From 0f4bd46ec252887f44f1f065b41867cac8f70dfb Mon Sep 17 00:00:00 2001 From: KOSAKI Motohiro Date: Tue, 22 Dec 2009 03:15:43 +0000 Subject: kmsg_dump: Dump on crash_kexec as well crash_kexec gets called before kmsg_dump(KMSG_DUMP_OOPS) if panic_on_oops is set, so the kernel log buffer is not stored for this case. This patch adds a KMSG_DUMP_KEXEC dump type which gets called when crash_kexec() is invoked. To avoid getting double dumps, the old KMSG_DUMP_PANIC is moved below crash_kexec(). The mtdoops driver is modified to handle KMSG_DUMP_KEXEC in the same way as a panic. Signed-off-by: KOSAKI Motohiro Acked-by: Simon Kagstrom Signed-off-by: David Woodhouse --- kernel/kexec.c | 4 ++++ kernel/panic.c | 3 ++- kernel/printk.c | 1 + 3 files changed, 7 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/kexec.c b/kernel/kexec.c index 433e9fcc1fc5..ae217488fef8 100644 --- a/kernel/kexec.c +++ b/kernel/kexec.c @@ -32,6 +32,7 @@ #include #include #include +#include #include #include @@ -1074,6 +1075,9 @@ void crash_kexec(struct pt_regs *regs) if (mutex_trylock(&kexec_mutex)) { if (kexec_crash_image) { struct pt_regs fixed_regs; + + kmsg_dump(KMSG_DUMP_KEXEC); + crash_setup_regs(&fixed_regs, regs); crash_save_vmcoreinfo(); machine_crash_shutdown(&fixed_regs); diff --git a/kernel/panic.c b/kernel/panic.c index 5827f7b97254..c787333282b8 100644 --- a/kernel/panic.c +++ b/kernel/panic.c @@ -75,7 +75,6 @@ NORET_TYPE void panic(const char * fmt, ...) dump_stack(); #endif - kmsg_dump(KMSG_DUMP_PANIC); /* * If we have crashed and we have a crash kernel loaded let it handle * everything else. @@ -83,6 +82,8 @@ NORET_TYPE void panic(const char * fmt, ...) */ crash_kexec(NULL); + kmsg_dump(KMSG_DUMP_PANIC); + /* * Note smp_send_stop is the usual smp shutdown function, which * unfortunately means it may not be hardened to work in a panic diff --git a/kernel/printk.c b/kernel/printk.c index 1ded8e7dd19b..2c9dc0b03a5e 100644 --- a/kernel/printk.c +++ b/kernel/printk.c @@ -1467,6 +1467,7 @@ EXPORT_SYMBOL_GPL(kmsg_dump_unregister); static const char const *kmsg_reasons[] = { [KMSG_DUMP_OOPS] = "oops", [KMSG_DUMP_PANIC] = "panic", + [KMSG_DUMP_KEXEC] = "kexec", }; static const char *kmsg_to_str(enum kmsg_dump_reason reason) -- cgit v1.2.3 From ea9d8e3f45404d411c00ae67b45cc35c58265bb7 Mon Sep 17 00:00:00 2001 From: Xiaotian Feng Date: Thu, 7 Jan 2010 11:22:44 +0800 Subject: clockevent: Don't remove broadcast device when cpu is dead Marc reported that the BUG_ON in clockevents_notify() triggers on his system. This happens because the kernel tries to remove an active clock event device (used for broadcasting) from the device list. The handling of devices which can be used as per cpu device and as a global broadcast device is suboptimal. The simplest solution for now (and for stable) is to check whether the device is used as global broadcast device, but this needs to be revisited. [ tglx: restored the cpuweight check and massaged the changelog ] Reported-by: Marc Dionne Tested-by: Marc Dionne Signed-off-by: Xiaotian Feng LKML-Reference: <1262834564-13033-1-git-send-email-dfeng@redhat.com> Signed-off-by: Thomas Gleixner Cc: stable@kernel.org --- kernel/time/clockevents.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/time/clockevents.c b/kernel/time/clockevents.c index 6f740d9f0948..d7395fdfb9f3 100644 --- a/kernel/time/clockevents.c +++ b/kernel/time/clockevents.c @@ -259,7 +259,8 @@ void clockevents_notify(unsigned long reason, void *arg) cpu = *((int *)arg); list_for_each_entry_safe(dev, tmp, &clockevent_devices, list) { if (cpumask_test_cpu(cpu, dev->cpumask) && - cpumask_weight(dev->cpumask) == 1) { + cpumask_weight(dev->cpumask) == 1 && + !tick_is_broadcast_device(dev)) { BUG_ON(dev->mode != CLOCK_EVT_MODE_UNUSED); list_del(&dev->list); } -- cgit v1.2.3 From 50b926e439620c469565e8be0f28be78f5fca1ce Mon Sep 17 00:00:00 2001 From: Mike Galbraith Date: Mon, 4 Jan 2010 14:44:56 +0100 Subject: sched: Fix vmark regression on big machines SD_PREFER_SIBLING is set at the CPU domain level if power saving isn't enabled, leading to many cache misses on large machines as we traverse looking for an idle shared cache to wake to. Change the enabler of select_idle_sibling() to SD_SHARE_PKG_RESOURCES, and enable same at the sibling domain level. Reported-by: Lin Ming Signed-off-by: Mike Galbraith Signed-off-by: Peter Zijlstra LKML-Reference: <1262612696.15495.15.camel@marge.simson.net> Signed-off-by: Ingo Molnar --- kernel/sched_fair.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 42ac3c9f66f6..8fe7ee81c552 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -1508,7 +1508,7 @@ static int select_task_rq_fair(struct task_struct *p, int sd_flag, int wake_flag * If there's an idle sibling in this domain, make that * the wake_affine target instead of the current cpu. */ - if (tmp->flags & SD_PREFER_SIBLING) + if (tmp->flags & SD_SHARE_PKG_RESOURCES) target = select_idle_sibling(p, tmp, target); if (target >= 0) { -- cgit v1.2.3 From 6d558c3ac9b6508d26fd5cadccce51fc9d726b1c Mon Sep 17 00:00:00 2001 From: Yong Zhang Date: Mon, 11 Jan 2010 14:21:25 +0800 Subject: sched: Reassign prev and switch_count when reacquire_kernel_lock() fail Assume A->B schedule is processing, if B have acquired BKL before and it need reschedule this time. Then on B's context, it will go to need_resched_nonpreemptible for reschedule. But at this time, prev and switch_count are related to A. It's wrong and will lead to incorrect scheduler statistics. Signed-off-by: Yong Zhang Signed-off-by: Peter Zijlstra LKML-Reference: <2674af741001102238w7b0ddcadref00d345e2181d11@mail.gmail.com> Signed-off-by: Ingo Molnar --- kernel/sched.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index c535cc4f6428..4508fe7048be 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -5530,8 +5530,11 @@ need_resched_nonpreemptible: post_schedule(rq); - if (unlikely(reacquire_kernel_lock(current) < 0)) + if (unlikely(reacquire_kernel_lock(current) < 0)) { + prev = rq->curr; + switch_count = &prev->nivcsw; goto need_resched_nonpreemptible; + } preempt_enable_no_resched(); if (need_resched()) -- cgit v1.2.3 From fe432200abb0d64f409895168d9ad8fbb9d8e6c6 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 18 Jan 2010 09:08:26 +0100 Subject: perf: Fix perf_event_do_pending() fallback callsite Paul questioned the context in which we should call perf_event_do_pending(). After looking at that I found that it should be called from IRQ context these days, however the fallback call-site is placed in softirq context. Ammend this by placing the callback in the IRQ timer path. Reported-by: Paul Mackerras Signed-off-by: Peter Zijlstra LKML-Reference: <1263374859.4244.192.camel@laptop> Signed-off-by: Ingo Molnar --- kernel/timer.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/timer.c b/kernel/timer.c index 15533b792397..c61a7949387f 100644 --- a/kernel/timer.c +++ b/kernel/timer.c @@ -1198,6 +1198,7 @@ void update_process_times(int user_tick) run_local_timers(); rcu_check_callbacks(cpu, user_tick); printk_tick(); + perf_event_do_pending(); scheduler_tick(); run_posix_cpu_timers(p); } @@ -1209,8 +1210,6 @@ static void run_timer_softirq(struct softirq_action *h) { struct tvec_base *base = __get_cpu_var(tvec_bases); - perf_event_do_pending(); - hrtimer_run_pending(); if (time_after_eq(jiffies, base->timer_jiffies)) -- cgit v1.2.3 From 22e190851f8709c48baf00ed9ce6144cdc54d025 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 18 Jan 2010 09:12:32 +0100 Subject: perf: Honour event state for aux stream data Anton reported that perf record kept receiving events even after calling ioctl(PERF_EVENT_IOC_DISABLE). It turns out that FORK,COMM and MMAP events didn't respect the disabled state and kept flowing in. Reported-by: Anton Blanchard Signed-off-by: Peter Zijlstra Tested-by: Anton Blanchard LKML-Reference: <1263459187.4244.265.camel@laptop> CC: stable@kernel.org Signed-off-by: Ingo Molnar --- kernel/perf_event.c | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'kernel') diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 603c0d8b5df1..d27746bd3a06 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -3268,6 +3268,9 @@ static void perf_event_task_output(struct perf_event *event, static int perf_event_task_match(struct perf_event *event) { + if (event->state != PERF_EVENT_STATE_ACTIVE) + return 0; + if (event->cpu != -1 && event->cpu != smp_processor_id()) return 0; @@ -3377,6 +3380,9 @@ static void perf_event_comm_output(struct perf_event *event, static int perf_event_comm_match(struct perf_event *event) { + if (event->state != PERF_EVENT_STATE_ACTIVE) + return 0; + if (event->cpu != -1 && event->cpu != smp_processor_id()) return 0; @@ -3494,6 +3500,9 @@ static void perf_event_mmap_output(struct perf_event *event, static int perf_event_mmap_match(struct perf_event *event, struct perf_mmap_event *mmap_event) { + if (event->state != PERF_EVENT_STATE_ACTIVE) + return 0; + if (event->cpu != -1 && event->cpu != smp_processor_id()) return 0; -- cgit v1.2.3 From fabf318e5e4bda0aca2b0d617b191884fda62703 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 21 Jan 2010 21:04:57 +0100 Subject: sched: Fix fork vs hotplug vs cpuset namespaces There are a number of issues: 1) TASK_WAKING vs cgroup_clone (cpusets) copy_process(): sched_fork() child->state = TASK_WAKING; /* waiting for wake_up_new_task() */ if (current->nsproxy != p->nsproxy) ns_cgroup_clone() cgroup_clone() mutex_lock(inode->i_mutex) mutex_lock(cgroup_mutex) cgroup_attach_task() ss->can_attach() ss->attach() [ -> cpuset_attach() ] cpuset_attach_task() set_cpus_allowed_ptr(); while (child->state == TASK_WAKING) cpu_relax(); will deadlock the system. 2) cgroup_clone (cpusets) vs copy_process So even if the above would work we still have: copy_process(): if (current->nsproxy != p->nsproxy) ns_cgroup_clone() cgroup_clone() mutex_lock(inode->i_mutex) mutex_lock(cgroup_mutex) cgroup_attach_task() ss->can_attach() ss->attach() [ -> cpuset_attach() ] cpuset_attach_task() set_cpus_allowed_ptr(); ... p->cpus_allowed = current->cpus_allowed over-writing the modified cpus_allowed. 3) fork() vs hotplug if we unplug the child's cpu after the sanity check when the child gets attached to the task_list but before wake_up_new_task() shit will meet with fan. Solve all these issues by moving fork cpu selection into wake_up_new_task(). Reported-by: Serge E. Hallyn Tested-by: Serge E. Hallyn Signed-off-by: Peter Zijlstra LKML-Reference: <1264106190.4283.1314.camel@laptop> Signed-off-by: Thomas Gleixner --- kernel/fork.c | 15 --------------- kernel/sched.c | 39 +++++++++++++++++++++++++++------------ 2 files changed, 27 insertions(+), 27 deletions(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index 5b2959b3ffc2..f88bd984df35 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1241,21 +1241,6 @@ static struct task_struct *copy_process(unsigned long clone_flags, /* Need tasklist lock for parent etc handling! */ write_lock_irq(&tasklist_lock); - /* - * The task hasn't been attached yet, so its cpus_allowed mask will - * not be changed, nor will its assigned CPU. - * - * The cpus_allowed mask of the parent may have changed after it was - * copied first time - so re-copy it here, then check the child's CPU - * to ensure it is on a valid CPU (and if not, just force it back to - * parent's CPU). This avoids alot of nasty races. - */ - p->cpus_allowed = current->cpus_allowed; - p->rt.nr_cpus_allowed = current->rt.nr_cpus_allowed; - if (unlikely(!cpu_isset(task_cpu(p), p->cpus_allowed) || - !cpu_online(task_cpu(p)))) - set_task_cpu(p, smp_processor_id()); - /* CLONE_PARENT re-uses the old parent */ if (clone_flags & (CLONE_PARENT|CLONE_THREAD)) { p->real_parent = current->real_parent; diff --git a/kernel/sched.c b/kernel/sched.c index 4508fe7048be..3a8fb30a91b1 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -2320,14 +2320,12 @@ static int select_fallback_rq(int cpu, struct task_struct *p) } /* - * Called from: + * Gets called from 3 sites (exec, fork, wakeup), since it is called without + * holding rq->lock we need to ensure ->cpus_allowed is stable, this is done + * by: * - * - fork, @p is stable because it isn't on the tasklist yet - * - * - exec, @p is unstable, retry loop - * - * - wake-up, we serialize ->cpus_allowed against TASK_WAKING so - * we should be good. + * exec: is unstable, retry loop + * fork & wake-up: serialize ->cpus_allowed against TASK_WAKING */ static inline int select_task_rq(struct task_struct *p, int sd_flags, int wake_flags) @@ -2620,9 +2618,6 @@ void sched_fork(struct task_struct *p, int clone_flags) if (p->sched_class->task_fork) p->sched_class->task_fork(p); -#ifdef CONFIG_SMP - cpu = select_task_rq(p, SD_BALANCE_FORK, 0); -#endif set_task_cpu(p, cpu); #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) @@ -2652,6 +2647,21 @@ void wake_up_new_task(struct task_struct *p, unsigned long clone_flags) { unsigned long flags; struct rq *rq; + int cpu = get_cpu(); + +#ifdef CONFIG_SMP + /* + * Fork balancing, do it here and not earlier because: + * - cpus_allowed can change in the fork path + * - any previously selected cpu might disappear through hotplug + * + * We still have TASK_WAKING but PF_STARTING is gone now, meaning + * ->cpus_allowed is stable, we have preemption disabled, meaning + * cpu_online_mask is stable. + */ + cpu = select_task_rq(p, SD_BALANCE_FORK, 0); + set_task_cpu(p, cpu); +#endif rq = task_rq_lock(p, &flags); BUG_ON(p->state != TASK_WAKING); @@ -2665,6 +2675,7 @@ void wake_up_new_task(struct task_struct *p, unsigned long clone_flags) p->sched_class->task_woken(rq, p); #endif task_rq_unlock(rq, &flags); + put_cpu(); } #ifdef CONFIG_PREEMPT_NOTIFIERS @@ -7139,14 +7150,18 @@ int set_cpus_allowed_ptr(struct task_struct *p, const struct cpumask *new_mask) * the ->cpus_allowed mask from under waking tasks, which would be * possible when we change rq->lock in ttwu(), so synchronize against * TASK_WAKING to avoid that. + * + * Make an exception for freshly cloned tasks, since cpuset namespaces + * might move the task about, we have to validate the target in + * wake_up_new_task() anyway since the cpu might have gone away. */ again: - while (p->state == TASK_WAKING) + while (p->state == TASK_WAKING && !(p->flags & PF_STARTING)) cpu_relax(); rq = task_rq_lock(p, &flags); - if (p->state == TASK_WAKING) { + if (p->state == TASK_WAKING && !(p->flags & PF_STARTING)) { task_rq_unlock(rq, &flags); goto again; } -- cgit v1.2.3 From 74bf4076f2ed79b5510440b72a561823a8852ec0 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 25 Jan 2010 15:11:53 -0500 Subject: tracing: Prevent kernel oops with corrupted buffer If the contents of the ftrace ring buffer gets corrupted and the trace file is read, it could create a kernel oops (usualy just killing the user task thread). This is caused by the checking of the pid in the buffer. If the pid is negative, it still references the cmdline cache array, which could point to an invalid address. The simple fix is to test for negative PIDs. Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 0df1b0f2cb9e..eac6875cb990 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -951,6 +951,11 @@ void trace_find_cmdline(int pid, char comm[]) return; } + if (WARN_ON_ONCE(pid < 0)) { + strcpy(comm, ""); + return; + } + if (pid > PID_MAX_DEFAULT) { strcpy(comm, "<...>"); return; -- cgit v1.2.3 From 7b7422a566aa0dc1e582ce263d4c7ff4a772700a Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 26 Jan 2010 12:51:10 +0100 Subject: clocksource: Prevent potential kgdb dead lock commit 0f8e8ef7 (clocksource: Simplify clocksource watchdog resume logic) introduced a potential kgdb dead lock. When the kernel is stopped by kgdb inside code which holds watchdog_lock then kgdb dead locks in clocksource_resume_watchdog(). clocksource_resume_watchdog() is called from kbdg via clocksource_touch_watchdog() to avoid that the clock source watchdog marks TSC unstable after the kernel has been stopped. Solve this by replacing spin_lock with a spin_trylock and just return in case the lock is held. Not resetting the watchdog might result in TSC becoming marked unstable, but that's an acceptable penalty for using kgdb. The timekeeping is anyway easily screwed up by kgdb when the system uses either jiffies or a clock source which wraps in short intervals (e.g. pm_timer wraps about every 4.6s), so we really do not have to worry about that occasional TSC marked unstable side effect. The second caller of clocksource_resume_watchdog() is clocksource_resume(). The trylock is safe here as well because the system is UP at this point, interrupts are disabled and nothing else can hold watchdog_lock(). Reported-by: Jason Wessel LKML-Reference: <1264480000-6997-4-git-send-email-jason.wessel@windriver.com> Cc: kgdb-bugreport@lists.sourceforge.net Cc: Martin Schwidefsky Cc: John Stultz Cc: Andrew Morton Signed-off-by: Thomas Gleixner --- kernel/time/clocksource.c | 18 +++++++++++++++--- 1 file changed, 15 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/time/clocksource.c b/kernel/time/clocksource.c index e85c23404d34..13700833c181 100644 --- a/kernel/time/clocksource.c +++ b/kernel/time/clocksource.c @@ -343,7 +343,19 @@ static void clocksource_resume_watchdog(void) { unsigned long flags; - spin_lock_irqsave(&watchdog_lock, flags); + /* + * We use trylock here to avoid a potential dead lock when + * kgdb calls this code after the kernel has been stopped with + * watchdog_lock held. When watchdog_lock is held we just + * return and accept, that the watchdog might trigger and mark + * the monitored clock source (usually TSC) unstable. + * + * This does not affect the other caller clocksource_resume() + * because at this point the kernel is UP, interrupts are + * disabled and nothing can hold watchdog_lock. + */ + if (!spin_trylock_irqsave(&watchdog_lock, flags)) + return; clocksource_reset_watchdog(); spin_unlock_irqrestore(&watchdog_lock, flags); } @@ -458,8 +470,8 @@ void clocksource_resume(void) * clocksource_touch_watchdog - Update watchdog * * Update the watchdog after exception contexts such as kgdb so as not - * to incorrectly trip the watchdog. - * + * to incorrectly trip the watchdog. This might fail when the kernel + * was stopped in code which holds watchdog_lock. */ void clocksource_touch_watchdog(void) { -- cgit v1.2.3 From 492a74f4210e15f4701422e2e1c4cd3c1e45ddae Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 25 Jan 2010 15:17:47 -0500 Subject: ring-buffer: Check if ring buffer iterator has stale data Usually reads of the ring buffer is performed by a single task. There are two types of reads from the ring buffer. One is a consuming read which will consume the entry that was read and the next read will be the entry that follows. The other is an iterator that will let the user read the contents of the ring buffer without modifying it. When an iterator is allocated, writes to the ring buffer are disabled to protect the iterator. The problem exists when consuming reads happen while an iterator is allocated. Specifically, the kind of read that swaps out an entire page (used by splice) and replaces it with a new read. If the iterator is on the page that is swapped out, then the next read may read from this swapped out page and return garbage. This patch adds a check when reading the iterator to make sure that the iterator contents are still valid. If a consuming read has taken place, the iterator is reset. Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'kernel') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index edefe3b2801b..503b630e0bda 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -464,6 +464,8 @@ struct ring_buffer_iter { struct ring_buffer_per_cpu *cpu_buffer; unsigned long head; struct buffer_page *head_page; + struct buffer_page *cache_reader_page; + unsigned long cache_read; u64 read_stamp; }; @@ -2716,6 +2718,8 @@ static void rb_iter_reset(struct ring_buffer_iter *iter) iter->read_stamp = cpu_buffer->read_stamp; else iter->read_stamp = iter->head_page->page->time_stamp; + iter->cache_reader_page = cpu_buffer->reader_page; + iter->cache_read = cpu_buffer->read; } /** @@ -3066,6 +3070,15 @@ rb_iter_peek(struct ring_buffer_iter *iter, u64 *ts) cpu_buffer = iter->cpu_buffer; buffer = cpu_buffer->buffer; + /* + * Check if someone performed a consuming read to + * the buffer. A consuming read invalidates the iterator + * and we need to reset the iterator in this case. + */ + if (unlikely(iter->cache_read != cpu_buffer->read || + iter->cache_reader_page != cpu_buffer->reader_page)) + rb_iter_reset(iter); + again: /* * We repeat when a timestamp is encountered. -- cgit v1.2.3 From 3c05d7482777f15e71bb4cb1ba78dee2800dfec6 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 26 Jan 2010 16:14:08 -0500 Subject: ring-buffer: Check for end of page in iterator If the iterator comes to an empty page for some reason, or if the page is emptied by a consuming read. The iterator code currently does not check if the iterator is pass the contents, and may return a false entry. This patch adds a check to the ring buffer iterator to test if the current page has been completely read and sets the iterator to the next page if necessary. Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 503b630e0bda..8c1b2d290718 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -3064,9 +3064,6 @@ rb_iter_peek(struct ring_buffer_iter *iter, u64 *ts) struct ring_buffer_event *event; int nr_loops = 0; - if (ring_buffer_iter_empty(iter)) - return NULL; - cpu_buffer = iter->cpu_buffer; buffer = cpu_buffer->buffer; @@ -3080,6 +3077,9 @@ rb_iter_peek(struct ring_buffer_iter *iter, u64 *ts) rb_iter_reset(iter); again: + if (ring_buffer_iter_empty(iter)) + return NULL; + /* * We repeat when a timestamp is encountered. * We can get multiple timestamps by nested interrupts or also @@ -3094,6 +3094,11 @@ rb_iter_peek(struct ring_buffer_iter *iter, u64 *ts) if (rb_per_cpu_empty(cpu_buffer)) return NULL; + if (iter->head >= local_read(&iter->head_page->page->commit)) { + rb_inc_iter(iter); + goto again; + } + event = rb_iter_head_event(iter); switch (event->type_len) { -- cgit v1.2.3 From 03688970347bfea32823953a7ce5886d1713205f Mon Sep 17 00:00:00 2001 From: Mike Frysinger Date: Fri, 22 Jan 2010 08:12:47 -0500 Subject: tracing/documentation: Cover new frame pointer semantics Update the graph tracer examples to cover the new frame pointer semantics (in terms of passing it along). Move the HAVE_FUNCTION_GRAPH_FP_TEST docs out of the Kconfig, into the right place, and expand on the details. Signed-off-by: Mike Frysinger LKML-Reference: <1264165967-18938-1-git-send-email-vapier@gentoo.org> Signed-off-by: Steven Rostedt --- kernel/trace/Kconfig | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 6c22d8a2f289..60e2ce0181ee 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -27,9 +27,7 @@ config HAVE_FUNCTION_GRAPH_TRACER config HAVE_FUNCTION_GRAPH_FP_TEST bool help - An arch may pass in a unique value (frame pointer) to both the - entering and exiting of a function. On exit, the value is compared - and if it does not match, then it will panic the kernel. + See Documentation/trace/ftrace-design.txt config HAVE_FUNCTION_TRACE_MCOUNT_TEST bool -- cgit v1.2.3 From 48d50674179981e41f432167b2441cec782d5484 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Tue, 26 Jan 2010 19:16:41 +0100 Subject: lockdep: Fix check_usage_backwards() error message Lockdep has found the real bug, but the output doesn't look right to me: > ========================================================= > [ INFO: possible irq lock inversion dependency detected ] > 2.6.33-rc5 #77 > --------------------------------------------------------- > emacs/1609 just changed the state of lock: > (&(&tty->ctrl_lock)->rlock){+.....}, at: [] tty_fasync+0xe8/0x190 > but this lock took another, HARDIRQ-unsafe lock in the past: > (&(&sighand->siglock)->rlock){-.....} "HARDIRQ-unsafe" and "this lock took another" looks wrong, afaics. > ... key at: [] __key.46539+0x0/0x8 > ... acquired at: > [] __lock_acquire+0x1056/0x15a0 > [] lock_acquire+0x9f/0x120 > [] _raw_spin_lock_irqsave+0x52/0x90 > [] __proc_set_tty+0x3e/0x150 > [] tty_open+0x51d/0x5e0 The stack-trace shows that this lock (ctrl_lock) was taken under ->siglock (which is hopefully irq-safe). This is a clear typo in check_usage_backwards() where we tell the print a fancy routine we're forwards. Signed-off-by: Oleg Nesterov Signed-off-by: Peter Zijlstra LKML-Reference: <20100126181641.GA10460@redhat.com> Signed-off-by: Ingo Molnar --- kernel/lockdep.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/lockdep.c b/kernel/lockdep.c index 5feaddcdbe49..c62ec14609b9 100644 --- a/kernel/lockdep.c +++ b/kernel/lockdep.c @@ -2147,7 +2147,7 @@ check_usage_backwards(struct task_struct *curr, struct held_lock *this, return ret; return print_irq_inversion_bug(curr, &root, target_entry, - this, 1, irqclass); + this, 0, irqclass); } void print_irqtrace_events(struct task_struct *curr) -- cgit v1.2.3 From 11854247e2c851e7ff9ce138e501c6cffc5a4217 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 21 Jan 2010 16:34:27 +0100 Subject: sched: Fix incorrect sanity check We moved to migrate on wakeup, which means that sleeping tasks could still be present on offline cpus. Amend the check to only test running tasks. Reported-by: Heiko Carstens Signed-off-by: Peter Zijlstra Signed-off-by: Ingo Molnar --- kernel/cpu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/cpu.c b/kernel/cpu.c index 1c8ddd6ee940..08e54e7beaed 100644 --- a/kernel/cpu.c +++ b/kernel/cpu.c @@ -151,7 +151,7 @@ static inline void check_for_tasks(int cpu) write_lock_irq(&tasklist_lock); for_each_process(p) { - if (task_cpu(p) == cpu && + if (task_cpu(p) == cpu && p->state == TASK_RUNNING && (!cputime_eq(p->utime, cputime_zero) || !cputime_eq(p->stime, cputime_zero))) printk(KERN_WARNING "Task %s (pid = %d) is on cpu %d\ -- cgit v1.2.3 From 9d3cfc4c1d17c6d3bc1373e3b954c56b92607755 Mon Sep 17 00:00:00 2001 From: Frans Pop Date: Mon, 25 Jan 2010 14:56:34 +0100 Subject: sched: Correct printk whitespace in warning from cpu down task check Due to an incorrect line break the output currently contains tabs. Also remove trailing space. The actual output that logcheck sent me looked like this: Task events/1 (pid = 10) is on cpu 1^I^I^I^I(state = 1, flags = 84208040) After this patch it becomes: Task events/1 (pid = 10) is on cpu 1 (state = 1, flags = 84208040) Signed-off-by: Frans Pop Signed-off-by: Peter Zijlstra LKML-Reference: <201001251456.34996.elendil@planet.nl> Signed-off-by: Ingo Molnar --- kernel/cpu.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/cpu.c b/kernel/cpu.c index 08e54e7beaed..677f25376a38 100644 --- a/kernel/cpu.c +++ b/kernel/cpu.c @@ -154,10 +154,10 @@ static inline void check_for_tasks(int cpu) if (task_cpu(p) == cpu && p->state == TASK_RUNNING && (!cputime_eq(p->utime, cputime_zero) || !cputime_eq(p->stime, cputime_zero))) - printk(KERN_WARNING "Task %s (pid = %d) is on cpu %d\ - (state = %ld, flags = %x) \n", - p->comm, task_pid_nr(p), cpu, - p->state, p->flags); + printk(KERN_WARNING "Task %s (pid = %d) is on cpu %d " + "(state = %ld, flags = %x)\n", + p->comm, task_pid_nr(p), cpu, + p->state, p->flags); } write_unlock_irq(&tasklist_lock); } -- cgit v1.2.3 From b23ff0e9330e4b11e18af984d50573598e10e7f9 Mon Sep 17 00:00:00 2001 From: Mahesh Salgaonkar Date: Thu, 21 Jan 2010 18:25:16 +0530 Subject: hw_breakpoints: Release the bp slot if arch_validate_hwbkpt_settings() fails. On a given architecture, when hardware breakpoint registration fails due to un-supported access type (read/write/execute), we lose the bp slot since register_perf_hw_breakpoint() does not release the bp slot on failure. Hence, any subsequent hardware breakpoint registration starts failing with 'no space left on device' error. This patch introduces error handling in register_perf_hw_breakpoint() function and releases bp slot on error. Signed-off-by: Mahesh Salgaonkar Cc: Ananth N Mavinakayanahalli Cc: K. Prasad Cc: Maneesh Soni LKML-Reference: <20100121125516.GA32521@in.ibm.com> Signed-off-by: Frederic Weisbecker --- kernel/hw_breakpoint.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'kernel') diff --git a/kernel/hw_breakpoint.c b/kernel/hw_breakpoint.c index 50dbd5999588..c030ae657f20 100644 --- a/kernel/hw_breakpoint.c +++ b/kernel/hw_breakpoint.c @@ -296,6 +296,10 @@ int register_perf_hw_breakpoint(struct perf_event *bp) if (!bp->attr.disabled || !bp->overflow_handler) ret = arch_validate_hwbkpt_settings(bp, bp->ctx->task); + /* if arch_validate_hwbkpt_settings() fails then release bp slot */ + if (ret) + release_bp_slot(bp); + return ret; } -- cgit v1.2.3 From cc0967490c1c3824bc5b75718b6ca8a51d9f2617 Mon Sep 17 00:00:00 2001 From: Jason Wessel Date: Thu, 28 Jan 2010 17:04:42 -0600 Subject: x86, hw_breakpoints, kgdb: Fix kgdb to use hw_breakpoint API In the 2.6.33 kernel, the hw_breakpoint API is now used for the performance event counters. The hw_breakpoint_handler() now consumes the hw breakpoints that were previously set by kgdb arch specific code. In order for kgdb to work in conjunction with this core API change, kgdb must use some of the low level functions of the hw_breakpoint API to install, uninstall, and deal with hw breakpoint reservations. The kgdb core required a change to call kgdb_disable_hw_debug anytime a slave cpu enters kgdb_wait() in order to keep all the hw breakpoints in sync as well as to prevent hitting a hw breakpoint while kgdb is active. During the architecture specific initialization of kgdb, it will pre-allocate 4 disabled (struct perf event **) structures. Kgdb will use these to manage the capabilities for the 4 hw breakpoint registers, per cpu. Right now the hw_breakpoint API does not have a way to ask how many breakpoints are available, on each CPU so it is possible that the install of a breakpoint might fail when kgdb restores the system to the run state. The intent of this patch is to first get the basic functionality of hw breakpoints working and leave it to the person debugging the kernel to understand what hw breakpoints are in use and what restrictions have been imposed as a result. Breakpoint constraints will be dealt with in a future patch. While atomic, the x86 specific kgdb code will call arch_uninstall_hw_breakpoint() and arch_install_hw_breakpoint() to manage the cpu specific hw breakpoints. The net result of these changes allow kgdb to use the same pool of hw_breakpoints that are used by the perf event API, but neither knows about future reservations for the available hw breakpoint slots. Signed-off-by: Jason Wessel Acked-by: Frederic Weisbecker Cc: kgdb-bugreport@lists.sourceforge.net Cc: K.Prasad Cc: Peter Zijlstra Cc: Alan Stern Cc: torvalds@linux-foundation.org LKML-Reference: <1264719883-7285-2-git-send-email-jason.wessel@windriver.com> Signed-off-by: Ingo Molnar --- kernel/kgdb.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'kernel') diff --git a/kernel/kgdb.c b/kernel/kgdb.c index 2eb517e23514..c7ade62e4ef0 100644 --- a/kernel/kgdb.c +++ b/kernel/kgdb.c @@ -583,6 +583,9 @@ static void kgdb_wait(struct pt_regs *regs) smp_wmb(); atomic_set(&cpu_in_kgdb[cpu], 1); + /* Disable any cpu specific hw breakpoints */ + kgdb_disable_hw_debug(regs); + /* Wait till primary CPU is done with debugging */ while (atomic_read(&passive_cpu_wait[cpu])) cpu_relax(); -- cgit v1.2.3 From 5352ae638e2d7d5c9b2e4d528676bbf2af6fd6f3 Mon Sep 17 00:00:00 2001 From: Jason Wessel Date: Thu, 28 Jan 2010 17:04:43 -0600 Subject: perf, hw_breakpoint, kgdb: Do not take mutex for kernel debugger This patch fixes the regression in functionality where the kernel debugger and the perf API do not nicely share hw breakpoint reservations. The kernel debugger cannot use any mutex_lock() calls because it can start the kernel running from an invalid context. A mutex free version of the reservation API needed to get created for the kernel debugger to safely update hw breakpoint reservations. The possibility for a breakpoint reservation to be concurrently processed at the time that kgdb interrupts the system is improbable. Should this corner case occur the end user is warned, and the kernel debugger will prohibit updating the hardware breakpoint reservations. Any time the kernel debugger reserves a hardware breakpoint it will be a system wide reservation. Signed-off-by: Jason Wessel Acked-by: Frederic Weisbecker Cc: kgdb-bugreport@lists.sourceforge.net Cc: K.Prasad Cc: Peter Zijlstra Cc: Alan Stern Cc: torvalds@linux-foundation.org LKML-Reference: <1264719883-7285-3-git-send-email-jason.wessel@windriver.com> Signed-off-by: Ingo Molnar --- kernel/hw_breakpoint.c | 52 ++++++++++++++++++++++++++++++++++++++++---------- 1 file changed, 42 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/hw_breakpoint.c b/kernel/hw_breakpoint.c index c030ae657f20..8a5c7d55ac9f 100644 --- a/kernel/hw_breakpoint.c +++ b/kernel/hw_breakpoint.c @@ -243,38 +243,70 @@ static void toggle_bp_slot(struct perf_event *bp, bool enable) * ((per_cpu(nr_bp_flexible, *) > 1) + max(per_cpu(nr_cpu_bp_pinned, *)) * + max(per_cpu(nr_task_bp_pinned, *))) < HBP_NUM */ -int reserve_bp_slot(struct perf_event *bp) +static int __reserve_bp_slot(struct perf_event *bp) { struct bp_busy_slots slots = {0}; - int ret = 0; - - mutex_lock(&nr_bp_mutex); fetch_bp_busy_slots(&slots, bp); /* Flexible counters need to keep at least one slot */ - if (slots.pinned + (!!slots.flexible) == HBP_NUM) { - ret = -ENOSPC; - goto end; - } + if (slots.pinned + (!!slots.flexible) == HBP_NUM) + return -ENOSPC; toggle_bp_slot(bp, true); -end: + return 0; +} + +int reserve_bp_slot(struct perf_event *bp) +{ + int ret; + + mutex_lock(&nr_bp_mutex); + + ret = __reserve_bp_slot(bp); + mutex_unlock(&nr_bp_mutex); return ret; } +static void __release_bp_slot(struct perf_event *bp) +{ + toggle_bp_slot(bp, false); +} + void release_bp_slot(struct perf_event *bp) { mutex_lock(&nr_bp_mutex); - toggle_bp_slot(bp, false); + __release_bp_slot(bp); mutex_unlock(&nr_bp_mutex); } +/* + * Allow the kernel debugger to reserve breakpoint slots without + * taking a lock using the dbg_* variant of for the reserve and + * release breakpoint slots. + */ +int dbg_reserve_bp_slot(struct perf_event *bp) +{ + if (mutex_is_locked(&nr_bp_mutex)) + return -1; + + return __reserve_bp_slot(bp); +} + +int dbg_release_bp_slot(struct perf_event *bp) +{ + if (mutex_is_locked(&nr_bp_mutex)) + return -1; + + __release_bp_slot(bp); + + return 0; +} int register_perf_hw_breakpoint(struct perf_event *bp) { -- cgit v1.2.3