[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/SIISelLowering.cpp:5585
+      Type *Ty = GV->getValueType();
+      if (DAG.getDataLayout().getTypeAllocSize(Ty).isZero()) {
+        assert(PtrVT == MVT::i32 && "32-bit pointer is expected.");
----------------
Should probably add a comment explaining this is dynamically allocated or something


================
Comment at: llvm/test/CodeGen/AMDGPU/hip.extern.shared.array.ll:3-10
+ at lds0 = addrspace(3) global [512 x float] undef
+ at lds1 = addrspace(3) global [256 x float] undef
+ at lds2 = addrspace(3) global [4096 x float] undef
+ at lds3 = addrspace(3) global [67 x i8] undef
+
+ at dynamic_shared0 = external addrspace(3) global [0 x float]
+ at dynamic_shared1 = external addrspace(3) global [0 x double]
----------------
None of these use an unnaturally high alignment


Repository:
  rG LLVM Github Monorepo

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

https://reviews.llvm.org/D82496