1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-31 16:02:52 +01:00
Commit Graph

500 Commits

Author SHA1 Message Date
Daniel Dunbar
acd46f61df Workaround SCEV non-determinism on this test, for now, to get buildbots back to
green. Dan, please revert this once the real problem is fixed.

llvm-svn: 105732
2010-06-09 17:54:40 +00:00
Dan Gohman
65954b2eb4 Optimize ScalarEvolution's SCEVComplexityCompare predicate: don't go
scrounging through SCEVUnknown contents and SCEVNAryExpr operands;
instead just do a simple deterministic comparison of the precomputed
hash data.

Also, since this is more precise, it eliminates the need for the slow
N^2 duplicate detection code.

llvm-svn: 105540
2010-06-07 19:06:13 +00:00
Dan Gohman
9c1b7fdc46 Add a comment to this test.
llvm-svn: 102387
2010-04-26 21:37:43 +00:00
Dan Gohman
231fe284cd ScalarEvolution support for <= and >= loops.
Also, generalize ScalarEvolutions's min and max recognition to handle
some new forms of min and max that this change makes more common.

llvm-svn: 102234
2010-04-24 03:09:42 +00:00
Chris Lattner
790231f95e fix some failures my callgraph dump format change broke.
llvm-svn: 102197
2010-04-23 18:38:40 +00:00
Dan Gohman
31d6b29bae Don't attempt to analyze values which are obviously undef. This fixes some
assertion failures in extreme cases.

llvm-svn: 102042
2010-04-22 01:35:11 +00:00
Dan Gohman
e87da8a25d Generalize ScalarEvolution's PHI analysis to handle loops that don't
have preheaders or dedicated exit blocks, as clients may not otherwise
need to run LoopSimplify.

llvm-svn: 101030
2010-04-12 07:49:36 +00:00
Dan Gohman
b2a87fa39d Pointers to zero-sized objects don't point to overlapping objects.
llvm-svn: 100789
2010-04-08 18:11:50 +00:00
Chris Lattner
23334439e9 add newlines at the end of files.
llvm-svn: 100705
2010-04-07 22:53:17 +00:00
Mon P Wang
484bbe6aa9 Reapply address space patch after fixing an issue in MemCopyOptimizer.
Added support for address spaces and added a isVolatile field to memcpy, memmove, and memset,
e.g., llvm.memcpy.i32(i8*, i8*, i32, i32) -> llvm.memcpy.p0i8.p0i8.i32(i8*, i8*, i32, i32, i1)

llvm-svn: 100304
2010-04-04 03:10:48 +00:00
Mon P Wang
0ccf050ca3 Revert r100191 since it breaks objc in clang
llvm-svn: 100199
2010-04-02 18:43:02 +00:00
Mon P Wang
a01350755e Reapply address space patch after fixing an issue in MemCopyOptimizer.
Added support for address spaces and added a isVolatile field to memcpy, memmove, and memset,
e.g., llvm.memcpy.i32(i8*, i8*, i32, i32) -> llvm.memcpy.p0i8.p0i8.i32(i8*, i8*, i32, i32, i1)

llvm-svn: 100191
2010-04-02 18:04:15 +00:00
Bob Wilson
aae933cc81 Revert Mon Ping's change 99928, since it broke all the llvm-gcc buildbots.
llvm-svn: 99948
2010-03-30 22:27:04 +00:00
Mon P Wang
9351ea594a Added support for address spaces and added a isVolatile field to memcpy, memmove, and memset,
e.g., llvm.memcpy.i32(i8*, i8*, i32, i32) -> llvm.memcpy.p0i8.p0i8.i32(i8*, i8*, i32, i32, i1)
A update of langref will occur in a subsequent checkin.

llvm-svn: 99928
2010-03-30 20:55:56 +00:00
Dan Gohman
a722142044 Avoid analyzing instructions in blocks not reachable from the entry block.
They are lots of trouble, and they don't matter. This fixes PR6559.

llvm-svn: 98103
2010-03-09 23:46:50 +00:00
Chris Lattner
5649a97c00 remove andersen's tests.
llvm-svn: 97490
2010-03-01 20:23:15 +00:00
Dan Gohman
9268d67079 Teach ScalarEvolution how to compute a tripcount for a loop with
true or false as its exit condition. These are usually eliminated by
SimplifyCFG, but the may be left around during a pass which wishes
to preserve the CFG.

llvm-svn: 96683
2010-02-19 18:12:07 +00:00
Dan Gohman
71fc5e8fce -disable-output is no longer needed with -analyze.
llvm-svn: 94574
2010-01-26 19:25:59 +00:00
Dan Gohman
5e06a05a16 Fix the the ceiling-division used in computing the MaxBECount so that it doesn't
have trouble with an intermediate add overflowing. Also, be more conservative
about the case where the induction variable in an SLT loop exit can step past
the RHS of the SLT and overflow in a single step.

Make getSignedRange more aggressive, to recover for some common cases which
the above fixes pessimized.

This addresses rdar://7561161.

llvm-svn: 94512
2010-01-26 04:40:18 +00:00
Tobias Grosser
24e0f1ed17 Fix PR6047
Nodes that had children outside of the post dominator tree (infinite loops)
where removed from the post dominator tree. This seems to be wrong. Leave them
in the tree.

llvm-svn: 93633
2010-01-16 13:38:07 +00:00
Dan Gohman
771144e807 Use WriteAsOperand instead of getName() to print loop header names,
so that unnamed blocks are handled.

llvm-svn: 93059
2010-01-09 18:17:45 +00:00
Dan Gohman
5fa04f2707 Delete useless trailing semicolons.
llvm-svn: 92740
2010-01-05 17:55:26 +00:00
Chris Lattner
6d180b4a2c gvn is optimizing this better now.
llvm-svn: 90696
2009-12-06 04:16:05 +00:00
Dan Gohman
6bb055cfcd Add a comment about A[i+(j+1)].
llvm-svn: 90185
2009-12-01 01:38:10 +00:00
Chris Lattner
911e5047d0 @test9 is a testcase for r89958. Before 89958, we misanalyzed the
first expression as P+4+4*i which we considered to possibly alias
P+4*j.  Now we correctly analyze the former one as P+1+4*i.

@test10 is a sanity test that verfies that we know that P+4+4*i != P+4*i.

llvm-svn: 89960
2009-11-26 19:25:46 +00:00
Chris Lattner
ce573daf09 Implement PR1143 (at -m64) by making basicaa look through extensions. We
previously already handled it at -m32 because there were no i32->i64 
extensions for addressing.

llvm-svn: 89959
2009-11-26 18:53:33 +00:00
Chris Lattner
d86a693b70 teach GetLinearExpression to be a bit more aggressive.
llvm-svn: 89955
2009-11-26 17:00:01 +00:00
Chris Lattner
993cb8c911 update status of this. basicaa is much improved now,
only missing the one form (in this testcase).  Dan, do you
consider this example to be important?

llvm-svn: 89953
2009-11-26 16:42:00 +00:00
Chris Lattner
9c88c96b3f Teach basicaa that x|c == x+c when the c bits of x are clear. This
allows us to compile the example in readme.txt into:

LBB1_1:                                                     ## %bb
	movl	4(%rdx,%rax), %ecx
	movl	%ecx, %esi
	imull	(%rdx,%rax), %esi
	imull	%esi, %ecx
	movl	%esi, 8(%rdx,%rax)
	imull	%ecx, %esi
	movl	%ecx, 12(%rdx,%rax)
	movl	%esi, 16(%rdx,%rax)
	imull	%ecx, %esi
	movl	%esi, 20(%rdx,%rax)
	addq	$16, %rax
	cmpq	$4000, %rax
	jne	LBB1_1

instead of:

LBB1_1: 
	movl	(%rdx,%rax), %ecx
	imull	4(%rdx,%rax), %ecx
	movl	%ecx, 8(%rdx,%rax)
	imull	4(%rdx,%rax), %ecx
	movl	%ecx, 12(%rdx,%rax)
	imull	8(%rdx,%rax), %ecx
	movl	%ecx, 16(%rdx,%rax)
	imull	12(%rdx,%rax), %ecx
	movl	%ecx, 20(%rdx,%rax)
	addq	$16, %rax
	cmpq	$4000, %rax
	jne	LBB1_1

GCC (4.2) doesn't seem to be able to eliminate the loads in this 
testcase either, it generates:

L2:
	movl	(%rdx), %eax
	imull	4(%rdx), %eax
	movl	%eax, 8(%rdx)
	imull	4(%rdx), %eax
	movl	%eax, 12(%rdx)
	imull	8(%rdx), %eax
	movl	%eax, 16(%rdx)
	imull	12(%rdx), %eax
	movl	%eax, 20(%rdx)
	addl	$4, %ecx
	addq	$16, %rdx
	cmpl	$1002, %ecx
	jne	L2

llvm-svn: 89952
2009-11-26 16:26:43 +00:00
Chris Lattner
677b93d4c8 teach basicaa that A[i] != A[i+1].
llvm-svn: 89951
2009-11-26 16:18:10 +00:00
Chris Lattner
82257f0385 rename test
llvm-svn: 89950
2009-11-26 16:08:41 +00:00
Chris Lattner
69e59e50f3 Change the other half of aliasGEP (which handles GEP differencing) to use DecomposeGEPExpression. This dramatically simplifies and shrinks the code by eliminating the horrible CheckGEPInstructions method, fixes a miscompilation (@test3) and makes the code more aggressive. In particular, we now handle the @test4 case, which is reduced from the SmallPtrSet constructor. Missing this caused us to emit a variable length memset instead of a fixed size one.
llvm-svn: 89922
2009-11-26 02:17:34 +00:00
Chris Lattner
862a3532d6 add a new random feature test
llvm-svn: 89921
2009-11-26 02:16:28 +00:00
Chris Lattner
632f60ccc9 remove a silly condition that doesn't make a lot of sense anymore.
llvm-svn: 89601
2009-11-22 16:15:59 +00:00
Victor Hernandez
3c98070f2c Update computeArraySize() to use ComputeMultiple() to determine the array size associated with a malloc; also extend PerformHeapAllocSRoA() to check if the optimized malloc's arg had its highest bit set, so that it is safe for ComputeMultiple() to look through sext instructions while determining the optimized malloc's array size
llvm-svn: 86676
2009-11-10 08:32:25 +00:00
Victor Hernandez
8736a8fca4 Re-commit r86077 now that r86290 fixes the 179.art and 175.vpr ARM regressions.
Here is the original commit message:

This commit updates malloc optimizations to operate on malloc calls that have constant int size arguments.

Update CreateMalloc so that its callers specify the size to allocate:
MallocInst-autoupgrade users use non-TargetData-computed allocation sizes.
Optimization uses use TargetData to compute the allocation size.

Now that malloc calls can have constant sizes, update isArrayMallocHelper() to use TargetData to determine the size of the malloced type and the size of malloced arrays.
Extend getMallocType() to support malloc calls that have non-bitcast uses.

Update OptimizeGlobalAddressOfMalloc() to optimize malloc calls that have non-bitcast uses.  The bitcast use of a malloc call has to be treated specially here because the uses of the bitcast need to be replaced and the bitcast needs to be erased (just like the malloc call) for OptimizeGlobalAddressOfMalloc() to work correctly.

Update PerformHeapAllocSRoA() to optimize malloc calls that have non-bitcast uses.  The bitcast use of the malloc is not handled specially here because ReplaceUsesOfMallocWithGlobal replaces through the bitcast use.

Update OptimizeOnceStoredGlobal() to not care about the malloc calls' bitcast use.

Update all globalopt malloc tests to not rely on autoupgraded-MallocInsts, but instead use explicit malloc calls with correct allocation sizes.

llvm-svn: 86311
2009-11-07 00:16:28 +00:00
Victor Hernandez
a5a12cd62e Revert r86077 because it caused crashes in 179.art and 175.vpr on ARM
llvm-svn: 86213
2009-11-06 01:33:24 +00:00
Victor Hernandez
21ec158c23 Update CreateMalloc so that its callers specify the size to allocate:
MallocInst-autoupgrade users use non-TargetData-computed allocation sizes.
Optimization uses use TargetData to compute the allocation size.

Now that malloc calls can have constant sizes, update isArrayMallocHelper() to use TargetData to determine the size of the malloced type and the size of malloced arrays.
Extend getMallocType() to support malloc calls that have non-bitcast uses.

Update OptimizeGlobalAddressOfMalloc() to optimize malloc calls that have non-bitcast uses.  The bitcast use of a malloc call has to be treated specially here because the uses of the bitcast need to be replaced and the bitcast needs to be erased (just like the malloc call) for OptimizeGlobalAddressOfMalloc() to work correctly.

Update PerformHeapAllocSRoA() to optimize malloc calls that have non-bitcast uses.  The bitcast use of the malloc is not handled specially here because ReplaceUsesOfMallocWithGlobal replaces through the bitcast use.

Update OptimizeOnceStoredGlobal() to not care about the malloc calls' bitcast use.

Update all globalopt malloc tests to not rely on autoupgraded-MallocInsts, but instead use explicit malloc calls with correct allocation sizes.

llvm-svn: 86077
2009-11-05 00:03:03 +00:00
Kenneth Uildriks
e711736014 Make opt default to not adding a target data string and update tests that depend on target data to supply it within the test
llvm-svn: 85900
2009-11-03 15:29:06 +00:00
Edward O'Callaghan
142c2fc0ae Convert Analysis tests to FileCheck in regards to PR5307.
llvm-svn: 85241
2009-10-27 14:54:46 +00:00
Dan Gohman
c8468855a7 Teach BasicAA how to analyze Select instructions, and make it more
aggressive on PHI instructions.

llvm-svn: 85158
2009-10-26 21:55:43 +00:00
Dan Gohman
06621f065a Update these tests to match what Loop::print now prints.
llvm-svn: 85021
2009-10-24 23:52:07 +00:00
Chris Lattner
eebed04912 fix test
llvm-svn: 84405
2009-10-18 05:03:00 +00:00
Chris Lattner
8db424f027 tighten up test3, add test3a for the converse
transform, which isn't happening yet.

llvm-svn: 84402
2009-10-18 04:55:26 +00:00
Chris Lattner
aff491fab6 tighten test2, add a test that it doesn't get transformed in the invalid edge case.
llvm-svn: 84401
2009-10-18 04:50:18 +00:00
Nick Lewycky
671ca16583 Merge tests into modref.ll. Also add a test for r84174 at Chris' behest!
llvm-svn: 84400
2009-10-18 04:41:36 +00:00
Nick Lewycky
d21c892821 Add a couple new testcases.
llvm-svn: 84385
2009-10-18 00:42:07 +00:00
Chris Lattner
1ed26100a4 replace a useless test with a useful one
llvm-svn: 84383
2009-10-17 23:59:51 +00:00
Nick Lewycky
0498027dec Make use of the result of the loads even though that means adding -instcombine.
llvm-svn: 84125
2009-10-14 19:02:13 +00:00
Evan Cheng
907fd9c3fb Another BasicAA fix. If a value does not alias a GEP's base pointer, then it
cannot alias the GEP. GEP pointer alias rule states this clearly:
A pointer value formed from a getelementptr instruction is associated with the
addresses associated with the first operand of the getelementptr.

llvm-svn: 84079
2009-10-14 06:41:49 +00:00
Evan Cheng
b193d9df99 Replace test with a simpler hand crafted one.
llvm-svn: 84069
2009-10-14 01:45:10 +00:00
Evan Cheng
3130aba687 Teach basic AA about PHI nodes. If all operands of a phi NoAlias another value than it's safe to declare the PHI NoAlias the value. Ditto for MustAlias.
llvm-svn: 84038
2009-10-13 22:02:20 +00:00
Chris Lattner
a433d8cb67 don't use dead loads as tests.
llvm-svn: 83985
2009-10-13 17:39:29 +00:00
Nick Lewycky
72f185f281 Teach BasicAA a little something about the atomic intrinsics: they can only
modify through the pointer they're given.

llvm-svn: 83959
2009-10-13 07:48:38 +00:00
Victor Hernandez
f772c0b8b2 Revert 82694 "Auto-upgrade malloc instructions to malloc calls." because it causes regressions in the nightly tests.
llvm-svn: 82784
2009-09-25 18:11:52 +00:00
Victor Hernandez
ff8027ced6 Auto-upgrade malloc instructions to malloc calls.
Reviewed by Devang Patel.

llvm-svn: 82694
2009-09-24 17:47:49 +00:00
Dan Gohman
723717853b Teach ScalarEvolution how to reason about no-wrap flags on loops
where the induction variable has a non-unit stride, such as {0,+,2}, and
there are expressions such as {1,+,2} inside the loop formed with
or or add nsw operators.

llvm-svn: 82151
2009-09-17 18:05:20 +00:00
Dan Gohman
c9307490ce Eliminate more redundant llvm-as calls.
llvm-svn: 81540
2009-09-11 18:17:12 +00:00
Dan Gohman
205b641954 Change tests from "opt %s" to "opt < %s" so that opt doesn't see the
input filename so that opt doesn't print the input filename in the
output so that grep lines in the tests don't unintentionally match
strings in the input filename.

llvm-svn: 81537
2009-09-11 18:01:28 +00:00
Dan Gohman
142428ce64 Eliminate more uses of llvm-as and llvm-dis.
llvm-svn: 81293
2009-09-09 00:09:15 +00:00
Dan Gohman
a3ab9b3b9e Convert a few more opt | llvm-dis to opt -S.
llvm-svn: 81261
2009-09-08 22:41:33 +00:00
Dan Gohman
c95df8b6d8 Use opt -S instead of piping bitcode output through llvm-dis.
llvm-svn: 81257
2009-09-08 22:34:10 +00:00
Dan Gohman
8d84372836 Change these tests to feed the assembly files to opt directly, instead
of using llvm-as, now that opt supports this.

llvm-svn: 81226
2009-09-08 16:50:01 +00:00
Andreas Neustifter
d842a676d6 Updated tests to use ProfileVerifer to test ProfileLoader and ProfileEstimator.
(Keep disabled test disabled until selfhosted build issue is resolved.)

llvm-svn: 81008
2009-09-04 17:21:59 +00:00
Daniel Dunbar
e1e99bbb67 Revert "--- Reverse-merging r80908 into '.':", I already "fixed" this.
llvm-svn: 80970
2009-09-03 23:40:10 +00:00
Bill Wendling
faf3fd8b85 --- Reverse-merging r80908 into '.':
D    test/Analysis/Profiling

--- Reverse-merging r80907 into '.':
U    lib/Analysis/ProfileInfoLoaderPass.cpp

Attempt to remove failure in the self-hosting build bot.

llvm-svn: 80966
2009-09-03 23:13:46 +00:00
Daniel Dunbar
23700a97d2 Disable some parts of the profiling-tool-chain test, which is currently failing
on a self-hosted build (although it seems to work on non-self hosted). I'll work
with Andreas to figure this out.

llvm-svn: 80947
2009-09-03 21:09:53 +00:00
Daniel Dunbar
eed0ecc21a Reapply profiling tests.
llvm-svn: 80908
2009-09-03 07:38:00 +00:00
Andreas Neustifter
3418f26ee5 Removed temporarily because of breaking Darwin builds.
(See http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090831/086214.html)

llvm-svn: 80799
2009-09-02 16:47:24 +00:00
Andreas Neustifter
3b25ffcef0 Changed profiling-tool-chain.ll test to use optimal-edge-profiling instead of
edge-profiling, this is more useful since the loading of the
optimal-edge-profiling is more complicated.
The edge-profiling is tested in edge-profiling.ll where only the
instrumentation is tested.

llvm-svn: 80791
2009-09-02 14:24:08 +00:00
Daniel Dunbar
3b7954c9bf Don't force the triple or data layout in this test. We just have to get them
from the host and hope that works.

llvm-svn: 80751
2009-09-02 02:43:11 +00:00
Chris Lattner
4c06cf3a4f rename test so that name reflects what it is testing for.
llvm-svn: 80519
2009-08-30 21:36:39 +00:00
Chris Lattner
0ad4b1ca5e convert to filecheck format.
llvm-svn: 80518
2009-08-30 21:36:06 +00:00
Torok Edwin
f1fbd8af64 rm needs -f
llvm-svn: 80363
2009-08-28 14:05:07 +00:00
Torok Edwin
b9c2b743ea Remove the llvmprof.out from the test output, otherwise running
make check in a non-clean directory causes it to fail (for example when running
make check twice), since execution counts will differ.

llvm-svn: 80362
2009-08-28 13:35:44 +00:00
Andreas Neustifter
c2d92270a2 Remove profiling output file because two consecutive runs of make check give
error.

llvm-svn: 80357
2009-08-28 10:38:26 +00:00
Andreas Neustifter
e4c8ff4447 Removed unnecessary file creation during test.
llvm-svn: 80356
2009-08-28 10:07:41 +00:00
Andreas Neustifter
da88343196 Pulled all tests into one test. Removed some redundant tests. Rename.
llvm-svn: 80355
2009-08-28 10:00:28 +00:00
Andreas Neustifter
233f8ec4db Readded test from r79615, this tests the complete profiling tool chain. Furhter
tests can test only parts of this system.

llvm-svn: 80348
2009-08-28 06:41:00 +00:00
Dan Gohman
cfb6234d13 Create a ScalarEvolution-based AliasAnalysis implementation.
This is a simple AliasAnalysis implementation which works by making
ScalarEvolution queries. ScalarEvolution has a more complete understanding
of arithmetic than BasicAA's collection of ad-hoc checks, so it handles
some cases that BasicAA misses, for example p[i] and p[i+1] within the
same iteration of a loop.

This is currently experimental. It may be that the main use for this pass
will be to help find cases where BasicAA can be profitably extended, or
to help in the development of the overall AliasAnalysis infrastructure,
however it's also possible that it could grow up to become a directly
useful pass.

llvm-svn: 80098
2009-08-26 14:53:06 +00:00
Andreas Neustifter
fb55ceecf3 Removed profiling test, lli not available on all platforms.
llvm-svn: 79633
2009-08-21 15:27:35 +00:00
Andreas Neustifter
b11ab6b931 Added tests for Profiling Infrastructure.
llvm-svn: 79615
2009-08-21 09:36:28 +00:00
Dan Gohman
7965c57a0f Loosen up the regex for this test so that it doesn't implicitly
depend on TargetData information.

llvm-svn: 79491
2009-08-19 23:19:36 +00:00
Dan Gohman
9f80d2be6b Make LLVM Assembly dramatically easier to read by aligning the comments,
using formatted_raw_ostream's PadToColumn.

Before:

bb1:            ; preds = %bb
  %2 = sext i32 %i.01 to i64            ; <i64> [#uses=1]
  %3 = getelementptr double* %p, i64 %2         ; <double*> [#uses=1]
  %4 = load double* %3, align 8         ; <double> [#uses=1]
  %5 = fmul double %4, 1.100000e+00             ; <double> [#uses=1]
  %6 = sext i32 %i.01 to i64            ; <i64> [#uses=1]
  %7 = getelementptr double* %p, i64 %6         ; <double*> [#uses=1]

After:

bb1:                                        ; preds = %bb
  %2 = sext i32 %i.01 to i64                ; <i64> [#uses=1]
  %3 = getelementptr double* %p, i64 %2     ; <double*> [#uses=1]
  %4 = load double* %3, align 8             ; <double> [#uses=1]
  %5 = fmul double %4, 1.100000e+00         ; <double> [#uses=1]
  %6 = sext i32 %i.01 to i64                ; <i64> [#uses=1]
  %7 = getelementptr double* %p, i64 %6     ; <double*> [#uses=1]

Several tests required whitespace adjustments.

llvm-svn: 78816
2009-08-12 17:23:50 +00:00
Andreas Bolka
facd1be2d6 Add another Strong-SIV testcase.
llvm-svn: 78446
2009-08-08 00:21:49 +00:00
Andreas Bolka
8c032970b9 Fix Strong-SIV testcase.
llvm-svn: 78384
2009-08-07 15:42:32 +00:00
Andreas Bolka
4d59f26912 ZIV tester for LDA.
llvm-svn: 78157
2009-08-05 04:26:05 +00:00
Andreas Bolka
1532dc3110 Fix LDA testcases.
llvm-svn: 78153
2009-08-05 04:03:29 +00:00
Andreas Bolka
d713824363 Expand LDA testcases.
llvm-svn: 77926
2009-08-02 23:28:14 +00:00
Andreas Bolka
7d6b4c24e6 Slightly reformat LDA tests to ease grepping.
llvm-svn: 77398
2009-07-28 23:40:40 +00:00
Dan Gohman
39c0d3b10a Change the assembly syntax for nsw, nuw, and exact, putting them
after their associated opcodes rather than before. This makes them
a little easier to read.

llvm-svn: 77194
2009-07-27 16:11:46 +00:00
Dan Gohman
c126330ef1 When attempting to sign-extend an addrec by interpreting
the step value as unsigned, the start value and the addrec
itself still need to be treated as signed.

llvm-svn: 77078
2009-07-25 16:03:30 +00:00
Dan Gohman
13aabe2ba4 Teach ScalarEvolution to make use of no-overflow flags when
analyzing add recurrences.

llvm-svn: 77034
2009-07-25 01:22:26 +00:00
Andreas Bolka
fa70ad653b FileCheck'ize and expand LDA testcases.
llvm-svn: 76880
2009-07-23 15:56:53 +00:00
Dan Gohman
190d29cb5c Replace the original ad-hoc code for determining whether (v pred w) implies
(x pred y) with more thorough code that does more complete canonicalization
before resorting to range checks. This helps it find more cases where
the canonicalized expressions match.

llvm-svn: 76671
2009-07-21 23:03:19 +00:00
Dan Gohman
7a836d6400 Add a testcase for PR4569, which is now fixed.
llvm-svn: 76526
2009-07-21 00:50:52 +00:00
Torok Edwin
bbccd95874 Introduce a pointertracking pass.
For now this only computes the allocated size of the memory pointed to by a
pointer, and offset a pointer from allocated pointer.
The actual checkLimits part will come later, after another round of review.

llvm-svn: 75657
2009-07-14 18:44:28 +00:00
Dan Gohman
5d53d0b73b Add testcases for PR4538, PR4537, and PR4534.
llvm-svn: 75533
2009-07-13 22:30:31 +00:00
Nick Lewycky
d9573d899d When comparing constants, consider a less wide constant to be "less complex"
than a wider one, before trying to compare their contents which will crash
if their sizes are different.

llvm-svn: 74792
2009-07-04 17:24:52 +00:00
Andreas Bolka
71ef361d1d Array accesses are independent if the underlying arrays differ.
llvm-svn: 74499
2009-06-30 02:12:10 +00:00
Andreas Bolka
bed5622fb7 Print pairwise dependence results, add testcases.
llvm-svn: 74402
2009-06-28 00:35:22 +00:00
Dan Gohman
60844e20f0 Add a testcase demoing some of ScalarEvolution's new trip count logic.
llvm-svn: 74049
2009-06-24 01:22:30 +00:00
Dan Gohman
058143a236 Fix a bug in the trip-count computation with And/Or. If either of the
sides is CouldNotCompute, the resulting exact count must be CouldNotCompute.

llvm-svn: 73920
2009-06-22 23:28:56 +00:00
Dan Gohman
b85ac02513 Fix llvm::ComputeNumSignBits to handle pointer types
conservatively correctly, instead of aborting.

llvm-svn: 73908
2009-06-22 22:02:32 +00:00
Dan Gohman
db7860aa40 Teach ScalarEvolution how to analyze loops with multiple exit
blocks, and also exit blocks with multiple conditions (combined
with (bitwise) ands and ors). It's often infeasible to compute an
exact trip count in such cases, but a useful upper bound can often
be found.

llvm-svn: 73866
2009-06-22 00:31:57 +00:00
Dan Gohman
eb0a67278b Fix ScalarEvolution's backedge-taken count computations to check for
overflow when computing a integer division to round up.

Thanks to Nick Lewycky for noticing this!

llvm-svn: 73862
2009-06-21 23:46:38 +00:00
Dan Gohman
e72fd546a2 Teach ScalarEvolution how to recognize another xor(and(x, C), C) case.
If C is a single bit and the and gets analyzed as a truncate and
zero-extend, the xor can be represnted as an add.

llvm-svn: 73664
2009-06-18 00:00:20 +00:00
Dan Gohman
50b7d0d843 Add -disable-output to a bunch of tests that don't care about the output.
llvm-svn: 73633
2009-06-17 20:56:26 +00:00
Dan Gohman
473789f75a Fix ScalarEvolution's Xor handling to not assume that an And
that gets recognized with a SCEVZeroExtendExpr must be an And
with a low-bits mask. With r73540, this is no longer the case.

llvm-svn: 73594
2009-06-17 01:22:39 +00:00
Dan Gohman
5f6f8101d5 Split the Add, Sub, and Mul instruction opcodes into separate
integer and floating-point opcodes, introducing
FAdd, FSub, and FMul.

For now, the AsmParser, BitcodeReader, and IRBuilder all preserve
backwards compatability, and the Core LLVM APIs preserve backwards
compatibility for IR producers. Most front-ends won't need to change
immediately.

This implements the first step of the plan outlined here:
http://nondot.org/sabre/LLVMNotes/IntegerOverflow.txt

llvm-svn: 72897
2009-06-04 22:49:04 +00:00
Dan Gohman
5dd6f54a5f Teach BasicAliasAnalysis to understand constant gep indices that fall
beyond their associated static array type.

I believe that this fixes a legitimate bug, because BasicAliasAnalysis
already has code to check for this condition that works for non-constant
indices, however it was missing the case of constant indices. With this
change, it checks for both.

This fixes PR4267, and miscompiles of SPEC 188.ammp and 464.h264.href.

llvm-svn: 72451
2009-05-27 01:48:27 +00:00
Dan Gohman
592d65ba06 Teach ScalarEvolution to recognize x^-1 in the case where non-demanded
bits have been stripped out by instcombine.

llvm-svn: 72010
2009-05-18 16:29:04 +00:00
Dan Gohman
2c9bd7e0cb Make ScalarEvolution::isLoopGuardedByCond work even when the edge
entering a loop is a non-split critical edge.

llvm-svn: 72004
2009-05-18 15:36:09 +00:00
Dan Gohman
904f081ce7 Add nounwind to a few tests.
llvm-svn: 72002
2009-05-18 15:16:49 +00:00
Eli Friedman
aec1764402 Allow scalar evolution to compute iteration counts for loops with a
pointer-based condition.  This fixes PR3171.

llvm-svn: 71354
2009-05-09 12:32:42 +00:00
Dan Gohman
141989d3c2 Fix bogus overflow checks by replacing them with actual
overflow checks.

llvm-svn: 71284
2009-05-08 23:11:16 +00:00
Dan Gohman
603f022049 Fold trunc casts into add-recurrence expressions, allowing the
add-recurrence to be exposed. Add a new SCEV folding rule to
help simplify expressions in the presence of these extra truncs.

llvm-svn: 71264
2009-05-08 21:03:19 +00:00
Dan Gohman
0dc2b769b0 When printing a SCEVUnknown with pointer type, don't print an
artificial "ptrtoint", as it tends to clutter up complicated
expressions. The cast operators now print both source and
destination types, which is usually sufficient.

llvm-svn: 70554
2009-05-01 17:02:22 +00:00
Dan Gohman
3c9f4f765c Extend ScalarEvolution's getBackedgeTakenCount to be able to
compute an upper-bound value for the trip count, in addition to
the actual trip count. Use this to allow getZeroExtendExpr and
getSignExtendExpr to fold casts in more cases.

This may eventually morph into a more general value-range
analysis capability; there are certainly plenty of places where
more complete value-range information would allow more folding.

llvm-svn: 70509
2009-04-30 20:47:05 +00:00
Dan Gohman
06aff30f01 Generalize the cast-of-addrec folding to handle folding of SCEVs like
(sext i8 {-128,+,1} to i64) to i64 {-128,+,1}, where the iteration
crosses from negative to positive, but is still safe if the trip
count is within range.

llvm-svn: 70421
2009-04-29 22:28:28 +00:00
Dan Gohman
9fa631c81a Fix this test to match the new output from scalar-evolution.
llvm-svn: 70410
2009-04-29 21:06:20 +00:00
Dan Gohman
55befacc69 Include the source type in SCEV cast expression debug output, and
print sext, zext, and trunc, instead of signextend, zeroextend,
and truncate, respectively, for consistency with the main IR.

llvm-svn: 70405
2009-04-29 20:27:52 +00:00
Dan Gohman
211c5de27d Fix a grammaro in a comment.
llvm-svn: 70331
2009-04-28 21:54:23 +00:00
Nick Lewycky
32cfba44df Simplify trunc(extend(x)) in SCEVs, just for completeness. Also fix some odd
whitespace in the same file.

llvm-svn: 69870
2009-04-23 05:15:08 +00:00
Owen Anderson
0c0498a365 Testcase for PR3909.
llvm-svn: 69868
2009-04-23 04:33:42 +00:00
Dan Gohman
19990f2310 When turning (ashr(shl(x, n), n)) into sext(trunc(x)), the width of the
type to truncate to should be the number of bits of the value that are
preserved, not the number that are clobbered with sign-extension.
This fixes regressions in ldecod.

llvm-svn: 69704
2009-04-21 20:18:36 +00:00
Chris Lattner
f8afc9ecda change this to test for an alias result more directly.
llvm-svn: 67046
2009-03-16 18:28:27 +00:00
Nick Lewycky
d271c19fea Add a replacement for 2009-02-12-GEPNoalias.ll that works without -debug.
llvm-svn: 67011
2009-03-14 19:40:09 +00:00
Chris Lattner
5972bc6234 remove a buggy test, it is not ok to use -debug in RUN line.
llvm-svn: 66918
2009-03-13 18:19:34 +00:00
Dan Gohman
06c6037d9f Update this test for the LoopInfo::print changes.
llvm-svn: 65597
2009-02-27 00:17:49 +00:00
Dan Gohman
dd9c79d45b Rename ScalarEvolution's getIterationCount to getBackedgeTakenCount,
to more accurately describe what it does. Expand its doxygen comment
to describe what the backedge-taken count is and how it differs
from the actual iteration count of the loop. Adjust names and
comments in associated code accordingly.

llvm-svn: 65382
2009-02-24 18:55:53 +00:00
Nick Lewycky
f6c8dc0825 BasicAA was making the assumption that a local allocation which hadn't escaped
couldn't ever be the return of call instruction. However, it's quite possible
that said local allocation is itself the return of a function call. That's
what malloc and calloc are for, actually.

llvm-svn: 64442
2009-02-13 07:06:27 +00:00
Owen Anderson
c418b82207 Finish making AliasAnalysis aware of the fact that most atomic intrinsics only dereference their arguments, and enhance
BasicAA to make use of this fact when computing ModRef info.

llvm-svn: 63718
2009-02-04 05:16:46 +00:00
Nick Lewycky
911bb6122c Wind SCEV back in time, to Nov 18th. This 'fixes' PR3275, PR3294, PR3295,
PR3296 and PR3302.

llvm-svn: 62160
2009-01-13 09:18:58 +00:00
Nick Lewycky
2c01a8db3d Don't try to analyze this "backward" case. This is overly conservative
pending a correct solution.

llvm-svn: 61589
2009-01-02 18:54:17 +00:00
Nick Lewycky
bb69bd55a4 Check that the function prototypes are correct before assuming that the
parameters are pointers.

llvm-svn: 61451
2008-12-27 16:20:53 +00:00
Nick Lewycky
8f96b51785 Resubmit support for the 'nocapture' attribute.
The problematic part of this patch is that we were out of attribute bits,
requiring some fancy bit hacking to make it fit (by shrinking alignment)
without breaking existing users or the file format.

This change will require users to rebuild llvm-gcc to match llvm.

llvm-svn: 61239
2008-12-19 06:39:12 +00:00
Bill Wendling
905350a341 Remove empty test.
llvm-svn: 61095
2008-12-16 19:07:17 +00:00
Bill Wendling
f807a68f2e Temporarily revert r61019, r61030, and r61040. These were breaking LLVM Release
builds.

llvm-svn: 61094
2008-12-16 19:06:48 +00:00
Nick Lewycky
1b0fc83809 Generalize support for analyzing loops to include SLE/SGE loop exit conditions
and support for non-unit strides with signed exit conditions.

llvm-svn: 61082
2008-12-16 08:30:01 +00:00
Chris Lattner
0e79aa6595 Teach basicaa to use the nocapture attribute when possible. When the
intrinsics are properly marked nocapture, the fixme should be addressed.

llvm-svn: 61040
2008-12-15 18:59:22 +00:00
Nick Lewycky
51228d6707 Revert my re-instated reverted commit, fixes the bootstrap build on x86-64 linux.
llvm-svn: 60951
2008-12-12 17:09:07 +00:00
Nick Lewycky
312d95be37 Sneaky, sneaky: move the -1 to the outside of the SMax. Reinstate the
optimization of SGE/SLE with unit stride, now that it works properly.

llvm-svn: 60881
2008-12-11 17:40:14 +00:00
Chris Lattner
e2b5854e41 Allow basicaa to walk through geps with identical indices in
parallel, allowing it to decide that P/Q must alias if A/B
must alias in things like:
 P = gep A, 0, i, 1
 Q = gep B, 0, i, 1

This allows GVN to delete 62 more instructions out of 403.gcc.

llvm-svn: 60820
2008-12-10 01:04:47 +00:00
Evan Cheng
35f9ef25f1 xfail this for now.
llvm-svn: 60777
2008-12-09 18:43:00 +00:00
Nick Lewycky
41060b1556 It's easy to handle SLE/SGE when the loop has a unit stride.
llvm-svn: 60748
2008-12-09 07:25:04 +00:00
Nick Lewycky
47fa9bd187 Extend the 'noalias' attribute to function return values. This is intended to
indicate functions that allocate, such as operator new, or list::insert. The
actual definition is slightly less strict (for now).

No changes to the bitcode reader/writer, asm printer or verifier were needed.

llvm-svn: 59934
2008-11-24 03:41:24 +00:00
Nick Lewycky
c573f70ae4 Add a utility function that detects whether a loop is guaranteed to be finite.
Use it to safely handle less-than-or-equals-to exit conditions in loops. These
also occur when the loop exit branch is exit on true because SCEV inverses the
icmp predicate.

Use it again to handle non-zero strides, but only with an unsigned comparison
in the exit condition.

llvm-svn: 59528
2008-11-18 15:10:54 +00:00
Nick Lewycky
1cddd8346f Don't brute-force analyze cubic or higher polynomials.
If this patch causes a performance regression for anyone, please let me know,
and it can be fixed in a different way with much more effort.

llvm-svn: 59384
2008-11-16 04:14:25 +00:00
Nick Lewycky
cc1b7622a5 Don't crash analyzing certain quadratics (addrec of {X,+,Y,+,1}).
We're still waiting on code that actually analyzes them properly.

llvm-svn: 58592
2008-11-03 02:43:49 +00:00
Duncan Sands
2c9f6ab06d Rationalize the names of passes that print information:
-callgraph => print-callgraph
    -callscc   => print-callgraph-sccs
    -cfgscc    => print-cfg-sccs
    -externalfnconstants => print-externalfnconstants
    -print               => print-function
    -print-alias-sets (no change)
    -print-callgraph     => dot-callgraph
    -print-cfg           => dot-cfg
    -print-cfg-only      => dot-cfg-only
    -print-dom-info (no change)
    -printm              => print-module
    -printusedtypes      => print-used-types

llvm-svn: 56487
2008-09-23 12:47:39 +00:00
Duncan Sands
b2b5ddfaf6 Remove the MarkModRef pass (use AddReadAttrs instead).
Unfortunately this means removing one regression test
of GlobalsModRef because I couldn't work out how to
perform it without MarkModRef.

llvm-svn: 56342
2008-09-19 08:23:44 +00:00
Duncan Sands
c2ff9ca568 Add a new pass AddReadAttrs which works out which functions
can get the readnone/readonly attributes, and gives them it.
The plan is to remove markmodref (which did the same thing
by querying GlobalsModRef) and delete the analogous
functionality from GlobalsModRef.

llvm-svn: 56341
2008-09-19 08:17:05 +00:00
Duncan Sands
aaa27f4261 Teach -callgraph to always print the callgraph (as the
description says it does), not just when -analyze is
used as well.  This means printing to stderr, so adjust
some tests.

llvm-svn: 56337
2008-09-19 07:57:09 +00:00
Dan Gohman
885e592e9b Finally re-apply r46959. This is made feasible by the combination
of r56230, r56232, and r56246.

llvm-svn: 56247
2008-09-16 18:52:57 +00:00
Dan Gohman
64622ff4d6 Fix spacing in the grep line for this test, following the recent
SCEV-whitespace changes.

llvm-svn: 56234
2008-09-16 01:37:08 +00:00
Dan Gohman
0045cec2ba Teach ScalarEvolution to consider loop preheaders in the search for
an if statement that guards a loop, to allow indvars to avoid smax
operations in more situations.

llvm-svn: 56232
2008-09-15 22:18:04 +00:00
Dan Gohman
6418cb3998 Fix WriteAsOperand to not emit a leading space character. Adjust
its callers to emit a space character before calling it when a
space is needed.

This fixes several spurious whitespace issues in
ScalarEvolution's debug dumps. See the test changes for
examples.

This also fixes odd space-after-tab indentation in the output
for switch statements, and changes calls from being printed like
this:
  call void @foo( i32 %x )
to this:
  call void @foo(i32 %x)

llvm-svn: 56196
2008-09-14 17:21:12 +00:00
Duncan Sands
8d5640ac90 Fix PR2792: treat volatile loads as writing memory somewhere.
Treat stores as reading memory, just to play safe.

llvm-svn: 56188
2008-09-13 12:45:50 +00:00
Duncan Sands
d8ab4c3485 Correct callgraph construction. It has two problems:
(1) code left over from the days of ConstantPointerRef:
if a use of a function is a GlobalValue then that is
not considered a reason to add an edge from the external
node, even though the use may be as an initializer for
an externally visible global!  There might be some point
to this behaviour when the use is by an alias (though the
code predated aliases by some centuries), but I think
PR2782 is a better way of handling that.  (2) If function
F calls function G, and also G is a parameter to the
call, then an F->G edge is not added to the callgraph.
While this doesn't seem to matter much, adding such an
edge makes the callgraph more regular.
In addition, the new code should be faster as well as
simpler.

llvm-svn: 55987
2008-09-09 12:40:47 +00:00
Duncan Sands
4a2e0c6591 Testcase for commits 55700 and 55714.
llvm-svn: 55715
2008-09-03 19:38:41 +00:00
Duncan Sands
06bd9f3885 Since onlyReadsMemory returns true if in fact
doesNotAccessMemory, check doesNotAccessMemory
first, since otherwise functions may be
marked readonly rather than readnone.

llvm-svn: 55697
2008-09-03 15:31:24 +00:00
Duncan Sands
0283a6c991 Cleanup GlobalsModRef a bit. When analysing the
callgraph, when one member of a SCC calls another
then the analysis would drop to mod-ref because
there is (usually) no function info for the callee
yet; fix this.  Teach the analysis about function
attributes, in particular the readonly attribute
(which requires being careful about globals).

llvm-svn: 55696
2008-09-03 12:55:42 +00:00
Owen Anderson
42c1d6c7b9 Remove GCSE and LoadVN from the testsuite.
llvm-svn: 54832
2008-08-16 00:00:54 +00:00
Dan Gohman
37da9e9dbd Extend ScalarEvolution's executesAtLeastOnce logic to be able to
continue past the first conditional branch when looking for a
relevant test. This helps it avoid using MAX expressions in
loop trip counts in more cases.

llvm-svn: 54697
2008-08-12 20:17:31 +00:00
Eli Friedman
ef366a16a2 PR2621: Improvements to the SCEV AddRec binomial expansion. This
version uses a new algorithm for evaluating the binomial coefficients 
which is significantly more efficient for AddRecs of more than 2 terms 
(see the comments in the code for details on how the algorithm works).  
It also fixes some bugs: it removes the arbitrary length restriction for 
AddRecs, it fixes the silent generation of incorrect code for AddRecs 
which require a wide calculation width, and it fixes an issue where we 
were incorrectly truncating the iteration count too far when evaluating 
an AddRec expression narrower than the induction variable.

There are still a few related issues I know of: I think there's 
still an issue with the SCEVExpander expansion of AddRec in terms of
the width of the induction variable used.  The hack to avoid generating 
too-wide integers shouldn't be necessary; instead, the callers should be 
considering the cost of the expansion before expanding it (in addition 
to not expanding too-wide integers, we might not want to expand 
expressions that are really expensive, especially when optimizing for 
size; calculating an length-17 32-bit AddRec currently generates about 250 
instructions of straight-line code on X86).  Also, for long 32-bit 
AddRecs on X86, CodeGen really sucks at scheduling the code.  I'm planning on 
filing follow-up PRs for these issues.

llvm-svn: 54332
2008-08-04 23:49:06 +00:00
Eli Friedman
a750724150 Another SCEV issue from PR2607; essentially the same issue, but this
time applying to the implicit comparison in smin expressions. The 
correct way to transform an inequality into the opposite 
inequality, either signed or unsigned, is with a not expression.

I looked through the SCEV code, and I don't think there are any more 
occurrences of this issue.

llvm-svn: 54194
2008-07-30 04:36:32 +00:00
Eli Friedman
81169f2e1b Fix for PR2607: SCEV miscomputing the loop count for loops with an
SGT exit condition.  Essentially, the correct way to flip an inequality 
in 2's complement is the not operator, not the negation operator.  
That said, the difference only affects cases involving INT_MIN.

Also, enhance the pre-test search logic to be a bit smarter about 
inequalities flipped with a not operator, so it can eliminate the smax 
from the iteration count for simple loops.

llvm-svn: 54184
2008-07-30 00:04:08 +00:00
Wojciech Matyjewicz
eea926ec20 Fix PR2088. Use modulo linear equation solver to compute loop iteration
count.

llvm-svn: 53810
2008-07-20 15:55:14 +00:00
Nick Lewycky
13166526c5 XFAIL this test.
llvm-svn: 53793
2008-07-19 15:52:06 +00:00
Wojciech Matyjewicz
852a8f47f1 While testing particular algorithms to compute loop iteration count the brute
force evaluation (ComputeIterationCountExhaustively) should be turned off.

It doesn't apply to trip-count2.ll because this file tests the brute force
evaluation.

The test for PR2364 (2008-05-25-NegativeStepToZero.ll) currently fails
showing that the patch for this bug doesn't work. I'll fix it in a few hours
with a patch for PR2088.

llvm-svn: 53792
2008-07-19 13:26:15 +00:00
Nick Lewycky
df9e9f0b0e Stop creating extraneous smax/umax in SCEV. This removes a regression where we
started complicating many loops ('for' loops, in fact).

llvm-svn: 53508
2008-07-12 07:41:32 +00:00
Chris Lattner
e987a3bdd1 If we are checking to see if the result of a call aliases a
pointer derived from a local allocation, if the local allocation
never escapes, the pointers can't alias.  This implements PR2436

llvm-svn: 52301
2008-06-16 06:19:11 +00:00
Nick Lewycky
0aa53f2b52 Crash less. The i64 restriction in BinomialCoefficient caused some problems
with code that was expecting different bit widths for different values.

Make getTruncateOrZeroExtend a method on ScalarEvolution, and use it.

llvm-svn: 52248
2008-06-13 04:38:55 +00:00
Matthijs Kooijman
98322ead14 Remove trailing whitespace after line continuations in test cases to them work.
This fixes two test cases that were not being run properly before.

llvm-svn: 52179
2008-06-10 15:07:07 +00:00
Matthijs Kooijman
82d762a948 Suppress the (stderr) output of -aa-eval, this fixes 5 tests.
llvm-svn: 52173
2008-06-10 12:39:15 +00:00
Wojciech Matyjewicz
06e4c8a420 Fixes PR2395. Looking for a constant in a GEP tail (when the first GEP
is longer than the second one) should stop after finding one. Added break 
instruction guarantees it. It also changes difference between offsets to 
absolute value of this difference in the condition.

llvm-svn: 51875
2008-06-02 17:26:12 +00:00
Owen Anderson
2a0090d9bc Move these tests into the proper directory.
llvm-svn: 51685
2008-05-29 16:30:29 +00:00
Nick Lewycky
0ba4adf4ef Whoops -- forgot PR reference on this test.
llvm-svn: 51569
2008-05-26 20:23:33 +00:00
Nick Lewycky
7116ad5a18 Use {} instead of "" in RUN lines.
llvm-svn: 51561
2008-05-26 01:27:08 +00:00
Nick Lewycky
f24743a6bb Don't treat values as signed when looking at loop steppings in HowForToNonZero.
llvm-svn: 51560
2008-05-25 23:43:32 +00:00
Dan Gohman
2412469191 Remove lingering references to .llx and .tr in the tests.
llvm-svn: 51500
2008-05-23 21:15:35 +00:00
Gabor Greif
807c2df887 sabre brings to my attention that the 'tr' suffix is also obsolete
llvm-svn: 51349
2008-05-20 21:00:03 +00:00
Gabor Greif
d8a4dbb5da Rename the last test with .llx extension to .ll, resolve duplicate test by renaming to isnan2. Now that no test has llx ending there is no need to search for them from dg.exp too.
llvm-svn: 51328
2008-05-20 19:52:04 +00:00
Owen Anderson
3e607df2f8 Fix this test. It was testing broken behavior in that it required ADCE to eliminate
a potentially infinite loop, which is undesirable.  Instead, test the LICM behavior
that we're really interested in.

llvm-svn: 51177
2008-05-16 04:25:09 +00:00
Owen Anderson
611b415d12 Fix PR1098 by correcting the postdominators analysis.
Patch by Florian Brandner.

llvm-svn: 50628
2008-05-04 21:07:35 +00:00
Chris Lattner
a9d8d647ca rename *.llx -> *.ll, last batch.
llvm-svn: 49971
2008-04-19 22:32:52 +00:00
Owen Anderson
cd1b9c4b43 Make GVN able to remove unnecessary calls to read-only functions again.
llvm-svn: 49842
2008-04-17 05:36:50 +00:00
Dale Johannesen
45e14f7753 Don't assume a tail call can't reference a byval
argument to the outer function, this isn't correct.

llvm-svn: 49731
2008-04-15 17:41:34 +00:00
Owen Anderson
a6d1d8dec2 The functionality being tested was removed because it was horribly unsafe.
llvm-svn: 49610
2008-04-13 09:51:06 +00:00
Duncan Sands
79af9d68ec Testcase for pr2169.
llvm-svn: 49344
2008-04-07 17:03:16 +00:00
Duncan Sands
8e40ac013e Testcase for PR2160.
llvm-svn: 48655
2008-03-21 20:22:11 +00:00
Daniel Berlin
0f20fa71ba Fix PR 2160 by making sure arguments to external functions get marked as pointing to anything
llvm-svn: 48509
2008-03-18 22:22:53 +00:00
Gabor Greif
fd0fb616f8 Fix http://llvm.org/bugs/show_bug.cgi?id=2104 by ordering lexicographically what gets printed. Be const-correct in PrintResults and uninline it too
llvm-svn: 47712
2008-02-28 08:38:45 +00:00
Evan Cheng
e283043860 Temporarily reverting 46959.
llvm-svn: 47542
2008-02-25 03:57:32 +00:00
Nick Lewycky
0b6cdd1eff Add 'umax' similar to 'smax' SCEV. Closes PR2003.
Parse reversed smax and umax as smin and umin and express them with negative
or binary-not SCEVs (which are really just subtract under the hood).

Parse 'xor %x, -1' as (-1 - %x).

Remove dead code (ConstantInt::get always returns a ConstantInt).

Don't use getIntegerSCEV(-1, Ty). The first value is an int, then it gets
passed into a uint64_t. Instead, create the -1 directly from
ConstantInt::getAllOnesValue().

llvm-svn: 47360
2008-02-20 06:48:22 +00:00
Tanya Lattner
f37f44f9ec Remove llvm-upgrade.
llvm-svn: 47110
2008-02-14 06:56:27 +00:00
Wojciech Matyjewicz
6df5450bc4 Now that ScalarEvolution::print writes to the correct stream, there is
no need to redirect stderr into stdout.

llvm-svn: 47009
2008-02-12 15:12:40 +00:00
Wojciech Matyjewicz
ae01857e92 Change negative grep into positive one in my yesterday's testcase.
llvm-svn: 47008
2008-02-12 15:10:35 +00:00
Wojciech Matyjewicz
b6e6dacc60 Fix PR2002. Suppose n is the initial value for the induction
variable (with step 1) and m is its final value. Then, the correct trip 
count is SMAX(m,n)-n. Previously, we used SMAX(0,m-n), but m-n may 
overflow and can't in general be interpreted as signed.

Patch by Nick Lewycky.

llvm-svn: 47007
2008-02-12 15:09:36 +00:00
Wojciech Matyjewicz
2874a19254 If the LHS of the comparison is a loop-invariant we also want to move it
to the RHS. This simple change allows to compute loop iteration count 
for loops with condition similar to the one in the testcase (which seems 
to be quite common).

llvm-svn: 46959
2008-02-11 18:37:34 +00:00
Wojciech Matyjewicz
76c2b22ee4 Fix PR1798 - an error in the evaluation of SCEVAddRecExpr at an
arbitrary iteration.

The patch:
1) changes SCEVSDivExpr into SCEVUDivExpr,
2) replaces PartialFact() function with BinomialCoefficient(); the 
computations (essentially, the division) in BinomialCoefficient() are 
performed with the apprioprate bitwidth necessary to avoid overflow; 
unsigned division is used instead of the signed one.

Computations in BinomialCoefficient() require support from the code 
generator for APInts. Currently, we use a hack rounding up the 
neccessary bitwidth to the nearest power of 2. The hack is easy to turn 
off in future.

One remaining issue: we assume the divisor of the binomial coefficient 
formula can be computed accurately using 16 bits. It means we can handle 
AddRecs of length up to 9. In future, we should use APInts to evaluate 
the divisor.

Thanks to Nicholas for cooperation!

llvm-svn: 46955
2008-02-11 11:03:14 +00:00
Chris Lattner
cd5013eb2f Teach basicaa that 'byval' arguments define a new memory location that
can't be aliased to other known objects.  This allows us to know that byval 
pointer args don't alias globals, etc.

llvm-svn: 46315
2008-01-24 18:00:32 +00:00
Nick Lewycky
e05651c6d7 Accept both %y, %x and %x, %y as valid answers.
llvm-svn: 45649
2008-01-06 03:12:44 +00:00
Chris Lattner
95f9743c43 Fix PR1782, patch by Wojtek Matyjewicz!
llvm-svn: 44733
2007-12-09 07:35:13 +00:00
Tanya Lattner
c33660d278 Fix bug in regression tests that ignored stderr output in RUN lines. Updated tests and fixed broken run lines.
XFAILed 3 arm regressions (will file bugs)

llvm-svn: 44389
2007-11-28 04:57:00 +00:00
Dan Gohman
dfda7ca2cc Change grep '' to grep {}.
Change 2>&1 | to |&.

llvm-svn: 44344
2007-11-27 00:10:35 +00:00
Owen Anderson
05275b7b14 Allow GVN to eliminate read-only function calls when it can detect that they are redundant.
llvm-svn: 44323
2007-11-26 02:26:36 +00:00
Nick Lewycky
1536b3f230 Add new SCEV, SCEVSMax. This allows LLVM to analyze do-while loops.
llvm-svn: 44319
2007-11-25 22:41:31 +00:00
Duncan Sands
265752ae8f Ding dong, the DoesntAccessMemoryFns and
OnlyReadsMemoryFns tables are dead!  We
get more, and more accurate, information
from gcc via the readnone and readonly
function attributes.

llvm-svn: 44288
2007-11-23 19:30:27 +00:00
Duncan Sands
703986d327 Teach alias analysis about readnone/readonly functions.
Based on a patch by Török Edwin.

llvm-svn: 44279
2007-11-22 21:43:27 +00:00
Nick Lewycky
e34dd3a90f Create nodes for inline asm so that we don't crash looking for the node later.
llvm-svn: 44267
2007-11-22 03:07:37 +00:00
Nick Lewycky
f1d30ded71 Be more careful when transforming | to +. Patch from Wojciech Matyjewicz.
llvm-svn: 44248
2007-11-20 08:24:44 +00:00
Anton Korobeynikov
c9ab2516fa Reverted r44163 per request
llvm-svn: 44177
2007-11-15 18:33:16 +00:00
Nick Lewycky
6690b9499b Fix handling of overflow in loop calculation by adding new UDiv SCEV. This SCEV
is disabled in the sense that it will refuse to create one from a UDiv
instruction, until the code is better tested.

llvm-svn: 44163
2007-11-15 06:30:50 +00:00
Chris Lattner
02a91dce21 Fix PR1774 and BasicAA/2007-11-05-SizeCrash.ll
llvm-svn: 43756
2007-11-06 05:58:42 +00:00
Owen Anderson
6cc4e90741 Fix for PR1741.
llvm-svn: 43326
2007-10-25 02:36:18 +00:00
Nick Lewycky
a44e458ddc Build the correct range for loops with unusual bounds. Fix from Jay Foad.
llvm-svn: 42394
2007-09-27 14:12:54 +00:00
Dan Gohman
794fa1f8f7 Convert tests using "| wc -l | grep ..." to use the count script.
llvm-svn: 41097
2007-08-15 13:36:28 +00:00
Nick Lewycky
8543faa5fe Add reference to problem report.
llvm-svn: 40889
2007-08-07 12:27:03 +00:00
Chandler Carruth
67d3119773 This resolves a regression of BasicAA which failed to find any memory information for overloaded intrinsics (PR1600). This resolves that issue, and improves the matching scheme to use a BitVector rather than a binary search.
llvm-svn: 40872
2007-08-06 20:57:16 +00:00
Nick Lewycky
9e45c33d84 Fix the dates on these tests. It's not September yet. Thanks Reid!
llvm-svn: 40869
2007-08-06 20:00:11 +00:00
Nick Lewycky
04447caa7b Let scalar-evolution analyze loops with an unsigned comparison for the exit
condition. Fixes 1597.

llvm-svn: 40867
2007-08-06 19:21:00 +00:00
Nick Lewycky
071a2c6a5e Don't assume it's safe to transform a loop just because it's dominated by any
comparison. Fixes bug 1598.

llvm-svn: 40866
2007-08-06 18:33:46 +00:00
Chris Lattner
38cfb16fe2 update for new domtree dump format
llvm-svn: 40857
2007-08-06 06:17:08 +00:00
Christopher Lamb
9a2a58c818 Teach BasicAA about noalias parameter attributes, but do it correctly this time.
llvm-svn: 40711
2007-08-02 01:18:14 +00:00
Christopher Lamb
5487dd14b4 Revert overly aggressive interpretation of noalias
llvm-svn: 40635
2007-07-31 16:18:07 +00:00
Devang Patel
1324500d25 Bunch of tests to check loop passes.
llvm-svn: 40629
2007-07-31 08:04:17 +00:00
Christopher Lamb
907892e47a Teach BasicAA about noalias function parameters. Passes all of DejaGNU and test-suite.
llvm-svn: 40624
2007-07-31 07:04:51 +00:00
Nick Lewycky
43e5fd5be7 Handle decrementing loops properly. Fixes PR1533.
Always pass the constant as the second parameter to HowManyLessThans.

Remove obsolete "isSigned" parameter.

llvm-svn: 39893
2007-07-16 02:08:00 +00:00
John Criswell
57e5ed4b5a Convert .cvsignore files
llvm-svn: 37801
2007-06-29 16:35:07 +00:00
Zhou Sheng
266fac909c Add two test cases to cover apintification change.
llvm-svn: 36476
2007-04-26 16:44:48 +00:00
Devang Patel
d78334d2e7 New test case.
llvm-svn: 36293
2007-04-21 00:34:37 +00:00
Devang Patel
158854c5fc New test.
llvm-svn: 36235
2007-04-18 00:51:43 +00:00
Reid Spencer
df17fa8ef9 For PR1319:
Remove && from the end of the lines to prevent tests from throwing run
lines into the background. Also, clean up places where the same command
is run multiple times by using a temporary file.

llvm-svn: 36142
2007-04-16 17:36:08 +00:00
Reid Spencer
43899915e9 For PR1319:
Fix syntax of tests to ensure grep pattern is properly quoted.

llvm-svn: 36134
2007-04-16 15:31:49 +00:00
Reid Spencer
d93b834b52 For PR1319:
Fix test syntax per new rules.

llvm-svn: 36133
2007-04-16 15:15:52 +00:00
Nick Lewycky
cb7d12a4a5 Fix this test to test what it was intended to test.
llvm-svn: 36072
2007-04-15 13:14:33 +00:00
Reid Spencer
2077c3789f For PR1319:
Upgrade to use new Tcl exec based test harness.

llvm-svn: 36066
2007-04-15 09:31:07 +00:00
Owen Anderson
9304434617 Update tests for the disappearance of -idom.
llvm-svn: 36064
2007-04-15 08:54:57 +00:00
Reid Spencer
cfa86314dd Changes to fix problems with "make check". Apparently you can redefine
functions and Tcl's just tickled with that. The fix is to give the "new"
test system a different interface function name.

llvm-svn: 36022
2007-04-14 22:51:29 +00:00
Reid Spencer
6ac77d03a8 Convert test cases to new llvm.exp version of llvm_runtest and fix tests
that it found to be broken.

llvm-svn: 36009
2007-04-14 19:27:03 +00:00
Reid Spencer
69d42ad9b6 For PR1319:
Convert to use new llvm.exp version of llvm_testrun

llvm-svn: 36008
2007-04-14 19:10:21 +00:00
Reid Spencer
56b310ae49 Make the llvm-runtest function much more amenable by eliminating all the
global variables that needed to be passed in. This makes it possible to
add new global variables with only a couple changes (Makefile and llvm-dg.exp)
instead of touching every single dg.exp file.

llvm-svn: 35918
2007-04-11 19:56:59 +00:00
Nick Lewycky
7f42a891cf Fix this testcase to fail if the bug were reintroduced.
llvm-svn: 35776
2007-04-08 21:49:13 +00:00
Owen Anderson
4b323657b9 Remove DomSet completely. This concludes work on PR1171.
llvm-svn: 35775
2007-04-08 21:30:05 +00:00
Reid Spencer
50ee6b8557 Remove use of implementation keyword.
llvm-svn: 35412
2007-03-28 02:38:26 +00:00
Reid Spencer
82293f34de For PR411:
Update these tests to not use the same name even though the type of the
value differs. After PR411 hits, type planes will be gone and it will be
illegal for a name to be used twice, regardless of type.

llvm-svn: 33660
2007-01-30 16:16:01 +00:00
Reid Spencer
6a31ec1259 For PR761:
Remove "target endian/pointersize" or add "target datalayout" to make
the test parse properly or set the datalayout because defaults changes.

For PR645:
Make global names use the @ prefix.

For llvm-upgrade changes:
Fix test cases or completely remove use of llvm-upgrade for test cases
that cannot survive the new renaming or upgrade capabilities.

llvm-svn: 33533
2007-01-26 08:25:06 +00:00
Reid Spencer
4572ce85b0 Regression is gone, don't try to find it on clean target.
llvm-svn: 33296
2007-01-17 07:59:14 +00:00