Quoting kevin.rogo...@intel.com (2018-01-09 12:21:29)
> From: Kevin Rogovin <kevin.rogo...@intel.com>
>
> Signed-off-by: Kevin Rogovin <kevin.rogo...@intel.com>
> ---
> +bool
> +brw_bo_padding_is_good(struct brw_bo *bo)
> +{
> + if (bo->padding_size > 0) {
> + struct drm_i915_gem_mmap mmap_arg = {
> + .handle = bo->gem_handle,
> + .offset = bo->size,
> + .size = bo->padding_size,
> + .flags = 0,
> + };
> + uint8_t *mapped;
> + int ret;
> + uint8_t expected_value;
> +
> + /* We cannot use brw_bo_map() because it maps precisely the range
> + * [0, bo->size) and we wish to map the range of the padding which
> + * is [bo->size, bo->size + bo->pading_size)
> + */
> + ret = drmIoctl(bo->bufmgr->fd, DRM_IOCTL_I915_GEM_MMAP, &mmap_arg);
> + if (ret != 0) {
> + DBG("Unable to map mapping buffer %d (%s) buffer for pad
> checking.\n",
> + bo->gem_handle, bo->name);
> + return false;
> + }
> +
> + mapped = (uint8_t*) (uintptr_t) mmap_arg.addr_ptr;
> + /* bah-humbug, we need to see the latest contents and
> + * if the bo is not cache coherent we likely need to
> + * invalidate the cache lines to get it.
> + */
> + if (!bo->cache_coherent && !bo->bufmgr->has_llc) {
> + gen_invalidate_range(mapped, bo->padding_size);
> + }
> +
> + expected_value = bo->gem_handle & 0xFF;
> + for (uint32_t i = 0; i < bo->padding_size; ++i) {
> + if (expected_value != mapped[i]) {
> + drm_munmap(mapped, bo->padding_size);
> + return false;
> + }
> + expected_value = next_noise_value(expected_value);
> + }
> + drm_munmap(mapped, bo->padding_size);
> + }
> + return true;
> +}
Looks good from the uabi pov and interaction with brw_bo.
Reviewed-by: Chris Wilson <ch...@chris-wilson.co.uk>
I've a few stylistic nitpicks, but that's better handled by
Kenneth and/or Ian, as I very rarely get the mesa coding style right
myself :)
-Chris
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
