[PATCH] D105206: [RISCV] Implement lround*/llround*/lrint*/llrint* with fcvt instruction with -fno-math-errno

Jessica Clarke via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Tue Jul 6 10:31:50 PDT 2021


jrtc27 added inline comments.


================
Comment at: llvm/test/CodeGen/RISCV/double-intrinsics-rv32.ll:2
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=riscv32 -mattr=+d -verify-machineinstrs < %s \
+; RUN:   | FileCheck -check-prefix=RV32IFD %s
----------------
jrtc27 wrote:
> craig.topper wrote:
> > jrtc27 wrote:
> > > We upstreamed support to UTC to let you pre-process with `sed 's/iXLEN/i32/' | llc ...` so you can have just the template rather than need two copies of the file. Helps keeps things a bit more manageable (though unfortunately you've already burned the time writing both copies...).
> > Is there an example test that does this?
> Not with update_llc_test_checks.py as it's relatively recent, but there are lots of sed-using tests in llvm/test/CodeGen/Hexagon, and a few in AArch64, ARM, WinEH and X86. It should "just work" if you do the obvious thing.
Well, llvm/test/tools/UpdateTestChecks/update_llc_test_checks/Inputs/pre-process.ll also exists as a test for this use case


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D105206/new/

https://reviews.llvm.org/D105206



More information about the llvm-commits mailing list