[PATCH] D155125: [amdgpu][lds] Introduce LDS frame size function attribute

[Matt Arsenault via Phabricator via llvm-commits]

arsenm added a comment.

Should document what it means in AMDGPUUsage



================
Comment at: llvm/lib/Target/AMDGPU/AMDGPUMachineFunction.cpp:177
+    uint32_t attrSize = 0;
+    StringRef S = F.getFnAttribute("amdgpu-lds-size").getValueAsString();
+    if (!S.empty())
----------------
Use getFnAttributeAsParsedInteger


================
Comment at: llvm/test/CodeGen/AMDGPU/lower-module-lds.ll:67
+; CHECK: attributes #0 = { "amdgpu-lds-size"="12" }
+; CHECK: attributes #1 = { "amdgpu-elide-module-lds" }
----------------
is this one going away?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D155125/new/

https://reviews.llvm.org/D155125