[SRU][J][PATCH 0/1] CVE-2025-38361
Cengiz Can
cengiz.can at canonical.com
Fri Sep 11 00:14:46 UTC 2026
https://ubuntu.com/security/CVE-2025-38361
[ Impact ]
In the Linux kernel, the following vulnerability has been resolved:
drm/amd/display: Check dce_hwseq before dereferencing it
[WHAT]
hws was checked for null earlier in dce110_blank_stream, indicating hws can be
null, and should be checked whenever it is used.
(cherry picked from commit 79db43611ff61280b6de58ce1305e0b2ecf675ad)
[ Fix ]
jammy/linux: backported from 79db43611ff6
The fix adds a null check on the dce_hwseq pointer (hws) in
dce110_blank_stream before it is dereferenced, matching the earlier null check
already present in the same function. In this tree the change landed in
dce110_hw_sequencer.c rather than the upstream hwss/dce110/dce110_hwseq.c
path. As this tree lacks link->skip_implict_edp_power_control, the
edp_backlight_control call was unconditional; only the `if (hws)` null guard
was added, preserving the fix's intent.
[ Test Plan ]
Build and boot tested.
[ Where Problems Could Occur ]
This change only touches the DCE 11.0 display code path in the amdgpu driver,
so any regression would be confined to systems with AMD GPUs using that
display hardware generation when blanking a stream (for example during mode
setting, display power-down, or eDP backlight control). A bad fix could skip
required backlight control or blanking work, potentially leaving a panel in an
incorrect state. Systems without AMD graphics, and AMD systems using other
display engine generations, are not affected.
[ Other Info ]
Kybele flow-v11-25-ga27c0fa6. Reference: be93df16/v1
More information about the kernel-team
mailing list