linux-kernel - Re: [REGRESSION] Black screen on HDMI power-cycle after commit 3471b9a31ce3 (7900XTX + LG C3)

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <CABXGCsOAExctLKVc-ATBgVXyU5UYujg+KrhwyLXX3cTdjxRDww@mail.gmail.com>
Date: Wed, 10 Dec 2025 13:49:07 +0500
From: Mikhail Gavrilov <mikhail.v.gavrilov@...il.com>
To: "Pillai, Aurabindo" <Aurabindo.Pillai@....com>
Cc: "Deucher, Alexander" <Alexander.Deucher@....com>, amd-gfx list <amd-gfx@...ts.freedesktop.org>, 
	Linux List Kernel Mailing <linux-kernel@...r.kernel.org>, 
	"Vojvodic, Relja (Reggie)" <Relja.Vojvodic@....com>, "Liu, Wenjing" <Wenjing.Liu@....com>
Subject: Re: [REGRESSION] Black screen on HDMI power-cycle after commit
 3471b9a31ce3 (7900XTX + LG C3)

On Tue, Dec 9, 2025 at 8:33 PM Pillai, Aurabindo
<Aurabindo.Pillai@....com> wrote:
>
> [AMD Official Use Only - AMD Internal Distribution Only]
>
> I see some errors unrelated to display, like:
>
> [  361.423085] amdgpu 0000:03:00.0: amdgpu: ring_buffer_start = 00000000bdad4a86; ring_buffer_end = 000000000059d4d9; write_frame = 00000000965e67c9
> [  361.423099] amdgpu 0000:03:00.0: amdgpu: write_frame is pointing to address out of bounds
> [  361.423188] amdgpu 0000:03:00.0: amdgpu: ring_buffer_start = 00000000bdad4a86; ring_buffer_end = 000000000059d4d9; write_frame = 00000000965e67c9
> [  361.423190] amdgpu 0000:03:00.0: amdgpu: write_frame is pointing to address out of bounds
>
> and
>
> [  361.519801] amdgpu 0000:03:00.0: amdgpu: device lost from bus!
> [  361.519814] amdgpu 0000:03:00.0: amdgpu: SMU: response:0xFFFFFFFF for index:36 param:0x00000001 message:SetWorkloadMask?
> [  361.519818] amdgpu 0000:03:00.0: amdgpu: Failed to set workload mask 0x00000001
> [  361.519826] amdgpu 0000:03:00.0: amdgpu: (-121) failed to disable fullscreen 3D power profile mode
>
> Do you have these errors with the patch reverted?
>
> Since the follow up patch did not work, please try partially reverting the culprit patch (the two hunks) to help us pinpoint which one fixes the issue for you?
>

Hi Aurabindo,

I have performed the partial revert tests as requested. Here are the results:

Reverting ONLY link_detection.c (read hunk): The issue persists. I
still get "No Signal" after replugging the HDMI cable. (Log attached:
dmesg-6.18.0-cb015814f8b6-hunk-1-read.zip)

Reverting ONLY link_ddc.c (write hunk): The issue is FIXED. The
display works correctly and wakes up immediately after replugging the
cable. (Log attached: dmesg-6.18.0-cb015814f8b6-hunk-2-write.zip)

Conclusion: It seems the regression is specifically caused by the
changes in drivers/gpu/drm/amd/display/dc/link/protocols/link_ddc.c
(the write_scdc_data function). The change in link_detection.c appears
to be unrelated to the failure on my setup.

-- 
Best Regards,
Mike Gavrilov.

Download attachment "dmesg-6.18.0-cb015814f8b6-hunk-1-read.zip" of type "application/zip" (29894 bytes)

Download attachment "dmesg-6.18.0-cb015814f8b6-hunk-2-write.zip" of type "application/zip" (63888 bytes)