AlexIndustrial/mesa
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

anv: Prepare anv_device_get_pat_entry() for discrete GPUs

Browse Source 
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25462>
This commit is contained in:
José Roberto de Souza
2023-11-22 07:36:36 -08:00
committed by Marge Bot
parent eb18a92ef9
commit 99ae565af2
1 changed files with 10 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/intel/vulkan/anv_device.c
+10
View File
@@ -5187,6 +5187,16 @@ const struct intel_device_info_pat_entry *
anv_device_get_pat_entry(struct anv_device *device,
enum anv_bo_alloc_flags alloc_flags)
{
/* PAT indexes has no actual effect in DG2 and DG1, smem caches will always
* be snopped by GPU and lmem will always be WC.
* This might change in future discrete platforms.
*/
if (anv_physical_device_has_vram(device->physical)) {
if (alloc_flags & ANV_BO_ALLOC_NO_LOCAL_MEM)
return &device->info->pat.cached_coherent;
return &device->info->pat.writecombining;
}
if (alloc_flags & (ANV_BO_ALLOC_HOST_CACHED_COHERENT))
return &device->info->pat.cached_coherent;
else if (alloc_flags & (ANV_BO_ALLOC_EXTERNAL | ANV_BO_ALLOC_SCANOUT))