Commit Graph

61717 Commits

Author SHA1 Message Date
Dan Gohman
8e0163ac1e Optimize the "bit test" code path for switch lowering in the
case where the bit mask has exactly one bit.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106716 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-24 02:06:24 +00:00
Jakob Stoklund Olesen
b8ac3b035b Revert "Replace a big gob of old coalescer logic with the new CoalescerPair class."
Whiny buildbots.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106710 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-24 00:52:22 +00:00
Gabor Greif
407014f9a5 use getNumArgOperands
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106709 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-24 00:48:48 +00:00
Gabor Greif
de9f5452d3 use ArgOperand API
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106707 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-24 00:44:01 +00:00
Devang Patel
e9916a302f Use ValueMap instead of DenseMap.
The ValueMapper used by various cloning utility maps MDNodes also.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106706 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-24 00:33:28 +00:00
Bruno Cardoso Lopes
6539dc6e6c Add AVX CMP{SS,SD}{rr,rm} instructions and encoding testcases
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106705 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-24 00:32:06 +00:00
Bruno Cardoso Lopes
34a491bd8e Move SSE and AVX shuffle, unpack and compare code to more appropriate places
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106702 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-24 00:15:50 +00:00
Jakob Stoklund Olesen
e9c59711d3 Replace a big gob of old coalescer logic with the new CoalescerPair class.
CoalescerPair can determine if a copy can be coalesced, and which register gets
merged away. The old logic in SimpleRegisterCoalescing had evolved into
something a bit too convoluted.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106701 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-24 00:12:39 +00:00
Devang Patel
774cca70b1 Use available typedef for " DenseMap<const Value*, Value*>".
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106699 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-24 00:00:42 +00:00
Devang Patel
29d3dd8a64 Cosmetic change.
Do not use "ValueMap" as a name for a local variable or an argument.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106698 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 23:55:51 +00:00
Gabor Greif
71339c965c use ArgOperand accessors
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106697 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 23:38:07 +00:00
Bill Wendling
4b722108e2 We are missing opportunites to use ldm. Take code like this:
void t(int *cp0, int *cp1, int *dp, int fmd) {
  int c0, c1, d0, d1, d2, d3;
  c0 = (*cp0++ & 0xffff) | ((*cp1++ << 16) & 0xffff0000);
  c1 = (*cp0++ & 0xffff) | ((*cp1++ << 16) & 0xffff0000);
  /* ... */
}

It code gens into something pretty bad. But with this change (analogous to the
X86 back-end), it will use ldm and generate few instructions.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106693 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 23:00:16 +00:00
Gabor Greif
8ff72b5344 use the new isFreeCall API and ArgOperand accessors
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106692 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 22:48:06 +00:00
Gabor Greif
02680f946b minor enhancement to llvm::isFreeCall API: return CallInst; no functional change
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106686 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 21:51:12 +00:00
Gabor Greif
e3401c4fae use ArgOperand getters
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106685 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 21:41:47 +00:00
Bruno Cardoso Lopes
e93e300ad0 Add AVX MOVMSK{PS,PD}rr instructions
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106683 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 21:30:27 +00:00
Bruno Cardoso Lopes
428256b818 Add tests for different AVX cmp opcodes, also teach the x86 asm parser to understand the vcmp instruction
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106678 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 21:10:57 +00:00
Eric Christopher
749bb7e2d9 Update according to feedback.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106677 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 20:49:35 +00:00
Rafael Espindola
11f403cf2b Add an extra-library-path option to the plugin. This is used to support
having a library both as bitcode and native code. We want to use the
bitcode first, but if codegen produces new undefined references we have to use
the native code to satisfy those references.

Gold has no notion of bitcode and native search directories, so instead it has
an API where the plugin can instruct it to look for the libraries it is passing
to it. This patch uses that API.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106674 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 20:20:59 +00:00
Bruno Cardoso Lopes
7dbfd07e32 Add AVX SHUF{PS,PD}{rr,rm} instructions
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106672 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 20:07:15 +00:00
Nico Weber
50b9efc2a8 Add support for the x86 instructions "pusha" and "popa".
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106671 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 20:00:58 +00:00
Bruno Cardoso Lopes
81cd7ffe45 Fix a tblgen bug.
Given the pattern below as an example:
list<dag> Pattern = [(set RC:$dst, (v4f32 (shufp:src3 RC:$src1,
                            (mem_frag addr:$src2))))];

The right reference resolving should lead to:
list<dag> Pattern = [(set VR128:$dst, (v4f32 (shufp:src3 VR128:$src1,
                            (mem_frag addr:$src2))))];
But was yielding:
list<dag> Pattern = [(set VR128:$dst, (v4f32 (shufp VR128:$src1,
                            (mem_frag addr:$src2))))];

Fix this by passing the right name when creating a new DagInit node.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106670 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 19:50:39 +00:00
Dale Johannesen
9b0d0a42a6 Reinstate correct test, remove the real invalidated test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106664 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 18:56:06 +00:00
Dale Johannesen
2d513dfb6f Remove tests invalidated by previous checkin.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106663 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 18:53:12 +00:00
Dale Johannesen
e39fdbe118 Do not do tail calls to external symbols. If the
branch turns out to be ARM-to-Thumb or vice versa
the linker cannot resolve this.  8120438.

If this optimization is going to be useful we probably
need a compiler flag "assume callees are same architecture"
or something like that.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106662 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 18:52:34 +00:00
Bill Wendling
151d26d15d MorphNodeTo doesn't preserve the memory operands. Because we're morphing a node
into the same node, but with different non-memory operands, we need to replace
the memory operands after it's finished morphing.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106643 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 18:16:24 +00:00
Daniel Dunbar
238e73f901 tests: Tweak lit.cfg to fix breakage with out-of-dir lookup.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106638 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 18:06:16 +00:00
Daniel Dunbar
cbe762b5d1 Revert r106263, "Fold the ShrinkDemandedOps pass into the regular DAGCombiner pass,"... it was causing both 'file' (with clang) and 176.gcc (with llvm-gcc) to be miscompiled.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106634 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 17:09:26 +00:00
Jim Grosbach
f1ab49e83a The generic DAG combiner can now fold atomic fences when needed, so switch
to using that.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106633 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 16:25:07 +00:00
Jim Grosbach
5def57aaf4 When using libcall expansions for the atomic intrinsics, the explicit
MEMBARRIER fences aren't necessary for ARM. Tell the combiner to fold them
away.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106631 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 16:08:49 +00:00
Jim Grosbach
9a526495e0 Some targets don't require the fencing MEMBARRIER instructions surrounding
atomic intrinsics, either because the use locking instructions for the
atomics, or because they perform the locking directly. Add support in the
DAG combiner to fold away the fences.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106630 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 16:07:42 +00:00
Jim Grosbach
b9cd499dec remove trailing whitespace
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106628 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 15:49:18 +00:00
Jakob Stoklund Olesen
014b8631c0 Add a few VNInfo data structure checks.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106627 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 15:34:36 +00:00
Gabor Greif
b37a64a726 use ArgOperand accessors
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106626 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 13:56:57 +00:00
Gabor Greif
7bbdf0c848 use ArgOperand accessors
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106623 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 13:09:06 +00:00
Gabor Greif
bb6eabf4f5 use helper to neatly access arguments
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106622 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 08:45:32 +00:00
Eric Christopher
18ebf74781 Update uses, defs, and comments for darwin tls patterns.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106621 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 08:01:49 +00:00
Nick Lewycky
ec4c02d252 Don't link against libm and libpthread which don't exist in BeOS/Haiku. Also,
Haiku like Linux provides <regex.h>, so use it. Patch by Paul Davey!


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106620 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 06:48:34 +00:00
Daniel Dunbar
d24c9d5f91 Revert r106066, "Create a more targeted fix for not sinking instructions into a range where it"... it causes bzip2 to be miscompiled by Clang.
Conflicts:

	lib/CodeGen/MachineSink.cpp

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106614 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 00:48:25 +00:00
Stuart Hastings
0ac9a6f9e1 Less incorrect handling of zero-length bitfields. Radars 7992077 and 8093043.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106611 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-23 00:31:14 +00:00
Eric Christopher
1008d35000 Get the addend correct for i386 pic.
Thanks Daniel!


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106608 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-22 23:51:47 +00:00
Bill Wendling
e45da4f156 Generate DWARF information during Apple-style build. They'll be stripped out
later on. But we need them saved in the symbols directory.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106604 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-22 23:44:15 +00:00
Dan Gohman
605c14fb11 Replace ScalarEvolution's private copy of getLoopPredecessor
with LoopInfo's public copy.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106603 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-22 23:43:28 +00:00
Bruno Cardoso Lopes
62a76c6401 Add AVX compare packed instructions
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106600 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-22 23:37:59 +00:00
Dan Gohman
dd2fa4bbea Loosen up this test so that it doesn't depend as much on register
allocation details.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106599 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-22 23:32:47 +00:00
Devang Patel
170f06ebe2 Revert 106592 for now. It causes clang-selfhost build failure.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106598 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-22 23:29:55 +00:00
Dan Gohman
caf71ab473 Fix OptimizeMax to handle an odd case where one of the max operands
is another max which folds. This fixes PR7454.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106594 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-22 23:07:13 +00:00
Bruno Cardoso Lopes
0caca3967b Reapply support for AVX unpack and interleave instructions, with
testcases this time.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106593 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-22 23:02:38 +00:00
Devang Patel
186b3d2606 If a metadata operand is seeded in value map and the metadata should also be seeded in value map. This is not limited to function local metadata.
Failure to seed metdata in such cases causes troubles when in a cloned module, metadata from a new module refers to values in old module. Usually this results in mysterious bugpoint crashes. For example,

 Checking to see if we can delete global inits: Unknown constant!
 UNREACHABLE executed at /d/g/llvm/lib/Bitcode/Writer/BitcodeWriter.cpp:904!



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106592 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-22 22:53:21 +00:00
Devang Patel
3bf329f495 While cloning a module, clone metadata attached with instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106591 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-22 22:50:42 +00:00