[RFC PATCH] drm/msm/a5xx: really check for A510 in a5xx_gpu_init

[RFC PATCH] drm/msm/a5xx: really check for A510 in a5xx_gpu_init

[Dmitry Baryshkov]

The commit 010c8bbad2cb ("drm: msm: adreno: Disable preemption on Adreno
510") added special handling for a510 (this SKU doesn't seem to support
preemption, so the driver should clamp nr_rings to 1). However the
gpu->revn is not yet set (it is set later, in adreno_gpu_init()) and
thus the condition is always false. Check config->rev instead.

Fixes: 010c8bbad2cb ("drm: msm: adreno: Disable preemption on Adreno 510")
Reported-by: Adam Skladowski <a39.skl@gmail.com>
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
---
 drivers/gpu/drm/msm/adreno/a5xx_gpu.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/msm/adreno/a5xx_gpu.c b/drivers/gpu/drm/msm/adreno/a5xx_gpu.c
index 1e8d2982d603..a99310b68793 100644
--- a/drivers/gpu/drm/msm/adreno/a5xx_gpu.c
+++ b/drivers/gpu/drm/msm/adreno/a5xx_gpu.c
@@ -1743,6 +1743,7 @@ struct msm_gpu *a5xx_gpu_init(struct drm_device *dev)
 {
 	struct msm_drm_private *priv = dev->dev_private;
 	struct platform_device *pdev = priv->gpu_pdev;
+	struct adreno_platform_config *config = pdev->dev.platform_data;
 	struct a5xx_gpu *a5xx_gpu = NULL;
 	struct adreno_gpu *adreno_gpu;
 	struct msm_gpu *gpu;
@@ -1769,7 +1770,7 @@ struct msm_gpu *a5xx_gpu_init(struct drm_device *dev)
 
 	nr_rings = 4;
 
-	if (adreno_is_a510(adreno_gpu))
+	if (adreno_cmp_rev(ADRENO_REV(5, 1, 0, ANY_ID), config->rev))
 		nr_rings = 1;
 
 	ret = adreno_gpu_init(dev, pdev, adreno_gpu, &funcs, nr_rings);
-- 
2.30.2

Re: [RFC PATCH] drm/msm/a5xx: really check for A510 in a5xx_gpu_init

[Adam Skl]

On 9.04.2023 03:13, Dmitry Baryshkov wrote:

> The commit 010c8bbad2cb ("drm: msm: adreno: Disable preemption on Adreno
> 510") added special handling for a510 (this SKU doesn't seem to support
> preemption, so the driver should clamp nr_rings to 1). However the
> gpu->revn is not yet set (it is set later, in adreno_gpu_init()) and
> thus the condition is always false. Check config->rev instead.
>
> Fixes: 010c8bbad2cb ("drm: msm: adreno: Disable preemption on Adreno 510")
> Reported-by: Adam Skladowski <a39.skl@gmail.com>
> Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
> ---
>  drivers/gpu/drm/msm/adreno/a5xx_gpu.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/msm/adreno/a5xx_gpu.c b/drivers/gpu/drm/msm/adreno/a5xx_gpu.c
> index 1e8d2982d603..a99310b68793 100644
> --- a/drivers/gpu/drm/msm/adreno/a5xx_gpu.c
> +++ b/drivers/gpu/drm/msm/adreno/a5xx_gpu.c
> @@ -1743,6 +1743,7 @@ struct msm_gpu *a5xx_gpu_init(struct drm_device *dev)
>  {
>  	struct msm_drm_private *priv = dev->dev_private;
>  	struct platform_device *pdev = priv->gpu_pdev;
> +	struct adreno_platform_config *config = pdev->dev.platform_data;
>  	struct a5xx_gpu *a5xx_gpu = NULL;
>  	struct adreno_gpu *adreno_gpu;
>  	struct msm_gpu *gpu;
> @@ -1769,7 +1770,7 @@ struct msm_gpu *a5xx_gpu_init(struct drm_device *dev)
>  
>  	nr_rings = 4;
>  
> -	if (adreno_is_a510(adreno_gpu))
> +	if (adreno_cmp_rev(ADRENO_REV(5, 1, 0, ANY_ID), config->rev))
>  		nr_rings = 1;
>  
>  	ret = adreno_gpu_init(dev, pdev, adreno_gpu, &funcs, nr_rings);

After testing on Leeco S2 it appears to work as intended now, thanks.

Tested-by: Adam Skladowski <a39.skl@gmail.com>

Re: [RFC PATCH] drm/msm/a5xx: really check for A510 in a5xx_gpu_init

[Stephen Boyd]

Quoting Dmitry Baryshkov (2023-04-08 18:13:29)
> The commit 010c8bbad2cb ("drm: msm: adreno: Disable preemption on Adreno
> 510") added special handling for a510 (this SKU doesn't seem to support
> preemption, so the driver should clamp nr_rings to 1). However the
> gpu->revn is not yet set (it is set later, in adreno_gpu_init()) and
> thus the condition is always false. Check config->rev instead.
>
> Fixes: 010c8bbad2cb ("drm: msm: adreno: Disable preemption on Adreno 510")
> Reported-by: Adam Skladowski <a39.skl@gmail.com>
> Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
> ---

Maybe as a followup you can put a WARN_ON_ONCE() inside a new function
that gets gpu->revn and warns if the value is 0?

Re: [RFC PATCH] drm/msm/a5xx: really check for A510 in a5xx_gpu_init

[Dmitry Baryshkov]

On Mon, 10 Apr 2023 at 22:28, Stephen Boyd <swboyd@chromium.org> wrote:
>
> Quoting Dmitry Baryshkov (2023-04-08 18:13:29)
> > The commit 010c8bbad2cb ("drm: msm: adreno: Disable preemption on Adreno
> > 510") added special handling for a510 (this SKU doesn't seem to support
> > preemption, so the driver should clamp nr_rings to 1). However the
> > gpu->revn is not yet set (it is set later, in adreno_gpu_init()) and
> > thus the condition is always false. Check config->rev instead.
> >
> > Fixes: 010c8bbad2cb ("drm: msm: adreno: Disable preemption on Adreno 510")
> > Reported-by: Adam Skladowski <a39.skl@gmail.com>
> > Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
> > ---
>
> Maybe as a followup you can put a WARN_ON_ONCE() inside a new function
> that gets gpu->revn and warns if the value is 0?

Sounds like a good idea.

-- 
With best wishes
Dmitry