[PATCH] D101938: [RISCV] Initial version of a demand based vsetvli insertion pass.
Craig Topper via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Mon May 17 14:48:37 PDT 2021
craig.topper added inline comments.
================
Comment at: llvm/lib/Target/RISCV/RISCVInsertVSETVLI.cpp:235
+
+ if (PrevVSETVLI && (MI.isCall() || MI.modifiesRegister(RISCV::VL) ||
+ MI.modifiesRegister(RISCV::VTYPE))) {
----------------
frasercrmck wrote:
> Not for this patch but should we make sure calls are marked as clobbering VL/VTYPE? Or is there a reason to keep them separate?
I think any call with a regmask operand probably does appear to clobber VL since the regmask marks non-clobbered registers and not include VL. But I think I wasn't sure all calls would end up with a regmask argument.
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D101938/new/
https://reviews.llvm.org/D101938
More information about the llvm-commits
mailing list