Cached memory accesses between the CPU and the GPU are not coherent on ARM. Use the DMA TTM allocator on this architecture to obtain coherent memory.
Signed-off-by: Alexandre Courbot acourbot@nvidia.com --- drivers/gpu/drm/nouveau/nouveau_ttm.c | 2 ++ 1 file changed, 2 insertions(+)
diff --git a/drivers/gpu/drm/nouveau/nouveau_ttm.c b/drivers/gpu/drm/nouveau/nouveau_ttm.c index 70893416d4b6..d77fc8a5167d 100644 --- a/drivers/gpu/drm/nouveau/nouveau_ttm.c +++ b/drivers/gpu/drm/nouveau/nouveau_ttm.c @@ -385,6 +385,8 @@ nouveau_ttm_init(struct nouveau_drm *drm) if (swiotlb_nr_tbl()) drm->ttm.populate_method = DMA; #endif + if (IS_ENABLED(CONFIG_ARM)) + drm->ttm.populate_method = DMA; #if defined(TTM_HAS_AGP) if (drm->agp.stat == ENABLED) drm->ttm.populate_method = AGP;