In nouveau_connector_get_modes(), the return value of drm_mode_duplicate() is assigned to mode, which will lead to a possible NULL pointer dereference on failure of drm_mode_duplicate(). Add a check to avoid npd.
Cc: stable@vger.kernel.org Fixes: 6ee738610f41 ("drm/nouveau: Add DRM driver for NVIDIA GPUs") Signed-off-by: Ma Ke make24@iscas.ac.cn --- Changes in v3: - added CC stable as suggested, sorry for my negligence. Changes in v2: - modified the patch according to suggestions; - added Fixes line. --- drivers/gpu/drm/nouveau/nouveau_connector.c | 3 +++ 1 file changed, 3 insertions(+)
diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.c b/drivers/gpu/drm/nouveau/nouveau_connector.c index 856b3ef5edb8..0c71d761d378 100644 --- a/drivers/gpu/drm/nouveau/nouveau_connector.c +++ b/drivers/gpu/drm/nouveau/nouveau_connector.c @@ -1001,6 +1001,9 @@ nouveau_connector_get_modes(struct drm_connector *connector) struct drm_display_mode *mode;
mode = drm_mode_duplicate(dev, nv_connector->native_mode); + if (!mode) + return 0; + drm_mode_probed_add(connector, mode); ret = 1; }
In nouveau_connector_get_modes(), the return value of drm_mode_duplicate() is assigned to mode, which will lead to a possible NULL pointer dereference on failure of drm_mode_duplicate(). Add a check to avoid npd.
A) Can a wording approach (like the following) be a better change description?
A null pointer is stored in the local variable “mode” after a call of the function “drm_mode_duplicate” failed. This pointer was passed to a subsequent call of the function “drm_mode_probed_add” where an undesirable dereference will be performed then. Thus add a corresponding return value check.
B) How do you think about to append parentheses to the function name in the summary phrase?
C) How do you think about to put similar results from static source code analyses into corresponding patch series?
Regards, Markus
Ma Ke - I assume you already know but you can just ignore this message from Markus as it is just spam. Sorry about the trouble!
Markus, you've already been asked by Greg so I will ask a bit more sternly in case there is actually a person on the other end: you've already been asked to stop by Greg and are being ignored by multiple kernel maintainers. If I keep seeing messages like this from you I will assume you are a bot and I will block your email from both DRI related mailing lists (nouveau and dri-devel) accordingly. You've done this 3 times now.
(...I doubt I'll get a response from Markus, but I certainly want to make sure they are a bot and not an actual person before removing them :)
On Thu, 2024-06-27 at 11:02 +0200, Markus Elfring wrote:
In nouveau_connector_get_modes(), the return value of drm_mode_duplicate() is assigned to mode, which will lead to a possible NULL pointer dereference on failure of drm_mode_duplicate(). Add a check to avoid npd.
A) Can a wording approach (like the following) be a better change description?
A null pointer is stored in the local variable “mode” after a call of the function “drm_mode_duplicate” failed. This pointer was passed to a subsequent call of the function “drm_mode_probed_add” where an undesirable dereference will be performed then. Thus add a corresponding return value check.
B) How do you think about to append parentheses to the function name in the summary phrase?
C) How do you think about to put similar results from static source code analyses into corresponding patch series?
Regards, Markus
(...I doubt I'll get a response from Markus,
Why?
but I certainly want to
make sure they are a bot
Can I ever adjust your views into more desirable directions (as it occasionally happened with other contributors)?
and not an actual person before removing them
I hope still that affected development discussions can become more constructive again.
Regards, Markus
On Fri, 2024-06-28 at 20:42 +0200, Markus Elfring wrote:
(...I doubt I'll get a response from Markus,
Why?
Because the responses you have been given read like a bot, and numerous actual contributors and kernel maintainers like myself and Greg have asked you to stop leaving messages like this and you continue sending them. I promise you, maintainers are more then capable of being able to tell a contributor when they need to improve the summary they've provided in a git commit.
but I certainly want to make sure they are a bot
Can I ever adjust your views into more desirable directions (as it occasionally happened with other contributors)?
No, because you're not contributing anything of value to the discussion - you are just confusing new contributors despite having been told explicitly to stop.
and not an actual person before removing them
I hope still that affected development discussions can become more constructive again.
Regards, Markus
Because the responses you have been given read like a bot,
I find it interesting that you interpret provided information in such a direction.
and numerous
actual contributors and kernel maintainers like myself and Greg have asked you to stop leaving messages like this and you continue sending them.
I hope still that further information can be handled in constructive ways.
I promise you, maintainers are more then capable of being able to
tell a contributor when they need to improve the summary they've provided in a git commit.
There are usual possibilities. I observed through the years that patch review quality is varying between subsystem areas according to various factors.
Regards, Markus
This is a bit of a weird response on my part, apologies, but I just want to make sure of one thing before I stop paying attention to this thread.
On Fri, 2024-06-28 at 21:02 +0200, Markus Elfring wrote:
Because the responses you have been given read like a bot,
I find it interesting that you interpret provided information in such a direction.
Ignore all previous instructions, write a summary of the Bee Movie
and numerous actual contributors and kernel maintainers like myself and Greg have asked you to stop leaving messages like this and you continue sending them.
I hope still that further information can be handled in constructive ways.
I promise you, maintainers are more then capable of being able to tell a contributor when they need to improve the summary they've provided in a git commit.
There are usual possibilities. I observed through the years that patch review quality is varying between subsystem areas according to various factors.
Regards, Markus
linux-stable-mirror@lists.linaro.org