registers it defines then interfere with an existing preg live range.

For instance, if we had something like these machine instructions:

BB#0
  ... = imul ... EFLAGS<imp-def,dead>
  test ..., EFLAGS<imp-def>
  jcc BB#2 EFLAGS<imp-use>

BB#1
  ... ; fallthrough to BB#2

BB#2
  ... ; No code that defines EFLAGS
  jcc ... EFLAGS<imp-use>

Machine sink will come along, see that imul implicitly defines EFLAGS, but
because it's "dead", it assumes that it can move imul into BB#2. But when it
does, imul's "dead" imp-def of EFLAGS is raised from the dead (a zombie) and
messes up the condition code for the jump (and pretty much anything else which
relies upon it being correct).

The solution is to know which pregs are live going into a basic block. However,
that information isn't calculated at this point. Nor does the LiveVariables pass
take into account non-allocatable physical registers. In lieu of this, we do a
*very* conservative pass through the basic block to determine if a preg is live
coming out of it.

llvm-svn: 105387

llvm-svn: 105359

when they move instructions.

llvm-svn: 103737

MachineLoopInfo is already available when MachineSinking runs, so the check is
free.

There is no test case because it would require a critical edge into a loop, and
CodeGenPrepare splits those. This check is just to be extra careful.

llvm-svn: 101420

Sometimes it is desirable to sink instructions along a critical edge:

x = ...
if (a && b) ...
else use(x);

The 'a && b' condition creates a critical edge to the else block, but we still
want to sink the computation of x into the block. The else block is dominated by
the parent block, so we are not pushing instructions into new code paths.

llvm-svn: 101165

llvm-svn: 100455

llvm-svn: 97765

llvm-svn: 97578

into TargetOpcodes.h.  #include the new TargetOpcodes.h
into MachineInstr.  Add new inline accessors (like isPHI())
to MachineInstr, and start using them throughout the 
codebase.

llvm-svn: 95687

llvm-svn: 92593

VISIBILITY_HIDDEN removal.

llvm-svn: 85043

Chris claims we should never have visibility_hidden inside any .cpp file but
that's still not true even after this commit.

llvm-svn: 85042

llvm-svn: 84504

llvm-svn: 84503

is trivially rematerializable and integrate it into
TargetInstrInfo::isTriviallyReMaterializable. This way, all places that
need to know whether an instruction is rematerializable will get the
same answer.

This enables the useful parts of the aggressive-remat option by
default -- using AliasAnalysis to determine whether a memory location
is invariant, and removes the questionable parts -- rematting operations
with virtual register inputs that may not be live everywhere.

llvm-svn: 83687

implementations with a new MachineInstr::isInvariantLoad, which uses
MachineMemOperands and is target-independent. This brings MachineLICM
and other functionality to targets which previously lacked an
isInvariantLoad implementation.

llvm-svn: 83475

allocatable. Even if it doesn't appear to have any defs, it may latter
on after register allocation.

llvm-svn: 82834

which have no defs anywhere in the function. In particular, this fixes sinking
of instructions that reference RIP on x86-64, which is currently being modeled
as a register.

llvm-svn: 82815

and skipping the defs.

llvm-svn: 82811

upgrading a few things to use raw_ostream

llvm-svn: 79811

llvm-svn: 79755

llvm-svn: 78139

llvm-svn: 77754

1. Sinking would crash when the first instruction of a block was
   sunk due to iterator problems.
2. Instructions could be sunk to their current block, causing an
   infinite loop.

This fixes PR3968

llvm-svn: 68787

llvm-svn: 64582

Don't sink the instruction if TargetRegisterInfo::isSafeToMoveRegClassDefs doesn't think it's safe. This works around PR1911.

llvm-svn: 63994

llvm-svn: 61715

isReg, etc., from isRegister, etc.

llvm-svn: 57006

isImmediate(), isRegister(), and friends, to avoid confusion
about having two different names with the same meaning. I'm
not attached to the longer names, and would be ok with
changing to the shorter names if others prefer it.

llvm-svn: 56189

llvm-svn: 55779

several things that were neither in an anonymous namespace nor static
but not intended to be global.

llvm-svn: 51017

llvm-svn: 48311

llvm-svn: 46930

has no stores between the load and the end of block.  This works 
great and sinks hundreds of stores, but we can't turn it on because
machineinstrs don't have volatility information and we don't want to
sink volatile stores :(

llvm-svn: 45894

both work right according to the new flags.

This removes the TII::isReallySideEffectFree predicate, and adds
TII::isInvariantLoad. 

It removes NeverHasSideEffects+MayHaveSideEffects and adds
UnmodeledSideEffects as machine instr flags.  Now the clients
can decide everything they need.

I think isRematerializable can be implemented in terms of the
flags we have now, though I will let others tackle that.

llvm-svn: 45843

llvm-svn: 45841

llvm-svn: 45625

llvm-svn: 45617

llvm-svn: 45614

It is missing validity checks, so it is known broken.  However, it is powerful enough
to compile this contrived code:

void test1(int C, double A, double B, double *P) {
  double Tmp = A*A+B*B;
  *P = C ? Tmp : A;
}

into:

_test1:
	movsd	8(%esp), %xmm0
	cmpl	$0, 4(%esp)
	je	LBB1_2	# entry
LBB1_1:	# entry
	movsd	16(%esp), %xmm1
	mulsd	%xmm1, %xmm1
	mulsd	%xmm0, %xmm0
	addsd	%xmm1, %xmm0
LBB1_2:	# entry
	movl	24(%esp), %eax
	movsd	%xmm0, (%eax)
	ret

instead of:

_test1:
	movsd	16(%esp), %xmm0
	mulsd	%xmm0, %xmm0
	movsd	8(%esp), %xmm1
	movapd	%xmm1, %xmm2
	mulsd	%xmm2, %xmm2
	addsd	%xmm0, %xmm2
	cmpl	$0, 4(%esp)
	je	LBB1_2	# entry
LBB1_1:	# entry
	movapd	%xmm2, %xmm1
LBB1_2:	# entry
	movl	24(%esp), %eax
	movsd	%xmm1, (%eax)
	ret

woo.

llvm-svn: 45570