[PATCH i-g-t v3 2/3] tests/intel/kms_dp_linktrain_fallback: fix MST traversal state leak between subtests

[Sowmiya S <sowmiya.s@intel.com>]

traversed_mst_output_count is file-static and was never reset between
subtests. In a full-binary run the count from dp-fallback causes
setup_mst_outputs() to skip all MST outputs in subsequent subtests.

Reset it at the top of run_lt_fallback_test(). The DSC path never
reaches setup_mst_outputs() so run_dsc_sst_fallaback_test() does not
need the reset.

Signed-off-by: Sowmiya S <sowmiya.s@intel.com>
---
 tests/intel/kms_dp_linktrain_fallback.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/tests/intel/kms_dp_linktrain_fallback.c b/tests/intel/kms_dp_linktrain_fallback.c
index 115183d2d..44b875ab9 100644
--- a/tests/intel/kms_dp_linktrain_fallback.c
+++ b/tests/intel/kms_dp_linktrain_fallback.c
@@ -431,6 +431,13 @@ static bool run_lt_fallback_test(data_t *data)
 	bool ran = false;
 	igt_output_t *output;
 
+	/*
+	 * Reset per invocation so MST traversal state from a previous subtest
+	 * (e.g. dp-fallback followed by dsc-fallback) doesn't leak and cause
+	 * MST siblings to be silently skipped as "already visited".
+	 */
+	traversed_mst_output_count = 0;
+
 	for_each_connected_output(&data->display, output) {
 		data->output = output;
 
-- 
2.51.0