[PATCH] D152431: [Inliner] Handle convergence control when inlining a call

[Matt Arsenault via Phabricator via llvm-commits]

arsenm added inline comments.


================
Comment at: llvm/lib/Analysis/ConvergenceControlUsage.cpp:34
+      }
+      if (CB->getOperandBundle(LLVMContext::OB_convergencectrl).has_value())
+        return ConvergenceControlUsage::ControlledOps;
----------------
Don't think you need has_value()


================
Comment at: llvm/test/Transforms/Inline/convergence-inline.ll:2
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; RUN: opt -passes='cgscc(inline)' -S %s | FileCheck %s
+
----------------
Pre-commit test with baseline checks


================
Comment at: llvm/test/Transforms/Inline/convergence-inline.ll:101
+  call void @convergent_callee(i32 0) [ "convergencectrl"(token %tok.loop) ]
+  br i1 undef, label %hdr, label %end
+
----------------
Don't use branch on undef 


================
Comment at: llvm/test/Transforms/Inline/convergence-inline.ll:106
+}
+
+declare void @f(i32) convergent
----------------
Add indirect with convergence bundle call? Also a recursive call?


================
Comment at: llvm/test/Transforms/Inline/convergence-inline.ll:108
+declare void @f(i32) convergent
+
+declare token @llvm.experimental.convergence.entry()
----------------
Test where a convergent call is passed as an argument to a call that also has a convergence token 


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D152431/new/

https://reviews.llvm.org/D152431