llvm-svn: 41013

Increase efficiency of sign_extend_inreg by using subregisters for truncation. As the README suggests sign_extend_subreg is selected to (sext(trunc)).

llvm-svn: 41010

llvm-svn: 41009

llvm-svn: 40772

llvm-svn: 40028

llvm-svn: 39998

llvm-svn: 39977

llvm-svn: 37239

llvm-svn: 36965

llvm-svn: 36799

llvm-svn: 36185

llvm-svn: 36028

llvm-svn: 35884

llvm-svn: 35867

llvm-svn: 35637

llvm-svn: 35634

llvm-svn: 35435

llvm-svn: 35241

llvm-svn: 35107

llvm-svn: 34837

llvm-svn: 34204

llvm-svn: 34202

llvm-svn: 34072

llvm-svn: 33423

llvm-svn: 33228

* PIC-aware internal structures in X86 Codegen have been refactored
* Visibility (default/weak) has been added
* Docs fixes (external weak linkage, visibility, formatting)

llvm-svn: 33136

llvm-svn: 32945

llvm-svn: 32842

llvm-svn: 32741

_pairtest:
        movl 8(%esp), %eax
        movl 4(%esp), %ecx
        movd %eax, %xmm0
        movd %ecx, %xmm1
        addss %xmm0, %xmm1
        movl 12(%esp), %eax
        movss %xmm1, (%eax)
        ret

instead of:

_pairtest:
        subl $12, %esp
        movl 20(%esp), %eax
        movl %eax, 4(%esp)
        movl 16(%esp), %eax
        movl %eax, (%esp)
        movss (%esp), %xmm0
        addss 4(%esp), %xmm0
        movl 24(%esp), %eax
        movss %xmm0, (%eax)
        addl $12, %esp
        ret

llvm-svn: 32427

llvm-svn: 31976

llvm-svn: 31730

llvm-svn: 31649

llvm-svn: 31122

llvm-svn: 30921

llvm-svn: 30769

llvm-svn: 30763

llvm-svn: 30559

llvm-svn: 30556

llvm-svn: 30515