[all-commits] [llvm/llvm-project] 7a1123: [X86] Correct the conversion of VALIGND/Q intrinsi...

[Craig Topper via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 7a112356e4a19c0e63f533548fa95a03a5537fb9
      https://github.com/llvm/llvm-project/commit/7a112356e4a19c0e63f533548fa95a03a5537fb9
  Author: Craig Topper <craig.topper at sifive.com>
  Date:   2021-06-26 (Sat, 26 Jun 2021)

  Changed paths:
    M clang/lib/CodeGen/CGBuiltin.cpp
    M clang/test/CodeGen/X86/avx512vl-builtins.c

  Log Message:
  -----------
  [X86] Correct the conversion of VALIGND/Q intrinsics to shufflevector.

We need to mask the immediate to the width of a single vector
rather than 2 vectors. If we use the width of 2 vectors then
any shift larger than the length of 1 vector is going to overflow
the shuffle indices.

Fixes PR50895.