[llvm-commits] [llvm] r55139 - in /llvm/trunk: lib/Target/X86/X86ISelLowering.cpp test/CodeGen/X86/2004-02-12-Memcpy.ll test/CodeGen/X86/byval3.ll test/CodeGen/X86/byval4.ll test/CodeGen/X86/byval5.ll test/CodeGen/X86/memset-2.ll test/CodeGen/X86/memset64-on-x86-32.ll

[Dale Johannesen]

On Aug 21, 2008, at 2:00 PMPDT, Evan Cheng wrote:

> Fix a number of byval / memcpy / memset related codegen issues.
> 1. x86-64 byval alignment should be max of 8 and alignment of type.  
> Previously the code was not doing what the commit message was saying.

I am sure this used to work, although that may be because the FE  
always passed in an alignment.  Does it help with 6112686?

> 2. Do not use byte repeat move and store operations. These are slow.

They are sometimes slower than alternatives and sometimes faster,  
depending on size, alignment, and exactly which hardware you have.
IME, whatever you do here, somebody will send us a case where we  
should do something else.  I'm not objecting to the patch.