[Intel-gfx] [PATCH i-g-t] igt/gem_tiled_blits: Show more errors

Chris Wilson chris at chris-wilson.co.uk
Thu May 31 17:29:32 UTC 2018


glk is failing gem_tiled_blits which is very odd as it doesn't use
fencing and so the tiling is all internal to the GPU. From the small
number of examples seen so far, it looks like just a single bit is being
flipped. Let's dump some values to see if it there is a larger pattern
here.

Furthermore since gem_linear_blits is also showing bitflips on glk, we
can rule out the impact of tiling altogether! It just becomes a question
of which piece of hw is broken...

References: https://bugs.freedesktop.org/show_bug.cgi?id=106608
Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
---
 tests/gem_linear_blits.c | 11 +++++++----
 tests/gem_tiled_blits.c  | 14 ++++++++------
 2 files changed, 15 insertions(+), 10 deletions(-)

diff --git a/tests/gem_linear_blits.c b/tests/gem_linear_blits.c
index 7d05fa865..8a8d2a6cc 100644
--- a/tests/gem_linear_blits.c
+++ b/tests/gem_linear_blits.c
@@ -141,16 +141,19 @@ create_bo(int fd, uint32_t val)
 static void
 check_bo(int fd, uint32_t handle, uint32_t val)
 {
+	int num_errors;
 	int i;
 
 	gem_read(fd, handle, 0, linear, sizeof(linear));
+
+	num_errors = 0;
 	for (i = 0; i < WIDTH*HEIGHT; i++) {
-		igt_assert_f(linear[i] == val,
-			     "Expected 0x%08x, found 0x%08x "
-			     "at offset 0x%08x\n",
-			     val, linear[i], i * 4);
+		if (linear[i] != val && num_errors++ < 32)
+			igt_warn("[%08x] Expected 0x%08x, found 0x%08x (difference 0x%08x)\n",
+				 i * 4, val, linear[i], val ^ linear[i]);
 		val++;
 	}
+	igt_assert_eq(num_errors, 0);
 }
 
 static void run_test(int fd, int count)
diff --git a/tests/gem_tiled_blits.c b/tests/gem_tiled_blits.c
index 0d472e3a1..26cd900b7 100644
--- a/tests/gem_tiled_blits.c
+++ b/tests/gem_tiled_blits.c
@@ -90,10 +90,11 @@ create_bo(uint32_t start_val)
 }
 
 static void
-check_bo(drm_intel_bo *bo, uint32_t start_val)
+check_bo(drm_intel_bo *bo, uint32_t val)
 {
 	drm_intel_bo *linear_bo;
 	uint32_t *linear;
+	int num_errors;
 	int i;
 
 	linear_bo = drm_intel_bo_alloc(bufmgr, "linear dst", 1024 * 1024, 4096);
@@ -103,13 +104,14 @@ check_bo(drm_intel_bo *bo, uint32_t start_val)
 	do_or_die(drm_intel_bo_map(linear_bo, 0));
 	linear = linear_bo->virtual;
 
+	num_errors = 0;
 	for (i = 0; i < 1024 * 1024 / 4; i++) {
-		igt_assert_f(linear[i] == start_val,
-			     "Expected 0x%08x, found 0x%08x "
-			     "at offset 0x%08x\n",
-			     start_val, linear[i], i * 4);
-		start_val++;
+		if (linear[i] != val && num_errors++ < 32)
+			igt_warn("[%08x] Expected 0x%08x, found 0x%08x (difference 0x%08x)\n",
+				 i * 4, val, linear[i], val ^ linear[i]);
+		val++;
 	}
+	igt_assert_eq(num_errors, 0);
 	drm_intel_bo_unmap(linear_bo);
 
 	drm_intel_bo_unreference(linear_bo);
-- 
2.17.1



More information about the Intel-gfx mailing list