Re: [PATCH v13 3/4] drm/atomic-helper: Re-order bridge chain pre-enable and post-disable

Re: [PATCH v13 3/4] drm/atomic-helper: Re-order bridge chain pre-enable and post-disable

[Aradhya Bhatia]

+rockchip maintainers

Hi Vicente, all,


I went through the drivers and the affected areas in the gru-kevin
chromebook pipeline last week, but nothing has stood out.


Pipeline:

rockchip,display-subsystem / rk3399-vop (Big/Lite) (CRTC) ->
rk3399-edp (Encoder) -> analogix_dp_core (Bridge) ->
sharp,lq123p (edp-panel)

I am unable to debug this further since I do not have the hardware.

I could use some help, especially from folks who understand the hardware
requirements better.


On 11/09/25 09:01, Aradhya Bhatia wrote:
> Hi Vicente,
> 
> Thank you for the bisection and reporting the issue.
> 
> On 10/09/25 16:17, Vicente Bergas wrote:
>> Hi,
>> this patch causes a regression. It has been reported in
>> https://bugzilla.kernel.org/show_bug.cgi?id=220554
>>
>> It affects the gru/kevin platform (arm64,RK3399) with the Panfrost DRM driver.
> 
> I believe the Panfrost DRM driver may only be for the GPU.
> 
> Based on the dts files in arm64/rockchip/, this is the pipeline of the
> gru-kevin setup that I understand.
> 
> 	rk3399-vop (Big/Lite) -> rk3399-edp -> sharp,lq123p (edp-panel)
> 
> The setup seems to be using the drm/rockchip drivers for the display
> controller and for the bridge.
> 
>>
>> When it boots in console mode, the blinking of the cursor keeps the display on.
>> If it is turned off via /sys/class/graphics/fbcon/cursor_blink, then
>> the display briefly shows each key press presented on screen for less
>> than one second and then powers off.
>>
>> When starting the graphical mode (wayland), if there are no
>> applications drawing on the screen, the only way to keep the display
>> on is by continuously moving the mouse.
>>
> 
> Okay!
> 
> I will have a look through the drivers. In the meanwhile, please do
> report back if you find any other observations.
> 
> 

-- 
Regards
Aradhya


_______________________________________________
Linux-rockchip mailing list
Linux-rockchip@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-rockchip

Re: [PATCH v13 3/4] drm/atomic-helper: Re-order bridge chain pre-enable and post-disable

[Vicente Bergas]

On Mon, Oct 6, 2025 at 5:30 PM Aradhya Bhatia <aradhya.bhatia@linux.dev> wrote:
>
> +rockchip maintainers
>
> Hi Vicente, all,

Hi everybody,
please, can some expert on this platform take a look at this bug?

Regards,
  Vicente.

> I went through the drivers and the affected areas in the gru-kevin
> chromebook pipeline last week, but nothing has stood out.
>
>
> Pipeline:
>
> rockchip,display-subsystem / rk3399-vop (Big/Lite) (CRTC) ->
> rk3399-edp (Encoder) -> analogix_dp_core (Bridge) ->
> sharp,lq123p (edp-panel)
>
> I am unable to debug this further since I do not have the hardware.
>
> I could use some help, especially from folks who understand the hardware
> requirements better.
>
>
> On 11/09/25 09:01, Aradhya Bhatia wrote:
> > Hi Vicente,
> >
> > Thank you for the bisection and reporting the issue.
> >
> > On 10/09/25 16:17, Vicente Bergas wrote:
> >> Hi,
> >> this patch causes a regression. It has been reported in
> >> https://bugzilla.kernel.org/show_bug.cgi?id=220554
> >>
> >> It affects the gru/kevin platform (arm64,RK3399) with the Panfrost DRM driver.
> >
> > I believe the Panfrost DRM driver may only be for the GPU.
> >
> > Based on the dts files in arm64/rockchip/, this is the pipeline of the
> > gru-kevin setup that I understand.
> >
> >       rk3399-vop (Big/Lite) -> rk3399-edp -> sharp,lq123p (edp-panel)
> >
> > The setup seems to be using the drm/rockchip drivers for the display
> > controller and for the bridge.
> >
> >>
> >> When it boots in console mode, the blinking of the cursor keeps the display on.
> >> If it is turned off via /sys/class/graphics/fbcon/cursor_blink, then
> >> the display briefly shows each key press presented on screen for less
> >> than one second and then powers off.
> >>
> >> When starting the graphical mode (wayland), if there are no
> >> applications drawing on the screen, the only way to keep the display
> >> on is by continuously moving the mouse.
> >>
> >
> > Okay!
> >
> > I will have a look through the drivers. In the meanwhile, please do
> > report back if you find any other observations.
> >
> >
>
> --
> Regards
> Aradhya
>

_______________________________________________
Linux-rockchip mailing list
Linux-rockchip@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-rockchip

Re: [PATCH v13 3/4] drm/atomic-helper: Re-order bridge chain pre-enable and post-disable

[Laurent Pinchart]

+Linus, Marek and Tomi.

On Mon, Dec 01, 2025 at 08:34:37PM +0100, Vicente Bergas wrote:
> On Mon, Oct 6, 2025 at 5:30 PM Aradhya Bhatia <aradhya.bhatia@linux.dev> wrote:
> >
> > +rockchip maintainers
> >
> > Hi Vicente, all,
> 
> Hi everybody,
> please, can some expert on this platform take a look at this bug?

There are similar issues with MCDE and R-Car DU, see
https://lore.kernel.org/all/CAD++jLkNCH=8VmwXh0UJS5QZ9wB-iP2kinytT+__fq0L1PzoZQ@mail.gmail.com/

I think we should revert commit c9b1150a68d9362a0827609fc0dc1664c0d8bfe1
and work on a clean solution. It broke too many drivers.

> > I went through the drivers and the affected areas in the gru-kevin
> > chromebook pipeline last week, but nothing has stood out.
> >
> >
> > Pipeline:
> >
> > rockchip,display-subsystem / rk3399-vop (Big/Lite) (CRTC) ->
> > rk3399-edp (Encoder) -> analogix_dp_core (Bridge) ->
> > sharp,lq123p (edp-panel)
> >
> > I am unable to debug this further since I do not have the hardware.
> >
> > I could use some help, especially from folks who understand the hardware
> > requirements better.
> >
> >
> > On 11/09/25 09:01, Aradhya Bhatia wrote:
> > > Hi Vicente,
> > >
> > > Thank you for the bisection and reporting the issue.
> > >
> > > On 10/09/25 16:17, Vicente Bergas wrote:
> > >> Hi,
> > >> this patch causes a regression. It has been reported in
> > >> https://bugzilla.kernel.org/show_bug.cgi?id=220554
> > >>
> > >> It affects the gru/kevin platform (arm64,RK3399) with the Panfrost DRM driver.
> > >
> > > I believe the Panfrost DRM driver may only be for the GPU.
> > >
> > > Based on the dts files in arm64/rockchip/, this is the pipeline of the
> > > gru-kevin setup that I understand.
> > >
> > >       rk3399-vop (Big/Lite) -> rk3399-edp -> sharp,lq123p (edp-panel)
> > >
> > > The setup seems to be using the drm/rockchip drivers for the display
> > > controller and for the bridge.
> > >
> > >>
> > >> When it boots in console mode, the blinking of the cursor keeps the display on.
> > >> If it is turned off via /sys/class/graphics/fbcon/cursor_blink, then
> > >> the display briefly shows each key press presented on screen for less
> > >> than one second and then powers off.
> > >>
> > >> When starting the graphical mode (wayland), if there are no
> > >> applications drawing on the screen, the only way to keep the display
> > >> on is by continuously moving the mouse.
> > >>
> > >
> > > Okay!
> > >
> > > I will have a look through the drivers. In the meanwhile, please do
> > > report back if you find any other observations.

-- 
Regards,

Laurent Pinchart

_______________________________________________
Linux-rockchip mailing list
Linux-rockchip@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-rockchip

Re: [PATCH v13 3/4] drm/atomic-helper: Re-order bridge chain pre-enable and post-disable

[Chaoyi Chen]

[-- Attachment #1: Type: text/plain, Size: 724 bytes --]

Hi Aradhya,

On 10/6/2025 11:30 PM, Aradhya Bhatia wrote:
> +rockchip maintainers
> 
> Hi Vicente, all,
> 
> 
> I went through the drivers and the affected areas in the gru-kevin
> chromebook pipeline last week, but nothing has stood out.
> 
> 
> Pipeline:
> 
> rockchip,display-subsystem / rk3399-vop (Big/Lite) (CRTC) ->
> rk3399-edp (Encoder) -> analogix_dp_core (Bridge) ->
> sharp,lq123p (edp-panel)
> 
> I am unable to debug this further since I do not have the hardware.
> 
> I could use some help, especially from folks who understand the hardware
> requirements better.
> 
>

Sorry for late reply. Could you please try the patch in the
attachment? I think this should work without PSR. Thank you.

-- 
Best, 
Chaoyi

[-- Attachment #2: 0001-drm-bridge-analogix_dp-Force-prepare-drm-panel-in-pr.patch --]
[-- Type: text/plain, Size: 1269 bytes --]

From ae2f62f1e738b22d6f0724e3dd14df3fde5ae434 Mon Sep 17 00:00:00 2001
From: Chaoyi Chen <chaoyi.chen@rock-chips.com>
Date: Wed, 3 Dec 2025 11:29:43 +0800
Subject: [PATCH] drm/bridge: analogix_dp: Force prepare drm panel in
 pre_enable

Signed-off-by: Chaoyi Chen <chaoyi.chen@rock-chips.com>
---
 drivers/gpu/drm/bridge/analogix/analogix_dp_core.c | 11 -----------
 1 file changed, 11 deletions(-)

diff --git a/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c b/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c
index efe534977d12..2c4415fc1e6e 100644
--- a/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c
+++ b/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c
@@ -1126,17 +1126,6 @@ static void analogix_dp_bridge_atomic_pre_enable(struct drm_bridge *bridge,
 						 struct drm_atomic_state *old_state)
 {
 	struct analogix_dp_device *dp = to_dp(bridge);
-	struct drm_crtc *crtc;
-	struct drm_crtc_state *old_crtc_state;
-
-	crtc = analogix_dp_get_new_crtc(dp, old_state);
-	if (!crtc)
-		return;
-
-	old_crtc_state = drm_atomic_get_old_crtc_state(old_state, crtc);
-	/* Don't touch the panel if we're coming back from PSR */
-	if (old_crtc_state && old_crtc_state->self_refresh_active)
-		return;
 
 	drm_panel_prepare(dp->plat_data->panel);
 }
-- 
2.51.1


[-- Attachment #3: Type: text/plain, Size: 170 bytes --]

_______________________________________________
Linux-rockchip mailing list
Linux-rockchip@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-rockchip