[PATCH] D82496: [amdgpu] Add codegen support for HIP dynamic shared memory.

[Matt Arsenault via Phabricator via llvm-commits]

arsenm added inline comments.


================
Comment at: llvm/lib/Target/AMDGPU/AMDGPUMachineFunction.h:34
+  /// Align for dynamic shared memory if any.
+  MaybeAlign DynLDSAlign;
 
----------------
arsenm wrote:
> This doesn't need to be MaybeAlign, just Align. Also expand to Dynamic?
> 
> Also should probably elaborate that this is used for the case where a dynamically sized global is used
This also needs to be added to the MachineFunctionInfo serialization


================
Comment at: llvm/lib/Target/AMDGPU/AMDGPUMachineFunction.h:61
   unsigned getLDSSize() const {
+    if (DynLDSAlign)
+      LDSSize = alignTo(LDSSize, *DynLDSAlign);
----------------
I think having this here actually breaks the calculation of the total size for all of the statically known globals


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D82496/new/

https://reviews.llvm.org/D82496