[PATCH] drm/msm/mdp5: drop cursor pipe for msm8x76

[PATCH] drm/msm/mdp5: drop cursor pipe for msm8x76

[Dmitry Baryshkov]

The MSM8976/56 family of platforms doe't have a separate cursor plane.
Instead the cursor is a part of the LM block. Drop the erroneously added
pipe configuration, letting LM to provide a legacy cursor instead.

Fixes: 1860f2a8b8b1 ("drm/msm/mdp5: Add configuration for msm8x76")
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
---
 drivers/gpu/drm/msm/disp/mdp5/mdp5_cfg.c | 9 ---------
 1 file changed, 9 deletions(-)

diff --git a/drivers/gpu/drm/msm/disp/mdp5/mdp5_cfg.c b/drivers/gpu/drm/msm/disp/mdp5/mdp5_cfg.c
index 7c91fc1915f3..389d74ff7516 100644
--- a/drivers/gpu/drm/msm/disp/mdp5/mdp5_cfg.c
+++ b/drivers/gpu/drm/msm/disp/mdp5/mdp5_cfg.c
@@ -698,15 +698,6 @@ static const struct mdp5_cfg_hw msm8x76_config = {
 			MDP_PIPE_CAP_SW_PIX_EXT	|
 			0,
 	},
-	.pipe_cursor = {
-		.count = 1,
-		.base = { 0x440DC },
-		.caps = MDP_PIPE_CAP_HFLIP	|
-			MDP_PIPE_CAP_VFLIP	|
-			MDP_PIPE_CAP_SW_PIX_EXT	|
-			MDP_PIPE_CAP_CURSOR	|
-			0,
-	},
 	.lm = {
 		.count = 2,
 		.base = { 0x44000, 0x45000 },

---
base-commit: c68a982815dcce5464e3bf2a31ac94f5146c04ca
change-id: 20260914-mdp5-fix-cursor-b80c4320e48a

Best regards,
--  
With best wishes
Dmitry

Re: [PATCH] drm/msm/mdp5: drop cursor pipe for msm8x76

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] Missing `flush_hw_mask` update when converting to legacy LM cursor breaks cursor updates.
--

commit e1dbf3d7b78aa6bf44fff588ff9b069eceea7f6b
Author: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>

drm/msm/mdp5: drop cursor pipe for msm8x76

This commit removes the separate cursor plane configuration for MSM8976/56
platforms. It drops the erroneously added pipe configuration to let the LM
block provide a legacy cursor instead.

> diff --git a/drivers/gpu/drm/msm/disp/mdp5/mdp5_cfg.c b/drivers/gpu/drm/msm/disp/mdp5/mdp5_cfg.c
> index 7c91fc1915f39..389d74ff75163 100644
> --- a/drivers/gpu/drm/msm/disp/mdp5/mdp5_cfg.c
> +++ b/drivers/gpu/drm/msm/disp/mdp5/mdp5_cfg.c
> @@ -698,15 +698,6 @@ static const struct mdp5_cfg_hw msm8x76_config = {
>  			MDP_PIPE_CAP_SW_PIX_EXT	|
>  			0,
>  	},
> -	.pipe_cursor = {
> -		.count = 1,
> -		.base = { 0x440DC },
> -		.caps = MDP_PIPE_CAP_HFLIP	|
> -			MDP_PIPE_CAP_VFLIP	|
> -			MDP_PIPE_CAP_SW_PIX_EXT	|
> -			MDP_PIPE_CAP_CURSOR	|
> -			0,
> -	},

[Severity: High]
Since this drops the cursor pipe to rely on the legacy LM cursor, does the
ctl.flush_hw_mask in msm8x76_config also need to be updated?

Currently, flush_hw_mask is set to 0xffffffff for this platform. When
standalone cursor updates use MDP5_CTL_FLUSH_CURSOR_0, fix_sw_flush() in
mdp5_ctl.c translates this to an LM flush for legacy cursors. However, this
fallback is only triggered if BIT_NEEDS_SW_FIX() determines the cursor flush
bit is excluded from the platform's flush_hw_mask.

Because the cursor bit is still included in the 0xffffffff mask, the
software fix evaluates to false. This prevents the LM from being flushed
and sends a hardware flush for a non-existent cursor block, which can
cause the cursor to lag or freeze during standalone updates.

Should flush_hw_mask be updated to exclude the cursor flush bits (for
example, 0xffffffff & ~MDP5_CTL_FLUSH_CURSOR_0)?

>  	.lm = {
>  		.count = 2,
>  		.base = { 0x44000, 0x45000 },

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260914-mdp5-fix-cursor-v1-1-9fba3ace6cb9@oss.qualcomm.com?part=1

Re: [PATCH] drm/msm/mdp5: drop cursor pipe for msm8x76

[Dmitry Baryshkov]

On Sun, Sep 13, 2026 at 11:38:43PM +0000, sashiko-bot@kernel.org wrote:
> Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
> - [High] Missing `flush_hw_mask` update when converting to legacy LM cursor breaks cursor updates.
> --
> 
> commit e1dbf3d7b78aa6bf44fff588ff9b069eceea7f6b
> Author: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
> 
> drm/msm/mdp5: drop cursor pipe for msm8x76
> 
> This commit removes the separate cursor plane configuration for MSM8976/56
> platforms. It drops the erroneously added pipe configuration to let the LM
> block provide a legacy cursor instead.
> 
> > diff --git a/drivers/gpu/drm/msm/disp/mdp5/mdp5_cfg.c b/drivers/gpu/drm/msm/disp/mdp5/mdp5_cfg.c
> > index 7c91fc1915f39..389d74ff75163 100644
> > --- a/drivers/gpu/drm/msm/disp/mdp5/mdp5_cfg.c
> > +++ b/drivers/gpu/drm/msm/disp/mdp5/mdp5_cfg.c
> > @@ -698,15 +698,6 @@ static const struct mdp5_cfg_hw msm8x76_config = {
> >  			MDP_PIPE_CAP_SW_PIX_EXT	|
> >  			0,
> >  	},
> > -	.pipe_cursor = {
> > -		.count = 1,
> > -		.base = { 0x440DC },
> > -		.caps = MDP_PIPE_CAP_HFLIP	|
> > -			MDP_PIPE_CAP_VFLIP	|
> > -			MDP_PIPE_CAP_SW_PIX_EXT	|
> > -			MDP_PIPE_CAP_CURSOR	|
> > -			0,
> > -	},
> 
> [Severity: High]
> Since this drops the cursor pipe to rely on the legacy LM cursor, does the
> ctl.flush_hw_mask in msm8x76_config also need to be updated?
> 
> Currently, flush_hw_mask is set to 0xffffffff for this platform. When
> standalone cursor updates use MDP5_CTL_FLUSH_CURSOR_0, fix_sw_flush() in
> mdp5_ctl.c translates this to an LM flush for legacy cursors. However, this
> fallback is only triggered if BIT_NEEDS_SW_FIX() determines the cursor flush
> bit is excluded from the platform's flush_hw_mask.
> 
> Because the cursor bit is still included in the 0xffffffff mask, the
> software fix evaluates to false. This prevents the LM from being flushed
> and sends a hardware flush for a non-existent cursor block, which can
> cause the cursor to lag or freeze during standalone updates.
> 
> Should flush_hw_mask be updated to exclude the cursor flush bits (for
> example, 0xffffffff & ~MDP5_CTL_FLUSH_CURSOR_0)?

This should not be necessary. The legacy cursor implemented in LM
doesn't update the HW mask.

> 
> >  	.lm = {
> >  		.count = 2,
> >  		.base = { 0x44000, 0x45000 },
> 
> -- 
> Sashiko AI review · https://sashiko.dev/#/patchset/20260914-mdp5-fix-cursor-v1-1-9fba3ace6cb9@oss.qualcomm.com?part=1

-- 
With best wishes
Dmitry

Re: [PATCH] drm/msm/mdp5: drop cursor pipe for msm8x76

[Konrad Dybcio]

On 9/14/26 1:24 AM, Dmitry Baryshkov wrote:
> The MSM8976/56 family of platforms doe't have a separate cursor plane.
> Instead the cursor is a part of the LM block. Drop the erroneously added
> pipe configuration, letting LM to provide a legacy cursor instead.
> 
> Fixes: 1860f2a8b8b1 ("drm/msm/mdp5: Add configuration for msm8x76")
> Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
> ---

The HPG agrees, the SWI disagrees.. possibly just a leftover register
though?

Konrad

Re: [PATCH] drm/msm/mdp5: drop cursor pipe for msm8x76

[Dmitry Baryshkov]

On Mon, 14 Sep 2026 02:24:24 +0300, Dmitry Baryshkov wrote:
> The MSM8976/56 family of platforms doe't have a separate cursor plane.
> Instead the cursor is a part of the LM block. Drop the erroneously added
> pipe configuration, letting LM to provide a legacy cursor instead.

Applied to ttt, thanks!

[1/1] drm/msm/mdp5: drop cursor pipe for msm8x76
      https://gitlab.freedesktop.org/lumag/msm/-/commit/c33cee4c07f1

Best regards,
-- 
With best wishes
Dmitry