projects / oota-llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 892803a) | patch
[X86] Allow x86 call frame optimization to fold more loads into pushes
authorMichael Kuperstein <michael.m.kuperstein@intel.com>
Wed, 12 Aug 2015 10:14:58 +0000 (10:14 +0000)
committerMichael Kuperstein <michael.m.kuperstein@intel.com>
Wed, 12 Aug 2015 10:14:58 +0000 (10:14 +0000)
commit426921ffc7c9c31a5ffe6e767d867d5e414dd898
tree08c75d8ec6828bda4a05615f457b9e5f2f1ea8cftree | snapshot
parent892803aa81bd5e1b54bfa041c09f2c551c64e72dcommit | diff
[X86] Allow x86 call frame optimization to fold more loads into pushes

This abstracts away the test for "when can we fold across a MachineInstruction"
into the the MI interface, and changes call-frame optimization use the same test
the peephole optimizer users.

Differential Revision: http://reviews.llvm.org/D11945

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@244729 91177308-0d34-0410-b5e6-96231b3b80d8
include/llvm/CodeGen/MachineInstr.h diff | blob | history
lib/CodeGen/MachineInstr.cpp diff | blob | history
lib/CodeGen/PeepholeOptimizer.cpp diff | blob | history
lib/Target/X86/X86CallFrameOptimization.cpp diff | blob | history
test/CodeGen/X86/movtopush.ll diff | blob | history
C/C++ LLVM-based compilers that forbids OOTA behaviors