diff options
| author | Hans Wennborg <hans@hanshq.net> | 2016-05-17 20:13:29 +0000 |
|---|---|---|
| committer | Hans Wennborg <hans@hanshq.net> | 2016-05-17 20:13:29 +0000 |
| commit | c3fb51171e90b47d714c569e0548552d4ea46a93 (patch) | |
| tree | e847d8403e77285712a6d1104fe971f8d9d56017 /llvm/test/CodeGen/X86/inalloca-invoke.ll | |
| parent | f0a41089ff2a580ad6665f62e8198b5efaa1ede2 (diff) | |
| download | bcm5719-llvm-c3fb51171e90b47d714c569e0548552d4ea46a93.tar.gz bcm5719-llvm-c3fb51171e90b47d714c569e0548552d4ea46a93.zip | |
X86: Avoid using _chkstk when lowering WIN_ALLOCA instructions
This patch moves the expansion of WIN_ALLOCA pseudo-instructions
into a separate pass that walks the CFG and lowers the instructions
based on a conservative estimate of the offset between the stack
pointer and the lowest accessed stack address.
The goal is to reduce binary size and run-time costs by removing
calls to _chkstk. While it doesn't fix all the code quality problems
with inalloca calls, it's an incremental improvement for PR27076.
Differential Revision: http://reviews.llvm.org/D20263
llvm-svn: 269828
Diffstat (limited to 'llvm/test/CodeGen/X86/inalloca-invoke.ll')
| -rw-r--r-- | llvm/test/CodeGen/X86/inalloca-invoke.ll | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/llvm/test/CodeGen/X86/inalloca-invoke.ll b/llvm/test/CodeGen/X86/inalloca-invoke.ll index 9a184e563b1..d90e5012ba4 100644 --- a/llvm/test/CodeGen/X86/inalloca-invoke.ll +++ b/llvm/test/CodeGen/X86/inalloca-invoke.ll @@ -21,7 +21,8 @@ blah: %beg = getelementptr %frame.reverse, %frame.reverse* %rev_args, i32 0, i32 0 %end = getelementptr %frame.reverse, %frame.reverse* %rev_args, i32 0, i32 1 -; CHECK: calll __chkstk +; CHECK: pushl %eax +; CHECK: subl $20, %esp ; CHECK: movl %esp, %[[beg:[^ ]*]] ; CHECK: leal 12(%[[beg]]), %[[end:[^ ]*]] |
