drm/xe/xe2hpg: Remove extra allocation of CCS pages for dgfx
authorAkshata Jahagirdar <akshata.jahagirdar@intel.com>
Mon, 8 Apr 2024 17:05:41 +0000 (22:35 +0530)
committerMatt Roper <matthew.d.roper@intel.com>
Tue, 9 Apr 2024 21:22:04 +0000 (14:22 -0700)
On Xe2 dGPU, compression is only supported with VRAM. When copying from
VRAM -> system memory the KMD uses mapping with uncompressed PAT
so the copy in system memory is guaranteed to be uncompressed.
When restoring such buffers from system memory -> VRAM the KMD can't
easily know which pages were originally compressed, so we always use
uncompressed -> uncompressed here.
so this means that there's no need for extra CCS storage on such
platforms.

v2: More description added to commit message

Signed-off-by: Akshata Jahagirdar <akshata.jahagirdar@intel.com>
Signed-off-by: Balasubramani Vivekanandan <balasubramani.vivekanandan@intel.com>
Reviewed-by: Himal Prasad Ghimiray <himal.prasad.ghimiray@intel.com>
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240408170545.3769566-8-balasubramani.vivekanandan@intel.com
drivers/gpu/drm/xe/xe_bo.c

index 6166bc7156560c53bbf867ffd1361c4cb9121e26..fdeb3691d3f6f11253099c346dbb1758203803a3 100644 (file)
@@ -2201,6 +2201,9 @@ bool xe_bo_needs_ccs_pages(struct xe_bo *bo)
 {
        struct xe_device *xe = xe_bo_device(bo);
 
+       if (GRAPHICS_VER(xe) >= 20 && IS_DGFX(xe))
+               return false;
+
        if (!xe_device_has_flat_ccs(xe) || bo->ttm.type != ttm_bo_type_device)
                return false;