[PATCH] D25268: Handle *_EXTEND_VECTOR_INREG during Integer Legalization

Simon Pilgrim via llvm-commits llvm-commits at lists.llvm.org
Thu Oct 6 11:51:02 PDT 2016


RKSimon accepted this revision.
RKSimon added a comment.
This revision is now accepted and ready to land.

That's some pretty ugly codegen, but at least it doesn't crash any more!

One minor request but otherwise LGTM.



> pirama wrote in promote-vec3.ll:1
> 'll add your suggested changes to this test.
> 
> I ran update_llc_test_checks and the checks seem to hard-code registers.  Is this fine?  (I've had issues in the past where hard-coded register values cause test failures on some bots where the RA picked a different set of registers).

You've set the triple so this should be fine. Although its more likely that future changes may alter the (awful) codegen, the fact that its generated from the script makes it trivial to update.

> promote-vec3.ll:5
> +; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx | FileCheck %s --check-prefix=AVX_ANY
> +; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefix=AVX_ANY
> +

Please can you add at least one x86_64 test case?

https://reviews.llvm.org/D25268





More information about the llvm-commits mailing list