From 4135038a582c20ffdadfcf6564852e0b72a20968 Mon Sep 17 00:00:00 2001 From: Marcin Slusarz Date: Fri, 28 Jan 2011 11:00:31 -0500 Subject: watchdog: Fix broken nowatchdog logic Passing nowatchdog to kernel disables 2 things: creation of watchdog threads AND initialization of percpu watchdog_hrtimer. As hrtimers are initialized only at boot it's not possible to enable watchdog later - for me all watchdog threads started to eat 100% of CPU time, but they could just crash. Additionally, even if these threads would start properly, watchdog_disable_all_cpus was guarded by no_watchdog check, so you couldn't disable watchdog. To fix this, remove no_watchdog variable and use already existing watchdog_enabled variable. Signed-off-by: Marcin Slusarz [ removed another no_watchdog instance ] Signed-off-by: Don Zickus Cc: Stephane Eranian Cc: Peter Zijlstra Cc: Frederic Weisbecker Cc: LKML-Reference: <1296230433-6261-1-git-send-email-dzickus@redhat.com> Signed-off-by: Ingo Molnar --- kernel/watchdog.c | 20 ++++++-------------- 1 file changed, 6 insertions(+), 14 deletions(-) (limited to 'kernel/watchdog.c') diff --git a/kernel/watchdog.c b/kernel/watchdog.c index d7ebdf4cea9..d9961ea1c3f 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -27,7 +27,7 @@ #include #include -int watchdog_enabled; +int watchdog_enabled = 1; int __read_mostly softlockup_thresh = 60; static DEFINE_PER_CPU(unsigned long, watchdog_touch_ts); @@ -43,9 +43,6 @@ static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts_saved); static DEFINE_PER_CPU(struct perf_event *, watchdog_ev); #endif -static int no_watchdog; - - /* boot commands */ /* * Should we panic when a soft-lockup or hard-lockup occurs: @@ -58,7 +55,7 @@ static int __init hardlockup_panic_setup(char *str) if (!strncmp(str, "panic", 5)) hardlockup_panic = 1; else if (!strncmp(str, "0", 1)) - no_watchdog = 1; + watchdog_enabled = 0; return 1; } __setup("nmi_watchdog=", hardlockup_panic_setup); @@ -77,7 +74,7 @@ __setup("softlockup_panic=", softlockup_panic_setup); static int __init nowatchdog_setup(char *str) { - no_watchdog = 1; + watchdog_enabled = 0; return 1; } __setup("nowatchdog", nowatchdog_setup); @@ -85,7 +82,7 @@ __setup("nowatchdog", nowatchdog_setup); /* deprecated */ static int __init nosoftlockup_setup(char *str) { - no_watchdog = 1; + watchdog_enabled = 0; return 1; } __setup("nosoftlockup", nosoftlockup_setup); @@ -476,9 +473,6 @@ static void watchdog_disable_all_cpus(void) { int cpu; - if (no_watchdog) - return; - for_each_online_cpu(cpu) watchdog_disable(cpu); @@ -530,7 +524,8 @@ cpu_callback(struct notifier_block *nfb, unsigned long action, void *hcpu) break; case CPU_ONLINE: case CPU_ONLINE_FROZEN: - err = watchdog_enable(hotcpu); + if (watchdog_enabled) + err = watchdog_enable(hotcpu); break; #ifdef CONFIG_HOTPLUG_CPU case CPU_UP_CANCELED: @@ -555,9 +550,6 @@ void __init lockup_detector_init(void) void *cpu = (void *)(long)smp_processor_id(); int err; - if (no_watchdog) - return; - err = cpu_callback(&cpu_nfb, CPU_UP_PREPARE, cpu); WARN_ON(notifier_to_errno(err)); -- cgit v1.2.3-70-g09d2 From 397357666de6b5b6adb5fa99f9758ec8cf30ac34 Mon Sep 17 00:00:00 2001 From: Marcin Slusarz Date: Fri, 28 Jan 2011 11:00:32 -0500 Subject: watchdog: Fix sysctl consistency If it was not possible to enable watchdog for any cpu, switch watchdog_enabled back to 0, because it's visible via kernel.watchdog sysctl. Signed-off-by: Marcin Slusarz Signed-off-by: Don Zickus Cc: Peter Zijlstra Cc: Frederic Weisbecker Cc: LKML-Reference: <1296230433-6261-2-git-send-email-dzickus@redhat.com> Signed-off-by: Ingo Molnar --- kernel/watchdog.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) (limited to 'kernel/watchdog.c') diff --git a/kernel/watchdog.c b/kernel/watchdog.c index d9961ea1c3f..c7e0049344b 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -429,9 +429,6 @@ static int watchdog_enable(int cpu) wake_up_process(p); } - /* if any cpu succeeds, watchdog is considered enabled for the system */ - watchdog_enabled = 1; - return 0; } @@ -459,12 +456,16 @@ static void watchdog_disable(int cpu) static void watchdog_enable_all_cpus(void) { int cpu; - int result = 0; + + watchdog_enabled = 0; for_each_online_cpu(cpu) - result += watchdog_enable(cpu); + if (!watchdog_enable(cpu)) + /* if any cpu succeeds, watchdog is considered + enabled for the system */ + watchdog_enabled = 1; - if (result) + if (!watchdog_enabled) printk(KERN_ERR "watchdog: failed to be enabled on some cpus\n"); } -- cgit v1.2.3-70-g09d2 From 9ffdc6c37df131f89d52001e0ef03091b158826f Mon Sep 17 00:00:00 2001 From: Marcin Slusarz Date: Fri, 28 Jan 2011 11:00:33 -0500 Subject: watchdog: Don't change watchdog state on read of sysctl Signed-off-by: Marcin Slusarz [ add {}'s to fix a warning ] Signed-off-by: Don Zickus Cc: Peter Zijlstra Cc: Frederic Weisbecker Cc: LKML-Reference: <1296230433-6261-3-git-send-email-dzickus@redhat.com> Signed-off-by: Ingo Molnar --- kernel/watchdog.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) (limited to 'kernel/watchdog.c') diff --git a/kernel/watchdog.c b/kernel/watchdog.c index c7e0049344b..f37f974aa81 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -493,10 +493,12 @@ int proc_dowatchdog_enabled(struct ctl_table *table, int write, { proc_dointvec(table, write, buffer, length, ppos); - if (watchdog_enabled) - watchdog_enable_all_cpus(); - else - watchdog_disable_all_cpus(); + if (write) { + if (watchdog_enabled) + watchdog_enable_all_cpus(); + else + watchdog_disable_all_cpus(); + } return 0; } -- cgit v1.2.3-70-g09d2 From 5651f7f47dbb1cf2b95a60582546db4ff508e2b4 Mon Sep 17 00:00:00 2001 From: Don Zickus Date: Wed, 9 Feb 2011 14:02:33 -0500 Subject: watchdog, nmi: Lower the severity of error messages During boot if the hardlockup detector fails to initialize, it complains very loudly. Some failures should be expected under certain situations, ie no lapics, or resource in-use. Tone those error messages down a bit. Keep the rest at a high level. Reported-by: Paul Bolle Tested-by: Paul Bolle Signed-off-by: Don Zickus Cc: Peter Zijlstra LKML-Reference: <1297278153-21111-1-git-send-email-dzickus@redhat.com> Signed-off-by: Ingo Molnar --- kernel/watchdog.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) (limited to 'kernel/watchdog.c') diff --git a/kernel/watchdog.c b/kernel/watchdog.c index f37f974aa81..18bb15776c5 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -363,8 +363,14 @@ static int watchdog_nmi_enable(int cpu) goto out_save; } - printk(KERN_ERR "NMI watchdog disabled for cpu%i: unable to create perf event: %ld\n", - cpu, PTR_ERR(event)); + + /* vary the KERN level based on the returned errno */ + if (PTR_ERR(event) == -EOPNOTSUPP) + printk(KERN_INFO "NMI watchdog disabled (cpu%i): not supported (no LAPIC?)\n", cpu); + else if (PTR_ERR(event) == -ENOENT) + printk(KERN_WARNING "NMI watchdog disabled (cpu%i): hardware events not enabled\n", cpu); + else + printk(KERN_ERR "NMI watchdog disabled (cpu%i): unable to create perf event: %ld\n", cpu, PTR_ERR(event)); return PTR_ERR(event); /* success path */ -- cgit v1.2.3-70-g09d2