| Message ID | 20180906000350.2478-4-drawat@vmware.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | vmwgfx as a new driver for igt-gpu-tools | expand |
Quoting Deepak Rawat (2018-09-06 01:03:48) > For a Xorg bug vmwgfx has a kernel workaround which reset the value of > mode::type. This will cause crtc state not to match what is expected. Seems suspect, I think we need a stronger reason to accept iface irregularities. -Chris
diff --git a/tests/kms_atomic.c b/tests/kms_atomic.c index ac02baf0..72714e12 100644 --- a/tests/kms_atomic.c +++ b/tests/kms_atomic.c @@ -237,7 +237,12 @@ static void crtc_commit(igt_pipe_t *pipe, igt_plane_t *plane, { igt_display_commit2(pipe->display, s); - crtc_check_current_state(pipe, pipe->values, plane->values, relax); + /* + * For a vmwgfx xorg driver bug kernel module reset the value of + * mode::type so crtc state check fails for legacy commit + */ + if (!is_vmwgfx_device(pipe->display->drm_fd) || !(s == COMMIT_LEGACY)) + crtc_check_current_state(pipe, pipe->values, plane->values, relax); plane_check_current_state(plane, plane->values, relax); }
For a Xorg bug vmwgfx has a kernel workaround which reset the value of mode::type. This will cause crtc state not to match what is expected. Signed-off-by: Deepak Rawat <drawat@vmware.com> --- tests/kms_atomic.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-)