llvm-svn: 46916

pabs{b,w,d} are not two address
fix extract-to-mem sse4 ops
add sse4 vector sign extend nodes

llvm-svn: 46915

llvm-svn: 46908

llvm-svn: 46903

llvm-svn: 46902

llvm-svn: 46901

common problem of putting two terminators in the same block.  I can't write
a testcase for this because the .ll parser rejects this before the verifier
can, but this can occur when generating IR.

llvm-svn: 46900

begin adding some methods to use it this way.

llvm-svn: 46899

llvm-svn: 46898

llvm-svn: 46896

llvm-svn: 46895

It's not always safe to fold movsd into xorpd, etc. Check the alignment of the load address first to make sure it's 16 byte aligned.

llvm-svn: 46893

128-bit vectors need it only when SSE is on.

llvm-svn: 46890

llvm-svn: 46877

llvm-svn: 46876

llvm-svn: 46866

llvm-svn: 46861

to return pointers instead of references, since this is always what
is needed.

llvm-svn: 46857

llvm-svn: 46854

as <unknown>. And make some minor adjustments to the MemOperand
dump format.

llvm-svn: 46853

Before:
_main:
        subq    $8, %rsp
        leaq    _X(%rip), %rax
        movsd   8(%rax), %xmm1
        movss   _X(%rip), %xmm0
        call    _t
        xorl    %ecx, %ecx
        movl    %ecx, %eax
        addq    $8, %rsp
        ret
Now:
_main:
        subq    $8, %rsp
        movsd   _X+8(%rip), %xmm1
        movss   _X(%rip), %xmm0
        call    _t
        xorl    %ecx, %ecx
        movl    %ecx, %eax
        addq    $8, %rsp
        ret

Notice there is another idiotic codegen issue that needs to be fixed asap:
xorl    %ecx, %ecx
movl    %ecx, %eax

llvm-svn: 46850

llvm-svn: 46849

llvm-svn: 46848

llvm-svn: 46836

llvm-svn: 46833

initializer problem, a minor tweak to the way the
DAGISelEmitter finds load/store nodes, and a renaming of the
new PseudoSourceValue objects.

llvm-svn: 46827

http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080128/057882.html

This is causing a miscompilation on PPC G5 and just now seeing it on iMac x86-64.

llvm-svn: 46822

llvm-svn: 46821

RegAllocaLocal still *requires* LiveVariables since it runs PHIElimination, followed by TwoAddress which requires LiveVariables. We cannot run LiveVariables on non-SSA code.

llvm-svn: 46813

Throttle the non-local dependence analysis for basic blocks with more than 50 predecessors. Added command line option to play with this threshold.

llvm-svn: 46790

llvm-svn: 46776

Move to getCALLSEQ_END to ensure CALLSEQ_END node produces a flag. This is consistent with the definition in td file.

llvm-svn: 46775

check more intelligent.  This speeds up mem2reg from 5.29s to 
0.79s on a synthetic testcase with tons of predecessors and
phi nodes.

llvm-svn: 46767

llvm-svn: 46764

If a vr is already marked alive in a bb, then it has PHI uses that are visited earlier, then it is not killed in the def block (i.e. not dead).

llvm-svn: 46763

llvm-svn: 46746

was incorrectly simplifying "x == (gep x, 1, i)" into false, even 
though i could be negative.  As it turns out, all the code to 
handle this already existed, we just need to disable the incorrect
optimization case and let the general case handle it.

llvm-svn: 46739

llvm-svn: 46738

llvm-svn: 46725

Dwarf requires variable entries to be in the source order. Right now, since we are recording variable information at isel time this means parameters would appear in the reverse order. The short term fix is to issue recordVariable() at asm printing time instead.

llvm-svn: 46724