[<prev] [next>] [day] [month] [year] [list]
Message-Id: <20231216104350.12375-1-tintinm2017@gmail.com>
Date: Sat, 16 Dec 2023 16:13:51 +0530
From: attreyee-muk <tintinm2017@...il.com>
To: maarten.lankhorst@...ux.intel.com,
mripard@...nel.org,
tzimmermann@...e.de,
airlied@...il.com,
daniel@...ll.ch
Cc: attreyee-muk <tintinm2017@...il.com>,
dri-devel@...ts.freedesktop.org,
linux-kernel@...r.kernel.org
Subject: [PATCH] drm/drm_gem.c: Fixed WARNING in drm_gem_object_handle_put_unlocked
Previously, I was sending out the patch in a wrong format, hence sending
this patch again describing my changes.
Looking at the code and the call trace reported by Syzbot (issue link : https://syzkaller.appspot.com/bug?extid=ef3256a360c02207a4cb), I understood
that the issue lies in the drm_gem.c file in
drm_gem_object_handle_put_unlocked(obj);.
As per my understanding, I believe that a warning is given out when the DRM GEM object handle gets
destroyed along with the function getting returned immediately.
To take care of the warning only, I think that the WARN_ON can be removed with
only the function getting returned when the object handle gets
destroyed, without a warning.
Or else, we can even have a message logging in the place of a direct warning which will give out the
information that the object handle has been destroyed.
I also think that the WARN_ON is given for debugging purposes and hence proposing a
solution if the warning can be replaced with a simple message log.
I am looking forward to getting help in this regard from the respected maintainers.
Thank You
Attreyee Mukherjee
Signed-off-by: Attreyee Mukherjee <tintinm2017@...il.com>
---
drivers/gpu/drm/drm_gem.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c
index 44a948b80ee1..b6be9400cd11 100644
--- a/drivers/gpu/drm/drm_gem.c
+++ b/drivers/gpu/drm/drm_gem.c
@@ -222,7 +222,7 @@ drm_gem_object_handle_put_unlocked(struct drm_gem_object *obj)
struct drm_device *dev = obj->dev;
bool final = false;
- if (WARN_ON(READ_ONCE(obj->handle_count) == 0))
+ if (READ_ONCE(obj->handle_count) == 0)
return;
/*
--
2.34.1
Powered by blists - more mailing lists