projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: e219e66) | patch
[ValueTracking] Fix computeKnownFPClass with canonicalize
authorPiotr Sobczak <Piotr.Sobczak@amd.com>
Mon, 15 May 2023 12:13:51 +0000 (14:13 +0200)
committerPiotr Sobczak <Piotr.Sobczak@amd.com>
Mon, 15 May 2023 12:50:35 +0000 (14:50 +0200)
commit7322d35476c34aa92720088554fb627154a99a6e
tree1c5fe4819fbf1696486fc7a5918fd30199fe8e28tree | snapshot
parente219e66e442ba630ff21fdd495df705b3d485d41commit | diff
[ValueTracking] Fix computeKnownFPClass with canonicalize

Update code that assumes llvm.canonicalize only handles scalars,
by adding a call to getScalarType().
This is fine, as the intrinsic is trivially vectorizable.

Introduced in D147870, and uncovered by D148065.

Differential Revision: https://reviews.llvm.org/D150556
llvm/lib/Analysis/ValueTracking.cpp diff | blob | history
llvm/test/Transforms/InstCombine/canonicalize.ll diff | blob | history
Domain: System / Toolchain;