On 10/18/21 19:45, Matthew Auld wrote:
Even though userptr objects are always coherent with the GPU, with no
way for userspace to change this with the set_caching ioctl, even on
non-LLC platforms, there is still the 'Bypass LCC' mocs setting, which
might permit reading the contents of main memory directly.
Signed-off-by: Matthew Auld <matthew.auld@xxxxxxxxx>
Cc: Thomas Hellström <thomas.hellstrom@xxxxxxxxxxxxxxx>
---
drivers/gpu/drm/i915/gem/i915_gem_userptr.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
Reviewed-by: Thomas Hellström <thomas.hellstrom@xxxxxxxxxxxxxxx>
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_userptr.c b/drivers/gpu/drm/i915/gem/i915_gem_userptr.c
index 887aca9e8dd2..3173c9f9a040 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_userptr.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_userptr.c
@@ -165,8 +165,11 @@ static int i915_gem_userptr_get_pages(struct drm_i915_gem_object *obj)
goto err;
}
- sg_page_sizes = i915_sg_dma_sizes(st->sgl);
+ WARN_ON_ONCE(!(obj->cache_coherent & I915_BO_CACHE_COHERENT_FOR_WRITE));
+ if (i915_gem_object_can_bypass_llc(obj))
+ obj->cache_dirty = true;
+ sg_page_sizes = i915_sg_dma_sizes(st->sgl);
__i915_gem_object_set_pages(obj, st, sg_page_sizes);
return 0;
