Re: [PATCH] drm/amd/display: Fix stale replay_events after mod_power stream removal

[Leo Li <sunpeng.li@amd.com>]

On 2026-09-17 06:42, Simon Polack wrote:
> [Some people who received this message don't often get email from spolack+git@mailbox.org. Learn why this is important at https://aka.ms/LearnAboutSenderIdentification ]
> 
> [Why]
> mod_power_remove_stream() shifts the remaining power_entity slots down
> but does not move replay_events, and mod_power_add_stream() does not
> initialize it. replay_events therefore stay bound to the map slot
> instead of the stream.
> 
> When several streams are disabled in one atomic commit,
> amdgpu_dm_mod_power_update_streams() removes them one after another.
> The eDP stream can then be looked up in a slot whose stale
> replay_events already have replay_event_hw_programming set, so
> amdgpu_dm_replay_set_event() returns early ("already in desired state")
> without calling mod_power_set_replay_event(). Replay is not disabled
> before the eDP panel is powered off. After DPMS on, the sink reports
> neither replay state nor frame lock (DPCD 0x378 = 0x00, no error bits),
> so the HPD IRQ recovery does not trigger and the panel stays black
> until a full modeset.
> 
> Seen with an eDP panel using FreeSync Replay plus two DP-MST displays:
> DPMS off/on of all outputs leaves eDP black, while DPMS of eDP alone
> works. Doing an eDP-only DPMS first makes the next all-output DPMS
> fail reliably.
> 
> [How]
> Shift replay_events together with the PSR cached fields in
> mod_power_remove_stream() and initialize it to replay_event_vsync in
> mod_power_add_stream(), matching the psr_event_vsync initial value used
> for PSR (both vsync events are driven together by
> amdgpu_dm_crtc_set_static_screen_optimze()).
> 
> Tested on 7.3.0-rc3 (238650ef6c7c): the reproducer above now recovers
> reliably, and Replay still engages when the screen is idle.
> 
> The issue was debugged with help from an AI assistant (Claude), which
> analysed ftrace/kprobe traces and the driver source, pointed to the
> missing replay_events handling and suggested this change. I collected
> the traces and built and tested the fix on the affected hardware.
> 
> Fixes: 4cef2ac4c795 ("drm/amd/display: Introduce power module on Linux")
> Assisted-by: Claude:claude-opus-5
> Signed-off-by: Simon Polack <spolack+git@mailbox.org>

Reviewed-by: Leo Li <sunpeng.li@amd.com>

Thanks!
> ---
> Based on v7.3-rc3 (238650ef6c7c), where it was tested; also applies to
> amd-staging-drm-next.
> 
> This does not address why stale replay_events can reach the eDP slot
> with replay_event_hw_programming already set in the first place; with
> per-stream bookkeeping that state is consistent again. Traces
> (kprobes/fprobes on the Replay path, eDP-only vs. all-output DPMS) and
> DPCD dumps are available on request.
> 
>  drivers/gpu/drm/amd/display/modules/power/power.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/drivers/gpu/drm/amd/display/modules/power/power.c b/drivers/gpu/drm/amd/display/modules/power/power.c
> index 2f9690e65ca9..d900be8cb5fc 100644
> --- a/drivers/gpu/drm/amd/display/modules/power/power.c
> +++ b/drivers/gpu/drm/amd/display/modules/power/power.c
> @@ -328,6 +328,7 @@ bool mod_power_add_stream(struct mod_power *mod_power,
>                 core_power->map[core_power->num_entities].psr_enabled = 0;
>                 core_power->map[core_power->num_entities].psr_events = psr_event_vsync;
>                 core_power->map[core_power->num_entities].psr_power_opt = 0;
> +               core_power->map[core_power->num_entities].replay_events = replay_event_vsync;
>                 core_power->num_entities++;
>                 return true;
>         }
> @@ -387,6 +388,7 @@ bool mod_power_remove_stream(struct mod_power *mod_power,
>                 core_power->map[i].psr_enabled = core_power->map[i + 1].psr_enabled;
>                 core_power->map[i].psr_events = core_power->map[i + 1].psr_events;
>                 core_power->map[i].psr_power_opt = core_power->map[i + 1].psr_power_opt;
> +               core_power->map[i].replay_events = core_power->map[i + 1].replay_events;
> 
>                 memcpy(core_power->map[i].psr_context, core_power->map[i + 1].psr_context, sizeof(struct mod_power_psr_context));
>                 memset(core_power->map[i + 1].psr_context, 0, sizeof(struct mod_power_psr_context));
> 
> base-commit: 238650ef6c7c7cca08e032527329424c9fbd70e5
> --
> 2.55.0