llvm-svn: 22523

llvm-svn: 22469

expand the code to work for all integer datatypes.  This should unbreak
alpha.

llvm-svn: 22464

the target natively supports.  This eliminates some special-case code from
the x86 backend and generates better code as well.

For an i8 to f64 conversion, before & after:

_x87 before:
        subl $2, %esp
        movb 6(%esp), %al
        movsbw %al, %ax
        movw %ax, (%esp)
        filds (%esp)
        addl $2, %esp
        ret

_x87 after:
        subl $2, %esp
        movsbw 6(%esp), %ax
        movw %ax, (%esp)
        filds (%esp)
        addl $2, %esp
        ret

_sse before:
        subl $12, %esp
        movb 16(%esp), %al
        movsbl %al, %eax
        cvtsi2sd %eax, %xmm0
        addl $12, %esp
        ret

_sse after:
        subl $12, %esp
        movsbl 16(%esp), %eax
        cvtsi2sd %eax, %xmm0
        addl $12, %esp
        ret

llvm-svn: 22452

SelectionDAGLegalize::ExpandLegalUINT_TO_FP method.

Add a new method, PromoteLegalUINT_TO_FP, which allows targets to request
that UINT_TO_FP operations be promoted to a larger input type.  This is
useful for targets that have some UINT_TO_FP or SINT_TO_FP operations but
not all of them (like X86).

The same should be done with SINT_TO_FP, but this patch does not do that
yet.

llvm-svn: 22447

llvm-svn: 22446

gv and Graphviz.

llvm-svn: 22434

llvm-svn: 22432

llvm-svn: 22429

It is not safe to call LegalizeOp on something that has already been legalized.
Instead, just force another iteration of legalization.

This could affect all platforms but X86, as this codepath is dynamically
dead on X86 (ISD::MEMSET and friends are legal).

llvm-svn: 22419

llvm-svn: 22417

This is the last MVTSDNode.

This allows us to eliminate a bunch of special case code for handling
MVTSDNodes.

llvm-svn: 22367

llvm-svn: 22366

MVTSDNode class.  This class is used to provide an operand to operators
that require an extra type.  We start by converting FP_ROUND_INREG and
SIGN_EXTEND_INREG over to using it.

llvm-svn: 22364

1. Pass Value*'s into lowering methods so that the proper pointers can be
   added to load/stores from the valist
2. Intrinsics that return void should only return a token chain, not a token
   chain/retval pair.
3. Rename LowerVAArgNext -> LowerVAArg, because VANext is long gone.

llvm-svn: 22338

1: Legalize operand in UINT_TO_FP expanision

2: SRA x, const i8 was not promoting the constant to shift amount type.
llvm-svn: 22337

and legalizing the extload.  Strange.  Should fix most alpha regressions.

llvm-svn: 22329

llvm-svn: 22320

llvm-svn: 22319

llvm-svn: 22315

tracking the instructions causing loads and stores provides more information than just the pointer being loaded or stored

llvm-svn: 22311

handling it in the 64 bit case.  The two code paths should probably be merged.

llvm-svn: 22302

See last commit for LangRef, this implements it on all targets.

llvm-svn: 22273

llvm-svn: 22254

integer types.  Add a couple checks to not perform these kinds of transform
on floating point values.

llvm-svn: 22228

long %bar(long %X) {
  %Y = mul long %X, 4294967297
  ret long %Y
}

to this:

l1_bar:
        mov %EAX, DWORD PTR [%ESP + 4]
        mov %EDX, %EAX
        add %EDX, DWORD PTR [%ESP + 8]
        ret

instead of:

l1_bar:
        mov %ECX, DWORD PTR [%ESP + 4]
        mov %EDX, 1
        mov %EAX, %ECX
        mul %EDX
        add %EDX, %ECX
        add %EDX, DWORD PTR [%ESP + 8]
        mov %EAX, %ECX
        ret

llvm-svn: 22044

llvm-svn: 22026

llvm-svn: 22020

codepath

llvm-svn: 22019

llvm-svn: 22018

llvm-svn: 22016

llvm-svn: 22015

special case hacks.

llvm-svn: 22014

llvm-svn: 22011

llvm-svn: 22010

llvm-svn: 22009

llvm-svn: 22004

llvm-svn: 22001

llvm-svn: 21993

llvm-svn: 21958