[llvm] [LLVM][NVPTX] Add NVPTX codegen support for clusterlaunchcontrol instruction (PR #134568)

[via llvm-commits]

================
@@ -1478,6 +1478,104 @@ similar but the latter uses generic addressing (see `Generic Addressing <https:/
 
 For more information, refer `PTX ISA <https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-st-bulk>`__.
 
+
+clusterlaunchcontrol Intrinsics
+-------------------------------
+
+'``llvm.nvvm.clusterlaunchcontrol.try_cancel*``' Intrinsics
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Syntax:
+"""""""
+
+.. code-block:: llvm
+
+  declare void @llvm.nvvm.clusterlaunchcontrol.try_cancel.async(ptr %addr, ptr %mbar)
+  declare void @llvm.nvvm.clusterlaunchcontrol.try_cancel.async.shared(ptr addrspace(3) %addr, ptr addrspace(3) %mbar)
+  declare void @llvm.nvvm.clusterlaunchcontrol.try_cancel.async.multicast(ptr %addr, ptr %mbar)
+  declare void @llvm.nvvm.clusterlaunchcontrol.try_cancel.async.multicast.shared(ptr addrspace(3) %addr, ptr addrspace(3) %mbar)
----------------
gonzalobg wrote:

Even though `.multicast` modifies data/barrier in `.shared::cluster`, for these instructions "`.shared`" needs to be `.shared::cta` and PTX makes that explicit by only allowing that.

What's the reason to not be explicit about that here? (is that an NVPTX backend convention that `.shared` always only means `.shared::cta`?)

https://github.com/llvm/llvm-project/pull/134568