From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-alma10-1.taild15c8.ts.net [100.103.45.18]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 0A0643A1D01; Sat, 17 Oct 2026 17:25:26 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=100.103.45.18 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1792257927; cv=none; b=gxsWfBQaI54KS377x2b3oyvgtceX3T6Ut1sk+ryU/kDzZbTPE+dmq72IIP/jELuRjbV4asCecgcUmsGx3iouYQH8tJ38TsVG9wdnnjLa7KGOqZ2BdXXLTF+KMn7BJNRHVjIy8/CVFdP+o9eLRAHNHr/ikCl0mthQRNT0S1zeNIM= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1792257927; c=relaxed/simple; bh=fot4ruM0UJvpVMDThG6X+/ApWaZpFa1071D16Bgb/Uk=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=fbfbf08cuUK1hIFbQhqzOT9Uih+AbYpNsfy0Z/yzaKJlYZnEQnljtLCjqwHoy4B/EtPIvfM3ljx289WGuGZWFrnlDKMBq/+zwTRrNJbilXJV6KYQeG1q998jjgYmwrooWx712W0Bw/oy777KEVZPpczrriBw9FNLCOJrYVZULzk= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=linuxfoundation.org header.i=@linuxfoundation.org header.b=TM0VRMNd; arc=none smtp.client-ip=100.103.45.18 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=linuxfoundation.org header.i=@linuxfoundation.org header.b="TM0VRMNd" Received: by smtp.kernel.org (Postfix) with ESMTPSA id 4B1D41F000C4; Sat, 17 Oct 2026 17:25:25 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linuxfoundation.org; s=korg; t=1792257925; bh=BKBTu+viAsrO+MU92pgru/zUDwjONVcZn6HdqdudW18=; h=From:To:Cc:Subject:Date:In-Reply-To:References; b=TM0VRMNdNud4UW3imzdL8W81Z339A3kqZYpUEzkuutZaz/m3I1V+xsBDLyK1oJUOU OVBpPlCeI+edMO/Loibxccmb2yptlcN8yW1jxWgPbPNha6aPd5Hb5BMKrIO2UtzLrN RuMQ/5G2fBLLbIkFxabyhiUeD29JiKWE834PO1eU= From: Greg Kroah-Hartman To: stable@vger.kernel.org Cc: Greg Kroah-Hartman , patches@lists.linux.dev, "Yi Lai" , syzbot+3c4321e10eea460eb606@syzkaller.appspotmail.com, Frederic Weisbecker , "Peter Zijlstra (Intel)" , Sasha Levin Subject: [PATCH 6.6 0716/1674] perf: Fix hang while freeing sigtrap event Date: Sat, 17 Oct 2026 17:49:48 +0200 Message-ID: <20261017160449.486404776@linuxfoundation.org> X-Mailer: git-send-email 2.56.0 In-Reply-To: <20261017160431.119607952@linuxfoundation.org> References: <20261017160431.119607952@linuxfoundation.org> User-Agent: quilt/0.69 X-stable: review X-Patchwork-Hint: ignore Precedence: bulk X-Mailing-List: patches@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: 8bit 6.6-stable review patch. If anyone has any objections, please let me know. ------------------ From: Frederic Weisbecker [ Upstream commit 56799bc035658738f362acec3e7647bb84e68933 ] Perf can hang while freeing a sigtrap event if a related deferred signal hadn't managed to be sent before the file got closed: perf_event_overflow() task_work_add(perf_pending_task) fput() task_work_add(____fput()) task_work_run() ____fput() perf_release() perf_event_release_kernel() _free_event() perf_pending_task_sync() task_work_cancel() -> FAILED rcuwait_wait_event() Once task_work_run() is running, the list of pending callbacks is removed from the task_struct and from this point on task_work_cancel() can't remove any pending and not yet started work items, hence the task_work_cancel() failure and the hang on rcuwait_wait_event(). Task work could be changed to remove one work at a time, so a work running on the current task can always cancel a pending one, however the wait / wake design is still subject to inverted dependencies when remote targets are involved, as pictured by Oleg: T1 T2 fd = perf_event_open(pid => T2->pid); fd = perf_event_open(pid => T1->pid); close(fd) close(fd) perf_event_overflow() perf_event_overflow() task_work_add(perf_pending_task) task_work_add(perf_pending_task) fput() fput() task_work_add(____fput()) task_work_add(____fput()) task_work_run() task_work_run() ____fput() ____fput() perf_release() perf_release() perf_event_release_kernel() perf_event_release_kernel() _free_event() _free_event() perf_pending_task_sync() perf_pending_task_sync() rcuwait_wait_event() rcuwait_wait_event() Therefore the only option left is to acquire the event reference count upon queueing the perf task work and release it from the task work, just like it was done before 3a5465418f5f ("perf: Fix event leak upon exec and file release") but without the leaks it fixed. Some adjustments are necessary to make it work: * A child event might dereference its parent upon freeing. Care must be taken to release the parent last. * Some places assuming the event doesn't have any reference held and therefore can be freed right away must instead put the reference and let the reference counting to its job. Reported-by: "Yi Lai" Closes: https://lore.kernel.org/all/Zx9Losv4YcJowaP%2F@ly-workstation/ Reported-by: syzbot+3c4321e10eea460eb606@syzkaller.appspotmail.com Closes: https://lore.kernel.org/all/673adf75.050a0220.87769.0024.GAE@google.com/ Fixes: 3a5465418f5f ("perf: Fix event leak upon exec and file release") Signed-off-by: Frederic Weisbecker Signed-off-by: Peter Zijlstra (Intel) Link: https://lkml.kernel.org/r/20250304135446.18905-1-frederic@kernel.org Signed-off-by: Sasha Levin --- include/linux/perf_event.h | 1 - kernel/events/core.c | 60 ++++++++++++-------------------------- 2 files changed, 18 insertions(+), 43 deletions(-) diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 58746c090859b..72642a021964e 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -786,7 +786,6 @@ struct perf_event { struct irq_work pending_irq; struct callback_head pending_task; unsigned int pending_work; - struct rcuwait pending_work_wait; atomic_t event_limit; diff --git a/kernel/events/core.c b/kernel/events/core.c index 3169742fd98ad..ad0e04e3e5777 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -2312,6 +2312,7 @@ event_sched_out(struct perf_event *event, struct perf_event_context *ctx) !event->pending_work && !task_work_add(current, &event->pending_task, TWA_RESUME)) { event->pending_work = 1; + WARN_ON_ONCE(!atomic_long_inc_not_zero(&event->refcount)); } else { local_dec(&event->ctx->nr_pending); } @@ -5226,35 +5227,9 @@ static bool exclusive_event_installable(struct perf_event *event, static void perf_addr_filters_splice(struct perf_event *event, struct list_head *head); -static void perf_pending_task_sync(struct perf_event *event) -{ - struct callback_head *head = &event->pending_task; - - if (!event->pending_work) - return; - /* - * If the task is queued to the current task's queue, we - * obviously can't wait for it to complete. Simply cancel it. - */ - if (task_work_cancel(current, head)) { - event->pending_work = 0; - local_dec(&event->ctx->nr_pending); - return; - } - - /* - * All accesses related to the event are within the same - * non-preemptible section in perf_pending_task(). The RCU - * grace period before the event is freed will make sure all - * those accesses are complete by then. - */ - rcuwait_wait_event(&event->pending_work_wait, !event->pending_work, TASK_UNINTERRUPTIBLE); -} - static void _free_event(struct perf_event *event) { irq_work_sync(&event->pending_irq); - perf_pending_task_sync(event); unaccount_event(event); @@ -5379,10 +5354,17 @@ static void perf_remove_from_owner(struct perf_event *event) static void put_event(struct perf_event *event) { + struct perf_event *parent; + if (!atomic_long_dec_and_test(&event->refcount)) return; + parent = event->parent; _free_event(event); + + /* Matches the refcount bump in inherit_event() */ + if (parent) + put_event(parent); } /* @@ -5466,11 +5448,6 @@ int perf_event_release_kernel(struct perf_event *event) if (tmp == child) { perf_remove_from_context(child, DETACH_GROUP); list_move(&child->child_list, &free_list); - /* - * This matches the refcount bump in inherit_event(); - * this can't be the last reference. - */ - put_event(event); } else { var = &ctx->refcount; } @@ -5496,7 +5473,8 @@ int perf_event_release_kernel(struct perf_event *event) void *var = &child->ctx->refcount; list_del(&child->child_list); - free_event(child); + /* Last reference unless ->pending_task work is pending */ + put_event(child); /* * Wake any perf_event_free_task() waiting for this event to be @@ -5507,7 +5485,11 @@ int perf_event_release_kernel(struct perf_event *event) } no_ctx: - put_event(event); /* Must be the 'last' reference */ + /* + * Last reference unless ->pending_task work is pending on this event + * or any of its children. + */ + put_event(event); return 0; } EXPORT_SYMBOL_GPL(perf_event_release_kernel); @@ -6912,11 +6894,6 @@ static void perf_pending_task(struct callback_head *head) struct perf_event *event = container_of(head, struct perf_event, pending_task); int rctx; - /* - * All accesses to the event must belong to the same implicit RCU read-side - * critical section as the ->pending_work reset. See comment in - * perf_pending_task_sync(). - */ preempt_disable_notrace(); /* * If we 'fail' here, that's OK, it means recursion is already disabled @@ -6928,12 +6905,13 @@ static void perf_pending_task(struct callback_head *head) event->pending_work = 0; perf_sigtrap(event); local_dec(&event->ctx->nr_pending); - rcuwait_wake_up(&event->pending_work_wait); } if (rctx >= 0) perf_swevent_put_recursion_context(rctx); preempt_enable_notrace(); + + put_event(event); } #ifdef CONFIG_GUEST_PERF_EVENTS @@ -12133,7 +12111,6 @@ perf_event_alloc(struct perf_event_attr *attr, int cpu, init_waitqueue_head(&event->waitq); init_irq_work(&event->pending_irq, perf_pending_irq); init_task_work(&event->pending_task, perf_pending_task); - rcuwait_init(&event->pending_work_wait); mutex_init(&event->mmap_mutex); raw_spin_lock_init(&event->addr_filters.lock); @@ -13292,8 +13269,7 @@ perf_event_exit_event(struct perf_event *event, * Kick perf_poll() for is_event_hup(); */ perf_event_wakeup(parent_event); - free_event(event); - put_event(parent_event); + put_event(event); return; } -- 2.53.0