Skip to contentSkip to conversation

perf: Fix system hang caused by cpu-clock

PATCH Applied Dapeng Mi 3 participants 6 messages Oct 15, 2025 – Nov 3, 2025

Subsystems (2)
Stable branch perf events
lore
Share & tools
lore

Dapeng Mi

author
#0 lore raw

PATCH 1/1 perf: Fix system hang caused by cpu-clock

A system hang issue caused by cpu-clock is reported and bisection indicates the commit 18dbcbfabfff ("perf: Fix the POLL_HUP delivery
breakage") causes this issue.

The root cause of the hang issue is that cpu-clock is a specific SW event which relies on the hrtimer. The __perf_event_overflow() is invoked from the hrtimer handler for cpu-clock event, and __perf_event_overflow() tries to call event stop callback (cpu_clock_event_stop()) to stop the event, and cpu_clock_event_stop() calls htimer_cancel() to cancel the hrtimer. But unfortunately the hrtimer callback is currently executing and then traps into deadlock.

To avoid this deadlock, use hrtimer_try_to_cancel() instead of hrtimer_cancel() to cancel the hrtimer, and set PERF_HES_STOPPED flag for the stopping events. perf_swevent_hrtimer() would stop the event hrtimer once it detects the PERF_HES_STOPPED flag.

Reported-by Octavia Togami <[email protected]>
Closes https://lore.kernel.org/all/CAHPNGSQpXEopYreir+uDDEbtXTBvBvi8c6fYXJvceqtgTPao3Q@mail.gmail.com/ open here
Suggested-by Peter Zijlstra <[email protected]>
Fixes 18dbcbfabfff ("perf: Fix the POLL_HUP delivery breakage")
Tested-by Octavia Togami <[email protected]>
Signed-off-by Dapeng Mi <[email protected]>
---
 kernel/events/core.c | 18 +++++++++++++-----
 1 file changed, 13 insertions(+), 5 deletions(-)
kernel/events/core.c +13 -5
index 7541f6f85fcb..f90105d5f26a 100644
@@ -11773,7 +11773,8 @@ static enum hrtimer_restart perf_swevent_hrtimer(struct hrtimer *hrtimer)
 
 	event = container_of(hrtimer, struct perf_event, hw.hrtimer);
 
-	if (event->state != PERF_EVENT_STATE_ACTIVE)
+	if (event->state != PERF_EVENT_STATE_ACTIVE ||
+	    event->hw.state & PERF_HES_STOPPED)
 		return HRTIMER_NORESTART;
 
 	event->pmu->read(event);

@@ -11819,15 +11820,18 @@ static void perf_swevent_cancel_hrtimer(struct perf_event *event)
 	struct hw_perf_event *hwc = &event->hw;
 
 	/*
-	 * The throttle can be triggered in the hrtimer handler.
-	 * The HRTIMER_NORESTART should be used to stop the timer,
-	 * rather than hrtimer_cancel(). See perf_swevent_hrtimer()
+	 * The event stop can be triggered in the hrtimer handler.
+	 * So use hrtimer_try_to_cancel() instead of hrtimer_cancel()
+	 * to stop the hrtimer() to avoid trapping into a dead loop.
+	 * Simultaneously the event would be set PERF_HES_STOPPED flag,
+	 * perf_swevent_hrtimer() would stop the event hrtimer once it
+	 * detects the PERF_HES_STOPPED flag.
 	 */
 	if (is_sampling_event(event) && (hwc->interrupts != MAX_INTERRUPTS)) {
 		ktime_t remaining = hrtimer_get_remaining(&hwc->hrtimer);
 		local64_set(&hwc->period_left, ktime_to_ns(remaining));
 
-		hrtimer_cancel(&hwc->hrtimer);
+		hrtimer_try_to_cancel(&hwc->hrtimer);
 	}
 }
 

@@ -11871,12 +11875,14 @@ static void cpu_clock_event_update(struct perf_event *event)
 
 static void cpu_clock_event_start(struct perf_event *event, int flags)
 {
+	event->hw.state = 0;
 	local64_set(&event->hw.prev_count, local_clock());
 	perf_swevent_start_hrtimer(event);
 }
 
 static void cpu_clock_event_stop(struct perf_event *event, int flags)
 {
+	event->hw.state = PERF_HES_STOPPED;
 	perf_swevent_cancel_hrtimer(event);
 	if (flags & PERF_EF_UPDATE)
 		cpu_clock_event_update(event);

@@ -11950,12 +11956,14 @@ static void task_clock_event_update(struct perf_event *event, u64 now)
 
 static void task_clock_event_start(struct perf_event *event, int flags)
 {
+	event->hw.state = 0;
 	local64_set(&event->hw.prev_count, event->ctx->time);
 	perf_swevent_start_hrtimer(event);
 }
 
 static void task_clock_event_stop(struct perf_event *event, int flags)
 {
+	event->hw.state = PERF_HES_STOPPED;
 	perf_swevent_cancel_hrtimer(event);
 	if (flags & PERF_EF_UPDATE)
 		task_clock_event_update(event, event->ctx->time);

base-commit: 3a8660878839faadb4f1a6dd72c3179c1df56787

Signature · 1 line
2.34.1
maintainer
#1 lore raw
Quoted 32 lines · Dapeng Mi

A system hang issue caused by cpu-clock is reported and bisection indicates the commit 18dbcbfabfff ("perf: Fix the POLL_HUP delivery
breakage") causes this issue.

The root cause of the hang issue is that cpu-clock is a specific SW event which relies on the hrtimer. The __perf_event_overflow() is invoked from the hrtimer handler for cpu-clock event, and __perf_event_overflow() tries to call event stop callback (cpu_clock_event_stop()) to stop the event, and cpu_clock_event_stop() calls htimer_cancel() to cancel the hrtimer. But unfortunately the hrtimer callback is currently executing and then traps into deadlock.

To avoid this deadlock, use hrtimer_try_to_cancel() instead of hrtimer_cancel() to cancel the hrtimer, and set PERF_HES_STOPPED flag for the stopping events. perf_swevent_hrtimer() would stop the event hrtimer once it detects the PERF_HES_STOPPED flag.

Reported-by Octavia Togami <[email protected]>
Closes https://lore.kernel.org/all/CAHPNGSQpXEopYreir+uDDEbtXTBvBvi8c6fYXJvceqtgTPao3Q@mail.gmail.com/ open here
Suggested-by Peter Zijlstra <[email protected]>
Fixes 18dbcbfabfff ("perf: Fix the POLL_HUP delivery breakage")
Tested-by Octavia Togami <[email protected]>
Signed-off-by Dapeng Mi <[email protected]>
---
 kernel/events/core.c | 18 +++++++++++++-----
 1 file changed, 13 insertions(+), 5 deletions(-)
kernel/events/core.c +0 -0
index 7541f6f85fcb..f90105d5f26a 100644

kernel/events/core.c static enum hrtimer_restart perf_swevent_hrtimer(struct hrtimer *hrtimer) · line 11780

Quoted 9 lines
kernel/events/core.c +2 -1
@@ -11773,7 +11773,8 @@ static enum hrtimer_restart perf_swevent_hrtimer(struct hrtimer *hrtimer)
 
 	event = container_of(hrtimer, struct perf_event, hw.hrtimer);
 
-	if (event->state != PERF_EVENT_STATE_ACTIVE)
+	if (event->state != PERF_EVENT_STATE_ACTIVE ||
+	    event->hw.state & PERF_HES_STOPPED)
 		return HRTIMER_NORESTART;
 
 	event->pmu->read(event);

I was wondering if we need a HES_STOPPED check after calling __perf_event_overflow(), but typically that will return 1 when it does the stop itself, which then already does NORESTART.

So yeah, I suppose this works. Let me go queue this up.

Thanks!

author
#2 lore raw
Quoted 41 lines · Peter Zijlstra
Quoted 40 lines · Dapeng Mi

A system hang issue caused by cpu-clock is reported and bisection indicates the commit 18dbcbfabfff ("perf: Fix the POLL_HUP delivery
breakage") causes this issue.

The root cause of the hang issue is that cpu-clock is a specific SW event which relies on the hrtimer. The __perf_event_overflow() is invoked from the hrtimer handler for cpu-clock event, and __perf_event_overflow() tries to call event stop callback (cpu_clock_event_stop()) to stop the event, and cpu_clock_event_stop() calls htimer_cancel() to cancel the hrtimer. But unfortunately the hrtimer callback is currently executing and then traps into deadlock.

To avoid this deadlock, use hrtimer_try_to_cancel() instead of hrtimer_cancel() to cancel the hrtimer, and set PERF_HES_STOPPED flag for the stopping events. perf_swevent_hrtimer() would stop the event hrtimer once it detects the PERF_HES_STOPPED flag.

Reported-by Octavia Togami <[email protected]>
Closes https://lore.kernel.org/all/CAHPNGSQpXEopYreir+uDDEbtXTBvBvi8c6fYXJvceqtgTPao3Q@mail.gmail.com/ open here
Suggested-by Peter Zijlstra <[email protected]>
Fixes 18dbcbfabfff ("perf: Fix the POLL_HUP delivery breakage")
Tested-by Octavia Togami <[email protected]>
Signed-off-by Dapeng Mi <[email protected]>
---
 kernel/events/core.c | 18 +++++++++++++-----
 1 file changed, 13 insertions(+), 5 deletions(-)
kernel/events/core.c +2 -1
index 7541f6f85fcb..f90105d5f26a 100644
@@ -11773,7 +11773,8 @@ static enum hrtimer_restart perf_swevent_hrtimer(struct hrtimer *hrtimer)
 
 	event = container_of(hrtimer, struct perf_event, hw.hrtimer);
 
-	if (event->state != PERF_EVENT_STATE_ACTIVE)
+	if (event->state != PERF_EVENT_STATE_ACTIVE ||
+	    event->hw.state & PERF_HES_STOPPED)
 		return HRTIMER_NORESTART;
Quoted 4 lines

kernel/events/core.c static enum hrtimer_restart perf_swevent_hrtimer(struct hrtimer *hrtimer) · line 11780

Quoted 1 line
kernel/events/core.c +0 -0
 	event->pmu->read(event);

I was wondering if we need a HES_STOPPED check after calling __perf_event_overflow(), but typically that will return 1 when it does the stop itself, which then already does NORESTART.

Yes.

Quoted 2 lines

So yeah, I suppose this works. Let me go queue this up.

Thanks for reviewing this patch.

Quoted 3 lines

Thanks!

merge bot Applied
#3 lore raw
In reply to Dapeng Mi
The following commit has been merged into the perf/urgent branch of tip:

The following commit has been merged into the perf/urgent branch of tip:

Commit-ID:     e33076f34f7a449c0e1808f15d88b2dd9a85979a
Gitweb:        https://git.kernel.org/tip/e33076f34f7a449c0e1808f15d88b2dd9a85979a
Author:        Dapeng Mi <[email protected]>
AuthorDate:    Wed, 15 Oct 2025 13:18:28 +08:00
Committer:     Peter Zijlstra <[email protected]>
CommitterDate: Mon, 03 Nov 2025 10:26:09 +01:00

perf: Fix system hang caused by cpu-clock

A system hang issue caused by cpu-clock is reported and bisection indicates the commit 18dbcbfabfff ("perf: Fix the POLL_HUP delivery
breakage") causes this issue.

The root cause of the hang issue is that cpu-clock is a specific SW event which relies on the hrtimer. The __perf_event_overflow() is invoked from the hrtimer handler for cpu-clock event, and __perf_event_overflow() tries to call event stop callback (cpu_clock_event_stop()) to stop the event, and cpu_clock_event_stop() calls htimer_cancel() to cancel the hrtimer. But unfortunately the hrtimer callback is currently executing and then traps into deadlock.

To avoid this deadlock, use hrtimer_try_to_cancel() instead of hrtimer_cancel() to cancel the hrtimer, and set PERF_HES_STOPPED flag for the stopping events. perf_swevent_hrtimer() would stop the event hrtimer once it detects the PERF_HES_STOPPED flag.

Closes https://lore.kernel.org/all/CAHPNGSQpXEopYreir+uDDEbtXTBvBvi8c6fYXJvceqtgTPao3Q@mail.gmail.com/ open here
Fixes 18dbcbfabfff ("perf: Fix the POLL_HUP delivery breakage")
Reported-by Octavia Togami <[email protected]>
Suggested-by Peter Zijlstra <[email protected]>
Signed-off-by Dapeng Mi <[email protected]>
Signed-off-by Peter Zijlstra (Intel) <[email protected]>
Tested-by Octavia Togami <[email protected]>
Link https://patch.msgid.link/[email protected] open here
---
 kernel/events/core.c | 18 +++++++++++++-----
 1 file changed, 13 insertions(+), 5 deletions(-)
kernel/events/core.c +13 -5
index 177e57c..6e4af97 100644
@@ -11773,7 +11773,8 @@ static enum hrtimer_restart perf_swevent_hrtimer(struct hrtimer *hrtimer)
 
 	event = container_of(hrtimer, struct perf_event, hw.hrtimer);
 
-	if (event->state != PERF_EVENT_STATE_ACTIVE)
+	if (event->state != PERF_EVENT_STATE_ACTIVE ||
+	    event->hw.state & PERF_HES_STOPPED)
 		return HRTIMER_NORESTART;
 
 	event->pmu->read(event);

@@ -11819,15 +11820,18 @@ static void perf_swevent_cancel_hrtimer(struct perf_event *event)
 	struct hw_perf_event *hwc = &event->hw;
 
 	/*
-	 * The throttle can be triggered in the hrtimer handler.
-	 * The HRTIMER_NORESTART should be used to stop the timer,
-	 * rather than hrtimer_cancel(). See perf_swevent_hrtimer()
+	 * The event stop can be triggered in the hrtimer handler.
+	 * So use hrtimer_try_to_cancel() instead of hrtimer_cancel()
+	 * to stop the hrtimer() to avoid trapping into a dead loop.
+	 * Simultaneously the event would be set PERF_HES_STOPPED flag,
+	 * perf_swevent_hrtimer() would stop the event hrtimer once it
+	 * detects the PERF_HES_STOPPED flag.
 	 */
 	if (is_sampling_event(event) && (hwc->interrupts != MAX_INTERRUPTS)) {
 		ktime_t remaining = hrtimer_get_remaining(&hwc->hrtimer);
 		local64_set(&hwc->period_left, ktime_to_ns(remaining));
 
-		hrtimer_cancel(&hwc->hrtimer);
+		hrtimer_try_to_cancel(&hwc->hrtimer);
 	}
 }
 

@@ -11871,12 +11875,14 @@ static void cpu_clock_event_update(struct perf_event *event)
 
 static void cpu_clock_event_start(struct perf_event *event, int flags)
 {
+	event->hw.state = 0;
 	local64_set(&event->hw.prev_count, local_clock());
 	perf_swevent_start_hrtimer(event);
 }
 
 static void cpu_clock_event_stop(struct perf_event *event, int flags)
 {
+	event->hw.state = PERF_HES_STOPPED;
 	perf_swevent_cancel_hrtimer(event);
 	if (flags & PERF_EF_UPDATE)
 		cpu_clock_event_update(event);

@@ -11950,12 +11956,14 @@ static void task_clock_event_update(struct perf_event *event, u64 now)
 
 static void task_clock_event_start(struct perf_event *event, int flags)
 {
+	event->hw.state = 0;
 	local64_set(&event->hw.prev_count, event->ctx->time);
 	perf_swevent_start_hrtimer(event);
 }
 
 static void task_clock_event_stop(struct perf_event *event, int flags)
 {
+	event->hw.state = PERF_HES_STOPPED;
 	perf_swevent_cancel_hrtimer(event);
 	if (flags & PERF_EF_UPDATE)
 		task_clock_event_update(event, event->ctx->time);
merge bot Applied
#4 lore raw

[tip: perf/urgent] perf/core: Fix system hang caused by cpu-clock usage

In reply to Dapeng Mi
The following commit has been merged into the perf/urgent branch of tip:

The following commit has been merged into the perf/urgent branch of tip:

Commit-ID:     e061eb22817cb15e65b91e46d3fa8cd5ae60f9f4
Gitweb:        https://git.kernel.org/tip/e061eb22817cb15e65b91e46d3fa8cd5ae60f9f4
Author:        Dapeng Mi <[email protected]>
AuthorDate:    Wed, 15 Oct 2025 13:18:28 +08:00
Committer:     Ingo Molnar <[email protected]>
CommitterDate: Mon, 03 Nov 2025 10:52:10 +01:00

perf/core: Fix system hang caused by cpu-clock usage

cpu-clock usage by the async-profiler tool can trigger a system hang, which got bisected back to the following commit by Octavia Togami:

18dbcbfabfff ("perf: Fix the POLL_HUP delivery breakage") causes this issue

The root cause of the hang is that cpu-clock is a special type of SW event which relies on hrtimers. The __perf_event_overflow() callback is invoked from the hrtimer handler for cpu-clock events, and __perf_event_overflow() tries to call cpu_clock_event_stop() to stop the event, which calls htimer_cancel() to cancel the hrtimer.

But that's a recursion into the hrtimer code from a hrtimer handler, which (unsurprisingly) deadlocks.

To fix this bug, use hrtimer_try_to_cancel() instead, and set the PERF_HES_STOPPED flag, which causes perf_swevent_hrtimer() to stop the event once it sees the PERF_HES_STOPPED flag.

[ mingo: Fixed the comments and improved the changelog. ]

Closes https://lore.kernel.org/all/CAHPNGSQpXEopYreir+uDDEbtXTBvBvi8c6fYXJvceqtgTPao3Q@mail.gmail.com/ open here
Fixes 18dbcbfabfff ("perf: Fix the POLL_HUP delivery breakage")
Reported-by Octavia Togami <[email protected]>
Suggested-by Peter Zijlstra <[email protected]>
Signed-off-by Dapeng Mi <[email protected]>
Signed-off-by Peter Zijlstra (Intel) <[email protected]>
Signed-off-by Ingo Molnar <[email protected]>
Tested-by Octavia Togami <[email protected]>
Link https://github.com/lucko/spark/issues/530
Link https://patch.msgid.link/[email protected] open here
---
 kernel/events/core.c | 20 +++++++++++++++-----
 1 file changed, 15 insertions(+), 5 deletions(-)
kernel/events/core.c +15 -5
index 177e57c..1fd347d 100644
@@ -11773,7 +11773,8 @@ static enum hrtimer_restart perf_swevent_hrtimer(struct hrtimer *hrtimer)
 
 	event = container_of(hrtimer, struct perf_event, hw.hrtimer);
 
-	if (event->state != PERF_EVENT_STATE_ACTIVE)
+	if (event->state != PERF_EVENT_STATE_ACTIVE ||
+	    event->hw.state & PERF_HES_STOPPED)
 		return HRTIMER_NORESTART;
 
 	event->pmu->read(event);

@@ -11819,15 +11820,20 @@ static void perf_swevent_cancel_hrtimer(struct perf_event *event)
 	struct hw_perf_event *hwc = &event->hw;
 
 	/*
-	 * The throttle can be triggered in the hrtimer handler.
-	 * The HRTIMER_NORESTART should be used to stop the timer,
-	 * rather than hrtimer_cancel(). See perf_swevent_hrtimer()
+	 * Careful: this function can be triggered in the hrtimer handler,
+	 * for cpu-clock events, so hrtimer_cancel() would cause a
+	 * deadlock.
+	 *
+	 * So use hrtimer_try_to_cancel() to try to stop the hrtimer,
+	 * and the cpu-clock handler also sets the PERF_HES_STOPPED flag,
+	 * which guarantees that perf_swevent_hrtimer() will stop the
+	 * hrtimer once it sees the PERF_HES_STOPPED flag.
 	 */
 	if (is_sampling_event(event) && (hwc->interrupts != MAX_INTERRUPTS)) {
 		ktime_t remaining = hrtimer_get_remaining(&hwc->hrtimer);
 		local64_set(&hwc->period_left, ktime_to_ns(remaining));
 
-		hrtimer_cancel(&hwc->hrtimer);
+		hrtimer_try_to_cancel(&hwc->hrtimer);
 	}
 }
 

@@ -11871,12 +11877,14 @@ static void cpu_clock_event_update(struct perf_event *event)
 
 static void cpu_clock_event_start(struct perf_event *event, int flags)
 {
+	event->hw.state = 0;
 	local64_set(&event->hw.prev_count, local_clock());
 	perf_swevent_start_hrtimer(event);
 }
 
 static void cpu_clock_event_stop(struct perf_event *event, int flags)
 {
+	event->hw.state = PERF_HES_STOPPED;
 	perf_swevent_cancel_hrtimer(event);
 	if (flags & PERF_EF_UPDATE)
 		cpu_clock_event_update(event);

@@ -11950,12 +11958,14 @@ static void task_clock_event_update(struct perf_event *event, u64 now)
 
 static void task_clock_event_start(struct perf_event *event, int flags)
 {
+	event->hw.state = 0;
 	local64_set(&event->hw.prev_count, event->ctx->time);
 	perf_swevent_start_hrtimer(event);
 }
 
 static void task_clock_event_stop(struct perf_event *event, int flags)
 {
+	event->hw.state = PERF_HES_STOPPED;
 	perf_swevent_cancel_hrtimer(event);
 	if (flags & PERF_EF_UPDATE)
 		task_clock_event_update(event, event->ctx->time);
merge bot Applied
#5 lore raw

[tip: perf/urgent] perf/core: Fix system hang caused by cpu-clock usage

In reply to Dapeng Mi
The following commit has been merged into the perf/urgent branch of tip:

The following commit has been merged into the perf/urgent branch of tip:

Commit-ID:     eb3182ef0405ff2f6668fd3e5ff9883f60ce8801
Gitweb:        https://git.kernel.org/tip/eb3182ef0405ff2f6668fd3e5ff9883f60ce8801
Author:        Dapeng Mi <[email protected]>
AuthorDate:    Wed, 15 Oct 2025 13:18:28 +08:00
Committer:     Ingo Molnar <[email protected]>
CommitterDate: Mon, 03 Nov 2025 11:04:19 +01:00

perf/core: Fix system hang caused by cpu-clock usage

cpu-clock usage by the async-profiler tool can trigger a system hang, which got bisected back to the following commit by Octavia Togami:

18dbcbfabfff ("perf: Fix the POLL_HUP delivery breakage") causes this issue

The root cause of the hang is that cpu-clock is a special type of SW event which relies on hrtimers. The __perf_event_overflow() callback is invoked from the hrtimer handler for cpu-clock events, and __perf_event_overflow() tries to call cpu_clock_event_stop() to stop the event, which calls htimer_cancel() to cancel the hrtimer.

But that's a recursion into the hrtimer code from a hrtimer handler, which (unsurprisingly) deadlocks.

To fix this bug, use hrtimer_try_to_cancel() instead, and set the PERF_HES_STOPPED flag, which causes perf_swevent_hrtimer() to stop the event once it sees the PERF_HES_STOPPED flag.

[ mingo: Fixed the comments and improved the changelog. ]

Closes https://lore.kernel.org/all/CAHPNGSQpXEopYreir+uDDEbtXTBvBvi8c6fYXJvceqtgTPao3Q@mail.gmail.com/ open here
Fixes 18dbcbfabfff ("perf: Fix the POLL_HUP delivery breakage")
Reported-by Octavia Togami <[email protected]>
Suggested-by Peter Zijlstra <[email protected]>
Signed-off-by Dapeng Mi <[email protected]>
Signed-off-by Peter Zijlstra (Intel) <[email protected]>
Signed-off-by Ingo Molnar <[email protected]>
Tested-by Octavia Togami <[email protected]>
Cc [email protected]
Link https://github.com/lucko/spark/issues/530
Link https://patch.msgid.link/[email protected] open here
---
 kernel/events/core.c | 20 +++++++++++++++-----
 1 file changed, 15 insertions(+), 5 deletions(-)
kernel/events/core.c +15 -5
index 177e57c..1fd347d 100644
@@ -11773,7 +11773,8 @@ static enum hrtimer_restart perf_swevent_hrtimer(struct hrtimer *hrtimer)
 
 	event = container_of(hrtimer, struct perf_event, hw.hrtimer);
 
-	if (event->state != PERF_EVENT_STATE_ACTIVE)
+	if (event->state != PERF_EVENT_STATE_ACTIVE ||
+	    event->hw.state & PERF_HES_STOPPED)
 		return HRTIMER_NORESTART;
 
 	event->pmu->read(event);

@@ -11819,15 +11820,20 @@ static void perf_swevent_cancel_hrtimer(struct perf_event *event)
 	struct hw_perf_event *hwc = &event->hw;
 
 	/*
-	 * The throttle can be triggered in the hrtimer handler.
-	 * The HRTIMER_NORESTART should be used to stop the timer,
-	 * rather than hrtimer_cancel(). See perf_swevent_hrtimer()
+	 * Careful: this function can be triggered in the hrtimer handler,
+	 * for cpu-clock events, so hrtimer_cancel() would cause a
+	 * deadlock.
+	 *
+	 * So use hrtimer_try_to_cancel() to try to stop the hrtimer,
+	 * and the cpu-clock handler also sets the PERF_HES_STOPPED flag,
+	 * which guarantees that perf_swevent_hrtimer() will stop the
+	 * hrtimer once it sees the PERF_HES_STOPPED flag.
 	 */
 	if (is_sampling_event(event) && (hwc->interrupts != MAX_INTERRUPTS)) {
 		ktime_t remaining = hrtimer_get_remaining(&hwc->hrtimer);
 		local64_set(&hwc->period_left, ktime_to_ns(remaining));
 
-		hrtimer_cancel(&hwc->hrtimer);
+		hrtimer_try_to_cancel(&hwc->hrtimer);
 	}
 }
 

@@ -11871,12 +11877,14 @@ static void cpu_clock_event_update(struct perf_event *event)
 
 static void cpu_clock_event_start(struct perf_event *event, int flags)
 {
+	event->hw.state = 0;
 	local64_set(&event->hw.prev_count, local_clock());
 	perf_swevent_start_hrtimer(event);
 }
 
 static void cpu_clock_event_stop(struct perf_event *event, int flags)
 {
+	event->hw.state = PERF_HES_STOPPED;
 	perf_swevent_cancel_hrtimer(event);
 	if (flags & PERF_EF_UPDATE)
 		cpu_clock_event_update(event);

@@ -11950,12 +11958,14 @@ static void task_clock_event_update(struct perf_event *event, u64 now)
 
 static void task_clock_event_start(struct perf_event *event, int flags)
 {
+	event->hw.state = 0;
 	local64_set(&event->hw.prev_count, event->ctx->time);
 	perf_swevent_start_hrtimer(event);
 }
 
 static void task_clock_event_stop(struct perf_event *event, int flags)
 {
+	event->hw.state = PERF_HES_STOPPED;
 	perf_swevent_cancel_hrtimer(event);
 	if (flags & PERF_EF_UPDATE)
 		task_clock_event_update(event, event->ctx->time);