[v2] hardlockup: detect hard lockups using secondary (buddy) CPUs

  From: Colin Cross <ccross@android.com>

Implement a hardlockup detector that doesn't doesn't need any extra
arch-specific support code to detect lockups. Instead of using
something arch-specific we will use the buddy system, where each CPU
watches out for another one. Specifically, each CPU will use its
softlockup hrtimer to check that the next CPU is processing hrtimer
interrupts by verifying that a counter is increasing.

NOTE: unlike the other hard lockup detectors, the buddy one can't
easily show what's happening on the CPU that locked up just by doing a
simple backtrace. It relies on some other mechanism in the system to
get information about the locked up CPUs. This could be support for
NMI backtraces like [1], it could be a mechanism for printing the PC
of locked CPUs at panic time like [2] / [3], or it could be something
else. Even though that means we still rely on arch-specific code, this
arch-specific code seems to often be implemented even on architectures
that don't have a hardlockup detector.

This style of hardlockup detector originated in some downstream
Android trees and has been rebased on / carried in ChromeOS trees for
quite a long time for use on arm and arm64 boards. Historically on
these boards we've leveraged mechanism [2] / [3] to get information
about hung CPUs, but we could move to [1].

Although the original motivation for the buddy system was for use on
systems without an arch-specific hardlockup detector, it can still be
useful to use even on systems that _do_ have an arch-specific
hardlockup detector. On x86, for instance, there is a 24-part patch
series [4] in progress switching the arch-specific hard lockup
detector from a scarce perf counter to a less-scarce hardware
resource. Potentially the buddy system could be a simpler alternative
to free up the perf counter but still get hard lockup detection.

Overall, pros (+) and cons (-) of the buddy system compared to an
arch-specific hardlockup detector:
+ Usable on systems that don't have an arch-specific hardlockup
  detector, like arm32 and arm64 (though it's being worked on for
  arm64 [5]).
+ May free up scarce hardware resources.
+ If a CPU totally goes out to lunch (can't process NMIs) the buddy
  system could still detect the problem (though it would be unlikely
  to be able to get a stack trace).
- If all CPUs are hard locked up at the same time the buddy system
  can't detect it.
- If we don't have SMP we can't use the buddy system.
- The buddy system needs an arch-specific mechanism (possibly NMI
  backtrace) to get info about the locked up CPU.

[1] https://lore.kernel.org/r/20230419225604.21204-1-dianders@chromium.org
[2] https://issuetracker.google.com/172213129
[3] https://docs.kernel.org/trace/coresight/coresight-cpu-debug.html
[4] https://lore.kernel.org/lkml/20230301234753.28582-1-ricardo.neri-calderon@linux.intel.com/
[5] https://lore.kernel.org/linux-arm-kernel/20220903093415.15850-1-lecopzer.chen@mediatek.com/

Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Matthias Kaehlcke <mka@chromium.org>
Signed-off-by: Guenter Roeck <groeck@chromium.org>
Signed-off-by: Tzung-Bi Shih <tzungbi@chromium.org>
Signed-off-by: Douglas Anderson <dianders@chromium.org>
---
This patch has been rebased in ChromeOS kernel trees many times, and
each time someone had to do work on it they added their
Signed-off-by. I've included those here. I've also left the author as
Colin Cross since the core code is still his.

I'll also note that the CC list is pretty giant, but that's what
get_maintainers came up with (plus a few other folks I thought would
be interested). As far as I can tell, there's no true MAINTAINER
listed for the existing watchdog code. Assuming people don't hate
this, maybe it would go through Andrew Morton's tree?

Changes in v2:
- cpu => CPU.
- Reworked description and Kconfig based on v1 discussion.
- No code changes

 include/linux/nmi.h         |  18 ++++-
 kernel/Makefile             |   1 +
 kernel/watchdog.c           |  24 ++++--
 kernel/watchdog_buddy_cpu.c | 141 ++++++++++++++++++++++++++++++++++++
 lib/Kconfig.debug           |  23 +++++-
 5 files changed, 196 insertions(+), 11 deletions(-)
 create mode 100644 kernel/watchdog_buddy_cpu.c
  

Hi--

On 4/28/23 16:37, Douglas Anderson wrote:
> From: Colin Cross <ccross@android.com>
> 
> Implement a hardlockup detector that doesn't doesn't need any extra
> arch-specific support code to detect lockups. Instead of using
> something arch-specific we will use the buddy system, where each CPU
> watches out for another one. Specifically, each CPU will use its
> softlockup hrtimer to check that the next CPU is processing hrtimer
> interrupts by verifying that a counter is increasing.
> 
> NOTE: unlike the other hard lockup detectors, the buddy one can't
> easily show what's happening on the CPU that locked up just by doing a
> simple backtrace. It relies on some other mechanism in the system to
> get information about the locked up CPUs. This could be support for
> NMI backtraces like [1], it could be a mechanism for printing the PC
> of locked CPUs at panic time like [2] / [3], or it could be something
> else. Even though that means we still rely on arch-specific code, this
> arch-specific code seems to often be implemented even on architectures
> that don't have a hardlockup detector.
> 
> This style of hardlockup detector originated in some downstream
> Android trees and has been rebased on / carried in ChromeOS trees for
> quite a long time for use on arm and arm64 boards. Historically on
> these boards we've leveraged mechanism [2] / [3] to get information
> about hung CPUs, but we could move to [1].
> 
> Although the original motivation for the buddy system was for use on
> systems without an arch-specific hardlockup detector, it can still be
> useful to use even on systems that _do_ have an arch-specific
> hardlockup detector. On x86, for instance, there is a 24-part patch
> series [4] in progress switching the arch-specific hard lockup
> detector from a scarce perf counter to a less-scarce hardware
> resource. Potentially the buddy system could be a simpler alternative
> to free up the perf counter but still get hard lockup detection.
> 
> Overall, pros (+) and cons (-) of the buddy system compared to an
> arch-specific hardlockup detector:
> + Usable on systems that don't have an arch-specific hardlockup
>   detector, like arm32 and arm64 (though it's being worked on for
>   arm64 [5]).
> + May free up scarce hardware resources.
> + If a CPU totally goes out to lunch (can't process NMIs) the buddy
>   system could still detect the problem (though it would be unlikely
>   to be able to get a stack trace).
> - If all CPUs are hard locked up at the same time the buddy system
>   can't detect it.
> - If we don't have SMP we can't use the buddy system.
> - The buddy system needs an arch-specific mechanism (possibly NMI
>   backtrace) to get info about the locked up CPU.
> 
> [1] https://lore.kernel.org/r/20230419225604.21204-1-dianders@chromium.org
> [2] https://issuetracker.google.com/172213129
> [3] https://docs.kernel.org/trace/coresight/coresight-cpu-debug.html
> [4] https://lore.kernel.org/lkml/20230301234753.28582-1-ricardo.neri-calderon@linux.intel.com/
> [5] https://lore.kernel.org/linux-arm-kernel/20220903093415.15850-1-lecopzer.chen@mediatek.com/
> 
> Signed-off-by: Colin Cross <ccross@android.com>
> Signed-off-by: Matthias Kaehlcke <mka@chromium.org>
> Signed-off-by: Guenter Roeck <groeck@chromium.org>
> Signed-off-by: Tzung-Bi Shih <tzungbi@chromium.org>
> Signed-off-by: Douglas Anderson <dianders@chromium.org>
> ---
> This patch has been rebased in ChromeOS kernel trees many times, and
> each time someone had to do work on it they added their
> Signed-off-by. I've included those here. I've also left the author as
> Colin Cross since the core code is still his.
> 
> I'll also note that the CC list is pretty giant, but that's what
> get_maintainers came up with (plus a few other folks I thought would
> be interested). As far as I can tell, there's no true MAINTAINER
> listed for the existing watchdog code. Assuming people don't hate
> this, maybe it would go through Andrew Morton's tree?
> 
> Changes in v2:
> - cpu => CPU.
> - Reworked description and Kconfig based on v1 discussion.

or at least some of the comments from v1. :(

> - No code changes
> 
>  include/linux/nmi.h         |  18 ++++-
>  kernel/Makefile             |   1 +
>  kernel/watchdog.c           |  24 ++++--
>  kernel/watchdog_buddy_cpu.c | 141 ++++++++++++++++++++++++++++++++++++
>  lib/Kconfig.debug           |  23 +++++-
>  5 files changed, 196 insertions(+), 11 deletions(-)
>  create mode 100644 kernel/watchdog_buddy_cpu.c
> 

> diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug
> index 39d1d93164bd..511eb14660b1 100644
> --- a/lib/Kconfig.debug
> +++ b/lib/Kconfig.debug

> @@ -1055,9 +1059,26 @@ config HARDLOCKUP_DETECTOR
>  	  chance to run.  The current stack trace is displayed upon detection
>  	  and the system will stay locked up.
>  
> +config HARDLOCKUP_DETECTOR_BUDDY_CPU
> +	bool "Buddy CPU hardlockup detector"
> +	depends on DEBUG_KERNEL && SMP
> +	depends on !HARDLOCKUP_DETECTOR && !HAVE_NMI_WATCHDOG
> +	depends on !S390
> +	select HARDLOCKUP_DETECTOR_CORE
> +	select SOFTLOCKUP_DETECTOR
> +	help
> +	  Say Y here to enable a hardlockup detector where CPUs check

Be consistent in the use of "CPU". Change 2 occurrences of "cpu"
below to "CPU".

> +	  each other for lockup. Each cpu uses its softlockup hrtimer
> +	  to check that the next cpu is processing hrtimer interrupts by
> +	  verifying that a counter is increasing.
> +
> +	  This hardlockup detector is useful on systems that don't have
> +	  an arch-specific hardlockup detector or if resources needed
> +	  for the hardlockup detector are better used for other things.
  

On Fri, Apr 28, 2023 at 4:41 PM Douglas Anderson <dianders@chromium.org> wrote:
>
> From: Colin Cross <ccross@android.com>
>
> Implement a hardlockup detector that doesn't doesn't need any extra
> arch-specific support code to detect lockups. Instead of using
> something arch-specific we will use the buddy system, where each CPU
> watches out for another one. Specifically, each CPU will use its
> softlockup hrtimer to check that the next CPU is processing hrtimer
> interrupts by verifying that a counter is increasing.
>
> NOTE: unlike the other hard lockup detectors, the buddy one can't
> easily show what's happening on the CPU that locked up just by doing a
> simple backtrace. It relies on some other mechanism in the system to
> get information about the locked up CPUs. This could be support for
> NMI backtraces like [1], it could be a mechanism for printing the PC
> of locked CPUs at panic time like [2] / [3], or it could be something
> else. Even though that means we still rely on arch-specific code, this
> arch-specific code seems to often be implemented even on architectures
> that don't have a hardlockup detector.
>
> This style of hardlockup detector originated in some downstream
> Android trees and has been rebased on / carried in ChromeOS trees for
> quite a long time for use on arm and arm64 boards. Historically on
> these boards we've leveraged mechanism [2] / [3] to get information
> about hung CPUs, but we could move to [1].
>
> Although the original motivation for the buddy system was for use on
> systems without an arch-specific hardlockup detector, it can still be
> useful to use even on systems that _do_ have an arch-specific
> hardlockup detector. On x86, for instance, there is a 24-part patch
> series [4] in progress switching the arch-specific hard lockup
> detector from a scarce perf counter to a less-scarce hardware
> resource. Potentially the buddy system could be a simpler alternative
> to free up the perf counter but still get hard lockup detection.
>
> Overall, pros (+) and cons (-) of the buddy system compared to an
> arch-specific hardlockup detector:
> + Usable on systems that don't have an arch-specific hardlockup
>   detector, like arm32 and arm64 (though it's being worked on for
>   arm64 [5]).
> + May free up scarce hardware resources.
> + If a CPU totally goes out to lunch (can't process NMIs) the buddy
>   system could still detect the problem (though it would be unlikely
>   to be able to get a stack trace).
> - If all CPUs are hard locked up at the same time the buddy system
>   can't detect it.
> - If we don't have SMP we can't use the buddy system.
> - The buddy system needs an arch-specific mechanism (possibly NMI
>   backtrace) to get info about the locked up CPU.

Thanks for this list, it is really useful! Is it worth mentioning the
behavior around idle? Could this approach potentially use more power?

Thanks,
Ian

> [1] https://lore.kernel.org/r/20230419225604.21204-1-dianders@chromium.org
> [2] https://issuetracker.google.com/172213129
> [3] https://docs.kernel.org/trace/coresight/coresight-cpu-debug.html
> [4] https://lore.kernel.org/lkml/20230301234753.28582-1-ricardo.neri-calderon@linux.intel.com/
> [5] https://lore.kernel.org/linux-arm-kernel/20220903093415.15850-1-lecopzer.chen@mediatek.com/
>
> Signed-off-by: Colin Cross <ccross@android.com>
> Signed-off-by: Matthias Kaehlcke <mka@chromium.org>
> Signed-off-by: Guenter Roeck <groeck@chromium.org>
> Signed-off-by: Tzung-Bi Shih <tzungbi@chromium.org>
> Signed-off-by: Douglas Anderson <dianders@chromium.org>
> ---
> This patch has been rebased in ChromeOS kernel trees many times, and
> each time someone had to do work on it they added their
> Signed-off-by. I've included those here. I've also left the author as
> Colin Cross since the core code is still his.
>
> I'll also note that the CC list is pretty giant, but that's what
> get_maintainers came up with (plus a few other folks I thought would
> be interested). As far as I can tell, there's no true MAINTAINER
> listed for the existing watchdog code. Assuming people don't hate
> this, maybe it would go through Andrew Morton's tree?
>
> Changes in v2:
> - cpu => CPU.
> - Reworked description and Kconfig based on v1 discussion.
> - No code changes
>
>  include/linux/nmi.h         |  18 ++++-
>  kernel/Makefile             |   1 +
>  kernel/watchdog.c           |  24 ++++--
>  kernel/watchdog_buddy_cpu.c | 141 ++++++++++++++++++++++++++++++++++++
>  lib/Kconfig.debug           |  23 +++++-
>  5 files changed, 196 insertions(+), 11 deletions(-)
>  create mode 100644 kernel/watchdog_buddy_cpu.c
>
> diff --git a/include/linux/nmi.h b/include/linux/nmi.h
> index 048c0b9aa623..35f6c5c2378b 100644
> --- a/include/linux/nmi.h
> +++ b/include/linux/nmi.h
> @@ -45,6 +45,8 @@ extern void touch_softlockup_watchdog(void);
>  extern void touch_softlockup_watchdog_sync(void);
>  extern void touch_all_softlockup_watchdogs(void);
>  extern unsigned int  softlockup_panic;
> +DECLARE_PER_CPU(unsigned long, hrtimer_interrupts);
> +DECLARE_PER_CPU(unsigned long, hrtimer_interrupts_saved);
>
>  extern int lockup_detector_online_cpu(unsigned int cpu);
>  extern int lockup_detector_offline_cpu(unsigned int cpu);
> @@ -81,14 +83,14 @@ static inline void reset_hung_task_detector(void) { }
>  #define NMI_WATCHDOG_ENABLED      (1 << NMI_WATCHDOG_ENABLED_BIT)
>  #define SOFT_WATCHDOG_ENABLED     (1 << SOFT_WATCHDOG_ENABLED_BIT)
>
> -#if defined(CONFIG_HARDLOCKUP_DETECTOR)
> +#if defined(CONFIG_HARDLOCKUP_DETECTOR_CORE)
>  extern void hardlockup_detector_disable(void);
>  extern unsigned int hardlockup_panic;
>  #else
>  static inline void hardlockup_detector_disable(void) {}
>  #endif
>
> -#if defined(CONFIG_HAVE_NMI_WATCHDOG) || defined(CONFIG_HARDLOCKUP_DETECTOR)
> +#if defined(CONFIG_HAVE_NMI_WATCHDOG) || defined(CONFIG_HARDLOCKUP_DETECTOR_CORE)
>  # define NMI_WATCHDOG_SYSCTL_PERM      0644
>  #else
>  # define NMI_WATCHDOG_SYSCTL_PERM      0444
> @@ -124,6 +126,14 @@ void watchdog_nmi_disable(unsigned int cpu);
>
>  void lockup_detector_reconfigure(void);
>
> +#ifdef CONFIG_HARDLOCKUP_DETECTOR_BUDDY_CPU
> +void buddy_cpu_touch_watchdog(void);
> +void watchdog_check_hardlockup(void);
> +#else
> +static inline void buddy_cpu_touch_watchdog(void) {}
> +static inline void watchdog_check_hardlockup(void) {}
> +#endif
> +
>  /**
>   * touch_nmi_watchdog - restart NMI watchdog timeout.
>   *
> @@ -134,6 +144,7 @@ void lockup_detector_reconfigure(void);
>  static inline void touch_nmi_watchdog(void)
>  {
>         arch_touch_nmi_watchdog();
> +       buddy_cpu_touch_watchdog();
>         touch_softlockup_watchdog();
>  }
>
> @@ -196,8 +207,7 @@ static inline bool trigger_single_cpu_backtrace(int cpu)
>  u64 hw_nmi_get_sample_period(int watchdog_thresh);
>  #endif
>
> -#if defined(CONFIG_HARDLOCKUP_CHECK_TIMESTAMP) && \
> -    defined(CONFIG_HARDLOCKUP_DETECTOR)
> +#if defined(CONFIG_HARDLOCKUP_CHECK_TIMESTAMP) && defined(CONFIG_HARDLOCKUP_DETECTOR_PERF)
>  void watchdog_update_hrtimer_threshold(u64 period);
>  #else
>  static inline void watchdog_update_hrtimer_threshold(u64 period) { }
> diff --git a/kernel/Makefile b/kernel/Makefile
> index 10ef068f598d..a2054f16f9f4 100644
> --- a/kernel/Makefile
> +++ b/kernel/Makefile
> @@ -91,6 +91,7 @@ obj-$(CONFIG_FAIL_FUNCTION) += fail_function.o
>  obj-$(CONFIG_KGDB) += debug/
>  obj-$(CONFIG_DETECT_HUNG_TASK) += hung_task.o
>  obj-$(CONFIG_LOCKUP_DETECTOR) += watchdog.o
> +obj-$(CONFIG_HARDLOCKUP_DETECTOR_BUDDY_CPU) += watchdog_buddy_cpu.o
>  obj-$(CONFIG_HARDLOCKUP_DETECTOR_PERF) += watchdog_hld.o
>  obj-$(CONFIG_SECCOMP) += seccomp.o
>  obj-$(CONFIG_RELAY) += relay.o
> diff --git a/kernel/watchdog.c b/kernel/watchdog.c
> index 8e61f21e7e33..1199043689ae 100644
> --- a/kernel/watchdog.c
> +++ b/kernel/watchdog.c
> @@ -29,7 +29,7 @@
>
>  static DEFINE_MUTEX(watchdog_mutex);
>
> -#if defined(CONFIG_HARDLOCKUP_DETECTOR) || defined(CONFIG_HAVE_NMI_WATCHDOG)
> +#if defined(CONFIG_HARDLOCKUP_DETECTOR_CORE) || defined(CONFIG_HAVE_NMI_WATCHDOG)
>  # define WATCHDOG_DEFAULT      (SOFT_WATCHDOG_ENABLED | NMI_WATCHDOG_ENABLED)
>  # define NMI_WATCHDOG_DEFAULT  1
>  #else
> @@ -47,7 +47,7 @@ static int __read_mostly nmi_watchdog_available;
>  struct cpumask watchdog_cpumask __read_mostly;
>  unsigned long *watchdog_cpumask_bits = cpumask_bits(&watchdog_cpumask);
>
> -#ifdef CONFIG_HARDLOCKUP_DETECTOR
> +#ifdef CONFIG_HARDLOCKUP_DETECTOR_CORE
>
>  # ifdef CONFIG_SMP
>  int __read_mostly sysctl_hardlockup_all_cpu_backtrace;
> @@ -85,7 +85,9 @@ static int __init hardlockup_panic_setup(char *str)
>  }
>  __setup("nmi_watchdog=", hardlockup_panic_setup);
>
> -#endif /* CONFIG_HARDLOCKUP_DETECTOR */
> +#endif /* CONFIG_HARDLOCKUP_DETECTOR_CORE */
> +
> +#ifdef CONFIG_HARDLOCKUP_DETECTOR
>
>  /*
>   * These functions can be overridden if an architecture implements its
> @@ -106,6 +108,13 @@ void __weak watchdog_nmi_disable(unsigned int cpu)
>         hardlockup_detector_perf_disable();
>  }
>
> +#else
> +
> +int __weak watchdog_nmi_enable(unsigned int cpu) { return 0; }
> +void __weak watchdog_nmi_disable(unsigned int cpu) { return; }
> +
> +#endif /* CONFIG_HARDLOCKUP_DETECTOR */
> +
>  /* Return 0, if a NMI watchdog is available. Error code otherwise */
>  int __weak __init watchdog_nmi_probe(void)
>  {
> @@ -179,8 +188,8 @@ static DEFINE_PER_CPU(unsigned long, watchdog_touch_ts);
>  static DEFINE_PER_CPU(unsigned long, watchdog_report_ts);
>  static DEFINE_PER_CPU(struct hrtimer, watchdog_hrtimer);
>  static DEFINE_PER_CPU(bool, softlockup_touch_sync);
> -static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts);
> -static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts_saved);
> +DEFINE_PER_CPU(unsigned long, hrtimer_interrupts);
> +DEFINE_PER_CPU(unsigned long, hrtimer_interrupts_saved);
>  static unsigned long soft_lockup_nmi_warn;
>
>  static int __init nowatchdog_setup(char *str)
> @@ -364,6 +373,9 @@ static enum hrtimer_restart watchdog_timer_fn(struct hrtimer *hrtimer)
>         /* kick the hardlockup detector */
>         watchdog_interrupt_count();
>
> +       /* test for hardlockups */
> +       watchdog_check_hardlockup();
> +
>         /* kick the softlockup detector */
>         if (completion_done(this_cpu_ptr(&softlockup_completion))) {
>                 reinit_completion(this_cpu_ptr(&softlockup_completion));
> @@ -820,7 +832,7 @@ static struct ctl_table watchdog_sysctls[] = {
>         },
>  #endif /* CONFIG_SMP */
>  #endif
> -#ifdef CONFIG_HARDLOCKUP_DETECTOR
> +#ifdef CONFIG_HARDLOCKUP_DETECTOR_CORE
>         {
>                 .procname       = "hardlockup_panic",
>                 .data           = &hardlockup_panic,
> diff --git a/kernel/watchdog_buddy_cpu.c b/kernel/watchdog_buddy_cpu.c
> new file mode 100644
> index 000000000000..db813b00e6ef
> --- /dev/null
> +++ b/kernel/watchdog_buddy_cpu.c
> @@ -0,0 +1,141 @@
> +// SPDX-License-Identifier: GPL-2.0
> +
> +#include <linux/cpu.h>
> +#include <linux/cpumask.h>
> +#include <linux/kernel.h>
> +#include <linux/nmi.h>
> +#include <linux/percpu-defs.h>
> +
> +static DEFINE_PER_CPU(bool, watchdog_touch);
> +static DEFINE_PER_CPU(bool, hard_watchdog_warn);
> +static cpumask_t __read_mostly watchdog_cpus;
> +
> +static unsigned long hardlockup_allcpu_dumped;
> +
> +int __init watchdog_nmi_probe(void)
> +{
> +       return 0;
> +}
> +
> +notrace void buddy_cpu_touch_watchdog(void)
> +{
> +       /*
> +        * Using __raw here because some code paths have
> +        * preemption enabled.  If preemption is enabled
> +        * then interrupts should be enabled too, in which
> +        * case we shouldn't have to worry about the watchdog
> +        * going off.
> +        */
> +       raw_cpu_write(watchdog_touch, true);
> +}
> +EXPORT_SYMBOL_GPL(buddy_cpu_touch_watchdog);
> +
> +static unsigned int watchdog_next_cpu(unsigned int cpu)
> +{
> +       cpumask_t cpus = watchdog_cpus;
> +       unsigned int next_cpu;
> +
> +       next_cpu = cpumask_next(cpu, &cpus);
> +       if (next_cpu >= nr_cpu_ids)
> +               next_cpu = cpumask_first(&cpus);
> +
> +       if (next_cpu == cpu)
> +               return nr_cpu_ids;
> +
> +       return next_cpu;
> +}
> +
> +int watchdog_nmi_enable(unsigned int cpu)
> +{
> +       /*
> +        * The new cpu will be marked online before the first hrtimer interrupt
> +        * runs on it.  If another cpu tests for a hardlockup on the new cpu
> +        * before it has run its first hrtimer, it will get a false positive.
> +        * Touch the watchdog on the new cpu to delay the first check for at
> +        * least 3 sampling periods to guarantee one hrtimer has run on the new
> +        * cpu.
> +        */
> +       per_cpu(watchdog_touch, cpu) = true;
> +       /* Match with smp_rmb() in watchdog_check_hardlockup() */
> +       smp_wmb();
> +       cpumask_set_cpu(cpu, &watchdog_cpus);
> +       return 0;
> +}
> +
> +void watchdog_nmi_disable(unsigned int cpu)
> +{
> +       unsigned int next_cpu = watchdog_next_cpu(cpu);
> +
> +       /*
> +        * Offlining this cpu will cause the cpu before this one to start
> +        * checking the one after this one.  If this cpu just finished checking
> +        * the next cpu and updating hrtimer_interrupts_saved, and then the
> +        * previous cpu checks it within one sample period, it will trigger a
> +        * false positive.  Touch the watchdog on the next cpu to prevent it.
> +        */
> +       if (next_cpu < nr_cpu_ids)
> +               per_cpu(watchdog_touch, next_cpu) = true;
> +       /* Match with smp_rmb() in watchdog_check_hardlockup() */
> +       smp_wmb();
> +       cpumask_clear_cpu(cpu, &watchdog_cpus);
> +}
> +
> +static int is_hardlockup_buddy_cpu(unsigned 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;
> +}
> +
> +void watchdog_check_hardlockup(void)
> +{
> +       unsigned int next_cpu;
> +
> +       /*
> +        * Test for hardlockups every 3 samples.  The sample period is
> +        *  watchdog_thresh * 2 / 5, so 3 samples gets us back to slightly over
> +        *  watchdog_thresh (over by 20%).
> +        */
> +       if (__this_cpu_read(hrtimer_interrupts) % 3 != 0)
> +               return;
> +
> +       /* check for a hardlockup on the next cpu */
> +       next_cpu = watchdog_next_cpu(smp_processor_id());
> +       if (next_cpu >= nr_cpu_ids)
> +               return;
> +
> +       /* Match with smp_wmb() in watchdog_nmi_enable() / watchdog_nmi_disable() */
> +       smp_rmb();
> +
> +       if (per_cpu(watchdog_touch, next_cpu) == true) {
> +               per_cpu(watchdog_touch, next_cpu) = false;
> +               return;
> +       }
> +
> +       if (is_hardlockup_buddy_cpu(next_cpu)) {
> +               /* only warn once */
> +               if (per_cpu(hard_watchdog_warn, next_cpu) == true)
> +                       return;
> +
> +               /*
> +                * Perform all-CPU dump only once to avoid multiple hardlockups
> +                * generating interleaving traces
> +                */
> +               if (sysctl_hardlockup_all_cpu_backtrace &&
> +                               !test_and_set_bit(0, &hardlockup_allcpu_dumped))
> +                       trigger_allbutself_cpu_backtrace();
> +
> +               if (hardlockup_panic)
> +                       panic("Watchdog detected hard LOCKUP on cpu %u", next_cpu);
> +               else
> +                       WARN(1, "Watchdog detected hard LOCKUP on cpu %u", next_cpu);
> +
> +               per_cpu(hard_watchdog_warn, next_cpu) = true;
> +       } else {
> +               per_cpu(hard_watchdog_warn, next_cpu) = false;
> +       }
> +}
> diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug
> index 39d1d93164bd..511eb14660b1 100644
> --- a/lib/Kconfig.debug
> +++ b/lib/Kconfig.debug
> @@ -1036,6 +1036,9 @@ config HARDLOCKUP_DETECTOR_PERF
>  config HARDLOCKUP_CHECK_TIMESTAMP
>         bool
>
> +config HARDLOCKUP_DETECTOR_CORE
> +       bool
> +
>  #
>  # arch/ can define HAVE_HARDLOCKUP_DETECTOR_ARCH to provide their own hard
>  # lockup detector rather than the perf based detector.
> @@ -1045,6 +1048,7 @@ config HARDLOCKUP_DETECTOR
>         depends on DEBUG_KERNEL && !S390
>         depends on HAVE_HARDLOCKUP_DETECTOR_PERF || HAVE_HARDLOCKUP_DETECTOR_ARCH
>         select LOCKUP_DETECTOR
> +       select HARDLOCKUP_DETECTOR_CORE
>         select HARDLOCKUP_DETECTOR_PERF if HAVE_HARDLOCKUP_DETECTOR_PERF
>         help
>           Say Y here to enable the kernel to act as a watchdog to detect
> @@ -1055,9 +1059,26 @@ config HARDLOCKUP_DETECTOR
>           chance to run.  The current stack trace is displayed upon detection
>           and the system will stay locked up.
>
> +config HARDLOCKUP_DETECTOR_BUDDY_CPU
> +       bool "Buddy CPU hardlockup detector"
> +       depends on DEBUG_KERNEL && SMP
> +       depends on !HARDLOCKUP_DETECTOR && !HAVE_NMI_WATCHDOG
> +       depends on !S390
> +       select HARDLOCKUP_DETECTOR_CORE
> +       select SOFTLOCKUP_DETECTOR
> +       help
> +         Say Y here to enable a hardlockup detector where CPUs check
> +         each other for lockup. Each cpu uses its softlockup hrtimer
> +         to check that the next cpu is processing hrtimer interrupts by
> +         verifying that a counter is increasing.
> +
> +         This hardlockup detector is useful on systems that don't have
> +         an arch-specific hardlockup detector or if resources needed
> +         for the hardlockup detector are better used for other things.
> +
>  config BOOTPARAM_HARDLOCKUP_PANIC
>         bool "Panic (Reboot) On Hard Lockups"
> -       depends on HARDLOCKUP_DETECTOR
> +       depends on HARDLOCKUP_DETECTOR_CORE
>         help
>           Say Y here to enable the kernel to panic on "hard lockups",
>           which are bugs that cause the kernel to loop in kernel
> --
> 2.40.1.495.gc816e09b53d-goog
>
  

Hi,


On Fri, Apr 28, 2023 at 5:36 PM Randy Dunlap <rdunlap@infradead.org> wrote:
>
> Hi--
>
> On 4/28/23 16:37, Douglas Anderson wrote:
> > From: Colin Cross <ccross@android.com>
> >
> > Implement a hardlockup detector that doesn't doesn't need any extra
> > arch-specific support code to detect lockups. Instead of using
> > something arch-specific we will use the buddy system, where each CPU
> > watches out for another one. Specifically, each CPU will use its
> > softlockup hrtimer to check that the next CPU is processing hrtimer
> > interrupts by verifying that a counter is increasing.
> >
> > NOTE: unlike the other hard lockup detectors, the buddy one can't
> > easily show what's happening on the CPU that locked up just by doing a
> > simple backtrace. It relies on some other mechanism in the system to
> > get information about the locked up CPUs. This could be support for
> > NMI backtraces like [1], it could be a mechanism for printing the PC
> > of locked CPUs at panic time like [2] / [3], or it could be something
> > else. Even though that means we still rely on arch-specific code, this
> > arch-specific code seems to often be implemented even on architectures
> > that don't have a hardlockup detector.
> >
> > This style of hardlockup detector originated in some downstream
> > Android trees and has been rebased on / carried in ChromeOS trees for
> > quite a long time for use on arm and arm64 boards. Historically on
> > these boards we've leveraged mechanism [2] / [3] to get information
> > about hung CPUs, but we could move to [1].
> >
> > Although the original motivation for the buddy system was for use on
> > systems without an arch-specific hardlockup detector, it can still be
> > useful to use even on systems that _do_ have an arch-specific
> > hardlockup detector. On x86, for instance, there is a 24-part patch
> > series [4] in progress switching the arch-specific hard lockup
> > detector from a scarce perf counter to a less-scarce hardware
> > resource. Potentially the buddy system could be a simpler alternative
> > to free up the perf counter but still get hard lockup detection.
> >
> > Overall, pros (+) and cons (-) of the buddy system compared to an
> > arch-specific hardlockup detector:
> > + Usable on systems that don't have an arch-specific hardlockup
> >   detector, like arm32 and arm64 (though it's being worked on for
> >   arm64 [5]).
> > + May free up scarce hardware resources.
> > + If a CPU totally goes out to lunch (can't process NMIs) the buddy
> >   system could still detect the problem (though it would be unlikely
> >   to be able to get a stack trace).
> > - If all CPUs are hard locked up at the same time the buddy system
> >   can't detect it.
> > - If we don't have SMP we can't use the buddy system.
> > - The buddy system needs an arch-specific mechanism (possibly NMI
> >   backtrace) to get info about the locked up CPU.
> >
> > [1] https://lore.kernel.org/r/20230419225604.21204-1-dianders@chromium.org
> > [2] https://issuetracker.google.com/172213129
> > [3] https://docs.kernel.org/trace/coresight/coresight-cpu-debug.html
> > [4] https://lore.kernel.org/lkml/20230301234753.28582-1-ricardo.neri-calderon@linux.intel.com/
> > [5] https://lore.kernel.org/linux-arm-kernel/20220903093415.15850-1-lecopzer.chen@mediatek.com/
> >
> > Signed-off-by: Colin Cross <ccross@android.com>
> > Signed-off-by: Matthias Kaehlcke <mka@chromium.org>
> > Signed-off-by: Guenter Roeck <groeck@chromium.org>
> > Signed-off-by: Tzung-Bi Shih <tzungbi@chromium.org>
> > Signed-off-by: Douglas Anderson <dianders@chromium.org>
> > ---
> > This patch has been rebased in ChromeOS kernel trees many times, and
> > each time someone had to do work on it they added their
> > Signed-off-by. I've included those here. I've also left the author as
> > Colin Cross since the core code is still his.
> >
> > I'll also note that the CC list is pretty giant, but that's what
> > get_maintainers came up with (plus a few other folks I thought would
> > be interested). As far as I can tell, there's no true MAINTAINER
> > listed for the existing watchdog code. Assuming people don't hate
> > this, maybe it would go through Andrew Morton's tree?
> >
> > Changes in v2:
> > - cpu => CPU.
> > - Reworked description and Kconfig based on v1 discussion.
>
> or at least some of the comments from v1. :(

Oh no! My email program confused me and I thought all of your cpu=>CPU
stuff was in the patch description, not in the Kconfig. I'll whip up a
quick v3.

-Doug
  

Hi,

On Sat, Apr 29, 2023 at 2:22 PM Ian Rogers <irogers@google.com> wrote:
>
> On Fri, Apr 28, 2023 at 4:41 PM Douglas Anderson <dianders@chromium.org> wrote:
> >
> > From: Colin Cross <ccross@android.com>
> >
> > Implement a hardlockup detector that doesn't doesn't need any extra
> > arch-specific support code to detect lockups. Instead of using
> > something arch-specific we will use the buddy system, where each CPU
> > watches out for another one. Specifically, each CPU will use its
> > softlockup hrtimer to check that the next CPU is processing hrtimer
> > interrupts by verifying that a counter is increasing.
> >
> > NOTE: unlike the other hard lockup detectors, the buddy one can't
> > easily show what's happening on the CPU that locked up just by doing a
> > simple backtrace. It relies on some other mechanism in the system to
> > get information about the locked up CPUs. This could be support for
> > NMI backtraces like [1], it could be a mechanism for printing the PC
> > of locked CPUs at panic time like [2] / [3], or it could be something
> > else. Even though that means we still rely on arch-specific code, this
> > arch-specific code seems to often be implemented even on architectures
> > that don't have a hardlockup detector.
> >
> > This style of hardlockup detector originated in some downstream
> > Android trees and has been rebased on / carried in ChromeOS trees for
> > quite a long time for use on arm and arm64 boards. Historically on
> > these boards we've leveraged mechanism [2] / [3] to get information
> > about hung CPUs, but we could move to [1].
> >
> > Although the original motivation for the buddy system was for use on
> > systems without an arch-specific hardlockup detector, it can still be
> > useful to use even on systems that _do_ have an arch-specific
> > hardlockup detector. On x86, for instance, there is a 24-part patch
> > series [4] in progress switching the arch-specific hard lockup
> > detector from a scarce perf counter to a less-scarce hardware
> > resource. Potentially the buddy system could be a simpler alternative
> > to free up the perf counter but still get hard lockup detection.
> >
> > Overall, pros (+) and cons (-) of the buddy system compared to an
> > arch-specific hardlockup detector:
> > + Usable on systems that don't have an arch-specific hardlockup
> >   detector, like arm32 and arm64 (though it's being worked on for
> >   arm64 [5]).
> > + May free up scarce hardware resources.
> > + If a CPU totally goes out to lunch (can't process NMIs) the buddy
> >   system could still detect the problem (though it would be unlikely
> >   to be able to get a stack trace).
> > - If all CPUs are hard locked up at the same time the buddy system
> >   can't detect it.
> > - If we don't have SMP we can't use the buddy system.
> > - The buddy system needs an arch-specific mechanism (possibly NMI
> >   backtrace) to get info about the locked up CPU.
>
> Thanks for this list, it is really useful! Is it worth mentioning the
> behavior around idle? Could this approach potentially use more power?

Sure, I'll add some text in there. If I'm analyzing the code properly,
my belief is that, if anything, the buddy detector should be better
for idle/power than some other detectors.

Specifically, note that the main "worker" of the buddy detector is
called from watchdog_timer_fn(). The timer function is the same one
that runs for other hard lockup detectors, but in those cases it
_only_ pets the watchdog of the running CPU. With the buddy detector
it pets the running CPU's watchdog and then checks on the buddy's
watchdog. There is no separate wakeup / interrupt that needs to run
periodically to look for hard lockups.

I'm about to send a v3 to fix the cpu=>CPU that I missed on v2. I'll
add text similar to the above to the commit message.

-Doug