The TLB on GFX8 stores each block of 8 PTEs where any of the valid bits
are set.

Fixes: 5255e146 ("drm/amdgpu: rework TLB flushing")
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Tested-by: NMichal Kubecek <mkubecek@suse.cz>
Signed-off-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>