wait: don't wake up freezable wait by freezing fake signal

Message ID 20230822074547.8037-1-ming.qian@nxp.com
State New
Headers
Series wait: don't wake up freezable wait by freezing fake signal |

Commit Message

Ming Qian Aug. 22, 2023, 7:45 a.m. UTC
  kernel may try to wake up task with fake signal when freezing task,
if the task is waiting event using wait_event_freezable,
it's better to freeze the wait, instead of breaking it.

otherwise the caller may need to retry the wait,
maybe like below code:

	if (rc == -ERESTARTSYS && freezing(current)) {
		clear_thread_flag(TIF_SIGPENDING);
		goto again;
	}

Signed-off-by: TaoJiang <tao.jiang_2@nxp.com>
Signed-off-by: Ming Qian <ming.qian@nxp.com>
---
 include/linux/wait.h |  9 +++++++--
 kernel/sched/wait.c  | 14 ++++++++++++++
 2 files changed, 21 insertions(+), 2 deletions(-)
  

Comments

Ming Qian Sept. 26, 2023, 2:43 a.m. UTC | #1
>From: Peter Zijlstra <peterz@infradead.org>
>Sent: 2023年9月25日 18:08
>To: Ming Qian <ming.qian@nxp.com>
>Cc: mingo@redhat.com; juri.lelli@redhat.com; vincent.guittot@linaro.org;
>dietmar.eggemann@arm.com; rostedt@goodmis.org; bsegall@google.com;
>mgorman@suse.de; bristot@redhat.com; vschneid@redhat.com; Eagle Zhou
><eagle.zhou@nxp.com>; Tao Jiang <tao.jiang_2@nxp.com>; dl-linux-imx
><linux-imx@nxp.com>; linux-kernel@vger.kernel.org; linux-arm-
>kernel@lists.infradead.org
>Subject: [EXT] Re: [PATCH] wait: don't wake up freezable wait by freezing fake
>signal
>
>Caution: This is an external email. Please take care when clicking links or
>opening attachments. When in doubt, report the message using the 'Report
>this email' button
>
>
>On Tue, Aug 22, 2023 at 03:45:47PM +0800, Ming Qian wrote:
>> kernel may try to wake up task with fake signal when freezing task, if
>> the task is waiting event using wait_event_freezable, it's better to
>> freeze the wait, instead of breaking it.
>>
>> otherwise the caller may need to retry the wait, maybe like below
>> code:
>>
>>       if (rc == -ERESTARTSYS && freezing(current)) {
>>               clear_thread_flag(TIF_SIGPENDING);
>>               goto again;
>>       }
>
>I'm not sure I get it -- is there an actual problem, or are you trying to optimize
>something?

Hi Peter,
    We have a driver that needs to interact with the daemon on nxp IMX8MP-EVK,  Initially we use wait_event_interruptible to wait for a response from the daemon, but in our suspend test, sometimes we met failure which was caused by the waiting interrupted.
    Then we use wait_event_freezable instead, but the problem remains.  The above retry  code should fix the problem, but we think it may be a common issue.
    So we try to improve the wait_event_freezable, then we can pass our suspend test without any rety.

Ming
  

Patch

diff --git a/include/linux/wait.h b/include/linux/wait.h
index 5ec7739400f4..975ee67f259a 100644
--- a/include/linux/wait.h
+++ b/include/linux/wait.h
@@ -287,6 +287,7 @@  static inline void wake_up_pollfree(struct wait_queue_head *wq_head)
 	 (state & (TASK_INTERRUPTIBLE | TASK_WAKEKILL)))
 
 extern void init_wait_entry(struct wait_queue_entry *wq_entry, int flags);
+bool freezing_wait(struct wait_queue_entry *wq_entry, int state);
 
 /*
  * The below macro ___wait_event() has an explicit shadow of the __ret
@@ -314,8 +315,12 @@  extern void init_wait_entry(struct wait_queue_entry *wq_entry, int flags);
 			break;							\
 										\
 		if (___wait_is_interruptible(state) && __int) {			\
-			__ret = __int;						\
-			goto __out;						\
+			if (unlikely(freezing_wait(&__wq_entry, state))) {	\
+				clear_thread_flag(TIF_SIGPENDING);		\
+			} else {						\
+				__ret = __int;					\
+				goto __out;					\
+			}							\
 		}								\
 										\
 		cmd;								\
diff --git a/kernel/sched/wait.c b/kernel/sched/wait.c
index 802d98cf2de3..37c3b2bc1fc9 100644
--- a/kernel/sched/wait.c
+++ b/kernel/sched/wait.c
@@ -5,6 +5,8 @@ 
  * (C) 2004 Nadia Yvette Chambers, Oracle
  */
 
+#include <linux/freezer.h>
+
 void __init_waitqueue_head(struct wait_queue_head *wq_head, const char *name, struct lock_class_key *key)
 {
 	spin_lock_init(&wq_head->lock);
@@ -484,3 +486,15 @@  int woken_wake_function(struct wait_queue_entry *wq_entry, unsigned mode, int sy
 	return default_wake_function(wq_entry, mode, sync, key);
 }
 EXPORT_SYMBOL(woken_wake_function);
+
+bool freezing_wait(struct wait_queue_entry *wq_entry, int state)
+{
+	if (!(state & TASK_FREEZABLE))
+		return false;
+	if (fatal_signal_pending(wq_entry->private))
+		return false;
+	if (unlikely(freezing(wq_entry->private)))
+		return true;
+	return false;
+}
+EXPORT_SYMBOL(freezing_wait);