[all-commits] [llvm/llvm-project] 8a4031: [RISCV] Treat __riscv_vsetvl_*(-1) as vlmax.

[Craig Topper via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 8a403166aa617d66bf420ec11350e9e0f4d1f02f
      https://github.com/llvm/llvm-project/commit/8a403166aa617d66bf420ec11350e9e0f4d1f02f
  Author: Craig Topper <craig.topper at sifive.com>
  Date:   2023-06-16 (Fri, 16 Jun 2023)

  Changed paths:
    M llvm/lib/Target/RISCV/RISCVISelDAGToDAG.cpp
    M llvm/test/CodeGen/RISCV/rvv/vsetvli-intrinsics.ll

  Log Message:
  -----------
  [RISCV] Treat __riscv_vsetvl_*(-1) as vlmax.

We already treat -1 passed to instruction intrinsics as vlmax, this
make vsetvli consistent.

Reviewed By: rogfer01

Differential Revision: https://reviews.llvm.org/D152954