[all-commits] [llvm/llvm-project] e4e69b: [RISCV] Split PseudoVSETVLI into 2 instructions to...

[Craig Topper via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: e4e69ba4d120be8abb8982bf3ae040a2ef6b6749
      https://github.com/llvm/llvm-project/commit/e4e69ba4d120be8abb8982bf3ae040a2ef6b6749
  Author: Craig Topper <craig.topper at sifive.com>
  Date:   2021-09-02 (Thu, 02 Sep 2021)

  Changed paths:
    M llvm/lib/Target/RISCV/RISCVExpandPseudoInsts.cpp
    M llvm/lib/Target/RISCV/RISCVISelDAGToDAG.cpp
    M llvm/lib/Target/RISCV/RISCVInsertVSETVLI.cpp
    M llvm/lib/Target/RISCV/RISCVInstrInfoVPseudos.td
    M llvm/test/CodeGen/RISCV/rvv/vsetvli-insert-crossbb.mir
    M llvm/test/CodeGen/RISCV/rvv/vsetvli-insert.mir

  Log Message:
  -----------
  [RISCV] Split PseudoVSETVLI into 2 instructions to allow different register classes for rs1.

X0 has special meaning for vsetvli, we need to make sure we never
create it a vsetvli that uses it by accident. This could happen
if the register coalescer coalesces a copy from X0 into this
instruction.

This patch splits the instruction so that we can have GPRNoX0
register class to use for the cases where we don't want the source
to be X0. The verifier won't let us explicitly use X0 on a GPRNoX0
operand so we need a separate pseudo for those cases.

I don't currently have a failing example for this. There was a
failure in D107957, but the coalescable copy from that example
should have been optimized away much earlier so I've fixed that.

This is not a complete fix. We still need to prevent the same
possible issue on the AVL operand of all of the vector instruction
pseudos. I don't want to make two versions of all of those so we
need to find a different solution for those. I have an idea I'm
going to try.

Differential Revision: https://reviews.llvm.org/D109110