Remove EFLAGS from the BLSI/BLSMSK/BLSR patterns. The nodes created by DAG combine don't contain an EFLAGS def.

llvm-svn: 170308

Simplify BMI ANDN matching to use patterns instead of a DAG combine. Also add ANDN to isDefConvertible.

llvm-svn: 170305

llvm-svn: 170304

Remove store forms of DEC/INC from isDefConvertible. Since they are stores they don't have a register def.

llvm-svn: 170303

We match the pattern "x >= y ? x-y : 0" into "subus x, y" and two special cases
if y is a constant. DAGCombiner canonicalizes those so we first have to undo the
canonicalization for those cases. The pattern occurs in gzip when the loop
vectorizer is enabled. Part of PR14613.

llvm-svn: 170273

Not all chips targeted by x86_64 have this feature, but a dramatically
increasing number do. Specifying a chip-specific tuning parameter will
continue to turn the feature on or off as appropriate for that
particular chip, but the generic flag should try to achieve the best
performance on the most widely available hardware. Today, the number of
chips with fast UA access dwarfs those without in the x86-64 space.

Note that this also brings LLVM's code generation for this '-march' flag
more in line with that of modern GCCs. Reviewed by Dan Gohman.

llvm-svn: 170269

TypeLegalizer: Do not generate target specific nodes with illegal types, because we cant type-legalize them.

llvm-svn: 170245

llvm-svn: 170052

mention the inline memcpy / memset expansion code is a mess?

This patch split the ZeroOrLdSrc argument into two: IsMemset and ZeroMemset.
The first indicates whether it is expanding a memset or a memcpy / memmove.
The later is whether the memset is a memset of zero. It's totally possible
(likely even) that targets may want to do different things for memcpy and
memset of zero.

llvm-svn: 169959

Also added more comments to explain why it is generally ok to return true.
- Rename getOptimalMemOpType argument IsZeroVal to ZeroOrLdSrc. It's meant to
be true for loaded source (memcpy) or zero constants (memset). The poor name
choice is probably some kind of legacy issue.

llvm-svn: 169954

f64 load / store on non-SSE2 x86 targets.

llvm-svn: 169944

llvm-svn: 169854

MVTs, instead of EVTs.

Accordingly, add bitsLT (and similar) to MVT.

llvm-svn: 169850

EVTs.

llvm-svn: 169848

of EVT.

llvm-svn: 169845

This shouldn't affect codegen for -O0 compiles as tail call markers are not
emitted in unoptimized compiles.  Testing with the external/internal nightly
test suite reveals no change in compile time performance.  Testing with -O1,
-O2 and -O3 with fast-isel enabled did not cause any compile-time or
execution-time failures.  All tests were performed on my x86 machine.
I'll monitor our arm testers to ensure no regressions occur there.

In an upcoming clang patch I will be marking the objc_autoreleaseReturnValue
and objc_retainAutoreleaseReturnValue as tail calls unconditionally.  While
it's theoretically true that this is just an optimization, it's an
optimization that we very much want to happen even at -O0, or else ARC
applications become substantially harder to debug.

Part of rdar://12553082

llvm-svn: 169796

1. Teach it to use overlapping unaligned load / store to copy / set the trailing
   bytes. e.g. On 86, use two pairs of movups / movaps for 17 - 31 byte copies.
2. Use f64 for memcpy / memset on targets where i64 is not legal but f64 is. e.g.
   x86 and ARM.
3. When memcpy from a constant string, do *not* replace the load with a constant
   if it's not possible to materialize an integer immediate with a single
   instruction (required a new target hook: TLI.isIntImmLegal()).
4. Use unaligned load / stores more aggressively if target hooks indicates they
   are "fast".
5. Update ARM target hooks to use unaligned load / stores. e.g. vld1.8 / vst1.8.
   Also increase the threshold to something reasonable (8 for memset, 4 pairs
   for memcpy).

This significantly improves Dhrystone, up to 50% on ARM iOS devices.

rdar://12760078

llvm-svn: 169791

Accidental commit... git svn betrayed me. Sorry for the noise.

llvm-svn: 169741

Summary:
Not all chips targeted by x86_64 have this feature, but a dramatically
increasing number do. Specifying a chip-specific tuning parameter will
continue to turn the feature on or off as appropriate for that
particular chip, but the generic flag should try to achieve the best
performance on the most widely available hardware. Today, the number of
chips with fast UA access dwarfs those without in the x86-64 space.

Note that this also brings LLVM's code generation for this '-march' flag
more in line with that of modern GCCs.

CC: llvm-commits

Differential Revision: http://llvm-reviews.chandlerc.com/D195

llvm-svn: 169740

Thanks to the PaX folks for noticing in review! We need some tests here,
any sugestions welcome...

llvm-svn: 169739

Intel chips.

The model number rules were determined by inspecting Intel's
documentation for their newer chip model numbers. My understanding is
that all of the newer Intel chips have fast unaligned memory access, but
if anyone is concerned about a particular chip, just shout.

No tests updated; it's not clear we have dedicated tests for the chips'
various features, but if anyone would like tests (or can point me at
some existing ones), I'm happy to oblige.

llvm-svn: 169730

- fix a bug which cause sigfault.
- add two testing cases which was causing crash

llvm-svn: 169687

There are still bugs in this pass, as well as other issues that are
being worked on, but the bugs are crashers that occur pretty easily in
the wild. Test cases have been sent to the original commit's review
thread.

This reverts the commits:
  r169671: Fix a logic error.
  r169604: Move the popcnt tests to an X86 subdirectory.
  r168931: Initial commit adding the pass.

llvm-svn: 169683

llvm-svn: 169651

the VSRI instruction before it since it does not affect the MSB.

Thanks Craig Topper for suggesting this.

llvm-svn: 169638

llvm-svn: 169624

understand target implementation of any_extend / extload, just generate
zero_extend in place of any_extend for liveouts when the target knows the
zero_extend will be implicit (e.g. ARM ldrb / ldrh) or folded (e.g. x86 movz).

rdar://12771555

llvm-svn: 169536

llvm-svn: 169534

llvm-svn: 169521

Remove intrinsic specific instructions for (V)MOVQUmr with patterns pointing to the normal instructions.

llvm-svn: 169482

llvm-svn: 169477

and extload's. If they are implemented as zero-extend, or implicitly
zero-extend, then this can enable more demanded bits optimizations. e.g.

define void @foo(i16* %ptr, i32 %a) nounwind {
entry:
  %tmp1 = icmp ult i32 %a, 100
  br i1 %tmp1, label %bb1, label %bb2
bb1:
  %tmp2 = load i16* %ptr, align 2
  br label %bb2
bb2:
  %tmp3 = phi i16 [ 0, %entry ], [ %tmp2, %bb1 ]
  %cmp = icmp ult i16 %tmp3, 24
  br i1 %cmp, label %bb3, label %exit
bb3:
  call void @bar() nounwind
  br label %exit
exit:
  ret void
}

This compiles to the followings before:
        push    {lr}
        mov     r2, #0
        cmp     r1, #99
        bhi     LBB0_2
@ BB#1:                                 @ %bb1
        ldrh    r2, [r0]
LBB0_2:                                 @ %bb2
        uxth    r0, r2
        cmp     r0, #23
        bhi     LBB0_4
@ BB#3:                                 @ %bb3
        bl      _bar
LBB0_4:                                 @ %exit
        pop     {lr}
        bx      lr

The uxth is not needed since ldrh implicitly zero-extend the high bits. With
this change it's eliminated.

rdar://12771555

llvm-svn: 169459

This is for the lldb team so most of but not all of the values are
to be printed as hex with this option.  Some small values like the
scale in an X86 address were requested to printed in decimal
without the leading 0x.

There may be some tweaks need to places that may still be in
decimal that they want in hex.  Specially for arm.  I made my best
guess.  Any tweaks from here should be simple.

I also did the best I know now with help from the C++ gurus
creating the cleanest formatImm() utility function and containing
the changes.  But if someone has a better idea to make something
cleaner I'm all ears and game for changing the implementation.

rdar://8109283

llvm-svn: 169393

Generate VPBLENDD for AVX2 and VPBLENDW for v16i16 type on AVX2.

llvm-svn: 169366

x ^ -1.

Patch by David Majnemer.
rdar://12755626

llvm-svn: 169339

textually as NativeClient. Also added a link to the native client project for
readers unfamiliar with it.

A Clang patch will follow shortly.

llvm-svn: 169291

missed in the first pass because the script didn't yet handle include
guards.

Note that the script is now able to handle all of these headers without
manual edits. =]

llvm-svn: 169224

Sooooo many of these had incorrect or strange main module includes.
I have manually inspected all of these, and fixed the main module
include to be the nearest plausible thing I could find. If you own or
care about any of these source files, I encourage you to take some time
and check that these edits were sensible. I can't have broken anything
(I strictly added headers, and reordered them, never removed), but they
may not be the headers you'd really like to identify as containing the
API being implemented.

Many forward declarations and missing includes were added to a header
files to allow them to parse cleanly when included first. The main
module rule does in fact have its merits. =]

llvm-svn: 169131

This revision attempts to recognize following population-count pattern:

 while(a) { c++; ... ; a &= a - 1; ... },
  where <c> and <a>could be used multiple times in the loop body.

 TODO: On X8664 and ARM, __buildin_ctpop() are not expanded to a efficent 
instruction sequence, which need to be improved in the following commits.

Reviewed by Nadav, really appreciate!

llvm-svn: 168931

The logic was not changed, only names.

llvm-svn: 168875