From 84e478c6f1eb9c4bfa1fff2f8108e9a061b46428 Mon Sep 17 00:00:00 2001 From: Don Zickus Date: Fri, 5 Feb 2010 21:47:05 -0500 Subject: nmi_watchdog: Config option to enable new nmi_watchdog These are the bits that enable the new nmi_watchdog and safely isolate the old nmi_watchdog. Only one or the other can run, not both at the same time. Signed-off-by: Don Zickus Cc: Linus Torvalds Cc: Andrew Morton Cc: gorcunov@gmail.com Cc: aris@redhat.com Cc: peterz@infradead.org LKML-Reference: <1265424425-31562-4-git-send-email-dzickus@redhat.com> Signed-off-by: Ingo Molnar --- lib/Kconfig.debug | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'lib') diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index 25c3ed594c5..f80b67e72aa 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -170,6 +170,19 @@ config DETECT_SOFTLOCKUP can be detected via the NMI-watchdog, on platforms that support it.) +config NMI_WATCHDOG + bool "Detect Hard Lockups with an NMI Watchdog" + depends on DEBUG_KERNEL && PERF_EVENTS + default y + help + Say Y here to enable the kernel to use the NMI as a watchdog + to detect hard lockups. This is useful when a cpu hangs for no + reason but can still respond to NMIs. A backtrace is displayed + for reviewing and reporting. + + The overhead should be minimal, just an extra NMI every few + seconds. + config BOOTPARAM_SOFTLOCKUP_PANIC bool "Panic (Reboot) On Soft Lockups" depends on DETECT_SOFTLOCKUP -- cgit v1.2.3-70-g09d2 From 8e7672cdb413af859086ffceaed68f7e1e8ea4c2 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Tue, 9 Feb 2010 06:11:00 +0100 Subject: nmi_watchdog: Only enable on x86 for now It wont even build on other platforms just yet - so restrict it to x86 for now. Cc: Don Zickus Cc: gorcunov@gmail.com Cc: aris@redhat.com Cc: peterz@infradead.org LKML-Reference: <1265424425-31562-4-git-send-email-dzickus@redhat.com> Signed-off-by: Ingo Molnar --- lib/Kconfig.debug | 1 + 1 file changed, 1 insertion(+) (limited to 'lib') diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index f80b67e72aa..acef88239e1 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -173,6 +173,7 @@ config DETECT_SOFTLOCKUP config NMI_WATCHDOG bool "Detect Hard Lockups with an NMI Watchdog" depends on DEBUG_KERNEL && PERF_EVENTS + depends on X86 default y help Say Y here to enable the kernel to use the NMI as a watchdog -- cgit v1.2.3-70-g09d2 From c3128fb6ad39b0edda6675d20585a64846cf89ea Mon Sep 17 00:00:00 2001 From: Don Zickus Date: Fri, 12 Feb 2010 17:19:18 -0500 Subject: nmi_watchdog: Use a boolean config flag for compiling Determines if an arch has setup arch specific perf_events and nmi_watchdog code. This should restrict compiles to only those arches ready. Signed-off-by: Don Zickus Cc: peterz@infradead.org Cc: gorcunov@gmail.com Cc: aris@redhat.com LKML-Reference: <1266013161-31197-1-git-send-email-dzickus@redhat.com> Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 1 + init/Kconfig | 5 +++++ lib/Kconfig.debug | 3 +-- 3 files changed, 7 insertions(+), 2 deletions(-) (limited to 'lib') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index cbcbfdee3ee..4f9685fa3a3 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -52,6 +52,7 @@ config X86 select HAVE_KERNEL_LZO select HAVE_HW_BREAKPOINT select PERF_EVENTS + select PERF_EVENTS_NMI select ANON_INODES select HAVE_ARCH_KMEMCHECK select HAVE_USER_RETURN_NOTIFIER diff --git a/init/Kconfig b/init/Kconfig index ada48441aff..7331a16dd82 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -946,6 +946,11 @@ config PERF_USE_VMALLOC help See tools/perf/design.txt for details +config PERF_EVENTS_NMI + bool + help + Arch has support for nmi_watchdog + menu "Kernel Performance Events And Counters" config PERF_EVENTS diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index acef88239e1..01a4d85ee74 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -172,8 +172,7 @@ config DETECT_SOFTLOCKUP config NMI_WATCHDOG bool "Detect Hard Lockups with an NMI Watchdog" - depends on DEBUG_KERNEL && PERF_EVENTS - depends on X86 + depends on DEBUG_KERNEL && PERF_EVENTS && PERF_EVENTS_NMI default y help Say Y here to enable the kernel to use the NMI as a watchdog -- cgit v1.2.3-70-g09d2 From c99c30feadf664ccd8590b96259cb9f3954bd246 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Sun, 28 Feb 2010 20:49:00 +0100 Subject: nmi_watchdog: Turn it off by default It was nice to enable it by default for testing - but before we push it upstream we want it to be off - so that people can opt-in gradually. Cc: Don Zickus Cc: peterz@infradead.org Cc: gorcunov@gmail.com Cc: aris@redhat.com LKML-Reference: <1266880143-24943-1-git-send-email-dzickus@redhat.com> Signed-off-by: Ingo Molnar --- lib/Kconfig.debug | 1 - 1 file changed, 1 deletion(-) (limited to 'lib') diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index 01a4d85ee74..e2e73cc17a9 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -173,7 +173,6 @@ config DETECT_SOFTLOCKUP config NMI_WATCHDOG bool "Detect Hard Lockups with an NMI Watchdog" depends on DEBUG_KERNEL && PERF_EVENTS && PERF_EVENTS_NMI - default y help Say Y here to enable the kernel to use the NMI as a watchdog to detect hard lockups. This is useful when a cpu hangs for no -- cgit v1.2.3-70-g09d2 From 58687acba59266735adb8ccd9b5b9aa2c7cd205b Mon Sep 17 00:00:00 2001 From: Don Zickus Date: Fri, 7 May 2010 17:11:44 -0400 Subject: lockup_detector: Combine nmi_watchdog and softlockup detector The new nmi_watchdog (which uses the perf event subsystem) is very similar in structure to the softlockup detector. Using Ingo's suggestion, I combined the two functionalities into one file: kernel/watchdog.c. Now both the nmi_watchdog (or hardlockup detector) and softlockup detector sit on top of the perf event subsystem, which is run every 60 seconds or so to see if there are any lockups. To detect hardlockups, cpus not responding to interrupts, I implemented an hrtimer that runs 5 times for every perf event overflow event. If that stops counting on a cpu, then the cpu is most likely in trouble. To detect softlockups, tasks not yielding to the scheduler, I used the previous kthread idea that now gets kicked every time the hrtimer fires. If the kthread isn't being scheduled neither is anyone else and the warning is printed to the console. I tested this on x86_64 and both the softlockup and hardlockup paths work. V2: - cleaned up the Kconfig and softlockup combination - surrounded hardlockup cases with #ifdef CONFIG_PERF_EVENTS_NMI - seperated out the softlockup case from perf event subsystem - re-arranged the enabling/disabling nmi watchdog from proc space - added cpumasks for hardlockup failure cases - removed fallback to soft events if no PMU exists for hard events V3: - comment cleanups - drop support for older softlockup code - per_cpu cleanups - completely remove software clock base hardlockup detector - use per_cpu masking on hard/soft lockup detection - #ifdef cleanups - rename config option NMI_WATCHDOG to LOCKUP_DETECTOR - documentation additions V4: - documentation fixes - convert per_cpu to __get_cpu_var - powerpc compile fixes V5: - split apart warn flags for hard and soft lockups TODO: - figure out how to make an arch-agnostic clock2cycles call (if possible) to feed into perf events as a sample period [fweisbec: merged conflict patch] Signed-off-by: Don Zickus Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Cyrill Gorcunov Cc: Eric Paris Cc: Randy Dunlap LKML-Reference: <1273266711-18706-2-git-send-email-dzickus@redhat.com> Signed-off-by: Frederic Weisbecker --- Documentation/kernel-parameters.txt | 2 + arch/x86/include/asm/nmi.h | 2 +- arch/x86/kernel/apic/Makefile | 4 +- arch/x86/kernel/apic/hw_nmi.c | 2 +- arch/x86/kernel/traps.c | 4 +- include/linux/nmi.h | 8 +- include/linux/sched.h | 6 + init/Kconfig | 5 +- kernel/Makefile | 3 +- kernel/sysctl.c | 21 +- kernel/watchdog.c | 592 ++++++++++++++++++++++++++++++++++++ lib/Kconfig.debug | 30 +- 12 files changed, 650 insertions(+), 29 deletions(-) create mode 100644 kernel/watchdog.c (limited to 'lib') diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index 839b21b0699..dfe8d1c226c 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -1777,6 +1777,8 @@ and is between 256 and 4096 characters. It is defined in the file nousb [USB] Disable the USB subsystem + nowatchdog [KNL] Disable the lockup detector. + nowb [ARM] nox2apic [X86-64,APIC] Do not enable x2APIC mode. diff --git a/arch/x86/include/asm/nmi.h b/arch/x86/include/asm/nmi.h index 5b41b0feb6d..932f0f86b4b 100644 --- a/arch/x86/include/asm/nmi.h +++ b/arch/x86/include/asm/nmi.h @@ -17,7 +17,7 @@ int do_nmi_callback(struct pt_regs *regs, int cpu); extern void die_nmi(char *str, struct pt_regs *regs, int do_panic); extern int check_nmi_watchdog(void); -#if !defined(CONFIG_NMI_WATCHDOG) +#if !defined(CONFIG_LOCKUP_DETECTOR) extern int nmi_watchdog_enabled; #endif extern int avail_to_resrv_perfctr_nmi_bit(unsigned int); diff --git a/arch/x86/kernel/apic/Makefile b/arch/x86/kernel/apic/Makefile index 1a4512e48d2..52f32e0ea19 100644 --- a/arch/x86/kernel/apic/Makefile +++ b/arch/x86/kernel/apic/Makefile @@ -3,10 +3,10 @@ # obj-$(CONFIG_X86_LOCAL_APIC) += apic.o apic_noop.o probe_$(BITS).o ipi.o -ifneq ($(CONFIG_NMI_WATCHDOG),y) +ifneq ($(CONFIG_LOCKUP_DETECTOR),y) obj-$(CONFIG_X86_LOCAL_APIC) += nmi.o endif -obj-$(CONFIG_NMI_WATCHDOG) += hw_nmi.o +obj-$(CONFIG_LOCKUP_DETECTOR) += hw_nmi.o obj-$(CONFIG_X86_IO_APIC) += io_apic.o obj-$(CONFIG_SMP) += ipi.o diff --git a/arch/x86/kernel/apic/hw_nmi.c b/arch/x86/kernel/apic/hw_nmi.c index e8b78a0be5d..79425f96fce 100644 --- a/arch/x86/kernel/apic/hw_nmi.c +++ b/arch/x86/kernel/apic/hw_nmi.c @@ -89,7 +89,7 @@ int hw_nmi_is_cpu_stuck(struct pt_regs *regs) u64 hw_nmi_get_sample_period(void) { - return cpu_khz * 1000; + return (u64)(cpu_khz) * 1000 * 60; } #ifdef ARCH_HAS_NMI_WATCHDOG diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c index bdc7fab3ef3..bd347c2b34d 100644 --- a/arch/x86/kernel/traps.c +++ b/arch/x86/kernel/traps.c @@ -406,7 +406,7 @@ static notrace __kprobes void default_do_nmi(struct pt_regs *regs) == NOTIFY_STOP) return; -#ifndef CONFIG_NMI_WATCHDOG +#ifndef CONFIG_LOCKUP_DETECTOR /* * Ok, so this is none of the documented NMI sources, * so it must be the NMI watchdog. @@ -414,7 +414,7 @@ static notrace __kprobes void default_do_nmi(struct pt_regs *regs) if (nmi_watchdog_tick(regs, reason)) return; if (!do_nmi_callback(regs, cpu)) -#endif /* !CONFIG_NMI_WATCHDOG */ +#endif /* !CONFIG_LOCKUP_DETECTOR */ unknown_nmi_error(reason, regs); #else unknown_nmi_error(reason, regs); diff --git a/include/linux/nmi.h b/include/linux/nmi.h index 22cc7960b64..abd48aacaf7 100644 --- a/include/linux/nmi.h +++ b/include/linux/nmi.h @@ -20,7 +20,7 @@ extern void touch_nmi_watchdog(void); extern void acpi_nmi_disable(void); extern void acpi_nmi_enable(void); #else -#ifndef CONFIG_NMI_WATCHDOG +#ifndef CONFIG_LOCKUP_DETECTOR static inline void touch_nmi_watchdog(void) { touch_softlockup_watchdog(); @@ -51,12 +51,12 @@ static inline bool trigger_all_cpu_backtrace(void) } #endif -#ifdef CONFIG_NMI_WATCHDOG +#ifdef CONFIG_LOCKUP_DETECTOR int hw_nmi_is_cpu_stuck(struct pt_regs *); u64 hw_nmi_get_sample_period(void); -extern int nmi_watchdog_enabled; +extern int watchdog_enabled; struct ctl_table; -extern int proc_nmi_enabled(struct ctl_table *, int , +extern int proc_dowatchdog_enabled(struct ctl_table *, int , void __user *, size_t *, loff_t *); #endif diff --git a/include/linux/sched.h b/include/linux/sched.h index dad7f668ebf..37efe8fa530 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -346,6 +346,12 @@ extern int proc_dohung_task_timeout_secs(struct ctl_table *table, int write, size_t *lenp, loff_t *ppos); #endif +#ifdef CONFIG_LOCKUP_DETECTOR +extern int proc_dowatchdog_thresh(struct ctl_table *table, int write, + void __user *buffer, + size_t *lenp, loff_t *ppos); +#endif + /* Attach to any functions which should be ignored in wchan output. */ #define __sched __attribute__((__section__(".sched.text"))) diff --git a/init/Kconfig b/init/Kconfig index c6c8903cb53..e44e25422f2 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -944,8 +944,11 @@ config PERF_USE_VMALLOC config PERF_EVENTS_NMI bool + depends on PERF_EVENTS help - Arch has support for nmi_watchdog + System hardware can generate an NMI using the perf event + subsystem. Also has support for calculating CPU cycle events + to determine how many clock cycles in a given period. menu "Kernel Performance Events And Counters" diff --git a/kernel/Makefile b/kernel/Makefile index d5c30060ac1..6adeafc3e25 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -76,9 +76,8 @@ obj-$(CONFIG_GCOV_KERNEL) += gcov/ obj-$(CONFIG_AUDIT_TREE) += audit_tree.o obj-$(CONFIG_KPROBES) += kprobes.o obj-$(CONFIG_KGDB) += kgdb.o -obj-$(CONFIG_DETECT_SOFTLOCKUP) += softlockup.o -obj-$(CONFIG_NMI_WATCHDOG) += nmi_watchdog.o obj-$(CONFIG_DETECT_HUNG_TASK) += hung_task.o +obj-$(CONFIG_LOCKUP_DETECTOR) += watchdog.o obj-$(CONFIG_GENERIC_HARDIRQS) += irq/ obj-$(CONFIG_SECCOMP) += seccomp.o obj-$(CONFIG_RCU_TORTURE_TEST) += rcutorture.o diff --git a/kernel/sysctl.c b/kernel/sysctl.c index a38af430f0d..0f9adda85f9 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -74,7 +74,7 @@ #include #endif -#ifdef CONFIG_NMI_WATCHDOG +#ifdef CONFIG_LOCKUP_DETECTOR #include #endif @@ -686,16 +686,25 @@ static struct ctl_table kern_table[] = { .mode = 0444, .proc_handler = proc_dointvec, }, -#if defined(CONFIG_NMI_WATCHDOG) +#if defined(CONFIG_LOCKUP_DETECTOR) { - .procname = "nmi_watchdog", - .data = &nmi_watchdog_enabled, + .procname = "watchdog", + .data = &watchdog_enabled, .maxlen = sizeof (int), .mode = 0644, - .proc_handler = proc_nmi_enabled, + .proc_handler = proc_dowatchdog_enabled, + }, + { + .procname = "watchdog_thresh", + .data = &softlockup_thresh, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = proc_dowatchdog_thresh, + .extra1 = &neg_one, + .extra2 = &sixty, }, #endif -#if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86) && !defined(CONFIG_NMI_WATCHDOG) +#if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86) && !defined(CONFIG_LOCKUP_DETECTOR) { .procname = "unknown_nmi_panic", .data = &unknown_nmi_panic, diff --git a/kernel/watchdog.c b/kernel/watchdog.c new file mode 100644 index 00000000000..6b7fad8497a --- /dev/null +++ b/kernel/watchdog.c @@ -0,0 +1,592 @@ +/* + * Detect hard and soft lockups on a system + * + * started by Don Zickus, Copyright (C) 2010 Red Hat, Inc. + * + * this code detects hard lockups: incidents in where on a CPU + * the kernel does not respond to anything except NMI. + * + * Note: Most of this code is borrowed heavily from softlockup.c, + * so thanks to Ingo for the initial implementation. + * Some chunks also taken from arch/x86/kernel/apic/nmi.c, thanks + * to those contributors as well. + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include + +int watchdog_enabled; +int __read_mostly softlockup_thresh = 60; + +static DEFINE_PER_CPU(unsigned long, watchdog_touch_ts); +static DEFINE_PER_CPU(struct task_struct *, softlockup_watchdog); +static DEFINE_PER_CPU(struct hrtimer, watchdog_hrtimer); +static DEFINE_PER_CPU(bool, softlockup_touch_sync); +static DEFINE_PER_CPU(bool, hard_watchdog_warn); +static DEFINE_PER_CPU(bool, soft_watchdog_warn); +#ifdef CONFIG_PERF_EVENTS_NMI +static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts); +static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts_saved); +static DEFINE_PER_CPU(struct perf_event *, watchdog_ev); +#endif + +static int __read_mostly did_panic; +static int __initdata no_watchdog; + + +/* boot commands */ +/* + * Should we panic when a soft-lockup or hard-lockup occurs: + */ +#ifdef CONFIG_PERF_EVENTS_NMI +static int hardlockup_panic; + +static int __init hardlockup_panic_setup(char *str) +{ + if (!strncmp(str, "panic", 5)) + hardlockup_panic = 1; + return 1; +} +__setup("nmi_watchdog=", hardlockup_panic_setup); +#endif + +unsigned int __read_mostly softlockup_panic = + CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC_VALUE; + +static int __init softlockup_panic_setup(char *str) +{ + softlockup_panic = simple_strtoul(str, NULL, 0); + + return 1; +} +__setup("softlockup_panic=", softlockup_panic_setup); + +static int __init nowatchdog_setup(char *str) +{ + no_watchdog = 1; + return 1; +} +__setup("nowatchdog", nowatchdog_setup); + +/* deprecated */ +static int __init nosoftlockup_setup(char *str) +{ + no_watchdog = 1; + return 1; +} +__setup("nosoftlockup", nosoftlockup_setup); +/* */ + + +/* + * Returns seconds, approximately. We don't need nanosecond + * resolution, and we don't need to waste time with a big divide when + * 2^30ns == 1.074s. + */ +static unsigned long get_timestamp(int this_cpu) +{ + return cpu_clock(this_cpu) >> 30LL; /* 2^30 ~= 10^9 */ +} + +static unsigned long get_sample_period(void) +{ + /* + * convert softlockup_thresh from seconds to ns + * the divide by 5 is to give hrtimer 5 chances to + * increment before the hardlockup detector generates + * a warning + */ + return softlockup_thresh / 5 * NSEC_PER_SEC; +} + +/* Commands for resetting the watchdog */ +static void __touch_watchdog(void) +{ + int this_cpu = raw_smp_processor_id(); + + __get_cpu_var(watchdog_touch_ts) = get_timestamp(this_cpu); +} + +void touch_watchdog(void) +{ + __get_cpu_var(watchdog_touch_ts) = 0; +} +EXPORT_SYMBOL(touch_watchdog); + +void touch_all_watchdog(void) +{ + int cpu; + + /* + * this is done lockless + * do we care if a 0 races with a timestamp? + * all it means is the softlock check starts one cycle later + */ + for_each_online_cpu(cpu) + per_cpu(watchdog_touch_ts, cpu) = 0; +} + +void touch_nmi_watchdog(void) +{ + touch_watchdog(); +} +EXPORT_SYMBOL(touch_nmi_watchdog); + +void touch_all_nmi_watchdog(void) +{ + touch_all_watchdog(); +} + +void touch_softlockup_watchdog(void) +{ + touch_watchdog(); +} + +void touch_all_softlockup_watchdogs(void) +{ + touch_all_watchdog(); +} + +void touch_softlockup_watchdog_sync(void) +{ + __raw_get_cpu_var(softlockup_touch_sync) = true; + __raw_get_cpu_var(watchdog_touch_ts) = 0; +} + +void softlockup_tick(void) +{ +} + +#ifdef CONFIG_PERF_EVENTS_NMI +/* watchdog detector functions */ +static int is_hardlockup(int cpu) +{ + unsigned long hrint = per_cpu(hrtimer_interrupts, cpu); + + if (per_cpu(hrtimer_interrupts_saved, cpu) == hrint) + return 1; + + per_cpu(hrtimer_interrupts_saved, cpu) = hrint; + return 0; +} +#endif + +static int is_softlockup(unsigned long touch_ts, int cpu) +{ + unsigned long now = get_timestamp(cpu); + + /* Warn about unreasonable delays: */ + if (time_after(now, touch_ts + softlockup_thresh)) + return now - touch_ts; + + return 0; +} + +static int +watchdog_panic(struct notifier_block *this, unsigned long event, void *ptr) +{ + did_panic = 1; + + return NOTIFY_DONE; +} + +static struct notifier_block panic_block = { + .notifier_call = watchdog_panic, +}; + +#ifdef CONFIG_PERF_EVENTS_NMI +static struct perf_event_attr wd_hw_attr = { + .type = PERF_TYPE_HARDWARE, + .config = PERF_COUNT_HW_CPU_CYCLES, + .size = sizeof(struct perf_event_attr), + .pinned = 1, + .disabled = 1, +}; + +/* Callback function for perf event subsystem */ +void watchdog_overflow_callback(struct perf_event *event, int nmi, + struct perf_sample_data *data, + struct pt_regs *regs) +{ + int this_cpu = smp_processor_id(); + unsigned long touch_ts = per_cpu(watchdog_touch_ts, this_cpu); + + if (touch_ts == 0) { + __touch_watchdog(); + return; + } + + /* check for a hardlockup + * This is done by making sure our timer interrupt + * is incrementing. The timer interrupt should have + * fired multiple times before we overflow'd. If it hasn't + * then this is a good indication the cpu is stuck + */ + if (is_hardlockup(this_cpu)) { + /* only print hardlockups once */ + if (__get_cpu_var(hard_watchdog_warn) == true) + return; + + if (hardlockup_panic) + panic("Watchdog detected hard LOCKUP on cpu %d", this_cpu); + else + WARN(1, "Watchdog detected hard LOCKUP on cpu %d", this_cpu); + + __get_cpu_var(hard_watchdog_warn) = true; + return; + } + + __get_cpu_var(hard_watchdog_warn) = false; + return; +} +static void watchdog_interrupt_count(void) +{ + __get_cpu_var(hrtimer_interrupts)++; +} +#else +static inline void watchdog_interrupt_count(void) { return; } +#endif /* CONFIG_PERF_EVENTS_NMI */ + +/* watchdog kicker functions */ +static enum hrtimer_restart watchdog_timer_fn(struct hrtimer *hrtimer) +{ + int this_cpu = smp_processor_id(); + unsigned long touch_ts = __get_cpu_var(watchdog_touch_ts); + struct pt_regs *regs = get_irq_regs(); + int duration; + + /* kick the hardlockup detector */ + watchdog_interrupt_count(); + + /* kick the softlockup detector */ + wake_up_process(__get_cpu_var(softlockup_watchdog)); + + /* .. and repeat */ + hrtimer_forward_now(hrtimer, ns_to_ktime(get_sample_period())); + + if (touch_ts == 0) { + if (unlikely(per_cpu(softlockup_touch_sync, this_cpu))) { + /* + * If the time stamp was touched atomically + * make sure the scheduler tick is up to date. + */ + per_cpu(softlockup_touch_sync, this_cpu) = false; + sched_clock_tick(); + } + __touch_watchdog(); + return HRTIMER_RESTART; + } + + /* check for a softlockup + * This is done by making sure a high priority task is + * being scheduled. The task touches the watchdog to + * indicate it is getting cpu time. If it hasn't then + * this is a good indication some task is hogging the cpu + */ + duration = is_softlockup(touch_ts, this_cpu); + if (unlikely(duration)) { + /* only warn once */ + if (__get_cpu_var(soft_watchdog_warn) == true) + return HRTIMER_RESTART; + + printk(KERN_ERR "BUG: soft lockup - CPU#%d stuck for %us! [%s:%d]\n", + this_cpu, duration, + current->comm, task_pid_nr(current)); + print_modules(); + print_irqtrace_events(current); + if (regs) + show_regs(regs); + else + dump_stack(); + + if (softlockup_panic) + panic("softlockup: hung tasks"); + __get_cpu_var(soft_watchdog_warn) = true; + } else + __get_cpu_var(soft_watchdog_warn) = false; + + return HRTIMER_RESTART; +} + + +/* + * The watchdog thread - touches the timestamp. + */ +static int watchdog(void *__bind_cpu) +{ + struct sched_param param = { .sched_priority = MAX_RT_PRIO-1 }; + struct hrtimer *hrtimer = &per_cpu(watchdog_hrtimer, (unsigned long)__bind_cpu); + + sched_setscheduler(current, SCHED_FIFO, ¶m); + + /* initialize timestamp */ + __touch_watchdog(); + + /* kick off the timer for the hardlockup detector */ + /* done here because hrtimer_start can only pin to smp_processor_id() */ + hrtimer_start(hrtimer, ns_to_ktime(get_sample_period()), + HRTIMER_MODE_REL_PINNED); + + set_current_state(TASK_INTERRUPTIBLE); + /* + * Run briefly once per second to reset the softlockup timestamp. + * If this gets delayed for more than 60 seconds then the + * debug-printout triggers in softlockup_tick(). + */ + while (!kthread_should_stop()) { + __touch_watchdog(); + schedule(); + + if (kthread_should_stop()) + break; + + set_current_state(TASK_INTERRUPTIBLE); + } + __set_current_state(TASK_RUNNING); + + return 0; +} + + +#ifdef CONFIG_PERF_EVENTS_NMI +static int watchdog_nmi_enable(int cpu) +{ + struct perf_event_attr *wd_attr; + struct perf_event *event = per_cpu(watchdog_ev, cpu); + + /* is it already setup and enabled? */ + if (event && event->state > PERF_EVENT_STATE_OFF) + goto out; + + /* it is setup but not enabled */ + if (event != NULL) + goto out_enable; + + /* Try to register using hardware perf events */ + wd_attr = &wd_hw_attr; + wd_attr->sample_period = hw_nmi_get_sample_period(); + event = perf_event_create_kernel_counter(wd_attr, cpu, -1, watchdog_overflow_callback); + if (!IS_ERR(event)) { + printk(KERN_INFO "NMI watchdog enabled, takes one hw-pmu counter.\n"); + goto out_save; + } + + printk(KERN_ERR "NMI watchdog failed to create perf event on cpu%i: %p\n", cpu, event); + return -1; + + /* success path */ +out_save: + per_cpu(watchdog_ev, cpu) = event; +out_enable: + perf_event_enable(per_cpu(watchdog_ev, cpu)); +out: + return 0; +} + +static void watchdog_nmi_disable(int cpu) +{ + struct perf_event *event = per_cpu(watchdog_ev, cpu); + + if (event) { + perf_event_disable(event); + per_cpu(watchdog_ev, cpu) = NULL; + + /* should be in cleanup, but blocks oprofile */ + perf_event_release_kernel(event); + } + return; +} +#else +static int watchdog_nmi_enable(int cpu) { return 0; } +static void watchdog_nmi_disable(int cpu) { return; } +#endif /* CONFIG_PERF_EVENTS_NMI */ + +/* prepare/enable/disable routines */ +static int watchdog_prepare_cpu(int cpu) +{ + struct hrtimer *hrtimer = &per_cpu(watchdog_hrtimer, cpu); + + WARN_ON(per_cpu(softlockup_watchdog, cpu)); + hrtimer_init(hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_REL); + hrtimer->function = watchdog_timer_fn; + + return 0; +} + +static int watchdog_enable(int cpu) +{ + struct task_struct *p = per_cpu(softlockup_watchdog, cpu); + + /* enable the perf event */ + if (watchdog_nmi_enable(cpu) != 0) + return -1; + + /* create the watchdog thread */ + if (!p) { + p = kthread_create(watchdog, (void *)(unsigned long)cpu, "watchdog/%d", cpu); + if (IS_ERR(p)) { + printk(KERN_ERR "softlockup watchdog for %i failed\n", cpu); + return -1; + } + kthread_bind(p, cpu); + per_cpu(watchdog_touch_ts, cpu) = 0; + per_cpu(softlockup_watchdog, cpu) = p; + wake_up_process(p); + } + + return 0; +} + +static void watchdog_disable(int cpu) +{ + struct task_struct *p = per_cpu(softlockup_watchdog, cpu); + struct hrtimer *hrtimer = &per_cpu(watchdog_hrtimer, cpu); + + /* + * cancel the timer first to stop incrementing the stats + * and waking up the kthread + */ + hrtimer_cancel(hrtimer); + + /* disable the perf event */ + watchdog_nmi_disable(cpu); + + /* stop the watchdog thread */ + if (p) { + per_cpu(softlockup_watchdog, cpu) = NULL; + kthread_stop(p); + } + + /* if any cpu succeeds, watchdog is considered enabled for the system */ + watchdog_enabled = 1; +} + +static void watchdog_enable_all_cpus(void) +{ + int cpu; + int result; + + for_each_online_cpu(cpu) + result += watchdog_enable(cpu); + + if (result) + printk(KERN_ERR "watchdog: failed to be enabled on some cpus\n"); + +} + +static void watchdog_disable_all_cpus(void) +{ + int cpu; + + for_each_online_cpu(cpu) + watchdog_disable(cpu); + + /* if all watchdogs are disabled, then they are disabled for the system */ + watchdog_enabled = 0; +} + + +/* sysctl functions */ +#ifdef CONFIG_SYSCTL +/* + * proc handler for /proc/sys/kernel/nmi_watchdog + */ + +int proc_dowatchdog_enabled(struct ctl_table *table, int write, + void __user *buffer, size_t *length, loff_t *ppos) +{ + proc_dointvec(table, write, buffer, length, ppos); + + if (watchdog_enabled) + watchdog_enable_all_cpus(); + else + watchdog_disable_all_cpus(); + return 0; +} + +int proc_dowatchdog_thresh(struct ctl_table *table, int write, + void __user *buffer, + size_t *lenp, loff_t *ppos) +{ + return proc_dointvec_minmax(table, write, buffer, lenp, ppos); +} + +/* stub functions */ +int proc_dosoftlockup_thresh(struct ctl_table *table, int write, + void __user *buffer, + size_t *lenp, loff_t *ppos) +{ + return proc_dowatchdog_thresh(table, write, buffer, lenp, ppos); +} +/* end of stub functions */ +#endif /* CONFIG_SYSCTL */ + + +/* + * Create/destroy watchdog threads as CPUs come and go: + */ +static int __cpuinit +cpu_callback(struct notifier_block *nfb, unsigned long action, void *hcpu) +{ + int hotcpu = (unsigned long)hcpu; + + switch (action) { + case CPU_UP_PREPARE: + case CPU_UP_PREPARE_FROZEN: + if (watchdog_prepare_cpu(hotcpu)) + return NOTIFY_BAD; + break; + case CPU_ONLINE: + case CPU_ONLINE_FROZEN: + if (watchdog_enable(hotcpu)) + return NOTIFY_BAD; + break; +#ifdef CONFIG_HOTPLUG_CPU + case CPU_UP_CANCELED: + case CPU_UP_CANCELED_FROZEN: + watchdog_disable(hotcpu); + break; + case CPU_DEAD: + case CPU_DEAD_FROZEN: + watchdog_disable(hotcpu); + break; +#endif /* CONFIG_HOTPLUG_CPU */ + } + return NOTIFY_OK; +} + +static struct notifier_block __cpuinitdata cpu_nfb = { + .notifier_call = cpu_callback +}; + +static int __init spawn_watchdog_task(void) +{ + void *cpu = (void *)(long)smp_processor_id(); + int err; + + if (no_watchdog) + return 0; + + err = cpu_callback(&cpu_nfb, CPU_UP_PREPARE, cpu); + WARN_ON(err == NOTIFY_BAD); + + 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_watchdog_task); diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index 220ae6063b6..49e285dcaf5 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -153,7 +153,7 @@ config DEBUG_SHIRQ points; some don't and need to be caught. config DETECT_SOFTLOCKUP - bool "Detect Soft Lockups" + bool depends on DEBUG_KERNEL && !S390 default y help @@ -171,17 +171,27 @@ config DETECT_SOFTLOCKUP can be detected via the NMI-watchdog, on platforms that support it.) -config NMI_WATCHDOG - bool "Detect Hard Lockups with an NMI Watchdog" - depends on DEBUG_KERNEL && PERF_EVENTS && PERF_EVENTS_NMI +config LOCKUP_DETECTOR + bool "Detect Hard and Soft Lockups" + depends on DEBUG_KERNEL + default DETECT_SOFTLOCKUP help - Say Y here to enable the kernel to use the NMI as a watchdog - to detect hard lockups. This is useful when a cpu hangs for no - reason but can still respond to NMIs. A backtrace is displayed - for reviewing and reporting. + Say Y here to enable the kernel to act as a watchdog to detect + hard and soft lockups. + + Softlockups are bugs that cause the kernel to loop in kernel + mode for more than 60 seconds, without giving other tasks a + chance to run. The current stack trace is displayed upon + detection and the system will stay locked up. + + Hardlockups are bugs that cause the CPU to loop in kernel mode + for more than 60 seconds, without letting other interrupts have a + chance to run. The current stack trace is displayed upon detection + and the system will stay locked up. - The overhead should be minimal, just an extra NMI every few - seconds. + The overhead should be minimal. A periodic hrtimer runs to + generate interrupts and kick the watchdog task every 10-12 seconds. + An NMI is generated every 60 seconds or so to check for hardlockups. config BOOTPARAM_SOFTLOCKUP_PANIC bool "Panic (Reboot) On Soft Lockups" -- cgit v1.2.3-70-g09d2 From 89d7ce2a2178e7f562f608b466a18c8c2ece87af Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 13 May 2010 00:27:20 +0200 Subject: lockup_detector: Make BOOTPARAM_SOFTLOCKUP_PANIC depend on LOCKUP_DETECTOR Panic on softlockups was still depending on the softlockup detector. But the latter has been merged into the lockup detector now. Let's update this config dependency. Signed-off-by: Frederic Weisbecker Cc: Don Zickus Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Cyrill Gorcunov Cc: Eric Paris Cc: Randy Dunlap --- lib/Kconfig.debug | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'lib') diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index 49e285dcaf5..755713a359e 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -195,7 +195,7 @@ config LOCKUP_DETECTOR config BOOTPARAM_SOFTLOCKUP_PANIC bool "Panic (Reboot) On Soft Lockups" - depends on DETECT_SOFTLOCKUP + depends on LOCKUP_DETECTOR help Say Y here to enable the kernel to panic on "soft lockups", which are bugs that cause the kernel to loop in kernel -- cgit v1.2.3-70-g09d2 From e16bb1d7fe07609bc8b0e4c043eff2f47ada78d8 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 15 May 2010 22:30:22 +0200 Subject: lockup_detector: Update some config We kept CONFIG_DETECT_SOFTLOCKUP around for compatibility with older configs. But it was enabled by default if CONFIG_DEBUG_KERNEL. So if we want to enable CONFIG_LOCKUP_DETECTOR on configs that had CONFIG_DETECT_SOFTLOCKUP, all we need is to have the same enabling by default if CONFIG_DEBUG_KERNEL. We can then remove CONFIG_DETECT_SOFTLOCKUP directly. So tag CONFIG_LOCKUP_DETECTOR as default y. This is what we want for most serious kernel debugging anyway. And also forbid the lockup detector in S390 as it was for the previous softlockup detector, event though the true reason for that is not outlined. Signed-off-by: Frederic Weisbecker Cc: Ingo Molnar Cc: Don Zickus Cc: Peter Zijlstra Cc: Cyrill Gorcunov --- lib/Kconfig.debug | 25 +++---------------------- 1 file changed, 3 insertions(+), 22 deletions(-) (limited to 'lib') diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index 755713a359e..3a18b0b856c 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -152,29 +152,10 @@ config DEBUG_SHIRQ Drivers ought to be able to handle interrupts coming in at those points; some don't and need to be caught. -config DETECT_SOFTLOCKUP - bool - depends on DEBUG_KERNEL && !S390 - default y - help - Say Y here to enable the kernel to detect "soft lockups", - which are bugs that cause the kernel to loop in kernel - mode for more than 60 seconds, without giving other tasks a - chance to run. - - When a soft-lockup is detected, the kernel will print the - current stack trace (which you should report), but the - system will stay locked up. This feature has negligible - overhead. - - (Note that "hard lockups" are separate type of bugs that - can be detected via the NMI-watchdog, on platforms that - support it.) - config LOCKUP_DETECTOR bool "Detect Hard and Soft Lockups" - depends on DEBUG_KERNEL - default DETECT_SOFTLOCKUP + depends on DEBUG_KERNEL && !S390 + default y help Say Y here to enable the kernel to act as a watchdog to detect hard and soft lockups. @@ -212,7 +193,7 @@ config BOOTPARAM_SOFTLOCKUP_PANIC config BOOTPARAM_SOFTLOCKUP_PANIC_VALUE int - depends on DETECT_SOFTLOCKUP + depends on LOCKUP_DETECTOR range 0 1 default 0 if !BOOTPARAM_SOFTLOCKUP_PANIC default 1 if BOOTPARAM_SOFTLOCKUP_PANIC -- cgit v1.2.3-70-g09d2 From 23637d477c1f53acbb176a02c241d60a25888fae Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 15 May 2010 23:15:20 +0200 Subject: lockup_detector: Introduce CONFIG_HARDLOCKUP_DETECTOR This new config is deemed to simplify even more the lockup detector dependencies and can make it easier to bring a smooth sorting between archs that support the new generic lockup detector and those that still have their own, especially for those that are in the middle of this migration. Instead of checking whether we have CONFIG_LOCKUP_DETECTOR + CONFIG_PERF_EVENTS_NMI each time an arch wants to know if it needs to build its own lockup detector, take a shortcut with this new config. It is enabled only if the hardlockup detection part of the whole lockup detector is on. Signed-off-by: Frederic Weisbecker Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Don Zickus Cc: Cyrill Gorcunov --- arch/Kconfig | 4 ++++ init/Kconfig | 7 ------- kernel/watchdog.c | 14 +++++++------- lib/Kconfig.debug | 3 +++ 4 files changed, 14 insertions(+), 14 deletions(-) (limited to 'lib') diff --git a/arch/Kconfig b/arch/Kconfig index 89b0efb5094..35084f28008 100644 --- a/arch/Kconfig +++ b/arch/Kconfig @@ -147,5 +147,9 @@ config HAVE_USER_RETURN_NOTIFIER config HAVE_PERF_EVENTS_NMI bool + help + System hardware can generate an NMI using the perf event + subsystem. Also has support for calculating CPU cycle events + to determine how many clock cycles in a given period. source "kernel/gcov/Kconfig" diff --git a/init/Kconfig b/init/Kconfig index ab733c32292..eb77e8ccde1 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -942,13 +942,6 @@ config PERF_USE_VMALLOC help See tools/perf/design.txt for details -config PERF_EVENTS_NMI - def_bool PERF_EVENTS && HAVE_PERF_EVENTS_NMI - help - System hardware can generate an NMI using the perf event - subsystem. Also has support for calculating CPU cycle events - to determine how many clock cycles in a given period. - menu "Kernel Performance Events And Counters" config PERF_EVENTS diff --git a/kernel/watchdog.c b/kernel/watchdog.c index be5e74e62be..83fb63155cb 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -37,7 +37,7 @@ static DEFINE_PER_CPU(struct hrtimer, watchdog_hrtimer); static DEFINE_PER_CPU(bool, softlockup_touch_sync); static DEFINE_PER_CPU(bool, hard_watchdog_warn); static DEFINE_PER_CPU(bool, soft_watchdog_warn); -#ifdef CONFIG_PERF_EVENTS_NMI +#ifdef CONFIG_HARDLOCKUP_DETECTOR static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts); static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts_saved); static DEFINE_PER_CPU(struct perf_event *, watchdog_ev); @@ -51,7 +51,7 @@ static int __initdata no_watchdog; /* * Should we panic when a soft-lockup or hard-lockup occurs: */ -#ifdef CONFIG_PERF_EVENTS_NMI +#ifdef CONFIG_HARDLOCKUP_DETECTOR static int hardlockup_panic; static int __init hardlockup_panic_setup(char *str) @@ -152,7 +152,7 @@ void touch_softlockup_watchdog_sync(void) __raw_get_cpu_var(watchdog_touch_ts) = 0; } -#ifdef CONFIG_PERF_EVENTS_NMI +#ifdef CONFIG_HARDLOCKUP_DETECTOR /* watchdog detector functions */ static int is_hardlockup(int cpu) { @@ -189,7 +189,7 @@ static struct notifier_block panic_block = { .notifier_call = watchdog_panic, }; -#ifdef CONFIG_PERF_EVENTS_NMI +#ifdef CONFIG_HARDLOCKUP_DETECTOR static struct perf_event_attr wd_hw_attr = { .type = PERF_TYPE_HARDWARE, .config = PERF_COUNT_HW_CPU_CYCLES, @@ -239,7 +239,7 @@ static void watchdog_interrupt_count(void) } #else static inline void watchdog_interrupt_count(void) { return; } -#endif /* CONFIG_PERF_EVENTS_NMI */ +#endif /* CONFIG_HARDLOCKUP_DETECTOR */ /* watchdog kicker functions */ static enum hrtimer_restart watchdog_timer_fn(struct hrtimer *hrtimer) @@ -342,7 +342,7 @@ static int watchdog(void *__bind_cpu) } -#ifdef CONFIG_PERF_EVENTS_NMI +#ifdef CONFIG_HARDLOCKUP_DETECTOR static int watchdog_nmi_enable(int cpu) { struct perf_event_attr *wd_attr; @@ -393,7 +393,7 @@ static void watchdog_nmi_disable(int cpu) #else static int watchdog_nmi_enable(int cpu) { return 0; } static void watchdog_nmi_disable(int cpu) { return; } -#endif /* CONFIG_PERF_EVENTS_NMI */ +#endif /* CONFIG_HARDLOCKUP_DETECTOR */ /* prepare/enable/disable routines */ static int watchdog_prepare_cpu(int cpu) diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index 3a18b0b856c..e65e47d5c5e 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -174,6 +174,9 @@ config LOCKUP_DETECTOR generate interrupts and kick the watchdog task every 10-12 seconds. An NMI is generated every 60 seconds or so to check for hardlockups. +config HARDLOCKUP_DETECTOR + def_bool LOCKUP_DETECTOR && PERF_EVENTS && HAVE_PERF_EVENTS_NMI + config BOOTPARAM_SOFTLOCKUP_PANIC bool "Panic (Reboot) On Soft Lockups" depends on LOCKUP_DETECTOR -- cgit v1.2.3-70-g09d2 From e35e7fb0e9ea557f7504ced6fe4ccf69e44b7f07 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 19 May 2010 11:36:49 +0200 Subject: lockup_detector: Don't enable the lockup detector by default The lockup detector is a new feature that now involves the nmi watchdog. Drop the default y and let the user choose. Signed-off-by: Frederic Weisbecker Cc: Don Zickus Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Cyrill Gorcunov --- lib/Kconfig.debug | 1 - 1 file changed, 1 deletion(-) (limited to 'lib') diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index e65e47d5c5e..63968a96844 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -155,7 +155,6 @@ config DEBUG_SHIRQ config LOCKUP_DETECTOR bool "Detect Hard and Soft Lockups" depends on DEBUG_KERNEL && !S390 - default y help Say Y here to enable the kernel to act as a watchdog to detect hard and soft lockups. -- cgit v1.2.3-70-g09d2