Add test that higher-granularity masks override lower-granularity ones

Stream-level masks should always override globally-set masks.
Next-kernel masks should always override both stream-level masks
and globally-set masks.

Tests reveal an issue with the next-kernel mask not overriding the
stream mask on CUDA 11.0+. CUDA appears to apply the per-stream
mask to the QMD/TMD after `launchCallback()` is triggered, making
it impossible to override as currently implemented.

1 files changed, 2 insertions, 0 deletions

diff --git a/libsmctrl_test_mask_shared.h b/libsmctrl_test_mask_shared.h
index f95757d..5a115b7 100644
--- a/libsmctrl_test_mask_shared.h
+++ b/libsmctrl_test_mask_shared.h
@@ -6,7 +6,9 @@ extern "C" {
 enum partitioning_type {
   PARTITION_GLOBAL,
   PARTITION_STREAM,
+  PARTITION_STREAM_OVERRIDE,
   PARTITION_NEXT,
+  PARTITION_NEXT_OVERRIDE,
 };
 
 extern int test_constrained_size_and_location(enum partitioning_type part_type);