[llvm-branch-commits] [llvm] [AMDGPU] Allocate and reserve registers for address space 13 objects (PR #214248)
Matt Arsenault via llvm-branch-commits
llvm-branch-commits at lists.llvm.org
Mon Nov 2 20:48:48 PST 2026
================
@@ -343,6 +346,17 @@ struct AMDGPUPromoteAllocaPass
TargetMachine &TM;
};
+/// Only allocates objects in the VGPR ("as memory") address space, which is
+/// required for functionality; used in place of AMDGPUPromoteAllocaPass when
+/// optimizations are disabled.
+struct AMDGPUVGPRAllocatePass : OptionalPassInfoMixin<AMDGPUVGPRAllocatePass> {
+ AMDGPUVGPRAllocatePass(TargetMachine &TM) : TM(TM) {}
----------------
arsenm wrote:
```suggestion
AMDGPUVGPRAllocatePass(GCNTargetMachine &TM) : TM(TM) {}
```
https://github.com/llvm/llvm-project/pull/214248
More information about the llvm-branch-commits
mailing list