2.6.36/fs/ext2/balloc.c 2010-10-21 13:07:48.000000000 +0200
2761 +++ linux-2.6.36-vs2.3.0.36.33/fs/ext2/balloc.c 2010-10-21 13:09:36.000000000 +0200
2762 @@ -702,7 +702,6 @@ ext2_try_to_allocate(struct super_block
2764 end = EXT2_BLOCKS_PER_GROUP(sb);
2767 BUG_ON(start > EXT2_BLOCKS_PER_GROUP(sb));
2770 diff -NurpP --minimal linux-2.6.36/fs/ext2/ext2.h linux-2.6.36-vs2.3.0.36.33/fs/ext2/ext2.h
2771 --- linux-2.6.36/fs/ext2/ext2.h 2010-10-21 13:07:48.000000000 +0200
2772 +++ linux-2.6.36-vs2.3.0.36.33/fs/ext2/ext2.h 2010-10-21 14:10:55.000000000 +0200
2773 @@ -127,6 +127,7 @@ extern void ext2_set_inode_flags(struct
2774 extern void ext2_get_inode_flags(struct ext2_inode_info *);
2775 extern int ext2_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
2776 u64 start, u64 len);
2777 +extern int ext2_sync_flags(struct inode *, int, int);
2780 extern long ext2_ioctl(struct file *, unsigned int, unsigned long);
2781 diff -NurpP --minimal linux-2.6.36/fs/ext2/file.c linux-2.6.36-vs2.3.0.36.33/fs/ext2/file.c
2782 --- linux-2.6.36/fs/ext2/file.c 2010-08-02 16:52:48.000000000 +0200
2783 +++ linux-2.6.36-vs2.3.0.36.33/fs/ext2/file.c 2010-10-21 13:09:36.000000000 +0200
2784 @@ -104,4 +104,5 @@ const struct inode_operations ext2_file_
2785 .setattr = ext2_setattr,
2786 .check_acl = ext2_check_acl,
2787 .fiemap = ext2_fiemap,
2788 + .sync_flags = ext2_sync_flags,
2790 diff -NurpP --minimal linux-2.6.36/fs/ext2/ialloc.c linux-2.6.36-vs2.3.0.36.33/fs/ext2/ialloc.c
2791 --- linux-2.6.36/fs/ext2/ialloc.c 2010-10-21 13:07:48.000000000 +0200
2792 +++ linux-2.6.36-vs2.3.0.36.33/fs/ext2/ialloc.c 2010-10-21 13:09:36.000000000 +0200
2794 #include <linux/backing-dev.h>
2795 #include <linux/buffer_head.h>
2796 #include <linux/random.h>
2797 +#include <linux/vs_tag.h>
2801 @@ -548,6 +549,7 @@ got:
2802 inode->i_mode = mode;
2803 inode->i_uid = current_fsuid();
2804 inode->i_gid = dir->i_gid;
2805 + inode->i_tag = dx_current_fstag(sb);
2807 inode_init_owner(inode, dir, mode);
2809 diff -NurpP --minimal linux-2.6.36/fs/ext2/inode.c linux-2.6.36-vs2.3.0.36.33/fs/ext2/inode.c
2810 --- linux-2.6.36/fs/ext2/inode.c 2010-10-21 13:07:48.000000000 +0200
2811 +++ linux-2.6.36-vs2.3.0.36.33/fs/ext2/inode.c 2010-10-21 13:09:36.000000000 +0200
2813 #include <linux/mpage.h>
2814 #include <linux/fiemap.h>
2815 #include <linux/namei.h>
2816 +#include <linux/vs_tag.h>
2820 @@ -1169,7 +1170,7 @@ static void ext2_truncate_blocks(struct
2822 if (ext2_inode_is_fast_symlink(inode))
2824 - if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
2825 + if (IS_APPEND(inode) || IS_IXORUNLINK(inode))
2827 __ext2_truncate_blocks(inode, offset);
2829 @@ -1258,36 +1259,61 @@ void ext2_set_inode_flags(struct inode *
2831 unsigned int flags = EXT2_I(inode)->i_flags;
2833 - inode->i_flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC);
2834 + inode->i_flags &= ~(S_IMMUTABLE | S_IXUNLINK |
2835 + S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
2838 + if (flags & EXT2_IMMUTABLE_FL)
2839 + inode->i_flags |= S_IMMUTABLE;
2840 + if (flags & EXT2_IXUNLINK_FL)
2841 + inode->i_flags |= S_IXUNLINK;
2843 if (flags & EXT2_SYNC_FL)
2844 inode->i_flags |= S_SYNC;
2845 if (flags & EXT2_APPEND_FL)
2846 inode->i_flags |= S_APPEND;
2847 - if (flags & EXT2_IMMUTABLE_FL)
2848 - inode->i_flags |= S_IMMUTABLE;
2849 if (flags & EXT2_NOATIME_FL)
2850 inode->i_flags |= S_NOATIME;
2851 if (flags & EXT2_DIRSYNC_FL)
2852 inode->i_flags |= S_DIRSYNC;
2854 + inode->i_vflags &= ~(V_BARRIER | V_COW);
2856 + if (flags & EXT2_BARRIER_FL)
2857 + inode->i_vflags |= V_BARRIER;
2858 + if (flags & EXT2_COW_FL)
2859 + inode->i_vflags |= V_COW;
2862 /* Propagate flags from i_flags to EXT2_I(inode)->i_flags */
2863 void ext2_get_inode_flags(struct ext2_inode_info *ei)
2865 unsigned int flags = ei->vfs_inode.i_flags;
2866 + unsigned int vflags = ei->vfs_inode.i_vflags;
2868 + ei->i_flags &= ~(EXT2_SYNC_FL | EXT2_APPEND_FL |
2869 + EXT2_IMMUTABLE_FL | EXT2_IXUNLINK_FL |
2870 + EXT2_NOATIME_FL | EXT2_DIRSYNC_FL |
2871 + EXT2_BARRIER_FL | EXT2_COW_FL);
2873 + if (flags & S_IMMUTABLE)
2874 + ei->i_flags |= EXT2_IMMUTABLE_FL;
2875 + if (flags & S_IXUNLINK)
2876 + ei->i_flags |= EXT2_IXUNLINK_FL;
2878 - ei->i_flags &= ~(EXT2_SYNC_FL|EXT2_APPEND_FL|
2879 - EXT2_IMMUTABLE_FL|EXT2_NOATIME_FL|EXT2_DIRSYNC_FL);
2881 ei->i_flags |= EXT2_SYNC_FL;
2882 if (flags & S_APPEND)
2883 ei->i_flags |= EXT2_APPEND_FL;
2884 - if (flags & S_IMMUTABLE)
2885 - ei->i_flags |= EXT2_IMMUTABLE_FL;
2886 if (flags & S_NOATIME)
2887 ei->i_flags |= EXT2_NOATIME_FL;
2888 if (flags & S_DIRSYNC)
2889 ei->i_flags |= EXT2_DIRSYNC_FL;
2891 + if (vflags & V_BARRIER)
2892 + ei->i_flags |= EXT2_BARRIER_FL;
2893 + if (vflags & V_COW)
2894 + ei->i_flags |= EXT2_COW_FL;
2897 struct inode *ext2_iget (struct super_block *sb, unsigned long ino)
2898 @@ -1297,6 +1323,8 @@ struct inode *ext2_iget (struct super_bl
2899 struct ext2_inode *raw_inode;
2900 struct inode *inode;
2906 inode = iget_locked(sb, ino);
2907 @@ -1315,12 +1343,17 @@ struct inode *ext2_iget (struct super_bl
2910 inode->i_mode = le16_to_cpu(raw_inode->i_mode);
2911 - inode->i_uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
2912 - inode->i_gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
2913 + uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
2914 + gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
2915 if (!(test_opt (inode->i_sb, NO_UID32))) {
2916 - inode->i_uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
2917 - inode->i_gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
2918 + uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
2919 + gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
2921 + inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
2922 + inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
2923 + inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid,
2924 + le16_to_cpu(raw_inode->i_raw_tag));
2926 inode->i_nlink = le16_to_cpu(raw_inode->i_links_count);
2927 inode->i_size = le32_to_cpu(raw_inode->i_size);
2928 inode->i_atime.tv_sec = (signed)le32_to_cpu(raw_inode->i_atime);
2929 @@ -1418,8 +1451,8 @@ static int __ext2_write_inode(struct ino
2930 struct ext2_inode_info *ei = EXT2_I(inode);
2931 struct super_block *sb = inode->i_sb;
2932 ino_t ino = inode->i_ino;
2933 - uid_t uid = inode->i_uid;
2934 - gid_t gid = inode->i_gid;
2935 + uid_t uid = TAGINO_UID(DX_TAG(inode), inode->i_uid, inode->i_tag);
2936 + gid_t gid = TAGINO_GID(DX_TAG(inode), inode->i_gid, inode->i_tag);
2937 struct buffer_head * bh;
2938 struct ext2_inode * raw_inode = ext2_get_inode(sb, ino, &bh);
2940 @@ -1455,6 +1488,9 @@ static int __ext2_write_inode(struct ino
2941 raw_inode->i_uid_high = 0;
2942 raw_inode->i_gid_high = 0;
2944 +#ifdef CONFIG_TAGGING_INTERN
2945 + raw_inode->i_raw_tag = cpu_to_le16(inode->i_tag);
2947 raw_inode->i_links_count = cpu_to_le16(inode->i_nlink);
2948 raw_inode->i_size = cpu_to_le32(inode->i_size);
2949 raw_inode->i_atime = cpu_to_le32(inode->i_atime.tv_sec);
2950 @@ -1544,7 +1580,8 @@ int ext2_setattr(struct dentry *dentry,
2951 if (is_quota_modification(inode, iattr))
2952 dquot_initialize(inode);
2953 if ((iattr->ia_valid & ATTR_UID && iattr->ia_uid != inode->i_uid) ||
2954 - (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid)) {
2955 + (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid) ||
2956 + (iattr->ia_valid & ATTR_TAG && iattr->ia_tag != inode->i_tag)) {
2957 error = dquot_transfer(inode, iattr);
2960 diff -NurpP --minimal linux-2.6.36/fs/ext2/ioctl.c linux-2.6.36-vs2.3.0.36.33/fs/ext2/ioctl.c
2961 --- linux-2.6.36/fs/ext2/ioctl.c 2009-09-10 15:26:21.000000000 +0200
2962 +++ linux-2.6.36-vs2.3.0.36.33/fs/ext2/ioctl.c 2010-10-21 13:09:36.000000000 +0200
2964 #include <asm/uaccess.h>
2967 +int ext2_sync_flags(struct inode *inode, int flags, int vflags)
2969 + inode->i_flags = flags;
2970 + inode->i_vflags = vflags;
2971 + ext2_get_inode_flags(EXT2_I(inode));
2972 + inode->i_ctime = CURRENT_TIME_SEC;
2973 + mark_inode_dirty(inode);
2977 long ext2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
2979 struct inode *inode = filp->f_dentry->d_inode;
2980 @@ -51,6 +61,11 @@ long ext2_ioctl(struct file *filp, unsig
2982 flags = ext2_mask_flags(inode->i_mode, flags);
2984 + if (IS_BARRIER(inode)) {
2985 + vxwprintk_task(1, "messing with the barrier.");
2989 mutex_lock(&inode->i_mutex);
2990 /* Is it quota file? Do not allow user to mess with it */
2991 if (IS_NOQUOTA(inode)) {
2992 @@ -66,7 +81,9 @@ long ext2_ioctl(struct file *filp, unsig
2994 * This test looks nicer. Thanks to Pauline Middelink
2996 - if ((flags ^ oldflags) & (EXT2_APPEND_FL | EXT2_IMMUTABLE_FL)) {
2997 + if ((oldflags & EXT2_IMMUTABLE_FL) ||
2998 + ((flags ^ oldflags) & (EXT2_APPEND_FL |
2999 + EXT2_IMMUTABLE_FL | EXT2_IXUNLINK_FL))) {
3000 if (!capable(CAP_LINUX_IMMUTABLE)) {
3001 mutex_unlock(&inode->i_mutex);
3003 @@ -74,7 +91,7 @@ long ext2_ioctl(struct file *filp, unsig
3007 - flags = flags & EXT2_FL_USER_MODIFIABLE;
3008 + flags &= EXT2_FL_USER_MODIFIABLE;
3009 flags |= oldflags & ~EXT2_FL_USER_MODIFIABLE;
3010 ei->i_flags = flags;
3011 mutex_unlock(&inode->i_mutex);
3012 diff -NurpP --minimal linux-2.6.36/fs/ext2/namei.c linux-2.6.36-vs2.3.0.36.33/fs/ext2/namei.c
3013 --- linux-2.6.36/fs/ext2/namei.c 2010-07-07 18:31:51.000000000 +0200
3014 +++ linux-2.6.36-vs2.3.0.36.33/fs/ext2/namei.c 2010-10-21 13:09:36.000000000 +0200
3017 #include <linux/pagemap.h>
3018 #include <linux/quotaops.h>
3019 +#include <linux/vs_tag.h>
3023 @@ -75,6 +76,7 @@ static struct dentry *ext2_lookup(struct
3024 return ERR_PTR(-EIO);
3026 return ERR_CAST(inode);
3027 + dx_propagate_tag(nd, inode);
3031 @@ -418,6 +420,7 @@ const struct inode_operations ext2_dir_i
3033 .setattr = ext2_setattr,
3034 .check_acl = ext2_check_acl,
3035 + .sync_flags = ext2_sync_flags,
3038 const struct inode_operations ext2_special_inode_operations = {
3039 diff -NurpP --minimal linux-2.6.36/fs/ext2/super.c linux-2.6.36-vs2.3.0.36.33/fs/ext2/super.c
3040 --- linux-2.6.36/fs/ext2/super.c 2010-10-21 13:07:48.000000000 +0200
3041 +++ linux-2.6.36-vs2.3.0.36.33/fs/ext2/super.c 2010-10-21 13:09:36.000000000 +0200
3042 @@ -378,7 +378,8 @@ enum {
3043 Opt_err_ro, Opt_nouid32, Opt_nocheck, Opt_debug,
3044 Opt_oldalloc, Opt_orlov, Opt_nobh, Opt_user_xattr, Opt_nouser_xattr,
3045 Opt_acl, Opt_noacl, Opt_xip, Opt_ignore, Opt_err, Opt_quota,
3046 - Opt_usrquota, Opt_grpquota, Opt_reservation, Opt_noreservation
3047 + Opt_usrquota, Opt_grpquota, Opt_reservation, Opt_noreservation,
3048 + Opt_tag, Opt_notag, Opt_tagid
3051 static const match_table_t tokens = {
3052 @@ -406,6 +407,9 @@ static const match_table_t tokens = {
3054 {Opt_noacl, "noacl"},
3057 + {Opt_notag, "notag"},
3058 + {Opt_tagid, "tagid=%u"},
3059 {Opt_grpquota, "grpquota"},
3060 {Opt_ignore, "noquota"},
3061 {Opt_quota, "quota"},
3062 @@ -476,6 +480,20 @@ static int parse_options(char *options,
3064 set_opt (sbi->s_mount_opt, NO_UID32);
3066 +#ifndef CONFIG_TAGGING_NONE
3068 + set_opt (sbi->s_mount_opt, TAGGED);
3071 + clear_opt (sbi->s_mount_opt, TAGGED);
3074 +#ifdef CONFIG_PROPAGATE
3077 + set_opt (sbi->s_mount_opt, TAGGED);
3081 clear_opt (sbi->s_mount_opt, CHECK);
3083 @@ -833,6 +851,8 @@ static int ext2_fill_super(struct super_
3084 if (!parse_options((char *) data, sb))
3087 + if (EXT2_SB(sb)->s_mount_opt & EXT2_MOUNT_TAGGED)
3088 + sb->s_flags |= MS_TAGGED;
3089 sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
3090 ((EXT2_SB(sb)->s_mount_opt & EXT2_MOUNT_POSIX_ACL) ?
3092 @@ -1205,6 +1225,14 @@ static int ext2_remount (struct super_bl
3096 + if ((sbi->s_mount_opt & EXT2_MOUNT_TAGGED) &&
3097 + !(sb->s_flags & MS_TAGGED)) {
3098 + printk("EXT2-fs: %s: tagging not permitted on remount.\n",
3101 + goto restore_opts;
3104 sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
3105 ((sbi->s_mount_opt & EXT2_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0);
3107 diff -NurpP --minimal linux-2.6.36/fs/ext3/file.c linux-2.6.36-vs2.3.0.36.33/fs/ext3/file.c
3108 --- linux-2.6.36/fs/ext3/file.c 2010-07-07 18:31:51.000000000 +0200
3109 +++ linux-2.6.36-vs2.3.0.36.33/fs/ext3/file.c 2010-10-21 13:09:36.000000000 +0200
3110 @@ -81,5 +81,6 @@ const struct inode_operations ext3_file_
3112 .check_acl = ext3_check_acl,
3113 .fiemap = ext3_fiemap,
3114 + .sync_flags = ext3_sync_flags,
3117 diff -NurpP --minimal linux-2.6.36/fs/ext3/ialloc.c linux-2.6.36-vs2.3.0.36.33/fs/ext3/ialloc.c
3118 --- linux-2.6.36/fs/ext3/ialloc.c 2010-10-21 13:07:48.000000000 +0200
3119 +++ linux-2.6.36-vs2.3.0.36.33/fs/ext3/ialloc.c 2010-10-21 13:09:36.000000000 +0200
3121 #include <linux/buffer_head.h>
3122 #include <linux/random.h>
3123 #include <linux/bitops.h>
3124 +#include <linux/vs_tag.h>
3126 #include <asm/byteorder.h>
3128 @@ -531,6 +532,7 @@ got:
3129 inode->i_mode = mode;
3130 inode->i_uid = current_fsuid();
3131 inode->i_gid = dir->i_gid;
3132 + inode->i_tag = dx_current_fstag(sb);
3134 inode_init_owner(inode, dir, mode);
3136 diff -NurpP --minimal linux-2.6.36/fs/ext3/inode.c linux-2.6.36-vs2.3.0.36.33/fs/ext3/inode.c
3137 --- linux-2.6.36/fs/ext3/inode.c 2010-10-21 13:07:48.000000000 +0200
3138 +++ linux-2.6.36-vs2.3.0.36.33/fs/ext3/inode.c 2010-10-21 13:09:36.000000000 +0200
3140 #include <linux/bio.h>
3141 #include <linux/fiemap.h>
3142 #include <linux/namei.h>
3143 +#include <linux/vs_tag.h>
3147 @@ -2387,7 +2388,7 @@ static void ext3_free_branches(handle_t
3149 int ext3_can_truncate(struct inode *inode)
3151 - if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
3152 + if (IS_APPEND(inode) || IS_IXORUNLINK(inode))
3154 if (S_ISREG(inode->i_mode))
3156 @@ -2772,36 +2773,60 @@ void ext3_set_inode_flags(struct inode *
3158 unsigned int flags = EXT3_I(inode)->i_flags;
3160 - inode->i_flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC);
3161 + inode->i_flags &= ~(S_IMMUTABLE | S_IXUNLINK |
3162 + S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
3164 + if (flags & EXT3_IMMUTABLE_FL)
3165 + inode->i_flags |= S_IMMUTABLE;
3166 + if (flags & EXT3_IXUNLINK_FL)
3167 + inode->i_flags |= S_IXUNLINK;
3169 if (flags & EXT3_SYNC_FL)
3170 inode->i_flags |= S_SYNC;
3171 if (flags & EXT3_APPEND_FL)
3172 inode->i_flags |= S_APPEND;
3173 - if (flags & EXT3_IMMUTABLE_FL)
3174 - inode->i_flags |= S_IMMUTABLE;
3175 if (flags & EXT3_NOATIME_FL)
3176 inode->i_flags |= S_NOATIME;
3177 if (flags & EXT3_DIRSYNC_FL)
3178 inode->i_flags |= S_DIRSYNC;
3180 + inode->i_vflags &= ~(V_BARRIER | V_COW);
3182 + if (flags & EXT3_BARRIER_FL)
3183 + inode->i_vflags |= V_BARRIER;
3184 + if (flags & EXT3_COW_FL)
3185 + inode->i_vflags |= V_COW;
3188 /* Propagate flags from i_flags to EXT3_I(inode)->i_flags */
3189 void ext3_get_inode_flags(struct ext3_inode_info *ei)
3191 unsigned int flags = ei->vfs_inode.i_flags;
3192 + unsigned int vflags = ei->vfs_inode.i_vflags;
3194 + ei->i_flags &= ~(EXT3_SYNC_FL | EXT3_APPEND_FL |
3195 + EXT3_IMMUTABLE_FL | EXT3_IXUNLINK_FL |
3196 + EXT3_NOATIME_FL | EXT3_DIRSYNC_FL |
3197 + EXT3_BARRIER_FL | EXT3_COW_FL);
3199 + if (flags & S_IMMUTABLE)
3200 + ei->i_flags |= EXT3_IMMUTABLE_FL;
3201 + if (flags & S_IXUNLINK)
3202 + ei->i_flags |= EXT3_IXUNLINK_FL;
3204 - ei->i_flags &= ~(EXT3_SYNC_FL|EXT3_APPEND_FL|
3205 - EXT3_IMMUTABLE_FL|EXT3_NOATIME_FL|EXT3_DIRSYNC_FL);
3207 ei->i_flags |= EXT3_SYNC_FL;
3208 if (flags & S_APPEND)
3209 ei->i_flags |= EXT3_APPEND_FL;
3210 - if (flags & S_IMMUTABLE)
3211 - ei->i_flags |= EXT3_IMMUTABLE_FL;
3212 if (flags & S_NOATIME)
3213 ei->i_flags |= EXT3_NOATIME_FL;
3214 if (flags & S_DIRSYNC)
3215 ei->i_flags |= EXT3_DIRSYNC_FL;
3217 + if (vflags & V_BARRIER)
3218 + ei->i_flags |= EXT3_BARRIER_FL;
3219 + if (vflags & V_COW)
3220 + ei->i_flags |= EXT3_COW_FL;
3223 struct inode *ext3_iget(struct super_block *sb, unsigned long ino)
3224 @@ -2815,6 +2840,8 @@ struct inode *ext3_iget(struct super_blo
3225 transaction_t *transaction;
3231 inode = iget_locked(sb, ino);
3233 @@ -2831,12 +2858,17 @@ struct inode *ext3_iget(struct super_blo
3235 raw_inode = ext3_raw_inode(&iloc);
3236 inode->i_mode = le16_to_cpu(raw_inode->i_mode);
3237 - inode->i_uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
3238 - inode->i_gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
3239 + uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
3240 + gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
3241 if(!(test_opt (inode->i_sb, NO_UID32))) {
3242 - inode->i_uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
3243 - inode->i_gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
3244 + uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
3245 + gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
3247 + inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
3248 + inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
3249 + inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid,
3250 + le16_to_cpu(raw_inode->i_raw_tag));
3252 inode->i_nlink = le16_to_cpu(raw_inode->i_links_count);
3253 inode->i_size = le32_to_cpu(raw_inode->i_size);
3254 inode->i_atime.tv_sec = (signed)le32_to_cpu(raw_inode->i_atime);
3255 @@ -2991,6 +3023,8 @@ static int ext3_do_update_inode(handle_t
3256 struct ext3_inode *raw_inode = ext3_raw_inode(iloc);
3257 struct ext3_inode_info *ei = EXT3_I(inode);
3258 struct buffer_head *bh = iloc->bh;
3259 + uid_t uid = TAGINO_UID(DX_TAG(inode), inode->i_uid, inode->i_tag);
3260 + gid_t gid = TAGINO_GID(DX_TAG(inode), inode->i_gid, inode->i_tag);
3261 int err = 0, rc, block;
3264 @@ -3005,29 +3039,32 @@ again:
3265 ext3_get_inode_flags(ei);
3266 raw_inode->i_mode = cpu_to_le16(inode->i_mode);
3267 if(!(test_opt(inode->i_sb, NO_UID32))) {
3268 - raw_inode->i_uid_low = cpu_to_le16(low_16_bits(inode->i_uid));
3269 - raw_inode->i_gid_low = cpu_to_le16(low_16_bits(inode->i_gid));
3270 + raw_inode->i_uid_low = cpu_to_le16(low_16_bits(uid));
3271 + raw_inode->i_gid_low = cpu_to_le16(low_16_bits(gid));
3273 * Fix up interoperability with old kernels. Otherwise, old inodes get
3274 * re-used with the upper 16 bits of the uid/gid intact
3277 raw_inode->i_uid_high =
3278 - cpu_to_le16(high_16_bits(inode->i_uid));
3279 + cpu_to_le16(high_16_bits(uid));
3280 raw_inode->i_gid_high =
3281 - cpu_to_le16(high_16_bits(inode->i_gid));
3282 + cpu_to_le16(high_16_bits(gid));
3284 raw_inode->i_uid_high = 0;
3285 raw_inode->i_gid_high = 0;
3288 raw_inode->i_uid_low =
3289 - cpu_to_le16(fs_high2lowuid(inode->i_uid));
3290 + cpu_to_le16(fs_high2lowuid(uid));
3291 raw_inode->i_gid_low =
3292 - cpu_to_le16(fs_high2lowgid(inode->i_gid));
3293 + cpu_to_le16(fs_high2lowgid(gid));
3294 raw_inode->i_uid_high = 0;
3295 raw_inode->i_gid_high = 0;
3297 +#ifdef CONFIG_TAGGING_INTERN
3298 + raw_inode->i_raw_tag = cpu_to_le16(inode->i_tag);
3300 raw_inode->i_links_count = cpu_to_le16(inode->i_nlink);
3301 raw_inode->i_size = cpu_to_le32(ei->i_disksize);
3302 raw_inode->i_atime = cpu_to_le32(inode->i_atime.tv_sec);
3303 @@ -3187,7 +3224,8 @@ int ext3_setattr(struct dentry *dentry,
3304 if (is_quota_modification(inode, attr))
3305 dquot_initialize(inode);
3306 if ((ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) ||
3307 - (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) {
3308 + (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid) ||
3309 + (ia_valid & ATTR_TAG && attr->ia_tag != inode->i_tag)) {
3312 /* (user+group)*(old+new) structure, inode write (sb,
3313 @@ -3209,6 +3247,8 @@ int ext3_setattr(struct dentry *dentry,
3314 inode->i_uid = attr->ia_uid;
3315 if (attr->ia_valid & ATTR_GID)
3316 inode->i_gid = attr->ia_gid;
3317 + if ((attr->ia_valid & ATTR_TAG) && IS_TAGGED(inode))
3318 + inode->i_tag = attr->ia_tag;
3319 error = ext3_mark_inode_dirty(handle, inode);
3320 ext3_journal_stop(handle);
3322 diff -NurpP --minimal linux-2.6.36/fs/ext3/ioctl.c linux-2.6.36-vs2.3.0.36.33/fs/ext3/ioctl.c
3323 --- linux-2.6.36/fs/ext3/ioctl.c 2009-06-11 17:13:03.000000000 +0200
3324 +++ linux-2.6.36-vs2.3.0.36.33/fs/ext3/ioctl.c 2010-10-21 13:09:36.000000000 +0200
3328 #include <linux/fs.h>
3329 +#include <linux/mount.h>
3330 #include <linux/jbd.h>
3331 #include <linux/capability.h>
3332 #include <linux/ext3_fs.h>
3334 #include <linux/compat.h>
3335 #include <asm/uaccess.h>
3338 +int ext3_sync_flags(struct inode *inode, int flags, int vflags)
3340 + handle_t *handle = NULL;
3341 + struct ext3_iloc iloc;
3344 + handle = ext3_journal_start(inode, 1);
3345 + if (IS_ERR(handle))
3346 + return PTR_ERR(handle);
3348 + if (IS_SYNC(inode))
3349 + handle->h_sync = 1;
3350 + err = ext3_reserve_inode_write(handle, inode, &iloc);
3354 + inode->i_flags = flags;
3355 + inode->i_vflags = vflags;
3356 + ext3_get_inode_flags(EXT3_I(inode));
3357 + inode->i_ctime = CURRENT_TIME_SEC;
3359 + err = ext3_mark_iloc_dirty(handle, inode, &iloc);
3361 + ext3_journal_stop(handle);
3365 long ext3_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
3367 struct inode *inode = filp->f_dentry->d_inode;
3368 @@ -50,6 +79,11 @@ long ext3_ioctl(struct file *filp, unsig
3370 flags = ext3_mask_flags(inode->i_mode, flags);
3372 + if (IS_BARRIER(inode)) {
3373 + vxwprintk_task(1, "messing with the barrier.");
3377 mutex_lock(&inode->i_mutex);
3379 /* Is it quota file? Do not allow user to mess with it */
3380 @@ -68,7 +102,9 @@ long ext3_ioctl(struct file *filp, unsig
3382 * This test looks nicer. Thanks to Pauline Middelink
3384 - if ((flags ^ oldflags) & (EXT3_APPEND_FL | EXT3_IMMUTABLE_FL)) {
3385 + if ((oldflags & EXT3_IMMUTABLE_FL) ||
3386 + ((flags ^ oldflags) & (EXT3_APPEND_FL |
3387 + EXT3_IMMUTABLE_FL | EXT3_IXUNLINK_FL))) {
3388 if (!capable(CAP_LINUX_IMMUTABLE))
3391 @@ -93,7 +129,7 @@ long ext3_ioctl(struct file *filp, unsig
3395 - flags = flags & EXT3_FL_USER_MODIFIABLE;
3396 + flags &= EXT3_FL_USER_MODIFIABLE;
3397 flags |= oldflags & ~EXT3_FL_USER_MODIFIABLE;
3398 ei->i_flags = flags;
3400 diff -NurpP --minimal linux-2.6.36/fs/ext3/namei.c linux-2.6.36-vs2.3.0.36.33/fs/ext3/namei.c
3401 --- linux-2.6.36/fs/ext3/namei.c 2010-10-21 13:07:48.000000000 +0200
3402 +++ linux-2.6.36-vs2.3.0.36.33/fs/ext3/namei.c 2010-10-21 13:09:36.000000000 +0200
3404 #include <linux/quotaops.h>
3405 #include <linux/buffer_head.h>
3406 #include <linux/bio.h>
3407 +#include <linux/vs_tag.h>
3411 @@ -912,6 +913,7 @@ restart:
3413 ll_rw_block(READ_META, 1, &bh);
3415 + dx_propagate_tag(nd, inode);
3417 if ((bh = bh_use[ra_ptr++]) == NULL)
3419 @@ -2465,6 +2467,7 @@ const struct inode_operations ext3_dir_i
3420 .removexattr = generic_removexattr,
3422 .check_acl = ext3_check_acl,
3423 + .sync_flags = ext3_sync_flags,
3426 const struct inode_operations ext3_special_inode_operations = {
3427 diff -NurpP --minimal linux-2.6.36/fs/ext3/super.c linux-2.6.36-vs2.3.0.36.33/fs/ext3/super.c
3428 --- linux-2.6.36/fs/ext3/super.c 2010-10-21 13:07:48.000000000 +0200
3429 +++ linux-2.6.36-vs2.3.0.36.33/fs/ext3/super.c 2010-10-21 13:09:36.000000000 +0200
3430 @@ -802,7 +802,8 @@ enum {
3431 Opt_usrjquota, Opt_grpjquota, Opt_offusrjquota, Opt_offgrpjquota,
3432 Opt_jqfmt_vfsold, Opt_jqfmt_vfsv0, Opt_jqfmt_vfsv1, Opt_quota,
3433 Opt_noquota, Opt_ignore, Opt_barrier, Opt_nobarrier, Opt_err,
3434 - Opt_resize, Opt_usrquota, Opt_grpquota
3435 + Opt_resize, Opt_usrquota, Opt_grpquota,
3436 + Opt_tag, Opt_notag, Opt_tagid
3439 static const match_table_t tokens = {
3440 @@ -859,6 +860,9 @@ static const match_table_t tokens = {
3441 {Opt_barrier, "barrier"},
3442 {Opt_nobarrier, "nobarrier"},
3443 {Opt_resize, "resize"},
3445 + {Opt_notag, "notag"},
3446 + {Opt_tagid, "tagid=%u"},
3450 @@ -1011,6 +1015,20 @@ static int parse_options (char *options,
3452 set_opt (sbi->s_mount_opt, NO_UID32);
3454 +#ifndef CONFIG_TAGGING_NONE
3456 + set_opt (sbi->s_mount_opt, TAGGED);
3459 + clear_opt (sbi->s_mount_opt, TAGGED);
3462 +#ifdef CONFIG_PROPAGATE
3465 + set_opt (sbi->s_mount_opt, TAGGED);
3469 clear_opt (sbi->s_mount_opt, CHECK);
3471 @@ -1699,6 +1717,9 @@ static int ext3_fill_super (struct super
3475 + if (EXT3_SB(sb)->s_mount_opt & EXT3_MOUNT_TAGGED)
3476 + sb->s_flags |= MS_TAGGED;
3478 sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
3479 (test_opt(sb, POSIX_ACL) ? MS_POSIXACL : 0);
3481 @@ -2564,6 +2585,14 @@ static int ext3_remount (struct super_bl
3482 if (test_opt(sb, ABORT))
3483 ext3_abort(sb, __func__, "Abort forced by user");
3485 + if ((sbi->s_mount_opt & EXT3_MOUNT_TAGGED) &&
3486 + !(sb->s_flags & MS_TAGGED)) {
3487 + printk("EXT3-fs: %s: tagging not permitted on remount.\n",
3490 + goto restore_opts;
3493 sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
3494 (test_opt(sb, POSIX_ACL) ? MS_POSIXACL : 0);
3496 diff -NurpP --minimal linux-2.6.36/fs/ext4/ext4.h linux-2.6.36-vs2.3.0.36.33/fs/ext4/ext4.h
3497 --- linux-2.6.36/fs/ext4/ext4.h 2010-10-21 13:07:48.000000000 +0200
3498 +++ linux-2.6.36-vs2.3.0.36.33/fs/ext4/ext4.h 2010-10-21 13:09:36.000000000 +0200
3499 @@ -323,8 +323,12 @@ struct flex_groups {
3500 #define EXT4_EXTENTS_FL 0x00080000 /* Inode uses extents */
3501 #define EXT4_EA_INODE_FL 0x00200000 /* Inode used for large EA */
3502 #define EXT4_EOFBLOCKS_FL 0x00400000 /* Blocks allocated beyond EOF */
3503 +#define EXT4_IXUNLINK_FL 0x08000000 /* Immutable invert on unlink */
3504 #define EXT4_RESERVED_FL 0x80000000 /* reserved for ext4 lib */
3506 +#define EXT4_BARRIER_FL 0x04000000 /* Barrier for chroot() */
3507 +#define EXT4_COW_FL 0x20000000 /* Copy on Write marker */
3509 #define EXT4_FL_USER_VISIBLE 0x004BDFFF /* User visible flags */
3510 #define EXT4_FL_USER_MODIFIABLE 0x004B80FF /* User modifiable flags */
3512 @@ -593,7 +597,8 @@ struct ext4_inode {
3513 __le16 l_i_file_acl_high;
3514 __le16 l_i_uid_high; /* these 2 fields */
3515 __le16 l_i_gid_high; /* were reserved2[0] */
3516 - __u32 l_i_reserved2;
3517 + __le16 l_i_tag; /* Context Tag */
3518 + __u16 l_i_reserved2;
3521 __le16 h_i_reserved1; /* Obsoleted fragment number/size which are removed in ext4 */
3522 @@ -707,6 +712,7 @@ do { \
3523 #define i_gid_low i_gid
3524 #define i_uid_high osd2.linux2.l_i_uid_high
3525 #define i_gid_high osd2.linux2.l_i_gid_high
3526 +#define i_raw_tag osd2.linux2.l_i_tag
3527 #define i_reserved2 osd2.linux2.l_i_reserved2
3529 #elif defined(__GNU__)
3530 @@ -885,6 +891,7 @@ struct ext4_inode_info {
3531 #define EXT4_MOUNT_JOURNAL_CHECKSUM 0x800000 /* Journal checksums */
3532 #define EXT4_MOUNT_JOURNAL_ASYNC_COMMIT 0x1000000 /* Journal Async Commit */
3533 #define EXT4_MOUNT_I_VERSION 0x2000000 /* i_version support */
3534 +#define EXT4_MOUNT_TAGGED 0x4000000 /* Enable Context Tags */
3535 #define EXT4_MOUNT_DELALLOC 0x8000000 /* Delalloc support */
3536 #define EXT4_MOUNT_DATA_ERR_ABORT 0x10000000 /* Abort on file data write */
3537 #define EXT4_MOUNT_BLOCK_VALIDITY 0x20000000 /* Block validity checking */
3538 @@ -1997,6 +2004,7 @@ extern int ext4_map_blocks(handle_t *han
3539 struct ext4_map_blocks *map, int flags);
3540 extern int ext4_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
3541 __u64 start, __u64 len);
3542 +extern int ext4_sync_flags(struct inode *, int, int);
3544 extern int ext4_move_extents(struct file *o_filp, struct file *d_filp,
3545 __u64 start_orig, __u64 start_donor,
3546 diff -NurpP --minimal linux-2.6.36/fs/ext4/file.c linux-2.6.36-vs2.3.0.36.33/fs/ext4/file.c
3547 --- linux-2.6.36/fs/ext4/file.c 2010-10-21 13:07:48.000000000 +0200
3548 +++ linux-2.6.36-vs2.3.0.36.33/fs/ext4/file.c 2010-10-21 13:09:36.000000000 +0200
3549 @@ -161,5 +161,6 @@ const struct inode_operations ext4_file_
3550 .check_acl = ext4_check_acl,
3551 .fallocate = ext4_fallocate,
3552 .fiemap = ext4_fiemap,
3553 + .sync_flags = ext4_sync_flags,
3556 diff -NurpP --minimal linux-2.6.36/fs/ext4/ialloc.c linux-2.6.36-vs2.3.0.36.33/fs/ext4/ialloc.c
3557 --- linux-2.6.36/fs/ext4/ialloc.c 2010-10-21 13:07:48.000000000 +0200
3558 +++ linux-2.6.36-vs2.3.0.36.33/fs/ext4/ialloc.c 2010-10-21 13:09:36.000000000 +0200
3560 #include <linux/random.h>
3561 #include <linux/bitops.h>
3562 #include <linux/blkdev.h>
3563 +#include <linux/vs_tag.h>
3564 #include <asm/byteorder.h>
3567 @@ -976,6 +977,7 @@ got:
3568 inode->i_mode = mode;
3569 inode->i_uid = current_fsuid();
3570 inode->i_gid = dir->i_gid;
3571 + inode->i_tag = dx_current_fstag(sb);
3573 inode_init_owner(inode, dir, mode);
3575 diff -NurpP --minimal linux-2.6.36/fs/ext4/inode.c linux-2.6.36-vs2.3.0.36.33/fs/ext4/inode.c
3576 --- linux-2.6.36/fs/ext4/inode.c 2010-10-21 13:07:48.000000000 +0200
3577 +++ linux-2.6.36-vs2.3.0.36.33/fs/ext4/inode.c 2010-10-21 13:09:36.000000000 +0200
3579 #include <linux/workqueue.h>
3580 #include <linux/kernel.h>
3581 #include <linux/slab.h>
3582 +#include <linux/vs_tag.h>
3584 #include "ext4_jbd2.h"
3586 @@ -4596,7 +4597,7 @@ static void ext4_free_branches(handle_t
3588 int ext4_can_truncate(struct inode *inode)
3590 - if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
3591 + if (IS_APPEND(inode) || IS_IXORUNLINK(inode))
3593 if (S_ISREG(inode->i_mode))
3595 @@ -4947,41 +4948,64 @@ void ext4_set_inode_flags(struct inode *
3597 unsigned int flags = EXT4_I(inode)->i_flags;
3599 - inode->i_flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC);
3600 + inode->i_flags &= ~(S_IMMUTABLE | S_IXUNLINK |
3601 + S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
3603 + if (flags & EXT4_IMMUTABLE_FL)
3604 + inode->i_flags |= S_IMMUTABLE;
3605 + if (flags & EXT4_IXUNLINK_FL)
3606 + inode->i_flags |= S_IXUNLINK;
3608 if (flags & EXT4_SYNC_FL)
3609 inode->i_flags |= S_SYNC;
3610 if (flags & EXT4_APPEND_FL)
3611 inode->i_flags |= S_APPEND;
3612 - if (flags & EXT4_IMMUTABLE_FL)
3613 - inode->i_flags |= S_IMMUTABLE;
3614 if (flags & EXT4_NOATIME_FL)
3615 inode->i_flags |= S_NOATIME;
3616 if (flags & EXT4_DIRSYNC_FL)
3617 inode->i_flags |= S_DIRSYNC;
3619 + inode->i_vflags &= ~(V_BARRIER | V_COW);
3621 + if (flags & EXT4_BARRIER_FL)
3622 + inode->i_vflags |= V_BARRIER;
3623 + if (flags & EXT4_COW_FL)
3624 + inode->i_vflags |= V_COW;
3627 /* Propagate flags from i_flags to EXT4_I(inode)->i_flags */
3628 void ext4_get_inode_flags(struct ext4_inode_info *ei)
3630 - unsigned int vfs_fl;
3631 + unsigned int vfs_fl, vfs_vf;
3632 unsigned long old_fl, new_fl;
3635 vfs_fl = ei->vfs_inode.i_flags;
3636 + vfs_vf = ei->vfs_inode.i_vflags;
3637 old_fl = ei->i_flags;
3638 new_fl = old_fl & ~(EXT4_SYNC_FL|EXT4_APPEND_FL|
3639 EXT4_IMMUTABLE_FL|EXT4_NOATIME_FL|
3641 + EXT4_DIRSYNC_FL|EXT4_BARRIER_FL|
3644 + if (vfs_fl & S_IMMUTABLE)
3645 + new_fl |= EXT4_IMMUTABLE_FL;
3646 + if (vfs_fl & S_IXUNLINK)
3647 + new_fl |= EXT4_IXUNLINK_FL;
3649 if (vfs_fl & S_SYNC)
3650 new_fl |= EXT4_SYNC_FL;
3651 if (vfs_fl & S_APPEND)
3652 new_fl |= EXT4_APPEND_FL;
3653 - if (vfs_fl & S_IMMUTABLE)
3654 - new_fl |= EXT4_IMMUTABLE_FL;
3655 if (vfs_fl & S_NOATIME)
3656 new_fl |= EXT4_NOATIME_FL;
3657 if (vfs_fl & S_DIRSYNC)
3658 new_fl |= EXT4_DIRSYNC_FL;
3660 + if (vfs_vf & V_BARRIER)
3661 + new_fl |= EXT4_BARRIER_FL;
3662 + if (vfs_vf & V_COW)
3663 + new_fl |= EXT4_COW_FL;
3664 } while (cmpxchg(&ei->i_flags, old_fl, new_fl) != old_fl);
3667 @@ -5017,6 +5041,8 @@ struct inode *ext4_iget(struct super_blo
3668 journal_t *journal = EXT4_SB(sb)->s_journal;
3674 inode = iget_locked(sb, ino);
3676 @@ -5032,12 +5058,16 @@ struct inode *ext4_iget(struct super_blo
3678 raw_inode = ext4_raw_inode(&iloc);
3679 inode->i_mode = le16_to_cpu(raw_inode->i_mode);
3680 - inode->i_uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
3681 - inode->i_gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
3682 + uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
3683 + gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
3684 if (!(test_opt(inode->i_sb, NO_UID32))) {
3685 - inode->i_uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
3686 - inode->i_gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
3687 + uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
3688 + gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
3690 + inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
3691 + inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
3692 + inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid,
3693 + le16_to_cpu(raw_inode->i_raw_tag));
3694 inode->i_nlink = le16_to_cpu(raw_inode->i_links_count);
3696 ei->i_state_flags = 0;
3697 @@ -5256,6 +5286,8 @@ static int ext4_do_update_inode(handle_t
3698 struct ext4_inode *raw_inode = ext4_raw_inode(iloc);
3699 struct ext4_inode_info *ei = EXT4_I(inode);
3700 struct buffer_head *bh = iloc->bh;
3701 + uid_t uid = TAGINO_UID(DX_TAG(inode), inode->i_uid, inode->i_tag);
3702 + gid_t gid = TAGINO_GID(DX_TAG(inode), inode->i_gid, inode->i_tag);
3703 int err = 0, rc, block;
3705 /* For fields not not tracking in the in-memory inode,
3706 @@ -5266,29 +5298,32 @@ static int ext4_do_update_inode(handle_t
3707 ext4_get_inode_flags(ei);
3708 raw_inode->i_mode = cpu_to_le16(inode->i_mode);
3709 if (!(test_opt(inode->i_sb, NO_UID32))) {
3710 - raw_inode->i_uid_low = cpu_to_le16(low_16_bits(inode->i_uid));
3711 - raw_inode->i_gid_low = cpu_to_le16(low_16_bits(inode->i_gid));
3712 + raw_inode->i_uid_low = cpu_to_le16(low_16_bits(uid));
3713 + raw_inode->i_gid_low = cpu_to_le16(low_16_bits(gid));
3715 * Fix up interoperability with old kernels. Otherwise, old inodes get
3716 * re-used with the upper 16 bits of the uid/gid intact
3719 raw_inode->i_uid_high =
3720 - cpu_to_le16(high_16_bits(inode->i_uid));
3721 + cpu_to_le16(high_16_bits(uid));
3722 raw_inode->i_gid_high =
3723 - cpu_to_le16(high_16_bits(inode->i_gid));
3724 + cpu_to_le16(high_16_bits(gid));
3726 raw_inode->i_uid_high = 0;
3727 raw_inode->i_gid_high = 0;
3730 raw_inode->i_uid_low =
3731 - cpu_to_le16(fs_high2lowuid(inode->i_uid));
3732 + cpu_to_le16(fs_high2lowuid(uid));
3733 raw_inode->i_gid_low =
3734 - cpu_to_le16(fs_high2lowgid(inode->i_gid));
3735 + cpu_to_le16(fs_high2lowgid(gid));
3736 raw_inode->i_uid_high = 0;
3737 raw_inode->i_gid_high = 0;
3739 +#ifdef CONFIG_TAGGING_INTERN
3740 + raw_inode->i_raw_tag = cpu_to_le16(inode->i_tag);
3742 raw_inode->i_links_count = cpu_to_le16(inode->i_nlink);
3744 EXT4_INODE_SET_XTIME(i_ctime, inode, raw_inode);
3745 @@ -5473,7 +5508,8 @@ int ext4_setattr(struct dentry *dentry,
3746 if (is_quota_modification(inode, attr))
3747 dquot_initialize(inode);
3748 if ((ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) ||
3749 - (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) {
3750 + (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid) ||
3751 + (ia_valid & ATTR_TAG && attr->ia_tag != inode->i_tag)) {
3754 /* (user+group)*(old+new) structure, inode write (sb,
3755 @@ -5495,6 +5531,8 @@ int ext4_setattr(struct dentry *dentry,
3756 inode->i_uid = attr->ia_uid;
3757 if (attr->ia_valid & ATTR_GID)
3758 inode->i_gid = attr->ia_gid;
3759 + if ((attr->ia_valid & ATTR_TAG) && IS_TAGGED(inode))
3760 + inode->i_tag = attr->ia_tag;
3761 error = ext4_mark_inode_dirty(handle, inode);
3762 ext4_journal_stop(handle);
3764 diff -NurpP --minimal linux-2.6.36/fs/ext4/ioctl.c linux-2.6.36-vs2.3.0.36.33/fs/ext4/ioctl.c
3765 --- linux-2.6.36/fs/ext4/ioctl.c 2010-08-02 16:52:48.000000000 +0200
3766 +++ linux-2.6.36-vs2.3.0.36.33/fs/ext4/ioctl.c 2010-10-21 13:09:36.000000000 +0200
3768 #include <linux/compat.h>
3769 #include <linux/mount.h>
3770 #include <linux/file.h>
3771 +#include <linux/vs_tag.h>
3772 #include <asm/uaccess.h>
3773 #include "ext4_jbd2.h"
3777 +int ext4_sync_flags(struct inode *inode, int flags, int vflags)
3779 + handle_t *handle = NULL;
3780 + struct ext4_iloc iloc;
3783 + handle = ext4_journal_start(inode, 1);
3784 + if (IS_ERR(handle))
3785 + return PTR_ERR(handle);
3787 + if (IS_SYNC(inode))
3788 + ext4_handle_sync(handle);
3789 + err = ext4_reserve_inode_write(handle, inode, &iloc);
3793 + inode->i_flags = flags;
3794 + inode->i_vflags = vflags;
3795 + ext4_get_inode_flags(EXT4_I(inode));
3796 + inode->i_ctime = ext4_current_time(inode);
3798 + err = ext4_mark_iloc_dirty(handle, inode, &iloc);
3800 + ext4_journal_stop(handle);
3804 long ext4_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
3806 struct inode *inode = filp->f_dentry->d_inode;
3807 @@ -50,6 +79,11 @@ long ext4_ioctl(struct file *filp, unsig
3809 flags = ext4_mask_flags(inode->i_mode, flags);
3811 + if (IS_BARRIER(inode)) {
3812 + vxwprintk_task(1, "messing with the barrier.");
3817 mutex_lock(&inode->i_mutex);
3818 /* Is it quota file? Do not allow user to mess with it */
3819 @@ -67,7 +101,9 @@ long ext4_ioctl(struct file *filp, unsig
3821 * This test looks nicer. Thanks to Pauline Middelink
3823 - if ((flags ^ oldflags) & (EXT4_APPEND_FL | EXT4_IMMUTABLE_FL)) {
3824 + if ((oldflags & EXT4_IMMUTABLE_FL) ||
3825 + ((flags ^ oldflags) & (EXT4_APPEND_FL |
3826 + EXT4_IMMUTABLE_FL | EXT4_IXUNLINK_FL))) {
3827 if (!capable(CAP_LINUX_IMMUTABLE))
3830 diff -NurpP --minimal linux-2.6.36/fs/ext4/namei.c linux-2.6.36-vs2.3.0.36.33/fs/ext4/namei.c
3831 --- linux-2.6.36/fs/ext4/namei.c 2010-10-21 13:07:48.000000000 +0200
3832 +++ linux-2.6.36-vs2.3.0.36.33/fs/ext4/namei.c 2010-10-21 13:09:36.000000000 +0200
3834 #include <linux/quotaops.h>
3835 #include <linux/buffer_head.h>
3836 #include <linux/bio.h>
3837 +#include <linux/vs_tag.h>
3839 #include "ext4_jbd2.h"
3841 @@ -912,6 +913,7 @@ restart:
3843 ll_rw_block(READ_META, 1, &bh);
3845 + dx_propagate_tag(nd, inode);
3847 if ((bh = bh_use[ra_ptr++]) == NULL)
3849 @@ -2522,6 +2524,7 @@ const struct inode_operations ext4_dir_i
3851 .check_acl = ext4_check_acl,
3852 .fiemap = ext4_fiemap,
3853 + .sync_flags = ext4_sync_flags,
3856 const struct inode_operations ext4_special_inode_operations = {
3857 diff -NurpP --minimal linux-2.6.36/fs/ext4/super.c linux-2.6.36-vs2.3.0.36.33/fs/ext4/super.c
3858 --- linux-2.6.36/fs/ext4/super.c 2010-10-21 13:07:48.000000000 +0200
3859 +++ linux-2.6.36-vs2.3.0.36.33/fs/ext4/super.c 2010-10-21 13:09:36.000000000 +0200
3860 @@ -1219,6 +1219,7 @@ enum {
3861 Opt_inode_readahead_blks, Opt_journal_ioprio,
3862 Opt_dioread_nolock, Opt_dioread_lock,
3863 Opt_discard, Opt_nodiscard,
3864 + Opt_tag, Opt_notag, Opt_tagid
3867 static const match_table_t tokens = {
3868 @@ -1289,6 +1290,9 @@ static const match_table_t tokens = {
3869 {Opt_dioread_lock, "dioread_lock"},
3870 {Opt_discard, "discard"},
3871 {Opt_nodiscard, "nodiscard"},
3873 + {Opt_notag, "notag"},
3874 + {Opt_tagid, "tagid=%u"},
3878 @@ -1457,6 +1461,20 @@ static int parse_options(char *options,
3880 set_opt(sbi->s_mount_opt, NO_UID32);
3882 +#ifndef CONFIG_TAGGING_NONE
3884 + set_opt (sbi->s_mount_opt, TAGGED);
3887 + clear_opt (sbi->s_mount_opt, TAGGED);
3890 +#ifdef CONFIG_PROPAGATE
3893 + set_opt (sbi->s_mount_opt, TAGGED);
3897 set_opt(sbi->s_mount_opt, DEBUG);
3899 @@ -2691,6 +2709,9 @@ static int ext4_fill_super(struct super_
3900 &journal_ioprio, NULL, 0))
3903 + if (EXT4_SB(sb)->s_mount_opt & EXT4_MOUNT_TAGGED)
3904 + sb->s_flags |= MS_TAGGED;
3906 sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
3907 (test_opt(sb, POSIX_ACL) ? MS_POSIXACL : 0);
3909 @@ -3753,6 +3774,14 @@ static int ext4_remount(struct super_blo
3910 if (sbi->s_mount_flags & EXT4_MF_FS_ABORTED)
3911 ext4_abort(sb, "Abort forced by user");
3913 + if ((sbi->s_mount_opt & EXT4_MOUNT_TAGGED) &&
3914 + !(sb->s_flags & MS_TAGGED)) {
3915 + printk("EXT4-fs: %s: tagging not permitted on remount.\n",
3918 + goto restore_opts;
3921 sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
3922 (test_opt(sb, POSIX_ACL) ? MS_POSIXACL : 0);
3924 diff -NurpP --minimal linux-2.6.36/fs/fcntl.c linux-2.6.36-vs2.3.0.36.33/fs/fcntl.c
3925 --- linux-2.6.36/fs/fcntl.c 2010-10-21 13:07:48.000000000 +0200
3926 +++ linux-2.6.36-vs2.3.0.36.33/fs/fcntl.c 2010-10-21 13:09:36.000000000 +0200
3928 #include <linux/signal.h>
3929 #include <linux/rcupdate.h>
3930 #include <linux/pid_namespace.h>
3931 +#include <linux/vs_limit.h>
3933 #include <asm/poll.h>
3934 #include <asm/siginfo.h>
3935 @@ -103,6 +104,8 @@ SYSCALL_DEFINE3(dup3, unsigned int, oldf
3938 filp_close(tofree, files);
3940 + vx_openfd_inc(newfd); /* fd was unused */
3944 @@ -434,6 +437,8 @@ SYSCALL_DEFINE3(fcntl, unsigned int, fd,
3948 + if (!vx_files_avail(1))
3951 err = security_file_fcntl(filp, cmd, arg);
3953 diff -NurpP --minimal linux-2.6.36/fs/file.c linux-2.6.36-vs2.3.0.36.33/fs/file.c
3954 --- linux-2.6.36/fs/file.c 2010-10-21 13:07:48.000000000 +0200
3955 +++ linux-2.6.36-vs2.3.0.36.33/fs/file.c 2010-10-21 13:09:36.000000000 +0200
3957 #include <linux/spinlock.h>
3958 #include <linux/rcupdate.h>
3959 #include <linux/workqueue.h>
3960 +#include <linux/vs_limit.h>
3962 struct fdtable_defer {
3964 @@ -355,6 +356,8 @@ struct files_struct *dup_fd(struct files
3965 struct file *f = *old_fds++;
3968 + /* TODO: sum it first for check and performance */
3969 + vx_openfd_inc(open_files - i);
3972 * The fd may be claimed in the fd bitmap but not yet
3973 @@ -462,6 +465,7 @@ repeat:
3975 FD_CLR(fd, fdt->close_on_exec);
3977 + vx_openfd_inc(fd);
3980 if (rcu_dereference_raw(fdt->fd[fd]) != NULL) {
3981 diff -NurpP --minimal linux-2.6.36/fs/file_table.c linux-2.6.36-vs2.3.0.36.33/fs/file_table.c
3982 --- linux-2.6.36/fs/file_table.c 2010-10-21 13:07:48.000000000 +0200
3983 +++ linux-2.6.36-vs2.3.0.36.33/fs/file_table.c 2010-10-21 14:10:28.000000000 +0200
3985 #include <linux/percpu_counter.h>
3986 #include <linux/percpu.h>
3987 #include <linux/ima.h>
3988 +#include <linux/vs_limit.h>
3989 +#include <linux/vs_context.h>
3991 #include <asm/atomic.h>
3993 @@ -135,6 +137,8 @@ struct file *get_empty_filp(void)
3994 spin_lock_init(&f->f_lock);
3995 eventpoll_init_file(f);
3996 /* f->f_version: 0 */
3997 + f->f_xid = vx_current_xid();
4002 @@ -251,6 +255,8 @@ static void __fput(struct file *file)
4003 cdev_put(inode->i_cdev);
4004 fops_put(file->f_op);
4005 put_pid(file->f_owner.pid);
4006 + vx_files_dec(file);
4008 file_sb_list_del(file);
4009 if (file->f_mode & FMODE_WRITE)
4010 drop_file_write_access(file);
4011 @@ -334,6 +340,8 @@ void put_filp(struct file *file)
4013 if (atomic_long_dec_and_test(&file->f_count)) {
4014 security_file_free(file);
4015 + vx_files_dec(file);
4017 file_sb_list_del(file);
4020 diff -NurpP --minimal linux-2.6.36/fs/fs_struct.c linux-2.6.36-vs2.3.0.36.33/fs/fs_struct.c
4021 --- linux-2.6.36/fs/fs_struct.c 2010-10-21 13:07:48.000000000 +0200
4022 +++ linux-2.6.36-vs2.3.0.36.33/fs/fs_struct.c 2010-10-21 13:56:13.000000000 +0200
4024 #include <linux/path.h>
4025 #include <linux/slab.h>
4026 #include <linux/fs_struct.h>
4027 +#include <linux/vserver/global.h>
4030 * Replace the fs->{rootmnt,root} with {mnt,dentry}. Put the old values.
4031 @@ -77,6 +78,7 @@ void free_fs_struct(struct fs_struct *fs
4033 path_put(&fs->root);
4035 + atomic_dec(&vs_global_fs);
4036 kmem_cache_free(fs_cachep, fs);
4039 @@ -107,6 +109,7 @@ struct fs_struct *copy_fs_struct(struct
4040 spin_lock_init(&fs->lock);
4041 fs->umask = old->umask;
4042 get_fs_root_and_pwd(old, &fs->root, &fs->pwd);
4043 + atomic_inc(&vs_global_fs);
4047 diff -NurpP --minimal linux-2.6.36/fs/gfs2/file.c linux-2.6.36-vs2.3.0.36.33/fs/gfs2/file.c
4048 --- linux-2.6.36/fs/gfs2/file.c 2010-10-21 13:07:49.000000000 +0200
4049 +++ linux-2.6.36-vs2.3.0.36.33/fs/gfs2/file.c 2010-10-21 13:09:36.000000000 +0200
4050 @@ -132,6 +132,9 @@ static const u32 fsflags_to_gfs2[32] = {
4051 [7] = GFS2_DIF_NOATIME,
4052 [12] = GFS2_DIF_EXHASH,
4053 [14] = GFS2_DIF_INHERIT_JDATA,
4054 + [27] = GFS2_DIF_IXUNLINK,
4055 + [26] = GFS2_DIF_BARRIER,
4056 + [29] = GFS2_DIF_COW,
4059 static const u32 gfs2_to_fsflags[32] = {
4060 @@ -141,6 +144,9 @@ static const u32 gfs2_to_fsflags[32] = {
4061 [gfs2fl_NoAtime] = FS_NOATIME_FL,
4062 [gfs2fl_ExHash] = FS_INDEX_FL,
4063 [gfs2fl_InheritJdata] = FS_JOURNAL_DATA_FL,
4064 + [gfs2fl_IXUnlink] = FS_IXUNLINK_FL,
4065 + [gfs2fl_Barrier] = FS_BARRIER_FL,
4066 + [gfs2fl_Cow] = FS_COW_FL,
4069 static int gfs2_get_flags(struct file *filp, u32 __user *ptr)
4070 @@ -171,10 +177,16 @@ void gfs2_set_inode_flags(struct inode *
4072 struct gfs2_inode *ip = GFS2_I(inode);
4073 unsigned int flags = inode->i_flags;
4074 + unsigned int vflags = inode->i_vflags;
4076 + flags &= ~(S_IMMUTABLE | S_IXUNLINK |
4077 + S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
4079 - flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC);
4080 if (ip->i_diskflags & GFS2_DIF_IMMUTABLE)
4081 flags |= S_IMMUTABLE;
4082 + if (ip->i_diskflags & GFS2_DIF_IXUNLINK)
4083 + flags |= S_IXUNLINK;
4085 if (ip->i_diskflags & GFS2_DIF_APPENDONLY)
4087 if (ip->i_diskflags & GFS2_DIF_NOATIME)
4088 @@ -182,6 +194,43 @@ void gfs2_set_inode_flags(struct inode *
4089 if (ip->i_diskflags & GFS2_DIF_SYNC)
4091 inode->i_flags = flags;
4093 + vflags &= ~(V_BARRIER | V_COW);
4095 + if (ip->i_diskflags & GFS2_DIF_BARRIER)
4096 + vflags |= V_BARRIER;
4097 + if (ip->i_diskflags & GFS2_DIF_COW)
4099 + inode->i_vflags = vflags;
4102 +void gfs2_get_inode_flags(struct inode *inode)
4104 + struct gfs2_inode *ip = GFS2_I(inode);
4105 + unsigned int flags = inode->i_flags;
4106 + unsigned int vflags = inode->i_vflags;
4108 + ip->i_diskflags &= ~(GFS2_DIF_APPENDONLY |
4109 + GFS2_DIF_NOATIME | GFS2_DIF_SYNC |
4110 + GFS2_DIF_IMMUTABLE | GFS2_DIF_IXUNLINK |
4111 + GFS2_DIF_BARRIER | GFS2_DIF_COW);
4113 + if (flags & S_IMMUTABLE)
4114 + ip->i_diskflags |= GFS2_DIF_IMMUTABLE;
4115 + if (flags & S_IXUNLINK)
4116 + ip->i_diskflags |= GFS2_DIF_IXUNLINK;
4118 + if (flags & S_APPEND)
4119 + ip->i_diskflags |= GFS2_DIF_APPENDONLY;
4120 + if (flags & S_NOATIME)
4121 + ip->i_diskflags |= GFS2_DIF_NOATIME;
4122 + if (flags & S_SYNC)
4123 + ip->i_diskflags |= GFS2_DIF_SYNC;
4125 + if (vflags & V_BARRIER)
4126 + ip->i_diskflags |= GFS2_DIF_BARRIER;
4127 + if (vflags & V_COW)
4128 + ip->i_diskflags |= GFS2_DIF_COW;
4131 /* Flags that can be set by user space */
4132 @@ -293,6 +342,37 @@ static int gfs2_set_flags(struct file *f
4133 return do_gfs2_set_flags(filp, gfsflags, ~GFS2_DIF_JDATA);
4136 +int gfs2_sync_flags(struct inode *inode, int flags, int vflags)
4138 + struct gfs2_inode *ip = GFS2_I(inode);
4139 + struct gfs2_sbd *sdp = GFS2_SB(inode);
4140 + struct buffer_head *bh;
4141 + struct gfs2_holder gh;
4144 + error = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
4147 + error = gfs2_trans_begin(sdp, RES_DINODE, 0);
4150 + error = gfs2_meta_inode_buffer(ip, &bh);
4152 + goto out_trans_end;
4153 + gfs2_trans_add_bh(ip->i_gl, bh, 1);
4154 + inode->i_flags = flags;
4155 + inode->i_vflags = vflags;
4156 + gfs2_get_inode_flags(inode);
4157 + gfs2_dinode_out(ip, bh->b_data);
4159 + gfs2_set_aops(inode);
4161 + gfs2_trans_end(sdp);
4163 + gfs2_glock_dq_uninit(&gh);
4167 static long gfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
4170 diff -NurpP --minimal linux-2.6.36/fs/gfs2/inode.h linux-2.6.36-vs2.3.0.36.33/fs/gfs2/inode.h
4171 --- linux-2.6.36/fs/gfs2/inode.h 2010-08-02 16:52:49.000000000 +0200
4172 +++ linux-2.6.36-vs2.3.0.36.33/fs/gfs2/inode.h 2010-10-21 13:09:36.000000000 +0200
4173 @@ -109,6 +109,7 @@ extern const struct file_operations gfs2
4174 extern const struct file_operations gfs2_dir_fops_nolock;
4176 extern void gfs2_set_inode_flags(struct inode *inode);
4177 +extern int gfs2_sync_flags(struct inode *inode, int flags, int vflags);
4179 #ifdef CONFIG_GFS2_FS_LOCKING_DLM
4180 extern const struct file_operations gfs2_file_fops;
4181 diff -NurpP --minimal linux-2.6.36/fs/gfs2/ops_inode.c linux-2.6.36-vs2.3.0.36.33/fs/gfs2/ops_inode.c
4182 --- linux-2.6.36/fs/gfs2/ops_inode.c 2010-10-21 13:07:49.000000000 +0200
4183 +++ linux-2.6.36-vs2.3.0.36.33/fs/gfs2/ops_inode.c 2010-10-21 13:09:36.000000000 +0200
4184 @@ -1352,6 +1352,7 @@ const struct inode_operations gfs2_file_
4185 .listxattr = gfs2_listxattr,
4186 .removexattr = gfs2_removexattr,
4187 .fiemap = gfs2_fiemap,
4188 + .sync_flags = gfs2_sync_flags,
4191 const struct inode_operations gfs2_dir_iops = {
4192 @@ -1372,6 +1373,7 @@ const struct inode_operations gfs2_dir_i
4193 .listxattr = gfs2_listxattr,
4194 .removexattr = gfs2_removexattr,
4195 .fiemap = gfs2_fiemap,
4196 + .sync_flags = gfs2_sync_flags,
4199 const struct inode_operations gfs2_symlink_iops = {
4200 diff -NurpP --minimal linux-2.6.36/fs/hfsplus/ioctl.c linux-2.6.36-vs2.3.0.36.33/fs/hfsplus/ioctl.c
4201 --- linux-2.6.36/fs/hfsplus/ioctl.c 2010-08-02 16:52:49.000000000 +0200
4202 +++ linux-2.6.36-vs2.3.0.36.33/fs/hfsplus/ioctl.c 2010-10-21 13:09:36.000000000 +0200
4204 #include <linux/sched.h>
4205 #include <linux/xattr.h>
4206 #include <linux/smp_lock.h>
4207 +// #include <linux/mount.h>
4208 #include <asm/uaccess.h>
4209 #include "hfsplus_fs.h"
4211 diff -NurpP --minimal linux-2.6.36/fs/inode.c linux-2.6.36-vs2.3.0.36.33/fs/inode.c
4212 --- linux-2.6.36/fs/inode.c 2010-10-21 13:07:49.000000000 +0200
4213 +++ linux-2.6.36-vs2.3.0.36.33/fs/inode.c 2010-10-21 13:09:36.000000000 +0200
4215 #include <linux/mount.h>
4216 #include <linux/async.h>
4217 #include <linux/posix_acl.h>
4218 +#include <linux/vs_tag.h>
4221 * This is needed for the following functions:
4222 @@ -130,6 +131,9 @@ int inode_init_always(struct super_block
4223 struct address_space *const mapping = &inode->i_data;
4227 + /* essential because of inode slab reuse */
4229 inode->i_blkbits = sb->s_blocksize_bits;
4231 atomic_set(&inode->i_count, 1);
4232 @@ -150,6 +154,7 @@ int inode_init_always(struct super_block
4233 inode->i_bdev = NULL;
4234 inode->i_cdev = NULL;
4236 + inode->i_mdev = 0;
4237 inode->dirtied_when = 0;
4239 if (security_inode_alloc(inode))
4240 @@ -393,6 +398,8 @@ static int invalidate_list(struct list_h
4244 +EXPORT_SYMBOL_GPL(__iget);
4247 * invalidate_inodes - discard the inodes on a device
4249 @@ -1528,9 +1535,11 @@ void init_special_inode(struct inode *in
4250 if (S_ISCHR(mode)) {
4251 inode->i_fop = &def_chr_fops;
4252 inode->i_rdev = rdev;
4253 + inode->i_mdev = rdev;
4254 } else if (S_ISBLK(mode)) {
4255 inode->i_fop = &def_blk_fops;
4256 inode->i_rdev = rdev;
4257 + inode->i_mdev = rdev;
4258 } else if (S_ISFIFO(mode))
4259 inode->i_fop = &def_fifo_fops;
4260 else if (S_ISSOCK(mode))
4261 @@ -1559,5 +1568,6 @@ void inode_init_owner(struct inode *inod
4263 inode->i_gid = current_fsgid();
4264 inode->i_mode = mode;
4265 + inode->i_tag = dx_current_fstag(inode->i_sb);
4267 EXPORT_SYMBOL(inode_init_owner);
4268 diff -NurpP --minimal linux-2.6.36/fs/ioctl.c linux-2.6.36-vs2.3.0.36.33/fs/ioctl.c
4269 --- linux-2.6.36/fs/ioctl.c 2010-10-21 13:07:49.000000000 +0200
4270 +++ linux-2.6.36-vs2.3.0.36.33/fs/ioctl.c 2010-10-21 13:09:36.000000000 +0200
4272 #include <linux/writeback.h>
4273 #include <linux/buffer_head.h>
4274 #include <linux/falloc.h>
4275 +#include <linux/proc_fs.h>
4276 +#include <linux/vserver/inode.h>
4277 +#include <linux/vs_tag.h>
4279 #include <asm/ioctls.h>
4281 diff -NurpP --minimal linux-2.6.36/fs/ioprio.c linux-2.6.36-vs2.3.0.36.33/fs/ioprio.c
4282 --- linux-2.6.36/fs/ioprio.c 2010-07-07 18:31:52.000000000 +0200
4283 +++ linux-2.6.36-vs2.3.0.36.33/fs/ioprio.c 2010-10-21 13:09:36.000000000 +0200
4285 #include <linux/syscalls.h>
4286 #include <linux/security.h>
4287 #include <linux/pid_namespace.h>
4288 +#include <linux/vs_base.h>
4290 int set_task_ioprio(struct task_struct *task, int ioprio)
4292 @@ -124,6 +125,8 @@ SYSCALL_DEFINE3(ioprio_set, int, which,
4294 pgrp = find_vpid(who);
4295 do_each_pid_thread(pgrp, PIDTYPE_PGID, p) {
4296 + if (!vx_check(p->xid, VS_ADMIN_P | VS_IDENT))
4298 ret = set_task_ioprio(p, ioprio);
4301 @@ -213,6 +216,8 @@ SYSCALL_DEFINE2(ioprio_get, int, which,
4303 pgrp = find_vpid(who);
4304 do_each_pid_thread(pgrp, PIDTYPE_PGID, p) {
4305 + if (!vx_check(p->xid, VS_ADMIN_P | VS_IDENT))
4307 tmpio = get_task_ioprio(p);
4310 diff -NurpP --minimal linux-2.6.36/fs/jfs/file.c linux-2.6.36-vs2.3.0.36.33/fs/jfs/file.c
4311 --- linux-2.6.36/fs/jfs/file.c 2010-10-21 13:07:50.000000000 +0200
4312 +++ linux-2.6.36-vs2.3.0.36.33/fs/jfs/file.c 2010-10-21 13:09:36.000000000 +0200
4313 @@ -102,7 +102,8 @@ int jfs_setattr(struct dentry *dentry, s
4314 if (is_quota_modification(inode, iattr))
4315 dquot_initialize(inode);
4316 if ((iattr->ia_valid & ATTR_UID && iattr->ia_uid != inode->i_uid) ||
4317 - (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid)) {
4318 + (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid) ||
4319 + (iattr->ia_valid & ATTR_TAG && iattr->ia_tag != inode->i_tag)) {
4320 rc = dquot_transfer(inode, iattr);
4323 @@ -133,6 +134,7 @@ const struct inode_operations jfs_file_i
4324 #ifdef CONFIG_JFS_POSIX_ACL
4325 .check_acl = jfs_check_acl,
4327 + .sync_flags = jfs_sync_flags,
4330 const struct file_operations jfs_file_operations = {
4331 diff -NurpP --minimal linux-2.6.36/fs/jfs/ioctl.c linux-2.6.36-vs2.3.0.36.33/fs/jfs/ioctl.c
4332 --- linux-2.6.36/fs/jfs/ioctl.c 2008-12-25 00:26:37.000000000 +0100
4333 +++ linux-2.6.36-vs2.3.0.36.33/fs/jfs/ioctl.c 2010-10-21 13:09:36.000000000 +0200
4335 #include <linux/mount.h>
4336 #include <linux/time.h>
4337 #include <linux/sched.h>
4338 +#include <linux/mount.h>
4339 #include <asm/current.h>
4340 #include <asm/uaccess.h>
4342 @@ -52,6 +53,16 @@ static long jfs_map_ext2(unsigned long f
4346 +int jfs_sync_flags(struct inode *inode, int flags, int vflags)
4348 + inode->i_flags = flags;
4349 + inode->i_vflags = vflags;
4350 + jfs_get_inode_flags(JFS_IP(inode));
4351 + inode->i_ctime = CURRENT_TIME_SEC;
4352 + mark_inode_dirty(inode);
4356 long jfs_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
4358 struct inode *inode = filp->f_dentry->d_inode;
4359 @@ -85,6 +96,11 @@ long jfs_ioctl(struct file *filp, unsign
4360 if (!S_ISDIR(inode->i_mode))
4361 flags &= ~JFS_DIRSYNC_FL;
4363 + if (IS_BARRIER(inode)) {
4364 + vxwprintk_task(1, "messing with the barrier.");
4368 /* Is it quota file? Do not allow user to mess with it */
4369 if (IS_NOQUOTA(inode)) {
4371 @@ -102,8 +118,8 @@ long jfs_ioctl(struct file *filp, unsign
4372 * the relevant capability.
4374 if ((oldflags & JFS_IMMUTABLE_FL) ||
4375 - ((flags ^ oldflags) &
4376 - (JFS_APPEND_FL | JFS_IMMUTABLE_FL))) {
4377 + ((flags ^ oldflags) & (JFS_APPEND_FL |
4378 + JFS_IMMUTABLE_FL | JFS_IXUNLINK_FL))) {
4379 if (!capable(CAP_LINUX_IMMUTABLE)) {
4380 mutex_unlock(&inode->i_mutex);
4382 @@ -111,7 +127,7 @@ long jfs_ioctl(struct file *filp, unsign
4386 - flags = flags & JFS_FL_USER_MODIFIABLE;
4387 + flags &= JFS_FL_USER_MODIFIABLE;
4388 flags |= oldflags & ~JFS_FL_USER_MODIFIABLE;
4389 jfs_inode->mode2 = flags;
4391 diff -NurpP --minimal linux-2.6.36/fs/jfs/jfs_dinode.h linux-2.6.36-vs2.3.0.36.33/fs/jfs/jfs_dinode.h
4392 --- linux-2.6.36/fs/jfs/jfs_dinode.h 2008-12-25 00:26:37.000000000 +0100
4393 +++ linux-2.6.36-vs2.3.0.36.33/fs/jfs/jfs_dinode.h 2010-10-21 13:09:36.000000000 +0200
4394 @@ -161,9 +161,13 @@ struct dinode {
4396 #define JFS_APPEND_FL 0x01000000 /* writes to file may only append */
4397 #define JFS_IMMUTABLE_FL 0x02000000 /* Immutable file */
4398 +#define JFS_IXUNLINK_FL 0x08000000 /* Immutable invert on unlink */
4400 -#define JFS_FL_USER_VISIBLE 0x03F80000
4401 -#define JFS_FL_USER_MODIFIABLE 0x03F80000
4402 +#define JFS_BARRIER_FL 0x04000000 /* Barrier for chroot() */
4403 +#define JFS_COW_FL 0x20000000 /* Copy on Write marker */
4405 +#define JFS_FL_USER_VISIBLE 0x07F80000
4406 +#define JFS_FL_USER_MODIFIABLE 0x07F80000
4407 #define JFS_FL_INHERIT 0x03C80000
4409 /* These are identical to EXT[23]_IOC_GETFLAGS/SETFLAGS */
4410 diff -NurpP --minimal linux-2.6.36/fs/jfs/jfs_filsys.h linux-2.6.36-vs2.3.0.36.33/fs/jfs/jfs_filsys.h
4411 --- linux-2.6.36/fs/jfs/jfs_filsys.h 2008-12-25 00:26:37.000000000 +0100
4412 +++ linux-2.6.36-vs2.3.0.36.33/fs/jfs/jfs_filsys.h 2010-10-21 13:09:36.000000000 +0200
4414 #define JFS_NAME_MAX 255
4415 #define JFS_PATH_MAX BPSIZE
4417 +#define JFS_TAGGED 0x00800000 /* Context Tagging */
4420 * file system state (superblock state)
4421 diff -NurpP --minimal linux-2.6.36/fs/jfs/jfs_imap.c linux-2.6.36-vs2.3.0.36.33/fs/jfs/jfs_imap.c
4422 --- linux-2.6.36/fs/jfs/jfs_imap.c 2010-07-07 18:31:52.000000000 +0200
4423 +++ linux-2.6.36-vs2.3.0.36.33/fs/jfs/jfs_imap.c 2010-10-21 13:09:36.000000000 +0200
4425 #include <linux/pagemap.h>
4426 #include <linux/quotaops.h>
4427 #include <linux/slab.h>
4428 +#include <linux/vs_tag.h>
4430 #include "jfs_incore.h"
4431 #include "jfs_inode.h"
4432 @@ -3060,6 +3061,8 @@ static int copy_from_dinode(struct dinod
4434 struct jfs_inode_info *jfs_ip = JFS_IP(ip);
4435 struct jfs_sb_info *sbi = JFS_SBI(ip->i_sb);
4439 jfs_ip->fileset = le32_to_cpu(dip->di_fileset);
4440 jfs_ip->mode2 = le32_to_cpu(dip->di_mode);
4441 @@ -3080,14 +3083,18 @@ static int copy_from_dinode(struct dinod
4443 ip->i_nlink = le32_to_cpu(dip->di_nlink);
4445 - jfs_ip->saved_uid = le32_to_cpu(dip->di_uid);
4446 + uid = le32_to_cpu(dip->di_uid);
4447 + gid = le32_to_cpu(dip->di_gid);
4448 + ip->i_tag = INOTAG_TAG(DX_TAG(ip), uid, gid, 0);
4450 + jfs_ip->saved_uid = INOTAG_UID(DX_TAG(ip), uid, gid);
4452 ip->i_uid = jfs_ip->saved_uid;
4454 ip->i_uid = sbi->uid;
4457 - jfs_ip->saved_gid = le32_to_cpu(dip->di_gid);
4458 + jfs_ip->saved_gid = INOTAG_GID(DX_TAG(ip), uid, gid);
4460 ip->i_gid = jfs_ip->saved_gid;
4462 @@ -3152,14 +3159,12 @@ static void copy_to_dinode(struct dinode
4463 dip->di_size = cpu_to_le64(ip->i_size);
4464 dip->di_nblocks = cpu_to_le64(PBLK2LBLK(ip->i_sb, ip->i_blocks));
4465 dip->di_nlink = cpu_to_le32(ip->i_nlink);
4466 - if (sbi->uid == -1)
4467 - dip->di_uid = cpu_to_le32(ip->i_uid);
4469 - dip->di_uid = cpu_to_le32(jfs_ip->saved_uid);
4470 - if (sbi->gid == -1)
4471 - dip->di_gid = cpu_to_le32(ip->i_gid);
4473 - dip->di_gid = cpu_to_le32(jfs_ip->saved_gid);
4475 + dip->di_uid = cpu_to_le32(TAGINO_UID(DX_TAG(ip),
4476 + (sbi->uid == -1) ? ip->i_uid : jfs_ip->saved_uid, ip->i_tag));
4477 + dip->di_gid = cpu_to_le32(TAGINO_GID(DX_TAG(ip),
4478 + (sbi->gid == -1) ? ip->i_gid : jfs_ip->saved_gid, ip->i_tag));
4480 jfs_get_inode_flags(jfs_ip);
4482 * mode2 is only needed for storing the higher order bits.
4483 diff -NurpP --minimal linux-2.6.36/fs/jfs/jfs_inode.c linux-2.6.36-vs2.3.0.36.33/fs/jfs/jfs_inode.c
4484 --- linux-2.6.36/fs/jfs/jfs_inode.c 2010-08-02 16:52:49.000000000 +0200
4485 +++ linux-2.6.36-vs2.3.0.36.33/fs/jfs/jfs_inode.c 2010-10-21 13:09:36.000000000 +0200
4488 #include <linux/fs.h>
4489 #include <linux/quotaops.h>
4490 +#include <linux/vs_tag.h>
4491 #include "jfs_incore.h"
4492 #include "jfs_inode.h"
4493 #include "jfs_filsys.h"
4494 @@ -30,29 +31,46 @@ void jfs_set_inode_flags(struct inode *i
4496 unsigned int flags = JFS_IP(inode)->mode2;
4498 - inode->i_flags &= ~(S_IMMUTABLE | S_APPEND |
4499 - S_NOATIME | S_DIRSYNC | S_SYNC);
4500 + inode->i_flags &= ~(S_IMMUTABLE | S_IXUNLINK |
4501 + S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
4503 if (flags & JFS_IMMUTABLE_FL)
4504 inode->i_flags |= S_IMMUTABLE;
4505 + if (flags & JFS_IXUNLINK_FL)
4506 + inode->i_flags |= S_IXUNLINK;
4508 + if (flags & JFS_SYNC_FL)
4509 + inode->i_flags |= S_SYNC;
4510 if (flags & JFS_APPEND_FL)
4511 inode->i_flags |= S_APPEND;
4512 if (flags & JFS_NOATIME_FL)
4513 inode->i_flags |= S_NOATIME;
4514 if (flags & JFS_DIRSYNC_FL)
4515 inode->i_flags |= S_DIRSYNC;
4516 - if (flags & JFS_SYNC_FL)
4517 - inode->i_flags |= S_SYNC;
4519 + inode->i_vflags &= ~(V_BARRIER | V_COW);
4521 + if (flags & JFS_BARRIER_FL)
4522 + inode->i_vflags |= V_BARRIER;
4523 + if (flags & JFS_COW_FL)
4524 + inode->i_vflags |= V_COW;
4527 void jfs_get_inode_flags(struct jfs_inode_info *jfs_ip)
4529 unsigned int flags = jfs_ip->vfs_inode.i_flags;
4530 + unsigned int vflags = jfs_ip->vfs_inode.i_vflags;
4532 + jfs_ip->mode2 &= ~(JFS_IMMUTABLE_FL | JFS_IXUNLINK_FL |
4533 + JFS_APPEND_FL | JFS_NOATIME_FL |
4534 + JFS_DIRSYNC_FL | JFS_SYNC_FL |
4535 + JFS_BARRIER_FL | JFS_COW_FL);
4537 - jfs_ip->mode2 &= ~(JFS_IMMUTABLE_FL | JFS_APPEND_FL | JFS_NOATIME_FL |
4538 - JFS_DIRSYNC_FL | JFS_SYNC_FL);
4539 if (flags & S_IMMUTABLE)
4540 jfs_ip->mode2 |= JFS_IMMUTABLE_FL;
4541 + if (flags & S_IXUNLINK)
4542 + jfs_ip->mode2 |= JFS_IXUNLINK_FL;
4544 if (flags & S_APPEND)
4545 jfs_ip->mode2 |= JFS_APPEND_FL;
4546 if (flags & S_NOATIME)
4547 @@ -61,6 +79,11 @@ void jfs_get_inode_flags(struct jfs_inod
4548 jfs_ip->mode2 |= JFS_DIRSYNC_FL;
4550 jfs_ip->mode2 |= JFS_SYNC_FL;
4552 + if (vflags & V_BARRIER)
4553 + jfs_ip->mode2 |= JFS_BARRIER_FL;
4554 + if (vflags & V_COW)
4555 + jfs_ip->mode2 |= JFS_COW_FL;
4559 diff -NurpP --minimal linux-2.6.36/fs/jfs/jfs_inode.h linux-2.6.36-vs2.3.0.36.33/fs/jfs/jfs_inode.h
4560 --- linux-2.6.36/fs/jfs/jfs_inode.h 2010-10-21 13:07:50.000000000 +0200
4561 +++ linux-2.6.36-vs2.3.0.36.33/fs/jfs/jfs_inode.h 2010-10-21 13:09:36.000000000 +0200
4562 @@ -39,6 +39,7 @@ extern struct dentry *jfs_fh_to_dentry(s
4563 extern struct dentry *jfs_fh_to_parent(struct super_block *sb, struct fid *fid,
4564 int fh_len, int fh_type);
4565 extern void jfs_set_inode_flags(struct inode *);
4566 +extern int jfs_sync_flags(struct inode *, int, int);
4567 extern int jfs_get_block(struct inode *, sector_t, struct buffer_head *, int);
4568 extern int jfs_setattr(struct dentry *, struct iattr *);
4570 diff -NurpP --minimal linux-2.6.36/fs/jfs/namei.c linux-2.6.36-vs2.3.0.36.33/fs/jfs/namei.c
4571 --- linux-2.6.36/fs/jfs/namei.c 2010-07-07 18:31:52.000000000 +0200
4572 +++ linux-2.6.36-vs2.3.0.36.33/fs/jfs/namei.c 2010-10-21 13:09:36.000000000 +0200
4574 #include <linux/ctype.h>
4575 #include <linux/quotaops.h>
4576 #include <linux/exportfs.h>
4577 +#include <linux/vs_tag.h>
4578 #include "jfs_incore.h"
4579 #include "jfs_superblock.h"
4580 #include "jfs_inode.h"
4581 @@ -1491,6 +1492,7 @@ static struct dentry *jfs_lookup(struct
4582 return ERR_CAST(ip);
4585 + dx_propagate_tag(nd, ip);
4586 dentry = d_splice_alias(ip, dentry);
4588 if (dentry && (JFS_SBI(dip->i_sb)->mntflag & JFS_OS2))
4589 @@ -1560,6 +1562,7 @@ const struct inode_operations jfs_dir_in
4590 #ifdef CONFIG_JFS_POSIX_ACL
4591 .check_acl = jfs_check_acl,
4593 + .sync_flags = jfs_sync_flags,
4596 const struct file_operations jfs_dir_operations = {
4597 diff -NurpP --minimal linux-2.6.36/fs/jfs/super.c linux-2.6.36-vs2.3.0.36.33/fs/jfs/super.c
4598 --- linux-2.6.36/fs/jfs/super.c 2010-10-21 13:07:50.000000000 +0200
4599 +++ linux-2.6.36-vs2.3.0.36.33/fs/jfs/super.c 2010-10-21 13:09:36.000000000 +0200
4600 @@ -195,7 +195,8 @@ static void jfs_put_super(struct super_b
4602 Opt_integrity, Opt_nointegrity, Opt_iocharset, Opt_resize,
4603 Opt_resize_nosize, Opt_errors, Opt_ignore, Opt_err, Opt_quota,
4604 - Opt_usrquota, Opt_grpquota, Opt_uid, Opt_gid, Opt_umask
4605 + Opt_usrquota, Opt_grpquota, Opt_uid, Opt_gid, Opt_umask,
4606 + Opt_tag, Opt_notag, Opt_tagid
4609 static const match_table_t tokens = {
4610 @@ -205,6 +206,10 @@ static const match_table_t tokens = {
4611 {Opt_resize, "resize=%u"},
4612 {Opt_resize_nosize, "resize"},
4613 {Opt_errors, "errors=%s"},
4615 + {Opt_notag, "notag"},
4616 + {Opt_tagid, "tagid=%u"},
4617 + {Opt_tag, "tagxid"},
4618 {Opt_ignore, "noquota"},
4619 {Opt_ignore, "quota"},
4620 {Opt_usrquota, "usrquota"},
4621 @@ -339,6 +344,20 @@ static int parse_options(char *options,
4625 +#ifndef CONFIG_TAGGING_NONE
4627 + *flag |= JFS_TAGGED;
4630 + *flag &= JFS_TAGGED;
4633 +#ifdef CONFIG_PROPAGATE
4636 + *flag |= JFS_TAGGED;
4640 printk("jfs: Unrecognized mount option \"%s\" "
4641 " or missing value\n", p);
4642 @@ -369,6 +388,12 @@ static int jfs_remount(struct super_bloc
4643 if (!parse_options(data, sb, &newLVSize, &flag)) {
4646 + if ((flag & JFS_TAGGED) && !(sb->s_flags & MS_TAGGED)) {
4647 + printk(KERN_ERR "JFS: %s: tagging not permitted on remount.\n",
4654 if (sb->s_flags & MS_RDONLY) {
4655 @@ -460,6 +485,9 @@ static int jfs_fill_super(struct super_b
4656 #ifdef CONFIG_JFS_POSIX_ACL
4657 sb->s_flags |= MS_POSIXACL;
4659 + /* map mount option tagxid */
4660 + if (sbi->flag & JFS_TAGGED)
4661 + sb->s_flags |= MS_TAGGED;
4664 printk(KERN_ERR "resize option for remount only\n");
4665 diff -NurpP --minimal linux-2.6.36/fs/libfs.c linux-2.6.36-vs2.3.0.36.33/fs/libfs.c
4666 --- linux-2.6.36/fs/libfs.c 2010-10-21 13:07:50.000000000 +0200
4667 +++ linux-2.6.36-vs2.3.0.36.33/fs/libfs.c 2010-10-21 13:09:36.000000000 +0200
4668 @@ -124,7 +124,8 @@ static inline unsigned char dt_type(stru
4669 * both impossible due to the lock on directory.
4672 -int dcache_readdir(struct file * filp, void * dirent, filldir_t filldir)
4673 +static inline int do_dcache_readdir_filter(struct file *filp,
4674 + void *dirent, filldir_t filldir, int (*filter)(struct dentry *dentry))
4676 struct dentry *dentry = filp->f_path.dentry;
4677 struct dentry *cursor = filp->private_data;
4678 @@ -157,6 +158,8 @@ int dcache_readdir(struct file * filp, v
4679 next = list_entry(p, struct dentry, d_u.d_child);
4680 if (d_unhashed(next) || !next->d_inode)
4682 + if (filter && !filter(next))
4685 spin_unlock(&dcache_lock);
4686 if (filldir(dirent, next->d_name.name,
4687 @@ -175,6 +178,18 @@ int dcache_readdir(struct file * filp, v
4691 +int dcache_readdir(struct file *filp, void *dirent, filldir_t filldir)
4693 + return do_dcache_readdir_filter(filp, dirent, filldir, NULL);
4696 +int dcache_readdir_filter(struct file *filp, void *dirent, filldir_t filldir,
4697 + int (*filter)(struct dentry *))
4699 + return do_dcache_readdir_filter(filp, dirent, filldir, filter);
4703 ssize_t generic_read_dir(struct file *filp, char __user *buf, size_t siz, loff_t *ppos)
4706 @@ -925,6 +940,7 @@ EXPORT_SYMBOL(dcache_dir_close);
4707 EXPORT_SYMBOL(dcache_dir_lseek);
4708 EXPORT_SYMBOL(dcache_dir_open);
4709 EXPORT_SYMBOL(dcache_readdir);
4710 +EXPORT_SYMBOL(dcache_readdir_filter);
4711 EXPORT_SYMBOL(generic_read_dir);
4712 EXPORT_SYMBOL(get_sb_pseudo);
4713 EXPORT_SYMBOL(simple_write_begin);
4714 diff -NurpP --minimal linux-2.6.36/fs/locks.c linux-2.6.36-vs2.3.0.36.33/fs/locks.c
4715 --- linux-2.6.36/fs/locks.c 2010-07-07 18:31:52.000000000 +0200
4716 +++ linux-2.6.36-vs2.3.0.36.33/fs/locks.c 2010-10-21 13:09:36.000000000 +0200
4718 #include <linux/time.h>
4719 #include <linux/rcupdate.h>
4720 #include <linux/pid_namespace.h>
4721 +#include <linux/vs_base.h>
4722 +#include <linux/vs_limit.h>
4724 #include <asm/uaccess.h>
4726 @@ -148,6 +150,8 @@ static struct kmem_cache *filelock_cache
4727 /* Allocate an empty lock structure. */
4728 static struct file_lock *locks_alloc_lock(void)
4730 + if (!vx_locks_avail(1))
4732 return kmem_cache_alloc(filelock_cache, GFP_KERNEL);
4735 @@ -174,6 +178,7 @@ static void locks_free_lock(struct file_
4736 BUG_ON(!list_empty(&fl->fl_block));
4737 BUG_ON(!list_empty(&fl->fl_link));
4740 locks_release_private(fl);
4741 kmem_cache_free(filelock_cache, fl);
4743 @@ -194,6 +199,7 @@ void locks_init_lock(struct file_lock *f
4744 fl->fl_start = fl->fl_end = 0;
4746 fl->fl_lmops = NULL;
4750 EXPORT_SYMBOL(locks_init_lock);
4751 @@ -248,6 +254,7 @@ void locks_copy_lock(struct file_lock *n
4752 new->fl_file = fl->fl_file;
4753 new->fl_ops = fl->fl_ops;
4754 new->fl_lmops = fl->fl_lmops;
4755 + new->fl_xid = fl->fl_xid;
4757 locks_copy_private(new, fl);
4759 @@ -286,6 +293,11 @@ static int flock_make_lock(struct file *
4760 fl->fl_flags = FL_FLOCK;
4762 fl->fl_end = OFFSET_MAX;
4764 + vxd_assert(filp->f_xid == vx_current_xid(),
4765 + "f_xid(%d) == current(%d)", filp->f_xid, vx_current_xid());
4766 + fl->fl_xid = filp->f_xid;
4771 @@ -451,6 +463,7 @@ static int lease_init(struct file *filp,
4773 fl->fl_owner = current->files;
4774 fl->fl_pid = current->tgid;
4775 + fl->fl_xid = vx_current_xid();
4778 fl->fl_flags = FL_LEASE;
4779 @@ -470,6 +483,11 @@ static struct file_lock *lease_alloc(str
4781 return ERR_PTR(error);
4783 + fl->fl_xid = vx_current_xid();
4785 + vxd_assert(filp->f_xid == fl->fl_xid,
4786 + "f_xid(%d) == fl_xid(%d)", filp->f_xid, fl->fl_xid);
4788 error = lease_init(filp, type, fl);
4790 locks_free_lock(fl);
4791 @@ -770,6 +788,7 @@ static int flock_lock_file(struct file *
4795 + new_fl->fl_xid = -1;
4797 for_each_lock(inode, before) {
4798 struct file_lock *fl = *before;
4799 @@ -790,6 +809,7 @@ find_conflict:
4801 locks_copy_lock(new_fl, request);
4802 locks_insert_lock(before, new_fl);
4803 + vx_locks_inc(new_fl);
4807 @@ -800,7 +820,8 @@ out:
4811 -static int __posix_lock_file(struct inode *inode, struct file_lock *request, struct file_lock *conflock)
4812 +static int __posix_lock_file(struct inode *inode, struct file_lock *request,
4813 + struct file_lock *conflock, xid_t xid)
4815 struct file_lock *fl;
4816 struct file_lock *new_fl = NULL;
4817 @@ -810,6 +831,8 @@ static int __posix_lock_file(struct inod
4818 struct file_lock **before;
4819 int error, added = 0;
4821 + vxd_assert(xid == vx_current_xid(),
4822 + "xid(%d) == current(%d)", xid, vx_current_xid());
4824 * We may need two file_lock structures for this operation,
4825 * so we get them in advance to avoid races.
4826 @@ -820,7 +843,11 @@ static int __posix_lock_file(struct inod
4827 (request->fl_type != F_UNLCK ||
4828 request->fl_start != 0 || request->fl_end != OFFSET_MAX)) {
4829 new_fl = locks_alloc_lock();
4830 + new_fl->fl_xid = xid;
4831 + vx_locks_inc(new_fl);
4832 new_fl2 = locks_alloc_lock();
4833 + new_fl2->fl_xid = xid;
4834 + vx_locks_inc(new_fl2);
4838 @@ -1019,7 +1046,8 @@ static int __posix_lock_file(struct inod
4839 int posix_lock_file(struct file *filp, struct file_lock *fl,
4840 struct file_lock *conflock)
4842 - return __posix_lock_file(filp->f_path.dentry->d_inode, fl, conflock);
4843 + return __posix_lock_file(filp->f_path.dentry->d_inode,
4844 + fl, conflock, filp->f_xid);
4846 EXPORT_SYMBOL(posix_lock_file);
4848 @@ -1109,7 +1137,7 @@ int locks_mandatory_area(int read_write,
4849 fl.fl_end = offset + count - 1;
4852 - error = __posix_lock_file(inode, &fl, NULL);
4853 + error = __posix_lock_file(inode, &fl, NULL, filp->f_xid);
4854 if (error != FILE_LOCK_DEFERRED)
4856 error = wait_event_interruptible(fl.fl_wait, !fl.fl_next);
4857 @@ -1425,6 +1453,7 @@ int generic_setlease(struct file *filp,
4859 locks_copy_lock(new_fl, lease);
4860 locks_insert_lock(before, new_fl);
4861 + vx_locks_inc(new_fl);
4865 @@ -1780,6 +1809,11 @@ int fcntl_setlk(unsigned int fd, struct
4866 if (file_lock == NULL)
4869 + vxd_assert(filp->f_xid == vx_current_xid(),
4870 + "f_xid(%d) == current(%d)", filp->f_xid, vx_current_xid());
4871 + file_lock->fl_xid = filp->f_xid;
4872 + vx_locks_inc(file_lock);
4875 * This might block, so we do it before checking the inode.
4877 @@ -1898,6 +1932,11 @@ int fcntl_setlk64(unsigned int fd, struc
4878 if (file_lock == NULL)
4881 + vxd_assert(filp->f_xid == vx_current_xid(),
4882 + "f_xid(%d) == current(%d)", filp->f_xid, vx_current_xid());
4883 + file_lock->fl_xid = filp->f_xid;
4884 + vx_locks_inc(file_lock);
4887 * This might block, so we do it before checking the inode.
4889 @@ -2163,8 +2202,11 @@ static int locks_show(struct seq_file *f
4891 lock_get_status(f, fl, (long)f->private, "");
4893 - list_for_each_entry(bfl, &fl->fl_block, fl_block)
4894 + list_for_each_entry(bfl, &fl->fl_block, fl_block) {
4895 + if (!vx_check(fl->fl_xid, VS_WATCH_P | VS_IDENT))
4897 lock_get_status(f, bfl, (long)f->private, " ->");
4902 diff -NurpP --minimal linux-2.6.36/fs/namei.c linux-2.6.36-vs2.3.0.36.33/fs/namei.c
4903 --- linux-2.6.36/fs/namei.c 2010-10-21 13:07:50.000000000 +0200
4904 +++ linux-2.6.36-vs2.3.0.36.33/fs/namei.c 2010-10-21 14:45:53.000000000 +0200
4906 #include <linux/fcntl.h>
4907 #include <linux/device_cgroup.h>
4908 #include <linux/fs_struct.h>
4909 +#include <linux/proc_fs.h>
4910 +#include <linux/vserver/inode.h>
4911 +#include <linux/vs_base.h>
4912 +#include <linux/vs_tag.h>
4913 +#include <linux/vs_cowbl.h>
4914 +#include <linux/vs_device.h>
4915 +#include <linux/vs_context.h>
4916 +#include <linux/pid_namespace.h>
4917 #include <asm/uaccess.h>
4919 #include "internal.h"
4920 @@ -166,6 +174,77 @@ void putname(const char *name)
4921 EXPORT_SYMBOL(putname);
4924 +static inline int dx_barrier(const struct inode *inode)
4926 + if (IS_BARRIER(inode) && !vx_check(0, VS_ADMIN | VS_WATCH)) {
4927 + vxwprintk_task(1, "did hit the barrier.");
4933 +static int __dx_permission(const struct inode *inode, int mask)
4935 + if (dx_barrier(inode))
4938 + if (inode->i_sb->s_magic == DEVPTS_SUPER_MAGIC) {
4939 + /* devpts is xid tagged */
4940 + if (S_ISDIR(inode->i_mode) ||
4941 + vx_check((xid_t)inode->i_tag, VS_IDENT | VS_WATCH_P))
4944 + else if (inode->i_sb->s_magic == PROC_SUPER_MAGIC) {
4945 + struct proc_dir_entry *de = PDE(inode);
4947 + if (de && !vx_hide_check(0, de->vx_flags))
4950 + if ((mask & (MAY_WRITE | MAY_APPEND))) {
4952 + struct task_struct *tsk;
4954 + if (vx_check(0, VS_ADMIN | VS_WATCH_P) ||
4955 + vx_flags(VXF_STATE_SETUP, 0))
4958 + pid = PROC_I(inode)->pid;
4962 + tsk = pid_task(pid, PIDTYPE_PID);
4963 + vxdprintk(VXD_CBIT(tag, 0), "accessing %p[#%u]",
4964 + tsk, (tsk ? vx_task_xid(tsk) : 0));
4965 + if (tsk && vx_check(vx_task_xid(tsk), VS_IDENT | VS_WATCH_P))
4969 + /* FIXME: Should we block some entries here? */
4974 + if (dx_notagcheck(inode->i_sb) ||
4975 + dx_check(inode->i_tag, DX_HOSTID | DX_ADMIN | DX_WATCH |
4984 +int dx_permission(const struct inode *inode, int mask)
4986 + int ret = __dx_permission(inode, mask);
4987 + if (unlikely(ret)) {
4988 + vxwprintk_task(1, "denied %x access to %s:%p[#%d,%lu]",
4989 + mask, inode->i_sb->s_id, inode, inode->i_tag,
4996 * This does basic POSIX ACL permission checking
4998 @@ -266,10 +345,14 @@ int inode_permission(struct inode *inode
5000 * Nobody gets write access to an immutable file.
5002 - if (IS_IMMUTABLE(inode))
5003 + if (IS_IMMUTABLE(inode) && !IS_COW(inode))
5007 + retval = dx_permission(inode, mask);
5011 if (inode->i_op->permission)
5012 retval = inode->i_op->permission(inode, mask);
5014 @@ -463,6 +546,9 @@ static int exec_permission(struct inode
5018 + if (dx_barrier(inode))
5021 if (inode->i_op->permission) {
5022 ret = inode->i_op->permission(inode, MAY_EXEC);
5024 @@ -672,7 +758,8 @@ static __always_inline void follow_dotdo
5026 if (nd->path.dentry == nd->root.dentry &&
5027 nd->path.mnt == nd->root.mnt) {
5029 + /* for sane '/' avoid follow_mount() */
5032 if (nd->path.dentry != nd->path.mnt->mnt_root) {
5033 /* rare case of legitimate dget_parent()... */
5034 @@ -725,7 +812,7 @@ static int do_lookup(struct nameidata *n
5036 struct vfsmount *mnt = nd->path.mnt;
5037 struct dentry *dentry, *parent;
5038 - struct inode *dir;
5039 + struct inode *dir, *inode;
5041 * See if the low-level filesystem might want
5042 * to use its own hash..
5043 @@ -747,12 +834,26 @@ static int do_lookup(struct nameidata *n
5045 if (dentry->d_op && dentry->d_op->d_revalidate)
5046 goto need_revalidate;
5048 + inode = dentry->d_inode;
5052 + if (__dx_permission(inode, MAY_ACCESS))
5056 path->dentry = dentry;
5057 __follow_mount(path);
5061 + vxwprintk_task(1, "did lookup hidden %s:%p[#%d,%lu] »%s/%.*s«.",
5062 + inode->i_sb->s_id, inode, inode->i_tag, inode->i_ino,
5063 + vxd_path(&nd->path), name->len, name->name);
5068 parent = nd->path.dentry;
5069 dir = parent->d_inode;
5070 @@ -1307,7 +1408,7 @@ static int may_delete(struct inode *dir,
5073 if (check_sticky(dir, victim->d_inode)||IS_APPEND(victim->d_inode)||
5074 - IS_IMMUTABLE(victim->d_inode) || IS_SWAPFILE(victim->d_inode))
5075 + IS_IXORUNLINK(victim->d_inode) || IS_SWAPFILE(victim->d_inode))
5078 if (!S_ISDIR(victim->d_inode->i_mode))
5079 @@ -1430,6 +1531,14 @@ int may_open(struct path *path, int acc_
5083 +#ifdef CONFIG_VSERVER_COWBL
5084 + if (IS_COW(inode) && (flag & FMODE_WRITE)) {
5085 + if (IS_COW_LINK(inode))
5087 + inode->i_flags &= ~(S_IXUNLINK|S_IMMUTABLE);
5088 + mark_inode_dirty(inode);
5091 error = inode_permission(inode, acc_mode);
5094 @@ -1538,7 +1647,8 @@ static int open_will_truncate(int flag,
5097 static struct file *finish_open(struct nameidata *nd,
5098 - int open_flag, int acc_mode)
5099 + int open_flag, int acc_mode,
5100 + const char *pathname)
5104 @@ -1551,6 +1661,23 @@ static struct file *finish_open(struct n
5107 error = may_open(&nd->path, acc_mode, open_flag);
5108 +#ifdef CONFIG_VSERVER_COWBL
5109 + if (error == -EMLINK) {
5110 + struct dentry *dentry;
5111 + dentry = cow_break_link(pathname);
5112 + if (IS_ERR(dentry)) {
5113 + error = PTR_ERR(dentry);
5117 + if (will_truncate)
5118 + mnt_drop_write(nd->path.mnt);
5119 + release_open_intent(nd);
5120 + path_put(&nd->path);
5121 + return ERR_PTR(-EMLINK);
5127 mnt_drop_write(nd->path.mnt);
5128 @@ -1719,7 +1846,7 @@ static struct file *do_last(struct namei
5129 if (S_ISDIR(path->dentry->d_inode->i_mode))
5132 - filp = finish_open(nd, open_flag, acc_mode);
5133 + filp = finish_open(nd, open_flag, acc_mode, pathname);
5137 @@ -1748,7 +1875,11 @@ struct file *do_filp_open(int dfd, const
5139 int flag = open_to_namei_flags(open_flag);
5140 int force_reval = 0;
5142 +#ifdef CONFIG_VSERVER_COWBL
5147 if (!(open_flag & O_CREAT))
5150 @@ -1814,6 +1945,13 @@ reval:
5151 if (!(open_flag & O_NOFOLLOW))
5152 nd.flags |= LOOKUP_FOLLOW;
5153 filp = do_last(&nd, &path, open_flag, acc_mode, mode, pathname);
5154 +#ifdef CONFIG_VSERVER_COWBL
5155 + if (unlikely(IS_ERR(filp) && PTR_ERR(filp) == -EMLINK)) {
5161 while (unlikely(!filp)) { /* trailing symlink */
5163 struct inode *inode = path.dentry->d_inode;
5164 @@ -1852,6 +1990,13 @@ reval:
5166 nd.flags &= ~LOOKUP_PARENT;
5167 filp = do_last(&nd, &path, open_flag, acc_mode, mode, pathname);
5168 +#ifdef CONFIG_VSERVER_COWBL
5169 + if (unlikely(IS_ERR(filp) && PTR_ERR(filp) == -EMLINK)) {
5175 if (inode->i_op->put_link)
5176 inode->i_op->put_link(holder.dentry, &nd, cookie);
5178 @@ -1952,9 +2097,17 @@ int vfs_mknod(struct inode *dir, struct
5182 - if ((S_ISCHR(mode) || S_ISBLK(mode)) && !capable(CAP_MKNOD))
5183 + if (!(S_ISCHR(mode) || S_ISBLK(mode)))
5186 + if (!capable(CAP_MKNOD))
5189 + if (S_ISCHR(mode) && !vs_chrdev_perm(dev, DATTR_CREATE))
5191 + if (S_ISBLK(mode) && !vs_blkdev_perm(dev, DATTR_CREATE))
5194 if (!dir->i_op->mknod)
5197 @@ -2419,7 +2572,7 @@ int vfs_link(struct dentry *old_dentry,
5199 * A link to an append-only or immutable file cannot be created.
5201 - if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
5202 + if (IS_APPEND(inode) || IS_IXORUNLINK(inode))
5204 if (!dir->i_op->link)
5206 @@ -2791,6 +2944,218 @@ int vfs_follow_link(struct nameidata *nd
5207 return __vfs_follow_link(nd, link);
5211 +#ifdef CONFIG_VSERVER_COWBL
5213 +#include <linux/file.h>
5216 +long do_cow_splice(struct file *in, struct file *out, size_t len)
5220 + return do_splice_direct(in, &ppos, out, len, 0);
5223 +struct dentry *cow_break_link(const char *pathname)
5225 + int ret, mode, pathlen, redo = 0;
5226 + struct nameidata old_nd, dir_nd;
5227 + struct path old_path, new_path;
5228 + struct dentry *dir, *res = NULL;
5229 + struct file *old_file;
5230 + struct file *new_file;
5231 + char *to, *path, pad='\251';
5234 + vxdprintk(VXD_CBIT(misc, 1), "cow_break_link(»%s«)", pathname);
5235 + path = kmalloc(PATH_MAX, GFP_KERNEL);
5240 + /* old_nd will have refs to dentry and mnt */
5241 + ret = path_lookup(pathname, LOOKUP_FOLLOW, &old_nd);
5242 + vxdprintk(VXD_CBIT(misc, 2), "path_lookup(old): %d", ret);
5244 + goto out_free_path;
5246 + old_path = old_nd.path;
5247 + mode = old_path.dentry->d_inode->i_mode;
5249 + to = d_path(&old_path, path, PATH_MAX-2);
5250 + pathlen = strlen(to);
5251 + vxdprintk(VXD_CBIT(misc, 2), "old path »%s« [»%.*s«:%d]", to,
5252 + old_path.dentry->d_name.len, old_path.dentry->d_name.name,
5253 + old_path.dentry->d_name.len);
5255 + to[pathlen + 1] = 0;
5257 + to[pathlen] = pad--;
5259 + if (pad <= '\240')
5262 + vxdprintk(VXD_CBIT(misc, 1), "temp copy »%s«", to);
5263 + /* dir_nd will have refs to dentry and mnt */
5264 + ret = path_lookup(to,
5265 + LOOKUP_PARENT | LOOKUP_OPEN | LOOKUP_CREATE, &dir_nd);
5266 + vxdprintk(VXD_CBIT(misc, 2),
5267 + "path_lookup(new): %d", ret);
5271 + /* this puppy downs the inode mutex */
5272 + new_path.dentry = lookup_create(&dir_nd, 0);
5273 + if (!new_path.dentry || IS_ERR(new_path.dentry)) {
5274 + vxdprintk(VXD_CBIT(misc, 2),
5275 + "lookup_create(new): %p", new_path.dentry);
5276 + mutex_unlock(&dir_nd.path.dentry->d_inode->i_mutex);
5277 + path_put(&dir_nd.path);
5280 + vxdprintk(VXD_CBIT(misc, 2),
5281 + "lookup_create(new): %p [»%.*s«:%d]", new_path.dentry,
5282 + new_path.dentry->d_name.len, new_path.dentry->d_name.name,
5283 + new_path.dentry->d_name.len);
5284 + dir = dir_nd.path.dentry;
5286 + ret = vfs_create(dir_nd.path.dentry->d_inode, new_path.dentry, mode, &dir_nd);
5287 + vxdprintk(VXD_CBIT(misc, 2),
5288 + "vfs_create(new): %d", ret);
5289 + if (ret == -EEXIST) {
5290 + mutex_unlock(&dir->d_inode->i_mutex);
5291 + dput(new_path.dentry);
5292 + path_put(&dir_nd.path);
5296 + goto out_unlock_new;
5298 + /* drop out early, ret passes ENOENT */
5300 + if ((redo = d_unhashed(old_path.dentry)))
5301 + goto out_unlock_new;
5303 + new_path.mnt = dir_nd.path.mnt;
5304 + dget(old_path.dentry);
5305 + mntget(old_path.mnt);
5306 + /* this one cleans up the dentry/mnt in case of failure */
5307 + old_file = dentry_open(old_path.dentry, old_path.mnt,
5308 + O_RDONLY, current_cred());
5309 + vxdprintk(VXD_CBIT(misc, 2),
5310 + "dentry_open(old): %p", old_file);
5311 + if (!old_file || IS_ERR(old_file)) {
5312 + res = IS_ERR(old_file) ? (void *) old_file : res;
5313 + goto out_unlock_new;
5316 + dget(new_path.dentry);
5317 + mntget(new_path.mnt);
5318 + /* this one cleans up the dentry/mnt in case of failure */
5319 + new_file = dentry_open(new_path.dentry, new_path.mnt,
5320 + O_WRONLY, current_cred());
5321 + vxdprintk(VXD_CBIT(misc, 2),
5322 + "dentry_open(new): %p", new_file);
5324 + ret = IS_ERR(new_file) ? PTR_ERR(new_file) : -ENOENT;
5325 + if (!new_file || IS_ERR(new_file))
5326 + goto out_fput_old;
5328 + size = i_size_read(old_file->f_dentry->d_inode);
5329 + ret = do_cow_splice(old_file, new_file, size);
5330 + vxdprintk(VXD_CBIT(misc, 2), "do_splice_direct: %d", ret);
5332 + goto out_fput_both;
5333 + } else if (ret < size) {
5335 + goto out_fput_both;
5337 + struct inode *old_inode = old_path.dentry->d_inode;
5338 + struct inode *new_inode = new_path.dentry->d_inode;
5339 + struct iattr attr = {
5340 + .ia_uid = old_inode->i_uid,
5341 + .ia_gid = old_inode->i_gid,
5342 + .ia_valid = ATTR_UID | ATTR_GID
5345 + setattr_copy(new_inode, &attr);
5346 + mark_inode_dirty(new_inode);
5349 + mutex_lock(&old_path.dentry->d_inode->i_sb->s_vfs_rename_mutex);
5351 + /* drop out late */
5353 + if ((redo = d_unhashed(old_path.dentry)))
5356 + vxdprintk(VXD_CBIT(misc, 2),
5357 + "vfs_rename: [»%*s«:%d] -> [»%*s«:%d]",
5358 + new_path.dentry->d_name.len, new_path.dentry->d_name.name,
5359 + new_path.dentry->d_name.len,
5360 + old_path.dentry->d_name.len, old_path.dentry->d_name.name,
5361 + old_path.dentry->d_name.len);
5362 + ret = vfs_rename(dir_nd.path.dentry->d_inode, new_path.dentry,
5363 + old_nd.path.dentry->d_parent->d_inode, old_path.dentry);
5364 + vxdprintk(VXD_CBIT(misc, 2), "vfs_rename: %d", ret);
5365 + res = new_path.dentry;
5368 + mutex_unlock(&old_path.dentry->d_inode->i_sb->s_vfs_rename_mutex);
5371 + vxdprintk(VXD_CBIT(misc, 3),
5372 + "fput(new_file=%p[#%ld])", new_file,
5373 + atomic_long_read(&new_file->f_count));
5377 + vxdprintk(VXD_CBIT(misc, 3),
5378 + "fput(old_file=%p[#%ld])", old_file,
5379 + atomic_long_read(&old_file->f_count));
5383 + mutex_unlock(&dir->d_inode->i_mutex);
5387 + /* error path cleanup */
5388 + vfs_unlink(dir->d_inode, new_path.dentry);
5389 + dput(new_path.dentry);
5393 + goto out_rel_both;
5394 + /* lookup dentry once again */
5395 + path_put(&old_nd.path);
5396 + ret = path_lookup(pathname, LOOKUP_FOLLOW, &old_nd);
5398 + goto out_rel_both;
5400 + new_path.dentry = old_nd.path.dentry;
5401 + vxdprintk(VXD_CBIT(misc, 2),
5402 + "path_lookup(redo): %p [»%.*s«:%d]", new_path.dentry,
5403 + new_path.dentry->d_name.len, new_path.dentry->d_name.name,
5404 + new_path.dentry->d_name.len);
5405 + dget(new_path.dentry);
5406 + res = new_path.dentry;
5409 + path_put(&dir_nd.path);
5411 + path_put(&old_nd.path);
5416 + res = ERR_PTR(ret);
5422 /* get the link contents into pagecache */
5423 static char *page_getlink(struct dentry * dentry, struct page **ppage)
5425 diff -NurpP --minimal linux-2.6.36/fs/namespace.c linux-2.6.36-vs2.3.0.36.33/fs/namespace.c
5426 --- linux-2.6.36/fs/namespace.c 2010-10-21 13:07:50.000000000 +0200
5427 +++ linux-2.6.36-vs2.3.0.36.33/fs/namespace.c 2010-10-21 15:11:02.000000000 +0200
5429 #include <linux/idr.h>
5430 #include <linux/fs_struct.h>
5431 #include <linux/fsnotify.h>
5432 +#include <linux/vs_base.h>
5433 +#include <linux/vs_context.h>
5434 +#include <linux/vs_tag.h>
5435 +#include <linux/vserver/space.h>
5436 +#include <linux/vserver/global.h>
5437 #include <asm/uaccess.h>
5438 #include <asm/unistd.h>
5440 @@ -601,6 +606,7 @@ static struct vfsmount *clone_mnt(struct
5441 mnt->mnt_root = dget(root);
5442 mnt->mnt_mountpoint = mnt->mnt_root;
5443 mnt->mnt_parent = mnt;
5444 + mnt->mnt_tag = old->mnt_tag;
5446 if (flag & CL_SLAVE) {
5447 list_add(&mnt->mnt_slave, &old->mnt_slave_list);
5448 @@ -699,6 +705,31 @@ static inline void mangle(struct seq_fil
5449 seq_escape(m, s, " \t\n\\");
5452 +static int mnt_is_reachable(struct vfsmount *mnt)
5455 + struct dentry *point;
5458 + if (mnt == mnt->mnt_ns->root)
5461 + br_read_lock(vfsmount_lock);
5462 + root = current->fs->root;
5463 + point = root.dentry;
5465 + while ((mnt != mnt->mnt_parent) && (mnt != root.mnt)) {
5466 + point = mnt->mnt_mountpoint;
5467 + mnt = mnt->mnt_parent;
5470 + ret = (mnt == root.mnt) && is_subdir(point, root.dentry);
5472 + br_read_unlock(vfsmount_lock);
5478 * Simple .show_options callback for filesystems which don't want to
5479 * implement more complex mount option showing.
5480 @@ -801,6 +832,8 @@ static int show_sb_opts(struct seq_file
5481 { MS_SYNCHRONOUS, ",sync" },
5482 { MS_DIRSYNC, ",dirsync" },
5483 { MS_MANDLOCK, ",mand" },
5484 + { MS_TAGGED, ",tag" },
5485 + { MS_NOTAGCHECK, ",notagcheck" },
5488 const struct proc_fs_info *fs_infop;
5489 @@ -847,10 +880,20 @@ static int show_vfsmnt(struct seq_file *
5491 struct path mnt_path = { .dentry = mnt->mnt_root, .mnt = mnt };
5493 - mangle(m, mnt->mnt_devname ? mnt->mnt_devname : "none");
5495 - seq_path(m, &mnt_path, " \t\n\\");
5497 + if (vx_flags(VXF_HIDE_MOUNT, 0))
5499 + if (!mnt_is_reachable(mnt) && !vx_check(0, VS_WATCH_P))
5502 + if (!vx_check(0, VS_ADMIN|VS_WATCH) &&
5503 + mnt == current->fs->root.mnt) {
5504 + seq_puts(m, "/dev/root / ");
5506 + mangle(m, mnt->mnt_devname ? mnt->mnt_devname : "none");
5508 + seq_path(m, &mnt_path, " \t\n\\");
5511 show_type(m, mnt->mnt_sb);
5512 seq_puts(m, __mnt_is_readonly(mnt) ? " ro" : " rw");
5513 err = show_sb_opts(m, mnt->mnt_sb);
5514 @@ -880,6 +923,11 @@ static int show_mountinfo(struct seq_fil
5515 struct path root = p->root;
5518 + if (vx_flags(VXF_HIDE_MOUNT, 0))
5520 + if (!mnt_is_reachable(mnt) && !vx_check(0, VS_WATCH_P))
5523 seq_printf(m, "%i %i %u:%u ", mnt->mnt_id, mnt->mnt_parent->mnt_id,
5524 MAJOR(sb->s_dev), MINOR(sb->s_dev));
5525 seq_dentry(m, mnt->mnt_root, " \t\n\\");
5526 @@ -938,17 +986,27 @@ static int show_vfsstat(struct seq_file
5527 struct path mnt_path = { .dentry = mnt->mnt_root, .mnt = mnt };
5531 - if (mnt->mnt_devname) {
5532 - seq_puts(m, "device ");
5533 - mangle(m, mnt->mnt_devname);
5535 - seq_puts(m, "no device");
5536 + if (vx_flags(VXF_HIDE_MOUNT, 0))
5538 + if (!mnt_is_reachable(mnt) && !vx_check(0, VS_WATCH_P))
5542 - seq_puts(m, " mounted on ");
5543 - seq_path(m, &mnt_path, " \t\n\\");
5545 + if (!vx_check(0, VS_ADMIN|VS_WATCH) &&
5546 + mnt == current->fs->root.mnt) {
5547 + seq_puts(m, "device /dev/root mounted on / ");
5550 + if (mnt->mnt_devname) {
5551 + seq_puts(m, "device ");
5552 + mangle(m, mnt->mnt_devname);
5554 + seq_puts(m, "no device");
5557 + seq_puts(m, " mounted on ");
5558 + seq_path(m, &mnt_path, " \t\n\\");
5562 /* file system type */
5563 seq_puts(m, "with fstype ");
5564 @@ -1194,7 +1252,7 @@ SYSCALL_DEFINE2(umount, char __user *, n
5568 - if (!capable(CAP_SYS_ADMIN))
5569 + if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT))
5572 retval = do_umount(path.mnt, flags);
5573 @@ -1220,7 +1278,7 @@ SYSCALL_DEFINE1(oldumount, char __user *
5575 static int mount_is_safe(struct path *path)
5577 - if (capable(CAP_SYS_ADMIN))
5578 + if (vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT))
5582 @@ -1510,7 +1568,7 @@ static int do_change_type(struct path *p
5586 - if (!capable(CAP_SYS_ADMIN))
5587 + if (!vx_capable(CAP_SYS_ADMIN, VXC_NAMESPACE))
5590 if (path->dentry != path->mnt->mnt_root)
5591 @@ -1541,11 +1599,13 @@ static int do_change_type(struct path *p
5592 * do loopback mount.
5594 static int do_loopback(struct path *path, char *old_name,
5596 + tag_t tag, unsigned long flags, int mnt_flags)
5598 struct path old_path;
5599 struct vfsmount *mnt = NULL;
5600 int err = mount_is_safe(path);
5601 + int recurse = flags & MS_REC;
5605 if (!old_name || !*old_name)
5606 @@ -1580,6 +1640,7 @@ static int do_loopback(struct path *path
5607 br_write_unlock(vfsmount_lock);
5608 release_mounts(&umount_list);
5610 + mnt->mnt_flags = mnt_flags;
5613 up_write(&namespace_sem);
5614 @@ -1610,12 +1671,12 @@ static int change_mount_flags(struct vfs
5615 * on it - tough luck.
5617 static int do_remount(struct path *path, int flags, int mnt_flags,
5619 + void *data, xid_t xid)
5622 struct super_block *sb = path->mnt->mnt_sb;
5624 - if (!capable(CAP_SYS_ADMIN))
5625 + if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_REMOUNT))
5628 if (!check_mnt(path->mnt))
5629 @@ -1659,7 +1720,7 @@ static int do_move_mount(struct path *pa
5630 struct path old_path, parent_path;
5633 - if (!capable(CAP_SYS_ADMIN))
5634 + if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT))
5636 if (!old_name || !*old_name)
5638 @@ -1741,7 +1802,7 @@ static int do_new_mount(struct path *pat
5641 /* we need capabilities... */
5642 - if (!capable(CAP_SYS_ADMIN))
5643 + if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT))
5647 @@ -2009,6 +2070,7 @@ long do_mount(char *dev_name, char *dir_
5654 if ((flags & MS_MGC_MSK) == MS_MGC_VAL)
5655 @@ -2036,6 +2098,12 @@ long do_mount(char *dev_name, char *dir_
5656 if (!(flags & MS_NOATIME))
5657 mnt_flags |= MNT_RELATIME;
5659 + if (dx_parse_tag(data_page, &tag, 1, &mnt_flags, &flags)) {
5660 + /* FIXME: bind and re-mounts get the tag flag? */
5661 + if (flags & (MS_BIND|MS_REMOUNT))
5662 + flags |= MS_TAGID;
5665 /* Separate the per-mountpoint flags */
5666 if (flags & MS_NOSUID)
5667 mnt_flags |= MNT_NOSUID;
5668 @@ -2052,15 +2120,17 @@ long do_mount(char *dev_name, char *dir_
5669 if (flags & MS_RDONLY)
5670 mnt_flags |= MNT_READONLY;
5672 + if (!capable(CAP_SYS_ADMIN))
5673 + mnt_flags |= MNT_NODEV;
5674 flags &= ~(MS_NOSUID | MS_NOEXEC | MS_NODEV | MS_ACTIVE | MS_BORN |
5675 MS_NOATIME | MS_NODIRATIME | MS_RELATIME| MS_KERNMOUNT |
5678 if (flags & MS_REMOUNT)
5679 retval = do_remount(&path, flags & ~MS_REMOUNT, mnt_flags,
5682 else if (flags & MS_BIND)
5683 - retval = do_loopback(&path, dev_name, flags & MS_REC);
5684 + retval = do_loopback(&path, dev_name, tag, flags, mnt_flags);
5685 else if (flags & (MS_SHARED | MS_PRIVATE | MS_SLAVE | MS_UNBINDABLE))
5686 retval = do_change_type(&path, flags);
5687 else if (flags & MS_MOVE)
5688 @@ -2139,6 +2209,7 @@ static struct mnt_namespace *dup_mnt_ns(
5689 q = next_mnt(q, new_ns->root);
5691 up_write(&namespace_sem);
5692 + atomic_inc(&vs_global_mnt_ns);
5696 @@ -2280,9 +2351,10 @@ SYSCALL_DEFINE2(pivot_root, const char _
5697 down_write(&namespace_sem);
5698 mutex_lock(&old.dentry->d_inode->i_mutex);
5700 - if (IS_MNT_SHARED(old.mnt) ||
5701 + if ((IS_MNT_SHARED(old.mnt) ||
5702 IS_MNT_SHARED(new.mnt->mnt_parent) ||
5703 - IS_MNT_SHARED(root.mnt->mnt_parent))
5704 + IS_MNT_SHARED(root.mnt->mnt_parent)) &&
5705 + !vx_flags(VXF_STATE_SETUP, 0))
5707 if (!check_mnt(root.mnt))
5709 @@ -2415,6 +2487,7 @@ void put_mnt_ns(struct mnt_namespace *ns
5710 br_write_unlock(vfsmount_lock);
5711 up_write(&namespace_sem);
5712 release_mounts(&umount_list);
5713 + atomic_dec(&vs_global_mnt_ns);
5716 EXPORT_SYMBOL(put_mnt_ns);
5717 diff -NurpP --minimal linux-2.6.36/fs/nfs/client.c linux-2.6.36-vs2.3.0.36.33/fs/nfs/client.c
5718 --- linux-2.6.36/fs/nfs/client.c 2010-10-21 13:07:50.000000000 +0200
5719 +++ linux-2.6.36-vs2.3.0.36.33/fs/nfs/client.c 2010-10-21 13:09:36.000000000 +0200
5720 @@ -740,6 +740,9 @@ static int nfs_init_server_rpcclient(str
5721 if (server->flags & NFS_MOUNT_SOFT)
5722 server->client->cl_softrtry = 1;
5724 + server->client->cl_tag = 0;
5725 + if (server->flags & NFS_MOUNT_TAGGED)
5726 + server->client->cl_tag = 1;
5730 @@ -911,6 +914,10 @@ static void nfs_server_set_fsinfo(struct
5731 server->acdirmin = server->acdirmax = 0;
5734 + /* FIXME: needs fsinfo
5735 + if (server->flags & NFS_MOUNT_TAGGED)
5736 + sb->s_flags |= MS_TAGGED; */
5738 server->maxfilesize = fsinfo->maxfilesize;
5740 /* We're airborne Set socket buffersize */
5741 diff -NurpP --minimal linux-2.6.36/fs/nfs/dir.c linux-2.6.36-vs2.3.0.36.33/fs/nfs/dir.c
5742 --- linux-2.6.36/fs/nfs/dir.c 2010-10-21 13:07:50.000000000 +0200
5743 +++ linux-2.6.36-vs2.3.0.36.33/fs/nfs/dir.c 2010-10-21 13:09:36.000000000 +0200
5745 #include <linux/namei.h>
5746 #include <linux/mount.h>
5747 #include <linux/sched.h>
5748 +#include <linux/vs_tag.h>
5750 #include "nfs4_fs.h"
5751 #include "delegation.h"
5752 @@ -986,6 +987,7 @@ static struct dentry *nfs_lookup(struct
5754 goto out_unblock_sillyrename;
5756 + dx_propagate_tag(nd, inode);
5758 res = d_materialise_unique(dentry, inode);
5760 diff -NurpP --minimal linux-2.6.36/fs/nfs/inode.c linux-2.6.36-vs2.3.0.36.33/fs/nfs/inode.c
5761 --- linux-2.6.36/fs/nfs/inode.c 2010-10-21 13:07:50.000000000 +0200
5762 +++ linux-2.6.36-vs2.3.0.36.33/fs/nfs/inode.c 2010-10-21 13:09:36.000000000 +0200
5764 #include <linux/inet.h>
5765 #include <linux/nfs_xdr.h>
5766 #include <linux/slab.h>
5767 +#include <linux/vs_tag.h>
5769 #include <asm/system.h>
5770 #include <asm/uaccess.h>
5771 @@ -267,6 +268,8 @@ nfs_fhget(struct super_block *sb, struct
5772 if (inode->i_state & I_NEW) {
5773 struct nfs_inode *nfsi = NFS_I(inode);
5774 unsigned long now = jiffies;
5778 /* We set i_ino for the few things that still rely on it,
5779 * such as stat(2) */
5780 @@ -315,8 +318,8 @@ nfs_fhget(struct super_block *sb, struct
5781 nfsi->change_attr = 0;
5784 - inode->i_uid = -2;
5785 - inode->i_gid = -2;
5788 inode->i_blocks = 0;
5789 memset(nfsi->cookieverf, 0, sizeof(nfsi->cookieverf));
5791 @@ -353,13 +356,13 @@ nfs_fhget(struct super_block *sb, struct
5792 else if (nfs_server_capable(inode, NFS_CAP_NLINK))
5793 nfsi->cache_validity |= NFS_INO_INVALID_ATTR;
5794 if (fattr->valid & NFS_ATTR_FATTR_OWNER)
5795 - inode->i_uid = fattr->uid;
5797 else if (nfs_server_capable(inode, NFS_CAP_OWNER))
5798 nfsi->cache_validity |= NFS_INO_INVALID_ATTR
5799 | NFS_INO_INVALID_ACCESS
5800 | NFS_INO_INVALID_ACL;
5801 if (fattr->valid & NFS_ATTR_FATTR_GROUP)
5802 - inode->i_gid = fattr->gid;
5804 else if (nfs_server_capable(inode, NFS_CAP_OWNER_GROUP))
5805 nfsi->cache_validity |= NFS_INO_INVALID_ATTR
5806 | NFS_INO_INVALID_ACCESS
5807 @@ -372,6 +375,11 @@ nfs_fhget(struct super_block *sb, struct
5809 inode->i_blocks = nfs_calc_block_size(fattr->du.nfs3.used);
5811 + inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
5812 + inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
5813 + inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid, 0);
5814 + /* maybe fattr->xid someday */
5816 nfsi->attrtimeo = NFS_MINATTRTIMEO(inode);
5817 nfsi->attrtimeo_timestamp = now;
5818 nfsi->access_cache = RB_ROOT;
5819 @@ -488,6 +496,8 @@ void nfs_setattr_update_inode(struct ino
5820 inode->i_uid = attr->ia_uid;
5821 if ((attr->ia_valid & ATTR_GID) != 0)
5822 inode->i_gid = attr->ia_gid;
5823 + if ((attr->ia_valid & ATTR_TAG) && IS_TAGGED(inode))
5824 + inode->i_tag = attr->ia_tag;
5825 NFS_I(inode)->cache_validity |= NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
5826 spin_unlock(&inode->i_lock);
5828 @@ -923,6 +933,9 @@ static int nfs_check_inode_attributes(st
5829 struct nfs_inode *nfsi = NFS_I(inode);
5830 loff_t cur_size, new_isize;
5831 unsigned long invalid = 0;
5837 /* Has the inode gone and changed behind our back? */
5838 @@ -946,13 +959,18 @@ static int nfs_check_inode_attributes(st
5839 invalid |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE;
5842 + uid = INOTAG_UID(DX_TAG(inode), fattr->uid, fattr->gid);
5843 + gid = INOTAG_GID(DX_TAG(inode), fattr->uid, fattr->gid);
5844 + tag = INOTAG_TAG(DX_TAG(inode), fattr->uid, fattr->gid, 0);
5846 /* Have any file permissions changed? */
5847 if ((fattr->valid & NFS_ATTR_FATTR_MODE) && (inode->i_mode & S_IALLUGO) != (fattr->mode & S_IALLUGO))
5848 invalid |= NFS_INO_INVALID_ATTR | NFS_INO_INVALID_ACCESS | NFS_INO_INVALID_ACL;
5849 - if ((fattr->valid & NFS_ATTR_FATTR_OWNER) && inode->i_uid != fattr->uid)
5850 + if ((fattr->valid & NFS_ATTR_FATTR_OWNER) && uid != fattr->uid)
5851 invalid |= NFS_INO_INVALID_ATTR | NFS_INO_INVALID_ACCESS | NFS_INO_INVALID_ACL;
5852 - if ((fattr->valid & NFS_ATTR_FATTR_GROUP) && inode->i_gid != fattr->gid)
5853 + if ((fattr->valid & NFS_ATTR_FATTR_GROUP) && gid != fattr->gid)
5854 invalid |= NFS_INO_INVALID_ATTR | NFS_INO_INVALID_ACCESS | NFS_INO_INVALID_ACL;
5855 + /* maybe check for tag too? */
5857 /* Has the link count changed? */
5858 if ((fattr->valid & NFS_ATTR_FATTR_NLINK) && inode->i_nlink != fattr->nlink)
5859 @@ -1187,6 +1205,9 @@ static int nfs_update_inode(struct inode
5860 unsigned long invalid = 0;
5861 unsigned long now = jiffies;
5862 unsigned long save_cache_validity;
5867 dfprintk(VFS, "NFS: %s(%s/%ld ct=%d info=0x%x)\n",
5868 __func__, inode->i_sb->s_id, inode->i_ino,
5869 @@ -1289,6 +1310,9 @@ static int nfs_update_inode(struct inode
5870 | NFS_INO_REVAL_PAGECACHE
5871 | NFS_INO_REVAL_FORCED);
5873 + uid = INOTAG_UID(DX_TAG(inode), fattr->uid, fattr->gid);
5874 + gid = INOTAG_GID(DX_TAG(inode), fattr->uid, fattr->gid);
5875 + tag = INOTAG_TAG(DX_TAG(inode), fattr->uid, fattr->gid, 0);
5877 if (fattr->valid & NFS_ATTR_FATTR_ATIME)
5878 memcpy(&inode->i_atime, &fattr->atime, sizeof(inode->i_atime));
5879 @@ -1310,9 +1334,9 @@ static int nfs_update_inode(struct inode
5880 | NFS_INO_REVAL_FORCED);
5882 if (fattr->valid & NFS_ATTR_FATTR_OWNER) {
5883 - if (inode->i_uid != fattr->uid) {
5884 + if (uid != fattr->uid) {
5885 invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
5886 - inode->i_uid = fattr->uid;
5889 } else if (server->caps & NFS_CAP_OWNER)
5890 invalid |= save_cache_validity & (NFS_INO_INVALID_ATTR
5891 @@ -1321,9 +1345,9 @@ static int nfs_update_inode(struct inode
5892 | NFS_INO_REVAL_FORCED);
5894 if (fattr->valid & NFS_ATTR_FATTR_GROUP) {
5895 - if (inode->i_gid != fattr->gid) {
5896 + if (gid != fattr->gid) {
5897 invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
5898 - inode->i_gid = fattr->gid;
5901 } else if (server->caps & NFS_CAP_OWNER_GROUP)
5902 invalid |= save_cache_validity & (NFS_INO_INVALID_ATTR
5903 @@ -1331,6 +1355,10 @@ static int nfs_update_inode(struct inode
5904 | NFS_INO_INVALID_ACL
5905 | NFS_INO_REVAL_FORCED);
5907 + inode->i_uid = uid;
5908 + inode->i_gid = gid;
5909 + inode->i_tag = tag;
5911 if (fattr->valid & NFS_ATTR_FATTR_NLINK) {
5912 if (inode->i_nlink != fattr->nlink) {
5913 invalid |= NFS_INO_INVALID_ATTR;
5914 diff -NurpP --minimal linux-2.6.36/fs/nfs/nfs3xdr.c linux-2.6.36-vs2.3.0.36.33/fs/nfs/nfs3xdr.c
5915 --- linux-2.6.36/fs/nfs/nfs3xdr.c 2010-10-21 13:07:50.000000000 +0200
5916 +++ linux-2.6.36-vs2.3.0.36.33/fs/nfs/nfs3xdr.c 2010-10-21 13:09:36.000000000 +0200
5918 #include <linux/nfs3.h>
5919 #include <linux/nfs_fs.h>
5920 #include <linux/nfsacl.h>
5921 +#include <linux/vs_tag.h>
5922 #include "internal.h"
5924 #define NFSDBG_FACILITY NFSDBG_XDR
5925 @@ -175,7 +176,7 @@ xdr_decode_fattr(__be32 *p, struct nfs_f
5928 static inline __be32 *
5929 -xdr_encode_sattr(__be32 *p, struct iattr *attr)
5930 +xdr_encode_sattr(__be32 *p, struct iattr *attr, int tag)
5932 if (attr->ia_valid & ATTR_MODE) {
5934 @@ -183,15 +184,17 @@ xdr_encode_sattr(__be32 *p, struct iattr
5938 - if (attr->ia_valid & ATTR_UID) {
5939 + if (attr->ia_valid & ATTR_UID ||
5940 + (tag && (attr->ia_valid & ATTR_TAG))) {
5942 - *p++ = htonl(attr->ia_uid);
5943 + *p++ = htonl(TAGINO_UID(tag, attr->ia_uid, attr->ia_tag));
5947 - if (attr->ia_valid & ATTR_GID) {
5948 + if (attr->ia_valid & ATTR_GID ||
5949 + (tag && (attr->ia_valid & ATTR_TAG))) {
5951 - *p++ = htonl(attr->ia_gid);
5952 + *p++ = htonl(TAGINO_GID(tag, attr->ia_gid, attr->ia_tag));
5956 @@ -278,7 +281,8 @@ static int
5957 nfs3_xdr_sattrargs(struct rpc_rqst *req, __be32 *p, struct nfs3_sattrargs *args)
5959 p = xdr_encode_fhandle(p, args->fh);
5960 - p = xdr_encode_sattr(p, args->sattr);
5961 + p = xdr_encode_sattr(p, args->sattr,
5962 + req->rq_task->tk_client->cl_tag);
5963 *p++ = htonl(args->guard);
5965 p = xdr_encode_time3(p, &args->guardtime);
5966 @@ -383,7 +387,8 @@ nfs3_xdr_createargs(struct rpc_rqst *req
5967 *p++ = args->verifier[0];
5968 *p++ = args->verifier[1];
5970 - p = xdr_encode_sattr(p, args->sattr);
5971 + p = xdr_encode_sattr(p, args->sattr,
5972 + req->rq_task->tk_client->cl_tag);
5974 req->rq_slen = xdr_adjust_iovec(req->rq_svec, p);
5976 @@ -397,7 +402,8 @@ nfs3_xdr_mkdirargs(struct rpc_rqst *req,
5978 p = xdr_encode_fhandle(p, args->fh);
5979 p = xdr_encode_array(p, args->name, args->len);
5980 - p = xdr_encode_sattr(p, args->sattr);
5981 + p = xdr_encode_sattr(p, args->sattr,
5982 + req->rq_task->tk_client->cl_tag);
5983 req->rq_slen = xdr_adjust_iovec(req->rq_svec, p);
5986 @@ -410,7 +416,8 @@ nfs3_xdr_symlinkargs(struct rpc_rqst *re
5988 p = xdr_encode_fhandle(p, args->fromfh);
5989 p = xdr_encode_array(p, args->fromname, args->fromlen);
5990 - p = xdr_encode_sattr(p, args->sattr);
5991 + p = xdr_encode_sattr(p, args->sattr,
5992 + req->rq_task->tk_client->cl_tag);
5993 *p++ = htonl(args->pathlen);
5994 req->rq_slen = xdr_adjust_iovec(req->rq_svec, p);
5996 @@ -428,7 +435,8 @@ nfs3_xdr_mknodargs(struct rpc_rqst *req,
5997 p = xdr_encode_fhandle(p, args->fh);
5998 p = xdr_encode_array(p, args->name, args->len);
5999 *p++ = htonl(args->type);
6000 - p = xdr_encode_sattr(p, args->sattr);
6001 + p = xdr_encode_sattr(p, args->sattr,
6002 + req->rq_task->tk_client->cl_tag);
6003 if (args->type == NF3CHR || args->type == NF3BLK) {
6004 *p++ = htonl(MAJOR(args->rdev));
6005 *p++ = htonl(MINOR(args->rdev));
6006 diff -NurpP --minimal linux-2.6.36/fs/nfs/nfsroot.c linux-2.6.36-vs2.3.0.36.33/fs/nfs/nfsroot.c
6007 --- linux-2.6.36/fs/nfs/nfsroot.c 2010-08-02 16:52:50.000000000 +0200
6008 +++ linux-2.6.36-vs2.3.0.36.33/fs/nfs/nfsroot.c 2010-10-21 13:09:36.000000000 +0200
6009 @@ -122,12 +122,12 @@ static int mount_port __initdata = 0; /
6011 /* Options that take integer arguments */
6012 Opt_port, Opt_rsize, Opt_wsize, Opt_timeo, Opt_retrans, Opt_acregmin,
6013 - Opt_acregmax, Opt_acdirmin, Opt_acdirmax,
6014 + Opt_acregmax, Opt_acdirmin, Opt_acdirmax, Opt_tagid,
6015 /* Options that take no arguments */
6016 Opt_soft, Opt_hard, Opt_intr,
6017 Opt_nointr, Opt_posix, Opt_noposix, Opt_cto, Opt_nocto, Opt_ac,
6018 Opt_noac, Opt_lock, Opt_nolock, Opt_v2, Opt_v3, Opt_udp, Opt_tcp,
6019 - Opt_acl, Opt_noacl,
6020 + Opt_acl, Opt_noacl, Opt_tag, Opt_notag,
6024 @@ -164,6 +164,9 @@ static const match_table_t tokens __init
6027 {Opt_noacl, "noacl"},
6029 + {Opt_notag, "notag"},
6030 + {Opt_tagid, "tagid=%u"},
6034 @@ -275,6 +278,20 @@ static int __init root_nfs_parse(char *n
6036 nfs_data.flags |= NFS_MOUNT_NOACL;
6038 +#ifndef CONFIG_TAGGING_NONE
6040 + nfs_data.flags |= NFS_MOUNT_TAGGED;
6043 + nfs_data.flags &= ~NFS_MOUNT_TAGGED;
6046 +#ifdef CONFIG_PROPAGATE
6049 + nfs_data.flags |= NFS_MOUNT_TAGGED;
6053 printk(KERN_WARNING "Root-NFS: unknown "
6055 diff -NurpP --minimal linux-2.6.36/fs/nfs/super.c linux-2.6.36-vs2.3.0.36.33/fs/nfs/super.c
6056 --- linux-2.6.36/fs/nfs/super.c 2010-10-21 13:07:50.000000000 +0200
6057 +++ linux-2.6.36-vs2.3.0.36.33/fs/nfs/super.c 2010-10-21 13:09:36.000000000 +0200
6059 #include <linux/nfs_xdr.h>
6060 #include <linux/magic.h>
6061 #include <linux/parser.h>
6062 +#include <linux/vs_tag.h>
6064 #include <asm/system.h>
6065 #include <asm/uaccess.h>
6066 @@ -617,6 +618,7 @@ static void nfs_show_mount_options(struc
6067 { NFS_MOUNT_NORDIRPLUS, ",nordirplus", "" },
6068 { NFS_MOUNT_UNSHARED, ",nosharecache", "" },
6069 { NFS_MOUNT_NORESVPORT, ",noresvport", "" },
6070 + { NFS_MOUNT_TAGGED, ",tag", "" },
6073 const struct proc_nfs_info *nfs_infop;
6074 diff -NurpP --minimal linux-2.6.36/fs/nfsd/auth.c linux-2.6.36-vs2.3.0.36.33/fs/nfsd/auth.c
6075 --- linux-2.6.36/fs/nfsd/auth.c 2010-02-25 11:52:05.000000000 +0100
6076 +++ linux-2.6.36-vs2.3.0.36.33/fs/nfsd/auth.c 2010-10-21 13:09:36.000000000 +0200
6078 /* Copyright (C) 1995, 1996 Olaf Kirch <okir@monad.swb.de> */
6080 #include <linux/sched.h>
6081 +#include <linux/vs_tag.h>
6085 @@ -36,6 +37,9 @@ int nfsd_setuser(struct svc_rqst *rqstp,
6087 new->fsuid = rqstp->rq_cred.cr_uid;
6088 new->fsgid = rqstp->rq_cred.cr_gid;
6089 + /* FIXME: this desperately needs a tag :)
6090 + new->xid = (xid_t)INOTAG_TAG(DX_TAG_NFSD, cred.cr_uid, cred.cr_gid, 0);
6093 rqgi = rqstp->rq_cred.cr_group_info;
6095 diff -NurpP --minimal linux-2.6.36/fs/nfsd/nfs3xdr.c linux-2.6.36-vs2.3.0.36.33/fs/nfsd/nfs3xdr.c
6096 --- linux-2.6.36/fs/nfsd/nfs3xdr.c 2010-02-25 11:52:05.000000000 +0100
6097 +++ linux-2.6.36-vs2.3.0.36.33/fs/nfsd/nfs3xdr.c 2010-10-21 13:09:36.000000000 +0200
6101 #include <linux/namei.h>
6102 +#include <linux/vs_tag.h>
6106 @@ -95,6 +96,8 @@ static __be32 *
6107 decode_sattr3(__be32 *p, struct iattr *iap)
6115 @@ -104,12 +107,15 @@ decode_sattr3(__be32 *p, struct iattr *i
6118 iap->ia_valid |= ATTR_UID;
6119 - iap->ia_uid = ntohl(*p++);
6120 + uid = ntohl(*p++);
6123 iap->ia_valid |= ATTR_GID;
6124 - iap->ia_gid = ntohl(*p++);
6125 + gid = ntohl(*p++);
6127 + iap->ia_uid = INOTAG_UID(DX_TAG_NFSD, uid, gid);
6128 + iap->ia_gid = INOTAG_GID(DX_TAG_NFSD, uid, gid);
6129 + iap->ia_tag = INOTAG_TAG(DX_TAG_NFSD, uid, gid, 0);
6133 @@ -165,8 +171,12 @@ encode_fattr3(struct svc_rqst *rqstp, __
6134 *p++ = htonl(nfs3_ftypes[(stat->mode & S_IFMT) >> 12]);
6135 *p++ = htonl((u32) stat->mode);
6136 *p++ = htonl((u32) stat->nlink);
6137 - *p++ = htonl((u32) nfsd_ruid(rqstp, stat->uid));
6138 - *p++ = htonl((u32) nfsd_rgid(rqstp, stat->gid));
6139 + *p++ = htonl((u32) nfsd_ruid(rqstp,
6140 + TAGINO_UID(0 /* FIXME: DX_TAG(dentry->d_inode) */,
6141 + stat->uid, stat->tag)));
6142 + *p++ = htonl((u32) nfsd_rgid(rqstp,
6143 + TAGINO_GID(0 /* FIXME: DX_TAG(dentry->d_inode) */,
6144 + stat->gid, stat->tag)));
6145 if (S_ISLNK(stat->mode) && stat->size > NFS3_MAXPATHLEN) {
6146 p = xdr_encode_hyper(p, (u64) NFS3_MAXPATHLEN);
6148 diff -NurpP --minimal linux-2.6.36/fs/nfsd/nfs4xdr.c linux-2.6.36-vs2.3.0.36.33/fs/nfsd/nfs4xdr.c
6149 --- linux-2.6.36/fs/nfsd/nfs4xdr.c 2010-10-21 13:07:50.000000000 +0200
6150 +++ linux-2.6.36-vs2.3.0.36.33/fs/nfsd/nfs4xdr.c 2010-10-21 13:09:36.000000000 +0200
6152 #include <linux/nfsd_idmap.h>
6153 #include <linux/nfs4_acl.h>
6154 #include <linux/sunrpc/svcauth_gss.h>
6155 +#include <linux/vs_tag.h>
6159 @@ -2057,14 +2058,18 @@ out_acl:
6160 WRITE32(stat.nlink);
6162 if (bmval1 & FATTR4_WORD1_OWNER) {
6163 - status = nfsd4_encode_user(rqstp, stat.uid, &p, &buflen);
6164 + status = nfsd4_encode_user(rqstp,
6165 + TAGINO_UID(DX_TAG(dentry->d_inode),
6166 + stat.uid, stat.tag), &p, &buflen);
6167 if (status == nfserr_resource)
6172 if (bmval1 & FATTR4_WORD1_OWNER_GROUP) {
6173 - status = nfsd4_encode_group(rqstp, stat.gid, &p, &buflen);
6174 + status = nfsd4_encode_group(rqstp,
6175 + TAGINO_GID(DX_TAG(dentry->d_inode),
6176 + stat.gid, stat.tag), &p, &buflen);
6177 if (status == nfserr_resource)
6180 diff -NurpP --minimal linux-2.6.36/fs/nfsd/nfsxdr.c linux-2.6.36-vs2.3.0.36.33/fs/nfsd/nfsxdr.c
6181 --- linux-2.6.36/fs/nfsd/nfsxdr.c 2010-02-25 11:52:05.000000000 +0100
6182 +++ linux-2.6.36-vs2.3.0.36.33/fs/nfsd/nfsxdr.c 2010-10-21 13:09:36.000000000 +0200
6187 +#include <linux/vs_tag.h>
6189 #define NFSDDBG_FACILITY NFSDDBG_XDR
6191 @@ -88,6 +89,8 @@ static __be32 *
6192 decode_sattr(__be32 *p, struct iattr *iap)
6200 @@ -101,12 +104,15 @@ decode_sattr(__be32 *p, struct iattr *ia
6202 if ((tmp = ntohl(*p++)) != (u32)-1) {
6203 iap->ia_valid |= ATTR_UID;
6204 - iap->ia_uid = tmp;
6207 if ((tmp = ntohl(*p++)) != (u32)-1) {
6208 iap->ia_valid |= ATTR_GID;
6209 - iap->ia_gid = tmp;
6212 + iap->ia_uid = INOTAG_UID(DX_TAG_NFSD, uid, gid);
6213 + iap->ia_gid = INOTAG_GID(DX_TAG_NFSD, uid, gid);
6214 + iap->ia_tag = INOTAG_TAG(DX_TAG_NFSD, uid, gid, 0);
6215 if ((tmp = ntohl(*p++)) != (u32)-1) {
6216 iap->ia_valid |= ATTR_SIZE;
6218 @@ -151,8 +157,10 @@ encode_fattr(struct svc_rqst *rqstp, __b
6219 *p++ = htonl(nfs_ftypes[type >> 12]);
6220 *p++ = htonl((u32) stat->mode);
6221 *p++ = htonl((u32) stat->nlink);
6222 - *p++ = htonl((u32) nfsd_ruid(rqstp, stat->uid));
6223 - *p++ = htonl((u32) nfsd_rgid(rqstp, stat->gid));
6224 + *p++ = htonl((u32) nfsd_ruid(rqstp,
6225 + TAGINO_UID(DX_TAG(dentry->d_inode), stat->uid, stat->tag)));
6226 + *p++ = htonl((u32) nfsd_rgid(rqstp,
6227 + TAGINO_GID(DX_TAG(dentry->d_inode), stat->gid, stat->tag)));
6229 if (S_ISLNK(type) && stat->size > NFS_MAXPATHLEN) {
6230 *p++ = htonl(NFS_MAXPATHLEN);
6231 diff -NurpP --minimal linux-2.6.36/fs/ocfs2/dlmglue.c linux-2.6.36-vs2.3.0.36.33/fs/ocfs2/dlmglue.c
6232 --- linux-2.6.36/fs/ocfs2/dlmglue.c 2010-10-21 13:07:50.000000000 +0200
6233 +++ linux-2.6.36-vs2.3.0.36.33/fs/ocfs2/dlmglue.c 2010-10-21 13:09:36.000000000 +0200
6234 @@ -2114,6 +2114,7 @@ static void __ocfs2_stuff_meta_lvb(struc
6235 lvb->lvb_iclusters = cpu_to_be32(oi->ip_clusters);
6236 lvb->lvb_iuid = cpu_to_be32(inode->i_uid);
6237 lvb->lvb_igid = cpu_to_be32(inode->i_gid);
6238 + lvb->lvb_itag = cpu_to_be16(inode->i_tag);
6239 lvb->lvb_imode = cpu_to_be16(inode->i_mode);
6240 lvb->lvb_inlink = cpu_to_be16(inode->i_nlink);
6241 lvb->lvb_iatime_packed =
6242 @@ -2168,6 +2169,7 @@ static void ocfs2_refresh_inode_from_lvb
6244 inode->i_uid = be32_to_cpu(lvb->lvb_iuid);
6245 inode->i_gid = be32_to_cpu(lvb->lvb_igid);
6246 + inode->i_tag = be16_to_cpu(lvb->lvb_itag);
6247 inode->i_mode = be16_to_cpu(lvb->lvb_imode);
6248 inode->i_nlink = be16_to_cpu(lvb->lvb_inlink);
6249 ocfs2_unpack_timespec(&inode->i_atime,
6250 diff -NurpP --minimal linux-2.6.36/fs/ocfs2/dlmglue.h linux-2.6.36-vs2.3.0.36.33/fs/ocfs2/dlmglue.h
6251 --- linux-2.6.36/fs/ocfs2/dlmglue.h 2010-10-21 13:07:50.000000000 +0200
6252 +++ linux-2.6.36-vs2.3.0.36.33/fs/ocfs2/dlmglue.h 2010-10-21 13:09:36.000000000 +0200
6253 @@ -46,7 +46,8 @@ struct ocfs2_meta_lvb {
6256 __be32 lvb_igeneration;
6257 - __be32 lvb_reserved2;
6259 + __be16 lvb_reserved2;
6262 #define OCFS2_QINFO_LVB_VERSION 1
6263 diff -NurpP --minimal linux-2.6.36/fs/ocfs2/file.c linux-2.6.36-vs2.3.0.36.33/fs/ocfs2/file.c
6264 --- linux-2.6.36/fs/ocfs2/file.c 2010-10-21 13:07:50.000000000 +0200
6265 +++ linux-2.6.36-vs2.3.0.36.33/fs/ocfs2/file.c 2010-10-21 13:09:36.000000000 +0200
6266 @@ -1138,13 +1138,15 @@ int ocfs2_setattr(struct dentry *dentry,
6267 mlog(0, "uid change: %d\n", attr->ia_uid);
6268 if (attr->ia_valid & ATTR_GID)
6269 mlog(0, "gid change: %d\n", attr->ia_gid);
6270 + if (attr->ia_valid & ATTR_TAG)
6271 + mlog(0, "tag change: %d\n", attr->ia_tag);
6272 if (attr->ia_valid & ATTR_SIZE)
6273 mlog(0, "size change...\n");
6274 if (attr->ia_valid & (ATTR_ATIME | ATTR_MTIME | ATTR_CTIME))
6275 mlog(0, "time change...\n");
6277 #define OCFS2_VALID_ATTRS (ATTR_ATIME | ATTR_MTIME | ATTR_CTIME | ATTR_SIZE \
6278 - | ATTR_GID | ATTR_UID | ATTR_MODE)
6279 + | ATTR_GID | ATTR_UID | ATTR_TAG | ATTR_MODE)
6280 if (!(attr->ia_valid & OCFS2_VALID_ATTRS)) {
6281 mlog(0, "can't handle attrs: 0x%x\n", attr->ia_valid);
6283 diff -NurpP --minimal linux-2.6.36/fs/ocfs2/inode.c linux-2.6.36-vs2.3.0.36.33/fs/ocfs2/inode.c
6284 --- linux-2.6.36/fs/ocfs2/inode.c 2010-10-21 13:07:50.000000000 +0200
6285 +++ linux-2.6.36-vs2.3.0.36.33/fs/ocfs2/inode.c 2010-10-21 13:09:36.000000000 +0200
6287 #include <linux/highmem.h>
6288 #include <linux/pagemap.h>
6289 #include <linux/quotaops.h>
6290 +#include <linux/vs_tag.h>
6292 #include <asm/byteorder.h>
6294 @@ -78,11 +79,13 @@ void ocfs2_set_inode_flags(struct inode
6296 unsigned int flags = OCFS2_I(inode)->ip_attr;
6298 - inode->i_flags &= ~(S_IMMUTABLE |
6299 + inode->i_flags &= ~(S_IMMUTABLE | S_IXUNLINK |
6300 S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
6302 if (flags & OCFS2_IMMUTABLE_FL)
6303 inode->i_flags |= S_IMMUTABLE;
6304 + if (flags & OCFS2_IXUNLINK_FL)
6305 + inode->i_flags |= S_IXUNLINK;
6307 if (flags & OCFS2_SYNC_FL)
6308 inode->i_flags |= S_SYNC;
6309 @@ -92,25 +95,44 @@ void ocfs2_set_inode_flags(struct inode
6310 inode->i_flags |= S_NOATIME;
6311 if (flags & OCFS2_DIRSYNC_FL)
6312 inode->i_flags |= S_DIRSYNC;
6314 + inode->i_vflags &= ~(V_BARRIER | V_COW);
6316 + if (flags & OCFS2_BARRIER_FL)
6317 + inode->i_vflags |= V_BARRIER;
6318 + if (flags & OCFS2_COW_FL)
6319 + inode->i_vflags |= V_COW;
6322 /* Propagate flags from i_flags to OCFS2_I(inode)->ip_attr */
6323 void ocfs2_get_inode_flags(struct ocfs2_inode_info *oi)
6325 unsigned int flags = oi->vfs_inode.i_flags;
6326 + unsigned int vflags = oi->vfs_inode.i_vflags;
6328 + oi->ip_attr &= ~(OCFS2_SYNC_FL | OCFS2_APPEND_FL |
6329 + OCFS2_IMMUTABLE_FL | OCFS2_IXUNLINK_FL |
6330 + OCFS2_NOATIME_FL | OCFS2_DIRSYNC_FL |
6331 + OCFS2_BARRIER_FL | OCFS2_COW_FL);
6333 + if (flags & S_IMMUTABLE)
6334 + oi->ip_attr |= OCFS2_IMMUTABLE_FL;
6335 + if (flags & S_IXUNLINK)
6336 + oi->ip_attr |= OCFS2_IXUNLINK_FL;
6338 - oi->ip_attr &= ~(OCFS2_SYNC_FL|OCFS2_APPEND_FL|
6339 - OCFS2_IMMUTABLE_FL|OCFS2_NOATIME_FL|OCFS2_DIRSYNC_FL);
6341 oi->ip_attr |= OCFS2_SYNC_FL;
6342 if (flags & S_APPEND)
6343 oi->ip_attr |= OCFS2_APPEND_FL;
6344 - if (flags & S_IMMUTABLE)
6345 - oi->ip_attr |= OCFS2_IMMUTABLE_FL;
6346 if (flags & S_NOATIME)
6347 oi->ip_attr |= OCFS2_NOATIME_FL;
6348 if (flags & S_DIRSYNC)
6349 oi->ip_attr |= OCFS2_DIRSYNC_FL;
6351 + if (vflags & V_BARRIER)
6352 + oi->ip_attr |= OCFS2_BARRIER_FL;
6353 + if (vflags & V_COW)
6354 + oi->ip_attr |= OCFS2_COW_FL;
6357 struct inode *ocfs2_ilookup(struct super_block *sb, u64 blkno)
6358 @@ -245,6 +267,8 @@ void ocfs2_populate_inode(struct inode *
6359 struct super_block *sb;
6360 struct ocfs2_super *osb;
6365 mlog_entry("(0x%p, size:%llu)\n", inode,
6366 (unsigned long long)le64_to_cpu(fe->i_size));
6367 @@ -276,8 +300,12 @@ void ocfs2_populate_inode(struct inode *
6368 inode->i_generation = le32_to_cpu(fe->i_generation);
6369 inode->i_rdev = huge_decode_dev(le64_to_cpu(fe->id1.dev1.i_rdev));
6370 inode->i_mode = le16_to_cpu(fe->i_mode);
6371 - inode->i_uid = le32_to_cpu(fe->i_uid);
6372 - inode->i_gid = le32_to_cpu(fe->i_gid);
6373 + uid = le32_to_cpu(fe->i_uid);
6374 + gid = le32_to_cpu(fe->i_gid);
6375 + inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
6376 + inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
6377 + inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid,
6378 + /* le16_to_cpu(raw_inode->i_raw_tag)i */ 0);
6380 /* Fast symlinks will have i_size but no allocated clusters. */
6381 if (S_ISLNK(inode->i_mode) && !fe->i_clusters)
6382 diff -NurpP --minimal linux-2.6.36/fs/ocfs2/inode.h linux-2.6.36-vs2.3.0.36.33/fs/ocfs2/inode.h
6383 --- linux-2.6.36/fs/ocfs2/inode.h 2010-10-21 13:07:50.000000000 +0200
6384 +++ linux-2.6.36-vs2.3.0.36.33/fs/ocfs2/inode.h 2010-10-21 13:09:36.000000000 +0200
6385 @@ -153,6 +153,7 @@ struct buffer_head *ocfs2_bread(struct i
6387 void ocfs2_set_inode_flags(struct inode *inode);
6388 void ocfs2_get_inode_flags(struct ocfs2_inode_info *oi);
6389 +int ocfs2_sync_flags(struct inode *inode, int, int);
6391 static inline blkcnt_t ocfs2_inode_sector_count(struct inode *inode)
6393 diff -NurpP --minimal linux-2.6.36/fs/ocfs2/ioctl.c linux-2.6.36-vs2.3.0.36.33/fs/ocfs2/ioctl.c
6394 --- linux-2.6.36/fs/ocfs2/ioctl.c 2010-02-25 11:52:06.000000000 +0100
6395 +++ linux-2.6.36-vs2.3.0.36.33/fs/ocfs2/ioctl.c 2010-10-21 13:09:36.000000000 +0200
6396 @@ -43,7 +43,41 @@ static int ocfs2_get_inode_attr(struct i
6400 -static int ocfs2_set_inode_attr(struct inode *inode, unsigned flags,
6401 +int ocfs2_sync_flags(struct inode *inode, int flags, int vflags)
6403 + struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
6404 + struct buffer_head *bh = NULL;
6405 + handle_t *handle = NULL;
6408 + status = ocfs2_inode_lock(inode, &bh, 1);
6410 + mlog_errno(status);
6413 + handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS);
6414 + if (IS_ERR(handle)) {
6415 + status = PTR_ERR(handle);
6416 + mlog_errno(status);
6420 + inode->i_flags = flags;
6421 + inode->i_vflags = vflags;
6422 + ocfs2_get_inode_flags(OCFS2_I(inode));
6424 + status = ocfs2_mark_inode_dirty(handle, inode, bh);
6426 + mlog_errno(status);
6428 + ocfs2_commit_trans(osb, handle);
6430 + ocfs2_inode_unlock(inode, 1);
6435 +int ocfs2_set_inode_attr(struct inode *inode, unsigned flags,
6438 struct ocfs2_inode_info *ocfs2_inode = OCFS2_I(inode);
6439 @@ -68,6 +102,11 @@ static int ocfs2_set_inode_attr(struct i
6440 if (!S_ISDIR(inode->i_mode))
6441 flags &= ~OCFS2_DIRSYNC_FL;
6443 + if (IS_BARRIER(inode)) {
6444 + vxwprintk_task(1, "messing with the barrier.");
6448 handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS);
6449 if (IS_ERR(handle)) {
6450 status = PTR_ERR(handle);
6451 @@ -109,6 +148,7 @@ bail:
6456 long ocfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
6458 struct inode *inode = filp->f_path.dentry->d_inode;
6459 diff -NurpP --minimal linux-2.6.36/fs/ocfs2/namei.c linux-2.6.36-vs2.3.0.36.33/fs/ocfs2/namei.c
6460 --- linux-2.6.36/fs/ocfs2/namei.c 2010-10-21 13:07:50.000000000 +0200
6461 +++ linux-2.6.36-vs2.3.0.36.33/fs/ocfs2/namei.c 2010-10-21 13:09:36.000000000 +0200
6463 #include <linux/slab.h>
6464 #include <linux/highmem.h>
6465 #include <linux/quotaops.h>
6466 +#include <linux/vs_tag.h>
6468 #define MLOG_MASK_PREFIX ML_NAMEI
6469 #include <cluster/masklog.h>
6470 @@ -486,6 +487,7 @@ static int __ocfs2_mknod_locked(struct i
6471 struct ocfs2_dinode *fe = NULL;
6472 struct ocfs2_extent_list *fel;
6478 @@ -523,8 +525,11 @@ static int __ocfs2_mknod_locked(struct i
6479 fe->i_suballoc_loc = cpu_to_le64(suballoc_loc);
6480 fe->i_suballoc_bit = cpu_to_le16(suballoc_bit);
6481 fe->i_suballoc_slot = cpu_to_le16(inode_ac->ac_alloc_slot);
6482 - fe->i_uid = cpu_to_le32(inode->i_uid);
6483 - fe->i_gid = cpu_to_le32(inode->i_gid);
6485 + tag = dx_current_fstag(osb->sb);
6486 + fe->i_uid = cpu_to_le32(TAGINO_UID(DX_TAG(inode), inode->i_uid, tag));
6487 + fe->i_gid = cpu_to_le32(TAGINO_GID(DX_TAG(inode), inode->i_gid, tag));
6488 + inode->i_tag = tag;
6489 fe->i_mode = cpu_to_le16(inode->i_mode);
6490 if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode))
6491 fe->id1.dev1.i_rdev = cpu_to_le64(huge_encode_dev(dev));
6492 diff -NurpP --minimal linux-2.6.36/fs/ocfs2/ocfs2_fs.h linux-2.6.36-vs2.3.0.36.33/fs/ocfs2/ocfs2_fs.h
6493 --- linux-2.6.36/fs/ocfs2/ocfs2_fs.h 2010-10-21 13:07:51.000000000 +0200
6494 +++ linux-2.6.36-vs2.3.0.36.33/fs/ocfs2/ocfs2_fs.h 2010-10-21 14:00:41.000000000 +0200
6495 @@ -258,6 +258,11 @@
6496 #define OCFS2_TOPDIR_FL FS_TOPDIR_FL /* Top of directory hierarchies*/
6497 #define OCFS2_RESERVED_FL FS_RESERVED_FL /* reserved for ext2 lib */
6499 +#define OCFS2_IXUNLINK_FL FS_IXUNLINK_FL /* Immutable invert on unlink */
6501 +#define OCFS2_BARRIER_FL FS_BARRIER_FL /* Barrier for chroot() */
6502 +#define OCFS2_COW_FL FS_COW_FL /* Copy on Write marker */
6504 #define OCFS2_FL_VISIBLE FS_FL_USER_VISIBLE /* User visible flags */
6505 #define OCFS2_FL_MODIFIABLE FS_FL_USER_MODIFIABLE /* User modifiable flags */
6507 diff -NurpP --minimal linux-2.6.36/fs/ocfs2/ocfs2.h linux-2.6.36-vs2.3.0.36.33/fs/ocfs2/ocfs2.h
6508 --- linux-2.6.36/fs/ocfs2/ocfs2.h 2010-08-02 16:52:51.000000000 +0200
6509 +++ linux-2.6.36-vs2.3.0.36.33/fs/ocfs2/ocfs2.h 2010-10-21 13:09:36.000000000 +0200
6510 @@ -256,6 +256,7 @@ enum ocfs2_mount_options
6512 OCFS2_MOUNT_USRQUOTA = 1 << 10, /* We support user quotas */
6513 OCFS2_MOUNT_GRPQUOTA = 1 << 11, /* We support group quotas */
6514 + OCFS2_MOUNT_TAGGED = 1 << 12, /* use tagging */
6517 #define OCFS2_OSB_SOFT_RO 0x0001
6518 diff -NurpP --minimal linux-2.6.36/fs/ocfs2/super.c linux-2.6.36-vs2.3.0.36.33/fs/ocfs2/super.c
6519 --- linux-2.6.36/fs/ocfs2/super.c 2010-10-21 13:07:51.000000000 +0200
6520 +++ linux-2.6.36-vs2.3.0.36.33/fs/ocfs2/super.c 2010-10-21 13:09:36.000000000 +0200
6521 @@ -179,6 +179,7 @@ enum {
6525 + Opt_tag, Opt_notag, Opt_tagid,
6529 @@ -207,6 +208,9 @@ static const match_table_t tokens = {
6530 {Opt_grpquota, "grpquota"},
6531 {Opt_resv_level, "resv_level=%u"},
6532 {Opt_dir_resv_level, "dir_resv_level=%u"},
6534 + {Opt_notag, "notag"},
6535 + {Opt_tagid, "tagid=%u"},
6539 @@ -617,6 +621,13 @@ static int ocfs2_remount(struct super_bl
6543 + if ((osb->s_mount_opt & OCFS2_MOUNT_TAGGED) !=
6544 + (parsed_options.mount_opt & OCFS2_MOUNT_TAGGED)) {
6546 + mlog(ML_ERROR, "Cannot change tagging on remount\n");
6550 if ((osb->s_mount_opt & OCFS2_MOUNT_HB_LOCAL) !=
6551 (parsed_options.mount_opt & OCFS2_MOUNT_HB_LOCAL)) {
6553 @@ -1153,6 +1164,9 @@ static int ocfs2_fill_super(struct super
6555 ocfs2_complete_mount_recovery(osb);
6557 + if (osb->s_mount_opt & OCFS2_MOUNT_TAGGED)
6558 + sb->s_flags |= MS_TAGGED;
6560 if (ocfs2_mount_local(osb))
6561 snprintf(nodestr, sizeof(nodestr), "local");
6563 @@ -1468,6 +1482,20 @@ static int ocfs2_parse_options(struct su
6564 option < OCFS2_MAX_RESV_LEVEL)
6565 mopt->dir_resv_level = option;
6567 +#ifndef CONFIG_TAGGING_NONE
6569 + mopt->mount_opt |= OCFS2_MOUNT_TAGGED;
6572 + mopt->mount_opt &= ~OCFS2_MOUNT_TAGGED;
6575 +#ifdef CONFIG_PROPAGATE
6578 + mopt->mount_opt |= OCFS2_MOUNT_TAGGED;
6583 "Unrecognized mount option \"%s\" "
6584 diff -NurpP --minimal linux-2.6.36/fs/open.c linux-2.6.36-vs2.3.0.36.33/fs/open.c
6585 --- linux-2.6.36/fs/open.c 2010-10-21 13:07:51.000000000 +0200
6586 +++ linux-2.6.36-vs2.3.0.36.33/fs/open.c 2010-10-21 13:55:41.000000000 +0200
6588 #include <linux/fs_struct.h>
6589 #include <linux/ima.h>
6590 #include <linux/dnotify.h>
6591 +#include <linux/vs_base.h>
6592 +#include <linux/vs_limit.h>
6593 +#include <linux/vs_tag.h>
6594 +#include <linux/vs_cowbl.h>
6595 +#include <linux/vserver/dlimit.h>
6597 #include "internal.h"
6599 @@ -481,6 +486,12 @@ SYSCALL_DEFINE3(fchmodat, int, dfd, cons
6600 error = user_path_at(dfd, filename, LOOKUP_FOLLOW, &path);
6604 +#ifdef CONFIG_VSERVER_COWBL
6605 + error = cow_check_and_break(&path);
6607 + goto dput_and_out;
6609 inode = path.dentry->d_inode;
6611 error = mnt_want_write(path.mnt);
6612 @@ -518,11 +529,11 @@ static int chown_common(struct path *pat
6613 newattrs.ia_valid = ATTR_CTIME;
6614 if (user != (uid_t) -1) {
6615 newattrs.ia_valid |= ATTR_UID;
6616 - newattrs.ia_uid = user;
6617 + newattrs.ia_uid = dx_map_uid(user);
6619 if (group != (gid_t) -1) {
6620 newattrs.ia_valid |= ATTR_GID;
6621 - newattrs.ia_gid = group;
6622 + newattrs.ia_gid = dx_map_gid(group);
6624 if (!S_ISDIR(inode->i_mode))
6625 newattrs.ia_valid |=
6626 @@ -547,6 +558,10 @@ SYSCALL_DEFINE3(chown, const char __user
6627 error = mnt_want_write(path.mnt);
6630 +#ifdef CONFIG_VSERVER_COWBL
6631 + error = cow_check_and_break(&path);
6634 error = chown_common(&path, user, group);
6635 mnt_drop_write(path.mnt);
6637 @@ -572,6 +587,10 @@ SYSCALL_DEFINE5(fchownat, int, dfd, cons
6638 error = mnt_want_write(path.mnt);
6641 +#ifdef CONFIG_VSERVER_COWBL
6642 + error = cow_check_and_break(&path);
6645 error = chown_common(&path, user, group);
6646 mnt_drop_write(path.mnt);
6648 @@ -591,6 +610,10 @@ SYSCALL_DEFINE3(lchown, const char __use
6649 error = mnt_want_write(path.mnt);
6652 +#ifdef CONFIG_VSERVER_COWBL
6653 + error = cow_check_and_break(&path);
6656 error = chown_common(&path, user, group);
6657 mnt_drop_write(path.mnt);
6659 @@ -837,6 +860,7 @@ static void __put_unused_fd(struct files
6660 __FD_CLR(fd, fdt->open_fds);
6661 if (fd < files->next_fd)
6662 files->next_fd = fd;
6663 + vx_openfd_dec(fd);
6666 void put_unused_fd(unsigned int fd)
6667 diff -NurpP --minimal linux-2.6.36/fs/proc/array.c linux-2.6.36-vs2.3.0.36.33/fs/proc/array.c
6668 --- linux-2.6.36/fs/proc/array.c 2010-08-02 16:52:51.000000000 +0200
6669 +++ linux-2.6.36-vs2.3.0.36.33/fs/proc/array.c 2010-10-21 13:09:36.000000000 +0200
6671 #include <linux/pid_namespace.h>
6672 #include <linux/ptrace.h>
6673 #include <linux/tracehook.h>
6674 +#include <linux/vs_context.h>
6675 +#include <linux/vs_network.h>
6677 #include <asm/pgtable.h>
6678 #include <asm/processor.h>
6679 @@ -170,6 +172,9 @@ static inline void task_state(struct seq
6681 ppid = pid_alive(p) ?
6682 task_tgid_nr_ns(rcu_dereference(p->real_parent), ns) : 0;
6683 + if (unlikely(vx_current_initpid(p->pid)))
6688 struct task_struct *tracer = tracehook_tracer_task(p);
6689 @@ -287,7 +292,7 @@ static inline void task_sig(struct seq_f
6692 static void render_cap_t(struct seq_file *m, const char *header,
6694 + struct vx_info *vxi, kernel_cap_t *a)
6698 @@ -312,10 +317,11 @@ static inline void task_cap(struct seq_f
6699 cap_bset = cred->cap_bset;
6702 - render_cap_t(m, "CapInh:\t", &cap_inheritable);
6703 - render_cap_t(m, "CapPrm:\t", &cap_permitted);
6704 - render_cap_t(m, "CapEff:\t", &cap_effective);
6705 - render_cap_t(m, "CapBnd:\t", &cap_bset);
6706 + /* FIXME: maybe move the p->vx_info masking to __task_cred() ? */
6707 + render_cap_t(m, "CapInh:\t", p->vx_info, &cap_inheritable);
6708 + render_cap_t(m, "CapPrm:\t", p->vx_info, &cap_permitted);
6709 + render_cap_t(m, "CapEff:\t", p->vx_info, &cap_effective);
6710 + render_cap_t(m, "CapBnd:\t", p->vx_info, &cap_bset);
6713 static inline void task_context_switch_counts(struct seq_file *m,
6714 @@ -337,6 +343,42 @@ static void task_cpus_allowed(struct seq
6715 seq_printf(m, "\n");
6718 +int proc_pid_nsproxy(struct seq_file *m, struct pid_namespace *ns,
6719 + struct pid *pid, struct task_struct *task)
6721 + seq_printf(m, "Proxy:\t%p(%c)\n"
6729 + (task->nsproxy == init_task.nsproxy ? 'I' : '-'),
6730 + atomic_read(&task->nsproxy->count),
6731 + task->nsproxy->uts_ns,
6732 + (task->nsproxy->uts_ns == init_task.nsproxy->uts_ns ? 'I' : '-'),
6733 + task->nsproxy->ipc_ns,
6734 + (task->nsproxy->ipc_ns == init_task.nsproxy->ipc_ns ? 'I' : '-'),
6735 + task->nsproxy->mnt_ns,
6736 + (task->nsproxy->mnt_ns == init_task.nsproxy->mnt_ns ? 'I' : '-'),
6737 + task->nsproxy->pid_ns,
6738 + (task->nsproxy->pid_ns == init_task.nsproxy->pid_ns ? 'I' : '-'),
6739 + task->nsproxy->net_ns,
6740 + (task->nsproxy->net_ns == init_task.nsproxy->net_ns ? 'I' : '-'));
6744 +void task_vs_id(struct seq_file *m, struct task_struct *task)
6746 + if (task_vx_flags(task, VXF_HIDE_VINFO, 0))
6749 + seq_printf(m, "VxID: %d\n", vx_task_xid(task));
6750 + seq_printf(m, "NxID: %d\n", nx_task_nid(task));
6754 int proc_pid_status(struct seq_file *m, struct pid_namespace *ns,
6755 struct pid *pid, struct task_struct *task)
6757 @@ -353,6 +395,7 @@ int proc_pid_status(struct seq_file *m,
6759 task_cpus_allowed(m, task);
6760 cpuset_task_status_allowed(m, task);
6761 + task_vs_id(m, task);
6762 #if defined(CONFIG_S390)
6763 task_show_regs(m, task);
6765 @@ -465,6 +508,17 @@ static int do_task_stat(struct seq_file
6766 /* convert nsec -> ticks */
6767 start_time = nsec_to_clock_t(start_time);
6769 + /* fixup start time for virt uptime */
6770 + if (vx_flags(VXF_VIRT_UPTIME, 0)) {
6771 + unsigned long long bias =
6772 + current->vx_info->cvirt.bias_clock;
6774 + if (start_time > bias)
6775 + start_time -= bias;
6780 seq_printf(m, "%d (%s) %c %d %d %d %d %d %u %lu \
6781 %lu %lu %lu %lu %lu %ld %ld %ld %ld %d 0 %llu %lu %ld %lu %lu %lu %lu %lu \
6782 %lu %lu %lu %lu %lu %lu %lu %lu %d %d %u %u %llu %lu %ld\n",
6783 diff -NurpP --minimal linux-2.6.36/fs/proc/base.c linux-2.6.36-vs2.3.0.36.33/fs/proc/base.c
6784 --- linux-2.6.36/fs/proc/base.c 2010-10-21 13:07:51.000000000 +0200
6785 +++ linux-2.6.36-vs2.3.0.36.33/fs/proc/base.c 2010-10-21 13:58:32.000000000 +0200
6787 #include <linux/pid_namespace.h>
6788 #include <linux/fs_struct.h>
6789 #include <linux/slab.h>
6790 +#include <linux/vs_context.h>
6791 +#include <linux/vs_network.h>
6792 #include "internal.h"
6795 @@ -1041,12 +1043,17 @@ static ssize_t oom_adjust_write(struct f
6799 - if (oom_adjust < task->signal->oom_adj && !capable(CAP_SYS_RESOURCE)) {
6800 + if (oom_adjust < task->signal->oom_adj &&
6801 + !vx_capable(CAP_SYS_RESOURCE, VXC_OOM_ADJUST)) {
6802 unlock_task_sighand(task, &flags);
6803 put_task_struct(task);
6807 + /* prevent guest processes from circumventing the oom killer */
6808 + if (vx_current_xid() && (oom_adjust == OOM_DISABLE))
6809 + oom_adjust = OOM_ADJUST_MIN;
6812 * Warn that /proc/pid/oom_adj is deprecated, see
6813 * Documentation/feature-removal-schedule.txt.
6814 @@ -1179,7 +1186,7 @@ static ssize_t proc_loginuid_write(struc
6818 - if (!capable(CAP_AUDIT_CONTROL))
6819 + if (!vx_capable(CAP_AUDIT_CONTROL, VXC_AUDIT_CONTROL))
6823 @@ -1617,6 +1624,8 @@ static struct inode *proc_pid_make_inode
6824 inode->i_gid = cred->egid;
6827 + /* procfs is xid tagged */
6828 + inode->i_tag = (tag_t)vx_task_xid(task);
6829 security_task_to_inode(task, inode);
6832 @@ -2167,6 +2176,13 @@ static struct dentry *proc_pident_lookup
6836 + /* TODO: maybe we can come up with a generic approach? */
6837 + if (task_vx_flags(task, VXF_HIDE_VINFO, 0) &&
6838 + (dentry->d_name.len == 5) &&
6839 + (!memcmp(dentry->d_name.name, "vinfo", 5) ||
6840 + !memcmp(dentry->d_name.name, "ninfo", 5)))
6844 * Yes, it does not scale. And it should not. Don't add
6845 * new entries into /proc/<tgid>/ without very good reasons.
6846 @@ -2574,7 +2590,7 @@ out_iput:
6847 static struct dentry *proc_base_lookup(struct inode *dir, struct dentry *dentry)
6849 struct dentry *error;
6850 - struct task_struct *task = get_proc_task(dir);
6851 + struct task_struct *task = get_proc_task_real(dir);
6852 const struct pid_entry *p, *last;
6854 error = ERR_PTR(-ENOENT);
6855 @@ -2664,6 +2680,9 @@ static int proc_pid_personality(struct s
6856 static const struct file_operations proc_task_operations;
6857 static const struct inode_operations proc_task_inode_operations;
6859 +extern int proc_pid_vx_info(struct task_struct *, char *);
6860 +extern int proc_pid_nx_info(struct task_struct *, char *);
6862 static const struct pid_entry tgid_base_stuff[] = {
6863 DIR("task", S_IRUGO|S_IXUGO, proc_task_inode_operations, proc_task_operations),
6864 DIR("fd", S_IRUSR|S_IXUSR, proc_fd_inode_operations, proc_fd_operations),
6865 @@ -2723,6 +2742,8 @@ static const struct pid_entry tgid_base_
6866 #ifdef CONFIG_CGROUPS
6867 REG("cgroup", S_IRUGO, proc_cgroup_operations),
6869 + INF("vinfo", S_IRUGO, proc_pid_vx_info),
6870 + INF("ninfo", S_IRUGO, proc_pid_nx_info),
6871 INF("oom_score", S_IRUGO, proc_oom_score),
6872 REG("oom_adj", S_IRUGO|S_IWUSR, proc_oom_adjust_operations),
6873 REG("oom_score_adj", S_IRUGO|S_IWUSR, proc_oom_score_adj_operations),
6874 @@ -2739,6 +2760,7 @@ static const struct pid_entry tgid_base_
6875 #ifdef CONFIG_TASK_IO_ACCOUNTING
6876 INF("io", S_IRUGO, proc_tgid_io_accounting),
6878 + ONE("nsproxy", S_IRUGO, proc_pid_nsproxy),
6881 static int proc_tgid_base_readdir(struct file * filp,
6882 @@ -2930,7 +2952,7 @@ retry:
6884 pid = find_ge_pid(iter.tgid, ns);
6886 - iter.tgid = pid_nr_ns(pid, ns);
6887 + iter.tgid = pid_unmapped_nr_ns(pid, ns);
6888 iter.task = pid_task(pid, PIDTYPE_PID);
6889 /* What we to know is if the pid we have find is the
6890 * pid of a thread_group_leader. Testing for task
6891 @@ -2960,7 +2982,7 @@ static int proc_pid_fill_cache(struct fi
6892 struct tgid_iter iter)
6894 char name[PROC_NUMBUF];
6895 - int len = snprintf(name, sizeof(name), "%d", iter.tgid);
6896 + int len = snprintf(name, sizeof(name), "%d", vx_map_tgid(iter.tgid));
6897 return proc_fill_cache(filp, dirent, filldir, name, len,
6898 proc_pid_instantiate, iter.task, NULL);
6900 @@ -2969,7 +2991,7 @@ static int proc_pid_fill_cache(struct fi
6901 int proc_pid_readdir(struct file * filp, void * dirent, filldir_t filldir)
6903 unsigned int nr = filp->f_pos - FIRST_PROCESS_ENTRY;
6904 - struct task_struct *reaper = get_proc_task(filp->f_path.dentry->d_inode);
6905 + struct task_struct *reaper = get_proc_task_real(filp->f_path.dentry->d_inode);
6906 struct tgid_iter iter;
6907 struct pid_namespace *ns;
6909 @@ -2989,6 +3011,8 @@ int proc_pid_readdir(struct file * filp,
6911 iter.tgid += 1, iter = next_tgid(ns, iter)) {
6912 filp->f_pos = iter.tgid + TGID_OFFSET;
6913 + if (!vx_proc_task_visible(iter.task))
6915 if (proc_pid_fill_cache(filp, dirent, filldir, iter) < 0) {
6916 put_task_struct(iter.task);
6918 @@ -3137,6 +3161,8 @@ static struct dentry *proc_task_lookup(s
6919 tid = name_to_int(dentry);
6922 + if (vx_current_initpid(tid))
6925 ns = dentry->d_sb->s_fs_info;
6927 diff -NurpP --minimal linux-2.6.36/fs/proc/generic.c linux-2.6.36-vs2.3.0.36.33/fs/proc/generic.c
6928 --- linux-2.6.36/fs/proc/generic.c 2010-10-21 13:07:51.000000000 +0200
6929 +++ linux-2.6.36-vs2.3.0.36.33/fs/proc/generic.c 2010-10-21 13:09:36.000000000 +0200
6931 #include <linux/bitops.h>
6932 #include <linux/spinlock.h>
6933 #include <linux/completion.h>
6934 +#include <linux/vserver/inode.h>
6935 #include <asm/uaccess.h>
6937 #include "internal.h"
6938 @@ -424,6 +425,8 @@ struct dentry *proc_lookup_de(struct pro
6939 for (de = de->subdir; de ; de = de->next) {
6940 if (de->namelen != dentry->d_name.len)
6942 + if (!vx_hide_check(0, de->vx_flags))
6944 if (!memcmp(dentry->d_name.name, de->name, de->namelen)) {
6947 @@ -432,6 +435,8 @@ struct dentry *proc_lookup_de(struct pro
6948 spin_unlock(&proc_subdir_lock);
6950 inode = proc_get_inode(dir->i_sb, ino, de);
6951 + /* generic proc entries belong to the host */
6956 @@ -509,6 +514,8 @@ int proc_readdir_de(struct proc_dir_entr
6958 /* filldir passes info to user space */
6960 + if (!vx_hide_check(0, de->vx_flags))
6962 spin_unlock(&proc_subdir_lock);
6963 if (filldir(dirent, de->name, de->namelen, filp->f_pos,
6964 de->low_ino, de->mode >> 12) < 0) {
6965 @@ -516,6 +523,7 @@ int proc_readdir_de(struct proc_dir_entr
6968 spin_lock(&proc_subdir_lock);
6973 @@ -630,6 +638,7 @@ static struct proc_dir_entry *__proc_cre
6975 atomic_set(&ent->count, 1);
6977 + ent->vx_flags = IATTR_PROC_DEFAULT;
6978 spin_lock_init(&ent->pde_unload_lock);
6979 ent->pde_unload_completion = NULL;
6980 INIT_LIST_HEAD(&ent->pde_openers);
6981 @@ -653,7 +662,8 @@ struct proc_dir_entry *proc_symlink(cons
6987 + ent->vx_flags = IATTR_PROC_SYMLINK;
6991 diff -NurpP --minimal linux-2.6.36/fs/proc/inode.c linux-2.6.36-vs2.3.0.36.33/fs/proc/inode.c
6992 --- linux-2.6.36/fs/proc/inode.c 2010-10-21 13:07:51.000000000 +0200
6993 +++ linux-2.6.36-vs2.3.0.36.33/fs/proc/inode.c 2010-10-21 13:09:36.000000000 +0200
6994 @@ -428,6 +428,8 @@ struct inode *proc_get_inode(struct supe
6995 inode->i_uid = de->uid;
6996 inode->i_gid = de->gid;
6999 + PROC_I(inode)->vx_flags = de->vx_flags;
7001 inode->i_size = de->size;
7003 diff -NurpP --minimal linux-2.6.36/fs/proc/internal.h linux-2.6.36-vs2.3.0.36.33/fs/proc/internal.h
7004 --- linux-2.6.36/fs/proc/internal.h 2010-02-25 11:52:06.000000000 +0100
7005 +++ linux-2.6.36-vs2.3.0.36.33/fs/proc/internal.h 2010-10-21 13:09:36.000000000 +0200
7009 #include <linux/proc_fs.h>
7010 +#include <linux/vs_pid.h>
7012 extern struct proc_dir_entry proc_root;
7013 #ifdef CONFIG_PROC_SYSCTL
7014 @@ -51,6 +52,9 @@ extern int proc_pid_status(struct seq_fi
7015 struct pid *pid, struct task_struct *task);
7016 extern int proc_pid_statm(struct seq_file *m, struct pid_namespace *ns,
7017 struct pid *pid, struct task_struct *task);
7018 +extern int proc_pid_nsproxy(struct seq_file *m, struct pid_namespace *ns,
7019 + struct pid *pid, struct task_struct *task);
7021 extern loff_t mem_lseek(struct file *file, loff_t offset, int orig);
7023 extern const struct file_operations proc_maps_operations;
7024 @@ -68,11 +72,16 @@ static inline struct pid *proc_pid(struc
7025 return PROC_I(inode)->pid;
7028 -static inline struct task_struct *get_proc_task(struct inode *inode)
7029 +static inline struct task_struct *get_proc_task_real(struct inode *inode)
7031 return get_pid_task(proc_pid(inode), PIDTYPE_PID);
7034 +static inline struct task_struct *get_proc_task(struct inode *inode)
7036 + return vx_get_proc_task(inode, proc_pid(inode));
7039 static inline int proc_fd(struct inode *inode)
7041 return PROC_I(inode)->fd;
7042 diff -NurpP --minimal linux-2.6.36/fs/proc/loadavg.c linux-2.6.36-vs2.3.0.36.33/fs/proc/loadavg.c
7043 --- linux-2.6.36/fs/proc/loadavg.c 2009-09-10 15:26:23.000000000 +0200
7044 +++ linux-2.6.36-vs2.3.0.36.33/fs/proc/loadavg.c 2010-10-21 13:09:36.000000000 +0200
7047 static int loadavg_proc_show(struct seq_file *m, void *v)
7049 + unsigned long running;
7050 + unsigned int threads;
7051 unsigned long avnrun[3];
7053 get_avenrun(avnrun, FIXED_1/200, 0);
7055 + if (vx_flags(VXF_VIRT_LOAD, 0)) {
7056 + struct vx_info *vxi = current_vx_info();
7058 + running = atomic_read(&vxi->cvirt.nr_running);
7059 + threads = atomic_read(&vxi->cvirt.nr_threads);
7061 + running = nr_running();
7062 + threads = nr_threads;
7065 seq_printf(m, "%lu.%02lu %lu.%02lu %lu.%02lu %ld/%d %d\n",
7066 LOAD_INT(avnrun[0]), LOAD_FRAC(avnrun[0]),
7067 LOAD_INT(avnrun[1]), LOAD_FRAC(avnrun[1]),
7068 LOAD_INT(avnrun[2]), LOAD_FRAC(avnrun[2]),
7069 - nr_running(), nr_threads,
7071 task_active_pid_ns(current)->last_pid);
7074 diff -NurpP --minimal linux-2.6.36/fs/proc/meminfo.c linux-2.6.36-vs2.3.0.36.33/fs/proc/meminfo.c
7075 --- linux-2.6.36/fs/proc/meminfo.c 2009-12-03 20:02:53.000000000 +0100
7076 +++ linux-2.6.36-vs2.3.0.36.33/fs/proc/meminfo.c 2010-10-21 13:09:36.000000000 +0200
7077 @@ -39,7 +39,8 @@ static int meminfo_proc_show(struct seq_
7078 allowed = ((totalram_pages - hugetlb_total_pages())
7079 * sysctl_overcommit_ratio / 100) + total_swap_pages;
7081 - cached = global_page_state(NR_FILE_PAGES) -
7082 + cached = vx_flags(VXF_VIRT_MEM, 0) ?
7083 + vx_vsi_cached(&i) : global_page_state(NR_FILE_PAGES) -
7084 total_swapcache_pages - i.bufferram;
7087 diff -NurpP --minimal linux-2.6.36/fs/proc/root.c linux-2.6.36-vs2.3.0.36.33/fs/proc/root.c
7088 --- linux-2.6.36/fs/proc/root.c 2010-08-02 16:52:51.000000000 +0200
7089 +++ linux-2.6.36-vs2.3.0.36.33/fs/proc/root.c 2010-10-21 13:09:36.000000000 +0200
7091 #include <linux/bitops.h>
7092 #include <linux/mount.h>
7093 #include <linux/pid_namespace.h>
7094 +#include <linux/vserver/inode.h>
7096 #include "internal.h"
7098 +struct proc_dir_entry *proc_virtual;
7100 +extern void proc_vx_init(void);
7102 static int proc_test_super(struct super_block *sb, void *data)
7104 return sb->s_fs_info == data;
7105 @@ -135,6 +140,7 @@ void __init proc_root_init(void)
7107 proc_mkdir("bus", NULL);
7112 static int proc_root_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat
7113 @@ -202,6 +208,7 @@ struct proc_dir_entry proc_root = {
7114 .proc_iops = &proc_root_inode_operations,
7115 .proc_fops = &proc_root_operations,
7116 .parent = &proc_root,
7117 + .vx_flags = IATTR_ADMIN | IATTR_WATCH,
7120 int pid_ns_prepare_proc(struct pid_namespace *ns)
7121 diff -NurpP --minimal linux-2.6.36/fs/proc/uptime.c linux-2.6.36-vs2.3.0.36.33/fs/proc/uptime.c
7122 --- linux-2.6.36/fs/proc/uptime.c 2009-12-03 20:02:53.000000000 +0100
7123 +++ linux-2.6.36-vs2.3.0.36.33/fs/proc/uptime.c 2010-10-21 13:09:36.000000000 +0200
7125 #include <linux/sched.h>
7126 #include <linux/seq_file.h>
7127 #include <linux/time.h>
7128 -#include <linux/kernel_stat.h>
7129 +#include <linux/vserver/cvirt.h>
7130 #include <asm/cputime.h>
7132 static int uptime_proc_show(struct seq_file *m, void *v)
7134 struct timespec uptime;
7135 struct timespec idle;
7137 - cputime_t idletime = cputime_zero;
7139 - for_each_possible_cpu(i)
7140 - idletime = cputime64_add(idletime, kstat_cpu(i).cpustat.idle);
7141 + cputime_t idletime = cputime_add(init_task.utime, init_task.stime);
7143 do_posix_clock_monotonic_gettime(&uptime);
7144 monotonic_to_bootbased(&uptime);
7145 cputime_to_timespec(idletime, &idle);
7147 + if (vx_flags(VXF_VIRT_UPTIME, 0))
7148 + vx_vsi_uptime(&uptime, &idle);
7150 seq_printf(m, "%lu.%02lu %lu.%02lu\n",
7151 (unsigned long) uptime.tv_sec,
7152 (uptime.tv_nsec / (NSEC_PER_SEC / 100)),
7153 diff -NurpP --minimal linux-2.6.36/fs/quota/dquot.c linux-2.6.36-vs2.3.0.36.33/fs/quota/dquot.c
7154 --- linux-2.6.36/fs/quota/dquot.c 2010-10-21 13:07:51.000000000 +0200
7155 +++ linux-2.6.36-vs2.3.0.36.33/fs/quota/dquot.c 2010-10-21 13:09:36.000000000 +0200
7156 @@ -1538,6 +1538,9 @@ int __dquot_alloc_space(struct inode *in
7157 int reserve = flags & DQUOT_SPACE_RESERVE;
7158 int nofail = flags & DQUOT_SPACE_NOFAIL;
7160 + if ((ret = dl_alloc_space(inode, number)))
7164 * First test before acquiring mutex - solves deadlocks when we
7165 * re-enter the quota code and are already holding the mutex
7166 @@ -1592,6 +1595,9 @@ int dquot_alloc_inode(const struct inode
7168 char warntype[MAXQUOTAS];
7170 + if ((ret = dl_alloc_inode(inode)))
7173 /* First test before acquiring mutex - solves deadlocks when we
7174 * re-enter the quota code and are already holding the mutex */
7175 if (!dquot_active(inode))
7176 @@ -1662,6 +1668,8 @@ void __dquot_free_space(struct inode *in
7177 char warntype[MAXQUOTAS];
7178 int reserve = flags & DQUOT_SPACE_RESERVE;
7180 + dl_free_space(inode, number);
7182 /* First test before acquiring mutex - solves deadlocks when we
7183 * re-enter the quota code and are already holding the mutex */
7184 if (!dquot_active(inode)) {
7185 @@ -1700,6 +1708,8 @@ void dquot_free_inode(const struct inode
7187 char warntype[MAXQUOTAS];
7189 + dl_free_inode(inode);
7191 /* First test before acquiring mutex - solves deadlocks when we
7192 * re-enter the quota code and are already holding the mutex */
7193 if (!dquot_active(inode))
7194 diff -NurpP --minimal linux-2.6.36/fs/quota/quota.c linux-2.6.36-vs2.3.0.36.33/fs/quota/quota.c
7195 --- linux-2.6.36/fs/quota/quota.c 2010-08-02 16:52:51.000000000 +0200
7196 +++ linux-2.6.36-vs2.3.0.36.33/fs/quota/quota.c 2010-10-21 13:09:36.000000000 +0200
7198 #include <linux/fs.h>
7199 #include <linux/namei.h>
7200 #include <linux/slab.h>
7201 +#include <linux/vs_context.h>
7202 #include <asm/current.h>
7203 #include <asm/uaccess.h>
7204 #include <linux/kernel.h>
7205 @@ -38,7 +39,7 @@ static int check_quotactl_permission(str
7209 - if (!capable(CAP_SYS_ADMIN))
7210 + if (!vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL))
7214 @@ -296,6 +297,46 @@ static int do_quotactl(struct super_bloc
7218 +#if defined(CONFIG_BLK_DEV_VROOT) || defined(CONFIG_BLK_DEV_VROOT_MODULE)
7220 +#include <linux/vroot.h>
7221 +#include <linux/major.h>
7222 +#include <linux/module.h>
7223 +#include <linux/kallsyms.h>
7224 +#include <linux/vserver/debug.h>
7226 +static vroot_grb_func *vroot_get_real_bdev = NULL;
7228 +static spinlock_t vroot_grb_lock = SPIN_LOCK_UNLOCKED;
7230 +int register_vroot_grb(vroot_grb_func *func) {
7233 + spin_lock(&vroot_grb_lock);
7234 + if (!vroot_get_real_bdev) {
7235 + vroot_get_real_bdev = func;
7238 + spin_unlock(&vroot_grb_lock);
7241 +EXPORT_SYMBOL(register_vroot_grb);
7243 +int unregister_vroot_grb(vroot_grb_func *func) {
7244 + int ret = -EINVAL;
7246 + spin_lock(&vroot_grb_lock);
7247 + if (vroot_get_real_bdev) {
7248 + vroot_get_real_bdev = NULL;
7251 + spin_unlock(&vroot_grb_lock);
7254 +EXPORT_SYMBOL(unregister_vroot_grb);
7259 * look up a superblock on which quota ops will be performed
7260 * - use the name of a block device to find the superblock thereon
7261 @@ -313,6 +354,22 @@ static struct super_block *quotactl_bloc
7264 return ERR_CAST(bdev);
7265 +#if defined(CONFIG_BLK_DEV_VROOT) || defined(CONFIG_BLK_DEV_VROOT_MODULE)
7266 + if (bdev && bdev->bd_inode &&
7267 + imajor(bdev->bd_inode) == VROOT_MAJOR) {
7268 + struct block_device *bdnew = (void *)-EINVAL;
7270 + if (vroot_get_real_bdev)
7271 + bdnew = vroot_get_real_bdev(bdev);
7273 + vxdprintk(VXD_CBIT(misc, 0),
7274 + "vroot_get_real_bdev not set");
7276 + if (IS_ERR(bdnew))
7277 + return ERR_PTR(PTR_ERR(bdnew));
7281 sb = get_super(bdev);
7284 diff -NurpP --minimal linux-2.6.36/fs/reiserfs/file.c linux-2.6.36-vs2.3.0.36.33/fs/reiserfs/file.c
7285 --- linux-2.6.36/fs/reiserfs/file.c 2010-10-21 13:07:51.000000000 +0200
7286 +++ linux-2.6.36-vs2.3.0.36.33/fs/reiserfs/file.c 2010-10-21 13:09:36.000000000 +0200
7287 @@ -313,4 +313,5 @@ const struct inode_operations reiserfs_f
7288 .listxattr = reiserfs_listxattr,
7289 .removexattr = reiserfs_removexattr,
7290 .permission = reiserfs_permission,
7291 + .sync_flags = reiserfs_sync_flags,
7293 diff -NurpP --minimal linux-2.6.36/fs/reiserfs/inode.c linux-2.6.36-vs2.3.0.36.33/fs/reiserfs/inode.c
7294 --- linux-2.6.36/fs/reiserfs/inode.c 2010-10-21 13:07:51.000000000 +0200
7295 +++ linux-2.6.36-vs2.3.0.36.33/fs/reiserfs/inode.c 2010-10-21 13:59:16.000000000 +0200
7297 #include <linux/writeback.h>
7298 #include <linux/quotaops.h>
7299 #include <linux/swap.h>
7300 +#include <linux/vs_tag.h>
7302 int reiserfs_commit_write(struct file *f, struct page *page,
7303 unsigned from, unsigned to);
7304 @@ -1133,6 +1134,8 @@ static void init_inode(struct inode *ino
7305 struct buffer_head *bh;
7306 struct item_head *ih;
7310 //int version = ITEM_VERSION_1;
7312 bh = PATH_PLAST_BUFFER(path);
7313 @@ -1153,12 +1156,13 @@ static void init_inode(struct inode *ino
7314 (struct stat_data_v1 *)B_I_PITEM(bh, ih);
7315 unsigned long blocks;
7317 + uid = sd_v1_uid(sd);
7318 + gid = sd_v1_gid(sd);
7320 set_inode_item_key_version(inode, KEY_FORMAT_3_5);
7321 set_inode_sd_version(inode, STAT_DATA_V1);
7322 inode->i_mode = sd_v1_mode(sd);
7323 inode->i_nlink = sd_v1_nlink(sd);
7324 - inode->i_uid = sd_v1_uid(sd);
7325 - inode->i_gid = sd_v1_gid(sd);
7326 inode->i_size = sd_v1_size(sd);
7327 inode->i_atime.tv_sec = sd_v1_atime(sd);
7328 inode->i_mtime.tv_sec = sd_v1_mtime(sd);
7329 @@ -1200,11 +1204,12 @@ static void init_inode(struct inode *ino
7330 // (directories and symlinks)
7331 struct stat_data *sd = (struct stat_data *)B_I_PITEM(bh, ih);
7333 + uid = sd_v2_uid(sd);
7334 + gid = sd_v2_gid(sd);
7336 inode->i_mode = sd_v2_mode(sd);
7337 inode->i_nlink = sd_v2_nlink(sd);
7338 - inode->i_uid = sd_v2_uid(sd);
7339 inode->i_size = sd_v2_size(sd);
7340 - inode->i_gid = sd_v2_gid(sd);
7341 inode->i_mtime.tv_sec = sd_v2_mtime(sd);
7342 inode->i_atime.tv_sec = sd_v2_atime(sd);
7343 inode->i_ctime.tv_sec = sd_v2_ctime(sd);
7344 @@ -1234,6 +1239,10 @@ static void init_inode(struct inode *ino
7345 sd_attrs_to_i_attrs(sd_v2_attrs(sd), inode);
7348 + inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
7349 + inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
7350 + inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid, 0);
7353 if (S_ISREG(inode->i_mode)) {
7354 inode->i_op = &reiserfs_file_inode_operations;
7355 @@ -1256,13 +1265,15 @@ static void init_inode(struct inode *ino
7356 static void inode2sd(void *sd, struct inode *inode, loff_t size)
7358 struct stat_data *sd_v2 = (struct stat_data *)sd;
7359 + uid_t uid = TAGINO_UID(DX_TAG(inode), inode->i_uid, inode->i_tag);
7360 + gid_t gid = TAGINO_GID(DX_TAG(inode), inode->i_gid, inode->i_tag);
7363 + set_sd_v2_uid(sd_v2, uid);
7364 + set_sd_v2_gid(sd_v2, gid);
7365 set_sd_v2_mode(sd_v2, inode->i_mode);
7366 set_sd_v2_nlink(sd_v2, inode->i_nlink);
7367 - set_sd_v2_uid(sd_v2, inode->i_uid);
7368 set_sd_v2_size(sd_v2, size);
7369 - set_sd_v2_gid(sd_v2, inode->i_gid);
7370 set_sd_v2_mtime(sd_v2, inode->i_mtime.tv_sec);
7371 set_sd_v2_atime(sd_v2, inode->i_atime.tv_sec);
7372 set_sd_v2_ctime(sd_v2, inode->i_ctime.tv_sec);
7373 @@ -2861,14 +2872,19 @@ int reiserfs_commit_write(struct file *f
7374 void sd_attrs_to_i_attrs(__u16 sd_attrs, struct inode *inode)
7376 if (reiserfs_attrs(inode->i_sb)) {
7377 - if (sd_attrs & REISERFS_SYNC_FL)
7378 - inode->i_flags |= S_SYNC;
7380 - inode->i_flags &= ~S_SYNC;
7381 if (sd_attrs & REISERFS_IMMUTABLE_FL)
7382 inode->i_flags |= S_IMMUTABLE;
7384 inode->i_flags &= ~S_IMMUTABLE;
7385 + if (sd_attrs & REISERFS_IXUNLINK_FL)
7386 + inode->i_flags |= S_IXUNLINK;
7388 + inode->i_flags &= ~S_IXUNLINK;
7390 + if (sd_attrs & REISERFS_SYNC_FL)
7391 + inode->i_flags |= S_SYNC;
7393 + inode->i_flags &= ~S_SYNC;
7394 if (sd_attrs & REISERFS_APPEND_FL)
7395 inode->i_flags |= S_APPEND;
7397 @@ -2881,6 +2897,15 @@ void sd_attrs_to_i_attrs(__u16 sd_attrs,
7398 REISERFS_I(inode)->i_flags |= i_nopack_mask;
7400 REISERFS_I(inode)->i_flags &= ~i_nopack_mask;
7402 + if (sd_attrs & REISERFS_BARRIER_FL)
7403 + inode->i_vflags |= V_BARRIER;
7405 + inode->i_vflags &= ~V_BARRIER;
7406 + if (sd_attrs & REISERFS_COW_FL)
7407 + inode->i_vflags |= V_COW;
7409 + inode->i_vflags &= ~V_COW;
7413 @@ -2891,6 +2916,11 @@ void i_attrs_to_sd_attrs(struct inode *i
7414 *sd_attrs |= REISERFS_IMMUTABLE_FL;
7416 *sd_attrs &= ~REISERFS_IMMUTABLE_FL;
7417 + if (inode->i_flags & S_IXUNLINK)
7418 + *sd_attrs |= REISERFS_IXUNLINK_FL;
7420 + *sd_attrs &= ~REISERFS_IXUNLINK_FL;
7422 if (inode->i_flags & S_SYNC)
7423 *sd_attrs |= REISERFS_SYNC_FL;
7425 @@ -2903,6 +2933,15 @@ void i_attrs_to_sd_attrs(struct inode *i
7426 *sd_attrs |= REISERFS_NOTAIL_FL;
7428 *sd_attrs &= ~REISERFS_NOTAIL_FL;
7430 + if (inode->i_vflags & V_BARRIER)
7431 + *sd_attrs |= REISERFS_BARRIER_FL;
7433 + *sd_attrs &= ~REISERFS_BARRIER_FL;
7434 + if (inode->i_vflags & V_COW)
7435 + *sd_attrs |= REISERFS_COW_FL;
7437 + *sd_attrs &= ~REISERFS_COW_FL;
7441 @@ -3146,7 +3185,8 @@ int reiserfs_setattr(struct dentry *dent
7444 if ((ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) ||
7445 - (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) {
7446 + (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid) ||
7447 + (ia_valid & ATTR_TAG && attr->ia_tag != inode->i_tag)) {
7448 struct reiserfs_transaction_handle th;
7451 @@ -3175,6 +3215,9 @@ int reiserfs_setattr(struct dentry *dent
7452 inode->i_uid = attr->ia_uid;
7453 if (attr->ia_valid & ATTR_GID)
7454 inode->i_gid = attr->ia_gid;
7455 + if ((attr->ia_valid & ATTR_TAG) &&
7457 + inode->i_tag = attr->ia_tag;
7458 mark_inode_dirty(inode);
7459 error = journal_end(&th, inode->i_sb, jbegin_count);
7461 diff -NurpP --minimal linux-2.6.36/fs/reiserfs/ioctl.c linux-2.6.36-vs2.3.0.36.33/fs/reiserfs/ioctl.c
7462 --- linux-2.6.36/fs/reiserfs/ioctl.c 2010-10-21 13:07:51.000000000 +0200
7463 +++ linux-2.6.36-vs2.3.0.36.33/fs/reiserfs/ioctl.c 2010-10-21 13:09:36.000000000 +0200
7465 #include <linux/mount.h>
7466 #include <linux/reiserfs_fs.h>
7467 #include <linux/time.h>
7468 +#include <linux/mount.h>
7469 #include <asm/uaccess.h>
7470 #include <linux/pagemap.h>
7471 #include <linux/smp_lock.h>
7472 #include <linux/compat.h>
7475 +int reiserfs_sync_flags(struct inode *inode, int flags, int vflags)
7477 + __u16 sd_attrs = 0;
7479 + inode->i_flags = flags;
7480 + inode->i_vflags = vflags;
7482 + i_attrs_to_sd_attrs(inode, &sd_attrs);
7483 + REISERFS_I(inode)->i_attrs = sd_attrs;
7484 + inode->i_ctime = CURRENT_TIME_SEC;
7485 + mark_inode_dirty(inode);
7490 * reiserfs_ioctl - handler for ioctl for inode
7491 * supported commands:
7493 long reiserfs_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
7495 struct inode *inode = filp->f_path.dentry->d_inode;
7496 - unsigned int flags;
7497 + unsigned int flags, oldflags;
7500 reiserfs_write_lock(inode->i_sb);
7501 @@ -48,6 +64,7 @@ long reiserfs_ioctl(struct file *filp, u
7503 flags = REISERFS_I(inode)->i_attrs;
7504 i_attrs_to_sd_attrs(inode, (__u16 *) & flags);
7505 + flags &= REISERFS_FL_USER_VISIBLE;
7506 err = put_user(flags, (int __user *)arg);
7508 case REISERFS_IOC_SETFLAGS:{
7509 @@ -68,6 +85,10 @@ long reiserfs_ioctl(struct file *filp, u
7513 + if (IS_BARRIER(inode)) {
7514 + vxwprintk_task(1, "messing with the barrier.");
7518 * Is it quota file? Do not allow user to mess with it
7520 @@ -92,6 +113,10 @@ long reiserfs_ioctl(struct file *filp, u
7525 + oldflags = REISERFS_I(inode)->i_attrs;
7526 + flags &= REISERFS_FL_USER_MODIFIABLE;
7527 + flags |= oldflags & ~REISERFS_FL_USER_MODIFIABLE;
7528 sd_attrs_to_i_attrs(flags, inode);
7529 REISERFS_I(inode)->i_attrs = flags;
7530 inode->i_ctime = CURRENT_TIME_SEC;
7531 diff -NurpP --minimal linux-2.6.36/fs/reiserfs/namei.c linux-2.6.36-vs2.3.0.36.33/fs/reiserfs/namei.c
7532 --- linux-2.6.36/fs/reiserfs/namei.c 2010-08-02 16:52:52.000000000 +0200
7533 +++ linux-2.6.36-vs2.3.0.36.33/fs/reiserfs/namei.c 2010-10-21 13:09:36.000000000 +0200
7535 #include <linux/reiserfs_acl.h>
7536 #include <linux/reiserfs_xattr.h>
7537 #include <linux/quotaops.h>
7538 +#include <linux/vs_tag.h>
7540 #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; }
7541 #define DEC_DIR_INODE_NLINK(i) if (i->i_nlink != 1) drop_nlink(i);
7542 @@ -362,6 +363,7 @@ static struct dentry *reiserfs_lookup(st
7543 if (retval == IO_ERROR) {
7544 return ERR_PTR(-EIO);
7546 + dx_propagate_tag(nd, inode);
7548 return d_splice_alias(inode, dentry);
7550 @@ -1532,6 +1534,7 @@ const struct inode_operations reiserfs_d
7551 .listxattr = reiserfs_listxattr,
7552 .removexattr = reiserfs_removexattr,
7553 .permission = reiserfs_permission,
7554 + .sync_flags = reiserfs_sync_flags,
7558 diff -NurpP --minimal linux-2.6.36/fs/reiserfs/super.c linux-2.6.36-vs2.3.0.36.33/fs/reiserfs/super.c
7559 --- linux-2.6.36/fs/reiserfs/super.c 2010-10-21 13:07:51.000000000 +0200
7560 +++ linux-2.6.36-vs2.3.0.36.33/fs/reiserfs/super.c 2010-10-21 13:09:36.000000000 +0200
7561 @@ -893,6 +893,14 @@ static int reiserfs_parse_options(struct
7562 {"user_xattr",.setmask = 1 << REISERFS_UNSUPPORTED_OPT},
7563 {"nouser_xattr",.clrmask = 1 << REISERFS_UNSUPPORTED_OPT},
7565 +#ifndef CONFIG_TAGGING_NONE
7566 + {"tagxid",.setmask = 1 << REISERFS_TAGGED},
7567 + {"tag",.setmask = 1 << REISERFS_TAGGED},
7568 + {"notag",.clrmask = 1 << REISERFS_TAGGED},
7570 +#ifdef CONFIG_PROPAGATE
7571 + {"tag",.arg_required = 'T',.values = NULL},
7573 #ifdef CONFIG_REISERFS_FS_POSIX_ACL
7574 {"acl",.setmask = 1 << REISERFS_POSIXACL},
7575 {"noacl",.clrmask = 1 << REISERFS_POSIXACL},
7576 @@ -1202,6 +1210,14 @@ static int reiserfs_remount(struct super
7577 handle_quota_files(s, qf_names, &qfmt);
7580 + if ((mount_options & (1 << REISERFS_TAGGED)) &&
7581 + !(s->s_flags & MS_TAGGED)) {
7582 + reiserfs_warning(s, "super-vs01",
7583 + "reiserfs: tagging not permitted on remount.");
7590 /* Add options that are safe here */
7591 @@ -1684,6 +1700,10 @@ static int reiserfs_fill_super(struct su
7595 + /* map mount option tagxid */
7596 + if (REISERFS_SB(s)->s_mount_opt & (1 << REISERFS_TAGGED))
7597 + s->s_flags |= MS_TAGGED;
7599 rs = SB_DISK_SUPER_BLOCK(s);
7600 /* Let's do basic sanity check to verify that underlying device is not
7601 smaller than the filesystem. If the check fails then abort and scream,
7602 diff -NurpP --minimal linux-2.6.36/fs/reiserfs/xattr.c linux-2.6.36-vs2.3.0.36.33/fs/reiserfs/xattr.c
7603 --- linux-2.6.36/fs/reiserfs/xattr.c 2010-08-02 16:52:52.000000000 +0200
7604 +++ linux-2.6.36-vs2.3.0.36.33/fs/reiserfs/xattr.c 2010-10-21 13:09:36.000000000 +0200
7606 #include <linux/errno.h>
7607 #include <linux/gfp.h>
7608 #include <linux/fs.h>
7609 +#include <linux/mount.h>
7610 #include <linux/file.h>
7611 #include <linux/pagemap.h>
7612 #include <linux/xattr.h>
7613 diff -NurpP --minimal linux-2.6.36/fs/stat.c linux-2.6.36-vs2.3.0.36.33/fs/stat.c
7614 --- linux-2.6.36/fs/stat.c 2010-10-21 13:07:51.000000000 +0200
7615 +++ linux-2.6.36-vs2.3.0.36.33/fs/stat.c 2010-10-21 13:09:36.000000000 +0200
7616 @@ -26,6 +26,7 @@ void generic_fillattr(struct inode *inod
7617 stat->nlink = inode->i_nlink;
7618 stat->uid = inode->i_uid;
7619 stat->gid = inode->i_gid;
7620 + stat->tag = inode->i_tag;
7621 stat->rdev = inode->i_rdev;
7622 stat->atime = inode->i_atime;
7623 stat->mtime = inode->i_mtime;
7624 diff -NurpP --minimal linux-2.6.36/fs/statfs.c linux-2.6.36-vs2.3.0.36.33/fs/statfs.c
7625 --- linux-2.6.36/fs/statfs.c 2010-10-21 13:07:51.000000000 +0200
7626 +++ linux-2.6.36-vs2.3.0.36.33/fs/statfs.c 2010-10-21 14:07:35.000000000 +0200
7628 #include <linux/statfs.h>
7629 #include <linux/security.h>
7630 #include <linux/uaccess.h>
7631 +#include <linux/vs_base.h>
7632 +#include <linux/vs_dlimit.h>
7634 static int flags_by_mnt(int mnt_flags)
7636 @@ -59,6 +61,8 @@ int statfs_by_dentry(struct dentry *dent
7637 retval = dentry->d_sb->s_op->statfs(dentry, buf);
7638 if (retval == 0 && buf->f_frsize == 0)
7639 buf->f_frsize = buf->f_bsize;
7640 + if (!vx_check(0, VS_ADMIN|VS_WATCH))
7641 + vx_vsi_statfs(dentry->d_sb, buf);
7645 diff -NurpP --minimal linux-2.6.36/fs/super.c linux-2.6.36-vs2.3.0.36.33/fs/super.c
7646 --- linux-2.6.36/fs/super.c 2010-10-21 13:07:51.000000000 +0200
7647 +++ linux-2.6.36-vs2.3.0.36.33/fs/super.c 2010-10-21 13:09:36.000000000 +0200
7649 #include <linux/idr.h>
7650 #include <linux/mutex.h>
7651 #include <linux/backing-dev.h>
7652 +#include <linux/devpts_fs.h>
7653 +#include <linux/proc_fs.h>
7654 +#include <linux/vs_context.h>
7655 #include "internal.h"
7658 @@ -918,12 +921,18 @@ struct vfsmount *
7659 vfs_kern_mount(struct file_system_type *type, int flags, const char *name, void *data)
7661 struct vfsmount *mnt;
7662 + struct super_block *sb;
7663 char *secdata = NULL;
7667 return ERR_PTR(-ENODEV);
7670 + if ((type->fs_flags & FS_BINARY_MOUNTDATA) &&
7671 + !vx_capable(CAP_SYS_ADMIN, VXC_BINARY_MOUNT))
7675 mnt = alloc_vfsmnt(name);
7677 @@ -945,11 +954,19 @@ vfs_kern_mount(struct file_system_type *
7678 error = type->get_sb(type, flags, name, data, mnt);
7680 goto out_free_secdata;
7681 - BUG_ON(!mnt->mnt_sb);
7682 - WARN_ON(!mnt->mnt_sb->s_bdi);
7686 + WARN_ON(!sb->s_bdi);
7687 mnt->mnt_sb->s_flags |= MS_BORN;
7689 - error = security_sb_kern_mount(mnt->mnt_sb, flags, secdata);
7691 + if (!vx_capable(CAP_SYS_ADMIN, VXC_BINARY_MOUNT) && !sb->s_bdev &&
7692 + (sb->s_magic != PROC_SUPER_MAGIC) &&
7693 + (sb->s_magic != DEVPTS_SUPER_MAGIC))
7696 + error = security_sb_kern_mount(sb, flags, secdata);
7700 diff -NurpP --minimal linux-2.6.36/fs/sysfs/mount.c linux-2.6.36-vs2.3.0.36.33/fs/sysfs/mount.c
7701 --- linux-2.6.36/fs/sysfs/mount.c 2010-10-21 13:07:51.000000000 +0200
7702 +++ linux-2.6.36-vs2.3.0.36.33/fs/sysfs/mount.c 2010-10-21 13:09:36.000000000 +0200
7703 @@ -47,7 +47,7 @@ static int sysfs_fill_super(struct super
7705 sb->s_blocksize = PAGE_CACHE_SIZE;
7706 sb->s_blocksize_bits = PAGE_CACHE_SHIFT;
7707 - sb->s_magic = SYSFS_MAGIC;
7708 + sb->s_magic = SYSFS_SUPER_MAGIC;
7709 sb->s_op = &sysfs_ops;
7710 sb->s_time_gran = 1;
7712 diff -NurpP --minimal linux-2.6.36/fs/utimes.c linux-2.6.36-vs2.3.0.36.33/fs/utimes.c
7713 --- linux-2.6.36/fs/utimes.c 2010-10-21 13:07:51.000000000 +0200
7714 +++ linux-2.6.36-vs2.3.0.36.33/fs/utimes.c 2010-10-21 13:09:36.000000000 +0200
7716 #include <linux/stat.h>
7717 #include <linux/utime.h>
7718 #include <linux/syscalls.h>
7719 +#include <linux/mount.h>
7720 +#include <linux/vs_cowbl.h>
7721 #include <asm/uaccess.h>
7722 #include <asm/unistd.h>
7724 @@ -52,12 +54,18 @@ static int utimes_common(struct path *pa
7727 struct iattr newattrs;
7728 - struct inode *inode = path->dentry->d_inode;
7729 + struct inode *inode;
7731 error = mnt_want_write(path->mnt);
7735 + error = cow_check_and_break(path);
7737 + goto mnt_drop_write_and_out;
7739 + inode = path->dentry->d_inode;
7741 if (times && times[0].tv_nsec == UTIME_NOW &&
7742 times[1].tv_nsec == UTIME_NOW)
7744 diff -NurpP --minimal linux-2.6.36/fs/xattr.c linux-2.6.36-vs2.3.0.36.33/fs/xattr.c
7745 --- linux-2.6.36/fs/xattr.c 2010-08-02 16:52:52.000000000 +0200
7746 +++ linux-2.6.36-vs2.3.0.36.33/fs/xattr.c 2010-10-21 13:09:36.000000000 +0200
7748 #include <linux/module.h>
7749 #include <linux/fsnotify.h>
7750 #include <linux/audit.h>
7751 +#include <linux/mount.h>
7752 #include <asm/uaccess.h>
7755 @@ -49,7 +50,7 @@ xattr_permission(struct inode *inode, co
7756 * The trusted.* namespace can only be accessed by a privileged user.
7758 if (!strncmp(name, XATTR_TRUSTED_PREFIX, XATTR_TRUSTED_PREFIX_LEN))
7759 - return (capable(CAP_SYS_ADMIN) ? 0 : -EPERM);
7760 + return (vx_capable(CAP_SYS_ADMIN, VXC_FS_TRUSTED) ? 0 : -EPERM);
7762 /* In user.* namespace, only regular files and directories can have
7763 * extended attributes. For sticky directories, only the owner and
7764 diff -NurpP --minimal linux-2.6.36/fs/xfs/linux-2.6/xfs_ioctl.c linux-2.6.36-vs2.3.0.36.33/fs/xfs/linux-2.6/xfs_ioctl.c
7765 --- linux-2.6.36/fs/xfs/linux-2.6/xfs_ioctl.c 2010-10-21 13:07:51.000000000 +0200
7766 +++ linux-2.6.36-vs2.3.0.36.33/fs/xfs/linux-2.6/xfs_ioctl.c 2010-10-21 14:02:37.000000000 +0200
7768 #include "xfs_bmap_btree.h"
7769 #include "xfs_dinode.h"
7770 #include "xfs_inode.h"
7771 -#include "xfs_ioctl.h"
7772 +// #include "xfs_ioctl.h"
7773 #include "xfs_rtalloc.h"
7774 #include "xfs_itable.h"
7775 #include "xfs_error.h"
7776 @@ -738,6 +738,10 @@ xfs_merge_ioc_xflags(
7777 xflags |= XFS_XFLAG_IMMUTABLE;
7779 xflags &= ~XFS_XFLAG_IMMUTABLE;
7780 + if (flags & FS_IXUNLINK_FL)
7781 + xflags |= XFS_XFLAG_IXUNLINK;
7783 + xflags &= ~XFS_XFLAG_IXUNLINK;
7784 if (flags & FS_APPEND_FL)
7785 xflags |= XFS_XFLAG_APPEND;
7787 @@ -766,6 +770,8 @@ xfs_di2lxflags(
7789 if (di_flags & XFS_DIFLAG_IMMUTABLE)
7790 flags |= FS_IMMUTABLE_FL;
7791 + if (di_flags & XFS_DIFLAG_IXUNLINK)
7792 + flags |= FS_IXUNLINK_FL;
7793 if (di_flags & XFS_DIFLAG_APPEND)
7794 flags |= FS_APPEND_FL;
7795 if (di_flags & XFS_DIFLAG_SYNC)
7796 @@ -826,6 +832,8 @@ xfs_set_diflags(
7797 di_flags = (ip->i_d.di_flags & XFS_DIFLAG_PREALLOC);
7798 if (xflags & XFS_XFLAG_IMMUTABLE)
7799 di_flags |= XFS_DIFLAG_IMMUTABLE;
7800 + if (xflags & XFS_XFLAG_IXUNLINK)
7801 + di_flags |= XFS_DIFLAG_IXUNLINK;
7802 if (xflags & XFS_XFLAG_APPEND)
7803 di_flags |= XFS_DIFLAG_APPEND;
7804 if (xflags & XFS_XFLAG_SYNC)
7805 @@ -868,6 +876,10 @@ xfs_diflags_to_linux(
7806 inode->i_flags |= S_IMMUTABLE;
7808 inode->i_flags &= ~S_IMMUTABLE;
7809 + if (xflags & XFS_XFLAG_IXUNLINK)
7810 + inode->i_flags |= S_IXUNLINK;
7812 + inode->i_flags &= ~S_IXUNLINK;
7813 if (xflags & XFS_XFLAG_APPEND)
7814 inode->i_flags |= S_APPEND;
7816 @@ -1341,10 +1353,18 @@ xfs_file_ioctl(
7817 case XFS_IOC_FSGETXATTRA:
7818 return xfs_ioc_fsgetxattr(ip, 1, arg);
7819 case XFS_IOC_FSSETXATTR:
7820 + if (IS_BARRIER(inode)) {
7821 + vxwprintk_task(1, "messing with the barrier.");
7822 + return -XFS_ERROR(EACCES);
7824 return xfs_ioc_fssetxattr(ip, filp, arg);
7825 case XFS_IOC_GETXFLAGS:
7826 return xfs_ioc_getxflags(ip, arg);
7827 case XFS_IOC_SETXFLAGS:
7828 + if (IS_BARRIER(inode)) {
7829 + vxwprintk_task(1, "messing with the barrier.");
7830 + return -XFS_ERROR(EACCES);
7832 return xfs_ioc_setxflags(ip, filp, arg);
7834 case XFS_IOC_FSSETDM: {
7835 diff -NurpP --minimal linux-2.6.36/fs/xfs/linux-2.6/xfs_ioctl.h linux-2.6.36-vs2.3.0.36.33/fs/xfs/linux-2.6/xfs_ioctl.h
7836 --- linux-2.6.36/fs/xfs/linux-2.6/xfs_ioctl.h 2010-07-07 18:31:54.000000000 +0200
7837 +++ linux-2.6.36-vs2.3.0.36.33/fs/xfs/linux-2.6/xfs_ioctl.h 2010-10-21 13:09:36.000000000 +0200
7838 @@ -70,6 +70,12 @@ xfs_handle_to_dentry(
7839 void __user *uhandle,
7844 + struct inode *inode,
7851 diff -NurpP --minimal linux-2.6.36/fs/xfs/linux-2.6/xfs_iops.c linux-2.6.36-vs2.3.0.36.33/fs/xfs/linux-2.6/xfs_iops.c
7852 --- linux-2.6.36/fs/xfs/linux-2.6/xfs_iops.c 2010-10-21 13:07:51.000000000 +0200
7853 +++ linux-2.6.36-vs2.3.0.36.33/fs/xfs/linux-2.6/xfs_iops.c 2010-10-21 13:09:36.000000000 +0200
7855 #include "xfs_bmap_btree.h"
7856 #include "xfs_dinode.h"
7857 #include "xfs_inode.h"
7858 +#include "xfs_ioctl.h"
7859 #include "xfs_bmap.h"
7860 #include "xfs_rtalloc.h"
7861 #include "xfs_error.h"
7863 #include <linux/falloc.h>
7864 #include <linux/fiemap.h>
7865 #include <linux/slab.h>
7866 +#include <linux/vs_tag.h>
7869 * Bring the timestamps in the XFS inode uptodate.
7870 @@ -499,6 +501,7 @@ xfs_vn_getattr(
7871 stat->nlink = ip->i_d.di_nlink;
7872 stat->uid = ip->i_d.di_uid;
7873 stat->gid = ip->i_d.di_gid;
7874 + stat->tag = ip->i_d.di_tag;
7875 stat->ino = ip->i_ino;
7876 stat->atime = inode->i_atime;
7877 stat->mtime = inode->i_mtime;
7878 @@ -687,6 +690,7 @@ static const struct inode_operations xfs
7879 .listxattr = xfs_vn_listxattr,
7880 .fallocate = xfs_vn_fallocate,
7881 .fiemap = xfs_vn_fiemap,
7882 + .sync_flags = xfs_sync_flags,
7885 static const struct inode_operations xfs_dir_inode_operations = {
7886 @@ -712,6 +716,7 @@ static const struct inode_operations xfs
7887 .getxattr = generic_getxattr,
7888 .removexattr = generic_removexattr,
7889 .listxattr = xfs_vn_listxattr,
7890 + .sync_flags = xfs_sync_flags,
7893 static const struct inode_operations xfs_dir_ci_inode_operations = {
7894 @@ -761,6 +766,10 @@ xfs_diflags_to_iflags(
7895 inode->i_flags |= S_IMMUTABLE;
7897 inode->i_flags &= ~S_IMMUTABLE;
7898 + if (ip->i_d.di_flags & XFS_DIFLAG_IXUNLINK)
7899 + inode->i_flags |= S_IXUNLINK;
7901 + inode->i_flags &= ~S_IXUNLINK;
7902 if (ip->i_d.di_flags & XFS_DIFLAG_APPEND)
7903 inode->i_flags |= S_APPEND;
7905 @@ -773,6 +782,15 @@ xfs_diflags_to_iflags(
7906 inode->i_flags |= S_NOATIME;
7908 inode->i_flags &= ~S_NOATIME;
7910 + if (ip->i_d.di_vflags & XFS_DIVFLAG_BARRIER)
7911 + inode->i_vflags |= V_BARRIER;
7913 + inode->i_vflags &= ~V_BARRIER;
7914 + if (ip->i_d.di_vflags & XFS_DIVFLAG_COW)
7915 + inode->i_vflags |= V_COW;
7917 + inode->i_vflags &= ~V_COW;
7921 @@ -801,6 +819,7 @@ xfs_setup_inode(
7922 inode->i_nlink = ip->i_d.di_nlink;
7923 inode->i_uid = ip->i_d.di_uid;
7924 inode->i_gid = ip->i_d.di_gid;
7925 + inode->i_tag = ip->i_d.di_tag;
7927 switch (inode->i_mode & S_IFMT) {
7929 diff -NurpP --minimal linux-2.6.36/fs/xfs/linux-2.6/xfs_linux.h linux-2.6.36-vs2.3.0.36.33/fs/xfs/linux-2.6/xfs_linux.h
7930 --- linux-2.6.36/fs/xfs/linux-2.6/xfs_linux.h 2010-10-21 13:07:51.000000000 +0200
7931 +++ linux-2.6.36-vs2.3.0.36.33/fs/xfs/linux-2.6/xfs_linux.h 2010-10-21 13:09:36.000000000 +0200
7934 #define current_cpu() (raw_smp_processor_id())
7935 #define current_pid() (current->pid)
7936 +#define current_fstag(cred,vp) (dx_current_fstag((vp)->i_sb))
7937 #define current_test_flags(f) (current->flags & (f))
7938 #define current_set_flags_nested(sp, f) \
7939 (*(sp) = current->flags, current->flags |= (f))
7940 diff -NurpP --minimal linux-2.6.36/fs/xfs/linux-2.6/xfs_super.c linux-2.6.36-vs2.3.0.36.33/fs/xfs/linux-2.6/xfs_super.c
7941 --- linux-2.6.36/fs/xfs/linux-2.6/xfs_super.c 2010-10-21 13:07:51.000000000 +0200
7942 +++ linux-2.6.36-vs2.3.0.36.33/fs/xfs/linux-2.6/xfs_super.c 2010-10-21 13:09:36.000000000 +0200
7943 @@ -113,6 +113,9 @@ mempool_t *xfs_ioend_pool;
7944 #define MNTOPT_QUOTANOENF "qnoenforce" /* same as uqnoenforce */
7945 #define MNTOPT_DELAYLOG "delaylog" /* Delayed loging enabled */
7946 #define MNTOPT_NODELAYLOG "nodelaylog" /* Delayed loging disabled */
7947 +#define MNTOPT_TAGXID "tagxid" /* context tagging for inodes */
7948 +#define MNTOPT_TAGGED "tag" /* context tagging for inodes */
7949 +#define MNTOPT_NOTAGTAG "notag" /* do not use context tagging */
7952 * Table driven mount option parser.
7953 @@ -121,10 +124,14 @@ mempool_t *xfs_ioend_pool;
7954 * in the future, too.
7957 + Opt_tag, Opt_notag,
7958 Opt_barrier, Opt_nobarrier, Opt_err
7961 static const match_table_t tokens = {
7962 + {Opt_tag, "tagxid"},
7964 + {Opt_notag, "notag"},
7965 {Opt_barrier, "barrier"},
7966 {Opt_nobarrier, "nobarrier"},
7968 @@ -371,6 +378,19 @@ xfs_parseargs(
7969 } else if (!strcmp(this_char, "irixsgid")) {
7971 "XFS: irixsgid is now a sysctl(2) variable, option is deprecated.");
7972 +#ifndef CONFIG_TAGGING_NONE
7973 + } else if (!strcmp(this_char, MNTOPT_TAGGED)) {
7974 + mp->m_flags |= XFS_MOUNT_TAGGED;
7975 + } else if (!strcmp(this_char, MNTOPT_NOTAGTAG)) {
7976 + mp->m_flags &= ~XFS_MOUNT_TAGGED;
7977 + } else if (!strcmp(this_char, MNTOPT_TAGXID)) {
7978 + mp->m_flags |= XFS_MOUNT_TAGGED;
7980 +#ifdef CONFIG_PROPAGATE
7981 + } else if (!strcmp(this_char, MNTOPT_TAGGED)) {
7983 + mp->m_flags |= XFS_MOUNT_TAGGED;
7987 "XFS: unknown mount option [%s].", this_char);
7988 @@ -1320,6 +1340,16 @@ xfs_fs_remount(
7990 mp->m_flags &= ~XFS_MOUNT_BARRIER;
7993 + if (!(sb->s_flags & MS_TAGGED)) {
7995 + "XFS: %s: tagging not permitted on remount.\n",
8004 * Logically we would return an error here to prevent
8005 @@ -1547,6 +1577,9 @@ xfs_fs_fill_super(
8007 goto out_filestream_unmount;
8009 + if (mp->m_flags & XFS_MOUNT_TAGGED)
8010 + sb->s_flags |= MS_TAGGED;
8012 sb->s_magic = XFS_SB_MAGIC;
8013 sb->s_blocksize = mp->m_sb.sb_blocksize;
8014 sb->s_blocksize_bits = ffs(sb->s_blocksize) - 1;
8015 diff -NurpP --minimal linux-2.6.36/fs/xfs/xfs_dinode.h linux-2.6.36-vs2.3.0.36.33/fs/xfs/xfs_dinode.h
8016 --- linux-2.6.36/fs/xfs/xfs_dinode.h 2009-06-11 17:13:09.000000000 +0200
8017 +++ linux-2.6.36-vs2.3.0.36.33/fs/xfs/xfs_dinode.h 2010-10-21 13:09:36.000000000 +0200
8018 @@ -50,7 +50,9 @@ typedef struct xfs_dinode {
8019 __be32 di_gid; /* owner's group id */
8020 __be32 di_nlink; /* number of links to file */
8021 __be16 di_projid; /* owner's project id */
8022 - __u8 di_pad[8]; /* unused, zeroed space */
8023 + __be16 di_tag; /* context tagging */
8024 + __be16 di_vflags; /* vserver specific flags */
8025 + __u8 di_pad[4]; /* unused, zeroed space */
8026 __be16 di_flushiter; /* incremented on flush */
8027 xfs_timestamp_t di_atime; /* time last accessed */
8028 xfs_timestamp_t di_mtime; /* time last modified */
8029 @@ -183,6 +185,8 @@ static inline void xfs_dinode_put_rdev(s
8030 #define XFS_DIFLAG_EXTSZINHERIT_BIT 12 /* inherit inode extent size */
8031 #define XFS_DIFLAG_NODEFRAG_BIT 13 /* do not reorganize/defragment */
8032 #define XFS_DIFLAG_FILESTREAM_BIT 14 /* use filestream allocator */
8033 +#define XFS_DIFLAG_IXUNLINK_BIT 15 /* Immutable inver on unlink */
8035 #define XFS_DIFLAG_REALTIME (1 << XFS_DIFLAG_REALTIME_BIT)
8036 #define XFS_DIFLAG_PREALLOC (1 << XFS_DIFLAG_PREALLOC_BIT)
8037 #define XFS_DIFLAG_NEWRTBM (1 << XFS_DIFLAG_NEWRTBM_BIT)
8038 @@ -198,6 +202,7 @@ static inline void xfs_dinode_put_rdev(s
8039 #define XFS_DIFLAG_EXTSZINHERIT (1 << XFS_DIFLAG_EXTSZINHERIT_BIT)
8040 #define XFS_DIFLAG_NODEFRAG (1 << XFS_DIFLAG_NODEFRAG_BIT)
8041 #define XFS_DIFLAG_FILESTREAM (1 << XFS_DIFLAG_FILESTREAM_BIT)
8042 +#define XFS_DIFLAG_IXUNLINK (1 << XFS_DIFLAG_IXUNLINK_BIT)
8044 #ifdef CONFIG_XFS_RT
8045 #define XFS_IS_REALTIME_INODE(ip) ((ip)->i_d.di_flags & XFS_DIFLAG_REALTIME)
8046 @@ -210,6 +215,10 @@ static inline void xfs_dinode_put_rdev(s
8047 XFS_DIFLAG_IMMUTABLE | XFS_DIFLAG_APPEND | XFS_DIFLAG_SYNC | \
8048 XFS_DIFLAG_NOATIME | XFS_DIFLAG_NODUMP | XFS_DIFLAG_RTINHERIT | \
8049 XFS_DIFLAG_PROJINHERIT | XFS_DIFLAG_NOSYMLINKS | XFS_DIFLAG_EXTSIZE | \
8050 - XFS_DIFLAG_EXTSZINHERIT | XFS_DIFLAG_NODEFRAG | XFS_DIFLAG_FILESTREAM)
8051 + XFS_DIFLAG_EXTSZINHERIT | XFS_DIFLAG_NODEFRAG | XFS_DIFLAG_FILESTREAM | \
8052 + XFS_DIFLAG_IXUNLINK)
8054 +#define XFS_DIVFLAG_BARRIER 0x01
8055 +#define XFS_DIVFLAG_COW 0x02
8057 #endif /* __XFS_DINODE_H__ */
8058 diff -NurpP --minimal linux-2.6.36/fs/xfs/xfs_fs.h linux-2.6.36-vs2.3.0.36.33/fs/xfs/xfs_fs.h
8059 --- linux-2.6.36/fs/xfs/xfs_fs.h 2010-10-21 13:07:52.000000000 +0200
8060 +++ linux-2.6.36-vs2.3.0.36.33/fs/xfs/xfs_fs.h 2010-10-21 13:09:36.000000000 +0200
8061 @@ -67,6 +67,9 @@ struct fsxattr {
8062 #define XFS_XFLAG_EXTSZINHERIT 0x00001000 /* inherit inode extent size */
8063 #define XFS_XFLAG_NODEFRAG 0x00002000 /* do not defragment */
8064 #define XFS_XFLAG_FILESTREAM 0x00004000 /* use filestream allocator */
8065 +#define XFS_XFLAG_IXUNLINK 0x00008000 /* immutable invert on unlink */
8066 +#define XFS_XFLAG_BARRIER 0x10000000 /* chroot() barrier */
8067 +#define XFS_XFLAG_COW 0x20000000 /* copy on write mark */
8068 #define XFS_XFLAG_HASATTR 0x80000000 /* no DIFLAG for this */
8071 @@ -295,7 +298,8 @@ typedef struct xfs_bstat {
8072 __u32 bs_gen; /* generation count */
8073 __u16 bs_projid; /* project id */
8074 __u16 bs_forkoff; /* inode fork offset in bytes */
8075 - unsigned char bs_pad[12]; /* pad space, unused */
8076 + __u16 bs_tag; /* context tagging */
8077 + unsigned char bs_pad[10]; /* pad space, unused */
8078 __u32 bs_dmevmask; /* DMIG event mask */
8079 __u16 bs_dmstate; /* DMIG state info */
8080 __u16 bs_aextents; /* attribute number of extents */
8081 diff -NurpP --minimal linux-2.6.36/fs/xfs/xfs_ialloc.c linux-2.6.36-vs2.3.0.36.33/fs/xfs/xfs_ialloc.c
8082 --- linux-2.6.36/fs/xfs/xfs_ialloc.c 2010-10-21 13:07:52.000000000 +0200
8083 +++ linux-2.6.36-vs2.3.0.36.33/fs/xfs/xfs_ialloc.c 2010-10-21 13:09:36.000000000 +0200
8085 #include "xfs_error.h"
8086 #include "xfs_bmap.h"
8090 * Allocation group level functions.
8092 diff -NurpP --minimal linux-2.6.36/fs/xfs/xfs_inode.c linux-2.6.36-vs2.3.0.36.33/fs/xfs/xfs_inode.c
8093 --- linux-2.6.36/fs/xfs/xfs_inode.c 2010-10-21 13:07:52.000000000 +0200
8094 +++ linux-2.6.36-vs2.3.0.36.33/fs/xfs/xfs_inode.c 2010-10-21 13:09:36.000000000 +0200
8095 @@ -245,6 +245,7 @@ xfs_inotobp(
8099 +#include <linux/vs_tag.h>
8102 * This routine is called to map an inode to the buffer containing
8103 @@ -650,15 +651,25 @@ xfs_iformat_btree(
8105 xfs_dinode_from_disk(
8107 - xfs_dinode_t *from)
8108 + xfs_dinode_t *from,
8111 + uint32_t uid, gid, tag;
8113 to->di_magic = be16_to_cpu(from->di_magic);
8114 to->di_mode = be16_to_cpu(from->di_mode);
8115 to->di_version = from ->di_version;
8116 to->di_format = from->di_format;
8117 to->di_onlink = be16_to_cpu(from->di_onlink);
8118 - to->di_uid = be32_to_cpu(from->di_uid);
8119 - to->di_gid = be32_to_cpu(from->di_gid);
8121 + uid = be32_to_cpu(from->di_uid);
8122 + gid = be32_to_cpu(from->di_gid);
8123 + tag = be16_to_cpu(from->di_tag);
8125 + to->di_uid = INOTAG_UID(tagged, uid, gid);
8126 + to->di_gid = INOTAG_GID(tagged, uid, gid);
8127 + to->di_tag = INOTAG_TAG(tagged, uid, gid, tag);
8129 to->di_nlink = be32_to_cpu(from->di_nlink);
8130 to->di_projid = be16_to_cpu(from->di_projid);
8131 memcpy(to->di_pad, from->di_pad, sizeof(to->di_pad));
8132 @@ -679,21 +690,26 @@ xfs_dinode_from_disk(
8133 to->di_dmevmask = be32_to_cpu(from->di_dmevmask);
8134 to->di_dmstate = be16_to_cpu(from->di_dmstate);
8135 to->di_flags = be16_to_cpu(from->di_flags);
8136 + to->di_vflags = be16_to_cpu(from->di_vflags);
8137 to->di_gen = be32_to_cpu(from->di_gen);
8143 - xfs_icdinode_t *from)
8144 + xfs_icdinode_t *from,
8147 to->di_magic = cpu_to_be16(from->di_magic);
8148 to->di_mode = cpu_to_be16(from->di_mode);
8149 to->di_version = from ->di_version;
8150 to->di_format = from->di_format;
8151 to->di_onlink = cpu_to_be16(from->di_onlink);
8152 - to->di_uid = cpu_to_be32(from->di_uid);
8153 - to->di_gid = cpu_to_be32(from->di_gid);
8155 + to->di_uid = cpu_to_be32(TAGINO_UID(tagged, from->di_uid, from->di_tag));
8156 + to->di_gid = cpu_to_be32(TAGINO_GID(tagged, from->di_gid, from->di_tag));
8157 + to->di_tag = cpu_to_be16(TAGINO_TAG(tagged, from->di_tag));
8159 to->di_nlink = cpu_to_be32(from->di_nlink);
8160 to->di_projid = cpu_to_be16(from->di_projid);
8161 memcpy(to->di_pad, from->di_pad, sizeof(to->di_pad));
8162 @@ -714,12 +730,14 @@ xfs_dinode_to_disk(
8163 to->di_dmevmask = cpu_to_be32(from->di_dmevmask);
8164 to->di_dmstate = cpu_to_be16(from->di_dmstate);
8165 to->di_flags = cpu_to_be16(from->di_flags);
8166 + to->di_vflags = cpu_to_be16(from->di_vflags);
8167 to->di_gen = cpu_to_be32(from->di_gen);
8172 - __uint16_t di_flags)
8173 + __uint16_t di_flags,
8174 + __uint16_t di_vflags)
8178 @@ -730,6 +748,8 @@ _xfs_dic2xflags(
8179 flags |= XFS_XFLAG_PREALLOC;
8180 if (di_flags & XFS_DIFLAG_IMMUTABLE)
8181 flags |= XFS_XFLAG_IMMUTABLE;
8182 + if (di_flags & XFS_DIFLAG_IXUNLINK)
8183 + flags |= XFS_XFLAG_IXUNLINK;
8184 if (di_flags & XFS_DIFLAG_APPEND)
8185 flags |= XFS_XFLAG_APPEND;
8186 if (di_flags & XFS_DIFLAG_SYNC)
8187 @@ -754,6 +774,10 @@ _xfs_dic2xflags(
8188 flags |= XFS_XFLAG_FILESTREAM;
8191 + if (di_vflags & XFS_DIVFLAG_BARRIER)
8192 + flags |= FS_BARRIER_FL;
8193 + if (di_vflags & XFS_DIVFLAG_COW)
8194 + flags |= FS_COW_FL;
8198 @@ -763,7 +787,7 @@ xfs_ip2xflags(
8200 xfs_icdinode_t *dic = &ip->i_d;
8202 - return _xfs_dic2xflags(dic->di_flags) |
8203 + return _xfs_dic2xflags(dic->di_flags, dic->di_vflags) |
8204 (XFS_IFORK_Q(ip) ? XFS_XFLAG_HASATTR : 0);
8207 @@ -771,7 +795,8 @@ uint
8211 - return _xfs_dic2xflags(be16_to_cpu(dip->di_flags)) |
8212 + return _xfs_dic2xflags(be16_to_cpu(dip->di_flags),
8213 + be16_to_cpu(dip->di_vflags)) |
8214 (XFS_DFORK_Q(dip) ? XFS_XFLAG_HASATTR : 0);
8217 @@ -804,7 +829,6 @@ xfs_iread(
8220 dip = (xfs_dinode_t *)xfs_buf_offset(bp, ip->i_imap.im_boffset);
8223 * If we got something that isn't an inode it means someone
8224 * (nfs or dmi) has a stale handle.
8225 @@ -829,7 +853,8 @@ xfs_iread(
8226 * Otherwise, just get the truly permanent information.
8229 - xfs_dinode_from_disk(&ip->i_d, dip);
8230 + xfs_dinode_from_disk(&ip->i_d, dip,
8231 + mp->m_flags & XFS_MOUNT_TAGGED);
8232 error = xfs_iformat(ip, dip);
8235 @@ -1027,6 +1052,7 @@ xfs_ialloc(
8236 ASSERT(ip->i_d.di_nlink == nlink);
8237 ip->i_d.di_uid = current_fsuid();
8238 ip->i_d.di_gid = current_fsgid();
8239 + ip->i_d.di_tag = current_fstag(cr, &ip->i_vnode);
8240 ip->i_d.di_projid = prid;
8241 memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad));
8243 @@ -1087,6 +1113,7 @@ xfs_ialloc(
8244 ip->i_d.di_dmevmask = 0;
8245 ip->i_d.di_dmstate = 0;
8246 ip->i_d.di_flags = 0;
8247 + ip->i_d.di_vflags = 0;
8248 flags = XFS_ILOG_CORE;
8249 switch (mode & S_IFMT) {
8251 @@ -2105,6 +2132,7 @@ xfs_ifree(
8253 ip->i_d.di_mode = 0; /* mark incore inode as free */
8254 ip->i_d.di_flags = 0;
8255 + ip->i_d.di_vflags = 0;
8256 ip->i_d.di_dmevmask = 0;
8257 ip->i_d.di_forkoff = 0; /* mark the attr fork not in use */
8258 ip->i_df.if_ext_max =
8259 @@ -2975,7 +3003,8 @@ xfs_iflush_int(
8260 * because if the inode is dirty at all the core must
8263 - xfs_dinode_to_disk(dip, &ip->i_d);
8264 + xfs_dinode_to_disk(dip, &ip->i_d,
8265 + mp->m_flags & XFS_MOUNT_TAGGED);
8267 /* Wrap, we never let the log put out DI_MAX_FLUSH */
8268 if (ip->i_d.di_flushiter == DI_MAX_FLUSH)
8269 diff -NurpP --minimal linux-2.6.36/fs/xfs/xfs_inode.h linux-2.6.36-vs2.3.0.36.33/fs/xfs/xfs_inode.h
8270 --- linux-2.6.36/fs/xfs/xfs_inode.h 2010-10-21 13:07:52.000000000 +0200
8271 +++ linux-2.6.36-vs2.3.0.36.33/fs/xfs/xfs_inode.h 2010-10-21 13:09:36.000000000 +0200
8272 @@ -135,7 +135,9 @@ typedef struct xfs_icdinode {
8273 __uint32_t di_gid; /* owner's group id */
8274 __uint32_t di_nlink; /* number of links to file */
8275 __uint16_t di_projid; /* owner's project id */
8276 - __uint8_t di_pad[8]; /* unused, zeroed space */
8277 + __uint16_t di_tag; /* context tagging */
8278 + __uint16_t di_vflags; /* vserver specific flags */
8279 + __uint8_t di_pad[4]; /* unused, zeroed space */
8280 __uint16_t di_flushiter; /* incremented on flush */
8281 xfs_ictimestamp_t di_atime; /* time last accessed */
8282 xfs_ictimestamp_t di_mtime; /* time last modified */
8283 @@ -509,7 +511,7 @@ int xfs_itobp(struct xfs_mount *, struc
8284 int xfs_iread(struct xfs_mount *, struct xfs_trans *,
8285 struct xfs_inode *, uint);
8286 void xfs_dinode_to_disk(struct xfs_dinode *,
8287 - struct xfs_icdinode *);
8288 + struct xfs_icdinode *, int);
8289 void xfs_idestroy_fork(struct xfs_inode *, int);
8290 void xfs_idata_realloc(struct xfs_inode *, int, int);
8291 void xfs_iroot_realloc(struct xfs_inode *, int, int);
8292 diff -NurpP --minimal linux-2.6.36/fs/xfs/xfs_itable.c linux-2.6.36-vs2.3.0.36.33/fs/xfs/xfs_itable.c
8293 --- linux-2.6.36/fs/xfs/xfs_itable.c 2010-10-21 13:07:52.000000000 +0200
8294 +++ linux-2.6.36-vs2.3.0.36.33/fs/xfs/xfs_itable.c 2010-10-21 13:09:36.000000000 +0200
8295 @@ -97,6 +97,7 @@ xfs_bulkstat_one_int(
8296 buf->bs_mode = dic->di_mode;
8297 buf->bs_uid = dic->di_uid;
8298 buf->bs_gid = dic->di_gid;
8299 + buf->bs_tag = dic->di_tag;
8300 buf->bs_size = dic->di_size;
8303 diff -NurpP --minimal linux-2.6.36/fs/xfs/xfs_log_recover.c linux-2.6.36-vs2.3.0.36.33/fs/xfs/xfs_log_recover.c
8304 --- linux-2.6.36/fs/xfs/xfs_log_recover.c 2010-10-21 13:07:52.000000000 +0200
8305 +++ linux-2.6.36-vs2.3.0.36.33/fs/xfs/xfs_log_recover.c 2010-10-21 14:02:02.000000000 +0200
8306 @@ -2450,7 +2450,8 @@ xlog_recover_do_inode_trans(
8309 /* The core is in in-core format */
8310 - xfs_dinode_to_disk(dip, item->ri_buf[1].i_addr);
8311 + xfs_dinode_to_disk(dip, item->ri_buf[1].i_addr,
8312 + mp->m_flags & XFS_MOUNT_TAGGED);
8314 /* the rest is in on-disk format */
8315 if (item->ri_buf[1].i_len > sizeof(struct xfs_icdinode)) {
8316 diff -NurpP --minimal linux-2.6.36/fs/xfs/xfs_mount.h linux-2.6.36-vs2.3.0.36.33/fs/xfs/xfs_mount.h
8317 --- linux-2.6.36/fs/xfs/xfs_mount.h 2010-10-21 13:07:52.000000000 +0200
8318 +++ linux-2.6.36-vs2.3.0.36.33/fs/xfs/xfs_mount.h 2010-10-21 13:09:36.000000000 +0200
8319 @@ -238,6 +238,7 @@ typedef struct xfs_mount {
8321 #define XFS_MOUNT_NOATTR2 (1ULL << 25) /* disable use of attr2 format */
8323 +#define XFS_MOUNT_TAGGED (1ULL << 31) /* context tagging */
8326 * Default minimum read and write sizes.
8327 diff -NurpP --minimal linux-2.6.36/fs/xfs/xfs_vnodeops.c linux-2.6.36-vs2.3.0.36.33/fs/xfs/xfs_vnodeops.c
8328 --- linux-2.6.36/fs/xfs/xfs_vnodeops.c 2010-10-21 13:07:52.000000000 +0200
8329 +++ linux-2.6.36-vs2.3.0.36.33/fs/xfs/xfs_vnodeops.c 2010-10-21 15:17:26.000000000 +0200
8331 #include "xfs_vnodeops.h"
8332 #include "xfs_trace.h"
8336 +xfs_get_inode_flags(
8339 + struct inode *inode = VFS_I(ip);
8340 + unsigned int flags = inode->i_flags;
8341 + unsigned int vflags = inode->i_vflags;
8343 + if (flags & S_IMMUTABLE)
8344 + ip->i_d.di_flags |= XFS_DIFLAG_IMMUTABLE;
8346 + ip->i_d.di_flags &= ~XFS_DIFLAG_IMMUTABLE;
8347 + if (flags & S_IXUNLINK)
8348 + ip->i_d.di_flags |= XFS_DIFLAG_IXUNLINK;
8350 + ip->i_d.di_flags &= ~XFS_DIFLAG_IXUNLINK;
8352 + if (vflags & V_BARRIER)
8353 + ip->i_d.di_vflags |= XFS_DIVFLAG_BARRIER;
8355 + ip->i_d.di_vflags &= ~XFS_DIVFLAG_BARRIER;
8356 + if (vflags & V_COW)
8357 + ip->i_d.di_vflags |= XFS_DIVFLAG_COW;
8359 + ip->i_d.di_vflags &= ~XFS_DIVFLAG_COW;
8364 + struct inode *inode,
8368 + struct xfs_inode *ip = XFS_I(inode);
8369 + struct xfs_mount *mp = ip->i_mount;
8370 + struct xfs_trans *tp;
8371 + unsigned int lock_flags = 0;
8374 + tp = xfs_trans_alloc(mp, XFS_TRANS_SETATTR_NOT_SIZE);
8375 + code = xfs_trans_reserve(tp, 0, XFS_ICHANGE_LOG_RES(mp), 0, 0, 0);
8379 + xfs_ilock(ip, XFS_ILOCK_EXCL);
8381 + xfs_trans_ijoin(tp, ip);
8383 + inode->i_flags = flags;
8384 + inode->i_vflags = vflags;
8385 + xfs_get_inode_flags(ip);
8387 + xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
8388 + xfs_ichgtime(ip, XFS_ICHGTIME_CHG);
8390 + XFS_STATS_INC(xs_ig_attrchg);
8392 + if (mp->m_flags & XFS_MOUNT_WSYNC)
8393 + xfs_trans_set_sync(tp);
8394 + code = xfs_trans_commit(tp, 0);
8395 + xfs_iunlock(ip, XFS_ILOCK_EXCL);
8399 + xfs_trans_cancel(tp, 0);
8401 + xfs_iunlock(ip, XFS_ILOCK_EXCL);
8408 struct xfs_inode *ip,
8409 @@ -65,6 +137,7 @@ xfs_setattr(
8410 uint commit_flags=0;
8411 uid_t uid=0, iuid=0;
8412 gid_t gid=0, igid=0;
8413 + tag_t tag=0, itag=0;
8414 struct xfs_dquot *udqp, *gdqp, *olddquot1, *olddquot2;
8415 int need_iolock = 1;
8417 @@ -147,7 +220,7 @@ xfs_setattr(
8419 * Change file ownership. Must be the owner or privileged.
8421 - if (mask & (ATTR_UID|ATTR_GID)) {
8422 + if (mask & (ATTR_UID|ATTR_GID|ATTR_TAG)) {
8424 * These IDs could have changed since we last looked at them.
8425 * But, we're assured that if the ownership did change
8426 @@ -156,8 +229,10 @@ xfs_setattr(
8428 iuid = ip->i_d.di_uid;
8429 igid = ip->i_d.di_gid;
8430 + itag = ip->i_d.di_tag;
8431 gid = (mask & ATTR_GID) ? iattr->ia_gid : igid;
8432 uid = (mask & ATTR_UID) ? iattr->ia_uid : iuid;
8433 + tag = (mask & ATTR_TAG) ? iattr->ia_tag : itag;
8436 * Do a quota reservation only if uid/gid is actually
8437 @@ -165,7 +240,8 @@ xfs_setattr(
8439 if (XFS_IS_QUOTA_RUNNING(mp) &&
8440 ((XFS_IS_UQUOTA_ON(mp) && iuid != uid) ||
8441 - (XFS_IS_GQUOTA_ON(mp) && igid != gid))) {
8442 + (XFS_IS_GQUOTA_ON(mp) && igid != gid) ||
8443 + (XFS_IS_GQUOTA_ON(mp) && itag != tag))) {
8445 code = xfs_qm_vop_chown_reserve(tp, ip, udqp, gdqp,
8446 capable(CAP_FOWNER) ?
8447 @@ -326,7 +402,7 @@ xfs_setattr(
8449 * Change file ownership. Must be the owner or privileged.
8451 - if (mask & (ATTR_UID|ATTR_GID)) {
8452 + if (mask & (ATTR_UID|ATTR_GID|ATTR_TAG)) {
8454 * CAP_FSETID overrides the following restrictions:
8456 @@ -342,6 +418,10 @@ xfs_setattr(
8457 * Change the ownerships and register quota modifications
8458 * in the transaction.
8460 + if (itag != tag) {
8461 + ip->i_d.di_tag = tag;
8462 + inode->i_tag = tag;
8465 if (XFS_IS_QUOTA_RUNNING(mp) && XFS_IS_UQUOTA_ON(mp)) {
8466 ASSERT(mask & ATTR_UID);
8467 diff -NurpP --minimal linux-2.6.36/fs/xfs/xfs_vnodeops.h linux-2.6.36-vs2.3.0.36.33/fs/xfs/xfs_vnodeops.h
8468 --- linux-2.6.36/fs/xfs/xfs_vnodeops.h 2010-07-07 18:31:54.000000000 +0200
8469 +++ linux-2.6.36-vs2.3.0.36.33/fs/xfs/xfs_vnodeops.h 2010-10-21 13:09:36.000000000 +0200
8470 @@ -14,6 +14,7 @@ struct xfs_inode;
8474 +int xfs_sync_xflags(struct xfs_inode *ip);
8475 int xfs_setattr(struct xfs_inode *ip, struct iattr *vap, int flags);
8476 #define XFS_ATTR_DMI 0x01 /* invocation from a DMI function */
8477 #define XFS_ATTR_NONBLOCK 0x02 /* return EAGAIN if operation would block */
8478 diff -NurpP --minimal linux-2.6.36/include/asm-generic/tlb.h linux-2.6.36-vs2.3.0.36.33/include/asm-generic/tlb.h
8479 --- linux-2.6.36/include/asm-generic/tlb.h 2009-09-10 15:26:24.000000000 +0200
8480 +++ linux-2.6.36-vs2.3.0.36.33/include/asm-generic/tlb.h 2010-10-21 13:09:36.000000000 +0200
8482 #define _ASM_GENERIC__TLB_H
8484 #include <linux/swap.h>
8485 +#include <linux/vs_memory.h>
8486 #include <asm/pgalloc.h>
8487 #include <asm/tlbflush.h>
8489 diff -NurpP --minimal linux-2.6.36/include/linux/capability.h linux-2.6.36-vs2.3.0.36.33/include/linux/capability.h
8490 --- linux-2.6.36/include/linux/capability.h 2010-10-21 13:07:52.000000000 +0200
8491 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/capability.h 2010-10-21 13:09:36.000000000 +0200
8492 @@ -280,6 +280,7 @@ struct cpu_vfs_cap_data {
8493 arbitrary SCSI commands */
8494 /* Allow setting encryption key on loopback filesystem */
8495 /* Allow setting zone reclaim policy */
8496 +/* Allow the selection of a security context */
8498 #define CAP_SYS_ADMIN 21
8500 @@ -352,7 +353,13 @@ struct cpu_vfs_cap_data {
8502 #define CAP_MAC_ADMIN 33
8504 -#define CAP_LAST_CAP CAP_MAC_ADMIN
8505 +/* Allow context manipulations */
8506 +/* Allow changing context info on files */
8508 +#define CAP_CONTEXT 34
8511 +#define CAP_LAST_CAP CAP_CONTEXT
8513 #define cap_valid(x) ((x) >= 0 && (x) <= CAP_LAST_CAP)
8515 diff -NurpP --minimal linux-2.6.36/include/linux/devpts_fs.h linux-2.6.36-vs2.3.0.36.33/include/linux/devpts_fs.h
8516 --- linux-2.6.36/include/linux/devpts_fs.h 2008-12-25 00:26:37.000000000 +0100
8517 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/devpts_fs.h 2010-10-21 13:09:36.000000000 +0200
8518 @@ -45,5 +45,4 @@ static inline void devpts_pty_kill(struc
8523 #endif /* _LINUX_DEVPTS_FS_H */
8524 diff -NurpP --minimal linux-2.6.36/include/linux/ext2_fs.h linux-2.6.36-vs2.3.0.36.33/include/linux/ext2_fs.h
8525 --- linux-2.6.36/include/linux/ext2_fs.h 2010-02-25 11:52:07.000000000 +0100
8526 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/ext2_fs.h 2010-10-21 13:09:36.000000000 +0200
8527 @@ -189,8 +189,12 @@ struct ext2_group_desc
8528 #define EXT2_NOTAIL_FL FS_NOTAIL_FL /* file tail should not be merged */
8529 #define EXT2_DIRSYNC_FL FS_DIRSYNC_FL /* dirsync behaviour (directories only) */
8530 #define EXT2_TOPDIR_FL FS_TOPDIR_FL /* Top of directory hierarchies*/
8531 +#define EXT2_IXUNLINK_FL FS_IXUNLINK_FL /* Immutable invert on unlink */
8532 #define EXT2_RESERVED_FL FS_RESERVED_FL /* reserved for ext2 lib */
8534 +#define EXT2_BARRIER_FL FS_BARRIER_FL /* Barrier for chroot() */
8535 +#define EXT2_COW_FL FS_COW_FL /* Copy on Write marker */
8537 #define EXT2_FL_USER_VISIBLE FS_FL_USER_VISIBLE /* User visible flags */
8538 #define EXT2_FL_USER_MODIFIABLE FS_FL_USER_MODIFIABLE /* User modifiable flags */
8540 @@ -274,7 +278,8 @@ struct ext2_inode {
8542 __le16 l_i_uid_high; /* these 2 fields */
8543 __le16 l_i_gid_high; /* were reserved2[0] */
8544 - __u32 l_i_reserved2;
8545 + __le16 l_i_tag; /* Context Tag */
8546 + __u16 l_i_reserved2;
8549 __u8 h_i_frag; /* Fragment number */
8550 @@ -303,6 +308,7 @@ struct ext2_inode {
8551 #define i_gid_low i_gid
8552 #define i_uid_high osd2.linux2.l_i_uid_high
8553 #define i_gid_high osd2.linux2.l_i_gid_high
8554 +#define i_raw_tag osd2.linux2.l_i_tag
8555 #define i_reserved2 osd2.linux2.l_i_reserved2
8558 @@ -347,6 +353,7 @@ struct ext2_inode {
8559 #define EXT2_MOUNT_USRQUOTA 0x020000 /* user quota */
8560 #define EXT2_MOUNT_GRPQUOTA 0x040000 /* group quota */
8561 #define EXT2_MOUNT_RESERVATION 0x080000 /* Preallocation */
8562 +#define EXT2_MOUNT_TAGGED (1<<24) /* Enable Context Tags */
8565 #define clear_opt(o, opt) o &= ~EXT2_MOUNT_##opt
8566 diff -NurpP --minimal linux-2.6.36/include/linux/ext3_fs.h linux-2.6.36-vs2.3.0.36.33/include/linux/ext3_fs.h
8567 --- linux-2.6.36/include/linux/ext3_fs.h 2010-10-21 13:07:53.000000000 +0200
8568 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/ext3_fs.h 2010-10-21 13:09:36.000000000 +0200
8569 @@ -173,10 +173,14 @@ struct ext3_group_desc
8570 #define EXT3_NOTAIL_FL 0x00008000 /* file tail should not be merged */
8571 #define EXT3_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
8572 #define EXT3_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/
8573 +#define EXT3_IXUNLINK_FL 0x08000000 /* Immutable invert on unlink */
8574 #define EXT3_RESERVED_FL 0x80000000 /* reserved for ext3 lib */
8576 -#define EXT3_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */
8577 -#define EXT3_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */
8578 +#define EXT3_BARRIER_FL 0x04000000 /* Barrier for chroot() */
8579 +#define EXT3_COW_FL 0x20000000 /* Copy on Write marker */
8581 +#define EXT3_FL_USER_VISIBLE 0x0103DFFF /* User visible flags */
8582 +#define EXT3_FL_USER_MODIFIABLE 0x010380FF /* User modifiable flags */
8584 /* Flags that should be inherited by new inodes from their parent. */
8585 #define EXT3_FL_INHERITED (EXT3_SECRM_FL | EXT3_UNRM_FL | EXT3_COMPR_FL |\
8586 @@ -312,7 +316,8 @@ struct ext3_inode {
8588 __le16 l_i_uid_high; /* these 2 fields */
8589 __le16 l_i_gid_high; /* were reserved2[0] */
8590 - __u32 l_i_reserved2;
8591 + __le16 l_i_tag; /* Context Tag */
8592 + __u16 l_i_reserved2;
8595 __u8 h_i_frag; /* Fragment number */
8596 @@ -343,6 +348,7 @@ struct ext3_inode {
8597 #define i_gid_low i_gid
8598 #define i_uid_high osd2.linux2.l_i_uid_high
8599 #define i_gid_high osd2.linux2.l_i_gid_high
8600 +#define i_raw_tag osd2.linux2.l_i_tag
8601 #define i_reserved2 osd2.linux2.l_i_reserved2
8603 #elif defined(__GNU__)
8604 @@ -405,6 +411,7 @@ struct ext3_inode {
8605 #define EXT3_MOUNT_GRPQUOTA 0x200000 /* "old" group quota */
8606 #define EXT3_MOUNT_DATA_ERR_ABORT 0x400000 /* Abort on file data write
8607 * error in ordered mode */
8608 +#define EXT3_MOUNT_TAGGED (1<<24) /* Enable Context Tags */
8610 /* Compatibility, for having both ext2_fs.h and ext3_fs.h included at once */
8611 #ifndef _LINUX_EXT2_FS_H
8612 @@ -908,6 +915,7 @@ extern void ext3_get_inode_flags(struct
8613 extern void ext3_set_aops(struct inode *inode);
8614 extern int ext3_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
8615 u64 start, u64 len);
8616 +extern int ext3_sync_flags(struct inode *, int, int);
8619 extern long ext3_ioctl(struct file *, unsigned int, unsigned long);
8620 diff -NurpP --minimal linux-2.6.36/include/linux/fs.h linux-2.6.36-vs2.3.0.36.33/include/linux/fs.h
8621 --- linux-2.6.36/include/linux/fs.h 2010-10-21 13:07:53.000000000 +0200
8622 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/fs.h 2010-10-21 13:09:36.000000000 +0200
8623 @@ -208,6 +208,9 @@ struct inodes_stat_t {
8624 #define MS_KERNMOUNT (1<<22) /* this is a kern_mount call */
8625 #define MS_I_VERSION (1<<23) /* Update inode I_version field */
8626 #define MS_STRICTATIME (1<<24) /* Always perform atime updates */
8627 +#define MS_TAGGED (1<<25) /* use generic inode tagging */
8628 +#define MS_TAGID (1<<26) /* use specific tag for this mount */
8629 +#define MS_NOTAGCHECK (1<<27) /* don't check tags */
8630 #define MS_BORN (1<<29)
8631 #define MS_ACTIVE (1<<30)
8632 #define MS_NOUSER (1<<31)
8633 @@ -235,6 +238,14 @@ struct inodes_stat_t {
8634 #define S_NOCMTIME 128 /* Do not update file c/mtime */
8635 #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
8636 #define S_PRIVATE 512 /* Inode is fs-internal */
8637 +#define S_IXUNLINK 1024 /* Immutable Invert on unlink */
8639 +/* Linux-VServer related Inode flags */
8643 +#define V_BARRIER 4 /* Barrier for chroot() */
8644 +#define V_COW 8 /* Copy on Write */
8647 * Note that nosuid etc flags are inode-specific: setting some file-system
8648 @@ -257,12 +268,15 @@ struct inodes_stat_t {
8649 #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
8650 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
8651 #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
8652 -#define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
8653 -#define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
8654 +#define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
8655 +#define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
8656 +#define IS_TAGGED(inode) __IS_FLG(inode, MS_TAGGED)
8658 #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
8659 #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
8660 #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
8661 +#define IS_IXUNLINK(inode) ((inode)->i_flags & S_IXUNLINK)
8662 +#define IS_IXORUNLINK(inode) ((IS_IXUNLINK(inode) ? S_IMMUTABLE : 0) ^ IS_IMMUTABLE(inode))
8663 #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
8665 #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
8666 @@ -270,6 +284,16 @@ struct inodes_stat_t {
8667 #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
8668 #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
8670 +#define IS_BARRIER(inode) (S_ISDIR((inode)->i_mode) && ((inode)->i_vflags & V_BARRIER))
8672 +#ifdef CONFIG_VSERVER_COWBL
8673 +# define IS_COW(inode) (IS_IXUNLINK(inode) && IS_IMMUTABLE(inode))
8674 +# define IS_COW_LINK(inode) (S_ISREG((inode)->i_mode) && ((inode)->i_nlink > 1))
8676 +# define IS_COW(inode) (0)
8677 +# define IS_COW_LINK(inode) (0)
8680 /* the read-only stuff doesn't really belong here, but any other place is
8681 probably as bad and I don't want to create yet another include file. */
8683 @@ -353,11 +377,14 @@ struct inodes_stat_t {
8684 #define FS_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/
8685 #define FS_EXTENT_FL 0x00080000 /* Extents */
8686 #define FS_DIRECTIO_FL 0x00100000 /* Use direct i/o */
8687 +#define FS_IXUNLINK_FL 0x08000000 /* Immutable invert on unlink */
8688 #define FS_RESERVED_FL 0x80000000 /* reserved for ext2 lib */
8690 -#define FS_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */
8691 -#define FS_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */
8692 +#define FS_BARRIER_FL 0x04000000 /* Barrier for chroot() */
8693 +#define FS_COW_FL 0x20000000 /* Copy on Write marker */
8695 +#define FS_FL_USER_VISIBLE 0x0103DFFF /* User visible flags */
8696 +#define FS_FL_USER_MODIFIABLE 0x010380FF /* User modifiable flags */
8698 #define SYNC_FILE_RANGE_WAIT_BEFORE 1
8699 #define SYNC_FILE_RANGE_WRITE 2
8700 @@ -437,6 +464,7 @@ typedef void (dio_iodone_t)(struct kiocb
8701 #define ATTR_KILL_PRIV (1 << 14)
8702 #define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
8703 #define ATTR_TIMES_SET (1 << 16)
8704 +#define ATTR_TAG (1 << 17)
8707 * This is the Inode Attributes structure, used for notify_change(). It
8708 @@ -452,6 +480,7 @@ struct iattr {
8714 struct timespec ia_atime;
8715 struct timespec ia_mtime;
8716 @@ -465,6 +494,9 @@ struct iattr {
8717 struct file *ia_file;
8720 +#define ATTR_FLAG_BARRIER 512 /* Barrier for chroot() */
8721 +#define ATTR_FLAG_IXUNLINK 1024 /* Immutable invert on unlink */
8724 * Includes for diskquotas.
8726 @@ -732,7 +764,9 @@ struct inode {
8727 unsigned int i_nlink;
8733 unsigned int i_blkbits;
8736 @@ -774,7 +808,8 @@ struct inode {
8737 unsigned long i_state;
8738 unsigned long dirtied_when; /* jiffies of first dirtying */
8740 - unsigned int i_flags;
8741 + unsigned short i_flags;
8742 + unsigned short i_vflags;
8744 atomic_t i_writecount;
8745 #ifdef CONFIG_SECURITY
8746 @@ -862,12 +897,12 @@ static inline void i_size_write(struct i
8748 static inline unsigned iminor(const struct inode *inode)
8750 - return MINOR(inode->i_rdev);
8751 + return MINOR(inode->i_mdev);
8754 static inline unsigned imajor(const struct inode *inode)
8756 - return MAJOR(inode->i_rdev);
8757 + return MAJOR(inode->i_mdev);
8760 extern struct block_device *I_BDEV(struct inode *inode);
8761 @@ -929,6 +964,7 @@ struct file {
8763 struct fown_struct f_owner;
8764 const struct cred *f_cred;
8766 struct file_ra_state f_ra;
8769 @@ -1068,6 +1104,7 @@ struct file_lock {
8770 struct file *fl_file;
8775 struct fasync_struct * fl_fasync; /* for lease break notifications */
8776 unsigned long fl_break_time; /* for nonblocking lease breaks */
8777 @@ -1536,6 +1573,7 @@ struct inode_operations {
8778 ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
8779 ssize_t (*listxattr) (struct dentry *, char *, size_t);
8780 int (*removexattr) (struct dentry *, const char *);
8781 + int (*sync_flags) (struct inode *, int, int);
8782 void (*truncate_range)(struct inode *, loff_t, loff_t);
8783 long (*fallocate)(struct inode *inode, int mode, loff_t offset,
8785 @@ -1556,6 +1594,7 @@ extern ssize_t vfs_readv(struct file *,
8786 unsigned long, loff_t *);
8787 extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
8788 unsigned long, loff_t *);
8789 +ssize_t vfs_sendfile(struct file *, struct file *, loff_t *, size_t, loff_t);
8791 struct super_operations {
8792 struct inode *(*alloc_inode)(struct super_block *sb);
8793 @@ -2344,6 +2383,7 @@ extern int dcache_dir_open(struct inode
8794 extern int dcache_dir_close(struct inode *, struct file *);
8795 extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
8796 extern int dcache_readdir(struct file *, void *, filldir_t);
8797 +extern int dcache_readdir_filter(struct file *, void *, filldir_t, int (*)(struct dentry *));
8798 extern int simple_setattr(struct dentry *, struct iattr *);
8799 extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
8800 extern int simple_statfs(struct dentry *, struct kstatfs *);
8801 diff -NurpP --minimal linux-2.6.36/include/linux/gfs2_ondisk.h linux-2.6.36-vs2.3.0.36.33/include/linux/gfs2_ondisk.h
8802 --- linux-2.6.36/include/linux/gfs2_ondisk.h 2010-07-07 18:31:55.000000000 +0200
8803 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/gfs2_ondisk.h 2010-10-21 13:09:36.000000000 +0200
8804 @@ -211,6 +211,9 @@ enum {
8808 + gfs2fl_IXUnlink = 16,
8809 + gfs2fl_Barrier = 17,
8811 gfs2fl_TruncInProg = 29,
8812 gfs2fl_InheritDirectio = 30,
8813 gfs2fl_InheritJdata = 31,
8814 @@ -227,6 +230,9 @@ enum {
8815 #define GFS2_DIF_NOATIME 0x00000080
8816 #define GFS2_DIF_SYNC 0x00000100
8817 #define GFS2_DIF_SYSTEM 0x00000200 /* New in gfs2 */
8818 +#define GFS2_DIF_IXUNLINK 0x00010000
8819 +#define GFS2_DIF_BARRIER 0x00020000
8820 +#define GFS2_DIF_COW 0x00040000
8821 #define GFS2_DIF_TRUNC_IN_PROG 0x20000000 /* New in gfs2 */
8822 #define GFS2_DIF_INHERIT_DIRECTIO 0x40000000
8823 #define GFS2_DIF_INHERIT_JDATA 0x80000000
8824 diff -NurpP --minimal linux-2.6.36/include/linux/if_tun.h linux-2.6.36-vs2.3.0.36.33/include/linux/if_tun.h
8825 --- linux-2.6.36/include/linux/if_tun.h 2010-08-02 16:52:54.000000000 +0200
8826 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/if_tun.h 2010-10-21 13:09:36.000000000 +0200
8828 #define TUNDETACHFILTER _IOW('T', 214, struct sock_fprog)
8829 #define TUNGETVNETHDRSZ _IOR('T', 215, int)
8830 #define TUNSETVNETHDRSZ _IOW('T', 216, int)
8831 +#define TUNSETNID _IOW('T', 217, int)
8833 /* TUNSETIFF ifr flags */
8834 #define IFF_TUN 0x0001
8835 diff -NurpP --minimal linux-2.6.36/include/linux/init_task.h linux-2.6.36-vs2.3.0.36.33/include/linux/init_task.h
8836 --- linux-2.6.36/include/linux/init_task.h 2010-08-02 16:52:54.000000000 +0200
8837 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/init_task.h 2010-10-21 13:09:36.000000000 +0200
8838 @@ -172,6 +172,10 @@ extern struct cred init_cred;
8840 INIT_TRACE_RECURSION \
8841 INIT_TASK_RCU_PREEMPT(tsk) \
8843 + .vx_info = NULL, \
8845 + .nx_info = NULL, \
8849 diff -NurpP --minimal linux-2.6.36/include/linux/ipc.h linux-2.6.36-vs2.3.0.36.33/include/linux/ipc.h
8850 --- linux-2.6.36/include/linux/ipc.h 2009-12-03 20:02:55.000000000 +0100
8851 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/ipc.h 2010-10-21 13:09:36.000000000 +0200
8852 @@ -91,6 +91,7 @@ struct kern_ipc_perm
8860 diff -NurpP --minimal linux-2.6.36/include/linux/Kbuild linux-2.6.36-vs2.3.0.36.33/include/linux/Kbuild
8861 --- linux-2.6.36/include/linux/Kbuild 2010-10-21 13:07:52.000000000 +0200
8862 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/Kbuild 2010-10-21 14:12:01.000000000 +0200
8863 @@ -15,6 +15,7 @@ header-y += netfilter_bridge/
8864 header-y += netfilter_ipv4/
8865 header-y += netfilter_ipv6/
8867 +header-y += vserver/
8870 objhdr-y += version.h
8871 diff -NurpP --minimal linux-2.6.36/include/linux/loop.h linux-2.6.36-vs2.3.0.36.33/include/linux/loop.h
8872 --- linux-2.6.36/include/linux/loop.h 2009-09-10 15:26:25.000000000 +0200
8873 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/loop.h 2010-10-21 13:09:36.000000000 +0200
8874 @@ -45,6 +45,7 @@ struct loop_device {
8875 struct loop_func_table *lo_encryption;
8877 uid_t lo_key_owner; /* Who set the key */
8879 int (*ioctl)(struct loop_device *, int cmd,
8882 diff -NurpP --minimal linux-2.6.36/include/linux/magic.h linux-2.6.36-vs2.3.0.36.33/include/linux/magic.h
8883 --- linux-2.6.36/include/linux/magic.h 2010-07-07 18:31:55.000000000 +0200
8884 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/magic.h 2010-10-21 13:09:36.000000000 +0200
8887 #define ADFS_SUPER_MAGIC 0xadf5
8888 #define AFFS_SUPER_MAGIC 0xadff
8889 -#define AFS_SUPER_MAGIC 0x5346414F
8890 +#define AFS_SUPER_MAGIC 0x5346414F
8891 #define AUTOFS_SUPER_MAGIC 0x0187
8892 #define CODA_SUPER_MAGIC 0x73757245
8893 #define CRAMFS_MAGIC 0x28cd3d45 /* some random number */
8895 #define NFS_SUPER_MAGIC 0x6969
8896 #define OPENPROM_SUPER_MAGIC 0x9fa1
8897 #define PROC_SUPER_MAGIC 0x9fa0
8898 +#define DEVPTS_SUPER_MAGIC 0x1cd1
8899 #define QNX4_SUPER_MAGIC 0x002f /* qnx4 fs detection */
8901 #define REISERFS_SUPER_MAGIC 0x52654973 /* used by gcc */
8902 diff -NurpP --minimal linux-2.6.36/include/linux/major.h linux-2.6.36-vs2.3.0.36.33/include/linux/major.h
8903 --- linux-2.6.36/include/linux/major.h 2009-09-10 15:26:25.000000000 +0200
8904 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/major.h 2010-10-21 13:09:36.000000000 +0200
8906 #define HD_MAJOR IDE0_MAJOR
8907 #define PTY_SLAVE_MAJOR 3
8909 +#define VROOT_MAJOR 4
8910 #define TTYAUX_MAJOR 5
8913 diff -NurpP --minimal linux-2.6.36/include/linux/memcontrol.h linux-2.6.36-vs2.3.0.36.33/include/linux/memcontrol.h
8914 --- linux-2.6.36/include/linux/memcontrol.h 2010-10-21 13:07:53.000000000 +0200
8915 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/memcontrol.h 2010-10-21 13:09:36.000000000 +0200
8916 @@ -77,6 +77,13 @@ int task_in_mem_cgroup(struct task_struc
8917 extern struct mem_cgroup *try_get_mem_cgroup_from_page(struct page *page);
8918 extern struct mem_cgroup *mem_cgroup_from_task(struct task_struct *p);
8920 +extern u64 mem_cgroup_res_read_u64(struct mem_cgroup *mem, int member);
8921 +extern u64 mem_cgroup_memsw_read_u64(struct mem_cgroup *mem, int member);
8923 +extern s64 mem_cgroup_stat_read_cache(struct mem_cgroup *mem);
8924 +extern s64 mem_cgroup_stat_read_anon(struct mem_cgroup *mem);
8925 +extern s64 mem_cgroup_stat_read_mapped(struct mem_cgroup *mem);
8928 int mm_match_cgroup(const struct mm_struct *mm, const struct mem_cgroup *cgroup)
8930 diff -NurpP --minimal linux-2.6.36/include/linux/mm_types.h linux-2.6.36-vs2.3.0.36.33/include/linux/mm_types.h
8931 --- linux-2.6.36/include/linux/mm_types.h 2010-10-21 13:07:53.000000000 +0200
8932 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/mm_types.h 2010-10-21 13:09:36.000000000 +0200
8933 @@ -269,6 +269,7 @@ struct mm_struct {
8935 /* Architecture-specific MM context */
8936 mm_context_t context;
8937 + struct vx_info *mm_vx_info;
8939 /* Swap token stuff */
8941 diff -NurpP --minimal linux-2.6.36/include/linux/mount.h linux-2.6.36-vs2.3.0.36.33/include/linux/mount.h
8942 --- linux-2.6.36/include/linux/mount.h 2010-10-21 13:07:53.000000000 +0200
8943 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/mount.h 2010-10-21 13:09:36.000000000 +0200
8944 @@ -46,6 +46,9 @@ struct mnt_namespace;
8946 #define MNT_INTERNAL 0x4000
8948 +#define MNT_TAGID 0x10000
8949 +#define MNT_NOTAG 0x20000
8952 struct list_head mnt_hash;
8953 struct vfsmount *mnt_parent; /* fs we are mounted on */
8954 @@ -84,6 +87,7 @@ struct vfsmount {
8958 + tag_t mnt_tag; /* tagging used for vfsmount */
8961 static inline int *get_mnt_writers_ptr(struct vfsmount *mnt)
8962 diff -NurpP --minimal linux-2.6.36/include/linux/net.h linux-2.6.36-vs2.3.0.36.33/include/linux/net.h
8963 --- linux-2.6.36/include/linux/net.h 2010-08-02 16:52:55.000000000 +0200
8964 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/net.h 2010-10-21 13:09:36.000000000 +0200
8965 @@ -71,6 +71,7 @@ struct net;
8966 #define SOCK_NOSPACE 2
8967 #define SOCK_PASSCRED 3
8968 #define SOCK_PASSSEC 4
8969 +#define SOCK_USER_SOCKET 5
8971 #ifndef ARCH_HAS_SOCKET_TYPES
8973 diff -NurpP --minimal linux-2.6.36/include/linux/nfs_mount.h linux-2.6.36-vs2.3.0.36.33/include/linux/nfs_mount.h
8974 --- linux-2.6.36/include/linux/nfs_mount.h 2010-10-21 13:07:54.000000000 +0200
8975 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/nfs_mount.h 2010-10-21 13:09:36.000000000 +0200
8976 @@ -63,7 +63,8 @@ struct nfs_mount_data {
8977 #define NFS_MOUNT_SECFLAVOUR 0x2000 /* 5 */
8978 #define NFS_MOUNT_NORDIRPLUS 0x4000 /* 5 */
8979 #define NFS_MOUNT_UNSHARED 0x8000 /* 5 */
8980 -#define NFS_MOUNT_FLAGMASK 0xFFFF
8981 +#define NFS_MOUNT_TAGGED 0x10000 /* context tagging */
8982 +#define NFS_MOUNT_FLAGMASK 0x1FFFF
8984 /* The following are for internal use only */
8985 #define NFS_MOUNT_LOOKUP_CACHE_NONEG 0x10000
8986 diff -NurpP --minimal linux-2.6.36/include/linux/nsproxy.h linux-2.6.36-vs2.3.0.36.33/include/linux/nsproxy.h
8987 --- linux-2.6.36/include/linux/nsproxy.h 2009-06-11 17:13:17.000000000 +0200
8988 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/nsproxy.h 2010-10-21 13:09:36.000000000 +0200
8991 #include <linux/spinlock.h>
8992 #include <linux/sched.h>
8993 +#include <linux/vserver/debug.h>
8995 struct mnt_namespace;
8996 struct uts_namespace;
8997 @@ -63,22 +64,33 @@ static inline struct nsproxy *task_nspro
9000 int copy_namespaces(unsigned long flags, struct task_struct *tsk);
9001 +struct nsproxy *copy_nsproxy(struct nsproxy *orig);
9002 void exit_task_namespaces(struct task_struct *tsk);
9003 void switch_task_namespaces(struct task_struct *tsk, struct nsproxy *new);
9004 void free_nsproxy(struct nsproxy *ns);
9005 int unshare_nsproxy_namespaces(unsigned long, struct nsproxy **,
9006 struct fs_struct *);
9008 -static inline void put_nsproxy(struct nsproxy *ns)
9009 +#define get_nsproxy(n) __get_nsproxy(n, __FILE__, __LINE__)
9011 +static inline void __get_nsproxy(struct nsproxy *ns,
9012 + const char *_file, int _line)
9014 - if (atomic_dec_and_test(&ns->count)) {
9017 + vxlprintk(VXD_CBIT(space, 0), "get_nsproxy(%p[%u])",
9018 + ns, atomic_read(&ns->count), _file, _line);
9019 + atomic_inc(&ns->count);
9022 -static inline void get_nsproxy(struct nsproxy *ns)
9023 +#define put_nsproxy(n) __put_nsproxy(n, __FILE__, __LINE__)
9025 +static inline void __put_nsproxy(struct nsproxy *ns,
9026 + const char *_file, int _line)
9028 - atomic_inc(&ns->count);
9029 + vxlprintk(VXD_CBIT(space, 0), "put_nsproxy(%p[%u])",
9030 + ns, atomic_read(&ns->count), _file, _line);
9031 + if (atomic_dec_and_test(&ns->count)) {
9036 #ifdef CONFIG_CGROUP_NS
9037 diff -NurpP --minimal linux-2.6.36/include/linux/pid.h linux-2.6.36-vs2.3.0.36.33/include/linux/pid.h
9038 --- linux-2.6.36/include/linux/pid.h 2009-03-24 14:22:43.000000000 +0100
9039 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/pid.h 2010-10-21 13:09:36.000000000 +0200
9040 @@ -8,7 +8,8 @@ enum pid_type
9050 @@ -160,6 +161,7 @@ static inline pid_t pid_nr(struct pid *p
9053 pid_t pid_nr_ns(struct pid *pid, struct pid_namespace *ns);
9054 +pid_t pid_unmapped_nr_ns(struct pid *pid, struct pid_namespace *ns);
9055 pid_t pid_vnr(struct pid *pid);
9057 #define do_each_pid_task(pid, type, task) \
9058 diff -NurpP --minimal linux-2.6.36/include/linux/proc_fs.h linux-2.6.36-vs2.3.0.36.33/include/linux/proc_fs.h
9059 --- linux-2.6.36/include/linux/proc_fs.h 2009-12-03 20:02:56.000000000 +0100
9060 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/proc_fs.h 2010-10-21 13:09:36.000000000 +0200
9061 @@ -56,6 +56,7 @@ struct proc_dir_entry {
9067 const struct inode_operations *proc_iops;
9069 @@ -250,12 +251,18 @@ kclist_add(struct kcore_list *new, void
9070 extern void kclist_add(struct kcore_list *, void *, size_t, int type);
9077 int (*proc_get_link)(struct inode *, struct path *);
9078 int (*proc_read)(struct task_struct *task, char *page);
9079 int (*proc_show)(struct seq_file *m,
9080 struct pid_namespace *ns, struct pid *pid,
9081 struct task_struct *task);
9082 + int (*proc_vs_read)(char *page);
9083 + int (*proc_vxi_read)(struct vx_info *vxi, char *page);
9084 + int (*proc_nxi_read)(struct nx_info *nxi, char *page);
9087 struct ctl_table_header;
9088 @@ -263,6 +270,7 @@ struct ctl_table;
9095 struct proc_dir_entry *pde;
9096 diff -NurpP --minimal linux-2.6.36/include/linux/quotaops.h linux-2.6.36-vs2.3.0.36.33/include/linux/quotaops.h
9097 --- linux-2.6.36/include/linux/quotaops.h 2010-10-21 13:07:54.000000000 +0200
9098 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/quotaops.h 2010-10-21 13:09:36.000000000 +0200
9100 #define _LINUX_QUOTAOPS_
9102 #include <linux/fs.h>
9103 +#include <linux/vs_dlimit.h>
9105 #define DQUOT_SPACE_WARN 0x1
9106 #define DQUOT_SPACE_RESERVE 0x2
9107 @@ -205,11 +206,12 @@ static inline void dquot_drop(struct ino
9109 static inline int dquot_alloc_inode(const struct inode *inode)
9112 + return dl_alloc_inode(inode);
9115 static inline void dquot_free_inode(const struct inode *inode)
9117 + dl_free_inode(inode);
9120 static inline int dquot_transfer(struct inode *inode, struct iattr *iattr)
9121 @@ -220,6 +222,10 @@ static inline int dquot_transfer(struct
9122 static inline int __dquot_alloc_space(struct inode *inode, qsize_t number,
9127 + if ((ret = dl_alloc_space(inode, number)))
9129 if (!(flags & DQUOT_SPACE_RESERVE))
9130 inode_add_bytes(inode, number);
9132 @@ -230,6 +236,7 @@ static inline void __dquot_free_space(st
9134 if (!(flags & DQUOT_SPACE_RESERVE))
9135 inode_sub_bytes(inode, number);
9136 + dl_free_space(inode, number);
9139 static inline int dquot_claim_space_nodirty(struct inode *inode, qsize_t number)
9140 diff -NurpP --minimal linux-2.6.36/include/linux/reboot.h linux-2.6.36-vs2.3.0.36.33/include/linux/reboot.h
9141 --- linux-2.6.36/include/linux/reboot.h 2010-07-07 18:31:56.000000000 +0200
9142 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/reboot.h 2010-10-21 13:09:36.000000000 +0200
9144 #define LINUX_REBOOT_CMD_RESTART2 0xA1B2C3D4
9145 #define LINUX_REBOOT_CMD_SW_SUSPEND 0xD000FCE2
9146 #define LINUX_REBOOT_CMD_KEXEC 0x45584543
9147 +#define LINUX_REBOOT_CMD_OOM 0xDEADBEEF
9151 diff -NurpP --minimal linux-2.6.36/include/linux/reiserfs_fs.h linux-2.6.36-vs2.3.0.36.33/include/linux/reiserfs_fs.h
9152 --- linux-2.6.36/include/linux/reiserfs_fs.h 2010-10-21 13:07:54.000000000 +0200
9153 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/reiserfs_fs.h 2010-10-21 13:09:36.000000000 +0200
9154 @@ -977,6 +977,11 @@ struct stat_data_v1 {
9155 #define REISERFS_COMPR_FL FS_COMPR_FL
9156 #define REISERFS_NOTAIL_FL FS_NOTAIL_FL
9158 +/* unfortunately reiserfs sdattr is only 16 bit */
9159 +#define REISERFS_IXUNLINK_FL (FS_IXUNLINK_FL >> 16)
9160 +#define REISERFS_BARRIER_FL (FS_BARRIER_FL >> 16)
9161 +#define REISERFS_COW_FL (FS_COW_FL >> 16)
9163 /* persistent flags that file inherits from the parent directory */
9164 #define REISERFS_INHERIT_MASK ( REISERFS_IMMUTABLE_FL | \
9165 REISERFS_SYNC_FL | \
9166 @@ -986,6 +991,9 @@ struct stat_data_v1 {
9167 REISERFS_COMPR_FL | \
9168 REISERFS_NOTAIL_FL )
9170 +#define REISERFS_FL_USER_VISIBLE 0x80FF
9171 +#define REISERFS_FL_USER_MODIFIABLE 0x80FF
9173 /* Stat Data on disk (reiserfs version of UFS disk inode minus the
9176 @@ -2071,6 +2079,7 @@ static inline void reiserfs_update_sd(st
9177 void sd_attrs_to_i_attrs(__u16 sd_attrs, struct inode *inode);
9178 void i_attrs_to_sd_attrs(struct inode *inode, __u16 * sd_attrs);
9179 int reiserfs_setattr(struct dentry *dentry, struct iattr *attr);
9180 +int reiserfs_sync_flags(struct inode *inode, int, int);
9183 void set_de_name_and_namelen(struct reiserfs_dir_entry *de);
9184 diff -NurpP --minimal linux-2.6.36/include/linux/reiserfs_fs_sb.h linux-2.6.36-vs2.3.0.36.33/include/linux/reiserfs_fs_sb.h
9185 --- linux-2.6.36/include/linux/reiserfs_fs_sb.h 2010-02-25 11:52:07.000000000 +0100
9186 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/reiserfs_fs_sb.h 2010-10-21 13:09:36.000000000 +0200
9187 @@ -476,6 +476,7 @@ enum reiserfs_mount_options {
9188 REISERFS_EXPOSE_PRIVROOT,
9189 REISERFS_BARRIER_NONE,
9190 REISERFS_BARRIER_FLUSH,
9193 /* Actions on error */
9194 REISERFS_ERROR_PANIC,
9195 diff -NurpP --minimal linux-2.6.36/include/linux/sched.h linux-2.6.36-vs2.3.0.36.33/include/linux/sched.h
9196 --- linux-2.6.36/include/linux/sched.h 2010-10-21 13:07:54.000000000 +0200
9197 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/sched.h 2010-10-21 13:09:36.000000000 +0200
9198 @@ -1339,6 +1339,14 @@ struct task_struct {
9202 +/* vserver context data */
9203 + struct vx_info *vx_info;
9204 + struct nx_info *nx_info;
9210 /* Thread group tracking */
9213 @@ -1573,6 +1581,11 @@ struct pid_namespace;
9214 pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type,
9215 struct pid_namespace *ns);
9217 +#include <linux/vserver/base.h>
9218 +#include <linux/vserver/context.h>
9219 +#include <linux/vserver/debug.h>
9220 +#include <linux/vserver/pid.h>
9222 static inline pid_t task_pid_nr(struct task_struct *tsk)
9225 @@ -1586,7 +1599,8 @@ static inline pid_t task_pid_nr_ns(struc
9227 static inline pid_t task_pid_vnr(struct task_struct *tsk)
9229 - return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
9230 + // return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
9231 + return vx_map_pid(__task_pid_nr_ns(tsk, PIDTYPE_PID, NULL));
9235 @@ -1599,7 +1613,7 @@ pid_t task_tgid_nr_ns(struct task_struct
9237 static inline pid_t task_tgid_vnr(struct task_struct *tsk)
9239 - return pid_vnr(task_tgid(tsk));
9240 + return vx_map_tgid(pid_vnr(task_tgid(tsk)));
9244 diff -NurpP --minimal linux-2.6.36/include/linux/shmem_fs.h linux-2.6.36-vs2.3.0.36.33/include/linux/shmem_fs.h
9245 --- linux-2.6.36/include/linux/shmem_fs.h 2010-10-21 13:07:55.000000000 +0200
9246 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/shmem_fs.h 2010-10-21 13:09:36.000000000 +0200
9249 #define SHMEM_NR_DIRECT 16
9251 +#define TMPFS_SUPER_MAGIC 0x01021994
9254 struct shmem_inode_info {
9256 unsigned long flags;
9257 diff -NurpP --minimal linux-2.6.36/include/linux/stat.h linux-2.6.36-vs2.3.0.36.33/include/linux/stat.h
9258 --- linux-2.6.36/include/linux/stat.h 2008-12-25 00:26:37.000000000 +0100
9259 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/stat.h 2010-10-21 13:09:36.000000000 +0200
9260 @@ -66,6 +66,7 @@ struct kstat {
9267 struct timespec atime;
9268 diff -NurpP --minimal linux-2.6.36/include/linux/sunrpc/auth.h linux-2.6.36-vs2.3.0.36.33/include/linux/sunrpc/auth.h
9269 --- linux-2.6.36/include/linux/sunrpc/auth.h 2010-10-21 13:07:55.000000000 +0200
9270 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/sunrpc/auth.h 2010-10-21 13:09:36.000000000 +0200
9276 struct group_info *group_info;
9277 unsigned char machine_cred : 1;
9279 diff -NurpP --minimal linux-2.6.36/include/linux/sunrpc/clnt.h linux-2.6.36-vs2.3.0.36.33/include/linux/sunrpc/clnt.h
9280 --- linux-2.6.36/include/linux/sunrpc/clnt.h 2010-10-21 13:07:55.000000000 +0200
9281 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/sunrpc/clnt.h 2010-10-21 13:09:36.000000000 +0200
9282 @@ -49,7 +49,8 @@ struct rpc_clnt {
9283 unsigned int cl_softrtry : 1,/* soft timeouts */
9284 cl_discrtry : 1,/* disconnect before retry */
9285 cl_autobind : 1,/* use getport() */
9286 - cl_chatty : 1;/* be verbose */
9287 + cl_chatty : 1,/* be verbose */
9288 + cl_tag : 1;/* context tagging */
9290 struct rpc_rtt * cl_rtt; /* RTO estimator data */
9291 const struct rpc_timeout *cl_timeout; /* Timeout strategy */
9292 diff -NurpP --minimal linux-2.6.36/include/linux/syscalls.h linux-2.6.36-vs2.3.0.36.33/include/linux/syscalls.h
9293 --- linux-2.6.36/include/linux/syscalls.h 2010-10-21 13:07:55.000000000 +0200
9294 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/syscalls.h 2010-10-21 13:09:36.000000000 +0200
9295 @@ -478,6 +478,8 @@ asmlinkage long sys_symlink(const char _
9296 asmlinkage long sys_unlink(const char __user *pathname);
9297 asmlinkage long sys_rename(const char __user *oldname,
9298 const char __user *newname);
9299 +asmlinkage long sys_copyfile(const char __user *from, const char __user *to,
9301 asmlinkage long sys_chmod(const char __user *filename, mode_t mode);
9302 asmlinkage long sys_fchmod(unsigned int fd, mode_t mode);
9304 diff -NurpP --minimal linux-2.6.36/include/linux/sysctl.h linux-2.6.36-vs2.3.0.36.33/include/linux/sysctl.h
9305 --- linux-2.6.36/include/linux/sysctl.h 2010-08-02 16:52:56.000000000 +0200
9306 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/sysctl.h 2010-10-21 13:09:36.000000000 +0200
9307 @@ -60,6 +60,7 @@ enum
9308 CTL_ABI=9, /* Binary emulation */
9309 CTL_CPU=10, /* CPU stuff (speed scaling, etc) */
9310 CTL_ARLAN=254, /* arlan wireless driver */
9311 + CTL_VSERVER=4242, /* Linux-VServer debug */
9312 CTL_S390DBF=5677, /* s390 debug */
9313 CTL_SUNRPC=7249, /* sunrpc debug */
9314 CTL_PM=9899, /* frv power management */
9315 @@ -94,6 +95,7 @@ enum
9317 KERN_PANIC=15, /* int: panic timeout */
9318 KERN_REALROOTDEV=16, /* real root device to mount after initrd */
9319 + KERN_VSHELPER=17, /* string: path to vshelper policy agent */
9321 KERN_SPARC_REBOOT=21, /* reboot command on Sparc */
9322 KERN_CTLALTDEL=22, /* int: allow ctl-alt-del to reboot */
9323 diff -NurpP --minimal linux-2.6.36/include/linux/sysfs.h linux-2.6.36-vs2.3.0.36.33/include/linux/sysfs.h
9324 --- linux-2.6.36/include/linux/sysfs.h 2010-10-21 13:07:55.000000000 +0200
9325 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/sysfs.h 2010-10-21 13:09:36.000000000 +0200
9327 #include <linux/kobject_ns.h>
9328 #include <asm/atomic.h>
9330 +#define SYSFS_SUPER_MAGIC 0x62656572
9335 diff -NurpP --minimal linux-2.6.36/include/linux/time.h linux-2.6.36-vs2.3.0.36.33/include/linux/time.h
9336 --- linux-2.6.36/include/linux/time.h 2010-10-21 13:07:55.000000000 +0200
9337 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/time.h 2010-10-21 13:09:36.000000000 +0200
9338 @@ -252,6 +252,9 @@ static __always_inline void timespec_add
9339 a->tv_sec += __iter_div_u64_rem(a->tv_nsec + ns, NSEC_PER_SEC, &ns);
9343 +#include <linux/vs_time.h>
9345 #endif /* __KERNEL__ */
9347 #define NFDBITS __NFDBITS
9348 diff -NurpP --minimal linux-2.6.36/include/linux/types.h linux-2.6.36-vs2.3.0.36.33/include/linux/types.h
9349 --- linux-2.6.36/include/linux/types.h 2010-10-21 13:07:55.000000000 +0200
9350 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/types.h 2010-10-21 13:09:36.000000000 +0200
9351 @@ -40,6 +40,9 @@ typedef __kernel_uid32_t uid_t;
9352 typedef __kernel_gid32_t gid_t;
9353 typedef __kernel_uid16_t uid16_t;
9354 typedef __kernel_gid16_t gid16_t;
9355 +typedef unsigned int xid_t;
9356 +typedef unsigned int nid_t;
9357 +typedef unsigned int tag_t;
9359 typedef unsigned long uintptr_t;
9361 diff -NurpP --minimal linux-2.6.36/include/linux/vroot.h linux-2.6.36-vs2.3.0.36.33/include/linux/vroot.h
9362 --- linux-2.6.36/include/linux/vroot.h 1970-01-01 01:00:00.000000000 +0100
9363 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vroot.h 2010-10-21 13:09:36.000000000 +0200
9367 + * include/linux/vroot.h
9369 + * written by Herbert Pötzl, 9/11/2002
9370 + * ported to 2.6 by Herbert Pötzl, 30/12/2004
9372 + * Copyright (C) 2002-2007 by Herbert Pötzl.
9373 + * Redistribution of this file is permitted under the
9374 + * GNU General Public License.
9377 +#ifndef _LINUX_VROOT_H
9378 +#define _LINUX_VROOT_H
9383 +/* Possible states of device */
9389 +struct vroot_device {
9393 + struct semaphore vr_ctl_mutex;
9394 + struct block_device *vr_device;
9399 +typedef struct block_device *(vroot_grb_func)(struct block_device *);
9401 +extern int register_vroot_grb(vroot_grb_func *);
9402 +extern int unregister_vroot_grb(vroot_grb_func *);
9404 +#endif /* __KERNEL__ */
9406 +#define MAX_VROOT_DEFAULT 8
9409 + * IOCTL commands --- we will commandeer 0x56 ('V')
9412 +#define VROOT_SET_DEV 0x5600
9413 +#define VROOT_CLR_DEV 0x5601
9415 +#endif /* _LINUX_VROOT_H */
9416 diff -NurpP --minimal linux-2.6.36/include/linux/vs_base.h linux-2.6.36-vs2.3.0.36.33/include/linux/vs_base.h
9417 --- linux-2.6.36/include/linux/vs_base.h 1970-01-01 01:00:00.000000000 +0100
9418 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vs_base.h 2010-10-21 13:09:36.000000000 +0200
9423 +#include "vserver/base.h"
9424 +#include "vserver/check.h"
9425 +#include "vserver/debug.h"
9428 +#warning duplicate inclusion
9430 diff -NurpP --minimal linux-2.6.36/include/linux/vs_context.h linux-2.6.36-vs2.3.0.36.33/include/linux/vs_context.h
9431 --- linux-2.6.36/include/linux/vs_context.h 1970-01-01 01:00:00.000000000 +0100
9432 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vs_context.h 2010-10-21 13:09:36.000000000 +0200
9434 +#ifndef _VS_CONTEXT_H
9435 +#define _VS_CONTEXT_H
9437 +#include "vserver/base.h"
9438 +#include "vserver/check.h"
9439 +#include "vserver/context.h"
9440 +#include "vserver/history.h"
9441 +#include "vserver/debug.h"
9443 +#include <linux/sched.h>
9446 +#define get_vx_info(i) __get_vx_info(i, __FILE__, __LINE__, __HERE__)
9448 +static inline struct vx_info *__get_vx_info(struct vx_info *vxi,
9449 + const char *_file, int _line, void *_here)
9454 + vxlprintk(VXD_CBIT(xid, 2), "get_vx_info(%p[#%d.%d])",
9455 + vxi, vxi ? vxi->vx_id : 0,
9456 + vxi ? atomic_read(&vxi->vx_usecnt) : 0,
9458 + __vxh_get_vx_info(vxi, _here);
9460 + atomic_inc(&vxi->vx_usecnt);
9465 +extern void free_vx_info(struct vx_info *);
9467 +#define put_vx_info(i) __put_vx_info(i, __FILE__, __LINE__, __HERE__)
9469 +static inline void __put_vx_info(struct vx_info *vxi,
9470 + const char *_file, int _line, void *_here)
9475 + vxlprintk(VXD_CBIT(xid, 2), "put_vx_info(%p[#%d.%d])",
9476 + vxi, vxi ? vxi->vx_id : 0,
9477 + vxi ? atomic_read(&vxi->vx_usecnt) : 0,
9479 + __vxh_put_vx_info(vxi, _here);
9481 + if (atomic_dec_and_test(&vxi->vx_usecnt))
9482 + free_vx_info(vxi);
9486 +#define init_vx_info(p, i) \
9487 + __init_vx_info(p, i, __FILE__, __LINE__, __HERE__)
9489 +static inline void __init_vx_info(struct vx_info **vxp, struct vx_info *vxi,
9490 + const char *_file, int _line, void *_here)
9493 + vxlprintk(VXD_CBIT(xid, 3),
9494 + "init_vx_info(%p[#%d.%d])",
9495 + vxi, vxi ? vxi->vx_id : 0,
9496 + vxi ? atomic_read(&vxi->vx_usecnt) : 0,
9498 + __vxh_init_vx_info(vxi, vxp, _here);
9500 + atomic_inc(&vxi->vx_usecnt);
9506 +#define set_vx_info(p, i) \
9507 + __set_vx_info(p, i, __FILE__, __LINE__, __HERE__)
9509 +static inline void __set_vx_info(struct vx_info **vxp, struct vx_info *vxi,
9510 + const char *_file, int _line, void *_here)
9512 + struct vx_info *vxo;
9517 + vxlprintk(VXD_CBIT(xid, 3), "set_vx_info(%p[#%d.%d])",
9518 + vxi, vxi ? vxi->vx_id : 0,
9519 + vxi ? atomic_read(&vxi->vx_usecnt) : 0,
9521 + __vxh_set_vx_info(vxi, vxp, _here);
9523 + atomic_inc(&vxi->vx_usecnt);
9524 + vxo = xchg(vxp, vxi);
9529 +#define clr_vx_info(p) __clr_vx_info(p, __FILE__, __LINE__, __HERE__)
9531 +static inline void __clr_vx_info(struct vx_info **vxp,
9532 + const char *_file, int _line, void *_here)
9534 + struct vx_info *vxo;
9536 + vxo = xchg(vxp, NULL);
9540 + vxlprintk(VXD_CBIT(xid, 3), "clr_vx_info(%p[#%d.%d])",
9541 + vxo, vxo ? vxo->vx_id : 0,
9542 + vxo ? atomic_read(&vxo->vx_usecnt) : 0,
9544 + __vxh_clr_vx_info(vxo, vxp, _here);
9546 + if (atomic_dec_and_test(&vxo->vx_usecnt))
9547 + free_vx_info(vxo);
9551 +#define claim_vx_info(v, p) \
9552 + __claim_vx_info(v, p, __FILE__, __LINE__, __HERE__)
9554 +static inline void __claim_vx_info(struct vx_info *vxi,
9555 + struct task_struct *task,
9556 + const char *_file, int _line, void *_here)
9558 + vxlprintk(VXD_CBIT(xid, 3), "claim_vx_info(%p[#%d.%d.%d]) %p",
9559 + vxi, vxi ? vxi->vx_id : 0,
9560 + vxi ? atomic_read(&vxi->vx_usecnt) : 0,
9561 + vxi ? atomic_read(&vxi->vx_tasks) : 0,
9562 + task, _file, _line);
9563 + __vxh_claim_vx_info(vxi, task, _here);
9565 + atomic_inc(&vxi->vx_tasks);
9569 +extern void unhash_vx_info(struct vx_info *);
9571 +#define release_vx_info(v, p) \
9572 + __release_vx_info(v, p, __FILE__, __LINE__, __HERE__)
9574 +static inline void __release_vx_info(struct vx_info *vxi,
9575 + struct task_struct *task,
9576 + const char *_file, int _line, void *_here)
9578 + vxlprintk(VXD_CBIT(xid, 3), "release_vx_info(%p[#%d.%d.%d]) %p",
9579 + vxi, vxi ? vxi->vx_id : 0,
9580 + vxi ? atomic_read(&vxi->vx_usecnt) : 0,
9581 + vxi ? atomic_read(&vxi->vx_tasks) : 0,
9582 + task, _file, _line);
9583 + __vxh_release_vx_info(vxi, task, _here);
9587 + if (atomic_dec_and_test(&vxi->vx_tasks))
9588 + unhash_vx_info(vxi);
9592 +#define task_get_vx_info(p) \
9593 + __task_get_vx_info(p, __FILE__, __LINE__, __HERE__)
9595 +static inline struct vx_info *__task_get_vx_info(struct task_struct *p,
9596 + const char *_file, int _line, void *_here)
9598 + struct vx_info *vxi;
9601 + vxlprintk(VXD_CBIT(xid, 5), "task_get_vx_info(%p)",
9603 + vxi = __get_vx_info(p->vx_info, _file, _line, _here);
9609 +static inline void __wakeup_vx_info(struct vx_info *vxi)
9611 + if (waitqueue_active(&vxi->vx_wait))
9612 + wake_up_interruptible(&vxi->vx_wait);
9616 +#define enter_vx_info(v, s) __enter_vx_info(v, s, __FILE__, __LINE__)
9618 +static inline void __enter_vx_info(struct vx_info *vxi,
9619 + struct vx_info_save *vxis, const char *_file, int _line)
9621 + vxlprintk(VXD_CBIT(xid, 5), "enter_vx_info(%p[#%d],%p) %p[#%d,%p]",
9622 + vxi, vxi ? vxi->vx_id : 0, vxis, current,
9623 + current->xid, current->vx_info, _file, _line);
9624 + vxis->vxi = xchg(¤t->vx_info, vxi);
9625 + vxis->xid = current->xid;
9626 + current->xid = vxi ? vxi->vx_id : 0;
9629 +#define leave_vx_info(s) __leave_vx_info(s, __FILE__, __LINE__)
9631 +static inline void __leave_vx_info(struct vx_info_save *vxis,
9632 + const char *_file, int _line)
9634 + vxlprintk(VXD_CBIT(xid, 5), "leave_vx_info(%p[#%d,%p]) %p[#%d,%p]",
9635 + vxis, vxis->xid, vxis->vxi, current,
9636 + current->xid, current->vx_info, _file, _line);
9637 + (void)xchg(¤t->vx_info, vxis->vxi);
9638 + current->xid = vxis->xid;
9642 +static inline void __enter_vx_admin(struct vx_info_save *vxis)
9644 + vxis->vxi = xchg(¤t->vx_info, NULL);
9645 + vxis->xid = xchg(¤t->xid, (xid_t)0);
9648 +static inline void __leave_vx_admin(struct vx_info_save *vxis)
9650 + (void)xchg(¤t->xid, vxis->xid);
9651 + (void)xchg(¤t->vx_info, vxis->vxi);
9654 +#define task_is_init(p) \
9655 + __task_is_init(p, __FILE__, __LINE__, __HERE__)
9657 +static inline int __task_is_init(struct task_struct *p,
9658 + const char *_file, int _line, void *_here)
9660 + int is_init = is_global_init(p);
9664 + is_init = p->vx_info->vx_initpid == p->pid;
9669 +extern void exit_vx_info(struct task_struct *, int);
9670 +extern void exit_vx_info_early(struct task_struct *, int);
9674 +#warning duplicate inclusion
9676 diff -NurpP --minimal linux-2.6.36/include/linux/vs_cowbl.h linux-2.6.36-vs2.3.0.36.33/include/linux/vs_cowbl.h
9677 --- linux-2.6.36/include/linux/vs_cowbl.h 1970-01-01 01:00:00.000000000 +0100
9678 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vs_cowbl.h 2010-10-21 13:09:36.000000000 +0200
9680 +#ifndef _VS_COWBL_H
9681 +#define _VS_COWBL_H
9683 +#include <linux/fs.h>
9684 +#include <linux/dcache.h>
9685 +#include <linux/namei.h>
9686 +#include <linux/slab.h>
9688 +extern struct dentry *cow_break_link(const char *pathname);
9690 +static inline int cow_check_and_break(struct path *path)
9692 + struct inode *inode = path->dentry->d_inode;
9695 + /* do we need this check? */
9696 + if (IS_RDONLY(inode))
9699 + if (IS_COW(inode)) {
9700 + if (IS_COW_LINK(inode)) {
9701 + struct dentry *new_dentry, *old_dentry = path->dentry;
9704 + buf = kmalloc(PATH_MAX, GFP_KERNEL);
9708 + pp = d_path(path, buf, PATH_MAX);
9709 + new_dentry = cow_break_link(pp);
9711 + if (!IS_ERR(new_dentry)) {
9712 + path->dentry = new_dentry;
9715 + error = PTR_ERR(new_dentry);
9717 + inode->i_flags &= ~(S_IXUNLINK | S_IMMUTABLE);
9718 + inode->i_ctime = CURRENT_TIME;
9719 + mark_inode_dirty(inode);
9726 +#warning duplicate inclusion
9728 diff -NurpP --minimal linux-2.6.36/include/linux/vs_cvirt.h linux-2.6.36-vs2.3.0.36.33/include/linux/vs_cvirt.h
9729 --- linux-2.6.36/include/linux/vs_cvirt.h 1970-01-01 01:00:00.000000000 +0100
9730 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vs_cvirt.h 2010-10-21 13:09:36.000000000 +0200
9732 +#ifndef _VS_CVIRT_H
9733 +#define _VS_CVIRT_H
9735 +#include "vserver/cvirt.h"
9736 +#include "vserver/context.h"
9737 +#include "vserver/base.h"
9738 +#include "vserver/check.h"
9739 +#include "vserver/debug.h"
9742 +static inline void vx_activate_task(struct task_struct *p)
9744 + struct vx_info *vxi;
9746 + if ((vxi = p->vx_info)) {
9747 + vx_update_load(vxi);
9748 + atomic_inc(&vxi->cvirt.nr_running);
9752 +static inline void vx_deactivate_task(struct task_struct *p)
9754 + struct vx_info *vxi;
9756 + if ((vxi = p->vx_info)) {
9757 + vx_update_load(vxi);
9758 + atomic_dec(&vxi->cvirt.nr_running);
9762 +static inline void vx_uninterruptible_inc(struct task_struct *p)
9764 + struct vx_info *vxi;
9766 + if ((vxi = p->vx_info))
9767 + atomic_inc(&vxi->cvirt.nr_uninterruptible);
9770 +static inline void vx_uninterruptible_dec(struct task_struct *p)
9772 + struct vx_info *vxi;
9774 + if ((vxi = p->vx_info))
9775 + atomic_dec(&vxi->cvirt.nr_uninterruptible);
9780 +#warning duplicate inclusion
9782 diff -NurpP --minimal linux-2.6.36/include/linux/vs_device.h linux-2.6.36-vs2.3.0.36.33/include/linux/vs_device.h
9783 --- linux-2.6.36/include/linux/vs_device.h 1970-01-01 01:00:00.000000000 +0100
9784 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vs_device.h 2010-10-21 13:09:36.000000000 +0200
9786 +#ifndef _VS_DEVICE_H
9787 +#define _VS_DEVICE_H
9789 +#include "vserver/base.h"
9790 +#include "vserver/device.h"
9791 +#include "vserver/debug.h"
9794 +#ifdef CONFIG_VSERVER_DEVICE
9796 +int vs_map_device(struct vx_info *, dev_t, dev_t *, umode_t);
9798 +#define vs_device_perm(v, d, m, p) \
9799 + ((vs_map_device(current_vx_info(), d, NULL, m) & (p)) == (p))
9804 +int vs_map_device(struct vx_info *vxi,
9805 + dev_t device, dev_t *target, umode_t mode)
9812 +#define vs_device_perm(v, d, m, p) ((p) == (p))
9817 +#define vs_map_chrdev(d, t, p) \
9818 + ((vs_map_device(current_vx_info(), d, t, S_IFCHR) & (p)) == (p))
9819 +#define vs_map_blkdev(d, t, p) \
9820 + ((vs_map_device(current_vx_info(), d, t, S_IFBLK) & (p)) == (p))
9822 +#define vs_chrdev_perm(d, p) \
9823 + vs_device_perm(current_vx_info(), d, S_IFCHR, p)
9824 +#define vs_blkdev_perm(d, p) \
9825 + vs_device_perm(current_vx_info(), d, S_IFBLK, p)
9829 +#warning duplicate inclusion
9831 diff -NurpP --minimal linux-2.6.36/include/linux/vs_dlimit.h linux-2.6.36-vs2.3.0.36.33/include/linux/vs_dlimit.h
9832 --- linux-2.6.36/include/linux/vs_dlimit.h 1970-01-01 01:00:00.000000000 +0100
9833 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vs_dlimit.h 2010-10-21 13:09:36.000000000 +0200
9835 +#ifndef _VS_DLIMIT_H
9836 +#define _VS_DLIMIT_H
9838 +#include <linux/fs.h>
9840 +#include "vserver/dlimit.h"
9841 +#include "vserver/base.h"
9842 +#include "vserver/debug.h"
9845 +#define get_dl_info(i) __get_dl_info(i, __FILE__, __LINE__)
9847 +static inline struct dl_info *__get_dl_info(struct dl_info *dli,
9848 + const char *_file, int _line)
9852 + vxlprintk(VXD_CBIT(dlim, 4), "get_dl_info(%p[#%d.%d])",
9853 + dli, dli ? dli->dl_tag : 0,
9854 + dli ? atomic_read(&dli->dl_usecnt) : 0,
9856 + atomic_inc(&dli->dl_usecnt);
9861 +#define free_dl_info(i) \
9862 + call_rcu(&(i)->dl_rcu, rcu_free_dl_info)
9864 +#define put_dl_info(i) __put_dl_info(i, __FILE__, __LINE__)
9866 +static inline void __put_dl_info(struct dl_info *dli,
9867 + const char *_file, int _line)
9871 + vxlprintk(VXD_CBIT(dlim, 4), "put_dl_info(%p[#%d.%d])",
9872 + dli, dli ? dli->dl_tag : 0,
9873 + dli ? atomic_read(&dli->dl_usecnt) : 0,
9875 + if (atomic_dec_and_test(&dli->dl_usecnt))
9876 + free_dl_info(dli);
9880 +#define __dlimit_char(d) ((d) ? '*' : ' ')
9882 +static inline int __dl_alloc_space(struct super_block *sb,
9883 + tag_t tag, dlsize_t nr, const char *file, int line)
9885 + struct dl_info *dli = NULL;
9890 + dli = locate_dl_info(sb, tag);
9894 + spin_lock(&dli->dl_lock);
9895 + ret = (dli->dl_space_used + nr > dli->dl_space_total);
9897 + dli->dl_space_used += nr;
9898 + spin_unlock(&dli->dl_lock);
9901 + vxlprintk(VXD_CBIT(dlim, 1),
9902 + "ALLOC (%p,#%d)%c %lld bytes (%d)",
9903 + sb, tag, __dlimit_char(dli), (long long)nr,
9905 + return ret ? -ENOSPC : 0;
9908 +static inline void __dl_free_space(struct super_block *sb,
9909 + tag_t tag, dlsize_t nr, const char *_file, int _line)
9911 + struct dl_info *dli = NULL;
9915 + dli = locate_dl_info(sb, tag);
9919 + spin_lock(&dli->dl_lock);
9920 + if (dli->dl_space_used > nr)
9921 + dli->dl_space_used -= nr;
9923 + dli->dl_space_used = 0;
9924 + spin_unlock(&dli->dl_lock);
9927 + vxlprintk(VXD_CBIT(dlim, 1),
9928 + "FREE (%p,#%d)%c %lld bytes",
9929 + sb, tag, __dlimit_char(dli), (long long)nr,
9933 +static inline int __dl_alloc_inode(struct super_block *sb,
9934 + tag_t tag, const char *_file, int _line)
9936 + struct dl_info *dli;
9939 + dli = locate_dl_info(sb, tag);
9943 + spin_lock(&dli->dl_lock);
9944 + dli->dl_inodes_used++;
9945 + ret = (dli->dl_inodes_used > dli->dl_inodes_total);
9946 + spin_unlock(&dli->dl_lock);
9949 + vxlprintk(VXD_CBIT(dlim, 0),
9950 + "ALLOC (%p,#%d)%c inode (%d)",
9951 + sb, tag, __dlimit_char(dli), ret, _file, _line);
9952 + return ret ? -ENOSPC : 0;
9955 +static inline void __dl_free_inode(struct super_block *sb,
9956 + tag_t tag, const char *_file, int _line)
9958 + struct dl_info *dli;
9960 + dli = locate_dl_info(sb, tag);
9964 + spin_lock(&dli->dl_lock);
9965 + if (dli->dl_inodes_used > 1)
9966 + dli->dl_inodes_used--;
9968 + dli->dl_inodes_used = 0;
9969 + spin_unlock(&dli->dl_lock);
9972 + vxlprintk(VXD_CBIT(dlim, 0),
9973 + "FREE (%p,#%d)%c inode",
9974 + sb, tag, __dlimit_char(dli), _file, _line);
9977 +static inline void __dl_adjust_block(struct super_block *sb, tag_t tag,
9978 + unsigned long long *free_blocks, unsigned long long *root_blocks,
9979 + const char *_file, int _line)
9981 + struct dl_info *dli;
9982 + uint64_t broot, bfree;
9984 + dli = locate_dl_info(sb, tag);
9988 + spin_lock(&dli->dl_lock);
9989 + broot = (dli->dl_space_total -
9990 + (dli->dl_space_total >> 10) * dli->dl_nrlmult)
9991 + >> sb->s_blocksize_bits;
9992 + bfree = (dli->dl_space_total - dli->dl_space_used)
9993 + >> sb->s_blocksize_bits;
9994 + spin_unlock(&dli->dl_lock);
9996 + vxlprintk(VXD_CBIT(dlim, 2),
9997 + "ADJUST: %lld,%lld on %lld,%lld [mult=%d]",
9998 + (long long)bfree, (long long)broot,
9999 + *free_blocks, *root_blocks, dli->dl_nrlmult,
10001 + if (free_blocks) {
10002 + if (*free_blocks > bfree)
10003 + *free_blocks = bfree;
10005 + if (root_blocks) {
10006 + if (*root_blocks > broot)
10007 + *root_blocks = broot;
10009 + put_dl_info(dli);
10012 +#define dl_prealloc_space(in, bytes) \
10013 + __dl_alloc_space((in)->i_sb, (in)->i_tag, (dlsize_t)(bytes), \
10014 + __FILE__, __LINE__ )
10016 +#define dl_alloc_space(in, bytes) \
10017 + __dl_alloc_space((in)->i_sb, (in)->i_tag, (dlsize_t)(bytes), \
10018 + __FILE__, __LINE__ )
10020 +#define dl_reserve_space(in, bytes) \
10021 + __dl_alloc_space((in)->i_sb, (in)->i_tag, (dlsize_t)(bytes), \
10022 + __FILE__, __LINE__ )
10024 +#define dl_claim_space(in, bytes) (0)
10026 +#define dl_release_space(in, bytes) \
10027 + __dl_free_space((in)->i_sb, (in)->i_tag, (dlsize_t)(bytes), \
10028 + __FILE__, __LINE__ )
10030 +#define dl_free_space(in, bytes) \
10031 + __dl_free_space((in)->i_sb, (in)->i_tag, (dlsize_t)(bytes), \
10032 + __FILE__, __LINE__ )
10036 +#define dl_alloc_inode(in) \
10037 + __dl_alloc_inode((in)->i_sb, (in)->i_tag, __FILE__, __LINE__ )
10039 +#define dl_free_inode(in) \
10040 + __dl_free_inode((in)->i_sb, (in)->i_tag, __FILE__, __LINE__ )
10043 +#define dl_adjust_block(sb, tag, fb, rb) \
10044 + __dl_adjust_block(sb, tag, fb, rb, __FILE__, __LINE__ )
10048 +#warning duplicate inclusion
10050 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/base.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/base.h
10051 --- linux-2.6.36/include/linux/vserver/base.h 1970-01-01 01:00:00.000000000 +0100
10052 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/base.h 2010-10-21 13:09:36.000000000 +0200
10054 +#ifndef _VX_BASE_H
10055 +#define _VX_BASE_H
10058 +/* context state changes */
10070 +#define vx_task_xid(t) ((t)->xid)
10072 +#define vx_current_xid() vx_task_xid(current)
10074 +#define current_vx_info() (current->vx_info)
10077 +#define nx_task_nid(t) ((t)->nid)
10079 +#define nx_current_nid() nx_task_nid(current)
10081 +#define current_nx_info() (current->nx_info)
10084 +/* generic flag merging */
10086 +#define vs_check_flags(v, m, f) (((v) & (m)) ^ (f))
10088 +#define vs_mask_flags(v, f, m) (((v) & ~(m)) | ((f) & (m)))
10090 +#define vs_mask_mask(v, f, m) (((v) & ~(m)) | ((v) & (f) & (m)))
10092 +#define vs_check_bit(v, n) ((v) & (1LL << (n)))
10095 +/* context flags */
10097 +#define __vx_flags(v) ((v) ? (v)->vx_flags : 0)
10099 +#define vx_current_flags() __vx_flags(current_vx_info())
10101 +#define vx_info_flags(v, m, f) \
10102 + vs_check_flags(__vx_flags(v), m, f)
10104 +#define task_vx_flags(t, m, f) \
10105 + ((t) && vx_info_flags((t)->vx_info, m, f))
10107 +#define vx_flags(m, f) vx_info_flags(current_vx_info(), m, f)
10110 +/* context caps */
10112 +#define __vx_ccaps(v) ((v) ? (v)->vx_ccaps : 0)
10114 +#define vx_current_ccaps() __vx_ccaps(current_vx_info())
10116 +#define vx_info_ccaps(v, c) (__vx_ccaps(v) & (c))
10118 +#define vx_ccaps(c) vx_info_ccaps(current_vx_info(), (c))
10122 +/* network flags */
10124 +#define __nx_flags(n) ((n) ? (n)->nx_flags : 0)
10126 +#define nx_current_flags() __nx_flags(current_nx_info())
10128 +#define nx_info_flags(n, m, f) \
10129 + vs_check_flags(__nx_flags(n), m, f)
10131 +#define task_nx_flags(t, m, f) \
10132 + ((t) && nx_info_flags((t)->nx_info, m, f))
10134 +#define nx_flags(m, f) nx_info_flags(current_nx_info(), m, f)
10137 +/* network caps */
10139 +#define __nx_ncaps(n) ((n) ? (n)->nx_ncaps : 0)
10141 +#define nx_current_ncaps() __nx_ncaps(current_nx_info())
10143 +#define nx_info_ncaps(n, c) (__nx_ncaps(n) & (c))
10145 +#define nx_ncaps(c) nx_info_ncaps(current_nx_info(), c)
10148 +/* context mask capabilities */
10150 +#define __vx_mcaps(v) ((v) ? (v)->vx_ccaps >> 32UL : ~0 )
10152 +#define vx_info_mcaps(v, c) (__vx_mcaps(v) & (c))
10154 +#define vx_mcaps(c) vx_info_mcaps(current_vx_info(), c)
10157 +/* context bcap mask */
10159 +#define __vx_bcaps(v) ((v)->vx_bcaps)
10161 +#define vx_current_bcaps() __vx_bcaps(current_vx_info())
10164 +/* mask given bcaps */
10166 +#define vx_info_mbcaps(v, c) ((v) ? cap_intersect(__vx_bcaps(v), c) : c)
10168 +#define vx_mbcaps(c) vx_info_mbcaps(current_vx_info(), c)
10171 +/* masked cap_bset */
10173 +#define vx_info_cap_bset(v) vx_info_mbcaps(v, current->cap_bset)
10175 +#define vx_current_cap_bset() vx_info_cap_bset(current_vx_info())
10178 +#define vx_info_mbcap(v, b) \
10179 + (!vx_info_flags(v, VXF_STATE_SETUP, 0) ? \
10180 + vx_info_bcaps(v, b) : (b))
10182 +#define task_vx_mbcap(t, b) \
10183 + vx_info_mbcap((t)->vx_info, (t)->b)
10185 +#define vx_mbcap(b) task_vx_mbcap(current, b)
10188 +#define vx_cap_raised(v, c, f) cap_raised(vx_info_mbcaps(v, c), f)
10190 +#define vx_capable(b, c) (capable(b) || \
10191 + (cap_raised(current_cap(), b) && vx_ccaps(c)))
10193 +#define nx_capable(b, c) (capable(b) || \
10194 + (cap_raised(current_cap(), b) && nx_ncaps(c)))
10196 +#define vx_task_initpid(t, n) \
10197 + ((t)->vx_info && \
10198 + ((t)->vx_info->vx_initpid == (n)))
10200 +#define vx_current_initpid(n) vx_task_initpid(current, n)
10203 +/* context unshare mask */
10205 +#define __vx_umask(v) ((v)->vx_umask)
10207 +#define vx_current_umask() __vx_umask(current_vx_info())
10209 +#define vx_can_unshare(b, f) (capable(b) || \
10210 + (cap_raised(current_cap(), b) && \
10211 + !((f) & ~vx_current_umask())))
10214 +#define __vx_state(v) ((v) ? ((v)->vx_state) : 0)
10216 +#define vx_info_state(v, m) (__vx_state(v) & (m))
10219 +#define __nx_state(n) ((n) ? ((n)->nx_state) : 0)
10221 +#define nx_info_state(n, m) (__nx_state(n) & (m))
10224 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/cacct_cmd.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/cacct_cmd.h
10225 --- linux-2.6.36/include/linux/vserver/cacct_cmd.h 1970-01-01 01:00:00.000000000 +0100
10226 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/cacct_cmd.h 2010-10-21 13:09:36.000000000 +0200
10228 +#ifndef _VX_CACCT_CMD_H
10229 +#define _VX_CACCT_CMD_H
10232 +/* virtual host info name commands */
10234 +#define VCMD_sock_stat VC_CMD(VSTAT, 5, 0)
10236 +struct vcmd_sock_stat_v0 {
10238 + uint32_t count[3];
10239 + uint64_t total[3];
10245 +#include <linux/compiler.h>
10247 +extern int vc_sock_stat(struct vx_info *, void __user *);
10249 +#endif /* __KERNEL__ */
10250 +#endif /* _VX_CACCT_CMD_H */
10251 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/cacct_def.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/cacct_def.h
10252 --- linux-2.6.36/include/linux/vserver/cacct_def.h 1970-01-01 01:00:00.000000000 +0100
10253 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/cacct_def.h 2010-10-21 13:09:36.000000000 +0200
10255 +#ifndef _VX_CACCT_DEF_H
10256 +#define _VX_CACCT_DEF_H
10258 +#include <asm/atomic.h>
10259 +#include <linux/vserver/cacct.h>
10262 +struct _vx_sock_acc {
10263 + atomic_long_t count;
10264 + atomic_long_t total;
10267 +/* context sub struct */
10269 +struct _vx_cacct {
10270 + struct _vx_sock_acc sock[VXA_SOCK_SIZE][3];
10271 + atomic_t slab[8];
10272 + atomic_t page[6][8];
10275 +#ifdef CONFIG_VSERVER_DEBUG
10277 +static inline void __dump_vx_cacct(struct _vx_cacct *cacct)
10281 + printk("\t_vx_cacct:");
10282 + for (i = 0; i < 6; i++) {
10283 + struct _vx_sock_acc *ptr = cacct->sock[i];
10285 + printk("\t [%d] =", i);
10286 + for (j = 0; j < 3; j++) {
10287 + printk(" [%d] = %8lu, %8lu", j,
10288 + atomic_long_read(&ptr[j].count),
10289 + atomic_long_read(&ptr[j].total));
10297 +#endif /* _VX_CACCT_DEF_H */
10298 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/cacct.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/cacct.h
10299 --- linux-2.6.36/include/linux/vserver/cacct.h 1970-01-01 01:00:00.000000000 +0100
10300 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/cacct.h 2010-10-21 13:09:36.000000000 +0200
10302 +#ifndef _VX_CACCT_H
10303 +#define _VX_CACCT_H
10306 +enum sock_acc_field {
10307 + VXA_SOCK_UNSPEC = 0,
10313 + VXA_SOCK_SIZE /* array size */
10316 +#endif /* _VX_CACCT_H */
10317 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/cacct_int.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/cacct_int.h
10318 --- linux-2.6.36/include/linux/vserver/cacct_int.h 1970-01-01 01:00:00.000000000 +0100
10319 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/cacct_int.h 2010-10-21 13:09:36.000000000 +0200
10321 +#ifndef _VX_CACCT_INT_H
10322 +#define _VX_CACCT_INT_H
10328 +unsigned long vx_sock_count(struct _vx_cacct *cacct, int type, int pos)
10330 + return atomic_long_read(&cacct->sock[type][pos].count);
10335 +unsigned long vx_sock_total(struct _vx_cacct *cacct, int type, int pos)
10337 + return atomic_long_read(&cacct->sock[type][pos].total);
10340 +#endif /* __KERNEL__ */
10341 +#endif /* _VX_CACCT_INT_H */
10342 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/check.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/check.h
10343 --- linux-2.6.36/include/linux/vserver/check.h 1970-01-01 01:00:00.000000000 +0100
10344 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/check.h 2010-10-21 13:09:36.000000000 +0200
10346 +#ifndef _VS_CHECK_H
10347 +#define _VS_CHECK_H
10350 +#define MAX_S_CONTEXT 65535 /* Arbitrary limit */
10352 +#ifdef CONFIG_VSERVER_DYNAMIC_IDS
10353 +#define MIN_D_CONTEXT 49152 /* dynamic contexts start here */
10355 +#define MIN_D_CONTEXT 65536
10358 +/* check conditions */
10360 +#define VS_ADMIN 0x0001
10361 +#define VS_WATCH 0x0002
10362 +#define VS_HIDE 0x0004
10363 +#define VS_HOSTID 0x0008
10365 +#define VS_IDENT 0x0010
10366 +#define VS_EQUIV 0x0020
10367 +#define VS_PARENT 0x0040
10368 +#define VS_CHILD 0x0080
10370 +#define VS_ARG_MASK 0x00F0
10372 +#define VS_DYNAMIC 0x0100
10373 +#define VS_STATIC 0x0200
10375 +#define VS_ATR_MASK 0x0F00
10377 +#ifdef CONFIG_VSERVER_PRIVACY
10378 +#define VS_ADMIN_P (0)
10379 +#define VS_WATCH_P (0)
10381 +#define VS_ADMIN_P VS_ADMIN
10382 +#define VS_WATCH_P VS_WATCH
10385 +#define VS_HARDIRQ 0x1000
10386 +#define VS_SOFTIRQ 0x2000
10387 +#define VS_IRQ 0x4000
10389 +#define VS_IRQ_MASK 0xF000
10391 +#include <linux/hardirq.h>
10394 + * check current context for ADMIN/WATCH and
10395 + * optionally against supplied argument
10397 +static inline int __vs_check(int cid, int id, unsigned int mode)
10399 + if (mode & VS_ARG_MASK) {
10400 + if ((mode & VS_IDENT) && (id == cid))
10403 + if (mode & VS_ATR_MASK) {
10404 + if ((mode & VS_DYNAMIC) &&
10405 + (id >= MIN_D_CONTEXT) &&
10406 + (id <= MAX_S_CONTEXT))
10408 + if ((mode & VS_STATIC) &&
10409 + (id > 1) && (id < MIN_D_CONTEXT))
10412 + if (mode & VS_IRQ_MASK) {
10413 + if ((mode & VS_IRQ) && unlikely(in_interrupt()))
10415 + if ((mode & VS_HARDIRQ) && unlikely(in_irq()))
10417 + if ((mode & VS_SOFTIRQ) && unlikely(in_softirq()))
10420 + return (((mode & VS_ADMIN) && (cid == 0)) ||
10421 + ((mode & VS_WATCH) && (cid == 1)) ||
10422 + ((mode & VS_HOSTID) && (id == 0)));
10425 +#define vx_check(c, m) __vs_check(vx_current_xid(), c, (m) | VS_IRQ)
10427 +#define vx_weak_check(c, m) ((m) ? vx_check(c, m) : 1)
10430 +#define nx_check(c, m) __vs_check(nx_current_nid(), c, m)
10432 +#define nx_weak_check(c, m) ((m) ? nx_check(c, m) : 1)
10435 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/context_cmd.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/context_cmd.h
10436 --- linux-2.6.36/include/linux/vserver/context_cmd.h 1970-01-01 01:00:00.000000000 +0100
10437 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/context_cmd.h 2010-10-21 13:09:36.000000000 +0200
10439 +#ifndef _VX_CONTEXT_CMD_H
10440 +#define _VX_CONTEXT_CMD_H
10443 +/* vinfo commands */
10445 +#define VCMD_task_xid VC_CMD(VINFO, 1, 0)
10448 +extern int vc_task_xid(uint32_t);
10450 +#endif /* __KERNEL__ */
10452 +#define VCMD_vx_info VC_CMD(VINFO, 5, 0)
10454 +struct vcmd_vx_info_v0 {
10456 + uint32_t initpid;
10457 + /* more to come */
10461 +extern int vc_vx_info(struct vx_info *, void __user *);
10463 +#endif /* __KERNEL__ */
10465 +#define VCMD_ctx_stat VC_CMD(VSTAT, 0, 0)
10467 +struct vcmd_ctx_stat_v0 {
10470 + /* more to come */
10474 +extern int vc_ctx_stat(struct vx_info *, void __user *);
10476 +#endif /* __KERNEL__ */
10478 +/* context commands */
10480 +#define VCMD_ctx_create_v0 VC_CMD(VPROC, 1, 0)
10481 +#define VCMD_ctx_create VC_CMD(VPROC, 1, 1)
10483 +struct vcmd_ctx_create {
10484 + uint64_t flagword;
10487 +#define VCMD_ctx_migrate_v0 VC_CMD(PROCMIG, 1, 0)
10488 +#define VCMD_ctx_migrate VC_CMD(PROCMIG, 1, 1)
10490 +struct vcmd_ctx_migrate {
10491 + uint64_t flagword;
10495 +extern int vc_ctx_create(uint32_t, void __user *);
10496 +extern int vc_ctx_migrate(struct vx_info *, void __user *);
10498 +#endif /* __KERNEL__ */
10501 +/* flag commands */
10503 +#define VCMD_get_cflags VC_CMD(FLAGS, 1, 0)
10504 +#define VCMD_set_cflags VC_CMD(FLAGS, 2, 0)
10506 +struct vcmd_ctx_flags_v0 {
10507 + uint64_t flagword;
10512 +extern int vc_get_cflags(struct vx_info *, void __user *);
10513 +extern int vc_set_cflags(struct vx_info *, void __user *);
10515 +#endif /* __KERNEL__ */
10518 +/* context caps commands */
10520 +#define VCMD_get_ccaps VC_CMD(FLAGS, 3, 1)
10521 +#define VCMD_set_ccaps VC_CMD(FLAGS, 4, 1)
10523 +struct vcmd_ctx_caps_v1 {
10529 +extern int vc_get_ccaps(struct vx_info *, void __user *);
10530 +extern int vc_set_ccaps(struct vx_info *, void __user *);
10532 +#endif /* __KERNEL__ */
10535 +/* bcaps commands */
10537 +#define VCMD_get_bcaps VC_CMD(FLAGS, 9, 0)
10538 +#define VCMD_set_bcaps VC_CMD(FLAGS, 10, 0)
10540 +struct vcmd_bcaps {
10546 +extern int vc_get_bcaps(struct vx_info *, void __user *);
10547 +extern int vc_set_bcaps(struct vx_info *, void __user *);
10549 +#endif /* __KERNEL__ */
10552 +/* umask commands */
10554 +#define VCMD_get_umask VC_CMD(FLAGS, 13, 0)
10555 +#define VCMD_set_umask VC_CMD(FLAGS, 14, 0)
10557 +struct vcmd_umask {
10563 +extern int vc_get_umask(struct vx_info *, void __user *);
10564 +extern int vc_set_umask(struct vx_info *, void __user *);
10566 +#endif /* __KERNEL__ */
10571 +#define VCMD_get_badness VC_CMD(MEMCTRL, 5, 0)
10572 +#define VCMD_set_badness VC_CMD(MEMCTRL, 6, 0)
10574 +struct vcmd_badness_v0 {
10579 +extern int vc_get_badness(struct vx_info *, void __user *);
10580 +extern int vc_set_badness(struct vx_info *, void __user *);
10582 +#endif /* __KERNEL__ */
10583 +#endif /* _VX_CONTEXT_CMD_H */
10584 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/context.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/context.h
10585 --- linux-2.6.36/include/linux/vserver/context.h 1970-01-01 01:00:00.000000000 +0100
10586 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/context.h 2010-10-21 13:09:36.000000000 +0200
10588 +#ifndef _VX_CONTEXT_H
10589 +#define _VX_CONTEXT_H
10591 +#include <linux/types.h>
10592 +#include <linux/capability.h>
10595 +/* context flags */
10597 +#define VXF_INFO_SCHED 0x00000002
10598 +#define VXF_INFO_NPROC 0x00000004
10599 +#define VXF_INFO_PRIVATE 0x00000008
10601 +#define VXF_INFO_INIT 0x00000010
10602 +#define VXF_INFO_HIDE 0x00000020
10603 +#define VXF_INFO_ULIMIT 0x00000040
10604 +#define VXF_INFO_NSPACE 0x00000080
10606 +#define VXF_SCHED_HARD 0x00000100
10607 +#define VXF_SCHED_PRIO 0x00000200
10608 +#define VXF_SCHED_PAUSE 0x00000400
10610 +#define VXF_VIRT_MEM 0x00010000
10611 +#define VXF_VIRT_UPTIME 0x00020000
10612 +#define VXF_VIRT_CPU 0x00040000
10613 +#define VXF_VIRT_LOAD 0x00080000
10614 +#define VXF_VIRT_TIME 0x00100000
10616 +#define VXF_HIDE_MOUNT 0x01000000
10617 +/* was VXF_HIDE_NETIF 0x02000000 */
10618 +#define VXF_HIDE_VINFO 0x04000000
10620 +#define VXF_STATE_SETUP (1ULL << 32)
10621 +#define VXF_STATE_INIT (1ULL << 33)
10622 +#define VXF_STATE_ADMIN (1ULL << 34)
10624 +#define VXF_SC_HELPER (1ULL << 36)
10625 +#define VXF_REBOOT_KILL (1ULL << 37)
10626 +#define VXF_PERSISTENT (1ULL << 38)
10628 +#define VXF_FORK_RSS (1ULL << 48)
10629 +#define VXF_PROLIFIC (1ULL << 49)
10631 +#define VXF_IGNEG_NICE (1ULL << 52)
10633 +#define VXF_ONE_TIME (0x0007ULL << 32)
10635 +#define VXF_INIT_SET (VXF_STATE_SETUP | VXF_STATE_INIT | VXF_STATE_ADMIN)
10638 +/* context migration */
10640 +#define VXM_SET_INIT 0x00000001
10641 +#define VXM_SET_REAPER 0x00000002
10643 +/* context caps */
10645 +#define VXC_CAP_MASK 0x00000000
10647 +#define VXC_SET_UTSNAME 0x00000001
10648 +#define VXC_SET_RLIMIT 0x00000002
10649 +#define VXC_FS_SECURITY 0x00000004
10650 +#define VXC_FS_TRUSTED 0x00000008
10651 +#define VXC_TIOCSTI 0x00000010
10653 +/* was VXC_RAW_ICMP 0x00000100 */
10654 +#define VXC_SYSLOG 0x00001000
10655 +#define VXC_OOM_ADJUST 0x00002000
10656 +#define VXC_AUDIT_CONTROL 0x00004000
10658 +#define VXC_SECURE_MOUNT 0x00010000
10659 +#define VXC_SECURE_REMOUNT 0x00020000
10660 +#define VXC_BINARY_MOUNT 0x00040000
10662 +#define VXC_QUOTA_CTL 0x00100000
10663 +#define VXC_ADMIN_MAPPER 0x00200000
10664 +#define VXC_ADMIN_CLOOP 0x00400000
10666 +#define VXC_KTHREAD 0x01000000
10667 +#define VXC_NAMESPACE 0x02000000
10672 +#include <linux/list.h>
10673 +#include <linux/spinlock.h>
10674 +#include <linux/rcupdate.h>
10676 +#include "limit_def.h"
10677 +#include "sched_def.h"
10678 +#include "cvirt_def.h"
10679 +#include "cacct_def.h"
10680 +#include "device_def.h"
10682 +#define VX_SPACES 2
10684 +struct _vx_info_pc {
10685 + struct _vx_sched_pc sched_pc;
10686 + struct _vx_cvirt_pc cvirt_pc;
10690 + struct hlist_node vx_hlist; /* linked list of contexts */
10691 + xid_t vx_id; /* context id */
10692 + atomic_t vx_usecnt; /* usage count */
10693 + atomic_t vx_tasks; /* tasks count */
10694 + struct vx_info *vx_parent; /* parent context */
10695 + int vx_state; /* context state */
10697 + unsigned long vx_nsmask[VX_SPACES]; /* assignment mask */
10698 + struct nsproxy *vx_nsproxy[VX_SPACES]; /* private namespaces */
10699 + struct fs_struct *vx_fs[VX_SPACES]; /* private namespace fs */
10701 + uint64_t vx_flags; /* context flags */
10702 + uint64_t vx_ccaps; /* context caps (vserver) */
10703 + kernel_cap_t vx_bcaps; /* bounding caps (system) */
10704 + unsigned long vx_umask; /* unshare mask (guest) */
10706 + struct task_struct *vx_reaper; /* guest reaper process */
10707 + pid_t vx_initpid; /* PID of guest init */
10708 + int64_t vx_badness_bias; /* OOM points bias */
10710 + struct _vx_limit limit; /* vserver limits */
10711 + struct _vx_sched sched; /* vserver scheduler */
10712 + struct _vx_cvirt cvirt; /* virtual/bias stuff */
10713 + struct _vx_cacct cacct; /* context accounting */
10715 + struct _vx_device dmap; /* default device map targets */
10717 +#ifndef CONFIG_SMP
10718 + struct _vx_info_pc info_pc; /* per cpu data */
10720 + struct _vx_info_pc *ptr_pc; /* per cpu array */
10723 + wait_queue_head_t vx_wait; /* context exit waitqueue */
10724 + int reboot_cmd; /* last sys_reboot() cmd */
10725 + int exit_code; /* last process exit code */
10727 + char vx_name[65]; /* vserver name */
10730 +#ifndef CONFIG_SMP
10731 +#define vx_ptr_pc(vxi) (&(vxi)->info_pc)
10732 +#define vx_per_cpu(vxi, v, id) vx_ptr_pc(vxi)->v
10734 +#define vx_ptr_pc(vxi) ((vxi)->ptr_pc)
10735 +#define vx_per_cpu(vxi, v, id) per_cpu_ptr(vx_ptr_pc(vxi), id)->v
10738 +#define vx_cpu(vxi, v) vx_per_cpu(vxi, v, smp_processor_id())
10741 +struct vx_info_save {
10742 + struct vx_info *vxi;
10747 +/* status flags */
10749 +#define VXS_HASHED 0x0001
10750 +#define VXS_PAUSED 0x0010
10751 +#define VXS_SHUTDOWN 0x0100
10752 +#define VXS_HELPER 0x1000
10753 +#define VXS_RELEASED 0x8000
10756 +extern void claim_vx_info(struct vx_info *, struct task_struct *);
10757 +extern void release_vx_info(struct vx_info *, struct task_struct *);
10759 +extern struct vx_info *lookup_vx_info(int);
10760 +extern struct vx_info *lookup_or_create_vx_info(int);
10762 +extern int get_xid_list(int, unsigned int *, int);
10763 +extern int xid_is_hashed(xid_t);
10765 +extern int vx_migrate_task(struct task_struct *, struct vx_info *, int);
10767 +extern long vs_state_change(struct vx_info *, unsigned int);
10770 +#endif /* __KERNEL__ */
10771 +#endif /* _VX_CONTEXT_H */
10772 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/cvirt_cmd.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/cvirt_cmd.h
10773 --- linux-2.6.36/include/linux/vserver/cvirt_cmd.h 1970-01-01 01:00:00.000000000 +0100
10774 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/cvirt_cmd.h 2010-10-21 13:09:36.000000000 +0200
10776 +#ifndef _VX_CVIRT_CMD_H
10777 +#define _VX_CVIRT_CMD_H
10780 +/* virtual host info name commands */
10782 +#define VCMD_set_vhi_name VC_CMD(VHOST, 1, 0)
10783 +#define VCMD_get_vhi_name VC_CMD(VHOST, 2, 0)
10785 +struct vcmd_vhi_name_v0 {
10791 +enum vhi_name_field {
10792 + VHIN_CONTEXT = 0,
10804 +#include <linux/compiler.h>
10806 +extern int vc_set_vhi_name(struct vx_info *, void __user *);
10807 +extern int vc_get_vhi_name(struct vx_info *, void __user *);
10809 +#endif /* __KERNEL__ */
10811 +#define VCMD_virt_stat VC_CMD(VSTAT, 3, 0)
10813 +struct vcmd_virt_stat_v0 {
10816 + uint32_t nr_threads;
10817 + uint32_t nr_running;
10818 + uint32_t nr_uninterruptible;
10819 + uint32_t nr_onhold;
10820 + uint32_t nr_forks;
10821 + uint32_t load[3];
10825 +extern int vc_virt_stat(struct vx_info *, void __user *);
10827 +#endif /* __KERNEL__ */
10828 +#endif /* _VX_CVIRT_CMD_H */
10829 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/cvirt_def.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/cvirt_def.h
10830 --- linux-2.6.36/include/linux/vserver/cvirt_def.h 1970-01-01 01:00:00.000000000 +0100
10831 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/cvirt_def.h 2010-10-21 14:26:47.000000000 +0200
10833 +#ifndef _VX_CVIRT_DEF_H
10834 +#define _VX_CVIRT_DEF_H
10836 +#include <linux/jiffies.h>
10837 +#include <linux/spinlock.h>
10838 +#include <linux/wait.h>
10839 +#include <linux/time.h>
10840 +#include <asm/atomic.h>
10843 +struct _vx_usage_stat {
10847 + uint64_t softirq;
10853 +struct _vx_syslog {
10854 + wait_queue_head_t log_wait;
10855 + spinlock_t logbuf_lock; /* lock for the log buffer */
10857 + unsigned long log_start; /* next char to be read by syslog() */
10858 + unsigned long con_start; /* next char to be sent to consoles */
10859 + unsigned long log_end; /* most-recently-written-char + 1 */
10860 + unsigned long logged_chars; /* #chars since last read+clear operation */
10862 + char log_buf[1024];
10866 +/* context sub struct */
10868 +struct _vx_cvirt {
10869 + atomic_t nr_threads; /* number of current threads */
10870 + atomic_t nr_running; /* number of running threads */
10871 + atomic_t nr_uninterruptible; /* number of uninterruptible threads */
10873 + atomic_t nr_onhold; /* processes on hold */
10874 + uint32_t onhold_last; /* jiffies when put on hold */
10876 + struct timespec bias_ts; /* time offset to the host */
10877 + struct timespec bias_idle;
10878 + struct timespec bias_uptime; /* context creation point */
10879 + uint64_t bias_clock; /* offset in clock_t */
10881 + spinlock_t load_lock; /* lock for the load averages */
10882 + atomic_t load_updates; /* nr of load updates done so far */
10883 + uint32_t load_last; /* last time load was calculated */
10884 + uint32_t load[3]; /* load averages 1,5,15 */
10886 + atomic_t total_forks; /* number of forks so far */
10888 + struct _vx_syslog syslog;
10891 +struct _vx_cvirt_pc {
10892 + struct _vx_usage_stat cpustat;
10896 +#ifdef CONFIG_VSERVER_DEBUG
10898 +static inline void __dump_vx_cvirt(struct _vx_cvirt *cvirt)
10900 + printk("\t_vx_cvirt:\n");
10901 + printk("\t threads: %4d, %4d, %4d, %4d\n",
10902 + atomic_read(&cvirt->nr_threads),
10903 + atomic_read(&cvirt->nr_running),
10904 + atomic_read(&cvirt->nr_uninterruptible),
10905 + atomic_read(&cvirt->nr_onhold));
10906 + /* add rest here */
10907 + printk("\t total_forks = %d\n", atomic_read(&cvirt->total_forks));
10912 +#endif /* _VX_CVIRT_DEF_H */
10913 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/cvirt.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/cvirt.h
10914 --- linux-2.6.36/include/linux/vserver/cvirt.h 1970-01-01 01:00:00.000000000 +0100
10915 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/cvirt.h 2010-10-21 13:09:36.000000000 +0200
10917 +#ifndef _VX_CVIRT_H
10918 +#define _VX_CVIRT_H
10925 +void vx_vsi_uptime(struct timespec *, struct timespec *);
10930 +void vx_update_load(struct vx_info *);
10933 +int vx_do_syslog(int, char __user *, int);
10935 +#endif /* __KERNEL__ */
10936 +#endif /* _VX_CVIRT_H */
10937 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/debug_cmd.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/debug_cmd.h
10938 --- linux-2.6.36/include/linux/vserver/debug_cmd.h 1970-01-01 01:00:00.000000000 +0100
10939 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/debug_cmd.h 2010-10-21 13:09:36.000000000 +0200
10941 +#ifndef _VX_DEBUG_CMD_H
10942 +#define _VX_DEBUG_CMD_H
10945 +/* debug commands */
10947 +#define VCMD_dump_history VC_CMD(DEBUG, 1, 0)
10949 +#define VCMD_read_history VC_CMD(DEBUG, 5, 0)
10950 +#define VCMD_read_monitor VC_CMD(DEBUG, 6, 0)
10952 +struct vcmd_read_history_v0 {
10955 + char __user *data;
10958 +struct vcmd_read_monitor_v0 {
10961 + char __user *data;
10967 +#ifdef CONFIG_COMPAT
10969 +#include <asm/compat.h>
10971 +struct vcmd_read_history_v0_x32 {
10974 + compat_uptr_t data_ptr;
10977 +struct vcmd_read_monitor_v0_x32 {
10980 + compat_uptr_t data_ptr;
10983 +#endif /* CONFIG_COMPAT */
10985 +extern int vc_dump_history(uint32_t);
10987 +extern int vc_read_history(uint32_t, void __user *);
10988 +extern int vc_read_monitor(uint32_t, void __user *);
10990 +#ifdef CONFIG_COMPAT
10992 +extern int vc_read_history_x32(uint32_t, void __user *);
10993 +extern int vc_read_monitor_x32(uint32_t, void __user *);
10995 +#endif /* CONFIG_COMPAT */
10997 +#endif /* __KERNEL__ */
10998 +#endif /* _VX_DEBUG_CMD_H */
10999 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/debug.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/debug.h
11000 --- linux-2.6.36/include/linux/vserver/debug.h 1970-01-01 01:00:00.000000000 +0100
11001 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/debug.h 2010-10-21 13:09:36.000000000 +0200
11003 +#ifndef _VX_DEBUG_H
11004 +#define _VX_DEBUG_H
11007 +#define VXD_CBIT(n, m) (vx_debug_ ## n & (1 << (m)))
11008 +#define VXD_CMIN(n, m) (vx_debug_ ## n > (m))
11009 +#define VXD_MASK(n, m) (vx_debug_ ## n & (m))
11011 +#define VXD_DEV(d) (d), (d)->bd_inode->i_ino, \
11012 + imajor((d)->bd_inode), iminor((d)->bd_inode)
11013 +#define VXF_DEV "%p[%lu,%d:%d]"
11016 +#define vxd_path(p) \
11017 + ({ static char _buffer[PATH_MAX]; \
11018 + d_path(p, _buffer, sizeof(_buffer)); })
11020 +#define vxd_cond_path(n) \
11021 + ((n) ? vxd_path(&(n)->path) : "<null>" )
11024 +#ifdef CONFIG_VSERVER_DEBUG
11026 +extern unsigned int vx_debug_switch;
11027 +extern unsigned int vx_debug_xid;
11028 +extern unsigned int vx_debug_nid;
11029 +extern unsigned int vx_debug_tag;
11030 +extern unsigned int vx_debug_net;
11031 +extern unsigned int vx_debug_limit;
11032 +extern unsigned int vx_debug_cres;
11033 +extern unsigned int vx_debug_dlim;
11034 +extern unsigned int vx_debug_quota;
11035 +extern unsigned int vx_debug_cvirt;
11036 +extern unsigned int vx_debug_space;
11037 +extern unsigned int vx_debug_misc;
11040 +#define VX_LOGLEVEL "vxD: "
11041 +#define VX_PROC_FMT "%p: "
11042 +#define VX_PROCESS current
11044 +#define vxdprintk(c, f, x...) \
11047 + printk(VX_LOGLEVEL VX_PROC_FMT f "\n", \
11048 + VX_PROCESS , ##x); \
11051 +#define vxlprintk(c, f, x...) \
11054 + printk(VX_LOGLEVEL f " @%s:%d\n", x); \
11057 +#define vxfprintk(c, f, x...) \
11060 + printk(VX_LOGLEVEL f " %s@%s:%d\n", x); \
11066 +void dump_vx_info(struct vx_info *, int);
11067 +void dump_vx_info_inactive(int);
11069 +#else /* CONFIG_VSERVER_DEBUG */
11071 +#define vx_debug_switch 0
11072 +#define vx_debug_xid 0
11073 +#define vx_debug_nid 0
11074 +#define vx_debug_tag 0
11075 +#define vx_debug_net 0
11076 +#define vx_debug_limit 0
11077 +#define vx_debug_cres 0
11078 +#define vx_debug_dlim 0
11079 +#define vx_debug_cvirt 0
11081 +#define vxdprintk(x...) do { } while (0)
11082 +#define vxlprintk(x...) do { } while (0)
11083 +#define vxfprintk(x...) do { } while (0)
11085 +#endif /* CONFIG_VSERVER_DEBUG */
11088 +#ifdef CONFIG_VSERVER_WARN
11090 +#define VX_WARNLEVEL KERN_WARNING "vxW: "
11091 +#define VX_WARN_TASK "[»%s«,%u:#%u|%u|%u] "
11092 +#define VX_WARN_XID "[xid #%u] "
11093 +#define VX_WARN_NID "[nid #%u] "
11094 +#define VX_WARN_TAG "[tag #%u] "
11096 +#define vxwprintk(c, f, x...) \
11099 + printk(VX_WARNLEVEL f "\n", ##x); \
11102 +#else /* CONFIG_VSERVER_WARN */
11104 +#define vxwprintk(x...) do { } while (0)
11106 +#endif /* CONFIG_VSERVER_WARN */
11108 +#define vxwprintk_task(c, f, x...) \
11109 + vxwprintk(c, VX_WARN_TASK f, \
11110 + current->comm, current->pid, \
11111 + current->xid, current->nid, current->tag, ##x)
11112 +#define vxwprintk_xid(c, f, x...) \
11113 + vxwprintk(c, VX_WARN_XID f, current->xid, x)
11114 +#define vxwprintk_nid(c, f, x...) \
11115 + vxwprintk(c, VX_WARN_NID f, current->nid, x)
11116 +#define vxwprintk_tag(c, f, x...) \
11117 + vxwprintk(c, VX_WARN_TAG f, current->tag, x)
11119 +#ifdef CONFIG_VSERVER_DEBUG
11120 +#define vxd_assert_lock(l) assert_spin_locked(l)
11121 +#define vxd_assert(c, f, x...) vxlprintk(!(c), \
11122 + "assertion [" f "] failed.", ##x, __FILE__, __LINE__)
11124 +#define vxd_assert_lock(l) do { } while (0)
11125 +#define vxd_assert(c, f, x...) do { } while (0)
11129 +#endif /* _VX_DEBUG_H */
11130 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/device_cmd.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/device_cmd.h
11131 --- linux-2.6.36/include/linux/vserver/device_cmd.h 1970-01-01 01:00:00.000000000 +0100
11132 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/device_cmd.h 2010-10-21 13:09:36.000000000 +0200
11134 +#ifndef _VX_DEVICE_CMD_H
11135 +#define _VX_DEVICE_CMD_H
11138 +/* device vserver commands */
11140 +#define VCMD_set_mapping VC_CMD(DEVICE, 1, 0)
11141 +#define VCMD_unset_mapping VC_CMD(DEVICE, 2, 0)
11143 +struct vcmd_set_mapping_v0 {
11144 + const char __user *device;
11145 + const char __user *target;
11152 +#ifdef CONFIG_COMPAT
11154 +#include <asm/compat.h>
11156 +struct vcmd_set_mapping_v0_x32 {
11157 + compat_uptr_t device_ptr;
11158 + compat_uptr_t target_ptr;
11162 +#endif /* CONFIG_COMPAT */
11164 +#include <linux/compiler.h>
11166 +extern int vc_set_mapping(struct vx_info *, void __user *);
11167 +extern int vc_unset_mapping(struct vx_info *, void __user *);
11169 +#ifdef CONFIG_COMPAT
11171 +extern int vc_set_mapping_x32(struct vx_info *, void __user *);
11172 +extern int vc_unset_mapping_x32(struct vx_info *, void __user *);
11174 +#endif /* CONFIG_COMPAT */
11176 +#endif /* __KERNEL__ */
11177 +#endif /* _VX_DEVICE_CMD_H */
11178 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/device_def.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/device_def.h
11179 --- linux-2.6.36/include/linux/vserver/device_def.h 1970-01-01 01:00:00.000000000 +0100
11180 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/device_def.h 2010-10-21 13:09:36.000000000 +0200
11182 +#ifndef _VX_DEVICE_DEF_H
11183 +#define _VX_DEVICE_DEF_H
11185 +#include <linux/types.h>
11187 +struct vx_dmap_target {
11192 +struct _vx_device {
11193 +#ifdef CONFIG_VSERVER_DEVICE
11194 + struct vx_dmap_target targets[2];
11198 +#endif /* _VX_DEVICE_DEF_H */
11199 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/device.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/device.h
11200 --- linux-2.6.36/include/linux/vserver/device.h 1970-01-01 01:00:00.000000000 +0100
11201 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/device.h 2010-10-21 13:09:36.000000000 +0200
11203 +#ifndef _VX_DEVICE_H
11204 +#define _VX_DEVICE_H
11207 +#define DATTR_CREATE 0x00000001
11208 +#define DATTR_OPEN 0x00000002
11210 +#define DATTR_REMAP 0x00000010
11212 +#define DATTR_MASK 0x00000013
11215 +#else /* _VX_DEVICE_H */
11216 +#warning duplicate inclusion
11217 +#endif /* _VX_DEVICE_H */
11218 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/dlimit_cmd.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/dlimit_cmd.h
11219 --- linux-2.6.36/include/linux/vserver/dlimit_cmd.h 1970-01-01 01:00:00.000000000 +0100
11220 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/dlimit_cmd.h 2010-10-21 13:09:36.000000000 +0200
11222 +#ifndef _VX_DLIMIT_CMD_H
11223 +#define _VX_DLIMIT_CMD_H
11226 +/* dlimit vserver commands */
11228 +#define VCMD_add_dlimit VC_CMD(DLIMIT, 1, 0)
11229 +#define VCMD_rem_dlimit VC_CMD(DLIMIT, 2, 0)
11231 +#define VCMD_set_dlimit VC_CMD(DLIMIT, 5, 0)
11232 +#define VCMD_get_dlimit VC_CMD(DLIMIT, 6, 0)
11234 +struct vcmd_ctx_dlimit_base_v0 {
11235 + const char __user *name;
11239 +struct vcmd_ctx_dlimit_v0 {
11240 + const char __user *name;
11241 + uint32_t space_used; /* used space in kbytes */
11242 + uint32_t space_total; /* maximum space in kbytes */
11243 + uint32_t inodes_used; /* used inodes */
11244 + uint32_t inodes_total; /* maximum inodes */
11245 + uint32_t reserved; /* reserved for root in % */
11249 +#define CDLIM_UNSET ((uint32_t)0UL)
11250 +#define CDLIM_INFINITY ((uint32_t)~0UL)
11251 +#define CDLIM_KEEP ((uint32_t)~1UL)
11253 +#define DLIME_UNIT 0
11254 +#define DLIME_KILO 1
11255 +#define DLIME_MEGA 2
11256 +#define DLIME_GIGA 3
11258 +#define DLIMF_SHIFT 0x10
11260 +#define DLIMS_USED 0
11261 +#define DLIMS_TOTAL 2
11264 +uint64_t dlimit_space_32to64(uint32_t val, uint32_t flags, int shift)
11266 + int exp = (flags & DLIMF_SHIFT) ?
11267 + (flags >> shift) & DLIME_GIGA : DLIME_KILO;
11268 + return ((uint64_t)val) << (10 * exp);
11272 +uint32_t dlimit_space_64to32(uint64_t val, uint32_t *flags, int shift)
11276 + if (*flags & DLIMF_SHIFT) {
11277 + while (val > (1LL << 32) && (exp < 3)) {
11281 + *flags &= ~(DLIME_GIGA << shift);
11282 + *flags |= exp << shift;
11290 +#ifdef CONFIG_COMPAT
11292 +#include <asm/compat.h>
11294 +struct vcmd_ctx_dlimit_base_v0_x32 {
11295 + compat_uptr_t name_ptr;
11299 +struct vcmd_ctx_dlimit_v0_x32 {
11300 + compat_uptr_t name_ptr;
11301 + uint32_t space_used; /* used space in kbytes */
11302 + uint32_t space_total; /* maximum space in kbytes */
11303 + uint32_t inodes_used; /* used inodes */
11304 + uint32_t inodes_total; /* maximum inodes */
11305 + uint32_t reserved; /* reserved for root in % */
11309 +#endif /* CONFIG_COMPAT */
11311 +#include <linux/compiler.h>
11313 +extern int vc_add_dlimit(uint32_t, void __user *);
11314 +extern int vc_rem_dlimit(uint32_t, void __user *);
11316 +extern int vc_set_dlimit(uint32_t, void __user *);
11317 +extern int vc_get_dlimit(uint32_t, void __user *);
11319 +#ifdef CONFIG_COMPAT
11321 +extern int vc_add_dlimit_x32(uint32_t, void __user *);
11322 +extern int vc_rem_dlimit_x32(uint32_t, void __user *);
11324 +extern int vc_set_dlimit_x32(uint32_t, void __user *);
11325 +extern int vc_get_dlimit_x32(uint32_t, void __user *);
11327 +#endif /* CONFIG_COMPAT */
11329 +#endif /* __KERNEL__ */
11330 +#endif /* _VX_DLIMIT_CMD_H */
11331 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/dlimit.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/dlimit.h
11332 --- linux-2.6.36/include/linux/vserver/dlimit.h 1970-01-01 01:00:00.000000000 +0100
11333 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/dlimit.h 2010-10-21 13:09:36.000000000 +0200
11335 +#ifndef _VX_DLIMIT_H
11336 +#define _VX_DLIMIT_H
11338 +#include "switch.h"
11343 +/* keep in sync with CDLIM_INFINITY */
11345 +#define DLIM_INFINITY (~0ULL)
11347 +#include <linux/spinlock.h>
11348 +#include <linux/rcupdate.h>
11350 +struct super_block;
11353 + struct hlist_node dl_hlist; /* linked list of contexts */
11354 + struct rcu_head dl_rcu; /* the rcu head */
11355 + tag_t dl_tag; /* context tag */
11356 + atomic_t dl_usecnt; /* usage count */
11357 + atomic_t dl_refcnt; /* reference count */
11359 + struct super_block *dl_sb; /* associated superblock */
11361 + spinlock_t dl_lock; /* protect the values */
11363 + unsigned long long dl_space_used; /* used space in bytes */
11364 + unsigned long long dl_space_total; /* maximum space in bytes */
11365 + unsigned long dl_inodes_used; /* used inodes */
11366 + unsigned long dl_inodes_total; /* maximum inodes */
11368 + unsigned int dl_nrlmult; /* non root limit mult */
11373 +extern void rcu_free_dl_info(struct rcu_head *);
11374 +extern void unhash_dl_info(struct dl_info *);
11376 +extern struct dl_info *locate_dl_info(struct super_block *, tag_t);
11381 +extern void vx_vsi_statfs(struct super_block *, struct kstatfs *);
11383 +typedef uint64_t dlsize_t;
11385 +#endif /* __KERNEL__ */
11386 +#else /* _VX_DLIMIT_H */
11387 +#warning duplicate inclusion
11388 +#endif /* _VX_DLIMIT_H */
11389 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/global.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/global.h
11390 --- linux-2.6.36/include/linux/vserver/global.h 1970-01-01 01:00:00.000000000 +0100
11391 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/global.h 2010-10-21 13:09:36.000000000 +0200
11393 +#ifndef _VX_GLOBAL_H
11394 +#define _VX_GLOBAL_H
11397 +extern atomic_t vx_global_ctotal;
11398 +extern atomic_t vx_global_cactive;
11400 +extern atomic_t nx_global_ctotal;
11401 +extern atomic_t nx_global_cactive;
11403 +extern atomic_t vs_global_nsproxy;
11404 +extern atomic_t vs_global_fs;
11405 +extern atomic_t vs_global_mnt_ns;
11406 +extern atomic_t vs_global_uts_ns;
11407 +extern atomic_t vs_global_user_ns;
11408 +extern atomic_t vs_global_pid_ns;
11411 +#endif /* _VX_GLOBAL_H */
11412 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/history.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/history.h
11413 --- linux-2.6.36/include/linux/vserver/history.h 1970-01-01 01:00:00.000000000 +0100
11414 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/history.h 2010-10-21 13:09:36.000000000 +0200
11416 +#ifndef _VX_HISTORY_H
11417 +#define _VX_HISTORY_H
11422 + VXH_THROW_OOPS = 1,
11426 + VXH_INIT_VX_INFO,
11429 + VXH_CLAIM_VX_INFO,
11430 + VXH_RELEASE_VX_INFO,
11431 + VXH_ALLOC_VX_INFO,
11432 + VXH_DEALLOC_VX_INFO,
11433 + VXH_HASH_VX_INFO,
11434 + VXH_UNHASH_VX_INFO,
11436 + VXH_LOOKUP_VX_INFO,
11437 + VXH_CREATE_VX_INFO,
11440 +struct _vxhe_vxi {
11441 + struct vx_info *ptr;
11447 +struct _vxhe_set_clr {
11451 +struct _vxhe_loc_lookup {
11455 +struct _vx_hist_entry {
11457 + unsigned short seq;
11458 + unsigned short type;
11459 + struct _vxhe_vxi vxi;
11461 + struct _vxhe_set_clr sc;
11462 + struct _vxhe_loc_lookup ll;
11466 +#ifdef CONFIG_VSERVER_HISTORY
11468 +extern unsigned volatile int vxh_active;
11470 +struct _vx_hist_entry *vxh_advance(void *loc);
11474 +void __vxh_copy_vxi(struct _vx_hist_entry *entry, struct vx_info *vxi)
11476 + entry->vxi.ptr = vxi;
11478 + entry->vxi.usecnt = atomic_read(&vxi->vx_usecnt);
11479 + entry->vxi.tasks = atomic_read(&vxi->vx_tasks);
11480 + entry->vxi.xid = vxi->vx_id;
11485 +#define __HERE__ current_text_addr()
11487 +#define __VXH_BODY(__type, __data, __here) \
11488 + struct _vx_hist_entry *entry; \
11490 + preempt_disable(); \
11491 + entry = vxh_advance(__here); \
11493 + entry->type = __type; \
11494 + preempt_enable();
11497 + /* pass vxi only */
11499 +#define __VXH_SMPL \
11500 + __vxh_copy_vxi(entry, vxi)
11503 +void __vxh_smpl(struct vx_info *vxi, int __type, void *__here)
11505 + __VXH_BODY(__type, __VXH_SMPL, __here)
11508 + /* pass vxi and data (void *) */
11510 +#define __VXH_DATA \
11511 + __vxh_copy_vxi(entry, vxi); \
11512 + entry->sc.data = data
11515 +void __vxh_data(struct vx_info *vxi, void *data,
11516 + int __type, void *__here)
11518 + __VXH_BODY(__type, __VXH_DATA, __here)
11521 + /* pass vxi and arg (long) */
11523 +#define __VXH_LONG \
11524 + __vxh_copy_vxi(entry, vxi); \
11525 + entry->ll.arg = arg
11528 +void __vxh_long(struct vx_info *vxi, long arg,
11529 + int __type, void *__here)
11531 + __VXH_BODY(__type, __VXH_LONG, __here)
11536 +void __vxh_throw_oops(void *__here)
11538 + __VXH_BODY(VXH_THROW_OOPS, {}, __here);
11539 + /* prevent further acquisition */
11544 +#define vxh_throw_oops() __vxh_throw_oops(__HERE__);
11546 +#define __vxh_get_vx_info(v, h) __vxh_smpl(v, VXH_GET_VX_INFO, h);
11547 +#define __vxh_put_vx_info(v, h) __vxh_smpl(v, VXH_PUT_VX_INFO, h);
11549 +#define __vxh_init_vx_info(v, d, h) \
11550 + __vxh_data(v, d, VXH_INIT_VX_INFO, h);
11551 +#define __vxh_set_vx_info(v, d, h) \
11552 + __vxh_data(v, d, VXH_SET_VX_INFO, h);
11553 +#define __vxh_clr_vx_info(v, d, h) \
11554 + __vxh_data(v, d, VXH_CLR_VX_INFO, h);
11556 +#define __vxh_claim_vx_info(v, d, h) \
11557 + __vxh_data(v, d, VXH_CLAIM_VX_INFO, h);
11558 +#define __vxh_release_vx_info(v, d, h) \
11559 + __vxh_data(v, d, VXH_RELEASE_VX_INFO, h);
11561 +#define vxh_alloc_vx_info(v) \
11562 + __vxh_smpl(v, VXH_ALLOC_VX_INFO, __HERE__);
11563 +#define vxh_dealloc_vx_info(v) \
11564 + __vxh_smpl(v, VXH_DEALLOC_VX_INFO, __HERE__);
11566 +#define vxh_hash_vx_info(v) \
11567 + __vxh_smpl(v, VXH_HASH_VX_INFO, __HERE__);
11568 +#define vxh_unhash_vx_info(v) \
11569 + __vxh_smpl(v, VXH_UNHASH_VX_INFO, __HERE__);
11571 +#define vxh_loc_vx_info(v, l) \
11572 + __vxh_long(v, l, VXH_LOC_VX_INFO, __HERE__);
11573 +#define vxh_lookup_vx_info(v, l) \
11574 + __vxh_long(v, l, VXH_LOOKUP_VX_INFO, __HERE__);
11575 +#define vxh_create_vx_info(v, l) \
11576 + __vxh_long(v, l, VXH_CREATE_VX_INFO, __HERE__);
11578 +extern void vxh_dump_history(void);
11581 +#else /* CONFIG_VSERVER_HISTORY */
11583 +#define __HERE__ 0
11585 +#define vxh_throw_oops() do { } while (0)
11587 +#define __vxh_get_vx_info(v, h) do { } while (0)
11588 +#define __vxh_put_vx_info(v, h) do { } while (0)
11590 +#define __vxh_init_vx_info(v, d, h) do { } while (0)
11591 +#define __vxh_set_vx_info(v, d, h) do { } while (0)
11592 +#define __vxh_clr_vx_info(v, d, h) do { } while (0)
11594 +#define __vxh_claim_vx_info(v, d, h) do { } while (0)
11595 +#define __vxh_release_vx_info(v, d, h) do { } while (0)
11597 +#define vxh_alloc_vx_info(v) do { } while (0)
11598 +#define vxh_dealloc_vx_info(v) do { } while (0)
11600 +#define vxh_hash_vx_info(v) do { } while (0)
11601 +#define vxh_unhash_vx_info(v) do { } while (0)
11603 +#define vxh_loc_vx_info(v, l) do { } while (0)
11604 +#define vxh_lookup_vx_info(v, l) do { } while (0)
11605 +#define vxh_create_vx_info(v, l) do { } while (0)
11607 +#define vxh_dump_history() do { } while (0)
11610 +#endif /* CONFIG_VSERVER_HISTORY */
11612 +#endif /* _VX_HISTORY_H */
11613 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/inode_cmd.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/inode_cmd.h
11614 --- linux-2.6.36/include/linux/vserver/inode_cmd.h 1970-01-01 01:00:00.000000000 +0100
11615 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/inode_cmd.h 2010-10-21 13:09:36.000000000 +0200
11617 +#ifndef _VX_INODE_CMD_H
11618 +#define _VX_INODE_CMD_H
11621 +/* inode vserver commands */
11623 +#define VCMD_get_iattr VC_CMD(INODE, 1, 1)
11624 +#define VCMD_set_iattr VC_CMD(INODE, 2, 1)
11626 +#define VCMD_fget_iattr VC_CMD(INODE, 3, 0)
11627 +#define VCMD_fset_iattr VC_CMD(INODE, 4, 0)
11629 +struct vcmd_ctx_iattr_v1 {
11630 + const char __user *name;
11636 +struct vcmd_ctx_fiattr_v0 {
11646 +#ifdef CONFIG_COMPAT
11648 +#include <asm/compat.h>
11650 +struct vcmd_ctx_iattr_v1_x32 {
11651 + compat_uptr_t name_ptr;
11657 +#endif /* CONFIG_COMPAT */
11659 +#include <linux/compiler.h>
11661 +extern int vc_get_iattr(void __user *);
11662 +extern int vc_set_iattr(void __user *);
11664 +extern int vc_fget_iattr(uint32_t, void __user *);
11665 +extern int vc_fset_iattr(uint32_t, void __user *);
11667 +#ifdef CONFIG_COMPAT
11669 +extern int vc_get_iattr_x32(void __user *);
11670 +extern int vc_set_iattr_x32(void __user *);
11672 +#endif /* CONFIG_COMPAT */
11674 +#endif /* __KERNEL__ */
11675 +#endif /* _VX_INODE_CMD_H */
11676 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/inode.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/inode.h
11677 --- linux-2.6.36/include/linux/vserver/inode.h 1970-01-01 01:00:00.000000000 +0100
11678 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/inode.h 2010-10-21 13:09:36.000000000 +0200
11680 +#ifndef _VX_INODE_H
11681 +#define _VX_INODE_H
11684 +#define IATTR_TAG 0x01000000
11686 +#define IATTR_ADMIN 0x00000001
11687 +#define IATTR_WATCH 0x00000002
11688 +#define IATTR_HIDE 0x00000004
11689 +#define IATTR_FLAGS 0x00000007
11691 +#define IATTR_BARRIER 0x00010000
11692 +#define IATTR_IXUNLINK 0x00020000
11693 +#define IATTR_IMMUTABLE 0x00040000
11694 +#define IATTR_COW 0x00080000
11699 +#ifdef CONFIG_VSERVER_PROC_SECURE
11700 +#define IATTR_PROC_DEFAULT ( IATTR_ADMIN | IATTR_HIDE )
11701 +#define IATTR_PROC_SYMLINK ( IATTR_ADMIN )
11703 +#define IATTR_PROC_DEFAULT ( IATTR_ADMIN )
11704 +#define IATTR_PROC_SYMLINK ( IATTR_ADMIN )
11707 +#define vx_hide_check(c, m) (((m) & IATTR_HIDE) ? vx_check(c, m) : 1)
11709 +#endif /* __KERNEL__ */
11711 +/* inode ioctls */
11713 +#define FIOC_GETXFLG _IOR('x', 5, long)
11714 +#define FIOC_SETXFLG _IOW('x', 6, long)
11716 +#else /* _VX_INODE_H */
11717 +#warning duplicate inclusion
11718 +#endif /* _VX_INODE_H */
11719 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/Kbuild linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/Kbuild
11720 --- linux-2.6.36/include/linux/vserver/Kbuild 1970-01-01 01:00:00.000000000 +0100
11721 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/Kbuild 2010-10-21 13:09:36.000000000 +0200
11724 +unifdef-y += context_cmd.h network_cmd.h space_cmd.h \
11725 + cacct_cmd.h cvirt_cmd.h limit_cmd.h dlimit_cmd.h \
11726 + inode_cmd.h tag_cmd.h sched_cmd.h signal_cmd.h \
11727 + debug_cmd.h device_cmd.h
11729 +unifdef-y += switch.h network.h monitor.h inode.h device.h
11731 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/limit_cmd.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/limit_cmd.h
11732 --- linux-2.6.36/include/linux/vserver/limit_cmd.h 1970-01-01 01:00:00.000000000 +0100
11733 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/limit_cmd.h 2010-10-21 13:09:36.000000000 +0200
11735 +#ifndef _VX_LIMIT_CMD_H
11736 +#define _VX_LIMIT_CMD_H
11739 +/* rlimit vserver commands */
11741 +#define VCMD_get_rlimit VC_CMD(RLIMIT, 1, 0)
11742 +#define VCMD_set_rlimit VC_CMD(RLIMIT, 2, 0)
11743 +#define VCMD_get_rlimit_mask VC_CMD(RLIMIT, 3, 0)
11744 +#define VCMD_reset_hits VC_CMD(RLIMIT, 7, 0)
11745 +#define VCMD_reset_minmax VC_CMD(RLIMIT, 9, 0)
11747 +struct vcmd_ctx_rlimit_v0 {
11749 + uint64_t minimum;
11750 + uint64_t softlimit;
11751 + uint64_t maximum;
11754 +struct vcmd_ctx_rlimit_mask_v0 {
11755 + uint32_t minimum;
11756 + uint32_t softlimit;
11757 + uint32_t maximum;
11760 +#define VCMD_rlimit_stat VC_CMD(VSTAT, 1, 0)
11762 +struct vcmd_rlimit_stat_v0 {
11766 + uint64_t minimum;
11767 + uint64_t maximum;
11770 +#define CRLIM_UNSET (0ULL)
11771 +#define CRLIM_INFINITY (~0ULL)
11772 +#define CRLIM_KEEP (~1ULL)
11776 +#ifdef CONFIG_IA32_EMULATION
11778 +struct vcmd_ctx_rlimit_v0_x32 {
11780 + uint64_t minimum;
11781 + uint64_t softlimit;
11782 + uint64_t maximum;
11783 +} __attribute__ ((packed));
11785 +#endif /* CONFIG_IA32_EMULATION */
11787 +#include <linux/compiler.h>
11789 +extern int vc_get_rlimit_mask(uint32_t, void __user *);
11790 +extern int vc_get_rlimit(struct vx_info *, void __user *);
11791 +extern int vc_set_rlimit(struct vx_info *, void __user *);
11792 +extern int vc_reset_hits(struct vx_info *, void __user *);
11793 +extern int vc_reset_minmax(struct vx_info *, void __user *);
11795 +extern int vc_rlimit_stat(struct vx_info *, void __user *);
11797 +#ifdef CONFIG_IA32_EMULATION
11799 +extern int vc_get_rlimit_x32(struct vx_info *, void __user *);
11800 +extern int vc_set_rlimit_x32(struct vx_info *, void __user *);
11802 +#endif /* CONFIG_IA32_EMULATION */
11804 +#endif /* __KERNEL__ */
11805 +#endif /* _VX_LIMIT_CMD_H */
11806 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/limit_def.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/limit_def.h
11807 --- linux-2.6.36/include/linux/vserver/limit_def.h 1970-01-01 01:00:00.000000000 +0100
11808 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/limit_def.h 2010-10-21 13:09:36.000000000 +0200
11810 +#ifndef _VX_LIMIT_DEF_H
11811 +#define _VX_LIMIT_DEF_H
11813 +#include <asm/atomic.h>
11814 +#include <asm/resource.h>
11816 +#include "limit.h"
11819 +struct _vx_res_limit {
11820 + rlim_t soft; /* Context soft limit */
11821 + rlim_t hard; /* Context hard limit */
11823 + rlim_atomic_t rcur; /* Current value */
11824 + rlim_t rmin; /* Context minimum */
11825 + rlim_t rmax; /* Context maximum */
11827 + atomic_t lhit; /* Limit hits */
11830 +/* context sub struct */
11832 +struct _vx_limit {
11833 + struct _vx_res_limit res[NUM_LIMITS];
11836 +#ifdef CONFIG_VSERVER_DEBUG
11838 +static inline void __dump_vx_limit(struct _vx_limit *limit)
11842 + printk("\t_vx_limit:");
11843 + for (i = 0; i < NUM_LIMITS; i++) {
11844 + printk("\t [%2d] = %8lu %8lu/%8lu, %8ld/%8ld, %8d\n",
11845 + i, (unsigned long)__rlim_get(limit, i),
11846 + (unsigned long)__rlim_rmin(limit, i),
11847 + (unsigned long)__rlim_rmax(limit, i),
11848 + (long)__rlim_soft(limit, i),
11849 + (long)__rlim_hard(limit, i),
11850 + atomic_read(&__rlim_lhit(limit, i)));
11856 +#endif /* _VX_LIMIT_DEF_H */
11857 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/limit.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/limit.h
11858 --- linux-2.6.36/include/linux/vserver/limit.h 1970-01-01 01:00:00.000000000 +0100
11859 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/limit.h 2010-10-21 13:09:36.000000000 +0200
11861 +#ifndef _VX_LIMIT_H
11862 +#define _VX_LIMIT_H
11864 +#define VLIMIT_NSOCK 16
11865 +#define VLIMIT_OPENFD 17
11866 +#define VLIMIT_ANON 18
11867 +#define VLIMIT_SHMEM 19
11868 +#define VLIMIT_SEMARY 20
11869 +#define VLIMIT_NSEMS 21
11870 +#define VLIMIT_DENTRY 22
11871 +#define VLIMIT_MAPPED 23
11876 +#define VLIM_NOCHECK ((1L << VLIMIT_DENTRY) | (1L << RLIMIT_RSS))
11878 +/* keep in sync with CRLIM_INFINITY */
11880 +#define VLIM_INFINITY (~0ULL)
11882 +#include <asm/atomic.h>
11883 +#include <asm/resource.h>
11885 +#ifndef RLIM_INFINITY
11886 +#warning RLIM_INFINITY is undefined
11889 +#define __rlim_val(l, r, v) ((l)->res[r].v)
11891 +#define __rlim_soft(l, r) __rlim_val(l, r, soft)
11892 +#define __rlim_hard(l, r) __rlim_val(l, r, hard)
11894 +#define __rlim_rcur(l, r) __rlim_val(l, r, rcur)
11895 +#define __rlim_rmin(l, r) __rlim_val(l, r, rmin)
11896 +#define __rlim_rmax(l, r) __rlim_val(l, r, rmax)
11898 +#define __rlim_lhit(l, r) __rlim_val(l, r, lhit)
11899 +#define __rlim_hit(l, r) atomic_inc(&__rlim_lhit(l, r))
11901 +typedef atomic_long_t rlim_atomic_t;
11902 +typedef unsigned long rlim_t;
11904 +#define __rlim_get(l, r) atomic_long_read(&__rlim_rcur(l, r))
11905 +#define __rlim_set(l, r, v) atomic_long_set(&__rlim_rcur(l, r), v)
11906 +#define __rlim_inc(l, r) atomic_long_inc(&__rlim_rcur(l, r))
11907 +#define __rlim_dec(l, r) atomic_long_dec(&__rlim_rcur(l, r))
11908 +#define __rlim_add(l, r, v) atomic_long_add(v, &__rlim_rcur(l, r))
11909 +#define __rlim_sub(l, r, v) atomic_long_sub(v, &__rlim_rcur(l, r))
11912 +#if (RLIM_INFINITY == VLIM_INFINITY)
11913 +#define VX_VLIM(r) ((long long)(long)(r))
11914 +#define VX_RLIM(v) ((rlim_t)(v))
11916 +#define VX_VLIM(r) (((r) == RLIM_INFINITY) \
11917 + ? VLIM_INFINITY : (long long)(r))
11918 +#define VX_RLIM(v) (((v) == VLIM_INFINITY) \
11919 + ? RLIM_INFINITY : (rlim_t)(v))
11924 +void vx_vsi_meminfo(struct sysinfo *);
11925 +void vx_vsi_swapinfo(struct sysinfo *);
11926 +long vx_vsi_cached(struct sysinfo *);
11928 +#define NUM_LIMITS 24
11930 +#endif /* __KERNEL__ */
11931 +#endif /* _VX_LIMIT_H */
11932 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/limit_int.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/limit_int.h
11933 --- linux-2.6.36/include/linux/vserver/limit_int.h 1970-01-01 01:00:00.000000000 +0100
11934 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/limit_int.h 2010-10-21 13:09:36.000000000 +0200
11936 +#ifndef _VX_LIMIT_INT_H
11937 +#define _VX_LIMIT_INT_H
11939 +#include "context.h"
11943 +#define VXD_RCRES_COND(r) VXD_CBIT(cres, r)
11944 +#define VXD_RLIMIT_COND(r) VXD_CBIT(limit, r)
11946 +extern const char *vlimit_name[NUM_LIMITS];
11948 +static inline void __vx_acc_cres(struct vx_info *vxi,
11949 + int res, int dir, void *_data, char *_file, int _line)
11951 + if (VXD_RCRES_COND(res))
11952 + vxlprintk(1, "vx_acc_cres[%5d,%s,%2d]: %5ld%s (%p)",
11953 + (vxi ? vxi->vx_id : -1), vlimit_name[res], res,
11954 + (vxi ? (long)__rlim_get(&vxi->limit, res) : 0),
11955 + (dir > 0) ? "++" : "--", _data, _file, _line);
11960 + __rlim_inc(&vxi->limit, res);
11962 + __rlim_dec(&vxi->limit, res);
11965 +static inline void __vx_add_cres(struct vx_info *vxi,
11966 + int res, int amount, void *_data, char *_file, int _line)
11968 + if (VXD_RCRES_COND(res))
11969 + vxlprintk(1, "vx_add_cres[%5d,%s,%2d]: %5ld += %5d (%p)",
11970 + (vxi ? vxi->vx_id : -1), vlimit_name[res], res,
11971 + (vxi ? (long)__rlim_get(&vxi->limit, res) : 0),
11972 + amount, _data, _file, _line);
11977 + __rlim_add(&vxi->limit, res, amount);
11981 +int __vx_cres_adjust_max(struct _vx_limit *limit, int res, rlim_t value)
11983 + int cond = (value > __rlim_rmax(limit, res));
11986 + __rlim_rmax(limit, res) = value;
11991 +int __vx_cres_adjust_min(struct _vx_limit *limit, int res, rlim_t value)
11993 + int cond = (value < __rlim_rmin(limit, res));
11996 + __rlim_rmin(limit, res) = value;
12001 +void __vx_cres_fixup(struct _vx_limit *limit, int res, rlim_t value)
12003 + if (!__vx_cres_adjust_max(limit, res, value))
12004 + __vx_cres_adjust_min(limit, res, value);
12009 + +1 ... no limit hit
12010 + -1 ... over soft limit
12011 + 0 ... over hard limit */
12013 +static inline int __vx_cres_avail(struct vx_info *vxi,
12014 + int res, int num, char *_file, int _line)
12016 + struct _vx_limit *limit;
12019 + if (VXD_RLIMIT_COND(res))
12020 + vxlprintk(1, "vx_cres_avail[%5d,%s,%2d]: %5ld/%5ld > %5ld + %5d",
12021 + (vxi ? vxi->vx_id : -1), vlimit_name[res], res,
12022 + (vxi ? (long)__rlim_soft(&vxi->limit, res) : -1),
12023 + (vxi ? (long)__rlim_hard(&vxi->limit, res) : -1),
12024 + (vxi ? (long)__rlim_get(&vxi->limit, res) : 0),
12025 + num, _file, _line);
12029 + limit = &vxi->limit;
12030 + value = __rlim_get(limit, res);
12032 + if (!__vx_cres_adjust_max(limit, res, value))
12033 + __vx_cres_adjust_min(limit, res, value);
12038 + if (__rlim_soft(limit, res) == RLIM_INFINITY)
12040 + if (value + num <= __rlim_soft(limit, res))
12043 + if (__rlim_hard(limit, res) == RLIM_INFINITY)
12045 + if (value + num <= __rlim_hard(limit, res))
12048 + __rlim_hit(limit, res);
12053 +static const int VLA_RSS[] = { RLIMIT_RSS, VLIMIT_ANON, VLIMIT_MAPPED, 0 };
12056 +rlim_t __vx_cres_array_sum(struct _vx_limit *limit, const int *array)
12058 + rlim_t value, sum = 0;
12061 + while ((res = *array++)) {
12062 + value = __rlim_get(limit, res);
12063 + __vx_cres_fixup(limit, res, value);
12070 +rlim_t __vx_cres_array_fixup(struct _vx_limit *limit, const int *array)
12072 + rlim_t value = __vx_cres_array_sum(limit, array + 1);
12073 + int res = *array;
12075 + if (value == __rlim_get(limit, res))
12078 + __rlim_set(limit, res, value);
12079 + /* now adjust min/max */
12080 + if (!__vx_cres_adjust_max(limit, res, value))
12081 + __vx_cres_adjust_min(limit, res, value);
12086 +static inline int __vx_cres_array_avail(struct vx_info *vxi,
12087 + const int *array, int num, char *_file, int _line)
12089 + struct _vx_limit *limit;
12090 + rlim_t value = 0;
12098 + limit = &vxi->limit;
12100 + value = __vx_cres_array_sum(limit, array + 1);
12102 + __rlim_set(limit, res, value);
12103 + __vx_cres_fixup(limit, res, value);
12105 + return __vx_cres_avail(vxi, res, num, _file, _line);
12109 +static inline void vx_limit_fixup(struct _vx_limit *limit, int id)
12114 + /* complex resources first */
12115 + if ((id < 0) || (id == RLIMIT_RSS))
12116 + __vx_cres_array_fixup(limit, VLA_RSS);
12118 + for (res = 0; res < NUM_LIMITS; res++) {
12119 + if ((id > 0) && (res != id))
12122 + value = __rlim_get(limit, res);
12123 + __vx_cres_fixup(limit, res, value);
12125 + /* not supposed to happen, maybe warn? */
12126 + if (__rlim_rmax(limit, res) > __rlim_hard(limit, res))
12127 + __rlim_rmax(limit, res) = __rlim_hard(limit, res);
12132 +#endif /* __KERNEL__ */
12133 +#endif /* _VX_LIMIT_INT_H */
12134 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/monitor.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/monitor.h
12135 --- linux-2.6.36/include/linux/vserver/monitor.h 1970-01-01 01:00:00.000000000 +0100
12136 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/monitor.h 2010-10-21 13:09:36.000000000 +0200
12138 +#ifndef _VX_MONITOR_H
12139 +#define _VX_MONITOR_H
12141 +#include <linux/types.h>
12148 + VXM_UPDATE = 0x20,
12152 + VXM_RQINFO_1 = 0x24,
12155 + VXM_ACTIVATE = 0x40,
12162 + VXM_MIGRATE = 0x48,
12165 + /* all other bits are flags */
12166 + VXM_SCHED = 0x80,
12169 +struct _vxm_update_1 {
12170 + uint32_t tokens_max;
12171 + uint32_t fill_rate;
12172 + uint32_t interval;
12175 +struct _vxm_update_2 {
12176 + uint32_t tokens_min;
12177 + uint32_t fill_rate;
12178 + uint32_t interval;
12181 +struct _vxm_rqinfo_1 {
12182 + uint16_t running;
12186 + uint32_t idle_tokens;
12189 +struct _vxm_rqinfo_2 {
12190 + uint32_t norm_time;
12191 + uint32_t idle_time;
12192 + uint32_t idle_skip;
12195 +struct _vxm_sched {
12197 + uint32_t norm_time;
12198 + uint32_t idle_time;
12201 +struct _vxm_task {
12206 +struct _vxm_event {
12215 + struct _vxm_task tsk;
12219 +struct _vx_mon_entry {
12223 + struct _vxm_event ev;
12224 + struct _vxm_sched sd;
12225 + struct _vxm_update_1 u1;
12226 + struct _vxm_update_2 u2;
12227 + struct _vxm_rqinfo_1 q1;
12228 + struct _vxm_rqinfo_2 q2;
12233 +#endif /* _VX_MONITOR_H */
12234 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/network_cmd.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/network_cmd.h
12235 --- linux-2.6.36/include/linux/vserver/network_cmd.h 1970-01-01 01:00:00.000000000 +0100
12236 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/network_cmd.h 2010-10-21 13:09:36.000000000 +0200
12238 +#ifndef _VX_NETWORK_CMD_H
12239 +#define _VX_NETWORK_CMD_H
12242 +/* vinfo commands */
12244 +#define VCMD_task_nid VC_CMD(VINFO, 2, 0)
12247 +extern int vc_task_nid(uint32_t);
12249 +#endif /* __KERNEL__ */
12251 +#define VCMD_nx_info VC_CMD(VINFO, 6, 0)
12253 +struct vcmd_nx_info_v0 {
12255 + /* more to come */
12259 +extern int vc_nx_info(struct nx_info *, void __user *);
12261 +#endif /* __KERNEL__ */
12263 +#include <linux/in.h>
12264 +#include <linux/in6.h>
12266 +#define VCMD_net_create_v0 VC_CMD(VNET, 1, 0)
12267 +#define VCMD_net_create VC_CMD(VNET, 1, 1)
12269 +struct vcmd_net_create {
12270 + uint64_t flagword;
12273 +#define VCMD_net_migrate VC_CMD(NETMIG, 1, 0)
12275 +#define VCMD_net_add VC_CMD(NETALT, 1, 0)
12276 +#define VCMD_net_remove VC_CMD(NETALT, 2, 0)
12278 +struct vcmd_net_addr_v0 {
12281 + struct in_addr ip[4];
12282 + struct in_addr mask[4];
12285 +#define VCMD_net_add_ipv4 VC_CMD(NETALT, 1, 1)
12286 +#define VCMD_net_remove_ipv4 VC_CMD(NETALT, 2, 1)
12288 +struct vcmd_net_addr_ipv4_v1 {
12291 + struct in_addr ip;
12292 + struct in_addr mask;
12295 +#define VCMD_net_add_ipv6 VC_CMD(NETALT, 3, 1)
12296 +#define VCMD_net_remove_ipv6 VC_CMD(NETALT, 4, 1)
12298 +struct vcmd_net_addr_ipv6_v1 {
12302 + struct in6_addr ip;
12303 + struct in6_addr mask;
12306 +#define VCMD_add_match_ipv4 VC_CMD(NETALT, 5, 0)
12307 +#define VCMD_get_match_ipv4 VC_CMD(NETALT, 6, 0)
12309 +struct vcmd_match_ipv4_v0 {
12314 + struct in_addr ip;
12315 + struct in_addr ip2;
12316 + struct in_addr mask;
12319 +#define VCMD_add_match_ipv6 VC_CMD(NETALT, 7, 0)
12320 +#define VCMD_get_match_ipv6 VC_CMD(NETALT, 8, 0)
12322 +struct vcmd_match_ipv6_v0 {
12327 + struct in6_addr ip;
12328 + struct in6_addr ip2;
12329 + struct in6_addr mask;
12334 +extern int vc_net_create(uint32_t, void __user *);
12335 +extern int vc_net_migrate(struct nx_info *, void __user *);
12337 +extern int vc_net_add(struct nx_info *, void __user *);
12338 +extern int vc_net_remove(struct nx_info *, void __user *);
12340 +extern int vc_net_add_ipv4(struct nx_info *, void __user *);
12341 +extern int vc_net_remove_ipv4(struct nx_info *, void __user *);
12343 +extern int vc_net_add_ipv6(struct nx_info *, void __user *);
12344 +extern int vc_net_remove_ipv6(struct nx_info *, void __user *);
12346 +extern int vc_add_match_ipv4(struct nx_info *, void __user *);
12347 +extern int vc_get_match_ipv4(struct nx_info *, void __user *);
12349 +extern int vc_add_match_ipv6(struct nx_info *, void __user *);
12350 +extern int vc_get_match_ipv6(struct nx_info *, void __user *);
12352 +#endif /* __KERNEL__ */
12355 +/* flag commands */
12357 +#define VCMD_get_nflags VC_CMD(FLAGS, 5, 0)
12358 +#define VCMD_set_nflags VC_CMD(FLAGS, 6, 0)
12360 +struct vcmd_net_flags_v0 {
12361 + uint64_t flagword;
12366 +extern int vc_get_nflags(struct nx_info *, void __user *);
12367 +extern int vc_set_nflags(struct nx_info *, void __user *);
12369 +#endif /* __KERNEL__ */
12372 +/* network caps commands */
12374 +#define VCMD_get_ncaps VC_CMD(FLAGS, 7, 0)
12375 +#define VCMD_set_ncaps VC_CMD(FLAGS, 8, 0)
12377 +struct vcmd_net_caps_v0 {
12383 +extern int vc_get_ncaps(struct nx_info *, void __user *);
12384 +extern int vc_set_ncaps(struct nx_info *, void __user *);
12386 +#endif /* __KERNEL__ */
12387 +#endif /* _VX_CONTEXT_CMD_H */
12388 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/network.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/network.h
12389 --- linux-2.6.36/include/linux/vserver/network.h 1970-01-01 01:00:00.000000000 +0100
12390 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/network.h 2010-10-21 13:09:36.000000000 +0200
12392 +#ifndef _VX_NETWORK_H
12393 +#define _VX_NETWORK_H
12395 +#include <linux/types.h>
12398 +#define MAX_N_CONTEXT 65535 /* Arbitrary limit */
12401 +/* network flags */
12403 +#define NXF_INFO_PRIVATE 0x00000008
12405 +#define NXF_SINGLE_IP 0x00000100
12406 +#define NXF_LBACK_REMAP 0x00000200
12407 +#define NXF_LBACK_ALLOW 0x00000400
12409 +#define NXF_HIDE_NETIF 0x02000000
12410 +#define NXF_HIDE_LBACK 0x04000000
12412 +#define NXF_STATE_SETUP (1ULL << 32)
12413 +#define NXF_STATE_ADMIN (1ULL << 34)
12415 +#define NXF_SC_HELPER (1ULL << 36)
12416 +#define NXF_PERSISTENT (1ULL << 38)
12418 +#define NXF_ONE_TIME (0x0005ULL << 32)
12421 +#define NXF_INIT_SET (__nxf_init_set())
12423 +static inline uint64_t __nxf_init_set(void) {
12424 + return NXF_STATE_ADMIN
12425 +#ifdef CONFIG_VSERVER_AUTO_LBACK
12426 + | NXF_LBACK_REMAP
12429 +#ifdef CONFIG_VSERVER_AUTO_SINGLE
12432 + | NXF_HIDE_NETIF;
12436 +/* network caps */
12438 +#define NXC_TUN_CREATE 0x00000001
12440 +#define NXC_RAW_ICMP 0x00000100
12443 +/* address types */
12445 +#define NXA_TYPE_IPV4 0x0001
12446 +#define NXA_TYPE_IPV6 0x0002
12448 +#define NXA_TYPE_NONE 0x0000
12449 +#define NXA_TYPE_ANY 0x00FF
12451 +#define NXA_TYPE_ADDR 0x0010
12452 +#define NXA_TYPE_MASK 0x0020
12453 +#define NXA_TYPE_RANGE 0x0040
12455 +#define NXA_MASK_ALL (NXA_TYPE_ADDR | NXA_TYPE_MASK | NXA_TYPE_RANGE)
12457 +#define NXA_MOD_BCAST 0x0100
12458 +#define NXA_MOD_LBACK 0x0200
12460 +#define NXA_LOOPBACK 0x1000
12462 +#define NXA_MASK_BIND (NXA_MASK_ALL | NXA_MOD_BCAST | NXA_MOD_LBACK)
12463 +#define NXA_MASK_SHOW (NXA_MASK_ALL | NXA_LOOPBACK)
12467 +#include <linux/list.h>
12468 +#include <linux/spinlock.h>
12469 +#include <linux/rcupdate.h>
12470 +#include <linux/in.h>
12471 +#include <linux/in6.h>
12472 +#include <asm/atomic.h>
12474 +struct nx_addr_v4 {
12475 + struct nx_addr_v4 *next;
12476 + struct in_addr ip[2];
12477 + struct in_addr mask;
12482 +struct nx_addr_v6 {
12483 + struct nx_addr_v6 *next;
12484 + struct in6_addr ip;
12485 + struct in6_addr mask;
12492 + struct hlist_node nx_hlist; /* linked list of nxinfos */
12493 + nid_t nx_id; /* vnet id */
12494 + atomic_t nx_usecnt; /* usage count */
12495 + atomic_t nx_tasks; /* tasks count */
12496 + int nx_state; /* context state */
12498 + uint64_t nx_flags; /* network flag word */
12499 + uint64_t nx_ncaps; /* network capabilities */
12501 + struct in_addr v4_lback; /* Loopback address */
12502 + struct in_addr v4_bcast; /* Broadcast address */
12503 + struct nx_addr_v4 v4; /* First/Single ipv4 address */
12504 +#ifdef CONFIG_IPV6
12505 + struct nx_addr_v6 v6; /* First/Single ipv6 address */
12507 + char nx_name[65]; /* network context name */
12511 +/* status flags */
12513 +#define NXS_HASHED 0x0001
12514 +#define NXS_SHUTDOWN 0x0100
12515 +#define NXS_RELEASED 0x8000
12517 +extern struct nx_info *lookup_nx_info(int);
12519 +extern int get_nid_list(int, unsigned int *, int);
12520 +extern int nid_is_hashed(nid_t);
12522 +extern int nx_migrate_task(struct task_struct *, struct nx_info *);
12524 +extern long vs_net_change(struct nx_info *, unsigned int);
12529 +#define NX_IPV4(n) ((n)->v4.type != NXA_TYPE_NONE)
12530 +#ifdef CONFIG_IPV6
12531 +#define NX_IPV6(n) ((n)->v6.type != NXA_TYPE_NONE)
12533 +#define NX_IPV6(n) (0)
12536 +#endif /* __KERNEL__ */
12537 +#endif /* _VX_NETWORK_H */
12538 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/percpu.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/percpu.h
12539 --- linux-2.6.36/include/linux/vserver/percpu.h 1970-01-01 01:00:00.000000000 +0100
12540 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/percpu.h 2010-10-21 13:09:36.000000000 +0200
12542 +#ifndef _VX_PERCPU_H
12543 +#define _VX_PERCPU_H
12545 +#include "cvirt_def.h"
12546 +#include "sched_def.h"
12548 +struct _vx_percpu {
12549 + struct _vx_cvirt_pc cvirt;
12550 + struct _vx_sched_pc sched;
12553 +#define PERCPU_PERCTX (sizeof(struct _vx_percpu))
12555 +#endif /* _VX_PERCPU_H */
12556 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/pid.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/pid.h
12557 --- linux-2.6.36/include/linux/vserver/pid.h 1970-01-01 01:00:00.000000000 +0100
12558 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/pid.h 2010-10-21 13:09:36.000000000 +0200
12560 +#ifndef _VSERVER_PID_H
12561 +#define _VSERVER_PID_H
12563 +/* pid faking stuff */
12565 +#define vx_info_map_pid(v, p) \
12566 + __vx_info_map_pid((v), (p), __func__, __FILE__, __LINE__)
12567 +#define vx_info_map_tgid(v,p) vx_info_map_pid(v,p)
12568 +#define vx_map_pid(p) vx_info_map_pid(current_vx_info(), p)
12569 +#define vx_map_tgid(p) vx_map_pid(p)
12571 +static inline int __vx_info_map_pid(struct vx_info *vxi, int pid,
12572 + const char *func, const char *file, int line)
12574 + if (vx_info_flags(vxi, VXF_INFO_INIT, 0)) {
12575 + vxfprintk(VXD_CBIT(cvirt, 2),
12576 + "vx_map_tgid: %p/%llx: %d -> %d",
12577 + vxi, (long long)vxi->vx_flags, pid,
12578 + (pid && pid == vxi->vx_initpid) ? 1 : pid,
12579 + func, file, line);
12582 + if (pid == vxi->vx_initpid)
12588 +#define vx_info_rmap_pid(v, p) \
12589 + __vx_info_rmap_pid((v), (p), __func__, __FILE__, __LINE__)
12590 +#define vx_rmap_pid(p) vx_info_rmap_pid(current_vx_info(), p)
12591 +#define vx_rmap_tgid(p) vx_rmap_pid(p)
12593 +static inline int __vx_info_rmap_pid(struct vx_info *vxi, int pid,
12594 + const char *func, const char *file, int line)
12596 + if (vx_info_flags(vxi, VXF_INFO_INIT, 0)) {
12597 + vxfprintk(VXD_CBIT(cvirt, 2),
12598 + "vx_rmap_tgid: %p/%llx: %d -> %d",
12599 + vxi, (long long)vxi->vx_flags, pid,
12600 + (pid == 1) ? vxi->vx_initpid : pid,
12601 + func, file, line);
12602 + if ((pid == 1) && vxi->vx_initpid)
12603 + return vxi->vx_initpid;
12604 + if (pid == vxi->vx_initpid)
12611 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/sched_cmd.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/sched_cmd.h
12612 --- linux-2.6.36/include/linux/vserver/sched_cmd.h 1970-01-01 01:00:00.000000000 +0100
12613 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/sched_cmd.h 2010-10-21 13:09:36.000000000 +0200
12615 +#ifndef _VX_SCHED_CMD_H
12616 +#define _VX_SCHED_CMD_H
12619 +/* sched vserver commands */
12621 +#define VCMD_set_sched_v2 VC_CMD(SCHED, 1, 2)
12622 +#define VCMD_set_sched_v3 VC_CMD(SCHED, 1, 3)
12623 +#define VCMD_set_sched_v4 VC_CMD(SCHED, 1, 4)
12625 +struct vcmd_set_sched_v2 {
12626 + int32_t fill_rate;
12627 + int32_t interval;
12629 + int32_t tokens_min;
12630 + int32_t tokens_max;
12631 + uint64_t cpu_mask;
12634 +struct vcmd_set_sched_v3 {
12635 + uint32_t set_mask;
12636 + int32_t fill_rate;
12637 + int32_t interval;
12639 + int32_t tokens_min;
12640 + int32_t tokens_max;
12641 + int32_t priority_bias;
12644 +struct vcmd_set_sched_v4 {
12645 + uint32_t set_mask;
12646 + int32_t fill_rate;
12647 + int32_t interval;
12649 + int32_t tokens_min;
12650 + int32_t tokens_max;
12651 + int32_t prio_bias;
12653 + int32_t bucket_id;
12656 +#define VCMD_set_sched VC_CMD(SCHED, 1, 5)
12657 +#define VCMD_get_sched VC_CMD(SCHED, 2, 5)
12659 +struct vcmd_sched_v5 {
12662 + int32_t bucket_id;
12663 + int32_t fill_rate[2];
12664 + int32_t interval[2];
12666 + int32_t tokens_min;
12667 + int32_t tokens_max;
12668 + int32_t prio_bias;
12671 +#define VXSM_FILL_RATE 0x0001
12672 +#define VXSM_INTERVAL 0x0002
12673 +#define VXSM_FILL_RATE2 0x0004
12674 +#define VXSM_INTERVAL2 0x0008
12675 +#define VXSM_TOKENS 0x0010
12676 +#define VXSM_TOKENS_MIN 0x0020
12677 +#define VXSM_TOKENS_MAX 0x0040
12678 +#define VXSM_PRIO_BIAS 0x0100
12680 +#define VXSM_IDLE_TIME 0x0200
12681 +#define VXSM_FORCE 0x0400
12683 +#define VXSM_V3_MASK 0x0173
12684 +#define VXSM_SET_MASK 0x01FF
12686 +#define VXSM_CPU_ID 0x1000
12687 +#define VXSM_BUCKET_ID 0x2000
12689 +#define VXSM_MSEC 0x4000
12691 +#define SCHED_KEEP (-2) /* only for v2 */
12695 +#include <linux/compiler.h>
12697 +extern int vc_set_sched_v2(struct vx_info *, void __user *);
12698 +extern int vc_set_sched_v3(struct vx_info *, void __user *);
12699 +extern int vc_set_sched_v4(struct vx_info *, void __user *);
12700 +extern int vc_set_sched(struct vx_info *, void __user *);
12701 +extern int vc_get_sched(struct vx_info *, void __user *);
12703 +#endif /* __KERNEL__ */
12705 +#define VCMD_sched_info VC_CMD(SCHED, 3, 0)
12707 +struct vcmd_sched_info {
12709 + int32_t bucket_id;
12710 + uint64_t user_msec;
12711 + uint64_t sys_msec;
12712 + uint64_t hold_msec;
12713 + uint32_t token_usec;
12714 + int32_t vavavoom;
12719 +extern int vc_sched_info(struct vx_info *, void __user *);
12721 +#endif /* __KERNEL__ */
12722 +#endif /* _VX_SCHED_CMD_H */
12723 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/sched_def.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/sched_def.h
12724 --- linux-2.6.36/include/linux/vserver/sched_def.h 1970-01-01 01:00:00.000000000 +0100
12725 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/sched_def.h 2010-10-21 13:09:36.000000000 +0200
12727 +#ifndef _VX_SCHED_DEF_H
12728 +#define _VX_SCHED_DEF_H
12730 +#include <linux/spinlock.h>
12731 +#include <linux/jiffies.h>
12732 +#include <linux/cpumask.h>
12733 +#include <asm/atomic.h>
12734 +#include <asm/param.h>
12737 +/* context sub struct */
12739 +struct _vx_sched {
12740 + spinlock_t tokens_lock; /* lock for token bucket */
12742 + int tokens; /* number of CPU tokens */
12743 + int fill_rate[2]; /* Fill rate: add X tokens... */
12744 + int interval[2]; /* Divisor: per Y jiffies */
12745 + int tokens_min; /* Limit: minimum for unhold */
12746 + int tokens_max; /* Limit: no more than N tokens */
12748 + int prio_bias; /* bias offset for priority */
12750 + unsigned update_mask; /* which features should be updated */
12751 + cpumask_t update; /* CPUs which should update */
12754 +struct _vx_sched_pc {
12755 + int tokens; /* number of CPU tokens */
12756 + int flags; /* bucket flags */
12758 + int fill_rate[2]; /* Fill rate: add X tokens... */
12759 + int interval[2]; /* Divisor: per Y jiffies */
12760 + int tokens_min; /* Limit: minimum for unhold */
12761 + int tokens_max; /* Limit: no more than N tokens */
12763 + int prio_bias; /* bias offset for priority */
12764 + int vavavoom; /* last calculated vavavoom */
12766 + unsigned long norm_time; /* last time accounted */
12767 + unsigned long idle_time; /* non linear time for fair sched */
12768 + unsigned long token_time; /* token time for accounting */
12769 + unsigned long onhold; /* jiffies when put on hold */
12771 + uint64_t user_ticks; /* token tick events */
12772 + uint64_t sys_ticks; /* token tick events */
12773 + uint64_t hold_ticks; /* token ticks paused */
12777 +#define VXSF_ONHOLD 0x0001
12778 +#define VXSF_IDLE_TIME 0x0100
12780 +#ifdef CONFIG_VSERVER_DEBUG
12782 +static inline void __dump_vx_sched(struct _vx_sched *sched)
12784 + printk("\t_vx_sched:\n");
12785 + printk("\t tokens: %4d/%4d, %4d/%4d, %4d, %4d\n",
12786 + sched->fill_rate[0], sched->interval[0],
12787 + sched->fill_rate[1], sched->interval[1],
12788 + sched->tokens_min, sched->tokens_max);
12789 + printk("\t priority = %4d\n", sched->prio_bias);
12794 +#endif /* _VX_SCHED_DEF_H */
12795 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/sched.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/sched.h
12796 --- linux-2.6.36/include/linux/vserver/sched.h 1970-01-01 01:00:00.000000000 +0100
12797 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/sched.h 2010-10-21 13:09:36.000000000 +0200
12799 +#ifndef _VX_SCHED_H
12800 +#define _VX_SCHED_H
12807 +void vx_vsi_uptime(struct timespec *, struct timespec *);
12812 +void vx_update_load(struct vx_info *);
12815 +int vx_tokens_recalc(struct _vx_sched_pc *,
12816 + unsigned long *, unsigned long *, int [2]);
12818 +void vx_update_sched_param(struct _vx_sched *sched,
12819 + struct _vx_sched_pc *sched_pc);
12821 +#endif /* __KERNEL__ */
12822 +#else /* _VX_SCHED_H */
12823 +#warning duplicate inclusion
12824 +#endif /* _VX_SCHED_H */
12825 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/signal_cmd.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/signal_cmd.h
12826 --- linux-2.6.36/include/linux/vserver/signal_cmd.h 1970-01-01 01:00:00.000000000 +0100
12827 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/signal_cmd.h 2010-10-21 13:09:36.000000000 +0200
12829 +#ifndef _VX_SIGNAL_CMD_H
12830 +#define _VX_SIGNAL_CMD_H
12833 +/* signalling vserver commands */
12835 +#define VCMD_ctx_kill VC_CMD(PROCTRL, 1, 0)
12836 +#define VCMD_wait_exit VC_CMD(EVENT, 99, 0)
12838 +struct vcmd_ctx_kill_v0 {
12843 +struct vcmd_wait_exit_v0 {
12844 + int32_t reboot_cmd;
12845 + int32_t exit_code;
12850 +extern int vc_ctx_kill(struct vx_info *, void __user *);
12851 +extern int vc_wait_exit(struct vx_info *, void __user *);
12853 +#endif /* __KERNEL__ */
12855 +/* process alteration commands */
12857 +#define VCMD_get_pflags VC_CMD(PROCALT, 5, 0)
12858 +#define VCMD_set_pflags VC_CMD(PROCALT, 6, 0)
12860 +struct vcmd_pflags_v0 {
12861 + uint32_t flagword;
12867 +extern int vc_get_pflags(uint32_t pid, void __user *);
12868 +extern int vc_set_pflags(uint32_t pid, void __user *);
12870 +#endif /* __KERNEL__ */
12871 +#endif /* _VX_SIGNAL_CMD_H */
12872 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/signal.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/signal.h
12873 --- linux-2.6.36/include/linux/vserver/signal.h 1970-01-01 01:00:00.000000000 +0100
12874 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/signal.h 2010-10-21 13:09:36.000000000 +0200
12876 +#ifndef _VX_SIGNAL_H
12877 +#define _VX_SIGNAL_H
12884 +int vx_info_kill(struct vx_info *, int, int);
12886 +#endif /* __KERNEL__ */
12887 +#else /* _VX_SIGNAL_H */
12888 +#warning duplicate inclusion
12889 +#endif /* _VX_SIGNAL_H */
12890 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/space_cmd.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/space_cmd.h
12891 --- linux-2.6.36/include/linux/vserver/space_cmd.h 1970-01-01 01:00:00.000000000 +0100
12892 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/space_cmd.h 2010-10-21 13:09:36.000000000 +0200
12894 +#ifndef _VX_SPACE_CMD_H
12895 +#define _VX_SPACE_CMD_H
12898 +#define VCMD_enter_space_v0 VC_CMD(PROCALT, 1, 0)
12899 +#define VCMD_enter_space_v1 VC_CMD(PROCALT, 1, 1)
12900 +#define VCMD_enter_space VC_CMD(PROCALT, 1, 2)
12902 +#define VCMD_set_space_v0 VC_CMD(PROCALT, 3, 0)
12903 +#define VCMD_set_space_v1 VC_CMD(PROCALT, 3, 1)
12904 +#define VCMD_set_space VC_CMD(PROCALT, 3, 2)
12906 +#define VCMD_get_space_mask_v0 VC_CMD(PROCALT, 4, 0)
12908 +#define VCMD_get_space_mask VC_CMD(VSPACE, 0, 1)
12909 +#define VCMD_get_space_default VC_CMD(VSPACE, 1, 0)
12912 +struct vcmd_space_mask_v1 {
12916 +struct vcmd_space_mask_v2 {
12924 +extern int vc_enter_space_v1(struct vx_info *, void __user *);
12925 +extern int vc_set_space_v1(struct vx_info *, void __user *);
12926 +extern int vc_enter_space(struct vx_info *, void __user *);
12927 +extern int vc_set_space(struct vx_info *, void __user *);
12928 +extern int vc_get_space_mask(void __user *, int);
12930 +#endif /* __KERNEL__ */
12931 +#endif /* _VX_SPACE_CMD_H */
12932 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/space.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/space.h
12933 --- linux-2.6.36/include/linux/vserver/space.h 1970-01-01 01:00:00.000000000 +0100
12934 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/space.h 2010-10-21 13:09:36.000000000 +0200
12936 +#ifndef _VX_SPACE_H
12937 +#define _VX_SPACE_H
12939 +#include <linux/types.h>
12943 +int vx_set_space(struct vx_info *vxi, unsigned long mask, unsigned index);
12945 +#else /* _VX_SPACE_H */
12946 +#warning duplicate inclusion
12947 +#endif /* _VX_SPACE_H */
12948 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/switch.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/switch.h
12949 --- linux-2.6.36/include/linux/vserver/switch.h 1970-01-01 01:00:00.000000000 +0100
12950 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/switch.h 2010-10-21 13:09:36.000000000 +0200
12952 +#ifndef _VX_SWITCH_H
12953 +#define _VX_SWITCH_H
12955 +#include <linux/types.h>
12958 +#define VC_CATEGORY(c) (((c) >> 24) & 0x3F)
12959 +#define VC_COMMAND(c) (((c) >> 16) & 0xFF)
12960 +#define VC_VERSION(c) ((c) & 0xFFF)
12962 +#define VC_CMD(c, i, v) ((((VC_CAT_ ## c) & 0x3F) << 24) \
12963 + | (((i) & 0xFF) << 16) | ((v) & 0xFFF))
12967 + Syscall Matrix V2.8
12969 + |VERSION|CREATE |MODIFY |MIGRATE|CONTROL|EXPERIM| |SPECIAL|SPECIAL|
12970 + |STATS |DESTROY|ALTER |CHANGE |LIMIT |TEST | | | |
12971 + |INFO |SETUP | |MOVE | | | | | |
12972 + -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
12973 + SYSTEM |VERSION|VSETUP |VHOST | | | | |DEVICE | |
12974 + HOST | 00| 01| 02| 03| 04| 05| | 06| 07|
12975 + -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
12976 + CPU | |VPROC |PROCALT|PROCMIG|PROCTRL| | |SCHED. | |
12977 + PROCESS| 08| 09| 10| 11| 12| 13| | 14| 15|
12978 + -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
12979 + MEMORY | | | | |MEMCTRL| | |SWAP | |
12980 + | 16| 17| 18| 19| 20| 21| | 22| 23|
12981 + -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
12982 + NETWORK| |VNET |NETALT |NETMIG |NETCTL | | |SERIAL | |
12983 + | 24| 25| 26| 27| 28| 29| | 30| 31|
12984 + -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
12985 + DISK | | | |TAGMIG |DLIMIT | | |INODE | |
12986 + VFS | 32| 33| 34| 35| 36| 37| | 38| 39|
12987 + -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
12988 + OTHER |VSTAT | | | | | | |VINFO | |
12989 + | 40| 41| 42| 43| 44| 45| | 46| 47|
12990 + =======+=======+=======+=======+=======+=======+=======+ +=======+=======+
12991 + SPECIAL|EVENT | | | |FLAGS | | |VSPACE | |
12992 + | 48| 49| 50| 51| 52| 53| | 54| 55|
12993 + -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
12994 + SPECIAL|DEBUG | | | |RLIMIT |SYSCALL| | |COMPAT |
12995 + | 56| 57| 58| 59| 60|TEST 61| | 62| 63|
12996 + -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
13000 +#define VC_CAT_VERSION 0
13002 +#define VC_CAT_VSETUP 1
13003 +#define VC_CAT_VHOST 2
13005 +#define VC_CAT_DEVICE 6
13007 +#define VC_CAT_VPROC 9
13008 +#define VC_CAT_PROCALT 10
13009 +#define VC_CAT_PROCMIG 11
13010 +#define VC_CAT_PROCTRL 12
13012 +#define VC_CAT_SCHED 14
13013 +#define VC_CAT_MEMCTRL 20
13015 +#define VC_CAT_VNET 25
13016 +#define VC_CAT_NETALT 26
13017 +#define VC_CAT_NETMIG 27
13018 +#define VC_CAT_NETCTRL 28
13020 +#define VC_CAT_TAGMIG 35
13021 +#define VC_CAT_DLIMIT 36
13022 +#define VC_CAT_INODE 38
13024 +#define VC_CAT_VSTAT 40
13025 +#define VC_CAT_VINFO 46
13026 +#define VC_CAT_EVENT 48
13028 +#define VC_CAT_FLAGS 52
13029 +#define VC_CAT_VSPACE 54
13030 +#define VC_CAT_DEBUG 56
13031 +#define VC_CAT_RLIMIT 60
13033 +#define VC_CAT_SYSTEST 61
13034 +#define VC_CAT_COMPAT 63
13036 +/* query version */
13038 +#define VCMD_get_version VC_CMD(VERSION, 0, 0)
13039 +#define VCMD_get_vci VC_CMD(VERSION, 1, 0)
13044 +#include <linux/errno.h>
13046 +#endif /* __KERNEL__ */
13048 +#endif /* _VX_SWITCH_H */
13050 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/tag_cmd.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/tag_cmd.h
13051 --- linux-2.6.36/include/linux/vserver/tag_cmd.h 1970-01-01 01:00:00.000000000 +0100
13052 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/tag_cmd.h 2010-10-21 13:09:36.000000000 +0200
13054 +#ifndef _VX_TAG_CMD_H
13055 +#define _VX_TAG_CMD_H
13058 +/* vinfo commands */
13060 +#define VCMD_task_tag VC_CMD(VINFO, 3, 0)
13063 +extern int vc_task_tag(uint32_t);
13065 +#endif /* __KERNEL__ */
13067 +/* context commands */
13069 +#define VCMD_tag_migrate VC_CMD(TAGMIG, 1, 0)
13072 +extern int vc_tag_migrate(uint32_t);
13074 +#endif /* __KERNEL__ */
13075 +#endif /* _VX_TAG_CMD_H */
13076 diff -NurpP --minimal linux-2.6.36/include/linux/vserver/tag.h linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/tag.h
13077 --- linux-2.6.36/include/linux/vserver/tag.h 1970-01-01 01:00:00.000000000 +0100
13078 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/tag.h 2010-10-21 13:09:36.000000000 +0200
13083 +#include <linux/types.h>
13086 +#define DX_TAG(in) (IS_TAGGED(in))
13089 +#ifdef CONFIG_TAG_NFSD
13090 +#define DX_TAG_NFSD 1
13092 +#define DX_TAG_NFSD 0
13096 +#ifdef CONFIG_TAGGING_NONE
13098 +#define MAX_UID 0xFFFFFFFF
13099 +#define MAX_GID 0xFFFFFFFF
13101 +#define INOTAG_TAG(cond, uid, gid, tag) (0)
13103 +#define TAGINO_UID(cond, uid, tag) (uid)
13104 +#define TAGINO_GID(cond, gid, tag) (gid)
13109 +#ifdef CONFIG_TAGGING_GID16
13111 +#define MAX_UID 0xFFFFFFFF
13112 +#define MAX_GID 0x0000FFFF
13114 +#define INOTAG_TAG(cond, uid, gid, tag) \
13115 + ((cond) ? (((gid) >> 16) & 0xFFFF) : 0)
13117 +#define TAGINO_UID(cond, uid, tag) (uid)
13118 +#define TAGINO_GID(cond, gid, tag) \
13119 + ((cond) ? (((gid) & 0xFFFF) | ((tag) << 16)) : (gid))
13124 +#ifdef CONFIG_TAGGING_ID24
13126 +#define MAX_UID 0x00FFFFFF
13127 +#define MAX_GID 0x00FFFFFF
13129 +#define INOTAG_TAG(cond, uid, gid, tag) \
13130 + ((cond) ? ((((uid) >> 16) & 0xFF00) | (((gid) >> 24) & 0xFF)) : 0)
13132 +#define TAGINO_UID(cond, uid, tag) \
13133 + ((cond) ? (((uid) & 0xFFFFFF) | (((tag) & 0xFF00) << 16)) : (uid))
13134 +#define TAGINO_GID(cond, gid, tag) \
13135 + ((cond) ? (((gid) & 0xFFFFFF) | (((tag) & 0x00FF) << 24)) : (gid))
13140 +#ifdef CONFIG_TAGGING_UID16
13142 +#define MAX_UID 0x0000FFFF
13143 +#define MAX_GID 0xFFFFFFFF
13145 +#define INOTAG_TAG(cond, uid, gid, tag) \
13146 + ((cond) ? (((uid) >> 16) & 0xFFFF) : 0)
13148 +#define TAGINO_UID(cond, uid, tag) \
13149 + ((cond) ? (((uid) & 0xFFFF) | ((tag) << 16)) : (uid))
13150 +#define TAGINO_GID(cond, gid, tag) (gid)
13155 +#ifdef CONFIG_TAGGING_INTERN
13157 +#define MAX_UID 0xFFFFFFFF
13158 +#define MAX_GID 0xFFFFFFFF
13160 +#define INOTAG_TAG(cond, uid, gid, tag) \
13161 + ((cond) ? (tag) : 0)
13163 +#define TAGINO_UID(cond, uid, tag) (uid)
13164 +#define TAGINO_GID(cond, gid, tag) (gid)
13169 +#ifndef CONFIG_TAGGING_NONE
13170 +#define dx_current_fstag(sb) \
13171 + ((sb)->s_flags & MS_TAGGED ? dx_current_tag() : 0)
13173 +#define dx_current_fstag(sb) (0)
13176 +#ifndef CONFIG_TAGGING_INTERN
13177 +#define TAGINO_TAG(cond, tag) (0)
13179 +#define TAGINO_TAG(cond, tag) ((cond) ? (tag) : 0)
13182 +#define INOTAG_UID(cond, uid, gid) \
13183 + ((cond) ? ((uid) & MAX_UID) : (uid))
13184 +#define INOTAG_GID(cond, uid, gid) \
13185 + ((cond) ? ((gid) & MAX_GID) : (gid))
13188 +static inline uid_t dx_map_uid(uid_t uid)
13190 + if ((uid > MAX_UID) && (uid != -1))
13192 + return (uid & MAX_UID);
13195 +static inline gid_t dx_map_gid(gid_t gid)
13197 + if ((gid > MAX_GID) && (gid != -1))
13199 + return (gid & MAX_GID);
13207 +#define dx_notagcheck(sb) ((sb) && ((sb)->s_flags & MS_NOTAGCHECK))
13209 +int dx_parse_tag(char *string, tag_t *tag, int remove, int *mnt_flags,
13210 + unsigned long *flags);
13212 +#ifdef CONFIG_PROPAGATE
13214 +void __dx_propagate_tag(struct nameidata *nd, struct inode *inode);
13216 +#define dx_propagate_tag(n, i) __dx_propagate_tag(n, i)
13219 +#define dx_propagate_tag(n, i) do { } while (0)
13222 +#endif /* _DX_TAG_H */
13223 diff -NurpP --minimal linux-2.6.36/include/linux/vs_inet6.h linux-2.6.36-vs2.3.0.36.33/include/linux/vs_inet6.h
13224 --- linux-2.6.36/include/linux/vs_inet6.h 1970-01-01 01:00:00.000000000 +0100
13225 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vs_inet6.h 2010-10-21 13:09:36.000000000 +0200
13227 +#ifndef _VS_INET6_H
13228 +#define _VS_INET6_H
13230 +#include "vserver/base.h"
13231 +#include "vserver/network.h"
13232 +#include "vserver/debug.h"
13234 +#include <net/ipv6.h>
13236 +#define NXAV6(a) &(a)->ip, &(a)->mask, (a)->prefix, (a)->type
13237 +#define NXAV6_FMT "[%pI6/%pI6/%d:%04x]"
13240 +#ifdef CONFIG_IPV6
13243 +int v6_addr_match(struct nx_addr_v6 *nxa,
13244 + const struct in6_addr *addr, uint16_t mask)
13248 + switch (nxa->type & mask) {
13249 + case NXA_TYPE_MASK:
13250 + ret = ipv6_masked_addr_cmp(&nxa->ip, &nxa->mask, addr);
13252 + case NXA_TYPE_ADDR:
13253 + ret = ipv6_addr_equal(&nxa->ip, addr);
13255 + case NXA_TYPE_ANY:
13259 + vxdprintk(VXD_CBIT(net, 0),
13260 + "v6_addr_match(%p" NXAV6_FMT ",%pI6,%04x) = %d",
13261 + nxa, NXAV6(nxa), addr, mask, ret);
13266 +int v6_addr_in_nx_info(struct nx_info *nxi,
13267 + const struct in6_addr *addr, uint16_t mask)
13269 + struct nx_addr_v6 *nxa;
13274 + for (nxa = &nxi->v6; nxa; nxa = nxa->next)
13275 + if (v6_addr_match(nxa, addr, mask))
13279 + vxdprintk(VXD_CBIT(net, 0),
13280 + "v6_addr_in_nx_info(%p[#%u],%pI6,%04x) = %d",
13281 + nxi, nxi ? nxi->nx_id : 0, addr, mask, ret);
13286 +int v6_nx_addr_match(struct nx_addr_v6 *nxa, struct nx_addr_v6 *addr, uint16_t mask)
13288 + /* FIXME: needs full range checks */
13289 + return v6_addr_match(nxa, &addr->ip, mask);
13293 +int v6_nx_addr_in_nx_info(struct nx_info *nxi, struct nx_addr_v6 *nxa, uint16_t mask)
13295 + struct nx_addr_v6 *ptr;
13297 + for (ptr = &nxi->v6; ptr; ptr = ptr->next)
13298 + if (v6_nx_addr_match(ptr, nxa, mask))
13305 + * Check if a given address matches for a socket
13307 + * nxi: the socket's nx_info if any
13308 + * addr: to be verified address
13311 +int v6_sock_addr_match (
13312 + struct nx_info *nxi,
13313 + struct inet_sock *inet,
13314 + struct in6_addr *addr)
13316 + struct sock *sk = &inet->sk;
13317 + struct in6_addr *saddr = inet6_rcv_saddr(sk);
13319 + if (!ipv6_addr_any(addr) &&
13320 + ipv6_addr_equal(saddr, addr))
13322 + if (ipv6_addr_any(saddr))
13323 + return v6_addr_in_nx_info(nxi, addr, -1);
13328 + * check if address is covered by socket
13330 + * sk: the socket to check against
13331 + * addr: the address in question (must be != 0)
13335 +int __v6_addr_match_socket(const struct sock *sk, struct nx_addr_v6 *nxa)
13337 + struct nx_info *nxi = sk->sk_nx_info;
13338 + struct in6_addr *saddr = inet6_rcv_saddr(sk);
13340 + vxdprintk(VXD_CBIT(net, 5),
13341 + "__v6_addr_in_socket(%p," NXAV6_FMT ") %p:%pI6 %p;%lx",
13342 + sk, NXAV6(nxa), nxi, saddr, sk->sk_socket,
13343 + (sk->sk_socket?sk->sk_socket->flags:0));
13345 + if (!ipv6_addr_any(saddr)) { /* direct address match */
13346 + return v6_addr_match(nxa, saddr, -1);
13347 + } else if (nxi) { /* match against nx_info */
13348 + return v6_nx_addr_in_nx_info(nxi, nxa, -1);
13349 + } else { /* unrestricted any socket */
13355 +/* inet related checks and helpers */
13359 +struct net_device;
13363 +#include <linux/netdevice.h>
13364 +#include <linux/inetdevice.h>
13365 +#include <net/inet_timewait_sock.h>
13368 +int dev_in_nx_info(struct net_device *, struct nx_info *);
13369 +int v6_dev_in_nx_info(struct net_device *, struct nx_info *);
13370 +int nx_v6_addr_conflict(struct nx_info *, struct nx_info *);
13375 +int v6_ifa_in_nx_info(struct inet6_ifaddr *ifa, struct nx_info *nxi)
13381 + return v6_addr_in_nx_info(nxi, &ifa->addr, -1);
13385 +int nx_v6_ifa_visible(struct nx_info *nxi, struct inet6_ifaddr *ifa)
13387 + vxdprintk(VXD_CBIT(net, 1), "nx_v6_ifa_visible(%p[#%u],%p) %d",
13388 + nxi, nxi ? nxi->nx_id : 0, ifa,
13389 + nxi ? v6_ifa_in_nx_info(ifa, nxi) : 0);
13391 + if (!nx_info_flags(nxi, NXF_HIDE_NETIF, 0))
13393 + if (v6_ifa_in_nx_info(ifa, nxi))
13399 +struct nx_v6_sock_addr {
13400 + struct in6_addr saddr; /* Address used for validation */
13401 + struct in6_addr baddr; /* Address used for socket bind */
13405 +int v6_map_sock_addr(struct inet_sock *inet, struct sockaddr_in6 *addr,
13406 + struct nx_v6_sock_addr *nsa)
13408 + // struct sock *sk = &inet->sk;
13409 + // struct nx_info *nxi = sk->sk_nx_info;
13410 + struct in6_addr saddr = addr->sin6_addr;
13411 + struct in6_addr baddr = saddr;
13413 + nsa->saddr = saddr;
13414 + nsa->baddr = baddr;
13419 +void v6_set_sock_addr(struct inet_sock *inet, struct nx_v6_sock_addr *nsa)
13421 + // struct sock *sk = &inet->sk;
13422 + // struct in6_addr *saddr = inet6_rcv_saddr(sk);
13424 + // *saddr = nsa->baddr;
13425 + // inet->inet_saddr = nsa->baddr;
13429 +int nx_info_has_v6(struct nx_info *nxi)
13433 + if (NX_IPV6(nxi))
13438 +#else /* CONFIG_IPV6 */
13441 +int nx_v6_dev_visible(struct nx_info *n, struct net_device *d)
13448 +int nx_v6_addr_conflict(struct nx_info *n, uint32_t a, const struct sock *s)
13454 +int v6_ifa_in_nx_info(struct in_ifaddr *a, struct nx_info *n)
13460 +int nx_info_has_v6(struct nx_info *nxi)
13465 +#endif /* CONFIG_IPV6 */
13467 +#define current_nx_info_has_v6() \
13468 + nx_info_has_v6(current_nx_info())
13471 +#warning duplicate inclusion
13473 diff -NurpP --minimal linux-2.6.36/include/linux/vs_inet.h linux-2.6.36-vs2.3.0.36.33/include/linux/vs_inet.h
13474 --- linux-2.6.36/include/linux/vs_inet.h 1970-01-01 01:00:00.000000000 +0100
13475 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vs_inet.h 2010-10-21 14:48:51.000000000 +0200
13477 +#ifndef _VS_INET_H
13478 +#define _VS_INET_H
13480 +#include "vserver/base.h"
13481 +#include "vserver/network.h"
13482 +#include "vserver/debug.h"
13484 +#define IPI_LOOPBACK htonl(INADDR_LOOPBACK)
13486 +#define NXAV4(a) NIPQUAD((a)->ip[0]), NIPQUAD((a)->ip[1]), \
13487 + NIPQUAD((a)->mask), (a)->type
13488 +#define NXAV4_FMT "[" NIPQUAD_FMT "-" NIPQUAD_FMT "/" NIPQUAD_FMT ":%04x]"
13490 +#define NIPQUAD(addr) \
13491 + ((unsigned char *)&addr)[0], \
13492 + ((unsigned char *)&addr)[1], \
13493 + ((unsigned char *)&addr)[2], \
13494 + ((unsigned char *)&addr)[3]
13496 +#define NIPQUAD_FMT "%u.%u.%u.%u"
13500 +int v4_addr_match(struct nx_addr_v4 *nxa, __be32 addr, uint16_t tmask)
13502 + __be32 ip = nxa->ip[0].s_addr;
13503 + __be32 mask = nxa->mask.s_addr;
13504 + __be32 bcast = ip | ~mask;
13507 + switch (nxa->type & tmask) {
13508 + case NXA_TYPE_MASK:
13509 + ret = (ip == (addr & mask));
13511 + case NXA_TYPE_ADDR:
13515 + /* fall through to broadcast */
13516 + case NXA_MOD_BCAST:
13517 + ret = ((tmask & NXA_MOD_BCAST) && (addr == bcast));
13519 + case NXA_TYPE_RANGE:
13520 + ret = ((nxa->ip[0].s_addr <= addr) &&
13521 + (nxa->ip[1].s_addr > addr));
13523 + case NXA_TYPE_ANY:
13528 + vxdprintk(VXD_CBIT(net, 0),
13529 + "v4_addr_match(%p" NXAV4_FMT "," NIPQUAD_FMT ",%04x) = %d",
13530 + nxa, NXAV4(nxa), NIPQUAD(addr), tmask, ret);
13535 +int v4_addr_in_nx_info(struct nx_info *nxi, __be32 addr, uint16_t tmask)
13537 + struct nx_addr_v4 *nxa;
13544 + /* allow 127.0.0.1 when remapping lback */
13545 + if ((tmask & NXA_LOOPBACK) &&
13546 + (addr == IPI_LOOPBACK) &&
13547 + nx_info_flags(nxi, NXF_LBACK_REMAP, 0))
13550 + /* check for lback address */
13551 + if ((tmask & NXA_MOD_LBACK) &&
13552 + (nxi->v4_lback.s_addr == addr))
13555 + /* check for broadcast address */
13556 + if ((tmask & NXA_MOD_BCAST) &&
13557 + (nxi->v4_bcast.s_addr == addr))
13560 + /* check for v4 addresses */
13561 + for (nxa = &nxi->v4; nxa; nxa = nxa->next)
13562 + if (v4_addr_match(nxa, addr, tmask))
13566 + vxdprintk(VXD_CBIT(net, 0),
13567 + "v4_addr_in_nx_info(%p[#%u]," NIPQUAD_FMT ",%04x) = %d",
13568 + nxi, nxi ? nxi->nx_id : 0, NIPQUAD(addr), tmask, ret);
13573 +int v4_nx_addr_match(struct nx_addr_v4 *nxa, struct nx_addr_v4 *addr, uint16_t mask)
13575 + /* FIXME: needs full range checks */
13576 + return v4_addr_match(nxa, addr->ip[0].s_addr, mask);
13580 +int v4_nx_addr_in_nx_info(struct nx_info *nxi, struct nx_addr_v4 *nxa, uint16_t mask)
13582 + struct nx_addr_v4 *ptr;
13584 + for (ptr = &nxi->v4; ptr; ptr = ptr->next)
13585 + if (v4_nx_addr_match(ptr, nxa, mask))
13590 +#include <net/inet_sock.h>
13593 + * Check if a given address matches for a socket
13595 + * nxi: the socket's nx_info if any
13596 + * addr: to be verified address
13599 +int v4_sock_addr_match (
13600 + struct nx_info *nxi,
13601 + struct inet_sock *inet,
13604 + __be32 saddr = inet->inet_rcv_saddr;
13605 + __be32 bcast = nxi ? nxi->v4_bcast.s_addr : INADDR_BROADCAST;
13607 + if (addr && (saddr == addr || bcast == addr))
13610 + return v4_addr_in_nx_info(nxi, addr, NXA_MASK_BIND);
13615 +/* inet related checks and helpers */
13619 +struct net_device;
13622 +#ifdef CONFIG_INET
13624 +#include <linux/netdevice.h>
13625 +#include <linux/inetdevice.h>
13626 +#include <net/inet_sock.h>
13627 +#include <net/inet_timewait_sock.h>
13630 +int dev_in_nx_info(struct net_device *, struct nx_info *);
13631 +int v4_dev_in_nx_info(struct net_device *, struct nx_info *);
13632 +int nx_v4_addr_conflict(struct nx_info *, struct nx_info *);
13636 + * check if address is covered by socket
13638 + * sk: the socket to check against
13639 + * addr: the address in question (must be != 0)
13643 +int __v4_addr_match_socket(const struct sock *sk, struct nx_addr_v4 *nxa)
13645 + struct nx_info *nxi = sk->sk_nx_info;
13646 + __be32 saddr = inet_rcv_saddr(sk);
13648 + vxdprintk(VXD_CBIT(net, 5),
13649 + "__v4_addr_in_socket(%p," NXAV4_FMT ") %p:" NIPQUAD_FMT " %p;%lx",
13650 + sk, NXAV4(nxa), nxi, NIPQUAD(saddr), sk->sk_socket,
13651 + (sk->sk_socket?sk->sk_socket->flags:0));
13653 + if (saddr) { /* direct address match */
13654 + return v4_addr_match(nxa, saddr, -1);
13655 + } else if (nxi) { /* match against nx_info */
13656 + return v4_nx_addr_in_nx_info(nxi, nxa, -1);
13657 + } else { /* unrestricted any socket */
13665 +int nx_dev_visible(struct nx_info *nxi, struct net_device *dev)
13667 + vxdprintk(VXD_CBIT(net, 1), "nx_dev_visible(%p[#%u],%p »%s«) %d",
13668 + nxi, nxi ? nxi->nx_id : 0, dev, dev->name,
13669 + nxi ? dev_in_nx_info(dev, nxi) : 0);
13671 + if (!nx_info_flags(nxi, NXF_HIDE_NETIF, 0))
13673 + if (dev_in_nx_info(dev, nxi))
13680 +int v4_ifa_in_nx_info(struct in_ifaddr *ifa, struct nx_info *nxi)
13686 + return v4_addr_in_nx_info(nxi, ifa->ifa_local, NXA_MASK_SHOW);
13690 +int nx_v4_ifa_visible(struct nx_info *nxi, struct in_ifaddr *ifa)
13692 + vxdprintk(VXD_CBIT(net, 1), "nx_v4_ifa_visible(%p[#%u],%p) %d",
13693 + nxi, nxi ? nxi->nx_id : 0, ifa,
13694 + nxi ? v4_ifa_in_nx_info(ifa, nxi) : 0);
13696 + if (!nx_info_flags(nxi, NXF_HIDE_NETIF, 0))
13698 + if (v4_ifa_in_nx_info(ifa, nxi))
13704 +struct nx_v4_sock_addr {
13705 + __be32 saddr; /* Address used for validation */
13706 + __be32 baddr; /* Address used for socket bind */
13710 +int v4_map_sock_addr(struct inet_sock *inet, struct sockaddr_in *addr,
13711 + struct nx_v4_sock_addr *nsa)
13713 + struct sock *sk = &inet->sk;
13714 + struct nx_info *nxi = sk->sk_nx_info;
13715 + __be32 saddr = addr->sin_addr.s_addr;
13716 + __be32 baddr = saddr;
13718 + vxdprintk(VXD_CBIT(net, 3),
13719 + "inet_bind(%p)* %p,%p;%lx " NIPQUAD_FMT,
13720 + sk, sk->sk_nx_info, sk->sk_socket,
13721 + (sk->sk_socket ? sk->sk_socket->flags : 0),
13725 + if (saddr == INADDR_ANY) {
13726 + if (nx_info_flags(nxi, NXF_SINGLE_IP, 0))
13727 + baddr = nxi->v4.ip[0].s_addr;
13728 + } else if (saddr == IPI_LOOPBACK) {
13729 + if (nx_info_flags(nxi, NXF_LBACK_REMAP, 0))
13730 + baddr = nxi->v4_lback.s_addr;
13731 + } else { /* normal address bind */
13732 + if (!v4_addr_in_nx_info(nxi, saddr, NXA_MASK_BIND))
13733 + return -EADDRNOTAVAIL;
13737 + vxdprintk(VXD_CBIT(net, 3),
13738 + "inet_bind(%p) " NIPQUAD_FMT ", " NIPQUAD_FMT,
13739 + sk, NIPQUAD(saddr), NIPQUAD(baddr));
13741 + nsa->saddr = saddr;
13742 + nsa->baddr = baddr;
13747 +void v4_set_sock_addr(struct inet_sock *inet, struct nx_v4_sock_addr *nsa)
13749 + inet->inet_saddr = nsa->baddr;
13750 + inet->inet_rcv_saddr = nsa->baddr;
13755 + * helper to simplify inet_lookup_listener
13757 + * nxi: the socket's nx_info if any
13758 + * addr: to be verified address
13759 + * saddr: socket address
13761 +static inline int v4_inet_addr_match (
13762 + struct nx_info *nxi,
13766 + if (addr && (saddr == addr))
13769 + return nxi ? v4_addr_in_nx_info(nxi, addr, NXA_MASK_BIND) : 1;
13773 +static inline __be32 nx_map_sock_lback(struct nx_info *nxi, __be32 addr)
13775 + if (nx_info_flags(nxi, NXF_HIDE_LBACK, 0) &&
13776 + (addr == nxi->v4_lback.s_addr))
13777 + return IPI_LOOPBACK;
13782 +int nx_info_has_v4(struct nx_info *nxi)
13786 + if (NX_IPV4(nxi))
13788 + if (nx_info_flags(nxi, NXF_LBACK_REMAP, 0))
13793 +#else /* CONFIG_INET */
13796 +int nx_dev_visible(struct nx_info *n, struct net_device *d)
13802 +int nx_v4_addr_conflict(struct nx_info *n, uint32_t a, const struct sock *s)
13808 +int v4_ifa_in_nx_info(struct in_ifaddr *a, struct nx_info *n)
13814 +int nx_info_has_v4(struct nx_info *nxi)
13819 +#endif /* CONFIG_INET */
13821 +#define current_nx_info_has_v4() \
13822 + nx_info_has_v4(current_nx_info())
13825 +// #warning duplicate inclusion
13827 diff -NurpP --minimal linux-2.6.36/include/linux/vs_limit.h linux-2.6.36-vs2.3.0.36.33/include/linux/vs_limit.h
13828 --- linux-2.6.36/include/linux/vs_limit.h 1970-01-01 01:00:00.000000000 +0100
13829 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vs_limit.h 2010-10-21 13:09:36.000000000 +0200
13831 +#ifndef _VS_LIMIT_H
13832 +#define _VS_LIMIT_H
13834 +#include "vserver/limit.h"
13835 +#include "vserver/base.h"
13836 +#include "vserver/context.h"
13837 +#include "vserver/debug.h"
13838 +#include "vserver/context.h"
13839 +#include "vserver/limit_int.h"
13842 +#define vx_acc_cres(v, d, p, r) \
13843 + __vx_acc_cres(v, r, d, p, __FILE__, __LINE__)
13845 +#define vx_acc_cres_cond(x, d, p, r) \
13846 + __vx_acc_cres(((x) == vx_current_xid()) ? current_vx_info() : 0, \
13847 + r, d, p, __FILE__, __LINE__)
13850 +#define vx_add_cres(v, a, p, r) \
13851 + __vx_add_cres(v, r, a, p, __FILE__, __LINE__)
13852 +#define vx_sub_cres(v, a, p, r) vx_add_cres(v, -(a), p, r)
13854 +#define vx_add_cres_cond(x, a, p, r) \
13855 + __vx_add_cres(((x) == vx_current_xid()) ? current_vx_info() : 0, \
13856 + r, a, p, __FILE__, __LINE__)
13857 +#define vx_sub_cres_cond(x, a, p, r) vx_add_cres_cond(x, -(a), p, r)
13860 +/* process and file limits */
13862 +#define vx_nproc_inc(p) \
13863 + vx_acc_cres((p)->vx_info, 1, p, RLIMIT_NPROC)
13865 +#define vx_nproc_dec(p) \
13866 + vx_acc_cres((p)->vx_info,-1, p, RLIMIT_NPROC)
13868 +#define vx_files_inc(f) \
13869 + vx_acc_cres_cond((f)->f_xid, 1, f, RLIMIT_NOFILE)
13871 +#define vx_files_dec(f) \
13872 + vx_acc_cres_cond((f)->f_xid,-1, f, RLIMIT_NOFILE)
13874 +#define vx_locks_inc(l) \
13875 + vx_acc_cres_cond((l)->fl_xid, 1, l, RLIMIT_LOCKS)
13877 +#define vx_locks_dec(l) \
13878 + vx_acc_cres_cond((l)->fl_xid,-1, l, RLIMIT_LOCKS)
13880 +#define vx_openfd_inc(f) \
13881 + vx_acc_cres(current_vx_info(), 1, (void *)(long)(f), VLIMIT_OPENFD)
13883 +#define vx_openfd_dec(f) \
13884 + vx_acc_cres(current_vx_info(),-1, (void *)(long)(f), VLIMIT_OPENFD)
13887 +#define vx_cres_avail(v, n, r) \
13888 + __vx_cres_avail(v, r, n, __FILE__, __LINE__)
13891 +#define vx_nproc_avail(n) \
13892 + vx_cres_avail(current_vx_info(), n, RLIMIT_NPROC)
13894 +#define vx_files_avail(n) \
13895 + vx_cres_avail(current_vx_info(), n, RLIMIT_NOFILE)
13897 +#define vx_locks_avail(n) \
13898 + vx_cres_avail(current_vx_info(), n, RLIMIT_LOCKS)
13900 +#define vx_openfd_avail(n) \
13901 + vx_cres_avail(current_vx_info(), n, VLIMIT_OPENFD)
13904 +/* dentry limits */
13906 +#define vx_dentry_inc(d) do { \
13907 + if (atomic_read(&d->d_count) == 1) \
13908 + vx_acc_cres(current_vx_info(), 1, d, VLIMIT_DENTRY); \
13911 +#define vx_dentry_dec(d) do { \
13912 + if (atomic_read(&d->d_count) == 0) \
13913 + vx_acc_cres(current_vx_info(),-1, d, VLIMIT_DENTRY); \
13916 +#define vx_dentry_avail(n) \
13917 + vx_cres_avail(current_vx_info(), n, VLIMIT_DENTRY)
13920 +/* socket limits */
13922 +#define vx_sock_inc(s) \
13923 + vx_acc_cres((s)->sk_vx_info, 1, s, VLIMIT_NSOCK)
13925 +#define vx_sock_dec(s) \
13926 + vx_acc_cres((s)->sk_vx_info,-1, s, VLIMIT_NSOCK)
13928 +#define vx_sock_avail(n) \
13929 + vx_cres_avail(current_vx_info(), n, VLIMIT_NSOCK)
13932 +/* ipc resource limits */
13934 +#define vx_ipcmsg_add(v, u, a) \
13935 + vx_add_cres(v, a, u, RLIMIT_MSGQUEUE)
13937 +#define vx_ipcmsg_sub(v, u, a) \
13938 + vx_sub_cres(v, a, u, RLIMIT_MSGQUEUE)
13940 +#define vx_ipcmsg_avail(v, a) \
13941 + vx_cres_avail(v, a, RLIMIT_MSGQUEUE)
13944 +#define vx_ipcshm_add(v, k, a) \
13945 + vx_add_cres(v, a, (void *)(long)(k), VLIMIT_SHMEM)
13947 +#define vx_ipcshm_sub(v, k, a) \
13948 + vx_sub_cres(v, a, (void *)(long)(k), VLIMIT_SHMEM)
13950 +#define vx_ipcshm_avail(v, a) \
13951 + vx_cres_avail(v, a, VLIMIT_SHMEM)
13954 +#define vx_semary_inc(a) \
13955 + vx_acc_cres(current_vx_info(), 1, a, VLIMIT_SEMARY)
13957 +#define vx_semary_dec(a) \
13958 + vx_acc_cres(current_vx_info(), -1, a, VLIMIT_SEMARY)
13961 +#define vx_nsems_add(a,n) \
13962 + vx_add_cres(current_vx_info(), n, a, VLIMIT_NSEMS)
13964 +#define vx_nsems_sub(a,n) \
13965 + vx_sub_cres(current_vx_info(), n, a, VLIMIT_NSEMS)
13969 +#warning duplicate inclusion
13971 diff -NurpP --minimal linux-2.6.36/include/linux/vs_memory.h linux-2.6.36-vs2.3.0.36.33/include/linux/vs_memory.h
13972 --- linux-2.6.36/include/linux/vs_memory.h 1970-01-01 01:00:00.000000000 +0100
13973 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vs_memory.h 2010-10-21 13:09:36.000000000 +0200
13975 +#ifndef _VS_MEMORY_H
13976 +#define _VS_MEMORY_H
13978 +#include "vserver/limit.h"
13979 +#include "vserver/base.h"
13980 +#include "vserver/context.h"
13981 +#include "vserver/debug.h"
13982 +#include "vserver/context.h"
13983 +#include "vserver/limit_int.h"
13986 + VXPT_UNKNOWN = 0,
13995 +#define vx_page_fault(mm, vma, type, ret)
13999 +void __vx_page_fault(struct mm_struct *mm,
14000 + struct vm_area_struct *vma, int type, int ret)
14002 + struct vx_info *vxi = mm->mm_vx_info;
14005 + static char *page_type[6] =
14006 + { "UNKNOWN", "ANON", "NONE", "FILE", "SWAP", "WRITE" };
14007 + static char *page_what[4] =
14008 + { "FAULT_OOM", "FAULT_SIGBUS", "FAULT_MINOR", "FAULT_MAJOR" };
14014 + what = (ret & 0x3);
14016 +/* printk("[%d] page[%d][%d] %2x %s %s\n", vxi->vx_id,
14017 + type, what, ret, page_type[type], page_what[what]);
14019 + if (ret & VM_FAULT_WRITE)
14021 + atomic_inc(&vxi->cacct.page[type][what]);
14024 +#define vx_page_fault(mm, vma, type, ret) __vx_page_fault(mm, vma, type, ret)
14028 +extern unsigned long vx_badness(struct task_struct *task, struct mm_struct *mm);
14031 +#warning duplicate inclusion
14033 diff -NurpP --minimal linux-2.6.36/include/linux/vs_network.h linux-2.6.36-vs2.3.0.36.33/include/linux/vs_network.h
14034 --- linux-2.6.36/include/linux/vs_network.h 1970-01-01 01:00:00.000000000 +0100
14035 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vs_network.h 2010-10-21 13:09:36.000000000 +0200
14037 +#ifndef _NX_VS_NETWORK_H
14038 +#define _NX_VS_NETWORK_H
14040 +#include "vserver/context.h"
14041 +#include "vserver/network.h"
14042 +#include "vserver/base.h"
14043 +#include "vserver/check.h"
14044 +#include "vserver/debug.h"
14046 +#include <linux/sched.h>
14049 +#define get_nx_info(i) __get_nx_info(i, __FILE__, __LINE__)
14051 +static inline struct nx_info *__get_nx_info(struct nx_info *nxi,
14052 + const char *_file, int _line)
14057 + vxlprintk(VXD_CBIT(nid, 2), "get_nx_info(%p[#%d.%d])",
14058 + nxi, nxi ? nxi->nx_id : 0,
14059 + nxi ? atomic_read(&nxi->nx_usecnt) : 0,
14062 + atomic_inc(&nxi->nx_usecnt);
14067 +extern void free_nx_info(struct nx_info *);
14069 +#define put_nx_info(i) __put_nx_info(i, __FILE__, __LINE__)
14071 +static inline void __put_nx_info(struct nx_info *nxi, const char *_file, int _line)
14076 + vxlprintk(VXD_CBIT(nid, 2), "put_nx_info(%p[#%d.%d])",
14077 + nxi, nxi ? nxi->nx_id : 0,
14078 + nxi ? atomic_read(&nxi->nx_usecnt) : 0,
14081 + if (atomic_dec_and_test(&nxi->nx_usecnt))
14082 + free_nx_info(nxi);
14086 +#define init_nx_info(p, i) __init_nx_info(p, i, __FILE__, __LINE__)
14088 +static inline void __init_nx_info(struct nx_info **nxp, struct nx_info *nxi,
14089 + const char *_file, int _line)
14092 + vxlprintk(VXD_CBIT(nid, 3),
14093 + "init_nx_info(%p[#%d.%d])",
14094 + nxi, nxi ? nxi->nx_id : 0,
14095 + nxi ? atomic_read(&nxi->nx_usecnt) : 0,
14098 + atomic_inc(&nxi->nx_usecnt);
14104 +#define set_nx_info(p, i) __set_nx_info(p, i, __FILE__, __LINE__)
14106 +static inline void __set_nx_info(struct nx_info **nxp, struct nx_info *nxi,
14107 + const char *_file, int _line)
14109 + struct nx_info *nxo;
14114 + vxlprintk(VXD_CBIT(nid, 3), "set_nx_info(%p[#%d.%d])",
14115 + nxi, nxi ? nxi->nx_id : 0,
14116 + nxi ? atomic_read(&nxi->nx_usecnt) : 0,
14119 + atomic_inc(&nxi->nx_usecnt);
14120 + nxo = xchg(nxp, nxi);
14124 +#define clr_nx_info(p) __clr_nx_info(p, __FILE__, __LINE__)
14126 +static inline void __clr_nx_info(struct nx_info **nxp,
14127 + const char *_file, int _line)
14129 + struct nx_info *nxo;
14131 + nxo = xchg(nxp, NULL);
14135 + vxlprintk(VXD_CBIT(nid, 3), "clr_nx_info(%p[#%d.%d])",
14136 + nxo, nxo ? nxo->nx_id : 0,
14137 + nxo ? atomic_read(&nxo->nx_usecnt) : 0,
14140 + if (atomic_dec_and_test(&nxo->nx_usecnt))
14141 + free_nx_info(nxo);
14145 +#define claim_nx_info(v, p) __claim_nx_info(v, p, __FILE__, __LINE__)
14147 +static inline void __claim_nx_info(struct nx_info *nxi,
14148 + struct task_struct *task, const char *_file, int _line)
14150 + vxlprintk(VXD_CBIT(nid, 3), "claim_nx_info(%p[#%d.%d.%d]) %p",
14151 + nxi, nxi ? nxi->nx_id : 0,
14152 + nxi?atomic_read(&nxi->nx_usecnt):0,
14153 + nxi?atomic_read(&nxi->nx_tasks):0,
14154 + task, _file, _line);
14156 + atomic_inc(&nxi->nx_tasks);
14160 +extern void unhash_nx_info(struct nx_info *);
14162 +#define release_nx_info(v, p) __release_nx_info(v, p, __FILE__, __LINE__)
14164 +static inline void __release_nx_info(struct nx_info *nxi,
14165 + struct task_struct *task, const char *_file, int _line)
14167 + vxlprintk(VXD_CBIT(nid, 3), "release_nx_info(%p[#%d.%d.%d]) %p",
14168 + nxi, nxi ? nxi->nx_id : 0,
14169 + nxi ? atomic_read(&nxi->nx_usecnt) : 0,
14170 + nxi ? atomic_read(&nxi->nx_tasks) : 0,
14171 + task, _file, _line);
14175 + if (atomic_dec_and_test(&nxi->nx_tasks))
14176 + unhash_nx_info(nxi);
14180 +#define task_get_nx_info(i) __task_get_nx_info(i, __FILE__, __LINE__)
14182 +static __inline__ struct nx_info *__task_get_nx_info(struct task_struct *p,
14183 + const char *_file, int _line)
14185 + struct nx_info *nxi;
14188 + vxlprintk(VXD_CBIT(nid, 5), "task_get_nx_info(%p)",
14189 + p, _file, _line);
14190 + nxi = __get_nx_info(p->nx_info, _file, _line);
14196 +static inline void exit_nx_info(struct task_struct *p)
14199 + release_nx_info(p->nx_info, p);
14204 +#warning duplicate inclusion
14206 diff -NurpP --minimal linux-2.6.36/include/linux/vs_pid.h linux-2.6.36-vs2.3.0.36.33/include/linux/vs_pid.h
14207 --- linux-2.6.36/include/linux/vs_pid.h 1970-01-01 01:00:00.000000000 +0100
14208 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vs_pid.h 2010-10-21 13:09:36.000000000 +0200
14213 +#include "vserver/base.h"
14214 +#include "vserver/check.h"
14215 +#include "vserver/context.h"
14216 +#include "vserver/debug.h"
14217 +#include "vserver/pid.h"
14218 +#include <linux/pid_namespace.h>
14221 +#define VXF_FAKE_INIT (VXF_INFO_INIT | VXF_STATE_INIT)
14224 +int vx_proc_task_visible(struct task_struct *task)
14226 + if ((task->pid == 1) &&
14227 + !vx_flags(VXF_FAKE_INIT, VXF_FAKE_INIT))
14228 + /* show a blend through init */
14230 + if (vx_check(vx_task_xid(task), VS_WATCH | VS_IDENT))
14237 +#define find_task_by_real_pid(pid) find_task_by_pid_ns(pid, &init_pid_ns)
14242 +struct task_struct *vx_find_proc_task_by_pid(int pid)
14244 + struct task_struct *task = find_task_by_real_pid(pid);
14246 + if (task && !vx_proc_task_visible(task)) {
14247 + vxdprintk(VXD_CBIT(misc, 6),
14248 + "dropping task (find) %p[#%u,%u] for %p[#%u,%u]",
14249 + task, task->xid, task->pid,
14250 + current, current->xid, current->pid);
14259 +struct task_struct *vx_get_proc_task(struct inode *inode, struct pid *pid)
14261 + struct task_struct *task = get_pid_task(pid, PIDTYPE_PID);
14263 + if (task && !vx_proc_task_visible(task)) {
14264 + vxdprintk(VXD_CBIT(misc, 6),
14265 + "dropping task (get) %p[#%u,%u] for %p[#%u,%u]",
14266 + task, task->xid, task->pid,
14267 + current, current->xid, current->pid);
14268 + put_task_struct(task);
14277 +struct task_struct *vx_child_reaper(struct task_struct *p)
14279 + struct vx_info *vxi = p->vx_info;
14280 + struct task_struct *reaper = child_reaper(p);
14285 + BUG_ON(!p->vx_info->vx_reaper);
14287 + /* child reaper for the guest reaper */
14288 + if (vxi->vx_reaper == p)
14291 + reaper = vxi->vx_reaper;
14293 + vxdprintk(VXD_CBIT(xid, 7),
14294 + "vx_child_reaper(%p[#%u,%u]) = %p[#%u,%u]",
14295 + p, p->xid, p->pid, reaper, reaper->xid, reaper->pid);
14303 +#warning duplicate inclusion
14305 diff -NurpP --minimal linux-2.6.36/include/linux/vs_sched.h linux-2.6.36-vs2.3.0.36.33/include/linux/vs_sched.h
14306 --- linux-2.6.36/include/linux/vs_sched.h 1970-01-01 01:00:00.000000000 +0100
14307 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vs_sched.h 2010-10-21 13:09:36.000000000 +0200
14309 +#ifndef _VS_SCHED_H
14310 +#define _VS_SCHED_H
14312 +#include "vserver/base.h"
14313 +#include "vserver/context.h"
14314 +#include "vserver/sched.h"
14317 +#define VAVAVOOM_RATIO 50
14319 +#define MAX_PRIO_BIAS 20
14320 +#define MIN_PRIO_BIAS -20
14323 +#ifdef CONFIG_VSERVER_HARDCPU
14326 + * effective_prio - return the priority that is based on the static
14327 + * priority but is modified by bonuses/penalties.
14329 + * We scale the actual sleep average [0 .... MAX_SLEEP_AVG]
14330 + * into a -4 ... 0 ... +4 bonus/penalty range.
14332 + * Additionally, we scale another amount based on the number of
14333 + * CPU tokens currently held by the context, if the process is
14334 + * part of a context (and the appropriate SCHED flag is set).
14335 + * This ranges from -5 ... 0 ... +15, quadratically.
14337 + * So, the total bonus is -9 .. 0 .. +19
14338 + * We use ~50% of the full 0...39 priority range so that:
14340 + * 1) nice +19 interactive tasks do not preempt nice 0 CPU hogs.
14341 + * 2) nice -20 CPU hogs do not get preempted by nice 0 tasks.
14342 + * unless that context is far exceeding its CPU allocation.
14344 + * Both properties are important to certain workloads.
14347 +int vx_effective_vavavoom(struct _vx_sched_pc *sched_pc, int max_prio)
14349 + int vavavoom, max;
14351 + /* lots of tokens = lots of vavavoom
14352 + * no tokens = no vavavoom */
14353 + if ((vavavoom = sched_pc->tokens) >= 0) {
14354 + max = sched_pc->tokens_max;
14355 + vavavoom = max - vavavoom;
14357 + vavavoom = max_prio * VAVAVOOM_RATIO / 100
14358 + * (vavavoom*vavavoom - (max >> 2)) / max;
14366 +int vx_adjust_prio(struct task_struct *p, int prio, int max_user)
14368 + struct vx_info *vxi = p->vx_info;
14369 + struct _vx_sched_pc *sched_pc;
14374 + sched_pc = &vx_cpu(vxi, sched_pc);
14375 + if (vx_info_flags(vxi, VXF_SCHED_PRIO, 0)) {
14376 + int vavavoom = vx_effective_vavavoom(sched_pc, max_user);
14378 + sched_pc->vavavoom = vavavoom;
14379 + prio += vavavoom;
14381 + prio += sched_pc->prio_bias;
14385 +#else /* !CONFIG_VSERVER_HARDCPU */
14388 +int vx_adjust_prio(struct task_struct *p, int prio, int max_user)
14390 + struct vx_info *vxi = p->vx_info;
14393 + prio += vx_cpu(vxi, sched_pc).prio_bias;
14397 +#endif /* CONFIG_VSERVER_HARDCPU */
14400 +static inline void vx_account_user(struct vx_info *vxi,
14401 + cputime_t cputime, int nice)
14405 + vx_cpu(vxi, sched_pc).user_ticks += cputime;
14408 +static inline void vx_account_system(struct vx_info *vxi,
14409 + cputime_t cputime, int idle)
14413 + vx_cpu(vxi, sched_pc).sys_ticks += cputime;
14417 +#warning duplicate inclusion
14419 diff -NurpP --minimal linux-2.6.36/include/linux/vs_socket.h linux-2.6.36-vs2.3.0.36.33/include/linux/vs_socket.h
14420 --- linux-2.6.36/include/linux/vs_socket.h 1970-01-01 01:00:00.000000000 +0100
14421 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vs_socket.h 2010-10-21 13:09:36.000000000 +0200
14423 +#ifndef _VS_SOCKET_H
14424 +#define _VS_SOCKET_H
14426 +#include "vserver/debug.h"
14427 +#include "vserver/base.h"
14428 +#include "vserver/cacct.h"
14429 +#include "vserver/context.h"
14430 +#include "vserver/tag.h"
14433 +/* socket accounting */
14435 +#include <linux/socket.h>
14437 +static inline int vx_sock_type(int family)
14439 + switch (family) {
14441 + return VXA_SOCK_UNSPEC;
14443 + return VXA_SOCK_UNIX;
14445 + return VXA_SOCK_INET;
14447 + return VXA_SOCK_INET6;
14449 + return VXA_SOCK_PACKET;
14451 + return VXA_SOCK_OTHER;
14455 +#define vx_acc_sock(v, f, p, s) \
14456 + __vx_acc_sock(v, f, p, s, __FILE__, __LINE__)
14458 +static inline void __vx_acc_sock(struct vx_info *vxi,
14459 + int family, int pos, int size, char *file, int line)
14462 + int type = vx_sock_type(family);
14464 + atomic_long_inc(&vxi->cacct.sock[type][pos].count);
14465 + atomic_long_add(size, &vxi->cacct.sock[type][pos].total);
14469 +#define vx_sock_recv(sk, s) \
14470 + vx_acc_sock((sk)->sk_vx_info, (sk)->sk_family, 0, s)
14471 +#define vx_sock_send(sk, s) \
14472 + vx_acc_sock((sk)->sk_vx_info, (sk)->sk_family, 1, s)
14473 +#define vx_sock_fail(sk, s) \
14474 + vx_acc_sock((sk)->sk_vx_info, (sk)->sk_family, 2, s)
14477 +#define sock_vx_init(s) do { \
14478 + (s)->sk_xid = 0; \
14479 + (s)->sk_vx_info = NULL; \
14482 +#define sock_nx_init(s) do { \
14483 + (s)->sk_nid = 0; \
14484 + (s)->sk_nx_info = NULL; \
14488 +#warning duplicate inclusion
14490 diff -NurpP --minimal linux-2.6.36/include/linux/vs_tag.h linux-2.6.36-vs2.3.0.36.33/include/linux/vs_tag.h
14491 --- linux-2.6.36/include/linux/vs_tag.h 1970-01-01 01:00:00.000000000 +0100
14492 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vs_tag.h 2010-10-21 13:09:36.000000000 +0200
14497 +#include <linux/vserver/tag.h>
14499 +/* check conditions */
14501 +#define DX_ADMIN 0x0001
14502 +#define DX_WATCH 0x0002
14503 +#define DX_HOSTID 0x0008
14505 +#define DX_IDENT 0x0010
14507 +#define DX_ARG_MASK 0x0010
14510 +#define dx_task_tag(t) ((t)->tag)
14512 +#define dx_current_tag() dx_task_tag(current)
14514 +#define dx_check(c, m) __dx_check(dx_current_tag(), c, m)
14516 +#define dx_weak_check(c, m) ((m) ? dx_check(c, m) : 1)
14520 + * check current context for ADMIN/WATCH and
14521 + * optionally against supplied argument
14523 +static inline int __dx_check(tag_t cid, tag_t id, unsigned int mode)
14525 + if (mode & DX_ARG_MASK) {
14526 + if ((mode & DX_IDENT) && (id == cid))
14529 + return (((mode & DX_ADMIN) && (cid == 0)) ||
14530 + ((mode & DX_WATCH) && (cid == 1)) ||
14531 + ((mode & DX_HOSTID) && (id == 0)));
14535 +int dx_permission(const struct inode *inode, int mask);
14539 +#warning duplicate inclusion
14541 diff -NurpP --minimal linux-2.6.36/include/linux/vs_time.h linux-2.6.36-vs2.3.0.36.33/include/linux/vs_time.h
14542 --- linux-2.6.36/include/linux/vs_time.h 1970-01-01 01:00:00.000000000 +0100
14543 +++ linux-2.6.36-vs2.3.0.36.33/include/linux/vs_time.h 2010-10-21 14:25:37.000000000 +0200
14545 +#ifndef _VS_TIME_H
14546 +#define _VS_TIME_H
14549 +/* time faking stuff */
14551 +#ifdef CONFIG_VSERVER_VTIME
14553 +extern void vx_adjust_timespec(struct timespec *ts);
14554 +extern int vx_settimeofday(struct timespec *ts);
14557 +#define vx_adjust_timespec(t) do { } while (0)
14558 +#define vx_settimeofday(t) do_settimeofday(t)
14562 +#warning duplicate inclusion
14564 diff -NurpP --minimal linux-2.6.36/include/net/addrconf.h linux-2.6.36-vs2.3.0.36.33/include/net/addrconf.h
14565 --- linux-2.6.36/include/net/addrconf.h 2010-10-21 13:07:56.000000000 +0200
14566 +++ linux-2.6.36-vs2.3.0.36.33/include/net/addrconf.h 2010-10-21 13:09:36.000000000 +0200
14567 @@ -84,7 +84,8 @@ extern int ipv6_dev_get_saddr(struct n
14568 struct net_device *dev,
14569 const struct in6_addr *daddr,
14570 unsigned int srcprefs,
14571 - struct in6_addr *saddr);
14572 + struct in6_addr *saddr,
14573 + struct nx_info *nxi);
14574 extern int ipv6_get_lladdr(struct net_device *dev,
14575 struct in6_addr *addr,
14576 unsigned char banned_flags);
14577 diff -NurpP --minimal linux-2.6.36/include/net/af_unix.h linux-2.6.36-vs2.3.0.36.33/include/net/af_unix.h
14578 --- linux-2.6.36/include/net/af_unix.h 2010-10-21 13:07:56.000000000 +0200
14579 +++ linux-2.6.36-vs2.3.0.36.33/include/net/af_unix.h 2010-10-21 13:09:36.000000000 +0200
14581 #include <linux/socket.h>
14582 #include <linux/un.h>
14583 #include <linux/mutex.h>
14584 +#include <linux/vs_base.h>
14585 #include <net/sock.h>
14587 extern void unix_inflight(struct file *fp);
14588 diff -NurpP --minimal linux-2.6.36/include/net/inet_timewait_sock.h linux-2.6.36-vs2.3.0.36.33/include/net/inet_timewait_sock.h
14589 --- linux-2.6.36/include/net/inet_timewait_sock.h 2010-08-02 16:52:56.000000000 +0200
14590 +++ linux-2.6.36-vs2.3.0.36.33/include/net/inet_timewait_sock.h 2010-10-21 13:09:36.000000000 +0200
14591 @@ -117,6 +117,10 @@ struct inet_timewait_sock {
14592 #define tw_hash __tw_common.skc_hash
14593 #define tw_prot __tw_common.skc_prot
14594 #define tw_net __tw_common.skc_net
14595 +#define tw_xid __tw_common.skc_xid
14596 +#define tw_vx_info __tw_common.skc_vx_info
14597 +#define tw_nid __tw_common.skc_nid
14598 +#define tw_nx_info __tw_common.skc_nx_info
14600 volatile unsigned char tw_substate;
14601 /* 3 bits hole, try to pack */
14602 diff -NurpP --minimal linux-2.6.36/include/net/route.h linux-2.6.36-vs2.3.0.36.33/include/net/route.h
14603 --- linux-2.6.36/include/net/route.h 2010-10-21 13:07:56.000000000 +0200
14604 +++ linux-2.6.36-vs2.3.0.36.33/include/net/route.h 2010-10-21 13:09:36.000000000 +0200
14605 @@ -145,6 +145,9 @@ static inline void ip_rt_put(struct rtab
14606 dst_release(&rt->dst);
14609 +#include <linux/vs_base.h>
14610 +#include <linux/vs_inet.h>
14612 #define IPTOS_RT_MASK (IPTOS_TOS_MASK & ~3)
14614 extern const __u8 ip_tos2prio[16];
14615 @@ -154,6 +157,9 @@ static inline char rt_tos2priority(u8 to
14616 return ip_tos2prio[IPTOS_TOS(tos)>>1];
14619 +extern int ip_v4_find_src(struct net *net, struct nx_info *,
14620 + struct rtable **, struct flowi *);
14622 static inline int ip_route_connect(struct rtable **rp, __be32 dst,
14623 __be32 src, u32 tos, int oif, u8 protocol,
14624 __be16 sport, __be16 dport, struct sock *sk,
14625 @@ -171,11 +177,24 @@ static inline int ip_route_connect(struc
14628 struct net *net = sock_net(sk);
14629 + struct nx_info *nx_info = current_nx_info();
14631 if (inet_sk(sk)->transparent)
14632 fl.flags |= FLOWI_FLAG_ANYSRC;
14634 - if (!dst || !src) {
14636 + nx_info = sk->sk_nx_info;
14638 + vxdprintk(VXD_CBIT(net, 4),
14639 + "ip_route_connect(%p) %p,%p;%lx",
14640 + sk, nx_info, sk->sk_socket,
14641 + (sk->sk_socket?sk->sk_socket->flags:0));
14643 + err = ip_v4_find_src(net, nx_info, rp, &fl);
14647 + if (!fl.fl4_dst || !fl.fl4_src) {
14648 err = __ip_route_output_key(net, rp, &fl);
14651 diff -NurpP --minimal linux-2.6.36/include/net/sock.h linux-2.6.36-vs2.3.0.36.33/include/net/sock.h
14652 --- linux-2.6.36/include/net/sock.h 2010-10-21 13:07:56.000000000 +0200
14653 +++ linux-2.6.36-vs2.3.0.36.33/include/net/sock.h 2010-10-21 13:09:36.000000000 +0200
14654 @@ -150,6 +150,10 @@ struct sock_common {
14655 #ifdef CONFIG_NET_NS
14656 struct net *skc_net;
14659 + struct vx_info *skc_vx_info;
14661 + struct nx_info *skc_nx_info;
14665 @@ -241,6 +245,10 @@ struct sock {
14666 #define sk_bind_node __sk_common.skc_bind_node
14667 #define sk_prot __sk_common.skc_prot
14668 #define sk_net __sk_common.skc_net
14669 +#define sk_xid __sk_common.skc_xid
14670 +#define sk_vx_info __sk_common.skc_vx_info
14671 +#define sk_nid __sk_common.skc_nid
14672 +#define sk_nx_info __sk_common.skc_nx_info
14673 kmemcheck_bitfield_begin(flags);
14674 unsigned int sk_shutdown : 2,
14676 diff -NurpP --minimal linux-2.6.36/init/Kconfig linux-2.6.36-vs2.3.0.36.33/init/Kconfig
14677 --- linux-2.6.36/init/Kconfig 2010-10-21 13:07:56.000000000 +0200
14678 +++ linux-2.6.36-vs2.3.0.36.33/init/Kconfig 2010-10-21 13:09:36.000000000 +0200
14679 @@ -476,6 +476,7 @@ config HAVE_UNSTABLE_SCHED_CLOCK
14681 boolean "Control Group support"
14685 This option adds support for grouping sets of processes together, for
14686 use with process control subsystems such as Cpusets, CFS, memory
14687 @@ -503,6 +504,7 @@ config CGROUP_DEBUG
14689 bool "Namespace cgroup subsystem"
14693 Provides a simple namespace cgroup subsystem to
14694 provide hierarchical naming of sets of namespaces,
14695 diff -NurpP --minimal linux-2.6.36/init/main.c linux-2.6.36-vs2.3.0.36.33/init/main.c
14696 --- linux-2.6.36/init/main.c 2010-10-21 13:07:56.000000000 +0200
14697 +++ linux-2.6.36-vs2.3.0.36.33/init/main.c 2010-10-21 13:10:12.000000000 +0200
14699 #include <linux/sfi.h>
14700 #include <linux/shmem_fs.h>
14701 #include <linux/slab.h>
14702 +#include <linux/vserver/percpu.h>
14704 #include <asm/io.h>
14705 #include <asm/bugs.h>
14706 diff -NurpP --minimal linux-2.6.36/ipc/mqueue.c linux-2.6.36-vs2.3.0.36.33/ipc/mqueue.c
14707 --- linux-2.6.36/ipc/mqueue.c 2010-10-21 13:07:56.000000000 +0200
14708 +++ linux-2.6.36-vs2.3.0.36.33/ipc/mqueue.c 2010-10-21 13:28:42.000000000 +0200
14710 #include <linux/pid.h>
14711 #include <linux/ipc_namespace.h>
14712 #include <linux/slab.h>
14713 +#include <linux/vs_context.h>
14714 +#include <linux/vs_limit.h>
14716 #include <net/sock.h>
14718 @@ -66,6 +68,7 @@ struct mqueue_inode_info {
14719 struct sigevent notify;
14720 struct pid* notify_owner;
14721 struct user_struct *user; /* user who created, for accounting */
14722 + struct vx_info *vxi;
14723 struct sock *notify_sock;
14724 struct sk_buff *notify_cookie;
14726 @@ -125,6 +128,7 @@ static struct inode *mqueue_get_inode(st
14727 if (S_ISREG(mode)) {
14728 struct mqueue_inode_info *info;
14729 struct task_struct *p = current;
14730 + struct vx_info *vxi = p->vx_info;
14731 unsigned long mq_bytes, mq_msg_tblsz;
14733 inode->i_fop = &mqueue_file_operations;
14734 @@ -138,6 +142,7 @@ static struct inode *mqueue_get_inode(st
14735 info->notify_owner = NULL;
14737 info->user = NULL; /* set when all is ok */
14738 + info->vxi = NULL;
14739 memset(&info->attr, 0, sizeof(info->attr));
14740 info->attr.mq_maxmsg = ipc_ns->mq_msg_max;
14741 info->attr.mq_msgsize = ipc_ns->mq_msgsize_max;
14742 @@ -156,16 +161,19 @@ static struct inode *mqueue_get_inode(st
14743 spin_lock(&mq_lock);
14744 if (u->mq_bytes + mq_bytes < u->mq_bytes ||
14745 u->mq_bytes + mq_bytes >
14746 - task_rlimit(p, RLIMIT_MSGQUEUE)) {
14747 + task_rlimit(p, RLIMIT_MSGQUEUE) ||
14748 + !vx_ipcmsg_avail(vxi, mq_bytes)) {
14749 spin_unlock(&mq_lock);
14750 /* mqueue_evict_inode() releases info->messages */
14753 u->mq_bytes += mq_bytes;
14754 + vx_ipcmsg_add(vxi, u, mq_bytes);
14755 spin_unlock(&mq_lock);
14758 info->user = get_uid(u);
14759 + info->vxi = get_vx_info(vxi);
14760 } else if (S_ISDIR(mode)) {
14762 /* Some things misbehave if size == 0 on a directory */
14763 @@ -267,8 +275,11 @@ static void mqueue_evict_inode(struct in
14764 + info->attr.mq_msgsize);
14767 + struct vx_info *vxi = info->vxi;
14769 spin_lock(&mq_lock);
14770 user->mq_bytes -= mq_bytes;
14771 + vx_ipcmsg_sub(vxi, user, mq_bytes);
14773 * get_ns_from_inode() ensures that the
14774 * (ipc_ns = sb->s_fs_info) is either a valid ipc_ns
14775 @@ -278,6 +289,7 @@ static void mqueue_evict_inode(struct in
14777 ipc_ns->mq_queues_count--;
14778 spin_unlock(&mq_lock);
14779 + put_vx_info(vxi);
14783 diff -NurpP --minimal linux-2.6.36/ipc/msg.c linux-2.6.36-vs2.3.0.36.33/ipc/msg.c
14784 --- linux-2.6.36/ipc/msg.c 2010-08-02 16:52:57.000000000 +0200
14785 +++ linux-2.6.36-vs2.3.0.36.33/ipc/msg.c 2010-10-21 13:09:36.000000000 +0200
14787 #include <linux/rwsem.h>
14788 #include <linux/nsproxy.h>
14789 #include <linux/ipc_namespace.h>
14790 +#include <linux/vs_base.h>
14792 #include <asm/current.h>
14793 #include <asm/uaccess.h>
14794 @@ -190,6 +191,7 @@ static int newque(struct ipc_namespace *
14796 msq->q_perm.mode = msgflg & S_IRWXUGO;
14797 msq->q_perm.key = key;
14798 + msq->q_perm.xid = vx_current_xid();
14800 msq->q_perm.security = NULL;
14801 retval = security_msg_queue_alloc(msq);
14802 diff -NurpP --minimal linux-2.6.36/ipc/namespace.c linux-2.6.36-vs2.3.0.36.33/ipc/namespace.c
14803 --- linux-2.6.36/ipc/namespace.c 2009-09-10 15:26:27.000000000 +0200
14804 +++ linux-2.6.36-vs2.3.0.36.33/ipc/namespace.c 2010-10-21 13:09:36.000000000 +0200
14806 #include <linux/slab.h>
14807 #include <linux/fs.h>
14808 #include <linux/mount.h>
14809 +#include <linux/vs_base.h>
14810 +#include <linux/vserver/global.h>
14814 diff -NurpP --minimal linux-2.6.36/ipc/sem.c linux-2.6.36-vs2.3.0.36.33/ipc/sem.c
14815 --- linux-2.6.36/ipc/sem.c 2010-10-21 13:07:56.000000000 +0200
14816 +++ linux-2.6.36-vs2.3.0.36.33/ipc/sem.c 2010-10-21 13:09:36.000000000 +0200
14818 #include <linux/rwsem.h>
14819 #include <linux/nsproxy.h>
14820 #include <linux/ipc_namespace.h>
14821 +#include <linux/vs_base.h>
14822 +#include <linux/vs_limit.h>
14824 #include <asm/uaccess.h>
14826 @@ -260,6 +262,7 @@ static int newary(struct ipc_namespace *
14828 sma->sem_perm.mode = (semflg & S_IRWXUGO);
14829 sma->sem_perm.key = key;
14830 + sma->sem_perm.xid = vx_current_xid();
14832 sma->sem_perm.security = NULL;
14833 retval = security_sem_alloc(sma);
14834 @@ -275,6 +278,9 @@ static int newary(struct ipc_namespace *
14837 ns->used_sems += nsems;
14838 + /* FIXME: obsoleted? */
14839 + vx_semary_inc(sma);
14840 + vx_nsems_add(sma, nsems);
14842 sma->sem_base = (struct sem *) &sma[1];
14844 @@ -730,6 +736,9 @@ static void freeary(struct ipc_namespace
14846 wake_up_sem_queue_do(&tasks);
14847 ns->used_sems -= sma->sem_nsems;
14848 + /* FIXME: obsoleted? */
14849 + vx_nsems_sub(sma, sma->sem_nsems);
14850 + vx_semary_dec(sma);
14851 security_sem_free(sma);
14852 ipc_rcu_putref(sma);
14854 diff -NurpP --minimal linux-2.6.36/ipc/shm.c linux-2.6.36-vs2.3.0.36.33/ipc/shm.c
14855 --- linux-2.6.36/ipc/shm.c 2010-08-02 16:52:57.000000000 +0200
14856 +++ linux-2.6.36-vs2.3.0.36.33/ipc/shm.c 2010-10-21 13:09:36.000000000 +0200
14858 #include <linux/nsproxy.h>
14859 #include <linux/mount.h>
14860 #include <linux/ipc_namespace.h>
14861 +#include <linux/vs_context.h>
14862 +#include <linux/vs_limit.h>
14864 #include <asm/uaccess.h>
14866 @@ -169,7 +171,12 @@ static void shm_open(struct vm_area_stru
14868 static void shm_destroy(struct ipc_namespace *ns, struct shmid_kernel *shp)
14870 - ns->shm_tot -= (shp->shm_segsz + PAGE_SIZE - 1) >> PAGE_SHIFT;
14871 + struct vx_info *vxi = lookup_vx_info(shp->shm_perm.xid);
14872 + int numpages = (shp->shm_segsz + PAGE_SIZE - 1) >> PAGE_SHIFT;
14874 + vx_ipcshm_sub(vxi, shp, numpages);
14875 + ns->shm_tot -= numpages;
14879 if (!is_file_hugepages(shp->shm_file))
14880 @@ -179,6 +186,7 @@ static void shm_destroy(struct ipc_names
14882 fput (shp->shm_file);
14883 security_shm_free(shp);
14884 + put_vx_info(vxi);
14885 ipc_rcu_putref(shp);
14888 @@ -349,11 +357,15 @@ static int newseg(struct ipc_namespace *
14889 if (ns->shm_tot + numpages > ns->shm_ctlall)
14892 + if (!vx_ipcshm_avail(current_vx_info(), numpages))
14895 shp = ipc_rcu_alloc(sizeof(*shp));
14899 shp->shm_perm.key = key;
14900 + shp->shm_perm.xid = vx_current_xid();
14901 shp->shm_perm.mode = (shmflg & S_IRWXUGO);
14902 shp->mlock_user = NULL;
14904 @@ -407,6 +419,7 @@ static int newseg(struct ipc_namespace *
14905 ns->shm_tot += numpages;
14906 error = shp->shm_perm.id;
14908 + vx_ipcshm_add(current_vx_info(), key, numpages);
14912 diff -NurpP --minimal linux-2.6.36/kernel/capability.c linux-2.6.36-vs2.3.0.36.33/kernel/capability.c
14913 --- linux-2.6.36/kernel/capability.c 2010-08-02 16:52:57.000000000 +0200
14914 +++ linux-2.6.36-vs2.3.0.36.33/kernel/capability.c 2010-10-21 13:09:36.000000000 +0200
14916 #include <linux/security.h>
14917 #include <linux/syscalls.h>
14918 #include <linux/pid_namespace.h>
14919 +#include <linux/vs_context.h>
14920 #include <asm/uaccess.h>
14923 @@ -119,6 +120,7 @@ static int cap_validate_magic(cap_user_h
14929 * The only thing that can change the capabilities of the current
14930 * process is the current process. As such, we can't be in this code
14931 @@ -289,6 +291,8 @@ error:
14935 +#include <linux/vserver/base.h>
14938 * capable - Determine if the current task has a superior capability in effect
14939 * @cap: The capability to be tested for
14940 @@ -301,6 +305,9 @@ error:
14942 int capable(int cap)
14944 + /* here for now so we don't require task locking */
14945 + if (vs_check_bit(VXC_CAP_MASK, cap) && !vx_mcaps(1L << cap))
14947 if (unlikely(!cap_valid(cap))) {
14948 printk(KERN_CRIT "capable() called with invalid cap=%u\n", cap);
14950 diff -NurpP --minimal linux-2.6.36/kernel/compat.c linux-2.6.36-vs2.3.0.36.33/kernel/compat.c
14951 --- linux-2.6.36/kernel/compat.c 2010-10-21 13:07:56.000000000 +0200
14952 +++ linux-2.6.36-vs2.3.0.36.33/kernel/compat.c 2010-10-21 14:22:43.000000000 +0200
14953 @@ -914,7 +914,7 @@ asmlinkage long compat_sys_stime(compat_
14957 - do_settimeofday(&tv);
14958 + vx_settimeofday(&tv);
14962 diff -NurpP --minimal linux-2.6.36/kernel/exit.c linux-2.6.36-vs2.3.0.36.33/kernel/exit.c
14963 --- linux-2.6.36/kernel/exit.c 2010-10-21 13:07:56.000000000 +0200
14964 +++ linux-2.6.36-vs2.3.0.36.33/kernel/exit.c 2010-10-21 13:09:36.000000000 +0200
14966 #include <linux/fs_struct.h>
14967 #include <linux/init_task.h>
14968 #include <linux/perf_event.h>
14969 +#include <linux/vs_limit.h>
14970 +#include <linux/vs_context.h>
14971 +#include <linux/vs_network.h>
14972 +#include <linux/vs_pid.h>
14973 #include <trace/events/sched.h>
14974 #include <linux/hw_breakpoint.h>
14976 @@ -487,9 +491,11 @@ static void close_files(struct files_str
14977 filp_close(file, files);
14980 + vx_openfd_dec(i);
14988 @@ -1023,11 +1029,16 @@ NORET_TYPE void do_exit(long code)
14990 validate_creds_for_do_exit(tsk);
14992 + /* needs to stay after exit_notify() */
14993 + exit_vx_info(tsk, code);
14994 + exit_nx_info(tsk);
14998 /* causes final put_task_struct in finish_task_switch(). */
14999 tsk->state = TASK_DEAD;
15001 + printk("bad task: %p [%lx]\n", current, current->state);
15003 /* Avoid "noreturn function does return". */
15005 diff -NurpP --minimal linux-2.6.36/kernel/fork.c linux-2.6.36-vs2.3.0.36.33/kernel/fork.c
15006 --- linux-2.6.36/kernel/fork.c 2010-10-21 13:07:56.000000000 +0200
15007 +++ linux-2.6.36-vs2.3.0.36.33/kernel/fork.c 2010-10-21 13:09:36.000000000 +0200
15009 #include <linux/perf_event.h>
15010 #include <linux/posix-timers.h>
15011 #include <linux/user-return-notifier.h>
15012 +#include <linux/vs_context.h>
15013 +#include <linux/vs_network.h>
15014 +#include <linux/vs_limit.h>
15015 +#include <linux/vs_memory.h>
15017 #include <asm/pgtable.h>
15018 #include <asm/pgalloc.h>
15019 @@ -160,6 +164,8 @@ void free_task(struct task_struct *tsk)
15020 account_kernel_stack(tsk->stack, -1);
15021 free_thread_info(tsk->stack);
15022 rt_mutex_debug_task_free(tsk);
15023 + clr_vx_info(&tsk->vx_info);
15024 + clr_nx_info(&tsk->nx_info);
15025 ftrace_graph_exit_task(tsk);
15026 free_task_struct(tsk);
15028 @@ -492,6 +498,7 @@ static struct mm_struct * mm_init(struct
15029 if (likely(!mm_alloc_pgd(mm))) {
15031 mmu_notifier_mm_init(mm);
15032 + set_vx_info(&mm->mm_vx_info, p->vx_info);
15036 @@ -525,6 +532,7 @@ void __mmdrop(struct mm_struct *mm)
15038 destroy_context(mm);
15039 mmu_notifier_mm_destroy(mm);
15040 + clr_vx_info(&mm->mm_vx_info);
15043 EXPORT_SYMBOL_GPL(__mmdrop);
15044 @@ -660,6 +668,7 @@ struct mm_struct *dup_mm(struct task_str
15047 memcpy(mm, oldmm, sizeof(*mm));
15048 + mm->mm_vx_info = NULL;
15050 /* Initializing for Swap token stuff */
15051 mm->token_priority = 0;
15052 @@ -698,6 +707,7 @@ fail_nocontext:
15053 * If init_new_context() failed, we cannot use mmput() to free the mm
15054 * because it calls destroy_context()
15056 + clr_vx_info(&mm->mm_vx_info);
15060 @@ -973,6 +983,8 @@ static struct task_struct *copy_process(
15062 struct task_struct *p;
15063 int cgroup_callbacks_done = 0;
15064 + struct vx_info *vxi;
15065 + struct nx_info *nxi;
15067 if ((clone_flags & (CLONE_NEWNS|CLONE_FS)) == (CLONE_NEWNS|CLONE_FS))
15068 return ERR_PTR(-EINVAL);
15069 @@ -1019,7 +1031,12 @@ static struct task_struct *copy_process(
15070 DEBUG_LOCKS_WARN_ON(!p->hardirqs_enabled);
15071 DEBUG_LOCKS_WARN_ON(!p->softirqs_enabled);
15073 + init_vx_info(&p->vx_info, current_vx_info());
15074 + init_nx_info(&p->nx_info, current_nx_info());
15077 + if (!vx_nproc_avail(1))
15078 + goto bad_fork_free;
15079 if (atomic_read(&p->real_cred->user->processes) >=
15080 task_rlimit(p, RLIMIT_NPROC)) {
15081 if (!capable(CAP_SYS_ADMIN) && !capable(CAP_SYS_RESOURCE) &&
15082 @@ -1284,6 +1301,18 @@ static struct task_struct *copy_process(
15085 spin_unlock(¤t->sighand->siglock);
15087 + /* p is copy of current */
15088 + vxi = p->vx_info;
15090 + claim_vx_info(vxi, p);
15091 + atomic_inc(&vxi->cvirt.nr_threads);
15092 + atomic_inc(&vxi->cvirt.total_forks);
15095 + nxi = p->nx_info;
15097 + claim_nx_info(nxi, p);
15098 write_unlock_irq(&tasklist_lock);
15099 proc_fork_connector(p);
15100 cgroup_post_fork(p);
15101 diff -NurpP --minimal linux-2.6.36/kernel/kthread.c linux-2.6.36-vs2.3.0.36.33/kernel/kthread.c
15102 --- linux-2.6.36/kernel/kthread.c 2010-10-21 13:07:57.000000000 +0200
15103 +++ linux-2.6.36-vs2.3.0.36.33/kernel/kthread.c 2010-10-21 13:42:40.000000000 +0200
15105 #include <linux/mutex.h>
15106 #include <linux/slab.h>
15107 #include <linux/freezer.h>
15108 +#include <linux/vs_pid.h>
15109 #include <trace/events/sched.h>
15111 static DEFINE_SPINLOCK(kthread_create_lock);
15112 diff -NurpP --minimal linux-2.6.36/kernel/Makefile linux-2.6.36-vs2.3.0.36.33/kernel/Makefile
15113 --- linux-2.6.36/kernel/Makefile 2010-10-21 13:07:56.000000000 +0200
15114 +++ linux-2.6.36-vs2.3.0.36.33/kernel/Makefile 2010-10-21 13:09:36.000000000 +0200
15115 @@ -25,6 +25,7 @@ CFLAGS_REMOVE_sched_clock.o = -pg
15116 CFLAGS_REMOVE_perf_event.o = -pg
15120 obj-$(CONFIG_FREEZER) += freezer.o
15121 obj-$(CONFIG_PROFILING) += profile.o
15122 obj-$(CONFIG_SYSCTL_SYSCALL_CHECK) += sysctl_check.o
15123 diff -NurpP --minimal linux-2.6.36/kernel/nsproxy.c linux-2.6.36-vs2.3.0.36.33/kernel/nsproxy.c
15124 --- linux-2.6.36/kernel/nsproxy.c 2010-07-07 18:31:57.000000000 +0200
15125 +++ linux-2.6.36-vs2.3.0.36.33/kernel/nsproxy.c 2010-10-21 13:09:36.000000000 +0200
15127 #include <linux/mnt_namespace.h>
15128 #include <linux/utsname.h>
15129 #include <linux/pid_namespace.h>
15130 +#include <linux/vserver/global.h>
15131 +#include <linux/vserver/debug.h>
15132 #include <net/net_namespace.h>
15133 #include <linux/ipc_namespace.h>
15135 @@ -43,8 +45,11 @@ static inline struct nsproxy *create_nsp
15136 struct nsproxy *nsproxy;
15138 nsproxy = kmem_cache_alloc(nsproxy_cachep, GFP_KERNEL);
15141 atomic_set(&nsproxy->count, 1);
15142 + atomic_inc(&vs_global_nsproxy);
15144 + vxdprintk(VXD_CBIT(space, 2), "create_nsproxy = %p[1]", nsproxy);
15148 @@ -53,41 +58,52 @@ static inline struct nsproxy *create_nsp
15149 * Return the newly created nsproxy. Do not attach this to the task,
15150 * leave it to the caller to do proper locking and attach it to task.
15152 -static struct nsproxy *create_new_namespaces(unsigned long flags,
15153 - struct task_struct *tsk, struct fs_struct *new_fs)
15154 +static struct nsproxy *unshare_namespaces(unsigned long flags,
15155 + struct nsproxy *orig, struct fs_struct *new_fs)
15157 struct nsproxy *new_nsp;
15160 + vxdprintk(VXD_CBIT(space, 4),
15161 + "unshare_namespaces(0x%08lx,%p,%p)",
15162 + flags, orig, new_fs);
15164 new_nsp = create_nsproxy();
15166 return ERR_PTR(-ENOMEM);
15168 - new_nsp->mnt_ns = copy_mnt_ns(flags, tsk->nsproxy->mnt_ns, new_fs);
15169 + new_nsp->mnt_ns = copy_mnt_ns(flags, orig->mnt_ns, new_fs);
15170 if (IS_ERR(new_nsp->mnt_ns)) {
15171 err = PTR_ERR(new_nsp->mnt_ns);
15175 - new_nsp->uts_ns = copy_utsname(flags, tsk->nsproxy->uts_ns);
15176 + new_nsp->uts_ns = copy_utsname(flags, orig->uts_ns);
15177 if (IS_ERR(new_nsp->uts_ns)) {
15178 err = PTR_ERR(new_nsp->uts_ns);
15182 - new_nsp->ipc_ns = copy_ipcs(flags, tsk->nsproxy->ipc_ns);
15183 + new_nsp->ipc_ns = copy_ipcs(flags, orig->ipc_ns);
15184 if (IS_ERR(new_nsp->ipc_ns)) {
15185 err = PTR_ERR(new_nsp->ipc_ns);
15189 - new_nsp->pid_ns = copy_pid_ns(flags, task_active_pid_ns(tsk));
15190 + new_nsp->pid_ns = copy_pid_ns(flags, orig->pid_ns);
15191 if (IS_ERR(new_nsp->pid_ns)) {
15192 err = PTR_ERR(new_nsp->pid_ns);
15196 - new_nsp->net_ns = copy_net_ns(flags, tsk->nsproxy->net_ns);
15198 + new_nsp->user_ns = copy_user_ns(flags, orig->user_ns);
15199 + if (IS_ERR(new_nsp->user_ns)) {
15200 + err = PTR_ERR(new_nsp->user_ns);
15204 + new_nsp->net_ns = copy_net_ns(flags, orig->net_ns);
15205 if (IS_ERR(new_nsp->net_ns)) {
15206 err = PTR_ERR(new_nsp->net_ns);
15208 @@ -112,6 +128,38 @@ out_ns:
15209 return ERR_PTR(err);
15212 +static struct nsproxy *create_new_namespaces(int flags, struct task_struct *tsk,
15213 + struct fs_struct *new_fs)
15215 + return unshare_namespaces(flags, tsk->nsproxy, new_fs);
15219 + * copies the nsproxy, setting refcount to 1, and grabbing a
15220 + * reference to all contained namespaces.
15222 +struct nsproxy *copy_nsproxy(struct nsproxy *orig)
15224 + struct nsproxy *ns = create_nsproxy();
15227 + memcpy(ns, orig, sizeof(struct nsproxy));
15228 + atomic_set(&ns->count, 1);
15231 + get_mnt_ns(ns->mnt_ns);
15233 + get_uts_ns(ns->uts_ns);
15235 + get_ipc_ns(ns->ipc_ns);
15237 + get_pid_ns(ns->pid_ns);
15239 + get_net(ns->net_ns);
15245 * called from clone. This now handles copy for nsproxy and all
15246 * namespaces therein.
15247 @@ -119,9 +167,12 @@ out_ns:
15248 int copy_namespaces(unsigned long flags, struct task_struct *tsk)
15250 struct nsproxy *old_ns = tsk->nsproxy;
15251 - struct nsproxy *new_ns;
15252 + struct nsproxy *new_ns = NULL;
15255 + vxdprintk(VXD_CBIT(space, 7), "copy_namespaces(0x%08lx,%p[%p])",
15256 + flags, tsk, old_ns);
15261 @@ -131,7 +182,7 @@ int copy_namespaces(unsigned long flags,
15262 CLONE_NEWPID | CLONE_NEWNET)))
15265 - if (!capable(CAP_SYS_ADMIN)) {
15266 + if (!vx_can_unshare(CAP_SYS_ADMIN, flags)) {
15270 @@ -158,6 +209,9 @@ int copy_namespaces(unsigned long flags,
15273 put_nsproxy(old_ns);
15274 + vxdprintk(VXD_CBIT(space, 3),
15275 + "copy_namespaces(0x%08lx,%p[%p]) = %d [%p]",
15276 + flags, tsk, old_ns, err, new_ns);
15280 @@ -171,7 +225,9 @@ void free_nsproxy(struct nsproxy *ns)
15281 put_ipc_ns(ns->ipc_ns);
15283 put_pid_ns(ns->pid_ns);
15284 - put_net(ns->net_ns);
15286 + put_net(ns->net_ns);
15287 + atomic_dec(&vs_global_nsproxy);
15288 kmem_cache_free(nsproxy_cachep, ns);
15291 @@ -184,11 +240,15 @@ int unshare_nsproxy_namespaces(unsigned
15295 + vxdprintk(VXD_CBIT(space, 4),
15296 + "unshare_nsproxy_namespaces(0x%08lx,[%p])",
15297 + unshare_flags, current->nsproxy);
15299 if (!(unshare_flags & (CLONE_NEWNS | CLONE_NEWUTS | CLONE_NEWIPC |
15303 - if (!capable(CAP_SYS_ADMIN))
15304 + if (!vx_can_unshare(CAP_SYS_ADMIN, unshare_flags))
15307 *new_nsp = create_new_namespaces(unshare_flags, current,
15308 diff -NurpP --minimal linux-2.6.36/kernel/pid.c linux-2.6.36-vs2.3.0.36.33/kernel/pid.c
15309 --- linux-2.6.36/kernel/pid.c 2010-10-21 13:07:57.000000000 +0200
15310 +++ linux-2.6.36-vs2.3.0.36.33/kernel/pid.c 2010-10-21 13:09:36.000000000 +0200
15312 #include <linux/pid_namespace.h>
15313 #include <linux/init_task.h>
15314 #include <linux/syscalls.h>
15315 +#include <linux/vs_pid.h>
15317 #define pid_hashfn(nr, ns) \
15318 hash_long((unsigned long)nr + (unsigned long)ns, pidhash_shift)
15319 @@ -339,7 +340,7 @@ EXPORT_SYMBOL_GPL(find_pid_ns);
15321 struct pid *find_vpid(int nr)
15323 - return find_pid_ns(nr, current->nsproxy->pid_ns);
15324 + return find_pid_ns(vx_rmap_pid(nr), current->nsproxy->pid_ns);
15326 EXPORT_SYMBOL_GPL(find_vpid);
15328 @@ -399,6 +400,9 @@ void transfer_pid(struct task_struct *ol
15329 struct task_struct *pid_task(struct pid *pid, enum pid_type type)
15331 struct task_struct *result = NULL;
15333 + if (type == PIDTYPE_REALPID)
15334 + type = PIDTYPE_PID;
15336 struct hlist_node *first;
15337 first = rcu_dereference_check(pid->tasks[type].first,
15338 @@ -416,7 +420,7 @@ EXPORT_SYMBOL(pid_task);
15340 struct task_struct *find_task_by_pid_ns(pid_t nr, struct pid_namespace *ns)
15342 - return pid_task(find_pid_ns(nr, ns), PIDTYPE_PID);
15343 + return pid_task(find_pid_ns(vx_rmap_pid(nr), ns), PIDTYPE_PID);
15346 struct task_struct *find_task_by_vpid(pid_t vnr)
15347 @@ -458,7 +462,7 @@ struct pid *find_get_pid(pid_t nr)
15349 EXPORT_SYMBOL_GPL(find_get_pid);
15351 -pid_t pid_nr_ns(struct pid *pid, struct pid_namespace *ns)
15352 +pid_t pid_unmapped_nr_ns(struct pid *pid, struct pid_namespace *ns)
15356 @@ -471,6 +475,11 @@ pid_t pid_nr_ns(struct pid *pid, struct
15360 +pid_t pid_nr_ns(struct pid *pid, struct pid_namespace *ns)
15362 + return vx_map_pid(pid_unmapped_nr_ns(pid, ns));
15365 pid_t pid_vnr(struct pid *pid)
15367 return pid_nr_ns(pid, current->nsproxy->pid_ns);
15368 diff -NurpP --minimal linux-2.6.36/kernel/pid_namespace.c linux-2.6.36-vs2.3.0.36.33/kernel/pid_namespace.c
15369 --- linux-2.6.36/kernel/pid_namespace.c 2010-07-07 18:31:57.000000000 +0200
15370 +++ linux-2.6.36-vs2.3.0.36.33/kernel/pid_namespace.c 2010-10-21 13:09:36.000000000 +0200
15372 #include <linux/err.h>
15373 #include <linux/acct.h>
15374 #include <linux/slab.h>
15375 +#include <linux/vserver/global.h>
15377 #define BITS_PER_PAGE (PAGE_SIZE*8)
15379 @@ -87,6 +88,7 @@ static struct pid_namespace *create_pid_
15382 kref_init(&ns->kref);
15383 + atomic_inc(&vs_global_pid_ns);
15385 ns->parent = get_pid_ns(parent_pid_ns);
15387 @@ -112,6 +114,7 @@ static void destroy_pid_namespace(struct
15389 for (i = 0; i < PIDMAP_ENTRIES; i++)
15390 kfree(ns->pidmap[i].page);
15391 + atomic_dec(&vs_global_pid_ns);
15392 kmem_cache_free(pid_ns_cachep, ns);
15395 diff -NurpP --minimal linux-2.6.36/kernel/posix-timers.c linux-2.6.36-vs2.3.0.36.33/kernel/posix-timers.c
15396 --- linux-2.6.36/kernel/posix-timers.c 2010-10-21 13:07:57.000000000 +0200
15397 +++ linux-2.6.36-vs2.3.0.36.33/kernel/posix-timers.c 2010-10-21 13:09:36.000000000 +0200
15399 #include <linux/wait.h>
15400 #include <linux/workqueue.h>
15401 #include <linux/module.h>
15402 +#include <linux/vs_context.h>
15405 * Management arrays for POSIX timers. Timers are kept in slab memory
15406 @@ -363,6 +364,7 @@ int posix_timer_event(struct k_itimer *t
15408 struct task_struct *task;
15409 int shared, ret = -1;
15412 * FIXME: if ->sigq is queued we can race with
15413 * dequeue_signal()->do_schedule_next_timer().
15414 @@ -379,10 +381,18 @@ int posix_timer_event(struct k_itimer *t
15416 task = pid_task(timr->it_pid, PIDTYPE_PID);
15418 + struct vx_info_save vxis;
15419 + struct vx_info *vxi;
15421 + vxi = get_vx_info(task->vx_info);
15422 + enter_vx_info(vxi, &vxis);
15423 shared = !(timr->it_sigev_notify & SIGEV_THREAD_ID);
15424 ret = send_sigqueue(timr->sigq, task, shared);
15425 + leave_vx_info(&vxis);
15426 + put_vx_info(vxi);
15430 /* If we failed to send the signal the timer stops. */
15433 diff -NurpP --minimal linux-2.6.36/kernel/printk.c linux-2.6.36-vs2.3.0.36.33/kernel/printk.c
15434 --- linux-2.6.36/kernel/printk.c 2010-10-21 13:07:57.000000000 +0200
15435 +++ linux-2.6.36-vs2.3.0.36.33/kernel/printk.c 2010-10-21 13:43:06.000000000 +0200
15437 #include <linux/syslog.h>
15438 #include <linux/cpu.h>
15439 #include <linux/notifier.h>
15440 +#include <linux/vs_cvirt.h>
15442 #include <asm/uaccess.h>
15444 @@ -266,18 +267,15 @@ int do_syslog(int type, char __user *buf
15445 unsigned i, j, limit, count;
15451 error = security_syslog(type, from_file);
15456 - case SYSLOG_ACTION_CLOSE: /* Close log */
15458 - case SYSLOG_ACTION_OPEN: /* Open log */
15460 - case SYSLOG_ACTION_READ: /* Read from log */
15461 + if ((type == SYSLOG_ACTION_READ) ||
15462 + (type == SYSLOG_ACTION_READ_ALL) ||
15463 + (type == SYSLOG_ACTION_READ_CLEAR)) {
15465 if (!buf || len < 0)
15467 @@ -288,6 +286,16 @@ int do_syslog(int type, char __user *buf
15472 + if (!vx_check(0, VS_ADMIN|VS_WATCH))
15473 + return vx_do_syslog(type, buf, len);
15476 + case SYSLOG_ACTION_CLOSE: /* Close log */
15478 + case SYSLOG_ACTION_OPEN: /* Open log */
15480 + case SYSLOG_ACTION_READ: /* Read from log */
15481 error = wait_event_interruptible(log_wait,
15482 (log_start - log_end));
15484 @@ -314,16 +322,6 @@ int do_syslog(int type, char __user *buf
15486 /* Read last kernel messages */
15487 case SYSLOG_ACTION_READ_ALL:
15489 - if (!buf || len < 0)
15494 - if (!access_ok(VERIFY_WRITE, buf, len)) {
15499 if (count > log_buf_len)
15500 count = log_buf_len;
15501 diff -NurpP --minimal linux-2.6.36/kernel/ptrace.c linux-2.6.36-vs2.3.0.36.33/kernel/ptrace.c
15502 --- linux-2.6.36/kernel/ptrace.c 2010-10-21 13:07:57.000000000 +0200
15503 +++ linux-2.6.36-vs2.3.0.36.33/kernel/ptrace.c 2010-10-21 13:09:36.000000000 +0200
15505 #include <linux/syscalls.h>
15506 #include <linux/uaccess.h>
15507 #include <linux/regset.h>
15508 +#include <linux/vs_context.h>
15512 @@ -150,6 +151,11 @@ int __ptrace_may_access(struct task_stru
15513 dumpable = get_dumpable(task->mm);
15514 if (!dumpable && !capable(CAP_SYS_PTRACE))
15516 + if (!vx_check(task->xid, VS_ADMIN_P|VS_IDENT))
15518 + if (!vx_check(task->xid, VS_IDENT) &&
15519 + !task_vx_flags(task, VXF_STATE_ADMIN, 0))
15522 return security_ptrace_access_check(task, mode);
15524 @@ -709,6 +715,10 @@ SYSCALL_DEFINE4(ptrace, long, request, l
15529 + if (!vx_check(vx_task_xid(child), VS_WATCH_P | VS_IDENT))
15530 + goto out_put_task_struct;
15532 if (request == PTRACE_ATTACH) {
15533 ret = ptrace_attach(child);
15535 diff -NurpP --minimal linux-2.6.36/kernel/sched.c linux-2.6.36-vs2.3.0.36.33/kernel/sched.c
15536 --- linux-2.6.36/kernel/sched.c 2010-10-21 13:07:57.000000000 +0200
15537 +++ linux-2.6.36-vs2.3.0.36.33/kernel/sched.c 2010-10-21 13:09:36.000000000 +0200
15539 #include <linux/ctype.h>
15540 #include <linux/ftrace.h>
15541 #include <linux/slab.h>
15542 +#include <linux/vs_sched.h>
15543 +#include <linux/vs_cvirt.h>
15545 #include <asm/tlb.h>
15546 #include <asm/irq_regs.h>
15547 @@ -3024,9 +3026,17 @@ static inline long calc_load_fold_idle(v
15549 void get_avenrun(unsigned long *loads, unsigned long offset, int shift)
15551 - loads[0] = (avenrun[0] + offset) << shift;
15552 - loads[1] = (avenrun[1] + offset) << shift;
15553 - loads[2] = (avenrun[2] + offset) << shift;
15554 + if (vx_flags(VXF_VIRT_LOAD, 0)) {
15555 + struct vx_info *vxi = current_vx_info();
15557 + loads[0] = (vxi->cvirt.load[0] + offset) << shift;
15558 + loads[1] = (vxi->cvirt.load[1] + offset) << shift;
15559 + loads[2] = (vxi->cvirt.load[2] + offset) << shift;
15561 + loads[0] = (avenrun[0] + offset) << shift;
15562 + loads[1] = (avenrun[1] + offset) << shift;
15563 + loads[2] = (avenrun[2] + offset) << shift;
15567 static unsigned long
15568 @@ -3321,16 +3331,19 @@ void account_user_time(struct task_struc
15569 cputime_t cputime_scaled)
15571 struct cpu_usage_stat *cpustat = &kstat_this_cpu.cpustat;
15572 + struct vx_info *vxi = p->vx_info; /* p is _always_ current */
15574 + int nice = (TASK_NICE(p) > 0);
15576 /* Add user time to process. */
15577 p->utime = cputime_add(p->utime, cputime);
15578 p->utimescaled = cputime_add(p->utimescaled, cputime_scaled);
15579 + vx_account_user(vxi, cputime, nice);
15580 account_group_user_time(p, cputime);
15582 /* Add user time to cpustat. */
15583 tmp = cputime_to_cputime64(cputime);
15584 - if (TASK_NICE(p) > 0)
15586 cpustat->nice = cputime64_add(cpustat->nice, tmp);
15588 cpustat->user = cputime64_add(cpustat->user, tmp);
15589 @@ -3381,6 +3394,7 @@ void account_system_time(struct task_str
15590 cputime_t cputime, cputime_t cputime_scaled)
15592 struct cpu_usage_stat *cpustat = &kstat_this_cpu.cpustat;
15593 + struct vx_info *vxi = p->vx_info; /* p is _always_ current */
15596 if ((p->flags & PF_VCPU) && (irq_count() - hardirq_offset == 0)) {
15597 @@ -3391,6 +3405,7 @@ void account_system_time(struct task_str
15598 /* Add system time to process. */
15599 p->stime = cputime_add(p->stime, cputime);
15600 p->stimescaled = cputime_add(p->stimescaled, cputime_scaled);
15601 + vx_account_system(vxi, cputime, 0 /* do we have idle time? */);
15602 account_group_system_time(p, cputime);
15604 /* Add system time to cpustat. */
15605 @@ -4477,7 +4492,7 @@ SYSCALL_DEFINE1(nice, int, increment)
15608 if (increment < 0 && !can_nice(current, nice))
15610 + return vx_flags(VXF_IGNEG_NICE, 0) ? 0 : -EPERM;
15612 retval = security_task_setnice(current, nice);
15614 diff -NurpP --minimal linux-2.6.36/kernel/sched_fair.c linux-2.6.36-vs2.3.0.36.33/kernel/sched_fair.c
15615 --- linux-2.6.36/kernel/sched_fair.c 2010-10-21 13:07:57.000000000 +0200
15616 +++ linux-2.6.36-vs2.3.0.36.33/kernel/sched_fair.c 2010-10-21 13:09:36.000000000 +0200
15617 @@ -782,6 +782,9 @@ enqueue_entity(struct cfs_rq *cfs_rq, st
15618 check_spread(cfs_rq, se);
15619 if (se != cfs_rq->curr)
15620 __enqueue_entity(cfs_rq, se);
15622 + if (entity_is_task(se))
15623 + vx_activate_task(task_of(se));
15626 static void __clear_buddies(struct cfs_rq *cfs_rq, struct sched_entity *se)
15627 @@ -825,6 +828,8 @@ dequeue_entity(struct cfs_rq *cfs_rq, st
15629 if (se != cfs_rq->curr)
15630 __dequeue_entity(cfs_rq, se);
15631 + if (entity_is_task(se))
15632 + vx_deactivate_task(task_of(se));
15633 account_entity_dequeue(cfs_rq, se);
15634 update_min_vruntime(cfs_rq);
15636 diff -NurpP --minimal linux-2.6.36/kernel/signal.c linux-2.6.36-vs2.3.0.36.33/kernel/signal.c
15637 --- linux-2.6.36/kernel/signal.c 2010-10-21 13:07:57.000000000 +0200
15638 +++ linux-2.6.36-vs2.3.0.36.33/kernel/signal.c 2010-10-21 13:09:36.000000000 +0200
15640 #include <linux/freezer.h>
15641 #include <linux/pid_namespace.h>
15642 #include <linux/nsproxy.h>
15643 +#include <linux/vs_context.h>
15644 +#include <linux/vs_pid.h>
15645 #define CREATE_TRACE_POINTS
15646 #include <trace/events/signal.h>
15648 @@ -646,9 +648,18 @@ static int check_kill_permission(int sig
15652 + vxdprintk(VXD_CBIT(misc, 7),
15653 + "check_kill_permission(%d,%p,%p[#%u,%u])",
15654 + sig, info, t, vx_task_xid(t), t->pid);
15656 if (!valid_signal(sig))
15659 +/* FIXME: needed? if so, why?
15660 + if ((info != SEND_SIG_NOINFO) &&
15661 + (is_si_special(info) || !si_fromuser(info)))
15664 if (!si_fromuser(info))
15667 @@ -678,6 +689,20 @@ static int check_kill_permission(int sig
15672 + if (t->pid == 1 && current->xid)
15676 + /* FIXME: we shouldn't return ESRCH ever, to avoid
15677 + loops, maybe ENOENT or EACCES? */
15678 + if (!vx_check(vx_task_xid(t), VS_WATCH_P | VS_IDENT)) {
15679 + vxdprintk(current->xid || VXD_CBIT(misc, 7),
15680 + "signal %d[%p] xid mismatch %p[#%u,%u] xid=#%u",
15681 + sig, info, t, vx_task_xid(t), t->pid, current->xid);
15685 return security_task_kill(t, info, sig, 0);
15688 @@ -1170,7 +1195,7 @@ int kill_pid_info(int sig, struct siginf
15691 p = pid_task(pid, PIDTYPE_PID);
15693 + if (p && vx_check(vx_task_xid(p), VS_IDENT)) {
15694 error = group_send_sig_info(sig, info, p);
15695 if (unlikely(error == -ESRCH))
15697 @@ -1210,7 +1235,7 @@ int kill_pid_info_as_uid(int sig, struct
15700 p = pid_task(pid, PIDTYPE_PID);
15702 + if (!p || !vx_check(vx_task_xid(p), VS_IDENT)) {
15706 @@ -1265,8 +1290,10 @@ static int kill_something_info(int sig,
15707 struct task_struct * p;
15709 for_each_process(p) {
15710 - if (task_pid_vnr(p) > 1 &&
15711 - !same_thread_group(p, current)) {
15712 + if (vx_check(vx_task_xid(p), VS_ADMIN|VS_IDENT) &&
15713 + task_pid_vnr(p) > 1 &&
15714 + !same_thread_group(p, current) &&
15715 + !vx_current_initpid(p->pid)) {
15716 int err = group_send_sig_info(sig, info, p);
15719 @@ -1933,6 +1960,11 @@ relock:
15720 !sig_kernel_only(signr))
15723 + /* virtual init is protected against user signals */
15724 + if ((info->si_code == SI_USER) &&
15725 + vx_current_initpid(current->pid))
15728 if (sig_kernel_stop(signr)) {
15730 * The default action is to stop all threads in
15731 diff -NurpP --minimal linux-2.6.36/kernel/softirq.c linux-2.6.36-vs2.3.0.36.33/kernel/softirq.c
15732 --- linux-2.6.36/kernel/softirq.c 2010-08-02 16:52:57.000000000 +0200
15733 +++ linux-2.6.36-vs2.3.0.36.33/kernel/softirq.c 2010-10-21 13:09:36.000000000 +0200
15735 #include <linux/ftrace.h>
15736 #include <linux/smp.h>
15737 #include <linux/tick.h>
15738 +#include <linux/vs_context.h>
15740 #define CREATE_TRACE_POINTS
15741 #include <trace/events/irq.h>
15742 diff -NurpP --minimal linux-2.6.36/kernel/sys.c linux-2.6.36-vs2.3.0.36.33/kernel/sys.c
15743 --- linux-2.6.36/kernel/sys.c 2010-10-21 13:07:57.000000000 +0200
15744 +++ linux-2.6.36-vs2.3.0.36.33/kernel/sys.c 2010-10-21 13:45:48.000000000 +0200
15746 #include <linux/syscalls.h>
15747 #include <linux/kprobes.h>
15748 #include <linux/user_namespace.h>
15749 +#include <linux/vs_pid.h>
15751 #include <asm/uaccess.h>
15752 #include <asm/io.h>
15753 @@ -131,7 +132,10 @@ static int set_one_prio(struct task_stru
15756 if (niceval < task_nice(p) && !can_nice(p, niceval)) {
15758 + if (vx_flags(VXF_IGNEG_NICE, 0))
15764 no_nice = security_task_setnice(p, niceval);
15765 @@ -181,6 +185,8 @@ SYSCALL_DEFINE3(setpriority, int, which,
15767 pgrp = task_pgrp(current);
15768 do_each_pid_thread(pgrp, PIDTYPE_PGID, p) {
15769 + if (!vx_check(p->xid, VS_ADMIN_P | VS_IDENT))
15771 error = set_one_prio(p, niceval, error);
15772 } while_each_pid_thread(pgrp, PIDTYPE_PGID, p);
15774 @@ -244,6 +250,8 @@ SYSCALL_DEFINE2(getpriority, int, which,
15776 pgrp = task_pgrp(current);
15777 do_each_pid_thread(pgrp, PIDTYPE_PGID, p) {
15778 + if (!vx_check(p->xid, VS_ADMIN_P | VS_IDENT))
15780 niceval = 20 - task_nice(p);
15781 if (niceval > retval)
15783 @@ -357,6 +365,8 @@ EXPORT_SYMBOL_GPL(kernel_power_off);
15785 static DEFINE_MUTEX(reboot_mutex);
15787 +long vs_reboot(unsigned int, void __user *);
15790 * Reboot system call: for obvious reasons only root may call it,
15791 * and even root needs to set up some magic numbers in the registers
15792 @@ -389,6 +399,9 @@ SYSCALL_DEFINE4(reboot, int, magic1, int
15793 if ((cmd == LINUX_REBOOT_CMD_POWER_OFF) && !pm_power_off)
15794 cmd = LINUX_REBOOT_CMD_HALT;
15796 + if (!vx_check(0, VS_ADMIN|VS_WATCH))
15797 + return vs_reboot(cmd, arg);
15799 mutex_lock(&reboot_mutex);
15801 case LINUX_REBOOT_CMD_RESTART:
15802 @@ -1169,7 +1182,7 @@ SYSCALL_DEFINE2(sethostname, char __user
15804 char tmp[__NEW_UTS_LEN];
15806 - if (!capable(CAP_SYS_ADMIN))
15807 + if (!vx_capable(CAP_SYS_ADMIN, VXC_SET_UTSNAME))
15809 if (len < 0 || len > __NEW_UTS_LEN)
15811 @@ -1218,7 +1231,7 @@ SYSCALL_DEFINE2(setdomainname, char __us
15813 char tmp[__NEW_UTS_LEN];
15815 - if (!capable(CAP_SYS_ADMIN))
15816 + if (!vx_capable(CAP_SYS_ADMIN, VXC_SET_UTSNAME))
15818 if (len < 0 || len > __NEW_UTS_LEN)
15820 @@ -1334,7 +1347,7 @@ int do_prlimit(struct task_struct *tsk,
15821 task_lock(tsk->group_leader);
15823 if (new_rlim->rlim_max > rlim->rlim_max &&
15824 - !capable(CAP_SYS_RESOURCE))
15825 + !vx_capable(CAP_SYS_RESOURCE, VXC_SET_RLIMIT))
15828 retval = security_task_setrlimit(tsk->group_leader,
15829 @@ -1383,7 +1396,7 @@ static int check_prlimit_permission(stru
15830 cred->gid != tcred->egid ||
15831 cred->gid != tcred->sgid ||
15832 cred->gid != tcred->gid) &&
15833 - !capable(CAP_SYS_RESOURCE)) {
15834 + !vx_capable(CAP_SYS_RESOURCE, VXC_SET_RLIMIT)) {
15838 diff -NurpP --minimal linux-2.6.36/kernel/sysctl_binary.c linux-2.6.36-vs2.3.0.36.33/kernel/sysctl_binary.c
15839 --- linux-2.6.36/kernel/sysctl_binary.c 2010-08-02 16:52:57.000000000 +0200
15840 +++ linux-2.6.36-vs2.3.0.36.33/kernel/sysctl_binary.c 2010-10-21 13:09:36.000000000 +0200
15841 @@ -73,6 +73,7 @@ static const struct bin_table bin_kern_t
15843 { CTL_INT, KERN_PANIC, "panic" },
15844 { CTL_INT, KERN_REALROOTDEV, "real-root-dev" },
15845 + { CTL_STR, KERN_VSHELPER, "vshelper" },
15847 { CTL_STR, KERN_SPARC_REBOOT, "reboot-cmd" },
15848 { CTL_INT, KERN_CTLALTDEL, "ctrl-alt-del" },
15849 diff -NurpP --minimal linux-2.6.36/kernel/sysctl.c linux-2.6.36-vs2.3.0.36.33/kernel/sysctl.c
15850 --- linux-2.6.36/kernel/sysctl.c 2010-10-21 13:07:57.000000000 +0200
15851 +++ linux-2.6.36-vs2.3.0.36.33/kernel/sysctl.c 2010-10-21 13:09:36.000000000 +0200
15853 #if defined(CONFIG_PROVE_LOCKING) || defined(CONFIG_LOCK_STAT)
15854 #include <linux/lockdep.h>
15856 +extern char vshelper_path[];
15857 #ifdef CONFIG_CHR_DEV_SG
15858 #include <scsi/sg.h>
15860 @@ -573,6 +574,13 @@ static struct ctl_table kern_table[] = {
15861 .proc_handler = proc_dostring,
15865 + .procname = "vshelper",
15866 + .data = &vshelper_path,
15869 + .proc_handler = &proc_dostring,
15871 #ifdef CONFIG_CHR_DEV_SG
15873 .procname = "sg-big-buff",
15874 diff -NurpP --minimal linux-2.6.36/kernel/time/timekeeping.c linux-2.6.36-vs2.3.0.36.33/kernel/time/timekeeping.c
15875 --- linux-2.6.36/kernel/time/timekeeping.c 2010-10-21 13:07:57.000000000 +0200
15876 +++ linux-2.6.36-vs2.3.0.36.33/kernel/time/timekeeping.c 2010-10-21 14:17:59.000000000 +0200
15877 @@ -229,6 +229,7 @@ void getnstimeofday(struct timespec *ts)
15878 } while (read_seqretry(&xtime_lock, seq));
15880 timespec_add_ns(ts, nsecs);
15881 + vx_adjust_timespec(ts);
15884 EXPORT_SYMBOL(getnstimeofday);
15885 diff -NurpP --minimal linux-2.6.36/kernel/time.c linux-2.6.36-vs2.3.0.36.33/kernel/time.c
15886 --- linux-2.6.36/kernel/time.c 2010-10-21 13:07:57.000000000 +0200
15887 +++ linux-2.6.36-vs2.3.0.36.33/kernel/time.c 2010-10-21 14:22:43.000000000 +0200
15888 @@ -92,7 +92,7 @@ SYSCALL_DEFINE1(stime, time_t __user *,
15892 - do_settimeofday(&tv);
15893 + vx_settimeofday(&tv);
15897 @@ -177,7 +177,7 @@ int do_sys_settimeofday(struct timespec
15898 /* SMP safe, again the code in arch/foo/time.c should
15899 * globally block out interrupts when it runs.
15901 - return do_settimeofday(tv);
15902 + return vx_settimeofday(tv);
15906 diff -NurpP --minimal linux-2.6.36/kernel/timer.c linux-2.6.36-vs2.3.0.36.33/kernel/timer.c
15907 --- linux-2.6.36/kernel/timer.c 2010-10-21 13:07:57.000000000 +0200
15908 +++ linux-2.6.36-vs2.3.0.36.33/kernel/timer.c 2010-10-21 13:09:36.000000000 +0200
15910 #include <linux/perf_event.h>
15911 #include <linux/sched.h>
15912 #include <linux/slab.h>
15913 +#include <linux/vs_base.h>
15914 +#include <linux/vs_cvirt.h>
15915 +#include <linux/vs_pid.h>
15916 +#include <linux/vserver/sched.h>
15918 #include <asm/uaccess.h>
15919 #include <asm/unistd.h>
15920 @@ -1332,12 +1336,6 @@ SYSCALL_DEFINE1(alarm, unsigned int, sec
15927 - * The Alpha uses getxpid, getxuid, and getxgid instead. Maybe this
15928 - * should be moved into arch/i386 instead?
15932 * sys_getpid - return the thread group id of the current process
15933 @@ -1366,10 +1364,23 @@ SYSCALL_DEFINE0(getppid)
15935 pid = task_tgid_vnr(current->real_parent);
15937 + return vx_map_pid(pid);
15944 + * The Alpha uses getxpid, getxuid, and getxgid instead.
15947 +asmlinkage long do_getxpid(long *ppid)
15949 + *ppid = sys_getppid();
15950 + return sys_getpid();
15953 +#else /* _alpha_ */
15955 SYSCALL_DEFINE0(getuid)
15957 /* Only we change this so SMP safe */
15958 diff -NurpP --minimal linux-2.6.36/kernel/user_namespace.c linux-2.6.36-vs2.3.0.36.33/kernel/user_namespace.c
15959 --- linux-2.6.36/kernel/user_namespace.c 2010-10-21 13:07:57.000000000 +0200
15960 +++ linux-2.6.36-vs2.3.0.36.33/kernel/user_namespace.c 2010-10-21 13:09:36.000000000 +0200
15962 #include <linux/user_namespace.h>
15963 #include <linux/highuid.h>
15964 #include <linux/cred.h>
15965 +#include <linux/vserver/global.h>
15968 * Create a new user namespace, deriving the creator from the user in the
15969 @@ -31,6 +32,7 @@ int create_user_ns(struct cred *new)
15972 kref_init(&ns->kref);
15973 + atomic_inc(&vs_global_user_ns);
15975 for (n = 0; n < UIDHASH_SZ; ++n)
15976 INIT_HLIST_HEAD(ns->uidhash_table + n);
15977 @@ -79,6 +81,8 @@ void free_user_ns(struct kref *kref)
15978 struct user_namespace *ns =
15979 container_of(kref, struct user_namespace, kref);
15981 + /* FIXME: maybe move into destroyer? */
15982 + atomic_dec(&vs_global_user_ns);
15983 INIT_WORK(&ns->destroyer, free_user_ns_work);
15984 schedule_work(&ns->destroyer);
15986 diff -NurpP --minimal linux-2.6.36/kernel/utsname.c linux-2.6.36-vs2.3.0.36.33/kernel/utsname.c
15987 --- linux-2.6.36/kernel/utsname.c 2009-09-10 15:26:28.000000000 +0200
15988 +++ linux-2.6.36-vs2.3.0.36.33/kernel/utsname.c 2010-10-21 13:09:36.000000000 +0200
15989 @@ -14,14 +14,17 @@
15990 #include <linux/utsname.h>
15991 #include <linux/err.h>
15992 #include <linux/slab.h>
15993 +#include <linux/vserver/global.h>
15995 static struct uts_namespace *create_uts_ns(void)
15997 struct uts_namespace *uts_ns;
15999 uts_ns = kmalloc(sizeof(struct uts_namespace), GFP_KERNEL);
16002 kref_init(&uts_ns->kref);
16003 + atomic_inc(&vs_global_uts_ns);
16008 @@ -71,5 +74,6 @@ void free_uts_ns(struct kref *kref)
16009 struct uts_namespace *ns;
16011 ns = container_of(kref, struct uts_namespace, kref);
16012 + atomic_dec(&vs_global_uts_ns);
16015 diff -NurpP --minimal linux-2.6.36/kernel/vserver/cacct.c linux-2.6.36-vs2.3.0.36.33/kernel/vserver/cacct.c
16016 --- linux-2.6.36/kernel/vserver/cacct.c 1970-01-01 01:00:00.000000000 +0100
16017 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/cacct.c 2010-10-21 13:09:36.000000000 +0200
16020 + * linux/kernel/vserver/cacct.c
16022 + * Virtual Server: Context Accounting
16024 + * Copyright (C) 2006-2007 Herbert Pötzl
16026 + * V0.01 added accounting stats
16030 +#include <linux/types.h>
16031 +#include <linux/vs_context.h>
16032 +#include <linux/vserver/cacct_cmd.h>
16033 +#include <linux/vserver/cacct_int.h>
16035 +#include <asm/errno.h>
16036 +#include <asm/uaccess.h>
16039 +int vc_sock_stat(struct vx_info *vxi, void __user *data)
16041 + struct vcmd_sock_stat_v0 vc_data;
16044 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
16047 + field = vc_data.field;
16048 + if ((field < 0) || (field >= VXA_SOCK_SIZE))
16051 + for (j = 0; j < 3; j++) {
16052 + vc_data.count[j] = vx_sock_count(&vxi->cacct, field, j);
16053 + vc_data.total[j] = vx_sock_total(&vxi->cacct, field, j);
16056 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
16061 diff -NurpP --minimal linux-2.6.36/kernel/vserver/cacct_init.h linux-2.6.36-vs2.3.0.36.33/kernel/vserver/cacct_init.h
16062 --- linux-2.6.36/kernel/vserver/cacct_init.h 1970-01-01 01:00:00.000000000 +0100
16063 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/cacct_init.h 2010-10-21 13:09:36.000000000 +0200
16067 +static inline void vx_info_init_cacct(struct _vx_cacct *cacct)
16072 + for (i = 0; i < VXA_SOCK_SIZE; i++) {
16073 + for (j = 0; j < 3; j++) {
16074 + atomic_long_set(&cacct->sock[i][j].count, 0);
16075 + atomic_long_set(&cacct->sock[i][j].total, 0);
16078 + for (i = 0; i < 8; i++)
16079 + atomic_set(&cacct->slab[i], 0);
16080 + for (i = 0; i < 5; i++)
16081 + for (j = 0; j < 4; j++)
16082 + atomic_set(&cacct->page[i][j], 0);
16085 +static inline void vx_info_exit_cacct(struct _vx_cacct *cacct)
16090 diff -NurpP --minimal linux-2.6.36/kernel/vserver/cacct_proc.h linux-2.6.36-vs2.3.0.36.33/kernel/vserver/cacct_proc.h
16091 --- linux-2.6.36/kernel/vserver/cacct_proc.h 1970-01-01 01:00:00.000000000 +0100
16092 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/cacct_proc.h 2010-10-21 13:09:36.000000000 +0200
16094 +#ifndef _VX_CACCT_PROC_H
16095 +#define _VX_CACCT_PROC_H
16097 +#include <linux/vserver/cacct_int.h>
16100 +#define VX_SOCKA_TOP \
16101 + "Type\t recv #/bytes\t\t send #/bytes\t\t fail #/bytes\n"
16103 +static inline int vx_info_proc_cacct(struct _vx_cacct *cacct, char *buffer)
16105 + int i, j, length = 0;
16106 + static char *type[VXA_SOCK_SIZE] = {
16107 + "UNSPEC", "UNIX", "INET", "INET6", "PACKET", "OTHER"
16110 + length += sprintf(buffer + length, VX_SOCKA_TOP);
16111 + for (i = 0; i < VXA_SOCK_SIZE; i++) {
16112 + length += sprintf(buffer + length, "%s:", type[i]);
16113 + for (j = 0; j < 3; j++) {
16114 + length += sprintf(buffer + length,
16115 + "\t%10lu/%-10lu",
16116 + vx_sock_count(cacct, i, j),
16117 + vx_sock_total(cacct, i, j));
16119 + buffer[length++] = '\n';
16122 + length += sprintf(buffer + length, "\n");
16123 + length += sprintf(buffer + length,
16124 + "slab:\t %8u %8u %8u %8u\n",
16125 + atomic_read(&cacct->slab[1]),
16126 + atomic_read(&cacct->slab[4]),
16127 + atomic_read(&cacct->slab[0]),
16128 + atomic_read(&cacct->slab[2]));
16130 + length += sprintf(buffer + length, "\n");
16131 + for (i = 0; i < 5; i++) {
16132 + length += sprintf(buffer + length,
16133 + "page[%d]: %8u %8u %8u %8u\t %8u %8u %8u %8u\n", i,
16134 + atomic_read(&cacct->page[i][0]),
16135 + atomic_read(&cacct->page[i][1]),
16136 + atomic_read(&cacct->page[i][2]),
16137 + atomic_read(&cacct->page[i][3]),
16138 + atomic_read(&cacct->page[i][4]),
16139 + atomic_read(&cacct->page[i][5]),
16140 + atomic_read(&cacct->page[i][6]),
16141 + atomic_read(&cacct->page[i][7]));
16146 +#endif /* _VX_CACCT_PROC_H */
16147 diff -NurpP --minimal linux-2.6.36/kernel/vserver/context.c linux-2.6.36-vs2.3.0.36.33/kernel/vserver/context.c
16148 --- linux-2.6.36/kernel/vserver/context.c 1970-01-01 01:00:00.000000000 +0100
16149 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/context.c 2010-10-21 14:39:59.000000000 +0200
16152 + * linux/kernel/vserver/context.c
16154 + * Virtual Server: Context Support
16156 + * Copyright (C) 2003-2007 Herbert Pötzl
16158 + * V0.01 context helper
16159 + * V0.02 vx_ctx_kill syscall command
16160 + * V0.03 replaced context_info calls
16161 + * V0.04 redesign of struct (de)alloc
16162 + * V0.05 rlimit basic implementation
16163 + * V0.06 task_xid and info commands
16164 + * V0.07 context flags and caps
16165 + * V0.08 switch to RCU based hash
16166 + * V0.09 revert to non RCU for now
16167 + * V0.10 and back to working RCU hash
16168 + * V0.11 and back to locking again
16169 + * V0.12 referenced context store
16170 + * V0.13 separate per cpu data
16171 + * V0.14 changed vcmds to vxi arg
16172 + * V0.15 added context stat
16173 + * V0.16 have __create claim() the vxi
16174 + * V0.17 removed older and legacy stuff
16178 +#include <linux/slab.h>
16179 +#include <linux/types.h>
16180 +#include <linux/security.h>
16181 +#include <linux/pid_namespace.h>
16183 +#include <linux/vserver/context.h>
16184 +#include <linux/vserver/network.h>
16185 +#include <linux/vserver/debug.h>
16186 +#include <linux/vserver/limit.h>
16187 +#include <linux/vserver/limit_int.h>
16188 +#include <linux/vserver/space.h>
16189 +#include <linux/init_task.h>
16190 +#include <linux/fs_struct.h>
16192 +#include <linux/vs_context.h>
16193 +#include <linux/vs_limit.h>
16194 +#include <linux/vs_pid.h>
16195 +#include <linux/vserver/context_cmd.h>
16197 +#include "cvirt_init.h"
16198 +#include "cacct_init.h"
16199 +#include "limit_init.h"
16200 +#include "sched_init.h"
16203 +atomic_t vx_global_ctotal = ATOMIC_INIT(0);
16204 +atomic_t vx_global_cactive = ATOMIC_INIT(0);
16207 +/* now inactive context structures */
16209 +static struct hlist_head vx_info_inactive = HLIST_HEAD_INIT;
16211 +static spinlock_t vx_info_inactive_lock = SPIN_LOCK_UNLOCKED;
16214 +/* __alloc_vx_info()
16216 + * allocate an initialized vx_info struct
16217 + * doesn't make it visible (hash) */
16219 +static struct vx_info *__alloc_vx_info(xid_t xid)
16221 + struct vx_info *new = NULL;
16224 + vxdprintk(VXD_CBIT(xid, 0), "alloc_vx_info(%d)*", xid);
16226 + /* would this benefit from a slab cache? */
16227 + new = kmalloc(sizeof(struct vx_info), GFP_KERNEL);
16231 + memset(new, 0, sizeof(struct vx_info));
16233 + new->ptr_pc = alloc_percpu(struct _vx_info_pc);
16234 + if (!new->ptr_pc)
16237 + new->vx_id = xid;
16238 + INIT_HLIST_NODE(&new->vx_hlist);
16239 + atomic_set(&new->vx_usecnt, 0);
16240 + atomic_set(&new->vx_tasks, 0);
16241 + new->vx_parent = NULL;
16242 + new->vx_state = 0;
16243 + init_waitqueue_head(&new->vx_wait);
16245 + /* prepare reaper */
16246 + get_task_struct(init_pid_ns.child_reaper);
16247 + new->vx_reaper = init_pid_ns.child_reaper;
16248 + new->vx_badness_bias = 0;
16250 + /* rest of init goes here */
16251 + vx_info_init_limit(&new->limit);
16252 + vx_info_init_sched(&new->sched);
16253 + vx_info_init_cvirt(&new->cvirt);
16254 + vx_info_init_cacct(&new->cacct);
16256 + /* per cpu data structures */
16257 + for_each_possible_cpu(cpu) {
16258 + vx_info_init_sched_pc(
16259 + &vx_per_cpu(new, sched_pc, cpu), cpu);
16260 + vx_info_init_cvirt_pc(
16261 + &vx_per_cpu(new, cvirt_pc, cpu), cpu);
16264 + new->vx_flags = VXF_INIT_SET;
16265 + cap_set_init_eff(new->vx_bcaps);
16266 + new->vx_ccaps = 0;
16267 + new->vx_umask = 0;
16269 + new->reboot_cmd = 0;
16270 + new->exit_code = 0;
16272 + // preconfig fs entries
16273 + for (index = 0; index < VX_SPACES; index++) {
16274 + spin_lock(&init_fs.lock);
16276 + spin_unlock(&init_fs.lock);
16277 + new->vx_fs[index] = &init_fs;
16280 + vxdprintk(VXD_CBIT(xid, 0),
16281 + "alloc_vx_info(%d) = %p", xid, new);
16282 + vxh_alloc_vx_info(new);
16283 + atomic_inc(&vx_global_ctotal);
16292 +/* __dealloc_vx_info()
16294 + * final disposal of vx_info */
16296 +static void __dealloc_vx_info(struct vx_info *vxi)
16298 +#ifdef CONFIG_VSERVER_WARN
16299 + struct vx_info_save vxis;
16302 + vxdprintk(VXD_CBIT(xid, 0),
16303 + "dealloc_vx_info(%p)", vxi);
16304 + vxh_dealloc_vx_info(vxi);
16306 +#ifdef CONFIG_VSERVER_WARN
16307 + enter_vx_info(vxi, &vxis);
16308 + vx_info_exit_limit(&vxi->limit);
16309 + vx_info_exit_sched(&vxi->sched);
16310 + vx_info_exit_cvirt(&vxi->cvirt);
16311 + vx_info_exit_cacct(&vxi->cacct);
16313 + for_each_possible_cpu(cpu) {
16314 + vx_info_exit_sched_pc(
16315 + &vx_per_cpu(vxi, sched_pc, cpu), cpu);
16316 + vx_info_exit_cvirt_pc(
16317 + &vx_per_cpu(vxi, cvirt_pc, cpu), cpu);
16319 + leave_vx_info(&vxis);
16323 + vxi->vx_state |= VXS_RELEASED;
16326 + free_percpu(vxi->ptr_pc);
16329 + atomic_dec(&vx_global_ctotal);
16332 +static void __shutdown_vx_info(struct vx_info *vxi)
16334 + struct nsproxy *nsproxy;
16335 + struct fs_struct *fs;
16340 + vxi->vx_state |= VXS_SHUTDOWN;
16341 + vs_state_change(vxi, VSC_SHUTDOWN);
16343 + for (index = 0; index < VX_SPACES; index++) {
16344 + nsproxy = xchg(&vxi->vx_nsproxy[index], NULL);
16346 + put_nsproxy(nsproxy);
16348 + fs = xchg(&vxi->vx_fs[index], NULL);
16349 + spin_lock(&fs->lock);
16350 + kill = !--fs->users;
16351 + spin_unlock(&fs->lock);
16353 + free_fs_struct(fs);
16357 +/* exported stuff */
16359 +void free_vx_info(struct vx_info *vxi)
16361 + unsigned long flags;
16364 + /* check for reference counts first */
16365 + BUG_ON(atomic_read(&vxi->vx_usecnt));
16366 + BUG_ON(atomic_read(&vxi->vx_tasks));
16368 + /* context must not be hashed */
16369 + BUG_ON(vx_info_state(vxi, VXS_HASHED));
16371 + /* context shutdown is mandatory */
16372 + BUG_ON(!vx_info_state(vxi, VXS_SHUTDOWN));
16374 + /* nsproxy and fs check */
16375 + for (index = 0; index < VX_SPACES; index++) {
16376 + BUG_ON(vxi->vx_nsproxy[index]);
16377 + BUG_ON(vxi->vx_fs[index]);
16380 + spin_lock_irqsave(&vx_info_inactive_lock, flags);
16381 + hlist_del(&vxi->vx_hlist);
16382 + spin_unlock_irqrestore(&vx_info_inactive_lock, flags);
16384 + __dealloc_vx_info(vxi);
16388 +/* hash table for vx_info hash */
16390 +#define VX_HASH_SIZE 13
16392 +static struct hlist_head vx_info_hash[VX_HASH_SIZE] =
16393 + { [0 ... VX_HASH_SIZE-1] = HLIST_HEAD_INIT };
16395 +static spinlock_t vx_info_hash_lock = SPIN_LOCK_UNLOCKED;
16398 +static inline unsigned int __hashval(xid_t xid)
16400 + return (xid % VX_HASH_SIZE);
16405 +/* __hash_vx_info()
16407 + * add the vxi to the global hash table
16408 + * requires the hash_lock to be held */
16410 +static inline void __hash_vx_info(struct vx_info *vxi)
16412 + struct hlist_head *head;
16414 + vxd_assert_lock(&vx_info_hash_lock);
16415 + vxdprintk(VXD_CBIT(xid, 4),
16416 + "__hash_vx_info: %p[#%d]", vxi, vxi->vx_id);
16417 + vxh_hash_vx_info(vxi);
16419 + /* context must not be hashed */
16420 + BUG_ON(vx_info_state(vxi, VXS_HASHED));
16422 + vxi->vx_state |= VXS_HASHED;
16423 + head = &vx_info_hash[__hashval(vxi->vx_id)];
16424 + hlist_add_head(&vxi->vx_hlist, head);
16425 + atomic_inc(&vx_global_cactive);
16428 +/* __unhash_vx_info()
16430 + * remove the vxi from the global hash table
16431 + * requires the hash_lock to be held */
16433 +static inline void __unhash_vx_info(struct vx_info *vxi)
16435 + unsigned long flags;
16437 + vxd_assert_lock(&vx_info_hash_lock);
16438 + vxdprintk(VXD_CBIT(xid, 4),
16439 + "__unhash_vx_info: %p[#%d.%d.%d]", vxi, vxi->vx_id,
16440 + atomic_read(&vxi->vx_usecnt), atomic_read(&vxi->vx_tasks));
16441 + vxh_unhash_vx_info(vxi);
16443 + /* context must be hashed */
16444 + BUG_ON(!vx_info_state(vxi, VXS_HASHED));
16445 + /* but without tasks */
16446 + BUG_ON(atomic_read(&vxi->vx_tasks));
16448 + vxi->vx_state &= ~VXS_HASHED;
16449 + hlist_del_init(&vxi->vx_hlist);
16450 + spin_lock_irqsave(&vx_info_inactive_lock, flags);
16451 + hlist_add_head(&vxi->vx_hlist, &vx_info_inactive);
16452 + spin_unlock_irqrestore(&vx_info_inactive_lock, flags);
16453 + atomic_dec(&vx_global_cactive);
16457 +/* __lookup_vx_info()
16459 + * requires the hash_lock to be held
16460 + * doesn't increment the vx_refcnt */
16462 +static inline struct vx_info *__lookup_vx_info(xid_t xid)
16464 + struct hlist_head *head = &vx_info_hash[__hashval(xid)];
16465 + struct hlist_node *pos;
16466 + struct vx_info *vxi;
16468 + vxd_assert_lock(&vx_info_hash_lock);
16469 + hlist_for_each(pos, head) {
16470 + vxi = hlist_entry(pos, struct vx_info, vx_hlist);
16472 + if (vxi->vx_id == xid)
16477 + vxdprintk(VXD_CBIT(xid, 0),
16478 + "__lookup_vx_info(#%u): %p[#%u]",
16479 + xid, vxi, vxi ? vxi->vx_id : 0);
16480 + vxh_lookup_vx_info(vxi, xid);
16485 +/* __create_vx_info()
16487 + * create the requested context
16488 + * get(), claim() and hash it */
16490 +static struct vx_info *__create_vx_info(int id)
16492 + struct vx_info *new, *vxi = NULL;
16494 + vxdprintk(VXD_CBIT(xid, 1), "create_vx_info(%d)*", id);
16496 + if (!(new = __alloc_vx_info(id)))
16497 + return ERR_PTR(-ENOMEM);
16499 + /* required to make dynamic xids unique */
16500 + spin_lock(&vx_info_hash_lock);
16502 + /* static context requested */
16503 + if ((vxi = __lookup_vx_info(id))) {
16504 + vxdprintk(VXD_CBIT(xid, 0),
16505 + "create_vx_info(%d) = %p (already there)", id, vxi);
16506 + if (vx_info_flags(vxi, VXF_STATE_SETUP, 0))
16507 + vxi = ERR_PTR(-EBUSY);
16509 + vxi = ERR_PTR(-EEXIST);
16512 + /* new context */
16513 + vxdprintk(VXD_CBIT(xid, 0),
16514 + "create_vx_info(%d) = %p (new)", id, new);
16515 + claim_vx_info(new, NULL);
16516 + __hash_vx_info(get_vx_info(new));
16517 + vxi = new, new = NULL;
16520 + spin_unlock(&vx_info_hash_lock);
16521 + vxh_create_vx_info(IS_ERR(vxi) ? NULL : vxi, id);
16523 + __dealloc_vx_info(new);
16528 +/* exported stuff */
16531 +void unhash_vx_info(struct vx_info *vxi)
16533 + __shutdown_vx_info(vxi);
16534 + spin_lock(&vx_info_hash_lock);
16535 + __unhash_vx_info(vxi);
16536 + spin_unlock(&vx_info_hash_lock);
16537 + __wakeup_vx_info(vxi);
16541 +/* lookup_vx_info()
16543 + * search for a vx_info and get() it
16544 + * negative id means current */
16546 +struct vx_info *lookup_vx_info(int id)
16548 + struct vx_info *vxi = NULL;
16551 + vxi = get_vx_info(current_vx_info());
16552 + } else if (id > 1) {
16553 + spin_lock(&vx_info_hash_lock);
16554 + vxi = get_vx_info(__lookup_vx_info(id));
16555 + spin_unlock(&vx_info_hash_lock);
16560 +/* xid_is_hashed()
16562 + * verify that xid is still hashed */
16564 +int xid_is_hashed(xid_t xid)
16568 + spin_lock(&vx_info_hash_lock);
16569 + hashed = (__lookup_vx_info(xid) != NULL);
16570 + spin_unlock(&vx_info_hash_lock);
16574 +#ifdef CONFIG_PROC_FS
16578 + * get a subset of hashed xids for proc
16579 + * assumes size is at least one */
16581 +int get_xid_list(int index, unsigned int *xids, int size)
16583 + int hindex, nr_xids = 0;
16585 + /* only show current and children */
16586 + if (!vx_check(0, VS_ADMIN | VS_WATCH)) {
16589 + xids[nr_xids] = vx_current_xid();
16593 + for (hindex = 0; hindex < VX_HASH_SIZE; hindex++) {
16594 + struct hlist_head *head = &vx_info_hash[hindex];
16595 + struct hlist_node *pos;
16597 + spin_lock(&vx_info_hash_lock);
16598 + hlist_for_each(pos, head) {
16599 + struct vx_info *vxi;
16604 + vxi = hlist_entry(pos, struct vx_info, vx_hlist);
16605 + xids[nr_xids] = vxi->vx_id;
16606 + if (++nr_xids >= size) {
16607 + spin_unlock(&vx_info_hash_lock);
16611 + /* keep the lock time short */
16612 + spin_unlock(&vx_info_hash_lock);
16619 +#ifdef CONFIG_VSERVER_DEBUG
16621 +void dump_vx_info_inactive(int level)
16623 + struct hlist_node *entry, *next;
16625 + hlist_for_each_safe(entry, next, &vx_info_inactive) {
16626 + struct vx_info *vxi =
16627 + list_entry(entry, struct vx_info, vx_hlist);
16629 + dump_vx_info(vxi, level);
16636 +int vx_migrate_user(struct task_struct *p, struct vx_info *vxi)
16638 + struct user_struct *new_user, *old_user;
16643 + if (vx_info_flags(vxi, VXF_INFO_PRIVATE, 0))
16646 + new_user = alloc_uid(vxi->vx_id, p->uid);
16650 + old_user = p->user;
16651 + if (new_user != old_user) {
16652 + atomic_inc(&new_user->processes);
16653 + atomic_dec(&old_user->processes);
16654 + p->user = new_user;
16656 + free_uid(old_user);
16662 +void vx_mask_cap_bset(struct vx_info *vxi, struct task_struct *p)
16664 + // p->cap_effective &= vxi->vx_cap_bset;
16665 + p->cap_effective =
16666 + cap_intersect(p->cap_effective, vxi->cap_bset);
16667 + // p->cap_inheritable &= vxi->vx_cap_bset;
16668 + p->cap_inheritable =
16669 + cap_intersect(p->cap_inheritable, vxi->cap_bset);
16670 + // p->cap_permitted &= vxi->vx_cap_bset;
16671 + p->cap_permitted =
16672 + cap_intersect(p->cap_permitted, vxi->cap_bset);
16677 +#include <linux/file.h>
16678 +#include <linux/fdtable.h>
16680 +static int vx_openfd_task(struct task_struct *tsk)
16682 + struct files_struct *files = tsk->files;
16683 + struct fdtable *fdt;
16684 + const unsigned long *bptr;
16685 + int count, total;
16687 + /* no rcu_read_lock() because of spin_lock() */
16688 + spin_lock(&files->file_lock);
16689 + fdt = files_fdtable(files);
16690 + bptr = fdt->open_fds->fds_bits;
16691 + count = fdt->max_fds / (sizeof(unsigned long) * 8);
16692 + for (total = 0; count > 0; count--) {
16694 + total += hweight_long(*bptr);
16697 + spin_unlock(&files->file_lock);
16702 +/* for *space compatibility */
16704 +asmlinkage long sys_unshare(unsigned long);
16707 + * migrate task to new context
16708 + * gets vxi, puts old_vxi on change
16709 + * optionally unshares namespaces (hack)
16712 +int vx_migrate_task(struct task_struct *p, struct vx_info *vxi, int unshare)
16714 + struct vx_info *old_vxi;
16720 + vxdprintk(VXD_CBIT(xid, 5),
16721 + "vx_migrate_task(%p,%p[#%d.%d])", p, vxi,
16722 + vxi->vx_id, atomic_read(&vxi->vx_usecnt));
16724 + if (vx_info_flags(vxi, VXF_INFO_PRIVATE, 0) &&
16725 + !vx_info_flags(vxi, VXF_STATE_SETUP, 0))
16728 + if (vx_info_state(vxi, VXS_SHUTDOWN))
16731 + old_vxi = task_get_vx_info(p);
16732 + if (old_vxi == vxi)
16735 +// if (!(ret = vx_migrate_user(p, vxi))) {
16740 + openfd = vx_openfd_task(p);
16743 + atomic_dec(&old_vxi->cvirt.nr_threads);
16744 + atomic_dec(&old_vxi->cvirt.nr_running);
16745 + __rlim_dec(&old_vxi->limit, RLIMIT_NPROC);
16746 + /* FIXME: what about the struct files here? */
16747 + __rlim_sub(&old_vxi->limit, VLIMIT_OPENFD, openfd);
16748 + /* account for the executable */
16749 + __rlim_dec(&old_vxi->limit, VLIMIT_DENTRY);
16751 + atomic_inc(&vxi->cvirt.nr_threads);
16752 + atomic_inc(&vxi->cvirt.nr_running);
16753 + __rlim_inc(&vxi->limit, RLIMIT_NPROC);
16754 + /* FIXME: what about the struct files here? */
16755 + __rlim_add(&vxi->limit, VLIMIT_OPENFD, openfd);
16756 + /* account for the executable */
16757 + __rlim_inc(&vxi->limit, VLIMIT_DENTRY);
16760 + release_vx_info(old_vxi, p);
16761 + clr_vx_info(&p->vx_info);
16763 + claim_vx_info(vxi, p);
16764 + set_vx_info(&p->vx_info, vxi);
16765 + p->xid = vxi->vx_id;
16767 + vxdprintk(VXD_CBIT(xid, 5),
16768 + "moved task %p into vxi:%p[#%d]",
16769 + p, vxi, vxi->vx_id);
16771 + // vx_mask_cap_bset(vxi, p);
16774 + /* hack for *spaces to provide compatibility */
16776 + struct nsproxy *old_nsp, *new_nsp;
16778 + ret = unshare_nsproxy_namespaces(
16779 + CLONE_NEWUTS | CLONE_NEWIPC | CLONE_NEWUSER,
16784 + old_nsp = xchg(&p->nsproxy, new_nsp);
16785 + vx_set_space(vxi,
16786 + CLONE_NEWUTS | CLONE_NEWIPC | CLONE_NEWUSER, 0);
16787 + put_nsproxy(old_nsp);
16791 + put_vx_info(old_vxi);
16795 +int vx_set_reaper(struct vx_info *vxi, struct task_struct *p)
16797 + struct task_struct *old_reaper;
16802 + vxdprintk(VXD_CBIT(xid, 6),
16803 + "vx_set_reaper(%p[#%d],%p[#%d,%d])",
16804 + vxi, vxi->vx_id, p, p->xid, p->pid);
16806 + old_reaper = vxi->vx_reaper;
16807 + if (old_reaper == p)
16810 + /* set new child reaper */
16811 + get_task_struct(p);
16812 + vxi->vx_reaper = p;
16813 + put_task_struct(old_reaper);
16817 +int vx_set_init(struct vx_info *vxi, struct task_struct *p)
16822 + vxdprintk(VXD_CBIT(xid, 6),
16823 + "vx_set_init(%p[#%d],%p[#%d,%d,%d])",
16824 + vxi, vxi->vx_id, p, p->xid, p->pid, p->tgid);
16826 + vxi->vx_flags &= ~VXF_STATE_INIT;
16827 + // vxi->vx_initpid = p->tgid;
16828 + vxi->vx_initpid = p->pid;
16832 +void vx_exit_init(struct vx_info *vxi, struct task_struct *p, int code)
16834 + vxdprintk(VXD_CBIT(xid, 6),
16835 + "vx_exit_init(%p[#%d],%p[#%d,%d,%d])",
16836 + vxi, vxi->vx_id, p, p->xid, p->pid, p->tgid);
16838 + vxi->exit_code = code;
16839 + vxi->vx_initpid = 0;
16843 +void vx_set_persistent(struct vx_info *vxi)
16845 + vxdprintk(VXD_CBIT(xid, 6),
16846 + "vx_set_persistent(%p[#%d])", vxi, vxi->vx_id);
16848 + get_vx_info(vxi);
16849 + claim_vx_info(vxi, NULL);
16852 +void vx_clear_persistent(struct vx_info *vxi)
16854 + vxdprintk(VXD_CBIT(xid, 6),
16855 + "vx_clear_persistent(%p[#%d])", vxi, vxi->vx_id);
16857 + release_vx_info(vxi, NULL);
16858 + put_vx_info(vxi);
16861 +void vx_update_persistent(struct vx_info *vxi)
16863 + if (vx_info_flags(vxi, VXF_PERSISTENT, 0))
16864 + vx_set_persistent(vxi);
16866 + vx_clear_persistent(vxi);
16870 +/* task must be current or locked */
16872 +void exit_vx_info(struct task_struct *p, int code)
16874 + struct vx_info *vxi = p->vx_info;
16877 + atomic_dec(&vxi->cvirt.nr_threads);
16880 + vxi->exit_code = code;
16881 + release_vx_info(vxi, p);
16885 +void exit_vx_info_early(struct task_struct *p, int code)
16887 + struct vx_info *vxi = p->vx_info;
16890 + if (vxi->vx_initpid == p->pid)
16891 + vx_exit_init(vxi, p, code);
16892 + if (vxi->vx_reaper == p)
16893 + vx_set_reaper(vxi, init_pid_ns.child_reaper);
16898 +/* vserver syscall commands below here */
16900 +/* taks xid and vx_info functions */
16902 +#include <asm/uaccess.h>
16905 +int vc_task_xid(uint32_t id)
16910 + struct task_struct *tsk;
16912 + read_lock(&tasklist_lock);
16913 + tsk = find_task_by_real_pid(id);
16914 + xid = (tsk) ? tsk->xid : -ESRCH;
16915 + read_unlock(&tasklist_lock);
16917 + xid = vx_current_xid();
16922 +int vc_vx_info(struct vx_info *vxi, void __user *data)
16924 + struct vcmd_vx_info_v0 vc_data;
16926 + vc_data.xid = vxi->vx_id;
16927 + vc_data.initpid = vxi->vx_initpid;
16929 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
16935 +int vc_ctx_stat(struct vx_info *vxi, void __user *data)
16937 + struct vcmd_ctx_stat_v0 vc_data;
16939 + vc_data.usecnt = atomic_read(&vxi->vx_usecnt);
16940 + vc_data.tasks = atomic_read(&vxi->vx_tasks);
16942 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
16948 +/* context functions */
16950 +int vc_ctx_create(uint32_t xid, void __user *data)
16952 + struct vcmd_ctx_create vc_data = { .flagword = VXF_INIT_SET };
16953 + struct vx_info *new_vxi;
16956 + if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
16959 + if ((xid > MAX_S_CONTEXT) || (xid < 2))
16962 + new_vxi = __create_vx_info(xid);
16963 + if (IS_ERR(new_vxi))
16964 + return PTR_ERR(new_vxi);
16966 + /* initial flags */
16967 + new_vxi->vx_flags = vc_data.flagword;
16970 + if (vs_state_change(new_vxi, VSC_STARTUP))
16973 + ret = vx_migrate_task(current, new_vxi, (!data));
16977 + /* return context id on success */
16978 + ret = new_vxi->vx_id;
16980 + /* get a reference for persistent contexts */
16981 + if ((vc_data.flagword & VXF_PERSISTENT))
16982 + vx_set_persistent(new_vxi);
16984 + release_vx_info(new_vxi, NULL);
16985 + put_vx_info(new_vxi);
16990 +int vc_ctx_migrate(struct vx_info *vxi, void __user *data)
16992 + struct vcmd_ctx_migrate vc_data = { .flagword = 0 };
16995 + if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
16998 + ret = vx_migrate_task(current, vxi, 0);
17001 + if (vc_data.flagword & VXM_SET_INIT)
17002 + ret = vx_set_init(vxi, current);
17005 + if (vc_data.flagword & VXM_SET_REAPER)
17006 + ret = vx_set_reaper(vxi, current);
17011 +int vc_get_cflags(struct vx_info *vxi, void __user *data)
17013 + struct vcmd_ctx_flags_v0 vc_data;
17015 + vc_data.flagword = vxi->vx_flags;
17017 + /* special STATE flag handling */
17018 + vc_data.mask = vs_mask_flags(~0ULL, vxi->vx_flags, VXF_ONE_TIME);
17020 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
17025 +int vc_set_cflags(struct vx_info *vxi, void __user *data)
17027 + struct vcmd_ctx_flags_v0 vc_data;
17028 + uint64_t mask, trigger;
17030 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
17033 + /* special STATE flag handling */
17034 + mask = vs_mask_mask(vc_data.mask, vxi->vx_flags, VXF_ONE_TIME);
17035 + trigger = (mask & vxi->vx_flags) ^ (mask & vc_data.flagword);
17037 + if (vxi == current_vx_info()) {
17038 + /* if (trigger & VXF_STATE_SETUP)
17039 + vx_mask_cap_bset(vxi, current); */
17040 + if (trigger & VXF_STATE_INIT) {
17043 + ret = vx_set_init(vxi, current);
17046 + ret = vx_set_reaper(vxi, current);
17052 + vxi->vx_flags = vs_mask_flags(vxi->vx_flags,
17053 + vc_data.flagword, mask);
17054 + if (trigger & VXF_PERSISTENT)
17055 + vx_update_persistent(vxi);
17061 +static inline uint64_t caps_from_cap_t(kernel_cap_t c)
17063 + uint64_t v = c.cap[0] | ((uint64_t)c.cap[1] << 32);
17065 + // printk("caps_from_cap_t(%08x:%08x) = %016llx\n", c.cap[1], c.cap[0], v);
17069 +static inline kernel_cap_t cap_t_from_caps(uint64_t v)
17071 + kernel_cap_t c = __cap_empty_set;
17073 + c.cap[0] = v & 0xFFFFFFFF;
17074 + c.cap[1] = (v >> 32) & 0xFFFFFFFF;
17076 + // printk("cap_t_from_caps(%016llx) = %08x:%08x\n", v, c.cap[1], c.cap[0]);
17081 +static int do_get_caps(struct vx_info *vxi, uint64_t *bcaps, uint64_t *ccaps)
17084 + *bcaps = caps_from_cap_t(vxi->vx_bcaps);
17086 + *ccaps = vxi->vx_ccaps;
17091 +int vc_get_ccaps(struct vx_info *vxi, void __user *data)
17093 + struct vcmd_ctx_caps_v1 vc_data;
17096 + ret = do_get_caps(vxi, NULL, &vc_data.ccaps);
17099 + vc_data.cmask = ~0ULL;
17101 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
17106 +static int do_set_caps(struct vx_info *vxi,
17107 + uint64_t bcaps, uint64_t bmask, uint64_t ccaps, uint64_t cmask)
17109 + uint64_t bcold = caps_from_cap_t(vxi->vx_bcaps);
17112 + printk("do_set_caps(%16llx, %16llx, %16llx, %16llx)\n",
17113 + bcaps, bmask, ccaps, cmask);
17115 + vxi->vx_bcaps = cap_t_from_caps(
17116 + vs_mask_flags(bcold, bcaps, bmask));
17117 + vxi->vx_ccaps = vs_mask_flags(vxi->vx_ccaps, ccaps, cmask);
17122 +int vc_set_ccaps(struct vx_info *vxi, void __user *data)
17124 + struct vcmd_ctx_caps_v1 vc_data;
17126 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
17129 + return do_set_caps(vxi, 0, 0, vc_data.ccaps, vc_data.cmask);
17132 +int vc_get_bcaps(struct vx_info *vxi, void __user *data)
17134 + struct vcmd_bcaps vc_data;
17137 + ret = do_get_caps(vxi, &vc_data.bcaps, NULL);
17140 + vc_data.bmask = ~0ULL;
17142 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
17147 +int vc_set_bcaps(struct vx_info *vxi, void __user *data)
17149 + struct vcmd_bcaps vc_data;
17151 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
17154 + return do_set_caps(vxi, vc_data.bcaps, vc_data.bmask, 0, 0);
17158 +int vc_get_umask(struct vx_info *vxi, void __user *data)
17160 + struct vcmd_umask vc_data;
17162 + vc_data.umask = vxi->vx_umask;
17163 + vc_data.mask = ~0ULL;
17165 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
17170 +int vc_set_umask(struct vx_info *vxi, void __user *data)
17172 + struct vcmd_umask vc_data;
17174 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
17177 + vxi->vx_umask = vs_mask_flags(vxi->vx_umask,
17178 + vc_data.umask, vc_data.mask);
17183 +int vc_get_badness(struct vx_info *vxi, void __user *data)
17185 + struct vcmd_badness_v0 vc_data;
17187 + vc_data.bias = vxi->vx_badness_bias;
17189 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
17194 +int vc_set_badness(struct vx_info *vxi, void __user *data)
17196 + struct vcmd_badness_v0 vc_data;
17198 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
17201 + vxi->vx_badness_bias = vc_data.bias;
17205 +#include <linux/module.h>
17207 +EXPORT_SYMBOL_GPL(free_vx_info);
17209 diff -NurpP --minimal linux-2.6.36/kernel/vserver/cvirt.c linux-2.6.36-vs2.3.0.36.33/kernel/vserver/cvirt.c
17210 --- linux-2.6.36/kernel/vserver/cvirt.c 1970-01-01 01:00:00.000000000 +0100
17211 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/cvirt.c 2010-10-21 14:47:08.000000000 +0200
17214 + * linux/kernel/vserver/cvirt.c
17216 + * Virtual Server: Context Virtualization
17218 + * Copyright (C) 2004-2007 Herbert Pötzl
17220 + * V0.01 broken out from limit.c
17221 + * V0.02 added utsname stuff
17222 + * V0.03 changed vcmds to vxi arg
17226 +#include <linux/types.h>
17227 +#include <linux/utsname.h>
17228 +#include <linux/vs_cvirt.h>
17229 +#include <linux/vserver/switch.h>
17230 +#include <linux/vserver/cvirt_cmd.h>
17232 +#include <asm/uaccess.h>
17235 +void vx_vsi_uptime(struct timespec *uptime, struct timespec *idle)
17237 + struct vx_info *vxi = current_vx_info();
17239 + set_normalized_timespec(uptime,
17240 + uptime->tv_sec - vxi->cvirt.bias_uptime.tv_sec,
17241 + uptime->tv_nsec - vxi->cvirt.bias_uptime.tv_nsec);
17244 + set_normalized_timespec(idle,
17245 + idle->tv_sec - vxi->cvirt.bias_idle.tv_sec,
17246 + idle->tv_nsec - vxi->cvirt.bias_idle.tv_nsec);
17250 +uint64_t vx_idle_jiffies(void)
17252 + return init_task.utime + init_task.stime;
17257 +static inline uint32_t __update_loadavg(uint32_t load,
17258 + int wsize, int delta, int n)
17260 + unsigned long long calc, prev;
17262 + /* just set it to n */
17263 + if (unlikely(delta >= wsize))
17264 + return (n << FSHIFT);
17266 + calc = delta * n;
17268 + prev = (wsize - delta);
17271 + do_div(calc, wsize);
17276 +void vx_update_load(struct vx_info *vxi)
17278 + uint32_t now, last, delta;
17279 + unsigned int nr_running, nr_uninterruptible;
17280 + unsigned int total;
17281 + unsigned long flags;
17283 + spin_lock_irqsave(&vxi->cvirt.load_lock, flags);
17286 + last = vxi->cvirt.load_last;
17287 + delta = now - last;
17289 + if (delta < 5*HZ)
17292 + nr_running = atomic_read(&vxi->cvirt.nr_running);
17293 + nr_uninterruptible = atomic_read(&vxi->cvirt.nr_uninterruptible);
17294 + total = nr_running + nr_uninterruptible;
17296 + vxi->cvirt.load[0] = __update_loadavg(vxi->cvirt.load[0],
17297 + 60*HZ, delta, total);
17298 + vxi->cvirt.load[1] = __update_loadavg(vxi->cvirt.load[1],
17299 + 5*60*HZ, delta, total);
17300 + vxi->cvirt.load[2] = __update_loadavg(vxi->cvirt.load[2],
17301 + 15*60*HZ, delta, total);
17303 + vxi->cvirt.load_last = now;
17305 + atomic_inc(&vxi->cvirt.load_updates);
17306 + spin_unlock_irqrestore(&vxi->cvirt.load_lock, flags);
17311 + * Commands to do_syslog:
17313 + * 0 -- Close the log. Currently a NOP.
17314 + * 1 -- Open the log. Currently a NOP.
17315 + * 2 -- Read from the log.
17316 + * 3 -- Read all messages remaining in the ring buffer.
17317 + * 4 -- Read and clear all messages remaining in the ring buffer
17318 + * 5 -- Clear ring buffer.
17319 + * 6 -- Disable printk's to console
17320 + * 7 -- Enable printk's to console
17321 + * 8 -- Set level of messages printed to console
17322 + * 9 -- Return number of unread characters in the log buffer
17323 + * 10 -- Return size of the log buffer
17325 +int vx_do_syslog(int type, char __user *buf, int len)
17328 + int do_clear = 0;
17329 + struct vx_info *vxi = current_vx_info();
17330 + struct _vx_syslog *log;
17334 + log = &vxi->cvirt.syslog;
17337 + case 0: /* Close log */
17338 + case 1: /* Open log */
17340 + case 2: /* Read from log */
17341 + error = wait_event_interruptible(log->log_wait,
17342 + (log->log_start - log->log_end));
17345 + spin_lock_irq(&log->logbuf_lock);
17346 + spin_unlock_irq(&log->logbuf_lock);
17348 + case 4: /* Read/clear last kernel messages */
17350 + /* fall through */
17351 + case 3: /* Read last kernel messages */
17354 + case 5: /* Clear ring buffer */
17357 + case 6: /* Disable logging to console */
17358 + case 7: /* Enable logging to console */
17359 + case 8: /* Set level of messages printed to console */
17362 + case 9: /* Number of chars in the log buffer */
17364 + case 10: /* Size of the log buffer */
17374 +/* virtual host info names */
17376 +static char *vx_vhi_name(struct vx_info *vxi, int id)
17378 + struct nsproxy *nsproxy;
17379 + struct uts_namespace *uts;
17381 + if (id == VHIN_CONTEXT)
17382 + return vxi->vx_name;
17384 + nsproxy = vxi->vx_nsproxy[0];
17388 + uts = nsproxy->uts_ns;
17393 + case VHIN_SYSNAME:
17394 + return uts->name.sysname;
17395 + case VHIN_NODENAME:
17396 + return uts->name.nodename;
17397 + case VHIN_RELEASE:
17398 + return uts->name.release;
17399 + case VHIN_VERSION:
17400 + return uts->name.version;
17401 + case VHIN_MACHINE:
17402 + return uts->name.machine;
17403 + case VHIN_DOMAINNAME:
17404 + return uts->name.domainname;
17411 +int vc_set_vhi_name(struct vx_info *vxi, void __user *data)
17413 + struct vcmd_vhi_name_v0 vc_data;
17416 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
17419 + name = vx_vhi_name(vxi, vc_data.field);
17423 + memcpy(name, vc_data.name, 65);
17427 +int vc_get_vhi_name(struct vx_info *vxi, void __user *data)
17429 + struct vcmd_vhi_name_v0 vc_data;
17432 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
17435 + name = vx_vhi_name(vxi, vc_data.field);
17439 + memcpy(vc_data.name, name, 65);
17440 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
17446 +int vc_virt_stat(struct vx_info *vxi, void __user *data)
17448 + struct vcmd_virt_stat_v0 vc_data;
17449 + struct _vx_cvirt *cvirt = &vxi->cvirt;
17450 + struct timespec uptime;
17452 + do_posix_clock_monotonic_gettime(&uptime);
17453 + set_normalized_timespec(&uptime,
17454 + uptime.tv_sec - cvirt->bias_uptime.tv_sec,
17455 + uptime.tv_nsec - cvirt->bias_uptime.tv_nsec);
17457 + vc_data.offset = timespec_to_ns(&cvirt->bias_ts);
17458 + vc_data.uptime = timespec_to_ns(&uptime);
17459 + vc_data.nr_threads = atomic_read(&cvirt->nr_threads);
17460 + vc_data.nr_running = atomic_read(&cvirt->nr_running);
17461 + vc_data.nr_uninterruptible = atomic_read(&cvirt->nr_uninterruptible);
17462 + vc_data.nr_onhold = atomic_read(&cvirt->nr_onhold);
17463 + vc_data.nr_forks = atomic_read(&cvirt->total_forks);
17464 + vc_data.load[0] = cvirt->load[0];
17465 + vc_data.load[1] = cvirt->load[1];
17466 + vc_data.load[2] = cvirt->load[2];
17468 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
17474 +#ifdef CONFIG_VSERVER_VTIME
17476 +/* virtualized time base */
17478 +void vx_adjust_timespec(struct timespec *ts)
17480 + struct vx_info *vxi;
17482 + if (!vx_flags(VXF_VIRT_TIME, 0))
17485 + vxi = current_vx_info();
17486 + ts->tv_sec += vxi->cvirt.bias_ts.tv_sec;
17487 + ts->tv_nsec += vxi->cvirt.bias_ts.tv_nsec;
17489 + if (ts->tv_nsec >= NSEC_PER_SEC) {
17491 + ts->tv_nsec -= NSEC_PER_SEC;
17492 + } else if (ts->tv_nsec < 0) {
17494 + ts->tv_nsec += NSEC_PER_SEC;
17498 +int vx_settimeofday(struct timespec *ts)
17500 + struct timespec ats, delta;
17501 + struct vx_info *vxi;
17503 + if (!vx_flags(VXF_VIRT_TIME, 0))
17504 + return do_settimeofday(ts);
17506 + getnstimeofday(&ats);
17507 + delta = timespec_sub(ts, &ats);
17509 + vxi = current_vx_info();
17510 + vxi->cvirt.bias_ts = timespec_add(&vxi->cvirt.bias_ts, &delta);
17516 diff -NurpP --minimal linux-2.6.36/kernel/vserver/cvirt_init.h linux-2.6.36-vs2.3.0.36.33/kernel/vserver/cvirt_init.h
17517 --- linux-2.6.36/kernel/vserver/cvirt_init.h 1970-01-01 01:00:00.000000000 +0100
17518 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/cvirt_init.h 2010-10-21 14:40:17.000000000 +0200
17522 +extern uint64_t vx_idle_jiffies(void);
17524 +static inline void vx_info_init_cvirt(struct _vx_cvirt *cvirt)
17526 + uint64_t idle_jiffies = vx_idle_jiffies();
17527 + uint64_t nsuptime;
17529 + do_posix_clock_monotonic_gettime(&cvirt->bias_uptime);
17530 + nsuptime = (unsigned long long)cvirt->bias_uptime.tv_sec
17531 + * NSEC_PER_SEC + cvirt->bias_uptime.tv_nsec;
17532 + cvirt->bias_clock = nsec_to_clock_t(nsuptime);
17533 + cvirt->bias_ts.tv_sec = 0;
17534 + cvirt->bias_ts.tv_nsec = 0;
17536 + jiffies_to_timespec(idle_jiffies, &cvirt->bias_idle);
17537 + atomic_set(&cvirt->nr_threads, 0);
17538 + atomic_set(&cvirt->nr_running, 0);
17539 + atomic_set(&cvirt->nr_uninterruptible, 0);
17540 + atomic_set(&cvirt->nr_onhold, 0);
17542 + spin_lock_init(&cvirt->load_lock);
17543 + cvirt->load_last = jiffies;
17544 + atomic_set(&cvirt->load_updates, 0);
17545 + cvirt->load[0] = 0;
17546 + cvirt->load[1] = 0;
17547 + cvirt->load[2] = 0;
17548 + atomic_set(&cvirt->total_forks, 0);
17550 + spin_lock_init(&cvirt->syslog.logbuf_lock);
17551 + init_waitqueue_head(&cvirt->syslog.log_wait);
17552 + cvirt->syslog.log_start = 0;
17553 + cvirt->syslog.log_end = 0;
17554 + cvirt->syslog.con_start = 0;
17555 + cvirt->syslog.logged_chars = 0;
17559 +void vx_info_init_cvirt_pc(struct _vx_cvirt_pc *cvirt_pc, int cpu)
17561 + // cvirt_pc->cpustat = { 0 };
17564 +static inline void vx_info_exit_cvirt(struct _vx_cvirt *cvirt)
17568 + vxwprintk_xid((value = atomic_read(&cvirt->nr_threads)),
17569 + "!!! cvirt: %p[nr_threads] = %d on exit.",
17571 + vxwprintk_xid((value = atomic_read(&cvirt->nr_running)),
17572 + "!!! cvirt: %p[nr_running] = %d on exit.",
17574 + vxwprintk_xid((value = atomic_read(&cvirt->nr_uninterruptible)),
17575 + "!!! cvirt: %p[nr_uninterruptible] = %d on exit.",
17577 + vxwprintk_xid((value = atomic_read(&cvirt->nr_onhold)),
17578 + "!!! cvirt: %p[nr_onhold] = %d on exit.",
17584 +void vx_info_exit_cvirt_pc(struct _vx_cvirt_pc *cvirt_pc, int cpu)
17589 diff -NurpP --minimal linux-2.6.36/kernel/vserver/cvirt_proc.h linux-2.6.36-vs2.3.0.36.33/kernel/vserver/cvirt_proc.h
17590 --- linux-2.6.36/kernel/vserver/cvirt_proc.h 1970-01-01 01:00:00.000000000 +0100
17591 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/cvirt_proc.h 2010-10-21 13:09:36.000000000 +0200
17593 +#ifndef _VX_CVIRT_PROC_H
17594 +#define _VX_CVIRT_PROC_H
17596 +#include <linux/nsproxy.h>
17597 +#include <linux/mnt_namespace.h>
17598 +#include <linux/ipc_namespace.h>
17599 +#include <linux/utsname.h>
17600 +#include <linux/ipc.h>
17604 +int vx_info_proc_nsproxy(struct nsproxy *nsproxy, char *buffer)
17606 + struct mnt_namespace *ns;
17607 + struct uts_namespace *uts;
17608 + struct ipc_namespace *ipc;
17609 + struct path path;
17610 + char *pstr, *root;
17616 + length += sprintf(buffer + length,
17617 + "NSProxy:\t%p [%p,%p,%p]\n",
17618 + nsproxy, nsproxy->mnt_ns,
17619 + nsproxy->uts_ns, nsproxy->ipc_ns);
17621 + ns = nsproxy->mnt_ns;
17625 + pstr = kmalloc(PATH_MAX, GFP_KERNEL);
17629 + path.mnt = ns->root;
17630 + path.dentry = ns->root->mnt_root;
17631 + root = d_path(&path, pstr, PATH_MAX - 2);
17632 + length += sprintf(buffer + length,
17633 + "Namespace:\t%p [#%u]\n"
17634 + "RootPath:\t%s\n",
17635 + ns, atomic_read(&ns->count),
17640 + uts = nsproxy->uts_ns;
17644 + length += sprintf(buffer + length,
17645 + "SysName:\t%.*s\n"
17646 + "NodeName:\t%.*s\n"
17647 + "Release:\t%.*s\n"
17648 + "Version:\t%.*s\n"
17649 + "Machine:\t%.*s\n"
17650 + "DomainName:\t%.*s\n",
17651 + __NEW_UTS_LEN, uts->name.sysname,
17652 + __NEW_UTS_LEN, uts->name.nodename,
17653 + __NEW_UTS_LEN, uts->name.release,
17654 + __NEW_UTS_LEN, uts->name.version,
17655 + __NEW_UTS_LEN, uts->name.machine,
17656 + __NEW_UTS_LEN, uts->name.domainname);
17659 + ipc = nsproxy->ipc_ns;
17663 + length += sprintf(buffer + length,
17664 + "SEMS:\t\t%d %d %d %d %d\n"
17665 + "MSG:\t\t%d %d %d\n"
17666 + "SHM:\t\t%lu %lu %d %d\n",
17667 + ipc->sem_ctls[0], ipc->sem_ctls[1],
17668 + ipc->sem_ctls[2], ipc->sem_ctls[3],
17670 + ipc->msg_ctlmax, ipc->msg_ctlmnb, ipc->msg_ctlmni,
17671 + (unsigned long)ipc->shm_ctlmax,
17672 + (unsigned long)ipc->shm_ctlall,
17673 + ipc->shm_ctlmni, ipc->shm_tot);
17680 +#include <linux/sched.h>
17682 +#define LOAD_INT(x) ((x) >> FSHIFT)
17683 +#define LOAD_FRAC(x) LOAD_INT(((x) & (FIXED_1 - 1)) * 100)
17686 +int vx_info_proc_cvirt(struct _vx_cvirt *cvirt, char *buffer)
17691 + length += sprintf(buffer + length,
17692 + "BiasUptime:\t%lu.%02lu\n",
17693 + (unsigned long)cvirt->bias_uptime.tv_sec,
17694 + (cvirt->bias_uptime.tv_nsec / (NSEC_PER_SEC / 100)));
17696 + a = cvirt->load[0] + (FIXED_1 / 200);
17697 + b = cvirt->load[1] + (FIXED_1 / 200);
17698 + c = cvirt->load[2] + (FIXED_1 / 200);
17699 + length += sprintf(buffer + length,
17700 + "nr_threads:\t%d\n"
17701 + "nr_running:\t%d\n"
17702 + "nr_unintr:\t%d\n"
17703 + "nr_onhold:\t%d\n"
17704 + "load_updates:\t%d\n"
17705 + "loadavg:\t%d.%02d %d.%02d %d.%02d\n"
17706 + "total_forks:\t%d\n",
17707 + atomic_read(&cvirt->nr_threads),
17708 + atomic_read(&cvirt->nr_running),
17709 + atomic_read(&cvirt->nr_uninterruptible),
17710 + atomic_read(&cvirt->nr_onhold),
17711 + atomic_read(&cvirt->load_updates),
17712 + LOAD_INT(a), LOAD_FRAC(a),
17713 + LOAD_INT(b), LOAD_FRAC(b),
17714 + LOAD_INT(c), LOAD_FRAC(c),
17715 + atomic_read(&cvirt->total_forks));
17720 +int vx_info_proc_cvirt_pc(struct _vx_cvirt_pc *cvirt_pc,
17721 + char *buffer, int cpu)
17727 +#endif /* _VX_CVIRT_PROC_H */
17728 diff -NurpP --minimal linux-2.6.36/kernel/vserver/debug.c linux-2.6.36-vs2.3.0.36.33/kernel/vserver/debug.c
17729 --- linux-2.6.36/kernel/vserver/debug.c 1970-01-01 01:00:00.000000000 +0100
17730 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/debug.c 2010-10-21 13:09:36.000000000 +0200
17733 + * kernel/vserver/debug.c
17735 + * Copyright (C) 2005-2007 Herbert Pötzl
17737 + * V0.01 vx_info dump support
17741 +#include <linux/module.h>
17743 +#include <linux/vserver/context.h>
17746 +void dump_vx_info(struct vx_info *vxi, int level)
17748 + printk("vx_info %p[#%d, %d.%d, %4x]\n", vxi, vxi->vx_id,
17749 + atomic_read(&vxi->vx_usecnt),
17750 + atomic_read(&vxi->vx_tasks),
17753 + __dump_vx_limit(&vxi->limit);
17754 + __dump_vx_sched(&vxi->sched);
17755 + __dump_vx_cvirt(&vxi->cvirt);
17756 + __dump_vx_cacct(&vxi->cacct);
17762 +EXPORT_SYMBOL_GPL(dump_vx_info);
17764 diff -NurpP --minimal linux-2.6.36/kernel/vserver/device.c linux-2.6.36-vs2.3.0.36.33/kernel/vserver/device.c
17765 --- linux-2.6.36/kernel/vserver/device.c 1970-01-01 01:00:00.000000000 +0100
17766 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/device.c 2010-10-21 13:09:36.000000000 +0200
17769 + * linux/kernel/vserver/device.c
17771 + * Linux-VServer: Device Support
17773 + * Copyright (C) 2006 Herbert Pötzl
17774 + * Copyright (C) 2007 Daniel Hokka Zakrisson
17776 + * V0.01 device mapping basics
17777 + * V0.02 added defaults
17781 +#include <linux/slab.h>
17782 +#include <linux/rcupdate.h>
17783 +#include <linux/fs.h>
17784 +#include <linux/namei.h>
17785 +#include <linux/hash.h>
17787 +#include <asm/errno.h>
17788 +#include <asm/uaccess.h>
17789 +#include <linux/vserver/base.h>
17790 +#include <linux/vserver/debug.h>
17791 +#include <linux/vserver/context.h>
17792 +#include <linux/vserver/device.h>
17793 +#include <linux/vserver/device_cmd.h>
17796 +#define DMAP_HASH_BITS 4
17799 +struct vs_mapping {
17801 + struct hlist_node hlist;
17802 + struct list_head list;
17804 +#define dm_hlist u.hlist
17805 +#define dm_list u.list
17808 + struct vx_dmap_target target;
17812 +static struct hlist_head dmap_main_hash[1 << DMAP_HASH_BITS];
17814 +static spinlock_t dmap_main_hash_lock = SPIN_LOCK_UNLOCKED;
17816 +static struct vx_dmap_target dmap_defaults[2] = {
17817 + { .flags = DATTR_OPEN },
17818 + { .flags = DATTR_OPEN },
17822 +struct kmem_cache *dmap_cachep __read_mostly;
17824 +int __init dmap_cache_init(void)
17826 + dmap_cachep = kmem_cache_create("dmap_cache",
17827 + sizeof(struct vs_mapping), 0,
17828 + SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
17832 +__initcall(dmap_cache_init);
17835 +static inline unsigned int __hashval(dev_t dev, int bits)
17837 + return hash_long((unsigned long)dev, bits);
17841 +/* __hash_mapping()
17842 + * add the mapping to the hash table
17844 +static inline void __hash_mapping(struct vx_info *vxi, struct vs_mapping *vdm)
17846 + spinlock_t *hash_lock = &dmap_main_hash_lock;
17847 + struct hlist_head *head, *hash = dmap_main_hash;
17848 + int device = vdm->device;
17850 + spin_lock(hash_lock);
17851 + vxdprintk(VXD_CBIT(misc, 8), "__hash_mapping: %p[#%d] %08x:%08x",
17852 + vxi, vxi ? vxi->vx_id : 0, device, vdm->target.target);
17854 + head = &hash[__hashval(device, DMAP_HASH_BITS)];
17855 + hlist_add_head(&vdm->dm_hlist, head);
17856 + spin_unlock(hash_lock);
17860 +static inline int __mode_to_default(umode_t mode)
17873 +/* __set_default()
17876 +static inline void __set_default(struct vx_info *vxi, umode_t mode,
17877 + struct vx_dmap_target *vdmt)
17879 + spinlock_t *hash_lock = &dmap_main_hash_lock;
17880 + spin_lock(hash_lock);
17883 + vxi->dmap.targets[__mode_to_default(mode)] = *vdmt;
17885 + dmap_defaults[__mode_to_default(mode)] = *vdmt;
17888 + spin_unlock(hash_lock);
17890 + vxdprintk(VXD_CBIT(misc, 8), "__set_default: %p[#%u] %08x %04x",
17891 + vxi, vxi ? vxi->vx_id : 0, vdmt->target, vdmt->flags);
17895 +/* __remove_default()
17896 + * remove a default
17898 +static inline int __remove_default(struct vx_info *vxi, umode_t mode)
17900 + spinlock_t *hash_lock = &dmap_main_hash_lock;
17901 + spin_lock(hash_lock);
17904 + vxi->dmap.targets[__mode_to_default(mode)].flags = 0;
17905 + else /* remove == reset */
17906 + dmap_defaults[__mode_to_default(mode)].flags = DATTR_OPEN | mode;
17908 + spin_unlock(hash_lock);
17913 +/* __find_mapping()
17914 + * find a mapping in the hash table
17916 + * caller must hold hash_lock
17918 +static inline int __find_mapping(xid_t xid, dev_t device, umode_t mode,
17919 + struct vs_mapping **local, struct vs_mapping **global)
17921 + struct hlist_head *hash = dmap_main_hash;
17922 + struct hlist_head *head = &hash[__hashval(device, DMAP_HASH_BITS)];
17923 + struct hlist_node *pos;
17924 + struct vs_mapping *vdm;
17930 + hlist_for_each(pos, head) {
17931 + vdm = hlist_entry(pos, struct vs_mapping, dm_hlist);
17933 + if ((vdm->device == device) &&
17934 + !((vdm->target.flags ^ mode) & S_IFMT)) {
17935 + if (vdm->xid == xid) {
17938 + } else if (global && vdm->xid == 0)
17943 + if (global && *global)
17950 +/* __lookup_mapping()
17951 + * find a mapping and store the result in target and flags
17953 +static inline int __lookup_mapping(struct vx_info *vxi,
17954 + dev_t device, dev_t *target, int *flags, umode_t mode)
17956 + spinlock_t *hash_lock = &dmap_main_hash_lock;
17957 + struct vs_mapping *vdm, *global;
17958 + struct vx_dmap_target *vdmt;
17960 + xid_t xid = vxi->vx_id;
17963 + spin_lock(hash_lock);
17964 + if (__find_mapping(xid, device, mode, &vdm, &global) > 0) {
17966 + vdmt = &vdm->target;
17970 + index = __mode_to_default(mode);
17971 + if (vxi && vxi->dmap.targets[index].flags) {
17973 + vdmt = &vxi->dmap.targets[index];
17974 + } else if (global) {
17976 + vdmt = &global->target;
17980 + vdmt = &dmap_defaults[index];
17984 + if (target && (vdmt->flags & DATTR_REMAP))
17985 + *target = vdmt->target;
17987 + *target = device;
17989 + *flags = vdmt->flags;
17991 + spin_unlock(hash_lock);
17997 +/* __remove_mapping()
17998 + * remove a mapping from the hash table
18000 +static inline int __remove_mapping(struct vx_info *vxi, dev_t device,
18003 + spinlock_t *hash_lock = &dmap_main_hash_lock;
18004 + struct vs_mapping *vdm = NULL;
18007 + spin_lock(hash_lock);
18009 + ret = __find_mapping((vxi ? vxi->vx_id : 0), device, mode, &vdm,
18011 + vxdprintk(VXD_CBIT(misc, 8), "__remove_mapping: %p[#%d] %08x %04x",
18012 + vxi, vxi ? vxi->vx_id : 0, device, mode);
18015 + hlist_del(&vdm->dm_hlist);
18018 + spin_unlock(hash_lock);
18020 + kmem_cache_free(dmap_cachep, vdm);
18026 +int vs_map_device(struct vx_info *vxi,
18027 + dev_t device, dev_t *target, umode_t mode)
18029 + int ret, flags = DATTR_MASK;
18033 + *target = device;
18036 + ret = __lookup_mapping(vxi, device, target, &flags, mode);
18037 + vxdprintk(VXD_CBIT(misc, 8), "vs_map_device: %08x target: %08x flags: %04x mode: %04x mapped=%d",
18038 + device, target ? *target : 0, flags, mode, ret);
18040 + return (flags & DATTR_MASK);
18045 +static int do_set_mapping(struct vx_info *vxi,
18046 + dev_t device, dev_t target, int flags, umode_t mode)
18049 + struct vs_mapping *new;
18051 + new = kmem_cache_alloc(dmap_cachep, GFP_KERNEL);
18055 + INIT_HLIST_NODE(&new->dm_hlist);
18056 + new->device = device;
18057 + new->target.target = target;
18058 + new->target.flags = flags | mode;
18059 + new->xid = (vxi ? vxi->vx_id : 0);
18061 + vxdprintk(VXD_CBIT(misc, 8), "do_set_mapping: %08x target: %08x flags: %04x", device, target, flags);
18062 + __hash_mapping(vxi, new);
18064 + struct vx_dmap_target new = {
18065 + .target = target,
18066 + .flags = flags | mode,
18068 + __set_default(vxi, mode, &new);
18074 +static int do_unset_mapping(struct vx_info *vxi,
18075 + dev_t device, dev_t target, int flags, umode_t mode)
18077 + int ret = -EINVAL;
18080 + ret = __remove_mapping(vxi, device, mode);
18084 + ret = __remove_default(vxi, mode);
18094 +static inline int __user_device(const char __user *name, dev_t *dev,
18097 + struct nameidata nd;
18104 + ret = user_lpath(name, &nd.path);
18107 + if (nd.path.dentry->d_inode) {
18108 + *dev = nd.path.dentry->d_inode->i_rdev;
18109 + *mode = nd.path.dentry->d_inode->i_mode;
18111 + path_put(&nd.path);
18115 +static inline int __mapping_mode(dev_t device, dev_t target,
18116 + umode_t device_mode, umode_t target_mode, umode_t *mode)
18119 + *mode = device_mode & S_IFMT;
18121 + *mode = target_mode & S_IFMT;
18125 + /* if both given, device and target mode have to match */
18126 + if (device && target &&
18127 + ((device_mode ^ target_mode) & S_IFMT))
18133 +static inline int do_mapping(struct vx_info *vxi, const char __user *device_path,
18134 + const char __user *target_path, int flags, int set)
18136 + dev_t device = ~0, target = ~0;
18137 + umode_t device_mode = 0, target_mode = 0, mode;
18140 + ret = __user_device(device_path, &device, &device_mode);
18143 + ret = __user_device(target_path, &target, &target_mode);
18147 + ret = __mapping_mode(device, target,
18148 + device_mode, target_mode, &mode);
18153 + return do_set_mapping(vxi, device, target,
18156 + return do_unset_mapping(vxi, device, target,
18161 +int vc_set_mapping(struct vx_info *vxi, void __user *data)
18163 + struct vcmd_set_mapping_v0 vc_data;
18165 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18168 + return do_mapping(vxi, vc_data.device, vc_data.target,
18169 + vc_data.flags, 1);
18172 +int vc_unset_mapping(struct vx_info *vxi, void __user *data)
18174 + struct vcmd_set_mapping_v0 vc_data;
18176 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18179 + return do_mapping(vxi, vc_data.device, vc_data.target,
18180 + vc_data.flags, 0);
18184 +#ifdef CONFIG_COMPAT
18186 +int vc_set_mapping_x32(struct vx_info *vxi, void __user *data)
18188 + struct vcmd_set_mapping_v0_x32 vc_data;
18190 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18193 + return do_mapping(vxi, compat_ptr(vc_data.device_ptr),
18194 + compat_ptr(vc_data.target_ptr), vc_data.flags, 1);
18197 +int vc_unset_mapping_x32(struct vx_info *vxi, void __user *data)
18199 + struct vcmd_set_mapping_v0_x32 vc_data;
18201 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18204 + return do_mapping(vxi, compat_ptr(vc_data.device_ptr),
18205 + compat_ptr(vc_data.target_ptr), vc_data.flags, 0);
18208 +#endif /* CONFIG_COMPAT */
18211 diff -NurpP --minimal linux-2.6.36/kernel/vserver/dlimit.c linux-2.6.36-vs2.3.0.36.33/kernel/vserver/dlimit.c
18212 --- linux-2.6.36/kernel/vserver/dlimit.c 1970-01-01 01:00:00.000000000 +0100
18213 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/dlimit.c 2010-10-21 13:09:36.000000000 +0200
18216 + * linux/kernel/vserver/dlimit.c
18218 + * Virtual Server: Context Disk Limits
18220 + * Copyright (C) 2004-2009 Herbert Pötzl
18222 + * V0.01 initial version
18223 + * V0.02 compat32 splitup
18224 + * V0.03 extended interface
18228 +#include <linux/statfs.h>
18229 +#include <linux/sched.h>
18230 +#include <linux/namei.h>
18231 +#include <linux/vs_tag.h>
18232 +#include <linux/vs_dlimit.h>
18233 +#include <linux/vserver/dlimit_cmd.h>
18234 +#include <linux/slab.h>
18235 +// #include <linux/gfp.h>
18237 +#include <asm/uaccess.h>
18239 +/* __alloc_dl_info()
18241 + * allocate an initialized dl_info struct
18242 + * doesn't make it visible (hash) */
18244 +static struct dl_info *__alloc_dl_info(struct super_block *sb, tag_t tag)
18246 + struct dl_info *new = NULL;
18248 + vxdprintk(VXD_CBIT(dlim, 5),
18249 + "alloc_dl_info(%p,%d)*", sb, tag);
18251 + /* would this benefit from a slab cache? */
18252 + new = kmalloc(sizeof(struct dl_info), GFP_KERNEL);
18256 + memset(new, 0, sizeof(struct dl_info));
18257 + new->dl_tag = tag;
18259 + INIT_RCU_HEAD(&new->dl_rcu);
18260 + INIT_HLIST_NODE(&new->dl_hlist);
18261 + spin_lock_init(&new->dl_lock);
18262 + atomic_set(&new->dl_refcnt, 0);
18263 + atomic_set(&new->dl_usecnt, 0);
18265 + /* rest of init goes here */
18267 + vxdprintk(VXD_CBIT(dlim, 4),
18268 + "alloc_dl_info(%p,%d) = %p", sb, tag, new);
18272 +/* __dealloc_dl_info()
18274 + * final disposal of dl_info */
18276 +static void __dealloc_dl_info(struct dl_info *dli)
18278 + vxdprintk(VXD_CBIT(dlim, 4),
18279 + "dealloc_dl_info(%p)", dli);
18281 + dli->dl_hlist.next = LIST_POISON1;
18282 + dli->dl_tag = -1;
18285 + BUG_ON(atomic_read(&dli->dl_usecnt));
18286 + BUG_ON(atomic_read(&dli->dl_refcnt));
18292 +/* hash table for dl_info hash */
18294 +#define DL_HASH_SIZE 13
18296 +struct hlist_head dl_info_hash[DL_HASH_SIZE];
18298 +static spinlock_t dl_info_hash_lock = SPIN_LOCK_UNLOCKED;
18301 +static inline unsigned int __hashval(struct super_block *sb, tag_t tag)
18303 + return ((tag ^ (unsigned long)sb) % DL_HASH_SIZE);
18308 +/* __hash_dl_info()
18310 + * add the dli to the global hash table
18311 + * requires the hash_lock to be held */
18313 +static inline void __hash_dl_info(struct dl_info *dli)
18315 + struct hlist_head *head;
18317 + vxdprintk(VXD_CBIT(dlim, 6),
18318 + "__hash_dl_info: %p[#%d]", dli, dli->dl_tag);
18319 + get_dl_info(dli);
18320 + head = &dl_info_hash[__hashval(dli->dl_sb, dli->dl_tag)];
18321 + hlist_add_head_rcu(&dli->dl_hlist, head);
18324 +/* __unhash_dl_info()
18326 + * remove the dli from the global hash table
18327 + * requires the hash_lock to be held */
18329 +static inline void __unhash_dl_info(struct dl_info *dli)
18331 + vxdprintk(VXD_CBIT(dlim, 6),
18332 + "__unhash_dl_info: %p[#%d]", dli, dli->dl_tag);
18333 + hlist_del_rcu(&dli->dl_hlist);
18334 + put_dl_info(dli);
18338 +/* __lookup_dl_info()
18340 + * requires the rcu_read_lock()
18341 + * doesn't increment the dl_refcnt */
18343 +static inline struct dl_info *__lookup_dl_info(struct super_block *sb, tag_t tag)
18345 + struct hlist_head *head = &dl_info_hash[__hashval(sb, tag)];
18346 + struct hlist_node *pos;
18347 + struct dl_info *dli;
18349 + hlist_for_each_entry_rcu(dli, pos, head, dl_hlist) {
18351 + if (dli->dl_tag == tag && dli->dl_sb == sb) {
18359 +struct dl_info *locate_dl_info(struct super_block *sb, tag_t tag)
18361 + struct dl_info *dli;
18364 + dli = get_dl_info(__lookup_dl_info(sb, tag));
18365 + vxdprintk(VXD_CBIT(dlim, 7),
18366 + "locate_dl_info(%p,#%d) = %p", sb, tag, dli);
18367 + rcu_read_unlock();
18371 +void rcu_free_dl_info(struct rcu_head *head)
18373 + struct dl_info *dli = container_of(head, struct dl_info, dl_rcu);
18374 + int usecnt, refcnt;
18376 + BUG_ON(!dli || !head);
18378 + usecnt = atomic_read(&dli->dl_usecnt);
18379 + BUG_ON(usecnt < 0);
18381 + refcnt = atomic_read(&dli->dl_refcnt);
18382 + BUG_ON(refcnt < 0);
18384 + vxdprintk(VXD_CBIT(dlim, 3),
18385 + "rcu_free_dl_info(%p)", dli);
18387 + __dealloc_dl_info(dli);
18389 + printk("!!! rcu didn't free\n");
18395 +static int do_addrem_dlimit(uint32_t id, const char __user *name,
18396 + uint32_t flags, int add)
18398 + struct path path;
18401 + ret = user_lpath(name, &path);
18403 + struct super_block *sb;
18404 + struct dl_info *dli;
18407 + if (!path.dentry->d_inode)
18408 + goto out_release;
18409 + if (!(sb = path.dentry->d_inode->i_sb))
18410 + goto out_release;
18413 + dli = __alloc_dl_info(sb, id);
18414 + spin_lock(&dl_info_hash_lock);
18417 + if (__lookup_dl_info(sb, id))
18419 + __hash_dl_info(dli);
18422 + spin_lock(&dl_info_hash_lock);
18423 + dli = __lookup_dl_info(sb, id);
18428 + __unhash_dl_info(dli);
18432 + spin_unlock(&dl_info_hash_lock);
18434 + __dealloc_dl_info(dli);
18441 +int vc_add_dlimit(uint32_t id, void __user *data)
18443 + struct vcmd_ctx_dlimit_base_v0 vc_data;
18445 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18448 + return do_addrem_dlimit(id, vc_data.name, vc_data.flags, 1);
18451 +int vc_rem_dlimit(uint32_t id, void __user *data)
18453 + struct vcmd_ctx_dlimit_base_v0 vc_data;
18455 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18458 + return do_addrem_dlimit(id, vc_data.name, vc_data.flags, 0);
18461 +#ifdef CONFIG_COMPAT
18463 +int vc_add_dlimit_x32(uint32_t id, void __user *data)
18465 + struct vcmd_ctx_dlimit_base_v0_x32 vc_data;
18467 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18470 + return do_addrem_dlimit(id,
18471 + compat_ptr(vc_data.name_ptr), vc_data.flags, 1);
18474 +int vc_rem_dlimit_x32(uint32_t id, void __user *data)
18476 + struct vcmd_ctx_dlimit_base_v0_x32 vc_data;
18478 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18481 + return do_addrem_dlimit(id,
18482 + compat_ptr(vc_data.name_ptr), vc_data.flags, 0);
18485 +#endif /* CONFIG_COMPAT */
18489 +int do_set_dlimit(uint32_t id, const char __user *name,
18490 + uint32_t space_used, uint32_t space_total,
18491 + uint32_t inodes_used, uint32_t inodes_total,
18492 + uint32_t reserved, uint32_t flags)
18494 + struct path path;
18497 + ret = user_lpath(name, &path);
18499 + struct super_block *sb;
18500 + struct dl_info *dli;
18503 + if (!path.dentry->d_inode)
18504 + goto out_release;
18505 + if (!(sb = path.dentry->d_inode->i_sb))
18506 + goto out_release;
18508 + /* sanity checks */
18509 + if ((reserved != CDLIM_KEEP &&
18510 + reserved > 100) ||
18511 + (inodes_used != CDLIM_KEEP &&
18512 + inodes_used > inodes_total) ||
18513 + (space_used != CDLIM_KEEP &&
18514 + space_used > space_total))
18515 + goto out_release;
18518 + dli = locate_dl_info(sb, id);
18520 + goto out_release;
18522 + spin_lock(&dli->dl_lock);
18524 + if (inodes_used != CDLIM_KEEP)
18525 + dli->dl_inodes_used = inodes_used;
18526 + if (inodes_total != CDLIM_KEEP)
18527 + dli->dl_inodes_total = inodes_total;
18528 + if (space_used != CDLIM_KEEP)
18529 + dli->dl_space_used = dlimit_space_32to64(
18530 + space_used, flags, DLIMS_USED);
18532 + if (space_total == CDLIM_INFINITY)
18533 + dli->dl_space_total = DLIM_INFINITY;
18534 + else if (space_total != CDLIM_KEEP)
18535 + dli->dl_space_total = dlimit_space_32to64(
18536 + space_total, flags, DLIMS_TOTAL);
18538 + if (reserved != CDLIM_KEEP)
18539 + dli->dl_nrlmult = (1 << 10) * (100 - reserved) / 100;
18541 + spin_unlock(&dli->dl_lock);
18543 + put_dl_info(dli);
18552 +int vc_set_dlimit(uint32_t id, void __user *data)
18554 + struct vcmd_ctx_dlimit_v0 vc_data;
18556 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18559 + return do_set_dlimit(id, vc_data.name,
18560 + vc_data.space_used, vc_data.space_total,
18561 + vc_data.inodes_used, vc_data.inodes_total,
18562 + vc_data.reserved, vc_data.flags);
18565 +#ifdef CONFIG_COMPAT
18567 +int vc_set_dlimit_x32(uint32_t id, void __user *data)
18569 + struct vcmd_ctx_dlimit_v0_x32 vc_data;
18571 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18574 + return do_set_dlimit(id, compat_ptr(vc_data.name_ptr),
18575 + vc_data.space_used, vc_data.space_total,
18576 + vc_data.inodes_used, vc_data.inodes_total,
18577 + vc_data.reserved, vc_data.flags);
18580 +#endif /* CONFIG_COMPAT */
18584 +int do_get_dlimit(uint32_t id, const char __user *name,
18585 + uint32_t *space_used, uint32_t *space_total,
18586 + uint32_t *inodes_used, uint32_t *inodes_total,
18587 + uint32_t *reserved, uint32_t *flags)
18589 + struct path path;
18592 + ret = user_lpath(name, &path);
18594 + struct super_block *sb;
18595 + struct dl_info *dli;
18598 + if (!path.dentry->d_inode)
18599 + goto out_release;
18600 + if (!(sb = path.dentry->d_inode->i_sb))
18601 + goto out_release;
18604 + dli = locate_dl_info(sb, id);
18606 + goto out_release;
18608 + spin_lock(&dli->dl_lock);
18609 + *inodes_used = dli->dl_inodes_used;
18610 + *inodes_total = dli->dl_inodes_total;
18612 + *space_used = dlimit_space_64to32(
18613 + dli->dl_space_used, flags, DLIMS_USED);
18615 + if (dli->dl_space_total == DLIM_INFINITY)
18616 + *space_total = CDLIM_INFINITY;
18618 + *space_total = dlimit_space_64to32(
18619 + dli->dl_space_total, flags, DLIMS_TOTAL);
18621 + *reserved = 100 - ((dli->dl_nrlmult * 100 + 512) >> 10);
18622 + spin_unlock(&dli->dl_lock);
18624 + put_dl_info(dli);
18635 +int vc_get_dlimit(uint32_t id, void __user *data)
18637 + struct vcmd_ctx_dlimit_v0 vc_data;
18640 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18643 + ret = do_get_dlimit(id, vc_data.name,
18644 + &vc_data.space_used, &vc_data.space_total,
18645 + &vc_data.inodes_used, &vc_data.inodes_total,
18646 + &vc_data.reserved, &vc_data.flags);
18650 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
18655 +#ifdef CONFIG_COMPAT
18657 +int vc_get_dlimit_x32(uint32_t id, void __user *data)
18659 + struct vcmd_ctx_dlimit_v0_x32 vc_data;
18662 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18665 + ret = do_get_dlimit(id, compat_ptr(vc_data.name_ptr),
18666 + &vc_data.space_used, &vc_data.space_total,
18667 + &vc_data.inodes_used, &vc_data.inodes_total,
18668 + &vc_data.reserved, &vc_data.flags);
18672 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
18677 +#endif /* CONFIG_COMPAT */
18680 +void vx_vsi_statfs(struct super_block *sb, struct kstatfs *buf)
18682 + struct dl_info *dli;
18683 + __u64 blimit, bfree, bavail;
18686 + dli = locate_dl_info(sb, dx_current_tag());
18690 + spin_lock(&dli->dl_lock);
18691 + if (dli->dl_inodes_total == (unsigned long)DLIM_INFINITY)
18694 + /* reduce max inodes available to limit */
18695 + if (buf->f_files > dli->dl_inodes_total)
18696 + buf->f_files = dli->dl_inodes_total;
18698 + ifree = dli->dl_inodes_total - dli->dl_inodes_used;
18699 + /* reduce free inodes to min */
18700 + if (ifree < buf->f_ffree)
18701 + buf->f_ffree = ifree;
18704 + if (dli->dl_space_total == DLIM_INFINITY)
18707 + blimit = dli->dl_space_total >> sb->s_blocksize_bits;
18709 + if (dli->dl_space_total < dli->dl_space_used)
18712 + bfree = (dli->dl_space_total - dli->dl_space_used)
18713 + >> sb->s_blocksize_bits;
18715 + bavail = ((dli->dl_space_total >> 10) * dli->dl_nrlmult);
18716 + if (bavail < dli->dl_space_used)
18719 + bavail = (bavail - dli->dl_space_used)
18720 + >> sb->s_blocksize_bits;
18722 + /* reduce max space available to limit */
18723 + if (buf->f_blocks > blimit)
18724 + buf->f_blocks = blimit;
18726 + /* reduce free space to min */
18727 + if (bfree < buf->f_bfree)
18728 + buf->f_bfree = bfree;
18730 + /* reduce avail space to min */
18731 + if (bavail < buf->f_bavail)
18732 + buf->f_bavail = bavail;
18735 + spin_unlock(&dli->dl_lock);
18736 + put_dl_info(dli);
18741 +#include <linux/module.h>
18743 +EXPORT_SYMBOL_GPL(locate_dl_info);
18744 +EXPORT_SYMBOL_GPL(rcu_free_dl_info);
18746 diff -NurpP --minimal linux-2.6.36/kernel/vserver/helper.c linux-2.6.36-vs2.3.0.36.33/kernel/vserver/helper.c
18747 --- linux-2.6.36/kernel/vserver/helper.c 1970-01-01 01:00:00.000000000 +0100
18748 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/helper.c 2010-10-21 13:09:36.000000000 +0200
18751 + * linux/kernel/vserver/helper.c
18753 + * Virtual Context Support
18755 + * Copyright (C) 2004-2007 Herbert Pötzl
18757 + * V0.01 basic helper
18761 +#include <linux/kmod.h>
18762 +#include <linux/reboot.h>
18763 +#include <linux/vs_context.h>
18764 +#include <linux/vs_network.h>
18765 +#include <linux/vserver/signal.h>
18768 +char vshelper_path[255] = "/sbin/vshelper";
18771 +static int do_vshelper(char *name, char *argv[], char *envp[], int sync)
18775 + if ((ret = call_usermodehelper(name, argv, envp, sync))) {
18776 + printk( KERN_WARNING
18777 + "%s: (%s %s) returned %s with %d\n",
18778 + name, argv[1], argv[2],
18779 + sync ? "sync" : "async", ret);
18781 + vxdprintk(VXD_CBIT(switch, 4),
18782 + "%s: (%s %s) returned %s with %d",
18783 + name, argv[1], argv[2], sync ? "sync" : "async", ret);
18788 + * vshelper path is set via /proc/sys
18789 + * invoked by vserver sys_reboot(), with
18790 + * the following arguments
18792 + * argv [0] = vshelper_path;
18793 + * argv [1] = action: "restart", "halt", "poweroff", ...
18794 + * argv [2] = context identifier
18796 + * envp [*] = type-specific parameters
18799 +long vs_reboot_helper(struct vx_info *vxi, int cmd, void __user *arg)
18801 + char id_buf[8], cmd_buf[16];
18802 + char uid_buf[16], pid_buf[16];
18805 + char *argv[] = {vshelper_path, NULL, id_buf, 0};
18806 + char *envp[] = {"HOME=/", "TERM=linux",
18807 + "PATH=/sbin:/usr/sbin:/bin:/usr/bin",
18808 + uid_buf, pid_buf, cmd_buf, 0};
18810 + if (vx_info_state(vxi, VXS_HELPER))
18812 + vxi->vx_state |= VXS_HELPER;
18814 + snprintf(id_buf, sizeof(id_buf)-1, "%d", vxi->vx_id);
18816 + snprintf(cmd_buf, sizeof(cmd_buf)-1, "VS_CMD=%08x", cmd);
18817 + snprintf(uid_buf, sizeof(uid_buf)-1, "VS_UID=%d", current_uid());
18818 + snprintf(pid_buf, sizeof(pid_buf)-1, "VS_PID=%d", current->pid);
18821 + case LINUX_REBOOT_CMD_RESTART:
18822 + argv[1] = "restart";
18825 + case LINUX_REBOOT_CMD_HALT:
18826 + argv[1] = "halt";
18829 + case LINUX_REBOOT_CMD_POWER_OFF:
18830 + argv[1] = "poweroff";
18833 + case LINUX_REBOOT_CMD_SW_SUSPEND:
18834 + argv[1] = "swsusp";
18837 + case LINUX_REBOOT_CMD_OOM:
18842 + vxi->vx_state &= ~VXS_HELPER;
18846 + ret = do_vshelper(vshelper_path, argv, envp, 0);
18847 + vxi->vx_state &= ~VXS_HELPER;
18848 + __wakeup_vx_info(vxi);
18849 + return (ret) ? -EPERM : 0;
18853 +long vs_reboot(unsigned int cmd, void __user *arg)
18855 + struct vx_info *vxi = current_vx_info();
18858 + vxdprintk(VXD_CBIT(misc, 5),
18859 + "vs_reboot(%p[#%d],%u)",
18860 + vxi, vxi ? vxi->vx_id : 0, cmd);
18862 + ret = vs_reboot_helper(vxi, cmd, arg);
18866 + vxi->reboot_cmd = cmd;
18867 + if (vx_info_flags(vxi, VXF_REBOOT_KILL, 0)) {
18869 + case LINUX_REBOOT_CMD_RESTART:
18870 + case LINUX_REBOOT_CMD_HALT:
18871 + case LINUX_REBOOT_CMD_POWER_OFF:
18872 + vx_info_kill(vxi, 0, SIGKILL);
18873 + vx_info_kill(vxi, 1, SIGKILL);
18881 +long vs_oom_action(unsigned int cmd)
18883 + struct vx_info *vxi = current_vx_info();
18886 + vxdprintk(VXD_CBIT(misc, 5),
18887 + "vs_oom_action(%p[#%d],%u)",
18888 + vxi, vxi ? vxi->vx_id : 0, cmd);
18890 + ret = vs_reboot_helper(vxi, cmd, NULL);
18894 + vxi->reboot_cmd = cmd;
18895 + if (vx_info_flags(vxi, VXF_REBOOT_KILL, 0)) {
18896 + vx_info_kill(vxi, 0, SIGKILL);
18897 + vx_info_kill(vxi, 1, SIGKILL);
18903 + * argv [0] = vshelper_path;
18904 + * argv [1] = action: "startup", "shutdown"
18905 + * argv [2] = context identifier
18907 + * envp [*] = type-specific parameters
18910 +long vs_state_change(struct vx_info *vxi, unsigned int cmd)
18912 + char id_buf[8], cmd_buf[16];
18913 + char *argv[] = {vshelper_path, NULL, id_buf, 0};
18914 + char *envp[] = {"HOME=/", "TERM=linux",
18915 + "PATH=/sbin:/usr/sbin:/bin:/usr/bin", cmd_buf, 0};
18917 + if (!vx_info_flags(vxi, VXF_SC_HELPER, 0))
18920 + snprintf(id_buf, sizeof(id_buf)-1, "%d", vxi->vx_id);
18921 + snprintf(cmd_buf, sizeof(cmd_buf)-1, "VS_CMD=%08x", cmd);
18924 + case VSC_STARTUP:
18925 + argv[1] = "startup";
18927 + case VSC_SHUTDOWN:
18928 + argv[1] = "shutdown";
18934 + return do_vshelper(vshelper_path, argv, envp, 1);
18939 + * argv [0] = vshelper_path;
18940 + * argv [1] = action: "netup", "netdown"
18941 + * argv [2] = context identifier
18943 + * envp [*] = type-specific parameters
18946 +long vs_net_change(struct nx_info *nxi, unsigned int cmd)
18948 + char id_buf[8], cmd_buf[16];
18949 + char *argv[] = {vshelper_path, NULL, id_buf, 0};
18950 + char *envp[] = {"HOME=/", "TERM=linux",
18951 + "PATH=/sbin:/usr/sbin:/bin:/usr/bin", cmd_buf, 0};
18953 + if (!nx_info_flags(nxi, NXF_SC_HELPER, 0))
18956 + snprintf(id_buf, sizeof(id_buf)-1, "%d", nxi->nx_id);
18957 + snprintf(cmd_buf, sizeof(cmd_buf)-1, "VS_CMD=%08x", cmd);
18961 + argv[1] = "netup";
18963 + case VSC_NETDOWN:
18964 + argv[1] = "netdown";
18970 + return do_vshelper(vshelper_path, argv, envp, 1);
18973 diff -NurpP --minimal linux-2.6.36/kernel/vserver/history.c linux-2.6.36-vs2.3.0.36.33/kernel/vserver/history.c
18974 --- linux-2.6.36/kernel/vserver/history.c 1970-01-01 01:00:00.000000000 +0100
18975 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/history.c 2010-10-21 13:09:36.000000000 +0200
18978 + * kernel/vserver/history.c
18980 + * Virtual Context History Backtrace
18982 + * Copyright (C) 2004-2007 Herbert Pötzl
18984 + * V0.01 basic structure
18985 + * V0.02 hash/unhash and trace
18986 + * V0.03 preemption fixes
18990 +#include <linux/module.h>
18991 +#include <asm/uaccess.h>
18993 +#include <linux/vserver/context.h>
18994 +#include <linux/vserver/debug.h>
18995 +#include <linux/vserver/debug_cmd.h>
18996 +#include <linux/vserver/history.h>
18999 +#ifdef CONFIG_VSERVER_HISTORY
19000 +#define VXH_SIZE CONFIG_VSERVER_HISTORY_SIZE
19002 +#define VXH_SIZE 64
19005 +struct _vx_history {
19006 + unsigned int counter;
19008 + struct _vx_hist_entry entry[VXH_SIZE + 1];
19012 +DEFINE_PER_CPU(struct _vx_history, vx_history_buffer);
19014 +unsigned volatile int vxh_active = 1;
19016 +static atomic_t sequence = ATOMIC_INIT(0);
19021 + * requires disabled preemption */
19023 +struct _vx_hist_entry *vxh_advance(void *loc)
19025 + unsigned int cpu = smp_processor_id();
19026 + struct _vx_history *hist = &per_cpu(vx_history_buffer, cpu);
19027 + struct _vx_hist_entry *entry;
19028 + unsigned int index;
19030 + index = vxh_active ? (hist->counter++ % VXH_SIZE) : VXH_SIZE;
19031 + entry = &hist->entry[index];
19033 + entry->seq = atomic_inc_return(&sequence);
19034 + entry->loc = loc;
19038 +EXPORT_SYMBOL_GPL(vxh_advance);
19041 +#define VXH_LOC_FMTS "(#%04x,*%d):%p"
19043 +#define VXH_LOC_ARGS(e) (e)->seq, cpu, (e)->loc
19046 +#define VXH_VXI_FMTS "%p[#%d,%d.%d]"
19048 +#define VXH_VXI_ARGS(e) (e)->vxi.ptr, \
19049 + (e)->vxi.ptr ? (e)->vxi.xid : 0, \
19050 + (e)->vxi.ptr ? (e)->vxi.usecnt : 0, \
19051 + (e)->vxi.ptr ? (e)->vxi.tasks : 0
19053 +void vxh_dump_entry(struct _vx_hist_entry *e, unsigned cpu)
19055 + switch (e->type) {
19056 + case VXH_THROW_OOPS:
19057 + printk( VXH_LOC_FMTS " oops \n", VXH_LOC_ARGS(e));
19060 + case VXH_GET_VX_INFO:
19061 + case VXH_PUT_VX_INFO:
19062 + printk( VXH_LOC_FMTS " %s_vx_info " VXH_VXI_FMTS "\n",
19064 + (e->type == VXH_GET_VX_INFO) ? "get" : "put",
19065 + VXH_VXI_ARGS(e));
19068 + case VXH_INIT_VX_INFO:
19069 + case VXH_SET_VX_INFO:
19070 + case VXH_CLR_VX_INFO:
19071 + printk( VXH_LOC_FMTS " %s_vx_info " VXH_VXI_FMTS " @%p\n",
19073 + (e->type == VXH_INIT_VX_INFO) ? "init" :
19074 + ((e->type == VXH_SET_VX_INFO) ? "set" : "clr"),
19075 + VXH_VXI_ARGS(e), e->sc.data);
19078 + case VXH_CLAIM_VX_INFO:
19079 + case VXH_RELEASE_VX_INFO:
19080 + printk( VXH_LOC_FMTS " %s_vx_info " VXH_VXI_FMTS " @%p\n",
19082 + (e->type == VXH_CLAIM_VX_INFO) ? "claim" : "release",
19083 + VXH_VXI_ARGS(e), e->sc.data);
19086 + case VXH_ALLOC_VX_INFO:
19087 + case VXH_DEALLOC_VX_INFO:
19088 + printk( VXH_LOC_FMTS " %s_vx_info " VXH_VXI_FMTS "\n",
19090 + (e->type == VXH_ALLOC_VX_INFO) ? "alloc" : "dealloc",
19091 + VXH_VXI_ARGS(e));
19094 + case VXH_HASH_VX_INFO:
19095 + case VXH_UNHASH_VX_INFO:
19096 + printk( VXH_LOC_FMTS " __%s_vx_info " VXH_VXI_FMTS "\n",
19098 + (e->type == VXH_HASH_VX_INFO) ? "hash" : "unhash",
19099 + VXH_VXI_ARGS(e));
19102 + case VXH_LOC_VX_INFO:
19103 + case VXH_LOOKUP_VX_INFO:
19104 + case VXH_CREATE_VX_INFO:
19105 + printk( VXH_LOC_FMTS " __%s_vx_info [#%d] -> " VXH_VXI_FMTS "\n",
19107 + (e->type == VXH_CREATE_VX_INFO) ? "create" :
19108 + ((e->type == VXH_LOC_VX_INFO) ? "loc" : "lookup"),
19109 + e->ll.arg, VXH_VXI_ARGS(e));
19114 +static void __vxh_dump_history(void)
19116 + unsigned int i, cpu;
19118 + printk("History:\tSEQ: %8x\tNR_CPUS: %d\n",
19119 + atomic_read(&sequence), NR_CPUS);
19121 + for (i = 0; i < VXH_SIZE; i++) {
19122 + for_each_online_cpu(cpu) {
19123 + struct _vx_history *hist =
19124 + &per_cpu(vx_history_buffer, cpu);
19125 + unsigned int index = (hist->counter - i) % VXH_SIZE;
19126 + struct _vx_hist_entry *entry = &hist->entry[index];
19128 + vxh_dump_entry(entry, cpu);
19133 +void vxh_dump_history(void)
19137 + local_irq_enable();
19139 + local_irq_disable();
19141 + __vxh_dump_history();
19145 +/* vserver syscall commands below here */
19148 +int vc_dump_history(uint32_t id)
19151 + __vxh_dump_history();
19158 +int do_read_history(struct __user _vx_hist_entry *data,
19159 + int cpu, uint32_t *index, uint32_t *count)
19161 + int pos, ret = 0;
19162 + struct _vx_history *hist = &per_cpu(vx_history_buffer, cpu);
19163 + int end = hist->counter;
19164 + int start = end - VXH_SIZE + 2;
19165 + int idx = *index;
19167 + /* special case: get current pos */
19173 + /* have we lost some data? */
19177 + for (pos = 0; (pos < *count) && (idx < end); pos++, idx++) {
19178 + struct _vx_hist_entry *entry =
19179 + &hist->entry[idx % VXH_SIZE];
19181 + /* send entry to userspace */
19182 + ret = copy_to_user(&data[pos], entry, sizeof(*entry));
19186 + /* save new index and count */
19189 + return ret ? ret : (*index < end);
19192 +int vc_read_history(uint32_t id, void __user *data)
19194 + struct vcmd_read_history_v0 vc_data;
19197 + if (id >= NR_CPUS)
19200 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19203 + ret = do_read_history((struct __user _vx_hist_entry *)vc_data.data,
19204 + id, &vc_data.index, &vc_data.count);
19206 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19211 +#ifdef CONFIG_COMPAT
19213 +int vc_read_history_x32(uint32_t id, void __user *data)
19215 + struct vcmd_read_history_v0_x32 vc_data;
19218 + if (id >= NR_CPUS)
19221 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19224 + ret = do_read_history((struct __user _vx_hist_entry *)
19225 + compat_ptr(vc_data.data_ptr),
19226 + id, &vc_data.index, &vc_data.count);
19228 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19233 +#endif /* CONFIG_COMPAT */
19235 diff -NurpP --minimal linux-2.6.36/kernel/vserver/inet.c linux-2.6.36-vs2.3.0.36.33/kernel/vserver/inet.c
19236 --- linux-2.6.36/kernel/vserver/inet.c 1970-01-01 01:00:00.000000000 +0100
19237 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/inet.c 2010-10-21 13:09:36.000000000 +0200
19240 +#include <linux/in.h>
19241 +#include <linux/inetdevice.h>
19242 +#include <linux/vs_inet.h>
19243 +#include <linux/vs_inet6.h>
19244 +#include <linux/vserver/debug.h>
19245 +#include <net/route.h>
19246 +#include <net/addrconf.h>
19249 +int nx_v4_addr_conflict(struct nx_info *nxi1, struct nx_info *nxi2)
19253 + if (!nxi1 || !nxi2 || nxi1 == nxi2)
19256 + struct nx_addr_v4 *ptr;
19258 + for (ptr = &nxi1->v4; ptr; ptr = ptr->next) {
19259 + if (v4_nx_addr_in_nx_info(nxi2, ptr, -1)) {
19266 + vxdprintk(VXD_CBIT(net, 2),
19267 + "nx_v4_addr_conflict(%p,%p): %d",
19268 + nxi1, nxi2, ret);
19274 +#ifdef CONFIG_IPV6
19276 +int nx_v6_addr_conflict(struct nx_info *nxi1, struct nx_info *nxi2)
19280 + if (!nxi1 || !nxi2 || nxi1 == nxi2)
19283 + struct nx_addr_v6 *ptr;
19285 + for (ptr = &nxi1->v6; ptr; ptr = ptr->next) {
19286 + if (v6_nx_addr_in_nx_info(nxi2, ptr, -1)) {
19293 + vxdprintk(VXD_CBIT(net, 2),
19294 + "nx_v6_addr_conflict(%p,%p): %d",
19295 + nxi1, nxi2, ret);
19302 +int v4_dev_in_nx_info(struct net_device *dev, struct nx_info *nxi)
19304 + struct in_device *in_dev;
19305 + struct in_ifaddr **ifap;
19306 + struct in_ifaddr *ifa;
19311 + in_dev = in_dev_get(dev);
19315 + for (ifap = &in_dev->ifa_list; (ifa = *ifap) != NULL;
19316 + ifap = &ifa->ifa_next) {
19317 + if (v4_addr_in_nx_info(nxi, ifa->ifa_local, NXA_MASK_SHOW)) {
19322 + in_dev_put(in_dev);
19328 +#ifdef CONFIG_IPV6
19330 +int v6_dev_in_nx_info(struct net_device *dev, struct nx_info *nxi)
19332 + struct inet6_dev *in_dev;
19333 + struct inet6_ifaddr *ifa;
19338 + in_dev = in6_dev_get(dev);
19342 + // for (ifap = &in_dev->addr_list; (ifa = *ifap) != NULL;
19343 + list_for_each_entry(ifa, &in_dev->addr_list, if_list) {
19344 + if (v6_addr_in_nx_info(nxi, &ifa->addr, -1)) {
19349 + in6_dev_put(in_dev);
19356 +int dev_in_nx_info(struct net_device *dev, struct nx_info *nxi)
19362 + if (nxi->v4.type && v4_dev_in_nx_info(dev, nxi))
19364 +#ifdef CONFIG_IPV6
19366 + if (nxi->v6.type && v6_dev_in_nx_info(dev, nxi))
19371 + vxdprintk(VXD_CBIT(net, 3),
19372 + "dev_in_nx_info(%p,%p[#%d]) = %d",
19373 + dev, nxi, nxi ? nxi->nx_id : 0, ret);
19377 +int ip_v4_find_src(struct net *net, struct nx_info *nxi,
19378 + struct rtable **rp, struct flowi *fl)
19383 + /* FIXME: handle lback only case */
19384 + if (!NX_IPV4(nxi))
19387 + vxdprintk(VXD_CBIT(net, 4),
19388 + "ip_v4_find_src(%p[#%u]) " NIPQUAD_FMT " -> " NIPQUAD_FMT,
19389 + nxi, nxi ? nxi->nx_id : 0,
19390 + NIPQUAD(fl->fl4_src), NIPQUAD(fl->fl4_dst));
19392 + /* single IP is unconditional */
19393 + if (nx_info_flags(nxi, NXF_SINGLE_IP, 0) &&
19394 + (fl->fl4_src == INADDR_ANY))
19395 + fl->fl4_src = nxi->v4.ip[0].s_addr;
19397 + if (fl->fl4_src == INADDR_ANY) {
19398 + struct nx_addr_v4 *ptr;
19399 + __be32 found = 0;
19402 + err = __ip_route_output_key(net, rp, fl);
19404 + found = (*rp)->rt_src;
19406 + vxdprintk(VXD_CBIT(net, 4),
19407 + "ip_v4_find_src(%p[#%u]) rok[%u]: " NIPQUAD_FMT,
19408 + nxi, nxi ? nxi->nx_id : 0, fl->oif, NIPQUAD(found));
19409 + if (v4_addr_in_nx_info(nxi, found, NXA_MASK_BIND))
19413 + for (ptr = &nxi->v4; ptr; ptr = ptr->next) {
19414 + __be32 primary = ptr->ip[0].s_addr;
19415 + __be32 mask = ptr->mask.s_addr;
19416 + __be32 neta = primary & mask;
19418 + vxdprintk(VXD_CBIT(net, 4), "ip_v4_find_src(%p[#%u]) chk: "
19419 + NIPQUAD_FMT "/" NIPQUAD_FMT "/" NIPQUAD_FMT,
19420 + nxi, nxi ? nxi->nx_id : 0, NIPQUAD(primary),
19421 + NIPQUAD(mask), NIPQUAD(neta));
19422 + if ((found & mask) != neta)
19425 + fl->fl4_src = primary;
19426 + err = __ip_route_output_key(net, rp, fl);
19427 + vxdprintk(VXD_CBIT(net, 4),
19428 + "ip_v4_find_src(%p[#%u]) rok[%u]: " NIPQUAD_FMT,
19429 + nxi, nxi ? nxi->nx_id : 0, fl->oif, NIPQUAD(primary));
19431 + found = (*rp)->rt_src;
19433 + if (found == primary)
19437 + /* still no source ip? */
19438 + found = ipv4_is_loopback(fl->fl4_dst)
19439 + ? IPI_LOOPBACK : nxi->v4.ip[0].s_addr;
19441 + /* assign src ip to flow */
19442 + fl->fl4_src = found;
19445 + if (!v4_addr_in_nx_info(nxi, fl->fl4_src, NXA_MASK_BIND))
19449 + if (nx_info_flags(nxi, NXF_LBACK_REMAP, 0)) {
19450 + if (ipv4_is_loopback(fl->fl4_dst))
19451 + fl->fl4_dst = nxi->v4_lback.s_addr;
19452 + if (ipv4_is_loopback(fl->fl4_src))
19453 + fl->fl4_src = nxi->v4_lback.s_addr;
19454 + } else if (ipv4_is_loopback(fl->fl4_dst) &&
19455 + !nx_info_flags(nxi, NXF_LBACK_ALLOW, 0))
19461 +EXPORT_SYMBOL_GPL(ip_v4_find_src);
19463 diff -NurpP --minimal linux-2.6.36/kernel/vserver/init.c linux-2.6.36-vs2.3.0.36.33/kernel/vserver/init.c
19464 --- linux-2.6.36/kernel/vserver/init.c 1970-01-01 01:00:00.000000000 +0100
19465 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/init.c 2010-10-21 13:09:36.000000000 +0200
19468 + * linux/kernel/init.c
19470 + * Virtual Server Init
19472 + * Copyright (C) 2004-2007 Herbert Pötzl
19474 + * V0.01 basic structure
19478 +#include <linux/init.h>
19480 +int vserver_register_sysctl(void);
19481 +void vserver_unregister_sysctl(void);
19484 +static int __init init_vserver(void)
19488 +#ifdef CONFIG_VSERVER_DEBUG
19489 + vserver_register_sysctl();
19495 +static void __exit exit_vserver(void)
19498 +#ifdef CONFIG_VSERVER_DEBUG
19499 + vserver_unregister_sysctl();
19504 +/* FIXME: GFP_ZONETYPES gone
19505 +long vx_slab[GFP_ZONETYPES]; */
19509 +module_init(init_vserver);
19510 +module_exit(exit_vserver);
19512 diff -NurpP --minimal linux-2.6.36/kernel/vserver/inode.c linux-2.6.36-vs2.3.0.36.33/kernel/vserver/inode.c
19513 --- linux-2.6.36/kernel/vserver/inode.c 1970-01-01 01:00:00.000000000 +0100
19514 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/inode.c 2010-10-21 14:46:31.000000000 +0200
19517 + * linux/kernel/vserver/inode.c
19519 + * Virtual Server: File System Support
19521 + * Copyright (C) 2004-2007 Herbert Pötzl
19523 + * V0.01 separated from vcontext V0.05
19524 + * V0.02 moved to tag (instead of xid)
19528 +#include <linux/tty.h>
19529 +#include <linux/proc_fs.h>
19530 +#include <linux/devpts_fs.h>
19531 +#include <linux/fs.h>
19532 +#include <linux/file.h>
19533 +#include <linux/mount.h>
19534 +#include <linux/parser.h>
19535 +#include <linux/namei.h>
19536 +#include <linux/vserver/inode.h>
19537 +#include <linux/vserver/inode_cmd.h>
19538 +#include <linux/vs_base.h>
19539 +#include <linux/vs_tag.h>
19541 +#include <asm/uaccess.h>
19544 +static int __vc_get_iattr(struct inode *in, uint32_t *tag, uint32_t *flags, uint32_t *mask)
19546 + struct proc_dir_entry *entry;
19548 + if (!in || !in->i_sb)
19551 + *flags = IATTR_TAG
19552 + | (IS_IMMUTABLE(in) ? IATTR_IMMUTABLE : 0)
19553 + | (IS_IXUNLINK(in) ? IATTR_IXUNLINK : 0)
19554 + | (IS_BARRIER(in) ? IATTR_BARRIER : 0)
19555 + | (IS_COW(in) ? IATTR_COW : 0);
19556 + *mask = IATTR_IXUNLINK | IATTR_IMMUTABLE | IATTR_COW;
19558 + if (S_ISDIR(in->i_mode))
19559 + *mask |= IATTR_BARRIER;
19561 + if (IS_TAGGED(in)) {
19562 + *tag = in->i_tag;
19563 + *mask |= IATTR_TAG;
19566 + switch (in->i_sb->s_magic) {
19567 + case PROC_SUPER_MAGIC:
19568 + entry = PROC_I(in)->pde;
19570 + /* check for specific inodes? */
19572 + *mask |= IATTR_FLAGS;
19574 + *flags |= (entry->vx_flags & IATTR_FLAGS);
19576 + *flags |= (PROC_I(in)->vx_flags & IATTR_FLAGS);
19579 + case DEVPTS_SUPER_MAGIC:
19580 + *tag = in->i_tag;
19581 + *mask |= IATTR_TAG;
19590 +int vc_get_iattr(void __user *data)
19592 + struct path path;
19593 + struct vcmd_ctx_iattr_v1 vc_data = { .tag = -1 };
19596 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19599 + ret = user_lpath(vc_data.name, &path);
19601 + ret = __vc_get_iattr(path.dentry->d_inode,
19602 + &vc_data.tag, &vc_data.flags, &vc_data.mask);
19608 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19613 +#ifdef CONFIG_COMPAT
19615 +int vc_get_iattr_x32(void __user *data)
19617 + struct path path;
19618 + struct vcmd_ctx_iattr_v1_x32 vc_data = { .tag = -1 };
19621 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19624 + ret = user_lpath(compat_ptr(vc_data.name_ptr), &path);
19626 + ret = __vc_get_iattr(path.dentry->d_inode,
19627 + &vc_data.tag, &vc_data.flags, &vc_data.mask);
19633 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19638 +#endif /* CONFIG_COMPAT */
19641 +int vc_fget_iattr(uint32_t fd, void __user *data)
19643 + struct file *filp;
19644 + struct vcmd_ctx_fiattr_v0 vc_data = { .tag = -1 };
19647 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19651 + if (!filp || !filp->f_dentry || !filp->f_dentry->d_inode)
19654 + ret = __vc_get_iattr(filp->f_dentry->d_inode,
19655 + &vc_data.tag, &vc_data.flags, &vc_data.mask);
19659 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19665 +static int __vc_set_iattr(struct dentry *de, uint32_t *tag, uint32_t *flags, uint32_t *mask)
19667 + struct inode *in = de->d_inode;
19668 + int error = 0, is_proc = 0, has_tag = 0;
19669 + struct iattr attr = { 0 };
19671 + if (!in || !in->i_sb)
19674 + is_proc = (in->i_sb->s_magic == PROC_SUPER_MAGIC);
19675 + if ((*mask & IATTR_FLAGS) && !is_proc)
19678 + has_tag = IS_TAGGED(in) ||
19679 + (in->i_sb->s_magic == DEVPTS_SUPER_MAGIC);
19680 + if ((*mask & IATTR_TAG) && !has_tag)
19683 + mutex_lock(&in->i_mutex);
19684 + if (*mask & IATTR_TAG) {
19685 + attr.ia_tag = *tag;
19686 + attr.ia_valid |= ATTR_TAG;
19689 + if (*mask & IATTR_FLAGS) {
19690 + struct proc_dir_entry *entry = PROC_I(in)->pde;
19691 + unsigned int iflags = PROC_I(in)->vx_flags;
19693 + iflags = (iflags & ~(*mask & IATTR_FLAGS))
19694 + | (*flags & IATTR_FLAGS);
19695 + PROC_I(in)->vx_flags = iflags;
19697 + entry->vx_flags = iflags;
19700 + if (*mask & (IATTR_IMMUTABLE | IATTR_IXUNLINK |
19701 + IATTR_BARRIER | IATTR_COW)) {
19702 + int iflags = in->i_flags;
19703 + int vflags = in->i_vflags;
19705 + if (*mask & IATTR_IMMUTABLE) {
19706 + if (*flags & IATTR_IMMUTABLE)
19707 + iflags |= S_IMMUTABLE;
19709 + iflags &= ~S_IMMUTABLE;
19711 + if (*mask & IATTR_IXUNLINK) {
19712 + if (*flags & IATTR_IXUNLINK)
19713 + iflags |= S_IXUNLINK;
19715 + iflags &= ~S_IXUNLINK;
19717 + if (S_ISDIR(in->i_mode) && (*mask & IATTR_BARRIER)) {
19718 + if (*flags & IATTR_BARRIER)
19719 + vflags |= V_BARRIER;
19721 + vflags &= ~V_BARRIER;
19723 + if (S_ISREG(in->i_mode) && (*mask & IATTR_COW)) {
19724 + if (*flags & IATTR_COW)
19727 + vflags &= ~V_COW;
19729 + if (in->i_op && in->i_op->sync_flags) {
19730 + error = in->i_op->sync_flags(in, iflags, vflags);
19736 + if (attr.ia_valid) {
19737 + if (in->i_op && in->i_op->setattr)
19738 + error = in->i_op->setattr(de, &attr);
19740 + error = inode_change_ok(in, &attr);
19742 + setattr_copy(in, &attr);
19743 + mark_inode_dirty(in);
19749 + mutex_unlock(&in->i_mutex);
19753 +int vc_set_iattr(void __user *data)
19755 + struct path path;
19756 + struct vcmd_ctx_iattr_v1 vc_data;
19759 + if (!capable(CAP_LINUX_IMMUTABLE))
19761 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19764 + ret = user_lpath(vc_data.name, &path);
19766 + ret = __vc_set_iattr(path.dentry,
19767 + &vc_data.tag, &vc_data.flags, &vc_data.mask);
19771 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19776 +#ifdef CONFIG_COMPAT
19778 +int vc_set_iattr_x32(void __user *data)
19780 + struct path path;
19781 + struct vcmd_ctx_iattr_v1_x32 vc_data;
19784 + if (!capable(CAP_LINUX_IMMUTABLE))
19786 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19789 + ret = user_lpath(compat_ptr(vc_data.name_ptr), &path);
19791 + ret = __vc_set_iattr(path.dentry,
19792 + &vc_data.tag, &vc_data.flags, &vc_data.mask);
19796 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19801 +#endif /* CONFIG_COMPAT */
19803 +int vc_fset_iattr(uint32_t fd, void __user *data)
19805 + struct file *filp;
19806 + struct vcmd_ctx_fiattr_v0 vc_data;
19809 + if (!capable(CAP_LINUX_IMMUTABLE))
19811 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19815 + if (!filp || !filp->f_dentry || !filp->f_dentry->d_inode)
19818 + ret = __vc_set_iattr(filp->f_dentry, &vc_data.tag,
19819 + &vc_data.flags, &vc_data.mask);
19823 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19829 +enum { Opt_notagcheck, Opt_tag, Opt_notag, Opt_tagid, Opt_err };
19831 +static match_table_t tokens = {
19832 + {Opt_notagcheck, "notagcheck"},
19833 +#ifdef CONFIG_PROPAGATE
19834 + {Opt_notag, "notag"},
19835 + {Opt_tag, "tag"},
19836 + {Opt_tagid, "tagid=%u"},
19842 +static void __dx_parse_remove(char *string, char *opt)
19844 + char *p = strstr(string, opt);
19848 + while (*q != '\0' && *q != ',')
19857 +int dx_parse_tag(char *string, tag_t *tag, int remove, int *mnt_flags,
19858 + unsigned long *flags)
19861 + substring_t args[MAX_OPT_ARGS];
19862 + int token, option = 0;
19863 + char *s, *p, *opts;
19867 + s = kstrdup(string, GFP_KERNEL | GFP_ATOMIC);
19872 + while ((p = strsep(&opts, ",")) != NULL) {
19873 + token = match_token(p, tokens, args);
19875 + vxdprintk(VXD_CBIT(tag, 7),
19876 + "dx_parse_tag(»%s«): %d:#%d",
19877 + p, token, option);
19880 +#ifdef CONFIG_PROPAGATE
19885 + __dx_parse_remove(s, "tag");
19886 + *mnt_flags |= MNT_TAGID;
19887 + set |= MNT_TAGID;
19891 + __dx_parse_remove(s, "notag");
19892 + *mnt_flags |= MNT_NOTAG;
19893 + set |= MNT_NOTAG;
19896 + if (tag && !match_int(args, &option))
19899 + __dx_parse_remove(s, "tagid");
19900 + *mnt_flags |= MNT_TAGID;
19901 + set |= MNT_TAGID;
19904 + case Opt_notagcheck:
19906 + __dx_parse_remove(s, "notagcheck");
19907 + *flags |= MS_NOTAGCHECK;
19908 + set |= MS_NOTAGCHECK;
19913 + strcpy(string, s);
19918 +#ifdef CONFIG_PROPAGATE
19920 +void __dx_propagate_tag(struct nameidata *nd, struct inode *inode)
19922 + tag_t new_tag = 0;
19923 + struct vfsmount *mnt;
19928 + mnt = nd->path.mnt;
19932 + propagate = (mnt->mnt_flags & MNT_TAGID);
19934 + new_tag = mnt->mnt_tag;
19936 + vxdprintk(VXD_CBIT(tag, 7),
19937 + "dx_propagate_tag(%p[#%lu.%d]): %d,%d",
19938 + inode, inode->i_ino, inode->i_tag,
19939 + new_tag, (propagate) ? 1 : 0);
19942 + inode->i_tag = new_tag;
19945 +#include <linux/module.h>
19947 +EXPORT_SYMBOL_GPL(__dx_propagate_tag);
19949 +#endif /* CONFIG_PROPAGATE */
19951 diff -NurpP --minimal linux-2.6.36/kernel/vserver/Kconfig linux-2.6.36-vs2.3.0.36.33/kernel/vserver/Kconfig
19952 --- linux-2.6.36/kernel/vserver/Kconfig 1970-01-01 01:00:00.000000000 +0100
19953 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/Kconfig 2010-10-21 13:09:36.000000000 +0200
19956 +# Linux VServer configuration
19959 +menu "Linux VServer"
19961 +config VSERVER_AUTO_LBACK
19962 + bool "Automatically Assign Loopback IP"
19965 + Automatically assign a guest specific loopback
19966 + IP and add it to the kernel network stack on
19969 +config VSERVER_AUTO_SINGLE
19970 + bool "Automatic Single IP Special Casing"
19971 + depends on EXPERIMENTAL
19974 + This allows network contexts with a single IP to
19975 + automatically remap 0.0.0.0 bindings to that IP,
19976 + avoiding further network checks and improving
19979 + (note: such guests do not allow to change the ip
19980 + on the fly and do not show loopback addresses)
19982 +config VSERVER_COWBL
19983 + bool "Enable COW Immutable Link Breaking"
19986 + This enables the COW (Copy-On-Write) link break code.
19987 + It allows you to treat unified files like normal files
19988 + when writing to them (which will implicitely break the
19989 + link and create a copy of the unified file)
19991 +config VSERVER_VTIME
19992 + bool "Enable Virtualized Guest Time"
19993 + depends on EXPERIMENTAL
19996 + This enables per guest time offsets to allow for
19997 + adjusting the system clock individually per guest.
19998 + this adds some overhead to the time functions and
19999 + therefore should not be enabled without good reason.
20001 +config VSERVER_DEVICE
20002 + bool "Enable Guest Device Mapping"
20003 + depends on EXPERIMENTAL
20006 + This enables generic device remapping.
20008 +config VSERVER_PROC_SECURE
20009 + bool "Enable Proc Security"
20010 + depends on PROC_FS
20013 + This configures ProcFS security to initially hide
20014 + non-process entries for all contexts except the main and
20015 + spectator context (i.e. for all guests), which is a secure
20018 + (note: on 1.2x the entries were visible by default)
20020 +config VSERVER_HARDCPU
20021 + bool "Enable Hard CPU Limits"
20024 + Activate the Hard CPU Limits
20026 + This will compile in code that allows the Token Bucket
20027 + Scheduler to put processes on hold when a context's
20028 + tokens are depleted (provided that its per-context
20029 + sched_hard flag is set).
20031 + Processes belonging to that context will not be able
20032 + to consume CPU resources again until a per-context
20033 + configured minimum of tokens has been reached.
20035 +config VSERVER_IDLETIME
20036 + bool "Avoid idle CPUs by skipping Time"
20037 + depends on VSERVER_HARDCPU
20040 + This option allows the scheduler to artificially
20041 + advance time (per cpu) when otherwise the idle
20042 + task would be scheduled, thus keeping the cpu
20043 + busy and sharing the available resources among
20044 + certain contexts.
20046 +config VSERVER_IDLELIMIT
20047 + bool "Limit the IDLE task"
20048 + depends on VSERVER_HARDCPU
20051 + Limit the idle slices, so the the next context
20052 + will be scheduled as soon as possible.
20054 + This might improve interactivity and latency, but
20055 + will also marginally increase scheduling overhead.
20058 + prompt "Persistent Inode Tagging"
20059 + default TAGGING_ID24
20061 + This adds persistent context information to filesystems
20062 + mounted with the tagxid option. Tagging is a requirement
20063 + for per-context disk limits and per-context quota.
20066 +config TAGGING_NONE
20069 + do not store per-context information in inodes.
20071 +config TAGGING_UID16
20072 + bool "UID16/GID32"
20074 + reduces UID to 16 bit, but leaves GID at 32 bit.
20076 +config TAGGING_GID16
20077 + bool "UID32/GID16"
20079 + reduces GID to 16 bit, but leaves UID at 32 bit.
20081 +config TAGGING_ID24
20082 + bool "UID24/GID24"
20084 + uses the upper 8bit from UID and GID for XID tagging
20085 + which leaves 24bit for UID/GID each, which should be
20086 + more than sufficient for normal use.
20088 +config TAGGING_INTERN
20089 + bool "UID32/GID32"
20091 + this uses otherwise reserved inode fields in the on
20092 + disk representation, which limits the use to a few
20093 + filesystems (currently ext2 and ext3)
20098 + bool "Tag NFSD User Auth and Files"
20101 + Enable this if you do want the in-kernel NFS
20102 + Server to use the tagging specified above.
20103 + (will require patched clients too)
20105 +config VSERVER_PRIVACY
20106 + bool "Honor Privacy Aspects of Guests"
20109 + When enabled, most context checks will disallow
20110 + access to structures assigned to a specific context,
20111 + like ptys or loop devices.
20113 +config VSERVER_CONTEXTS
20114 + int "Maximum number of Contexts (1-65533)" if EMBEDDED
20116 + default "768" if 64BIT
20119 + This setting will optimize certain data structures
20120 + and memory allocations according to the expected
20123 + note: this is not a strict upper limit.
20125 +config VSERVER_WARN
20126 + bool "VServer Warnings"
20129 + This enables various runtime warnings, which will
20130 + notify about potential manipulation attempts or
20131 + resource shortage. It is generally considered to
20132 + be a good idea to have that enabled.
20134 +config VSERVER_DEBUG
20135 + bool "VServer Debugging Code"
20138 + Set this to yes if you want to be able to activate
20139 + debugging output at runtime. It adds a very small
20140 + overhead to all vserver related functions and
20141 + increases the kernel size by about 20k.
20143 +config VSERVER_HISTORY
20144 + bool "VServer History Tracing"
20145 + depends on VSERVER_DEBUG
20148 + Set this to yes if you want to record the history of
20149 + linux-vserver activities, so they can be replayed in
20150 + the event of a kernel panic or oops.
20152 +config VSERVER_HISTORY_SIZE
20153 + int "Per-CPU History Size (32-65536)"
20154 + depends on VSERVER_HISTORY
20158 + This allows you to specify the number of entries in
20159 + the per-CPU history buffer.
20161 +config VSERVER_MONITOR
20162 + bool "VServer Scheduling Monitor"
20163 + depends on VSERVER_DISABLED
20166 + Set this to yes if you want to record the scheduling
20167 + decisions, so that they can be relayed to userspace
20168 + for detailed analysis.
20170 +config VSERVER_MONITOR_SIZE
20171 + int "Per-CPU Monitor Queue Size (32-65536)"
20172 + depends on VSERVER_MONITOR
20176 + This allows you to specify the number of entries in
20177 + the per-CPU scheduling monitor buffer.
20179 +config VSERVER_MONITOR_SYNC
20180 + int "Per-CPU Monitor Sync Interval (0-65536)"
20181 + depends on VSERVER_MONITOR
20185 + This allows you to specify the interval in ticks
20186 + when a time sync entry is inserted.
20188 +config VSERVER_LEGACY_MEM
20189 + bool "Legacy Memory Limits"
20192 + This provides fake memory limits to keep
20193 + older tools happy in the face of memory
20203 + select NAMESPACES
20209 +config VSERVER_SECURITY
20211 + depends on SECURITY
20213 + select SECURITY_CAPABILITIES
20215 diff -NurpP --minimal linux-2.6.36/kernel/vserver/limit.c linux-2.6.36-vs2.3.0.36.33/kernel/vserver/limit.c
20216 --- linux-2.6.36/kernel/vserver/limit.c 1970-01-01 01:00:00.000000000 +0100
20217 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/limit.c 2010-10-21 13:09:36.000000000 +0200
20220 + * linux/kernel/vserver/limit.c
20222 + * Virtual Server: Context Limits
20224 + * Copyright (C) 2004-2010 Herbert Pötzl
20226 + * V0.01 broken out from vcontext V0.05
20227 + * V0.02 changed vcmds to vxi arg
20228 + * V0.03 added memory cgroup support
20232 +#include <linux/sched.h>
20233 +#include <linux/module.h>
20234 +#include <linux/memcontrol.h>
20235 +#include <linux/res_counter.h>
20236 +#include <linux/vs_limit.h>
20237 +#include <linux/vserver/limit.h>
20238 +#include <linux/vserver/limit_cmd.h>
20240 +#include <asm/uaccess.h>
20243 +const char *vlimit_name[NUM_LIMITS] = {
20244 +#ifdef CONFIG_VSERVER_LEGACY_MEM
20245 + [RLIMIT_RSS] = "RSS",
20246 + [RLIMIT_AS] = "VM",
20247 +#endif /* CONFIG_VSERVER_LEGACY_MEM */
20248 + [RLIMIT_CPU] = "CPU",
20249 + [RLIMIT_NPROC] = "NPROC",
20250 + [RLIMIT_NOFILE] = "NOFILE",
20251 + [RLIMIT_LOCKS] = "LOCKS",
20252 + [RLIMIT_SIGPENDING] = "SIGP",
20253 + [RLIMIT_MSGQUEUE] = "MSGQ",
20255 + [VLIMIT_NSOCK] = "NSOCK",
20256 + [VLIMIT_OPENFD] = "OPENFD",
20257 + [VLIMIT_SHMEM] = "SHMEM",
20258 + [VLIMIT_DENTRY] = "DENTRY",
20261 +EXPORT_SYMBOL_GPL(vlimit_name);
20263 +#define MASK_ENTRY(x) (1 << (x))
20265 +const struct vcmd_ctx_rlimit_mask_v0 vlimit_mask = {
20268 + , /* softlimit */
20269 +#ifdef CONFIG_VSERVER_LEGACY_MEM
20270 + MASK_ENTRY( RLIMIT_RSS ) |
20271 +#endif /* CONFIG_VSERVER_LEGACY_MEM */
20274 +#ifdef CONFIG_VSERVER_LEGACY_MEM
20275 + MASK_ENTRY( RLIMIT_RSS ) |
20276 + MASK_ENTRY( RLIMIT_AS ) |
20277 +#endif /* CONFIG_VSERVER_LEGACY_MEM */
20278 + MASK_ENTRY( RLIMIT_NPROC ) |
20279 + MASK_ENTRY( RLIMIT_NOFILE ) |
20280 + MASK_ENTRY( RLIMIT_LOCKS ) |
20281 + MASK_ENTRY( RLIMIT_MSGQUEUE ) |
20283 + MASK_ENTRY( VLIMIT_NSOCK ) |
20284 + MASK_ENTRY( VLIMIT_OPENFD ) |
20285 + MASK_ENTRY( VLIMIT_SHMEM ) |
20286 + MASK_ENTRY( VLIMIT_DENTRY ) |
20289 + /* accounting only */
20290 +uint32_t account_mask =
20291 + MASK_ENTRY( VLIMIT_SEMARY ) |
20292 + MASK_ENTRY( VLIMIT_NSEMS ) |
20293 + MASK_ENTRY( VLIMIT_MAPPED ) |
20297 +static int is_valid_vlimit(int id)
20299 + uint32_t mask = vlimit_mask.minimum |
20300 + vlimit_mask.softlimit | vlimit_mask.maximum;
20301 + return mask & (1 << id);
20304 +static int is_accounted_vlimit(int id)
20306 + if (is_valid_vlimit(id))
20308 + return account_mask & (1 << id);
20312 +static inline uint64_t vc_get_soft(struct vx_info *vxi, int id)
20314 + rlim_t limit = __rlim_soft(&vxi->limit, id);
20315 + return VX_VLIM(limit);
20318 +static inline uint64_t vc_get_hard(struct vx_info *vxi, int id)
20320 + rlim_t limit = __rlim_hard(&vxi->limit, id);
20321 + return VX_VLIM(limit);
20324 +static int do_get_rlimit(struct vx_info *vxi, uint32_t id,
20325 + uint64_t *minimum, uint64_t *softlimit, uint64_t *maximum)
20327 + if (!is_valid_vlimit(id))
20331 + *minimum = CRLIM_UNSET;
20333 + *softlimit = vc_get_soft(vxi, id);
20335 + *maximum = vc_get_hard(vxi, id);
20339 +int vc_get_rlimit(struct vx_info *vxi, void __user *data)
20341 + struct vcmd_ctx_rlimit_v0 vc_data;
20344 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20347 + ret = do_get_rlimit(vxi, vc_data.id,
20348 + &vc_data.minimum, &vc_data.softlimit, &vc_data.maximum);
20352 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
20357 +static int do_set_rlimit(struct vx_info *vxi, uint32_t id,
20358 + uint64_t minimum, uint64_t softlimit, uint64_t maximum)
20360 + if (!is_valid_vlimit(id))
20363 + if (maximum != CRLIM_KEEP)
20364 + __rlim_hard(&vxi->limit, id) = VX_RLIM(maximum);
20365 + if (softlimit != CRLIM_KEEP)
20366 + __rlim_soft(&vxi->limit, id) = VX_RLIM(softlimit);
20368 + /* clamp soft limit */
20369 + if (__rlim_soft(&vxi->limit, id) > __rlim_hard(&vxi->limit, id))
20370 + __rlim_soft(&vxi->limit, id) = __rlim_hard(&vxi->limit, id);
20375 +int vc_set_rlimit(struct vx_info *vxi, void __user *data)
20377 + struct vcmd_ctx_rlimit_v0 vc_data;
20379 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20382 + return do_set_rlimit(vxi, vc_data.id,
20383 + vc_data.minimum, vc_data.softlimit, vc_data.maximum);
20386 +#ifdef CONFIG_IA32_EMULATION
20388 +int vc_set_rlimit_x32(struct vx_info *vxi, void __user *data)
20390 + struct vcmd_ctx_rlimit_v0_x32 vc_data;
20392 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20395 + return do_set_rlimit(vxi, vc_data.id,
20396 + vc_data.minimum, vc_data.softlimit, vc_data.maximum);
20399 +int vc_get_rlimit_x32(struct vx_info *vxi, void __user *data)
20401 + struct vcmd_ctx_rlimit_v0_x32 vc_data;
20404 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20407 + ret = do_get_rlimit(vxi, vc_data.id,
20408 + &vc_data.minimum, &vc_data.softlimit, &vc_data.maximum);
20412 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
20417 +#endif /* CONFIG_IA32_EMULATION */
20420 +int vc_get_rlimit_mask(uint32_t id, void __user *data)
20422 + if (copy_to_user(data, &vlimit_mask, sizeof(vlimit_mask)))
20428 +static inline void vx_reset_hits(struct _vx_limit *limit)
20432 + for (lim = 0; lim < NUM_LIMITS; lim++) {
20433 + atomic_set(&__rlim_lhit(limit, lim), 0);
20437 +int vc_reset_hits(struct vx_info *vxi, void __user *data)
20439 + vx_reset_hits(&vxi->limit);
20443 +static inline void vx_reset_minmax(struct _vx_limit *limit)
20448 + for (lim = 0; lim < NUM_LIMITS; lim++) {
20449 + value = __rlim_get(limit, lim);
20450 + __rlim_rmax(limit, lim) = value;
20451 + __rlim_rmin(limit, lim) = value;
20455 +int vc_reset_minmax(struct vx_info *vxi, void __user *data)
20457 + vx_reset_minmax(&vxi->limit);
20462 +int vc_rlimit_stat(struct vx_info *vxi, void __user *data)
20464 + struct vcmd_rlimit_stat_v0 vc_data;
20465 + struct _vx_limit *limit = &vxi->limit;
20468 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20472 + if (!is_accounted_vlimit(id))
20475 + vx_limit_fixup(limit, id);
20476 + vc_data.hits = atomic_read(&__rlim_lhit(limit, id));
20477 + vc_data.value = __rlim_get(limit, id);
20478 + vc_data.minimum = __rlim_rmin(limit, id);
20479 + vc_data.maximum = __rlim_rmax(limit, id);
20481 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
20487 +void vx_vsi_meminfo(struct sysinfo *val)
20489 +#ifdef CONFIG_CGROUP_MEM_RES_CTLR
20490 + struct mem_cgroup *mcg = mem_cgroup_from_task(current);
20491 + u64 res_limit, res_usage;
20496 + res_limit = mem_cgroup_res_read_u64(mcg, RES_LIMIT);
20497 + res_usage = mem_cgroup_res_read_u64(mcg, RES_USAGE);
20499 + if (res_limit != RESOURCE_MAX)
20500 + val->totalram = (res_limit >> PAGE_SHIFT);
20501 + val->freeram = val->totalram - (res_usage >> PAGE_SHIFT);
20502 + val->bufferram = 0;
20503 + val->totalhigh = 0;
20504 + val->freehigh = 0;
20505 +#endif /* CONFIG_CGROUP_MEM_RES_CTLR */
20509 +void vx_vsi_swapinfo(struct sysinfo *val)
20511 +#ifdef CONFIG_CGROUP_MEM_RES_CTLR
20512 +#ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP
20513 + struct mem_cgroup *mcg = mem_cgroup_from_task(current);
20514 + u64 res_limit, res_usage, memsw_limit, memsw_usage;
20515 + s64 swap_limit, swap_usage;
20520 + res_limit = mem_cgroup_res_read_u64(mcg, RES_LIMIT);
20521 + res_usage = mem_cgroup_res_read_u64(mcg, RES_USAGE);
20522 + memsw_limit = mem_cgroup_memsw_read_u64(mcg, RES_LIMIT);
20523 + memsw_usage = mem_cgroup_memsw_read_u64(mcg, RES_USAGE);
20525 + if (res_limit == RESOURCE_MAX)
20528 + swap_limit = memsw_limit - res_limit;
20529 + if (memsw_limit != RESOURCE_MAX)
20530 + val->totalswap = swap_limit >> PAGE_SHIFT;
20532 + swap_usage = memsw_usage - res_usage;
20533 + val->freeswap = (swap_usage < swap_limit) ?
20534 + val->totalswap - (swap_usage >> PAGE_SHIFT) : 0;
20535 +#else /* !CONFIG_CGROUP_MEM_RES_CTLR_SWAP */
20536 + val->totalswap = 0;
20537 + val->freeswap = 0;
20538 +#endif /* !CONFIG_CGROUP_MEM_RES_CTLR_SWAP */
20539 +#endif /* CONFIG_CGROUP_MEM_RES_CTLR */
20543 +long vx_vsi_cached(struct sysinfo *val)
20545 +#ifdef CONFIG_CGROUP_MEM_RES_CTLR
20546 + struct mem_cgroup *mcg = mem_cgroup_from_task(current);
20548 + return mem_cgroup_stat_read_cache(mcg);
20555 +unsigned long vx_badness(struct task_struct *task, struct mm_struct *mm)
20557 + struct vx_info *vxi = mm->mm_vx_info;
20558 + unsigned long points;
20564 + points = vxi->vx_badness_bias;
20566 + v = __vx_cres_array_fixup(&vxi->limit, VLA_RSS);
20567 + w = __rlim_soft(&vxi->limit, RLIMIT_RSS);
20568 + points += (v > w) ? (v - w) : 0;
20573 diff -NurpP --minimal linux-2.6.36/kernel/vserver/limit_init.h linux-2.6.36-vs2.3.0.36.33/kernel/vserver/limit_init.h
20574 --- linux-2.6.36/kernel/vserver/limit_init.h 1970-01-01 01:00:00.000000000 +0100
20575 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/limit_init.h 2010-10-21 13:09:36.000000000 +0200
20579 +static inline void vx_info_init_limit(struct _vx_limit *limit)
20583 + for (lim = 0; lim < NUM_LIMITS; lim++) {
20584 + __rlim_soft(limit, lim) = RLIM_INFINITY;
20585 + __rlim_hard(limit, lim) = RLIM_INFINITY;
20586 + __rlim_set(limit, lim, 0);
20587 + atomic_set(&__rlim_lhit(limit, lim), 0);
20588 + __rlim_rmin(limit, lim) = 0;
20589 + __rlim_rmax(limit, lim) = 0;
20593 +static inline void vx_info_exit_limit(struct _vx_limit *limit)
20598 + for (lim = 0; lim < NUM_LIMITS; lim++) {
20599 + if ((1 << lim) & VLIM_NOCHECK)
20601 + value = __rlim_get(limit, lim);
20602 + vxwprintk_xid(value,
20603 + "!!! limit: %p[%s,%d] = %ld on exit.",
20604 + limit, vlimit_name[lim], lim, (long)value);
20608 diff -NurpP --minimal linux-2.6.36/kernel/vserver/limit_proc.h linux-2.6.36-vs2.3.0.36.33/kernel/vserver/limit_proc.h
20609 --- linux-2.6.36/kernel/vserver/limit_proc.h 1970-01-01 01:00:00.000000000 +0100
20610 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/limit_proc.h 2010-10-21 13:09:36.000000000 +0200
20612 +#ifndef _VX_LIMIT_PROC_H
20613 +#define _VX_LIMIT_PROC_H
20615 +#include <linux/vserver/limit_int.h>
20618 +#define VX_LIMIT_FMT ":\t%8ld\t%8ld/%8ld\t%8lld/%8lld\t%6d\n"
20619 +#define VX_LIMIT_TOP \
20620 + "Limit\t current\t min/max\t\t soft/hard\t\thits\n"
20622 +#define VX_LIMIT_ARG(r) \
20623 + (unsigned long)__rlim_get(limit, r), \
20624 + (unsigned long)__rlim_rmin(limit, r), \
20625 + (unsigned long)__rlim_rmax(limit, r), \
20626 + VX_VLIM(__rlim_soft(limit, r)), \
20627 + VX_VLIM(__rlim_hard(limit, r)), \
20628 + atomic_read(&__rlim_lhit(limit, r))
20630 +static inline int vx_info_proc_limit(struct _vx_limit *limit, char *buffer)
20632 + vx_limit_fixup(limit, -1);
20633 + return sprintf(buffer, VX_LIMIT_TOP
20634 + "PROC" VX_LIMIT_FMT
20635 + "VM" VX_LIMIT_FMT
20636 + "VML" VX_LIMIT_FMT
20637 + "RSS" VX_LIMIT_FMT
20638 + "ANON" VX_LIMIT_FMT
20639 + "RMAP" VX_LIMIT_FMT
20640 + "FILES" VX_LIMIT_FMT
20641 + "OFD" VX_LIMIT_FMT
20642 + "LOCKS" VX_LIMIT_FMT
20643 + "SOCK" VX_LIMIT_FMT
20644 + "MSGQ" VX_LIMIT_FMT
20645 + "SHM" VX_LIMIT_FMT
20646 + "SEMA" VX_LIMIT_FMT
20647 + "SEMS" VX_LIMIT_FMT
20648 + "DENT" VX_LIMIT_FMT,
20649 + VX_LIMIT_ARG(RLIMIT_NPROC),
20650 + VX_LIMIT_ARG(RLIMIT_AS),
20651 + VX_LIMIT_ARG(RLIMIT_MEMLOCK),
20652 + VX_LIMIT_ARG(RLIMIT_RSS),
20653 + VX_LIMIT_ARG(VLIMIT_ANON),
20654 + VX_LIMIT_ARG(VLIMIT_MAPPED),
20655 + VX_LIMIT_ARG(RLIMIT_NOFILE),
20656 + VX_LIMIT_ARG(VLIMIT_OPENFD),
20657 + VX_LIMIT_ARG(RLIMIT_LOCKS),
20658 + VX_LIMIT_ARG(VLIMIT_NSOCK),
20659 + VX_LIMIT_ARG(RLIMIT_MSGQUEUE),
20660 + VX_LIMIT_ARG(VLIMIT_SHMEM),
20661 + VX_LIMIT_ARG(VLIMIT_SEMARY),
20662 + VX_LIMIT_ARG(VLIMIT_NSEMS),
20663 + VX_LIMIT_ARG(VLIMIT_DENTRY));
20666 +#endif /* _VX_LIMIT_PROC_H */
20669 diff -NurpP --minimal linux-2.6.36/kernel/vserver/Makefile linux-2.6.36-vs2.3.0.36.33/kernel/vserver/Makefile
20670 --- linux-2.6.36/kernel/vserver/Makefile 1970-01-01 01:00:00.000000000 +0100
20671 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/Makefile 2010-10-21 13:09:36.000000000 +0200
20674 +# Makefile for the Linux vserver routines.
20678 +obj-y += vserver.o
20680 +vserver-y := switch.o context.o space.o sched.o network.o inode.o \
20681 + limit.o cvirt.o cacct.o signal.o helper.o init.o \
20684 +vserver-$(CONFIG_INET) += inet.o
20685 +vserver-$(CONFIG_PROC_FS) += proc.o
20686 +vserver-$(CONFIG_VSERVER_DEBUG) += sysctl.o debug.o
20687 +vserver-$(CONFIG_VSERVER_HISTORY) += history.o
20688 +vserver-$(CONFIG_VSERVER_MONITOR) += monitor.o
20689 +vserver-$(CONFIG_VSERVER_DEVICE) += device.o
20691 diff -NurpP --minimal linux-2.6.36/kernel/vserver/monitor.c linux-2.6.36-vs2.3.0.36.33/kernel/vserver/monitor.c
20692 --- linux-2.6.36/kernel/vserver/monitor.c 1970-01-01 01:00:00.000000000 +0100
20693 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/monitor.c 2010-10-21 13:09:36.000000000 +0200
20696 + * kernel/vserver/monitor.c
20698 + * Virtual Context Scheduler Monitor
20700 + * Copyright (C) 2006-2007 Herbert Pötzl
20702 + * V0.01 basic design
20706 +#include <linux/module.h>
20707 +#include <linux/jiffies.h>
20708 +#include <asm/uaccess.h>
20709 +#include <asm/atomic.h>
20711 +#include <linux/vserver/monitor.h>
20712 +#include <linux/vserver/debug_cmd.h>
20715 +#ifdef CONFIG_VSERVER_MONITOR
20716 +#define VXM_SIZE CONFIG_VSERVER_MONITOR_SIZE
20718 +#define VXM_SIZE 64
20721 +struct _vx_monitor {
20722 + unsigned int counter;
20724 + struct _vx_mon_entry entry[VXM_SIZE+1];
20728 +DEFINE_PER_CPU(struct _vx_monitor, vx_monitor_buffer);
20730 +unsigned volatile int vxm_active = 1;
20732 +static atomic_t sequence = ATOMIC_INIT(0);
20737 + * requires disabled preemption */
20739 +struct _vx_mon_entry *vxm_advance(int cpu)
20741 + struct _vx_monitor *mon = &per_cpu(vx_monitor_buffer, cpu);
20742 + struct _vx_mon_entry *entry;
20743 + unsigned int index;
20745 + index = vxm_active ? (mon->counter++ % VXM_SIZE) : VXM_SIZE;
20746 + entry = &mon->entry[index];
20748 + entry->ev.seq = atomic_inc_return(&sequence);
20749 + entry->ev.jif = jiffies;
20753 +EXPORT_SYMBOL_GPL(vxm_advance);
20756 +int do_read_monitor(struct __user _vx_mon_entry *data,
20757 + int cpu, uint32_t *index, uint32_t *count)
20759 + int pos, ret = 0;
20760 + struct _vx_monitor *mon = &per_cpu(vx_monitor_buffer, cpu);
20761 + int end = mon->counter;
20762 + int start = end - VXM_SIZE + 2;
20763 + int idx = *index;
20765 + /* special case: get current pos */
20771 + /* have we lost some data? */
20775 + for (pos = 0; (pos < *count) && (idx < end); pos++, idx++) {
20776 + struct _vx_mon_entry *entry =
20777 + &mon->entry[idx % VXM_SIZE];
20779 + /* send entry to userspace */
20780 + ret = copy_to_user(&data[pos], entry, sizeof(*entry));
20784 + /* save new index and count */
20787 + return ret ? ret : (*index < end);
20790 +int vc_read_monitor(uint32_t id, void __user *data)
20792 + struct vcmd_read_monitor_v0 vc_data;
20795 + if (id >= NR_CPUS)
20798 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20801 + ret = do_read_monitor((struct __user _vx_mon_entry *)vc_data.data,
20802 + id, &vc_data.index, &vc_data.count);
20804 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
20809 +#ifdef CONFIG_COMPAT
20811 +int vc_read_monitor_x32(uint32_t id, void __user *data)
20813 + struct vcmd_read_monitor_v0_x32 vc_data;
20816 + if (id >= NR_CPUS)
20819 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20822 + ret = do_read_monitor((struct __user _vx_mon_entry *)
20823 + compat_ptr(vc_data.data_ptr),
20824 + id, &vc_data.index, &vc_data.count);
20826 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
20831 +#endif /* CONFIG_COMPAT */
20833 diff -NurpP --minimal linux-2.6.36/kernel/vserver/network.c linux-2.6.36-vs2.3.0.36.33/kernel/vserver/network.c
20834 --- linux-2.6.36/kernel/vserver/network.c 1970-01-01 01:00:00.000000000 +0100
20835 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/network.c 2010-10-21 13:09:36.000000000 +0200
20838 + * linux/kernel/vserver/network.c
20840 + * Virtual Server: Network Support
20842 + * Copyright (C) 2003-2007 Herbert Pötzl
20844 + * V0.01 broken out from vcontext V0.05
20845 + * V0.02 cleaned up implementation
20846 + * V0.03 added equiv nx commands
20847 + * V0.04 switch to RCU based hash
20848 + * V0.05 and back to locking again
20849 + * V0.06 changed vcmds to nxi arg
20850 + * V0.07 have __create claim() the nxi
20854 +#include <linux/err.h>
20855 +#include <linux/slab.h>
20856 +#include <linux/rcupdate.h>
20858 +#include <linux/vs_network.h>
20859 +#include <linux/vs_pid.h>
20860 +#include <linux/vserver/network_cmd.h>
20863 +atomic_t nx_global_ctotal = ATOMIC_INIT(0);
20864 +atomic_t nx_global_cactive = ATOMIC_INIT(0);
20866 +static struct kmem_cache *nx_addr_v4_cachep = NULL;
20867 +static struct kmem_cache *nx_addr_v6_cachep = NULL;
20870 +static int __init init_network(void)
20872 + nx_addr_v4_cachep = kmem_cache_create("nx_v4_addr_cache",
20873 + sizeof(struct nx_addr_v4), 0,
20874 + SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
20875 + nx_addr_v6_cachep = kmem_cache_create("nx_v6_addr_cache",
20876 + sizeof(struct nx_addr_v6), 0,
20877 + SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
20882 +/* __alloc_nx_addr_v4() */
20884 +static inline struct nx_addr_v4 *__alloc_nx_addr_v4(void)
20886 + struct nx_addr_v4 *nxa = kmem_cache_alloc(
20887 + nx_addr_v4_cachep, GFP_KERNEL);
20889 + if (!IS_ERR(nxa))
20890 + memset(nxa, 0, sizeof(*nxa));
20894 +/* __dealloc_nx_addr_v4() */
20896 +static inline void __dealloc_nx_addr_v4(struct nx_addr_v4 *nxa)
20898 + kmem_cache_free(nx_addr_v4_cachep, nxa);
20901 +/* __dealloc_nx_addr_v4_all() */
20903 +static inline void __dealloc_nx_addr_v4_all(struct nx_addr_v4 *nxa)
20906 + struct nx_addr_v4 *next = nxa->next;
20908 + __dealloc_nx_addr_v4(nxa);
20914 +#ifdef CONFIG_IPV6
20916 +/* __alloc_nx_addr_v6() */
20918 +static inline struct nx_addr_v6 *__alloc_nx_addr_v6(void)
20920 + struct nx_addr_v6 *nxa = kmem_cache_alloc(
20921 + nx_addr_v6_cachep, GFP_KERNEL);
20923 + if (!IS_ERR(nxa))
20924 + memset(nxa, 0, sizeof(*nxa));
20928 +/* __dealloc_nx_addr_v6() */
20930 +static inline void __dealloc_nx_addr_v6(struct nx_addr_v6 *nxa)
20932 + kmem_cache_free(nx_addr_v6_cachep, nxa);
20935 +/* __dealloc_nx_addr_v6_all() */
20937 +static inline void __dealloc_nx_addr_v6_all(struct nx_addr_v6 *nxa)
20940 + struct nx_addr_v6 *next = nxa->next;
20942 + __dealloc_nx_addr_v6(nxa);
20947 +#endif /* CONFIG_IPV6 */
20949 +/* __alloc_nx_info()
20951 + * allocate an initialized nx_info struct
20952 + * doesn't make it visible (hash) */
20954 +static struct nx_info *__alloc_nx_info(nid_t nid)
20956 + struct nx_info *new = NULL;
20958 + vxdprintk(VXD_CBIT(nid, 1), "alloc_nx_info(%d)*", nid);
20960 + /* would this benefit from a slab cache? */
20961 + new = kmalloc(sizeof(struct nx_info), GFP_KERNEL);
20965 + memset(new, 0, sizeof(struct nx_info));
20966 + new->nx_id = nid;
20967 + INIT_HLIST_NODE(&new->nx_hlist);
20968 + atomic_set(&new->nx_usecnt, 0);
20969 + atomic_set(&new->nx_tasks, 0);
20970 + new->nx_state = 0;
20972 + new->nx_flags = NXF_INIT_SET;
20974 + /* rest of init goes here */
20976 + new->v4_lback.s_addr = htonl(INADDR_LOOPBACK);
20977 + new->v4_bcast.s_addr = htonl(INADDR_BROADCAST);
20979 + vxdprintk(VXD_CBIT(nid, 0),
20980 + "alloc_nx_info(%d) = %p", nid, new);
20981 + atomic_inc(&nx_global_ctotal);
20985 +/* __dealloc_nx_info()
20987 + * final disposal of nx_info */
20989 +static void __dealloc_nx_info(struct nx_info *nxi)
20991 + vxdprintk(VXD_CBIT(nid, 0),
20992 + "dealloc_nx_info(%p)", nxi);
20994 + nxi->nx_hlist.next = LIST_POISON1;
20997 + BUG_ON(atomic_read(&nxi->nx_usecnt));
20998 + BUG_ON(atomic_read(&nxi->nx_tasks));
21000 + __dealloc_nx_addr_v4_all(nxi->v4.next);
21002 + nxi->nx_state |= NXS_RELEASED;
21004 + atomic_dec(&nx_global_ctotal);
21007 +static void __shutdown_nx_info(struct nx_info *nxi)
21009 + nxi->nx_state |= NXS_SHUTDOWN;
21010 + vs_net_change(nxi, VSC_NETDOWN);
21013 +/* exported stuff */
21015 +void free_nx_info(struct nx_info *nxi)
21017 + /* context shutdown is mandatory */
21018 + BUG_ON(nxi->nx_state != NXS_SHUTDOWN);
21020 + /* context must not be hashed */
21021 + BUG_ON(nxi->nx_state & NXS_HASHED);
21023 + BUG_ON(atomic_read(&nxi->nx_usecnt));
21024 + BUG_ON(atomic_read(&nxi->nx_tasks));
21026 + __dealloc_nx_info(nxi);
21030 +void __nx_set_lback(struct nx_info *nxi)
21032 + int nid = nxi->nx_id;
21033 + __be32 lback = htonl(INADDR_LOOPBACK ^ ((nid & 0xFFFF) << 8));
21035 + nxi->v4_lback.s_addr = lback;
21038 +extern int __nx_inet_add_lback(__be32 addr);
21039 +extern int __nx_inet_del_lback(__be32 addr);
21042 +/* hash table for nx_info hash */
21044 +#define NX_HASH_SIZE 13
21046 +struct hlist_head nx_info_hash[NX_HASH_SIZE];
21048 +static spinlock_t nx_info_hash_lock = SPIN_LOCK_UNLOCKED;
21051 +static inline unsigned int __hashval(nid_t nid)
21053 + return (nid % NX_HASH_SIZE);
21058 +/* __hash_nx_info()
21060 + * add the nxi to the global hash table
21061 + * requires the hash_lock to be held */
21063 +static inline void __hash_nx_info(struct nx_info *nxi)
21065 + struct hlist_head *head;
21067 + vxd_assert_lock(&nx_info_hash_lock);
21068 + vxdprintk(VXD_CBIT(nid, 4),
21069 + "__hash_nx_info: %p[#%d]", nxi, nxi->nx_id);
21071 + /* context must not be hashed */
21072 + BUG_ON(nx_info_state(nxi, NXS_HASHED));
21074 + nxi->nx_state |= NXS_HASHED;
21075 + head = &nx_info_hash[__hashval(nxi->nx_id)];
21076 + hlist_add_head(&nxi->nx_hlist, head);
21077 + atomic_inc(&nx_global_cactive);
21080 +/* __unhash_nx_info()
21082 + * remove the nxi from the global hash table
21083 + * requires the hash_lock to be held */
21085 +static inline void __unhash_nx_info(struct nx_info *nxi)
21087 + vxd_assert_lock(&nx_info_hash_lock);
21088 + vxdprintk(VXD_CBIT(nid, 4),
21089 + "__unhash_nx_info: %p[#%d.%d.%d]", nxi, nxi->nx_id,
21090 + atomic_read(&nxi->nx_usecnt), atomic_read(&nxi->nx_tasks));
21092 + /* context must be hashed */
21093 + BUG_ON(!nx_info_state(nxi, NXS_HASHED));
21094 + /* but without tasks */
21095 + BUG_ON(atomic_read(&nxi->nx_tasks));
21097 + nxi->nx_state &= ~NXS_HASHED;
21098 + hlist_del(&nxi->nx_hlist);
21099 + atomic_dec(&nx_global_cactive);
21103 +/* __lookup_nx_info()
21105 + * requires the hash_lock to be held
21106 + * doesn't increment the nx_refcnt */
21108 +static inline struct nx_info *__lookup_nx_info(nid_t nid)
21110 + struct hlist_head *head = &nx_info_hash[__hashval(nid)];
21111 + struct hlist_node *pos;
21112 + struct nx_info *nxi;
21114 + vxd_assert_lock(&nx_info_hash_lock);
21115 + hlist_for_each(pos, head) {
21116 + nxi = hlist_entry(pos, struct nx_info, nx_hlist);
21118 + if (nxi->nx_id == nid)
21123 + vxdprintk(VXD_CBIT(nid, 0),
21124 + "__lookup_nx_info(#%u): %p[#%u]",
21125 + nid, nxi, nxi ? nxi->nx_id : 0);
21130 +/* __create_nx_info()
21132 + * create the requested context
21133 + * get(), claim() and hash it */
21135 +static struct nx_info *__create_nx_info(int id)
21137 + struct nx_info *new, *nxi = NULL;
21139 + vxdprintk(VXD_CBIT(nid, 1), "create_nx_info(%d)*", id);
21141 + if (!(new = __alloc_nx_info(id)))
21142 + return ERR_PTR(-ENOMEM);
21144 + /* required to make dynamic xids unique */
21145 + spin_lock(&nx_info_hash_lock);
21147 + /* static context requested */
21148 + if ((nxi = __lookup_nx_info(id))) {
21149 + vxdprintk(VXD_CBIT(nid, 0),
21150 + "create_nx_info(%d) = %p (already there)", id, nxi);
21151 + if (nx_info_flags(nxi, NXF_STATE_SETUP, 0))
21152 + nxi = ERR_PTR(-EBUSY);
21154 + nxi = ERR_PTR(-EEXIST);
21157 + /* new context */
21158 + vxdprintk(VXD_CBIT(nid, 0),
21159 + "create_nx_info(%d) = %p (new)", id, new);
21160 + claim_nx_info(new, NULL);
21161 + __nx_set_lback(new);
21162 + __hash_nx_info(get_nx_info(new));
21163 + nxi = new, new = NULL;
21166 + spin_unlock(&nx_info_hash_lock);
21168 + __dealloc_nx_info(new);
21174 +/* exported stuff */
21177 +void unhash_nx_info(struct nx_info *nxi)
21179 + __shutdown_nx_info(nxi);
21180 + spin_lock(&nx_info_hash_lock);
21181 + __unhash_nx_info(nxi);
21182 + spin_unlock(&nx_info_hash_lock);
21185 +/* lookup_nx_info()
21187 + * search for a nx_info and get() it
21188 + * negative id means current */
21190 +struct nx_info *lookup_nx_info(int id)
21192 + struct nx_info *nxi = NULL;
21195 + nxi = get_nx_info(current_nx_info());
21196 + } else if (id > 1) {
21197 + spin_lock(&nx_info_hash_lock);
21198 + nxi = get_nx_info(__lookup_nx_info(id));
21199 + spin_unlock(&nx_info_hash_lock);
21204 +/* nid_is_hashed()
21206 + * verify that nid is still hashed */
21208 +int nid_is_hashed(nid_t nid)
21212 + spin_lock(&nx_info_hash_lock);
21213 + hashed = (__lookup_nx_info(nid) != NULL);
21214 + spin_unlock(&nx_info_hash_lock);
21219 +#ifdef CONFIG_PROC_FS
21223 + * get a subset of hashed nids for proc
21224 + * assumes size is at least one */
21226 +int get_nid_list(int index, unsigned int *nids, int size)
21228 + int hindex, nr_nids = 0;
21230 + /* only show current and children */
21231 + if (!nx_check(0, VS_ADMIN | VS_WATCH)) {
21234 + nids[nr_nids] = nx_current_nid();
21238 + for (hindex = 0; hindex < NX_HASH_SIZE; hindex++) {
21239 + struct hlist_head *head = &nx_info_hash[hindex];
21240 + struct hlist_node *pos;
21242 + spin_lock(&nx_info_hash_lock);
21243 + hlist_for_each(pos, head) {
21244 + struct nx_info *nxi;
21249 + nxi = hlist_entry(pos, struct nx_info, nx_hlist);
21250 + nids[nr_nids] = nxi->nx_id;
21251 + if (++nr_nids >= size) {
21252 + spin_unlock(&nx_info_hash_lock);
21256 + /* keep the lock time short */
21257 + spin_unlock(&nx_info_hash_lock);
21266 + * migrate task to new network
21267 + * gets nxi, puts old_nxi on change
21270 +int nx_migrate_task(struct task_struct *p, struct nx_info *nxi)
21272 + struct nx_info *old_nxi;
21278 + vxdprintk(VXD_CBIT(nid, 5),
21279 + "nx_migrate_task(%p,%p[#%d.%d.%d])",
21280 + p, nxi, nxi->nx_id,
21281 + atomic_read(&nxi->nx_usecnt),
21282 + atomic_read(&nxi->nx_tasks));
21284 + if (nx_info_flags(nxi, NXF_INFO_PRIVATE, 0) &&
21285 + !nx_info_flags(nxi, NXF_STATE_SETUP, 0))
21288 + if (nx_info_state(nxi, NXS_SHUTDOWN))
21291 + /* maybe disallow this completely? */
21292 + old_nxi = task_get_nx_info(p);
21293 + if (old_nxi == nxi)
21298 + clr_nx_info(&p->nx_info);
21299 + claim_nx_info(nxi, p);
21300 + set_nx_info(&p->nx_info, nxi);
21301 + p->nid = nxi->nx_id;
21304 + vxdprintk(VXD_CBIT(nid, 5),
21305 + "moved task %p into nxi:%p[#%d]",
21306 + p, nxi, nxi->nx_id);
21309 + release_nx_info(old_nxi, p);
21312 + put_nx_info(old_nxi);
21317 +void nx_set_persistent(struct nx_info *nxi)
21319 + vxdprintk(VXD_CBIT(nid, 6),
21320 + "nx_set_persistent(%p[#%d])", nxi, nxi->nx_id);
21322 + get_nx_info(nxi);
21323 + claim_nx_info(nxi, NULL);
21326 +void nx_clear_persistent(struct nx_info *nxi)
21328 + vxdprintk(VXD_CBIT(nid, 6),
21329 + "nx_clear_persistent(%p[#%d])", nxi, nxi->nx_id);
21331 + release_nx_info(nxi, NULL);
21332 + put_nx_info(nxi);
21335 +void nx_update_persistent(struct nx_info *nxi)
21337 + if (nx_info_flags(nxi, NXF_PERSISTENT, 0))
21338 + nx_set_persistent(nxi);
21340 + nx_clear_persistent(nxi);
21343 +/* vserver syscall commands below here */
21345 +/* taks nid and nx_info functions */
21347 +#include <asm/uaccess.h>
21350 +int vc_task_nid(uint32_t id)
21355 + struct task_struct *tsk;
21357 + read_lock(&tasklist_lock);
21358 + tsk = find_task_by_real_pid(id);
21359 + nid = (tsk) ? tsk->nid : -ESRCH;
21360 + read_unlock(&tasklist_lock);
21362 + nid = nx_current_nid();
21367 +int vc_nx_info(struct nx_info *nxi, void __user *data)
21369 + struct vcmd_nx_info_v0 vc_data;
21371 + vc_data.nid = nxi->nx_id;
21373 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
21379 +/* network functions */
21381 +int vc_net_create(uint32_t nid, void __user *data)
21383 + struct vcmd_net_create vc_data = { .flagword = NXF_INIT_SET };
21384 + struct nx_info *new_nxi;
21387 + if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
21390 + if ((nid > MAX_S_CONTEXT) || (nid < 2))
21393 + new_nxi = __create_nx_info(nid);
21394 + if (IS_ERR(new_nxi))
21395 + return PTR_ERR(new_nxi);
21397 + /* initial flags */
21398 + new_nxi->nx_flags = vc_data.flagword;
21401 + if (vs_net_change(new_nxi, VSC_NETUP))
21404 + ret = nx_migrate_task(current, new_nxi);
21408 + /* return context id on success */
21409 + ret = new_nxi->nx_id;
21411 + /* get a reference for persistent contexts */
21412 + if ((vc_data.flagword & NXF_PERSISTENT))
21413 + nx_set_persistent(new_nxi);
21415 + release_nx_info(new_nxi, NULL);
21416 + put_nx_info(new_nxi);
21421 +int vc_net_migrate(struct nx_info *nxi, void __user *data)
21423 + return nx_migrate_task(current, nxi);
21428 +int do_add_v4_addr(struct nx_info *nxi, __be32 ip, __be32 ip2, __be32 mask,
21429 + uint16_t type, uint16_t flags)
21431 + struct nx_addr_v4 *nxa = &nxi->v4;
21433 + if (NX_IPV4(nxi)) {
21434 + /* locate last entry */
21435 + for (; nxa->next; nxa = nxa->next);
21436 + nxa->next = __alloc_nx_addr_v4();
21440 + return PTR_ERR(nxa);
21443 + if (nxi->v4.next)
21444 + /* remove single ip for ip list */
21445 + nxi->nx_flags &= ~NXF_SINGLE_IP;
21447 + nxa->ip[0].s_addr = ip;
21448 + nxa->ip[1].s_addr = ip2;
21449 + nxa->mask.s_addr = mask;
21450 + nxa->type = type;
21451 + nxa->flags = flags;
21456 +int vc_net_add(struct nx_info *nxi, void __user *data)
21458 + struct vcmd_net_addr_v0 vc_data;
21459 + int index, ret = 0;
21461 + if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
21464 + switch (vc_data.type) {
21465 + case NXA_TYPE_IPV4:
21466 + if ((vc_data.count < 1) || (vc_data.count > 4))
21470 + while (index < vc_data.count) {
21471 + ret = do_add_v4_addr(nxi, vc_data.ip[index].s_addr, 0,
21472 + vc_data.mask[index].s_addr, NXA_TYPE_ADDR, 0);
21480 + case NXA_TYPE_IPV4|NXA_MOD_BCAST:
21481 + nxi->v4_bcast = vc_data.ip[0];
21485 + case NXA_TYPE_IPV4|NXA_MOD_LBACK:
21486 + nxi->v4_lback = vc_data.ip[0];
21497 +int vc_net_remove(struct nx_info *nxi, void __user *data)
21499 + struct vcmd_net_addr_v0 vc_data;
21501 + if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
21504 + switch (vc_data.type) {
21505 + case NXA_TYPE_ANY:
21506 + __dealloc_nx_addr_v4_all(xchg(&nxi->v4.next, NULL));
21507 + memset(&nxi->v4, 0, sizeof(nxi->v4));
21517 +int vc_net_add_ipv4(struct nx_info *nxi, void __user *data)
21519 + struct vcmd_net_addr_ipv4_v1 vc_data;
21521 + if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
21524 + switch (vc_data.type) {
21525 + case NXA_TYPE_ADDR:
21526 + case NXA_TYPE_RANGE:
21527 + case NXA_TYPE_MASK:
21528 + return do_add_v4_addr(nxi, vc_data.ip.s_addr, 0,
21529 + vc_data.mask.s_addr, vc_data.type, vc_data.flags);
21531 + case NXA_TYPE_ADDR | NXA_MOD_BCAST:
21532 + nxi->v4_bcast = vc_data.ip;
21535 + case NXA_TYPE_ADDR | NXA_MOD_LBACK:
21536 + nxi->v4_lback = vc_data.ip;
21545 +int vc_net_remove_ipv4(struct nx_info *nxi, void __user *data)
21547 + struct vcmd_net_addr_ipv4_v1 vc_data;
21549 + if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
21552 + switch (vc_data.type) {
21553 +/* case NXA_TYPE_ADDR:
21556 + case NXA_TYPE_ANY:
21557 + __dealloc_nx_addr_v4_all(xchg(&nxi->v4.next, NULL));
21558 + memset(&nxi->v4, 0, sizeof(nxi->v4));
21568 +#ifdef CONFIG_IPV6
21570 +int do_add_v6_addr(struct nx_info *nxi,
21571 + struct in6_addr *ip, struct in6_addr *mask,
21572 + uint32_t prefix, uint16_t type, uint16_t flags)
21574 + struct nx_addr_v6 *nxa = &nxi->v6;
21576 + if (NX_IPV6(nxi)) {
21577 + /* locate last entry */
21578 + for (; nxa->next; nxa = nxa->next);
21579 + nxa->next = __alloc_nx_addr_v6();
21583 + return PTR_ERR(nxa);
21587 + nxa->mask = *mask;
21588 + nxa->prefix = prefix;
21589 + nxa->type = type;
21590 + nxa->flags = flags;
21595 +int vc_net_add_ipv6(struct nx_info *nxi, void __user *data)
21597 + struct vcmd_net_addr_ipv6_v1 vc_data;
21599 + if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
21602 + switch (vc_data.type) {
21603 + case NXA_TYPE_ADDR:
21604 + case NXA_TYPE_MASK:
21605 + return do_add_v6_addr(nxi, &vc_data.ip, &vc_data.mask,
21606 + vc_data.prefix, vc_data.type, vc_data.flags);
21613 +int vc_net_remove_ipv6(struct nx_info *nxi, void __user *data)
21615 + struct vcmd_net_addr_ipv6_v1 vc_data;
21617 + if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
21620 + switch (vc_data.type) {
21621 + case NXA_TYPE_ANY:
21622 + __dealloc_nx_addr_v6_all(xchg(&nxi->v6.next, NULL));
21623 + memset(&nxi->v6, 0, sizeof(nxi->v6));
21632 +#endif /* CONFIG_IPV6 */
21635 +int vc_get_nflags(struct nx_info *nxi, void __user *data)
21637 + struct vcmd_net_flags_v0 vc_data;
21639 + vc_data.flagword = nxi->nx_flags;
21641 + /* special STATE flag handling */
21642 + vc_data.mask = vs_mask_flags(~0ULL, nxi->nx_flags, NXF_ONE_TIME);
21644 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
21649 +int vc_set_nflags(struct nx_info *nxi, void __user *data)
21651 + struct vcmd_net_flags_v0 vc_data;
21652 + uint64_t mask, trigger;
21654 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
21657 + /* special STATE flag handling */
21658 + mask = vs_mask_mask(vc_data.mask, nxi->nx_flags, NXF_ONE_TIME);
21659 + trigger = (mask & nxi->nx_flags) ^ (mask & vc_data.flagword);
21661 + nxi->nx_flags = vs_mask_flags(nxi->nx_flags,
21662 + vc_data.flagword, mask);
21663 + if (trigger & NXF_PERSISTENT)
21664 + nx_update_persistent(nxi);
21669 +int vc_get_ncaps(struct nx_info *nxi, void __user *data)
21671 + struct vcmd_net_caps_v0 vc_data;
21673 + vc_data.ncaps = nxi->nx_ncaps;
21674 + vc_data.cmask = ~0ULL;
21676 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
21681 +int vc_set_ncaps(struct nx_info *nxi, void __user *data)
21683 + struct vcmd_net_caps_v0 vc_data;
21685 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
21688 + nxi->nx_ncaps = vs_mask_flags(nxi->nx_ncaps,
21689 + vc_data.ncaps, vc_data.cmask);
21694 +#include <linux/module.h>
21696 +module_init(init_network);
21698 +EXPORT_SYMBOL_GPL(free_nx_info);
21699 +EXPORT_SYMBOL_GPL(unhash_nx_info);
21701 diff -NurpP --minimal linux-2.6.36/kernel/vserver/proc.c linux-2.6.36-vs2.3.0.36.33/kernel/vserver/proc.c
21702 --- linux-2.6.36/kernel/vserver/proc.c 1970-01-01 01:00:00.000000000 +0100
21703 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/proc.c 2010-10-21 13:09:36.000000000 +0200
21706 + * linux/kernel/vserver/proc.c
21708 + * Virtual Context Support
21710 + * Copyright (C) 2003-2007 Herbert Pötzl
21712 + * V0.01 basic structure
21713 + * V0.02 adaptation vs1.3.0
21714 + * V0.03 proc permissions
21715 + * V0.04 locking/generic
21716 + * V0.05 next generation procfs
21717 + * V0.06 inode validation
21718 + * V0.07 generic rewrite vid
21719 + * V0.08 remove inode type
21723 +#include <linux/proc_fs.h>
21724 +#include <linux/fs_struct.h>
21725 +#include <linux/mount.h>
21726 +#include <asm/unistd.h>
21728 +#include <linux/vs_context.h>
21729 +#include <linux/vs_network.h>
21730 +#include <linux/vs_cvirt.h>
21732 +#include <linux/in.h>
21733 +#include <linux/inetdevice.h>
21734 +#include <linux/vs_inet.h>
21735 +#include <linux/vs_inet6.h>
21737 +#include <linux/vserver/global.h>
21739 +#include "cvirt_proc.h"
21740 +#include "cacct_proc.h"
21741 +#include "limit_proc.h"
21742 +#include "sched_proc.h"
21743 +#include "vci_config.h"
21746 +static inline char *print_cap_t(char *buffer, kernel_cap_t *c)
21750 + CAP_FOR_EACH_U32(__capi) {
21751 + buffer += sprintf(buffer, "%08x",
21752 + c->cap[(_KERNEL_CAPABILITY_U32S-1) - __capi]);
21758 +static struct proc_dir_entry *proc_virtual;
21760 +static struct proc_dir_entry *proc_virtnet;
21763 +/* first the actual feeds */
21766 +static int proc_vci(char *buffer)
21768 + return sprintf(buffer,
21769 + "VCIVersion:\t%04x:%04x\n"
21770 + "VCISyscall:\t%d\n"
21771 + "VCIKernel:\t%08x\n",
21772 + VCI_VERSION >> 16,
21773 + VCI_VERSION & 0xFFFF,
21775 + vci_kernel_config());
21778 +static int proc_virtual_info(char *buffer)
21780 + return proc_vci(buffer);
21783 +static int proc_virtual_status(char *buffer)
21785 + return sprintf(buffer,
21787 + "#CActive:\t%d\n"
21788 + "#NSProxy:\t%d\t%d %d %d %d %d %d\n"
21789 + "#InitTask:\t%d\t%d %d\n",
21790 + atomic_read(&vx_global_ctotal),
21791 + atomic_read(&vx_global_cactive),
21792 + atomic_read(&vs_global_nsproxy),
21793 + atomic_read(&vs_global_fs),
21794 + atomic_read(&vs_global_mnt_ns),
21795 + atomic_read(&vs_global_uts_ns),
21796 + atomic_read(&nr_ipc_ns),
21797 + atomic_read(&vs_global_user_ns),
21798 + atomic_read(&vs_global_pid_ns),
21799 + atomic_read(&init_task.usage),
21800 + atomic_read(&init_task.nsproxy->count),
21801 + init_task.fs->users);
21805 +int proc_vxi_info(struct vx_info *vxi, char *buffer)
21809 + length = sprintf(buffer,
21817 + vxi->vx_badness_bias);
21821 +int proc_vxi_status(struct vx_info *vxi, char *buffer)
21823 + char *orig = buffer;
21825 + buffer += sprintf(buffer,
21828 + "Flags:\t%016llx\n",
21829 + atomic_read(&vxi->vx_usecnt),
21830 + atomic_read(&vxi->vx_tasks),
21831 + (unsigned long long)vxi->vx_flags);
21833 + buffer += sprintf(buffer, "BCaps:\t");
21834 + buffer = print_cap_t(buffer, &vxi->vx_bcaps);
21835 + buffer += sprintf(buffer, "\n");
21837 + buffer += sprintf(buffer,
21838 + "CCaps:\t%016llx\n"
21839 + "Spaces:\t%08lx %08lx\n",
21840 + (unsigned long long)vxi->vx_ccaps,
21841 + vxi->vx_nsmask[0], vxi->vx_nsmask[1]);
21842 + return buffer - orig;
21845 +int proc_vxi_limit(struct vx_info *vxi, char *buffer)
21847 + return vx_info_proc_limit(&vxi->limit, buffer);
21850 +int proc_vxi_sched(struct vx_info *vxi, char *buffer)
21854 + length = vx_info_proc_sched(&vxi->sched, buffer);
21855 + for_each_online_cpu(cpu) {
21856 + length += vx_info_proc_sched_pc(
21857 + &vx_per_cpu(vxi, sched_pc, cpu),
21858 + buffer + length, cpu);
21863 +int proc_vxi_nsproxy0(struct vx_info *vxi, char *buffer)
21865 + return vx_info_proc_nsproxy(vxi->vx_nsproxy[0], buffer);
21868 +int proc_vxi_nsproxy1(struct vx_info *vxi, char *buffer)
21870 + return vx_info_proc_nsproxy(vxi->vx_nsproxy[1], buffer);
21873 +int proc_vxi_cvirt(struct vx_info *vxi, char *buffer)
21877 + vx_update_load(vxi);
21878 + length = vx_info_proc_cvirt(&vxi->cvirt, buffer);
21879 + for_each_online_cpu(cpu) {
21880 + length += vx_info_proc_cvirt_pc(
21881 + &vx_per_cpu(vxi, cvirt_pc, cpu),
21882 + buffer + length, cpu);
21887 +int proc_vxi_cacct(struct vx_info *vxi, char *buffer)
21889 + return vx_info_proc_cacct(&vxi->cacct, buffer);
21893 +static int proc_virtnet_info(char *buffer)
21895 + return proc_vci(buffer);
21898 +static int proc_virtnet_status(char *buffer)
21900 + return sprintf(buffer,
21902 + "#CActive:\t%d\n",
21903 + atomic_read(&nx_global_ctotal),
21904 + atomic_read(&nx_global_cactive));
21907 +int proc_nxi_info(struct nx_info *nxi, char *buffer)
21909 + struct nx_addr_v4 *v4a;
21910 +#ifdef CONFIG_IPV6
21911 + struct nx_addr_v6 *v6a;
21915 + length = sprintf(buffer,
21918 + "Bcast:\t" NIPQUAD_FMT "\n"
21919 + "Lback:\t" NIPQUAD_FMT "\n",
21922 + NIPQUAD(nxi->v4_bcast.s_addr),
21923 + NIPQUAD(nxi->v4_lback.s_addr));
21925 + if (!NX_IPV4(nxi))
21927 + for (i = 0, v4a = &nxi->v4; v4a; i++, v4a = v4a->next)
21928 + length += sprintf(buffer + length, "%d:\t" NXAV4_FMT "\n",
21931 +#ifdef CONFIG_IPV6
21932 + if (!NX_IPV6(nxi))
21934 + for (i = 0, v6a = &nxi->v6; v6a; i++, v6a = v6a->next)
21935 + length += sprintf(buffer + length, "%d:\t" NXAV6_FMT "\n",
21942 +int proc_nxi_status(struct nx_info *nxi, char *buffer)
21946 + length = sprintf(buffer,
21949 + "Flags:\t%016llx\n"
21950 + "NCaps:\t%016llx\n",
21951 + atomic_read(&nxi->nx_usecnt),
21952 + atomic_read(&nxi->nx_tasks),
21953 + (unsigned long long)nxi->nx_flags,
21954 + (unsigned long long)nxi->nx_ncaps);
21960 +/* here the inode helpers */
21966 + struct inode_operations *iop;
21967 + struct file_operations *fop;
21968 + union proc_op op;
21971 +static struct inode *vs_proc_make_inode(struct super_block *sb, struct vs_entry *p)
21973 + struct inode *inode = new_inode(sb);
21978 + inode->i_mode = p->mode;
21980 + inode->i_op = p->iop;
21982 + inode->i_fop = p->fop;
21984 + inode->i_nlink = (p->mode & S_IFDIR) ? 2 : 1;
21985 + inode->i_flags |= S_IMMUTABLE;
21987 + inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
21989 + inode->i_uid = 0;
21990 + inode->i_gid = 0;
21991 + inode->i_tag = 0;
21996 +static struct dentry *vs_proc_instantiate(struct inode *dir,
21997 + struct dentry *dentry, int id, void *ptr)
21999 + struct vs_entry *p = ptr;
22000 + struct inode *inode = vs_proc_make_inode(dir->i_sb, p);
22001 + struct dentry *error = ERR_PTR(-EINVAL);
22006 + PROC_I(inode)->op = p->op;
22007 + PROC_I(inode)->fd = id;
22008 + d_add(dentry, inode);
22016 +typedef struct dentry *instantiate_t(struct inode *, struct dentry *, int, void *);
22019 + * Fill a directory entry.
22021 + * If possible create the dcache entry and derive our inode number and
22022 + * file type from dcache entry.
22024 + * Since all of the proc inode numbers are dynamically generated, the inode
22025 + * numbers do not exist until the inode is cache. This means creating the
22026 + * the dcache entry in readdir is necessary to keep the inode numbers
22027 + * reported by readdir in sync with the inode numbers reported
22030 +static int proc_fill_cache(struct file *filp, void *dirent, filldir_t filldir,
22031 + char *name, int len, instantiate_t instantiate, int id, void *ptr)
22033 + struct dentry *child, *dir = filp->f_dentry;
22034 + struct inode *inode;
22035 + struct qstr qname;
22037 + unsigned type = DT_UNKNOWN;
22039 + qname.name = name;
22041 + qname.hash = full_name_hash(name, len);
22043 + child = d_lookup(dir, &qname);
22045 + struct dentry *new;
22046 + new = d_alloc(dir, &qname);
22048 + child = instantiate(dir->d_inode, new, id, ptr);
22055 + if (!child || IS_ERR(child) || !child->d_inode)
22056 + goto end_instantiate;
22057 + inode = child->d_inode;
22059 + ino = inode->i_ino;
22060 + type = inode->i_mode >> 12;
22065 + ino = find_inode_number(dir, &qname);
22068 + return filldir(dirent, name, len, filp->f_pos, ino, type);
22073 +/* get and revalidate vx_info/xid */
22076 +struct vx_info *get_proc_vx_info(struct inode *inode)
22078 + return lookup_vx_info(PROC_I(inode)->fd);
22081 +static int proc_xid_revalidate(struct dentry *dentry, struct nameidata *nd)
22083 + struct inode *inode = dentry->d_inode;
22084 + xid_t xid = PROC_I(inode)->fd;
22086 + if (!xid || xid_is_hashed(xid))
22093 +/* get and revalidate nx_info/nid */
22095 +static int proc_nid_revalidate(struct dentry *dentry, struct nameidata *nd)
22097 + struct inode *inode = dentry->d_inode;
22098 + nid_t nid = PROC_I(inode)->fd;
22100 + if (!nid || nid_is_hashed(nid))
22108 +#define PROC_BLOCK_SIZE (PAGE_SIZE - 1024)
22110 +static ssize_t proc_vs_info_read(struct file *file, char __user *buf,
22111 + size_t count, loff_t *ppos)
22113 + struct inode *inode = file->f_dentry->d_inode;
22114 + unsigned long page;
22115 + ssize_t length = 0;
22117 + if (count > PROC_BLOCK_SIZE)
22118 + count = PROC_BLOCK_SIZE;
22120 + /* fade that out as soon as stable */
22121 + WARN_ON(PROC_I(inode)->fd);
22123 + if (!(page = __get_free_page(GFP_KERNEL)))
22126 + BUG_ON(!PROC_I(inode)->op.proc_vs_read);
22127 + length = PROC_I(inode)->op.proc_vs_read((char *)page);
22130 + length = simple_read_from_buffer(buf, count, ppos,
22131 + (char *)page, length);
22137 +static ssize_t proc_vx_info_read(struct file *file, char __user *buf,
22138 + size_t count, loff_t *ppos)
22140 + struct inode *inode = file->f_dentry->d_inode;
22141 + struct vx_info *vxi = NULL;
22142 + xid_t xid = PROC_I(inode)->fd;
22143 + unsigned long page;
22144 + ssize_t length = 0;
22146 + if (count > PROC_BLOCK_SIZE)
22147 + count = PROC_BLOCK_SIZE;
22149 + /* fade that out as soon as stable */
22151 + vxi = lookup_vx_info(xid);
22155 + length = -ENOMEM;
22156 + if (!(page = __get_free_page(GFP_KERNEL)))
22159 + BUG_ON(!PROC_I(inode)->op.proc_vxi_read);
22160 + length = PROC_I(inode)->op.proc_vxi_read(vxi, (char *)page);
22163 + length = simple_read_from_buffer(buf, count, ppos,
22164 + (char *)page, length);
22168 + put_vx_info(vxi);
22173 +static ssize_t proc_nx_info_read(struct file *file, char __user *buf,
22174 + size_t count, loff_t *ppos)
22176 + struct inode *inode = file->f_dentry->d_inode;
22177 + struct nx_info *nxi = NULL;
22178 + nid_t nid = PROC_I(inode)->fd;
22179 + unsigned long page;
22180 + ssize_t length = 0;
22182 + if (count > PROC_BLOCK_SIZE)
22183 + count = PROC_BLOCK_SIZE;
22185 + /* fade that out as soon as stable */
22187 + nxi = lookup_nx_info(nid);
22191 + length = -ENOMEM;
22192 + if (!(page = __get_free_page(GFP_KERNEL)))
22195 + BUG_ON(!PROC_I(inode)->op.proc_nxi_read);
22196 + length = PROC_I(inode)->op.proc_nxi_read(nxi, (char *)page);
22199 + length = simple_read_from_buffer(buf, count, ppos,
22200 + (char *)page, length);
22204 + put_nx_info(nxi);
22211 +/* here comes the lower level */
22214 +#define NOD(NAME, MODE, IOP, FOP, OP) { \
22215 + .len = sizeof(NAME) - 1, \
22216 + .name = (NAME), \
22224 +#define DIR(NAME, MODE, OTYPE) \
22225 + NOD(NAME, (S_IFDIR | (MODE)), \
22226 + &proc_ ## OTYPE ## _inode_operations, \
22227 + &proc_ ## OTYPE ## _file_operations, { } )
22229 +#define INF(NAME, MODE, OTYPE) \
22230 + NOD(NAME, (S_IFREG | (MODE)), NULL, \
22231 + &proc_vs_info_file_operations, \
22232 + { .proc_vs_read = &proc_##OTYPE } )
22234 +#define VINF(NAME, MODE, OTYPE) \
22235 + NOD(NAME, (S_IFREG | (MODE)), NULL, \
22236 + &proc_vx_info_file_operations, \
22237 + { .proc_vxi_read = &proc_##OTYPE } )
22239 +#define NINF(NAME, MODE, OTYPE) \
22240 + NOD(NAME, (S_IFREG | (MODE)), NULL, \
22241 + &proc_nx_info_file_operations, \
22242 + { .proc_nxi_read = &proc_##OTYPE } )
22245 +static struct file_operations proc_vs_info_file_operations = {
22246 + .read = proc_vs_info_read,
22249 +static struct file_operations proc_vx_info_file_operations = {
22250 + .read = proc_vx_info_read,
22253 +static struct dentry_operations proc_xid_dentry_operations = {
22254 + .d_revalidate = proc_xid_revalidate,
22257 +static struct vs_entry vx_base_stuff[] = {
22258 + VINF("info", S_IRUGO, vxi_info),
22259 + VINF("status", S_IRUGO, vxi_status),
22260 + VINF("limit", S_IRUGO, vxi_limit),
22261 + VINF("sched", S_IRUGO, vxi_sched),
22262 + VINF("nsproxy", S_IRUGO, vxi_nsproxy0),
22263 + VINF("nsproxy1",S_IRUGO, vxi_nsproxy1),
22264 + VINF("cvirt", S_IRUGO, vxi_cvirt),
22265 + VINF("cacct", S_IRUGO, vxi_cacct),
22272 +static struct dentry *proc_xid_instantiate(struct inode *dir,
22273 + struct dentry *dentry, int id, void *ptr)
22275 + dentry->d_op = &proc_xid_dentry_operations;
22276 + return vs_proc_instantiate(dir, dentry, id, ptr);
22279 +static struct dentry *proc_xid_lookup(struct inode *dir,
22280 + struct dentry *dentry, struct nameidata *nd)
22282 + struct vs_entry *p = vx_base_stuff;
22283 + struct dentry *error = ERR_PTR(-ENOENT);
22285 + for (; p->name; p++) {
22286 + if (p->len != dentry->d_name.len)
22288 + if (!memcmp(dentry->d_name.name, p->name, p->len))
22294 + error = proc_xid_instantiate(dir, dentry, PROC_I(dir)->fd, p);
22299 +static int proc_xid_readdir(struct file *filp,
22300 + void *dirent, filldir_t filldir)
22302 + struct dentry *dentry = filp->f_dentry;
22303 + struct inode *inode = dentry->d_inode;
22304 + struct vs_entry *p = vx_base_stuff;
22305 + int size = sizeof(vx_base_stuff) / sizeof(struct vs_entry);
22309 + pos = filp->f_pos;
22312 + ino = inode->i_ino;
22313 + if (filldir(dirent, ".", 1, pos, ino, DT_DIR) < 0)
22316 + /* fall through */
22318 + ino = parent_ino(dentry);
22319 + if (filldir(dirent, "..", 2, pos, ino, DT_DIR) < 0)
22322 + /* fall through */
22325 + if (index >= size)
22327 + for (p += index; p->name; p++) {
22328 + if (proc_fill_cache(filp, dirent, filldir, p->name, p->len,
22329 + vs_proc_instantiate, PROC_I(inode)->fd, p))
22335 + filp->f_pos = pos;
22341 +static struct file_operations proc_nx_info_file_operations = {
22342 + .read = proc_nx_info_read,
22345 +static struct dentry_operations proc_nid_dentry_operations = {
22346 + .d_revalidate = proc_nid_revalidate,
22349 +static struct vs_entry nx_base_stuff[] = {
22350 + NINF("info", S_IRUGO, nxi_info),
22351 + NINF("status", S_IRUGO, nxi_status),
22356 +static struct dentry *proc_nid_instantiate(struct inode *dir,
22357 + struct dentry *dentry, int id, void *ptr)
22359 + dentry->d_op = &proc_nid_dentry_operations;
22360 + return vs_proc_instantiate(dir, dentry, id, ptr);
22363 +static struct dentry *proc_nid_lookup(struct inode *dir,
22364 + struct dentry *dentry, struct nameidata *nd)
22366 + struct vs_entry *p = nx_base_stuff;
22367 + struct dentry *error = ERR_PTR(-ENOENT);
22369 + for (; p->name; p++) {
22370 + if (p->len != dentry->d_name.len)
22372 + if (!memcmp(dentry->d_name.name, p->name, p->len))
22378 + error = proc_nid_instantiate(dir, dentry, PROC_I(dir)->fd, p);
22383 +static int proc_nid_readdir(struct file *filp,
22384 + void *dirent, filldir_t filldir)
22386 + struct dentry *dentry = filp->f_dentry;
22387 + struct inode *inode = dentry->d_inode;
22388 + struct vs_entry *p = nx_base_stuff;
22389 + int size = sizeof(nx_base_stuff) / sizeof(struct vs_entry);
22393 + pos = filp->f_pos;
22396 + ino = inode->i_ino;
22397 + if (filldir(dirent, ".", 1, pos, ino, DT_DIR) < 0)
22400 + /* fall through */
22402 + ino = parent_ino(dentry);
22403 + if (filldir(dirent, "..", 2, pos, ino, DT_DIR) < 0)
22406 + /* fall through */
22409 + if (index >= size)
22411 + for (p += index; p->name; p++) {
22412 + if (proc_fill_cache(filp, dirent, filldir, p->name, p->len,
22413 + vs_proc_instantiate, PROC_I(inode)->fd, p))
22419 + filp->f_pos = pos;
22424 +#define MAX_MULBY10 ((~0U - 9) / 10)
22426 +static inline int atovid(const char *str, int len)
22431 + while (len-- > 0) {
22436 + if (vid >= MAX_MULBY10)
22446 +/* now the upper level (virtual) */
22449 +static struct file_operations proc_xid_file_operations = {
22450 + .read = generic_read_dir,
22451 + .readdir = proc_xid_readdir,
22454 +static struct inode_operations proc_xid_inode_operations = {
22455 + .lookup = proc_xid_lookup,
22458 +static struct vs_entry vx_virtual_stuff[] = {
22459 + INF("info", S_IRUGO, virtual_info),
22460 + INF("status", S_IRUGO, virtual_status),
22461 + DIR(NULL, S_IRUGO | S_IXUGO, xid),
22465 +static struct dentry *proc_virtual_lookup(struct inode *dir,
22466 + struct dentry *dentry, struct nameidata *nd)
22468 + struct vs_entry *p = vx_virtual_stuff;
22469 + struct dentry *error = ERR_PTR(-ENOENT);
22472 + for (; p->name; p++) {
22473 + if (p->len != dentry->d_name.len)
22475 + if (!memcmp(dentry->d_name.name, p->name, p->len))
22479 + goto instantiate;
22481 + id = atovid(dentry->d_name.name, dentry->d_name.len);
22482 + if ((id < 0) || !xid_is_hashed(id))
22486 + error = proc_xid_instantiate(dir, dentry, id, p);
22491 +static struct file_operations proc_nid_file_operations = {
22492 + .read = generic_read_dir,
22493 + .readdir = proc_nid_readdir,
22496 +static struct inode_operations proc_nid_inode_operations = {
22497 + .lookup = proc_nid_lookup,
22500 +static struct vs_entry nx_virtnet_stuff[] = {
22501 + INF("info", S_IRUGO, virtnet_info),
22502 + INF("status", S_IRUGO, virtnet_status),
22503 + DIR(NULL, S_IRUGO | S_IXUGO, nid),
22507 +static struct dentry *proc_virtnet_lookup(struct inode *dir,
22508 + struct dentry *dentry, struct nameidata *nd)
22510 + struct vs_entry *p = nx_virtnet_stuff;
22511 + struct dentry *error = ERR_PTR(-ENOENT);
22514 + for (; p->name; p++) {
22515 + if (p->len != dentry->d_name.len)
22517 + if (!memcmp(dentry->d_name.name, p->name, p->len))
22521 + goto instantiate;
22523 + id = atovid(dentry->d_name.name, dentry->d_name.len);
22524 + if ((id < 0) || !nid_is_hashed(id))
22528 + error = proc_nid_instantiate(dir, dentry, id, p);
22534 +#define PROC_MAXVIDS 32
22536 +int proc_virtual_readdir(struct file *filp,
22537 + void *dirent, filldir_t filldir)
22539 + struct dentry *dentry = filp->f_dentry;
22540 + struct inode *inode = dentry->d_inode;
22541 + struct vs_entry *p = vx_virtual_stuff;
22542 + int size = sizeof(vx_virtual_stuff) / sizeof(struct vs_entry);
22544 + unsigned int xid_array[PROC_MAXVIDS];
22545 + char buf[PROC_NUMBUF];
22546 + unsigned int nr_xids, i;
22549 + pos = filp->f_pos;
22552 + ino = inode->i_ino;
22553 + if (filldir(dirent, ".", 1, pos, ino, DT_DIR) < 0)
22556 + /* fall through */
22558 + ino = parent_ino(dentry);
22559 + if (filldir(dirent, "..", 2, pos, ino, DT_DIR) < 0)
22562 + /* fall through */
22565 + if (index >= size)
22567 + for (p += index; p->name; p++) {
22568 + if (proc_fill_cache(filp, dirent, filldir, p->name, p->len,
22569 + vs_proc_instantiate, 0, p))
22574 + index = pos - size;
22575 + p = &vx_virtual_stuff[size - 1];
22576 + nr_xids = get_xid_list(index, xid_array, PROC_MAXVIDS);
22577 + for (i = 0; i < nr_xids; i++) {
22578 + int n, xid = xid_array[i];
22579 + unsigned int j = PROC_NUMBUF;
22583 + buf[--j] = '0' + (n % 10);
22586 + if (proc_fill_cache(filp, dirent, filldir,
22587 + buf + j, PROC_NUMBUF - j,
22588 + vs_proc_instantiate, xid, p))
22594 + filp->f_pos = pos;
22598 +static int proc_virtual_getattr(struct vfsmount *mnt,
22599 + struct dentry *dentry, struct kstat *stat)
22601 + struct inode *inode = dentry->d_inode;
22603 + generic_fillattr(inode, stat);
22604 + stat->nlink = 2 + atomic_read(&vx_global_cactive);
22608 +static struct file_operations proc_virtual_dir_operations = {
22609 + .read = generic_read_dir,
22610 + .readdir = proc_virtual_readdir,
22613 +static struct inode_operations proc_virtual_dir_inode_operations = {
22614 + .getattr = proc_virtual_getattr,
22615 + .lookup = proc_virtual_lookup,
22622 +int proc_virtnet_readdir(struct file *filp,
22623 + void *dirent, filldir_t filldir)
22625 + struct dentry *dentry = filp->f_dentry;
22626 + struct inode *inode = dentry->d_inode;
22627 + struct vs_entry *p = nx_virtnet_stuff;
22628 + int size = sizeof(nx_virtnet_stuff) / sizeof(struct vs_entry);
22630 + unsigned int nid_array[PROC_MAXVIDS];
22631 + char buf[PROC_NUMBUF];
22632 + unsigned int nr_nids, i;
22635 + pos = filp->f_pos;
22638 + ino = inode->i_ino;
22639 + if (filldir(dirent, ".", 1, pos, ino, DT_DIR) < 0)
22642 + /* fall through */
22644 + ino = parent_ino(dentry);
22645 + if (filldir(dirent, "..", 2, pos, ino, DT_DIR) < 0)
22648 + /* fall through */
22651 + if (index >= size)
22653 + for (p += index; p->name; p++) {
22654 + if (proc_fill_cache(filp, dirent, filldir, p->name, p->len,
22655 + vs_proc_instantiate, 0, p))
22660 + index = pos - size;
22661 + p = &nx_virtnet_stuff[size - 1];
22662 + nr_nids = get_nid_list(index, nid_array, PROC_MAXVIDS);
22663 + for (i = 0; i < nr_nids; i++) {
22664 + int n, nid = nid_array[i];
22665 + unsigned int j = PROC_NUMBUF;
22669 + buf[--j] = '0' + (n % 10);
22672 + if (proc_fill_cache(filp, dirent, filldir,
22673 + buf + j, PROC_NUMBUF - j,
22674 + vs_proc_instantiate, nid, p))
22680 + filp->f_pos = pos;
22684 +static int proc_virtnet_getattr(struct vfsmount *mnt,
22685 + struct dentry *dentry, struct kstat *stat)
22687 + struct inode *inode = dentry->d_inode;
22689 + generic_fillattr(inode, stat);
22690 + stat->nlink = 2 + atomic_read(&nx_global_cactive);
22694 +static struct file_operations proc_virtnet_dir_operations = {
22695 + .read = generic_read_dir,
22696 + .readdir = proc_virtnet_readdir,
22699 +static struct inode_operations proc_virtnet_dir_inode_operations = {
22700 + .getattr = proc_virtnet_getattr,
22701 + .lookup = proc_virtnet_lookup,
22706 +void proc_vx_init(void)
22708 + struct proc_dir_entry *ent;
22710 + ent = proc_mkdir("virtual", 0);
22712 + ent->proc_fops = &proc_virtual_dir_operations;
22713 + ent->proc_iops = &proc_virtual_dir_inode_operations;
22715 + proc_virtual = ent;
22717 + ent = proc_mkdir("virtnet", 0);
22719 + ent->proc_fops = &proc_virtnet_dir_operations;
22720 + ent->proc_iops = &proc_virtnet_dir_inode_operations;
22722 + proc_virtnet = ent;
22728 +/* per pid info */
22731 +int proc_pid_vx_info(struct task_struct *p, char *buffer)
22733 + struct vx_info *vxi;
22734 + char *orig = buffer;
22736 + buffer += sprintf(buffer, "XID:\t%d\n", vx_task_xid(p));
22738 + vxi = task_get_vx_info(p);
22742 + buffer += sprintf(buffer, "BCaps:\t");
22743 + buffer = print_cap_t(buffer, &vxi->vx_bcaps);
22744 + buffer += sprintf(buffer, "\n");
22745 + buffer += sprintf(buffer, "CCaps:\t%016llx\n",
22746 + (unsigned long long)vxi->vx_ccaps);
22747 + buffer += sprintf(buffer, "CFlags:\t%016llx\n",
22748 + (unsigned long long)vxi->vx_flags);
22749 + buffer += sprintf(buffer, "CIPid:\t%d\n", vxi->vx_initpid);
22751 + put_vx_info(vxi);
22753 + return buffer - orig;
22757 +int proc_pid_nx_info(struct task_struct *p, char *buffer)
22759 + struct nx_info *nxi;
22760 + struct nx_addr_v4 *v4a;
22761 +#ifdef CONFIG_IPV6
22762 + struct nx_addr_v6 *v6a;
22764 + char *orig = buffer;
22767 + buffer += sprintf(buffer, "NID:\t%d\n", nx_task_nid(p));
22769 + nxi = task_get_nx_info(p);
22773 + buffer += sprintf(buffer, "NCaps:\t%016llx\n",
22774 + (unsigned long long)nxi->nx_ncaps);
22775 + buffer += sprintf(buffer, "NFlags:\t%016llx\n",
22776 + (unsigned long long)nxi->nx_flags);
22778 + buffer += sprintf(buffer,
22779 + "V4Root[bcast]:\t" NIPQUAD_FMT "\n",
22780 + NIPQUAD(nxi->v4_bcast.s_addr));
22781 + buffer += sprintf (buffer,
22782 + "V4Root[lback]:\t" NIPQUAD_FMT "\n",
22783 + NIPQUAD(nxi->v4_lback.s_addr));
22784 + if (!NX_IPV4(nxi))
22786 + for (i = 0, v4a = &nxi->v4; v4a; i++, v4a = v4a->next)
22787 + buffer += sprintf(buffer, "V4Root[%d]:\t" NXAV4_FMT "\n",
22790 +#ifdef CONFIG_IPV6
22791 + if (!NX_IPV6(nxi))
22793 + for (i = 0, v6a = &nxi->v6; v6a; i++, v6a = v6a->next)
22794 + buffer += sprintf(buffer, "V6Root[%d]:\t" NXAV6_FMT "\n",
22798 + put_nx_info(nxi);
22800 + return buffer - orig;
22803 diff -NurpP --minimal linux-2.6.36/kernel/vserver/sched.c linux-2.6.36-vs2.3.0.36.33/kernel/vserver/sched.c
22804 --- linux-2.6.36/kernel/vserver/sched.c 1970-01-01 01:00:00.000000000 +0100
22805 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/sched.c 2010-10-21 13:09:36.000000000 +0200
22808 + * linux/kernel/vserver/sched.c
22810 + * Virtual Server: Scheduler Support
22812 + * Copyright (C) 2004-2007 Herbert Pötzl
22814 + * V0.01 adapted Sam Vilains version to 2.6.3
22815 + * V0.02 removed legacy interface
22816 + * V0.03 changed vcmds to vxi arg
22817 + * V0.04 removed older and legacy interfaces
22821 +#include <linux/vs_context.h>
22822 +#include <linux/vs_sched.h>
22823 +#include <linux/vserver/sched_cmd.h>
22825 +#include <asm/uaccess.h>
22828 +#define vxd_check_range(val, min, max) do { \
22829 + vxlprintk((val < min) || (val > max), \
22830 + "check_range(%ld,%ld,%ld)", \
22831 + (long)val, (long)min, (long)max, \
22832 + __FILE__, __LINE__); \
22836 +void vx_update_sched_param(struct _vx_sched *sched,
22837 + struct _vx_sched_pc *sched_pc)
22839 + unsigned int set_mask = sched->update_mask;
22841 + if (set_mask & VXSM_FILL_RATE)
22842 + sched_pc->fill_rate[0] = sched->fill_rate[0];
22843 + if (set_mask & VXSM_INTERVAL)
22844 + sched_pc->interval[0] = sched->interval[0];
22845 + if (set_mask & VXSM_FILL_RATE2)
22846 + sched_pc->fill_rate[1] = sched->fill_rate[1];
22847 + if (set_mask & VXSM_INTERVAL2)
22848 + sched_pc->interval[1] = sched->interval[1];
22849 + if (set_mask & VXSM_TOKENS)
22850 + sched_pc->tokens = sched->tokens;
22851 + if (set_mask & VXSM_TOKENS_MIN)
22852 + sched_pc->tokens_min = sched->tokens_min;
22853 + if (set_mask & VXSM_TOKENS_MAX)
22854 + sched_pc->tokens_max = sched->tokens_max;
22855 + if (set_mask & VXSM_PRIO_BIAS)
22856 + sched_pc->prio_bias = sched->prio_bias;
22858 + if (set_mask & VXSM_IDLE_TIME)
22859 + sched_pc->flags |= VXSF_IDLE_TIME;
22861 + sched_pc->flags &= ~VXSF_IDLE_TIME;
22864 + sched_pc->norm_time = jiffies;
22869 + * recalculate the context's scheduling tokens
22871 + * ret > 0 : number of tokens available
22872 + * ret < 0 : on hold, check delta_min[]
22873 + * -1 only jiffies
22874 + * -2 also idle time
22877 +int vx_tokens_recalc(struct _vx_sched_pc *sched_pc,
22878 + unsigned long *norm_time, unsigned long *idle_time, int delta_min[2])
22882 + int flags = sched_pc->flags;
22884 + /* how much time did pass? */
22885 + delta = *norm_time - sched_pc->norm_time;
22886 + // printk("@ %ld, %ld, %ld\n", *norm_time, sched_pc->norm_time, jiffies);
22887 + vxd_check_range(delta, 0, INT_MAX);
22889 + if (delta >= sched_pc->interval[0]) {
22890 + long tokens, integral;
22892 + /* calc integral token part */
22893 + tokens = delta / sched_pc->interval[0];
22894 + integral = tokens * sched_pc->interval[0];
22895 + tokens *= sched_pc->fill_rate[0];
22896 +#ifdef CONFIG_VSERVER_HARDCPU
22897 + delta_min[0] = delta - integral;
22898 + vxd_check_range(delta_min[0], 0, sched_pc->interval[0]);
22900 + /* advance time */
22901 + sched_pc->norm_time += delta;
22904 + sched_pc->tokens += tokens;
22905 + sched_pc->token_time += tokens;
22907 + delta_min[0] = delta;
22909 +#ifdef CONFIG_VSERVER_IDLETIME
22910 + if (!(flags & VXSF_IDLE_TIME))
22913 + /* how much was the idle skip? */
22914 + delta = *idle_time - sched_pc->idle_time;
22915 + vxd_check_range(delta, 0, INT_MAX);
22917 + if (delta >= sched_pc->interval[1]) {
22918 + long tokens, integral;
22920 + /* calc fair share token part */
22921 + tokens = delta / sched_pc->interval[1];
22922 + integral = tokens * sched_pc->interval[1];
22923 + tokens *= sched_pc->fill_rate[1];
22924 + delta_min[1] = delta - integral;
22925 + vxd_check_range(delta_min[1], 0, sched_pc->interval[1]);
22927 + /* advance idle time */
22928 + sched_pc->idle_time += integral;
22931 + sched_pc->tokens += tokens;
22932 + sched_pc->token_time += tokens;
22934 + delta_min[1] = delta;
22938 + /* clip at maximum */
22939 + if (sched_pc->tokens > sched_pc->tokens_max)
22940 + sched_pc->tokens = sched_pc->tokens_max;
22941 + tokens = sched_pc->tokens;
22943 + if ((flags & VXSF_ONHOLD)) {
22944 + /* can we unhold? */
22945 + if (tokens >= sched_pc->tokens_min) {
22946 + flags &= ~VXSF_ONHOLD;
22947 + sched_pc->hold_ticks +=
22948 + *norm_time - sched_pc->onhold;
22952 + /* put on hold? */
22953 + if (tokens <= 0) {
22954 + flags |= VXSF_ONHOLD;
22955 + sched_pc->onhold = *norm_time;
22959 + sched_pc->flags = flags;
22963 + tokens = sched_pc->tokens_min - tokens;
22964 + sched_pc->flags = flags;
22965 + // BUG_ON(tokens < 0); probably doesn't hold anymore
22967 +#ifdef CONFIG_VSERVER_HARDCPU
22968 + /* next interval? */
22969 + if (!sched_pc->fill_rate[0])
22970 + delta_min[0] = HZ;
22971 + else if (tokens > sched_pc->fill_rate[0])
22972 + delta_min[0] += sched_pc->interval[0] *
22973 + tokens / sched_pc->fill_rate[0];
22975 + delta_min[0] = sched_pc->interval[0] - delta_min[0];
22976 + vxd_check_range(delta_min[0], 0, INT_MAX);
22978 +#ifdef CONFIG_VSERVER_IDLETIME
22979 + if (!(flags & VXSF_IDLE_TIME))
22982 + /* next interval? */
22983 + if (!sched_pc->fill_rate[1])
22984 + delta_min[1] = HZ;
22985 + else if (tokens > sched_pc->fill_rate[1])
22986 + delta_min[1] += sched_pc->interval[1] *
22987 + tokens / sched_pc->fill_rate[1];
22989 + delta_min[1] = sched_pc->interval[1] - delta_min[1];
22990 + vxd_check_range(delta_min[1], 0, INT_MAX);
22995 +#endif /* CONFIG_VSERVER_IDLETIME */
22998 +#endif /* CONFIG_VSERVER_HARDCPU */
23001 +static inline unsigned long msec_to_ticks(unsigned long msec)
23003 + return msecs_to_jiffies(msec);
23006 +static inline unsigned long ticks_to_msec(unsigned long ticks)
23008 + return jiffies_to_msecs(ticks);
23011 +static inline unsigned long ticks_to_usec(unsigned long ticks)
23013 + return jiffies_to_usecs(ticks);
23017 +static int do_set_sched(struct vx_info *vxi, struct vcmd_sched_v5 *data)
23019 + unsigned int set_mask = data->mask;
23020 + unsigned int update_mask;
23023 + /* Sanity check data values */
23024 + if (data->tokens_max <= 0)
23025 + data->tokens_max = HZ;
23026 + if (data->tokens_min < 0)
23027 + data->tokens_min = HZ / 3;
23028 + if (data->tokens_min >= data->tokens_max)
23029 + data->tokens_min = data->tokens_max;
23031 + if (data->prio_bias > MAX_PRIO_BIAS)
23032 + data->prio_bias = MAX_PRIO_BIAS;
23033 + if (data->prio_bias < MIN_PRIO_BIAS)
23034 + data->prio_bias = MIN_PRIO_BIAS;
23036 + spin_lock(&vxi->sched.tokens_lock);
23038 + /* sync up on delayed updates */
23039 + for_each_cpu_mask(cpu, vxi->sched.update)
23040 + vx_update_sched_param(&vxi->sched,
23041 + &vx_per_cpu(vxi, sched_pc, cpu));
23043 + if (set_mask & VXSM_FILL_RATE)
23044 + vxi->sched.fill_rate[0] = data->fill_rate[0];
23045 + if (set_mask & VXSM_FILL_RATE2)
23046 + vxi->sched.fill_rate[1] = data->fill_rate[1];
23047 + if (set_mask & VXSM_INTERVAL)
23048 + vxi->sched.interval[0] = (set_mask & VXSM_MSEC) ?
23049 + msec_to_ticks(data->interval[0]) : data->interval[0];
23050 + if (set_mask & VXSM_INTERVAL2)
23051 + vxi->sched.interval[1] = (set_mask & VXSM_MSEC) ?
23052 + msec_to_ticks(data->interval[1]) : data->interval[1];
23053 + if (set_mask & VXSM_TOKENS)
23054 + vxi->sched.tokens = data->tokens;
23055 + if (set_mask & VXSM_TOKENS_MIN)
23056 + vxi->sched.tokens_min = data->tokens_min;
23057 + if (set_mask & VXSM_TOKENS_MAX)
23058 + vxi->sched.tokens_max = data->tokens_max;
23059 + if (set_mask & VXSM_PRIO_BIAS)
23060 + vxi->sched.prio_bias = data->prio_bias;
23062 + /* Sanity check rate/interval */
23063 + for (i = 0; i < 2; i++) {
23064 + if (data->fill_rate[i] < 0)
23065 + data->fill_rate[i] = 0;
23066 + if (data->interval[i] <= 0)
23067 + data->interval[i] = HZ;
23070 + update_mask = vxi->sched.update_mask & VXSM_SET_MASK;
23071 + update_mask |= (set_mask & (VXSM_SET_MASK | VXSM_IDLE_TIME));
23072 + vxi->sched.update_mask = update_mask;
23076 + if (set_mask & VXSM_CPU_ID) {
23077 + vxi->sched.update = cpumask_of_cpu(data->cpu_id);
23078 + cpus_and(vxi->sched.update, cpu_online_map,
23079 + vxi->sched.update);
23081 + vxi->sched.update = cpu_online_map;
23083 + /* forced reload? */
23084 + if (set_mask & VXSM_FORCE) {
23085 + for_each_cpu_mask(cpu, vxi->sched.update)
23086 + vx_update_sched_param(&vxi->sched,
23087 + &vx_per_cpu(vxi, sched_pc, cpu));
23088 + vxi->sched.update = CPU_MASK_NONE;
23091 + /* on UP we update immediately */
23092 + vx_update_sched_param(&vxi->sched,
23093 + &vx_per_cpu(vxi, sched_pc, 0));
23096 + spin_unlock(&vxi->sched.tokens_lock);
23101 +#define COPY_IDS(C) C(cpu_id); C(bucket_id)
23102 +#define COPY_PRI(C) C(prio_bias)
23103 +#define COPY_TOK(C) C(tokens); C(tokens_min); C(tokens_max)
23104 +#define COPY_FRI(C) C(fill_rate[0]); C(interval[0]); \
23105 + C(fill_rate[1]); C(interval[1]);
23107 +#define COPY_VALUE(name) vc_data.name = data->name
23109 +static int do_set_sched_v4(struct vx_info *vxi, struct vcmd_set_sched_v4 *data)
23111 + struct vcmd_sched_v5 vc_data;
23113 + vc_data.mask = data->set_mask;
23114 + COPY_IDS(COPY_VALUE);
23115 + COPY_PRI(COPY_VALUE);
23116 + COPY_TOK(COPY_VALUE);
23117 + vc_data.fill_rate[0] = vc_data.fill_rate[1] = data->fill_rate;
23118 + vc_data.interval[0] = vc_data.interval[1] = data->interval;
23119 + return do_set_sched(vxi, &vc_data);
23122 +int vc_set_sched_v4(struct vx_info *vxi, void __user *data)
23124 + struct vcmd_set_sched_v4 vc_data;
23126 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
23129 + return do_set_sched_v4(vxi, &vc_data);
23132 + /* latest interface is v5 */
23134 +int vc_set_sched(struct vx_info *vxi, void __user *data)
23136 + struct vcmd_sched_v5 vc_data;
23138 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
23141 + return do_set_sched(vxi, &vc_data);
23145 +#define COPY_PRI(C) C(prio_bias)
23146 +#define COPY_TOK(C) C(tokens); C(tokens_min); C(tokens_max)
23147 +#define COPY_FRI(C) C(fill_rate[0]); C(interval[0]); \
23148 + C(fill_rate[1]); C(interval[1]);
23150 +#define COPY_VALUE(name) vc_data.name = data->name
23153 +int vc_get_sched(struct vx_info *vxi, void __user *data)
23155 + struct vcmd_sched_v5 vc_data;
23157 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
23160 + if (vc_data.mask & VXSM_CPU_ID) {
23161 + int cpu = vc_data.cpu_id;
23162 + struct _vx_sched_pc *data;
23164 + if (!cpu_possible(cpu))
23167 + data = &vx_per_cpu(vxi, sched_pc, cpu);
23168 + COPY_TOK(COPY_VALUE);
23169 + COPY_PRI(COPY_VALUE);
23170 + COPY_FRI(COPY_VALUE);
23172 + if (data->flags & VXSF_IDLE_TIME)
23173 + vc_data.mask |= VXSM_IDLE_TIME;
23175 + struct _vx_sched *data = &vxi->sched;
23177 + COPY_TOK(COPY_VALUE);
23178 + COPY_PRI(COPY_VALUE);
23179 + COPY_FRI(COPY_VALUE);
23182 + if (vc_data.mask & VXSM_MSEC) {
23183 + vc_data.interval[0] = ticks_to_msec(vc_data.interval[0]);
23184 + vc_data.interval[1] = ticks_to_msec(vc_data.interval[1]);
23187 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
23193 +int vc_sched_info(struct vx_info *vxi, void __user *data)
23195 + struct vcmd_sched_info vc_data;
23198 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
23201 + cpu = vc_data.cpu_id;
23202 + if (!cpu_possible(cpu))
23206 + struct _vx_sched_pc *sched_pc =
23207 + &vx_per_cpu(vxi, sched_pc, cpu);
23209 + vc_data.user_msec = ticks_to_msec(sched_pc->user_ticks);
23210 + vc_data.sys_msec = ticks_to_msec(sched_pc->sys_ticks);
23211 + vc_data.hold_msec = ticks_to_msec(sched_pc->hold_ticks);
23212 + vc_data.vavavoom = sched_pc->vavavoom;
23214 + vc_data.token_usec = ticks_to_usec(1);
23216 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
23221 diff -NurpP --minimal linux-2.6.36/kernel/vserver/sched_init.h linux-2.6.36-vs2.3.0.36.33/kernel/vserver/sched_init.h
23222 --- linux-2.6.36/kernel/vserver/sched_init.h 1970-01-01 01:00:00.000000000 +0100
23223 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/sched_init.h 2010-10-21 13:09:36.000000000 +0200
23226 +static inline void vx_info_init_sched(struct _vx_sched *sched)
23228 + static struct lock_class_key tokens_lock_key;
23230 + /* scheduling; hard code starting values as constants */
23231 + sched->fill_rate[0] = 1;
23232 + sched->interval[0] = 4;
23233 + sched->fill_rate[1] = 1;
23234 + sched->interval[1] = 8;
23235 + sched->tokens = HZ >> 2;
23236 + sched->tokens_min = HZ >> 4;
23237 + sched->tokens_max = HZ >> 1;
23238 + sched->tokens_lock = SPIN_LOCK_UNLOCKED;
23239 + sched->prio_bias = 0;
23241 + lockdep_set_class(&sched->tokens_lock, &tokens_lock_key);
23245 +void vx_info_init_sched_pc(struct _vx_sched_pc *sched_pc, int cpu)
23247 + sched_pc->fill_rate[0] = 1;
23248 + sched_pc->interval[0] = 4;
23249 + sched_pc->fill_rate[1] = 1;
23250 + sched_pc->interval[1] = 8;
23251 + sched_pc->tokens = HZ >> 2;
23252 + sched_pc->tokens_min = HZ >> 4;
23253 + sched_pc->tokens_max = HZ >> 1;
23254 + sched_pc->prio_bias = 0;
23255 + sched_pc->vavavoom = 0;
23256 + sched_pc->token_time = 0;
23257 + sched_pc->idle_time = 0;
23258 + sched_pc->norm_time = jiffies;
23260 + sched_pc->user_ticks = 0;
23261 + sched_pc->sys_ticks = 0;
23262 + sched_pc->hold_ticks = 0;
23265 +static inline void vx_info_exit_sched(struct _vx_sched *sched)
23271 +void vx_info_exit_sched_pc(struct _vx_sched_pc *sched_pc, int cpu)
23275 diff -NurpP --minimal linux-2.6.36/kernel/vserver/sched_proc.h linux-2.6.36-vs2.3.0.36.33/kernel/vserver/sched_proc.h
23276 --- linux-2.6.36/kernel/vserver/sched_proc.h 1970-01-01 01:00:00.000000000 +0100
23277 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/sched_proc.h 2010-10-21 13:09:36.000000000 +0200
23279 +#ifndef _VX_SCHED_PROC_H
23280 +#define _VX_SCHED_PROC_H
23284 +int vx_info_proc_sched(struct _vx_sched *sched, char *buffer)
23288 + length += sprintf(buffer,
23289 + "FillRate:\t%8d,%d\n"
23290 + "Interval:\t%8d,%d\n"
23291 + "TokensMin:\t%8d\n"
23292 + "TokensMax:\t%8d\n"
23293 + "PrioBias:\t%8d\n",
23294 + sched->fill_rate[0],
23295 + sched->fill_rate[1],
23296 + sched->interval[0],
23297 + sched->interval[1],
23298 + sched->tokens_min,
23299 + sched->tokens_max,
23300 + sched->prio_bias);
23305 +int vx_info_proc_sched_pc(struct _vx_sched_pc *sched_pc,
23306 + char *buffer, int cpu)
23310 + length += sprintf(buffer + length,
23311 + "cpu %d: %lld %lld %lld %ld %ld", cpu,
23312 + (unsigned long long)sched_pc->user_ticks,
23313 + (unsigned long long)sched_pc->sys_ticks,
23314 + (unsigned long long)sched_pc->hold_ticks,
23315 + sched_pc->token_time,
23316 + sched_pc->idle_time);
23317 + length += sprintf(buffer + length,
23318 + " %c%c %d %d %d %d/%d %d/%d",
23319 + (sched_pc->flags & VXSF_ONHOLD) ? 'H' : 'R',
23320 + (sched_pc->flags & VXSF_IDLE_TIME) ? 'I' : '-',
23321 + sched_pc->tokens,
23322 + sched_pc->tokens_min,
23323 + sched_pc->tokens_max,
23324 + sched_pc->fill_rate[0],
23325 + sched_pc->interval[0],
23326 + sched_pc->fill_rate[1],
23327 + sched_pc->interval[1]);
23328 + length += sprintf(buffer + length,
23330 + sched_pc->prio_bias,
23331 + sched_pc->vavavoom);
23335 +#endif /* _VX_SCHED_PROC_H */
23336 diff -NurpP --minimal linux-2.6.36/kernel/vserver/signal.c linux-2.6.36-vs2.3.0.36.33/kernel/vserver/signal.c
23337 --- linux-2.6.36/kernel/vserver/signal.c 1970-01-01 01:00:00.000000000 +0100
23338 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/signal.c 2010-10-21 13:09:36.000000000 +0200
23341 + * linux/kernel/vserver/signal.c
23343 + * Virtual Server: Signal Support
23345 + * Copyright (C) 2003-2007 Herbert Pötzl
23347 + * V0.01 broken out from vcontext V0.05
23348 + * V0.02 changed vcmds to vxi arg
23349 + * V0.03 adjusted siginfo for kill
23353 +#include <asm/uaccess.h>
23355 +#include <linux/vs_context.h>
23356 +#include <linux/vs_pid.h>
23357 +#include <linux/vserver/signal_cmd.h>
23360 +int vx_info_kill(struct vx_info *vxi, int pid, int sig)
23362 + int retval, count = 0;
23363 + struct task_struct *p;
23364 + struct siginfo *sip = SEND_SIG_PRIV;
23367 + vxdprintk(VXD_CBIT(misc, 4),
23368 + "vx_info_kill(%p[#%d],%d,%d)*",
23369 + vxi, vxi->vx_id, pid, sig);
23370 + read_lock(&tasklist_lock);
23374 + for_each_process(p) {
23377 + if (vx_task_xid(p) != vxi->vx_id || p->pid <= 1 ||
23378 + (pid && vxi->vx_initpid == p->pid))
23381 + err = group_send_sig_info(sig, sip, p);
23383 + if (err != -EPERM)
23389 + if (vxi->vx_initpid) {
23390 + pid = vxi->vx_initpid;
23391 + /* for now, only SIGINT to private init ... */
23392 + if (!vx_info_flags(vxi, VXF_STATE_ADMIN, 0) &&
23393 + /* ... as long as there are tasks left */
23394 + (atomic_read(&vxi->vx_tasks) > 1))
23397 + /* fallthrough */
23399 + p = find_task_by_real_pid(pid);
23401 + if (vx_task_xid(p) == vxi->vx_id)
23402 + retval = group_send_sig_info(sig, sip, p);
23406 + read_unlock(&tasklist_lock);
23407 + vxdprintk(VXD_CBIT(misc, 4),
23408 + "vx_info_kill(%p[#%d],%d,%d,%ld) = %d",
23409 + vxi, vxi->vx_id, pid, sig, (long)sip, retval);
23413 +int vc_ctx_kill(struct vx_info *vxi, void __user *data)
23415 + struct vcmd_ctx_kill_v0 vc_data;
23417 + if (copy_from_user(&vc_data, data, sizeof(vc_data)))
23420 + /* special check to allow guest shutdown */
23421 + if (!vx_info_flags(vxi, VXF_STATE_ADMIN, 0) &&
23422 + /* forbid killall pid=0 when init is present */
23423 + (((vc_data.pid < 1) && vxi->vx_initpid) ||
23424 + (vc_data.pid > 1)))
23427 + return vx_info_kill(vxi, vc_data.pid, vc_data.sig);
23431 +static int __wait_exit(struct vx_info *vxi)
23433 + DECLARE_WAITQUEUE(wait, current);
23436 + add_wait_queue(&vxi->vx_wait, &wait);
23437 + set_current_state(TASK_INTERRUPTIBLE);
23440 + if (vx_info_state(vxi,
23441 + VXS_SHUTDOWN | VXS_HASHED | VXS_HELPER) == VXS_SHUTDOWN)
23443 + if (signal_pending(current)) {
23444 + ret = -ERESTARTSYS;
23451 + set_current_state(TASK_RUNNING);
23452 + remove_wait_queue(&vxi->vx_wait, &wait);
23458 +int vc_wait_exit(struct vx_info *vxi, void __user *data)
23460 + struct vcmd_wait_exit_v0 vc_data;
23463 + ret = __wait_exit(vxi);
23464 + vc_data.reboot_cmd = vxi->reboot_cmd;
23465 + vc_data.exit_code = vxi->exit_code;
23467 + if (copy_to_user(data, &vc_data, sizeof(vc_data)))
23472 diff -NurpP --minimal linux-2.6.36/kernel/vserver/space.c linux-2.6.36-vs2.3.0.36.33/kernel/vserver/space.c
23473 --- linux-2.6.36/kernel/vserver/space.c 1970-01-01 01:00:00.000000000 +0100
23474 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/space.c 2010-10-21 14:41:06.000000000 +0200
23477 + * linux/kernel/vserver/space.c
23479 + * Virtual Server: Context Space Support
23481 + * Copyright (C) 2003-2007 Herbert Pötzl
23483 + * V0.01 broken out from context.c 0.07
23484 + * V0.02 added task locking for namespace
23485 + * V0.03 broken out vx_enter_namespace
23486 + * V0.04 added *space support and commands
23490 +#include <linux/utsname.h>
23491 +#include <linux/nsproxy.h>
23492 +#include <linux/err.h>
23493 +#include <linux/fs_struct.h>
23494 +#include <asm/uaccess.h>
23496 +#include <linux/vs_context.h>
23497 +#include <linux/vserver/space.h>
23498 +#include <linux/vserver/space_cmd.h>
23500 +atomic_t vs_global_nsproxy = ATOMIC_INIT(0);
23501 +atomic_t vs_global_fs = ATOMIC_INIT(0);
23502 +atomic_t vs_global_mnt_ns = ATOMIC_INIT(0);
23503 +atomic_t vs_global_uts_ns = ATOMIC_INIT(0);
23504 +atomic_t vs_global_user_ns = ATOMIC_INIT(0);
23505 +atomic_t vs_global_pid_ns = ATOMIC_INIT(0);
23508 +/* namespace functions */
23510 +#include <linux/mnt_namespace.h>
23511 +#include <linux/user_namespace.h>
23512 +#include <linux/pid_namespace.h>
23513 +#include <linux/ipc_namespace.h>
23514 +#include <net/net_namespace.h>
23517 +static const struct vcmd_space_mask_v1 space_mask_v0 = {
23518 + .mask = CLONE_FS |
23526 +static const struct vcmd_space_mask_v1 space_mask = {
23527 + .mask = CLONE_FS |
23532 +#ifdef CONFIG_PID_NS
23535 +#ifdef CONFIG_NET_NS
23541 +static const struct vcmd_space_mask_v1 default_space_mask = {
23542 + .mask = CLONE_FS |
23547 +#ifdef CONFIG_PID_NS
23554 + * build a new nsproxy mix
23555 + * assumes that both proxies are 'const'
23556 + * does not touch nsproxy refcounts
23557 + * will hold a reference on the result.
23560 +struct nsproxy *vs_mix_nsproxy(struct nsproxy *old_nsproxy,
23561 + struct nsproxy *new_nsproxy, unsigned long mask)
23563 + struct mnt_namespace *old_ns;
23564 + struct uts_namespace *old_uts;
23565 + struct ipc_namespace *old_ipc;
23566 +#ifdef CONFIG_PID_NS
23567 + struct pid_namespace *old_pid;
23569 +#ifdef CONFIG_NET_NS
23570 + struct net *old_net;
23572 + struct nsproxy *nsproxy;
23574 + nsproxy = copy_nsproxy(old_nsproxy);
23578 + if (mask & CLONE_NEWNS) {
23579 + old_ns = nsproxy->mnt_ns;
23580 + nsproxy->mnt_ns = new_nsproxy->mnt_ns;
23581 + if (nsproxy->mnt_ns)
23582 + get_mnt_ns(nsproxy->mnt_ns);
23586 + if (mask & CLONE_NEWUTS) {
23587 + old_uts = nsproxy->uts_ns;
23588 + nsproxy->uts_ns = new_nsproxy->uts_ns;
23589 + if (nsproxy->uts_ns)
23590 + get_uts_ns(nsproxy->uts_ns);
23594 + if (mask & CLONE_NEWIPC) {
23595 + old_ipc = nsproxy->ipc_ns;
23596 + nsproxy->ipc_ns = new_nsproxy->ipc_ns;
23597 + if (nsproxy->ipc_ns)
23598 + get_ipc_ns(nsproxy->ipc_ns);
23602 +#ifdef CONFIG_PID_NS
23603 + if (mask & CLONE_NEWPID) {
23604 + old_pid = nsproxy->pid_ns;
23605 + nsproxy->pid_ns = new_nsproxy->pid_ns;
23606 + if (nsproxy->pid_ns)
23607 + get_pid_ns(nsproxy->pid_ns);
23611 +#ifdef CONFIG_NET_NS
23612 + if (mask & CLONE_NEWNET) {
23613 + old_net = nsproxy->net_ns;
23614 + nsproxy->net_ns = new_nsproxy->net_ns;
23615 + if (nsproxy->net_ns)
23616 + get_net(nsproxy->net_ns);
23621 + put_mnt_ns(old_ns);
23623 + put_uts_ns(old_uts);
23625 + put_ipc_ns(old_ipc);
23626 +#ifdef CONFIG_PID_NS
23628 + put_pid_ns(old_pid);
23630 +#ifdef CONFIG_NET_NS
23632 + put_net(old_net);
23640 + * merge two nsproxy structs into a new one.
23641 + * will hold a reference on the result.
23645 +struct nsproxy *__vs_merge_nsproxy(struct nsproxy *old,
23646 + struct nsproxy *proxy, unsigned long mask)
23648 + struct nsproxy null_proxy = { .mnt_ns = NULL };
23654 + /* vs_mix_nsproxy returns with reference */
23655 + return vs_mix_nsproxy(old ? old : &null_proxy,
23658 + get_nsproxy(proxy);
23663 +int vx_enter_space(struct vx_info *vxi, unsigned long mask, unsigned index)
23665 + struct nsproxy *proxy, *proxy_cur, *proxy_new;
23666 + struct fs_struct *fs_cur, *fs = NULL;
23667 + int ret, kill = 0;
23669 + vxdprintk(VXD_CBIT(space, 8), "vx_enter_space(%p[#%u],0x%08lx,%d)",
23670 + vxi, vxi->vx_id, mask, index);
23672 + if (vx_info_flags(vxi, VXF_INFO_PRIVATE, 0))
23676 + mask = vxi->vx_nsmask[index];
23678 + if ((mask & vxi->vx_nsmask[index]) != mask)
23681 + if (mask & CLONE_FS) {
23682 + fs = copy_fs_struct(vxi->vx_fs[index]);
23686 + proxy = vxi->vx_nsproxy[index];
23688 + vxdprintk(VXD_CBIT(space, 9),
23689 + "vx_enter_space(%p[#%u],0x%08lx,%d) -> (%p,%p)",
23690 + vxi, vxi->vx_id, mask, index, proxy, fs);
23692 + task_lock(current);
23693 + fs_cur = current->fs;
23695 + if (mask & CLONE_FS) {
23696 + spin_lock(&fs_cur->lock);
23697 + current->fs = fs;
23698 + kill = !--fs_cur->users;
23699 + spin_unlock(&fs_cur->lock);
23702 + proxy_cur = current->nsproxy;
23703 + get_nsproxy(proxy_cur);
23704 + task_unlock(current);
23707 + free_fs_struct(fs_cur);
23709 + proxy_new = __vs_merge_nsproxy(proxy_cur, proxy, mask);
23710 + if (IS_ERR(proxy_new)) {
23711 + ret = PTR_ERR(proxy_new);
23715 + proxy_new = xchg(¤t->nsproxy, proxy_new);
23719 + put_nsproxy(proxy_new);
23722 + put_nsproxy(proxy_cur);
23727 +int vx_set_space(struct vx_info *vxi, unsigned long mask, unsigned index)
23729 + struct nsproxy *proxy_vxi, *proxy_cur, *proxy_new;
23730 + struct fs_struct *fs_vxi, *fs;
23731 + int ret, kill = 0;
23733 + vxdprintk(VXD_CBIT(space, 8), "vx_set_space(%p[#%u],0x%08lx,%d)",
23734 + vxi, vxi->vx_id, mask, index);
23737 + mask = default_space_mask.mask;
23739 + if ((mask & space_mask.mask) != mask)
23742 + proxy_vxi = vxi->vx_nsproxy[index];
23743 + fs_vxi = vxi->vx_fs[index];
23745 + if (mask & CLONE_FS) {
23746 + fs = copy_fs_struct(current->fs);
23751 + task_lock(current);
23753 + if (mask & CLONE_FS) {
23754 + spin_lock(&fs_vxi->lock);
23755 + vxi->vx_fs[index] = fs;
23756 + kill = !--fs_vxi->users;
23757 + spin_unlock(&fs_vxi->lock);
23760 + proxy_cur = current->nsproxy;
23761 + get_nsproxy(proxy_cur);
23762 + task_unlock(current);
23765 + free_fs_struct(fs_vxi);
23767 + proxy_new = __vs_merge_nsproxy(proxy_vxi, proxy_cur, mask);
23768 + if (IS_ERR(proxy_new)) {
23769 + ret = PTR_ERR(proxy_new);
23773 + proxy_new = xchg(&vxi->vx_nsproxy[index], proxy_new);
23774 + vxi->vx_nsmask[index] |= mask;
23778 + put_nsproxy(proxy_new);
23781 + put_nsproxy(proxy_cur);
23786 +int vc_enter_space_v1(struct vx_info *vxi, void __user *data)
23788 + struct vcmd_space_mask_v1 vc_data = { .mask = 0 };
23790 + if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
23793 + return vx_enter_space(vxi, vc_data.mask, 0);
23796 +int vc_enter_space(struct vx_info *vxi, void __user *data)
23798 + struct vcmd_space_mask_v2 vc_data = { .mask = 0 };
23800 + if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
23803 + if (vc_data.index >= VX_SPACES)
23806 + return vx_enter_space(vxi, vc_data.mask, vc_data.index);
23809 +int vc_set_space_v1(struct vx_info *vxi, void __user *data)
23811 + struct vcmd_space_mask_v1 vc_data = { .mask = 0 };
23813 + if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
23816 + return vx_set_space(vxi, vc_data.mask, 0);
23819 +int vc_set_space(struct vx_info *vxi, void __user *data)
23821 + struct vcmd_space_mask_v2 vc_data = { .mask = 0 };
23823 + if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
23826 + if (vc_data.index >= VX_SPACES)
23829 + return vx_set_space(vxi, vc_data.mask, vc_data.index);
23832 +int vc_get_space_mask(void __user *data, int type)
23834 + const struct vcmd_space_mask_v1 *mask;
23837 + mask = &space_mask_v0;
23838 + else if (type == 1)
23839 + mask = &space_mask;
23841 + mask = &default_space_mask;
23843 + vxdprintk(VXD_CBIT(space, 10),
23844 + "vc_get_space_mask(%d) = %08llx", type, mask->mask);
23846 + if (copy_to_user(data, mask, sizeof(*mask)))
23851 diff -NurpP --minimal linux-2.6.36/kernel/vserver/switch.c linux-2.6.36-vs2.3.0.36.33/kernel/vserver/switch.c
23852 --- linux-2.6.36/kernel/vserver/switch.c 1970-01-01 01:00:00.000000000 +0100
23853 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/switch.c 2010-10-21 13:09:36.000000000 +0200
23856 + * linux/kernel/vserver/switch.c
23858 + * Virtual Server: Syscall Switch
23860 + * Copyright (C) 2003-2007 Herbert Pötzl
23862 + * V0.01 syscall switch
23863 + * V0.02 added signal to context
23864 + * V0.03 added rlimit functions
23865 + * V0.04 added iattr, task/xid functions
23866 + * V0.05 added debug/history stuff
23867 + * V0.06 added compat32 layer
23868 + * V0.07 vcmd args and perms
23869 + * V0.08 added status commands
23870 + * V0.09 added tag commands
23871 + * V0.10 added oom bias
23872 + * V0.11 added device commands
23876 +#include <linux/vs_context.h>
23877 +#include <linux/vs_network.h>
23878 +#include <linux/vserver/switch.h>
23880 +#include "vci_config.h"
23884 +int vc_get_version(uint32_t id)
23886 + return VCI_VERSION;
23890 +int vc_get_vci(uint32_t id)
23892 + return vci_kernel_config();
23895 +#include <linux/vserver/context_cmd.h>
23896 +#include <linux/vserver/cvirt_cmd.h>
23897 +#include <linux/vserver/cacct_cmd.h>
23898 +#include <linux/vserver/limit_cmd.h>
23899 +#include <linux/vserver/network_cmd.h>
23900 +#include <linux/vserver/sched_cmd.h>
23901 +#include <linux/vserver/debug_cmd.h>
23902 +#include <linux/vserver/inode_cmd.h>
23903 +#include <linux/vserver/dlimit_cmd.h>
23904 +#include <linux/vserver/signal_cmd.h>
23905 +#include <linux/vserver/space_cmd.h>
23906 +#include <linux/vserver/tag_cmd.h>
23907 +#include <linux/vserver/device_cmd.h>
23909 +#include <linux/vserver/inode.h>
23910 +#include <linux/vserver/dlimit.h>
23913 +#ifdef CONFIG_COMPAT
23914 +#define __COMPAT(name, id, data, compat) \
23915 + (compat) ? name ## _x32(id, data) : name(id, data)
23916 +#define __COMPAT_NO_ID(name, data, compat) \
23917 + (compat) ? name ## _x32(data) : name(data)
23919 +#define __COMPAT(name, id, data, compat) \
23921 +#define __COMPAT_NO_ID(name, data, compat) \
23927 +long do_vcmd(uint32_t cmd, uint32_t id,
23928 + struct vx_info *vxi, struct nx_info *nxi,
23929 + void __user *data, int compat)
23933 + case VCMD_get_version:
23934 + return vc_get_version(id);
23935 + case VCMD_get_vci:
23936 + return vc_get_vci(id);
23938 + case VCMD_task_xid:
23939 + return vc_task_xid(id);
23940 + case VCMD_vx_info:
23941 + return vc_vx_info(vxi, data);
23943 + case VCMD_task_nid:
23944 + return vc_task_nid(id);
23945 + case VCMD_nx_info:
23946 + return vc_nx_info(nxi, data);
23948 + case VCMD_task_tag:
23949 + return vc_task_tag(id);
23951 + case VCMD_set_space_v1:
23952 + return vc_set_space_v1(vxi, data);
23953 + /* this is version 2 */
23954 + case VCMD_set_space:
23955 + return vc_set_space(vxi, data);
23957 + case VCMD_get_space_mask_v0:
23958 + return vc_get_space_mask(data, 0);
23959 + /* this is version 1 */
23960 + case VCMD_get_space_mask:
23961 + return vc_get_space_mask(data, 1);
23963 + case VCMD_get_space_default:
23964 + return vc_get_space_mask(data, -1);
23966 +#ifdef CONFIG_IA32_EMULATION
23967 + case VCMD_get_rlimit:
23968 + return __COMPAT(vc_get_rlimit, vxi, data, compat);
23969 + case VCMD_set_rlimit:
23970 + return __COMPAT(vc_set_rlimit, vxi, data, compat);
23972 + case VCMD_get_rlimit:
23973 + return vc_get_rlimit(vxi, data);
23974 + case VCMD_set_rlimit:
23975 + return vc_set_rlimit(vxi, data);
23977 + case VCMD_get_rlimit_mask:
23978 + return vc_get_rlimit_mask(id, data);
23979 + case VCMD_reset_hits:
23980 + return vc_reset_hits(vxi, data);
23981 + case VCMD_reset_minmax:
23982 + return vc_reset_minmax(vxi, data);
23984 + case VCMD_get_vhi_name:
23985 + return vc_get_vhi_name(vxi, data);
23986 + case VCMD_set_vhi_name:
23987 + return vc_set_vhi_name(vxi, data);
23989 + case VCMD_ctx_stat:
23990 + return vc_ctx_stat(vxi, data);
23991 + case VCMD_virt_stat:
23992 + return vc_virt_stat(vxi, data);
23993 + case VCMD_sock_stat:
23994 + return vc_sock_stat(vxi, data);
23995 + case VCMD_rlimit_stat:
23996 + return vc_rlimit_stat(vxi, data);
23998 + case VCMD_set_cflags:
23999 + return vc_set_cflags(vxi, data);
24000 + case VCMD_get_cflags:
24001 + return vc_get_cflags(vxi, data);
24003 + /* this is version 1 */
24004 + case VCMD_set_ccaps:
24005 + return vc_set_ccaps(vxi, data);
24006 + /* this is version 1 */
24007 + case VCMD_get_ccaps:
24008 + return vc_get_ccaps(vxi, data);
24009 + case VCMD_set_bcaps:
24010 + return vc_set_bcaps(vxi, data);
24011 + case VCMD_get_bcaps:
24012 + return vc_get_bcaps(vxi, data);
24014 + case VCMD_set_badness:
24015 + return vc_set_badness(vxi, data);
24016 + case VCMD_get_badness:
24017 + return vc_get_badness(vxi, data);
24019 + case VCMD_set_nflags:
24020 + return vc_set_nflags(nxi, data);
24021 + case VCMD_get_nflags:
24022 + return vc_get_nflags(nxi, data);
24024 + case VCMD_set_ncaps:
24025 + return vc_set_ncaps(nxi, data);
24026 + case VCMD_get_ncaps:
24027 + return vc_get_ncaps(nxi, data);
24029 + case VCMD_set_sched_v4:
24030 + return vc_set_sched_v4(vxi, data);
24031 + /* this is version 5 */
24032 + case VCMD_set_sched:
24033 + return vc_set_sched(vxi, data);
24034 + case VCMD_get_sched:
24035 + return vc_get_sched(vxi, data);
24036 + case VCMD_sched_info:
24037 + return vc_sched_info(vxi, data);
24039 + case VCMD_add_dlimit:
24040 + return __COMPAT(vc_add_dlimit, id, data, compat);
24041 + case VCMD_rem_dlimit:
24042 + return __COMPAT(vc_rem_dlimit, id, data, compat);
24043 + case VCMD_set_dlimit:
24044 + return __COMPAT(vc_set_dlimit, id, data, compat);
24045 + case VCMD_get_dlimit:
24046 + return __COMPAT(vc_get_dlimit, id, data, compat);
24048 + case VCMD_ctx_kill:
24049 + return vc_ctx_kill(vxi, data);
24051 + case VCMD_wait_exit:
24052 + return vc_wait_exit(vxi, data);
24054 + case VCMD_get_iattr:
24055 + return __COMPAT_NO_ID(vc_get_iattr, data, compat);
24056 + case VCMD_set_iattr:
24057 + return __COMPAT_NO_ID(vc_set_iattr, data, compat);
24059 + case VCMD_fget_iattr:
24060 + return vc_fget_iattr(id, data);
24061 + case VCMD_fset_iattr:
24062 + return vc_fset_iattr(id, data);
24064 + case VCMD_enter_space_v0:
24065 + return vc_enter_space_v1(vxi, NULL);
24066 + case VCMD_enter_space_v1:
24067 + return vc_enter_space_v1(vxi, data);
24068 + /* this is version 2 */
24069 + case VCMD_enter_space:
24070 + return vc_enter_space(vxi, data);
24072 + case VCMD_ctx_create_v0:
24073 + return vc_ctx_create(id, NULL);
24074 + case VCMD_ctx_create:
24075 + return vc_ctx_create(id, data);
24076 + case VCMD_ctx_migrate_v0:
24077 + return vc_ctx_migrate(vxi, NULL);
24078 + case VCMD_ctx_migrate:
24079 + return vc_ctx_migrate(vxi, data);
24081 + case VCMD_net_create_v0:
24082 + return vc_net_create(id, NULL);
24083 + case VCMD_net_create:
24084 + return vc_net_create(id, data);
24085 + case VCMD_net_migrate:
24086 + return vc_net_migrate(nxi, data);
24088 + case VCMD_tag_migrate:
24089 + return vc_tag_migrate(id);
24091 + case VCMD_net_add:
24092 + return vc_net_add(nxi, data);
24093 + case VCMD_net_remove:
24094 + return vc_net_remove(nxi, data);
24096 + case VCMD_net_add_ipv4:
24097 + return vc_net_add_ipv4(nxi, data);
24098 + case VCMD_net_remove_ipv4:
24099 + return vc_net_remove_ipv4(nxi, data);
24100 +#ifdef CONFIG_IPV6
24101 + case VCMD_net_add_ipv6:
24102 + return vc_net_add_ipv6(nxi, data);
24103 + case VCMD_net_remove_ipv6:
24104 + return vc_net_remove_ipv6(nxi, data);
24106 +/* case VCMD_add_match_ipv4:
24107 + return vc_add_match_ipv4(nxi, data);
24108 + case VCMD_get_match_ipv4:
24109 + return vc_get_match_ipv4(nxi, data);
24110 +#ifdef CONFIG_IPV6
24111 + case VCMD_add_match_ipv6:
24112 + return vc_add_match_ipv6(nxi, data);
24113 + case VCMD_get_match_ipv6:
24114 + return vc_get_match_ipv6(nxi, data);
24117 +#ifdef CONFIG_VSERVER_DEVICE
24118 + case VCMD_set_mapping:
24119 + return __COMPAT(vc_set_mapping, vxi, data, compat);
24120 + case VCMD_unset_mapping:
24121 + return __COMPAT(vc_unset_mapping, vxi, data, compat);
24123 +#ifdef CONFIG_VSERVER_HISTORY
24124 + case VCMD_dump_history:
24125 + return vc_dump_history(id);
24126 + case VCMD_read_history:
24127 + return __COMPAT(vc_read_history, id, data, compat);
24129 +#ifdef CONFIG_VSERVER_MONITOR
24130 + case VCMD_read_monitor:
24131 + return __COMPAT(vc_read_monitor, id, data, compat);
24134 + vxwprintk_task(1, "unimplemented VCMD_%02d_%d[%d]",
24135 + VC_CATEGORY(cmd), VC_COMMAND(cmd), VC_VERSION(cmd));
24141 +#define __VCMD(vcmd, _perm, _args, _flags) \
24142 + case VCMD_ ## vcmd: perm = _perm; \
24143 + args = _args; flags = _flags; break
24146 +#define VCA_NONE 0x00
24147 +#define VCA_VXI 0x01
24148 +#define VCA_NXI 0x02
24150 +#define VCF_NONE 0x00
24151 +#define VCF_INFO 0x01
24152 +#define VCF_ADMIN 0x02
24153 +#define VCF_ARES 0x06 /* includes admin */
24154 +#define VCF_SETUP 0x08
24156 +#define VCF_ZIDOK 0x10 /* zero id okay */
24160 +long do_vserver(uint32_t cmd, uint32_t id, void __user *data, int compat)
24163 + int permit = -1, state = 0;
24164 + int perm = -1, args = 0, flags = 0;
24165 + struct vx_info *vxi = NULL;
24166 + struct nx_info *nxi = NULL;
24169 + /* unpriviledged commands */
24170 + __VCMD(get_version, 0, VCA_NONE, 0);
24171 + __VCMD(get_vci, 0, VCA_NONE, 0);
24172 + __VCMD(get_rlimit_mask, 0, VCA_NONE, 0);
24173 + __VCMD(get_space_mask_v0,0, VCA_NONE, 0);
24174 + __VCMD(get_space_mask, 0, VCA_NONE, 0);
24175 + __VCMD(get_space_default,0, VCA_NONE, 0);
24177 + /* info commands */
24178 + __VCMD(task_xid, 2, VCA_NONE, 0);
24179 + __VCMD(reset_hits, 2, VCA_VXI, 0);
24180 + __VCMD(reset_minmax, 2, VCA_VXI, 0);
24181 + __VCMD(vx_info, 3, VCA_VXI, VCF_INFO);
24182 + __VCMD(get_bcaps, 3, VCA_VXI, VCF_INFO);
24183 + __VCMD(get_ccaps, 3, VCA_VXI, VCF_INFO);
24184 + __VCMD(get_cflags, 3, VCA_VXI, VCF_INFO);
24185 + __VCMD(get_badness, 3, VCA_VXI, VCF_INFO);
24186 + __VCMD(get_vhi_name, 3, VCA_VXI, VCF_INFO);
24187 + __VCMD(get_rlimit, 3, VCA_VXI, VCF_INFO);
24189 + __VCMD(ctx_stat, 3, VCA_VXI, VCF_INFO);
24190 + __VCMD(virt_stat, 3, VCA_VXI, VCF_INFO);
24191 + __VCMD(sock_stat, 3, VCA_VXI, VCF_INFO);
24192 + __VCMD(rlimit_stat, 3, VCA_VXI, VCF_INFO);
24194 + __VCMD(task_nid, 2, VCA_NONE, 0);
24195 + __VCMD(nx_info, 3, VCA_NXI, VCF_INFO);
24196 + __VCMD(get_ncaps, 3, VCA_NXI, VCF_INFO);
24197 + __VCMD(get_nflags, 3, VCA_NXI, VCF_INFO);
24199 + __VCMD(task_tag, 2, VCA_NONE, 0);
24201 + __VCMD(get_iattr, 2, VCA_NONE, 0);
24202 + __VCMD(fget_iattr, 2, VCA_NONE, 0);
24203 + __VCMD(get_dlimit, 3, VCA_NONE, VCF_INFO);
24204 + __VCMD(get_sched, 3, VCA_VXI, VCF_INFO);
24205 + __VCMD(sched_info, 3, VCA_VXI, VCF_INFO | VCF_ZIDOK);
24207 + /* lower admin commands */
24208 + __VCMD(wait_exit, 4, VCA_VXI, VCF_INFO);
24209 + __VCMD(ctx_create_v0, 5, VCA_NONE, 0);
24210 + __VCMD(ctx_create, 5, VCA_NONE, 0);
24211 + __VCMD(ctx_migrate_v0, 5, VCA_VXI, VCF_ADMIN);
24212 + __VCMD(ctx_migrate, 5, VCA_VXI, VCF_ADMIN);
24213 + __VCMD(enter_space_v0, 5, VCA_VXI, VCF_ADMIN);
24214 + __VCMD(enter_space_v1, 5, VCA_VXI, VCF_ADMIN);
24215 + __VCMD(enter_space, 5, VCA_VXI, VCF_ADMIN);
24217 + __VCMD(net_create_v0, 5, VCA_NONE, 0);
24218 + __VCMD(net_create, 5, VCA_NONE, 0);
24219 + __VCMD(net_migrate, 5, VCA_NXI, VCF_ADMIN);
24221 + __VCMD(tag_migrate, 5, VCA_NONE, VCF_ADMIN);
24223 + /* higher admin commands */
24224 + __VCMD(ctx_kill, 6, VCA_VXI, VCF_ARES);
24225 + __VCMD(set_space_v1, 7, VCA_VXI, VCF_ARES | VCF_SETUP);
24226 + __VCMD(set_space, 7, VCA_VXI, VCF_ARES | VCF_SETUP);
24228 + __VCMD(set_ccaps, 7, VCA_VXI, VCF_ARES | VCF_SETUP);
24229 + __VCMD(set_bcaps, 7, VCA_VXI, VCF_ARES | VCF_SETUP);
24230 + __VCMD(set_cflags, 7, VCA_VXI, VCF_ARES | VCF_SETUP);
24231 + __VCMD(set_badness, 7, VCA_VXI, VCF_ARES | VCF_SETUP);
24233 + __VCMD(set_vhi_name, 7, VCA_VXI, VCF_ARES | VCF_SETUP);
24234 + __VCMD(set_rlimit, 7, VCA_VXI, VCF_ARES | VCF_SETUP);
24235 + __VCMD(set_sched, 7, VCA_VXI, VCF_ARES | VCF_SETUP);
24236 + __VCMD(set_sched_v4, 7, VCA_VXI, VCF_ARES | VCF_SETUP);
24238 + __VCMD(set_ncaps, 7, VCA_NXI, VCF_ARES | VCF_SETUP);
24239 + __VCMD(set_nflags, 7, VCA_NXI, VCF_ARES | VCF_SETUP);
24240 + __VCMD(net_add, 8, VCA_NXI, VCF_ARES | VCF_SETUP);
24241 + __VCMD(net_remove, 8, VCA_NXI, VCF_ARES | VCF_SETUP);
24242 + __VCMD(net_add_ipv4, 8, VCA_NXI, VCF_ARES | VCF_SETUP);
24243 + __VCMD(net_remove_ipv4, 8, VCA_NXI, VCF_ARES | VCF_SETUP);
24244 +#ifdef CONFIG_IPV6
24245 + __VCMD(net_add_ipv6, 8, VCA_NXI, VCF_ARES | VCF_SETUP);
24246 + __VCMD(net_remove_ipv6, 8, VCA_NXI, VCF_ARES | VCF_SETUP);
24248 + __VCMD(set_iattr, 7, VCA_NONE, 0);
24249 + __VCMD(fset_iattr, 7, VCA_NONE, 0);
24250 + __VCMD(set_dlimit, 7, VCA_NONE, VCF_ARES);
24251 + __VCMD(add_dlimit, 8, VCA_NONE, VCF_ARES);
24252 + __VCMD(rem_dlimit, 8, VCA_NONE, VCF_ARES);
24254 +#ifdef CONFIG_VSERVER_DEVICE
24255 + __VCMD(set_mapping, 8, VCA_VXI, VCF_ARES|VCF_ZIDOK);
24256 + __VCMD(unset_mapping, 8, VCA_VXI, VCF_ARES|VCF_ZIDOK);
24258 + /* debug level admin commands */
24259 +#ifdef CONFIG_VSERVER_HISTORY
24260 + __VCMD(dump_history, 9, VCA_NONE, 0);
24261 + __VCMD(read_history, 9, VCA_NONE, 0);
24263 +#ifdef CONFIG_VSERVER_MONITOR
24264 + __VCMD(read_monitor, 9, VCA_NONE, 0);
24271 + vxdprintk(VXD_CBIT(switch, 0),
24272 + "vc: VCMD_%02d_%d[%d], %d,%p [%d,%d,%x,%x]",
24273 + VC_CATEGORY(cmd), VC_COMMAND(cmd),
24274 + VC_VERSION(cmd), id, data, compat,
24275 + perm, args, flags);
24282 + if (!capable(CAP_CONTEXT))
24286 + /* moved here from the individual commands */
24288 + if ((perm > 1) && !capable(CAP_SYS_ADMIN))
24292 + /* vcmd involves resource management */
24294 + if ((flags & VCF_ARES) && !capable(CAP_SYS_RESOURCE))
24298 + /* various legacy exceptions */
24300 + /* will go away when spectator is a cap */
24301 + case VCMD_ctx_migrate_v0:
24302 + case VCMD_ctx_migrate:
24304 + current->xid = 1;
24310 + /* will go away when spectator is a cap */
24311 + case VCMD_net_migrate:
24313 + current->nid = 1;
24320 + /* vcmds are fine by default */
24323 + /* admin type vcmds require admin ... */
24324 + if (flags & VCF_ADMIN)
24325 + permit = vx_check(0, VS_ADMIN) ? 1 : 0;
24327 + /* ... but setup type vcmds override that */
24328 + if (!permit && (flags & VCF_SETUP))
24329 + permit = vx_flags(VXF_STATE_SETUP, 0) ? 2 : 0;
24337 + if (!id && (flags & VCF_ZIDOK))
24341 + if (args & VCA_VXI) {
24342 + vxi = lookup_vx_info(id);
24346 + if ((flags & VCF_ADMIN) &&
24347 + /* special case kill for shutdown */
24348 + (cmd != VCMD_ctx_kill) &&
24349 + /* can context be administrated? */
24350 + !vx_info_flags(vxi, VXF_STATE_ADMIN, 0)) {
24356 + if (args & VCA_NXI) {
24357 + nxi = lookup_nx_info(id);
24361 + if ((flags & VCF_ADMIN) &&
24362 + /* can context be administrated? */
24363 + !nx_info_flags(nxi, NXF_STATE_ADMIN, 0)) {
24370 + ret = do_vcmd(cmd, id, vxi, nxi, data, compat);
24373 + if ((args & VCA_NXI) && nxi)
24374 + put_nx_info(nxi);
24376 + if ((args & VCA_VXI) && vxi)
24377 + put_vx_info(vxi);
24379 + vxdprintk(VXD_CBIT(switch, 1),
24380 + "vc: VCMD_%02d_%d[%d] = %08lx(%ld) [%d,%d]",
24381 + VC_CATEGORY(cmd), VC_COMMAND(cmd),
24382 + VC_VERSION(cmd), ret, ret, state, permit);
24387 +sys_vserver(uint32_t cmd, uint32_t id, void __user *data)
24389 + return do_vserver(cmd, id, data, 0);
24392 +#ifdef CONFIG_COMPAT
24395 +sys32_vserver(uint32_t cmd, uint32_t id, void __user *data)
24397 + return do_vserver(cmd, id, data, 1);
24400 +#endif /* CONFIG_COMPAT */
24401 diff -NurpP --minimal linux-2.6.36/kernel/vserver/sysctl.c linux-2.6.36-vs2.3.0.36.33/kernel/vserver/sysctl.c
24402 --- linux-2.6.36/kernel/vserver/sysctl.c 1970-01-01 01:00:00.000000000 +0100
24403 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/sysctl.c 2010-10-21 13:09:36.000000000 +0200
24406 + * kernel/vserver/sysctl.c
24408 + * Virtual Context Support
24410 + * Copyright (C) 2004-2007 Herbert Pötzl
24412 + * V0.01 basic structure
24416 +#include <linux/module.h>
24417 +#include <linux/ctype.h>
24418 +#include <linux/sysctl.h>
24419 +#include <linux/parser.h>
24420 +#include <asm/uaccess.h>
24423 + CTL_DEBUG_ERROR = 0,
24424 + CTL_DEBUG_SWITCH = 1,
24439 +unsigned int vx_debug_switch = 0;
24440 +unsigned int vx_debug_xid = 0;
24441 +unsigned int vx_debug_nid = 0;
24442 +unsigned int vx_debug_tag = 0;
24443 +unsigned int vx_debug_net = 0;
24444 +unsigned int vx_debug_limit = 0;
24445 +unsigned int vx_debug_cres = 0;
24446 +unsigned int vx_debug_dlim = 0;
24447 +unsigned int vx_debug_quota = 0;
24448 +unsigned int vx_debug_cvirt = 0;
24449 +unsigned int vx_debug_space = 0;
24450 +unsigned int vx_debug_misc = 0;
24453 +static struct ctl_table_header *vserver_table_header;
24454 +static ctl_table vserver_root_table[];
24457 +void vserver_register_sysctl(void)
24459 + if (!vserver_table_header) {
24460 + vserver_table_header = register_sysctl_table(vserver_root_table);
24465 +void vserver_unregister_sysctl(void)
24467 + if (vserver_table_header) {
24468 + unregister_sysctl_table(vserver_table_header);
24469 + vserver_table_header = NULL;
24474 +static int proc_dodebug(ctl_table *table, int write,
24475 + void __user *buffer, size_t *lenp, loff_t *ppos)
24477 + char tmpbuf[20], *p, c;
24478 + unsigned int value;
24479 + size_t left, len;
24481 + if ((*ppos && !write) || !*lenp) {
24489 + if (!access_ok(VERIFY_READ, buffer, left))
24491 + p = (char *)buffer;
24492 + while (left && __get_user(c, p) >= 0 && isspace(c))
24497 + if (left > sizeof(tmpbuf) - 1)
24499 + if (copy_from_user(tmpbuf, p, left))
24501 + tmpbuf[left] = '\0';
24503 + for (p = tmpbuf, value = 0; '0' <= *p && *p <= '9'; p++, left--)
24504 + value = 10 * value + (*p - '0');
24505 + if (*p && !isspace(*p))
24507 + while (left && isspace(*p))
24509 + *(unsigned int *)table->data = value;
24511 + if (!access_ok(VERIFY_WRITE, buffer, left))
24513 + len = sprintf(tmpbuf, "%d", *(unsigned int *)table->data);
24516 + if (__copy_to_user(buffer, tmpbuf, len))
24518 + if ((left -= len) > 0) {
24519 + if (put_user('\n', (char *)buffer + len))
24533 +#define CTL_ENTRY(ctl, name) \
24535 + .procname = #name, \
24536 + .data = &vx_ ## name, \
24537 + .maxlen = sizeof(int), \
24539 + .proc_handler = &proc_dodebug, \
24540 + .extra1 = &zero, \
24541 + .extra2 = &zero, \
24544 +static ctl_table vserver_debug_table[] = {
24545 + CTL_ENTRY(CTL_DEBUG_SWITCH, debug_switch),
24546 + CTL_ENTRY(CTL_DEBUG_XID, debug_xid),
24547 + CTL_ENTRY(CTL_DEBUG_NID, debug_nid),
24548 + CTL_ENTRY(CTL_DEBUG_TAG, debug_tag),
24549 + CTL_ENTRY(CTL_DEBUG_NET, debug_net),
24550 + CTL_ENTRY(CTL_DEBUG_LIMIT, debug_limit),
24551 + CTL_ENTRY(CTL_DEBUG_CRES, debug_cres),
24552 + CTL_ENTRY(CTL_DEBUG_DLIM, debug_dlim),
24553 + CTL_ENTRY(CTL_DEBUG_QUOTA, debug_quota),
24554 + CTL_ENTRY(CTL_DEBUG_CVIRT, debug_cvirt),
24555 + CTL_ENTRY(CTL_DEBUG_SPACE, debug_space),
24556 + CTL_ENTRY(CTL_DEBUG_MISC, debug_misc),
24560 +static ctl_table vserver_root_table[] = {
24562 + .procname = "vserver",
24564 + .child = vserver_debug_table
24570 +static match_table_t tokens = {
24571 + { CTL_DEBUG_SWITCH, "switch=%x" },
24572 + { CTL_DEBUG_XID, "xid=%x" },
24573 + { CTL_DEBUG_NID, "nid=%x" },
24574 + { CTL_DEBUG_TAG, "tag=%x" },
24575 + { CTL_DEBUG_NET, "net=%x" },
24576 + { CTL_DEBUG_LIMIT, "limit=%x" },
24577 + { CTL_DEBUG_CRES, "cres=%x" },
24578 + { CTL_DEBUG_DLIM, "dlim=%x" },
24579 + { CTL_DEBUG_QUOTA, "quota=%x" },
24580 + { CTL_DEBUG_CVIRT, "cvirt=%x" },
24581 + { CTL_DEBUG_SPACE, "space=%x" },
24582 + { CTL_DEBUG_MISC, "misc=%x" },
24583 + { CTL_DEBUG_ERROR, NULL }
24586 +#define HANDLE_CASE(id, name, val) \
24587 + case CTL_DEBUG_ ## id: \
24588 + vx_debug_ ## name = val; \
24589 + printk("vs_debug_" #name "=0x%x\n", val); \
24593 +static int __init vs_debug_setup(char *str)
24598 + printk("vs_debug_setup(%s)\n", str);
24599 + while ((p = strsep(&str, ",")) != NULL) {
24600 + substring_t args[MAX_OPT_ARGS];
24601 + unsigned int value;
24606 + token = match_token(p, tokens, args);
24607 + value = (token > 0) ? simple_strtoul(args[0].from, NULL, 0) : 0;
24610 + HANDLE_CASE(SWITCH, switch, value);
24611 + HANDLE_CASE(XID, xid, value);
24612 + HANDLE_CASE(NID, nid, value);
24613 + HANDLE_CASE(TAG, tag, value);
24614 + HANDLE_CASE(NET, net, value);
24615 + HANDLE_CASE(LIMIT, limit, value);
24616 + HANDLE_CASE(CRES, cres, value);
24617 + HANDLE_CASE(DLIM, dlim, value);
24618 + HANDLE_CASE(QUOTA, quota, value);
24619 + HANDLE_CASE(CVIRT, cvirt, value);
24620 + HANDLE_CASE(SPACE, space, value);
24621 + HANDLE_CASE(MISC, misc, value);
24630 +__setup("vsdebug=", vs_debug_setup);
24634 +EXPORT_SYMBOL_GPL(vx_debug_switch);
24635 +EXPORT_SYMBOL_GPL(vx_debug_xid);
24636 +EXPORT_SYMBOL_GPL(vx_debug_nid);
24637 +EXPORT_SYMBOL_GPL(vx_debug_net);
24638 +EXPORT_SYMBOL_GPL(vx_debug_limit);
24639 +EXPORT_SYMBOL_GPL(vx_debug_cres);
24640 +EXPORT_SYMBOL_GPL(vx_debug_dlim);
24641 +EXPORT_SYMBOL_GPL(vx_debug_quota);
24642 +EXPORT_SYMBOL_GPL(vx_debug_cvirt);
24643 +EXPORT_SYMBOL_GPL(vx_debug_space);
24644 +EXPORT_SYMBOL_GPL(vx_debug_misc);
24646 diff -NurpP --minimal linux-2.6.36/kernel/vserver/tag.c linux-2.6.36-vs2.3.0.36.33/kernel/vserver/tag.c
24647 --- linux-2.6.36/kernel/vserver/tag.c 1970-01-01 01:00:00.000000000 +0100
24648 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/tag.c 2010-10-21 13:09:36.000000000 +0200
24651 + * linux/kernel/vserver/tag.c
24653 + * Virtual Server: Shallow Tag Space
24655 + * Copyright (C) 2007 Herbert Pötzl
24657 + * V0.01 basic implementation
24661 +#include <linux/sched.h>
24662 +#include <linux/vserver/debug.h>
24663 +#include <linux/vs_pid.h>
24664 +#include <linux/vs_tag.h>
24666 +#include <linux/vserver/tag_cmd.h>
24669 +int dx_migrate_task(struct task_struct *p, tag_t tag)
24674 + vxdprintk(VXD_CBIT(tag, 5),
24675 + "dx_migrate_task(%p[#%d],#%d)", p, p->tag, tag);
24681 + vxdprintk(VXD_CBIT(tag, 5),
24682 + "moved task %p into [#%d]", p, tag);
24686 +/* vserver syscall commands below here */
24688 +/* taks xid and vx_info functions */
24691 +int vc_task_tag(uint32_t id)
24696 + struct task_struct *tsk;
24697 + read_lock(&tasklist_lock);
24698 + tsk = find_task_by_real_pid(id);
24699 + tag = (tsk) ? tsk->tag : -ESRCH;
24700 + read_unlock(&tasklist_lock);
24702 + tag = dx_current_tag();
24707 +int vc_tag_migrate(uint32_t tag)
24709 + return dx_migrate_task(current, tag & 0xFFFF);
24713 diff -NurpP --minimal linux-2.6.36/kernel/vserver/vci_config.h linux-2.6.36-vs2.3.0.36.33/kernel/vserver/vci_config.h
24714 --- linux-2.6.36/kernel/vserver/vci_config.h 1970-01-01 01:00:00.000000000 +0100
24715 +++ linux-2.6.36-vs2.3.0.36.33/kernel/vserver/vci_config.h 2010-10-21 13:09:36.000000000 +0200
24718 +/* interface version */
24720 +#define VCI_VERSION 0x00020306
24724 + VCI_KCBIT_NO_DYNAMIC = 0,
24726 + VCI_KCBIT_PROC_SECURE = 4,
24727 + VCI_KCBIT_HARDCPU = 5,
24728 + VCI_KCBIT_IDLELIMIT = 6,
24729 + VCI_KCBIT_IDLETIME = 7,
24731 + VCI_KCBIT_COWBL = 8,
24732 + VCI_KCBIT_FULLCOWBL = 9,
24733 + VCI_KCBIT_SPACES = 10,
24734 + VCI_KCBIT_NETV2 = 11,
24735 + VCI_KCBIT_MEMCG = 12,
24737 + VCI_KCBIT_DEBUG = 16,
24738 + VCI_KCBIT_HISTORY = 20,
24739 + VCI_KCBIT_TAGGED = 24,
24740 + VCI_KCBIT_PPTAG = 28,
24742 + VCI_KCBIT_MORE = 31,
24746 +static inline uint32_t vci_kernel_config(void)
24749 + (1 << VCI_KCBIT_NO_DYNAMIC) |
24751 + /* configured features */
24752 +#ifdef CONFIG_VSERVER_PROC_SECURE
24753 + (1 << VCI_KCBIT_PROC_SECURE) |
24755 +#ifdef CONFIG_VSERVER_HARDCPU
24756 + (1 << VCI_KCBIT_HARDCPU) |
24758 +#ifdef CONFIG_VSERVER_IDLELIMIT
24759 + (1 << VCI_KCBIT_IDLELIMIT) |
24761 +#ifdef CONFIG_VSERVER_IDLETIME
24762 + (1 << VCI_KCBIT_IDLETIME) |
24764 +#ifdef CONFIG_VSERVER_COWBL
24765 + (1 << VCI_KCBIT_COWBL) |
24766 + (1 << VCI_KCBIT_FULLCOWBL) |
24768 + (1 << VCI_KCBIT_SPACES) |
24769 + (1 << VCI_KCBIT_NETV2) |
24770 +#ifdef CONFIG_CGROUP_MEM_RES_CTLR
24771 + (1 << VCI_KCBIT_MEMCG) |
24774 + /* debug options */
24775 +#ifdef CONFIG_VSERVER_DEBUG
24776 + (1 << VCI_KCBIT_DEBUG) |
24778 +#ifdef CONFIG_VSERVER_HISTORY
24779 + (1 << VCI_KCBIT_HISTORY) |
24782 + /* inode context tagging */
24783 +#if defined(CONFIG_TAGGING_NONE)
24784 + (0 << VCI_KCBIT_TAGGED) |
24785 +#elif defined(CONFIG_TAGGING_UID16)
24786 + (1 << VCI_KCBIT_TAGGED) |
24787 +#elif defined(CONFIG_TAGGING_GID16)
24788 + (2 << VCI_KCBIT_TAGGED) |
24789 +#elif defined(CONFIG_TAGGING_ID24)
24790 + (3 << VCI_KCBIT_TAGGED) |
24791 +#elif defined(CONFIG_TAGGING_INTERN)
24792 + (4 << VCI_KCBIT_TAGGED) |
24793 +#elif defined(CONFIG_TAGGING_RUNTIME)
24794 + (5 << VCI_KCBIT_TAGGED) |
24796 + (7 << VCI_KCBIT_TAGGED) |
24798 + (1 << VCI_KCBIT_PPTAG) |
24802 diff -NurpP --minimal linux-2.6.36/mm/filemap_xip.c linux-2.6.36-vs2.3.0.36.33/mm/filemap_xip.c
24803 --- linux-2.6.36/mm/filemap_xip.c 2010-07-07 18:31:58.000000000 +0200
24804 +++ linux-2.6.36-vs2.3.0.36.33/mm/filemap_xip.c 2010-10-21 13:09:36.000000000 +0200
24806 #include <linux/seqlock.h>
24807 #include <linux/mutex.h>
24808 #include <linux/gfp.h>
24809 +#include <linux/vs_memory.h>
24810 #include <asm/tlbflush.h>
24811 #include <asm/io.h>
24813 diff -NurpP --minimal linux-2.6.36/mm/fremap.c linux-2.6.36-vs2.3.0.36.33/mm/fremap.c
24814 --- linux-2.6.36/mm/fremap.c 2010-10-21 13:07:57.000000000 +0200
24815 +++ linux-2.6.36-vs2.3.0.36.33/mm/fremap.c 2010-10-21 13:09:36.000000000 +0200
24817 #include <linux/module.h>
24818 #include <linux/syscalls.h>
24819 #include <linux/mmu_notifier.h>
24820 +#include <linux/vs_memory.h>
24822 #include <asm/mmu_context.h>
24823 #include <asm/cacheflush.h>
24824 diff -NurpP --minimal linux-2.6.36/mm/hugetlb.c linux-2.6.36-vs2.3.0.36.33/mm/hugetlb.c
24825 --- linux-2.6.36/mm/hugetlb.c 2010-10-21 13:07:57.000000000 +0200
24826 +++ linux-2.6.36-vs2.3.0.36.33/mm/hugetlb.c 2010-10-21 13:09:36.000000000 +0200
24829 #include <linux/hugetlb.h>
24830 #include <linux/node.h>
24831 +#include <linux/vs_memory.h>
24832 #include "internal.h"
24834 const unsigned long hugetlb_zero = 0, hugetlb_infinity = ~0UL;
24835 diff -NurpP --minimal linux-2.6.36/mm/memcontrol.c linux-2.6.36-vs2.3.0.36.33/mm/memcontrol.c
24836 --- linux-2.6.36/mm/memcontrol.c 2010-10-21 13:07:57.000000000 +0200
24837 +++ linux-2.6.36-vs2.3.0.36.33/mm/memcontrol.c 2010-10-21 13:09:36.000000000 +0200
24838 @@ -638,6 +638,31 @@ struct mem_cgroup *mem_cgroup_from_task(
24839 struct mem_cgroup, css);
24842 +u64 mem_cgroup_res_read_u64(struct mem_cgroup *mem, int member)
24844 + return res_counter_read_u64(&mem->res, member);
24847 +u64 mem_cgroup_memsw_read_u64(struct mem_cgroup *mem, int member)
24849 + return res_counter_read_u64(&mem->memsw, member);
24852 +s64 mem_cgroup_stat_read_cache(struct mem_cgroup *mem)
24854 + return mem_cgroup_read_stat(mem, MEM_CGROUP_STAT_CACHE);
24857 +s64 mem_cgroup_stat_read_anon(struct mem_cgroup *mem)
24859 + return mem_cgroup_read_stat(mem, MEM_CGROUP_STAT_RSS);
24862 +s64 mem_cgroup_stat_read_mapped(struct mem_cgroup *mem)
24864 + return mem_cgroup_read_stat(mem, MEM_CGROUP_STAT_FILE_MAPPED);
24867 static struct mem_cgroup *try_get_mem_cgroup_from_mm(struct mm_struct *mm)
24869 struct mem_cgroup *mem = NULL;
24870 diff -NurpP --minimal linux-2.6.36/mm/memory.c linux-2.6.36-vs2.3.0.36.33/mm/memory.c
24871 --- linux-2.6.36/mm/memory.c 2010-10-21 13:07:57.000000000 +0200
24872 +++ linux-2.6.36-vs2.3.0.36.33/mm/memory.c 2010-10-21 13:09:36.000000000 +0200
24873 @@ -3145,6 +3145,7 @@ static inline int handle_pte_fault(struc
24877 + int ret = 0, type = VXPT_UNKNOWN;
24880 if (!pte_present(entry)) {
24881 @@ -3169,9 +3170,12 @@ static inline int handle_pte_fault(struc
24882 if (unlikely(!pte_same(*pte, entry)))
24884 if (flags & FAULT_FLAG_WRITE) {
24885 - if (!pte_write(entry))
24886 - return do_wp_page(mm, vma, address,
24887 + if (!pte_write(entry)) {
24888 + ret = do_wp_page(mm, vma, address,
24889 pte, pmd, ptl, entry);
24890 + type = VXPT_WRITE;
24893 entry = pte_mkdirty(entry);
24895 entry = pte_mkyoung(entry);
24896 @@ -3189,7 +3193,10 @@ static inline int handle_pte_fault(struc
24899 pte_unmap_unlock(pte, ptl);
24903 + vx_page_fault(mm, vma, type, ret);
24908 diff -NurpP --minimal linux-2.6.36/mm/mremap.c linux-2.6.36-vs2.3.0.36.33/mm/mremap.c
24909 --- linux-2.6.36/mm/mremap.c 2010-07-07 18:31:58.000000000 +0200
24910 +++ linux-2.6.36-vs2.3.0.36.33/mm/mremap.c 2010-10-21 13:09:36.000000000 +0200
24912 #include <linux/security.h>
24913 #include <linux/syscalls.h>
24914 #include <linux/mmu_notifier.h>
24915 +#include <linux/vs_memory.h>
24917 #include <asm/uaccess.h>
24918 #include <asm/cacheflush.h>
24919 diff -NurpP --minimal linux-2.6.36/mm/oom_kill.c linux-2.6.36-vs2.3.0.36.33/mm/oom_kill.c
24920 --- linux-2.6.36/mm/oom_kill.c 2010-10-21 13:07:57.000000000 +0200
24921 +++ linux-2.6.36-vs2.3.0.36.33/mm/oom_kill.c 2010-10-21 14:57:29.000000000 +0200
24923 #include <linux/memcontrol.h>
24924 #include <linux/mempolicy.h>
24925 #include <linux/security.h>
24926 +#include <linux/reboot.h>
24927 +#include <linux/vs_memory.h>
24928 +#include <linux/vs_context.h>
24930 int sysctl_panic_on_oom;
24931 int sysctl_oom_kill_allocating_task;
24932 @@ -124,11 +127,18 @@ struct task_struct *find_lock_task_mm(st
24933 static bool oom_unkillable_task(struct task_struct *p,
24934 const struct mem_cgroup *mem, const nodemask_t *nodemask)
24936 - if (is_global_init(p))
24937 + unsigned xid = vx_current_xid();
24939 + /* skip the init task, global and per guest */
24940 + if (task_is_init(p))
24942 if (p->flags & PF_KTHREAD)
24945 + /* skip other guest and host processes if oom in guest */
24946 + if (xid && vx_task_xid(p) != xid)
24949 /* When mem_cgroup_out_of_memory() and p is not member of the group */
24950 if (mem && !task_in_mem_cgroup(p, mem))
24952 @@ -209,6 +219,18 @@ unsigned int oom_badness(struct task_str
24953 points += p->signal->oom_score_adj;
24956 + * add points for context badness and
24957 + * reduce badness for processes belonging to
24958 + * a different context
24961 + points += vx_badness(p, p->mm);
24963 + if ((vx_current_xid() > 1) &&
24964 + vx_current_xid() != vx_task_xid(p))
24968 * Never return 0 for an eligible task that may be killed since it's
24969 * possible that no single user task uses more than 0.1% of memory and
24970 * no single admin tasks uses more than 3.0%.
24971 @@ -407,8 +429,8 @@ static int oom_kill_task(struct task_str
24975 - pr_err("Killed process %d (%s) total-vm:%lukB, anon-rss:%lukB, file-rss:%lukB\n",
24976 - task_pid_nr(p), p->comm, K(p->mm->total_vm),
24977 + pr_err("Killed process %d:#%u (%s) total-vm:%lukB, anon-rss:%lukB, file-rss:%lukB\n",
24978 + task_pid_nr(p), p->xid, p->comm, K(p->mm->total_vm),
24979 K(get_mm_counter(p->mm, MM_ANONPAGES)),
24980 K(get_mm_counter(p->mm, MM_FILEPAGES)));
24982 @@ -452,8 +474,8 @@ static int oom_kill_process(struct task_
24986 - pr_err("%s: Kill process %d (%s) score %d or sacrifice child\n",
24987 - message, task_pid_nr(p), p->comm, points);
24988 + pr_err("%s: Kill process %d:#%u (%s) score %d or sacrifice child\n",
24989 + message, task_pid_nr(p), p->xid, p->comm, points);
24993 @@ -542,6 +564,8 @@ int unregister_oom_notifier(struct notif
24995 EXPORT_SYMBOL_GPL(unregister_oom_notifier);
24997 +long vs_oom_action(unsigned int);
25000 * Try to acquire the OOM killer lock for the zones in zonelist. Returns zero
25001 * if a parallel OOM killing is already taking place that includes a zone in
25002 @@ -701,7 +725,12 @@ retry:
25004 dump_header(NULL, gfp_mask, order, NULL, mpol_mask);
25005 read_unlock(&tasklist_lock);
25006 - panic("Out of memory and no killable processes...\n");
25008 + /* avoid panic for guest OOM */
25009 + if (current->xid)
25010 + vs_oom_action(LINUX_REBOOT_CMD_OOM);
25012 + panic("Out of memory and no killable processes...\n");
25015 if (oom_kill_process(p, gfp_mask, order, points, totalpages, NULL,
25016 diff -NurpP --minimal linux-2.6.36/mm/page_alloc.c linux-2.6.36-vs2.3.0.36.33/mm/page_alloc.c
25017 --- linux-2.6.36/mm/page_alloc.c 2010-10-21 13:07:57.000000000 +0200
25018 +++ linux-2.6.36-vs2.3.0.36.33/mm/page_alloc.c 2010-10-21 13:09:36.000000000 +0200
25020 #include <linux/compaction.h>
25021 #include <trace/events/kmem.h>
25022 #include <linux/ftrace_event.h>
25023 +#include <linux/vs_base.h>
25024 +#include <linux/vs_limit.h>
25026 #include <asm/tlbflush.h>
25027 #include <asm/div64.h>
25028 @@ -2329,6 +2331,9 @@ void si_meminfo(struct sysinfo *val)
25029 val->totalhigh = totalhigh_pages;
25030 val->freehigh = nr_free_highpages();
25031 val->mem_unit = PAGE_SIZE;
25033 + if (vx_flags(VXF_VIRT_MEM, 0))
25034 + vx_vsi_meminfo(val);
25037 EXPORT_SYMBOL(si_meminfo);
25038 @@ -2349,6 +2354,9 @@ void si_meminfo_node(struct sysinfo *val
25041 val->mem_unit = PAGE_SIZE;
25043 + if (vx_flags(VXF_VIRT_MEM, 0))
25044 + vx_vsi_meminfo(val);
25048 diff -NurpP --minimal linux-2.6.36/mm/rmap.c linux-2.6.36-vs2.3.0.36.33/mm/rmap.c
25049 --- linux-2.6.36/mm/rmap.c 2010-10-21 13:07:57.000000000 +0200
25050 +++ linux-2.6.36-vs2.3.0.36.33/mm/rmap.c 2010-10-21 13:19:11.000000000 +0200
25052 #include <linux/mmu_notifier.h>
25053 #include <linux/migrate.h>
25054 #include <linux/hugetlb.h>
25055 +#include <linux/vs_memory.h>
25057 #include <asm/tlbflush.h>
25059 diff -NurpP --minimal linux-2.6.36/mm/shmem.c linux-2.6.36-vs2.3.0.36.33/mm/shmem.c
25060 --- linux-2.6.36/mm/shmem.c 2010-10-21 13:07:57.000000000 +0200
25061 +++ linux-2.6.36-vs2.3.0.36.33/mm/shmem.c 2010-10-21 13:09:36.000000000 +0200
25062 @@ -1815,7 +1815,7 @@ static int shmem_statfs(struct dentry *d
25064 struct shmem_sb_info *sbinfo = SHMEM_SB(dentry->d_sb);
25066 - buf->f_type = TMPFS_MAGIC;
25067 + buf->f_type = TMPFS_SUPER_MAGIC;
25068 buf->f_bsize = PAGE_CACHE_SIZE;
25069 buf->f_namelen = NAME_MAX;
25070 if (sbinfo->max_blocks) {
25071 @@ -2377,7 +2377,7 @@ int shmem_fill_super(struct super_block
25072 sb->s_maxbytes = SHMEM_MAX_BYTES;
25073 sb->s_blocksize = PAGE_CACHE_SIZE;
25074 sb->s_blocksize_bits = PAGE_CACHE_SHIFT;
25075 - sb->s_magic = TMPFS_MAGIC;
25076 + sb->s_magic = TMPFS_SUPER_MAGIC;
25077 sb->s_op = &shmem_ops;
25078 sb->s_time_gran = 1;
25079 #ifdef CONFIG_TMPFS_POSIX_ACL
25080 diff -NurpP --minimal linux-2.6.36/mm/slab.c linux-2.6.36-vs2.3.0.36.33/mm/slab.c
25081 --- linux-2.6.36/mm/slab.c 2010-10-21 13:07:57.000000000 +0200
25082 +++ linux-2.6.36-vs2.3.0.36.33/mm/slab.c 2010-10-21 13:09:36.000000000 +0200
25083 @@ -407,6 +407,8 @@ static void kmem_list3_init(struct kmem_
25084 #define STATS_INC_FREEMISS(x) do { } while (0)
25087 +#include "slab_vs.h"
25092 @@ -3346,6 +3348,7 @@ retry:
25094 obj = slab_get_obj(cachep, slabp, nodeid);
25095 check_slabp(cachep, slabp);
25096 + vx_slab_alloc(cachep, flags);
25097 l3->free_objects--;
25098 /* move slabp to correct slabp list: */
25099 list_del(&slabp->list);
25100 @@ -3423,6 +3426,7 @@ __cache_alloc_node(struct kmem_cache *ca
25101 /* ___cache_alloc_node can fall back to other nodes */
25102 ptr = ____cache_alloc_node(cachep, flags, nodeid);
25104 + vx_slab_alloc(cachep, flags);
25105 local_irq_restore(save_flags);
25106 ptr = cache_alloc_debugcheck_after(cachep, flags, ptr, caller);
25107 kmemleak_alloc_recursive(ptr, obj_size(cachep), 1, cachep->flags,
25108 @@ -3609,6 +3613,7 @@ static inline void __cache_free(struct k
25110 kmemleak_free_recursive(objp, cachep->flags);
25111 objp = cache_free_debugcheck(cachep, objp, __builtin_return_address(0));
25112 + vx_slab_free(cachep);
25114 kmemcheck_slab_free(cachep, objp, obj_size(cachep));
25116 diff -NurpP --minimal linux-2.6.36/mm/slab_vs.h linux-2.6.36-vs2.3.0.36.33/mm/slab_vs.h
25117 --- linux-2.6.36/mm/slab_vs.h 1970-01-01 01:00:00.000000000 +0100
25118 +++ linux-2.6.36-vs2.3.0.36.33/mm/slab_vs.h 2010-10-21 13:09:36.000000000 +0200
25121 +#include <linux/vserver/context.h>
25123 +#include <linux/vs_context.h>
25126 +void vx_slab_alloc(struct kmem_cache *cachep, gfp_t flags)
25128 + int what = gfp_zone(cachep->gfpflags);
25129 + struct vx_info *vxi = current_vx_info();
25134 + atomic_add(cachep->buffer_size, &vxi->cacct.slab[what]);
25138 +void vx_slab_free(struct kmem_cache *cachep)
25140 + int what = gfp_zone(cachep->gfpflags);
25141 + struct vx_info *vxi = current_vx_info();
25146 + atomic_sub(cachep->buffer_size, &vxi->cacct.slab[what]);
25149 diff -NurpP --minimal linux-2.6.36/mm/swapfile.c linux-2.6.36-vs2.3.0.36.33/mm/swapfile.c
25150 --- linux-2.6.36/mm/swapfile.c 2010-10-21 13:07:57.000000000 +0200
25151 +++ linux-2.6.36-vs2.3.0.36.33/mm/swapfile.c 2010-10-21 13:09:36.000000000 +0200
25153 #include <asm/tlbflush.h>
25154 #include <linux/swapops.h>
25155 #include <linux/page_cgroup.h>
25156 +#include <linux/vs_base.h>
25157 +#include <linux/vs_memory.h>
25159 static bool swap_count_continued(struct swap_info_struct *, pgoff_t,
25161 @@ -1747,6 +1749,16 @@ static int swap_show(struct seq_file *sw
25163 if (si == SEQ_START_TOKEN) {
25164 seq_puts(swap,"Filename\t\t\t\tType\t\tSize\tUsed\tPriority\n");
25165 + if (vx_flags(VXF_VIRT_MEM, 0)) {
25166 + struct sysinfo si;
25168 + vx_vsi_swapinfo(&si);
25169 + if (si.totalswap < (1 << 10))
25171 + seq_printf(swap, "%s\t\t\t\t\t%s\t%lu\t%lu\t%d\n",
25172 + "hdv0", "partition", si.totalswap >> 10,
25173 + (si.totalswap - si.freeswap) >> 10, -1);
25178 @@ -2131,6 +2143,8 @@ void si_swapinfo(struct sysinfo *val)
25179 val->freeswap = nr_swap_pages + nr_to_be_unused;
25180 val->totalswap = total_swap_pages + nr_to_be_unused;
25181 spin_unlock(&swap_lock);
25182 + if (vx_flags(VXF_VIRT_MEM, 0))
25183 + vx_vsi_swapinfo(val);
25187 diff -NurpP --minimal linux-2.6.36/net/core/dev.c linux-2.6.36-vs2.3.0.36.33/net/core/dev.c
25188 --- linux-2.6.36/net/core/dev.c 2010-10-21 13:07:58.000000000 +0200
25189 +++ linux-2.6.36-vs2.3.0.36.33/net/core/dev.c 2010-10-21 14:12:56.000000000 +0200
25190 @@ -127,6 +127,7 @@
25191 #include <linux/in.h>
25192 #include <linux/jhash.h>
25193 #include <linux/random.h>
25194 +#include <linux/vs_inet.h>
25195 #include <trace/events/napi.h>
25196 #include <linux/pci.h>
25198 @@ -607,7 +608,8 @@ struct net_device *__dev_get_by_name(str
25199 struct hlist_head *head = dev_name_hash(net, name);
25201 hlist_for_each_entry(dev, p, head, name_hlist)
25202 - if (!strncmp(dev->name, name, IFNAMSIZ))
25203 + if (!strncmp(dev->name, name, IFNAMSIZ) &&
25204 + nx_dev_visible(current_nx_info(), dev))
25208 @@ -633,7 +635,8 @@ struct net_device *dev_get_by_name_rcu(s
25209 struct hlist_head *head = dev_name_hash(net, name);
25211 hlist_for_each_entry_rcu(dev, p, head, name_hlist)
25212 - if (!strncmp(dev->name, name, IFNAMSIZ))
25213 + if (!strncmp(dev->name, name, IFNAMSIZ) &&
25214 + nx_dev_visible(current_nx_info(), dev))
25218 @@ -684,7 +687,8 @@ struct net_device *__dev_get_by_index(st
25219 struct hlist_head *head = dev_index_hash(net, ifindex);
25221 hlist_for_each_entry(dev, p, head, index_hlist)
25222 - if (dev->ifindex == ifindex)
25223 + if ((dev->ifindex == ifindex) &&
25224 + nx_dev_visible(current_nx_info(), dev))
25228 @@ -709,7 +713,8 @@ struct net_device *dev_get_by_index_rcu(
25229 struct hlist_head *head = dev_index_hash(net, ifindex);
25231 hlist_for_each_entry_rcu(dev, p, head, index_hlist)
25232 - if (dev->ifindex == ifindex)
25233 + if ((dev->ifindex == ifindex) &&
25234 + nx_dev_visible(current_nx_info(), dev))
25238 @@ -762,10 +767,12 @@ struct net_device *dev_getbyhwaddr(struc
25242 - for_each_netdev(net, dev)
25243 + for_each_netdev(net, dev) {
25244 if (dev->type == type &&
25245 - !memcmp(dev->dev_addr, ha, dev->addr_len))
25246 + !memcmp(dev->dev_addr, ha, dev->addr_len) &&
25247 + nx_dev_visible(current_nx_info(), dev))
25253 @@ -776,9 +783,11 @@ struct net_device *__dev_getfirstbyhwtyp
25254 struct net_device *dev;
25257 - for_each_netdev(net, dev)
25258 - if (dev->type == type)
25259 + for_each_netdev(net, dev) {
25260 + if ((dev->type == type) &&
25261 + nx_dev_visible(current_nx_info(), dev))
25267 @@ -896,6 +905,8 @@ static int __dev_alloc_name(struct net *
25269 if (i < 0 || i >= max_netdevices)
25271 + if (!nx_dev_visible(current_nx_info(), d))
25274 /* avoid cases where sscanf is not exact inverse of printf */
25275 snprintf(buf, IFNAMSIZ, name, i);
25276 @@ -3658,6 +3669,8 @@ static int dev_ifconf(struct net *net, c
25279 for_each_netdev(net, dev) {
25280 + if (!nx_dev_visible(current_nx_info(), dev))
25282 for (i = 0; i < NPROTO; i++) {
25283 if (gifconf_list[i]) {
25285 @@ -3729,6 +3742,10 @@ static void dev_seq_printf_stats(struct
25286 struct rtnl_link_stats64 temp;
25287 const struct rtnl_link_stats64 *stats = dev_get_stats(dev, &temp);
25289 + /* device visible inside network context? */
25290 + if (!nx_dev_visible(current_nx_info(), dev))
25293 seq_printf(seq, "%6s: %7llu %7llu %4llu %4llu %4llu %5llu %10llu %9llu "
25294 "%8llu %7llu %4llu %4llu %4llu %5llu %7llu %10llu\n",
25295 dev->name, stats->rx_bytes, stats->rx_packets,
25296 diff -NurpP --minimal linux-2.6.36/net/core/rtnetlink.c linux-2.6.36-vs2.3.0.36.33/net/core/rtnetlink.c
25297 --- linux-2.6.36/net/core/rtnetlink.c 2010-10-21 13:07:58.000000000 +0200
25298 +++ linux-2.6.36-vs2.3.0.36.33/net/core/rtnetlink.c 2010-10-21 13:09:36.000000000 +0200
25299 @@ -927,6 +927,8 @@ static int rtnl_dump_ifinfo(struct sk_bu
25300 hlist_for_each_entry(dev, node, head, index_hlist) {
25303 + if (!nx_dev_visible(skb->sk->sk_nx_info, dev))
25305 if (rtnl_fill_ifinfo(skb, dev, RTM_NEWLINK,
25306 NETLINK_CB(cb->skb).pid,
25307 cb->nlh->nlmsg_seq, 0,
25308 @@ -1643,6 +1645,9 @@ void rtmsg_ifinfo(int type, struct net_d
25309 struct sk_buff *skb;
25310 int err = -ENOBUFS;
25312 + if (!nx_dev_visible(current_nx_info(), dev))
25315 skb = nlmsg_new(if_nlmsg_size(dev), GFP_KERNEL);
25318 diff -NurpP --minimal linux-2.6.36/net/core/sock.c linux-2.6.36-vs2.3.0.36.33/net/core/sock.c
25319 --- linux-2.6.36/net/core/sock.c 2010-10-21 13:07:58.000000000 +0200
25320 +++ linux-2.6.36-vs2.3.0.36.33/net/core/sock.c 2010-10-21 13:09:36.000000000 +0200
25321 @@ -127,6 +127,10 @@
25322 #include <net/cls_cgroup.h>
25324 #include <linux/filter.h>
25325 +#include <linux/vs_socket.h>
25326 +#include <linux/vs_limit.h>
25327 +#include <linux/vs_context.h>
25328 +#include <linux/vs_network.h>
25331 #include <net/tcp.h>
25332 @@ -1046,6 +1050,8 @@ static struct sock *sk_prot_alloc(struct
25334 sk_tx_queue_clear(sk);
25336 + sock_vx_init(sk);
25337 + sock_nx_init(sk);
25341 @@ -1142,6 +1148,11 @@ static void __sk_free(struct sock *sk)
25342 put_cred(sk->sk_peer_cred);
25343 put_pid(sk->sk_peer_pid);
25344 put_net(sock_net(sk));
25346 + clr_vx_info(&sk->sk_vx_info);
25348 + clr_nx_info(&sk->sk_nx_info);
25350 sk_prot_free(sk->sk_prot_creator, sk);
25353 @@ -1189,6 +1200,8 @@ struct sock *sk_clone(const struct sock
25356 get_net(sock_net(newsk));
25357 + sock_vx_init(newsk);
25358 + sock_nx_init(newsk);
25359 sk_node_init(&newsk->sk_node);
25360 sock_lock_init(newsk);
25361 bh_lock_sock(newsk);
25362 @@ -1244,6 +1257,12 @@ struct sock *sk_clone(const struct sock
25364 atomic_set(&newsk->sk_refcnt, 2);
25366 + set_vx_info(&newsk->sk_vx_info, sk->sk_vx_info);
25367 + newsk->sk_xid = sk->sk_xid;
25368 + vx_sock_inc(newsk);
25369 + set_nx_info(&newsk->sk_nx_info, sk->sk_nx_info);
25370 + newsk->sk_nid = sk->sk_nid;
25373 * Increment the counter in the same struct proto as the master
25374 * sock (sk_refcnt_debug_inc uses newsk->sk_prot->socks, that
25375 @@ -1986,6 +2005,12 @@ void sock_init_data(struct socket *sock,
25377 sk->sk_stamp = ktime_set(-1L, 0);
25379 + set_vx_info(&sk->sk_vx_info, current_vx_info());
25380 + sk->sk_xid = vx_current_xid();
25382 + set_nx_info(&sk->sk_nx_info, current_nx_info());
25383 + sk->sk_nid = nx_current_nid();
25386 * Before updating sk_refcnt, we must commit prior changes to memory
25387 * (Documentation/RCU/rculist_nulls.txt for details)
25388 diff -NurpP --minimal linux-2.6.36/net/ipv4/af_inet.c linux-2.6.36-vs2.3.0.36.33/net/ipv4/af_inet.c
25389 --- linux-2.6.36/net/ipv4/af_inet.c 2010-10-21 13:07:58.000000000 +0200
25390 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv4/af_inet.c 2010-10-21 13:09:36.000000000 +0200
25391 @@ -116,6 +116,7 @@
25392 #ifdef CONFIG_IP_MROUTE
25393 #include <linux/mroute.h>
25395 +#include <linux/vs_limit.h>
25398 /* The inetsw table contains everything that inet_create needs to
25399 @@ -327,9 +328,13 @@ lookup_protocol:
25403 + if ((protocol == IPPROTO_ICMP) &&
25404 + nx_capable(CAP_NET_RAW, NXC_RAW_ICMP))
25407 if (sock->type == SOCK_RAW && !kern && !capable(CAP_NET_RAW))
25408 goto out_rcu_unlock;
25411 err = -EAFNOSUPPORT;
25412 if (!inet_netns_ok(net, protocol))
25413 goto out_rcu_unlock;
25414 @@ -453,6 +458,7 @@ int inet_bind(struct socket *sock, struc
25415 struct sockaddr_in *addr = (struct sockaddr_in *)uaddr;
25416 struct sock *sk = sock->sk;
25417 struct inet_sock *inet = inet_sk(sk);
25418 + struct nx_v4_sock_addr nsa;
25419 unsigned short snum;
25422 @@ -466,7 +472,11 @@ int inet_bind(struct socket *sock, struc
25423 if (addr_len < sizeof(struct sockaddr_in))
25426 - chk_addr_ret = inet_addr_type(sock_net(sk), addr->sin_addr.s_addr);
25427 + err = v4_map_sock_addr(inet, addr, &nsa);
25431 + chk_addr_ret = inet_addr_type(sock_net(sk), nsa.saddr);
25433 /* Not specified by any standard per-se, however it breaks too
25434 * many applications when removed. It is unfortunate since
25435 @@ -478,7 +488,7 @@ int inet_bind(struct socket *sock, struc
25436 err = -EADDRNOTAVAIL;
25437 if (!sysctl_ip_nonlocal_bind &&
25438 !(inet->freebind || inet->transparent) &&
25439 - addr->sin_addr.s_addr != htonl(INADDR_ANY) &&
25440 + nsa.saddr != htonl(INADDR_ANY) &&
25441 chk_addr_ret != RTN_LOCAL &&
25442 chk_addr_ret != RTN_MULTICAST &&
25443 chk_addr_ret != RTN_BROADCAST)
25444 @@ -503,7 +513,7 @@ int inet_bind(struct socket *sock, struc
25445 if (sk->sk_state != TCP_CLOSE || inet->inet_num)
25446 goto out_release_sock;
25448 - inet->inet_rcv_saddr = inet->inet_saddr = addr->sin_addr.s_addr;
25449 + v4_set_sock_addr(inet, &nsa);
25450 if (chk_addr_ret == RTN_MULTICAST || chk_addr_ret == RTN_BROADCAST)
25451 inet->inet_saddr = 0; /* Use device */
25453 @@ -705,11 +715,13 @@ int inet_getname(struct socket *sock, st
25456 sin->sin_port = inet->inet_dport;
25457 - sin->sin_addr.s_addr = inet->inet_daddr;
25458 + sin->sin_addr.s_addr =
25459 + nx_map_sock_lback(sk->sk_nx_info, inet->inet_daddr);
25461 __be32 addr = inet->inet_rcv_saddr;
25463 addr = inet->inet_saddr;
25464 + addr = nx_map_sock_lback(sk->sk_nx_info, addr);
25465 sin->sin_port = inet->inet_sport;
25466 sin->sin_addr.s_addr = addr;
25468 diff -NurpP --minimal linux-2.6.36/net/ipv4/devinet.c linux-2.6.36-vs2.3.0.36.33/net/ipv4/devinet.c
25469 --- linux-2.6.36/net/ipv4/devinet.c 2010-10-21 13:07:58.000000000 +0200
25470 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv4/devinet.c 2010-10-21 13:09:36.000000000 +0200
25471 @@ -417,6 +417,7 @@ struct in_device *inetdev_by_index(struc
25473 EXPORT_SYMBOL(inetdev_by_index);
25476 /* Called only from RTNL semaphored context. No locks. */
25478 struct in_ifaddr *inet_ifa_byprefix(struct in_device *in_dev, __be32 prefix,
25479 @@ -659,6 +660,8 @@ int devinet_ioctl(struct net *net, unsig
25481 in_dev = __in_dev_get_rtnl(dev);
25483 + struct nx_info *nxi = current_nx_info();
25485 if (tryaddrmatch) {
25486 /* Matthias Andree */
25487 /* compare label and address (4.4BSD style) */
25488 @@ -667,6 +670,8 @@ int devinet_ioctl(struct net *net, unsig
25489 This is checked above. */
25490 for (ifap = &in_dev->ifa_list; (ifa = *ifap) != NULL;
25491 ifap = &ifa->ifa_next) {
25492 + if (!nx_v4_ifa_visible(nxi, ifa))
25494 if (!strcmp(ifr.ifr_name, ifa->ifa_label) &&
25495 sin_orig.sin_addr.s_addr ==
25496 ifa->ifa_address) {
25497 @@ -679,9 +684,12 @@ int devinet_ioctl(struct net *net, unsig
25498 comparing just the label */
25500 for (ifap = &in_dev->ifa_list; (ifa = *ifap) != NULL;
25501 - ifap = &ifa->ifa_next)
25502 + ifap = &ifa->ifa_next) {
25503 + if (!nx_v4_ifa_visible(nxi, ifa))
25505 if (!strcmp(ifr.ifr_name, ifa->ifa_label))
25511 @@ -833,6 +841,8 @@ static int inet_gifconf(struct net_devic
25514 for (ifa = in_dev->ifa_list; ifa; ifa = ifa->ifa_next) {
25515 + if (!nx_v4_ifa_visible(current_nx_info(), ifa))
25518 done += sizeof(ifr);
25520 @@ -1182,6 +1192,7 @@ static int inet_dump_ifaddr(struct sk_bu
25521 struct net_device *dev;
25522 struct in_device *in_dev;
25523 struct in_ifaddr *ifa;
25524 + struct sock *sk = skb->sk;
25525 struct hlist_head *head;
25526 struct hlist_node *node;
25528 @@ -1204,6 +1215,8 @@ static int inet_dump_ifaddr(struct sk_bu
25530 for (ifa = in_dev->ifa_list, ip_idx = 0; ifa;
25531 ifa = ifa->ifa_next, ip_idx++) {
25532 + if (sk && !nx_v4_ifa_visible(sk->sk_nx_info, ifa))
25534 if (ip_idx < s_ip_idx)
25536 if (inet_fill_ifaddr(skb, ifa,
25537 diff -NurpP --minimal linux-2.6.36/net/ipv4/fib_hash.c linux-2.6.36-vs2.3.0.36.33/net/ipv4/fib_hash.c
25538 --- linux-2.6.36/net/ipv4/fib_hash.c 2010-07-07 18:31:59.000000000 +0200
25539 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv4/fib_hash.c 2010-10-21 13:09:36.000000000 +0200
25540 @@ -1017,7 +1017,7 @@ static int fib_seq_show(struct seq_file
25541 prefix = f->fn_key;
25542 mask = FZ_MASK(iter->zone);
25543 flags = fib_flag_trans(fa->fa_type, mask, fi);
25545 + if (fi && nx_dev_visible(current_nx_info(), fi->fib_dev))
25547 "%s\t%08X\t%08X\t%04X\t%d\t%u\t%d\t%08X\t%d\t%u\t%u%n",
25548 fi->fib_dev ? fi->fib_dev->name : "*", prefix,
25549 diff -NurpP --minimal linux-2.6.36/net/ipv4/inet_connection_sock.c linux-2.6.36-vs2.3.0.36.33/net/ipv4/inet_connection_sock.c
25550 --- linux-2.6.36/net/ipv4/inet_connection_sock.c 2010-10-21 13:07:58.000000000 +0200
25551 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv4/inet_connection_sock.c 2010-10-21 13:09:36.000000000 +0200
25552 @@ -52,10 +52,40 @@ void inet_get_local_port_range(int *low,
25554 EXPORT_SYMBOL(inet_get_local_port_range);
25556 +int ipv4_rcv_saddr_equal(const struct sock *sk1, const struct sock *sk2)
25558 + __be32 sk1_rcv_saddr = inet_rcv_saddr(sk1),
25559 + sk2_rcv_saddr = inet_rcv_saddr(sk2);
25561 + if (inet_v6_ipv6only(sk2))
25564 + if (sk1_rcv_saddr &&
25566 + sk1_rcv_saddr == sk2_rcv_saddr)
25569 + if (sk1_rcv_saddr &&
25570 + !sk2_rcv_saddr &&
25571 + v4_addr_in_nx_info(sk2->sk_nx_info, sk1_rcv_saddr, NXA_MASK_BIND))
25574 + if (sk2_rcv_saddr &&
25575 + !sk1_rcv_saddr &&
25576 + v4_addr_in_nx_info(sk1->sk_nx_info, sk2_rcv_saddr, NXA_MASK_BIND))
25579 + if (!sk1_rcv_saddr &&
25580 + !sk2_rcv_saddr &&
25581 + nx_v4_addr_conflict(sk1->sk_nx_info, sk2->sk_nx_info))
25587 int inet_csk_bind_conflict(const struct sock *sk,
25588 const struct inet_bind_bucket *tb)
25590 - const __be32 sk_rcv_saddr = inet_rcv_saddr(sk);
25592 struct hlist_node *node;
25593 int reuse = sk->sk_reuse;
25594 @@ -75,9 +105,7 @@ int inet_csk_bind_conflict(const struct
25595 sk->sk_bound_dev_if == sk2->sk_bound_dev_if)) {
25596 if (!reuse || !sk2->sk_reuse ||
25597 sk2->sk_state == TCP_LISTEN) {
25598 - const __be32 sk2_rcv_saddr = inet_rcv_saddr(sk2);
25599 - if (!sk2_rcv_saddr || !sk_rcv_saddr ||
25600 - sk2_rcv_saddr == sk_rcv_saddr)
25601 + if (ipv4_rcv_saddr_equal(sk, sk2))
25605 diff -NurpP --minimal linux-2.6.36/net/ipv4/inet_diag.c linux-2.6.36-vs2.3.0.36.33/net/ipv4/inet_diag.c
25606 --- linux-2.6.36/net/ipv4/inet_diag.c 2010-07-07 18:31:59.000000000 +0200
25607 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv4/inet_diag.c 2010-10-21 13:09:36.000000000 +0200
25609 #include <linux/stddef.h>
25611 #include <linux/inet_diag.h>
25612 +#include <linux/vs_network.h>
25613 +#include <linux/vs_inet.h>
25615 static const struct inet_diag_handler **inet_diag_table;
25617 @@ -119,8 +121,10 @@ static int inet_csk_diag_fill(struct soc
25619 r->id.idiag_sport = inet->inet_sport;
25620 r->id.idiag_dport = inet->inet_dport;
25621 - r->id.idiag_src[0] = inet->inet_rcv_saddr;
25622 - r->id.idiag_dst[0] = inet->inet_daddr;
25623 + r->id.idiag_src[0] = nx_map_sock_lback(sk->sk_nx_info,
25624 + inet->inet_rcv_saddr);
25625 + r->id.idiag_dst[0] = nx_map_sock_lback(sk->sk_nx_info,
25626 + inet->inet_daddr);
25628 #if defined(CONFIG_IPV6) || defined (CONFIG_IPV6_MODULE)
25629 if (r->idiag_family == AF_INET6) {
25630 @@ -205,8 +209,8 @@ static int inet_twsk_diag_fill(struct in
25631 r->id.idiag_cookie[1] = (u32)(((unsigned long)tw >> 31) >> 1);
25632 r->id.idiag_sport = tw->tw_sport;
25633 r->id.idiag_dport = tw->tw_dport;
25634 - r->id.idiag_src[0] = tw->tw_rcv_saddr;
25635 - r->id.idiag_dst[0] = tw->tw_daddr;
25636 + r->id.idiag_src[0] = nx_map_sock_lback(tw->tw_nx_info, tw->tw_rcv_saddr);
25637 + r->id.idiag_dst[0] = nx_map_sock_lback(tw->tw_nx_info, tw->tw_daddr);
25638 r->idiag_state = tw->tw_substate;
25639 r->idiag_timer = 3;
25640 r->idiag_expires = DIV_ROUND_UP(tmo * 1000, HZ);
25641 @@ -263,6 +267,7 @@ static int inet_diag_get_exact(struct sk
25644 if (req->idiag_family == AF_INET) {
25645 + /* TODO: lback */
25646 sk = inet_lookup(&init_net, hashinfo, req->id.idiag_dst[0],
25647 req->id.idiag_dport, req->id.idiag_src[0],
25648 req->id.idiag_sport, req->id.idiag_if);
25649 @@ -505,6 +510,7 @@ static int inet_csk_diag_dump(struct soc
25653 + /* TODO: lback */
25654 entry.saddr = &inet->inet_rcv_saddr;
25655 entry.daddr = &inet->inet_daddr;
25657 @@ -541,6 +547,7 @@ static int inet_twsk_diag_dump(struct in
25661 + /* TODO: lback */
25662 entry.saddr = &tw->tw_rcv_saddr;
25663 entry.daddr = &tw->tw_daddr;
25665 @@ -587,8 +594,8 @@ static int inet_diag_fill_req(struct sk_
25667 r->id.idiag_sport = inet->inet_sport;
25668 r->id.idiag_dport = ireq->rmt_port;
25669 - r->id.idiag_src[0] = ireq->loc_addr;
25670 - r->id.idiag_dst[0] = ireq->rmt_addr;
25671 + r->id.idiag_src[0] = nx_map_sock_lback(sk->sk_nx_info, ireq->loc_addr);
25672 + r->id.idiag_dst[0] = nx_map_sock_lback(sk->sk_nx_info, ireq->rmt_addr);
25673 r->idiag_expires = jiffies_to_msecs(tmo);
25674 r->idiag_rqueue = 0;
25675 r->idiag_wqueue = 0;
25676 @@ -658,6 +665,7 @@ static int inet_diag_dump_reqs(struct sk
25680 + /* TODO: lback */
25682 #if defined(CONFIG_IPV6) || defined (CONFIG_IPV6_MODULE)
25683 (entry.family == AF_INET6) ?
25684 @@ -728,6 +736,8 @@ static int inet_diag_dump(struct sk_buff
25685 sk_nulls_for_each(sk, node, &ilb->head) {
25686 struct inet_sock *inet = inet_sk(sk);
25688 + if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
25693 @@ -794,6 +804,8 @@ skip_listen_ht:
25694 sk_nulls_for_each(sk, node, &head->chain) {
25695 struct inet_sock *inet = inet_sk(sk);
25697 + if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
25701 if (!(r->idiag_states & (1 << sk->sk_state)))
25702 @@ -818,6 +830,8 @@ next_normal:
25703 inet_twsk_for_each(tw, node,
25706 + if (!nx_check(tw->tw_nid, VS_WATCH_P | VS_IDENT))
25710 if (r->id.idiag_sport != tw->tw_sport &&
25711 diff -NurpP --minimal linux-2.6.36/net/ipv4/inet_hashtables.c linux-2.6.36-vs2.3.0.36.33/net/ipv4/inet_hashtables.c
25712 --- linux-2.6.36/net/ipv4/inet_hashtables.c 2010-10-21 13:07:58.000000000 +0200
25713 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv4/inet_hashtables.c 2010-10-21 13:09:36.000000000 +0200
25716 #include <net/inet_connection_sock.h>
25717 #include <net/inet_hashtables.h>
25718 +#include <net/route.h>
25719 #include <net/ip.h>
25722 @@ -132,6 +133,11 @@ static inline int compute_score(struct s
25723 if (rcv_saddr != daddr)
25727 + /* block non nx_info ips */
25728 + if (!v4_addr_in_nx_info(sk->sk_nx_info,
25729 + daddr, NXA_MASK_BIND))
25732 if (sk->sk_bound_dev_if) {
25733 if (sk->sk_bound_dev_if != dif)
25734 @@ -149,7 +155,6 @@ static inline int compute_score(struct s
25735 * wildcarded during the search since they can never be otherwise.
25739 struct sock *__inet_lookup_listener(struct net *net,
25740 struct inet_hashinfo *hashinfo,
25741 const __be32 daddr, const unsigned short hnum,
25742 @@ -172,6 +177,7 @@ begin:
25748 * if the nulls value we got at the end of this lookup is
25749 * not the expected one, we must restart lookup.
25750 diff -NurpP --minimal linux-2.6.36/net/ipv4/netfilter/nf_nat_helper.c linux-2.6.36-vs2.3.0.36.33/net/ipv4/netfilter/nf_nat_helper.c
25751 --- linux-2.6.36/net/ipv4/netfilter/nf_nat_helper.c 2010-07-07 18:31:59.000000000 +0200
25752 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv4/netfilter/nf_nat_helper.c 2010-10-21 13:09:36.000000000 +0200
25754 #include <net/route.h>
25756 #include <linux/netfilter_ipv4.h>
25757 +#include <net/route.h>
25758 #include <net/netfilter/nf_conntrack.h>
25759 #include <net/netfilter/nf_conntrack_helper.h>
25760 #include <net/netfilter/nf_conntrack_ecache.h>
25761 diff -NurpP --minimal linux-2.6.36/net/ipv4/netfilter.c linux-2.6.36-vs2.3.0.36.33/net/ipv4/netfilter.c
25762 --- linux-2.6.36/net/ipv4/netfilter.c 2010-10-21 13:07:58.000000000 +0200
25763 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv4/netfilter.c 2010-10-21 13:09:36.000000000 +0200
25765 #include <linux/ip.h>
25766 #include <linux/skbuff.h>
25767 #include <linux/gfp.h>
25768 -#include <net/route.h>
25769 +// #include <net/route.h>
25770 #include <net/xfrm.h>
25771 #include <net/ip.h>
25772 #include <net/netfilter/nf_queue.h>
25773 diff -NurpP --minimal linux-2.6.36/net/ipv4/raw.c linux-2.6.36-vs2.3.0.36.33/net/ipv4/raw.c
25774 --- linux-2.6.36/net/ipv4/raw.c 2010-10-21 13:07:58.000000000 +0200
25775 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv4/raw.c 2010-10-21 13:09:36.000000000 +0200
25776 @@ -116,7 +116,7 @@ static struct sock *__raw_v4_lookup(stru
25778 if (net_eq(sock_net(sk), net) && inet->inet_num == num &&
25779 !(inet->inet_daddr && inet->inet_daddr != raddr) &&
25780 - !(inet->inet_rcv_saddr && inet->inet_rcv_saddr != laddr) &&
25781 + v4_sock_addr_match(sk->sk_nx_info, inet, laddr) &&
25782 !(sk->sk_bound_dev_if && sk->sk_bound_dev_if != dif))
25783 goto found; /* gotcha */
25785 @@ -383,6 +383,12 @@ static int raw_send_hdrinc(struct sock *
25786 icmp_out_count(net, ((struct icmphdr *)
25787 skb_transport_header(skb))->type);
25790 + if (!nx_check(0, VS_ADMIN) && !capable(CAP_NET_RAW) &&
25791 + sk->sk_nx_info &&
25792 + !v4_addr_in_nx_info(sk->sk_nx_info, iph->saddr, NXA_MASK_BIND))
25795 err = NF_HOOK(NFPROTO_IPV4, NF_INET_LOCAL_OUT, skb, NULL,
25796 rt->dst.dev, dst_output);
25798 @@ -563,6 +569,13 @@ static int raw_sendmsg(struct kiocb *ioc
25801 security_sk_classify_flow(sk, &fl);
25802 + if (sk->sk_nx_info) {
25803 + err = ip_v4_find_src(sock_net(sk),
25804 + sk->sk_nx_info, &rt, &fl);
25809 err = ip_route_output_flow(sock_net(sk), &rt, &fl, sk, 1);
25812 @@ -635,17 +648,19 @@ static int raw_bind(struct sock *sk, str
25814 struct inet_sock *inet = inet_sk(sk);
25815 struct sockaddr_in *addr = (struct sockaddr_in *) uaddr;
25816 + struct nx_v4_sock_addr nsa = { 0 };
25820 if (sk->sk_state != TCP_CLOSE || addr_len < sizeof(struct sockaddr_in))
25822 - chk_addr_ret = inet_addr_type(sock_net(sk), addr->sin_addr.s_addr);
25823 + v4_map_sock_addr(inet, addr, &nsa);
25824 + chk_addr_ret = inet_addr_type(sock_net(sk), nsa.saddr);
25825 ret = -EADDRNOTAVAIL;
25826 - if (addr->sin_addr.s_addr && chk_addr_ret != RTN_LOCAL &&
25827 + if (nsa.saddr && chk_addr_ret != RTN_LOCAL &&
25828 chk_addr_ret != RTN_MULTICAST && chk_addr_ret != RTN_BROADCAST)
25830 - inet->inet_rcv_saddr = inet->inet_saddr = addr->sin_addr.s_addr;
25831 + v4_set_sock_addr(inet, &nsa);
25832 if (chk_addr_ret == RTN_MULTICAST || chk_addr_ret == RTN_BROADCAST)
25833 inet->inet_saddr = 0; /* Use device */
25835 @@ -697,7 +712,8 @@ static int raw_recvmsg(struct kiocb *ioc
25836 /* Copy the address. */
25838 sin->sin_family = AF_INET;
25839 - sin->sin_addr.s_addr = ip_hdr(skb)->saddr;
25840 + sin->sin_addr.s_addr =
25841 + nx_map_sock_lback(sk->sk_nx_info, ip_hdr(skb)->saddr);
25843 memset(&sin->sin_zero, 0, sizeof(sin->sin_zero));
25845 @@ -875,7 +891,8 @@ static struct sock *raw_get_first(struct
25846 struct hlist_node *node;
25848 sk_for_each(sk, node, &state->h->ht[state->bucket])
25849 - if (sock_net(sk) == seq_file_net(seq))
25850 + if ((sock_net(sk) == seq_file_net(seq)) &&
25851 + nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
25855 @@ -891,7 +908,8 @@ static struct sock *raw_get_next(struct
25859 - } while (sk && sock_net(sk) != seq_file_net(seq));
25860 + } while (sk && ((sock_net(sk) != seq_file_net(seq)) ||
25861 + !nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT)));
25863 if (!sk && ++state->bucket < RAW_HTABLE_SIZE) {
25864 sk = sk_head(&state->h->ht[state->bucket]);
25865 @@ -950,7 +968,10 @@ static void raw_sock_seq_show(struct seq
25867 seq_printf(seq, "%4d: %08X:%04X %08X:%04X"
25868 " %02X %08X:%08X %02X:%08lX %08X %5d %8d %lu %d %p %d\n",
25869 - i, src, srcp, dest, destp, sp->sk_state,
25871 + nx_map_sock_lback(current_nx_info(), src), srcp,
25872 + nx_map_sock_lback(current_nx_info(), dest), destp,
25874 sk_wmem_alloc_get(sp),
25875 sk_rmem_alloc_get(sp),
25876 0, 0L, 0, sock_i_uid(sp), 0, sock_i_ino(sp),
25877 diff -NurpP --minimal linux-2.6.36/net/ipv4/tcp.c linux-2.6.36-vs2.3.0.36.33/net/ipv4/tcp.c
25878 --- linux-2.6.36/net/ipv4/tcp.c 2010-10-21 13:07:58.000000000 +0200
25879 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv4/tcp.c 2010-10-21 13:09:36.000000000 +0200
25880 @@ -266,6 +266,7 @@
25881 #include <linux/crypto.h>
25882 #include <linux/time.h>
25883 #include <linux/slab.h>
25884 +#include <linux/in.h>
25886 #include <net/icmp.h>
25887 #include <net/tcp.h>
25888 diff -NurpP --minimal linux-2.6.36/net/ipv4/tcp_ipv4.c linux-2.6.36-vs2.3.0.36.33/net/ipv4/tcp_ipv4.c
25889 --- linux-2.6.36/net/ipv4/tcp_ipv4.c 2010-10-21 13:07:58.000000000 +0200
25890 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv4/tcp_ipv4.c 2010-10-21 13:09:36.000000000 +0200
25891 @@ -2016,6 +2016,12 @@ static void *listening_get_next(struct s
25892 req = req->dl_next;
25895 + vxdprintk(VXD_CBIT(net, 6),
25896 + "sk,req: %p [#%d] (from %d)", req->sk,
25897 + (req->sk)?req->sk->sk_nid:0, nx_current_nid());
25899 + !nx_check(req->sk->sk_nid, VS_WATCH_P | VS_IDENT))
25901 if (req->rsk_ops->family == st->family) {
25904 @@ -2041,6 +2047,10 @@ get_req:
25907 sk_nulls_for_each_from(sk, node) {
25908 + vxdprintk(VXD_CBIT(net, 6), "sk: %p [#%d] (from %d)",
25909 + sk, sk->sk_nid, nx_current_nid());
25910 + if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
25912 if (sk->sk_family == st->family && net_eq(sock_net(sk), net)) {
25915 @@ -2115,6 +2125,11 @@ static void *established_get_first(struc
25917 spin_lock_bh(lock);
25918 sk_nulls_for_each(sk, node, &tcp_hashinfo.ehash[st->bucket].chain) {
25919 + vxdprintk(VXD_CBIT(net, 6),
25920 + "sk,egf: %p [#%d] (from %d)",
25921 + sk, sk->sk_nid, nx_current_nid());
25922 + if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
25924 if (sk->sk_family != st->family ||
25925 !net_eq(sock_net(sk), net)) {
25927 @@ -2125,6 +2140,11 @@ static void *established_get_first(struc
25928 st->state = TCP_SEQ_STATE_TIME_WAIT;
25929 inet_twsk_for_each(tw, node,
25930 &tcp_hashinfo.ehash[st->bucket].twchain) {
25931 + vxdprintk(VXD_CBIT(net, 6),
25932 + "tw: %p [#%d] (from %d)",
25933 + tw, tw->tw_nid, nx_current_nid());
25934 + if (!nx_check(tw->tw_nid, VS_WATCH_P | VS_IDENT))
25936 if (tw->tw_family != st->family ||
25937 !net_eq(twsk_net(tw), net)) {
25939 @@ -2154,7 +2174,9 @@ static void *established_get_next(struct
25943 - while (tw && (tw->tw_family != st->family || !net_eq(twsk_net(tw), net))) {
25944 + while (tw && (tw->tw_family != st->family ||
25945 + !net_eq(twsk_net(tw), net) ||
25946 + !nx_check(tw->tw_nid, VS_WATCH_P | VS_IDENT))) {
25950 @@ -2178,6 +2200,11 @@ get_tw:
25951 sk = sk_nulls_next(sk);
25953 sk_nulls_for_each_from(sk, node) {
25954 + vxdprintk(VXD_CBIT(net, 6),
25955 + "sk,egn: %p [#%d] (from %d)",
25956 + sk, sk->sk_nid, nx_current_nid());
25957 + if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
25959 if (sk->sk_family == st->family && net_eq(sock_net(sk), net))
25962 @@ -2387,9 +2414,9 @@ static void get_openreq4(struct sock *sk
25963 seq_printf(f, "%4d: %08X:%04X %08X:%04X"
25964 " %02X %08X:%08X %02X:%08lX %08X %5d %8d %u %d %p%n",
25967 + nx_map_sock_lback(current_nx_info(), ireq->loc_addr),
25968 ntohs(inet_sk(sk)->inet_sport),
25970 + nx_map_sock_lback(current_nx_info(), ireq->rmt_addr),
25971 ntohs(ireq->rmt_port),
25973 0, 0, /* could print option size, but that is af dependent. */
25974 @@ -2441,7 +2468,10 @@ static void get_tcp4_sock(struct sock *s
25976 seq_printf(f, "%4d: %08X:%04X %08X:%04X %02X %08X:%08X %02X:%08lX "
25977 "%08X %5d %8d %lu %d %p %lu %lu %u %u %d%n",
25978 - i, src, srcp, dest, destp, sk->sk_state,
25980 + nx_map_sock_lback(current_nx_info(), src), srcp,
25981 + nx_map_sock_lback(current_nx_info(), dest), destp,
25983 tp->write_seq - tp->snd_una,
25986 @@ -2476,7 +2506,10 @@ static void get_timewait4_sock(struct in
25988 seq_printf(f, "%4d: %08X:%04X %08X:%04X"
25989 " %02X %08X:%08X %02X:%08lX %08X %5d %8d %d %d %p%n",
25990 - i, src, srcp, dest, destp, tw->tw_substate, 0, 0,
25992 + nx_map_sock_lback(current_nx_info(), src), srcp,
25993 + nx_map_sock_lback(current_nx_info(), dest), destp,
25994 + tw->tw_substate, 0, 0,
25995 3, jiffies_to_clock_t(ttd), 0, 0, 0, 0,
25996 atomic_read(&tw->tw_refcnt), tw, len);
25998 diff -NurpP --minimal linux-2.6.36/net/ipv4/tcp_minisocks.c linux-2.6.36-vs2.3.0.36.33/net/ipv4/tcp_minisocks.c
25999 --- linux-2.6.36/net/ipv4/tcp_minisocks.c 2010-10-21 13:07:58.000000000 +0200
26000 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv4/tcp_minisocks.c 2010-10-21 13:09:36.000000000 +0200
26002 #include <linux/slab.h>
26003 #include <linux/sysctl.h>
26004 #include <linux/workqueue.h>
26005 +#include <linux/vs_limit.h>
26006 +#include <linux/vs_socket.h>
26007 +#include <linux/vs_context.h>
26008 #include <net/tcp.h>
26009 #include <net/inet_common.h>
26010 #include <net/xfrm.h>
26011 @@ -290,6 +293,11 @@ void tcp_time_wait(struct sock *sk, int
26012 tcptw->tw_ts_recent = tp->rx_opt.ts_recent;
26013 tcptw->tw_ts_recent_stamp = tp->rx_opt.ts_recent_stamp;
26015 + tw->tw_xid = sk->sk_xid;
26016 + tw->tw_vx_info = NULL;
26017 + tw->tw_nid = sk->sk_nid;
26018 + tw->tw_nx_info = NULL;
26020 #if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
26021 if (tw->tw_family == PF_INET6) {
26022 struct ipv6_pinfo *np = inet6_sk(sk);
26023 diff -NurpP --minimal linux-2.6.36/net/ipv4/udp.c linux-2.6.36-vs2.3.0.36.33/net/ipv4/udp.c
26024 --- linux-2.6.36/net/ipv4/udp.c 2010-10-21 13:07:59.000000000 +0200
26025 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv4/udp.c 2010-10-21 13:09:36.000000000 +0200
26026 @@ -296,14 +296,7 @@ fail:
26028 EXPORT_SYMBOL(udp_lib_get_port);
26030 -static int ipv4_rcv_saddr_equal(const struct sock *sk1, const struct sock *sk2)
26032 - struct inet_sock *inet1 = inet_sk(sk1), *inet2 = inet_sk(sk2);
26034 - return (!ipv6_only_sock(sk2) &&
26035 - (!inet1->inet_rcv_saddr || !inet2->inet_rcv_saddr ||
26036 - inet1->inet_rcv_saddr == inet2->inet_rcv_saddr));
26038 +extern int ipv4_rcv_saddr_equal(const struct sock *, const struct sock *);
26040 static unsigned int udp4_portaddr_hash(struct net *net, __be32 saddr,
26042 @@ -338,6 +331,11 @@ static inline int compute_score(struct s
26043 if (inet->inet_rcv_saddr != daddr)
26047 + /* block non nx_info ips */
26048 + if (!v4_addr_in_nx_info(sk->sk_nx_info,
26049 + daddr, NXA_MASK_BIND))
26052 if (inet->inet_daddr) {
26053 if (inet->inet_daddr != saddr)
26054 @@ -441,6 +439,7 @@ exact_match:
26059 /* UDP is nearly always wildcards out the wazoo, it makes no sense to try
26060 * harder than this. -DaveM
26062 @@ -486,6 +485,11 @@ begin:
26063 sk_nulls_for_each_rcu(sk, node, &hslot->head) {
26064 score = compute_score(sk, net, saddr, hnum, sport,
26065 daddr, dport, dif);
26066 + /* FIXME: disabled?
26067 + if (score == 9) {
26071 if (score > badness) {
26074 @@ -499,6 +503,7 @@ begin:
26075 if (get_nulls_value(node) != slot)
26080 if (unlikely(!atomic_inc_not_zero(&result->sk_refcnt)))
26082 @@ -508,6 +513,7 @@ begin:
26090 @@ -550,8 +556,7 @@ static inline struct sock *udp_v4_mcast_
26091 udp_sk(s)->udp_port_hash != hnum ||
26092 (inet->inet_daddr && inet->inet_daddr != rmt_addr) ||
26093 (inet->inet_dport != rmt_port && inet->inet_dport) ||
26094 - (inet->inet_rcv_saddr &&
26095 - inet->inet_rcv_saddr != loc_addr) ||
26096 + !v4_sock_addr_match(sk->sk_nx_info, inet, loc_addr) ||
26097 ipv6_only_sock(s) ||
26098 (s->sk_bound_dev_if && s->sk_bound_dev_if != dif))
26100 @@ -900,8 +905,13 @@ int udp_sendmsg(struct kiocb *iocb, stru
26101 { .sport = inet->inet_sport,
26102 .dport = dport } } };
26103 struct net *net = sock_net(sk);
26104 + struct nx_info *nxi = sk->sk_nx_info;
26106 security_sk_classify_flow(sk, &fl);
26107 + err = ip_v4_find_src(net, nxi, &rt, &fl);
26111 err = ip_route_output_flow(net, &rt, &fl, sk, 1);
26113 if (err == -ENETUNREACH)
26114 @@ -1183,7 +1193,8 @@ try_again:
26116 sin->sin_family = AF_INET;
26117 sin->sin_port = udp_hdr(skb)->source;
26118 - sin->sin_addr.s_addr = ip_hdr(skb)->saddr;
26119 + sin->sin_addr.s_addr = nx_map_sock_lback(
26120 + skb->sk->sk_nx_info, ip_hdr(skb)->saddr);
26121 memset(sin->sin_zero, 0, sizeof(sin->sin_zero));
26123 if (inet->cmsg_flags)
26124 @@ -1923,6 +1934,8 @@ static struct sock *udp_get_first(struct
26125 sk_nulls_for_each(sk, node, &hslot->head) {
26126 if (!net_eq(sock_net(sk), net))
26128 + if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
26130 if (sk->sk_family == state->family)
26133 @@ -1940,7 +1953,9 @@ static struct sock *udp_get_next(struct
26136 sk = sk_nulls_next(sk);
26137 - } while (sk && (!net_eq(sock_net(sk), net) || sk->sk_family != state->family));
26138 + } while (sk && (!net_eq(sock_net(sk), net) ||
26139 + sk->sk_family != state->family ||
26140 + !nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT)));
26143 if (state->bucket <= state->udp_table->mask)
26144 @@ -2047,7 +2062,10 @@ static void udp4_format_sock(struct sock
26146 seq_printf(f, "%5d: %08X:%04X %08X:%04X"
26147 " %02X %08X:%08X %02X:%08lX %08X %5d %8d %lu %d %p %d%n",
26148 - bucket, src, srcp, dest, destp, sp->sk_state,
26150 + nx_map_sock_lback(current_nx_info(), src), srcp,
26151 + nx_map_sock_lback(current_nx_info(), dest), destp,
26153 sk_wmem_alloc_get(sp),
26154 sk_rmem_alloc_get(sp),
26155 0, 0L, 0, sock_i_uid(sp), 0, sock_i_ino(sp),
26156 diff -NurpP --minimal linux-2.6.36/net/ipv6/addrconf.c linux-2.6.36-vs2.3.0.36.33/net/ipv6/addrconf.c
26157 --- linux-2.6.36/net/ipv6/addrconf.c 2010-10-21 13:07:59.000000000 +0200
26158 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv6/addrconf.c 2010-10-21 13:09:36.000000000 +0200
26161 #include <linux/proc_fs.h>
26162 #include <linux/seq_file.h>
26163 +#include <linux/vs_network.h>
26164 +#include <linux/vs_inet6.h>
26166 /* Set to 3 to get tracing... */
26167 #define ACONF_DEBUG 2
26168 @@ -1119,7 +1121,7 @@ out:
26170 int ipv6_dev_get_saddr(struct net *net, struct net_device *dst_dev,
26171 const struct in6_addr *daddr, unsigned int prefs,
26172 - struct in6_addr *saddr)
26173 + struct in6_addr *saddr, struct nx_info *nxi)
26175 struct ipv6_saddr_score scores[2],
26176 *score = &scores[0], *hiscore = &scores[1];
26177 @@ -1191,6 +1193,8 @@ int ipv6_dev_get_saddr(struct net *net,
26181 + if (!v6_addr_in_nx_info(nxi, &score->ifa->addr, -1))
26185 bitmap_zero(score->scorebits, IPV6_SADDR_RULE_MAX);
26186 @@ -3077,7 +3081,10 @@ static void if6_seq_stop(struct seq_file
26187 static int if6_seq_show(struct seq_file *seq, void *v)
26189 struct inet6_ifaddr *ifp = (struct inet6_ifaddr *)v;
26190 - seq_printf(seq, "%pi6 %02x %02x %02x %02x %8s\n",
26192 + if (nx_check(0, VS_ADMIN|VS_WATCH) ||
26193 + v6_addr_in_nx_info(current_nx_info(), &ifp->addr, -1))
26194 + seq_printf(seq, "%pi6 %02x %02x %02x %02x %8s\n",
26196 ifp->idev->dev->ifindex,
26198 @@ -3585,6 +3592,11 @@ static int in6_dump_addrs(struct inet6_d
26199 struct ifacaddr6 *ifaca;
26201 int ip_idx = *p_ip_idx;
26202 + struct nx_info *nxi = skb->sk ? skb->sk->sk_nx_info : NULL;
26204 + /* disable ipv6 on non v6 guests */
26205 + if (nxi && !nx_info_has_v6(nxi))
26208 read_lock_bh(&idev->lock);
26210 @@ -3595,6 +3607,8 @@ static int in6_dump_addrs(struct inet6_d
26211 list_for_each_entry(ifa, &idev->addr_list, if_list) {
26212 if (++ip_idx < s_ip_idx)
26214 + if (!v6_addr_in_nx_info(nxi, &ifa->addr, -1))
26216 err = inet6_fill_ifaddr(skb, ifa,
26217 NETLINK_CB(cb->skb).pid,
26218 cb->nlh->nlmsg_seq,
26219 @@ -3611,6 +3625,8 @@ static int in6_dump_addrs(struct inet6_d
26220 ifmca = ifmca->next, ip_idx++) {
26221 if (ip_idx < s_ip_idx)
26223 + if (!v6_addr_in_nx_info(nxi, &ifmca->mca_addr, -1))
26225 err = inet6_fill_ifmcaddr(skb, ifmca,
26226 NETLINK_CB(cb->skb).pid,
26227 cb->nlh->nlmsg_seq,
26228 @@ -3626,6 +3642,8 @@ static int in6_dump_addrs(struct inet6_d
26229 ifaca = ifaca->aca_next, ip_idx++) {
26230 if (ip_idx < s_ip_idx)
26232 + if (!v6_addr_in_nx_info(nxi, &ifaca->aca_addr, -1))
26234 err = inet6_fill_ifacaddr(skb, ifaca,
26235 NETLINK_CB(cb->skb).pid,
26236 cb->nlh->nlmsg_seq,
26237 @@ -3973,6 +3991,11 @@ static int inet6_dump_ifinfo(struct sk_b
26238 struct inet6_dev *idev;
26239 struct hlist_head *head;
26240 struct hlist_node *node;
26241 + struct nx_info *nxi = skb->sk ? skb->sk->sk_nx_info : NULL;
26243 + /* FIXME: maybe disable ipv6 on non v6 guests?
26244 + if (skb->sk && skb->sk->sk_vx_info)
26245 + return skb->len; */
26248 s_idx = cb->args[1];
26249 @@ -3984,6 +4007,8 @@ static int inet6_dump_ifinfo(struct sk_b
26250 hlist_for_each_entry_rcu(dev, node, head, index_hlist) {
26253 + if (!v6_dev_in_nx_info(dev, nxi))
26255 idev = __in6_dev_get(dev);
26258 diff -NurpP --minimal linux-2.6.36/net/ipv6/af_inet6.c linux-2.6.36-vs2.3.0.36.33/net/ipv6/af_inet6.c
26259 --- linux-2.6.36/net/ipv6/af_inet6.c 2010-10-21 13:07:59.000000000 +0200
26260 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv6/af_inet6.c 2010-10-21 13:09:36.000000000 +0200
26262 #include <linux/netdevice.h>
26263 #include <linux/icmpv6.h>
26264 #include <linux/netfilter_ipv6.h>
26265 +#include <linux/vs_inet.h>
26266 +#include <linux/vs_inet6.h>
26268 #include <net/ip.h>
26269 #include <net/ipv6.h>
26270 @@ -160,9 +162,12 @@ lookup_protocol:
26274 + if ((protocol == IPPROTO_ICMPV6) &&
26275 + nx_capable(CAP_NET_RAW, NXC_RAW_ICMP))
26277 if (sock->type == SOCK_RAW && !kern && !capable(CAP_NET_RAW))
26278 goto out_rcu_unlock;
26281 sock->ops = answer->ops;
26282 answer_prot = answer->prot;
26283 answer_no_check = answer->no_check;
26284 @@ -261,6 +266,7 @@ int inet6_bind(struct socket *sock, stru
26285 struct inet_sock *inet = inet_sk(sk);
26286 struct ipv6_pinfo *np = inet6_sk(sk);
26287 struct net *net = sock_net(sk);
26288 + struct nx_v6_sock_addr nsa;
26290 unsigned short snum;
26292 @@ -272,6 +278,11 @@ int inet6_bind(struct socket *sock, stru
26294 if (addr_len < SIN6_LEN_RFC2133)
26297 + err = v6_map_sock_addr(inet, addr, &nsa);
26301 addr_type = ipv6_addr_type(&addr->sin6_addr);
26302 if ((addr_type & IPV6_ADDR_MULTICAST) && sock->type == SOCK_STREAM)
26304 @@ -303,6 +314,7 @@ int inet6_bind(struct socket *sock, stru
26305 /* Reproduce AF_INET checks to make the bindings consitant */
26306 v4addr = addr->sin6_addr.s6_addr32[3];
26307 chk_addr_ret = inet_addr_type(net, v4addr);
26309 if (!sysctl_ip_nonlocal_bind &&
26310 !(inet->freebind || inet->transparent) &&
26311 v4addr != htonl(INADDR_ANY) &&
26312 @@ -312,6 +324,10 @@ int inet6_bind(struct socket *sock, stru
26313 err = -EADDRNOTAVAIL;
26316 + if (!v4_addr_in_nx_info(sk->sk_nx_info, v4addr, NXA_MASK_BIND)) {
26317 + err = -EADDRNOTAVAIL;
26321 if (addr_type != IPV6_ADDR_ANY) {
26322 struct net_device *dev = NULL;
26323 @@ -338,6 +354,11 @@ int inet6_bind(struct socket *sock, stru
26327 + if (!v6_addr_in_nx_info(sk->sk_nx_info, &addr->sin6_addr, -1)) {
26328 + err = -EADDRNOTAVAIL;
26332 /* ipv4 addr of the socket is invalid. Only the
26333 * unspecified and mapped address have a v4 equivalent.
26335 @@ -353,6 +374,9 @@ int inet6_bind(struct socket *sock, stru
26339 + /* what's that for? */
26340 + v6_set_sock_addr(inet, &nsa);
26342 inet->inet_rcv_saddr = v4addr;
26343 inet->inet_saddr = v4addr;
26345 @@ -454,9 +478,11 @@ int inet6_getname(struct socket *sock, s
26347 sin->sin6_port = inet->inet_dport;
26348 ipv6_addr_copy(&sin->sin6_addr, &np->daddr);
26349 + /* FIXME: remap lback? */
26351 sin->sin6_flowinfo = np->flow_label;
26353 + /* FIXME: remap lback? */
26354 if (ipv6_addr_any(&np->rcv_saddr))
26355 ipv6_addr_copy(&sin->sin6_addr, &np->saddr);
26357 diff -NurpP --minimal linux-2.6.36/net/ipv6/fib6_rules.c linux-2.6.36-vs2.3.0.36.33/net/ipv6/fib6_rules.c
26358 --- linux-2.6.36/net/ipv6/fib6_rules.c 2010-10-21 13:07:59.000000000 +0200
26359 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv6/fib6_rules.c 2010-10-21 13:09:36.000000000 +0200
26360 @@ -89,7 +89,7 @@ static int fib6_rule_action(struct fib_r
26361 ip6_dst_idev(&rt->dst)->dev,
26363 rt6_flags2srcprefs(flags),
26367 if (!ipv6_prefix_equal(&saddr, &r->src.addr,
26369 diff -NurpP --minimal linux-2.6.36/net/ipv6/inet6_hashtables.c linux-2.6.36-vs2.3.0.36.33/net/ipv6/inet6_hashtables.c
26370 --- linux-2.6.36/net/ipv6/inet6_hashtables.c 2010-02-25 11:52:10.000000000 +0100
26371 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv6/inet6_hashtables.c 2010-10-21 13:09:36.000000000 +0200
26374 #include <linux/module.h>
26375 #include <linux/random.h>
26376 +#include <linux/vs_inet6.h>
26378 #include <net/inet_connection_sock.h>
26379 #include <net/inet_hashtables.h>
26380 @@ -82,7 +83,6 @@ struct sock *__inet6_lookup_established(
26381 unsigned int slot = hash & hashinfo->ehash_mask;
26382 struct inet_ehash_bucket *head = &hashinfo->ehash[slot];
26387 sk_nulls_for_each_rcu(sk, node, &head->chain) {
26388 @@ -94,7 +94,7 @@ begin:
26396 if (get_nulls_value(node) != slot)
26397 @@ -140,6 +140,9 @@ static int inline compute_score(struct s
26398 if (!ipv6_addr_equal(&np->rcv_saddr, daddr))
26402 + if (!v6_addr_in_nx_info(sk->sk_nx_info, daddr, -1))
26405 if (sk->sk_bound_dev_if) {
26406 if (sk->sk_bound_dev_if != dif)
26407 diff -NurpP --minimal linux-2.6.36/net/ipv6/ip6_output.c linux-2.6.36-vs2.3.0.36.33/net/ipv6/ip6_output.c
26408 --- linux-2.6.36/net/ipv6/ip6_output.c 2010-10-21 13:07:59.000000000 +0200
26409 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv6/ip6_output.c 2010-10-21 13:09:36.000000000 +0200
26410 @@ -938,7 +938,7 @@ static int ip6_dst_lookup_tail(struct so
26411 err = ipv6_dev_get_saddr(net, ip6_dst_idev(*dst)->dev,
26413 sk ? inet6_sk(sk)->srcprefs : 0,
26415 + &fl->fl6_src, sk->sk_nx_info);
26417 goto out_err_release;
26419 diff -NurpP --minimal linux-2.6.36/net/ipv6/Kconfig linux-2.6.36-vs2.3.0.36.33/net/ipv6/Kconfig
26420 --- linux-2.6.36/net/ipv6/Kconfig 2010-08-02 16:52:59.000000000 +0200
26421 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv6/Kconfig 2010-10-21 13:09:36.000000000 +0200
26424 # IPv6 as module will cause a CRASH if you try to unload it
26426 - tristate "The IPv6 protocol"
26428 + bool "The IPv6 protocol"
26431 This is complemental support for the IP version 6.
26432 You will still be able to do traditional IPv4 networking as well.
26433 diff -NurpP --minimal linux-2.6.36/net/ipv6/ndisc.c linux-2.6.36-vs2.3.0.36.33/net/ipv6/ndisc.c
26434 --- linux-2.6.36/net/ipv6/ndisc.c 2010-10-21 13:07:59.000000000 +0200
26435 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv6/ndisc.c 2010-10-21 13:09:36.000000000 +0200
26436 @@ -591,7 +591,7 @@ static void ndisc_send_na(struct net_dev
26438 if (ipv6_dev_get_saddr(dev_net(dev), dev, daddr,
26439 inet6_sk(dev_net(dev)->ipv6.ndisc_sk)->srcprefs,
26441 + &tmpaddr, NULL /* FIXME: ? */ ))
26443 src_addr = &tmpaddr;
26445 diff -NurpP --minimal linux-2.6.36/net/ipv6/raw.c linux-2.6.36-vs2.3.0.36.33/net/ipv6/raw.c
26446 --- linux-2.6.36/net/ipv6/raw.c 2010-10-21 13:07:59.000000000 +0200
26447 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv6/raw.c 2010-10-21 13:09:36.000000000 +0200
26449 #include <linux/icmpv6.h>
26450 #include <linux/netfilter.h>
26451 #include <linux/netfilter_ipv6.h>
26452 +#include <linux/vs_inet6.h>
26453 #include <linux/skbuff.h>
26454 #include <asm/uaccess.h>
26455 #include <asm/ioctls.h>
26456 @@ -283,6 +284,13 @@ static int rawv6_bind(struct sock *sk, s
26460 + if (!v6_addr_in_nx_info(sk->sk_nx_info, &addr->sin6_addr, -1)) {
26461 + err = -EADDRNOTAVAIL;
26467 /* ipv4 addr of the socket is invalid. Only the
26468 * unspecified and mapped address have a v4 equivalent.
26470 diff -NurpP --minimal linux-2.6.36/net/ipv6/route.c linux-2.6.36-vs2.3.0.36.33/net/ipv6/route.c
26471 --- linux-2.6.36/net/ipv6/route.c 2010-10-21 13:07:59.000000000 +0200
26472 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv6/route.c 2010-10-21 13:09:36.000000000 +0200
26473 @@ -2272,7 +2272,8 @@ static int rt6_fill_node(struct net *net
26474 struct inet6_dev *idev = ip6_dst_idev(&rt->dst);
26475 struct in6_addr saddr_buf;
26476 if (ipv6_dev_get_saddr(net, idev ? idev->dev : NULL,
26477 - dst, 0, &saddr_buf) == 0)
26478 + dst, 0, &saddr_buf,
26479 + (skb->sk ? skb->sk->sk_nx_info : NULL)) == 0)
26480 NLA_PUT(skb, RTA_PREFSRC, 16, &saddr_buf);
26483 diff -NurpP --minimal linux-2.6.36/net/ipv6/tcp_ipv6.c linux-2.6.36-vs2.3.0.36.33/net/ipv6/tcp_ipv6.c
26484 --- linux-2.6.36/net/ipv6/tcp_ipv6.c 2010-10-21 13:07:59.000000000 +0200
26485 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv6/tcp_ipv6.c 2010-10-21 13:09:36.000000000 +0200
26488 #include <linux/crypto.h>
26489 #include <linux/scatterlist.h>
26490 +#include <linux/vs_inet6.h>
26492 static void tcp_v6_send_reset(struct sock *sk, struct sk_buff *skb);
26493 static void tcp_v6_reqsk_send_ack(struct sock *sk, struct sk_buff *skb,
26494 @@ -160,8 +161,15 @@ static int tcp_v6_connect(struct sock *s
26495 * connect() to INADDR_ANY means loopback (BSD'ism).
26498 - if(ipv6_addr_any(&usin->sin6_addr))
26499 - usin->sin6_addr.s6_addr[15] = 0x1;
26500 + if(ipv6_addr_any(&usin->sin6_addr)) {
26501 + struct nx_info *nxi = sk->sk_nx_info;
26503 + if (nxi && nx_info_has_v6(nxi))
26504 + /* FIXME: remap lback? */
26505 + usin->sin6_addr = nxi->v6.ip;
26507 + usin->sin6_addr.s6_addr[15] = 0x1;
26510 addr_type = ipv6_addr_type(&usin->sin6_addr);
26512 diff -NurpP --minimal linux-2.6.36/net/ipv6/udp.c linux-2.6.36-vs2.3.0.36.33/net/ipv6/udp.c
26513 --- linux-2.6.36/net/ipv6/udp.c 2010-10-21 13:07:59.000000000 +0200
26514 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv6/udp.c 2010-10-21 13:09:36.000000000 +0200
26515 @@ -48,13 +48,14 @@
26517 #include <linux/proc_fs.h>
26518 #include <linux/seq_file.h>
26519 +#include <linux/vs_inet6.h>
26520 #include "udp_impl.h"
26522 int ipv6_rcv_saddr_equal(const struct sock *sk, const struct sock *sk2)
26524 const struct in6_addr *sk_rcv_saddr6 = &inet6_sk(sk)->rcv_saddr;
26525 const struct in6_addr *sk2_rcv_saddr6 = inet6_rcv_saddr(sk2);
26526 - __be32 sk1_rcv_saddr = inet_sk(sk)->inet_rcv_saddr;
26527 + __be32 sk_rcv_saddr = inet_sk(sk)->inet_rcv_saddr;
26528 __be32 sk2_rcv_saddr = inet_rcv_saddr(sk2);
26529 int sk_ipv6only = ipv6_only_sock(sk);
26530 int sk2_ipv6only = inet_v6_ipv6only(sk2);
26531 @@ -62,24 +63,49 @@ int ipv6_rcv_saddr_equal(const struct so
26532 int addr_type2 = sk2_rcv_saddr6 ? ipv6_addr_type(sk2_rcv_saddr6) : IPV6_ADDR_MAPPED;
26534 /* if both are mapped, treat as IPv4 */
26535 - if (addr_type == IPV6_ADDR_MAPPED && addr_type2 == IPV6_ADDR_MAPPED)
26536 - return (!sk2_ipv6only &&
26537 - (!sk1_rcv_saddr || !sk2_rcv_saddr ||
26538 - sk1_rcv_saddr == sk2_rcv_saddr));
26539 + if (addr_type == IPV6_ADDR_MAPPED && addr_type2 == IPV6_ADDR_MAPPED) {
26540 + if (!sk2_ipv6only &&
26541 + (!sk_rcv_saddr || !sk2_rcv_saddr ||
26542 + sk_rcv_saddr == sk2_rcv_saddr))
26548 if (addr_type2 == IPV6_ADDR_ANY &&
26549 !(sk2_ipv6only && addr_type == IPV6_ADDR_MAPPED))
26553 if (addr_type == IPV6_ADDR_ANY &&
26554 !(sk_ipv6only && addr_type2 == IPV6_ADDR_MAPPED))
26558 if (sk2_rcv_saddr6 &&
26559 ipv6_addr_equal(sk_rcv_saddr6, sk2_rcv_saddr6))
26566 + if (!sk_rcv_saddr && !sk2_rcv_saddr)
26567 + return nx_v4_addr_conflict(sk->sk_nx_info, sk2->sk_nx_info);
26568 + if (!sk2_rcv_saddr)
26569 + return v4_addr_in_nx_info(sk->sk_nx_info, sk2_rcv_saddr, -1);
26570 + if (!sk_rcv_saddr)
26571 + return v4_addr_in_nx_info(sk2->sk_nx_info, sk_rcv_saddr, -1);
26574 + if (addr_type2 == IPV6_ADDR_ANY && addr_type == IPV6_ADDR_ANY)
26575 + return nx_v6_addr_conflict(sk->sk_nx_info, sk2->sk_nx_info);
26576 + else if (addr_type2 == IPV6_ADDR_ANY)
26577 + return v6_addr_in_nx_info(sk2->sk_nx_info, sk_rcv_saddr6, -1);
26578 + else if (addr_type == IPV6_ADDR_ANY) {
26579 + if (addr_type2 == IPV6_ADDR_MAPPED)
26580 + return nx_v4_addr_conflict(sk->sk_nx_info, sk2->sk_nx_info);
26582 + return v6_addr_in_nx_info(sk->sk_nx_info, sk2_rcv_saddr6, -1);
26587 static unsigned int udp6_portaddr_hash(struct net *net,
26588 @@ -143,6 +169,10 @@ static inline int compute_score(struct s
26589 if (!ipv6_addr_equal(&np->rcv_saddr, daddr))
26593 + /* block non nx_info ips */
26594 + if (!v6_addr_in_nx_info(sk->sk_nx_info, daddr, -1))
26597 if (!ipv6_addr_any(&np->daddr)) {
26598 if (!ipv6_addr_equal(&np->daddr, saddr))
26599 diff -NurpP --minimal linux-2.6.36/net/ipv6/xfrm6_policy.c linux-2.6.36-vs2.3.0.36.33/net/ipv6/xfrm6_policy.c
26600 --- linux-2.6.36/net/ipv6/xfrm6_policy.c 2010-08-02 16:53:00.000000000 +0200
26601 +++ linux-2.6.36-vs2.3.0.36.33/net/ipv6/xfrm6_policy.c 2010-10-21 13:09:36.000000000 +0200
26602 @@ -62,7 +62,7 @@ static int xfrm6_get_saddr(struct net *n
26603 dev = ip6_dst_idev(dst)->dev;
26604 ipv6_dev_get_saddr(dev_net(dev), dev,
26605 (struct in6_addr *)&daddr->a6, 0,
26606 - (struct in6_addr *)&saddr->a6);
26607 + (struct in6_addr *)&saddr->a6, NULL);
26611 diff -NurpP --minimal linux-2.6.36/net/netlink/af_netlink.c linux-2.6.36-vs2.3.0.36.33/net/netlink/af_netlink.c
26612 --- linux-2.6.36/net/netlink/af_netlink.c 2010-10-21 13:08:01.000000000 +0200
26613 +++ linux-2.6.36-vs2.3.0.36.33/net/netlink/af_netlink.c 2010-10-21 13:09:36.000000000 +0200
26615 #include <linux/types.h>
26616 #include <linux/audit.h>
26617 #include <linux/mutex.h>
26618 +#include <linux/vs_context.h>
26619 +#include <linux/vs_network.h>
26620 +#include <linux/vs_limit.h>
26622 #include <net/net_namespace.h>
26623 #include <net/sock.h>
26624 @@ -1928,6 +1931,8 @@ static struct sock *netlink_seq_socket_i
26625 sk_for_each(s, node, &hash->table[j]) {
26626 if (sock_net(s) != seq_file_net(seq))
26628 + if (!nx_check(s->sk_nid, VS_WATCH_P | VS_IDENT))
26632 iter->hash_idx = j;
26633 @@ -1962,7 +1967,8 @@ static void *netlink_seq_next(struct seq
26637 - } while (s && sock_net(s) != seq_file_net(seq));
26638 + } while (s && (sock_net(s) != seq_file_net(seq) ||
26639 + !nx_check(s->sk_nid, VS_WATCH_P | VS_IDENT)));
26643 @@ -1974,7 +1980,8 @@ static void *netlink_seq_next(struct seq
26645 for (; j <= hash->mask; j++) {
26646 s = sk_head(&hash->table[j]);
26647 - while (s && sock_net(s) != seq_file_net(seq))
26648 + while (s && (sock_net(s) != seq_file_net(seq) ||
26649 + !nx_check(s->sk_nid, VS_WATCH_P | VS_IDENT)))
26653 diff -NurpP --minimal linux-2.6.36/net/sctp/ipv6.c linux-2.6.36-vs2.3.0.36.33/net/sctp/ipv6.c
26654 --- linux-2.6.36/net/sctp/ipv6.c 2010-08-02 16:53:01.000000000 +0200
26655 +++ linux-2.6.36-vs2.3.0.36.33/net/sctp/ipv6.c 2010-10-21 13:09:36.000000000 +0200
26656 @@ -304,7 +304,8 @@ static void sctp_v6_get_saddr(struct sct
26657 dst ? ip6_dst_idev(dst)->dev : NULL,
26658 &daddr->v6.sin6_addr,
26659 inet6_sk(&sk->inet.sk)->srcprefs,
26660 - &saddr->v6.sin6_addr);
26661 + &saddr->v6.sin6_addr,
26662 + asoc->base.sk->sk_nx_info);
26663 SCTP_DEBUG_PRINTK("saddr from ipv6_get_saddr: %pI6\n",
26664 &saddr->v6.sin6_addr);
26666 diff -NurpP --minimal linux-2.6.36/net/socket.c linux-2.6.36-vs2.3.0.36.33/net/socket.c
26667 --- linux-2.6.36/net/socket.c 2010-10-21 13:08:01.000000000 +0200
26668 +++ linux-2.6.36-vs2.3.0.36.33/net/socket.c 2010-10-21 13:09:36.000000000 +0200
26671 #include <net/sock.h>
26672 #include <linux/netfilter.h>
26673 +#include <linux/vs_base.h>
26674 +#include <linux/vs_socket.h>
26675 +#include <linux/vs_inet.h>
26676 +#include <linux/vs_inet6.h>
26678 #include <linux/if_tun.h>
26679 #include <linux/ipv6_route.h>
26680 @@ -551,7 +555,7 @@ static inline int __sock_sendmsg(struct
26681 struct msghdr *msg, size_t size)
26683 struct sock_iocb *si = kiocb_to_siocb(iocb);
26687 sock_update_classid(sock->sk);
26689 @@ -564,7 +568,22 @@ static inline int __sock_sendmsg(struct
26693 - return sock->ops->sendmsg(iocb, sock, msg, size);
26694 + len = sock->ops->sendmsg(iocb, sock, msg, size);
26697 + vx_sock_send(sock->sk, size);
26699 + vx_sock_fail(sock->sk, size);
26701 + vxdprintk(VXD_CBIT(net, 7),
26702 + "__sock_sendmsg: %p[%p,%p,%p;%d/%d]:%d/%d",
26704 + (sock->sk)?sock->sk->sk_nx_info:0,
26705 + (sock->sk)?sock->sk->sk_vx_info:0,
26706 + (sock->sk)?sock->sk->sk_xid:0,
26707 + (sock->sk)?sock->sk->sk_nid:0,
26708 + (unsigned int)size, len);
26712 int sock_sendmsg(struct socket *sock, struct msghdr *msg, size_t size)
26713 @@ -681,6 +700,7 @@ static inline int __sock_recvmsg_nosec(s
26714 struct msghdr *msg, size_t size, int flags)
26716 struct sock_iocb *si = kiocb_to_siocb(iocb);
26719 sock_update_classid(sock->sk);
26721 @@ -690,7 +710,18 @@ static inline int __sock_recvmsg_nosec(s
26725 - return sock->ops->recvmsg(iocb, sock, msg, size, flags);
26726 + len = sock->ops->recvmsg(iocb, sock, msg, size, flags);
26727 + if ((len >= 0) && sock->sk)
26728 + vx_sock_recv(sock->sk, len);
26729 + vxdprintk(VXD_CBIT(net, 7),
26730 + "__sock_recvmsg: %p[%p,%p,%p;%d/%d]:%d/%d",
26732 + (sock->sk)?sock->sk->sk_nx_info:0,
26733 + (sock->sk)?sock->sk->sk_vx_info:0,
26734 + (sock->sk)?sock->sk->sk_xid:0,
26735 + (sock->sk)?sock->sk->sk_nid:0,
26736 + (unsigned int)size, len);
26740 static inline int __sock_recvmsg(struct kiocb *iocb, struct socket *sock,
26741 @@ -1159,6 +1190,13 @@ static int __sock_create(struct net *net
26742 if (type < 0 || type >= SOCK_MAX)
26745 + if (!nx_check(0, VS_ADMIN)) {
26746 + if (family == PF_INET && !current_nx_info_has_v4())
26747 + return -EAFNOSUPPORT;
26748 + if (family == PF_INET6 && !current_nx_info_has_v6())
26749 + return -EAFNOSUPPORT;
26754 This uglymoron is moved from INET layer to here to avoid
26755 @@ -1293,6 +1331,7 @@ SYSCALL_DEFINE3(socket, int, family, int
26759 + set_bit(SOCK_USER_SOCKET, &sock->flags);
26760 retval = sock_map_fd(sock, flags & (O_CLOEXEC | O_NONBLOCK));
26763 @@ -1334,10 +1373,12 @@ SYSCALL_DEFINE4(socketpair, int, family,
26764 err = sock_create(family, type, protocol, &sock1);
26767 + set_bit(SOCK_USER_SOCKET, &sock1->flags);
26769 err = sock_create(family, type, protocol, &sock2);
26771 goto out_release_1;
26772 + set_bit(SOCK_USER_SOCKET, &sock2->flags);
26774 err = sock1->ops->socketpair(sock1, sock2);
26776 diff -NurpP --minimal linux-2.6.36/net/sunrpc/auth.c linux-2.6.36-vs2.3.0.36.33/net/sunrpc/auth.c
26777 --- linux-2.6.36/net/sunrpc/auth.c 2010-10-21 13:08:01.000000000 +0200
26778 +++ linux-2.6.36-vs2.3.0.36.33/net/sunrpc/auth.c 2010-10-21 13:09:36.000000000 +0200
26780 #include <linux/hash.h>
26781 #include <linux/sunrpc/clnt.h>
26782 #include <linux/spinlock.h>
26783 +#include <linux/vs_tag.h>
26786 # define RPCDBG_FACILITY RPCDBG_AUTH
26787 @@ -425,6 +426,7 @@ rpcauth_lookupcred(struct rpc_auth *auth
26788 memset(&acred, 0, sizeof(acred));
26789 acred.uid = cred->fsuid;
26790 acred.gid = cred->fsgid;
26791 + acred.tag = dx_current_tag();
26792 acred.group_info = get_group_info(((struct cred *)cred)->group_info);
26794 ret = auth->au_ops->lookup_cred(auth, &acred, flags);
26795 @@ -465,6 +467,7 @@ rpcauth_bind_root_cred(struct rpc_task *
26796 struct auth_cred acred = {
26799 + .tag = dx_current_tag(),
26802 dprintk("RPC: %5u looking up %s cred\n",
26803 diff -NurpP --minimal linux-2.6.36/net/sunrpc/auth_unix.c linux-2.6.36-vs2.3.0.36.33/net/sunrpc/auth_unix.c
26804 --- linux-2.6.36/net/sunrpc/auth_unix.c 2010-10-21 13:08:01.000000000 +0200
26805 +++ linux-2.6.36-vs2.3.0.36.33/net/sunrpc/auth_unix.c 2010-10-21 13:09:36.000000000 +0200
26806 @@ -12,12 +12,14 @@
26807 #include <linux/module.h>
26808 #include <linux/sunrpc/clnt.h>
26809 #include <linux/sunrpc/auth.h>
26810 +#include <linux/vs_tag.h>
26812 #define NFS_NGROUPS 16
26815 struct rpc_cred uc_base;
26818 gid_t uc_gids[NFS_NGROUPS];
26820 #define uc_uid uc_base.cr_uid
26821 @@ -78,6 +80,7 @@ unx_create_cred(struct rpc_auth *auth, s
26822 groups = NFS_NGROUPS;
26824 cred->uc_gid = acred->gid;
26825 + cred->uc_tag = acred->tag;
26826 for (i = 0; i < groups; i++)
26827 cred->uc_gids[i] = GROUP_AT(acred->group_info, i);
26828 if (i < NFS_NGROUPS)
26829 @@ -119,7 +122,9 @@ unx_match(struct auth_cred *acred, struc
26833 - if (cred->uc_uid != acred->uid || cred->uc_gid != acred->gid)
26834 + if (cred->uc_uid != acred->uid ||
26835 + cred->uc_gid != acred->gid ||
26836 + cred->uc_tag != acred->tag)
26839 if (acred->group_info != NULL)
26840 @@ -142,7 +147,7 @@ unx_marshal(struct rpc_task *task, __be3
26841 struct rpc_clnt *clnt = task->tk_client;
26842 struct unx_cred *cred = container_of(task->tk_rqstp->rq_cred, struct unx_cred, uc_base);
26843 __be32 *base, *hold;
26847 *p++ = htonl(RPC_AUTH_UNIX);
26849 @@ -152,9 +157,12 @@ unx_marshal(struct rpc_task *task, __be3
26850 * Copy the UTS nodename captured when the client was created.
26852 p = xdr_encode_array(p, clnt->cl_nodename, clnt->cl_nodelen);
26853 + tag = task->tk_client->cl_tag;
26855 - *p++ = htonl((u32) cred->uc_uid);
26856 - *p++ = htonl((u32) cred->uc_gid);
26857 + *p++ = htonl((u32) TAGINO_UID(tag,
26858 + cred->uc_uid, cred->uc_tag));
26859 + *p++ = htonl((u32) TAGINO_GID(tag,
26860 + cred->uc_gid, cred->uc_tag));
26862 for (i = 0; i < 16 && cred->uc_gids[i] != (gid_t) NOGROUP; i++)
26863 *p++ = htonl((u32) cred->uc_gids[i]);
26864 diff -NurpP --minimal linux-2.6.36/net/sunrpc/clnt.c linux-2.6.36-vs2.3.0.36.33/net/sunrpc/clnt.c
26865 --- linux-2.6.36/net/sunrpc/clnt.c 2010-10-21 13:08:01.000000000 +0200
26866 +++ linux-2.6.36-vs2.3.0.36.33/net/sunrpc/clnt.c 2010-10-21 13:09:36.000000000 +0200
26868 #include <linux/utsname.h>
26869 #include <linux/workqueue.h>
26870 #include <linux/in6.h>
26871 +#include <linux/vs_cvirt.h>
26873 #include <linux/sunrpc/clnt.h>
26874 #include <linux/sunrpc/rpc_pipe_fs.h>
26875 @@ -358,6 +359,9 @@ struct rpc_clnt *rpc_create(struct rpc_c
26876 if (!(args->flags & RPC_CLNT_CREATE_QUIET))
26877 clnt->cl_chatty = 1;
26879 + /* TODO: handle RPC_CLNT_CREATE_TAGGED
26880 + if (args->flags & RPC_CLNT_CREATE_TAGGED)
26881 + clnt->cl_tag = 1; */
26884 EXPORT_SYMBOL_GPL(rpc_create);
26885 diff -NurpP --minimal linux-2.6.36/net/unix/af_unix.c linux-2.6.36-vs2.3.0.36.33/net/unix/af_unix.c
26886 --- linux-2.6.36/net/unix/af_unix.c 2010-10-21 13:08:01.000000000 +0200
26887 +++ linux-2.6.36-vs2.3.0.36.33/net/unix/af_unix.c 2010-10-21 13:09:36.000000000 +0200
26888 @@ -114,6 +114,8 @@
26889 #include <linux/mount.h>
26890 #include <net/checksum.h>
26891 #include <linux/security.h>
26892 +#include <linux/vs_context.h>
26893 +#include <linux/vs_limit.h>
26895 static struct hlist_head unix_socket_table[UNIX_HASH_SIZE + 1];
26896 static DEFINE_SPINLOCK(unix_table_lock);
26897 @@ -258,6 +260,8 @@ static struct sock *__unix_find_socket_b
26898 if (!net_eq(sock_net(s), net))
26901 + if (!nx_check(s->sk_nid, VS_WATCH_P | VS_IDENT))
26903 if (u->addr->len == len &&
26904 !memcmp(u->addr->name, sunname, len))
26906 @@ -2148,6 +2152,8 @@ static struct sock *unix_seq_idx(struct
26907 for (s = first_unix_socket(&iter->i); s; s = next_unix_socket(&iter->i, s)) {
26908 if (sock_net(s) != seq_file_net(seq))
26910 + if (!nx_check(s->sk_nid, VS_WATCH_P | VS_IDENT))
26915 @@ -2172,7 +2178,8 @@ static void *unix_seq_next(struct seq_fi
26916 sk = first_unix_socket(&iter->i);
26918 sk = next_unix_socket(&iter->i, sk);
26919 - while (sk && (sock_net(sk) != seq_file_net(seq)))
26920 + while (sk && (sock_net(sk) != seq_file_net(seq) ||
26921 + !nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT)))
26922 sk = next_unix_socket(&iter->i, sk);
26925 diff -NurpP --minimal linux-2.6.36/scripts/checksyscalls.sh linux-2.6.36-vs2.3.0.36.33/scripts/checksyscalls.sh
26926 --- linux-2.6.36/scripts/checksyscalls.sh 2010-10-21 13:08:01.000000000 +0200
26927 +++ linux-2.6.36-vs2.3.0.36.33/scripts/checksyscalls.sh 2010-10-21 13:09:36.000000000 +0200
26928 @@ -193,7 +193,6 @@ cat << EOF
26929 #define __IGNORE_afs_syscall
26930 #define __IGNORE_getpmsg
26931 #define __IGNORE_putpmsg
26932 -#define __IGNORE_vserver
26936 diff -NurpP --minimal linux-2.6.36/security/commoncap.c linux-2.6.36-vs2.3.0.36.33/security/commoncap.c
26937 --- linux-2.6.36/security/commoncap.c 2010-10-21 13:08:02.000000000 +0200
26938 +++ linux-2.6.36-vs2.3.0.36.33/security/commoncap.c 2010-10-21 13:09:36.000000000 +0200
26940 #include <linux/prctl.h>
26941 #include <linux/securebits.h>
26942 #include <linux/syslog.h>
26943 +#include <linux/vs_context.h>
26946 * If a non-root user executes a setuid-root binary in
26947 @@ -53,7 +54,7 @@ static void warn_setuid_and_fcaps_mixed(
26949 int cap_netlink_send(struct sock *sk, struct sk_buff *skb)
26951 - NETLINK_CB(skb).eff_cap = current_cap();
26952 + NETLINK_CB(skb).eff_cap = vx_mbcaps(current_cap());
26956 @@ -63,6 +64,7 @@ int cap_netlink_recv(struct sk_buff *skb
26961 EXPORT_SYMBOL(cap_netlink_recv);
26964 @@ -83,7 +85,22 @@ EXPORT_SYMBOL(cap_netlink_recv);
26965 int cap_capable(struct task_struct *tsk, const struct cred *cred, int cap,
26968 - return cap_raised(cred->cap_effective, cap) ? 0 : -EPERM;
26969 + struct vx_info *vxi = tsk->vx_info;
26972 + printk("cap_capable() VXF_STATE_SETUP = %llx, raised = %x, eff = %08x:%08x\n",
26973 + vx_info_flags(vxi, VXF_STATE_SETUP, 0),
26974 + cap_raised(tsk->cap_effective, cap),
26975 + tsk->cap_effective.cap[1], tsk->cap_effective.cap[0]);
26978 + /* special case SETUP */
26979 + if (vx_info_flags(vxi, VXF_STATE_SETUP, 0) &&
26980 + /* FIXME: maybe use cred instead? */
26981 + cap_raised(tsk->cred->cap_effective, cap))
26984 + return vx_cap_raised(vxi, cred->cap_effective, cap) ? 0 : -EPERM;
26988 @@ -571,7 +588,7 @@ int cap_inode_setxattr(struct dentry *de
26990 if (!strncmp(name, XATTR_SECURITY_PREFIX,
26991 sizeof(XATTR_SECURITY_PREFIX) - 1) &&
26992 - !capable(CAP_SYS_ADMIN))
26993 + !vx_capable(CAP_SYS_ADMIN, VXC_FS_SECURITY))
26997 @@ -597,7 +614,7 @@ int cap_inode_removexattr(struct dentry
26999 if (!strncmp(name, XATTR_SECURITY_PREFIX,
27000 sizeof(XATTR_SECURITY_PREFIX) - 1) &&
27001 - !capable(CAP_SYS_ADMIN))
27002 + !vx_capable(CAP_SYS_ADMIN, VXC_FS_SECURITY))
27006 @@ -899,7 +916,8 @@ int cap_syslog(int type, bool from_file)
27007 if (type != SYSLOG_ACTION_OPEN && from_file)
27009 if ((type != SYSLOG_ACTION_READ_ALL &&
27010 - type != SYSLOG_ACTION_SIZE_BUFFER) && !capable(CAP_SYS_ADMIN))
27011 + type != SYSLOG_ACTION_SIZE_BUFFER) &&
27012 + !vx_capable(CAP_SYS_ADMIN, VXC_SYSLOG))
27016 @@ -951,3 +969,4 @@ int cap_file_mmap(struct file *file, uns
27021 diff -NurpP --minimal linux-2.6.36/security/selinux/hooks.c linux-2.6.36-vs2.3.0.36.33/security/selinux/hooks.c
27022 --- linux-2.6.36/security/selinux/hooks.c 2010-10-21 13:08:02.000000000 +0200
27023 +++ linux-2.6.36-vs2.3.0.36.33/security/selinux/hooks.c 2010-10-21 13:09:36.000000000 +0200
27025 #include <linux/dccp.h>
27026 #include <linux/quota.h>
27027 #include <linux/un.h> /* for Unix socket types */
27028 -#include <net/af_unix.h> /* for Unix socket types */
27029 #include <linux/parser.h>
27030 #include <linux/nfs_mount.h>
27031 #include <net/ipv6.h>
27032 diff -NurpP --minimal linux-2.6.36-vs2.3.0.36.33/fs/namei.c linux-2.6.36-vs2.3.0.36.34/fs/namei.c
27033 --- linux-2.6.36-vs2.3.0.36.33/fs/namei.c 2010-10-21 14:45:53.000000000 +0200
27034 +++ linux-2.6.36-vs2.3.0.36.34/fs/namei.c 2010-10-21 17:45:49.000000000 +0200
27035 @@ -848,7 +848,8 @@ done:
27039 - vxwprintk_task(1, "did lookup hidden %s:%p[#%d,%lu] »%s/%.*s«.",
27040 + vxwprintk_task(1,
27041 + "did lookup hidden %s:%p[#%d,%lu] " VS_Q("%s/%.*s") ".",
27042 inode->i_sb->s_id, inode, inode->i_tag, inode->i_ino,
27043 vxd_path(&nd->path), name->len, name->name);
27045 @@ -2968,7 +2969,8 @@ struct dentry *cow_break_link(const char
27046 char *to, *path, pad='\251';
27049 - vxdprintk(VXD_CBIT(misc, 1), "cow_break_link(»%s«)", pathname);
27050 + vxdprintk(VXD_CBIT(misc, 1),
27051 + "cow_break_link(" VS_Q("%s") ")", pathname);
27052 path = kmalloc(PATH_MAX, GFP_KERNEL);
27055 @@ -2985,7 +2987,8 @@ struct dentry *cow_break_link(const char
27057 to = d_path(&old_path, path, PATH_MAX-2);
27058 pathlen = strlen(to);
27059 - vxdprintk(VXD_CBIT(misc, 2), "old path »%s« [»%.*s«:%d]", to,
27060 + vxdprintk(VXD_CBIT(misc, 2),
27061 + "old path " VS_Q("%s") " [" VS_Q("%.*s") ":%d]", to,
27062 old_path.dentry->d_name.len, old_path.dentry->d_name.name,
27063 old_path.dentry->d_name.len);
27065 @@ -2996,7 +2999,7 @@ retry:
27069 - vxdprintk(VXD_CBIT(misc, 1), "temp copy »%s«", to);
27070 + vxdprintk(VXD_CBIT(misc, 1), "temp copy " VS_Q("%s"), to);
27071 /* dir_nd will have refs to dentry and mnt */
27072 ret = path_lookup(to,
27073 LOOKUP_PARENT | LOOKUP_OPEN | LOOKUP_CREATE, &dir_nd);
27074 @@ -3015,7 +3018,8 @@ retry:
27077 vxdprintk(VXD_CBIT(misc, 2),
27078 - "lookup_create(new): %p [»%.*s«:%d]", new_path.dentry,
27079 + "lookup_create(new): %p [" VS_Q("%.*s") ":%d]",
27081 new_path.dentry->d_name.len, new_path.dentry->d_name.name,
27082 new_path.dentry->d_name.len);
27083 dir = dir_nd.path.dentry;
27084 @@ -3091,7 +3095,7 @@ retry:
27087 vxdprintk(VXD_CBIT(misc, 2),
27088 - "vfs_rename: [»%*s«:%d] -> [»%*s«:%d]",
27089 + "vfs_rename: [" VS_Q("%*s") ":%d] -> [" VS_Q("%*s") ":%d]",
27090 new_path.dentry->d_name.len, new_path.dentry->d_name.name,
27091 new_path.dentry->d_name.len,
27092 old_path.dentry->d_name.len, old_path.dentry->d_name.name,
27093 @@ -3136,7 +3140,8 @@ out_redo:
27095 new_path.dentry = old_nd.path.dentry;
27096 vxdprintk(VXD_CBIT(misc, 2),
27097 - "path_lookup(redo): %p [»%.*s«:%d]", new_path.dentry,
27098 + "path_lookup(redo): %p [" VS_Q("%.*s") ":%d]",
27100 new_path.dentry->d_name.len, new_path.dentry->d_name.name,
27101 new_path.dentry->d_name.len);
27102 dget(new_path.dentry);
27103 diff -NurpP --minimal linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/debug.h linux-2.6.36-vs2.3.0.36.34/include/linux/vserver/debug.h
27104 --- linux-2.6.36-vs2.3.0.36.33/include/linux/vserver/debug.h 2010-10-21 13:09:36.000000000 +0200
27105 +++ linux-2.6.36-vs2.3.0.36.34/include/linux/vserver/debug.h 2010-10-21 18:35:25.000000000 +0200
27107 imajor((d)->bd_inode), iminor((d)->bd_inode)
27108 #define VXF_DEV "%p[%lu,%d:%d]"
27110 +#if defined(CONFIG_QUOTES_UTF8)
27111 +#define VS_Q_LQM "\xc2\xbb"
27112 +#define VS_Q_RQM "\xc2\xab"
27113 +#elif defined(CONFIG_QUOTES_ASCII)
27114 +#define VS_Q_LQM "\x27"
27115 +#define VS_Q_RQM "\x27"
27117 +#define VS_Q_LQM "\xbb"
27118 +#define VS_Q_RQM "\xab"
27121 +#define VS_Q(f) VS_Q_LQM f VS_Q_RQM
27124 #define vxd_path(p) \
27125 ({ static char _buffer[PATH_MAX]; \
27126 @@ -86,7 +99,7 @@ void dump_vx_info_inactive(int);
27127 #ifdef CONFIG_VSERVER_WARN
27129 #define VX_WARNLEVEL KERN_WARNING "vxW: "
27130 -#define VX_WARN_TASK "[»%s«,%u:#%u|%u|%u] "
27131 +#define VX_WARN_TASK "[" VS_Q("%s") ",%u:#%u|%u|%u] "
27132 #define VX_WARN_XID "[xid #%u] "
27133 #define VX_WARN_NID "[nid #%u] "
27134 #define VX_WARN_TAG "[tag #%u] "
27135 diff -NurpP --minimal linux-2.6.36-vs2.3.0.36.33/include/linux/vs_inet.h linux-2.6.36-vs2.3.0.36.34/include/linux/vs_inet.h
27136 --- linux-2.6.36-vs2.3.0.36.33/include/linux/vs_inet.h 2010-10-21 14:48:51.000000000 +0200
27137 +++ linux-2.6.36-vs2.3.0.36.34/include/linux/vs_inet.h 2010-10-21 17:41:35.000000000 +0200
27138 @@ -188,7 +188,8 @@ int __v4_addr_match_socket(const struct
27140 int nx_dev_visible(struct nx_info *nxi, struct net_device *dev)
27142 - vxdprintk(VXD_CBIT(net, 1), "nx_dev_visible(%p[#%u],%p »%s«) %d",
27143 + vxdprintk(VXD_CBIT(net, 1),
27144 + "nx_dev_visible(%p[#%u],%p " VS_Q("%s") ") %d",
27145 nxi, nxi ? nxi->nx_id : 0, dev, dev->name,
27146 nxi ? dev_in_nx_info(dev, nxi) : 0);
27148 diff -NurpP --minimal linux-2.6.36-vs2.3.0.36.33/kernel/vserver/inode.c linux-2.6.36-vs2.3.0.36.34/kernel/vserver/inode.c
27149 --- linux-2.6.36-vs2.3.0.36.33/kernel/vserver/inode.c 2010-10-21 14:46:31.000000000 +0200
27150 +++ linux-2.6.36-vs2.3.0.36.34/kernel/vserver/inode.c 2010-10-21 17:40:44.000000000 +0200
27151 @@ -358,7 +358,7 @@ int dx_parse_tag(char *string, tag_t *ta
27152 token = match_token(p, tokens, args);
27154 vxdprintk(VXD_CBIT(tag, 7),
27155 - "dx_parse_tag(»%s«): %d:#%d",
27156 + "dx_parse_tag(" VS_Q("%s") "): %d:#%d",
27160 diff -NurpP --minimal linux-2.6.36-vs2.3.0.36.33/kernel/vserver/Kconfig linux-2.6.36-vs2.3.0.36.34/kernel/vserver/Kconfig
27161 --- linux-2.6.36-vs2.3.0.36.33/kernel/vserver/Kconfig 2010-10-21 13:09:36.000000000 +0200
27162 +++ linux-2.6.36-vs2.3.0.36.34/kernel/vserver/Kconfig 2010-10-21 18:34:45.000000000 +0200
27163 @@ -239,6 +239,29 @@ config VSERVER_LEGACY_MEM
27164 older tools happy in the face of memory
27168 + prompt "Quotes used in debug and warn messages"
27169 + default QUOTES_ISO8859
27171 +config QUOTES_ISO8859
27172 + bool "Extended ASCII (ISO 8859) angle quotes"
27174 + This uses the extended ASCII characters \xbb
27175 + and \xab for quoting file and process names.
27177 +config QUOTES_UTF8
27178 + bool "UTF-8 angle quotes"
27180 + This uses the the UTF-8 sequences for angle
27181 + quotes to quote file and process names.
27183 +config QUOTES_ASCII
27184 + bool "ASCII single quotes"
27186 + This uses the ASCII single quote character
27187 + (\x27) to quote file and process names.
27193 diff -NurpP --minimal linux-2.6.36-vs2.3.0.36.34/kernel/vserver/Kconfig linux-2.6.36-vs2.3.0.36.35/kernel/vserver/Kconfig
27194 --- linux-2.6.36-vs2.3.0.36.34/kernel/vserver/Kconfig 2010-10-21 18:34:45.000000000 +0200
27195 +++ linux-2.6.36-vs2.3.0.36.35/kernel/vserver/Kconfig 2010-10-21 19:04:20.000000000 +0200
27196 @@ -204,33 +204,6 @@ config VSERVER_HISTORY_SIZE
27197 This allows you to specify the number of entries in
27198 the per-CPU history buffer.
27200 -config VSERVER_MONITOR
27201 - bool "VServer Scheduling Monitor"
27202 - depends on VSERVER_DISABLED
27205 - Set this to yes if you want to record the scheduling
27206 - decisions, so that they can be relayed to userspace
27207 - for detailed analysis.
27209 -config VSERVER_MONITOR_SIZE
27210 - int "Per-CPU Monitor Queue Size (32-65536)"
27211 - depends on VSERVER_MONITOR
27215 - This allows you to specify the number of entries in
27216 - the per-CPU scheduling monitor buffer.
27218 -config VSERVER_MONITOR_SYNC
27219 - int "Per-CPU Monitor Sync Interval (0-65536)"
27220 - depends on VSERVER_MONITOR
27224 - This allows you to specify the interval in ticks
27225 - when a time sync entry is inserted.
27227 config VSERVER_LEGACY_MEM
27228 bool "Legacy Memory Limits"
27230 diff -NurpP --minimal linux-2.6.36-vs2.3.0.36.34/kernel/vserver/monitor.c linux-2.6.36-vs2.3.0.36.35/kernel/vserver/monitor.c
27231 --- linux-2.6.36-vs2.3.0.36.34/kernel/vserver/monitor.c 2010-10-21 13:09:36.000000000 +0200
27232 +++ linux-2.6.36-vs2.3.0.36.35/kernel/vserver/monitor.c 1970-01-01 01:00:00.000000000 +0100
27235 - * kernel/vserver/monitor.c
27237 - * Virtual Context Scheduler Monitor
27239 - * Copyright (C) 2006-2007 Herbert Pötzl
27241 - * V0.01 basic design
27245 -#include <linux/module.h>
27246 -#include <linux/jiffies.h>
27247 -#include <asm/uaccess.h>
27248 -#include <asm/atomic.h>
27250 -#include <linux/vserver/monitor.h>
27251 -#include <linux/vserver/debug_cmd.h>
27254 -#ifdef CONFIG_VSERVER_MONITOR
27255 -#define VXM_SIZE CONFIG_VSERVER_MONITOR_SIZE
27257 -#define VXM_SIZE 64
27260 -struct _vx_monitor {
27261 - unsigned int counter;
27263 - struct _vx_mon_entry entry[VXM_SIZE+1];
27267 -DEFINE_PER_CPU(struct _vx_monitor, vx_monitor_buffer);
27269 -unsigned volatile int vxm_active = 1;
27271 -static atomic_t sequence = ATOMIC_INIT(0);
27276 - * requires disabled preemption */
27278 -struct _vx_mon_entry *vxm_advance(int cpu)
27280 - struct _vx_monitor *mon = &per_cpu(vx_monitor_buffer, cpu);
27281 - struct _vx_mon_entry *entry;
27282 - unsigned int index;
27284 - index = vxm_active ? (mon->counter++ % VXM_SIZE) : VXM_SIZE;
27285 - entry = &mon->entry[index];
27287 - entry->ev.seq = atomic_inc_return(&sequence);
27288 - entry->ev.jif = jiffies;
27292 -EXPORT_SYMBOL_GPL(vxm_advance);
27295 -int do_read_monitor(struct __user _vx_mon_entry *data,
27296 - int cpu, uint32_t *index, uint32_t *count)
27298 - int pos, ret = 0;
27299 - struct _vx_monitor *mon = &per_cpu(vx_monitor_buffer, cpu);
27300 - int end = mon->counter;
27301 - int start = end - VXM_SIZE + 2;
27302 - int idx = *index;
27304 - /* special case: get current pos */
27310 - /* have we lost some data? */
27314 - for (pos = 0; (pos < *count) && (idx < end); pos++, idx++) {
27315 - struct _vx_mon_entry *entry =
27316 - &mon->entry[idx % VXM_SIZE];
27318 - /* send entry to userspace */
27319 - ret = copy_to_user(&data[pos], entry, sizeof(*entry));
27323 - /* save new index and count */
27326 - return ret ? ret : (*index < end);
27329 -int vc_read_monitor(uint32_t id, void __user *data)
27331 - struct vcmd_read_monitor_v0 vc_data;
27334 - if (id >= NR_CPUS)
27337 - if (copy_from_user(&vc_data, data, sizeof(vc_data)))
27340 - ret = do_read_monitor((struct __user _vx_mon_entry *)vc_data.data,
27341 - id, &vc_data.index, &vc_data.count);
27343 - if (copy_to_user(data, &vc_data, sizeof(vc_data)))
27348 -#ifdef CONFIG_COMPAT
27350 -int vc_read_monitor_x32(uint32_t id, void __user *data)
27352 - struct vcmd_read_monitor_v0_x32 vc_data;
27355 - if (id >= NR_CPUS)
27358 - if (copy_from_user(&vc_data, data, sizeof(vc_data)))
27361 - ret = do_read_monitor((struct __user _vx_mon_entry *)
27362 - compat_ptr(vc_data.data_ptr),
27363 - id, &vc_data.index, &vc_data.count);
27365 - if (copy_to_user(data, &vc_data, sizeof(vc_data)))
27370 -#endif /* CONFIG_COMPAT */
27372 diff -NurpP --minimal linux-2.6.36-vs2.3.0.36.34/kernel/vserver/switch.c linux-2.6.36-vs2.3.0.36.35/kernel/vserver/switch.c
27373 --- linux-2.6.36-vs2.3.0.36.34/kernel/vserver/switch.c 2010-10-21 13:09:36.000000000 +0200
27374 +++ linux-2.6.36-vs2.3.0.36.35/kernel/vserver/switch.c 2010-10-21 19:05:54.000000000 +0200
27375 @@ -272,10 +272,6 @@ long do_vcmd(uint32_t cmd, uint32_t id,
27376 case VCMD_read_history:
27377 return __COMPAT(vc_read_history, id, data, compat);
27379 -#ifdef CONFIG_VSERVER_MONITOR
27380 - case VCMD_read_monitor:
27381 - return __COMPAT(vc_read_monitor, id, data, compat);
27384 vxwprintk_task(1, "unimplemented VCMD_%02d_%d[%d]",
27385 VC_CATEGORY(cmd), VC_COMMAND(cmd), VC_VERSION(cmd));
27386 @@ -406,9 +402,6 @@ long do_vserver(uint32_t cmd, uint32_t i
27387 __VCMD(dump_history, 9, VCA_NONE, 0);
27388 __VCMD(read_history, 9, VCA_NONE, 0);
27390 -#ifdef CONFIG_VSERVER_MONITOR
27391 - __VCMD(read_monitor, 9, VCA_NONE, 0);