Commit Graph

10458 Commits

Author SHA1 Message Date
Dan Gohman
5fa417c790 Fix memdep's code for reasoning about dependences between two calls. A Ref
response from getModRefInfo is not useful here. Instead, check for identical
calls only in the NoModRef case.

Reapply r110270, and strengthen it to compensate for the memdep changes.
When both calls are readonly, there is no dependence between them.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110382 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-05 22:09:15 +00:00
Devang Patel
baefea4498 Move x86 specific tests into test/CodeGen/X86.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110372 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-05 20:25:37 +00:00
Bob Wilson
a1d410d512 Add an ARM RSCrr instruction for disassembly only.
Partial fix for PR7792.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110361 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-05 18:59:36 +00:00
Bob Wilson
cff7178844 Add an ARM RSBrr instruction for disassembly only.
Partial fix for PR7792.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110358 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-05 18:23:43 +00:00
Dan Gohman
a34c885d51 Move x86-specific tests out of test/Transforms/LoopStrengthReduce and
into test/CodeGen/X86, so that they aren't run when the x86 target is
not enabled.

Fix uglygep.ll to not be x86-specific.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110343 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-05 17:04:15 +00:00
Daniel Dunbar
feaac8f7ff tests: CodeGen/X86/GC tests require X86.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110338 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-05 15:45:33 +00:00
Daniel Dunbar
079515f382 tests: Mark MC/AsmParser tests as requiring x86 for now -- almost all of them
rely on using a specific x86 triple to test what they want to test.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110337 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-05 15:44:15 +00:00
Rafael Espindola
6d50ed6321 check-lit was failing again on F13 64 bits :-(
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110311 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-05 03:35:01 +00:00
Dan Gohman
34fa82f7e8 Revert r110270 for now. It appears to uncover a memdep bug.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110293 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-05 00:43:10 +00:00
Bob Wilson
1d9125a6ff ARM "rrx" shift operands do not have an immediate. PR7790.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110292 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-05 00:34:42 +00:00
Dan Gohman
907857d746 The trouble with testing for "ModRef" and "NoModRef" is that
one is a suffix of the other, and FileCheck accepts superstrings.
Adjust the output to avoid this problem.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110280 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-04 23:37:55 +00:00
Bill Wendling
af33b7b4a7 The lower invoke pass needs to have unreachable code elimination run after it
because it could create such things. This fixes a MingW buildbot test failure.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110279 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-04 23:36:02 +00:00
Dan Gohman
3dcc91ee8c The two-callsite form of AliasAnalysis::getModRefInfo is documented
to return Ref if the left callsite only reads memory read or written
by the right callsite; fix BasicAliasAnalysis to implement this.

Add AliasAnalysisEvaluator support for testing the two-callsite
form of getModRefInfo.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110270 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-04 22:56:29 +00:00
Eli Friedman
7752442bfa PR7814: Truncates cannot be ignored for signed comparisons.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110268 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-04 22:40:58 +00:00
Stuart Hastings
009944e774 Test case for r110250. Radar 8264670. Test case by Fariborz Jahanian!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110254 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-04 22:05:38 +00:00
Bill Wendling
547b6ed16e Testcase for r110248.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110249 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-04 21:56:30 +00:00
Devang Patel
225764b7a5 Test case for combination of r110234 & r110235.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110238 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-04 18:42:46 +00:00
Dan Gohman
1e926f2d3a These tests are no longer stored in *CVS*.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110201 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-04 15:58:01 +00:00
Stuart Hastings
b884666a38 call-imm.ll test case regex fix. Patch by Dimitry Andric!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110199 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-04 15:31:35 +00:00
Kalle Raiskila
bc2697cca0 Make SPU backend handle insertelement and
store for "half vectors"


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110198 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-04 13:59:48 +00:00
Bob Wilson
67b453b0d1 Combine NEON VABD (absolute difference) intrinsics with ADDs to make VABA
(absolute difference with accumulate) intrinsics.  Radar 8228576.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110170 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-04 00:12:08 +00:00
Dan Gohman
0fd353376b Make instcombine set explicit alignments on load or store
instructions with alignment 0, so that subsequent passes don't
need to bother checking the TargetData ABI size manually.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110128 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-03 18:20:32 +00:00
Jakob Stoklund Olesen
0055f97f72 OK, that's it. This test is going away now. But don't worry, I am taking it to a
nice farm in the country where it can play with other tests. And bunnies.

It is not clear what is being tested, and the revision history shows a bunch of
random changes to the expected instruction count. Clearly, we are just fudging
it to pass whenever it fails.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110118 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-03 17:21:14 +00:00
Peter Collingbourne
3bababf880 Add an atomic lowering pass
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110113 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-03 16:19:16 +00:00
Michael J. Spencer
237f8fe5df MC: Fix symbol fragment offsets in COFF.
Patch by Cameron Esfahani!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110104 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-03 05:02:46 +00:00
Michael J. Spencer
a69494ee5e Revert "MC: Fix symbol fragment offsets in COFF."
This reverts commit r110100

Wrong path caps.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110103 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-03 04:53:28 +00:00
Michael J. Spencer
ef1f8c3cc3 MC: Fix symbol fragment offsets in COFF.
Patch by Cameron Esfahani!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110100 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-03 04:43:24 +00:00
Stuart Hastings
8eab2c2f54 Diabolical hack to make a test compatible with clang. (Thanks to Dale!) Radar 8246180.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110081 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-02 23:29:03 +00:00
Dan Gohman
a8afb2a623 Add a lint check for indirectbr with no successors.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110074 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-02 23:06:43 +00:00
Stuart Hastings
32a9e831c2 Testcase for r110043. Radar 8246180.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110070 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-02 22:09:53 +00:00
Kalle Raiskila
e1c9159f63 More SPU v2f32 stuff added: insertelement and shuffle.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110038 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-02 11:22:10 +00:00
Kalle Raiskila
c9fda996fc Add preliminary v2f32 support for SPU. Like with v2i32, we just
duplicate the instructions and operate on half vectors. 

Also reorder code in SPUInstrInfo.td for better coherency.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110037 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-02 10:25:47 +00:00
Owen Anderson
61378363b7 Re-apply the infamous r108614, with a fix pointed out by Dirk Steinke.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110036 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-02 09:32:13 +00:00
Kalle Raiskila
82fe467ca5 Add preliminary v2i32 support for SPU backend. As there are no
such registers in SPU, this support boils down to "emulating" 
them by duplicating instructions on the general purpose registers. 

This adds the most basic operations on v2i32: passing parameters,
addition, subtraction, multiplication and a few others.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110035 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-02 08:54:39 +00:00
Daniel Dunbar
27a9939903 Fix comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110006 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-02 01:25:20 +00:00
Daniel Dunbar
e3536b89b4 tests: Kill off custom targets which were just there for TestRunner.sh.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110003 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-02 00:52:44 +00:00
Daniel Dunbar
ed56fc363f tests: Deprecate TestRunner.sh, and have it just invoke 'llvm-lit' (which will
need to be in your path). Please move to using 'llvm-lit' if you are still using
TestRunner.sh.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110002 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-02 00:52:41 +00:00
Eli Friedman
e3837014d6 PR7781: Fix incorrect shifting in PPCTargetLowering::LowerBUILD_VECTOR.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@109998 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-02 00:18:19 +00:00
Daniel Dunbar
666b402243 tests: Make 'lit' the default test tool. You can still use 'make check-dg' to
run the tests using DejaGNU, but not for much longer. This is a last call for
DejaGNU supporters, if no one complains soon the DejaGNU support is going to
die.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@109997 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-02 00:05:18 +00:00
Eli Friedman
348e02600e PR7774: Fix undefined shifts in Alpha backend. As a bonus, this actually
improves the generated code in some cases.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@109985 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-01 21:13:28 +00:00
Bob Wilson
6945ef380c Revert new AVX intrinsic tests. They are breaking buildbots and Bruno is
away from a computer now.
--- Reverse-merging r109881 into '.':
D    test/CodeGen/X86/avx-intrinsics-x86.ll
D    test/CodeGen/X86/avx-intrinsics-x86_64.ll


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@109959 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-31 22:36:03 +00:00
Daniel Dunbar
10b173a1e7 Speculatively revert r108614, "Another attempt at getting the clang self-host to
like my instcombine patch.", in an attempt to fix Clang i386 bootstrap.
 - Also PR7719.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@109953 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-31 19:51:11 +00:00
Bob Wilson
98e1479575 Add support for disassembling VMVN (immediate) instructions. PR7747.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@109946 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-31 05:57:44 +00:00
Dale Johannesen
38cb1381b5 PPC doesn't supported VLA with large alignment. This was
formerly rejected by the FE, so asserted in the BE; now the FE only
warns, so we treat it as a legitimate fatal error in PPC BE.
This means the test for the feature won't pass, so it's xfail'd.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@109892 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-30 21:09:48 +00:00
Bruno Cardoso Lopes
74273b8822 A *bunch* of tests for AVX intrinsics
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@109881 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-30 19:57:56 +00:00
Bob Wilson
45f57a65b6 Attempt to fix the llvm-gcc-powerpc-darwin9 buildbot.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@109876 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-30 18:52:47 +00:00
Eli Friedman
ad78a88711 Fix for bug reported by Evzen Muller on llvm-commits: make sure to correctly
check the range of the constant when optimizing a comparison between a
constant and a sign_extend_inreg node.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@109854 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-30 06:44:31 +00:00
Jim Grosbach
6ccfc507dc Many Thumb2 instructions can reference the full ARM register set (i.e.,
have 4 bits per register in the operand encoding), but have undefined
behavior when the operand value is 13 or 15 (SP and PC, respectively).
The trivial coalescer in linear scan sometimes will merge a copy from
SP into a subsequent instruction which uses the copy, and if that
instruction cannot legally reference SP, we get bad code such as:
  mls r0,r9,r0,sp
instead of:
  mov r2, sp
  mls r0, r9, r0, r2

This patch adds a new register class for use by Thumb2 that excludes
the problematic registers (SP and PC) and is used instead of GPR
for those operands which cannot legally reference PC or SP. The
trivial coalescer explicitly requires that the register class
of the destination for the COPY instruction contain the source
register for the COPY to be considered for coalescing. This prevents
errant instructions like that above.

PR7499




git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@109842 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-30 02:41:01 +00:00
Eric Christopher
07f4eda3a5 Fix this up per llvm-gcc r109819.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@109820 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-29 23:20:29 +00:00
Benjamin Kramer
38ca1078b4 Remove XFAIL, test doesn't leak anymore.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@109801 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-29 20:36:36 +00:00