llvm-svn: 47662

stack slot and store if the  SINT_TO_FP is actually legal.  This allows
us to compile:

double a(double b) {return (unsigned)b;}

to:

_a:
	cvttsd2siq	%xmm0, %rax
	movl	%eax, %eax
	cvtsi2sdq	%rax, %xmm0
	ret

instead of:

_a:
	subq	$8, %rsp
	cvttsd2siq	%xmm0, %rax
	movl	%eax, %eax
	cvtsi2sdq	%rax, %xmm0
	addq	$8, %rsp
	ret

crazy.

llvm-svn: 47660

llvm-svn: 47659

_test:
	movl	%edi, %eax
	ret

instead of:

_test:
        movl    $4294967295, %ecx
        movq    %rdi, %rax
        andq    %rcx, %rax
        ret

It would be great to write this as a Pat pattern that used subregs 
instead of a 'pseudo' instruction, but I don't know how to do that
in td files.

llvm-svn: 47658

llvm-svn: 47657

llvm-svn: 47656

ComputeMaskedBits to use the APInt form, and remove the
non-APInt form.

llvm-svn: 47654

llvm-svn: 47653

llvm-svn: 47652

Change several cases in SimplifyDemandedMask that don't ever do any
simplifying to reuse the logic in ComputeMaskedBits instead of
duplicating it.

llvm-svn: 47648

llvm-svn: 47644

llvm-svn: 47642

llvm-svn: 47641

llvm-svn: 47640

in an invalid transformation.

llvm-svn: 47639

llvm-svn: 47637

llvm-svn: 47635

llvm-svn: 47634

llvm-svn: 47633

instead of init'ing it maximally to zeros on entry.  getFreePhysReg
is pretty hot and only a few elements are typically used.  This speeds
up linscan by 5% on 176.gcc.

llvm-svn: 47631

llvm-svn: 47629

llvm-svn: 47628

would have been a Godsend here!

llvm-svn: 47625

llvm-svn: 47623

first commit of new LTO system.  It is not hooked up in the llvm/tools/Makefile, so no one will build it be default yet

llvm-svn: 47621

llvm-svn: 47619

llvm-svn: 47616

llvm-svn: 47615

Remove unnecessary getOperand/setOperand overriders.
Simplify getReturnValue()

llvm-svn: 47614

llvm-svn: 47612

Unify to ReturnInst::init() member functions.

llvm-svn: 47611

try to simplify them.

llvm-svn: 47610

llvm-svn: 47608

llvm-svn: 47607

llvm-svn: 47606

also fixes cfrac, flops, and 175.vpr

llvm-svn: 47605

Really feed llvm-as with the testcase, do not let it read from stdin. This fixes the hangs seen on solaris10.

llvm-svn: 47604

by supplying a dummy
"verbose" procedure

llvm-svn: 47603

(on solaris10, which are:
  CodeGen/PowerPC/frounds.ll
  Transforms/InstCombine/2008-02-23-MulSub.ll)

I needed a tool to figure out which one is the guilty.

To this end I have added a verbosity
option to the test/Makefile.

It can be invoked thus:

gmake check TESTSUITE=CodeGen/PowerPC VERBOSE="-v -v"

(The number of "-v"s specifies the verbosity level.
 Instead of "-v" other aliases can be specified,
 please consult the dejagnu docs for info.)

At level >= 2 following line is logged for each
test, before running it:

ABOUT TO RUN: <test>.ll

llvm-svn: 47602

CodeGen/PowerPC/illegal-element-type.ll): suppose
a node X is processed, and processing maps it to
a node Y.  Then X continues to exist in the DAG,
but with no users.  While processing some other
node, a new node may be created that happens to
be equal to X, and thus X will be reused rather
than a truly new node.  This can cause X to
"magically reappear", and since it is in the
Processed state in will not be reprocessed, so
at the end of type legalization the illegal node
X can still be present.  The solution is to replace
X with Y whenever X gets resurrected like this.

llvm-svn: 47601