linux-kernel - [BUG] gpu: drm: amd: noretry=0 causes failure in amdgpu_device_ip

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>] [day] [month] [year] [list]

Message-ID: <47e950989fe0fc297a2272139d69a5a7c4c98de5.camel@uvos.xyz>
Date:   Wed, 01 Nov 2023 13:35:47 +0100
From:   Carl Klemm <carl@...s.xyz>
To:     christian.koenig@....com, alexander.deucher@....com
Cc:     dri-devel@...ts.freedesktop.org, linux-kernel@...r.kernel.org
Subject: [BUG] gpu: drm: amd: noretry=0 causes failure in
 amdgpu_device_ip_resume on vega10

Hi,

When migrateing from 5.15 to 6.5.9 i noticed that noretry no longer
function on vega10 (Instinct MI25). The device will fail to start:

[   40.080411] amdgpu: fw load failed
[   40.083816] amdgpu: smu firmware loading failed
[   40.088350] amdgpu 0000:83:00.0: amdgpu: amdgpu_device_ip_resume
failed (-22).

I have also repoduced the same issue on 6.1.55
It is also possible that the issue was caused by newer gpu firmware,
instead of the change in kernel. The issue was repduced with the
firmware from linux-firmware-20230804.

for full dmesg see: https://uvos.xyz/noretry.dmesg.log

The same system also contains 2 vega20 and 1 navi21 device that both
work fine with noretry=0. For more information on the system this
problem was encountered on see this rocminfo dump:
https://uvos.xyz/rocminfo.log

Regards,
Carl