diff -NurpP --minimal linux-source-2.6.22.orig/arch/alpha/Kconfig linux-source-2.6.22/arch/alpha/Kconfig --- linux-source-2.6.22.orig/arch/alpha/Kconfig 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/alpha/Kconfig 2007-10-24 16:14:05.791544531 +0200 @@ -662,6 +662,8 @@ config DUMMY_CONSOLE depends on VGA_HOSE default y +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-source-2.6.22.orig/arch/alpha/kernel/asm-offsets.c linux-source-2.6.22/arch/alpha/kernel/asm-offsets.c --- linux-source-2.6.22.orig/arch/alpha/kernel/asm-offsets.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/alpha/kernel/asm-offsets.c 2007-10-24 16:14:05.791544531 +0200 @@ -36,6 +36,7 @@ void foo(void) DEFINE(PT_PTRACED, PT_PTRACED); DEFINE(CLONE_VM, CLONE_VM); DEFINE(CLONE_UNTRACED, CLONE_UNTRACED); + DEFINE(CLONE_KTHREAD, CLONE_KTHREAD); DEFINE(SIGCHLD, SIGCHLD); BLANK(); diff -NurpP --minimal linux-source-2.6.22.orig/arch/alpha/kernel/entry.S linux-source-2.6.22/arch/alpha/kernel/entry.S --- linux-source-2.6.22.orig/arch/alpha/kernel/entry.S 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/alpha/kernel/entry.S 2007-10-24 16:14:05.791544531 +0200 @@ -643,7 +643,7 @@ kernel_thread: stq $2, 152($sp) /* HAE */ /* Shuffle FLAGS to the front; add CLONE_VM. */ - ldi $1, CLONE_VM|CLONE_UNTRACED + ldi $1, CLONE_VM | CLONE_UNTRACED | CLONE_KTHREAD; or $18, $1, $16 bsr $26, sys_clone @@ -872,24 +872,15 @@ sys_getxgid: .globl sys_getxpid .ent sys_getxpid sys_getxpid: + lda $sp, -16($sp) + stq $26, 0($sp) .prologue 0 - ldq $2, TI_TASK($8) - /* See linux/kernel/timer.c sys_getppid for discussion - about this loop. */ - ldq $3, TASK_GROUP_LEADER($2) - ldq $4, TASK_REAL_PARENT($3) - ldl $0, TASK_TGID($2) -1: ldl $1, TASK_TGID($4) -#ifdef CONFIG_SMP - mov $4, $5 - mb - ldq $3, TASK_GROUP_LEADER($2) - ldq $4, TASK_REAL_PARENT($3) - cmpeq $4, $5, $5 - beq $5, 1b -#endif - stq $1, 80($sp) + lda $16, 96($sp) + jsr $26, do_getxpid + ldq $26, 0($sp) + + lda $sp, 16($sp) ret .end sys_getxpid diff -NurpP --minimal linux-source-2.6.22.orig/arch/alpha/kernel/osf_sys.c linux-source-2.6.22/arch/alpha/kernel/osf_sys.c --- linux-source-2.6.22.orig/arch/alpha/kernel/osf_sys.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/alpha/kernel/osf_sys.c 2007-10-24 16:14:05.791544531 +0200 @@ -883,7 +883,7 @@ osf_gettimeofday(struct timeval32 __user { if (tv) { struct timeval ktv; - do_gettimeofday(&ktv); + vx_gettimeofday(&ktv); if (put_tv32(tv, &ktv)) return -EFAULT; } diff -NurpP --minimal linux-source-2.6.22.orig/arch/alpha/kernel/ptrace.c linux-source-2.6.22/arch/alpha/kernel/ptrace.c --- linux-source-2.6.22.orig/arch/alpha/kernel/ptrace.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/alpha/kernel/ptrace.c 2007-10-24 16:14:05.791544531 +0200 @@ -15,6 +15,7 @@ #include #include #include +#include #include #include @@ -283,6 +284,11 @@ do_sys_ptrace(long request, long pid, lo goto out_notsk; } + if (!vx_check(vx_task_xid(child), VS_WATCH_P | VS_IDENT)) { + ret = -EPERM; + goto out; + } + if (request == PTRACE_ATTACH) { ret = ptrace_attach(child); goto out; diff -NurpP --minimal linux-source-2.6.22.orig/arch/alpha/kernel/semaphore.c linux-source-2.6.22/arch/alpha/kernel/semaphore.c --- linux-source-2.6.22.orig/arch/alpha/kernel/semaphore.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/alpha/kernel/semaphore.c 2007-10-24 16:14:05.791544531 +0200 @@ -68,8 +68,8 @@ __down_failed(struct semaphore *sem) DECLARE_WAITQUEUE(wait, tsk); #ifdef CONFIG_DEBUG_SEMAPHORE - printk("%s(%d): down failed(%p)\n", - tsk->comm, tsk->pid, sem); + printk("%s(%d:#%u): down failed(%p)\n", + tsk->comm, tsk->pid, tsk->xid, sem); #endif tsk->state = TASK_UNINTERRUPTIBLE; @@ -97,8 +97,8 @@ __down_failed(struct semaphore *sem) wake_up(&sem->wait); #ifdef CONFIG_DEBUG_SEMAPHORE - printk("%s(%d): down acquired(%p)\n", - tsk->comm, tsk->pid, sem); + printk("%s(%d:#%u): down acquired(%p)\n", + tsk->comm, tsk->pid, tsk->xid, sem); #endif } @@ -110,8 +110,8 @@ __down_failed_interruptible(struct semap long ret = 0; #ifdef CONFIG_DEBUG_SEMAPHORE - printk("%s(%d): down failed(%p)\n", - tsk->comm, tsk->pid, sem); + printk("%s(%d:#%u): down failed(%p)\n", + tsk->comm, tsk->pid, tsk->xid, sem); #endif tsk->state = TASK_INTERRUPTIBLE; diff -NurpP --minimal linux-source-2.6.22.orig/arch/alpha/kernel/systbls.S linux-source-2.6.22/arch/alpha/kernel/systbls.S --- linux-source-2.6.22.orig/arch/alpha/kernel/systbls.S 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/alpha/kernel/systbls.S 2007-10-24 16:14:05.791544531 +0200 @@ -446,7 +446,7 @@ sys_call_table: .quad sys_stat64 /* 425 */ .quad sys_lstat64 .quad sys_fstat64 - .quad sys_ni_syscall /* sys_vserver */ + .quad sys_vserver /* sys_vserver */ .quad sys_ni_syscall /* sys_mbind */ .quad sys_ni_syscall /* sys_get_mempolicy */ .quad sys_ni_syscall /* sys_set_mempolicy */ diff -NurpP --minimal linux-source-2.6.22.orig/arch/alpha/kernel/traps.c linux-source-2.6.22/arch/alpha/kernel/traps.c --- linux-source-2.6.22.orig/arch/alpha/kernel/traps.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/alpha/kernel/traps.c 2007-10-24 16:14:05.791544531 +0200 @@ -182,7 +182,8 @@ die_if_kernel(char * str, struct pt_regs #ifdef CONFIG_SMP printk("CPU %d ", hard_smp_processor_id()); #endif - printk("%s(%d): %s %ld\n", current->comm, current->pid, str, err); + printk("%s(%d[#%u]): %s %ld\n", current->comm, + current->pid, current->xid, str, err); dik_show_regs(regs, r9_15); dik_show_trace((unsigned long *)(regs+1)); dik_show_code((unsigned int *)regs->pc); diff -NurpP --minimal linux-source-2.6.22.orig/arch/alpha/mm/fault.c linux-source-2.6.22/arch/alpha/mm/fault.c --- linux-source-2.6.22.orig/arch/alpha/mm/fault.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/alpha/mm/fault.c 2007-10-24 16:14:05.791544531 +0200 @@ -197,8 +197,8 @@ do_page_fault(unsigned long address, uns down_read(&mm->mmap_sem); goto survive; } - printk(KERN_ALERT "VM: killing process %s(%d)\n", - current->comm, current->pid); + printk(KERN_ALERT "VM: killing process %s(%d:#%u)\n", + current->comm, current->pid, current->xid); if (!user_mode(regs)) goto no_context; do_exit(SIGKILL); diff -NurpP --minimal linux-source-2.6.22.orig/arch/arm/Kconfig linux-source-2.6.22/arch/arm/Kconfig --- linux-source-2.6.22.orig/arch/arm/Kconfig 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/arm/Kconfig 2007-10-24 16:14:05.791544531 +0200 @@ -1042,6 +1042,8 @@ source "arch/arm/oprofile/Kconfig" source "arch/arm/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-source-2.6.22.orig/arch/arm/kernel/calls.S linux-source-2.6.22/arch/arm/kernel/calls.S --- linux-source-2.6.22.orig/arch/arm/kernel/calls.S 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/arm/kernel/calls.S 2007-10-24 16:14:05.791544531 +0200 @@ -322,7 +322,7 @@ /* 310 */ CALL(sys_request_key) CALL(sys_keyctl) CALL(ABI(sys_semtimedop, sys_oabi_semtimedop)) -/* vserver */ CALL(sys_ni_syscall) + CALL(sys_vserver) CALL(sys_ioprio_set) /* 315 */ CALL(sys_ioprio_get) CALL(sys_inotify_init) diff -NurpP --minimal linux-source-2.6.22.orig/arch/arm/kernel/process.c linux-source-2.6.22/arch/arm/kernel/process.c --- linux-source-2.6.22.orig/arch/arm/kernel/process.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/arm/kernel/process.c 2007-10-24 16:14:05.791544531 +0200 @@ -261,7 +261,8 @@ void __show_regs(struct pt_regs *regs) void show_regs(struct pt_regs * regs) { printk("\n"); - printk("Pid: %d, comm: %20s\n", current->pid, current->comm); + printk("Pid: %d[#%u], comm: %20s\n", + current->pid, current->xid, current->comm); __show_regs(regs); __backtrace(); } @@ -423,7 +424,8 @@ pid_t kernel_thread(int (*fn)(void *), v regs.ARM_pc = (unsigned long)kernel_thread_helper; regs.ARM_cpsr = SVC_MODE; - return do_fork(flags|CLONE_VM|CLONE_UNTRACED, 0, ®s, 0, NULL, NULL); + return do_fork(flags | CLONE_VM | CLONE_UNTRACED | CLONE_KTHREAD, + 0, ®s, 0, NULL, NULL); } EXPORT_SYMBOL(kernel_thread); diff -NurpP --minimal linux-source-2.6.22.orig/arch/arm/kernel/traps.c linux-source-2.6.22/arch/arm/kernel/traps.c --- linux-source-2.6.22.orig/arch/arm/kernel/traps.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/arm/kernel/traps.c 2007-10-24 16:14:05.791544531 +0200 @@ -222,8 +222,8 @@ static void __die(const char *str, int e str, err, ++die_counter); print_modules(); __show_regs(regs); - printk("Process %s (pid: %d, stack limit = 0x%p)\n", - tsk->comm, tsk->pid, thread + 1); + printk("Process %s (pid: %d:#%u, stack limit = 0x%p)\n", + tsk->comm, tsk->pid, tsk->xid, thread + 1); if (!user_mode(regs) || in_interrupt()) { dump_mem("Stack: ", regs->ARM_sp, diff -NurpP --minimal linux-source-2.6.22.orig/arch/arm/mm/fault.c linux-source-2.6.22/arch/arm/mm/fault.c --- linux-source-2.6.22.orig/arch/arm/mm/fault.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/arm/mm/fault.c 2007-10-24 16:14:05.791544531 +0200 @@ -266,7 +266,8 @@ do_page_fault(unsigned long addr, unsign * happened to us that made us unable to handle * the page fault gracefully. */ - printk("VM: killing process %s\n", tsk->comm); + printk("VM: killing process %s(%d:#%u)\n", + tsk->comm, tsk->pid, tsk->xid); do_exit(SIGKILL); return 0; diff -NurpP --minimal linux-source-2.6.22.orig/arch/arm26/Kconfig linux-source-2.6.22/arch/arm26/Kconfig --- linux-source-2.6.22.orig/arch/arm26/Kconfig 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/arm26/Kconfig 2007-10-24 16:14:05.791544531 +0200 @@ -243,6 +243,8 @@ source "drivers/usb/Kconfig" source "arch/arm26/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-source-2.6.22.orig/arch/arm26/kernel/calls.S linux-source-2.6.22/arch/arm26/kernel/calls.S --- linux-source-2.6.22.orig/arch/arm26/kernel/calls.S 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/arm26/kernel/calls.S 2007-10-24 16:14:05.791544531 +0200 @@ -257,6 +257,11 @@ __syscall_start: .long sys_lremovexattr .long sys_fremovexattr .long sys_tkill + + .rept 313 - (. - __syscall_start) / 4 + .long sys_ni_syscall + .endr + .long sys_vserver /* 313 */ __syscall_end: .rept NR_syscalls - (__syscall_end - __syscall_start) / 4 diff -NurpP --minimal linux-source-2.6.22.orig/arch/arm26/kernel/process.c linux-source-2.6.22/arch/arm26/kernel/process.c --- linux-source-2.6.22.orig/arch/arm26/kernel/process.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/arm26/kernel/process.c 2007-10-24 16:14:05.791544531 +0200 @@ -365,7 +365,8 @@ pid_t kernel_thread(int (*fn)(void *), v regs.ARM_r3 = (unsigned long)do_exit; regs.ARM_pc = (unsigned long)kernel_thread_helper | MODE_SVC26; - return do_fork(flags|CLONE_VM|CLONE_UNTRACED, 0, ®s, 0, NULL, NULL); + return do_fork(flags | CLONE_VM | CLONE_UNTRACED | CLONE_KTHREAD, + 0, ®s, 0, NULL, NULL); } EXPORT_SYMBOL(kernel_thread); diff -NurpP --minimal linux-source-2.6.22.orig/arch/arm26/kernel/traps.c linux-source-2.6.22/arch/arm26/kernel/traps.c --- linux-source-2.6.22.orig/arch/arm26/kernel/traps.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/arm26/kernel/traps.c 2007-10-24 16:14:05.791544531 +0200 @@ -185,8 +185,9 @@ NORET_TYPE void die(const char *str, str printk("Internal error: %s: %x\n", str, err); printk("CPU: %d\n", smp_processor_id()); show_regs(regs); - printk("Process %s (pid: %d, stack limit = 0x%p)\n", - current->comm, current->pid, end_of_stack(tsk)); + printk("Process %s (pid: %d[#%u], stack limit = 0x%p)\n", + current->comm, current->pid, + current->xid, end_of_stack(tsk)); if (!user_mode(regs) || in_interrupt()) { __dump_stack(tsk, (unsigned long)(regs + 1)); diff -NurpP --minimal linux-source-2.6.22.orig/arch/cris/arch-v10/kernel/process.c linux-source-2.6.22/arch/cris/arch-v10/kernel/process.c --- linux-source-2.6.22.orig/arch/cris/arch-v10/kernel/process.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/cris/arch-v10/kernel/process.c 2007-10-24 16:14:05.791544531 +0200 @@ -103,7 +103,8 @@ int kernel_thread(int (*fn)(void *), voi regs.dccr = 1 << I_DCCR_BITNR; /* Ok, create the new process.. */ - return do_fork(flags | CLONE_VM | CLONE_UNTRACED, 0, ®s, 0, NULL, NULL); + return do_fork(flags | CLONE_VM | CLONE_UNTRACED | CLONE_KTHREAD, + 0, ®s, 0, NULL, NULL); } /* setup the child's kernel stack with a pt_regs and switch_stack on it. diff -NurpP --minimal linux-source-2.6.22.orig/arch/cris/arch-v32/kernel/process.c linux-source-2.6.22/arch/cris/arch-v32/kernel/process.c --- linux-source-2.6.22.orig/arch/cris/arch-v32/kernel/process.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/cris/arch-v32/kernel/process.c 2007-10-24 16:14:05.791544531 +0200 @@ -120,7 +120,8 @@ kernel_thread(int (*fn)(void *), void * regs.ccs = 1 << (I_CCS_BITNR + CCS_SHIFT); /* Create the new process. */ - return do_fork(flags | CLONE_VM | CLONE_UNTRACED, 0, ®s, 0, NULL, NULL); + return do_fork(flags | CLONE_VM | CLONE_UNTRACED | CLONE_KTHREAD, + 0, ®s, 0, NULL, NULL); } /* diff -NurpP --minimal linux-source-2.6.22.orig/arch/cris/Kconfig linux-source-2.6.22/arch/cris/Kconfig --- linux-source-2.6.22.orig/arch/cris/Kconfig 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/cris/Kconfig 2007-10-24 16:14:05.791544531 +0200 @@ -200,6 +200,8 @@ source "drivers/usb/Kconfig" source "arch/cris/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-source-2.6.22.orig/arch/frv/kernel/kernel_thread.S linux-source-2.6.22/arch/frv/kernel/kernel_thread.S --- linux-source-2.6.22.orig/arch/frv/kernel/kernel_thread.S 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/frv/kernel/kernel_thread.S 2007-10-24 16:14:05.791544531 +0200 @@ -13,6 +13,8 @@ #include #define CLONE_VM 0x00000100 /* set if VM shared between processes */ +#define CLONE_KTHREAD 0x10000000 /* kernel thread */ +#define CLONE_KT (CLONE_VM | CLONE_KTHREAD) /* kernel thread flags */ #define KERN_ERR "<3>" .section .rodata @@ -37,7 +39,7 @@ kernel_thread: # start by forking the current process, but with shared VM setlos.p #__NR_clone,gr7 ; syscall number - ori gr10,#CLONE_VM,gr8 ; first syscall arg [clone_flags] + ori gr10,#CLONE_KT,gr8 ; first syscall arg [clone_flags] sethi.p #0xe4e4,gr9 ; second syscall arg [newsp] setlo #0xe4e4,gr9 setlos.p #0,gr10 ; third syscall arg [parent_tidptr] diff -NurpP --minimal linux-source-2.6.22.orig/arch/h8300/Kconfig linux-source-2.6.22/arch/h8300/Kconfig --- linux-source-2.6.22.orig/arch/h8300/Kconfig 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/h8300/Kconfig 2007-10-24 16:14:05.791544531 +0200 @@ -222,6 +222,8 @@ source "fs/Kconfig" source "arch/h8300/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-source-2.6.22.orig/arch/h8300/kernel/process.c linux-source-2.6.22/arch/h8300/kernel/process.c --- linux-source-2.6.22.orig/arch/h8300/kernel/process.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/h8300/kernel/process.c 2007-10-24 16:14:05.791544531 +0200 @@ -134,7 +134,7 @@ int kernel_thread(int (*fn)(void *), voi fs = get_fs(); set_fs (KERNEL_DS); - clone_arg = flags | CLONE_VM; + clone_arg = flags | CLONE_VM | CLONE_KTHREAD; __asm__("mov.l sp,er3\n\t" "sub.l er2,er2\n\t" "mov.l %2,er1\n\t" diff -NurpP --minimal linux-source-2.6.22.orig/arch/i386/Kconfig linux-source-2.6.22/arch/i386/Kconfig --- linux-source-2.6.22.orig/arch/i386/Kconfig 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/i386/Kconfig 2007-10-24 16:14:05.791544531 +0200 @@ -1230,6 +1230,8 @@ endmenu source "arch/i386/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-source-2.6.22.orig/arch/i386/kernel/process.c linux-source-2.6.22/arch/i386/kernel/process.c --- linux-source-2.6.22.orig/arch/i386/kernel/process.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/i386/kernel/process.c 2007-10-24 16:14:05.791544531 +0200 @@ -302,8 +302,10 @@ void show_regs(struct pt_regs * regs) unsigned long cr0 = 0L, cr2 = 0L, cr3 = 0L, cr4 = 0L; printk("\n"); - printk("Pid: %d, comm: %20s\n", current->pid, current->comm); - printk("EIP: %04x:[<%08lx>] CPU: %d\n",0xffff & regs->xcs,regs->eip, smp_processor_id()); + printk("Pid: %d[#%u], comm: %20s\n", + current->pid, current->xid, current->comm); + printk("EIP: %04x:[<%08lx>] CPU: %d\n", + 0xffff & regs->xcs,regs->eip, smp_processor_id()); print_symbol("EIP is at %s\n", regs->eip); if (user_mode_vm(regs)) @@ -355,7 +357,8 @@ int kernel_thread(int (*fn)(void *), voi regs.eflags = X86_EFLAGS_IF | X86_EFLAGS_SF | X86_EFLAGS_PF | 0x2; /* Ok, create the new process.. */ - return do_fork(flags | CLONE_VM | CLONE_UNTRACED, 0, ®s, 0, NULL, NULL); + return do_fork(flags | CLONE_VM | CLONE_UNTRACED | CLONE_KTHREAD, + 0, ®s, 0, NULL, NULL); } EXPORT_SYMBOL(kernel_thread); diff -NurpP --minimal linux-source-2.6.22.orig/arch/i386/kernel/syscall_table.S linux-source-2.6.22/arch/i386/kernel/syscall_table.S --- linux-source-2.6.22.orig/arch/i386/kernel/syscall_table.S 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/i386/kernel/syscall_table.S 2007-10-24 16:14:05.791544531 +0200 @@ -272,7 +272,7 @@ ENTRY(sys_call_table) .long sys_tgkill /* 270 */ .long sys_utimes .long sys_fadvise64_64 - .long sys_ni_syscall /* sys_vserver */ + .long sys_vserver .long sys_mbind .long sys_get_mempolicy .long sys_set_mempolicy diff -NurpP --minimal linux-source-2.6.22.orig/arch/i386/kernel/sysenter.c linux-source-2.6.22/arch/i386/kernel/sysenter.c --- linux-source-2.6.22.orig/arch/i386/kernel/sysenter.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/arch/i386/kernel/sysenter.c 2007-10-24 16:14:05.791544531 +0200 @@ -17,6 +17,7 @@ #include #include #include +#include #include #include diff -NurpP --minimal linux-source-2.6.22.orig/arch/i386/kernel/traps.c linux-source-2.6.22/arch/i386/kernel/traps.c --- linux-source-2.6.22.orig/arch/i386/kernel/traps.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/i386/kernel/traps.c 2007-10-24 16:14:05.791544531 +0200 @@ -56,6 +56,8 @@ #include #include +#include +#include #include "mach_traps.h" @@ -303,8 +305,8 @@ void show_registers(struct pt_regs *regs regs->esi, regs->edi, regs->ebp, esp); printk(KERN_EMERG "ds: %04x es: %04x fs: %04x gs: %04x ss: %04x\n", regs->xds & 0xffff, regs->xes & 0xffff, regs->xfs & 0xffff, gs, ss); - printk(KERN_EMERG "Process %.*s (pid: %d, ti=%p task=%p task.ti=%p)", - TASK_COMM_LEN, current->comm, current->pid, + printk(KERN_EMERG "Process %.*s (pid: %d[#%u], ti=%p task=%p task.ti=%p)", + TASK_COMM_LEN, current->comm, current->pid, current->xid, current_thread_info(), current, task_thread_info(current)); /* * When in-kernel, we also print out the stack and code at the @@ -375,6 +377,8 @@ void die(const char * str, struct pt_reg oops_enter(); + vxh_throw_oops(); + if (die.lock_owner != raw_smp_processor_id()) { console_verbose(); spin_lock_irqsave(&die.lock, flags); @@ -412,9 +416,9 @@ void die(const char * str, struct pt_reg if (nl) printk("\n"); if (notify_die(DIE_OOPS, str, regs, err, - current->thread.trap_no, SIGSEGV) != - NOTIFY_STOP) { + current->thread.trap_no, SIGSEGV) != NOTIFY_STOP) { show_registers(regs); + vxh_dump_history(); /* Executive summary in case the oops scrolled away */ esp = (unsigned long) (®s->esp); savesegment(ss, ss); diff -NurpP --minimal linux-source-2.6.22.orig/arch/i386/mm/fault.c linux-source-2.6.22/arch/i386/mm/fault.c --- linux-source-2.6.22.orig/arch/i386/mm/fault.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/arch/i386/mm/fault.c 2007-10-24 16:14:05.791544531 +0200 @@ -587,7 +587,8 @@ out_of_memory: down_read(&mm->mmap_sem); goto survive; } - printk("VM: killing process %s\n", tsk->comm); + printk("VM: killing process %s(%d:#%u)\n", + tsk->comm, tsk->pid, tsk->xid); if (error_code & 4) do_exit(SIGKILL); goto no_context; diff -NurpP --minimal linux-source-2.6.22.orig/arch/ia64/ia32/binfmt_elf32.c linux-source-2.6.22/arch/ia64/ia32/binfmt_elf32.c --- linux-source-2.6.22.orig/arch/ia64/ia32/binfmt_elf32.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/ia64/ia32/binfmt_elf32.c 2007-10-24 16:14:05.791544531 +0200 @@ -233,7 +233,8 @@ ia32_setup_arg_pages (struct linux_binpr kmem_cache_free(vm_area_cachep, mpnt); return ret; } - current->mm->stack_vm = current->mm->total_vm = vma_pages(mpnt); + vx_vmpages_sub(current->mm, current->mm->total_vm - vma_pages(mpnt)); + current->mm->stack_vm = current->mm->total_vm; } for (i = 0 ; i < MAX_ARG_PAGES ; i++) { diff -NurpP --minimal linux-source-2.6.22.orig/arch/ia64/ia32/ia32_entry.S linux-source-2.6.22/arch/ia64/ia32/ia32_entry.S --- linux-source-2.6.22.orig/arch/ia64/ia32/ia32_entry.S 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/ia64/ia32/ia32_entry.S 2007-10-24 16:14:05.791544531 +0200 @@ -446,7 +446,7 @@ ia32_syscall_table: data8 sys_tgkill /* 270 */ data8 compat_sys_utimes data8 sys32_fadvise64_64 - data8 sys_ni_syscall + data8 sys32_vserver data8 sys_ni_syscall data8 sys_ni_syscall /* 275 */ data8 sys_ni_syscall diff -NurpP --minimal linux-source-2.6.22.orig/arch/ia64/ia32/sys_ia32.c linux-source-2.6.22/arch/ia64/ia32/sys_ia32.c --- linux-source-2.6.22.orig/arch/ia64/ia32/sys_ia32.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/ia64/ia32/sys_ia32.c 2007-10-24 16:14:05.791544531 +0200 @@ -1182,7 +1182,7 @@ sys32_gettimeofday (struct compat_timeva { if (tv) { struct timeval ktv; - do_gettimeofday(&ktv); + vx_gettimeofday(&ktv); if (put_tv32(tv, &ktv)) return -EFAULT; } diff -NurpP --minimal linux-source-2.6.22.orig/arch/ia64/Kconfig linux-source-2.6.22/arch/ia64/Kconfig --- linux-source-2.6.22.orig/arch/ia64/Kconfig 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/ia64/Kconfig 2007-10-24 16:14:05.791544531 +0200 @@ -592,6 +592,8 @@ endmenu source "arch/ia64/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-source-2.6.22.orig/arch/ia64/kernel/asm-offsets.c linux-source-2.6.22/arch/ia64/kernel/asm-offsets.c --- linux-source-2.6.22.orig/arch/ia64/kernel/asm-offsets.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/ia64/kernel/asm-offsets.c 2007-10-24 16:14:05.791544531 +0200 @@ -192,6 +192,7 @@ void foo(void) /* for assembly files which can't include sched.h: */ DEFINE(IA64_CLONE_VFORK, CLONE_VFORK); DEFINE(IA64_CLONE_VM, CLONE_VM); + DEFINE(IA64_CLONE_KTHREAD, CLONE_KTHREAD); BLANK(); DEFINE(IA64_CPUINFO_NSEC_PER_CYC_OFFSET, diff -NurpP --minimal linux-source-2.6.22.orig/arch/ia64/kernel/entry.S linux-source-2.6.22/arch/ia64/kernel/entry.S --- linux-source-2.6.22.orig/arch/ia64/kernel/entry.S 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/ia64/kernel/entry.S 2007-10-24 16:14:05.791544531 +0200 @@ -1547,7 +1547,7 @@ sys_call_table: data8 sys_mq_notify data8 sys_mq_getsetattr data8 sys_kexec_load - data8 sys_ni_syscall // reserved for vserver + data8 sys_vserver data8 sys_waitid // 1270 data8 sys_add_key data8 sys_request_key diff -NurpP --minimal linux-source-2.6.22.orig/arch/ia64/kernel/perfmon.c linux-source-2.6.22/arch/ia64/kernel/perfmon.c --- linux-source-2.6.22.orig/arch/ia64/kernel/perfmon.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/ia64/kernel/perfmon.c 2007-10-24 16:14:05.791544531 +0200 @@ -40,6 +40,7 @@ #include #include #include +#include #include #include @@ -2395,7 +2396,7 @@ pfm_smpl_buffer_alloc(struct task_struct */ insert_vm_struct(mm, vma); - mm->total_vm += size >> PAGE_SHIFT; + vx_vmpages_add(mm, size >> PAGE_SHIFT); vm_stat_account(vma->vm_mm, vma->vm_flags, vma->vm_file, vma_pages(vma)); up_write(&task->mm->mmap_sem); diff -NurpP --minimal linux-source-2.6.22.orig/arch/ia64/kernel/process.c linux-source-2.6.22/arch/ia64/kernel/process.c --- linux-source-2.6.22.orig/arch/ia64/kernel/process.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/ia64/kernel/process.c 2007-10-24 16:14:05.791544531 +0200 @@ -105,7 +105,8 @@ show_regs (struct pt_regs *regs) unsigned long ip = regs->cr_iip + ia64_psr(regs)->ri; print_modules(); - printk("\nPid: %d, CPU %d, comm: %20s\n", current->pid, smp_processor_id(), current->comm); + printk("\nPid: %d[#%u], CPU %d, comm: %20s\n", + current->pid, current->xid, smp_processor_id(), current->comm); printk("psr : %016lx ifs : %016lx ip : [<%016lx>] %s\n", regs->cr_ipsr, regs->cr_ifs, ip, print_tainted()); print_symbol("ip is at %s\n", ip); @@ -696,7 +697,8 @@ kernel_thread (int (*fn)(void *), void * regs.sw.ar_fpsr = regs.pt.ar_fpsr = ia64_getreg(_IA64_REG_AR_FPSR); regs.sw.ar_bspstore = (unsigned long) current + IA64_RBS_OFFSET; regs.sw.pr = (1 << PRED_KERNEL_STACK); - return do_fork(flags | CLONE_VM | CLONE_UNTRACED, 0, ®s.pt, 0, NULL, NULL); + return do_fork(flags | CLONE_VM | CLONE_UNTRACED | CLONE_KTHREAD, + 0, ®s.pt, 0, NULL, NULL); } EXPORT_SYMBOL(kernel_thread); diff -NurpP --minimal linux-source-2.6.22.orig/arch/ia64/kernel/ptrace.c linux-source-2.6.22/arch/ia64/kernel/ptrace.c --- linux-source-2.6.22.orig/arch/ia64/kernel/ptrace.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/ia64/kernel/ptrace.c 2007-10-24 16:14:05.791544531 +0200 @@ -17,6 +17,7 @@ #include #include #include +#include #include #include @@ -1443,6 +1444,9 @@ sys_ptrace (long request, pid_t pid, uns read_unlock(&tasklist_lock); if (!child) goto out; + if (!vx_check(vx_task_xid(child), VS_WATCH_P | VS_IDENT)) + goto out_tsk; + ret = -EPERM; if (pid == 1) /* no messing around with init! */ goto out_tsk; diff -NurpP --minimal linux-source-2.6.22.orig/arch/ia64/kernel/traps.c linux-source-2.6.22/arch/ia64/kernel/traps.c --- linux-source-2.6.22.orig/arch/ia64/kernel/traps.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/ia64/kernel/traps.c 2007-10-24 16:14:05.791544531 +0200 @@ -60,8 +60,9 @@ die (const char *str, struct pt_regs *re put_cpu(); if (++die.lock_owner_depth < 3) { - printk("%s[%d]: %s %ld [%d]\n", - current->comm, current->pid, str, err, ++die_counter); + printk("%s[%d[#%u]]: %s %ld [%d]\n", + current->comm, current->pid, current->xid, + str, err, ++die_counter); (void) notify_die(DIE_OOPS, (char *)str, regs, err, 255, SIGSEGV); show_regs(regs); } else @@ -313,8 +314,9 @@ handle_fpu_swa (int fp_fault, struct pt_ if ((last.count & 15) < 5 && (ia64_fetchadd(1, &last.count, acq) & 15) < 5) { last.time = current_jiffies + 5 * HZ; printk(KERN_WARNING - "%s(%d): floating-point assist fault at ip %016lx, isr %016lx\n", - current->comm, current->pid, regs->cr_iip + ia64_psr(regs)->ri, isr); + "%s(%d[#%u]): floating-point assist fault at ip %016lx, isr %016lx\n", + current->comm, current->pid, current->xid, + regs->cr_iip + ia64_psr(regs)->ri, isr); } } } diff -NurpP --minimal linux-source-2.6.22.orig/arch/ia64/mm/fault.c linux-source-2.6.22/arch/ia64/mm/fault.c --- linux-source-2.6.22.orig/arch/ia64/mm/fault.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/ia64/mm/fault.c 2007-10-24 16:14:05.791544531 +0200 @@ -10,6 +10,7 @@ #include #include #include +#include #include #include diff -NurpP --minimal linux-source-2.6.22.orig/arch/m32r/kernel/process.c linux-source-2.6.22/arch/m32r/kernel/process.c --- linux-source-2.6.22.orig/arch/m32r/kernel/process.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/m32r/kernel/process.c 2007-10-24 16:14:05.791544531 +0200 @@ -211,8 +211,8 @@ int kernel_thread(int (*fn)(void *), voi regs.psw = M32R_PSW_BIE; /* Ok, create the new process. */ - return do_fork(flags | CLONE_VM | CLONE_UNTRACED, 0, ®s, 0, NULL, - NULL); + return do_fork(flags | CLONE_VM | CLONE_UNTRACED | CLONE_KTHREAD, + 0, ®s, 0, NULL, NULL); } /* diff -NurpP --minimal linux-source-2.6.22.orig/arch/m32r/kernel/traps.c linux-source-2.6.22/arch/m32r/kernel/traps.c --- linux-source-2.6.22.orig/arch/m32r/kernel/traps.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/m32r/kernel/traps.c 2007-10-24 16:14:05.791544531 +0200 @@ -195,8 +195,9 @@ static void show_registers(struct pt_reg } else { printk("SPI: %08lx\n", sp); } - printk("Process %s (pid: %d, process nr: %d, stackpage=%08lx)", - current->comm, current->pid, 0xffff & i, 4096+(unsigned long)current); + printk("Process %s (pid: %d[#%u], process nr: %d, stackpage=%08lx)", + current->comm, current->pid, current->xid, + 0xffff & i, 4096+(unsigned long)current); /* * When in-kernel, we also print out the stack and code at the diff -NurpP --minimal linux-source-2.6.22.orig/arch/m68k/Kconfig linux-source-2.6.22/arch/m68k/Kconfig --- linux-source-2.6.22.orig/arch/m68k/Kconfig 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/m68k/Kconfig 2007-10-24 16:14:05.791544531 +0200 @@ -678,6 +678,8 @@ source "fs/Kconfig" source "arch/m68k/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-source-2.6.22.orig/arch/m68k/kernel/process.c linux-source-2.6.22/arch/m68k/kernel/process.c --- linux-source-2.6.22.orig/arch/m68k/kernel/process.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/m68k/kernel/process.c 2007-10-24 16:14:05.791544531 +0200 @@ -159,7 +159,8 @@ int kernel_thread(int (*fn)(void *), voi { register long retval __asm__ ("d0"); - register long clone_arg __asm__ ("d1") = flags | CLONE_VM | CLONE_UNTRACED; + register long clone_arg __asm__ ("d1") = + flags | CLONE_VM | CLONE_UNTRACED | CLONE_KTHREAD; retval = __NR_clone; __asm__ __volatile__ diff -NurpP --minimal linux-source-2.6.22.orig/arch/m68k/kernel/ptrace.c linux-source-2.6.22/arch/m68k/kernel/ptrace.c --- linux-source-2.6.22.orig/arch/m68k/kernel/ptrace.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/m68k/kernel/ptrace.c 2007-10-24 16:14:05.791544531 +0200 @@ -18,6 +18,7 @@ #include #include #include +#include #include #include @@ -278,6 +279,8 @@ long arch_ptrace(struct task_struct *chi ret = ptrace_request(child, request, addr, data); break; } + if (!vx_check(vx_task_xid(child), VS_WATCH_P | VS_IDENT)) + goto out_tsk; return ret; out_eio: diff -NurpP --minimal linux-source-2.6.22.orig/arch/m68k/kernel/traps.c linux-source-2.6.22/arch/m68k/kernel/traps.c --- linux-source-2.6.22.orig/arch/m68k/kernel/traps.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/m68k/kernel/traps.c 2007-10-24 16:14:05.791544531 +0200 @@ -899,8 +899,8 @@ void show_registers(struct pt_regs *regs printk("d4: %08lx d5: %08lx a0: %08lx a1: %08lx\n", regs->d4, regs->d5, regs->a0, regs->a1); - printk("Process %s (pid: %d, task=%p)\n", - current->comm, current->pid, current); + printk("Process %s (pid: %d[#%u], task=%p)\n", + current->comm, current->pid, current->xid, current); addr = (unsigned long)&fp->un; printk("Frame format=%X ", regs->format); switch (regs->format) { diff -NurpP --minimal linux-source-2.6.22.orig/arch/m68knommu/Kconfig linux-source-2.6.22/arch/m68knommu/Kconfig --- linux-source-2.6.22.orig/arch/m68knommu/Kconfig 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/m68knommu/Kconfig 2007-10-24 16:14:05.791544531 +0200 @@ -670,6 +670,8 @@ source "fs/Kconfig" source "arch/m68knommu/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-source-2.6.22.orig/arch/m68knommu/kernel/process.c linux-source-2.6.22/arch/m68knommu/kernel/process.c --- linux-source-2.6.22.orig/arch/m68knommu/kernel/process.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/m68knommu/kernel/process.c 2007-10-24 16:14:05.791544531 +0200 @@ -122,7 +122,7 @@ void show_regs(struct pt_regs * regs) int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags) { int retval; - long clone_arg = flags | CLONE_VM; + long clone_arg = flags | CLONE_VM | CLONE_KTHREAD; mm_segment_t fs; fs = get_fs(); diff -NurpP --minimal linux-source-2.6.22.orig/arch/m68knommu/kernel/traps.c linux-source-2.6.22/arch/m68knommu/kernel/traps.c --- linux-source-2.6.22.orig/arch/m68knommu/kernel/traps.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/m68knommu/kernel/traps.c 2007-10-24 16:14:05.791544531 +0200 @@ -80,8 +80,9 @@ void die_if_kernel(char *str, struct pt_ printk(KERN_EMERG "d4: %08lx d5: %08lx a0: %08lx a1: %08lx\n", fp->d4, fp->d5, fp->a0, fp->a1); - printk(KERN_EMERG "Process %s (pid: %d, stackpage=%08lx)\n", - current->comm, current->pid, PAGE_SIZE+(unsigned long)current); + printk(KERN_EMERG "Process %s (pid: %d[#%u], stackpage=%08lx)\n", + current->comm, current->pid, current->xid, + PAGE_SIZE+(unsigned long)current); show_stack(NULL, (unsigned long *)fp); do_exit(SIGSEGV); } diff -NurpP --minimal linux-source-2.6.22.orig/arch/mips/Kconfig linux-source-2.6.22/arch/mips/Kconfig --- linux-source-2.6.22.orig/arch/mips/Kconfig 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/mips/Kconfig 2007-10-24 16:14:05.791544531 +0200 @@ -1959,6 +1959,8 @@ source "arch/mips/oprofile/Kconfig" source "arch/mips/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-source-2.6.22.orig/arch/mips/kernel/linux32.c linux-source-2.6.22/arch/mips/kernel/linux32.c --- linux-source-2.6.22.orig/arch/mips/kernel/linux32.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/mips/kernel/linux32.c 2007-10-24 16:14:05.791544531 +0200 @@ -229,7 +229,7 @@ sys32_gettimeofday(struct compat_timeval { if (tv) { struct timeval ktv; - do_gettimeofday(&ktv); + vx_gettimeofday(&ktv); if (put_tv32(tv, &ktv)) return -EFAULT; } diff -NurpP --minimal linux-source-2.6.22.orig/arch/mips/kernel/process.c linux-source-2.6.22/arch/mips/kernel/process.c --- linux-source-2.6.22.orig/arch/mips/kernel/process.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/mips/kernel/process.c 2007-10-24 16:14:05.791544531 +0200 @@ -236,7 +236,8 @@ long kernel_thread(int (*fn)(void *), vo #endif /* Ok, create the new process.. */ - return do_fork(flags | CLONE_VM | CLONE_UNTRACED, 0, ®s, 0, NULL, NULL); + return do_fork(flags | CLONE_VM | CLONE_UNTRACED | CLONE_KTHREAD, + 0, ®s, 0, NULL, NULL); } /* diff -NurpP --minimal linux-source-2.6.22.orig/arch/mips/kernel/ptrace.c linux-source-2.6.22/arch/mips/kernel/ptrace.c --- linux-source-2.6.22.orig/arch/mips/kernel/ptrace.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/mips/kernel/ptrace.c 2007-10-24 16:14:05.791544531 +0200 @@ -25,6 +25,7 @@ #include #include #include +#include #include #include @@ -171,6 +172,9 @@ long arch_ptrace(struct task_struct *chi { int ret; + if (!vx_check(vx_task_xid(child), VS_WATCH_P | VS_IDENT)) + goto out; + switch (request) { /* when I and D space are separate, these will need to be fixed. */ case PTRACE_PEEKTEXT: /* read word at location addr. */ diff -NurpP --minimal linux-source-2.6.22.orig/arch/mips/kernel/scall32-o32.S linux-source-2.6.22/arch/mips/kernel/scall32-o32.S --- linux-source-2.6.22.orig/arch/mips/kernel/scall32-o32.S 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/mips/kernel/scall32-o32.S 2007-10-24 16:14:05.791544531 +0200 @@ -619,7 +619,7 @@ einval: li v0, -EINVAL sys sys_mq_timedreceive 5 sys sys_mq_notify 2 /* 4275 */ sys sys_mq_getsetattr 3 - sys sys_ni_syscall 0 /* sys_vserver */ + sys sys_vserver 3 sys sys_waitid 5 sys sys_ni_syscall 0 /* available, was setaltroot */ sys sys_add_key 5 /* 4280 */ diff -NurpP --minimal linux-source-2.6.22.orig/arch/mips/kernel/scall64-64.S linux-source-2.6.22/arch/mips/kernel/scall64-64.S --- linux-source-2.6.22.orig/arch/mips/kernel/scall64-64.S 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/mips/kernel/scall64-64.S 2007-10-24 16:14:05.791544531 +0200 @@ -434,7 +434,7 @@ sys_call_table: PTR sys_mq_timedreceive PTR sys_mq_notify PTR sys_mq_getsetattr /* 5235 */ - PTR sys_ni_syscall /* sys_vserver */ + PTR sys_vserver PTR sys_waitid PTR sys_ni_syscall /* available, was setaltroot */ PTR sys_add_key diff -NurpP --minimal linux-source-2.6.22.orig/arch/mips/kernel/scall64-n32.S linux-source-2.6.22/arch/mips/kernel/scall64-n32.S --- linux-source-2.6.22.orig/arch/mips/kernel/scall64-n32.S 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/mips/kernel/scall64-n32.S 2007-10-24 16:14:05.791544531 +0200 @@ -360,7 +360,7 @@ EXPORT(sysn32_call_table) PTR compat_sys_mq_timedreceive PTR compat_sys_mq_notify PTR compat_sys_mq_getsetattr - PTR sys_ni_syscall /* 6240, sys_vserver */ + PTR sys32_vserver /* 6240 */ PTR compat_sys_waitid PTR sys_ni_syscall /* available, was setaltroot */ PTR sys_add_key diff -NurpP --minimal linux-source-2.6.22.orig/arch/mips/kernel/scall64-o32.S linux-source-2.6.22/arch/mips/kernel/scall64-o32.S --- linux-source-2.6.22.orig/arch/mips/kernel/scall64-o32.S 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/mips/kernel/scall64-o32.S 2007-10-24 16:14:05.791544531 +0200 @@ -482,7 +482,7 @@ sys_call_table: PTR compat_sys_mq_timedreceive PTR compat_sys_mq_notify /* 4275 */ PTR compat_sys_mq_getsetattr - PTR sys_ni_syscall /* sys_vserver */ + PTR sys32_vserver PTR sys32_waitid PTR sys_ni_syscall /* available, was setaltroot */ PTR sys_add_key /* 4280 */ diff -NurpP --minimal linux-source-2.6.22.orig/arch/mips/kernel/traps.c linux-source-2.6.22/arch/mips/kernel/traps.c --- linux-source-2.6.22.orig/arch/mips/kernel/traps.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/mips/kernel/traps.c 2007-10-24 16:14:05.791544531 +0200 @@ -302,8 +302,9 @@ void show_registers(struct pt_regs *regs { show_regs(regs); print_modules(); - printk("Process %s (pid: %d, threadinfo=%p, task=%p)\n", - current->comm, current->pid, current_thread_info(), current); + printk("Process %s (pid: %d:#%u, threadinfo=%p, task=%p)\n", + current->comm, current->pid, current->xid, + current_thread_info(), current); show_stacktrace(current, regs); show_code((unsigned int *) regs->cp0_epc); printk("\n"); diff -NurpP --minimal linux-source-2.6.22.orig/arch/mips/mm/fault.c linux-source-2.6.22/arch/mips/mm/fault.c --- linux-source-2.6.22.orig/arch/mips/mm/fault.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/mips/mm/fault.c 2007-10-24 16:14:05.791544531 +0200 @@ -179,7 +179,8 @@ out_of_memory: down_read(&mm->mmap_sem); goto survive; } - printk("VM: killing process %s\n", tsk->comm); + printk("VM: killing process %s(%d:#%u)\n", + tsk->comm, tsk->pid, tsk->xid); if (user_mode(regs)) do_exit(SIGKILL); goto no_context; diff -NurpP --minimal linux-source-2.6.22.orig/arch/parisc/Kconfig linux-source-2.6.22/arch/parisc/Kconfig --- linux-source-2.6.22.orig/arch/parisc/Kconfig 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/parisc/Kconfig 2007-10-24 16:14:05.791544531 +0200 @@ -271,6 +271,8 @@ source "arch/parisc/oprofile/Kconfig" source "arch/parisc/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-source-2.6.22.orig/arch/parisc/kernel/entry.S linux-source-2.6.22/arch/parisc/kernel/entry.S --- linux-source-2.6.22.orig/arch/parisc/kernel/entry.S 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/parisc/kernel/entry.S 2007-10-24 16:14:05.791544531 +0200 @@ -761,6 +761,7 @@ END(fault_vector_11) #define CLONE_VM 0x100 /* Must agree with */ #define CLONE_UNTRACED 0x00800000 +#define CLONE_KTHREAD 0x10000000 .import do_fork ENTRY(__kernel_thread) diff -NurpP --minimal linux-source-2.6.22.orig/arch/parisc/kernel/process.c linux-source-2.6.22/arch/parisc/kernel/process.c --- linux-source-2.6.22.orig/arch/parisc/kernel/process.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/parisc/kernel/process.c 2007-10-24 16:14:05.791544531 +0200 @@ -173,7 +173,7 @@ pid_t kernel_thread(int (*fn)(void *), v * kernel_thread can become a #define. */ - return __kernel_thread(fn, arg, flags); + return __kernel_thread(fn, arg, flags | CLONE_KTHREAD); } EXPORT_SYMBOL(kernel_thread); diff -NurpP --minimal linux-source-2.6.22.orig/arch/parisc/kernel/syscall_table.S linux-source-2.6.22/arch/parisc/kernel/syscall_table.S --- linux-source-2.6.22.orig/arch/parisc/kernel/syscall_table.S 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/arch/parisc/kernel/syscall_table.S 2007-10-24 16:14:05.791544531 +0200 @@ -361,7 +361,7 @@ ENTRY_COMP(mbind) /* 260 */ ENTRY_COMP(get_mempolicy) ENTRY_COMP(set_mempolicy) - ENTRY_SAME(ni_syscall) /* 263: reserved for vserver */ + ENTRY_DIFF(vserver) ENTRY_SAME(add_key) ENTRY_SAME(request_key) /* 265 */ ENTRY_SAME(keyctl) diff -NurpP --minimal linux-source-2.6.22.orig/arch/parisc/kernel/sys_parisc32.c linux-source-2.6.22/arch/parisc/kernel/sys_parisc32.c --- linux-source-2.6.22.orig/arch/parisc/kernel/sys_parisc32.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/arch/parisc/kernel/sys_parisc32.c 2007-10-24 16:14:05.791544531 +0200 @@ -204,11 +204,11 @@ static inline long get_ts32(struct times asmlinkage int sys32_gettimeofday(struct compat_timeval __user *tv, struct timezone __user *tz) { - extern void do_gettimeofday(struct timeval *tv); + extern void vx_gettimeofday(struct timeval *tv); if (tv) { struct timeval ktv; - do_gettimeofday(&ktv); + vx_gettimeofday(&ktv); if (put_compat_timeval(tv, &ktv)) return -EFAULT; } diff -NurpP --minimal linux-source-2.6.22.orig/arch/parisc/kernel/traps.c linux-source-2.6.22/arch/parisc/kernel/traps.c --- linux-source-2.6.22.orig/arch/parisc/kernel/traps.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/parisc/kernel/traps.c 2007-10-24 16:14:05.791544531 +0200 @@ -218,8 +218,9 @@ void die_if_kernel(char *str, struct pt_ if (err == 0) return; /* STFU */ - printk(KERN_CRIT "%s (pid %d): %s (code %ld) at " RFMT "\n", - current->comm, current->pid, str, err, regs->iaoq[0]); + printk(KERN_CRIT "%s (pid %d:#%u): %s (code %ld) at " RFMT "\n", + current->comm, current->pid, current->xid, + str, err, regs->iaoq[0]); #ifdef PRINT_USER_FAULTS /* XXX for debugging only */ show_regs(regs); @@ -251,8 +252,8 @@ KERN_CRIT " || | pdc_console_restart(); if (err) - printk(KERN_CRIT "%s (pid %d): %s (code %ld)\n", - current->comm, current->pid, str, err); + printk(KERN_CRIT "%s (pid %d:#%u): %s (code %ld)\n", + current->comm, current->pid, current->xid, str, err); /* Wot's wrong wif bein' racy? */ if (current->thread.flags & PARISC_KERNEL_DEATH) { diff -NurpP --minimal linux-source-2.6.22.orig/arch/parisc/mm/fault.c linux-source-2.6.22/arch/parisc/mm/fault.c --- linux-source-2.6.22.orig/arch/parisc/mm/fault.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/parisc/mm/fault.c 2007-10-24 16:14:05.791544531 +0200 @@ -209,8 +209,9 @@ bad_area: #ifdef PRINT_USER_FAULTS printk(KERN_DEBUG "\n"); - printk(KERN_DEBUG "do_page_fault() pid=%d command='%s' type=%lu address=0x%08lx\n", - tsk->pid, tsk->comm, code, address); + printk(KERN_DEBUG "do_page_fault() pid=%d:#%u " + "command='%s' type=%lu address=0x%08lx\n", + tsk->pid, tsk->xid, tsk->comm, code, address); if (vma) { printk(KERN_DEBUG "vm_start = 0x%08lx, vm_end = 0x%08lx\n", vma->vm_start, vma->vm_end); @@ -260,7 +261,8 @@ no_context: out_of_memory: up_read(&mm->mmap_sem); - printk(KERN_CRIT "VM: killing process %s\n", current->comm); + printk(KERN_CRIT "VM: killing process %s(%d:#%u)\n", + current->comm, current->pid, current->xid); if (user_mode(regs)) do_exit(SIGKILL); goto no_context; diff -NurpP --minimal linux-source-2.6.22.orig/arch/powerpc/Kconfig linux-source-2.6.22/arch/powerpc/Kconfig --- linux-source-2.6.22.orig/arch/powerpc/Kconfig 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/powerpc/Kconfig 2007-10-24 16:14:05.791544531 +0200 @@ -906,6 +906,8 @@ endmenu source "arch/powerpc/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" config KEYS_COMPAT diff -NurpP --minimal linux-source-2.6.22.orig/arch/powerpc/kernel/asm-offsets.c linux-source-2.6.22/arch/powerpc/kernel/asm-offsets.c --- linux-source-2.6.22.orig/arch/powerpc/kernel/asm-offsets.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/powerpc/kernel/asm-offsets.c 2007-10-24 16:14:05.791544531 +0200 @@ -250,6 +250,7 @@ int main(void) DEFINE(CLONE_VM, CLONE_VM); DEFINE(CLONE_UNTRACED, CLONE_UNTRACED); + DEFINE(CLONE_KTHREAD, CLONE_KTHREAD); #ifndef CONFIG_PPC64 DEFINE(MM_PGD, offsetof(struct mm_struct, pgd)); diff -NurpP --minimal linux-source-2.6.22.orig/arch/powerpc/kernel/irq.c linux-source-2.6.22/arch/powerpc/kernel/irq.c --- linux-source-2.6.22.orig/arch/powerpc/kernel/irq.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/powerpc/kernel/irq.c 2007-10-24 16:14:05.791544531 +0200 @@ -53,6 +53,7 @@ #include #include #include +#include #include #include diff -NurpP --minimal linux-source-2.6.22.orig/arch/powerpc/kernel/misc_32.S linux-source-2.6.22/arch/powerpc/kernel/misc_32.S --- linux-source-2.6.22.orig/arch/powerpc/kernel/misc_32.S 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/arch/powerpc/kernel/misc_32.S 2007-10-24 16:14:05.791544531 +0200 @@ -766,7 +766,7 @@ _GLOBAL(kernel_thread) mr r30,r3 /* function */ mr r31,r4 /* argument */ ori r3,r5,CLONE_VM /* flags */ - oris r3,r3,CLONE_UNTRACED>>16 + oris r3,r3,(CLONE_UNTRACED|CLONE_KTHREAD)>>16 li r4,0 /* new sp (unused) */ li r0,__NR_clone sc diff -NurpP --minimal linux-source-2.6.22.orig/arch/powerpc/kernel/misc_64.S linux-source-2.6.22/arch/powerpc/kernel/misc_64.S --- linux-source-2.6.22.orig/arch/powerpc/kernel/misc_64.S 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/powerpc/kernel/misc_64.S 2007-10-24 16:14:05.791544531 +0200 @@ -434,7 +434,7 @@ _GLOBAL(kernel_thread) mr r29,r3 mr r30,r4 ori r3,r5,CLONE_VM /* flags */ - oris r3,r3,(CLONE_UNTRACED>>16) + oris r3,r3,(CLONE_UNTRACED|CLONE_KTHREAD)>>16 li r4,0 /* new sp (unused) */ li r0,__NR_clone sc diff -NurpP --minimal linux-source-2.6.22.orig/arch/powerpc/kernel/process.c linux-source-2.6.22/arch/powerpc/kernel/process.c --- linux-source-2.6.22.orig/arch/powerpc/kernel/process.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/arch/powerpc/kernel/process.c 2007-10-24 16:14:05.791544531 +0200 @@ -422,8 +422,9 @@ void show_regs(struct pt_regs * regs) trap = TRAP(regs); if (trap == 0x300 || trap == 0x600) printk("DAR: "REG", DSISR: "REG"\n", regs->dar, regs->dsisr); - printk("TASK = %p[%d] '%s' THREAD: %p", - current, current->pid, current->comm, task_thread_info(current)); + printk("TASK = %p[%d,#%u] '%s' THREAD: %p", + current, current->pid, current->xid, + current->comm, task_thread_info(current)); #ifdef CONFIG_SMP printk(" CPU: %d", smp_processor_id()); diff -NurpP --minimal linux-source-2.6.22.orig/arch/powerpc/kernel/sys_ppc32.c linux-source-2.6.22/arch/powerpc/kernel/sys_ppc32.c --- linux-source-2.6.22.orig/arch/powerpc/kernel/sys_ppc32.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/powerpc/kernel/sys_ppc32.c 2007-10-24 16:14:05.791544531 +0200 @@ -205,7 +205,7 @@ asmlinkage long compat_sys_gettimeofday( { if (tv) { struct timeval ktv; - do_gettimeofday(&ktv); + vx_gettimeofday(&ktv); if (put_tv32(tv, &ktv)) return -EFAULT; } diff -NurpP --minimal linux-source-2.6.22.orig/arch/powerpc/kernel/traps.c linux-source-2.6.22/arch/powerpc/kernel/traps.c --- linux-source-2.6.22.orig/arch/powerpc/kernel/traps.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/powerpc/kernel/traps.c 2007-10-24 16:14:05.791544531 +0200 @@ -878,8 +878,9 @@ void nonrecoverable_exception(struct pt_ void trace_syscall(struct pt_regs *regs) { - printk("Task: %p(%d), PC: %08lX/%08lX, Syscall: %3ld, Result: %s%ld %s\n", - current, current->pid, regs->nip, regs->link, regs->gpr[0], + printk("Task: %p(%d[#%u]), PC: %08lX/%08lX, Syscall: %3ld, Result: %s%ld %s\n", + current, current->pid, current->xid, + regs->nip, regs->link, regs->gpr[0], regs->ccr&0x10000000?"Error=":"", regs->gpr[3], print_tainted()); } diff -NurpP --minimal linux-source-2.6.22.orig/arch/powerpc/kernel/vdso.c linux-source-2.6.22/arch/powerpc/kernel/vdso.c --- linux-source-2.6.22.orig/arch/powerpc/kernel/vdso.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/powerpc/kernel/vdso.c 2007-10-24 16:14:05.791544531 +0200 @@ -21,6 +21,7 @@ #include #include #include +#include #include #include diff -NurpP --minimal linux-source-2.6.22.orig/arch/powerpc/mm/fault.c linux-source-2.6.22/arch/powerpc/mm/fault.c --- linux-source-2.6.22.orig/arch/powerpc/mm/fault.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/powerpc/mm/fault.c 2007-10-24 16:14:05.791544531 +0200 @@ -378,7 +378,8 @@ out_of_memory: down_read(&mm->mmap_sem); goto survive; } - printk("VM: killing process %s\n", current->comm); + printk("VM: killing process %s(%d:#%u)\n", + current->comm, current->pid, current->xid); if (user_mode(regs)) do_exit(SIGKILL); return SIGKILL; diff -NurpP --minimal linux-source-2.6.22.orig/arch/ppc/Kconfig linux-source-2.6.22/arch/ppc/Kconfig --- linux-source-2.6.22.orig/arch/ppc/Kconfig 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/ppc/Kconfig 2007-10-24 16:14:05.791544531 +0200 @@ -1455,6 +1455,8 @@ source "arch/powerpc/oprofile/Kconfig" source "arch/ppc/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-source-2.6.22.orig/arch/ppc/kernel/asm-offsets.c linux-source-2.6.22/arch/ppc/kernel/asm-offsets.c --- linux-source-2.6.22.orig/arch/ppc/kernel/asm-offsets.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/ppc/kernel/asm-offsets.c 2007-10-24 16:14:05.791544531 +0200 @@ -120,6 +120,7 @@ main(void) DEFINE(TRAP, STACK_FRAME_OVERHEAD+offsetof(struct pt_regs, trap)); DEFINE(CLONE_VM, CLONE_VM); DEFINE(CLONE_UNTRACED, CLONE_UNTRACED); + DEFINE(CLONE_KTHREAD, CLONE_KTHREAD); DEFINE(MM_PGD, offsetof(struct mm_struct, pgd)); /* About the CPU features table */ diff -NurpP --minimal linux-source-2.6.22.orig/arch/ppc/kernel/misc.S linux-source-2.6.22/arch/ppc/kernel/misc.S --- linux-source-2.6.22.orig/arch/ppc/kernel/misc.S 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/ppc/kernel/misc.S 2007-10-24 16:14:05.791544531 +0200 @@ -848,7 +848,7 @@ _GLOBAL(kernel_thread) mr r30,r3 /* function */ mr r31,r4 /* argument */ ori r3,r5,CLONE_VM /* flags */ - oris r3,r3,CLONE_UNTRACED>>16 + oris r3,r3,(CLONE_UNTRACED|CLONE_KTHREAD)>>16 li r4,0 /* new sp (unused) */ li r0,__NR_clone sc diff -NurpP --minimal linux-source-2.6.22.orig/arch/ppc/kernel/traps.c linux-source-2.6.22/arch/ppc/kernel/traps.c --- linux-source-2.6.22.orig/arch/ppc/kernel/traps.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/ppc/kernel/traps.c 2007-10-24 16:14:05.791544531 +0200 @@ -696,8 +696,9 @@ void nonrecoverable_exception(struct pt_ void trace_syscall(struct pt_regs *regs) { - printk("Task: %p(%d), PC: %08lX/%08lX, Syscall: %3ld, Result: %s%ld %s\n", - current, current->pid, regs->nip, regs->link, regs->gpr[0], + printk("Task: %p(%d[#%u]), PC: %08lX/%08lX, Syscall: %3ld, Result: %s%ld %s\n", + current, current->pid, current->xid, + regs->nip, regs->link, regs->gpr[0], regs->ccr&0x10000000?"Error=":"", regs->gpr[3], print_tainted()); } diff -NurpP --minimal linux-source-2.6.22.orig/arch/ppc/mm/fault.c linux-source-2.6.22/arch/ppc/mm/fault.c --- linux-source-2.6.22.orig/arch/ppc/mm/fault.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/ppc/mm/fault.c 2007-10-24 16:14:05.791544531 +0200 @@ -296,7 +296,8 @@ out_of_memory: down_read(&mm->mmap_sem); goto survive; } - printk("VM: killing process %s\n", current->comm); + printk("VM: killing process %s(%d:#%u)\n", + current->comm, current->pid, current->xid); if (user_mode(regs)) do_exit(SIGKILL); return SIGKILL; diff -NurpP --minimal linux-source-2.6.22.orig/arch/s390/Kconfig linux-source-2.6.22/arch/s390/Kconfig --- linux-source-2.6.22.orig/arch/s390/Kconfig 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/s390/Kconfig 2007-10-24 16:14:05.791544531 +0200 @@ -551,6 +551,8 @@ endmenu source "arch/s390/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-source-2.6.22.orig/arch/s390/kernel/compat_linux.c linux-source-2.6.22/arch/s390/kernel/compat_linux.c --- linux-source-2.6.22.orig/arch/s390/kernel/compat_linux.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/s390/kernel/compat_linux.c 2007-10-24 16:14:05.791544531 +0200 @@ -567,7 +567,7 @@ asmlinkage long sys32_gettimeofday(struc { if (tv) { struct timeval ktv; - do_gettimeofday(&ktv); + vx_gettimeofday(&ktv); if (put_tv32(tv, &ktv)) return -EFAULT; } diff -NurpP --minimal linux-source-2.6.22.orig/arch/s390/kernel/process.c linux-source-2.6.22/arch/s390/kernel/process.c --- linux-source-2.6.22.orig/arch/s390/kernel/process.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/s390/kernel/process.c 2007-10-24 16:14:05.791544531 +0200 @@ -164,9 +164,9 @@ void show_regs(struct pt_regs *regs) struct task_struct *tsk = current; printk("CPU: %d %s\n", task_thread_info(tsk)->cpu, print_tainted()); - printk("Process %s (pid: %d, task: %p, ksp: %p)\n", - current->comm, current->pid, (void *) tsk, - (void *) tsk->thread.ksp); + printk("Process %s (pid: %d[#%u], task: %p, ksp: %p)\n", + current->comm, current->pid, current->xid, + (void *) tsk, (void *) tsk->thread.ksp); show_registers(regs); /* Show stack backtrace if pt_regs is from kernel mode */ @@ -197,7 +197,7 @@ int kernel_thread(int (*fn)(void *), voi regs.orig_gpr2 = -1; /* Ok, create the new process.. */ - return do_fork(flags | CLONE_VM | CLONE_UNTRACED, + return do_fork(flags | CLONE_VM | CLONE_UNTRACED | CLONE_KTHREAD, 0, ®s, 0, NULL, NULL); } diff -NurpP --minimal linux-source-2.6.22.orig/arch/s390/kernel/ptrace.c linux-source-2.6.22/arch/s390/kernel/ptrace.c --- linux-source-2.6.22.orig/arch/s390/kernel/ptrace.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/s390/kernel/ptrace.c 2007-10-24 16:14:05.791544531 +0200 @@ -33,6 +33,7 @@ #include #include #include +#include #include #include @@ -725,7 +726,13 @@ sys_ptrace(long request, long pid, long goto out; } + if (!vx_check(vx_task_xid(child), VS_WATCH_P | VS_IDENT)) { + ret = -EPERM; + goto out_tsk; + } + ret = do_ptrace(child, request, addr, data); +out_tsk: put_task_struct(child); out: unlock_kernel(); diff -NurpP --minimal linux-source-2.6.22.orig/arch/s390/kernel/syscalls.S linux-source-2.6.22/arch/s390/kernel/syscalls.S --- linux-source-2.6.22.orig/arch/s390/kernel/syscalls.S 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/s390/kernel/syscalls.S 2007-10-24 16:14:05.791544531 +0200 @@ -271,7 +271,7 @@ SYSCALL(sys_clock_settime,sys_clock_sett SYSCALL(sys_clock_gettime,sys_clock_gettime,sys32_clock_gettime_wrapper) /* 260 */ SYSCALL(sys_clock_getres,sys_clock_getres,sys32_clock_getres_wrapper) SYSCALL(sys_clock_nanosleep,sys_clock_nanosleep,sys32_clock_nanosleep_wrapper) -NI_SYSCALL /* reserved for vserver */ +SYSCALL(sys_vserver,sys_vserver,sys32_vserver) SYSCALL(s390_fadvise64_64,sys_ni_syscall,sys32_fadvise64_64_wrapper) SYSCALL(sys_statfs64,sys_statfs64,compat_sys_statfs64_wrapper) SYSCALL(sys_fstatfs64,sys_fstatfs64,compat_sys_fstatfs64_wrapper) diff -NurpP --minimal linux-source-2.6.22.orig/arch/s390/mm/fault.c linux-source-2.6.22/arch/s390/mm/fault.c --- linux-source-2.6.22.orig/arch/s390/mm/fault.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/s390/mm/fault.c 2007-10-24 16:14:05.791544531 +0200 @@ -216,7 +216,8 @@ static int do_out_of_memory(struct pt_re down_read(&mm->mmap_sem); return 1; } - printk("VM: killing process %s\n", tsk->comm); + printk("VM: killing process %s(%d:#%u)\n", + tsk->comm, tsk->pid, tsk->xid); if (regs->psw.mask & PSW_MASK_PSTATE) do_exit(SIGKILL); do_no_context(regs, error_code, address); diff -NurpP --minimal linux-source-2.6.22.orig/arch/sh/Kconfig linux-source-2.6.22/arch/sh/Kconfig --- linux-source-2.6.22.orig/arch/sh/Kconfig 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/sh/Kconfig 2007-10-24 16:14:05.791544531 +0200 @@ -723,6 +723,8 @@ source "arch/sh/oprofile/Kconfig" source "arch/sh/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-source-2.6.22.orig/arch/sh/kernel/irq.c linux-source-2.6.22/arch/sh/kernel/irq.c --- linux-source-2.6.22.orig/arch/sh/kernel/irq.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/sh/kernel/irq.c 2007-10-24 16:14:05.791544531 +0200 @@ -12,6 +12,7 @@ #include #include #include +#include #include #include #include diff -NurpP --minimal linux-source-2.6.22.orig/arch/sh/kernel/process.c linux-source-2.6.22/arch/sh/kernel/process.c --- linux-source-2.6.22.orig/arch/sh/kernel/process.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/sh/kernel/process.c 2007-10-24 16:14:05.791544531 +0200 @@ -118,7 +118,8 @@ void machine_power_off(void) void show_regs(struct pt_regs * regs) { printk("\n"); - printk("Pid : %d, Comm: %20s\n", current->pid, current->comm); + printk("Pid : %d:#%u, Comm: %20s\n", + current->pid, current->xid, current->comm); print_symbol("PC is at %s\n", instruction_pointer(regs)); printk("PC : %08lx SP : %08lx SR : %08lx ", regs->pc, regs->regs[15], regs->sr); @@ -179,7 +180,7 @@ int kernel_thread(int (*fn)(void *), voi regs.sr = (1 << 30); /* Ok, create the new process.. */ - return do_fork(flags | CLONE_VM | CLONE_UNTRACED, 0, + return do_fork(flags | CLONE_VM | CLONE_UNTRACED | CLONE_KTHREAD, 0, ®s, 0, NULL, NULL); } diff -NurpP --minimal linux-source-2.6.22.orig/arch/sh/kernel/vsyscall/vsyscall.c linux-source-2.6.22/arch/sh/kernel/vsyscall/vsyscall.c --- linux-source-2.6.22.orig/arch/sh/kernel/vsyscall/vsyscall.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/sh/kernel/vsyscall/vsyscall.c 2007-10-24 16:14:06.291582822 +0200 @@ -18,6 +18,7 @@ #include #include #include +#include /* * Should the kernel map a VDSO page into processes and pass its diff -NurpP --minimal linux-source-2.6.22.orig/arch/sh/mm/fault.c linux-source-2.6.22/arch/sh/mm/fault.c --- linux-source-2.6.22.orig/arch/sh/mm/fault.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/sh/mm/fault.c 2007-10-24 16:14:06.291582822 +0200 @@ -201,7 +201,8 @@ out_of_memory: down_read(&mm->mmap_sem); goto survive; } - printk("VM: killing process %s\n", tsk->comm); + printk("VM: killing process %s(%d:#%u)\n", + tsk->comm, tsk->pid, tsk->xid); if (user_mode(regs)) do_exit(SIGKILL); goto no_context; diff -NurpP --minimal linux-source-2.6.22.orig/arch/sh64/kernel/process.c linux-source-2.6.22/arch/sh64/kernel/process.c --- linux-source-2.6.22.orig/arch/sh64/kernel/process.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/sh64/kernel/process.c 2007-10-24 16:14:06.291582822 +0200 @@ -400,8 +400,8 @@ int kernel_thread(int (*fn)(void *), voi regs.pc = (unsigned long)kernel_thread_helper; regs.sr = (1 << 30); - return do_fork(flags | CLONE_VM | CLONE_UNTRACED, 0, - ®s, 0, NULL, NULL); + return do_fork(flags | CLONE_VM | CLONE_UNTRACED | CLONE_KTHREAD, + 0, ®s, 0, NULL, NULL); } /* diff -NurpP --minimal linux-source-2.6.22.orig/arch/sh64/mm/fault.c linux-source-2.6.22/arch/sh64/mm/fault.c --- linux-source-2.6.22.orig/arch/sh64/mm/fault.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/sh64/mm/fault.c 2007-10-24 16:14:06.291582822 +0200 @@ -81,7 +81,7 @@ static inline void print_vma(struct vm_a static inline void print_task(struct task_struct *tsk) { - printk("Task pid %d\n", tsk->pid); + printk("Task pid %d:#%u\n", tsk->pid, tsk->xid); } static pte_t *lookup_pte(struct mm_struct *mm, unsigned long address) @@ -330,7 +330,8 @@ out_of_memory: down_read(&mm->mmap_sem); goto survive; } - printk("VM: killing process %s\n", tsk->comm); + printk("VM: killing process %s(%d:#%u)\n", + tsk->comm, tsk->pid, tsk->xid); if (user_mode(regs)) do_exit(SIGKILL); goto no_context; diff -NurpP --minimal linux-source-2.6.22.orig/arch/sparc/Kconfig linux-source-2.6.22/arch/sparc/Kconfig --- linux-source-2.6.22.orig/arch/sparc/Kconfig 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/sparc/Kconfig 2007-10-24 16:14:06.291582822 +0200 @@ -317,6 +317,8 @@ endmenu source "arch/sparc/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-source-2.6.22.orig/arch/sparc/kernel/process.c linux-source-2.6.22/arch/sparc/kernel/process.c --- linux-source-2.6.22.orig/arch/sparc/kernel/process.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/arch/sparc/kernel/process.c 2007-10-24 16:14:06.291582822 +0200 @@ -707,7 +707,8 @@ pid_t kernel_thread(int (*fn)(void *), v /* Notreached by child. */ "1: mov %%o0, %0\n\t" : "=r" (retval) : - "i" (__NR_clone), "r" (flags | CLONE_VM | CLONE_UNTRACED), + "i" (__NR_clone), "r" (flags | + CLONE_VM | CLONE_UNTRACED | CLONE_KTHREAD), "i" (__NR_exit), "r" (fn), "r" (arg) : "g1", "g2", "g3", "o0", "o1", "memory", "cc"); return retval; diff -NurpP --minimal linux-source-2.6.22.orig/arch/sparc/kernel/ptrace.c linux-source-2.6.22/arch/sparc/kernel/ptrace.c --- linux-source-2.6.22.orig/arch/sparc/kernel/ptrace.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/sparc/kernel/ptrace.c 2007-10-24 16:14:06.291582822 +0200 @@ -19,6 +19,7 @@ #include #include #include +#include #include #include @@ -302,6 +303,10 @@ asmlinkage void do_ptrace(struct pt_regs pt_error_return(regs, -ret); goto out; } + if (!vx_check(vx_task_xid(child), VS_WATCH_P | VS_IDENT)) { + pt_error_return(regs, ESRCH); + goto out_tsk; + } if ((current->personality == PER_SUNOS && request == PTRACE_SUNATTACH) || (current->personality != PER_SUNOS && request == PTRACE_ATTACH)) { diff -NurpP --minimal linux-source-2.6.22.orig/arch/sparc/kernel/systbls.S linux-source-2.6.22/arch/sparc/kernel/systbls.S --- linux-source-2.6.22.orig/arch/sparc/kernel/systbls.S 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/sparc/kernel/systbls.S 2007-10-24 16:14:06.291582822 +0200 @@ -71,7 +71,7 @@ sys_call_table: /*250*/ .long sparc_mremap, sys_sysctl, sys_getsid, sys_fdatasync, sys_nfsservctl /*255*/ .long sys_sync_file_range, sys_clock_settime, sys_clock_gettime, sys_clock_getres, sys_clock_nanosleep /*260*/ .long sys_sched_getaffinity, sys_sched_setaffinity, sys_timer_settime, sys_timer_gettime, sys_timer_getoverrun -/*265*/ .long sys_timer_delete, sys_timer_create, sys_nis_syscall, sys_io_setup, sys_io_destroy +/*265*/ .long sys_timer_delete, sys_timer_create, sys_vserver, sys_io_setup, sys_io_destroy /*270*/ .long sys_io_submit, sys_io_cancel, sys_io_getevents, sys_mq_open, sys_mq_unlink /*275*/ .long sys_mq_timedsend, sys_mq_timedreceive, sys_mq_notify, sys_mq_getsetattr, sys_waitid /*280*/ .long sys_tee, sys_add_key, sys_request_key, sys_keyctl, sys_openat diff -NurpP --minimal linux-source-2.6.22.orig/arch/sparc/kernel/traps.c linux-source-2.6.22/arch/sparc/kernel/traps.c --- linux-source-2.6.22.orig/arch/sparc/kernel/traps.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/sparc/kernel/traps.c 2007-10-24 16:14:06.291582822 +0200 @@ -99,7 +99,8 @@ void die_if_kernel(char *str, struct pt_ " /_| \\__/ |_\\\n" " \\__U_/\n"); - printk("%s(%d): %s [#%d]\n", current->comm, current->pid, str, ++die_counter); + printk("%s(%d[#%u]): %s [#%d]\n", current->comm, + current->pid, current->xid, str, ++die_counter); show_regs(regs); __SAVE; __SAVE; __SAVE; __SAVE; diff -NurpP --minimal linux-source-2.6.22.orig/arch/sparc/mm/fault.c linux-source-2.6.22/arch/sparc/mm/fault.c --- linux-source-2.6.22.orig/arch/sparc/mm/fault.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/sparc/mm/fault.c 2007-10-24 16:14:06.291582822 +0200 @@ -367,7 +367,8 @@ no_context: */ out_of_memory: up_read(&mm->mmap_sem); - printk("VM: killing process %s\n", tsk->comm); + printk("VM: killing process %s(%d:#%u)\n", + tsk->comm, tsk->pid, tsk->xid); if (from_user) do_exit(SIGKILL); goto no_context; diff -NurpP --minimal linux-source-2.6.22.orig/arch/sparc64/Kconfig linux-source-2.6.22/arch/sparc64/Kconfig --- linux-source-2.6.22.orig/arch/sparc64/Kconfig 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/arch/sparc64/Kconfig 2007-10-24 16:14:06.291582822 +0200 @@ -466,6 +466,8 @@ endmenu source "arch/sparc64/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-source-2.6.22.orig/arch/sparc64/kernel/binfmt_aout32.c linux-source-2.6.22/arch/sparc64/kernel/binfmt_aout32.c --- linux-source-2.6.22.orig/arch/sparc64/kernel/binfmt_aout32.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/sparc64/kernel/binfmt_aout32.c 2007-10-24 16:14:06.291582822 +0200 @@ -27,6 +27,7 @@ #include #include #include +#include #include #include diff -NurpP --minimal linux-source-2.6.22.orig/arch/sparc64/kernel/process.c linux-source-2.6.22/arch/sparc64/kernel/process.c --- linux-source-2.6.22.orig/arch/sparc64/kernel/process.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/arch/sparc64/kernel/process.c 2007-10-24 16:14:06.291582822 +0200 @@ -718,7 +718,8 @@ pid_t kernel_thread(int (*fn)(void *), v /* Notreached by child. */ "1:" : "=r" (retval) : - "i" (__NR_clone), "r" (flags | CLONE_VM | CLONE_UNTRACED), + "i" (__NR_clone), "r" (flags | + CLONE_VM | CLONE_UNTRACED | CLONE_KTHREAD), "i" (__NR_exit), "r" (fn), "r" (arg) : "g1", "g2", "g3", "o0", "o1", "memory", "cc"); return retval; diff -NurpP --minimal linux-source-2.6.22.orig/arch/sparc64/kernel/ptrace.c linux-source-2.6.22/arch/sparc64/kernel/ptrace.c --- linux-source-2.6.22.orig/arch/sparc64/kernel/ptrace.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/sparc64/kernel/ptrace.c 2007-10-24 16:14:06.291582822 +0200 @@ -22,6 +22,7 @@ #include #include #include +#include #include #include @@ -215,6 +216,10 @@ asmlinkage void do_ptrace(struct pt_regs pt_error_return(regs, -ret); goto out; } + if (!vx_check(vx_task_xid(child), VS_WATCH_P | VS_IDENT)) { + pt_error_return(regs, ESRCH); + goto out_tsk; + } if ((current->personality == PER_SUNOS && request == PTRACE_SUNATTACH) || (current->personality != PER_SUNOS && request == PTRACE_ATTACH)) { diff -NurpP --minimal linux-source-2.6.22.orig/arch/sparc64/kernel/sys_sparc32.c linux-source-2.6.22/arch/sparc64/kernel/sys_sparc32.c --- linux-source-2.6.22.orig/arch/sparc64/kernel/sys_sparc32.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/sparc64/kernel/sys_sparc32.c 2007-10-24 16:14:06.291582822 +0200 @@ -729,7 +729,7 @@ asmlinkage long sys32_gettimeofday(struc { if (tv) { struct timeval ktv; - do_gettimeofday(&ktv); + vx_gettimeofday(&ktv); if (put_tv32(tv, &ktv)) return -EFAULT; } diff -NurpP --minimal linux-source-2.6.22.orig/arch/sparc64/kernel/systbls.S linux-source-2.6.22/arch/sparc64/kernel/systbls.S --- linux-source-2.6.22.orig/arch/sparc64/kernel/systbls.S 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/sparc64/kernel/systbls.S 2007-10-24 16:14:06.291582822 +0200 @@ -72,7 +72,7 @@ sys_call_table32: /*250*/ .word sys32_mremap, sys32_sysctl, sys32_getsid, sys_fdatasync, sys32_nfsservctl .word sys32_sync_file_range, compat_sys_clock_settime, compat_sys_clock_gettime, compat_sys_clock_getres, sys32_clock_nanosleep /*260*/ .word compat_sys_sched_getaffinity, compat_sys_sched_setaffinity, sys32_timer_settime, compat_sys_timer_gettime, sys_timer_getoverrun - .word sys_timer_delete, compat_sys_timer_create, sys_ni_syscall, compat_sys_io_setup, sys_io_destroy + .word sys_timer_delete, compat_sys_timer_create, sys32_vserver, compat_sys_io_setup, sys_io_destroy /*270*/ .word sys32_io_submit, sys_io_cancel, compat_sys_io_getevents, sys32_mq_open, sys_mq_unlink .word compat_sys_mq_timedsend, compat_sys_mq_timedreceive, compat_sys_mq_notify, compat_sys_mq_getsetattr, compat_sys_waitid /*280*/ .word sys32_tee, sys_add_key, sys_request_key, sys_keyctl, compat_sys_openat @@ -144,7 +144,7 @@ sys_call_table: /*250*/ .word sys64_mremap, sys_sysctl, sys_getsid, sys_fdatasync, sys_nfsservctl .word sys_sync_file_range, sys_clock_settime, sys_clock_gettime, sys_clock_getres, sys_clock_nanosleep /*260*/ .word sys_sched_getaffinity, sys_sched_setaffinity, sys_timer_settime, sys_timer_gettime, sys_timer_getoverrun - .word sys_timer_delete, sys_timer_create, sys_ni_syscall, sys_io_setup, sys_io_destroy + .word sys_timer_delete, sys_timer_create, sys_vserver, sys_io_setup, sys_io_destroy /*270*/ .word sys_io_submit, sys_io_cancel, sys_io_getevents, sys_mq_open, sys_mq_unlink .word sys_mq_timedsend, sys_mq_timedreceive, sys_mq_notify, sys_mq_getsetattr, sys_waitid /*280*/ .word sys_tee, sys_add_key, sys_request_key, sys_keyctl, sys_openat diff -NurpP --minimal linux-source-2.6.22.orig/arch/sparc64/kernel/traps.c linux-source-2.6.22/arch/sparc64/kernel/traps.c --- linux-source-2.6.22.orig/arch/sparc64/kernel/traps.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/arch/sparc64/kernel/traps.c 2007-10-24 16:14:06.291582822 +0200 @@ -2225,7 +2225,8 @@ void die_if_kernel(char *str, struct pt_ " /_| \\__/ |_\\\n" " \\__U_/\n"); - printk("%s(%d): %s [#%d]\n", current->comm, current->pid, str, ++die_counter); + printk("%s(%d[#%u]): %s [#%d]\n", current->comm, + current->pid, current->xid, str, ++die_counter); notify_die(DIE_OOPS, str, regs, 0, 255, SIGSEGV); __asm__ __volatile__("flushw"); __show_regs(regs); diff -NurpP --minimal linux-source-2.6.22.orig/arch/sparc64/mm/fault.c linux-source-2.6.22/arch/sparc64/mm/fault.c --- linux-source-2.6.22.orig/arch/sparc64/mm/fault.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/arch/sparc64/mm/fault.c 2007-10-24 16:14:06.291582822 +0200 @@ -463,7 +463,8 @@ handle_kernel_fault: out_of_memory: insn = get_fault_insn(regs, insn); up_read(&mm->mmap_sem); - printk("VM: killing process %s\n", current->comm); + printk("VM: killing process %s(%d:#%u)\n", + current->comm, current->pid, current->xid); if (!(regs->tstate & TSTATE_PRIV)) do_exit(SIGKILL); goto handle_kernel_fault; diff -NurpP --minimal linux-source-2.6.22.orig/arch/sparc64/solaris/fs.c linux-source-2.6.22/arch/sparc64/solaris/fs.c --- linux-source-2.6.22.orig/arch/sparc64/solaris/fs.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/sparc64/solaris/fs.c 2007-10-24 16:14:06.291582822 +0200 @@ -368,7 +368,7 @@ static int report_statvfs(struct vfsmoun int j = strlen (p); if (j > 15) j = 15; - if (IS_RDONLY(inode)) i = 1; + if (IS_RDONLY(inode) || MNT_IS_RDONLY(mnt)) i = 1; if (mnt->mnt_flags & MNT_NOSUID) i |= 2; if (!sysv_valid_dev(inode->i_sb->s_dev)) return -EOVERFLOW; @@ -404,7 +404,7 @@ static int report_statvfs64(struct vfsmo int j = strlen (p); if (j > 15) j = 15; - if (IS_RDONLY(inode)) i = 1; + if (IS_RDONLY(inode) || MNT_IS_RDONLY(mnt)) i = 1; if (mnt->mnt_flags & MNT_NOSUID) i |= 2; if (!sysv_valid_dev(inode->i_sb->s_dev)) return -EOVERFLOW; diff -NurpP --minimal linux-source-2.6.22.orig/arch/um/Kconfig linux-source-2.6.22/arch/um/Kconfig --- linux-source-2.6.22.orig/arch/um/Kconfig 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/um/Kconfig 2007-10-24 16:14:06.291582822 +0200 @@ -316,6 +316,8 @@ source "drivers/connector/Kconfig" source "fs/Kconfig" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-source-2.6.22.orig/arch/um/kernel/trap.c linux-source-2.6.22/arch/um/kernel/trap.c --- linux-source-2.6.22.orig/arch/um/kernel/trap.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/um/kernel/trap.c 2007-10-24 16:14:06.291582822 +0200 @@ -210,7 +210,8 @@ unsigned long segv(struct faultinfo fi, current->thread.arch.faultinfo = fi; force_sig_info(SIGBUS, &si, current); } else if (err == -ENOMEM) { - printk("VM: killing process %s\n", current->comm); + printk("VM: killing process %s(%d:#%u)\n", + current->comm, current->pid, current->xid); do_exit(SIGKILL); } else { BUG_ON(err != -EFAULT); diff -NurpP --minimal linux-source-2.6.22.orig/arch/v850/Kconfig linux-source-2.6.22/arch/v850/Kconfig --- linux-source-2.6.22.orig/arch/v850/Kconfig 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/v850/Kconfig 2007-10-24 16:14:06.291582822 +0200 @@ -333,6 +333,8 @@ source "drivers/usb/Kconfig" source "arch/v850/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-source-2.6.22.orig/arch/v850/kernel/process.c linux-source-2.6.22/arch/v850/kernel/process.c --- linux-source-2.6.22.orig/arch/v850/kernel/process.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/v850/kernel/process.c 2007-10-24 16:14:06.291582822 +0200 @@ -82,7 +82,7 @@ int kernel_thread (int (*fn)(void *), vo /* Clone this thread. Note that we don't pass the clone syscall's second argument -- it's ignored for calls from kernel mode (the child's SP is always set to the top of the kernel stack). */ - arg0 = flags | CLONE_VM; + arg0 = flags | CLONE_VM | CLONE_KTHREAD; syscall = __NR_clone; asm volatile ("trap " SYSCALL_SHORT_TRAP : "=r" (ret), "=r" (syscall) diff -NurpP --minimal linux-source-2.6.22.orig/arch/v850/kernel/ptrace.c linux-source-2.6.22/arch/v850/kernel/ptrace.c --- linux-source-2.6.22.orig/arch/v850/kernel/ptrace.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/v850/kernel/ptrace.c 2007-10-24 16:14:06.291582822 +0200 @@ -23,6 +23,7 @@ #include #include #include +#include #include #include @@ -116,6 +117,9 @@ long arch_ptrace(struct task_struct *chi { int rval; + if (!vx_check(vx_task_xid(child), VS_WATCH_P | VS_IDENT)) + goto out; + switch (request) { unsigned long val, copied; diff -NurpP --minimal linux-source-2.6.22.orig/arch/x86_64/ia32/ia32_aout.c linux-source-2.6.22/arch/x86_64/ia32/ia32_aout.c --- linux-source-2.6.22.orig/arch/x86_64/ia32/ia32_aout.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/x86_64/ia32/ia32_aout.c 2007-10-24 16:14:06.291582822 +0200 @@ -25,6 +25,7 @@ #include #include #include +#include #include #include diff -NurpP --minimal linux-source-2.6.22.orig/arch/x86_64/ia32/ia32_binfmt.c linux-source-2.6.22/arch/x86_64/ia32/ia32_binfmt.c --- linux-source-2.6.22.orig/arch/x86_64/ia32/ia32_binfmt.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/x86_64/ia32/ia32_binfmt.c 2007-10-24 16:14:06.291582822 +0200 @@ -324,7 +324,8 @@ int ia32_setup_arg_pages(struct linux_bi kmem_cache_free(vm_area_cachep, mpnt); return ret; } - mm->stack_vm = mm->total_vm = vma_pages(mpnt); + vx_vmpages_sub(mm, mm->total_vm - vma_pages(mpnt)); + mm->stack_vm = mm->total_vm; } for (i = 0 ; i < MAX_ARG_PAGES ; i++) { diff -NurpP --minimal linux-source-2.6.22.orig/arch/x86_64/ia32/ia32entry.S linux-source-2.6.22/arch/x86_64/ia32/ia32entry.S --- linux-source-2.6.22.orig/arch/x86_64/ia32/ia32entry.S 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/arch/x86_64/ia32/ia32entry.S 2007-10-24 16:14:06.291582822 +0200 @@ -680,7 +680,7 @@ ia32_sys_call_table: .quad sys_tgkill /* 270 */ .quad compat_sys_utimes .quad sys32_fadvise64_64 - .quad quiet_ni_syscall /* sys_vserver */ + .quad sys32_vserver .quad sys_mbind .quad compat_sys_get_mempolicy /* 275 */ .quad sys_set_mempolicy diff -NurpP --minimal linux-source-2.6.22.orig/arch/x86_64/ia32/syscall32.c linux-source-2.6.22/arch/x86_64/ia32/syscall32.c --- linux-source-2.6.22.orig/arch/x86_64/ia32/syscall32.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/x86_64/ia32/syscall32.c 2007-10-24 16:14:06.291582822 +0200 @@ -10,6 +10,7 @@ #include #include #include +#include #include #include #include diff -NurpP --minimal linux-source-2.6.22.orig/arch/x86_64/ia32/sys_ia32.c linux-source-2.6.22/arch/x86_64/ia32/sys_ia32.c --- linux-source-2.6.22.orig/arch/x86_64/ia32/sys_ia32.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/x86_64/ia32/sys_ia32.c 2007-10-24 16:14:06.291582822 +0200 @@ -454,7 +454,7 @@ sys32_gettimeofday(struct compat_timeval { if (tv) { struct timeval ktv; - do_gettimeofday(&ktv); + vx_gettimeofday(&ktv); if (put_tv32(tv, &ktv)) return -EFAULT; } diff -NurpP --minimal linux-source-2.6.22.orig/arch/x86_64/Kconfig linux-source-2.6.22/arch/x86_64/Kconfig --- linux-source-2.6.22.orig/arch/x86_64/Kconfig 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/x86_64/Kconfig 2007-10-24 16:14:06.291582822 +0200 @@ -782,6 +782,8 @@ endmenu source "arch/x86_64/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-source-2.6.22.orig/arch/x86_64/kernel/process.c linux-source-2.6.22/arch/x86_64/kernel/process.c --- linux-source-2.6.22.orig/arch/x86_64/kernel/process.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/x86_64/kernel/process.c 2007-10-24 16:14:06.291582822 +0200 @@ -54,7 +54,8 @@ asmlinkage extern void ret_from_fork(void); -unsigned long kernel_thread_flags = CLONE_VM | CLONE_UNTRACED; +unsigned long kernel_thread_flags = + CLONE_VM | CLONE_UNTRACED | CLONE_KTHREAD; unsigned long boot_option_idle_override = 0; EXPORT_SYMBOL(boot_option_idle_override); @@ -310,8 +311,8 @@ void __show_regs(struct pt_regs * regs) printk("\n"); print_modules(); - printk("Pid: %d, comm: %.20s %s %s %.*s\n", - current->pid, current->comm, print_tainted(), + printk("Pid: %d:#%u, comm: %.20s %s %s %.*s\n", + current->pid, current->xid, current->comm, print_tainted(), init_utsname()->release, (int)strcspn(init_utsname()->version, " "), init_utsname()->version); diff -NurpP --minimal linux-source-2.6.22.orig/arch/x86_64/kernel/traps.c linux-source-2.6.22/arch/x86_64/kernel/traps.c --- linux-source-2.6.22.orig/arch/x86_64/kernel/traps.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/x86_64/kernel/traps.c 2007-10-24 16:14:06.291582822 +0200 @@ -413,8 +413,9 @@ void show_registers(struct pt_regs *regs rsp = regs->rsp; printk("CPU %d ", cpu); __show_regs(regs); - printk("Process %s (pid: %d, threadinfo %p, task %p)\n", - cur->comm, cur->pid, task_thread_info(cur), cur); + printk("Process %s (pid: %d[#%u], threadinfo %p, task %p)\n", + cur->comm, cur->pid, cur->xid, + task_thread_info(cur), cur); /* * When in-kernel, we also print out the stack and code at the @@ -580,8 +581,8 @@ static void __kprobes do_trap(int trapnr if (exception_trace && unhandled_signal(tsk, signr)) printk(KERN_INFO - "%s[%d] trap %s rip:%lx rsp:%lx error:%lx\n", - tsk->comm, tsk->pid, str, + "%s[%d:#%u] trap %s rip:%lx rsp:%lx error:%lx\n", + tsk->comm, tsk->pid, tsk->xid, str, regs->rip, regs->rsp, error_code); if (info) @@ -684,8 +685,8 @@ asmlinkage void __kprobes do_general_pro if (exception_trace && unhandled_signal(tsk, SIGSEGV)) printk(KERN_INFO - "%s[%d] general protection rip:%lx rsp:%lx error:%lx\n", - tsk->comm, tsk->pid, + "%s[%d:#%u] general protection rip:%lx rsp:%lx error:%lx\n", + tsk->comm, tsk->pid, tsk->xid, regs->rip, regs->rsp, error_code); force_sig(SIGSEGV, tsk); diff -NurpP --minimal linux-source-2.6.22.orig/arch/x86_64/mm/fault.c linux-source-2.6.22/arch/x86_64/mm/fault.c --- linux-source-2.6.22.orig/arch/x86_64/mm/fault.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/arch/x86_64/mm/fault.c 2007-10-24 16:14:06.291582822 +0200 @@ -497,10 +497,10 @@ bad_area_nosemaphore: if (exception_trace && unhandled_signal(tsk, SIGSEGV)) { printk( - "%s%s[%d]: segfault at %016lx rip %016lx rsp %016lx error %lx\n", + "%s%s[%d:#%u]: segfault at %016lx rip %016lx rsp %016lx error %lx\n", tsk->pid > 1 ? KERN_INFO : KERN_EMERG, - tsk->comm, tsk->pid, address, regs->rip, - regs->rsp, error_code); + tsk->comm, tsk->pid, tsk->xid, address, + regs->rip, regs->rsp, error_code); } tsk->thread.cr2 = address; @@ -567,7 +567,8 @@ out_of_memory: yield(); goto again; } - printk("VM: killing process %s\n", tsk->comm); + printk("VM: killing process %s(%d:#%u)\n", + tsk->comm, tsk->pid, tsk->xid); if (error_code & 4) do_exit(SIGKILL); goto no_context; diff -NurpP --minimal linux-source-2.6.22.orig/Documentation/vserver/debug.txt linux-source-2.6.22/Documentation/vserver/debug.txt --- linux-source-2.6.22.orig/Documentation/vserver/debug.txt 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/Documentation/vserver/debug.txt 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,154 @@ + +debug_cvirt: + + 2 4 "vx_map_tgid: %p/%llx: %d -> %d" + "vx_rmap_tgid: %p/%llx: %d -> %d" + +debug_dlim: + + 0 1 "ALLOC (%p,#%d)%c inode (%d)" + "FREE (%p,#%d)%c inode" + 1 2 "ALLOC (%p,#%d)%c %lld bytes (%d)" + "FREE (%p,#%d)%c %lld bytes" + 2 4 "ADJUST: %lld,%lld on %ld,%ld [mult=%d]" + 3 8 "ext3_has_free_blocks(%p): %lu<%lu+1, %c, %u!=%u r=%d" + "ext3_has_free_blocks(%p): free=%lu, root=%lu" + "rcu_free_dl_info(%p)" + 4 10 "alloc_dl_info(%p,%d) = %p" + "dealloc_dl_info(%p)" + "get_dl_info(%p[#%d.%d])" + "put_dl_info(%p[#%d.%d])" + 5 20 "alloc_dl_info(%p,%d)*" + 6 40 "__hash_dl_info: %p[#%d]" + "__unhash_dl_info: %p[#%d]" + 7 80 "locate_dl_info(%p,#%d) = %p" + +debug_misc: + + 0 1 "destroy_dqhash: %p [#0x%08x] c=%d" + "new_dqhash: %p [#0x%08x]" + "vroot[%d]_clr_dev: dev=%p[%lu,%d:%d]" + "vroot[%d]_get_real_bdev: dev=%p[%lu,%d:%d]" + "vroot[%d]_set_dev: dev=%p[%lu,%d:%d]" + "vroot_get_real_bdev not set" + 1 2 "cow_break_link(»%s«)" + "temp copy »%s«" + 2 4 "dentry_open(new): %p" + "dentry_open(old): %p" + "lookup_create(new): %p" + "old path »%s«" + "path_lookup(old): %d" + "vfs_create(new): %d" + "vfs_rename: %d" + "vfs_sendfile: %d" + 3 8 "fput(new_file=%p[#%d])" + "fput(old_file=%p[#%d])" + 4 10 "vx_info_kill(%p[#%d],%d,%d) = %d" + "vx_info_kill(%p[#%d],%d,%d)*" + 5 20 "vs_reboot(%p[#%d],%d)" + 6 40 "dropping task %p[#%u,%u] for %p[#%u,%u]" + +debug_net: + + 2 4 "nx_addr_conflict(%p,%p) %d.%d,%d.%d" + 3 8 "inet_bind(%p) %d.%d.%d.%d, %d.%d.%d.%d, %d.%d.%d.%d" + "inet_bind(%p)* %p,%p;%lx %d.%d.%d.%d" + 4 10 "ip_route_connect(%p) %p,%p;%lx" + 5 20 "__addr_in_socket(%p,%d.%d.%d.%d) %p:%d.%d.%d.%d %p;%lx" + 6 40 "sk,egf: %p [#%d] (from %d)" + "sk,egn: %p [#%d] (from %d)" + "sk,req: %p [#%d] (from %d)" + "sk: %p [#%d] (from %d)" + "tw: %p [#%d] (from %d)" + 7 80 "__sock_recvmsg: %p[%p,%p,%p;%d]:%d/%d" + "__sock_sendmsg: %p[%p,%p,%p;%d]:%d/%d" + +debug_nid: + + 0 1 "__lookup_nx_info(#%u): %p[#%u]" + "alloc_nx_info(%d) = %p" + "create_nx_info(%d) (dynamic rejected)" + "create_nx_info(%d) = %p (already there)" + "create_nx_info(%d) = %p (new)" + "dealloc_nx_info(%p)" + 1 2 "alloc_nx_info(%d)*" + "create_nx_info(%d)*" + 2 4 "get_nx_info(%p[#%d.%d])" + "put_nx_info(%p[#%d.%d])" + 3 8 "claim_nx_info(%p[#%d.%d.%d]) %p" + "clr_nx_info(%p[#%d.%d])" + "init_nx_info(%p[#%d.%d])" + "release_nx_info(%p[#%d.%d.%d]) %p" + "set_nx_info(%p[#%d.%d])" + 4 10 "__hash_nx_info: %p[#%d]" + "__nx_dynamic_id: [#%d]" + "__unhash_nx_info: %p[#%d.%d.%d]" + 5 20 "moved task %p into nxi:%p[#%d]" + "nx_migrate_task(%p,%p[#%d.%d.%d])" + "task_get_nx_info(%p)" + 6 40 "nx_clear_persistent(%p[#%d])" + +debug_quota: + + 0 1 "quota_sync_dqh(%p,%d) discard inode %p" + 1 2 "quota_sync_dqh(%p,%d)" + "sync_dquots(%p,%d)" + "sync_dquots_dqh(%p,%d)" + 3 8 "do_quotactl(%p,%d,cmd=%d,id=%d,%p)" + +debug_switch: + + 0 1 "vc: VCMD_%02d_%d[%d], %d,%p [%d,%d,%x,%x]" + 1 2 "vc: VCMD_%02d_%d[%d] = %08lx(%ld) [%d,%d]" + 4 10 "%s: (%s %s) returned %s with %d" + +debug_tag: + + 7 80 "dx_parse_tag(»%s«): %d:#%d" + "dx_propagate_tag(%p[#%lu.%d]): %d,%d" + +debug_xid: + + 0 1 "__lookup_vx_info(#%u): %p[#%u]" + "alloc_vx_info(%d) = %p" + "alloc_vx_info(%d)*" + "create_vx_info(%d) (dynamic rejected)" + "create_vx_info(%d) = %p (already there)" + "create_vx_info(%d) = %p (new)" + "dealloc_vx_info(%p)" + "loc_vx_info(%d) = %p (found)" + "loc_vx_info(%d) = %p (new)" + "loc_vx_info(%d) = %p (not available)" + 1 2 "create_vx_info(%d)*" + "loc_vx_info(%d)*" + 2 4 "get_vx_info(%p[#%d.%d])" + "put_vx_info(%p[#%d.%d])" + 3 8 "claim_vx_info(%p[#%d.%d.%d]) %p" + "clr_vx_info(%p[#%d.%d])" + "init_vx_info(%p[#%d.%d])" + "release_vx_info(%p[#%d.%d.%d]) %p" + "set_vx_info(%p[#%d.%d])" + 4 10 "__hash_vx_info: %p[#%d]" + "__unhash_vx_info: %p[#%d.%d.%d]" + "__vx_dynamic_id: [#%d]" + 5 20 "enter_vx_info(%p[#%d],%p) %p[#%d,%p]" + "leave_vx_info(%p[#%d,%p]) %p[#%d,%p]" + "moved task %p into vxi:%p[#%d]" + "task_get_vx_info(%p)" + "vx_migrate_task(%p,%p[#%d.%d])" + 6 40 "vx_clear_persistent(%p[#%d])" + "vx_exit_init(%p[#%d],%p[#%d,%d,%d])" + "vx_set_init(%p[#%d],%p[#%d,%d,%d])" + "vx_set_persistent(%p[#%d])" + "vx_set_reaper(%p[#%d],%p[#%d,%d])" + 7 80 "vx_child_reaper(%p[#%u,%u]) = %p[#%u,%u]" + + +debug_limit: + + n 2^n "vx_acc_cres[%5d,%s,%2d]: %5d%s" + "vx_cres_avail[%5d,%s,%2d]: %5ld > %5d + %5d" + + m 2^m "vx_acc_page[%5d,%s,%2d]: %5d%s" + "vx_acc_pages[%5d,%s,%2d]: %5d += %5d" + "vx_pages_avail[%5d,%s,%2d]: %5ld > %5d + %5d" diff -NurpP --minimal linux-source-2.6.22.orig/drivers/block/Kconfig linux-source-2.6.22/drivers/block/Kconfig --- linux-source-2.6.22.orig/drivers/block/Kconfig 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/drivers/block/Kconfig 2007-10-24 16:14:06.291582822 +0200 @@ -311,6 +311,13 @@ config BLK_DEV_CRYPTOLOOP instead, which can be configured to be on-disk compatible with the cryptoloop device. +config BLK_DEV_VROOT + tristate "Virtual Root device support" + depends on QUOTACTL + ---help--- + Saying Y here will allow you to use quota/fs ioctls on a shared + partition within a virtual server without compromising security. + config BLK_DEV_NBD tristate "Network block device support" depends on NET diff -NurpP --minimal linux-source-2.6.22.orig/drivers/block/loop.c linux-source-2.6.22/drivers/block/loop.c --- linux-source-2.6.22.orig/drivers/block/loop.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/drivers/block/loop.c 2007-10-24 16:14:06.291582822 +0200 @@ -74,6 +74,7 @@ #include #include #include +#include #include @@ -790,6 +791,7 @@ static int loop_set_fd(struct loop_devic lo->lo_blocksize = lo_blocksize; lo->lo_device = bdev; lo->lo_flags = lo_flags; + lo->lo_xid = vx_current_xid(); lo->lo_backing_file = file; lo->transfer = transfer_none; lo->ioctl = NULL; @@ -909,6 +911,7 @@ static int loop_clr_fd(struct loop_devic lo->lo_encrypt_key_size = 0; lo->lo_flags = 0; lo->lo_thread = NULL; + lo->lo_xid = 0; memset(lo->lo_encrypt_key, 0, LO_KEY_SIZE); memset(lo->lo_crypt_name, 0, LO_NAME_SIZE); memset(lo->lo_file_name, 0, LO_NAME_SIZE); @@ -930,7 +933,7 @@ loop_set_status(struct loop_device *lo, struct loop_func_table *xfer; if (lo->lo_encrypt_key_size && lo->lo_key_owner != current->uid && - !capable(CAP_SYS_ADMIN)) + !vx_capable(CAP_SYS_ADMIN, VXC_ADMIN_CLOOP)) return -EPERM; if (lo->lo_state != Lo_bound) return -ENXIO; @@ -1010,7 +1013,8 @@ loop_get_status(struct loop_device *lo, memcpy(info->lo_crypt_name, lo->lo_crypt_name, LO_NAME_SIZE); info->lo_encrypt_type = lo->lo_encryption ? lo->lo_encryption->number : 0; - if (lo->lo_encrypt_key_size && capable(CAP_SYS_ADMIN)) { + if (lo->lo_encrypt_key_size && + vx_capable(CAP_SYS_ADMIN, VXC_ADMIN_CLOOP)) { info->lo_encrypt_key_size = lo->lo_encrypt_key_size; memcpy(info->lo_encrypt_key, lo->lo_encrypt_key, lo->lo_encrypt_key_size); @@ -1321,6 +1325,9 @@ static int lo_open(struct inode *inode, { struct loop_device *lo = inode->i_bdev->bd_disk->private_data; + if (!vx_check(lo->lo_xid, VS_IDENT|VS_HOSTID)) + return -EACCES; + mutex_lock(&lo->lo_ctl_mutex); lo->lo_refcnt++; mutex_unlock(&lo->lo_ctl_mutex); diff -NurpP --minimal linux-source-2.6.22.orig/drivers/block/Makefile linux-source-2.6.22/drivers/block/Makefile --- linux-source-2.6.22.orig/drivers/block/Makefile 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/drivers/block/Makefile 2007-10-24 16:14:06.291582822 +0200 @@ -29,4 +29,5 @@ obj-$(CONFIG_BLK_DEV_CRYPTOLOOP) += cryp obj-$(CONFIG_VIODASD) += viodasd.o obj-$(CONFIG_BLK_DEV_SX8) += sx8.o obj-$(CONFIG_BLK_DEV_UB) += ub.o +obj-$(CONFIG_BLK_DEV_VROOT) += vroot.o diff -NurpP --minimal linux-source-2.6.22.orig/drivers/block/vroot.c linux-source-2.6.22/drivers/block/vroot.c --- linux-source-2.6.22.orig/drivers/block/vroot.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/drivers/block/vroot.c 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,281 @@ +/* + * linux/drivers/block/vroot.c + * + * written by Herbert Pötzl, 9/11/2002 + * ported to 2.6.10 by Herbert Pötzl, 30/12/2004 + * + * based on the loop.c code by Theodore Ts'o. + * + * Copyright (C) 2002-2007 by Herbert Pötzl. + * Redistribution of this file is permitted under the + * GNU General Public License. + * + */ + +#include +#include +#include +#include +#include + +#include +#include + + +static int max_vroot = 8; + +static struct vroot_device *vroot_dev; +static struct gendisk **disks; + + +static int vroot_set_dev( + struct vroot_device *vr, + struct file *vr_file, + struct block_device *bdev, + unsigned int arg) +{ + struct block_device *real_bdev; + struct file *file; + struct inode *inode; + int error; + + error = -EBUSY; + if (vr->vr_state != Vr_unbound) + goto out; + + error = -EBADF; + file = fget(arg); + if (!file) + goto out; + + error = -EINVAL; + inode = file->f_dentry->d_inode; + + + if (S_ISBLK(inode->i_mode)) { + real_bdev = inode->i_bdev; + vr->vr_device = real_bdev; + __iget(real_bdev->bd_inode); + } else + goto out_fput; + + vxdprintk(VXD_CBIT(misc, 0), + "vroot[%d]_set_dev: dev=" VXF_DEV, + vr->vr_number, VXD_DEV(real_bdev)); + + vr->vr_state = Vr_bound; + error = 0; + + out_fput: + fput(file); + out: + return error; +} + +static int vroot_clr_dev( + struct vroot_device *vr, + struct file *vr_file, + struct block_device *bdev) +{ + struct block_device *real_bdev; + + if (vr->vr_state != Vr_bound) + return -ENXIO; + if (vr->vr_refcnt > 1) /* we needed one fd for the ioctl */ + return -EBUSY; + + real_bdev = vr->vr_device; + + vxdprintk(VXD_CBIT(misc, 0), + "vroot[%d]_clr_dev: dev=" VXF_DEV, + vr->vr_number, VXD_DEV(real_bdev)); + + bdput(real_bdev); + vr->vr_state = Vr_unbound; + vr->vr_device = NULL; + return 0; +} + + +static int vr_ioctl(struct inode *inode, struct file *file, + unsigned int cmd, unsigned long arg) +{ + struct vroot_device *vr = inode->i_bdev->bd_disk->private_data; + int err; + + down(&vr->vr_ctl_mutex); + switch (cmd) { + case VROOT_SET_DEV: + err = vroot_set_dev(vr, file, inode->i_bdev, arg); + break; + case VROOT_CLR_DEV: + err = vroot_clr_dev(vr, file, inode->i_bdev); + break; + default: + err = -EINVAL; + break; + } + up(&vr->vr_ctl_mutex); + return err; +} + +static int vr_open(struct inode *inode, struct file *file) +{ + struct vroot_device *vr = inode->i_bdev->bd_disk->private_data; + + down(&vr->vr_ctl_mutex); + vr->vr_refcnt++; + up(&vr->vr_ctl_mutex); + return 0; +} + +static int vr_release(struct inode *inode, struct file *file) +{ + struct vroot_device *vr = inode->i_bdev->bd_disk->private_data; + + down(&vr->vr_ctl_mutex); + --vr->vr_refcnt; + up(&vr->vr_ctl_mutex); + return 0; +} + +static struct block_device_operations vr_fops = { + .owner = THIS_MODULE, + .open = vr_open, + .release = vr_release, + .ioctl = vr_ioctl, +}; + +struct block_device *__vroot_get_real_bdev(struct block_device *bdev) +{ + struct inode *inode = bdev->bd_inode; + struct vroot_device *vr; + struct block_device *real_bdev; + int minor = iminor(inode); + + vr = &vroot_dev[minor]; + real_bdev = vr->vr_device; + + vxdprintk(VXD_CBIT(misc, 0), + "vroot[%d]_get_real_bdev: dev=" VXF_DEV, + vr->vr_number, VXD_DEV(real_bdev)); + + if (vr->vr_state != Vr_bound) + return ERR_PTR(-ENXIO); + + __iget(real_bdev->bd_inode); + return real_bdev; +} + +/* + * And now the modules code and kernel interface. + */ + +module_param(max_vroot, int, 0); + +MODULE_PARM_DESC(max_vroot, "Maximum number of vroot devices (1-256)"); +MODULE_LICENSE("GPL"); +MODULE_ALIAS_BLOCKDEV_MAJOR(VROOT_MAJOR); + +MODULE_AUTHOR ("Herbert Pötzl"); +MODULE_DESCRIPTION ("Virtual Root Device Mapper"); + + +int __init vroot_init(void) +{ + int err, i; + + if (max_vroot < 1 || max_vroot > 256) { + max_vroot = MAX_VROOT_DEFAULT; + printk(KERN_WARNING "vroot: invalid max_vroot " + "(must be between 1 and 256), " + "using default (%d)\n", max_vroot); + } + + if (register_blkdev(VROOT_MAJOR, "vroot")) + return -EIO; + + err = -ENOMEM; + vroot_dev = kmalloc(max_vroot * sizeof(struct vroot_device), GFP_KERNEL); + if (!vroot_dev) + goto out_mem1; + memset(vroot_dev, 0, max_vroot * sizeof(struct vroot_device)); + + disks = kmalloc(max_vroot * sizeof(struct gendisk *), GFP_KERNEL); + if (!disks) + goto out_mem2; + + for (i = 0; i < max_vroot; i++) { + disks[i] = alloc_disk(1); + if (!disks[i]) + goto out_mem3; + } + + for (i = 0; i < max_vroot; i++) { + struct vroot_device *vr = &vroot_dev[i]; + struct gendisk *disk = disks[i]; + + memset(vr, 0, sizeof(*vr)); + init_MUTEX(&vr->vr_ctl_mutex); + vr->vr_number = i; + disk->major = VROOT_MAJOR; + disk->first_minor = i; + disk->fops = &vr_fops; + sprintf(disk->disk_name, "vroot%d", i); + disk->private_data = vr; + } + + err = register_vroot_grb(&__vroot_get_real_bdev); + if (err) + goto out_mem3; + + for (i = 0; i < max_vroot; i++) + add_disk(disks[i]); + printk(KERN_INFO "vroot: loaded (max %d devices)\n", max_vroot); + return 0; + +out_mem3: + while (i--) + put_disk(disks[i]); + kfree(disks); +out_mem2: + kfree(vroot_dev); +out_mem1: + unregister_blkdev(VROOT_MAJOR, "vroot"); + printk(KERN_ERR "vroot: ran out of memory\n"); + return err; +} + +void vroot_exit(void) +{ + int i; + + if (unregister_vroot_grb(&__vroot_get_real_bdev)) + printk(KERN_WARNING "vroot: cannot unregister grb\n"); + + for (i = 0; i < max_vroot; i++) { + del_gendisk(disks[i]); + put_disk(disks[i]); + } + if (unregister_blkdev(VROOT_MAJOR, "vroot")) + printk(KERN_WARNING "vroot: cannot unregister blkdev\n"); + + kfree(disks); + kfree(vroot_dev); +} + +module_init(vroot_init); +module_exit(vroot_exit); + +#ifndef MODULE + +static int __init max_vroot_setup(char *str) +{ + max_vroot = simple_strtol(str, NULL, 0); + return 1; +} + +__setup("max_vroot=", max_vroot_setup); + +#endif + diff -NurpP --minimal linux-source-2.6.22.orig/drivers/char/sysrq.c linux-source-2.6.22/drivers/char/sysrq.c --- linux-source-2.6.22.orig/drivers/char/sysrq.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/drivers/char/sysrq.c 2007-10-24 16:14:06.291582822 +0200 @@ -36,6 +36,7 @@ #include #include #include +#include #include #include @@ -309,6 +310,21 @@ static struct sysrq_key_op sysrq_unrt_op .enable_mask = SYSRQ_ENABLE_RTNICE, }; + +#ifdef CONFIG_VSERVER_DEBUG +static void sysrq_handle_vxinfo(int key, struct tty_struct *tty) +{ + dump_vx_info_inactive((key == 'x')?0:1); +} + +static struct sysrq_key_op sysrq_showvxinfo_op = { + .handler = sysrq_handle_vxinfo, + .help_msg = "conteXt", + .action_msg = "Show Context Info", + .enable_mask = SYSRQ_ENABLE_DUMP, +}; +#endif + /* Key Operations table and lock */ static DEFINE_SPINLOCK(sysrq_key_table_lock); @@ -357,7 +373,11 @@ static struct sysrq_key_op *sysrq_key_ta /* x: May be registered on ppc/powerpc for xmon */ NULL, /* x */ NULL, /* y */ - NULL /* z */ +#ifdef CONFIG_VSERVER_DEBUG + &sysrq_showvxinfo_op, /* z */ +#else + NULL, /* z */ +#endif }; /* key2index calculation, -1 on invalid index */ @@ -369,6 +389,8 @@ static int sysrq_key_table_key2index(int retval = key - '0'; else if ((key >= 'a') && (key <= 'z')) retval = key + 10 - 'a'; + else if ((key >= 'A') && (key <= 'Z')) + retval = key + 10 - 'A'; else retval = -1; return retval; diff -NurpP --minimal linux-source-2.6.22.orig/drivers/char/tty_io.c linux-source-2.6.22/drivers/char/tty_io.c --- linux-source-2.6.22.orig/drivers/char/tty_io.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/drivers/char/tty_io.c 2007-10-24 16:14:06.291582822 +0200 @@ -103,6 +103,7 @@ #include #include +#include #undef TTY_DEBUG_HANGUP @@ -3049,13 +3050,15 @@ unlock: static int tiocgpgrp(struct tty_struct *tty, struct tty_struct *real_tty, pid_t __user *p) { + pid_t pgrp; /* * (tty == real_tty) is a cheap way of * testing if the tty is NOT a master pty. */ if (tty == real_tty && current->signal->tty != real_tty) return -ENOTTY; - return put_user(pid_nr(real_tty->pgrp), p); + pgrp = vx_map_pid(pid_nr(real_tty->pgrp)); + return put_user(pgrp, p); } /** @@ -3086,6 +3089,7 @@ static int tiocspgrp(struct tty_struct * return -ENOTTY; if (get_user(pgrp_nr, p)) return -EFAULT; + pgrp_nr = vx_rmap_pid(pgrp_nr); if (pgrp_nr < 0) return -EINVAL; rcu_read_lock(); diff -NurpP --minimal linux-source-2.6.22.orig/drivers/infiniband/hw/ipath/ipath_user_pages.c linux-source-2.6.22/drivers/infiniband/hw/ipath/ipath_user_pages.c --- linux-source-2.6.22.orig/drivers/infiniband/hw/ipath/ipath_user_pages.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/drivers/infiniband/hw/ipath/ipath_user_pages.c 2007-10-24 16:14:06.291582822 +0200 @@ -33,6 +33,7 @@ #include #include +#include #include "ipath_kernel.h" @@ -61,7 +62,8 @@ static int __get_user_pages(unsigned lon lock_limit = current->signal->rlim[RLIMIT_MEMLOCK].rlim_cur >> PAGE_SHIFT; - if (num_pages > lock_limit) { + if (num_pages > lock_limit || + !vx_vmlocked_avail(current->mm, num_pages)) { ret = -ENOMEM; goto bail; } @@ -78,7 +80,7 @@ static int __get_user_pages(unsigned lon goto bail_release; } - current->mm->locked_vm += num_pages; + vx_vmlocked_add(current->mm, num_pages); ret = 0; goto bail; @@ -203,7 +205,7 @@ void ipath_release_user_pages(struct pag __ipath_release_user_pages(p, num_pages, 1); - current->mm->locked_vm -= num_pages; + vx_vmlocked_sub(current->mm, num_pages); up_write(¤t->mm->mmap_sem); } @@ -220,7 +222,7 @@ static void user_pages_account(struct wo container_of(_work, struct ipath_user_pages_work, work); down_write(&work->mm->mmap_sem); - work->mm->locked_vm -= work->num_pages; + vx_vmlocked_sub(work->mm, work->num_pages); up_write(&work->mm->mmap_sem); mmput(work->mm); kfree(work); diff -NurpP --minimal linux-source-2.6.22.orig/drivers/md/dm.c linux-source-2.6.22/drivers/md/dm.c --- linux-source-2.6.22.orig/drivers/md/dm.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/drivers/md/dm.c 2007-10-24 16:14:06.291582822 +0200 @@ -21,6 +21,7 @@ #include #include #include +#include #define DM_MSG_PREFIX "core" @@ -77,6 +78,7 @@ struct mapped_device { rwlock_t map_lock; atomic_t holders; atomic_t open_count; + xid_t xid; unsigned long flags; @@ -223,6 +225,7 @@ static void __exit dm_exit(void) static int dm_blk_open(struct inode *inode, struct file *file) { struct mapped_device *md; + int ret = -ENXIO; spin_lock(&_minor_lock); @@ -231,18 +234,19 @@ static int dm_blk_open(struct inode *ino goto out; if (test_bit(DMF_FREEING, &md->flags) || - test_bit(DMF_DELETING, &md->flags)) { - md = NULL; + test_bit(DMF_DELETING, &md->flags)) + goto out; + + ret = -EACCES; + if (!vx_check(md->xid, VS_IDENT|VS_HOSTID)) goto out; - } dm_get(md); atomic_inc(&md->open_count); - + ret = 0; out: spin_unlock(&_minor_lock); - - return md ? 0 : -ENXIO; + return ret; } static int dm_blk_close(struct inode *inode, struct file *file) @@ -438,6 +442,14 @@ int dm_set_geometry(struct mapped_device return 0; } +/* + * Get the xid associated with a dm device + */ +xid_t dm_get_xid(struct mapped_device *md) +{ + return md->xid; +} + /*----------------------------------------------------------------- * CRUD START: * A more elegant soln is in the works that uses the queue @@ -1000,6 +1012,7 @@ static struct mapped_device *alloc_dev(i atomic_set(&md->holders, 1); atomic_set(&md->open_count, 0); atomic_set(&md->event_nr, 0); + md->xid = vx_current_xid(); md->queue = blk_alloc_queue(GFP_KERNEL); if (!md->queue) diff -NurpP --minimal linux-source-2.6.22.orig/drivers/md/dm.h linux-source-2.6.22/drivers/md/dm.h --- linux-source-2.6.22.orig/drivers/md/dm.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/drivers/md/dm.h 2007-10-24 16:14:06.291582822 +0200 @@ -91,6 +91,8 @@ void dm_put_target_type(struct target_ty int dm_target_iterate(void (*iter_func)(struct target_type *tt, void *param), void *param); +xid_t dm_get_xid(struct mapped_device *md); + /*----------------------------------------------------------------- * Useful inlines. *---------------------------------------------------------------*/ diff -NurpP --minimal linux-source-2.6.22.orig/drivers/md/dm-ioctl.c linux-source-2.6.22/drivers/md/dm-ioctl.c --- linux-source-2.6.22.orig/drivers/md/dm-ioctl.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/drivers/md/dm-ioctl.c 2007-10-24 16:14:06.291582822 +0200 @@ -15,6 +15,7 @@ #include #include #include +#include #include @@ -100,7 +101,8 @@ static struct hash_cell *__get_name_cell unsigned int h = hash_str(str); list_for_each_entry (hc, _name_buckets + h, name_list) - if (!strcmp(hc->name, str)) { + if (vx_check(dm_get_xid(hc->md), VS_WATCH_P | VS_IDENT) && + !strcmp(hc->name, str)) { dm_get(hc->md); return hc; } @@ -114,7 +116,8 @@ static struct hash_cell *__get_uuid_cell unsigned int h = hash_str(str); list_for_each_entry (hc, _uuid_buckets + h, uuid_list) - if (!strcmp(hc->uuid, str)) { + if (vx_check(dm_get_xid(hc->md), VS_WATCH_P | VS_IDENT) && + !strcmp(hc->uuid, str)) { dm_get(hc->md); return hc; } @@ -349,6 +352,9 @@ typedef int (*ioctl_fn)(struct dm_ioctl static int remove_all(struct dm_ioctl *param, size_t param_size) { + if (!vx_check(0, VS_ADMIN)) + return -EPERM; + dm_hash_remove_all(1); param->data_size = 0; return 0; @@ -396,6 +402,8 @@ static int list_devices(struct dm_ioctl */ for (i = 0; i < NUM_BUCKETS; i++) { list_for_each_entry (hc, _name_buckets + i, name_list) { + if (!vx_check(dm_get_xid(hc->md), VS_WATCH_P | VS_IDENT)) + continue; needed += sizeof(struct dm_name_list); needed += strlen(hc->name) + 1; needed += ALIGN_MASK; @@ -419,6 +427,8 @@ static int list_devices(struct dm_ioctl */ for (i = 0; i < NUM_BUCKETS; i++) { list_for_each_entry (hc, _name_buckets + i, name_list) { + if (!vx_check(dm_get_xid(hc->md), VS_WATCH_P | VS_IDENT)) + continue; if (old_nl) old_nl->next = (uint32_t) ((void *) nl - (void *) old_nl); @@ -609,10 +619,11 @@ static struct hash_cell *__find_device_h if (!md) goto out; - mdptr = dm_get_mdptr(md); + if (vx_check(dm_get_xid(md), VS_WATCH_P | VS_IDENT)) + mdptr = dm_get_mdptr(md); + if (!mdptr) dm_put(md); - out: return mdptr; } @@ -1409,8 +1420,8 @@ static int ctl_ioctl(struct inode *inode ioctl_fn fn = NULL; size_t param_size; - /* only root can play with this */ - if (!capable(CAP_SYS_ADMIN)) + /* only root and certain contexts can play with this */ + if (!vx_capable(CAP_SYS_ADMIN, VXC_ADMIN_MAPPER)) return -EACCES; if (_IOC_TYPE(command) != DM_IOCTL) diff -NurpP --minimal linux-source-2.6.22.orig/fs/attr.c linux-source-2.6.22/fs/attr.c --- linux-source-2.6.22.orig/fs/attr.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/fs/attr.c 2007-10-24 16:14:06.291582822 +0200 @@ -14,6 +14,9 @@ #include #include #include +#include +#include +#include /* Taken over from the old code... */ @@ -55,6 +58,30 @@ int inode_change_ok(struct inode *inode, if (current->fsuid != inode->i_uid && !capable(CAP_FOWNER)) goto error; } + + /* Check for evil vserver activity */ + if (vx_check(0, VS_ADMIN)) + goto fine; + + if (IS_BARRIER(inode)) { + vxwprintk(1, "xid=%d messing with the barrier.", + vx_current_xid()); + goto error; + } + switch (inode->i_sb->s_magic) { + case PROC_SUPER_MAGIC: + /* maybe allow that in the future? */ + vxwprintk(1, "xid=%d messing with the procfs.", + vx_current_xid()); + goto error; + case DEVPTS_SUPER_MAGIC: + /* devpts is xid tagged */ + if (vx_check((xid_t)inode->i_tag, VS_IDENT)) + goto fine; + vxwprintk(1, "xid=%d messing with the devpts.", + vx_current_xid()); + goto error; + } fine: retval = 0; error: @@ -78,6 +105,8 @@ int inode_setattr(struct inode * inode, inode->i_uid = attr->ia_uid; if (ia_valid & ATTR_GID) inode->i_gid = attr->ia_gid; + if ((ia_valid & ATTR_TAG) && IS_TAGGED(inode)) + inode->i_tag = attr->ia_tag; if (ia_valid & ATTR_ATIME) inode->i_atime = timespec_trunc(attr->ia_atime, inode->i_sb->s_time_gran); @@ -153,7 +182,8 @@ int notify_change(struct dentry *dentry, error = security_inode_setattr(dentry, mnt, attr); if (!error) { if ((ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) || - (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) + (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid) || + (ia_valid & ATTR_TAG && attr->ia_tag != inode->i_tag)) error = DQUOT_TRANSFER(inode, attr) ? -EDQUOT : 0; if (!error) error = inode_setattr(inode, attr); diff -NurpP --minimal linux-source-2.6.22.orig/fs/binfmt_aout.c linux-source-2.6.22/fs/binfmt_aout.c --- linux-source-2.6.22.orig/fs/binfmt_aout.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/binfmt_aout.c 2007-10-24 16:14:06.291582822 +0200 @@ -24,6 +24,7 @@ #include #include #include +#include #include #include diff -NurpP --minimal linux-source-2.6.22.orig/fs/binfmt_elf.c linux-source-2.6.22/fs/binfmt_elf.c --- linux-source-2.6.22.orig/fs/binfmt_elf.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/binfmt_elf.c 2007-10-24 16:14:06.291582822 +0200 @@ -39,6 +39,7 @@ #include #include #include +#include #include #include #include diff -NurpP --minimal linux-source-2.6.22.orig/fs/binfmt_flat.c linux-source-2.6.22/fs/binfmt_flat.c --- linux-source-2.6.22.orig/fs/binfmt_flat.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/binfmt_flat.c 2007-10-24 16:14:06.291582822 +0200 @@ -36,6 +36,7 @@ #include #include #include +#include #include #include diff -NurpP --minimal linux-source-2.6.22.orig/fs/binfmt_som.c linux-source-2.6.22/fs/binfmt_som.c --- linux-source-2.6.22.orig/fs/binfmt_som.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/binfmt_som.c 2007-10-24 16:14:06.291582822 +0200 @@ -28,6 +28,7 @@ #include #include #include +#include #include #include diff -NurpP --minimal linux-source-2.6.22.orig/fs/dcache.c linux-source-2.6.22/fs/dcache.c --- linux-source-2.6.22.orig/fs/dcache.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/fs/dcache.c 2007-10-24 16:14:06.291582822 +0200 @@ -31,6 +31,7 @@ #include #include #include +#include #include "internal.h" @@ -176,6 +177,7 @@ void dput(struct dentry *dentry) if (!dentry) return; + vx_dentry_dec(dentry); repeat: if (atomic_read(&dentry->d_count) == 1) might_sleep(); @@ -189,6 +191,8 @@ repeat: return; } + vx_dentry_dec(dentry); + /* * AV: ->d_delete() is _NOT_ allowed to block now. */ @@ -288,6 +292,7 @@ static inline struct dentry * __dget_loc if (!list_empty(&dentry->d_lru)) { dentry_stat.nr_unused--; list_del_init(&dentry->d_lru); + vx_dentry_inc(dentry); } return dentry; } @@ -898,6 +903,9 @@ struct dentry *d_alloc(struct dentry * p struct dentry *dentry; char *dname; + if (!vx_dentry_avail(1)) + return NULL; + dentry = kmem_cache_alloc(dentry_cache, GFP_KERNEL); if (!dentry) return NULL; @@ -946,6 +954,7 @@ struct dentry *d_alloc(struct dentry * p if (parent) list_add(&dentry->d_u.d_child, &parent->d_subdirs); dentry_stat.nr_dentry++; + vx_dentry_inc(dentry); spin_unlock(&dcache_lock); return dentry; @@ -1295,6 +1304,7 @@ struct dentry * __d_lookup(struct dentry if (!d_unhashed(dentry)) { atomic_inc(&dentry->d_count); + vx_dentry_inc(dentry); found = dentry; } spin_unlock(&dentry->d_lock); diff -NurpP --minimal linux-source-2.6.22.orig/fs/devpts/inode.c linux-source-2.6.22/fs/devpts/inode.c --- linux-source-2.6.22.orig/fs/devpts/inode.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/devpts/inode.c 2007-10-24 16:14:06.291582822 +0200 @@ -20,8 +20,22 @@ #include #include #include +#include -#define DEVPTS_SUPER_MAGIC 0x1cd1 + +static int devpts_permission(struct inode *inode, int mask, struct nameidata *nd) +{ + int ret = -EACCES; + + /* devpts is xid tagged */ + if (vx_check((xid_t)inode->i_tag, VS_WATCH_P | VS_IDENT)) + ret = generic_permission(inode, mask, NULL); + return ret; +} + +static struct inode_operations devpts_file_inode_operations = { + .permission = devpts_permission, +}; static struct vfsmount *devpts_mnt; static struct dentry *devpts_root; @@ -92,6 +106,25 @@ static int devpts_remount(struct super_b return 0; } +static int devpts_filter(struct dentry *de) +{ + /* devpts is xid tagged */ + return vx_check((xid_t)de->d_inode->i_tag, VS_WATCH_P | VS_IDENT); +} + +static int devpts_readdir(struct file * filp, void * dirent, filldir_t filldir) +{ + return dcache_readdir_filter(filp, dirent, filldir, devpts_filter); +} + +static struct file_operations devpts_dir_operations = { + .open = dcache_dir_open, + .release = dcache_dir_close, + .llseek = dcache_dir_lseek, + .read = generic_read_dir, + .readdir = devpts_readdir, +}; + static const struct super_operations devpts_sops = { .statfs = simple_statfs, .remount_fs = devpts_remount, @@ -117,8 +150,10 @@ devpts_fill_super(struct super_block *s, inode->i_uid = inode->i_gid = 0; inode->i_mode = S_IFDIR | S_IRUGO | S_IXUGO | S_IWUSR; inode->i_op = &simple_dir_inode_operations; - inode->i_fop = &simple_dir_operations; + inode->i_fop = &devpts_dir_operations; inode->i_nlink = 2; + /* devpts is xid tagged */ + inode->i_tag = (tag_t)vx_current_xid(); devpts_root = s->s_root = d_alloc_root(inode); if (s->s_root) @@ -176,6 +211,9 @@ int devpts_pty_new(struct tty_struct *tt inode->i_gid = config.setgid ? config.gid : current->fsgid; inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME; init_special_inode(inode, S_IFCHR|config.mode, device); + /* devpts is xid tagged */ + inode->i_tag = (tag_t)vx_current_xid(); + inode->i_op = &devpts_file_inode_operations; inode->i_private = tty; dentry = get_node(number); diff -NurpP --minimal linux-source-2.6.22.orig/fs/ecryptfs/inode.c linux-source-2.6.22/fs/ecryptfs/inode.c --- linux-source-2.6.22.orig/fs/ecryptfs/inode.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/fs/ecryptfs/inode.c 2007-10-24 16:14:06.291582822 +0200 @@ -434,7 +434,7 @@ static int ecryptfs_link(struct dentry * lower_dir_dentry = lock_parent(lower_new_dentry); rc = vfs_link(lower_old_dentry, lower_old_mnt, lower_dir_dentry->d_inode, lower_new_dentry, - lower_new_mnt); + lower_new_mnt, NULL); if (rc || !lower_new_dentry->d_inode) goto out_lock; rc = ecryptfs_interpose(lower_new_dentry, new_dentry, dir->i_sb, 0); @@ -463,7 +463,7 @@ static int ecryptfs_unlink(struct inode struct inode *lower_dir_inode = ecryptfs_inode_to_lower(dir); lock_parent(lower_dentry); - rc = vfs_unlink(lower_dir_inode, lower_dentry, lower_mnt); + rc = vfs_unlink(lower_dir_inode, lower_dentry, lower_mnt, NULL); if (rc) { printk(KERN_ERR "Error in vfs_unlink; rc = [%d]\n", rc); goto out_unlock; @@ -502,7 +502,7 @@ static int ecryptfs_symlink(struct inode goto out_lock; } rc = vfs_symlink(lower_dir_dentry->d_inode, lower_dentry, lower_mnt, - encoded_symname, mode); + encoded_symname, mode, NULL); kfree(encoded_symname); if (rc || !lower_dentry->d_inode) goto out_lock; @@ -530,7 +530,7 @@ static int ecryptfs_mkdir(struct inode * lower_mnt = ecryptfs_dentry_to_lower_mnt(dentry); lower_dir_dentry = lock_parent(lower_dentry); rc = vfs_mkdir(lower_dir_dentry->d_inode, lower_dentry, lower_mnt, - mode); + mode, NULL); if (rc || !lower_dentry->d_inode) goto out; rc = ecryptfs_interpose(lower_dentry, dentry, dir->i_sb, 0); @@ -558,7 +558,7 @@ static int ecryptfs_rmdir(struct inode * dget(dentry); lower_dir_dentry = lock_parent(lower_dentry); dget(lower_dentry); - rc = vfs_rmdir(lower_dir_dentry->d_inode, lower_dentry, lower_mnt); + rc = vfs_rmdir(lower_dir_dentry->d_inode, lower_dentry, lower_mnt, NULL); dput(lower_dentry); if (!rc) d_delete(lower_dentry); @@ -583,7 +583,7 @@ ecryptfs_mknod(struct inode *dir, struct lower_mnt = ecryptfs_dentry_to_lower_mnt(dentry); lower_dir_dentry = lock_parent(lower_dentry); rc = vfs_mknod(lower_dir_dentry->d_inode, lower_dentry, lower_mnt, mode, - dev); + dev, NULL); if (rc || !lower_dentry->d_inode) goto out; rc = ecryptfs_interpose(lower_dentry, dentry, dir->i_sb, 0); diff -NurpP --minimal linux-source-2.6.22.orig/fs/exec.c linux-source-2.6.22/fs/exec.c --- linux-source-2.6.22.orig/fs/exec.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/fs/exec.c 2007-10-24 16:14:06.291582822 +0200 @@ -51,6 +51,7 @@ #include #include #include +#include #include #include @@ -440,7 +441,8 @@ int setup_arg_pages(struct linux_binprm kmem_cache_free(vm_area_cachep, mpnt); return ret; } - mm->stack_vm = mm->total_vm = vma_pages(mpnt); + vx_vmpages_sub(mm, mm->total_vm - vma_pages(mpnt)); + mm->stack_vm = mm->total_vm; } for (i = 0 ; i < MAX_ARG_PAGES ; i++) { @@ -1336,7 +1338,7 @@ static int format_corename(char *corenam /* UNIX time of coredump */ case 't': { struct timeval tv; - do_gettimeofday(&tv); + vx_gettimeofday(&tv); rc = snprintf(out_ptr, out_end - out_ptr, "%lu", tv.tv_sec); if (rc > out_end - out_ptr) diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext2/balloc.c linux-source-2.6.22/fs/ext2/balloc.c --- linux-source-2.6.22.orig/fs/ext2/balloc.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext2/balloc.c 2007-10-24 16:14:06.291582822 +0200 @@ -16,6 +16,8 @@ #include #include #include +#include +#include /* * balloc.c contains the blocks allocation and deallocation routines @@ -102,12 +104,13 @@ static int reserve_blocks(struct super_b { struct ext2_sb_info *sbi = EXT2_SB(sb); struct ext2_super_block *es = sbi->s_es; - unsigned free_blocks; - unsigned root_blocks; + unsigned long long free_blocks, root_blocks; free_blocks = percpu_counter_read_positive(&sbi->s_freeblocks_counter); root_blocks = le32_to_cpu(es->s_r_blocks_count); + DLIMIT_ADJUST_BLOCK(sb, dx_current_tag(), &free_blocks, &root_blocks); + if (free_blocks < count) count = free_blocks; @@ -258,6 +261,7 @@ do_more: } error_return: brelse(bitmap_bh); + DLIMIT_FREE_BLOCK(inode, freed); release_blocks(sb, freed); DQUOT_FREE_BLOCK(inode, freed); } @@ -361,6 +365,10 @@ int ext2_new_block(struct inode *inode, *err = -ENOSPC; goto out_dquot; } + if (DLIMIT_ALLOC_BLOCK(inode, es_alloc)) { + *err = -ENOSPC; + goto out_dlimit; + } ext2_debug ("goal=%lu.\n", goal); @@ -508,6 +516,8 @@ got_block: *err = 0; out_release: group_release_blocks(sb, group_no, desc, gdp_bh, group_alloc); + DLIMIT_FREE_BLOCK(inode, es_alloc); +out_dlimit: release_blocks(sb, es_alloc); out_dquot: DQUOT_FREE_BLOCK(inode, dq_alloc); diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext2/ext2.h linux-source-2.6.22/fs/ext2/ext2.h --- linux-source-2.6.22.orig/fs/ext2/ext2.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext2/ext2.h 2007-10-24 16:14:06.291582822 +0200 @@ -167,6 +167,7 @@ extern const struct file_operations ext2 extern const struct address_space_operations ext2_aops; extern const struct address_space_operations ext2_aops_xip; extern const struct address_space_operations ext2_nobh_aops; +extern int ext2_sync_flags(struct inode *inode); /* namei.c */ extern const struct inode_operations ext2_dir_inode_operations; diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext2/file.c linux-source-2.6.22/fs/ext2/file.c --- linux-source-2.6.22.orig/fs/ext2/file.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext2/file.c 2007-10-24 16:14:06.291582822 +0200 @@ -54,6 +54,7 @@ const struct file_operations ext2_file_o .release = ext2_release_file, .fsync = ext2_sync_file, .sendfile = generic_file_sendfile, + .sendpage = generic_file_sendpage, .splice_read = generic_file_splice_read, .splice_write = generic_file_splice_write, }; @@ -85,4 +86,5 @@ const struct inode_operations ext2_file_ #endif .setattr = ext2_setattr, .permission = ext2_permission, + .sync_flags = ext2_sync_flags, }; diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext2/ialloc.c linux-source-2.6.22/fs/ext2/ialloc.c --- linux-source-2.6.22.orig/fs/ext2/ialloc.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext2/ialloc.c 2007-10-24 16:14:06.291582822 +0200 @@ -17,6 +17,8 @@ #include #include #include +#include +#include #include "ext2.h" #include "xattr.h" #include "acl.h" @@ -125,6 +127,7 @@ void ext2_free_inode (struct inode * ino ext2_xattr_delete_inode(inode); DQUOT_FREE_INODE(inode); DQUOT_DROP(inode); + DLIMIT_FREE_INODE(inode); } es = EXT2_SB(sb)->s_es; @@ -464,6 +467,11 @@ struct inode *ext2_new_inode(struct inod if (!inode) return ERR_PTR(-ENOMEM); + inode->i_tag = dx_current_fstag(sb); + if (DLIMIT_ALLOC_INODE(inode)) { + err = -ENOSPC; + goto fail_dlim; + } ei = EXT2_I(inode); sbi = EXT2_SB(sb); es = sbi->s_es; @@ -577,7 +585,8 @@ got: inode->i_blocks = 0; inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME_SEC; memset(ei->i_data, 0, sizeof(ei->i_data)); - ei->i_flags = EXT2_I(dir)->i_flags & ~EXT2_BTREE_FL; + ei->i_flags = EXT2_I(dir)->i_flags & + ~(EXT2_BTREE_FL|EXT2_IUNLINK_FL|EXT2_BARRIER_FL); if (S_ISLNK(mode)) ei->i_flags &= ~(EXT2_IMMUTABLE_FL|EXT2_APPEND_FL); /* dirsync is only applied to directories */ @@ -625,12 +634,15 @@ fail_free_drop: fail_drop: DQUOT_DROP(inode); + DLIMIT_FREE_INODE(inode); inode->i_flags |= S_NOQUOTA; inode->i_nlink = 0; iput(inode); return ERR_PTR(err); fail: + DLIMIT_FREE_INODE(inode); +fail_dlim: make_bad_inode(inode); iput(inode); return ERR_PTR(err); diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext2/inode.c linux-source-2.6.22/fs/ext2/inode.c --- linux-source-2.6.22.orig/fs/ext2/inode.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext2/inode.c 2007-10-24 16:14:06.291582822 +0200 @@ -31,6 +31,7 @@ #include #include #include +#include #include "ext2.h" #include "acl.h" #include "xip.h" @@ -913,7 +914,7 @@ void ext2_truncate (struct inode * inode return; if (ext2_inode_is_fast_symlink(inode)) return; - if (IS_APPEND(inode) || IS_IMMUTABLE(inode)) + if (IS_APPEND(inode) || IS_IXORUNLINK(inode)) return; ext2_discard_prealloc(inode); @@ -1042,13 +1043,20 @@ void ext2_set_inode_flags(struct inode * { unsigned int flags = EXT2_I(inode)->i_flags; - inode->i_flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC); + inode->i_flags &= ~(S_IMMUTABLE | S_IUNLINK | S_BARRIER | + S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC); + + if (flags & EXT2_IMMUTABLE_FL) + inode->i_flags |= S_IMMUTABLE; + if (flags & EXT2_IUNLINK_FL) + inode->i_flags |= S_IUNLINK; + if (flags & EXT2_BARRIER_FL) + inode->i_flags |= S_BARRIER; + if (flags & EXT2_SYNC_FL) inode->i_flags |= S_SYNC; if (flags & EXT2_APPEND_FL) inode->i_flags |= S_APPEND; - if (flags & EXT2_IMMUTABLE_FL) - inode->i_flags |= S_IMMUTABLE; if (flags & EXT2_NOATIME_FL) inode->i_flags |= S_NOATIME; if (flags & EXT2_DIRSYNC_FL) @@ -1074,12 +1082,37 @@ void ext2_get_inode_flags(struct ext2_in ei->i_flags |= EXT2_DIRSYNC_FL; } +int ext2_sync_flags(struct inode *inode) +{ + unsigned int oldflags, newflags; + + oldflags = EXT2_I(inode)->i_flags; + newflags = oldflags & ~(EXT2_IMMUTABLE_FL | + EXT2_IUNLINK_FL | EXT2_BARRIER_FL); + + if (IS_IMMUTABLE(inode)) + newflags |= EXT2_IMMUTABLE_FL; + if (IS_IUNLINK(inode)) + newflags |= EXT2_IUNLINK_FL; + if (IS_BARRIER(inode)) + newflags |= EXT2_BARRIER_FL; + + if (oldflags ^ newflags) { + EXT2_I(inode)->i_flags = newflags; + inode->i_ctime = CURRENT_TIME; + mark_inode_dirty(inode); + } + return 0; +} + void ext2_read_inode (struct inode * inode) { struct ext2_inode_info *ei = EXT2_I(inode); ino_t ino = inode->i_ino; struct buffer_head * bh; struct ext2_inode * raw_inode = ext2_get_inode(inode->i_sb, ino, &bh); + uid_t uid; + gid_t gid; int n; #ifdef CONFIG_EXT2_FS_POSIX_ACL @@ -1090,12 +1123,17 @@ void ext2_read_inode (struct inode * ino goto bad_inode; inode->i_mode = le16_to_cpu(raw_inode->i_mode); - inode->i_uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low); - inode->i_gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low); + uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low); + gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low); if (!(test_opt (inode->i_sb, NO_UID32))) { - inode->i_uid |= le16_to_cpu(raw_inode->i_uid_high) << 16; - inode->i_gid |= le16_to_cpu(raw_inode->i_gid_high) << 16; + uid |= le16_to_cpu(raw_inode->i_uid_high) << 16; + gid |= le16_to_cpu(raw_inode->i_gid_high) << 16; } + inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid); + inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid); + inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid, + le16_to_cpu(raw_inode->i_raw_tag)); + inode->i_nlink = le16_to_cpu(raw_inode->i_links_count); inode->i_size = le32_to_cpu(raw_inode->i_size); inode->i_atime.tv_sec = (signed)le32_to_cpu(raw_inode->i_atime); @@ -1192,8 +1230,8 @@ static int ext2_update_inode(struct inod struct ext2_inode_info *ei = EXT2_I(inode); struct super_block *sb = inode->i_sb; ino_t ino = inode->i_ino; - uid_t uid = inode->i_uid; - gid_t gid = inode->i_gid; + uid_t uid = TAGINO_UID(DX_TAG(inode), inode->i_uid, inode->i_tag); + gid_t gid = TAGINO_GID(DX_TAG(inode), inode->i_gid, inode->i_tag); struct buffer_head * bh; struct ext2_inode * raw_inode = ext2_get_inode(sb, ino, &bh); int n; @@ -1229,6 +1267,9 @@ static int ext2_update_inode(struct inod raw_inode->i_uid_high = 0; raw_inode->i_gid_high = 0; } +#ifdef CONFIG_TAGGING_INTERN + raw_inode->i_raw_tag = cpu_to_le16(inode->i_tag); +#endif raw_inode->i_links_count = cpu_to_le16(inode->i_nlink); raw_inode->i_size = cpu_to_le32(inode->i_size); raw_inode->i_atime = cpu_to_le32(inode->i_atime.tv_sec); @@ -1315,7 +1356,8 @@ int ext2_setattr(struct dentry *dentry, if (error) return error; if ((iattr->ia_valid & ATTR_UID && iattr->ia_uid != inode->i_uid) || - (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid)) { + (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid) || + (iattr->ia_valid & ATTR_TAG && iattr->ia_tag != inode->i_tag)) { error = DQUOT_TRANSFER(inode, iattr) ? -EDQUOT : 0; if (error) return error; diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext2/ioctl.c linux-source-2.6.22/fs/ext2/ioctl.c --- linux-source-2.6.22.orig/fs/ext2/ioctl.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext2/ioctl.c 2007-10-24 16:14:06.291582822 +0200 @@ -13,6 +13,7 @@ #include #include #include +#include #include #include @@ -33,7 +34,8 @@ int ext2_ioctl (struct inode * inode, st case EXT2_IOC_SETFLAGS: { unsigned int oldflags; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER)) @@ -54,7 +56,9 @@ int ext2_ioctl (struct inode * inode, st * * This test looks nicer. Thanks to Pauline Middelink */ - if ((flags ^ oldflags) & (EXT2_APPEND_FL | EXT2_IMMUTABLE_FL)) { + if ((oldflags & EXT2_IMMUTABLE_FL) || + ((flags ^ oldflags) & (EXT2_APPEND_FL | + EXT2_IMMUTABLE_FL | EXT2_IUNLINK_FL))) { if (!capable(CAP_LINUX_IMMUTABLE)) { mutex_unlock(&inode->i_mutex); return -EPERM; @@ -76,7 +80,8 @@ int ext2_ioctl (struct inode * inode, st case EXT2_IOC_SETVERSION: if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER)) return -EPERM; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if (get_user(inode->i_generation, (int __user *) arg)) return -EFAULT; diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext2/namei.c linux-source-2.6.22/fs/ext2/namei.c --- linux-source-2.6.22.orig/fs/ext2/namei.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext2/namei.c 2007-10-24 16:14:06.291582822 +0200 @@ -31,6 +31,7 @@ */ #include +#include #include "ext2.h" #include "xattr.h" #include "acl.h" @@ -66,6 +67,7 @@ static struct dentry *ext2_lookup(struct inode = iget(dir->i_sb, ino); if (!inode) return ERR_PTR(-EACCES); + dx_propagate_tag(nd, inode); } return d_splice_alias(inode, dentry); } @@ -391,6 +393,7 @@ const struct inode_operations ext2_dir_i #endif .setattr = ext2_setattr, .permission = ext2_permission, + .sync_flags = ext2_sync_flags, }; const struct inode_operations ext2_special_inode_operations = { @@ -402,4 +405,5 @@ const struct inode_operations ext2_speci #endif .setattr = ext2_setattr, .permission = ext2_permission, + .sync_flags = ext2_sync_flags, }; diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext2/super.c linux-source-2.6.22/fs/ext2/super.c --- linux-source-2.6.22.orig/fs/ext2/super.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext2/super.c 2007-10-24 16:14:06.291582822 +0200 @@ -321,7 +321,7 @@ enum { Opt_err_ro, Opt_nouid32, Opt_nocheck, Opt_debug, Opt_oldalloc, Opt_orlov, Opt_nobh, Opt_user_xattr, Opt_nouser_xattr, Opt_acl, Opt_noacl, Opt_xip, Opt_ignore, Opt_err, Opt_quota, - Opt_usrquota, Opt_grpquota + Opt_usrquota, Opt_grpquota, Opt_tag, Opt_notag, Opt_tagid }; static match_table_t tokens = { @@ -349,6 +349,10 @@ static match_table_t tokens = { {Opt_acl, "acl"}, {Opt_noacl, "noacl"}, {Opt_xip, "xip"}, + {Opt_tag, "tag"}, + {Opt_notag, "notag"}, + {Opt_tagid, "tagid=%u"}, + {Opt_tag, "tagxid"}, {Opt_grpquota, "grpquota"}, {Opt_ignore, "noquota"}, {Opt_quota, "quota"}, @@ -417,6 +421,20 @@ static int parse_options (char * options case Opt_nouid32: set_opt (sbi->s_mount_opt, NO_UID32); break; +#ifndef CONFIG_TAGGING_NONE + case Opt_tag: + set_opt (sbi->s_mount_opt, TAGGED); + break; + case Opt_notag: + clear_opt (sbi->s_mount_opt, TAGGED); + break; +#endif +#ifdef CONFIG_PROPAGATE + case Opt_tagid: + /* use args[0] */ + set_opt (sbi->s_mount_opt, TAGGED); + break; +#endif case Opt_nocheck: clear_opt (sbi->s_mount_opt, CHECK); break; @@ -727,6 +745,8 @@ static int ext2_fill_super(struct super_ if (!parse_options ((char *) data, sbi)) goto failed_mount; + if (EXT2_SB(sb)->s_mount_opt & EXT2_MOUNT_TAGGED) + sb->s_flags |= MS_TAGGED; sb->s_flags = (sb->s_flags & ~MS_POSIXACL) | ((EXT2_SB(sb)->s_mount_opt & EXT2_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0); @@ -1035,6 +1055,13 @@ static int ext2_remount (struct super_bl goto restore_opts; } + if ((sbi->s_mount_opt & EXT2_MOUNT_TAGGED) && + !(sb->s_flags & MS_TAGGED)) { + printk("EXT2-fs: %s: tagging not permitted on remount.\n", + sb->s_id); + return -EINVAL; + } + sb->s_flags = (sb->s_flags & ~MS_POSIXACL) | ((sbi->s_mount_opt & EXT2_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0); diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext2/symlink.c linux-source-2.6.22/fs/ext2/symlink.c --- linux-source-2.6.22.orig/fs/ext2/symlink.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext2/symlink.c 2007-10-24 16:14:06.291582822 +0200 @@ -38,6 +38,7 @@ const struct inode_operations ext2_symli .listxattr = ext2_listxattr, .removexattr = generic_removexattr, #endif + .sync_flags = ext2_sync_flags, }; const struct inode_operations ext2_fast_symlink_inode_operations = { @@ -49,4 +50,5 @@ const struct inode_operations ext2_fast_ .listxattr = ext2_listxattr, .removexattr = generic_removexattr, #endif + .sync_flags = ext2_sync_flags, }; diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext2/xattr.c linux-source-2.6.22/fs/ext2/xattr.c --- linux-source-2.6.22.orig/fs/ext2/xattr.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext2/xattr.c 2007-10-24 16:14:06.291582822 +0200 @@ -60,6 +60,7 @@ #include #include #include +#include #include "ext2.h" #include "xattr.h" #include "acl.h" @@ -641,8 +642,12 @@ ext2_xattr_set2(struct inode *inode, str the inode. */ ea_bdebug(new_bh, "reusing block"); + error = -ENOSPC; + if (DLIMIT_ALLOC_BLOCK(inode, 1)) + goto cleanup; error = -EDQUOT; if (DQUOT_ALLOC_BLOCK(inode, 1)) { + DLIMIT_FREE_BLOCK(inode, 1); unlock_buffer(new_bh); goto cleanup; } @@ -736,6 +741,7 @@ ext2_xattr_set2(struct inode *inode, str le32_to_cpu(HDR(old_bh)->h_refcount) - 1); if (ce) mb_cache_entry_release(ce); + DLIMIT_FREE_BLOCK(inode, 1); DQUOT_FREE_BLOCK(inode, 1); mark_buffer_dirty(old_bh); ea_bdebug(old_bh, "refcount now=%d", @@ -800,6 +806,7 @@ ext2_xattr_delete_inode(struct inode *in mark_buffer_dirty(bh); if (IS_SYNC(inode)) sync_dirty_buffer(bh); + DLIMIT_FREE_BLOCK(inode, 1); DQUOT_FREE_BLOCK(inode, 1); } EXT2_I(inode)->i_file_acl = 0; diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext3/balloc.c linux-source-2.6.22/fs/ext3/balloc.c --- linux-source-2.6.22.orig/fs/ext3/balloc.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext3/balloc.c 2007-10-24 16:14:06.291582822 +0200 @@ -19,6 +19,8 @@ #include #include #include +#include +#include /* * balloc.c contains the blocks allocation and deallocation routines @@ -613,8 +615,10 @@ void ext3_free_blocks(handle_t *handle, return; } ext3_free_blocks_sb(handle, sb, block, count, &dquot_freed_blocks); - if (dquot_freed_blocks) + if (dquot_freed_blocks) { + DLIMIT_FREE_BLOCK(inode, dquot_freed_blocks); DQUOT_FREE_BLOCK(inode, dquot_freed_blocks); + } return; } @@ -1353,18 +1357,33 @@ out: * * Check if filesystem has at least 1 free block available for allocation. */ -static int ext3_has_free_blocks(struct ext3_sb_info *sbi) +static int ext3_has_free_blocks(struct super_block *sb) { - ext3_fsblk_t free_blocks, root_blocks; + struct ext3_sb_info *sbi = EXT3_SB(sb); + unsigned long long free_blocks, root_blocks; + int cond; free_blocks = percpu_counter_read_positive(&sbi->s_freeblocks_counter); root_blocks = le32_to_cpu(sbi->s_es->s_r_blocks_count); - if (free_blocks < root_blocks + 1 && !capable(CAP_SYS_RESOURCE) && + + vxdprintk(VXD_CBIT(dlim, 3), + "ext3_has_free_blocks(%p): free=%llu, root=%llu", + sb, free_blocks, root_blocks); + + DLIMIT_ADJUST_BLOCK(sb, dx_current_tag(), &free_blocks, &root_blocks); + + cond = (free_blocks < root_blocks + 1 && + !capable(CAP_SYS_RESOURCE) && sbi->s_resuid != current->fsuid && - (sbi->s_resgid == 0 || !in_group_p (sbi->s_resgid))) { - return 0; - } - return 1; + (sbi->s_resgid == 0 || !in_group_p (sbi->s_resgid))); + + vxdprintk(VXD_CBIT(dlim, 3), + "ext3_has_free_blocks(%p): %llu<%llu+1, %c, %u!=%u r=%d", + sb, free_blocks, root_blocks, + !capable(CAP_SYS_RESOURCE)?'1':'0', + sbi->s_resuid, current->fsuid, cond?0:1); + + return (cond ? 0 : 1); } /** @@ -1381,7 +1400,7 @@ static int ext3_has_free_blocks(struct e */ int ext3_should_retry_alloc(struct super_block *sb, int *retries) { - if (!ext3_has_free_blocks(EXT3_SB(sb)) || (*retries)++ > 3) + if (!ext3_has_free_blocks(sb) || (*retries)++ > 3) return 0; jbd_debug(1, "%s: retrying operation after ENOSPC\n", sb->s_id); @@ -1444,6 +1463,8 @@ ext3_fsblk_t ext3_new_blocks(handle_t *h *errp = -EDQUOT; return 0; } + if (DLIMIT_ALLOC_BLOCK(inode, num)) + goto out_dlimit; sbi = EXT3_SB(sb); es = EXT3_SB(sb)->s_es; @@ -1460,7 +1481,7 @@ ext3_fsblk_t ext3_new_blocks(handle_t *h if (block_i && ((windowsz = block_i->rsv_window_node.rsv_goal_size) > 0)) my_rsv = &block_i->rsv_window_node; - if (!ext3_has_free_blocks(sbi)) { + if (!ext3_has_free_blocks(sb)) { *errp = -ENOSPC; goto out; } @@ -1647,12 +1668,16 @@ allocated: *errp = 0; brelse(bitmap_bh); DQUOT_FREE_BLOCK(inode, *count-num); + DLIMIT_FREE_BLOCK(inode, *count-num); *count = num; return ret_block; io_error: *errp = -EIO; out: + if (!performed_allocation) + DLIMIT_FREE_BLOCK(inode, *count); +out_dlimit: if (fatal) { *errp = fatal; ext3_std_error(sb, fatal); diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext3/file.c linux-source-2.6.22/fs/ext3/file.c --- linux-source-2.6.22.orig/fs/ext3/file.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext3/file.c 2007-10-24 16:14:06.291582822 +0200 @@ -121,6 +121,7 @@ const struct file_operations ext3_file_o .release = ext3_release_file, .fsync = ext3_sync_file, .sendfile = generic_file_sendfile, + .sendpage = generic_file_sendpage, .splice_read = generic_file_splice_read, .splice_write = generic_file_splice_write, }; @@ -135,5 +136,6 @@ const struct inode_operations ext3_file_ .removexattr = generic_removexattr, #endif .permission = ext3_permission, + .sync_flags = ext3_sync_flags, }; diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext3/ialloc.c linux-source-2.6.22/fs/ext3/ialloc.c --- linux-source-2.6.22.orig/fs/ext3/ialloc.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext3/ialloc.c 2007-10-24 16:14:06.291582822 +0200 @@ -23,6 +23,8 @@ #include #include #include +#include +#include #include @@ -127,6 +129,7 @@ void ext3_free_inode (handle_t *handle, ext3_xattr_delete_inode(handle, inode); DQUOT_FREE_INODE(inode); DQUOT_DROP(inode); + DLIMIT_FREE_INODE(inode); is_directory = S_ISDIR(inode->i_mode); @@ -445,6 +448,12 @@ struct inode *ext3_new_inode(handle_t *h inode = new_inode(sb); if (!inode) return ERR_PTR(-ENOMEM); + + inode->i_tag = dx_current_fstag(sb); + if (DLIMIT_ALLOC_INODE(inode)) { + err = -ENOSPC; + goto out_dlimit; + } ei = EXT3_I(inode); sbi = EXT3_SB(sb); @@ -566,7 +575,8 @@ got: ei->i_dir_start_lookup = 0; ei->i_disksize = 0; - ei->i_flags = EXT3_I(dir)->i_flags & ~EXT3_INDEX_FL; + ei->i_flags = EXT3_I(dir)->i_flags & + ~(EXT3_INDEX_FL|EXT3_IUNLINK_FL|EXT3_BARRIER_FL); if (S_ISLNK(mode)) ei->i_flags &= ~(EXT3_IMMUTABLE_FL|EXT3_APPEND_FL); /* dirsync only applies to directories */ @@ -621,6 +631,8 @@ got: fail: ext3_std_error(sb, err); out: + DLIMIT_FREE_INODE(inode); +out_dlimit: iput(inode); ret = ERR_PTR(err); really_out: @@ -632,6 +644,7 @@ fail_free_drop: fail_drop: DQUOT_DROP(inode); + DLIMIT_FREE_INODE(inode); inode->i_flags |= S_NOQUOTA; inode->i_nlink = 0; iput(inode); diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext3/inode.c linux-source-2.6.22/fs/ext3/inode.c --- linux-source-2.6.22.orig/fs/ext3/inode.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext3/inode.c 2007-10-24 16:14:06.291582822 +0200 @@ -36,6 +36,7 @@ #include #include #include +#include #include "xattr.h" #include "acl.h" @@ -2237,7 +2238,7 @@ void ext3_truncate(struct inode *inode) return; if (ext3_inode_is_fast_symlink(inode)) return; - if (IS_APPEND(inode) || IS_IMMUTABLE(inode)) + if (IS_APPEND(inode) || IS_IXORUNLINK(inode)) return; /* @@ -2559,13 +2560,20 @@ void ext3_set_inode_flags(struct inode * { unsigned int flags = EXT3_I(inode)->i_flags; - inode->i_flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC); + inode->i_flags &= ~(S_IMMUTABLE | S_IUNLINK | S_BARRIER | + S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC); + + if (flags & EXT3_IMMUTABLE_FL) + inode->i_flags |= S_IMMUTABLE; + if (flags & EXT3_IUNLINK_FL) + inode->i_flags |= S_IUNLINK; + if (flags & EXT3_BARRIER_FL) + inode->i_flags |= S_BARRIER; + if (flags & EXT3_SYNC_FL) inode->i_flags |= S_SYNC; if (flags & EXT3_APPEND_FL) inode->i_flags |= S_APPEND; - if (flags & EXT3_IMMUTABLE_FL) - inode->i_flags |= S_IMMUTABLE; if (flags & EXT3_NOATIME_FL) inode->i_flags |= S_NOATIME; if (flags & EXT3_DIRSYNC_FL) @@ -2591,6 +2599,45 @@ void ext3_get_inode_flags(struct ext3_in ei->i_flags |= EXT3_DIRSYNC_FL; } +int ext3_sync_flags(struct inode *inode) +{ + unsigned int oldflags, newflags; + int err = 0; + + oldflags = EXT3_I(inode)->i_flags; + newflags = oldflags & ~(EXT3_IMMUTABLE_FL | + EXT3_IUNLINK_FL | EXT3_BARRIER_FL); + + if (IS_IMMUTABLE(inode)) + newflags |= EXT3_IMMUTABLE_FL; + if (IS_IUNLINK(inode)) + newflags |= EXT3_IUNLINK_FL; + if (IS_BARRIER(inode)) + newflags |= EXT3_BARRIER_FL; + + if (oldflags ^ newflags) { + handle_t *handle; + struct ext3_iloc iloc; + + handle = ext3_journal_start(inode, 1); + if (IS_ERR(handle)) + return PTR_ERR(handle); + if (IS_SYNC(inode)) + handle->h_sync = 1; + err = ext3_reserve_inode_write(handle, inode, &iloc); + if (err) + goto flags_err; + + EXT3_I(inode)->i_flags = newflags; + inode->i_ctime = CURRENT_TIME; + + err = ext3_mark_iloc_dirty(handle, inode, &iloc); + flags_err: + ext3_journal_stop(handle); + } + return err; +} + void ext3_read_inode(struct inode * inode) { struct ext3_iloc iloc; @@ -2598,6 +2645,8 @@ void ext3_read_inode(struct inode * inod struct ext3_inode_info *ei = EXT3_I(inode); struct buffer_head *bh; int block; + uid_t uid; + gid_t gid; #ifdef CONFIG_EXT3_FS_POSIX_ACL ei->i_acl = EXT3_ACL_NOT_CACHED; @@ -2610,12 +2659,17 @@ void ext3_read_inode(struct inode * inod bh = iloc.bh; raw_inode = ext3_raw_inode(&iloc); inode->i_mode = le16_to_cpu(raw_inode->i_mode); - inode->i_uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low); - inode->i_gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low); + uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low); + gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low); if(!(test_opt (inode->i_sb, NO_UID32))) { - inode->i_uid |= le16_to_cpu(raw_inode->i_uid_high) << 16; - inode->i_gid |= le16_to_cpu(raw_inode->i_gid_high) << 16; + uid |= le16_to_cpu(raw_inode->i_uid_high) << 16; + gid |= le16_to_cpu(raw_inode->i_gid_high) << 16; } + inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid); + inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid); + inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid, + le16_to_cpu(raw_inode->i_raw_tag)); + inode->i_nlink = le16_to_cpu(raw_inode->i_links_count); inode->i_size = le32_to_cpu(raw_inode->i_size); inode->i_atime.tv_sec = (signed)le32_to_cpu(raw_inode->i_atime); @@ -2741,6 +2795,8 @@ static int ext3_do_update_inode(handle_t struct ext3_inode *raw_inode = ext3_raw_inode(iloc); struct ext3_inode_info *ei = EXT3_I(inode); struct buffer_head *bh = iloc->bh; + uid_t uid = TAGINO_UID(DX_TAG(inode), inode->i_uid, inode->i_tag); + gid_t gid = TAGINO_GID(DX_TAG(inode), inode->i_gid, inode->i_tag); int err = 0, rc, block; /* For fields not not tracking in the in-memory inode, @@ -2751,29 +2807,32 @@ static int ext3_do_update_inode(handle_t ext3_get_inode_flags(ei); raw_inode->i_mode = cpu_to_le16(inode->i_mode); if(!(test_opt(inode->i_sb, NO_UID32))) { - raw_inode->i_uid_low = cpu_to_le16(low_16_bits(inode->i_uid)); - raw_inode->i_gid_low = cpu_to_le16(low_16_bits(inode->i_gid)); + raw_inode->i_uid_low = cpu_to_le16(low_16_bits(uid)); + raw_inode->i_gid_low = cpu_to_le16(low_16_bits(gid)); /* * Fix up interoperability with old kernels. Otherwise, old inodes get * re-used with the upper 16 bits of the uid/gid intact */ if(!ei->i_dtime) { raw_inode->i_uid_high = - cpu_to_le16(high_16_bits(inode->i_uid)); + cpu_to_le16(high_16_bits(uid)); raw_inode->i_gid_high = - cpu_to_le16(high_16_bits(inode->i_gid)); + cpu_to_le16(high_16_bits(gid)); } else { raw_inode->i_uid_high = 0; raw_inode->i_gid_high = 0; } } else { raw_inode->i_uid_low = - cpu_to_le16(fs_high2lowuid(inode->i_uid)); + cpu_to_le16(fs_high2lowuid(uid)); raw_inode->i_gid_low = - cpu_to_le16(fs_high2lowgid(inode->i_gid)); + cpu_to_le16(fs_high2lowgid(gid)); raw_inode->i_uid_high = 0; raw_inode->i_gid_high = 0; } +#ifdef CONFIG_TAGGING_INTERN + raw_inode->i_raw_tag = cpu_to_le16(inode->i_tag); +#endif raw_inode->i_links_count = cpu_to_le16(inode->i_nlink); raw_inode->i_size = cpu_to_le32(ei->i_disksize); raw_inode->i_atime = cpu_to_le32(inode->i_atime.tv_sec); @@ -2926,7 +2985,8 @@ int ext3_setattr(struct dentry *dentry, return error; if ((ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) || - (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) { + (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid) || + (ia_valid & ATTR_TAG && attr->ia_tag != inode->i_tag)) { handle_t *handle; /* (user+group)*(old+new) structure, inode write (sb, @@ -2948,6 +3008,8 @@ int ext3_setattr(struct dentry *dentry, inode->i_uid = attr->ia_uid; if (attr->ia_valid & ATTR_GID) inode->i_gid = attr->ia_gid; + if ((attr->ia_valid & ATTR_TAG) && IS_TAGGED(inode)) + inode->i_tag = attr->ia_tag; error = ext3_mark_inode_dirty(handle, inode); ext3_journal_stop(handle); } diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext3/ioctl.c linux-source-2.6.22/fs/ext3/ioctl.c --- linux-source-2.6.22.orig/fs/ext3/ioctl.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext3/ioctl.c 2007-10-24 16:14:06.291582822 +0200 @@ -8,6 +8,7 @@ */ #include +#include #include #include #include @@ -15,6 +16,7 @@ #include #include #include +#include #include int ext3_ioctl (struct inode * inode, struct file * filp, unsigned int cmd, @@ -38,7 +40,8 @@ int ext3_ioctl (struct inode * inode, st unsigned int oldflags; unsigned int jflag; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER)) @@ -62,7 +65,9 @@ int ext3_ioctl (struct inode * inode, st * * This test looks nicer. Thanks to Pauline Middelink */ - if ((flags ^ oldflags) & (EXT3_APPEND_FL | EXT3_IMMUTABLE_FL)) { + if ((oldflags & EXT3_IMMUTABLE_FL) || + ((flags ^ oldflags) & (EXT3_APPEND_FL | + EXT3_IMMUTABLE_FL | EXT3_IUNLINK_FL))) { if (!capable(CAP_LINUX_IMMUTABLE)) { mutex_unlock(&inode->i_mutex); return -EPERM; @@ -124,7 +129,8 @@ flags_err: if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER)) return -EPERM; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if (get_user(generation, (int __user *) arg)) return -EFAULT; @@ -178,7 +184,8 @@ flags_err: if (!test_opt(inode->i_sb, RESERVATION) ||!S_ISREG(inode->i_mode)) return -ENOTTY; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER)) @@ -213,7 +220,8 @@ flags_err: if (!capable(CAP_SYS_RESOURCE)) return -EPERM; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if (get_user(n_blocks_count, (__u32 __user *)arg)) @@ -234,7 +242,8 @@ flags_err: if (!capable(CAP_SYS_RESOURCE)) return -EPERM; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if (copy_from_user(&input, (struct ext3_new_group_input __user *)arg, @@ -249,6 +258,38 @@ flags_err: return err; } +#if defined(CONFIG_VSERVER_LEGACY) && !defined(CONFIG_TAGGING_NONE) + case EXT3_IOC_SETTAG: { + handle_t *handle; + struct ext3_iloc iloc; + int tag; + int err; + + /* fixme: if stealth, return -ENOTTY */ + if (!capable(CAP_CONTEXT)) + return -EPERM; + if (IS_RDONLY(inode)) + return -EROFS; + if (!(inode->i_sb->s_flags & MS_TAGGED)) + return -ENOSYS; + if (get_user(tag, (int __user *) arg)) + return -EFAULT; + + handle = ext3_journal_start(inode, 1); + if (IS_ERR(handle)) + return PTR_ERR(handle); + err = ext3_reserve_inode_write(handle, inode, &iloc); + if (err) + return err; + + inode->i_tag = (tag & 0xFFFF); + inode->i_ctime = CURRENT_TIME; + + err = ext3_mark_iloc_dirty(handle, inode, &iloc); + ext3_journal_stop(handle); + return err; + } +#endif default: return -ENOTTY; diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext3/namei.c linux-source-2.6.22/fs/ext3/namei.c --- linux-source-2.6.22.orig/fs/ext3/namei.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/fs/ext3/namei.c 2007-10-24 16:14:06.291582822 +0200 @@ -36,6 +36,7 @@ #include #include #include +#include #include "namei.h" #include "xattr.h" @@ -1052,6 +1053,7 @@ static struct dentry *ext3_lookup(struct if (!inode) return ERR_PTR(-EACCES); + dx_propagate_tag(nd, inode); } return d_splice_alias(inode, dentry); } @@ -2443,6 +2445,7 @@ const struct inode_operations ext3_dir_i .removexattr = generic_removexattr, #endif .permission = ext3_permission, + .sync_flags = ext3_sync_flags, }; const struct inode_operations ext3_special_inode_operations = { @@ -2454,4 +2457,5 @@ const struct inode_operations ext3_speci .removexattr = generic_removexattr, #endif .permission = ext3_permission, + .sync_flags = ext3_sync_flags, }; diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext3/super.c linux-source-2.6.22/fs/ext3/super.c --- linux-source-2.6.22.orig/fs/ext3/super.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext3/super.c 2007-10-24 16:14:06.291582822 +0200 @@ -674,7 +674,7 @@ enum { Opt_usrjquota, Opt_grpjquota, Opt_offusrjquota, Opt_offgrpjquota, Opt_jqfmt_vfsold, Opt_jqfmt_vfsv0, Opt_quota, Opt_noquota, Opt_ignore, Opt_barrier, Opt_err, Opt_resize, Opt_usrquota, - Opt_grpquota + Opt_grpquota, Opt_tag, Opt_notag, Opt_tagid }; static match_table_t tokens = { @@ -724,6 +724,10 @@ static match_table_t tokens = { {Opt_quota, "quota"}, {Opt_usrquota, "usrquota"}, {Opt_barrier, "barrier=%u"}, + {Opt_tag, "tag"}, + {Opt_notag, "notag"}, + {Opt_tagid, "tagid=%u"}, + {Opt_tag, "tagxid"}, {Opt_err, NULL}, {Opt_resize, "resize"}, }; @@ -817,6 +821,20 @@ static int parse_options (char *options, case Opt_nouid32: set_opt (sbi->s_mount_opt, NO_UID32); break; +#ifndef CONFIG_TAGGING_NONE + case Opt_tag: + set_opt (sbi->s_mount_opt, TAGGED); + break; + case Opt_notag: + clear_opt (sbi->s_mount_opt, TAGGED); + break; +#endif +#ifdef CONFIG_PROPAGATE + case Opt_tagid: + /* use args[0] */ + set_opt (sbi->s_mount_opt, TAGGED); + break; +#endif case Opt_nocheck: clear_opt (sbi->s_mount_opt, CHECK); break; @@ -1487,6 +1505,9 @@ static int ext3_fill_super (struct super NULL, 0)) goto failed_mount; + if (EXT3_SB(sb)->s_mount_opt & EXT3_MOUNT_TAGGED) + sb->s_flags |= MS_TAGGED; + sb->s_flags = (sb->s_flags & ~MS_POSIXACL) | ((sbi->s_mount_opt & EXT3_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0); @@ -2302,6 +2323,12 @@ static int ext3_remount (struct super_bl if (sbi->s_mount_opt & EXT3_MOUNT_ABORT) ext3_abort(sb, __FUNCTION__, "Abort forced by user"); + if ((sbi->s_mount_opt & EXT3_MOUNT_TAGGED) && + !(sb->s_flags & MS_TAGGED)) { + printk("EXT3-fs: %s: tagging not permitted on remount.\n", + sb->s_id); + return -EINVAL; + } sb->s_flags = (sb->s_flags & ~MS_POSIXACL) | ((sbi->s_mount_opt & EXT3_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0); diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext3/symlink.c linux-source-2.6.22/fs/ext3/symlink.c --- linux-source-2.6.22.orig/fs/ext3/symlink.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext3/symlink.c 2007-10-24 16:14:06.291582822 +0200 @@ -40,6 +40,7 @@ const struct inode_operations ext3_symli .listxattr = ext3_listxattr, .removexattr = generic_removexattr, #endif + .sync_flags = ext3_sync_flags, }; const struct inode_operations ext3_fast_symlink_inode_operations = { @@ -51,4 +52,5 @@ const struct inode_operations ext3_fast_ .listxattr = ext3_listxattr, .removexattr = generic_removexattr, #endif + .sync_flags = ext3_sync_flags, }; diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext3/xattr.c linux-source-2.6.22/fs/ext3/xattr.c --- linux-source-2.6.22.orig/fs/ext3/xattr.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext3/xattr.c 2007-10-24 16:14:06.291582822 +0200 @@ -58,6 +58,7 @@ #include #include #include +#include #include "xattr.h" #include "acl.h" @@ -497,6 +498,7 @@ ext3_xattr_release_block(handle_t *handl error = ext3_journal_dirty_metadata(handle, bh); if (IS_SYNC(inode)) handle->h_sync = 1; + DLIMIT_FREE_BLOCK(inode, 1); DQUOT_FREE_BLOCK(inode, 1); ea_bdebug(bh, "refcount now=%d; releasing", le32_to_cpu(BHDR(bh)->h_refcount)); @@ -771,11 +773,14 @@ inserted: if (new_bh == bs->bh) ea_bdebug(new_bh, "keeping"); else { + error = -ENOSPC; + if (DLIMIT_ALLOC_BLOCK(inode, 1)) + goto cleanup; /* The old block is released after updating the inode. */ error = -EDQUOT; if (DQUOT_ALLOC_BLOCK(inode, 1)) - goto cleanup; + goto cleanup_dlimit; error = ext3_journal_get_write_access(handle, new_bh); if (error) @@ -852,6 +857,8 @@ cleanup: cleanup_dquot: DQUOT_FREE_BLOCK(inode, 1); +cleanup_dlimit: + DLIMIT_FREE_BLOCK(inode, 1); goto cleanup; bad_block: diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext4/balloc.c linux-source-2.6.22/fs/ext4/balloc.c --- linux-source-2.6.22.orig/fs/ext4/balloc.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext4/balloc.c 2007-10-24 16:14:06.291582822 +0200 @@ -19,6 +19,8 @@ #include #include #include +#include +#include /* * balloc.c contains the blocks allocation and deallocation routines @@ -630,8 +632,10 @@ void ext4_free_blocks(handle_t *handle, return; } ext4_free_blocks_sb(handle, sb, block, count, &dquot_freed_blocks); - if (dquot_freed_blocks) + if (dquot_freed_blocks) { + DLIMIT_FREE_BLOCK(inode, dquot_freed_blocks); DQUOT_FREE_BLOCK(inode, dquot_freed_blocks); + } return; } @@ -1370,18 +1374,33 @@ out: * * Check if filesystem has at least 1 free block available for allocation. */ -static int ext4_has_free_blocks(struct ext4_sb_info *sbi) +static int ext4_has_free_blocks(struct super_block *sb) { + struct ext4_sb_info *sbi = EXT4_SB(sb); ext4_fsblk_t free_blocks, root_blocks; + int cond; free_blocks = percpu_counter_read_positive(&sbi->s_freeblocks_counter); root_blocks = ext4_r_blocks_count(sbi->s_es); - if (free_blocks < root_blocks + 1 && !capable(CAP_SYS_RESOURCE) && + + vxdprintk(VXD_CBIT(dlim, 3), + "ext4_has_free_blocks(%p): free=%llu, root=%llu", + sb, free_blocks, root_blocks); + + DLIMIT_ADJUST_BLOCK(sb, dx_current_tag(), &free_blocks, &root_blocks); + + cond = (free_blocks < root_blocks + 1 && + !capable(CAP_SYS_RESOURCE) && sbi->s_resuid != current->fsuid && - (sbi->s_resgid == 0 || !in_group_p (sbi->s_resgid))) { - return 0; - } - return 1; + (sbi->s_resgid == 0 || !in_group_p (sbi->s_resgid))); + + vxdprintk(VXD_CBIT(dlim, 3), + "ext4_has_free_blocks(%p): %llu<%llu+1, %c, %u!=%u r=%d", + sb, free_blocks, root_blocks, + !capable(CAP_SYS_RESOURCE)?'1':'0', + sbi->s_resuid, current->fsuid, cond?0:1); + + return (cond ? 0 : 1); } /** @@ -1398,7 +1417,7 @@ static int ext4_has_free_blocks(struct e */ int ext4_should_retry_alloc(struct super_block *sb, int *retries) { - if (!ext4_has_free_blocks(EXT4_SB(sb)) || (*retries)++ > 3) + if (!ext4_has_free_blocks(sb) || (*retries)++ > 3) return 0; jbd_debug(1, "%s: retrying operation after ENOSPC\n", sb->s_id); @@ -1461,6 +1480,8 @@ ext4_fsblk_t ext4_new_blocks(handle_t *h *errp = -EDQUOT; return 0; } + if (DLIMIT_ALLOC_BLOCK(inode, num)) + goto out_dlimit; sbi = EXT4_SB(sb); es = EXT4_SB(sb)->s_es; @@ -1477,7 +1498,7 @@ ext4_fsblk_t ext4_new_blocks(handle_t *h if (block_i && ((windowsz = block_i->rsv_window_node.rsv_goal_size) > 0)) my_rsv = &block_i->rsv_window_node; - if (!ext4_has_free_blocks(sbi)) { + if (!ext4_has_free_blocks(sb)) { *errp = -ENOSPC; goto out; } @@ -1661,12 +1682,16 @@ allocated: *errp = 0; brelse(bitmap_bh); DQUOT_FREE_BLOCK(inode, *count-num); + DLIMIT_FREE_BLOCK(inode, *count-num); *count = num; return ret_block; io_error: *errp = -EIO; out: + if (!performed_allocation) + DLIMIT_FREE_BLOCK(inode, *count); +out_dlimit: if (fatal) { *errp = fatal; ext4_std_error(sb, fatal); diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext4/file.c linux-source-2.6.22/fs/ext4/file.c --- linux-source-2.6.22.orig/fs/ext4/file.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext4/file.c 2007-10-24 16:14:06.291582822 +0200 @@ -121,6 +121,7 @@ const struct file_operations ext4_file_o .release = ext4_release_file, .fsync = ext4_sync_file, .sendfile = generic_file_sendfile, + .sendpage = generic_file_sendpage, .splice_read = generic_file_splice_read, .splice_write = generic_file_splice_write, }; @@ -135,5 +136,6 @@ const struct inode_operations ext4_file_ .removexattr = generic_removexattr, #endif .permission = ext4_permission, + .sync_flags = ext4_sync_flags, }; diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext4/ialloc.c linux-source-2.6.22/fs/ext4/ialloc.c --- linux-source-2.6.22.orig/fs/ext4/ialloc.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext4/ialloc.c 2007-10-24 16:14:06.291582822 +0200 @@ -24,6 +24,8 @@ #include #include #include +#include +#include #include #include "xattr.h" @@ -127,6 +129,7 @@ void ext4_free_inode (handle_t *handle, ext4_xattr_delete_inode(handle, inode); DQUOT_FREE_INODE(inode); DQUOT_DROP(inode); + DLIMIT_FREE_INODE(inode); is_directory = S_ISDIR(inode->i_mode); @@ -448,6 +451,12 @@ struct inode *ext4_new_inode(handle_t *h inode = new_inode(sb); if (!inode) return ERR_PTR(-ENOMEM); + + inode->i_tag = dx_current_fstag(sb); + if (DLIMIT_ALLOC_INODE(inode)) { + err = -ENOSPC; + goto out_dlimit; + } ei = EXT4_I(inode); sbi = EXT4_SB(sb); @@ -569,7 +578,8 @@ got: ei->i_dir_start_lookup = 0; ei->i_disksize = 0; - ei->i_flags = EXT4_I(dir)->i_flags & ~EXT4_INDEX_FL; + ei->i_flags = EXT4_I(dir)->i_flags & + ~(EXT4_INDEX_FL|EXT4_IUNLINK_FL|EXT4_BARRIER_FL); if (S_ISLNK(mode)) ei->i_flags &= ~(EXT4_IMMUTABLE_FL|EXT4_APPEND_FL); /* dirsync only applies to directories */ @@ -635,6 +645,8 @@ got: fail: ext4_std_error(sb, err); out: + DLIMIT_FREE_INODE(inode); +out_dlimit: iput(inode); ret = ERR_PTR(err); really_out: @@ -646,6 +658,7 @@ fail_free_drop: fail_drop: DQUOT_DROP(inode); + DLIMIT_FREE_INODE(inode); inode->i_flags |= S_NOQUOTA; inode->i_nlink = 0; iput(inode); diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext4/inode.c linux-source-2.6.22/fs/ext4/inode.c --- linux-source-2.6.22.orig/fs/ext4/inode.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext4/inode.c 2007-10-24 16:14:06.291582822 +0200 @@ -36,6 +36,7 @@ #include #include #include +#include #include "xattr.h" #include "acl.h" @@ -2244,7 +2245,7 @@ void ext4_truncate(struct inode *inode) return; if (ext4_inode_is_fast_symlink(inode)) return; - if (IS_APPEND(inode) || IS_IMMUTABLE(inode)) + if (IS_APPEND(inode) || IS_IXORUNLINK(inode)) return; /* @@ -2570,19 +2571,65 @@ void ext4_set_inode_flags(struct inode * { unsigned int flags = EXT4_I(inode)->i_flags; - inode->i_flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC); + inode->i_flags &= ~(S_IMMUTABLE | S_IUNLINK | S_BARRIER | + S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC); + + if (flags & EXT4_IMMUTABLE_FL) + inode->i_flags |= S_IMMUTABLE; + if (flags & EXT4_IUNLINK_FL) + inode->i_flags |= S_IUNLINK; + if (flags & EXT4_BARRIER_FL) + inode->i_flags |= S_BARRIER; + if (flags & EXT4_SYNC_FL) inode->i_flags |= S_SYNC; if (flags & EXT4_APPEND_FL) inode->i_flags |= S_APPEND; - if (flags & EXT4_IMMUTABLE_FL) - inode->i_flags |= S_IMMUTABLE; if (flags & EXT4_NOATIME_FL) inode->i_flags |= S_NOATIME; if (flags & EXT4_DIRSYNC_FL) inode->i_flags |= S_DIRSYNC; } +int ext4_sync_flags(struct inode *inode) +{ + unsigned int oldflags, newflags; + int err = 0; + + oldflags = EXT4_I(inode)->i_flags; + newflags = oldflags & ~(EXT4_IMMUTABLE_FL | + EXT4_IUNLINK_FL | EXT4_BARRIER_FL); + + if (IS_IMMUTABLE(inode)) + newflags |= EXT4_IMMUTABLE_FL; + if (IS_IUNLINK(inode)) + newflags |= EXT4_IUNLINK_FL; + if (IS_BARRIER(inode)) + newflags |= EXT4_BARRIER_FL; + + if (oldflags ^ newflags) { + handle_t *handle; + struct ext4_iloc iloc; + + handle = ext4_journal_start(inode, 1); + if (IS_ERR(handle)) + return PTR_ERR(handle); + if (IS_SYNC(inode)) + handle->h_sync = 1; + err = ext4_reserve_inode_write(handle, inode, &iloc); + if (err) + goto flags_err; + + EXT4_I(inode)->i_flags = newflags; + inode->i_ctime = CURRENT_TIME; + + err = ext4_mark_iloc_dirty(handle, inode, &iloc); + flags_err: + ext4_journal_stop(handle); + } + return err; +} + void ext4_read_inode(struct inode * inode) { struct ext4_iloc iloc; @@ -2590,6 +2637,8 @@ void ext4_read_inode(struct inode * inod struct ext4_inode_info *ei = EXT4_I(inode); struct buffer_head *bh; int block; + uid_t uid; + gid_t gid; #ifdef CONFIG_EXT4DEV_FS_POSIX_ACL ei->i_acl = EXT4_ACL_NOT_CACHED; @@ -2602,12 +2651,17 @@ void ext4_read_inode(struct inode * inod bh = iloc.bh; raw_inode = ext4_raw_inode(&iloc); inode->i_mode = le16_to_cpu(raw_inode->i_mode); - inode->i_uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low); - inode->i_gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low); + uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low); + gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low); if(!(test_opt (inode->i_sb, NO_UID32))) { - inode->i_uid |= le16_to_cpu(raw_inode->i_uid_high) << 16; - inode->i_gid |= le16_to_cpu(raw_inode->i_gid_high) << 16; + uid |= le16_to_cpu(raw_inode->i_uid_high) << 16; + gid |= le16_to_cpu(raw_inode->i_gid_high) << 16; } + inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid); + inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid); + inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid, + le16_to_cpu(raw_inode->i_raw_tag)); + inode->i_nlink = le16_to_cpu(raw_inode->i_links_count); inode->i_size = le32_to_cpu(raw_inode->i_size); inode->i_atime.tv_sec = (signed)le32_to_cpu(raw_inode->i_atime); @@ -2737,6 +2791,8 @@ static int ext4_do_update_inode(handle_t struct ext4_inode *raw_inode = ext4_raw_inode(iloc); struct ext4_inode_info *ei = EXT4_I(inode); struct buffer_head *bh = iloc->bh; + uid_t uid = TAGINO_UID(DX_TAG(inode), inode->i_uid, inode->i_tag); + gid_t gid = TAGINO_GID(DX_TAG(inode), inode->i_gid, inode->i_tag); int err = 0, rc, block; /* For fields not not tracking in the in-memory inode, @@ -2746,29 +2802,32 @@ static int ext4_do_update_inode(handle_t raw_inode->i_mode = cpu_to_le16(inode->i_mode); if(!(test_opt(inode->i_sb, NO_UID32))) { - raw_inode->i_uid_low = cpu_to_le16(low_16_bits(inode->i_uid)); - raw_inode->i_gid_low = cpu_to_le16(low_16_bits(inode->i_gid)); + raw_inode->i_uid_low = cpu_to_le16(low_16_bits(uid)); + raw_inode->i_gid_low = cpu_to_le16(low_16_bits(gid)); /* * Fix up interoperability with old kernels. Otherwise, old inodes get * re-used with the upper 16 bits of the uid/gid intact */ if(!ei->i_dtime) { raw_inode->i_uid_high = - cpu_to_le16(high_16_bits(inode->i_uid)); + cpu_to_le16(high_16_bits(uid)); raw_inode->i_gid_high = - cpu_to_le16(high_16_bits(inode->i_gid)); + cpu_to_le16(high_16_bits(gid)); } else { raw_inode->i_uid_high = 0; raw_inode->i_gid_high = 0; } } else { raw_inode->i_uid_low = - cpu_to_le16(fs_high2lowuid(inode->i_uid)); + cpu_to_le16(fs_high2lowuid(uid)); raw_inode->i_gid_low = - cpu_to_le16(fs_high2lowgid(inode->i_gid)); + cpu_to_le16(fs_high2lowgid(gid)); raw_inode->i_uid_high = 0; raw_inode->i_gid_high = 0; } +#ifdef CONFIG_TAGGING_INTERN + raw_inode->i_raw_tag = cpu_to_le16(inode->i_tag); +#endif raw_inode->i_links_count = cpu_to_le16(inode->i_nlink); raw_inode->i_size = cpu_to_le32(ei->i_disksize); raw_inode->i_atime = cpu_to_le32(inode->i_atime.tv_sec); @@ -2925,7 +2984,8 @@ int ext4_setattr(struct dentry *dentry, return error; if ((ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) || - (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) { + (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid) || + (ia_valid & ATTR_TAG && attr->ia_tag != inode->i_tag)) { handle_t *handle; /* (user+group)*(old+new) structure, inode write (sb, @@ -2947,6 +3007,8 @@ int ext4_setattr(struct dentry *dentry, inode->i_uid = attr->ia_uid; if (attr->ia_valid & ATTR_GID) inode->i_gid = attr->ia_gid; + if ((attr->ia_valid & ATTR_TAG) && IS_TAGGED(inode)) + inode->i_tag = attr->ia_tag; error = ext4_mark_inode_dirty(handle, inode); ext4_journal_stop(handle); } diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext4/ioctl.c linux-source-2.6.22/fs/ext4/ioctl.c --- linux-source-2.6.22.orig/fs/ext4/ioctl.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext4/ioctl.c 2007-10-24 16:14:06.291582822 +0200 @@ -8,6 +8,7 @@ */ #include +#include #include #include #include @@ -15,6 +16,7 @@ #include #include #include +#include #include int ext4_ioctl (struct inode * inode, struct file * filp, unsigned int cmd, @@ -37,7 +39,8 @@ int ext4_ioctl (struct inode * inode, st unsigned int oldflags; unsigned int jflag; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER)) @@ -61,7 +64,9 @@ int ext4_ioctl (struct inode * inode, st * * This test looks nicer. Thanks to Pauline Middelink */ - if ((flags ^ oldflags) & (EXT4_APPEND_FL | EXT4_IMMUTABLE_FL)) { + if ((oldflags & EXT4_IMMUTABLE_FL) || + ((flags ^ oldflags) & (EXT4_APPEND_FL | + EXT4_IMMUTABLE_FL | EXT4_IUNLINK_FL))) { if (!capable(CAP_LINUX_IMMUTABLE)) { mutex_unlock(&inode->i_mutex); return -EPERM; @@ -123,7 +128,8 @@ flags_err: if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER)) return -EPERM; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if (get_user(generation, (int __user *) arg)) return -EFAULT; @@ -177,7 +183,8 @@ flags_err: if (!test_opt(inode->i_sb, RESERVATION) ||!S_ISREG(inode->i_mode)) return -ENOTTY; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER)) @@ -212,7 +219,8 @@ flags_err: if (!capable(CAP_SYS_RESOURCE)) return -EPERM; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if (get_user(n_blocks_count, (__u32 __user *)arg)) @@ -233,7 +241,8 @@ flags_err: if (!capable(CAP_SYS_RESOURCE)) return -EPERM; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if (copy_from_user(&input, (struct ext4_new_group_input __user *)arg, @@ -248,6 +257,39 @@ flags_err: return err; } +#if defined(CONFIG_VSERVER_LEGACY) && !defined(CONFIG_TAGGING_NONE) + case EXT4_IOC_SETTAG: { + handle_t *handle; + struct ext4_iloc iloc; + int tag; + int err; + + /* fixme: if stealth, return -ENOTTY */ + if (!capable(CAP_CONTEXT)) + return -EPERM; + if (IS_RDONLY(inode)) + return -EROFS; + if (!(inode->i_sb->s_flags & MS_TAGGED)) + return -ENOSYS; + if (get_user(tag, (int __user *) arg)) + return -EFAULT; + + handle = ext4_journal_start(inode, 1); + if (IS_ERR(handle)) + return PTR_ERR(handle); + err = ext4_reserve_inode_write(handle, inode, &iloc); + if (err) + return err; + + inode->i_tag = (tag & 0xFFFF); + inode->i_ctime = CURRENT_TIME; + + err = ext4_mark_iloc_dirty(handle, inode, &iloc); + ext4_journal_stop(handle); + return err; + } +#endif + default: return -ENOTTY; } diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext4/namei.c linux-source-2.6.22/fs/ext4/namei.c --- linux-source-2.6.22.orig/fs/ext4/namei.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/fs/ext4/namei.c 2007-10-24 16:14:06.291582822 +0200 @@ -36,6 +36,7 @@ #include #include #include +#include #include "namei.h" #include "xattr.h" @@ -1050,6 +1051,7 @@ static struct dentry *ext4_lookup(struct if (!inode) return ERR_PTR(-EACCES); + dx_propagate_tag(nd, inode); } return d_splice_alias(inode, dentry); } @@ -2441,6 +2443,7 @@ const struct inode_operations ext4_dir_i .removexattr = generic_removexattr, #endif .permission = ext4_permission, + .sync_flags = ext4_sync_flags, }; const struct inode_operations ext4_special_inode_operations = { @@ -2452,4 +2455,5 @@ const struct inode_operations ext4_speci .removexattr = generic_removexattr, #endif .permission = ext4_permission, + .sync_flags = ext4_sync_flags, }; diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext4/super.c linux-source-2.6.22/fs/ext4/super.c --- linux-source-2.6.22.orig/fs/ext4/super.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext4/super.c 2007-10-24 16:14:06.291582822 +0200 @@ -725,7 +725,7 @@ enum { Opt_usrjquota, Opt_grpjquota, Opt_offusrjquota, Opt_offgrpjquota, Opt_jqfmt_vfsold, Opt_jqfmt_vfsv0, Opt_quota, Opt_noquota, Opt_ignore, Opt_barrier, Opt_err, Opt_resize, Opt_usrquota, - Opt_grpquota, Opt_extents, + Opt_grpquota, Opt_extents, Opt_tag, Opt_notag, Opt_tagid }; static match_table_t tokens = { @@ -776,6 +776,10 @@ static match_table_t tokens = { {Opt_usrquota, "usrquota"}, {Opt_barrier, "barrier=%u"}, {Opt_extents, "extents"}, + {Opt_tag, "tag"}, + {Opt_notag, "notag"}, + {Opt_tagid, "tagid=%u"}, + {Opt_tag, "tagxid"}, {Opt_err, NULL}, {Opt_resize, "resize"}, }; @@ -869,6 +873,20 @@ static int parse_options (char *options, case Opt_nouid32: set_opt (sbi->s_mount_opt, NO_UID32); break; +#ifndef CONFIG_TAGGING_NONE + case Opt_tag: + set_opt (sbi->s_mount_opt, TAGGED); + break; + case Opt_notag: + clear_opt (sbi->s_mount_opt, TAGGED); + break; +#endif +#ifdef CONFIG_PROPAGATE + case Opt_tagid: + /* use args[0] */ + set_opt (sbi->s_mount_opt, TAGGED); + break; +#endif case Opt_nocheck: clear_opt (sbi->s_mount_opt, CHECK); break; @@ -1546,6 +1564,9 @@ static int ext4_fill_super (struct super NULL, 0)) goto failed_mount; + if (EXT4_SB(sb)->s_mount_opt & EXT4_MOUNT_TAGGED) + sb->s_flags |= MS_TAGGED; + sb->s_flags = (sb->s_flags & ~MS_POSIXACL) | ((sbi->s_mount_opt & EXT4_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0); @@ -2377,6 +2398,12 @@ static int ext4_remount (struct super_bl if (sbi->s_mount_opt & EXT4_MOUNT_ABORT) ext4_abort(sb, __FUNCTION__, "Abort forced by user"); + if ((sbi->s_mount_opt & EXT4_MOUNT_TAGGED) && + !(sb->s_flags & MS_TAGGED)) { + printk("EXT4-fs: %s: tagging not permitted on remount.\n", + sb->s_id); + return -EINVAL; + } sb->s_flags = (sb->s_flags & ~MS_POSIXACL) | ((sbi->s_mount_opt & EXT4_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0); diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext4/symlink.c linux-source-2.6.22/fs/ext4/symlink.c --- linux-source-2.6.22.orig/fs/ext4/symlink.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext4/symlink.c 2007-10-24 16:14:06.291582822 +0200 @@ -40,6 +40,7 @@ const struct inode_operations ext4_symli .listxattr = ext4_listxattr, .removexattr = generic_removexattr, #endif + .sync_flags = ext4_sync_flags, }; const struct inode_operations ext4_fast_symlink_inode_operations = { @@ -51,4 +52,5 @@ const struct inode_operations ext4_fast_ .listxattr = ext4_listxattr, .removexattr = generic_removexattr, #endif + .sync_flags = ext4_sync_flags, }; diff -NurpP --minimal linux-source-2.6.22.orig/fs/ext4/xattr.c linux-source-2.6.22/fs/ext4/xattr.c --- linux-source-2.6.22.orig/fs/ext4/xattr.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ext4/xattr.c 2007-10-24 16:14:06.291582822 +0200 @@ -58,6 +58,7 @@ #include #include #include +#include #include "xattr.h" #include "acl.h" @@ -496,6 +497,7 @@ ext4_xattr_release_block(handle_t *handl error = ext4_journal_dirty_metadata(handle, bh); if (IS_SYNC(inode)) handle->h_sync = 1; + DLIMIT_FREE_BLOCK(inode, 1); DQUOT_FREE_BLOCK(inode, 1); ea_bdebug(bh, "refcount now=%d; releasing", le32_to_cpu(BHDR(bh)->h_refcount)); @@ -769,11 +771,14 @@ inserted: if (new_bh == bs->bh) ea_bdebug(new_bh, "keeping"); else { + error = -ENOSPC; + if (DLIMIT_ALLOC_BLOCK(inode, 1)) + goto cleanup; /* The old block is released after updating the inode. */ error = -EDQUOT; if (DQUOT_ALLOC_BLOCK(inode, 1)) - goto cleanup; + goto cleanup_dlimit; error = ext4_journal_get_write_access(handle, new_bh); if (error) @@ -850,6 +855,8 @@ cleanup: cleanup_dquot: DQUOT_FREE_BLOCK(inode, 1); +cleanup_dlimit: + DLIMIT_FREE_BLOCK(inode, 1); goto cleanup; bad_block: diff -NurpP --minimal linux-source-2.6.22.orig/fs/fcntl.c linux-source-2.6.22/fs/fcntl.c --- linux-source-2.6.22.orig/fs/fcntl.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/fcntl.c 2007-10-24 16:14:06.291582822 +0200 @@ -18,6 +18,7 @@ #include #include #include +#include #include #include @@ -84,6 +85,8 @@ repeat: error = -EMFILE; if (newfd >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur) goto out; + if (!vx_files_avail(1)) + goto out; error = expand_files(files, newfd); if (error < 0) @@ -124,6 +127,7 @@ static int dupfd(struct file *file, unsi FD_SET(fd, fdt->open_fds); FD_CLR(fd, fdt->close_on_exec); spin_unlock(&files->file_lock); + vx_openfd_inc(fd); fd_install(fd, file); } else { spin_unlock(&files->file_lock); @@ -176,6 +180,9 @@ asmlinkage long sys_dup2(unsigned int ol if (tofree) filp_close(tofree, files); + else + vx_openfd_inc(newfd); /* fd was unused */ + err = newfd; out: return err; diff -NurpP --minimal linux-source-2.6.22.orig/fs/file_table.c linux-source-2.6.22/fs/file_table.c --- linux-source-2.6.22.orig/fs/file_table.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/file_table.c 2007-10-24 16:14:06.291582822 +0200 @@ -20,6 +20,8 @@ #include #include #include +#include +#include #include @@ -119,6 +121,8 @@ struct file *get_empty_filp(void) f->f_gid = tsk->fsgid; eventpoll_init_file(f); /* f->f_version: 0 */ + f->f_xid = vx_current_xid(); + vx_files_inc(f); return f; over: @@ -174,6 +178,8 @@ void fastcall __fput(struct file *file) if (file->f_mode & FMODE_WRITE) put_write_access(inode); put_pid(file->f_owner.pid); + vx_files_dec(file); + file->f_xid = 0; file_kill(file); file->f_path.dentry = NULL; file->f_path.mnt = NULL; @@ -239,6 +245,8 @@ void put_filp(struct file *file) { if (atomic_dec_and_test(&file->f_count)) { security_file_free(file); + vx_files_dec(file); + file->f_xid = 0; file_kill(file); file_free(file); } diff -NurpP --minimal linux-source-2.6.22.orig/fs/hfsplus/ioctl.c linux-source-2.6.22/fs/hfsplus/ioctl.c --- linux-source-2.6.22.orig/fs/hfsplus/ioctl.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/hfsplus/ioctl.c 2007-10-24 16:14:06.291582822 +0200 @@ -16,6 +16,7 @@ #include #include #include +#include #include #include "hfsplus_fs.h" @@ -35,7 +36,8 @@ int hfsplus_ioctl(struct inode *inode, s flags |= FS_NODUMP_FL; /* EXT2_NODUMP_FL */ return put_user(flags, (int __user *)arg); case HFSPLUS_IOC_EXT2_SETFLAGS: { - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER)) diff -NurpP --minimal linux-source-2.6.22.orig/fs/inode.c linux-source-2.6.22/fs/inode.c --- linux-source-2.6.22.orig/fs/inode.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/inode.c 2007-10-24 16:14:06.291582822 +0200 @@ -115,6 +115,9 @@ static struct inode *alloc_inode(struct struct address_space * const mapping = &inode->i_data; inode->i_sb = sb; + + /* essential because of inode slab reuse */ + inode->i_tag = 0; inode->i_blkbits = sb->s_blocksize_bits; inode->i_flags = 0; atomic_set(&inode->i_count, 1); @@ -231,6 +234,8 @@ void __iget(struct inode * inode) inodes_stat.nr_unused--; } +EXPORT_SYMBOL_GPL(__iget); + /** * clear_inode - clear an inode * @inode: inode to clear diff -NurpP --minimal linux-source-2.6.22.orig/fs/ioctl.c linux-source-2.6.22/fs/ioctl.c --- linux-source-2.6.22.orig/fs/ioctl.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ioctl.c 2007-10-24 16:14:06.291582822 +0200 @@ -13,10 +13,19 @@ #include #include #include +#include +#include +#include #include #include + +#ifdef CONFIG_VSERVER_LEGACY +extern int vx_proc_ioctl(struct inode *, struct file *, + unsigned int, unsigned long); +#endif + static long do_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) { @@ -152,6 +161,48 @@ int vfs_ioctl(struct file *filp, unsigne else error = -ENOTTY; break; +#ifdef CONFIG_VSERVER_LEGACY +#ifndef CONFIG_TAGGING_NONE + case FIOC_GETTAG: { + struct inode *inode = filp->f_dentry->d_inode; + + /* fixme: if stealth, return -ENOTTY */ + error = -EPERM; + if (capable(CAP_CONTEXT)) + error = put_user(inode->i_tag, (int __user *) arg); + break; + } + case FIOC_SETTAG: { + struct inode *inode = filp->f_dentry->d_inode; + int tag; + + /* fixme: if stealth, return -ENOTTY */ + error = -EPERM; + if (!capable(CAP_CONTEXT)) + break; + error = -EROFS; + if (IS_RDONLY(inode)) + break; + error = -ENOSYS; + if (!(inode->i_sb->s_flags & MS_TAGGED)) + break; + error = -EFAULT; + if (get_user(tag, (int __user *) arg)) + break; + error = 0; + inode->i_tag = (tag & 0xFFFF); + inode->i_ctime = CURRENT_TIME; + mark_inode_dirty(inode); + break; + } +#endif + case FIOC_GETXFLG: + case FIOC_SETXFLG: + error = -ENOTTY; + if (filp->f_dentry->d_inode->i_sb->s_magic == PROC_SUPER_MAGIC) + error = vx_proc_ioctl(filp->f_dentry->d_inode, filp, cmd, arg); + break; +#endif default: if (S_ISREG(filp->f_path.dentry->d_inode->i_mode)) error = file_ioctl(filp, cmd, arg); diff -NurpP --minimal linux-source-2.6.22.orig/fs/ioprio.c linux-source-2.6.22/fs/ioprio.c --- linux-source-2.6.22.orig/fs/ioprio.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ioprio.c 2007-10-24 16:14:06.291582822 +0200 @@ -25,6 +25,7 @@ #include #include #include +#include static int set_task_ioprio(struct task_struct *task, int ioprio) { @@ -103,6 +104,8 @@ asmlinkage long sys_ioprio_set(int which else pgrp = find_pid(who); do_each_pid_task(pgrp, PIDTYPE_PGID, p) { + if (!vx_check(p->xid, VS_ADMIN_P | VS_IDENT)) + continue; ret = set_task_ioprio(p, ioprio); if (ret) break; @@ -112,7 +115,7 @@ asmlinkage long sys_ioprio_set(int which if (!who) user = current->user; else - user = find_user(who); + user = find_user(vx_current_xid(), who); if (!user) break; @@ -190,6 +193,8 @@ asmlinkage long sys_ioprio_get(int which else pgrp = find_pid(who); do_each_pid_task(pgrp, PIDTYPE_PGID, p) { + if (!vx_check(p->xid, VS_ADMIN_P | VS_IDENT)) + continue; tmpio = get_task_ioprio(p); if (tmpio < 0) continue; @@ -203,7 +208,7 @@ asmlinkage long sys_ioprio_get(int which if (!who) user = current->user; else - user = find_user(who); + user = find_user(vx_current_xid(), who); if (!user) break; diff -NurpP --minimal linux-source-2.6.22.orig/fs/jfs/acl.c linux-source-2.6.22/fs/jfs/acl.c --- linux-source-2.6.22.orig/fs/jfs/acl.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/jfs/acl.c 2007-10-24 16:14:06.291582822 +0200 @@ -232,7 +232,8 @@ int jfs_setattr(struct dentry *dentry, s return rc; if ((iattr->ia_valid & ATTR_UID && iattr->ia_uid != inode->i_uid) || - (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid)) { + (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid) || + (iattr->ia_valid & ATTR_TAG && iattr->ia_tag != inode->i_tag)) { if (DQUOT_TRANSFER(inode, iattr)) return -EDQUOT; } diff -NurpP --minimal linux-source-2.6.22.orig/fs/jfs/file.c linux-source-2.6.22/fs/jfs/file.c --- linux-source-2.6.22.orig/fs/jfs/file.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/jfs/file.c 2007-10-24 16:14:06.291582822 +0200 @@ -98,6 +98,7 @@ const struct inode_operations jfs_file_i .setattr = jfs_setattr, .permission = jfs_permission, #endif + .sync_flags = jfs_sync_flags, }; const struct file_operations jfs_file_operations = { @@ -109,6 +110,7 @@ const struct file_operations jfs_file_op .aio_write = generic_file_aio_write, .mmap = generic_file_mmap, .sendfile = generic_file_sendfile, + .sendpage = generic_file_sendpage, .splice_read = generic_file_splice_read, .splice_write = generic_file_splice_write, .fsync = jfs_fsync, diff -NurpP --minimal linux-source-2.6.22.orig/fs/jfs/inode.c linux-source-2.6.22/fs/jfs/inode.c --- linux-source-2.6.22.orig/fs/jfs/inode.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/jfs/inode.c 2007-10-24 16:14:06.291582822 +0200 @@ -22,6 +22,7 @@ #include #include #include +#include #include "jfs_incore.h" #include "jfs_inode.h" #include "jfs_filsys.h" @@ -143,6 +144,7 @@ void jfs_delete_inode(struct inode *inod DQUOT_INIT(inode); DQUOT_FREE_INODE(inode); DQUOT_DROP(inode); + DLIMIT_FREE_INODE(inode); } clear_inode(inode); diff -NurpP --minimal linux-source-2.6.22.orig/fs/jfs/ioctl.c linux-source-2.6.22/fs/jfs/ioctl.c --- linux-source-2.6.22.orig/fs/jfs/ioctl.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/jfs/ioctl.c 2007-10-24 16:14:06.291582822 +0200 @@ -10,6 +10,7 @@ #include #include #include +#include #include #include @@ -66,7 +67,8 @@ int jfs_ioctl(struct inode * inode, stru case JFS_IOC_SETFLAGS: { unsigned int oldflags; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER)) @@ -87,8 +89,8 @@ int jfs_ioctl(struct inode * inode, stru * the relevant capability. */ if ((oldflags & JFS_IMMUTABLE_FL) || - ((flags ^ oldflags) & - (JFS_APPEND_FL | JFS_IMMUTABLE_FL))) { + ((flags ^ oldflags) & (JFS_APPEND_FL | + JFS_IMMUTABLE_FL | JFS_IUNLINK_FL))) { if (!capable(CAP_LINUX_IMMUTABLE)) return -EPERM; } diff -NurpP --minimal linux-source-2.6.22.orig/fs/jfs/jfs_dinode.h linux-source-2.6.22/fs/jfs/jfs_dinode.h --- linux-source-2.6.22.orig/fs/jfs/jfs_dinode.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/jfs/jfs_dinode.h 2007-10-24 16:14:06.291582822 +0200 @@ -162,9 +162,12 @@ struct dinode { #define JFS_APPEND_FL 0x01000000 /* writes to file may only append */ #define JFS_IMMUTABLE_FL 0x02000000 /* Immutable file */ -#define JFS_FL_USER_VISIBLE 0x03F80000 +#define JFS_BARRIER_FL 0x04000000 /* Barrier for chroot() */ +#define JFS_IUNLINK_FL 0x08000000 /* Immutable unlink */ + +#define JFS_FL_USER_VISIBLE 0x0FF80000 #define JFS_FL_USER_MODIFIABLE 0x03F80000 -#define JFS_FL_INHERIT 0x03C80000 +#define JFS_FL_INHERIT 0x0BC80000 /* These are identical to EXT[23]_IOC_GETFLAGS/SETFLAGS */ #define JFS_IOC_GETFLAGS _IOR('f', 1, long) diff -NurpP --minimal linux-source-2.6.22.orig/fs/jfs/jfs_dtree.c linux-source-2.6.22/fs/jfs/jfs_dtree.c --- linux-source-2.6.22.orig/fs/jfs/jfs_dtree.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/jfs/jfs_dtree.c 2007-10-24 16:14:06.291582822 +0200 @@ -102,6 +102,7 @@ #include #include +#include #include "jfs_incore.h" #include "jfs_superblock.h" #include "jfs_filsys.h" @@ -383,10 +384,10 @@ static u32 add_index(tid_t tid, struct i */ if (DQUOT_ALLOC_BLOCK(ip, sbi->nbperpage)) goto clean_up; - if (dbAlloc(ip, 0, sbi->nbperpage, &xaddr)) { - DQUOT_FREE_BLOCK(ip, sbi->nbperpage); - goto clean_up; - } + if (DLIMIT_ALLOC_BLOCK(ip, sbi->nbperpage)) + goto clean_up_dquot; + if (dbAlloc(ip, 0, sbi->nbperpage, &xaddr)) + goto clean_up_dlimit; /* * Save the table, we're going to overwrite it with the @@ -479,6 +480,12 @@ static u32 add_index(tid_t tid, struct i return index; + clean_up_dlimit: + DLIMIT_FREE_BLOCK(ip, sbi->nbperpage); + + clean_up_dquot: + DQUOT_FREE_BLOCK(ip, sbi->nbperpage); + clean_up: jfs_ip->next_index--; @@ -952,6 +959,7 @@ static int dtSplitUp(tid_t tid, struct tlock *tlck; struct lv *lv; int quota_allocation = 0; + int dlimit_allocation = 0; /* get split page */ smp = split->mp; @@ -1036,6 +1044,12 @@ static int dtSplitUp(tid_t tid, } quota_allocation += n; + if (DLIMIT_ALLOC_BLOCK(ip, n)) { + rc = -ENOSPC; + goto extendOut; + } + dlimit_allocation += n; + if ((rc = dbReAlloc(sbi->ipbmap, xaddr, (s64) xlen, (s64) n, &nxaddr))) goto extendOut; @@ -1309,6 +1323,9 @@ static int dtSplitUp(tid_t tid, freeKeyName: kfree(key.name); + /* Rollback dlimit allocation */ + if (rc && dlimit_allocation) + DLIMIT_FREE_BLOCK(ip, dlimit_allocation); /* Rollback quota allocation */ if (rc && quota_allocation) DQUOT_FREE_BLOCK(ip, quota_allocation); @@ -1376,6 +1393,12 @@ static int dtSplitPage(tid_t tid, struct release_metapage(rmp); return -EDQUOT; } + /* Allocate blocks to dlimit. */ + if (DLIMIT_ALLOC_BLOCK(ip, lengthPXD(pxd))) { + DQUOT_FREE_BLOCK(ip, lengthPXD(pxd)); + release_metapage(rmp); + return -ENOSPC; + } jfs_info("dtSplitPage: ip:0x%p smp:0x%p rmp:0x%p", ip, smp, rmp); @@ -1926,6 +1949,12 @@ static int dtSplitRoot(tid_t tid, release_metapage(rmp); return -EDQUOT; } + /* Allocate blocks to dlimit. */ + if (DLIMIT_ALLOC_BLOCK(ip, lengthPXD(pxd))) { + DQUOT_FREE_BLOCK(ip, lengthPXD(pxd)); + release_metapage(rmp); + return -ENOSPC; + } BT_MARK_DIRTY(rmp, ip); /* @@ -2292,6 +2321,8 @@ static int dtDeleteUp(tid_t tid, struct xlen = lengthPXD(&fp->header.self); + /* Free dlimit allocation. */ + DLIMIT_FREE_BLOCK(ip, xlen); /* Free quota allocation. */ DQUOT_FREE_BLOCK(ip, xlen); @@ -2368,6 +2399,8 @@ static int dtDeleteUp(tid_t tid, struct xlen = lengthPXD(&p->header.self); + /* Free dlimit allocation */ + DLIMIT_FREE_BLOCK(ip, xlen); /* Free quota allocation */ DQUOT_FREE_BLOCK(ip, xlen); diff -NurpP --minimal linux-source-2.6.22.orig/fs/jfs/jfs_extent.c linux-source-2.6.22/fs/jfs/jfs_extent.c --- linux-source-2.6.22.orig/fs/jfs/jfs_extent.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/jfs/jfs_extent.c 2007-10-24 16:14:06.291582822 +0200 @@ -18,6 +18,7 @@ #include #include +#include #include "jfs_incore.h" #include "jfs_inode.h" #include "jfs_superblock.h" @@ -147,6 +148,14 @@ extAlloc(struct inode *ip, s64 xlen, s64 return -EDQUOT; } + /* Allocate blocks to dlimit. */ + if (DLIMIT_ALLOC_BLOCK(ip, nxlen)) { + DQUOT_FREE_BLOCK(ip, nxlen); + dbFree(ip, nxaddr, (s64) nxlen); + mutex_unlock(&JFS_IP(ip)->commit_mutex); + return -ENOSPC; + } + /* determine the value of the extent flag */ xflag = abnr ? XAD_NOTRECORDED : 0; @@ -164,6 +173,7 @@ extAlloc(struct inode *ip, s64 xlen, s64 */ if (rc) { dbFree(ip, nxaddr, nxlen); + DLIMIT_FREE_BLOCK(ip, nxlen); DQUOT_FREE_BLOCK(ip, nxlen); mutex_unlock(&JFS_IP(ip)->commit_mutex); return (rc); @@ -261,6 +271,13 @@ int extRealloc(struct inode *ip, s64 nxl mutex_unlock(&JFS_IP(ip)->commit_mutex); return -EDQUOT; } + /* Allocate blocks to dlimit. */ + if (DLIMIT_ALLOC_BLOCK(ip, nxlen)) { + DQUOT_FREE_BLOCK(ip, nxlen); + dbFree(ip, nxaddr, (s64) nxlen); + up(&JFS_IP(ip)->commit_sem); + return -ENOSPC; + } delta = nxlen - xlen; @@ -297,6 +314,7 @@ int extRealloc(struct inode *ip, s64 nxl /* extend the extent */ if ((rc = xtExtend(0, ip, xoff + xlen, (int) nextend, 0))) { dbFree(ip, xaddr + xlen, delta); + DLIMIT_FREE_BLOCK(ip, nxlen); DQUOT_FREE_BLOCK(ip, nxlen); goto exit; } @@ -308,6 +326,7 @@ int extRealloc(struct inode *ip, s64 nxl */ if ((rc = xtTailgate(0, ip, xoff, (int) ntail, nxaddr, 0))) { dbFree(ip, nxaddr, nxlen); + DLIMIT_FREE_BLOCK(ip, nxlen); DQUOT_FREE_BLOCK(ip, nxlen); goto exit; } diff -NurpP --minimal linux-source-2.6.22.orig/fs/jfs/jfs_filsys.h linux-source-2.6.22/fs/jfs/jfs_filsys.h --- linux-source-2.6.22.orig/fs/jfs/jfs_filsys.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/jfs/jfs_filsys.h 2007-10-24 16:14:06.291582822 +0200 @@ -264,6 +264,7 @@ #define JFS_NAME_MAX 255 #define JFS_PATH_MAX BPSIZE +#define JFS_TAGGED 0x00800000 /* Context Tagging */ /* * file system state (superblock state) diff -NurpP --minimal linux-source-2.6.22.orig/fs/jfs/jfs_imap.c linux-source-2.6.22/fs/jfs/jfs_imap.c --- linux-source-2.6.22.orig/fs/jfs/jfs_imap.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/jfs/jfs_imap.c 2007-10-24 16:14:06.291582822 +0200 @@ -45,6 +45,7 @@ #include #include #include +#include #include "jfs_incore.h" #include "jfs_inode.h" @@ -3075,6 +3076,8 @@ static int copy_from_dinode(struct dinod { struct jfs_inode_info *jfs_ip = JFS_IP(ip); struct jfs_sb_info *sbi = JFS_SBI(ip->i_sb); + uid_t uid; + gid_t gid; jfs_ip->fileset = le32_to_cpu(dip->di_fileset); jfs_ip->mode2 = le32_to_cpu(dip->di_mode); @@ -3095,14 +3098,18 @@ static int copy_from_dinode(struct dinod } ip->i_nlink = le32_to_cpu(dip->di_nlink); - jfs_ip->saved_uid = le32_to_cpu(dip->di_uid); + uid = le32_to_cpu(dip->di_uid); + gid = le32_to_cpu(dip->di_gid); + ip->i_tag = INOTAG_TAG(DX_TAG(ip), uid, gid, 0); + + jfs_ip->saved_uid = INOTAG_UID(DX_TAG(ip), uid, gid); if (sbi->uid == -1) ip->i_uid = jfs_ip->saved_uid; else { ip->i_uid = sbi->uid; } - jfs_ip->saved_gid = le32_to_cpu(dip->di_gid); + jfs_ip->saved_gid = INOTAG_GID(DX_TAG(ip), uid, gid); if (sbi->gid == -1) ip->i_gid = jfs_ip->saved_gid; else { @@ -3167,14 +3174,12 @@ static void copy_to_dinode(struct dinode dip->di_size = cpu_to_le64(ip->i_size); dip->di_nblocks = cpu_to_le64(PBLK2LBLK(ip->i_sb, ip->i_blocks)); dip->di_nlink = cpu_to_le32(ip->i_nlink); - if (sbi->uid == -1) - dip->di_uid = cpu_to_le32(ip->i_uid); - else - dip->di_uid = cpu_to_le32(jfs_ip->saved_uid); - if (sbi->gid == -1) - dip->di_gid = cpu_to_le32(ip->i_gid); - else - dip->di_gid = cpu_to_le32(jfs_ip->saved_gid); + + dip->di_uid = cpu_to_le32(TAGINO_UID(DX_TAG(ip), + (sbi->uid == -1) ? ip->i_uid : jfs_ip->saved_uid, ip->i_tag)); + dip->di_gid = cpu_to_le32(TAGINO_GID(DX_TAG(ip), + (sbi->gid == -1) ? ip->i_gid : jfs_ip->saved_gid, ip->i_tag)); + jfs_get_inode_flags(jfs_ip); /* * mode2 is only needed for storing the higher order bits. diff -NurpP --minimal linux-source-2.6.22.orig/fs/jfs/jfs_inode.c linux-source-2.6.22/fs/jfs/jfs_inode.c --- linux-source-2.6.22.orig/fs/jfs/jfs_inode.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/jfs/jfs_inode.c 2007-10-24 16:14:06.291582822 +0200 @@ -18,6 +18,8 @@ #include #include +#include +#include #include "jfs_incore.h" #include "jfs_inode.h" #include "jfs_filsys.h" @@ -30,19 +32,47 @@ void jfs_set_inode_flags(struct inode *i { unsigned int flags = JFS_IP(inode)->mode2; - inode->i_flags &= ~(S_IMMUTABLE | S_APPEND | - S_NOATIME | S_DIRSYNC | S_SYNC); + inode->i_flags &= ~(S_IMMUTABLE | S_IUNLINK | S_BARRIER | + S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC); if (flags & JFS_IMMUTABLE_FL) inode->i_flags |= S_IMMUTABLE; + if (flags & JFS_IUNLINK_FL) + inode->i_flags |= S_IUNLINK; + if (flags & JFS_BARRIER_FL) + inode->i_flags |= S_BARRIER; + + if (flags & JFS_SYNC_FL) + inode->i_flags |= S_SYNC; if (flags & JFS_APPEND_FL) inode->i_flags |= S_APPEND; if (flags & JFS_NOATIME_FL) inode->i_flags |= S_NOATIME; if (flags & JFS_DIRSYNC_FL) inode->i_flags |= S_DIRSYNC; - if (flags & JFS_SYNC_FL) - inode->i_flags |= S_SYNC; +} + +int jfs_sync_flags(struct inode *inode) +{ + unsigned int oldflags, newflags; + + oldflags = JFS_IP(inode)->mode2; + newflags = oldflags & ~(JFS_IMMUTABLE_FL | + JFS_IUNLINK_FL | JFS_BARRIER_FL); + + if (IS_IMMUTABLE(inode)) + newflags |= JFS_IMMUTABLE_FL; + if (IS_IUNLINK(inode)) + newflags |= JFS_IUNLINK_FL; + if (IS_BARRIER(inode)) + newflags |= JFS_BARRIER_FL; + + if (oldflags ^ newflags) { + JFS_IP(inode)->mode2 = newflags; + inode->i_ctime = CURRENT_TIME; + mark_inode_dirty(inode); + } + return 0; } void jfs_get_inode_flags(struct jfs_inode_info *jfs_ip) @@ -108,10 +138,17 @@ struct inode *ialloc(struct inode *paren jfs_inode->saved_uid = inode->i_uid; jfs_inode->saved_gid = inode->i_gid; + inode->i_tag = dx_current_fstag(sb); + if (DLIMIT_ALLOC_INODE(inode)) { + iput(inode); + return ERR_PTR(-ENOSPC); + } + /* * Allocate inode to quota. */ if (DQUOT_ALLOC_INODE(inode)) { + DLIMIT_FREE_INODE(inode); DQUOT_DROP(inode); inode->i_flags |= S_NOQUOTA; inode->i_nlink = 0; diff -NurpP --minimal linux-source-2.6.22.orig/fs/jfs/jfs_inode.h linux-source-2.6.22/fs/jfs/jfs_inode.h --- linux-source-2.6.22.orig/fs/jfs/jfs_inode.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/jfs/jfs_inode.h 2007-10-24 16:14:06.291582822 +0200 @@ -33,6 +33,7 @@ extern void jfs_free_zero_link(struct in extern struct dentry *jfs_get_parent(struct dentry *dentry); extern void jfs_get_inode_flags(struct jfs_inode_info *); extern void jfs_set_inode_flags(struct inode *); +extern int jfs_sync_flags(struct inode *); extern int jfs_get_block(struct inode *, sector_t, struct buffer_head *, int); extern const struct address_space_operations jfs_aops; diff -NurpP --minimal linux-source-2.6.22.orig/fs/jfs/jfs_xtree.c linux-source-2.6.22/fs/jfs/jfs_xtree.c --- linux-source-2.6.22.orig/fs/jfs/jfs_xtree.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/jfs/jfs_xtree.c 2007-10-24 16:14:06.291582822 +0200 @@ -21,6 +21,7 @@ #include #include +#include #include "jfs_incore.h" #include "jfs_filsys.h" #include "jfs_metapage.h" @@ -846,7 +847,12 @@ int xtInsert(tid_t tid, /* transaction hint = 0; if ((rc = DQUOT_ALLOC_BLOCK(ip, xlen))) goto out; + if ((rc = DLIMIT_ALLOC_BLOCK(ip, xlen))) { + DQUOT_FREE_BLOCK(ip, xlen); + goto out; + } if ((rc = dbAlloc(ip, hint, (s64) xlen, &xaddr))) { + DLIMIT_FREE_BLOCK(ip, xlen); DQUOT_FREE_BLOCK(ip, xlen); goto out; } @@ -876,6 +882,7 @@ int xtInsert(tid_t tid, /* transaction /* undo data extent allocation */ if (*xaddrp == 0) { dbFree(ip, xaddr, (s64) xlen); + DLIMIT_FREE_BLOCK(ip, xlen); DQUOT_FREE_BLOCK(ip, xlen); } return rc; @@ -1236,6 +1243,7 @@ xtSplitPage(tid_t tid, struct inode *ip, struct tlock *tlck; struct xtlock *sxtlck = NULL, *rxtlck = NULL; int quota_allocation = 0; + int dlimit_allocation = 0; smp = split->mp; sp = XT_PAGE(ip, smp); @@ -1255,6 +1263,13 @@ xtSplitPage(tid_t tid, struct inode *ip, quota_allocation += lengthPXD(pxd); + /* Allocate blocks to dlimit. */ + if (DLIMIT_ALLOC_BLOCK(ip, lengthPXD(pxd))) { + rc = -ENOSPC; + goto clean_up; + } + dlimit_allocation += lengthPXD(pxd); + /* * allocate the new right page for the split */ @@ -1456,6 +1471,9 @@ xtSplitPage(tid_t tid, struct inode *ip, clean_up: + /* Rollback dlimit allocation. */ + if (dlimit_allocation) + DLIMIT_FREE_BLOCK(ip, dlimit_allocation); /* Rollback quota allocation. */ if (quota_allocation) DQUOT_FREE_BLOCK(ip, quota_allocation); @@ -1520,6 +1538,12 @@ xtSplitRoot(tid_t tid, release_metapage(rmp); return -EDQUOT; } + /* Allocate blocks to dlimit. */ + if (DLIMIT_ALLOC_BLOCK(ip, lengthPXD(pxd))) { + DQUOT_FREE_BLOCK(ip, lengthPXD(pxd)); + release_metapage(rmp); + return -ENOSPC; + } jfs_info("xtSplitRoot: ip:0x%p rmp:0x%p", ip, rmp); @@ -3951,6 +3975,8 @@ s64 xtTruncate(tid_t tid, struct inode * else ip->i_size = newsize; + /* update dlimit allocation to reflect freed blocks */ + DLIMIT_FREE_BLOCK(ip, nfreed); /* update quota allocation to reflect freed blocks */ DQUOT_FREE_BLOCK(ip, nfreed); diff -NurpP --minimal linux-source-2.6.22.orig/fs/jfs/namei.c linux-source-2.6.22/fs/jfs/namei.c --- linux-source-2.6.22.orig/fs/jfs/namei.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/jfs/namei.c 2007-10-24 16:14:06.291582822 +0200 @@ -20,6 +20,7 @@ #include #include #include +#include #include "jfs_incore.h" #include "jfs_superblock.h" #include "jfs_inode.h" @@ -1469,6 +1470,7 @@ static struct dentry *jfs_lookup(struct return ERR_PTR(-EACCES); } + dx_propagate_tag(nd, ip); dentry = d_splice_alias(ip, dentry); if (dentry && (JFS_SBI(dip->i_sb)->mntflag & JFS_OS2)) @@ -1521,6 +1523,7 @@ const struct inode_operations jfs_dir_in .setattr = jfs_setattr, .permission = jfs_permission, #endif + .sync_flags = jfs_sync_flags, }; const struct file_operations jfs_dir_operations = { diff -NurpP --minimal linux-source-2.6.22.orig/fs/jfs/super.c linux-source-2.6.22/fs/jfs/super.c --- linux-source-2.6.22.orig/fs/jfs/super.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/jfs/super.c 2007-10-24 16:14:06.291582822 +0200 @@ -194,7 +194,8 @@ static void jfs_put_super(struct super_b enum { Opt_integrity, Opt_nointegrity, Opt_iocharset, Opt_resize, Opt_resize_nosize, Opt_errors, Opt_ignore, Opt_err, Opt_quota, - Opt_usrquota, Opt_grpquota, Opt_uid, Opt_gid, Opt_umask + Opt_usrquota, Opt_grpquota, Opt_uid, Opt_gid, Opt_umask, + Opt_tag, Opt_notag, Opt_tagid }; static match_table_t tokens = { @@ -204,6 +205,10 @@ static match_table_t tokens = { {Opt_resize, "resize=%u"}, {Opt_resize_nosize, "resize"}, {Opt_errors, "errors=%s"}, + {Opt_tag, "tag"}, + {Opt_notag, "notag"}, + {Opt_tagid, "tagid=%u"}, + {Opt_tag, "tagxid"}, {Opt_ignore, "noquota"}, {Opt_ignore, "quota"}, {Opt_usrquota, "usrquota"}, @@ -338,6 +343,20 @@ static int parse_options(char *options, } break; } +#ifndef CONFIG_TAGGING_NONE + case Opt_tag: + *flag |= JFS_TAGGED; + break; + case Opt_notag: + *flag &= JFS_TAGGED; + break; +#endif +#ifdef CONFIG_PROPAGATE + case Opt_tagid: + /* use args[0] */ + *flag |= JFS_TAGGED; + break; +#endif default: printk("jfs: Unrecognized mount option \"%s\" " " or missing value\n", p); @@ -368,6 +387,13 @@ static int jfs_remount(struct super_bloc if (!parse_options(data, sb, &newLVSize, &flag)) { return -EINVAL; } + + if ((flag & JFS_TAGGED) && !(sb->s_flags & MS_TAGGED)) { + printk(KERN_ERR "JFS: %s: tagging not permitted on remount.\n", + sb->s_id); + return -EINVAL; + } + if (newLVSize) { if (sb->s_flags & MS_RDONLY) { printk(KERN_ERR @@ -439,6 +465,9 @@ static int jfs_fill_super(struct super_b #ifdef CONFIG_JFS_POSIX_ACL sb->s_flags |= MS_POSIXACL; #endif + /* map mount option tagxid */ + if (sbi->flag & JFS_TAGGED) + sb->s_flags |= MS_TAGGED; if (newLVSize) { printk(KERN_ERR "resize option for remount only\n"); diff -NurpP --minimal linux-source-2.6.22.orig/fs/jfs/xattr.c linux-source-2.6.22/fs/jfs/xattr.c --- linux-source-2.6.22.orig/fs/jfs/xattr.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/jfs/xattr.c 2007-10-24 16:14:06.291582822 +0200 @@ -23,6 +23,7 @@ #include #include #include +#include #include "jfs_incore.h" #include "jfs_superblock.h" #include "jfs_dmap.h" @@ -263,9 +264,16 @@ static int ea_write(struct inode *ip, st if (DQUOT_ALLOC_BLOCK(ip, nblocks)) { return -EDQUOT; } + /* Allocate new blocks to dlimit. */ + if (DLIMIT_ALLOC_BLOCK(ip, nblocks)) { + DQUOT_FREE_BLOCK(ip, nblocks); + return -ENOSPC; + } rc = dbAlloc(ip, INOHINT(ip), nblocks, &blkno); if (rc) { + /*Rollback dlimit allocation. */ + DLIMIT_FREE_BLOCK(ip, nblocks); /*Rollback quota allocation. */ DQUOT_FREE_BLOCK(ip, nblocks); return rc; @@ -332,6 +340,8 @@ static int ea_write(struct inode *ip, st failed: /* Rollback quota allocation. */ + DLIMIT_FREE_BLOCK(ip, nblocks); + /* Rollback quota allocation. */ DQUOT_FREE_BLOCK(ip, nblocks); dbFree(ip, blkno, nblocks); @@ -468,6 +478,7 @@ static int ea_get(struct inode *inode, s s64 blkno; int rc; int quota_allocation = 0; + int dlimit_allocation = 0; /* When fsck.jfs clears a bad ea, it doesn't clear the size */ if (ji->ea.flag == 0) @@ -543,6 +554,12 @@ static int ea_get(struct inode *inode, s quota_allocation = blocks_needed; + /* Allocate new blocks to dlimit. */ + rc = -ENOSPC; + if (DLIMIT_ALLOC_BLOCK(inode, blocks_needed)) + goto clean_up; + dlimit_allocation = blocks_needed; + rc = dbAlloc(inode, INOHINT(inode), (s64) blocks_needed, &blkno); if (rc) @@ -599,6 +616,9 @@ static int ea_get(struct inode *inode, s return ea_size; clean_up: + /* Rollback dlimit allocation */ + if (dlimit_allocation) + DLIMIT_FREE_BLOCK(inode, dlimit_allocation); /* Rollback quota allocation */ if (quota_allocation) DQUOT_FREE_BLOCK(inode, quota_allocation); @@ -675,8 +695,10 @@ static int ea_put(tid_t tid, struct inod } /* If old blocks exist, they must be removed from quota allocation. */ - if (old_blocks) + if (old_blocks) { + DLIMIT_FREE_BLOCK(inode, old_blocks); DQUOT_FREE_BLOCK(inode, old_blocks); + } inode->i_ctime = CURRENT_TIME; diff -NurpP --minimal linux-source-2.6.22.orig/fs/libfs.c linux-source-2.6.22/fs/libfs.c --- linux-source-2.6.22.orig/fs/libfs.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/libfs.c 2007-10-24 16:14:06.291582822 +0200 @@ -124,7 +124,8 @@ static inline unsigned char dt_type(stru * both impossible due to the lock on directory. */ -int dcache_readdir(struct file * filp, void * dirent, filldir_t filldir) +static inline int do_dcache_readdir_filter(struct file *filp, + void *dirent, filldir_t filldir, int (*filter)(struct dentry *dentry)) { struct dentry *dentry = filp->f_path.dentry; struct dentry *cursor = filp->private_data; @@ -157,6 +158,8 @@ int dcache_readdir(struct file * filp, v next = list_entry(p, struct dentry, d_u.d_child); if (d_unhashed(next) || !next->d_inode) continue; + if (filter && !filter(next)) + continue; spin_unlock(&dcache_lock); if (filldir(dirent, next->d_name.name, @@ -175,6 +178,18 @@ int dcache_readdir(struct file * filp, v return 0; } +int dcache_readdir(struct file *filp, void *dirent, filldir_t filldir) +{ + return do_dcache_readdir_filter(filp, dirent, filldir, NULL); +} + +int dcache_readdir_filter(struct file *filp, void *dirent, filldir_t filldir, + int (*filter)(struct dentry *)) +{ + return do_dcache_readdir_filter(filp, dirent, filldir, filter); +} + + ssize_t generic_read_dir(struct file *filp, char __user *buf, size_t siz, loff_t *ppos) { return -EISDIR; @@ -640,6 +655,7 @@ EXPORT_SYMBOL(dcache_dir_close); EXPORT_SYMBOL(dcache_dir_lseek); EXPORT_SYMBOL(dcache_dir_open); EXPORT_SYMBOL(dcache_readdir); +EXPORT_SYMBOL(dcache_readdir_filter); EXPORT_SYMBOL(generic_read_dir); EXPORT_SYMBOL(get_sb_pseudo); EXPORT_SYMBOL(simple_commit_write); diff -NurpP --minimal linux-source-2.6.22.orig/fs/locks.c linux-source-2.6.22/fs/locks.c --- linux-source-2.6.22.orig/fs/locks.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/fs/locks.c 2007-10-24 16:14:06.291582822 +0200 @@ -125,6 +125,8 @@ #include #include #include +#include +#include #include #include @@ -147,6 +149,8 @@ static struct kmem_cache *filelock_cache /* Allocate an empty lock structure. */ static struct file_lock *locks_alloc_lock(void) { + if (!vx_locks_avail(1)) + return NULL; return kmem_cache_alloc(filelock_cache, GFP_KERNEL); } @@ -172,6 +176,7 @@ static void locks_free_lock(struct file_ BUG_ON(!list_empty(&fl->fl_block)); BUG_ON(!list_empty(&fl->fl_link)); + vx_locks_dec(fl); locks_release_private(fl); kmem_cache_free(filelock_cache, fl); } @@ -191,6 +196,7 @@ void locks_init_lock(struct file_lock *f fl->fl_start = fl->fl_end = 0; fl->fl_ops = NULL; fl->fl_lmops = NULL; + fl->fl_xid = -1; } EXPORT_SYMBOL(locks_init_lock); @@ -244,6 +250,7 @@ void locks_copy_lock(struct file_lock *n new->fl_file = fl->fl_file; new->fl_ops = fl->fl_ops; new->fl_lmops = fl->fl_lmops; + new->fl_xid = fl->fl_xid; locks_copy_private(new, fl); } @@ -282,6 +289,11 @@ static int flock_make_lock(struct file * fl->fl_flags = FL_FLOCK; fl->fl_type = type; fl->fl_end = OFFSET_MAX; + + vxd_assert(filp->f_xid == vx_current_xid(), + "f_xid(%d) == current(%d)", filp->f_xid, vx_current_xid()); + fl->fl_xid = filp->f_xid; + vx_locks_inc(fl); *lock = fl; return 0; @@ -447,6 +459,7 @@ static int lease_init(struct file *filp, fl->fl_owner = current->files; fl->fl_pid = current->tgid; + fl->fl_xid = vx_current_xid(); fl->fl_file = filp; fl->fl_flags = FL_LEASE; @@ -466,6 +479,11 @@ static int lease_alloc(struct file *filp if (fl == NULL) goto out; + fl->fl_xid = vx_current_xid(); + if (filp) + vxd_assert(filp->f_xid == fl->fl_xid, + "f_xid(%d) == fl_xid(%d)", filp->f_xid, fl->fl_xid); + vx_locks_inc(fl); error = lease_init(filp, type, fl); if (error) { locks_free_lock(fl); @@ -769,6 +787,7 @@ static int flock_lock_file(struct file * if (found) cond_resched(); + new_fl->fl_xid = -1; find_conflict: for_each_lock(inode, before) { struct file_lock *fl = *before; @@ -787,6 +806,7 @@ find_conflict: goto out; locks_copy_lock(new_fl, request); locks_insert_lock(before, new_fl); + vx_locks_inc(new_fl); new_fl = NULL; error = 0; @@ -797,7 +817,8 @@ out: return error; } -static int __posix_lock_file(struct inode *inode, struct file_lock *request, struct file_lock *conflock) +static int __posix_lock_file(struct inode *inode, struct file_lock *request, + struct file_lock *conflock, xid_t xid) { struct file_lock *fl; struct file_lock *new_fl = NULL; @@ -807,6 +828,8 @@ static int __posix_lock_file(struct inod struct file_lock **before; int error, added = 0; + vxd_assert(xid == vx_current_xid(), + "xid(%d) == current(%d)", xid, vx_current_xid()); /* * We may need two file_lock structures for this operation, * so we get them in advance to avoid races. @@ -817,7 +840,11 @@ static int __posix_lock_file(struct inod (request->fl_type != F_UNLCK || request->fl_start != 0 || request->fl_end != OFFSET_MAX)) { new_fl = locks_alloc_lock(); + new_fl->fl_xid = xid; + vx_locks_inc(new_fl); new_fl2 = locks_alloc_lock(); + new_fl2->fl_xid = xid; + vx_locks_inc(new_fl2); } lock_kernel(); @@ -1016,7 +1043,8 @@ static int __posix_lock_file(struct inod int posix_lock_file(struct file *filp, struct file_lock *fl, struct file_lock *conflock) { - return __posix_lock_file(filp->f_path.dentry->d_inode, fl, conflock); + return __posix_lock_file(filp->f_path.dentry->d_inode, + fl, conflock, filp->f_xid); } EXPORT_SYMBOL(posix_lock_file); @@ -1106,7 +1134,7 @@ int locks_mandatory_area(int read_write, fl.fl_end = offset + count - 1; for (;;) { - error = __posix_lock_file(inode, &fl, NULL); + error = __posix_lock_file(inode, &fl, NULL, filp->f_xid); if (error != -EAGAIN) break; if (!(fl.fl_flags & FL_SLEEP)) @@ -1410,8 +1438,8 @@ static int __setlease(struct file *filp, goto out; locks_copy_lock(fl, lease); - locks_insert_lock(before, fl); + vx_locks_inc(fl); *flp = fl; error = 0; @@ -1738,6 +1766,11 @@ int fcntl_setlk(unsigned int fd, struct if (file_lock == NULL) return -ENOLCK; + vxd_assert(filp->f_xid == vx_current_xid(), + "f_xid(%d) == current(%d)", filp->f_xid, vx_current_xid()); + file_lock->fl_xid = filp->f_xid; + vx_locks_inc(file_lock); + /* * This might block, so we do it before checking the inode. */ @@ -1864,6 +1897,11 @@ int fcntl_setlk64(unsigned int fd, struc if (file_lock == NULL) return -ENOLCK; + vxd_assert(filp->f_xid == vx_current_xid(), + "f_xid(%d) == current(%d)", filp->f_xid, vx_current_xid()); + file_lock->fl_xid = filp->f_xid; + vx_locks_inc(file_lock); + /* * This might block, so we do it before checking the inode. */ @@ -2168,6 +2206,10 @@ int get_locks_status(char *buffer, char list_for_each(tmp, &file_lock_list) { struct list_head *btmp; struct file_lock *fl = list_entry(tmp, struct file_lock, fl_link); + + if (!vx_check(fl->fl_xid, VS_WATCH_P | VS_IDENT)) + continue; + lock_get_status(q, fl, ++i, ""); move_lock_status(&q, &pos, offset); diff -NurpP --minimal linux-source-2.6.22.orig/fs/namei.c linux-source-2.6.22/fs/namei.c --- linux-source-2.6.22.orig/fs/namei.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/fs/namei.c 2007-10-24 16:14:06.291582822 +0200 @@ -31,6 +31,11 @@ #include #include #include +#include +#include +#include +#include +#include #include #include @@ -224,6 +229,31 @@ int generic_permission(struct inode *ino return -EACCES; } +static inline int dx_barrier(struct inode *inode) +{ + if (IS_BARRIER(inode) && !vx_check(0, VS_ADMIN)) { + vxwprintk(1, "xid=%d did hit the barrier.", + vx_current_xid()); + return 1; + } + return 0; +} + +static inline int dx_permission(struct inode *inode, int mask, struct nameidata *nd) +{ + if (dx_barrier(inode)) + return -EACCES; + if (inode->i_tag == 0) + return 0; + if (dx_check(inode->i_tag, DX_ADMIN|DX_WATCH|DX_IDENT)) + return 0; + + vxwprintk(1, "xid=%d denied access to %p[#%d,%lu] »%s«.", + vx_current_xid(), inode, inode->i_tag, inode->i_ino, + vxd_cond_path(nd)); + return -EACCES; +} + int permission(struct inode *inode, int mask, struct nameidata *nd) { umode_t mode = inode->i_mode; @@ -234,14 +264,14 @@ int permission(struct inode *inode, int /* * Nobody gets write access to a read-only fs. */ - if (IS_RDONLY(inode) && + if ((IS_RDONLY(inode) || (nd && MNT_IS_RDONLY(nd->mnt))) && (S_ISREG(mode) || S_ISDIR(mode) || S_ISLNK(mode))) return -EROFS; /* * Nobody gets write access to an immutable file. */ - if (IS_IMMUTABLE(inode)) + if (IS_IMMUTABLE(inode) && !IS_COW(inode)) return -EACCES; } @@ -257,6 +287,8 @@ int permission(struct inode *inode, int /* Ordinary permission routines do not understand MAY_APPEND. */ submask = mask & ~MAY_APPEND; + if ((retval = dx_permission(inode, mask, nd))) + return retval; if (inode->i_op && inode->i_op->permission) retval = inode->i_op->permission(inode, submask, nd); else @@ -438,6 +470,8 @@ static int exec_permission_lite(struct i { umode_t mode = inode->i_mode; + if (dx_barrier(inode)) + return -EACCES; if (inode->i_op && inode->i_op->permission) return -EAGAIN; @@ -738,7 +772,8 @@ static __always_inline void follow_dotdo if (nd->dentry == fs->root && nd->mnt == fs->rootmnt) { read_unlock(&fs->lock); - break; + /* for sane '/' avoid follow_mount() */ + return; } read_unlock(&fs->lock); spin_lock(&dcache_lock); @@ -775,16 +810,34 @@ static int do_lookup(struct nameidata *n { struct vfsmount *mnt = nd->mnt; struct dentry *dentry = __d_lookup(nd->dentry, name); + struct inode *inode; if (!dentry) goto need_lookup; if (dentry->d_op && dentry->d_op->d_revalidate) goto need_revalidate; + inode = dentry->d_inode; + if (!inode) + goto done; + if (inode->i_sb->s_magic == PROC_SUPER_MAGIC) { + struct proc_dir_entry *de = PDE(inode); + + if (de && !vx_hide_check(0, de->vx_flags)) + goto hidden; + } + if (!dx_check(inode->i_tag, DX_WATCH|DX_ADMIN|DX_HOSTID|DX_IDENT)) + goto hidden; done: path->mnt = mnt; path->dentry = dentry; __follow_mount(path); return 0; +hidden: + vxwprintk(1, "xid=%d did lookup hidden %p[#%d,%lu] »%s«.", + vx_current_xid(), inode, inode->i_tag, inode->i_ino, + vxd_path(dentry, mnt)); + dput(dentry); + return -ENOENT; need_lookup: dentry = real_lookup(nd->dentry, name, nd); @@ -1409,7 +1462,8 @@ static inline int check_sticky(struct in * 10. We don't allow removal of NFS sillyrenamed files; it's handled by * nfs_async_unlink(). */ -static int may_delete(struct inode *dir,struct dentry *victim,int isdir) +static int may_delete(struct inode *dir, struct dentry *victim, + int isdir, struct nameidata *nd) { int error; @@ -1423,13 +1477,13 @@ static int may_delete(struct inode *dir, if (nd) nd->flags |= LOOKUP_CONTINUE; #endif - error = permission(dir,MAY_WRITE | MAY_EXEC, NULL); + error = permission(dir,MAY_WRITE | MAY_EXEC, nd); if (error) return error; if (IS_APPEND(dir)) return -EPERM; if (check_sticky(dir, victim->d_inode)||IS_APPEND(victim->d_inode)|| - IS_IMMUTABLE(victim->d_inode)) + IS_IXORUNLINK(victim->d_inode)) return -EPERM; if (isdir) { if (!S_ISDIR(victim->d_inode->i_mode)) @@ -1562,6 +1616,14 @@ int may_open(struct nameidata *nd, int a if (S_ISDIR(inode->i_mode) && (flag & FMODE_WRITE)) return -EISDIR; +#ifdef CONFIG_VSERVER_COWBL + if (IS_COW(inode) && (flag & FMODE_WRITE)) { + if (IS_COW_LINK(inode)) + return -EMLINK; + inode->i_flags &= ~(S_IUNLINK|S_IMMUTABLE); + mark_inode_dirty(inode); + } +#endif error = vfs_permission(nd, acc_mode); if (error) return error; @@ -1578,7 +1640,8 @@ int may_open(struct nameidata *nd, int a return -EACCES; flag &= ~O_TRUNC; - } else if (IS_RDONLY(inode) && (flag & FMODE_WRITE)) + } else if ((IS_RDONLY(inode) || MNT_IS_RDONLY(nd->mnt)) + && (flag & FMODE_WRITE)) return -EROFS; /* * An append-only file must be opened in append mode for writing. @@ -1667,6 +1730,11 @@ int open_namei(int dfd, const char *path struct dentry *dir; int count = 0; +#ifdef CONFIG_VSERVER_COWBL + int rflag = flag; + int rmode = mode; +restart: +#endif acc_mode = ACC_MODE(flag); /* O_TRUNC implies we need access checks for write permissions */ @@ -1760,6 +1828,22 @@ do_last: goto exit; ok: error = may_open(nd, acc_mode, flag); +#ifdef CONFIG_VSERVER_COWBL + if (error == -EMLINK) { + struct dentry *dentry; + dentry = cow_break_link(pathname); + if (IS_ERR(dentry)) { + error = PTR_ERR(dentry); + goto exit; + } + dput(dentry); + release_open_intent(nd); + path_release(nd); + flag = rflag; + mode = rmode; + goto restart; + } +#endif if (error) goto exit; return 0; @@ -1872,9 +1956,9 @@ fail: EXPORT_SYMBOL_GPL(lookup_create); int vfs_mknod(struct inode *dir, struct dentry *dentry, struct vfsmount *mnt, - int mode, dev_t dev) + int mode, dev_t dev, struct nameidata *nd) { - int error = may_create(dir, dentry, NULL); + int error = may_create(dir, dentry, nd); if (error) return error; @@ -1925,11 +2009,11 @@ asmlinkage long sys_mknodat(int dfd, con break; case S_IFCHR: case S_IFBLK: error = vfs_mknod(nd.dentry->d_inode, dentry, nd.mnt, - mode, new_decode_dev(dev)); + mode, new_decode_dev(dev), &nd); break; case S_IFIFO: case S_IFSOCK: error = vfs_mknod(nd.dentry->d_inode, dentry, nd.mnt, - mode, 0); + mode, 0, &nd); break; case S_IFDIR: error = -EPERM; @@ -1953,9 +2037,9 @@ asmlinkage long sys_mknod(const char __u } int vfs_mkdir(struct inode *dir, struct dentry *dentry, struct vfsmount *mnt, - int mode) + int mode, struct nameidata *nd) { - int error = may_create(dir, dentry, NULL); + int error = may_create(dir, dentry, nd); if (error) return error; @@ -1997,7 +2081,7 @@ asmlinkage long sys_mkdirat(int dfd, con if (!IS_POSIXACL(nd.dentry->d_inode)) mode &= ~current->fs->umask; - error = vfs_mkdir(nd.dentry->d_inode, dentry, nd.mnt, mode); + error = vfs_mkdir(nd.dentry->d_inode, dentry, nd.mnt, mode, &nd); dput(dentry); out_unlock: mutex_unlock(&nd.dentry->d_inode->i_mutex); @@ -2040,9 +2124,9 @@ void dentry_unhash(struct dentry *dentry spin_unlock(&dcache_lock); } -int vfs_rmdir(struct inode *dir, struct dentry *dentry,struct vfsmount *mnt) +int vfs_rmdir(struct inode *dir, struct dentry *dentry,struct vfsmount *mnt, struct nameidata *nd) { - int error = may_delete(dir, dentry, 1); + int error = may_delete(dir, dentry, 1, nd); if (error) return error; @@ -2105,7 +2189,7 @@ static long do_rmdir(int dfd, const char error = PTR_ERR(dentry); if (IS_ERR(dentry)) goto exit2; - error = vfs_rmdir(nd.dentry->d_inode, dentry, nd.mnt); + error = vfs_rmdir(nd.dentry->d_inode, dentry, nd.mnt, &nd); dput(dentry); exit2: mutex_unlock(&nd.dentry->d_inode->i_mutex); @@ -2121,9 +2205,9 @@ asmlinkage long sys_rmdir(const char __u return do_rmdir(AT_FDCWD, pathname); } -int vfs_unlink(struct inode *dir, struct dentry *dentry, struct vfsmount *mnt) +int vfs_unlink(struct inode *dir, struct dentry *dentry, struct vfsmount *mnt, struct nameidata *nd) { - int error = may_delete(dir, dentry, 0); + int error = may_delete(dir, dentry, 0, nd); if (error) return error; @@ -2185,7 +2269,7 @@ static long do_unlinkat(int dfd, const c inode = dentry->d_inode; if (inode) atomic_inc(&inode->i_count); - error = vfs_unlink(nd.dentry->d_inode, dentry, nd.mnt); + error = vfs_unlink(nd.dentry->d_inode, dentry, nd.mnt, &nd); exit2: dput(dentry); } @@ -2221,9 +2305,9 @@ asmlinkage long sys_unlink(const char __ } int vfs_symlink(struct inode *dir, struct dentry *dentry, struct vfsmount *mnt, - const char *oldname, int mode) + const char *oldname, int mode, struct nameidata *nd) { - int error = may_create(dir, dentry, NULL); + int error = may_create(dir, dentry, nd); if (error) return error; @@ -2268,7 +2352,7 @@ asmlinkage long sys_symlinkat(const char goto out_unlock; error = vfs_symlink(nd.dentry->d_inode, dentry, nd.mnt, from, - S_IALLUGO); + S_IALLUGO, &nd); dput(dentry); out_unlock: mutex_unlock(&nd.dentry->d_inode->i_mutex); @@ -2285,7 +2369,7 @@ asmlinkage long sys_symlink(const char _ return sys_symlinkat(oldname, AT_FDCWD, newname); } -int vfs_link(struct dentry *old_dentry, struct vfsmount *old_mnt, struct inode *dir, struct dentry *new_dentry, struct vfsmount *new_mnt) +int vfs_link(struct dentry *old_dentry, struct vfsmount *old_mnt, struct inode *dir, struct dentry *new_dentry, struct vfsmount *new_mnt, struct nameidata *nd) { struct inode *inode = old_dentry->d_inode; int error; @@ -2293,7 +2377,7 @@ int vfs_link(struct dentry *old_dentry, if (!inode) return -ENOENT; - error = may_create(dir, new_dentry, NULL); + error = may_create(dir, new_dentry, nd); if (error) return error; @@ -2303,7 +2387,7 @@ int vfs_link(struct dentry *old_dentry, /* * A link to an append-only or immutable file cannot be created. */ - if (IS_APPEND(inode) || IS_IMMUTABLE(inode)) + if (IS_APPEND(inode) || IS_IXORUNLINK(inode)) return -EPERM; if (!dir->i_op || !dir->i_op->link) return -EPERM; @@ -2365,7 +2449,7 @@ asmlinkage long sys_linkat(int olddfd, c if (IS_ERR(new_dentry)) goto out_unlock; error = vfs_link(old_nd.dentry, old_nd.mnt, nd.dentry->d_inode, - new_dentry, nd.mnt); + new_dentry, nd.mnt, &nd); dput(new_dentry); out_unlock: mutex_unlock(&nd.dentry->d_inode->i_mutex); @@ -2502,14 +2586,14 @@ int vfs_rename(struct inode *old_dir, st if (old_dentry->d_inode == new_dentry->d_inode) return 0; - error = may_delete(old_dir, old_dentry, is_dir); + error = may_delete(old_dir, old_dentry, is_dir, NULL); if (error) return error; if (!new_dentry->d_inode) error = may_create(new_dir, new_dentry, NULL); else - error = may_delete(new_dir, new_dentry, is_dir); + error = may_delete(new_dir, new_dentry, is_dir, NULL); if (error) return error; @@ -2589,6 +2673,9 @@ static int do_rename(int olddfd, const c error = -EINVAL; if (old_dentry == trap) goto exit4; + error = -EROFS; + if (MNT_IS_RDONLY(newnd.mnt)) + goto exit4; new_dentry = lookup_hash(&newnd); error = PTR_ERR(new_dentry); if (IS_ERR(new_dentry)) @@ -2682,6 +2769,153 @@ int vfs_follow_link(struct nameidata *nd return __vfs_follow_link(nd, link); } + +#ifdef CONFIG_VSERVER_COWBL + +#include + +struct dentry *cow_break_link(const char *pathname) +{ + int ret, mode, pathlen; + struct nameidata old_nd, dir_nd; + struct dentry *old_dentry, *new_dentry; + struct dentry *res = ERR_PTR(-EMLINK); + struct vfsmount *old_mnt, *new_mnt; + struct file *old_file; + struct file *new_file; + char *to, *path, pad='\251'; + loff_t size; + + vxdprintk(VXD_CBIT(misc, 1), "cow_break_link(»%s«)", pathname); + path = kmalloc(PATH_MAX, GFP_KERNEL); + if (!path) + goto out; + + ret = path_lookup(pathname, LOOKUP_FOLLOW, &old_nd); + vxdprintk(VXD_CBIT(misc, 2), "path_lookup(old): %d", ret); + if (ret < 0) + goto out_free_path; + + old_dentry = old_nd.dentry; + old_mnt = old_nd.mnt; + mode = old_dentry->d_inode->i_mode; + + to = d_path(old_dentry, old_mnt, path, PATH_MAX-2); + pathlen = strlen(to); + vxdprintk(VXD_CBIT(misc, 2), "old path »%s«", to); + + to[pathlen+1] = 0; +retry: + to[pathlen] = pad--; + if (pad <= '\240') + goto out_rel_old; + + vxdprintk(VXD_CBIT(misc, 1), "temp copy »%s«", to); + ret = path_lookup(to, + LOOKUP_PARENT|LOOKUP_OPEN|LOOKUP_CREATE, &dir_nd); + vxdprintk(VXD_CBIT(misc, 2), + "path_lookup(new): %d", ret); + if (ret < 0) + goto retry; + + /* this puppy downs the inode sem */ + new_dentry = lookup_create(&dir_nd, 0); + vxdprintk(VXD_CBIT(misc, 2), + "lookup_create(new): %p", new_dentry); + if (!new_dentry || IS_ERR(new_dentry)) { + path_release(&dir_nd); + goto retry; + } + + ret = vfs_create(dir_nd.dentry->d_inode, new_dentry, mode, &dir_nd); + vxdprintk(VXD_CBIT(misc, 2), + "vfs_create(new): %d", ret); + if (ret == -EEXIST) { + mutex_unlock(&dir_nd.dentry->d_inode->i_mutex); + dput(new_dentry); + path_release(&dir_nd); + goto retry; + } + else if (ret < 0) { + res = ERR_PTR(ret); + goto out_rel_both; + } + + new_mnt = dir_nd.mnt; + + dget(old_dentry); + mntget(old_mnt); + /* this one cleans up the dentry in case of failure */ + old_file = dentry_open(old_dentry, old_mnt, O_RDONLY); + vxdprintk(VXD_CBIT(misc, 2), + "dentry_open(old): %p", old_file); + if (!old_file || IS_ERR(old_file)) { + res = IS_ERR(old_file) ? (void *) old_file : res; + goto out_rel_both; + } + + dget(new_dentry); + mntget(new_mnt); + /* this one cleans up the dentry in case of failure */ + new_file = dentry_open(new_dentry, new_mnt, O_WRONLY); + vxdprintk(VXD_CBIT(misc, 2), + "dentry_open(new): %p", new_file); + if (!new_file || IS_ERR(new_file)) { + res = IS_ERR(new_file) ? (void *) new_file : res; + goto out_fput_old; + } + + size = i_size_read(old_file->f_dentry->d_inode); + ret = vfs_sendfile(new_file, old_file, NULL, size, 0); + vxdprintk(VXD_CBIT(misc, 2), "vfs_sendfile: %d", ret); + if (ret < 0) { + res = ERR_PTR(ret); + goto out_fput_both; + } + else if (ret < size) { + res = ERR_PTR(-ENOSPC); + goto out_fput_both; + } + + ret = vfs_rename(dir_nd.dentry->d_inode, new_dentry, new_mnt, + old_nd.dentry->d_parent->d_inode, old_dentry, old_mnt); + vxdprintk(VXD_CBIT(misc, 2), "vfs_rename: %d", ret); + if (!ret) { + res = new_dentry; + dget(new_dentry); + } + else + res = ERR_PTR(ret); + +out_fput_both: + vxdprintk(VXD_CBIT(misc, 3), + "fput(new_file=%p[#%d])", new_file, + atomic_read(&new_file->f_count)); + fput(new_file); + +out_fput_old: + vxdprintk(VXD_CBIT(misc, 3), + "fput(old_file=%p[#%d])", old_file, + atomic_read(&old_file->f_count)); + fput(old_file); + +out_rel_both: + mutex_unlock(&dir_nd.dentry->d_inode->i_mutex); + if (IS_ERR(res)) + vfs_unlink(dir_nd.dentry->d_inode, new_dentry, new_mnt, &dir_nd); + dput(new_dentry); + + path_release(&dir_nd); +out_rel_old: + path_release(&old_nd); +out_free_path: + kfree(path); +out: + return res; +} + +#endif + /* get the link contents into pagecache */ static char *page_getlink(struct dentry * dentry, struct page **ppage) { diff -NurpP --minimal linux-source-2.6.22.orig/fs/namespace.c linux-source-2.6.22/fs/namespace.c --- linux-source-2.6.22.orig/fs/namespace.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/fs/namespace.c 2007-10-24 16:14:06.291582822 +0200 @@ -25,6 +25,11 @@ #include #include #include +#include +#include +#include +#include +#include #include #include #include "pnode.h" @@ -241,6 +246,7 @@ static struct vfsmount *clone_mnt(struct mnt->mnt_root = dget(root); mnt->mnt_mountpoint = mnt->mnt_root; mnt->mnt_parent = mnt; + mnt->mnt_tag = old->mnt_tag; if (flag & CL_SLAVE) { list_add(&mnt->mnt_slave, &old->mnt_slave_list); @@ -349,48 +355,91 @@ static inline void mangle(struct seq_fil seq_escape(m, s, " \t\n\\"); } +static int mnt_is_reachable(struct vfsmount *mnt) +{ + struct vfsmount *root_mnt; + struct dentry *root, *point; + int ret; + + if (mnt == mnt->mnt_ns->root) + return 1; + + spin_lock(&vfsmount_lock); + root_mnt = current->fs->rootmnt; + root = current->fs->root; + point = root; + + while ((mnt != mnt->mnt_parent) && (mnt != root_mnt)) { + point = mnt->mnt_mountpoint; + mnt = mnt->mnt_parent; + } + + ret = (mnt == root_mnt) && is_subdir(point, root); + + spin_unlock(&vfsmount_lock); + + return ret; +} + static int show_vfsmnt(struct seq_file *m, void *v) { struct vfsmount *mnt = v; int err = 0; static struct proc_fs_info { - int flag; - char *str; + int s_flag; + int mnt_flag; + char *set_str; + char *unset_str; } fs_info[] = { - { MS_SYNCHRONOUS, ",sync" }, - { MS_DIRSYNC, ",dirsync" }, - { MS_MANDLOCK, ",mand" }, - { 0, NULL } - }; - static struct proc_fs_info mnt_info[] = { - { MNT_NOSUID, ",nosuid" }, - { MNT_NODEV, ",nodev" }, - { MNT_NOEXEC, ",noexec" }, - { MNT_NOATIME, ",noatime" }, - { MNT_NODIRATIME, ",nodiratime" }, - { MNT_RELATIME, ",relatime" }, - { 0, NULL } + { MS_RDONLY, MNT_RDONLY, "ro", "rw" }, + { MS_SYNCHRONOUS, 0, ",sync", NULL }, + { MS_DIRSYNC, 0, ",dirsync", NULL }, + { MS_MANDLOCK, 0, ",mand", NULL }, + { MS_TAGGED, 0, ",tag", NULL }, + { MS_NOATIME, MNT_NOATIME, ",noatime", NULL }, + { MS_NODIRATIME, MNT_NODIRATIME, ",nodiratime", NULL }, + { MS_RELATIME, MNT_RELATIME, ",relatime", NULL }, + { 0, MNT_NOSUID, ",nosuid", NULL }, + { 0, MNT_NODEV, ",nodev", NULL }, + { 0, MNT_NOEXEC, ",noexec", NULL }, + { 0, 0, NULL, NULL } }; - struct proc_fs_info *fs_infop; + struct proc_fs_info *p; + unsigned long s_flags = mnt->mnt_sb->s_flags; + int mnt_flags = mnt->mnt_flags; - mangle(m, mnt->mnt_devname ? mnt->mnt_devname : "none"); - seq_putc(m, ' '); - seq_path(m, mnt, mnt->mnt_root, " \t\n\\"); - seq_putc(m, ' '); - mangle(m, mnt->mnt_sb->s_type->name); - if (mnt->mnt_sb->s_subtype && mnt->mnt_sb->s_subtype[0]) { - seq_putc(m, '.'); - mangle(m, mnt->mnt_sb->s_subtype); - } - seq_puts(m, mnt->mnt_sb->s_flags & MS_RDONLY ? " ro" : " rw"); - for (fs_infop = fs_info; fs_infop->flag; fs_infop++) { - if (mnt->mnt_sb->s_flags & fs_infop->flag) - seq_puts(m, fs_infop->str); + if (vx_flags(VXF_HIDE_MOUNT, 0)) + return 0; + if (!mnt_is_reachable(mnt) && !vx_check(0, VS_WATCH_P)) + return 0; + + if (!vx_check(0, VS_ADMIN|VS_WATCH) && + mnt == current->fs->rootmnt) { + seq_puts(m, "/dev/root / "); + } else { + mangle(m, mnt->mnt_devname ? mnt->mnt_devname : "none"); + seq_putc(m, ' '); + seq_path(m, mnt, mnt->mnt_root, " \t\n\\"); + seq_putc(m, ' '); + + if (mnt->mnt_sb->s_subtype && mnt->mnt_sb->s_subtype[0]) { + seq_putc(m, '.'); + mangle(m, mnt->mnt_sb->s_subtype); + } } - for (fs_infop = mnt_info; fs_infop->flag; fs_infop++) { - if (mnt->mnt_flags & fs_infop->flag) - seq_puts(m, fs_infop->str); + mangle(m, mnt->mnt_sb->s_type->name); + seq_putc(m, ' '); + for (p = fs_info; (p->s_flag | p->mnt_flag) ; p++) { + if ((s_flags & p->s_flag) || (mnt_flags & p->mnt_flag)) { + if (p->set_str) + seq_puts(m, p->set_str); + } else { + if (p->unset_str) + seq_puts(m, p->unset_str); + } } + if (mnt->mnt_flags & MNT_TAGID) + seq_printf(m, ",tag=%d", mnt->mnt_tag); if (mnt->mnt_sb->s_op->show_options) err = mnt->mnt_sb->s_op->show_options(m, mnt); seq_puts(m, " 0 0\n"); @@ -409,17 +458,27 @@ static int show_vfsstat(struct seq_file struct vfsmount *mnt = v; int err = 0; - /* device */ - if (mnt->mnt_devname) { - seq_puts(m, "device "); - mangle(m, mnt->mnt_devname); - } else - seq_puts(m, "no device"); + if (vx_flags(VXF_HIDE_MOUNT, 0)) + return 0; + if (!mnt_is_reachable(mnt) && !vx_check(0, VS_WATCH_P)) + return 0; - /* mount point */ - seq_puts(m, " mounted on "); - seq_path(m, mnt, mnt->mnt_root, " \t\n\\"); - seq_putc(m, ' '); + if (!vx_check(0, VS_ADMIN|VS_WATCH) && + mnt == current->fs->rootmnt) { + seq_puts(m, "device /dev/root mounted on / "); + } else { + /* device */ + if (mnt->mnt_devname) { + seq_puts(m, "device "); + mangle(m, mnt->mnt_devname); + } else + seq_puts(m, "no device"); + + /* mount point */ + seq_puts(m, " mounted on "); + seq_path(m, mnt, mnt->mnt_root, " \t\n\\"); + seq_putc(m, ' '); + } /* file system type */ seq_puts(m, "with fstype "); @@ -649,7 +708,7 @@ asmlinkage long sys_umount(char __user * goto dput_and_out; retval = -EPERM; - if (!capable(CAP_SYS_ADMIN)) + if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT)) goto dput_and_out; retval = do_umount(nd.mnt, flags); @@ -673,7 +732,7 @@ asmlinkage long sys_oldumount(char __use static int mount_is_safe(struct nameidata *nd) { - if (capable(CAP_SYS_ADMIN)) + if (vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT)) return 0; return -EPERM; #ifdef notyet @@ -905,11 +964,13 @@ static int do_change_type(struct nameida /* * do loopback mount. */ -static int do_loopback(struct nameidata *nd, char *old_name, int recurse) +static int do_loopback(struct nameidata *nd, char *old_name, tag_t tag, + unsigned long flags, int mnt_flags) { struct nameidata old_nd; struct vfsmount *mnt = NULL; int err = mount_is_safe(nd); + int recurse = flags & MS_REC; if (err) return err; if (!old_name || !*old_name) @@ -935,6 +996,12 @@ static int do_loopback(struct nameidata if (!mnt) goto out; + mnt->mnt_flags = mnt_flags; + if (flags & MS_TAGID) { + mnt->mnt_tag = tag; + mnt->mnt_flags |= MNT_TAGID; + } + err = graft_tree(mnt, nd); if (err) { LIST_HEAD(umount_list); @@ -943,6 +1010,7 @@ static int do_loopback(struct nameidata spin_unlock(&vfsmount_lock); release_mounts(&umount_list); } + mnt->mnt_flags = mnt_flags; out: up_write(&namespace_sem); @@ -956,12 +1024,12 @@ out: * on it - tough luck. */ static int do_remount(struct nameidata *nd, int flags, int mnt_flags, - void *data) + void *data, xid_t xid) { int err; struct super_block *sb = nd->mnt->mnt_sb; - if (!capable(CAP_SYS_ADMIN)) + if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_REMOUNT)) return -EPERM; if (!check_mnt(nd->mnt)) @@ -995,7 +1063,7 @@ static int do_move_mount(struct nameidat struct nameidata old_nd, parent_nd; struct vfsmount *p; int err = 0; - if (!capable(CAP_SYS_ADMIN)) + if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT)) return -EPERM; if (!old_name || !*old_name) return -EINVAL; @@ -1075,7 +1143,7 @@ static int do_new_mount(struct nameidata return -EINVAL; /* we need capabilities... */ - if (!capable(CAP_SYS_ADMIN)) + if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT)) return -EPERM; mnt = do_kern_mount(type, flags, name, data); @@ -1387,6 +1455,7 @@ long do_mount(char *dev_name, char *dir_ struct nameidata nd; int retval = 0; int mnt_flags = 0; + tag_t tag = 0; /* Discard magic */ if ((flags & MS_MGC_MSK) == MS_MGC_VAL) @@ -1402,7 +1471,19 @@ long do_mount(char *dev_name, char *dir_ if (data_page) ((char *)data_page)[PAGE_SIZE - 1] = 0; +#ifdef CONFIG_PROPAGATE + retval = dx_parse_tag(data_page, &tag, 1); + if (retval) { + mnt_flags |= MNT_TAGID; + /* bind and re-mounts get the tag flag */ + if (flags & (MS_BIND|MS_REMOUNT)) + flags |= MS_TAGID; + } +#endif + /* Separate the per-mountpoint flags */ + if (flags & MS_RDONLY) + mnt_flags |= MNT_RDONLY; if (flags & MS_NOSUID) mnt_flags |= MNT_NOSUID; if (flags & MS_NODEV) @@ -1416,6 +1497,8 @@ long do_mount(char *dev_name, char *dir_ if (flags & MS_RELATIME) mnt_flags |= MNT_RELATIME; + if (!capable(CAP_SYS_ADMIN)) + mnt_flags |= MNT_NODEV; flags &= ~(MS_NOSUID | MS_NOEXEC | MS_NODEV | MS_ACTIVE | MS_NOATIME | MS_NODIRATIME | MS_RELATIME); @@ -1430,9 +1513,9 @@ long do_mount(char *dev_name, char *dir_ if (flags & MS_REMOUNT) retval = do_remount(&nd, flags & ~MS_REMOUNT, mnt_flags, - data_page); + data_page, tag); else if (flags & MS_BIND) - retval = do_loopback(&nd, dev_name, flags & MS_REC); + retval = do_loopback(&nd, dev_name, tag, flags, mnt_flags); else if (flags & (MS_SHARED | MS_PRIVATE | MS_SLAVE | MS_UNBINDABLE)) retval = do_change_type(&nd, flags); else if (flags & MS_MOVE) @@ -1505,6 +1588,7 @@ static struct mnt_namespace *dup_mnt_ns( q = next_mnt(q, new_ns->root); } up_write(&namespace_sem); + atomic_inc(&vs_global_mnt_ns); if (rootmnt) mntput(rootmnt); @@ -1867,6 +1951,7 @@ void __put_mnt_ns(struct mnt_namespace * spin_unlock(&vfsmount_lock); up_write(&namespace_sem); release_mounts(&umount_list); + atomic_dec(&vs_global_mnt_ns); kfree(ns); } diff -NurpP --minimal linux-source-2.6.22.orig/fs/nfs/client.c linux-source-2.6.22/fs/nfs/client.c --- linux-source-2.6.22.orig/fs/nfs/client.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/fs/nfs/client.c 2007-10-24 16:14:06.291582822 +0200 @@ -541,6 +541,9 @@ static int nfs_init_server_rpcclient(str if (server->flags & NFS4_MOUNT_INTR) server->client->cl_intr = 1; + server->client->cl_tag = 0; + if (server->flags & NFS_MOUNT_TAGGED) + server->client->cl_tag = 1; return 0; } @@ -698,6 +701,10 @@ static void nfs_server_set_fsinfo(struct server->acdirmin = server->acdirmax = 0; } + /* FIXME: needs fsinfo + if (server->flags & NFS_MOUNT_TAGGED) + sb->s_flags |= MS_TAGGED; */ + server->maxfilesize = fsinfo->maxfilesize; /* We're airborne Set socket buffersize */ diff -NurpP --minimal linux-source-2.6.22.orig/fs/nfs/dir.c linux-source-2.6.22/fs/nfs/dir.c --- linux-source-2.6.22.orig/fs/nfs/dir.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/nfs/dir.c 2007-10-24 16:14:06.291582822 +0200 @@ -34,6 +34,7 @@ #include #include #include +#include #include "nfs4_fs.h" #include "delegation.h" @@ -956,6 +957,7 @@ static struct dentry *nfs_lookup(struct if (IS_ERR(res)) goto out_unlock; + dx_propagate_tag(nd, inode); no_entry: res = d_materialise_unique(dentry, inode); if (res != NULL) { @@ -998,7 +1000,8 @@ static int is_atomic_open(struct inode * if (nd->flags & LOOKUP_DIRECTORY) return 0; /* Are we trying to write to a read only partition? */ - if (IS_RDONLY(dir) && (nd->intent.open.flags & (O_CREAT|O_TRUNC|FMODE_WRITE))) + if ((IS_RDONLY(dir) || MNT_IS_RDONLY(nd->mnt)) && + (nd->intent.open.flags & (O_CREAT|O_TRUNC|FMODE_WRITE))) return 0; return 1; } diff -NurpP --minimal linux-source-2.6.22.orig/fs/nfs/inode.c linux-source-2.6.22/fs/nfs/inode.c --- linux-source-2.6.22.orig/fs/nfs/inode.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/nfs/inode.c 2007-10-24 16:14:06.291582822 +0200 @@ -37,6 +37,7 @@ #include #include #include +#include #include #include @@ -285,8 +286,10 @@ nfs_fhget(struct super_block *sb, struct nfsi->change_attr = fattr->change_attr; inode->i_size = nfs_size_to_loff_t(fattr->size); inode->i_nlink = fattr->nlink; - inode->i_uid = fattr->uid; - inode->i_gid = fattr->gid; + inode->i_uid = INOTAG_UID(DX_TAG(inode), fattr->uid, fattr->gid); + inode->i_gid = INOTAG_GID(DX_TAG(inode), fattr->uid, fattr->gid); + inode->i_tag = INOTAG_TAG(DX_TAG(inode), fattr->uid, fattr->gid, 0); + /* maybe fattr->xid someday */ if (fattr->valid & (NFS_ATTR_FATTR_V3 | NFS_ATTR_FATTR_V4)) { /* * report the blocks in 512byte units @@ -377,6 +380,8 @@ void nfs_setattr_update_inode(struct ino inode->i_uid = attr->ia_uid; if ((attr->ia_valid & ATTR_GID) != 0) inode->i_gid = attr->ia_gid; + if ((attr->ia_valid & ATTR_TAG) && IS_TAGGED(inode)) + inode->i_tag = attr->ia_tag; spin_lock(&inode->i_lock); NFS_I(inode)->cache_validity |= NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL; spin_unlock(&inode->i_lock); @@ -825,6 +830,9 @@ static int nfs_check_inode_attributes(st struct nfs_inode *nfsi = NFS_I(inode); loff_t cur_size, new_isize; int data_unstable; + uid_t uid; + gid_t gid; + tag_t tag; /* Has the inode gone and changed behind our back? */ @@ -852,10 +860,15 @@ static int nfs_check_inode_attributes(st if (cur_size != new_isize && nfsi->npages == 0) nfsi->cache_validity |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE; + uid = INOTAG_UID(DX_TAG(inode), fattr->uid, fattr->gid); + gid = INOTAG_GID(DX_TAG(inode), fattr->uid, fattr->gid); + tag = INOTAG_TAG(DX_TAG(inode), fattr->uid, fattr->gid, 0); + /* Have any file permissions changed? */ if ((inode->i_mode & S_IALLUGO) != (fattr->mode & S_IALLUGO) - || inode->i_uid != fattr->uid - || inode->i_gid != fattr->gid) + || inode->i_uid != uid + || inode->i_gid != gid + || inode->i_tag != tag) nfsi->cache_validity |= NFS_INO_INVALID_ATTR | NFS_INO_INVALID_ACCESS | NFS_INO_INVALID_ACL; /* Has the link count changed? */ @@ -946,6 +959,9 @@ static int nfs_update_inode(struct inode unsigned int invalid = 0; unsigned long now = jiffies; int data_stable; + uid_t uid; + gid_t gid; + tag_t tag; dfprintk(VFS, "NFS: %s(%s/%ld ct=%d info=0x%x)\n", __FUNCTION__, inode->i_sb->s_id, inode->i_ino, @@ -1022,15 +1038,21 @@ static int nfs_update_inode(struct inode } memcpy(&inode->i_atime, &fattr->atime, sizeof(inode->i_atime)); + uid = INOTAG_UID(DX_TAG(inode), fattr->uid, fattr->gid); + gid = INOTAG_GID(DX_TAG(inode), fattr->uid, fattr->gid); + tag = INOTAG_TAG(DX_TAG(inode), fattr->uid, fattr->gid, 0); + if ((inode->i_mode & S_IALLUGO) != (fattr->mode & S_IALLUGO) || - inode->i_uid != fattr->uid || - inode->i_gid != fattr->gid) + inode->i_uid != uid || + inode->i_gid != gid || + inode->i_tag != tag) invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL; inode->i_mode = fattr->mode; inode->i_nlink = fattr->nlink; - inode->i_uid = fattr->uid; - inode->i_gid = fattr->gid; + inode->i_uid = uid; + inode->i_gid = gid; + inode->i_tag = tag; if (fattr->valid & (NFS_ATTR_FATTR_V3 | NFS_ATTR_FATTR_V4)) { /* diff -NurpP --minimal linux-source-2.6.22.orig/fs/nfs/nfs3xdr.c linux-source-2.6.22/fs/nfs/nfs3xdr.c --- linux-source-2.6.22.orig/fs/nfs/nfs3xdr.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/nfs/nfs3xdr.c 2007-10-24 16:14:06.291582822 +0200 @@ -22,6 +22,7 @@ #include #include #include +#include #include "internal.h" #define NFSDBG_FACILITY NFSDBG_XDR @@ -178,7 +179,7 @@ xdr_decode_fattr(__be32 *p, struct nfs_f } static inline __be32 * -xdr_encode_sattr(__be32 *p, struct iattr *attr) +xdr_encode_sattr(__be32 *p, struct iattr *attr, int tag) { if (attr->ia_valid & ATTR_MODE) { *p++ = xdr_one; @@ -186,15 +187,17 @@ xdr_encode_sattr(__be32 *p, struct iattr } else { *p++ = xdr_zero; } - if (attr->ia_valid & ATTR_UID) { + if (attr->ia_valid & ATTR_UID || + (tag && (attr->ia_valid & ATTR_TAG))) { *p++ = xdr_one; - *p++ = htonl(attr->ia_uid); + *p++ = htonl(TAGINO_UID(tag, attr->ia_uid, attr->ia_tag)); } else { *p++ = xdr_zero; } - if (attr->ia_valid & ATTR_GID) { + if (attr->ia_valid & ATTR_GID || + (tag && (attr->ia_valid & ATTR_TAG))) { *p++ = xdr_one; - *p++ = htonl(attr->ia_gid); + *p++ = htonl(TAGINO_GID(tag, attr->ia_gid, attr->ia_tag)); } else { *p++ = xdr_zero; } @@ -279,7 +282,8 @@ static int nfs3_xdr_sattrargs(struct rpc_rqst *req, __be32 *p, struct nfs3_sattrargs *args) { p = xdr_encode_fhandle(p, args->fh); - p = xdr_encode_sattr(p, args->sattr); + p = xdr_encode_sattr(p, args->sattr, + req->rq_task->tk_client->cl_tag); *p++ = htonl(args->guard); if (args->guard) p = xdr_encode_time3(p, &args->guardtime); @@ -370,7 +374,8 @@ nfs3_xdr_createargs(struct rpc_rqst *req *p++ = args->verifier[0]; *p++ = args->verifier[1]; } else - p = xdr_encode_sattr(p, args->sattr); + p = xdr_encode_sattr(p, args->sattr, + req->rq_task->tk_client->cl_tag); req->rq_slen = xdr_adjust_iovec(req->rq_svec, p); return 0; @@ -384,7 +389,8 @@ nfs3_xdr_mkdirargs(struct rpc_rqst *req, { p = xdr_encode_fhandle(p, args->fh); p = xdr_encode_array(p, args->name, args->len); - p = xdr_encode_sattr(p, args->sattr); + p = xdr_encode_sattr(p, args->sattr, + req->rq_task->tk_client->cl_tag); req->rq_slen = xdr_adjust_iovec(req->rq_svec, p); return 0; } @@ -397,7 +403,8 @@ nfs3_xdr_symlinkargs(struct rpc_rqst *re { p = xdr_encode_fhandle(p, args->fromfh); p = xdr_encode_array(p, args->fromname, args->fromlen); - p = xdr_encode_sattr(p, args->sattr); + p = xdr_encode_sattr(p, args->sattr, + req->rq_task->tk_client->cl_tag); *p++ = htonl(args->pathlen); req->rq_slen = xdr_adjust_iovec(req->rq_svec, p); @@ -415,7 +422,8 @@ nfs3_xdr_mknodargs(struct rpc_rqst *req, p = xdr_encode_fhandle(p, args->fh); p = xdr_encode_array(p, args->name, args->len); *p++ = htonl(args->type); - p = xdr_encode_sattr(p, args->sattr); + p = xdr_encode_sattr(p, args->sattr, + req->rq_task->tk_client->cl_tag); if (args->type == NF3CHR || args->type == NF3BLK) { *p++ = htonl(MAJOR(args->rdev)); *p++ = htonl(MINOR(args->rdev)); diff -NurpP --minimal linux-source-2.6.22.orig/fs/nfs/nfsroot.c linux-source-2.6.22/fs/nfs/nfsroot.c --- linux-source-2.6.22.orig/fs/nfs/nfsroot.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/nfs/nfsroot.c 2007-10-24 16:14:06.291582822 +0200 @@ -118,12 +118,12 @@ static int mount_port __initdata = 0; / enum { /* Options that take integer arguments */ Opt_port, Opt_rsize, Opt_wsize, Opt_timeo, Opt_retrans, Opt_acregmin, - Opt_acregmax, Opt_acdirmin, Opt_acdirmax, + Opt_acregmax, Opt_acdirmin, Opt_acdirmax, Opt_tagid, /* Options that take no arguments */ Opt_soft, Opt_hard, Opt_intr, Opt_nointr, Opt_posix, Opt_noposix, Opt_cto, Opt_nocto, Opt_ac, Opt_noac, Opt_lock, Opt_nolock, Opt_v2, Opt_v3, Opt_udp, Opt_tcp, - Opt_acl, Opt_noacl, + Opt_acl, Opt_noacl, Opt_tag, Opt_notag, /* Error token */ Opt_err }; @@ -160,6 +160,10 @@ static match_table_t __initdata tokens = {Opt_tcp, "tcp"}, {Opt_acl, "acl"}, {Opt_noacl, "noacl"}, + {Opt_tag, "tag"}, + {Opt_notag, "notag"}, + {Opt_tagid, "tagid=%u"}, + {Opt_tag, "tagxid"}, {Opt_err, NULL} }; @@ -274,6 +278,20 @@ static int __init root_nfs_parse(char *n case Opt_noacl: nfs_data.flags |= NFS_MOUNT_NOACL; break; +#ifndef CONFIG_TAGGING_NONE + case Opt_tag: + nfs_data.flags |= NFS_MOUNT_TAGGED; + break; + case Opt_notag: + nfs_data.flags &= ~NFS_MOUNT_TAGGED; + break; +#endif +#ifdef CONFIG_PROPAGATE + case Opt_tagid: + /* use args[0] */ + nfs_data.flags |= NFS_MOUNT_TAGGED; + break; +#endif default: printk(KERN_WARNING "Root-NFS: unknown " "option: %s\n", p); diff -NurpP --minimal linux-source-2.6.22.orig/fs/nfs/super.c linux-source-2.6.22/fs/nfs/super.c --- linux-source-2.6.22.orig/fs/nfs/super.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/fs/nfs/super.c 2007-10-24 16:14:06.291582822 +0200 @@ -45,6 +45,7 @@ #include #include #include +#include #include #include @@ -291,6 +292,7 @@ static void nfs_show_mount_options(struc { NFS_MOUNT_NONLM, ",nolock", "" }, { NFS_MOUNT_NOACL, ",noacl", "" }, { NFS_MOUNT_NORDIRPLUS, ",nordirplus", "" }, + { NFS_MOUNT_TAGGED, ",tag", "" }, { 0, NULL, NULL } }; const struct proc_nfs_info *nfs_infop; diff -NurpP --minimal linux-source-2.6.22.orig/fs/nfsd/auth.c linux-source-2.6.22/fs/nfsd/auth.c --- linux-source-2.6.22.orig/fs/nfsd/auth.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/nfsd/auth.c 2007-10-24 16:14:06.291582822 +0200 @@ -9,6 +9,7 @@ #include #include #include +#include #define CAP_NFSD_MASK (CAP_FS_MASK|CAP_TO_MASK(CAP_SYS_RESOURCE)) @@ -41,19 +42,22 @@ int nfsd_setuser(struct svc_rqst *rqstp, get_group_info(cred.cr_group_info); if (cred.cr_uid != (uid_t) -1) - current->fsuid = cred.cr_uid; + current->fsuid = INOTAG_UID(DX_TAG_NFSD, cred.cr_uid, cred.cr_gid); else current->fsuid = exp->ex_anon_uid; if (cred.cr_gid != (gid_t) -1) - current->fsgid = cred.cr_gid; + current->fsgid = INOTAG_GID(DX_TAG_NFSD, cred.cr_uid, cred.cr_gid); else current->fsgid = exp->ex_anon_gid; + /* this desperately needs a tag :) */ + current->xid = (xid_t)INOTAG_TAG(DX_TAG_NFSD, cred.cr_uid, cred.cr_gid, 0); + if (!cred.cr_group_info) return -ENOMEM; ret = set_current_groups(cred.cr_group_info); put_group_info(cred.cr_group_info); - if ((cred.cr_uid)) { + if (INOTAG_UID(DX_TAG_NFSD, cred.cr_uid, cred.cr_gid)) { cap_t(current->cap_effective) &= ~CAP_NFSD_MASK; } else { cap_t(current->cap_effective) |= (CAP_NFSD_MASK & diff -NurpP --minimal linux-source-2.6.22.orig/fs/nfsd/nfs3xdr.c linux-source-2.6.22/fs/nfsd/nfs3xdr.c --- linux-source-2.6.22.orig/fs/nfsd/nfs3xdr.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/nfsd/nfs3xdr.c 2007-10-24 16:14:06.291582822 +0200 @@ -21,6 +21,7 @@ #include #include #include +#include #define NFSDDBG_FACILITY NFSDDBG_XDR @@ -107,6 +108,8 @@ static __be32 * decode_sattr3(__be32 *p, struct iattr *iap) { u32 tmp; + uid_t uid = 0; + gid_t gid = 0; iap->ia_valid = 0; @@ -116,12 +119,15 @@ decode_sattr3(__be32 *p, struct iattr *i } if (*p++) { iap->ia_valid |= ATTR_UID; - iap->ia_uid = ntohl(*p++); + uid = ntohl(*p++); } if (*p++) { iap->ia_valid |= ATTR_GID; - iap->ia_gid = ntohl(*p++); + gid = ntohl(*p++); } + iap->ia_uid = INOTAG_UID(DX_TAG_NFSD, uid, gid); + iap->ia_gid = INOTAG_GID(DX_TAG_NFSD, uid, gid); + iap->ia_tag = INOTAG_TAG(DX_TAG_NFSD, uid, gid, 0); if (*p++) { u64 newsize; @@ -180,8 +186,10 @@ encode_fattr3(struct svc_rqst *rqstp, __ *p++ = htonl(nfs3_ftypes[(stat->mode & S_IFMT) >> 12]); *p++ = htonl((u32) stat->mode); *p++ = htonl((u32) stat->nlink); - *p++ = htonl((u32) nfsd_ruid(rqstp, stat->uid)); - *p++ = htonl((u32) nfsd_rgid(rqstp, stat->gid)); + *p++ = htonl((u32) nfsd_ruid(rqstp, + TAGINO_UID(DX_TAG(dentry->d_inode), stat->uid, stat->tag))); + *p++ = htonl((u32) nfsd_rgid(rqstp, + TAGINO_GID(DX_TAG(dentry->d_inode), stat->gid, stat->tag))); if (S_ISLNK(stat->mode) && stat->size > NFS3_MAXPATHLEN) { p = xdr_encode_hyper(p, (u64) NFS3_MAXPATHLEN); } else { diff -NurpP --minimal linux-source-2.6.22.orig/fs/nfsd/nfs4recover.c linux-source-2.6.22/fs/nfsd/nfs4recover.c --- linux-source-2.6.22.orig/fs/nfsd/nfs4recover.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/fs/nfsd/nfs4recover.c 2007-10-24 16:14:06.291582822 +0200 @@ -157,7 +157,7 @@ nfsd4_create_clid_dir(struct nfs4_client goto out_put; } status = vfs_mkdir(rec_dir.dentry->d_inode, dentry, rec_dir.mnt, - S_IRWXU); + S_IRWXU, NULL); out_put: dput(dentry); out_unlock: @@ -261,7 +261,7 @@ nfsd4_remove_clid_file(struct dentry *di return -EINVAL; } mutex_lock_nested(&dir->d_inode->i_mutex, I_MUTEX_PARENT); - status = vfs_unlink(dir->d_inode, dentry, rec_dir.mnt); + status = vfs_unlink(dir->d_inode, dentry, rec_dir.mnt, NULL); mutex_unlock(&dir->d_inode->i_mutex); return status; } @@ -276,7 +276,7 @@ nfsd4_clear_clid_dir(struct dentry *dir, * a kernel from the future.... */ nfsd4_list_rec_dir(dentry, nfsd4_remove_clid_file); mutex_lock_nested(&dir->d_inode->i_mutex, I_MUTEX_PARENT); - status = vfs_rmdir(dir->d_inode, dentry, rec_dir.mnt); + status = vfs_rmdir(dir->d_inode, dentry, rec_dir.mnt, NULL); mutex_unlock(&dir->d_inode->i_mutex); return status; } diff -NurpP --minimal linux-source-2.6.22.orig/fs/nfsd/nfs4xdr.c linux-source-2.6.22/fs/nfsd/nfs4xdr.c --- linux-source-2.6.22.orig/fs/nfsd/nfs4xdr.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/fs/nfsd/nfs4xdr.c 2007-10-24 16:14:06.291582822 +0200 @@ -56,6 +56,7 @@ #include #include #include +#include #define NFSDDBG_FACILITY NFSDDBG_XDR @@ -1727,14 +1728,18 @@ out_acl: WRITE32(stat.nlink); } if (bmval1 & FATTR4_WORD1_OWNER) { - status = nfsd4_encode_user(rqstp, stat.uid, &p, &buflen); + status = nfsd4_encode_user(rqstp, + TAGINO_UID(DX_TAG(dentry->d_inode), + stat.uid, stat.tag), &p, &buflen); if (status == nfserr_resource) goto out_resource; if (status) goto out; } if (bmval1 & FATTR4_WORD1_OWNER_GROUP) { - status = nfsd4_encode_group(rqstp, stat.gid, &p, &buflen); + status = nfsd4_encode_group(rqstp, + TAGINO_GID(DX_TAG(dentry->d_inode), + stat.gid, stat.tag), &p, &buflen); if (status == nfserr_resource) goto out_resource; if (status) diff -NurpP --minimal linux-source-2.6.22.orig/fs/nfsd/nfsxdr.c linux-source-2.6.22/fs/nfsd/nfsxdr.c --- linux-source-2.6.22.orig/fs/nfsd/nfsxdr.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/nfsd/nfsxdr.c 2007-10-24 16:14:06.291582822 +0200 @@ -15,6 +15,7 @@ #include #include #include +#include #define NFSDDBG_FACILITY NFSDDBG_XDR @@ -97,6 +98,8 @@ static __be32 * decode_sattr(__be32 *p, struct iattr *iap) { u32 tmp, tmp1; + uid_t uid = 0; + gid_t gid = 0; iap->ia_valid = 0; @@ -110,12 +113,15 @@ decode_sattr(__be32 *p, struct iattr *ia } if ((tmp = ntohl(*p++)) != (u32)-1) { iap->ia_valid |= ATTR_UID; - iap->ia_uid = tmp; + uid = tmp; } if ((tmp = ntohl(*p++)) != (u32)-1) { iap->ia_valid |= ATTR_GID; - iap->ia_gid = tmp; + gid = tmp; } + iap->ia_uid = INOTAG_UID(DX_TAG_NFSD, uid, gid); + iap->ia_gid = INOTAG_GID(DX_TAG_NFSD, uid, gid); + iap->ia_tag = INOTAG_TAG(DX_TAG_NFSD, uid, gid, 0); if ((tmp = ntohl(*p++)) != (u32)-1) { iap->ia_valid |= ATTR_SIZE; iap->ia_size = tmp; @@ -160,8 +166,10 @@ encode_fattr(struct svc_rqst *rqstp, __b *p++ = htonl(nfs_ftypes[type >> 12]); *p++ = htonl((u32) stat->mode); *p++ = htonl((u32) stat->nlink); - *p++ = htonl((u32) nfsd_ruid(rqstp, stat->uid)); - *p++ = htonl((u32) nfsd_rgid(rqstp, stat->gid)); + *p++ = htonl((u32) nfsd_ruid(rqstp, + TAGINO_UID(DX_TAG(dentry->d_inode), stat->uid, stat->tag))); + *p++ = htonl((u32) nfsd_rgid(rqstp, + TAGINO_GID(DX_TAG(dentry->d_inode), stat->gid, stat->tag))); if (S_ISLNK(type) && stat->size > NFS_MAXPATHLEN) { *p++ = htonl(NFS_MAXPATHLEN); diff -NurpP --minimal linux-source-2.6.22.orig/fs/nfsd/vfs.c linux-source-2.6.22/fs/nfsd/vfs.c --- linux-source-2.6.22.orig/fs/nfsd/vfs.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/fs/nfsd/vfs.c 2007-10-24 16:14:06.291582822 +0200 @@ -1195,14 +1195,14 @@ nfsd_create(struct svc_rqst *rqstp, stru host_err = vfs_create(dirp, dchild, iap->ia_mode, NULL); break; case S_IFDIR: - host_err = vfs_mkdir(dirp, dchild, exp->ex_mnt, iap->ia_mode); + host_err = vfs_mkdir(dirp, dchild, exp->ex_mnt, iap->ia_mode, NULL); break; case S_IFCHR: case S_IFBLK: case S_IFIFO: case S_IFSOCK: host_err = vfs_mknod(dirp, dchild, exp->ex_mnt, iap->ia_mode, - rdev); + rdev, NULL); break; default: printk("nfsd: bad file type %o in nfsd_create\n", type); @@ -1479,12 +1479,12 @@ nfsd_symlink(struct svc_rqst *rqstp, str strncpy(path_alloced, path, plen); path_alloced[plen] = 0; host_err = vfs_symlink(dentry->d_inode, dnew, - exp->ex_mnt, path_alloced, mode); + exp->ex_mnt, path_alloced, mode, NULL); kfree(path_alloced); } } else host_err = vfs_symlink(dentry->d_inode, dnew, exp->ex_mnt, path, - mode); + mode, NULL); if (!host_err) { if (EX_ISSYNC(exp)) @@ -1544,7 +1544,7 @@ nfsd_link(struct svc_rqst *rqstp, struct dest = dold->d_inode; host_err = vfs_link(dold, tfhp->fh_export->ex_mnt, dirp, - dnew, ffhp->fh_export->ex_mnt); + dnew, ffhp->fh_export->ex_mnt, NULL); if (!host_err) { if (EX_ISSYNC(ffhp->fh_export)) { err = nfserrno(nfsd_sync_dir(ddir)); @@ -1712,9 +1712,9 @@ nfsd_unlink(struct svc_rqst *rqstp, stru host_err = -EPERM; } else #endif - host_err = vfs_unlink(dirp, rdentry, exp->ex_mnt); + host_err = vfs_unlink(dirp, rdentry, exp->ex_mnt, NULL); } else { /* It's RMDIR */ - host_err = vfs_rmdir(dirp, rdentry, exp->ex_mnt); + host_err = vfs_rmdir(dirp, rdentry, exp->ex_mnt, NULL); } dput(rdentry); @@ -1825,7 +1825,8 @@ nfsd_permission(struct svc_export *exp, */ if (!(acc & MAY_LOCAL_ACCESS)) if (acc & (MAY_WRITE | MAY_SATTR | MAY_TRUNC)) { - if (EX_RDONLY(exp) || IS_RDONLY(inode)) + if (EX_RDONLY(exp) || IS_RDONLY(inode) + || MNT_IS_RDONLY(exp->ex_mnt)) return nfserr_rofs; if (/* (acc & MAY_WRITE) && */ IS_IMMUTABLE(inode)) return nfserr_perm; diff -NurpP --minimal linux-source-2.6.22.orig/fs/ocfs2/dlm/dlmfs.c linux-source-2.6.22/fs/ocfs2/dlm/dlmfs.c --- linux-source-2.6.22.orig/fs/ocfs2/dlm/dlmfs.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ocfs2/dlm/dlmfs.c 2007-10-24 16:14:06.291582822 +0200 @@ -43,6 +43,7 @@ #include #include #include +#include #include @@ -331,6 +332,7 @@ static struct inode *dlmfs_get_root_inod inode->i_mode = mode; inode->i_uid = current->fsuid; inode->i_gid = current->fsgid; + inode->i_tag = dx_current_fstag(sb); inode->i_blocks = 0; inode->i_mapping->backing_dev_info = &dlmfs_backing_dev_info; inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME; @@ -357,6 +359,7 @@ static struct inode *dlmfs_get_inode(str inode->i_mode = mode; inode->i_uid = current->fsuid; inode->i_gid = current->fsgid; + inode->i_tag = dx_current_fstag(sb); inode->i_blocks = 0; inode->i_mapping->backing_dev_info = &dlmfs_backing_dev_info; inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME; diff -NurpP --minimal linux-source-2.6.22.orig/fs/ocfs2/dlmglue.c linux-source-2.6.22/fs/ocfs2/dlmglue.c --- linux-source-2.6.22.orig/fs/ocfs2/dlmglue.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ocfs2/dlmglue.c 2007-10-24 16:14:06.291582822 +0200 @@ -1475,6 +1475,7 @@ static void __ocfs2_stuff_meta_lvb(struc lvb->lvb_iclusters = cpu_to_be32(oi->ip_clusters); lvb->lvb_iuid = cpu_to_be32(inode->i_uid); lvb->lvb_igid = cpu_to_be32(inode->i_gid); + lvb->lvb_itag = cpu_to_be16(inode->i_tag); lvb->lvb_imode = cpu_to_be16(inode->i_mode); lvb->lvb_inlink = cpu_to_be16(inode->i_nlink); lvb->lvb_iatime_packed = @@ -1527,6 +1528,7 @@ static void ocfs2_refresh_inode_from_lvb inode->i_uid = be32_to_cpu(lvb->lvb_iuid); inode->i_gid = be32_to_cpu(lvb->lvb_igid); + inode->i_tag = be16_to_cpu(lvb->lvb_itag); inode->i_mode = be16_to_cpu(lvb->lvb_imode); inode->i_nlink = be16_to_cpu(lvb->lvb_inlink); ocfs2_unpack_timespec(&inode->i_atime, diff -NurpP --minimal linux-source-2.6.22.orig/fs/ocfs2/dlmglue.h linux-source-2.6.22/fs/ocfs2/dlmglue.h --- linux-source-2.6.22.orig/fs/ocfs2/dlmglue.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ocfs2/dlmglue.h 2007-10-24 16:14:06.291582822 +0200 @@ -34,7 +34,7 @@ struct ocfs2_meta_lvb { __u8 lvb_version; __u8 lvb_reserved0; - __be16 lvb_reserved1; + __be16 lvb_itag; __be32 lvb_iclusters; __be32 lvb_iuid; __be32 lvb_igid; diff -NurpP --minimal linux-source-2.6.22.orig/fs/ocfs2/file.c linux-source-2.6.22/fs/ocfs2/file.c --- linux-source-2.6.22.orig/fs/ocfs2/file.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/fs/ocfs2/file.c 2007-10-24 16:14:06.291582822 +0200 @@ -943,13 +943,15 @@ int ocfs2_setattr(struct dentry *dentry, mlog(0, "uid change: %d\n", attr->ia_uid); if (attr->ia_valid & ATTR_GID) mlog(0, "gid change: %d\n", attr->ia_gid); + if (attr->ia_valid & ATTR_TAG) + mlog(0, "tag change: %d\n", attr->ia_tag); if (attr->ia_valid & ATTR_SIZE) mlog(0, "size change...\n"); if (attr->ia_valid & (ATTR_ATIME | ATTR_MTIME | ATTR_CTIME)) mlog(0, "time change...\n"); #define OCFS2_VALID_ATTRS (ATTR_ATIME | ATTR_MTIME | ATTR_CTIME | ATTR_SIZE \ - | ATTR_GID | ATTR_UID | ATTR_MODE) + | ATTR_GID | ATTR_UID | ATTR_TAG | ATTR_MODE) if (!(attr->ia_valid & OCFS2_VALID_ATTRS)) { mlog(0, "can't handle attrs: 0x%x\n", attr->ia_valid); return 0; @@ -1805,6 +1807,7 @@ bail: const struct inode_operations ocfs2_file_iops = { .setattr = ocfs2_setattr, .getattr = ocfs2_getattr, + .sync_flags = ocfs2_sync_flags, .permission = ocfs2_permission, }; diff -NurpP --minimal linux-source-2.6.22.orig/fs/ocfs2/inode.c linux-source-2.6.22/fs/ocfs2/inode.c --- linux-source-2.6.22.orig/fs/ocfs2/inode.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ocfs2/inode.c 2007-10-24 16:14:06.291582822 +0200 @@ -28,6 +28,7 @@ #include #include #include +#include #include @@ -42,6 +43,7 @@ #include "file.h" #include "heartbeat.h" #include "inode.h" +#include "ioctl.h" #include "journal.h" #include "namei.h" #include "suballoc.h" @@ -77,6 +79,10 @@ void ocfs2_set_inode_flags(struct inode if (flags & OCFS2_IMMUTABLE_FL) inode->i_flags |= S_IMMUTABLE; + if (flags & OCFS2_IUNLINK_FL) + inode->i_flags |= S_IUNLINK; + if (flags & OCFS2_BARRIER_FL) + inode->i_flags |= S_BARRIER; if (flags & OCFS2_SYNC_FL) inode->i_flags |= S_SYNC; @@ -107,6 +113,27 @@ void ocfs2_get_inode_flags(struct ocfs2_ oi->ip_attr |= OCFS2_DIRSYNC_FL; } +int ocfs2_sync_flags(struct inode *inode) +{ + unsigned int oldflags, newflags; + + oldflags = OCFS2_I(inode)->ip_flags; + newflags = oldflags & ~(OCFS2_IMMUTABLE_FL | + OCFS2_IUNLINK_FL | OCFS2_BARRIER_FL); + + if (IS_IMMUTABLE(inode)) + newflags |= OCFS2_IMMUTABLE_FL; + if (IS_IUNLINK(inode)) + newflags |= OCFS2_IUNLINK_FL; + if (IS_BARRIER(inode)) + newflags |= OCFS2_BARRIER_FL; + + if (oldflags ^ newflags) + return ocfs2_set_inode_attr(inode, + newflags, OCFS2_FL_MASK); + return 0; +} + struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 blkno, int flags) { struct inode *inode = NULL; @@ -212,6 +239,8 @@ int ocfs2_populate_inode(struct inode *i struct super_block *sb; struct ocfs2_super *osb; int status = -EINVAL; + uid_t uid; + gid_t gid; mlog_entry("(0x%p, size:%llu)\n", inode, (unsigned long long)le64_to_cpu(fe->i_size)); @@ -246,8 +275,12 @@ int ocfs2_populate_inode(struct inode *i inode->i_generation = le32_to_cpu(fe->i_generation); inode->i_rdev = huge_decode_dev(le64_to_cpu(fe->id1.dev1.i_rdev)); inode->i_mode = le16_to_cpu(fe->i_mode); - inode->i_uid = le32_to_cpu(fe->i_uid); - inode->i_gid = le32_to_cpu(fe->i_gid); + uid = le32_to_cpu(fe->i_uid); + gid = le32_to_cpu(fe->i_gid); + inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid); + inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid); + inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid, + /* le16_to_cpu(raw_inode->i_raw_tag)i */ 0); /* Fast symlinks will have i_size but no allocated clusters. */ if (S_ISLNK(inode->i_mode) && !fe->i_clusters) @@ -1224,8 +1257,11 @@ int ocfs2_mark_inode_dirty(handle_t *han fe->i_size = cpu_to_le64(i_size_read(inode)); fe->i_links_count = cpu_to_le16(inode->i_nlink); - fe->i_uid = cpu_to_le32(inode->i_uid); - fe->i_gid = cpu_to_le32(inode->i_gid); + fe->i_uid = cpu_to_le32(TAGINO_UID(DX_TAG(inode), + inode->i_uid, inode->i_tag)); + fe->i_gid = cpu_to_le32(TAGINO_GID(DX_TAG(inode), + inode->i_gid, inode->i_tag)); + /* i_tag = = cpu_to_le16(inode->i_tag); */ fe->i_mode = cpu_to_le16(inode->i_mode); fe->i_atime = cpu_to_le64(inode->i_atime.tv_sec); fe->i_atime_nsec = cpu_to_le32(inode->i_atime.tv_nsec); @@ -1253,15 +1289,24 @@ leave: void ocfs2_refresh_inode(struct inode *inode, struct ocfs2_dinode *fe) { + uid_t uid; + gid_t gid; + spin_lock(&OCFS2_I(inode)->ip_lock); OCFS2_I(inode)->ip_clusters = le32_to_cpu(fe->i_clusters); OCFS2_I(inode)->ip_attr = le32_to_cpu(fe->i_attr); + /* OCFS2_I(inode)->ip_flags &= ~OCFS2_FL_MASK; + OCFS2_I(inode)->ip_flags |= le32_to_cpu(fe->i_flags) & OCFS2_FL_MASK; */ ocfs2_set_inode_flags(inode); i_size_write(inode, le64_to_cpu(fe->i_size)); inode->i_nlink = le16_to_cpu(fe->i_links_count); - inode->i_uid = le32_to_cpu(fe->i_uid); - inode->i_gid = le32_to_cpu(fe->i_gid); + uid = le32_to_cpu(fe->i_uid); + gid = le32_to_cpu(fe->i_gid); + inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid); + inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid); + inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid, + /* le16_to_cpu(raw_inode->i_raw_tag)i */ 0); inode->i_mode = le16_to_cpu(fe->i_mode); if (S_ISLNK(inode->i_mode) && le32_to_cpu(fe->i_clusters) == 0) inode->i_blocks = 0; diff -NurpP --minimal linux-source-2.6.22.orig/fs/ocfs2/inode.h linux-source-2.6.22/fs/ocfs2/inode.h --- linux-source-2.6.22.orig/fs/ocfs2/inode.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ocfs2/inode.h 2007-10-24 16:14:06.291582822 +0200 @@ -142,6 +142,7 @@ int ocfs2_aio_write(struct file *file, s void ocfs2_set_inode_flags(struct inode *inode); void ocfs2_get_inode_flags(struct ocfs2_inode_info *oi); +int ocfs2_sync_flags(struct inode *inode); static inline blkcnt_t ocfs2_inode_sector_count(struct inode *inode) { diff -NurpP --minimal linux-source-2.6.22.orig/fs/ocfs2/ioctl.c linux-source-2.6.22/fs/ocfs2/ioctl.c --- linux-source-2.6.22.orig/fs/ocfs2/ioctl.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ocfs2/ioctl.c 2007-10-24 16:14:06.291582822 +0200 @@ -39,7 +39,7 @@ static int ocfs2_get_inode_attr(struct i return status; } -static int ocfs2_set_inode_attr(struct inode *inode, unsigned flags, +int ocfs2_set_inode_attr(struct inode *inode, unsigned flags, unsigned mask) { struct ocfs2_inode_info *ocfs2_inode = OCFS2_I(inode); diff -NurpP --minimal linux-source-2.6.22.orig/fs/ocfs2/ioctl.h linux-source-2.6.22/fs/ocfs2/ioctl.h --- linux-source-2.6.22.orig/fs/ocfs2/ioctl.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ocfs2/ioctl.h 2007-10-24 16:14:06.291582822 +0200 @@ -10,6 +10,9 @@ #ifndef OCFS2_IOCTL_H #define OCFS2_IOCTL_H +int ocfs2_set_inode_attr(struct inode *inode, unsigned flags, + unsigned mask); + int ocfs2_ioctl(struct inode * inode, struct file * filp, unsigned int cmd, unsigned long arg); long ocfs2_compat_ioctl(struct file *file, unsigned cmd, unsigned long arg); diff -NurpP --minimal linux-source-2.6.22.orig/fs/ocfs2/namei.c linux-source-2.6.22/fs/ocfs2/namei.c --- linux-source-2.6.22.orig/fs/ocfs2/namei.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ocfs2/namei.c 2007-10-24 16:14:06.291582822 +0200 @@ -40,6 +40,7 @@ #include #include #include +#include #define MLOG_MASK_PREFIX ML_NAMEI #include @@ -483,6 +484,9 @@ static int ocfs2_mknod_locked(struct ocf u64 fe_blkno = 0; u16 suballoc_bit; struct inode *inode = NULL; + uid_t uid; + gid_t gid; + tag_t tag; mlog_entry("(0x%p, 0x%p, %d, %lu, '%.*s')\n", dir, dentry, mode, (unsigned long)dev, dentry->d_name.len, @@ -542,13 +546,19 @@ static int ocfs2_mknod_locked(struct ocf fe->i_blkno = cpu_to_le64(fe_blkno); fe->i_suballoc_bit = cpu_to_le16(suballoc_bit); fe->i_suballoc_slot = cpu_to_le16(osb->slot_num); - fe->i_uid = cpu_to_le32(current->fsuid); + + tag = dx_current_fstag(osb->sb); + uid = current->fsuid; if (dir->i_mode & S_ISGID) { - fe->i_gid = cpu_to_le32(dir->i_gid); + gid = dir->i_gid; if (S_ISDIR(mode)) mode |= S_ISGID; } else - fe->i_gid = cpu_to_le32(current->fsgid); + gid = current->fsgid; + + fe->i_uid = cpu_to_le32(TAGINO_UID(DX_TAG(inode), uid, tag)); + fe->i_gid = cpu_to_le32(TAGINO_GID(DX_TAG(inode), gid, tag)); + inode->i_tag = tag; fe->i_mode = cpu_to_le16(mode); if (S_ISCHR(mode) || S_ISBLK(mode)) fe->id1.dev1.i_rdev = cpu_to_le64(huge_encode_dev(dev)); @@ -2316,5 +2326,6 @@ const struct inode_operations ocfs2_dir_ .rename = ocfs2_rename, .setattr = ocfs2_setattr, .getattr = ocfs2_getattr, + .sync_flags = ocfs2_sync_flags, .permission = ocfs2_permission, }; diff -NurpP --minimal linux-source-2.6.22.orig/fs/ocfs2/ocfs2_fs.h linux-source-2.6.22/fs/ocfs2/ocfs2_fs.h --- linux-source-2.6.22.orig/fs/ocfs2/ocfs2_fs.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ocfs2/ocfs2_fs.h 2007-10-24 16:14:06.291582822 +0200 @@ -152,8 +152,12 @@ #define OCFS2_NOATIME_FL (0x00000080) /* do not update atime */ #define OCFS2_DIRSYNC_FL (0x00010000) /* dirsync behaviour (directories only) */ +#define OCFS2_BARRIER_FL (0x04000000) /* Barrier for chroot() */ +#define OCFS2_IUNLINK_FL (0x08000000) /* Immutable unlink */ + #define OCFS2_FL_VISIBLE (0x000100FF) /* User visible flags */ #define OCFS2_FL_MODIFIABLE (0x000100FF) /* User modifiable flags */ +#define OCFS2_FL_MASK (0x0F0100FF) /* * Extent record flags (e_node.leaf.flags) diff -NurpP --minimal linux-source-2.6.22.orig/fs/ocfs2/ocfs2.h linux-source-2.6.22/fs/ocfs2/ocfs2.h --- linux-source-2.6.22.orig/fs/ocfs2/ocfs2.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ocfs2/ocfs2.h 2007-10-24 16:14:06.291582822 +0200 @@ -170,6 +170,7 @@ enum ocfs2_mount_options OCFS2_MOUNT_NOINTR = 1 << 2, /* Don't catch signals */ OCFS2_MOUNT_ERRORS_PANIC = 1 << 3, /* Panic on errors */ OCFS2_MOUNT_DATA_WRITEBACK = 1 << 4, /* No data ordering */ + OCFS2_MOUNT_TAGGED = 1 << 8, /* use tagging */ }; #define OCFS2_OSB_SOFT_RO 0x0001 diff -NurpP --minimal linux-source-2.6.22.orig/fs/ocfs2/super.c linux-source-2.6.22/fs/ocfs2/super.c --- linux-source-2.6.22.orig/fs/ocfs2/super.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/ocfs2/super.c 2007-10-24 16:14:06.291582822 +0200 @@ -140,6 +140,7 @@ enum { Opt_data_ordered, Opt_data_writeback, Opt_atime_quantum, + Opt_tag, Opt_notag, Opt_tagid, Opt_err, }; @@ -154,6 +155,10 @@ static match_table_t tokens = { {Opt_data_ordered, "data=ordered"}, {Opt_data_writeback, "data=writeback"}, {Opt_atime_quantum, "atime_quantum=%u"}, + {Opt_tag, "tag"}, + {Opt_tag, "tagxid"}, + {Opt_notag, "notag"}, + {Opt_tagid, "tagid=%u"}, {Opt_err, NULL} }; @@ -362,6 +367,14 @@ static int ocfs2_remount(struct super_bl goto out; } + printk("ocfs2_remount: %lx,%lx\n", osb->s_mount_opt, sb->s_flags); + if ((parsed_options & OCFS2_MOUNT_TAGGED) && + !(sb->s_flags & MS_TAGGED)) { + ret = -EINVAL; + mlog(ML_ERROR, "Cannot change tagging on remount\n"); + goto out; + } + if ((osb->s_mount_opt & OCFS2_MOUNT_HB_LOCAL) != (parsed_options & OCFS2_MOUNT_HB_LOCAL)) { ret = -EINVAL; @@ -654,6 +667,9 @@ static int ocfs2_fill_super(struct super ocfs2_complete_mount_recovery(osb); + if (osb->s_mount_opt & OCFS2_MOUNT_TAGGED) + sb->s_flags |= MS_TAGGED; + if (ocfs2_mount_local(osb)) snprintf(nodestr, sizeof(nodestr), "local"); else @@ -782,6 +798,20 @@ static int ocfs2_parse_options(struct su else osb->s_atime_quantum = OCFS2_DEFAULT_ATIME_QUANTUM; break; +#ifndef CONFIG_TAGGING_NONE + case Opt_tag: + *mount_opt |= OCFS2_MOUNT_TAGGED; + break; + case Opt_notag: + *mount_opt &= ~OCFS2_MOUNT_TAGGED; + break; +#endif +#ifdef CONFIG_PROPAGATE + case Opt_tagid: + /* use args[0] */ + *mount_opt |= OCFS2_MOUNT_TAGGED; + break; +#endif default: mlog(ML_ERROR, "Unrecognized mount option \"%s\" " diff -NurpP --minimal linux-source-2.6.22.orig/fs/open.c linux-source-2.6.22/fs/open.c --- linux-source-2.6.22.orig/fs/open.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/fs/open.c 2007-10-24 16:14:06.291582822 +0200 @@ -26,22 +26,31 @@ #include #include #include +#include +#include +#include +#include +#include int vfs_statfs(struct dentry *dentry, struct kstatfs *buf) { int retval = -ENODEV; if (dentry) { + struct super_block *sb = dentry->d_sb; + retval = -ENOSYS; - if (dentry->d_sb->s_op->statfs) { + if (sb->s_op->statfs) { memset(buf, 0, sizeof(*buf)); retval = security_sb_statfs(dentry); if (retval) return retval; - retval = dentry->d_sb->s_op->statfs(dentry, buf); + retval = sb->s_op->statfs(dentry, buf); if (retval == 0 && buf->f_frsize == 0) buf->f_frsize = buf->f_bsize; } + if (!vx_check(0, VS_ADMIN|VS_WATCH)) + vx_vsi_statfs(sb, buf); } return retval; } @@ -248,7 +257,7 @@ static long do_sys_truncate(const char _ goto dput_and_out; error = -EROFS; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || MNT_IS_RDONLY(nd.mnt)) goto dput_and_out; error = -EPERM; @@ -398,7 +407,7 @@ asmlinkage long sys_faccessat(int dfd, c special_file(nd.dentry->d_inode->i_mode)) goto out_path_release; - if(IS_RDONLY(nd.dentry->d_inode)) + if(IS_RDONLY(nd.dentry->d_inode) || MNT_IS_RDONLY(nd.mnt)) res = -EROFS; out_path_release: @@ -510,7 +519,7 @@ asmlinkage long sys_fchmod(unsigned int audit_inode(NULL, inode); err = -EROFS; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || MNT_IS_RDONLY(file->f_vfsmnt)) goto out_putf; err = -EPERM; if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) @@ -542,11 +551,11 @@ asmlinkage long sys_fchmodat(int dfd, co error = __user_walk_fd(dfd, filename, LOOKUP_FOLLOW, &nd); if (error) goto out; - inode = nd.dentry->d_inode; - error = -EROFS; - if (IS_RDONLY(inode)) + error = cow_check_and_break(&nd); + if (error) goto dput_and_out; + inode = nd.dentry->d_inode; error = -EPERM; if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) @@ -584,7 +593,7 @@ static int chown_common(struct dentry * goto out; } error = -EROFS; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || MNT_IS_RDONLY(mnt)) goto out; error = -EPERM; if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) @@ -592,11 +601,11 @@ static int chown_common(struct dentry * newattrs.ia_valid = ATTR_CTIME; if (user != (uid_t) -1) { newattrs.ia_valid |= ATTR_UID; - newattrs.ia_uid = user; + newattrs.ia_uid = dx_map_uid(user); } if (group != (gid_t) -1) { newattrs.ia_valid |= ATTR_GID; - newattrs.ia_gid = group; + newattrs.ia_gid = dx_map_gid(group); } if (!S_ISDIR(inode->i_mode)) newattrs.ia_valid |= ATTR_KILL_SUID|ATTR_KILL_SGID; @@ -619,7 +628,11 @@ asmlinkage long sys_chown(const char __u error = user_path_walk(filename, &nd); if (error) goto out; - error = chown_common(nd.dentry, nd.mnt, user, group, NULL); +#ifdef CONFIG_VSERVER_COWBL + error = cow_check_and_break(&nd); + if (!error) +#endif + error = chown_common(nd.dentry, nd.mnt, user, group, NULL); path_release(&nd); out: return error; @@ -639,7 +652,11 @@ asmlinkage long sys_fchownat(int dfd, co error = __user_walk_fd(dfd, filename, follow, &nd); if (error) goto out; - error = chown_common(nd.dentry, nd.mnt, user, group, NULL); +#ifdef CONFIG_VSERVER_COWBL + error = cow_check_and_break(&nd); + if (!error) +#endif + error = chown_common(nd.dentry, nd.mnt, user, group, NULL); path_release(&nd); out: return error; @@ -653,7 +670,11 @@ asmlinkage long sys_lchown(const char __ error = user_path_walk_link(filename, &nd); if (error) goto out; - error = chown_common(nd.dentry, nd.mnt, user, group, NULL); +#ifdef CONFIG_VSERVER_COWBL + error = cow_check_and_break(&nd); + if (!error) +#endif + error = chown_common(nd.dentry, nd.mnt, user, group, NULL); path_release(&nd); out: return error; @@ -899,6 +920,7 @@ repeat: FD_SET(fd, fdt->open_fds); FD_CLR(fd, fdt->close_on_exec); files->next_fd = fd + 1; + vx_openfd_inc(fd); #if 1 /* Sanity check */ if (fdt->fd[fd] != NULL) { @@ -921,6 +943,7 @@ static void __put_unused_fd(struct files __FD_CLR(fd, fdt->open_fds); if (fd < files->next_fd) files->next_fd = fd; + vx_openfd_dec(fd); } void fastcall put_unused_fd(unsigned int fd) diff -NurpP --minimal linux-source-2.6.22.orig/fs/proc/array.c linux-source-2.6.22/fs/proc/array.c --- linux-source-2.6.22.orig/fs/proc/array.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/proc/array.c 2007-10-24 16:14:06.291582822 +0200 @@ -75,6 +75,8 @@ #include #include #include +#include +#include #include #include @@ -134,8 +136,9 @@ static const char *task_state_array[] = "D (disk sleep)", /* 2 */ "T (stopped)", /* 4 */ "T (tracing stop)", /* 8 */ - "Z (zombie)", /* 16 */ - "X (dead)" /* 32 */ + "H (on hold)", /* 16 */ + "Z (zombie)", /* 32 */ + "X (dead)", /* 64 */ }; static inline const char * get_task_state(struct task_struct *tsk) @@ -144,7 +147,8 @@ static inline const char * get_task_stat TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE | TASK_STOPPED | - TASK_TRACED)) | + TASK_TRACED | + TASK_ONHOLD)) | (tsk->exit_state & (EXIT_ZOMBIE | EXIT_DEAD)); const char **p = &task_state_array[0]; @@ -161,8 +165,16 @@ static inline char * task_state(struct t struct group_info *group_info; int g; struct fdtable *fdt = NULL; + pid_t pid, ptgid, tppid, tgid; rcu_read_lock(); + tgid = vx_map_tgid(p->tgid); + pid = vx_map_pid(p->pid); + ptgid = vx_map_pid(pid_alive(p) ? + rcu_dereference(p->real_parent)->tgid : 0); + tppid = vx_map_pid(pid_alive(p) && p->ptrace ? + rcu_dereference(p->parent)->pid : 0); + buffer += sprintf(buffer, "State:\t%s\n" "SleepAVG:\t%lu%%\n" @@ -174,9 +186,7 @@ static inline char * task_state(struct t "Gid:\t%d\t%d\t%d\t%d\n", get_task_state(p), (p->sleep_avg/1024)*100/(1020000000/1024), - p->tgid, p->pid, - pid_alive(p) ? rcu_dereference(p->real_parent)->tgid : 0, - pid_alive(p) && p->ptrace ? rcu_dereference(p->parent)->pid : 0, + tgid, pid, (pid > 1) ? ptgid : 0, tppid, p->uid, p->euid, p->suid, p->fsuid, p->gid, p->egid, p->sgid, p->fsgid); @@ -283,17 +293,26 @@ static inline char * task_sig(struct tas static inline char *task_cap(struct task_struct *p, char *buffer) { - return buffer + sprintf(buffer, "CapInh:\t%016x\n" - "CapPrm:\t%016x\n" - "CapEff:\t%016x\n", - cap_t(p->cap_inheritable), - cap_t(p->cap_permitted), - cap_t(p->cap_effective)); + struct vx_info *vxi = p->vx_info; + + return buffer + sprintf(buffer, + "CapInh:\t%016x\n" + "CapPrm:\t%016x\n" + "CapEff:\t%016x\n", + (unsigned)vx_info_mbcap(vxi, p->cap_inheritable), + (unsigned)vx_info_mbcap(vxi, p->cap_permitted), + (unsigned)vx_info_mbcap(vxi, p->cap_effective)); } int proc_pid_status(struct task_struct *task, char * buffer) { char * orig = buffer; +#ifdef CONFIG_VSERVER_LEGACY + struct vx_info *vxi; +#endif +#ifdef CONFIG_VSERVER_LEGACYNET + struct nx_info *nxi; +#endif struct mm_struct *mm = get_task_mm(task); buffer = task_name(task, buffer); @@ -306,6 +325,46 @@ int proc_pid_status(struct task_struct * buffer = task_sig(task, buffer); buffer = task_cap(task, buffer); buffer = cpuset_task_status_allowed(task, buffer); + + if (task_vx_flags(task, VXF_HIDE_VINFO, 0)) + goto skip; +#ifdef CONFIG_VSERVER_LEGACY + buffer += sprintf (buffer,"s_context: %d\n", vx_task_xid(task)); + vxi = task_get_vx_info(task); + if (vxi) { + buffer += sprintf (buffer,"ctxflags: %08llx\n" + ,(unsigned long long)vxi->vx_flags); + buffer += sprintf (buffer,"initpid: %d\n" + ,vxi->vx_initpid); + } else { + buffer += sprintf (buffer,"ctxflags: none\n"); + buffer += sprintf (buffer,"initpid: none\n"); + } + put_vx_info(vxi); +#else + buffer += sprintf (buffer,"VxID: %d\n", vx_task_xid(task)); +#endif +#ifdef CONFIG_VSERVER_LEGACYNET + nxi = task_get_nx_info(task); + if (nxi) { + int i; + + buffer += sprintf (buffer,"ipv4root:"); + for (i=0; inbipv4; i++){ + buffer += sprintf (buffer," %08x/%08x" + ,nxi->ipv4[i] + ,nxi->mask[i]); + } + *buffer++ = '\n'; + buffer += sprintf (buffer,"ipv4root_bcast: %08x\n" + ,nxi->v4_bcast); + } else { + buffer += sprintf (buffer,"ipv4root: 0\n"); + buffer += sprintf (buffer,"ipv4root_bcast: 0\n"); + } + put_nx_info(nxi); +#endif +skip: #if defined(CONFIG_S390) buffer = task_show_regs(task, buffer); #endif @@ -320,7 +379,7 @@ static int do_task_stat(struct task_stru sigset_t sigign, sigcatch; char state; int res; - pid_t ppid = 0, pgid = -1, sid = -1; + pid_t pid = 0, ppid = 0, pgid = -1, sid = -1; int num_threads = 0; struct mm_struct *mm; unsigned long long start_time; @@ -382,8 +441,10 @@ static int do_task_stat(struct task_stru } sid = signal_session(sig); - pgid = process_group(task); - ppid = rcu_dereference(task->real_parent)->tgid; + pid = vx_info_map_pid(task->vx_info, task->pid); + pgid = vx_info_map_pid(task->vx_info, process_group(task)); + ppid = (pid > 1) ? vx_info_map_tgid(task->vx_info, + rcu_dereference(task->real_parent)->tgid) : 0; unlock_task_sighand(task, &flags); } @@ -410,10 +471,21 @@ static int do_task_stat(struct task_stru /* convert nsec -> ticks */ start_time = nsec_to_clock_t(start_time); + /* fixup start time for virt uptime */ + if (vx_flags(VXF_VIRT_UPTIME, 0)) { + unsigned long long bias = + current->vx_info->cvirt.bias_clock; + + if (start_time > bias) + start_time -= bias; + else + start_time = 0; + } + res = sprintf(buffer,"%d (%s) %c %d %d %d %d %d %u %lu \ %lu %lu %lu %lu %lu %ld %ld %ld %ld %d 0 %llu %lu %ld %lu %lu %lu %lu %lu \ %lu %lu %lu %lu %lu %lu %lu %lu %d %d %u %u %llu\n", - task->pid, + pid, tcomm, state, ppid, diff -NurpP --minimal linux-source-2.6.22.orig/fs/proc/base.c linux-source-2.6.22/fs/proc/base.c --- linux-source-2.6.22.orig/fs/proc/base.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/proc/base.c 2007-10-24 16:14:06.291582822 +0200 @@ -73,6 +73,9 @@ #include #include #include +#include +#include + #include "internal.h" /* NOTE: @@ -1049,6 +1052,8 @@ static struct inode *proc_pid_make_inode inode->i_uid = task->euid; inode->i_gid = task->egid; } + /* procfs is xid tagged */ + inode->i_tag = (tag_t)vx_task_xid(task); security_task_to_inode(task, inode); out: @@ -1595,6 +1600,13 @@ static struct dentry *proc_pident_lookup if (!task) goto out_no_task; + /* TODO: maybe we can come up with a generic approach? */ + if (task_vx_flags(task, VXF_HIDE_VINFO, 0) && + (dentry->d_name.len == 5) && + (!memcmp(dentry->d_name.name, "vinfo", 5) || + !memcmp(dentry->d_name.name, "ninfo", 5))) + goto out; + /* * Yes, it does not scale. And it should not. Don't add * new entries into /proc// without very good reasons. @@ -1790,14 +1802,14 @@ static int proc_self_readlink(struct den int buflen) { char tmp[PROC_NUMBUF]; - sprintf(tmp, "%d", current->tgid); + sprintf(tmp, "%d", vx_map_tgid(current->tgid)); return vfs_readlink(dentry,buffer,buflen,tmp); } static void *proc_self_follow_link(struct dentry *dentry, struct nameidata *nd) { char tmp[PROC_NUMBUF]; - sprintf(tmp, "%d", current->tgid); + sprintf(tmp, "%d", vx_map_tgid(current->tgid)); return ERR_PTR(vfs_follow_link(nd,tmp)); } @@ -1891,7 +1903,7 @@ out_iput: static struct dentry *proc_base_lookup(struct inode *dir, struct dentry *dentry) { struct dentry *error; - struct task_struct *task = get_proc_task(dir); + struct task_struct *task = get_proc_task_real(dir); const struct pid_entry *p, *last; error = ERR_PTR(-ENOENT); @@ -1956,6 +1968,9 @@ static int proc_pid_io_accounting(struct static const struct file_operations proc_task_operations; static const struct inode_operations proc_task_inode_operations; +extern int proc_pid_vx_info(struct task_struct *, char *); +extern int proc_pid_nx_info(struct task_struct *, char *); + static const struct pid_entry tgid_base_stuff[] = { DIR("task", S_IRUGO|S_IXUGO, task), DIR("fd", S_IRUSR|S_IXUSR, fd), @@ -1995,6 +2010,8 @@ static const struct pid_entry tgid_base_ #ifdef CONFIG_CPUSETS REG("cpuset", S_IRUGO, cpuset), #endif + INF("vinfo", S_IRUGO, pid_vx_info), + INF("ninfo", S_IRUGO, pid_nx_info), INF("oom_score", S_IRUGO, oom_score), REG("oom_adj", S_IRUGO|S_IWUSR, oom_adjust), #ifdef CONFIG_AUDITSYSCALL @@ -2143,7 +2160,7 @@ struct dentry *proc_pid_lookup(struct in goto out; rcu_read_lock(); - task = find_task_by_pid(tgid); + task = vx_find_proc_task_by_pid(tgid); if (task) get_task_struct(task); rcu_read_unlock(); @@ -2207,7 +2224,7 @@ static int proc_pid_fill_cache(struct fi int proc_pid_readdir(struct file * filp, void * dirent, filldir_t filldir) { unsigned int nr = filp->f_pos - FIRST_PROCESS_ENTRY; - struct task_struct *reaper = get_proc_task(filp->f_path.dentry->d_inode); + struct task_struct *reaper = get_proc_task_real(filp->f_path.dentry->d_inode); struct task_struct *task; int tgid; @@ -2226,6 +2243,8 @@ int proc_pid_readdir(struct file * filp, put_task_struct(task), task = next_tgid(tgid + 1)) { tgid = task->pid; filp->f_pos = tgid + TGID_OFFSET; + if (!vx_proc_task_visible(task)) + continue; if (proc_pid_fill_cache(filp, dirent, filldir, task, tgid) < 0) { put_task_struct(task); goto out; @@ -2352,9 +2371,11 @@ static struct dentry *proc_task_lookup(s tid = name_to_int(dentry); if (tid == ~0U) goto out; + if (vx_current_initpid(tid)) + goto out; rcu_read_lock(); - task = find_task_by_pid(tid); + task = vx_find_proc_task_by_pid(tid); if (task) get_task_struct(task); rcu_read_unlock(); @@ -2499,7 +2520,7 @@ static int proc_task_readdir(struct file for (task = first_tid(leader, tid, pos - 2); task; task = next_tid(task), pos++) { - tid = task->pid; + tid = vx_map_pid(task->pid); if (proc_task_fill_cache(filp, dirent, filldir, task, tid) < 0) { /* returning this tgid failed, save it as the first * pid for the next readir call */ diff -NurpP --minimal linux-source-2.6.22.orig/fs/proc/generic.c linux-source-2.6.22/fs/proc/generic.c --- linux-source-2.6.22.orig/fs/proc/generic.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/proc/generic.c 2007-10-24 16:14:06.291582822 +0200 @@ -20,6 +20,7 @@ #include #include #include +#include #include #include "internal.h" @@ -395,6 +396,8 @@ struct dentry *proc_lookup(struct inode for (de = de->subdir; de ; de = de->next) { if (de->namelen != dentry->d_name.len) continue; + if (!vx_hide_check(0, de->vx_flags)) + continue; if (!memcmp(dentry->d_name.name, de->name, de->namelen)) { unsigned int ino = de->low_ino; @@ -402,6 +405,8 @@ struct dentry *proc_lookup(struct inode spin_unlock(&proc_subdir_lock); error = -EINVAL; inode = proc_get_inode(dir->i_sb, ino, de); + /* generic proc entries belong to the host */ + inode->i_tag = 0; spin_lock(&proc_subdir_lock); break; } @@ -482,6 +487,8 @@ int proc_readdir(struct file * filp, /* filldir passes info to user space */ de_get(de); + if (!vx_hide_check(0, de->vx_flags)) + goto skip; spin_unlock(&proc_subdir_lock); if (filldir(dirent, de->name, de->namelen, filp->f_pos, de->low_ino, de->mode >> 12) < 0) { @@ -489,6 +496,7 @@ int proc_readdir(struct file * filp, goto out; } spin_lock(&proc_subdir_lock); + skip: filp->f_pos++; next = de->next; de_put(de); @@ -613,6 +621,7 @@ static struct proc_dir_entry *proc_creat ent->namelen = len; ent->mode = mode; ent->nlink = nlink; + ent->vx_flags = IATTR_PROC_DEFAULT; out: return ent; } @@ -633,7 +642,8 @@ struct proc_dir_entry *proc_symlink(cons kfree(ent->data); kfree(ent); ent = NULL; - } + } else + ent->vx_flags = IATTR_PROC_SYMLINK; } else { kfree(ent); ent = NULL; diff -NurpP --minimal linux-source-2.6.22.orig/fs/proc/inode.c linux-source-2.6.22/fs/proc/inode.c --- linux-source-2.6.22.orig/fs/proc/inode.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/proc/inode.c 2007-10-24 16:14:06.291582822 +0200 @@ -160,6 +160,8 @@ struct inode *proc_get_inode(struct supe inode->i_uid = de->uid; inode->i_gid = de->gid; } + if (de->vx_flags) + PROC_I(inode)->vx_flags = de->vx_flags; if (de->size) inode->i_size = de->size; if (de->nlink) diff -NurpP --minimal linux-source-2.6.22.orig/fs/proc/internal.h linux-source-2.6.22/fs/proc/internal.h --- linux-source-2.6.22.orig/fs/proc/internal.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/proc/internal.h 2007-10-24 16:14:06.291582822 +0200 @@ -10,6 +10,7 @@ */ #include +#include #ifdef CONFIG_PROC_SYSCTL extern int proc_sys_init(void); @@ -64,11 +65,16 @@ static inline struct pid *proc_pid(struc return PROC_I(inode)->pid; } -static inline struct task_struct *get_proc_task(struct inode *inode) +static inline struct task_struct *get_proc_task_real(struct inode *inode) { return get_pid_task(proc_pid(inode), PIDTYPE_PID); } +static inline struct task_struct *get_proc_task(struct inode *inode) +{ + return vx_get_proc_task(inode, proc_pid(inode)); +} + static inline int proc_fd(struct inode *inode) { return PROC_I(inode)->fd; diff -NurpP --minimal linux-source-2.6.22.orig/fs/proc/proc_misc.c linux-source-2.6.22/fs/proc/proc_misc.c --- linux-source-2.6.22.orig/fs/proc/proc_misc.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/fs/proc/proc_misc.c 2007-10-24 16:14:06.291582822 +0200 @@ -53,6 +53,8 @@ #include #include "internal.h" +#include + #define LOAD_INT(x) ((x) >> FSHIFT) #define LOAD_FRAC(x) LOAD_INT(((x) & (FIXED_1-1)) * 100) /* @@ -82,17 +84,32 @@ static int proc_calc_metrics(char *page, static int loadavg_read_proc(char *page, char **start, off_t off, int count, int *eof, void *data) { + unsigned int running, threads; int a, b, c; int len; - a = avenrun[0] + (FIXED_1/200); - b = avenrun[1] + (FIXED_1/200); - c = avenrun[2] + (FIXED_1/200); - len = sprintf(page,"%d.%02d %d.%02d %d.%02d %ld/%d %d\n", + if (vx_flags(VXF_VIRT_LOAD, 0)) { + struct vx_info *vxi = current->vx_info; + + a = vxi->cvirt.load[0] + (FIXED_1/200); + b = vxi->cvirt.load[1] + (FIXED_1/200); + c = vxi->cvirt.load[2] + (FIXED_1/200); + + running = atomic_read(&vxi->cvirt.nr_running); + threads = atomic_read(&vxi->cvirt.nr_threads); + } else { + a = avenrun[0] + (FIXED_1/200); + b = avenrun[1] + (FIXED_1/200); + c = avenrun[2] + (FIXED_1/200); + + running = nr_running(); + threads = nr_threads; + } + len = sprintf(page,"%d.%02d %d.%02d %d.%02d %d/%d %d\n", LOAD_INT(a), LOAD_FRAC(a), LOAD_INT(b), LOAD_FRAC(b), LOAD_INT(c), LOAD_FRAC(c), - nr_running(), nr_threads, current->nsproxy->pid_ns->last_pid); + running, threads, current->nsproxy->pid_ns->last_pid); return proc_calc_metrics(page, start, off, count, eof, len); } @@ -106,6 +123,9 @@ static int uptime_read_proc(char *page, do_posix_clock_monotonic_gettime(&uptime); cputime_to_timespec(idletime, &idle); + if (vx_flags(VXF_VIRT_UPTIME, 0)) + vx_vsi_uptime(&uptime, &idle); + len = sprintf(page,"%lu.%02lu %lu.%02lu\n", (unsigned long) uptime.tv_sec, (uptime.tv_nsec / (NSEC_PER_SEC / 100)), @@ -137,7 +157,7 @@ static int meminfo_read_proc(char *page, cached = global_page_state(NR_FILE_PAGES) - total_swapcache_pages - i.bufferram; - if (cached < 0) + if (cached < 0 || vx_flags(VXF_VIRT_MEM, 0)) cached = 0; get_vmalloc_info(&vmi); diff -NurpP --minimal linux-source-2.6.22.orig/fs/proc/root.c linux-source-2.6.22/fs/proc/root.c --- linux-source-2.6.22.orig/fs/proc/root.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/proc/root.c 2007-10-24 16:14:06.291582822 +0200 @@ -22,6 +22,9 @@ #include "internal.h" struct proc_dir_entry *proc_net, *proc_net_stat, *proc_bus, *proc_root_fs, *proc_root_driver; +struct proc_dir_entry *proc_virtual; + +extern void proc_vx_init(void); static int proc_get_sb(struct file_system_type *fs_type, int flags, const char *dev_name, void *data, struct vfsmount *mnt) @@ -79,6 +82,7 @@ void __init proc_root_init(void) proc_device_tree_init(); #endif proc_bus = proc_mkdir("bus", NULL); + proc_vx_init(); proc_sys_init(); } diff -NurpP --minimal linux-source-2.6.22.orig/fs/quota.c linux-source-2.6.22/fs/quota.c --- linux-source-2.6.22.orig/fs/quota.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/quota.c 2007-10-24 16:14:06.291582822 +0200 @@ -16,6 +16,7 @@ #include #include #include +#include /* Check validity of generic quotactl commands */ static int generic_quotactl_valid(struct super_block *sb, int type, int cmd, qid_t id) @@ -80,11 +81,11 @@ static int generic_quotactl_valid(struct if (cmd == Q_GETQUOTA) { if (((type == USRQUOTA && current->euid != id) || (type == GRPQUOTA && !in_egroup_p(id))) && - !capable(CAP_SYS_ADMIN)) + !vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL)) return -EPERM; } else if (cmd != Q_GETFMT && cmd != Q_SYNC && cmd != Q_GETINFO) - if (!capable(CAP_SYS_ADMIN)) + if (!vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL)) return -EPERM; return 0; @@ -131,10 +132,10 @@ static int xqm_quotactl_valid(struct sup if (cmd == Q_XGETQUOTA) { if (((type == XQM_USRQUOTA && current->euid != id) || (type == XQM_GRPQUOTA && !in_egroup_p(id))) && - !capable(CAP_SYS_ADMIN)) + !vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL)) return -EPERM; } else if (cmd != Q_XGETQSTAT && cmd != Q_XQUOTASYNC) { - if (!capable(CAP_SYS_ADMIN)) + if (!vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL)) return -EPERM; } @@ -327,6 +328,46 @@ static int do_quotactl(struct super_bloc return 0; } +#if defined(CONFIG_BLK_DEV_VROOT) || defined(CONFIG_BLK_DEV_VROOT_MODULE) + +#include +#include +#include +#include +#include + +static vroot_grb_func *vroot_get_real_bdev = NULL; + +static spinlock_t vroot_grb_lock = SPIN_LOCK_UNLOCKED; + +int register_vroot_grb(vroot_grb_func *func) { + int ret = -EBUSY; + + spin_lock(&vroot_grb_lock); + if (!vroot_get_real_bdev) { + vroot_get_real_bdev = func; + ret = 0; + } + spin_unlock(&vroot_grb_lock); + return ret; +} +EXPORT_SYMBOL(register_vroot_grb); + +int unregister_vroot_grb(vroot_grb_func *func) { + int ret = -EINVAL; + + spin_lock(&vroot_grb_lock); + if (vroot_get_real_bdev) { + vroot_get_real_bdev = NULL; + ret = 0; + } + spin_unlock(&vroot_grb_lock); + return ret; +} +EXPORT_SYMBOL(unregister_vroot_grb); + +#endif + /* * look up a superblock on which quota ops will be performed * - use the name of a block device to find the superblock thereon @@ -344,6 +385,22 @@ static inline struct super_block *quotac putname(tmp); if (IS_ERR(bdev)) return ERR_PTR(PTR_ERR(bdev)); +#if defined(CONFIG_BLK_DEV_VROOT) || defined(CONFIG_BLK_DEV_VROOT_MODULE) + if (bdev && bdev->bd_inode && + imajor(bdev->bd_inode) == VROOT_MAJOR) { + struct block_device *bdnew = (void *)-EINVAL; + + if (vroot_get_real_bdev) + bdnew = vroot_get_real_bdev(bdev); + else + vxdprintk(VXD_CBIT(misc, 0), + "vroot_get_real_bdev not set"); + bdput(bdev); + if (IS_ERR(bdnew)) + return ERR_PTR(PTR_ERR(bdnew)); + bdev = bdnew; + } +#endif sb = get_super(bdev); bdput(bdev); if (!sb) diff -NurpP --minimal linux-source-2.6.22.orig/fs/read_write.c linux-source-2.6.22/fs/read_write.c --- linux-source-2.6.22.orig/fs/read_write.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/read_write.c 2007-10-24 16:14:06.291582822 +0200 @@ -701,12 +701,77 @@ sys_writev(unsigned long fd, const struc return ret; } +ssize_t vfs_sendfile(struct file *out_file, struct file *in_file, loff_t *ppos, + size_t count, loff_t max) +{ + struct inode * in_inode, * out_inode; + loff_t pos; + ssize_t ret; + + /* verify in_file */ + in_inode = in_file->f_path.dentry->d_inode; + if (!in_inode) + return -EINVAL; + if (!in_file->f_op || !in_file->f_op->sendfile) + return -EINVAL; + + if (!ppos) + ppos = &in_file->f_pos; + else + if (!(in_file->f_mode & FMODE_PREAD)) + return -ESPIPE; + + ret = rw_verify_area(READ, in_file, ppos, count); + if (ret < 0) + return ret; + count = ret; + + /* verify out_file */ + out_inode = out_file->f_path.dentry->d_inode; + if (!out_inode) + return -EINVAL; + if (!out_file->f_op || !out_file->f_op->sendpage) + return -EINVAL; + + ret = rw_verify_area(WRITE, out_file, &out_file->f_pos, count); + if (ret < 0) + return ret; + count = ret; + + ret = security_file_permission (out_file, MAY_WRITE); + if (ret) + return ret; + + if (!max) + max = min(in_inode->i_sb->s_maxbytes, out_inode->i_sb->s_maxbytes); + + pos = *ppos; + if (unlikely(pos < 0)) + return -EINVAL; + if (unlikely(pos + count > max)) { + if (pos >= max) + return -EOVERFLOW; + count = max - pos; + } + + ret = in_file->f_op->sendfile(in_file, ppos, count, file_send_actor, out_file); + + if (ret > 0) { + add_rchar(current, ret); + add_wchar(current, ret); + } + + if (*ppos > max) + return -EOVERFLOW; + return ret; +} + +EXPORT_SYMBOL(vfs_sendfile); + static ssize_t do_sendfile(int out_fd, int in_fd, loff_t *ppos, size_t count, loff_t max) { struct file * in_file, * out_file; - struct inode * in_inode, * out_inode; - loff_t pos; ssize_t retval; int fput_needed_in, fput_needed_out; @@ -719,22 +784,6 @@ static ssize_t do_sendfile(int out_fd, i goto out; if (!(in_file->f_mode & FMODE_READ)) goto fput_in; - retval = -EINVAL; - in_inode = in_file->f_path.dentry->d_inode; - if (!in_inode) - goto fput_in; - if (!in_file->f_op || !in_file->f_op->sendfile) - goto fput_in; - retval = -ESPIPE; - if (!ppos) - ppos = &in_file->f_pos; - else - if (!(in_file->f_mode & FMODE_PREAD)) - goto fput_in; - retval = rw_verify_area(READ, in_file, ppos, count); - if (retval < 0) - goto fput_in; - count = retval; retval = security_file_permission (in_file, MAY_READ); if (retval) @@ -749,45 +798,11 @@ static ssize_t do_sendfile(int out_fd, i goto fput_in; if (!(out_file->f_mode & FMODE_WRITE)) goto fput_out; - retval = -EINVAL; - if (!out_file->f_op || !out_file->f_op->sendpage) - goto fput_out; - out_inode = out_file->f_path.dentry->d_inode; - retval = rw_verify_area(WRITE, out_file, &out_file->f_pos, count); - if (retval < 0) - goto fput_out; - count = retval; - - retval = security_file_permission (out_file, MAY_WRITE); - if (retval) - goto fput_out; - - if (!max) - max = min(in_inode->i_sb->s_maxbytes, out_inode->i_sb->s_maxbytes); - - pos = *ppos; - retval = -EINVAL; - if (unlikely(pos < 0)) - goto fput_out; - if (unlikely(pos + count > max)) { - retval = -EOVERFLOW; - if (pos >= max) - goto fput_out; - count = max - pos; - } - - retval = in_file->f_op->sendfile(in_file, ppos, count, file_send_actor, out_file); - if (retval > 0) { - add_rchar(current, retval); - add_wchar(current, retval); - } + retval = vfs_sendfile(out_file, in_file, ppos, count, max); inc_syscr(current); inc_syscw(current); - if (*ppos > max) - retval = -EOVERFLOW; - fput_out: fput_light(out_file, fput_needed_out); fput_in: diff -NurpP --minimal linux-source-2.6.22.orig/fs/reiserfs/bitmap.c linux-source-2.6.22/fs/reiserfs/bitmap.c --- linux-source-2.6.22.orig/fs/reiserfs/bitmap.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/reiserfs/bitmap.c 2007-10-24 16:14:06.291582822 +0200 @@ -13,6 +13,7 @@ #include #include #include +#include #define PREALLOCATION_SIZE 9 @@ -425,8 +426,10 @@ static void _reiserfs_free_block(struct set_sb_free_blocks(rs, sb_free_blocks(rs) + 1); journal_mark_dirty(th, s, sbh); - if (for_unformatted) + if (for_unformatted) { + DLIMIT_FREE_BLOCK(inode, 1); DQUOT_FREE_BLOCK_NODIRTY(inode, 1); + } } void reiserfs_free_block(struct reiserfs_transaction_handle *th, @@ -1034,6 +1037,7 @@ static inline int blocknrs_and_prealloc_ b_blocknr_t finish = SB_BLOCK_COUNT(s) - 1; int passno = 0; int nr_allocated = 0; + int blocks; determine_prealloc_size(hint); if (!hint->formatted_node) { @@ -1043,19 +1047,30 @@ static inline int blocknrs_and_prealloc_ "reiserquota: allocating %d blocks id=%u", amount_needed, hint->inode->i_uid); #endif - quota_ret = - DQUOT_ALLOC_BLOCK_NODIRTY(hint->inode, amount_needed); - if (quota_ret) /* Quota exceeded? */ + quota_ret = DQUOT_ALLOC_BLOCK_NODIRTY(hint->inode, + amount_needed); + if (quota_ret) return QUOTA_EXCEEDED; + if (DLIMIT_ALLOC_BLOCK(hint->inode, amount_needed)) { + DQUOT_FREE_BLOCK_NODIRTY(hint->inode, + amount_needed); + return NO_DISK_SPACE; + } + if (hint->preallocate && hint->prealloc_size) { #ifdef REISERQUOTA_DEBUG reiserfs_debug(s, REISERFS_DEBUG_CODE, "reiserquota: allocating (prealloc) %d blocks id=%u", hint->prealloc_size, hint->inode->i_uid); #endif - quota_ret = - DQUOT_PREALLOC_BLOCK_NODIRTY(hint->inode, - hint->prealloc_size); + quota_ret = DQUOT_PREALLOC_BLOCK_NODIRTY(hint->inode, + hint->prealloc_size); + if (!quota_ret && + DLIMIT_ALLOC_BLOCK(hint->inode, hint->prealloc_size)) { + DQUOT_FREE_BLOCK_NODIRTY(hint->inode, + hint->prealloc_size); + quota_ret = 1; + } if (quota_ret) hint->preallocate = hint->prealloc_size = 0; } @@ -1087,7 +1102,10 @@ static inline int blocknrs_and_prealloc_ nr_allocated, hint->inode->i_uid); #endif - DQUOT_FREE_BLOCK_NODIRTY(hint->inode, amount_needed + hint->prealloc_size - nr_allocated); /* Free not allocated blocks */ + /* Free not allocated blocks */ + blocks = amount_needed + hint->prealloc_size - nr_allocated; + DLIMIT_FREE_BLOCK(hint->inode, blocks); + DQUOT_FREE_BLOCK_NODIRTY(hint->inode, blocks); } while (nr_allocated--) reiserfs_free_block(hint->th, hint->inode, @@ -1118,10 +1136,10 @@ static inline int blocknrs_and_prealloc_ REISERFS_I(hint->inode)->i_prealloc_count, hint->inode->i_uid); #endif - DQUOT_FREE_BLOCK_NODIRTY(hint->inode, amount_needed + - hint->prealloc_size - nr_allocated - - REISERFS_I(hint->inode)-> - i_prealloc_count); + blocks = amount_needed + hint->prealloc_size - nr_allocated - + REISERFS_I(hint->inode)->i_prealloc_count; + DLIMIT_FREE_BLOCK(hint->inode, blocks); + DQUOT_FREE_BLOCK_NODIRTY(hint->inode, blocks); } return CARRY_ON; diff -NurpP --minimal linux-source-2.6.22.orig/fs/reiserfs/file.c linux-source-2.6.22/fs/reiserfs/file.c --- linux-source-2.6.22.orig/fs/reiserfs/file.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/fs/reiserfs/file.c 2007-10-24 16:14:06.291582822 +0200 @@ -1532,6 +1532,7 @@ const struct file_operations reiserfs_fi .release = reiserfs_file_release, .fsync = reiserfs_sync_file, .sendfile = generic_file_sendfile, + .sendpage = generic_file_sendpage, .aio_read = generic_file_aio_read, .aio_write = generic_file_aio_write, .splice_read = generic_file_splice_read, @@ -1546,4 +1547,5 @@ const struct inode_operations reiserfs_f .listxattr = reiserfs_listxattr, .removexattr = reiserfs_removexattr, .permission = reiserfs_permission, + .sync_flags = reiserfs_sync_flags, }; diff -NurpP --minimal linux-source-2.6.22.orig/fs/reiserfs/inode.c linux-source-2.6.22/fs/reiserfs/inode.c --- linux-source-2.6.22.orig/fs/reiserfs/inode.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/reiserfs/inode.c 2007-10-24 16:14:06.291582822 +0200 @@ -16,6 +16,8 @@ #include #include #include +#include +#include static int reiserfs_commit_write(struct file *f, struct page *page, unsigned from, unsigned to); @@ -50,6 +52,7 @@ void reiserfs_delete_inode(struct inode * stat data deletion */ if (!err) DQUOT_FREE_INODE(inode); + DLIMIT_FREE_INODE(inode); if (journal_end(&th, inode->i_sb, jbegin_count)) goto out; @@ -1112,6 +1115,8 @@ static void init_inode(struct inode *ino struct buffer_head *bh; struct item_head *ih; __u32 rdev; + uid_t uid; + gid_t gid; //int version = ITEM_VERSION_1; bh = PATH_PLAST_BUFFER(path); @@ -1135,12 +1140,13 @@ static void init_inode(struct inode *ino (struct stat_data_v1 *)B_I_PITEM(bh, ih); unsigned long blocks; + uid = sd_v1_uid(sd); + gid = sd_v1_gid(sd); + set_inode_item_key_version(inode, KEY_FORMAT_3_5); set_inode_sd_version(inode, STAT_DATA_V1); inode->i_mode = sd_v1_mode(sd); inode->i_nlink = sd_v1_nlink(sd); - inode->i_uid = sd_v1_uid(sd); - inode->i_gid = sd_v1_gid(sd); inode->i_size = sd_v1_size(sd); inode->i_atime.tv_sec = sd_v1_atime(sd); inode->i_mtime.tv_sec = sd_v1_mtime(sd); @@ -1182,11 +1188,12 @@ static void init_inode(struct inode *ino // (directories and symlinks) struct stat_data *sd = (struct stat_data *)B_I_PITEM(bh, ih); + uid = sd_v2_uid(sd); + gid = sd_v2_gid(sd); + inode->i_mode = sd_v2_mode(sd); inode->i_nlink = sd_v2_nlink(sd); - inode->i_uid = sd_v2_uid(sd); inode->i_size = sd_v2_size(sd); - inode->i_gid = sd_v2_gid(sd); inode->i_mtime.tv_sec = sd_v2_mtime(sd); inode->i_atime.tv_sec = sd_v2_atime(sd); inode->i_ctime.tv_sec = sd_v2_ctime(sd); @@ -1216,6 +1223,10 @@ static void init_inode(struct inode *ino sd_attrs_to_i_attrs(sd_v2_attrs(sd), inode); } + inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid); + inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid); + inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid, 0); + pathrelse(path); if (S_ISREG(inode->i_mode)) { inode->i_op = &reiserfs_file_inode_operations; @@ -1238,13 +1249,15 @@ static void init_inode(struct inode *ino static void inode2sd(void *sd, struct inode *inode, loff_t size) { struct stat_data *sd_v2 = (struct stat_data *)sd; + uid_t uid = TAGINO_UID(DX_TAG(inode), inode->i_uid, inode->i_tag); + gid_t gid = TAGINO_GID(DX_TAG(inode), inode->i_gid, inode->i_tag); __u16 flags; + set_sd_v2_uid(sd_v2, uid); + set_sd_v2_gid(sd_v2, gid); set_sd_v2_mode(sd_v2, inode->i_mode); set_sd_v2_nlink(sd_v2, inode->i_nlink); - set_sd_v2_uid(sd_v2, inode->i_uid); set_sd_v2_size(sd_v2, size); - set_sd_v2_gid(sd_v2, inode->i_gid); set_sd_v2_mtime(sd_v2, inode->i_mtime.tv_sec); set_sd_v2_atime(sd_v2, inode->i_atime.tv_sec); set_sd_v2_ctime(sd_v2, inode->i_ctime.tv_sec); @@ -1775,6 +1788,10 @@ int reiserfs_new_inode(struct reiserfs_t BUG_ON(!th->t_trans_id); + if (DLIMIT_ALLOC_INODE(inode)) { + err = -ENOSPC; + goto out_bad_dlimit; + } if (DQUOT_ALLOC_INODE(inode)) { err = -EDQUOT; goto out_end_trans; @@ -1960,6 +1977,9 @@ int reiserfs_new_inode(struct reiserfs_t DQUOT_FREE_INODE(inode); out_end_trans: + DLIMIT_FREE_INODE(inode); + + out_bad_dlimit: journal_end(th, th->t_super, th->t_blocks_allocated); /* Drop can be outside and it needs more credits so it's better to have it outside */ DQUOT_DROP(inode); @@ -2690,6 +2710,14 @@ void sd_attrs_to_i_attrs(__u16 sd_attrs, inode->i_flags |= S_IMMUTABLE; else inode->i_flags &= ~S_IMMUTABLE; + if (sd_attrs & REISERFS_IUNLINK_FL) + inode->i_flags |= S_IUNLINK; + else + inode->i_flags &= ~S_IUNLINK; + if (sd_attrs & REISERFS_BARRIER_FL) + inode->i_flags |= S_BARRIER; + else + inode->i_flags &= ~S_BARRIER; if (sd_attrs & REISERFS_APPEND_FL) inode->i_flags |= S_APPEND; else @@ -2712,6 +2740,14 @@ void i_attrs_to_sd_attrs(struct inode *i *sd_attrs |= REISERFS_IMMUTABLE_FL; else *sd_attrs &= ~REISERFS_IMMUTABLE_FL; + if (inode->i_flags & S_IUNLINK) + *sd_attrs |= REISERFS_IUNLINK_FL; + else + *sd_attrs &= ~REISERFS_IUNLINK_FL; + if (inode->i_flags & S_BARRIER) + *sd_attrs |= REISERFS_BARRIER_FL; + else + *sd_attrs &= ~REISERFS_BARRIER_FL; if (inode->i_flags & S_SYNC) *sd_attrs |= REISERFS_SYNC_FL; else @@ -2891,6 +2927,22 @@ static ssize_t reiserfs_direct_IO(int rw reiserfs_get_blocks_direct_io, NULL); } +int reiserfs_sync_flags(struct inode *inode) +{ + u16 oldflags, newflags; + + oldflags = REISERFS_I(inode)->i_attrs; + newflags = oldflags; + i_attrs_to_sd_attrs(inode, &newflags); + + if (oldflags ^ newflags) { + REISERFS_I(inode)->i_attrs = newflags; + inode->i_ctime = CURRENT_TIME_SEC; + mark_inode_dirty(inode); + } + return 0; +} + int reiserfs_setattr(struct dentry *dentry, struct iattr *attr) { struct inode *inode = dentry->d_inode; @@ -2940,9 +2992,11 @@ int reiserfs_setattr(struct dentry *dent } error = inode_change_ok(inode, attr); + if (!error) { if ((ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) || - (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) { + (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid) || + (ia_valid & ATTR_TAG && attr->ia_tag != inode->i_tag)) { error = reiserfs_chown_xattrs(inode, attr); if (!error) { @@ -2972,6 +3026,9 @@ int reiserfs_setattr(struct dentry *dent inode->i_uid = attr->ia_uid; if (attr->ia_valid & ATTR_GID) inode->i_gid = attr->ia_gid; + if ((attr->ia_valid & ATTR_TAG) && + IS_TAGGED(inode)) + inode->i_tag = attr->ia_tag; mark_inode_dirty(inode); error = journal_end(&th, inode->i_sb, jbegin_count); diff -NurpP --minimal linux-source-2.6.22.orig/fs/reiserfs/ioctl.c linux-source-2.6.22/fs/reiserfs/ioctl.c --- linux-source-2.6.22.orig/fs/reiserfs/ioctl.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/reiserfs/ioctl.c 2007-10-24 16:14:06.291582822 +0200 @@ -4,6 +4,7 @@ #include #include +#include #include #include #include @@ -24,7 +25,7 @@ static int reiserfs_unpack(struct inode int reiserfs_ioctl(struct inode *inode, struct file *filp, unsigned int cmd, unsigned long arg) { - unsigned int flags; + unsigned int flags, oldflags; switch (cmd) { case REISERFS_IOC_UNPACK: @@ -43,12 +44,14 @@ int reiserfs_ioctl(struct inode *inode, flags = REISERFS_I(inode)->i_attrs; i_attrs_to_sd_attrs(inode, (__u16 *) & flags); + flags &= REISERFS_FL_USER_VISIBLE; return put_user(flags, (int __user *)arg); case REISERFS_IOC_SETFLAGS:{ if (!reiserfs_attrs(inode->i_sb)) return -ENOTTY; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if ((current->fsuid != inode->i_uid) @@ -58,10 +61,12 @@ int reiserfs_ioctl(struct inode *inode, if (get_user(flags, (int __user *)arg)) return -EFAULT; - if (((flags ^ REISERFS_I(inode)-> - i_attrs) & (REISERFS_IMMUTABLE_FL | - REISERFS_APPEND_FL)) - && !capable(CAP_LINUX_IMMUTABLE)) + oldflags = REISERFS_I(inode) -> i_attrs; + if (((oldflags & REISERFS_IMMUTABLE_FL) || + ((flags ^ oldflags) & + (REISERFS_IMMUTABLE_FL | REISERFS_IUNLINK_FL | + REISERFS_APPEND_FL))) && + !capable(CAP_LINUX_IMMUTABLE)) return -EPERM; if ((flags & REISERFS_NOTAIL_FL) && @@ -72,6 +77,9 @@ int reiserfs_ioctl(struct inode *inode, if (result) return result; } + + flags = flags & REISERFS_FL_USER_MODIFIABLE; + flags |= oldflags & ~REISERFS_FL_USER_MODIFIABLE; sd_attrs_to_i_attrs(flags, inode); REISERFS_I(inode)->i_attrs = flags; inode->i_ctime = CURRENT_TIME_SEC; @@ -83,7 +91,8 @@ int reiserfs_ioctl(struct inode *inode, case REISERFS_IOC_SETVERSION: if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER)) return -EPERM; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if (get_user(inode->i_generation, (int __user *)arg)) return -EFAULT; diff -NurpP --minimal linux-source-2.6.22.orig/fs/reiserfs/namei.c linux-source-2.6.22/fs/reiserfs/namei.c --- linux-source-2.6.22.orig/fs/reiserfs/namei.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/reiserfs/namei.c 2007-10-24 16:14:06.291582822 +0200 @@ -17,6 +17,7 @@ #include #include #include +#include #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; } #define DEC_DIR_INODE_NLINK(i) if (i->i_nlink != 1) drop_nlink(i); @@ -360,6 +361,7 @@ static struct dentry *reiserfs_lookup(st reiserfs_write_unlock(dir->i_sb); return ERR_PTR(-EACCES); } + dx_propagate_tag(nd, inode); /* Propogate the priv_object flag so we know we're in the priv tree */ if (is_reiserfs_priv_object(dir)) @@ -595,6 +597,7 @@ static int new_inode_init(struct inode * } else { inode->i_gid = current->fsgid; } + inode->i_tag = dx_current_fstag(inode->i_sb); DQUOT_INIT(inode); return 0; } @@ -1541,6 +1544,7 @@ const struct inode_operations reiserfs_d .listxattr = reiserfs_listxattr, .removexattr = reiserfs_removexattr, .permission = reiserfs_permission, + .sync_flags = reiserfs_sync_flags, }; /* @@ -1557,6 +1561,7 @@ const struct inode_operations reiserfs_s .listxattr = reiserfs_listxattr, .removexattr = reiserfs_removexattr, .permission = reiserfs_permission, + .sync_flags = reiserfs_sync_flags, }; @@ -1570,5 +1575,6 @@ const struct inode_operations reiserfs_s .listxattr = reiserfs_listxattr, .removexattr = reiserfs_removexattr, .permission = reiserfs_permission, + .sync_flags = reiserfs_sync_flags, }; diff -NurpP --minimal linux-source-2.6.22.orig/fs/reiserfs/stree.c linux-source-2.6.22/fs/reiserfs/stree.c --- linux-source-2.6.22.orig/fs/reiserfs/stree.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/reiserfs/stree.c 2007-10-24 16:14:06.291582822 +0200 @@ -55,6 +55,7 @@ #include #include #include +#include /* Does the buffer contain a disk block which is in the tree. */ inline int B_IS_IN_TREE(const struct buffer_head *p_s_bh) @@ -1296,6 +1297,7 @@ int reiserfs_delete_item(struct reiserfs "reiserquota delete_item(): freeing %u, id=%u type=%c", quota_cut_bytes, p_s_inode->i_uid, head2type(&s_ih)); #endif + DLIMIT_FREE_SPACE(p_s_inode, quota_cut_bytes); DQUOT_FREE_SPACE_NODIRTY(p_s_inode, quota_cut_bytes); /* Return deleted body length */ @@ -1384,6 +1386,7 @@ void reiserfs_delete_solid_item(struct r #endif DQUOT_FREE_SPACE_NODIRTY(inode, quota_cut_bytes); + DLIMIT_FREE_SPACE(inode, quota_cut_bytes); } break; } @@ -1737,6 +1740,7 @@ int reiserfs_cut_from_item(struct reiser "reiserquota cut_from_item(): freeing %u id=%u type=%c", quota_cut_bytes, p_s_inode->i_uid, '?'); #endif + DLIMIT_FREE_SPACE(p_s_inode, quota_cut_bytes); DQUOT_FREE_SPACE_NODIRTY(p_s_inode, quota_cut_bytes); return n_ret_value; } @@ -1978,6 +1982,11 @@ int reiserfs_paste_into_item(struct reis pathrelse(p_s_search_path); return -EDQUOT; } + if (DLIMIT_ALLOC_SPACE(inode, n_pasted_size)) { + DQUOT_FREE_SPACE_NODIRTY(inode, n_pasted_size); + pathrelse(p_s_search_path); + return -ENOSPC; + } init_tb_struct(th, &s_paste_balance, th->t_super, p_s_search_path, n_pasted_size); #ifdef DISPLACE_NEW_PACKING_LOCALITIES @@ -2030,6 +2039,7 @@ int reiserfs_paste_into_item(struct reis n_pasted_size, inode->i_uid, key2type(&(p_s_key->on_disk_key))); #endif + DLIMIT_FREE_SPACE(inode, n_pasted_size); DQUOT_FREE_SPACE_NODIRTY(inode, n_pasted_size); return retval; } @@ -2067,6 +2077,11 @@ int reiserfs_insert_item(struct reiserfs pathrelse(p_s_path); return -EDQUOT; } + if (DLIMIT_ALLOC_SPACE(inode, quota_bytes)) { + DQUOT_FREE_SPACE_NODIRTY(inode, quota_bytes); + pathrelse(p_s_path); + return -ENOSPC; + } } init_tb_struct(th, &s_ins_balance, th->t_super, p_s_path, IH_SIZE + ih_item_len(p_s_ih)); @@ -2114,7 +2129,9 @@ int reiserfs_insert_item(struct reiserfs "reiserquota insert_item(): freeing %u id=%u type=%c", quota_bytes, inode->i_uid, head2type(p_s_ih)); #endif - if (inode) + if (inode) { + DLIMIT_FREE_SPACE(inode, quota_bytes); DQUOT_FREE_SPACE_NODIRTY(inode, quota_bytes); + } return retval; } diff -NurpP --minimal linux-source-2.6.22.orig/fs/reiserfs/super.c linux-source-2.6.22/fs/reiserfs/super.c --- linux-source-2.6.22.orig/fs/reiserfs/super.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/reiserfs/super.c 2007-10-24 16:14:06.291582822 +0200 @@ -882,6 +882,14 @@ static int reiserfs_parse_options(struct {"user_xattr",.setmask = 1 << REISERFS_UNSUPPORTED_OPT}, {"nouser_xattr",.clrmask = 1 << REISERFS_UNSUPPORTED_OPT}, #endif +#ifndef CONFIG_TAGGING_NONE + {"tagxid",.setmask = 1 << REISERFS_TAGGED}, + {"tag",.setmask = 1 << REISERFS_TAGGED}, + {"notag",.clrmask = 1 << REISERFS_TAGGED}, +#endif +#ifdef CONFIG_PROPAGATE + {"tag",.arg_required = 'T',.values = NULL}, +#endif #ifdef CONFIG_REISERFS_FS_POSIX_ACL {"acl",.setmask = 1 << REISERFS_POSIXACL}, {"noacl",.clrmask = 1 << REISERFS_POSIXACL}, @@ -1143,6 +1151,12 @@ static int reiserfs_remount(struct super return -EINVAL; } + if ((mount_options & (1 << REISERFS_TAGGED)) && + !(s->s_flags & MS_TAGGED)) { + reiserfs_warning(s, "reiserfs: tagging not permitted on remount."); + return -EINVAL; + } + handle_attrs(s); /* Add options that are safe here */ @@ -1591,6 +1605,10 @@ static int reiserfs_fill_super(struct su goto error; } + /* map mount option tagxid */ + if (REISERFS_SB(s)->s_mount_opt & (1 << REISERFS_TAGGED)) + s->s_flags |= MS_TAGGED; + rs = SB_DISK_SUPER_BLOCK(s); /* Let's do basic sanity check to verify that underlying device is not smaller than the filesystem. If the check fails then abort and scream, diff -NurpP --minimal linux-source-2.6.22.orig/fs/reiserfs/xattr.c linux-source-2.6.22/fs/reiserfs/xattr.c --- linux-source-2.6.22.orig/fs/reiserfs/xattr.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/fs/reiserfs/xattr.c 2007-10-24 16:14:06.291582822 +0200 @@ -35,6 +35,7 @@ #include #include #include +#include #include #include #include @@ -775,7 +776,7 @@ int reiserfs_delete_xattrs(struct inode if (dir->d_inode->i_nlink <= 2) { root = get_xa_root(inode->i_sb, XATTR_REPLACE); reiserfs_write_lock_xattrs(inode->i_sb); - err = vfs_rmdir(root->d_inode, dir, NULL); + err = vfs_rmdir(root->d_inode, dir, NULL, NULL); reiserfs_write_unlock_xattrs(inode->i_sb); dput(root); } else { diff -NurpP --minimal linux-source-2.6.22.orig/fs/stat.c linux-source-2.6.22/fs/stat.c --- linux-source-2.6.22.orig/fs/stat.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/fs/stat.c 2007-10-24 16:14:06.291582822 +0200 @@ -26,6 +26,7 @@ void generic_fillattr(struct inode *inod stat->nlink = inode->i_nlink; stat->uid = inode->i_uid; stat->gid = inode->i_gid; + stat->tag = inode->i_tag; stat->rdev = inode->i_rdev; stat->atime = inode->i_atime; stat->mtime = inode->i_mtime; diff -NurpP --minimal linux-source-2.6.22.orig/fs/super.c linux-source-2.6.22/fs/super.c --- linux-source-2.6.22.orig/fs/super.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/super.c 2007-10-24 16:14:06.291582822 +0200 @@ -37,6 +37,9 @@ #include #include #include +#include +#include +#include #include @@ -860,12 +863,18 @@ struct vfsmount * vfs_kern_mount(struct file_system_type *type, int flags, const char *name, void *data) { struct vfsmount *mnt; + struct super_block *sb; char *secdata = NULL; int error; if (!type) return ERR_PTR(-ENODEV); + error = -EPERM; + if ((type->fs_flags & FS_BINARY_MOUNTDATA) && + !vx_capable(CAP_SYS_ADMIN, VXC_BINARY_MOUNT)) + goto out; + error = -ENOMEM; mnt = alloc_vfsmnt(name); if (!mnt) @@ -885,7 +894,14 @@ vfs_kern_mount(struct file_system_type * if (error < 0) goto out_free_secdata; - error = security_sb_kern_mount(mnt->mnt_sb, secdata); + sb = mnt->mnt_sb; + error = -EPERM; + if (!vx_capable(CAP_SYS_ADMIN, VXC_BINARY_MOUNT) && !sb->s_bdev && + (sb->s_magic != PROC_SUPER_MAGIC) && + (sb->s_magic != DEVPTS_SUPER_MAGIC)) + goto out_sb; + + error = security_sb_kern_mount(sb, secdata); if (error) goto out_sb; diff -NurpP --minimal linux-source-2.6.22.orig/fs/sysfs/mount.c linux-source-2.6.22/fs/sysfs/mount.c --- linux-source-2.6.22.orig/fs/sysfs/mount.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/sysfs/mount.c 2007-10-24 16:14:06.291582822 +0200 @@ -12,8 +12,6 @@ #include "sysfs.h" -/* Random magic number */ -#define SYSFS_MAGIC 0x62656572 struct vfsmount *sysfs_mount; struct super_block * sysfs_sb = NULL; @@ -48,7 +46,7 @@ static int sysfs_fill_super(struct super sb->s_blocksize = PAGE_CACHE_SIZE; sb->s_blocksize_bits = PAGE_CACHE_SHIFT; - sb->s_magic = SYSFS_MAGIC; + sb->s_magic = SYSFS_SUPER_MAGIC; sb->s_op = &sysfs_ops; sb->s_time_gran = 1; sysfs_sb = sb; diff -NurpP --minimal linux-source-2.6.22.orig/fs/utimes.c linux-source-2.6.22/fs/utimes.c --- linux-source-2.6.22.orig/fs/utimes.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/fs/utimes.c 2007-10-24 16:14:06.291582822 +0200 @@ -6,6 +6,8 @@ #include #include #include +#include +#include #include #include @@ -70,6 +72,9 @@ long do_utimes(int dfd, char __user *fil if (error) goto out; + error = cow_check_and_break(&nd); + if (error) + goto dput_and_out; path.dentry = nd.dentry; path.mnt = nd.mnt; } diff -NurpP --minimal linux-source-2.6.22.orig/fs/xattr.c linux-source-2.6.22/fs/xattr.c --- linux-source-2.6.22.orig/fs/xattr.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/fs/xattr.c 2007-10-24 16:14:06.291582822 +0200 @@ -17,6 +17,7 @@ #include #include #include +#include #include @@ -225,6 +226,9 @@ setxattr(struct dentry *dentry, struct v } } + if (MNT_IS_RDONLY(mnt)) + return -EROFS; + error = vfs_setxattr(dentry, mnt, kname, kvalue, size, flags, file); kfree(kvalue); return error; @@ -451,6 +455,9 @@ removexattr(struct dentry *dentry, struc if (error < 0) return error; + if (MNT_IS_RDONLY(mnt)) + return -EROFS; + return vfs_removexattr(dentry, mnt, kname, file); } diff -NurpP --minimal linux-source-2.6.22.orig/fs/xfs/linux-2.6/xfs_file.c linux-source-2.6.22/fs/xfs/linux-2.6/xfs_file.c --- linux-source-2.6.22.orig/fs/xfs/linux-2.6/xfs_file.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/xfs/linux-2.6/xfs_file.c 2007-10-24 16:14:06.291582822 +0200 @@ -453,6 +453,7 @@ const struct file_operations xfs_file_op .aio_read = xfs_file_aio_read, .aio_write = xfs_file_aio_write, .sendfile = xfs_file_sendfile, + .sendpage = generic_file_sendpage, .splice_read = xfs_file_splice_read, .splice_write = xfs_file_splice_write, .unlocked_ioctl = xfs_file_ioctl, @@ -476,6 +477,7 @@ const struct file_operations xfs_invis_f .aio_read = xfs_file_aio_read_invis, .aio_write = xfs_file_aio_write_invis, .sendfile = xfs_file_sendfile_invis, + .sendpage = generic_file_sendpage, .splice_read = xfs_file_splice_read_invis, .splice_write = xfs_file_splice_write_invis, .unlocked_ioctl = xfs_file_ioctl_invis, diff -NurpP --minimal linux-source-2.6.22.orig/fs/xfs/linux-2.6/xfs_ioctl.c linux-source-2.6.22/fs/xfs/linux-2.6/xfs_ioctl.c --- linux-source-2.6.22.orig/fs/xfs/linux-2.6/xfs_ioctl.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/xfs/linux-2.6/xfs_ioctl.c 2007-10-24 16:14:06.291582822 +0200 @@ -1128,6 +1128,10 @@ xfs_di2lxflags( if (di_flags & XFS_DIFLAG_IMMUTABLE) flags |= FS_IMMUTABLE_FL; + if (di_flags & XFS_DIFLAG_IUNLINK) + flags |= FS_IUNLINK_FL; + if (di_flags & XFS_DIFLAG_BARRIER) + flags |= FS_BARRIER_FL; if (di_flags & XFS_DIFLAG_APPEND) flags |= FS_APPEND_FL; if (di_flags & XFS_DIFLAG_SYNC) diff -NurpP --minimal linux-source-2.6.22.orig/fs/xfs/linux-2.6/xfs_iops.c linux-source-2.6.22/fs/xfs/linux-2.6/xfs_iops.c --- linux-source-2.6.22.orig/fs/xfs/linux-2.6/xfs_iops.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/xfs/linux-2.6/xfs_iops.c 2007-10-24 16:14:06.291582822 +0200 @@ -51,6 +51,7 @@ #include #include #include +#include /* * Get a XFS inode from a given vnode. @@ -400,6 +401,7 @@ xfs_vn_lookup( d_add(dentry, NULL); return NULL; } + dx_propagate_tag(nd, vn_to_inode(cvp)); return d_splice_alias(vn_to_inode(cvp), dentry); } @@ -657,6 +659,10 @@ xfs_vn_setattr( int flags = 0; int error; + error = inode_change_ok(inode, attr); + if (error) + return error; + if (ia_valid & ATTR_UID) { vattr.va_mask |= XFS_AT_UID; vattr.va_uid = attr->ia_uid; @@ -665,6 +671,10 @@ xfs_vn_setattr( vattr.va_mask |= XFS_AT_GID; vattr.va_gid = attr->ia_gid; } + if ((ia_valid & ATTR_TAG) && IS_TAGGED(inode)) { + vattr.va_mask |= XFS_AT_TAG; + vattr.va_tag = attr->ia_tag; + } if (ia_valid & ATTR_SIZE) { vattr.va_mask |= XFS_AT_SIZE; vattr.va_size = attr->ia_size; @@ -710,6 +720,42 @@ xfs_vn_truncate( } STATIC int +xfs_vn_sync_flags(struct inode *inode) +{ + unsigned int oldflags, newflags; + int flags = 0; + int error; + bhv_vattr_t vattr; + bhv_vnode_t *vp = vn_from_inode(inode); + + memset(&vattr, 0, sizeof vattr); + + vattr.va_mask = XFS_AT_XFLAGS; + error = bhv_vop_getattr(vp, &vattr, 0, NULL); + + if (error) + return error; + oldflags = vattr.va_xflags; + newflags = oldflags & ~(XFS_XFLAG_IMMUTABLE | + XFS_XFLAG_IUNLINK | XFS_XFLAG_BARRIER); + + if (IS_IMMUTABLE(inode)) + newflags |= XFS_XFLAG_IMMUTABLE; + if (IS_IUNLINK(inode)) + newflags |= XFS_XFLAG_IUNLINK; + if (IS_BARRIER(inode)) + newflags |= XFS_XFLAG_BARRIER; + + if (oldflags ^ newflags) { + vattr.va_xflags = newflags; + vattr.va_mask |= XFS_AT_XFLAGS; + error = bhv_vop_setattr(vp, &vattr, flags, NULL); + } + vn_revalidate(vp); + return error; +} + +STATIC int xfs_vn_setxattr( struct dentry *dentry, const char *name, @@ -822,6 +868,7 @@ const struct inode_operations xfs_inode_ .getxattr = xfs_vn_getxattr, .listxattr = xfs_vn_listxattr, .removexattr = xfs_vn_removexattr, + .sync_flags = xfs_vn_sync_flags, }; const struct inode_operations xfs_dir_inode_operations = { @@ -841,6 +888,7 @@ const struct inode_operations xfs_dir_in .getxattr = xfs_vn_getxattr, .listxattr = xfs_vn_listxattr, .removexattr = xfs_vn_removexattr, + .sync_flags = xfs_vn_sync_flags, }; const struct inode_operations xfs_symlink_inode_operations = { @@ -854,4 +902,5 @@ const struct inode_operations xfs_symlin .getxattr = xfs_vn_getxattr, .listxattr = xfs_vn_listxattr, .removexattr = xfs_vn_removexattr, + .sync_flags = xfs_vn_sync_flags, }; diff -NurpP --minimal linux-source-2.6.22.orig/fs/xfs/linux-2.6/xfs_linux.h linux-source-2.6.22/fs/xfs/linux-2.6/xfs_linux.h --- linux-source-2.6.22.orig/fs/xfs/linux-2.6/xfs_linux.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/xfs/linux-2.6/xfs_linux.h 2007-10-24 16:14:06.291582822 +0200 @@ -129,6 +129,7 @@ #define current_pid() (current->pid) #define current_fsuid(cred) (current->fsuid) #define current_fsgid(cred) (current->fsgid) +#define current_fstag(cred,vp) (dx_current_fstag(vn_to_inode(vp)->i_sb)) #define current_test_flags(f) (current->flags & (f)) #define current_set_flags_nested(sp, f) \ (*(sp) = current->flags, current->flags |= (f)) diff -NurpP --minimal linux-source-2.6.22.orig/fs/xfs/linux-2.6/xfs_super.c linux-source-2.6.22/fs/xfs/linux-2.6/xfs_super.c --- linux-source-2.6.22.orig/fs/xfs/linux-2.6/xfs_super.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/xfs/linux-2.6/xfs_super.c 2007-10-24 16:14:06.291582822 +0200 @@ -157,6 +157,7 @@ xfs_revalidate_inode( inode->i_nlink = ip->i_d.di_nlink; inode->i_uid = ip->i_d.di_uid; inode->i_gid = ip->i_d.di_gid; + inode->i_tag = ip->i_d.di_tag; switch (inode->i_mode & S_IFMT) { case S_IFBLK: @@ -184,6 +185,14 @@ xfs_revalidate_inode( inode->i_flags |= S_IMMUTABLE; else inode->i_flags &= ~S_IMMUTABLE; + if (ip->i_d.di_flags & XFS_DIFLAG_IUNLINK) + inode->i_flags |= S_IUNLINK; + else + inode->i_flags &= ~S_IUNLINK; + if (ip->i_d.di_flags & XFS_DIFLAG_BARRIER) + inode->i_flags |= S_BARRIER; + else + inode->i_flags &= ~S_BARRIER; if (ip->i_d.di_flags & XFS_DIFLAG_APPEND) inode->i_flags |= S_APPEND; else @@ -712,6 +721,12 @@ xfs_fs_remount( int error; error = bhv_vfs_parseargs(vfsp, options, args, 1); + if ((args->flags2 & XFSMNT2_TAGGED) && + !(sb->s_flags & MS_TAGGED)) { + printk("XFS: %s: tagging not permitted on remount.\n", + sb->s_id); + error = EINVAL; + } if (!error) error = bhv_vfs_mntupdate(vfsp, flags, args); kmem_free(args, sizeof(*args)); diff -NurpP --minimal linux-source-2.6.22.orig/fs/xfs/linux-2.6/xfs_vnode.c linux-source-2.6.22/fs/xfs/linux-2.6/xfs_vnode.c --- linux-source-2.6.22.orig/fs/xfs/linux-2.6/xfs_vnode.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/xfs/linux-2.6/xfs_vnode.c 2007-10-24 16:14:06.291582822 +0200 @@ -119,6 +119,7 @@ vn_revalidate_core( inode->i_nlink = vap->va_nlink; inode->i_uid = vap->va_uid; inode->i_gid = vap->va_gid; + inode->i_tag = vap->va_tag; inode->i_blocks = vap->va_nblocks; inode->i_mtime = vap->va_mtime; inode->i_ctime = vap->va_ctime; @@ -126,6 +127,14 @@ vn_revalidate_core( inode->i_flags |= S_IMMUTABLE; else inode->i_flags &= ~S_IMMUTABLE; + if (vap->va_xflags & XFS_XFLAG_IUNLINK) + inode->i_flags |= S_IUNLINK; + else + inode->i_flags &= ~S_IUNLINK; + if (vap->va_xflags & XFS_XFLAG_BARRIER) + inode->i_flags |= S_BARRIER; + else + inode->i_flags &= ~S_BARRIER; if (vap->va_xflags & XFS_XFLAG_APPEND) inode->i_flags |= S_APPEND; else diff -NurpP --minimal linux-source-2.6.22.orig/fs/xfs/linux-2.6/xfs_vnode.h linux-source-2.6.22/fs/xfs/linux-2.6/xfs_vnode.h --- linux-source-2.6.22.orig/fs/xfs/linux-2.6/xfs_vnode.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/xfs/linux-2.6/xfs_vnode.h 2007-10-24 16:14:06.291582822 +0200 @@ -350,6 +350,7 @@ typedef struct bhv_vattr { xfs_nlink_t va_nlink; /* number of references to file */ uid_t va_uid; /* owner user id */ gid_t va_gid; /* owner group id */ + tag_t va_tag; /* owner group id */ xfs_ino_t va_nodeid; /* file id */ xfs_off_t va_size; /* file size in bytes */ u_long va_blocksize; /* blocksize preferred for i/o */ @@ -398,13 +399,15 @@ typedef struct bhv_vattr { #define XFS_AT_PROJID 0x04000000 #define XFS_AT_SIZE_NOPERM 0x08000000 #define XFS_AT_GENCOUNT 0x10000000 +#define XFS_AT_TAG 0x20000000 #define XFS_AT_ALL (XFS_AT_TYPE|XFS_AT_MODE|XFS_AT_UID|XFS_AT_GID|\ XFS_AT_FSID|XFS_AT_NODEID|XFS_AT_NLINK|XFS_AT_SIZE|\ XFS_AT_ATIME|XFS_AT_MTIME|XFS_AT_CTIME|XFS_AT_RDEV|\ XFS_AT_BLKSIZE|XFS_AT_NBLOCKS|XFS_AT_VCODE|XFS_AT_MAC|\ XFS_AT_ACL|XFS_AT_CAP|XFS_AT_INF|XFS_AT_XFLAGS|XFS_AT_EXTSIZE|\ - XFS_AT_NEXTENTS|XFS_AT_ANEXTENTS|XFS_AT_PROJID|XFS_AT_GENCOUNT) + XFS_AT_NEXTENTS|XFS_AT_ANEXTENTS|XFS_AT_PROJID|XFS_AT_GENCOUNT\ + XFS_AT_TAG) #define XFS_AT_STAT (XFS_AT_TYPE|XFS_AT_MODE|XFS_AT_UID|XFS_AT_GID|\ XFS_AT_FSID|XFS_AT_NODEID|XFS_AT_NLINK|XFS_AT_SIZE|\ diff -NurpP --minimal linux-source-2.6.22.orig/fs/xfs/quota/xfs_qm_syscalls.c linux-source-2.6.22/fs/xfs/quota/xfs_qm_syscalls.c --- linux-source-2.6.22.orig/fs/xfs/quota/xfs_qm_syscalls.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/xfs/quota/xfs_qm_syscalls.c 2007-10-24 16:14:06.291582822 +0200 @@ -17,6 +17,7 @@ */ #include +#include #include "xfs.h" #include "xfs_fs.h" @@ -211,7 +212,7 @@ xfs_qm_scall_quotaoff( xfs_qoff_logitem_t *qoffstart; int nculprits; - if (!force && !capable(CAP_SYS_ADMIN)) + if (!force && !vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL)) return XFS_ERROR(EPERM); /* * No file system can have quotas enabled on disk but not in core. @@ -380,7 +381,7 @@ xfs_qm_scall_trunc_qfiles( int error; xfs_inode_t *qip; - if (!capable(CAP_SYS_ADMIN)) + if (!vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL)) return XFS_ERROR(EPERM); error = 0; if (!XFS_SB_VERSION_HASQUOTA(&mp->m_sb) || flags == 0) { @@ -425,7 +426,7 @@ xfs_qm_scall_quotaon( uint accflags; __int64_t sbflags; - if (!capable(CAP_SYS_ADMIN)) + if (!vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL)) return XFS_ERROR(EPERM); flags &= (XFS_ALL_QUOTA_ACCT | XFS_ALL_QUOTA_ENFD); @@ -594,7 +595,7 @@ xfs_qm_scall_setqlim( int error; xfs_qcnt_t hard, soft; - if (!capable(CAP_SYS_ADMIN)) + if (!vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL)) return XFS_ERROR(EPERM); if ((newlim->d_fieldmask & diff -NurpP --minimal linux-source-2.6.22.orig/fs/xfs/xfs_clnt.h linux-source-2.6.22/fs/xfs/xfs_clnt.h --- linux-source-2.6.22.orig/fs/xfs/xfs_clnt.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/xfs/xfs_clnt.h 2007-10-24 16:14:06.291582822 +0200 @@ -99,5 +99,7 @@ struct xfs_mount_args { */ #define XFSMNT2_COMPAT_IOSIZE 0x00000001 /* don't report large preferred * I/O size in stat(2) */ +#define XFSMNT2_TAGGED 0x80000000 /* context tagging */ + #endif /* __XFS_CLNT_H__ */ diff -NurpP --minimal linux-source-2.6.22.orig/fs/xfs/xfs_dinode.h linux-source-2.6.22/fs/xfs/xfs_dinode.h --- linux-source-2.6.22.orig/fs/xfs/xfs_dinode.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/xfs/xfs_dinode.h 2007-10-24 16:14:06.291582822 +0200 @@ -53,7 +53,8 @@ typedef struct xfs_dinode_core __uint32_t di_gid; /* owner's group id */ __uint32_t di_nlink; /* number of links to file */ __uint16_t di_projid; /* owner's project id */ - __uint8_t di_pad[8]; /* unused, zeroed space */ + __uint16_t di_tag; /* context tagging */ + __uint8_t di_pad[6]; /* unused, zeroed space */ __uint16_t di_flushiter; /* incremented on flush */ xfs_timestamp_t di_atime; /* time last accessed */ xfs_timestamp_t di_mtime; /* time last modified */ @@ -257,6 +258,9 @@ typedef enum xfs_dinode_fmt #define XFS_DIFLAG_EXTSIZE_BIT 11 /* inode extent size allocator hint */ #define XFS_DIFLAG_EXTSZINHERIT_BIT 12 /* inherit inode extent size */ #define XFS_DIFLAG_NODEFRAG_BIT 13 /* do not reorganize/defragment */ +#define XFS_DIFLAG_BARRIER_BIT 14 /* chroot() barrier */ +#define XFS_DIFLAG_IUNLINK_BIT 15 /* immutable unlink */ + #define XFS_DIFLAG_REALTIME (1 << XFS_DIFLAG_REALTIME_BIT) #define XFS_DIFLAG_PREALLOC (1 << XFS_DIFLAG_PREALLOC_BIT) #define XFS_DIFLAG_NEWRTBM (1 << XFS_DIFLAG_NEWRTBM_BIT) @@ -271,12 +275,15 @@ typedef enum xfs_dinode_fmt #define XFS_DIFLAG_EXTSIZE (1 << XFS_DIFLAG_EXTSIZE_BIT) #define XFS_DIFLAG_EXTSZINHERIT (1 << XFS_DIFLAG_EXTSZINHERIT_BIT) #define XFS_DIFLAG_NODEFRAG (1 << XFS_DIFLAG_NODEFRAG_BIT) +#define XFS_DIFLAG_BARRIER (1 << XFS_DIFLAG_BARRIER_BIT) +#define XFS_DIFLAG_IUNLINK (1 << XFS_DIFLAG_IUNLINK_BIT) #define XFS_DIFLAG_ANY \ (XFS_DIFLAG_REALTIME | XFS_DIFLAG_PREALLOC | XFS_DIFLAG_NEWRTBM | \ XFS_DIFLAG_IMMUTABLE | XFS_DIFLAG_APPEND | XFS_DIFLAG_SYNC | \ XFS_DIFLAG_NOATIME | XFS_DIFLAG_NODUMP | XFS_DIFLAG_RTINHERIT | \ XFS_DIFLAG_PROJINHERIT | XFS_DIFLAG_NOSYMLINKS | XFS_DIFLAG_EXTSIZE | \ - XFS_DIFLAG_EXTSZINHERIT | XFS_DIFLAG_NODEFRAG) + XFS_DIFLAG_EXTSZINHERIT | XFS_DIFLAG_NODEFRAG | XFS_DIFLAG_BARRIER | \ + XFS_DIFLAG_IUNLINK) #endif /* __XFS_DINODE_H__ */ diff -NurpP --minimal linux-source-2.6.22.orig/fs/xfs/xfs_fs.h linux-source-2.6.22/fs/xfs/xfs_fs.h --- linux-source-2.6.22.orig/fs/xfs/xfs_fs.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/xfs/xfs_fs.h 2007-10-24 16:14:06.291582822 +0200 @@ -66,6 +66,8 @@ struct fsxattr { #define XFS_XFLAG_EXTSIZE 0x00000800 /* extent size allocator hint */ #define XFS_XFLAG_EXTSZINHERIT 0x00001000 /* inherit inode extent size */ #define XFS_XFLAG_NODEFRAG 0x00002000 /* do not defragment */ +#define XFS_XFLAG_BARRIER 0x00004000 /* chroot() barrier */ +#define XFS_XFLAG_IUNLINK 0x00008000 /* immutable unlink */ #define XFS_XFLAG_HASATTR 0x80000000 /* no DIFLAG for this */ /* @@ -294,7 +296,8 @@ typedef struct xfs_bstat { __s32 bs_extents; /* number of extents */ __u32 bs_gen; /* generation count */ __u16 bs_projid; /* project id */ - unsigned char bs_pad[14]; /* pad space, unused */ + __u16 bs_tag; /* context tagging */ + unsigned char bs_pad[12]; /* pad space, unused */ __u32 bs_dmevmask; /* DMIG event mask */ __u16 bs_dmstate; /* DMIG state info */ __u16 bs_aextents; /* attribute number of extents */ diff -NurpP --minimal linux-source-2.6.22.orig/fs/xfs/xfs_inode.c linux-source-2.6.22/fs/xfs/xfs_inode.c --- linux-source-2.6.22.orig/fs/xfs/xfs_inode.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/xfs/xfs_inode.c 2007-10-24 16:14:06.291582822 +0200 @@ -49,6 +49,7 @@ #include "xfs_quota.h" #include "xfs_acl.h" +#include kmem_zone_t *xfs_ifork_zone; kmem_zone_t *xfs_inode_zone; @@ -736,20 +737,35 @@ xfs_xlate_dinode_core( xfs_dinode_core_t *buf_core = (xfs_dinode_core_t *)buf; xfs_dinode_core_t *mem_core = (xfs_dinode_core_t *)dip; xfs_arch_t arch = ARCH_CONVERT; + uint32_t uid = 0, gid = 0; + uint16_t tag = 0; ASSERT(dir); + if (dir < 0) { + tag = mem_core->di_tag; + /* FIXME: supposed to use superblock flag */ + uid = TAGINO_UID(1, mem_core->di_uid, tag); + gid = TAGINO_GID(1, mem_core->di_gid, tag); + tag = TAGINO_TAG(1, tag); + } + INT_XLATE(buf_core->di_magic, mem_core->di_magic, dir, arch); INT_XLATE(buf_core->di_mode, mem_core->di_mode, dir, arch); INT_XLATE(buf_core->di_version, mem_core->di_version, dir, arch); INT_XLATE(buf_core->di_format, mem_core->di_format, dir, arch); INT_XLATE(buf_core->di_onlink, mem_core->di_onlink, dir, arch); - INT_XLATE(buf_core->di_uid, mem_core->di_uid, dir, arch); - INT_XLATE(buf_core->di_gid, mem_core->di_gid, dir, arch); + INT_XLATE(buf_core->di_uid, uid, dir, arch); + INT_XLATE(buf_core->di_gid, gid, dir, arch); + INT_XLATE(buf_core->di_tag, tag, dir, arch); INT_XLATE(buf_core->di_nlink, mem_core->di_nlink, dir, arch); INT_XLATE(buf_core->di_projid, mem_core->di_projid, dir, arch); if (dir > 0) { + /* FIXME: supposed to use superblock flag */ + mem_core->di_uid = INOTAG_UID(1, uid, gid); + mem_core->di_gid = INOTAG_GID(1, uid, gid); + mem_core->di_tag = INOTAG_TAG(1, uid, gid, tag); memcpy(mem_core->di_pad, buf_core->di_pad, sizeof(buf_core->di_pad)); } else { @@ -797,6 +813,10 @@ _xfs_dic2xflags( flags |= XFS_XFLAG_PREALLOC; if (di_flags & XFS_DIFLAG_IMMUTABLE) flags |= XFS_XFLAG_IMMUTABLE; + if (di_flags & XFS_DIFLAG_IUNLINK) + flags |= XFS_XFLAG_IUNLINK; + if (di_flags & XFS_DIFLAG_BARRIER) + flags |= XFS_XFLAG_BARRIER; if (di_flags & XFS_DIFLAG_APPEND) flags |= XFS_XFLAG_APPEND; if (di_flags & XFS_DIFLAG_SYNC) @@ -1129,6 +1149,7 @@ xfs_ialloc( ASSERT(ip->i_d.di_nlink == nlink); ip->i_d.di_uid = current_fsuid(cr); ip->i_d.di_gid = current_fsgid(cr); + ip->i_d.di_tag = current_fstag(cr, vp); ip->i_d.di_projid = prid; memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad)); diff -NurpP --minimal linux-source-2.6.22.orig/fs/xfs/xfs_itable.c linux-source-2.6.22/fs/xfs/xfs_itable.c --- linux-source-2.6.22.orig/fs/xfs/xfs_itable.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/xfs/xfs_itable.c 2007-10-24 16:14:06.291582822 +0200 @@ -89,6 +89,7 @@ xfs_bulkstat_one_iget( buf->bs_mode = dic->di_mode; buf->bs_uid = dic->di_uid; buf->bs_gid = dic->di_gid; + buf->bs_tag = dic->di_tag; buf->bs_size = dic->di_size; vn_atime_to_bstime(vp, &buf->bs_atime); buf->bs_mtime.tv_sec = dic->di_mtime.t_sec; @@ -163,6 +164,7 @@ xfs_bulkstat_one_dinode( buf->bs_mode = INT_GET(dic->di_mode, ARCH_CONVERT); buf->bs_uid = INT_GET(dic->di_uid, ARCH_CONVERT); buf->bs_gid = INT_GET(dic->di_gid, ARCH_CONVERT); + buf->bs_tag = INT_GET(dic->di_tag, ARCH_CONVERT); buf->bs_size = INT_GET(dic->di_size, ARCH_CONVERT); buf->bs_atime.tv_sec = INT_GET(dic->di_atime.t_sec, ARCH_CONVERT); buf->bs_atime.tv_nsec = INT_GET(dic->di_atime.t_nsec, ARCH_CONVERT); diff -NurpP --minimal linux-source-2.6.22.orig/fs/xfs/xfs_mount.h linux-source-2.6.22/fs/xfs/xfs_mount.h --- linux-source-2.6.22.orig/fs/xfs/xfs_mount.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/xfs/xfs_mount.h 2007-10-24 16:14:06.291582822 +0200 @@ -464,6 +464,7 @@ typedef struct xfs_mount { #define XFS_MOUNT_NO_PERCPU_SB (1ULL << 23) /* don't use per-cpu superblock counters */ +#define XFS_MOUNT_TAGGED (1ULL << 31) /* context tagging */ /* * Default minimum read and write sizes. diff -NurpP --minimal linux-source-2.6.22.orig/fs/xfs/xfs_vfsops.c linux-source-2.6.22/fs/xfs/xfs_vfsops.c --- linux-source-2.6.22.orig/fs/xfs/xfs_vfsops.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/xfs/xfs_vfsops.c 2007-10-24 16:14:06.291582822 +0200 @@ -300,6 +300,8 @@ xfs_start_flags( if (ap->flags2 & XFSMNT2_COMPAT_IOSIZE) mp->m_flags |= XFS_MOUNT_COMPAT_IOSIZE; + if (ap->flags2 & XFSMNT2_TAGGED) + mp->m_flags |= XFS_MOUNT_TAGGED; /* * no recovery flag requires a read-only mount @@ -394,6 +396,8 @@ xfs_finish_flags( return XFS_ERROR(EINVAL); } + if (ap->flags2 & XFSMNT2_TAGGED) + vfs->vfs_super->s_flags |= MS_TAGGED; return 0; } @@ -1645,6 +1649,9 @@ xfs_vget( * in stat(). */ #define MNTOPT_ATTR2 "attr2" /* do use attr2 attribute format */ #define MNTOPT_NOATTR2 "noattr2" /* do not use attr2 attribute format */ +#define MNTOPT_TAGXID "tagxid" /* context tagging for inodes */ +#define MNTOPT_TAGGED "tag" /* context tagging for inodes */ +#define MNTOPT_NOTAGTAG "notag" /* do not use context tagging */ STATIC unsigned long suffix_strtoul(char *s, char **endp, unsigned int base) @@ -1831,6 +1838,19 @@ xfs_parseargs( args->flags |= XFSMNT_ATTR2; } else if (!strcmp(this_char, MNTOPT_NOATTR2)) { args->flags &= ~XFSMNT_ATTR2; +#ifndef CONFIG_TAGGING_NONE + } else if (!strcmp(this_char, MNTOPT_TAGGED)) { + args->flags2 |= XFSMNT2_TAGGED; + } else if (!strcmp(this_char, MNTOPT_NOTAGTAG)) { + args->flags2 &= ~XFSMNT2_TAGGED; + } else if (!strcmp(this_char, MNTOPT_TAGXID)) { + args->flags2 |= XFSMNT2_TAGGED; +#endif +#ifdef CONFIG_PROPAGATE + } else if (!strcmp(this_char, MNTOPT_TAGGED)) { + /* use value */ + args->flags2 |= XFSMNT2_TAGGED; +#endif } else if (!strcmp(this_char, "osyncisdsync")) { /* no-op, this is now the default */ cmn_err(CE_WARN, diff -NurpP --minimal linux-source-2.6.22.orig/fs/xfs/xfs_vnodeops.c linux-source-2.6.22/fs/xfs/xfs_vnodeops.c --- linux-source-2.6.22.orig/fs/xfs/xfs_vnodeops.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/fs/xfs/xfs_vnodeops.c 2007-10-24 16:14:06.291582822 +0200 @@ -159,6 +159,7 @@ xfs_getattr( vap->va_mode = ip->i_d.di_mode; vap->va_uid = ip->i_d.di_uid; vap->va_gid = ip->i_d.di_gid; + vap->va_tag = ip->i_d.di_tag; vap->va_projid = ip->i_d.di_projid; /* @@ -259,6 +260,7 @@ xfs_setattr( uint commit_flags=0; uid_t uid=0, iuid=0; gid_t gid=0, igid=0; + tag_t tag=0, itag=0; int timeflags = 0; bhv_vnode_t *vp; xfs_prid_t projid=0, iprojid=0; @@ -315,6 +317,7 @@ xfs_setattr( (mask & (XFS_AT_UID|XFS_AT_GID|XFS_AT_PROJID))) { uint qflags = 0; + /* TODO: handle tagging? */ if ((mask & XFS_AT_UID) && XFS_IS_UQUOTA_ON(mp)) { uid = vap->va_uid; qflags |= XFS_QMOPT_UQUOTA; @@ -394,6 +397,8 @@ xfs_setattr( if (mask & (XFS_AT_MODE|XFS_AT_XFLAGS|XFS_AT_EXTSIZE|XFS_AT_UID| XFS_AT_GID|XFS_AT_PROJID)) { + /* TODO: handle tagging? */ + /* * CAP_FOWNER overrides the following restrictions: * @@ -442,7 +447,7 @@ xfs_setattr( * and can change the group id only to a group of which he * or she is a member. */ - if (mask & (XFS_AT_UID|XFS_AT_GID|XFS_AT_PROJID)) { + if (mask & (XFS_AT_UID|XFS_AT_GID|XFS_AT_TAG|XFS_AT_PROJID)) { /* * These IDs could have changed since we last looked at them. * But, we're assured that if the ownership did change @@ -450,10 +455,12 @@ xfs_setattr( * would have changed also. */ iuid = ip->i_d.di_uid; - iprojid = ip->i_d.di_projid; igid = ip->i_d.di_gid; - gid = (mask & XFS_AT_GID) ? vap->va_gid : igid; + itag = ip->i_d.di_tag; + iprojid = ip->i_d.di_projid; uid = (mask & XFS_AT_UID) ? vap->va_uid : iuid; + gid = (mask & XFS_AT_GID) ? vap->va_gid : igid; + tag = (mask & XFS_AT_TAG) ? vap->va_tag : itag; projid = (mask & XFS_AT_PROJID) ? (xfs_prid_t)vap->va_projid : iprojid; @@ -481,6 +488,7 @@ xfs_setattr( if ((XFS_IS_UQUOTA_ON(mp) && iuid != uid) || (XFS_IS_PQUOTA_ON(mp) && iprojid != projid) || (XFS_IS_GQUOTA_ON(mp) && igid != gid)) { + /* TODO: handle tagging? */ ASSERT(tp); code = XFS_QM_DQVOPCHOWNRESV(mp, tp, ip, udqp, gdqp, capable(CAP_FOWNER) ? @@ -706,7 +714,7 @@ xfs_setattr( * and can change the group id only to a group of which he * or she is a member. */ - if (mask & (XFS_AT_UID|XFS_AT_GID|XFS_AT_PROJID)) { + if (mask & (XFS_AT_UID|XFS_AT_GID|XFS_AT_TAG|XFS_AT_PROJID)) { /* * CAP_FSETID overrides the following restrictions: * @@ -722,6 +730,9 @@ xfs_setattr( * Change the ownerships and register quota modifications * in the transaction. */ + if (itag != tag) { + ip->i_d.di_tag = tag; + } if (iuid != uid) { if (XFS_IS_UQUOTA_ON(mp)) { ASSERT(mask & XFS_AT_UID); @@ -802,6 +813,10 @@ xfs_setattr( di_flags = (ip->i_d.di_flags & XFS_DIFLAG_PREALLOC); if (vap->va_xflags & XFS_XFLAG_IMMUTABLE) di_flags |= XFS_DIFLAG_IMMUTABLE; + if (vap->va_xflags & XFS_XFLAG_IUNLINK) + di_flags |= XFS_DIFLAG_IUNLINK; + if (vap->va_xflags & XFS_XFLAG_BARRIER) + di_flags |= XFS_DIFLAG_BARRIER; if (vap->va_xflags & XFS_XFLAG_APPEND) di_flags |= XFS_DIFLAG_APPEND; if (vap->va_xflags & XFS_XFLAG_SYNC) diff -NurpP --minimal linux-source-2.6.22.orig/include/asm-arm/tlb.h linux-source-2.6.22/include/asm-arm/tlb.h --- linux-source-2.6.22.orig/include/asm-arm/tlb.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/asm-arm/tlb.h 2007-10-24 16:14:06.291582822 +0200 @@ -28,6 +28,7 @@ #else /* !CONFIG_MMU */ #include +#include /* * TLB handling. This allows us to remove pages from the page diff -NurpP --minimal linux-source-2.6.22.orig/include/asm-arm26/tlb.h linux-source-2.6.22/include/asm-arm26/tlb.h --- linux-source-2.6.22.orig/include/asm-arm26/tlb.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/asm-arm26/tlb.h 2007-10-24 16:14:06.291582822 +0200 @@ -3,6 +3,7 @@ #include #include +#include /* * TLB handling. This allows us to remove pages from the page diff -NurpP --minimal linux-source-2.6.22.orig/include/asm-arm26/unistd.h linux-source-2.6.22/include/asm-arm26/unistd.h --- linux-source-2.6.22.orig/include/asm-arm26/unistd.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/asm-arm26/unistd.h 2007-10-24 16:14:06.291582822 +0200 @@ -302,6 +302,8 @@ #define __NR_mq_getsetattr (__NR_SYSCALL_BASE+279) #define __NR_waitid (__NR_SYSCALL_BASE+280) +#define __NR_vserver (__NR_SYSCALL_BASE+313) + /* * The following SWIs are ARM private. FIXME - make appropriate for arm26 */ diff -NurpP --minimal linux-source-2.6.22.orig/include/asm-generic/tlb.h linux-source-2.6.22/include/asm-generic/tlb.h --- linux-source-2.6.22.orig/include/asm-generic/tlb.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/asm-generic/tlb.h 2007-10-24 16:14:06.291582822 +0200 @@ -14,6 +14,7 @@ #define _ASM_GENERIC__TLB_H #include +#include #include #include diff -NurpP --minimal linux-source-2.6.22.orig/include/asm-ia64/tlb.h linux-source-2.6.22/include/asm-ia64/tlb.h --- linux-source-2.6.22.orig/include/asm-ia64/tlb.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/asm-ia64/tlb.h 2007-10-24 16:14:06.291582822 +0200 @@ -40,6 +40,7 @@ #include #include #include +#include #include #include diff -NurpP --minimal linux-source-2.6.22.orig/include/asm-powerpc/systbl.h linux-source-2.6.22/include/asm-powerpc/systbl.h --- linux-source-2.6.22.orig/include/asm-powerpc/systbl.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/asm-powerpc/systbl.h 2007-10-24 16:14:06.291582822 +0200 @@ -260,7 +260,7 @@ COMPAT_SYS_SPU(fstatfs64) SYSX(sys_ni_syscall, ppc_fadvise64_64, ppc_fadvise64_64) PPC_SYS_SPU(rtas) OLDSYS(debug_setcontext) -SYSCALL(ni_syscall) +SYSX(sys_vserver, sys32_vserver, sys_vserver) COMPAT_SYS(migrate_pages) COMPAT_SYS(mbind) COMPAT_SYS(get_mempolicy) diff -NurpP --minimal linux-source-2.6.22.orig/include/asm-powerpc/unistd.h linux-source-2.6.22/include/asm-powerpc/unistd.h --- linux-source-2.6.22.orig/include/asm-powerpc/unistd.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/asm-powerpc/unistd.h 2007-10-24 16:14:06.291582822 +0200 @@ -275,7 +275,7 @@ #endif #define __NR_rtas 255 #define __NR_sys_debug_setcontext 256 -/* Number 257 is reserved for vserver */ +#define __NR_vserver 257 #define __NR_migrate_pages 258 #define __NR_mbind 259 #define __NR_get_mempolicy 260 diff -NurpP --minimal linux-source-2.6.22.orig/include/asm-s390/unistd.h linux-source-2.6.22/include/asm-s390/unistd.h --- linux-source-2.6.22.orig/include/asm-s390/unistd.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/asm-s390/unistd.h 2007-10-24 16:14:06.291582822 +0200 @@ -202,7 +202,7 @@ #define __NR_clock_gettime (__NR_timer_create+6) #define __NR_clock_getres (__NR_timer_create+7) #define __NR_clock_nanosleep (__NR_timer_create+8) -/* Number 263 is reserved for vserver */ +#define __NR_vserver 263 #define __NR_statfs64 265 #define __NR_fstatfs64 266 #define __NR_remap_file_pages 267 diff -NurpP --minimal linux-source-2.6.22.orig/include/asm-sparc/unistd.h linux-source-2.6.22/include/asm-sparc/unistd.h --- linux-source-2.6.22.orig/include/asm-sparc/unistd.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/asm-sparc/unistd.h 2007-10-24 16:14:06.291582822 +0200 @@ -283,7 +283,7 @@ #define __NR_timer_getoverrun 264 #define __NR_timer_delete 265 #define __NR_timer_create 266 -/* #define __NR_vserver 267 Reserved for VSERVER */ +#define __NR_vserver 267 #define __NR_io_setup 268 #define __NR_io_destroy 269 #define __NR_io_submit 270 diff -NurpP --minimal linux-source-2.6.22.orig/include/asm-sparc64/tlb.h linux-source-2.6.22/include/asm-sparc64/tlb.h --- linux-source-2.6.22.orig/include/asm-sparc64/tlb.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/asm-sparc64/tlb.h 2007-10-24 16:14:06.291582822 +0200 @@ -3,6 +3,7 @@ #include #include +#include #include #include #include diff -NurpP --minimal linux-source-2.6.22.orig/include/asm-sparc64/unistd.h linux-source-2.6.22/include/asm-sparc64/unistd.h --- linux-source-2.6.22.orig/include/asm-sparc64/unistd.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/asm-sparc64/unistd.h 2007-10-24 16:14:06.291582822 +0200 @@ -285,7 +285,7 @@ #define __NR_timer_getoverrun 264 #define __NR_timer_delete 265 #define __NR_timer_create 266 -/* #define __NR_vserver 267 Reserved for VSERVER */ +#define __NR_vserver 267 #define __NR_io_setup 268 #define __NR_io_destroy 269 #define __NR_io_submit 270 diff -NurpP --minimal linux-source-2.6.22.orig/include/asm-x86_64/unistd.h linux-source-2.6.22/include/asm-x86_64/unistd.h --- linux-source-2.6.22.orig/include/asm-x86_64/unistd.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/asm-x86_64/unistd.h 2007-10-24 16:14:06.291582822 +0200 @@ -532,7 +532,7 @@ __SYSCALL(__NR_tgkill, sys_tgkill) #define __NR_utimes 235 __SYSCALL(__NR_utimes, sys_utimes) #define __NR_vserver 236 -__SYSCALL(__NR_vserver, sys_ni_syscall) +__SYSCALL(__NR_vserver, sys_vserver) #define __NR_mbind 237 __SYSCALL(__NR_mbind, sys_mbind) #define __NR_set_mempolicy 238 diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/capability.h linux-source-2.6.22/include/linux/capability.h --- linux-source-2.6.22.orig/include/linux/capability.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/capability.h 2007-10-24 16:14:06.291582822 +0200 @@ -237,6 +237,7 @@ typedef __u32 kernel_cap_t; arbitrary SCSI commands */ /* Allow setting encryption key on loopback filesystem */ /* Allow setting zone reclaim policy */ +/* Allow the selection of a security context */ #define CAP_SYS_ADMIN 21 @@ -290,6 +291,11 @@ typedef __u32 kernel_cap_t; #define CAP_AUDIT_CONTROL 30 +/* Allow context manipulations */ +/* Allow changing context info on files */ + +#define CAP_CONTEXT 31 + #ifdef __KERNEL__ /* * Bounding set diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/devpts_fs.h linux-source-2.6.22/include/linux/devpts_fs.h --- linux-source-2.6.22.orig/include/linux/devpts_fs.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/devpts_fs.h 2007-10-24 16:14:06.291582822 +0200 @@ -30,5 +30,7 @@ static inline void devpts_pty_kill(int n #endif +#define DEVPTS_SUPER_MAGIC 0x00001cd1 + #endif /* _LINUX_DEVPTS_FS_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/ext2_fs.h linux-source-2.6.22/include/linux/ext2_fs.h --- linux-source-2.6.22.orig/include/linux/ext2_fs.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/ext2_fs.h 2007-10-24 16:14:06.291582822 +0200 @@ -188,6 +188,8 @@ struct ext2_group_desc #define EXT2_NOTAIL_FL FS_NOTAIL_FL /* file tail should not be merged */ #define EXT2_DIRSYNC_FL FS_DIRSYNC_FL /* dirsync behaviour (directories only) */ #define EXT2_TOPDIR_FL FS_TOPDIR_FL /* Top of directory hierarchies*/ +#define EXT2_BARRIER_FL FS_BARRIER_FL /* Barrier for chroot() */ +#define EXT2_IUNLINK_FL FS_IUNLINK_FL /* Immutable unlink */ #define EXT2_RESERVED_FL FS_RESERVED_FL /* reserved for ext2 lib */ #define EXT2_FL_USER_VISIBLE FS_FL_USER_VISIBLE /* User visible flags */ @@ -244,7 +246,7 @@ struct ext2_inode { struct { __u8 l_i_frag; /* Fragment number */ __u8 l_i_fsize; /* Fragment size */ - __u16 i_pad1; + __u16 l_i_tag; /* Context Tag */ __le16 l_i_uid_high; /* these 2 fields */ __le16 l_i_gid_high; /* were reserved2[0] */ __u32 l_i_reserved2; @@ -276,6 +278,7 @@ struct ext2_inode { #define i_gid_low i_gid #define i_uid_high osd2.linux2.l_i_uid_high #define i_gid_high osd2.linux2.l_i_gid_high +#define i_raw_tag osd2.linux2.l_i_tag #define i_reserved2 osd2.linux2.l_i_reserved2 #endif @@ -317,8 +320,9 @@ struct ext2_inode { #define EXT2_MOUNT_XATTR_USER 0x004000 /* Extended user attributes */ #define EXT2_MOUNT_POSIX_ACL 0x008000 /* POSIX Access Control Lists */ #define EXT2_MOUNT_XIP 0x010000 /* Execute in place */ -#define EXT2_MOUNT_USRQUOTA 0x020000 /* user quota */ -#define EXT2_MOUNT_GRPQUOTA 0x040000 /* group quota */ +#define EXT2_MOUNT_USRQUOTA 0x020000 /* user quota */ +#define EXT2_MOUNT_GRPQUOTA 0x040000 /* group quota */ +#define EXT2_MOUNT_TAGGED (1<<24) /* Enable Context Tags */ #define clear_opt(o, opt) o &= ~EXT2_MOUNT_##opt diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/ext3_fs.h linux-source-2.6.22/include/linux/ext3_fs.h --- linux-source-2.6.22.orig/include/linux/ext3_fs.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/ext3_fs.h 2007-10-24 16:14:06.291582822 +0200 @@ -177,10 +177,20 @@ struct ext3_group_desc #define EXT3_NOTAIL_FL 0x00008000 /* file tail should not be merged */ #define EXT3_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */ #define EXT3_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/ +#define EXT3_BARRIER_FL 0x04000000 /* Barrier for chroot() */ +#define EXT3_IUNLINK_FL 0x08000000 /* Immutable unlink */ #define EXT3_RESERVED_FL 0x80000000 /* reserved for ext3 lib */ +#ifdef CONFIG_VSERVER_LEGACY +#define EXT3_FL_USER_VISIBLE 0x0803DFFF /* User visible flags */ +#define EXT3_FL_USER_MODIFIABLE 0x080380FF /* User modifiable flags */ +#else #define EXT3_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */ #define EXT3_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */ +#endif +#ifdef CONFIG_VSERVER_LEGACY +#define EXT3_IOC_SETTAG FIOC_SETTAGJ +#endif /* * Inode dynamic state flags @@ -296,7 +306,7 @@ struct ext3_inode { struct { __u8 l_i_frag; /* Fragment number */ __u8 l_i_fsize; /* Fragment size */ - __u16 i_pad1; + __u16 l_i_tag; /* Context Tag */ __le16 l_i_uid_high; /* these 2 fields */ __le16 l_i_gid_high; /* were reserved2[0] */ __u32 l_i_reserved2; @@ -330,6 +340,7 @@ struct ext3_inode { #define i_gid_low i_gid #define i_uid_high osd2.linux2.l_i_uid_high #define i_gid_high osd2.linux2.l_i_gid_high +#define i_raw_tag osd2.linux2.l_i_tag #define i_reserved2 osd2.linux2.l_i_reserved2 #elif defined(__GNU__) @@ -384,6 +395,7 @@ struct ext3_inode { #define EXT3_MOUNT_QUOTA 0x80000 /* Some quota option set */ #define EXT3_MOUNT_USRQUOTA 0x100000 /* "old" user quota */ #define EXT3_MOUNT_GRPQUOTA 0x200000 /* "old" group quota */ +#define EXT3_MOUNT_TAGGED (1<<24) /* Enable Context Tags */ /* Compatibility, for having both ext2_fs.h and ext3_fs.h included at once */ #ifndef _LINUX_EXT2_FS_H @@ -812,6 +824,7 @@ struct buffer_head * ext3_bread (handle_ int ext3_get_blocks_handle(handle_t *handle, struct inode *inode, sector_t iblock, unsigned long maxblocks, struct buffer_head *bh_result, int create, int extend_disksize); +extern int ext3_sync_flags(struct inode *inode); extern void ext3_read_inode (struct inode *); extern int ext3_write_inode (struct inode *, int); diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/ext4_fs.h linux-source-2.6.22/include/linux/ext4_fs.h --- linux-source-2.6.22.orig/include/linux/ext4_fs.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/ext4_fs.h 2007-10-24 16:14:06.291582822 +0200 @@ -189,11 +189,21 @@ struct ext4_group_desc #define EXT4_NOTAIL_FL 0x00008000 /* file tail should not be merged */ #define EXT4_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */ #define EXT4_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/ -#define EXT4_RESERVED_FL 0x80000000 /* reserved for ext4 lib */ #define EXT4_EXTENTS_FL 0x00080000 /* Inode uses extents */ +#define EXT4_BARRIER_FL 0x04000000 /* Barrier for chroot() */ +#define EXT4_IUNLINK_FL 0x08000000 /* Immutable unlink */ +#define EXT4_RESERVED_FL 0x80000000 /* reserved for ext4 lib */ +#ifdef CONFIG_VSERVER_LEGACY +#define EXT4_FL_USER_VISIBLE 0x080BDFFF /* User visible flags */ +#define EXT4_FL_USER_MODIFIABLE 0x080380FF /* User modifiable flags */ +#else #define EXT4_FL_USER_VISIBLE 0x000BDFFF /* User visible flags */ #define EXT4_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */ +#endif +#ifdef CONFIG_VSERVER_LEGACY +#define EXT4_IOC_SETTAG FIOC_SETTAGJ +#endif /* * Inode dynamic state flags @@ -312,7 +322,8 @@ struct ext4_inode { __le16 l_i_file_acl_high; __le16 l_i_uid_high; /* these 2 fields */ __le16 l_i_gid_high; /* were reserved2[0] */ - __u32 l_i_reserved2; + __u16 l_i_tag; /* Context Tag */ + __u16 l_i_reserved2; } linux2; struct { __u8 h_i_frag; /* Fragment number */ @@ -344,6 +355,7 @@ struct ext4_inode { #define i_gid_low i_gid #define i_uid_high osd2.linux2.l_i_uid_high #define i_gid_high osd2.linux2.l_i_gid_high +#define i_raw_tag osd2.linux2.l_i_tag #define i_reserved2 osd2.linux2.l_i_reserved2 #elif defined(__GNU__) @@ -400,6 +412,7 @@ struct ext4_inode { #define EXT4_MOUNT_USRQUOTA 0x100000 /* "old" user quota */ #define EXT4_MOUNT_GRPQUOTA 0x200000 /* "old" group quota */ #define EXT4_MOUNT_EXTENTS 0x400000 /* Extents support */ +#define EXT4_MOUNT_TAGGED (1<<24) /* Enable Context Tags */ /* Compatibility, for having both ext2_fs.h and ext4_fs.h included at once */ #ifndef _LINUX_EXT2_FS_H @@ -850,6 +863,7 @@ struct buffer_head * ext4_bread (handle_ int ext4_get_blocks_handle(handle_t *handle, struct inode *inode, sector_t iblock, unsigned long maxblocks, struct buffer_head *bh_result, int create, int extend_disksize); +extern int ext4_sync_flags(struct inode *inode); extern void ext4_read_inode (struct inode *); extern int ext4_write_inode (struct inode *, int); diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/fs.h linux-source-2.6.22/include/linux/fs.h --- linux-source-2.6.22.orig/include/linux/fs.h 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/include/linux/fs.h 2007-10-24 16:14:06.291582822 +0200 @@ -123,6 +123,8 @@ extern int dir_notify_enable; #define MS_SLAVE (1<<19) /* change to slave */ #define MS_SHARED (1<<20) /* change to shared */ #define MS_RELATIME (1<<21) /* Update atime relative to mtime/ctime. */ +#define MS_TAGGED (1<<24) /* use generic inode tagging */ +#define MS_TAGID (1<<25) /* use specific tag for this mount */ #define MS_ACTIVE (1<<30) #define MS_NOUSER (1<<31) @@ -149,6 +151,8 @@ extern int dir_notify_enable; #define S_NOCMTIME 128 /* Do not update file c/mtime */ #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */ #define S_PRIVATE 512 /* Inode is fs-internal */ +#define S_BARRIER 1024 /* Barrier for chroot() */ +#define S_IUNLINK 2048 /* Immutable unlink */ /* * Note that nosuid etc flags are inode-specific: setting some file-system @@ -165,24 +169,36 @@ extern int dir_notify_enable; */ #define __IS_FLG(inode,flg) ((inode)->i_sb->s_flags & (flg)) -#define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY) +#define IS_RDONLY(inode) __IS_FLG(inode, MS_RDONLY) #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \ ((inode)->i_flags & S_SYNC)) #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \ ((inode)->i_flags & (S_SYNC|S_DIRSYNC))) #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK) -#define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME) +#define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME) +#define IS_TAGGED(inode) __IS_FLG(inode, MS_TAGGED) #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA) #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND) #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE) +#define IS_IUNLINK(inode) ((inode)->i_flags & S_IUNLINK) +#define IS_IXORUNLINK(inode) ((IS_IUNLINK(inode) ? S_IMMUTABLE : 0) ^ IS_IMMUTABLE(inode)) #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL) +#define IS_BARRIER(inode) (S_ISDIR((inode)->i_mode) && ((inode)->i_flags & S_BARRIER)) #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD) #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME) #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE) #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE) +#ifdef CONFIG_VSERVER_COWBL +# define IS_COW(inode) (IS_IUNLINK(inode) && IS_IMMUTABLE(inode)) +# define IS_COW_LINK(inode) (S_ISREG((inode)->i_mode) && ((inode)->i_nlink > 1)) +#else +# define IS_COW(inode) (0) +# define IS_COW_LINK(inode) (0) +#endif + /* the read-only stuff doesn't really belong here, but any other place is probably as bad and I don't want to create yet another include file. */ @@ -256,11 +272,17 @@ extern int dir_notify_enable; #define FS_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/ #define FS_EXTENT_FL 0x00080000 /* Extents */ #define FS_DIRECTIO_FL 0x00100000 /* Use direct i/o */ +#define FS_BARRIER_FL 0x04000000 /* Barrier for chroot() */ +#define FS_IUNLINK_FL 0x08000000 /* Immutable unlink */ #define FS_RESERVED_FL 0x80000000 /* reserved for ext2 lib */ +#ifdef CONFIG_VSERVER_LEGACY +#define FS_FL_USER_VISIBLE 0x0803DFFF /* User visible flags */ +#define FS_FL_USER_MODIFIABLE 0x080380FF /* User modifiable flags */ +#else #define FS_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */ #define FS_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */ - +#endif #define SYNC_FILE_RANGE_WAIT_BEFORE 1 #define SYNC_FILE_RANGE_WRITE 2 @@ -327,6 +349,7 @@ typedef void (dio_iodone_t)(struct kiocb #define ATTR_KILL_SUID 2048 #define ATTR_KILL_SGID 4096 #define ATTR_FILE 8192 +#define ATTR_TAG 16384 /* * This is the Inode Attributes structure, used for notify_change(). It @@ -342,6 +365,7 @@ struct iattr { umode_t ia_mode; uid_t ia_uid; gid_t ia_gid; + tag_t ia_tag; loff_t ia_size; struct timespec ia_atime; struct timespec ia_mtime; @@ -358,6 +382,9 @@ struct iattr { struct file *ia_file; }; +#define ATTR_FLAG_BARRIER 512 /* Barrier for chroot() */ +#define ATTR_FLAG_IUNLINK 1024 /* Immutable unlink */ + /* * Includes for diskquotas. */ @@ -540,6 +567,7 @@ struct inode { unsigned int i_nlink; uid_t i_uid; gid_t i_gid; + tag_t i_tag; dev_t i_rdev; unsigned long i_version; loff_t i_size; @@ -731,6 +759,7 @@ struct file { loff_t f_pos; struct fown_struct f_owner; unsigned int f_uid, f_gid; + xid_t f_xid; struct file_ra_state f_ra; unsigned long f_version; @@ -814,6 +843,7 @@ struct file_lock { unsigned char fl_type; loff_t fl_start; loff_t fl_end; + xid_t fl_xid; struct fasync_struct * fl_fasync; /* for lease break notifications */ unsigned long fl_break_time; /* for nonblocking lease breaks */ @@ -996,12 +1026,12 @@ extern void unlock_super(struct super_bl */ extern int vfs_permission(struct nameidata *, int); extern int vfs_create(struct inode *, struct dentry *, int, struct nameidata *); -extern int vfs_mkdir(struct inode *, struct dentry *, struct vfsmount *, int); -extern int vfs_mknod(struct inode *, struct dentry *, struct vfsmount *, int, dev_t); -extern int vfs_symlink(struct inode *, struct dentry *, struct vfsmount *, const char *, int); -extern int vfs_link(struct dentry *, struct vfsmount *, struct inode *, struct dentry *, struct vfsmount *); -extern int vfs_rmdir(struct inode *, struct dentry *, struct vfsmount *); -extern int vfs_unlink(struct inode *, struct dentry *, struct vfsmount *); +extern int vfs_mkdir(struct inode *, struct dentry *, struct vfsmount *, int, struct nameidata *); +extern int vfs_mknod(struct inode *, struct dentry *, struct vfsmount *, int, dev_t, struct nameidata *); +extern int vfs_symlink(struct inode *, struct dentry *, struct vfsmount *, const char *, int, struct nameidata *); +extern int vfs_link(struct dentry *, struct vfsmount *, struct inode *, struct dentry *, struct vfsmount *, struct nameidata *); +extern int vfs_rmdir(struct inode *, struct dentry *, struct vfsmount *, struct nameidata *); +extern int vfs_unlink(struct inode *, struct dentry *, struct vfsmount *, struct nameidata *); extern int vfs_rename(struct inode *, struct dentry *, struct vfsmount *, struct inode *, struct dentry *, struct vfsmount *); /* @@ -1141,6 +1171,7 @@ struct inode_operations { ssize_t (*listxattr) (struct dentry *, char *, size_t); int (*removexattr) (struct dentry *, const char *); void (*truncate_range)(struct inode *, loff_t, loff_t); + int (*sync_flags) (struct inode *); }; struct seq_file; @@ -1156,6 +1187,7 @@ extern ssize_t vfs_readv(struct file *, unsigned long, loff_t *); extern ssize_t vfs_writev(struct file *, const struct iovec __user *, unsigned long, loff_t *); +ssize_t vfs_sendfile(struct file *, struct file *, loff_t *, size_t, loff_t); /* * NOTE: write_inode, delete_inode, clear_inode, put_inode can be called @@ -1766,6 +1798,7 @@ extern ssize_t generic_file_buffered_wri extern ssize_t do_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos); extern ssize_t do_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos); extern ssize_t generic_file_sendfile(struct file *, loff_t *, size_t, read_actor_t, void *); +extern ssize_t generic_file_sendpage(struct file *, struct page *, int, size_t, loff_t *, int); extern void do_generic_mapping_read(struct address_space *mapping, struct file_ra_state *, struct file *, loff_t *, read_descriptor_t *, read_actor_t); @@ -1901,6 +1934,7 @@ extern int dcache_dir_open(struct inode extern int dcache_dir_close(struct inode *, struct file *); extern loff_t dcache_dir_lseek(struct file *, loff_t, int); extern int dcache_readdir(struct file *, void *, filldir_t); +extern int dcache_readdir_filter(struct file *, void *, filldir_t, int (*)(struct dentry *)); extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *); extern int simple_statfs(struct dentry *, struct kstatfs *); extern int simple_link(struct dentry *, struct inode *, struct dentry *); diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/init_task.h linux-source-2.6.22/include/linux/init_task.h --- linux-source-2.6.22.orig/include/linux/init_task.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/init_task.h 2007-10-24 16:14:06.291582822 +0200 @@ -169,6 +169,10 @@ extern struct group_info init_groups; }, \ INIT_TRACE_IRQFLAGS \ INIT_LOCKDEP \ + .xid = 0, \ + .vx_info = NULL, \ + .nid = 0, \ + .nx_info = NULL, \ } diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/ipc.h linux-source-2.6.22/include/linux/ipc.h --- linux-source-2.6.22.orig/include/linux/ipc.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/ipc.h 2007-10-24 16:14:06.291582822 +0200 @@ -63,6 +63,7 @@ struct kern_ipc_perm key_t key; uid_t uid; gid_t gid; + xid_t xid; uid_t cuid; gid_t cgid; mode_t mode; diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/Kbuild linux-source-2.6.22/include/linux/Kbuild --- linux-source-2.6.22.orig/include/linux/Kbuild 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/include/linux/Kbuild 2007-10-24 16:14:06.291582822 +0200 @@ -352,3 +352,6 @@ unifdef-y += xattr.h unifdef-y += xfrm.h objhdr-y += version.h + +header-y += vserver/ + diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/loop.h linux-source-2.6.22/include/linux/loop.h --- linux-source-2.6.22.orig/include/linux/loop.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/loop.h 2007-10-24 16:14:06.291582822 +0200 @@ -45,6 +45,7 @@ struct loop_device { struct loop_func_table *lo_encryption; __u32 lo_init[2]; uid_t lo_key_owner; /* Who set the key */ + xid_t lo_xid; int (*ioctl)(struct loop_device *, int cmd, unsigned long arg); diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/major.h linux-source-2.6.22/include/linux/major.h --- linux-source-2.6.22.orig/include/linux/major.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/major.h 2007-10-24 16:14:06.291582822 +0200 @@ -15,6 +15,7 @@ #define HD_MAJOR IDE0_MAJOR #define PTY_SLAVE_MAJOR 3 #define TTY_MAJOR 4 +#define VROOT_MAJOR 4 #define TTYAUX_MAJOR 5 #define LP_MAJOR 6 #define VCS_MAJOR 7 diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/mount.h linux-source-2.6.22/include/linux/mount.h --- linux-source-2.6.22.orig/include/linux/mount.h 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/include/linux/mount.h 2007-10-24 16:14:06.291582822 +0200 @@ -28,12 +28,16 @@ struct mnt_namespace; #define MNT_NOATIME 0x08 #define MNT_NODIRATIME 0x10 #define MNT_RELATIME 0x20 +#define MNT_RDONLY 0x40 + +#define MNT_IS_RDONLY(m) ((m) && ((m)->mnt_flags & MNT_RDONLY)) #define MNT_SHRINKABLE 0x100 #define MNT_SHARED 0x1000 /* if the vfsmount is a shared mount */ #define MNT_UNBINDABLE 0x2000 /* if the vfsmount is a unbindable mount */ #define MNT_PNODE_MASK 0x3000 /* propagation flag mask */ +#define MNT_TAGID 0x8000 struct vfsmount { struct list_head mnt_hash; @@ -61,6 +65,7 @@ struct vfsmount { atomic_t mnt_count; int mnt_expiry_mark; /* true if marked for expiry */ int mnt_pinned; + tag_t mnt_tag; /* tagging used for vfsmount */ }; static inline struct vfsmount *mntget(struct vfsmount *mnt) diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/net.h linux-source-2.6.22/include/linux/net.h --- linux-source-2.6.22.orig/include/linux/net.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/net.h 2007-10-24 16:14:06.291582822 +0200 @@ -63,6 +63,7 @@ typedef enum { #define SOCK_NOSPACE 2 #define SOCK_PASSCRED 3 #define SOCK_PASSSEC 4 +#define SOCK_USER_SOCKET 5 #ifndef ARCH_HAS_SOCKET_TYPES /** diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/nfs_mount.h linux-source-2.6.22/include/linux/nfs_mount.h --- linux-source-2.6.22.orig/include/linux/nfs_mount.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/nfs_mount.h 2007-10-24 16:14:06.291582822 +0200 @@ -62,6 +62,7 @@ struct nfs_mount_data { #define NFS_MOUNT_STRICTLOCK 0x1000 /* reserved for NFSv4 */ #define NFS_MOUNT_SECFLAVOUR 0x2000 /* 5 */ #define NFS_MOUNT_NORDIRPLUS 0x4000 /* 5 */ +#define NFS_MOUNT_TAGGED 0x8000 /* context tagging */ #define NFS_MOUNT_FLAGMASK 0xFFFF #endif diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/nsproxy.h linux-source-2.6.22/include/linux/nsproxy.h --- linux-source-2.6.22.orig/include/linux/nsproxy.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/nsproxy.h 2007-10-24 16:14:06.291582822 +0200 @@ -3,6 +3,7 @@ #include #include +#include struct mnt_namespace; struct uts_namespace; @@ -32,26 +33,46 @@ struct nsproxy { extern struct nsproxy init_nsproxy; int copy_namespaces(int flags, struct task_struct *tsk); +struct nsproxy *copy_nsproxy(struct nsproxy *orig); void get_task_namespaces(struct task_struct *tsk); void free_nsproxy(struct nsproxy *ns); int unshare_nsproxy_namespaces(unsigned long, struct nsproxy **, struct fs_struct *); -static inline void put_nsproxy(struct nsproxy *ns) +#define get_nsproxy(n) __get_nsproxy(n, __FILE__, __LINE__) + +static inline void __get_nsproxy(struct nsproxy *ns, + const char *_file, int _line) { + vxlprintk(VXD_CBIT(space, 0), "get_nsproxy(%p[%u])", + ns, atomic_read(&ns->count), _file, _line); + atomic_inc(&ns->count); +} + +#define put_nsproxy(n) __put_nsproxy(n, __FILE__, __LINE__) + +static inline void __put_nsproxy(struct nsproxy *ns, + const char *_file, int _line) +{ + vxlprintk(VXD_CBIT(space, 0), "put_nsproxy(%p[%u])", + ns, atomic_read(&ns->count), _file, _line); if (atomic_dec_and_test(&ns->count)) { free_nsproxy(ns); } } -static inline void exit_task_namespaces(struct task_struct *p) +#define exit_task_namespaces(p) __exit_task_namespaces(p, __FILE__, __LINE__) + +static inline void __exit_task_namespaces(struct task_struct *p, + const char *_file, int _line) { struct nsproxy *ns = p->nsproxy; if (ns) { task_lock(p); p->nsproxy = NULL; task_unlock(p); - put_nsproxy(ns); + __put_nsproxy(ns, _file, _line); } } + #endif diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/pid.h linux-source-2.6.22/include/linux/pid.h --- linux-source-2.6.22.orig/include/linux/pid.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/pid.h 2007-10-24 16:14:06.291582822 +0200 @@ -8,7 +8,8 @@ enum pid_type PIDTYPE_PID, PIDTYPE_PGID, PIDTYPE_SID, - PIDTYPE_MAX + PIDTYPE_MAX, + PIDTYPE_REALPID }; /* diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/proc_fs.h linux-source-2.6.22/include/linux/proc_fs.h --- linux-source-2.6.22.orig/include/linux/proc_fs.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/proc_fs.h 2007-10-24 16:14:06.291582822 +0200 @@ -54,6 +54,7 @@ struct proc_dir_entry { nlink_t nlink; uid_t uid; gid_t gid; + int vx_flags; loff_t size; const struct inode_operations *proc_iops; const struct file_operations *proc_fops; @@ -246,13 +247,20 @@ static inline void kclist_add(struct kco extern void kclist_add(struct kcore_list *, void *, size_t); #endif +struct vx_info; +struct nx_info; + union proc_op { int (*proc_get_link)(struct inode *, struct dentry **, struct vfsmount **); int (*proc_read)(struct task_struct *task, char *page); + int (*proc_vs_read)(char *page); + int (*proc_vxi_read)(struct vx_info *vxi, char *page); + int (*proc_nxi_read)(struct nx_info *nxi, char *page); }; struct proc_inode { struct pid *pid; + int vx_flags; int fd; union proc_op op; struct proc_dir_entry *pde; diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/reiserfs_fs.h linux-source-2.6.22/include/linux/reiserfs_fs.h --- linux-source-2.6.22.orig/include/linux/reiserfs_fs.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/reiserfs_fs.h 2007-10-24 16:14:06.291582822 +0200 @@ -821,6 +821,10 @@ struct stat_data_v1 { #define REISERFS_COMPR_FL FS_COMPR_FL #define REISERFS_NOTAIL_FL FS_NOTAIL_FL +/* unfortunately reiserfs sdattr is only 16 bit */ +#define REISERFS_BARRIER_FL (FS_BARRIER_FL >> 16) +#define REISERFS_IUNLINK_FL (FS_IUNLINK_FL >> 16) + /* persistent flags that file inherits from the parent directory */ #define REISERFS_INHERIT_MASK ( REISERFS_IMMUTABLE_FL | \ REISERFS_SYNC_FL | \ @@ -830,6 +834,14 @@ struct stat_data_v1 { REISERFS_COMPR_FL | \ REISERFS_NOTAIL_FL ) +#ifdef CONFIG_VSERVER_LEGACY +#define REISERFS_FL_USER_VISIBLE (REISERFS_IUNLINK_FL|0x80FF) +#define REISERFS_FL_USER_MODIFIABLE (REISERFS_IUNLINK_FL|0x80FF) +#else +#define REISERFS_FL_USER_VISIBLE 0x80FF +#define REISERFS_FL_USER_MODIFIABLE 0x80FF +#endif + /* Stat Data on disk (reiserfs version of UFS disk inode minus the address blocks) */ struct stat_data { @@ -1901,6 +1913,7 @@ static inline void reiserfs_update_sd(st void sd_attrs_to_i_attrs(__u16 sd_attrs, struct inode *inode); void i_attrs_to_sd_attrs(struct inode *inode, __u16 * sd_attrs); int reiserfs_setattr(struct dentry *dentry, struct iattr *attr); +int reiserfs_sync_flags(struct inode *inode); /* namei.c */ void set_de_name_and_namelen(struct reiserfs_dir_entry *de); diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/reiserfs_fs_sb.h linux-source-2.6.22/include/linux/reiserfs_fs_sb.h --- linux-source-2.6.22.orig/include/linux/reiserfs_fs_sb.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/reiserfs_fs_sb.h 2007-10-24 16:14:06.291582822 +0200 @@ -458,6 +458,7 @@ enum reiserfs_mount_options { REISERFS_POSIXACL, REISERFS_BARRIER_NONE, REISERFS_BARRIER_FLUSH, + REISERFS_TAGGED, /* Actions on error */ REISERFS_ERROR_PANIC, diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/sched.h linux-source-2.6.22/include/linux/sched.h --- linux-source-2.6.22.orig/include/linux/sched.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/sched.h 2007-10-24 16:14:06.291582822 +0200 @@ -26,6 +26,7 @@ #define CLONE_STOPPED 0x02000000 /* Start in stopped state */ #define CLONE_NEWUTS 0x04000000 /* New utsname group? */ #define CLONE_NEWIPC 0x08000000 /* New ipcs */ +#define CLONE_KTHREAD 0x10000000 /* clone a kernel thread */ /* * Scheduling policies @@ -94,7 +95,7 @@ struct bio; * List of flags we want to share for kernel threads, * if only because they are not used by them anyway. */ -#define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND) +#define CLONE_KERNEL (CLONE_FS|CLONE_FILES|CLONE_SIGHAND|CLONE_KTHREAD) /* * These are the constant used to fake the fixed-point load-average @@ -146,12 +147,13 @@ extern unsigned long weighted_cpuload(co #define TASK_UNINTERRUPTIBLE 2 #define TASK_STOPPED 4 #define TASK_TRACED 8 +#define TASK_ONHOLD 16 /* in tsk->exit_state */ -#define EXIT_ZOMBIE 16 -#define EXIT_DEAD 32 +#define EXIT_ZOMBIE 32 +#define EXIT_DEAD 64 /* in tsk->state again */ -#define TASK_NONINTERACTIVE 64 -#define TASK_DEAD 128 +#define TASK_NONINTERACTIVE 128 +#define TASK_DEAD 256 #define __set_task_state(tsk, state_value) \ do { (tsk)->state = (state_value); } while (0) @@ -287,27 +289,30 @@ extern void arch_unmap_area_topdown(stru * The mm counters are not protected by its page_table_lock, * so must be incremented atomically. */ -#define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value) -#define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member)) -#define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member) -#define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member) -#define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member) typedef atomic_long_t mm_counter_t; +#define __set_mm_counter(mm, member, value) \ + atomic_long_set(&(mm)->_##member, value) +#define get_mm_counter(mm, member) \ + ((unsigned long)atomic_long_read(&(mm)->_##member)) #else /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */ /* * The mm counters are protected by its page_table_lock, * so can be incremented directly. */ -#define set_mm_counter(mm, member, value) (mm)->_##member = (value) -#define get_mm_counter(mm, member) ((mm)->_##member) -#define add_mm_counter(mm, member, value) (mm)->_##member += (value) -#define inc_mm_counter(mm, member) (mm)->_##member++ -#define dec_mm_counter(mm, member) (mm)->_##member-- typedef unsigned long mm_counter_t; +#define __set_mm_counter(mm, member, value) (mm)->_##member = (value) +#define get_mm_counter(mm, member) ((mm)->_##member) #endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */ +#define set_mm_counter(mm, member, value) \ + vx_ ## member ## pages_sub((mm), (get_mm_counter(mm, member) - value)) +#define add_mm_counter(mm, member, value) \ + vx_ ## member ## pages_add((mm), (value)) +#define inc_mm_counter(mm, member) vx_ ## member ## pages_inc((mm)) +#define dec_mm_counter(mm, member) vx_ ## member ## pages_dec((mm)) + #define get_mm_rss(mm) \ (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss)) #define update_hiwater_rss(mm) do { \ @@ -365,6 +370,7 @@ struct mm_struct { /* Architecture-specific MM context */ mm_context_t context; + struct vx_info *mm_vx_info; /* Swap token stuff */ /* @@ -570,9 +576,10 @@ struct user_struct { /* Hash table maintenance information */ struct list_head uidhash_list; uid_t uid; + xid_t xid; }; -extern struct user_struct *find_user(uid_t); +extern struct user_struct *find_user(xid_t, uid_t); extern struct user_struct root_user; #define INIT_USER (&root_user) @@ -969,6 +976,14 @@ struct task_struct { void *security; struct audit_context *audit_context; + +/* vserver context data */ + struct vx_info *vx_info; + struct nx_info *nx_info; + + xid_t xid; + nid_t nid; + seccomp_t seccomp; /* Thread group tracking */ @@ -1290,12 +1305,16 @@ extern struct task_struct init_task; extern struct mm_struct init_mm; -#define find_task_by_pid(nr) find_task_by_pid_type(PIDTYPE_PID, nr) +#define find_task_by_real_pid(nr) \ + find_task_by_pid_type(PIDTYPE_REALPID, nr) +#define find_task_by_pid(nr) \ + find_task_by_pid_type(PIDTYPE_PID, nr) + extern struct task_struct *find_task_by_pid_type(int type, int pid); extern void __set_special_pids(pid_t session, pid_t pgrp); /* per-UID process charging. */ -extern struct user_struct * alloc_uid(uid_t); +extern struct user_struct * alloc_uid(xid_t, uid_t); static inline struct user_struct *get_uid(struct user_struct *u) { atomic_inc(&u->__count); diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/shmem_fs.h linux-source-2.6.22/include/linux/shmem_fs.h --- linux-source-2.6.22.orig/include/linux/shmem_fs.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/shmem_fs.h 2007-10-24 16:14:06.291582822 +0200 @@ -8,6 +8,9 @@ #define SHMEM_NR_DIRECT 16 +#define TMPFS_SUPER_MAGIC 0x01021994 + + struct shmem_inode_info { spinlock_t lock; unsigned long flags; diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/stat.h linux-source-2.6.22/include/linux/stat.h --- linux-source-2.6.22.orig/include/linux/stat.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/stat.h 2007-10-24 16:14:06.291582822 +0200 @@ -66,6 +66,7 @@ struct kstat { unsigned int nlink; uid_t uid; gid_t gid; + tag_t tag; dev_t rdev; loff_t size; struct timespec atime; diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/sunrpc/auth.h linux-source-2.6.22/include/linux/sunrpc/auth.h --- linux-source-2.6.22.orig/include/linux/sunrpc/auth.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/sunrpc/auth.h 2007-10-24 16:14:06.291582822 +0200 @@ -24,6 +24,7 @@ struct auth_cred { uid_t uid; gid_t gid; + tag_t tag; struct group_info *group_info; }; diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/sunrpc/clnt.h linux-source-2.6.22/include/linux/sunrpc/clnt.h --- linux-source-2.6.22.orig/include/linux/sunrpc/clnt.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/sunrpc/clnt.h 2007-10-24 16:14:06.291582822 +0200 @@ -43,7 +43,8 @@ struct rpc_clnt { cl_discrtry : 1,/* disconnect before retry */ cl_autobind : 1,/* use getport() */ cl_oneshot : 1,/* dispose after use */ - cl_dead : 1;/* abandoned */ + cl_dead : 1,/* abandoned */ + cl_tag : 1;/* context tagging */ struct rpc_rtt * cl_rtt; /* RTO estimator data */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/syscalls.h linux-source-2.6.22/include/linux/syscalls.h --- linux-source-2.6.22.orig/include/linux/syscalls.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/syscalls.h 2007-10-24 16:14:06.291582822 +0200 @@ -294,6 +294,8 @@ asmlinkage long sys_symlink(const char _ asmlinkage long sys_unlink(const char __user *pathname); asmlinkage long sys_rename(const char __user *oldname, const char __user *newname); +asmlinkage long sys_copyfile(const char __user *from, const char __user *to, + umode_t mode); asmlinkage long sys_chmod(const char __user *filename, mode_t mode); asmlinkage long sys_fchmod(unsigned int fd, mode_t mode); diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/sysctl.h linux-source-2.6.22/include/linux/sysctl.h --- linux-source-2.6.22.orig/include/linux/sysctl.h 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/include/linux/sysctl.h 2007-10-24 16:14:06.291582822 +0200 @@ -106,6 +106,7 @@ enum KERN_CAP_BSET=14, /* int: capability bounding set */ KERN_PANIC=15, /* int: panic timeout */ KERN_REALROOTDEV=16, /* real root device to mount after initrd */ + KERN_VSHELPER=17, /* string: path to vshelper policy agent */ KERN_SPARC_REBOOT=21, /* reboot command on Sparc */ KERN_CTLALTDEL=22, /* int: allow ctl-alt-del to reboot */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/sysfs.h linux-source-2.6.22/include/linux/sysfs.h --- linux-source-2.6.22.orig/include/linux/sysfs.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/sysfs.h 2007-10-24 16:14:06.291582822 +0200 @@ -15,6 +15,8 @@ #include #include +#define SYSFS_SUPER_MAGIC 0x62656572 + struct kobject; struct module; struct nameidata; diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/time.h linux-source-2.6.22/include/linux/time.h --- linux-source-2.6.22.orig/include/linux/time.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/time.h 2007-10-24 16:14:06.291582822 +0200 @@ -176,6 +176,9 @@ static inline void timespec_add_ns(struc } a->tv_nsec = ns; } + +#include + #endif /* __KERNEL__ */ #define NFDBITS __NFDBITS diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/types.h linux-source-2.6.22/include/linux/types.h --- linux-source-2.6.22.orig/include/linux/types.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/linux/types.h 2007-10-24 16:14:06.291582822 +0200 @@ -39,6 +39,9 @@ typedef __kernel_uid32_t uid_t; typedef __kernel_gid32_t gid_t; typedef __kernel_uid16_t uid16_t; typedef __kernel_gid16_t gid16_t; +typedef unsigned int xid_t; +typedef unsigned int nid_t; +typedef unsigned int tag_t; #ifdef CONFIG_UID16 /* This is defined by include/asm-{arch}/posix_types.h */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vroot.h linux-source-2.6.22/include/linux/vroot.h --- linux-source-2.6.22.orig/include/linux/vroot.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vroot.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,51 @@ + +/* + * include/linux/vroot.h + * + * written by Herbert Pötzl, 9/11/2002 + * ported to 2.6 by Herbert Pötzl, 30/12/2004 + * + * Copyright (C) 2002-2007 by Herbert Pötzl. + * Redistribution of this file is permitted under the + * GNU General Public License. + */ + +#ifndef _LINUX_VROOT_H +#define _LINUX_VROOT_H + + +#ifdef __KERNEL__ + +/* Possible states of device */ +enum { + Vr_unbound, + Vr_bound, +}; + +struct vroot_device { + int vr_number; + int vr_refcnt; + + struct semaphore vr_ctl_mutex; + struct block_device *vr_device; + int vr_state; +}; + + +typedef struct block_device *(vroot_grb_func)(struct block_device *); + +extern int register_vroot_grb(vroot_grb_func *); +extern int unregister_vroot_grb(vroot_grb_func *); + +#endif /* __KERNEL__ */ + +#define MAX_VROOT_DEFAULT 8 + +/* + * IOCTL commands --- we will commandeer 0x56 ('V') + */ + +#define VROOT_SET_DEV 0x5600 +#define VROOT_CLR_DEV 0x5601 + +#endif /* _LINUX_VROOT_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vs_base.h linux-source-2.6.22/include/linux/vs_base.h --- linux-source-2.6.22.orig/include/linux/vs_base.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vs_base.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,9 @@ +#ifndef _VS_BASE_H +#define _VS_BASE_H + +#include "vserver/base.h" +#include "vserver/debug.h" + +#else +#warning duplicate inclusion +#endif diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vs_context.h linux-source-2.6.22/include/linux/vs_context.h --- linux-source-2.6.22.orig/include/linux/vs_context.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vs_context.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,224 @@ +#ifndef _VS_CONTEXT_H +#define _VS_CONTEXT_H + +#include "vserver/base.h" +#include "vserver/context.h" +#include "vserver/history.h" +#include "vserver/debug.h" + + +#define get_vx_info(i) __get_vx_info(i, __FILE__, __LINE__, __HERE__) + +static inline struct vx_info *__get_vx_info(struct vx_info *vxi, + const char *_file, int _line, void *_here) +{ + if (!vxi) + return NULL; + + vxlprintk(VXD_CBIT(xid, 2), "get_vx_info(%p[#%d.%d])", + vxi, vxi ? vxi->vx_id : 0, + vxi ? atomic_read(&vxi->vx_usecnt) : 0, + _file, _line); + __vxh_get_vx_info(vxi, _here); + + atomic_inc(&vxi->vx_usecnt); + return vxi; +} + + +extern void free_vx_info(struct vx_info *); + +#define put_vx_info(i) __put_vx_info(i, __FILE__, __LINE__, __HERE__) + +static inline void __put_vx_info(struct vx_info *vxi, + const char *_file, int _line, void *_here) +{ + if (!vxi) + return; + + vxlprintk(VXD_CBIT(xid, 2), "put_vx_info(%p[#%d.%d])", + vxi, vxi ? vxi->vx_id : 0, + vxi ? atomic_read(&vxi->vx_usecnt) : 0, + _file, _line); + __vxh_put_vx_info(vxi, _here); + + if (atomic_dec_and_test(&vxi->vx_usecnt)) + free_vx_info(vxi); +} + + +#define init_vx_info(p, i) \ + __init_vx_info(p, i, __FILE__, __LINE__, __HERE__) + +static inline void __init_vx_info(struct vx_info **vxp, struct vx_info *vxi, + const char *_file, int _line, void *_here) +{ + if (vxi) { + vxlprintk(VXD_CBIT(xid, 3), + "init_vx_info(%p[#%d.%d])", + vxi, vxi ? vxi->vx_id : 0, + vxi ? atomic_read(&vxi->vx_usecnt) : 0, + _file, _line); + __vxh_init_vx_info(vxi, vxp, _here); + + atomic_inc(&vxi->vx_usecnt); + } + *vxp = vxi; +} + + +#define set_vx_info(p, i) \ + __set_vx_info(p, i, __FILE__, __LINE__, __HERE__) + +static inline void __set_vx_info(struct vx_info **vxp, struct vx_info *vxi, + const char *_file, int _line, void *_here) +{ + struct vx_info *vxo; + + if (!vxi) + return; + + vxlprintk(VXD_CBIT(xid, 3), "set_vx_info(%p[#%d.%d])", + vxi, vxi ? vxi->vx_id : 0, + vxi ? atomic_read(&vxi->vx_usecnt) : 0, + _file, _line); + __vxh_set_vx_info(vxi, vxp, _here); + + atomic_inc(&vxi->vx_usecnt); + vxo = xchg(vxp, vxi); + BUG_ON(vxo); +} + + +#define clr_vx_info(p) __clr_vx_info(p, __FILE__, __LINE__, __HERE__) + +static inline void __clr_vx_info(struct vx_info **vxp, + const char *_file, int _line, void *_here) +{ + struct vx_info *vxo; + + vxo = xchg(vxp, NULL); + if (!vxo) + return; + + vxlprintk(VXD_CBIT(xid, 3), "clr_vx_info(%p[#%d.%d])", + vxo, vxo ? vxo->vx_id : 0, + vxo ? atomic_read(&vxo->vx_usecnt) : 0, + _file, _line); + __vxh_clr_vx_info(vxo, vxp, _here); + + if (atomic_dec_and_test(&vxo->vx_usecnt)) + free_vx_info(vxo); +} + + +#define claim_vx_info(v, p) \ + __claim_vx_info(v, p, __FILE__, __LINE__, __HERE__) + +static inline void __claim_vx_info(struct vx_info *vxi, + struct task_struct *task, + const char *_file, int _line, void *_here) +{ + vxlprintk(VXD_CBIT(xid, 3), "claim_vx_info(%p[#%d.%d.%d]) %p", + vxi, vxi ? vxi->vx_id : 0, + vxi ? atomic_read(&vxi->vx_usecnt) : 0, + vxi ? atomic_read(&vxi->vx_tasks) : 0, + task, _file, _line); + __vxh_claim_vx_info(vxi, task, _here); + + atomic_inc(&vxi->vx_tasks); +} + + +extern void unhash_vx_info(struct vx_info *); + +#define release_vx_info(v, p) \ + __release_vx_info(v, p, __FILE__, __LINE__, __HERE__) + +static inline void __release_vx_info(struct vx_info *vxi, + struct task_struct *task, + const char *_file, int _line, void *_here) +{ + vxlprintk(VXD_CBIT(xid, 3), "release_vx_info(%p[#%d.%d.%d]) %p", + vxi, vxi ? vxi->vx_id : 0, + vxi ? atomic_read(&vxi->vx_usecnt) : 0, + vxi ? atomic_read(&vxi->vx_tasks) : 0, + task, _file, _line); + __vxh_release_vx_info(vxi, task, _here); + + might_sleep(); + + if (atomic_dec_and_test(&vxi->vx_tasks)) + unhash_vx_info(vxi); +} + + +#define task_get_vx_info(p) \ + __task_get_vx_info(p, __FILE__, __LINE__, __HERE__) + +static inline struct vx_info *__task_get_vx_info(struct task_struct *p, + const char *_file, int _line, void *_here) +{ + struct vx_info *vxi; + + task_lock(p); + vxlprintk(VXD_CBIT(xid, 5), "task_get_vx_info(%p)", + p, _file, _line); + vxi = __get_vx_info(p->vx_info, _file, _line, _here); + task_unlock(p); + return vxi; +} + + +static inline void __wakeup_vx_info(struct vx_info *vxi) +{ + if (waitqueue_active(&vxi->vx_wait)) + wake_up_interruptible(&vxi->vx_wait); +} + + +#define enter_vx_info(v, s) __enter_vx_info(v, s, __FILE__, __LINE__) + +static inline void __enter_vx_info(struct vx_info *vxi, + struct vx_info_save *vxis, const char *_file, int _line) +{ + vxlprintk(VXD_CBIT(xid, 5), "enter_vx_info(%p[#%d],%p) %p[#%d,%p]", + vxi, vxi ? vxi->vx_id : 0, vxis, current, + current->xid, current->vx_info, _file, _line); + vxis->vxi = xchg(¤t->vx_info, vxi); + vxis->xid = current->xid; + current->xid = vxi ? vxi->vx_id : 0; +} + +#define leave_vx_info(s) __leave_vx_info(s, __FILE__, __LINE__) + +static inline void __leave_vx_info(struct vx_info_save *vxis, + const char *_file, int _line) +{ + vxlprintk(VXD_CBIT(xid, 5), "leave_vx_info(%p[#%d,%p]) %p[#%d,%p]", + vxis, vxis->xid, vxis->vxi, current, + current->xid, current->vx_info, _file, _line); + (void)xchg(¤t->vx_info, vxis->vxi); + current->xid = vxis->xid; +} + + +static inline void __enter_vx_admin(struct vx_info_save *vxis) +{ + vxis->vxi = xchg(¤t->vx_info, NULL); + vxis->xid = xchg(¤t->xid, (xid_t)0); +} + +static inline void __leave_vx_admin(struct vx_info_save *vxis) +{ + (void)xchg(¤t->xid, vxis->xid); + (void)xchg(¤t->vx_info, vxis->vxi); +} + +extern void exit_vx_info(struct task_struct *, int); +extern void exit_vx_info_early(struct task_struct *, int); + + +#else +#warning duplicate inclusion +#endif diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vs_cowbl.h linux-source-2.6.22/include/linux/vs_cowbl.h --- linux-source-2.6.22.orig/include/linux/vs_cowbl.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vs_cowbl.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,44 @@ +#ifndef _VS_COWBL_H +#define _VS_COWBL_H + +#include +#include +#include + +extern struct dentry *cow_break_link(const char *pathname); + +static inline int cow_check_and_break(struct nameidata *nd) +{ + struct inode *inode = nd->dentry->d_inode; + int error = 0; + if (IS_RDONLY(inode) || MNT_IS_RDONLY(nd->mnt)) + return -EROFS; + if (IS_COW(inode)) { + if (IS_COW_LINK(inode)) { + struct dentry *new_dentry, *old_dentry = nd->dentry; + char *path, *buf; + + buf = kmalloc(PATH_MAX, GFP_KERNEL); + if (!buf) { + return -ENOMEM; + } + path = d_path(nd->dentry, nd->mnt, buf, PATH_MAX); + new_dentry = cow_break_link(path); + kfree(buf); + if (!IS_ERR(new_dentry)) { + nd->dentry = new_dentry; + dput(old_dentry); + } else + error = PTR_ERR(new_dentry); + } else { + inode->i_flags &= ~(S_IUNLINK | S_IMMUTABLE); + inode->i_ctime = CURRENT_TIME; + mark_inode_dirty(inode); + } + } + return error; +} + +#else +#warning duplicate inclusion +#endif diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vs_cvirt.h linux-source-2.6.22/include/linux/vs_cvirt.h --- linux-source-2.6.22.orig/include/linux/vs_cvirt.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vs_cvirt.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,49 @@ +#ifndef _VS_CVIRT_H +#define _VS_CVIRT_H + +#include "vserver/cvirt.h" +#include "vserver/context.h" +#include "vserver/base.h" +#include "vserver/debug.h" + + +static inline void vx_activate_task(struct task_struct *p) +{ + struct vx_info *vxi; + + if ((vxi = p->vx_info)) { + vx_update_load(vxi); + atomic_inc(&vxi->cvirt.nr_running); + } +} + +static inline void vx_deactivate_task(struct task_struct *p) +{ + struct vx_info *vxi; + + if ((vxi = p->vx_info)) { + vx_update_load(vxi); + atomic_dec(&vxi->cvirt.nr_running); + } +} + +static inline void vx_uninterruptible_inc(struct task_struct *p) +{ + struct vx_info *vxi; + + if ((vxi = p->vx_info)) + atomic_inc(&vxi->cvirt.nr_uninterruptible); +} + +static inline void vx_uninterruptible_dec(struct task_struct *p) +{ + struct vx_info *vxi; + + if ((vxi = p->vx_info)) + atomic_dec(&vxi->cvirt.nr_uninterruptible); +} + + +#else +#warning duplicate inclusion +#endif diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vs_dlimit.h linux-source-2.6.22/include/linux/vs_dlimit.h --- linux-source-2.6.22.orig/include/linux/vs_dlimit.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vs_dlimit.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,209 @@ +#ifndef _VS_DLIMIT_H +#define _VS_DLIMIT_H + +#include "vserver/dlimit.h" +#include "vserver/base.h" +#include "vserver/debug.h" + + +#define get_dl_info(i) __get_dl_info(i, __FILE__, __LINE__) + +static inline struct dl_info *__get_dl_info(struct dl_info *dli, + const char *_file, int _line) +{ + if (!dli) + return NULL; + vxlprintk(VXD_CBIT(dlim, 4), "get_dl_info(%p[#%d.%d])", + dli, dli ? dli->dl_tag : 0, + dli ? atomic_read(&dli->dl_usecnt) : 0, + _file, _line); + atomic_inc(&dli->dl_usecnt); + return dli; +} + + +#define free_dl_info(i) \ + call_rcu(&(i)->dl_rcu, rcu_free_dl_info) + +#define put_dl_info(i) __put_dl_info(i, __FILE__, __LINE__) + +static inline void __put_dl_info(struct dl_info *dli, + const char *_file, int _line) +{ + if (!dli) + return; + vxlprintk(VXD_CBIT(dlim, 4), "put_dl_info(%p[#%d.%d])", + dli, dli ? dli->dl_tag : 0, + dli ? atomic_read(&dli->dl_usecnt) : 0, + _file, _line); + if (atomic_dec_and_test(&dli->dl_usecnt)) + free_dl_info(dli); +} + + +#define __dlimit_char(d) ((d) ? '*' : ' ') + +static inline int __dl_alloc_space(struct super_block *sb, + tag_t tag, dlsize_t nr, const char *file, int line) +{ + struct dl_info *dli = NULL; + int ret = 0; + + if (nr == 0) + goto out; + dli = locate_dl_info(sb, tag); + if (!dli) + goto out; + + spin_lock(&dli->dl_lock); + ret = (dli->dl_space_used + nr > dli->dl_space_total); + if (!ret) + dli->dl_space_used += nr; + spin_unlock(&dli->dl_lock); + put_dl_info(dli); +out: + vxlprintk(VXD_CBIT(dlim, 1), + "ALLOC (%p,#%d)%c %lld bytes (%d)", + sb, tag, __dlimit_char(dli), (long long)nr, + ret, file, line); + return ret; +} + +static inline void __dl_free_space(struct super_block *sb, + tag_t tag, dlsize_t nr, const char *_file, int _line) +{ + struct dl_info *dli = NULL; + + if (nr == 0) + goto out; + dli = locate_dl_info(sb, tag); + if (!dli) + goto out; + + spin_lock(&dli->dl_lock); + if (dli->dl_space_used > nr) + dli->dl_space_used -= nr; + else + dli->dl_space_used = 0; + spin_unlock(&dli->dl_lock); + put_dl_info(dli); +out: + vxlprintk(VXD_CBIT(dlim, 1), + "FREE (%p,#%d)%c %lld bytes", + sb, tag, __dlimit_char(dli), (long long)nr, + _file, _line); +} + +static inline int __dl_alloc_inode(struct super_block *sb, + tag_t tag, const char *_file, int _line) +{ + struct dl_info *dli; + int ret = 0; + + dli = locate_dl_info(sb, tag); + if (!dli) + goto out; + + spin_lock(&dli->dl_lock); + ret = (dli->dl_inodes_used >= dli->dl_inodes_total); + if (!ret) + dli->dl_inodes_used++; + spin_unlock(&dli->dl_lock); + put_dl_info(dli); +out: + vxlprintk(VXD_CBIT(dlim, 0), + "ALLOC (%p,#%d)%c inode (%d)", + sb, tag, __dlimit_char(dli), ret, _file, _line); + return ret; +} + +static inline void __dl_free_inode(struct super_block *sb, + tag_t tag, const char *_file, int _line) +{ + struct dl_info *dli; + + dli = locate_dl_info(sb, tag); + if (!dli) + goto out; + + spin_lock(&dli->dl_lock); + if (dli->dl_inodes_used > 1) + dli->dl_inodes_used--; + else + dli->dl_inodes_used = 0; + spin_unlock(&dli->dl_lock); + put_dl_info(dli); +out: + vxlprintk(VXD_CBIT(dlim, 0), + "FREE (%p,#%d)%c inode", + sb, tag, __dlimit_char(dli), _file, _line); +} + +static inline void __dl_adjust_block(struct super_block *sb, tag_t tag, + unsigned long long *free_blocks, unsigned long long *root_blocks, + const char *_file, int _line) +{ + struct dl_info *dli; + uint64_t broot, bfree; + + dli = locate_dl_info(sb, tag); + if (!dli) + return; + + spin_lock(&dli->dl_lock); + broot = (dli->dl_space_total - + (dli->dl_space_total >> 10) * dli->dl_nrlmult) + >> sb->s_blocksize_bits; + bfree = (dli->dl_space_total - dli->dl_space_used) + >> sb->s_blocksize_bits; + spin_unlock(&dli->dl_lock); + + vxlprintk(VXD_CBIT(dlim, 2), + "ADJUST: %lld,%lld on %lld,%lld [mult=%d]", + (long long)bfree, (long long)broot, + *free_blocks, *root_blocks, dli->dl_nrlmult, + _file, _line); + if (free_blocks) { + if (*free_blocks > bfree) + *free_blocks = bfree; + } + if (root_blocks) { + if (*root_blocks > broot) + *root_blocks = broot; + } + put_dl_info(dli); +} + +#define DLIMIT_ALLOC_SPACE(in, bytes) \ + __dl_alloc_space((in)->i_sb, (in)->i_tag, (dlsize_t)(bytes), \ + __FILE__, __LINE__ ) + +#define DLIMIT_FREE_SPACE(in, bytes) \ + __dl_free_space((in)->i_sb, (in)->i_tag, (dlsize_t)(bytes), \ + __FILE__, __LINE__ ) + +#define DLIMIT_ALLOC_BLOCK(in, nr) \ + __dl_alloc_space((in)->i_sb, (in)->i_tag, \ + ((dlsize_t)(nr)) << (in)->i_sb->s_blocksize_bits, \ + __FILE__, __LINE__ ) + +#define DLIMIT_FREE_BLOCK(in, nr) \ + __dl_free_space((in)->i_sb, (in)->i_tag, \ + ((dlsize_t)(nr)) << (in)->i_sb->s_blocksize_bits, \ + __FILE__, __LINE__ ) + + +#define DLIMIT_ALLOC_INODE(in) \ + __dl_alloc_inode((in)->i_sb, (in)->i_tag, __FILE__, __LINE__ ) + +#define DLIMIT_FREE_INODE(in) \ + __dl_free_inode((in)->i_sb, (in)->i_tag, __FILE__, __LINE__ ) + + +#define DLIMIT_ADJUST_BLOCK(sb, tag, fb, rb) \ + __dl_adjust_block(sb, tag, fb, rb, __FILE__, __LINE__ ) + + +#else +#warning duplicate inclusion +#endif diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/base.h linux-source-2.6.22/include/linux/vserver/base.h --- linux-source-2.6.22.orig/include/linux/vserver/base.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/base.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,236 @@ +#ifndef _VX_BASE_H +#define _VX_BASE_H + + +/* context state changes */ + +enum { + VSC_STARTUP = 1, + VSC_SHUTDOWN, + + VSC_NETUP, + VSC_NETDOWN, +}; + + +#define MAX_S_CONTEXT 65535 /* Arbitrary limit */ + +#ifdef CONFIG_VSERVER_DYNAMIC_IDS +#define MIN_D_CONTEXT 49152 /* dynamic contexts start here */ +#else +#define MIN_D_CONTEXT 65536 +#endif + +/* check conditions */ + +#define VS_ADMIN 0x0001 +#define VS_WATCH 0x0002 +#define VS_HIDE 0x0004 +#define VS_HOSTID 0x0008 + +#define VS_IDENT 0x0010 +#define VS_EQUIV 0x0020 +#define VS_PARENT 0x0040 +#define VS_CHILD 0x0080 + +#define VS_ARG_MASK 0x00F0 + +#define VS_DYNAMIC 0x0100 +#define VS_STATIC 0x0200 + +#define VS_ATR_MASK 0x0F00 + +#ifdef CONFIG_VSERVER_PRIVACY +#define VS_ADMIN_P (0) +#define VS_WATCH_P (0) +#else +#define VS_ADMIN_P VS_ADMIN +#define VS_WATCH_P VS_WATCH +#endif + +#define VS_HARDIRQ 0x1000 +#define VS_SOFTIRQ 0x2000 +#define VS_IRQ 0x4000 + +#define VS_IRQ_MASK 0xF000 + +#include + +/* + * check current context for ADMIN/WATCH and + * optionally against supplied argument + */ +static inline int __vs_check(int cid, int id, unsigned int mode) +{ + if (mode & VS_ARG_MASK) { + if ((mode & VS_IDENT) && (id == cid)) + return 1; + } + if (mode & VS_ATR_MASK) { + if ((mode & VS_DYNAMIC) && + (id >= MIN_D_CONTEXT) && + (id <= MAX_S_CONTEXT)) + return 1; + if ((mode & VS_STATIC) && + (id > 1) && (id < MIN_D_CONTEXT)) + return 1; + } + if (mode & VS_IRQ_MASK) { + if ((mode & VS_IRQ) && unlikely(in_interrupt())) + return 1; + if ((mode & VS_HARDIRQ) && unlikely(in_irq())) + return 1; + if ((mode & VS_SOFTIRQ) && unlikely(in_softirq())) + return 1; + } + return (((mode & VS_ADMIN) && (cid == 0)) || + ((mode & VS_WATCH) && (cid == 1)) || + ((mode & VS_HOSTID) && (id == 0))); +} + +#define vx_task_xid(t) ((t)->xid) + +#define vx_current_xid() vx_task_xid(current) + +#define current_vx_info() (current->vx_info) + + +#define vx_check(c, m) __vs_check(vx_current_xid(), c, (m) | VS_IRQ) + +#define vx_weak_check(c, m) ((m) ? vx_check(c, m) : 1) + + +#define nx_task_nid(t) ((t)->nid) + +#define nx_current_nid() nx_task_nid(current) + +#define current_nx_info() (current->nx_info) + + +#define nx_check(c, m) __vs_check(nx_current_nid(), c, m) + +#define nx_weak_check(c, m) ((m) ? nx_check(c, m) : 1) + + + +/* generic flag merging */ + +#define vs_check_flags(v, m, f) (((v) & (m)) ^ (f)) + +#define vs_mask_flags(v, f, m) (((v) & ~(m)) | ((f) & (m))) + +#define vs_mask_mask(v, f, m) (((v) & ~(m)) | ((v) & (f) & (m))) + +#define vs_check_bit(v, n) ((v) & (1LL << (n))) + + +/* context flags */ + +#define __vx_flags(v) ((v) ? (v)->vx_flags : 0) + +#define vx_current_flags() __vx_flags(current->vx_info) + +#define vx_info_flags(v, m, f) \ + vs_check_flags(__vx_flags(v), m, f) + +#define task_vx_flags(t, m, f) \ + ((t) && vx_info_flags((t)->vx_info, m, f)) + +#define vx_flags(m, f) vx_info_flags(current->vx_info, m, f) + + +/* context caps */ + +#define __vx_ccaps(v) ((v) ? (v)->vx_ccaps : 0) + +#define vx_current_ccaps() __vx_ccaps(current->vx_info) + +#define vx_info_ccaps(v, c) (__vx_ccaps(v) & (c)) + +#define vx_ccaps(c) vx_info_ccaps(current->vx_info, (c)) + + + +/* network flags */ + +#define __nx_flags(v) ((v) ? (v)->nx_flags : 0) + +#define nx_current_flags() __nx_flags(current->nx_info) + +#define nx_info_flags(v, m, f) \ + vs_check_flags(__nx_flags(v), m, f) + +#define task_nx_flags(t, m, f) \ + ((t) && nx_info_flags((t)->nx_info, m, f)) + +#define nx_flags(m, f) nx_info_flags(current->nx_info, m, f) + + +/* network caps */ + +#define __nx_ncaps(v) ((v) ? (v)->nx_ncaps : 0) + +#define nx_current_ncaps() __nx_ncaps(current->nx_info) + +#define nx_info_ncaps(v, c) (__nx_ncaps(v) & (c)) + +#define nx_ncaps(c) nx_info_ncaps(current->nx_info, c) + + +/* context mask capabilities */ + +#define __vx_mcaps(v) ((v) ? (v)->vx_ccaps >> 32UL : ~0 ) + +#define vx_info_mcaps(v, c) (__vx_mcaps(v) & (c)) + +#define vx_mcaps(c) vx_info_mcaps(current->vx_info, c) + + +/* context bcap mask */ + +#define __vx_bcaps(v) ((v) ? (v)->vx_bcaps : ~0 ) + +#define vx_current_bcaps() __vx_bcaps(current->vx_info) + +#define vx_info_bcaps(v, c) (__vx_bcaps(v) & (c)) + +#define vx_bcaps(c) vx_info_bcaps(current->vx_info, c) + + +#define vx_info_cap_bset(v) ((v) ? (v)->vx_cap_bset : cap_bset) + +#define vx_current_cap_bset() vx_info_cap_bset(current->vx_info) + + +#define __vx_info_mbcap(v, b) \ + (!vx_info_flags(v, VXF_STATE_SETUP, 0) ? \ + vx_info_bcaps(v, b) : (b)) + +#define vx_info_mbcap(v, b) __vx_info_mbcap(v, cap_t(b)) + +#define task_vx_mbcap(t, b) \ + vx_info_mbcap((t)->vx_info, (t)->b) + +#define vx_mbcap(b) task_vx_mbcap(current, b) + +#define vx_cap_raised(v, c, f) (vx_info_mbcap(v, c) & CAP_TO_MASK(f)) + +#define vx_capable(b, c) (capable(b) || \ + (cap_raised(current->cap_effective, b) && vx_ccaps(c))) + + +#define vx_current_initpid(n) \ + (current->vx_info && \ + (current->vx_info->vx_initpid == (n))) + + +#define __vx_state(v) ((v) ? ((v)->vx_state) : 0) + +#define vx_info_state(v, m) (__vx_state(v) & (m)) + + +#define __nx_state(v) ((v) ? ((v)->nx_state) : 0) + +#define nx_info_state(v, m) (__nx_state(v) & (m)) + +#endif diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/cacct_cmd.h linux-source-2.6.22/include/linux/vserver/cacct_cmd.h --- linux-source-2.6.22.orig/include/linux/vserver/cacct_cmd.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/cacct_cmd.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,23 @@ +#ifndef _VX_CACCT_CMD_H +#define _VX_CACCT_CMD_H + + +/* virtual host info name commands */ + +#define VCMD_sock_stat VC_CMD(VSTAT, 5, 0) + +struct vcmd_sock_stat_v0 { + uint32_t field; + uint32_t count[3]; + uint64_t total[3]; +}; + + +#ifdef __KERNEL__ + +#include + +extern int vc_sock_stat(struct vx_info *, void __user *); + +#endif /* __KERNEL__ */ +#endif /* _VX_CACCT_CMD_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/cacct_def.h linux-source-2.6.22/include/linux/vserver/cacct_def.h --- linux-source-2.6.22.orig/include/linux/vserver/cacct_def.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/cacct_def.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,43 @@ +#ifndef _VX_CACCT_DEF_H +#define _VX_CACCT_DEF_H + +#include +#include + + +struct _vx_sock_acc { + atomic_long_t count; + atomic_long_t total; +}; + +/* context sub struct */ + +struct _vx_cacct { + struct _vx_sock_acc sock[VXA_SOCK_SIZE][3]; + atomic_t slab[8]; + atomic_t page[6][8]; +}; + +#ifdef CONFIG_VSERVER_DEBUG + +static inline void __dump_vx_cacct(struct _vx_cacct *cacct) +{ + int i, j; + + printk("\t_vx_cacct:"); + for (i = 0; i < 6; i++) { + struct _vx_sock_acc *ptr = cacct->sock[i]; + + printk("\t [%d] =", i); + for (j = 0; j < 3; j++) { + printk(" [%d] = %8lu, %8lu", j, + atomic_long_read(&ptr[j].count), + atomic_long_read(&ptr[j].total)); + } + printk("\n"); + } +} + +#endif + +#endif /* _VX_CACCT_DEF_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/cacct.h linux-source-2.6.22/include/linux/vserver/cacct.h --- linux-source-2.6.22.orig/include/linux/vserver/cacct.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/cacct.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,15 @@ +#ifndef _VX_CACCT_H +#define _VX_CACCT_H + + +enum sock_acc_field { + VXA_SOCK_UNSPEC = 0, + VXA_SOCK_UNIX, + VXA_SOCK_INET, + VXA_SOCK_INET6, + VXA_SOCK_PACKET, + VXA_SOCK_OTHER, + VXA_SOCK_SIZE /* array size */ +}; + +#endif /* _VX_CACCT_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/cacct_int.h linux-source-2.6.22/include/linux/vserver/cacct_int.h --- linux-source-2.6.22.orig/include/linux/vserver/cacct_int.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/cacct_int.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,21 @@ +#ifndef _VX_CACCT_INT_H +#define _VX_CACCT_INT_H + + +#ifdef __KERNEL__ + +static inline +unsigned long vx_sock_count(struct _vx_cacct *cacct, int type, int pos) +{ + return atomic_long_read(&cacct->sock[type][pos].count); +} + + +static inline +unsigned long vx_sock_total(struct _vx_cacct *cacct, int type, int pos) +{ + return atomic_long_read(&cacct->sock[type][pos].total); +} + +#endif /* __KERNEL__ */ +#endif /* _VX_CACCT_INT_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/context_cmd.h linux-source-2.6.22/include/linux/vserver/context_cmd.h --- linux-source-2.6.22.orig/include/linux/vserver/context_cmd.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/context_cmd.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,123 @@ +#ifndef _VX_CONTEXT_CMD_H +#define _VX_CONTEXT_CMD_H + + +/* vinfo commands */ + +#define VCMD_task_xid VC_CMD(VINFO, 1, 0) + +#ifdef __KERNEL__ +extern int vc_task_xid(uint32_t, void __user *); + +#endif /* __KERNEL__ */ + +#define VCMD_vx_info VC_CMD(VINFO, 5, 0) + +struct vcmd_vx_info_v0 { + uint32_t xid; + uint32_t initpid; + /* more to come */ +}; + +#ifdef __KERNEL__ +extern int vc_vx_info(struct vx_info *, void __user *); + +#endif /* __KERNEL__ */ + +#define VCMD_ctx_stat VC_CMD(VSTAT, 0, 0) + +struct vcmd_ctx_stat_v0 { + uint32_t usecnt; + uint32_t tasks; + /* more to come */ +}; + +#ifdef __KERNEL__ +extern int vc_ctx_stat(struct vx_info *, void __user *); + +#endif /* __KERNEL__ */ + +/* context commands */ + +#define VCMD_ctx_create_v0 VC_CMD(VPROC, 1, 0) +#define VCMD_ctx_create VC_CMD(VPROC, 1, 1) + +struct vcmd_ctx_create { + uint64_t flagword; +}; + +#define VCMD_ctx_migrate_v0 VC_CMD(PROCMIG, 1, 0) +#define VCMD_ctx_migrate VC_CMD(PROCMIG, 1, 1) + +struct vcmd_ctx_migrate { + uint64_t flagword; +}; + +#ifdef __KERNEL__ +extern int vc_ctx_create(uint32_t, void __user *); +extern int vc_ctx_migrate(struct vx_info *, void __user *); + +#endif /* __KERNEL__ */ + + +/* flag commands */ + +#define VCMD_get_cflags VC_CMD(FLAGS, 1, 0) +#define VCMD_set_cflags VC_CMD(FLAGS, 2, 0) + +struct vcmd_ctx_flags_v0 { + uint64_t flagword; + uint64_t mask; +}; + +#ifdef __KERNEL__ +extern int vc_get_cflags(struct vx_info *, void __user *); +extern int vc_set_cflags(struct vx_info *, void __user *); + +#endif /* __KERNEL__ */ + + +/* context caps commands */ + +#define VCMD_get_ccaps_v0 VC_CMD(FLAGS, 3, 0) +#define VCMD_set_ccaps_v0 VC_CMD(FLAGS, 4, 0) + +struct vcmd_ctx_caps_v0 { + uint64_t bcaps; + uint64_t ccaps; + uint64_t cmask; +}; + +#define VCMD_get_ccaps VC_CMD(FLAGS, 3, 1) +#define VCMD_set_ccaps VC_CMD(FLAGS, 4, 1) + +struct vcmd_ctx_caps_v1 { + uint64_t ccaps; + uint64_t cmask; +}; + +#ifdef __KERNEL__ +extern int vc_get_ccaps_v0(struct vx_info *, void __user *); +extern int vc_set_ccaps_v0(struct vx_info *, void __user *); +extern int vc_get_ccaps(struct vx_info *, void __user *); +extern int vc_set_ccaps(struct vx_info *, void __user *); + +#endif /* __KERNEL__ */ + + +/* bcaps commands */ + +#define VCMD_get_bcaps VC_CMD(FLAGS, 9, 0) +#define VCMD_set_bcaps VC_CMD(FLAGS, 10, 0) + +struct vcmd_bcaps { + uint64_t bcaps; + uint64_t bmask; +}; + +#ifdef __KERNEL__ +extern int vc_get_bcaps(struct vx_info *, void __user *); +extern int vc_set_bcaps(struct vx_info *, void __user *); + +#endif /* __KERNEL__ */ +#endif /* _VX_CONTEXT_CMD_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/context.h linux-source-2.6.22/include/linux/vserver/context.h --- linux-source-2.6.22.orig/include/linux/vserver/context.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/context.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,172 @@ +#ifndef _VX_CONTEXT_H +#define _VX_CONTEXT_H + +#include +#include + + +#define VX_DYNAMIC_ID ((uint32_t)-1) /* id for dynamic context */ + +/* context flags */ + +#define VXF_INFO_SCHED 0x00000002 +#define VXF_INFO_NPROC 0x00000004 +#define VXF_INFO_PRIVATE 0x00000008 + +#define VXF_INFO_INIT 0x00000010 +#define VXF_INFO_HIDE 0x00000020 +#define VXF_INFO_ULIMIT 0x00000040 +#define VXF_INFO_NSPACE 0x00000080 + +#define VXF_SCHED_HARD 0x00000100 +#define VXF_SCHED_PRIO 0x00000200 +#define VXF_SCHED_PAUSE 0x00000400 + +#define VXF_VIRT_MEM 0x00010000 +#define VXF_VIRT_UPTIME 0x00020000 +#define VXF_VIRT_CPU 0x00040000 +#define VXF_VIRT_LOAD 0x00080000 +#define VXF_VIRT_TIME 0x00100000 + +#define VXF_HIDE_MOUNT 0x01000000 +#define VXF_HIDE_NETIF 0x02000000 +#define VXF_HIDE_VINFO 0x04000000 + +#define VXF_STATE_SETUP (1ULL << 32) +#define VXF_STATE_INIT (1ULL << 33) +#define VXF_STATE_ADMIN (1ULL << 34) + +#define VXF_SC_HELPER (1ULL << 36) +#define VXF_REBOOT_KILL (1ULL << 37) +#define VXF_PERSISTENT (1ULL << 38) + +#define VXF_FORK_RSS (1ULL << 48) +#define VXF_PROLIFIC (1ULL << 49) + +#define VXF_IGNEG_NICE (1ULL << 52) + +#define VXF_ONE_TIME (0x0007ULL << 32) + +#define VXF_INIT_SET (VXF_STATE_SETUP | VXF_STATE_INIT | VXF_STATE_ADMIN) + + +/* context migration */ + +#define VXM_SET_INIT 0x00000001 +#define VXM_SET_REAPER 0x00000002 + +/* context caps */ + +#define VXC_CAP_MASK 0x00000000 + +#define VXC_SET_UTSNAME 0x00000001 +#define VXC_SET_RLIMIT 0x00000002 + +#define VXC_RAW_ICMP 0x00000100 +#define VXC_SYSLOG 0x00001000 + +#define VXC_SECURE_MOUNT 0x00010000 +#define VXC_SECURE_REMOUNT 0x00020000 +#define VXC_BINARY_MOUNT 0x00040000 + +#define VXC_QUOTA_CTL 0x00100000 +#define VXC_ADMIN_MAPPER 0x00200000 +#define VXC_ADMIN_CLOOP 0x00400000 + + +#ifdef __KERNEL__ + +#include +#include +#include + +#include "limit_def.h" +#include "sched_def.h" +#include "cvirt_def.h" +#include "cacct_def.h" + +struct _vx_info_pc { + struct _vx_sched_pc sched_pc; + struct _vx_cvirt_pc cvirt_pc; +}; + +struct vx_info { + struct hlist_node vx_hlist; /* linked list of contexts */ + xid_t vx_id; /* context id */ + atomic_t vx_usecnt; /* usage count */ + atomic_t vx_tasks; /* tasks count */ + struct vx_info *vx_parent; /* parent context */ + int vx_state; /* context state */ + + unsigned long vx_nsmask; /* assignment mask */ + struct nsproxy *vx_nsproxy; /* private namespace */ + struct fs_struct *vx_fs; /* private namespace fs */ + + uint64_t vx_flags; /* context flags */ + uint64_t vx_bcaps; /* bounding caps (system) */ + uint64_t vx_ccaps; /* context caps (vserver) */ + kernel_cap_t vx_cap_bset; /* the guest's bset */ + + struct task_struct *vx_reaper; /* guest reaper process */ + pid_t vx_initpid; /* PID of guest init */ + + struct _vx_limit limit; /* vserver limits */ + struct _vx_sched sched; /* vserver scheduler */ + struct _vx_cvirt cvirt; /* virtual/bias stuff */ + struct _vx_cacct cacct; /* context accounting */ + +#ifndef CONFIG_SMP + struct _vx_info_pc info_pc; /* per cpu data */ +#else + struct _vx_info_pc *ptr_pc; /* per cpu array */ +#endif + + wait_queue_head_t vx_wait; /* context exit waitqueue */ + int reboot_cmd; /* last sys_reboot() cmd */ + int exit_code; /* last process exit code */ + + char vx_name[65]; /* vserver name */ +}; + +#ifndef CONFIG_SMP +#define vx_ptr_pc(vxi) (&(vxi)->info_pc) +#define vx_per_cpu(vxi, v, id) vx_ptr_pc(vxi)->v +#else +#define vx_ptr_pc(vxi) ((vxi)->ptr_pc) +#define vx_per_cpu(vxi, v, id) per_cpu_ptr(vx_ptr_pc(vxi), id)->v +#endif + +#define vx_cpu(vxi, v) vx_per_cpu(vxi, v, smp_processor_id()) + + +struct vx_info_save { + struct vx_info *vxi; + xid_t xid; +}; + + +/* status flags */ + +#define VXS_HASHED 0x0001 +#define VXS_PAUSED 0x0010 +#define VXS_SHUTDOWN 0x0100 +#define VXS_HELPER 0x1000 +#define VXS_RELEASED 0x8000 + + +extern void claim_vx_info(struct vx_info *, struct task_struct *); +extern void release_vx_info(struct vx_info *, struct task_struct *); + +extern struct vx_info *lookup_vx_info(int); +extern struct vx_info *lookup_or_create_vx_info(int); + +extern int get_xid_list(int, unsigned int *, int); +extern int xid_is_hashed(xid_t); + +extern int vx_migrate_task(struct task_struct *, struct vx_info *, int); + +extern long vs_state_change(struct vx_info *, unsigned int); + + +#endif /* __KERNEL__ */ +#endif /* _VX_CONTEXT_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/cvirt_cmd.h linux-source-2.6.22/include/linux/vserver/cvirt_cmd.h --- linux-source-2.6.22.orig/include/linux/vserver/cvirt_cmd.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/cvirt_cmd.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,53 @@ +#ifndef _VX_CVIRT_CMD_H +#define _VX_CVIRT_CMD_H + + +/* virtual host info name commands */ + +#define VCMD_set_vhi_name VC_CMD(VHOST, 1, 0) +#define VCMD_get_vhi_name VC_CMD(VHOST, 2, 0) + +struct vcmd_vhi_name_v0 { + uint32_t field; + char name[65]; +}; + + +enum vhi_name_field { + VHIN_CONTEXT = 0, + VHIN_SYSNAME, + VHIN_NODENAME, + VHIN_RELEASE, + VHIN_VERSION, + VHIN_MACHINE, + VHIN_DOMAINNAME, +}; + + +#ifdef __KERNEL__ + +#include + +extern int vc_set_vhi_name(struct vx_info *, void __user *); +extern int vc_get_vhi_name(struct vx_info *, void __user *); + +#endif /* __KERNEL__ */ + +#define VCMD_virt_stat VC_CMD(VSTAT, 3, 0) + +struct vcmd_virt_stat_v0 { + uint64_t offset; + uint64_t uptime; + uint32_t nr_threads; + uint32_t nr_running; + uint32_t nr_uninterruptible; + uint32_t nr_onhold; + uint32_t nr_forks; + uint32_t load[3]; +}; + +#ifdef __KERNEL__ +extern int vc_virt_stat(struct vx_info *, void __user *); + +#endif /* __KERNEL__ */ +#endif /* _VX_CVIRT_CMD_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/cvirt_def.h linux-source-2.6.22/include/linux/vserver/cvirt_def.h --- linux-source-2.6.22.orig/include/linux/vserver/cvirt_def.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/cvirt_def.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,81 @@ +#ifndef _VX_CVIRT_DEF_H +#define _VX_CVIRT_DEF_H + +#include +#include +#include +#include +#include +#include + + +struct _vx_usage_stat { + uint64_t user; + uint64_t nice; + uint64_t system; + uint64_t softirq; + uint64_t irq; + uint64_t idle; + uint64_t iowait; +}; + +struct _vx_syslog { + wait_queue_head_t log_wait; + spinlock_t logbuf_lock; /* lock for the log buffer */ + + unsigned long log_start; /* next char to be read by syslog() */ + unsigned long con_start; /* next char to be sent to consoles */ + unsigned long log_end; /* most-recently-written-char + 1 */ + unsigned long logged_chars; /* #chars since last read+clear operation */ + + char log_buf[1024]; +}; + + +/* context sub struct */ + +struct _vx_cvirt { + atomic_t nr_threads; /* number of current threads */ + atomic_t nr_running; /* number of running threads */ + atomic_t nr_uninterruptible; /* number of uninterruptible threads */ + + atomic_t nr_onhold; /* processes on hold */ + uint32_t onhold_last; /* jiffies when put on hold */ + + struct timeval bias_tv; /* time offset to the host */ + struct timespec bias_idle; + struct timespec bias_uptime; /* context creation point */ + uint64_t bias_clock; /* offset in clock_t */ + + spinlock_t load_lock; /* lock for the load averages */ + atomic_t load_updates; /* nr of load updates done so far */ + uint32_t load_last; /* last time load was calculated */ + uint32_t load[3]; /* load averages 1,5,15 */ + + atomic_t total_forks; /* number of forks so far */ + + struct _vx_syslog syslog; +}; + +struct _vx_cvirt_pc { + struct _vx_usage_stat cpustat; +}; + + +#ifdef CONFIG_VSERVER_DEBUG + +static inline void __dump_vx_cvirt(struct _vx_cvirt *cvirt) +{ + printk("\t_vx_cvirt:\n"); + printk("\t threads: %4d, %4d, %4d, %4d\n", + atomic_read(&cvirt->nr_threads), + atomic_read(&cvirt->nr_running), + atomic_read(&cvirt->nr_uninterruptible), + atomic_read(&cvirt->nr_onhold)); + /* add rest here */ + printk("\t total_forks = %d\n", atomic_read(&cvirt->total_forks)); +} + +#endif + +#endif /* _VX_CVIRT_DEF_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/cvirt.h linux-source-2.6.22/include/linux/vserver/cvirt.h --- linux-source-2.6.22.orig/include/linux/vserver/cvirt.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/cvirt.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,20 @@ +#ifndef _VX_CVIRT_H +#define _VX_CVIRT_H + + +#ifdef __KERNEL__ + +struct timespec; + +void vx_vsi_uptime(struct timespec *, struct timespec *); + + +struct vx_info; + +void vx_update_load(struct vx_info *); + + +int vx_do_syslog(int, char __user *, int); + +#endif /* __KERNEL__ */ +#endif /* _VX_CVIRT_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/debug_cmd.h linux-source-2.6.22/include/linux/vserver/debug_cmd.h --- linux-source-2.6.22.orig/include/linux/vserver/debug_cmd.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/debug_cmd.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,58 @@ +#ifndef _VX_DEBUG_CMD_H +#define _VX_DEBUG_CMD_H + + +/* debug commands */ + +#define VCMD_dump_history VC_CMD(DEBUG, 1, 0) + +#define VCMD_read_history VC_CMD(DEBUG, 5, 0) +#define VCMD_read_monitor VC_CMD(DEBUG, 6, 0) + +struct vcmd_read_history_v0 { + uint32_t index; + uint32_t count; + char __user *data; +}; + +struct vcmd_read_monitor_v0 { + uint32_t index; + uint32_t count; + char __user *data; +}; + + +#ifdef __KERNEL__ + +#ifdef CONFIG_COMPAT + +#include + +struct vcmd_read_history_v0_x32 { + uint32_t index; + uint32_t count; + compat_uptr_t data_ptr; +}; + +struct vcmd_read_monitor_v0_x32 { + uint32_t index; + uint32_t count; + compat_uptr_t data_ptr; +}; + +#endif /* CONFIG_COMPAT */ + +extern int vc_dump_history(uint32_t); + +extern int vc_read_history(uint32_t, void __user *); +extern int vc_read_monitor(uint32_t, void __user *); + +#ifdef CONFIG_COMPAT + +extern int vc_read_history_x32(uint32_t, void __user *); +extern int vc_read_monitor_x32(uint32_t, void __user *); + +#endif /* CONFIG_COMPAT */ + +#endif /* __KERNEL__ */ +#endif /* _VX_DEBUG_CMD_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/debug.h linux-source-2.6.22/include/linux/vserver/debug.h --- linux-source-2.6.22.orig/include/linux/vserver/debug.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/debug.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,113 @@ +#ifndef _VX_DEBUG_H +#define _VX_DEBUG_H + + +#define VXD_CBIT(n, m) (vx_debug_ ## n & (1 << (m))) +#define VXD_CMIN(n, m) (vx_debug_ ## n > (m)) +#define VXD_MASK(n, m) (vx_debug_ ## n & (m)) + +#define VXD_DEV(d) (d), (d)->bd_inode->i_ino, \ + imajor((d)->bd_inode), iminor((d)->bd_inode) +#define VXF_DEV "%p[%lu,%d:%d]" + + +#define __FUNC__ __func__ + + +#define vxd_path(d, m) \ + ({ static char _buffer[PATH_MAX]; \ + d_path(d, m, _buffer, sizeof(_buffer)); }) + +#define vxd_cond_path(n) \ + ((n) ? vxd_path((n)->dentry, (n)->mnt) : "" ) + + +#ifdef CONFIG_VSERVER_DEBUG + +extern unsigned int vx_debug_switch; +extern unsigned int vx_debug_xid; +extern unsigned int vx_debug_nid; +extern unsigned int vx_debug_tag; +extern unsigned int vx_debug_net; +extern unsigned int vx_debug_limit; +extern unsigned int vx_debug_cres; +extern unsigned int vx_debug_dlim; +extern unsigned int vx_debug_quota; +extern unsigned int vx_debug_cvirt; +extern unsigned int vx_debug_space; +extern unsigned int vx_debug_misc; + + +#define VX_LOGLEVEL "vxD: " + +#define vxdprintk(c, f, x...) \ + do { \ + if (c) \ + printk(VX_LOGLEVEL f "\n" , ##x); \ + } while (0) + +#define vxlprintk(c, f, x...) \ + do { \ + if (c) \ + printk(VX_LOGLEVEL f " @%s:%d\n", x); \ + } while (0) + +#define vxfprintk(c, f, x...) \ + do { \ + if (c) \ + printk(VX_LOGLEVEL f " %s@%s:%d\n", x); \ + } while (0) + + +struct vx_info; + +void dump_vx_info(struct vx_info *, int); +void dump_vx_info_inactive(int); + +#else /* CONFIG_VSERVER_DEBUG */ + +#define vx_debug_switch 0 +#define vx_debug_xid 0 +#define vx_debug_nid 0 +#define vx_debug_tag 0 +#define vx_debug_net 0 +#define vx_debug_limit 0 +#define vx_debug_cres 0 +#define vx_debug_dlim 0 +#define vx_debug_cvirt 0 + +#define vxdprintk(x...) do { } while (0) +#define vxlprintk(x...) do { } while (0) +#define vxfprintk(x...) do { } while (0) + +#endif /* CONFIG_VSERVER_DEBUG */ + + +#ifdef CONFIG_VSERVER_WARN + +#define VX_WARNLEVEL KERN_WARNING "vxW: " + +#define vxwprintk(c, f, x...) \ + do { \ + if (c) \ + printk(VX_WARNLEVEL f "\n" , ##x); \ + } while (0) + +#else /* CONFIG_VSERVER_WARN */ + +#define vxwprintk(x...) do { } while (0) + +#endif /* CONFIG_VSERVER_WARN */ + + +#ifdef CONFIG_VSERVER_DEBUG +#define vxd_assert_lock(l) assert_spin_locked(l) +#define vxd_assert(c, f, x...) vxlprintk(!(c), \ + "assertion [" f "] failed.", ##x, __FILE__, __LINE__) +#else +#define vxd_assert_lock(l) do { } while (0) +#define vxd_assert(c, f, x...) do { } while (0) +#endif + + +#endif /* _VX_DEBUG_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/dlimit_cmd.h linux-source-2.6.22/include/linux/vserver/dlimit_cmd.h --- linux-source-2.6.22.orig/include/linux/vserver/dlimit_cmd.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/dlimit_cmd.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,74 @@ +#ifndef _VX_DLIMIT_CMD_H +#define _VX_DLIMIT_CMD_H + + +/* dlimit vserver commands */ + +#define VCMD_add_dlimit VC_CMD(DLIMIT, 1, 0) +#define VCMD_rem_dlimit VC_CMD(DLIMIT, 2, 0) + +#define VCMD_set_dlimit VC_CMD(DLIMIT, 5, 0) +#define VCMD_get_dlimit VC_CMD(DLIMIT, 6, 0) + +struct vcmd_ctx_dlimit_base_v0 { + const char __user *name; + uint32_t flags; +}; + +struct vcmd_ctx_dlimit_v0 { + const char __user *name; + uint32_t space_used; /* used space in kbytes */ + uint32_t space_total; /* maximum space in kbytes */ + uint32_t inodes_used; /* used inodes */ + uint32_t inodes_total; /* maximum inodes */ + uint32_t reserved; /* reserved for root in % */ + uint32_t flags; +}; + +#define CDLIM_UNSET ((uint32_t)0UL) +#define CDLIM_INFINITY ((uint32_t)~0UL) +#define CDLIM_KEEP ((uint32_t)~1UL) + +#ifdef __KERNEL__ + +#ifdef CONFIG_COMPAT + +#include + +struct vcmd_ctx_dlimit_base_v0_x32 { + compat_uptr_t name_ptr; + uint32_t flags; +}; + +struct vcmd_ctx_dlimit_v0_x32 { + compat_uptr_t name_ptr; + uint32_t space_used; /* used space in kbytes */ + uint32_t space_total; /* maximum space in kbytes */ + uint32_t inodes_used; /* used inodes */ + uint32_t inodes_total; /* maximum inodes */ + uint32_t reserved; /* reserved for root in % */ + uint32_t flags; +}; + +#endif /* CONFIG_COMPAT */ + +#include + +extern int vc_add_dlimit(uint32_t, void __user *); +extern int vc_rem_dlimit(uint32_t, void __user *); + +extern int vc_set_dlimit(uint32_t, void __user *); +extern int vc_get_dlimit(uint32_t, void __user *); + +#ifdef CONFIG_COMPAT + +extern int vc_add_dlimit_x32(uint32_t, void __user *); +extern int vc_rem_dlimit_x32(uint32_t, void __user *); + +extern int vc_set_dlimit_x32(uint32_t, void __user *); +extern int vc_get_dlimit_x32(uint32_t, void __user *); + +#endif /* CONFIG_COMPAT */ + +#endif /* __KERNEL__ */ +#endif /* _VX_DLIMIT_CMD_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/dlimit.h linux-source-2.6.22/include/linux/vserver/dlimit.h --- linux-source-2.6.22.orig/include/linux/vserver/dlimit.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/dlimit.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,53 @@ +#ifndef _VX_DLIMIT_H +#define _VX_DLIMIT_H + +#include "switch.h" + + +#ifdef __KERNEL__ + +/* keep in sync with CDLIM_INFINITY */ + +#define DLIM_INFINITY (~0ULL) + +#include + +struct super_block; + +struct dl_info { + struct hlist_node dl_hlist; /* linked list of contexts */ + struct rcu_head dl_rcu; /* the rcu head */ + tag_t dl_tag; /* context tag */ + atomic_t dl_usecnt; /* usage count */ + atomic_t dl_refcnt; /* reference count */ + + struct super_block *dl_sb; /* associated superblock */ + + spinlock_t dl_lock; /* protect the values */ + + unsigned long long dl_space_used; /* used space in bytes */ + unsigned long long dl_space_total; /* maximum space in bytes */ + unsigned long dl_inodes_used; /* used inodes */ + unsigned long dl_inodes_total; /* maximum inodes */ + + unsigned int dl_nrlmult; /* non root limit mult */ +}; + +struct rcu_head; + +extern void rcu_free_dl_info(struct rcu_head *); +extern void unhash_dl_info(struct dl_info *); + +extern struct dl_info *locate_dl_info(struct super_block *, tag_t); + + +struct kstatfs; + +extern void vx_vsi_statfs(struct super_block *, struct kstatfs *); + +typedef uint64_t dlsize_t; + +#endif /* __KERNEL__ */ +#else /* _VX_DLIMIT_H */ +#warning duplicate inclusion +#endif /* _VX_DLIMIT_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/global.h linux-source-2.6.22/include/linux/vserver/global.h --- linux-source-2.6.22.orig/include/linux/vserver/global.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/global.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,18 @@ +#ifndef _VX_GLOBAL_H +#define _VX_GLOBAL_H + + +extern atomic_t vx_global_ctotal; +extern atomic_t vx_global_cactive; + +extern atomic_t nx_global_ctotal; +extern atomic_t nx_global_cactive; + +extern atomic_t vs_global_nsproxy; +extern atomic_t vs_global_fs; +extern atomic_t vs_global_mnt_ns; +extern atomic_t vs_global_uts_ns; +extern atomic_t vs_global_ipc_ns; + + +#endif /* _VX_GLOBAL_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/history.h linux-source-2.6.22/include/linux/vserver/history.h --- linux-source-2.6.22.orig/include/linux/vserver/history.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/history.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,197 @@ +#ifndef _VX_HISTORY_H +#define _VX_HISTORY_H + + +enum { + VXH_UNUSED = 0, + VXH_THROW_OOPS = 1, + + VXH_GET_VX_INFO, + VXH_PUT_VX_INFO, + VXH_INIT_VX_INFO, + VXH_SET_VX_INFO, + VXH_CLR_VX_INFO, + VXH_CLAIM_VX_INFO, + VXH_RELEASE_VX_INFO, + VXH_ALLOC_VX_INFO, + VXH_DEALLOC_VX_INFO, + VXH_HASH_VX_INFO, + VXH_UNHASH_VX_INFO, + VXH_LOC_VX_INFO, + VXH_LOOKUP_VX_INFO, + VXH_CREATE_VX_INFO, +}; + +struct _vxhe_vxi { + struct vx_info *ptr; + unsigned xid; + unsigned usecnt; + unsigned tasks; +}; + +struct _vxhe_set_clr { + void *data; +}; + +struct _vxhe_loc_lookup { + unsigned arg; +}; + +struct _vx_hist_entry { + void *loc; + unsigned short seq; + unsigned short type; + struct _vxhe_vxi vxi; + union { + struct _vxhe_set_clr sc; + struct _vxhe_loc_lookup ll; + }; +}; + +#ifdef CONFIG_VSERVER_HISTORY + +extern unsigned volatile int vxh_active; + +struct _vx_hist_entry *vxh_advance(void *loc); + + +static inline +void __vxh_copy_vxi(struct _vx_hist_entry *entry, struct vx_info *vxi) +{ + entry->vxi.ptr = vxi; + if (vxi) { + entry->vxi.usecnt = atomic_read(&vxi->vx_usecnt); + entry->vxi.tasks = atomic_read(&vxi->vx_tasks); + entry->vxi.xid = vxi->vx_id; + } +} + + +#define __HERE__ current_text_addr() + +#define __VXH_BODY(__type, __data, __here) \ + struct _vx_hist_entry *entry; \ + \ + preempt_disable(); \ + entry = vxh_advance(__here); \ + __data; \ + entry->type = __type; \ + preempt_enable(); + + + /* pass vxi only */ + +#define __VXH_SMPL \ + __vxh_copy_vxi(entry, vxi) + +static inline +void __vxh_smpl(struct vx_info *vxi, int __type, void *__here) +{ + __VXH_BODY(__type, __VXH_SMPL, __here) +} + + /* pass vxi and data (void *) */ + +#define __VXH_DATA \ + __vxh_copy_vxi(entry, vxi); \ + entry->sc.data = data + +static inline +void __vxh_data(struct vx_info *vxi, void *data, + int __type, void *__here) +{ + __VXH_BODY(__type, __VXH_DATA, __here) +} + + /* pass vxi and arg (long) */ + +#define __VXH_LONG \ + __vxh_copy_vxi(entry, vxi); \ + entry->ll.arg = arg + +static inline +void __vxh_long(struct vx_info *vxi, long arg, + int __type, void *__here) +{ + __VXH_BODY(__type, __VXH_LONG, __here) +} + + +static inline +void __vxh_throw_oops(void *__here) +{ + __VXH_BODY(VXH_THROW_OOPS, {}, __here); + /* prevent further acquisition */ + vxh_active = 0; +} + + +#define vxh_throw_oops() __vxh_throw_oops(__HERE__); + +#define __vxh_get_vx_info(v, h) __vxh_smpl(v, VXH_GET_VX_INFO, h); +#define __vxh_put_vx_info(v, h) __vxh_smpl(v, VXH_PUT_VX_INFO, h); + +#define __vxh_init_vx_info(v, d, h) \ + __vxh_data(v, d, VXH_INIT_VX_INFO, h); +#define __vxh_set_vx_info(v, d, h) \ + __vxh_data(v, d, VXH_SET_VX_INFO, h); +#define __vxh_clr_vx_info(v, d, h) \ + __vxh_data(v, d, VXH_CLR_VX_INFO, h); + +#define __vxh_claim_vx_info(v, d, h) \ + __vxh_data(v, d, VXH_CLAIM_VX_INFO, h); +#define __vxh_release_vx_info(v, d, h) \ + __vxh_data(v, d, VXH_RELEASE_VX_INFO, h); + +#define vxh_alloc_vx_info(v) \ + __vxh_smpl(v, VXH_ALLOC_VX_INFO, __HERE__); +#define vxh_dealloc_vx_info(v) \ + __vxh_smpl(v, VXH_DEALLOC_VX_INFO, __HERE__); + +#define vxh_hash_vx_info(v) \ + __vxh_smpl(v, VXH_HASH_VX_INFO, __HERE__); +#define vxh_unhash_vx_info(v) \ + __vxh_smpl(v, VXH_UNHASH_VX_INFO, __HERE__); + +#define vxh_loc_vx_info(v, l) \ + __vxh_long(v, l, VXH_LOC_VX_INFO, __HERE__); +#define vxh_lookup_vx_info(v, l) \ + __vxh_long(v, l, VXH_LOOKUP_VX_INFO, __HERE__); +#define vxh_create_vx_info(v, l) \ + __vxh_long(v, l, VXH_CREATE_VX_INFO, __HERE__); + +extern void vxh_dump_history(void); + + +#else /* CONFIG_VSERVER_HISTORY */ + +#define __HERE__ 0 + +#define vxh_throw_oops() do { } while (0) + +#define __vxh_get_vx_info(v, h) do { } while (0) +#define __vxh_put_vx_info(v, h) do { } while (0) + +#define __vxh_init_vx_info(v, d, h) do { } while (0) +#define __vxh_set_vx_info(v, d, h) do { } while (0) +#define __vxh_clr_vx_info(v, d, h) do { } while (0) + +#define __vxh_claim_vx_info(v, d, h) do { } while (0) +#define __vxh_release_vx_info(v, d, h) do { } while (0) + +#define vxh_alloc_vx_info(v) do { } while (0) +#define vxh_dealloc_vx_info(v) do { } while (0) + +#define vxh_hash_vx_info(v) do { } while (0) +#define vxh_unhash_vx_info(v) do { } while (0) + +#define vxh_loc_vx_info(v, l) do { } while (0) +#define vxh_lookup_vx_info(v, l) do { } while (0) +#define vxh_create_vx_info(v, l) do { } while (0) + +#define vxh_dump_history() do { } while (0) + + +#endif /* CONFIG_VSERVER_HISTORY */ + +#endif /* _VX_HISTORY_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/inode_cmd.h linux-source-2.6.22/include/linux/vserver/inode_cmd.h --- linux-source-2.6.22.orig/include/linux/vserver/inode_cmd.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/inode_cmd.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,59 @@ +#ifndef _VX_INODE_CMD_H +#define _VX_INODE_CMD_H + + +/* inode vserver commands */ + +#define VCMD_get_iattr VC_CMD(INODE, 1, 1) +#define VCMD_set_iattr VC_CMD(INODE, 2, 1) + +#define VCMD_fget_iattr VC_CMD(INODE, 3, 0) +#define VCMD_fset_iattr VC_CMD(INODE, 4, 0) + +struct vcmd_ctx_iattr_v1 { + const char __user *name; + uint32_t xid; + uint32_t flags; + uint32_t mask; +}; + +struct vcmd_ctx_fiattr_v0 { + uint32_t xid; + uint32_t flags; + uint32_t mask; +}; + + +#ifdef __KERNEL__ + + +#ifdef CONFIG_COMPAT + +#include + +struct vcmd_ctx_iattr_v1_x32 { + compat_uptr_t name_ptr; + uint32_t xid; + uint32_t flags; + uint32_t mask; +}; + +#endif /* CONFIG_COMPAT */ + +#include + +extern int vc_get_iattr(void __user *); +extern int vc_set_iattr(void __user *); + +extern int vc_fget_iattr(uint32_t, void __user *); +extern int vc_fset_iattr(uint32_t, void __user *); + +#ifdef CONFIG_COMPAT + +extern int vc_get_iattr_x32(void __user *); +extern int vc_set_iattr_x32(void __user *); + +#endif /* CONFIG_COMPAT */ + +#endif /* __KERNEL__ */ +#endif /* _VX_INODE_CMD_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/inode.h linux-source-2.6.22/include/linux/vserver/inode.h --- linux-source-2.6.22.orig/include/linux/vserver/inode.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/inode.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,38 @@ +#ifndef _VX_INODE_H +#define _VX_INODE_H + + +#define IATTR_TAG 0x01000000 + +#define IATTR_ADMIN 0x00000001 +#define IATTR_WATCH 0x00000002 +#define IATTR_HIDE 0x00000004 +#define IATTR_FLAGS 0x00000007 + +#define IATTR_BARRIER 0x00010000 +#define IATTR_IUNLINK 0x00020000 +#define IATTR_IMMUTABLE 0x00040000 + +#ifdef __KERNEL__ + + +#ifdef CONFIG_VSERVER_PROC_SECURE +#define IATTR_PROC_DEFAULT ( IATTR_ADMIN | IATTR_HIDE ) +#define IATTR_PROC_SYMLINK ( IATTR_ADMIN ) +#else +#define IATTR_PROC_DEFAULT ( IATTR_ADMIN ) +#define IATTR_PROC_SYMLINK ( IATTR_ADMIN ) +#endif + +#define vx_hide_check(c, m) (((m) & IATTR_HIDE) ? vx_check(c, m) : 1) + +#endif /* __KERNEL__ */ + +/* inode ioctls */ + +#define FIOC_GETXFLG _IOR('x', 5, long) +#define FIOC_SETXFLG _IOW('x', 6, long) + +#else /* _VX_INODE_H */ +#warning duplicate inclusion +#endif /* _VX_INODE_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/Kbuild linux-source-2.6.22/include/linux/vserver/Kbuild --- linux-source-2.6.22.orig/include/linux/vserver/Kbuild 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/Kbuild 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,9 @@ + +unifdef-y += context_cmd.h network_cmd.h space_cmd.h \ + cacct_cmd.h cvirt_cmd.h limit_cmd.h dlimit_cmd.h \ + inode_cmd.h sched_cmd.h signal_cmd.h debug_cmd.h + +unifdef-y += switch.h network.h monitor.h inode.h + +unifdef-y += legacy.h + diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/legacy.h linux-source-2.6.22/include/linux/vserver/legacy.h --- linux-source-2.6.22.orig/include/linux/vserver/legacy.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/legacy.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,49 @@ +#ifndef _VX_LEGACY_H +#define _VX_LEGACY_H + +#include "switch.h" + + +/* compatibiliy vserver commands */ + +#define VCMD_new_s_context VC_CMD(COMPAT, 1, 1) +#define VCMD_set_ipv4root VC_CMD(COMPAT, 2, 3) + +#define VCMD_create_context VC_CMD(VSETUP, 1, 0) + +/* compatibiliy vserver arguments */ + +struct vcmd_new_s_context_v1 { + uint32_t remove_cap; + uint32_t flags; +}; + +struct vcmd_set_ipv4root_v3 { + /* number of pairs in id */ + uint32_t broadcast; + struct { + uint32_t ip; + uint32_t mask; + } nx_mask_pair[NB_IPV4ROOT]; +}; + + +#define VX_INFO_LOCK 1 /* Can't request a new vx_id */ +#define VX_INFO_NPROC 4 /* Limit number of processes in a context */ +#define VX_INFO_PRIVATE 8 /* Noone can join this security context */ +#define VX_INFO_INIT 16 /* This process wants to become the */ + /* logical process 1 of the security */ + /* context */ +#define VX_INFO_HIDEINFO 32 /* Hide some information in /proc */ +#define VX_INFO_ULIMIT 64 /* Use ulimit of the current process */ + /* to become the global limits */ + /* of the context */ +#define VX_INFO_NAMESPACE 128 /* save private namespace */ + + +#ifdef __KERNEL__ +extern int vc_new_s_context(uint32_t, void __user *); +extern int vc_set_ipv4root(uint32_t, void __user *); + +#endif /* __KERNEL__ */ +#endif /* _VX_LEGACY_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/limit_cmd.h linux-source-2.6.22/include/linux/vserver/limit_cmd.h --- linux-source-2.6.22.orig/include/linux/vserver/limit_cmd.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/limit_cmd.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,69 @@ +#ifndef _VX_LIMIT_CMD_H +#define _VX_LIMIT_CMD_H + + +/* rlimit vserver commands */ + +#define VCMD_get_rlimit VC_CMD(RLIMIT, 1, 0) +#define VCMD_set_rlimit VC_CMD(RLIMIT, 2, 0) +#define VCMD_get_rlimit_mask VC_CMD(RLIMIT, 3, 0) +#define VCMD_reset_minmax VC_CMD(RLIMIT, 9, 0) + +struct vcmd_ctx_rlimit_v0 { + uint32_t id; + uint64_t minimum; + uint64_t softlimit; + uint64_t maximum; +}; + +struct vcmd_ctx_rlimit_mask_v0 { + uint32_t minimum; + uint32_t softlimit; + uint32_t maximum; +}; + +#define VCMD_rlimit_stat VC_CMD(VSTAT, 1, 0) + +struct vcmd_rlimit_stat_v0 { + uint32_t id; + uint32_t hits; + uint64_t value; + uint64_t minimum; + uint64_t maximum; +}; + +#define CRLIM_UNSET (0ULL) +#define CRLIM_INFINITY (~0ULL) +#define CRLIM_KEEP (~1ULL) + +#ifdef __KERNEL__ + +#ifdef CONFIG_IA32_EMULATION + +struct vcmd_ctx_rlimit_v0_x32 { + uint32_t id; + uint64_t minimum; + uint64_t softlimit; + uint64_t maximum; +} __attribute__ ((aligned (4))); + +#endif /* CONFIG_IA32_EMULATION */ + +#include + +extern int vc_get_rlimit_mask(uint32_t, void __user *); +extern int vc_get_rlimit(struct vx_info *, void __user *); +extern int vc_set_rlimit(struct vx_info *, void __user *); +extern int vc_reset_minmax(struct vx_info *, void __user *); + +extern int vc_rlimit_stat(struct vx_info *, void __user *); + +#ifdef CONFIG_IA32_EMULATION + +extern int vc_get_rlimit_x32(struct vx_info *, void __user *); +extern int vc_set_rlimit_x32(struct vx_info *, void __user *); + +#endif /* CONFIG_IA32_EMULATION */ + +#endif /* __KERNEL__ */ +#endif /* _VX_LIMIT_CMD_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/limit_def.h linux-source-2.6.22/include/linux/vserver/limit_def.h --- linux-source-2.6.22.orig/include/linux/vserver/limit_def.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/limit_def.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,47 @@ +#ifndef _VX_LIMIT_DEF_H +#define _VX_LIMIT_DEF_H + +#include +#include + +#include "limit.h" + + +struct _vx_res_limit { + rlim_t soft; /* Context soft limit */ + rlim_t hard; /* Context hard limit */ + + rlim_atomic_t rcur; /* Current value */ + rlim_t rmin; /* Context minimum */ + rlim_t rmax; /* Context maximum */ + + atomic_t lhit; /* Limit hits */ +}; + +/* context sub struct */ + +struct _vx_limit { + struct _vx_res_limit res[NUM_LIMITS]; +}; + +#ifdef CONFIG_VSERVER_DEBUG + +static inline void __dump_vx_limit(struct _vx_limit *limit) +{ + int i; + + printk("\t_vx_limit:"); + for (i = 0; i < NUM_LIMITS; i++) { + printk("\t [%2d] = %8lu %8lu/%8lu, %8ld/%8ld, %8d\n", + i, (unsigned long)__rlim_get(limit, i), + (unsigned long)__rlim_rmin(limit, i), + (unsigned long)__rlim_rmax(limit, i), + (long)__rlim_soft(limit, i), + (long)__rlim_hard(limit, i), + atomic_read(&__rlim_lhit(limit, i))); + } +} + +#endif + +#endif /* _VX_LIMIT_DEF_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/limit.h linux-source-2.6.22/include/linux/vserver/limit.h --- linux-source-2.6.22.orig/include/linux/vserver/limit.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/limit.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,68 @@ +#ifndef _VX_LIMIT_H +#define _VX_LIMIT_H + + +#define VLIMIT_NSOCK 16 +#define VLIMIT_OPENFD 17 +#define VLIMIT_ANON 18 +#define VLIMIT_SHMEM 19 +#define VLIMIT_SEMARY 20 +#define VLIMIT_NSEMS 21 +#define VLIMIT_DENTRY 22 +#define VLIMIT_MAPPED 23 + + +#ifdef __KERNEL__ + +#define VLIM_NOCHECK ((1L << VLIMIT_DENTRY) | (1L << RLIMIT_RSS)) + +/* keep in sync with CRLIM_INFINITY */ + +#define VLIM_INFINITY (~0ULL) + +#ifndef RLIM_INFINITY +#warning RLIM_INFINITY is undefined +#endif + +#define __rlim_val(l, r, v) ((l)->res[r].v) + +#define __rlim_soft(l, r) __rlim_val(l, r, soft) +#define __rlim_hard(l, r) __rlim_val(l, r, hard) + +#define __rlim_rcur(l, r) __rlim_val(l, r, rcur) +#define __rlim_rmin(l, r) __rlim_val(l, r, rmin) +#define __rlim_rmax(l, r) __rlim_val(l, r, rmax) + +#define __rlim_lhit(l, r) __rlim_val(l, r, lhit) +#define __rlim_hit(l, r) atomic_inc(&__rlim_lhit(l, r)) + +typedef atomic_long_t rlim_atomic_t; +typedef unsigned long rlim_t; + +#define __rlim_get(l, r) atomic_long_read(&__rlim_rcur(l, r)) +#define __rlim_set(l, r, v) atomic_long_set(&__rlim_rcur(l, r), v) +#define __rlim_inc(l, r) atomic_long_inc(&__rlim_rcur(l, r)) +#define __rlim_dec(l, r) atomic_long_dec(&__rlim_rcur(l, r)) +#define __rlim_add(l, r, v) atomic_long_add(v, &__rlim_rcur(l, r)) +#define __rlim_sub(l, r, v) atomic_long_sub(v, &__rlim_rcur(l, r)) + + +#if (RLIM_INFINITY == VLIM_INFINITY) +#define VX_VLIM(r) ((long long)(long)(r)) +#define VX_RLIM(v) ((rlim_t)(v)) +#else +#define VX_VLIM(r) (((r) == RLIM_INFINITY) \ + ? VLIM_INFINITY : (long long)(r)) +#define VX_RLIM(v) (((v) == VLIM_INFINITY) \ + ? RLIM_INFINITY : (rlim_t)(v)) +#endif + +struct sysinfo; + +void vx_vsi_meminfo(struct sysinfo *); +void vx_vsi_swapinfo(struct sysinfo *); + +#define NUM_LIMITS 24 + +#endif /* __KERNEL__ */ +#endif /* _VX_LIMIT_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/limit_int.h linux-source-2.6.22/include/linux/vserver/limit_int.h --- linux-source-2.6.22.orig/include/linux/vserver/limit_int.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/limit_int.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,198 @@ +#ifndef _VX_LIMIT_INT_H +#define _VX_LIMIT_INT_H + +#include "context.h" + +#ifdef __KERNEL__ + +#define VXD_RCRES_COND(r) VXD_CBIT(cres, r) +#define VXD_RLIMIT_COND(r) VXD_CBIT(limit, r) + +extern const char *vlimit_name[NUM_LIMITS]; + +static inline void __vx_acc_cres(struct vx_info *vxi, + int res, int dir, void *_data, char *_file, int _line) +{ + if (VXD_RCRES_COND(res)) + vxlprintk(1, "vx_acc_cres[%5d,%s,%2d]: %5ld%s (%p)", + (vxi ? vxi->vx_id : -1), vlimit_name[res], res, + (vxi ? (long)__rlim_get(&vxi->limit, res) : 0), + (dir > 0) ? "++" : "--", _data, _file, _line); + if (!vxi) + return; + + if (dir > 0) + __rlim_inc(&vxi->limit, res); + else + __rlim_dec(&vxi->limit, res); +} + +static inline void __vx_add_cres(struct vx_info *vxi, + int res, int amount, void *_data, char *_file, int _line) +{ + if (VXD_RCRES_COND(res)) + vxlprintk(1, "vx_add_cres[%5d,%s,%2d]: %5ld += %5d (%p)", + (vxi ? vxi->vx_id : -1), vlimit_name[res], res, + (vxi ? (long)__rlim_get(&vxi->limit, res) : 0), + amount, _data, _file, _line); + if (amount == 0) + return; + if (!vxi) + return; + __rlim_add(&vxi->limit, res, amount); +} + +static inline +int __vx_cres_adjust_max(struct _vx_limit *limit, int res, rlim_t value) +{ + int cond = (value > __rlim_rmax(limit, res)); + + if (cond) + __rlim_rmax(limit, res) = value; + return cond; +} + +static inline +int __vx_cres_adjust_min(struct _vx_limit *limit, int res, rlim_t value) +{ + int cond = (value < __rlim_rmin(limit, res)); + + if (cond) + __rlim_rmin(limit, res) = value; + return cond; +} + +static inline +void __vx_cres_fixup(struct _vx_limit *limit, int res, rlim_t value) +{ + if (!__vx_cres_adjust_max(limit, res, value)) + __vx_cres_adjust_min(limit, res, value); +} + + +/* return values: + +1 ... no limit hit + -1 ... over soft limit + 0 ... over hard limit */ + +static inline int __vx_cres_avail(struct vx_info *vxi, + int res, int num, char *_file, int _line) +{ + struct _vx_limit *limit; + rlim_t value; + + if (VXD_RLIMIT_COND(res)) + vxlprintk(1, "vx_cres_avail[%5d,%s,%2d]: %5ld/%5ld > %5ld + %5d", + (vxi ? vxi->vx_id : -1), vlimit_name[res], res, + (vxi ? (long)__rlim_soft(&vxi->limit, res) : -1), + (vxi ? (long)__rlim_hard(&vxi->limit, res) : -1), + (vxi ? (long)__rlim_get(&vxi->limit, res) : 0), + num, _file, _line); + if (!vxi) + return 1; + + limit = &vxi->limit; + value = __rlim_get(limit, res); + + if (!__vx_cres_adjust_max(limit, res, value)) + __vx_cres_adjust_min(limit, res, value); + + if (num == 0) + return 1; + + if (__rlim_soft(limit, res) == RLIM_INFINITY) + return -1; + if (value + num <= __rlim_soft(limit, res)) + return -1; + + if (__rlim_hard(limit, res) == RLIM_INFINITY) + return 1; + if (value + num <= __rlim_hard(limit, res)) + return 1; + + __rlim_hit(limit, res); + return 0; +} + + +static const int VLA_RSS[] = { RLIMIT_RSS, VLIMIT_ANON, VLIMIT_MAPPED, 0 }; + +static inline +rlim_t __vx_cres_array_sum(struct _vx_limit *limit, const int *array) +{ + rlim_t value, sum = 0; + int res; + + while ((res = *array++)) { + value = __rlim_get(limit, res); + __vx_cres_fixup(limit, res, value); + sum += value; + } + return sum; +} + +static inline +rlim_t __vx_cres_array_fixup(struct _vx_limit *limit, const int *array) +{ + rlim_t value = __vx_cres_array_sum(limit, array + 1); + int res = *array; + + if (value == __rlim_get(limit, res)) + return value; + + __rlim_set(limit, res, value); + /* now adjust min/max */ + if (!__vx_cres_adjust_max(limit, res, value)) + __vx_cres_adjust_min(limit, res, value); + + return value; +} + +static inline int __vx_cres_array_avail(struct vx_info *vxi, + const int *array, int num, char *_file, int _line) +{ + struct _vx_limit *limit; + rlim_t value = 0; + int res; + + if (num == 0) + return 1; + if (!vxi) + return 1; + + limit = &vxi->limit; + res = *array; + value = __vx_cres_array_sum(limit, array + 1); + + __rlim_set(limit, res, value); + __vx_cres_fixup(limit, res, value); + + return __vx_cres_avail(vxi, res, num, _file, _line); +} + + +static inline void vx_limit_fixup(struct _vx_limit *limit, int id) +{ + rlim_t value; + int res; + + /* complex resources first */ + if ((id < 0) || (id == RLIMIT_RSS)) + __vx_cres_array_fixup(limit, VLA_RSS); + + for (res = 0; res < NUM_LIMITS; res++) { + if ((id > 0) && (res != id)) + continue; + + value = __rlim_get(limit, res); + __vx_cres_fixup(limit, res, value); + + /* not supposed to happen, maybe warn? */ + if (__rlim_rmax(limit, res) > __rlim_hard(limit, res)) + __rlim_rmax(limit, res) = __rlim_hard(limit, res); + } +} + + +#endif /* __KERNEL__ */ +#endif /* _VX_LIMIT_INT_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/monitor.h linux-source-2.6.22/include/linux/vserver/monitor.h --- linux-source-2.6.22.orig/include/linux/vserver/monitor.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/monitor.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,95 @@ +#ifndef _VX_MONITOR_H +#define _VX_MONITOR_H + + +enum { + VXM_UNUSED = 0, + + VXM_SYNC = 0x10, + + VXM_UPDATE = 0x20, + VXM_UPDATE_1, + VXM_UPDATE_2, + + VXM_RQINFO_1 = 0x24, + VXM_RQINFO_2, + + VXM_ACTIVATE = 0x40, + VXM_DEACTIVATE, + VXM_IDLE, + + VXM_HOLD = 0x44, + VXM_UNHOLD, + + VXM_MIGRATE = 0x48, + VXM_RESCHED, + + /* all other bits are flags */ + VXM_SCHED = 0x80, +}; + +struct _vxm_update_1 { + uint32_t tokens_max; + uint32_t fill_rate; + uint32_t interval; +}; + +struct _vxm_update_2 { + uint32_t tokens_min; + uint32_t fill_rate; + uint32_t interval; +}; + +struct _vxm_rqinfo_1 { + uint16_t running; + uint16_t onhold; + uint16_t iowait; + uint16_t uintr; + uint32_t idle_tokens; +}; + +struct _vxm_rqinfo_2 { + uint32_t norm_time; + uint32_t idle_time; + uint32_t idle_skip; +}; + +struct _vxm_sched { + uint32_t tokens; + uint32_t norm_time; + uint32_t idle_time; +}; + +struct _vxm_task { + uint16_t pid; + uint16_t state; +}; + +struct _vxm_event { + uint32_t jif; + union { + uint32_t seq; + uint32_t sec; + }; + union { + uint32_t tokens; + uint32_t nsec; + struct _vxm_task tsk; + }; +}; + +struct _vx_mon_entry { + uint16_t type; + uint16_t xid; + union { + struct _vxm_event ev; + struct _vxm_sched sd; + struct _vxm_update_1 u1; + struct _vxm_update_2 u2; + struct _vxm_rqinfo_1 q1; + struct _vxm_rqinfo_2 q2; + }; +}; + + +#endif /* _VX_MONITOR_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/network_cmd.h linux-source-2.6.22/include/linux/vserver/network_cmd.h --- linux-source-2.6.22.orig/include/linux/vserver/network_cmd.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/network_cmd.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,89 @@ +#ifndef _VX_NETWORK_CMD_H +#define _VX_NETWORK_CMD_H + + +/* vinfo commands */ + +#define VCMD_task_nid VC_CMD(VINFO, 2, 0) + +#ifdef __KERNEL__ +extern int vc_task_nid(uint32_t, void __user *); + +#endif /* __KERNEL__ */ + +#define VCMD_nx_info VC_CMD(VINFO, 6, 0) + +struct vcmd_nx_info_v0 { + uint32_t nid; + /* more to come */ +}; + +#ifdef __KERNEL__ +extern int vc_nx_info(struct nx_info *, void __user *); + +#endif /* __KERNEL__ */ + +#define VCMD_net_create_v0 VC_CMD(VNET, 1, 0) +#define VCMD_net_create VC_CMD(VNET, 1, 1) + +struct vcmd_net_create { + uint64_t flagword; +}; + +#define VCMD_net_migrate VC_CMD(NETMIG, 1, 0) + +#define VCMD_net_add VC_CMD(NETALT, 1, 0) +#define VCMD_net_remove VC_CMD(NETALT, 2, 0) + +struct vcmd_net_addr_v0 { + uint16_t type; + uint16_t count; + uint32_t ip[4]; + uint32_t mask[4]; + /* more to come */ +}; + + +#ifdef __KERNEL__ +extern int vc_net_create(uint32_t, void __user *); +extern int vc_net_migrate(struct nx_info *, void __user *); + +extern int vc_net_add(struct nx_info *, void __user *); +extern int vc_net_remove(struct nx_info *, void __user *); + +#endif /* __KERNEL__ */ + + +/* flag commands */ + +#define VCMD_get_nflags VC_CMD(FLAGS, 5, 0) +#define VCMD_set_nflags VC_CMD(FLAGS, 6, 0) + +struct vcmd_net_flags_v0 { + uint64_t flagword; + uint64_t mask; +}; + +#ifdef __KERNEL__ +extern int vc_get_nflags(struct nx_info *, void __user *); +extern int vc_set_nflags(struct nx_info *, void __user *); + +#endif /* __KERNEL__ */ + + +/* network caps commands */ + +#define VCMD_get_ncaps VC_CMD(FLAGS, 7, 0) +#define VCMD_set_ncaps VC_CMD(FLAGS, 8, 0) + +struct vcmd_net_caps_v0 { + uint64_t ncaps; + uint64_t cmask; +}; + +#ifdef __KERNEL__ +extern int vc_get_ncaps(struct nx_info *, void __user *); +extern int vc_set_ncaps(struct nx_info *, void __user *); + +#endif /* __KERNEL__ */ +#endif /* _VX_CONTEXT_CMD_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/network.h linux-source-2.6.22/include/linux/vserver/network.h --- linux-source-2.6.22.orig/include/linux/vserver/network.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/network.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,142 @@ +#ifndef _VX_NETWORK_H +#define _VX_NETWORK_H + +#include + + +#define MAX_N_CONTEXT 65535 /* Arbitrary limit */ + +#define NX_DYNAMIC_ID ((uint32_t)-1) /* id for dynamic context */ + +#define NB_IPV4ROOT 16 + + +/* network flags */ + +#define NXF_INFO_PRIVATE 0x00000008 + +#define NXF_STATE_SETUP (1ULL << 32) +#define NXF_STATE_ADMIN (1ULL << 34) + +#define NXF_SC_HELPER (1ULL << 36) +#define NXF_PERSISTENT (1ULL << 38) + +#define NXF_ONE_TIME (0x0005ULL << 32) + +#define NXF_INIT_SET (NXF_STATE_ADMIN) + + +/* address types */ + +#define NXA_TYPE_IPV4 1 +#define NXA_TYPE_IPV6 2 + +#define NXA_MOD_BCAST (1 << 8) + +#define NXA_TYPE_ANY ((uint16_t)-1) + + +#ifdef __KERNEL__ + +#include +#include +#include +#include + + +struct nx_info { + struct hlist_node nx_hlist; /* linked list of nxinfos */ + nid_t nx_id; /* vnet id */ + atomic_t nx_usecnt; /* usage count */ + atomic_t nx_tasks; /* tasks count */ + int nx_state; /* context state */ + + uint64_t nx_flags; /* network flag word */ + uint64_t nx_ncaps; /* network capabilities */ + + int nbipv4; + __u32 ipv4[NB_IPV4ROOT]; /* Process can only bind to these IPs */ + /* The first one is used to connect */ + /* and for bind any service */ + /* The other must be used explicity */ + __u32 mask[NB_IPV4ROOT]; /* Netmask for each ipv4 */ + /* Used to select the proper source */ + /* address for sockets */ + __u32 v4_bcast; /* Broadcast address to receive UDP */ + + char nx_name[65]; /* network context name */ +}; + + +/* status flags */ + +#define NXS_HASHED 0x0001 +#define NXS_SHUTDOWN 0x0100 +#define NXS_RELEASED 0x8000 + +/* check conditions */ + +#define NX_ADMIN 0x0001 +#define NX_WATCH 0x0002 +#define NX_BLEND 0x0004 +#define NX_HOSTID 0x0008 + +#define NX_IDENT 0x0010 +#define NX_EQUIV 0x0020 +#define NX_PARENT 0x0040 +#define NX_CHILD 0x0080 + +#define NX_ARG_MASK 0x00F0 + +#define NX_DYNAMIC 0x0100 +#define NX_STATIC 0x0200 + +#define NX_ATR_MASK 0x0F00 + + +extern struct nx_info *lookup_nx_info(int); + +extern int get_nid_list(int, unsigned int *, int); +extern int nid_is_hashed(nid_t); + +extern int nx_migrate_task(struct task_struct *, struct nx_info *); + +extern long vs_net_change(struct nx_info *, unsigned int); + +struct in_ifaddr; +struct net_device; + +#ifdef CONFIG_INET +int ifa_in_nx_info(struct in_ifaddr *, struct nx_info *); +int dev_in_nx_info(struct net_device *, struct nx_info *); + +#else /* CONFIG_INET */ +static inline +int ifa_in_nx_info(struct in_ifaddr *a, struct nx_info *n) +{ + return 1; +} + +static inline +int dev_in_nx_info(struct net_device *d, struct nx_info *n) +{ + return 1; +} +#endif /* CONFIG_INET */ + +struct sock; + +#ifdef CONFIG_INET +int nx_addr_conflict(struct nx_info *, uint32_t, const struct sock *); +#else /* CONFIG_INET */ +static inline +int nx_addr_conflict(struct nx_info *n, uint32_t a, const struct sock *s) +{ + return 1; +} +#endif /* CONFIG_INET */ + +#endif /* __KERNEL__ */ +#else /* _VX_NETWORK_H */ +#warning duplicate inclusion +#endif /* _VX_NETWORK_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/percpu.h linux-source-2.6.22/include/linux/vserver/percpu.h --- linux-source-2.6.22.orig/include/linux/vserver/percpu.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/percpu.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,14 @@ +#ifndef _VX_PERCPU_H +#define _VX_PERCPU_H + +#include "cvirt_def.h" +#include "sched_def.h" + +struct _vx_percpu { + struct _vx_cvirt_pc cvirt; + struct _vx_sched_pc sched; +}; + +#define PERCPU_PERCTX (sizeof(struct _vx_percpu)) + +#endif /* _VX_PERCPU_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/sched_cmd.h linux-source-2.6.22/include/linux/vserver/sched_cmd.h --- linux-source-2.6.22.orig/include/linux/vserver/sched_cmd.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/sched_cmd.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,108 @@ +#ifndef _VX_SCHED_CMD_H +#define _VX_SCHED_CMD_H + + +/* sched vserver commands */ + +#define VCMD_set_sched_v2 VC_CMD(SCHED, 1, 2) +#define VCMD_set_sched_v3 VC_CMD(SCHED, 1, 3) +#define VCMD_set_sched_v4 VC_CMD(SCHED, 1, 4) + +struct vcmd_set_sched_v2 { + int32_t fill_rate; + int32_t interval; + int32_t tokens; + int32_t tokens_min; + int32_t tokens_max; + uint64_t cpu_mask; +}; + +struct vcmd_set_sched_v3 { + uint32_t set_mask; + int32_t fill_rate; + int32_t interval; + int32_t tokens; + int32_t tokens_min; + int32_t tokens_max; + int32_t priority_bias; +}; + +struct vcmd_set_sched_v4 { + uint32_t set_mask; + int32_t fill_rate; + int32_t interval; + int32_t tokens; + int32_t tokens_min; + int32_t tokens_max; + int32_t prio_bias; + int32_t cpu_id; + int32_t bucket_id; +}; + +#define VCMD_set_sched VC_CMD(SCHED, 1, 5) +#define VCMD_get_sched VC_CMD(SCHED, 2, 5) + +struct vcmd_sched_v5 { + uint32_t mask; + int32_t cpu_id; + int32_t bucket_id; + int32_t fill_rate[2]; + int32_t interval[2]; + int32_t tokens; + int32_t tokens_min; + int32_t tokens_max; + int32_t prio_bias; +}; + +#define VXSM_FILL_RATE 0x0001 +#define VXSM_INTERVAL 0x0002 +#define VXSM_FILL_RATE2 0x0004 +#define VXSM_INTERVAL2 0x0008 +#define VXSM_TOKENS 0x0010 +#define VXSM_TOKENS_MIN 0x0020 +#define VXSM_TOKENS_MAX 0x0040 +#define VXSM_PRIO_BIAS 0x0100 + +#define VXSM_IDLE_TIME 0x0200 +#define VXSM_FORCE 0x0400 + +#define VXSM_V3_MASK 0x0173 +#define VXSM_SET_MASK 0x01FF + +#define VXSM_CPU_ID 0x1000 +#define VXSM_BUCKET_ID 0x2000 + +#define VXSM_MSEC 0x4000 + +#define SCHED_KEEP (-2) /* only for v2 */ + +#ifdef __KERNEL__ + +#include + +extern int vc_set_sched_v2(struct vx_info *, void __user *); +extern int vc_set_sched_v3(struct vx_info *, void __user *); +extern int vc_set_sched_v4(struct vx_info *, void __user *); +extern int vc_set_sched(struct vx_info *, void __user *); +extern int vc_get_sched(struct vx_info *, void __user *); + +#endif /* __KERNEL__ */ + +#define VCMD_sched_info VC_CMD(SCHED, 3, 0) + +struct vcmd_sched_info { + int32_t cpu_id; + int32_t bucket_id; + uint64_t user_msec; + uint64_t sys_msec; + uint64_t hold_msec; + uint32_t token_usec; + int32_t vavavoom; +}; + +#ifdef __KERNEL__ + +extern int vc_sched_info(struct vx_info *, void __user *); + +#endif /* __KERNEL__ */ +#endif /* _VX_SCHED_CMD_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/sched_def.h linux-source-2.6.22/include/linux/vserver/sched_def.h --- linux-source-2.6.22.orig/include/linux/vserver/sched_def.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/sched_def.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,68 @@ +#ifndef _VX_SCHED_DEF_H +#define _VX_SCHED_DEF_H + +#include +#include +#include +#include +#include + + +/* context sub struct */ + +struct _vx_sched { + spinlock_t tokens_lock; /* lock for token bucket */ + + int tokens; /* number of CPU tokens */ + int fill_rate[2]; /* Fill rate: add X tokens... */ + int interval[2]; /* Divisor: per Y jiffies */ + int tokens_min; /* Limit: minimum for unhold */ + int tokens_max; /* Limit: no more than N tokens */ + + int prio_bias; /* bias offset for priority */ + + unsigned update_mask; /* which features should be updated */ + cpumask_t update; /* CPUs which should update */ +}; + +struct _vx_sched_pc { + int tokens; /* number of CPU tokens */ + int flags; /* bucket flags */ + + int fill_rate[2]; /* Fill rate: add X tokens... */ + int interval[2]; /* Divisor: per Y jiffies */ + int tokens_min; /* Limit: minimum for unhold */ + int tokens_max; /* Limit: no more than N tokens */ + + int prio_bias; /* bias offset for priority */ + int vavavoom; /* last calculated vavavoom */ + + unsigned long norm_time; /* last time accounted */ + unsigned long idle_time; /* non linear time for fair sched */ + unsigned long token_time; /* token time for accounting */ + unsigned long onhold; /* jiffies when put on hold */ + + uint64_t user_ticks; /* token tick events */ + uint64_t sys_ticks; /* token tick events */ + uint64_t hold_ticks; /* token ticks paused */ +}; + + +#define VXSF_ONHOLD 0x0001 +#define VXSF_IDLE_TIME 0x0100 + +#ifdef CONFIG_VSERVER_DEBUG + +static inline void __dump_vx_sched(struct _vx_sched *sched) +{ + printk("\t_vx_sched:\n"); + printk("\t tokens: %4d/%4d, %4d/%4d, %4d, %4d\n", + sched->fill_rate[0], sched->interval[0], + sched->fill_rate[1], sched->interval[1], + sched->tokens_min, sched->tokens_max); + printk("\t priority = %4d\n", sched->prio_bias); +} + +#endif + +#endif /* _VX_SCHED_DEF_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/sched.h linux-source-2.6.22/include/linux/vserver/sched.h --- linux-source-2.6.22.orig/include/linux/vserver/sched.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/sched.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,26 @@ +#ifndef _VX_SCHED_H +#define _VX_SCHED_H + + +#ifdef __KERNEL__ + +struct timespec; + +void vx_vsi_uptime(struct timespec *, struct timespec *); + + +struct vx_info; + +void vx_update_load(struct vx_info *); + + +int vx_tokens_recalc(struct _vx_sched_pc *, + unsigned long *, unsigned long *, int [2]); + +void vx_update_sched_param(struct _vx_sched *sched, + struct _vx_sched_pc *sched_pc); + +#endif /* __KERNEL__ */ +#else /* _VX_SCHED_H */ +#warning duplicate inclusion +#endif /* _VX_SCHED_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/signal_cmd.h linux-source-2.6.22/include/linux/vserver/signal_cmd.h --- linux-source-2.6.22.orig/include/linux/vserver/signal_cmd.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/signal_cmd.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,43 @@ +#ifndef _VX_SIGNAL_CMD_H +#define _VX_SIGNAL_CMD_H + + +/* signalling vserver commands */ + +#define VCMD_ctx_kill VC_CMD(PROCTRL, 1, 0) +#define VCMD_wait_exit VC_CMD(EVENT, 99, 0) + +struct vcmd_ctx_kill_v0 { + int32_t pid; + int32_t sig; +}; + +struct vcmd_wait_exit_v0 { + int32_t reboot_cmd; + int32_t exit_code; +}; + +#ifdef __KERNEL__ + +extern int vc_ctx_kill(struct vx_info *, void __user *); +extern int vc_wait_exit(struct vx_info *, void __user *); + +#endif /* __KERNEL__ */ + +/* process alteration commands */ + +#define VCMD_get_pflags VC_CMD(PROCALT, 5, 0) +#define VCMD_set_pflags VC_CMD(PROCALT, 6, 0) + +struct vcmd_pflags_v0 { + uint32_t flagword; + uint32_t mask; +}; + +#ifdef __KERNEL__ + +extern int vc_get_pflags(uint32_t pid, void __user *); +extern int vc_set_pflags(uint32_t pid, void __user *); + +#endif /* __KERNEL__ */ +#endif /* _VX_SIGNAL_CMD_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/signal.h linux-source-2.6.22/include/linux/vserver/signal.h --- linux-source-2.6.22.orig/include/linux/vserver/signal.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/signal.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,14 @@ +#ifndef _VX_SIGNAL_H +#define _VX_SIGNAL_H + + +#ifdef __KERNEL__ + +struct vx_info; + +int vx_info_kill(struct vx_info *, int, int); + +#endif /* __KERNEL__ */ +#else /* _VX_SIGNAL_H */ +#warning duplicate inclusion +#endif /* _VX_SIGNAL_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/space_cmd.h linux-source-2.6.22/include/linux/vserver/space_cmd.h --- linux-source-2.6.22.orig/include/linux/vserver/space_cmd.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/space_cmd.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,26 @@ +#ifndef _VX_SPACE_CMD_H +#define _VX_SPACE_CMD_H + + +#define VCMD_enter_space_v0 VC_CMD(PROCALT, 1, 0) +#define VCMD_enter_space VC_CMD(PROCALT, 1, 1) + +#define VCMD_set_space_v0 VC_CMD(PROCALT, 3, 0) +#define VCMD_set_space VC_CMD(PROCALT, 3, 1) + +#define VCMD_get_space_mask VC_CMD(PROCALT, 4, 0) + + +struct vcmd_space_mask { + uint64_t mask; +}; + + +#ifdef __KERNEL__ + +extern int vc_enter_space(struct vx_info *, void __user *); +extern int vc_set_space(struct vx_info *, void __user *); +extern int vc_get_space_mask(struct vx_info *, void __user *); + +#endif /* __KERNEL__ */ +#endif /* _VX_SPACE_CMD_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/space.h linux-source-2.6.22/include/linux/vserver/space.h --- linux-source-2.6.22.orig/include/linux/vserver/space.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/space.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,13 @@ +#ifndef _VX_SPACE_H +#define _VX_SPACE_H + + +#include + +struct vx_info; + +int vx_set_space(struct vx_info *vxi, unsigned long mask); + +#else /* _VX_SPACE_H */ +#warning duplicate inclusion +#endif /* _VX_SPACE_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/switch.h linux-source-2.6.22/include/linux/vserver/switch.h --- linux-source-2.6.22.orig/include/linux/vserver/switch.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/switch.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,95 @@ +#ifndef _VX_SWITCH_H +#define _VX_SWITCH_H + +#include + + +#define VC_CATEGORY(c) (((c) >> 24) & 0x3F) +#define VC_COMMAND(c) (((c) >> 16) & 0xFF) +#define VC_VERSION(c) ((c) & 0xFFF) + +#define VC_CMD(c, i, v) ((((VC_CAT_ ## c) & 0x3F) << 24) \ + | (((i) & 0xFF) << 16) | ((v) & 0xFFF)) + +/* + + Syscall Matrix V2.8 + + |VERSION|CREATE |MODIFY |MIGRATE|CONTROL|EXPERIM| |SPECIAL|SPECIAL| + |STATS |DESTROY|ALTER |CHANGE |LIMIT |TEST | | | | + |INFO |SETUP | |MOVE | | | | | | + -------+-------+-------+-------+-------+-------+-------+ +-------+-------+ + SYSTEM |VERSION|VSETUP |VHOST | | | | |DEVICES| | + HOST | 00| 01| 02| 03| 04| 05| | 06| 07| + -------+-------+-------+-------+-------+-------+-------+ +-------+-------+ + CPU | |VPROC |PROCALT|PROCMIG|PROCTRL| | |SCHED. | | + PROCESS| 08| 09| 10| 11| 12| 13| | 14| 15| + -------+-------+-------+-------+-------+-------+-------+ +-------+-------+ + MEMORY | | | | | | | |SWAP | | + | 16| 17| 18| 19| 20| 21| | 22| 23| + -------+-------+-------+-------+-------+-------+-------+ +-------+-------+ + NETWORK| |VNET |NETALT |NETMIG |NETCTL | | |SERIAL | | + | 24| 25| 26| 27| 28| 29| | 30| 31| + -------+-------+-------+-------+-------+-------+-------+ +-------+-------+ + DISK | | | | |DLIMIT | | |INODE | | + VFS | 32| 33| 34| 35| 36| 37| | 38| 39| + -------+-------+-------+-------+-------+-------+-------+ +-------+-------+ + OTHER |VSTAT | | | | | | |VINFO | | + | 40| 41| 42| 43| 44| 45| | 46| 47| + =======+=======+=======+=======+=======+=======+=======+ +=======+=======+ + SPECIAL|EVENT | | | |FLAGS | | | | | + | 48| 49| 50| 51| 52| 53| | 54| 55| + -------+-------+-------+-------+-------+-------+-------+ +-------+-------+ + SPECIAL|DEBUG | | | |RLIMIT |SYSCALL| | |COMPAT | + | 56| 57| 58| 59| 60|TEST 61| | 62| 63| + -------+-------+-------+-------+-------+-------+-------+ +-------+-------+ + +*/ + +#define VC_CAT_VERSION 0 + +#define VC_CAT_VSETUP 1 +#define VC_CAT_VHOST 2 + +#define VC_CAT_VPROC 9 +#define VC_CAT_PROCALT 10 +#define VC_CAT_PROCMIG 11 +#define VC_CAT_PROCTRL 12 + +#define VC_CAT_SCHED 14 + +#define VC_CAT_VNET 25 +#define VC_CAT_NETALT 26 +#define VC_CAT_NETMIG 27 +#define VC_CAT_NETCTRL 28 + +#define VC_CAT_DLIMIT 36 +#define VC_CAT_INODE 38 + +#define VC_CAT_VSTAT 40 +#define VC_CAT_VINFO 46 +#define VC_CAT_EVENT 48 + +#define VC_CAT_FLAGS 52 +#define VC_CAT_DEBUG 56 +#define VC_CAT_RLIMIT 60 + +#define VC_CAT_SYSTEST 61 +#define VC_CAT_COMPAT 63 + +/* query version */ + +#define VCMD_get_version VC_CMD(VERSION, 0, 0) +#define VCMD_get_vci VC_CMD(VERSION, 1, 0) + + +#ifdef __KERNEL__ + +#include + + +#else /* __KERNEL__ */ +#define __user +#endif /* __KERNEL__ */ + +#endif /* _VX_SWITCH_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vserver/tag.h linux-source-2.6.22/include/linux/vserver/tag.h --- linux-source-2.6.22.orig/include/linux/vserver/tag.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vserver/tag.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,140 @@ +#ifndef _DX_TAG_H +#define _DX_TAG_H + + +#define DX_TAG(in) (IS_TAGGED(in)) + + +#ifdef CONFIG_DX_TAG_NFSD +#define DX_TAG_NFSD 1 +#else +#define DX_TAG_NFSD 0 +#endif + + +#ifdef CONFIG_TAGGING_NONE + +#define MAX_UID 0xFFFFFFFF +#define MAX_GID 0xFFFFFFFF + +#define INOTAG_TAG(cond, uid, gid, tag) (0) + +#define TAGINO_UID(cond, uid, tag) (uid) +#define TAGINO_GID(cond, gid, tag) (gid) + +#endif + + +#ifdef CONFIG_TAGGING_GID16 + +#define MAX_UID 0xFFFFFFFF +#define MAX_GID 0x0000FFFF + +#define INOTAG_TAG(cond, uid, gid, tag) \ + ((cond) ? (((gid) >> 16) & 0xFFFF) : 0) + +#define TAGINO_UID(cond, uid, tag) (uid) +#define TAGINO_GID(cond, gid, tag) \ + ((cond) ? (((gid) & 0xFFFF) | ((tag) << 16)) : (gid)) + +#endif + + +#ifdef CONFIG_TAGGING_ID24 + +#define MAX_UID 0x00FFFFFF +#define MAX_GID 0x00FFFFFF + +#define INOTAG_TAG(cond, uid, gid, tag) \ + ((cond) ? ((((uid) >> 16) & 0xFF00) | (((gid) >> 24) & 0xFF)) : 0) + +#define TAGINO_UID(cond, uid, tag) \ + ((cond) ? (((uid) & 0xFFFFFF) | (((tag) & 0xFF00) << 16)) : (uid)) +#define TAGINO_GID(cond, gid, tag) \ + ((cond) ? (((gid) & 0xFFFFFF) | (((tag) & 0x00FF) << 24)) : (gid)) + +#endif + + +#ifdef CONFIG_TAGGING_UID16 + +#define MAX_UID 0x0000FFFF +#define MAX_GID 0xFFFFFFFF + +#define INOTAG_TAG(cond, uid, gid, tag) \ + ((cond) ? (((uid) >> 16) & 0xFFFF) : 0) + +#define TAGINO_UID(cond, uid, tag) \ + ((cond) ? (((uid) & 0xFFFF) | ((tag) << 16)) : (uid)) +#define TAGINO_GID(cond, gid, tag) (gid) + +#endif + + +#ifdef CONFIG_TAGGING_INTERN + +#define MAX_UID 0xFFFFFFFF +#define MAX_GID 0xFFFFFFFF + +#define INOTAG_TAG(cond, uid, gid, tag) \ + ((cond) ? (tag) : 0) + +#define TAGINO_UID(cond, uid, tag) (uid) +#define TAGINO_GID(cond, gid, tag) (gid) + +#endif + + +#ifndef CONFIG_TAGGING_NONE +#define dx_current_fstag(sb) \ + ((sb)->s_flags & MS_TAGGED ? dx_current_tag() : 0) +#else +#define dx_current_fstag(sb) (0) +#endif + +#ifndef CONFIG_TAGGING_INTERN +#define TAGINO_TAG(cond, tag) (0) +#else +#define TAGINO_TAG(cond, tag) ((cond) ? (tag) : 0) +#endif + +#define INOTAG_UID(cond, uid, gid) \ + ((cond) ? ((uid) & MAX_UID) : (uid)) +#define INOTAG_GID(cond, uid, gid) \ + ((cond) ? ((gid) & MAX_GID) : (gid)) + + +static inline uid_t dx_map_uid(uid_t uid) +{ + if ((uid > MAX_UID) && (uid != -1)) + uid = -2; + return (uid & MAX_UID); +} + +static inline gid_t dx_map_gid(gid_t gid) +{ + if ((gid > MAX_GID) && (gid != -1)) + gid = -2; + return (gid & MAX_GID); +} + + +#ifdef CONFIG_VSERVER_LEGACY +#define FIOC_GETTAG _IOR('x', 1, long) +#define FIOC_SETTAG _IOW('x', 2, long) +#define FIOC_SETTAGJ _IOW('x', 3, long) +#endif + +#ifdef CONFIG_PROPAGATE + +int dx_parse_tag(char *string, tag_t *tag, int remove); + +void __dx_propagate_tag(struct nameidata *nd, struct inode *inode); + +#define dx_propagate_tag(n, i) __dx_propagate_tag(n, i) + +#else +#define dx_propagate_tag(n, i) do { } while (0) +#endif + +#endif /* _DX_TAG_H */ diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vs_limit.h linux-source-2.6.22/include/linux/vs_limit.h --- linux-source-2.6.22.orig/include/linux/vs_limit.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vs_limit.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,140 @@ +#ifndef _VS_LIMIT_H +#define _VS_LIMIT_H + +#include "vserver/limit.h" +#include "vserver/base.h" +#include "vserver/context.h" +#include "vserver/debug.h" +#include "vserver/context.h" +#include "vserver/limit_int.h" + + +#define vx_acc_cres(v, d, p, r) \ + __vx_acc_cres(v, r, d, p, __FILE__, __LINE__) + +#define vx_acc_cres_cond(x, d, p, r) \ + __vx_acc_cres(((x) == vx_current_xid()) ? current->vx_info : 0, \ + r, d, p, __FILE__, __LINE__) + + +#define vx_add_cres(v, a, p, r) \ + __vx_add_cres(v, r, a, p, __FILE__, __LINE__) +#define vx_sub_cres(v, a, p, r) vx_add_cres(v, -(a), p, r) + +#define vx_add_cres_cond(x, a, p, r) \ + __vx_add_cres(((x) == vx_current_xid()) ? current->vx_info : 0, \ + r, a, p, __FILE__, __LINE__) +#define vx_sub_cres_cond(x, a, p, r) vx_add_cres_cond(x, -(a), p, r) + + +/* process and file limits */ + +#define vx_nproc_inc(p) \ + vx_acc_cres((p)->vx_info, 1, p, RLIMIT_NPROC) + +#define vx_nproc_dec(p) \ + vx_acc_cres((p)->vx_info,-1, p, RLIMIT_NPROC) + +#define vx_files_inc(f) \ + vx_acc_cres_cond((f)->f_xid, 1, f, RLIMIT_NOFILE) + +#define vx_files_dec(f) \ + vx_acc_cres_cond((f)->f_xid,-1, f, RLIMIT_NOFILE) + +#define vx_locks_inc(l) \ + vx_acc_cres_cond((l)->fl_xid, 1, l, RLIMIT_LOCKS) + +#define vx_locks_dec(l) \ + vx_acc_cres_cond((l)->fl_xid,-1, l, RLIMIT_LOCKS) + +#define vx_openfd_inc(f) \ + vx_acc_cres(current->vx_info, 1, (void *)(long)(f), VLIMIT_OPENFD) + +#define vx_openfd_dec(f) \ + vx_acc_cres(current->vx_info,-1, (void *)(long)(f), VLIMIT_OPENFD) + + +#define vx_cres_avail(v, n, r) \ + __vx_cres_avail(v, r, n, __FILE__, __LINE__) + + +#define vx_nproc_avail(n) \ + vx_cres_avail(current->vx_info, n, RLIMIT_NPROC) + +#define vx_files_avail(n) \ + vx_cres_avail(current->vx_info, n, RLIMIT_NOFILE) + +#define vx_locks_avail(n) \ + vx_cres_avail(current->vx_info, n, RLIMIT_LOCKS) + +#define vx_openfd_avail(n) \ + vx_cres_avail(current->vx_info, n, VLIMIT_OPENFD) + + +/* dentry limits */ + +#define vx_dentry_inc(d) do { \ + if (atomic_read(&d->d_count) == 1) \ + vx_acc_cres(current->vx_info, 1, d, VLIMIT_DENTRY); \ + } while (0) + +#define vx_dentry_dec(d) do { \ + if (atomic_read(&d->d_count) == 0) \ + vx_acc_cres(current->vx_info,-1, d, VLIMIT_DENTRY); \ + } while (0) + +#define vx_dentry_avail(n) \ + vx_cres_avail(current->vx_info, n, VLIMIT_DENTRY) + + +/* socket limits */ + +#define vx_sock_inc(s) \ + vx_acc_cres((s)->sk_vx_info, 1, s, VLIMIT_NSOCK) + +#define vx_sock_dec(s) \ + vx_acc_cres((s)->sk_vx_info,-1, s, VLIMIT_NSOCK) + +#define vx_sock_avail(n) \ + vx_cres_avail(current->vx_info, n, VLIMIT_NSOCK) + + +/* ipc resource limits */ + +#define vx_ipcmsg_add(v, u, a) \ + vx_add_cres(v, a, u, RLIMIT_MSGQUEUE) + +#define vx_ipcmsg_sub(v, u, a) \ + vx_sub_cres(v, a, u, RLIMIT_MSGQUEUE) + +#define vx_ipcmsg_avail(v, a) \ + vx_cres_avail(v, a, RLIMIT_MSGQUEUE) + + +#define vx_ipcshm_add(v, k, a) \ + vx_add_cres(v, a, (void *)(long)(k), VLIMIT_SHMEM) + +#define vx_ipcshm_sub(v, k, a) \ + vx_sub_cres(v, a, (void *)(long)(k), VLIMIT_SHMEM) + +#define vx_ipcshm_avail(v, a) \ + vx_cres_avail(v, a, VLIMIT_SHMEM) + + +#define vx_semary_inc(a) \ + vx_acc_cres(current->vx_info, 1, a, VLIMIT_SEMARY) + +#define vx_semary_dec(a) \ + vx_acc_cres(current->vx_info, -1, a, VLIMIT_SEMARY) + + +#define vx_nsems_add(a,n) \ + vx_add_cres(current->vx_info, n, a, VLIMIT_NSEMS) + +#define vx_nsems_sub(a,n) \ + vx_sub_cres(current->vx_info, n, a, VLIMIT_NSEMS) + + +#else +#warning duplicate inclusion +#endif diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vs_memory.h linux-source-2.6.22/include/linux/vs_memory.h --- linux-source-2.6.22.orig/include/linux/vs_memory.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vs_memory.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,159 @@ +#ifndef _VS_MEMORY_H +#define _VS_MEMORY_H + +#include "vserver/limit.h" +#include "vserver/base.h" +#include "vserver/context.h" +#include "vserver/debug.h" +#include "vserver/context.h" +#include "vserver/limit_int.h" + + +#define __acc_add_long(a, v) (*(v) += (a)) +#define __acc_inc_long(v) (++*(v)) +#define __acc_dec_long(v) (--*(v)) + +#if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS +#define __acc_add_atomic(a, v) atomic_long_add(a, v) +#define __acc_inc_atomic(v) atomic_long_inc(v) +#define __acc_dec_atomic(v) atomic_long_dec(v) +#else /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */ +#define __acc_add_atomic(a, v) __acc_add_long(a, v) +#define __acc_inc_atomic(v) __acc_inc_long(v) +#define __acc_dec_atomic(v) __acc_dec_long(v) +#endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */ + + +#define vx_acc_page(m, d, v, r) do { \ + if ((d) > 0) \ + __acc_inc_long(&(m)->v); \ + else \ + __acc_dec_long(&(m)->v); \ + __vx_acc_cres(m->mm_vx_info, r, d, m, __FILE__, __LINE__); \ +} while (0) + +#define vx_acc_page_atomic(m, d, v, r) do { \ + if ((d) > 0) \ + __acc_inc_atomic(&(m)->v); \ + else \ + __acc_dec_atomic(&(m)->v); \ + __vx_acc_cres(m->mm_vx_info, r, d, m, __FILE__, __LINE__); \ +} while (0) + + +#define vx_acc_pages(m, p, v, r) do { \ + unsigned long __p = (p); \ + __acc_add_long(__p, &(m)->v); \ + __vx_add_cres(m->mm_vx_info, r, __p, m, __FILE__, __LINE__); \ +} while (0) + +#define vx_acc_pages_atomic(m, p, v, r) do { \ + unsigned long __p = (p); \ + __acc_add_atomic(__p, &(m)->v); \ + __vx_add_cres(m->mm_vx_info, r, __p, m, __FILE__, __LINE__); \ +} while (0) + + + +#define vx_acc_vmpage(m, d) \ + vx_acc_page(m, d, total_vm, RLIMIT_AS) +#define vx_acc_vmlpage(m, d) \ + vx_acc_page(m, d, locked_vm, RLIMIT_MEMLOCK) +#define vx_acc_file_rsspage(m, d) \ + vx_acc_page_atomic(m, d, _file_rss, VLIMIT_MAPPED) +#define vx_acc_anon_rsspage(m, d) \ + vx_acc_page_atomic(m, d, _anon_rss, VLIMIT_ANON) + +#define vx_acc_vmpages(m, p) \ + vx_acc_pages(m, p, total_vm, RLIMIT_AS) +#define vx_acc_vmlpages(m, p) \ + vx_acc_pages(m, p, locked_vm, RLIMIT_MEMLOCK) +#define vx_acc_file_rsspages(m, p) \ + vx_acc_pages_atomic(m, p, _file_rss, VLIMIT_MAPPED) +#define vx_acc_anon_rsspages(m, p) \ + vx_acc_pages_atomic(m, p, _anon_rss, VLIMIT_ANON) + +#define vx_pages_add(s, r, p) __vx_add_cres(s, r, p, 0, __FILE__, __LINE__) +#define vx_pages_sub(s, r, p) vx_pages_add(s, r, -(p)) + +#define vx_vmpages_inc(m) vx_acc_vmpage(m, 1) +#define vx_vmpages_dec(m) vx_acc_vmpage(m, -1) +#define vx_vmpages_add(m, p) vx_acc_vmpages(m, p) +#define vx_vmpages_sub(m, p) vx_acc_vmpages(m, -(p)) + +#define vx_vmlocked_inc(m) vx_acc_vmlpage(m, 1) +#define vx_vmlocked_dec(m) vx_acc_vmlpage(m, -1) +#define vx_vmlocked_add(m, p) vx_acc_vmlpages(m, p) +#define vx_vmlocked_sub(m, p) vx_acc_vmlpages(m, -(p)) + +#define vx_file_rsspages_inc(m) vx_acc_file_rsspage(m, 1) +#define vx_file_rsspages_dec(m) vx_acc_file_rsspage(m, -1) +#define vx_file_rsspages_add(m, p) vx_acc_file_rsspages(m, p) +#define vx_file_rsspages_sub(m, p) vx_acc_file_rsspages(m, -(p)) + +#define vx_anon_rsspages_inc(m) vx_acc_anon_rsspage(m, 1) +#define vx_anon_rsspages_dec(m) vx_acc_anon_rsspage(m, -1) +#define vx_anon_rsspages_add(m, p) vx_acc_anon_rsspages(m, p) +#define vx_anon_rsspages_sub(m, p) vx_acc_anon_rsspages(m, -(p)) + + +#define vx_pages_avail(m, p, r) \ + __vx_cres_avail((m)->mm_vx_info, r, p, __FILE__, __LINE__) + +#define vx_vmpages_avail(m, p) vx_pages_avail(m, p, RLIMIT_AS) +#define vx_vmlocked_avail(m, p) vx_pages_avail(m, p, RLIMIT_MEMLOCK) +#define vx_anon_avail(m, p) vx_pages_avail(m, p, VLIMIT_ANON) +#define vx_mapped_avail(m, p) vx_pages_avail(m, p, VLIMIT_MAPPED) + +#define vx_rss_avail(m, p) \ + __vx_cres_array_avail((m)->mm_vx_info, VLA_RSS, p, __FILE__, __LINE__) + + +enum { + VXPT_UNKNOWN = 0, + VXPT_ANON, + VXPT_NONE, + VXPT_FILE, + VXPT_SWAP, + VXPT_WRITE +}; + +#if 0 +#define vx_page_fault(mm, vma, type, ret) +#else + +static inline +void __vx_page_fault(struct mm_struct *mm, + struct vm_area_struct *vma, int type, int ret) +{ + struct vx_info *vxi = mm->mm_vx_info; + int what; +/* + static char *page_type[6] = + { "UNKNOWN", "ANON", "NONE", "FILE", "SWAP", "WRITE" }; + static char *page_what[4] = + { "FAULT_OOM", "FAULT_SIGBUS", "FAULT_MINOR", "FAULT_MAJOR" }; +*/ + + if (!vxi) + return; + + what = (ret & 0x3); + +/* printk("[%d] page[%d][%d] %2x %s %s\n", vxi->vx_id, + type, what, ret, page_type[type], page_what[what]); +*/ + if (ret & VM_FAULT_WRITE) + what |= 0x4; + atomic_inc(&vxi->cacct.page[type][what]); +} + +#define vx_page_fault(mm, vma, type, ret) __vx_page_fault(mm, vma, type, ret) +#endif + + +extern unsigned long vx_badness(struct task_struct *task, struct mm_struct *mm); + +#else +#warning duplicate inclusion +#endif diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vs_network.h linux-source-2.6.22/include/linux/vs_network.h --- linux-source-2.6.22.orig/include/linux/vs_network.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vs_network.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,185 @@ +#ifndef _NX_VS_NETWORK_H +#define _NX_VS_NETWORK_H + +#include "vserver/context.h" +#include "vserver/network.h" +#include "vserver/base.h" +#include "vserver/debug.h" + + +#define get_nx_info(i) __get_nx_info(i, __FILE__, __LINE__) + +static inline struct nx_info *__get_nx_info(struct nx_info *nxi, + const char *_file, int _line) +{ + if (!nxi) + return NULL; + + vxlprintk(VXD_CBIT(nid, 2), "get_nx_info(%p[#%d.%d])", + nxi, nxi ? nxi->nx_id : 0, + nxi ? atomic_read(&nxi->nx_usecnt) : 0, + _file, _line); + + atomic_inc(&nxi->nx_usecnt); + return nxi; +} + + +extern void free_nx_info(struct nx_info *); + +#define put_nx_info(i) __put_nx_info(i, __FILE__, __LINE__) + +static inline void __put_nx_info(struct nx_info *nxi, const char *_file, int _line) +{ + if (!nxi) + return; + + vxlprintk(VXD_CBIT(nid, 2), "put_nx_info(%p[#%d.%d])", + nxi, nxi ? nxi->nx_id : 0, + nxi ? atomic_read(&nxi->nx_usecnt) : 0, + _file, _line); + + if (atomic_dec_and_test(&nxi->nx_usecnt)) + free_nx_info(nxi); +} + + +#define init_nx_info(p, i) __init_nx_info(p, i, __FILE__, __LINE__) + +static inline void __init_nx_info(struct nx_info **nxp, struct nx_info *nxi, + const char *_file, int _line) +{ + if (nxi) { + vxlprintk(VXD_CBIT(nid, 3), + "init_nx_info(%p[#%d.%d])", + nxi, nxi ? nxi->nx_id : 0, + nxi ? atomic_read(&nxi->nx_usecnt) : 0, + _file, _line); + + atomic_inc(&nxi->nx_usecnt); + } + *nxp = nxi; +} + + +#define set_nx_info(p, i) __set_nx_info(p, i, __FILE__, __LINE__) + +static inline void __set_nx_info(struct nx_info **nxp, struct nx_info *nxi, + const char *_file, int _line) +{ + struct nx_info *nxo; + + if (!nxi) + return; + + vxlprintk(VXD_CBIT(nid, 3), "set_nx_info(%p[#%d.%d])", + nxi, nxi ? nxi->nx_id : 0, + nxi ? atomic_read(&nxi->nx_usecnt) : 0, + _file, _line); + + atomic_inc(&nxi->nx_usecnt); + nxo = xchg(nxp, nxi); + BUG_ON(nxo); +} + +#define clr_nx_info(p) __clr_nx_info(p, __FILE__, __LINE__) + +static inline void __clr_nx_info(struct nx_info **nxp, + const char *_file, int _line) +{ + struct nx_info *nxo; + + nxo = xchg(nxp, NULL); + if (!nxo) + return; + + vxlprintk(VXD_CBIT(nid, 3), "clr_nx_info(%p[#%d.%d])", + nxo, nxo ? nxo->nx_id : 0, + nxo ? atomic_read(&nxo->nx_usecnt) : 0, + _file, _line); + + if (atomic_dec_and_test(&nxo->nx_usecnt)) + free_nx_info(nxo); +} + + +#define claim_nx_info(v, p) __claim_nx_info(v, p, __FILE__, __LINE__) + +static inline void __claim_nx_info(struct nx_info *nxi, + struct task_struct *task, const char *_file, int _line) +{ + vxlprintk(VXD_CBIT(nid, 3), "claim_nx_info(%p[#%d.%d.%d]) %p", + nxi, nxi ? nxi->nx_id : 0, + nxi?atomic_read(&nxi->nx_usecnt):0, + nxi?atomic_read(&nxi->nx_tasks):0, + task, _file, _line); + + atomic_inc(&nxi->nx_tasks); +} + + +extern void unhash_nx_info(struct nx_info *); + +#define release_nx_info(v, p) __release_nx_info(v, p, __FILE__, __LINE__) + +static inline void __release_nx_info(struct nx_info *nxi, + struct task_struct *task, const char *_file, int _line) +{ + vxlprintk(VXD_CBIT(nid, 3), "release_nx_info(%p[#%d.%d.%d]) %p", + nxi, nxi ? nxi->nx_id : 0, + nxi ? atomic_read(&nxi->nx_usecnt) : 0, + nxi ? atomic_read(&nxi->nx_tasks) : 0, + task, _file, _line); + + might_sleep(); + + if (atomic_dec_and_test(&nxi->nx_tasks)) + unhash_nx_info(nxi); +} + + +#define task_get_nx_info(i) __task_get_nx_info(i, __FILE__, __LINE__) + +static __inline__ struct nx_info *__task_get_nx_info(struct task_struct *p, + const char *_file, int _line) +{ + struct nx_info *nxi; + + task_lock(p); + vxlprintk(VXD_CBIT(nid, 5), "task_get_nx_info(%p)", + p, _file, _line); + nxi = __get_nx_info(p->nx_info, _file, _line); + task_unlock(p); + return nxi; +} + + + + +static inline int addr_in_nx_info(struct nx_info *nxi, uint32_t addr) +{ + int n, i; + + if (!nxi) + return 1; + + n = nxi->nbipv4; + if (n && (nxi->ipv4[0] == 0)) + return 1; + for (i = 0; i < n; i++) { + if (nxi->ipv4[i] == addr) + return 1; + } + return 0; +} + +static inline void exit_nx_info(struct task_struct *p) +{ + if (p->nx_info) + release_nx_info(p->nx_info, p); +} + + +#else +#warning duplicate inclusion +#endif diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vs_pid.h linux-source-2.6.22/include/linux/vs_pid.h --- linux-source-2.6.22.orig/include/linux/vs_pid.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vs_pid.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,133 @@ +#ifndef _VS_PID_H +#define _VS_PID_H + +#include "vserver/base.h" +#include "vserver/context.h" +#include "vserver/debug.h" +#include + + +/* pid faking stuff */ + + +#define vx_info_map_pid(v, p) \ + __vx_info_map_pid((v), (p), __FUNC__, __FILE__, __LINE__) +#define vx_info_map_tgid(v,p) vx_info_map_pid(v,p) +#define vx_map_pid(p) vx_info_map_pid(current->vx_info, p) +#define vx_map_tgid(p) vx_map_pid(p) + +static inline int __vx_info_map_pid(struct vx_info *vxi, int pid, + const char *func, const char *file, int line) +{ + if (vx_info_flags(vxi, VXF_INFO_INIT, 0)) { + vxfprintk(VXD_CBIT(cvirt, 2), + "vx_map_tgid: %p/%llx: %d -> %d", + vxi, (long long)vxi->vx_flags, pid, + (pid && pid == vxi->vx_initpid) ? 1 : pid, + func, file, line); + if (pid == 0) + return 0; + if (pid == vxi->vx_initpid) + return 1; + } + return pid; +} + +#define vx_info_rmap_pid(v, p) \ + __vx_info_rmap_pid((v), (p), __FUNC__, __FILE__, __LINE__) +#define vx_rmap_pid(p) vx_info_rmap_pid(current->vx_info, p) +#define vx_rmap_tgid(p) vx_rmap_pid(p) + +static inline int __vx_info_rmap_pid(struct vx_info *vxi, int pid, + const char *func, const char *file, int line) +{ + if (vx_info_flags(vxi, VXF_INFO_INIT, 0)) { + vxfprintk(VXD_CBIT(cvirt, 2), + "vx_rmap_tgid: %p/%llx: %d -> %d", + vxi, (long long)vxi->vx_flags, pid, + (pid == 1) ? vxi->vx_initpid : pid, + func, file, line); + if ((pid == 1) && vxi->vx_initpid) + return vxi->vx_initpid; + if (pid == vxi->vx_initpid) + return ~0U; + } + return pid; +} + + +#define VXF_FAKE_INIT (VXF_INFO_INIT | VXF_STATE_INIT) + +static inline +int vx_proc_task_visible(struct task_struct *task) +{ + if ((task->pid == 1) && + !vx_flags(VXF_FAKE_INIT, VXF_FAKE_INIT)) + /* show a blend through init */ + goto visible; + if (vx_check(vx_task_xid(task), VS_WATCH | VS_IDENT)) + goto visible; + return 0; +visible: + return 1; +} + +static inline +struct task_struct *vx_find_proc_task_by_pid(int pid) +{ + struct task_struct *task = find_task_by_real_pid(pid); + + if (task && !vx_proc_task_visible(task)) { + vxdprintk(VXD_CBIT(misc, 6), + "dropping task (find) %p[#%u,%u] for %p[#%u,%u]", + task, task->xid, task->pid, + current, current->xid, current->pid); + task = NULL; + } + return task; +} + +static inline +struct task_struct *vx_get_proc_task(struct inode *inode, struct pid *pid) +{ + struct task_struct *task = get_pid_task(pid, PIDTYPE_PID); + + if (task && !vx_proc_task_visible(task)) { + vxdprintk(VXD_CBIT(misc, 6), + "dropping task (get) %p[#%u,%u] for %p[#%u,%u]", + task, task->xid, task->pid, + current, current->xid, current->pid); + put_task_struct(task); + task = NULL; + } + return task; +} + + +static inline +struct task_struct *vx_child_reaper(struct task_struct *p) +{ + struct vx_info *vxi = p->vx_info; + struct task_struct *reaper = child_reaper(p); + + if (!vxi) + goto out; + + BUG_ON(!p->vx_info->vx_reaper); + + /* child reaper for the guest reaper */ + if (vxi->vx_reaper == p) + goto out; + + reaper = vxi->vx_reaper; +out: + vxdprintk(VXD_CBIT(xid, 7), + "vx_child_reaper(%p[#%u,%u]) = %p[#%u,%u]", + p, p->xid, p->pid, reaper, reaper->xid, reaper->pid); + return reaper; +} + + +#else +#warning duplicate inclusion +#endif diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vs_sched.h linux-source-2.6.22/include/linux/vs_sched.h --- linux-source-2.6.22.orig/include/linux/vs_sched.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vs_sched.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,110 @@ +#ifndef _VS_SCHED_H +#define _VS_SCHED_H + +#include "vserver/base.h" +#include "vserver/context.h" +#include "vserver/sched.h" + + +#define VAVAVOOM_RATIO 50 + +#define MAX_PRIO_BIAS 20 +#define MIN_PRIO_BIAS -20 + + +#ifdef CONFIG_VSERVER_HARDCPU + +/* + * effective_prio - return the priority that is based on the static + * priority but is modified by bonuses/penalties. + * + * We scale the actual sleep average [0 .... MAX_SLEEP_AVG] + * into a -4 ... 0 ... +4 bonus/penalty range. + * + * Additionally, we scale another amount based on the number of + * CPU tokens currently held by the context, if the process is + * part of a context (and the appropriate SCHED flag is set). + * This ranges from -5 ... 0 ... +15, quadratically. + * + * So, the total bonus is -9 .. 0 .. +19 + * We use ~50% of the full 0...39 priority range so that: + * + * 1) nice +19 interactive tasks do not preempt nice 0 CPU hogs. + * 2) nice -20 CPU hogs do not get preempted by nice 0 tasks. + * unless that context is far exceeding its CPU allocation. + * + * Both properties are important to certain workloads. + */ +static inline +int vx_effective_vavavoom(struct _vx_sched_pc *sched_pc, int max_prio) +{ + int vavavoom, max; + + /* lots of tokens = lots of vavavoom + * no tokens = no vavavoom */ + if ((vavavoom = sched_pc->tokens) >= 0) { + max = sched_pc->tokens_max; + vavavoom = max - vavavoom; + max = max * max; + vavavoom = max_prio * VAVAVOOM_RATIO / 100 + * (vavavoom*vavavoom - (max >> 2)) / max; + return vavavoom; + } + return 0; +} + + +static inline +int vx_adjust_prio(struct task_struct *p, int prio, int max_user) +{ + struct vx_info *vxi = p->vx_info; + struct _vx_sched_pc *sched_pc; + + if (!vxi) + return prio; + + sched_pc = &vx_cpu(vxi, sched_pc); + if (vx_info_flags(vxi, VXF_SCHED_PRIO, 0)) { + int vavavoom = vx_effective_vavavoom(sched_pc, max_user); + + sched_pc->vavavoom = vavavoom; + prio += vavavoom; + } + prio += sched_pc->prio_bias; + return prio; +} + +#else /* !CONFIG_VSERVER_HARDCPU */ + +static inline +int vx_adjust_prio(struct task_struct *p, int prio, int max_user) +{ + struct vx_info *vxi = p->vx_info; + + if (vxi) + prio += vx_cpu(vxi, sched_pc).prio_bias; + return prio; +} + +#endif /* CONFIG_VSERVER_HARDCPU */ + + +static inline void vx_account_user(struct vx_info *vxi, + cputime_t cputime, int nice) +{ + if (!vxi) + return; + vx_cpu(vxi, sched_pc).user_ticks += cputime; +} + +static inline void vx_account_system(struct vx_info *vxi, + cputime_t cputime, int idle) +{ + if (!vxi) + return; + vx_cpu(vxi, sched_pc).sys_ticks += cputime; +} + +#else +#warning duplicate inclusion +#endif diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vs_socket.h linux-source-2.6.22/include/linux/vs_socket.h --- linux-source-2.6.22.orig/include/linux/vs_socket.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vs_socket.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,67 @@ +#ifndef _VS_SOCKET_H +#define _VS_SOCKET_H + +#include "vserver/debug.h" +#include "vserver/base.h" +#include "vserver/cacct.h" +#include "vserver/context.h" + + +/* socket accounting */ + +#include + +static inline int vx_sock_type(int family) +{ + switch (family) { + case PF_UNSPEC: + return VXA_SOCK_UNSPEC; + case PF_UNIX: + return VXA_SOCK_UNIX; + case PF_INET: + return VXA_SOCK_INET; + case PF_INET6: + return VXA_SOCK_INET6; + case PF_PACKET: + return VXA_SOCK_PACKET; + default: + return VXA_SOCK_OTHER; + } +} + +#define vx_acc_sock(v, f, p, s) \ + __vx_acc_sock(v, f, p, s, __FILE__, __LINE__) + +static inline void __vx_acc_sock(struct vx_info *vxi, + int family, int pos, int size, char *file, int line) +{ + if (vxi) { + int type = vx_sock_type(family); + + atomic_long_inc(&vxi->cacct.sock[type][pos].count); + atomic_long_add(size, &vxi->cacct.sock[type][pos].total); + } +} + +#define vx_sock_recv(sk, s) \ + vx_acc_sock((sk)->sk_vx_info, (sk)->sk_family, 0, s) +#define vx_sock_send(sk, s) \ + vx_acc_sock((sk)->sk_vx_info, (sk)->sk_family, 1, s) +#define vx_sock_fail(sk, s) \ + vx_acc_sock((sk)->sk_vx_info, (sk)->sk_family, 2, s) + + +#define sock_vx_init(s) do { \ + (s)->sk_xid = 0; \ + (s)->sk_vx_info = NULL; \ + } while (0) + +#define sock_nx_init(s) do { \ + (s)->sk_nid = 0; \ + (s)->sk_nx_info = NULL; \ + } while (0) + + +#else +#warning duplicate inclusion +#endif diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vs_tag.h linux-source-2.6.22/include/linux/vs_tag.h --- linux-source-2.6.22.orig/include/linux/vs_tag.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vs_tag.h 2007-10-24 16:14:06.291582822 +0200 @@ -0,0 +1,43 @@ +#ifndef _VS_TAG_H +#define _VS_TAG_H + +#include + +/* check conditions */ + +#define DX_ADMIN 0x0001 +#define DX_WATCH 0x0002 +#define DX_HOSTID 0x0008 + +#define DX_IDENT 0x0010 + +#define DX_ARG_MASK 0x0010 + + +#define dx_task_tag(t) ((t)->xid) + +#define dx_current_tag() dx_task_tag(current) + +#define dx_check(c, m) __dx_check(dx_current_tag(), c, m) + +#define dx_weak_check(c, m) ((m) ? dx_check(c, m) : 1) + + +/* + * check current context for ADMIN/WATCH and + * optionally against supplied argument + */ +static inline int __dx_check(tag_t cid, tag_t id, unsigned int mode) +{ + if (mode & DX_ARG_MASK) { + if ((mode & DX_IDENT) && (id == cid)) + return 1; + } + return (((mode & DX_ADMIN) && (cid == 0)) || + ((mode & DX_WATCH) && (cid == 1)) || + ((mode & DX_HOSTID) && (id == 0))); +} + +#else +#warning duplicate inclusion +#endif diff -NurpP --minimal linux-source-2.6.22.orig/include/linux/vs_time.h linux-source-2.6.22/include/linux/vs_time.h --- linux-source-2.6.22.orig/include/linux/vs_time.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/include/linux/vs_time.h 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,19 @@ +#ifndef _VS_TIME_H +#define _VS_TIME_H + + +/* time faking stuff */ + +#ifdef CONFIG_VSERVER_VTIME + +extern void vx_gettimeofday(struct timeval *tv); +extern int vx_settimeofday(struct timespec *ts); + +#else +#define vx_gettimeofday(t) do_gettimeofday(t) +#define vx_settimeofday(t) do_settimeofday(t) +#endif + +#else +#warning duplicate inclusion +#endif diff -NurpP --minimal linux-source-2.6.22.orig/include/net/af_unix.h linux-source-2.6.22/include/net/af_unix.h --- linux-source-2.6.22.orig/include/net/af_unix.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/net/af_unix.h 2007-10-24 16:14:06.791620780 +0200 @@ -4,6 +4,7 @@ #include #include #include +#include #include extern void unix_inflight(struct file *fp); @@ -17,9 +18,9 @@ extern spinlock_t unix_table_lock; extern atomic_t unix_tot_inflight; -static inline struct sock *first_unix_socket(int *i) +static inline struct sock *next_unix_socket_table(int *i) { - for (*i = 0; *i <= UNIX_HASH_SIZE; (*i)++) { + for ((*i)++; *i <= UNIX_HASH_SIZE; (*i)++) { if (!hlist_empty(&unix_socket_table[*i])) return __sk_head(&unix_socket_table[*i]); } @@ -28,16 +29,19 @@ static inline struct sock *first_unix_so static inline struct sock *next_unix_socket(int *i, struct sock *s) { - struct sock *next = sk_next(s); - /* More in this chain? */ - if (next) - return next; - /* Look for next non-empty chain. */ - for ((*i)++; *i <= UNIX_HASH_SIZE; (*i)++) { - if (!hlist_empty(&unix_socket_table[*i])) - return __sk_head(&unix_socket_table[*i]); - } - return NULL; + do { + if (s) + s = sk_next(s); + if (!s) + s = next_unix_socket_table(i); + } while (s && !nx_check(s->sk_nid, VS_WATCH_P | VS_IDENT)); + return s; +} + +static inline struct sock *first_unix_socket(int *i) +{ + *i = 0; + return next_unix_socket(i, NULL); } #define forall_unix_sockets(i, s) \ diff -NurpP --minimal linux-source-2.6.22.orig/include/net/inet_hashtables.h linux-source-2.6.22/include/net/inet_hashtables.h --- linux-source-2.6.22.orig/include/net/inet_hashtables.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/net/inet_hashtables.h 2007-10-24 16:14:06.791620780 +0200 @@ -271,6 +271,26 @@ static inline int inet_iif(const struct return ((struct rtable *)skb->dst)->rt_iif; } +/* + * Check if a given address matches for an inet socket + * + * nxi: the socket's nx_info if any + * addr: to be verified address + * saddr: socket addresses + */ +static inline int inet_addr_match ( + struct nx_info *nxi, + uint32_t addr, + uint32_t saddr) +{ + if (addr && (saddr == addr)) + return 1; + if (!saddr) + return addr_in_nx_info(nxi, addr); + return 0; +} + + extern struct sock *__inet_lookup_listener(struct inet_hashinfo *hashinfo, const __be32 daddr, const unsigned short hnum, diff -NurpP --minimal linux-source-2.6.22.orig/include/net/inet_sock.h linux-source-2.6.22/include/net/inet_sock.h --- linux-source-2.6.22.orig/include/net/inet_sock.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/net/inet_sock.h 2007-10-24 16:14:06.791620780 +0200 @@ -113,6 +113,7 @@ struct inet_sock { /* Socket demultiplex comparisons on incoming packets. */ __be32 daddr; __be32 rcv_saddr; + __be32 rcv_saddr2; /* Second bound ipv4 addr, for ipv4root */ __be16 dport; __u16 num; __be32 saddr; diff -NurpP --minimal linux-source-2.6.22.orig/include/net/inet_timewait_sock.h linux-source-2.6.22/include/net/inet_timewait_sock.h --- linux-source-2.6.22.orig/include/net/inet_timewait_sock.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/net/inet_timewait_sock.h 2007-10-24 16:14:06.791620780 +0200 @@ -115,6 +115,10 @@ struct inet_timewait_sock { #define tw_refcnt __tw_common.skc_refcnt #define tw_hash __tw_common.skc_hash #define tw_prot __tw_common.skc_prot +#define tw_xid __tw_common.skc_xid +#define tw_vx_info __tw_common.skc_vx_info +#define tw_nid __tw_common.skc_nid +#define tw_nx_info __tw_common.skc_nx_info volatile unsigned char tw_substate; /* 3 bits hole, try to pack */ unsigned char tw_rcv_wscale; diff -NurpP --minimal linux-source-2.6.22.orig/include/net/route.h linux-source-2.6.22/include/net/route.h --- linux-source-2.6.22.orig/include/net/route.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/net/route.h 2007-10-24 16:14:06.791620780 +0200 @@ -27,12 +27,16 @@ #include #include #include +#include #include #include #include #include #include #include +#include +#include +#include #ifndef __KERNEL__ #warning This file is not supposed to be used outside of kernel. @@ -143,6 +147,59 @@ static inline char rt_tos2priority(u8 to return ip_tos2prio[IPTOS_TOS(tos)>>1]; } +#define IPI_LOOPBACK htonl(INADDR_LOOPBACK) + +static inline int ip_find_src(struct nx_info *nxi, struct rtable **rp, struct flowi *fl) +{ + int err; + int i, n = nxi->nbipv4; + u32 ipv4root = nxi->ipv4[0]; + + if (ipv4root == 0) + return 0; + + if (fl->fl4_src == 0) { + if (n > 1) { + u32 foundsrc; + + err = __ip_route_output_key(rp, fl); + if (err) { + fl->fl4_src = ipv4root; + err = __ip_route_output_key(rp, fl); + } + if (err) + return err; + + foundsrc = (*rp)->rt_src; + ip_rt_put(*rp); + + for (i=0; imask[i]; + u32 ipv4 = nxi->ipv4[i]; + u32 net4 = ipv4 & mask; + + if (foundsrc == ipv4) { + fl->fl4_src = ipv4; + break; + } + if (!fl->fl4_src && (foundsrc & mask) == net4) + fl->fl4_src = ipv4; + } + } + if (fl->fl4_src == 0) + fl->fl4_src = (fl->fl4_dst == IPI_LOOPBACK) + ? IPI_LOOPBACK : ipv4root; + } else { + for (i=0; iipv4[i] == fl->fl4_src) + break; + } + if (i == n) + return -EPERM; + } + return 0; +} + static inline int ip_route_connect(struct rtable **rp, __be32 dst, __be32 src, u32 tos, int oif, u8 protocol, __be16 sport, __be16 dport, struct sock *sk, @@ -158,7 +215,27 @@ static inline int ip_route_connect(struc .dport = dport } } }; int err; - if (!dst || !src) { + struct nx_info *nx_info = current->nx_info; + + if (sk) + nx_info = sk->sk_nx_info; + vxdprintk(VXD_CBIT(net, 4), + "ip_route_connect(%p) %p,%p;%lx", + sk, nx_info, sk->sk_socket, + (sk->sk_socket?sk->sk_socket->flags:0)); + + if (nx_info) { + err = ip_find_src(nx_info, rp, &fl); + if (err) + return err; + if (fl.fl4_dst == IPI_LOOPBACK && !nx_check(0, VS_ADMIN)) + fl.fl4_dst = nx_info->ipv4[0]; +#ifdef CONFIG_VSERVER_REMAP_SADDR + if (fl.fl4_src == IPI_LOOPBACK && !nx_check(0, VS_ADMIN)) + fl.fl4_src = nx_info->ipv4[0]; +#endif + } + if (!fl.fl4_dst || !fl.fl4_src) { err = __ip_route_output_key(rp, &fl); if (err) return err; diff -NurpP --minimal linux-source-2.6.22.orig/include/net/sock.h linux-source-2.6.22/include/net/sock.h --- linux-source-2.6.22.orig/include/net/sock.h 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/include/net/sock.h 2007-10-24 16:14:06.791620780 +0200 @@ -119,6 +119,10 @@ struct sock_common { atomic_t skc_refcnt; unsigned int skc_hash; struct proto *skc_prot; + xid_t skc_xid; + struct vx_info *skc_vx_info; + nid_t skc_nid; + struct nx_info *skc_nx_info; }; /** @@ -195,6 +199,10 @@ struct sock { #define sk_refcnt __sk_common.skc_refcnt #define sk_hash __sk_common.skc_hash #define sk_prot __sk_common.skc_prot +#define sk_xid __sk_common.skc_xid +#define sk_vx_info __sk_common.skc_vx_info +#define sk_nid __sk_common.skc_nid +#define sk_nx_info __sk_common.skc_nx_info unsigned char sk_shutdown : 2, sk_no_check : 2, sk_userlocks : 4; diff -NurpP --minimal linux-source-2.6.22.orig/init/main.c linux-source-2.6.22/init/main.c --- linux-source-2.6.22.orig/init/main.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/init/main.c 2007-10-24 16:14:06.791620780 +0200 @@ -55,6 +55,7 @@ #include #include #include +#include #include #include @@ -366,12 +367,14 @@ EXPORT_SYMBOL(__per_cpu_offset); static void __init setup_per_cpu_areas(void) { - unsigned long size, i; + unsigned long size, vspc, i; char *ptr; unsigned long nr_possible_cpus = num_possible_cpus(); + vspc = PERCPU_PERCTX * CONFIG_VSERVER_CONTEXTS; + /* Copy section for each CPU (we discard the original) */ - size = ALIGN(PERCPU_ENOUGH_ROOM, PAGE_SIZE); + size = ALIGN(PERCPU_ENOUGH_ROOM + vspc, PAGE_SIZE); ptr = alloc_bootmem_pages(size * nr_possible_cpus); for_each_possible_cpu(i) { diff -NurpP --minimal linux-source-2.6.22.orig/ipc/mqueue.c linux-source-2.6.22/ipc/mqueue.c --- linux-source-2.6.22.orig/ipc/mqueue.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/ipc/mqueue.c 2007-10-24 16:14:06.791620780 +0200 @@ -29,6 +29,8 @@ #include #include #include +#include +#include #include #include "util.h" @@ -151,17 +153,20 @@ static struct inode *mqueue_get_inode(st spin_lock(&mq_lock); if (u->mq_bytes + mq_bytes < u->mq_bytes || u->mq_bytes + mq_bytes > - p->signal->rlim[RLIMIT_MSGQUEUE].rlim_cur) { + p->signal->rlim[RLIMIT_MSGQUEUE].rlim_cur || + !vx_ipcmsg_avail(p->vx_info, mq_bytes)) { spin_unlock(&mq_lock); goto out_inode; } u->mq_bytes += mq_bytes; + vx_ipcmsg_add(p->vx_info, u, mq_bytes); spin_unlock(&mq_lock); info->messages = kmalloc(mq_msg_tblsz, GFP_KERNEL); if (!info->messages) { spin_lock(&mq_lock); u->mq_bytes -= mq_bytes; + vx_ipcmsg_sub(p->vx_info, u, mq_bytes); spin_unlock(&mq_lock); goto out_inode; } @@ -257,10 +262,14 @@ static void mqueue_delete_inode(struct i (info->attr.mq_maxmsg * info->attr.mq_msgsize)); user = info->user; if (user) { + struct vx_info *vxi = lookup_vx_info(user->xid); + spin_lock(&mq_lock); user->mq_bytes -= mq_bytes; + vx_ipcmsg_sub(vxi, user, mq_bytes); queues_count--; spin_unlock(&mq_lock); + put_vx_info(vxi); free_uid(user); } } @@ -748,7 +757,7 @@ asmlinkage long sys_mq_unlink(const char if (inode) atomic_inc(&inode->i_count); - err = vfs_unlink(dentry->d_parent->d_inode, dentry, mqueue_mnt); + err = vfs_unlink(dentry->d_parent->d_inode, dentry, mqueue_mnt, NULL); out_err: dput(dentry); diff -NurpP --minimal linux-source-2.6.22.orig/ipc/msg.c linux-source-2.6.22/ipc/msg.c --- linux-source-2.6.22.orig/ipc/msg.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/ipc/msg.c 2007-10-24 16:14:06.791620780 +0200 @@ -36,6 +36,7 @@ #include #include #include +#include #include #include @@ -149,6 +150,7 @@ static int newque (struct ipc_namespace msq->q_perm.mode = msgflg & S_IRWXUGO; msq->q_perm.key = key; + msq->q_perm.xid = vx_current_xid(); msq->q_perm.security = NULL; retval = security_msg_queue_alloc(msq); diff -NurpP --minimal linux-source-2.6.22.orig/ipc/sem.c linux-source-2.6.22/ipc/sem.c --- linux-source-2.6.22.orig/ipc/sem.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/ipc/sem.c 2007-10-24 16:14:06.791620780 +0200 @@ -82,6 +82,8 @@ #include #include #include +#include +#include #include #include "util.h" @@ -229,6 +231,7 @@ static int newary (struct ipc_namespace sma->sem_perm.mode = (semflg & S_IRWXUGO); sma->sem_perm.key = key; + sma->sem_perm.xid = vx_current_xid(); sma->sem_perm.security = NULL; retval = security_sem_alloc(sma); @@ -244,6 +247,9 @@ static int newary (struct ipc_namespace return -ENOSPC; } ns->used_sems += nsems; + /* FIXME: obsoleted? */ + vx_semary_inc(sma); + vx_nsems_add(sma, nsems); sma->sem_id = sem_buildid(ns, id, sma->sem_perm.seq); sma->sem_base = (struct sem *) &sma[1]; @@ -525,6 +531,9 @@ static void freeary (struct ipc_namespac sem_unlock(sma); ns->used_sems -= sma->sem_nsems; + /* FIXME: obsoleted? */ + vx_nsems_sub(sma, sma->sem_nsems); + vx_semary_dec(sma); size = sizeof (*sma) + sma->sem_nsems * sizeof (struct sem); security_sem_free(sma); ipc_rcu_putref(sma); diff -NurpP --minimal linux-source-2.6.22.orig/ipc/shm.c linux-source-2.6.22/ipc/shm.c --- linux-source-2.6.22.orig/ipc/shm.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/ipc/shm.c 2007-10-24 16:14:06.791620780 +0200 @@ -38,6 +38,8 @@ #include #include #include +#include +#include #include @@ -185,7 +187,12 @@ static void shm_open(struct vm_area_stru */ static void shm_destroy(struct ipc_namespace *ns, struct shmid_kernel *shp) { - ns->shm_tot -= (shp->shm_segsz + PAGE_SIZE - 1) >> PAGE_SHIFT; + struct vx_info *vxi = lookup_vx_info(shp->shm_perm.xid); + int numpages = (shp->shm_segsz + PAGE_SIZE - 1) >> PAGE_SHIFT; + + vx_ipcshm_sub(vxi, shp, numpages); + ns->shm_tot -= numpages; + shm_rmid(ns, shp->id); shm_unlock(shp); if (!is_file_hugepages(shp->shm_file)) @@ -195,6 +202,7 @@ static void shm_destroy(struct ipc_names shp->mlock_user); fput (shp->shm_file); security_shm_free(shp); + put_vx_info(vxi); ipc_rcu_putref(shp); } @@ -351,11 +359,15 @@ static int newseg (struct ipc_namespace if (ns->shm_tot + numpages > ns->shm_ctlall) return -ENOSPC; + if (!vx_ipcshm_avail(current->vx_info, numpages)) + return -ENOSPC; + shp = ipc_rcu_alloc(sizeof(*shp)); if (!shp) return -ENOMEM; shp->shm_perm.key = key; + shp->shm_perm.xid = vx_current_xid(); shp->shm_perm.mode = (shmflg & S_IRWXUGO); shp->mlock_user = NULL; @@ -406,6 +418,7 @@ static int newseg (struct ipc_namespace file->f_dentry->d_inode->i_ino = shp->id; ns->shm_tot += numpages; + vx_ipcshm_add(current->vx_info, key, numpages); shm_unlock(shp); return shp->id; diff -NurpP --minimal linux-source-2.6.22.orig/ipc/util.c linux-source-2.6.22/ipc/util.c --- linux-source-2.6.22.orig/ipc/util.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/ipc/util.c 2007-10-24 16:14:06.791620780 +0200 @@ -32,6 +32,8 @@ #include #include #include +#include +#include #include @@ -72,6 +74,7 @@ static struct ipc_namespace *clone_ipc_n goto err_shm; kref_init(&ns->kref); + atomic_inc(&vs_global_ipc_ns); return ns; err_shm: @@ -108,6 +111,7 @@ void free_ipc_ns(struct kref *kref) sem_exit_ns(ns); msg_exit_ns(ns); shm_exit_ns(ns); + atomic_dec(&vs_global_ipc_ns); kfree(ns); } #else diff -NurpP --minimal linux-source-2.6.22.orig/kernel/capability.c linux-source-2.6.22/kernel/capability.c --- linux-source-2.6.22.orig/kernel/capability.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/kernel/capability.c 2007-10-24 16:14:06.791620780 +0200 @@ -12,6 +12,7 @@ #include #include #include +#include #include unsigned securebits = SECUREBITS_DEFAULT; /* systemwide security settings */ @@ -103,6 +104,8 @@ static inline int cap_set_pg(int pgrp_nr pgrp = find_pid(pgrp_nr); do_each_pid_task(pgrp, PIDTYPE_PGID, g) { + if (!vx_check(g->xid, VS_ADMIN_P | VS_IDENT)) + continue; target = g; while_each_thread(g, target) { if (!security_capset_check(target, effective, @@ -246,8 +249,12 @@ int __capable(struct task_struct *t, int } EXPORT_SYMBOL(__capable); +#include int capable(int cap) { + /* here for now so we don't require task locking */ + if (vs_check_bit(VXC_CAP_MASK, cap) && !vx_mcaps(1L << cap)) + return 0; return __capable(current, cap); } EXPORT_SYMBOL(capable); diff -NurpP --minimal linux-source-2.6.22.orig/kernel/compat.c linux-source-2.6.22/kernel/compat.c --- linux-source-2.6.22.orig/kernel/compat.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/kernel/compat.c 2007-10-24 16:14:06.791620780 +0200 @@ -846,7 +846,7 @@ asmlinkage long compat_sys_time(compat_t compat_time_t i; struct timeval tv; - do_gettimeofday(&tv); + vx_gettimeofday(&tv); i = tv.tv_sec; if (tloc) { @@ -870,7 +870,7 @@ asmlinkage long compat_sys_stime(compat_ if (err) return err; - do_settimeofday(&tv); + vx_settimeofday(&tv); return 0; } diff -NurpP --minimal linux-source-2.6.22.orig/kernel/exit.c linux-source-2.6.22/kernel/exit.c --- linux-source-2.6.22.orig/kernel/exit.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/kernel/exit.c 2007-10-24 16:14:06.791620780 +0200 @@ -44,6 +44,11 @@ #include #include #include +#include +#include +#include +#include +#include #include #include @@ -443,9 +448,11 @@ static void close_files(struct files_str filp_close(file, files); cond_resched(); } + vx_openfd_dec(i); } i++; set >>= 1; + cond_resched(); } } } @@ -525,6 +532,7 @@ static inline void __put_fs_struct(struc dput(fs->altroot); mntput(fs->altrootmnt); } + atomic_dec(&vs_global_fs); kmem_cache_free(fs_cachep, fs); } } @@ -596,6 +604,14 @@ static void exit_mm(struct task_struct * static inline void choose_new_parent(struct task_struct *p, struct task_struct *reaper) { + /* check for reaper context */ + vxwprintk((p->xid != reaper->xid) && (reaper != child_reaper(p)), + "rogue reaper: %p[%d,#%u] <> %p[%d,#%u]", + p, p->pid, p->xid, reaper, reaper->pid, reaper->xid); + + if (p == reaper) + reaper = vx_child_reaper(p); + /* * Make sure we're not reparenting to ourselves and that * the parent is not a zombie. @@ -687,7 +703,7 @@ forget_original_parent(struct task_struc do { reaper = next_thread(reaper); if (reaper == father) { - reaper = child_reaper(father); + reaper = vx_child_reaper(father); break; } } while (reaper->exit_state); @@ -964,6 +980,8 @@ fastcall NORET_TYPE void do_exit(long co tsk->exit_code = code; proc_exit_connector(tsk); exit_task_namespaces(tsk); + /* needs to stay before exit_notify() */ + exit_vx_info_early(tsk, code); exit_notify(tsk); #ifdef CONFIG_NUMA mpol_free(tsk->mempolicy); @@ -994,6 +1012,10 @@ fastcall NORET_TYPE void do_exit(long co if (tsk->splice_pipe) __free_pipe_info(tsk->splice_pipe); + /* needs to stay after exit_notify() */ + exit_vx_info(tsk, code); + exit_nx_info(tsk); + preempt_disable(); /* causes final put_task_struct in finish_task_switch(). */ tsk->state = TASK_DEAD; diff -NurpP --minimal linux-source-2.6.22.orig/kernel/fork.c linux-source-2.6.22/kernel/fork.c --- linux-source-2.6.22.orig/kernel/fork.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/kernel/fork.c 2007-10-24 16:14:06.791620780 +0200 @@ -49,6 +49,11 @@ #include #include #include +#include +#include +#include +#include +#include #include #include @@ -108,6 +113,8 @@ void free_task(struct task_struct *tsk) { free_thread_info(tsk->stack); rt_mutex_debug_task_free(tsk); + clr_vx_info(&tsk->vx_info); + clr_nx_info(&tsk->nx_info); free_task_struct(tsk); } EXPORT_SYMBOL(free_task); @@ -215,6 +222,8 @@ static inline int dup_mmap(struct mm_str mm->free_area_cache = oldmm->mmap_base; mm->cached_hole_size = ~0UL; mm->map_count = 0; + __set_mm_counter(mm, file_rss, 0); + __set_mm_counter(mm, anon_rss, 0); cpus_clear(mm->cpu_vm_mask); mm->mm_rb = RB_ROOT; rb_link = &mm->mm_rb.rb_node; @@ -226,7 +235,7 @@ static inline int dup_mmap(struct mm_str if (mpnt->vm_flags & VM_DONTCOPY) { long pages = vma_pages(mpnt); - mm->total_vm -= pages; + vx_vmpages_sub(mm, pages); vm_stat_account(mm, mpnt->vm_flags, mpnt->vm_file, -pages); continue; @@ -335,8 +344,8 @@ static struct mm_struct * mm_init(struct INIT_LIST_HEAD(&mm->mmlist); mm->core_waiters = 0; mm->nr_ptes = 0; - set_mm_counter(mm, file_rss, 0); - set_mm_counter(mm, anon_rss, 0); + __set_mm_counter(mm, file_rss, 0); + __set_mm_counter(mm, anon_rss, 0); spin_lock_init(&mm->page_table_lock); rwlock_init(&mm->ioctx_list_lock); mm->ioctx_list = NULL; @@ -345,6 +354,7 @@ static struct mm_struct * mm_init(struct if (likely(!mm_alloc_pgd(mm))) { mm->def_flags = 0; + set_vx_info(&mm->mm_vx_info, current->vx_info); return mm; } free_mm(mm); @@ -376,6 +386,7 @@ void fastcall __mmdrop(struct mm_struct BUG_ON(mm == &init_mm); mm_free_pgd(mm); destroy_context(mm); + clr_vx_info(&mm->mm_vx_info); free_mm(mm); } @@ -490,6 +501,7 @@ static struct mm_struct *dup_mm(struct t goto fail_nomem; memcpy(mm, oldmm, sizeof(*mm)); + mm->mm_vx_info = NULL; /* Initializing for Swap token stuff */ mm->token_priority = 0; @@ -521,6 +533,7 @@ fail_nocontext: * If init_new_context() failed, we cannot use mmput() to free the mm * because it calls destroy_context() */ + clr_vx_info(&mm->mm_vx_info); mm_free_pgd(mm); free_mm(mm); return NULL; @@ -591,6 +604,7 @@ static inline struct fs_struct *__copy_f fs->altroot = NULL; } read_unlock(&old->lock); + atomic_inc(&vs_global_fs); } return fs; } @@ -709,6 +723,8 @@ static struct files_struct *dup_fd(struc struct file *f = *old_fds++; if (f) { get_file(f); + /* TODO: sum it first for check and performance */ + vx_openfd_inc(open_files - i); } else { /* * The fd may be claimed in the fd bitmap but not yet @@ -961,6 +977,8 @@ static struct task_struct *copy_process( { int retval; struct task_struct *p = NULL; + struct vx_info *vxi; + struct nx_info *nxi; if ((clone_flags & (CLONE_NEWNS|CLONE_FS)) == (CLONE_NEWNS|CLONE_FS)) return ERR_PTR(-EINVAL); @@ -995,12 +1013,30 @@ static struct task_struct *copy_process( DEBUG_LOCKS_WARN_ON(!p->hardirqs_enabled); DEBUG_LOCKS_WARN_ON(!p->softirqs_enabled); #endif + init_vx_info(&p->vx_info, current->vx_info); + init_nx_info(&p->nx_info, current->nx_info); + + /* check vserver memory */ + if (p->mm && !(clone_flags & CLONE_VM)) { + if (vx_vmpages_avail(p->mm, p->mm->total_vm)) + vx_pages_add(p->vx_info, RLIMIT_AS, p->mm->total_vm); + else + goto bad_fork_free; + } + if (p->mm && vx_flags(VXF_FORK_RSS, 0)) { + if (!vx_rss_avail(p->mm, get_mm_counter(p->mm, file_rss))) + goto bad_fork_cleanup_vm; + } + retval = -EAGAIN; + if (!vx_nproc_avail(1)) + goto bad_fork_cleanup_vm; + if (atomic_read(&p->user->processes) >= p->signal->rlim[RLIMIT_NPROC].rlim_cur) { if (!capable(CAP_SYS_ADMIN) && !capable(CAP_SYS_RESOURCE) && p->user != &root_user) - goto bad_fork_free; + goto bad_fork_cleanup_vm; } atomic_inc(&p->user->__count); @@ -1265,6 +1301,18 @@ static struct task_struct *copy_process( total_forks++; spin_unlock(¤t->sighand->siglock); + + /* p is copy of current */ + vxi = p->vx_info; + if (vxi) { + claim_vx_info(vxi, p); + atomic_inc(&vxi->cvirt.nr_threads); + atomic_inc(&vxi->cvirt.total_forks); + vx_nproc_inc(p); + } + nxi = p->nx_info; + if (nxi) + claim_nx_info(nxi, p); write_unlock_irq(&tasklist_lock); proc_fork_connector(p); return p; @@ -1306,6 +1354,9 @@ bad_fork_cleanup_count: put_group_info(p->group_info); atomic_dec(&p->user->processes); free_uid(p->user); +bad_fork_cleanup_vm: + if (p->mm && !(clone_flags & CLONE_VM)) + vx_pages_sub(p->vx_info, RLIMIT_AS, p->mm->total_vm); bad_fork_free: free_task(p); fork_out: @@ -1367,6 +1418,15 @@ long do_fork(unsigned long clone_flags, if (!pid) return -EAGAIN; + + /* kernel threads are host only */ + if ((clone_flags & CLONE_KTHREAD) && !vx_check(0, VS_ADMIN)) { + vxwprintk(1, "xid=%d tried to spawn a kernel thread.", + vx_current_xid()); + free_pid(pid); + return -EPERM; + } + nr = pid->nr; if (unlikely(current->ptrace)) { trace = fork_traceflag (clone_flags); diff -NurpP --minimal linux-source-2.6.22.orig/kernel/kthread.c linux-source-2.6.22/kernel/kthread.c --- linux-source-2.6.22.orig/kernel/kthread.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/kernel/kthread.c 2007-10-24 16:14:06.791620780 +0200 @@ -96,7 +96,7 @@ static void create_kthread(struct kthrea } else { wait_for_completion(&create->started); read_lock(&tasklist_lock); - create->result = find_task_by_pid(pid); + create->result = find_task_by_real_pid(pid); read_unlock(&tasklist_lock); } complete(&create->done); diff -NurpP --minimal linux-source-2.6.22.orig/kernel/Makefile linux-source-2.6.22/kernel/Makefile --- linux-source-2.6.22.orig/kernel/Makefile 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/kernel/Makefile 2007-10-24 16:14:06.791620780 +0200 @@ -10,6 +10,8 @@ obj-y = sched.o fork.o exec_domain.o kthread.o wait.o kfifo.o sys_ni.o posix-cpu-timers.o mutex.o \ hrtimer.o rwsem.o latency.o nsproxy.o srcu.o die_notifier.o +obj-y += vserver/ + obj-$(CONFIG_STACKTRACE) += stacktrace.o obj-y += time/ obj-$(CONFIG_DEBUG_MUTEXES) += mutex-debug.o diff -NurpP --minimal linux-source-2.6.22.orig/kernel/nsproxy.c linux-source-2.6.22/kernel/nsproxy.c --- linux-source-2.6.22.orig/kernel/nsproxy.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/kernel/nsproxy.c 2007-10-24 16:14:06.791620780 +0200 @@ -20,14 +20,11 @@ #include #include #include +#include +#include struct nsproxy init_nsproxy = INIT_NSPROXY(init_nsproxy); -static inline void get_nsproxy(struct nsproxy *ns) -{ - atomic_inc(&ns->count); -} - void get_task_namespaces(struct task_struct *tsk) { struct nsproxy *ns = tsk->nsproxy; @@ -46,6 +43,9 @@ static inline struct nsproxy *clone_nspr ns = kmemdup(orig, sizeof(struct nsproxy), GFP_KERNEL); if (ns) atomic_set(&ns->count, 1); + vxdprintk(VXD_CBIT(space, 2), "clone_nsproxy(%p[%u] = %p[1]", + orig, atomic_read(&orig->count), ns); + atomic_inc(&vs_global_nsproxy); return ns; } @@ -54,28 +54,32 @@ static inline struct nsproxy *clone_nspr * Return the newly created nsproxy. Do not attach this to the task, * leave it to the caller to do proper locking and attach it to task. */ -static struct nsproxy *create_new_namespaces(int flags, struct task_struct *tsk, +static struct nsproxy *unshare_namespaces(int flags, struct nsproxy *orig, struct fs_struct *new_fs) { struct nsproxy *new_nsp; - new_nsp = clone_nsproxy(tsk->nsproxy); + vxdprintk(VXD_CBIT(space, 4), + "unshare_namespaces(0x%08x,%p,%p)", + flags, orig, new_fs); + + new_nsp = clone_nsproxy(orig); if (!new_nsp) return ERR_PTR(-ENOMEM); - new_nsp->mnt_ns = copy_mnt_ns(flags, tsk->nsproxy->mnt_ns, new_fs); + new_nsp->mnt_ns = copy_mnt_ns(flags, orig->mnt_ns, new_fs); if (IS_ERR(new_nsp->mnt_ns)) goto out_ns; - new_nsp->uts_ns = copy_utsname(flags, tsk->nsproxy->uts_ns); + new_nsp->uts_ns = copy_utsname(flags, orig->uts_ns); if (IS_ERR(new_nsp->uts_ns)) goto out_uts; - new_nsp->ipc_ns = copy_ipcs(flags, tsk->nsproxy->ipc_ns); + new_nsp->ipc_ns = copy_ipcs(flags, orig->ipc_ns); if (IS_ERR(new_nsp->ipc_ns)) goto out_ipc; - new_nsp->pid_ns = copy_pid_ns(flags, tsk->nsproxy->pid_ns); + new_nsp->pid_ns = copy_pid_ns(flags, orig->pid_ns); if (IS_ERR(new_nsp->pid_ns)) goto out_pid; @@ -95,6 +99,33 @@ out_ns: return ERR_PTR(-ENOMEM); } +static struct nsproxy *create_new_namespaces(int flags, struct task_struct *tsk, + struct fs_struct *new_fs) +{ + return unshare_namespaces(flags, tsk->nsproxy, new_fs); +} + +/* + * copies the nsproxy, setting refcount to 1, and grabbing a + * reference to all contained namespaces. + */ +struct nsproxy *copy_nsproxy(struct nsproxy *orig) +{ + struct nsproxy *ns = clone_nsproxy(orig); + + if (ns) { + if (ns->mnt_ns) + get_mnt_ns(ns->mnt_ns); + if (ns->uts_ns) + get_uts_ns(ns->uts_ns); + if (ns->ipc_ns) + get_ipc_ns(ns->ipc_ns); + if (ns->pid_ns) + get_pid_ns(ns->pid_ns); + } + return ns; +} + /* * called from clone. This now handles copy for nsproxy and all * namespaces therein. @@ -102,9 +133,12 @@ out_ns: int copy_namespaces(int flags, struct task_struct *tsk) { struct nsproxy *old_ns = tsk->nsproxy; - struct nsproxy *new_ns; + struct nsproxy *new_ns = NULL; int err = 0; + vxdprintk(VXD_CBIT(space, 7), "copy_namespaces(0x%08x,%p[%p])", + flags, tsk, old_ns); + if (!old_ns) return 0; @@ -127,6 +161,9 @@ int copy_namespaces(int flags, struct ta tsk->nsproxy = new_ns; out: put_nsproxy(old_ns); + vxdprintk(VXD_CBIT(space, 3), + "copy_namespaces(0x%08x,%p[%p]) = %d [%p]", + flags, tsk, old_ns, err, new_ns); return err; } @@ -140,6 +177,7 @@ void free_nsproxy(struct nsproxy *ns) put_ipc_ns(ns->ipc_ns); if (ns->pid_ns) put_pid_ns(ns->pid_ns); + atomic_dec(&vs_global_nsproxy); kfree(ns); } @@ -152,6 +190,10 @@ int unshare_nsproxy_namespaces(unsigned { int err = 0; + vxdprintk(VXD_CBIT(space, 4), + "unshare_nsproxy_namespaces(0x%08lx,[%p])", + unshare_flags, current->nsproxy); + if (!(unshare_flags & (CLONE_NEWNS | CLONE_NEWUTS | CLONE_NEWIPC))) return 0; diff -NurpP --minimal linux-source-2.6.22.orig/kernel/pid.c linux-source-2.6.22/kernel/pid.c --- linux-source-2.6.22.orig/kernel/pid.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/kernel/pid.c 2007-10-24 16:14:06.791620780 +0200 @@ -28,6 +28,7 @@ #include #include #include +#include #define pid_hashfn(nr) hash_long((unsigned long)nr, pidhash_shift) static struct hlist_head *pid_hash; @@ -295,6 +296,9 @@ void fastcall transfer_pid(struct task_s struct task_struct * fastcall pid_task(struct pid *pid, enum pid_type type) { struct task_struct *result = NULL; + + if (type == PIDTYPE_REALPID) + type = PIDTYPE_PID; if (pid) { struct hlist_node *first; first = rcu_dereference(pid->tasks[type].first); @@ -309,7 +313,17 @@ struct task_struct * fastcall pid_task(s */ struct task_struct *find_task_by_pid_type(int type, int nr) { - return pid_task(find_pid(nr), type); + struct task_struct *task; + + if (type == PIDTYPE_PID) + nr = vx_rmap_pid(nr); + + task = pid_task(find_pid(nr), type); + if (task && (type != PIDTYPE_REALPID) && + /* maybe VS_WATCH_P in the future? */ + !vx_check(task->xid, VS_WATCH|VS_IDENT)) + return NULL; + return task; } EXPORT_SYMBOL(find_task_by_pid_type); diff -NurpP --minimal linux-source-2.6.22.orig/kernel/posix-timers.c linux-source-2.6.22/kernel/posix-timers.c --- linux-source-2.6.22.orig/kernel/posix-timers.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/kernel/posix-timers.c 2007-10-24 16:14:06.791620780 +0200 @@ -47,6 +47,7 @@ #include #include #include +#include /* * Management arrays for POSIX timers. Timers are kept in slab memory @@ -297,6 +298,12 @@ void do_schedule_next_timer(struct sigin int posix_timer_event(struct k_itimer *timr,int si_private) { + struct vx_info_save vxis; + struct vx_info *vxi; + int ret; + + vxi = task_get_vx_info(timr->it_process); + enter_vx_info(vxi, &vxis); memset(&timr->sigq->info, 0, sizeof(siginfo_t)); timr->sigq->info.si_sys_private = si_private; /* Send signal to the process that owns this timer.*/ @@ -309,11 +316,11 @@ int posix_timer_event(struct k_itimer *t if (timr->it_sigev_notify & SIGEV_THREAD_ID) { struct task_struct *leader; - int ret = send_sigqueue(timr->it_sigev_signo, timr->sigq, - timr->it_process); + ret = send_sigqueue(timr->it_sigev_signo, timr->sigq, + timr->it_process); if (likely(ret >= 0)) - return ret; + goto out; timr->it_sigev_notify = SIGEV_SIGNAL; leader = timr->it_process->group_leader; @@ -321,8 +328,12 @@ int posix_timer_event(struct k_itimer *t timr->it_process = leader; } - return send_group_sigqueue(timr->it_sigev_signo, timr->sigq, + ret = send_group_sigqueue(timr->it_sigev_signo, timr->sigq, timr->it_process); +out: + leave_vx_info(&vxis); + put_vx_info(vxi); + return ret; } EXPORT_SYMBOL_GPL(posix_timer_event); @@ -402,7 +413,7 @@ static struct task_struct * good_sigeven struct task_struct *rtn = current->group_leader; if ((event->sigev_notify & SIGEV_THREAD_ID ) && - (!(rtn = find_task_by_pid(event->sigev_notify_thread_id)) || + (!(rtn = find_task_by_real_pid(event->sigev_notify_thread_id)) || rtn->tgid != current->tgid || (event->sigev_notify & ~SIGEV_THREAD_ID) != SIGEV_SIGNAL)) return NULL; diff -NurpP --minimal linux-source-2.6.22.orig/kernel/printk.c linux-source-2.6.22/kernel/printk.c --- linux-source-2.6.22.orig/kernel/printk.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/kernel/printk.c 2007-10-24 16:14:06.791620780 +0200 @@ -31,6 +31,7 @@ #include #include #include +#include #include @@ -182,18 +183,13 @@ int do_syslog(int type, char __user *buf unsigned long i, j, limit, count; int do_clear = 0; char c; - int error = 0; + int error; error = security_syslog(type); if (error) return error; - switch (type) { - case 0: /* Close log */ - break; - case 1: /* Open log */ - break; - case 2: /* Read from log */ + if ((type >= 2) && (type <= 4)) { error = -EINVAL; if (!buf || len < 0) goto out; @@ -204,6 +200,16 @@ int do_syslog(int type, char __user *buf error = -EFAULT; goto out; } + } + if (!vx_check(0, VS_ADMIN|VS_WATCH)) + return vx_do_syslog(type, buf, len); + + switch (type) { + case 0: /* Close log */ + break; + case 1: /* Open log */ + break; + case 2: /* Read from log */ error = wait_event_interruptible(log_wait, (log_start - log_end)); if (error) @@ -228,16 +234,6 @@ int do_syslog(int type, char __user *buf do_clear = 1; /* FALL THRU */ case 3: /* Read last kernel messages */ - error = -EINVAL; - if (!buf || len < 0) - goto out; - error = 0; - if (!len) - goto out; - if (!access_ok(VERIFY_WRITE, buf, len)) { - error = -EFAULT; - goto out; - } count = len; if (count > log_buf_len) count = log_buf_len; diff -NurpP --minimal linux-source-2.6.22.orig/kernel/ptrace.c linux-source-2.6.22/kernel/ptrace.c --- linux-source-2.6.22.orig/kernel/ptrace.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/kernel/ptrace.c 2007-10-24 16:14:06.791620780 +0200 @@ -19,6 +19,7 @@ #include #include #include +#include #include #include @@ -145,6 +146,11 @@ static int may_attach(struct task_struct dumpable = task->mm->dumpable; if (!dumpable && !capable(CAP_SYS_PTRACE)) return -EPERM; + if (!vx_check(task->xid, VS_ADMIN_P|VS_IDENT)) + return -EPERM; + if (!vx_check(task->xid, VS_IDENT) && + !task_vx_flags(task, VXF_STATE_ADMIN, 0)) + return -EACCES; return security_ptrace(current, task); } @@ -471,6 +477,10 @@ asmlinkage long sys_ptrace(long request, goto out; } + ret = -EPERM; + if (!vx_check(vx_task_xid(child), VS_WATCH_P | VS_IDENT)) + goto out_put_task_struct; + if (request == PTRACE_ATTACH) { ret = ptrace_attach(child); goto out_put_task_struct; diff -NurpP --minimal linux-source-2.6.22.orig/kernel/sched.c linux-source-2.6.22/kernel/sched.c --- linux-source-2.6.22.orig/kernel/sched.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/kernel/sched.c 2007-10-24 16:14:06.791620780 +0200 @@ -56,6 +56,8 @@ #include #include +#include +#include /* * Scheduler clock - returns current time in nanosec units. @@ -281,6 +283,16 @@ struct rq { struct task_struct *migration_thread; struct list_head migration_queue; #endif + unsigned long norm_time; + unsigned long idle_time; +#ifdef CONFIG_VSERVER_IDLETIME + int idle_skip; +#endif +#ifdef CONFIG_VSERVER_HARDCPU + struct list_head hold_queue; + unsigned long nr_onhold; + int idle_tokens; +#endif #ifdef CONFIG_SCHEDSTATS /* latency stats */ @@ -714,6 +726,7 @@ sched_info_switch(struct task_struct *pr */ static void dequeue_task(struct task_struct *p, struct prio_array *array) { + BUG_ON(p->state & TASK_ONHOLD); array->nr_active--; list_del(&p->run_list); if (list_empty(array->queue + p->prio)) @@ -722,6 +735,7 @@ static void dequeue_task(struct task_str static void enqueue_task(struct task_struct *p, struct prio_array *array) { + BUG_ON(p->state & TASK_ONHOLD); sched_info_queued(p); list_add_tail(&p->run_list, array->queue + p->prio); __set_bit(p->prio, array->bitmap); @@ -735,12 +749,14 @@ static void enqueue_task(struct task_str */ static void requeue_task(struct task_struct *p, struct prio_array *array) { + BUG_ON(p->state & TASK_ONHOLD); list_move_tail(&p->run_list, array->queue + p->prio); } static inline void enqueue_task_head(struct task_struct *p, struct prio_array *array) { + BUG_ON(p->state & TASK_ONHOLD); list_add(&p->run_list, array->queue + p->prio); __set_bit(p->prio, array->bitmap); array->nr_active++; @@ -769,6 +785,10 @@ static inline int __normal_prio(struct t bonus = CURRENT_BONUS(p) - MAX_BONUS / 2; prio = p->static_prio - bonus; + + /* adjust effective priority */ + prio = vx_adjust_prio(p, prio, MAX_USER_PRIO); + if (prio < MAX_RT_PRIO) prio = MAX_RT_PRIO; if (prio > MAX_PRIO-1) @@ -878,6 +898,9 @@ static int effective_prio(struct task_st return p->prio; } +#include "sched_mon.h" + + /* * __activate_task - move a task to the runqueue. */ @@ -887,6 +910,7 @@ static void __activate_task(struct task_ if (batch_task(p)) target = rq->expired; + vxm_activate_task(p, rq); enqueue_task(p, target); inc_nr_running(p, rq); } @@ -896,6 +920,7 @@ static void __activate_task(struct task_ */ static inline void __activate_idle_task(struct task_struct *p, struct rq *rq) { + vxm_activate_idle(p, rq); enqueue_task_head(p, rq->active); inc_nr_running(p, rq); } @@ -1030,19 +1055,30 @@ static void activate_task(struct task_st } p->timestamp = now; out: + vx_activate_task(p); __activate_task(p, rq); } /* - * deactivate_task - remove a task from the runqueue. + * __deactivate_task - remove a task from the runqueue. */ -static void deactivate_task(struct task_struct *p, struct rq *rq) +static void __deactivate_task(struct task_struct *p, struct rq *rq) { dec_nr_running(p, rq); dequeue_task(p, p->array); + vxm_deactivate_task(p, rq); p->array = NULL; } +static inline +void deactivate_task(struct task_struct *p, struct rq *rq) +{ + vx_deactivate_task(p); + __deactivate_task(p, rq); +} + +#include "sched_hard.h" + /* * resched_task - mark a task 'to be rescheduled now'. * @@ -1129,6 +1165,7 @@ migrate_task(struct task_struct *p, int { struct rq *rq = task_rq(p); + vxm_migrate_task(p, rq, dest_cpu); /* * If the task is not on a runqueue (and not running), then * it is sufficient to simply update the task's cpu field. @@ -1518,6 +1555,12 @@ static int try_to_wake_up(struct task_st rq = task_rq_lock(p, &flags); old_state = p->state; + + /* we need to unhold suspended tasks */ + if (old_state & TASK_ONHOLD) { + vx_unhold_task(p, rq); + old_state = p->state; + } if (!(old_state & state)) goto out; @@ -1625,6 +1668,7 @@ out_activate: #endif /* CONFIG_SMP */ if (old_state == TASK_UNINTERRUPTIBLE) { rq->nr_uninterruptible--; + vx_uninterruptible_dec(p); /* * Tasks on involuntary sleep don't earn * sleep_avg beyond just interactive state. @@ -1676,7 +1720,7 @@ int fastcall wake_up_state(struct task_s return try_to_wake_up(p, state, 0); } -static void task_running_tick(struct rq *rq, struct task_struct *p); +static void task_running_tick(struct rq *rq, struct task_struct *p, int cpu); /* * Perform scheduler related setup for a newly forked process p. * p is forked by current. @@ -1737,7 +1781,7 @@ void fastcall sched_fork(struct task_str * runqueue lock is not a problem. */ current->time_slice = 1; - task_running_tick(cpu_rq(cpu), current); + task_running_tick(cpu_rq(cpu), current, cpu); } local_irq_enable(); put_cpu(); @@ -1772,6 +1816,7 @@ void fastcall wake_up_new_task(struct ta p->prio = effective_prio(p); + vx_activate_task(p); if (likely(cpu == this_cpu)) { if (!(clone_flags & CLONE_VM)) { /* @@ -1783,6 +1828,7 @@ void fastcall wake_up_new_task(struct ta __activate_task(p, rq); else { p->prio = current->prio; + BUG_ON(p->state & TASK_ONHOLD); p->normal_prio = current->normal_prio; list_add_tail(&p->run_list, ¤t->run_list); p->array = current->array; @@ -3351,13 +3397,16 @@ static inline int expired_starving(struc void account_user_time(struct task_struct *p, cputime_t cputime) { struct cpu_usage_stat *cpustat = &kstat_this_cpu.cpustat; + struct vx_info *vxi = p->vx_info; /* p is _always_ current */ cputime64_t tmp; + int nice = (TASK_NICE(p) > 0); p->utime = cputime_add(p->utime, cputime); + vx_account_user(vxi, cputime, nice); /* Add user time to cpustat. */ tmp = cputime_to_cputime64(cputime); - if (TASK_NICE(p) > 0) + if (nice) cpustat->nice = cputime64_add(cpustat->nice, tmp); else cpustat->user = cputime64_add(cpustat->user, tmp); @@ -3373,10 +3422,12 @@ void account_system_time(struct task_str cputime_t cputime) { struct cpu_usage_stat *cpustat = &kstat_this_cpu.cpustat; + struct vx_info *vxi = p->vx_info; /* p is _always_ current */ struct rq *rq = this_rq(); cputime64_t tmp; p->stime = cputime_add(p->stime, cputime); + vx_account_system(vxi, cputime, (p == rq->idle)); /* Add system time to cpustat. */ tmp = cputime_to_cputime64(cputime); @@ -3415,7 +3466,7 @@ void account_steal_time(struct task_stru cpustat->steal = cputime64_add(cpustat->steal, tmp); } -static void task_running_tick(struct rq *rq, struct task_struct *p) +static void task_running_tick(struct rq *rq, struct task_struct *p, int cpu) { if (p->array != rq->active) { /* Task has expired but was not scheduled yet */ @@ -3445,7 +3496,7 @@ static void task_running_tick(struct rq } goto out_unlock; } - if (!--p->time_slice) { + if (vx_need_resched(p, --p->time_slice, cpu)) { dequeue_task(p, rq->active); set_tsk_need_resched(p); p->prio = effective_prio(p); @@ -3506,9 +3557,12 @@ void scheduler_tick(void) struct rq *rq = cpu_rq(cpu); update_cpu_clock(p, rq, now); + vxm_sync(now, cpu); - if (!idle_at_tick) - task_running_tick(rq, p); + if (idle_at_tick) + vx_idle_resched(rq); + else + task_running_tick(rq, p, cpu); #ifdef CONFIG_SMP update_load(rq); rq->idle_at_tick = idle_at_tick; @@ -3630,14 +3684,25 @@ need_resched_nonpreemptible: unlikely(signal_pending(prev)))) prev->state = TASK_RUNNING; else { - if (prev->state == TASK_UNINTERRUPTIBLE) + if (prev->state == TASK_UNINTERRUPTIBLE) { rq->nr_uninterruptible++; + vx_uninterruptible_inc(prev); + } deactivate_task(prev, rq); } } cpu = smp_processor_id(); + vx_set_rq_time(rq, jiffies); +try_unhold: + vx_try_unhold(rq, cpu); +pick_next: + if (unlikely(!rq->nr_running)) { + /* can we skip idle time? */ + if (vx_try_skip(rq, cpu)) + goto try_unhold; + idle_balance(cpu, rq); if (!rq->nr_running) { next = rq->idle; @@ -3663,6 +3728,10 @@ need_resched_nonpreemptible: queue = array->queue + idx; next = list_entry(queue->next, struct task_struct, run_list); + /* check before we schedule this context */ + if (!vx_schedule(next, rq, cpu)) + goto pick_next; + if (!rt_task(next) && interactive_sleep(next->sleep_type)) { unsigned long long delta = now - next->timestamp; if (unlikely((long long)(now - next->timestamp) < 0)) @@ -4263,7 +4332,7 @@ asmlinkage long sys_nice(int increment) nice = 19; if (increment < 0 && !can_nice(current, nice)) - return -EPERM; + return vx_flags(VXF_IGNEG_NICE, 0) ? 0 : -EPERM; retval = security_task_setnice(current, nice); if (retval) @@ -4435,6 +4504,7 @@ recheck: oldprio = p->prio; __setscheduler(p, policy, param->sched_priority); if (array) { + vx_activate_task(p); __activate_task(p, rq); /* * Reschedule if we are currently running on this runqueue and @@ -5188,6 +5258,7 @@ static int __migrate_task(struct task_st p->timestamp = p->timestamp - rq_src->most_recent_timestamp + rq_dest->most_recent_timestamp; deactivate_task(p, rq_src); + vx_activate_task(p); __activate_task(p, rq_dest); if (TASK_PREEMPTS_CURR(p, rq_dest)) resched_task(rq_dest->curr); @@ -7058,7 +7129,10 @@ void __init sched_init(void) INIT_LIST_HEAD(&rq->migration_queue); #endif atomic_set(&rq->nr_iowait, 0); - +#ifdef CONFIG_VSERVER_HARDCPU + INIT_LIST_HEAD(&rq->hold_queue); + rq->nr_onhold = 0; +#endif for (j = 0; j < 2; j++) { array = rq->arrays + j; for (k = 0; k < MAX_PRIO; k++) { @@ -7144,6 +7218,7 @@ void normalize_rt_tasks(void) deactivate_task(p, task_rq(p)); __setscheduler(p, SCHED_NORMAL, 0); if (array) { + vx_activate_task(p); __activate_task(p, task_rq(p)); resched_task(rq->curr); } diff -NurpP --minimal linux-source-2.6.22.orig/kernel/sched_hard.h linux-source-2.6.22/kernel/sched_hard.h --- linux-source-2.6.22.orig/kernel/sched_hard.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/sched_hard.h 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,324 @@ + +#ifdef CONFIG_VSERVER_IDLELIMIT + +/* + * vx_idle_resched - reschedule after maxidle + */ +static inline +void vx_idle_resched(struct rq *rq) +{ + /* maybe have a better criterion for paused */ + if (!--rq->idle_tokens && !list_empty(&rq->hold_queue)) + set_need_resched(); +} + +#else /* !CONFIG_VSERVER_IDLELIMIT */ + +#define vx_idle_resched(rq) + +#endif /* CONFIG_VSERVER_IDLELIMIT */ + + + +#ifdef CONFIG_VSERVER_IDLETIME + +#define vx_set_rq_min_skip(rq, min) \ + (rq)->idle_skip = (min) + +#define vx_save_min_skip(ret, min, val) \ + __vx_save_min_skip(ret, min, val) + +static inline +void __vx_save_min_skip(int ret, int *min, int val) +{ + if (ret > -2) + return; + if ((*min > val) || !*min) + *min = val; +} + +static inline +int vx_try_skip(struct rq *rq, int cpu) +{ + /* artificially advance time */ + if (rq->idle_skip > 0) { + vxdprintk(list_empty(&rq->hold_queue), + "hold queue empty on cpu %d", cpu); + rq->idle_time += rq->idle_skip; + vxm_idle_skip(rq, cpu); + return 1; + } + return 0; +} + +#else /* !CONFIG_VSERVER_IDLETIME */ + +#define vx_set_rq_min_skip(rq, min) \ + ({ int dummy = (min); dummy; }) + +#define vx_save_min_skip(ret, min, val) + +static inline +int vx_try_skip(struct rq *rq, int cpu) +{ + return 0; +} + +#endif /* CONFIG_VSERVER_IDLETIME */ + + + +#ifdef CONFIG_VSERVER_HARDCPU + +#define vx_set_rq_max_idle(rq, max) \ + (rq)->idle_tokens = (max) + +#define vx_save_max_idle(ret, min, val) \ + __vx_save_max_idle(ret, min, val) + +static inline +void __vx_save_max_idle(int ret, int *min, int val) +{ + if (*min > val) + *min = val; +} + + +/* + * vx_hold_task - put a task on the hold queue + */ +static inline +void vx_hold_task(struct task_struct *p, struct rq *rq) +{ + __deactivate_task(p, rq); + p->state |= TASK_ONHOLD; + /* a new one on hold */ + rq->nr_onhold++; + vxm_hold_task(p, rq); + list_add_tail(&p->run_list, &rq->hold_queue); +} + +/* + * vx_unhold_task - put a task back to the runqueue + */ +static inline +void vx_unhold_task(struct task_struct *p, struct rq *rq) +{ + list_del(&p->run_list); + /* one less waiting */ + rq->nr_onhold--; + p->state &= ~TASK_ONHOLD; + enqueue_task(p, rq->expired); + inc_nr_running(p, rq); + vxm_unhold_task(p, rq); + + if (p->static_prio < rq->best_expired_prio) + rq->best_expired_prio = p->static_prio; +} + +unsigned long nr_onhold(void) +{ + unsigned long i, sum = 0; + + for_each_online_cpu(i) + sum += cpu_rq(i)->nr_onhold; + + return sum; +} + + + +static inline +int __vx_tokens_avail(struct _vx_sched_pc *sched_pc) +{ + return sched_pc->tokens; +} + +static inline +void __vx_consume_token(struct _vx_sched_pc *sched_pc) +{ + sched_pc->tokens--; +} + +static inline +int vx_need_resched(struct task_struct *p, int slice, int cpu) +{ + struct vx_info *vxi = p->vx_info; + + if (vx_info_flags(vxi, VXF_SCHED_HARD|VXF_SCHED_PRIO, 0)) { + struct _vx_sched_pc *sched_pc = + &vx_per_cpu(vxi, sched_pc, cpu); + int tokens; + + /* maybe we can simplify that to decrement + the token counter unconditional? */ + + if ((tokens = __vx_tokens_avail(sched_pc)) > 0) + __vx_consume_token(sched_pc); + + /* for tokens > 0, one token was consumed */ + if (tokens < 2) + slice = 0; + } + vxm_need_resched(p, slice, cpu); + return (slice == 0); +} + + +#define vx_set_rq_time(rq, time) do { \ + rq->norm_time = time; \ +} while (0) + + +static inline +void vx_try_unhold(struct rq *rq, int cpu) +{ + struct vx_info *vxi = NULL; + struct list_head *l, *n; + int maxidle = HZ; + int minskip = 0; + + /* nothing to do? what about pause? */ + if (list_empty(&rq->hold_queue)) + return; + + list_for_each_safe(l, n, &rq->hold_queue) { + int ret, delta_min[2]; + struct _vx_sched_pc *sched_pc; + struct task_struct *p; + + p = list_entry(l, struct task_struct, run_list); + /* don't bother with same context */ + if (vxi == p->vx_info) + continue; + + vxi = p->vx_info; + /* ignore paused contexts */ + if (vx_info_flags(vxi, VXF_SCHED_PAUSE, 0)) + continue; + + sched_pc = &vx_per_cpu(vxi, sched_pc, cpu); + + /* recalc tokens */ + vxm_sched_info(sched_pc, vxi, cpu); + ret = vx_tokens_recalc(sched_pc, + &rq->norm_time, &rq->idle_time, delta_min); + vxm_tokens_recalc(sched_pc, rq, vxi, cpu); + + if (ret > 0) { + /* we found a runable context */ + vx_unhold_task(p, rq); + break; + } + vx_save_max_idle(ret, &maxidle, delta_min[0]); + vx_save_min_skip(ret, &minskip, delta_min[1]); + } + vx_set_rq_max_idle(rq, maxidle); + vx_set_rq_min_skip(rq, minskip); + vxm_rq_max_min(rq, cpu); +} + + +static inline +int vx_schedule(struct task_struct *next, struct rq *rq, int cpu) +{ + struct vx_info *vxi = next->vx_info; + struct _vx_sched_pc *sched_pc; + int delta_min[2]; + int flags, ret; + + if (!vxi) + return 1; + + flags = vxi->vx_flags; + + if (unlikely(vs_check_flags(flags, VXF_SCHED_PAUSE, 0))) + goto put_on_hold; + if (!vs_check_flags(flags, VXF_SCHED_HARD | VXF_SCHED_PRIO, 0)) + return 1; + + sched_pc = &vx_per_cpu(vxi, sched_pc, cpu); +#ifdef CONFIG_SMP + /* update scheduler params */ + if (cpu_isset(cpu, vxi->sched.update)) { + vx_update_sched_param(&vxi->sched, sched_pc); + vxm_update_sched(sched_pc, vxi, cpu); + cpu_clear(cpu, vxi->sched.update); + } +#endif + vxm_sched_info(sched_pc, vxi, cpu); + ret = vx_tokens_recalc(sched_pc, + &rq->norm_time, &rq->idle_time, delta_min); + vxm_tokens_recalc(sched_pc, rq, vxi, cpu); + + if (!vs_check_flags(flags, VXF_SCHED_HARD, 0)) + return 1; + + if (unlikely(ret < 0)) { + vx_save_max_idle(ret, &rq->idle_tokens, delta_min[0]); + vx_save_min_skip(ret, &rq->idle_skip, delta_min[1]); + vxm_rq_max_min(rq, cpu); + put_on_hold: + vx_hold_task(next, rq); + return 0; + } + return 1; +} + + +#else /* CONFIG_VSERVER_HARDCPU */ + +static inline +void vx_hold_task(struct task_struct *p, struct rq *rq) +{ + return; +} + +static inline +void vx_unhold_task(struct task_struct *p, struct rq *rq) +{ + return; +} + +unsigned long nr_onhold(void) +{ + return 0; +} + + +static inline +int vx_need_resched(struct task_struct *p, int slice, int cpu) +{ + return (slice == 0); +} + + +#define vx_set_rq_time(rq, time) + +static inline +void vx_try_unhold(struct rq *rq, int cpu) +{ + return; +} + +static inline +int vx_schedule(struct task_struct *next, struct rq *rq, int cpu) +{ + struct vx_info *vxi = next->vx_info; + struct _vx_sched_pc *sched_pc; + int delta_min[2]; + int ret; + + if (!vx_info_flags(vxi, VXF_SCHED_PRIO, 0)) + return 1; + + sched_pc = &vx_per_cpu(vxi, sched_pc, cpu); + vxm_sched_info(sched_pc, vxi, cpu); + ret = vx_tokens_recalc(sched_pc, + &rq->norm_time, &rq->idle_time, delta_min); + vxm_tokens_recalc(sched_pc, rq, vxi, cpu); + return 1; +} + +#endif /* CONFIG_VSERVER_HARDCPU */ + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/sched_mon.h linux-source-2.6.22/kernel/sched_mon.h --- linux-source-2.6.22.orig/kernel/sched_mon.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/sched_mon.h 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,200 @@ + +#include + +#ifdef CONFIG_VSERVER_MONITOR + +#ifdef CONFIG_VSERVER_HARDCPU +#define HARDCPU(x) (x) +#else +#define HARDCPU(x) (0) +#endif + +#ifdef CONFIG_VSERVER_IDLETIME +#define IDLETIME(x) (x) +#else +#define IDLETIME(x) (0) +#endif + +struct _vx_mon_entry *vxm_advance(int cpu); + + +static inline +void __vxm_basic(struct _vx_mon_entry *entry, xid_t xid, int type) +{ + entry->type = type; + entry->xid = xid; +} + +static inline +void __vxm_sync(int cpu) +{ + struct _vx_mon_entry *entry = vxm_advance(cpu); + + __vxm_basic(entry, 0, VXM_SYNC); + entry->ev.sec = xtime.tv_sec; + entry->ev.nsec = xtime.tv_nsec; +} + +static inline +void __vxm_task(struct task_struct *p, int type) +{ + struct _vx_mon_entry *entry = vxm_advance(task_cpu(p)); + + __vxm_basic(entry, p->xid, type); + entry->ev.tsk.pid = p->pid; + entry->ev.tsk.state = p->state; +} + +static inline +void __vxm_sched(struct _vx_sched_pc *s, struct vx_info *vxi, int cpu) +{ + struct _vx_mon_entry *entry = vxm_advance(cpu); + + __vxm_basic(entry, vxi->vx_id, (VXM_SCHED | s->flags)); + entry->sd.tokens = s->tokens; + entry->sd.norm_time = s->norm_time; + entry->sd.idle_time = s->idle_time; +} + +static inline +void __vxm_rqinfo1(struct rq *q, int cpu) +{ + struct _vx_mon_entry *entry = vxm_advance(cpu); + + entry->type = VXM_RQINFO_1; + entry->xid = ((unsigned long)q >> 16) & 0xffff; + entry->q1.running = q->nr_running; + entry->q1.onhold = HARDCPU(q->nr_onhold); + entry->q1.iowait = atomic_read(&q->nr_iowait); + entry->q1.uintr = q->nr_uninterruptible; + entry->q1.idle_tokens = IDLETIME(q->idle_tokens); +} + +static inline +void __vxm_rqinfo2(struct rq *q, int cpu) +{ + struct _vx_mon_entry *entry = vxm_advance(cpu); + + entry->type = VXM_RQINFO_2; + entry->xid = (unsigned long)q & 0xffff; + entry->q2.norm_time = q->norm_time; + entry->q2.idle_time = q->idle_time; + entry->q2.idle_skip = IDLETIME(q->idle_skip); +} + +static inline +void __vxm_update(struct _vx_sched_pc *s, struct vx_info *vxi, int cpu) +{ + struct _vx_mon_entry *entry = vxm_advance(cpu); + + __vxm_basic(entry, vxi->vx_id, VXM_UPDATE); + entry->ev.tokens = s->tokens; +} + +static inline +void __vxm_update1(struct _vx_sched_pc *s, struct vx_info *vxi, int cpu) +{ + struct _vx_mon_entry *entry = vxm_advance(cpu); + + __vxm_basic(entry, vxi->vx_id, VXM_UPDATE_1); + entry->u1.tokens_max = s->tokens_max; + entry->u1.fill_rate = s->fill_rate[0]; + entry->u1.interval = s->interval[0]; +} + +static inline +void __vxm_update2(struct _vx_sched_pc *s, struct vx_info *vxi, int cpu) +{ + struct _vx_mon_entry *entry = vxm_advance(cpu); + + __vxm_basic(entry, vxi->vx_id, VXM_UPDATE_2); + entry->u2.tokens_min = s->tokens_min; + entry->u2.fill_rate = s->fill_rate[1]; + entry->u2.interval = s->interval[1]; +} + + +#define vxm_activate_task(p,q) __vxm_task(p, VXM_ACTIVATE) +#define vxm_activate_idle(p,q) __vxm_task(p, VXM_IDLE) +#define vxm_deactivate_task(p,q) __vxm_task(p, VXM_DEACTIVATE) +#define vxm_hold_task(p,q) __vxm_task(p, VXM_HOLD) +#define vxm_unhold_task(p,q) __vxm_task(p, VXM_UNHOLD) + +static inline +void vxm_migrate_task(struct task_struct *p, struct rq *rq, int dest) +{ + __vxm_task(p, VXM_MIGRATE); + __vxm_rqinfo1(rq, task_cpu(p)); + __vxm_rqinfo2(rq, task_cpu(p)); +} + +static inline +void vxm_idle_skip(struct rq *rq, int cpu) +{ + __vxm_rqinfo1(rq, cpu); + __vxm_rqinfo2(rq, cpu); +} + +static inline +void vxm_need_resched(struct task_struct *p, int slice, int cpu) +{ + if (slice) + return; + + __vxm_task(p, VXM_RESCHED); +} + +static inline +void vxm_sync(unsigned long now, int cpu) +{ + if (!CONFIG_VSERVER_MONITOR_SYNC || + (now % CONFIG_VSERVER_MONITOR_SYNC)) + return; + + __vxm_sync(cpu); +} + +#define vxm_sched_info(s,v,c) __vxm_sched(s,v,c) + +static inline +void vxm_tokens_recalc(struct _vx_sched_pc *s, struct rq *rq, + struct vx_info *vxi, int cpu) +{ + __vxm_sched(s, vxi, cpu); + __vxm_rqinfo2(rq, cpu); +} + +static inline +void vxm_update_sched(struct _vx_sched_pc *s, struct vx_info *vxi, int cpu) +{ + __vxm_sched(s, vxi, cpu); + __vxm_update(s, vxi, cpu); + __vxm_update1(s, vxi, cpu); + __vxm_update2(s, vxi, cpu); +} + +static inline +void vxm_rq_max_min(struct rq *rq, int cpu) +{ + __vxm_rqinfo1(rq, cpu); + __vxm_rqinfo2(rq, cpu); +} + +#else /* CONFIG_VSERVER_MONITOR */ + +#define vxm_activate_task(t,q) do { } while (0) +#define vxm_activate_idle(t,q) do { } while (0) +#define vxm_deactivate_task(t,q) do { } while (0) +#define vxm_hold_task(t,q) do { } while (0) +#define vxm_unhold_task(t,q) do { } while (0) +#define vxm_migrate_task(t,q,d) do { } while (0) +#define vxm_idle_skip(q,c) do { } while (0) +#define vxm_need_resched(t,s,c) do { } while (0) +#define vxm_sync(s,c) do { } while (0) +#define vxm_sched_info(s,v,c) do { } while (0) +#define vxm_tokens_recalc(s,q,v,c) do { } while (0) +#define vxm_update_sched(s,v,c) do { } while (0) +#define vxm_rq_max_min(q,c) do { } while (0) + +#endif /* CONFIG_VSERVER_MONITOR */ + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/signal.c linux-source-2.6.22/kernel/signal.c --- linux-source-2.6.22.orig/kernel/signal.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/kernel/signal.c 2007-10-24 16:14:06.791620780 +0200 @@ -26,6 +26,8 @@ #include #include #include +#include +#include #include #include @@ -523,19 +525,34 @@ static int check_kill_permission(int sig if (!valid_signal(sig)) return error; + if ((info != SEND_SIG_NOINFO) && + (is_si_special(info) || !SI_FROMUSER(info))) + goto skip; + + vxdprintk(VXD_CBIT(misc, 7), + "check_kill_permission(%d,%p,%p[#%u,%u])", + sig, info, t, vx_task_xid(t), t->pid); + error = audit_signal_info(sig, t); /* Let audit system see the signal */ if (error) return error; error = -EPERM; - if ((info == SEND_SIG_NOINFO || (!is_si_special(info) && SI_FROMUSER(info))) - && ((sig != SIGCONT) || + if (((sig != SIGCONT) || (process_session(current) != process_session(t))) && (current->euid ^ t->suid) && (current->euid ^ t->uid) && (current->uid ^ t->suid) && (current->uid ^ t->uid) && !capable(CAP_KILL)) return error; + error = -ESRCH; + if (!vx_check(vx_task_xid(t), VS_WATCH_P | VS_IDENT)) { + vxdprintk(current->xid || VXD_CBIT(misc, 7), + "signal %d[%p] xid mismatch %p[#%u,%u] xid=#%u", + sig, info, t, vx_task_xid(t), t->pid, current->xid); + return error; + } +skip: return security_task_kill(t, info, sig, 0); } @@ -1043,7 +1060,7 @@ int kill_pid_info(int sig, struct siginf p = pid_task(pid, PIDTYPE_PID); error = -ESRCH; - if (p) + if (p && vx_check(vx_task_xid(p), VS_IDENT)) error = group_send_sig_info(sig, info, p); if (unlikely(sig_needs_tasklist(sig))) @@ -1057,7 +1074,7 @@ kill_proc_info(int sig, struct siginfo * { int error; rcu_read_lock(); - error = kill_pid_info(sig, info, find_pid(pid)); + error = kill_pid_info(sig, info, find_pid(vx_rmap_pid(pid))); rcu_read_unlock(); return error; } @@ -1118,7 +1135,8 @@ static int kill_something_info(int sig, read_lock(&tasklist_lock); for_each_process(p) { - if (p->pid > 1 && p->tgid != current->tgid) { + if (vx_check(vx_task_xid(p), VS_ADMIN_P|VS_IDENT) && + p->pid > 1 && p->tgid != current->tgid) { int err = group_send_sig_info(sig, info, p); ++count; if (err != -EPERM) @@ -1814,6 +1832,11 @@ relock: if (current == child_reaper(current)) continue; + /* virtual init is protected against user signals */ + if ((info->si_code == SI_USER) && + vx_current_initpid(current->pid)) + continue; + if (sig_kernel_stop(signr)) { /* * The default action is to stop all threads in diff -NurpP --minimal linux-source-2.6.22.orig/kernel/softirq.c linux-source-2.6.22/kernel/softirq.c --- linux-source-2.6.22.orig/kernel/softirq.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/kernel/softirq.c 2007-10-24 16:14:06.791620780 +0200 @@ -18,6 +18,7 @@ #include #include #include +#include #include /* diff -NurpP --minimal linux-source-2.6.22.orig/kernel/sys.c linux-source-2.6.22/kernel/sys.c --- linux-source-2.6.22.orig/kernel/sys.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/kernel/sys.c 2007-10-24 16:14:06.791620780 +0200 @@ -35,6 +35,7 @@ #include #include #include +#include #include #include @@ -638,7 +639,10 @@ static int set_one_prio(struct task_stru goto out; } if (niceval < task_nice(p) && !can_nice(p, niceval)) { - error = -EACCES; + if (vx_flags(VXF_IGNEG_NICE, 0)) + error = 0; + else + error = -EACCES; goto out; } no_nice = security_task_setnice(p, niceval); @@ -686,6 +690,8 @@ asmlinkage long sys_setpriority(int whic else pgrp = task_pgrp(current); do_each_pid_task(pgrp, PIDTYPE_PGID, p) { + if (!vx_check(p->xid, VS_ADMIN_P | VS_IDENT)) + continue; error = set_one_prio(p, niceval, error); } while_each_pid_task(pgrp, PIDTYPE_PGID, p); break; @@ -694,7 +700,8 @@ asmlinkage long sys_setpriority(int whic if (!who) who = current->uid; else - if ((who != current->uid) && !(user = find_user(who))) + if ((who != current->uid) && + !(user = find_user(vx_current_xid(), who))) goto out_unlock; /* No processes for this user */ do_each_thread(g, p) @@ -746,6 +753,8 @@ asmlinkage long sys_getpriority(int whic else pgrp = task_pgrp(current); do_each_pid_task(pgrp, PIDTYPE_PGID, p) { + if (!vx_check(p->xid, VS_ADMIN_P | VS_IDENT)) + continue; niceval = 20 - task_nice(p); if (niceval > retval) retval = niceval; @@ -756,7 +765,8 @@ asmlinkage long sys_getpriority(int whic if (!who) who = current->uid; else - if ((who != current->uid) && !(user = find_user(who))) + if ((who != current->uid) && + !(user = find_user(vx_current_xid(), who))) goto out_unlock; /* No processes for this user */ do_each_thread(g, p) @@ -869,6 +879,9 @@ void kernel_power_off(void) machine_power_off(); } EXPORT_SYMBOL_GPL(kernel_power_off); + +long vs_reboot(unsigned int, void __user *); + /* * Reboot system call: for obvious reasons only root may call it, * and even root needs to set up some magic numbers in the registers @@ -899,6 +912,9 @@ asmlinkage long sys_reboot(int magic1, i if ((cmd == LINUX_REBOOT_CMD_POWER_OFF) && !pm_power_off) cmd = LINUX_REBOOT_CMD_HALT; + if (!vx_check(0, VS_ADMIN|VS_WATCH)) + return vs_reboot(cmd, arg); + lock_kernel(); switch (cmd) { case LINUX_REBOOT_CMD_RESTART: @@ -1078,7 +1094,7 @@ static int set_user(uid_t new_ruid, int { struct user_struct *new_user; - new_user = alloc_uid(new_ruid); + new_user = alloc_uid(vx_current_xid(), new_ruid); if (!new_user) return -EAGAIN; @@ -1432,15 +1448,18 @@ asmlinkage long sys_setpgid(pid_t pid, p { struct task_struct *p; struct task_struct *group_leader = current->group_leader; + pid_t rpgid; int err = -EINVAL; if (!pid) - pid = group_leader->pid; + pid = vx_map_pid(group_leader->pid); if (!pgid) pgid = pid; if (pgid < 0) return -EINVAL; + rpgid = vx_rmap_pid(pgid); + /* From this point forward we keep holding onto the tasklist lock * so that our parent does not change from under us. -DaveM */ @@ -1474,20 +1493,20 @@ asmlinkage long sys_setpgid(pid_t pid, p if (pgid != pid) { struct task_struct *g = - find_task_by_pid_type(PIDTYPE_PGID, pgid); + find_task_by_pid_type(PIDTYPE_PGID, rpgid); if (!g || task_session(g) != task_session(group_leader)) goto out; } - err = security_task_setpgid(p, pgid); + err = security_task_setpgid(p, rpgid); if (err) goto out; - if (process_group(p) != pgid) { + if (process_group(p) != rpgid) { detach_pid(p, PIDTYPE_PGID); - p->signal->pgrp = pgid; - attach_pid(p, PIDTYPE_PGID, find_pid(pgid)); + p->signal->pgrp = rpgid; + attach_pid(p, PIDTYPE_PGID, find_pid(rpgid)); } err = 0; @@ -1500,7 +1519,7 @@ out: asmlinkage long sys_getpgid(pid_t pid) { if (!pid) - return process_group(current); + return vx_rmap_pid(process_group(current)); else { int retval; struct task_struct *p; @@ -1512,7 +1531,7 @@ asmlinkage long sys_getpgid(pid_t pid) if (p) { retval = security_task_getpgid(p); if (!retval) - retval = process_group(p); + retval = vx_rmap_pid(process_group(p)); } read_unlock(&tasklist_lock); return retval; @@ -1863,7 +1882,7 @@ asmlinkage long sys_sethostname(char __u int errno; char tmp[__NEW_UTS_LEN]; - if (!capable(CAP_SYS_ADMIN)) + if (!vx_capable(CAP_SYS_ADMIN, VXC_SET_UTSNAME)) return -EPERM; if (len < 0 || len > __NEW_UTS_LEN) return -EINVAL; @@ -1908,7 +1927,7 @@ asmlinkage long sys_setdomainname(char _ int errno; char tmp[__NEW_UTS_LEN]; - if (!capable(CAP_SYS_ADMIN)) + if (!vx_capable(CAP_SYS_ADMIN, VXC_SET_UTSNAME)) return -EPERM; if (len < 0 || len > __NEW_UTS_LEN) return -EINVAL; @@ -1975,7 +1994,7 @@ asmlinkage long sys_setrlimit(unsigned i return -EINVAL; old_rlim = current->signal->rlim + resource; if ((new_rlim.rlim_max > old_rlim->rlim_max) && - !capable(CAP_SYS_RESOURCE)) + !vx_capable(CAP_SYS_RESOURCE, VXC_SET_RLIMIT)) return -EPERM; if (resource == RLIMIT_NOFILE && new_rlim.rlim_max > NR_OPEN) return -EPERM; diff -NurpP --minimal linux-source-2.6.22.orig/kernel/sysctl.c linux-source-2.6.22/kernel/sysctl.c --- linux-source-2.6.22.orig/kernel/sysctl.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/kernel/sysctl.c 2007-10-24 16:14:06.791620780 +0200 @@ -89,6 +89,7 @@ static int ngroups_max = NGROUPS_MAX; #ifdef CONFIG_KMOD extern char modprobe_path[]; #endif +extern char vshelper_path[]; #ifdef CONFIG_CHR_DEV_SG extern int sg_big_buff; #endif @@ -343,6 +344,15 @@ static ctl_table kern_table[] = { .strategy = &sysctl_string, }, #endif + { + .ctl_name = KERN_VSHELPER, + .procname = "vshelper", + .data = &vshelper_path, + .maxlen = 256, + .mode = 0644, + .proc_handler = &proc_dostring, + .strategy = &sysctl_string, + }, #ifdef CONFIG_CHR_DEV_SG { .ctl_name = KERN_SG_BIG_BUFF, diff -NurpP --minimal linux-source-2.6.22.orig/kernel/time.c linux-source-2.6.22/kernel/time.c --- linux-source-2.6.22.orig/kernel/time.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/kernel/time.c 2007-10-24 16:14:06.791620780 +0200 @@ -60,7 +60,7 @@ asmlinkage long sys_time(time_t __user * time_t i; struct timeval tv; - do_gettimeofday(&tv); + vx_gettimeofday(&tv); i = tv.tv_sec; if (tloc) { @@ -91,7 +91,7 @@ asmlinkage long sys_stime(time_t __user if (err) return err; - do_settimeofday(&tv); + vx_settimeofday(&tv); return 0; } @@ -101,7 +101,7 @@ asmlinkage long sys_gettimeofday(struct { if (likely(tv != NULL)) { struct timeval ktv; - do_gettimeofday(&ktv); + vx_gettimeofday(&ktv); if (copy_to_user(tv, &ktv, sizeof(ktv))) return -EFAULT; } @@ -175,7 +175,7 @@ int do_sys_settimeofday(struct timespec /* SMP safe, again the code in arch/foo/time.c should * globally block out interrupts when it runs. */ - return do_settimeofday(tv); + return vx_settimeofday(tv); } return 0; } @@ -388,7 +388,7 @@ void getnstimeofday(struct timespec *tv) { struct timeval x; - do_gettimeofday(&x); + vx_gettimeofday(&x); tv->tv_sec = x.tv_sec; tv->tv_nsec = x.tv_usec * NSEC_PER_USEC; } diff -NurpP --minimal linux-source-2.6.22.orig/kernel/timer.c linux-source-2.6.22/kernel/timer.c --- linux-source-2.6.22.orig/kernel/timer.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/kernel/timer.c 2007-10-24 16:14:06.791620780 +0200 @@ -36,6 +36,10 @@ #include #include #include +#include +#include +#include +#include #include #include @@ -921,12 +925,6 @@ asmlinkage unsigned long sys_alarm(unsig #endif -#ifndef __alpha__ - -/* - * The Alpha uses getxpid, getxuid, and getxgid instead. Maybe this - * should be moved into arch/i386 instead? - */ /** * sys_getpid - return the thread group id of the current process @@ -939,7 +937,7 @@ asmlinkage unsigned long sys_alarm(unsig */ asmlinkage long sys_getpid(void) { - return current->tgid; + return vx_map_tgid(current->tgid); } /* @@ -955,10 +953,23 @@ asmlinkage long sys_getppid(void) rcu_read_lock(); pid = rcu_dereference(current->real_parent)->tgid; rcu_read_unlock(); + return vx_map_pid(pid); +} - return pid; +#ifdef __alpha__ + +/* + * The Alpha uses getxpid, getxuid, and getxgid instead. + */ + +asmlinkage long do_getxpid(long *ppid) +{ + *ppid = sys_getppid(); + return sys_getpid(); } +#else /* _alpha_ */ + asmlinkage long sys_getuid(void) { /* Only we change this so SMP safe */ @@ -1118,6 +1129,8 @@ int do_sysinfo(struct sysinfo *info) tp.tv_nsec = tp.tv_nsec - NSEC_PER_SEC; tp.tv_sec++; } + if (vx_flags(VXF_VIRT_UPTIME, 0)) + vx_vsi_uptime(&tp, NULL); info->uptime = tp.tv_sec + (tp.tv_nsec ? 1 : 0); info->loads[0] = avenrun[0] << (SI_LOAD_SHIFT - FSHIFT); diff -NurpP --minimal linux-source-2.6.22.orig/kernel/user.c linux-source-2.6.22/kernel/user.c --- linux-source-2.6.22.orig/kernel/user.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/kernel/user.c 2007-10-24 16:14:06.791620780 +0200 @@ -23,8 +23,8 @@ #define UIDHASH_BITS (CONFIG_BASE_SMALL ? 3 : 8) #define UIDHASH_SZ (1 << UIDHASH_BITS) #define UIDHASH_MASK (UIDHASH_SZ - 1) -#define __uidhashfn(uid) (((uid >> UIDHASH_BITS) + uid) & UIDHASH_MASK) -#define uidhashentry(uid) (uidhash_table + __uidhashfn((uid))) +#define __uidhashfn(xid,uid) ((((uid) >> UIDHASH_BITS) + ((uid)^(xid))) & UIDHASH_MASK) +#define uidhashentry(xid,uid) (uidhash_table + __uidhashfn((xid),(uid))) static struct kmem_cache *uid_cachep; static struct list_head uidhash_table[UIDHASH_SZ]; @@ -66,7 +66,7 @@ static inline void uid_hash_remove(struc list_del(&up->uidhash_list); } -static inline struct user_struct *uid_hash_find(uid_t uid, struct list_head *hashent) +static inline struct user_struct *uid_hash_find(xid_t xid, uid_t uid, struct list_head *hashent) { struct list_head *up; @@ -75,7 +75,7 @@ static inline struct user_struct *uid_ha user = list_entry(up, struct user_struct, uidhash_list); - if(user->uid == uid) { + if(user->uid == uid && user->xid == xid) { atomic_inc(&user->__count); return user; } @@ -90,13 +90,13 @@ static inline struct user_struct *uid_ha * * If the user_struct could not be found, return NULL. */ -struct user_struct *find_user(uid_t uid) +struct user_struct *find_user(xid_t xid, uid_t uid) { struct user_struct *ret; unsigned long flags; spin_lock_irqsave(&uidhash_lock, flags); - ret = uid_hash_find(uid, uidhashentry(uid)); + ret = uid_hash_find(xid, uid, uidhashentry(xid, uid)); spin_unlock_irqrestore(&uidhash_lock, flags); return ret; } @@ -120,13 +120,13 @@ void free_uid(struct user_struct *up) } } -struct user_struct * alloc_uid(uid_t uid) +struct user_struct * alloc_uid(xid_t xid, uid_t uid) { - struct list_head *hashent = uidhashentry(uid); + struct list_head *hashent = uidhashentry(xid, uid); struct user_struct *up; spin_lock_irq(&uidhash_lock); - up = uid_hash_find(uid, hashent); + up = uid_hash_find(xid, uid, hashent); spin_unlock_irq(&uidhash_lock); if (!up) { @@ -136,6 +136,7 @@ struct user_struct * alloc_uid(uid_t uid if (!new) return NULL; new->uid = uid; + new->xid = xid; atomic_set(&new->__count, 1); atomic_set(&new->processes, 0); atomic_set(&new->files, 0); @@ -158,7 +159,7 @@ struct user_struct * alloc_uid(uid_t uid * on adding the same user already.. */ spin_lock_irq(&uidhash_lock); - up = uid_hash_find(uid, hashent); + up = uid_hash_find(xid, uid, hashent); if (up) { key_put(new->uid_keyring); key_put(new->session_keyring); @@ -215,7 +216,7 @@ static int __init uid_cache_init(void) /* Insert the root user immediately (init already runs as root) */ spin_lock_irq(&uidhash_lock); - uid_hash_insert(&root_user, uidhashentry(0)); + uid_hash_insert(&root_user, uidhashentry(0,0)); spin_unlock_irq(&uidhash_lock); return 0; diff -NurpP --minimal linux-source-2.6.22.orig/kernel/utsname.c linux-source-2.6.22/kernel/utsname.c --- linux-source-2.6.22.orig/kernel/utsname.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/kernel/utsname.c 2007-10-24 16:14:06.791620780 +0200 @@ -13,6 +13,7 @@ #include #include #include +#include /* * Clone a new ns copying an original utsname, setting refcount to 1 @@ -27,6 +28,7 @@ static struct uts_namespace *clone_uts_n if (ns) { memcpy(&ns->name, &old_ns->name, sizeof(ns->name)); kref_init(&ns->kref); + atomic_inc(&vs_global_uts_ns); } return ns; } @@ -58,5 +60,6 @@ void free_uts_ns(struct kref *kref) struct uts_namespace *ns; ns = container_of(kref, struct uts_namespace, kref); + atomic_dec(&vs_global_uts_ns); kfree(ns); } diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/cacct.c linux-source-2.6.22/kernel/vserver/cacct.c --- linux-source-2.6.22.orig/kernel/vserver/cacct.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/cacct.c 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,44 @@ +/* + * linux/kernel/vserver/cacct.c + * + * Virtual Server: Context Accounting + * + * Copyright (C) 2006-2007 Herbert Pötzl + * + * V0.01 added accounting stats + * + */ + +#include +#include +#include +#include +#include +#include + +#include +#include + + +int vc_sock_stat(struct vx_info *vxi, void __user *data) +{ + struct vcmd_sock_stat_v0 vc_data; + int j, field; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + field = vc_data.field; + if ((field < 0) || (field >= VXA_SOCK_SIZE)) + return -EINVAL; + + for (j = 0; j < 3; j++) { + vc_data.count[j] = vx_sock_count(&vxi->cacct, field, j); + vc_data.total[j] = vx_sock_total(&vxi->cacct, field, j); + } + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/cacct_init.h linux-source-2.6.22/kernel/vserver/cacct_init.h --- linux-source-2.6.22.orig/kernel/vserver/cacct_init.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/cacct_init.h 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,25 @@ + + +static inline void vx_info_init_cacct(struct _vx_cacct *cacct) +{ + int i, j; + + + for (i = 0; i < VXA_SOCK_SIZE; i++) { + for (j = 0; j < 3; j++) { + atomic_set(&cacct->sock[i][j].count, 0); + atomic_set(&cacct->sock[i][j].total, 0); + } + } + for (i = 0; i < 8; i++) + atomic_set(&cacct->slab[i], 0); + for (i = 0; i < 5; i++) + for (j = 0; j < 4; j++) + atomic_set(&cacct->page[i][j], 0); +} + +static inline void vx_info_exit_cacct(struct _vx_cacct *cacct) +{ + return; +} + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/cacct_proc.h linux-source-2.6.22/kernel/vserver/cacct_proc.h --- linux-source-2.6.22.orig/kernel/vserver/cacct_proc.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/cacct_proc.h 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,53 @@ +#ifndef _VX_CACCT_PROC_H +#define _VX_CACCT_PROC_H + +#include + + +#define VX_SOCKA_TOP \ + "Type\t recv #/bytes\t\t send #/bytes\t\t fail #/bytes\n" + +static inline int vx_info_proc_cacct(struct _vx_cacct *cacct, char *buffer) +{ + int i, j, length = 0; + static char *type[VXA_SOCK_SIZE] = { + "UNSPEC", "UNIX", "INET", "INET6", "PACKET", "OTHER" + }; + + length += sprintf(buffer + length, VX_SOCKA_TOP); + for (i = 0; i < VXA_SOCK_SIZE; i++) { + length += sprintf(buffer + length, "%s:", type[i]); + for (j = 0; j < 3; j++) { + length += sprintf(buffer + length, + "\t%10lu/%-10lu", + vx_sock_count(cacct, i, j), + vx_sock_total(cacct, i, j)); + } + buffer[length++] = '\n'; + } + + length += sprintf(buffer + length, "\n"); + length += sprintf(buffer + length, + "slab:\t %8u %8u %8u %8u\n", + atomic_read(&cacct->slab[1]), + atomic_read(&cacct->slab[4]), + atomic_read(&cacct->slab[0]), + atomic_read(&cacct->slab[2])); + + length += sprintf(buffer + length, "\n"); + for (i = 0; i < 5; i++) { + length += sprintf(buffer + length, + "page[%d]: %8u %8u %8u %8u\t %8u %8u %8u %8u\n", i, + atomic_read(&cacct->page[i][0]), + atomic_read(&cacct->page[i][1]), + atomic_read(&cacct->page[i][2]), + atomic_read(&cacct->page[i][3]), + atomic_read(&cacct->page[i][4]), + atomic_read(&cacct->page[i][5]), + atomic_read(&cacct->page[i][6]), + atomic_read(&cacct->page[i][7])); + } + return length; +} + +#endif /* _VX_CACCT_PROC_H */ diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/context.c linux-source-2.6.22/kernel/vserver/context.c --- linux-source-2.6.22.orig/kernel/vserver/context.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/context.c 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,1106 @@ +/* + * linux/kernel/vserver/context.c + * + * Virtual Server: Context Support + * + * Copyright (C) 2003-2007 Herbert Pötzl + * + * V0.01 context helper + * V0.02 vx_ctx_kill syscall command + * V0.03 replaced context_info calls + * V0.04 redesign of struct (de)alloc + * V0.05 rlimit basic implementation + * V0.06 task_xid and info commands + * V0.07 context flags and caps + * V0.08 switch to RCU based hash + * V0.09 revert to non RCU for now + * V0.10 and back to working RCU hash + * V0.11 and back to locking again + * V0.12 referenced context store + * V0.13 separate per cpu data + * V0.14 changed vcmds to vxi arg + * V0.15 added context stat + * V0.16 have __create claim() the vxi + * + */ + +#include +#include +#include +#include + +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include + +#include +#include + +#include "cvirt_init.h" +#include "cacct_init.h" +#include "limit_init.h" +#include "sched_init.h" + + +atomic_t vx_global_ctotal = ATOMIC_INIT(0); +atomic_t vx_global_cactive = ATOMIC_INIT(0); + + +/* now inactive context structures */ + +static struct hlist_head vx_info_inactive = HLIST_HEAD_INIT; + +static spinlock_t vx_info_inactive_lock = SPIN_LOCK_UNLOCKED; + + +/* __alloc_vx_info() + + * allocate an initialized vx_info struct + * doesn't make it visible (hash) */ + +static struct vx_info *__alloc_vx_info(xid_t xid) +{ + struct vx_info *new = NULL; + int cpu; + + vxdprintk(VXD_CBIT(xid, 0), "alloc_vx_info(%d)*", xid); + + /* would this benefit from a slab cache? */ + new = kmalloc(sizeof(struct vx_info), GFP_KERNEL); + if (!new) + return 0; + + memset(new, 0, sizeof(struct vx_info)); +#ifdef CONFIG_SMP + new->ptr_pc = alloc_percpu(struct _vx_info_pc); + if (!new->ptr_pc) + goto error; +#endif + new->vx_id = xid; + INIT_HLIST_NODE(&new->vx_hlist); + atomic_set(&new->vx_usecnt, 0); + atomic_set(&new->vx_tasks, 0); + new->vx_parent = NULL; + new->vx_state = 0; + init_waitqueue_head(&new->vx_wait); + + /* prepare reaper */ + get_task_struct(init_pid_ns.child_reaper); + new->vx_reaper = init_pid_ns.child_reaper; + + /* rest of init goes here */ + vx_info_init_limit(&new->limit); + vx_info_init_sched(&new->sched); + vx_info_init_cvirt(&new->cvirt); + vx_info_init_cacct(&new->cacct); + + /* per cpu data structures */ + for_each_possible_cpu(cpu) { + vx_info_init_sched_pc( + &vx_per_cpu(new, sched_pc, cpu), cpu); + vx_info_init_cvirt_pc( + &vx_per_cpu(new, cvirt_pc, cpu), cpu); + } + + new->vx_flags = VXF_INIT_SET; + new->vx_bcaps = CAP_INIT_EFF_SET; + new->vx_ccaps = 0; + new->vx_cap_bset = cap_bset; + + new->reboot_cmd = 0; + new->exit_code = 0; + + vxdprintk(VXD_CBIT(xid, 0), + "alloc_vx_info(%d) = %p", xid, new); + vxh_alloc_vx_info(new); + atomic_inc(&vx_global_ctotal); + return new; +#ifdef CONFIG_SMP +error: + kfree(new); + return 0; +#endif +} + +/* __dealloc_vx_info() + + * final disposal of vx_info */ + +static void __dealloc_vx_info(struct vx_info *vxi) +{ + int cpu; + + vxdprintk(VXD_CBIT(xid, 0), + "dealloc_vx_info(%p)", vxi); + vxh_dealloc_vx_info(vxi); + + vxi->vx_id = -1; + + vx_info_exit_limit(&vxi->limit); + vx_info_exit_sched(&vxi->sched); + vx_info_exit_cvirt(&vxi->cvirt); + vx_info_exit_cacct(&vxi->cacct); + + for_each_possible_cpu(cpu) { + vx_info_exit_sched_pc( + &vx_per_cpu(vxi, sched_pc, cpu), cpu); + vx_info_exit_cvirt_pc( + &vx_per_cpu(vxi, cvirt_pc, cpu), cpu); + } + + vxi->vx_state |= VXS_RELEASED; + +#ifdef CONFIG_SMP + free_percpu(vxi->ptr_pc); +#endif + kfree(vxi); + atomic_dec(&vx_global_ctotal); +} + +static void __shutdown_vx_info(struct vx_info *vxi) +{ + struct nsproxy *nsproxy; + struct fs_struct *fs; + + might_sleep(); + + vxi->vx_state |= VXS_SHUTDOWN; + vs_state_change(vxi, VSC_SHUTDOWN); + + nsproxy = xchg(&vxi->vx_nsproxy, NULL); + fs = xchg(&vxi->vx_fs, NULL); + + if (nsproxy) + put_nsproxy(nsproxy); + if (fs) + put_fs_struct(fs); +} + +/* exported stuff */ + +void free_vx_info(struct vx_info *vxi) +{ + unsigned long flags; + + /* check for reference counts first */ + BUG_ON(atomic_read(&vxi->vx_usecnt)); + BUG_ON(atomic_read(&vxi->vx_tasks)); + + /* context must not be hashed */ + BUG_ON(vx_info_state(vxi, VXS_HASHED)); + + /* context shutdown is mandatory */ + BUG_ON(!vx_info_state(vxi, VXS_SHUTDOWN)); + + BUG_ON(vxi->vx_nsproxy); + BUG_ON(vxi->vx_fs); + + spin_lock_irqsave(&vx_info_inactive_lock, flags); + hlist_del(&vxi->vx_hlist); + spin_unlock_irqrestore(&vx_info_inactive_lock, flags); + + __dealloc_vx_info(vxi); +} + + +/* hash table for vx_info hash */ + +#define VX_HASH_SIZE 13 + +static struct hlist_head vx_info_hash[VX_HASH_SIZE] = + { [0 ... VX_HASH_SIZE-1] = HLIST_HEAD_INIT }; + +static spinlock_t vx_info_hash_lock = SPIN_LOCK_UNLOCKED; + + +static inline unsigned int __hashval(xid_t xid) +{ + return (xid % VX_HASH_SIZE); +} + + + +/* __hash_vx_info() + + * add the vxi to the global hash table + * requires the hash_lock to be held */ + +static inline void __hash_vx_info(struct vx_info *vxi) +{ + struct hlist_head *head; + + vxd_assert_lock(&vx_info_hash_lock); + vxdprintk(VXD_CBIT(xid, 4), + "__hash_vx_info: %p[#%d]", vxi, vxi->vx_id); + vxh_hash_vx_info(vxi); + + /* context must not be hashed */ + BUG_ON(vx_info_state(vxi, VXS_HASHED)); + + vxi->vx_state |= VXS_HASHED; + head = &vx_info_hash[__hashval(vxi->vx_id)]; + hlist_add_head(&vxi->vx_hlist, head); + atomic_inc(&vx_global_cactive); +} + +/* __unhash_vx_info() + + * remove the vxi from the global hash table + * requires the hash_lock to be held */ + +static inline void __unhash_vx_info(struct vx_info *vxi) +{ + unsigned long flags; + + vxd_assert_lock(&vx_info_hash_lock); + vxdprintk(VXD_CBIT(xid, 4), + "__unhash_vx_info: %p[#%d.%d.%d]", vxi, vxi->vx_id, + atomic_read(&vxi->vx_usecnt), atomic_read(&vxi->vx_tasks)); + vxh_unhash_vx_info(vxi); + + /* context must be hashed */ + BUG_ON(!vx_info_state(vxi, VXS_HASHED)); + /* but without tasks */ + BUG_ON(atomic_read(&vxi->vx_tasks)); + + vxi->vx_state &= ~VXS_HASHED; + hlist_del_init(&vxi->vx_hlist); + spin_lock_irqsave(&vx_info_inactive_lock, flags); + hlist_add_head(&vxi->vx_hlist, &vx_info_inactive); + spin_unlock_irqrestore(&vx_info_inactive_lock, flags); + atomic_dec(&vx_global_cactive); +} + + +/* __lookup_vx_info() + + * requires the hash_lock to be held + * doesn't increment the vx_refcnt */ + +static inline struct vx_info *__lookup_vx_info(xid_t xid) +{ + struct hlist_head *head = &vx_info_hash[__hashval(xid)]; + struct hlist_node *pos; + struct vx_info *vxi; + + vxd_assert_lock(&vx_info_hash_lock); + hlist_for_each(pos, head) { + vxi = hlist_entry(pos, struct vx_info, vx_hlist); + + if (vxi->vx_id == xid) + goto found; + } + vxi = NULL; +found: + vxdprintk(VXD_CBIT(xid, 0), + "__lookup_vx_info(#%u): %p[#%u]", + xid, vxi, vxi ? vxi->vx_id : 0); + vxh_lookup_vx_info(vxi, xid); + return vxi; +} + + +/* __vx_dynamic_id() + + * find unused dynamic xid + * requires the hash_lock to be held */ + +static inline xid_t __vx_dynamic_id(void) +{ + static xid_t seq = MAX_S_CONTEXT; + xid_t barrier = seq; + + vxd_assert_lock(&vx_info_hash_lock); + do { + if (++seq > MAX_S_CONTEXT) + seq = MIN_D_CONTEXT; + if (!__lookup_vx_info(seq)) { + vxdprintk(VXD_CBIT(xid, 4), + "__vx_dynamic_id: [#%d]", seq); + return seq; + } + } while (barrier != seq); + return 0; +} + +#ifdef CONFIG_VSERVER_LEGACY + +/* __loc_vx_info() + + * locate or create the requested context + * get() it and if new hash it */ + +static struct vx_info *__loc_vx_info(int id, int *err) +{ + struct vx_info *new, *vxi = NULL; + + vxdprintk(VXD_CBIT(xid, 1), "loc_vx_info(%d)*", id); + + if (!(new = __alloc_vx_info(id))) { + *err = -ENOMEM; + return NULL; + } + + /* required to make dynamic xids unique */ + spin_lock(&vx_info_hash_lock); + + /* dynamic context requested */ + if (id == VX_DYNAMIC_ID) { +#ifdef CONFIG_VSERVER_DYNAMIC_IDS + id = __vx_dynamic_id(); + if (!id) { + printk(KERN_ERR "no dynamic context available.\n"); + goto out_unlock; + } + new->vx_id = id; +#else + printk(KERN_ERR "dynamic contexts disabled.\n"); + goto out_unlock; +#endif + } + /* existing context requested */ + else if ((vxi = __lookup_vx_info(id))) { + /* context in setup is not available */ + if (vxi->vx_flags & VXF_STATE_SETUP) { + vxdprintk(VXD_CBIT(xid, 0), + "loc_vx_info(%d) = %p (not available)", id, vxi); + vxi = NULL; + *err = -EBUSY; + } else { + vxdprintk(VXD_CBIT(xid, 0), + "loc_vx_info(%d) = %p (found)", id, vxi); + get_vx_info(vxi); + *err = 0; + } + goto out_unlock; + } + + /* new context requested */ + vxdprintk(VXD_CBIT(xid, 0), + "loc_vx_info(%d) = %p (new)", id, new); + __hash_vx_info(get_vx_info(new)); + vxi = new, new = NULL; + *err = 1; + +out_unlock: + spin_unlock(&vx_info_hash_lock); + vxh_loc_vx_info(vxi, id); + if (new) + __dealloc_vx_info(new); + return vxi; +} + +#endif + +/* __create_vx_info() + + * create the requested context + * get(), claim() and hash it */ + +static struct vx_info *__create_vx_info(int id) +{ + struct vx_info *new, *vxi = NULL; + + vxdprintk(VXD_CBIT(xid, 1), "create_vx_info(%d)*", id); + + if (!(new = __alloc_vx_info(id))) + return ERR_PTR(-ENOMEM); + + /* required to make dynamic xids unique */ + spin_lock(&vx_info_hash_lock); + + /* dynamic context requested */ + if (id == VX_DYNAMIC_ID) { +#ifdef CONFIG_VSERVER_DYNAMIC_IDS + id = __vx_dynamic_id(); + if (!id) { + printk(KERN_ERR "no dynamic context available.\n"); + vxi = ERR_PTR(-EAGAIN); + goto out_unlock; + } + new->vx_id = id; +#else + printk(KERN_ERR "dynamic contexts disabled.\n"); + vxi = ERR_PTR(-EINVAL); + goto out_unlock; +#endif + } + /* static context requested */ + else if ((vxi = __lookup_vx_info(id))) { + vxdprintk(VXD_CBIT(xid, 0), + "create_vx_info(%d) = %p (already there)", id, vxi); + if (vx_info_flags(vxi, VXF_STATE_SETUP, 0)) + vxi = ERR_PTR(-EBUSY); + else + vxi = ERR_PTR(-EEXIST); + goto out_unlock; + } +#ifdef CONFIG_VSERVER_DYNAMIC_IDS + /* dynamic xid creation blocker */ + else if (id >= MIN_D_CONTEXT) { + vxdprintk(VXD_CBIT(xid, 0), + "create_vx_info(%d) (dynamic rejected)", id); + vxi = ERR_PTR(-EINVAL); + goto out_unlock; + } +#endif + + /* new context */ + vxdprintk(VXD_CBIT(xid, 0), + "create_vx_info(%d) = %p (new)", id, new); + claim_vx_info(new, NULL); + __hash_vx_info(get_vx_info(new)); + vxi = new, new = NULL; + +out_unlock: + spin_unlock(&vx_info_hash_lock); + vxh_create_vx_info(IS_ERR(vxi) ? NULL : vxi, id); + if (new) + __dealloc_vx_info(new); + return vxi; +} + + +/* exported stuff */ + + +void unhash_vx_info(struct vx_info *vxi) +{ + __shutdown_vx_info(vxi); + spin_lock(&vx_info_hash_lock); + __unhash_vx_info(vxi); + spin_unlock(&vx_info_hash_lock); + __wakeup_vx_info(vxi); +} + + +/* lookup_vx_info() + + * search for a vx_info and get() it + * negative id means current */ + +struct vx_info *lookup_vx_info(int id) +{ + struct vx_info *vxi = NULL; + + if (id < 0) { + vxi = get_vx_info(current->vx_info); + } else if (id > 1) { + spin_lock(&vx_info_hash_lock); + vxi = get_vx_info(__lookup_vx_info(id)); + spin_unlock(&vx_info_hash_lock); + } + return vxi; +} + +/* xid_is_hashed() + + * verify that xid is still hashed */ + +int xid_is_hashed(xid_t xid) +{ + int hashed; + + spin_lock(&vx_info_hash_lock); + hashed = (__lookup_vx_info(xid) != NULL); + spin_unlock(&vx_info_hash_lock); + return hashed; +} + +#ifdef CONFIG_VSERVER_LEGACY + +struct vx_info *lookup_or_create_vx_info(int id) +{ + int err; + + return __loc_vx_info(id, &err); +} + +#endif + +#ifdef CONFIG_PROC_FS + +/* get_xid_list() + + * get a subset of hashed xids for proc + * assumes size is at least one */ + +int get_xid_list(int index, unsigned int *xids, int size) +{ + int hindex, nr_xids = 0; + + /* only show current and children */ + if (!vx_check(0, VS_ADMIN | VS_WATCH)) { + if (index > 0) + return 0; + xids[nr_xids] = vx_current_xid(); + return 1; + } + + for (hindex = 0; hindex < VX_HASH_SIZE; hindex++) { + struct hlist_head *head = &vx_info_hash[hindex]; + struct hlist_node *pos; + + spin_lock(&vx_info_hash_lock); + hlist_for_each(pos, head) { + struct vx_info *vxi; + + if (--index > 0) + continue; + + vxi = hlist_entry(pos, struct vx_info, vx_hlist); + xids[nr_xids] = vxi->vx_id; + if (++nr_xids >= size) { + spin_unlock(&vx_info_hash_lock); + goto out; + } + } + /* keep the lock time short */ + spin_unlock(&vx_info_hash_lock); + } +out: + return nr_xids; +} +#endif + +#ifdef CONFIG_VSERVER_DEBUG + +void dump_vx_info_inactive(int level) +{ + struct hlist_node *entry, *next; + + hlist_for_each_safe(entry, next, &vx_info_inactive) { + struct vx_info *vxi = + list_entry(entry, struct vx_info, vx_hlist); + + dump_vx_info(vxi, level); + } +} + +#endif + +int vx_migrate_user(struct task_struct *p, struct vx_info *vxi) +{ + struct user_struct *new_user, *old_user; + + if (!p || !vxi) + BUG(); + + if (vx_info_flags(vxi, VXF_INFO_PRIVATE, 0)) + return -EACCES; + + new_user = alloc_uid(vxi->vx_id, p->uid); + if (!new_user) + return -ENOMEM; + + old_user = p->user; + if (new_user != old_user) { + atomic_inc(&new_user->processes); + atomic_dec(&old_user->processes); + p->user = new_user; + } + free_uid(old_user); + return 0; +} + +void vx_mask_cap_bset(struct vx_info *vxi, struct task_struct *p) +{ + p->cap_effective &= vxi->vx_cap_bset; + p->cap_inheritable &= vxi->vx_cap_bset; + p->cap_permitted &= vxi->vx_cap_bset; +} + + +#include + +static int vx_openfd_task(struct task_struct *tsk) +{ + struct files_struct *files = tsk->files; + struct fdtable *fdt; + const unsigned long *bptr; + int count, total; + + /* no rcu_read_lock() because of spin_lock() */ + spin_lock(&files->file_lock); + fdt = files_fdtable(files); + bptr = fdt->open_fds->fds_bits; + count = fdt->max_fds / (sizeof(unsigned long) * 8); + for (total = 0; count > 0; count--) { + if (*bptr) + total += hweight_long(*bptr); + bptr++; + } + spin_unlock(&files->file_lock); + return total; +} + + +/* for *space compatibility */ + +asmlinkage long sys_unshare(unsigned long); + +/* + * migrate task to new context + * gets vxi, puts old_vxi on change + * optionally unshares namespaces (hack) + */ + +int vx_migrate_task(struct task_struct *p, struct vx_info *vxi, int unshare) +{ + struct vx_info *old_vxi; + int ret = 0; + + if (!p || !vxi) + BUG(); + + vxdprintk(VXD_CBIT(xid, 5), + "vx_migrate_task(%p,%p[#%d.%d])", p, vxi, + vxi->vx_id, atomic_read(&vxi->vx_usecnt)); + + if (vx_info_flags(vxi, VXF_INFO_PRIVATE, 0) && + !vx_info_flags(vxi, VXF_STATE_SETUP, 0)) + return -EACCES; + + if (vx_info_state(vxi, VXS_SHUTDOWN)) + return -EFAULT; + + old_vxi = task_get_vx_info(p); + if (old_vxi == vxi) + goto out; + + if (!(ret = vx_migrate_user(p, vxi))) { + int openfd; + + task_lock(p); + openfd = vx_openfd_task(p); + + if (old_vxi) { + atomic_dec(&old_vxi->cvirt.nr_threads); + atomic_dec(&old_vxi->cvirt.nr_running); + __rlim_dec(&old_vxi->limit, RLIMIT_NPROC); + /* FIXME: what about the struct files here? */ + __rlim_sub(&old_vxi->limit, VLIMIT_OPENFD, openfd); + /* account for the executable */ + __rlim_dec(&old_vxi->limit, VLIMIT_DENTRY); + } + atomic_inc(&vxi->cvirt.nr_threads); + atomic_inc(&vxi->cvirt.nr_running); + __rlim_inc(&vxi->limit, RLIMIT_NPROC); + /* FIXME: what about the struct files here? */ + __rlim_add(&vxi->limit, VLIMIT_OPENFD, openfd); + /* account for the executable */ + __rlim_inc(&vxi->limit, VLIMIT_DENTRY); + + if (old_vxi) { + release_vx_info(old_vxi, p); + clr_vx_info(&p->vx_info); + } + claim_vx_info(vxi, p); + set_vx_info(&p->vx_info, vxi); + p->xid = vxi->vx_id; + + vxdprintk(VXD_CBIT(xid, 5), + "moved task %p into vxi:%p[#%d]", + p, vxi, vxi->vx_id); + + vx_mask_cap_bset(vxi, p); + task_unlock(p); + + /* hack for *spaces to provide compatibility */ + if (unshare) { + struct nsproxy *old_nsp, *new_nsp; + + ret = unshare_nsproxy_namespaces( + CLONE_NEWUTS | CLONE_NEWIPC, + &new_nsp, NULL); + if (ret) + goto out; + + old_nsp = xchg(&p->nsproxy, new_nsp); + vx_set_space(vxi, CLONE_NEWUTS | CLONE_NEWIPC); + put_nsproxy(old_nsp); + } + } +out: + put_vx_info(old_vxi); + return ret; +} + +int vx_set_reaper(struct vx_info *vxi, struct task_struct *p) +{ + struct task_struct *old_reaper; + + if (!vxi) + return -EINVAL; + + vxdprintk(VXD_CBIT(xid, 6), + "vx_set_reaper(%p[#%d],%p[#%d,%d])", + vxi, vxi->vx_id, p, p->xid, p->pid); + + old_reaper = vxi->vx_reaper; + if (old_reaper == p) + return 0; + + /* set new child reaper */ + get_task_struct(p); + vxi->vx_reaper = p; + put_task_struct(old_reaper); + return 0; +} + +int vx_set_init(struct vx_info *vxi, struct task_struct *p) +{ + if (!vxi) + return -EINVAL; + + vxdprintk(VXD_CBIT(xid, 6), + "vx_set_init(%p[#%d],%p[#%d,%d,%d])", + vxi, vxi->vx_id, p, p->xid, p->pid, p->tgid); + + vxi->vx_flags &= ~VXF_STATE_INIT; + vxi->vx_initpid = p->tgid; + return 0; +} + +void vx_exit_init(struct vx_info *vxi, struct task_struct *p, int code) +{ + vxdprintk(VXD_CBIT(xid, 6), + "vx_exit_init(%p[#%d],%p[#%d,%d,%d])", + vxi, vxi->vx_id, p, p->xid, p->pid, p->tgid); + + vxi->exit_code = code; + vxi->vx_initpid = 0; +} + + +void vx_set_persistent(struct vx_info *vxi) +{ + vxdprintk(VXD_CBIT(xid, 6), + "vx_set_persistent(%p[#%d])", vxi, vxi->vx_id); + + get_vx_info(vxi); + claim_vx_info(vxi, NULL); +} + +void vx_clear_persistent(struct vx_info *vxi) +{ + vxdprintk(VXD_CBIT(xid, 6), + "vx_clear_persistent(%p[#%d])", vxi, vxi->vx_id); + + release_vx_info(vxi, NULL); + put_vx_info(vxi); +} + +void vx_update_persistent(struct vx_info *vxi) +{ + if (vx_info_flags(vxi, VXF_PERSISTENT, 0)) + vx_set_persistent(vxi); + else + vx_clear_persistent(vxi); +} + + +/* task must be current or locked */ + +void exit_vx_info(struct task_struct *p, int code) +{ + struct vx_info *vxi = p->vx_info; + + if (vxi) { + atomic_dec(&vxi->cvirt.nr_threads); + vx_nproc_dec(p); + + vxi->exit_code = code; + release_vx_info(vxi, p); + } +} + +void exit_vx_info_early(struct task_struct *p, int code) +{ + struct vx_info *vxi = p->vx_info; + + if (vxi) { + if (vxi->vx_initpid == p->tgid) + vx_exit_init(vxi, p, code); + if (vxi->vx_reaper == p) + vx_set_reaper(vxi, init_pid_ns.child_reaper); + } +} + + +/* vserver syscall commands below here */ + +/* taks xid and vx_info functions */ + +#include + + +int vc_task_xid(uint32_t id, void __user *data) +{ + xid_t xid; + + if (id) { + struct task_struct *tsk; + + if (!vx_check(0, VS_ADMIN | VS_WATCH)) + return -EPERM; + + read_lock(&tasklist_lock); + tsk = find_task_by_real_pid(id); + xid = (tsk) ? tsk->xid : -ESRCH; + read_unlock(&tasklist_lock); + } else + xid = vx_current_xid(); + return xid; +} + + +int vc_vx_info(struct vx_info *vxi, void __user *data) +{ + struct vcmd_vx_info_v0 vc_data; + + vc_data.xid = vxi->vx_id; + vc_data.initpid = vxi->vx_initpid; + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + + +int vc_ctx_stat(struct vx_info *vxi, void __user *data) +{ + struct vcmd_ctx_stat_v0 vc_data; + + vc_data.usecnt = atomic_read(&vxi->vx_usecnt); + vc_data.tasks = atomic_read(&vxi->vx_tasks); + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + + +/* context functions */ + +int vc_ctx_create(uint32_t xid, void __user *data) +{ + struct vcmd_ctx_create vc_data = { .flagword = VXF_INIT_SET }; + struct vx_info *new_vxi; + int ret; + + if (data && copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + if ((xid > MAX_S_CONTEXT) && (xid != VX_DYNAMIC_ID)) + return -EINVAL; + if (xid < 2) + return -EINVAL; + + new_vxi = __create_vx_info(xid); + if (IS_ERR(new_vxi)) + return PTR_ERR(new_vxi); + + /* initial flags */ + new_vxi->vx_flags = vc_data.flagword; + + ret = -ENOEXEC; + if (vs_state_change(new_vxi, VSC_STARTUP)) + goto out; + + ret = vx_migrate_task(current, new_vxi, (!data)); + if (ret) + goto out; + + /* return context id on success */ + ret = new_vxi->vx_id; + + /* get a reference for persistent contexts */ + if ((vc_data.flagword & VXF_PERSISTENT)) + vx_set_persistent(new_vxi); +out: + release_vx_info(new_vxi, NULL); + put_vx_info(new_vxi); + return ret; +} + + +int vc_ctx_migrate(struct vx_info *vxi, void __user *data) +{ + struct vcmd_ctx_migrate vc_data = { .flagword = 0 }; + int ret; + + if (data && copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + ret = vx_migrate_task(current, vxi, 0); + if (ret) + return ret; + if (vc_data.flagword & VXM_SET_INIT) + ret = vx_set_init(vxi, current); + if (ret) + return ret; + if (vc_data.flagword & VXM_SET_REAPER) + ret = vx_set_reaper(vxi, current); + return ret; +} + + +int vc_get_cflags(struct vx_info *vxi, void __user *data) +{ + struct vcmd_ctx_flags_v0 vc_data; + + vc_data.flagword = vxi->vx_flags; + + /* special STATE flag handling */ + vc_data.mask = vs_mask_flags(~0ULL, vxi->vx_flags, VXF_ONE_TIME); + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + +int vc_set_cflags(struct vx_info *vxi, void __user *data) +{ + struct vcmd_ctx_flags_v0 vc_data; + uint64_t mask, trigger; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + /* special STATE flag handling */ + mask = vs_mask_mask(vc_data.mask, vxi->vx_flags, VXF_ONE_TIME); + trigger = (mask & vxi->vx_flags) ^ (mask & vc_data.flagword); + + if (vxi == current->vx_info) { + if (trigger & VXF_STATE_SETUP) + vx_mask_cap_bset(vxi, current); + if (trigger & VXF_STATE_INIT) { + int ret; + + ret = vx_set_init(vxi, current); + if (ret) + return ret; + ret = vx_set_reaper(vxi, current); + if (ret) + return ret; + } + } + + vxi->vx_flags = vs_mask_flags(vxi->vx_flags, + vc_data.flagword, mask); + if (trigger & VXF_PERSISTENT) + vx_update_persistent(vxi); + + return 0; +} + +static int do_get_caps(struct vx_info *vxi, uint64_t *bcaps, uint64_t *ccaps) +{ + if (bcaps) + *bcaps = vxi->vx_bcaps; + if (ccaps) + *ccaps = vxi->vx_ccaps; + + return 0; +} + +int vc_get_ccaps_v0(struct vx_info *vxi, void __user *data) +{ + struct vcmd_ctx_caps_v0 vc_data; + int ret; + + ret = do_get_caps(vxi, &vc_data.bcaps, &vc_data.ccaps); + if (ret) + return ret; + vc_data.cmask = ~0ULL; + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + +int vc_get_ccaps(struct vx_info *vxi, void __user *data) +{ + struct vcmd_ctx_caps_v1 vc_data; + int ret; + + ret = do_get_caps(vxi, NULL, &vc_data.ccaps); + if (ret) + return ret; + vc_data.cmask = ~0ULL; + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + +static int do_set_caps(struct vx_info *vxi, + uint64_t bcaps, uint64_t bmask, uint64_t ccaps, uint64_t cmask) +{ + vxi->vx_bcaps = vs_mask_flags(vxi->vx_bcaps, bcaps, bmask); + vxi->vx_ccaps = vs_mask_flags(vxi->vx_ccaps, ccaps, cmask); + + return 0; +} + +int vc_set_ccaps_v0(struct vx_info *vxi, void __user *data) +{ + struct vcmd_ctx_caps_v0 vc_data; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + /* simulate old &= behaviour for bcaps */ + return do_set_caps(vxi, 0, ~vc_data.bcaps, + vc_data.ccaps, vc_data.cmask); +} + +int vc_set_ccaps(struct vx_info *vxi, void __user *data) +{ + struct vcmd_ctx_caps_v1 vc_data; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + return do_set_caps(vxi, 0, 0, vc_data.ccaps, vc_data.cmask); +} + +int vc_get_bcaps(struct vx_info *vxi, void __user *data) +{ + struct vcmd_bcaps vc_data; + int ret; + + ret = do_get_caps(vxi, &vc_data.bcaps, NULL); + if (ret) + return ret; + vc_data.bmask = ~0ULL; + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + +int vc_set_bcaps(struct vx_info *vxi, void __user *data) +{ + struct vcmd_bcaps vc_data; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + return do_set_caps(vxi, vc_data.bcaps, vc_data.bmask, 0, 0); +} + +#include + +EXPORT_SYMBOL_GPL(free_vx_info); + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/cvirt.c linux-source-2.6.22/kernel/vserver/cvirt.c --- linux-source-2.6.22.orig/kernel/vserver/cvirt.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/cvirt.c 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,304 @@ +/* + * linux/kernel/vserver/cvirt.c + * + * Virtual Server: Context Virtualization + * + * Copyright (C) 2004-2007 Herbert Pötzl + * + * V0.01 broken out from limit.c + * V0.02 added utsname stuff + * V0.03 changed vcmds to vxi arg + * + */ + +#include +#include +#include +#include +#include +#include +#include + +#include +#include + + +void vx_vsi_uptime(struct timespec *uptime, struct timespec *idle) +{ + struct vx_info *vxi = current->vx_info; + + set_normalized_timespec(uptime, + uptime->tv_sec - vxi->cvirt.bias_uptime.tv_sec, + uptime->tv_nsec - vxi->cvirt.bias_uptime.tv_nsec); + if (!idle) + return; + set_normalized_timespec(idle, + idle->tv_sec - vxi->cvirt.bias_idle.tv_sec, + idle->tv_nsec - vxi->cvirt.bias_idle.tv_nsec); + return; +} + +uint64_t vx_idle_jiffies(void) +{ + return init_task.utime + init_task.stime; +} + + + +static inline uint32_t __update_loadavg(uint32_t load, + int wsize, int delta, int n) +{ + unsigned long long calc, prev; + + /* just set it to n */ + if (unlikely(delta >= wsize)) + return (n << FSHIFT); + + calc = delta * n; + calc <<= FSHIFT; + prev = (wsize - delta); + prev *= load; + calc += prev; + do_div(calc, wsize); + return calc; +} + + +void vx_update_load(struct vx_info *vxi) +{ + uint32_t now, last, delta; + unsigned int nr_running, nr_uninterruptible; + unsigned int total; + unsigned long flags; + + spin_lock_irqsave(&vxi->cvirt.load_lock, flags); + + now = jiffies; + last = vxi->cvirt.load_last; + delta = now - last; + + if (delta < 5*HZ) + goto out; + + nr_running = atomic_read(&vxi->cvirt.nr_running); + nr_uninterruptible = atomic_read(&vxi->cvirt.nr_uninterruptible); + total = nr_running + nr_uninterruptible; + + vxi->cvirt.load[0] = __update_loadavg(vxi->cvirt.load[0], + 60*HZ, delta, total); + vxi->cvirt.load[1] = __update_loadavg(vxi->cvirt.load[1], + 5*60*HZ, delta, total); + vxi->cvirt.load[2] = __update_loadavg(vxi->cvirt.load[2], + 15*60*HZ, delta, total); + + vxi->cvirt.load_last = now; +out: + atomic_inc(&vxi->cvirt.load_updates); + spin_unlock_irqrestore(&vxi->cvirt.load_lock, flags); +} + + +/* + * Commands to do_syslog: + * + * 0 -- Close the log. Currently a NOP. + * 1 -- Open the log. Currently a NOP. + * 2 -- Read from the log. + * 3 -- Read all messages remaining in the ring buffer. + * 4 -- Read and clear all messages remaining in the ring buffer + * 5 -- Clear ring buffer. + * 6 -- Disable printk's to console + * 7 -- Enable printk's to console + * 8 -- Set level of messages printed to console + * 9 -- Return number of unread characters in the log buffer + * 10 -- Return size of the log buffer + */ +int vx_do_syslog(int type, char __user *buf, int len) +{ + int error = 0; + int do_clear = 0; + struct vx_info *vxi = current->vx_info; + struct _vx_syslog *log; + + if (!vxi) + return -EINVAL; + log = &vxi->cvirt.syslog; + + switch (type) { + case 0: /* Close log */ + case 1: /* Open log */ + break; + case 2: /* Read from log */ + error = wait_event_interruptible(log->log_wait, + (log->log_start - log->log_end)); + if (error) + break; + spin_lock_irq(&log->logbuf_lock); + spin_unlock_irq(&log->logbuf_lock); + break; + case 4: /* Read/clear last kernel messages */ + do_clear = 1; + /* fall through */ + case 3: /* Read last kernel messages */ + return 0; + + case 5: /* Clear ring buffer */ + return 0; + + case 6: /* Disable logging to console */ + case 7: /* Enable logging to console */ + case 8: /* Set level of messages printed to console */ + break; + + case 9: /* Number of chars in the log buffer */ + return 0; + case 10: /* Size of the log buffer */ + return 0; + default: + error = -EINVAL; + break; + } + return error; +} + + +/* virtual host info names */ + +static char *vx_vhi_name(struct vx_info *vxi, int id) +{ + struct nsproxy *nsproxy; + struct uts_namespace *uts; + + + if (id == VHIN_CONTEXT) + return vxi->vx_name; + + nsproxy = vxi->vx_nsproxy; + if (!nsproxy) + return NULL; + + uts = nsproxy->uts_ns; + if (!uts) + return NULL; + + switch (id) { + case VHIN_SYSNAME: + return uts->name.sysname; + case VHIN_NODENAME: + return uts->name.nodename; + case VHIN_RELEASE: + return uts->name.release; + case VHIN_VERSION: + return uts->name.version; + case VHIN_MACHINE: + return uts->name.machine; + case VHIN_DOMAINNAME: + return uts->name.domainname; + default: + return NULL; + } + return NULL; +} + +int vc_set_vhi_name(struct vx_info *vxi, void __user *data) +{ + struct vcmd_vhi_name_v0 vc_data; + char *name; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + name = vx_vhi_name(vxi, vc_data.field); + if (!name) + return -EINVAL; + + memcpy(name, vc_data.name, 65); + return 0; +} + +int vc_get_vhi_name(struct vx_info *vxi, void __user *data) +{ + struct vcmd_vhi_name_v0 vc_data; + char *name; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + name = vx_vhi_name(vxi, vc_data.field); + if (!name) + return -EINVAL; + + memcpy(vc_data.name, name, 65); + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + + +int vc_virt_stat(struct vx_info *vxi, void __user *data) +{ + struct vcmd_virt_stat_v0 vc_data; + struct _vx_cvirt *cvirt = &vxi->cvirt; + struct timespec uptime; + + do_posix_clock_monotonic_gettime(&uptime); + set_normalized_timespec(&uptime, + uptime.tv_sec - cvirt->bias_uptime.tv_sec, + uptime.tv_nsec - cvirt->bias_uptime.tv_nsec); + + vc_data.offset = timeval_to_ns(&cvirt->bias_tv); + vc_data.uptime = timespec_to_ns(&uptime); + vc_data.nr_threads = atomic_read(&cvirt->nr_threads); + vc_data.nr_running = atomic_read(&cvirt->nr_running); + vc_data.nr_uninterruptible = atomic_read(&cvirt->nr_uninterruptible); + vc_data.nr_onhold = atomic_read(&cvirt->nr_onhold); + vc_data.nr_forks = atomic_read(&cvirt->total_forks); + vc_data.load[0] = cvirt->load[0]; + vc_data.load[1] = cvirt->load[1]; + vc_data.load[2] = cvirt->load[2]; + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + + +#ifdef CONFIG_VSERVER_VTIME + +/* virtualized time base */ + +void vx_gettimeofday(struct timeval *tv) +{ + do_gettimeofday(tv); + if (!vx_flags(VXF_VIRT_TIME, 0)) + return; + + tv->tv_sec += current->vx_info->cvirt.bias_tv.tv_sec; + tv->tv_usec += current->vx_info->cvirt.bias_tv.tv_usec; + + if (tv->tv_usec >= USEC_PER_SEC) { + tv->tv_sec++; + tv->tv_usec -= USEC_PER_SEC; + } else if (tv->tv_usec < 0) { + tv->tv_sec--; + tv->tv_usec += USEC_PER_SEC; + } +} + +int vx_settimeofday(struct timespec *ts) +{ + struct timeval tv; + + if (!vx_flags(VXF_VIRT_TIME, 0)) + return do_settimeofday(ts); + + do_gettimeofday(&tv); + current->vx_info->cvirt.bias_tv.tv_sec = + ts->tv_sec - tv.tv_sec; + current->vx_info->cvirt.bias_tv.tv_usec = + (ts->tv_nsec/NSEC_PER_USEC) - tv.tv_usec; + return 0; +} + +#endif + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/cvirt_init.h linux-source-2.6.22/kernel/vserver/cvirt_init.h --- linux-source-2.6.22.orig/kernel/vserver/cvirt_init.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/cvirt_init.h 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,69 @@ + + +extern uint64_t vx_idle_jiffies(void); + +static inline void vx_info_init_cvirt(struct _vx_cvirt *cvirt) +{ + uint64_t idle_jiffies = vx_idle_jiffies(); + uint64_t nsuptime; + + do_posix_clock_monotonic_gettime(&cvirt->bias_uptime); + nsuptime = (unsigned long long)cvirt->bias_uptime.tv_sec + * NSEC_PER_SEC + cvirt->bias_uptime.tv_nsec; + cvirt->bias_clock = nsec_to_clock_t(nsuptime); + cvirt->bias_tv.tv_sec = 0; + cvirt->bias_tv.tv_usec = 0; + + jiffies_to_timespec(idle_jiffies, &cvirt->bias_idle); + atomic_set(&cvirt->nr_threads, 0); + atomic_set(&cvirt->nr_running, 0); + atomic_set(&cvirt->nr_uninterruptible, 0); + atomic_set(&cvirt->nr_onhold, 0); + + spin_lock_init(&cvirt->load_lock); + cvirt->load_last = jiffies; + atomic_set(&cvirt->load_updates, 0); + cvirt->load[0] = 0; + cvirt->load[1] = 0; + cvirt->load[2] = 0; + atomic_set(&cvirt->total_forks, 0); + + spin_lock_init(&cvirt->syslog.logbuf_lock); + init_waitqueue_head(&cvirt->syslog.log_wait); + cvirt->syslog.log_start = 0; + cvirt->syslog.log_end = 0; + cvirt->syslog.con_start = 0; + cvirt->syslog.logged_chars = 0; +} + +static inline +void vx_info_init_cvirt_pc(struct _vx_cvirt_pc *cvirt_pc, int cpu) +{ + // cvirt_pc->cpustat = { 0 }; +} + +static inline void vx_info_exit_cvirt(struct _vx_cvirt *cvirt) +{ + int value; + + vxwprintk((value = atomic_read(&cvirt->nr_threads)), + "!!! cvirt: %p[nr_threads] = %d on exit.", + cvirt, value); + vxwprintk((value = atomic_read(&cvirt->nr_running)), + "!!! cvirt: %p[nr_running] = %d on exit.", + cvirt, value); + vxwprintk((value = atomic_read(&cvirt->nr_uninterruptible)), + "!!! cvirt: %p[nr_uninterruptible] = %d on exit.", + cvirt, value); + vxwprintk((value = atomic_read(&cvirt->nr_onhold)), + "!!! cvirt: %p[nr_onhold] = %d on exit.", + cvirt, value); + return; +} + +static inline +void vx_info_exit_cvirt_pc(struct _vx_cvirt_pc *cvirt_pc, int cpu) +{ + return; +} + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/cvirt_proc.h linux-source-2.6.22/kernel/vserver/cvirt_proc.h --- linux-source-2.6.22.orig/kernel/vserver/cvirt_proc.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/cvirt_proc.h 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,133 @@ +#ifndef _VX_CVIRT_PROC_H +#define _VX_CVIRT_PROC_H + +#include +#include +#include +#include + + +static inline +int vx_info_proc_nsproxy(struct nsproxy *nsproxy, char *buffer) +{ + struct mnt_namespace *ns; + struct uts_namespace *uts; + struct ipc_namespace *ipc; + struct vfsmount *mnt; + char *path, *root; + int length = 0; + + if (!nsproxy) + goto out; + + length += sprintf(buffer + length, + "NSProxy:\t%p [%p,%p,%p]\n", + nsproxy, nsproxy->mnt_ns, + nsproxy->uts_ns, nsproxy->ipc_ns); + + ns = nsproxy->mnt_ns; + if (!ns) + goto skip_ns; + + path = kmalloc(PATH_MAX, GFP_KERNEL); + if (!path) + goto skip_ns; + + mnt = ns->root; + root = d_path(mnt->mnt_root, mnt->mnt_parent, path, PATH_MAX - 2); + length += sprintf(buffer + length, + "Namespace:\t%p [#%u]\n" + "RootPath:\t%s\n", + ns, atomic_read(&ns->count), + root); + kfree(path); +skip_ns: + + uts = nsproxy->uts_ns; + if (!uts) + goto skip_uts; + + length += sprintf(buffer + length, + "SysName:\t%.*s\n" + "NodeName:\t%.*s\n" + "Release:\t%.*s\n" + "Version:\t%.*s\n" + "Machine:\t%.*s\n" + "DomainName:\t%.*s\n", + __NEW_UTS_LEN, uts->name.sysname, + __NEW_UTS_LEN, uts->name.nodename, + __NEW_UTS_LEN, uts->name.release, + __NEW_UTS_LEN, uts->name.version, + __NEW_UTS_LEN, uts->name.machine, + __NEW_UTS_LEN, uts->name.domainname); +skip_uts: + + ipc = nsproxy->ipc_ns; + if (!ipc) + goto skip_ipc; + + length += sprintf(buffer + length, + "SEMS:\t\t%d %d %d %d %d\n" + "MSG:\t\t%d %d %d\n" + "SHM:\t\t%lu %lu %d %d\n", + ipc->sem_ctls[0], ipc->sem_ctls[1], + ipc->sem_ctls[2], ipc->sem_ctls[3], + ipc->used_sems, + ipc->msg_ctlmax, ipc->msg_ctlmnb, ipc->msg_ctlmni, + (unsigned long)ipc->shm_ctlmax, + (unsigned long)ipc->shm_ctlall, + ipc->shm_ctlmni, ipc->shm_tot); +skip_ipc: +out: + return length; +} + + +#include + +#define LOAD_INT(x) ((x) >> FSHIFT) +#define LOAD_FRAC(x) LOAD_INT(((x) & (FIXED_1 - 1)) * 100) + +static inline +int vx_info_proc_cvirt(struct _vx_cvirt *cvirt, char *buffer) +{ + int length = 0; + int a, b, c; + + length += sprintf(buffer + length, + "BiasUptime:\t%lu.%02lu\n", + (unsigned long)cvirt->bias_uptime.tv_sec, + (cvirt->bias_uptime.tv_nsec / (NSEC_PER_SEC / 100))); + + a = cvirt->load[0] + (FIXED_1 / 200); + b = cvirt->load[1] + (FIXED_1 / 200); + c = cvirt->load[2] + (FIXED_1 / 200); + length += sprintf(buffer + length, + "nr_threads:\t%d\n" + "nr_running:\t%d\n" + "nr_unintr:\t%d\n" + "nr_onhold:\t%d\n" + "load_updates:\t%d\n" + "loadavg:\t%d.%02d %d.%02d %d.%02d\n" + "total_forks:\t%d\n", + atomic_read(&cvirt->nr_threads), + atomic_read(&cvirt->nr_running), + atomic_read(&cvirt->nr_uninterruptible), + atomic_read(&cvirt->nr_onhold), + atomic_read(&cvirt->load_updates), + LOAD_INT(a), LOAD_FRAC(a), + LOAD_INT(b), LOAD_FRAC(b), + LOAD_INT(c), LOAD_FRAC(c), + atomic_read(&cvirt->total_forks)); + return length; +} + +static inline +int vx_info_proc_cvirt_pc(struct _vx_cvirt_pc *cvirt_pc, + char *buffer, int cpu) +{ + int length = 0; + return length; +} + +#endif /* _VX_CVIRT_PROC_H */ diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/debug.c linux-source-2.6.22/kernel/vserver/debug.c --- linux-source-2.6.22.orig/kernel/vserver/debug.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/debug.c 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,35 @@ +/* + * kernel/vserver/debug.c + * + * Copyright (C) 2005-2007 Herbert Pötzl + * + * V0.01 vx_info dump support + * + */ + +#include +#include +#include +#include + +#include + + +void dump_vx_info(struct vx_info *vxi, int level) +{ + printk("vx_info %p[#%d, %d.%d, %4x]\n", vxi, vxi->vx_id, + atomic_read(&vxi->vx_usecnt), + atomic_read(&vxi->vx_tasks), + vxi->vx_state); + if (level > 0) { + __dump_vx_limit(&vxi->limit); + __dump_vx_sched(&vxi->sched); + __dump_vx_cvirt(&vxi->cvirt); + __dump_vx_cacct(&vxi->cacct); + } + printk("---\n"); +} + + +EXPORT_SYMBOL_GPL(dump_vx_info); + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/dlimit.c linux-source-2.6.22/kernel/vserver/dlimit.c --- linux-source-2.6.22.orig/kernel/vserver/dlimit.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/dlimit.c 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,526 @@ +/* + * linux/kernel/vserver/dlimit.c + * + * Virtual Server: Context Disk Limits + * + * Copyright (C) 2004-2007 Herbert Pötzl + * + * V0.01 initial version + * V0.02 compat32 splitup + * + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include + +/* __alloc_dl_info() + + * allocate an initialized dl_info struct + * doesn't make it visible (hash) */ + +static struct dl_info *__alloc_dl_info(struct super_block *sb, tag_t tag) +{ + struct dl_info *new = NULL; + + vxdprintk(VXD_CBIT(dlim, 5), + "alloc_dl_info(%p,%d)*", sb, tag); + + /* would this benefit from a slab cache? */ + new = kmalloc(sizeof(struct dl_info), GFP_KERNEL); + if (!new) + return 0; + + memset(new, 0, sizeof(struct dl_info)); + new->dl_tag = tag; + new->dl_sb = sb; + INIT_RCU_HEAD(&new->dl_rcu); + INIT_HLIST_NODE(&new->dl_hlist); + spin_lock_init(&new->dl_lock); + atomic_set(&new->dl_refcnt, 0); + atomic_set(&new->dl_usecnt, 0); + + /* rest of init goes here */ + + vxdprintk(VXD_CBIT(dlim, 4), + "alloc_dl_info(%p,%d) = %p", sb, tag, new); + return new; +} + +/* __dealloc_dl_info() + + * final disposal of dl_info */ + +static void __dealloc_dl_info(struct dl_info *dli) +{ + vxdprintk(VXD_CBIT(dlim, 4), + "dealloc_dl_info(%p)", dli); + + dli->dl_hlist.next = LIST_POISON1; + dli->dl_tag = -1; + dli->dl_sb = 0; + + BUG_ON(atomic_read(&dli->dl_usecnt)); + BUG_ON(atomic_read(&dli->dl_refcnt)); + + kfree(dli); +} + + +/* hash table for dl_info hash */ + +#define DL_HASH_SIZE 13 + +struct hlist_head dl_info_hash[DL_HASH_SIZE]; + +static spinlock_t dl_info_hash_lock = SPIN_LOCK_UNLOCKED; + + +static inline unsigned int __hashval(struct super_block *sb, tag_t tag) +{ + return ((tag ^ (unsigned long)sb) % DL_HASH_SIZE); +} + + + +/* __hash_dl_info() + + * add the dli to the global hash table + * requires the hash_lock to be held */ + +static inline void __hash_dl_info(struct dl_info *dli) +{ + struct hlist_head *head; + + vxdprintk(VXD_CBIT(dlim, 6), + "__hash_dl_info: %p[#%d]", dli, dli->dl_tag); + get_dl_info(dli); + head = &dl_info_hash[__hashval(dli->dl_sb, dli->dl_tag)]; + hlist_add_head_rcu(&dli->dl_hlist, head); +} + +/* __unhash_dl_info() + + * remove the dli from the global hash table + * requires the hash_lock to be held */ + +static inline void __unhash_dl_info(struct dl_info *dli) +{ + vxdprintk(VXD_CBIT(dlim, 6), + "__unhash_dl_info: %p[#%d]", dli, dli->dl_tag); + hlist_del_rcu(&dli->dl_hlist); + put_dl_info(dli); +} + + +/* __lookup_dl_info() + + * requires the rcu_read_lock() + * doesn't increment the dl_refcnt */ + +static inline struct dl_info *__lookup_dl_info(struct super_block *sb, tag_t tag) +{ + struct hlist_head *head = &dl_info_hash[__hashval(sb, tag)]; + struct hlist_node *pos; + struct dl_info *dli; + + hlist_for_each_entry_rcu(dli, pos, head, dl_hlist) { + + if (dli->dl_tag == tag && dli->dl_sb == sb) { + return dli; + } + } + return NULL; +} + + +struct dl_info *locate_dl_info(struct super_block *sb, tag_t tag) +{ + struct dl_info *dli; + + rcu_read_lock(); + dli = get_dl_info(__lookup_dl_info(sb, tag)); + vxdprintk(VXD_CBIT(dlim, 7), + "locate_dl_info(%p,#%d) = %p", sb, tag, dli); + rcu_read_unlock(); + return dli; +} + +void rcu_free_dl_info(struct rcu_head *head) +{ + struct dl_info *dli = container_of(head, struct dl_info, dl_rcu); + int usecnt, refcnt; + + BUG_ON(!dli || !head); + + usecnt = atomic_read(&dli->dl_usecnt); + BUG_ON(usecnt < 0); + + refcnt = atomic_read(&dli->dl_refcnt); + BUG_ON(refcnt < 0); + + vxdprintk(VXD_CBIT(dlim, 3), + "rcu_free_dl_info(%p)", dli); + if (!usecnt) + __dealloc_dl_info(dli); + else + printk("!!! rcu didn't free\n"); +} + + + + +static int do_addrem_dlimit(uint32_t id, const char __user *name, + uint32_t flags, int add) +{ + struct nameidata nd; + int ret; + + ret = user_path_walk_link(name, &nd); + if (!ret) { + struct super_block *sb; + struct dl_info *dli; + + ret = -EINVAL; + if (!nd.dentry->d_inode) + goto out_release; + if (!(sb = nd.dentry->d_inode->i_sb)) + goto out_release; + + if (add) { + dli = __alloc_dl_info(sb, id); + spin_lock(&dl_info_hash_lock); + + ret = -EEXIST; + if (__lookup_dl_info(sb, id)) + goto out_unlock; + __hash_dl_info(dli); + dli = NULL; + } else { + spin_lock(&dl_info_hash_lock); + dli = __lookup_dl_info(sb, id); + + ret = -ESRCH; + if (!dli) + goto out_unlock; + __unhash_dl_info(dli); + } + ret = 0; + out_unlock: + spin_unlock(&dl_info_hash_lock); + if (add && dli) + __dealloc_dl_info(dli); + out_release: + path_release(&nd); + } + return ret; +} + +int vc_add_dlimit(uint32_t id, void __user *data) +{ + struct vcmd_ctx_dlimit_base_v0 vc_data; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + return do_addrem_dlimit(id, vc_data.name, vc_data.flags, 1); +} + +int vc_rem_dlimit(uint32_t id, void __user *data) +{ + struct vcmd_ctx_dlimit_base_v0 vc_data; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + return do_addrem_dlimit(id, vc_data.name, vc_data.flags, 0); +} + +#ifdef CONFIG_COMPAT + +int vc_add_dlimit_x32(uint32_t id, void __user *data) +{ + struct vcmd_ctx_dlimit_base_v0_x32 vc_data; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + return do_addrem_dlimit(id, + compat_ptr(vc_data.name_ptr), vc_data.flags, 1); +} + +int vc_rem_dlimit_x32(uint32_t id, void __user *data) +{ + struct vcmd_ctx_dlimit_base_v0_x32 vc_data; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + return do_addrem_dlimit(id, + compat_ptr(vc_data.name_ptr), vc_data.flags, 0); +} + +#endif /* CONFIG_COMPAT */ + + +static inline +int do_set_dlimit(uint32_t id, const char __user *name, + uint32_t space_used, uint32_t space_total, + uint32_t inodes_used, uint32_t inodes_total, + uint32_t reserved, uint32_t flags) +{ + struct nameidata nd; + int ret; + + ret = user_path_walk_link(name, &nd); + if (!ret) { + struct super_block *sb; + struct dl_info *dli; + + ret = -EINVAL; + if (!nd.dentry->d_inode) + goto out_release; + if (!(sb = nd.dentry->d_inode->i_sb)) + goto out_release; + if ((reserved != CDLIM_KEEP && + reserved > 100) || + (inodes_used != CDLIM_KEEP && + inodes_used > inodes_total) || + (space_used != CDLIM_KEEP && + space_used > space_total)) + goto out_release; + + ret = -ESRCH; + dli = locate_dl_info(sb, id); + if (!dli) + goto out_release; + + spin_lock(&dli->dl_lock); + + if (inodes_used != CDLIM_KEEP) + dli->dl_inodes_used = inodes_used; + if (inodes_total != CDLIM_KEEP) + dli->dl_inodes_total = inodes_total; + if (space_used != CDLIM_KEEP) { + dli->dl_space_used = space_used; + dli->dl_space_used <<= 10; + } + if (space_total == CDLIM_INFINITY) + dli->dl_space_total = DLIM_INFINITY; + else if (space_total != CDLIM_KEEP) { + dli->dl_space_total = space_total; + dli->dl_space_total <<= 10; + } + if (reserved != CDLIM_KEEP) + dli->dl_nrlmult = (1 << 10) * (100 - reserved) / 100; + + spin_unlock(&dli->dl_lock); + + put_dl_info(dli); + ret = 0; + + out_release: + path_release(&nd); + } + return ret; +} + +int vc_set_dlimit(uint32_t id, void __user *data) +{ + struct vcmd_ctx_dlimit_v0 vc_data; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + return do_set_dlimit(id, vc_data.name, + vc_data.space_used, vc_data.space_total, + vc_data.inodes_used, vc_data.inodes_total, + vc_data.reserved, vc_data.flags); +} + +#ifdef CONFIG_COMPAT + +int vc_set_dlimit_x32(uint32_t id, void __user *data) +{ + struct vcmd_ctx_dlimit_v0_x32 vc_data; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + return do_set_dlimit(id, compat_ptr(vc_data.name_ptr), + vc_data.space_used, vc_data.space_total, + vc_data.inodes_used, vc_data.inodes_total, + vc_data.reserved, vc_data.flags); +} + +#endif /* CONFIG_COMPAT */ + + +static inline +int do_get_dlimit(uint32_t id, const char __user *name, + uint32_t *space_used, uint32_t *space_total, + uint32_t *inodes_used, uint32_t *inodes_total, + uint32_t *reserved, uint32_t *flags) +{ + struct nameidata nd; + int ret; + + ret = user_path_walk_link(name, &nd); + if (!ret) { + struct super_block *sb; + struct dl_info *dli; + + ret = -EINVAL; + if (!nd.dentry->d_inode) + goto out_release; + if (!(sb = nd.dentry->d_inode->i_sb)) + goto out_release; + + ret = -ESRCH; + dli = locate_dl_info(sb, id); + if (!dli) + goto out_release; + + spin_lock(&dli->dl_lock); + *inodes_used = dli->dl_inodes_used; + *inodes_total = dli->dl_inodes_total; + *space_used = dli->dl_space_used >> 10; + if (dli->dl_space_total == DLIM_INFINITY) + *space_total = CDLIM_INFINITY; + else + *space_total = dli->dl_space_total >> 10; + + *reserved = 100 - ((dli->dl_nrlmult * 100 + 512) >> 10); + spin_unlock(&dli->dl_lock); + + put_dl_info(dli); + ret = -EFAULT; + + ret = 0; + out_release: + path_release(&nd); + } + return ret; +} + + +int vc_get_dlimit(uint32_t id, void __user *data) +{ + struct vcmd_ctx_dlimit_v0 vc_data; + int ret; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + ret = do_get_dlimit(id, vc_data.name, + &vc_data.space_used, &vc_data.space_total, + &vc_data.inodes_used, &vc_data.inodes_total, + &vc_data.reserved, &vc_data.flags); + if (ret) + return ret; + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + +#ifdef CONFIG_COMPAT + +int vc_get_dlimit_x32(uint32_t id, void __user *data) +{ + struct vcmd_ctx_dlimit_v0_x32 vc_data; + int ret; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + ret = do_get_dlimit(id, compat_ptr(vc_data.name_ptr), + &vc_data.space_used, &vc_data.space_total, + &vc_data.inodes_used, &vc_data.inodes_total, + &vc_data.reserved, &vc_data.flags); + if (ret) + return ret; + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + +#endif /* CONFIG_COMPAT */ + + +void vx_vsi_statfs(struct super_block *sb, struct kstatfs *buf) +{ + struct dl_info *dli; + __u64 blimit, bfree, bavail; + __u32 ifree; + + dli = locate_dl_info(sb, dx_current_tag()); + if (!dli) + return; + + spin_lock(&dli->dl_lock); + if (dli->dl_inodes_total == (unsigned long)DLIM_INFINITY) + goto no_ilim; + + /* reduce max inodes available to limit */ + if (buf->f_files > dli->dl_inodes_total) + buf->f_files = dli->dl_inodes_total; + + ifree = dli->dl_inodes_total - dli->dl_inodes_used; + /* reduce free inodes to min */ + if (ifree < buf->f_ffree) + buf->f_ffree = ifree; + +no_ilim: + if (dli->dl_space_total == DLIM_INFINITY) + goto no_blim; + + blimit = dli->dl_space_total >> sb->s_blocksize_bits; + + if (dli->dl_space_total < dli->dl_space_used) + bfree = 0; + else + bfree = (dli->dl_space_total - dli->dl_space_used) + >> sb->s_blocksize_bits; + + bavail = ((dli->dl_space_total >> 10) * dli->dl_nrlmult); + if (bavail < dli->dl_space_used) + bavail = 0; + else + bavail = (bavail - dli->dl_space_used) + >> sb->s_blocksize_bits; + + /* reduce max space available to limit */ + if (buf->f_blocks > blimit) + buf->f_blocks = blimit; + + /* reduce free space to min */ + if (bfree < buf->f_bfree) + buf->f_bfree = bfree; + + /* reduce avail space to min */ + if (bavail < buf->f_bavail) + buf->f_bavail = bavail; + +no_blim: + spin_unlock(&dli->dl_lock); + put_dl_info(dli); + + return; +} + +#include + +EXPORT_SYMBOL_GPL(locate_dl_info); +EXPORT_SYMBOL_GPL(rcu_free_dl_info); + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/helper.c linux-source-2.6.22/kernel/vserver/helper.c --- linux-source-2.6.22.orig/kernel/vserver/helper.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/helper.c 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,208 @@ +/* + * linux/kernel/vserver/helper.c + * + * Virtual Context Support + * + * Copyright (C) 2004-2007 Herbert Pötzl + * + * V0.01 basic helper + * + */ + +#include +#include +#include +#include +#include +#include +#include + +#include +#include + + +char vshelper_path[255] = "/sbin/vshelper"; + + +static int do_vshelper(char *name, char *argv[], char *envp[], int sync) +{ + int ret; + + if ((ret = call_usermodehelper(name, argv, envp, sync))) { + printk( KERN_WARNING + "%s: (%s %s) returned %s with %d\n", + name, argv[1], argv[2], + sync ? "sync" : "async", ret); + } + vxdprintk(VXD_CBIT(switch, 4), + "%s: (%s %s) returned %s with %d", + name, argv[1], argv[2], sync ? "sync" : "async", ret); + return ret; +} + +/* + * vshelper path is set via /proc/sys + * invoked by vserver sys_reboot(), with + * the following arguments + * + * argv [0] = vshelper_path; + * argv [1] = action: "restart", "halt", "poweroff", ... + * argv [2] = context identifier + * + * envp [*] = type-specific parameters + */ + +long vs_reboot_helper(struct vx_info *vxi, int cmd, void __user *arg) +{ + char id_buf[8], cmd_buf[16]; + char uid_buf[16], pid_buf[16]; + int ret; + + char *argv[] = {vshelper_path, NULL, id_buf, 0}; + char *envp[] = {"HOME=/", "TERM=linux", + "PATH=/sbin:/usr/sbin:/bin:/usr/bin", + uid_buf, pid_buf, cmd_buf, 0}; + + if (vx_info_state(vxi, VXS_HELPER)) + return -EAGAIN; + vxi->vx_state |= VXS_HELPER; + + snprintf(id_buf, sizeof(id_buf)-1, "%d", vxi->vx_id); + + snprintf(cmd_buf, sizeof(cmd_buf)-1, "VS_CMD=%08x", cmd); + snprintf(uid_buf, sizeof(uid_buf)-1, "VS_UID=%d", current->uid); + snprintf(pid_buf, sizeof(pid_buf)-1, "VS_PID=%d", current->pid); + + switch (cmd) { + case LINUX_REBOOT_CMD_RESTART: + argv[1] = "restart"; + break; + + case LINUX_REBOOT_CMD_HALT: + argv[1] = "halt"; + break; + + case LINUX_REBOOT_CMD_POWER_OFF: + argv[1] = "poweroff"; + break; + + case LINUX_REBOOT_CMD_SW_SUSPEND: + argv[1] = "swsusp"; + break; + + default: + vxi->vx_state &= ~VXS_HELPER; + return 0; + } + +#ifndef CONFIG_VSERVER_LEGACY + ret = do_vshelper(vshelper_path, argv, envp, 1); +#else + ret = do_vshelper(vshelper_path, argv, envp, 0); +#endif + vxi->vx_state &= ~VXS_HELPER; + __wakeup_vx_info(vxi); + return (ret) ? -EPERM : 0; +} + + +long vs_reboot(unsigned int cmd, void __user *arg) +{ + struct vx_info *vxi = current->vx_info; + long ret = 0; + + vxdprintk(VXD_CBIT(misc, 5), + "vs_reboot(%p[#%d],%d)", + vxi, vxi ? vxi->vx_id : 0, cmd); + + ret = vs_reboot_helper(vxi, cmd, arg); + if (ret) + return ret; + + vxi->reboot_cmd = cmd; + if (vx_info_flags(vxi, VXF_REBOOT_KILL, 0)) { + switch (cmd) { + case LINUX_REBOOT_CMD_RESTART: + case LINUX_REBOOT_CMD_HALT: + case LINUX_REBOOT_CMD_POWER_OFF: + vx_info_kill(vxi, 0, SIGKILL); + vx_info_kill(vxi, 1, SIGKILL); + default: + break; + } + } + return 0; +} + + +/* + * argv [0] = vshelper_path; + * argv [1] = action: "startup", "shutdown" + * argv [2] = context identifier + * + * envp [*] = type-specific parameters + */ + +long vs_state_change(struct vx_info *vxi, unsigned int cmd) +{ + char id_buf[8], cmd_buf[16]; + char *argv[] = {vshelper_path, NULL, id_buf, 0}; + char *envp[] = {"HOME=/", "TERM=linux", + "PATH=/sbin:/usr/sbin:/bin:/usr/bin", cmd_buf, 0}; + + if (!vx_info_flags(vxi, VXF_SC_HELPER, 0)) + return 0; + + snprintf(id_buf, sizeof(id_buf)-1, "%d", vxi->vx_id); + snprintf(cmd_buf, sizeof(cmd_buf)-1, "VS_CMD=%08x", cmd); + + switch (cmd) { + case VSC_STARTUP: + argv[1] = "startup"; + break; + case VSC_SHUTDOWN: + argv[1] = "shutdown"; + break; + default: + return 0; + } + + return do_vshelper(vshelper_path, argv, envp, 1); +} + + +/* + * argv [0] = vshelper_path; + * argv [1] = action: "netup", "netdown" + * argv [2] = context identifier + * + * envp [*] = type-specific parameters + */ + +long vs_net_change(struct nx_info *nxi, unsigned int cmd) +{ + char id_buf[8], cmd_buf[16]; + char *argv[] = {vshelper_path, NULL, id_buf, 0}; + char *envp[] = {"HOME=/", "TERM=linux", + "PATH=/sbin:/usr/sbin:/bin:/usr/bin", cmd_buf, 0}; + + if (!nx_info_flags(nxi, NXF_SC_HELPER, 0)) + return 0; + + snprintf(id_buf, sizeof(id_buf)-1, "%d", nxi->nx_id); + snprintf(cmd_buf, sizeof(cmd_buf)-1, "VS_CMD=%08x", cmd); + + switch (cmd) { + case VSC_NETUP: + argv[1] = "netup"; + break; + case VSC_NETDOWN: + argv[1] = "netdown"; + break; + default: + return 0; + } + + return do_vshelper(vshelper_path, argv, envp, 1); +} + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/history.c linux-source-2.6.22/kernel/vserver/history.c --- linux-source-2.6.22.orig/kernel/vserver/history.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/history.c 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,264 @@ +/* + * kernel/vserver/history.c + * + * Virtual Context History Backtrace + * + * Copyright (C) 2004-2007 Herbert Pötzl + * + * V0.01 basic structure + * V0.02 hash/unhash and trace + * V0.03 preemption fixes + * + */ + +#include +#include +#include +#include + +#include +#include +#include + +#include +#include +#include +#include + + +#ifdef CONFIG_VSERVER_HISTORY +#define VXH_SIZE CONFIG_VSERVER_HISTORY_SIZE +#else +#define VXH_SIZE 64 +#endif + +struct _vx_history { + unsigned int counter; + + struct _vx_hist_entry entry[VXH_SIZE + 1]; +}; + + +DEFINE_PER_CPU(struct _vx_history, vx_history_buffer); + +unsigned volatile int vxh_active = 1; + +static atomic_t sequence = ATOMIC_INIT(0); + + +/* vxh_advance() + + * requires disabled preemption */ + +struct _vx_hist_entry *vxh_advance(void *loc) +{ + unsigned int cpu = smp_processor_id(); + struct _vx_history *hist = &per_cpu(vx_history_buffer, cpu); + struct _vx_hist_entry *entry; + unsigned int index; + + index = vxh_active ? (hist->counter++ % VXH_SIZE) : VXH_SIZE; + entry = &hist->entry[index]; + + entry->seq = atomic_inc_return(&sequence); + entry->loc = loc; + return entry; +} + +EXPORT_SYMBOL_GPL(vxh_advance); + + +#define VXH_LOC_FMTS "(#%04x,*%d):%p" + +#define VXH_LOC_ARGS(e) (e)->seq, cpu, (e)->loc + + +#define VXH_VXI_FMTS "%p[#%d,%d.%d]" + +#define VXH_VXI_ARGS(e) (e)->vxi.ptr, \ + (e)->vxi.ptr ? (e)->vxi.xid : 0, \ + (e)->vxi.ptr ? (e)->vxi.usecnt : 0, \ + (e)->vxi.ptr ? (e)->vxi.tasks : 0 + +void vxh_dump_entry(struct _vx_hist_entry *e, unsigned cpu) +{ + switch (e->type) { + case VXH_THROW_OOPS: + printk( VXH_LOC_FMTS " oops \n", VXH_LOC_ARGS(e)); + break; + + case VXH_GET_VX_INFO: + case VXH_PUT_VX_INFO: + printk( VXH_LOC_FMTS " %s_vx_info " VXH_VXI_FMTS "\n", + VXH_LOC_ARGS(e), + (e->type == VXH_GET_VX_INFO) ? "get" : "put", + VXH_VXI_ARGS(e)); + break; + + case VXH_INIT_VX_INFO: + case VXH_SET_VX_INFO: + case VXH_CLR_VX_INFO: + printk( VXH_LOC_FMTS " %s_vx_info " VXH_VXI_FMTS " @%p\n", + VXH_LOC_ARGS(e), + (e->type == VXH_INIT_VX_INFO) ? "init" : + ((e->type == VXH_SET_VX_INFO) ? "set" : "clr"), + VXH_VXI_ARGS(e), e->sc.data); + break; + + case VXH_CLAIM_VX_INFO: + case VXH_RELEASE_VX_INFO: + printk( VXH_LOC_FMTS " %s_vx_info " VXH_VXI_FMTS " @%p\n", + VXH_LOC_ARGS(e), + (e->type == VXH_CLAIM_VX_INFO) ? "claim" : "release", + VXH_VXI_ARGS(e), e->sc.data); + break; + + case VXH_ALLOC_VX_INFO: + case VXH_DEALLOC_VX_INFO: + printk( VXH_LOC_FMTS " %s_vx_info " VXH_VXI_FMTS "\n", + VXH_LOC_ARGS(e), + (e->type == VXH_ALLOC_VX_INFO) ? "alloc" : "dealloc", + VXH_VXI_ARGS(e)); + break; + + case VXH_HASH_VX_INFO: + case VXH_UNHASH_VX_INFO: + printk( VXH_LOC_FMTS " __%s_vx_info " VXH_VXI_FMTS "\n", + VXH_LOC_ARGS(e), + (e->type == VXH_HASH_VX_INFO) ? "hash" : "unhash", + VXH_VXI_ARGS(e)); + break; + + case VXH_LOC_VX_INFO: + case VXH_LOOKUP_VX_INFO: + case VXH_CREATE_VX_INFO: + printk( VXH_LOC_FMTS " __%s_vx_info [#%d] -> " VXH_VXI_FMTS "\n", + VXH_LOC_ARGS(e), + (e->type == VXH_CREATE_VX_INFO) ? "create" : + ((e->type == VXH_LOC_VX_INFO) ? "loc" : "lookup"), + e->ll.arg, VXH_VXI_ARGS(e)); + break; + } +} + +static void __vxh_dump_history(void) +{ + unsigned int i, cpu; + + printk("History:\tSEQ: %8x\tNR_CPUS: %d\n", + atomic_read(&sequence), NR_CPUS); + + for (i = 0; i < VXH_SIZE; i++) { + for_each_online_cpu(cpu) { + struct _vx_history *hist = + &per_cpu(vx_history_buffer, cpu); + unsigned int index = (hist->counter - i) % VXH_SIZE; + struct _vx_hist_entry *entry = &hist->entry[index]; + + vxh_dump_entry(entry, cpu); + } + } +} + +void vxh_dump_history(void) +{ + vxh_active = 0; +#ifdef CONFIG_SMP + local_irq_enable(); + smp_send_stop(); + local_irq_disable(); +#endif + __vxh_dump_history(); +} + + +/* vserver syscall commands below here */ + + +int vc_dump_history(uint32_t id) +{ + vxh_active = 0; + __vxh_dump_history(); + vxh_active = 1; + + return 0; +} + + +int do_read_history(struct __user _vx_hist_entry *data, + int cpu, uint32_t *index, uint32_t *count) +{ + int pos, ret = 0; + struct _vx_history *hist = &per_cpu(vx_history_buffer, cpu); + int end = hist->counter; + int start = end - VXH_SIZE + 2; + int idx = *index; + + /* special case: get current pos */ + if (!*count) { + *index = end; + return 0; + } + + /* have we lost some data? */ + if (idx < start) + idx = start; + + for (pos = 0; (pos < *count) && (idx < end); pos++, idx++) { + struct _vx_hist_entry *entry = + &hist->entry[idx % VXH_SIZE]; + + /* send entry to userspace */ + ret = copy_to_user(&data[pos], entry, sizeof(*entry)); + if (ret) + break; + } + /* save new index and count */ + *index = idx; + *count = pos; + return ret ? ret : (*index < end); +} + +int vc_read_history(uint32_t id, void __user *data) +{ + struct vcmd_read_history_v0 vc_data; + int ret; + + if (id >= NR_CPUS) + return -EINVAL; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + ret = do_read_history((struct __user _vx_hist_entry *)vc_data.data, + id, &vc_data.index, &vc_data.count); + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return ret; +} + +#ifdef CONFIG_COMPAT + +int vc_read_history_x32(uint32_t id, void __user *data) +{ + struct vcmd_read_history_v0_x32 vc_data; + int ret; + + if (id >= NR_CPUS) + return -EINVAL; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + ret = do_read_history((struct __user _vx_hist_entry *) + compat_ptr(vc_data.data_ptr), + id, &vc_data.index, &vc_data.count); + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return ret; +} + +#endif /* CONFIG_COMPAT */ + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/init.c linux-source-2.6.22/kernel/vserver/init.c --- linux-source-2.6.22.orig/kernel/vserver/init.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/init.c 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,47 @@ +/* + * linux/kernel/init.c + * + * Virtual Server Init + * + * Copyright (C) 2004-2007 Herbert Pötzl + * + * V0.01 basic structure + * + */ + +#include +#include +#include + +int vserver_register_sysctl(void); +void vserver_unregister_sysctl(void); + + +static int __init init_vserver(void) +{ + int ret = 0; + +#ifdef CONFIG_VSERVER_DEBUG + vserver_register_sysctl(); +#endif + return ret; +} + + +static void __exit exit_vserver(void) +{ + +#ifdef CONFIG_VSERVER_DEBUG + vserver_unregister_sysctl(); +#endif + return; +} + +/* FIXME: GFP_ZONETYPES gone +long vx_slab[GFP_ZONETYPES]; */ +long vx_area; + + +module_init(init_vserver); +module_exit(exit_vserver); + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/inode.c linux-source-2.6.22/kernel/vserver/inode.c --- linux-source-2.6.22.orig/kernel/vserver/inode.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/inode.c 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,417 @@ +/* + * linux/kernel/vserver/inode.c + * + * Virtual Server: File System Support + * + * Copyright (C) 2004-2007 Herbert Pötzl + * + * V0.01 separated from vcontext V0.05 + * + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include + + +static int __vc_get_iattr(struct inode *in, uint32_t *tag, uint32_t *flags, uint32_t *mask) +{ + struct proc_dir_entry *entry; + + if (!in || !in->i_sb) + return -ESRCH; + + *flags = IATTR_TAG + | (IS_BARRIER(in) ? IATTR_BARRIER : 0) + | (IS_IUNLINK(in) ? IATTR_IUNLINK : 0) + | (IS_IMMUTABLE(in) ? IATTR_IMMUTABLE : 0); + *mask = IATTR_IUNLINK | IATTR_IMMUTABLE; + + if (S_ISDIR(in->i_mode)) + *mask |= IATTR_BARRIER; + + if (IS_TAGGED(in)) { + *tag = in->i_tag; + *mask |= IATTR_TAG; + } + + switch (in->i_sb->s_magic) { + case PROC_SUPER_MAGIC: + entry = PROC_I(in)->pde; + + /* check for specific inodes? */ + if (entry) + *mask |= IATTR_FLAGS; + if (entry) + *flags |= (entry->vx_flags & IATTR_FLAGS); + else + *flags |= (PROC_I(in)->vx_flags & IATTR_FLAGS); + break; + + case DEVPTS_SUPER_MAGIC: + *tag = in->i_tag; + *mask |= IATTR_TAG; + break; + + default: + break; + } + return 0; +} + +int vc_get_iattr(void __user *data) +{ + struct nameidata nd; + struct vcmd_ctx_iattr_v1 vc_data = { .xid = -1 }; + int ret; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + ret = user_path_walk_link(vc_data.name, &nd); + if (!ret) { + ret = __vc_get_iattr(nd.dentry->d_inode, + &vc_data.xid, &vc_data.flags, &vc_data.mask); + path_release(&nd); + } + if (ret) + return ret; + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + ret = -EFAULT; + return ret; +} + +#ifdef CONFIG_COMPAT + +int vc_get_iattr_x32(void __user *data) +{ + struct nameidata nd; + struct vcmd_ctx_iattr_v1_x32 vc_data = { .xid = -1 }; + int ret; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + ret = user_path_walk_link(compat_ptr(vc_data.name_ptr), &nd); + if (!ret) { + ret = __vc_get_iattr(nd.dentry->d_inode, + &vc_data.xid, &vc_data.flags, &vc_data.mask); + path_release(&nd); + } + if (ret) + return ret; + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + ret = -EFAULT; + return ret; +} + +#endif /* CONFIG_COMPAT */ + + +int vc_fget_iattr(uint32_t fd, void __user *data) +{ + struct file *filp; + struct vcmd_ctx_fiattr_v0 vc_data = { .xid = -1 }; + int ret; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + filp = fget(fd); + if (!filp || !filp->f_dentry || !filp->f_dentry->d_inode) + return -EBADF; + + ret = __vc_get_iattr(filp->f_dentry->d_inode, + &vc_data.xid, &vc_data.flags, &vc_data.mask); + + fput(filp); + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + ret = -EFAULT; + return ret; +} + + +static int __vc_set_iattr(struct dentry *de, uint32_t *tag, uint32_t *flags, uint32_t *mask) +{ + struct inode *in = de->d_inode; + int error = 0, is_proc = 0, has_tag = 0; + struct iattr attr = { 0 }; + + if (!in || !in->i_sb) + return -ESRCH; + + is_proc = (in->i_sb->s_magic == PROC_SUPER_MAGIC); + if ((*mask & IATTR_FLAGS) && !is_proc) + return -EINVAL; + + has_tag = IS_TAGGED(in) || + (in->i_sb->s_magic == DEVPTS_SUPER_MAGIC); + if ((*mask & IATTR_TAG) && !has_tag) + return -EINVAL; + + mutex_lock(&in->i_mutex); + if (*mask & IATTR_TAG) { + attr.ia_tag = *tag; + attr.ia_valid |= ATTR_TAG; + } + + if (*mask & IATTR_FLAGS) { + struct proc_dir_entry *entry = PROC_I(in)->pde; + unsigned int iflags = PROC_I(in)->vx_flags; + + iflags = (iflags & ~(*mask & IATTR_FLAGS)) + | (*flags & IATTR_FLAGS); + PROC_I(in)->vx_flags = iflags; + if (entry) + entry->vx_flags = iflags; + } + + if (*mask & (IATTR_BARRIER | IATTR_IUNLINK | IATTR_IMMUTABLE)) { + if (*mask & IATTR_IMMUTABLE) { + if (*flags & IATTR_IMMUTABLE) + in->i_flags |= S_IMMUTABLE; + else + in->i_flags &= ~S_IMMUTABLE; + } + if (*mask & IATTR_IUNLINK) { + if (*flags & IATTR_IUNLINK) + in->i_flags |= S_IUNLINK; + else + in->i_flags &= ~S_IUNLINK; + } + if (S_ISDIR(in->i_mode) && (*mask & IATTR_BARRIER)) { + if (*flags & IATTR_BARRIER) + in->i_flags |= S_BARRIER; + else + in->i_flags &= ~S_BARRIER; + } + if (in->i_op && in->i_op->sync_flags) { + error = in->i_op->sync_flags(in); + if (error) + goto out; + } + } + + if (attr.ia_valid) { + if (in->i_op && in->i_op->setattr) + error = in->i_op->setattr(de, &attr); + else { + error = inode_change_ok(in, &attr); + if (!error) + error = inode_setattr(in, &attr); + } + } + +out: + mutex_unlock(&in->i_mutex); + return error; +} + +int vc_set_iattr(void __user *data) +{ + struct nameidata nd; + struct vcmd_ctx_iattr_v1 vc_data; + int ret; + + if (!capable(CAP_LINUX_IMMUTABLE)) + return -EPERM; + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + ret = user_path_walk_link(vc_data.name, &nd); + if (!ret) { + ret = __vc_set_iattr(nd.dentry, + &vc_data.xid, &vc_data.flags, &vc_data.mask); + path_release(&nd); + } + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + ret = -EFAULT; + return ret; +} + +#ifdef CONFIG_COMPAT + +int vc_set_iattr_x32(void __user *data) +{ + struct nameidata nd; + struct vcmd_ctx_iattr_v1_x32 vc_data; + int ret; + + if (!capable(CAP_LINUX_IMMUTABLE)) + return -EPERM; + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + ret = user_path_walk_link(compat_ptr(vc_data.name_ptr), &nd); + if (!ret) { + ret = __vc_set_iattr(nd.dentry, + &vc_data.xid, &vc_data.flags, &vc_data.mask); + path_release(&nd); + } + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + ret = -EFAULT; + return ret; +} + +#endif /* CONFIG_COMPAT */ + +int vc_fset_iattr(uint32_t fd, void __user *data) +{ + struct file *filp; + struct vcmd_ctx_fiattr_v0 vc_data; + int ret; + + if (!capable(CAP_LINUX_IMMUTABLE)) + return -EPERM; + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + filp = fget(fd); + if (!filp || !filp->f_dentry || !filp->f_dentry->d_inode) + return -EBADF; + + ret = __vc_set_iattr(filp->f_dentry, &vc_data.xid, + &vc_data.flags, &vc_data.mask); + + fput(filp); + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return ret; +} + +#ifdef CONFIG_VSERVER_LEGACY + +#define PROC_DYNAMIC_FIRST 0xF0000000UL + +int vx_proc_ioctl(struct inode *inode, struct file *filp, + unsigned int cmd, unsigned long arg) +{ + struct proc_dir_entry *entry; + int error = 0; + int flags; + + if (inode->i_ino < PROC_DYNAMIC_FIRST) + return -ENOTTY; + + entry = PROC_I(inode)->pde; + if (!entry) + return -ENOTTY; + + switch(cmd) { + case FIOC_GETXFLG: { + /* fixme: if stealth, return -ENOTTY */ + error = -EPERM; + flags = entry->vx_flags; + if (capable(CAP_CONTEXT)) + error = put_user(flags, (int __user *) arg); + break; + } + case FIOC_SETXFLG: { + /* fixme: if stealth, return -ENOTTY */ + error = -EPERM; + if (!capable(CAP_CONTEXT)) + break; + error = -EROFS; + if (IS_RDONLY(inode)) + break; + error = -EFAULT; + if (get_user(flags, (int __user *) arg)) + break; + error = 0; + entry->vx_flags = flags; + break; + } + default: + return -ENOTTY; + } + return error; +} +#endif /* CONFIG_VSERVER_LEGACY */ + +#ifdef CONFIG_PROPAGATE + +int dx_parse_tag(char *string, tag_t *tag, int remove) +{ + static match_table_t tokens = { + {1, "tagid=%u"}, + {0, NULL} + }; + substring_t args[MAX_OPT_ARGS]; + int token, option = 0; + + if (!string) + return 0; + + token = match_token(string, tokens, args); + if (token && tag && !match_int(args, &option)) + *tag = option; + + vxdprintk(VXD_CBIT(tag, 7), + "dx_parse_tag(»%s«): %d:#%d", + string, token, option); + + if ((token == 1) && remove) { + char *p = strstr(string, "tagid="); + char *q = p; + + if (p) { + while (*q != '\0' && *q != ',') + q++; + while (*q) + *p++ = *q++; + while (*p) + *p++ = '\0'; + } + } + return token; +} + +void __dx_propagate_tag(struct nameidata *nd, struct inode *inode) +{ + tag_t new_tag = 0; + struct vfsmount *mnt; + int propagate; + + if (!nd) + return; + mnt = nd->mnt; + if (!mnt) + return; + + propagate = (mnt->mnt_flags & MNT_TAGID); + if (propagate) + new_tag = mnt->mnt_tag; + + vxdprintk(VXD_CBIT(tag, 7), + "dx_propagate_tag(%p[#%lu.%d]): %d,%d", + inode, inode->i_ino, inode->i_tag, + new_tag, (propagate) ? 1 : 0); + + if (propagate) + inode->i_tag = new_tag; +} + +#include + +EXPORT_SYMBOL_GPL(__dx_propagate_tag); + +#endif /* CONFIG_PROPAGATE */ + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/Kconfig linux-source-2.6.22/kernel/vserver/Kconfig --- linux-source-2.6.22.orig/kernel/vserver/Kconfig 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/Kconfig 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,284 @@ +# +# Linux VServer configuration +# + +menu "Linux VServer" + +config VSERVER_LEGACY + bool "Enable Legacy Kernel API" + depends on EXPERIMENTAL + default n + help + This enables the legacy API used in vs1.xx, maintaining + compatibility with older vserver tools, and guest images + that are configured using the legacy method. + +config VSERVER_LEGACY_VERSION + bool "Show a Legacy Version ID" + depends on VSERVER_LEGACY + default n + help + This shows a special legacy version to very old tools + which do not handle the current version correctly. + + Warning: recent tools are not able to utilize the + newer APIs when this is enabled, so some features will + not be available. Better avoid it, unless you really, + really need it for backwards compatibility. + +config VSERVER_DYNAMIC_IDS + bool "Enable dynamic context IDs" + depends on EXPERIMENTAL && VSERVER_LEGACY + default n + help + This enables support of in kernel dynamic context IDs, + which is deprecated and will probably be removed in the + next release. + +config VSERVER_LEGACYNET + bool "Enable Legacy Networking Kernel API" + depends on EXPERIMENTAL + default n + help + This enables the legacy networking API which is used + by older tools (pre 0.30.210) to set up the network + context (chbind). + +config VSERVER_REMAP_SADDR + bool "Remap Source IP Address" + depends on EXPERIMENTAL + default n + help + This allows to remap the source IP address of 'local' + connections from 127.0.0.1 to the first assigned + guest IP. + +config VSERVER_COWBL + bool "Enable COW Immutable Link Breaking" + default y + help + This enables the COW (Copy-On-Write) link break code. + It allows you to treat unified files like normal files + when writing to them (which will implicitely break the + link and create a copy of the unified file) + +config VSERVER_VTIME + bool "Enable Virtualized Guest Time" + depends on EXPERIMENTAL + default n + help + This enables per guest time offsets to allow for + adjusting the system clock individually per guest. + this adds some overhead to the time functions and + therefore should not be enabled without good reason. + +config VSERVER_PROC_SECURE + bool "Enable Proc Security" + depends on PROC_FS + default y + help + This configures ProcFS security to initially hide + non-process entries for all contexts except the main and + spectator context (i.e. for all guests), which is a secure + default. + + (note: on 1.2x the entries were visible by default) + +config VSERVER_HARDCPU + bool "Enable Hard CPU Limits" + default y + help + Activate the Hard CPU Limits + + This will compile in code that allows the Token Bucket + Scheduler to put processes on hold when a context's + tokens are depleted (provided that its per-context + sched_hard flag is set). + + Processes belonging to that context will not be able + to consume CPU resources again until a per-context + configured minimum of tokens has been reached. + +config VSERVER_IDLETIME + bool "Avoid idle CPUs by skipping Time" + depends on VSERVER_HARDCPU + default y + help + This option allows the scheduler to artificially + advance time (per cpu) when otherwise the idle + task would be scheduled, thus keeping the cpu + busy and sharing the available resources among + certain contexts. + +config VSERVER_IDLELIMIT + bool "Limit the IDLE task" + depends on VSERVER_HARDCPU + default n + help + Limit the idle slices, so the the next context + will be scheduled as soon as possible. + + This might improve interactivity and latency, but + will also marginally increase scheduling overhead. + +choice + prompt "Persistent Inode Tagging" + default TAGGING_ID24 + help + This adds persistent context information to filesystems + mounted with the tagxid option. Tagging is a requirement + for per-context disk limits and per-context quota. + + +config TAGGING_NONE + bool "Disabled" + help + do not store per-context information in inodes. + +config TAGGING_UID16 + bool "UID16/GID32" + help + reduces UID to 16 bit, but leaves GID at 32 bit. + +config TAGGING_GID16 + bool "UID32/GID16" + help + reduces GID to 16 bit, but leaves UID at 32 bit. + +config TAGGING_ID24 + bool "UID24/GID24" + help + uses the upper 8bit from UID and GID for XID tagging + which leaves 24bit for UID/GID each, which should be + more than sufficient for normal use. + +config TAGGING_INTERN + bool "UID32/GID32" + help + this uses otherwise reserved inode fields in the on + disk representation, which limits the use to a few + filesystems (currently ext2 and ext3) + +endchoice + +config TAG_NFSD + bool "Tag NFSD User Auth and Files" + default n + help + Enable this if you do want the in-kernel NFS + Server to use the tagging specified above. + (will require patched clients too) + +config PROPAGATE + bool "Enable Inode Tag Propagation" + default n + depends on EXPERIMENTAL + help + This allows for the tagid= mount option to specify + a tagid which is to be used for the entire mount + tree. + +config VSERVER_PRIVACY + bool "Honor Privacy Aspects of Guests" + default y + help + When enabled, most context checks will disallow + access to structures assigned to a specific context, + like ptys or loop devices. + +config VSERVER_CONTEXTS + int "Maximum number of Contexts (1-65533)" if EMBEDDED + range 1 65533 + default "768" if 64BIT + default "256" + help + This setting will optimize certain data structures + and memory allocations according to the expected + maximum. + + note: this is not a strict upper limit. + +config VSERVER_WARN + bool "VServer Warnings" + default y + help + This enables various runtime warnings, which will + notify about potential manipulation attempts or + resource shortage. It is generally considered to + be a good idea to have that enabled. + +config VSERVER_DEBUG + bool "VServer Debugging Code" + default n + help + Set this to yes if you want to be able to activate + debugging output at runtime. It adds a very small + overhead to all vserver related functions and + increases the kernel size by about 20k. + +config VSERVER_HISTORY + bool "VServer History Tracing" + depends on VSERVER_DEBUG + default n + help + Set this to yes if you want to record the history of + linux-vserver activities, so they can be replayed in + the event of a kernel panic or oops. + +config VSERVER_HISTORY_SIZE + int "Per-CPU History Size (32-65536)" + depends on VSERVER_HISTORY + range 32 65536 + default 64 + help + This allows you to specify the number of entries in + the per-CPU history buffer. + +config VSERVER_MONITOR + bool "VServer Scheduling Monitor" + depends on VSERVER_DEBUG + default n + help + Set this to yes if you want to record the scheduling + decisions, so that they can be relayed to userspace + for detailed analysis. + +config VSERVER_MONITOR_SIZE + int "Per-CPU Monitor Queue Size (32-65536)" + depends on VSERVER_MONITOR + range 32 65536 + default 1024 + help + This allows you to specify the number of entries in + the per-CPU scheduling monitor buffer. + +config VSERVER_MONITOR_SYNC + int "Per-CPU Monitor Sync Interval (0-65536)" + depends on VSERVER_MONITOR + range 0 65536 + default 256 + help + This allows you to specify the interval in ticks + when a time sync entry is inserted. + +endmenu + + +config VSERVER + bool + default y + select UTS_NS + select SYSVIPC + select IPC_NS + +config VSERVER_SECURITY + bool + depends on SECURITY + default y + select SECURITY_CAPABILITIES + +config VSERVER_NGNET + bool + depends on EXPERIMENTAL && !VSERVER_LEGACYNET + default y + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/legacy.c linux-source-2.6.22/kernel/vserver/legacy.c --- linux-source-2.6.22.orig/kernel/vserver/legacy.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/legacy.c 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,114 @@ +/* + * linux/kernel/vserver/legacy.c + * + * Virtual Server: Legacy Funtions + * + * Copyright (C) 2001-2003 Jacques Gelinas + * Copyright (C) 2003-2007 Herbert Pötzl + * + * V0.01 broken out from vcontext.c V0.05 + * V0.02 updated to spaces *sigh* + * + */ + +#include +#include +#include +#include +#include +// #include + +#include +#include + + +extern int vx_set_init(struct vx_info *, struct task_struct *); + +static int vx_set_initpid(struct vx_info *vxi, int pid) +{ + struct task_struct *init; + + init = find_task_by_real_pid(pid); + if (!init) + return -ESRCH; + return vx_set_init(vxi, init); +} + +int vc_new_s_context(uint32_t ctx, void __user *data) +{ + int ret = -ENOMEM; + struct vcmd_new_s_context_v1 vc_data; + struct vx_info *new_vxi; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + /* legacy hack, will be removed soon */ + if (ctx == -2) { + /* assign flags and initpid */ + if (!current->vx_info) + return -EINVAL; + ret = 0; + if (vc_data.flags & VX_INFO_INIT) + ret = vx_set_initpid(current->vx_info, current->tgid); + if (ret == 0) { + /* We keep the same vx_id, but lower the capabilities */ + current->vx_info->vx_bcaps &= (~vc_data.remove_cap); + ret = vx_current_xid(); + current->vx_info->vx_flags |= vc_data.flags; + } + return ret; + } + + if (!vx_check(0, VS_ADMIN) || !capable(CAP_SYS_ADMIN) + /* might make sense in the future, or not ... */ + || vx_flags(VX_INFO_PRIVATE, 0)) + return -EPERM; + + /* ugly hack for Spectator */ + if (ctx == 1) { + current->xid = 1; + return 0; + } + + if (((ctx > MAX_S_CONTEXT) && (ctx != VX_DYNAMIC_ID)) || + (ctx == 0)) + return -EINVAL; + + if ((ctx == VX_DYNAMIC_ID) || (ctx < MIN_D_CONTEXT)) + new_vxi = lookup_or_create_vx_info(ctx); + else + new_vxi = lookup_vx_info(ctx); + + if (!new_vxi) + return -EINVAL; + + ret = -EPERM; + if (!vx_info_flags(new_vxi, VXF_STATE_SETUP, 0) && + vx_info_flags(new_vxi, VX_INFO_PRIVATE, 0)) + goto out_put; + + ret = vx_migrate_task(current, new_vxi, + vx_info_flags(new_vxi, VXF_STATE_SETUP, 0)); + new_vxi->vx_flags &= ~VXF_STATE_SETUP; + + if (ret == 0) { + current->vx_info->vx_bcaps &= (~vc_data.remove_cap); + new_vxi->vx_flags |= vc_data.flags; + if (vc_data.flags & VX_INFO_INIT) + vx_set_initpid(new_vxi, current->tgid); + if (vc_data.flags & VX_INFO_NAMESPACE) + vx_set_space(new_vxi, CLONE_NEWNS | CLONE_FS); + if (vc_data.flags & VX_INFO_NPROC) + __rlim_set(&new_vxi->limit, RLIMIT_NPROC, + current->signal->rlim[RLIMIT_NPROC].rlim_max); + + /* tweak some defaults for legacy */ + new_vxi->vx_flags |= (VXF_HIDE_NETIF | VXF_INFO_INIT); + ret = new_vxi->vx_id; + } +out_put: + put_vx_info(new_vxi); + return ret; +} + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/legacynet.c linux-source-2.6.22/kernel/vserver/legacynet.c --- linux-source-2.6.22.orig/kernel/vserver/legacynet.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/legacynet.c 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,85 @@ + +/* + * linux/kernel/vserver/legacynet.c + * + * Virtual Server: Legacy Network Funtions + * + * Copyright (C) 2001-2003 Jacques Gelinas + * Copyright (C) 2003-2007 Herbert Pötzl + * + * V0.01 broken out from legacy.c + * + */ + +#include +#include +#include +#include +// #include +#include + +#include +#include + + +extern struct nx_info *create_nx_info(void); + +/* set ipv4 root (syscall) */ + +int vc_set_ipv4root(uint32_t nbip, void __user *data) +{ + int i, err = -EPERM; + struct vcmd_set_ipv4root_v3 vc_data; + struct nx_info *new_nxi, *nxi = current->nx_info; + + if (nbip < 0 || nbip > NB_IPV4ROOT) + return -EINVAL; + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + if (!nxi || nxi->ipv4[0] == 0 || capable(CAP_NET_ADMIN)) + /* We are allowed to change everything */ + err = 0; + else if (nxi) { + int found = 0; + + /* We are allowed to select a subset of the currently + installed IP numbers. No new one are allowed + We can't change the broadcast address though */ + for (i = 0; i < nbip; i++) { + int j; + __u32 nxip = vc_data.nx_mask_pair[i].ip; + for (j = 0; j < nxi->nbipv4; j++) { + if (nxip == nxi->ipv4[j]) { + found++; + break; + } + } + } + if ((found == nbip) && + (vc_data.broadcast == nxi->v4_bcast)) + err = 0; + } + if (err) + return err; + + new_nxi = create_nx_info(); + if (IS_ERR(new_nxi)) + return -EINVAL; + + new_nxi->nbipv4 = nbip; + for (i = 0; i < nbip; i++) { + new_nxi->ipv4[i] = vc_data.nx_mask_pair[i].ip; + new_nxi->mask[i] = vc_data.nx_mask_pair[i].mask; + } + new_nxi->v4_bcast = vc_data.broadcast; + if (nxi) + printk("!!! switching nx_info %p->%p\n", nxi, new_nxi); + + nx_migrate_task(current, new_nxi); + release_nx_info(new_nxi, NULL); + put_nx_info(new_nxi); + return 0; +} + + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/limit.c linux-source-2.6.22/kernel/vserver/limit.c --- linux-source-2.6.22.orig/kernel/vserver/limit.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/limit.c 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,319 @@ +/* + * linux/kernel/vserver/limit.c + * + * Virtual Server: Context Limits + * + * Copyright (C) 2004-2007 Herbert Pötzl + * + * V0.01 broken out from vcontext V0.05 + * V0.02 changed vcmds to vxi arg + * + */ + +#include +#include +#include +#include +#include +#include + +#include +#include + + +const char *vlimit_name[NUM_LIMITS] = { + [RLIMIT_CPU] = "CPU", + [RLIMIT_RSS] = "RSS", + [RLIMIT_NPROC] = "NPROC", + [RLIMIT_NOFILE] = "NOFILE", + [RLIMIT_MEMLOCK] = "VML", + [RLIMIT_AS] = "VM", + [RLIMIT_LOCKS] = "LOCKS", + [RLIMIT_SIGPENDING] = "SIGP", + [RLIMIT_MSGQUEUE] = "MSGQ", + + [VLIMIT_NSOCK] = "NSOCK", + [VLIMIT_OPENFD] = "OPENFD", + [VLIMIT_ANON] = "ANON", + [VLIMIT_SHMEM] = "SHMEM", + [VLIMIT_DENTRY] = "DENTRY", +}; + +EXPORT_SYMBOL_GPL(vlimit_name); + +#define MASK_ENTRY(x) (1 << (x)) + +const struct vcmd_ctx_rlimit_mask_v0 vlimit_mask = { + /* minimum */ + 0 + , /* softlimit */ + MASK_ENTRY( RLIMIT_RSS ) | + MASK_ENTRY( VLIMIT_ANON ) | + 0 + , /* maximum */ + MASK_ENTRY( RLIMIT_RSS ) | + MASK_ENTRY( RLIMIT_NPROC ) | + MASK_ENTRY( RLIMIT_NOFILE ) | + MASK_ENTRY( RLIMIT_MEMLOCK ) | + MASK_ENTRY( RLIMIT_AS ) | + MASK_ENTRY( RLIMIT_LOCKS ) | + MASK_ENTRY( RLIMIT_MSGQUEUE ) | + + MASK_ENTRY( VLIMIT_NSOCK ) | + MASK_ENTRY( VLIMIT_OPENFD ) | + MASK_ENTRY( VLIMIT_ANON ) | + MASK_ENTRY( VLIMIT_SHMEM ) | + MASK_ENTRY( VLIMIT_DENTRY ) | + 0 +}; + /* accounting only */ +uint32_t account_mask = + MASK_ENTRY( VLIMIT_SEMARY ) | + MASK_ENTRY( VLIMIT_NSEMS ) | + MASK_ENTRY( VLIMIT_MAPPED ) | + 0; + + +static int is_valid_vlimit(int id) +{ + uint32_t mask = vlimit_mask.minimum | + vlimit_mask.softlimit | vlimit_mask.maximum; + return mask & (1 << id); +} + +static int is_accounted_vlimit(int id) +{ + if (is_valid_vlimit(id)) + return 1; + return account_mask & (1 << id); +} + + +static inline uint64_t vc_get_soft(struct vx_info *vxi, int id) +{ + rlim_t limit = __rlim_soft(&vxi->limit, id); + return VX_VLIM(limit); +} + +static inline uint64_t vc_get_hard(struct vx_info *vxi, int id) +{ + rlim_t limit = __rlim_hard(&vxi->limit, id); + return VX_VLIM(limit); +} + +static int do_get_rlimit(struct vx_info *vxi, uint32_t id, + uint64_t *minimum, uint64_t *softlimit, uint64_t *maximum) +{ + if (!is_valid_vlimit(id)) + return -EINVAL; + + if (minimum) + *minimum = CRLIM_UNSET; + if (softlimit) + *softlimit = vc_get_soft(vxi, id); + if (maximum) + *maximum = vc_get_hard(vxi, id); + return 0; +} + +int vc_get_rlimit(struct vx_info *vxi, void __user *data) +{ + struct vcmd_ctx_rlimit_v0 vc_data; + int ret; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + ret = do_get_rlimit(vxi, vc_data.id, + &vc_data.minimum, &vc_data.softlimit, &vc_data.maximum); + if (ret) + return ret; + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + +static int do_set_rlimit(struct vx_info *vxi, uint32_t id, + uint64_t minimum, uint64_t softlimit, uint64_t maximum) +{ + if (!is_valid_vlimit(id)) + return -EINVAL; + + if (maximum != CRLIM_KEEP) + __rlim_hard(&vxi->limit, id) = VX_RLIM(maximum); + if (softlimit != CRLIM_KEEP) + __rlim_soft(&vxi->limit, id) = VX_RLIM(softlimit); + + /* clamp soft limit */ + if (__rlim_soft(&vxi->limit, id) > __rlim_hard(&vxi->limit, id)) + __rlim_soft(&vxi->limit, id) = __rlim_hard(&vxi->limit, id); + + return 0; +} + +int vc_set_rlimit(struct vx_info *vxi, void __user *data) +{ + struct vcmd_ctx_rlimit_v0 vc_data; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + return do_set_rlimit(vxi, vc_data.id, + vc_data.minimum, vc_data.softlimit, vc_data.maximum); +} + +#ifdef CONFIG_IA32_EMULATION + +int vc_set_rlimit_x32(struct vx_info *vxi, void __user *data) +{ + struct vcmd_ctx_rlimit_v0_x32 vc_data; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + return do_set_rlimit(vxi, vc_data.id, + vc_data.minimum, vc_data.softlimit, vc_data.maximum); +} + +int vc_get_rlimit_x32(struct vx_info *vxi, void __user *data) +{ + struct vcmd_ctx_rlimit_v0_x32 vc_data; + int ret; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + ret = do_get_rlimit(vxi, vc_data.id, + &vc_data.minimum, &vc_data.softlimit, &vc_data.maximum); + if (ret) + return ret; + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + +#endif /* CONFIG_IA32_EMULATION */ + + +int vc_get_rlimit_mask(uint32_t id, void __user *data) +{ + if (copy_to_user(data, &vlimit_mask, sizeof(vlimit_mask))) + return -EFAULT; + return 0; +} + + +static inline void vx_reset_minmax(struct _vx_limit *limit) +{ + rlim_t value; + int lim; + + for (lim = 0; lim < NUM_LIMITS; lim++) { + value = __rlim_get(limit, lim); + __rlim_rmax(limit, lim) = value; + __rlim_rmin(limit, lim) = value; + } +} + + +int vc_reset_minmax(struct vx_info *vxi, void __user *data) +{ + vx_reset_minmax(&vxi->limit); + return 0; +} + + +int vc_rlimit_stat(struct vx_info *vxi, void __user *data) +{ + struct vcmd_rlimit_stat_v0 vc_data; + struct _vx_limit *limit = &vxi->limit; + int id; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + id = vc_data.id; + if (!is_accounted_vlimit(id)) + return -EINVAL; + + vx_limit_fixup(limit, id); + vc_data.hits = atomic_read(&__rlim_lhit(limit, id)); + vc_data.value = __rlim_get(limit, id); + vc_data.minimum = __rlim_rmin(limit, id); + vc_data.maximum = __rlim_rmax(limit, id); + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + + +void vx_vsi_meminfo(struct sysinfo *val) +{ + struct vx_info *vxi = current->vx_info; + unsigned long totalram, freeram; + rlim_t v; + + /* we blindly accept the max */ + v = __rlim_soft(&vxi->limit, RLIMIT_RSS); + totalram = (v != RLIM_INFINITY) ? v : val->totalram; + + /* total minus used equals free */ + v = __vx_cres_array_fixup(&vxi->limit, VLA_RSS); + freeram = (v < totalram) ? totalram - v : 0; + + val->totalram = totalram; + val->freeram = freeram; + val->bufferram = 0; + val->totalhigh = 0; + val->freehigh = 0; + return; +} + +void vx_vsi_swapinfo(struct sysinfo *val) +{ + struct vx_info *vxi = current->vx_info; + unsigned long totalswap, freeswap; + rlim_t v, w; + + v = __rlim_soft(&vxi->limit, RLIMIT_RSS); + if (v == RLIM_INFINITY) { + val->freeswap = val->totalswap; + return; + } + + /* we blindly accept the max */ + w = __rlim_hard(&vxi->limit, RLIMIT_RSS); + totalswap = (w != RLIM_INFINITY) ? (w - v) : val->totalswap; + + /* currently 'used' swap */ + w = __vx_cres_array_fixup(&vxi->limit, VLA_RSS); + w -= (w > v) ? v : w; + + /* total minus used equals free */ + freeswap = (w < totalswap) ? totalswap - w : 0; + + val->totalswap = totalswap; + val->freeswap = freeswap; + return; +} + + +unsigned long vx_badness(struct task_struct *task, struct mm_struct *mm) +{ + struct vx_info *vxi = mm->mm_vx_info; + unsigned long points; + rlim_t v, w; + + if (!vxi) + return 0; + + v = __vx_cres_array_fixup(&vxi->limit, VLA_RSS); + w = __rlim_soft(&vxi->limit, RLIMIT_RSS); + points = (v > w) ? (v - w) : 0; + + return points; +} + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/limit_init.h linux-source-2.6.22/kernel/vserver/limit_init.h --- linux-source-2.6.22.orig/kernel/vserver/limit_init.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/limit_init.h 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,33 @@ + + +static inline void vx_info_init_limit(struct _vx_limit *limit) +{ + int lim; + + for (lim = 0; lim < NUM_LIMITS; lim++) { + __rlim_soft(limit, lim) = RLIM_INFINITY; + __rlim_hard(limit, lim) = RLIM_INFINITY; + __rlim_set(limit, lim, 0); + atomic_set(&__rlim_lhit(limit, lim), 0); + __rlim_rmin(limit, lim) = 0; + __rlim_rmax(limit, lim) = 0; + } +} + +static inline void vx_info_exit_limit(struct _vx_limit *limit) +{ +#ifdef CONFIG_VSERVER_WARN + rlim_t value; + int lim; + + for (lim = 0; lim < NUM_LIMITS; lim++) { + if ((1 << lim) & VLIM_NOCHECK) + continue; + value = __rlim_get(limit, lim); + vxwprintk(value, + "!!! limit: %p[%s,%d] = %ld on exit.", + limit, vlimit_name[lim], lim, (long)value); + } +#endif +} + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/limit_proc.h linux-source-2.6.22/kernel/vserver/limit_proc.h --- linux-source-2.6.22.orig/kernel/vserver/limit_proc.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/limit_proc.h 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,57 @@ +#ifndef _VX_LIMIT_PROC_H +#define _VX_LIMIT_PROC_H + +#include + + +#define VX_LIMIT_FMT ":\t%8ld\t%8ld/%8ld\t%8lld/%8lld\t%6d\n" +#define VX_LIMIT_TOP \ + "Limit\t current\t min/max\t\t soft/hard\t\thits\n" + +#define VX_LIMIT_ARG(r) \ + (unsigned long)__rlim_get(limit, r), \ + (unsigned long)__rlim_rmin(limit, r), \ + (unsigned long)__rlim_rmax(limit, r), \ + VX_VLIM(__rlim_soft(limit, r)), \ + VX_VLIM(__rlim_hard(limit, r)), \ + atomic_read(&__rlim_lhit(limit, r)) + +static inline int vx_info_proc_limit(struct _vx_limit *limit, char *buffer) +{ + vx_limit_fixup(limit, -1); + return sprintf(buffer, VX_LIMIT_TOP + "PROC" VX_LIMIT_FMT + "VM" VX_LIMIT_FMT + "VML" VX_LIMIT_FMT + "RSS" VX_LIMIT_FMT + "ANON" VX_LIMIT_FMT + "RMAP" VX_LIMIT_FMT + "FILES" VX_LIMIT_FMT + "OFD" VX_LIMIT_FMT + "LOCKS" VX_LIMIT_FMT + "SOCK" VX_LIMIT_FMT + "MSGQ" VX_LIMIT_FMT + "SHM" VX_LIMIT_FMT + "SEMA" VX_LIMIT_FMT + "SEMS" VX_LIMIT_FMT + "DENT" VX_LIMIT_FMT, + VX_LIMIT_ARG(RLIMIT_NPROC), + VX_LIMIT_ARG(RLIMIT_AS), + VX_LIMIT_ARG(RLIMIT_MEMLOCK), + VX_LIMIT_ARG(RLIMIT_RSS), + VX_LIMIT_ARG(VLIMIT_ANON), + VX_LIMIT_ARG(VLIMIT_MAPPED), + VX_LIMIT_ARG(RLIMIT_NOFILE), + VX_LIMIT_ARG(VLIMIT_OPENFD), + VX_LIMIT_ARG(RLIMIT_LOCKS), + VX_LIMIT_ARG(VLIMIT_NSOCK), + VX_LIMIT_ARG(RLIMIT_MSGQUEUE), + VX_LIMIT_ARG(VLIMIT_SHMEM), + VX_LIMIT_ARG(VLIMIT_SEMARY), + VX_LIMIT_ARG(VLIMIT_NSEMS), + VX_LIMIT_ARG(VLIMIT_DENTRY)); +} + +#endif /* _VX_LIMIT_PROC_H */ + + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/Makefile linux-source-2.6.22/kernel/vserver/Makefile --- linux-source-2.6.22.orig/kernel/vserver/Makefile 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/Makefile 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,17 @@ +# +# Makefile for the Linux vserver routines. +# + + +obj-y += vserver.o + +vserver-y := switch.o context.o space.o sched.o network.o inode.o \ + limit.o cvirt.o cacct.o signal.o helper.o init.o dlimit.o + +vserver-$(CONFIG_PROC_FS) += proc.o +vserver-$(CONFIG_VSERVER_DEBUG) += sysctl.o debug.o +vserver-$(CONFIG_VSERVER_LEGACY) += legacy.o +vserver-$(CONFIG_VSERVER_LEGACYNET) += legacynet.o +vserver-$(CONFIG_VSERVER_HISTORY) += history.o +vserver-$(CONFIG_VSERVER_MONITOR) += monitor.o + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/monitor.c linux-source-2.6.22/kernel/vserver/monitor.c --- linux-source-2.6.22.orig/kernel/vserver/monitor.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/monitor.c 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,143 @@ +/* + * kernel/vserver/monitor.c + * + * Virtual Context Scheduler Monitor + * + * Copyright (C) 2006-2007 Herbert Pötzl + * + * V0.01 basic design + * + */ + +#include +#include +#include +#include +#include + +#include +#include +#include + +#include +#include + + +#ifdef CONFIG_VSERVER_MONITOR +#define VXM_SIZE CONFIG_VSERVER_MONITOR_SIZE +#else +#define VXM_SIZE 64 +#endif + +struct _vx_monitor { + unsigned int counter; + + struct _vx_mon_entry entry[VXM_SIZE+1]; +}; + + +DEFINE_PER_CPU(struct _vx_monitor, vx_monitor_buffer); + +unsigned volatile int vxm_active = 1; + +static atomic_t sequence = ATOMIC_INIT(0); + + +/* vxm_advance() + + * requires disabled preemption */ + +struct _vx_mon_entry *vxm_advance(int cpu) +{ + struct _vx_monitor *mon = &per_cpu(vx_monitor_buffer, cpu); + struct _vx_mon_entry *entry; + unsigned int index; + + index = vxm_active ? (mon->counter++ % VXM_SIZE) : VXM_SIZE; + entry = &mon->entry[index]; + + entry->ev.seq = atomic_inc_return(&sequence); + entry->ev.jif = jiffies; + return entry; +} + +EXPORT_SYMBOL_GPL(vxm_advance); + + +int do_read_monitor(struct __user _vx_mon_entry *data, + int cpu, uint32_t *index, uint32_t *count) +{ + int pos, ret = 0; + struct _vx_monitor *mon = &per_cpu(vx_monitor_buffer, cpu); + int end = mon->counter; + int start = end - VXM_SIZE + 2; + int idx = *index; + + /* special case: get current pos */ + if (!*count) { + *index = end; + return 0; + } + + /* have we lost some data? */ + if (idx < start) + idx = start; + + for (pos = 0; (pos < *count) && (idx < end); pos++, idx++) { + struct _vx_mon_entry *entry = + &mon->entry[idx % VXM_SIZE]; + + /* send entry to userspace */ + ret = copy_to_user(&data[pos], entry, sizeof(*entry)); + if (ret) + break; + } + /* save new index and count */ + *index = idx; + *count = pos; + return ret ? ret : (*index < end); +} + +int vc_read_monitor(uint32_t id, void __user *data) +{ + struct vcmd_read_monitor_v0 vc_data; + int ret; + + if (id >= NR_CPUS) + return -EINVAL; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + ret = do_read_monitor((struct __user _vx_mon_entry *)vc_data.data, + id, &vc_data.index, &vc_data.count); + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return ret; +} + +#ifdef CONFIG_COMPAT + +int vc_read_monitor_x32(uint32_t id, void __user *data) +{ + struct vcmd_read_monitor_v0_x32 vc_data; + int ret; + + if (id >= NR_CPUS) + return -EINVAL; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + ret = do_read_monitor((struct __user _vx_mon_entry *) + compat_ptr(vc_data.data_ptr), + id, &vc_data.index, &vc_data.count); + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return ret; +} + +#endif /* CONFIG_COMPAT */ + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/network.c linux-source-2.6.22/kernel/vserver/network.c --- linux-source-2.6.22.orig/kernel/vserver/network.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/network.c 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,771 @@ +/* + * linux/kernel/vserver/network.c + * + * Virtual Server: Network Support + * + * Copyright (C) 2003-2007 Herbert Pötzl + * + * V0.01 broken out from vcontext V0.05 + * V0.02 cleaned up implementation + * V0.03 added equiv nx commands + * V0.04 switch to RCU based hash + * V0.05 and back to locking again + * V0.06 changed vcmds to nxi arg + * V0.07 have __create claim() the nxi + * + */ + +#include +#include +#include + +#include +#include +#include + + +atomic_t nx_global_ctotal = ATOMIC_INIT(0); +atomic_t nx_global_cactive = ATOMIC_INIT(0); + + +/* __alloc_nx_info() + + * allocate an initialized nx_info struct + * doesn't make it visible (hash) */ + +static struct nx_info *__alloc_nx_info(nid_t nid) +{ + struct nx_info *new = NULL; + + vxdprintk(VXD_CBIT(nid, 1), "alloc_nx_info(%d)*", nid); + + /* would this benefit from a slab cache? */ + new = kmalloc(sizeof(struct nx_info), GFP_KERNEL); + if (!new) + return 0; + + memset(new, 0, sizeof(struct nx_info)); + new->nx_id = nid; + INIT_HLIST_NODE(&new->nx_hlist); + atomic_set(&new->nx_usecnt, 0); + atomic_set(&new->nx_tasks, 0); + new->nx_state = 0; + + new->nx_flags = NXF_INIT_SET; + + /* rest of init goes here */ + + vxdprintk(VXD_CBIT(nid, 0), + "alloc_nx_info(%d) = %p", nid, new); + atomic_inc(&nx_global_ctotal); + return new; +} + +/* __dealloc_nx_info() + + * final disposal of nx_info */ + +static void __dealloc_nx_info(struct nx_info *nxi) +{ + vxdprintk(VXD_CBIT(nid, 0), + "dealloc_nx_info(%p)", nxi); + + nxi->nx_hlist.next = LIST_POISON1; + nxi->nx_id = -1; + + BUG_ON(atomic_read(&nxi->nx_usecnt)); + BUG_ON(atomic_read(&nxi->nx_tasks)); + + nxi->nx_state |= NXS_RELEASED; + kfree(nxi); + atomic_dec(&nx_global_ctotal); +} + +static void __shutdown_nx_info(struct nx_info *nxi) +{ + nxi->nx_state |= NXS_SHUTDOWN; + vs_net_change(nxi, VSC_NETDOWN); +} + +/* exported stuff */ + +void free_nx_info(struct nx_info *nxi) +{ + /* context shutdown is mandatory */ + BUG_ON(nxi->nx_state != NXS_SHUTDOWN); + + /* context must not be hashed */ + BUG_ON(nxi->nx_state & NXS_HASHED); + + BUG_ON(atomic_read(&nxi->nx_usecnt)); + BUG_ON(atomic_read(&nxi->nx_tasks)); + + __dealloc_nx_info(nxi); +} + + +/* hash table for nx_info hash */ + +#define NX_HASH_SIZE 13 + +struct hlist_head nx_info_hash[NX_HASH_SIZE]; + +static spinlock_t nx_info_hash_lock = SPIN_LOCK_UNLOCKED; + + +static inline unsigned int __hashval(nid_t nid) +{ + return (nid % NX_HASH_SIZE); +} + + + +/* __hash_nx_info() + + * add the nxi to the global hash table + * requires the hash_lock to be held */ + +static inline void __hash_nx_info(struct nx_info *nxi) +{ + struct hlist_head *head; + + vxd_assert_lock(&nx_info_hash_lock); + vxdprintk(VXD_CBIT(nid, 4), + "__hash_nx_info: %p[#%d]", nxi, nxi->nx_id); + + /* context must not be hashed */ + BUG_ON(nx_info_state(nxi, NXS_HASHED)); + + nxi->nx_state |= NXS_HASHED; + head = &nx_info_hash[__hashval(nxi->nx_id)]; + hlist_add_head(&nxi->nx_hlist, head); + atomic_inc(&nx_global_cactive); +} + +/* __unhash_nx_info() + + * remove the nxi from the global hash table + * requires the hash_lock to be held */ + +static inline void __unhash_nx_info(struct nx_info *nxi) +{ + vxd_assert_lock(&nx_info_hash_lock); + vxdprintk(VXD_CBIT(nid, 4), + "__unhash_nx_info: %p[#%d.%d.%d]", nxi, nxi->nx_id, + atomic_read(&nxi->nx_usecnt), atomic_read(&nxi->nx_tasks)); + + /* context must be hashed */ + BUG_ON(!nx_info_state(nxi, NXS_HASHED)); + /* but without tasks */ + BUG_ON(atomic_read(&nxi->nx_tasks)); + + nxi->nx_state &= ~NXS_HASHED; + hlist_del(&nxi->nx_hlist); + atomic_dec(&nx_global_cactive); +} + + +/* __lookup_nx_info() + + * requires the hash_lock to be held + * doesn't increment the nx_refcnt */ + +static inline struct nx_info *__lookup_nx_info(nid_t nid) +{ + struct hlist_head *head = &nx_info_hash[__hashval(nid)]; + struct hlist_node *pos; + struct nx_info *nxi; + + vxd_assert_lock(&nx_info_hash_lock); + hlist_for_each(pos, head) { + nxi = hlist_entry(pos, struct nx_info, nx_hlist); + + if (nxi->nx_id == nid) + goto found; + } + nxi = NULL; +found: + vxdprintk(VXD_CBIT(nid, 0), + "__lookup_nx_info(#%u): %p[#%u]", + nid, nxi, nxi ? nxi->nx_id : 0); + return nxi; +} + + +/* __nx_dynamic_id() + + * find unused dynamic nid + * requires the hash_lock to be held */ + +static inline nid_t __nx_dynamic_id(void) +{ + static nid_t seq = MAX_N_CONTEXT; + nid_t barrier = seq; + + vxd_assert_lock(&nx_info_hash_lock); + do { + if (++seq > MAX_N_CONTEXT) + seq = MIN_D_CONTEXT; + if (!__lookup_nx_info(seq)) { + vxdprintk(VXD_CBIT(nid, 4), + "__nx_dynamic_id: [#%d]", seq); + return seq; + } + } while (barrier != seq); + return 0; +} + +/* __create_nx_info() + + * create the requested context + * get(), claim() and hash it */ + +static struct nx_info *__create_nx_info(int id) +{ + struct nx_info *new, *nxi = NULL; + + vxdprintk(VXD_CBIT(nid, 1), "create_nx_info(%d)*", id); + + if (!(new = __alloc_nx_info(id))) + return ERR_PTR(-ENOMEM); + + /* required to make dynamic xids unique */ + spin_lock(&nx_info_hash_lock); + + /* dynamic context requested */ + if (id == NX_DYNAMIC_ID) { +#ifdef CONFIG_VSERVER_DYNAMIC_IDS + id = __nx_dynamic_id(); + if (!id) { + printk(KERN_ERR "no dynamic context available.\n"); + nxi = ERR_PTR(-EAGAIN); + goto out_unlock; + } + new->nx_id = id; +#else + printk(KERN_ERR "dynamic contexts disabled.\n"); + nxi = ERR_PTR(-EINVAL); + goto out_unlock; +#endif + } + /* static context requested */ + else if ((nxi = __lookup_nx_info(id))) { + vxdprintk(VXD_CBIT(nid, 0), + "create_nx_info(%d) = %p (already there)", id, nxi); + if (nx_info_flags(nxi, NXF_STATE_SETUP, 0)) + nxi = ERR_PTR(-EBUSY); + else + nxi = ERR_PTR(-EEXIST); + goto out_unlock; + } + /* dynamic nid creation blocker */ + else if (id >= MIN_D_CONTEXT) { + vxdprintk(VXD_CBIT(nid, 0), + "create_nx_info(%d) (dynamic rejected)", id); + nxi = ERR_PTR(-EINVAL); + goto out_unlock; + } + + /* new context */ + vxdprintk(VXD_CBIT(nid, 0), + "create_nx_info(%d) = %p (new)", id, new); + claim_nx_info(new, NULL); + __hash_nx_info(get_nx_info(new)); + nxi = new, new = NULL; + +out_unlock: + spin_unlock(&nx_info_hash_lock); + if (new) + __dealloc_nx_info(new); + return nxi; +} + + + +/* exported stuff */ + + +void unhash_nx_info(struct nx_info *nxi) +{ + __shutdown_nx_info(nxi); + spin_lock(&nx_info_hash_lock); + __unhash_nx_info(nxi); + spin_unlock(&nx_info_hash_lock); +} + +#ifdef CONFIG_VSERVER_LEGACYNET + +struct nx_info *create_nx_info(void) +{ + return __create_nx_info(NX_DYNAMIC_ID); +} + +#endif + +/* lookup_nx_info() + + * search for a nx_info and get() it + * negative id means current */ + +struct nx_info *lookup_nx_info(int id) +{ + struct nx_info *nxi = NULL; + + if (id < 0) { + nxi = get_nx_info(current->nx_info); + } else if (id > 1) { + spin_lock(&nx_info_hash_lock); + nxi = get_nx_info(__lookup_nx_info(id)); + spin_unlock(&nx_info_hash_lock); + } + return nxi; +} + +/* nid_is_hashed() + + * verify that nid is still hashed */ + +int nid_is_hashed(nid_t nid) +{ + int hashed; + + spin_lock(&nx_info_hash_lock); + hashed = (__lookup_nx_info(nid) != NULL); + spin_unlock(&nx_info_hash_lock); + return hashed; +} + + +#ifdef CONFIG_PROC_FS + +/* get_nid_list() + + * get a subset of hashed nids for proc + * assumes size is at least one */ + +int get_nid_list(int index, unsigned int *nids, int size) +{ + int hindex, nr_nids = 0; + + /* only show current and children */ + if (!nx_check(0, VS_ADMIN | VS_WATCH)) { + if (index > 0) + return 0; + nids[nr_nids] = nx_current_nid(); + return 1; + } + + for (hindex = 0; hindex < NX_HASH_SIZE; hindex++) { + struct hlist_head *head = &nx_info_hash[hindex]; + struct hlist_node *pos; + + spin_lock(&nx_info_hash_lock); + hlist_for_each(pos, head) { + struct nx_info *nxi; + + if (--index > 0) + continue; + + nxi = hlist_entry(pos, struct nx_info, nx_hlist); + nids[nr_nids] = nxi->nx_id; + if (++nr_nids >= size) { + spin_unlock(&nx_info_hash_lock); + goto out; + } + } + /* keep the lock time short */ + spin_unlock(&nx_info_hash_lock); + } +out: + return nr_nids; +} +#endif + + +/* + * migrate task to new network + * gets nxi, puts old_nxi on change + */ + +int nx_migrate_task(struct task_struct *p, struct nx_info *nxi) +{ + struct nx_info *old_nxi; + int ret = 0; + + if (!p || !nxi) + BUG(); + + vxdprintk(VXD_CBIT(nid, 5), + "nx_migrate_task(%p,%p[#%d.%d.%d])", + p, nxi, nxi->nx_id, + atomic_read(&nxi->nx_usecnt), + atomic_read(&nxi->nx_tasks)); + + if (nx_info_flags(nxi, NXF_INFO_PRIVATE, 0) && + !nx_info_flags(nxi, NXF_STATE_SETUP, 0)) + return -EACCES; + + if (nx_info_state(nxi, NXS_SHUTDOWN)) + return -EFAULT; + + /* maybe disallow this completely? */ + old_nxi = task_get_nx_info(p); + if (old_nxi == nxi) + goto out; + + task_lock(p); + if (old_nxi) + clr_nx_info(&p->nx_info); + claim_nx_info(nxi, p); + set_nx_info(&p->nx_info, nxi); + p->nid = nxi->nx_id; + task_unlock(p); + + vxdprintk(VXD_CBIT(nid, 5), + "moved task %p into nxi:%p[#%d]", + p, nxi, nxi->nx_id); + + if (old_nxi) + release_nx_info(old_nxi, p); + ret = 0; +out: + put_nx_info(old_nxi); + return ret; +} + + +#ifdef CONFIG_INET + +#include +#include + +int ifa_in_nx_info(struct in_ifaddr *ifa, struct nx_info *nxi) +{ + if (!nxi) + return 1; + if (!ifa) + return 0; + return addr_in_nx_info(nxi, ifa->ifa_local); +} + +int dev_in_nx_info(struct net_device *dev, struct nx_info *nxi) +{ + struct in_device *in_dev; + struct in_ifaddr **ifap; + struct in_ifaddr *ifa; + int ret = 0; + + if (!nxi) + return 1; + + if (!dev) + goto out; + in_dev = in_dev_get(dev); + if (!in_dev) + goto out; + + for (ifap = &in_dev->ifa_list; (ifa = *ifap) != NULL; + ifap = &ifa->ifa_next) { + if (addr_in_nx_info(nxi, ifa->ifa_local)) { + ret = 1; + break; + } + } + in_dev_put(in_dev); +out: + return ret; +} + +/* + * check if address is covered by socket + * + * sk: the socket to check against + * addr: the address in question (must be != 0) + */ +static inline int __addr_in_socket(const struct sock *sk, uint32_t addr) +{ + struct nx_info *nxi = sk->sk_nx_info; + uint32_t saddr = inet_rcv_saddr(sk); + + vxdprintk(VXD_CBIT(net, 5), + "__addr_in_socket(%p," NIPQUAD_FMT ") %p:" NIPQUAD_FMT " %p;%lx", + sk, NIPQUAD(addr), nxi, NIPQUAD(saddr), sk->sk_socket, + (sk->sk_socket ? sk->sk_socket->flags : 0)); + + if (saddr) { + /* direct address match */ + return (saddr == addr); + } else if (nxi) { + /* match against nx_info */ + return addr_in_nx_info(nxi, addr); + } else { + /* unrestricted any socket */ + return 1; + } +} + + +int nx_addr_conflict(struct nx_info *nxi, uint32_t addr, const struct sock *sk) +{ + vxdprintk(VXD_CBIT(net, 2), + "nx_addr_conflict(%p,%p) " NIPQUAD_FMT, + nxi, sk, NIPQUAD(addr)); + + if (addr) { + /* check real address */ + return __addr_in_socket(sk, addr); + } else if (nxi) { + /* check against nx_info */ + int i, n = nxi->nbipv4; + + for (i = 0; i < n; i++) + if (__addr_in_socket(sk, nxi->ipv4[i])) + return 1; + return 0; + } else { + /* check against any */ + return 1; + } +} + +#endif /* CONFIG_INET */ + +void nx_set_persistent(struct nx_info *nxi) +{ + vxdprintk(VXD_CBIT(nid, 6), + "nx_set_persistent(%p[#%d])", nxi, nxi->nx_id); + + get_nx_info(nxi); + claim_nx_info(nxi, NULL); +} + +void nx_clear_persistent(struct nx_info *nxi) +{ + vxdprintk(VXD_CBIT(nid, 6), + "nx_clear_persistent(%p[#%d])", nxi, nxi->nx_id); + + release_nx_info(nxi, NULL); + put_nx_info(nxi); +} + +void nx_update_persistent(struct nx_info *nxi) +{ + if (nx_info_flags(nxi, NXF_PERSISTENT, 0)) + nx_set_persistent(nxi); + else + nx_clear_persistent(nxi); +} + +/* vserver syscall commands below here */ + +/* taks nid and nx_info functions */ + +#include + + +int vc_task_nid(uint32_t id, void __user *data) +{ + nid_t nid; + + if (id) { + struct task_struct *tsk; + + if (!nx_check(0, VS_ADMIN | VS_WATCH)) + return -EPERM; + + read_lock(&tasklist_lock); + tsk = find_task_by_real_pid(id); + nid = (tsk) ? tsk->nid : -ESRCH; + read_unlock(&tasklist_lock); + } else + nid = nx_current_nid(); + return nid; +} + + +int vc_nx_info(struct nx_info *nxi, void __user *data) +{ + struct vcmd_nx_info_v0 vc_data; + + vc_data.nid = nxi->nx_id; + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + + +/* network functions */ + +int vc_net_create(uint32_t nid, void __user *data) +{ + struct vcmd_net_create vc_data = { .flagword = NXF_INIT_SET }; + struct nx_info *new_nxi; + int ret; + + if (data && copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + if ((nid > MAX_S_CONTEXT) && (nid != NX_DYNAMIC_ID)) + return -EINVAL; + if (nid < 2) + return -EINVAL; + + new_nxi = __create_nx_info(nid); + if (IS_ERR(new_nxi)) + return PTR_ERR(new_nxi); + + /* initial flags */ + new_nxi->nx_flags = vc_data.flagword; + + ret = -ENOEXEC; + if (vs_net_change(new_nxi, VSC_NETUP)) + goto out; + + ret = nx_migrate_task(current, new_nxi); + if (ret) + goto out; + + /* return context id on success */ + ret = new_nxi->nx_id; + + /* get a reference for persistent contexts */ + if ((vc_data.flagword & NXF_PERSISTENT)) + nx_set_persistent(new_nxi); +out: + release_nx_info(new_nxi, NULL); + put_nx_info(new_nxi); + return ret; +} + + +int vc_net_migrate(struct nx_info *nxi, void __user *data) +{ + return nx_migrate_task(current, nxi); +} + +int vc_net_add(struct nx_info *nxi, void __user *data) +{ + struct vcmd_net_addr_v0 vc_data; + int index, pos, ret = 0; + + if (data && copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + switch (vc_data.type) { + case NXA_TYPE_IPV4: + if ((vc_data.count < 1) || (vc_data.count > 4)) + return -EINVAL; + break; + + default: + break; + } + + switch (vc_data.type) { + case NXA_TYPE_IPV4: + index = 0; + while ((index < vc_data.count) && + ((pos = nxi->nbipv4) < NB_IPV4ROOT)) { + nxi->ipv4[pos] = vc_data.ip[index]; + nxi->mask[pos] = vc_data.mask[index]; + index++; + nxi->nbipv4++; + } + ret = index; + break; + + case NXA_TYPE_IPV4|NXA_MOD_BCAST: + nxi->v4_bcast = vc_data.ip[0]; + ret = 1; + break; + + default: + ret = -EINVAL; + break; + } + return ret; +} + +int vc_net_remove(struct nx_info *nxi, void __user *data) +{ + struct vcmd_net_addr_v0 vc_data; + + if (data && copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + switch (vc_data.type) { + case NXA_TYPE_ANY: + nxi->nbipv4 = 0; + break; + + default: + return -EINVAL; + } + return 0; +} + +int vc_get_nflags(struct nx_info *nxi, void __user *data) +{ + struct vcmd_net_flags_v0 vc_data; + + vc_data.flagword = nxi->nx_flags; + + /* special STATE flag handling */ + vc_data.mask = vs_mask_flags(~0ULL, nxi->nx_flags, NXF_ONE_TIME); + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + +int vc_set_nflags(struct nx_info *nxi, void __user *data) +{ + struct vcmd_net_flags_v0 vc_data; + uint64_t mask, trigger; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + /* special STATE flag handling */ + mask = vs_mask_mask(vc_data.mask, nxi->nx_flags, NXF_ONE_TIME); + trigger = (mask & nxi->nx_flags) ^ (mask & vc_data.flagword); + + nxi->nx_flags = vs_mask_flags(nxi->nx_flags, + vc_data.flagword, mask); + if (trigger & NXF_PERSISTENT) + nx_update_persistent(nxi); + + return 0; +} + +int vc_get_ncaps(struct nx_info *nxi, void __user *data) +{ + struct vcmd_net_caps_v0 vc_data; + + vc_data.ncaps = nxi->nx_ncaps; + vc_data.cmask = ~0ULL; + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + +int vc_set_ncaps(struct nx_info *nxi, void __user *data) +{ + struct vcmd_net_caps_v0 vc_data; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + nxi->nx_ncaps = vs_mask_flags(nxi->nx_ncaps, + vc_data.ncaps, vc_data.cmask); + return 0; +} + + +#include + +EXPORT_SYMBOL_GPL(free_nx_info); +EXPORT_SYMBOL_GPL(unhash_nx_info); + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/proc.c linux-source-2.6.22/kernel/vserver/proc.c --- linux-source-2.6.22.orig/kernel/vserver/proc.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/proc.c 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,1033 @@ +/* + * linux/kernel/vserver/proc.c + * + * Virtual Context Support + * + * Copyright (C) 2003-2007 Herbert Pötzl + * + * V0.01 basic structure + * V0.02 adaptation vs1.3.0 + * V0.03 proc permissions + * V0.04 locking/generic + * V0.05 next generation procfs + * V0.06 inode validation + * V0.07 generic rewrite vid + * V0.08 remove inode type + * + */ + +#include +#include +#include +#include +#include +#include + +#include +#include + +#include +#include + +#include "cvirt_proc.h" +#include "cacct_proc.h" +#include "limit_proc.h" +#include "sched_proc.h" +#include "vci_config.h" + +static struct proc_dir_entry *proc_virtual; + +static struct proc_dir_entry *proc_virtnet; + + +/* first the actual feeds */ + + +static int proc_vci(char *buffer) +{ + return sprintf(buffer, + "VCIVersion:\t%04x:%04x\n" + "VCISyscall:\t%d\n" + "VCIKernel:\t%08x\n", + VCI_VERSION >> 16, + VCI_VERSION & 0xFFFF, + __NR_vserver, + vci_kernel_config()); +} + +static int proc_virtual_info(char *buffer) +{ + return proc_vci(buffer); +} + +static int proc_virtual_status(char *buffer) +{ + return sprintf(buffer, + "#CTotal:\t%d\n" + "#CActive:\t%d\n" + "#NSProxy:\t%d\t%d %d %d %d\n", + atomic_read(&vx_global_ctotal), + atomic_read(&vx_global_cactive), + atomic_read(&vs_global_nsproxy), + atomic_read(&vs_global_fs), + atomic_read(&vs_global_mnt_ns), + atomic_read(&vs_global_uts_ns), + atomic_read(&vs_global_ipc_ns)); +} + + +int proc_vxi_info(struct vx_info *vxi, char *buffer) +{ + int length; + + length = sprintf(buffer, + "ID:\t%d\n" + "Info:\t%p\n" + "Init:\t%d\n", + vxi->vx_id, + vxi, + vxi->vx_initpid); + return length; +} + +int proc_vxi_status(struct vx_info *vxi, char *buffer) +{ + int length; + + length = sprintf(buffer, + "UseCnt:\t%d\n" + "Tasks:\t%d\n" + "Flags:\t%016llx\n" + "BCaps:\t%016llx\n" + "CCaps:\t%016llx\n" + "Spaces:\t%08lx\n", + atomic_read(&vxi->vx_usecnt), + atomic_read(&vxi->vx_tasks), + (unsigned long long)vxi->vx_flags, + (unsigned long long)vxi->vx_bcaps, + (unsigned long long)vxi->vx_ccaps, + vxi->vx_nsmask); + return length; +} + +int proc_vxi_limit(struct vx_info *vxi, char *buffer) +{ + return vx_info_proc_limit(&vxi->limit, buffer); +} + +int proc_vxi_sched(struct vx_info *vxi, char *buffer) +{ + int cpu, length; + + length = vx_info_proc_sched(&vxi->sched, buffer); + for_each_online_cpu(cpu) { + length += vx_info_proc_sched_pc( + &vx_per_cpu(vxi, sched_pc, cpu), + buffer + length, cpu); + } + return length; +} + +int proc_vxi_nsproxy(struct vx_info *vxi, char *buffer) +{ + return vx_info_proc_nsproxy(vxi->vx_nsproxy, buffer); +} + +int proc_vxi_cvirt(struct vx_info *vxi, char *buffer) +{ + int cpu, length; + + vx_update_load(vxi); + length = vx_info_proc_cvirt(&vxi->cvirt, buffer); + for_each_online_cpu(cpu) { + length += vx_info_proc_cvirt_pc( + &vx_per_cpu(vxi, cvirt_pc, cpu), + buffer + length, cpu); + } + return length; +} + +int proc_vxi_cacct(struct vx_info *vxi, char *buffer) +{ + return vx_info_proc_cacct(&vxi->cacct, buffer); +} + + +static int proc_virtnet_info(char *buffer) +{ + return proc_vci(buffer); +} + +static int proc_virtnet_status(char *buffer) +{ + return sprintf(buffer, + "#CTotal:\t%d\n" + "#CActive:\t%d\n", + atomic_read(&nx_global_ctotal), + atomic_read(&nx_global_cactive)); +} + +int proc_nxi_info(struct nx_info *nxi, char *buffer) +{ + int length, i; + + length = sprintf(buffer, + "ID:\t%d\n" + "Info:\t%p\n", + nxi->nx_id, + nxi); + for (i = 0; i < nxi->nbipv4; i++) { + length += sprintf(buffer + length, + "%d:\t" NIPQUAD_FMT "/" NIPQUAD_FMT "\n", i, + NIPQUAD(nxi->ipv4[i]), NIPQUAD(nxi->mask[i])); + } + return length; +} + +int proc_nxi_status(struct nx_info *nxi, char *buffer) +{ + int length; + + length = sprintf(buffer, + "UseCnt:\t%d\n" + "Tasks:\t%d\n" + "Flags:\t%016llx\n" + "NCaps:\t%016llx\n", + atomic_read(&nxi->nx_usecnt), + atomic_read(&nxi->nx_tasks), + (unsigned long long)nxi->nx_flags, + (unsigned long long)nxi->nx_ncaps); + return length; +} + + + +/* here the inode helpers */ + +struct vs_entry { + int len; + char *name; + mode_t mode; + struct inode_operations *iop; + struct file_operations *fop; + union proc_op op; +}; + +static struct inode *vs_proc_make_inode(struct super_block *sb, struct vs_entry *p) +{ + struct inode *inode = new_inode(sb); + + if (!inode) + goto out; + + inode->i_mode = p->mode; + if (p->iop) + inode->i_op = p->iop; + if (p->fop) + inode->i_fop = p->fop; + + inode->i_nlink = (p->mode & S_IFDIR) ? 2 : 1; + inode->i_flags |= S_IMMUTABLE; + + inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME; + + inode->i_uid = 0; + inode->i_gid = 0; + inode->i_tag = 0; +out: + return inode; +} + +static struct dentry *vs_proc_instantiate(struct inode *dir, + struct dentry *dentry, int id, void *ptr) +{ + struct vs_entry *p = ptr; + struct inode *inode = vs_proc_make_inode(dir->i_sb, p); + struct dentry *error = ERR_PTR(-EINVAL); + + if (!inode) + goto out; + + PROC_I(inode)->op = p->op; + PROC_I(inode)->fd = id; + d_add(dentry, inode); + error = NULL; +out: + return error; +} + +/* Lookups */ + +typedef struct dentry *instantiate_t(struct inode *, struct dentry *, int, void *); + +/* + * Fill a directory entry. + * + * If possible create the dcache entry and derive our inode number and + * file type from dcache entry. + * + * Since all of the proc inode numbers are dynamically generated, the inode + * numbers do not exist until the inode is cache. This means creating the + * the dcache entry in readdir is necessary to keep the inode numbers + * reported by readdir in sync with the inode numbers reported + * by stat. + */ +static int proc_fill_cache(struct file *filp, void *dirent, filldir_t filldir, + char *name, int len, instantiate_t instantiate, int id, void *ptr) +{ + struct dentry *child, *dir = filp->f_dentry; + struct inode *inode; + struct qstr qname; + ino_t ino = 0; + unsigned type = DT_UNKNOWN; + + qname.name = name; + qname.len = len; + qname.hash = full_name_hash(name, len); + + child = d_lookup(dir, &qname); + if (!child) { + struct dentry *new; + new = d_alloc(dir, &qname); + if (new) { + child = instantiate(dir->d_inode, new, id, ptr); + if (child) + dput(new); + else + child = new; + } + } + if (!child || IS_ERR(child) || !child->d_inode) + goto end_instantiate; + inode = child->d_inode; + if (inode) { + ino = inode->i_ino; + type = inode->i_mode >> 12; + } + dput(child); +end_instantiate: + if (!ino) + ino = find_inode_number(dir, &qname); + if (!ino) + ino = 1; + return filldir(dirent, name, len, filp->f_pos, ino, type); +} + + + +/* get and revalidate vx_info/xid */ + +static inline +struct vx_info *get_proc_vx_info(struct inode *inode) +{ + return lookup_vx_info(PROC_I(inode)->fd); +} + +static int proc_xid_revalidate(struct dentry *dentry, struct nameidata *nd) +{ + struct inode *inode = dentry->d_inode; + xid_t xid = PROC_I(inode)->fd; + + if (!xid || xid_is_hashed(xid)) + return 1; + d_drop(dentry); + return 0; +} + + +/* get and revalidate nx_info/nid */ + +static int proc_nid_revalidate(struct dentry *dentry, struct nameidata *nd) +{ + struct inode *inode = dentry->d_inode; + nid_t nid = PROC_I(inode)->fd; + + if (!nid || nid_is_hashed(nid)) + return 1; + d_drop(dentry); + return 0; +} + + + +#define PROC_BLOCK_SIZE (PAGE_SIZE - 1024) + +static ssize_t proc_vs_info_read(struct file *file, char __user *buf, + size_t count, loff_t *ppos) +{ + struct inode *inode = file->f_dentry->d_inode; + unsigned long page; + ssize_t length = 0; + + if (count > PROC_BLOCK_SIZE) + count = PROC_BLOCK_SIZE; + + /* fade that out as soon as stable */ + WARN_ON(PROC_I(inode)->fd); + + if (!(page = __get_free_page(GFP_KERNEL))) + return -ENOMEM; + + BUG_ON(!PROC_I(inode)->op.proc_vs_read); + length = PROC_I(inode)->op.proc_vs_read((char *)page); + + if (length >= 0) + length = simple_read_from_buffer(buf, count, ppos, + (char *)page, length); + + free_page(page); + return length; +} + +static ssize_t proc_vx_info_read(struct file *file, char __user *buf, + size_t count, loff_t *ppos) +{ + struct inode *inode = file->f_dentry->d_inode; + struct vx_info *vxi = NULL; + xid_t xid = PROC_I(inode)->fd; + unsigned long page; + ssize_t length = 0; + + if (count > PROC_BLOCK_SIZE) + count = PROC_BLOCK_SIZE; + + /* fade that out as soon as stable */ + WARN_ON(!xid); + vxi = lookup_vx_info(xid); + if (!vxi) + goto out; + + length = -ENOMEM; + if (!(page = __get_free_page(GFP_KERNEL))) + goto out_put; + + BUG_ON(!PROC_I(inode)->op.proc_vxi_read); + length = PROC_I(inode)->op.proc_vxi_read(vxi, (char *)page); + + if (length >= 0) + length = simple_read_from_buffer(buf, count, ppos, + (char *)page, length); + + free_page(page); +out_put: + put_vx_info(vxi); +out: + return length; +} + +static ssize_t proc_nx_info_read(struct file *file, char __user *buf, + size_t count, loff_t *ppos) +{ + struct inode *inode = file->f_dentry->d_inode; + struct nx_info *nxi = NULL; + nid_t nid = PROC_I(inode)->fd; + unsigned long page; + ssize_t length = 0; + + if (count > PROC_BLOCK_SIZE) + count = PROC_BLOCK_SIZE; + + /* fade that out as soon as stable */ + WARN_ON(!nid); + nxi = lookup_nx_info(nid); + if (!nxi) + goto out; + + length = -ENOMEM; + if (!(page = __get_free_page(GFP_KERNEL))) + goto out_put; + + BUG_ON(!PROC_I(inode)->op.proc_nxi_read); + length = PROC_I(inode)->op.proc_nxi_read(nxi, (char *)page); + + if (length >= 0) + length = simple_read_from_buffer(buf, count, ppos, + (char *)page, length); + + free_page(page); +out_put: + put_nx_info(nxi); +out: + return length; +} + + + +/* here comes the lower level */ + + +#define NOD(NAME, MODE, IOP, FOP, OP) { \ + .len = sizeof(NAME) - 1, \ + .name = (NAME), \ + .mode = MODE, \ + .iop = IOP, \ + .fop = FOP, \ + .op = OP, \ +} + + +#define DIR(NAME, MODE, OTYPE) \ + NOD(NAME, (S_IFDIR | (MODE)), \ + &proc_ ## OTYPE ## _inode_operations, \ + &proc_ ## OTYPE ## _file_operations, { } ) + +#define INF(NAME, MODE, OTYPE) \ + NOD(NAME, (S_IFREG | (MODE)), NULL, \ + &proc_vs_info_file_operations, \ + { .proc_vs_read = &proc_##OTYPE } ) + +#define VINF(NAME, MODE, OTYPE) \ + NOD(NAME, (S_IFREG | (MODE)), NULL, \ + &proc_vx_info_file_operations, \ + { .proc_vxi_read = &proc_##OTYPE } ) + +#define NINF(NAME, MODE, OTYPE) \ + NOD(NAME, (S_IFREG | (MODE)), NULL, \ + &proc_nx_info_file_operations, \ + { .proc_nxi_read = &proc_##OTYPE } ) + + +static struct file_operations proc_vs_info_file_operations = { + .read = proc_vs_info_read, +}; + +static struct file_operations proc_vx_info_file_operations = { + .read = proc_vx_info_read, +}; + +static struct dentry_operations proc_xid_dentry_operations = { + .d_revalidate = proc_xid_revalidate, +}; + +static struct vs_entry vx_base_stuff[] = { + VINF("info", S_IRUGO, vxi_info), + VINF("status", S_IRUGO, vxi_status), + VINF("limit", S_IRUGO, vxi_limit), + VINF("sched", S_IRUGO, vxi_sched), + VINF("nsproxy", S_IRUGO, vxi_nsproxy), + VINF("cvirt", S_IRUGO, vxi_cvirt), + VINF("cacct", S_IRUGO, vxi_cacct), + {} +}; + + + + +static struct dentry *proc_xid_instantiate(struct inode *dir, + struct dentry *dentry, int id, void *ptr) +{ + dentry->d_op = &proc_xid_dentry_operations; + return vs_proc_instantiate(dir, dentry, id, ptr); +} + +static struct dentry *proc_xid_lookup(struct inode *dir, + struct dentry *dentry, struct nameidata *nd) +{ + struct vs_entry *p = vx_base_stuff; + struct dentry *error = ERR_PTR(-ENOENT); + + for (; p->name; p++) { + if (p->len != dentry->d_name.len) + continue; + if (!memcmp(dentry->d_name.name, p->name, p->len)) + break; + } + if (!p->name) + goto out; + + error = proc_xid_instantiate(dir, dentry, PROC_I(dir)->fd, p); +out: + return error; +} + +static int proc_xid_readdir(struct file *filp, + void *dirent, filldir_t filldir) +{ + struct dentry *dentry = filp->f_dentry; + struct inode *inode = dentry->d_inode; + struct vs_entry *p = vx_base_stuff; + int size = sizeof(vx_base_stuff) / sizeof(struct vs_entry); + int pos, index; + u64 ino; + + pos = filp->f_pos; + switch (pos) { + case 0: + ino = inode->i_ino; + if (filldir(dirent, ".", 1, pos, ino, DT_DIR) < 0) + goto out; + pos++; + /* fall through */ + case 1: + ino = parent_ino(dentry); + if (filldir(dirent, "..", 2, pos, ino, DT_DIR) < 0) + goto out; + pos++; + /* fall through */ + default: + index = pos - 2; + if (index >= size) + goto out; + for (p += index; p->name; p++) { + if (proc_fill_cache(filp, dirent, filldir, p->name, p->len, + vs_proc_instantiate, PROC_I(inode)->fd, p)) + goto out; + pos++; + } + } +out: + filp->f_pos = pos; + return 1; +} + + + +static struct file_operations proc_nx_info_file_operations = { + .read = proc_nx_info_read, +}; + +static struct dentry_operations proc_nid_dentry_operations = { + .d_revalidate = proc_nid_revalidate, +}; + +static struct vs_entry nx_base_stuff[] = { + NINF("info", S_IRUGO, nxi_info), + NINF("status", S_IRUGO, nxi_status), + {} +}; + + +static struct dentry *proc_nid_instantiate(struct inode *dir, + struct dentry *dentry, int id, void *ptr) +{ + dentry->d_op = &proc_nid_dentry_operations; + return vs_proc_instantiate(dir, dentry, id, ptr); +} + +static struct dentry *proc_nid_lookup(struct inode *dir, + struct dentry *dentry, struct nameidata *nd) +{ + struct vs_entry *p = nx_base_stuff; + struct dentry *error = ERR_PTR(-ENOENT); + + for (; p->name; p++) { + if (p->len != dentry->d_name.len) + continue; + if (!memcmp(dentry->d_name.name, p->name, p->len)) + break; + } + if (!p->name) + goto out; + + error = proc_nid_instantiate(dir, dentry, PROC_I(dir)->fd, p); +out: + return error; +} + +static int proc_nid_readdir(struct file *filp, + void *dirent, filldir_t filldir) +{ + struct dentry *dentry = filp->f_dentry; + struct inode *inode = dentry->d_inode; + struct vs_entry *p = nx_base_stuff; + int size = sizeof(nx_base_stuff) / sizeof(struct vs_entry); + int pos, index; + u64 ino; + + pos = filp->f_pos; + switch (pos) { + case 0: + ino = inode->i_ino; + if (filldir(dirent, ".", 1, pos, ino, DT_DIR) < 0) + goto out; + pos++; + /* fall through */ + case 1: + ino = parent_ino(dentry); + if (filldir(dirent, "..", 2, pos, ino, DT_DIR) < 0) + goto out; + pos++; + /* fall through */ + default: + index = pos - 2; + if (index >= size) + goto out; + for (p += index; p->name; p++) { + if (proc_fill_cache(filp, dirent, filldir, p->name, p->len, + vs_proc_instantiate, PROC_I(inode)->fd, p)) + goto out; + pos++; + } + } +out: + filp->f_pos = pos; + return 1; +} + + +#define MAX_MULBY10 ((~0U - 9) / 10) + +static inline int atovid(const char *str, int len) +{ + int vid, c; + + vid = 0; + while (len-- > 0) { + c = *str - '0'; + str++; + if (c > 9) + return -1; + if (vid >= MAX_MULBY10) + return -1; + vid *= 10; + vid += c; + if (!vid) + return -1; + } + return vid; +} + +/* now the upper level (virtual) */ + + +static struct file_operations proc_xid_file_operations = { + .read = generic_read_dir, + .readdir = proc_xid_readdir, +}; + +static struct inode_operations proc_xid_inode_operations = { + .lookup = proc_xid_lookup, +}; + +static struct vs_entry vx_virtual_stuff[] = { + INF("info", S_IRUGO, virtual_info), + INF("status", S_IRUGO, virtual_status), + DIR(NULL, S_IRUGO | S_IXUGO, xid), +}; + + +static struct dentry *proc_virtual_lookup(struct inode *dir, + struct dentry *dentry, struct nameidata *nd) +{ + struct vs_entry *p = vx_virtual_stuff; + struct dentry *error = ERR_PTR(-ENOENT); + int id = 0; + + for (; p->name; p++) { + if (p->len != dentry->d_name.len) + continue; + if (!memcmp(dentry->d_name.name, p->name, p->len)) + break; + } + if (p->name) + goto instantiate; + + id = atovid(dentry->d_name.name, dentry->d_name.len); + if ((id < 0) || !xid_is_hashed(id)) + goto out; + +instantiate: + error = proc_xid_instantiate(dir, dentry, id, p); +out: + return error; +} + +static struct file_operations proc_nid_file_operations = { + .read = generic_read_dir, + .readdir = proc_nid_readdir, +}; + +static struct inode_operations proc_nid_inode_operations = { + .lookup = proc_nid_lookup, +}; + +static struct vs_entry nx_virtnet_stuff[] = { + INF("info", S_IRUGO, virtnet_info), + INF("status", S_IRUGO, virtnet_status), + DIR(NULL, S_IRUGO | S_IXUGO, nid), +}; + + +static struct dentry *proc_virtnet_lookup(struct inode *dir, + struct dentry *dentry, struct nameidata *nd) +{ + struct vs_entry *p = nx_virtnet_stuff; + struct dentry *error = ERR_PTR(-ENOENT); + int id = 0; + + for (; p->name; p++) { + if (p->len != dentry->d_name.len) + continue; + if (!memcmp(dentry->d_name.name, p->name, p->len)) + break; + } + if (p->name) + goto instantiate; + + id = atovid(dentry->d_name.name, dentry->d_name.len); + if ((id < 0) || !nid_is_hashed(id)) + goto out; + +instantiate: + error = proc_nid_instantiate(dir, dentry, id, p); +out: + return error; +} + + + +#define PROC_NUMBUF 10 +#define PROC_MAXVIDS 32 + +int proc_virtual_readdir(struct file *filp, + void *dirent, filldir_t filldir) +{ + struct dentry *dentry = filp->f_dentry; + struct inode *inode = dentry->d_inode; + struct vs_entry *p = vx_virtual_stuff; + int size = sizeof(vx_virtual_stuff) / sizeof(struct vs_entry); + int pos, index; + unsigned int xid_array[PROC_MAXVIDS]; + char buf[PROC_NUMBUF]; + unsigned int nr_xids, i; + u64 ino; + + pos = filp->f_pos; + switch (pos) { + case 0: + ino = inode->i_ino; + if (filldir(dirent, ".", 1, pos, ino, DT_DIR) < 0) + goto out; + pos++; + /* fall through */ + case 1: + ino = parent_ino(dentry); + if (filldir(dirent, "..", 2, pos, ino, DT_DIR) < 0) + goto out; + pos++; + /* fall through */ + default: + index = pos - 2; + if (index >= size) + goto entries; + for (p += index; p->name; p++) { + if (proc_fill_cache(filp, dirent, filldir, p->name, p->len, + vs_proc_instantiate, 0, p)) + goto out; + pos++; + } + entries: + index = pos - size; + p = &vx_virtual_stuff[size - 1]; + nr_xids = get_xid_list(index, xid_array, PROC_MAXVIDS); + for (i = 0; i < nr_xids; i++) { + int n, xid = xid_array[i]; + unsigned int j = PROC_NUMBUF; + + n = xid; + do + buf[--j] = '0' + (n % 10); + while (n /= 10); + + if (proc_fill_cache(filp, dirent, filldir, + buf + j, PROC_NUMBUF - j, + vs_proc_instantiate, xid, p)) + goto out; + pos++; + } + } +out: + filp->f_pos = pos; + return 0; +} + +static int proc_virtual_getattr(struct vfsmount *mnt, + struct dentry *dentry, struct kstat *stat) +{ + struct inode *inode = dentry->d_inode; + + generic_fillattr(inode, stat); + stat->nlink = 2 + atomic_read(&vx_global_cactive); + return 0; +} + +static struct file_operations proc_virtual_dir_operations = { + .read = generic_read_dir, + .readdir = proc_virtual_readdir, +}; + +static struct inode_operations proc_virtual_dir_inode_operations = { + .getattr = proc_virtual_getattr, + .lookup = proc_virtual_lookup, +}; + + + + + +int proc_virtnet_readdir(struct file *filp, + void *dirent, filldir_t filldir) +{ + struct dentry *dentry = filp->f_dentry; + struct inode *inode = dentry->d_inode; + struct vs_entry *p = nx_virtnet_stuff; + int size = sizeof(nx_virtnet_stuff) / sizeof(struct vs_entry); + int pos, index; + unsigned int nid_array[PROC_MAXVIDS]; + char buf[PROC_NUMBUF]; + unsigned int nr_nids, i; + u64 ino; + + pos = filp->f_pos; + switch (pos) { + case 0: + ino = inode->i_ino; + if (filldir(dirent, ".", 1, pos, ino, DT_DIR) < 0) + goto out; + pos++; + /* fall through */ + case 1: + ino = parent_ino(dentry); + if (filldir(dirent, "..", 2, pos, ino, DT_DIR) < 0) + goto out; + pos++; + /* fall through */ + default: + index = pos - 2; + if (index >= size) + goto entries; + for (p += index; p->name; p++) { + if (proc_fill_cache(filp, dirent, filldir, p->name, p->len, + vs_proc_instantiate, 0, p)) + goto out; + pos++; + } + entries: + index = pos - size; + p = &nx_virtnet_stuff[size - 1]; + nr_nids = get_nid_list(index, nid_array, PROC_MAXVIDS); + for (i = 0; i < nr_nids; i++) { + int n, nid = nid_array[i]; + unsigned int j = PROC_NUMBUF; + + n = nid; + do + buf[--j] = '0' + (n % 10); + while (n /= 10); + + if (proc_fill_cache(filp, dirent, filldir, + buf + j, PROC_NUMBUF - j, + vs_proc_instantiate, nid, p)) + goto out; + pos++; + } + } +out: + filp->f_pos = pos; + return 0; +} + +static int proc_virtnet_getattr(struct vfsmount *mnt, + struct dentry *dentry, struct kstat *stat) +{ + struct inode *inode = dentry->d_inode; + + generic_fillattr(inode, stat); + stat->nlink = 2 + atomic_read(&nx_global_cactive); + return 0; +} + +static struct file_operations proc_virtnet_dir_operations = { + .read = generic_read_dir, + .readdir = proc_virtnet_readdir, +}; + +static struct inode_operations proc_virtnet_dir_inode_operations = { + .getattr = proc_virtnet_getattr, + .lookup = proc_virtnet_lookup, +}; + + + +void proc_vx_init(void) +{ + struct proc_dir_entry *ent; + + ent = proc_mkdir("virtual", 0); + if (ent) { + ent->proc_fops = &proc_virtual_dir_operations; + ent->proc_iops = &proc_virtual_dir_inode_operations; + } + proc_virtual = ent; + + ent = proc_mkdir("virtnet", 0); + if (ent) { + ent->proc_fops = &proc_virtnet_dir_operations; + ent->proc_iops = &proc_virtnet_dir_inode_operations; + } + proc_virtnet = ent; +} + + + + +/* per pid info */ + + +int proc_pid_vx_info(struct task_struct *p, char *buffer) +{ + struct vx_info *vxi; + char *orig = buffer; + + buffer += sprintf(buffer, "XID:\t%d\n", vx_task_xid(p)); + + vxi = task_get_vx_info(p); + if (!vxi) + goto out; + + buffer += sprintf(buffer, "BCaps:\t%016llx\n", + (unsigned long long)vxi->vx_bcaps); + buffer += sprintf(buffer, "CCaps:\t%016llx\n", + (unsigned long long)vxi->vx_ccaps); + buffer += sprintf(buffer, "CFlags:\t%016llx\n", + (unsigned long long)vxi->vx_flags); + buffer += sprintf(buffer, "CIPid:\t%d\n", vxi->vx_initpid); + + put_vx_info(vxi); +out: + return buffer - orig; +} + + +int proc_pid_nx_info(struct task_struct *p, char *buffer) +{ + struct nx_info *nxi; + char *orig = buffer; + int i; + + buffer += sprintf(buffer, "NID:\t%d\n", nx_task_nid(p)); + + nxi = task_get_nx_info(p); + if (!nxi) + goto out; + + buffer += sprintf(buffer, "NCaps:\t%016llx\n", + (unsigned long long)nxi->nx_ncaps); + buffer += sprintf(buffer, "NFlags:\t%016llx\n", + (unsigned long long)nxi->nx_flags); + + for (i = 0; i < nxi->nbipv4; i++){ + buffer += sprintf(buffer, + "V4Root[%d]:\t" NIPQUAD_FMT "/" NIPQUAD_FMT "\n", i, + NIPQUAD(nxi->ipv4[i]), + NIPQUAD(nxi->mask[i])); + } + buffer += sprintf(buffer, + "V4Root[bcast]:\t" NIPQUAD_FMT "\n", + NIPQUAD(nxi->v4_bcast)); + + put_nx_info(nxi); +out: + return buffer - orig; +} + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/sched.c linux-source-2.6.22/kernel/vserver/sched.c --- linux-source-2.6.22.orig/kernel/vserver/sched.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/sched.c 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,447 @@ +/* + * linux/kernel/vserver/sched.c + * + * Virtual Server: Scheduler Support + * + * Copyright (C) 2004-2007 Herbert Pötzl + * + * V0.01 adapted Sam Vilains version to 2.6.3 + * V0.02 removed legacy interface + * V0.03 changed vcmds to vxi arg + * + */ + +#include +#include +#include +#include + +#include +#include + +#define vxd_check_range(val, min, max) do { \ + vxlprintk((val < min) || (val > max), \ + "check_range(%ld,%ld,%ld)", \ + (long)val, (long)min, (long)max, \ + __FILE__, __LINE__); \ + } while (0) + + +void vx_update_sched_param(struct _vx_sched *sched, + struct _vx_sched_pc *sched_pc) +{ + unsigned int set_mask = sched->update_mask; + + if (set_mask & VXSM_FILL_RATE) + sched_pc->fill_rate[0] = sched->fill_rate[0]; + if (set_mask & VXSM_INTERVAL) + sched_pc->interval[0] = sched->interval[0]; + if (set_mask & VXSM_FILL_RATE2) + sched_pc->fill_rate[1] = sched->fill_rate[1]; + if (set_mask & VXSM_INTERVAL2) + sched_pc->interval[1] = sched->interval[1]; + if (set_mask & VXSM_TOKENS) + sched_pc->tokens = sched->tokens; + if (set_mask & VXSM_TOKENS_MIN) + sched_pc->tokens_min = sched->tokens_min; + if (set_mask & VXSM_TOKENS_MAX) + sched_pc->tokens_max = sched->tokens_max; + if (set_mask & VXSM_PRIO_BIAS) + sched_pc->prio_bias = sched->prio_bias; + + if (set_mask & VXSM_IDLE_TIME) + sched_pc->flags |= VXSF_IDLE_TIME; + else + sched_pc->flags &= ~VXSF_IDLE_TIME; + + /* reset time */ + sched_pc->norm_time = jiffies; +} + + +/* + * recalculate the context's scheduling tokens + * + * ret > 0 : number of tokens available + * ret < 0 : on hold, check delta_min[] + * -1 only jiffies + * -2 also idle time + * + */ +int vx_tokens_recalc(struct _vx_sched_pc *sched_pc, + unsigned long *norm_time, unsigned long *idle_time, int delta_min[2]) +{ + long delta; + long tokens = 0; + int flags = sched_pc->flags; + + /* how much time did pass? */ + delta = *norm_time - sched_pc->norm_time; + vxd_check_range(delta, 0, INT_MAX); + + if (delta >= sched_pc->interval[0]) { + long tokens, integral; + + /* calc integral token part */ + tokens = delta / sched_pc->interval[0]; + integral = tokens * sched_pc->interval[0]; + tokens *= sched_pc->fill_rate[0]; +#ifdef CONFIG_VSERVER_HARDCPU + delta_min[0] = delta - integral; + vxd_check_range(delta_min[0], 0, sched_pc->interval[0]); +#endif + /* advance time */ + sched_pc->norm_time += delta; + + /* add tokens */ + sched_pc->tokens += tokens; + sched_pc->token_time += tokens; + } else + delta_min[0] = delta; + +#ifdef CONFIG_VSERVER_IDLETIME + if (!(flags & VXSF_IDLE_TIME)) + goto skip_idle; + + /* how much was the idle skip? */ + delta = *idle_time - sched_pc->idle_time; + vxd_check_range(delta, 0, INT_MAX); + + if (delta >= sched_pc->interval[1]) { + long tokens, integral; + + /* calc fair share token part */ + tokens = delta / sched_pc->interval[1]; + integral = tokens * sched_pc->interval[1]; + tokens *= sched_pc->fill_rate[1]; + delta_min[1] = delta - integral; + vxd_check_range(delta_min[1], 0, sched_pc->interval[1]); + + /* advance idle time */ + sched_pc->idle_time += integral; + + /* add tokens */ + sched_pc->tokens += tokens; + sched_pc->token_time += tokens; + } else + delta_min[1] = delta; +skip_idle: +#endif + + /* clip at maximum */ + if (sched_pc->tokens > sched_pc->tokens_max) + sched_pc->tokens = sched_pc->tokens_max; + tokens = sched_pc->tokens; + + if ((flags & VXSF_ONHOLD)) { + /* can we unhold? */ + if (tokens >= sched_pc->tokens_min) { + flags &= ~VXSF_ONHOLD; + sched_pc->hold_ticks += + *norm_time - sched_pc->onhold; + } else + goto on_hold; + } else { + /* put on hold? */ + if (tokens <= 0) { + flags |= VXSF_ONHOLD; + sched_pc->onhold = *norm_time; + goto on_hold; + } + } + sched_pc->flags = flags; + return tokens; + +on_hold: + tokens = sched_pc->tokens_min - tokens; + sched_pc->flags = flags; + BUG_ON(tokens < 0); + +#ifdef CONFIG_VSERVER_HARDCPU + /* next interval? */ + if (!sched_pc->fill_rate[0]) + delta_min[0] = HZ; + else if (tokens > sched_pc->fill_rate[0]) + delta_min[0] += sched_pc->interval[0] * + tokens / sched_pc->fill_rate[0]; + else + delta_min[0] = sched_pc->interval[0] - delta_min[0]; + vxd_check_range(delta_min[0], 0, INT_MAX); + +#ifdef CONFIG_VSERVER_IDLETIME + if (!(flags & VXSF_IDLE_TIME)) + return -1; + + /* next interval? */ + if (!sched_pc->fill_rate[1]) + delta_min[1] = HZ; + else if (tokens > sched_pc->fill_rate[1]) + delta_min[1] += sched_pc->interval[1] * + tokens / sched_pc->fill_rate[1]; + else + delta_min[1] = sched_pc->interval[1] - delta_min[1]; + vxd_check_range(delta_min[1], 0, INT_MAX); + + return -2; +#else + return -1; +#endif /* CONFIG_VSERVER_IDLETIME */ +#else + return 0; +#endif /* CONFIG_VSERVER_HARDCPU */ +} + +static inline unsigned long msec_to_ticks(unsigned long msec) +{ + return msecs_to_jiffies(msec); +} + +static inline unsigned long ticks_to_msec(unsigned long ticks) +{ + return jiffies_to_msecs(ticks); +} + +static inline unsigned long ticks_to_usec(unsigned long ticks) +{ + return jiffies_to_usecs(ticks); +} + + +static int do_set_sched(struct vx_info *vxi, struct vcmd_sched_v5 *data) +{ + unsigned int set_mask = data->mask; + unsigned int update_mask; + int i, cpu; + + /* Sanity check data values */ + if (data->tokens_max <= 0) + data->tokens_max = HZ; + if (data->tokens_min < 0) + data->tokens_min = HZ / 3; + if (data->tokens_min >= data->tokens_max) + data->tokens_min = data->tokens_max; + + if (data->prio_bias > MAX_PRIO_BIAS) + data->prio_bias = MAX_PRIO_BIAS; + if (data->prio_bias < MIN_PRIO_BIAS) + data->prio_bias = MIN_PRIO_BIAS; + + spin_lock(&vxi->sched.tokens_lock); + + /* sync up on delayed updates */ + for_each_cpu_mask(cpu, vxi->sched.update) + vx_update_sched_param(&vxi->sched, + &vx_per_cpu(vxi, sched_pc, cpu)); + + if (set_mask & VXSM_FILL_RATE) + vxi->sched.fill_rate[0] = data->fill_rate[0]; + if (set_mask & VXSM_FILL_RATE2) + vxi->sched.fill_rate[1] = data->fill_rate[1]; + if (set_mask & VXSM_INTERVAL) + vxi->sched.interval[0] = (set_mask & VXSM_MSEC) ? + msec_to_ticks(data->interval[0]) : data->interval[0]; + if (set_mask & VXSM_INTERVAL2) + vxi->sched.interval[1] = (set_mask & VXSM_MSEC) ? + msec_to_ticks(data->interval[1]) : data->interval[1]; + if (set_mask & VXSM_TOKENS) + vxi->sched.tokens = data->tokens; + if (set_mask & VXSM_TOKENS_MIN) + vxi->sched.tokens_min = data->tokens_min; + if (set_mask & VXSM_TOKENS_MAX) + vxi->sched.tokens_max = data->tokens_max; + if (set_mask & VXSM_PRIO_BIAS) + vxi->sched.prio_bias = data->prio_bias; + + /* Sanity check rate/interval */ + for (i = 0; i < 2; i++) { + if (data->fill_rate[i] < 0) + data->fill_rate[i] = 0; + if (data->interval[i] <= 0) + data->interval[i] = HZ; + } + + update_mask = vxi->sched.update_mask & VXSM_SET_MASK; + update_mask |= (set_mask & (VXSM_SET_MASK | VXSM_IDLE_TIME)); + vxi->sched.update_mask = update_mask; +#ifdef CONFIG_SMP + rmb(); + if (set_mask & VXSM_CPU_ID) { + vxi->sched.update = cpumask_of_cpu(data->cpu_id); + cpus_and(vxi->sched.update, cpu_online_map, + vxi->sched.update); + } else + vxi->sched.update = cpu_online_map; + + /* forced reload? */ + if (set_mask & VXSM_FORCE) { + for_each_cpu_mask(cpu, vxi->sched.update) + vx_update_sched_param(&vxi->sched, + &vx_per_cpu(vxi, sched_pc, cpu)); + vxi->sched.update = CPU_MASK_NONE; + } +#else + /* on UP we update immediately */ + vx_update_sched_param(&vxi->sched, + &vx_per_cpu(vxi, sched_pc, 0)); +#endif + + spin_unlock(&vxi->sched.tokens_lock); + return 0; +} + +#define COPY_IDS(C) C(cpu_id); C(bucket_id) +#define COPY_PRI(C) C(prio_bias) +#define COPY_TOK(C) C(tokens); C(tokens_min); C(tokens_max) +#define COPY_FRI(C) C(fill_rate[0]); C(interval[0]); \ + C(fill_rate[1]); C(interval[1]); + +#define COPY_VALUE(name) vc_data.name = data->name + +static int do_set_sched_v4(struct vx_info *vxi, struct vcmd_set_sched_v4 *data) +{ + struct vcmd_sched_v5 vc_data; + + vc_data.mask = data->set_mask; + COPY_IDS(COPY_VALUE); + COPY_PRI(COPY_VALUE); + COPY_TOK(COPY_VALUE); + vc_data.fill_rate[0] = vc_data.fill_rate[1] = data->fill_rate; + vc_data.interval[0] = vc_data.interval[1] = data->interval; + return do_set_sched(vxi, &vc_data); +} + +#ifdef CONFIG_VSERVER_LEGACY + +#define COPY_MASK_V2(name, mask) \ + if (vc_data.name != SCHED_KEEP) { \ + vc_data_v4.name = vc_data.name; \ + vc_data_v4.set_mask |= mask; \ + } + +int vc_set_sched_v2(struct vx_info *vxi, void __user *data) +{ + struct vcmd_set_sched_v2 vc_data; + struct vcmd_set_sched_v4 vc_data_v4 = { .set_mask = 0 }; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + COPY_MASK_V2(fill_rate, VXSM_FILL_RATE); + COPY_MASK_V2(interval, VXSM_INTERVAL); + COPY_MASK_V2(tokens, VXSM_TOKENS); + COPY_MASK_V2(tokens_min, VXSM_TOKENS_MIN); + COPY_MASK_V2(tokens_max, VXSM_TOKENS_MAX); + vc_data_v4.bucket_id = 0; + + do_set_sched_v4(vxi, &vc_data_v4); + return 0; +} +#endif + +int vc_set_sched_v3(struct vx_info *vxi, void __user *data) +{ + struct vcmd_set_sched_v3 vc_data; + struct vcmd_set_sched_v4 vc_data_v4; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + /* structures are binary compatible */ + memcpy(&vc_data_v4, &vc_data, sizeof(vc_data)); + vc_data_v4.set_mask &= VXSM_V3_MASK; + vc_data_v4.bucket_id = 0; + + return do_set_sched_v4(vxi, &vc_data_v4); +} + +int vc_set_sched_v4(struct vx_info *vxi, void __user *data) +{ + struct vcmd_set_sched_v4 vc_data; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + return do_set_sched_v4(vxi, &vc_data); +} + + /* latest interface is v5 */ + +int vc_set_sched(struct vx_info *vxi, void __user *data) +{ + struct vcmd_sched_v5 vc_data; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + return do_set_sched(vxi, &vc_data); +} + + +int vc_get_sched(struct vx_info *vxi, void __user *data) +{ + struct vcmd_sched_v5 vc_data; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + if (vc_data.mask & VXSM_CPU_ID) { + int cpu = vc_data.cpu_id; + struct _vx_sched_pc *data; + + if (!cpu_possible(cpu)) + return -EINVAL; + + data = &vx_per_cpu(vxi, sched_pc, cpu); + COPY_TOK(COPY_VALUE); + COPY_PRI(COPY_VALUE); + COPY_FRI(COPY_VALUE); + + if (data->flags & VXSF_IDLE_TIME) + vc_data.mask |= VXSM_IDLE_TIME; + } else { + struct _vx_sched *data = &vxi->sched; + + COPY_TOK(COPY_VALUE); + COPY_PRI(COPY_VALUE); + COPY_FRI(COPY_VALUE); + } + + if (vc_data.mask & VXSM_MSEC) { + vc_data.interval[0] = ticks_to_msec(vc_data.interval[0]); + vc_data.interval[1] = ticks_to_msec(vc_data.interval[1]); + } + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + + +int vc_sched_info(struct vx_info *vxi, void __user *data) +{ + struct vcmd_sched_info vc_data; + int cpu; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + cpu = vc_data.cpu_id; + if (!cpu_possible(cpu)) + return -EINVAL; + + if (vxi) { + struct _vx_sched_pc *sched_pc = + &vx_per_cpu(vxi, sched_pc, cpu); + + vc_data.user_msec = ticks_to_msec(sched_pc->user_ticks); + vc_data.sys_msec = ticks_to_msec(sched_pc->sys_ticks); + vc_data.hold_msec = ticks_to_msec(sched_pc->hold_ticks); + vc_data.vavavoom = sched_pc->vavavoom; + } + vc_data.token_usec = ticks_to_usec(1); + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/sched_init.h linux-source-2.6.22/kernel/vserver/sched_init.h --- linux-source-2.6.22.orig/kernel/vserver/sched_init.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/sched_init.h 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,50 @@ + +static inline void vx_info_init_sched(struct _vx_sched *sched) +{ + static struct lock_class_key tokens_lock_key; + + /* scheduling; hard code starting values as constants */ + sched->fill_rate[0] = 1; + sched->interval[0] = 4; + sched->fill_rate[1] = 1; + sched->interval[1] = 8; + sched->tokens = HZ >> 2; + sched->tokens_min = HZ >> 4; + sched->tokens_max = HZ >> 1; + sched->tokens_lock = SPIN_LOCK_UNLOCKED; + sched->prio_bias = 0; + + lockdep_set_class(&sched->tokens_lock, &tokens_lock_key); +} + +static inline +void vx_info_init_sched_pc(struct _vx_sched_pc *sched_pc, int cpu) +{ + sched_pc->fill_rate[0] = 1; + sched_pc->interval[0] = 4; + sched_pc->fill_rate[1] = 1; + sched_pc->interval[1] = 8; + sched_pc->tokens = HZ >> 2; + sched_pc->tokens_min = HZ >> 4; + sched_pc->tokens_max = HZ >> 1; + sched_pc->prio_bias = 0; + sched_pc->vavavoom = 0; + sched_pc->token_time = 0; + sched_pc->idle_time = 0; + sched_pc->norm_time = jiffies; + + sched_pc->user_ticks = 0; + sched_pc->sys_ticks = 0; + sched_pc->hold_ticks = 0; +} + +static inline void vx_info_exit_sched(struct _vx_sched *sched) +{ + return; +} + +static inline +void vx_info_exit_sched_pc(struct _vx_sched_pc *sched_pc, int cpu) +{ + return; +} diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/sched_proc.h linux-source-2.6.22/kernel/vserver/sched_proc.h --- linux-source-2.6.22.orig/kernel/vserver/sched_proc.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/sched_proc.h 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,57 @@ +#ifndef _VX_SCHED_PROC_H +#define _VX_SCHED_PROC_H + + +static inline +int vx_info_proc_sched(struct _vx_sched *sched, char *buffer) +{ + int length = 0; + + length += sprintf(buffer, + "FillRate:\t%8d,%d\n" + "Interval:\t%8d,%d\n" + "TokensMin:\t%8d\n" + "TokensMax:\t%8d\n" + "PrioBias:\t%8d\n", + sched->fill_rate[0], + sched->fill_rate[1], + sched->interval[0], + sched->interval[1], + sched->tokens_min, + sched->tokens_max, + sched->prio_bias); + return length; +} + +static inline +int vx_info_proc_sched_pc(struct _vx_sched_pc *sched_pc, + char *buffer, int cpu) +{ + int length = 0; + + length += sprintf(buffer + length, + "cpu %d: %lld %lld %lld %ld %ld", cpu, + (unsigned long long)sched_pc->user_ticks, + (unsigned long long)sched_pc->sys_ticks, + (unsigned long long)sched_pc->hold_ticks, + sched_pc->token_time, + sched_pc->idle_time); + length += sprintf(buffer + length, + " %c%c %d %d %d %d/%d %d/%d", + (sched_pc->flags & VXSF_ONHOLD) ? 'H' : 'R', + (sched_pc->flags & VXSF_IDLE_TIME) ? 'I' : '-', + sched_pc->tokens, + sched_pc->tokens_min, + sched_pc->tokens_max, + sched_pc->fill_rate[0], + sched_pc->interval[0], + sched_pc->fill_rate[1], + sched_pc->interval[1]); + length += sprintf(buffer + length, + " %d %d\n", + sched_pc->prio_bias, + sched_pc->vavavoom); + return length; +} + +#endif /* _VX_SCHED_PROC_H */ diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/signal.c linux-source-2.6.22/kernel/vserver/signal.c --- linux-source-2.6.22.orig/kernel/vserver/signal.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/signal.c 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,134 @@ +/* + * linux/kernel/vserver/signal.c + * + * Virtual Server: Signal Support + * + * Copyright (C) 2003-2007 Herbert Pötzl + * + * V0.01 broken out from vcontext V0.05 + * V0.02 changed vcmds to vxi arg + * V0.03 adjusted siginfo for kill + * + */ + +#include + +#include +#include + +#include +#include + + +int vx_info_kill(struct vx_info *vxi, int pid, int sig) +{ + int retval, count = 0; + struct task_struct *p; + struct siginfo *sip = SEND_SIG_PRIV; + + retval = -ESRCH; + vxdprintk(VXD_CBIT(misc, 4), + "vx_info_kill(%p[#%d],%d,%d)*", + vxi, vxi->vx_id, pid, sig); + read_lock(&tasklist_lock); + switch (pid) { + case 0: + case -1: + for_each_process(p) { + int err = 0; + + if (vx_task_xid(p) != vxi->vx_id || p->pid <= 1 || + (pid && vxi->vx_initpid == p->pid)) + continue; + + err = group_send_sig_info(sig, sip, p); + ++count; + if (err != -EPERM) + retval = err; + } + break; + + case 1: + if (vxi->vx_initpid) { + pid = vxi->vx_initpid; + /* for now, only SIGINT to private init ... */ + if (!vx_info_flags(vxi, VXF_STATE_ADMIN, 0) && + /* ... as long as there are tasks left */ + (atomic_read(&vxi->vx_tasks) > 1)) + sig = SIGINT; + } + /* fallthrough */ + default: + p = find_task_by_real_pid(pid); + if (p) { + if (vx_task_xid(p) == vxi->vx_id) + retval = group_send_sig_info(sig, sip, p); + } + break; + } + read_unlock(&tasklist_lock); + vxdprintk(VXD_CBIT(misc, 4), + "vx_info_kill(%p[#%d],%d,%d,%ld) = %d", + vxi, vxi->vx_id, pid, sig, (long)sip, retval); + return retval; +} + +int vc_ctx_kill(struct vx_info *vxi, void __user *data) +{ + struct vcmd_ctx_kill_v0 vc_data; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + /* special check to allow guest shutdown */ + if (!vx_info_flags(vxi, VXF_STATE_ADMIN, 0) && + /* forbid killall pid=0 when init is present */ + (((vc_data.pid < 1) && vxi->vx_initpid) || + (vc_data.pid > 1))) + return -EACCES; + + return vx_info_kill(vxi, vc_data.pid, vc_data.sig); +} + + +static int __wait_exit(struct vx_info *vxi) +{ + DECLARE_WAITQUEUE(wait, current); + int ret = 0; + + add_wait_queue(&vxi->vx_wait, &wait); + set_current_state(TASK_INTERRUPTIBLE); + +wait: + if (vx_info_state(vxi, + VXS_SHUTDOWN | VXS_HASHED | VXS_HELPER) == VXS_SHUTDOWN) + goto out; + if (signal_pending(current)) { + ret = -ERESTARTSYS; + goto out; + } + schedule(); + goto wait; + +out: + set_current_state(TASK_RUNNING); + remove_wait_queue(&vxi->vx_wait, &wait); + return ret; +} + + + +int vc_wait_exit(struct vx_info *vxi, void __user *data) +{ + struct vcmd_wait_exit_v0 vc_data; + int ret; + + ret = __wait_exit(vxi); + vc_data.reboot_cmd = vxi->reboot_cmd; + vc_data.exit_code = vxi->exit_code; + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + ret = -EFAULT; + return ret; +} + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/space.c linux-source-2.6.22/kernel/vserver/space.c --- linux-source-2.6.22.orig/kernel/vserver/space.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/space.c 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,281 @@ +/* + * linux/kernel/vserver/space.c + * + * Virtual Server: Context Space Support + * + * Copyright (C) 2003-2007 Herbert Pötzl + * + * V0.01 broken out from context.c 0.07 + * V0.02 added task locking for namespace + * V0.03 broken out vx_enter_namespace + * V0.04 added *space support and commands + * + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include + + +atomic_t vs_global_nsproxy = ATOMIC_INIT(0); +atomic_t vs_global_fs = ATOMIC_INIT(0); +atomic_t vs_global_mnt_ns = ATOMIC_INIT(0); +atomic_t vs_global_uts_ns = ATOMIC_INIT(0); +atomic_t vs_global_ipc_ns = ATOMIC_INIT(0); + + +/* namespace functions */ + +#include + +const struct vcmd_space_mask space_mask = { + .mask = CLONE_NEWNS | + CLONE_NEWUTS | + CLONE_NEWIPC | + CLONE_FS +}; + + +/* + * build a new nsproxy mix + * assumes that both proxies are 'const' + * does not touch nsproxy refcounts + * will hold a reference on the result. + */ + +struct nsproxy *vs_mix_nsproxy(struct nsproxy *old_nsproxy, + struct nsproxy *new_nsproxy, unsigned long mask) +{ + struct mnt_namespace *old_ns; + struct uts_namespace *old_uts; + struct ipc_namespace *old_ipc; + struct nsproxy *nsproxy; + + nsproxy = copy_nsproxy(old_nsproxy); + if (!nsproxy) + goto out; + + if (mask & CLONE_NEWNS) { + old_ns = nsproxy->mnt_ns; + nsproxy->mnt_ns = new_nsproxy->mnt_ns; + if (nsproxy->mnt_ns) + get_mnt_ns(nsproxy->mnt_ns); + } else + old_ns = NULL; + + if (mask & CLONE_NEWUTS) { + old_uts = nsproxy->uts_ns; + nsproxy->uts_ns = new_nsproxy->uts_ns; + if (nsproxy->uts_ns) + get_uts_ns(nsproxy->uts_ns); + } else + old_uts = NULL; + + if (mask & CLONE_NEWIPC) { + old_ipc = nsproxy->ipc_ns; + nsproxy->ipc_ns = new_nsproxy->ipc_ns; + if (nsproxy->ipc_ns) + get_ipc_ns(nsproxy->ipc_ns); + } else + old_ipc = NULL; + + if (old_ns) + put_mnt_ns(old_ns); + if (old_uts) + put_uts_ns(old_uts); + if (old_ipc) + put_ipc_ns(old_ipc); +out: + return nsproxy; +} + + +/* + * merge two nsproxy structs into a new one. + * will hold a reference on the result. + */ + +static inline +struct nsproxy *__vs_merge_nsproxy(struct nsproxy *old, + struct nsproxy *proxy, unsigned long mask) +{ + struct nsproxy null_proxy = { .mnt_ns = NULL }; + + if (!proxy) + return NULL; + + if (mask) { + /* vs_mix_nsproxy returns with reference */ + return vs_mix_nsproxy(old ? old : &null_proxy, + proxy, mask); + } + get_nsproxy(proxy); + return proxy; +} + +/* + * merge two fs structs into a new one. + * will take a reference on the result. + */ + +static inline +struct fs_struct *__vs_merge_fs(struct fs_struct *old, + struct fs_struct *fs, unsigned long mask) +{ + if (!(mask & CLONE_FS)) { + if (old) + atomic_inc(&old->count); + return old; + } + + if (!fs) + return NULL; + + return copy_fs_struct(fs); +} + + +int vx_enter_space(struct vx_info *vxi, unsigned long mask) +{ + struct nsproxy *proxy, *proxy_cur, *proxy_new; + struct fs_struct *fs, *fs_cur, *fs_new; + int ret; + + if (vx_info_flags(vxi, VXF_INFO_PRIVATE, 0)) + return -EACCES; + + if (!mask) + mask = vxi->vx_nsmask; + + if ((mask & vxi->vx_nsmask) != mask) + return -EINVAL; + + proxy = vxi->vx_nsproxy; + fs = vxi->vx_fs; + + task_lock(current); + fs_cur = current->fs; + atomic_inc(&fs_cur->count); + proxy_cur = current->nsproxy; + get_nsproxy(proxy_cur); + task_unlock(current); + + fs_new = __vs_merge_fs(fs_cur, fs, mask); + if (IS_ERR(fs_new)) { + ret = PTR_ERR(fs_new); + goto out_put; + } + + proxy_new = __vs_merge_nsproxy(proxy_cur, proxy, mask); + if (IS_ERR(proxy_new)) { + ret = PTR_ERR(proxy_new); + goto out_put_fs; + } + + fs_new = xchg(¤t->fs, fs_new); + proxy_new = xchg(¤t->nsproxy, proxy_new); + ret = 0; + + if (proxy_new) + put_nsproxy(proxy_new); +out_put_fs: + if (fs_new) + put_fs_struct(fs_new); +out_put: + if (proxy_cur) + put_nsproxy(proxy_cur); + if (fs_cur) + put_fs_struct(fs_cur); + return ret; +} + + +int vx_set_space(struct vx_info *vxi, unsigned long mask) +{ + struct nsproxy *proxy_vxi, *proxy_cur, *proxy_new; + struct fs_struct *fs_vxi, *fs_cur, *fs_new; + int ret; + + if (!mask) + mask = space_mask.mask; + + if ((mask & space_mask.mask) != mask) + return -EINVAL; + + proxy_vxi = vxi->vx_nsproxy; + fs_vxi = vxi->vx_fs; + + task_lock(current); + fs_cur = current->fs; + atomic_inc(&fs_cur->count); + proxy_cur = current->nsproxy; + get_nsproxy(proxy_cur); + task_unlock(current); + + fs_new = __vs_merge_fs(fs_vxi, fs_cur, mask); + if (IS_ERR(fs_new)) { + ret = PTR_ERR(fs_new); + goto out_put; + } + + proxy_new = __vs_merge_nsproxy(proxy_vxi, proxy_cur, mask); + if (IS_ERR(proxy_new)) { + ret = PTR_ERR(proxy_new); + goto out_put_fs; + } + + fs_new = xchg(&vxi->vx_fs, fs_new); + proxy_new = xchg(&vxi->vx_nsproxy, proxy_new); + vxi->vx_nsmask |= mask; + ret = 0; + + if (proxy_new) + put_nsproxy(proxy_new); +out_put_fs: + if (fs_new) + put_fs_struct(fs_new); +out_put: + if (proxy_cur) + put_nsproxy(proxy_cur); + if (fs_cur) + put_fs_struct(fs_cur); + return ret; +} + + +int vc_enter_space(struct vx_info *vxi, void __user *data) +{ + struct vcmd_space_mask vc_data = { .mask = 0 }; + + if (data && copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + return vx_enter_space(vxi, vc_data.mask); +} + +int vc_set_space(struct vx_info *vxi, void __user *data) +{ + struct vcmd_space_mask vc_data = { .mask = 0 }; + + if (data && copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + return vx_set_space(vxi, vc_data.mask); +} + +int vc_get_space_mask(struct vx_info *vxi, void __user *data) +{ + if (copy_to_user(data, &space_mask, sizeof(space_mask))) + return -EFAULT; + return 0; +} + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/switch.c linux-source-2.6.22/kernel/vserver/switch.c --- linux-source-2.6.22.orig/kernel/vserver/switch.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/switch.c 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,538 @@ +/* + * linux/kernel/vserver/switch.c + * + * Virtual Server: Syscall Switch + * + * Copyright (C) 2003-2007 Herbert Pötzl + * + * V0.01 syscall switch + * V0.02 added signal to context + * V0.03 added rlimit functions + * V0.04 added iattr, task/xid functions + * V0.05 added debug/history stuff + * V0.06 added compat32 layer + * V0.07 vcmd args and perms + * V0.08 added status commands + * + */ + +#include +#include +#include +#include + +#include +#include +#include + +#include "vci_config.h" + +static inline +int vc_get_version(uint32_t id) +{ +#ifdef CONFIG_VSERVER_LEGACY_VERSION + if (id == 63) + return VCI_LEGACY_VERSION; +#endif + return VCI_VERSION; +} + +static inline +int vc_get_vci(uint32_t id) +{ + return vci_kernel_config(); +} + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include + + +#ifdef CONFIG_COMPAT +#define __COMPAT(name, id, data, compat) \ + (compat) ? name ## _x32(id, data) : name(id, data) +#define __COMPAT_NO_ID(name, data, compat) \ + (compat) ? name ## _x32(data) : name(data) +#else +#define __COMPAT(name, id, data, compat) \ + name(id, data) +#define __COMPAT_NO_ID(name, data, compat) \ + name(data) +#endif + + +static inline +long do_vcmd(uint32_t cmd, uint32_t id, + struct vx_info *vxi, struct nx_info *nxi, + void __user *data, int compat) +{ + switch (cmd) { + + case VCMD_get_version: + return vc_get_version(id); + case VCMD_get_vci: + return vc_get_vci(id); + + case VCMD_task_xid: + return vc_task_xid(id, data); + case VCMD_vx_info: + return vc_vx_info(vxi, data); + + case VCMD_task_nid: + return vc_task_nid(id, data); + case VCMD_nx_info: + return vc_nx_info(nxi, data); + + case VCMD_set_space_v0: + /* this is version 1 */ + case VCMD_set_space: + return vc_set_space(vxi, data); + + case VCMD_get_space_mask: + return vc_get_space_mask(vxi, data); + +#ifdef CONFIG_IA32_EMULATION + case VCMD_get_rlimit: + return __COMPAT(vc_get_rlimit, vxi, data, compat); + case VCMD_set_rlimit: + return __COMPAT(vc_set_rlimit, vxi, data, compat); +#else + case VCMD_get_rlimit: + return vc_get_rlimit(vxi, data); + case VCMD_set_rlimit: + return vc_set_rlimit(vxi, data); +#endif + case VCMD_get_rlimit_mask: + return vc_get_rlimit_mask(id, data); + case VCMD_reset_minmax: + return vc_reset_minmax(vxi, data); + + case VCMD_get_vhi_name: + return vc_get_vhi_name(vxi, data); + case VCMD_set_vhi_name: + return vc_set_vhi_name(vxi, data); + + case VCMD_ctx_stat: + return vc_ctx_stat(vxi, data); + case VCMD_virt_stat: + return vc_virt_stat(vxi, data); + case VCMD_sock_stat: + return vc_sock_stat(vxi, data); + case VCMD_rlimit_stat: + return vc_rlimit_stat(vxi, data); + + case VCMD_set_cflags: + return vc_set_cflags(vxi, data); + case VCMD_get_cflags: + return vc_get_cflags(vxi, data); + + case VCMD_set_ccaps_v0: + return vc_set_ccaps_v0(vxi, data); + /* this is version 1 */ + case VCMD_set_ccaps: + return vc_set_ccaps(vxi, data); + case VCMD_get_ccaps_v0: + return vc_get_ccaps_v0(vxi, data); + /* this is version 1 */ + case VCMD_get_ccaps: + return vc_get_ccaps(vxi, data); + case VCMD_set_bcaps: + return vc_set_bcaps(vxi, data); + case VCMD_get_bcaps: + return vc_get_bcaps(vxi, data); + + case VCMD_set_nflags: + return vc_set_nflags(nxi, data); + case VCMD_get_nflags: + return vc_get_nflags(nxi, data); + + case VCMD_set_ncaps: + return vc_set_ncaps(nxi, data); + case VCMD_get_ncaps: + return vc_get_ncaps(nxi, data); + +#ifdef CONFIG_VSERVER_LEGACY + case VCMD_set_sched_v2: + return vc_set_sched_v2(vxi, data); +#endif + case VCMD_set_sched_v3: + return vc_set_sched_v3(vxi, data); + case VCMD_set_sched_v4: + return vc_set_sched_v4(vxi, data); + /* this is version 5 */ + case VCMD_set_sched: + return vc_set_sched(vxi, data); + case VCMD_get_sched: + return vc_get_sched(vxi, data); + case VCMD_sched_info: + return vc_sched_info(vxi, data); + + case VCMD_add_dlimit: + return __COMPAT(vc_add_dlimit, id, data, compat); + case VCMD_rem_dlimit: + return __COMPAT(vc_rem_dlimit, id, data, compat); + case VCMD_set_dlimit: + return __COMPAT(vc_set_dlimit, id, data, compat); + case VCMD_get_dlimit: + return __COMPAT(vc_get_dlimit, id, data, compat); + + case VCMD_ctx_kill: + return vc_ctx_kill(vxi, data); + + case VCMD_wait_exit: + return vc_wait_exit(vxi, data); + +#ifdef CONFIG_VSERVER_LEGACY + case VCMD_create_context: + return vc_ctx_create(id, NULL); +#endif + + case VCMD_get_iattr: + return __COMPAT_NO_ID(vc_get_iattr, data, compat); + case VCMD_set_iattr: + return __COMPAT_NO_ID(vc_set_iattr, data, compat); + + case VCMD_fget_iattr: + return vc_fget_iattr(id, data); + case VCMD_fset_iattr: + return vc_fset_iattr(id, data); + + case VCMD_enter_space_v0: + return vc_enter_space(vxi, NULL); + /* this is version 1 */ + case VCMD_enter_space: + return vc_enter_space(vxi, data); + + case VCMD_ctx_create_v0: + return vc_ctx_create(id, NULL); + case VCMD_ctx_create: + return vc_ctx_create(id, data); + case VCMD_ctx_migrate_v0: + return vc_ctx_migrate(vxi, NULL); + case VCMD_ctx_migrate: + return vc_ctx_migrate(vxi, data); + + case VCMD_net_create_v0: + return vc_net_create(id, NULL); + case VCMD_net_create: + return vc_net_create(id, data); + case VCMD_net_migrate: + return vc_net_migrate(nxi, data); + case VCMD_net_add: + return vc_net_add(nxi, data); + case VCMD_net_remove: + return vc_net_remove(nxi, data); + +#ifdef CONFIG_VSERVER_HISTORY + case VCMD_dump_history: + return vc_dump_history(id); + case VCMD_read_history: + return __COMPAT(vc_read_history, id, data, compat); +#endif +#ifdef CONFIG_VSERVER_MONITOR + case VCMD_read_monitor: + return __COMPAT(vc_read_monitor, id, data, compat); +#endif +#ifdef CONFIG_VSERVER_LEGACY + case VCMD_new_s_context: + return vc_new_s_context(id, data); +#endif +#ifdef CONFIG_VSERVER_LEGACYNET + case VCMD_set_ipv4root: + return vc_set_ipv4root(id, data); +#endif + default: + vxwprintk(1, "unimplemented VCMD_%02d_%d[%d]", + VC_CATEGORY(cmd), VC_COMMAND(cmd), VC_VERSION(cmd)); + } + return -ENOSYS; +} + + +#define __VCMD(vcmd, _perm, _args, _flags) \ + case VCMD_ ## vcmd: perm = _perm; \ + args = _args; flags = _flags; break + + +#define VCA_NONE 0x00 +#define VCA_VXI 0x01 +#define VCA_NXI 0x02 + +#define VCF_NONE 0x00 +#define VCF_INFO 0x01 +#define VCF_ADMIN 0x02 +#define VCF_ARES 0x06 /* includes admin */ +#define VCF_SETUP 0x08 + +#define VCF_ZIDOK 0x10 /* zero id okay */ + + +static inline +long do_vserver(uint32_t cmd, uint32_t id, void __user *data, int compat) +{ + long ret; + int permit = -1, state = 0; + int perm = -1, args = 0, flags = 0; + struct vx_info *vxi = NULL; + struct nx_info *nxi = NULL; + + switch (cmd) { + /* unpriviledged commands */ + __VCMD(get_version, 0, VCA_NONE, 0); + __VCMD(get_vci, 0, VCA_NONE, 0); + __VCMD(get_rlimit_mask, 0, VCA_NONE, 0); + __VCMD(get_space_mask, 0, VCA_NONE, 0); + + /* info commands */ + __VCMD(task_xid, 2, VCA_NONE, 0); + __VCMD(reset_minmax, 2, VCA_VXI, 0); + __VCMD(vx_info, 3, VCA_VXI, VCF_INFO); + __VCMD(get_bcaps, 3, VCA_VXI, VCF_INFO); + __VCMD(get_ccaps_v0, 3, VCA_VXI, VCF_INFO); + __VCMD(get_ccaps, 3, VCA_VXI, VCF_INFO); + __VCMD(get_cflags, 3, VCA_VXI, VCF_INFO); + __VCMD(get_vhi_name, 3, VCA_VXI, VCF_INFO); + __VCMD(get_rlimit, 3, VCA_VXI, VCF_INFO); + + __VCMD(ctx_stat, 3, VCA_VXI, VCF_INFO); + __VCMD(virt_stat, 3, VCA_VXI, VCF_INFO); + __VCMD(sock_stat, 3, VCA_VXI, VCF_INFO); + __VCMD(rlimit_stat, 3, VCA_VXI, VCF_INFO); + + __VCMD(task_nid, 2, VCA_NONE, 0); + __VCMD(nx_info, 3, VCA_NXI, VCF_INFO); + __VCMD(get_ncaps, 3, VCA_NXI, VCF_INFO); + __VCMD(get_nflags, 3, VCA_NXI, VCF_INFO); + + __VCMD(get_iattr, 2, VCA_NONE, 0); + __VCMD(fget_iattr, 2, VCA_NONE, 0); + __VCMD(get_dlimit, 3, VCA_NONE, VCF_INFO); + __VCMD(get_sched, 3, VCA_VXI, VCF_INFO); + __VCMD(sched_info, 3, VCA_VXI, VCF_INFO | VCF_ZIDOK); + + /* lower admin commands */ + __VCMD(wait_exit, 4, VCA_VXI, VCF_INFO); + __VCMD(ctx_create_v0, 5, VCA_NONE, 0); + __VCMD(ctx_create, 5, VCA_NONE, 0); + __VCMD(ctx_migrate_v0, 5, VCA_VXI, VCF_ADMIN); + __VCMD(ctx_migrate, 5, VCA_VXI, VCF_ADMIN); + __VCMD(enter_space_v0, 5, VCA_VXI, VCF_ADMIN); + __VCMD(enter_space, 5, VCA_VXI, VCF_ADMIN); + + __VCMD(net_create_v0, 5, VCA_NONE, 0); + __VCMD(net_create, 5, VCA_NONE, 0); + __VCMD(net_migrate, 5, VCA_NXI, VCF_ADMIN); + + /* higher admin commands */ + __VCMD(ctx_kill, 6, VCA_VXI, VCF_ARES); + __VCMD(set_space_v0, 7, VCA_VXI, VCF_ARES | VCF_SETUP); + __VCMD(set_space, 7, VCA_VXI, VCF_ARES | VCF_SETUP); + + __VCMD(set_ccaps_v0, 7, VCA_VXI, VCF_ARES | VCF_SETUP); + __VCMD(set_ccaps, 7, VCA_VXI, VCF_ARES | VCF_SETUP); + __VCMD(set_bcaps, 7, VCA_VXI, VCF_ARES | VCF_SETUP); + __VCMD(set_cflags, 7, VCA_VXI, VCF_ARES | VCF_SETUP); + + __VCMD(set_vhi_name, 7, VCA_VXI, VCF_ARES | VCF_SETUP); + __VCMD(set_rlimit, 7, VCA_VXI, VCF_ARES | VCF_SETUP); + __VCMD(set_sched, 7, VCA_VXI, VCF_ARES | VCF_SETUP); + __VCMD(set_sched_v2, 7, VCA_VXI, VCF_ARES | VCF_SETUP); + __VCMD(set_sched_v3, 7, VCA_VXI, VCF_ARES | VCF_SETUP); + __VCMD(set_sched_v4, 7, VCA_VXI, VCF_ARES | VCF_SETUP); + + __VCMD(set_ncaps, 7, VCA_NXI, VCF_ARES | VCF_SETUP); + __VCMD(set_nflags, 7, VCA_NXI, VCF_ARES | VCF_SETUP); + __VCMD(net_add, 8, VCA_NXI, VCF_ARES | VCF_SETUP); + __VCMD(net_remove, 8, VCA_NXI, VCF_ARES | VCF_SETUP); + + __VCMD(set_iattr, 7, VCA_NONE, 0); + __VCMD(fset_iattr, 7, VCA_NONE, 0); + __VCMD(set_dlimit, 7, VCA_NONE, VCF_ARES); + __VCMD(add_dlimit, 8, VCA_NONE, VCF_ARES); + __VCMD(rem_dlimit, 8, VCA_NONE, VCF_ARES); + + /* debug level admin commands */ +#ifdef CONFIG_VSERVER_HISTORY + __VCMD(dump_history, 9, VCA_NONE, 0); + __VCMD(read_history, 9, VCA_NONE, 0); +#endif +#ifdef CONFIG_VSERVER_MONITOR + __VCMD(read_monitor, 9, VCA_NONE, 0); +#endif + + /* legacy commands */ +#ifdef CONFIG_VSERVER_LEGACY + __VCMD(new_s_context, 1, VCA_NONE, 0); + __VCMD(create_context, 5, VCA_NONE, 0); +#endif +#ifdef CONFIG_VSERVER_LEGACYNET + __VCMD(set_ipv4root, 5, VCA_NONE, 0); +#endif + default: + perm = -1; + } + + vxdprintk(VXD_CBIT(switch, 0), + "vc: VCMD_%02d_%d[%d], %d,%p [%d,%d,%x,%x]", + VC_CATEGORY(cmd), VC_COMMAND(cmd), + VC_VERSION(cmd), id, data, compat, + perm, args, flags); + + ret = -ENOSYS; + if (perm < 0) + goto out; + + state = 1; +#ifdef CONFIG_VSERVER_LEGACY + if (!capable(CAP_CONTEXT) && + /* dirty hack for capremove */ + !(cmd == VCMD_new_s_context && id == -2)) + goto out; +#else + if (!capable(CAP_CONTEXT)) + goto out; +#endif + + state = 2; + /* moved here from the individual commands */ + ret = -EPERM; + if ((perm > 1) && !capable(CAP_SYS_ADMIN)) + goto out; + + state = 3; + /* vcmd involves resource management */ + ret = -EPERM; + if ((flags & VCF_ARES) && !capable(CAP_SYS_RESOURCE)) + goto out; + + state = 4; + /* various legacy exceptions */ + switch (cmd) { +#ifdef CONFIG_VSERVER_LEGACY + case VCMD_set_cflags: + case VCMD_set_ccaps_v0: + ret = 0; + if (vx_check(0, VS_WATCH)) + goto out; + break; + + case VCMD_ctx_create_v0: +#endif + /* will go away when spectator is a cap */ + case VCMD_ctx_migrate_v0: + case VCMD_ctx_migrate: + if (id == 1) { + current->xid = 1; + ret = 1; + goto out; + } + break; + + /* will go away when spectator is a cap */ + case VCMD_net_migrate: + if (id == 1) { + current->nid = 1; + ret = 1; + goto out; + } + break; + + /* legacy special casing */ + case VCMD_set_space_v0: + id = -1; + break; + } + + /* vcmds are fine by default */ + permit = 1; + + /* admin type vcmds require admin ... */ + if (flags & VCF_ADMIN) + permit = vx_check(0, VS_ADMIN) ? 1 : 0; + + /* ... but setup type vcmds override that */ + if (!permit && (flags & VCF_SETUP)) + permit = vx_flags(VXF_STATE_SETUP, 0) ? 2 : 0; + + state = 5; + ret = -EPERM; + if (!permit) + goto out; + + state = 6; + if (!id && (flags & VCF_ZIDOK)) + goto skip_id; + + ret = -ESRCH; + if (args & VCA_VXI) { + vxi = lookup_vx_info(id); + if (!vxi) + goto out; + + if ((flags & VCF_ADMIN) && + /* special case kill for shutdown */ + (cmd != VCMD_ctx_kill) && + /* can context be administrated? */ + !vx_info_flags(vxi, VXF_STATE_ADMIN, 0)) { + ret = -EACCES; + goto out_vxi; + } + } + state = 7; + if (args & VCA_NXI) { + nxi = lookup_nx_info(id); + if (!nxi) + goto out_vxi; + + if ((flags & VCF_ADMIN) && + /* can context be administrated? */ + !nx_info_flags(nxi, NXF_STATE_ADMIN, 0)) { + ret = -EACCES; + goto out_nxi; + } + } +skip_id: + state = 8; + ret = do_vcmd(cmd, id, vxi, nxi, data, compat); + +out_nxi: + if ((args & VCA_NXI) && nxi) + put_nx_info(nxi); +out_vxi: + if ((args & VCA_VXI) && vxi) + put_vx_info(vxi); +out: + vxdprintk(VXD_CBIT(switch, 1), + "vc: VCMD_%02d_%d[%d] = %08lx(%ld) [%d,%d]", + VC_CATEGORY(cmd), VC_COMMAND(cmd), + VC_VERSION(cmd), ret, ret, state, permit); + return ret; +} + +asmlinkage long +sys_vserver(uint32_t cmd, uint32_t id, void __user *data) +{ + return do_vserver(cmd, id, data, 0); +} + +#ifdef CONFIG_COMPAT + +asmlinkage long +sys32_vserver(uint32_t cmd, uint32_t id, void __user *data) +{ + return do_vserver(cmd, id, data, 1); +} + +#endif /* CONFIG_COMPAT */ diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/sysctl.c linux-source-2.6.22/kernel/vserver/sysctl.c --- linux-source-2.6.22.orig/kernel/vserver/sysctl.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/sysctl.c 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,248 @@ +/* + * kernel/vserver/sysctl.c + * + * Virtual Context Support + * + * Copyright (C) 2004-2007 Herbert Pötzl + * + * V0.01 basic structure + * + */ + +#include +#include +#include +#include +#include +#include +#include + +#include +#include + + +#define CTL_VSERVER 4242 /* unused? */ + +enum { + CTL_DEBUG_ERROR = 0, + CTL_DEBUG_SWITCH = 1, + CTL_DEBUG_XID, + CTL_DEBUG_NID, + CTL_DEBUG_TAG, + CTL_DEBUG_NET, + CTL_DEBUG_LIMIT, + CTL_DEBUG_CRES, + CTL_DEBUG_DLIM, + CTL_DEBUG_QUOTA, + CTL_DEBUG_CVIRT, + CTL_DEBUG_SPACE, + CTL_DEBUG_MISC, +}; + + +unsigned int vx_debug_switch = 0; +unsigned int vx_debug_xid = 0; +unsigned int vx_debug_nid = 0; +unsigned int vx_debug_tag = 0; +unsigned int vx_debug_net = 0; +unsigned int vx_debug_limit = 0; +unsigned int vx_debug_cres = 0; +unsigned int vx_debug_dlim = 0; +unsigned int vx_debug_quota = 0; +unsigned int vx_debug_cvirt = 0; +unsigned int vx_debug_space = 0; +unsigned int vx_debug_misc = 0; + + +static struct ctl_table_header *vserver_table_header; +static ctl_table vserver_table[]; + + +void vserver_register_sysctl(void) +{ + if (!vserver_table_header) { + vserver_table_header = register_sysctl_table(vserver_table); + } + +} + +void vserver_unregister_sysctl(void) +{ + if (vserver_table_header) { + unregister_sysctl_table(vserver_table_header); + vserver_table_header = NULL; + } +} + + +static int proc_dodebug(ctl_table *table, int write, + struct file *filp, void __user *buffer, size_t *lenp, loff_t *ppos) +{ + char tmpbuf[20], *p, c; + unsigned int value; + size_t left, len; + + if ((*ppos && !write) || !*lenp) { + *lenp = 0; + return 0; + } + + left = *lenp; + + if (write) { + if (!access_ok(VERIFY_READ, buffer, left)) + return -EFAULT; + p = (char *)buffer; + while (left && __get_user(c, p) >= 0 && isspace(c)) + left--, p++; + if (!left) + goto done; + + if (left > sizeof(tmpbuf) - 1) + return -EINVAL; + if (copy_from_user(tmpbuf, p, left)) + return -EFAULT; + tmpbuf[left] = '\0'; + + for (p = tmpbuf, value = 0; '0' <= *p && *p <= '9'; p++, left--) + value = 10 * value + (*p - '0'); + if (*p && !isspace(*p)) + return -EINVAL; + while (left && isspace(*p)) + left--, p++; + *(unsigned int *)table->data = value; + } else { + if (!access_ok(VERIFY_WRITE, buffer, left)) + return -EFAULT; + len = sprintf(tmpbuf, "%d", *(unsigned int *)table->data); + if (len > left) + len = left; + if (__copy_to_user(buffer, tmpbuf, len)) + return -EFAULT; + if ((left -= len) > 0) { + if (put_user('\n', (char *)buffer + len)) + return -EFAULT; + left--; + } + } + +done: + *lenp -= left; + *ppos += *lenp; + return 0; +} + + +#define CTL_ENTRY(ctl, name) \ + { \ + .ctl_name = ctl, \ + .procname = #name, \ + .data = &vx_ ## name, \ + .maxlen = sizeof(int), \ + .mode = 0644, \ + .proc_handler = &proc_dodebug \ + } + +static ctl_table debug_table[] = { + CTL_ENTRY(CTL_DEBUG_SWITCH, debug_switch), + CTL_ENTRY(CTL_DEBUG_XID, debug_xid), + CTL_ENTRY(CTL_DEBUG_NID, debug_nid), + CTL_ENTRY(CTL_DEBUG_TAG, debug_tag), + CTL_ENTRY(CTL_DEBUG_NET, debug_net), + CTL_ENTRY(CTL_DEBUG_LIMIT, debug_limit), + CTL_ENTRY(CTL_DEBUG_CRES, debug_cres), + CTL_ENTRY(CTL_DEBUG_DLIM, debug_dlim), + CTL_ENTRY(CTL_DEBUG_QUOTA, debug_quota), + CTL_ENTRY(CTL_DEBUG_CVIRT, debug_cvirt), + CTL_ENTRY(CTL_DEBUG_SPACE, debug_space), + CTL_ENTRY(CTL_DEBUG_MISC, debug_misc), + { .ctl_name = 0 } +}; + +static ctl_table vserver_table[] = { + { + .ctl_name = CTL_VSERVER, + .procname = "vserver", + .mode = 0555, + .child = debug_table + }, + { .ctl_name = 0 } +}; + + +static match_table_t tokens = { + { CTL_DEBUG_SWITCH, "switch=%x" }, + { CTL_DEBUG_XID, "xid=%x" }, + { CTL_DEBUG_NID, "nid=%x" }, + { CTL_DEBUG_TAG, "tag=%x" }, + { CTL_DEBUG_NET, "net=%x" }, + { CTL_DEBUG_LIMIT, "limit=%x" }, + { CTL_DEBUG_CRES, "cres=%x" }, + { CTL_DEBUG_DLIM, "dlim=%x" }, + { CTL_DEBUG_QUOTA, "quota=%x" }, + { CTL_DEBUG_CVIRT, "cvirt=%x" }, + { CTL_DEBUG_SPACE, "space=%x" }, + { CTL_DEBUG_MISC, "misc=%x" }, + { CTL_DEBUG_ERROR, NULL } +}; + +#define HANDLE_CASE(id, name, val) \ + case CTL_DEBUG_ ## id: \ + vx_debug_ ## name = val; \ + printk("vs_debug_" #name "=0x%x\n", val); \ + break + + +static int __init vs_debug_setup(char *str) +{ + char *p; + int token; + + printk("vs_debug_setup(%s)\n", str); + while ((p = strsep(&str, ",")) != NULL) { + substring_t args[MAX_OPT_ARGS]; + unsigned int value; + + if (!*p) + continue; + + token = match_token(p, tokens, args); + value = (token > 0) ? simple_strtoul(args[0].from, NULL, 0) : 0; + + switch (token) { + HANDLE_CASE(SWITCH, switch, value); + HANDLE_CASE(XID, xid, value); + HANDLE_CASE(NID, nid, value); + HANDLE_CASE(TAG, tag, value); + HANDLE_CASE(NET, net, value); + HANDLE_CASE(LIMIT, limit, value); + HANDLE_CASE(CRES, cres, value); + HANDLE_CASE(DLIM, dlim, value); + HANDLE_CASE(QUOTA, quota, value); + HANDLE_CASE(CVIRT, cvirt, value); + HANDLE_CASE(SPACE, space, value); + HANDLE_CASE(MISC, misc, value); + default: + return -EINVAL; + break; + } + } + return 1; +} + +__setup("vsdebug=", vs_debug_setup); + + + +EXPORT_SYMBOL_GPL(vx_debug_switch); +EXPORT_SYMBOL_GPL(vx_debug_xid); +EXPORT_SYMBOL_GPL(vx_debug_nid); +EXPORT_SYMBOL_GPL(vx_debug_net); +EXPORT_SYMBOL_GPL(vx_debug_limit); +EXPORT_SYMBOL_GPL(vx_debug_cres); +EXPORT_SYMBOL_GPL(vx_debug_dlim); +EXPORT_SYMBOL_GPL(vx_debug_quota); +EXPORT_SYMBOL_GPL(vx_debug_cvirt); +EXPORT_SYMBOL_GPL(vx_debug_space); +EXPORT_SYMBOL_GPL(vx_debug_misc); + diff -NurpP --minimal linux-source-2.6.22.orig/kernel/vserver/vci_config.h linux-source-2.6.22/kernel/vserver/vci_config.h --- linux-source-2.6.22.orig/kernel/vserver/vci_config.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/kernel/vserver/vci_config.h 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,92 @@ + +/* interface version */ + +#define VCI_VERSION 0x00020200 +#define VCI_LEGACY_VERSION 0x000100FF + + +enum { + VCI_KCBIT_NO_DYNAMIC = 0, + VCI_KCBIT_LEGACY = 1, + VCI_KCBIT_LEGACYNET = 2, + VCI_KCBIT_NGNET = 3, + + VCI_KCBIT_PROC_SECURE = 4, + VCI_KCBIT_HARDCPU = 5, + VCI_KCBIT_IDLELIMIT = 6, + VCI_KCBIT_IDLETIME = 7, + + VCI_KCBIT_COWBL = 8, + VCI_KCBIT_FULLCOWBL = 9, + VCI_KCBIT_SPACES = 10, + + VCI_KCBIT_LEGACY_VERSION = 15, + VCI_KCBIT_DEBUG = 16, + VCI_KCBIT_HISTORY = 20, + VCI_KCBIT_TAGGED = 24, +}; + + +static inline uint32_t vci_kernel_config(void) +{ + return + /* various legacy options */ +#ifndef CONFIG_VSERVER_DYNAMIC_IDS + (1 << VCI_KCBIT_NO_DYNAMIC) | +#endif +#ifdef CONFIG_VSERVER_LEGACY + (1 << VCI_KCBIT_LEGACY) | +#endif +#ifdef CONFIG_VSERVER_LEGACYNET + (1 << VCI_KCBIT_LEGACYNET) | +#endif +#ifdef CONFIG_VSERVER_LEGACY_VERSION + (1 << VCI_KCBIT_LEGACY_VERSION) | +#endif + + /* configured features */ +#ifdef CONFIG_VSERVER_PROC_SECURE + (1 << VCI_KCBIT_PROC_SECURE) | +#endif +#ifdef CONFIG_VSERVER_HARDCPU + (1 << VCI_KCBIT_HARDCPU) | +#endif +#ifdef CONFIG_VSERVER_IDLELIMIT + (1 << VCI_KCBIT_IDLELIMIT) | +#endif +#ifdef CONFIG_VSERVER_IDLETIME + (1 << VCI_KCBIT_IDLETIME) | +#endif +#ifdef CONFIG_VSERVER_COWBL + (1 << VCI_KCBIT_COWBL) | + (1 << VCI_KCBIT_FULLCOWBL) | +#endif + (1 << VCI_KCBIT_SPACES) | + + /* debug options */ +#ifdef CONFIG_VSERVER_DEBUG + (1 << VCI_KCBIT_DEBUG) | +#endif +#ifdef CONFIG_VSERVER_HISTORY + (1 << VCI_KCBIT_HISTORY) | +#endif + + /* inode context tagging */ +#if defined(CONFIG_TAGGING_NONE) + (0 << VCI_KCBIT_TAGGED) | +#elif defined(CONFIG_TAGGING_UID16) + (1 << VCI_KCBIT_TAGGED) | +#elif defined(CONFIG_TAGGING_GID16) + (2 << VCI_KCBIT_TAGGED) | +#elif defined(CONFIG_TAGGING_ID24) + (3 << VCI_KCBIT_TAGGED) | +#elif defined(CONFIG_TAGGING_INTERN) + (4 << VCI_KCBIT_TAGGED) | +#elif defined(CONFIG_TAGGING_RUNTIME) + (5 << VCI_KCBIT_TAGGED) | +#else + (7 << VCI_KCBIT_TAGGED) | +#endif + 0; +} + diff -NurpP --minimal linux-source-2.6.22.orig/Makefile linux-source-2.6.22/Makefile --- linux-source-2.6.22.orig/Makefile 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/Makefile 2007-10-24 16:14:53.795211165 +0200 @@ -1,7 +1,7 @@ VERSION = 2 PATCHLEVEL = 6 SUBLEVEL = 22 -EXTRAVERSION = .9 +EXTRAVERSION = NAME = Holy Dancing Manatees, Batman! # *DOCUMENTATION* diff -NurpP --minimal linux-source-2.6.22.orig/mm/filemap.c linux-source-2.6.22/mm/filemap.c --- linux-source-2.6.22.orig/mm/filemap.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/mm/filemap.c 2007-10-24 16:14:06.791620780 +0200 @@ -1245,6 +1245,31 @@ int file_send_actor(read_descriptor_t * return written; } +/* FIXME: It would be as simple as this, if we had a (void __user*) to write. + * We already have a kernel buffer, so it should be even simpler, right? ;) + * + * Yes, sorta. After duplicating the complete path of generic_file_write(), + * at least some special cases could be removed, so the copy is simpler than + * the original. But it remains a copy, so overall complexity increases. + */ +static ssize_t +generic_kernel_file_write(struct file *, const char *, size_t, loff_t *); + +ssize_t generic_file_sendpage(struct file *file, struct page *page, + int offset, size_t size, loff_t *ppos, int more) +{ + ssize_t ret; + char *kaddr; + + kaddr = kmap(page); + ret = generic_kernel_file_write(file, kaddr + offset, size, ppos); + kunmap(page); + + return ret; +} + +EXPORT_SYMBOL(generic_file_sendpage); + ssize_t generic_file_sendfile(struct file *in_file, loff_t *ppos, size_t count, read_actor_t actor, void *target) { @@ -1924,6 +1949,19 @@ int remove_suid(struct path *path) } EXPORT_SYMBOL(remove_suid); +static inline size_t +filemap_copy_from_kernel(struct page *page, unsigned long offset, + const char *buf, unsigned bytes) +{ + char *kaddr; + + kaddr = kmap(page); + memcpy(kaddr + offset, buf, bytes); + kunmap(page); + + return bytes; +} + size_t __filemap_copy_from_user_iovec_inatomic(char *vaddr, const struct iovec *iov, size_t base, size_t bytes) @@ -2235,6 +2273,175 @@ zero_length_segment: } EXPORT_SYMBOL(generic_file_buffered_write); +static inline void +filemap_set_next_kvec(const struct kvec **iovp, size_t *basep, size_t bytes) +{ + const struct kvec *iov = *iovp; + size_t base = *basep; + + while (bytes) { + int copy = min(bytes, iov->iov_len - base); + + bytes -= copy; + base += copy; + if (iov->iov_len == base) { + iov++; + base = 0; + } + } + *iovp = iov; + *basep = base; +} + +/* + * TODO: + * This largely tries to copy generic_file_aio_write_nolock(), although it + * doesn't have to be nearly as generic. A real cleanup should either + * merge this into generic_file_aio_write_nolock() as well or keep it special + * and remove as much code as possible. + */ +static ssize_t +generic_kernel_file_aio_write_nolock(struct kiocb *iocb, const struct kvec*iov, + unsigned long nr_segs, loff_t *ppos) +{ + struct file *file = iocb->ki_filp; + struct address_space * mapping = file->f_mapping; + const struct address_space_operations *a_ops = mapping->a_ops; + size_t ocount; /* original count */ + size_t count; /* after file limit checks */ + struct inode *inode = mapping->host; + long status = 0; + loff_t pos; + struct page *page; + struct page *cached_page = NULL; + const int isblk = S_ISBLK(inode->i_mode); + ssize_t written; + ssize_t err; + size_t bytes; + struct pagevec lru_pvec; + const struct kvec *cur_iov = iov; /* current kvec */ + size_t iov_base = 0; /* offset in the current kvec */ + unsigned long seg; + char *buf; + + ocount = 0; + for (seg = 0; seg < nr_segs; seg++) { + const struct kvec *iv = &iov[seg]; + + /* + * If any segment has a negative length, or the cumulative + * length ever wraps negative then return -EINVAL. + */ + ocount += iv->iov_len; + if (unlikely((ssize_t)(ocount|iv->iov_len) < 0)) + return -EINVAL; + } + + count = ocount; + pos = *ppos; + pagevec_init(&lru_pvec, 0); + + /* We can write back this queue in page reclaim */ + current->backing_dev_info = mapping->backing_dev_info; + written = 0; + + err = generic_write_checks(file, &pos, &count, isblk); + if (err) + goto out; + + + if (count == 0) + goto out; + + remove_suid(file->f_dentry); + file_update_time(file); + + /* There is no sane reason to use O_DIRECT */ + BUG_ON(file->f_flags & O_DIRECT); + + buf = iov->iov_base; + do { + unsigned long index; + unsigned long offset; + size_t copied; + + offset = (pos & (PAGE_CACHE_SIZE -1)); /* Within page */ + index = pos >> PAGE_CACHE_SHIFT; + bytes = PAGE_CACHE_SIZE - offset; + if (bytes > count) + bytes = count; + + page = __grab_cache_page(mapping,index,&cached_page,&lru_pvec); + if (!page) { + status = -ENOMEM; + break; + } + + status = a_ops->prepare_write(file, page, offset, offset+bytes); + if (unlikely(status)) { + loff_t isize = i_size_read(inode); + /* + * prepare_write() may have instantiated a few blocks + * outside i_size. Trim these off again. + */ + unlock_page(page); + page_cache_release(page); + if (pos + bytes > isize) + vmtruncate(inode, isize); + break; + } + + BUG_ON(nr_segs != 1); + copied = filemap_copy_from_kernel(page, offset, buf, bytes); + + flush_dcache_page(page); + status = a_ops->commit_write(file, page, offset, offset+bytes); + if (likely(copied > 0)) { + if (!status) + status = copied; + + if (status >= 0) { + written += status; + count -= status; + pos += status; + buf += status; + if (unlikely(nr_segs > 1)) + filemap_set_next_kvec(&cur_iov, + &iov_base, status); + } + } + if (unlikely(copied != bytes)) + if (status >= 0) + status = -EFAULT; + unlock_page(page); + mark_page_accessed(page); + page_cache_release(page); + if (status < 0) + break; + balance_dirty_pages_ratelimited(mapping); + cond_resched(); + } while (count); + *ppos = pos; + + if (cached_page) + page_cache_release(cached_page); + + /* + * For now, when the user asks for O_SYNC, we'll actually give O_DSYNC + */ + if (status >= 0) { + if ((file->f_flags & O_SYNC) || IS_SYNC(inode)) + status = generic_osync_inode(inode, mapping, + OSYNC_METADATA|OSYNC_DATA); + } + + err = written ? written : status; +out: + pagevec_lru_add(&lru_pvec); + current->backing_dev_info = 0; + return err; +} + static ssize_t __generic_file_aio_write_nolock(struct kiocb *iocb, const struct iovec *iov, unsigned long nr_segs, loff_t *ppos) @@ -2335,6 +2542,36 @@ out: return written ? written : err; } +static ssize_t +generic_kernel_file_write_nolock(struct file *file, const struct kvec *iov, + unsigned long nr_segs, loff_t *ppos) +{ + struct kiocb kiocb; + ssize_t ret; + + init_sync_kiocb(&kiocb, file); + ret = generic_kernel_file_aio_write_nolock(&kiocb, iov, nr_segs, ppos); + if (ret == -EIOCBQUEUED) + ret = wait_on_sync_kiocb(&kiocb); + return ret; +} + +static ssize_t generic_kernel_file_write(struct file *file, const char *buf, + size_t count, loff_t *ppos) +{ + struct inode *inode = file->f_mapping->host; + ssize_t err; + struct kvec local_iov = { .iov_base = (char *) buf, + .iov_len = count }; + + mutex_lock(&inode->i_mutex); + err = generic_kernel_file_write_nolock(file, &local_iov, 1, ppos); + mutex_unlock(&inode->i_mutex); + + return err; +} + + ssize_t generic_file_aio_write_nolock(struct kiocb *iocb, const struct iovec *iov, unsigned long nr_segs, loff_t pos) { diff -NurpP --minimal linux-source-2.6.22.orig/mm/filemap_xip.c linux-source-2.6.22/mm/filemap_xip.c --- linux-source-2.6.22.orig/mm/filemap_xip.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/mm/filemap_xip.c 2007-10-24 16:14:06.791620780 +0200 @@ -14,6 +14,7 @@ #include #include #include +#include #include #include "filemap.h" diff -NurpP --minimal linux-source-2.6.22.orig/mm/fremap.c linux-source-2.6.22/mm/fremap.c --- linux-source-2.6.22.orig/mm/fremap.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/mm/fremap.c 2007-10-24 16:14:06.791620780 +0200 @@ -15,6 +15,7 @@ #include #include #include +#include #include #include @@ -74,6 +75,8 @@ int install_page(struct mm_struct *mm, s err = -ENOMEM; if (page_mapcount(page) > INT_MAX/2) goto unlock; + if (!vx_rss_avail(mm, 1)) + goto unlock; if (pte_none(*pte) || !zap_pte(mm, vma, addr, pte)) inc_mm_counter(mm, file_rss); diff -NurpP --minimal linux-source-2.6.22.orig/mm/hugetlb.c linux-source-2.6.22/mm/hugetlb.c --- linux-source-2.6.22.orig/mm/hugetlb.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/mm/hugetlb.c 2007-10-24 16:14:06.791620780 +0200 @@ -19,6 +19,7 @@ #include #include +#include #include "internal.h" const unsigned long hugetlb_zero = 0, hugetlb_infinity = ~0UL; diff -NurpP --minimal linux-source-2.6.22.orig/mm/memory.c linux-source-2.6.22/mm/memory.c --- linux-source-2.6.22.orig/mm/memory.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/mm/memory.c 2007-10-24 16:14:06.791620780 +0200 @@ -498,6 +498,9 @@ static int copy_pte_range(struct mm_stru int progress = 0; int rss[2]; + if (!vx_rss_avail(dst_mm, ((end - addr)/PAGE_SIZE + 1))) + return -ENOMEM; + again: rss[1] = rss[0] = 0; dst_pte = pte_alloc_map_lock(dst_mm, dst_pmd, addr, &dst_ptl); @@ -2157,6 +2160,11 @@ static int do_swap_page(struct mm_struct count_vm_event(PGMAJFAULT); } + if (!vx_rss_avail(mm, 1)) { + ret = VM_FAULT_OOM; + goto out; + } + delayacct_clear_flag(DELAYACCT_PF_SWAPIN); mark_page_accessed(page); lock_page(page); @@ -2229,6 +2237,8 @@ static int do_anonymous_page(struct mm_s /* Allocate our own private page. */ pte_unmap(page_table); + if (!vx_rss_avail(mm, 1)) + goto oom; if (unlikely(anon_vma_prepare(vma))) goto oom; page = alloc_zeroed_user_highpage(vma, address); @@ -2302,6 +2312,9 @@ static int do_no_page(struct mm_struct * pte_unmap(page_table); BUG_ON(vma->vm_flags & VM_PFNMAP); + if (!vx_rss_avail(mm, 1)) + return VM_FAULT_OOM; + if (vma->vm_file) { mapping = vma->vm_file->f_mapping; sequence = mapping->truncate_count; @@ -2527,6 +2540,7 @@ static inline int handle_pte_fault(struc { pte_t entry; spinlock_t *ptl; + int ret, type = VXPT_UNKNOWN; entry = *pte; if (!pte_present(entry)) { @@ -2555,9 +2569,12 @@ static inline int handle_pte_fault(struc if (unlikely(!pte_same(*pte, entry))) goto unlock; if (write_access) { - if (!pte_write(entry)) - return do_wp_page(mm, vma, address, + if (!pte_write(entry)) { + ret = do_wp_page(mm, vma, address, pte, pmd, ptl, entry); + type = VXPT_WRITE; + goto out; + } entry = pte_mkdirty(entry); } entry = pte_mkyoung(entry); @@ -2576,7 +2593,10 @@ static inline int handle_pte_fault(struc } unlock: pte_unmap_unlock(pte, ptl); - return VM_FAULT_MINOR; + ret = VM_FAULT_MINOR; +out: + vx_page_fault(mm, vma, type, ret); + return ret; } /* diff -NurpP --minimal linux-source-2.6.22.orig/mm/mlock.c linux-source-2.6.22/mm/mlock.c --- linux-source-2.6.22.orig/mm/mlock.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/mm/mlock.c 2007-10-24 16:14:06.791620780 +0200 @@ -12,6 +12,7 @@ #include #include #include +#include int can_do_mlock(void) { @@ -76,7 +77,7 @@ success: ret = make_pages_present(start, end); } - mm->locked_vm -= pages; + vx_vmlocked_sub(mm, pages); out: if (ret == -ENOMEM) ret = -EAGAIN; @@ -134,7 +135,7 @@ static int do_mlock(unsigned long start, asmlinkage long sys_mlock(unsigned long start, size_t len) { - unsigned long locked; + unsigned long locked, grow; unsigned long lock_limit; int error = -ENOMEM; @@ -145,8 +146,10 @@ asmlinkage long sys_mlock(unsigned long len = PAGE_ALIGN(len + (start & ~PAGE_MASK)); start &= PAGE_MASK; - locked = len >> PAGE_SHIFT; - locked += current->mm->locked_vm; + grow = len >> PAGE_SHIFT; + if (!vx_vmlocked_avail(current->mm, grow)) + goto out; + locked = current->mm->locked_vm + grow; lock_limit = current->signal->rlim[RLIMIT_MEMLOCK].rlim_cur; lock_limit >>= PAGE_SHIFT; @@ -154,6 +157,7 @@ asmlinkage long sys_mlock(unsigned long /* check against resource limits */ if ((locked <= lock_limit) || capable(CAP_IPC_LOCK)) error = do_mlock(start, len, 1); +out: up_write(¤t->mm->mmap_sem); return error; } @@ -213,6 +217,8 @@ asmlinkage long sys_mlockall(int flags) lock_limit >>= PAGE_SHIFT; ret = -ENOMEM; + if (!vx_vmlocked_avail(current->mm, current->mm->total_vm)) + goto out; if (!(flags & MCL_CURRENT) || (current->mm->total_vm <= lock_limit) || capable(CAP_IPC_LOCK)) ret = do_mlockall(flags); diff -NurpP --minimal linux-source-2.6.22.orig/mm/mmap.c linux-source-2.6.22/mm/mmap.c --- linux-source-2.6.22.orig/mm/mmap.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/mm/mmap.c 2007-10-24 16:14:06.791620780 +0200 @@ -1144,10 +1144,10 @@ munmap_back: kmem_cache_free(vm_area_cachep, vma); } out: - mm->total_vm += len >> PAGE_SHIFT; + vx_vmpages_add(mm, len >> PAGE_SHIFT); vm_stat_account(mm, vm_flags, file, len >> PAGE_SHIFT); if (vm_flags & VM_LOCKED) { - mm->locked_vm += len >> PAGE_SHIFT; + vx_vmlocked_add(mm, len >> PAGE_SHIFT); make_pages_present(addr, addr + len); } if (flags & MAP_POPULATE) { @@ -1502,9 +1502,9 @@ static int acct_stack_growth(struct vm_a return -ENOMEM; /* Ok, everything looks good - let it rip */ - mm->total_vm += grow; + vx_vmpages_add(mm, grow); if (vma->vm_flags & VM_LOCKED) - mm->locked_vm += grow; + vx_vmlocked_add(mm, grow); vm_stat_account(mm, vma->vm_flags, vma->vm_file, grow); return 0; } @@ -1662,9 +1662,9 @@ static void remove_vma_list(struct mm_st do { long nrpages = vma_pages(vma); - mm->total_vm -= nrpages; + vx_vmpages_sub(mm, nrpages); if (vma->vm_flags & VM_LOCKED) - mm->locked_vm -= nrpages; + vx_vmlocked_sub(mm, nrpages); vm_stat_account(mm, vma->vm_flags, vma->vm_file, -nrpages); vma = remove_vma(vma); } while (vma); @@ -1903,6 +1903,8 @@ unsigned long do_brk(unsigned long addr, lock_limit >>= PAGE_SHIFT; if (locked > lock_limit && !capable(CAP_IPC_LOCK)) return -EAGAIN; + if (!vx_vmlocked_avail(mm, len >> PAGE_SHIFT)) + return -ENOMEM; } /* @@ -1929,7 +1931,8 @@ unsigned long do_brk(unsigned long addr, if (mm->map_count > sysctl_max_map_count) return -ENOMEM; - if (security_vm_enough_memory(len >> PAGE_SHIFT)) + if (security_vm_enough_memory(len >> PAGE_SHIFT) || + !vx_vmpages_avail(mm, len >> PAGE_SHIFT)) return -ENOMEM; /* Can we just expand an old private anonymous mapping? */ @@ -1955,9 +1958,9 @@ unsigned long do_brk(unsigned long addr, (VM_READ|VM_WRITE|VM_EXEC|VM_SHARED)]; vma_link(mm, vma, prev, rb_link, rb_parent); out: - mm->total_vm += len >> PAGE_SHIFT; + vx_vmpages_add(mm, len >> PAGE_SHIFT); if (flags & VM_LOCKED) { - mm->locked_vm += len >> PAGE_SHIFT; + vx_vmlocked_add(mm, len >> PAGE_SHIFT); make_pages_present(addr, addr + len); } return addr; @@ -1986,6 +1989,11 @@ void exit_mmap(struct mm_struct *mm) free_pgtables(&tlb, vma, FIRST_USER_ADDRESS, 0); tlb_finish_mmu(tlb, 0, end); + set_mm_counter(mm, file_rss, 0); + set_mm_counter(mm, anon_rss, 0); + vx_vmpages_sub(mm, mm->total_vm); + vx_vmlocked_sub(mm, mm->locked_vm); + /* * Walk the list again, actually closing and freeing it, * with preemption enabled, without holding any MM locks. @@ -2025,7 +2033,8 @@ int insert_vm_struct(struct mm_struct * if (__vma && __vma->vm_start < vma->vm_end) return -ENOMEM; if ((vma->vm_flags & VM_ACCOUNT) && - security_vm_enough_memory(vma_pages(vma))) + (security_vm_enough_memory(vma_pages(vma)) || + !vx_vmpages_avail(mm, vma_pages(vma)))) return -ENOMEM; vma_link(mm, vma, prev, rb_link, rb_parent); return 0; @@ -2098,6 +2107,8 @@ int may_expand_vm(struct mm_struct *mm, if (cur + npages > lim) return 0; + if (!vx_vmpages_avail(mm, npages)) + return 0; return 1; } @@ -2168,7 +2179,6 @@ int install_special_mapping(struct mm_st return -ENOMEM; } - mm->total_vm += len >> PAGE_SHIFT; - + vx_vmpages_add(mm, len >> PAGE_SHIFT); return 0; } diff -NurpP --minimal linux-source-2.6.22.orig/mm/mremap.c linux-source-2.6.22/mm/mremap.c --- linux-source-2.6.22.orig/mm/mremap.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/mm/mremap.c 2007-10-24 16:14:06.791620780 +0200 @@ -18,6 +18,7 @@ #include #include #include +#include #include #include @@ -212,7 +213,7 @@ static unsigned long move_vma(struct vm_ * If this were a serious issue, we'd add a flag to do_munmap(). */ hiwater_vm = mm->hiwater_vm; - mm->total_vm += new_len >> PAGE_SHIFT; + vx_vmpages_add(mm, new_len >> PAGE_SHIFT); vm_stat_account(mm, vma->vm_flags, vma->vm_file, new_len>>PAGE_SHIFT); if (do_munmap(mm, old_addr, old_len) < 0) { @@ -230,7 +231,7 @@ static unsigned long move_vma(struct vm_ } if (vm_flags & VM_LOCKED) { - mm->locked_vm += new_len >> PAGE_SHIFT; + vx_vmlocked_add(mm, new_len >> PAGE_SHIFT); if (new_len > old_len) make_pages_present(new_addr + old_len, new_addr + new_len); @@ -337,6 +338,9 @@ unsigned long do_mremap(unsigned long ad ret = -EAGAIN; if (locked > lock_limit && !capable(CAP_IPC_LOCK)) goto out; + if (!vx_vmlocked_avail(current->mm, + (new_len - old_len) >> PAGE_SHIFT)) + goto out; } if (!may_expand_vm(mm, (new_len - old_len) >> PAGE_SHIFT)) { ret = -ENOMEM; @@ -365,10 +369,10 @@ unsigned long do_mremap(unsigned long ad vma_adjust(vma, vma->vm_start, addr + new_len, vma->vm_pgoff, NULL); - mm->total_vm += pages; + vx_vmpages_add(mm, pages); vm_stat_account(mm, vma->vm_flags, vma->vm_file, pages); if (vma->vm_flags & VM_LOCKED) { - mm->locked_vm += pages; + vx_vmlocked_add(mm, pages); make_pages_present(addr + old_len, addr + new_len); } diff -NurpP --minimal linux-source-2.6.22.orig/mm/nommu.c linux-source-2.6.22/mm/nommu.c --- linux-source-2.6.22.orig/mm/nommu.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/mm/nommu.c 2007-10-24 16:14:06.791620780 +0200 @@ -936,7 +936,7 @@ unsigned long do_mmap_pgoff(struct file realalloc += kobjsize(vma); askedalloc += sizeof(*vma); - current->mm->total_vm += len >> PAGE_SHIFT; + vx_vmpages_add(current->mm, len >> PAGE_SHIFT); add_nommu_vma(vma); @@ -1061,7 +1061,7 @@ int do_munmap(struct mm_struct *mm, unsi kfree(vml); update_hiwater_vm(mm); - mm->total_vm -= len >> PAGE_SHIFT; + vx_vmpages_sub(mm, len >> PAGE_SHIFT); #ifdef DEBUG show_process_blocks(); @@ -1093,7 +1093,7 @@ void exit_mmap(struct mm_struct * mm) printk("Exit_mmap:\n"); #endif - mm->total_vm = 0; + vx_vmpages_sub(mm, mm->total_vm); while ((tmp = mm->context.vmlist)) { mm->context.vmlist = tmp->next; diff -NurpP --minimal linux-source-2.6.22.orig/mm/oom_kill.c linux-source-2.6.22/mm/oom_kill.c --- linux-source-2.6.22.orig/mm/oom_kill.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/mm/oom_kill.c 2007-10-24 16:14:06.791620780 +0200 @@ -24,6 +24,7 @@ #include #include #include +#include int sysctl_panic_on_oom; /* #define DEBUG */ @@ -66,6 +67,12 @@ unsigned long badness(struct task_struct points = mm->total_vm; /* + * add points for context badness + */ + + points += vx_badness(p, mm); + + /* * After this unlock we can no longer dereference local variable `mm' */ task_unlock(p); @@ -156,8 +163,8 @@ unsigned long badness(struct task_struct } #ifdef DEBUG - printk(KERN_DEBUG "OOMkill: task %d (%s) got %d points\n", - p->pid, p->comm, points); + printk(KERN_DEBUG "OOMkill: task %d:#%u (%s) got %d points\n", + p->pid, p->xid, p->comm, points); #endif return points; } @@ -288,7 +295,8 @@ static void __oom_kill_task(struct task_ } if (verbose) - printk(KERN_ERR "Killed process %d (%s)\n", p->pid, p->comm); + printk(KERN_ERR "Killed process %d:#%u (%s)\n", + p->pid, p->xid, p->comm); /* * We give our sacrificial lamb high priority and access to @@ -358,8 +366,8 @@ static int oom_kill_process(struct task_ return 0; } - printk(KERN_ERR "%s: kill process %d (%s) score %li or a child\n", - message, p->pid, p->comm, points); + printk(KERN_ERR "%s: kill process %d:#%u (%s) score %li or a child\n", + message, p->pid, p->xid, p->comm, points); /* Try to kill a child first */ list_for_each(tsk, &p->children) { diff -NurpP --minimal linux-source-2.6.22.orig/mm/page_alloc.c linux-source-2.6.22/mm/page_alloc.c --- linux-source-2.6.22.orig/mm/page_alloc.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/mm/page_alloc.c 2007-10-24 16:14:06.791620780 +0200 @@ -41,6 +41,8 @@ #include #include #include +#include +#include #include #include @@ -1488,6 +1490,9 @@ void si_meminfo(struct sysinfo *val) val->totalhigh = totalhigh_pages; val->freehigh = nr_free_highpages(); val->mem_unit = PAGE_SIZE; + + if (vx_flags(VXF_VIRT_MEM, 0)) + vx_vsi_meminfo(val); } EXPORT_SYMBOL(si_meminfo); @@ -1508,6 +1513,9 @@ void si_meminfo_node(struct sysinfo *val val->freehigh = 0; #endif val->mem_unit = PAGE_SIZE; + + if (vx_flags(VXF_VIRT_MEM, 0)) + vx_vsi_meminfo(val); } #endif diff -NurpP --minimal linux-source-2.6.22.orig/mm/rmap.c linux-source-2.6.22/mm/rmap.c --- linux-source-2.6.22.orig/mm/rmap.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/mm/rmap.c 2007-10-24 16:14:06.791620780 +0200 @@ -48,6 +48,7 @@ #include #include #include +#include #include diff -NurpP --minimal linux-source-2.6.22.orig/mm/shmem.c linux-source-2.6.22/mm/shmem.c --- linux-source-2.6.22.orig/mm/shmem.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/mm/shmem.c 2007-10-24 16:14:06.791620780 +0200 @@ -55,7 +55,6 @@ #include /* This magic number is used in glibc for posix shared memory */ -#define TMPFS_MAGIC 0x01021994 #define ENTRIES_PER_PAGE (PAGE_CACHE_SIZE/sizeof(unsigned long)) #define ENTRIES_PER_PAGEPAGE (ENTRIES_PER_PAGE*ENTRIES_PER_PAGE) @@ -1734,7 +1733,7 @@ static int shmem_statfs(struct dentry *d { struct shmem_sb_info *sbinfo = SHMEM_SB(dentry->d_sb); - buf->f_type = TMPFS_MAGIC; + buf->f_type = TMPFS_SUPER_MAGIC; buf->f_bsize = PAGE_CACHE_SIZE; buf->f_namelen = NAME_MAX; spin_lock(&sbinfo->stat_lock); @@ -2308,7 +2307,7 @@ static int shmem_fill_super(struct super sb->s_maxbytes = SHMEM_MAX_BYTES; sb->s_blocksize = PAGE_CACHE_SIZE; sb->s_blocksize_bits = PAGE_CACHE_SHIFT; - sb->s_magic = TMPFS_MAGIC; + sb->s_magic = TMPFS_SUPER_MAGIC; sb->s_op = &shmem_ops; sb->s_time_gran = 1; #ifdef CONFIG_TMPFS_POSIX_ACL diff -NurpP --minimal linux-source-2.6.22.orig/mm/slab.c linux-source-2.6.22/mm/slab.c --- linux-source-2.6.22.orig/mm/slab.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/mm/slab.c 2007-10-24 16:14:06.791620780 +0200 @@ -510,6 +510,8 @@ struct kmem_cache { #define STATS_INC_FREEMISS(x) do { } while (0) #endif +#include "slab_vs.h" + #if DEBUG /* @@ -3314,6 +3316,7 @@ retry: obj = slab_get_obj(cachep, slabp, nodeid); check_slabp(cachep, slabp); + vx_slab_alloc(cachep, flags); l3->free_objects--; /* move slabp to correct slabp list: */ list_del(&slabp->list); @@ -3386,6 +3389,7 @@ __cache_alloc_node(struct kmem_cache *ca /* ___cache_alloc_node can fall back to other nodes */ ptr = ____cache_alloc_node(cachep, flags, nodeid); out: + vx_slab_alloc(cachep, flags); local_irq_restore(save_flags); ptr = cache_alloc_debugcheck_after(cachep, flags, ptr, caller); @@ -3551,6 +3555,7 @@ static inline void __cache_free(struct k check_irq_off(); objp = cache_free_debugcheck(cachep, objp, __builtin_return_address(0)); + vx_slab_free(cachep); if (cache_free_alien(cachep, objp)) return; diff -NurpP --minimal linux-source-2.6.22.orig/mm/slab_vs.h linux-source-2.6.22/mm/slab_vs.h --- linux-source-2.6.22.orig/mm/slab_vs.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-source-2.6.22/mm/slab_vs.h 2007-10-24 16:14:06.791620780 +0200 @@ -0,0 +1,27 @@ + +#include + +#include + +static inline +void vx_slab_alloc(struct kmem_cache *cachep, gfp_t flags) +{ + int what = gfp_zone(cachep->gfpflags); + + if (!current->vx_info) + return; + + atomic_add(cachep->buffer_size, ¤t->vx_info->cacct.slab[what]); +} + +static inline +void vx_slab_free(struct kmem_cache *cachep) +{ + int what = gfp_zone(cachep->gfpflags); + + if (!current->vx_info) + return; + + atomic_sub(cachep->buffer_size, ¤t->vx_info->cacct.slab[what]); +} + diff -NurpP --minimal linux-source-2.6.22.orig/mm/swapfile.c linux-source-2.6.22/mm/swapfile.c --- linux-source-2.6.22.orig/mm/swapfile.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/mm/swapfile.c 2007-10-24 16:14:06.791620780 +0200 @@ -31,6 +31,8 @@ #include #include #include +#include +#include DEFINE_SPINLOCK(swap_lock); unsigned int nr_swapfiles; @@ -1712,6 +1714,8 @@ void si_swapinfo(struct sysinfo *val) val->freeswap = nr_swap_pages + nr_to_be_unused; val->totalswap = total_swap_pages + nr_to_be_unused; spin_unlock(&swap_lock); + if (vx_flags(VXF_VIRT_MEM, 0)) + vx_vsi_swapinfo(val); } /* diff -NurpP --minimal linux-source-2.6.22.orig/net/core/dev.c linux-source-2.6.22/net/core/dev.c --- linux-source-2.6.22.orig/net/core/dev.c 2007-10-15 00:33:08.000000000 +0200 +++ linux-source-2.6.22/net/core/dev.c 2007-10-24 16:14:06.791620780 +0200 @@ -117,6 +117,8 @@ #include #include #include +#include /* remove with NXF_HIDE_NETIF */ +#include /* * The list of packet types we will receive (as opposed to discard) @@ -2120,6 +2122,9 @@ static int dev_ifconf(char __user *arg) total = 0; for_each_netdev(dev) { + if (vx_flags(VXF_HIDE_NETIF, 0) && + !dev_in_nx_info(dev, current->nx_info)) + continue; for (i = 0; i < NPROTO; i++) { if (gifconf_list[i]) { int done; @@ -2183,6 +2188,10 @@ void dev_seq_stop(struct seq_file *seq, static void dev_seq_printf_stats(struct seq_file *seq, struct net_device *dev) { struct net_device_stats *stats = dev->get_stats(dev); + struct nx_info *nxi = current->nx_info; + + if (vx_flags(VXF_HIDE_NETIF, 0) && !dev_in_nx_info(dev, nxi)) + return; seq_printf(seq, "%6s:%8lu %7lu %4lu %4lu %4lu %5lu %10lu %9lu " "%8lu %7lu %4lu %4lu %4lu %5lu %7lu %10lu\n", diff -NurpP --minimal linux-source-2.6.22.orig/net/core/rtnetlink.c linux-source-2.6.22/net/core/rtnetlink.c --- linux-source-2.6.22.orig/net/core/rtnetlink.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/net/core/rtnetlink.c 2007-10-24 16:14:06.791620780 +0200 @@ -35,6 +35,7 @@ #include #include #include +#include /* remove with NXF_HIDE_NETIF */ #include #include @@ -537,6 +538,9 @@ static int rtnl_dump_ifinfo(struct sk_bu idx = 0; for_each_netdev(dev) { + if (vx_info_flags(skb->sk->sk_vx_info, VXF_HIDE_NETIF, 0) && + !dev_in_nx_info(dev, skb->sk->sk_nx_info)) + continue; if (idx < s_idx) goto cont; if (rtnl_fill_ifinfo(skb, dev, RTM_NEWLINK, diff -NurpP --minimal linux-source-2.6.22.orig/net/core/sock.c linux-source-2.6.22/net/core/sock.c --- linux-source-2.6.22.orig/net/core/sock.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/net/core/sock.c 2007-10-24 16:14:06.791620780 +0200 @@ -125,6 +125,9 @@ #include #include +#include +#include +#include #ifdef CONFIG_INET #include @@ -869,6 +872,8 @@ struct sock *sk_alloc(int family, gfp_t sk->sk_prot = sk->sk_prot_creator = prot; sock_lock_init(sk); } + sock_vx_init(sk); + sock_nx_init(sk); if (security_sk_alloc(sk, family, priority)) goto out_free; @@ -907,6 +912,11 @@ void sk_free(struct sock *sk) __FUNCTION__, atomic_read(&sk->sk_omem_alloc)); security_sk_free(sk); + vx_sock_dec(sk); + clr_vx_info(&sk->sk_vx_info); + sk->sk_xid = -1; + clr_nx_info(&sk->sk_nx_info); + sk->sk_nid = -1; if (sk->sk_prot_creator->slab != NULL) kmem_cache_free(sk->sk_prot_creator->slab, sk); else @@ -924,6 +934,8 @@ struct sock *sk_clone(const struct sock sock_copy(newsk, sk); /* SANITY */ + sock_vx_init(newsk); + sock_nx_init(newsk); sk_node_init(&newsk->sk_node); sock_lock_init(newsk); bh_lock_sock(newsk); @@ -969,6 +981,12 @@ struct sock *sk_clone(const struct sock newsk->sk_priority = 0; atomic_set(&newsk->sk_refcnt, 2); + set_vx_info(&newsk->sk_vx_info, sk->sk_vx_info); + newsk->sk_xid = sk->sk_xid; + vx_sock_inc(newsk); + set_nx_info(&newsk->sk_nx_info, sk->sk_nx_info); + newsk->sk_nid = sk->sk_nid; + /* * Increment the counter in the same struct proto as the master * sock (sk_refcnt_debug_inc uses newsk->sk_prot->socks, that @@ -1551,6 +1569,11 @@ void sock_init_data(struct socket *sock, sk->sk_stamp = ktime_set(-1L, -1L); + set_vx_info(&sk->sk_vx_info, current->vx_info); + sk->sk_xid = vx_current_xid(); + vx_sock_inc(sk); + set_nx_info(&sk->sk_nx_info, current->nx_info); + sk->sk_nid = nx_current_nid(); atomic_set(&sk->sk_refcnt, 1); } diff -NurpP --minimal linux-source-2.6.22.orig/net/ipv4/af_inet.c linux-source-2.6.22/net/ipv4/af_inet.c --- linux-source-2.6.22.orig/net/ipv4/af_inet.c 2007-10-15 00:33:08.000000000 +0200 +++ linux-source-2.6.22/net/ipv4/af_inet.c 2007-10-24 16:14:06.791620780 +0200 @@ -115,6 +115,7 @@ #ifdef CONFIG_IP_MROUTE #include #endif +#include DEFINE_SNMP_STAT(struct linux_mib, net_statistics) __read_mostly; @@ -308,9 +309,11 @@ lookup_protocol: } err = -EPERM; + if ((protocol == IPPROTO_ICMP) && vx_ccaps(VXC_RAW_ICMP)) + goto override; if (answer->capability > 0 && !capable(answer->capability)) goto out_rcu_unlock; - +override: sock->ops = answer->ops; answer_prot = answer->prot; answer_no_check = answer->no_check; @@ -427,6 +430,10 @@ int inet_bind(struct socket *sock, struc unsigned short snum; int chk_addr_ret; int err; + __u32 s_addr; /* Address used for validation */ + __u32 s_addr1; /* Address used for socket */ + __u32 s_addr2; /* Broadcast address for the socket */ + struct nx_info *nxi = sk->sk_nx_info; /* If the socket has its own bind function then use it. (RAW) */ if (sk->sk_prot->bind) { @@ -437,7 +444,40 @@ int inet_bind(struct socket *sock, struc if (addr_len < sizeof(struct sockaddr_in)) goto out; - chk_addr_ret = inet_addr_type(addr->sin_addr.s_addr); + s_addr = addr->sin_addr.s_addr; + s_addr1 = s_addr; + s_addr2 = 0xffffffffl; + + vxdprintk(VXD_CBIT(net, 3), + "inet_bind(%p)* %p,%p;%lx " NIPQUAD_FMT, + sk, sk->sk_nx_info, sk->sk_socket, + (sk->sk_socket?sk->sk_socket->flags:0), + NIPQUAD(s_addr)); + if (nxi) { + __u32 v4_bcast = nxi->v4_bcast; + __u32 ipv4root = nxi->ipv4[0]; + int nbipv4 = nxi->nbipv4; + + if (s_addr == 0) { + /* bind to any for 1-n */ + s_addr = ipv4root; + s_addr1 = (nbipv4 > 1) ? 0 : s_addr; + s_addr2 = v4_bcast; + } else if (s_addr == IPI_LOOPBACK) { + /* rewrite localhost to ipv4root */ + s_addr = ipv4root; + s_addr1 = ipv4root; + } else if (s_addr != v4_bcast) { + /* normal address bind */ + if (!addr_in_nx_info(nxi, s_addr)) + return -EADDRNOTAVAIL; + } + } + chk_addr_ret = inet_addr_type(s_addr); + + vxdprintk(VXD_CBIT(net, 3), + "inet_bind(%p) " NIPQUAD_FMT ", " NIPQUAD_FMT ", " NIPQUAD_FMT, + sk, NIPQUAD(s_addr), NIPQUAD(s_addr1), NIPQUAD(s_addr2)); /* Not specified by any standard per-se, however it breaks too * many applications when removed. It is unfortunate since @@ -449,7 +489,7 @@ int inet_bind(struct socket *sock, struc err = -EADDRNOTAVAIL; if (!sysctl_ip_nonlocal_bind && !inet->freebind && - addr->sin_addr.s_addr != INADDR_ANY && + s_addr != INADDR_ANY && chk_addr_ret != RTN_LOCAL && chk_addr_ret != RTN_MULTICAST && chk_addr_ret != RTN_BROADCAST) @@ -474,7 +514,8 @@ int inet_bind(struct socket *sock, struc if (sk->sk_state != TCP_CLOSE || inet->num) goto out_release_sock; - inet->rcv_saddr = inet->saddr = addr->sin_addr.s_addr; + inet->rcv_saddr = inet->saddr = s_addr1; + inet->rcv_saddr2 = s_addr2; if (chk_addr_ret == RTN_MULTICAST || chk_addr_ret == RTN_BROADCAST) inet->saddr = 0; /* Use device */ diff -NurpP --minimal linux-source-2.6.22.orig/net/ipv4/devinet.c linux-source-2.6.22/net/ipv4/devinet.c --- linux-source-2.6.22.orig/net/ipv4/devinet.c 2007-10-15 00:33:08.000000000 +0200 +++ linux-source-2.6.22/net/ipv4/devinet.c 2007-10-24 16:14:06.791620780 +0200 @@ -56,6 +56,7 @@ #include #endif #include +#include #include #include @@ -676,6 +677,9 @@ int devinet_ioctl(unsigned int cmd, void *colon = ':'; if ((in_dev = __in_dev_get_rtnl(dev)) != NULL) { + struct nx_info *nxi = current->nx_info; + int hide_netif = vx_flags(VXF_HIDE_NETIF, 0); + if (tryaddrmatch) { /* Matthias Andree */ /* compare label and address (4.4BSD style) */ @@ -684,6 +688,8 @@ int devinet_ioctl(unsigned int cmd, void This is checked above. */ for (ifap = &in_dev->ifa_list; (ifa = *ifap) != NULL; ifap = &ifa->ifa_next) { + if (hide_netif && !ifa_in_nx_info(ifa, nxi)) + continue; if (!strcmp(ifr.ifr_name, ifa->ifa_label) && sin_orig.sin_addr.s_addr == ifa->ifa_address) { @@ -696,9 +702,12 @@ int devinet_ioctl(unsigned int cmd, void comparing just the label */ if (!ifa) { for (ifap = &in_dev->ifa_list; (ifa = *ifap) != NULL; - ifap = &ifa->ifa_next) + ifap = &ifa->ifa_next) { + if (hide_netif && !ifa_in_nx_info(ifa, nxi)) + continue; if (!strcmp(ifr.ifr_name, ifa->ifa_label)) break; + } } } @@ -849,6 +858,9 @@ static int inet_gifconf(struct net_devic goto out; for (; ifa; ifa = ifa->ifa_next) { + if (vx_flags(VXF_HIDE_NETIF, 0) && + !ifa_in_nx_info(ifa, current->nx_info)) + continue; if (!buf) { done += sizeof(ifr); continue; @@ -1179,6 +1191,7 @@ static int inet_dump_ifaddr(struct sk_bu struct net_device *dev; struct in_device *in_dev; struct in_ifaddr *ifa; + struct sock *sk = skb->sk; int s_ip_idx, s_idx = cb->args[0]; s_ip_idx = ip_idx = cb->args[1]; @@ -1193,6 +1206,9 @@ static int inet_dump_ifaddr(struct sk_bu for (ifa = in_dev->ifa_list, ip_idx = 0; ifa; ifa = ifa->ifa_next, ip_idx++) { + if (sk && vx_info_flags(sk->sk_vx_info, VXF_HIDE_NETIF, 0) && + !ifa_in_nx_info(ifa, sk->sk_nx_info)) + continue; if (ip_idx < s_ip_idx) continue; if (inet_fill_ifaddr(skb, ifa, NETLINK_CB(cb->skb).pid, diff -NurpP --minimal linux-source-2.6.22.orig/net/ipv4/fib_hash.c linux-source-2.6.22/net/ipv4/fib_hash.c --- linux-source-2.6.22.orig/net/ipv4/fib_hash.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/net/ipv4/fib_hash.c 2007-10-24 16:14:06.791620780 +0200 @@ -34,6 +34,7 @@ #include #include #include +#include #include #include @@ -982,6 +983,8 @@ static unsigned fib_flag_trans(int type, return flags; } +extern int dev_in_nx_info(struct net_device *, struct nx_info *); + /* * This outputs /proc/net/route. * @@ -1012,7 +1015,8 @@ static int fib_seq_show(struct seq_file prefix = f->fn_key; mask = FZ_MASK(iter->zone); flags = fib_flag_trans(fa->fa_type, mask, fi); - if (fi) + if (fi && (!vx_flags(VXF_HIDE_NETIF, 0) || + dev_in_nx_info(fi->fib_dev, current->nx_info))) snprintf(bf, sizeof(bf), "%s\t%08X\t%08X\t%04X\t%d\t%u\t%d\t%08X\t%d\t%u\t%u", fi->fib_dev ? fi->fib_dev->name : "*", prefix, diff -NurpP --minimal linux-source-2.6.22.orig/net/ipv4/inet_connection_sock.c linux-source-2.6.22/net/ipv4/inet_connection_sock.c --- linux-source-2.6.22.orig/net/ipv4/inet_connection_sock.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/net/ipv4/inet_connection_sock.c 2007-10-24 16:14:06.791620780 +0200 @@ -37,7 +37,6 @@ int sysctl_local_port_range[2] = { 32768 int inet_csk_bind_conflict(const struct sock *sk, const struct inet_bind_bucket *tb) { - const __be32 sk_rcv_saddr = inet_rcv_saddr(sk); struct sock *sk2; struct hlist_node *node; int reuse = sk->sk_reuse; @@ -50,9 +49,8 @@ int inet_csk_bind_conflict(const struct sk->sk_bound_dev_if == sk2->sk_bound_dev_if)) { if (!reuse || !sk2->sk_reuse || sk2->sk_state == TCP_LISTEN) { - const __be32 sk2_rcv_saddr = inet_rcv_saddr(sk2); - if (!sk2_rcv_saddr || !sk_rcv_saddr || - sk2_rcv_saddr == sk_rcv_saddr) + if (nx_addr_conflict(sk->sk_nx_info, + inet_rcv_saddr(sk), sk2)) break; } } diff -NurpP --minimal linux-source-2.6.22.orig/net/ipv4/inet_diag.c linux-source-2.6.22/net/ipv4/inet_diag.c --- linux-source-2.6.22.orig/net/ipv4/inet_diag.c 2007-10-15 00:33:08.000000000 +0200 +++ linux-source-2.6.22/net/ipv4/inet_diag.c 2007-10-24 16:14:06.791620780 +0200 @@ -697,6 +697,8 @@ static int inet_diag_dump(struct sk_buff sk_for_each(sk, node, &hashinfo->listening_hash[i]) { struct inet_sock *inet = inet_sk(sk); + if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT)) + continue; if (num < s_num) { num++; continue; @@ -757,6 +759,8 @@ skip_listen_ht: sk_for_each(sk, node, &head->chain) { struct inet_sock *inet = inet_sk(sk); + if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT)) + continue; if (num < s_num) goto next_normal; if (!(r->idiag_states & (1 << sk->sk_state))) @@ -781,6 +785,8 @@ next_normal: inet_twsk_for_each(tw, node, &head->twchain) { + if (!nx_check(tw->tw_nid, VS_WATCH_P | VS_IDENT)) + continue; if (num < s_num) goto next_dying; if (r->id.idiag_sport != tw->tw_sport && diff -NurpP --minimal linux-source-2.6.22.orig/net/ipv4/inet_hashtables.c linux-source-2.6.22/net/ipv4/inet_hashtables.c --- linux-source-2.6.22.orig/net/ipv4/inet_hashtables.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/net/ipv4/inet_hashtables.c 2007-10-24 16:14:06.791620780 +0200 @@ -140,11 +140,10 @@ static struct sock *inet_lookup_listener const __be32 rcv_saddr = inet->rcv_saddr; int score = sk->sk_family == PF_INET ? 1 : 0; - if (rcv_saddr) { - if (rcv_saddr != daddr) - continue; + if (inet_addr_match(sk->sk_nx_info, daddr, rcv_saddr)) score += 2; - } + else + continue; if (sk->sk_bound_dev_if) { if (sk->sk_bound_dev_if != dif) continue; @@ -175,7 +174,7 @@ struct sock *__inet_lookup_listener(stru const struct inet_sock *inet = inet_sk((sk = __sk_head(head))); if (inet->num == hnum && !sk->sk_node.next && - (!inet->rcv_saddr || inet->rcv_saddr == daddr) && + inet_addr_match(sk->sk_nx_info, daddr, inet->rcv_saddr) && (sk->sk_family == PF_INET || !ipv6_only_sock(sk)) && !sk->sk_bound_dev_if) goto sherry_cache; diff -NurpP --minimal linux-source-2.6.22.orig/net/ipv4/raw.c linux-source-2.6.22/net/ipv4/raw.c --- linux-source-2.6.22.orig/net/ipv4/raw.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/net/ipv4/raw.c 2007-10-24 16:14:06.791620780 +0200 @@ -101,6 +101,27 @@ static void raw_v4_unhash(struct sock *s write_unlock_bh(&raw_v4_lock); } + +/* + * Check if a given address matches for a socket + * + * nxi: the socket's nx_info if any + * addr: to be verified address + * saddr/baddr: socket addresses + */ +static inline int raw_addr_match ( + struct nx_info *nxi, + uint32_t addr, + uint32_t saddr, + uint32_t baddr) +{ + if (addr && (saddr == addr || baddr == addr)) + return 1; + if (!saddr) + return addr_in_nx_info(nxi, addr); + return 0; +} + struct sock *__raw_v4_lookup(struct sock *sk, unsigned short num, __be32 raddr, __be32 laddr, int dif) @@ -112,7 +133,8 @@ struct sock *__raw_v4_lookup(struct sock if (inet->num == num && !(inet->daddr && inet->daddr != raddr) && - !(inet->rcv_saddr && inet->rcv_saddr != laddr) && + raw_addr_match(sk->sk_nx_info, laddr, + inet->rcv_saddr, inet->rcv_saddr2) && !(sk->sk_bound_dev_if && sk->sk_bound_dev_if != dif)) goto found; /* gotcha */ } @@ -314,6 +336,11 @@ static int raw_send_hdrinc(struct sock * iph->check = ip_fast_csum((unsigned char *)iph, iph->ihl); } + err = -EPERM; + if (!nx_check(0, VS_ADMIN) && !capable(CAP_NET_RAW) + && (!addr_in_nx_info(sk->sk_nx_info, iph->saddr))) + goto error_free; + err = NF_HOOK(PF_INET, NF_IP_LOCAL_OUT, skb, NULL, rt->u.dst.dev, dst_output); if (err > 0) @@ -325,6 +352,7 @@ out: error_fault: err = -EFAULT; +error_free: kfree_skb(skb); error: IP_INC_STATS(IPSTATS_MIB_OUTDISCARDS); @@ -491,6 +519,12 @@ static int raw_sendmsg(struct kiocb *ioc } security_sk_classify_flow(sk, &fl); + if (sk->sk_nx_info) { + err = ip_find_src(sk->sk_nx_info, &rt, &fl); + + if (err) + goto done; + } err = ip_route_output_flow(&rt, &fl, sk, 1); } if (err) @@ -795,7 +829,8 @@ static struct sock *raw_get_first(struct struct hlist_node *node; sk_for_each(sk, node, &raw_v4_htable[state->bucket]) - if (sk->sk_family == PF_INET) + if (sk->sk_family == PF_INET && + nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT)) goto found; } sk = NULL; @@ -811,7 +846,8 @@ static struct sock *raw_get_next(struct sk = sk_next(sk); try_again: ; - } while (sk && sk->sk_family != PF_INET); + } while (sk && (sk->sk_family != PF_INET || + !nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))); if (!sk && ++state->bucket < RAWV4_HTABLE_SIZE) { sk = sk_head(&raw_v4_htable[state->bucket]); diff -NurpP --minimal linux-source-2.6.22.orig/net/ipv4/tcp.c linux-source-2.6.22/net/ipv4/tcp.c --- linux-source-2.6.22.orig/net/ipv4/tcp.c 2007-10-15 00:33:08.000000000 +0200 +++ linux-source-2.6.22/net/ipv4/tcp.c 2007-10-24 16:14:06.791620780 +0200 @@ -258,6 +258,7 @@ #include #include #include +#include #include #include diff -NurpP --minimal linux-source-2.6.22.orig/net/ipv4/tcp_ipv4.c linux-source-2.6.22/net/ipv4/tcp_ipv4.c --- linux-source-2.6.22.orig/net/ipv4/tcp_ipv4.c 2007-10-15 00:33:08.000000000 +0200 +++ linux-source-2.6.22/net/ipv4/tcp_ipv4.c 2007-10-24 16:14:06.791620780 +0200 @@ -1975,6 +1975,12 @@ static void *listening_get_next(struct s req = req->dl_next; while (1) { while (req) { + vxdprintk(VXD_CBIT(net, 6), + "sk,req: %p [#%d] (from %d)", req->sk, + (req->sk)?req->sk->sk_nid:0, nx_current_nid()); + if (req->sk && + !nx_check(req->sk->sk_nid, VS_WATCH_P | VS_IDENT)) + continue; if (req->rsk_ops->family == st->family) { cur = req; goto out; @@ -1999,6 +2005,10 @@ get_req: } get_sk: sk_for_each_from(sk, node) { + vxdprintk(VXD_CBIT(net, 6), "sk: %p [#%d] (from %d)", + sk, sk->sk_nid, nx_current_nid()); + if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT)) + continue; if (sk->sk_family == st->family) { cur = sk; goto out; @@ -2050,18 +2060,26 @@ static void *established_get_first(struc read_lock(&tcp_hashinfo.ehash[st->bucket].lock); sk_for_each(sk, node, &tcp_hashinfo.ehash[st->bucket].chain) { - if (sk->sk_family != st->family) { + vxdprintk(VXD_CBIT(net, 6), + "sk,egf: %p [#%d] (from %d)", + sk, sk->sk_nid, nx_current_nid()); + if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT)) + continue; + if (sk->sk_family != st->family) continue; - } rc = sk; goto out; } st->state = TCP_SEQ_STATE_TIME_WAIT; inet_twsk_for_each(tw, node, &tcp_hashinfo.ehash[st->bucket].twchain) { - if (tw->tw_family != st->family) { + vxdprintk(VXD_CBIT(net, 6), + "tw: %p [#%d] (from %d)", + tw, tw->tw_nid, nx_current_nid()); + if (!nx_check(tw->tw_nid, VS_WATCH_P | VS_IDENT)) + continue; + if (tw->tw_family != st->family) continue; - } rc = tw; goto out; } @@ -2085,7 +2103,8 @@ static void *established_get_next(struct tw = cur; tw = tw_next(tw); get_tw: - while (tw && tw->tw_family != st->family) { + while (tw && (tw->tw_family != st->family || + !nx_check(tw->tw_nid, VS_WATCH_P | VS_IDENT))) { tw = tw_next(tw); } if (tw) { @@ -2109,6 +2128,11 @@ get_tw: sk = sk_next(sk); sk_for_each_from(sk, node) { + vxdprintk(VXD_CBIT(net, 6), + "sk,egn: %p [#%d] (from %d)", + sk, sk->sk_nid, nx_current_nid()); + if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT)) + continue; if (sk->sk_family == st->family) goto found; } diff -NurpP --minimal linux-source-2.6.22.orig/net/ipv4/tcp_minisocks.c linux-source-2.6.22/net/ipv4/tcp_minisocks.c --- linux-source-2.6.22.orig/net/ipv4/tcp_minisocks.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/net/ipv4/tcp_minisocks.c 2007-10-24 16:14:06.791620780 +0200 @@ -28,6 +28,10 @@ #include #include +#include +#include +#include + #ifdef CONFIG_SYSCTL #define SYNC_INIT 0 /* let the user enable it */ #else @@ -293,6 +297,11 @@ void tcp_time_wait(struct sock *sk, int tcptw->tw_ts_recent = tp->rx_opt.ts_recent; tcptw->tw_ts_recent_stamp = tp->rx_opt.ts_recent_stamp; + tw->tw_xid = sk->sk_xid; + tw->tw_vx_info = NULL; + tw->tw_nid = sk->sk_nid; + tw->tw_nx_info = NULL; + #if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) if (tw->tw_family == PF_INET6) { struct ipv6_pinfo *np = inet6_sk(sk); diff -NurpP --minimal linux-source-2.6.22.orig/net/ipv4/udp.c linux-source-2.6.22/net/ipv4/udp.c --- linux-source-2.6.22.orig/net/ipv4/udp.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/net/ipv4/udp.c 2007-10-24 16:14:06.791620780 +0200 @@ -221,11 +221,8 @@ int udp_get_port(struct sock *sk, unsign int ipv4_rcv_saddr_equal(const struct sock *sk1, const struct sock *sk2) { - struct inet_sock *inet1 = inet_sk(sk1), *inet2 = inet_sk(sk2); - return ( !ipv6_only_sock(sk2) && - (!inet1->rcv_saddr || !inet2->rcv_saddr || - inet1->rcv_saddr == inet2->rcv_saddr )); + nx_addr_conflict(sk1->sk_nx_info, inet_rcv_saddr(sk1), sk2)); } static inline int udp_v4_get_port(struct sock *sk, unsigned short snum) @@ -246,15 +243,22 @@ static struct sock *__udp4_lib_lookup(__ int badness = -1; read_lock(&udp_hash_lock); + sk_for_each(sk, node, &udptable[hnum & (UDP_HTABLE_SIZE - 1)]) { struct inet_sock *inet = inet_sk(sk); if (sk->sk_hash == hnum && !ipv6_only_sock(sk)) { int score = (sk->sk_family == PF_INET ? 1 : 0); + if (inet->rcv_saddr) { if (inet->rcv_saddr != daddr) continue; score+=2; + } else if (sk->sk_nx_info) { + if (addr_in_nx_info(sk->sk_nx_info, daddr)) + score+=2; + else + continue; } if (inet->daddr) { if (inet->daddr != saddr) @@ -280,6 +284,7 @@ static struct sock *__udp4_lib_lookup(__ } } } + if (result) sock_hold(result); read_unlock(&udp_hash_lock); @@ -301,7 +306,8 @@ static inline struct sock *udp_v4_mcast_ if (s->sk_hash != hnum || (inet->daddr && inet->daddr != rmt_addr) || (inet->dport != rmt_port && inet->dport) || - (inet->rcv_saddr && inet->rcv_saddr != loc_addr) || + (inet->rcv_saddr && inet->rcv_saddr != loc_addr && + inet->rcv_saddr2 && inet->rcv_saddr2 != loc_addr) || ipv6_only_sock(s) || (s->sk_bound_dev_if && s->sk_bound_dev_if != dif)) continue; @@ -631,7 +637,20 @@ int udp_sendmsg(struct kiocb *iocb, stru .uli_u = { .ports = { .sport = inet->sport, .dport = dport } } }; + struct nx_info *nxi = sk->sk_nx_info; + security_sk_classify_flow(sk, &fl); + if (nxi) { + err = ip_find_src(nxi, &rt, &fl); + if (err) + goto out; + if (daddr == IPI_LOOPBACK && !nx_check(0, VS_ADMIN)) + daddr = fl.fl4_dst = nxi->ipv4[0]; +#ifdef CONFIG_VSERVER_REMAP_SADDR + if (saddr == IPI_LOOPBACK && !nx_check(0, VS_ADMIN)) + saddr = fl.fl4_src = nxi->ipv4[0]; +#endif + } err = ip_route_output_flow(&rt, &fl, sk, 1); if (err) { if (err == -ENETUNREACH) @@ -1551,7 +1570,8 @@ static struct sock *udp_get_first(struct for (state->bucket = 0; state->bucket < UDP_HTABLE_SIZE; ++state->bucket) { struct hlist_node *node; sk_for_each(sk, node, state->hashtable + state->bucket) { - if (sk->sk_family == state->family) + if (sk->sk_family == state->family && + nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT)) goto found; } } @@ -1568,7 +1588,8 @@ static struct sock *udp_get_next(struct sk = sk_next(sk); try_again: ; - } while (sk && sk->sk_family != state->family); + } while (sk && (sk->sk_family != state->family || + !nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))); if (!sk && ++state->bucket < UDP_HTABLE_SIZE) { sk = sk_head(state->hashtable + state->bucket); diff -NurpP --minimal linux-source-2.6.22.orig/net/ipv6/addrconf.c linux-source-2.6.22/net/ipv6/addrconf.c --- linux-source-2.6.22.orig/net/ipv6/addrconf.c 2007-10-15 00:33:08.000000000 +0200 +++ linux-source-2.6.22/net/ipv6/addrconf.c 2007-10-24 16:14:06.791620780 +0200 @@ -2772,7 +2772,10 @@ static void if6_seq_stop(struct seq_file static int if6_seq_show(struct seq_file *seq, void *v) { struct inet6_ifaddr *ifp = (struct inet6_ifaddr *)v; - seq_printf(seq, + + /* no ipv6 inside a vserver for now */ + if (nx_check(0, VS_ADMIN|VS_WATCH)) + seq_printf(seq, NIP6_SEQFMT " %02x %02x %02x %02x %8s\n", NIP6(ifp->addr), ifp->idev->dev->ifindex, @@ -3255,6 +3258,10 @@ static int inet6_dump_addr(struct sk_buf struct ifmcaddr6 *ifmca; struct ifacaddr6 *ifaca; + /* no ipv6 inside a vserver for now */ + if (skb->sk && skb->sk->sk_vx_info) + return skb->len; + s_idx = cb->args[0]; s_ip_idx = ip_idx = cb->args[1]; @@ -3578,6 +3585,10 @@ static int inet6_dump_ifinfo(struct sk_b struct net_device *dev; struct inet6_dev *idev; + /* no ipv6 inside a vserver for now */ + if (skb->sk && skb->sk->sk_vx_info) + return skb->len; + read_lock(&dev_base_lock); idx = 0; for_each_netdev(dev) { diff -NurpP --minimal linux-source-2.6.22.orig/net/netlink/af_netlink.c linux-source-2.6.22/net/netlink/af_netlink.c --- linux-source-2.6.22.orig/net/netlink/af_netlink.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/net/netlink/af_netlink.c 2007-10-24 16:14:06.791620780 +0200 @@ -56,6 +56,9 @@ #include #include #include +#include +#include +#include #include #include diff -NurpP --minimal linux-source-2.6.22.orig/net/socket.c linux-source-2.6.22/net/socket.c --- linux-source-2.6.22.orig/net/socket.c 2007-10-15 00:33:08.000000000 +0200 +++ linux-source-2.6.22/net/socket.c 2007-10-24 16:14:06.791620780 +0200 @@ -92,6 +92,8 @@ #include #include +#include +#include static int sock_no_open(struct inode *irrelevant, struct file *dontcare); static ssize_t sock_aio_read(struct kiocb *iocb, const struct iovec *iov, @@ -543,7 +545,7 @@ static inline int __sock_sendmsg(struct struct msghdr *msg, size_t size) { struct sock_iocb *si = kiocb_to_siocb(iocb); - int err; + int err, len; si->sock = sock; si->scm = NULL; @@ -554,7 +556,22 @@ static inline int __sock_sendmsg(struct if (err) return err; - return sock->ops->sendmsg(iocb, sock, msg, size); + len = sock->ops->sendmsg(iocb, sock, msg, size); + if (sock->sk) { + if (len == size) + vx_sock_send(sock->sk, size); + else + vx_sock_fail(sock->sk, size); + } + vxdprintk(VXD_CBIT(net, 7), + "__sock_sendmsg: %p[%p,%p,%p;%d/%d]:%d/%d", + sock, sock->sk, + (sock->sk)?sock->sk->sk_nx_info:0, + (sock->sk)?sock->sk->sk_vx_info:0, + (sock->sk)?sock->sk->sk_xid:0, + (sock->sk)?sock->sk->sk_nid:0, + (unsigned int)size, len); + return len; } int sock_sendmsg(struct socket *sock, struct msghdr *msg, size_t size) @@ -623,7 +640,7 @@ EXPORT_SYMBOL_GPL(__sock_recv_timestamp) static inline int __sock_recvmsg(struct kiocb *iocb, struct socket *sock, struct msghdr *msg, size_t size, int flags) { - int err; + int err, len; struct sock_iocb *si = kiocb_to_siocb(iocb); si->sock = sock; @@ -636,7 +653,18 @@ static inline int __sock_recvmsg(struct if (err) return err; - return sock->ops->recvmsg(iocb, sock, msg, size, flags); + len = sock->ops->recvmsg(iocb, sock, msg, size, flags); + if ((len >= 0) && sock->sk) + vx_sock_recv(sock->sk, len); + vxdprintk(VXD_CBIT(net, 7), + "__sock_recvmsg: %p[%p,%p,%p;%d/%d]:%d/%d", + sock, sock->sk, + (sock->sk)?sock->sk->sk_nx_info:0, + (sock->sk)?sock->sk->sk_vx_info:0, + (sock->sk)?sock->sk->sk_xid:0, + (sock->sk)?sock->sk->sk_nid:0, + (unsigned int)size, len); + return len; } int sock_recvmsg(struct socket *sock, struct msghdr *msg, @@ -1090,6 +1118,10 @@ static int __sock_create(int family, int if (type < 0 || type >= SOCK_MAX) return -EINVAL; + /* disable IPv6 inside vservers for now */ + if (family == PF_INET6 && !nx_check(0, VS_ADMIN)) + return -EAFNOSUPPORT; + /* Compatibility. This uglymoron is moved from INET layer to here to avoid @@ -1207,6 +1239,7 @@ asmlinkage long sys_socket(int family, i if (retval < 0) goto out; + set_bit(SOCK_USER_SOCKET, &sock->flags); retval = sock_map_fd(sock); if (retval < 0) goto out_release; @@ -1239,10 +1272,12 @@ asmlinkage long sys_socketpair(int famil err = sock_create(family, type, protocol, &sock1); if (err < 0) goto out; + set_bit(SOCK_USER_SOCKET, &sock1->flags); err = sock_create(family, type, protocol, &sock2); if (err < 0) goto out_release_1; + set_bit(SOCK_USER_SOCKET, &sock2->flags); err = sock1->ops->socketpair(sock1, sock2); if (err < 0) diff -NurpP --minimal linux-source-2.6.22.orig/net/sunrpc/auth.c linux-source-2.6.22/net/sunrpc/auth.c --- linux-source-2.6.22.orig/net/sunrpc/auth.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/net/sunrpc/auth.c 2007-10-24 16:14:06.791620780 +0200 @@ -13,6 +13,7 @@ #include #include #include +#include #ifdef RPC_DEBUG # define RPCDBG_FACILITY RPCDBG_AUTH @@ -263,6 +264,7 @@ rpcauth_lookupcred(struct rpc_auth *auth struct auth_cred acred = { .uid = current->fsuid, .gid = current->fsgid, + .tag = dx_current_tag(), .group_info = current->group_info, }; struct rpc_cred *ret; @@ -282,6 +284,7 @@ rpcauth_bindcred(struct rpc_task *task) struct auth_cred acred = { .uid = current->fsuid, .gid = current->fsgid, + .tag = dx_current_tag(), .group_info = current->group_info, }; struct rpc_cred *ret; diff -NurpP --minimal linux-source-2.6.22.orig/net/sunrpc/auth_unix.c linux-source-2.6.22/net/sunrpc/auth_unix.c --- linux-source-2.6.22.orig/net/sunrpc/auth_unix.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/net/sunrpc/auth_unix.c 2007-10-24 16:14:06.791620780 +0200 @@ -11,12 +11,14 @@ #include #include #include +#include #define NFS_NGROUPS 16 struct unx_cred { struct rpc_cred uc_base; gid_t uc_gid; + tag_t uc_tag; gid_t uc_gids[NFS_NGROUPS]; }; #define uc_uid uc_base.cr_uid @@ -79,6 +81,7 @@ unx_create_cred(struct rpc_auth *auth, s if (flags & RPCAUTH_LOOKUP_ROOTCREDS) { cred->uc_uid = 0; cred->uc_gid = 0; + cred->uc_tag = dx_current_tag(); cred->uc_gids[0] = NOGROUP; } else { int groups = acred->group_info->ngroups; @@ -87,6 +90,7 @@ unx_create_cred(struct rpc_auth *auth, s cred->uc_uid = acred->uid; cred->uc_gid = acred->gid; + cred->uc_tag = acred->tag; for (i = 0; i < groups; i++) cred->uc_gids[i] = GROUP_AT(acred->group_info, i); if (i < NFS_NGROUPS) @@ -118,7 +122,8 @@ unx_match(struct auth_cred *acred, struc int groups; if (cred->uc_uid != acred->uid - || cred->uc_gid != acred->gid) + || cred->uc_gid != acred->gid + || cred->uc_tag != acred->tag) return 0; groups = acred->group_info->ngroups; @@ -144,7 +149,7 @@ unx_marshal(struct rpc_task *task, __be3 struct rpc_clnt *clnt = task->tk_client; struct unx_cred *cred = (struct unx_cred *) task->tk_msg.rpc_cred; __be32 *base, *hold; - int i; + int i, tag; *p++ = htonl(RPC_AUTH_UNIX); base = p++; @@ -154,9 +159,12 @@ unx_marshal(struct rpc_task *task, __be3 * Copy the UTS nodename captured when the client was created. */ p = xdr_encode_array(p, clnt->cl_nodename, clnt->cl_nodelen); + tag = task->tk_client->cl_tag; - *p++ = htonl((u32) cred->uc_uid); - *p++ = htonl((u32) cred->uc_gid); + *p++ = htonl((u32) TAGINO_UID(tag, + cred->uc_uid, cred->uc_tag)); + *p++ = htonl((u32) TAGINO_GID(tag, + cred->uc_gid, cred->uc_tag)); hold = p++; for (i = 0; i < 16 && cred->uc_gids[i] != (gid_t) NOGROUP; i++) *p++ = htonl((u32) cred->uc_gids[i]); diff -NurpP --minimal linux-source-2.6.22.orig/net/sunrpc/clnt.c linux-source-2.6.22/net/sunrpc/clnt.c --- linux-source-2.6.22.orig/net/sunrpc/clnt.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/net/sunrpc/clnt.c 2007-10-24 16:14:06.791620780 +0200 @@ -30,6 +30,7 @@ #include #include #include +#include #include #include @@ -249,7 +250,9 @@ struct rpc_clnt *rpc_create(struct rpc_c clnt->cl_oneshot = 1; if (args->flags & RPC_CLNT_CREATE_DISCRTRY) clnt->cl_discrtry = 1; - + /* TODO: handle RPC_CLNT_CREATE_TAGGED + if (args->flags & RPC_CLNT_CREATE_TAGGED) + clnt->cl_tag = 1; */ return clnt; } EXPORT_SYMBOL_GPL(rpc_create); diff -NurpP --minimal linux-source-2.6.22.orig/net/unix/af_unix.c linux-source-2.6.22/net/unix/af_unix.c --- linux-source-2.6.22.orig/net/unix/af_unix.c 2007-10-15 00:33:08.000000000 +0200 +++ linux-source-2.6.22/net/unix/af_unix.c 2007-10-24 16:14:06.791620780 +0200 @@ -115,6 +115,8 @@ #include #include #include +#include +#include int sysctl_unix_max_dgram_qlen __read_mostly = 10; @@ -252,6 +254,8 @@ static struct sock *__unix_find_socket_b sk_for_each(s, node, &unix_socket_table[hash ^ type]) { struct unix_sock *u = unix_sk(s); + if (!nx_check(s->sk_nid, VS_WATCH_P | VS_IDENT)) + continue; if (u->addr->len == len && !memcmp(u->addr->name, sunname, len)) goto found; @@ -807,7 +811,7 @@ static int unix_bind(struct socket *sock */ mode = S_IFSOCK | (SOCK_INODE(sock)->i_mode & ~current->fs->umask); - err = vfs_mknod(nd.dentry->d_inode, dentry, nd.mnt, mode, 0); + err = vfs_mknod(nd.dentry->d_inode, dentry, nd.mnt, mode, 0, NULL); if (err) goto out_mknod_dput; mutex_unlock(&nd.dentry->d_inode->i_mutex); diff -NurpP --minimal linux-source-2.6.22.orig/net/x25/af_x25.c linux-source-2.6.22/net/x25/af_x25.c --- linux-source-2.6.22.orig/net/x25/af_x25.c 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/net/x25/af_x25.c 2007-10-24 16:14:06.791620780 +0200 @@ -500,7 +500,10 @@ static int x25_create(struct socket *soc x25 = x25_sk(sk); - sock_init_data(sock, sk); + sk->sk_socket = sock; + sk->sk_type = sock->type; + sk->sk_sleep = &sock->wait; + sock->sk = sk; x25_init_timers(sk); diff -NurpP --minimal linux-source-2.6.22.orig/security/commoncap.c linux-source-2.6.22/security/commoncap.c --- linux-source-2.6.22.orig/security/commoncap.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/security/commoncap.c 2007-10-24 16:14:06.791620780 +0200 @@ -22,10 +22,11 @@ #include #include #include +#include int cap_netlink_send(struct sock *sk, struct sk_buff *skb) { - NETLINK_CB(skb).eff_cap = current->cap_effective; + cap_t(NETLINK_CB(skb).eff_cap) = vx_mbcap(cap_effective); return 0; } @@ -43,7 +44,7 @@ EXPORT_SYMBOL(cap_netlink_recv); int cap_capable (struct task_struct *tsk, int cap) { /* Derived from include/linux/sched.h:capable. */ - if (cap_raised(tsk->cap_effective, cap)) + if (vx_cap_raised(tsk->vx_info, tsk->cap_effective, cap)) return 0; return -EPERM; } @@ -141,7 +142,8 @@ void cap_bprm_apply_creds (struct linux_ /* Derived from fs/exec.c:compute_creds. */ kernel_cap_t new_permitted, working; - new_permitted = cap_intersect (bprm->cap_permitted, cap_bset); + new_permitted = cap_intersect (bprm->cap_permitted, + vx_current_cap_bset()); working = cap_intersect (bprm->cap_inheritable, current->cap_inheritable); new_permitted = cap_combine (new_permitted, working); @@ -311,7 +313,8 @@ void cap_task_reparent_to_init (struct t int cap_syslog (int type) { - if ((type != 3 && type != 10) && !capable(CAP_SYS_ADMIN)) + if ((type != 3 && type != 10) && + !vx_capable(CAP_SYS_ADMIN, VXC_SYSLOG)) return -EPERM; return 0; } diff -NurpP --minimal linux-source-2.6.22.orig/security/dummy.c linux-source-2.6.22/security/dummy.c --- linux-source-2.6.22.orig/security/dummy.c 2007-10-15 00:33:09.000000000 +0200 +++ linux-source-2.6.22/security/dummy.c 2007-10-24 16:14:06.791620780 +0200 @@ -28,6 +28,7 @@ #include #include #include +#include static int dummy_ptrace (struct task_struct *parent, struct task_struct *child) { @@ -692,7 +693,7 @@ static int dummy_sem_semop (struct sem_a static int dummy_netlink_send (struct sock *sk, struct sk_buff *skb) { - NETLINK_CB(skb).eff_cap = current->cap_effective; + cap_t(NETLINK_CB(skb).eff_cap) = vx_mbcap(cap_effective); return 0; } diff -NurpP --minimal linux-source-2.6.22.orig/security/Kconfig linux-source-2.6.22/security/Kconfig --- linux-source-2.6.22.orig/security/Kconfig 2007-07-09 01:32:17.000000000 +0200 +++ linux-source-2.6.22/security/Kconfig 2007-10-24 16:14:06.791620780 +0200 @@ -6,6 +6,7 @@ menu "Security options" config KEYS bool "Enable access key retention support" + depends on !VSERVER_SECURITY help This option provides support for retaining authentication tokens and access keys in the kernel.