and xor. The 32-bit move immediates can be hoisted out of loops by machine
LICM but the isel hacks were preventing them.

Instead, let peephole optimization pass recognize registers that are defined by
immediates and the ARM target hook will fold the immediates in.

Other changes include 1) do not fold and / xor into cmp to isel TST / TEQ
instructions if there are multiple uses. This happens when the 'and' is live
out, machine sink would have sinked the computation and that ends up pessimizing
code. The peephole pass would recognize situations where the 'and' can be
toggled to define CPSR and eliminate the comparison anyway.

2) Move peephole pass to after machine LICM, sink, and CSE to avoid blocking
important optimizations.

rdar://8663787, rdar://8241368

llvm-svn: 119548

fixed physical registers.  Start moving fp comparison
aliases to the .td file (which default to using %st1 if
nothing is specified).

llvm-svn: 118352

add fixed immediate values.  Move the aad and aam aliases to
use this, and document it.

llvm-svn: 118350

result instruction operand numbering matched the result pattern.

Fixing this allows us to move the xchg/test aliases to the .td file.

llvm-svn: 118334

operand list instead of the operand list redundantly declared on the alias
or instruction.

With this change, we finally remove the ins/outs list on the alias.  Before:
  def : InstAlias<(outs GR16:$dst), (ins GR8 :$src),
                  "movsx $src, $dst",
                  (MOVSX16rr8W GR16:$dst, GR8:$src)>;
After:
  def : InstAlias<"movsx $src, $dst",
                  (MOVSX16rr8W GR16:$dst, GR8:$src)>;

This also makes the alias mechanism more general and powerful, which will
be exploited in subsequent patches.

llvm-svn: 118329

file instead of the asmmatcher.

llvm-svn: 118324

filling them in one at a time.  Previously this iterated over the
asmoperands, which left the problem of "holes".  The new approach
simplifies things.

llvm-svn: 118104

aliases installed and working.  They now work when the
matched pattern and the result instruction have exactly
the same operand list.

This is now enough for us to define proper aliases for
movzx and movsx, implementing rdar://8017633 and PR7459.

Note that we do not accept instructions like:
  movzx 0(%rsp), %rsi

GAS accepts this instruction, but it doesn't make any
sense because we don't know the size of the memory
operand.  It could be 8/16/32 bits.

llvm-svn: 117901

todo: the result field.

llvm-svn: 117894

CodeGenInstruction into its own helper class.  No functionality change.

llvm-svn: 117893

llvm-svn: 117892

simplify CodeGenInstruction. No functionality change.

llvm-svn: 117891

CodeGenInstruction::FlattenAsmStringVariants method.  Use it
to simplify the code in AsmWriterInst, which now no longer 
needs to worry about variants.

llvm-svn: 117886

operand values. This is useful for operands which require additional trickery
to encode into the instruction. For example, the ARM shifted immediate and
shifted register operands.

llvm-svn: 116353

try to match them by name first. If there is no by-name match, fall back to
assuming they are in order (this was the previous behavior).

llvm-svn: 116211

llvm-svn: 116068

reflection, this isn't going to achieve the purpose I intended it for.  Back to the drawing board!

llvm-svn: 114710

Not intended functionality change, as nothing uses this yet.

llvm-svn: 114702

useful after all.

llvm-svn: 110531

need the Compare flag after all.

--- Reverse-merging r109901 into '.':
U    include/llvm/Target/TargetInstrDesc.h
U    include/llvm/Target/Target.td
U    utils/TableGen/InstrInfoEmitter.cpp
U    utils/TableGen/CodeGenInstruction.cpp
U    utils/TableGen/CodeGenInstruction.h

llvm-svn: 110424

later to identify and possibly remove superfluous compare instructions -- those
that are testing for and setting a status flag that should already be set.

llvm-svn: 109901

from two places in CodeGenDAGPatterns.cpp, and
use it in DAGISelMatcherGen.cpp instead of using
an incorrect predicate that happened to get lucky
on our current targets.

llvm-svn: 99726

llvm-svn: 98879

Add checking that the input/output operand list in spelled right.

llvm-svn: 98865

llvm-svn: 97348

llvm-svn: 95808

for representing constraint info semantically instead of
as a c expression that will be blatted out to the .inc
file.  Fix X86RecognizableInstr to use this instead of
parsing C code :).

llvm-svn: 95753

bunch of associated comments, because it doesn't have anything to do
with DAGs or scheduling. This is another step in decoupling MachineInstr
emitting from scheduling.

llvm-svn: 85517

set, these flags indicate the instructions source / def operands have special
register allocation requirement that are not captured in their register classes.
Post-allocation passes (e.g. post-alloc scheduler) should not change their
allocations. e.g. ARM::LDRD require the two definitions to be allocated
even / odd register pair.

llvm-svn: 83196

llvm-svn: 60487

instruction to execute. This can be used for transformations (like two-address
conversion) to remat an instruction instead of generating a "move"
instruction. The idea is to decrease the live ranges and register pressure and
all that jazz.

llvm-svn: 51660

llvm-svn: 48381

x86 backend where instructions were not marked maystore/mayload, and perf issues where
instructions were not marked neverHasSideEffects.  It would be really nice if we could
write patterns for copy instructions.

I have audited all the x86 instructions down to MOVDQAmr.  The flags on others and on
other targets are probably not right in all cases, but no clients currently use this
info that are enabled by default.

llvm-svn: 45829

This is currently not set by anything.

llvm-svn: 45748

Evan, please review the comments I added to getNumDefs to make sure
that they are accurate, thx.

llvm-svn: 45687

just unconditionally use the def name of the instruction.

llvm-svn: 45684

llvm-svn: 45667

llvm-svn: 45656

llvm-svn: 45645

llvm-svn: 45419