Commit Graph

71176 Commits

Author SHA1 Message Date
Jakob Stoklund Olesen
0caa420042 Fix SystemZ tests
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128686 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 23:02:12 +00:00
Nadav Rotem
0ff8a4fa35 Instcombile optimization: extractelement(cast) -> cast(extractelement)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128683 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 22:57:29 +00:00
Jakob Stoklund Olesen
ca6fd009ad Fix ARM tests to be register allocator independent.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128680 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 22:14:03 +00:00
Benjamin Kramer
7ebdc37d2b InstCombine: APFloat can't perform arithmetic on PPC double doubles, don't even try.
Thanks Eli!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128676 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 21:35:49 +00:00
Johnny Chen
0c630b5f38 Add a test case for a malformed LDC/LDC2 instructions with PUDW = 0b0000, which
amounts to an UNDEFINED instruction.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128668 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 20:54:30 +00:00
Evan Cheng
463d358f1d Distribute (A + B) * C to (A * C) + (B * C) to make use of NEON multiplier
accumulator forwarding:
vadd d3, d0, d1
vmul d3, d3, d2
=>
vmul d3, d0, d2
vmla d3, d1, d2


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128665 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 19:38:48 +00:00
Johnny Chen
a52d7da1d8 Fix single word and unsigned byte data transfer instruction encodings so that
Inst{4} = 0.

rdar://problem/9213022


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128662 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 19:28:35 +00:00
Jakob Stoklund Olesen
a6f7499244 Fix Mips, Sparc, and XCore tests that were dependent on register allocation.
Add an extra run with -regalloc=basic to keep them honest.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128654 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 18:42:43 +00:00
Akira Hatanaka
1d6b38d9d3 Added support for FP conditional move instructions and fixed bugs in handling of FP comparisons.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128650 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 18:26:17 +00:00
Nick Lewycky
5bb3eced65 Pick better examples. "fpext float 3.1415 to double" won't parse because 3.1415
isn't an exact float. Also "fpext float 1.0 to float" is invalid IR because
it's not performing an extension.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128647 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 18:20:19 +00:00
Jakob Stoklund Olesen
280ea1a746 Don't completely eliminate identity copies that also modify super register liveness.
Turn them into noop KILL instructions instead. This lets the scavenger know when
super-registers are killed and defined.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128645 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 17:55:25 +00:00
Johnny Chen
8901e6ff3d Add BLXi to the instruction table for disassembly purpose.
A8.6.23 BLX (immediate)

rdar://problem/9212921


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128644 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 17:53:50 +00:00
Jakob Stoklund Olesen
02ae9f2f27 Allow kill flags on two-address instructions. They are harmless.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128643 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 17:52:41 +00:00
Jakob Stoklund Olesen
8e53aca51a Mark all uses as <undef> when joining a copy.
This way, shrinkToUses() will ignore the instruction that is about to be
deleted, and we avoid leaving invalid live ranges that SplitKit doesn't like.

Fix a misunderstanding in MachineVerifier about <def,undef> operands. The
<undef> flag is valid on def operands where it has the same meaning as <undef>
on a use operand. It only applies to sub-register defines which also read the
full register.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128642 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 17:23:25 +00:00
Daniel Dunbar
48a36158ec Remove stray empty test file.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128640 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 17:01:56 +00:00
Devang Patel
e62dfcf4b3 Remove dead code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128639 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 16:53:49 +00:00
Bruno Cardoso Lopes
b41aaab5a1 Revert r128632 again, until I figure out what break the tests
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128635 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 15:54:36 +00:00
Jakob Stoklund Olesen
314a3ef502 Fix bug found by valgrind.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128634 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 15:14:11 +00:00
Richard Osborne
e8f3533323 Add XCore intrinsics for initializing / starting / synchronizing threads.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128633 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 15:13:13 +00:00
Bruno Cardoso Lopes
bcd3a9cd84 Reapply r128585 without generating a lib depedency cycle. An updated log:
- Implement asm parsing support for LDRT, LDRBT, STRT, STRBT and
  {STR,LDC}{2}_{PRE,POST} fixing the encoding wherever is possible.
- Move all instructions which use am2offset without a pattern to use
  addrmode2.
- Add a new encoding bit to describe the index mode used and teach
  printAddrMode2Operand to check by the addressing mode which index
  mode to print.
- Testcases

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128632 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 14:52:28 +00:00
Michael J. Spencer
56584fcbfd Fix whitespace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128631 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 13:06:39 +00:00
Michael J. Spencer
c9c08fb3a7 Switch FileRemover from PathV1 to V2.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128630 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 13:04:19 +00:00
NAKAMURA Takumi
dc5198bac7 lib/CodeGen/LiveIntervalAnalysis.cpp: [PR9590] Don't use std::pow(float,float) here.
We don't expect the real "powf()" on some hosts (and powf() would be available on other hosts).
For consistency, std::pow(double,double) may be called instead.
Or, precision issue might attack us, to see unstable regalloc and stack coloring.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128629 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 12:11:33 +00:00
Benjamin Kramer
00e00d693c InstCombine: Fix transform to use the swapped predicate.
Thanks Frits!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128628 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 10:46:03 +00:00
Benjamin Kramer
68b4bd0a63 InstCombine: fold fcmp (fneg x), (fneg y) -> fcmp x, y
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128627 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 10:12:22 +00:00
Benjamin Kramer
0db50189dc InstCombine: fold fcmp pred (fneg x), C -> fcmp swap(pred) x, -C
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128626 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 10:12:15 +00:00
Benjamin Kramer
b194bdc03b InstCombine: Shrink "fcmp (fpext x), C" to "fcmp x, C" if C can be losslessly converted to the type of x.
Fixes PR9592.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128625 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 10:12:07 +00:00
Benjamin Kramer
cd0274ca18 InstCombine: fold fcmp (fpext x), (fpext y) -> fcmp x, y.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128624 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 10:11:58 +00:00
Duncan Sands
021cf96ffc Will not compile without the spec!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128623 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 10:03:32 +00:00
Duncan Sands
f202c43de1 Strip trailing whitespace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128622 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 09:58:51 +00:00
Bill Wendling
b38aa9af72 Testcase for r128619 (PR9571).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128620 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 08:13:57 +00:00
Jakob Stoklund Olesen
312babc93f Pick a conservative register class when creating a small live range for remat.
The rematerialized instruction may require a more constrained register class
than the register being spilled. In the test case, the spilled register has been
inflated to the DPR register class, but we are rematerializing a load of the
ssub_0 sub-register which only exists for DPR_VFP2 registers.

The register class is reinflated after spilling, so the conservative choice is
only temporary.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128610 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 03:54:44 +00:00
Matt Beaumont-Gay
e4345c9977 Revert "- Implement asm parsing support for LDRT, LDRBT, STRT, STRBT and"
This revision introduced a dependency cycle, as nlewycky mentioned by email.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128597 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 00:39:16 +00:00
Nick Lewycky
9cdf388d26 Fix typo in generated HTML.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128594 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 00:23:57 +00:00
Bob Wilson
bbe7c65390 Use intrinsics for Neon vmull operations. Radar 9208957.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128591 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 00:09:35 +00:00
Owen Anderson
6a7d36a320 Somehow we managed to forget to encode the lane index for a large swathe of NEON instructions. With this fix, the entire test-suite passes with the Thumb integrated assembler.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128587 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-30 23:45:29 +00:00
Evan Cheng
ee2e0e347e Don't try to create zero-sized stack objects.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128586 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-30 23:44:13 +00:00
Bruno Cardoso Lopes
40829ed6f5 - Implement asm parsing support for LDRT, LDRBT, STRT, STRBT and
{STR,LDC}{2}_PRE.
- Fixed the encoding in some places.
- Some of those instructions were using am2offset and now use addrmode2.
Codegen isn't affected, instructions which use SelectAddrMode2Offset were not
touched.
- Teach printAddrMode2Operand to check by the addressing mode which index
mode to print.
- This is a work in progress, more work to come. The idea is to change places
which use am2offset to use addrmode2 instead, as to unify assembly parser.
- Add testcases for assembly parser

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128585 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-30 23:32:32 +00:00
Cameron Zwarich
c0e6d780cd Add a ARM-specific SD node for VBSL so that forms with a constant first operand
can be recognized. This fixes <rdar://problem/9183078>.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128584 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-30 23:01:21 +00:00
Jim Grosbach
f5b23368d2 Tidy up. Whitespace and 80-columns.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128583 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-30 22:38:13 +00:00
Bill Wendling
31d244ead5 * The DSE code that tested for overlapping needed to take into account the fact
that one of the numbers is signed while the other is unsigned. This could lead
  to a wrong result when the signed was promoted to an unsigned int.

* Add the data layout line to the testcase so that it will test the appropriate
  thing.

Patch by David Terei!


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128577 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-30 21:37:19 +00:00
Akira Hatanaka
81bd78b897 fixed typo
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128574 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-30 21:15:35 +00:00
Jakob Stoklund Olesen
b516e00b43 Don't add the same analysis implementation pair twice.
Prevent infinite growth of the list.

Patch by José Fonseca!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128568 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-30 18:32:53 +00:00
Jakob Stoklund Olesen
03ef449917 Reset StringMap's NumTombstones on clears and rehashes.
StringMap was not properly updating NumTombstones after a clear or rehash.

This was not fatal until now because the table was growing faster than
NumTombstones could, but with the previous change of preventing infinite
growth of the table the invariant (NumItems + NumTombstones <= NumBuckets)
stopped being observed, causing infinite loops in certain situations.

Patch by José Fonseca!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128567 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-30 18:32:51 +00:00
Jakob Stoklund Olesen
e10fff6f88 Prevent infinite growth of SmallPtrSet instances.
Rehash but don't grow when full of tombstones.

Patch by José Fonseca!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128566 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-30 18:32:48 +00:00
Jakob Stoklund Olesen
aea4fe2862 Prevent infinite growth of SmallMap instances.
Rehash but don't grow when full of tombstones.

Patch by José Fonseca!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128565 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-30 18:32:44 +00:00
Jakob Stoklund Olesen
414fdbdb01 Prevent infinite growth of the DenseMap.
When the hash function uses object pointers all free entries eventually
become tombstones as they are used at least once, regardless of the size.

DenseMap cannot function with zero empty keys, so it double size to get
get ridof the tombstones.

However DenseMap never shrinks automatically unless it is cleared, so
the net result is that certain tables grow infinitely.

The solution is to make a fresh copy of the table without tombstones
instead of doubling size, by simply calling grow with the current size.

Patch by José Fonseca!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128564 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-30 18:32:41 +00:00
Jakob Stoklund Olesen
efe65ce25c Fix evil VirtRegRewriter bug.
The rewriter can keep track of multiple stack slots in the same register if they
happen to have the same value. When an instruction modifies a stack slot by
defining a register that is mapped to a stack slot, other stack slots in that
register are no longer valid.

This is a very rare problem, and I don't have a simple test case. I get the
impression that VirtRegRewriter knows it is about to be deleted, inventing a
last opaque problem.

<rdar://problem/9204040>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128562 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-30 18:14:07 +00:00
Jakob Stoklund Olesen
eb1c5dcb12 Teach VirtRegRewriter about the new virtual register numbers. No functional change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128561 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-30 18:14:04 +00:00
Benjamin Kramer
8398512f89 Avoid turning a floating point division with a constant power of two into a denormal multiplication.
Some platforms may treat denormals as zero, on other platforms multiplication
with a subnormal is slower than dividing by a normal.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128555 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-30 17:02:54 +00:00