From e59494f441c834ca7aaa0e6fa6678ddbd3d72743 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 16 Jul 2008 00:13:45 -0400 Subject: ftrace: fix 4d3702b6 (post-v2.6.26): WARNING: at kernel/lockdep.c:2731 check_flags (ftrace) On Wed, 16 Jul 2008, Vegard Nossum wrote: > When booting 4d3702b6, I got this huge thing: > > Testing tracer wakeup: <4>------------[ cut here ]------------ > WARNING: at kernel/lockdep.c:2731 check_flags+0x123/0x160() > Modules linked in: > Pid: 1, comm: swapper Not tainted 2.6.26-crashing-02127-g4d3702b6 #30 > [] warn_on_slowpath+0x59/0xb0 > [] ? ftrace_call+0x5/0x8 > [] ? native_read_tsc+0x0/0x20 > [] ? sub_preempt_count+0x12/0xf0 > [] ? trace_hardirqs_off+0xb/0x10 > [] ? __lock_acquire+0x2cc/0x1120 > [] ? trace_hardirqs_off+0xb/0x10 > [] ? mcount_call+0x5/0xa > [] check_flags+0x123/0x160 > [] lock_acquire+0x51/0xd0 > [] ? ftrace_call+0x5/0x8 > [] _spin_lock_irqsave+0x5f/0xa0 > [] ? ftrace_record_ip+0xf5/0x220 > [] ? debug_locks_off+0x3/0x50 > [] ftrace_record_ip+0xf5/0x220 > [] mcount_call+0x5/0xa > [] ? debug_locks_off+0x8/0x50 > [] check_flags+0xf7/0x160 > [] lock_acquire+0x51/0xd0 > [] ? ftrace_call+0x5/0x8 > [] _spin_lock_irqsave+0x5f/0xa0 > [] ? wakeup_tracer_call+0x6d/0xf0 > [] ? _local_bh_enable+0x62/0xb0 > [] ? sub_preempt_count+0xd/0xf0 > [] wakeup_tracer_call+0x6d/0xf0 > [] ? __do_softirq+0xf4/0x110 > [] ? wakeup_tracer_call+0x91/0xf0 > [] ftrace_call+0x5/0x8 > [] ? __do_softirq+0xf4/0x110 > [] ? sub_preempt_count+0x12/0xf0 > [] _local_bh_enable+0x62/0xb0 > [] __do_softirq+0xf4/0x110 > [] do_softirq+0xad/0xb0 > [] irq_exit+0xa5/0xb0 > [] smp_apic_timer_interrupt+0x66/0xa0 > [] ? trace_hardirqs_off_thunk+0xc/0x10 > [] apic_timer_interrupt+0x2d/0x34 > [] ? find_usage_backwards+0xb/0xf0 > [] ? _spin_unlock_irqrestore+0x69/0x80 > [] tg_shares_up+0x132/0x1d0 > [] walk_tg_tree+0x62/0xa0 > [] ? tg_shares_up+0x0/0x1d0 > [] ? tg_nop+0x0/0x10 > [] update_shares+0x5d/0x80 > [] try_to_wake_up+0x6f/0x280 > [] ? __ftrace_modify_code+0x0/0xc0 > [] ? __ftrace_modify_code+0x0/0xc0 > [] wake_up_process+0x14/0x20 > [] kthread_create+0x66/0xb0 > [] ? do_stop+0x0/0x200 > [] ? __stop_machine_run+0x30/0xb0 > [] __stop_machine_run+0x50/0xb0 > [] ? do_stop+0x0/0x200 > [] ? __ftrace_modify_code+0x0/0xc0 > [] ? mutex_unlock+0xd/0x10 > [] stop_machine_run+0x2c/0x60 > [] unregister_ftrace_function+0x103/0x180 > [] stop_wakeup_tracer+0x17/0x60 > [] wakeup_tracer_ctrl_update+0xf/0x30 > [] trace_selftest_startup_wakeup+0xb5/0x130 > [] ? trace_wakeup_test_thread+0x0/0x70 > [] register_tracer+0x135/0x1b0 > [] init_wakeup_tracer+0xd/0xf > [] kernel_init+0x1a9/0x2ce > [] ? _spin_unlock_irq+0x3b/0x60 > [] ? trace_hardirqs_on_thunk+0xc/0x10 > [] ? init_wakeup_tracer+0x0/0xf > [] ? trace_hardirqs_on_caller+0x126/0x180 > [] ? trace_hardirqs_on_thunk+0xc/0x10 > [] ? restore_nocheck_notrace+0x0/0xe > [] ? kernel_init+0x0/0x2ce > [] ? kernel_init+0x0/0x2ce > [] kernel_thread_helper+0x7/0x10 > ======================= > ---[ end trace a7919e7f17c0a725 ]--- > irq event stamp: 579530 > hardirqs last enabled at (579528): [] trace_hardirqs_on+0xb/0x10 > hardirqs last disabled at (579529): [] trace_hardirqs_off+0xb/0x10 > softirqs last enabled at (579530): [] __do_softirq+0xf4/0x110 > softirqs last disabled at (579517): [] do_softirq+0xad/0xb0 > irq event stamp: 579530 > hardirqs last enabled at (579528): [] trace_hardirqs_on+0xb/0x10 > hardirqs last disabled at (579529): [] trace_hardirqs_off+0xb/0x10 > softirqs last enabled at (579530): [] __do_softirq+0xf4/0x110 > softirqs last disabled at (579517): [] do_softirq+0xad/0xb0 > PASSED > > Incidentally, the kernel also hung while I was typing in this report. Things get weird between lockdep and ftrace because ftrace can be called within lockdep internal code (via the mcount pointer) and lockdep can be called with ftrace (via spin_locks). Signed-off-by: Steven Rostedt Tested-by: Vegard Nossum Signed-off-by: Ingo Molnar --- kernel/trace/trace_sched_wakeup.c | 27 ++++++++++++++++----------- 1 file changed, 16 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_sched_wakeup.c b/kernel/trace/trace_sched_wakeup.c index 3c8d61df4474..e303ccb62cdf 100644 --- a/kernel/trace/trace_sched_wakeup.c +++ b/kernel/trace/trace_sched_wakeup.c @@ -26,7 +26,8 @@ static struct task_struct *wakeup_task; static int wakeup_cpu; static unsigned wakeup_prio = -1; -static DEFINE_SPINLOCK(wakeup_lock); +static raw_spinlock_t wakeup_lock = + (raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED; static void __wakeup_reset(struct trace_array *tr); @@ -56,7 +57,8 @@ wakeup_tracer_call(unsigned long ip, unsigned long parent_ip) if (unlikely(disabled != 1)) goto out; - spin_lock_irqsave(&wakeup_lock, flags); + local_irq_save(flags); + __raw_spin_lock(&wakeup_lock); if (unlikely(!wakeup_task)) goto unlock; @@ -71,7 +73,8 @@ wakeup_tracer_call(unsigned long ip, unsigned long parent_ip) trace_function(tr, data, ip, parent_ip, flags); unlock: - spin_unlock_irqrestore(&wakeup_lock, flags); + __raw_spin_unlock(&wakeup_lock); + local_irq_restore(flags); out: atomic_dec(&data->disabled); @@ -145,7 +148,8 @@ wakeup_sched_switch(void *private, void *rq, struct task_struct *prev, if (likely(disabled != 1)) goto out; - spin_lock_irqsave(&wakeup_lock, flags); + local_irq_save(flags); + __raw_spin_lock(&wakeup_lock); /* We could race with grabbing wakeup_lock */ if (unlikely(!tracer_enabled || next != wakeup_task)) @@ -174,7 +178,8 @@ wakeup_sched_switch(void *private, void *rq, struct task_struct *prev, out_unlock: __wakeup_reset(tr); - spin_unlock_irqrestore(&wakeup_lock, flags); + __raw_spin_unlock(&wakeup_lock); + local_irq_restore(flags); out: atomic_dec(&tr->data[cpu]->disabled); } @@ -209,8 +214,6 @@ static void __wakeup_reset(struct trace_array *tr) struct trace_array_cpu *data; int cpu; - assert_spin_locked(&wakeup_lock); - for_each_possible_cpu(cpu) { data = tr->data[cpu]; tracing_reset(data); @@ -229,9 +232,11 @@ static void wakeup_reset(struct trace_array *tr) { unsigned long flags; - spin_lock_irqsave(&wakeup_lock, flags); + local_irq_save(flags); + __raw_spin_lock(&wakeup_lock); __wakeup_reset(tr); - spin_unlock_irqrestore(&wakeup_lock, flags); + __raw_spin_unlock(&wakeup_lock); + local_irq_restore(flags); } static void @@ -252,7 +257,7 @@ wakeup_check_start(struct trace_array *tr, struct task_struct *p, goto out; /* interrupts should be off from try_to_wake_up */ - spin_lock(&wakeup_lock); + __raw_spin_lock(&wakeup_lock); /* check for races. */ if (!tracer_enabled || p->prio >= wakeup_prio) @@ -274,7 +279,7 @@ wakeup_check_start(struct trace_array *tr, struct task_struct *p, CALLER_ADDR1, CALLER_ADDR2, flags); out_locked: - spin_unlock(&wakeup_lock); + __raw_spin_unlock(&wakeup_lock); out: atomic_dec(&tr->data[cpu]->disabled); } -- cgit v1.2.3 From b8f8c3cf0a4ac0632ec3f0e15e9dc0c29de917af Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 18 Jul 2008 17:27:28 +0200 Subject: nohz: prevent tick stop outside of the idle loop Jack Ren and Eric Miao tracked down the following long standing problem in the NOHZ code: scheduler switch to idle task enable interrupts Window starts here ----> interrupt happens (does not set NEED_RESCHED) irq_exit() stops the tick ----> interrupt happens (does set NEED_RESCHED) return from schedule() cpu_idle(): preempt_disable(); Window ends here The interrupts can happen at any point inside the race window. The first interrupt stops the tick, the second one causes the scheduler to rerun and switch away from idle again and we end up with the tick disabled. The fact that it needs two interrupts where the first one does not set NEED_RESCHED and the second one does made the bug obscure and extremly hard to reproduce and analyse. Kudos to Jack and Eric. Solution: Limit the NOHZ functionality to the idle loop to make sure that we can not run into such a situation ever again. cpu_idle() { preempt_disable(); while(1) { tick_nohz_stop_sched_tick(1); <- tell NOHZ code that we are in the idle loop while (!need_resched()) halt(); tick_nohz_restart_sched_tick(); <- disables NOHZ mode preempt_enable_no_resched(); schedule(); preempt_disable(); } } In hindsight we should have done this forever, but ... /me grabs a large brown paperbag. Debugged-by: Jack Ren , Debugged-by: eric miao Signed-off-by: Thomas Gleixner --- kernel/softirq.c | 2 +- kernel/time/tick-sched.c | 12 ++++++++++-- 2 files changed, 11 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/softirq.c b/kernel/softirq.c index 36e061740047..05f248039d77 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -312,7 +312,7 @@ void irq_exit(void) #ifdef CONFIG_NO_HZ /* Make sure that timer wheel updates are propagated */ if (!in_interrupt() && idle_cpu(smp_processor_id()) && !need_resched()) - tick_nohz_stop_sched_tick(); + tick_nohz_stop_sched_tick(0); rcu_irq_exit(); #endif preempt_enable_no_resched(); diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index 86baa4f0dfe4..ee962d11107b 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -195,7 +195,7 @@ u64 get_cpu_idle_time_us(int cpu, u64 *last_update_time) * Called either from the idle loop or from irq_exit() when an idle period was * just interrupted by an interrupt which did not cause a reschedule. */ -void tick_nohz_stop_sched_tick(void) +void tick_nohz_stop_sched_tick(int inidle) { unsigned long seq, last_jiffies, next_jiffies, delta_jiffies, flags; struct tick_sched *ts; @@ -224,6 +224,11 @@ void tick_nohz_stop_sched_tick(void) if (unlikely(ts->nohz_mode == NOHZ_MODE_INACTIVE)) goto end; + if (!inidle && !ts->inidle) + goto end; + + ts->inidle = 1; + if (need_resched()) goto end; @@ -372,11 +377,14 @@ void tick_nohz_restart_sched_tick(void) local_irq_disable(); tick_nohz_stop_idle(cpu); - if (!ts->tick_stopped) { + if (!ts->inidle || !ts->tick_stopped) { + ts->inidle = 0; local_irq_enable(); return; } + ts->inidle = 0; + rcu_exit_nohz(); /* Update jiffies first */ -- cgit v1.2.3 From 8df185a95c9b84fc0c3c02224e64fdc5b83bae34 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Tue, 8 Jul 2008 15:55:48 -0700 Subject: kthread: reduce stack pressure in create_kthread and kthreadd * Replace: set_cpus_allowed(..., CPU_MASK_ALL) with: set_cpus_allowed_ptr(..., CPU_MASK_ALL_PTR) to remove excessive stack requirements when NR_CPUS=4096. Signed-off-by: Mike Travis Cc: Andrew Morton Signed-off-by: Ingo Molnar --- kernel/kthread.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/kthread.c b/kernel/kthread.c index ac3fb7326641..6111c27491b1 100644 --- a/kernel/kthread.c +++ b/kernel/kthread.c @@ -106,7 +106,7 @@ static void create_kthread(struct kthread_create_info *create) */ sched_setscheduler(create->result, SCHED_NORMAL, ¶m); set_user_nice(create->result, KTHREAD_NICE_LEVEL); - set_cpus_allowed(create->result, CPU_MASK_ALL); + set_cpus_allowed_ptr(create->result, CPU_MASK_ALL_PTR); } complete(&create->done); } @@ -233,7 +233,7 @@ int kthreadd(void *unused) set_task_comm(tsk, "kthreadd"); ignore_signals(tsk); set_user_nice(tsk, KTHREAD_NICE_LEVEL); - set_cpus_allowed(tsk, CPU_MASK_ALL); + set_cpus_allowed_ptr(tsk, CPU_MASK_ALL_PTR); current->flags |= PF_NOFREEZE | PF_FREEZER_NOSIG; -- cgit v1.2.3 From 1e01cb0c6ff7e9ddb6547551794c6aa82785a7cb Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 15 Jul 2008 09:53:37 -0400 Subject: ftrace: only trace preempt off with preempt tracer When PREEMPT_TRACER and IRQSOFF_TRACER are both configured and irqsoff tracer is running, the preempt_off sections might also be traced. Thanks to Andrew Morton for pointing out my mistake of spin_lock disabling interrupts while he was reviewing ftrace.txt. Seems that my example I used actually hit this bug. Signed-off-by: Steven Rostedt Cc: Linus Torvalds Cc: Andrew Morton Cc: Peter Zijlstra Signed-off-by: Ingo Molnar --- kernel/trace/trace_irqsoff.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_irqsoff.c b/kernel/trace/trace_irqsoff.c index 421d6fe3650e..b1e4a89b08eb 100644 --- a/kernel/trace/trace_irqsoff.c +++ b/kernel/trace/trace_irqsoff.c @@ -337,12 +337,14 @@ EXPORT_SYMBOL(trace_hardirqs_off_caller); #ifdef CONFIG_PREEMPT_TRACER void trace_preempt_on(unsigned long a0, unsigned long a1) { - stop_critical_timing(a0, a1); + if (preempt_trace()) + stop_critical_timing(a0, a1); } void trace_preempt_off(unsigned long a0, unsigned long a1) { - start_critical_timing(a0, a1); + if (preempt_trace()) + start_critical_timing(a0, a1); } #endif /* CONFIG_PREEMPT_TRACER */ -- cgit v1.2.3 From 422037bafde8083acc3c539ceba3dfc60a04110c Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Wed, 23 Jul 2008 11:16:38 +0200 Subject: sched: fix hrtick & generic-ipi dependency Andrew Morton reported this s390 allmodconfig build failure: kernel/built-in.o: In function `hrtick_start_fair': sched.c:(.text+0x69c6): undefined reference to `__smp_call_function_single' the reason is that s390 is not a generic-ipi SMP platform yet, while the hrtick code relies on it. Fix the dependency. Signed-off-by: Ingo Molnar --- kernel/Kconfig.hz | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/Kconfig.hz b/kernel/Kconfig.hz index 2a202a846757..382dd5a8b2d7 100644 --- a/kernel/Kconfig.hz +++ b/kernel/Kconfig.hz @@ -55,4 +55,4 @@ config HZ default 1000 if HZ_1000 config SCHED_HRTICK - def_bool HIGH_RES_TIMERS + def_bool HIGH_RES_TIMERS && USE_GENERIC_SMP_HELPERS -- cgit v1.2.3 From 1986b0cb1671ea39178b4e2b00461109728fc935 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Thu, 24 Jul 2008 08:10:02 +0200 Subject: ftrace: remove latency-tracer leftover remove the :vim=ft=help tag from trace files. I used them years ago to syntax-highlight traces and forgot about this hack. Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 3 --- 1 file changed, 3 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 868e121c8e38..fc20e09a6cb1 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1203,9 +1203,6 @@ static void *s_next(struct seq_file *m, void *v, loff_t *pos) iter->pos = *pos; - if (last_ent && !ent) - seq_puts(m, "\n\nvim:ft=help\n"); - return ent; } -- cgit v1.2.3 From 58838cf3ca3337d76141c33d6c68376490263468 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 24 Jul 2008 12:43:13 +0200 Subject: sched: clean up compiler warning Reported-by: Daniel Walker Signed-off-by: Peter Zijlstra Signed-off-by: Ingo Molnar --- kernel/sched_rt.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c index 147004c651c0..93ac8ee08271 100644 --- a/kernel/sched_rt.c +++ b/kernel/sched_rt.c @@ -253,7 +253,7 @@ static int do_balance_runtime(struct rt_rq *rt_rq) diff = iter->rt_runtime - iter->rt_time; if (diff > 0) { - do_div(diff, weight); + diff = div_u64((u64)diff, weight); if (rt_rq->rt_runtime + diff > rt_period) diff = rt_period - rt_rq->rt_runtime; iter->rt_runtime -= diff; -- cgit v1.2.3 From c748e1340e0de3fa7fed86f8bdf499be9242afff Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Wed, 23 Jul 2008 21:27:03 -0700 Subject: mm/vmstat.c: proper externs This patch adds proper extern declarations for five variables in include/linux/vmstat.h Signed-off-by: Adrian Bunk Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sysctl.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 2a7b9d88706b..1f7b3b76a166 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -43,6 +43,7 @@ #include #include #include +#include #include #include #include @@ -80,7 +81,6 @@ extern int sysctl_drop_caches; extern int percpu_pagelist_fraction; extern int compat_log; extern int maps_protect; -extern int sysctl_stat_interval; extern int latencytop_enabled; extern int sysctl_nr_open_min, sysctl_nr_open_max; #ifdef CONFIG_RCU_TORTURE_TEST -- cgit v1.2.3 From a1e78772d72b2616ed20e54896e68e0e7044854e Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Wed, 23 Jul 2008 21:27:23 -0700 Subject: hugetlb: reserve huge pages for reliable MAP_PRIVATE hugetlbfs mappings until fork() This patch reserves huge pages at mmap() time for MAP_PRIVATE mappings in a similar manner to the reservations taken for MAP_SHARED mappings. The reserve count is accounted both globally and on a per-VMA basis for private mappings. This guarantees that a process that successfully calls mmap() will successfully fault all pages in the future unless fork() is called. The characteristics of private mappings of hugetlbfs files behaviour after this patch are; 1. The process calling mmap() is guaranteed to succeed all future faults until it forks(). 2. On fork(), the parent may die due to SIGKILL on writes to the private mapping if enough pages are not available for the COW. For reasonably reliable behaviour in the face of a small huge page pool, children of hugepage-aware processes should not reference the mappings; such as might occur when fork()ing to exec(). 3. On fork(), the child VMAs inherit no reserves. Reads on pages already faulted by the parent will succeed. Successful writes will depend on enough huge pages being free in the pool. 4. Quotas of the hugetlbfs mount are checked at reserve time for the mapper and at fault time otherwise. Before this patch, all reads or writes in the child potentially needs page allocations that can later lead to the death of the parent. This applies to reads and writes of uninstantiated pages as well as COW. After the patch it is only a write to an instantiated page that causes problems. Signed-off-by: Mel Gorman Acked-by: Adam Litke Cc: Andy Whitcroft Cc: William Lee Irwin III Cc: Hugh Dickins Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/fork.c | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index adefc1131f27..552c8d8e77ad 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -33,6 +33,7 @@ #include #include #include +#include #include #include #include @@ -306,6 +307,14 @@ static int dup_mmap(struct mm_struct *mm, struct mm_struct *oldmm) spin_unlock(&file->f_mapping->i_mmap_lock); } + /* + * Clear hugetlb-related page reserves for children. This only + * affects MAP_PRIVATE mappings. Faults generated by the child + * are not guaranteed to succeed, even if read-only + */ + if (is_vm_hugetlb_page(tmp)) + reset_vma_resv_huge_pages(tmp); + /* * Link in the new vma and copy the page table entries. */ -- cgit v1.2.3 From e5ff215941d59f8ae6bf58f6428dc5c26745a612 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Wed, 23 Jul 2008 21:27:42 -0700 Subject: hugetlb: multiple hstates for multiple page sizes Add basic support for more than one hstate in hugetlbfs. This is the key to supporting multiple hugetlbfs page sizes at once. - Rather than a single hstate, we now have an array, with an iterator - default_hstate continues to be the struct hstate which we use by default - Add functions for architectures to register new hstates [akpm@linux-foundation.org: coding-style fixes] Acked-by: Adam Litke Acked-by: Nishanth Aravamudan Signed-off-by: Andi Kleen Signed-off-by: Nick Piggin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sysctl.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 1f7b3b76a166..1a8299d1fe59 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -959,7 +959,7 @@ static struct ctl_table vm_table[] = { #ifdef CONFIG_HUGETLB_PAGE { .procname = "nr_hugepages", - .data = &max_huge_pages, + .data = NULL, .maxlen = sizeof(unsigned long), .mode = 0644, .proc_handler = &hugetlb_sysctl_handler, @@ -985,10 +985,12 @@ static struct ctl_table vm_table[] = { { .ctl_name = CTL_UNNUMBERED, .procname = "nr_overcommit_hugepages", - .data = &sysctl_overcommit_huge_pages, - .maxlen = sizeof(sysctl_overcommit_huge_pages), + .data = NULL, + .maxlen = sizeof(unsigned long), .mode = 0644, .proc_handler = &hugetlb_overcommit_handler, + .extra1 = (void *)&hugetlb_zero, + .extra2 = (void *)&hugetlb_infinity, }, #endif { -- cgit v1.2.3 From ab763c7112ce0e2559c73f921617c81dc7287ca6 Mon Sep 17 00:00:00 2001 From: "Andrew G. Morgan" Date: Wed, 23 Jul 2008 21:28:25 -0700 Subject: security: filesystem capabilities refactor kernel code To date, we've tried hard to confine filesystem support for capabilities to the security modules. This has left a lot of the code in kernel/capability.c in a state where it looks like it supports something that filesystem support for capabilities actually suppresses when the LSM security/commmoncap.c code runs. What is left is a lot of code that uses sub-optimal locking in the main kernel With this change we refactor the main kernel code and make it explicit which locks are needed and that the only remaining kernel races in this area are associated with non-filesystem capability code. Signed-off-by: Andrew G. Morgan Acked-by: Serge Hallyn Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/capability.c | 338 ++++++++++++++++++++++++++++++++++------------------ 1 file changed, 221 insertions(+), 117 deletions(-) (limited to 'kernel') diff --git a/kernel/capability.c b/kernel/capability.c index 901e0fdc3fff..0101e847603e 100644 --- a/kernel/capability.c +++ b/kernel/capability.c @@ -115,11 +115,208 @@ static int cap_validate_magic(cap_user_header_t header, unsigned *tocopy) return 0; } +#ifndef CONFIG_SECURITY_FILE_CAPABILITIES + +/* + * Without filesystem capability support, we nominally support one process + * setting the capabilities of another + */ +static inline int cap_get_target_pid(pid_t pid, kernel_cap_t *pEp, + kernel_cap_t *pIp, kernel_cap_t *pPp) +{ + struct task_struct *target; + int ret; + + spin_lock(&task_capability_lock); + read_lock(&tasklist_lock); + + if (pid && pid != task_pid_vnr(current)) { + target = find_task_by_vpid(pid); + if (!target) { + ret = -ESRCH; + goto out; + } + } else + target = current; + + ret = security_capget(target, pEp, pIp, pPp); + +out: + read_unlock(&tasklist_lock); + spin_unlock(&task_capability_lock); + + return ret; +} + +/* + * cap_set_pg - set capabilities for all processes in a given process + * group. We call this holding task_capability_lock and tasklist_lock. + */ +static inline int cap_set_pg(int pgrp_nr, kernel_cap_t *effective, + kernel_cap_t *inheritable, + kernel_cap_t *permitted) +{ + struct task_struct *g, *target; + int ret = -EPERM; + int found = 0; + struct pid *pgrp; + + spin_lock(&task_capability_lock); + read_lock(&tasklist_lock); + + pgrp = find_vpid(pgrp_nr); + do_each_pid_task(pgrp, PIDTYPE_PGID, g) { + target = g; + while_each_thread(g, target) { + if (!security_capset_check(target, effective, + inheritable, permitted)) { + security_capset_set(target, effective, + inheritable, permitted); + ret = 0; + } + found = 1; + } + } while_each_pid_task(pgrp, PIDTYPE_PGID, g); + + read_unlock(&tasklist_lock); + spin_unlock(&task_capability_lock); + + if (!found) + ret = 0; + return ret; +} + /* - * For sys_getproccap() and sys_setproccap(), any of the three - * capability set pointers may be NULL -- indicating that that set is - * uninteresting and/or not to be changed. + * cap_set_all - set capabilities for all processes other than init + * and self. We call this holding task_capability_lock and tasklist_lock. */ +static inline int cap_set_all(kernel_cap_t *effective, + kernel_cap_t *inheritable, + kernel_cap_t *permitted) +{ + struct task_struct *g, *target; + int ret = -EPERM; + int found = 0; + + spin_lock(&task_capability_lock); + read_lock(&tasklist_lock); + + do_each_thread(g, target) { + if (target == current + || is_container_init(target->group_leader)) + continue; + found = 1; + if (security_capset_check(target, effective, inheritable, + permitted)) + continue; + ret = 0; + security_capset_set(target, effective, inheritable, permitted); + } while_each_thread(g, target); + + read_unlock(&tasklist_lock); + spin_unlock(&task_capability_lock); + + if (!found) + ret = 0; + + return ret; +} + +/* + * Given the target pid does not refer to the current process we + * need more elaborate support... (This support is not present when + * filesystem capabilities are configured.) + */ +static inline int do_sys_capset_other_tasks(pid_t pid, kernel_cap_t *effective, + kernel_cap_t *inheritable, + kernel_cap_t *permitted) +{ + struct task_struct *target; + int ret; + + if (!capable(CAP_SETPCAP)) + return -EPERM; + + if (pid == -1) /* all procs other than current and init */ + return cap_set_all(effective, inheritable, permitted); + + else if (pid < 0) /* all procs in process group */ + return cap_set_pg(-pid, effective, inheritable, permitted); + + /* target != current */ + spin_lock(&task_capability_lock); + read_lock(&tasklist_lock); + + target = find_task_by_vpid(pid); + if (!target) + ret = -ESRCH; + else { + ret = security_capset_check(target, effective, inheritable, + permitted); + + /* having verified that the proposed changes are legal, + we now put them into effect. */ + if (!ret) + security_capset_set(target, effective, inheritable, + permitted); + } + + read_unlock(&tasklist_lock); + spin_unlock(&task_capability_lock); + + return ret; +} + +#else /* ie., def CONFIG_SECURITY_FILE_CAPABILITIES */ + +/* + * If we have configured with filesystem capability support, then the + * only thing that can change the capabilities of the current process + * is the current process. As such, we can't be in this code at the + * same time as we are in the process of setting capabilities in this + * process. The net result is that we can limit our use of locks to + * when we are reading the caps of another process. + */ +static inline int cap_get_target_pid(pid_t pid, kernel_cap_t *pEp, + kernel_cap_t *pIp, kernel_cap_t *pPp) +{ + int ret; + + if (pid && (pid != task_pid_vnr(current))) { + struct task_struct *target; + + spin_lock(&task_capability_lock); + read_lock(&tasklist_lock); + + target = find_task_by_vpid(pid); + if (!target) + ret = -ESRCH; + else + ret = security_capget(target, pEp, pIp, pPp); + + read_unlock(&tasklist_lock); + spin_unlock(&task_capability_lock); + } else + ret = security_capget(current, pEp, pIp, pPp); + + return ret; +} + +/* + * With filesystem capability support configured, the kernel does not + * permit the changing of capabilities in one process by another + * process. (CAP_SETPCAP has much less broad semantics when configured + * this way.) + */ +static inline int do_sys_capset_other_tasks(pid_t pid, + kernel_cap_t *effective, + kernel_cap_t *inheritable, + kernel_cap_t *permitted) +{ + return -EPERM; +} + +#endif /* ie., ndef CONFIG_SECURITY_FILE_CAPABILITIES */ /* * Atomically modify the effective capabilities returning the original @@ -155,7 +352,6 @@ asmlinkage long sys_capget(cap_user_header_t header, cap_user_data_t dataptr) { int ret = 0; pid_t pid; - struct task_struct *target; unsigned tocopy; kernel_cap_t pE, pI, pP; @@ -169,23 +365,7 @@ asmlinkage long sys_capget(cap_user_header_t header, cap_user_data_t dataptr) if (pid < 0) return -EINVAL; - spin_lock(&task_capability_lock); - read_lock(&tasklist_lock); - - if (pid && pid != task_pid_vnr(current)) { - target = find_task_by_vpid(pid); - if (!target) { - ret = -ESRCH; - goto out; - } - } else - target = current; - - ret = security_capget(target, &pE, &pI, &pP); - -out: - read_unlock(&tasklist_lock); - spin_unlock(&task_capability_lock); + ret = cap_get_target_pid(pid, &pE, &pI, &pP); if (!ret) { struct __user_cap_data_struct kdata[_KERNEL_CAPABILITY_U32S]; @@ -216,7 +396,6 @@ out: * before modification is attempted and the application * fails. */ - if (copy_to_user(dataptr, kdata, tocopy * sizeof(struct __user_cap_data_struct))) { return -EFAULT; @@ -226,70 +405,8 @@ out: return ret; } -/* - * cap_set_pg - set capabilities for all processes in a given process - * group. We call this holding task_capability_lock and tasklist_lock. - */ -static inline int cap_set_pg(int pgrp_nr, kernel_cap_t *effective, - kernel_cap_t *inheritable, - kernel_cap_t *permitted) -{ - struct task_struct *g, *target; - int ret = -EPERM; - int found = 0; - struct pid *pgrp; - - pgrp = find_vpid(pgrp_nr); - do_each_pid_task(pgrp, PIDTYPE_PGID, g) { - target = g; - while_each_thread(g, target) { - if (!security_capset_check(target, effective, - inheritable, - permitted)) { - security_capset_set(target, effective, - inheritable, - permitted); - ret = 0; - } - found = 1; - } - } while_each_pid_task(pgrp, PIDTYPE_PGID, g); - - if (!found) - ret = 0; - return ret; -} - -/* - * cap_set_all - set capabilities for all processes other than init - * and self. We call this holding task_capability_lock and tasklist_lock. - */ -static inline int cap_set_all(kernel_cap_t *effective, - kernel_cap_t *inheritable, - kernel_cap_t *permitted) -{ - struct task_struct *g, *target; - int ret = -EPERM; - int found = 0; - - do_each_thread(g, target) { - if (target == current || is_container_init(target->group_leader)) - continue; - found = 1; - if (security_capset_check(target, effective, inheritable, - permitted)) - continue; - ret = 0; - security_capset_set(target, effective, inheritable, permitted); - } while_each_thread(g, target); - - if (!found) - ret = 0; - return ret; -} - /** - * sys_capset - set capabilities for a process or a group of processes + * sys_capset - set capabilities for a process or (*) a group of processes * @header: pointer to struct that contains capability version and * target pid data * @data: pointer to struct that contains the effective, permitted, @@ -313,7 +430,6 @@ asmlinkage long sys_capset(cap_user_header_t header, const cap_user_data_t data) struct __user_cap_data_struct kdata[_KERNEL_CAPABILITY_U32S]; unsigned i, tocopy; kernel_cap_t inheritable, permitted, effective; - struct task_struct *target; int ret; pid_t pid; @@ -324,9 +440,6 @@ asmlinkage long sys_capset(cap_user_header_t header, const cap_user_data_t data) if (get_user(pid, &header->pid)) return -EFAULT; - if (pid && pid != task_pid_vnr(current) && !capable(CAP_SETPCAP)) - return -EPERM; - if (copy_from_user(&kdata, data, tocopy * sizeof(struct __user_cap_data_struct))) { return -EFAULT; @@ -344,40 +457,31 @@ asmlinkage long sys_capset(cap_user_header_t header, const cap_user_data_t data) i++; } - spin_lock(&task_capability_lock); - read_lock(&tasklist_lock); - - if (pid > 0 && pid != task_pid_vnr(current)) { - target = find_task_by_vpid(pid); - if (!target) { - ret = -ESRCH; - goto out; - } - } else - target = current; - - ret = 0; - - /* having verified that the proposed changes are legal, - we now put them into effect. */ - if (pid < 0) { - if (pid == -1) /* all procs other than current and init */ - ret = cap_set_all(&effective, &inheritable, &permitted); + if (pid && (pid != task_pid_vnr(current))) + ret = do_sys_capset_other_tasks(pid, &effective, &inheritable, + &permitted); + else { + /* + * This lock is required even when filesystem + * capability support is configured - it protects the + * sys_capget() call from returning incorrect data in + * the case that the targeted process is not the + * current one. + */ + spin_lock(&task_capability_lock); - else /* all procs in process group */ - ret = cap_set_pg(-pid, &effective, &inheritable, - &permitted); - } else { - ret = security_capset_check(target, &effective, &inheritable, + ret = security_capset_check(current, &effective, &inheritable, &permitted); + /* + * Having verified that the proposed changes are + * legal, we now put them into effect. + */ if (!ret) - security_capset_set(target, &effective, &inheritable, + security_capset_set(current, &effective, &inheritable, &permitted); + spin_unlock(&task_capability_lock); } -out: - read_unlock(&tasklist_lock); - spin_unlock(&task_capability_lock); return ret; } -- cgit v1.2.3 From 0d63081d418c73cc187c893069e0f24c4c6eecd3 Mon Sep 17 00:00:00 2001 From: Pavel Machek Date: Wed, 23 Jul 2008 21:28:32 -0700 Subject: swsusp: provide users with a hint about the no_console_suspend option Tell the user about the no_console_suspend option, so that we don't have to tell each bug reporter personally. [akpm@linux-foundation.org: clarify the text a little] Signed-off-by: Pavel Machek Cc: "Rafael J. Wysocki" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/printk.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/printk.c b/kernel/printk.c index 07ad9e7f7a66..3f7a2a94583b 100644 --- a/kernel/printk.c +++ b/kernel/printk.c @@ -933,7 +933,7 @@ void suspend_console(void) { if (!console_suspend_enabled) return; - printk("Suspending console(s)\n"); + printk("Suspending console(s) (use no_console_suspend to debug)\n"); acquire_console_sem(); console_suspended = 1; } -- cgit v1.2.3 From 77437fd4e61f87cc94d9314baa5cbf50e3ccdf54 Mon Sep 17 00:00:00 2001 From: David Brownell Date: Wed, 23 Jul 2008 21:28:33 -0700 Subject: pm: boot time suspend selftest Boot-time test for system suspend states (STR or standby). The generic RTC framework triggers wakeup alarms, which are used to exit those states. - Measures some aspects of suspend time ... this uses "jiffies" until someone converts it to use a timebase that works properly even while timer IRQs are disabled. - Triggered by a command line parameter. By default nothing even vaguely troublesome will happen, but "test_suspend=mem" will give you a brief STR test during system boot. (Or you may need to use "test_suspend=standby" instead, if your hardware needs that.) This isn't without problems. It fires early enough during boot that for example both PCMCIA and MMC stacks have misbehaved. The workaround in those cases was to boot without such media cards inserted. [matthltc@us.ibm.com: fix compile failure in boot time suspend selftest] Signed-off-by: David Brownell Cc: Ingo Molnar Cc: Pavel Machek Cc: "Rafael J. Wysocki" Signed-off-by: Matt Helsley Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/power/Kconfig | 11 +++ kernel/power/main.c | 194 ++++++++++++++++++++++++++++++++++++++++++++++++++- 2 files changed, 204 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/power/Kconfig b/kernel/power/Kconfig index 59dfdf1e1d20..dcd165f92a88 100644 --- a/kernel/power/Kconfig +++ b/kernel/power/Kconfig @@ -94,6 +94,17 @@ config SUSPEND powered and thus its contents are preserved, such as the suspend-to-RAM state (e.g. the ACPI S3 state). +config PM_TEST_SUSPEND + bool "Test suspend/resume and wakealarm during bootup" + depends on SUSPEND && PM_DEBUG && RTC_LIB=y + ---help--- + This option will let you suspend your machine during bootup, and + make it wake up a few seconds later using an RTC wakeup alarm. + Enable this with a kernel parameter like "test_suspend=mem". + + You probably want to have your system's RTC driver statically + linked, ensuring that it's available when this test runs. + config SUSPEND_FREEZER bool "Enable freezer for suspend to RAM/standby" \ if ARCH_WANTS_FREEZER_CONTROL || BROKEN diff --git a/kernel/power/main.c b/kernel/power/main.c index 3398f4651aa1..95bff23ecdaa 100644 --- a/kernel/power/main.c +++ b/kernel/power/main.c @@ -132,6 +132,61 @@ static inline int suspend_test(int level) { return 0; } #ifdef CONFIG_SUSPEND +#ifdef CONFIG_PM_TEST_SUSPEND + +/* + * We test the system suspend code by setting an RTC wakealarm a short + * time in the future, then suspending. Suspending the devices won't + * normally take long ... some systems only need a few milliseconds. + * + * The time it takes is system-specific though, so when we test this + * during system bootup we allow a LOT of time. + */ +#define TEST_SUSPEND_SECONDS 5 + +static unsigned long suspend_test_start_time; + +static void suspend_test_start(void) +{ + /* FIXME Use better timebase than "jiffies", ideally a clocksource. + * What we want is a hardware counter that will work correctly even + * during the irqs-are-off stages of the suspend/resume cycle... + */ + suspend_test_start_time = jiffies; +} + +static void suspend_test_finish(const char *label) +{ + long nj = jiffies - suspend_test_start_time; + unsigned msec; + + msec = jiffies_to_msecs(abs(nj)); + pr_info("PM: %s took %d.%03d seconds\n", label, + msec / 1000, msec % 1000); + + /* Warning on suspend means the RTC alarm period needs to be + * larger -- the system was sooo slooowwww to suspend that the + * alarm (should have) fired before the system went to sleep! + * + * Warning on either suspend or resume also means the system + * has some performance issues. The stack dump of a WARN_ON + * is more likely to get the right attention than a printk... + */ + WARN_ON(msec > (TEST_SUSPEND_SECONDS * 1000)); +} + +#else + +static void suspend_test_start(void) +{ +} + +static void suspend_test_finish(const char *label) +{ +} + +#endif + /* This is just an arbitrary number */ #define FREE_PAGE_NUMBER (100) @@ -266,12 +321,13 @@ int suspend_devices_and_enter(suspend_state_t state) goto Close; } suspend_console(); + suspend_test_start(); error = device_suspend(PMSG_SUSPEND); if (error) { printk(KERN_ERR "PM: Some devices failed to suspend\n"); goto Recover_platform; } - + suspend_test_finish("suspend devices"); if (suspend_test(TEST_DEVICES)) goto Recover_platform; @@ -293,7 +349,9 @@ int suspend_devices_and_enter(suspend_state_t state) if (suspend_ops->finish) suspend_ops->finish(); Resume_devices: + suspend_test_start(); device_resume(PMSG_RESUME); + suspend_test_finish("resume devices"); resume_console(); Close: if (suspend_ops->end) @@ -521,3 +579,137 @@ static int __init pm_init(void) } core_initcall(pm_init); + + +#ifdef CONFIG_PM_TEST_SUSPEND + +#include + +/* + * To test system suspend, we need a hands-off mechanism to resume the + * system. RTCs wake alarms are a common self-contained mechanism. + */ + +static void __init test_wakealarm(struct rtc_device *rtc, suspend_state_t state) +{ + static char err_readtime[] __initdata = + KERN_ERR "PM: can't read %s time, err %d\n"; + static char err_wakealarm [] __initdata = + KERN_ERR "PM: can't set %s wakealarm, err %d\n"; + static char err_suspend[] __initdata = + KERN_ERR "PM: suspend test failed, error %d\n"; + static char info_test[] __initdata = + KERN_INFO "PM: test RTC wakeup from '%s' suspend\n"; + + unsigned long now; + struct rtc_wkalrm alm; + int status; + + /* this may fail if the RTC hasn't been initialized */ + status = rtc_read_time(rtc, &alm.time); + if (status < 0) { + printk(err_readtime, rtc->dev.bus_id, status); + return; + } + rtc_tm_to_time(&alm.time, &now); + + memset(&alm, 0, sizeof alm); + rtc_time_to_tm(now + TEST_SUSPEND_SECONDS, &alm.time); + alm.enabled = true; + + status = rtc_set_alarm(rtc, &alm); + if (status < 0) { + printk(err_wakealarm, rtc->dev.bus_id, status); + return; + } + + if (state == PM_SUSPEND_MEM) { + printk(info_test, pm_states[state]); + status = pm_suspend(state); + if (status == -ENODEV) + state = PM_SUSPEND_STANDBY; + } + if (state == PM_SUSPEND_STANDBY) { + printk(info_test, pm_states[state]); + status = pm_suspend(state); + } + if (status < 0) + printk(err_suspend, status); +} + +static int __init has_wakealarm(struct device *dev, void *name_ptr) +{ + struct rtc_device *candidate = to_rtc_device(dev); + + if (!candidate->ops->set_alarm) + return 0; + if (!device_may_wakeup(candidate->dev.parent)) + return 0; + + *(char **)name_ptr = dev->bus_id; + return 1; +} + +/* + * Kernel options like "test_suspend=mem" force suspend/resume sanity tests + * at startup time. They're normally disabled, for faster boot and because + * we can't know which states really work on this particular system. + */ +static suspend_state_t test_state __initdata = PM_SUSPEND_ON; + +static char warn_bad_state[] __initdata = + KERN_WARNING "PM: can't test '%s' suspend state\n"; + +static int __init setup_test_suspend(char *value) +{ + unsigned i; + + /* "=mem" ==> "mem" */ + value++; + for (i = 0; i < PM_SUSPEND_MAX; i++) { + if (!pm_states[i]) + continue; + if (strcmp(pm_states[i], value) != 0) + continue; + test_state = (__force suspend_state_t) i; + return 0; + } + printk(warn_bad_state, value); + return 0; +} +__setup("test_suspend", setup_test_suspend); + +static int __init test_suspend(void) +{ + static char warn_no_rtc[] __initdata = + KERN_WARNING "PM: no wakealarm-capable RTC driver is ready\n"; + + char *pony = NULL; + struct rtc_device *rtc = NULL; + + /* PM is initialized by now; is that state testable? */ + if (test_state == PM_SUSPEND_ON) + goto done; + if (!valid_state(test_state)) { + printk(warn_bad_state, pm_states[test_state]); + goto done; + } + + /* RTCs have initialized by now too ... can we use one? */ + class_find_device(rtc_class, NULL, &pony, has_wakealarm); + if (pony) + rtc = rtc_class_open(pony); + if (!rtc) { + printk(warn_no_rtc); + goto done; + } + + /* go for it */ + test_wakealarm(rtc, test_state); + rtc_class_close(rtc); +done: + return 0; +} +late_initcall(test_suspend); + +#endif /* CONFIG_PM_TEST_SUSPEND */ -- cgit v1.2.3 From 0d83304c7e7bd3b05be90281b3a47841bc8f057a Mon Sep 17 00:00:00 2001 From: Akinobu Mita Date: Wed, 23 Jul 2008 21:28:38 -0700 Subject: pm: hibernation: simplify memory bitmap This patch simplifies the memory bitmap manipulations. - remove the member size in struct bm_block It is not necessary for struct bm_block to have the number of bit chunks that can be calculated by using end_pfn and start_pfn. - use find_next_bit() for memory_bm_next_pfn No need to invent the bitmap library only for the memory bitmap. Signed-off-by: Akinobu Mita Signed-off-by: Rafael J. Wysocki Acked-by: Pavel Machek Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/power/snapshot.c | 88 ++++++++++++------------------------------------- 1 file changed, 21 insertions(+), 67 deletions(-) (limited to 'kernel') diff --git a/kernel/power/snapshot.c b/kernel/power/snapshot.c index 5f91a07c4eac..5d2ab836e998 100644 --- a/kernel/power/snapshot.c +++ b/kernel/power/snapshot.c @@ -205,8 +205,7 @@ static void chain_free(struct chain_allocator *ca, int clear_page_nosave) * objects. The main list's elements are of type struct zone_bitmap * and each of them corresonds to one zone. For each zone bitmap * object there is a list of objects of type struct bm_block that - * represent each blocks of bit chunks in which information is - * stored. + * represent each blocks of bitmap in which information is stored. * * struct memory_bitmap contains a pointer to the main list of zone * bitmap objects, a struct bm_position used for browsing the bitmap, @@ -224,26 +223,27 @@ static void chain_free(struct chain_allocator *ca, int clear_page_nosave) * pfns that correspond to the start and end of the represented zone. * * struct bm_block contains a pointer to the memory page in which - * information is stored (in the form of a block of bit chunks - * of type unsigned long each). It also contains the pfns that - * correspond to the start and end of the represented memory area and - * the number of bit chunks in the block. + * information is stored (in the form of a block of bitmap) + * It also contains the pfns that correspond to the start and end of + * the represented memory area. */ #define BM_END_OF_MAP (~0UL) -#define BM_CHUNKS_PER_BLOCK (PAGE_SIZE / sizeof(long)) -#define BM_BITS_PER_CHUNK (sizeof(long) << 3) #define BM_BITS_PER_BLOCK (PAGE_SIZE << 3) struct bm_block { struct bm_block *next; /* next element of the list */ unsigned long start_pfn; /* pfn represented by the first bit */ unsigned long end_pfn; /* pfn represented by the last bit plus 1 */ - unsigned int size; /* number of bit chunks */ - unsigned long *data; /* chunks of bits representing pages */ + unsigned long *data; /* bitmap representing pages */ }; +static inline unsigned long bm_block_bits(struct bm_block *bb) +{ + return bb->end_pfn - bb->start_pfn; +} + struct zone_bitmap { struct zone_bitmap *next; /* next element of the list */ unsigned long start_pfn; /* minimal pfn in this zone */ @@ -257,7 +257,6 @@ struct zone_bitmap { struct bm_position { struct zone_bitmap *zone_bm; struct bm_block *block; - int chunk; int bit; }; @@ -272,12 +271,6 @@ struct memory_bitmap { /* Functions that operate on memory bitmaps */ -static inline void memory_bm_reset_chunk(struct memory_bitmap *bm) -{ - bm->cur.chunk = 0; - bm->cur.bit = -1; -} - static void memory_bm_position_reset(struct memory_bitmap *bm) { struct zone_bitmap *zone_bm; @@ -285,7 +278,7 @@ static void memory_bm_position_reset(struct memory_bitmap *bm) zone_bm = bm->zone_bm_list; bm->cur.zone_bm = zone_bm; bm->cur.block = zone_bm->bm_blocks; - memory_bm_reset_chunk(bm); + bm->cur.bit = 0; } static void memory_bm_free(struct memory_bitmap *bm, int clear_nosave_free); @@ -394,12 +387,10 @@ memory_bm_create(struct memory_bitmap *bm, gfp_t gfp_mask, int safe_needed) bb->start_pfn = pfn; if (nr >= BM_BITS_PER_BLOCK) { pfn += BM_BITS_PER_BLOCK; - bb->size = BM_CHUNKS_PER_BLOCK; nr -= BM_BITS_PER_BLOCK; } else { /* This is executed only once in the loop */ pfn += nr; - bb->size = DIV_ROUND_UP(nr, BM_BITS_PER_CHUNK); } bb->end_pfn = pfn; bb = bb->next; @@ -478,8 +469,8 @@ static int memory_bm_find_bit(struct memory_bitmap *bm, unsigned long pfn, } zone_bm->cur_block = bb; pfn -= bb->start_pfn; - *bit_nr = pfn % BM_BITS_PER_CHUNK; - *addr = bb->data + pfn / BM_BITS_PER_CHUNK; + *bit_nr = pfn; + *addr = bb->data; return 0; } @@ -528,36 +519,6 @@ static int memory_bm_test_bit(struct memory_bitmap *bm, unsigned long pfn) return test_bit(bit, addr); } -/* Two auxiliary functions for memory_bm_next_pfn */ - -/* Find the first set bit in the given chunk, if there is one */ - -static inline int next_bit_in_chunk(int bit, unsigned long *chunk_p) -{ - bit++; - while (bit < BM_BITS_PER_CHUNK) { - if (test_bit(bit, chunk_p)) - return bit; - - bit++; - } - return -1; -} - -/* Find a chunk containing some bits set in given block of bits */ - -static inline int next_chunk_in_block(int n, struct bm_block *bb) -{ - n++; - while (n < bb->size) { - if (bb->data[n]) - return n; - - n++; - } - return -1; -} - /** * memory_bm_next_pfn - find the pfn that corresponds to the next set bit * in the bitmap @bm. If the pfn cannot be found, BM_END_OF_MAP is @@ -571,40 +532,33 @@ static unsigned long memory_bm_next_pfn(struct memory_bitmap *bm) { struct zone_bitmap *zone_bm; struct bm_block *bb; - int chunk; int bit; do { bb = bm->cur.block; do { - chunk = bm->cur.chunk; bit = bm->cur.bit; - do { - bit = next_bit_in_chunk(bit, bb->data + chunk); - if (bit >= 0) - goto Return_pfn; - - chunk = next_chunk_in_block(chunk, bb); - bit = -1; - } while (chunk >= 0); + bit = find_next_bit(bb->data, bm_block_bits(bb), bit); + if (bit < bm_block_bits(bb)) + goto Return_pfn; + bb = bb->next; bm->cur.block = bb; - memory_bm_reset_chunk(bm); + bm->cur.bit = 0; } while (bb); zone_bm = bm->cur.zone_bm->next; if (zone_bm) { bm->cur.zone_bm = zone_bm; bm->cur.block = zone_bm->bm_blocks; - memory_bm_reset_chunk(bm); + bm->cur.bit = 0; } } while (zone_bm); memory_bm_position_reset(bm); return BM_END_OF_MAP; Return_pfn: - bm->cur.chunk = chunk; - bm->cur.bit = bit; - return bb->start_pfn + chunk * BM_BITS_PER_CHUNK + bit; + bm->cur.bit = bit + 1; + return bb->start_pfn + bit; } /** -- cgit v1.2.3 From c1a220e7acf8ad2c03504891f4a70cd9c32c904b Mon Sep 17 00:00:00 2001 From: Zhang Rui Date: Wed, 23 Jul 2008 21:28:39 -0700 Subject: pm: introduce new interfaces schedule_work_on() and queue_work_on() This interface allows adding a job on a specific cpu. Although a work struct on a cpu will be scheduled to other cpu if the cpu dies, there is a recursion if a work task tries to offline the cpu it's running on. we need to schedule the task to a specific cpu in this case. http://bugzilla.kernel.org/show_bug.cgi?id=10897 [oleg@tv-sign.ru: cleanups] Signed-off-by: Zhang Rui Tested-by: Rus Signed-off-by: Rafael J. Wysocki Acked-by: Pavel Machek Signed-off-by: Oleg Nesterov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/workqueue.c | 39 ++++++++++++++++++++++++++++++++++++++- 1 file changed, 38 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/workqueue.c b/kernel/workqueue.c index a6d36346d10a..6fd158b21026 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -140,7 +140,6 @@ static void insert_work(struct cpu_workqueue_struct *cwq, wake_up(&cwq->more_work); } -/* Preempt must be disabled. */ static void __queue_work(struct cpu_workqueue_struct *cwq, struct work_struct *work) { @@ -175,6 +174,31 @@ int queue_work(struct workqueue_struct *wq, struct work_struct *work) } EXPORT_SYMBOL_GPL(queue_work); +/** + * queue_work_on - queue work on specific cpu + * @cpu: CPU number to execute work on + * @wq: workqueue to use + * @work: work to queue + * + * Returns 0 if @work was already on a queue, non-zero otherwise. + * + * We queue the work to a specific CPU, the caller must ensure it + * can't go away. + */ +int +queue_work_on(int cpu, struct workqueue_struct *wq, struct work_struct *work) +{ + int ret = 0; + + if (!test_and_set_bit(WORK_STRUCT_PENDING, work_data_bits(work))) { + BUG_ON(!list_empty(&work->entry)); + __queue_work(wq_per_cpu(wq, cpu), work); + ret = 1; + } + return ret; +} +EXPORT_SYMBOL_GPL(queue_work_on); + static void delayed_work_timer_fn(unsigned long __data) { struct delayed_work *dwork = (struct delayed_work *)__data; @@ -553,6 +577,19 @@ int schedule_work(struct work_struct *work) } EXPORT_SYMBOL(schedule_work); +/* + * schedule_work_on - put work task on a specific cpu + * @cpu: cpu to put the work task on + * @work: job to be done + * + * This puts a job on a specific cpu + */ +int schedule_work_on(int cpu, struct work_struct *work) +{ + return queue_work_on(cpu, keventd_wq, work); +} +EXPORT_SYMBOL(schedule_work_on); + /** * schedule_delayed_work - put work task in global workqueue after delay * @dwork: job to be done -- cgit v1.2.3 From 2f15fc4bdf91eb399da3f47a09c55831d9f22826 Mon Sep 17 00:00:00 2001 From: Zhang Rui Date: Wed, 23 Jul 2008 21:28:40 -0700 Subject: pm: schedule sysrq poweroff on boot cpu schedule sysrq poweroff on boot cpu. sysrq poweroff needs to disable nonboot cpus, and we need to run this on boot cpu to avoid any recursion. http://bugzilla.kernel.org/show_bug.cgi?id=10897 [kosaki.motohiro@jp.fujitsu.com: build fix] Signed-off-by: Zhang Rui Tested-by: Rus Signed-off-by: Rafael J. Wysocki Acked-by: Pavel Machek Signed-off-by: KOSAKI Motohiro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/power/poweroff.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/power/poweroff.c b/kernel/power/poweroff.c index 678ec736076b..72016f051477 100644 --- a/kernel/power/poweroff.c +++ b/kernel/power/poweroff.c @@ -10,6 +10,7 @@ #include #include #include +#include /* * When the user hits Sys-Rq o to power down the machine this is the @@ -25,7 +26,8 @@ static DECLARE_WORK(poweroff_work, do_poweroff); static void handle_poweroff(int key, struct tty_struct *tty) { - schedule_work(&poweroff_work); + /* run sysrq poweroff on boot cpu */ + schedule_work_on(first_cpu(cpu_online_map), &poweroff_work); } static struct sysrq_key_op sysrq_poweroff_op = { -- cgit v1.2.3 From f0af566da6e9a4a2f5a83c5a70f3d0a772050e21 Mon Sep 17 00:00:00 2001 From: David Howells Date: Wed, 23 Jul 2008 21:28:44 -0700 Subject: pm: fix try_to_freeze_tasks()'s use of do_div() Fix try_to_freeze_tasks()'s use of do_div() on an s64 by making elapsed_csecs64 a u64 instead and dividing that. Possibly this should be guarded lest the interval calculation turn up negative, but the possible negativity of the result of the division is cast away anyway. This was introduced by patch 438e2ce68dfd4af4cfcec2f873564fb921db4bb5. Signed-off-by: David Howells Acked-by: "Rafael J. Wysocki" Acked-by: Pavel Machek Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/power/process.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/power/process.c b/kernel/power/process.c index 5fb87652f214..278946aecaf0 100644 --- a/kernel/power/process.c +++ b/kernel/power/process.c @@ -149,7 +149,7 @@ static int try_to_freeze_tasks(bool sig_only) unsigned long end_time; unsigned int todo; struct timeval start, end; - s64 elapsed_csecs64; + u64 elapsed_csecs64; unsigned int elapsed_csecs; do_gettimeofday(&start); -- cgit v1.2.3 From 82736f4d1d2b7063b829cc93171a6e5aea8a9c49 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Uwe=20Kleine-K=C3=B6nig?= Date: Wed, 23 Jul 2008 21:28:54 -0700 Subject: generic irqs: handle failure of irqchip->set_type in setup_irq MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit set_type returns an int indicating success or failure, but up to now setup_irq ignores that. In my case this resulted in a machine hang: gpio-keys requested IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING, but arm/ns9xxx can only trigger on one direction so set_type didn't touch the configuration which happens do default to a level sensitiveness and returned -EINVAL. setup_irq ignored that and unmasked the irq. This resulted in an endless triggering of the gpio-key interrupt service routine which effectively killed the machine. With this patch applied setup_irq propagates the error to the caller. Note that before in the case chip && !chip->set_type && !chip->name a NULL pointer was feed to printk. This is fixed, too. Signed-off-by: Uwe Kleine-König Cc: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/irq/manage.c | 64 +++++++++++++++++++++++++++++++++++------------------ 1 file changed, 42 insertions(+), 22 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index 3cfc0fefb5ee..5bc6e5ecc493 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -308,6 +308,30 @@ void compat_irq_chip_set_default_handler(struct irq_desc *desc) desc->handle_irq = NULL; } +static int __irq_set_trigger(struct irq_chip *chip, unsigned int irq, + unsigned long flags) +{ + int ret; + + if (!chip || !chip->set_type) { + /* + * IRQF_TRIGGER_* but the PIC does not support multiple + * flow-types? + */ + pr_warning("No set_type function for IRQ %d (%s)\n", irq, + chip ? (chip->name ? : "unknown") : "unknown"); + return 0; + } + + ret = chip->set_type(irq, flags & IRQF_TRIGGER_MASK); + + if (ret) + pr_err("setting flow type for irq %u failed (%pF)\n", + irq, chip->set_type); + + return ret; +} + /* * Internal function to register an irqaction - typically used to * allocate special interrupts that are part of the architecture. @@ -319,6 +343,7 @@ int setup_irq(unsigned int irq, struct irqaction *new) const char *old_name = NULL; unsigned long flags; int shared = 0; + int ret; if (irq >= NR_IRQS) return -EINVAL; @@ -376,35 +401,23 @@ int setup_irq(unsigned int irq, struct irqaction *new) shared = 1; } - *p = new; - - /* Exclude IRQ from balancing */ - if (new->flags & IRQF_NOBALANCING) - desc->status |= IRQ_NO_BALANCING; - if (!shared) { irq_chip_set_defaults(desc->chip); -#if defined(CONFIG_IRQ_PER_CPU) - if (new->flags & IRQF_PERCPU) - desc->status |= IRQ_PER_CPU; -#endif - /* Setup the type (level, edge polarity) if configured: */ if (new->flags & IRQF_TRIGGER_MASK) { - if (desc->chip->set_type) - desc->chip->set_type(irq, - new->flags & IRQF_TRIGGER_MASK); - else - /* - * IRQF_TRIGGER_* but the PIC does not support - * multiple flow-types? - */ - printk(KERN_WARNING "No IRQF_TRIGGER set_type " - "function for IRQ %d (%s)\n", irq, - desc->chip->name); + ret = __irq_set_trigger(desc->chip, irq, new->flags); + + if (ret) { + spin_unlock_irqrestore(&desc->lock, flags); + return ret; + } } else compat_irq_chip_set_default_handler(desc); +#if defined(CONFIG_IRQ_PER_CPU) + if (new->flags & IRQF_PERCPU) + desc->status |= IRQ_PER_CPU; +#endif desc->status &= ~(IRQ_AUTODETECT | IRQ_WAITING | IRQ_INPROGRESS | IRQ_SPURIOUS_DISABLED); @@ -423,6 +436,13 @@ int setup_irq(unsigned int irq, struct irqaction *new) /* Set default affinity mask once everything is setup */ irq_select_affinity(irq); } + + *p = new; + + /* Exclude IRQ from balancing */ + if (new->flags & IRQF_NOBALANCING) + desc->status |= IRQ_NO_BALANCING; + /* Reset broken irq detection when installing new handler */ desc->irq_count = 0; desc->irqs_unhandled = 0; -- cgit v1.2.3 From aaca0bdca573f3f51ea03139f9c7289541e7bca3 Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Wed, 23 Jul 2008 21:29:20 -0700 Subject: flag parameters: paccept This patch is by far the most complex in the series. It adds a new syscall paccept. This syscall differs from accept in that it adds (at the userlevel) two additional parameters: - a signal mask - a flags value The flags parameter can be used to set flag like SOCK_CLOEXEC. This is imlpemented here as well. Some people argued that this is a property which should be inherited from the file desriptor for the server but this is against POSIX. Additionally, we really want the signal mask parameter as well (similar to pselect, ppoll, etc). So an interface change in inevitable. The flag value is the same as for socket and socketpair. I think diverging here will only create confusion. Similar to the filesystem interfaces where the use of the O_* constants differs, it is acceptable here. The signal mask is handled as for pselect etc. The mask is temporarily installed for the thread and removed before the call returns. I modeled the code after pselect. If there is a problem it's likely also in pselect. For architectures which use socketcall I maintained this interface instead of adding a system call. The symmetry shouldn't be broken. The following test must be adjusted for architectures other than x86 and x86-64 and in case the syscall numbers changed. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ #include #include #include #include #include #include #include #include #include #ifndef __NR_paccept # ifdef __x86_64__ # define __NR_paccept 288 # elif defined __i386__ # define SYS_PACCEPT 18 # define USE_SOCKETCALL 1 # else # error "need __NR_paccept" # endif #endif #ifdef USE_SOCKETCALL # define paccept(fd, addr, addrlen, mask, flags) \ ({ long args[6] = { \ (long) fd, (long) addr, (long) addrlen, (long) mask, 8, (long) flags }; \ syscall (__NR_socketcall, SYS_PACCEPT, args); }) #else # define paccept(fd, addr, addrlen, mask, flags) \ syscall (__NR_paccept, fd, addr, addrlen, mask, 8, flags) #endif #define PORT 57392 #define SOCK_CLOEXEC O_CLOEXEC static pthread_barrier_t b; static void * tf (void *arg) { pthread_barrier_wait (&b); int s = socket (AF_INET, SOCK_STREAM, 0); struct sockaddr_in sin; sin.sin_family = AF_INET; sin.sin_addr.s_addr = htonl (INADDR_LOOPBACK); sin.sin_port = htons (PORT); connect (s, (const struct sockaddr *) &sin, sizeof (sin)); close (s); pthread_barrier_wait (&b); s = socket (AF_INET, SOCK_STREAM, 0); sin.sin_port = htons (PORT); connect (s, (const struct sockaddr *) &sin, sizeof (sin)); close (s); pthread_barrier_wait (&b); pthread_barrier_wait (&b); sleep (2); pthread_kill ((pthread_t) arg, SIGUSR1); return NULL; } static void handler (int s) { } int main (void) { pthread_barrier_init (&b, NULL, 2); struct sockaddr_in sin; pthread_t th; if (pthread_create (&th, NULL, tf, (void *) pthread_self ()) != 0) { puts ("pthread_create failed"); return 1; } int s = socket (AF_INET, SOCK_STREAM, 0); int reuse = 1; setsockopt (s, SOL_SOCKET, SO_REUSEADDR, &reuse, sizeof (reuse)); sin.sin_family = AF_INET; sin.sin_addr.s_addr = htonl (INADDR_LOOPBACK); sin.sin_port = htons (PORT); bind (s, (struct sockaddr *) &sin, sizeof (sin)); listen (s, SOMAXCONN); pthread_barrier_wait (&b); int s2 = paccept (s, NULL, 0, NULL, 0); if (s2 < 0) { puts ("paccept(0) failed"); return 1; } int coe = fcntl (s2, F_GETFD); if (coe & FD_CLOEXEC) { puts ("paccept(0) set close-on-exec-flag"); return 1; } close (s2); pthread_barrier_wait (&b); s2 = paccept (s, NULL, 0, NULL, SOCK_CLOEXEC); if (s2 < 0) { puts ("paccept(SOCK_CLOEXEC) failed"); return 1; } coe = fcntl (s2, F_GETFD); if ((coe & FD_CLOEXEC) == 0) { puts ("paccept(SOCK_CLOEXEC) does not set close-on-exec flag"); return 1; } close (s2); pthread_barrier_wait (&b); struct sigaction sa; sa.sa_handler = handler; sa.sa_flags = 0; sigemptyset (&sa.sa_mask); sigaction (SIGUSR1, &sa, NULL); sigset_t ss; pthread_sigmask (SIG_SETMASK, NULL, &ss); sigaddset (&ss, SIGUSR1); pthread_sigmask (SIG_SETMASK, &ss, NULL); sigdelset (&ss, SIGUSR1); alarm (4); pthread_barrier_wait (&b); errno = 0 ; s2 = paccept (s, NULL, 0, &ss, 0); if (s2 != -1 || errno != EINTR) { puts ("paccept did not fail with EINTR"); return 1; } close (s); puts ("OK"); return 0; } ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ [akpm@linux-foundation.org: make it compile] [akpm@linux-foundation.org: add sys_ni stub] Signed-off-by: Ulrich Drepper Acked-by: Davide Libenzi Cc: Michael Kerrisk Cc: Cc: "David S. Miller" Cc: Roland McGrath Cc: Kyle McMartin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sys_ni.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/sys_ni.c b/kernel/sys_ni.c index 0fea0ee12da9..2f0b8a2e600f 100644 --- a/kernel/sys_ni.c +++ b/kernel/sys_ni.c @@ -31,6 +31,7 @@ cond_syscall(sys_socketpair); cond_syscall(sys_bind); cond_syscall(sys_listen); cond_syscall(sys_accept); +cond_syscall(sys_paccept); cond_syscall(sys_connect); cond_syscall(sys_getsockname); cond_syscall(sys_getpeername); -- cgit v1.2.3 From 9deb27baedb79759c3ab9435a7d8b841842d56e9 Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Wed, 23 Jul 2008 21:29:24 -0700 Subject: flag parameters: signalfd This patch adds the new signalfd4 syscall. It extends the old signalfd syscall by one parameter which is meant to hold a flag value. In this patch the only flag support is SFD_CLOEXEC which causes the close-on-exec flag for the returned file descriptor to be set. A new name SFD_CLOEXEC is introduced which in this implementation must have the same value as O_CLOEXEC. The following test must be adjusted for architectures other than x86 and x86-64 and in case the syscall numbers changed. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ #include #include #include #include #include #ifndef __NR_signalfd4 # ifdef __x86_64__ # define __NR_signalfd4 289 # elif defined __i386__ # define __NR_signalfd4 327 # else # error "need __NR_signalfd4" # endif #endif #define SFD_CLOEXEC O_CLOEXEC int main (void) { sigset_t ss; sigemptyset (&ss); sigaddset (&ss, SIGUSR1); int fd = syscall (__NR_signalfd4, -1, &ss, 8, 0); if (fd == -1) { puts ("signalfd4(0) failed"); return 1; } int coe = fcntl (fd, F_GETFD); if (coe == -1) { puts ("fcntl failed"); return 1; } if (coe & FD_CLOEXEC) { puts ("signalfd4(0) set close-on-exec flag"); return 1; } close (fd); fd = syscall (__NR_signalfd4, -1, &ss, 8, SFD_CLOEXEC); if (fd == -1) { puts ("signalfd4(SFD_CLOEXEC) failed"); return 1; } coe = fcntl (fd, F_GETFD); if (coe == -1) { puts ("fcntl failed"); return 1; } if ((coe & FD_CLOEXEC) == 0) { puts ("signalfd4(SFD_CLOEXEC) does not set close-on-exec flag"); return 1; } close (fd); puts ("OK"); return 0; } ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ [akpm@linux-foundation.org: add sys_ni stub] Signed-off-by: Ulrich Drepper Acked-by: Davide Libenzi Cc: Michael Kerrisk Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sys_ni.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/sys_ni.c b/kernel/sys_ni.c index 2f0b8a2e600f..8627c89ae9e8 100644 --- a/kernel/sys_ni.c +++ b/kernel/sys_ni.c @@ -156,6 +156,7 @@ cond_syscall(sys_ioprio_get); /* New file descriptors */ cond_syscall(sys_signalfd); +cond_syscall(sys_signalfd4); cond_syscall(compat_sys_signalfd); cond_syscall(sys_timerfd_create); cond_syscall(sys_timerfd_settime); -- cgit v1.2.3 From b087498eb5605673b0f260a7620d91818cd72304 Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Wed, 23 Jul 2008 21:29:25 -0700 Subject: flag parameters: eventfd This patch adds the new eventfd2 syscall. It extends the old eventfd syscall by one parameter which is meant to hold a flag value. In this patch the only flag support is EFD_CLOEXEC which causes the close-on-exec flag for the returned file descriptor to be set. A new name EFD_CLOEXEC is introduced which in this implementation must have the same value as O_CLOEXEC. The following test must be adjusted for architectures other than x86 and x86-64 and in case the syscall numbers changed. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ #include #include #include #include #ifndef __NR_eventfd2 # ifdef __x86_64__ # define __NR_eventfd2 290 # elif defined __i386__ # define __NR_eventfd2 328 # else # error "need __NR_eventfd2" # endif #endif #define EFD_CLOEXEC O_CLOEXEC int main (void) { int fd = syscall (__NR_eventfd2, 1, 0); if (fd == -1) { puts ("eventfd2(0) failed"); return 1; } int coe = fcntl (fd, F_GETFD); if (coe == -1) { puts ("fcntl failed"); return 1; } if (coe & FD_CLOEXEC) { puts ("eventfd2(0) sets close-on-exec flag"); return 1; } close (fd); fd = syscall (__NR_eventfd2, 1, EFD_CLOEXEC); if (fd == -1) { puts ("eventfd2(EFD_CLOEXEC) failed"); return 1; } coe = fcntl (fd, F_GETFD); if (coe == -1) { puts ("fcntl failed"); return 1; } if ((coe & FD_CLOEXEC) == 0) { puts ("eventfd2(EFD_CLOEXEC) does not set close-on-exec flag"); return 1; } close (fd); puts ("OK"); return 0; } ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ [akpm@linux-foundation.org: add sys_ni stub] Signed-off-by: Ulrich Drepper Acked-by: Davide Libenzi Cc: Michael Kerrisk Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sys_ni.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/sys_ni.c b/kernel/sys_ni.c index 8627c89ae9e8..2a361ccdc7ca 100644 --- a/kernel/sys_ni.c +++ b/kernel/sys_ni.c @@ -164,3 +164,4 @@ cond_syscall(sys_timerfd_gettime); cond_syscall(compat_sys_timerfd_settime); cond_syscall(compat_sys_timerfd_gettime); cond_syscall(sys_eventfd); +cond_syscall(sys_eventfd2); -- cgit v1.2.3 From 4006553b06306b34054529477b06b68a1c66249b Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Wed, 23 Jul 2008 21:29:32 -0700 Subject: flag parameters: inotify_init This patch introduces the new syscall inotify_init1 (note: the 1 stands for the one parameter the syscall takes, as opposed to no parameter before). The values accepted for this parameter are function-specific and defined in the inotify.h header. Here the values must match the O_* flags, though. In this patch CLOEXEC support is introduced. The following test must be adjusted for architectures other than x86 and x86-64 and in case the syscall numbers changed. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ #include #include #include #include #ifndef __NR_inotify_init1 # ifdef __x86_64__ # define __NR_inotify_init1 294 # elif defined __i386__ # define __NR_inotify_init1 332 # else # error "need __NR_inotify_init1" # endif #endif #define IN_CLOEXEC O_CLOEXEC int main (void) { int fd; fd = syscall (__NR_inotify_init1, 0); if (fd == -1) { puts ("inotify_init1(0) failed"); return 1; } int coe = fcntl (fd, F_GETFD); if (coe == -1) { puts ("fcntl failed"); return 1; } if (coe & FD_CLOEXEC) { puts ("inotify_init1(0) set close-on-exit"); return 1; } close (fd); fd = syscall (__NR_inotify_init1, IN_CLOEXEC); if (fd == -1) { puts ("inotify_init1(IN_CLOEXEC) failed"); return 1; } coe = fcntl (fd, F_GETFD); if (coe == -1) { puts ("fcntl failed"); return 1; } if ((coe & FD_CLOEXEC) == 0) { puts ("inotify_init1(O_CLOEXEC) does not set close-on-exit"); return 1; } close (fd); puts ("OK"); return 0; } ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ [akpm@linux-foundation.org: add sys_ni stub] Signed-off-by: Ulrich Drepper Acked-by: Davide Libenzi Cc: Michael Kerrisk Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sys_ni.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/sys_ni.c b/kernel/sys_ni.c index 2a361ccdc7ca..bd66ac5406f3 100644 --- a/kernel/sys_ni.c +++ b/kernel/sys_ni.c @@ -96,6 +96,7 @@ cond_syscall(sys_keyctl); cond_syscall(compat_sys_keyctl); cond_syscall(compat_sys_socketcall); cond_syscall(sys_inotify_init); +cond_syscall(sys_inotify_init1); cond_syscall(sys_inotify_add_watch); cond_syscall(sys_inotify_rm_watch); cond_syscall(sys_migrate_pages); -- cgit v1.2.3 From be61a86d7237dd80510615f38ae21d6e1e98660c Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Wed, 23 Jul 2008 21:29:40 -0700 Subject: flag parameters: NONBLOCK in pipe This patch adds O_NONBLOCK support to pipe2. It is minimally more involved than the patches for eventfd et.al but still trivial. The interfaces of the create_write_pipe and create_read_pipe helper functions were changed and the one other caller as well. The following test must be adjusted for architectures other than x86 and x86-64 and in case the syscall numbers changed. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ #include #include #include #include #ifndef __NR_pipe2 # ifdef __x86_64__ # define __NR_pipe2 293 # elif defined __i386__ # define __NR_pipe2 331 # else # error "need __NR_pipe2" # endif #endif int main (void) { int fds[2]; if (syscall (__NR_pipe2, fds, 0) == -1) { puts ("pipe2(0) failed"); return 1; } for (int i = 0; i < 2; ++i) { int fl = fcntl (fds[i], F_GETFL); if (fl == -1) { puts ("fcntl failed"); return 1; } if (fl & O_NONBLOCK) { printf ("pipe2(0) set non-blocking mode for fds[%d]\n", i); return 1; } close (fds[i]); } if (syscall (__NR_pipe2, fds, O_NONBLOCK) == -1) { puts ("pipe2(O_NONBLOCK) failed"); return 1; } for (int i = 0; i < 2; ++i) { int fl = fcntl (fds[i], F_GETFL); if (fl == -1) { puts ("fcntl failed"); return 1; } if ((fl & O_NONBLOCK) == 0) { printf ("pipe2(O_NONBLOCK) does not set non-blocking mode for fds[%d]\n", i); return 1; } close (fds[i]); } puts ("OK"); return 0; } ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Signed-off-by: Ulrich Drepper Acked-by: Davide Libenzi Cc: Michael Kerrisk Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/kmod.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/kmod.c b/kernel/kmod.c index 90d7af1c1655..2989f67c4446 100644 --- a/kernel/kmod.c +++ b/kernel/kmod.c @@ -417,12 +417,12 @@ int call_usermodehelper_stdinpipe(struct subprocess_info *sub_info, { struct file *f; - f = create_write_pipe(); + f = create_write_pipe(0); if (IS_ERR(f)) return PTR_ERR(f); *filp = f; - f = create_read_pipe(f); + f = create_read_pipe(f, 0); if (IS_ERR(f)) { free_write_pipe(*filp); return PTR_ERR(f); -- cgit v1.2.3 From 5df439ef06d4173357711a04740aa8bfcf50d621 Mon Sep 17 00:00:00 2001 From: Wang Chen Date: Fri, 25 Jul 2008 01:45:23 -0700 Subject: flag parameters: fix compile error of sys_epoll_create1 GEN .version CHK include/linux/compile.h UPD include/linux/compile.h CC init/version.o LD init/built-in.o LD vmlinux arch/x86/kernel/built-in.o: In function `sys_call_table': (.rodata+0x8a4): undefined reference to `sys_epoll_create1' make: *** [vmlinux] Error 1 Signed-off-by: Wang Chen Cc: Ulrich Drepper Cc: Davide Libenzi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sys_ni.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/sys_ni.c b/kernel/sys_ni.c index bd66ac5406f3..55eca1594da9 100644 --- a/kernel/sys_ni.c +++ b/kernel/sys_ni.c @@ -57,6 +57,7 @@ cond_syscall(compat_sys_set_robust_list); cond_syscall(sys_get_robust_list); cond_syscall(compat_sys_get_robust_list); cond_syscall(sys_epoll_create); +cond_syscall(sys_epoll_create1); cond_syscall(sys_epoll_ctl); cond_syscall(sys_epoll_wait); cond_syscall(sys_epoll_pwait); -- cgit v1.2.3 From 2fc9c4e18f94431e7eb77d97edb2a995b46fba55 Mon Sep 17 00:00:00 2001 From: Vegard Nossum Date: Fri, 25 Jul 2008 01:45:34 -0700 Subject: kallsyms: fix potential overflow in binary search This will probably never trigger... but it won't hurt to be careful. http://googleresearch.blogspot.com/2006/06/extra-extra-read-all-about-it-nearly.html Signed-off-by: Vegard Nossum Cc: Joshua Bloch Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/kallsyms.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/kallsyms.c b/kernel/kallsyms.c index 6fc0040f3e3a..38fc10ac7541 100644 --- a/kernel/kallsyms.c +++ b/kernel/kallsyms.c @@ -176,7 +176,7 @@ static unsigned long get_symbol_pos(unsigned long addr, high = kallsyms_num_syms; while (high - low > 1) { - mid = (low + high) / 2; + mid = low + (high - low) / 2; if (kallsyms_addresses[mid] <= addr) low = mid; else -- cgit v1.2.3 From b03f6489f9f27dc519a4c60ebf39cc7b8a58eae7 Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Fri, 25 Jul 2008 01:45:35 -0700 Subject: build kernel/profile.o only when requested Build kernel/profile.o only if CONFIG_PROFILING is enabled. This makes CONFIG_PROFILING=n kernels smaller. As a bonus, some profile_tick() calls and one branch from schedule() are now eliminated with CONFIG_PROFILING=n (but I doubt these are measurable effects). This patch changes the effects of CONFIG_PROFILING=n, but I don't think having more than two choices would be the better choice. This patch also adds the name of the first parameter to the prototypes of profile_{hits,tick}() since I anyway had to add them for the dummy functions. Signed-off-by: Adrian Bunk Cc: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/Makefile | 3 ++- kernel/profile.c | 4 ---- 2 files changed, 2 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/Makefile b/kernel/Makefile index 15ab63ffe64d..54f69837d35a 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -2,7 +2,7 @@ # Makefile for the linux kernel. # -obj-y = sched.o fork.o exec_domain.o panic.o printk.o profile.o \ +obj-y = sched.o fork.o exec_domain.o panic.o printk.o \ cpu.o exit.o itimer.o time.o softirq.o resource.o \ sysctl.o capability.o ptrace.o timer.o user.o \ signal.o sys.o kmod.o workqueue.o pid.o \ @@ -24,6 +24,7 @@ CFLAGS_REMOVE_sched_clock.o = -pg CFLAGS_REMOVE_sched.o = -mno-spe -pg endif +obj-$(CONFIG_PROFILING) += profile.o obj-$(CONFIG_SYSCTL_SYSCALL_CHECK) += sysctl_check.o obj-$(CONFIG_STACKTRACE) += stacktrace.o obj-y += time/ diff --git a/kernel/profile.c b/kernel/profile.c index 58926411eb2a..cd26bed4cc26 100644 --- a/kernel/profile.c +++ b/kernel/profile.c @@ -112,8 +112,6 @@ void __init profile_init(void) /* Profile event notifications */ -#ifdef CONFIG_PROFILING - static BLOCKING_NOTIFIER_HEAD(task_exit_notifier); static ATOMIC_NOTIFIER_HEAD(task_free_notifier); static BLOCKING_NOTIFIER_HEAD(munmap_notifier); @@ -203,8 +201,6 @@ void unregister_timer_hook(int (*hook)(struct pt_regs *)) } EXPORT_SYMBOL_GPL(unregister_timer_hook); -#endif /* CONFIG_PROFILING */ - #ifdef CONFIG_SMP /* -- cgit v1.2.3 From ac331d158e198d2a91a5b0a3ec4ca9991fdb57af Mon Sep 17 00:00:00 2001 From: KOSAKI Motohiro Date: Fri, 25 Jul 2008 01:45:38 -0700 Subject: call_usermodehelper(): increase reliability Presently call_usermodehelper_setup() uses GFP_ATOMIC. but it can return NULL _very_ easily. GFP_ATOMIC is needed only when we can't sleep. and, GFP_KERNEL is robust and better. thus, I add gfp_mask argument to call_usermodehelper_setup(). So, its callers pass the gfp_t as below: call_usermodehelper() and call_usermodehelper_keys(): depend on 'wait' argument. call_usermodehelper_pipe(): always GFP_KERNEL because always run under process context. orderly_poweroff(): pass to GFP_ATOMIC because may run under interrupt context. Signed-off-by: KOSAKI Motohiro Cc: "Paul Menage" Reviewed-by: Li Zefan Acked-by: Jeremy Fitzhardinge Cc: Rusty Russell Cc: Andi Kleen Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/kmod.c | 9 +++++---- kernel/sys.c | 2 +- 2 files changed, 6 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/kmod.c b/kernel/kmod.c index 2989f67c4446..2456d1a0befb 100644 --- a/kernel/kmod.c +++ b/kernel/kmod.c @@ -352,16 +352,17 @@ static inline void register_pm_notifier_callback(void) {} * @path: path to usermode executable * @argv: arg vector for process * @envp: environment for process + * @gfp_mask: gfp mask for memory allocation * * Returns either %NULL on allocation failure, or a subprocess_info * structure. This should be passed to call_usermodehelper_exec to * exec the process and free the structure. */ -struct subprocess_info *call_usermodehelper_setup(char *path, - char **argv, char **envp) +struct subprocess_info *call_usermodehelper_setup(char *path, char **argv, + char **envp, gfp_t gfp_mask) { struct subprocess_info *sub_info; - sub_info = kzalloc(sizeof(struct subprocess_info), GFP_ATOMIC); + sub_info = kzalloc(sizeof(struct subprocess_info), gfp_mask); if (!sub_info) goto out; @@ -494,7 +495,7 @@ int call_usermodehelper_pipe(char *path, char **argv, char **envp, struct subprocess_info *sub_info; int ret; - sub_info = call_usermodehelper_setup(path, argv, envp); + sub_info = call_usermodehelper_setup(path, argv, envp, GFP_KERNEL); if (sub_info == NULL) return -ENOMEM; diff --git a/kernel/sys.c b/kernel/sys.c index 14e97282eb6c..6c2188046048 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -1795,7 +1795,7 @@ int orderly_poweroff(bool force) goto out; } - info = call_usermodehelper_setup(argv[0], argv, envp); + info = call_usermodehelper_setup(argv[0], argv, envp, GFP_ATOMIC); if (info == NULL) { argv_free(argv); goto out; -- cgit v1.2.3 From b69c49b78457f681ecfb3147bd968434ee6559c1 Mon Sep 17 00:00:00 2001 From: FUJITA Tomonori Date: Fri, 25 Jul 2008 01:45:40 -0700 Subject: clean up duplicated alloc/free_thread_info We duplicate alloc/free_thread_info defines on many platforms (the majority uses __get_free_pages/free_pages). This patch defines common defines and removes these duplicated defines. __HAVE_ARCH_THREAD_INFO_ALLOCATOR is introduced for platforms that do something different. Signed-off-by: FUJITA Tomonori Acked-by: Russell King Cc: Pekka Enberg Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/fork.c | 17 +++++++++++++++++ 1 file changed, 17 insertions(+) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index 552c8d8e77ad..5a5d6fef341d 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -93,6 +93,23 @@ int nr_processes(void) static struct kmem_cache *task_struct_cachep; #endif +#ifndef __HAVE_ARCH_THREAD_INFO_ALLOCATOR +static inline struct thread_info *alloc_thread_info(struct task_struct *tsk) +{ +#ifdef CONFIG_DEBUG_STACK_USAGE + gfp_t mask = GFP_KERNEL | __GFP_ZERO; +#else + gfp_t mask = GFP_KERNEL; +#endif + return (struct thread_info *)__get_free_pages(mask, THREAD_SIZE_ORDER); +} + +static inline void free_thread_info(struct thread_info *ti) +{ + free_pages((unsigned long)ti, THREAD_SIZE_ORDER); +} +#endif + /* SLAB cache for signal_struct structures (tsk->signal) */ static struct kmem_cache *signal_cachep; -- cgit v1.2.3 From a8f18b909c0a3f22630846207035c8b84bb252b8 Mon Sep 17 00:00:00 2001 From: Arjan van de Ven Date: Fri, 25 Jul 2008 01:45:53 -0700 Subject: Add a WARN() macro; this is WARN_ON() + printk arguments Add a WARN() macro that acts like WARN_ON(), with the added feature that it takes a printk like argument that is printed as part of the warning message. [akpm@linux-foundation.org: fix printk arguments] [akpm@linux-foundation.org: coding-style fixes] Signed-off-by: Arjan van de Ven Cc: Greg KH Cc: Jiri Slaby Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/panic.c | 22 ++++++++++++++++++++++ 1 file changed, 22 insertions(+) (limited to 'kernel') diff --git a/kernel/panic.c b/kernel/panic.c index 425567f45b9f..12c5a0a6c89b 100644 --- a/kernel/panic.c +++ b/kernel/panic.c @@ -318,6 +318,28 @@ void warn_on_slowpath(const char *file, int line) add_taint(TAINT_WARN); } EXPORT_SYMBOL(warn_on_slowpath); + + +void warn_slowpath(const char *file, int line, const char *fmt, ...) +{ + va_list args; + char function[KSYM_SYMBOL_LEN]; + unsigned long caller = (unsigned long)__builtin_return_address(0); + sprint_symbol(function, caller); + + printk(KERN_WARNING "------------[ cut here ]------------\n"); + printk(KERN_WARNING "WARNING: at %s:%d %s()\n", file, + line, function); + va_start(args, fmt); + vprintk(fmt, args); + va_end(args); + + print_modules(); + dump_stack(); + print_oops_end_marker(); + add_taint(TAINT_WARN); +} +EXPORT_SYMBOL(warn_slowpath); #endif #ifdef CONFIG_CC_STACKPROTECTOR -- cgit v1.2.3 From 7a2c477069fbd32f91598f05334003979b987a39 Mon Sep 17 00:00:00 2001 From: Arjan van de Ven Date: Fri, 25 Jul 2008 01:45:54 -0700 Subject: kernel/irq/manage.c: replace a printk + WARN_ON() to a WARN() Replace a printk+WARN_ON() by a WARN(); this increases the chance of the string making it into the bugreport (ie: it goes inside the ---[ cut here ]--- section) Signed-off-by: Arjan van de Ven Cc: Thomas Gleixner Cc: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/irq/manage.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index 5bc6e5ecc493..f8914b92b664 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -260,9 +260,7 @@ int set_irq_wake(unsigned int irq, unsigned int on) } } else { if (desc->wake_depth == 0) { - printk(KERN_WARNING "Unbalanced IRQ %d " - "wake disable\n", irq); - WARN_ON(1); + WARN(1, "Unbalanced IRQ %d wake disable\n", irq); } else if (--desc->wake_depth == 0) { ret = set_irq_wake_real(irq, on); if (ret) -- cgit v1.2.3 From 717115e1a5856b57af0f71e1df7149108294fc10 Mon Sep 17 00:00:00 2001 From: Dave Young Date: Fri, 25 Jul 2008 01:45:58 -0700 Subject: printk ratelimiting rewrite All ratelimit user use same jiffies and burst params, so some messages (callbacks) will be lost. For example: a call printk_ratelimit(5 * HZ, 1) b call printk_ratelimit(5 * HZ, 1) before the 5*HZ timeout of a, then b will will be supressed. - rewrite __ratelimit, and use a ratelimit_state as parameter. Thanks for hints from andrew. - Add WARN_ON_RATELIMIT, update rcupreempt.h - remove __printk_ratelimit - use __ratelimit in net_ratelimit Signed-off-by: Dave Young Cc: "David S. Miller" Cc: "Paul E. McKenney" Cc: Dave Young Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/printk.c | 17 +++-------------- kernel/sysctl.c | 4 ++-- 2 files changed, 5 insertions(+), 16 deletions(-) (limited to 'kernel') diff --git a/kernel/printk.c b/kernel/printk.c index 3f7a2a94583b..a7f7559c5f6c 100644 --- a/kernel/printk.c +++ b/kernel/printk.c @@ -1308,6 +1308,8 @@ void tty_write_message(struct tty_struct *tty, char *msg) } #if defined CONFIG_PRINTK + +DEFINE_RATELIMIT_STATE(printk_ratelimit_state, 5 * HZ, 10); /* * printk rate limiting, lifted from the networking subsystem. * @@ -1315,22 +1317,9 @@ void tty_write_message(struct tty_struct *tty, char *msg) * every printk_ratelimit_jiffies to make a denial-of-service * attack impossible. */ -int __printk_ratelimit(int ratelimit_jiffies, int ratelimit_burst) -{ - return __ratelimit(ratelimit_jiffies, ratelimit_burst); -} -EXPORT_SYMBOL(__printk_ratelimit); - -/* minimum time in jiffies between messages */ -int printk_ratelimit_jiffies = 5 * HZ; - -/* number of messages we send before ratelimiting */ -int printk_ratelimit_burst = 10; - int printk_ratelimit(void) { - return __printk_ratelimit(printk_ratelimit_jiffies, - printk_ratelimit_burst); + return __ratelimit(&printk_ratelimit_state); } EXPORT_SYMBOL(printk_ratelimit); diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 1a8299d1fe59..35a50db9b6ce 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -624,7 +624,7 @@ static struct ctl_table kern_table[] = { { .ctl_name = KERN_PRINTK_RATELIMIT, .procname = "printk_ratelimit", - .data = &printk_ratelimit_jiffies, + .data = &printk_ratelimit_state.interval, .maxlen = sizeof(int), .mode = 0644, .proc_handler = &proc_dointvec_jiffies, @@ -633,7 +633,7 @@ static struct ctl_table kern_table[] = { { .ctl_name = KERN_PRINTK_RATELIMIT_BURST, .procname = "printk_ratelimit_burst", - .data = &printk_ratelimit_burst, + .data = &printk_ratelimit_state.burst, .maxlen = sizeof(int), .mode = 0644, .proc_handler = &proc_dointvec, -- cgit v1.2.3 From ef53d9c5e4da147ecaa43c44c5e5945eb83970a2 Mon Sep 17 00:00:00 2001 From: Srinivasa D S Date: Fri, 25 Jul 2008 01:46:04 -0700 Subject: kprobes: improve kretprobe scalability with hashed locking Currently list of kretprobe instances are stored in kretprobe object (as used_instances,free_instances) and in kretprobe hash table. We have one global kretprobe lock to serialise the access to these lists. This causes only one kretprobe handler to execute at a time. Hence affects system performance, particularly on SMP systems and when return probe is set on lot of functions (like on all systemcalls). Solution proposed here gives fine-grain locks that performs better on SMP system compared to present kretprobe implementation. Solution: 1) Instead of having one global lock to protect kretprobe instances present in kretprobe object and kretprobe hash table. We will have two locks, one lock for protecting kretprobe hash table and another lock for kretporbe object. 2) We hold lock present in kretprobe object while we modify kretprobe instance in kretprobe object and we hold per-hash-list lock while modifying kretprobe instances present in that hash list. To prevent deadlock, we never grab a per-hash-list lock while holding a kretprobe lock. 3) We can remove used_instances from struct kretprobe, as we can track used instances of kretprobe instances using kretprobe hash table. Time duration for kernel compilation ("make -j 8") on a 8-way ppc64 system with return probes set on all systemcalls looks like this. cacheline non-cacheline Un-patched kernel aligned patch aligned patch =============================================================================== real 9m46.784s 9m54.412s 10m2.450s user 40m5.715s 40m7.142s 40m4.273s sys 2m57.754s 2m58.583s 3m17.430s =========================================================== Time duration for kernel compilation ("make -j 8) on the same system, when kernel is not probed. ========================= real 9m26.389s user 40m8.775s sys 2m7.283s ========================= Signed-off-by: Srinivasa DS Signed-off-by: Jim Keniston Acked-by: Ananth N Mavinakayanahalli Cc: Anil S Keshavamurthy Cc: David S. Miller Cc: Masami Hiramatsu Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/kprobes.c | 127 ++++++++++++++++++++++++++++++++++++++----------------- 1 file changed, 89 insertions(+), 38 deletions(-) (limited to 'kernel') diff --git a/kernel/kprobes.c b/kernel/kprobes.c index 1485ca8d0e00..cb0b3bde3617 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -62,6 +62,7 @@ addr = ((kprobe_opcode_t *)(kallsyms_lookup_name(name))) #endif +static int kprobes_initialized; static struct hlist_head kprobe_table[KPROBE_TABLE_SIZE]; static struct hlist_head kretprobe_inst_table[KPROBE_TABLE_SIZE]; @@ -69,8 +70,15 @@ static struct hlist_head kretprobe_inst_table[KPROBE_TABLE_SIZE]; static bool kprobe_enabled; DEFINE_MUTEX(kprobe_mutex); /* Protects kprobe_table */ -DEFINE_SPINLOCK(kretprobe_lock); /* Protects kretprobe_inst_table */ static DEFINE_PER_CPU(struct kprobe *, kprobe_instance) = NULL; +static struct { + spinlock_t lock ____cacheline_aligned; +} kretprobe_table_locks[KPROBE_TABLE_SIZE]; + +static spinlock_t *kretprobe_table_lock_ptr(unsigned long hash) +{ + return &(kretprobe_table_locks[hash].lock); +} /* * Normally, functions that we'd want to prohibit kprobes in, are marked @@ -368,26 +376,53 @@ void __kprobes kprobes_inc_nmissed_count(struct kprobe *p) return; } -/* Called with kretprobe_lock held */ void __kprobes recycle_rp_inst(struct kretprobe_instance *ri, struct hlist_head *head) { + struct kretprobe *rp = ri->rp; + /* remove rp inst off the rprobe_inst_table */ hlist_del(&ri->hlist); - if (ri->rp) { - /* remove rp inst off the used list */ - hlist_del(&ri->uflist); - /* put rp inst back onto the free list */ - INIT_HLIST_NODE(&ri->uflist); - hlist_add_head(&ri->uflist, &ri->rp->free_instances); + INIT_HLIST_NODE(&ri->hlist); + if (likely(rp)) { + spin_lock(&rp->lock); + hlist_add_head(&ri->hlist, &rp->free_instances); + spin_unlock(&rp->lock); } else /* Unregistering */ hlist_add_head(&ri->hlist, head); } -struct hlist_head __kprobes *kretprobe_inst_table_head(struct task_struct *tsk) +void kretprobe_hash_lock(struct task_struct *tsk, + struct hlist_head **head, unsigned long *flags) +{ + unsigned long hash = hash_ptr(tsk, KPROBE_HASH_BITS); + spinlock_t *hlist_lock; + + *head = &kretprobe_inst_table[hash]; + hlist_lock = kretprobe_table_lock_ptr(hash); + spin_lock_irqsave(hlist_lock, *flags); +} + +void kretprobe_table_lock(unsigned long hash, unsigned long *flags) { - return &kretprobe_inst_table[hash_ptr(tsk, KPROBE_HASH_BITS)]; + spinlock_t *hlist_lock = kretprobe_table_lock_ptr(hash); + spin_lock_irqsave(hlist_lock, *flags); +} + +void kretprobe_hash_unlock(struct task_struct *tsk, unsigned long *flags) +{ + unsigned long hash = hash_ptr(tsk, KPROBE_HASH_BITS); + spinlock_t *hlist_lock; + + hlist_lock = kretprobe_table_lock_ptr(hash); + spin_unlock_irqrestore(hlist_lock, *flags); +} + +void kretprobe_table_unlock(unsigned long hash, unsigned long *flags) +{ + spinlock_t *hlist_lock = kretprobe_table_lock_ptr(hash); + spin_unlock_irqrestore(hlist_lock, *flags); } /* @@ -401,17 +436,21 @@ void __kprobes kprobe_flush_task(struct task_struct *tk) struct kretprobe_instance *ri; struct hlist_head *head, empty_rp; struct hlist_node *node, *tmp; - unsigned long flags = 0; + unsigned long hash, flags = 0; - INIT_HLIST_HEAD(&empty_rp); - spin_lock_irqsave(&kretprobe_lock, flags); - head = kretprobe_inst_table_head(tk); + if (unlikely(!kprobes_initialized)) + /* Early boot. kretprobe_table_locks not yet initialized. */ + return; + + hash = hash_ptr(tk, KPROBE_HASH_BITS); + head = &kretprobe_inst_table[hash]; + kretprobe_table_lock(hash, &flags); hlist_for_each_entry_safe(ri, node, tmp, head, hlist) { if (ri->task == tk) recycle_rp_inst(ri, &empty_rp); } - spin_unlock_irqrestore(&kretprobe_lock, flags); - + kretprobe_table_unlock(hash, &flags); + INIT_HLIST_HEAD(&empty_rp); hlist_for_each_entry_safe(ri, node, tmp, &empty_rp, hlist) { hlist_del(&ri->hlist); kfree(ri); @@ -423,24 +462,29 @@ static inline void free_rp_inst(struct kretprobe *rp) struct kretprobe_instance *ri; struct hlist_node *pos, *next; - hlist_for_each_entry_safe(ri, pos, next, &rp->free_instances, uflist) { - hlist_del(&ri->uflist); + hlist_for_each_entry_safe(ri, pos, next, &rp->free_instances, hlist) { + hlist_del(&ri->hlist); kfree(ri); } } static void __kprobes cleanup_rp_inst(struct kretprobe *rp) { - unsigned long flags; + unsigned long flags, hash; struct kretprobe_instance *ri; struct hlist_node *pos, *next; + struct hlist_head *head; + /* No race here */ - spin_lock_irqsave(&kretprobe_lock, flags); - hlist_for_each_entry_safe(ri, pos, next, &rp->used_instances, uflist) { - ri->rp = NULL; - hlist_del(&ri->uflist); + for (hash = 0; hash < KPROBE_TABLE_SIZE; hash++) { + kretprobe_table_lock(hash, &flags); + head = &kretprobe_inst_table[hash]; + hlist_for_each_entry_safe(ri, pos, next, head, hlist) { + if (ri->rp == rp) + ri->rp = NULL; + } + kretprobe_table_unlock(hash, &flags); } - spin_unlock_irqrestore(&kretprobe_lock, flags); free_rp_inst(rp); } @@ -831,32 +875,37 @@ static int __kprobes pre_handler_kretprobe(struct kprobe *p, struct pt_regs *regs) { struct kretprobe *rp = container_of(p, struct kretprobe, kp); - unsigned long flags = 0; + unsigned long hash, flags = 0; + struct kretprobe_instance *ri; /*TODO: consider to only swap the RA after the last pre_handler fired */ - spin_lock_irqsave(&kretprobe_lock, flags); + hash = hash_ptr(current, KPROBE_HASH_BITS); + spin_lock_irqsave(&rp->lock, flags); if (!hlist_empty(&rp->free_instances)) { - struct kretprobe_instance *ri; - ri = hlist_entry(rp->free_instances.first, - struct kretprobe_instance, uflist); + struct kretprobe_instance, hlist); + hlist_del(&ri->hlist); + spin_unlock_irqrestore(&rp->lock, flags); + ri->rp = rp; ri->task = current; if (rp->entry_handler && rp->entry_handler(ri, regs)) { - spin_unlock_irqrestore(&kretprobe_lock, flags); + spin_unlock_irqrestore(&rp->lock, flags); return 0; } arch_prepare_kretprobe(ri, regs); /* XXX(hch): why is there no hlist_move_head? */ - hlist_del(&ri->uflist); - hlist_add_head(&ri->uflist, &ri->rp->used_instances); - hlist_add_head(&ri->hlist, kretprobe_inst_table_head(ri->task)); - } else + INIT_HLIST_NODE(&ri->hlist); + kretprobe_table_lock(hash, &flags); + hlist_add_head(&ri->hlist, &kretprobe_inst_table[hash]); + kretprobe_table_unlock(hash, &flags); + } else { rp->nmissed++; - spin_unlock_irqrestore(&kretprobe_lock, flags); + spin_unlock_irqrestore(&rp->lock, flags); + } return 0; } @@ -892,7 +941,7 @@ static int __kprobes __register_kretprobe(struct kretprobe *rp, rp->maxactive = NR_CPUS; #endif } - INIT_HLIST_HEAD(&rp->used_instances); + spin_lock_init(&rp->lock); INIT_HLIST_HEAD(&rp->free_instances); for (i = 0; i < rp->maxactive; i++) { inst = kmalloc(sizeof(struct kretprobe_instance) + @@ -901,8 +950,8 @@ static int __kprobes __register_kretprobe(struct kretprobe *rp, free_rp_inst(rp); return -ENOMEM; } - INIT_HLIST_NODE(&inst->uflist); - hlist_add_head(&inst->uflist, &rp->free_instances); + INIT_HLIST_NODE(&inst->hlist); + hlist_add_head(&inst->hlist, &rp->free_instances); } rp->nmissed = 0; @@ -1009,6 +1058,7 @@ static int __init init_kprobes(void) for (i = 0; i < KPROBE_TABLE_SIZE; i++) { INIT_HLIST_HEAD(&kprobe_table[i]); INIT_HLIST_HEAD(&kretprobe_inst_table[i]); + spin_lock_init(&(kretprobe_table_locks[i].lock)); } /* @@ -1050,6 +1100,7 @@ static int __init init_kprobes(void) err = arch_init_kprobes(); if (!err) err = register_die_notifier(&kprobe_exceptions_nb); + kprobes_initialized = (err == 0); if (!err) init_test_probes(); -- cgit v1.2.3 From 8b6dd986823a8d92ed9f54baa5cef8604d9d9d44 Mon Sep 17 00:00:00 2001 From: Abhishek Sagar Date: Fri, 25 Jul 2008 01:46:05 -0700 Subject: kprobes: remove redundant config check I noticed that there's a CONFIG_KPROBES check inside kernel/kprobes.c, which is redundant. Signed-off-by: Abhishek Sagar Acked-by: Masami Hiramatsu Cc: Ananth N Mavinakayanahalli Cc: Anil S Keshavamurthy Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/kprobes.c | 5 ----- 1 file changed, 5 deletions(-) (limited to 'kernel') diff --git a/kernel/kprobes.c b/kernel/kprobes.c index cb0b3bde3617..75bc2cd9ebc6 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -1337,13 +1337,8 @@ EXPORT_SYMBOL_GPL(register_jprobe); EXPORT_SYMBOL_GPL(unregister_jprobe); EXPORT_SYMBOL_GPL(register_jprobes); EXPORT_SYMBOL_GPL(unregister_jprobes); -#ifdef CONFIG_KPROBES EXPORT_SYMBOL_GPL(jprobe_return); -#endif - -#ifdef CONFIG_KPROBES EXPORT_SYMBOL_GPL(register_kretprobe); EXPORT_SYMBOL_GPL(unregister_kretprobe); EXPORT_SYMBOL_GPL(register_kretprobes); EXPORT_SYMBOL_GPL(unregister_kretprobes); -#endif -- cgit v1.2.3 From 7e9abd89cbdf9b73d327d8173343abce9022609b Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Fri, 25 Jul 2008 01:46:54 -0700 Subject: cgroup: use read lock to guard find_existing_css_set() The function does not modify anything (except the temporary css template), so it's sufficient to hold read lock. Signed-off-by: Li Zefan Acked-by: Paul Menage Cc: Balbir Singh Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cgroup.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 15ac0e1e4f4d..f50edadfdd86 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -415,11 +415,11 @@ static struct css_set *find_css_set( /* First see if we already have a cgroup group that matches * the desired set */ - write_lock(&css_set_lock); + read_lock(&css_set_lock); res = find_existing_css_set(oldcg, cgrp, template); if (res) get_css_set(res); - write_unlock(&css_set_lock); + read_unlock(&css_set_lock); if (res) return res; -- cgit v1.2.3 From 71cbb949d17d4d776abd547135feb7f3282405c8 Mon Sep 17 00:00:00 2001 From: KOSAKI Motohiro Date: Fri, 25 Jul 2008 01:46:55 -0700 Subject: cgroup: list_for_each cleanup -------------------------- while() { list_entry(); ... } -------------------------- is equivalent to following code. -------------------------- list_for_each_entry(){ ... } -------------------------- later can review easily more. this patch is just clean up. it doesn't have any behavor change. Signed-off-by: KOSAKI Motohiro Cc: Paul Menage Cc: Li Zefan Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cgroup.c | 44 +++++++++++++++++++++----------------------- 1 file changed, 21 insertions(+), 23 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index f50edadfdd86..6836a9063634 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -241,17 +241,20 @@ static int use_task_css_set_links; */ static void unlink_css_set(struct css_set *cg) { + struct cg_cgroup_link *link; + struct cg_cgroup_link *saved_link; + write_lock(&css_set_lock); hlist_del(&cg->hlist); css_set_count--; - while (!list_empty(&cg->cg_links)) { - struct cg_cgroup_link *link; - link = list_entry(cg->cg_links.next, - struct cg_cgroup_link, cg_link_list); + + list_for_each_entry_safe(link, saved_link, &cg->cg_links, + cg_link_list) { list_del(&link->cg_link_list); list_del(&link->cgrp_link_list); kfree(link); } + write_unlock(&css_set_lock); } @@ -363,15 +366,14 @@ static struct css_set *find_existing_css_set( static int allocate_cg_links(int count, struct list_head *tmp) { struct cg_cgroup_link *link; + struct cg_cgroup_link *saved_link; int i; INIT_LIST_HEAD(tmp); for (i = 0; i < count; i++) { link = kmalloc(sizeof(*link), GFP_KERNEL); if (!link) { - while (!list_empty(tmp)) { - link = list_entry(tmp->next, - struct cg_cgroup_link, - cgrp_link_list); + list_for_each_entry_safe(link, saved_link, tmp, + cgrp_link_list) { list_del(&link->cgrp_link_list); kfree(link); } @@ -384,11 +386,10 @@ static int allocate_cg_links(int count, struct list_head *tmp) static void free_cg_links(struct list_head *tmp) { - while (!list_empty(tmp)) { - struct cg_cgroup_link *link; - link = list_entry(tmp->next, - struct cg_cgroup_link, - cgrp_link_list); + struct cg_cgroup_link *link; + struct cg_cgroup_link *saved_link; + + list_for_each_entry_safe(link, saved_link, tmp, cgrp_link_list) { list_del(&link->cgrp_link_list); kfree(link); } @@ -1093,6 +1094,8 @@ static void cgroup_kill_sb(struct super_block *sb) { struct cgroupfs_root *root = sb->s_fs_info; struct cgroup *cgrp = &root->top_cgroup; int ret; + struct cg_cgroup_link *link; + struct cg_cgroup_link *saved_link; BUG_ON(!root); @@ -1112,10 +1115,9 @@ static void cgroup_kill_sb(struct super_block *sb) { * root cgroup */ write_lock(&css_set_lock); - while (!list_empty(&cgrp->css_sets)) { - struct cg_cgroup_link *link; - link = list_entry(cgrp->css_sets.next, - struct cg_cgroup_link, cgrp_link_list); + + list_for_each_entry_safe(link, saved_link, &cgrp->css_sets, + cgrp_link_list) { list_del(&link->cg_link_list); list_del(&link->cgrp_link_list); kfree(link); @@ -1756,15 +1758,11 @@ int cgroup_add_files(struct cgroup *cgrp, int cgroup_task_count(const struct cgroup *cgrp) { int count = 0; - struct list_head *l; + struct cg_cgroup_link *link; read_lock(&css_set_lock); - l = cgrp->css_sets.next; - while (l != &cgrp->css_sets) { - struct cg_cgroup_link *link = - list_entry(l, struct cg_cgroup_link, cgrp_link_list); + list_for_each_entry(link, &cgrp->css_sets, cgrp_link_list) { count += atomic_read(&link->cg->ref.refcount); - l = l->next; } read_unlock(&css_set_lock); return count; -- cgit v1.2.3 From 8947f9d5b361ce927be6d5c11fed57905b7a4100 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Fri, 25 Jul 2008 01:46:56 -0700 Subject: cgroups: annotate two variables with __read_mostly - need_forkexit_callback will be read only after system boot. - use_task_css_set_links will be read only after it's set. And these 2 variables are checked when a new process is forked. Signed-off-by: Li Zefan Acked-by: Paul Menage Acked-by: KOSAKI Motohiro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cgroup.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 6836a9063634..70d083c6fb6b 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -118,7 +118,7 @@ static int root_count; * extra work in the fork/exit path if none of the subsystems need to * be called. */ -static int need_forkexit_callback; +static int need_forkexit_callback __read_mostly; static int need_mm_owner_callback __read_mostly; /* convenient tests for these bits */ @@ -220,7 +220,7 @@ static struct hlist_head *css_set_hash(struct cgroup_subsys_state *css[]) * task until after the first call to cgroup_iter_start(). This * reduces the fork()/exit() overhead for people who have cgroups * compiled into their kernel but not actually in use */ -static int use_task_css_set_links; +static int use_task_css_set_links __read_mostly; /* When we create or destroy a css_set, the operation simply * takes/releases a reference count on all the cgroups referenced -- cgit v1.2.3 From db3b14978abc02041046ed8353f0899cb58ffffc Mon Sep 17 00:00:00 2001 From: Paul Menage Date: Fri, 25 Jul 2008 01:46:58 -0700 Subject: cgroup files: add write_string cgroup control file method This patch adds a write_string() method for cgroups control files. The semantics are that a buffer is copied from userspace to kernelspace and the handler function invoked on that buffer. The buffer is guaranteed to be nul-terminated, and no longer than max_write_len (defaulting to 64 bytes if unspecified). Later patches will convert existing raw file write handlers in control group subsystems to use this method. Signed-off-by: Paul Menage Cc: Paul Jackson Cc: Pavel Emelyanov Acked-by: Balbir Singh Acked-by: Serge Hallyn Cc: KAMEZAWA Hiroyuki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cgroup.c | 35 +++++++++++++++++++++++++++++++++++ 1 file changed, 35 insertions(+) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 70d083c6fb6b..3a99cc2df860 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1363,6 +1363,39 @@ static ssize_t cgroup_write_X64(struct cgroup *cgrp, struct cftype *cft, return retval; } +static ssize_t cgroup_write_string(struct cgroup *cgrp, struct cftype *cft, + struct file *file, + const char __user *userbuf, + size_t nbytes, loff_t *unused_ppos) +{ + char local_buffer[64]; + int retval = 0; + size_t max_bytes = cft->max_write_len; + char *buffer = local_buffer; + + if (!max_bytes) + max_bytes = sizeof(local_buffer) - 1; + if (nbytes >= max_bytes) + return -E2BIG; + /* Allocate a dynamic buffer if we need one */ + if (nbytes >= sizeof(local_buffer)) { + buffer = kmalloc(nbytes + 1, GFP_KERNEL); + if (buffer == NULL) + return -ENOMEM; + } + if (nbytes && copy_from_user(buffer, userbuf, nbytes)) + return -EFAULT; + + buffer[nbytes] = 0; /* nul-terminate */ + strstrip(buffer); + retval = cft->write_string(cgrp, cft, buffer); + if (!retval) + retval = nbytes; + if (buffer != local_buffer) + kfree(buffer); + return retval; +} + static ssize_t cgroup_common_file_write(struct cgroup *cgrp, struct cftype *cft, struct file *file, @@ -1440,6 +1473,8 @@ static ssize_t cgroup_file_write(struct file *file, const char __user *buf, return cft->write(cgrp, cft, file, buf, nbytes, ppos); if (cft->write_u64 || cft->write_s64) return cgroup_write_X64(cgrp, cft, file, buf, nbytes, ppos); + if (cft->write_string) + return cgroup_write_string(cgrp, cft, file, buf, nbytes, ppos); if (cft->trigger) { int ret = cft->trigger(cgrp, (unsigned int)cft->private); return ret ? ret : nbytes; -- cgit v1.2.3 From e788e066c651b1bbf4a927dc95395c1aa13be436 Mon Sep 17 00:00:00 2001 From: Paul Menage Date: Fri, 25 Jul 2008 01:46:59 -0700 Subject: cgroup files: move the release_agent file to use typed handlers Adds cgroup_release_agent_write() and cgroup_release_agent_show() methods to handle writing/reading the path to a cgroup hierarchy's release agent. As a result, cgroup_common_file_read() is now unnecessary. As part of the change, a previously-tolerated race in cgroup_release_agent() is avoided by copying the current release_agent_path prior to calling call_usermode_helper(). Signed-off-by: Paul Menage Cc: Paul Jackson Cc: Pavel Emelyanov Cc: Balbir Singh Acked-by: Serge Hallyn Cc: KAMEZAWA Hiroyuki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cgroup.c | 125 ++++++++++++++++++++++++++------------------------------ 1 file changed, 57 insertions(+), 68 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 3a99cc2df860..0120b5d67a73 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -89,11 +89,7 @@ struct cgroupfs_root { /* Hierarchy-specific flags */ unsigned long flags; - /* The path to use for release notifications. No locking - * between setting and use - so if userspace updates this - * while child cgroups exist, you could miss a - * notification. We ensure that it's always a valid - * NUL-terminated string */ + /* The path to use for release notifications. */ char release_agent_path[PATH_MAX]; }; @@ -1329,6 +1325,45 @@ enum cgroup_filetype { FILE_RELEASE_AGENT, }; +/** + * cgroup_lock_live_group - take cgroup_mutex and check that cgrp is alive. + * @cgrp: the cgroup to be checked for liveness + * + * Returns true (with lock held) on success, or false (with no lock + * held) on failure. + */ +int cgroup_lock_live_group(struct cgroup *cgrp) +{ + mutex_lock(&cgroup_mutex); + if (cgroup_is_removed(cgrp)) { + mutex_unlock(&cgroup_mutex); + return false; + } + return true; +} + +static int cgroup_release_agent_write(struct cgroup *cgrp, struct cftype *cft, + const char *buffer) +{ + BUILD_BUG_ON(sizeof(cgrp->root->release_agent_path) < PATH_MAX); + if (!cgroup_lock_live_group(cgrp)) + return -ENODEV; + strcpy(cgrp->root->release_agent_path, buffer); + mutex_unlock(&cgroup_mutex); + return 0; +} + +static int cgroup_release_agent_show(struct cgroup *cgrp, struct cftype *cft, + struct seq_file *seq) +{ + if (!cgroup_lock_live_group(cgrp)) + return -ENODEV; + seq_puts(seq, cgrp->root->release_agent_path); + seq_putc(seq, '\n'); + mutex_unlock(&cgroup_mutex); + return 0; +} + static ssize_t cgroup_write_X64(struct cgroup *cgrp, struct cftype *cft, struct file *file, const char __user *userbuf, @@ -1443,10 +1478,6 @@ static ssize_t cgroup_common_file_write(struct cgroup *cgrp, else clear_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags); break; - case FILE_RELEASE_AGENT: - BUILD_BUG_ON(sizeof(cgrp->root->release_agent_path) < PATH_MAX); - strcpy(cgrp->root->release_agent_path, buffer); - break; default: retval = -EINVAL; goto out2; @@ -1506,49 +1537,6 @@ static ssize_t cgroup_read_s64(struct cgroup *cgrp, struct cftype *cft, return simple_read_from_buffer(buf, nbytes, ppos, tmp, len); } -static ssize_t cgroup_common_file_read(struct cgroup *cgrp, - struct cftype *cft, - struct file *file, - char __user *buf, - size_t nbytes, loff_t *ppos) -{ - enum cgroup_filetype type = cft->private; - char *page; - ssize_t retval = 0; - char *s; - - if (!(page = (char *)__get_free_page(GFP_KERNEL))) - return -ENOMEM; - - s = page; - - switch (type) { - case FILE_RELEASE_AGENT: - { - struct cgroupfs_root *root; - size_t n; - mutex_lock(&cgroup_mutex); - root = cgrp->root; - n = strnlen(root->release_agent_path, - sizeof(root->release_agent_path)); - n = min(n, (size_t) PAGE_SIZE); - strncpy(s, root->release_agent_path, n); - mutex_unlock(&cgroup_mutex); - s += n; - break; - } - default: - retval = -EINVAL; - goto out; - } - *s++ = '\n'; - - retval = simple_read_from_buffer(buf, nbytes, ppos, page, s - page); -out: - free_page((unsigned long)page); - return retval; -} - static ssize_t cgroup_file_read(struct file *file, char __user *buf, size_t nbytes, loff_t *ppos) { @@ -1606,6 +1594,7 @@ int cgroup_seqfile_release(struct inode *inode, struct file *file) static struct file_operations cgroup_seqfile_operations = { .read = seq_read, + .write = cgroup_file_write, .llseek = seq_lseek, .release = cgroup_seqfile_release, }; @@ -2283,8 +2272,9 @@ static struct cftype files[] = { static struct cftype cft_release_agent = { .name = "release_agent", - .read = cgroup_common_file_read, - .write = cgroup_common_file_write, + .read_seq_string = cgroup_release_agent_show, + .write_string = cgroup_release_agent_write, + .max_write_len = PATH_MAX, .private = FILE_RELEASE_AGENT, }; @@ -3111,27 +3101,24 @@ static void cgroup_release_agent(struct work_struct *work) while (!list_empty(&release_list)) { char *argv[3], *envp[3]; int i; - char *pathbuf; + char *pathbuf = NULL, *agentbuf = NULL; struct cgroup *cgrp = list_entry(release_list.next, struct cgroup, release_list); list_del_init(&cgrp->release_list); spin_unlock(&release_list_lock); pathbuf = kmalloc(PAGE_SIZE, GFP_KERNEL); - if (!pathbuf) { - spin_lock(&release_list_lock); - continue; - } - - if (cgroup_path(cgrp, pathbuf, PAGE_SIZE) < 0) { - kfree(pathbuf); - spin_lock(&release_list_lock); - continue; - } + if (!pathbuf) + goto continue_free; + if (cgroup_path(cgrp, pathbuf, PAGE_SIZE) < 0) + goto continue_free; + agentbuf = kstrdup(cgrp->root->release_agent_path, GFP_KERNEL); + if (!agentbuf) + goto continue_free; i = 0; - argv[i++] = cgrp->root->release_agent_path; - argv[i++] = (char *)pathbuf; + argv[i++] = agentbuf; + argv[i++] = pathbuf; argv[i] = NULL; i = 0; @@ -3145,8 +3132,10 @@ static void cgroup_release_agent(struct work_struct *work) * be a slow process */ mutex_unlock(&cgroup_mutex); call_usermodehelper(argv[0], argv, envp, UMH_WAIT_EXEC); - kfree(pathbuf); mutex_lock(&cgroup_mutex); + continue_free: + kfree(pathbuf); + kfree(agentbuf); spin_lock(&release_list_lock); } spin_unlock(&release_list_lock); -- cgit v1.2.3 From 84eea842886ac35020be6043e04748ed22014359 Mon Sep 17 00:00:00 2001 From: Paul Menage Date: Fri, 25 Jul 2008 01:47:00 -0700 Subject: cgroups: misc cleanups to write_string patchset This patch contains cleanups suggested by reviewers for the recent write_string() patchset: - pair cgroup_lock_live_group() with cgroup_unlock() in cgroup.c for clarity, rather than directly unlocking cgroup_mutex. - make the return type of cgroup_lock_live_group() a bool - use a #define'd constant for the local buffer size in read/write functions Signed-off-by: Paul Menage Cc: Paul Jackson Cc: Pavel Emelyanov Cc: Balbir Singh Acked-by: Serge Hallyn Cc: KAMEZAWA Hiroyuki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cgroup.c | 21 ++++++++++++--------- 1 file changed, 12 insertions(+), 9 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 0120b5d67a73..a14122ecaa5e 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1329,10 +1329,10 @@ enum cgroup_filetype { * cgroup_lock_live_group - take cgroup_mutex and check that cgrp is alive. * @cgrp: the cgroup to be checked for liveness * - * Returns true (with lock held) on success, or false (with no lock - * held) on failure. + * On success, returns true; the lock should be later released with + * cgroup_unlock(). On failure returns false with no lock held. */ -int cgroup_lock_live_group(struct cgroup *cgrp) +bool cgroup_lock_live_group(struct cgroup *cgrp) { mutex_lock(&cgroup_mutex); if (cgroup_is_removed(cgrp)) { @@ -1349,7 +1349,7 @@ static int cgroup_release_agent_write(struct cgroup *cgrp, struct cftype *cft, if (!cgroup_lock_live_group(cgrp)) return -ENODEV; strcpy(cgrp->root->release_agent_path, buffer); - mutex_unlock(&cgroup_mutex); + cgroup_unlock(); return 0; } @@ -1360,16 +1360,19 @@ static int cgroup_release_agent_show(struct cgroup *cgrp, struct cftype *cft, return -ENODEV; seq_puts(seq, cgrp->root->release_agent_path); seq_putc(seq, '\n'); - mutex_unlock(&cgroup_mutex); + cgroup_unlock(); return 0; } +/* A buffer size big enough for numbers or short strings */ +#define CGROUP_LOCAL_BUFFER_SIZE 64 + static ssize_t cgroup_write_X64(struct cgroup *cgrp, struct cftype *cft, struct file *file, const char __user *userbuf, size_t nbytes, loff_t *unused_ppos) { - char buffer[64]; + char buffer[CGROUP_LOCAL_BUFFER_SIZE]; int retval = 0; char *end; @@ -1403,7 +1406,7 @@ static ssize_t cgroup_write_string(struct cgroup *cgrp, struct cftype *cft, const char __user *userbuf, size_t nbytes, loff_t *unused_ppos) { - char local_buffer[64]; + char local_buffer[CGROUP_LOCAL_BUFFER_SIZE]; int retval = 0; size_t max_bytes = cft->max_write_len; char *buffer = local_buffer; @@ -1518,7 +1521,7 @@ static ssize_t cgroup_read_u64(struct cgroup *cgrp, struct cftype *cft, char __user *buf, size_t nbytes, loff_t *ppos) { - char tmp[64]; + char tmp[CGROUP_LOCAL_BUFFER_SIZE]; u64 val = cft->read_u64(cgrp, cft); int len = sprintf(tmp, "%llu\n", (unsigned long long) val); @@ -1530,7 +1533,7 @@ static ssize_t cgroup_read_s64(struct cgroup *cgrp, struct cftype *cft, char __user *buf, size_t nbytes, loff_t *ppos) { - char tmp[64]; + char tmp[CGROUP_LOCAL_BUFFER_SIZE]; s64 val = cft->read_s64(cgrp, cft); int len = sprintf(tmp, "%lld\n", (long long) val); -- cgit v1.2.3 From 6379c106152388f7ea45d6dda63edda0e9181fc8 Mon Sep 17 00:00:00 2001 From: Paul Menage Date: Fri, 25 Jul 2008 01:47:01 -0700 Subject: cgroup files: move notify_on_release file to separate write handler This patch moves the write handler for the cgroups notify_on_release file into a separate handler. This handler requires no cgroups locking since it relies on atomic bitops for synchronization. Signed-off-by: Paul Menage Cc: Paul Jackson Cc: Pavel Emelyanov Cc: Balbir Singh Cc: Serge Hallyn Cc: KAMEZAWA Hiroyuki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cgroup.c | 21 +++++++++++++-------- 1 file changed, 13 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index a14122ecaa5e..d597d3015786 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1474,13 +1474,6 @@ static ssize_t cgroup_common_file_write(struct cgroup *cgrp, case FILE_TASKLIST: retval = attach_task_by_pid(cgrp, buffer); break; - case FILE_NOTIFY_ON_RELEASE: - clear_bit(CGRP_RELEASABLE, &cgrp->flags); - if (simple_strtoul(buffer, NULL, 10) != 0) - set_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags); - else - clear_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags); - break; default: retval = -EINVAL; goto out2; @@ -2252,6 +2245,18 @@ static u64 cgroup_read_notify_on_release(struct cgroup *cgrp, return notify_on_release(cgrp); } +static int cgroup_write_notify_on_release(struct cgroup *cgrp, + struct cftype *cft, + u64 val) +{ + clear_bit(CGRP_RELEASABLE, &cgrp->flags); + if (val) + set_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags); + else + clear_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags); + return 0; +} + /* * for the common functions, 'private' gives the type of file */ @@ -2268,7 +2273,7 @@ static struct cftype files[] = { { .name = "notify_on_release", .read_u64 = cgroup_read_notify_on_release, - .write = cgroup_common_file_write, + .write_u64 = cgroup_write_notify_on_release, .private = FILE_NOTIFY_ON_RELEASE, }, }; -- cgit v1.2.3 From af351026aafc8da16518a02b41c66d3e0c1cdef4 Mon Sep 17 00:00:00 2001 From: Paul Menage Date: Fri, 25 Jul 2008 01:47:01 -0700 Subject: cgroup files: turn attach_task_by_pid directly into a cgroup write handler This patch changes attach_task_by_pid() to take a u64 rather than a string; as a result it can be called directly as a control groups write_u64 handler, and cgroup_common_file_write() can be removed. Signed-off-by: Paul Menage Cc: Paul Jackson Cc: Pavel Emelyanov Cc: Balbir Singh Cc: Serge Hallyn Cc: KAMEZAWA Hiroyuki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cgroup.c | 80 ++++++++++----------------------------------------------- 1 file changed, 14 insertions(+), 66 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index d597d3015786..86b71e714e13 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -504,10 +504,6 @@ static struct css_set *find_css_set( * knows that the cgroup won't be removed, as cgroup_rmdir() * needs that mutex. * - * The cgroup_common_file_write handler for operations that modify - * the cgroup hierarchy holds cgroup_mutex across the entire operation, - * single threading all such cgroup modifications across the system. - * * The fork and exit callbacks cgroup_fork() and cgroup_exit(), don't * (usually) take cgroup_mutex. These are the two most performance * critical pieces of code here. The exception occurs on cgroup_exit(), @@ -1279,18 +1275,14 @@ int cgroup_attach_task(struct cgroup *cgrp, struct task_struct *tsk) } /* - * Attach task with pid 'pid' to cgroup 'cgrp'. Call with - * cgroup_mutex, may take task_lock of task + * Attach task with pid 'pid' to cgroup 'cgrp'. Call with cgroup_mutex + * held. May take task_lock of task */ -static int attach_task_by_pid(struct cgroup *cgrp, char *pidbuf) +static int attach_task_by_pid(struct cgroup *cgrp, u64 pid) { - pid_t pid; struct task_struct *tsk; int ret; - if (sscanf(pidbuf, "%d", &pid) != 1) - return -EIO; - if (pid) { rcu_read_lock(); tsk = find_task_by_vpid(pid); @@ -1316,6 +1308,16 @@ static int attach_task_by_pid(struct cgroup *cgrp, char *pidbuf) return ret; } +static int cgroup_tasks_write(struct cgroup *cgrp, struct cftype *cft, u64 pid) +{ + int ret; + if (!cgroup_lock_live_group(cgrp)) + return -ENODEV; + ret = attach_task_by_pid(cgrp, pid); + cgroup_unlock(); + return ret; +} + /* The various types of files and directories in a cgroup file system */ enum cgroup_filetype { FILE_ROOT, @@ -1434,60 +1436,6 @@ static ssize_t cgroup_write_string(struct cgroup *cgrp, struct cftype *cft, return retval; } -static ssize_t cgroup_common_file_write(struct cgroup *cgrp, - struct cftype *cft, - struct file *file, - const char __user *userbuf, - size_t nbytes, loff_t *unused_ppos) -{ - enum cgroup_filetype type = cft->private; - char *buffer; - int retval = 0; - - if (nbytes >= PATH_MAX) - return -E2BIG; - - /* +1 for nul-terminator */ - buffer = kmalloc(nbytes + 1, GFP_KERNEL); - if (buffer == NULL) - return -ENOMEM; - - if (copy_from_user(buffer, userbuf, nbytes)) { - retval = -EFAULT; - goto out1; - } - buffer[nbytes] = 0; /* nul-terminate */ - strstrip(buffer); /* strip -just- trailing whitespace */ - - mutex_lock(&cgroup_mutex); - - /* - * This was already checked for in cgroup_file_write(), but - * check again now we're holding cgroup_mutex. - */ - if (cgroup_is_removed(cgrp)) { - retval = -ENODEV; - goto out2; - } - - switch (type) { - case FILE_TASKLIST: - retval = attach_task_by_pid(cgrp, buffer); - break; - default: - retval = -EINVAL; - goto out2; - } - - if (retval == 0) - retval = nbytes; -out2: - mutex_unlock(&cgroup_mutex); -out1: - kfree(buffer); - return retval; -} - static ssize_t cgroup_file_write(struct file *file, const char __user *buf, size_t nbytes, loff_t *ppos) { @@ -2265,7 +2213,7 @@ static struct cftype files[] = { .name = "tasks", .open = cgroup_tasks_open, .read = cgroup_tasks_read, - .write = cgroup_common_file_write, + .write_u64 = cgroup_tasks_write, .release = cgroup_tasks_release, .private = FILE_TASKLIST, }, -- cgit v1.2.3 From e37123953292146445c8629b3950d0513fd10ae2 Mon Sep 17 00:00:00 2001 From: Paul Menage Date: Fri, 25 Jul 2008 01:47:02 -0700 Subject: cgroup files: remove cpuset_common_file_write() This patch tweaks the signatures of the update_cpumask() and update_nodemask() functions so that they can be called directly as handlers for the new cgroups write_string() method. This allows cpuset_common_file_write() to be removed. Signed-off-by: Paul Menage Cc: Paul Jackson Cc: Pavel Emelyanov Cc: Balbir Singh Cc: Serge Hallyn Cc: KAMEZAWA Hiroyuki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cpuset.c | 109 ++++++++++++++++++-------------------------------------- 1 file changed, 35 insertions(+), 74 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index d5738910c34c..276ce7e4f1ab 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -227,10 +227,6 @@ static struct cpuset top_cpuset = { * The task_struct fields mems_allowed and mems_generation may only * be accessed in the context of that task, so require no locks. * - * The cpuset_common_file_write handler for operations that modify - * the cpuset hierarchy holds cgroup_mutex across the entire operation, - * single threading all such cpuset modifications across the system. - * * The cpuset_common_file_read() handlers only hold callback_mutex across * small pieces of code, such as when reading out possibly multi-word * cpumasks and nodemasks. @@ -772,7 +768,7 @@ static void cpuset_change_cpumask(struct task_struct *tsk, * @cs: the cpuset to consider * @buf: buffer of cpu numbers written to this cpuset */ -static int update_cpumask(struct cpuset *cs, char *buf) +static int update_cpumask(struct cpuset *cs, const char *buf) { struct cpuset trialcs; struct cgroup_scanner scan; @@ -792,7 +788,6 @@ static int update_cpumask(struct cpuset *cs, char *buf) * that parsing. The validate_change() call ensures that cpusets * with tasks have cpus. */ - buf = strstrip(buf); if (!*buf) { cpus_clear(trialcs.cpus_allowed); } else { @@ -902,7 +897,7 @@ static void cpuset_migrate_mm(struct mm_struct *mm, const nodemask_t *from, static void *cpuset_being_rebound; -static int update_nodemask(struct cpuset *cs, char *buf) +static int update_nodemask(struct cpuset *cs, const char *buf) { struct cpuset trialcs; nodemask_t oldmem; @@ -929,7 +924,6 @@ static int update_nodemask(struct cpuset *cs, char *buf) * that parsing. The validate_change() call ensures that cpusets * with tasks have memory. */ - buf = strstrip(buf); if (!*buf) { nodes_clear(trialcs.mems_allowed); } else { @@ -1256,72 +1250,14 @@ typedef enum { FILE_SPREAD_SLAB, } cpuset_filetype_t; -static ssize_t cpuset_common_file_write(struct cgroup *cont, - struct cftype *cft, - struct file *file, - const char __user *userbuf, - size_t nbytes, loff_t *unused_ppos) -{ - struct cpuset *cs = cgroup_cs(cont); - cpuset_filetype_t type = cft->private; - char *buffer; - int retval = 0; - - /* Crude upper limit on largest legitimate cpulist user might write. */ - if (nbytes > 100U + 6 * max(NR_CPUS, MAX_NUMNODES)) - return -E2BIG; - - /* +1 for nul-terminator */ - buffer = kmalloc(nbytes + 1, GFP_KERNEL); - if (!buffer) - return -ENOMEM; - - if (copy_from_user(buffer, userbuf, nbytes)) { - retval = -EFAULT; - goto out1; - } - buffer[nbytes] = 0; /* nul-terminate */ - - cgroup_lock(); - - if (cgroup_is_removed(cont)) { - retval = -ENODEV; - goto out2; - } - - switch (type) { - case FILE_CPULIST: - retval = update_cpumask(cs, buffer); - break; - case FILE_MEMLIST: - retval = update_nodemask(cs, buffer); - break; - default: - retval = -EINVAL; - goto out2; - } - - if (retval == 0) - retval = nbytes; -out2: - cgroup_unlock(); -out1: - kfree(buffer); - return retval; -} - static int cpuset_write_u64(struct cgroup *cgrp, struct cftype *cft, u64 val) { int retval = 0; struct cpuset *cs = cgroup_cs(cgrp); cpuset_filetype_t type = cft->private; - cgroup_lock(); - - if (cgroup_is_removed(cgrp)) { - cgroup_unlock(); + if (!cgroup_lock_live_group(cgrp)) return -ENODEV; - } switch (type) { case FILE_CPU_EXCLUSIVE: @@ -1367,12 +1303,9 @@ static int cpuset_write_s64(struct cgroup *cgrp, struct cftype *cft, s64 val) struct cpuset *cs = cgroup_cs(cgrp); cpuset_filetype_t type = cft->private; - cgroup_lock(); - - if (cgroup_is_removed(cgrp)) { - cgroup_unlock(); + if (!cgroup_lock_live_group(cgrp)) return -ENODEV; - } + switch (type) { case FILE_SCHED_RELAX_DOMAIN_LEVEL: retval = update_relax_domain_level(cs, val); @@ -1385,6 +1318,32 @@ static int cpuset_write_s64(struct cgroup *cgrp, struct cftype *cft, s64 val) return retval; } +/* + * Common handling for a write to a "cpus" or "mems" file. + */ +static int cpuset_write_resmask(struct cgroup *cgrp, struct cftype *cft, + const char *buf) +{ + int retval = 0; + + if (!cgroup_lock_live_group(cgrp)) + return -ENODEV; + + switch (cft->private) { + case FILE_CPULIST: + retval = update_cpumask(cgroup_cs(cgrp), buf); + break; + case FILE_MEMLIST: + retval = update_nodemask(cgroup_cs(cgrp), buf); + break; + default: + retval = -EINVAL; + break; + } + cgroup_unlock(); + return retval; +} + /* * These ascii lists should be read in a single call, by using a user * buffer large enough to hold the entire map. If read in smaller @@ -1504,14 +1463,16 @@ static struct cftype files[] = { { .name = "cpus", .read = cpuset_common_file_read, - .write = cpuset_common_file_write, + .write_string = cpuset_write_resmask, + .max_write_len = (100U + 6 * NR_CPUS), .private = FILE_CPULIST, }, { .name = "mems", .read = cpuset_common_file_read, - .write = cpuset_common_file_write, + .write_string = cpuset_write_resmask, + .max_write_len = (100U + 6 * MAX_NUMNODES), .private = FILE_MEMLIST, }, -- cgit v1.2.3 From 856c13aa1ff6136c1968414fdea5938ea9d5ebf2 Mon Sep 17 00:00:00 2001 From: Paul Menage Date: Fri, 25 Jul 2008 01:47:04 -0700 Subject: cgroup files: convert res_counter_write() to be a cgroups write_string() handler Currently res_counter_write() is a raw file handler even though it's ultimately taking a number, since in some cases it wants to pre-process the string when converting it to a number. This patch converts res_counter_write() from a raw file handler to a write_string() handler; this allows some of the boilerplate copying/locking/checking to be removed, and simplies the cleanup path, since these functions are now performed by the cgroups framework. [lizf@cn.fujitsu.com: build fix] Signed-off-by: Paul Menage Cc: Paul Jackson Cc: Pavel Emelyanov Cc: Balbir Singh Cc: Serge Hallyn Cc: KAMEZAWA Hiroyuki Signed-off-by: Li Zefan Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/res_counter.c | 48 +++++++++++++++++++++--------------------------- 1 file changed, 21 insertions(+), 27 deletions(-) (limited to 'kernel') diff --git a/kernel/res_counter.c b/kernel/res_counter.c index d3c61b4ebef2..f275c8eca772 100644 --- a/kernel/res_counter.c +++ b/kernel/res_counter.c @@ -13,6 +13,7 @@ #include #include #include +#include void res_counter_init(struct res_counter *counter) { @@ -102,44 +103,37 @@ u64 res_counter_read_u64(struct res_counter *counter, int member) return *res_counter_member(counter, member); } -ssize_t res_counter_write(struct res_counter *counter, int member, - const char __user *userbuf, size_t nbytes, loff_t *pos, - int (*write_strategy)(char *st_buf, unsigned long long *val)) +int res_counter_memparse_write_strategy(const char *buf, + unsigned long long *res) { - int ret; - char *buf, *end; - unsigned long flags; - unsigned long long tmp, *val; - - buf = kmalloc(nbytes + 1, GFP_KERNEL); - ret = -ENOMEM; - if (buf == NULL) - goto out; + char *end; + /* FIXME - make memparse() take const char* args */ + *res = memparse((char *)buf, &end); + if (*end != '\0') + return -EINVAL; - buf[nbytes] = '\0'; - ret = -EFAULT; - if (copy_from_user(buf, userbuf, nbytes)) - goto out_free; + *res = PAGE_ALIGN(*res); + return 0; +} - ret = -EINVAL; +int res_counter_write(struct res_counter *counter, int member, + const char *buf, write_strategy_fn write_strategy) +{ + char *end; + unsigned long flags; + unsigned long long tmp, *val; - strstrip(buf); if (write_strategy) { - if (write_strategy(buf, &tmp)) { - goto out_free; - } + if (write_strategy(buf, &tmp)) + return -EINVAL; } else { tmp = simple_strtoull(buf, &end, 10); if (*end != '\0') - goto out_free; + return -EINVAL; } spin_lock_irqsave(&counter->lock, flags); val = res_counter_member(counter, member); *val = tmp; spin_unlock_irqrestore(&counter->lock, flags); - ret = nbytes; -out_free: - kfree(buf); -out: - return ret; + return 0; } -- cgit v1.2.3 From e885dcde75685e09f23cffae1f6d5169c105b8a0 Mon Sep 17 00:00:00 2001 From: "Serge E. Hallyn" Date: Fri, 25 Jul 2008 01:47:06 -0700 Subject: cgroup_clone: use pid of newly created task for new cgroup cgroup_clone creates a new cgroup with the pid of the task. This works correctly for unshare, but for clone cgroup_clone is called from copy_namespaces inside copy_process, which happens before the new pid is created. As a result, the new cgroup was created with current's pid. This patch: 1. Moves the call inside copy_process to after the new pid is created 2. Passes the struct pid into ns_cgroup_clone (as it is not yet attached to the task) 3. Passes a name from ns_cgroup_clone() into cgroup_clone() so as to keep cgroup_clone() itself simpler 4. Uses pid_vnr() to get the process id value, so that the pid used to name the new cgroup is always the pid as it would be known to the task which did the cloning or unsharing. I think that is the most intuitive thing to do. This way, task t1 does clone(CLONE_NEWPID) to get t2, which does clone(CLONE_NEWPID) to get t3, then the cgroup for t3 will be named for the pid by which t2 knows t3. (Thanks to Dan Smith for finding the main bug) Changelog: June 11: Incorporate Paul Menage's feedback: don't pass NULL to ns_cgroup_clone from unshare, and reduce patch size by using 'nodename' in cgroup_clone. June 10: Original version [akpm@linux-foundation.org: build fix] [akpm@linux-foundation.org: coding-style fixes] Signed-off-by: Serge Hallyn Acked-by: Paul Menage Tested-by: Dan Smith Cc: Balbir Singh Cc: KAMEZAWA Hiroyuki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cgroup.c | 7 +++---- kernel/fork.c | 6 ++++++ kernel/ns_cgroup.c | 8 ++++++-- kernel/nsproxy.c | 8 +------- 4 files changed, 16 insertions(+), 13 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 86b71e714e13..66ec9fd21e0c 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -2848,16 +2848,17 @@ void cgroup_exit(struct task_struct *tsk, int run_callbacks) * cgroup_clone - clone the cgroup the given subsystem is attached to * @tsk: the task to be moved * @subsys: the given subsystem + * @nodename: the name for the new cgroup * * Duplicate the current cgroup in the hierarchy that the given * subsystem is attached to, and move this task into the new * child. */ -int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *subsys) +int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *subsys, + char *nodename) { struct dentry *dentry; int ret = 0; - char nodename[MAX_CGROUP_TYPE_NAMELEN]; struct cgroup *parent, *child; struct inode *inode; struct css_set *cg; @@ -2882,8 +2883,6 @@ int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *subsys) cg = tsk->cgroups; parent = task_cgroup(tsk, subsys->subsys_id); - snprintf(nodename, MAX_CGROUP_TYPE_NAMELEN, "%d", tsk->pid); - /* Pin the hierarchy */ atomic_inc(&parent->root->sb->s_active); diff --git a/kernel/fork.c b/kernel/fork.c index 5a5d6fef341d..228f80c9155a 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1107,6 +1107,12 @@ static struct task_struct *copy_process(unsigned long clone_flags, if (clone_flags & CLONE_THREAD) p->tgid = current->tgid; + if (current->nsproxy != p->nsproxy) { + retval = ns_cgroup_clone(p, pid); + if (retval) + goto bad_fork_free_pid; + } + p->set_child_tid = (clone_flags & CLONE_CHILD_SETTID) ? child_tidptr : NULL; /* * Clear TID on mm_release()? diff --git a/kernel/ns_cgroup.c b/kernel/ns_cgroup.c index 48d7ed6fc3a4..43c2111cd54d 100644 --- a/kernel/ns_cgroup.c +++ b/kernel/ns_cgroup.c @@ -7,6 +7,7 @@ #include #include #include +#include #include #include @@ -24,9 +25,12 @@ static inline struct ns_cgroup *cgroup_to_ns( struct ns_cgroup, css); } -int ns_cgroup_clone(struct task_struct *task) +int ns_cgroup_clone(struct task_struct *task, struct pid *pid) { - return cgroup_clone(task, &ns_subsys); + char name[PROC_NUMBUF]; + + snprintf(name, PROC_NUMBUF, "%d", pid_vnr(pid)); + return cgroup_clone(task, &ns_subsys, name); } /* diff --git a/kernel/nsproxy.c b/kernel/nsproxy.c index adc785146a1c..21575fc46d05 100644 --- a/kernel/nsproxy.c +++ b/kernel/nsproxy.c @@ -157,12 +157,6 @@ int copy_namespaces(unsigned long flags, struct task_struct *tsk) goto out; } - err = ns_cgroup_clone(tsk); - if (err) { - put_nsproxy(new_ns); - goto out; - } - tsk->nsproxy = new_ns; out: @@ -209,7 +203,7 @@ int unshare_nsproxy_namespaces(unsigned long unshare_flags, goto out; } - err = ns_cgroup_clone(current); + err = ns_cgroup_clone(current, task_pid(current)); if (err) put_nsproxy(*new_nsp); -- cgit v1.2.3 From 0b2f630a28d53b5a2082a5275bc3334b10373508 Mon Sep 17 00:00:00 2001 From: Miao Xie Date: Fri, 25 Jul 2008 01:47:21 -0700 Subject: cpusets: restructure the function update_cpumask() and update_nodemask() Extract two functions from update_cpumask() and update_nodemask().They will be used later for updating tasks' cpus_allowed and mems_allowed after CPU/NODE offline/online. [lizf@cn.fujitsu.com: build fix] Signed-off-by: Miao Xie Acked-by: Paul Jackson Cc: David Rientjes Cc: Li Zefan Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cpuset.c | 181 ++++++++++++++++++++++++++++++++++---------------------- 1 file changed, 109 insertions(+), 72 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 276ce7e4f1ab..7326d51eefe1 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -763,6 +763,37 @@ static void cpuset_change_cpumask(struct task_struct *tsk, set_cpus_allowed_ptr(tsk, &((cgroup_cs(scan->cg))->cpus_allowed)); } +/** + * update_tasks_cpumask - Update the cpumasks of tasks in the cpuset. + * @cs: the cpuset in which each task's cpus_allowed mask needs to be changed + * + * Called with cgroup_mutex held + * + * The cgroup_scan_tasks() function will scan all the tasks in a cgroup, + * calling callback functions for each. + * + * Return 0 if successful, -errno if not. + */ +static int update_tasks_cpumask(struct cpuset *cs) +{ + struct cgroup_scanner scan; + struct ptr_heap heap; + int retval; + + retval = heap_init(&heap, PAGE_SIZE, GFP_KERNEL, &started_after); + if (retval) + return retval; + + scan.cg = cs->css.cgroup; + scan.test_task = cpuset_test_cpumask; + scan.process_task = cpuset_change_cpumask; + scan.heap = &heap; + retval = cgroup_scan_tasks(&scan); + + heap_free(&heap); + return retval; +} + /** * update_cpumask - update the cpus_allowed mask of a cpuset and all tasks in it * @cs: the cpuset to consider @@ -771,8 +802,6 @@ static void cpuset_change_cpumask(struct task_struct *tsk, static int update_cpumask(struct cpuset *cs, const char *buf) { struct cpuset trialcs; - struct cgroup_scanner scan; - struct ptr_heap heap; int retval; int is_load_balanced; @@ -806,10 +835,6 @@ static int update_cpumask(struct cpuset *cs, const char *buf) if (cpus_equal(cs->cpus_allowed, trialcs.cpus_allowed)) return 0; - retval = heap_init(&heap, PAGE_SIZE, GFP_KERNEL, &started_after); - if (retval) - return retval; - is_load_balanced = is_sched_load_balance(&trialcs); mutex_lock(&callback_mutex); @@ -820,12 +845,9 @@ static int update_cpumask(struct cpuset *cs, const char *buf) * Scan tasks in the cpuset, and update the cpumasks of any * that need an update. */ - scan.cg = cs->css.cgroup; - scan.test_task = cpuset_test_cpumask; - scan.process_task = cpuset_change_cpumask; - scan.heap = &heap; - cgroup_scan_tasks(&scan); - heap_free(&heap); + retval = update_tasks_cpumask(cs); + if (retval < 0) + return retval; if (is_load_balanced) rebuild_sched_domains(); @@ -881,73 +903,25 @@ static void cpuset_migrate_mm(struct mm_struct *mm, const nodemask_t *from, mutex_unlock(&callback_mutex); } -/* - * Handle user request to change the 'mems' memory placement - * of a cpuset. Needs to validate the request, update the - * cpusets mems_allowed and mems_generation, and for each - * task in the cpuset, rebind any vma mempolicies and if - * the cpuset is marked 'memory_migrate', migrate the tasks - * pages to the new memory. - * - * Call with cgroup_mutex held. May take callback_mutex during call. - * Will take tasklist_lock, scan tasklist for tasks in cpuset cs, - * lock each such tasks mm->mmap_sem, scan its vma's and rebind - * their mempolicies to the cpusets new mems_allowed. - */ - static void *cpuset_being_rebound; -static int update_nodemask(struct cpuset *cs, const char *buf) +/** + * update_tasks_nodemask - Update the nodemasks of tasks in the cpuset. + * @cs: the cpuset in which each task's mems_allowed mask needs to be changed + * @oldmem: old mems_allowed of cpuset cs + * + * Called with cgroup_mutex held + * Return 0 if successful, -errno if not. + */ +static int update_tasks_nodemask(struct cpuset *cs, const nodemask_t *oldmem) { - struct cpuset trialcs; - nodemask_t oldmem; struct task_struct *p; struct mm_struct **mmarray; int i, n, ntasks; int migrate; int fudge; - int retval; struct cgroup_iter it; - - /* - * top_cpuset.mems_allowed tracks node_stats[N_HIGH_MEMORY]; - * it's read-only - */ - if (cs == &top_cpuset) - return -EACCES; - - trialcs = *cs; - - /* - * An empty mems_allowed is ok iff there are no tasks in the cpuset. - * Since nodelist_parse() fails on an empty mask, we special case - * that parsing. The validate_change() call ensures that cpusets - * with tasks have memory. - */ - if (!*buf) { - nodes_clear(trialcs.mems_allowed); - } else { - retval = nodelist_parse(buf, trialcs.mems_allowed); - if (retval < 0) - goto done; - - if (!nodes_subset(trialcs.mems_allowed, - node_states[N_HIGH_MEMORY])) - return -EINVAL; - } - oldmem = cs->mems_allowed; - if (nodes_equal(oldmem, trialcs.mems_allowed)) { - retval = 0; /* Too easy - nothing to do */ - goto done; - } - retval = validate_change(cs, &trialcs); - if (retval < 0) - goto done; - - mutex_lock(&callback_mutex); - cs->mems_allowed = trialcs.mems_allowed; - cs->mems_generation = cpuset_mems_generation++; - mutex_unlock(&callback_mutex); + int retval; cpuset_being_rebound = cs; /* causes mpol_dup() rebind */ @@ -1014,7 +988,7 @@ static int update_nodemask(struct cpuset *cs, const char *buf) mpol_rebind_mm(mm, &cs->mems_allowed); if (migrate) - cpuset_migrate_mm(mm, &oldmem, &cs->mems_allowed); + cpuset_migrate_mm(mm, oldmem, &cs->mems_allowed); mmput(mm); } @@ -1026,6 +1000,70 @@ done: return retval; } +/* + * Handle user request to change the 'mems' memory placement + * of a cpuset. Needs to validate the request, update the + * cpusets mems_allowed and mems_generation, and for each + * task in the cpuset, rebind any vma mempolicies and if + * the cpuset is marked 'memory_migrate', migrate the tasks + * pages to the new memory. + * + * Call with cgroup_mutex held. May take callback_mutex during call. + * Will take tasklist_lock, scan tasklist for tasks in cpuset cs, + * lock each such tasks mm->mmap_sem, scan its vma's and rebind + * their mempolicies to the cpusets new mems_allowed. + */ +static int update_nodemask(struct cpuset *cs, const char *buf) +{ + struct cpuset trialcs; + nodemask_t oldmem; + int retval; + + /* + * top_cpuset.mems_allowed tracks node_stats[N_HIGH_MEMORY]; + * it's read-only + */ + if (cs == &top_cpuset) + return -EACCES; + + trialcs = *cs; + + /* + * An empty mems_allowed is ok iff there are no tasks in the cpuset. + * Since nodelist_parse() fails on an empty mask, we special case + * that parsing. The validate_change() call ensures that cpusets + * with tasks have memory. + */ + if (!*buf) { + nodes_clear(trialcs.mems_allowed); + } else { + retval = nodelist_parse(buf, trialcs.mems_allowed); + if (retval < 0) + goto done; + + if (!nodes_subset(trialcs.mems_allowed, + node_states[N_HIGH_MEMORY])) + return -EINVAL; + } + oldmem = cs->mems_allowed; + if (nodes_equal(oldmem, trialcs.mems_allowed)) { + retval = 0; /* Too easy - nothing to do */ + goto done; + } + retval = validate_change(cs, &trialcs); + if (retval < 0) + goto done; + + mutex_lock(&callback_mutex); + cs->mems_allowed = trialcs.mems_allowed; + cs->mems_generation = cpuset_mems_generation++; + mutex_unlock(&callback_mutex); + + retval = update_tasks_nodemask(cs, &oldmem); +done: + return retval; +} + int current_cpuset_is_being_rebound(void) { return task_cs(current) == cpuset_being_rebound; @@ -1935,7 +1973,6 @@ void __init cpuset_init_smp(void) } /** - * cpuset_cpus_allowed - return cpus_allowed mask from a tasks cpuset. * @tsk: pointer to task_struct from which to obtain cpuset->cpus_allowed. * @pmask: pointer to cpumask_t variable to receive cpus_allowed set. -- cgit v1.2.3 From f9b4fb8dabf38fb456c97f01aace07cb6e7c1723 Mon Sep 17 00:00:00 2001 From: Miao Xie Date: Fri, 25 Jul 2008 01:47:22 -0700 Subject: cpusets: update task's cpus_allowed and mems_allowed after CPU/NODE offline/online The bug is that a task may run on the cpu/node which is not in its cpuset.cpus/ cpuset.mems. It can be reproduced by the following commands: ----------------------------------- # mkdir /dev/cpuset # mount -t cpuset xxx /dev/cpuset # mkdir /dev/cpuset/0 # echo 0-1 > /dev/cpuset/0/cpus # echo 0 > /dev/cpuset/0/mems # echo $$ > /dev/cpuset/0/tasks # echo 0 > /sys/devices/system/cpu/cpu1/online # echo 1 > /sys/devices/system/cpu/cpu1/online ----------------------------------- There is only CPU0 in cpuset.cpus, but the task in this cpuset runs on both CPU0 and CPU1. It is because the task's cpu_allowed didn't get updated after we did CPU offline/online manipulation. Similar for mem_allowed. This patch fixes this bug expect for root cpuset. Because there is a problem about root cpuset, in that whether it is necessary to update all the tasks in root cpuset or not after cpu/node offline/online. If updating, some kernel threads which is bound into a specified cpu will be unbound. If not updating, there is a bug in root cpuset. This bug is also caused by offline/online manipulation. For example, there is a dual-cpu machine. we create a sub cpuset in root cpuset and assign 1 to its cpus. And then we attach some tasks into this sub cpuset. After this, we offline CPU1. Now, the tasks in this new cpuset are moved into root cpuset automatically because there is no cpu in sub cpuset. Then we online CPU1, we find all the tasks which doesn't belong to root cpuset originally just run on CPU0. Maybe we need to add a flag in the task_struct to mark which task can't be unbound? Signed-off-by: Miao Xie Acked-by: Paul Jackson Cc: Li Zefan Cc: Paul Jackson Cc: Paul Menage Cc: David Rientjes Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cpuset.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 7326d51eefe1..6eae6639e851 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -1851,6 +1851,7 @@ static void scan_for_empty_cpusets(const struct cpuset *root) struct cpuset *child; /* scans child cpusets of cp */ struct list_head queue; struct cgroup *cont; + nodemask_t oldmems; INIT_LIST_HEAD(&queue); @@ -1870,6 +1871,8 @@ static void scan_for_empty_cpusets(const struct cpuset *root) nodes_subset(cp->mems_allowed, node_states[N_HIGH_MEMORY])) continue; + oldmems = cp->mems_allowed; + /* Remove offline cpus and mems from this cpuset. */ mutex_lock(&callback_mutex); cpus_and(cp->cpus_allowed, cp->cpus_allowed, cpu_online_map); @@ -1881,6 +1884,10 @@ static void scan_for_empty_cpusets(const struct cpuset *root) if (cpus_empty(cp->cpus_allowed) || nodes_empty(cp->mems_allowed)) remove_tasks_in_empty_cpuset(cp); + else { + update_tasks_cpumask(cp); + update_tasks_nodemask(cp, &oldmems); + } } } -- cgit v1.2.3 From c372e817afc629fea9ff6321313325ed0b4a855b Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Fri, 25 Jul 2008 01:47:23 -0700 Subject: cpuset: avoid unnecessary sched domains rebuilding When changing 'sched_relax_domain_level', don't rebuild sched domains if 'cpus' is empty or 'sched_load_balance' is not set. Also make the comments of rebuild_sched_domains() more readable. Signed-off-by: Li Zefan Cc: Hidetoshi Seto Cc: Paul Jackson Cc: Paul Menage Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cpuset.c | 18 ++++++++++++------ 1 file changed, 12 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 6eae6639e851..60d2c4702c6c 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -496,11 +496,16 @@ update_domain_attr(struct sched_domain_attr *dattr, struct cpuset *c) /* * rebuild_sched_domains() * - * If the flag 'sched_load_balance' of any cpuset with non-empty - * 'cpus' changes, or if the 'cpus' allowed changes in any cpuset - * which has that flag enabled, or if any cpuset with a non-empty - * 'cpus' is removed, then call this routine to rebuild the - * scheduler's dynamic sched domains. + * This routine will be called to rebuild the scheduler's dynamic + * sched domains: + * - if the flag 'sched_load_balance' of any cpuset with non-empty + * 'cpus' changes, + * - or if the 'cpus' allowed changes in any cpuset which has that + * flag enabled, + * - or if the 'sched_relax_domain_level' of any cpuset which has + * that flag enabled and with non-empty 'cpus' changes, + * - or if any cpuset with non-empty 'cpus' is removed, + * - or if a cpu gets offlined. * * This routine builds a partial partition of the systems CPUs * (the set of non-overlappping cpumask_t's in the array 'part' @@ -1076,7 +1081,8 @@ static int update_relax_domain_level(struct cpuset *cs, s64 val) if (val != cs->relax_domain_level) { cs->relax_domain_level = val; - rebuild_sched_domains(); + if (!cpus_empty(cs->cpus_allowed) && is_sched_load_balance(cs)) + rebuild_sched_domains(); } return 0; -- cgit v1.2.3 From 489a5393a20dcbf91104052120eb2eff8791b61b Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Fri, 25 Jul 2008 01:47:23 -0700 Subject: cpuset: don't pass empty cpumasks to partition_sched_domains() I create lots of empty cpusets(empty cpumasks) and turn off the "sched_load_balance" in top cpuset. I found that all these empty cpumasks are passed to partition_sched_domains() in rebuild_sched_domains(), it's very time-consuming for partition_sched_domains() and it's not need. It also reduce memory consumed and some works in rebuild_sched_domains() too. Signed-off-by: Lai Jiangshan Acked-by: Paul Menage Cc: Paul Jackson Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cpuset.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 60d2c4702c6c..531b235e546f 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -610,8 +610,13 @@ void rebuild_sched_domains(void) while (__kfifo_get(q, (void *)&cp, sizeof(cp))) { struct cgroup *cont; struct cpuset *child; /* scans child cpusets of cp */ + + if (cpus_empty(cp->cpus_allowed)) + continue; + if (is_sched_load_balance(cp)) csa[csn++] = cp; + list_for_each_entry(cont, &cp->css.cgroup->children, sibling) { child = cgroup_cs(cont); __kfifo_put(q, (void *)&child, sizeof(cp)); -- cgit v1.2.3 From 02412483777651a26b19a75e49c2a451a174ca9c Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Fri, 25 Jul 2008 01:47:24 -0700 Subject: cpuset: code-cleanup for started_after cgroup(cgroup_scan_tasks) will initialize heap->gt for us. This patch removes started_after() and its helper-function. Signed-off-by: Lai Jiangshan Acked-by: Paul Menage Cc: Paul Jackson Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cpuset.c | 37 ++++++------------------------------- 1 file changed, 6 insertions(+), 31 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 531b235e546f..ebbc9b082e48 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -709,36 +709,6 @@ done: /* Don't kfree(dattr) -- partition_sched_domains() does that. */ } -static inline int started_after_time(struct task_struct *t1, - struct timespec *time, - struct task_struct *t2) -{ - int start_diff = timespec_compare(&t1->start_time, time); - if (start_diff > 0) { - return 1; - } else if (start_diff < 0) { - return 0; - } else { - /* - * Arbitrarily, if two processes started at the same - * time, we'll say that the lower pointer value - * started first. Note that t2 may have exited by now - * so this may not be a valid pointer any longer, but - * that's fine - it still serves to distinguish - * between two tasks started (effectively) - * simultaneously. - */ - return t1 > t2; - } -} - -static inline int started_after(void *p1, void *p2) -{ - struct task_struct *t1 = p1; - struct task_struct *t2 = p2; - return started_after_time(t1, &t2->start_time, t2); -} - /** * cpuset_test_cpumask - test a task's cpus_allowed versus its cpuset's * @tsk: task to test @@ -790,7 +760,12 @@ static int update_tasks_cpumask(struct cpuset *cs) struct ptr_heap heap; int retval; - retval = heap_init(&heap, PAGE_SIZE, GFP_KERNEL, &started_after); + /* + * cgroup_scan_tasks() will initialize heap->gt for us. + * heap_init() is still needed here for we should not change + * cs->cpus_allowed when heap_init() fails. + */ + retval = heap_init(&heap, PAGE_SIZE, GFP_KERNEL, NULL); if (retval) return retval; -- cgit v1.2.3 From da5ef6bb96158b0fc0d808704237a453af449124 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Fri, 25 Jul 2008 01:47:25 -0700 Subject: cpuset: two minor code-cleanups In cpuset_update_task_memory_state() local variable struct task_struct *tsk = current; And local variable tsk is used 14 times and statement task_cs(tsk) is used twice in this function. So using task_cs(tsk) instead of task_cs(current) is better for readability. And "(struct cgroup_scanner *)&scan" is not good for readability also. (and "container_of" is used in cpuset_do_move_task(), not "(cpuset_hotplug_scanner *)scan") Signed-off-by: Lai Jiangshan Acked-by: Paul Menage Cc: Paul Jackson Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cpuset.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index ebbc9b082e48..91cf85b36dd5 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -365,7 +365,7 @@ void cpuset_update_task_memory_state(void) my_cpusets_mem_gen = top_cpuset.mems_generation; } else { rcu_read_lock(); - my_cpusets_mem_gen = task_cs(current)->mems_generation; + my_cpusets_mem_gen = task_cs(tsk)->mems_generation; rcu_read_unlock(); } @@ -1777,7 +1777,7 @@ static void move_member_tasks_to_cpuset(struct cpuset *from, struct cpuset *to) scan.scan.heap = NULL; scan.to = to->css.cgroup; - if (cgroup_scan_tasks((struct cgroup_scanner *)&scan)) + if (cgroup_scan_tasks(&scan.scan)) printk(KERN_ERR "move_member_tasks_to_cpuset: " "cgroup_scan_tasks failed\n"); } -- cgit v1.2.3 From 4b7a1304267bff68260ae861784b27130e805be3 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:26 -0700 Subject: posix timers: timer_delete: remove the bogus "->it_process != NULL" check sys_timer_delete() and itimer_delete() check "timer->it_process != NULL", this looks completely bogus. ->it_process == NULL means that this timer is already under destruction or it is not fully initialized, this must not happen. sys_timer_delete: the timer is locked, and lock_timer() can't succeed if ->it_process == NULL. itimer_delete: it is called by exit_itimers() when there are no other threads which can play with signal_struct->posix_timers. Signed-off-by: Oleg Nesterov Acked-by: Roland McGrath Cc: john stultz Cc: Thomas Gleixner Cc: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/posix-timers.c | 18 ++++++++---------- 1 file changed, 8 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/posix-timers.c b/kernel/posix-timers.c index dbd8398ddb0b..17f53266fb67 100644 --- a/kernel/posix-timers.c +++ b/kernel/posix-timers.c @@ -856,11 +856,10 @@ retry_delete: * This keeps any tasks waiting on the spin lock from thinking * they got something (see the lock code above). */ - if (timer->it_process) { - if (timer->it_sigev_notify == (SIGEV_SIGNAL|SIGEV_THREAD_ID)) - put_task_struct(timer->it_process); - timer->it_process = NULL; - } + if (timer->it_sigev_notify == (SIGEV_SIGNAL|SIGEV_THREAD_ID)) + put_task_struct(timer->it_process); + timer->it_process = NULL; + unlock_timer(timer, flags); release_posix_timer(timer, IT_ID_SET); return 0; @@ -885,11 +884,10 @@ retry_delete: * This keeps any tasks waiting on the spin lock from thinking * they got something (see the lock code above). */ - if (timer->it_process) { - if (timer->it_sigev_notify == (SIGEV_SIGNAL|SIGEV_THREAD_ID)) - put_task_struct(timer->it_process); - timer->it_process = NULL; - } + if (timer->it_sigev_notify == (SIGEV_SIGNAL|SIGEV_THREAD_ID)) + put_task_struct(timer->it_process); + timer->it_process = NULL; + unlock_timer(timer, flags); release_posix_timer(timer, IT_ID_SET); } -- cgit v1.2.3 From 96347e7759e2e433c427defa0fa1adfc8cce6226 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:27 -0700 Subject: posix timers: release_posix_timer: kill the bogus put_task_struct(->it_process); release_posix_timer() can't be called with ->it_process != NULL. Once sys_timer_create() sets ->it_process it must not call release_posix_timer(), otherwise we can race with another thread doing sys_timer_delete(), this timer is visible to idr_find() and unlocked. The same is true for two other callers (actually, for any possible caller), sys_timer_delete() and itimer_delete(). They must clear ->it_process before unlock_timer() + release_posix_timer(). Signed-off-by: Oleg Nesterov Acked-by: Roland McGrath Cc: john stultz Cc: Thomas Gleixner Cc: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/posix-timers.c | 3 --- 1 file changed, 3 deletions(-) (limited to 'kernel') diff --git a/kernel/posix-timers.c b/kernel/posix-timers.c index 17f53266fb67..9a21681aa80f 100644 --- a/kernel/posix-timers.c +++ b/kernel/posix-timers.c @@ -449,9 +449,6 @@ static void release_posix_timer(struct k_itimer *tmr, int it_id_set) spin_unlock_irqrestore(&idr_lock, flags); } sigqueue_free(tmr->sigq); - if (unlikely(tmr->it_process) && - tmr->it_sigev_notify == (SIGEV_SIGNAL|SIGEV_THREAD_ID)) - put_task_struct(tmr->it_process); kmem_cache_free(posix_timers_cache, tmr); } -- cgit v1.2.3 From 6715ca451cfff1c9ce4b33ad9918a1dacf43997c Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:27 -0700 Subject: signals: collect_signal: remove the unneeded sigismember() check collect_signal() checks sigismember(&list->signal, sig), this is not needed. This "sig" was just found by next_signal(), so it must be valid. We have a (completely broken) call to ->notifier in between, but it must not play with sigpending->signal bits or unlock ->siglock. Signed-off-by: Oleg Nesterov Acked-by: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/signal.c | 3 --- 1 file changed, 3 deletions(-) (limited to 'kernel') diff --git a/kernel/signal.c b/kernel/signal.c index 6c0958e52ea7..c5b9aabb1550 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -343,9 +343,6 @@ static int collect_signal(int sig, struct sigpending *list, siginfo_t *info) struct sigqueue *q, *first = NULL; int still_pending = 0; - if (unlikely(!sigismember(&list->signal, sig))) - return 0; - /* * Collect the siginfo appropriate to this signal. Check if * there is another siginfo for the same signal. -- cgit v1.2.3 From d4434207616980885205c605697868c0f07e4378 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:28 -0700 Subject: signals: collect_signal: simplify the "still_pending" logic Factor out sigdelset() calls and remove the "still_pending" variable. Signed-off-by: Oleg Nesterov Acked-by: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/signal.c | 15 ++++++--------- 1 file changed, 6 insertions(+), 9 deletions(-) (limited to 'kernel') diff --git a/kernel/signal.c b/kernel/signal.c index c5b9aabb1550..50ad439377b2 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -341,7 +341,6 @@ unblock_all_signals(void) static int collect_signal(int sig, struct sigpending *list, siginfo_t *info) { struct sigqueue *q, *first = NULL; - int still_pending = 0; /* * Collect the siginfo appropriate to this signal. Check if @@ -349,26 +348,24 @@ static int collect_signal(int sig, struct sigpending *list, siginfo_t *info) */ list_for_each_entry(q, &list->list, list) { if (q->info.si_signo == sig) { - if (first) { - still_pending = 1; - break; - } + if (first) + goto still_pending; first = q; } } + + sigdelset(&list->signal, sig); + if (first) { +still_pending: list_del_init(&first->list); copy_siginfo(info, &first->info); __sigqueue_free(first); - if (!still_pending) - sigdelset(&list->signal, sig); } else { - /* Ok, it wasn't in the queue. This must be a fast-pathed signal or we must have been out of queue space. So zero out the info. */ - sigdelset(&list->signal, sig); info->si_signo = sig; info->si_errno = 0; info->si_code = 0; -- cgit v1.2.3 From 100360f03077663b7bef3af44805b6cf700c3bee Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:29 -0700 Subject: signals: change collect_signal() to return void With the recent changes collect_signal() always returns true. Change it to return void and update the single caller. Signed-off-by: Oleg Nesterov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/signal.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/signal.c b/kernel/signal.c index 50ad439377b2..fea236fe0b50 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -338,7 +338,7 @@ unblock_all_signals(void) spin_unlock_irqrestore(¤t->sighand->siglock, flags); } -static int collect_signal(int sig, struct sigpending *list, siginfo_t *info) +static void collect_signal(int sig, struct sigpending *list, siginfo_t *info) { struct sigqueue *q, *first = NULL; @@ -372,7 +372,6 @@ still_pending: info->si_pid = 0; info->si_uid = 0; } - return 1; } static int __dequeue_signal(struct sigpending *pending, sigset_t *mask, @@ -390,8 +389,7 @@ static int __dequeue_signal(struct sigpending *pending, sigset_t *mask, } } - if (!collect_signal(sig, pending, info)) - sig = 0; + collect_signal(sig, pending, info); } return sig; -- cgit v1.2.3 From 3854a771821c970065e3203a0b40ddc4101538cc Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:29 -0700 Subject: __exit_signal: don't take rcu lock There is no reason for rcu_read_lock() in __exit_signal(). tsk->sighand can only be changed if tsk does exec, obviously this is not possible. Signed-off-by: Oleg Nesterov Cc: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/exit.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'kernel') diff --git a/kernel/exit.c b/kernel/exit.c index 93d2711b9381..a7799d8a6404 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -85,7 +85,6 @@ static void __exit_signal(struct task_struct *tsk) BUG_ON(!sig); BUG_ON(!atomic_read(&sig->count)); - rcu_read_lock(); sighand = rcu_dereference(tsk->sighand); spin_lock(&sighand->siglock); @@ -136,7 +135,6 @@ static void __exit_signal(struct task_struct *tsk) tsk->signal = NULL; tsk->sighand = NULL; spin_unlock(&sighand->siglock); - rcu_read_unlock(); __cleanup_sighand(sighand); clear_tsk_thread_flag(tsk,TIF_SIGPENDING); -- cgit v1.2.3 From 92413d771e7123304fb4b9efd2a00cccc946e383 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:30 -0700 Subject: signals: dequeue_signal: don't check SIGNAL_GROUP_EXIT when setting SIGNAL_STOP_DEQUEUED dequeue_signal() checks SIGNAL_GROUP_EXIT before setting SIGNAL_STOP_DEQUEUED. This was added by 788e05a67c343fa22f2ae1d3ca264e7f15c25eaf a long ago to avoid the coredump/SIGSTOP race. Since then the related code was changed, and now this subtle check is both incomplete and unneeded at the same time. It is incomplete because nowadays exec() doesn't set SIGNAL_GROUP_EXIT, so in fact we should check signal_group_exit() to avoid a similar race. Fortunately, we doesn't need the check at all. The only function which relies on SIGNAL_STOP_DEQUEUED is do_signal_stop(), and it ignores this flag if signal_group_exit() == T, this covers the SIGNAL_GROUP_EXIT case. Signed-off-by: Oleg Nesterov Acked-by: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/signal.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/signal.c b/kernel/signal.c index fea236fe0b50..15f901a26ec7 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -454,8 +454,7 @@ int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info) * is to alert stop-signal processing code when another * processor has come along and cleared the flag. */ - if (!(tsk->signal->flags & SIGNAL_GROUP_EXIT)) - tsk->signal->flags |= SIGNAL_STOP_DEQUEUED; + tsk->signal->flags |= SIGNAL_STOP_DEQUEUED; } if ((info->si_code & __SI_MASK) == __SI_TIMER && info->si_sys_private) { /* -- cgit v1.2.3 From 2b201a9eddf509e8e935b45e573648e36f4b623f Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:31 -0700 Subject: signals: do_signal_stop: kill the SIGNAL_UNKILLABLE check fae5fa44f1fd079ffbed8e0add929dd7bbd1347f changed do_signal_stop() to check SIGNAL_UNKILLABLE, this wasn't needed. If signal_group_exit() == F, the signal sent to SIGNAL_UNKILLABLE task must be already filtered out by the caller, get_signal_to_deliver(). And if signal_group_exit() == T we are not going to stop. Signed-off-by: Oleg Nesterov Acked-by: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/signal.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/signal.c b/kernel/signal.c index 15f901a26ec7..0514da573f22 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -1649,8 +1649,7 @@ static int do_signal_stop(int signr) } else { struct task_struct *t; - if (unlikely((sig->flags & (SIGNAL_STOP_DEQUEUED | SIGNAL_UNKILLABLE)) - != SIGNAL_STOP_DEQUEUED) || + if (!likely(sig->flags & SIGNAL_STOP_DEQUEUED) || unlikely(signal_group_exit(sig))) return 0; /* -- cgit v1.2.3 From d8878ba3f05ae5bbfad5a6e72e5121c0ea35f989 Mon Sep 17 00:00:00 2001 From: Michael Kerrisk Date: Fri, 25 Jul 2008 01:47:32 -0700 Subject: signals: make siginfo_t si_utime + si_sstime report times in USER_HZ, not HZ In the switch to configurable HZ in 2.6, the treatment of the si_utime and si_stime fields that are exposed to userland via the siginfo structure looks to have been botched. As things stand, these fields report times in units of HZ, so that userland gets information that varies depending on the HZ that the kernel was configured with. This patch changes the reported values to use USER_HZ units. Signed-off-by: Michael Kerrisk Acked-by: Oleg Nesterov Cc: Thomas Gleixner Cc: Ingo Molnar Cc: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/signal.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/signal.c b/kernel/signal.c index 0514da573f22..ba60eeeb63aa 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -1370,10 +1370,9 @@ void do_notify_parent(struct task_struct *tsk, int sig) info.si_uid = tsk->uid; - /* FIXME: find out whether or not this is supposed to be c*time. */ - info.si_utime = cputime_to_jiffies(cputime_add(tsk->utime, + info.si_utime = cputime_to_clock_t(cputime_add(tsk->utime, tsk->signal->utime)); - info.si_stime = cputime_to_jiffies(cputime_add(tsk->stime, + info.si_stime = cputime_to_clock_t(cputime_add(tsk->stime, tsk->signal->stime)); info.si_status = tsk->exit_code & 0x7f; @@ -1441,9 +1440,8 @@ static void do_notify_parent_cldstop(struct task_struct *tsk, int why) info.si_uid = tsk->uid; - /* FIXME: find out whether or not this is supposed to be c*time. */ - info.si_utime = cputime_to_jiffies(tsk->utime); - info.si_stime = cputime_to_jiffies(tsk->stime); + info.si_utime = cputime_to_clock_t(tsk->utime); + info.si_stime = cputime_to_clock_t(tsk->stime); info.si_code = why; switch (why) { -- cgit v1.2.3 From bc64efd220dcd4449aef8dd2564d73127b583b09 Mon Sep 17 00:00:00 2001 From: Gustavo Fernando Padovan Date: Fri, 25 Jul 2008 01:47:33 -0700 Subject: kernel/signal.c: change vars pid and tgid types to pid_t Change the type of pid and tgid variables from int to the POSIX type pid_t. Signed-off-by: Gustavo F. Padovan Cc: Oleg Nesterov Cc: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/signal.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/signal.c b/kernel/signal.c index ba60eeeb63aa..fdab7b363fa7 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -1116,7 +1116,7 @@ EXPORT_SYMBOL_GPL(kill_pid_info_as_uid); * is probably wrong. Should make it like BSD or SYSV. */ -static int kill_something_info(int sig, struct siginfo *info, int pid) +static int kill_something_info(int sig, struct siginfo *info, pid_t pid) { int ret; @@ -2184,7 +2184,7 @@ sys_rt_sigtimedwait(const sigset_t __user *uthese, } asmlinkage long -sys_kill(int pid, int sig) +sys_kill(pid_t pid, int sig) { struct siginfo info; @@ -2197,7 +2197,7 @@ sys_kill(int pid, int sig) return kill_something_info(sig, &info, pid); } -static int do_tkill(int tgid, int pid, int sig) +static int do_tkill(pid_t tgid, pid_t pid, int sig) { int error; struct siginfo info; @@ -2243,7 +2243,7 @@ static int do_tkill(int tgid, int pid, int sig) * exists but it's not belonging to the target process anymore. This * method solves the problem of threads exiting and PIDs getting reused. */ -asmlinkage long sys_tgkill(int tgid, int pid, int sig) +asmlinkage long sys_tgkill(pid_t tgid, pid_t pid, int sig) { /* This is only valid for single tasks */ if (pid <= 0 || tgid <= 0) @@ -2256,7 +2256,7 @@ asmlinkage long sys_tgkill(int tgid, int pid, int sig) * Send a signal to only one task, even if it's a CLONE_THREAD task. */ asmlinkage long -sys_tkill(int pid, int sig) +sys_tkill(pid_t pid, int sig) { /* This is only valid for single tasks */ if (pid <= 0) @@ -2266,7 +2266,7 @@ sys_tkill(int pid, int sig) } asmlinkage long -sys_rt_sigqueueinfo(int pid, int sig, siginfo_t __user *uinfo) +sys_rt_sigqueueinfo(pid_t pid, int sig, siginfo_t __user *uinfo) { siginfo_t info; -- cgit v1.2.3 From 3d749b9e676b26584a47e75c235aa6f69d0697ae Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:37 -0700 Subject: ptrace: simplify ptrace_stop()->sigkill_pending() path 1. SIGKILL can't be blocked, remove this check from sigkill_pending(). 2. When ptrace_stop() sees sigkill_pending() == T, it can just return. Kill "int killed" and simplify the code. This also is more correct, the tracer shouldn't see us in TASK_TRACED if we are not going to stop. I strongly believe this code needs further changes. We should do the "was this task killed" check unconditionally, currently it depends on arch_ptrace_stop_needed(). On the other hand, sigkill_pending() isn't very clever. If the task was killed tkill(SIGKILL), the signal can be already dequeued if the caller is do_exit(). Signed-off-by: Oleg Nesterov Cc: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/signal.c | 12 +++++------- 1 file changed, 5 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/signal.c b/kernel/signal.c index fdab7b363fa7..39c1706edf03 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -1496,9 +1496,8 @@ static inline int may_ptrace_stop(void) */ static int sigkill_pending(struct task_struct *tsk) { - return ((sigismember(&tsk->pending.signal, SIGKILL) || - sigismember(&tsk->signal->shared_pending.signal, SIGKILL)) && - !unlikely(sigismember(&tsk->blocked, SIGKILL))); + return sigismember(&tsk->pending.signal, SIGKILL) || + sigismember(&tsk->signal->shared_pending.signal, SIGKILL); } /* @@ -1514,8 +1513,6 @@ static int sigkill_pending(struct task_struct *tsk) */ static void ptrace_stop(int exit_code, int clear_code, siginfo_t *info) { - int killed = 0; - if (arch_ptrace_stop_needed(exit_code, info)) { /* * The arch code has something special to do before a @@ -1531,7 +1528,8 @@ static void ptrace_stop(int exit_code, int clear_code, siginfo_t *info) spin_unlock_irq(¤t->sighand->siglock); arch_ptrace_stop(exit_code, info); spin_lock_irq(¤t->sighand->siglock); - killed = sigkill_pending(current); + if (sigkill_pending(current)) + return; } /* @@ -1548,7 +1546,7 @@ static void ptrace_stop(int exit_code, int clear_code, siginfo_t *info) __set_current_state(TASK_TRACED); spin_unlock_irq(¤t->sighand->siglock); read_lock(&tasklist_lock); - if (!unlikely(killed) && may_ptrace_stop()) { + if (may_ptrace_stop()) { do_notify_parent_cldstop(current, CLD_TRAPPED); read_unlock(&tasklist_lock); schedule(); -- cgit v1.2.3 From 7b34e4283c685f5cc6ba6d30e939906eee0d4bcf Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:37 -0700 Subject: introduce PF_KTHREAD flag Introduce the new PF_KTHREAD flag to mark the kernel threads. It is set by INIT_TASK() and copied to the forked childs (we could set it in kthreadd() along with PF_NOFREEZE instead). daemonize() was changed as well. In that case testing of PF_KTHREAD is racy, but daemonize() is hopeless anyway. This flag is cleared in do_execve(), before search_binary_handler(). Probably not the best place, we can do this in exec_mmap() or in start_thread(), or clear it along with PF_FORKNOEXEC. But I think this doesn't matter in practice, and if do_execve() fails kthread should die soon. Signed-off-by: Oleg Nesterov Cc: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/exit.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/exit.c b/kernel/exit.c index a7799d8a6404..28a44a2612dc 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -430,7 +430,7 @@ void daemonize(const char *name, ...) * We don't want to have TIF_FREEZE set if the system-wide hibernation * or suspend transition begins right now. */ - current->flags |= PF_NOFREEZE; + current->flags |= (PF_NOFREEZE | PF_KTHREAD); if (current->nsproxy != &init_nsproxy) { get_nsproxy(&init_nsproxy); -- cgit v1.2.3 From 246bb0b1deb29726990620d8b5e55ca29f331362 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:38 -0700 Subject: kill PF_BORROWED_MM in favour of PF_KTHREAD Kill PF_BORROWED_MM. Change use_mm/unuse_mm to not play with ->flags, and do s/PF_BORROWED_MM/PF_KTHREAD/ for a couple of other users. No functional changes yet. But this allows us to do further fixes/cleanups. oom_kill/ptrace/etc often check "p->mm != NULL" to filter out the kthreads, this is wrong because of use_mm(). The problem with PF_BORROWED_MM is that we need task_lock() to avoid races. With this patch we can check PF_KTHREAD directly, or use a simple lockless helper: /* The result must not be dereferenced !!! */ struct mm_struct *__get_task_mm(struct task_struct *tsk) { if (tsk->flags & PF_KTHREAD) return NULL; return tsk->mm; } Note also ecard_task(). It runs with ->mm != NULL, but it's the kernel thread without PF_BORROWED_MM. Signed-off-by: Oleg Nesterov Cc: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/fork.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index 228f80c9155a..eeaec6893b0d 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -474,7 +474,7 @@ EXPORT_SYMBOL_GPL(mmput); /** * get_task_mm - acquire a reference to the task's mm * - * Returns %NULL if the task has no mm. Checks PF_BORROWED_MM (meaning + * Returns %NULL if the task has no mm. Checks PF_KTHREAD (meaning * this kernel workthread has transiently adopted a user mm with use_mm, * to do its AIO) is not set and if so returns a reference to it, after * bumping up the use count. User must release the mm via mmput() @@ -487,7 +487,7 @@ struct mm_struct *get_task_mm(struct task_struct *task) task_lock(task); mm = task->mm; if (mm) { - if (task->flags & PF_BORROWED_MM) + if (task->flags & PF_KTHREAD) mm = NULL; else atomic_inc(&mm->mm_users); -- cgit v1.2.3 From 32ecb1f26dd50eeaac4e3f4dea4541c97848e459 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:41 -0700 Subject: coredump: turn mm->core_startup_done into the pointer to struct core_state mm->core_startup_done points to "struct completion startup_done" allocated on the coredump_wait()'s stack. Introduce the new structure, core_state, which holds this "struct completion". This way we can add more info visible to the threads participating in coredump without enlarging mm_struct. No changes in affected .o files. Signed-off-by: Oleg Nesterov Cc: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/exit.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/exit.c b/kernel/exit.c index 28a44a2612dc..f7fa21dbced4 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -680,7 +680,7 @@ static void exit_mm(struct task_struct * tsk) up_read(&mm->mmap_sem); down_write(&mm->mmap_sem); if (!--mm->core_waiters) - complete(mm->core_startup_done); + complete(&mm->core_state->startup); up_write(&mm->mmap_sem); wait_for_completion(&mm->core_done); -- cgit v1.2.3 From 999d9fc1670bc082928b93b11d1f2e0e417d973c Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:41 -0700 Subject: coredump: move mm->core_waiters into struct core_state Move mm->core_waiters into "struct core_state" allocated on stack. This shrinks mm_struct a little bit and allows further changes. This patch mostly does s/core_waiters/core_state. The only essential change is that coredump_wait() must clear mm->core_state before return. The coredump_wait()'s path is uglified and .text grows by 30 bytes, this is fixed by the next patch. Signed-off-by: Oleg Nesterov Cc: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/exit.c | 8 ++++---- kernel/fork.c | 2 +- kernel/signal.c | 4 ++-- 3 files changed, 7 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/exit.c b/kernel/exit.c index f7fa21dbced4..988e232254e9 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -670,16 +670,16 @@ static void exit_mm(struct task_struct * tsk) return; /* * Serialize with any possible pending coredump. - * We must hold mmap_sem around checking core_waiters + * We must hold mmap_sem around checking core_state * and clearing tsk->mm. The core-inducing thread - * will increment core_waiters for each thread in the + * will increment ->nr_threads for each thread in the * group with ->mm != NULL. */ down_read(&mm->mmap_sem); - if (mm->core_waiters) { + if (mm->core_state) { up_read(&mm->mmap_sem); down_write(&mm->mmap_sem); - if (!--mm->core_waiters) + if (!--mm->core_state->nr_threads) complete(&mm->core_state->startup); up_write(&mm->mmap_sem); diff --git a/kernel/fork.c b/kernel/fork.c index eeaec6893b0d..813d5c89b9d5 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -400,7 +400,7 @@ static struct mm_struct * mm_init(struct mm_struct * mm, struct task_struct *p) INIT_LIST_HEAD(&mm->mmlist); mm->flags = (current->mm) ? current->mm->flags : MMF_DUMP_FILTER_DEFAULT; - mm->core_waiters = 0; + mm->core_state = NULL; mm->nr_ptes = 0; set_mm_counter(mm, file_rss, 0); set_mm_counter(mm, anon_rss, 0); diff --git a/kernel/signal.c b/kernel/signal.c index 39c1706edf03..5c7b7eaa0dc6 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -1480,10 +1480,10 @@ static inline int may_ptrace_stop(void) * is a deadlock situation, and pointless because our tracer * is dead so don't allow us to stop. * If SIGKILL was already sent before the caller unlocked - * ->siglock we must see ->core_waiters != 0. Otherwise it + * ->siglock we must see ->core_state != NULL. Otherwise it * is safe to enter schedule(). */ - if (unlikely(current->mm->core_waiters) && + if (unlikely(current->mm->core_state) && unlikely(current->mm == current->parent->mm)) return 0; -- cgit v1.2.3 From c5f1cc8c1828486a61ab3e575da6e2c62b34d399 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:42 -0700 Subject: coredump: turn core_state->nr_threads into atomic_t Turn core_state->nr_threads into atomic_t and kill now unneeded down_write(&mm->mmap_sem) in exit_mm(). Signed-off-by: Oleg Nesterov Cc: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/exit.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/exit.c b/kernel/exit.c index 988e232254e9..63d82957baae 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -678,10 +678,9 @@ static void exit_mm(struct task_struct * tsk) down_read(&mm->mmap_sem); if (mm->core_state) { up_read(&mm->mmap_sem); - down_write(&mm->mmap_sem); - if (!--mm->core_state->nr_threads) + + if (atomic_dec_and_test(&mm->core_state->nr_threads)) complete(&mm->core_state->startup); - up_write(&mm->mmap_sem); wait_for_completion(&mm->core_done); down_read(&mm->mmap_sem); -- cgit v1.2.3 From b564daf806d492dd4f7afe9b6c83b8d35d137669 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:44 -0700 Subject: coredump: construct the list of coredumping threads at startup time binfmt->core_dump() has to iterate over the all threads in system in order to find the coredumping threads and construct the list using the GFP_ATOMIC allocations. With this patch each thread allocates the list node on exit_mm()'s stack and adds itself to the list. This allows us to do further changes: - simplify ->core_dump() - change exit_mm() to clear ->mm first, then wait for ->core_done. this makes the coredumping process visible to oom_kill - kill mm->core_done Signed-off-by: Oleg Nesterov Acked-by: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/exit.c | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/exit.c b/kernel/exit.c index 63d82957baae..b66f0d55c791 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -664,6 +664,7 @@ assign_new_owner: static void exit_mm(struct task_struct * tsk) { struct mm_struct *mm = tsk->mm; + struct core_state *core_state; mm_release(tsk, mm); if (!mm) @@ -676,11 +677,19 @@ static void exit_mm(struct task_struct * tsk) * group with ->mm != NULL. */ down_read(&mm->mmap_sem); - if (mm->core_state) { + core_state = mm->core_state; + if (core_state) { + struct core_thread self; up_read(&mm->mmap_sem); - if (atomic_dec_and_test(&mm->core_state->nr_threads)) - complete(&mm->core_state->startup); + self.task = tsk; + self.next = xchg(&core_state->dumper.next, &self); + /* + * Implies mb(), the result of xchg() must be visible + * to core_state->dumper. + */ + if (atomic_dec_and_test(&core_state->nr_threads)) + complete(&core_state->startup); wait_for_completion(&mm->core_done); down_read(&mm->mmap_sem); -- cgit v1.2.3 From a94e2d408eaedbd85aae259621d46fafc10479a2 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:46 -0700 Subject: coredump: kill mm->core_done Now that we have core_state->dumper list we can use it to wake up the sub-threads waiting for the coredump completion. This uglifies the code and .text grows by 47 bytes, but otoh mm_struct lessens by sizeof(struct completion). Also, with this change we can decouple exit_mm() from the coredumping code. Signed-off-by: Oleg Nesterov Cc: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/exit.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/exit.c b/kernel/exit.c index b66f0d55c791..8a4d4d12e294 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -691,7 +691,13 @@ static void exit_mm(struct task_struct * tsk) if (atomic_dec_and_test(&core_state->nr_threads)) complete(&core_state->startup); - wait_for_completion(&mm->core_done); + for (;;) { + set_task_state(tsk, TASK_UNINTERRUPTIBLE); + if (!self.task) /* see coredump_finish() */ + break; + schedule(); + } + __set_task_state(tsk, TASK_RUNNING); down_read(&mm->mmap_sem); } atomic_inc(&mm->mm_count); -- cgit v1.2.3 From 1a4d9b0aa0d3c50314e57525a5e5ec2cfc48b4c8 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:47 -0700 Subject: workqueues: insert_work: use "list_head *" instead of "int tail" insert_work() inserts the new work_struct before or after cwq->worklist, depending on the "int tail" parameter. Change it to accept "list_head *" instead, this shrinks .text a bit and allows us to insert the barrier after specific work_struct. Signed-off-by: Oleg Nesterov Cc: Jarek Poplawski Cc: Max Krasnyansky Cc: Peter Zijlstra Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/workqueue.c | 17 +++++++---------- 1 file changed, 7 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 6fd158b21026..d9a2d65cc63e 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -125,7 +125,7 @@ struct cpu_workqueue_struct *get_wq_data(struct work_struct *work) } static void insert_work(struct cpu_workqueue_struct *cwq, - struct work_struct *work, int tail) + struct work_struct *work, struct list_head *head) { set_wq_data(work, cwq); /* @@ -133,10 +133,7 @@ static void insert_work(struct cpu_workqueue_struct *cwq, * result of list_add() below, see try_to_grab_pending(). */ smp_wmb(); - if (tail) - list_add_tail(&work->entry, &cwq->worklist); - else - list_add(&work->entry, &cwq->worklist); + list_add_tail(&work->entry, head); wake_up(&cwq->more_work); } @@ -146,7 +143,7 @@ static void __queue_work(struct cpu_workqueue_struct *cwq, unsigned long flags; spin_lock_irqsave(&cwq->lock, flags); - insert_work(cwq, work, 1); + insert_work(cwq, work, &cwq->worklist); spin_unlock_irqrestore(&cwq->lock, flags); } @@ -361,14 +358,14 @@ static void wq_barrier_func(struct work_struct *work) } static void insert_wq_barrier(struct cpu_workqueue_struct *cwq, - struct wq_barrier *barr, int tail) + struct wq_barrier *barr, struct list_head *head) { INIT_WORK(&barr->work, wq_barrier_func); __set_bit(WORK_STRUCT_PENDING, work_data_bits(&barr->work)); init_completion(&barr->done); - insert_work(cwq, &barr->work, tail); + insert_work(cwq, &barr->work, head); } static int flush_cpu_workqueue(struct cpu_workqueue_struct *cwq) @@ -388,7 +385,7 @@ static int flush_cpu_workqueue(struct cpu_workqueue_struct *cwq) active = 0; spin_lock_irq(&cwq->lock); if (!list_empty(&cwq->worklist) || cwq->current_work != NULL) { - insert_wq_barrier(cwq, &barr, 1); + insert_wq_barrier(cwq, &barr, &cwq->worklist); active = 1; } spin_unlock_irq(&cwq->lock); @@ -473,7 +470,7 @@ static void wait_on_cpu_work(struct cpu_workqueue_struct *cwq, spin_lock_irq(&cwq->lock); if (unlikely(cwq->current_work == work)) { - insert_wq_barrier(cwq, &barr, 0); + insert_wq_barrier(cwq, &barr, cwq->worklist.next); running = 1; } spin_unlock_irq(&cwq->lock); -- cgit v1.2.3 From db700897224b5ebdf852f2d38920ce428940d059 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:49 -0700 Subject: workqueues: implement flush_work() Most of users of flush_workqueue() can be changed to use cancel_work_sync(), but sometimes we really need to wait for the completion and cancelling is not an option. schedule_on_each_cpu() is good example. Add the new helper, flush_work(work), which waits for the completion of the specific work_struct. More precisely, it "flushes" the result of of the last queue_work() which is visible to the caller. For example, this code queue_work(wq, work); /* WINDOW */ queue_work(wq, work); flush_work(work); doesn't necessary work "as expected". What can happen in the WINDOW above is - wq starts the execution of work->func() - the caller migrates to another CPU now, after the 2nd queue_work() this work is active on the previous CPU, and at the same time it is queued on another. In this case flush_work(work) may return before the first work->func() completes. It is trivial to add another helper int flush_work_sync(struct work_struct *work) { return flush_work(work) || wait_on_work(work); } which works "more correctly", but it has to iterate over all CPUs and thus it much slower than flush_work(). Signed-off-by: Oleg Nesterov Acked-by: Max Krasnyansky Acked-by: Jarek Poplawski Cc: Peter Zijlstra Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/workqueue.c | 46 ++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 46 insertions(+) (limited to 'kernel') diff --git a/kernel/workqueue.c b/kernel/workqueue.c index d9a2d65cc63e..ee41cf857d55 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -423,6 +423,52 @@ void flush_workqueue(struct workqueue_struct *wq) } EXPORT_SYMBOL_GPL(flush_workqueue); +/** + * flush_work - block until a work_struct's callback has terminated + * @work: the work which is to be flushed + * + * It is expected that, prior to calling flush_work(), the caller has + * arranged for the work to not be requeued, otherwise it doesn't make + * sense to use this function. + */ +int flush_work(struct work_struct *work) +{ + struct cpu_workqueue_struct *cwq; + struct list_head *prev; + struct wq_barrier barr; + + might_sleep(); + cwq = get_wq_data(work); + if (!cwq) + return 0; + + prev = NULL; + spin_lock_irq(&cwq->lock); + if (!list_empty(&work->entry)) { + /* + * See the comment near try_to_grab_pending()->smp_rmb(). + * If it was re-queued under us we are not going to wait. + */ + smp_rmb(); + if (unlikely(cwq != get_wq_data(work))) + goto out; + prev = &work->entry; + } else { + if (cwq->current_work != work) + goto out; + prev = &cwq->worklist; + } + insert_wq_barrier(cwq, &barr, prev->next); +out: + spin_unlock_irq(&cwq->lock); + if (!prev) + return 0; + + wait_for_completion(&barr.done); + return 1; +} +EXPORT_SYMBOL_GPL(flush_work); + /* * Upon a successful return (>= 0), the caller "owns" WORK_STRUCT_PENDING bit, * so this work can't be re-armed in any way. -- cgit v1.2.3 From 8616a89ab761239c963eea3a63be383f127cc7e8 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:49 -0700 Subject: workqueues: schedule_on_each_cpu: use flush_work() Change schedule_on_each_cpu() to use flush_work() instead of flush_workqueue(), this way we don't wait for other work_struct's which can be queued meanwhile. Signed-off-by: Oleg Nesterov Cc: Jarek Poplawski Cc: Max Krasnyansky Cc: Peter Zijlstra Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/workqueue.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/workqueue.c b/kernel/workqueue.c index ee41cf857d55..5fbffd302eb5 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -690,7 +690,8 @@ int schedule_on_each_cpu(work_func_t func) set_bit(WORK_STRUCT_PENDING, work_data_bits(work)); __queue_work(per_cpu_ptr(keventd_wq->cpu_wq, cpu), work); } - flush_workqueue(keventd_wq); + for_each_online_cpu(cpu) + flush_work(per_cpu_ptr(works, cpu)); put_online_cpus(); free_percpu(works); return 0; -- cgit v1.2.3 From 3da1c84c00c7e5fa8348336bd8c342f9128b0f14 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:50 -0700 Subject: workqueues: make get_online_cpus() useable for work->func() workqueue_cpu_callback(CPU_DEAD) flushes cwq->thread under cpu_maps_update_begin(). This means that the multithreaded workqueues can't use get_online_cpus() due to the possible deadlock, very bad and very old problem. Introduce the new state, CPU_POST_DEAD, which is called after cpu_hotplug_done() but before cpu_maps_update_done(). Change workqueue_cpu_callback() to use CPU_POST_DEAD instead of CPU_DEAD. This means that create/destroy functions can't rely on get_online_cpus() any longer and should take cpu_add_remove_lock instead. [akpm@linux-foundation.org: fix CONFIG_SMP=n] Signed-off-by: Oleg Nesterov Acked-by: Gautham R Shenoy Cc: Heiko Carstens Cc: Max Krasnyansky Cc: Paul Jackson Cc: Paul Menage Cc: Peter Zijlstra Cc: Vegard Nossum Cc: Martin Schwidefsky Cc: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cpu.c | 5 +++++ kernel/workqueue.c | 18 +++++++++--------- 2 files changed, 14 insertions(+), 9 deletions(-) (limited to 'kernel') diff --git a/kernel/cpu.c b/kernel/cpu.c index 2cc409ce0a8f..10ba5f1004a5 100644 --- a/kernel/cpu.c +++ b/kernel/cpu.c @@ -285,6 +285,11 @@ out_allowed: set_cpus_allowed_ptr(current, &old_allowed); out_release: cpu_hotplug_done(); + if (!err) { + if (raw_notifier_call_chain(&cpu_chain, CPU_POST_DEAD | mod, + hcpu) == NOTIFY_BAD) + BUG(); + } return err; } diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 5fbffd302eb5..828e58230cbc 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -828,7 +828,7 @@ struct workqueue_struct *__create_workqueue_key(const char *name, err = create_workqueue_thread(cwq, singlethread_cpu); start_workqueue_thread(cwq, -1); } else { - get_online_cpus(); + cpu_maps_update_begin(); spin_lock(&workqueue_lock); list_add(&wq->list, &workqueues); spin_unlock(&workqueue_lock); @@ -840,7 +840,7 @@ struct workqueue_struct *__create_workqueue_key(const char *name, err = create_workqueue_thread(cwq, cpu); start_workqueue_thread(cwq, cpu); } - put_online_cpus(); + cpu_maps_update_done(); } if (err) { @@ -854,8 +854,8 @@ EXPORT_SYMBOL_GPL(__create_workqueue_key); static void cleanup_workqueue_thread(struct cpu_workqueue_struct *cwq) { /* - * Our caller is either destroy_workqueue() or CPU_DEAD, - * get_online_cpus() protects cwq->thread. + * Our caller is either destroy_workqueue() or CPU_POST_DEAD, + * cpu_add_remove_lock protects cwq->thread. */ if (cwq->thread == NULL) return; @@ -865,7 +865,7 @@ static void cleanup_workqueue_thread(struct cpu_workqueue_struct *cwq) flush_cpu_workqueue(cwq); /* - * If the caller is CPU_DEAD and cwq->worklist was not empty, + * If the caller is CPU_POST_DEAD and cwq->worklist was not empty, * a concurrent flush_workqueue() can insert a barrier after us. * However, in that case run_workqueue() won't return and check * kthread_should_stop() until it flushes all work_struct's. @@ -889,14 +889,14 @@ void destroy_workqueue(struct workqueue_struct *wq) const cpumask_t *cpu_map = wq_cpu_map(wq); int cpu; - get_online_cpus(); + cpu_maps_update_begin(); spin_lock(&workqueue_lock); list_del(&wq->list); spin_unlock(&workqueue_lock); for_each_cpu_mask_nr(cpu, *cpu_map) cleanup_workqueue_thread(per_cpu_ptr(wq->cpu_wq, cpu)); - put_online_cpus(); + cpu_maps_update_done(); free_percpu(wq->cpu_wq); kfree(wq); @@ -935,7 +935,7 @@ static int __devinit workqueue_cpu_callback(struct notifier_block *nfb, case CPU_UP_CANCELED: start_workqueue_thread(cwq, -1); - case CPU_DEAD: + case CPU_POST_DEAD: cleanup_workqueue_thread(cwq); break; } @@ -943,7 +943,7 @@ static int __devinit workqueue_cpu_callback(struct notifier_block *nfb, switch (action) { case CPU_UP_CANCELED: - case CPU_DEAD: + case CPU_POST_DEAD: cpu_clear(cpu, cpu_populated_map); } -- cgit v1.2.3 From a67da70dc0955580665f5444f318b92e69a3c272 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:52 -0700 Subject: workqueues: lockdep annotations for flush_work() Add lockdep annotations to flush_work() and update the comment. Signed-off-by: Oleg Nesterov Cc: Jarek Poplawski Acked-by: Johannes Berg Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/workqueue.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'kernel') diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 828e58230cbc..4fcb75b98443 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -427,6 +427,8 @@ EXPORT_SYMBOL_GPL(flush_workqueue); * flush_work - block until a work_struct's callback has terminated * @work: the work which is to be flushed * + * Returns false if @work has already terminated. + * * It is expected that, prior to calling flush_work(), the caller has * arranged for the work to not be requeued, otherwise it doesn't make * sense to use this function. @@ -442,6 +444,9 @@ int flush_work(struct work_struct *work) if (!cwq) return 0; + lock_acquire(&cwq->wq->lockdep_map, 0, 0, 0, 2, _THIS_IP_); + lock_release(&cwq->wq->lockdep_map, 1, _THIS_IP_); + prev = NULL; spin_lock_irq(&cwq->lock); if (!list_empty(&work->entry)) { -- cgit v1.2.3 From ef1ca236b8d645349ed6569598ae3f6c1b9511c0 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:53 -0700 Subject: workqueues: queue_work() can use queue_work_on() queue_work() can use queue_work_on() to avoid the code duplication. Signed-off-by: Oleg Nesterov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/workqueue.c | 11 ++++------- 1 file changed, 4 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 4fcb75b98443..fe08a8512ddd 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -159,14 +159,11 @@ static void __queue_work(struct cpu_workqueue_struct *cwq, */ int queue_work(struct workqueue_struct *wq, struct work_struct *work) { - int ret = 0; + int ret; + + ret = queue_work_on(get_cpu(), wq, work); + put_cpu(); - if (!test_and_set_bit(WORK_STRUCT_PENDING, work_data_bits(work))) { - BUG_ON(!list_empty(&work->entry)); - __queue_work(wq_per_cpu(wq, get_cpu()), work); - put_cpu(); - ret = 1; - } return ret; } EXPORT_SYMBOL_GPL(queue_work); -- cgit v1.2.3 From 8de6d308bab4f67fcf953562f9f08f9527cad72d Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:53 -0700 Subject: workqueues: schedule_on_each_cpu() can use schedule_work_on() schedule_on_each_cpu() can use schedule_work_on() to avoid the code duplication. Signed-off-by: Oleg Nesterov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/workqueue.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/workqueue.c b/kernel/workqueue.c index fe08a8512ddd..7cf430372f89 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -689,8 +689,7 @@ int schedule_on_each_cpu(work_func_t func) struct work_struct *work = per_cpu_ptr(works, cpu); INIT_WORK(work, func); - set_bit(WORK_STRUCT_PENDING, work_data_bits(work)); - __queue_work(per_cpu_ptr(keventd_wq->cpu_wq, cpu), work); + schedule_work_on(cpu, work); } for_each_online_cpu(cpu) flush_work(per_cpu_ptr(works, cpu)); -- cgit v1.2.3 From 8448502cfc915f70e3f8923849ade27d472044cb Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:54 -0700 Subject: workqueues: do CPU_UP_CANCELED if CPU_UP_PREPARE fails The bug was pointed out by Akinobu Mita , and this patch is based on his original patch. workqueue_cpu_callback(CPU_UP_PREPARE) expects that if it returns NOTIFY_BAD, _cpu_up() will send CPU_UP_CANCELED then. However, this is not true since "cpu hotplug: cpu: deliver CPU_UP_CANCELED only to NOTIFY_OKed callbacks with CPU_UP_PREPARE" commit: a0d8cdb652d35af9319a9e0fb7134de2a276c636 The callback which has returned NOTIFY_BAD will not receive CPU_UP_CANCELED. Change the code to fulfil the CPU_UP_CANCELED logic if CPU_UP_PREPARE fails. Signed-off-by: Oleg Nesterov Reported-by: Akinobu Mita Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/workqueue.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 7cf430372f89..ec7e4f62aaff 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -911,6 +911,7 @@ static int __devinit workqueue_cpu_callback(struct notifier_block *nfb, unsigned int cpu = (unsigned long)hcpu; struct cpu_workqueue_struct *cwq; struct workqueue_struct *wq; + int ret = NOTIFY_OK; action &= ~CPU_TASKS_FROZEN; @@ -918,7 +919,7 @@ static int __devinit workqueue_cpu_callback(struct notifier_block *nfb, case CPU_UP_PREPARE: cpu_set(cpu, cpu_populated_map); } - +undo: list_for_each_entry(wq, &workqueues, list) { cwq = per_cpu_ptr(wq->cpu_wq, cpu); @@ -928,7 +929,9 @@ static int __devinit workqueue_cpu_callback(struct notifier_block *nfb, break; printk(KERN_ERR "workqueue [%s] for %i failed\n", wq->name, cpu); - return NOTIFY_BAD; + action = CPU_UP_CANCELED; + ret = NOTIFY_BAD; + goto undo; case CPU_ONLINE: start_workqueue_thread(cwq, cpu); @@ -948,7 +951,7 @@ static int __devinit workqueue_cpu_callback(struct notifier_block *nfb, cpu_clear(cpu, cpu_populated_map); } - return NOTIFY_OK; + return ret; } void __init init_workqueues(void) -- cgit v1.2.3 From 339caf2a224fc9af0f01686bf287dda32c6efca6 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Fri, 25 Jul 2008 01:48:31 -0700 Subject: proc: misplaced export of find_get_pid Move EXPORT_SYMBOL right after the func Signed-off-by: David Sterba Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/pid.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/pid.c b/kernel/pid.c index 30bd5d4b2ac7..753fd90d9ec1 100644 --- a/kernel/pid.c +++ b/kernel/pid.c @@ -435,6 +435,7 @@ struct pid *find_get_pid(pid_t nr) return pid; } +EXPORT_SYMBOL_GPL(find_get_pid); pid_t pid_nr_ns(struct pid *pid, struct pid_namespace *ns) { @@ -497,7 +498,6 @@ struct pid *find_ge_pid(int nr, struct pid_namespace *ns) return pid; } -EXPORT_SYMBOL_GPL(find_get_pid); /* * The pid hash table is scaled according to the amount of memory in the -- cgit v1.2.3 From 99541c23cd32bacf1a591ca537a7c0cb9053ad7e Mon Sep 17 00:00:00 2001 From: Alexey Dobriyan Date: Fri, 25 Jul 2008 01:48:31 -0700 Subject: sysctl: check for bogus modes Catch, e. g., 644/0644 typo. Signed-off-by: Alexey Dobriyan Acked-by: "Eric W. Biederman" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sysctl_check.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'kernel') diff --git a/kernel/sysctl_check.c b/kernel/sysctl_check.c index c09350d564f2..c35da23ab8fb 100644 --- a/kernel/sysctl_check.c +++ b/kernel/sysctl_check.c @@ -1532,6 +1532,8 @@ int sysctl_check_table(struct nsproxy *namespaces, struct ctl_table *table) sysctl_check_leaf(namespaces, table, &fail); } sysctl_check_bin_path(table, &fail); + if (table->mode > 0777) + set_fail(&fail, table, "bogus .mode"); if (fail) { set_fail(&fail, table, NULL); error = -EINVAL; -- cgit v1.2.3 From 19b0cfcca41dd772065671ad0584e1cea0f3fd13 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 25 Jul 2008 01:48:35 -0700 Subject: pidns: remove now unused kill_proc function This function operated on a pid_t to kill a task, which is no longer valid in a containerized system. It has finally lost all its users and we can safely remove it from the tree. Signed-off-by: Pavel Emelyanov Cc: Oleg Nesterov Cc: "Eric W. Biederman" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/signal.c | 12 ------------ 1 file changed, 12 deletions(-) (limited to 'kernel') diff --git a/kernel/signal.c b/kernel/signal.c index 5c7b7eaa0dc6..82c3545596c5 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -1228,17 +1228,6 @@ int kill_pid(struct pid *pid, int sig, int priv) } EXPORT_SYMBOL(kill_pid); -int -kill_proc(pid_t pid, int sig, int priv) -{ - int ret; - - rcu_read_lock(); - ret = kill_pid_info(sig, __si_special(priv), find_pid(pid)); - rcu_read_unlock(); - return ret; -} - /* * These functions support sending signals using preallocated sigqueue * structures. This is needed "because realtime applications cannot @@ -1906,7 +1895,6 @@ EXPORT_SYMBOL(recalc_sigpending); EXPORT_SYMBOL_GPL(dequeue_signal); EXPORT_SYMBOL(flush_signals); EXPORT_SYMBOL(force_sig); -EXPORT_SYMBOL(kill_proc); EXPORT_SYMBOL(ptrace_notify); EXPORT_SYMBOL(send_sig); EXPORT_SYMBOL(send_sig_info); -- cgit v1.2.3 From e49859e71e0318b564de1546bdc30fab738f9deb Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 25 Jul 2008 01:48:36 -0700 Subject: pidns: remove now unused find_pid function. This one had the only users so far - the kill_proc, which is removed, so drop this (invalid in namespaced world) call too. And of course - erase all references on it from comments. Signed-off-by: Pavel Emelyanov Cc: Oleg Nesterov Cc: "Eric W. Biederman" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/pid.c | 8 +------- 1 file changed, 1 insertion(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/pid.c b/kernel/pid.c index 753fd90d9ec1..064e76afa507 100644 --- a/kernel/pid.c +++ b/kernel/pid.c @@ -309,12 +309,6 @@ struct pid *find_vpid(int nr) } EXPORT_SYMBOL_GPL(find_vpid); -struct pid *find_pid(int nr) -{ - return find_pid_ns(nr, &init_pid_ns); -} -EXPORT_SYMBOL_GPL(find_pid); - /* * attach_pid() must be called with the tasklist_lock write-held. */ @@ -483,7 +477,7 @@ EXPORT_SYMBOL(task_session_nr_ns); /* * Used by proc to find the first pid that is greater then or equal to nr. * - * If there is a pid at nr this function is exactly the same as find_pid. + * If there is a pid at nr this function is exactly the same as find_pid_ns. */ struct pid *find_ge_pid(int nr, struct pid_namespace *ns) { -- cgit v1.2.3 From 28325df0d9339b7f3aba9c45174d4586223ef46b Mon Sep 17 00:00:00 2001 From: Mathieu Desnoyers Date: Fri, 25 Jul 2008 01:48:38 -0700 Subject: markers: use rcu_barrier_sched() and call_rcu_sched() rcu_barrier_sched() and call_rcu_sched() were introduced in 2.6.26 for the Markers. Change the marker code to use them. It can be seen as a fix since the marker code was using an ugly, temporary, #ifdef hack to work around CONFIG_PREEMPT_RCU. Signed-off-by: Mathieu Desnoyers Acked-by: Paul McKenney Cc: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/marker.c | 25 ++++++++----------------- 1 file changed, 8 insertions(+), 17 deletions(-) (limited to 'kernel') diff --git a/kernel/marker.c b/kernel/marker.c index 1abfb923b761..971da5317903 100644 --- a/kernel/marker.c +++ b/kernel/marker.c @@ -441,7 +441,7 @@ static int remove_marker(const char *name) hlist_del(&e->hlist); /* Make sure the call_rcu has been executed */ if (e->rcu_pending) - rcu_barrier(); + rcu_barrier_sched(); kfree(e); return 0; } @@ -476,7 +476,7 @@ static int marker_set_format(struct marker_entry **entry, const char *format) hlist_del(&(*entry)->hlist); /* Make sure the call_rcu has been executed */ if ((*entry)->rcu_pending) - rcu_barrier(); + rcu_barrier_sched(); kfree(*entry); *entry = e; trace_mark(core_marker_format, "name %s format %s", @@ -655,7 +655,7 @@ int marker_probe_register(const char *name, const char *format, * make sure it's executed now. */ if (entry->rcu_pending) - rcu_barrier(); + rcu_barrier_sched(); old = marker_entry_add_probe(entry, probe, probe_private); if (IS_ERR(old)) { ret = PTR_ERR(old); @@ -670,10 +670,7 @@ int marker_probe_register(const char *name, const char *format, entry->rcu_pending = 1; /* write rcu_pending before calling the RCU callback */ smp_wmb(); -#ifdef CONFIG_PREEMPT_RCU - synchronize_sched(); /* Until we have the call_rcu_sched() */ -#endif - call_rcu(&entry->rcu, free_old_closure); + call_rcu_sched(&entry->rcu, free_old_closure); end: mutex_unlock(&markers_mutex); return ret; @@ -704,7 +701,7 @@ int marker_probe_unregister(const char *name, if (!entry) goto end; if (entry->rcu_pending) - rcu_barrier(); + rcu_barrier_sched(); old = marker_entry_remove_probe(entry, probe, probe_private); mutex_unlock(&markers_mutex); marker_update_probes(); /* may update entry */ @@ -716,10 +713,7 @@ int marker_probe_unregister(const char *name, entry->rcu_pending = 1; /* write rcu_pending before calling the RCU callback */ smp_wmb(); -#ifdef CONFIG_PREEMPT_RCU - synchronize_sched(); /* Until we have the call_rcu_sched() */ -#endif - call_rcu(&entry->rcu, free_old_closure); + call_rcu_sched(&entry->rcu, free_old_closure); remove_marker(name); /* Ignore busy error message */ ret = 0; end: @@ -786,7 +780,7 @@ int marker_probe_unregister_private_data(marker_probe_func *probe, goto end; } if (entry->rcu_pending) - rcu_barrier(); + rcu_barrier_sched(); old = marker_entry_remove_probe(entry, NULL, probe_private); mutex_unlock(&markers_mutex); marker_update_probes(); /* may update entry */ @@ -797,10 +791,7 @@ int marker_probe_unregister_private_data(marker_probe_func *probe, entry->rcu_pending = 1; /* write rcu_pending before calling the RCU callback */ smp_wmb(); -#ifdef CONFIG_PREEMPT_RCU - synchronize_sched(); /* Until we have the call_rcu_sched() */ -#endif - call_rcu(&entry->rcu, free_old_closure); + call_rcu_sched(&entry->rcu, free_old_closure); remove_marker(entry->name); /* Ignore busy error message */ end: mutex_unlock(&markers_mutex); -- cgit v1.2.3 From a89cc1959d0ea5f36bf7421dc97b34f03809637d Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Fri, 25 Jul 2008 01:48:39 -0700 Subject: markers: fix sparse integer as NULL pointer warning kernel/trace/trace_sysprof.c:164:20: warning: Using plain integer as NULL pointer Signed-off-by: Harvey Harrison Cc: Mathieu Desnoyers Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/trace/trace_sysprof.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/trace/trace_sysprof.c b/kernel/trace/trace_sysprof.c index 63528086337c..ce2d723c10e1 100644 --- a/kernel/trace/trace_sysprof.c +++ b/kernel/trace/trace_sysprof.c @@ -161,7 +161,7 @@ static void timer_notify(struct pt_regs *regs, int cpu) __trace_special(tr, data, 2, regs->ip, 0); while (i < sample_max_depth) { - frame.next_fp = 0; + frame.next_fp = NULL; frame.return_address = 0; if (!copy_stack_frame(fp, &frame)) break; -- cgit v1.2.3 From 7394f0f6c0baab650ea9194cb1be847df646fb57 Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Fri, 25 Jul 2008 01:48:40 -0700 Subject: unexport uts_sem With the removal of the Solaris binary emulation the export of uts_sem became unused. Signed-off-by: Adrian Bunk Acked-by: David S. Miller Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sys.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'kernel') diff --git a/kernel/sys.c b/kernel/sys.c index 6c2188046048..0c9d3fa1f5ff 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -1343,8 +1343,6 @@ EXPORT_SYMBOL(in_egroup_p); DECLARE_RWSEM(uts_sem); -EXPORT_SYMBOL(uts_sem); - asmlinkage long sys_newuname(struct new_utsname __user * name) { int errno = 0; -- cgit v1.2.3 From 49b5cf34727a6c1be1568ab28e89a2d9a6bf51e0 Mon Sep 17 00:00:00 2001 From: Jonathan Lim Date: Fri, 25 Jul 2008 01:48:40 -0700 Subject: accounting: account for user time when updating memory integrals Adapt acct_update_integrals() to include user time when calculating the time difference. The units of acct_rss_mem1 and acct_vm_mem1 are also changed from pages-jiffies to pages-usecs to avoid calling jiffies_to_usecs() in xacct_add_tsk() which might overflow. Signed-off-by: Jonathan Lim Cc: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sched.c | 2 ++ kernel/tsacct.c | 21 ++++++++++++++------- 2 files changed, 16 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index 6acf749d3336..0047bd9b96aa 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4046,6 +4046,8 @@ void account_user_time(struct task_struct *p, cputime_t cputime) cpustat->nice = cputime64_add(cpustat->nice, tmp); else cpustat->user = cputime64_add(cpustat->user, tmp); + /* Account for user time used */ + acct_update_integrals(p); } /* diff --git a/kernel/tsacct.c b/kernel/tsacct.c index 4ab1b584961b..1da6990af8e0 100644 --- a/kernel/tsacct.c +++ b/kernel/tsacct.c @@ -84,9 +84,9 @@ void xacct_add_tsk(struct taskstats *stats, struct task_struct *p) { struct mm_struct *mm; - /* convert pages-jiffies to Mbyte-usec */ - stats->coremem = jiffies_to_usecs(p->acct_rss_mem1) * PAGE_SIZE / MB; - stats->virtmem = jiffies_to_usecs(p->acct_vm_mem1) * PAGE_SIZE / MB; + /* convert pages-usec to Mbyte-usec */ + stats->coremem = p->acct_rss_mem1 * PAGE_SIZE / MB; + stats->virtmem = p->acct_vm_mem1 * PAGE_SIZE / MB; mm = get_task_mm(p); if (mm) { /* adjust to KB unit */ @@ -118,12 +118,19 @@ void xacct_add_tsk(struct taskstats *stats, struct task_struct *p) void acct_update_integrals(struct task_struct *tsk) { if (likely(tsk->mm)) { - long delta = cputime_to_jiffies( - cputime_sub(tsk->stime, tsk->acct_stimexpd)); + cputime_t time, dtime; + struct timeval value; + u64 delta; + + time = tsk->stime + tsk->utime; + dtime = cputime_sub(time, tsk->acct_timexpd); + jiffies_to_timeval(cputime_to_jiffies(dtime), &value); + delta = value.tv_sec; + delta = delta * USEC_PER_SEC + value.tv_usec; if (delta == 0) return; - tsk->acct_stimexpd = tsk->stime; + tsk->acct_timexpd = time; tsk->acct_rss_mem1 += delta * get_mm_rss(tsk->mm); tsk->acct_vm_mem1 += delta * tsk->mm->total_vm; } @@ -135,7 +142,7 @@ void acct_update_integrals(struct task_struct *tsk) */ void acct_clear_integrals(struct task_struct *tsk) { - tsk->acct_stimexpd = 0; + tsk->acct_timexpd = 0; tsk->acct_rss_mem1 = 0; tsk->acct_vm_mem1 = 0; } -- cgit v1.2.3 From 081e4c8a75692c21f3a119a81ca3270081879d0e Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 25 Jul 2008 01:48:42 -0700 Subject: bsdacct: rename acct_gbls to bsd_acct_struct After I fixed access to task->tgid in kernel/acct.c, Oleg pointed out some bad side effects with this accounting vs pid namespaces interaction. I.e. when some task in pid namespace sets this accounting up, this blocks all the others from doing the same. Restricting this to init namespace only could help, but didn't look a graceful solution. So here is the approach to make this accounting work with pid namespaces properly. The idea is simple - when a task dies it accounts itself in each namespace it is visible from and which set the accounting up. For example here are the commands run and the output of lastcomm from init and sub namespaces: init_ns# accton pacct sub_ns# accton pacct (this is a different file - sub ns is run in a chroot-ed environment) init_ns# cat /dev/null sub_ns# ls /dev/null init_ns# accton sub_ns# accton sub_ns# lastcomm -f pacct ls 0 [136,0] 0.00 secs Thu May 15 10:30 accton 0 [136,0] 0.00 secs Thu May 15 10:30 init_ns# lastcomm -f pacct accton root pts/0 0.00 secs Thu May 15 14:30 << got from sub cat root pts/1 0.00 secs Thu May 15 14:30 ls root pts/0 0.00 secs Thu May 15 14:30 << got from sub accton root pts/1 0.00 secs Thu May 15 14:30 That was the summary, the details are in patches. This patch: It will be visible in pid_namespace.h file, so fix its name to look better outside the acct.c file. Signed-off-by: Pavel Emelyanov Cc: Balbir Singh Cc: "Eric W. Biederman" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/acct.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/acct.c b/kernel/acct.c index 91e1cfd734d2..ee3e605190f9 100644 --- a/kernel/acct.c +++ b/kernel/acct.c @@ -82,7 +82,7 @@ static void do_acct_process(struct pid_namespace *ns, struct file *); * can be placed in the same cache line as the lock. This primes * the cache line to have the data after getting the lock. */ -struct acct_glbs { +struct bsd_acct_struct { spinlock_t lock; volatile int active; volatile int needcheck; @@ -91,7 +91,7 @@ struct acct_glbs { struct timer_list timer; }; -static struct acct_glbs acct_globals __cacheline_aligned = +static struct bsd_acct_struct acct_globals __cacheline_aligned = {__SPIN_LOCK_UNLOCKED(acct_globals.lock)}; /* -- cgit v1.2.3 From 84406c153a5bfa5d8b428a0933e9d39db6b59a75 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 25 Jul 2008 01:48:42 -0700 Subject: pidns: use kzalloc when allocating new pid_namespace struct It makes many fields initialization implicit helping in auto-setting #ifdef-ed fields (bsd-acct related pointer will be such). Signed-off-by: Pavel Emelyanov Cc: Balbir Singh Cc: "Eric W. Biederman" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/pid_namespace.c | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/pid_namespace.c b/kernel/pid_namespace.c index 98702b4b8851..06331cc1c3f5 100644 --- a/kernel/pid_namespace.c +++ b/kernel/pid_namespace.c @@ -71,7 +71,7 @@ static struct pid_namespace *create_pid_namespace(unsigned int level) struct pid_namespace *ns; int i; - ns = kmem_cache_alloc(pid_ns_cachep, GFP_KERNEL); + ns = kmem_cache_zalloc(pid_ns_cachep, GFP_KERNEL); if (ns == NULL) goto out; @@ -84,17 +84,13 @@ static struct pid_namespace *create_pid_namespace(unsigned int level) goto out_free_map; kref_init(&ns->kref); - ns->last_pid = 0; - ns->child_reaper = NULL; ns->level = level; set_bit(0, ns->pidmap[0].page); atomic_set(&ns->pidmap[0].nr_free, BITS_PER_PAGE - 1); - for (i = 1; i < PIDMAP_ENTRIES; i++) { - ns->pidmap[i].page = NULL; + for (i = 1; i < PIDMAP_ENTRIES; i++) atomic_set(&ns->pidmap[i].nr_free, BITS_PER_PAGE); - } return ns; -- cgit v1.2.3 From 1c552858ac2b1732a99d234d46b98098baef41ff Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 25 Jul 2008 01:48:44 -0700 Subject: bsdacct: "truthify" a comment near acct_process The acct_process does not accept any arguments actually. Signed-off-by: Pavel Emelyanov Cc: Balbir Singh Cc: "Eric W. Biederman" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/acct.c | 1 - 1 file changed, 1 deletion(-) (limited to 'kernel') diff --git a/kernel/acct.c b/kernel/acct.c index ee3e605190f9..d9ee1838b4d4 100644 --- a/kernel/acct.c +++ b/kernel/acct.c @@ -579,7 +579,6 @@ void acct_collect(long exitcode, int group_dead) /** * acct_process - now just a wrapper around do_acct_process - * @exitcode: task exit code * * handles process accounting for an exiting task */ -- cgit v1.2.3 From e59a04a7aa5ce2483470aee4f2eb79ba6b9afe8b Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 25 Jul 2008 01:48:44 -0700 Subject: bsdacct: make check timer accept a bsd_acct_struct argument We're going to have many bsd_acct_struct instances, not just one, so the timer (currently working with a global one) has to know which one to work with. Use a handy setup_timer macro for it (thanks to Oleg for one). Signed-off-by: Pavel Emelyanov Cc: Balbir Singh Cc: "Eric W. Biederman" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/acct.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/acct.c b/kernel/acct.c index d9ee1838b4d4..05f8bc094a4b 100644 --- a/kernel/acct.c +++ b/kernel/acct.c @@ -97,9 +97,10 @@ static struct bsd_acct_struct acct_globals __cacheline_aligned = /* * Called whenever the timer says to check the free space. */ -static void acct_timeout(unsigned long unused) +static void acct_timeout(unsigned long x) { - acct_globals.needcheck = 1; + struct bsd_acct_struct *acct = (struct bsd_acct_struct *)x; + acct->needcheck = 1; } /* @@ -193,8 +194,8 @@ static void acct_file_reopen(struct file *file) acct_globals.needcheck = 0; acct_globals.active = 1; /* It's been deleted if it was used before so this is safe */ - init_timer(&acct_globals.timer); - acct_globals.timer.function = acct_timeout; + setup_timer(&acct_globals.timer, acct_timeout, + (unsigned long)&acct_globals); acct_globals.timer.expires = jiffies + ACCT_TIMEOUT*HZ; add_timer(&acct_globals.timer); } -- cgit v1.2.3 From a75d97976517dcda69150fd81d6be86ae63324a1 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 25 Jul 2008 01:48:45 -0700 Subject: bsdacct: turn the acct_lock from on-the-struct to global Don't use per-bsd-acct-struct lock, but work with a global one. This lock is taken for short periods, so it doesn't seem it'll become a bottleneck, but it will allow us to easily avoid many locking difficulties in the future. So this is a mostly s/acct_globals.lock/acct_lock/ over the file. Signed-off-by: Pavel Emelyanov Cc: Balbir Singh Cc: "Eric W. Biederman" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/acct.c | 42 +++++++++++++++++++++--------------------- 1 file changed, 21 insertions(+), 21 deletions(-) (limited to 'kernel') diff --git a/kernel/acct.c b/kernel/acct.c index 05f8bc094a4b..fc71c1304977 100644 --- a/kernel/acct.c +++ b/kernel/acct.c @@ -83,7 +83,6 @@ static void do_acct_process(struct pid_namespace *ns, struct file *); * the cache line to have the data after getting the lock. */ struct bsd_acct_struct { - spinlock_t lock; volatile int active; volatile int needcheck; struct file *file; @@ -91,8 +90,9 @@ struct bsd_acct_struct { struct timer_list timer; }; -static struct bsd_acct_struct acct_globals __cacheline_aligned = - {__SPIN_LOCK_UNLOCKED(acct_globals.lock)}; +static DEFINE_SPINLOCK(acct_lock); + +static struct bsd_acct_struct acct_globals __cacheline_aligned; /* * Called whenever the timer says to check the free space. @@ -114,11 +114,11 @@ static int check_free_space(struct file *file) sector_t resume; sector_t suspend; - spin_lock(&acct_globals.lock); + spin_lock(&acct_lock); res = acct_globals.active; if (!file || !acct_globals.needcheck) goto out; - spin_unlock(&acct_globals.lock); + spin_unlock(&acct_lock); /* May block */ if (vfs_statfs(file->f_path.dentry, &sbuf)) @@ -140,7 +140,7 @@ static int check_free_space(struct file *file) * If some joker switched acct_globals.file under us we'ld better be * silent and _not_ touch anything. */ - spin_lock(&acct_globals.lock); + spin_lock(&acct_lock); if (file != acct_globals.file) { if (act) res = act>0; @@ -165,7 +165,7 @@ static int check_free_space(struct file *file) add_timer(&acct_globals.timer); res = acct_globals.active; out: - spin_unlock(&acct_globals.lock); + spin_unlock(&acct_lock); return res; } @@ -173,7 +173,7 @@ out: * Close the old accounting file (if currently open) and then replace * it with file (if non-NULL). * - * NOTE: acct_globals.lock MUST be held on entry and exit. + * NOTE: acct_lock MUST be held on entry and exit. */ static void acct_file_reopen(struct file *file) { @@ -201,11 +201,11 @@ static void acct_file_reopen(struct file *file) } if (old_acct) { mnt_unpin(old_acct->f_path.mnt); - spin_unlock(&acct_globals.lock); + spin_unlock(&acct_lock); do_acct_process(old_ns, old_acct); filp_close(old_acct, NULL); put_pid_ns(old_ns); - spin_lock(&acct_globals.lock); + spin_lock(&acct_lock); } } @@ -235,10 +235,10 @@ static int acct_on(char *name) return error; } - spin_lock(&acct_globals.lock); + spin_lock(&acct_lock); mnt_pin(file->f_path.mnt); acct_file_reopen(file); - spin_unlock(&acct_globals.lock); + spin_unlock(&acct_lock); mntput(file->f_path.mnt); /* it's pinned, now give up active reference */ @@ -272,9 +272,9 @@ asmlinkage long sys_acct(const char __user *name) } else { error = security_acct(NULL); if (!error) { - spin_lock(&acct_globals.lock); + spin_lock(&acct_lock); acct_file_reopen(NULL); - spin_unlock(&acct_globals.lock); + spin_unlock(&acct_lock); } } return error; @@ -289,10 +289,10 @@ asmlinkage long sys_acct(const char __user *name) */ void acct_auto_close_mnt(struct vfsmount *m) { - spin_lock(&acct_globals.lock); + spin_lock(&acct_lock); if (acct_globals.file && acct_globals.file->f_path.mnt == m) acct_file_reopen(NULL); - spin_unlock(&acct_globals.lock); + spin_unlock(&acct_lock); } /** @@ -304,12 +304,12 @@ void acct_auto_close_mnt(struct vfsmount *m) */ void acct_auto_close(struct super_block *sb) { - spin_lock(&acct_globals.lock); + spin_lock(&acct_lock); if (acct_globals.file && acct_globals.file->f_path.mnt->mnt_sb == sb) { acct_file_reopen(NULL); } - spin_unlock(&acct_globals.lock); + spin_unlock(&acct_lock); } /* @@ -594,15 +594,15 @@ void acct_process(void) if (!acct_globals.file) return; - spin_lock(&acct_globals.lock); + spin_lock(&acct_lock); file = acct_globals.file; if (unlikely(!file)) { - spin_unlock(&acct_globals.lock); + spin_unlock(&acct_lock); return; } get_file(file); ns = get_pid_ns(acct_globals.ns); - spin_unlock(&acct_globals.lock); + spin_unlock(&acct_lock); do_acct_process(ns, file); fput(file); -- cgit v1.2.3 From 6248b1b342005a428b1247b4e89249da1528d88d Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 25 Jul 2008 01:48:46 -0700 Subject: bsdacct: make internal code work with passed bsd_acct_struct, not global This adds the appropriate pointer to all the internal (i.e. static) functions that work with global acct instance. API calls pass a global instance to them (while we still have such). Mostly this is a s/acct_globals./acct->/ over the file. Signed-off-by: Pavel Emelyanov Cc: Balbir Singh Cc: "Eric W. Biederman" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/acct.c | 77 ++++++++++++++++++++++++++++++----------------------------- 1 file changed, 39 insertions(+), 38 deletions(-) (limited to 'kernel') diff --git a/kernel/acct.c b/kernel/acct.c index fc71c1304977..72d4760c8da8 100644 --- a/kernel/acct.c +++ b/kernel/acct.c @@ -75,7 +75,8 @@ int acct_parm[3] = {4, 2, 30}; /* * External references and all of the globals. */ -static void do_acct_process(struct pid_namespace *ns, struct file *); +static void do_acct_process(struct bsd_acct_struct *acct, + struct pid_namespace *ns, struct file *); /* * This structure is used so that all the data protected by lock @@ -106,7 +107,7 @@ static void acct_timeout(unsigned long x) /* * Check the amount of free space and suspend/resume accordingly. */ -static int check_free_space(struct file *file) +static int check_free_space(struct bsd_acct_struct *acct, struct file *file) { struct kstatfs sbuf; int res; @@ -115,8 +116,8 @@ static int check_free_space(struct file *file) sector_t suspend; spin_lock(&acct_lock); - res = acct_globals.active; - if (!file || !acct_globals.needcheck) + res = acct->active; + if (!file || !acct->needcheck) goto out; spin_unlock(&acct_lock); @@ -137,33 +138,33 @@ static int check_free_space(struct file *file) act = 0; /* - * If some joker switched acct_globals.file under us we'ld better be + * If some joker switched acct->file under us we'ld better be * silent and _not_ touch anything. */ spin_lock(&acct_lock); - if (file != acct_globals.file) { + if (file != acct->file) { if (act) res = act>0; goto out; } - if (acct_globals.active) { + if (acct->active) { if (act < 0) { - acct_globals.active = 0; + acct->active = 0; printk(KERN_INFO "Process accounting paused\n"); } } else { if (act > 0) { - acct_globals.active = 1; + acct->active = 1; printk(KERN_INFO "Process accounting resumed\n"); } } - del_timer(&acct_globals.timer); - acct_globals.needcheck = 0; - acct_globals.timer.expires = jiffies + ACCT_TIMEOUT*HZ; - add_timer(&acct_globals.timer); - res = acct_globals.active; + del_timer(&acct->timer); + acct->needcheck = 0; + acct->timer.expires = jiffies + ACCT_TIMEOUT*HZ; + add_timer(&acct->timer); + res = acct->active; out: spin_unlock(&acct_lock); return res; @@ -175,34 +176,33 @@ out: * * NOTE: acct_lock MUST be held on entry and exit. */ -static void acct_file_reopen(struct file *file) +static void acct_file_reopen(struct bsd_acct_struct *acct, struct file *file) { struct file *old_acct = NULL; struct pid_namespace *old_ns = NULL; - if (acct_globals.file) { - old_acct = acct_globals.file; - old_ns = acct_globals.ns; - del_timer(&acct_globals.timer); - acct_globals.active = 0; - acct_globals.needcheck = 0; - acct_globals.file = NULL; + if (acct->file) { + old_acct = acct->file; + old_ns = acct->ns; + del_timer(&acct->timer); + acct->active = 0; + acct->needcheck = 0; + acct->file = NULL; } if (file) { - acct_globals.file = file; - acct_globals.ns = get_pid_ns(task_active_pid_ns(current)); - acct_globals.needcheck = 0; - acct_globals.active = 1; + acct->file = file; + acct->ns = get_pid_ns(task_active_pid_ns(current)); + acct->needcheck = 0; + acct->active = 1; /* It's been deleted if it was used before so this is safe */ - setup_timer(&acct_globals.timer, acct_timeout, - (unsigned long)&acct_globals); - acct_globals.timer.expires = jiffies + ACCT_TIMEOUT*HZ; - add_timer(&acct_globals.timer); + setup_timer(&acct->timer, acct_timeout, (unsigned long)acct); + acct->timer.expires = jiffies + ACCT_TIMEOUT*HZ; + add_timer(&acct->timer); } if (old_acct) { mnt_unpin(old_acct->f_path.mnt); spin_unlock(&acct_lock); - do_acct_process(old_ns, old_acct); + do_acct_process(acct, old_ns, old_acct); filp_close(old_acct, NULL); put_pid_ns(old_ns); spin_lock(&acct_lock); @@ -237,7 +237,7 @@ static int acct_on(char *name) spin_lock(&acct_lock); mnt_pin(file->f_path.mnt); - acct_file_reopen(file); + acct_file_reopen(&acct_globals, file); spin_unlock(&acct_lock); mntput(file->f_path.mnt); /* it's pinned, now give up active reference */ @@ -273,7 +273,7 @@ asmlinkage long sys_acct(const char __user *name) error = security_acct(NULL); if (!error) { spin_lock(&acct_lock); - acct_file_reopen(NULL); + acct_file_reopen(&acct_globals, NULL); spin_unlock(&acct_lock); } } @@ -291,7 +291,7 @@ void acct_auto_close_mnt(struct vfsmount *m) { spin_lock(&acct_lock); if (acct_globals.file && acct_globals.file->f_path.mnt == m) - acct_file_reopen(NULL); + acct_file_reopen(&acct_globals, NULL); spin_unlock(&acct_lock); } @@ -307,7 +307,7 @@ void acct_auto_close(struct super_block *sb) spin_lock(&acct_lock); if (acct_globals.file && acct_globals.file->f_path.mnt->mnt_sb == sb) { - acct_file_reopen(NULL); + acct_file_reopen(&acct_globals, NULL); } spin_unlock(&acct_lock); } @@ -426,7 +426,8 @@ static u32 encode_float(u64 value) /* * do_acct_process does all actual work. Caller holds the reference to file. */ -static void do_acct_process(struct pid_namespace *ns, struct file *file) +static void do_acct_process(struct bsd_acct_struct *acct, + struct pid_namespace *ns, struct file *file) { struct pacct_struct *pacct = ¤t->signal->pacct; acct_t ac; @@ -441,7 +442,7 @@ static void do_acct_process(struct pid_namespace *ns, struct file *file) * First check to see if there is enough free_space to continue * the process accounting system. */ - if (!check_free_space(file)) + if (!check_free_space(acct, file)) return; /* @@ -604,7 +605,7 @@ void acct_process(void) ns = get_pid_ns(acct_globals.ns); spin_unlock(&acct_lock); - do_acct_process(ns, file); + do_acct_process(&acct_globals, ns, file); fput(file); put_pid_ns(ns); } -- cgit v1.2.3 From 0b6b030fc30d169bb406b34b4fc60d99dde4a9c6 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 25 Jul 2008 01:48:47 -0700 Subject: bsdacct: switch from global bsd_acct_struct instance to per-pidns one Allocate the structure on the first call to sys_acct(). After this each namespace, that ordered the accounting, will live with this structure till its own death. Two notes - routines, that close the accounting on fs umount time use the init_pid_ns's acct by now; - accounting routine accounts to dying task's namespace (also by now). Signed-off-by: Pavel Emelyanov Cc: Balbir Singh Cc: "Eric W. Biederman" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/acct.c | 84 +++++++++++++++++++++++++++++++++++++++----------- kernel/pid_namespace.c | 2 ++ 2 files changed, 68 insertions(+), 18 deletions(-) (limited to 'kernel') diff --git a/kernel/acct.c b/kernel/acct.c index 72d4760c8da8..febbbc67157e 100644 --- a/kernel/acct.c +++ b/kernel/acct.c @@ -93,8 +93,6 @@ struct bsd_acct_struct { static DEFINE_SPINLOCK(acct_lock); -static struct bsd_acct_struct acct_globals __cacheline_aligned; - /* * Called whenever the timer says to check the free space. */ @@ -176,7 +174,8 @@ out: * * NOTE: acct_lock MUST be held on entry and exit. */ -static void acct_file_reopen(struct bsd_acct_struct *acct, struct file *file) +static void acct_file_reopen(struct bsd_acct_struct *acct, struct file *file, + struct pid_namespace *ns) { struct file *old_acct = NULL; struct pid_namespace *old_ns = NULL; @@ -188,10 +187,11 @@ static void acct_file_reopen(struct bsd_acct_struct *acct, struct file *file) acct->active = 0; acct->needcheck = 0; acct->file = NULL; + acct->ns = NULL; } if (file) { acct->file = file; - acct->ns = get_pid_ns(task_active_pid_ns(current)); + acct->ns = ns; acct->needcheck = 0; acct->active = 1; /* It's been deleted if it was used before so this is safe */ @@ -204,7 +204,6 @@ static void acct_file_reopen(struct bsd_acct_struct *acct, struct file *file) spin_unlock(&acct_lock); do_acct_process(acct, old_ns, old_acct); filp_close(old_acct, NULL); - put_pid_ns(old_ns); spin_lock(&acct_lock); } } @@ -213,6 +212,8 @@ static int acct_on(char *name) { struct file *file; int error; + struct pid_namespace *ns; + struct bsd_acct_struct *acct = NULL; /* Difference from BSD - they don't do O_APPEND */ file = filp_open(name, O_WRONLY|O_APPEND|O_LARGEFILE, 0); @@ -229,18 +230,34 @@ static int acct_on(char *name) return -EIO; } + ns = task_active_pid_ns(current); + if (ns->bacct == NULL) { + acct = kzalloc(sizeof(struct bsd_acct_struct), GFP_KERNEL); + if (acct == NULL) { + filp_close(file, NULL); + return -ENOMEM; + } + } + error = security_acct(file); if (error) { + kfree(acct); filp_close(file, NULL); return error; } spin_lock(&acct_lock); + if (ns->bacct == NULL) { + ns->bacct = acct; + acct = NULL; + } + mnt_pin(file->f_path.mnt); - acct_file_reopen(&acct_globals, file); + acct_file_reopen(ns->bacct, file, ns); spin_unlock(&acct_lock); mntput(file->f_path.mnt); /* it's pinned, now give up active reference */ + kfree(acct); return 0; } @@ -270,10 +287,16 @@ asmlinkage long sys_acct(const char __user *name) error = acct_on(tmp); putname(tmp); } else { + struct bsd_acct_struct *acct; + + acct = task_active_pid_ns(current)->bacct; + if (acct == NULL) + return 0; + error = security_acct(NULL); if (!error) { spin_lock(&acct_lock); - acct_file_reopen(&acct_globals, NULL); + acct_file_reopen(acct, NULL, NULL); spin_unlock(&acct_lock); } } @@ -289,9 +312,15 @@ asmlinkage long sys_acct(const char __user *name) */ void acct_auto_close_mnt(struct vfsmount *m) { + struct bsd_acct_struct *acct; + + acct = init_pid_ns.bacct; + if (acct == NULL) + return; + spin_lock(&acct_lock); - if (acct_globals.file && acct_globals.file->f_path.mnt == m) - acct_file_reopen(&acct_globals, NULL); + if (acct->file && acct->file->f_path.mnt == m) + acct_file_reopen(acct, NULL, NULL); spin_unlock(&acct_lock); } @@ -304,10 +333,29 @@ void acct_auto_close_mnt(struct vfsmount *m) */ void acct_auto_close(struct super_block *sb) { + struct bsd_acct_struct *acct; + + acct = init_pid_ns.bacct; + if (acct == NULL) + return; + spin_lock(&acct_lock); - if (acct_globals.file && - acct_globals.file->f_path.mnt->mnt_sb == sb) { - acct_file_reopen(&acct_globals, NULL); + if (acct->file && acct->file->f_path.mnt->mnt_sb == sb) + acct_file_reopen(acct, NULL, NULL); + spin_unlock(&acct_lock); +} + +void acct_exit_ns(struct pid_namespace *ns) +{ + struct bsd_acct_struct *acct; + + spin_lock(&acct_lock); + acct = ns->bacct; + if (acct != NULL) { + if (acct->file != NULL) + acct_file_reopen(acct, NULL, NULL); + + kfree(acct); } spin_unlock(&acct_lock); } @@ -587,25 +635,25 @@ void acct_collect(long exitcode, int group_dead) void acct_process(void) { struct file *file = NULL; - struct pid_namespace *ns; + struct pid_namespace *ns = task_active_pid_ns(current); + struct bsd_acct_struct *acct; + acct = ns->bacct; /* * accelerate the common fastpath: */ - if (!acct_globals.file) + if (!acct || !acct->file) return; spin_lock(&acct_lock); - file = acct_globals.file; + file = acct->file; if (unlikely(!file)) { spin_unlock(&acct_lock); return; } get_file(file); - ns = get_pid_ns(acct_globals.ns); spin_unlock(&acct_lock); - do_acct_process(&acct_globals, ns, file); + do_acct_process(acct, ns, file); fput(file); - put_pid_ns(ns); } diff --git a/kernel/pid_namespace.c b/kernel/pid_namespace.c index 06331cc1c3f5..ea567b78d1aa 100644 --- a/kernel/pid_namespace.c +++ b/kernel/pid_namespace.c @@ -12,6 +12,7 @@ #include #include #include +#include #define BITS_PER_PAGE (PAGE_SIZE*8) @@ -181,6 +182,7 @@ void zap_pid_ns_processes(struct pid_namespace *pid_ns) /* Child reaper for the pid namespace is going away */ pid_ns->child_reaper = NULL; + acct_exit_ns(pid_ns); return; } -- cgit v1.2.3 From b5a7174875ea570cc675f2c503e800db8efdd6a7 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 25 Jul 2008 01:48:47 -0700 Subject: bsdacct: turn acct off for all pidns-s on umount time All the bsd_acct_strcts with opened accounting are linked into a global list. So, the acct_auto_close(_mnt) walks one and drops the accounting for each. Signed-off-by: Pavel Emelyanov Cc: Balbir Singh Cc: "Eric W. Biederman" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/acct.c | 28 ++++++++++++++++------------ 1 file changed, 16 insertions(+), 12 deletions(-) (limited to 'kernel') diff --git a/kernel/acct.c b/kernel/acct.c index febbbc67157e..7fc9f9dd1e9e 100644 --- a/kernel/acct.c +++ b/kernel/acct.c @@ -89,9 +89,11 @@ struct bsd_acct_struct { struct file *file; struct pid_namespace *ns; struct timer_list timer; + struct list_head list; }; static DEFINE_SPINLOCK(acct_lock); +static LIST_HEAD(acct_list); /* * Called whenever the timer says to check the free space. @@ -188,12 +190,14 @@ static void acct_file_reopen(struct bsd_acct_struct *acct, struct file *file, acct->needcheck = 0; acct->file = NULL; acct->ns = NULL; + list_del(&acct->list); } if (file) { acct->file = file; acct->ns = ns; acct->needcheck = 0; acct->active = 1; + list_add(&acct->list, &acct_list); /* It's been deleted if it was used before so this is safe */ setup_timer(&acct->timer, acct_timeout, (unsigned long)acct); acct->timer.expires = jiffies + ACCT_TIMEOUT*HZ; @@ -314,13 +318,13 @@ void acct_auto_close_mnt(struct vfsmount *m) { struct bsd_acct_struct *acct; - acct = init_pid_ns.bacct; - if (acct == NULL) - return; - spin_lock(&acct_lock); - if (acct->file && acct->file->f_path.mnt == m) - acct_file_reopen(acct, NULL, NULL); +restart: + list_for_each_entry(acct, &acct_list, list) + if (acct->file && acct->file->f_path.mnt == m) { + acct_file_reopen(acct, NULL, NULL); + goto restart; + } spin_unlock(&acct_lock); } @@ -335,13 +339,13 @@ void acct_auto_close(struct super_block *sb) { struct bsd_acct_struct *acct; - acct = init_pid_ns.bacct; - if (acct == NULL) - return; - spin_lock(&acct_lock); - if (acct->file && acct->file->f_path.mnt->mnt_sb == sb) - acct_file_reopen(acct, NULL, NULL); +restart: + list_for_each_entry(acct, &acct_list, list) + if (acct->file && acct->file->f_path.mnt->mnt_sb == sb) { + acct_file_reopen(acct, NULL, NULL); + goto restart; + } spin_unlock(&acct_lock); } -- cgit v1.2.3 From 7d1e13505be8c2bd2207894f4e0f069e1f9b51c9 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 25 Jul 2008 01:48:48 -0700 Subject: bsdacct: account dying tasks in all relevant namespaces This just makes the acct_proces walk the pid namespaces from current up to the top and account a task in each with the accounting turned on. ns->parent access if safe lockless, since current it still alive and holds its namespace, which in turn holds its parent. Signed-off-by: Pavel Emelyanov Cc: Balbir Singh Cc: "Eric W. Biederman" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/acct.c | 21 ++++++++++++++------- 1 file changed, 14 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/acct.c b/kernel/acct.c index 7fc9f9dd1e9e..0feba97e114e 100644 --- a/kernel/acct.c +++ b/kernel/acct.c @@ -631,15 +631,9 @@ void acct_collect(long exitcode, int group_dead) spin_unlock_irq(¤t->sighand->siglock); } -/** - * acct_process - now just a wrapper around do_acct_process - * - * handles process accounting for an exiting task - */ -void acct_process(void) +static void acct_process_in_ns(struct pid_namespace *ns) { struct file *file = NULL; - struct pid_namespace *ns = task_active_pid_ns(current); struct bsd_acct_struct *acct; acct = ns->bacct; @@ -661,3 +655,16 @@ void acct_process(void) do_acct_process(acct, ns, file); fput(file); } + +/** + * acct_process - now just a wrapper around do_acct_process + * + * handles process accounting for an exiting task + */ +void acct_process(void) +{ + struct pid_namespace *ns; + + for (ns = task_active_pid_ns(current); ns != NULL; ns = ns->parent) + acct_process_in_ns(ns); +} -- cgit v1.2.3 From 0c18d7a5df82524e634637c3aec24d4cba096442 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 25 Jul 2008 01:48:49 -0700 Subject: bsdacct: fix and add comments around acct_process() Fix the one describing what this function is and add one more - about locking absence around pid namespaces loop. Signed-off-by: Pavel Emelyanov Cc: Randy Dunlap Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/acct.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/acct.c b/kernel/acct.c index 0feba97e114e..dd68b9059418 100644 --- a/kernel/acct.c +++ b/kernel/acct.c @@ -657,7 +657,8 @@ static void acct_process_in_ns(struct pid_namespace *ns) } /** - * acct_process - now just a wrapper around do_acct_process + * acct_process - now just a wrapper around acct_process_in_ns, + * which in turn is a wrapper around do_acct_process. * * handles process accounting for an exiting task */ @@ -665,6 +666,11 @@ void acct_process(void) { struct pid_namespace *ns; + /* + * This loop is safe lockless, since current is still + * alive and holds its namespace, which in turn holds + * its parent. + */ for (ns = task_active_pid_ns(current); ns != NULL; ns = ns->parent) acct_process_in_ns(ns); } -- cgit v1.2.3 From 297c5d92634c809cef23d73e7b2556f2528ff7e2 Mon Sep 17 00:00:00 2001 From: Andrea Righi Date: Fri, 25 Jul 2008 01:48:49 -0700 Subject: task IO accounting: provide distinct tgid/tid I/O statistics Report per-thread I/O statistics in /proc/pid/task/tid/io and aggregate parent I/O statistics in /proc/pid/io. This approach follows the same model used to account per-process and per-thread CPU times. As a practial application, this allows for example to quickly find the top I/O consumer when a process spawns many child threads that perform the actual I/O work, because the aggregated I/O statistics can always be found in /proc/pid/io. [ Oleg Nesterov points out that we should check that the task is still alive before we iterate over the threads, but also says that we can do that fixup on top of this later. - Linus ] Acked-by: Balbir Singh Signed-off-by: Andrea Righi Cc: Matt Heaton Cc: Shailabh Nagar Acked-by-with-comments: Oleg Nesterov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/exit.c | 27 +++++++++++++++++++++++++++ kernel/fork.c | 6 ++++++ 2 files changed, 33 insertions(+) (limited to 'kernel') diff --git a/kernel/exit.c b/kernel/exit.c index 8a4d4d12e294..ad933bb29ec7 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -120,6 +120,18 @@ static void __exit_signal(struct task_struct *tsk) sig->nivcsw += tsk->nivcsw; sig->inblock += task_io_get_inblock(tsk); sig->oublock += task_io_get_oublock(tsk); +#ifdef CONFIG_TASK_XACCT + sig->rchar += tsk->rchar; + sig->wchar += tsk->wchar; + sig->syscr += tsk->syscr; + sig->syscw += tsk->syscw; +#endif /* CONFIG_TASK_XACCT */ +#ifdef CONFIG_TASK_IO_ACCOUNTING + sig->ioac.read_bytes += tsk->ioac.read_bytes; + sig->ioac.write_bytes += tsk->ioac.write_bytes; + sig->ioac.cancelled_write_bytes += + tsk->ioac.cancelled_write_bytes; +#endif /* CONFIG_TASK_IO_ACCOUNTING */ sig->sum_sched_runtime += tsk->se.sum_exec_runtime; sig = NULL; /* Marker for below. */ } @@ -1366,6 +1378,21 @@ static int wait_task_zombie(struct task_struct *p, int options, psig->coublock += task_io_get_oublock(p) + sig->oublock + sig->coublock; +#ifdef CONFIG_TASK_XACCT + psig->rchar += p->rchar + sig->rchar; + psig->wchar += p->wchar + sig->wchar; + psig->syscr += p->syscr + sig->syscr; + psig->syscw += p->syscw + sig->syscw; +#endif /* CONFIG_TASK_XACCT */ +#ifdef CONFIG_TASK_IO_ACCOUNTING + psig->ioac.read_bytes += + p->ioac.read_bytes + sig->ioac.read_bytes; + psig->ioac.write_bytes += + p->ioac.write_bytes + sig->ioac.write_bytes; + psig->ioac.cancelled_write_bytes += + p->ioac.cancelled_write_bytes + + sig->ioac.cancelled_write_bytes; +#endif /* CONFIG_TASK_IO_ACCOUNTING */ spin_unlock_irq(&p->parent->sighand->siglock); } diff --git a/kernel/fork.c b/kernel/fork.c index 813d5c89b9d5..b99d73e971a4 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -812,6 +812,12 @@ static int copy_signal(unsigned long clone_flags, struct task_struct *tsk) sig->nvcsw = sig->nivcsw = sig->cnvcsw = sig->cnivcsw = 0; sig->min_flt = sig->maj_flt = sig->cmin_flt = sig->cmaj_flt = 0; sig->inblock = sig->oublock = sig->cinblock = sig->coublock = 0; +#ifdef CONFIG_TASK_XACCT + sig->rchar = sig->wchar = sig->syscr = sig->syscw = 0; +#endif +#ifdef CONFIG_TASK_IO_ACCOUNTING + memset(&sig->ioac, 0, sizeof(sig->ioac)); +#endif sig->sum_sched_runtime = 0; INIT_LIST_HEAD(&sig->cpu_timers[0]); INIT_LIST_HEAD(&sig->cpu_timers[1]); -- cgit v1.2.3 From 3e85ba034deec351f02cb55ff225bbd616463841 Mon Sep 17 00:00:00 2001 From: David Howells Date: Fri, 25 Jul 2008 01:48:50 -0700 Subject: tsacct: fix bacct_add_tsk()'s use of do_div() Fix bacct_add_tsk()'s use of do_div() on an s64 by making ac_etime a u64 instead and dividing that. Possibly this should be guarded lest the interval calculation turn up negative, but the possible negativity of the result of the division is cast away, and it shouldn't end up negative anyway. This was introduced by patch f3cef7a99469afc159fec3a61b42dc7ca5b6824f. Signed-off-by: David Howells Cc: Jay Lan Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/tsacct.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/tsacct.c b/kernel/tsacct.c index 1da6990af8e0..3da47ccdc5e5 100644 --- a/kernel/tsacct.c +++ b/kernel/tsacct.c @@ -28,14 +28,14 @@ void bacct_add_tsk(struct taskstats *stats, struct task_struct *tsk) { struct timespec uptime, ts; - s64 ac_etime; + u64 ac_etime; BUILD_BUG_ON(TS_COMM_LEN < TASK_COMM_LEN); /* calculate task elapsed time in timespec */ do_posix_clock_monotonic_gettime(&uptime); ts = timespec_sub(uptime, tsk->start_time); - /* rebase elapsed time to usec */ + /* rebase elapsed time to usec (should never be negative) */ ac_etime = timespec_to_ns(&ts); do_div(ac_etime, NSEC_PER_USEC); stats->ac_etime = ac_etime; -- cgit v1.2.3 From 873b47717732c2f33a4b14de02571a4295a02f0c Mon Sep 17 00:00:00 2001 From: Keika Kobayashi Date: Fri, 25 Jul 2008 01:48:52 -0700 Subject: per-task-delay-accounting: add memory reclaim delay Sometimes, application responses become bad under heavy memory load. Applications take a bit time to reclaim memory. The statistics, how long memory reclaim takes, will be useful to measure memory usage. This patch adds accounting memory reclaim to per-task-delay-accounting for accounting the time of do_try_to_free_pages(). - When System is under low memory load, memory reclaim may not occur. $ free total used free shared buffers cached Mem: 8197800 1577300 6620500 0 4808 1516724 -/+ buffers/cache: 55768 8142032 Swap: 16386292 0 16386292 $ vmstat 1 procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu---- r b swpd free buff cache si so bi bo in cs us sy id wa 0 0 0 5069748 10612 3014060 0 0 0 0 3 26 0 0 100 0 0 0 0 5069748 10612 3014060 0 0 0 0 4 22 0 0 100 0 0 0 0 5069748 10612 3014060 0 0 0 0 3 18 0 0 100 0 Measure the time of tar command. $ ls -s test.dat 1501472 test.dat $ time tar cvf test.tar test.dat real 0m13.388s user 0m0.116s sys 0m5.304s $ ./delayget -d -p CPU count real total virtual total delay total 428 5528345500 5477116080 62749891 IO count delay total 338 8078977189 SWAP count delay total 0 0 RECLAIM count delay total 0 0 - When system is under heavy memory load memory reclaim may occur. $ vmstat 1 procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu---- r b swpd free buff cache si so bi bo in cs us sy id wa 0 0 7159032 49724 1812 3012 0 0 0 0 3 24 0 0 100 0 0 0 7159032 49724 1812 3012 0 0 0 0 4 24 0 0 100 0 0 0 7159032 49848 1812 3012 0 0 0 0 3 22 0 0 100 0 In this case, one process uses more 8G memory by execution of malloc() and memset(). $ time tar cvf test.tar test.dat real 1m38.563s <- increased by 85 sec user 0m0.140s sys 0m7.060s $ ./delayget -d -p CPU count real total virtual total delay total 9021 7140446250 7315277975 923201824 IO count delay total 8965 90466349669 SWAP count delay total 3 21036367 RECLAIM count delay total 740 61011951153 In the later case, the value of RECLAIM is increasing. So, taskstats can show how much memory reclaim influences TAT. Signed-off-by: Keika Kobayashi Acked-by: Balbir Singh Acked-by: KOSAKI Motohiro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/delayacct.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'kernel') diff --git a/kernel/delayacct.c b/kernel/delayacct.c index 10e43fd8b721..84b6782a2ce4 100644 --- a/kernel/delayacct.c +++ b/kernel/delayacct.c @@ -165,3 +165,16 @@ __u64 __delayacct_blkio_ticks(struct task_struct *tsk) return ret; } +void __delayacct_freepages_start(void) +{ + delayacct_start(¤t->delays->freepages_start); +} + +void __delayacct_freepages_end(void) +{ + delayacct_end(¤t->delays->freepages_start, + ¤t->delays->freepages_end, + ¤t->delays->freepages_delay, + ¤t->delays->freepages_count); +} + -- cgit v1.2.3 From 016ae219b920c4e606088761d3d6070cdf8ba706 Mon Sep 17 00:00:00 2001 From: Keika Kobayashi Date: Fri, 25 Jul 2008 01:48:53 -0700 Subject: per-task-delay-accounting: update taskstats for memory reclaim delay Add members for memory reclaim delay to taskstats, and accumulate them in __delayacct_add_tsk() . Signed-off-by: Keika Kobayashi Cc: Hiroshi Shimamoto Cc: Balbir Singh Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/delayacct.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'kernel') diff --git a/kernel/delayacct.c b/kernel/delayacct.c index 84b6782a2ce4..b3179dad71be 100644 --- a/kernel/delayacct.c +++ b/kernel/delayacct.c @@ -145,8 +145,11 @@ int __delayacct_add_tsk(struct taskstats *d, struct task_struct *tsk) d->blkio_delay_total = (tmp < d->blkio_delay_total) ? 0 : tmp; tmp = d->swapin_delay_total + tsk->delays->swapin_delay; d->swapin_delay_total = (tmp < d->swapin_delay_total) ? 0 : tmp; + tmp = d->freepages_delay_total + tsk->delays->freepages_delay; + d->freepages_delay_total = (tmp < d->freepages_delay_total) ? 0 : tmp; d->blkio_count += tsk->delays->blkio_count; d->swapin_count += tsk->delays->swapin_count; + d->freepages_count += tsk->delays->freepages_count; spin_unlock_irqrestore(&tsk->delays->lock, flags); done: -- cgit v1.2.3 From b81f3ea92ba1fa676775677679889dc2a7f03c8b Mon Sep 17 00:00:00 2001 From: Vegard Nossum Date: Fri, 25 Jul 2008 01:48:55 -0700 Subject: taskstats: remove initialization of static per-cpu variable Cc: Shailabh Nagar Signed-off-by: Vegard Nossum Cc: Balbir Singh Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/taskstats.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/taskstats.c b/kernel/taskstats.c index 06b17547f4e7..bd6be76303cf 100644 --- a/kernel/taskstats.c +++ b/kernel/taskstats.c @@ -35,7 +35,7 @@ */ #define TASKSTATS_CPUMASK_MAXLEN (100+6*NR_CPUS) -static DEFINE_PER_CPU(__u32, taskstats_seqnum) = { 0 }; +static DEFINE_PER_CPU(__u32, taskstats_seqnum); static int family_registered; struct kmem_cache *taskstats_cache; -- cgit v1.2.3 From 9b81361631bbb1d85c99ddec677d42afe516737b Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 25 Jul 2008 13:02:37 +0200 Subject: signalfd: fix undefined reference to `compat_sys_signalfd4' when !CONFIG_SIGNALFD fix: arch/x86/ia32/built-in.o: In function `ia32_sys_call_table': (.rodata+0xa38): undefined reference to `compat_sys_signalfd4' on !CONFIG_SIGNALFD. Signed-off-by: Ingo Molnar Signed-off-by: Linus Torvalds --- kernel/sys_ni.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/sys_ni.c b/kernel/sys_ni.c index 55eca1594da9..08d6e1bb99ac 100644 --- a/kernel/sys_ni.c +++ b/kernel/sys_ni.c @@ -160,6 +160,7 @@ cond_syscall(sys_ioprio_get); cond_syscall(sys_signalfd); cond_syscall(sys_signalfd4); cond_syscall(compat_sys_signalfd); +cond_syscall(compat_sys_signalfd4); cond_syscall(sys_timerfd_create); cond_syscall(sys_timerfd_settime); cond_syscall(sys_timerfd_gettime); -- cgit v1.2.3 From 1fe371044b21b226b96a9dd959e971b50b28c78e Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Sat, 26 Jul 2008 15:09:47 +0200 Subject: ftrace: fix modular build fix: ERROR: "start_critical_timings" [drivers/acpi/processor.ko] undefined! ERROR: "stop_critical_timings" [drivers/acpi/processor.ko] undefined! Signed-off-by: Ingo Molnar --- kernel/trace/trace_irqsoff.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'kernel') diff --git a/kernel/trace/trace_irqsoff.c b/kernel/trace/trace_irqsoff.c index b1e4a89b08eb..ece6cfb649fa 100644 --- a/kernel/trace/trace_irqsoff.c +++ b/kernel/trace/trace_irqsoff.c @@ -253,12 +253,14 @@ void start_critical_timings(void) if (preempt_trace() || irq_trace()) start_critical_timing(CALLER_ADDR0, CALLER_ADDR1); } +EXPORT_SYMBOL_GPL(start_critical_timings); void stop_critical_timings(void) { if (preempt_trace() || irq_trace()) stop_critical_timing(CALLER_ADDR0, CALLER_ADDR1); } +EXPORT_SYMBOL_GPL(stop_critical_timings); #ifdef CONFIG_IRQSOFF_TRACER #ifdef CONFIG_PROVE_LOCKING -- cgit v1.2.3 From a2e2e3577c3ef2b5dbb866e97e612aae4adfa32f Mon Sep 17 00:00:00 2001 From: David Brownell Date: Fri, 25 Jul 2008 19:44:38 -0700 Subject: pm selftest: rtc paranoia Cope with a quirk of some RTCs (notably ACPI ones) which aren't guaranteed to implement oneshot behavior when they woke the system from sleeep: forcibly disable the alarm, just in case. Signed-off-by: David Brownell Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/power/main.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'kernel') diff --git a/kernel/power/main.c b/kernel/power/main.c index 95bff23ecdaa..0b7476f5d2a6 100644 --- a/kernel/power/main.c +++ b/kernel/power/main.c @@ -635,6 +635,13 @@ static void __init test_wakealarm(struct rtc_device *rtc, suspend_state_t state) } if (status < 0) printk(err_suspend, status); + + /* Some platforms can't detect that the alarm triggered the + * wakeup, or (accordingly) disable it after it afterwards. + * It's supposed to give oneshot behavior; cope. + */ + alm.enabled = false; + rtc_set_alarm(rtc, &alm); } static int __init has_wakealarm(struct device *dev, void *name_ptr) -- cgit v1.2.3 From 7fccf0326536c1b245b98740d489abb9aab69a12 Mon Sep 17 00:00:00 2001 From: WANG Cong Date: Fri, 25 Jul 2008 19:45:02 -0700 Subject: kernel/kexec.c: make 'kimage_terminate' void Since kimage_terminate() always returns 0, make it void. Signed-off-by: WANG Cong Signed-off-by: "Eric W. Biederman" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/kexec.c | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/kexec.c b/kernel/kexec.c index 1c5fcacbcf33..6db42ff8d520 100644 --- a/kernel/kexec.c +++ b/kernel/kexec.c @@ -589,14 +589,12 @@ static void kimage_free_extra_pages(struct kimage *image) kimage_free_page_list(&image->unuseable_pages); } -static int kimage_terminate(struct kimage *image) +static void kimage_terminate(struct kimage *image) { if (*image->entry != 0) image->entry++; *image->entry = IND_DONE; - - return 0; } #define for_each_kimage_entry(image, ptr, entry) \ @@ -997,9 +995,7 @@ asmlinkage long sys_kexec_load(unsigned long entry, unsigned long nr_segments, if (result) goto out; } - result = kimage_terminate(image); - if (result) - goto out; + kimage_terminate(image); } /* Install the new kernel, and Uninstall the old */ image = xchg(dest_image, image); -- cgit v1.2.3 From 3ab83521378268044a448113c6aa9a9e245f4d2f Mon Sep 17 00:00:00 2001 From: Huang Ying Date: Fri, 25 Jul 2008 19:45:07 -0700 Subject: kexec jump This patch provides an enhancement to kexec/kdump. It implements the following features: - Backup/restore memory used by the original kernel before/after kexec. - Save/restore CPU state before/after kexec. The features of this patch can be used as a general method to call program in physical mode (paging turning off). This can be used to call BIOS code under Linux. kexec-tools needs to be patched to support kexec jump. The patches and the precompiled kexec can be download from the following URL: source: http://khibernation.sourceforge.net/download/release_v10/kexec-tools/kexec-tools-src_git_kh10.tar.bz2 patches: http://khibernation.sourceforge.net/download/release_v10/kexec-tools/kexec-tools-patches_git_kh10.tar.bz2 binary: http://khibernation.sourceforge.net/download/release_v10/kexec-tools/kexec_git_kh10 Usage example of calling some physical mode code and return: 1. Compile and install patched kernel with following options selected: CONFIG_X86_32=y CONFIG_KEXEC=y CONFIG_PM=y CONFIG_KEXEC_JUMP=y 2. Build patched kexec-tool or download the pre-built one. 3. Build some physical mode executable named such as "phy_mode" 4. Boot kernel compiled in step 1. 5. Load physical mode executable with /sbin/kexec. The shell command line can be as follow: /sbin/kexec --load-preserve-context --args-none phy_mode 6. Call physical mode executable with following shell command line: /sbin/kexec -e Implementation point: To support jumping without reserving memory. One shadow backup page (source page) is allocated for each page used by kexeced code image (destination page). When do kexec_load, the image of kexeced code is loaded into source pages, and before executing, the destination pages and the source pages are swapped, so the contents of destination pages are backupped. Before jumping to the kexeced code image and after jumping back to the original kernel, the destination pages and the source pages are swapped too. C ABI (calling convention) is used as communication protocol between kernel and called code. A flag named KEXEC_PRESERVE_CONTEXT for sys_kexec_load is added to indicate that the loaded kernel image is used for jumping back. Now, only the i386 architecture is supported. Signed-off-by: Huang Ying Acked-by: Vivek Goyal Cc: "Eric W. Biederman" Cc: Pavel Machek Cc: Nigel Cunningham Cc: "Rafael J. Wysocki" Cc: Ingo Molnar Cc: Thomas Gleixner Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/kexec.c | 57 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++ kernel/sys.c | 31 ++++++++----------------------- 2 files changed, 65 insertions(+), 23 deletions(-) (limited to 'kernel') diff --git a/kernel/kexec.c b/kernel/kexec.c index 6db42ff8d520..a0d920915b38 100644 --- a/kernel/kexec.c +++ b/kernel/kexec.c @@ -24,6 +24,8 @@ #include #include #include +#include +#include #include #include @@ -242,6 +244,12 @@ static int kimage_normal_alloc(struct kimage **rimage, unsigned long entry, goto out; } + image->swap_page = kimage_alloc_control_pages(image, 0); + if (!image->swap_page) { + printk(KERN_ERR "Could not allocate swap buffer\n"); + goto out; + } + result = 0; out: if (result == 0) @@ -986,6 +994,8 @@ asmlinkage long sys_kexec_load(unsigned long entry, unsigned long nr_segments, if (result) goto out; + if (flags & KEXEC_PRESERVE_CONTEXT) + image->preserve_context = 1; result = machine_kexec_prepare(image); if (result) goto out; @@ -1411,3 +1421,50 @@ static int __init crash_save_vmcoreinfo_init(void) } module_init(crash_save_vmcoreinfo_init) + +/** + * kernel_kexec - reboot the system + * + * Move into place and start executing a preloaded standalone + * executable. If nothing was preloaded return an error. + */ +int kernel_kexec(void) +{ + int error = 0; + + if (xchg(&kexec_lock, 1)) + return -EBUSY; + if (!kexec_image) { + error = -EINVAL; + goto Unlock; + } + + if (kexec_image->preserve_context) { +#ifdef CONFIG_KEXEC_JUMP + local_irq_disable(); + save_processor_state(); +#endif + } else { + blocking_notifier_call_chain(&reboot_notifier_list, + SYS_RESTART, NULL); + system_state = SYSTEM_RESTART; + device_shutdown(); + sysdev_shutdown(); + printk(KERN_EMERG "Starting new kernel\n"); + machine_shutdown(); + } + + machine_kexec(kexec_image); + + if (kexec_image->preserve_context) { +#ifdef CONFIG_KEXEC_JUMP + restore_processor_state(); + local_irq_enable(); +#endif + } + + Unlock: + xchg(&kexec_lock, 0); + + return error; +} diff --git a/kernel/sys.c b/kernel/sys.c index 0c9d3fa1f5ff..c01858090a98 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -301,26 +301,6 @@ void kernel_restart(char *cmd) } EXPORT_SYMBOL_GPL(kernel_restart); -/** - * kernel_kexec - reboot the system - * - * Move into place and start executing a preloaded standalone - * executable. If nothing was preloaded return an error. - */ -static void kernel_kexec(void) -{ -#ifdef CONFIG_KEXEC - struct kimage *image; - image = xchg(&kexec_image, NULL); - if (!image) - return; - kernel_restart_prepare(NULL); - printk(KERN_EMERG "Starting new kernel\n"); - machine_shutdown(); - machine_kexec(image); -#endif -} - static void kernel_shutdown_prepare(enum system_states state) { blocking_notifier_call_chain(&reboot_notifier_list, @@ -425,10 +405,15 @@ asmlinkage long sys_reboot(int magic1, int magic2, unsigned int cmd, void __user kernel_restart(buffer); break; +#ifdef CONFIG_KEXEC case LINUX_REBOOT_CMD_KEXEC: - kernel_kexec(); - unlock_kernel(); - return -EINVAL; + { + int ret; + ret = kernel_kexec(); + unlock_kernel(); + return ret; + } +#endif #ifdef CONFIG_HIBERNATION case LINUX_REBOOT_CMD_SW_SUSPEND: -- cgit v1.2.3 From 89081d17f7bb81d89fa1aa9b70f821c5cf4d39e9 Mon Sep 17 00:00:00 2001 From: Huang Ying Date: Fri, 25 Jul 2008 19:45:10 -0700 Subject: kexec jump: save/restore device state This patch implements devices state save/restore before after kexec. This patch together with features in kexec_jump patch can be used for following: - A simple hibernation implementation without ACPI support. You can kexec a hibernating kernel, save the memory image of original system and shutdown the system. When resuming, you restore the memory image of original system via ordinary kexec load then jump back. - Kernel/system debug through making system snapshot. You can make system snapshot, jump back, do some thing and make another system snapshot. - Cooperative multi-kernel/system. With kexec jump, you can switch between several kernels/systems quickly without boot process except the first time. This appears like swap a whole kernel/system out/in. - A general method to call program in physical mode (paging turning off). This can be used to invoke BIOS code under Linux. The following user-space tools can be used with kexec jump: - kexec-tools needs to be patched to support kexec jump. The patches and the precompiled kexec can be download from the following URL: source: http://khibernation.sourceforge.net/download/release_v10/kexec-tools/kexec-tools-src_git_kh10.tar.bz2 patches: http://khibernation.sourceforge.net/download/release_v10/kexec-tools/kexec-tools-patches_git_kh10.tar.bz2 binary: http://khibernation.sourceforge.net/download/release_v10/kexec-tools/kexec_git_kh10 - makedumpfile with patches are used as memory image saving tool, it can exclude free pages from original kernel memory image file. The patches and the precompiled makedumpfile can be download from the following URL: source: http://khibernation.sourceforge.net/download/release_v10/makedumpfile/makedumpfile-src_cvs_kh10.tar.bz2 patches: http://khibernation.sourceforge.net/download/release_v10/makedumpfile/makedumpfile-patches_cvs_kh10.tar.bz2 binary: http://khibernation.sourceforge.net/download/release_v10/makedumpfile/makedumpfile_cvs_kh10 - An initramfs image can be used as the root file system of kexeced kernel. An initramfs image built with "BuildRoot" can be downloaded from the following URL: initramfs image: http://khibernation.sourceforge.net/download/release_v10/initramfs/rootfs_cvs_kh10.gz All user space tools above are included in the initramfs image. Usage example of simple hibernation: 1. Compile and install patched kernel with following options selected: CONFIG_X86_32=y CONFIG_RELOCATABLE=y CONFIG_KEXEC=y CONFIG_CRASH_DUMP=y CONFIG_PM=y CONFIG_HIBERNATION=y CONFIG_KEXEC_JUMP=y 2. Build an initramfs image contains kexec-tool and makedumpfile, or download the pre-built initramfs image, called rootfs.gz in following text. 3. Prepare a partition to save memory image of original kernel, called hibernating partition in following text. 4. Boot kernel compiled in step 1 (kernel A). 5. In the kernel A, load kernel compiled in step 1 (kernel B) with /sbin/kexec. The shell command line can be as follow: /sbin/kexec --load-preserve-context /boot/bzImage --mem-min=0x100000 --mem-max=0xffffff --initrd=rootfs.gz 6. Boot the kernel B with following shell command line: /sbin/kexec -e 7. The kernel B will boot as normal kexec. In kernel B the memory image of kernel A can be saved into hibernating partition as follow: jump_back_entry=`cat /proc/cmdline | tr ' ' '\n' | grep kexec_jump_back_entry | cut -d '='` echo $jump_back_entry > kexec_jump_back_entry cp /proc/vmcore dump.elf Then you can shutdown the machine as normal. 8. Boot kernel compiled in step 1 (kernel C). Use the rootfs.gz as root file system. 9. In kernel C, load the memory image of kernel A as follow: /sbin/kexec -l --args-none --entry=`cat kexec_jump_back_entry` dump.elf 10. Jump back to the kernel A as follow: /sbin/kexec -e Then, kernel A is resumed. Implementation point: To support jumping between two kernels, before jumping to (executing) the new kernel and jumping back to the original kernel, the devices are put into quiescent state, and the state of devices and CPU is saved. After jumping back from kexeced kernel and jumping to the new kernel, the state of devices and CPU are restored accordingly. The devices/CPU state save/restore code of software suspend is called to implement corresponding function. Known issues: - Because the segment number supported by sys_kexec_load is limited, hibernation image with many segments may not be load. This is planned to be eliminated by adding a new flag to sys_kexec_load to make a image can be loaded with multiple sys_kexec_load invoking. Now, only the i386 architecture is supported. Signed-off-by: Huang Ying Acked-by: Vivek Goyal Cc: "Eric W. Biederman" Cc: Pavel Machek Cc: Nigel Cunningham Cc: "Rafael J. Wysocki" Cc: Ingo Molnar Cc: Thomas Gleixner Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/kexec.c | 39 +++++++++++++++++++++++++++++++++++++++ kernel/power/power.h | 2 -- 2 files changed, 39 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/kexec.c b/kernel/kexec.c index a0d920915b38..c8a4370e2a34 100644 --- a/kernel/kexec.c +++ b/kernel/kexec.c @@ -26,6 +26,10 @@ #include #include #include +#include +#include +#include +#include #include #include @@ -1441,7 +1445,31 @@ int kernel_kexec(void) if (kexec_image->preserve_context) { #ifdef CONFIG_KEXEC_JUMP + mutex_lock(&pm_mutex); + pm_prepare_console(); + error = freeze_processes(); + if (error) { + error = -EBUSY; + goto Restore_console; + } + suspend_console(); + error = device_suspend(PMSG_FREEZE); + if (error) + goto Resume_console; + error = disable_nonboot_cpus(); + if (error) + goto Resume_devices; local_irq_disable(); + /* At this point, device_suspend() has been called, + * but *not* device_power_down(). We *must* + * device_power_down() now. Otherwise, drivers for + * some devices (e.g. interrupt controllers) become + * desynchronized with the actual state of the + * hardware at resume time, and evil weirdness ensues. + */ + error = device_power_down(PMSG_FREEZE); + if (error) + goto Enable_irqs; save_processor_state(); #endif } else { @@ -1459,7 +1487,18 @@ int kernel_kexec(void) if (kexec_image->preserve_context) { #ifdef CONFIG_KEXEC_JUMP restore_processor_state(); + device_power_up(PMSG_RESTORE); + Enable_irqs: local_irq_enable(); + enable_nonboot_cpus(); + Resume_devices: + device_resume(PMSG_RESTORE); + Resume_console: + resume_console(); + thaw_processes(); + Restore_console: + pm_restore_console(); + mutex_unlock(&pm_mutex); #endif } diff --git a/kernel/power/power.h b/kernel/power/power.h index 700f44ec8406..acc0c101dbd5 100644 --- a/kernel/power/power.h +++ b/kernel/power/power.h @@ -53,8 +53,6 @@ extern int hibernation_platform_enter(void); extern int pfn_is_nosave(unsigned long); -extern struct mutex pm_mutex; - #define power_attr(_name) \ static struct kobj_attribute _name##_attr = { \ .attr = { \ -- cgit v1.2.3 From 7babe8db99d305340cf4828ce1f5a1481d5622ef Mon Sep 17 00:00:00 2001 From: Eduard - Gabriel Munteanu Date: Fri, 25 Jul 2008 19:45:11 -0700 Subject: Full conversion to early_initcall() interface, remove old interface A previous patch added the early_initcall(), to allow a cleaner hooking of pre-SMP initcalls. Now we remove the older interface, converting all existing users to the new one. [akpm@linux-foundation.org: cleanups] [akpm@linux-foundation.org: build fix] [kosaki.motohiro@jp.fujitsu.com: warning fix] [kosaki.motohiro@jp.fujitsu.com: warning fix] Signed-off-by: Eduard - Gabriel Munteanu Cc: Tom Zanussi Signed-off-by: KOSAKI Motohiro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sched.c | 5 ++++- kernel/smp.c | 4 +++- kernel/softirq.c | 3 ++- kernel/softlockup.c | 25 ++++++++++++++++++++++--- 4 files changed, 31 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index 0047bd9b96aa..fde1a1026359 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -6389,7 +6389,7 @@ static struct notifier_block __cpuinitdata migration_notifier = { .priority = 10 }; -void __init migration_init(void) +static int __init migration_init(void) { void *cpu = (void *)(long)smp_processor_id(); int err; @@ -6399,7 +6399,10 @@ void __init migration_init(void) BUG_ON(err == NOTIFY_BAD); migration_call(&migration_notifier, CPU_ONLINE, cpu); register_cpu_notifier(&migration_notifier); + + return err; } +early_initcall(migration_init); #endif #ifdef CONFIG_SMP diff --git a/kernel/smp.c b/kernel/smp.c index 462c785ca1ee..96fc7c0edc59 100644 --- a/kernel/smp.c +++ b/kernel/smp.c @@ -33,7 +33,7 @@ struct call_single_queue { spinlock_t lock; }; -void __cpuinit init_call_single_data(void) +static int __cpuinit init_call_single_data(void) { int i; @@ -43,7 +43,9 @@ void __cpuinit init_call_single_data(void) spin_lock_init(&q->lock); INIT_LIST_HEAD(&q->list); } + return 0; } +early_initcall(init_call_single_data); static void csd_flag_wait(struct call_single_data *data) { diff --git a/kernel/softirq.c b/kernel/softirq.c index f6b03d56c2bf..c506f266a6b9 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -630,7 +630,7 @@ static struct notifier_block __cpuinitdata cpu_nfb = { .notifier_call = cpu_callback }; -__init int spawn_ksoftirqd(void) +static __init int spawn_ksoftirqd(void) { void *cpu = (void *)(long)smp_processor_id(); int err = cpu_callback(&cpu_nfb, CPU_UP_PREPARE, cpu); @@ -640,6 +640,7 @@ __init int spawn_ksoftirqd(void) register_cpu_notifier(&cpu_nfb); return 0; } +early_initcall(spawn_ksoftirqd); #ifdef CONFIG_SMP /* diff --git a/kernel/softlockup.c b/kernel/softlockup.c index 7bd8d1aadd5d..b75b492fbfcf 100644 --- a/kernel/softlockup.c +++ b/kernel/softlockup.c @@ -338,14 +338,33 @@ static struct notifier_block __cpuinitdata cpu_nfb = { .notifier_call = cpu_callback }; -__init void spawn_softlockup_task(void) +static int __initdata nosoftlockup; + +static int __init nosoftlockup_setup(char *str) +{ + nosoftlockup = 1; + return 1; +} +__setup("nosoftlockup", nosoftlockup_setup); + +static int __init spawn_softlockup_task(void) { void *cpu = (void *)(long)smp_processor_id(); - int err = cpu_callback(&cpu_nfb, CPU_UP_PREPARE, cpu); + int err; - BUG_ON(err == NOTIFY_BAD); + if (nosoftlockup) + return 0; + + err = cpu_callback(&cpu_nfb, CPU_UP_PREPARE, cpu); + if (err == NOTIFY_BAD) { + BUG(); + return 1; + } cpu_callback(&cpu_nfb, CPU_ONLINE, cpu); register_cpu_notifier(&cpu_nfb); atomic_notifier_chain_register(&panic_notifier_list, &panic_block); + + return 0; } +early_initcall(spawn_softlockup_task); -- cgit v1.2.3 From 20d8b67c06fa5e74f44e80b0a0fd68c8327f7c6a Mon Sep 17 00:00:00 2001 From: Eduard - Gabriel Munteanu Date: Fri, 25 Jul 2008 19:45:12 -0700 Subject: relay: add buffer-only channels; useful for early logging Allows one to create and use a channel with no associated files. Files can be initialized later. This is useful in scenarios such as logging in early code, before VFS is up. Therefore, such channels can be created and used as soon as kmem_cache_init() completed. This is needed by kmemtrace to do tracing in early kernel code. [kosaki.motohiro@jp.fujitsu.com: build fix] Signed-off-by: Eduard - Gabriel Munteanu Cc: Tom Zanussi Signed-off-by: KOSAKI Motohiro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/relay.c | 170 +++++++++++++++++++++++++++++++++++++++++++++++---------- 1 file changed, 141 insertions(+), 29 deletions(-) (limited to 'kernel') diff --git a/kernel/relay.c b/kernel/relay.c index 7de644cdec43..04006ef970b8 100644 --- a/kernel/relay.c +++ b/kernel/relay.c @@ -407,6 +407,35 @@ void relay_reset(struct rchan *chan) } EXPORT_SYMBOL_GPL(relay_reset); +static inline void relay_set_buf_dentry(struct rchan_buf *buf, + struct dentry *dentry) +{ + buf->dentry = dentry; + buf->dentry->d_inode->i_size = buf->early_bytes; +} + +static struct dentry *relay_create_buf_file(struct rchan *chan, + struct rchan_buf *buf, + unsigned int cpu) +{ + struct dentry *dentry; + char *tmpname; + + tmpname = kzalloc(NAME_MAX + 1, GFP_KERNEL); + if (!tmpname) + return NULL; + snprintf(tmpname, NAME_MAX, "%s%d", chan->base_filename, cpu); + + /* Create file in fs */ + dentry = chan->cb->create_buf_file(tmpname, chan->parent, + S_IRUSR, buf, + &chan->is_global); + + kfree(tmpname); + + return dentry; +} + /* * relay_open_buf - create a new relay channel buffer * @@ -416,45 +445,34 @@ static struct rchan_buf *relay_open_buf(struct rchan *chan, unsigned int cpu) { struct rchan_buf *buf = NULL; struct dentry *dentry; - char *tmpname; if (chan->is_global) return chan->buf[0]; - tmpname = kzalloc(NAME_MAX + 1, GFP_KERNEL); - if (!tmpname) - goto end; - snprintf(tmpname, NAME_MAX, "%s%d", chan->base_filename, cpu); - buf = relay_create_buf(chan); if (!buf) - goto free_name; + return NULL; + + if (chan->has_base_filename) { + dentry = relay_create_buf_file(chan, buf, cpu); + if (!dentry) + goto free_buf; + relay_set_buf_dentry(buf, dentry); + } buf->cpu = cpu; __relay_reset(buf, 1); - /* Create file in fs */ - dentry = chan->cb->create_buf_file(tmpname, chan->parent, S_IRUSR, - buf, &chan->is_global); - if (!dentry) - goto free_buf; - - buf->dentry = dentry; - if(chan->is_global) { chan->buf[0] = buf; buf->cpu = 0; } - goto free_name; + return buf; free_buf: relay_destroy_buf(buf); - buf = NULL; -free_name: - kfree(tmpname); -end: - return buf; + return NULL; } /** @@ -537,8 +555,8 @@ static int __cpuinit relay_hotcpu_callback(struct notifier_block *nb, /** * relay_open - create a new relay channel - * @base_filename: base name of files to create - * @parent: dentry of parent directory, %NULL for root directory + * @base_filename: base name of files to create, %NULL for buffering only + * @parent: dentry of parent directory, %NULL for root directory or buffer * @subbuf_size: size of sub-buffers * @n_subbufs: number of sub-buffers * @cb: client callback functions @@ -560,8 +578,6 @@ struct rchan *relay_open(const char *base_filename, { unsigned int i; struct rchan *chan; - if (!base_filename) - return NULL; if (!(subbuf_size && n_subbufs)) return NULL; @@ -576,7 +592,10 @@ struct rchan *relay_open(const char *base_filename, chan->alloc_size = FIX_SIZE(subbuf_size * n_subbufs); chan->parent = parent; chan->private_data = private_data; - strlcpy(chan->base_filename, base_filename, NAME_MAX); + if (base_filename) { + chan->has_base_filename = 1; + strlcpy(chan->base_filename, base_filename, NAME_MAX); + } setup_callbacks(chan, cb); kref_init(&chan->kref); @@ -604,6 +623,94 @@ free_bufs: } EXPORT_SYMBOL_GPL(relay_open); +struct rchan_percpu_buf_dispatcher { + struct rchan_buf *buf; + struct dentry *dentry; +}; + +/* Called in atomic context. */ +static void __relay_set_buf_dentry(void *info) +{ + struct rchan_percpu_buf_dispatcher *p = info; + + relay_set_buf_dentry(p->buf, p->dentry); +} + +/** + * relay_late_setup_files - triggers file creation + * @chan: channel to operate on + * @base_filename: base name of files to create + * @parent: dentry of parent directory, %NULL for root directory + * + * Returns 0 if successful, non-zero otherwise. + * + * Use to setup files for a previously buffer-only channel. + * Useful to do early tracing in kernel, before VFS is up, for example. + */ +int relay_late_setup_files(struct rchan *chan, + const char *base_filename, + struct dentry *parent) +{ + int err = 0; + unsigned int i, curr_cpu; + unsigned long flags; + struct dentry *dentry; + struct rchan_percpu_buf_dispatcher disp; + + if (!chan || !base_filename) + return -EINVAL; + + strlcpy(chan->base_filename, base_filename, NAME_MAX); + + mutex_lock(&relay_channels_mutex); + /* Is chan already set up? */ + if (unlikely(chan->has_base_filename)) + return -EEXIST; + chan->has_base_filename = 1; + chan->parent = parent; + curr_cpu = get_cpu(); + /* + * The CPU hotplug notifier ran before us and created buffers with + * no files associated. So it's safe to call relay_setup_buf_file() + * on all currently online CPUs. + */ + for_each_online_cpu(i) { + if (unlikely(!chan->buf[i])) { + printk(KERN_ERR "relay_late_setup_files: CPU %u " + "has no buffer, it must have!\n", i); + BUG(); + err = -EINVAL; + break; + } + + dentry = relay_create_buf_file(chan, chan->buf[i], i); + if (unlikely(!dentry)) { + err = -EINVAL; + break; + } + + if (curr_cpu == i) { + local_irq_save(flags); + relay_set_buf_dentry(chan->buf[i], dentry); + local_irq_restore(flags); + } else { + disp.buf = chan->buf[i]; + disp.dentry = dentry; + smp_mb(); + /* relay_channels_mutex must be held, so wait. */ + err = smp_call_function_single(i, + __relay_set_buf_dentry, + &disp, 1); + } + if (unlikely(err)) + break; + } + put_cpu(); + mutex_unlock(&relay_channels_mutex); + + return err; +} + /** * relay_switch_subbuf - switch to a new sub-buffer * @buf: channel buffer @@ -627,8 +734,13 @@ size_t relay_switch_subbuf(struct rchan_buf *buf, size_t length) old_subbuf = buf->subbufs_produced % buf->chan->n_subbufs; buf->padding[old_subbuf] = buf->prev_padding; buf->subbufs_produced++; - buf->dentry->d_inode->i_size += buf->chan->subbuf_size - - buf->padding[old_subbuf]; + if (buf->dentry) + buf->dentry->d_inode->i_size += + buf->chan->subbuf_size - + buf->padding[old_subbuf]; + else + buf->early_bytes += buf->chan->subbuf_size - + buf->padding[old_subbuf]; smp_mb(); if (waitqueue_active(&buf->read_wait)) /* @@ -1237,4 +1349,4 @@ static __init int relay_init(void) return 0; } -module_init(relay_init); +early_initcall(relay_init); -- cgit v1.2.3 From 51cc50685a4275c6a02653670af9f108a64e01cf Mon Sep 17 00:00:00 2001 From: Alexey Dobriyan Date: Fri, 25 Jul 2008 19:45:34 -0700 Subject: SL*B: drop kmem cache argument from constructor Kmem cache passed to constructor is only needed for constructors that are themselves multiplexeres. Nobody uses this "feature", nor does anybody uses passed kmem cache in non-trivial way, so pass only pointer to object. Non-trivial places are: arch/powerpc/mm/init_64.c arch/powerpc/mm/hugetlbpage.c This is flag day, yes. Signed-off-by: Alexey Dobriyan Acked-by: Pekka Enberg Acked-by: Christoph Lameter Cc: Jon Tollefson Cc: Nick Piggin Cc: Matt Mackall [akpm@linux-foundation.org: fix arch/powerpc/mm/hugetlbpage.c] [akpm@linux-foundation.org: fix mm/slab.c] [akpm@linux-foundation.org: fix ubifs] Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/fork.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index b99d73e971a4..80e83e459b17 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1442,7 +1442,7 @@ long do_fork(unsigned long clone_flags, #define ARCH_MIN_MMSTRUCT_ALIGN 0 #endif -static void sighand_ctor(struct kmem_cache *cachep, void *data) +static void sighand_ctor(void *data) { struct sighand_struct *sighand = data; -- cgit v1.2.3 From b8c512f6190e313df69060bae4a161c5c044e272 Mon Sep 17 00:00:00 2001 From: Arjan van de Ven Date: Fri, 25 Jul 2008 19:45:36 -0700 Subject: Use WARN() in kernel/irq/manage.c Replace a printk+WARN_ON() by a WARN(); this increases the chance of the string making it into the bugreport (ie: it goes inside the ---[ cut here ]--- section) Signed-off-by: Arjan van de Ven Cc: Thomas Gleixner Cc: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/irq/manage.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index f8914b92b664..152abfd3589f 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -177,8 +177,7 @@ static void __enable_irq(struct irq_desc *desc, unsigned int irq) { switch (desc->depth) { case 0: - printk(KERN_WARNING "Unbalanced enable for IRQ %d\n", irq); - WARN_ON(1); + WARN(1, KERN_WARNING "Unbalanced enable for IRQ %d\n", irq); break; case 1: { unsigned int status = desc->status & ~IRQ_DISABLED; -- cgit v1.2.3 From 261c40c1191ad8d7a2e49fa2bb5f6a84e3d44b10 Mon Sep 17 00:00:00 2001 From: Arjan van de Ven Date: Fri, 25 Jul 2008 19:45:37 -0700 Subject: use WARN() in kernel/irq/chip.c Use WARN() instead of a printk+WARN_ON() pair; this way the message becomes part of the warning section for better reporting/collection. Signed-off-by: Arjan van de Ven Cc: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/irq/chip.c | 12 ++++-------- 1 file changed, 4 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index 964964baefa2..3cd441ebf5d2 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -28,8 +28,7 @@ void dynamic_irq_init(unsigned int irq) unsigned long flags; if (irq >= NR_IRQS) { - printk(KERN_ERR "Trying to initialize invalid IRQ%d\n", irq); - WARN_ON(1); + WARN(1, KERN_ERR "Trying to initialize invalid IRQ%d\n", irq); return; } @@ -62,8 +61,7 @@ void dynamic_irq_cleanup(unsigned int irq) unsigned long flags; if (irq >= NR_IRQS) { - printk(KERN_ERR "Trying to cleanup invalid IRQ%d\n", irq); - WARN_ON(1); + WARN(1, KERN_ERR "Trying to cleanup invalid IRQ%d\n", irq); return; } @@ -71,9 +69,8 @@ void dynamic_irq_cleanup(unsigned int irq) spin_lock_irqsave(&desc->lock, flags); if (desc->action) { spin_unlock_irqrestore(&desc->lock, flags); - printk(KERN_ERR "Destroying IRQ%d without calling free_irq\n", + WARN(1, KERN_ERR "Destroying IRQ%d without calling free_irq\n", irq); - WARN_ON(1); return; } desc->msi_desc = NULL; @@ -96,8 +93,7 @@ int set_irq_chip(unsigned int irq, struct irq_chip *chip) unsigned long flags; if (irq >= NR_IRQS) { - printk(KERN_ERR "Trying to install chip for IRQ%d\n", irq); - WARN_ON(1); + WARN(1, KERN_ERR "Trying to install chip for IRQ%d\n", irq); return -EINVAL; } -- cgit v1.2.3 From ff1188646c6870f336e910fb894eeed74f50471f Mon Sep 17 00:00:00 2001 From: Roland McGrath Date: Fri, 25 Jul 2008 19:45:45 -0700 Subject: tracehook: unexport ptrace_notify The ptrace_notify() function should not be called by any modules. It was only ever exported to be called by binfmt exec functions. But that is no longer necessary since fs/exec.c deals with that generically now. There should be no calls to ptrace_notify() from outside the core kernel. Signed-off-by: Roland McGrath Cc: Oleg Nesterov Reviewed-by: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/signal.c | 1 - 1 file changed, 1 deletion(-) (limited to 'kernel') diff --git a/kernel/signal.c b/kernel/signal.c index 82c3545596c5..8715c18b27b9 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -1895,7 +1895,6 @@ EXPORT_SYMBOL(recalc_sigpending); EXPORT_SYMBOL_GPL(dequeue_signal); EXPORT_SYMBOL(flush_signals); EXPORT_SYMBOL(force_sig); -EXPORT_SYMBOL(ptrace_notify); EXPORT_SYMBOL(send_sig); EXPORT_SYMBOL(send_sig_info); EXPORT_SYMBOL(sigprocmask); -- cgit v1.2.3 From 30199f5a46aee204bf437a4f5b0740f3efe448b7 Mon Sep 17 00:00:00 2001 From: Roland McGrath Date: Fri, 25 Jul 2008 19:45:46 -0700 Subject: tracehook: exit This moves the PTRACE_EVENT_EXIT tracing into a tracehook.h inline, tracehook_report_exec(). The change has no effect, just clean-up. Signed-off-by: Roland McGrath Cc: Oleg Nesterov Reviewed-by: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/exit.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/exit.c b/kernel/exit.c index ad933bb29ec7..c3691cbc220a 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -46,6 +46,7 @@ #include #include #include +#include #include #include @@ -1029,10 +1030,7 @@ NORET_TYPE void do_exit(long code) if (unlikely(!tsk->pid)) panic("Attempted to kill the idle task!"); - if (unlikely(current->ptrace & PT_TRACE_EXIT)) { - current->ptrace_message = code; - ptrace_notify((PTRACE_EVENT_EXIT << 8) | SIGTRAP); - } + tracehook_report_exit(&code); /* * We're taking recursive faults here in do_exit. Safest is to just -- cgit v1.2.3 From 09a05394fe2448a4139b014936330af23fa7ec83 Mon Sep 17 00:00:00 2001 From: Roland McGrath Date: Fri, 25 Jul 2008 19:45:47 -0700 Subject: tracehook: clone This moves all the ptrace initialization and tracing logic for task creation into tracehook.h and ptrace.h inlines. It reorganizes the code slightly, but should not change any behavior. There are four tracehook entry points, at each important stage of task creation. This keeps the interface from the core fork.c code fairly clean, while supporting the complex setup required for ptrace or something like it. Signed-off-by: Roland McGrath Cc: Oleg Nesterov Reviewed-by: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/fork.c | 69 ++++++++++++++++++++++++----------------------------------- 1 file changed, 28 insertions(+), 41 deletions(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index 80e83e459b17..b42f8ed23611 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -37,6 +37,7 @@ #include #include #include +#include #include #include #include @@ -865,8 +866,7 @@ static void copy_flags(unsigned long clone_flags, struct task_struct *p) new_flags &= ~PF_SUPERPRIV; new_flags |= PF_FORKNOEXEC; - if (!(clone_flags & CLONE_PTRACE)) - p->ptrace = 0; + new_flags |= PF_STARTING; p->flags = new_flags; clear_freeze_flag(p); } @@ -907,7 +907,8 @@ static struct task_struct *copy_process(unsigned long clone_flags, struct pt_regs *regs, unsigned long stack_size, int __user *child_tidptr, - struct pid *pid) + struct pid *pid, + int trace) { int retval; struct task_struct *p; @@ -1163,8 +1164,6 @@ static struct task_struct *copy_process(unsigned long clone_flags, */ p->group_leader = p; INIT_LIST_HEAD(&p->thread_group); - INIT_LIST_HEAD(&p->ptrace_entry); - INIT_LIST_HEAD(&p->ptraced); /* Now that the task is set up, run cgroup callbacks if * necessary. We need to run them before the task is visible @@ -1195,7 +1194,6 @@ static struct task_struct *copy_process(unsigned long clone_flags, p->real_parent = current->real_parent; else p->real_parent = current; - p->parent = p->real_parent; spin_lock(¤t->sighand->siglock); @@ -1237,8 +1235,7 @@ static struct task_struct *copy_process(unsigned long clone_flags, if (likely(p->pid)) { list_add_tail(&p->sibling, &p->real_parent->children); - if (unlikely(p->ptrace & PT_PTRACED)) - __ptrace_link(p, current->parent); + tracehook_finish_clone(p, clone_flags, trace); if (thread_group_leader(p)) { if (clone_flags & CLONE_NEWPID) @@ -1323,29 +1320,13 @@ struct task_struct * __cpuinit fork_idle(int cpu) struct pt_regs regs; task = copy_process(CLONE_VM, 0, idle_regs(®s), 0, NULL, - &init_struct_pid); + &init_struct_pid, 0); if (!IS_ERR(task)) init_idle(task, cpu); return task; } -static int fork_traceflag(unsigned clone_flags) -{ - if (clone_flags & CLONE_UNTRACED) - return 0; - else if (clone_flags & CLONE_VFORK) { - if (current->ptrace & PT_TRACE_VFORK) - return PTRACE_EVENT_VFORK; - } else if ((clone_flags & CSIGNAL) != SIGCHLD) { - if (current->ptrace & PT_TRACE_CLONE) - return PTRACE_EVENT_CLONE; - } else if (current->ptrace & PT_TRACE_FORK) - return PTRACE_EVENT_FORK; - - return 0; -} - /* * Ok, this is the main fork-routine. * @@ -1380,14 +1361,14 @@ long do_fork(unsigned long clone_flags, } } - if (unlikely(current->ptrace)) { - trace = fork_traceflag (clone_flags); - if (trace) - clone_flags |= CLONE_PTRACE; - } + /* + * When called from kernel_thread, don't do user tracing stuff. + */ + if (likely(user_mode(regs))) + trace = tracehook_prepare_clone(clone_flags); p = copy_process(clone_flags, stack_start, regs, stack_size, - child_tidptr, NULL); + child_tidptr, NULL, trace); /* * Do this prior waking up the new thread - the thread pointer * might get invalid after that point, if the thread exits quickly. @@ -1405,24 +1386,30 @@ long do_fork(unsigned long clone_flags, init_completion(&vfork); } - if ((p->ptrace & PT_PTRACED) || (clone_flags & CLONE_STOPPED)) { + tracehook_report_clone(trace, regs, clone_flags, nr, p); + + /* + * We set PF_STARTING at creation in case tracing wants to + * use this to distinguish a fully live task from one that + * hasn't gotten to tracehook_report_clone() yet. Now we + * clear it and set the child going. + */ + p->flags &= ~PF_STARTING; + + if (unlikely(clone_flags & CLONE_STOPPED)) { /* * We'll start up with an immediate SIGSTOP. */ sigaddset(&p->pending.signal, SIGSTOP); set_tsk_thread_flag(p, TIF_SIGPENDING); - } - - if (!(clone_flags & CLONE_STOPPED)) - wake_up_new_task(p, clone_flags); - else __set_task_state(p, TASK_STOPPED); - - if (unlikely (trace)) { - current->ptrace_message = nr; - ptrace_notify ((trace << 8) | SIGTRAP); + } else { + wake_up_new_task(p, clone_flags); } + tracehook_report_clone_complete(trace, regs, + clone_flags, nr, p); + if (clone_flags & CLONE_VFORK) { freezer_do_not_count(); wait_for_completion(&vfork); -- cgit v1.2.3 From daded34be96b1975ff8539ff62ad8b158ce7d842 Mon Sep 17 00:00:00 2001 From: Roland McGrath Date: Fri, 25 Jul 2008 19:45:47 -0700 Subject: tracehook: vfork-done This moves the PTRACE_EVENT_VFORK_DONE tracing into a tracehook.h inline, tracehook_report_vfork_done(). The change has no effect, just clean-up. Signed-off-by: Roland McGrath Cc: Oleg Nesterov Reviewed-by: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/fork.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index b42f8ed23611..abb3ed6298f6 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1414,10 +1414,7 @@ long do_fork(unsigned long clone_flags, freezer_do_not_count(); wait_for_completion(&vfork); freezer_count(); - if (unlikely (current->ptrace & PT_TRACE_VFORK_DONE)) { - current->ptrace_message = nr; - ptrace_notify ((PTRACE_EVENT_VFORK_DONE << 8) | SIGTRAP); - } + tracehook_report_vfork_done(p, nr); } } else { nr = PTR_ERR(p); -- cgit v1.2.3 From dae33574dcf5211e1f43c7e45fa29f73ba3e00cb Mon Sep 17 00:00:00 2001 From: Roland McGrath Date: Fri, 25 Jul 2008 19:45:48 -0700 Subject: tracehook: release_task This moves the ptrace-related logic from release_task into tracehook.h and ptrace.h inlines. It provides clean hooks both before and after locking tasklist_lock, for future tracing logic to do more cleanup without the lock. This also changes release_task() itself in the rare "zap_leader" case to set the leader to EXIT_DEAD before iterating. This maintains the invariant that release_task() only ever handles a task in EXIT_DEAD. This is a common-sense invariant that is already always true except in this one arcane case of zombie leader whose parent ignores SIGCHLD. This change is harmless and only costs one store in this one rare case. It keeps the expected state more consisently sane, which is nicer when debugging weirdness in release_task(). It also lets some future code in the tracehook entry points rely on this invariant for bookkeeping. Signed-off-by: Roland McGrath Cc: Oleg Nesterov Reviewed-by: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/exit.c | 21 +++++++++------------ 1 file changed, 9 insertions(+), 12 deletions(-) (limited to 'kernel') diff --git a/kernel/exit.c b/kernel/exit.c index c3691cbc220a..da28745f7c38 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -163,27 +163,17 @@ static void delayed_put_task_struct(struct rcu_head *rhp) put_task_struct(container_of(rhp, struct task_struct, rcu)); } -/* - * Do final ptrace-related cleanup of a zombie being reaped. - * - * Called with write_lock(&tasklist_lock) held. - */ -static void ptrace_release_task(struct task_struct *p) -{ - BUG_ON(!list_empty(&p->ptraced)); - ptrace_unlink(p); - BUG_ON(!list_empty(&p->ptrace_entry)); -} void release_task(struct task_struct * p) { struct task_struct *leader; int zap_leader; repeat: + tracehook_prepare_release_task(p); atomic_dec(&p->user->processes); proc_flush_task(p); write_lock_irq(&tasklist_lock); - ptrace_release_task(p); + tracehook_finish_release_task(p); __exit_signal(p); /* @@ -205,6 +195,13 @@ repeat: * that case. */ zap_leader = task_detached(leader); + + /* + * This maintains the invariant that release_task() + * only runs on a task in EXIT_DEAD, just for sanity. + */ + if (zap_leader) + leader->exit_state = EXIT_DEAD; } write_unlock_irq(&tasklist_lock); -- cgit v1.2.3 From 35de254dc60f91004b3b5ebb1fc7b2c3093d6032 Mon Sep 17 00:00:00 2001 From: Roland McGrath Date: Fri, 25 Jul 2008 19:45:51 -0700 Subject: tracehook: tracehook_consider_ignored_signal This defines tracehook_consider_ignored_signal() has a fine-grained hook for deciding to prevent the normal short-circuit of sending an ignored signal, as ptrace does. There is no change, only cleanup. Signed-off-by: Roland McGrath Cc: Oleg Nesterov Reviewed-by: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/signal.c | 27 ++++++++++++++++----------- 1 file changed, 16 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/signal.c b/kernel/signal.c index 8715c18b27b9..9efd1cee6d0b 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -22,6 +22,7 @@ #include #include #include +#include #include #include #include @@ -39,24 +40,21 @@ static struct kmem_cache *sigqueue_cachep; -static int __sig_ignored(struct task_struct *t, int sig) +static void __user *sig_handler(struct task_struct *t, int sig) { - void __user *handler; + return t->sighand->action[sig - 1].sa.sa_handler; +} +static int sig_handler_ignored(void __user *handler, int sig) +{ /* Is it explicitly or implicitly ignored? */ - - handler = t->sighand->action[sig - 1].sa.sa_handler; return handler == SIG_IGN || (handler == SIG_DFL && sig_kernel_ignore(sig)); } static int sig_ignored(struct task_struct *t, int sig) { - /* - * Tracers always want to know about signals.. - */ - if (t->ptrace & PT_PTRACED) - return 0; + void __user *handler; /* * Blocked signals are never ignored, since the @@ -66,7 +64,14 @@ static int sig_ignored(struct task_struct *t, int sig) if (sigismember(&t->blocked, sig) || sigismember(&t->real_blocked, sig)) return 0; - return __sig_ignored(t, sig); + handler = sig_handler(t, sig); + if (!sig_handler_ignored(handler, sig)) + return 0; + + /* + * Tracers may want to know about even ignored signals. + */ + return !tracehook_consider_ignored_signal(t, sig, handler); } /* @@ -2298,7 +2303,7 @@ int do_sigaction(int sig, struct k_sigaction *act, struct k_sigaction *oact) * (for example, SIGCHLD), shall cause the pending signal to * be discarded, whether or not it is blocked" */ - if (__sig_ignored(t, sig)) { + if (sig_handler_ignored(sig_handler(t, sig), sig)) { sigemptyset(&mask); sigaddset(&mask, sig); rm_from_queue_full(&mask, &t->signal->shared_pending); -- cgit v1.2.3 From 445a91d2fe3667fb8fc251433645f686933cf56a Mon Sep 17 00:00:00 2001 From: Roland McGrath Date: Fri, 25 Jul 2008 19:45:52 -0700 Subject: tracehook: tracehook_consider_fatal_signal This defines tracehook_consider_fatal_signal() has a fine-grained hook for deciding to skip the special cases for a fatal signal, as ptrace does. There is no change, only cleanup. Signed-off-by: Roland McGrath Cc: Oleg Nesterov Reviewed-by: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/signal.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/signal.c b/kernel/signal.c index 9efd1cee6d0b..1a942ce32ba0 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -300,12 +300,12 @@ flush_signal_handlers(struct task_struct *t, int force_default) int unhandled_signal(struct task_struct *tsk, int sig) { + void __user *handler = tsk->sighand->action[sig-1].sa.sa_handler; if (is_global_init(tsk)) return 1; - if (tsk->ptrace & PT_PTRACED) + if (handler != SIG_IGN && handler != SIG_DFL) return 0; - return (tsk->sighand->action[sig-1].sa.sa_handler == SIG_IGN) || - (tsk->sighand->action[sig-1].sa.sa_handler == SIG_DFL); + return !tracehook_consider_fatal_signal(tsk, sig, handler); } @@ -761,7 +761,8 @@ static void complete_signal(int sig, struct task_struct *p, int group) if (sig_fatal(p, sig) && !(signal->flags & (SIGNAL_UNKILLABLE | SIGNAL_GROUP_EXIT)) && !sigismember(&t->real_blocked, sig) && - (sig == SIGKILL || !(t->ptrace & PT_PTRACED))) { + (sig == SIGKILL || + !tracehook_consider_fatal_signal(t, sig, SIG_DFL))) { /* * This signal will be fatal to the whole group. */ -- cgit v1.2.3 From 7bcf6a2ca5f639b038c48711ebe6c4eca2036641 Mon Sep 17 00:00:00 2001 From: Roland McGrath Date: Fri, 25 Jul 2008 19:45:53 -0700 Subject: tracehook: get_signal_to_deliver This defines the tracehook_get_signal() hook to allow tracing code to slip in before normal signal dequeuing. This lays the groundwork for new tracing features that can inject synthetic signals outside the normal queue or control the disposition of delivered signals. The calling convention lets tracehook_get_signal() decide both exactly what will happen and what signal number to report in the handler/exit. Signed-off-by: Roland McGrath Cc: Oleg Nesterov Reviewed-by: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/signal.c | 38 +++++++++++++++++++++++++++----------- 1 file changed, 27 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/signal.c b/kernel/signal.c index 1a942ce32ba0..10b31ecdd9c8 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -1754,17 +1754,33 @@ relock: do_signal_stop(0)) goto relock; - signr = dequeue_signal(current, ¤t->blocked, info); - if (!signr) - break; /* will return 0 */ + /* + * Tracing can induce an artifical signal and choose sigaction. + * The return value in @signr determines the default action, + * but @info->si_signo is the signal number we will report. + */ + signr = tracehook_get_signal(current, regs, info, return_ka); + if (unlikely(signr < 0)) + goto relock; + if (unlikely(signr != 0)) + ka = return_ka; + else { + signr = dequeue_signal(current, ¤t->blocked, + info); - if (signr != SIGKILL) { - signr = ptrace_signal(signr, info, regs, cookie); if (!signr) - continue; + break; /* will return 0 */ + + if (signr != SIGKILL) { + signr = ptrace_signal(signr, info, + regs, cookie); + if (!signr) + continue; + } + + ka = &sighand->action[signr-1]; } - ka = &sighand->action[signr-1]; if (ka->sa.sa_handler == SIG_IGN) /* Do nothing. */ continue; if (ka->sa.sa_handler != SIG_DFL) { @@ -1812,7 +1828,7 @@ relock: spin_lock_irq(&sighand->siglock); } - if (likely(do_signal_stop(signr))) { + if (likely(do_signal_stop(info->si_signo))) { /* It released the siglock. */ goto relock; } @@ -1833,7 +1849,7 @@ relock: if (sig_kernel_coredump(signr)) { if (print_fatal_signals) - print_fatal_signal(regs, signr); + print_fatal_signal(regs, info->si_signo); /* * If it was able to dump core, this kills all * other threads in the group and synchronizes with @@ -1842,13 +1858,13 @@ relock: * first and our do_group_exit call below will use * that value and ignore the one we pass it. */ - do_coredump((long)signr, signr, regs); + do_coredump(info->si_signo, info->si_signo, regs); } /* * Death signals, no core dump. */ - do_group_exit(signr); + do_group_exit(info->si_signo); /* NOTREACHED */ } spin_unlock_irq(&sighand->siglock); -- cgit v1.2.3 From fa00b80b3c41a845b3d56f866fb40a2e98754c51 Mon Sep 17 00:00:00 2001 From: Roland McGrath Date: Fri, 25 Jul 2008 19:45:54 -0700 Subject: tracehook: job control This defines the tracehook_notify_jctl() hook to formalize the ptrace effects on the job control notifications. There is no change, only cleanup. Signed-off-by: Roland McGrath Cc: Oleg Nesterov Reviewed-by: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/signal.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/signal.c b/kernel/signal.c index 10b31ecdd9c8..e9e699f4b1bd 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -596,9 +596,6 @@ static int check_kill_permission(int sig, struct siginfo *info, return security_task_kill(t, info, sig, 0); } -/* forward decl */ -static void do_notify_parent_cldstop(struct task_struct *tsk, int why); - /* * Handle magic process-wide effects of stop/continue signals. Unlike * the signal actions, these happen immediately at signal-generation @@ -1605,7 +1602,7 @@ finish_stop(int stop_count) * a group stop in progress and we are the last to stop, * report to the parent. When ptraced, every thread reports itself. */ - if (stop_count == 0 || (current->ptrace & PT_PTRACED)) { + if (tracehook_notify_jctl(stop_count == 0, CLD_STOPPED)) { read_lock(&tasklist_lock); do_notify_parent_cldstop(current, CLD_STOPPED); read_unlock(&tasklist_lock); @@ -1741,6 +1738,9 @@ relock: signal->flags &= ~SIGNAL_CLD_MASK; spin_unlock_irq(&sighand->siglock); + if (unlikely(!tracehook_notify_jctl(1, why))) + goto relock; + read_lock(&tasklist_lock); do_notify_parent_cldstop(current->group_leader, why); read_unlock(&tasklist_lock); @@ -1906,7 +1906,7 @@ void exit_signals(struct task_struct *tsk) out: spin_unlock_irq(&tsk->sighand->siglock); - if (unlikely(group_stop)) { + if (unlikely(group_stop) && tracehook_notify_jctl(1, CLD_STOPPED)) { read_lock(&tasklist_lock); do_notify_parent_cldstop(tsk, CLD_STOPPED); read_unlock(&tasklist_lock); -- cgit v1.2.3 From 2b2a1ff64afbadac842bbc58c5166962cf4f7664 Mon Sep 17 00:00:00 2001 From: Roland McGrath Date: Fri, 25 Jul 2008 19:45:54 -0700 Subject: tracehook: death This moves the ptrace logic in task death (exit_notify) into tracehook.h inlines. Some code is rearranged slightly to make things nicer. There is no change, only cleanup. There is one hook called with the tasklist_lock write-locked, as ptrace needs. There is also a new hook called after exit_state changes and without locks. This is a better place for tracing work to be in the future, since it doesn't delay the whole system with locking. Signed-off-by: Roland McGrath Cc: Oleg Nesterov Reviewed-by: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/exit.c | 26 +++++++++----------------- kernel/signal.c | 10 +++++++--- 2 files changed, 16 insertions(+), 20 deletions(-) (limited to 'kernel') diff --git a/kernel/exit.c b/kernel/exit.c index da28745f7c38..6cdf60712bd2 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -885,7 +885,8 @@ static void forget_original_parent(struct task_struct *father) */ static void exit_notify(struct task_struct *tsk, int group_dead) { - int state; + int signal; + void *cookie; /* * This does two things: @@ -922,22 +923,11 @@ static void exit_notify(struct task_struct *tsk, int group_dead) !capable(CAP_KILL)) tsk->exit_signal = SIGCHLD; - /* If something other than our normal parent is ptracing us, then - * send it a SIGCHLD instead of honoring exit_signal. exit_signal - * only has special meaning to our real parent. - */ - if (!task_detached(tsk) && thread_group_empty(tsk)) { - int signal = ptrace_reparented(tsk) ? - SIGCHLD : tsk->exit_signal; - do_notify_parent(tsk, signal); - } else if (tsk->ptrace) { - do_notify_parent(tsk, SIGCHLD); - } + signal = tracehook_notify_death(tsk, &cookie, group_dead); + if (signal > 0) + signal = do_notify_parent(tsk, signal); - state = EXIT_ZOMBIE; - if (task_detached(tsk) && likely(!tsk->ptrace)) - state = EXIT_DEAD; - tsk->exit_state = state; + tsk->exit_state = signal < 0 ? EXIT_DEAD : EXIT_ZOMBIE; /* mt-exec, de_thread() is waiting for us */ if (thread_group_leader(tsk) && @@ -947,8 +937,10 @@ static void exit_notify(struct task_struct *tsk, int group_dead) write_unlock_irq(&tasklist_lock); + tracehook_report_death(tsk, signal, cookie, group_dead); + /* If the process is dead, release it - nobody will wait for it */ - if (state == EXIT_DEAD) + if (signal < 0) release_task(tsk); } diff --git a/kernel/signal.c b/kernel/signal.c index e9e699f4b1bd..0e862d3130ff 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -1326,9 +1326,11 @@ static inline void __wake_up_parent(struct task_struct *p, /* * Let a parent know about the death of a child. * For a stopped/continued status change, use do_notify_parent_cldstop instead. + * + * Returns -1 if our parent ignored us and so we've switched to + * self-reaping, or else @sig. */ - -void do_notify_parent(struct task_struct *tsk, int sig) +int do_notify_parent(struct task_struct *tsk, int sig) { struct siginfo info; unsigned long flags; @@ -1399,12 +1401,14 @@ void do_notify_parent(struct task_struct *tsk, int sig) */ tsk->exit_signal = -1; if (psig->action[SIGCHLD-1].sa.sa_handler == SIG_IGN) - sig = 0; + sig = -1; } if (valid_signal(sig) && sig > 0) __group_send_sig_info(sig, &info, tsk->parent); __wake_up_parent(tsk, tsk->parent); spin_unlock_irqrestore(&psig->siglock, flags); + + return sig; } static void do_notify_parent_cldstop(struct task_struct *tsk, int why) -- cgit v1.2.3 From b787f7ba677840da16a2228c16571ce8a1fcb799 Mon Sep 17 00:00:00 2001 From: Roland McGrath Date: Fri, 25 Jul 2008 19:45:55 -0700 Subject: tracehook: force signal_pending() This defines a new hook tracehook_force_sigpending() that lets tracing code decide to force TIF_SIGPENDING on in recalc_sigpending(). This is not used yet, so it compiles away to nothing for now. It lays the groundwork for new tracing code that can interrupt a task synthetically without actually sending a signal. Signed-off-by: Roland McGrath Cc: Oleg Nesterov Reviewed-by: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/signal.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/signal.c b/kernel/signal.c index 0e862d3130ff..954f77d7e3bc 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -134,7 +134,9 @@ void recalc_sigpending_and_wake(struct task_struct *t) void recalc_sigpending(void) { - if (!recalc_sigpending_tsk(current) && !freezing(current)) + if (unlikely(tracehook_force_sigpending())) + set_thread_flag(TIF_SIGPENDING); + else if (!recalc_sigpending_tsk(current) && !freezing(current)) clear_thread_flag(TIF_SIGPENDING); } -- cgit v1.2.3 From 85ba2d862e521375a8ee01526c5c46b1f24bb4af Mon Sep 17 00:00:00 2001 From: Roland McGrath Date: Fri, 25 Jul 2008 19:45:58 -0700 Subject: tracehook: wait_task_inactive This extends wait_task_inactive() with a new argument so it can be used in a "soft" mode where it will check for the task changing state unexpectedly and back off. There is no change to existing callers. This lays the groundwork to allow robust, noninvasive tracing that can try to sample a blocked thread but back off safely if it wakes up. Signed-off-by: Roland McGrath Cc: Oleg Nesterov Reviewed-by: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/kthread.c | 2 +- kernel/ptrace.c | 2 +- kernel/sched.c | 29 +++++++++++++++++++++++++++-- 3 files changed, 29 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/kthread.c b/kernel/kthread.c index 6111c27491b1..96cff2f8710b 100644 --- a/kernel/kthread.c +++ b/kernel/kthread.c @@ -176,7 +176,7 @@ void kthread_bind(struct task_struct *k, unsigned int cpu) return; } /* Must have done schedule() in kthread() before we set_task_cpu */ - wait_task_inactive(k); + wait_task_inactive(k, 0); set_task_cpu(k, cpu); k->cpus_allowed = cpumask_of_cpu(cpu); k->rt.nr_cpus_allowed = 1; diff --git a/kernel/ptrace.c b/kernel/ptrace.c index 8392a9da6450..082b3fcb32a0 100644 --- a/kernel/ptrace.c +++ b/kernel/ptrace.c @@ -107,7 +107,7 @@ int ptrace_check_attach(struct task_struct *child, int kill) read_unlock(&tasklist_lock); if (!ret && !kill) - wait_task_inactive(child); + ret = wait_task_inactive(child, TASK_TRACED) ? 0 : -ESRCH; /* All systems go.. */ return ret; diff --git a/kernel/sched.c b/kernel/sched.c index fde1a1026359..0236958addcb 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1867,16 +1867,24 @@ migrate_task(struct task_struct *p, int dest_cpu, struct migration_req *req) /* * wait_task_inactive - wait for a thread to unschedule. * + * If @match_state is nonzero, it's the @p->state value just checked and + * not expected to change. If it changes, i.e. @p might have woken up, + * then return zero. When we succeed in waiting for @p to be off its CPU, + * we return a positive number (its total switch count). If a second call + * a short while later returns the same number, the caller can be sure that + * @p has remained unscheduled the whole time. + * * The caller must ensure that the task *will* unschedule sometime soon, * else this function might spin for a *long* time. This function can't * be called with interrupts off, or it may introduce deadlock with * smp_call_function() if an IPI is sent by the same process we are * waiting to become inactive. */ -void wait_task_inactive(struct task_struct *p) +unsigned long wait_task_inactive(struct task_struct *p, long match_state) { unsigned long flags; int running, on_rq; + unsigned long ncsw; struct rq *rq; for (;;) { @@ -1899,8 +1907,11 @@ void wait_task_inactive(struct task_struct *p) * return false if the runqueue has changed and p * is actually now running somewhere else! */ - while (task_running(rq, p)) + while (task_running(rq, p)) { + if (match_state && unlikely(p->state != match_state)) + return 0; cpu_relax(); + } /* * Ok, time to look more closely! We need the rq @@ -1910,8 +1921,20 @@ void wait_task_inactive(struct task_struct *p) rq = task_rq_lock(p, &flags); running = task_running(rq, p); on_rq = p->se.on_rq; + ncsw = 0; + if (!match_state || p->state == match_state) { + ncsw = p->nivcsw + p->nvcsw; + if (unlikely(!ncsw)) + ncsw = 1; + } task_rq_unlock(rq, &flags); + /* + * If it changed from the expected state, bail out now. + */ + if (unlikely(!ncsw)) + break; + /* * Was it really running after all now that we * checked with the proper locks actually held? @@ -1944,6 +1967,8 @@ void wait_task_inactive(struct task_struct *p) */ break; } + + return ncsw; } /*** -- cgit v1.2.3 From 96930a6365c99c160138a395566e360b27348b8f Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Fri, 25 Jul 2008 19:46:21 -0700 Subject: make cgroup_seqfile_release() static cgroup_seqfile_release() can become static. Signed-off-by: Adrian Bunk Acked-by: Paul Menage Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cgroup.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 66ec9fd21e0c..89bd6fb7894f 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1529,7 +1529,7 @@ static int cgroup_seqfile_show(struct seq_file *m, void *arg) return cft->read_seq_string(state->cgroup, cft, m); } -int cgroup_seqfile_release(struct inode *inode, struct file *file) +static int cgroup_seqfile_release(struct inode *inode, struct file *file) { struct seq_file *seq = file->private_data; kfree(seq->private); -- cgit v1.2.3 From 734550921e9b7ab924a43aa3d0bd4239dac4fbf1 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Mon, 14 Jul 2008 21:22:20 -0400 Subject: [PATCH] beginning of sysctl cleanup - ctl_table_set New object: set of sysctls [currently - root and per-net-ns]. Contains: pointer to parent set, list of tables and "should I see this set?" method (->is_seen(set)). Current lists of tables are subsumed by that; net-ns contains such a beast. ->lookup() for ctl_table_root returns pointer to ctl_table_set instead of that to ->list of that ctl_table_set. [folded compile fixes by rdd for configs without sysctl] Signed-off-by: Al Viro --- kernel/sysctl.c | 41 +++++++++++++++++++++++++++++++---------- 1 file changed, 31 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 35a50db9b6ce..8ee4a0619fbb 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -160,12 +160,13 @@ static struct ctl_table root_table[]; static struct ctl_table_root sysctl_table_root; static struct ctl_table_header root_table_header = { .ctl_table = root_table, - .ctl_entry = LIST_HEAD_INIT(sysctl_table_root.header_list), + .ctl_entry = LIST_HEAD_INIT(sysctl_table_root.default_set.list), .root = &sysctl_table_root, + .set = &sysctl_table_root.default_set, }; static struct ctl_table_root sysctl_table_root = { .root_list = LIST_HEAD_INIT(sysctl_table_root.root_list), - .header_list = LIST_HEAD_INIT(root_table_header.ctl_entry), + .default_set.list = LIST_HEAD_INIT(root_table_header.ctl_entry), }; static struct ctl_table kern_table[]; @@ -1403,14 +1404,20 @@ void sysctl_head_finish(struct ctl_table_header *head) spin_unlock(&sysctl_lock); } +static struct ctl_table_set * +lookup_header_set(struct ctl_table_root *root, struct nsproxy *namespaces) +{ + struct ctl_table_set *set = &root->default_set; + if (root->lookup) + set = root->lookup(root, namespaces); + return set; +} + static struct list_head * lookup_header_list(struct ctl_table_root *root, struct nsproxy *namespaces) { - struct list_head *header_list; - header_list = &root->header_list; - if (root->lookup) - header_list = root->lookup(root, namespaces); - return header_list; + struct ctl_table_set *set = lookup_header_set(root, namespaces); + return &set->list; } struct ctl_table_header *__sysctl_head_next(struct nsproxy *namespaces, @@ -1720,7 +1727,6 @@ struct ctl_table_header *__register_sysctl_paths( struct nsproxy *namespaces, const struct ctl_path *path, struct ctl_table *table) { - struct list_head *header_list; struct ctl_table_header *header; struct ctl_table *new, **prevp; unsigned int n, npath; @@ -1772,8 +1778,8 @@ struct ctl_table_header *__register_sysctl_paths( } #endif spin_lock(&sysctl_lock); - header_list = lookup_header_list(root, namespaces); - list_add_tail(&header->ctl_entry, header_list); + header->set = lookup_header_set(root, namespaces); + list_add_tail(&header->ctl_entry, &header->set->list); spin_unlock(&sysctl_lock); return header; @@ -1832,6 +1838,15 @@ void unregister_sysctl_table(struct ctl_table_header * header) kfree(header); } +void setup_sysctl_set(struct ctl_table_set *p, + struct ctl_table_set *parent, + int (*is_seen)(struct ctl_table_set *)) +{ + INIT_LIST_HEAD(&p->list); + p->parent = parent ? parent : &sysctl_table_root.default_set; + p->is_seen = is_seen; +} + #else /* !CONFIG_SYSCTL */ struct ctl_table_header *register_sysctl_table(struct ctl_table * table) { @@ -1848,6 +1863,12 @@ void unregister_sysctl_table(struct ctl_table_header * table) { } +void setup_sysctl_set(struct ctl_table_set *p, + struct ctl_table_set *parent, + int (*is_seen)(struct ctl_table_set *)) +{ +} + #endif /* CONFIG_SYSCTL */ /* -- cgit v1.2.3 From f7e6ced4061da509f737541ca4dbd44d83a6e82f Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 15 Jul 2008 01:44:23 -0400 Subject: [PATCH] allow delayed freeing of ctl_table_header Refcount the sucker; instead of freeing it by the end of unregistration just drop the refcount and free only when it hits zero. Make sure that we _always_ make ->unregistering non-NULL in start_unregistering(). That allows anybody to get a reference to such puppy, preventing its freeing and reuse. It does *not* block unregistration. Anybody who holds such a reference can * try to grab a "use" reference (ctl_head_grab()); that will succeeds if and only if it hadn't entered unregistration yet. If it succeeds, we can use it in all normal ways until we release the "use" reference (with ctl_head_finish()). Note that this relies on having ->unregistering become non-NULL in all cases when one starts to unregister the sucker. * keep pointers to ctl_table entries; they *can* be freed if the entire thing is unregistered. However, if ctl_head_grab() succeeds, we know that unregistration had not happened (and will not happen until ctl_head_finish()) and such pointers can be used safely. IOW, now we can have inodes under /proc/sys keep references to ctl_table entries, protecting them with references to ctl_table_header and grabbing the latter for the duration of operations that require access to ctl_table. That won't cause deadlocks, since unregistration will not be stopped by mere keeping a reference to ctl_table_header. Signed-off-by: Al Viro --- kernel/sysctl.c | 37 ++++++++++++++++++++++++++++++++++++- 1 file changed, 36 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 8ee4a0619fbb..60d9357e7172 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -1387,6 +1387,9 @@ static void start_unregistering(struct ctl_table_header *p) spin_unlock(&sysctl_lock); wait_for_completion(&wait); spin_lock(&sysctl_lock); + } else { + /* anything non-NULL; we'll never dereference it */ + p->unregistering = ERR_PTR(-EINVAL); } /* * do not remove from the list until nobody holds it; walking the @@ -1395,6 +1398,32 @@ static void start_unregistering(struct ctl_table_header *p) list_del_init(&p->ctl_entry); } +void sysctl_head_get(struct ctl_table_header *head) +{ + spin_lock(&sysctl_lock); + head->count++; + spin_unlock(&sysctl_lock); +} + +void sysctl_head_put(struct ctl_table_header *head) +{ + spin_lock(&sysctl_lock); + if (!--head->count) + kfree(head); + spin_unlock(&sysctl_lock); +} + +struct ctl_table_header *sysctl_head_grab(struct ctl_table_header *head) +{ + if (!head) + BUG(); + spin_lock(&sysctl_lock); + if (!use_table(head)) + head = ERR_PTR(-ENOENT); + spin_unlock(&sysctl_lock); + return head; +} + void sysctl_head_finish(struct ctl_table_header *head) { if (!head) @@ -1771,6 +1800,7 @@ struct ctl_table_header *__register_sysctl_paths( header->unregistering = NULL; header->root = root; sysctl_set_parent(NULL, header->ctl_table); + header->count = 1; #ifdef CONFIG_SYSCTL_SYSCALL_CHECK if (sysctl_check_table(namespaces, header->ctl_table)) { kfree(header); @@ -1834,8 +1864,9 @@ void unregister_sysctl_table(struct ctl_table_header * header) spin_lock(&sysctl_lock); start_unregistering(header); + if (!--header->count) + kfree(header); spin_unlock(&sysctl_lock); - kfree(header); } void setup_sysctl_set(struct ctl_table_set *p, @@ -1869,6 +1900,10 @@ void setup_sysctl_set(struct ctl_table_set *p, { } +void sysctl_head_put(struct ctl_table_header *head) +{ +} + #endif /* CONFIG_SYSCTL */ /* -- cgit v1.2.3 From ae7edecc9b8810770a8e5cb9a466ea4bdcfa8401 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 15 Jul 2008 06:33:31 -0400 Subject: [PATCH] sysctl: keep track of tree relationships In a sense, that's the heart of the series. It's based on the following property of the trees we are actually asked to add: they can be split into stem that is already covered by registered trees and crown that is entirely new. IOW, if a/b and a/c/d are introduced by our tree, then a/c is also introduced by it. That allows to associate tree and table entry with each node in the union; while directory nodes might be covered by many trees, only one will cover the node by its crown. And that will allow much saner logics for /proc/sys in the next patches. This patch introduces the data structures needed to keep track of that. When adding a sysctl table, we find a "parent" one. Which is to say, find the deepest node on its stem that already is present in one of the tables from our table set or its ancestor sets. That table will be our parent and that node in it - attachment point. Add our table to list anchored in parent, have it refer the parent and contents of attachment point. Also remember where its crown lives. Signed-off-by: Al Viro --- kernel/sysctl.c | 63 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 63 insertions(+) (limited to 'kernel') diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 60d9357e7172..c9a0af887033 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -1680,6 +1680,52 @@ static __init int sysctl_init(void) core_initcall(sysctl_init); +static int is_branch_in(struct ctl_table *branch, struct ctl_table *table) +{ + struct ctl_table *p; + const char *s = branch->procname; + + /* branch should have named subdirectory as its first element */ + if (!s || !branch->child) + return 0; + + /* ... and nothing else */ + if (branch[1].procname || branch[1].ctl_name) + return 0; + + /* table should contain subdirectory with the same name */ + for (p = table; p->procname || p->ctl_name; p++) { + if (!p->child) + continue; + if (p->procname && strcmp(p->procname, s) == 0) + return 1; + } + return 0; +} + +/* see if attaching q to p would be an improvement */ +static void try_attach(struct ctl_table_header *p, struct ctl_table_header *q) +{ + struct ctl_table *to = p->ctl_table, *by = q->ctl_table; + int is_better = 0; + int not_in_parent = !p->attached_by; + + while (is_branch_in(by, to)) { + if (by == q->attached_by) + is_better = 1; + if (to == p->attached_by) + not_in_parent = 1; + by = by->child; + to = to->child; + } + + if (is_better && not_in_parent) { + q->attached_by = by; + q->attached_to = to; + q->parent = p; + } +} + /** * __register_sysctl_paths - register a sysctl hierarchy * @root: List of sysctl headers to register on @@ -1759,6 +1805,7 @@ struct ctl_table_header *__register_sysctl_paths( struct ctl_table_header *header; struct ctl_table *new, **prevp; unsigned int n, npath; + struct ctl_table_set *set; /* Count the path components */ for (npath = 0; path[npath].ctl_name || path[npath].procname; ++npath) @@ -1809,6 +1856,18 @@ struct ctl_table_header *__register_sysctl_paths( #endif spin_lock(&sysctl_lock); header->set = lookup_header_set(root, namespaces); + header->attached_by = header->ctl_table; + header->attached_to = root_table; + header->parent = &root_table_header; + for (set = header->set; set; set = set->parent) { + struct ctl_table_header *p; + list_for_each_entry(p, &set->list, ctl_entry) { + if (p->unregistering) + continue; + try_attach(p, header); + } + } + header->parent->count++; list_add_tail(&header->ctl_entry, &header->set->list); spin_unlock(&sysctl_lock); @@ -1864,6 +1923,10 @@ void unregister_sysctl_table(struct ctl_table_header * header) spin_lock(&sysctl_lock); start_unregistering(header); + if (!--header->parent->count) { + WARN_ON(1); + kfree(header->parent); + } if (!--header->count) kfree(header); spin_unlock(&sysctl_lock); -- cgit v1.2.3 From 9043476f726802f4b00c96d0c4f418dde48d1304 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 15 Jul 2008 08:54:06 -0400 Subject: [PATCH] sanitize proc_sysctl * keep references to ctl_table_head and ctl_table in /proc/sys inodes * grab the former during operations, use the latter for access to entry if that succeeds * have ->d_compare() check if table should be seen for one who does lookup; that allows us to avoid flipping inodes - if we have the same name resolve to different things, we'll just keep several dentries and ->d_compare() will reject the wrong ones. * have ->lookup() and ->readdir() scan the table of our inode first, then walk all ctl_table_header and scan ->attached_by for those that are attached to our directory. * implement ->getattr(). * get rid of insane amounts of tree-walking * get rid of the need to know dentry in ->permission() and of the contortions induced by that. Signed-off-by: Al Viro --- kernel/sysctl.c | 15 +++++++++++++++ 1 file changed, 15 insertions(+) (limited to 'kernel') diff --git a/kernel/sysctl.c b/kernel/sysctl.c index c9a0af887033..ff5abcca5ddf 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -1932,6 +1932,21 @@ void unregister_sysctl_table(struct ctl_table_header * header) spin_unlock(&sysctl_lock); } +int sysctl_is_seen(struct ctl_table_header *p) +{ + struct ctl_table_set *set = p->set; + int res; + spin_lock(&sysctl_lock); + if (p->unregistering) + res = 0; + else if (!set->is_seen) + res = 1; + else + res = set->is_seen(set); + spin_unlock(&sysctl_lock); + return res; +} + void setup_sysctl_set(struct ctl_table_set *p, struct ctl_table_set *parent, int (*is_seen)(struct ctl_table_set *)) -- cgit v1.2.3 From e6305c43eda10ebfd2ad9e35d6e172ccc7bb3695 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 15 Jul 2008 21:03:57 -0400 Subject: [PATCH] sanitize ->permission() prototype * kill nameidata * argument; map the 3 bits in ->flags anybody cares about to new MAY_... ones and pass with the mask. * kill redundant gfs2_iop_permission() * sanitize ecryptfs_permission() * fix remaining places where ->permission() instances might barf on new MAY_... found in mask. The obvious next target in that direction is permission(9) folded fix for nfs_permission() breakage from Miklos Szeredi Signed-off-by: Al Viro --- kernel/sysctl.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/sysctl.c b/kernel/sysctl.c index ff5abcca5ddf..911d846f0503 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -1516,9 +1516,9 @@ static int do_sysctl_strategy(struct ctl_table_root *root, int op = 0, rc; if (oldval) - op |= 004; + op |= MAY_READ; if (newval) - op |= 002; + op |= MAY_WRITE; if (sysctl_perm(root, table, op)) return -EPERM; @@ -1560,7 +1560,7 @@ repeat: if (n == table->ctl_name) { int error; if (table->child) { - if (sysctl_perm(root, table, 001)) + if (sysctl_perm(root, table, MAY_EXEC)) return -EPERM; name++; nlen--; @@ -1635,7 +1635,7 @@ static int test_perm(int mode, int op) mode >>= 6; else if (in_egroup_p(0)) mode >>= 3; - if ((mode & op & 0007) == op) + if ((op & ~mode & (MAY_READ|MAY_WRITE|MAY_EXEC)) == 0) return 0; return -EACCES; } @@ -1645,7 +1645,7 @@ int sysctl_perm(struct ctl_table_root *root, struct ctl_table *table, int op) int error; int mode; - error = security_sysctl(table, op); + error = security_sysctl(table, op & (MAY_READ | MAY_WRITE | MAY_EXEC)); if (error) return error; -- cgit v1.2.3 From 7f2da1e7d0330395e5e9e350b879b98a1ea495df Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 10 May 2008 20:44:54 -0400 Subject: [PATCH] kill altroot long overdue... Signed-off-by: Al Viro --- kernel/exec_domain.c | 1 - kernel/exit.c | 2 -- kernel/fork.c | 7 ------- 3 files changed, 10 deletions(-) (limited to 'kernel') diff --git a/kernel/exec_domain.c b/kernel/exec_domain.c index c1ef192aa655..0d407e886735 100644 --- a/kernel/exec_domain.c +++ b/kernel/exec_domain.c @@ -168,7 +168,6 @@ __set_personality(u_long personality) current->personality = personality; oep = current_thread_info()->exec_domain; current_thread_info()->exec_domain = ep; - set_fs_altroot(); module_put(oep->module); return 0; diff --git a/kernel/exit.c b/kernel/exit.c index 6cdf60712bd2..0caf590548a0 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -565,8 +565,6 @@ void put_fs_struct(struct fs_struct *fs) if (atomic_dec_and_test(&fs->count)) { path_put(&fs->root); path_put(&fs->pwd); - if (fs->altroot.dentry) - path_put(&fs->altroot); kmem_cache_free(fs_cachep, fs); } } diff --git a/kernel/fork.c b/kernel/fork.c index abb3ed6298f6..5e050c1317c4 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -657,13 +657,6 @@ static struct fs_struct *__copy_fs_struct(struct fs_struct *old) path_get(&old->root); fs->pwd = old->pwd; path_get(&old->pwd); - if (old->altroot.dentry) { - fs->altroot = old->altroot; - path_get(&old->altroot); - } else { - fs->altroot.mnt = NULL; - fs->altroot.dentry = NULL; - } read_unlock(&old->lock); } return fs; -- cgit v1.2.3 From 3f8206d496e9e9495afb1d4e70d29712b4d403c9 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 26 Jul 2008 03:46:43 -0400 Subject: [PATCH] get rid of indirect users of namei.h fs.h needs path.h, not namei.h; nfs_fs.h doesn't need it at all. Several places in the tree needed direct include. Signed-off-by: Al Viro --- kernel/cgroup.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 89bd6fb7894f..657f8f8d93a5 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -45,6 +45,7 @@ #include #include #include +#include #include -- cgit v1.2.3