- Fix a subtle bug in RemoveCopyByCommutingDef. ALR is the live range where the source is defined; BLR is the live range which is defined by the copy.
If ALR and BLR overlaps and end of BLR extends beyond end of ALR, e.g.                                                                                                 
 A = or A, B                                                                                                                                                            
 ...                                                                                                                                                                    
 B = A                                                                                                                                                                  
 ...                                                                                                                                                                    
 C = A<kill>                                                                                                                                                            
 ...                                                                                                                                                                    
   = B                                                                                                                                                                  
                                                                                                                                                                        
then do not add kills of A to the newly created B interval.
- Also fix some kill info update bug.

llvm-svn: 48141

llvm-svn: 48140

Move StrongPHIElimination after live interval analysis.  This will make things happier down the road.

llvm-svn: 48138

Avoid creating BUILD_VECTOR of all zero elements of "non-normalized" type (e.g. v8i16 on x86) after legalizer. Instruction selection does not expect to see them. In all likelihood this can only be an issue in a bugpoint reduced test case.

llvm-svn: 48136

Change insert/extract subreg instructions to be able to be used in TableGen patterns.
Use the above features to reimplement an x86-64 pseudo instruction as a pattern.

llvm-svn: 48130

llvm-svn: 48128

llvm-svn: 48123

field to 32 bits, thus enabling correct handling of ByVal
structs bigger than 0x1ffff.  Abstract interface a bit.
Fixes gcc.c-torture/execute/pr23135.c and 
gcc.c-torture/execute/pr28982b.c in gcc testsuite (were ICE'ing
on ppc32, quietly producing wrong code on x86-32.)

llvm-svn: 48122

two regression tests:
test/CodeGen/PowerPC/2007-10-21-LocalRegAllocAssert.ll
test/CodeGen/PowerPC/2007-10-21-LocalRegAllocAssert2.ll

llvm-svn: 48120

cell really does support cross-regclass moves, because R3 is in lots of different regclasses, and the code is not consistent when it comes to value tracking.

llvm-svn: 48119

llvm-svn: 48118

llvm-svn: 48117

- "Redefinition of I" (iterator masks previous definition)
- include missing header file

Patch by Argiris Kirtzidis!

llvm-svn: 48115

llvm-svn: 48112

llvm-svn: 48111

llvm-svn: 48109

they are produced by calls (which are known exact) and by cross block copies
which are known to be produced by extends.

This improves:

define double @test2() {
	%tmp85 = call double asm sideeffect "fld0", "={st(0)}"()
	ret double %tmp85
}

from:

_test2:
	subl	$20, %esp
	# InlineAsm Start
	fld0
	# InlineAsm End
	fstpl	8(%esp)
	movsd	8(%esp), %xmm0
	movsd	%xmm0, (%esp)
	fldl	(%esp)
	addl	$20, %esp
	#FP_REG_KILL
	ret

to:

_test2:
	# InlineAsm Start
	fld0
	# InlineAsm End
	#FP_REG_KILL
	ret

by avoiding a f64 <-> f80 trip

llvm-svn: 48108

an RFP register class.

Teach ScheduleDAG how to handle CopyToReg with different src/dst 
reg classes.

This allows us to compile trivial inline asms that expect stuff
on the top of x87-fp stack.

llvm-svn: 48107

llvm-svn: 48106

in different register classes, e.g. copy of ST(0) to RFP*.  This gets
some really trivial inline asm working that plops things on the top of
stack (PR879)

llvm-svn: 48105

RST -> RFP{32/64/80}.  We only handle ST(0) for now.

llvm-svn: 48104

llvm-svn: 48101

llvm-svn: 48100

Secondly, we have to check whether the branch is actually pointing to the block
with the unwind in it. We could have gotten here because of the unwind_to alone.

llvm-svn: 48099

codegen yet because these can't be spilled (they don't exist until after RA).

llvm-svn: 48098

llvm-svn: 48097

at all.

llvm-svn: 48096

llvm-svn: 48094

isel'ing value preserving FP roundings from one fp stack reg to another
into a noop, instead of stack traffic.

llvm-svn: 48093

into a vector of zeros or undef, and when the top part is obviously
zero, we can just use movd + shuffle.  This allows us to compile
vec_set-B.ll into:

_test3:
	movl	$1234567, %eax
	andl	4(%esp), %eax
	movd	%eax, %xmm0
	ret

instead of:

_test3:
	subl	$28, %esp
	movl	$1234567, %eax
	andl	32(%esp), %eax
	movl	%eax, (%esp)
	movl	$0, 4(%esp)
	movq	(%esp), %xmm0
	addl	$28, %esp
	ret

llvm-svn: 48090

and also update the cloning interface's major user, the loop optimizations.

llvm-svn: 48088

llvm-svn: 48086

Add the ability to remove just one instance of a BB from a phi node. This fixes
the compile error in the tree now.

llvm-svn: 48085

removal of invoke, PR1269.

llvm-svn: 48084

llvm-svn: 48064

#include <xmmintrin.h>
__m128i doload64(short x) {return _mm_set_epi16(0,0,0,0,0,0,0,1);}

into:
	movl	$1, %eax
	movd	%eax, %xmm0
	ret

instead of a constant pool load.

llvm-svn: 48063

of BUILD_VECTORS that only have two unique elements:

1. The previous code was nondeterminstic, because it walked a map in
   SDOperand order, which isn't determinstic.
2. The previous code didn't handle the case when one element was undef
   very well.  Now we ensure that the generated shuffle mask has the
   undef vector on the RHS (instead of potentially being on the LHS)
   and that any elements that refer to it are themselves undef.  This
   allows us to compile CodeGen/X86/vec_set-9.ll into:

_test3:
	movd	%rdi, %xmm0
	punpcklqdq	%xmm0, %xmm0
	ret

instead of:

_test3:
	movd	%rdi, %xmm1
	#IMPLICIT_DEF %xmm0
	punpcklqdq	%xmm1, %xmm0
	ret

... saving a register.

llvm-svn: 48060

_test3:
	movd	%rdi, %xmm1
	#IMPLICIT_DEF %xmm0
	punpcklqdq	%xmm1, %xmm0
	ret

instead of:

_test3:
	#IMPLICIT_DEF %rax
	movd	%rax, %xmm0
	movd	%rdi, %xmm1
	punpcklqdq	%xmm1, %xmm0
	ret

This is still not ideal.  There is no reason to two xmm regs.

llvm-svn: 48058

2) Don't try to insert an i64 value into the low part of a 
   vector with movq on an x86-32 target.  This allows us to 
   compile:

__m128i doload64(short x) {return _mm_set_epi16(0,0,0,0,0,0,0,1);}

into:

_doload64:
	movaps	LCPI1_0, %xmm0
	ret

instead of:

_doload64:
	subl	$28, %esp
	movl	$0, 4(%esp)
	movl	$1, (%esp)
	movq	(%esp), %xmm0
	addl	$28, %esp
	ret

llvm-svn: 48057

SCALAR_TO_VECTOR on paths that end up not using it.

llvm-svn: 48056