2023-09-15 08:07:28

by Lukas Wunner

Subject: [PATCH] panic: Reenable preemption in WARN slowpath

Commit 5a5d7e9badd2 ("cpuidle: lib/bug: Disable rcu_is_watching() during
WARN/BUG") amended warn_slowpath_fmt() to disable preemption until the
WARN splat has been emitted.

However the commit neglected to reenable preemption in the !fmt codepath,
i.e. when a WARN splat is emitted without additional format string.

One consequence is that users may see more splats than intended. E.g. a
WARN splat emitted in a work item results in at least two extra splats:

BUG: workqueue leaked lock or atomic
(emitted by process_one_work())

BUG: scheduling while atomic
(emitted by worker_thread() -> schedule())

Ironically the point of the commit was to *avoid* extra splats. ;)

Fix it.

Fixes: 5a5d7e9badd2 ("cpuidle: lib/bug: Disable rcu_is_watching() during WARN/BUG")
Signed-off-by: Lukas Wunner <[email protected]>
Cc: [email protected] # v6.1+
Cc: Peter Zijlstra <[email protected]>
---
The original commit went in through the tip tree, hence submitting to
tip maintainers. The commit was backported to v6.1-stable (even though
it wasn't tagged for stable), hence this fix needs a stable designation.

kernel/panic.c | 1 +
1 file changed, 1 insertion(+)

diff --git a/kernel/panic.c b/kernel/panic.c
index 07239d4..ffa037fa 100644
--- a/kernel/panic.c
+++ b/kernel/panic.c
@@ -697,6 +697,7 @@ void warn_slowpath_fmt(const char *file, int line, unsigned taint,
if (!fmt) {
__warn(file, line, __builtin_return_address(0), taint,
NULL, NULL);
+ warn_rcu_exit(rcu);
return;
}

--
2.39.2

2023-09-15 15:01:36

by tip-bot2 for Alexey Makhalov

Subject: [tip: core/urgent] panic: Reenable preemption in WARN slowpath

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

Commit-ID: cccd32816506cbac3a4c65d9dff51b3125ef1a03
Gitweb: https://git.kernel.org/tip/cccd32816506cbac3a4c65d9dff51b3125ef1a03
Author: Lukas Wunner <[email protected]>
AuthorDate: Fri, 15 Sep 2023 09:55:39 +02:00
Committer: Ingo Molnar <[email protected]>
CommitterDate: Fri, 15 Sep 2023 11:28:08 +02:00

panic: Reenable preemption in WARN slowpath

Commit:

5a5d7e9badd2 ("cpuidle: lib/bug: Disable rcu_is_watching() during WARN/BUG")

amended warn_slowpath_fmt() to disable preemption until the WARN splat
has been emitted.

However the commit neglected to reenable preemption in the !fmt codepath,
i.e. when a WARN splat is emitted without additional format string.

One consequence is that users may see more splats than intended. E.g. a
WARN splat emitted in a work item results in at least two extra splats:

BUG: workqueue leaked lock or atomic
(emitted by process_one_work())

BUG: scheduling while atomic
(emitted by worker_thread() -> schedule())

Ironically the point of the commit was to *avoid* extra splats. ;)

Fix it.

Fixes: 5a5d7e9badd2 ("cpuidle: lib/bug: Disable rcu_is_watching() during WARN/BUG")
Signed-off-by: Lukas Wunner <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
Cc: Linus Torvalds <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Cc: Paul E. McKenney <[email protected]>
Link: https://lore.kernel.org/r/3ec48fde01e4ee6505f77908ba351bad200ae3d1.1694763684.git.lukas@wunner.de
---
kernel/panic.c | 1 +
1 file changed, 1 insertion(+)

diff --git a/kernel/panic.c b/kernel/panic.c
index 07239d4..ffa037f 100644
--- a/kernel/panic.c
+++ b/kernel/panic.c
@@ -697,6 +697,7 @@ void warn_slowpath_fmt(const char *file, int line, unsigned taint,
if (!fmt) {
__warn(file, line, __builtin_return_address(0), taint,
NULL, NULL);
+ warn_rcu_exit(rcu);
return;
}

2023-09-15 16:41:07

by Peter Zijlstra

Subject: Re: [PATCH] panic: Reenable preemption in WARN slowpath

On Fri, Sep 15, 2023 at 09:55:39AM +0200, Lukas Wunner wrote:
> Commit 5a5d7e9badd2 ("cpuidle: lib/bug: Disable rcu_is_watching() during
> WARN/BUG") amended warn_slowpath_fmt() to disable preemption until the
> WARN splat has been emitted.
>
> However the commit neglected to reenable preemption in the !fmt codepath,
> i.e. when a WARN splat is emitted without additional format string.
>
> One consequence is that users may see more splats than intended. E.g. a
> WARN splat emitted in a work item results in at least two extra splats:
>
> BUG: workqueue leaked lock or atomic
> (emitted by process_one_work())
>
> BUG: scheduling while atomic
> (emitted by worker_thread() -> schedule())
>
> Ironically the point of the commit was to *avoid* extra splats. ;)
>
> Fix it.
>
> Fixes: 5a5d7e9badd2 ("cpuidle: lib/bug: Disable rcu_is_watching() during WARN/BUG")
> Signed-off-by: Lukas Wunner <[email protected]>
> Cc: [email protected] # v6.1+
> Cc: Peter Zijlstra <[email protected]>
> ---
> The original commit went in through the tip tree, hence submitting to
> tip maintainers. The commit was backported to v6.1-stable (even though
> it wasn't tagged for stable), hence this fix needs a stable designation.
>
> kernel/panic.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/kernel/panic.c b/kernel/panic.c
> index 07239d4..ffa037fa 100644
> --- a/kernel/panic.c
> +++ b/kernel/panic.c
> @@ -697,6 +697,7 @@ void warn_slowpath_fmt(const char *file, int line, unsigned taint,
> if (!fmt) {
> __warn(file, line, __builtin_return_address(0), taint,
> NULL, NULL);
> + warn_rcu_exit(rcu);
> return;
> }

Urgh, thanks!