fb14181b18
turn IPSCCP back on now that the iterator invalidation bug is fixed.
...
llvm-svn: 85858
2009-11-03 03:42:51 +00:00
d195f9e5c3
Turn neon reg-reg moves fixup code into separate pass. This should reduce the compile time.
...
llvm-svn: 85850
2009-11-03 01:04:26 +00:00
fbe0256b23
Revert r85049, it is causing PR5367
...
llvm-svn: 85847
2009-11-03 00:24:48 +00:00
b389f2a04d
Revert previous change to a comment. The BlockAddresses go in the
...
constant pool so they don't get wrapped separately.
llvm-svn: 85844
2009-11-03 00:02:05 +00:00
b70ef3c8c7
fix a nasty iterator invalidation bug from my conversion from
...
std::map to DenseMap, exposed on release llvm-gcc bootstrap.
llvm-svn: 85840
2009-11-02 23:25:39 +00:00
a8a58efc03
Revert 85799 for now. It might be breaking llvm-gcc driver.
...
llvm-svn: 85827
2009-11-02 21:49:14 +00:00
1c66e8a6b7
Put BlockAddresses into ARM constant pools.
...
llvm-svn: 85824
2009-11-02 20:59:23 +00:00
d9dfc2e752
Fix ARMAsmParser::ParseMemoryOffsetReg() where the parameter OffsetRegNum should
...
have been passed as a reference.
llvm-svn: 85823
2009-11-02 20:14:39 +00:00
a15cc59dcb
revert r8579[56], which are causing unhappiness in buildbot land.
...
llvm-svn: 85818
2009-11-02 19:31:10 +00:00
df022f1449
Set bit instead of calling pow() to compute 2 << n
...
llvm-svn: 85814
2009-11-02 18:51:28 +00:00
a38a4dfea7
typo
...
llvm-svn: 85812
2009-11-02 18:28:45 +00:00
a3d794ebbb
disable IPSCCP support for multiple return values, it is buggy, so just
...
disable it until I can fix it.
llvm-svn: 85810
2009-11-02 18:22:51 +00:00
5ac6f244fd
Fix schedule model for BFC.
...
llvm-svn: 85809
2009-11-02 17:28:36 +00:00
4c00a524eb
Hyphenate some comments.
...
llvm-svn: 85808
2009-11-02 17:10:37 +00:00
00822aabf6
Chain dependencies used to enforce memory order should have latency of 0 (except for true dependency of Store followed by aliased Load... we estimate that case with a single cycle of latency assuming the hardware will bypass)
...
llvm-svn: 85807
2009-11-02 17:06:28 +00:00
433ab09ca3
Add support for BlockAddress values in ARM constant pools.
...
llvm-svn: 85806
2009-11-02 16:59:06 +00:00
e90a4aa703
Prune unnecessary include.
...
llvm-svn: 85805
2009-11-02 16:58:31 +00:00
2729543984
Initilize the machine LICM CSE map upon the first time an instruction is hoisted to
...
the loop preheader. Add instructions which are already in the preheader block that
may be common expressions of those that are hoisted out. These does get a few more
instructions CSE'ed.
llvm-svn: 85799
2009-11-02 08:09:49 +00:00
115cfc07d5
These are done / no longer care.
...
llvm-svn: 85798
2009-11-02 07:58:25 +00:00
f6f1b37f9f
Add an entry.
...
llvm-svn: 85797
2009-11-02 07:51:19 +00:00
9d49f0c858
improve IPSCCP to be able to propagate the result of "!mayBeOverridden"
...
function to calls of that function, regardless of whether it has local
linkage or has its address taken. Not escaping should only affect
whether we make an aggressive assumption about the arguments to a
function, not whether we can track the result of it.
llvm-svn: 85795
2009-11-02 07:33:59 +00:00
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
ab7c709f43
This fixes functions like
...
void f (int a1, int a2, int a3, int a4, int a5,...)
In ARMTargetLowering::LowerFormalArguments if the function has 4 or
more regular arguments we used to set VarArgsFrameIndex using an
offset of 0, which is only correct if the function has exactly 4
regular arguments.
llvm-svn: 85590
2009-10-30 14:33:14 +00:00
1cf0b03064
Add ARM codegen for indirect branches.
...
clang/test/CodeGen/indirect-goto.c runs! (unoptimized)
llvm-svn: 85577
2009-10-30 05:45:42 +00:00
b0cad70c24
Most stack straces don't need 3 digits worth of levels.
...
llvm-svn: 85575
2009-10-30 02:45:10 +00:00
64997909a6
Don't delete blocks which have their address taken.
...
llvm-svn: 85572
2009-10-30 02:13:27 +00:00
e0a8b8fc56
Mention if a block has its address taken in debug output.
...
llvm-svn: 85571
2009-10-30 02:08:26 +00:00
43c574068b
Simplify this code and avoid an extra space character in the output.
...
llvm-svn: 85568
2009-10-30 02:01:10 +00:00
4246f00e0c
Add support for BlockAddress static initializers.
...
llvm-svn: 85562
2009-10-30 01:45:18 +00:00
2f8ccf2bf4
Add a FIXME comment.
...
llvm-svn: 85559
2009-10-30 01:38:20 +00:00
c83dc4b506
Add some comments.
...
llvm-svn: 85558
2009-10-30 01:34:35 +00:00
f7c4299312
Initial x86 support for BlockAddresses.
...
llvm-svn: 85557
2009-10-30 01:28:02 +00:00
6c9388011b
Initial target-independent CodeGen support for BlockAddresses.
...
llvm-svn: 85556
2009-10-30 01:27:03 +00:00
8578068302
Dial back the realignment a bit.
...
llvm-svn: 85546
2009-10-30 00:08:40 +00:00
faa7660fd0
Between scheduling regions, correctly maintain anti-dep breaking state so that we don't incorrectly rename registers that span these regions.
...
llvm-svn: 85537
2009-10-29 23:30:59 +00:00
05efd893db
Remove some unnecessary spaces in debug output.
...
llvm-svn: 85536
2009-10-29 23:30:06 +00:00
554a75a973
Move some code from being emitted as boilerplate duplicated in every
...
*ISelDAGToDAG.cpp to being regular code in SelectionDAGISel.cpp.
llvm-svn: 85530
2009-10-29 22:30:23 +00:00
9f1b2d4619
Fix a couple of bugs in aggressive anti-dep breaking.
...
llvm-svn: 85522
2009-10-29 19:17:04 +00:00
94f8f87890
Refactor complicated predicate into a separate function.
...
llvm-svn: 85519
2009-10-29 18:40:06 +00:00
78c327f171
First bitcase use may not lead to a dbg.declare intrinsic. Iterate uses until one find's dbg.declare intrinsic.
...
Patch by Sunae Seo.
llvm-svn: 85518
2009-10-29 18:20:34 +00:00
453d64c9f5
Rename usesCustomDAGSchedInserter to usesCustomInserter, and update a
...
bunch of associated comments, because it doesn't have anything to do
with DAGs or scheduling. This is another step in decoupling MachineInstr
emitting from scheduling.
llvm-svn: 85517
2009-10-29 18:10:34 +00:00
104f57cae8
Refactor the code for unfolding a load into a separate function.
...
llvm-svn: 85515
2009-10-29 17:47:20 +00:00
c44768ae69
Reapply r85338.
...
llvm-svn: 85514
2009-10-29 17:39:46 +00:00
5e1db97573
Fix MSVC build.
...
llvm-svn: 85505
2009-10-29 12:55:32 +00:00
b43a43a8fd
Apply some cleanups. No functionality changes.
...
llvm-svn: 85498
2009-10-29 07:35:15 +00:00
6747b4ca30
add sanity check for indbr.
...
llvm-svn: 85496
2009-10-29 05:53:32 +00:00
312748848f
just for the hell of it, allow globalopt to statically evaluate
...
static constructors with indirect gotos :)
llvm-svn: 85495
2009-10-29 05:51:50 +00:00
0c778f70e9
add interpreter support for indirect goto / blockaddress. The interpreter
...
now correctly runs clang's test/CodeGen/indirect-goto.c. The JIT will abort
on it until someone feels compelled to implement this.
llvm-svn: 85488
2009-10-29 05:26:09 +00:00
0590837d34
fix 80-col.
...
llvm-svn: 85480
2009-10-29 04:41:24 +00:00
d74ffbde67
Explicitly convert to double to suppress Visual C++ 2008 build error C2668 pow is ambiguous call to overloaded function
...
llvm-svn: 85478
2009-10-29 03:43:06 +00:00
b352d76480
To get more thorough testing from llc-beta nightly runs, do dynamic stack
...
realignment regardless of whether it's strictly necessary.
llvm-svn: 85476
2009-10-29 02:41:21 +00:00
cad78a7634
When the function is doing dynamic stack realignment, the spill slot will be
...
indexed via the stack pointer, even if a frame pointer is present. Update the
heuristic to place it nearest the stack pointer in that case, rather than
nearest the frame pointer.
llvm-svn: 85474
2009-10-29 02:33:47 +00:00
ee8b951e73
teach various passes about blockaddress. We no longer
...
crash on any clang tests.
llvm-svn: 85465
2009-10-29 01:21:20 +00:00
8829fc3a18
When there is a 2-instruction spill sequence, record
...
the second (store) instruction in SpillSlotToUsesMap
consistently. I don't think this matters functionally,
but it's cleaner and Evan wants it this way.
llvm-svn: 85463
2009-10-29 01:15:40 +00:00
3505c94923
Don't put in these EH changes.
...
llvm-svn: 85460
2009-10-29 00:37:35 +00:00
be060382e9
teach ValueMapper about BlockAddress', making bugpoint a lot more useful.
...
llvm-svn: 85458
2009-10-29 00:31:02 +00:00
cf5a47d63d
unindent massive blocks, no functionality change.
...
llvm-svn: 85457
2009-10-29 00:28:30 +00:00
a8eceedb82
Reverting r85338 for now. It's causing a bootstrap failure on PPC darwin9.
...
--- Reverse-merging r85338 into '.':
U lib/CodeGen/SimpleRegisterCoalescing.cpp
U lib/CodeGen/SimpleRegisterCoalescing.h
llvm-svn: 85454
2009-10-29 00:22:16 +00:00
3903320a48
Add a hasAddressTaken for BasicBlock.
...
llvm-svn: 85449
2009-10-29 00:09:08 +00:00
3ab552ec74
Reimplement BranchFolding change to avoid tail merging for a 1 instruction
...
common tail, except when the OptimizeForSize function attribute is present.
Radar 7338114.
llvm-svn: 85441
2009-10-28 22:10:20 +00:00
14f162d9dc
When we generate spill code, then decide we don't need
...
to spill after all, we weren't handling 2-instruction
spill sequences correctly (PPC Altivec). We need to
remove the store in this case. Removing the other
instruction(s) would be goodness but is not needed for
correctness, and isn't done here. 7331562.
llvm-svn: 85437
2009-10-28 21:56:18 +00:00
1fd4c577d2
Make sure we return the right sized type here.
...
llvm-svn: 85436
2009-10-28 21:32:16 +00:00
97b9312663
Revert r85346 change to control tail merging by CodeGenOpt::Level.
...
I'm going to redo this using the OptimizeForSize function attribute.
llvm-svn: 85426
2009-10-28 20:46:46 +00:00
0d025421cd
Extend getMallocArraySize() to determine the array size if the malloc argument is:
...
ArraySize * ElementSize
ElementSize * ArraySize
ArraySize << log2(ElementSize)
ElementSize << log2(ArraySize)
Refactor isArrayMallocHelper and delete isSafeToGetMallocArraySize, so that there is only 1 copy of the malloc array determining logic.
Update users of getMallocArraySize() to not bother calling isArrayMalloc() as well.
llvm-svn: 85421
2009-10-28 20:18:55 +00:00
e30ed53c05
Make AntiDepReg.h internal.
...
llvm-svn: 85412
2009-10-28 18:29:54 +00:00
73789b848d
Add a Thumb BRIND pattern. Change the ARM BRIND assembly to separate the
...
opcode and operand with a tab. Check for these instructions in the usual
places.
llvm-svn: 85411
2009-10-28 18:26:41 +00:00
6203c6868f
fconsts and fconstd are obviously re-materializable.
...
llvm-svn: 85410
2009-10-28 18:19:56 +00:00
294aea709e
Cleanup now that frame index scavenging via post-pass is working for ARM and Thumb2.
...
llvm-svn: 85406
2009-10-28 17:33:28 +00:00
ffd561bc2d
llvm.dbg.global_variables do not exist anymore.
...
llvm-svn: 85402
2009-10-28 16:51:52 +00:00
fbaac77c70
add a new 'SetCurrentDebugType' API (requested by Andrew Haley for JIT
...
stuff) to programmatically control the current debug flavor. While
I'm at it, doxygenate Debug.h and clean it up.
llvm-svn: 85395
2009-10-28 15:32:19 +00:00
14ca753e28
Don't call SDNode::isPredecessorOf when it isn't necessary. If the load's
...
chains have no users, they can't be predecessors of the condition.
llvm-svn: 85394
2009-10-28 15:28:02 +00:00
57780dfdfc
Simplify this code: if the unfolded load can't be hoisted, just delete
...
the new instructions and leave the old one in place.
llvm-svn: 85393
2009-10-28 15:23:36 +00:00
1042ca112f
No newline at end of file.
...
llvm-svn: 85390
2009-10-28 15:04:53 +00:00
ecc60b80b0
Update CMake file.
...
llvm-svn: 85389
2009-10-28 13:29:18 +00:00
2b2bd28973
Treat lifetime begin/end markers as allocations/frees respectively for the
...
purposes for GVN/DSE.
llvm-svn: 85383
2009-10-28 07:05:35 +00:00
175308c43e
Add ABCD, a generalized implementation of the Elimination of Array Bounds
...
Checks on Demand algorithm which looks at arbitrary branches instead of loop
iterations. This is GSoC work by Andre Tavares with only editorial changes
applied!
llvm-svn: 85382
2009-10-28 07:03:15 +00:00
ec6d7c945d
Give ARMISD::EH_SJLJ_LONGJMP and EH_SJLJ_SETJMP names.
...
llvm-svn: 85381
2009-10-28 06:55:03 +00:00
fc16e5a98f
Be more careful about invariance reasoning on "store" queries. Stores still need
...
to depend on Ref and ModRef calls within the invariant region.
llvm-svn: 85380
2009-10-28 06:30:52 +00:00
f64db3e1d0
X86 palignr intrinsics immediate field is in bits. ISel must transform it into bytes.
...
llvm-svn: 85379
2009-10-28 06:30:34 +00:00
d0e86d57c1
Add trivial support for the invariance intrinsics to memdep. This logic is
...
purely local for now.
llvm-svn: 85378
2009-10-28 06:18:42 +00:00
5956dc87d3
add bitcode reader support for blockaddress. We can now fully
...
round trip blockaddress through .ll and .bc files, so add a testcase.
There are still a bunch of places in the optimizer and other places
that need to be updated to work with these constructs, but at least
the basics are in now.
llvm-svn: 85377
2009-10-28 05:53:48 +00:00
f540d74b58
bitcode writer support for blockaddress.
...
llvm-svn: 85376
2009-10-28 05:24:40 +00:00
a91a563530
Previously, all operands to Constant were themselves constant.
...
In the new world order, BlockAddress can have a BasicBlock operand.
This doesn't permute much, because if you have a ConstantExpr (or
anything more specific than Constant) we still know the operand has
to be a Constant.
llvm-svn: 85375
2009-10-28 05:14:34 +00:00
2cb85b4020
'static const void *X = &&y' can only be put in the
...
readonly section if a reference to the containing function
is valid in the readonly section.
llvm-svn: 85370
2009-10-28 04:12:16 +00:00
cd139c0373
Rewrite SelectionDAG::isPredecessorOf to be iterative instead of
...
recursive to avoid consuming extraordinary amounts of stack space
when processing tall graphs.
llvm-svn: 85369
2009-10-28 03:44:30 +00:00
3432c62d93
full asmparser support for blockaddress. We can now do:
...
$ llvm-as foo.ll -d -disable-output
which reads and prints the .ll file. BC encoding is the
next project. Testcase will go in once that works.
llvm-svn: 85368
2009-10-28 03:39:23 +00:00
214cc709bc
asmprinter support for BlockAddress.
...
llvm-svn: 85367
2009-10-28 03:38:12 +00:00
3a39530445
when we tear down a module, we need to be careful to
...
zap BlockAddress values.
llvm-svn: 85366
2009-10-28 03:37:35 +00:00
1b44f10ff4
Teach MachineLICM to unfold loads from constant memory from
...
otherwise unhoistable instructions in order to allow the loads
to be hoisted.
llvm-svn: 85364
2009-10-28 03:21:57 +00:00
4a609f3cef
Use fconsts and fconstd to materialize small fp constants.
...
llvm-svn: 85362
2009-10-28 01:44:26 +00:00
83896a59e1
Add a second ValueType argument to isFPImmLegal.
...
llvm-svn: 85361
2009-10-28 01:43:28 +00:00
4b46cbfc23
Mark dead physregdefs dead immediately. This helps MachineSink and
...
MachineLICM and other things which run before LiveVariables is run.
llvm-svn: 85360
2009-10-28 01:13:53 +00:00
75d6a4afd2
Allow constants of different types to share constant pool entries
...
if they have compatible encodings.
llvm-svn: 85359
2009-10-28 01:12:16 +00:00
deab9dee12
Remove getIEEEFloatParts and getIEEEDoubleParts. They are not needed.
...
llvm-svn: 85358
2009-10-28 01:08:17 +00:00
9c6bc1f563
Update SystemZ to use PSW following the way x86 uses EFLAGS. Besides
...
eliminating a use of MVT::Flag, this is needed for an upcoming CodeGen
change.
This unfortunately requires SystemZ to switch to the list-burr
scheduler, in order to handle the physreg defs properly, however
that's what LLVM has available at this time.
llvm-svn: 85357
2009-10-28 00:55:57 +00:00
e4b80c9bee
Add an indirect branch pattern for ARM. Testcase will be coming soon.
...
llvm-svn: 85355
2009-10-28 00:37:03 +00:00
d04cb6d0fa
rename indbr -> indirectbr to appease the residents of #llvm.
...
llvm-svn: 85351
2009-10-28 00:19:10 +00:00
31b132c0b7
IR support for the new BlockAddress constant kind. This is
...
untested and there is no way to use it, next up: doing battle
with asmparser.
llvm-svn: 85349
2009-10-28 00:01:44 +00:00
9693f9d465
Record CodeGen optimization level in the BranchFolding pass so that we can
...
use it to control tail merging when there is a tradeoff between performance
and code size. When there is only 1 instruction in the common tail, we have
been merging. That can be good for code size but is a definite loss for
performance. Now we will avoid tail merging in that case when the
optimization level is "Aggressive", i.e., "-O3". Radar 7338114.
Since the IfConversion pass invokes BranchFolding, it too needs to know
the optimization level. Note that I removed the RegisterPass instantiation
for IfConversion because it required a default constructor. If someone
wants to keep that for some reason, we can add a default constructor with
a hard-wired optimization level.
llvm-svn: 85346
2009-10-27 23:49:38 +00:00
f6ee7bea88
Rename lib/VMCore/ConstantsContext.h:ValueMap<> to ConstantUniqueMap<> to avoid
...
colliding with llvm/ADT/ValueMap.h:ValueMap<>.
llvm-svn: 85344
2009-10-27 23:45:55 +00:00
2e5198ff09
Add new note.
...
llvm-svn: 85341
2009-10-27 23:30:07 +00:00
cd3939ef0e
Fixed a bug in the coalescer where intervals were occasionally merged despite a real interference. This fixes rdar://problem/7157961.
...
llvm-svn: 85338
2009-10-27 23:16:58 +00:00
5607d2cb54
Enable virtual register based frame index scavenging by default for ARM & T2.
...
llvm-svn: 85335
2009-10-27 22:52:29 +00:00
fd2730ee8c
Move and clarify note.
...
llvm-svn: 85334
2009-10-27 22:48:31 +00:00
c1403a0024
Infrastructure for dynamic stack realignment on ARM. For now, this is off by
...
default behind a command line option. This will enable better performance for
vectors on NEON enabled processors.
llvm-svn: 85333
2009-10-27 22:45:39 +00:00
2974f63cb5
Note corrected.
...
llvm-svn: 85332
2009-10-27 22:43:24 +00:00
cd4d148040
Modify note.
...
llvm-svn: 85331
2009-10-27 22:40:45 +00:00
a205402c16
Add a note.
...
llvm-svn: 85329
2009-10-27 22:34:43 +00:00
11cf3f4a27
Factor out redundancy from clone() implementations.
...
llvm-svn: 85327
2009-10-27 22:16:29 +00:00
a5e078b677
Update the MachineBasicBlock CFG for an indirect branch.
...
llvm-svn: 85325
2009-10-27 22:10:34 +00:00
a4374e66f0
Add CodeGen support for indirect branches.
...
llvm-svn: 85323
2009-10-27 21:56:26 +00:00
6d8b15bebc
make the build build.
...
llvm-svn: 85319
2009-10-27 21:43:39 +00:00
67c902147e
Add new APFloat methods that return sign, exp, and mantissa of ieee float and double values.
...
llvm-svn: 85318
2009-10-27 21:35:42 +00:00
c5c281ea44
Random updates to passes for indbr, I need blockaddress before I can do much more.
...
llvm-svn: 85316
2009-10-27 21:27:42 +00:00
e8628a0206
cppbackend support for indbr
...
llvm-svn: 85312
2009-10-27 21:24:48 +00:00
42c979e8d3
CBE support for indbr.
...
llvm-svn: 85311
2009-10-27 21:21:06 +00:00
cde65ec581
Similar to r85280, do not clear the "S" bit for RSBri and RSBrs.
...
llvm-svn: 85299
2009-10-27 20:51:49 +00:00
cfeaa48642
Do not held on to DenseMap slot accross map insertion. The insertion may cause the map to grow rending the slot invalid.
...
Use this opportunity to use ValueMap instead of DenseMap.
llvm-svn: 85298
2009-10-27 20:47:17 +00:00
4f36affe5f
Set condition code bits of BL and BLr9 to 0b1110 (ALways) to distinguish between
...
BL_pred and BLr9_pred.
llvm-svn: 85297
2009-10-27 20:45:15 +00:00
26076a8f10
don't use stdio
...
llvm-svn: 85296
2009-10-27 20:42:54 +00:00
4567db45b8
Change the JIT to compile eagerly by default as agreed in
...
http://llvm.org/PR5184 , and beef up the comments to describe what both options
do and the risks of lazy compilation in the presence of threads.
llvm-svn: 85295
2009-10-27 20:30:28 +00:00
f390e04a47
Rename MallocFreeHelper as MemoryBuiltins
...
llvm-svn: 85286
2009-10-27 20:05:49 +00:00
16993aa30b
Do away with addLegalFPImmediate. Add a target hook isFPImmLegal which returns true if the fp immediate can be natively codegened by target.
...
llvm-svn: 85281
2009-10-27 19:56:55 +00:00
b5044ce5a9
Do not clear the "S" bit for RSCri and RSCrs. They inherit from the "sI"
...
instruction format that already takes care of setting this.
llvm-svn: 85280
2009-10-27 19:52:03 +00:00
3ed871fe62
add enough support for indirect branch for the feature test to pass
...
(assembler,asmprinter, bc reader+writer) and document it. Codegen
currently aborts on it.
llvm-svn: 85274
2009-10-27 19:13:16 +00:00
df5dcdaa10
Explicitly specify 0b00, i.e, zero rotation, as the rotate filed (Inst{11-10})
...
for the r/rr fragment of the multiclass AI_unary_rrot/AI_bin_rrot.
llvm-svn: 85271
2009-10-27 18:44:24 +00:00
a789f17ffc
Remove unnecessary gotos to fall-thru successors.
...
llvm-svn: 85257
2009-10-27 17:40:24 +00:00
b678a56fef
Test commit. Added '.' to the comment line.
...
llvm-svn: 85255
2009-10-27 17:25:15 +00:00
26a7ae4fba
Type.h doesn't need to #include LLVMContext.h
...
llvm-svn: 85254
2009-10-27 17:08:31 +00:00
0997991252
pseudosourcevalue is also still using getGlobalContext(), so it isn't
...
thread safe either.
llvm-svn: 85253
2009-10-27 17:02:08 +00:00
fb22a85baf
apparently the X86 JIT isn't fully contextized, it is still using getGlobalContext() :(
...
llvm-svn: 85252
2009-10-27 17:01:03 +00:00
d90d169f9d
Correctly align double arguments in the stack.
...
llvm-svn: 85235
2009-10-27 14:09:44 +00:00
c6b3b25f94
Fix a pretty serious misfeature of the inliner: if it inlines a function
...
with multiple return values it inserts a PHI to merge them all together.
However, if the return values are all the same, it ends up with a pointless
PHI and this pointless PHI happens to really block SRoA from happening in
at least a silly C++ example written by Doug, but probably others. This
fixes rdar://7339069.
llvm-svn: 85206
2009-10-27 05:39:41 +00:00
2b0a49a682
VS build fix, patch by Marius Wachtler.
...
llvm-svn: 85197
2009-10-27 02:14:13 +00:00
e3e08e0f90
Fix OProfileJITEventListener after r85182.
...
llvm-svn: 85192
2009-10-27 01:06:51 +00:00
7a50b280c1
Add objectsize intrinsic and hook it up through codegen. Doesn't
...
do anything than return "I don't know" at the moment.
llvm-svn: 85189
2009-10-27 00:52:25 +00:00
538984c1c3
Now VFP instructions.
...
llvm-svn: 85186
2009-10-27 00:20:49 +00:00
f808106bbe
Add braces to avoid ambiguous else.
...
llvm-svn: 85185
2009-10-27 00:11:02 +00:00
b02bdb4552
Change Thumb1 and Thumb2 instructions to separate opcode from operands with a tab instead of a space.
...
llvm-svn: 85184
2009-10-27 00:08:59 +00:00
bf43f6543d
Automatically do the equivalent of freeMachineCodeForFunction(F) when F is
...
being destroyed. This allows users to run global optimizations like globaldce
even after some functions have been jitted.
This patch also removes the Function* parameter to
JITEventListener::NotifyFreeingMachineCode() since it can cause that to be
called when the Function is partially destroyed. This change will be even more
helpful later when I think we'll want to allow machine code to actually outlive
its Function.
llvm-svn: 85182
2009-10-27 00:03:05 +00:00
762195bd01
Rename MallocHelper as MallocFreeHelper, since it now also identifies calls to free()
...
llvm-svn: 85181
2009-10-26 23:58:56 +00:00
03b5de67b0
Add a straight-forward implementation of SCCVN for aggressively eliminating scalar redundancies.
...
llvm-svn: 85179
2009-10-26 23:55:47 +00:00
13edef55b6
Change ARM asm strings to separate opcode from operands with a tab instead of a space.
...
llvm-svn: 85178
2009-10-26 23:45:59 +00:00
de5ad42aa1
Remove FreeInst.
...
Remove LowerAllocations pass.
Update some more passes to treate free calls just like they were treating FreeInst.
llvm-svn: 85176
2009-10-26 23:43:48 +00:00
a33fa47141
Try to get ahead of Johnny Chen and pro-actively add some more ARM encoding
...
bits. Johnny, please review -- I do not have a good track record of getting
these right.
llvm-svn: 85173
2009-10-26 22:59:12 +00:00
1de6a1f7d2
Fix ARM encoding typo: Opcod3 is not passed to ASuI parent class.
...
Patch by Johnny Chen.
llvm-svn: 85169
2009-10-26 22:42:13 +00:00
a6aba77e0a
Add more ARM instruction encodings for 's' bit set and "rs" register encoding
...
bits. Patch by Johnny Chen.
llvm-svn: 85167
2009-10-26 22:34:44 +00:00
e056d1077e
Allow the aggressive anti-dep breaker to process the same region multiple times. This is necessary because new anti-dependencies are exposed when "current" ones are broken.
...
llvm-svn: 85166
2009-10-26 22:31:16 +00:00
34e38afa96
Simplify this code. LoopDeletion doesn't need to explicit check that
...
the loop exiting block dominates the latch block; if ScalarEvolution
can prove that the trip-count is finite, that's sufficient.
llvm-svn: 85165
2009-10-26 22:18:58 +00:00
672927f393
Code that checks WillNotOverflowSignedAdd before creating an Add
...
can safely use the NSW bit on the Add.
llvm-svn: 85164
2009-10-26 22:14:22 +00:00
ce8f626f82
Update CMake files.
...
llvm-svn: 85161
2009-10-26 22:06:01 +00:00
3b7ba5f35b
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
661ea989e9
Define virtual destructor in *.cpp file.
...
llvm-svn: 85146
2009-10-26 19:41:00 +00:00
de11f36ab7
Add aggressive anti-dependence breaker. Currently it is not the default for any target. Enable with -break-anti-dependencies=all.
...
llvm-svn: 85145
2009-10-26 19:32:42 +00:00
6a1d9eace9
Check in the experimental GEP splitter pass. This pass splits complex
...
GEPs (more than one non-zero index) into simple GEPs (at most one
non-zero index). In some simple experiments using this it's not
uncommon to see 3% overall code size wins, because it exposes
redundancies that can be eliminated, however it's tricky to use
because instcombine aggressively undoes the work that this pass does.
llvm-svn: 85144
2009-10-26 19:12:14 +00:00
0ae1b2961d
Revert r85134, it breaks mingw build
...
llvm-svn: 85138
2009-10-26 18:40:24 +00:00
9aba0d9988
When checking whether a def of an aliased register is dead, ask the
...
machineinstr whether the aliased register is dead, rather than the original
register is dead. This allows it to get the correct answer when examining
an instruction like this:
CALLpcrel32 <ga:foo>, %AL<imp-def>, %EAX<imp-def,dead>
where EAX is dead but a subregister of it is still live. This fixes PR5294.
llvm-svn: 85135
2009-10-26 18:26:18 +00:00
752aea6513
Make PIC16 overlay a loadable pass.
...
llvm-svn: 85134
2009-10-26 18:22:59 +00:00
d6dd2a0840
Do not use expensive sort().
...
llvm-svn: 85130
2009-10-26 17:09:00 +00:00
8370485db9
Break anti-dependence breaking out into its own class.
...
llvm-svn: 85127
2009-10-26 16:59:04 +00:00
c80b5c8707
Add support to encode type info using llvm::Constant.
...
Patch by Talin!
llvm-svn: 85126
2009-10-26 16:54:35 +00:00
6a10d5ebd3
Fix a typo in a comment.
...
llvm-svn: 85120
2009-10-26 15:55:24 +00:00
683eed3286
reapply r85085 with a bugfix to avoid infinite looping.
...
All of the 'demorgan' related xforms need to use
dyn_castNotVal, not m_Not.
llvm-svn: 85119
2009-10-26 15:40:07 +00:00
d632f89596
Make LSR's OptimizeShadowIV ignore induction variables with negative
...
strides for now, because it doesn't handle them correctly. This fixes a
miscompile of SingleSource/Benchmarks/Misc-C++/ray.
This problem was usually hidden because indvars transforms such induction
variables into negations of canonical induction variables.
llvm-svn: 85118
2009-10-26 15:32:57 +00:00
b81cb2a9a8
- Revert some changes from 85044, 85045, and 85047 that broke x86_64 tests and
...
bootstrapping. It's not safe to leave identity subreg_to_reg and insert_subreg
around.
- Relax register scavenging to allow use of partially "not-live" registers. It's
common for targets to operate on registers where the top bits are undef. e.g.
s0 =
d0 = insert_subreg d0<undef>, s0, 1
...
= d0
When the insert_subreg is eliminated by the coalescer, the scavenger used to
complain. The previous fix was to keep to insert_subreg around. But that's
brittle and it's overly conservative when we want to use the scavenger to
allocate registers. It's actually legal and desirable for other instructions
to use the "undef" part of d0. e.g.
s0 =
d0 = insert_subreg d0<undef>, s0, 1
...
s1 =
= s1
= d0
We probably need add a "partial-undef" marker on machine operand so the
machine verifier would not complain.
llvm-svn: 85091
2009-10-26 04:56:07 +00:00
8014a728b9
Revert 85085. It causes infinite looping during llvm-gcc build.
...
llvm-svn: 85090
2009-10-26 03:51:32 +00:00
56869f22c4
Move DataTypes.h to include/llvm/System, update all users. This breaks the last
...
direct inclusion edge from System to Support.
llvm-svn: 85086
2009-10-26 01:35:46 +00:00
2e6564d6ff
Implement PR3266 & PR5276, folding:
...
not (or (icmp, icmp)) -> and(icmp, icmp)
llvm-svn: 85085
2009-10-26 01:06:31 +00:00
e38317f432
fix PR5295 where the .ll parser didn't reject a function after a global
...
or global after a function with conflicting names. Update some testcases
that were accidentally depending on this behavior.
llvm-svn: 85081
2009-10-25 23:22:50 +00:00
3d913cda5f
Suppress -Asserts warning.
...
llvm-svn: 85078
2009-10-25 23:11:06 +00:00
f4c2a9f02a
fix PR5186: the JIT shouldn't try to codegen available_externally
...
functions it should just look them up like declarations.
llvm-svn: 85077
2009-10-25 23:06:42 +00:00
083524ae5c
Remove unused includes.
...
llvm-svn: 85074
2009-10-25 22:38:41 +00:00
5bde1cbd93
of -> or
...
llvm-svn: 85065
2009-10-25 19:14:48 +00:00
5a9788b979
80-column cleanup
...
llvm-svn: 85064
2009-10-25 18:55:46 +00:00
b18a46806b
Reapply 85006 with a minor fix.
...
llvm-svn: 85052
2009-10-25 08:14:11 +00:00
2e7dee5f23
Add ARM getMatchingSuperRegClass to handle S / D / Q cross regclass coalescing.
...
llvm-svn: 85049
2009-10-25 07:53:28 +00:00
5d1b849658
Don't forget subreg indices when folding load / store.
...
llvm-svn: 85048
2009-10-25 07:52:27 +00:00
4c799e65d7
Use isIdentityCopy. Fix a bozo bug (flipped condition) in InvalidateRegDef.
...
llvm-svn: 85047
2009-10-25 07:51:47 +00:00
5d2245b210
Code clean up.
...
llvm-svn: 85046
2009-10-25 07:49:57 +00:00
19348415c4
Do not delete identity insert_subreg even if dest is virtual. Let later passes delete them. This avoids register scavenger complain.
...
llvm-svn: 85045
2009-10-25 07:48:51 +00:00
974e12b2d3
Remove includes of Support/Compiler.h that are no longer needed after the
...
VISIBILITY_HIDDEN removal.
llvm-svn: 85043
2009-10-25 06:57:41 +00:00
02d5f77d26
Remove VISIBILITY_HIDDEN from class/struct found inside anonymous namespaces.
...
Chris claims we should never have visibility_hidden inside any .cpp file but
that's still not true even after this commit.
llvm-svn: 85042
2009-10-25 06:33:48 +00:00
13b8b56dd4
this is done.
...
llvm-svn: 85041
2009-10-25 06:17:51 +00:00
d8e8fb4d64
Teach FoldBitCast to be able to handle bitcasts from (e.g.) i128 -> <4 x float>.
...
This allows us to simplify this:
union vec2d {
double e[2];
double v __attribute__((vector_size(16)));
};
typedef union vec2d vec2d;
static vec2d a={{1,2}}, b={{3,4}};
vec2d foo () {
return (vec2d){ .v = a.v + b.v * (vec2d){{5,5}}.v };
}
down to:
define %0 @foo() nounwind ssp {
entry:
%mrv5 = insertvalue %0 undef, double 1.600000e+01, 0 ; <%0> [#uses=1]
%mrv6 = insertvalue %0 %mrv5, double 2.200000e+01, 1 ; <%0> [#uses=1]
ret %0 %mrv6
}
instead of:
define %0 @foo() nounwind ssp {
entry:
%mrv5 = insertvalue %0 undef, double extractelement (<2 x double> fadd (<2 x double> fmul (<2 x double> bitcast (<1 x i128> <i128 85174437667405312423031577302488055808> to <2 x double>), <2 x double> <double 3.000000e+00, double 4.000000e+00>), <2 x double> <double 1.000000e+00, double 2.000000e+00>), i32 0), 0 ; <%0> [#uses=1]
%mrv6 = insertvalue %0 %mrv5, double extractelement (<2 x double> fadd (<2 x double> fmul (<2 x double> bitcast (<1 x i128> <i128 85174437667405312423031577302488055808> to <2 x double>), <2 x double> <double 3.000000e+00, double 4.000000e+00>), <2 x double> <double 1.000000e+00, double 2.000000e+00>), i32 1), 1 ; <%0> [#uses=1]
ret %0 %mrv6
}
llvm-svn: 85040
2009-10-25 06:15:37 +00:00
9d051246eb
move FoldBitCast earlier in the file, and use it instead of
...
ConstantExpr::getBitCast in various places.
llvm-svn: 85039
2009-10-25 06:08:26 +00:00
c5fd5ad4aa
refactor FoldBitCast to reduce nesting and to always return a constantexpr
...
instead of returning null on failure. No functionality change.
llvm-svn: 85038
2009-10-25 06:02:57 +00:00
54d7179a25
Remove ICmpInst::isSignedPredicate which was a reimplementation
...
CmpInst::isSigned.
llvm-svn: 85037
2009-10-25 05:20:17 +00:00
7494b3b1d4
Sink isTrueWhenEqual from ICmpInst to CmpInst. Add a matching isFalseWhenEqual
...
which is equal to !isTrueWhenEqual for ints but not for floats.
llvm-svn: 85036
2009-10-25 03:50:03 +00:00
5027d359e7
When the scavenger is looking for a good candidate location to restore from a
...
spill, it should avoid doing so inside the live range of a virtual register.
llvm-svn: 85026
2009-10-25 00:45:07 +00:00
ef41a1ce3c
MapValue doesn't needs its LLVMContext argument.
...
llvm-svn: 85020
2009-10-24 23:37:16 +00:00
8f4078ba39
Rename isLoopExit to isLoopExiting, for consistency with the wording
...
used elsewhere - an exit block is a block outside the loop branched to
from within the loop. An exiting block is a block inside the loop that
branches out.
llvm-svn: 85019
2009-10-24 23:34:26 +00:00
b979794e4b
Rewrite LoopRotation's SSA updating code using SSAUpdater.
...
llvm-svn: 85016
2009-10-24 23:19:52 +00:00
2cbca4e229
Make DominanceFrontierBase::print's output prettier.
...
llvm-svn: 85011
2009-10-24 20:01:11 +00:00
9e8d42f295
Revert back 85006 for now as it breaks PIC16 tests.
...
llvm-svn: 85008
2009-10-24 18:19:41 +00:00
055c3305a1
Adding support for placing global objects in shared data memory.
...
llvm-svn: 85006
2009-10-24 18:02:44 +00:00
3db7bd27c2
various cleanups suggested by Duncan
...
llvm-svn: 84993
2009-10-24 05:27:19 +00:00
9e2d5b3b8e
fix PR5287, a serious regression from my previous patches. Thanks to
...
Duncan for the nice tiny testcase.
llvm-svn: 84992
2009-10-24 05:22:15 +00:00
e297149e26
Auto-upgrade free instructions to calls to the builtin free function.
...
Update all analysis passes and transforms to treat free calls just like FreeInst.
Remove RaiseAllocations and all its tests since FreeInst no longer needs to be raised.
llvm-svn: 84987
2009-10-24 04:23:03 +00:00
46ed1f8341
80 col violation.
...
llvm-svn: 84986
2009-10-24 02:07:42 +00:00
b4a1347be2
Add some asserts to catch copyRegToReg() fails early
...
llvm-svn: 84983
2009-10-24 00:27:00 +00:00
e2871d69db
Restrict Thumb1 register allocation to low registers, even for instructions that
...
can access the hi regs. Our prologue and epilogue code doesn't know how to
properly handle save/restore of the hi regs, so things go badly when we alloc
them.
llvm-svn: 84982
2009-10-24 00:19:24 +00:00
d638bc310a
Identity copies should not contribute to spill weight.
...
llvm-svn: 84978
2009-10-23 23:09:19 +00:00
22b2c011f9
FIXME no longer applies. R12 and R3 are available for allocation
...
llvm-svn: 84977
2009-10-23 23:07:42 +00:00
d0fc8f809a
Fix http://llvm.org/PR4822 : allow module deletion after a function has been
...
compiled.
When functions are compiled, they accumulate references in the JITResolver's
stub maps. This patch removes those references when the functions are
destroyed. It's illegal to destroy a Function when any thread may still try to
call its machine code.
This patch also updates r83987 to use ValueMap instead of explicit CallbackVHs
and fixes a couple "do stuff inside assert()" bugs from r84522.
llvm-svn: 84975
2009-10-23 22:37:43 +00:00
8acf2956b8
Remove AllocationInst. Since MallocInst went away, AllocaInst is the only subclass of AllocationInst, so it no longer is necessary.
...
llvm-svn: 84969
2009-10-23 21:09:37 +00:00
4ef112be62
APInt-ify the gep scaling code, so that it correctly handles the case where
...
the scale overflows pointer-sized arithmetic. This fixes PR5281.
llvm-svn: 84954
2009-10-23 17:57:43 +00:00
41d00ac45b
Make LoopDeletion check the maximum backedge taken count, rather than the
...
exact backedge taken count, when checking for infinite loops. This allows
it to delete loops with multiple exit conditions.
llvm-svn: 84952
2009-10-23 17:10:01 +00:00
851193b873
some stuff is done, we still have constantexpr simplification to do.
...
llvm-svn: 84943
2009-10-23 07:00:55 +00:00
ccf1e84779
teach libanalysis to simplify vector loads with bitcast sources. This
...
implements something out of Target/README.txt producing:
_foo: ## @foo
movl 4(%esp), %eax
movapd LCPI1_0, %xmm0
movapd %xmm0, (%eax)
ret $4
instead of:
_foo: ## @foo
movl 4(%esp), %eax
movapd _b, %xmm0
mulpd LCPI1_0, %xmm0
addpd _a, %xmm0
movapd %xmm0, (%eax)
ret $4
llvm-svn: 84942
2009-10-23 06:57:37 +00:00
59f94c01dd
enhance FoldReinterpretLoadFromConstPtr to handle loads of up to 32
...
bytes (i256).
llvm-svn: 84941
2009-10-23 06:50:36 +00:00
ed00b80bf8
teach libanalysis to fold int and fp loads from almost arbitrary
...
non-type-safe constant initializers. This sort of thing happens
quite a bit for 4-byte loads out of string constants, unions,
bitfields, and an interesting endianness check from sqlite, which
is something like this:
const int sqlite3one = 1;
# define SQLITE_BIGENDIAN (*(char *)(&sqlite3one)==0)
# define SQLITE_LITTLEENDIAN (*(char *)(&sqlite3one)==1)
# define SQLITE_UTF16NATIVE (SQLITE_BIGENDIAN?SQLITE_UTF16BE:SQLITE_UTF16LE)
all of these macros now constant fold away.
This implements PR3152 and is based on a patch started by Eli, but heavily
modified and extended.
llvm-svn: 84936
2009-10-23 06:23:49 +00:00
8b86efefec
X86 needs critical path anti-dependency breaking.
...
llvm-svn: 84931
2009-10-23 05:57:35 +00:00
0d670b5b92
Commit fixes for half precision I noted in review, so
...
they don't get lost; I don't think the originator has
write access.
llvm-svn: 84928
2009-10-23 04:02:51 +00:00
2981d1b93d
Minor code cleanup.
...
llvm-svn: 84919
2009-10-23 00:59:10 +00:00
6aea1a2a5a
Neuter stack protectors by only checking character arrays. This is what GCC
...
does.
llvm-svn: 84916
2009-10-23 00:01:05 +00:00
02ad4cb32e
Allow the target to select the level of anti-dependence breaking that should be performed by the post-RA scheduler. The default is none.
...
llvm-svn: 84911
2009-10-22 23:19:17 +00:00
b5cd6233b1
Use 'waitpid' instead of 'wait'. Basing Program::Wait() on 'wait()' prevents it being correct within a multithreaded context.
...
This address: PR 5277 (Program::Wait is unsafe to call from multiple threads).
Note: If waitpid() turns out to be non-portable, we can add more autoconf magic, or look into
another solution.
llvm-svn: 84903
2009-10-22 22:16:17 +00:00
ab304c9237
Random include cleanup.
...
llvm-svn: 84898
2009-10-22 21:49:41 +00:00
86dfc11cdf
Fix OProfileJITEventListener after r84054 renamed CompileUnit to Scope.
...
llvm-svn: 84895
2009-10-22 20:57:35 +00:00
4f0b8d7861
Tidying up some code and comments. No functionality change.
...
llvm-svn: 84894
2009-10-22 20:48:59 +00:00
1155fdf6dc
Hide MetadataContext implementation details.
...
llvm-svn: 84886
2009-10-22 19:36:54 +00:00
6da5dbf3c2
Fix getMDs() interface such that it does not expose implementation details.
...
llvm-svn: 84885
2009-10-22 18:55:16 +00:00
084679e70a
Using TrackingVH instead of WeakVH or WeakMetadataVH.
...
llvm-svn: 84884
2009-10-22 18:25:28 +00:00
09c319e269
Sort handler names to ensure deterministic behavior.
...
llvm-svn: 84878
2009-10-22 17:40:37 +00:00
9d763cc3f8
Revert 84843. Evan, this was breaking some of the if-conversion tests.
...
llvm-svn: 84868
2009-10-22 16:52:21 +00:00
7e8017cce7
Verify that the function and exception table have been allocated
...
before freeing them.
llvm-svn: 84859
2009-10-22 14:35:57 +00:00
c77f8634e5
Shift art to the right to keep GCC from complaining about multi-line comments.
...
llvm-svn: 84849
2009-10-22 09:28:49 +00:00
3615b9bef3
Move if-conversion before post-regalloc scheduling so the predicated instruction get scheduled properly.
...
llvm-svn: 84843
2009-10-22 06:48:32 +00:00
943f4f41f2
Load / store multiple was missing opportunites when the load / store bundles are at the end of the bb. Test case is already in, the bug is exposed by subsequent commit.
...
llvm-svn: 84842
2009-10-22 06:47:35 +00:00
cf7e8947e9
move another load optimization from instcombine -> libanalysis.
...
llvm-svn: 84841
2009-10-22 06:44:07 +00:00
51d2f70e32
move 'loading i32 from string' optimization from instcombine
...
to libanalysis. Instcombine shrinking... does this even
make sense???
llvm-svn: 84840
2009-10-22 06:38:35 +00:00
1664a4fd86
Move some constant folding logic for loads out of instcombine into
...
Analysis/ConstantFolding.cpp. This doesn't change the behavior of
instcombine but makes other clients of ConstantFoldInstruction
able to handle loads. This was partially extracted from Eli's patch
in PR3152.
llvm-svn: 84836
2009-10-22 06:25:11 +00:00
5457a96b63
Trim more includes.
...
llvm-svn: 84832
2009-10-22 05:11:00 +00:00
344fcd9d61
Trim include.
...
llvm-svn: 84831
2009-10-22 05:08:49 +00:00
1448799377
fix warning.
...
llvm-svn: 84826
2009-10-22 03:42:27 +00:00
0fffb49d56
Fix getHandlerNames() interface. Now it populate clinet supplied small vector with handler names.
...
llvm-svn: 84820
2009-10-22 01:01:24 +00:00
0f55e9ce2e
Don't generate sbfx / ubfx with negative lsb field. Patch by David Conrad.
...
llvm-svn: 84813
2009-10-22 00:40:00 +00:00
6746d43a3f
Use StringRef to construct MDString.
...
llvm-svn: 84811
2009-10-22 00:22:05 +00:00
c7a962d3b3
fix PR5262.
...
llvm-svn: 84810
2009-10-22 00:17:26 +00:00
f1ad2c618a
Use special DAG-to-DAG preprocessing to allow mem-mem instructions to be selected.
...
Yay for ASCII graphics!
llvm-svn: 84808
2009-10-22 00:16:00 +00:00
8626367e38
Fix null pointer dereference.
...
llvm-svn: 84806
2009-10-22 00:15:17 +00:00
dcb99d310a
Remove meaningless const.
...
Pass StringRef by value.
llvm-svn: 84804
2009-10-22 00:10:15 +00:00
ff97acd8f1
Revert the main portion of r31856. It was causing BranchFolding
...
to break up CFG diamonds by banishing one of the blocks to the end of
the function, which is bad for code density and branch size.
This does pessimize MultiSource/Benchmarks/Ptrdist/yacr2, the
benchmark cited as the reason for the change, however I've examined
the code and it looks more like a case of gaming a particular
branch than of being generally applicable.
llvm-svn: 84803
2009-10-22 00:03:58 +00:00
27e0be274e
Derive metadata hierarchy from Value instead of User.
...
llvm-svn: 84801
2009-10-21 23:57:35 +00:00
966526cbfb
revert r84754, it isn't the right approach. Edwin, please propose
...
patches for fixes like this instead of committing them directly.
llvm-svn: 84799
2009-10-21 23:41:58 +00:00
f2e74df12e
Missing piece of the ARM frame index post-scavenging conditionalization
...
llvm-svn: 84798
2009-10-21 23:40:56 +00:00
8b051b8240
Fix thinko noticed by Chris.
...
llvm-svn: 84797
2009-10-21 23:29:32 +00:00
1cfb958b00
Rename msasm to alignstack per review.
...
llvm-svn: 84795
2009-10-21 23:28:00 +00:00
22146fbc5f
Remove pointless return; at end of function.
...
llvm-svn: 84794
2009-10-21 23:27:54 +00:00
c16c8c707b
The spill restore needs to be resolved to the SP/FP just like the spill
...
llvm-svn: 84792
2009-10-21 22:59:56 +00:00
05536f50dc
Conditionalize ARM/T2 frame index post-scavenging while working out fixes
...
for a few bugs.
llvm-svn: 84791
2009-10-21 22:59:24 +00:00
b378347796
Simplify code. No intended functionality change.
...
llvm-svn: 84790
2009-10-21 22:55:51 +00:00
2505c1e17a
Use StringRef.
...
llvm-svn: 84786
2009-10-21 21:57:13 +00:00
854530a7dd
Most of the NEON shuffle instructions do not support 64-bit element types.
...
llvm-svn: 84785
2009-10-21 21:36:27 +00:00
49914e6eb3
Do not use SmallVector to store MDNode elements.
...
llvm-svn: 84784
2009-10-21 21:25:09 +00:00
b45d621ab1
Revert r84764, it breaks mingw build
...
llvm-svn: 84783
2009-10-21 21:15:18 +00:00
a93ca3c637
Improve handling of immediates by splitting 32-bit immediates into two 16-bit
...
immediate operands when they will fit into the using instruction.
llvm-svn: 84778
2009-10-21 20:44:34 +00:00
9e92857b38
Add DAG printing for RMW stuff debugging
...
llvm-svn: 84776
2009-10-21 19:18:28 +00:00
ccfa3e31f7
RMW preprocessing stuff was incorrect. Grab the stuff from x86 backend and disable some tests until it will be clever enough to handle them.
...
llvm-svn: 84775
2009-10-21 19:17:55 +00:00
5399c2d05e
Implement branch folding
...
llvm-svn: 84774
2009-10-21 19:17:18 +00:00
2983dcb1f2
Cosmetic changes, no functionality changes
...
llvm-svn: 84773
2009-10-21 19:16:49 +00:00
be9e179104
Make changes to rev 84292 as requested by Chris Lattner.
...
Most changes are cleanup, but there is 1 correctness fix:
I fixed InstCombine so that the icmp is removed only if the malloc call is removed (which requires explicit removal because the Worklist won't DCE any calls since they can have side-effects).
llvm-svn: 84772
2009-10-21 19:11:40 +00:00
0db964a3a0
Fix NEON VST2LN instruction encoding.
...
Patch by Johnny Chen.
llvm-svn: 84767
2009-10-21 17:54:01 +00:00
87671da29a
Revert 84732. It was the wrong fix.
...
llvm-svn: 84766
2009-10-21 17:52:34 +00:00
3eb5d33b65
Incorporate various suggestions Chris gave during metadata review.
...
- i < getNumElements() instead of getNumElements() > i
- Make setParent() private
- Fix use of resizeOperands
- Reset HasMetadata bit after removing all metadata attached to an instruction
- Efficient use of iterators
llvm-svn: 84765
2009-10-21 17:33:41 +00:00
c5804f3cb6
Build shared lib instead of an archive.
...
llvm-svn: 84764
2009-10-21 17:27:23 +00:00
bc30143451
Cleanup of frame index scavenging. Better code flow and more accurately
...
handles T2 and ARM use cases.
llvm-svn: 84761
2009-10-21 15:26:21 +00:00
1539a352a6
Fix PR5262: when folding select into PHI, make sure all operands are available
...
in the PHI's Basic Block. This uses a conservative approach, because we don't
have dominator info in instcombine.
llvm-svn: 84754
2009-10-21 10:49:00 +00:00
47ea743618
Add a pass to overlay pic16 data sections for function frame and automatic
...
variables. This pass can be invoked by llvm-ld or opt to traverse over the call graph
to detect what function frames and their automatic variables can be overlaid.
Currently this builds an archive , but needs to be changed to a loadable module.
llvm-svn: 84753
2009-10-21 10:42:44 +00:00
786b15fe12
Match more patterns to movt.
...
llvm-svn: 84751
2009-10-21 08:15:52 +00:00
f781bd8947
Need a comma after imp-use.
...
llvm-svn: 84749
2009-10-21 07:56:02 +00:00
05f77d5006
De-bork CMake build
...
llvm-svn: 84744
2009-10-21 06:01:54 +00:00
8ed7bef409
make GVN work better when TD is not around:
...
"In the existing code, if the load and the value to replace it with are
of different types *and* target data is available, it tries to use the
target data to coerce the replacement value to the type of the load.
Otherwise, it skips all effort to handle the type mismatch and just
feeds the wrongly-typed replacement value to replaceAllUsesWith, which
triggers an assertion.
The patch replaces it with an outer if checking for type mismatch, and
an inner if-else that checks whether target data is available and, if
not, returns false rather than trying to replace the load."
Patch by Kenneth Uildriks!
llvm-svn: 84739
2009-10-21 04:11:19 +00:00
175d04c90f
tidy
...
llvm-svn: 84738
2009-10-21 04:10:24 +00:00
5b5cb92816
Fix some more NEON instruction encoding problems.
...
Thanks to Johnny Chen for discovering the problem.
llvm-svn: 84732
2009-10-21 02:27:20 +00:00
1d7f7d21dc
Do not remove dead metadata for now.
...
llvm-svn: 84731
2009-10-21 02:21:34 +00:00
bd3650cc84
Leave some NEON instruction encoding bits unspecified instead of setting
...
a default value of zero. This is important for decoding the instructions.
Patch by Johnny Chen, with some changes from me, too.
llvm-svn: 84730
2009-10-21 02:15:46 +00:00
682a2d154a
Revert r84658 and r84691. They were causing llvm-gcc bootstrap to fail.
...
llvm-svn: 84727
2009-10-21 01:44:44 +00:00
bc69313909
IPSCCP is missing stuff.
...
llvm-svn: 84725
2009-10-21 01:10:37 +00:00
08fa03c740
Delete the MacOSJITEventListener per echristo's request. It was disabled by
...
default and didn't work anyway.
llvm-svn: 84720
2009-10-21 00:43:48 +00:00
7099d0c19c
Add note
...
llvm-svn: 84713
2009-10-21 00:14:15 +00:00
e43af4a085
Be crazy and assert in case of unsupported modifier passed.
...
llvm-svn: 84712
2009-10-21 00:13:58 +00:00
11074fa73e
Handle external symbols
...
llvm-svn: 84711
2009-10-21 00:13:42 +00:00
cc55b9086d
Distinguish between pcrel imm operands and 'normal' ones. Fix fixes gross weirdness of asmprinting.
...
llvm-svn: 84710
2009-10-21 00:13:25 +00:00
94ba9c27b0
Add basic block operands & jump kinds
...
llvm-svn: 84709
2009-10-21 00:13:05 +00:00
8a06a4e5c8
Ignore all implicit reg operands
...
llvm-svn: 84708
2009-10-21 00:12:44 +00:00
196b0e5431
Add a workaround for different memops prefixes
...
llvm-svn: 84707
2009-10-21 00:12:27 +00:00
4e32bff382
Checkpoint MCInst printer. We (almostly) able to print global / JT / constpool entries
...
llvm-svn: 84706
2009-10-21 00:12:08 +00:00
7fbc0a50d5
Add simple operand printing stuff
...
llvm-svn: 84704
2009-10-21 00:11:27 +00:00
6e78029d82
Add experimental MSP430 MCInstLowering stuff
...
llvm-svn: 84703
2009-10-21 00:11:08 +00:00
daaa2f0c31
Wire up MSP430 printMCInst() method
...
llvm-svn: 84702
2009-10-21 00:10:47 +00:00
b6e2e4fd50
Add MSP430 InstPrinter stub
...
llvm-svn: 84701
2009-10-21 00:10:30 +00:00
13de81ac40
Use proper target data
...
llvm-svn: 84700
2009-10-21 00:10:00 +00:00
4388beb884
Respect src register allocation requirements when breaking anti-dependencies. Remove some dead code.
...
llvm-svn: 84691
2009-10-20 22:50:43 +00:00
0c35dbdf03
Cosmetic changes.
...
s/validName/isValidName/g
s/with an Instruction/to an Instruction/g
s/RegisterMDKind/registerMDKind/g
llvm-svn: 84689
2009-10-20 22:50:27 +00:00
a470eac6a1
Fix -Asserts warning.
...
llvm-svn: 84687
2009-10-20 22:10:05 +00:00
a6faf60831
Fix invalid for vector types fneg(bitconvert(x)) => bitconvert(x ^ sign)
...
transform.
llvm-svn: 84683
2009-10-20 21:37:45 +00:00
16edcd724a
Oops. Backing out 84681 - needs to wait for the indexing patch.
...
llvm-svn: 84682
2009-10-20 21:28:22 +00:00
d8938d1ca8
Added some debugging output to pre-alloc splitting.
...
llvm-svn: 84681
2009-10-20 21:25:13 +00:00
83f5c835ff
Fix another place that calls Loop::contains a lot to construct a sorted
...
container of the blocks and do efficient lookups. This makes
isLoopSimplifyForm much faster on large loops, fixing a significant
compile-time issue in builds with assertions enabled.
llvm-svn: 84673
2009-10-20 20:41:13 +00:00
cccf5084a3
Disable by default while debugging
...
llvm-svn: 84669
2009-10-20 20:31:31 +00:00
7f903681ac
alternate fix for PR5258 which avoids worklist problems, with reduced testcase.
...
llvm-svn: 84667
2009-10-20 20:27:49 +00:00
f3a2b6499e
add cmd line opt to disable frame index reuse for ARM and T2. debug aid.
...
llvm-svn: 84664
2009-10-20 20:19:50 +00:00
b6b8ec769c
Restore LoopUnswitch's block-oriented threshold. LoopUnswitch now checks both
...
the estimated code size and the number of blocks when deciding whether to
do a non-trivial unswitch. This protects it from some very undesirable
worst-case behavior on large numbers of loop-unswitchable conditions, such
as in the testcase in PR5259.
llvm-svn: 84661
2009-10-20 20:06:09 +00:00
baf6dd26ea
Checkpoint more aggressive anti-dependency breaking for post-ra scheduler.
...
llvm-svn: 84658
2009-10-20 19:54:44 +00:00
d2ba321930
Better handle instructions that re-def a scratch register
...
llvm-svn: 84657
2009-10-20 19:52:35 +00:00
3d9d78463c
Following r84485, add Defs = [EFLAGS] to the 32-bit lock instructions too.
...
llvm-svn: 84652
2009-10-20 18:14:49 +00:00
27c669242a
Move the Function*->allocated blocks map from the JITMemoryManager to the
...
JITEmitter.
I'm gradually making Functions auto-remove themselves from the JIT when they're
destroyed. In this case, the Function needs to be removed from the JITEmitter,
but the map recording which Functions need to be removed lived behind the
JITMemoryManager interface, which made things difficult.
This patch replaces the deallocateMemForFunction(Function*) method with a pair
of methods deallocateFunctionBody(void *) and deallocateExceptionTable(void *)
corresponding to the two startFoo/endFoo pairs.
llvm-svn: 84651
2009-10-20 18:13:21 +00:00
5f2581a37a
Register re-use for scavenged frame indices must check for re-deginition
...
of the register in the instruction which kills the scavenged value.
llvm-svn: 84641
2009-10-20 16:33:57 +00:00
4a43e3068d
Make TranslateX86CC return COND_INVALID instead of aborting when it
...
encounters an OEQ or UNE comparison, and update its callers to check
for this return status and recover. This fixes a problem resulting from
the LowerOperation hooks being called from LegalizeVectorOps, because
LegalizeVectorOps only lowers vectors, so OEQ and UNE comparisons may
still be at large. This fixes PR5092.
llvm-svn: 84640
2009-10-20 16:22:37 +00:00
cf10ec951d
Fix PR5258, jump-threading creating invalid PHIs.
...
When an incoming value for a PHI is updated, we must also updated all other
incoming values for the same BB to match, otherwise we create invalid PHIs.
llvm-svn: 84638
2009-10-20 15:42:00 +00:00
729d92bd74
Fix PR4313: IPSCCP was not setting the lattice value for the invoke instruction
...
when the invoke had multiple return values: it set the lattice value only on the
extractvalue.
This caused the invoke's lattice value to remain the default (undefined), and
later propagated to extractvalue's operand, which incorrectly introduces
undefined behavior.
llvm-svn: 84637
2009-10-20 15:15:09 +00:00
3301207a15
Random #include pruning.
...
llvm-svn: 84632
2009-10-20 11:44:38 +00:00
8296e62244
This file is replaeced by PIC16Section.h.
...
llvm-svn: 84628
2009-10-20 09:16:32 +00:00
9351e4f4b2
implement some more easy hooks.
...
llvm-svn: 84614
2009-10-20 06:22:33 +00:00
60d5131653
Implement some hooks, make printOperand abort if unknown modifiers are
...
present.
llvm-svn: 84613
2009-10-20 06:15:28 +00:00
227767b4e5
t2MOVi32imm is currently always lowered by the Thumb2ITBlockPass.
...
llvm-svn: 84611
2009-10-20 05:58:02 +00:00
f0b3d15cfe
Wire up the ARM MCInst printer, for llvm-mc.
...
llvm-svn: 84600
2009-10-20 05:15:36 +00:00
c0964a571b
Re-apply r84295, with fixes to how the loop "top" and "bottom" blocks are
...
tracked. Instead of trying to manually keep track of these locations
while doing complex modifications, just recompute them when they're needed.
This fixes a bug in which the TopMBB and BotMBB were not correctly updated,
leading to invalid transformations.
llvm-svn: 84598
2009-10-20 04:50:37 +00:00
ccc2e0cbda
Trim unnecessary includes.
...
llvm-svn: 84597
2009-10-20 04:23:20 +00:00
d383c2f6ba
Add getTopBlock and getBottomBlock member functions to MachineLoopInfo.
...
llvm-svn: 84596
2009-10-20 04:16:37 +00:00
34f040a575
Now that all ARM subtargets use frame index scavenging, the Thumb1 requires*
...
functions are not needed.
llvm-svn: 84587
2009-10-20 01:32:47 +00:00
fa4bcaec7c
If the physical register being spilled does not have an interval, spill its sub-registers instead.
...
llvm-svn: 84586
2009-10-20 01:31:09 +00:00
84f6235b6f
Enable post-pass frame index register scavenging for ARM and Thumb2
...
llvm-svn: 84585
2009-10-20 01:26:58 +00:00
484d2e9491
lower ARM::MOVi32imm properly.
...
llvm-svn: 84583
2009-10-20 01:11:37 +00:00
43c5589a7e
add support for external symbols. The mc instprinter can now handle
...
reasonable code like Codegen/ARM/2009-02-27-SpillerBug.ll, producing
identical output except for superior formatting of constant pool entries.
llvm-svn: 84582
2009-10-20 00:56:16 +00:00
1b06acbd70
get fancy: support basic block operands. Yay for jumps.
...
llvm-svn: 84579
2009-10-20 00:52:47 +00:00
85ab670644
add supprort for the 'sbit' operand, MOVi apparently has one.
...
llvm-svn: 84577
2009-10-20 00:46:11 +00:00
19c52201bd
add support for instruction predicates.
...
llvm-svn: 84575
2009-10-20 00:42:49 +00:00
2f69ed8f4c
implement printSORegOperand, add lowering for the nasty and despicable MOVi2pieces :)
...
llvm-svn: 84573
2009-10-20 00:40:56 +00:00
772b2f84eb
Refs: A8-598.
...
Leave Inst{11-8}, which represents the starting byte index of the extracted
result in the concatenation of the operands and is left unspecified.
Patch by Johnny Chen.
llvm-svn: 84572
2009-10-20 00:38:19 +00:00
68f495caad
Add missing encoding bits to NLdSt class of instructions.
...
Patch by Johnny Chen.
llvm-svn: 84570
2009-10-20 00:19:08 +00:00
0b4a59fc07
X86 should ignore implicit regs when lowering to MCInst also,
...
no functionality change.
llvm-svn: 84567
2009-10-19 23:35:57 +00:00
bd531262f8
handle addmode4 modifiers, fix a fixme in printRegisterList
...
by ignoring all implicit regs when lowering.
llvm-svn: 84566
2009-10-19 23:31:43 +00:00
d99b6974b9
simplify by using the twine form of GetOrCreateSymbol
...
llvm-svn: 84565
2009-10-19 23:05:23 +00:00
f5f263f1b4
Enable allocation of R3 in Thumb1
...
llvm-svn: 84563
2009-10-19 22:57:03 +00:00
d91c11091d
use EmitLabel instead of text emission
...
llvm-svn: 84562
2009-10-19 22:51:16 +00:00
86dfd73c38
add a twine version of MCContext::GetOrCreateSymbol.
...
llvm-svn: 84561
2009-10-19 22:49:00 +00:00
186c6b0834
lower the ARM::CONSTPOOL_ENTRY pseudo op, giving us constant pool entries
...
like:
@ BB#1:
.align 2
LCPI1_0:
.long L_.str-(LPC0+8)
Note that proper indentation of the label :)
llvm-svn: 84558
2009-10-19 22:33:05 +00:00
5a5a3bc5d4
Adjust the scavenge register spilling to allow the target to choose an
...
appropriate restore location for the spill as well as perform the actual
save and restore.
The Thumb1 target uses this to make sure R12 is not clobbered while a spilled
scavenger register is live there.
llvm-svn: 84554
2009-10-19 22:27:30 +00:00
add5749ac8
add MCInstLower support for lowering ARM::PICADD, a pseudo op for pic stuffola.
...
llvm-svn: 84553
2009-10-19 22:23:04 +00:00
168ad6985e
Refactor lookup_or_add to contain _MUCH_ less duplicated code. Add support for
...
numbering first class aggregate instructions while we're at it.
llvm-svn: 84547
2009-10-19 22:14:22 +00:00
ef2979b1da
add register list and hacked up addrmode #4 support, we now get this:
...
_main:
stmsp! sp!, {r7, lr}
mov r7, sp
sub sp, sp, #4
mov r0, #0
str r0, [sp]
ldr r0, LCPI1_0
bl _printf
ldr r0, [sp]
mov sp, r7
ldmsp! sp!, {r7, pc}
Note the unhappy ldm/stm because of modifiers being ignored.
llvm-svn: 84546
2009-10-19 22:09:23 +00:00
f264f8a21c
revert r84540, fixing build breakage I didn't see because of
...
broken makefile deps :(
llvm-svn: 84544
2009-10-19 21:59:25 +00:00
7ddfdc4133
add addrmode2 support, getting us up to:
...
_main:
stm ,
mov r7, sp
sub sp, sp, #4
mov r0, #0
str r0, [sp]
ldr r0, LCPI1_0
bl _printf
ldr r0, [sp]
mov sp, r7
ldm ,
llvm-svn: 84543
2009-10-19 21:57:05 +00:00
889a6217fa
add jump tables, constant pools and some trivial global
...
lowering stuff. We can now compile hello world to:
_main:
stm ,
mov r7, sp
sub sp, sp, #4
mov r0, #0
str r0,
ldr r0,
bl _printf
ldr r0,
mov sp, r7
ldm ,
Almost looks like arm code :)
llvm-svn: 84542
2009-10-19 21:53:00 +00:00
5c704d505c
Malloc calls are marked NoAlias, so the code below the isMalloc() check makes it redundant. Removing the isMalloc() check.
...
llvm-svn: 84541
2009-10-19 21:47:22 +00:00
e3796a0fee
pass mangler in as a reference instead of a pointer.
...
llvm-svn: 84540
2009-10-19 21:45:31 +00:00
8bd9b9d3fa
reduce #includes
...
llvm-svn: 84536
2009-10-19 21:23:15 +00:00
89d47205b5
add printing support for SOImm operands, getting us to:
...
_main:
stm ,
mov r7, sp
sub sp, sp, #4
mov r0, #0
str r0,
llvm-svn: 84535
2009-10-19 21:21:39 +00:00
1059b5b32d
Simplify some code.
...
llvm-svn: 84533
2009-10-19 21:14:57 +00:00
93e3ef6536
wire up some basic printOperand goodness, giving us stuff like this before
...
we abort:
_main:
stm ,
mov r7, sp
sub sp, sp,
mov r0,
str r0,
llvm-svn: 84532
2009-10-19 20:59:55 +00:00
78393d79a0
add the files that go with the previous rev
...
llvm-svn: 84531
2009-10-19 20:21:05 +00:00
71eb077c15
wire up skeletal support for having llc print instructions
...
through mcinst lowering -> mcinstprinter, when llc is passed
the -enable-arm-mcinst-printer flag. Currently this
is very "aborty".
llvm-svn: 84530
2009-10-19 20:20:46 +00:00
5dab84ca9c
Banish ConstantsLock. It's serving no purpose other than slowing things down
...
at the moment.
llvm-svn: 84529
2009-10-19 20:11:52 +00:00
296544e15f
wire up ARM's printMCInst method. Now llvm-mc should be able to produce
...
"something" when printing MCInsts, it will just be missing all the
operand info.
llvm-svn: 84528
2009-10-19 19:59:05 +00:00
a290778859
stub out a minimal ARMInstPrinter.
...
llvm-svn: 84527
2009-10-19 19:56:26 +00:00
5db7b6a5d4
remove strings from instructions who are never asmprinted.
...
All of these "subreg32" modifier instructions are handled
explicitly by the MCInst lowering phase. If they got to
the asmprinter, they would explode. They should eventually
be replace with correct use of subregs.
llvm-svn: 84526
2009-10-19 19:51:42 +00:00
7d2edada7f
Clean up the JITResolver stub/callsite<->function maps.
...
The JITResolver maps Functions to their canonical stubs and all callsites for
lazily-compiled functions to their target Functions. To make Function
destruction work, I'm going to need to remove all callsites on destruction, so
this patch also adds the reverse mapping for that.
There was an incorrect assumption in here that the only stub for a function
would be the one caused by needing to lazily compile it, while x86-64 far calls
and dlsym-stubs could also cause such stubs, but I didn't look for a test case
that the assumption broke.
This also adds DenseMapInfo<AssertingVH> so I can use DenseMaps instead of
std::maps.
llvm-svn: 84522
2009-10-19 18:49:59 +00:00
8d2e3a28d6
simplify code, reducing string thrashing.
...
llvm-svn: 84521
2009-10-19 18:49:14 +00:00
3334deb19b
switch hidden gv stubs to use MachineModuleInfoMachO instead of a custom map.
...
llvm-svn: 84520
2009-10-19 18:44:38 +00:00
6462adcd20
use MachineModuleInfoMachO for non-lazy gv stubs instead of a private map.
...
llvm-svn: 84519
2009-10-19 18:38:33 +00:00
8383c3d7f3
Revert r84295, this unbreaks llvm-gcc bootstrap on x86-64/linux
...
llvm-svn: 84516
2009-10-19 18:21:09 +00:00
24fcb826bd
remove dead map
...
llvm-svn: 84513
2009-10-19 18:11:25 +00:00
84a493ca21
don't bother trying to avoid emitting redundant constant pool alignment directives.
...
llvm-svn: 84512
2009-10-19 18:08:02 +00:00
66ebfab3ab
remove accidental comment.
...
llvm-svn: 84510
2009-10-19 18:03:41 +00:00
e6da1826e8
emit .subsections_via_symbols through MCStreamer instead of textually.
...
llvm-svn: 84509
2009-10-19 18:03:08 +00:00
ee9399acaa
cleanup doFinalization -> EmitEndOfAsmFile.
...
llvm-svn: 84508
2009-10-19 17:59:19 +00:00
5ca7b345b9
PR 5245 - The imediate size target flag was not set on 3A-prefixed SSSE3 instructions.
...
llvm-svn: 84506
2009-10-19 17:31:16 +00:00
8f986672a1
Fix SplitBlockPredecessors' LoopInfo updating code to handle the case
...
where a loop's header is being split and it has predecessors which are not
contained by the most-nested loop which contains the loop.
This fixes PR5235.
llvm-svn: 84505
2009-10-19 16:04:50 +00:00
b0db9917af
Fix a typo in a comment.
...
llvm-svn: 84504
2009-10-19 14:56:05 +00:00
20a327b31f
Change a few instance variables to be local variables.
...
llvm-svn: 84503
2009-10-19 14:52:05 +00:00
511d2e26dd
Change instnamer to name arguments "arg" instead of "tmp" for clarity, and
...
to name basic blocks "bb" instead of "BB", for consistency.
llvm-svn: 84502
2009-10-19 14:47:32 +00:00
033f01c922
Fix PR5247, "lock addq" pattern (and other atomics), it DOES modify EFLAGS.
...
LLC was scheduling compares before the adds causing wrong branches to be taken
in programs, resulting in misoptimized code wherever atomic adds where used.
llvm-svn: 84485
2009-10-19 11:00:58 +00:00
b0c23e81e7
various cleanups.
...
llvm-svn: 84471
2009-10-19 07:10:59 +00:00
aa1332ed72
simplify.
...
llvm-svn: 84465
2009-10-19 05:51:03 +00:00
8d58f2f8ea
clean up after metadata changes.
...
llvm-svn: 84463
2009-10-19 05:31:10 +00:00
18df82a20c
Add support for matching shuffle patterns with palignr.
...
llvm-svn: 84459
2009-10-19 02:17:23 +00:00
f0236e011e
Spill slots cannot alias.
...
llvm-svn: 84432
2009-10-18 19:58:47 +00:00
c436631a9c
Turn on post-alloc scheduling for x86.
...
llvm-svn: 84431
2009-10-18 19:57:27 +00:00
936d87b39d
Oops. I forgot to change the tests first. Disable post-alloc scheduling.
...
llvm-svn: 84425
2009-10-18 18:31:31 +00:00
0e9d9ca855
-Revert parts of 84326 and 84411. Distinquishing between fixed and non-fixed
...
stack slots and giving them different PseudoSourceValue's did not fix the
problem of post-alloc scheduling miscompiling llvm itself.
- Apply Dan's conservative workaround by assuming any non fixed stack slots can
alias other memory locations. This means a load from spill slot #1 cannot
move above a store of spill slot #2 .
- Enable post-alloc scheduling for x86 at optimization leverl Default and above.
llvm-svn: 84424
2009-10-18 18:16:27 +00:00
0b8db2dab7
Only fixed stack objects and spill slots should be get FixedStack PseudoSourceValue.
...
llvm-svn: 84411
2009-10-18 06:27:36 +00:00
c05471ea5b
remove some nonascii weird stuff
...
llvm-svn: 84410
2009-10-18 05:27:44 +00:00
39afe7e657
add some fixme's
...
llvm-svn: 84408
2009-10-18 05:09:15 +00:00
1fa98f0d74
remove the IndMemRemPass, which only made sense for when malloc/free were intrinsic
...
instructions.
llvm-svn: 84404
2009-10-18 05:02:09 +00:00
f4591d9a72
fix the other issue with ID's, hopefully really fixing the linux build.
...
llvm-svn: 84403
2009-10-18 04:58:34 +00:00
7883fc2ab1
fix some problems with ID definitions, which will hopefully fix the build bots.
...
llvm-svn: 84399
2009-10-18 04:27:14 +00:00
9e6882c8de
add function passes for printing various dominator datastructures
...
accessible through opt. Patch by Tobias Grosser!
llvm-svn: 84397
2009-10-18 04:10:40 +00:00
da42426e04
make DOTGraphTraits public, patch by Tobias Grosser!
...
llvm-svn: 84396
2009-10-18 04:09:11 +00:00
bb1410ea8b
Fix test/Bindings/Ocaml/vmcore.ml. When IRBuilder::CreateMalloc was removed,
...
LLVMBuildMalloc was reimplemented but with the bug that it didn't insert the
resulting instruction.
llvm-svn: 84374
2009-10-17 23:52:26 +00:00
b2647b9e67
inline isGEP away.
...
llvm-svn: 84373
2009-10-17 23:48:54 +00:00
54279632f3
Fix my -Asserts warning fix.
...
llvm-svn: 84372
2009-10-17 23:15:04 +00:00
f67d297eda
Teach vm core to more aggressively fold 'trunc' constantexprs,
...
allowing it to simplify the crazy constantexprs in the testcases
down to something sensible. This allows -std-compile-opts to
completely "devirtualize" the pointers to member functions in
the testcase from PR5176.
llvm-svn: 84368
2009-10-17 21:53:27 +00:00
04f049f2f5
Move UnescapeString to a static function for its sole client; its inefficient and broken.
...
llvm-svn: 84358
2009-10-17 20:43:42 +00:00
59449581b2
Remove llvm::EscapeString, raw_ostream::write_escaped is much faster.
...
llvm-svn: 84357
2009-10-17 20:43:29 +00:00
8eff29d805
Use raw_ostream::write_escaped instead of EscapeString.
...
llvm-svn: 84356
2009-10-17 20:43:19 +00:00
4108c43fd1
Add raw_ostream::write_escaped, for writing escaped strings.
...
llvm-svn: 84355
2009-10-17 20:43:08 +00:00
867f67301b
First draft of the OptionPreprocessor.
...
More to follow...
llvm-svn: 84352
2009-10-17 20:09:29 +00:00
e63c2cb3b7
Emit newlines at the end of instructions too.
...
llvm-svn: 84348
2009-10-17 19:43:45 +00:00
77668004c4
Move StringMap's string has function into StringExtras.h
...
llvm-svn: 84344
2009-10-17 18:21:06 +00:00
c099f1a0c5
Remove unnecessary include.
...
llvm-svn: 84336
2009-10-17 18:11:57 +00:00
47dfc9ae1b
Suppress -Asserts warning.
...
llvm-svn: 84327
2009-10-17 09:33:00 +00:00
4729191bb2
Distinquish stack slots from other stack objects. They (and fixed objects) get FixedStack PseudoSourceValues.
...
llvm-svn: 84326
2009-10-17 09:20:14 +00:00
8759585aba
Revert 84315 for now. Re-thinking the patch.
...
llvm-svn: 84321
2009-10-17 07:53:04 +00:00
0818d87ed1
Rename getFixedStack to getStackObject. The stack objects represented are not
...
necessarily fixed. Only those will negative frame indices are "fixed."
llvm-svn: 84315
2009-10-17 06:22:26 +00:00
a6e4db8ff7
80 col violation.
...
llvm-svn: 84311
2009-10-17 06:05:11 +00:00
88b36f1140
Simplify some code (first hunk) and fix PR5208 (second hunk) by
...
updating the callgraph when introducing a call.
llvm-svn: 84310
2009-10-17 05:39:39 +00:00
650997fb0b
Delete an obsolete comment.
...
llvm-svn: 84300
2009-10-17 01:37:38 +00:00
a3aaf85e23
Remove MallocInst from LLVM Instructions.
...
llvm-svn: 84299
2009-10-17 01:18:07 +00:00
0d3d9ee03e
Enhance CodePlacementOpt's unconditional intra-loop branch elimination logic
...
to be more general and understand more varieties of loops.
Teach CodePlacementOpt to reorganize the basic blocks of a loop so that
they are contiguous. This also includes a fair amount of logic for preserving
fall-through edges while doing so. This fixes a BranchFolding-ism where blocks
which can't be made to use a fall-through edge and don't conveniently fit
anywhere nearby get tossed out to the end of the function.
llvm-svn: 84295
2009-10-17 00:32:43 +00:00
c7d6a8327c
Autoupgrade malloc insts to malloc calls.
...
Update testcases that rely on malloc insts being present.
Also prematurely remove MallocInst handling from IndMemRemoval and RaiseAllocations to help pass tests in this incremental step.
llvm-svn: 84292
2009-10-17 00:00:19 +00:00
264da3274e
HeapAllocSRoA also needs to check if malloc array size can be computed.
...
llvm-svn: 84288
2009-10-16 23:12:25 +00:00
b1baaf5ab9
Allow widening of extract subvector
...
llvm-svn: 84279
2009-10-16 22:05:48 +00:00
12524fd2e5
Do not emit name entry for a pointer type.
...
llvm-svn: 84276
2009-10-16 21:27:43 +00:00
007ceb4603
Change createPostRAScheduler so it can be turned off at llc -O1.
...
llvm-svn: 84273
2009-10-16 21:06:15 +00:00
99429a00ff
Move zext and sext casts fed by loads into the same block as the
...
load, to help SelectionDAG fold them into the loads, unless
conditions are unfavorable.
llvm-svn: 84271
2009-10-16 20:59:35 +00:00
8f842d31fd
Parse PHI instruction with attached metadata.
...
llvm-svn: 84264
2009-10-16 18:45:49 +00:00
d0099a94db
If there is not any llvm instruction associated with each lexical scope encoded in debug info then create such scope on demand for variable info.
...
llvm-svn: 84262
2009-10-16 18:18:03 +00:00
6f834660c9
Invert isSafeToGetMallocArraySize check because we return NULL when we don't know the size.
...
Thanks to Duncan Sands for noticing this bug.
llvm-svn: 84261
2009-10-16 18:08:17 +00:00
0058c7bcb0
Strip trailing white space.
...
llvm-svn: 84256
2009-10-16 15:20:13 +00:00
ca019e704c
Update CMake file.
...
llvm-svn: 84252
2009-10-16 10:29:08 +00:00
89259d7929
Cleaned up some code. No functionality change.
...
llvm-svn: 84251
2009-10-16 08:58:34 +00:00
343d17d762
I am no spelling bee.
...
llvm-svn: 84250
2009-10-16 06:18:09 +00:00
03da4dba59
Enable post-alloc scheduling for all ARM variants except for Thumb1.
...
llvm-svn: 84249
2009-10-16 06:11:08 +00:00
8b614768f8
If post-alloc scheduler is not enabled, it should return false, not true.
...
llvm-svn: 84248
2009-10-16 06:10:34 +00:00
47062ce503
Indent code.
...
llvm-svn: 84247
2009-10-16 05:42:28 +00:00
8cd7c92613
Add comment.
...
llvm-svn: 84246
2009-10-16 05:33:58 +00:00
9c096840a0
80 column violation.
...
llvm-svn: 84244
2009-10-16 05:18:39 +00:00
01404ecec7
Fix more NEON instruction encodings.
...
Patch by Johnny Chen.
llvm-svn: 84243
2009-10-16 03:58:44 +00:00
4794b2b27b
Add half precision floating point support (float16) to APFloat,
...
patch by Peter Johnson! (PR5195)
llvm-svn: 84239
2009-10-16 02:13:51 +00:00
27f2049305
add haiku support, patch by Paul Davey!
...
llvm-svn: 84238
2009-10-16 02:06:30 +00:00
5d9134747b
MC: Set symbol values in MachO MCStreamer.
...
llvm-svn: 84236
2009-10-16 01:58:23 +00:00
e024def785
Minor formatting tweaks.
...
llvm-svn: 84235
2009-10-16 01:58:15 +00:00
8e6d889d4c
MC: Switch assembler API to using MCExpr instead of MCValue.
...
llvm-svn: 84234
2009-10-16 01:58:03 +00:00
c3bd60e393
MC: Remove unneeded context argument to MCExpr::Evaluate*.
...
llvm-svn: 84233
2009-10-16 01:57:52 +00:00
5f339244fb
MC: Tweak variable assignment diagnostics, and make reassignment of non-absolute
...
variables and symbols invalid.
llvm-svn: 84232
2009-10-16 01:57:39 +00:00
d20cda028a
MC: When parsing a variable reference, substitute absolute variables immediately
...
since they are allowed to be redefined.
llvm-svn: 84230
2009-10-16 01:34:54 +00:00
17b9027b5c
MC: Move assembler variable values from MCContext to MCSymbol.
...
llvm-svn: 84229
2009-10-16 01:33:57 +00:00
584a1b25ef
MC: Switch MCContext value table to storing MCExprs.
...
llvm-svn: 84228
2009-10-16 01:33:11 +00:00
8ec25937f1
When checking aliases between phi sources and V2, we know the sources are not themselves phi nodes. However, V2 may be. Call aliasCheck with V2 first to potentially eliminate a std::swap call.
...
llvm-svn: 84226
2009-10-16 00:33:09 +00:00
419160bd79
Revert svn r80498 and replace it with a different solution. The only problem
...
I can see with the original code was that I forgot that this runs after
type legalization and hence the result type will always be i32. (Custom
legalization of EXTRACT_VECTOR_ELT is only enabled for vector types with
8- and 16-bit elements.)
Regarding the FIXME comment: any information about sign and zero-extension
should be captured by separate extension operations. The DAG combiner should
handle those to produce either VGETLANEu or VGETLANEs, and that seems to be
working now. If there are cases that we're missing, let me know.
llvm-svn: 84218
2009-10-15 23:12:05 +00:00
49e417c52c
Dllexport stuff cleanup:
...
1. Emit external function type information for all COFF targets since it's
a feature of object format
2. Emit linker directives only for cygming (since this is ld-specific stuff)
llvm-svn: 84214
2009-10-15 22:36:18 +00:00
3f23601b00
Branches must be the last instruction in a Thumb2 IT block. Approved by Evan Cheng.
...
llvm-svn: 84212
2009-10-15 22:25:32 +00:00
4138b11c93
Fix encoding bits for N3VLInt3_QHS multiclass with 8-bit elements.
...
Patch by Johnny Chen.
llvm-svn: 84206
2009-10-15 21:57:47 +00:00
644de27659
Fix ARM memory operand parsing of post indexing with just a base register, that
...
is just "[Rn]" and no tailing comma with an offset, etc.
llvm-svn: 84205
2009-10-15 21:42:45 +00:00
2f9be50774
Fix a potential performance problem in placing ARM constant pools.
...
In the case where there are no good places to put constants and we fall back
upon inserting unconditional branches to make new blocks, allow all constant
pool references in range of those blocks to put constants there, even if that
means resetting the "high water marks" for those references. This will still
terminate because you can't keep splitting blocks forever, and in the bad
cases where we have to split blocks, it is important to avoid splitting more
than necessary.
llvm-svn: 84202
2009-10-15 20:49:47 +00:00
146dcf2ab5
More bits of the ARM target assembler for llvm-mc, code added to parse labels
...
as expressions, code for parsing a few arm specific directives (still needs
the MCStreamer calls for these). Some clean up of the operand parsing code
and adding some comments.
llvm-svn: 84201
2009-10-15 20:48:48 +00:00
e4a2117161
Remove X86Subtarget::IsLinux. It's no longer being used.
...
llvm-svn: 84200
2009-10-15 20:23:21 +00:00
13020b1faf
Fix bug where array malloc with unexpected computation of the size argument resulted in MallocHelper
...
identifying the malloc as a non-array malloc. This broke GlobalOpt's optimization of stores of mallocs
to global variables.
The fix is to classify malloc's into 3 categories:
1. non-array mallocs
2. array mallocs whose array size can be determined
3. mallocs that cannot be determined to be of type 1 or 2 and cannot be optimized
getMallocArraySize() returns NULL for category 3, and all users of this function must avoid their
malloc optimization if this function returns NULL.
Eventually, currently unexpected codegen for computing the malloc's size argument will be supported in
isArrayMalloc() and getMallocArraySize(), extending malloc optimizations to those examples.
llvm-svn: 84199
2009-10-15 20:14:52 +00:00
74c35c126a
Add files Sanjiv forgot.
...
llvm-svn: 84196
2009-10-15 19:46:34 +00:00
a07cae65e3
Re-apply 84180 with the fixed test case.
...
llvm-svn: 84195
2009-10-15 19:26:25 +00:00
460ceae432
Move Blackfin intrinsics into the Target/Blackfin directory.
...
llvm-svn: 84194
2009-10-15 18:50:52 +00:00
e4197250cc
Report errors correctly for unselected target intrinsics.
...
llvm-svn: 84193
2009-10-15 18:50:03 +00:00
923b5aa973
Clean up TargetIntrinsicInfo API. Add pure virtual methods.
...
llvm-svn: 84192
2009-10-15 18:49:26 +00:00
d708a4908c
Add missing break statements! Thanks to Duncan Sands for pointing this out!
...
llvm-svn: 84191
2009-10-15 18:48:58 +00:00
3b5bdd536b
Revert "Complete Rewrite of AsmPrinter, TargetObjectFile based on new
...
PIC16Section class", it breaks globals.ll.
llvm-svn: 84184
2009-10-15 15:02:14 +00:00
917db86b42
Complete Rewrite of AsmPrinter, TargetObjectFile based on new PIC16Section class
...
derived from MCSection.
llvm-svn: 84180
2009-10-15 10:10:43 +00:00
7357636091
Few changes to comply with new DebugInfo Metadata representation.
...
llvm-svn: 84179
2009-10-15 09:48:25 +00:00
9cec2380f2
Teach basicaa about memcpy/memmove/memset. The length argument can be used to
...
improve alias results if constant, and the source pointer can't be modified.
llvm-svn: 84175
2009-10-15 07:11:24 +00:00
d7216cf74c
Teach BasicAA to use the size parameter of the memory use marker intrinsics.
...
llvm-svn: 84174
2009-10-15 06:12:11 +00:00
68ead6c7a8
Be smarter about reusing constant pool entries.
...
llvm-svn: 84173
2009-10-15 05:52:29 +00:00
b4f2a85fe4
Fix another problem with ARM constant pools. Radar 7303551.
...
When ARMConstantIslandPass cannot find any good locations (i.e., "water") to
place constants, it falls back to inserting unconditional branches to make a
place to put them. My recent change exposed a problem in this area. We may
sometimes append to the same block more than one unconditional branch. The
symptoms of this are that the generated assembly has a branch to an undefined
label and running llc with -debug will cause a seg fault.
This happens more easily since my change to prevent CPEs from moving from
lower to higher addresses as the algorithm iterates, but it could have
happened before. The end of the block may be in range for various constant
pool references, but the insertion point for new CPEs is not right at the end
of the block -- it is at the end of the CPEs that have already been placed
at the end of the block. The insertion point could be out of range. When
that happens, the fallback code will always append another unconditional
branch if the end of the block is in range.
The fix is to only append an unconditional branch if the block does not
already end with one. I also removed a check to see if the constant pool load
instruction is at the end of the block, since that is redundant with
checking if the end of the block is in-range.
There is more to be done here, but I think this fixes the immediate problem.
llvm-svn: 84172
2009-10-15 05:10:36 +00:00
c855b45b78
only try to fold constantexpr operands when the worklist is first populated,
...
don't bother every time going around the main worklist. This speeds up a
release-asserts opt -std-compile-opts on 403.gcc by about 4% (1.5s). It
seems to speed up the most expensive instances of instcombine by ~10%.
llvm-svn: 84171
2009-10-15 04:59:28 +00:00
dd1f68a10c
don't bother calling ConstantFoldInstruction unless there is a use of the
...
instruction (which disqualifies stores, unreachable, etc) and at least the
first operand is a constant. This filters out a lot of obvious cases that
can't be folded. Also, switch the IRBuilder to a TargetFolder, which tries
harder.
llvm-svn: 84170
2009-10-15 04:13:44 +00:00
fcb27ad54c
Take advantage of TargetData when available; we know that the atomic intrinsics
...
only dereference the element they point to directly with no pointer arithmetic.
llvm-svn: 84159
2009-10-15 00:36:35 +00:00
c9af381df8
Make CodePlacementOpt align loops, rather than loop headers. The
...
header is just the entry block to the loop, and it needn't be at
the top of the loop in the code layout.
Remove the code that suppressed loop alignment for outer loops,
so that outer loops are aligned.
llvm-svn: 84158
2009-10-15 00:36:22 +00:00
2f61e0946a
When LiveVariables is adding implicit-def to model "partial dead", add the earlyclobber marker if the superreg def has it.
...
llvm-svn: 84153
2009-10-14 23:39:27 +00:00
70b1fa5a24
Print earlyclobber for implicit-defs as well.
...
llvm-svn: 84152
2009-10-14 23:37:31 +00:00
cfcf6bc70d
Fix instruction encoding bits for NEON VPADAL.
...
Patch by Johnny Chen.
llvm-svn: 84146
2009-10-14 21:43:17 +00:00
ad03cf02f6
Remove unused variables to fix build warning.
...
llvm-svn: 84144
2009-10-14 21:40:45 +00:00
b1d6fde13e
Make loop not recalc getNumOperands() each time around
...
llvm-svn: 84138
2009-10-14 21:22:39 +00:00
6875c5ebe4
Add support to record DbgScope as inlined scope.
...
llvm-svn: 84134
2009-10-14 21:08:09 +00:00
02f5588f62
quiet compiler warning
...
llvm-svn: 84133
2009-10-14 21:07:11 +00:00
94068707e1
Inst{11-8} for vshl should be 0b0101, not 0b1111.
...
Refs: A7-17 & A8-750.
Patch by Johnny Chen.
llvm-svn: 84131
2009-10-14 20:31:01 +00:00
1a791eedbf
Set instruction encoding bits 4 and 7 for ARM register-register and
...
register-shifted-register instructions. Patch by Johnny Chen.
llvm-svn: 84124
2009-10-14 19:00:24 +00:00
c350cdf3b3
Refactor code to select NEON VST intrinsics.
...
llvm-svn: 84122
2009-10-14 18:32:29 +00:00
92f8619923
Use isVoidTy()
...
llvm-svn: 84118
2009-10-14 17:29:00 +00:00
12b4799787
Refactor code to select NEON VLD intrinsics.
...
llvm-svn: 84117
2009-10-14 17:28:52 +00:00
ebaa76ed0d
Add copyMD to copy metadata from one instruction to another instruction.
...
llvm-svn: 84113
2009-10-14 17:02:49 +00:00
93117bc499
More refactoring. NEON vst lane intrinsics can share almost all the code for
...
vld lane intrinsics.
llvm-svn: 84110
2009-10-14 16:46:45 +00:00
4145e3ac8d
Refactor code for selecting NEON load lane intrinsics.
...
llvm-svn: 84109
2009-10-14 16:19:03 +00:00
8e6ccb65df
I don't see any point in having both eh.selector.i32 and eh.selector.i64,
...
so get rid of eh.selector.i64 and rename eh.selector.i32 to eh.selector.
Likewise for eh.typeid.for. This aligns us with gcc, which always uses a
32 bit value for the selector on all platforms. My understanding is that
the register allocator used to assert if the selector intrinsic size didn't
match the pointer size, and this was the reason for introducing the two
variants. However my testing shows that this is no longer the case (I
fixed some bugs in selector lowering yesterday, and some more today in the
fastisel path; these might have caused the original problems).
llvm-svn: 84106
2009-10-14 16:11:37 +00:00
6b9044db01
make instcombine's instruction sinking more aggressive in the
...
presence of PHI nodes.
llvm-svn: 84103
2009-10-14 15:21:58 +00:00
b3ccb6498c
Clear VisitedPHIs after use.
...
llvm-svn: 84080
2009-10-14 06:46:26 +00:00
c1eed9d120
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
f92f5557ee
More code clean up based on patch feedback.
...
llvm-svn: 84074
2009-10-14 05:22:03 +00:00
31565b3e9f
Change VisitedPHIs into an instance variable that's freed by each alias() call.
...
llvm-svn: 84072
2009-10-14 05:05:02 +00:00
0be8c2b0e3
Make isSafeToClobberEFLAGS more aggressive. Teach it to scan backwards
...
(for uses marked kill and defs marked dead) a few instructions in
addition to forwards. Also, increase the maximum number of instructions
to scan, as it appears to help in a fair number of cases.
llvm-svn: 84061
2009-10-14 00:08:59 +00:00
df777edea9
This remat entry is basically done. There are hooks to allow targets
...
to remat non-load instructions as loads, and the remat code now uses
the UnmodeledSideEffects flags, MachineMemOperands, and similar things
to decide which instructions are valid for rematerialization.
llvm-svn: 84060
2009-10-14 00:02:01 +00:00
0aa63c97ce
Add a few README.txt items.
...
llvm-svn: 84059
2009-10-13 23:58:05 +00:00
3a80daced0
Correct comment about ARM immediates using '#' not '$' and TODO for modifiers.
...
llvm-svn: 84055
2009-10-13 23:33:38 +00:00
d7ebfe3963
s/DebugLoc.CompileUnit/DebugLoc.Scope/g
...
s/DebugLoc.InlinedLoc/DebugLoc.InlinedAtLoc/g
llvm-svn: 84054
2009-10-13 23:28:53 +00:00
a677136900
Check void type before using RAUWd.
...
llvm-svn: 84049
2009-10-13 22:56:32 +00:00
b62d160b3c
More Neon clean-up: avoid the need for custom-lowering vld/st-lane intrinsics
...
by creating TargetConstants during instruction selection instead of during
legalization.
llvm-svn: 84042
2009-10-13 22:29:24 +00:00
f50799412c
More bits of the ARM target assembler for llvm-mc to parse immediates.
...
Also fixed a couple of coding style things that crept in. And added more
to the temporary hacked up ARMAsmParser::MatchInstruction() method for testing.
llvm-svn: 84040
2009-10-13 22:19:02 +00:00
c10e88db22
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
1fdbe1152d
NEON VLD/VST are now fully implemented. For operations that expand to
...
multiple instructions, the expansion is done during selection so there is
no need to do anything special during legalization.
llvm-svn: 84036
2009-10-13 21:55:24 +00:00
115741ba79
Do not check use_empty() before replaceAllUsesWith(). This gives ValueHandles a chance to get properly updated.
...
llvm-svn: 84033
2009-10-13 21:41:20 +00:00
d162dbac7f
Keep track of stubs that are created. This fixes PR5162 and probably PR4822 and
...
4406. Patch by Nick Lewycky!
llvm-svn: 84032
2009-10-13 21:32:57 +00:00
18a956cb4a
Introduce new convenience methods for sign extending or
...
truncating an SDValue (depending on whether the target
type is bigger or smaller than the value's type); or zero
extending or truncating it. Use it in a few places (this
seems to be a popular operation, but I only modified cases
of it in SelectionDAGBuild). In particular, the eh_selector
lowering was doing this wrong due to a repeated rather than
inverted test, fixed with this change.
llvm-svn: 84027
2009-10-13 21:04:12 +00:00
3b51560ae4
Revise ARM inline assembly memory operands to require the memory address to
...
be in a register. The previous use of ARM address mode 2 was completely
arbitrary and inappropriate for Thumb. Radar 7137468.
llvm-svn: 84022
2009-10-13 20:50:28 +00:00
fd04c74bc0
Add an "msasm" flag to inline asm as suggested in PR 5125.
...
A little ugliness is accepted to keep the binary file format
compatible. No functional change yet.
llvm-svn: 84020
2009-10-13 20:46:56 +00:00
7460e0822f
Fix method name in comment, per Bob Wilson.
...
llvm-svn: 84017
2009-10-13 20:25:58 +00:00
2dc6f8de03
Use the new CodeMetrics class to compute code size instead of
...
manually counting instructions.
llvm-svn: 84016
2009-10-13 20:12:23 +00:00
abb728d3f4
Compute a full cost value even when a setjmp call is found.
...
llvm-svn: 84015
2009-10-13 20:10:10 +00:00
2ccea5d13f
Split code not specific to Function inlining out into a separate class,
...
named CodeMetrics. Move it to be a non-nested class. Rename RegionInfo
back to FunctionInfo.
llvm-svn: 84013
2009-10-13 19:58:07 +00:00
4c98efb917
Add debugging output.
...
llvm-svn: 84011
2009-10-13 19:16:03 +00:00
423e42b371
Add ARMv6T2 SBFX/UBFX instructions. Approved by Anton Korobeynikov.
...
llvm-svn: 84009
2009-10-13 18:59:48 +00:00
f34311779c
Update CMake file (lexically order files).
...
llvm-svn: 84008
2009-10-13 18:57:27 +00:00
78cb102490
Do not write empty METADATA_ATTACHMENT record.
...
llvm-svn: 84006
2009-10-13 18:51:28 +00:00
0b26c9e0b3
Remove unnecessary assert.
...
llvm-svn: 84004
2009-10-13 18:51:03 +00:00
a8d176d62f
Update the other CMake file.
...
llvm-svn: 84003
2009-10-13 18:50:54 +00:00
52b1745a96
Parse GEP with attached custom metadata. This happens during libprofile_rt.bca build.
...
llvm-svn: 84002
2009-10-13 18:49:55 +00:00
113d959f1b
Update CMake file.
...
llvm-svn: 84001
2009-10-13 18:48:07 +00:00
f1f3dd3c3c
Refactor some code. No functionality changes.
...
llvm-svn: 84000
2009-10-13 18:42:04 +00:00
54463e837a
Commit the removal of this file, which is now moved to lib/Analysis.
...
llvm-svn: 83999
2009-10-13 18:37:20 +00:00
4552e3cd73
Move the InlineCost code from Transforms/Utils to Analysis.
...
llvm-svn: 83998
2009-10-13 18:30:07 +00:00
5b3e05bcaa
Start refactoring the inline cost estimation code so that it can be used
...
for purposes other than inlining.
llvm-svn: 83997
2009-10-13 18:24:11 +00:00
19788ca686
change simplifycfg to not duplicate 'unwind' instructions. Hopefully
...
this will increase the likelihood of common code getting sunk towards
the unwind.
llvm-svn: 83996
2009-10-13 18:13:05 +00:00
71ca652475
Make LoopUnswitch's cost estimation count Instructions, rather than
...
BasicBlocks, so that it doesn't blindly procede in the presence of
large individual BasicBlocks. This addresses a class of code-size
expansion problems.
llvm-svn: 83992
2009-10-13 17:50:43 +00:00
f98e981cd0
Make the ExecutionEngine automatically remove global mappings on when their
...
GlobalValue is destroyed. Function destruction still leaks machine code and
can crash on leaked stubs, but this is some progress.
llvm-svn: 83987
2009-10-13 17:42:08 +00:00
c5cc659cfa
"there is not any instruction with attached debug info in this module" does not mean "there is no debug info in this module". :)
...
llvm-svn: 83984
2009-10-13 17:35:35 +00:00
453a06e3ac
Add some ARM instruction encoding bits.
...
Patch by Johnny Chen.
llvm-svn: 83983
2009-10-13 17:35:30 +00:00
d26a26ae7e
Fix regression introduced by r83894.
...
llvm-svn: 83982
2009-10-13 17:29:13 +00:00
e6f26a7415
Copy metadata when value is RAUW'd. It is debatable whether this is the right approach for custom metadata data in general. However, right now the only custom data user, "dbg", expects this behavior while FE is constructing llvm IR with debug info.
...
llvm-svn: 83977
2009-10-13 17:00:54 +00:00
0bc673de0d
Fix a tab. Thanks to Johnny Chen for pointing it out.
...
llvm-svn: 83973
2009-10-13 15:27:23 +00:00
9bc890425a
Force memory use markers to have a ConstantInt for the size argument.
...
llvm-svn: 83960
2009-10-13 07:57:33 +00:00
e2782c7614
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
cdf01b5d82
Fix a -Asserts warning.
...
llvm-svn: 83950
2009-10-13 06:47:08 +00:00
70e8505eb1
Memory dependence analysis was incorrectly stopping to scan for stores to a pointer at bitcast uses of a malloc call.
...
It should continue scanning until the malloc call, and this patch fixes that.
llvm-svn: 83931
2009-10-13 01:42:53 +00:00
0f58bec599
Find enclosing subprogram info.
...
llvm-svn: 83922
2009-10-12 23:11:24 +00:00
0af2a420cd
Set default location for a function if it is not set.
...
llvm-svn: 83921
2009-10-12 23:10:55 +00:00
11b32384f2
Fix two warnings about unused variables that are only used in assert() calls.
...
llvm-svn: 83917
2009-10-12 22:51:49 +00:00
5b07a903d4
Delete a comment that makes no sense to me. The statement that moving a CPE
...
before its reference is only supported on ARM has not been true for a while.
In fact, until recently, that was only supported for Thumb. Besides that,
CPEs are always a multiple of 4 bytes in size, so inserting a CPE should have
no effect on Thumb alignment.
llvm-svn: 83916
2009-10-12 22:49:05 +00:00
3f0b9b3db4
Fix a problem in the code where ARMAsmParser::ParseShift() second argument
...
should have been a pointer to a reference.
llvm-svn: 83915
2009-10-12 22:39:54 +00:00
f815861591
Make licm debug message readable.
...
llvm-svn: 83908
2009-10-12 22:25:23 +00:00
3250e7769f
Change CreateNewWater method to return NewMBB by reference.
...
llvm-svn: 83905
2009-10-12 21:39:43 +00:00
cc121aa750
Last week, ARMConstantIslandPass was failing to converge for the
...
MultiSource/Benchmarks/MiBench/automotive-susan test. The failure has
since been masked by an unrelated change (just randomly), so I don't have
a testcase for this now. Radar 7291928.
The situation where this happened is that a constant pool entry (CPE) was
placed at a lower address than the load that referenced it. There were in
fact 2 CPEs placed at adjacent addresses and referenced by 2 loads that were
close together in the code. The distance from the loads to the CPEs was
right at the limit of what they could handle, so that only one of the CPEs
could be placed within range. On every iteration, the first CPE was found
to be out of range, causing a new CPE to be inserted. The second CPE had
been in range but the newly inserted entry pushed it too far away. Thus the
second CPE was also replaced by a new entry, which in turn pushed the first
CPE out of range. Etc.
Judging from some comments in the code, the initial implementation of this
pass did not support CPEs placed _before_ their references. In the case
where the CPE is placed at a higher address, the key to making the algorithm
terminate is that new CPEs are only inserted at the end of a group of adjacent
CPEs. This is implemented by removing a basic block from the "WaterList"
once it has been used, and then adding the newly inserted CPE block to the
list so that the next insertion will come after it. This avoids the ping-pong
effect where CPEs are repeatedly moved to the beginning of a group of
adjacent CPEs. This does not work when going backwards, however, because the
entries at the end of an adjacent group of CPEs are closer than the CPEs
earlier in the group.
To make this pass terminate, we need to maintain a property that changes can
only happen in some sort of monotonic fashion. The fix used here is to require
that the CPE for a particular constant pool load can only move to lower
addresses. This is a very simple change to the code and should not cause
any significant degradation in the results.
llvm-svn: 83902
2009-10-12 21:23:15 +00:00
e4adae267e
Another minor clean-up.
...
llvm-svn: 83897
2009-10-12 20:45:53 +00:00
196bf32ab0
Remove redundant parameter.
...
llvm-svn: 83894
2009-10-12 20:37:23 +00:00
3a7326e705
Use early exit to reduce indentation.
...
llvm-svn: 83874
2009-10-12 19:04:03 +00:00
3af34312d4
Change to return a value by reference.
...
llvm-svn: 83873
2009-10-12 19:01:12 +00:00
c7a3cf4066
Add a typedef for an iterator.
...
llvm-svn: 83872
2009-10-12 18:52:13 +00:00
06243d7bf2
Revert the kludge in 76703. I got a clean
...
bootstrap of FSF-style PPC, so there is some
reason to believe the original bug (which was
never analyzed) has been fixed, probably by
82266.
llvm-svn: 83871
2009-10-12 18:49:00 +00:00
4c9f0e8f53
Fix warning.
...
llvm-svn: 83870
2009-10-12 18:45:32 +00:00
406ac811ab
Fix http://llvm.org/PR5160 , to let CallbackVHs modify other ValueHandles on the
...
same Value without breaking things.
llvm-svn: 83861
2009-10-12 17:43:32 +00:00
32eed95a78
Remove a redundant member variable.
...
llvm-svn: 83857
2009-10-12 16:44:10 +00:00
a698d7ac3c
Don't forget to mark RAX as live-out of the function when arranging for
...
it to hold the address of an sret return value, for x86-64 ABI purposes.
Also, fix the test that was originally intended to test this to actually
test it, using FileCheck.
llvm-svn: 83853
2009-10-12 16:36:12 +00:00
a3ed9edd40
More heuristics for Combiner-AA. Still catches all important cases, but
...
compile time penalty on gnugo, the worst case in MultiSource, is down to
about 2.5% from 30%
llvm-svn: 83824
2009-10-12 05:53:58 +00:00
8227b0512f
Haiku porting patches, Credit to Paul Davey.
...
llvm-svn: 83823
2009-10-12 04:57:20 +00:00
0840c823e4
Fix PR5087, patch by Jakub Staszak!
...
llvm-svn: 83822
2009-10-12 04:22:44 +00:00
40cf28d6eb
add some more hooks to the C bindings, patch by Kenneth Uildriks!
...
llvm-svn: 83821
2009-10-12 04:01:02 +00:00
8abd572dae
populate instcombine's initial worklist more carefully, causing
...
it to visit instructions from the start of the function to the
end of the function in the first path. This greatly speeds up
some pathological cases (e.g. PR5150).
Try #3 , this time with some unneeded debug info stuff removed
which was causing dead pointers to be added to the worklist.
llvm-svn: 83818
2009-10-12 03:58:40 +00:00
8ce6b36c86
revert r83814 for now, it is making the llvm-gcc bootstrap unhappy.
...
llvm-svn: 83817
2009-10-11 23:56:08 +00:00
78d6310429
populate instcombine's initial worklist more carefully, causing
...
it to visit instructions from the start of the function to the
end of the function in the first path. This greatly speeds up
some pathological cases (e.g. PR5150).
llvm-svn: 83814
2009-10-11 23:17:43 +00:00
4b38ce9f25
Add missed mem-mem move patterns
...
llvm-svn: 83812
2009-10-11 23:03:53 +00:00
415c3dc501
Add MSP430 mem-mem insts support. Patch by Brian Lucas with some my refinements
...
llvm-svn: 83811
2009-10-11 23:03:28 +00:00
2c2deae5ac
remove some harmful code that would turn an insertelement on an undef
...
into a shuffle even if it was used by another insertelement. If the
visitation order of instcombine was wrong, this would turn a chain of
insertelements into a chain of shufflevectors, which was quite painful.
Since CollectShuffleElements handles these cases, the code can just
be nuked.
llvm-svn: 83810
2009-10-11 23:02:46 +00:00
c6cdbfbfdd
teach instcombine to simplify xor's harder, catching the
...
new testcase.
llvm-svn: 83799
2009-10-11 22:22:13 +00:00
6e6ac47125
cleanups
...
llvm-svn: 83797
2009-10-11 22:00:32 +00:00
1639234775
cleanup, no functionality change.
...
llvm-svn: 83795
2009-10-11 21:36:10 +00:00
fd27f8a5b3
generalize a transformation even more: we don't care whether the
...
input the the mul is a zext from bool, just that it is all zeros
other than the low bit. This fixes some phase ordering issues
that would cause us to miss some xforms in mul.ll when the worklist
is visited differently.
llvm-svn: 83794
2009-10-11 21:29:45 +00:00
406cb75c6b
simplify a transformation by making it more general.
...
llvm-svn: 83792
2009-10-11 21:22:21 +00:00
f39f4f928a
temporarily revert previous patch
...
llvm-svn: 83791
2009-10-11 21:05:34 +00:00
bb058d3a23
populate instcombine's initial worklist more carefully, causing
...
it to visit instructions from the start of the function to the
end of the function in the first path. This greatly speeds up
some pathological cases (e.g. PR5150).
llvm-svn: 83790
2009-10-11 21:04:37 +00:00
8b3081350e
Remove CleanupDbgInfo, instcombine does this and its not worth duplicating it
...
here.
llvm-svn: 83789
2009-10-11 19:58:35 +00:00
907ec36943
LICM shouldn't sink/delete debug information. Fix this and add a testcase.
...
For now the metadata of sinked/hoisted instructions is still wrong, but that'll
be fixed when instructions will have debug metadata directly attached.
llvm-svn: 83786
2009-10-11 19:15:54 +00:00
6bce6bbf40
Implement 'm' memory operand properly
...
llvm-svn: 83785
2009-10-11 19:14:21 +00:00
a58a3f930a
Implement proper asmprinting for the globals. This eliminates bogus "call" modifier and also adds support for offsets wrt globals.
...
llvm-svn: 83784
2009-10-11 19:14:02 +00:00
3525a4a268
Implement asm printing for inline asm memory operands
...
llvm-svn: 83783
2009-10-11 19:13:34 +00:00
85c85c5e04
when folding duplicate conditions, delete the
...
now-probably-dead instruction tree feeding it.
llvm-svn: 83778
2009-10-11 18:39:58 +00:00
e374382b8f
implement rdar://7293527, a trivial instcombine that llvm-gcc
...
gets but clang doesn't, because it is implemented in GCC's
fold routine.
llvm-svn: 83761
2009-10-11 07:53:15 +00:00
97b1405207
implement a transformation in jump threading that is currently
...
done by condprop, but do it in a much more general form. The
basic idea is that we can do a limited form of tail duplication
in the case when we have a branch on a phi. Moving the branch
up in to the predecessor block makes instruction selection
much easier and encourages chained jump threadings.
llvm-svn: 83759
2009-10-11 07:24:57 +00:00
6ce85e85f5
restructure some code, no functionality change.
...
llvm-svn: 83756
2009-10-11 04:40:21 +00:00
f466bc84c9
factor some code better and move a function, no functionality change.
...
llvm-svn: 83755
2009-10-11 04:33:43 +00:00
f99a74e24b
make jump threading on a phi with undef inputs happen.
...
llvm-svn: 83754
2009-10-11 04:18:15 +00:00
71d353dd48
rewrite LCSSA to use SSAUpdate, to only return true if it modifies
...
the IR, and to implement the FIXME'd optimization.
llvm-svn: 83748
2009-10-11 02:53:37 +00:00
101dde30ed
clean up and simplify some code. Don't use setvector when things will be
...
inserted only once, just use vector. Don't compute ExitBlocks unless we
need it, change std::sort to array_pod_sort.
llvm-svn: 83747
2009-10-11 01:07:15 +00:00
b6c65faa64
switch GVN to use SSAUpdater. Besides removing a lot of complexity
...
from GVN, this also speeds it up, inserts fewer PHI nodes (see the
testcase) and allows it to remove more loads (due to fewer PHI nodes
standing in the way).
llvm-svn: 83746
2009-10-10 23:50:30 +00:00
9c382cebc5
add a simple helper method.
...
llvm-svn: 83745
2009-10-10 23:41:48 +00:00
249265de06
add ability for clients of SSAUpdater to find out about the
...
PHI nodes inserted.
llvm-svn: 83744
2009-10-10 23:15:24 +00:00
89d2a5c4f3
remove dead code
...
llvm-svn: 83742
2009-10-10 23:04:12 +00:00
67cdd8b567
add the ability to get a rewritten value from the middle of a block,
...
not just at the end. Add a big comment explaining when this could
be useful (which never happens for jump threading).
llvm-svn: 83741
2009-10-10 23:00:11 +00:00
e474a8d3a7
rename GetValueInBlock -> GetValueAtEndOfBlock to better reflect
...
what it does.
llvm-svn: 83740
2009-10-10 22:41:58 +00:00
5b8826b4da
It seems that OR operation does not affect status reg at all.
...
Remove impdef of SRW. This fixes PR4779
llvm-svn: 83739
2009-10-10 22:17:47 +00:00
65e69a77e1
use a typedef instead of spelling out an insane type. Yay for auto someday.
...
llvm-svn: 83707
2009-10-10 09:09:20 +00:00
84095071ea
Change jump threading to use the new SSAUpdater class instead of
...
DemoteRegToStack. This makes it more efficient (because it isn't
creating a ton of load/stores that are eventually removed by a later
mem2reg), and more slightly more effective (because those load/stores
don't get in the way of threading).
llvm-svn: 83706
2009-10-10 09:05:58 +00:00
60d4e69c81
Implement an efficient and fully general SSA update mechanism that
...
works on unstructured CFGs. This implements PR217, our oldest open PR.
llvm-svn: 83705
2009-10-10 09:04:27 +00:00
f30a2b0c86
random tidying
...
llvm-svn: 83701
2009-10-10 06:22:45 +00:00
b8120770b4
Create a new InstrEmitter class for translating SelectionDAG nodes
...
into MachineInstrs. This is mostly just moving the code from
ScheduleDAGSDNodesEmit.cpp into a new class. This decouples MachineInstr
emitting from scheduling.
llvm-svn: 83699
2009-10-10 01:32:21 +00:00
a22f2d8614
Make getMachineNode return a MachineSDNode* instead of a generic SDNode*
...
since it won't do any folding. This will help avoid some inconvenient
casting.
llvm-svn: 83698
2009-10-10 01:29:16 +00:00
1faa11521e
Remove a no-longer-necessary #include.
...
llvm-svn: 83697
2009-10-10 00:36:09 +00:00
e919de5acf
Replace X86's CanRematLoadWithDispOperand by calling the target-independent
...
MachineInstr::isInvariantLoad instead, which has the benefit of being
more complete.
llvm-svn: 83696
2009-10-10 00:34:18 +00:00
26e9b89b7c
Fix a missing initialization of PostRAScheduler's AA member.
...
llvm-svn: 83695
2009-10-10 00:15:38 +00:00
918ec53c64
The ScheduleDAG framework now requires an AliasAnalysis argument, though
...
it isn't needed in the ScheduleDAGSDNodes schedulers.
llvm-svn: 83691
2009-10-09 23:33:48 +00:00
4a72f7ab53
Mark the LDR instruction with isReMaterializable, as it is rematerializable
...
when loading from an invariant memory location.
llvm-svn: 83688
2009-10-09 23:28:27 +00:00
87b02d5bbc
Factor out LiveIntervalAnalysis' code to determine whether an instruction
...
is trivially rematerializable and integrate it into
TargetInstrInfo::isTriviallyReMaterializable. This way, all places that
need to know whether an instruction is rematerializable will get the
same answer.
This enables the useful parts of the aggressive-remat option by
default -- using AliasAnalysis to determine whether a memory location
is invariant, and removes the questionable parts -- rematting operations
with virtual register inputs that may not be live everywhere.
llvm-svn: 83687
2009-10-09 23:27:56 +00:00
df45c7f642
Extract scope information from the variable itself, instead of relying on alloca or llvm.dbg.declare location.
...
While recording beginning of a function, use scope info from the first location entry instead of just relying on first location entry itself.
llvm-svn: 83684
2009-10-09 22:42:28 +00:00
307c053f2e
ExecutionEngine::clearGlobalMappingsFromModule failed to remove reverse
...
mappings, which could cause errors and assert-failures. This patch fixes that,
adds a test, and refactors the global-mapping-removal code into a single place.
llvm-svn: 83678
2009-10-09 22:10:27 +00:00
f4eb6777e5
Add a const qualifier.
...
llvm-svn: 83677
2009-10-09 22:09:05 +00:00
96a5b87ae2
Use names instead of numbers for some of the magic
...
constants used in inlining heuristics (especially
those used in more than one file). No functional change.
llvm-svn: 83675
2009-10-09 21:42:02 +00:00
a2b99107c4
Added another bit of the ARM target assembler to llvm-mc to parse register
...
lists. Changed ARMAsmParser::MatchRegisterName to return -1 instead of 0 on
errors so 0-15 values could be returned as register numbers. Also added the
rest of the arm register names to the currently hacked up version to allow more
testing. Some changes to ARMAsmParser::ParseOperand to give different errors
for things not yet supported and some additions to the hacked
ARMAsmParser::MatchInstruction to allow more testing for now.
llvm-svn: 83673
2009-10-09 21:12:28 +00:00
e19c1810d7
isTriviallyReMaterializable checks the
...
TargetInstrDesc::isRematerializable flag, so it isn't necessary to do
this check in its callers.
llvm-svn: 83671
2009-10-09 21:02:10 +00:00
7d9dffb413
Fix the x86 test-shrink optimization so that it doesn't shrink comparisons
...
when one of the bits being tested would end up being the sign bit in the
narrower type, and a signed comparison is being performed, since this would
change the result of the signed comparison. This fixes PR5132.
llvm-svn: 83670
2009-10-09 20:35:19 +00:00
dd76bb23d1
Add basic infrastructure and x86 support for preserving MachineMemOperand
...
information when unfolding memory references.
llvm-svn: 83656
2009-10-09 18:10:05 +00:00
fafa1fe2ad
Check invalid debug info for enums. This may happen when underlyng enum is optimized away. Eventually DwarfChecker will clean this up during llvm verification stage.
...
llvm-svn: 83655
2009-10-09 17:51:49 +00:00
26070c5cff
when previous scratch register is killed, flag the value as no longer tracking
...
llvm-svn: 83653
2009-10-09 17:33:33 +00:00
dc4893abfe
Revert r83606 and add comments explaining why it isn't safe.
...
llvm-svn: 83649
2009-10-09 16:35:06 +00:00
7ddb844a2d
Give Dan and my recent changes, machine LICM is now code size neutral.
...
llvm-svn: 83624
2009-10-09 06:31:25 +00:00
ccd4545ecb
Fix a logic error that caused non-rematable loop invariants loads to be licm'ed out of loop.
...
llvm-svn: 83622
2009-10-09 06:21:52 +00:00
be27d61cbc
Reset kill markers after live interval is reconstructed.
...
llvm-svn: 83608
2009-10-09 01:17:11 +00:00
d65387a49f
Preserve HasNSW and HasNUW when constructing SCEVs for Add and Mul
...
instructions.
llvm-svn: 83606
2009-10-09 00:41:22 +00:00
3059924bdd
When considering whether to inline Callee into Caller,
...
and that will make Caller too big to inline, see if it
might be better to inline Caller into its callers instead.
This situation is described in PR 2973, although I haven't
tried the specific case in SPASS.
llvm-svn: 83602
2009-10-09 00:11:32 +00:00
816fe0a4aa
Add the ability to track HasNSW and HasNUW on more kinds of SCEV expressions.
...
llvm-svn: 83601
2009-10-09 00:10:36 +00:00
84e7967fae
Add codegen support for NEON vst4lane intrinsics with 128-bit vectors.
...
llvm-svn: 83600
2009-10-09 00:01:36 +00:00
c409030838
Add codegen support for NEON vst3lane intrinsics with 128-bit vectors.
...
llvm-svn: 83598
2009-10-08 23:51:31 +00:00
b851eb356a
Add codegen support for NEON vst2lane intrinsics with 128-bit vectors.
...
llvm-svn: 83596
2009-10-08 23:38:24 +00:00
38ba47225a
Add codegen support for NEON vld4lane intrinsics with 128-bit vectors.
...
Also fix some copy-and-paste errors in previous changes.
llvm-svn: 83590
2009-10-08 22:53:57 +00:00
e23984fbbe
Remove code that makes no sense.
...
llvm-svn: 83589
2009-10-08 22:42:35 +00:00
cf54e934f8
Add codegen support for NEON vld3lane intrinsics with 128-bit vectors.
...
llvm-svn: 83585
2009-10-08 22:27:33 +00:00
da3c79698c
Update CMake build yet again after a source file was removed
...
llvm-svn: 83575
2009-10-08 21:24:34 +00:00
3dc625cdce
It's possible for a global variable to be optimized out of a metadata object. So
...
we should allow a "null" with this dyn_cast.
llvm-svn: 83573
2009-10-08 20:52:51 +00:00
222b86cd54
Use lower16 / upper16 imm modifiers to asmprint 32-bit imms splitted via movt/movw pair.
...
llvm-svn: 83572
2009-10-08 20:43:22 +00:00
e1969dcd11
Clear variable debug info map at the end of the function.
...
llvm-svn: 83571
2009-10-08 20:41:17 +00:00
c2728f44a9
Add codegen support for NEON vld2lane intrinsics with 128-bit vectors.
...
llvm-svn: 83568
2009-10-08 18:56:10 +00:00
fac9476589
Clean up some unnecessary initializations.
...
llvm-svn: 83566
2009-10-08 18:52:56 +00:00
4facd965bd
Clean up a comment (indentation was wrong).
...
llvm-svn: 83565
2009-10-08 18:51:31 +00:00
2a45a65511
Add a SelectionDAG getTargetInsertSubreg convenience function,
...
similar to getTargetExtractSubreg.
llvm-svn: 83564
2009-10-08 18:49:46 +00:00
20b2a77765
Do not record line number to implicitly mark start of function if function has arguments. Extra line number entries trip gdb in some cases.
...
llvm-svn: 83563
2009-10-08 18:48:03 +00:00
02fda54e8f
Add missing names for the XCore specific LADD and LSUB nodes.
...
llvm-svn: 83556
2009-10-08 17:14:57 +00:00
09984279fd
Add a form of addPreserved which takes a string argument, to allow passes
...
to declare that they preserve other passes without needing to pull in
additional header file or library dependencies. Convert MachineFunctionPass
and CodeGenLICM to make use of this.
llvm-svn: 83555
2009-10-08 17:00:02 +00:00
4e13316bf9
Add some peepholes for signed comparisons using ashr X, X, 32.
...
llvm-svn: 83549
2009-10-08 15:38:17 +00:00
78765deded
remove LoopVR pass. According to Nick:
...
"LoopVR's logic was copied into ScalarEvolution::getUnsignedRange and
::getSignedRange. Please delete LoopVR."
llvm-svn: 83531
2009-10-08 06:42:44 +00:00
b6b0ab6117
Add codegen support for NEON vst4 intrinsics with <1 x i64> vectors.
...
llvm-svn: 83526
2009-10-08 05:18:18 +00:00
534ea5ae32
Cleanup up unused R3LiveIn tracking.
...
llvm-svn: 83522
2009-10-08 01:50:26 +00:00
c0615aa17f
Re-enable register scavenging in Thumb1 by default.
...
llvm-svn: 83521
2009-10-08 01:46:59 +00:00
48a805bc6c
bugfix. The target may use virtual registers that aren't tracked for re-use but are allocated by the scavenger. The re-use algorithm needs to watch for that.
...
llvm-svn: 83519
2009-10-08 01:09:45 +00:00
71387b4b2f
Add codegen support for NEON vst3 intrinsics with <1 x i64> vectors.
...
llvm-svn: 83518
2009-10-08 00:28:28 +00:00
d4f5670096
Add codegen support for NEON vst2 intrinsics with <1 x i64> vectors.
...
llvm-svn: 83513
2009-10-08 00:21:01 +00:00
dafd08ea7e
In instcombine's debug output, avoid printing ADD for instructions that are
...
already on the worklist, and print Visited when an instruction is about to be
visited. Net, on one input, this reduced the output size by at least 9x.
llvm-svn: 83510
2009-10-08 00:12:24 +00:00
32cc4ec304
Add codegen support for NEON vld4 intrinsics with <1 x i64> vectors.
...
llvm-svn: 83508
2009-10-07 23:54:04 +00:00
5ef3c6d9f4
Add codegen support for NEON vld3 intrinsics with <1 x i64> vectors.
...
llvm-svn: 83506
2009-10-07 23:39:57 +00:00
763be1a248
Add codegen support for NEON vld2 intrinsics with <1 x i64> vectors.
...
llvm-svn: 83502
2009-10-07 22:57:01 +00:00
456735c54b
reverting thumb1 scavenging default due to test failure while I figure out what's up.
...
llvm-svn: 83501
2009-10-07 22:49:41 +00:00
63fe7ff7d9
second half of lazy liveness removal.
...
llvm-svn: 83500
2009-10-07 22:49:30 +00:00
e32fe29d29
Fix handling of x86 'R' constraint.
...
llvm-svn: 83499
2009-10-07 22:47:20 +00:00
267fa622fa
Enable thumb1 register scavenging by default.
...
llvm-svn: 83496
2009-10-07 22:26:31 +00:00
37cf79e5a6
Enable thumb1 register scavenging by default.
...
llvm-svn: 83494
2009-10-07 22:26:14 +00:00
55571bd7bd
Extract subprogram and compile unit information from the debug info attached to an instruction.
...
llvm-svn: 83491
2009-10-07 22:04:08 +00:00
50820a2677
Add some instruction encoding bits for NEON load/store instructions.
...
llvm-svn: 83490
2009-10-07 21:53:04 +00:00
5b741f3d14
80-column and whitespace fixes.
...
llvm-svn: 83489
2009-10-07 21:14:25 +00:00
818b6b96cb
Fixed MCSectionMachO::ParseSectionSpecifier to allow an attribute of "none" so
...
that a symbol stub section with no attributes can be parsed as in:
.section __TEXT,__picsymbolstub4,symbol_stubs,none,16
llvm-svn: 83488
2009-10-07 20:57:20 +00:00
e7ef4a9a6b
Add codegen support for NEON vst4 intrinsics with 128-bit vectors.
...
llvm-svn: 83486
2009-10-07 20:49:18 +00:00
23464866ad
Add codegen support for NEON vst3 intrinsics with 128-bit vectors.
...
llvm-svn: 83484
2009-10-07 20:30:08 +00:00
63849cbd32
grammar
...
llvm-svn: 83483
2009-10-07 19:08:36 +00:00
3dcb5377ef
Add codegen support for NEON vst2 intrinsics with 128-bit vectors.
...
llvm-svn: 83482
2009-10-07 18:47:39 +00:00
2af2253e49
add initializers for clarity. Add missing assignment of PrevLastUseOp.
...
llvm-svn: 83481
2009-10-07 18:44:24 +00:00
de1cb40eb4
Remove LazyLiveness from the tree. It doesn't work right now, and I'm not going to have the time
...
to finish it any time soon. If someone's interested it, they can resurrect it from SVN history.
llvm-svn: 83480
2009-10-07 18:40:17 +00:00
ab3a9474d6
Add codegen support for NEON vld4 intrinsics with 128-bit vectors.
...
llvm-svn: 83479
2009-10-07 18:09:32 +00:00
2207e5fc7b
Add another bit of the ARM target assembler to llvm-mc to parse registers
...
with writeback, things like "sp!", etc. Also added some more stuff to the
temporarily hacked methods ARMAsmParser::MatchRegisterName and
ARMAsmParser::MatchInstruction to allow more parser testing.
llvm-svn: 83477
2009-10-07 18:01:35 +00:00
9b60992120
Replace some code for aggressive-remat with MachineInstr::isInvariantLoad, and
...
teach it how to recognize invariant physical registers.
llvm-svn: 83476
2009-10-07 17:47:20 +00:00
be8137b0b4
Replace TargetInstrInfo::isInvariantLoad and its target-specific
...
implementations with a new MachineInstr::isInvariantLoad, which uses
MachineMemOperands and is target-independent. This brings MachineLICM
and other functionality to targets which previously lacked an
isInvariantLoad implementation.
llvm-svn: 83475
2009-10-07 17:38:06 +00:00
db9493ce68
Add a few simple MachineVerifier checks for MachineMemOperands.
...
llvm-svn: 83474
2009-10-07 17:36:00 +00:00
6bbefc2f67
Add codegen support for NEON vld3 intrinsics with 128-bit vectors.
...
llvm-svn: 83471
2009-10-07 17:24:55 +00:00
99e80228a9
Rearrange code for selecting vld2 intrinsics. No functionality change.
...
This is just to be more consistent with the forthcoming code for vld3/4.
llvm-svn: 83470
2009-10-07 17:23:09 +00:00
fa14dd430c
Add register-reuse to frame-index register scavenging. When a target uses
...
a virtual register to eliminate a frame index, it can return that register
and the constant stored there to PEI to track. When scavenging to allocate
for those registers, PEI then tracks the last-used register and value, and
if it is still available and matches the value for the next index, reuses
the existing value rather and removes the re-materialization instructions.
Fancier tracking and adjustment of scavenger allocations to keep more
values live for longer is possible, but not yet implemented and would likely
be better done via a different, less special-purpose, approach to the
problem.
eliminateFrameIndex() is modified so the target implementations can return
the registers they wish to be tracked for reuse.
ARM Thumb1 implements and utilizes the new mechanism. All other targets are
simply modified to adjust for the changed eliminateFrameIndex() prototype.
llvm-svn: 83467
2009-10-07 17:12:56 +00:00
bcdb02587e
Do not assume that the module is set.
...
llvm-svn: 83462
2009-10-07 16:37:55 +00:00
a4560bad97
Add PR to this FIXME, looks like I didn't commit this change after all.
...
llvm-svn: 83457
2009-10-07 09:22:55 +00:00
986742a14f
Make getPointerTo return a const PointerType* rather than
...
an unqualified PointerType* because it seems more correct.
llvm-svn: 83454
2009-10-07 07:35:19 +00:00
e666bc9f64
Add FreeInst to the "is a call" check for Insts that are calls, but
...
not intrinsics.
llvm-svn: 83441
2009-10-07 00:54:08 +00:00
f8e4410994
Fix this comment. The loop header is the loop entry point.
...
llvm-svn: 83437
2009-10-07 00:33:10 +00:00
75b59fb055
Add PseudoSourceValues for constpool stuff on ELF (Darwin should use something similar)
...
and register spills.
llvm-svn: 83435
2009-10-07 00:06:35 +00:00
6ba26317ce
While we still have a MallocInst treat it as a call like any other
...
for inlining.
When MallocInst goes away this code will be subsumed as part of
calls and work just fine...
llvm-svn: 83434
2009-10-07 00:02:18 +00:00
febe39b488
Added bits of the ARM target assembler to llvm-mc to parse some load instruction
...
operands. Some parsing of arm memory operands for preindexing and postindexing
forms including with register controled shifts. This is a work in progress.
llvm-svn: 83424
2009-10-06 22:26:42 +00:00
e6b778d5ff
Add codegen support for NEON vld2 operations on quad registers.
...
llvm-svn: 83422
2009-10-06 22:01:59 +00:00
74b3d284f2
Use copyRegToReg hook to copy registers.
...
llvm-svn: 83421
2009-10-06 22:01:15 +00:00
e2595b5be6
r83391 was completely broken since Twines keep references to their inputs, and
...
some of the inputs were temporaries. Here's a real fix for the miscompilation.
Thanks to sabre for pointing out the problem.
llvm-svn: 83417
2009-10-06 21:45:26 +00:00
dc7d1ce575
Fix a comment typo.
...
Patch by Johnny Chen.
llvm-svn: 83407
2009-10-06 20:18:46 +00:00
08506ee6b0
Bugfix for the CommaSeparated option. The original code was adding the whole
...
string at the end of the list, instead of the last comma-separated string.
llvm-svn: 83405
2009-10-06 19:55:53 +00:00
2275a7dfef
Update CMake file.
...
llvm-svn: 83404
2009-10-06 19:45:38 +00:00
4598eb6214
Add support to handle debug info attached to an instruction.
...
This is not yet enabled.
llvm-svn: 83400
2009-10-06 18:37:31 +00:00
10d3dc569b
Instead of printing unnecessary basic block labels as labels in
...
verbose-asm mode, print comments instead. This eliminates a non-comment
difference between verbose-asm mode and non-verbose-asm mode.
Also, factor out the relevant code out of all the targets and into
target-independent code.
llvm-svn: 83392
2009-10-06 17:38:38 +00:00
92e3a06623
Fix PR5112, a miscompilation on gcc-4.0.3. Patch by Collin Winter!
...
llvm-svn: 83391
2009-10-06 17:25:50 +00:00
a893f5bdf5
remove predicate simplifier, it never got the last bugs beaten
...
out of it, and jump threading, condprop and gvn are now getting
most of the benefit. This was approved by Nicholas and Nicolas.
llvm-svn: 83390
2009-10-06 16:59:46 +00:00
692f6e7f9d
Remove xs1b predicate since it is no longer needed to differentiate betweem
...
xs1a and xs1b.
llvm-svn: 83383
2009-10-06 16:17:57 +00:00
d7b887410d
Remove xs1a subtarget. xs1a is a preproduction device used in
...
early development boards which is no longer supported in the
XMOS toolchain.
llvm-svn: 83381
2009-10-06 16:01:09 +00:00
29e8555056
Default to the xs1b subtarget
...
llvm-svn: 83380
2009-10-06 15:41:52 +00:00
9ed7b16bf3
Introduce and use convenience methods for getting pointer types
...
where the element is of a basic builtin type. For example, to get
an i8* use getInt8PtrTy.
llvm-svn: 83379
2009-10-06 15:40:36 +00:00
fa21fe7c3d
grammar
...
llvm-svn: 83378
2009-10-06 15:03:44 +00:00
7d838bb66e
Fix cut-n-pasto.
...
llvm-svn: 83367
2009-10-06 03:15:38 +00:00
eb43b17074
Update processDebugLoc() to handle requests to process debug info, before and after emitting instructions.
...
llvm-svn: 83364
2009-10-06 03:04:58 +00:00
051454a16f
Update processDebugLoc() so that it can be used to process debug info before and after printing an instruction.
...
llvm-svn: 83363
2009-10-06 02:19:11 +00:00
2980a22028
Remove dead code.
...
llvm-svn: 83362
2009-10-06 02:01:32 +00:00
8db360da7b
Add utility routine to set begin and end labels for DbgScopes.
...
This will be used by processDebugLoc().
llvm-svn: 83361
2009-10-06 01:50:42 +00:00
849e59abb2
Remove unintentional function decl.
...
llvm-svn: 83356
2009-10-06 01:31:35 +00:00
475d32a987
Add utility routine to collect variable debug info. This is not yet used.
...
llvm-svn: 83355
2009-10-06 01:26:37 +00:00
e97fe329e9
Fix http://llvm.org/PR5116 by rolling back r60822. This passes `make unittests
...
check-lit` on both x86-64 Linux and x86-32 Darwin.
llvm-svn: 83353
2009-10-06 00:35:55 +00:00
bb802206d2
Set default location for the function if it is not already set.
...
This code is not yet enabled.
llvm-svn: 83349
2009-10-06 00:09:08 +00:00
d859d86538
Existence of a compile unit for input source file is a good indicator to check debug info's presence in a module.
...
llvm-svn: 83348
2009-10-06 00:03:14 +00:00
1c9eef72b4
If subprogram die is not available then construct new one.
...
This can happen if debug info is processed lazily.
llvm-svn: 83347
2009-10-05 23:59:00 +00:00
4c420eca41
Adjust context for the global variables that are not at file scope, e.g.
...
void foo() { static int bar = 42; }
Here, foo's DIE is parent of bar's DIE.
llvm-svn: 83344
2009-10-05 23:40:42 +00:00
4144a82154
Set address while constructing DIE.
...
llvm-svn: 83343
2009-10-05 23:22:08 +00:00
44e5dbcaf0
Extend ConstantFolding to understand signed overflow variants
...
llvm-svn: 83338
2009-10-05 22:53:52 +00:00
2dfb5da6bb
In Thumb1, the register scavenger is not always able to use an emergency
...
spill slot. When frame references are via the frame pointer, they will be
negative, but Thumb1 load/store instructions only allow positive immediate
offsets. Instead, Thumb1 will spill to R12.
llvm-svn: 83336
2009-10-05 22:30:23 +00:00
c5b8846891
Don't treat malloc calls with non-matching prototype as malloc.
...
Fixes second part of PR5130, miscompilation in FreeBSD kernel, where malloc takes 3 params,
and *does* initialize memory.
llvm-svn: 83324
2009-10-05 21:15:43 +00:00
50d75a6099
No newline at end of files.
...
llvm-svn: 83318
2009-10-05 18:43:19 +00:00
2089d16140
Gracefully handle various scopes while recording source line info.
...
llvm-svn: 83317
2009-10-05 18:03:19 +00:00
e525d9ddc0
Remove an unnnecessary LLVMContext argument in
...
ConstantFoldLoadThroughGEPConstantExpr.
llvm-svn: 83311
2009-10-05 16:36:26 +00:00
238cf49812
Use Use::operator= instead of Use::set, for consistency.
...
llvm-svn: 83310
2009-10-05 16:31:55 +00:00
2728569a38
Remove explicit enum integer values. They don't appear to be needed, and
...
they make it less convenient to add new entries.
llvm-svn: 83308
2009-10-05 15:52:08 +00:00
774149a878
Add RIP to GR64_NOREX. This fixed a MachineVerifier error when RIP
...
is used in an operand which requires GR64_NOREX.
llvm-svn: 83307
2009-10-05 15:42:08 +00:00
fdd8790718
strength reduce a ton of type equality tests to check the typeid (Through
...
the new predicates I added) instead of going through a context and doing a
pointer comparison. Besides being cheaper, this allows a smart compiler
to turn the if sequence into a switch.
llvm-svn: 83297
2009-10-05 05:54:46 +00:00
59d939894b
teach the optimizer how to constant fold uadd/usub intrinsics.
...
llvm-svn: 83295
2009-10-05 05:26:04 +00:00
351534f9b1
simplify this code a bunch.
...
llvm-svn: 83294
2009-10-05 05:06:24 +00:00
9ca7c0913a
code simplifications.
...
llvm-svn: 83292
2009-10-05 05:00:35 +00:00
463716d559
instcombine shouldn't delete all null checks for mallocs.
...
This fixes PR5130.
llvm-svn: 83290
2009-10-05 02:47:47 +00:00
ff00c3562a
stop MachineFunctionPass from claiming that it preserves LoopDependence info,
...
which causes dependence info to be linked into lli.
llvm-svn: 83289
2009-10-05 02:35:05 +00:00
cb08fb7b83
remove llvm-db: it is completely broken and if anyone wants to do a debugger,
...
they should not base it on llvm-db (which not following almost any "best practices").
llvm-svn: 83288
2009-10-05 02:29:51 +00:00
b5049bebb3
Do away with the strange use of BitVectors in SSI, and just use normal sets. This makes the code much more C++/LLVM-ish.
...
llvm-svn: 83286
2009-10-04 18:49:55 +00:00
63c733faf9
Whitespace and formatting.
...
llvm-svn: 83285
2009-10-04 18:18:39 +00:00
286feb16a9
Fix a typo in the comment.
...
llvm-svn: 83283
2009-10-04 17:52:13 +00:00
a62bf10651
SSI needs to require DT and DF transitively, since it uses them outside of its runOnFunction.
...
Similarly, it can be marked setPreservesAll, since it does no work in its runOnFunction.
llvm-svn: 83282
2009-10-04 17:47:39 +00:00
bb4ed2394b
Allow -inline-threshold override default threshold even if compiling to optimize for size.
...
llvm-svn: 83274
2009-10-04 06:13:54 +00:00
ae4617c974
Requires element types in a constant initializer to match the element types of
...
of the constant. This reverts r6544 and r7428.
llvm-svn: 83270
2009-10-03 19:30:43 +00:00
d76b9b766c
Add a comment to describe letters used in multiclass name suffixes.
...
llvm-svn: 83257
2009-10-03 04:44:16 +00:00
a9abf57409
Fix encoding problem for VMLS instruction.
...
Thanks to Johnny Chen for pointing this out!
llvm-svn: 83256
2009-10-03 04:41:21 +00:00
a7780905e7
Oops. Renamed remaining MachineInstrIndex references.
...
llvm-svn: 83255
2009-10-03 04:31:31 +00:00
920301ecc5
Renamed MachineInstrIndex to LiveIndex.
...
llvm-svn: 83254
2009-10-03 04:21:37 +00:00
f01ee30328
MingW build fixes
...
- MingW needs -lpsapi (in ${LIBS}) linked after -lLLVMSystem.
Noticed by Ronald Pijnacker!
- Some parts of the System library must be build with exceptions on windows.
Based on a patch by Jay Foad!
llvm-svn: 83251
2009-10-02 19:36:31 +00:00
3b008a3a65
Fix a use-after-free in post-ra-scheduling.
...
MI->addOperand invalidates references to it's operands, avoid touching
the operand after a new one was added.
llvm-svn: 83249
2009-10-02 15:59:52 +00:00
32a47ea7b6
getFunctionAlignment should return log2 alignment.
...
llvm-svn: 83242
2009-10-02 06:57:25 +00:00
b659dff4eb
Forgot about ARM::tPUSH. It also has a new writeback operand.
...
llvm-svn: 83237
2009-10-02 05:03:07 +00:00
2dcee28a61
Move load / store multiple before post-alloc scheduling.
...
llvm-svn: 83236
2009-10-02 04:57:15 +00:00
d725159d9d
All callee-saved registers are live-out of a return block.
...
llvm-svn: 83223
2009-10-01 23:28:47 +00:00
1cc6dd97da
Remove neonfp attribute and instead set default based on CPU string. Add -arm-use-neon-fp to override the default.
...
llvm-svn: 83218
2009-10-01 22:19:57 +00:00
70cb67fcf4
Expand api out in the usual inserter way, though, I do have a
...
question, can we get rid of the BasicBlock versions of all inserters
and use Head == 0 to indicate the old case when GetInsertBlock == 0?
llvm-svn: 83216
2009-10-01 22:08:58 +00:00
9a051a5922
Restore the -post-RA-scheduler flag as an override for the target specification. Remove -mattr for setting PostRAScheduler enable and instead use CPU string.
...
llvm-svn: 83215
2009-10-01 21:46:35 +00:00
6f012d83f2
ARM::tPOP and tPOP_RET each has an extra writeback operand now.
...
llvm-svn: 83214
2009-10-01 20:54:53 +00:00
75cc16c0f2
Add support to extract lexical scope information from DebugLoc attached with an machine instruction.
...
This is not yet enabled.
llvm-svn: 83210
2009-10-01 20:31:14 +00:00
be3039e776
Use MachineFrameInfo.getPristineRegs() to determine which callee-saved registers are available for anti-dependency breaking. Some cleanup.
...
llvm-svn: 83208
2009-10-01 19:45:32 +00:00
787f94c28d
Record first and last instruction of a scope in DbgScope.
...
llvm-svn: 83207
2009-10-01 18:25:23 +00:00
4fb2891396
Observe hasExtraSrcRegAllocReq and hasExtraDefRegAllocReq. Do not change
...
operands of instructions with these properties while breaking anti-dep.
llvm-svn: 83198
2009-10-01 08:26:23 +00:00
1b2b64f618
Add hasExtraSrcRegAllocReq and hasExtraDefRegAllocReq flags to ld / st multiple,
...
ld / st pairs, etc.
llvm-svn: 83197
2009-10-01 08:22:27 +00:00
d846fbf20d
Remove GVNPRE.cpp from the CMake makefile
...
llvm-svn: 83194
2009-10-01 05:30:05 +00:00
5f3cc06cd2
remove the GVNPRE pass. It has been subsumed by the GVN pass.
...
Ok'd by Owen.
llvm-svn: 83193
2009-10-01 02:18:36 +00:00
4bcd523acb
Update ARM JIT emitter to account for ld/st multiple changes.
...
llvm-svn: 83192
2009-10-01 01:39:21 +00:00
3bbc6c3ae6
Change ld/st multiples to explicitly model the writeback to base register. This fixes most of the -ldstopti-before-sched2 regressions.
...
llvm-svn: 83191
2009-10-01 01:33:39 +00:00
34986f12e6
Add another MDNode into DebugLocTuple. This will be used to keep track of inlined functions.
...
llvm-svn: 83190
2009-10-01 01:15:28 +00:00
4dbca6dfd4
If location info is attached with an instruction then keep track of alloca slots used by a variable. This info will be used by AsmPrinter to emit debug info for variables.
...
llvm-svn: 83189
2009-10-01 01:03:26 +00:00
e0709cfc92
Use MachineInstr as an processDebugLoc() argument.
...
This will allow processDebugLoc() to handle scopes for DWARF debug info.
llvm-svn: 83183
2009-09-30 23:12:50 +00:00
3256c751f5
Use MDNode * directly as an RecordSourceLine() argument.
...
llvm-svn: 83182
2009-09-30 22:51:28 +00:00
5c13c2d392
Remove dead code.
...
llvm-svn: 83181
2009-09-30 22:43:52 +00:00
a49be76f43
Add isFOO() helpers. Fix getDirectory() and getFilename() for DIScope.
...
llvm-svn: 83180
2009-09-30 22:34:41 +00:00
b2120755a2
Use OutStreamer.SwitchSection instead of writing out textual section directives.
...
Add a new TargetLoweringObjectFileMachO::getConstTextCoalSection method to
get access to that section.
llvm-svn: 83178
2009-09-30 22:25:37 +00:00
b633d7a665
Add a new virtual EmitStartOfAsmFile method to the AsmPrinter and use this
...
to emit target-specific things at the beginning of the asm output. This
fixes a problem for PPC, where the text sections are not being kept together
as expected. The base class doInitialization code calls DW->BeginModule()
which emits a bunch of DWARF section directives. The PPC doInitialization
code then emits all the TEXT section directives, with the intention that they
will be kept together. But as I understand it, the Darwin assembler treats
the default TEXT section as a special case and moves it to the beginning of
the file, which means that all those DWARF sections are in the middle of
the text. With this change, the EmitStartOfAsmFile hook is called before
the DWARF section directives are emitted, so that all the PPC text section
directives come out right at the beginning of the file.
llvm-svn: 83176
2009-09-30 22:06:26 +00:00
64c8d5a004
Fix a comment typo.
...
llvm-svn: 83174
2009-09-30 21:44:42 +00:00
48575e3f2e
Check for null MDNode element while printing comment.
...
llvm-svn: 83172
2009-09-30 21:26:51 +00:00
53904987ae
Fix a comment.
...
llvm-svn: 83171
2009-09-30 21:26:13 +00:00
699702e0a8
The AsmPrinter base class contains a DwarfWriter member, so there's no need
...
for derived AsmPrinters to add another one. In some cases, fixing this
removes the need to override the doInitialization method.
llvm-svn: 83170
2009-09-30 21:24:45 +00:00
8ce67f83bc
Assert that ConstantArrays are created with correctly-typed elements.
...
llvm-svn: 83168
2009-09-30 21:08:08 +00:00
ea0bb8f555
Fix this code so that it doesn't try to iterate through a std::vector
...
while calling changeImmediateDominator, which removes elements from the
vector. This fixes PR5097.
llvm-svn: 83166
2009-09-30 20:54:16 +00:00
cea8dab1d1
Silence comparison always false warning in -Asserts mode.
...
llvm-svn: 83164
2009-09-30 20:43:07 +00:00
c87197784a
Add additional assert() to verify no extraneous use of a scavenged register.
...
llvm-svn: 83163
2009-09-30 20:35:36 +00:00
711ab5b8ec
Print tag name for MDNodes that are used to encode debug info.
...
llvm-svn: 83160
2009-09-30 20:16:54 +00:00
8ff5c19ebd
Fix integer overflow in instruction scheduling. This can happen if we have
...
basic blocks that are so long that their size overflows a short.
Also assert that overflow does not happen in the future, as requested by Evan.
This fixes PR4401.
llvm-svn: 83159
2009-09-30 20:15:38 +00:00
a2fe1a6811
Clarify comment phrasing.
...
llvm-svn: 83148
2009-09-30 15:23:38 +00:00
ce5a8ca3ef
Add a option which would move ld/st multiple pass before post-alloc scheduling.
...
llvm-svn: 83145
2009-09-30 08:53:01 +00:00
f305ead1cc
Add a target hook to add pre- post-regalloc scheduling passes.
...
llvm-svn: 83144
2009-09-30 08:49:50 +00:00
6d59690c3e
Fix compile error as debug interface changed.
...
By the way, this code is buggy. You can't keep a map<MDNode *, something>
because the MDNode may be destroyed and reused for something else.
llvm-svn: 83141
2009-09-30 04:50:26 +00:00
882f4c11ed
replace TRI->isVirtualRegister() with TargetRegisterInfo::isVirtualRegister()
...
per customary usage
llvm-svn: 83137
2009-09-30 01:47:59 +00:00
70ce8a03b1
When checking whether we need to reserve a register for the scavenger,
...
the size of the saved frame pointer needs to be taken into account.
llvm-svn: 83136
2009-09-30 01:43:29 +00:00
bcad0c8421
Add "isBarrier = 1" to return instructions.
...
Patch by Sylvere Teissier.
llvm-svn: 83135
2009-09-30 01:35:11 +00:00
cdd3e35005
fix compiler warning
...
llvm-svn: 83132
2009-09-30 00:37:40 +00:00
20e5f5ed79
For Darwin, emit all the text section directives together before the dwarf
...
section directives. This causes the assembler to put the text sections at
the beginning of the object file, which helps work around a limitation of the
Darwin ARM relocations. Radar 7255355.
llvm-svn: 83127
2009-09-30 00:23:42 +00:00
c2105298cc
Simplify.
...
llvm-svn: 83123
2009-09-30 00:14:40 +00:00
17199b56b0
Remove -post-RA-schedule flag and add a TargetSubtarget method to enable post-register-allocation scheduling. By default it is off. For ARM, enable/disable with -mattr=+/-postrasched. Enable by default for cortex-a8.
...
llvm-svn: 83122
2009-09-30 00:10:16 +00:00
14cf8ecf0b
Add a way for a frontend to generate more complex dwarf location
...
information. This allows arbitrary code involving DW_OP_plus_uconst
and DW_OP_deref. The scheme allows for easy extention to include,
any, or all of the DW_OP_ opcodes. I thought about just exposing all
of them, but, wasn't sure if people wanted the dwarf opcodes exposed
in the api. Is that a layering violation?
With this scheme, the entire existing block scheme used by llvm-gcc
can be switched over to the new scheme. I think that would be
cleaner, as then the compiler specific bits are not present in llvm
proper. Before the old code can be yanked however, similar code in
clang would have to be removed.
Next up, more testing.
llvm-svn: 83120
2009-09-30 00:08:22 +00:00
fa6847f099
minor cleanup and add clarifying comment
...
llvm-svn: 83117
2009-09-29 23:17:20 +00:00
7b4d52b170
Lookup handler name only when assertions are enabled.
...
llvm-svn: 83114
2009-09-29 22:05:52 +00:00
b4034364d6
Add removeMD().
...
llvm-svn: 83107
2009-09-29 20:42:25 +00:00
5bf7a49fd7
Only one custom meadata of each kind can be attached with an instruction.
...
llvm-svn: 83105
2009-09-29 20:30:57 +00:00
3ea3fe6594
Additional check for regno==0
...
llvm-svn: 83103
2009-09-29 20:11:10 +00:00
561977940c
Use assertion instead of early exit to catch malformed custom metadata store.
...
llvm-svn: 83102
2009-09-29 20:01:19 +00:00
5d58383ea9
Remove unnecessary cast.
...
llvm-svn: 83100
2009-09-29 19:56:13 +00:00
b296942f6d
Remove std::string uses from DebugInfo interface.
...
llvm-svn: 83083
2009-09-29 18:40:58 +00:00
a4a7f44cb5
Simplify the tracking of virtual frame index registers. Ranges cannot overlap,
...
so a simple "current register" will suffice. Also add some additional
sanity-checking assertions to make sure things are as we expect.
llvm-svn: 83081
2009-09-29 18:23:15 +00:00
8fc22227bb
Moving register scavenging to a post pass results in virtual registers in
...
the instruction we're scavenging for. The scavenger needs to know to avoid
them when analyzing register usage.
llvm-svn: 83077
2009-09-29 17:24:37 +00:00
139c3dba53
Fix PR4687. Pre ARMv5te does not support ldrd / strd. Patch by John Tytgat.
...
llvm-svn: 83058
2009-09-29 07:07:30 +00:00
ba4a6fdd17
Parse custom metadata attached with an instruction.
...
llvm-svn: 83033
2009-09-29 00:01:14 +00:00
5264202a38
Adjust processFunctionBeforeCalleeSavedScan() to correctly reserve a stack
...
slot for the register scavenger when compiling Thumb1 functions.
llvm-svn: 83023
2009-09-28 22:08:06 +00:00
f919bd6651
Add C API calls for building FNeg operations. Patch by KS Sreeram!
...
llvm-svn: 83021
2009-09-28 21:51:41 +00:00
2d85eef974
s/class Metadata/class MetadataContext/g
...
llvm-svn: 83019
2009-09-28 21:41:20 +00:00
b1a4477f1f
Do not use global typedef for MDKindID.
...
llvm-svn: 83016
2009-09-28 21:14:55 +00:00
f230656a02
When extending the operands of an addrec, iterate through all
...
the operands, rather than trying to partition them into a start
and a step. This handles non-affine add recurrences correctly.
llvm-svn: 83011
2009-09-28 21:01:47 +00:00
565371b4c9
Do not hardcode metadata names.
...
llvm-svn: 83010
2009-09-28 20:56:00 +00:00
4854ef0023
Fix Thumb2 IT block pass bug. t2MOVi32imm may not be the start of a IT block.
...
llvm-svn: 83008
2009-09-28 20:47:15 +00:00
0bb5af345a
Use KILL instead of IMPLICIT_DEF in LowerSubregs pass.
...
llvm-svn: 83007
2009-09-28 20:32:46 +00:00
dc9efe8078
Introduce the TargetInstrInfo::KILL machine instruction and get rid of the
...
unused DECLARE instruction.
KILL is not yet used anywhere, it will replace TargetInstrInfo::IMPLICIT_DEF
in the places where IMPLICIT_DEF is just used to alter liveness of physical
registers.
llvm-svn: 83006
2009-09-28 20:32:26 +00:00
d1415a8e56
Create a README.txt for lib/Analysis, and add an entry.
...
llvm-svn: 83001
2009-09-28 18:38:53 +00:00
59c0c1388f
Remove dead code.
...
llvm-svn: 82999
2009-09-28 18:31:56 +00:00
6905f15256
Use VerifySchedule instead of doing the work manually.
...
llvm-svn: 82995
2009-09-28 16:09:41 +00:00
86dc886584
Fix this debug output to handle the case where the loop has been deleted.
...
llvm-svn: 82994
2009-09-28 15:40:01 +00:00
0bd312afd8
Include the name of the loop header in debug messages.
...
llvm-svn: 82993
2009-09-28 15:07:18 +00:00
7d3b0be05b
Remove a redundant #ifndef and add an assertion string.
...
llvm-svn: 82991
2009-09-28 14:38:19 +00:00
9a7320c711
Convert LoopSimplify and LoopExtractor from FunctionPass to LoopPass.
...
llvm-svn: 82990
2009-09-28 14:37:51 +00:00
2dd957fff6
Pass the optimization level when constructing the ARM instruction selector.
...
Otherwise, it is always set to "default", which prevents debug info from
even being generated during isel. Radar 7250345.
llvm-svn: 82988
2009-09-28 14:30:20 +00:00
83e0d481ae
Make ARM and Thumb2 32-bit immediate materialization into a single 32-bit pseudo
...
instruction. This makes it re-materializable.
Thumb2 will split it back out into two instructions so IT pass will generate the
right mask. Also, this expose opportunies to optimize the movw to a 16-bit move.
llvm-svn: 82982
2009-09-28 09:14:39 +00:00
c30d816d7a
Fix thinko in my recent movt commit: it's not safe to remat movt, since it has input reg argument.
...
Disable rematting of it for now.
llvm-svn: 82975
2009-09-28 07:26:46 +00:00
0261b5d2d2
The select instruction is not neccesarily in the same block as the
...
phi nodes. Make sure to phi translate from the right block.
This fixes a llvm-building-llvm failure on GVN-PRE.cpp
llvm-svn: 82970
2009-09-28 06:49:44 +00:00
e0c5313493
Coalescer should not delete extract_subreg, insert_subreg, and subreg_to_reg of
...
physical registers. This is especially critical for the later two since they
start the live interval of a super-register. e.g.
%DO<def> = INSERT_SUBREG %D0<undef>, %S0<kill>, 1
If this instruction is eliminated, the register scavenger will not be happy as
D0 is not defined previously.
This fixes PR5055.
llvm-svn: 82968
2009-09-28 05:28:43 +00:00
08d86dac71
Remove temporary debugging hack.
...
llvm-svn: 82953
2009-09-28 00:44:15 +00:00
4dbb301f17
Move the dominator verification code out of special code embedded within
...
the PassManager code into a regular verifyAnalysis method.
Also, reorganize loop verification. Make the LoopPass infrastructure
call verifyLoop as needed instead of having LoopInfo::verifyAnalysis
check every loop in the function after each looop pass. Add a new
command-line argument, -verify-loop-info, to enable the expensive
full checking.
llvm-svn: 82952
2009-09-28 00:27:48 +00:00
2963777d0c
Move this assert to check the condition as soon as it is known.
...
llvm-svn: 82951
2009-09-28 00:10:28 +00:00
277e767578
Extend the StartPassTimer and StopPassTimer functions so that the
...
code that stops the timer doesn't have to search to find the timer
object before it stops the timer. This avoids a lock acquisition
and a few other things done with the timer running.
llvm-svn: 82949
2009-09-28 00:07:05 +00:00
7c2b1e71c1
Use movt/movw pair to materialize 32 bit constants on ARMv6T2+.
...
This should be better than single load from constpool.
llvm-svn: 82948
2009-09-27 23:52:58 +00:00
b0934cdd1c
Fix an old copy+pasto.
...
llvm-svn: 82947
2009-09-27 23:52:07 +00:00
adde5dfde2
Extract the code for inserting a loop into the loop queue into
...
a separate function.
llvm-svn: 82946
2009-09-27 23:49:43 +00:00
37a99664ff
When a loop is deleted, immediately release all of the active
...
LoopPasses for that loop. This avoids trouble with the PassManager
trying to call verifyAnalysis on them, and frees up some memory
sooner rather than later.
llvm-svn: 82945
2009-09-27 23:43:07 +00:00
5e8ba5d514
Extract the code for releasing a pass into a separate function, and
...
tidy it up a little.
llvm-svn: 82944
2009-09-27 23:38:27 +00:00
93e06a57c9
Remove the "metadata*" type and simplify the code it complicated. This was only
...
used to support GlobalVariables storing MDNodes, back when they were derived
from Constant before the introduction of NamedMDNode, but never removed.
llvm-svn: 82943
2009-09-27 23:27:42 +00:00
a5fc03562f
LBRX no longer has an explicit SrcValueSDNode operand, so the type
...
operand is now at index 2, rather than 3. This fixes the
"Invalid child # of SDNode!" failures on PowerPC.
llvm-svn: 82942
2009-09-27 23:17:47 +00:00
4425660b1f
simplify some code.
...
llvm-svn: 82936
2009-09-27 21:46:50 +00:00
b2e88cd01c
The bitcast case is not needed here: instcombine turns icmp(bitcast(x), null) -> icmp(x, null) already.
...
llvm-svn: 82935
2009-09-27 21:42:46 +00:00
8b4d3dfbbf
calls are already unmovable, malloc doesn't need a special case.
...
llvm-svn: 82933
2009-09-27 21:36:19 +00:00
f9e0c7f84b
calls to external functions are already marked overdefined, special casing
...
malloc isn't needed.
llvm-svn: 82932
2009-09-27 21:35:11 +00:00
5abb1e4cd2
calls are already handled, malloc doesn't need a special case.
...
llvm-svn: 82931
2009-09-27 21:33:46 +00:00
28a5f25d87
Round out the API for the new optimization flags.
...
llvm-svn: 82930
2009-09-27 21:33:04 +00:00
466d57f6c1
calls are rejected above, no need to special case malloc here.
...
llvm-svn: 82929
2009-09-27 21:31:39 +00:00
43d0db70ac
remove special handling of bitcast(malloc), it will be handled
...
when the loop inspects the bitcast operand.
llvm-svn: 82928
2009-09-27 21:29:28 +00:00
a8627272c1
unlike the malloc instruction, "malloc" calls do not claim to be readonly, just nounwind.
...
llvm-svn: 82927
2009-09-27 21:23:38 +00:00
b391e87263
allow pushing icmps through phis with multiple uses and across critical edges.
...
These are important to push up to encourage jump threading. This shrinks 176.gcc a bit.
llvm-svn: 82923
2009-09-27 20:46:36 +00:00
ae289632ef
Enhance the previous fix for PR4895 to allow more values than just
...
simple constants for the true/false value of the select. We now
do phi translation etc. This really fixes PR4895 :)
llvm-svn: 82917
2009-09-27 20:18:49 +00:00
facb867af3
implement PR4895, by making FoldOpIntoPhi handle select conditions
...
that are phi nodes. Also tighten up FoldOpIntoPhi to treat constantexpr
operands to phis just like other variables, avoiding moving constantexpr
computations around.
Patch by Daniel Dunbar.
llvm-svn: 82913
2009-09-27 19:57:57 +00:00
336e2bd91b
Use explicit structs instead of std::pair to map callee saved regs to spill slots.
...
llvm-svn: 82909
2009-09-27 17:58:47 +00:00
cb663d749d
Delete a bogus comment.
...
llvm-svn: 82908
2009-09-27 17:50:44 +00:00
2aab867048
Fix SCEVExpander's canonical addrec expansion code to work on loops that
...
aren't in canonical loop-simplify form, since it doesn't itself depend
on LoopSimplify. This means handling loops without preheaders and loops
with multiple backedges.
llvm-svn: 82905
2009-09-27 17:46:40 +00:00
0e70af36c0
Grab an LLVM Context from an instruction that exists rather than one
...
that is deleted in some situations. This fixes a use-after-free.
llvm-svn: 82903
2009-09-27 16:10:30 +00:00
fc20b67e80
Tell ScalarEvolution to forget everything it knows about a loop before
...
rotating the loop, since loop rotation is a very significant change.
llvm-svn: 82901
2009-09-27 15:37:03 +00:00
4338f905b7
Micro-optimize DerivedType::dropAllTypeUses.
...
llvm-svn: 82900
2009-09-27 15:32:21 +00:00
6daa565cf9
Instead of testing whether an instruction dominates the loop preheader,
...
test whether it properly dominates the loop header. This is equivalent
when the loop has a preheader, and has the advantage of working when
the loop doesn't have a preheader. Since IVUsers doesn't Require
LoopSimplify, the loop isn't guaranteed to have a preheader.
llvm-svn: 82899
2009-09-27 15:30:00 +00:00
f39a3a9288
Add a properlyDominates member function to ScalarEvolution.
...
llvm-svn: 82898
2009-09-27 15:26:03 +00:00
0523507a2f
Remove a redundant #include.
...
llvm-svn: 82897
2009-09-27 15:21:52 +00:00
819d15c7d9
Avoid using mutex locks if not in multithreaded mode by using a SmartScopedMutex
...
in RegisterStatistic.
llvm-svn: 82896
2009-09-27 11:08:03 +00:00
a6b9cab822
Enable pre-regalloc load / store multiple pass for Thumb2.
...
llvm-svn: 82893
2009-09-27 09:46:04 +00:00
14d1cccc2b
Remove dead code from this function and optimize. Update its corresponding
...
LangRef entry too.
llvm-svn: 82890
2009-09-27 07:55:32 +00:00
42fb7452df
Instruction::clone does not need to take an LLVMContext&. Remove that and
...
update all the callers.
llvm-svn: 82889
2009-09-27 07:38:41 +00:00
37d8015dc1
remove support for "NoSub" from regex. It seems like a minor optimization
...
and makes the API more annoying. Add a Regex::getNumMatches() method.
llvm-svn: 82877
2009-09-26 21:27:04 +00:00
62995c71a2
Fix SimplifyLibCalls to transfer attributes from callees rather than
...
calls, since direct calls don't always reflect the attributes of their
callees.
llvm-svn: 82867
2009-09-26 18:10:13 +00:00
5bafe38916
Fix a case where ScalarEvolution was expanding pointer arithmetic
...
to inttoptr/ptrtoint unnecessarily.
llvm-svn: 82864
2009-09-26 16:11:57 +00:00
129de48982
For the NSWSub support in the builder to actually be useable,
...
there need to be corresponding changes to the constant folders,
done in this patch.
llvm-svn: 82862
2009-09-26 15:35:35 +00:00
832800aa6f
Convert comparisons like (x == infinity) to (x >= infinity) on targets
...
where FCMP_OEQ is not legal and FCMP_OGE is, such as x86.
llvm-svn: 82861
2009-09-26 15:24:17 +00:00
6a3bdd872c
Really remove this option.
...
llvm-svn: 82838
2009-09-26 02:49:49 +00:00
d0fe5abc23
Remove a couple of unused command line options.
...
llvm-svn: 82837
2009-09-26 02:45:45 +00:00
4a949408fb
Add comment.
...
llvm-svn: 82836
2009-09-26 02:43:36 +00:00
2f5bdcb7c2
Don't hoist or sink instructions with physreg uses if the physreg is
...
allocatable. Even if it doesn't appear to have any defs, it may latter
on after register allocation.
llvm-svn: 82834
2009-09-26 02:34:00 +00:00
e30d63f1d8
Unbreak MachineLICM for instructions that reference RIP on x86-64 too.
...
llvm-svn: 82825
2009-09-25 23:58:45 +00:00
394468dc8e
Rename ConstantFP's getInf to getInfinity.
...
llvm-svn: 82823
2009-09-25 23:40:21 +00:00
5a6b11cb71
Move MachineMemOperand::getAlignment out of line, to avoid needing
...
MathExtras.h in MachineMemOperand.h.
llvm-svn: 82822
2009-09-25 23:33:20 +00:00
f6dbd6f26e
Remove unused variable.
...
llvm-svn: 82821
2009-09-25 23:26:56 +00:00
5ffd53892d
Transform pow(x, 0.5) to (x == -inf ? inf : fabs(sqrt(x))), which is
...
typically faster then doing a general pow.
llvm-svn: 82819
2009-09-25 23:10:17 +00:00
feb5021134
Add a ConstantFP::getInf utility function for creating infinity ConstantFPs.
...
llvm-svn: 82818
2009-09-25 23:00:48 +00:00
a317687e85
Fix MachineSink to be able to sink instructions that use physical registers
...
which have no defs anywhere in the function. In particular, this fixes sinking
of instructions that reference RIP on x86-64, which is currently being modeled
as a register.
llvm-svn: 82815
2009-09-25 22:53:29 +00:00
a3090655d5
Regenerate
...
llvm-svn: 82814
2009-09-25 22:53:17 +00:00
e603710d11
Simplify a few more uses of reg_iterator.
...
llvm-svn: 82812
2009-09-25 22:26:13 +00:00
5edd45a76b
Simplify this code by using use_iterator instead of reg_iterator
...
and skipping the defs.
llvm-svn: 82811
2009-09-25 22:24:52 +00:00
fd6aad7f1c
Code clean up and prepare for Thumb2 support. No functionality changes.
...
llvm-svn: 82805
2009-09-25 21:44:53 +00:00
3872b3c13e
Flip -disable-post-RA-scheduler to -post-RA-scheduler.
...
llvm-svn: 82803
2009-09-25 21:38:11 +00:00
f6a987b784
Handle sqrt in CannotBeNegativeZero. absf and absl
...
appear to be misspellings, removed in favor of fabs*.
llvm-svn: 82796
2009-09-25 20:54:50 +00:00
48b185d6f7
Improve MachineMemOperand handling.
...
- Allocate MachineMemOperands and MachineMemOperand lists in MachineFunctions.
This eliminates MachineInstr's std::list member and allows the data to be
created by isel and live for the remainder of codegen, avoiding a lot of
copying and unnecessary translation. This also shrinks MemSDNode.
- Delete MemOperandSDNode. Introduce MachineSDNode which has dedicated
fields for MachineMemOperands.
- Change MemSDNode to have a MachineMemOperand member instead of its own
fields with the same information. This introduces some redundancy, but
it's more consistent with what MachineInstr will eventually want.
- Ignore alignment when searching for redundant loads for CSE, but remember
the greatest alignment.
Target-specific code which previously used MemOperandSDNodes with generic
SDNodes now use MemIntrinsicSDNodes, with opcodes in a designated range
so that the SelectionDAG framework knows that MachineMemOperand information
is available.
llvm-svn: 82794
2009-09-25 20:36:54 +00:00
32f71d714b
Rename getTargetNode to getMachineNode, for consistency with the
...
naming scheme used in SelectionDAG, where there are multiple kinds
of "target" nodes, but "machine" nodes are nodes which represent
a MachineInstr.
llvm-svn: 82790
2009-09-25 18:54:59 +00:00
bea6848f9d
Finish scheduling itineraries for NEON.
...
llvm-svn: 82788
2009-09-25 18:38:29 +00:00
e6ff7662b6
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
a318d91a1e
Make sure sin, cos, sqrt calls are marked readonly
...
before producing FSIN, FCOS, FSQRT. If they aren't
so marked we have to assume they might set errno.
llvm-svn: 82781
2009-09-25 18:00:35 +00:00
c72134269f
Generate FSQRT from calls to the sqrt function, which
...
allows appropriate backends to generate a sqrt instruction.
On x86, this isn't done at -O0 because we go through
FastISel instead. This is a behavior change from before
this series of sqrt patches started. I think this is OK
considering that compile speed is most important at -O0, but
could be convinced otherwise.
llvm-svn: 82778
2009-09-25 17:23:22 +00:00
f7fe0132c9
Add some comments to clarify things that I discovered this week.
...
llvm-svn: 82773
2009-09-25 16:34:46 +00:00
d60367c198
pr4926: ARM requires the stack pointer to be aligned, even for leaf functions.
...
For the AAPCS ABI, SP must always be 4-byte aligned, and at any "public
interface" it must be 8-byte aligned. For the older ARM APCS ABI, the stack
alignment is just always 4 bytes. For X86, we currently align SP at
entry to a function (e.g., to 16 bytes for Darwin), but no stack alignment
is needed at other times, such as for a leaf function.
After discussing this with Dan, I decided to go with the approach of adding
a new "TransientStackAlignment" field to TargetFrameInfo. This value
specifies the stack alignment that must be maintained even in between calls.
It defaults to 1 except for ARM, where it is 4. (Some other targets may
also want to set this if they have similar stack requirements. It's not
currently required for PPC because it sets targetHandlesStackFrameRounding
and handles the alignment in target-specific code.) The existing StackAlignment
value specifies the alignment upon entry to a function, which is how we've
been using it anyway.
llvm-svn: 82767
2009-09-25 14:41:49 +00:00
18150d5abc
Fix combiner-aa issue with bases which are different, but can alias.
...
Previously, it treated GV+28 GV+0 as different bases, and assumed they could
not alias.
llvm-svn: 82753
2009-09-25 06:05:26 +00:00
28328db571
Don't try to use pre-indexed addressing with sthbrx/stwbrx
...
instructions. This fixes a PowerPC bug exposed by some unrelated
changes I'm working on.
llvm-svn: 82743
2009-09-25 00:57:30 +00:00
ebdfe4af62
Add a version of dumpr() that has a SelectionDAG* argument.
...
llvm-svn: 82742
2009-09-25 00:34:34 +00:00
372e9a389b
Start of revamping the register scavenging in PEI. ARM Thumb1 is the driving
...
interest for this, as it currently reserves a register rather than using
the scavenger for matierializing constants as needed.
Instead of scavenging registers on the fly while eliminating frame indices,
new virtual registers are created, and then a scavenged collectively in a
post-pass over the function. This isolates the bits that need to interact
with the scavenger, and sets the stage for more intelligent use, and reuse,
of scavenged registers.
For the time being, this is disabled by default. Once the bugs are worked out,
the current scavenging calls in replaceFrameIndices() will be removed and
the post-pass scavenging will be the default. Until then,
-enable-frame-index-scavenging enables the new code. Currently, only the
Thumb1 back end is set up to use it.
llvm-svn: 82734
2009-09-24 23:52:18 +00:00
944fa25934
Delete space after function name, before (, reflow a comment and
...
delete a few blank lines.
llvm-svn: 82729
2009-09-24 23:21:26 +00:00
d6f9a2f90b
Fix spacing.
...
llvm-svn: 82727
2009-09-24 23:11:08 +00:00
f08d2db928
add and document regex support for FileCheck. You can now do stuff like:
...
; CHECK: movl {{%e[a-z][xi]}}, %eax
or whatever.
llvm-svn: 82717
2009-09-24 21:47:32 +00:00
abab11abb2
unconditionally request MMI
...
llvm-svn: 82716
2009-09-24 21:45:57 +00:00
688f991b38
wrap long lines.
...
llvm-svn: 82715
2009-09-24 21:44:20 +00:00
bf97147a7e
Make the end-of-itinerary mark explicit. Some cleanup.
...
llvm-svn: 82709
2009-09-24 20:22:50 +00:00
1ce83629df
tidy up, fix a memory leak in Regex::isValid
...
llvm-svn: 82707
2009-09-24 20:15:51 +00:00
21bd8c9fc5
Constant propagating byval pointer is safe if function is readonly.
...
llvm-svn: 82700
2009-09-24 18:33:42 +00:00
46cd467310
Auto-upgrade malloc instructions to malloc calls.
...
Reviewed by Devang Patel.
llvm-svn: 82694
2009-09-24 17:47:49 +00:00
5fe313d6e0
Fix a hypothetical problem for targets with StackGrowsUp and a non-zero
...
LocalAreaOffset. (We don't have any of those right now.)
PEI::calculateFrameObjectOffsets includes the absolute value of the
LocalAreaOffset in the cumulative offset value used to calculate the
stack frame size. It then adds the raw value of the LocalAreaOffset
to the stack size. For a StackGrowsDown target, that raw value is negative
and has the effect of cancelling out the absolute value that was added
earlier, but that obviously won't work for a StackGrowsUp target. Change
to subtract the absolute value of the LocalAreaOffset.
llvm-svn: 82693
2009-09-24 16:42:27 +00:00
4e6f2e46e8
Move parent assertion check before metadata deletion.
...
llvm-svn: 82692
2009-09-24 16:19:11 +00:00
f95a450ef9
Don't constant propagate byval pointers, since they are not really pointers, but
...
rather structs passed by value.
This fixes PR5038.
llvm-svn: 82689
2009-09-24 09:47:18 +00:00
87d8f2b9d5
unconditionally compute MMI even if the target doesn't support EH or Debug info, because the target may use it for other things, this fixes PR5036
...
llvm-svn: 82684
2009-09-24 05:44:53 +00:00
4e37f877b5
reapply r82348 with a fix, thanks Jeffrey.
...
llvm-svn: 82683
2009-09-24 05:38:36 +00:00
26ea28eb5f
Fix PR5024 with a big hammer: disable the double-def assertion in the scavenger.
...
LiveVariables add implicit kills to correctly track partial register kills. This works well enough and is fairly accurate. But coalescer can make it impossible to maintain these markers. e.g.
BL <ga:sss1>, %R0<kill,undef>, %S0<kill>, %R0<imp-def>, %R1<imp-def,dead>, %R2<imp-def,dead>, %R3<imp-def,dead>, %R12<imp-def,dead>, %LR<imp-def,dead>, %D0<imp-def>, ...
...
%reg1031<def> = FLDS <cp#1>, 0, 14, %reg0, Mem:LD4[ConstantPool]
...
%S0<def> = FCPYS %reg1031<kill>, 14, %reg0, %D0<imp-use,kill>
When reg1031 and S0 are coalesced, the copy (FCPYS) will be eliminated the the implicit-kill of D0 is lost. In this case it's possible to move the marker to the FLDS. But in many cases, this is not possible. Suppose
%reg1031<def> = FOO <cp#1>, %D0<imp-def>
...
%S0<def> = FCPYS %reg1031<kill>, 14, %reg0, %D0<imp-use,kill>
When FCPYS goes away, the definition of S0 is the "FOO" instruction. However, transferring the D0 implicit-kill to FOO doesn't work since it is the def of D0 itself. We need to fix this in another time by introducing a "kill" pseudo instruction to track liveness.
Disabling the assertion is not ideal, but machine verifier is doing that job now. It's important to know double-def is not a miscomputation since it means a register should be free but it's not tracked as free. It's a performance issue instead.
llvm-svn: 82677
2009-09-24 02:27:09 +00:00
a21aac3811
Clean up LiveVariables and change how it deals with partial updates and kills. This also eliminate the horrible check which scan forward to the end of the basic block. It should be faster and more accurate.
...
llvm-svn: 82676
2009-09-24 02:15:22 +00:00
a75d6bf3a8
Roll back r82348, which introduced an infinite loop in ParseCStringVector() that
...
a trivial unittest would have caught. This revision also adds the trivial
unittest.
llvm-svn: 82675
2009-09-24 01:14:07 +00:00
fb1b55bc9c
A minor improvment in accuracy to inline cost
...
computation, and some cosmetics.
llvm-svn: 82660
2009-09-23 22:05:24 +00:00
afcaf79603
Checkpoint NEON scheduling itineraries.
...
llvm-svn: 82657
2009-09-23 21:38:08 +00:00
203d53ed79
Use getStoreSize() instead of getStoreSizeInBits()/8.
...
llvm-svn: 82656
2009-09-23 21:07:02 +00:00
08c0a95ac6
Rename several variables from EVT to more descriptive names, now that EVT
...
is also the name of their type, as declarations like "EVT EVT" look
really odd.
llvm-svn: 82654
2009-09-23 21:02:20 +00:00
75b9c2741d
Fix verification of explicit operands.
...
The machine code verifier did not check for explicit operands correctly. It
used MachineInstr::getNumExplicitOperands, but that method may cheat and use
the declared count in the TargetInstrDesc.
Now we check the explicit operands one at a time in visitMachineOperand.
llvm-svn: 82652
2009-09-23 20:57:55 +00:00
abfdbbfd1a
Remove BlackfinRegisterInfo::getFrameIndexOffset since it is the same as the
...
default implementation. Update comment on the default version, which made it
sound like most targets override it. Currently only X86 and SystemZ override
this method.
llvm-svn: 82651
2009-09-23 20:57:02 +00:00
52e819c7cf
Delete attached metadata when an instruction is deleted.
...
llvm-svn: 82647
2009-09-23 20:33:51 +00:00
29a44df5f8
ARM does not support offset folding (yet). Disable it for now.
...
This fixes PR5031. Unfortunately, there is no small testcase :(
llvm-svn: 82643
2009-09-23 19:04:09 +00:00
c2c86cede6
Edit a comment.
...
llvm-svn: 82641
2009-09-23 18:53:19 +00:00
add586526e
Do not leave behind metadata while cloning an instruction.
...
llvm-svn: 82638
2009-09-23 18:32:25 +00:00
a4c98a3e34
Fix bug in kill flag updating for post-register-allocation scheduling. When the kill flag of a superreg needs to be cleared because there are one or more subregs live, we instead add implicit-defs of those subregs and leave the kill flag on the superreg. This allows us to end the live-range of the superreg without ending the live-ranges of the subregs.
...
llvm-svn: 82629
2009-09-23 16:35:25 +00:00
43355aec61
Fix a struct/class mismatch.
...
llvm-svn: 82622
2009-09-23 11:48:57 +00:00
262f86ed90
Fix PR5024. LiveVariables physical register defs should *commit* only after all
...
of the defs are processed.
Also fix a implicit_def propagation bug: a implicit_def of a physical register
should be applied to uses of the sub-registers.
llvm-svn: 82616
2009-09-23 06:28:31 +00:00
b2abe07457
Fix a obvious logic error.
...
llvm-svn: 82610
2009-09-23 05:23:19 +00:00
8bcc6445c7
errorstr can be null, don't unconditionally set it. Only report that
...
"the jit has not been linked in" if the interpreter failed.
This fixes a unit test failure.
llvm-svn: 82601
2009-09-23 02:03:49 +00:00
41fa2bd112
Make EngineBuilder return more error codes, by KS Sreeram.
...
llvm-svn: 82600
2009-09-23 01:46:04 +00:00
c0353bfff5
Give MachineMemOperand an operator<<, factoring out code from
...
two different places for printing MachineMemOperands.
Drop the virtual from Value::dump and instead give Value a
protected virtual hook that can be overridden by subclasses
to implement custom printing. This lets printing be more
consistent, and simplifies printing of PseudoSourceValue
values.
llvm-svn: 82599
2009-09-23 01:33:16 +00:00
1439957928
Fix X86's unfoldMemoryOperand to properly handle MachineMemOperands.
...
llvm-svn: 82597
2009-09-23 01:29:41 +00:00
8808063181
This is overly constraining with respect to clang.
...
llvm-svn: 82591
2009-09-23 00:13:30 +00:00
ddc2ce4fe2
No need to verify that malloc's return type is i8*.
...
llvm-svn: 82561
2009-09-22 18:50:03 +00:00
74b79566f6
Use early returns.
...
llvm-svn: 82554
2009-09-22 16:47:52 +00:00
b2942cb5cb
... missed hiding a variable for MSVC only.
...
llvm-svn: 82552
2009-09-22 16:10:35 +00:00
1bdedd3876
Hide MSVC specific CRT interaction behind _MSC_VER.
...
llvm-svn: 82551
2009-09-22 15:58:35 +00:00
88e98fec70
Remove the GetProcessId() call from Win32/Program.inc, take 2.
...
GetProcessId() was introduced only in Windows XP, and we want to support earlier
versions.
llvm-svn: 82548
2009-09-22 15:40:32 +00:00
86707c9913
Fix commento.
...
llvm-svn: 82544
2009-09-22 10:30:34 +00:00
4c7b0ca05b
Add a magic LLVM_DISABLE_CRT_DEBUG environment variable which we check in RegisterHandler and use to disable the Win32 crash dialogs. These are a major blocker to any kind of automated testing.
...
Also, tweak the 'lit' test runner to set this variable unconditionally.
llvm-svn: 82537
2009-09-22 09:50:28 +00:00
08d1e41c10
Fix PR5024. LiveVariables::FindLastPartialDef should return a set of sub-registers that were defined by the last partial def, not just a single sub-register.
...
llvm-svn: 82535
2009-09-22 08:34:46 +00:00
1a0a423b38
Use Compiler.h macro instead of __attribute__.
...
llvm-svn: 82532
2009-09-22 07:38:23 +00:00
7b446a001d
Revert "Get rid of GetProcessId in Win32/Program.inc.", this breaks
...
ExecuteAndWait.
llvm-svn: 82522
2009-09-22 04:44:56 +00:00
78faee0664
Switch FoldingSet::AddString to StringRef based API.
...
- This also fixes a dereference of std::string::end, which makes MSVC unhappy and was causing all the static analyzer clang tests to fail.
llvm-svn: 82517
2009-09-22 03:34:53 +00:00
c827d9ee0e
Workaround what I believe is an MSVC bug where it emits a definition for a
...
static const class member into each translation unit, with external linkage???
- If someone understands this issue better, please clue me in, I haven't
consulted the standard yet.
llvm-svn: 82516
2009-09-22 03:34:40 +00:00
0356483539
Switch DIDescriptor to use a TrackingVH. - This makes it much safer to work with debug info, since it was extraordinarily easy to have dangling pointers thanks to MDNode uniquing.
...
llvm-svn: 82507
2009-09-22 02:03:18 +00:00
70d4fb0d97
Add a TrackingVH value handle.
...
This is designed for tracking a value even when it might move (like WeakVH), but it is an error to delete the referenced value (unlike WeakVH0. TrackingVH is templated like AssertingVH on the tracked Value subclass, it is an error to RAUW a tracked value to an incompatible type.
For implementation reasons the latter error is only diagnosed on accesses to a mis-RAUWed TrackingVH, because we don't want a virtual interface in a templated class.
The former error is also only diagnosed on access, so that clients are allowed to delete a tracked value, as long as they don't use it. This makes it easier for the client to reason about destruction.
llvm-svn: 82506
2009-09-22 02:02:33 +00:00
0dfed43a5b
Fix a pasto. Also simplify for Bill's benefit.
...
llvm-svn: 82505
2009-09-22 01:48:19 +00:00
8c500100f6
Minor bug fix. LowerSubregs should translate
...
%S0<def> = EXTRACT_SUBREG %Q0<kill>, 1
to
%S0<def> = IMPLICIT_DEF %Q0<imp-use,kill>
Implicit_def does not *read* any register so the operand should be marked "implicit". The missing "implicit" marker on the operand is wrong, but it doesn't actually break anything.
llvm-svn: 82503
2009-09-22 00:29:40 +00:00
cf295039e4
Fix PR5023: The instruction form of DominatorTree::dominates did not
...
take into consideration that the result of an invoke is only valid in
the normal dest, not the unwind dest. This caused 'PHINode::hasConstantValue'
to return true in an invalid situation, causing mem2reg to delete a phi that
was actually needed. This caused a crash building 483.xalancbmk.
llvm-svn: 82491
2009-09-21 22:39:35 +00:00
22151ce5e9
move DominatorTree::dominates for instructions out of line,
...
no functionality change.
llvm-svn: 82490
2009-09-21 22:30:50 +00:00
7d08da6b2a
tidy up
...
llvm-svn: 82489
2009-09-21 22:27:34 +00:00
e3ce1e2a37
tidy up
...
llvm-svn: 82488
2009-09-21 22:26:02 +00:00
255f416470
Clean up spill weight computation. Also some changes to give loop induction
...
variable increment / decrement slighter high priority.
This has major impact on some micro-benchmarks. On MultiSource/Applications
and spec tests, it's a minor win. It also reduce 256.bzip instruction count
by 8%, 55 on 164.gzip on i386 / Darwin.
llvm-svn: 82485
2009-09-21 21:12:25 +00:00
5090273367
Add Cortex-A8 VFP model.
...
llvm-svn: 82483
2009-09-21 20:52:17 +00:00
e7c8242baa
Change MachineMemOperand's alignment value to be the alignment of
...
the base pointer, without the offset. This matches MemSDNode's
new alignment behavior, and holds more interesting information.
llvm-svn: 82473
2009-09-21 19:47:04 +00:00
69499b13fd
Add support for rematerializing FsFLD0SS and FsFLD0SD as constant-pool
...
loads in order to reduce register pressure.
llvm-svn: 82470
2009-09-21 18:30:38 +00:00
48ade83e60
Recognize SSE min and max opportunities in even more cases.
...
And fix a bug with the behavior of min/max instructions formed from
fcmp uge comparisons.
Also, use FiniteOnlyFPMath() for this code instead of UnsafeFPMath,
as it is more specific.
llvm-svn: 82466
2009-09-21 18:03:22 +00:00
757acfea63
Fix the offset values for these memoperands. For frame objects, the
...
PseudoSourceValue already effectively represents the offset from the
frame base, so the actual offset should not be added to it.
llvm-svn: 82465
2009-09-21 17:58:09 +00:00
247053867e
big endian systems shift by bits too, hopefully this will fix the ppc
...
bootstrap problems.
llvm-svn: 82464
2009-09-21 17:55:47 +00:00
43d6830ea0
Nick pointed out that DominanceFrontier and DominanceTree are preserved
...
by setPreservesCFG().
llvm-svn: 82463
2009-09-21 17:54:42 +00:00
af57ae3da4
Remove the special-case for constants in PHI nodes; it's not really
...
helpful, and it didn't correctly handle the case of constants input
to PHIs for backedges.
llvm-svn: 82462
2009-09-21 17:53:35 +00:00
9045f235d2
fix PR5016, a crash I introduced in GVN handing first class
...
arrays and structs, which cannot be bitcast to integers.
llvm-svn: 82460
2009-09-21 17:24:04 +00:00
a3c45bda22
Fix this assertion string to mention subreg_to_reg.
...
llvm-svn: 82455
2009-09-21 15:18:33 +00:00
2c7e72cf42
initialize SymbolsCanStartWithDigit to false by default
...
llvm-svn: 82454
2009-09-21 14:11:56 +00:00
da99be6fae
Update cmake.
...
llvm-svn: 82449
2009-09-21 08:08:29 +00:00
f6eb7d83d1
Verify that phi instructions refer to MBBs in the CFG.
...
The machine code verifier no longer tolerates phi instructions with noop
operands. All MBBs on a phi instruction must be in the CFG.
llvm-svn: 82448
2009-09-21 07:19:08 +00:00
4d8af2f1ae
enable non-local analysis and PRE of large store -> little load.
...
This doesn't kick in too much because of phi translation issues,
but this can be resolved in the future.
llvm-svn: 82447
2009-09-21 06:48:08 +00:00
0cdc17eb50
convert an std::pair to an explicit struct.
...
llvm-svn: 82446
2009-09-21 06:30:24 +00:00
d28f90897a
move some functions, add a comment.
...
llvm-svn: 82444
2009-09-21 06:24:16 +00:00
9d7fb29522
split HandleLoadFromClobberingStore in two pieces: one that does the
...
analysis, one that does the xform.
llvm-svn: 82443
2009-09-21 06:22:46 +00:00
61858787b1
add a note
...
llvm-svn: 82442
2009-09-21 06:04:07 +00:00
bc528b16c2
Register the MachineModuleInfo for the ARM JIT, and update JITDwarfEmitter to
...
assert if the setModuleInfo hasn't been called.
llvm-svn: 82441
2009-09-21 05:58:35 +00:00
0a9616d906
Improve GVN to be able to forward substitute a small load
...
from a piece of a large store when both are in the same block.
This allows clang to compile the testcase in PR4216 to this code:
_test_bitfield:
movl 4(%esp), %eax
movl %eax, %ecx
andl $-65536, %ecx
orl $32962, %eax
andl $40186, %eax
orl %ecx, %eax
ret
This is not ideal, but is a whole lot better than the code produced
by llvm-gcc:
_test_bitfield:
movw $-32574, %ax
orw 4(%esp), %ax
andw $-25350, %ax
movw %ax, 4(%esp)
movw 7(%esp), %cx
shlw $8, %cx
movzbl 6(%esp), %edx
orw %cx, %dx
movzwl %dx, %ecx
shll $16, %ecx
movzwl %ax, %eax
orl %ecx, %eax
ret
and dramatically better than that produced by gcc 4.2:
_test_bitfield:
pushl %ebx
call L3
"L00000000001$pb":
L3:
popl %ebx
movl 8(%esp), %eax
leal 0(,%eax,4), %edx
sarb $7, %dl
movl %eax, %ecx
andl $7168, %ecx
andl $-7201, %ebx
movzbl %dl, %edx
andl $1, %edx
sall $5, %edx
orl %ecx, %ebx
orl %edx, %ebx
andl $24, %eax
andl $-58336, %ebx
orl %eax, %ebx
orl $32962, %ebx
movl %ebx, %eax
popl %ebx
ret
llvm-svn: 82439
2009-09-21 05:57:11 +00:00
fccbd0afc6
Fix PR4986. "r1024 = insert_subreg r1024, undef, 2" cannot be turned in an implicit_def. Instead, it's an identity copy so it should be eliminated. Also make sure to update livevariable kill information.
...
llvm-svn: 82436
2009-09-21 04:32:32 +00:00
9858859fd6
one case handled, expanded another testcase inline.
...
llvm-svn: 82427
2009-09-21 02:53:57 +00:00
1eefa9c427
formatting cleanups, no functionality change.
...
llvm-svn: 82426
2009-09-21 02:42:51 +00:00
9a10db8c46
Implement the JIT side of the GDB JIT debugging interface. To enable this
...
feature, either build the JIT in debug mode to enable it by default or pass
-jit-emit-debug to lli.
Right now, the only debug information that this communicates to GDB is call
frame information, since it's already being generated to support exceptions in
the JIT. Eventually, when DWARF generation isn't tied so tightly to AsmPrinter,
it will be easy to push that information to GDB through this interface.
Here's a step-by-step breakdown of how the feature works:
- The JIT generates the machine code and DWARF call frame info
(.eh_frame/.debug_frame) for a function into memory.
- The JIT copies that info into an in-memory ELF file with a symbol for the
function.
- The JIT creates a code entry pointing to the ELF buffer and adds it to a
linked list hanging off of a global descriptor at a special symbol that GDB
knows about.
- The JIT calls a function marked noinline that GDB knows about and has put an
internal breakpoint in.
- GDB catches the breakpoint and reads the global descriptor to look for new
code.
- When sees there is new code, it reads the ELF from the inferior's memory and
adds it to itself as an object file.
- The JIT continues, and the next time we stop the program, we are able to
produce a proper backtrace.
Consider running the following program through the JIT:
#include <stdio.h>
void baz(short z) {
long w = z + 1;
printf("%d, %x\n", w, *((int*)NULL)); // SEGFAULT here
}
void bar(short y) {
int z = y + 1;
baz(z);
}
void foo(char x) {
short y = x + 1;
bar(y);
}
int main(int argc, char** argv) {
char x = 1;
foo(x);
}
Here is a backtrace before this patch:
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x2aaaabdfbd10 (LWP 25476)]
0x00002aaaabe7d1a8 in ?? ()
(gdb) bt
#0 0x00002aaaabe7d1a8 in ?? ()
#1 0x0000000000000003 in ?? ()
#2 0x0000000000000004 in ?? ()
#3 0x00032aaaabe7cfd0 in ?? ()
#4 0x00002aaaabe7d12c in ?? ()
#5 0x00022aaa00000003 in ?? ()
#6 0x00002aaaabe7d0aa in ?? ()
#7 0x01000002abe7cff0 in ?? ()
#8 0x00002aaaabe7d02c in ?? ()
#9 0x0100000000000001 in ?? ()
#10 0x00000000014388e0 in ?? ()
#11 0x00007fff00000001 in ?? ()
#12 0x0000000000b870a2 in llvm::JIT::runFunction (this=0x1405b70,
F=0x14024e0, ArgValues=@0x7fffffffe050)
at /home/rnk/llvm-gdb/lib/ExecutionEngine/JIT/JIT.cpp:395
#13 0x0000000000baa4c5 in llvm::ExecutionEngine::runFunctionAsMain
(this=0x1405b70, Fn=0x14024e0, argv=@0x13f06f8, envp=0x7fffffffe3b0)
at /home/rnk/llvm-gdb/lib/ExecutionEngine/ExecutionEngine.cpp:377
#14 0x00000000007ebd52 in main (argc=2, argv=0x7fffffffe398,
envp=0x7fffffffe3b0) at /home/rnk/llvm-gdb/tools/lli/lli.cpp:208
And a backtrace after this patch:
Program received signal SIGSEGV, Segmentation fault.
0x00002aaaabe7d1a8 in baz ()
(gdb) bt
#0 0x00002aaaabe7d1a8 in baz ()
#1 0x00002aaaabe7d12c in bar ()
#2 0x00002aaaabe7d0aa in foo ()
#3 0x00002aaaabe7d02c in main ()
#4 0x0000000000b870a2 in llvm::JIT::runFunction (this=0x1405b70,
F=0x14024e0, ArgValues=...)
at /home/rnk/llvm-gdb/lib/ExecutionEngine/JIT/JIT.cpp:395
#5 0x0000000000baa4c5 in llvm::ExecutionEngine::runFunctionAsMain
(this=0x1405b70, Fn=0x14024e0, argv=..., envp=0x7fffffffe3c0)
at /home/rnk/llvm-gdb/lib/ExecutionEngine/ExecutionEngine.cpp:377
#6 0x00000000007ebd52 in main (argc=2, argv=0x7fffffffe3a8,
envp=0x7fffffffe3c0) at /home/rnk/llvm-gdb/tools/lli/lli.cpp:208
llvm-svn: 82418
2009-09-20 23:52:43 +00:00
5a3a854205
simplify as daniel suggests
...
llvm-svn: 82415
2009-09-20 22:56:43 +00:00
7e6d56ebc5
Revert r82404, it is causing a bootstrap miscompile. This is very very
...
scary, as it indicates a lurking bug. yay.
llvm-svn: 82411
2009-09-20 22:44:26 +00:00
eea16a168a
improve memdep to eliminate bitcasts (and aliases, and noop geps)
...
early for the stated reasons: this allows it to find more
equivalences and depend less on code layout.
llvm-svn: 82404
2009-09-20 21:00:18 +00:00
a0aa8fb6a6
Move CoerceAvailableValueToLoadType earlier in GVN.cpp. Hook it up
...
so that nonlocal and partially redundant loads can use it as well.
The testcase shows examples of craziness this can handle. This triggers
*many* times in 176.gcc.
llvm-svn: 82403
2009-09-20 20:09:34 +00:00
7c62d8a1a8
change the interface to CoerceAvailableValueToLoadType to be
...
more generic.
llvm-svn: 82402
2009-09-20 19:31:14 +00:00
1dd48c34e5
enhance GVN to forward substitute a stored value to a load
...
(and load -> load) when the base pointers must alias but when
they are different types. This occurs very very frequently in
176.gcc and other code that uses bitfields a lot.
llvm-svn: 82399
2009-09-20 19:03:47 +00:00
fc2d846dc8
update an entry, delete an entry which has been fixed.
...
llvm-svn: 82398
2009-09-20 17:37:38 +00:00
bb1a1bd2bd
tidy up
...
llvm-svn: 82397
2009-09-20 17:32:21 +00:00
692a3ea0b7
--- Reverse-merging r82282 into '.':
...
U lib/CodeGen/AsmPrinter/DwarfException.cpp
U lib/CodeGen/AsmPrinter/DwarfException.h
--- Reverse-merging r82274 into '.':
U lib/Target/TargetLoweringObjectFile.cpp
G lib/CodeGen/AsmPrinter/DwarfException.cpp
These revisions were breaking everything.
llvm-svn: 82396
2009-09-20 09:13:15 +00:00
13306a1fff
remove a temporary hack.
...
llvm-svn: 82395
2009-09-20 07:47:59 +00:00
5159bbaf9e
rename X86ATTAsmPrinter.cpp -> X86AsmPrinter.cpp likewise the .h file.
...
llvm-svn: 82394
2009-09-20 07:41:30 +00:00
40d9a5319a
move target registry stuff to X86ATTAsmPrinter.cpp
...
llvm-svn: 82393
2009-09-20 07:39:06 +00:00
b95a9cd6a7
simplify this now that createX86CodePrinterPass is trivial
...
llvm-svn: 82392
2009-09-20 07:37:51 +00:00
288a95da0c
rename X86ATTAsmPrinter class -> X86AsmPrinter
...
llvm-svn: 82391
2009-09-20 07:35:34 +00:00
00b4508bb0
remove the asmstring, it is now dead. Improve comment.
...
llvm-svn: 82390
2009-09-20 07:32:00 +00:00
9b3ed87506
Peer through zext and sext to eliminate them when it is safe to do so.
...
llvm-svn: 82389
2009-09-20 07:31:25 +00:00
a93dcf1bc0
kill off printPICLabel now, it's specialness is handled by
...
the MachineInstr ->MCInst lowering process, not in the
asmprinter.
llvm-svn: 82388
2009-09-20 07:28:26 +00:00
609fbed49f
delete X86IntelAsmPrinter! Now -x86-asm-syntax just switches
...
the instruction syntax, not the entire asmprinter.
llvm-svn: 82387
2009-09-20 07:25:17 +00:00
b0225ba289
Fold 'icmp eq (icmp), true' into an xor(icmp).
...
llvm-svn: 82386
2009-09-20 07:21:39 +00:00
4479034537
Add an intel syntax MCInstPrinter implementation. You can now
...
transcode from AT&T to intel syntax with "llvm-mc foo.s -output-asm-variant=1"
llvm-svn: 82385
2009-09-20 07:17:49 +00:00
78f908c877
tidy up
...
llvm-svn: 82384
2009-09-20 07:16:54 +00:00
a15f0044a0
eliminate a use of strtoul.
...
llvm-svn: 82382
2009-09-20 06:58:54 +00:00
17ec6b11d2
split random COFF asmprinter state out to X86COFFMachineModuleInfo.h.
...
Make dllexport directives come out in determinstic order.
llvm-svn: 82381
2009-09-20 06:45:52 +00:00
ff550aa36d
Correct the comment; this applies to fcmp too.
...
llvm-svn: 82380
2009-09-20 06:27:35 +00:00
0f8348e85b
Remove tab, again.
...
llvm-svn: 82379
2009-09-20 06:26:34 +00:00
28260409f2
Teach the constant folder how to not a cmpinst.
...
llvm-svn: 82378
2009-09-20 06:24:51 +00:00
56efff07f5
smallvectorize getExtraOptionNames
...
llvm-svn: 82377
2009-09-20 06:21:43 +00:00
131dca9c48
minor cleanups.
...
llvm-svn: 82375
2009-09-20 06:18:38 +00:00
00f53807c3
strength reduce further StringRef-> const char*, saving another 620 bytes.
...
llvm-svn: 82372
2009-09-20 05:53:47 +00:00
4a03452077
Try turning icmp(bitcast(x), bitcast(y)) into icmp(bitcast(bitcast(x)), y) in
...
the hopes that the two bitcasts will merge.
llvm-svn: 82371
2009-09-20 05:48:50 +00:00
b1f2e101db
switch an std::string to StringRef, shaving 400 bytes off CommandLine.o
...
llvm-svn: 82370
2009-09-20 05:48:01 +00:00
2b31b53d97
Remove tabs I added.
...
llvm-svn: 82369
2009-09-20 05:47:45 +00:00
6ec8caf003
the switch from std::map -> StringMap caused --help output to be in
...
non-sorted order, restore the sort.
llvm-svn: 82368
2009-09-20 05:37:24 +00:00
8d0309daa0
eliminate the duplicate detection loop, moving it into the loop that populates the Opts vector in the first place.
...
llvm-svn: 82367
2009-09-20 05:22:52 +00:00
64dbb5ca5a
Eliminate a masochistic "algorithm" loop, shrinking CommandLine.o from 71524->70700 bytes.
...
llvm-svn: 82366
2009-09-20 05:18:28 +00:00
28610b9878
don't use count + insert, just do insert + failure. Also, instead of deleting from
...
the middle of a vector, swap the last element in and pop_back. Also saves 330 bytes :)
llvm-svn: 82365
2009-09-20 05:15:12 +00:00
41f8b0b7a6
switch to SmallPtrSet instead of std::set, saving 1K from the
...
release-asserts .o file (72900->71856).
llvm-svn: 82364
2009-09-20 05:12:14 +00:00
f74e28abfa
change an std::sort to an array_pod_sort call, shrinking CommandLine.o by 9%.
...
llvm-svn: 82363
2009-09-20 05:06:23 +00:00
e7c1e210c7
Several changes together in a murky mess:
...
1. Change some "\n" -> '\n'.
2. eliminte some std::string's by using raw_ostream::indent.
3. move a bunch of code out of the main arg parser routine into
a new static HandlePrefixedOrGroupedOption function.
4. Greatly simplify the implementation of getOptionPred, and make
it avoid splitting prefix options at = when that doesn't match
a non-prefix option.
llvm-svn: 82362
2009-09-20 05:03:30 +00:00
9e085545f8
Clean up the usage of evaluateICmpRelation's return value.
...
Add another line to the ConstantExprFold test to demonstrate the GEPs may not
wrap around in either the signed or unsigned senses.
llvm-svn: 82361
2009-09-20 04:27:06 +00:00
ecbb126e34
Fix refacto, this code was expecting to stride past the argument prefix.
...
llvm-svn: 82360
2009-09-20 04:03:41 +00:00
6058b51f8c
Strip trailing whitespace.
...
llvm-svn: 82359
2009-09-20 04:03:34 +00:00
2949a2398c
Remove dead store by taking a guess at what Chris meant. I wasn't able to
...
design a testcase that would tickle this behaviour.
llvm-svn: 82357
2009-09-20 03:48:46 +00:00
f878d70720
Still one more thing wrong here...
...
llvm-svn: 82356
2009-09-20 02:27:06 +00:00
7d6781b0fe
Tabs -> spaces, and remove trailing whitespace.
...
llvm-svn: 82355
2009-09-20 02:20:51 +00:00
0f899601f3
Here's fun! It turns out that these filter functions can be internal. If they're
...
internal, they shouldn't use the indirect pointer stuff. In the case of
throw_rethrow_test, it was marked as 'internal' and calculated its own offset to
its contents.
llvm-svn: 82354
2009-09-20 02:19:49 +00:00
595b3dfcbe
Delete dead code. sext and zext can not turn integers into pointers. Further,
...
the optimization described in the comment is only valid with target data.
llvm-svn: 82353
2009-09-20 02:11:47 +00:00
5a3fa4ef33
convert argname to StringRef, simplifying LookupOption.
...
llvm-svn: 82352
2009-09-20 02:02:24 +00:00
0a40a975dd
convert 'Value' to StringRef which makes it easier to
...
maintain the "null is unspecified, empty is empty" semantics.
llvm-svn: 82351
2009-09-20 01:53:12 +00:00
40fef8032e
Change CommaSeparated processing to do it with StringRef instead of temporary std::strings.
...
This requires StringRef'izing ProvideOption which I also did.
llvm-svn: 82350
2009-09-20 01:49:31 +00:00
e0332983fd
Value* were never meant to be const. Removing constness from the constant
...
folder removes a lot of const_casting and requires no changes to clang or
llvm-gcc.
llvm-svn: 82349
2009-09-20 01:35:59 +00:00
77c2724360
rewrite ParseCStringVector in terms of stringref.
...
llvm-svn: 82348
2009-09-20 01:33:46 +00:00
372a8ae403
move a couple non-trivial methods out of line, add new
...
find_first_of/find_first_of methods.
llvm-svn: 82347
2009-09-20 01:22:16 +00:00
1b88fbdaa3
coding style cleanup
...
llvm-svn: 82346
2009-09-20 01:11:23 +00:00
3b8adaf488
convert a bunch more stuff to use StringRef. The ArgName arguments are now
...
stringref because they may not be nul terminated. For options like -Lfoo
this now avoids a O(n) temporary std::strings where N is the length of
the string after -L.
llvm-svn: 82345
2009-09-20 00:40:49 +00:00
a894053a9b
When computing live intervals for earlyclobber operands,
...
we pushed the beginning of the interval back 1, so the
interval would overlap with inputs that die. We were
also pushing the end of the interval back 1, though,
which means the earlyclobber didn't overlap with other
output operands. Don't do this. PR 4964.
llvm-svn: 82342
2009-09-20 00:36:41 +00:00
ca2552d9f9
avoid a bunch of malloc thrashing for PositinoalVals by eliminating
...
a std::vector and a bunch of std::string temporaries.
llvm-svn: 82341
2009-09-20 00:07:40 +00:00
605109d151
Teach the constant folder how to handle a few simple i1 cases.
...
llvm-svn: 82340
2009-09-20 00:04:02 +00:00
fa9c6f43a0
Avoid some temporary strings.
...
llvm-svn: 82339
2009-09-19 23:59:02 +00:00
84c1527b6b
add some more overloads of StringRef::getAsInteger for
...
common and useful integer types.
llvm-svn: 82338
2009-09-19 23:58:48 +00:00
85689b2065
Revert r82274. It's causing failures in the CINT2006 benchmarks.
...
llvm-svn: 82336
2009-09-19 22:02:37 +00:00
be22ec4cfc
Fix indentation.
...
llvm-svn: 82333
2009-09-19 20:40:14 +00:00
c418d6b106
Strip trailing whitespace.
...
llvm-svn: 82332
2009-09-19 20:40:05 +00:00
1303c0ab86
Remove the default value for ConstantStruct::get's isPacked parameter and
...
update the code which was broken by this.
llvm-svn: 82327
2009-09-19 20:30:26 +00:00
68ee70035e
provide a "strtoull" operation that works on StringRef's.
...
llvm-svn: 82322
2009-09-19 19:47:14 +00:00
3b8bd05081
Add a comment explaining why you would ever want to do this.
...
llvm-svn: 82319
2009-09-19 19:00:06 +00:00
aecd74d895
convert a bunch of std::strings to use StringRef. This should eliminate
...
a massive number of temporary strings created when parsing a command line.
More still left to eliminate.
llvm-svn: 82318
2009-09-19 18:55:05 +00:00
7e6deb1cb4
Lett users of sparse propagation do their own thing with phi nodes if they want
...
to. This can be combined with LCSSA or SSI form to store more information on a
PHINode than can be computed by looking at its incoming values.
llvm-svn: 82317
2009-09-19 18:33:36 +00:00
1636b7ef47
The flag "--dot-cfg-only" is at the moment equivalent to the flag "--dot-cfg".
...
It prints the content of all bbs, instead of printing empty bbs to make the
CFG more readable. Fix this. Patch by Tobias Grosser.
llvm-svn: 82315
2009-09-19 11:25:44 +00:00
b82b5514fe
Fix funky comments.
...
llvm-svn: 82314
2009-09-19 10:09:15 +00:00
543d9b2fba
Try to speed up the slowest parts of the CommandLine library
...
- Replace std::map<std::string with StringMap
- Eliminate unnecessary std::string copies
- ~10% speed-up for clang's testsuite on my machine (debug build)
llvm-svn: 82312
2009-09-19 10:01:45 +00:00
9827ad39a7
Fix PR4926. When target hook EmitInstrWithCustomInserter() insert new basic blocks and update CFG, it should also inform sdisel of the changes so the phi source operands will come from the right basic blocks.
...
llvm-svn: 82311
2009-09-19 09:51:03 +00:00
5d034499ad
Enhance transform passes so that they apply the same tranforms to malloc calls as to MallocInst.
...
Reviewed by Dan Gohman.
llvm-svn: 82300
2009-09-18 22:35:49 +00:00
8c33d67fbf
Fix a comment typo and some whitespace.
...
llvm-svn: 82285
2009-09-18 21:43:11 +00:00
3c21a35908
Fix a typo in an assertion message.
...
llvm-svn: 82284
2009-09-18 21:42:44 +00:00
053237109a
Factor out label difference creation.
...
llvm-svn: 82282
2009-09-18 21:37:56 +00:00
537d8d99be
Enhance analysis passes so that they apply the same analysis to malloc calls as to MallocInst.
...
Reviewed by Eli Friedman.
llvm-svn: 82281
2009-09-18 21:34:51 +00:00
43f2cd7757
It's inefficient to have place the exception tables (which contain the LSDA)
...
into the __DATA section. At launch time, dyld has to update most of the section
to fix up the type info pointers. It's better to place it into the __TEXT
section and use pc-rel indirect pointer encodings. Similar to the personality
routine.
llvm-svn: 82274
2009-09-18 21:14:36 +00:00
270d0f986f
Enhance EmitInstrWithCustomInserter() so target can specify CFG changes that sdisel will use to properly complete phi nodes.
...
Not functionality change yet.
llvm-svn: 82273
2009-09-18 21:02:19 +00:00
7ba874bb5e
Fix cmake build, which has a different -I that
...
causes the "../foo" to not find the file
llvm-svn: 82270
2009-09-18 20:35:59 +00:00
3a78ce3a56
Make a new X8632_MachoTargetObjectFile TLOF implementation whose
...
getSymbolForDwarfGlobalReference is smart enough to know that it
needs to register the stub it references with MachineModuleInfoMachO,
so that it gets emitted at the end of the file.
Move stub emission from X86ATTAsmPrinter::doFinalization to the
new X86ATTAsmPrinter::EmitEndOfAsmFile asmprinter hook. The important
thing here is that EmitEndOfAsmFile is called *after* the ehframes are
emitted, so we get all the stubs.
This allows us to remove a gross hack from the asmprinter where it would
"just know" that it needed to output stubs for personality functions.
Now this is all driven from a consistent interface.
The testcase change is just reordering the expected output now that the
stubs come out after the ehframe instead of before.
This also unblocks other changes that Bill wants to make.
llvm-svn: 82269
2009-09-18 20:22:52 +00:00
71a15b1316
add a new hook to allow targets to splat stuff at the end of the file.
...
Overriding doFinalization is pretty lame.
llvm-svn: 82268
2009-09-18 20:17:03 +00:00
5e9a5c3664
Model the carry bit on ppc32. Without this we could
...
move a SUBFC (etc.) below the SUBFE (etc.) that consumed
the carry bit. Add missing ADDIC8, noticed along the way.
llvm-svn: 82266
2009-09-18 20:15:22 +00:00
722b1eefdb
Add support for using the FLAGS result of or, xor, and and instructions
...
on x86, to avoid explicit test instructions. A few existing tests changed
due to arbitrary register allocation differences.
llvm-svn: 82263
2009-09-18 19:59:53 +00:00
8e31aa773f
Added RCL and RCR (rotate left and right with a
...
carry bit) instructions to the Intel instruction
tables.
llvm-svn: 82260
2009-09-18 19:35:23 +00:00
af206b8c88
Write and read metadata attachments.
...
llvm-svn: 82259
2009-09-18 19:26:43 +00:00
788eaabd18
Update malloc call creation code (AllocType is now the element type of the malloc, not the resulting type).
...
In getMallocArraySize(), fix bug in the case that array size is the product of 2 constants.
Extend isArrayMalloc() and getMallocArraySize() to handle case where malloc is used as char array.
Ensure that ArraySize in LowerAllocations::runOnBasicBlock() is correct type.
Extend Instruction::isSafeToSpeculativelyExecute() to handle malloc calls.
Add verification for malloc calls.
Reviewed by Dan Gohman.
llvm-svn: 82257
2009-09-18 19:20:02 +00:00
e133923abe
duncan points out the EH selector values are signed.
...
llvm-svn: 82245
2009-09-18 18:34:29 +00:00
1e64038bcb
This file can need access to the X86 instruction enums when the table exceeds 32-bits.
...
llvm-svn: 82235
2009-09-18 18:08:55 +00:00
592638ae05
Allow symbols to start from the digit if target requests it. This allows, e.g. pinning
...
variables to specified absolute address. Make use of this feature for MSP430.
This unbreaks PR4776.
llvm-svn: 82227
2009-09-18 16:57:42 +00:00
bf4c56d82b
Stop using alloca.
...
llvm-svn: 82225
2009-09-18 16:46:16 +00:00
f4db6396e0
Revert r82214. It broke 403.gcc on x86_64 / Darwin.
...
llvm-svn: 82215
2009-09-18 08:26:06 +00:00
6ba1931d60
Fix a bug in sdisel switch lowering code. When it updates the phi nodes in switch successor blocks, it can introduce multiple phi operands of the same value from different blocks (and may not be on the predecessor list).
...
This can be seen on CodeGen/Generic/2006-09-06-SwitchLowering.ll. But it's not known to cause any real regression (but I have added an assertion for it now).
llvm-svn: 82214
2009-09-18 08:16:04 +00:00
6a3260e004
Add newlines.
...
llvm-svn: 82206
2009-09-18 07:36:47 +00:00
1bd81314e7
tolerate llvm.eh.selector.i64 on 32-bit systems and llvm.eh.selector.i32 on
...
64-bit systems.
llvm-svn: 82180
2009-09-17 23:54:54 +00:00
ea8a4b984c
Fix parsing of optional metadata for 'load', 'store' and 'alloc' instructions.
...
llvm-svn: 82175
2009-09-17 23:04:48 +00:00
a6ebba270d
pass machinemoduleinfo down into getSymbolForDwarfGlobalReference,
...
currently unused.
llvm-svn: 82157
2009-09-17 18:49:52 +00:00
36bad00bef
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
9632c14949
grammar
...
llvm-svn: 82150
2009-09-17 17:57:26 +00:00
5ff90ed33e
Initialize HasMetadata to zero.
...
llvm-svn: 82145
2009-09-17 14:51:57 +00:00
4ba30d9913
Remove simple regalloc. It has bit rotted.
...
llvm-svn: 82127
2009-09-17 05:48:07 +00:00
ac6271e3f4
add a version of the APFloat constructor that initializes to 0.0
...
llvm-svn: 82110
2009-09-17 01:08:43 +00:00
f56b0482c4
Fix PR4910: Broken logic in coalescer means when a physical register liveness is being shortened, the sub-registers were not. The symptom is the register allocator could not find a free register for this particular test.
...
llvm-svn: 82108
2009-09-17 00:57:15 +00:00
c7012fa1e8
Some platforms may need malloc.h for alloca.
...
llvm-svn: 82100
2009-09-17 00:14:44 +00:00
487d1c8138
Update CMake.
...
llvm-svn: 82097
2009-09-17 00:06:48 +00:00
cc774e7476
Added the LODS (load byte into register, usually
...
as part string parsing) instructions to the Intel
instruction tables.
llvm-svn: 82089
2009-09-16 22:59:28 +00:00
4498168753
Add StringRef::{rfind, rsplit}
...
llvm-svn: 82087
2009-09-16 22:38:48 +00:00
73645ac10c
Added the LAR (load segment access rights)
...
instructions to the Intel instruction tables.
llvm-svn: 82084
2009-09-16 21:55:34 +00:00
7865f79011
Added the LOOP family of instructions to the Intel
...
instruction tables.
llvm-svn: 82083
2009-09-16 21:50:07 +00:00
ca968cf1f0
Added an alternate form of register-register CMP
...
to the Intel instruction tables.
llvm-svn: 82081
2009-09-16 21:11:23 +00:00
44b3a87f78
Fix typo.
...
llvm-svn: 82080
2009-09-16 21:09:07 +00:00
852c9b6627
At iSel time, update DebugLoc based on debug info attached with an instruction.
...
llvm-svn: 82077
2009-09-16 20:39:11 +00:00
0f64d71d99
Add a new pass for doing late hoisting of floating-point and vector
...
constants out of loops. These aren't covered by the regular LICM
pass, because in LLVM IR constants don't require separate
instructions. They're not always covered by the MachineLICM pass
either, because it doesn't know how to unfold folded constant-pool
loads. This is somewhat experimental at this point, and off by
default.
llvm-svn: 82076
2009-09-16 20:25:11 +00:00
dec23fd825
Print debug info attached with an instruction.
...
llvm-svn: 82075
2009-09-16 20:21:17 +00:00
5d8cfb217c
Expand vector floating-point conversions not supported by NEON.
...
llvm-svn: 82074
2009-09-16 20:20:44 +00:00
820640d39f
Provide a way to extract location info from DILocation.
...
llvm-svn: 82064
2009-09-16 18:20:05 +00:00
0d9950e315
Parse debug info attached with an instruction.
...
llvm-svn: 82063
2009-09-16 18:18:06 +00:00
d5497a4bf8
Add llvm::Metadata to manage metadata used in a context.
...
This interface will be used to attach metadata with an instruction.
llvm-svn: 82060
2009-09-16 18:09:00 +00:00
c0edda3184
Fix incorrect assert that should be a user error for code like 'mov $0, %%eax'.
...
llvm-svn: 82054
2009-09-16 17:18:29 +00:00
bd0050810c
Change FoldPHIArgBinOpIntoPHI to decline folding if it would introduce two
...
phis, similar to the FoldPHIArgGEPIntoPHI change.
Also, delete some comments that don't reflect the code.
llvm-svn: 82053
2009-09-16 16:50:24 +00:00
6f3d4e900b
Don't sort the vector when it is empty. This should fix some expensive checking
...
failures.
llvm-svn: 82040
2009-09-16 11:43:12 +00:00
41c1103273
Reapplied r81355 with the problems fixed.
...
(See http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090907/086737.html and
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090907/086746.html )
llvm-svn: 82039
2009-09-16 11:35:50 +00:00
230c717e0b
updated lib/CodeGen/CMakeLists.txt to unbreak cmake build after r82018
...
llvm-svn: 82038
2009-09-16 10:18:36 +00:00
f8cb758ba8
Preserve ProfileInfo during CodeGenPrepare.
...
llvm-svn: 82034
2009-09-16 09:26:52 +00:00
05f4039763
move FnStubs/GVSTubs/HiddenGVStub handling out of the X86 asmprinter
...
and use MachineModuleInfoMachO instead.
llvm-svn: 82022
2009-09-16 06:25:03 +00:00
3828c2fa80
revert a hunk of r82018 that wasn't supposed to go in yet.
...
llvm-svn: 82020
2009-09-16 06:04:53 +00:00
f3f54ffc89
add a new MachineModuleInfoMachO class, which is the per-module
...
stuff common across all macho targets.
llvm-svn: 82018
2009-09-16 06:03:48 +00:00
07fc26d63c
the pointer MMI keeps will start out with object-file format specific stuff
...
llvm-svn: 82012
2009-09-16 05:26:00 +00:00
273215733c
tidy up
...
llvm-svn: 82011
2009-09-16 05:25:43 +00:00
46dcaadb4a
rearrange X86ATTAsmPrinter::doFinalization, making a scan of
...
the global variable list only happen for COFF targets.
llvm-svn: 82010
2009-09-16 05:20:33 +00:00
8df79f2681
remove the AsmPrinter::printMCInst hook hack now that
...
we have MCInstPrinter.
llvm-svn: 82006
2009-09-16 04:57:15 +00:00
b15840a6d6
fix cmake build
...
llvm-svn: 81999
2009-09-16 04:44:00 +00:00
1ae49ee7ca
Do not try and sink a load whose chain result has more than one use, when
...
trying to create RMW opportunities in the x86 backend. This can cause a
cycle to appear in the graph, since the other uses may eventually feed into
the TokenFactor we are sinking the load below.
llvm-svn: 81996
2009-09-16 03:20:46 +00:00
771a1f1720
Added the ENTER instruction, which sets up a stack
...
frame, to the Intel instruction tables.
llvm-svn: 81995
2009-09-16 02:57:13 +00:00
0fdee53266
Added the definitions for one-bit left shifts to
...
the Intel instruction tables.
The patterns will stay blank because ADD reg, reg
is faster, but having the encoding available is
useful for the disassembler.
llvm-svn: 81994
2009-09-16 02:28:43 +00:00
3b7ce109ec
Don't sink gep operators through phi nodes if the result would require
...
more than one phi, since that leads to higher register pressure on
entry to the phi. This is especially problematic when the phi is in
a loop header, as it increases register pressure throughout the loop.
llvm-svn: 81993
2009-09-16 02:01:52 +00:00
1c28c4db58
Removed a few instructions that were already
...
covered by other definitions.
llvm-svn: 81992
2009-09-16 01:54:38 +00:00
b866602f06
Big change #1 for personality function references:
...
Eliminate the PersonalityPrefix/Suffix & NeedsIndirectEncoding
fields from MAI: they aren't part of the asm syntax, they are
related to the structure of the object file.
To replace their functionality, add a new
TLOF::getSymbolForDwarfGlobalReference method which asks targets
to decide how to reference a global from EH in a pc-relative way.
The default implementation just returns the symbol. The default
darwin implementation references the symbol through an indirect
$non_lazy_ptr stub. The bizarro x86-64 darwin specialization
handles the weird "foo@GOTPCREL+4" hack.
DwarfException.cpp now uses this to emit the reference to the
symbol in the right way, and this also eliminates another
horrible hack from DwarfException.cpp:
- if (strcmp(MAI->getPersonalitySuffix(), "+4@GOTPCREL"))
- O << "-" << MAI->getPCSymbol();
llvm-svn: 81991
2009-09-16 01:46:41 +00:00
0d3f6962a1
remove a dead variable.
...
llvm-svn: 81985
2009-09-16 01:29:11 +00:00
6a833f6806
add a helper method for creating MCSymbol and MCSymbolRefExpr at
...
the same time.
llvm-svn: 81984
2009-09-16 01:26:31 +00:00
e739ac8900
Added a variety of floating-point and SSE instructions.
...
All of these do not have patterns (they're for the
disassembler).
Many of the floating-point instructions will probably
be rolled into definitions that have patterns, and may
eventually be superseded by mdefs. So I put them
together and left a comment.
llvm-svn: 81979
2009-09-16 01:13:52 +00:00
6b99ae882b
inline AsmPrinter::getCurrentFunctionEHName into its only caller.
...
llvm-svn: 81970
2009-09-16 00:35:39 +00:00
6cc46577f4
Expand some more vector operations not supported by Neon.
...
llvm-svn: 81969
2009-09-16 00:32:15 +00:00
d106abfce7
Eliminate AsmPrinter::EmitExternalGlobal, inlining its (now)
...
one implementation into its one caller. This eliminates a totally
awesome and gratuitous hack where we casted a Function* to
GlobalVariable*.
llvm-svn: 81967
2009-09-16 00:17:39 +00:00
4ed397c141
Neon does not support vector divide or remainder. Expand them.
...
llvm-svn: 81966
2009-09-16 00:17:28 +00:00
4e465598f0
eliminate the PPC backend's implementation of EmitExternalGlobal
...
and use PersonalityPrefix/Suffix to achieve the same effect (like
the x86 backend).
This changes the code generated for ppc static mode, but guess what,
we were generating this before:
.byte 0x9B ; Personality (indirect pcrel sdata4)
.long ___gxx_personality_v0-. ; Personality
which is not correct! (it is not an 'indirect' reference).
llvm-svn: 81965
2009-09-16 00:14:19 +00:00
dd5a989034
eliminate the horrid AsmPrinter::getGlobalLinkName method, inlining
...
it into all of its call sites and simplifying them.
llvm-svn: 81962
2009-09-16 00:08:41 +00:00
3045f9e57b
simplify some code
...
llvm-svn: 81961
2009-09-16 00:08:07 +00:00
194a2518e5
Expand all v2f64 arithmetic operations for Neon.
...
Radar 7200803. (This should also fix the
SingleSource/UnitTests/Vector/sumarray-dbl test.)
llvm-svn: 81959
2009-09-15 23:55:57 +00:00
f8f6479388
Added far return instructions (that is, returns to
...
code in other segments) to the Intel instruction
tables.
llvm-svn: 81953
2009-09-15 23:37:51 +00:00
2251293788
remove some horrible MAI hooks which fortunately turn out to be always empty.
...
llvm-svn: 81946
2009-09-15 23:11:32 +00:00
7b40df7589
strength reduce a call to PrintRelDirective(true).
...
llvm-svn: 81942
2009-09-15 22:58:35 +00:00
f2471ec967
add hooks to hang target-specific goop off MachineModuleInfo,
...
move MachineFunctionInfo virtual method out of line to give it
a home.
llvm-svn: 81940
2009-09-15 22:44:26 +00:00
fbb88b180c
Do not add the SVOffset to the Node CSE ID. The same pointer argument cannot have different
...
SVOffsets.
llvm-svn: 81937
2009-09-15 22:30:11 +00:00
17f54d0f42
Expand on comment.
...
llvm-svn: 81928
2009-09-15 21:56:46 +00:00
7a27b9342c
Updated comments per Eli's suggestion.
...
llvm-svn: 81923
2009-09-15 21:43:27 +00:00
79b570de9d
Added register-to-register ADD instructions to the
...
Intel tables, where the source operand is
specified by the R/M field and the destination
operand by the Reg field.
llvm-svn: 81914
2009-09-15 20:53:57 +00:00
316b4a0206
Drop the raw_ostream required buffer size to 1.
...
- As best I can tell, we have eliminated all the code which used to require a
larger buffer size.
llvm-svn: 81912
2009-09-15 20:31:46 +00:00
9367ec8644
Fix -Asserts warning.
...
llvm-svn: 81909
2009-09-15 20:31:12 +00:00
178135c88b
Better solution for tracking both the original alignment of the access, and the current alignment based
...
on the source value offset. This avoids increasing the size of mem nodes.
llvm-svn: 81897
2009-09-15 19:05:41 +00:00
e8ea876cc3
Added a new register class for segment registers
...
to the Intel register table.
Added 16- and 64-bit MOVs to and from the segment
registers to the Intel instruction tables.
llvm-svn: 81895
2009-09-15 18:47:29 +00:00
c092b76d63
Change the marker byte for stubs from 0xcd to 0xce (another form of
...
interrupt instruction, which shouldn't arise any other way). 0xcd is
also used by JITMemoryManager to initialize the buffer to garbage,
which means it could appear following a noreturn call even when
that is not a stub, confusing X86CompilationCallback2. PR 4929.
llvm-svn: 81888
2009-09-15 18:32:14 +00:00
db4916a123
fix PR4984 by ensuring that fastisel adds properly sign extended GEP displacement
...
values to machineinstrs.
llvm-svn: 81886
2009-09-15 18:27:02 +00:00
7ff4a94c4b
add missing file
...
llvm-svn: 81881
2009-09-15 18:03:13 +00:00
967bf27de2
Handle AddrMode4 for Thumb2 in rewriteT2FrameIndex. This occurs for
...
VLDM/VSTM instructions, and without this check, the code assumes that an
offset is allowed, as it would be with VLDR/VSTR. The asm printer,
however, silently drops the offset, producing incorrect code. Since the
address register in this case is either the stack or frame pointer, the
spill location ends up conflicting with some other stack slot or with
outgoing arguments on the stack.
llvm-svn: 81879
2009-09-15 17:56:18 +00:00
f3369c22a7
Fix superreg use in ARMAsmPrinter. Approved by Anton Korobeynikov.
...
llvm-svn: 81878
2009-09-15 17:53:11 +00:00
840c700654
several major improvements to the sparc backend: support for weak linkage
...
and PIC codegen. Patch by Venkatraman Govindaraju!
llvm-svn: 81877
2009-09-15 17:46:24 +00:00
94262db4d4
Teach ValueTracking how to look through GlobalAliases. GlobalAliases are
...
not folded in the constant folder because the constant folder doesn't
simplify ConstantExpr operands.
llvm-svn: 81864
2009-09-15 16:14:44 +00:00
0ea0e1b4af
Fix an accidental inversion of the inbounds flag.
...
llvm-svn: 81862
2009-09-15 16:00:30 +00:00
e4532f30b1
When a constant's type is refined, update the constant in place
...
instead of cloning and RAUWing it.
- Make AbstractTypeUser a friend of Value so that it can offer
its subclasses a way to update a Value's type in place. This
is better than a universally visible setType method on Value,
and it's sufficient for the immediate need.
- Eliminate the constant "convert" functions. This eliminates a
lot of logic duplication, and fixes a complicated bug where a
constant can't actually be cloned during the type refinement
process because some of the types that its folder needs are
half-destroyed, being in the middle of refinement themselves.
- Move the getValType functions from being static overloaded
functions in Constants.cpp to be members of class template
specializations in ConstantsContext.h. This means that the
code ends up getting instantiated twice, however it also
makes it possible to eliminate all "convert" functions, so
it's not a big net code size increase. And if desired, the
duplicate instantiations could be eliminated with some
reorganization.
llvm-svn: 81861
2009-09-15 15:58:07 +00:00
7465cd769c
Add more newlines to make up for the ones removed from the end of instructions.
...
llvm-svn: 81851
2009-09-15 07:08:25 +00:00
a38ff3edae
Another try at early partial coalescing. Identity phi source copies (their sources are defined by phi join def) are coalesced. And the phi join copy is backward copy propagated into the other copies.
...
Still miscompiling some tests. :-(
llvm-svn: 81849
2009-09-15 06:45:16 +00:00
c5cb3bd927
Forbid arrays of function-type and structures with function-typed fields.
...
While I'm there, change code that does:
SomeTy == Type::getFooType(Context)
into:
SomeTy->getTypeID() == FooTyID
to decrease the amount of useless type creation which may involve locking, etc.
llvm-svn: 81846
2009-09-15 06:28:26 +00:00
ee8f74f5d1
fix PR4963: folding insertvalue would sometimes turn a packed struct into
...
an unpacked one.
llvm-svn: 81845
2009-09-15 06:28:12 +00:00
e0987215f0
add a new CallGraphNode::replaceCallEdge method and use it from
...
argpromote to avoid invalidating an iterator. This fixes PR4977.
All clang tests now pass with expensive checking (on my system
at least).
llvm-svn: 81843
2009-09-15 05:40:35 +00:00
e9a4992399
add newline to debug dump
...
llvm-svn: 81840
2009-09-15 05:14:57 +00:00
d4d966f500
make -debug-pass=Executions show information about what call graph nodes
...
are in the SCC for each execution of a CGSCC pass.
llvm-svn: 81838
2009-09-15 05:03:04 +00:00
f82f27be3f
add some missing quotes in debug output
...
llvm-svn: 81836
2009-09-15 04:45:26 +00:00
8e780355b7
Remove invalid add_dependencies line to unbreak the CMake build.
...
llvm-svn: 81827
2009-09-15 04:06:36 +00:00
0adf07eb9e
Get rid of GetProcessId in Win32/Program.inc.
...
GetProcessId was introduced only in XP. As a bonus, this change makes Program
objects copyable, since Program is now basically a PID.
llvm-svn: 81826
2009-09-15 03:39:45 +00:00
520a6856ba
Don't pull a load through a callseq_start if the load's chain
...
has multiple uses, as one of the other uses may be on a path
to a different node above the callseq_start, because that
leads to a cyclic graph. This problem is exposed when
-combiner-global-alias-analysis is used. This fixes PR4880.
llvm-svn: 81821
2009-09-15 01:22:01 +00:00
d41f8fd2b3
Remove incorrect CSE code from r81813.
...
llvm-svn: 81819
2009-09-15 00:38:09 +00:00
050e0cdb9a
Modified the Intel instruction tables to include
...
versions of CALL and JMP with segmented addresses
provided in-line, as pairs of immediates.
llvm-svn: 81818
2009-09-15 00:35:17 +00:00
ccab317837
Added the first bits of the ARM target assembler to llvm-mc. For now it only
...
parses the .word directive as 4 bytes and ARMAsmParser::ParseInstruction will
give an error is called. Broke out the test of the .word directive into two
different test cases, one for x86 and one for arm.
llvm-svn: 81817
2009-09-15 00:27:25 +00:00
879d8f1c3e
Substantially speed up combiner-aa in the following ways:
...
1. Switch from an std::set to a SmallPtrSet for visited chain nodes.
2. Do not force the recursive flattening of token factor nodes, regardless of
use count.
3. Immediately process newly created TokenFactor nodes.
Also, improve combiner-aa by teaching it that loads to non-overlapping offsets
of relatively aligned objects cannot alias.
These changes result in a >5x speedup for combiner-aa on most testcases.
llvm-svn: 81816
2009-09-15 00:18:30 +00:00
01c1e1152d
Teach the legalizer to propagate the original alignment of loads and store when
...
it splits them.
llvm-svn: 81815
2009-09-15 00:14:28 +00:00
65829a4ccb
On x86-64, the 32-bit cmov doesn't actually clear the high 32-bit of
...
its result if the condition is false.
llvm-svn: 81814
2009-09-15 00:14:11 +00:00
02a685a914
Add an "original alignment" field to load and store nodes. This enables the
...
DAG Combiner to disambiguate chains for loads and stores of types which are
broken up by the Legalizer into smaller pieces.
llvm-svn: 81813
2009-09-15 00:13:12 +00:00
f9eafce3af
When extending a memset range past the front, set the alignment of the
...
memset region to the alignment of the new start address.
llvm-svn: 81810
2009-09-14 23:39:10 +00:00
Chris Lattner