Bugfix stream-mask override, support old CUDA, and start Hopper support

Use a different callback to intercept the TMD/QMD later in the
launch pipeline.

Major improvements:
- Fix bug with next mask not overriding stream mask on CUDA 11.0+
- Add CUDA 6.5-10.2 support for next- and global-granularity
  partitioning masks on x86_64 and aarch64 Jetson
- Remove libdl dependency
- Partially support TMD/QMD Version 4 (Hopper)

Minor improvements:
- Check for sufficient CUDA version before before attempting to
  apply a next-granularity partitioning mask
- Only check for sufficient CUDA version on the first call to
  `libsmctrl_set_next_mask()` or `libsmctrl_set_global_mask()`,
  rather than checking every time (lowers overheads)
- Check that TMD version is sufficient before modifying it
- Improve documentation

Issues:
- Partitioning mask bits have a different meaning in TMD/QMD
  Version 4 and require floorsweeping and remapping information to
  properly construct. This information will be forthcoming in
  future releases of libsmctrl and nvdebug.

1 files changed, 3 insertions, 3 deletions

diff --git a/libsmctrl.h b/libsmctrl.h
index 6285de6..b85c0c7 100644
--- a/libsmctrl.h
+++ b/libsmctrl.h
@@ -1,5 +1,5 @@
 /**
- * Copyright 2024 Joshua Bakita
+ * Copyright 2022-2024 Joshua Bakita
  * Library to control TPC masks on CUDA launches. Co-opts preexisting debug
  * logic in the CUDA driver library, and thus requires a build with -lcuda.
  */
@@ -15,7 +15,7 @@ typedef unsigned __int128 uint128_t;
 
 // Set global default TPC mask for all kernels, incl. CUDA-internal ones
 // @param mask   A bitmask of enabled/disabled TPCs (see Notes on Bitmasks)
-// Supported: CUDA 10.2, and CUDA 11.0 - CUDA 12.1
+// Supported: CUDA 6.5 - CUDA 12.6
 extern void libsmctrl_set_global_mask(uint64_t mask);
 // Set default TPC mask for all kernels launched via `stream`
 // (overrides global mask)
@@ -27,7 +27,7 @@ extern void libsmctrl_set_stream_mask_ext(void* stream, uint128_t mask);
 // Set TPC mask for the next kernel launch from the caller's CPU thread
 // (overrides global and per-stream masks, applies only to next launch).
 // @param mask   A bitmask of enabled/disabled TPCs (see Notes on Bitmasks)
-// Supported: CUDA 11.0 - CUDA 12.1
+// Supported: CUDA 6.5 - CUDA 12.6
 extern void libsmctrl_set_next_mask(uint64_t mask);
 
 /**