[i-g-t] tests/kms_color: Fix all ctm tests

Bhanuprakash Modem bhanuprakash.modem at intel.com
Sun Sep 25 12:54:03 UTC 2022


Fix the logic to generate the expected result to deal with the
hardware rounding issues.

Signed-off-by: Bhanuprakash Modem <bhanuprakash.modem at intel.com>
---
 tests/kms_color.c | 12 ++----------
 1 file changed, 2 insertions(+), 10 deletions(-)

diff --git a/tests/kms_color.c b/tests/kms_color.c
index ab285af8..f887f734 100644
--- a/tests/kms_color.c
+++ b/tests/kms_color.c
@@ -456,11 +456,6 @@ static bool test_pipe_ctm(data_t *data,
 			  color_t *after,
 			  double *ctm_matrix)
 {
-	const double ctm_identity[] = {
-		1.0, 0.0, 0.0,
-		0.0, 1.0, 0.0,
-		0.0, 0.0, 1.0
-	};
 	gamma_lut_t *degamma_linear, *gamma_linear;
 	igt_output_t *output = data->output;
 	bool ret = true;
@@ -509,13 +504,10 @@ static bool test_pipe_ctm(data_t *data,
 		disable_degamma(primary->pipe);
 		disable_gamma(primary->pipe);
 	}
-
 	disable_ctm(primary->pipe);
-	igt_display_commit(&data->display);
 
 	paint_rectangles(data, mode, after, &fb);
 	igt_plane_set_fb(primary, &fb);
-	set_ctm(primary->pipe, ctm_identity);
 	igt_display_commit(&data->display);
 	igt_wait_for_vblank(data->drm_fd,
 			    display->pipes[primary->pipe->pipe].crtc_offset);
@@ -534,7 +526,7 @@ static bool test_pipe_ctm(data_t *data,
 	 * Verify that the CRC of the software computed output is
 	 * equal to the CRC of the CTM matrix transformation output.
 	 */
-	ret &= igt_skip_crc_compare || igt_check_crc_equal(&crc_software, &crc_hardware);
+	ret = igt_skip_crc_compare || igt_check_crc_equal(&crc_software, &crc_hardware);
 
 	igt_plane_set_fb(primary, NULL);
 	igt_output_set_pipe(output, PIPE_NONE);
@@ -770,7 +762,7 @@ run_ctm_tests_for_pipe(data_t *data, enum pipe p,
 			expected_colors[0].r =
 				expected_colors[1].g =
 				expected_colors[2].b =
-				0.5 + delta * (i - 2);
+				ctm[0] + delta * (i - (iter / 2));
 			if (test_pipe_ctm(data, data->primary, red_green_blue,
 					  expected_colors, ctm)) {
 				success = true;
-- 
2.37.3



More information about the Intel-gfx-trybot mailing list