[Intel-gfx] [PATCH libdrm] intel: drm_intel_bo_gem_create_from_* on platforms w/o HW tiling

[Eric Engestrom]

On Monday, 2020-01-20 18:43:43 +0200, Imre Deak wrote:
> Platforms without a HW detiler doesn't support the get_tiling IOCTL.
> Fix the drm_intel_bo_gem_create_from_* functions assuming the default
> no-tiling, no-swizzling setting for the GEM buffer in this case.
> 
> Signed-off-by: Imre Deak <imre.deak at intel.com>
> ---
>  intel/intel_bufmgr_gem.c | 42 +++++++++++++++++++++++++---------------
>  1 file changed, 26 insertions(+), 16 deletions(-)
> 
> diff --git a/intel/intel_bufmgr_gem.c b/intel/intel_bufmgr_gem.c
> index fbf48730..fc249ef1 100644
> --- a/intel/intel_bufmgr_gem.c
> +++ b/intel/intel_bufmgr_gem.c
> @@ -1069,6 +1069,27 @@ check_bo_alloc_userptr(drm_intel_bufmgr *bufmgr,
>  					  tiling_mode, stride, size, flags);
>  }
>  
> +static int get_tiling_mode(drm_intel_bufmgr_gem *bufmgr_gem,
> +			   uint32_t gem_handle,
> +			   uint32_t *tiling_mode,
> +			   uint32_t *swizzle_mode)
> +{
> +	struct drm_i915_gem_get_tiling get_tiling;
> +	int ret;
> +
> +	memclear(get_tiling);
> +	ret = drmIoctl(bufmgr_gem->fd,
> +		       DRM_IOCTL_I915_GEM_GET_TILING,
> +		       &get_tiling);

You're missing `get_tiling.handle = gem_handle;`

Or better yet, just initialise `get_tiling` and get rid of the memclear():
  struct drm_i915_gem_get_tiling get_tiling = {
    .handle = gem_handle,
  };

With either fix:
Reviewed-by: Eric Engestrom <eric at engestrom.ch>

FYI, I've posted the following MR for the equivalent Mesa changes:
https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3497

> +	if (ret != 0 && errno != EOPNOTSUPP)
> +		return ret;
> +
> +	*tiling_mode = get_tiling.tiling_mode;
> +	*swizzle_mode = get_tiling.swizzle_mode;
> +
> +	return 0;
> +}
> +
>  /**
>   * Returns a drm_intel_bo wrapping the given buffer object handle.
>   *
> @@ -1084,7 +1105,6 @@ drm_intel_bo_gem_create_from_name(drm_intel_bufmgr *bufmgr,
>  	drm_intel_bo_gem *bo_gem;
>  	int ret;
>  	struct drm_gem_open open_arg;
> -	struct drm_i915_gem_get_tiling get_tiling;
>  
>  	/* At the moment most applications only have a few named bo.
>  	 * For instance, in a DRI client only the render buffers passed
> @@ -1146,16 +1166,11 @@ drm_intel_bo_gem_create_from_name(drm_intel_bufmgr *bufmgr,
>  	HASH_ADD(name_hh, bufmgr_gem->name_table,
>  		 global_name, sizeof(bo_gem->global_name), bo_gem);
>  
> -	memclear(get_tiling);
> -	get_tiling.handle = bo_gem->gem_handle;
> -	ret = drmIoctl(bufmgr_gem->fd,
> -		       DRM_IOCTL_I915_GEM_GET_TILING,
> -		       &get_tiling);
> +	ret = get_tiling_mode(bufmgr_gem, bo_gem->gem_handle,
> +			      &bo_gem->tiling_mode, &bo_gem->swizzle_mode);
>  	if (ret != 0)
>  		goto err_unref;
>  
> -	bo_gem->tiling_mode = get_tiling.tiling_mode;
> -	bo_gem->swizzle_mode = get_tiling.swizzle_mode;
>  	/* XXX stride is unknown */
>  	drm_intel_bo_gem_set_in_aperture_size(bufmgr_gem, bo_gem, 0);
>  	DBG("bo_create_from_handle: %d (%s)\n", handle, bo_gem->name);
> @@ -2634,7 +2649,6 @@ drm_intel_bo_gem_create_from_prime(drm_intel_bufmgr *bufmgr, int prime_fd, int s
>  	int ret;
>  	uint32_t handle;
>  	drm_intel_bo_gem *bo_gem;
> -	struct drm_i915_gem_get_tiling get_tiling;
>  
>  	pthread_mutex_lock(&bufmgr_gem->lock);
>  	ret = drmPrimeFDToHandle(bufmgr_gem->fd, prime_fd, &handle);
> @@ -2688,15 +2702,11 @@ drm_intel_bo_gem_create_from_prime(drm_intel_bufmgr *bufmgr, int prime_fd, int s
>  	bo_gem->has_error = false;
>  	bo_gem->reusable = false;
>  
> -	memclear(get_tiling);
> -	get_tiling.handle = bo_gem->gem_handle;
> -	if (drmIoctl(bufmgr_gem->fd,
> -		     DRM_IOCTL_I915_GEM_GET_TILING,
> -		     &get_tiling))
> +	ret = get_tiling_mode(bufmgr_gem, handle,
> +			      &bo_gem->tiling_mode, &bo_gem->swizzle_mode);
> +	if (ret)
>  		goto err;
>  
> -	bo_gem->tiling_mode = get_tiling.tiling_mode;
> -	bo_gem->swizzle_mode = get_tiling.swizzle_mode;
>  	/* XXX stride is unknown */
>  	drm_intel_bo_gem_set_in_aperture_size(bufmgr_gem, bo_gem, 0);
>