From: Nicholas Kazlauskas <[email protected]> [Why] Race condition between notification of driver idle and the command being processed. We could theoretically enter idle between the submission and the wait for idle that occurs after.
[How] Switch the notification to NO_WAIT to avoid the RPTR access. Reviewed-by: Sung joon Kim <[email protected]> Acked-by: Rodrigo Siqueira <[email protected]> Signed-off-by: Nicholas Kazlauskas <[email protected]> --- drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c index 7724dcadecba..dadeaa9c92dd 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c +++ b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c @@ -1209,7 +1209,8 @@ static void dc_dmub_srv_notify_idle(const struct dc *dc, bool allow_idle) } /* NOTE: This does not use the "wake" interface since this is part of the wake path. */ - dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); + /* We also do not perform a wait since DMCUB could enter idle after the notification. */ + dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_NO_WAIT); } static void dc_dmub_srv_exit_low_power_state(const struct dc *dc) -- 2.42.0
