[PATCH v5 11/18] watchdog/hardlockup: Move perf hardlockup watchdog petting to watchdog.c
From: Douglas Anderson
Date: Fri May 19 2023 - 13:22:04 EST
In preparation for the buddy hardlockup detector, which wants the same
petting logic as the current perf hardlockup detector, move the code
to watchdog.c. While doing this, rename the global variable to match
others nearby. As part of this change we have to change the code to
account for the fact that the CPU we're running on might be different
than the one we're checking.
Currently the code in watchdog.c is guarded by
CONFIG_HARDLOCKUP_DETECTOR_PERF, which makes this change seem
silly. However, a future patch will change this.
Signed-off-by: Douglas Anderson <dianders@xxxxxxxxxxxx>
---
Changes in v5:
- Fixed wrong __this_cpu to per_cpu (oops).
- Move side effect (timestamp check ordering) to its own patch.
- watchdog_hardlockup_touch => watchdog_hardlockup_touched.
Changes in v4:
- ("Move perf hardlockup watchdog petting to watchdog.c") new for v4.
include/linux/nmi.h | 5 +++--
kernel/watchdog.c | 19 +++++++++++++++++++
kernel/watchdog_perf.c | 19 -------------------
3 files changed, 22 insertions(+), 21 deletions(-)
diff --git a/include/linux/nmi.h b/include/linux/nmi.h
index 92aa568c0c42..e286a2a1902d 100644
--- a/include/linux/nmi.h
+++ b/include/linux/nmi.h
@@ -88,7 +88,10 @@ static inline void hardlockup_detector_disable(void) {}
#endif
#if defined(CONFIG_HARDLOCKUP_DETECTOR_PERF)
+void arch_touch_nmi_watchdog(void);
void watchdog_hardlockup_check(unsigned int cpu, struct pt_regs *regs);
+#elif !defined(CONFIG_HAVE_NMI_WATCHDOG)
+static inline void arch_touch_nmi_watchdog(void) { }
#endif
#if defined(CONFIG_HAVE_NMI_WATCHDOG) || defined(CONFIG_HARDLOCKUP_DETECTOR)
@@ -98,7 +101,6 @@ void watchdog_hardlockup_check(unsigned int cpu, struct pt_regs *regs);
#endif
#if defined(CONFIG_HARDLOCKUP_DETECTOR_PERF)
-extern void arch_touch_nmi_watchdog(void);
extern void hardlockup_detector_perf_stop(void);
extern void hardlockup_detector_perf_restart(void);
extern void hardlockup_detector_perf_disable(void);
@@ -113,7 +115,6 @@ static inline void hardlockup_detector_perf_enable(void) { }
static inline void hardlockup_detector_perf_cleanup(void) { }
# if !defined(CONFIG_HAVE_NMI_WATCHDOG)
static inline int hardlockup_detector_perf_init(void) { return -ENODEV; }
-static inline void arch_touch_nmi_watchdog(void) {}
# else
static inline int hardlockup_detector_perf_init(void) { return 0; }
# endif
diff --git a/kernel/watchdog.c b/kernel/watchdog.c
index 2552e224f76a..64d7d2a0a7df 100644
--- a/kernel/watchdog.c
+++ b/kernel/watchdog.c
@@ -90,8 +90,22 @@ __setup("nmi_watchdog=", hardlockup_panic_setup);
static DEFINE_PER_CPU(atomic_t, hrtimer_interrupts);
static DEFINE_PER_CPU(int, hrtimer_interrupts_saved);
static DEFINE_PER_CPU(bool, watchdog_hardlockup_warned);
+static DEFINE_PER_CPU(bool, watchdog_hardlockup_touched);
static unsigned long watchdog_hardlockup_all_cpu_dumped;
+notrace void arch_touch_nmi_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_hardlockup_touched, true);
+}
+EXPORT_SYMBOL(arch_touch_nmi_watchdog);
+
static bool is_hardlockup(unsigned int cpu)
{
int hrint = atomic_read(&per_cpu(hrtimer_interrupts, cpu));
@@ -116,6 +130,11 @@ static void watchdog_hardlockup_kick(void)
void watchdog_hardlockup_check(unsigned int cpu, struct pt_regs *regs)
{
+ if (per_cpu(watchdog_hardlockup_touched, cpu)) {
+ per_cpu(watchdog_hardlockup_touched, cpu) = false;
+ return;
+ }
+
/*
* Check for a hardlockup by making sure the CPU's timer
* interrupt is incrementing. The timer interrupt should have
diff --git a/kernel/watchdog_perf.c b/kernel/watchdog_perf.c
index 4e60e8023515..547917ebd5d3 100644
--- a/kernel/watchdog_perf.c
+++ b/kernel/watchdog_perf.c
@@ -20,26 +20,12 @@
#include <asm/irq_regs.h>
#include <linux/perf_event.h>
-static DEFINE_PER_CPU(bool, watchdog_nmi_touch);
static DEFINE_PER_CPU(struct perf_event *, watchdog_ev);
static DEFINE_PER_CPU(struct perf_event *, dead_event);
static struct cpumask dead_events_mask;
static atomic_t watchdog_cpus = ATOMIC_INIT(0);
-notrace void arch_touch_nmi_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_nmi_touch, true);
-}
-EXPORT_SYMBOL(arch_touch_nmi_watchdog);
-
#ifdef CONFIG_HARDLOCKUP_CHECK_TIMESTAMP
static DEFINE_PER_CPU(ktime_t, last_timestamp);
static DEFINE_PER_CPU(unsigned int, nmi_rearmed);
@@ -115,11 +101,6 @@ static void watchdog_overflow_callback(struct perf_event *event,
if (!watchdog_check_timestamp())
return;
- if (__this_cpu_read(watchdog_nmi_touch) == true) {
- __this_cpu_write(watchdog_nmi_touch, false);
- return;
- }
-
watchdog_hardlockup_check(smp_processor_id(), regs);
}
--
2.40.1.698.g37aff9b760-goog