Re: [PATCH v3 3/3] drm/panfrost: Add initial panfrost driver

[Steven Price]

On 09/04/2019 21:54, Rob Herring wrote:
> This adds the initial driver for panfrost which supports Arm Mali
> Midgard and Bifrost family of GPUs. Currently, only the T860 and
> T760 Midgard GPUs have been tested.
> 
> v2:
> - Add GPU reset on job hangs (Tomeu)
> - Add RuntimePM and devfreq support (Tomeu)
> - Fix T760 support (Tomeu)
> - Add a TODO file (Rob, Tomeu)
> - Support multiple in fences (Tomeu)
> - Drop support for shared fences (Tomeu)
> - Fill in MMU de-init (Rob)
> - Move register definitions back to single header (Rob)
> - Clean-up hardcoded job submit todos (Rob)
> - Implement feature setup based on features/issues (Rob)
> - Add remaining Midgard DT compatible strings (Rob)
> 
> v3:
> - Add support for reset lines (Neil)
> - Add a MAINTAINERS entry (Rob)
> - Call dma_set_mask_and_coherent (Rob)
> - Do MMU invalidate on map and unmap. Restructure to do a single
>   operation per map/unmap call. (Rob)
> - Add a missing explicit padding to struct drm_panfrost_create_bo (Rob)
> - Fix 0-day error: "panfrost_devfreq.c:151:9-16: ERROR: PTR_ERR applied after initialization to constant on line 150"
> - Drop HW_FEATURE_AARCH64_MMU conditional (Rob)
> - s/DRM_PANFROST_PARAM_GPU_ID/DRM_PANFROST_PARAM_GPU_PROD_ID/ (Rob)
> - Check drm_gem_shmem_prime_import_sg_table() error code (Rob)
> - Re-order power on sequence (Rob)
> - Move panfrost_acquire_object_fences() before scheduling job (Rob)
> - Add NULL checks on array pointers in job clean-up (Rob)
> - Rework devfreq (Tomeu)
> - Fix devfreq init with no regulator (Rob)
> - Various WS and comments clean-up (Rob)
> 
> Cc: Maarten Lankhorst <maarten.lankhorst@xxxxxxxxxxxxxxx>
> Cc: Maxime Ripard <maxime.ripard@xxxxxxxxxxx>
> Cc: Sean Paul <sean@xxxxxxxxxx>
> Cc: David Airlie <airlied@xxxxxxxx>
> Cc: Daniel Vetter <daniel@xxxxxxxx>
> Cc: Lyude Paul <lyude@xxxxxxxxxx>
> Reviewed-by: Alyssa Rosenzweig <alyssa@xxxxxxxxxxxxx>
> Reviewed-by: Eric Anholt <eric@xxxxxxxxxx>
> Signed-off-by: Marty E. Plummer <hanetzer@xxxxxxxxxxxxx>
> Signed-off-by: Tomeu Vizoso <tomeu.vizoso@xxxxxxxxxxxxx>
> Signed-off-by: Neil Armstrong <narmstrong@xxxxxxxxxxxx>
> Signed-off-by: Rob Herring <robh@xxxxxxxxxx>
> ---

This looks like it has addressed my comments, so:

Reviewed-by: Steven Price <steven.price@xxxxxxx>


There's a couple of nits below if you do respin, but these shouldn't
block merging the driver.

[...]
> +static int panfrost_ioctl_submit(struct drm_device *dev, void *data,
> +		struct drm_file *file)
> +{
> +	struct panfrost_device *pfdev = dev->dev_private;
> +	struct drm_panfrost_submit *args = data;
> +	struct drm_syncobj *sync_out;
> +	struct panfrost_job *job;
> +	int ret = 0;
> +
> +	job = kcalloc(1, sizeof(*job), GFP_KERNEL);

Why not simply kzalloc()? I think you're going to want a new ioctl
if/when adding support for multiple job submissions in one ioctl.

[...]

> +static int panfrost_probe(struct platform_device *pdev)
> +{
> +	struct panfrost_device *pfdev;
> +	struct drm_device *ddev;
> +	int err;
> +
> +	pfdev = devm_kzalloc(&pdev->dev, sizeof(*pfdev), GFP_KERNEL);
> +	if (!pfdev)
> +		return -ENOMEM;
> +
> +	pfdev->pdev = pdev;
> +	pfdev->dev = &pdev->dev;
> +
> +	platform_set_drvdata(pdev, pfdev);
> +	dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(40));

The hardware reports the correct mask, from kbase:

> 	err = dma_set_mask(kbdev->dev,
> 			DMA_BIT_MASK(kbdev->gpu_props.mmu.pa_bits));
> 	if (err)
> 		goto dma_set_mask_failed;
> 
> 	err = dma_set_coherent_mask(kbdev->dev,
> 			DMA_BIT_MASK(kbdev->gpu_props.mmu.pa_bits));
> 	if (err)
> 		goto dma_set_mask_failed;

Although I'm not sure if it has ever varied.

Thanks,

Steve