For aligned load/store instructions, it's only required to know whether a
authorJim Grosbach <grosbach@apple.com>
Tue, 19 Jan 2010 18:31:11 +0000 (18:31 +0000)
committerJim Grosbach <grosbach@apple.com>
Tue, 19 Jan 2010 18:31:11 +0000 (18:31 +0000)
commite45ab8a0a90e4f3a59d8c38038ae3e495ee1fef3
treec2cb2db266d2fc51246754d560fdf5b1ba792e8d
parentc0404b3715151d020dda074740bbdaa794f9328a
For aligned load/store instructions, it's only required to know whether a
function can support dynamic stack realignment. That's a much easier question
to answer at instruction selection stage than whether the function actually
will have dynamic alignment prologue. This allows the removal of the
stack alignment heuristic pass, and improves code quality for cases where
the heuristic would result in dynamic alignment code being generated when
it was not strictly necessary.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@93885 91177308-0d34-0410-b5e6-96231b3b80d8
include/llvm/CodeGen/Passes.h
lib/CodeGen/MaxStackAlignment.cpp [deleted file]
lib/Target/ARM/ARMBaseInstrInfo.cpp
lib/Target/ARM/ARMBaseRegisterInfo.cpp
lib/Target/ARM/ARMBaseRegisterInfo.h
lib/Target/ARM/ARMTargetMachine.cpp
lib/Target/X86/X86InstrInfo.cpp
lib/Target/X86/X86RegisterInfo.cpp
lib/Target/X86/X86RegisterInfo.h
lib/Target/X86/X86TargetMachine.cpp