llvm-svn: 28440

FORMAL_ARGUMENTS nodes include a token operand.

llvm-svn: 28439

return argument pops the hidden struct pointer if present, not the caller.

For example, in this testcase:

struct X { int D, E, F, G; };
struct X bar() {
  struct X a;
  a.D = 0;
  a.E = 1;
  a.F = 2;
  a.G = 3;
  return a;
}
void foo(struct X *P) {
  *P = bar();
}

We used to emit:

_foo:
        subl $28, %esp
        movl 32(%esp), %eax
        movl %eax, (%esp)
        call _bar
        addl $28, %esp
        ret
_bar:
        movl 4(%esp), %eax
        movl $0, (%eax)
        movl $1, 4(%eax)
        movl $2, 8(%eax)
        movl $3, 12(%eax)
        ret

This is correct on Linux/X86 but not Darwin/X86.  With this patch, we now
emit:

_foo:
        subl $28, %esp
        movl 32(%esp), %eax
        movl %eax, (%esp)
        call _bar
***     addl $24, %esp
        ret
_bar:
        movl 4(%esp), %eax
        movl $0, (%eax)
        movl $1, 4(%eax)
        movl $2, 8(%eax)
        movl $3, 12(%eax)
***     ret $4

For the record, GCC emits (which is functionally equivalent to our new code):

_bar:
        movl    4(%esp), %eax
        movl    $3, 12(%eax)
        movl    $2, 8(%eax)
        movl    $1, 4(%eax)
        movl    $0, (%eax)
        ret     $4
_foo:
        pushl   %esi
        subl    $40, %esp
        movl    48(%esp), %esi
        leal    16(%esp), %eax
        movl    %eax, (%esp)
        call    _bar
        subl    $4, %esp
        movl    16(%esp), %eax
        movl    %eax, (%esi)
        movl    20(%esp), %eax
        movl    %eax, 4(%esi)
        movl    24(%esp), %eax
        movl    %eax, 8(%esi)
        movl    28(%esp), %eax
        movl    %eax, 12(%esi)
        addl    $40, %esp
        popl    %esi
        ret

This fixes SingleSource/Benchmarks/CoyoteBench/fftbench with LLC and the
JIT, and fixes the X86-backend portion of PR729.  The CBE still needs to
be updated.

llvm-svn: 28438

llvm-svn: 28409

llvm-svn: 28357

arguments at once.

llvm-svn: 28327

llvm-svn: 28326

llvm-svn: 28324

it doesn't currently use/maintain the chain properly.  Also, make the
X86ISelLowering.cpp file 80-col clean.

llvm-svn: 28320

llvm-svn: 28265

llvm-svn: 28139

llvm-svn: 28131

Refactor TargetMachine, pushing handling of TargetData into the target-specific subclasses.  This has one caller-visible change: getTargetData() now returns a pointer instead of a reference.

This fixes PR 759.

llvm-svn: 28074

passing by value.

llvm-svn: 28015

from each vector. e.g.
        shuffle(G1, G2, 7, 1, 5, 2)
==>
        movaps _G2, %xmm0
        shufps $151, _G1, %xmm0
        shufps $216, %xmm0, %xmm0

llvm-svn: 28011

FORMAL_ARGUMENTS SDOperand in the return result vector.

llvm-svn: 28009

for now.

Patch contributed by Alexander Friedman.

llvm-svn: 27994

llvm-svn: 27992

llvm-svn: 27989

llvm-svn: 27988

- Fixed vararg support.

llvm-svn: 27985

llvm-svn: 27980

llvm-svn: 27975

llvm-svn: 27972

llvm-svn: 27961

llvm-svn: 27959

llvm-svn: 27953

x86 and ppc for 100% dense switch statements when relocations are non-PIC.
This support will be extended and enhanced in the coming days to support
PIC, and less dense forms of jump tables.

llvm-svn: 27947

Don't do all the lowering stuff for 2-wide build_vector's. Also, minor optimization for shuffle of undef.

llvm-svn: 27946

Fix a performance regression. Use {p}shuf* when there are only two distinct elements in a build_vector.

llvm-svn: 27945

movd always clear the top 96 bits and movss does so when it's loading the
value from memory.
The net result is codegen for 4-wide shuffles is much improved. It is near
optimal if one or more elements is a zero. e.g.

__m128i test(int a, int b) {
  return _mm_set_epi32(0, 0, b, a);
}

compiles to

_test:
	movd 8(%esp), %xmm1
	movd 4(%esp), %xmm0
	punpckldq %xmm1, %xmm0
	ret

compare to gcc:

_test:
	subl	$12, %esp
	movd	20(%esp), %xmm0
	movd	16(%esp), %xmm1
	punpckldq	%xmm0, %xmm1
	movq	%xmm1, %xmm0
	movhps	LC0, %xmm0
	addl	$12, %esp
	ret

or icc:

_test:
        movd      4(%esp), %xmm0                                #5.10
        movd      8(%esp), %xmm3                                #5.10
        xorl      %eax, %eax                                    #5.10
        movd      %eax, %xmm1                                   #5.10
        punpckldq %xmm1, %xmm0                                  #5.10
        movd      %eax, %xmm2                                   #5.10
        punpckldq %xmm2, %xmm3                                  #5.10
        punpckldq %xmm3, %xmm0                                  #5.10
        ret                                                     #5.10

There are still room for improvement, for example the FP variant of the above example:

__m128 test(float a, float b) {
  return _mm_set_ps(0.0, 0.0, b, a);
}

_test:
	movss 8(%esp), %xmm1
	movss 4(%esp), %xmm0
	unpcklps %xmm1, %xmm0
	xorps %xmm1, %xmm1
	movlhps %xmm1, %xmm0
	ret

The xorps and movlhps are unnecessary. This will require post legalizer optimization to handle.

llvm-svn: 27939

scalar value.

e.g.
        _mm_set_epi32(0, a, 0, 0);
==>
	movd 4(%esp), %xmm0
	pshufd $69, %xmm0, %xmm0

        _mm_set_epi8(0, 0, 0, 0, 0, a, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0);
==>
	movzbw 4(%esp), %ax
	movzwl %ax, %eax
	pxor %xmm0, %xmm0
	pinsrw $5, %eax, %xmm0

llvm-svn: 27923

to a vector shuffle.
- VECTOR_SHUFFLE lowering change in preparation for more efficient codegen
of vector shuffle with zero (or any splat) vector.

llvm-svn: 27875

but i64 is not. If possible, change a i64 op to a f64 (e.g. load, constant)
and then cast it back.

llvm-svn: 27849

llvm-svn: 27847

  instructions.
- Fixed a commute vector_shuff bug.

llvm-svn: 27845

llvm-svn: 27840

llvm-svn: 27782

llvm-svn: 27779

llvm-svn: 27768