looks like is happening:

Without the peephole optimizer:
  (1)   sub     r6, r6, #32
        orr     r12, r12, lr, lsl r9
        orr     r2, r2, r3, lsl r10
  (x)   cmp     r6, #0
        ldr     r9, LCPI2_10
        ldr     r10, LCPI2_11
  (2)   sub     r8, r8, #32
  (a)   movge   r12, lr, lsr r6
  (y)   cmp     r8, #0
LPC2_10:
        ldr     lr, [pc, r10]
  (b)   movge   r2, r3, lsr r8

With the peephole optimizer:
        ldr     r9, LCPI2_10
        ldr     r10, LCPI2_11
  (1*)  subs    r6, r6, #32
  (2*)  subs    r8, r8, #32
  (a*)  movge   r12, lr, lsr r6
  (b*)  movge   r2, r3, lsr r8

(1) is used by (x) for the conditional move at (a). (2) is used by (y) for the
conditional move at (b). After the peephole optimizer, these the flags resulting
from (1*) are ignored and only the flags from (2*) are considered for both
conditional moves.

llvm-svn: 117876

llvm-svn: 117867

llvm-svn: 117765

a basic block.

llvm-svn: 117764

elsewhere.

llvm-svn: 117763

llvm-svn: 117762

llvm-svn: 117761

llvm-svn: 117745

llvm-svn: 117720

llvm-svn: 117677

operand and one of them has a single use that is a live out copy, favor the
one that is live out. Otherwise it will be difficult to eliminate the copy
if the instruction is a loop induction variable update. e.g.

BB:
sub r1, r3, #1
str r0, [r2, r3]
mov r3, r1
cmp
bne BB

=>

BB:
str r0, [r2, r3]
sub r3, r3, #1
cmp
bne BB

This fixed the recent 256.bzip2 regression.

llvm-svn: 117675

llvm-svn: 117673

llvm-svn: 117671

We don't want unused values forming their own equivalence classes, so we lump
them all together in one class, and then merge them with the class of the last
used value.

llvm-svn: 117670

llvm-svn: 117669

Inline asm multiple alternative constraints development phase 2 - improved basic logic, added initial platform support.

llvm-svn: 117667

llvm-svn: 117643

EquvivalenceClasses.h except it looks like overkill when elements are continuous
integers.

llvm-svn: 117631

multiplicity.

llvm-svn: 117630

llvm-svn: 117629

llvm-svn: 117615

llvm-svn: 117602

in SSAUpdaterImpl.h

Verifying live intervals revealed that the old method was completely wrong, and
we need an iterative approach to calculating PHI placemant. Fortunately, we have
MachineDominators available, so we don't have to compute that over and over
like SSAUpdaterImpl.h must.

Live-out values are cached between calls to mapValue() and computed in a greedy
way, so most calls will be working with very small block sets.

Thanks to Bob for explaining how this should work.

llvm-svn: 117599

proper SSA updating.

This doesn't cause MachineDominators to be recomputed since we are already
requiring MachineLoopInfo which uses dominators as well.

llvm-svn: 117598

split.

llvm-svn: 117597

llvm-svn: 117588

llvm-svn: 117563

Also do some minor refactoring to reduce indentation.

llvm-svn: 117558

llvm-svn: 117531

llvm-svn: 117520

by the number of defs first for it to match the instruction itinerary.

llvm-svn: 117518

Putting r117193 back except for the compile time cost. Rather than assuming fallthroughs uses all registers, just gather the union of all successor liveins.

llvm-svn: 117506

There are currently 100 references to COFF::IMAGE_SCN in 6 files
and 11 different functions. Section to attribute mapping really
needs to happen in one place to avoid problems like this.

llvm-svn: 117473

llvm-svn: 117472

llvm-svn: 117453

live out.

This doesn't prevent us from inserting a loop preheader later on, if that is
better.

llvm-svn: 117424

Critical edges going into a loop are not as bad as critical exits. We can handle
them by splitting the critical edge, or by having both inside and outside
registers live out of the predecessor.

llvm-svn: 117423

Only virtuals should be requires to be connected.

llvm-svn: 117422

memory, so a MachineMemOperand is useful (not propagated
into the MachineInstr yet).  No functional change except
for dump output.

llvm-svn: 117413

them, but hopefully we won't. And this is not the right data structure
to do it anyway.

llvm-svn: 117412