]> www.infradead.org Git - users/jedix/linux-maple.git/commit
drm/amd/display: Avoid race between dcn35_set_drr() and dc_state_destruct()
authorTobias Jakobi <tjakobi@math.uni-bielefeld.de>
Mon, 2 Sep 2024 09:40:27 +0000 (11:40 +0200)
committerAlex Deucher <alexander.deucher@amd.com>
Tue, 10 Sep 2024 20:31:44 +0000 (16:31 -0400)
commite835d5144f5ef78e4f8828c63e2f0d61144f283a
treef16840c422b2d6fbbcea66d8e1f9d30189273f4e
parenta7aeb03888b92304e2fc7d4d1c242f54a312561b
drm/amd/display: Avoid race between dcn35_set_drr() and dc_state_destruct()

dc_state_destruct() nulls the resource context of the DC state. The pipe
context passed to dcn35_set_drr() is a member of this resource context.

If dc_state_destruct() is called parallel to the IRQ processing (which
calls dcn35_set_drr() at some point), we can end up using already nulled
function callback fields of struct stream_resource.

The logic in dcn35_set_drr() already tries to avoid this, by checking tg
against NULL. But if the nulling happens exactly after the NULL check and
before the next access, then we get a race.

Avoid this by copying tg first to a local variable, and then use this
variable for all the operations. This should work, as long as nobody
frees the resource pool where the timing generators live.

Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/3142
Fixes: 06ad7e164256 ("drm/amd/display: Destroy DC context while keeping DML and DML2")
Signed-off-by: Tobias Jakobi <tjakobi@math.uni-bielefeld.de>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 0607a50c004798a96e62c089a4c34c220179dcb5)
Cc: stable@vger.kernel.org
drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c