arsenm wrote: > I have a conceptual objection: I don't think we can do both of these things: > > 1. Replace unused inputs of all intrinsics with poison > 2. Propagate poison from any argument, for all intrinsics > > So how should we handle this in general? Is it better to replace unused > inputs with "freeze poison"?
2 is not true. Poison does not unconditionally propagate through all intrinsic operands. These are special cases, we know the operation does not read the input. These should remain a constant poison https://github.com/llvm/llvm-project/pull/131287 _______________________________________________ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
