[PATCH v4] perf/core: Fix missing wakeup when waiting for context reference

[PATCH v4] perf/core: Fix missing wakeup when waiting for context reference

[Haifeng Xu]

In our production environment, we found many hung tasks which are
blocked for more than 18 hours. Their call traces are like this:

[346278.191038] __schedule+0x2d8/0x890
[346278.191046] schedule+0x4e/0xb0
[346278.191049] perf_event_free_task+0x220/0x270
[346278.191056] ? init_wait_var_entry+0x50/0x50
[346278.191060] copy_process+0x663/0x18d0
[346278.191068] kernel_clone+0x9d/0x3d0
[346278.191072] __do_sys_clone+0x5d/0x80
[346278.191076] __x64_sys_clone+0x25/0x30
[346278.191079] do_syscall_64+0x5c/0xc0
[346278.191083] ? syscall_exit_to_user_mode+0x27/0x50
[346278.191086] ? do_syscall_64+0x69/0xc0
[346278.191088] ? irqentry_exit_to_user_mode+0x9/0x20
[346278.191092] ? irqentry_exit+0x19/0x30
[346278.191095] ? exc_page_fault+0x89/0x160
[346278.191097] ? asm_exc_page_fault+0x8/0x30
[346278.191102] entry_SYSCALL_64_after_hwframe+0x44/0xae

The task was waiting for the refcount become to 1, but from the vmcore,
we found the refcount has already been 1. It seems that the task didn't
get woken up by perf_event_release_kernel() and got stuck forever. The
below scenario may cause the problem.

Thread A					Thread B
...						...
perf_event_free_task				perf_event_release_kernel
						   ...
						   acquire event->child_mutex
						   ...
						   get_ctx
   ...						   release event->child_mutex
   acquire ctx->mutex
   ...
   perf_free_event (acquire/release event->child_mutex)
   ...
   release ctx->mutex
   wait_var_event
						   acquire ctx->mutex
						   acquire event->child_mutex
						   # move existing events to free_list
						   release event->child_mutex
						   release ctx->mutex
						   put_ctx
...						...

In this case, all events of the ctx have been freed, so we couldn't
find the ctx in free_list and Thread A will miss the wakeup. It's thus
necessary to add a wakeup after dropping the reference.

Fixes: 1cf8dfe8a661 ("perf/core: Fix race between close() and fork()")
Cc: stable@vger.kernel.org
Signed-off-by: Haifeng Xu <haifeng.xu@shopee.com>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Mark Rutland <mark.rutland@arm.com>
---
Changes since v1:
- Add the fixed tag.
- Simplify v1's patch. (Frederic)

Changes since v2:
- Use Reviewed-by tag instead of Signed-off-by tag.

Changes since v3:
- Add Acked-by tag.
- Cc stable@vger.kernel.org. (Mark)
---
 kernel/events/core.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)

diff --git a/kernel/events/core.c b/kernel/events/core.c
index 4f0c45ab8d7d..15c35070db6a 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -5340,6 +5340,7 @@ int perf_event_release_kernel(struct perf_event *event)
 again:
 	mutex_lock(&event->child_mutex);
 	list_for_each_entry(child, &event->child_list, child_list) {
+		void *var = NULL;
 
 		/*
 		 * Cannot change, child events are not migrated, see the
@@ -5380,11 +5381,23 @@ int perf_event_release_kernel(struct perf_event *event)
 			 * this can't be the last reference.
 			 */
 			put_event(event);
+		} else {
+			var = &ctx->refcount;
 		}
 
 		mutex_unlock(&event->child_mutex);
 		mutex_unlock(&ctx->mutex);
 		put_ctx(ctx);
+
+		if (var) {
+			/*
+			 * If perf_event_free_task() has deleted all events from the
+			 * ctx while the child_mutex got released above, make sure to
+			 * notify about the preceding put_ctx().
+			 */
+			smp_mb(); /* pairs with wait_var_event() */
+			wake_up_var(var);
+		}
 		goto again;
 	}
 	mutex_unlock(&event->child_mutex);
-- 
2.25.1

Re: [PATCH v4] perf/core: Fix missing wakeup when waiting for context reference

[Peter Zijlstra]

On Mon, May 13, 2024 at 10:39:48AM +0000, Haifeng Xu wrote:
> In our production environment, we found many hung tasks which are
> blocked for more than 18 hours. Their call traces are like this:
> 
> [346278.191038] __schedule+0x2d8/0x890
> [346278.191046] schedule+0x4e/0xb0
> [346278.191049] perf_event_free_task+0x220/0x270
> [346278.191056] ? init_wait_var_entry+0x50/0x50
> [346278.191060] copy_process+0x663/0x18d0
> [346278.191068] kernel_clone+0x9d/0x3d0
> [346278.191072] __do_sys_clone+0x5d/0x80
> [346278.191076] __x64_sys_clone+0x25/0x30
> [346278.191079] do_syscall_64+0x5c/0xc0
> [346278.191083] ? syscall_exit_to_user_mode+0x27/0x50
> [346278.191086] ? do_syscall_64+0x69/0xc0
> [346278.191088] ? irqentry_exit_to_user_mode+0x9/0x20
> [346278.191092] ? irqentry_exit+0x19/0x30
> [346278.191095] ? exc_page_fault+0x89/0x160
> [346278.191097] ? asm_exc_page_fault+0x8/0x30
> [346278.191102] entry_SYSCALL_64_after_hwframe+0x44/0xae
> 
> The task was waiting for the refcount become to 1, but from the vmcore,
> we found the refcount has already been 1. It seems that the task didn't
> get woken up by perf_event_release_kernel() and got stuck forever. The
> below scenario may cause the problem.
> 
> Thread A					Thread B
> ...						...
> perf_event_free_task				perf_event_release_kernel
> 						   ...
> 						   acquire event->child_mutex
> 						   ...
> 						   get_ctx
>    ...						   release event->child_mutex
>    acquire ctx->mutex
>    ...
>    perf_free_event (acquire/release event->child_mutex)
>    ...
>    release ctx->mutex
>    wait_var_event
> 						   acquire ctx->mutex
> 						   acquire event->child_mutex
> 						   # move existing events to free_list
> 						   release event->child_mutex
> 						   release ctx->mutex
> 						   put_ctx
> ...						...
> 
> In this case, all events of the ctx have been freed, so we couldn't
> find the ctx in free_list and Thread A will miss the wakeup. It's thus
> necessary to add a wakeup after dropping the reference.
> 
> Fixes: 1cf8dfe8a661 ("perf/core: Fix race between close() and fork()")
> Cc: stable@vger.kernel.org
> Signed-off-by: Haifeng Xu <haifeng.xu@shopee.com>
> Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
> Acked-by: Mark Rutland <mark.rutland@arm.com>

Thanks!, I'll hang onto this until after the merge window and then stick
it in tip/perf/urgent or somesuch.

Re: [PATCH v4] perf/core: Fix missing wakeup when waiting for context reference

[Mark Rutland]

On Thu, May 16, 2024 at 10:51:06AM +0200, Peter Zijlstra wrote:
> Thanks!, I'll hang onto this until after the merge window and then stick
> it in tip/perf/urgent or somesuch.

Just to check -- I couldn't spot this in tip/perf/urgent just now; are
you still happy to pick this up?

Mark.

[tip: perf/urgent] perf/core: Fix missing wakeup when waiting for context reference

[tip-bot2 for Haifeng Xu]

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

Commit-ID:     74751ef5c1912ebd3e65c3b65f45587e05ce5d36
Gitweb:        https://git.kernel.org/tip/74751ef5c1912ebd3e65c3b65f45587e05ce5d36
Author:        Haifeng Xu <haifeng.xu@shopee.com>
AuthorDate:    Mon, 13 May 2024 10:39:48 
Committer:     Peter Zijlstra <peterz@infradead.org>
CommitterDate: Wed, 05 Jun 2024 15:52:33 +02:00

perf/core: Fix missing wakeup when waiting for context reference

In our production environment, we found many hung tasks which are
blocked for more than 18 hours. Their call traces are like this:

[346278.191038] __schedule+0x2d8/0x890
[346278.191046] schedule+0x4e/0xb0
[346278.191049] perf_event_free_task+0x220/0x270
[346278.191056] ? init_wait_var_entry+0x50/0x50
[346278.191060] copy_process+0x663/0x18d0
[346278.191068] kernel_clone+0x9d/0x3d0
[346278.191072] __do_sys_clone+0x5d/0x80
[346278.191076] __x64_sys_clone+0x25/0x30
[346278.191079] do_syscall_64+0x5c/0xc0
[346278.191083] ? syscall_exit_to_user_mode+0x27/0x50
[346278.191086] ? do_syscall_64+0x69/0xc0
[346278.191088] ? irqentry_exit_to_user_mode+0x9/0x20
[346278.191092] ? irqentry_exit+0x19/0x30
[346278.191095] ? exc_page_fault+0x89/0x160
[346278.191097] ? asm_exc_page_fault+0x8/0x30
[346278.191102] entry_SYSCALL_64_after_hwframe+0x44/0xae

The task was waiting for the refcount become to 1, but from the vmcore,
we found the refcount has already been 1. It seems that the task didn't
get woken up by perf_event_release_kernel() and got stuck forever. The
below scenario may cause the problem.

Thread A					Thread B
...						...
perf_event_free_task				perf_event_release_kernel
						   ...
						   acquire event->child_mutex
						   ...
						   get_ctx
   ...						   release event->child_mutex
   acquire ctx->mutex
   ...
   perf_free_event (acquire/release event->child_mutex)
   ...
   release ctx->mutex
   wait_var_event
						   acquire ctx->mutex
						   acquire event->child_mutex
						   # move existing events to free_list
						   release event->child_mutex
						   release ctx->mutex
						   put_ctx
...						...

In this case, all events of the ctx have been freed, so we couldn't
find the ctx in free_list and Thread A will miss the wakeup. It's thus
necessary to add a wakeup after dropping the reference.

Fixes: 1cf8dfe8a661 ("perf/core: Fix race between close() and fork()")
Signed-off-by: Haifeng Xu <haifeng.xu@shopee.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/20240513103948.33570-1-haifeng.xu@shopee.com
---
 kernel/events/core.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)

diff --git a/kernel/events/core.c b/kernel/events/core.c
index f0128c5..8f908f0 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -5384,6 +5384,7 @@ int perf_event_release_kernel(struct perf_event *event)
 again:
 	mutex_lock(&event->child_mutex);
 	list_for_each_entry(child, &event->child_list, child_list) {
+		void *var = NULL;
 
 		/*
 		 * Cannot change, child events are not migrated, see the
@@ -5424,11 +5425,23 @@ again:
 			 * this can't be the last reference.
 			 */
 			put_event(event);
+		} else {
+			var = &ctx->refcount;
 		}
 
 		mutex_unlock(&event->child_mutex);
 		mutex_unlock(&ctx->mutex);
 		put_ctx(ctx);
+
+		if (var) {
+			/*
+			 * If perf_event_free_task() has deleted all events from the
+			 * ctx while the child_mutex got released above, make sure to
+			 * notify about the preceding put_ctx().
+			 */
+			smp_mb(); /* pairs with wait_var_event() */
+			wake_up_var(var);
+		}
 		goto again;
 	}
 	mutex_unlock(&event->child_mutex);