Maybe that wasn't you or whatever, but your last patch that did this got
some CVE assigned to it that didn't really make any sense, given this is
just a null deref that'd end up as an oops?
It can only happen if the kzalloc in drm_mode_create fails.
I imagine that the `continue` is not the best cause of action anyway,
it's probably not worth adding some broken connector with a bunch of
missing modes.
It's just going to make things appear more broken and be unexpected to
userspace.
Maybe this is some new thing that we do now I am out of the loop on...
- Joshie 🐸✨
On 6/23/24 10:20 AM, Joshua Ashton wrote:
Are you planning on submitting a bogus CVE for this patch too?
- Joshie 🐸✨
On June 22, 2024 9:22:19 AM GMT+01:00, Ma Ke <make24@xxxxxxxxxxx> wrote:
In amdgpu_connector_add_common_modes(), the return value of drm_cvt_mode()
is assigned to mode, which will lead to a NULL pointer dereference on
failure of drm_cvt_mode(). Add a check to avoid npd.
Signed-off-by: Ma Ke <make24@xxxxxxxxxxx>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c
index 9caba10315a8..6cf946adb6fe 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c
@@ -458,6 +458,8 @@ static void amdgpu_connector_add_common_modes(struct drm_encoder *encoder,
continue;
mode = drm_cvt_mode(dev, common_modes[i].w, common_modes[i].h, 60, false, false, false);
+ if (!mode)
+ continue;
drm_mode_probed_add(connector, mode);
}
}
