[PATCH] D155267: [DAGCombiner] Change foldAndOrOfSETCC() to optimize and/or patterns with floating points.

Thu Aug 31 03:31:49 PDT 2023

mstorsjo added a comment.

In D155267#4629782 <https://reviews.llvm.org/D155267#4629782>, @kmitropoulou wrote:

> @mstorsjo : I could not reproduce the failure with the provided test in my workspace.

That's weird - I just recompiled clang at 5ec13535235d07eafd64058551bc495f87c283b1 <https://reviews.llvm.org/rG5ec13535235d07eafd64058551bc495f87c283b1>, and reproduced it again.

> After some testing, I found that the problem is due to the fact that FMINNUM_IEEE and FMAXNUM_IEEE are not supported in AArch64. I added new lit tests AArch64/combine_andor_with_cmps.ll in order to address the problem.
>
> Please let me know if the fix (https://reviews.llvm.org/D159240) does not address the problems that you encountered.

I tested applying D159240 <https://reviews.llvm.org/D159240> on the commit before 5ec13535235d07eafd64058551bc495f87c283b1 <https://reviews.llvm.org/rG5ec13535235d07eafd64058551bc495f87c283b1>, and the reduced testcase compiles fine now at least, so I presume the issue is fixed. Thanks!

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D155267/new/

https://reviews.llvm.org/D155267