llvm-svn: 82472

llvm-svn: 82471

loads in order to reduce register pressure.

llvm-svn: 82470

of enumeration type, providing the various unused enumerators as options.

llvm-svn: 82467

And fix a bug with the behavior of min/max instructions formed from
fcmp uge comparisons.

Also, use FiniteOnlyFPMath() for this code instead of UnsafeFPMath,
as it is more specific.

llvm-svn: 82466

PseudoSourceValue already effectively represents the offset from the
frame base, so the actual offset should not be added to it.

llvm-svn: 82465

bootstrap problems.

llvm-svn: 82464

by setPreservesCFG().

llvm-svn: 82463

helpful, and it didn't correctly handle the case of constants input
to PHIs for backedges.

llvm-svn: 82462

arrays and structs, which cannot be bitcast to integers.

llvm-svn: 82460

real work is performed within Sema. Addresses Chris's comments, but
still retains the heavyweight list-of-multimaps data structure.

llvm-svn: 82459

llvm-svn: 82456

llvm-svn: 82455

llvm-svn: 82454

llvm-svn: 82449

The machine code verifier no longer tolerates phi instructions with noop
operands. All MBBs on a phi instruction must be in the CFG.

llvm-svn: 82448

This doesn't kick in too much because of phi translation issues,
but this can be resolved in the future.

llvm-svn: 82447

llvm-svn: 82446

llvm-svn: 82444

analysis, one that does the xform.

llvm-svn: 82443

llvm-svn: 82442

assert if the setModuleInfo hasn't been called.

llvm-svn: 82441

llvm-svn: 82440

from a piece of a large store when both are in the same block.

This allows clang to compile the testcase in PR4216 to this code:

_test_bitfield:
	movl	4(%esp), %eax
	movl	%eax, %ecx
	andl	$-65536, %ecx
	orl	$32962, %eax
	andl	$40186, %eax
	orl	%ecx, %eax
	ret

This is not ideal, but is a whole lot better than the code produced
by llvm-gcc:

_test_bitfield:
	movw	$-32574, %ax
	orw	4(%esp), %ax
	andw	$-25350, %ax
	movw	%ax, 4(%esp)
	movw	7(%esp), %cx
	shlw	$8, %cx
	movzbl	6(%esp), %edx
	orw	%cx, %dx
	movzwl	%dx, %ecx
	shll	$16, %ecx
	movzwl	%ax, %eax
	orl	%ecx, %eax
	ret

and dramatically better than that produced by gcc 4.2:

_test_bitfield:
	pushl	%ebx
	call	L3
"L00000000001$pb":
L3:
	popl	%ebx
	movl	8(%esp), %eax
	leal	0(,%eax,4), %edx
	sarb	$7, %dl
	movl	%eax, %ecx
	andl	$7168, %ecx
	andl	$-7201, %ebx
	movzbl	%dl, %edx
	andl	$1, %edx
	sall	$5, %edx
	orl	%ecx, %ebx
	orl	%edx, %ebx
	andl	$24, %eax
	andl	$-58336, %ebx
	orl	%eax, %ebx
	orl	$32962, %ebx
	movl	%ebx, %eax
	popl	%ebx
	ret

llvm-svn: 82439

llvm-svn: 82438

llvm-svn: 82437

Fix PR4986. "r1024 = insert_subreg r1024, undef, 2" cannot be turned in an implicit_def. Instead, it's an identity copy so it should be eliminated. Also make sure to update livevariable kill information.

llvm-svn: 82436

llvm-svn: 82435

is.
 - The problem is that formatted_ostream forces its underlying buffer to be
   unbuffered, so if some client happens to wrap a formatted_ostream around
   something, but still use the underlying stream, then we can end up writing on
   a fully unbuffered output (which was never intended to be unbuffered).

 - This makes clang (and presumably llvm-gcc) -emit-llvm -S a mere 10x faster.

llvm-svn: 82434

llvm-svn: 82433

 - Currently this requires us to fake an input file.

 - This allows Sema to be keep all the logic for how to pull decls out of the external AST source and how to handle things like tentative definitions.

llvm-svn: 82432

 - It isn't really clear what to do with the preprocessor here, but this is more sensible.

llvm-svn: 82431

llvm-svn: 82430

Add Diagnostic to Indexer, and have it keep its own FileManager instead of taking an external reference (which was leaked in the case of the CIndex library).

llvm-svn: 82429

llvm-svn: 82427

llvm-svn: 82426

llvm-svn: 82425

; CHECK: foo
; CHECK-NOT: foo
; CHECK: bar

would always fail.

llvm-svn: 82424

llvm-svn: 82423

run if appropriate target is supported.

llvm-svn: 82419