regression/bisected/6.8 commit dfc03588cf8ce2af8ef810cd226dc98ee4fbac38 leads to slab-out-of-bounds in dm_resume+0x2068/0x24f0 [amdgpu]

[Mikhail Gavrilov]

Hi,
I already fill the bug report about slab-out-of-bounds in
dm_resume+0x208c/0x2600. It usually happens when the system resumes
from a suspended state.
But it was unclear if it was regression or not.
Today I decided to bisect it.
And first bad commit is:
commit dfc03588cf8ce2af8ef810cd226dc98ee4fbac38
Author: Alex Hung <alex.hung@xxxxxxx>
Date:   Fri Dec 1 06:25:24 2023 -0700

    drm/amd/display: Initialize writeback connector

    [WHAT]
    Create a drm_writeback_connector when connector signal equals
    SIGNAL_TYPE_VIRTUAL.

    Tested-by: Daniel Wheeler <daniel.wheeler@xxxxxxx>
    Reviewed-by: Harry Wentland <harry.wentland@xxxxxxx>
    Signed-off-by: Alex Hung <alex.hung@xxxxxxx>
    Signed-off-by: Alex Deucher <alexander.deucher@xxxxxxx>

 drivers/gpu/drm/amd/display/amdgpu_dm/Makefile     |   3 +-
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c  |  20 +-
 .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_wb.c   | 209 +++++++++++++++++++++
 .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_wb.h   |  35 ++++
 4 files changed, 265 insertions(+), 2 deletions(-)

So this is definitely a regression in the 6.8 release cycle.
But unfortunately I can't easily revert the commit for double checking.

Bisect log and all kernel logs from each step I attached here.
Also attached build config.

Alex, can you look into it?

[1] https://gitlab.freedesktop.org/drm/amd/-/issues/3199

-- 
Best Regards,
Mike Gavrilov.

Attachment: kernel-logs.zip
Description: Zip archive

Attachment: git-bisect-log-slab-out-of-bounds-in-dm_resume+0x2068.zip
Description: Zip archive

Attachment: .config.zip
Description: Zip archive