linux-kernel - [PATCH 0/2] drm/amd: fix VRR race condition during IRQ handling

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-next>] [day] [month] [year] [list]

Message-ID: <cover.1725269643.git.tjakobi@math.uni-bielefeld.de>
Date: Mon,  2 Sep 2024 11:40:25 +0200
From: tjakobi@...h.uni-bielefeld.de
To: amd-gfx@...ts.freedesktop.org,
	dri-devel@...ts.freedesktop.org,
	linux-kernel@...r.kernel.org
Cc: Tobias Jakobi <tjakobi@...h.uni-bielefeld.de>
Subject: [PATCH 0/2] drm/amd: fix VRR race condition during IRQ handling

From: Tobias Jakobi <tjakobi@...h.uni-bielefeld.de>

Hello,

this fixes a nasty race condition in the set_drr() callbacks for DCN10
and DCN35 that has existed now since quite some time, see this GitLab
issue for reference.

https://gitlab.freedesktop.org/drm/amd/-/issues/3142

The report just focuses von DCN10, but the same problem also exists in
the DCN35 code.

With best wishes,
Tobias

Tobias Jakobi (2):
  drm/amd/display: Avoid race between dcn10_set_drr() and
    dc_state_destruct()
  drm/amd/display: Avoid race between dcn35_set_drr() and
    dc_state_destruct()

 .../amd/display/dc/hwss/dcn10/dcn10_hwseq.c   | 20 +++++++++++--------
 .../amd/display/dc/hwss/dcn35/dcn35_hwseq.c   | 20 +++++++++++--------
 2 files changed, 24 insertions(+), 16 deletions(-)

-- 
2.44.2