[all-commits] [llvm/llvm-project] 1cf9ac: [TLI] Use AArch64 vector calling convention for Ar...

[David Sherwood via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 1cf9acdb750612985674a0349ed607768e1edc6d
      https://github.com/llvm/llvm-project/commit/1cf9acdb750612985674a0349ed607768e1edc6d
  Author: David Sherwood <david.sherwood at arm.com>
  Date:   2025-04-17 (Thu, 17 Apr 2025)

  Changed paths:
    M llvm/include/llvm/Analysis/TargetLibraryInfo.h
    M llvm/include/llvm/Analysis/VecFuncs.def
    M llvm/lib/Analysis/TargetLibraryInfo.cpp
    M llvm/lib/Transforms/Utils/InjectTLIMappings.cpp
    M llvm/test/Transforms/Util/add-TLI-mappings.ll
    M llvm/unittests/Analysis/ReplaceWithVecLibTest.cpp

  Log Message:
  -----------
  [TLI] Use AArch64 vector calling convention for ArmPL routines (#135790)

Some ArmPL vector math routines that operate on fixed-width
vector types obey the aarch64_vector_pcs calling convention,
e.g. armpl_vsinq_f64, etc. This patch adds support for
functions in the TLI to specify an optional calling
convention to be added to the declared function. Once this
patch lands I intend to follow up with a vectoriser patch
that adds this optional calling convention when invoking the
vector library variant from the loop vectoriser.



To unsubscribe from these emails, change your notification settings at https://github.com/llvm/llvm-project/settings/notifications