[Intel-gfx] [PATCH igt v2] igt/drv_hangman: Skip aliased I915_EXEC_BSD

[Chris Wilson]

If the system has bsd2, we do not know which ring the kernel will alias
I915_EXEC_BSD onto and so we do not what the matching string should be.
Skip the unknown.

v2: Deny the aliased I915_EXEC_BSD exists at all; be specific!

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=103324
Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
CC: Michał Winiarski <michal.winiarski at intel.com>
---
 tests/drv_hangman.c | 15 +++++++++++++++
 1 file changed, 15 insertions(+)

diff --git a/tests/drv_hangman.c b/tests/drv_hangman.c
index 0551ec16..40c82257 100644
--- a/tests/drv_hangman.c
+++ b/tests/drv_hangman.c
@@ -184,6 +184,7 @@ static void test_error_state_capture(unsigned ring_id,
 	uint64_t offset;
 
 	igt_require(gem_has_ring(device, ring_id));
+
 	clear_error_state();
 
 	hang = igt_hang_ctx(device, 0, ring_id, HANG_ALLOW_CAPTURE, &offset);
@@ -254,6 +255,20 @@ igt_main
 		if (e->exec_id == 0)
 			continue;
 
+		/*
+		 * If the device has 2 BSD rings then due to obtuse aliasing
+		 * in the API, we can not determine which ring I915_EXEC_BSD
+		 * will map to, and so must skip the test; as the matching name
+		 * may be either bsd or bsd2 depending on the kernel/test
+		 * ordering.
+		 *
+		 * Here we are not checking that executing on every ABI engine
+		 * results in a detectable hang, but that a hang generated
+		 * from a specific HW engine gives an indentifiable result.
+		 */
+		if (e->exec_id == I915_EXEC_BSD && e->flags == 0)
+			continue;
+
 		igt_subtest_f("error-state-capture-%s", e->name)
 			test_error_state_capture(e->exec_id | e->flags,
 						 e->full_name);
-- 
2.15.0.rc2