Commit Graph

33478 Commits

Author SHA1 Message Date
Chris Lattner
5638dc618a Use the libanalysis 'ConstantFoldLoadFromConstPtr' function
instead of reinventing SCCP-specific logic.  This gives us
new powers.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85789 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-02 06:06:14 +00:00
Chris Lattner
2a0433beea 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.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85788 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-02 05:55:40 +00:00
Evan Cheng
7baae87d8f Unbreak ARMBaseRegisterInfo::copyRegToReg.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85787 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-02 04:44:55 +00:00
Chris Lattner
794c15dc71 fix a bug exposed by moving SRoA earlier which caused a crash building kc++
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85786 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-02 04:37:17 +00:00
Chris Lattner
89112b608f only IPSCCP incoming arguments if the function is executable, this fixes
an assertion on the buildbot.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85784 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-02 03:25:55 +00:00
Chris Lattner
36c9952b1b add a new ValueState::getConstantInt() helper, use it to
simplify some code.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85783 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-02 03:21:36 +00:00
Edward O'Callaghan
bd5dc45296 Fix malloc.h is deprecated warning on DragonFly BSD.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85782 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-02 03:20:57 +00:00
Edward O'Callaghan
438b00b2ad Fix for warning seen on DF-BSD, Victor, please fix this to use a shift instead of pow()
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85781 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-02 03:14:31 +00:00
Chris Lattner
38871e4c3b tidy up some more: remove some extraneous inline specifiers, return harder.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85780 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-02 03:03:42 +00:00
Edward O'Callaghan
2afd0723a2 Apply fix for PR5135, Credit to Andreas Neustifter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85779 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-02 02:55:39 +00:00
Chris Lattner
8db50122a4 eliminate the SCCPSolver::getValueMapping method.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85778 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-02 02:54:24 +00:00
Chris Lattner
301a2792be fix failures introduced in r85774
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85777 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-02 02:48:17 +00:00
Chris Lattner
cc4f60b113 factor duplicated code into a new DeleteInstructionInBlock
function, eliminate temporary (and pointless) smallvector.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85776 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-02 02:47:51 +00:00
Chris Lattner
2f09625a9f Chris used to use '...' instead of proper grammar.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85775 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-02 02:33:50 +00:00
Chris Lattner
ea0db070a1 remove some extraneous llvmcontext stuff.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85774 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-02 02:30:06 +00:00
Chris Lattner
7927220aef change LatticeVal to use PointerIntPair to save some space.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85773 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-02 02:20:32 +00:00
Chris Lattner
7ebbabf500 fix instcombine to only do store sinking when the alignments
of the two loads agree.  Propagate that onto the new store.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85772 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-02 02:06:37 +00:00
Anton Korobeynikov
2ae0eec1c0 Handle splats of undefs properly. This includes the testcase for PR5364 as well.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85767 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-02 00:12:06 +00:00
Anton Korobeynikov
3a639a07ea 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!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85766 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-02 00:11:39 +00:00
Anton Korobeynikov
2e1da9fea4 64-bit FP loads & stores operate on both NEON and VFP pipelines.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85765 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-02 00:11:06 +00:00
Anton Korobeynikov
f95215f551 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)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85764 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-02 00:10:38 +00:00
Evan Cheng
f57b1baa44 Add PseudoSourceValue::mayAlias. It returns true if the object can ever alias any LLVM IR value.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85762 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 23:50:04 +00:00
Evan Cheng
e3ce8aab0a Fix a couple more places where we are creating ld / st instructions without memoperands.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85746 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 22:04:35 +00:00
Evan Cheng
48d8afab73 Make use of imm12 version of Thumb2 ldr / str instructions more aggressively.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85743 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 21:12:51 +00:00
Chris Lattner
a664bb7bdc 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85738 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 20:07:07 +00:00
Chris Lattner
751a362c22 split load sinking out to its own function, like gep sinking.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85737 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 20:04:24 +00:00
Chris Lattner
e3c6281a98 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85736 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 19:50:13 +00:00
Duncan Sands
b046964101 Add a missing closing parenthesis, and tweak to fit in 80
columns.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85732 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 19:12:43 +00:00
Chris Lattner
ec4c7b9766 cleanups, switch GlobalDCE to SmallPtrSet instead of std::set
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85730 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 19:03:42 +00:00
Chris Lattner
c4775e4b97 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85728 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 18:42:03 +00:00
Chris Lattner
280f3fedbf improve comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85725 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 18:17:37 +00:00
Chris Lattner
66588707ac add a comment about why we don't allow inlining indbr.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85724 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 18:16:30 +00:00
Chris Lattner
4a7642ec9c the verifier shouldn't modify the IR.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85722 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 18:11:50 +00:00
Douglas Gregor
076124ef26 Reverting 85714, 85715, 85716, which are breaking the build
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85717 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 16:42:53 +00:00
Dan Gohman
2166f62290 Add a function to Passes.h to allow clients to create instances
of the ScalarEvolution pass without needing to #include ScalarEvolution.h.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85716 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 15:28:36 +00:00
Dan Gohman
34e9d7b6be Don't #include Pass.h from CallGraph.h.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85715 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 15:23:35 +00:00
Dan Gohman
b2fae7560d Remove the #include of Pass.h from PassManager.h. This breaks a significant
#include dependency, as frontends commonly pull in PassManager.h.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85714 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 15:20:19 +00:00
Chris Lattner
b2710041fb 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85709 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 06:11:53 +00:00
Chris Lattner
882029269e 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85708 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 04:57:33 +00:00
Chris Lattner
0b649dd91f 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85707 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 04:23:20 +00:00
Chris Lattner
660a4f389c fix an issue where the verifier would reject a function whose entry
block had its address taken even if the blockaddress was dead.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85706 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 04:08:01 +00:00
Chris Lattner
da9c281121 if CostMetrics says to never duplicate some code, don't unswitch a loop.
This prevents unswitching from duplicating indbr's.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85705 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 03:42:55 +00:00
Chris Lattner
0a4c6789d5 constant fold indirectbr(blockaddress(%bb)) -> br label %bb.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85704 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 03:40:38 +00:00
Chris Lattner
43f44aa160 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85703 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 03:25:03 +00:00
Chris Lattner
b93a23a532 pull check for return inst out of loop, never inline a callee that contains
an indirectbr.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85702 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 03:07:53 +00:00
Chris Lattner
d0ec2352ae Fix BlockAddress::replaceUsesOfWithOnConstant to correctly
maintain the block use count in SubclassData.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85701 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 03:03:03 +00:00
Chris Lattner
2a749d3647 implement linker support for BlockAddress.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85700 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 02:46:39 +00:00
Chris Lattner
cdfc940912 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85699 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-01 01:27:45 +00:00
Evan Cheng
de17fb6e4d Use cbz and cbnz instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85698 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 23:46:45 +00:00
Jim Grosbach
8cd0a8cb82 vml[as].f32 cause stalls in following advanced SIMD instructions. Avoid using
them for scalar floating point operations for now.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85697 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 22:57:36 +00:00
Chris Lattner
fb6e701f08 Make sure PRE doesn't split crit edges from indirectbr.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85692 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 22:11:15 +00:00
Chris Lattner
a4b04210d4 llvm::SplitEdge should refuse to split an edge from an indirectbr.
Fix CodeGenPrepare to not try to split edges from indirectbr.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85690 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 22:04:43 +00:00
Chris Lattner
1b98ff3e4f update the comment above llvm::SplitCriticalEdge, and make
it abort on IndirectBrInst as describe in the comment.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85688 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 21:51:10 +00:00
Jim Grosbach
bcf2f2c159 Expand 64-bit logical shift right inline
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85687 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 21:42:19 +00:00
Jim Grosbach
b4a976c304 Expand 64-bit arithmetic shift right inline
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85685 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 21:00:56 +00:00
Dan Gohman
b953c944b6 Fix a missing newline in the dwarf output code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85684 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 20:59:09 +00:00
Dan Gohman
0ba90f3e34 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85682 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 20:19:03 +00:00
Dan Gohman
499a9377a3 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85681 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 20:17:39 +00:00
Chris Lattner
c1dc35094e adjust a couple xforms to work with null bb's in BlockAddress.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85680 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 20:13:24 +00:00
Chris Lattner
7d83ebcadd 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).


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85678 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 20:08:37 +00:00
Jim Grosbach
c2b879fcfe 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85675 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 19:38:01 +00:00
Chris Lattner
28f3d34a9b add a comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85671 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 17:48:31 +00:00
Dan Gohman
438b583dbd 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85670 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 17:33:01 +00:00
Dan Gohman
dca094100b Remove redundant code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85668 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 16:16:41 +00:00
Dan Gohman
f230d8ad15 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85667 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 16:08:00 +00:00
Dan Gohman
4c7279ac72 Rename forgetLoopBackedgeTakenCount to forgetLoop, because it
clears out more information than just the stored backedge taken count.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85664 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 15:04:55 +00:00
Dan Gohman
a2aabe38ae Replace LoopUnrollPass.cpp's custom code-size estimation code using
the new common CodeMetrics code.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85663 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 14:54:17 +00:00
Dan Gohman
b56c966019 Simplify this code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85662 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 14:46:50 +00:00
Dan Gohman
3f9432ebca Remove an unnecessary #include.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85661 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 14:39:43 +00:00
Dan Gohman
13d63f2d26 Update CMakeLists for recent renames.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85660 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 14:38:25 +00:00
Dan Gohman
928a75f057 Rename UnrollLoop.cpp to LoopUnroll.cpp, and LoopUnroll.cpp to
LoopUnrollPass.cpp, for consistency with other passes which are
similarly split.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85659 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 14:37:31 +00:00
Dan Gohman
9f476e3179 Remove CodeGenLICM. It's largely obsoleted by MachineLICM's new ability
to unfold loop-invariant loads.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85657 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 14:35:41 +00:00
Dan Gohman
576fd76a68 Make ScalarEvolutionAliasAnalysis slightly more aggressive, by making an
underlying alias call even for non-identified-object values.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85656 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 14:32:25 +00:00
Dan Gohman
5b097014f2 Reapply r85634, with the bug fixed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85655 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 14:22:52 +00:00
Dan Gohman
75832d7d4f When discarding SrcValue information, discard all of it so that code
that uses this information knows to behave conservatively.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85654 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 14:14:04 +00:00
Dan Gohman
60d686df54 Fix 80-column violation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85653 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 14:12:53 +00:00
Eric Christopher
cdfa666c30 Fix warning with gcc-4.0 and signed/unsigned.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85648 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 09:24:35 +00:00
Evan Cheng
9eda68988e It's safe to remat t2LDRpci; Add PseudoSourceValue to load / store's to enable more machine licm. More changes coming.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85643 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 03:39:36 +00:00
Evan Cheng
a4006c6dbf Revert 85634. It's breaking consumer-typeset (and others).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85641 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 01:28:06 +00:00
Dan Gohman
1410b7620a Add assertion checks here to turn silent miscompiles into aborts.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85639 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-30 23:59:06 +00:00
Dan Gohman
4cddfd9b6f 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85638 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-30 23:57:47 +00:00
Dan Gohman
0e747beb3c Add a comment about a missed opportunity.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85635 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-30 23:15:43 +00:00
Dan Gohman
1107c63efe 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85634 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-30 23:15:21 +00:00
Kevin Enderby
9c41fa87ea 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85632 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-30 22:55:57 +00:00
Devang Patel
5e8e2d7ad9 If string field is empty then return NULL.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85630 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-30 22:52:47 +00:00
Chris Lattner
dac8bde235 if basic blocks are destroyed while there are *just* BlockAddress' hanging
around, then zap them.  This is analogous to dangling constantexprs hanging
off functions.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85627 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-30 22:39:36 +00:00
Dan Gohman
2c63566e40 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85626 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-30 22:39:04 +00:00
Chris Lattner
3821176b2e make hasAddressTaken() constant time by storing a refcount in BB's subclass data.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85625 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-30 22:33:29 +00:00
Bob Wilson
57f224a5a4 Add a note about Robert Muth's alternate jump table implementation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85624 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-30 22:22:46 +00:00
Dan Gohman
8e42e4baae 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85623 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-30 22:22:22 +00:00
Dan Gohman
0115e164ba 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85622 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-30 22:18:41 +00:00
Chris Lattner
287db0c23c it isn't valid to take the address of the entry block.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85621 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-30 22:15:48 +00:00
Devang Patel
7cb7e12c08 If a type is derived from a derived type then calculate size appropriately.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85619 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-30 22:09:30 +00:00
Bob Wilson
929ffa2414 Fix a comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85610 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-30 20:13:25 +00:00
Evan Cheng
2f192c256c Add option to createGVNPass to disable PRE.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85609 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-30 20:12:24 +00:00
Lang Hames
788fa17af8 Stop the iterator in ValueLiveAt from potentially running off the end of the interval.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85599 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-30 18:12:09 +00:00
Rafael Espindola
c1382b745f 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.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85590 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-30 14:33:14 +00:00
Bob Wilson
ddb16df912 Add ARM codegen for indirect branches.
clang/test/CodeGen/indirect-goto.c runs! (unoptimized)


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85577 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-30 05:45:42 +00:00
Dan Gohman
0b81e19d19 Most stack straces don't need 3 digits worth of levels.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85575 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-30 02:45:10 +00:00
Dan Gohman
ab918103d2 Don't delete blocks which have their address taken.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85572 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-30 02:13:27 +00:00