[Freedreno] [PATCH] drm/msm: Do not run snapshot on non-DPU devices

[abhinavk at codeaurora.org]

Hi Fabio

Thanks for confirming.

Although I have no issues with your change, I am curious why even msm is 
probing and/or binding.
Your device tree should not be having any mdp/dpu nodes then.

Thanks

Abhinav
On 2021-09-16 04:42, Fabio Estevam wrote:
> Hi Abhinav,
> 
> On Wed, Sep 15, 2021 at 11:22 PM <abhinavk at codeaurora.org> wrote:
> 
>> Are you not using DPU or are you not using mdp4/mdp5 as well? Even if
>> you are using any of mdps, kms should
>> not be NULL. Hence wanted to check the test case.
> 
> I am running i.MX53, which is an NXP SoC, not Qualcomm's.
> 
> It does not use DPU, nor MDP4/5 and kms is NULL in this case.
> 
> Some debug prints to confirm:
> 
> --- a/drivers/gpu/drm/msm/msm_drv.c
> +++ b/drivers/gpu/drm/msm/msm_drv.c
> @@ -557,18 +557,22 @@ static int msm_drm_init(struct device *dev,
> const struct drm_driver *drv)
>         case KMS_MDP4:
>                 kms = mdp4_kms_init(ddev);
>                 priv->kms = kms;
> +               pr_err("******** KMS_MDP4\n");
>                 break;
>         case KMS_MDP5:
>                 kms = mdp5_kms_init(ddev);
> +               pr_err("******** KMS_MDP5\n");
>                 break;
>         case KMS_DPU:
>                 kms = dpu_kms_init(ddev);
> +               pr_err("******** KMS_DPU\n");
>                 priv->kms = kms;
>                 break;
>         default:
>                 /* valid only for the dummy headless case, where 
> of_node=NULL */
>                 WARN_ON(dev->of_node);
>                 kms = NULL;
> +               pr_err("******** KMS is NULL\n");
>                 break;
>         }
> 
> # dmesg | grep KMS
> [    3.153215] ******** KMS is NULL