47837c5182
don't mark the arguments of prototype overdefined, they will never be queried.
...
llvm-svn: 85793
2009-11-02 06:34:04 +00:00
5503328332
restore some code I removed in r85788, refactor it into
...
a shared place instead of duplicating it 4 times.
llvm-svn: 85792
2009-11-02 06:28:16 +00:00
4910b656b2
remove some confused code that dates from when we had
...
"multiple return values" but not "first class aggregates"
llvm-svn: 85791
2009-11-02 06:17:06 +00:00
809aee2f40
avoid redundant lookups in BBExecutable, and make it a SmallPtrSet.
...
llvm-svn: 85790
2009-11-02 06:11:23 +00:00
e77c9aa04a
Use the libanalysis 'ConstantFoldLoadFromConstPtr' function
...
instead of reinventing SCCP-specific logic. This gives us
new powers.
llvm-svn: 85789
2009-11-02 06:06:14 +00:00
f548403989
switch the main 'ValueState' map from being an std::map to being
...
a DenseMap. Doing this required being aware of subtle iterator
invalidation issues, but it provides a big speedup. In a
release-asserts build, this sped up optimizing 403.gcc from
1.34s -> 0.79s (IPSCCP) and 1.11s -> 0.44s (SCCP).
This commit also conflates in a bunch of general cleanups, sorry.
llvm-svn: 85788
2009-11-02 05:55:40 +00:00
1708b06c0e
Unbreak ARMBaseRegisterInfo::copyRegToReg.
...
llvm-svn: 85787
2009-11-02 04:44:55 +00:00
4e849162ef
fix a bug exposed by moving SRoA earlier which caused a crash building kc++
...
llvm-svn: 85786
2009-11-02 04:37:17 +00:00
e82b087ae6
only IPSCCP incoming arguments if the function is executable, this fixes
...
an assertion on the buildbot.
llvm-svn: 85784
2009-11-02 03:25:55 +00:00
9e97fbe114
add a new ValueState::getConstantInt() helper, use it to
...
simplify some code.
llvm-svn: 85783
2009-11-02 03:21:36 +00:00
2c7457c0f1
Fix malloc.h is deprecated warning on DragonFly BSD.
...
llvm-svn: 85782
2009-11-02 03:20:57 +00:00
0cc7777ce8
Fix for warning seen on DF-BSD, Victor, please fix this to use a shift instead of pow()
...
llvm-svn: 85781
2009-11-02 03:14:31 +00:00
7ccf1a6df6
tidy up some more: remove some extraneous inline specifiers, return harder.
...
llvm-svn: 85780
2009-11-02 03:03:42 +00:00
78eedecade
Apply fix for PR5135, Credit to Andreas Neustifter.
...
llvm-svn: 85779
2009-11-02 02:55:39 +00:00
b5a13d4c90
eliminate the SCCPSolver::getValueMapping method.
...
llvm-svn: 85778
2009-11-02 02:54:24 +00:00
c49ae9912a
fix failures introduced in r85774
...
llvm-svn: 85777
2009-11-02 02:48:17 +00:00
e405ed9651
factor duplicated code into a new DeleteInstructionInBlock
...
function, eliminate temporary (and pointless) smallvector.
llvm-svn: 85776
2009-11-02 02:47:51 +00:00
a3c39d394d
Chris used to use '...' instead of proper grammar.
...
llvm-svn: 85775
2009-11-02 02:33:50 +00:00
6df5cec72f
remove some extraneous llvmcontext stuff.
...
llvm-svn: 85774
2009-11-02 02:30:06 +00:00
efdd2bbce6
change LatticeVal to use PointerIntPair to save some space.
...
llvm-svn: 85773
2009-11-02 02:20:32 +00:00
3cd6a61b27
fix instcombine to only do store sinking when the alignments
...
of the two loads agree. Propagate that onto the new store.
llvm-svn: 85772
2009-11-02 02:06:37 +00:00
4d23754b14
Handle splats of undefs properly. This includes the testcase for PR5364 as well.
...
llvm-svn: 85767
2009-11-02 00:12:06 +00:00
0f38d989bd
Do not infer the target type for COPY_TO_REGCLASS from dest regclass, this won't work if it can contain several types. Require explicit result type for the node for now. This fixes PR5364.
...
PS: It seems that blackfin usage of copy_to_regclass is completely bogus!
llvm-svn: 85766
2009-11-02 00:11:39 +00:00
8cce1eb6aa
64-bit FP loads & stores operate on both NEON and VFP pipelines.
...
llvm-svn: 85765
2009-11-02 00:11:06 +00:00
14635da94b
Use NEON reg-reg moves, where profitable. This reduces "domain-cross" stalls, when we used to mix vfp and neon code (the former were used for reg-reg moves)
...
llvm-svn: 85764
2009-11-02 00:10:38 +00:00
ea68c7c9a8
Add PseudoSourceValue::mayAlias. It returns true if the object can ever alias any LLVM IR value.
...
llvm-svn: 85762
2009-11-01 23:50:04 +00:00
1a4492be97
Fix a couple more places where we are creating ld / st instructions without memoperands.
...
llvm-svn: 85746
2009-11-01 22:04:35 +00:00
43219997b6
Make use of imm12 version of Thumb2 ldr / str instructions more aggressively.
...
llvm-svn: 85743
2009-11-01 21:12:51 +00:00
328ef89bd1
when merging two loads, make sure to take the min of their alignment,
...
not the max. This didn't matter until the previous patch because
instcombine would refuse to sink loads with differenting alignments.
llvm-svn: 85738
2009-11-01 20:07:07 +00:00
2a249e267a
split load sinking out to its own function, like gep sinking.
...
llvm-svn: 85737
2009-11-01 20:04:24 +00:00
0b40a8bc0e
fix a bug noticed by inspection: when instcombine sinks loads through
...
phis, it didn't preserve the alignment of the load. This is a missed
optimization of the alignment is high and a miscompilation when the
alignment is low.
llvm-svn: 85736
2009-11-01 19:50:13 +00:00
f25d301311
Add a missing closing parenthesis, and tweak to fit in 80
...
columns.
llvm-svn: 85732
2009-11-01 19:12:43 +00:00
b5d9c8c708
cleanups, switch GlobalDCE to SmallPtrSet instead of std::set
...
llvm-svn: 85730
2009-11-01 19:03:42 +00:00
37536b90e1
remove a bunch of locking from LLVMContextImpl. Since only one thread
...
can be banging on a context at a time, this isn't needed. Owen, please
review.
llvm-svn: 85728
2009-11-01 18:42:03 +00:00
249f96e339
improve comment.
...
llvm-svn: 85725
2009-11-01 18:17:37 +00:00
1756673f58
add a comment about why we don't allow inlining indbr.
...
llvm-svn: 85724
2009-11-01 18:16:30 +00:00
253bc77513
the verifier shouldn't modify the IR.
...
llvm-svn: 85722
2009-11-01 18:11:50 +00:00
291f6145b8
Reverting 85714, 85715, 85716, which are breaking the build
...
llvm-svn: 85717
2009-11-01 16:42:53 +00:00
2b12b95d5d
Add a function to Passes.h to allow clients to create instances
...
of the ScalarEvolution pass without needing to #include ScalarEvolution.h.
llvm-svn: 85716
2009-11-01 15:28:36 +00:00
b49c889d53
Don't #include Pass.h from CallGraph.h.
...
llvm-svn: 85715
2009-11-01 15:23:35 +00:00
576ac96367
Remove the #include of Pass.h from PassManager.h. This breaks a significant
...
#include dependency, as frontends commonly pull in PassManager.h.
llvm-svn: 85714
2009-11-01 15:20:19 +00:00
1a8b80ed5a
teach ipsccp and ipconstprop that a blockaddress doesn't 'take the address' of a function
...
in a way that should prevent ip constprop. This allows clang/test/CodeGen/indirect-goto.c
to pass with the new indirect goto lowering.
llvm-svn: 85709
2009-11-01 06:11:53 +00:00
a1dc101f66
change llvm::MergeBlockIntoPredecessor to not merge two blocks BB1->BB2
...
when BB2 has its address taken. Since it ends up doing BB2->rauw(BB1),
this can cause the address of the entry block to be taken. Since it is
generally undesirable to nuke blocks whose address is taken, even when
we can, just unconditionally stop this xform.
llvm-svn: 85708
2009-11-01 04:57:33 +00:00
746139b736
strengthen an assumption: RevectorBlockTo knows that PredBB
...
ended in an uncond branch because the pass requires BreakCriticalEdges.
However, BCE doesn't eliminate critical adges from indbrs.
llvm-svn: 85707
2009-11-01 04:23:20 +00:00
274717419c
fix an issue where the verifier would reject a function whose entry
...
block had its address taken even if the blockaddress was dead.
llvm-svn: 85706
2009-11-01 04:08:01 +00:00
7a8db3a41a
if CostMetrics says to never duplicate some code, don't unswitch a loop.
...
This prevents unswitching from duplicating indbr's.
llvm-svn: 85705
2009-11-01 03:42:55 +00:00
54a4b84012
constant fold indirectbr(blockaddress(%bb)) -> br label %bb.
...
llvm-svn: 85704
2009-11-01 03:40:38 +00:00
50ba5c3dc2
improve x86 codegen support for blockaddress. We now compile
...
the testcase into:
_test1: ## @test1
## BB#0: ## %entry
leaq L_test1_bb6(%rip), %rax
jmpq *%rax
L_test1_bb: ## Address Taken
LBB1_1: ## %bb
movb $1, %al
ret
L_test1_bb6: ## Address Taken
LBB1_2: ## %bb6
movb $2, %al
ret
Note, it is very very strange that BlockAddressSDNode doesn't carry
around TargetFlags. Dan, please fix this.
llvm-svn: 85703
2009-11-01 03:25:03 +00:00
4578f8ea07
pull check for return inst out of loop, never inline a callee that contains
...
an indirectbr.
llvm-svn: 85702
2009-11-01 03:07:53 +00:00
c559a9f686
Fix BlockAddress::replaceUsesOfWithOnConstant to correctly
...
maintain the block use count in SubclassData.
llvm-svn: 85701
2009-11-01 03:03:03 +00:00
140a754ab2
implement linker support for BlockAddress.
...
llvm-svn: 85700
2009-11-01 02:46:39 +00:00
aa99c94e2a
Revert 85678/85680. The decision is to stay with the current form of
...
indirectbr, thus we don't need "blockaddr(@func, null)". Eliminate it
for simplicity.
llvm-svn: 85699
2009-11-01 01:27:45 +00:00
6f29ad9170
Use cbz and cbnz instructions.
...
llvm-svn: 85698
2009-10-31 23:46:45 +00:00
5cba8de2c8
vml[as].f32 cause stalls in following advanced SIMD instructions. Avoid using
...
them for scalar floating point operations for now.
llvm-svn: 85697
2009-10-31 22:57:36 +00:00
a546dcf418
Make sure PRE doesn't split crit edges from indirectbr.
...
llvm-svn: 85692
2009-10-31 22:11:15 +00:00
c872b09676
llvm::SplitEdge should refuse to split an edge from an indirectbr.
...
Fix CodeGenPrepare to not try to split edges from indirectbr.
llvm-svn: 85690
2009-10-31 22:04:43 +00:00
ba364b0a9a
update the comment above llvm::SplitCriticalEdge, and make
...
it abort on IndirectBrInst as describe in the comment.
llvm-svn: 85688
2009-10-31 21:51:10 +00:00
8fe6fd702d
Expand 64-bit logical shift right inline
...
llvm-svn: 85687
2009-10-31 21:42:19 +00:00
624fcb286e
Expand 64-bit arithmetic shift right inline
...
llvm-svn: 85685
2009-10-31 21:00:56 +00:00
ae6bb61594
Fix a missing newline in the dwarf output code.
...
llvm-svn: 85684
2009-10-31 20:59:09 +00:00
34341e69c4
Make -print-machineinstrs more readable.
...
- Be consistent when referring to MachineBasicBlocks: BB#0.
- Be consistent when referring to virtual registers: %reg1024.
- Be consistent when referring to unknown physical registers: %physreg10.
- Be consistent when referring to known physical registers: %RAX
- Be consistent when referring to register 0: %reg0
- Be consistent when printing alignments: align=16
- Print jump table contents.
- Don't print host addresses, in general.
- and various other cleanups.
llvm-svn: 85682
2009-10-31 20:19:03 +00:00
18dc1c5c9e
Factor out more code into addCommonCodeGenPasses. The JIT wasn't
...
previously running CodePlacementOpt. Also print headers before
each dump in -print-machineinstrs mode, so that it's clear which
dump is which.
llvm-svn: 85681
2009-10-31 20:17:39 +00:00
3c89c53f35
adjust a couple xforms to work with null bb's in BlockAddress.
...
llvm-svn: 85680
2009-10-31 20:13:24 +00:00
b2c0e243c5
Make blockaddress(@func, null) be valid, and make 'deleting a basic
...
block with a blockaddress still referring to it' replace the invalid
blockaddress with a new blockaddress(@func, null) instead of a
inttoptr(1).
This changes the bitcode encoding format, and still needs codegen
support (this should produce a non-zero value, referring to the entry
block of the function would also be quite reasonable).
llvm-svn: 85678
2009-10-31 20:08:37 +00:00
5d994048dd
Expand 64 bit left shift inline rather than using the libcall. For now, this
...
is unconditional. Making it still use the libcall when optimizing for size
would be a good adjustment.
llvm-svn: 85675
2009-10-31 19:38:01 +00:00
a742b8f94f
add a comment.
...
llvm-svn: 85671
2009-10-31 17:48:31 +00:00
2d02ff8cbb
Revert r85667. LoopUnroll currently can't call utility functions which
...
auto-update the DominatorTree because it doesn't keep the DominatorTree
current while it works.
llvm-svn: 85670
2009-10-31 17:33:01 +00:00
144694bcb7
Remove redundant code.
...
llvm-svn: 85668
2009-10-31 16:16:41 +00:00
041e2dbad1
Merge the enhancements from LoopUnroll's FoldBlockIntoPredecessor into
...
MergeBlockIntoPredecessor. This makes SimplifyCFG slightly more aggressive,
and makes it unnecessary for LoopUnroll to have its own copy of this code.
llvm-svn: 85667
2009-10-31 16:08:00 +00:00
880c92ac1c
Rename forgetLoopBackedgeTakenCount to forgetLoop, because it
...
clears out more information than just the stored backedge taken count.
llvm-svn: 85664
2009-10-31 15:04:55 +00:00
969e83a4ff
Replace LoopUnrollPass.cpp's custom code-size estimation code using
...
the new common CodeMetrics code.
llvm-svn: 85663
2009-10-31 14:54:17 +00:00
fa8969f70e
Simplify this code.
...
llvm-svn: 85662
2009-10-31 14:46:50 +00:00
af94015c18
Remove an unnecessary #include.
...
llvm-svn: 85661
2009-10-31 14:39:43 +00:00
f35b6640f6
Update CMakeLists for recent renames.
...
llvm-svn: 85660
2009-10-31 14:38:25 +00:00
f70e76c435
Rename UnrollLoop.cpp to LoopUnroll.cpp, and LoopUnroll.cpp to
...
LoopUnrollPass.cpp, for consistency with other passes which are
similarly split.
llvm-svn: 85659
2009-10-31 14:37:31 +00:00
fb7f0e57b6
Remove CodeGenLICM. It's largely obsoleted by MachineLICM's new ability
...
to unfold loop-invariant loads.
llvm-svn: 85657
2009-10-31 14:35:41 +00:00
ac45c9171d
Make ScalarEvolutionAliasAnalysis slightly more aggressive, by making an
...
underlying alias call even for non-identified-object values.
llvm-svn: 85656
2009-10-31 14:32:25 +00:00
930aa9d3d2
Reapply r85634, with the bug fixed.
...
llvm-svn: 85655
2009-10-31 14:22:52 +00:00
ba8735d25a
When discarding SrcValue information, discard all of it so that code
...
that uses this information knows to behave conservatively.
llvm-svn: 85654
2009-10-31 14:14:04 +00:00
97abba5b6f
Fix 80-column violation.
...
llvm-svn: 85653
2009-10-31 14:12:53 +00:00
a0ca9e944f
Fix warning with gcc-4.0 and signed/unsigned.
...
llvm-svn: 85648
2009-10-31 09:24:35 +00:00
cdbb70c065
It's safe to remat t2LDRpci; Add PseudoSourceValue to load / store's to enable more machine licm. More changes coming.
...
llvm-svn: 85643
2009-10-31 03:39:36 +00:00
c16d8f2054
Revert 85634. It's breaking consumer-typeset (and others).
...
llvm-svn: 85641
2009-10-31 01:28:06 +00:00
060ee82dab
Add assertion checks here to turn silent miscompiles into aborts.
...
llvm-svn: 85639
2009-10-30 23:59:06 +00:00
d814e32e57
Don't mark registers dead here when processing nodes with MVT::Flag
...
results. This works around a problem affecting targets which rely on
MVT::Flag to handle physical register defs.
llvm-svn: 85638
2009-10-30 23:57:47 +00:00
7f7d97eb73
Add a comment about a missed opportunity.
...
llvm-svn: 85635
2009-10-30 23:15:43 +00:00
5bec30ca5d
Optimize around the fact that pred_iterator is slow: instead of sorting
...
PHI operands by the predecessor order, sort them by the order used by the
first PHI in the block. This is still suffucient to expose duplicates.
llvm-svn: 85634
2009-10-30 23:15:21 +00:00
8be42bd09f
Updates to the ARM target assembler for llvm-mc per review comments from
...
Daniel Dunbar.
- Reordered the fields in the ARMOperand Mem struct to make the struct smaller.
Making bool's into 1 bit fields and put the MCExpr* fields adjacent to each
other.
- Fixed a number of places in ARMAsmParser.cpp so they have doxygen comments.
- Change the name of ARMAsmParser::ParseRegister() to MaybeParseRegister and
added the bool ParseWriteBack parameter.
- Changed ARMAsmParser::ParseMemory() to call MaybeParseRegister().
- Added ARMAsmParser::ParseMemoryOffsetReg to factor out parsing the offset of a
memory operand. And use it for both parsing both preindexed and post indexing
addressing forms in ARMAsmParser::ParseMemory.
- Changed the first argument to ParseShift() to a reference.
- Changed ParseShift() to check for Rrx first and return to reduce nesting.
llvm-svn: 85632
2009-10-30 22:55:57 +00:00
072b31a0bf
If string field is empty then return NULL.
...
llvm-svn: 85630
2009-10-30 22:52:47 +00:00
dd5d035302
if basic blocks are destroyed while there are *just* BlockAddress' hanging
...
around, then zap them. This is analogous to dangling constantexprs hanging
off functions.
llvm-svn: 85627
2009-10-30 22:39:36 +00:00
1a95106602
Teach SimplifyCFG how to eliminate duplicate PHI nodes within a block.
...
This reduces codesize on a variety of codes by 1-2% on x86-64. It also
helps clean up after SSAUpdater.
llvm-svn: 85626
2009-10-30 22:39:04 +00:00
74eb5d71a3
make hasAddressTaken() constant time by storing a refcount in BB's subclass data.
...
llvm-svn: 85625
2009-10-30 22:33:29 +00:00
c7415bf536
Add a note about Robert Muth's alternate jump table implementation.
...
llvm-svn: 85624
2009-10-30 22:22:46 +00:00
13e41edc71
Sort the incoming values in PHI nodes to match the predecessor order.
...
This helps expose duplicate PHIs, which will make it easier for them
to be eliminated.
llvm-svn: 85623
2009-10-30 22:22:22 +00:00
49fa51d936
Fix MachineLICM to use the correct virtual register class when
...
unfolding loads for hoisting. getOpcodeAfterMemoryUnfold returns the
opcode of the original operation without the load, not the load
itself, MachineLICM needs to know the operand index in order to get
the correct register class. Extend getOpcodeAfterMemoryUnfold to
return this information.
llvm-svn: 85622
2009-10-30 22:18:41 +00:00
112caed500
it isn't valid to take the address of the entry block.
...
llvm-svn: 85621
2009-10-30 22:15:48 +00:00
13cb281013
If a type is derived from a derived type then calculate size appropriately.
...
llvm-svn: 85619
2009-10-30 22:09:30 +00:00
6b00f4b7a8
Fix a comment.
...
llvm-svn: 85610
2009-10-30 20:13:25 +00:00
5a6b9c40d6
Add option to createGVNPass to disable PRE.
...
llvm-svn: 85609
2009-10-30 20:12:24 +00:00
94966ca926
Stop the iterator in ValueLiveAt from potentially running off the end of the interval.
...
llvm-svn: 85599
2009-10-30 18:12:09 +00:00
Chris Lattner