[PATCH] D94585: [IndirectFunctions] Skip propagating attributes to address taken functions

[Stanislav Mekhanoshin via Phabricator via llvm-commits]

rampitec added inline comments.


================
Comment at: llvm/lib/Target/AMDGPU/AMDGPUPropagateAttributes.cpp:242
+      if (F.hasAddressTaken()) {
+        if (!Roots.count(&F) && !NewRoots.count(&F)) {
+          NewRoots.insert(&F);
----------------
Just

```
if (!Roots.count(&F))
  NewRoots.insert(&F);
```
NewRoots is a set.


================
Comment at: llvm/test/CodeGen/AMDGPU/propagate-attributes-common-callees.ll:11
+define float @common_callee.gc(i32 %a) {
+  %add = add i32 %a, 6
+  %mul = mul nsw i32 %add, 9
----------------
95% of the code can be removed, it is not needed for the test. Here and in other places. You only need empty functions and calls.


================
Comment at: llvm/test/CodeGen/AMDGPU/propagate-attributes-direct-indirect.ll:77
+
+attributes #0 = { "amdgpu-flat-work-group-size"="1,256" "target-features"="-wavefrontsize16,-wavefrontsize32,+wavefrontsize64,+xnack" }
+attributes #1 = { convergent norecurse nounwind mustprogress
----------------
It still can be cleaned. +xnack is not needed. Most of the stuff in #1 is not needed too. Here and in other places.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D94585/new/

https://reviews.llvm.org/D94585