ef7c1fd909
Unbreak my CMake build. Say you'll link again.
...
llvm-svn: 80842
2009-09-02 22:45:31 +00:00
c180029be4
Revert 80839 for now. It causes test failures.
...
llvm-svn: 80841
2009-09-02 21:49:26 +00:00
36309cd391
Use CallbackVH, instead of WeakVH, to hold MDNode elements.
...
Use FoldingSetNode to unique MDNodes in a context.
Use CallbackVH hooks to update context's MDNodeSet appropriately.
llvm-svn: 80839
2009-09-02 21:22:09 +00:00
f0da41c3e4
More missed vdup patterns
...
llvm-svn: 80838
2009-09-02 21:21:28 +00:00
c76bfb777e
Switch llc from ParseBitcodeFile to ParseIRFile. This lets llc
...
transparently read either LLVM Assembly or LLVM Bitcode files.
llvm-svn: 80829
2009-09-02 19:35:19 +00:00
a7e8799c72
Add a comment noting the memory ownership rules.
...
llvm-svn: 80827
2009-09-02 19:21:56 +00:00
b7cb99bf50
Add convenience functions for reading in LLVM IR that autodetect
...
and LLVM Assembly and LLVM Bitcode and automatically call the
corresponding reader.
llvm-svn: 80809
2009-09-02 17:54:06 +00:00
7bd37b47c8
switch from std::string to SmallString + raw_svector_ostream.
...
llvm-svn: 80807
2009-09-02 17:37:38 +00:00
74f4ca7cb2
split mcinst lowering stuff out to its own file.
...
llvm-svn: 80806
2009-09-02 17:35:12 +00:00
957b1316e6
Fix the syntax of add/sub/mul nsw/nuw and sdiv exact.
...
llvm-svn: 80805
2009-09-02 17:31:42 +00:00
ae58e72018
Add const qualifiers for isBitcodeWrapper, and add new functions
...
isRawBitcode and isBitcode to allow clients to test whether a given
memory buffer holds a bitcode image.
llvm-svn: 80804
2009-09-02 17:21:29 +00:00
77ac99dc31
Refactor common code from ParseAssemblyString and ParseAssemblyFile,
...
to expose a low-level interface for parsing from an existing MemoryBuffer.
llvm-svn: 80803
2009-09-02 17:18:19 +00:00
87d69e80b1
Add const qualifiers to dominates' arguments.
...
llvm-svn: 80801
2009-09-02 17:05:05 +00:00
69e2afe030
Removed temporarily because of breaking Darwin builds.
...
(See http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090831/086214.html )
llvm-svn: 80799
2009-09-02 16:47:24 +00:00
bc751989f1
Rearrange code to eliminate redundancy and avoid gotos.
...
llvm-svn: 80798
2009-09-02 16:35:35 +00:00
2baa6a38d2
plug another leak in LLParser::PerFunctionState::SetInstName()
...
llvm-svn: 80792
2009-09-02 15:02:57 +00:00
da6b0fa8ed
Changed profiling-tool-chain.ll test to use optimal-edge-profiling instead of
...
edge-profiling, this is more useful since the loading of the
optimal-edge-profiling is more complicated.
The edge-profiling is tested in edge-profiling.ll where only the
instrumentation is tested.
llvm-svn: 80791
2009-09-02 14:24:08 +00:00
2fcee70aeb
plug memory leak in LLParser::PerFunctionState::SetInstName() by deleting a value after replacing it
...
llvm-svn: 80790
2009-09-02 14:22:03 +00:00
ae866b0c66
Sort edges in MaximumSpanningTree more stable in case of equal weight.
...
(See http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090824/085890.html )
llvm-svn: 80789
2009-09-02 14:03:11 +00:00
964fa2bdac
Changed set of BlocksToInstrument to set of InsertedBlocks that do not have to
...
be instrumented.
llvm-svn: 80788
2009-09-02 13:59:05 +00:00
4469c164d0
Code cleanups and added comments.
...
llvm-svn: 80781
2009-09-02 12:38:39 +00:00
82131d6ad7
Opaque types didn't work if llvm_is_multithreaded().
...
AlwaysOpaqueTy is always NULL at this point, and it causes an assertion failure.
Fix it by using the just constructed tmp instead.
llvm-svn: 80780
2009-09-02 12:23:05 +00:00
18e03dd70c
Fix DbgStopPointInst->getFileName/getDirectory, broken by the MDNodification in
...
r80406, and readd a -print-dbginfo test.
llvm-svn: 80778
2009-09-02 11:13:56 +00:00
68c5f477fa
Retype from unsigned to CallingConv::ID accordingly. Approved by Bob Wilson.
...
llvm-svn: 80773
2009-09-02 08:44:58 +00:00
defbbc9105
Fix month.
...
llvm-svn: 80769
2009-09-02 06:34:22 +00:00
4916267c97
fix PR4815: some cases where DeleteDeadInstruction can delete
...
the instruction BBI points to.
llvm-svn: 80768
2009-09-02 06:31:02 +00:00
09a79dcfdf
clean up this code a bit.
...
llvm-svn: 80767
2009-09-02 06:15:37 +00:00
2dd09dbdf7
eliminate VISIBILITY_HIDDEN from Transforms/Scalar. PR4861
...
llvm-svn: 80766
2009-09-02 06:11:42 +00:00
d5f4fcceae
refactor select 'sched insertion' out to its own method.
...
llvm-svn: 80764
2009-09-02 05:57:00 +00:00
cc8c581a5b
Add support for modeling whether or not the processor has support for
...
conditional moves as a subtarget feature. This is the easy part of
PR4841.
llvm-svn: 80763
2009-09-02 05:53:04 +00:00
950e60ea97
fix PR4848 an infinite loop when indexing down through a recursive gep
...
and we get the original pointer type. This doesn't mean that we're
at the first pointer being indexed. Correct the predicate.
llvm-svn: 80762
2009-09-02 05:35:45 +00:00
64b5842986
fix PR4837, some bugs folding vector compares. These
...
return a vector of i1, not i1 itself.
llvm-svn: 80761
2009-09-02 05:12:37 +00:00
65fb597793
revert my patch, duncan points out what is wrong with my logic. Add
...
a comment so that I don't change this in the future :)
llvm-svn: 80760
2009-09-02 04:39:04 +00:00
8f23276431
one more try at making this simpler, hopefully it won't break everything :)
...
llvm-svn: 80759
2009-09-02 04:34:06 +00:00
5632d96176
Complicate Chris's simplification, avoiding complaints
...
about singular iterators when building with expensive
checks turned on.
llvm-svn: 80757
2009-09-02 03:48:41 +00:00
0935339f81
Don't force the triple or data layout in this test. We just have to get them
...
from the host and hope that works.
llvm-svn: 80751
2009-09-02 02:43:11 +00:00
24a60cd2eb
Add a comment.
...
llvm-svn: 80749
2009-09-02 01:14:16 +00:00
a8ee2e7496
Fixed the asmstrings for 8-bit, 16-bit, and 32-bit ADD %rAX, imm instructions.
...
Added a 64-bit ADD %RAX, imm32 instruction.
Added all 4 forms for AND %rAX, imm and CMP %rAX, imm.
llvm-svn: 80746
2009-09-02 00:55:49 +00:00
10a5dabfd5
Fix PR4845: r77946 completely broke x86_64 Darwin (or any situation where the
...
desired triplet is a sub-target, e.g. thumbv7 vs. arm host). Reverting the
patch isn't quite right either since the previous behavior does not allow the
triplet to be overridden with -march.
llvm-svn: 80742
2009-09-02 00:19:03 +00:00
34455c9c65
Disable uniqueness test for now.
...
llvm-svn: 80741
2009-09-02 00:16:33 +00:00
8c7f86e698
For now disable MDNode uniquing. This fixes llvm-gcc bootstrap failure on certain Mac OS X 10.5. I am working on a proper fix.
...
llvm-svn: 80738
2009-09-01 23:56:42 +00:00
fcd618bf1d
Avoid calling removeVirtualRegisterKilled which iterates over the operands
...
to find the kill, since we already have the operand.
llvm-svn: 80736
2009-09-01 23:18:46 +00:00
efe96918f8
Refactor some code into separate functions. No functional changes.
...
llvm-svn: 80733
2009-09-01 22:51:08 +00:00
71e77ea89b
Move use of LV inside condition that guards for null LV.
...
llvm-svn: 80731
2009-09-01 22:19:00 +00:00
2a275102d3
Fix build warning.
...
llvm-svn: 80730
2009-09-01 22:07:12 +00:00
ff0e622d1c
Simplify.
...
llvm-svn: 80729
2009-09-01 22:07:06 +00:00
9ab9e0aa25
Fix what I believe is a copy-n-pasto introduced in r78129.
...
- Bruno, please check!!
llvm-svn: 80728
2009-09-01 22:07:00 +00:00
6c3843866c
X86/Encoding: Support ExternalSymbol operands in emitDisplacementField (for consistency).
...
llvm-svn: 80727
2009-09-01 22:06:53 +00:00
f7a14aa43d
Remove Offset from ExternalSybmol MachineOperands, this is unused (and at least partly unsupported, in X86 encoding at least).
...
llvm-svn: 80726
2009-09-01 22:06:46 +00:00
6190bb01be
debug intrinsics do not go in the callgraph, this fixes a couple
...
clang regtest failures.
llvm-svn: 80724
2009-09-01 21:37:50 +00:00
339c82df85
Fix a regression I introduced in r80708, found by llvm-test.
...
llvm-svn: 80718
2009-09-01 20:33:43 +00:00
92e18c456a
Fix ELF Writter related memory leaks
...
llvm-svn: 80717
2009-09-01 19:25:52 +00:00
e3cdafc70b
OptimalEdgeProfiling: Reading in Profiles.
...
This enables LLVM to read the OptimalEdgeProfiles.
llvm-svn: 80715
2009-09-01 19:08:51 +00:00
e7fa9fa3cc
Addedum to r80712, forgot to add files.
...
llvm-svn: 80713
2009-09-01 19:05:58 +00:00
759094e323
OptimalEdgeProfiling: Creation of profiles.
...
This adds the instrumentation and runtime part of OptimalEdgeProfiling.
llvm-svn: 80712
2009-09-01 19:03:44 +00:00
4fde5c756d
Small fix in ProfileEstimator that eliminates duplicated code.
...
llvm-svn: 80711
2009-09-01 19:01:59 +00:00
e4459e466f
reduce size of SmallString to something more reasonable
...
llvm-svn: 80710
2009-09-01 18:55:08 +00:00
9b463729d7
remove CallGraphNode::replaceCallSite, it is redundant with other APIs.
...
llvm-svn: 80708
2009-09-01 18:52:39 +00:00
d7797754d4
Add support for generating code for vst{234}lane intrinsics.
...
llvm-svn: 80707
2009-09-01 18:51:56 +00:00
f61b0fb5d0
cleanup/simplify
...
llvm-svn: 80706
2009-09-01 18:50:55 +00:00
39dc89b458
Fix incorrect declarations of intrinsics in this test.
...
llvm-svn: 80705
2009-09-01 18:50:43 +00:00
c5fcbdc7ee
Use raw_ostream instead of sstream
...
llvm-svn: 80704
2009-09-01 18:49:12 +00:00
8900f3ec57
remove a bunch of explicit code previously needed to update the
...
callgraph. This is now dead because RAUW does the job.
llvm-svn: 80703
2009-09-01 18:44:06 +00:00
7f6516949b
Add hidden flags to allow binary search of post-RA scheduling errors.
...
llvm-svn: 80702
2009-09-01 18:34:03 +00:00
5f582b7290
RRX reads CPSR.
...
llvm-svn: 80699
2009-09-01 18:32:09 +00:00
e33167ac35
doxygenate RefreshCallGraph, add a new 'verification mode', and run it after
...
CGSCC passes make change to ensure they are updating the callgraph correctly
(when assertions are enabled).
llvm-svn: 80698
2009-09-01 18:32:03 +00:00
4623a26ec3
Fix a typo in a comment.
...
llvm-svn: 80697
2009-09-01 18:29:01 +00:00
09b2e51711
Added TEST %rAX, $imm instructions to the Intel tables. These are required for the X86 disassembler.
...
llvm-svn: 80696
2009-09-01 18:14:18 +00:00
3f7b3d149f
simpler solution to iterator invalidation "problem" found
...
by expensive checking.
llvm-svn: 80695
2009-09-01 18:13:40 +00:00
1145e33bc6
enhance memcpy opt to turn memmoves into memcpy when the src/dest
...
don't alias. Remove an old and poorly reduced testcase that fails
with this transform for reasons unrelated to the original test.
llvm-svn: 80693
2009-09-01 17:56:32 +00:00
0f20a5b338
Reapply 80278
...
Add MO flags to simplify the printing of relocations.
Remove the support for printing large code model relocs (which
aren't supported anyway).
llvm-svn: 80691
2009-09-01 17:27:58 +00:00
5afbf2b541
revert inadvertant change from previous commit
...
llvm-svn: 80689
2009-09-01 17:19:13 +00:00
b5557a7b42
random code cleanups, no functionality change.
...
llvm-svn: 80682
2009-09-01 17:09:55 +00:00
1543d133db
Update CMake files.
...
llvm-svn: 80680
2009-09-01 17:01:02 +00:00
9372f4d7d5
Simply LSDA lable emission to use a direct special-case output instead of
...
EmitLabel()
llvm-svn: 80677
2009-09-01 16:43:35 +00:00
1f8303bfce
Do not manipulate invalid iterators. This fixes the
...
llvm-gcc build when expensive checking is turned on.
llvm-svn: 80671
2009-09-01 15:13:02 +00:00
a19e971f0c
Further refactoring of PIC16 Obj file code.
...
llvm-svn: 80670
2009-09-01 10:47:31 +00:00
a43ad3ddbd
Update CMakeLists.
...
llvm-svn: 80669
2009-09-01 10:24:10 +00:00
eb5a9d34d6
Preparation for Optimal Edge Profiling:
...
Add statistics for regular edge profiling, this enables the comparation of the
number of edges inserted by regular and optimal edge profiling.
llvm-svn: 80668
2009-09-01 10:08:39 +00:00
8bcd5dcf5d
Preparation for Optimal Edge Profiling:
...
Optimal edge profiling is only possible when blocks with no predecessors get an
virtual edge (BB,0) that counts the execution frequencies of this
function-exiting blocks.
This patch makes the necessary changes before actually enabling optimal edge profiling.
llvm-svn: 80667
2009-09-01 10:06:05 +00:00
b683b6e2ce
Preparation for Optimal Edge Profiling:
...
This adds a pass to verify the current profile against the flow conditions.
This is very helpful when later on trying to perserve the profiling information
during all passes.
llvm-svn: 80666
2009-09-01 08:48:42 +00:00
026e34c4f7
testcase for PR3601
...
llvm-svn: 80664
2009-09-01 06:33:49 +00:00
063d06527e
Change CallGraphNode to maintain it's Function as an AssertingVH
...
for sanity. This didn't turn up any bugs.
Change CallGraphNode to maintain its "callsite" information in the
call edges list as a WeakVH instead of as an instruction*. This fixes
a broad class of dangling pointer bugs, and makes CallGraph have a number
of useful invariants again. This fixes the class of problem indicated
by PR4029 and PR3601.
llvm-svn: 80663
2009-09-01 06:31:31 +00:00
aa57e1085e
Add virtual destructor.
...
llvm-svn: 80660
2009-09-01 05:04:28 +00:00
ff69320427
Add test for vld{234}_lane instructions.
...
llvm-svn: 80658
2009-09-01 04:27:10 +00:00
da9817cddd
Generate code for vld{234}_lane intrinsics.
...
llvm-svn: 80656
2009-09-01 04:26:28 +00:00
33b408a10f
Fix pr4843: When an instruction has multiple destination registers that are
...
tied to different source registers, the TwoAddressInstructionPass needs to
be smarter. Change it to check before replacing a source register whether
that source register is tied to a different destination register, and if so,
defer handling it until a subsequent iteration.
llvm-svn: 80654
2009-09-01 04:18:40 +00:00
4fa0595b94
llvm-mc: Store MCSymbolData value as a pointer (to make MSVC happy).
...
llvm-svn: 80652
2009-09-01 04:09:03 +00:00
f09e8d5497
SJLJ is arm/darwin only for now. force the triple for the test
...
llvm-svn: 80651
2009-09-01 02:34:49 +00:00
345a5ae6e4
Fix compiler warnings
...
llvm-svn: 80650
2009-09-01 02:05:03 +00:00
20eac92d88
Clean up LSDA name generation and use for SJLJ exception handling. This
...
makes an eggregious hack somewhat more palatable. Bringing the LSDA forward
and making it a GV available for reference would be even better, but is
beyond the scope of what I'm looking to solve at this point.
Objective C++ code could generate function names that broke the previous
scheme. This fixes that.
llvm-svn: 80649
2009-09-01 01:57:56 +00:00
b1bff47718
Introduce DILocation.
...
llvm-svn: 80648
2009-09-01 01:14:15 +00:00
71b42f2d40
Add getDirectory() and getFilename() interface to DIScope.
...
llvm-svn: 80647
2009-09-01 00:53:21 +00:00
b2de5fa689
Subprogram is a scope. Derive DISubprogram from DIScope.
...
llvm-svn: 80637
2009-08-31 22:47:13 +00:00
869529c58c
Rename DIBlock as DILexicalBlock.
...
llvm-svn: 80633
2009-08-31 22:00:15 +00:00
dba696afc0
Don't use an iterator which is potentially invalidated.
...
llvm-svn: 80632
2009-08-31 21:58:28 +00:00
e98396d3fc
Use early exit and reduce indentation.
...
llvm-svn: 80631
2009-08-31 21:54:55 +00:00
53b6a02dd1
If the tied registers are already the same, there is no need to change
...
them. Move the code to make that change inside the conditional.
llvm-svn: 80630
2009-08-31 21:54:16 +00:00
e7717e7b2a
Add triple parsing support for XCore.
...
llvm-svn: 80629
2009-08-31 21:51:36 +00:00
e4cfcc3081
Derive DICompileUnit from DIScope.
...
llvm-svn: 80627
2009-08-31 21:34:44 +00:00
c87c1e2cfd
Add flag to mark structs for Apple Block "byref" variables; also add code to
...
modify the type and location debug information for these variables to match the
programmer's expectations.
llvm-svn: 80625
2009-08-31 21:19:37 +00:00
cc2f1eb7fe
Extend the ValuesAtScope cache to cover all expressions, not just
...
SCEVUnknowns, as the non-SCEVUnknown cases in the getSCEVAtScope code
can also end up repeatedly climing through the same expression trees,
which can be unusably slow when the trees are very tall.
Also, add a quick check for SCEV pointer equality to the main
SCEV comparison routine, as the full comparison code can be expensive
in the case of large expression trees.
These fix compile-time problems in some pathlogical cases.
llvm-svn: 80623
2009-08-31 21:15:23 +00:00
f04db57674
restore semantics of operator* (removing a FIXME I had to introduce in r80224)
...
llvm-svn: 80622
2009-08-31 20:54:23 +00:00
c8985204d9
Don't mark a register live at an undef use.
...
llvm-svn: 80621
2009-08-31 20:47:02 +00:00
d9b6115f27
Introduce DIScope.
...
llvm-svn: 80620
2009-08-31 20:44:45 +00:00
88a26fb0f1
Oops. Fix inverted logic in assertion check.
...
llvm-svn: 80618
2009-08-31 20:27:49 +00:00
4f835f1d7d
Remove .n suffix for some 16-bit opcodes now that Darwin assembler is fixed.
...
llvm-svn: 80615
2009-08-31 20:14:07 +00:00
b0215efc45
X86/exp-asm-printer: Lower MachineOperand::MO_JumpTableIndex to MCOperand.
...
- Down to 7 failures on 403.gcc.
llvm-svn: 80605
2009-08-31 19:14:05 +00:00
443f849b6b
Stop printing old asm printing code inline with -experimental-asm-printer (this allows diffing and assembling the .s)
...
llvm-svn: 80604
2009-08-31 19:13:56 +00:00
524630915d
Avoid unnecessary +0 in experimental-asm-printer.
...
llvm-svn: 80603
2009-08-31 19:13:47 +00:00
9fda4bd998
Simplify isDerivedType() and other predicate interface.
...
llvm-svn: 80602
2009-08-31 18:49:10 +00:00
6627c4ec82
Output a hex value, because all of the others are hex.
...
llvm-svn: 80601
2009-08-31 18:26:48 +00:00
70e2f093a4
Fix some misspellings of XTARGET.
...
llvm-svn: 80598
2009-08-31 18:05:23 +00:00
a1aed7bb40
Mark test as passing on all x86, which it should,
...
although I don't think anyone cares about this
feature except Darwin. PR 4825.
llvm-svn: 80596
2009-08-31 17:49:20 +00:00
a9262bb19c
cleanups pointed out by duncan
...
llvm-svn: 80595
2009-08-31 17:08:30 +00:00
9cf8bcb69d
Revert commit 80428. It completely broke exception
...
handling on x86-32 linux.
llvm-svn: 80592
2009-08-31 16:45:16 +00:00
d18e668fbc
Free the constants that have no uses in ~LLVMContext.
...
This fixes leaks from LLVMContext in multithreaded apps.
Since constants are only deleted if they have no uses, it is safe to not delete
a Module on shutdown, as many single-threaded tools do.
Multithreaded apps should however delete the Module before destroying the
Context to ensure that there are no leaks (assuming they use a different context
for each thread).
llvm-svn: 80590
2009-08-31 16:14:59 +00:00
487d576020
Fix ExplicitSymbols leak.
...
llvm-svn: 80589
2009-08-31 16:12:29 +00:00
3c62f5095a
Normalize makefile comments and sort cmake file lists.
...
llvm-svn: 80584
2009-08-31 13:05:24 +00:00
897ffadd2d
llvm-mc: Pass values to MCStreamer as MCExprs, not MCValues.
...
llvm-svn: 80578
2009-08-31 08:09:28 +00:00
b7b2097aac
llvm-mc: Simplify EmitAssignment ('.set' is identical to '=').
...
llvm-svn: 80577
2009-08-31 08:09:09 +00:00
b12b33051d
llvm-mc: Remove MCAsmParser::Parse[Paren]RelocatableExpression.
...
llvm-svn: 80576
2009-08-31 08:08:50 +00:00
73da11ebda
llvm-mc: Switch MCInst to storing an MCExpr* instead of an MCValue.
...
Also, use MCInst::print instead of custom code in MCAsmPrinter.
llvm-svn: 80575
2009-08-31 08:08:38 +00:00
7c82d56930
llvm-mc: Add MCAsmParser::Parse[Paren]Expression forms which return an MCExpr.
...
llvm-svn: 80574
2009-08-31 08:08:17 +00:00
945e257e99
llvm-mc: Make MCSymbolData symbol member const.
...
llvm-svn: 80573
2009-08-31 08:08:06 +00:00
02c7254f6d
llvm-mc: Add MCContext to MCAssembler.
...
llvm-svn: 80572
2009-08-31 08:07:55 +00:00
940cda2590
llvm-mc: Add MCAsmParser::getContext.
...
llvm-svn: 80571
2009-08-31 08:07:44 +00:00
b34a2b9da1
llvm-mc: Add MCExpr::{dump,print}.
...
llvm-svn: 80570
2009-08-31 08:07:33 +00:00
f363645da2
llvm-mc: Switch MCExpr construction to using static member functions, and taking the MCContext (which now owns all MCExprs).
...
llvm-svn: 80569
2009-08-31 08:07:22 +00:00
aecc936ea0
llvm-mc: Add some doxyment markers.
...
llvm-svn: 80568
2009-08-31 08:07:08 +00:00
115e4d6d7b
llvm-mc: Move AsmExpr into MC lib (as MCExpr).
...
llvm-svn: 80567
2009-08-31 08:06:59 +00:00
eedcd84803
Step #1 to giving Callgraph some sane invariants. The problems with callgraph
...
stem from the fact that we have two types of passes that need to update it:
1. callgraphscc and module passes that are explicitly aware of it
2. Functionpasses (and loop passes etc) that are interlaced with CGSCC passes
by the CGSCC Passmgr.
In the case of #1 , we can reasonably expect the passes to update the call
graph just like any analysis. However, functionpasses are not and generally
should not be CG aware. This has caused us no end of problems, so this takes
a new approach. Logically, the CGSCC Pass manager can rescan every function
after it runs a function pass over it to see if the functionpass made any
updates to the IR that affect the callgraph. This allows it to catch new calls
introduced by the functionpass.
In practice, doing this would be slow. This implementation keeps track of
whether or not the current scc is dirtied by a function pass, and, if so,
delays updating the callgraph until it is actually needed again. This was
we avoid extraneous rescans, but we still have good invariants when the
callgraph is needed.
Step #2 of the "give Callgraph some sane invariants" is to change CallGraphNode
to use a CallBackVH for the callsite entry of the CallGraphNode. This way
we can immediately remove entries from the callgraph when a FunctionPass is
active instead of having dangling pointers. The current pass tries to tolerate
these dangling pointers, but it is just an evil hack.
This is related to PR3601/4835/4029. This also reverts r80541, a hack working
around the sad lack of invariants.
llvm-svn: 80566
2009-08-31 07:23:46 +00:00
ff5f1e4d70
fix some cases where instcombine would change hte IR but not return true
...
from runOnFunction
llvm-svn: 80562
2009-08-31 06:57:37 +00:00
7e4fbddf39
cleanups, factor some code out to a helper function
...
llvm-svn: 80542
2009-08-31 06:01:21 +00:00
b822abd37d
fix a crash building SPASS by tolerating a callsite that doesn't exist
...
in the callgraph, see the big comment at the top of the testcase.
llvm-svn: 80541
2009-08-31 05:46:59 +00:00
9e50747958
comment and simplify some code.
...
llvm-svn: 80540
2009-08-31 05:34:32 +00:00
70ebbc59f3
add -debug output
...
llvm-svn: 80539
2009-08-31 05:22:48 +00:00
19dd315e67
improve -debug output, so that -debug is more likely to print when
...
instcombine is changing stuff.
llvm-svn: 80538
2009-08-31 05:17:58 +00:00
4e3e930743
fix a bug I introduced with my 'instcombine builder' refactoring
...
changes: SimplifyDemandedBits can't use the builder yet because it
has the wrong insertion point. This fixes a crash building
MultiSource/Benchmarks/PAQ8p
llvm-svn: 80537
2009-08-31 04:36:22 +00:00
2f2110affa
simplify some code by making the SCCNodes set contain Function*'s
...
instead of CallGraphNode*'s. This also papers over a callgraph
problem where a pass (in this case, MemCpyOpt) introduces a new
function into the module (llvm.memset.i64) but doesn't add it to
the call graph (nor should it, since it is a function pass).
While it might be a good idea for MemCpyOpt to not synthesize
functions in a runOnFunction(), there is no need for FunctionAttrs
to be boneheaded, so fix it there. This fixes an assertion building
176.gcc.
llvm-svn: 80535
2009-08-31 04:09:04 +00:00
edcbfe8529
only print the override triple if it exists!
...
llvm-svn: 80534
2009-08-31 03:22:35 +00:00
081375bb08
Fix PR4834, a tricky case where the inliner would resolve an
...
indirect function pointer, inline it, then go to delete the body.
The problem is that the callgraph had other references to the function,
though the inliner had no way to know it, so we got a dangling pointer
and an invalid iterator out of the deal.
The fix to this is pretty simple: stop the inliner from deleting the
function by knowing that there are references to it. Do this by making
CallGraphNodes contain a refcount. This requires moving deletion of
available_externally functions to the module-level cleanup sweep where
it belongs.
llvm-svn: 80533
2009-08-31 03:15:49 +00:00
4275cf2efa
use an accessor instead of poking internals of a node.
...
llvm-svn: 80532
2009-08-31 02:24:20 +00:00
c6db3ee94b
CMake: updated library dependencies.
...
llvm-svn: 80531
2009-08-31 01:58:50 +00:00
ce713134b8
PR4747
...
Shared landing pads run into trouble with SJLJ, as the dispatch table is
mapped to call sites, and merging the pads will throw that off. There needs
to be a one-to-one mapping of landing pad exception table entries to invoke
call points.
Detecting the shared pad during lowering of SJLJ info insn't sufficient, as
the dispatch function may still need separate destinations to properly
handle phi-nodes.
llvm-svn: 80530
2009-08-31 01:35:03 +00:00
b1c3df239d
update unit test for previous change.
...
llvm-svn: 80528
2009-08-31 00:28:46 +00:00
305b115a87
Fix some nasty callgraph dangling pointer problems in
...
argpromotion and structretpromote. Basically, when replacing
a function, they used the 'changeFunction' api which changes
the entry in the function map (and steals/reuses the callgraph
node).
This has some interesting effects: first, the problem is that it doesn't
update the "callee" edges in any callees of the function in the call graph.
Second, this covers for a major problem in all the CGSCC pass stuff, which
is that it is completely broken when functions are deleted if they *don't*
reuse a CGN. (there is a cute little fixme about this though :).
This patch changes the protocol that CGSCC passes must obey: now the CGSCC
pass manager copies the SCC and preincrements its iterator to avoid passes
invalidating it. This allows CGSCC passes to mutate the current SCC. However
multiple passes may be run on that SCC, so if passes do this, they are now
required to *update* the SCC to be current when they return.
Other less interesting parts of this patch are that it makes passes update
the CG more directly, eliminates changeFunction, and requires clients of
replaceCallSite to specify the new callee CGN if they are changing it.
llvm-svn: 80527
2009-08-31 00:19:58 +00:00
a64b1c1d4c
Fix header comment for bindings/ocaml/llvm/Makefile.
...
llvm-svn: 80526
2009-08-30 23:41:20 +00:00
f34cb0cab4
Make sure we specify no arguments for context functions.
...
llvm-svn: 80525
2009-08-30 23:38:06 +00:00
69349a55e7
add a dump() method on callgraph.
...
llvm-svn: 80524
2009-08-30 22:24:32 +00:00
394b95c005
rename test
...
llvm-svn: 80523
2009-08-30 22:14:17 +00:00
71d51a2d79
merge all sinking tests into one and convert them to filecheck.
...
llvm-svn: 80522
2009-08-30 22:13:26 +00:00
362654f51c
convert scalar_promote to filecheck style and merge 2003-12-13-VolatilePromote.ll into it.
...
llvm-svn: 80521
2009-08-30 22:08:19 +00:00
b284f7b1d9
eliminate some uses of prcontext. Any help here would be appreciated :)
...
llvm-svn: 80520
2009-08-30 21:45:23 +00:00
2d3e0a35bd
rename test so that name reflects what it is testing for.
...
llvm-svn: 80519
2009-08-30 21:36:39 +00:00
c7d5796af8
convert to filecheck format.
...
llvm-svn: 80518
2009-08-30 21:36:06 +00:00
9e371f4b8b
suck a bunch more gep tests into getelementptr.ll and filecheckize them all.
...
llvm-svn: 80517
2009-08-30 21:31:34 +00:00
60745fe8e6
Tweak comment.
...
llvm-svn: 80516
2009-08-30 21:14:05 +00:00
bf75ffd1e2
Fix some possible-use-of-uninitialized warnings.
...
llvm-svn: 80515
2009-08-30 21:13:58 +00:00
89a9bb77e2
consolodate various GEP tests into getelementptr.ll using filecheck.
...
llvm-svn: 80514
2009-08-30 21:02:36 +00:00
2708a312db
another huge testcase, this time from 'gs' in llvm-test.
...
llvm-svn: 80513
2009-08-30 21:02:02 +00:00
3e1b67f62f
remove another poorly-reduced testcase which came from ldecod in llvm-test.
...
llvm-svn: 80512
2009-08-30 21:01:14 +00:00
8a2905e132
this testcase is 500 lines long and is distilled from bzip2, just
...
remove it.
llvm-svn: 80511
2009-08-30 21:00:11 +00:00
4698e19cba
convert to filecheck
...
llvm-svn: 80510
2009-08-30 20:48:15 +00:00
73913f4cd3
Fix PR4748: don't fold gep(bitcast(x)) into bitcast(gep) when x
...
is itself a bitcast. Since we have gep(bitcast(bitcast(y))) in this
case, just wait for the two bitcasts to get zapped. This prevents
instcombine from confusing some aliasing stuff, and allows it to
directly eliminate the load in the testcase.
llvm-svn: 80508
2009-08-30 20:38:21 +00:00
c2f2cf896e
misc cleanup
...
llvm-svn: 80507
2009-08-30 20:36:46 +00:00
a3e620caba
add getPointerAddressSpace() to GEP instruction, use the method
...
in a few scalar xforms to simplify things.
llvm-svn: 80506
2009-08-30 20:06:40 +00:00
c856539edf
eliminate InsertCastBefore, use the builder instead.
...
llvm-svn: 80505
2009-08-30 20:01:10 +00:00
606da5fed8
eliminate InsertBitCastBefore, just use the builder instead.
...
llvm-svn: 80504
2009-08-30 19:47:22 +00:00
ce3b538a7b
add a "getPointerAddressSpace" helper method to LoadInst and StoreInst.
...
llvm-svn: 80503
2009-08-30 19:45:21 +00:00
3681144bd8
Add missed pattern
...
llvm-svn: 80502
2009-08-30 19:06:39 +00:00
5966341a2e
convert a bunch more calls to InsertNewInstBefore to use
...
the new Instcombine builder.
llvm-svn: 80501
2009-08-30 18:50:58 +00:00
8326d529da
fix typo
...
llvm-svn: 80500
2009-08-30 17:53:59 +00:00
947365f73f
hopefully unbreak the build by making this-> explicit for dependent
...
base class lookup.
llvm-svn: 80499
2009-08-30 17:44:10 +00:00
eab572a8ff
EXTRACT_VECTOR_ELEMENT can have result type different from element type.
...
Remove the assertion and generalize the code for ARM NEON stuff.
llvm-svn: 80498
2009-08-30 17:14:54 +00:00
e14d4cdb5f
Add regular expression matching support, based on OpenBSD regexec()/regcomp()
...
implementation.
llvm-svn: 80493
2009-08-30 08:24:09 +00:00
022a582de2
give instcombine a custom IRBuilder that adds new instructions to the
...
workslist and is set to insert new instructions before the current one.
Convert a bunch of stuff that used to call InsertNewInstBefore over to
use it, greatly simplifying code and making it more natural.
There is still a lot more to go, but this is a good start.
llvm-svn: 80492
2009-08-30 07:44:24 +00:00
03686805e8
Update test.
...
llvm-svn: 80490
2009-08-30 07:01:09 +00:00
a0c89ee1da
add a new InstCombineWorklist::AddValue method that works even
...
if the operand is not an instruction.
Simplify most uses of AddOperandsToWorkList to use AddValue and
inline it into the one remaining callsite.
llvm-svn: 80488
2009-08-30 06:27:41 +00:00
bacd05c2eb
move AddUsersToWorkList to the worklist processing class, make the
...
argument stronger typed.
llvm-svn: 80487
2009-08-30 06:22:51 +00:00
795bfdbb55
rename AddUsesToWorkList -> AddOperandsToWorkList. The
...
former looks too much like AddUsersToWorkList and keeps
confusing me.
Remove AddSoonDeadInstToWorklist and change its two callers
to do the same thing in a simpler way.
llvm-svn: 80486
2009-08-30 06:20:05 +00:00
6e8b1a2530
llvm-mc/X86: Encode constant MCValue's correctly.
...
llvm-svn: 80485
2009-08-30 06:17:49 +00:00
6a715dccdf
llvm-mc: MCStreamer cleanups. - Remove EmitLocalSymbol, this is unsupported for now.
...
- Switch Emit{CommonSymbol,Zerofill} to take alignment in bytes (for consistency).
llvm-svn: 80484
2009-08-30 06:17:16 +00:00
905976b1db
inline the trivial AddToWorkList/RemoveFromWorkList methods
...
into their callers. simplify ReplaceInstUsesWith. Make
EraseInstFromFunction only add operands to the worklist if
there aren't too many of them (this was a scalability win
for crazy programs that was only infrequently enforced).
Switch more code to using EraseInstFromFunction instead of
duplicating it inline. Change some fcmp/icmp optimizations
to modify fcmp/icmp in place instead of creating a new one
and deleting the old one just to change the predicate.
llvm-svn: 80483
2009-08-30 06:13:40 +00:00
93ad6170fd
fix a bug I introduced in r80478 found by the build bot.
...
llvm-svn: 80482
2009-08-30 05:56:44 +00:00
97fd3599e1
refactor instcombine's worklist processing stuff out to its own class.
...
llvm-svn: 80481
2009-08-30 05:55:36 +00:00
e08503cd28
make DenseMap::clear() early exit if there is nothing to do.
...
llvm-svn: 80480
2009-08-30 05:55:04 +00:00
520bd0e689
some minor cleanups to IRBuilder, factor the insertion
...
hook out of the main IRBuilder class to allow clients to
override it.
llvm-svn: 80479
2009-08-30 05:42:59 +00:00
b2995e1eb1
more cleanups: remove some redundant code, and simplify some
...
other places.
llvm-svn: 80478
2009-08-30 05:30:55 +00:00
06c687b59e
eliminate the temporary SrcGEPOperands smallvector.
...
llvm-svn: 80477
2009-08-30 05:08:50 +00:00
e26bf17423
simplify/detangle some control flow.
...
llvm-svn: 80476
2009-08-30 05:00:50 +00:00
d7b6e913fe
simplify and cleanup some code, remove some code that just
...
does constant folding of gep's: this is already handled in
a more general way.
No functionality change.
llvm-svn: 80475
2009-08-30 04:49:01 +00:00
65c6a85cfd
default count-aa to -print-all. The whole reason to use count-aa is
...
to see what queries are being made by a transformation, we might as well
default to printing them.
llvm-svn: 80474
2009-08-30 04:25:40 +00:00
dc09af3ef5
Nuke moribund "std::string" version of EOL(..., Encoding).
...
llvm-svn: 80466
2009-08-30 00:28:57 +00:00
0dfe73ac9e
Remove an unnecessary Context argument.
...
llvm-svn: 80454
2009-08-29 23:39:38 +00:00
399d6ae4f7
Minor logic simplification.
...
llvm-svn: 80453
2009-08-29 23:37:49 +00:00
311d06902a
Add some comments.
...
llvm-svn: 80452
2009-08-29 23:36:57 +00:00
062d378603
Cleanup whitespace and indentation.
...
llvm-svn: 80451
2009-08-29 23:35:16 +00:00
07d0a55a9b
Remove some unused fields.
...
llvm-svn: 80450
2009-08-29 23:34:14 +00:00
ca73326f56
CMOV_GR8 clobbers EFLAGS when its expansion involves an xor to set
...
a register to 0. This fixes PR4814.
llvm-svn: 80445
2009-08-29 22:19:15 +00:00
b83f691931
Inline empty destructor.
...
llvm-svn: 80431
2009-08-29 13:38:21 +00:00
06a6057bbe
Fix warning about non-virtual destructor.
...
llvm-svn: 80429
2009-08-29 12:31:38 +00:00
39bb29f7fe
- Add target lowering methods to get the preferred format for the FDE and LSDA
...
encodings.
- Make some of the values emitted by the FDEs dependent upon the pointer
size. This is in line with how GCC does things. And it has the benefit of
working for Darwin in 64-bit mode now.
llvm-svn: 80428
2009-08-29 12:20:54 +00:00
f8b28e4327
Add a form of EOL which emits the text version of a DWARF format encoding. This
...
doesn't handle all values of the formatting. Those can be added as needed.
llvm-svn: 80427
2009-08-29 12:17:53 +00:00
57b946c94b
PR4795: Remove EEVT::isFP, isInt and isVec types used by TableGen's type
...
inferencing. As far as I can tell, these are equivalent to the existing
MVT::fAny, iAny and vAny types, and having both of them makes it harder
to reason about and modify the type inferencing code.
The specific problem in PR4795 occurs when updating a vAny type to be fAny
or iAny, or vice versa. Both iAny and fAny include vector types -- they
intersect with the set of types represented by vAny. When merging them,
choose fAny/iAny to represent the intersection. This is not perfect, since
fAny/iAny also include scalar types, but it is good enough for TableGen's
type inferencing.
llvm-svn: 80423
2009-08-29 05:53:25 +00:00
743dc0ef40
Make the augmentation size and next set of bytes agree on size,
...
and make the reference pointer size as it should be.
Fixes an abort on a testcase derived from libunwind's personality
test in 64-bit.
llvm-svn: 80414
2009-08-29 01:12:46 +00:00
9e955171ca
Fix OProfile support after r80406 changed the DebugInfo interface from
...
GlobalVariables to MDNodes.
llvm-svn: 80411
2009-08-29 00:44:16 +00:00
7cb103dc8b
Another stab at fixing up register kill flags after post-RA scheduling.
...
llvm-svn: 80410
2009-08-29 00:11:13 +00:00
ece642a54c
Do not assert on too wide splats we don't support.
...
llvm-svn: 80409
2009-08-29 00:08:18 +00:00
cd41d07f29
Add missed extract_element pattern
...
llvm-svn: 80408
2009-08-28 23:41:26 +00:00
80ae34974b
Reapply 79977.
...
Use MDNodes to encode debug info in llvm IR.
llvm-svn: 80406
2009-08-28 23:24:31 +00:00
43b9ca6f42
Let Darwin linker auto-synthesize stubs and lazy-pointers. This deletes a bunch of nasty code in ARM asm printer.
...
llvm-svn: 80404
2009-08-28 23:18:09 +00:00
d517ac0682
Nuke trailing whitespace.
...
llvm-svn: 80401
2009-08-28 22:33:43 +00:00
8d799c5144
Fix creation of .bca libraries with EXPORTED_SYMBOLS_LIST, this was putting the
...
llvm-ld shell wrapper script in the archive, not the relinked object!
Also, rename the temp file to avoid conflicts.
llvm-svn: 80373
2009-08-28 16:14:46 +00:00
7026421d61
Read Makefile.config before using LLVMGCC.
...
See http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090824/085828.html
llvm-svn: 80372
2009-08-28 16:13:45 +00:00
d70eebee51
install-sh chmods to 0755 by default, and this causes 'git diff' to show
...
that all the Makefiles changed mode.
Fix this by tellint install-sh to chmod
only to 0644, these are not executable files after all!
llvm-svn: 80371
2009-08-28 16:12:48 +00:00
f43ab91486
Short-term workaround for frame-related weirdness on win64.
...
Some other minor win64 fixes as well.
Patch by Michael Beck!
llvm-svn: 80370
2009-08-28 16:06:41 +00:00
e8a5863c3f
rm needs -f
...
llvm-svn: 80363
2009-08-28 14:05:07 +00:00
483691a185
Remove the llvmprof.out from the test output, otherwise running
...
make check in a non-clean directory causes it to fail (for example when running
make check twice), since execution counts will differ.
llvm-svn: 80362
2009-08-28 13:35:44 +00:00
991beb9aaf
Preparation for Optimal Edge Profiling:
...
This implements the maximum spanning tree algorithm on CFGs according to
weights given by the ProfileEstimator. This is then used to implement Optimal
Edge Profiling.
llvm-svn: 80358
2009-08-28 11:28:24 +00:00
ed3a6bedca
Remove profiling output file because two consecutive runs of make check give
...
error.
llvm-svn: 80357
2009-08-28 10:38:26 +00:00
d8e04ab883
Removed unnecessary file creation during test.
...
llvm-svn: 80356
2009-08-28 10:07:41 +00:00
52afe22566
Pulled all tests into one test. Removed some redundant tests. Rename.
...
llvm-svn: 80355
2009-08-28 10:00:28 +00:00
3033db2448
Fix -Asserts warning, round two.
...
llvm-svn: 80354
2009-08-28 08:08:22 +00:00
2a2459a291
llvm-mc: .lsym is more unsupported than unimplemented, pending a use case appearing.
...
Also, all one of the file level flags are implemented.
llvm-svn: 80352
2009-08-28 07:08:47 +00:00
2701eee0ad
llvm-mc: Support .comm emission.
...
llvm-svn: 80351
2009-08-28 07:08:35 +00:00
4047b53a40
Print a nl before pic labels so they start at a new line. This makes assembly more readable.
...
llvm-svn: 80350
2009-08-28 06:59:37 +00:00
5155fc1b69
Since all std::cout is gone, also remove iostream include.
...
(See http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090824/085620.html )
llvm-svn: 80349
2009-08-28 06:48:25 +00:00
d7f3569555
Readded test from r79615, this tests the complete profiling tool chain. Furhter
...
tests can test only parts of this system.
llvm-svn: 80348
2009-08-28 06:41:00 +00:00
42a39d038a
llvm-mc: Support .zerofill emission.
...
- I'm still trying to figure out the cleanest way to implement this and match the assembler, currently there are some substantial differences.
llvm-svn: 80347
2009-08-28 05:49:21 +00:00
a468a12dfa
llvm-mc: Tweak section alignment and size computation to match 'as' closer.
...
llvm-svn: 80345
2009-08-28 05:49:04 +00:00
d82ef0e923
llvm-mc: Factor getSectionData out of SwitchSection.
...
llvm-svn: 80344
2009-08-28 05:48:54 +00:00
96abad185d
llvm-mc: Emit .lcomm as .zerofill.
...
llvm-svn: 80343
2009-08-28 05:48:46 +00:00
e48a69b2a3
llvm-mc: Unique zero fill sections.
...
llvm-svn: 80342
2009-08-28 05:48:29 +00:00
cf72e1c03e
llvm-mc: Add const to EmitZeroFill section argument.
...
llvm-svn: 80341
2009-08-28 05:48:22 +00:00
93665faad0
llvm-mc: Fix thinko in emitting values.
...
llvm-svn: 80340
2009-08-28 05:48:10 +00:00
4b07ea1902
Add MathExtras.h OffsetToAlignment, like RoundUpToAlignment but returns the
...
offset to the next aligned integer.
llvm-svn: 80339
2009-08-28 05:48:04 +00:00
d46e3466e7
Fix -Asserts warning.
...
llvm-svn: 80338
2009-08-28 05:47:56 +00:00
0e8901803c
finish a half formed thought :)
...
llvm-svn: 80334
2009-08-28 04:48:54 +00:00
2ece038e6e
Another NO_RUNTIME_LIBRARIES tweak...
...
llvm-svn: 80331
2009-08-28 03:06:28 +00:00
e45cdcdca2
Tweak NO_RUNTIME_LIBS.
...
llvm-svn: 80330
2009-08-28 02:20:39 +00:00
51a25b06f7
Mark Andersen's as experimental.
...
llvm-svn: 80328
2009-08-28 00:45:47 +00:00
bda82c20f3
Fix PR3913, patch by Jakub Staszak!
...
llvm-svn: 80327
2009-08-28 00:43:14 +00:00
6da267de23
v4, v5 does not support sxtb / sxth.
...
llvm-svn: 80322
2009-08-28 00:31:43 +00:00
2cb0c3b5cb
Disable optional bindings for Apple-style builds.
...
llvm-svn: 80319
2009-08-28 00:21:13 +00:00
d581639645
Revert 76080. This broke some powerpc cross compiles.
...
It also makes the llvmCore build dependent on whatever version of llvm-gcc
happens to be installed on the build machine.
llvm-svn: 80316
2009-08-28 00:10:15 +00:00
8d38b8d20c
Revert r80305, I forgot a dependent change.
...
--- Reverse-merging r80305 into '.':
U tools/llvm-mc/AsmParser.cpp
llvm-svn: 80309
2009-08-27 23:58:10 +00:00
3082c01d68
Closure is a very generic name. Use AppleBlock instead.
...
llvm-svn: 80307
2009-08-27 23:51:51 +00:00
ba33ff294e
llvm-mc: Unique sections in .zerofill.
...
llvm-svn: 80305
2009-08-27 23:45:06 +00:00
57d9b1e19a
eliminate all 80-col violations that I have introduced in my recent checkins (and some others more)
...
llvm-svn: 80304
2009-08-27 23:44:33 +00:00
e36c50f163
Don't build runtime libraries in an Apple style build.
...
llvm-svn: 80303
2009-08-27 23:43:28 +00:00
e91191630b
Revert 80278 for now, it caused a lot of MIPS tests to fail
...
llvm-svn: 80280
2009-08-27 19:57:56 +00:00
18c21daafb
Revamp our friend Mips :)
...
Add MO flags to simplify the printing of relocations.
Remove the support for printing large code model relocs (which
aren't supported anyway).
llvm-svn: 80278
2009-08-27 19:40:40 +00:00
931ec72ebf
Don't mark CMOV_GR8 as two-address, or commutable, since it's a pseudo.
...
llvm-svn: 80271
2009-08-27 18:16:24 +00:00
ed10d7c441
Adjust the MachineBasicBlock verifier rules to be more
...
tolerant of blocks that end with "unreachable".
llvm-svn: 80270
2009-08-27 18:14:26 +00:00
69baa5139b
Add FIXME for when we support more specific XMM registers.
...
llvm-svn: 80269
2009-08-27 18:08:16 +00:00
799c92bcfb
Nuke trailing whitespace.
...
llvm-svn: 80268
2009-08-27 18:07:15 +00:00
bc1d841a63
Use stripPointerCasts instead of doing the same manually.
...
llvm-svn: 80267
2009-08-27 18:02:03 +00:00
ea398633b2
Minor code simplification.
...
llvm-svn: 80266
2009-08-27 17:59:08 +00:00
4ce5ade09d
Teach getUnderlyingObject and skipPointerCasts about GlobalAliases.
...
llvm-svn: 80265
2009-08-27 17:55:13 +00:00
aba809befd
Be somewhat more consistent about const qualifiers.
...
llvm-svn: 80264
2009-08-27 17:54:15 +00:00
1f59f0146f
Global Aliases are not identifiable objects.
...
llvm-svn: 80263
2009-08-27 17:52:56 +00:00
05f11351ab
Handle TargetData with const.
...
llvm-svn: 80262
2009-08-27 17:51:25 +00:00
20ef7728af
Strip trailing whitespace from blank lines.
...
llvm-svn: 80259
2009-08-27 17:29:49 +00:00
962c374781
re-apply r80197, now that iterator.h is not mentioned any more
...
llvm-svn: 80254
2009-08-27 17:07:35 +00:00
ba53af58f0
Hopefully the final missing part :(
...
scalar_to_vector is fully legal now
llvm-svn: 80251
2009-08-27 16:25:49 +00:00
076f105d86
Forgot about actual change :)
...
llvm-svn: 80250
2009-08-27 16:10:17 +00:00
205cac837f
scalar_to_vector is fully legal now (implemented as subreg accesses)
...
llvm-svn: 80249
2009-08-27 16:04:47 +00:00
7b8f61f3c8
Remove an entry from ValueMap before removing elements.
...
llvm-svn: 80247
2009-08-27 15:32:38 +00:00
d0b0262edf
Ok, sometimes it's profitable to turn scalar_to_vector stuff into subreg access.
...
Add a testcase.
llvm-svn: 80246
2009-08-27 14:51:42 +00:00
58ebae4acd
Transform float scalar_to_vector into subreg accesses.
...
No idea whether this is profitable or not.
llvm-svn: 80245
2009-08-27 14:38:44 +00:00
209baa5c48
STRD and LDRD require ARMv5TE, not just ARMv5T.
...
See http://llvm.org/PR4687 for more info and links.
llvm-svn: 80244
2009-08-27 14:14:21 +00:00
53fd9f392d
Inverse logic to increase portability.
...
llvm-svn: 80240
2009-08-27 12:02:34 +00:00
99c20d1459
To make mcc16 run correctly on mac.
...
llvm-svn: 80239
2009-08-27 11:54:38 +00:00
c7c5f9fa86
llvm-mc/Mach-O: Add MCCodeEmitter support, for encoding instructions.
...
- No relocations yet, of course.
llvm-svn: 80235
2009-08-27 08:17:51 +00:00
981a71c302
llvm-mc/X86: Implement single instruction encoding interface for MC.
...
- Note, this is a gigantic hack, with the sole purpose of unblocking further
work on the assembler (its also possible to test the mathcer more completely
now).
- Despite being a hack, its actually good enough to work over all of 403.gcc
(although some encodings are probably incorrect). This is a testament to the
beauty of X86's MachineInstr, no doubt! ;)
llvm-svn: 80234
2009-08-27 08:12:55 +00:00
c4f0e7e28d
llvm-mc: Print encodings after the instruction, and only when we have an asm
...
printer.
llvm-svn: 80233
2009-08-27 07:58:57 +00:00
6ec97df98f
X86: Mark EH_RETURN as code-gen-only.
...
llvm-svn: 80232
2009-08-27 07:58:05 +00:00
212b6d8a57
Add {MCInst,MCOperand}::{print,dump}
...
llvm-svn: 80231
2009-08-27 07:57:12 +00:00
d18dd1d5f8
llvm-mc: Only show instruction encodings with --show-encoding.
...
llvm-svn: 80230
2009-08-27 07:56:39 +00:00
1c62958901
the buildbots revealed one more breakage. fix. (why didn't I see this?)
...
llvm-svn: 80227
2009-08-27 06:59:20 +00:00
95225a2fe6
Clean up the minor mess I caused with removing iterator.h. I shall take care of 80-col violations and the FIXME later. (Thanks goodness that I live in another continent, so the monkeypox did not strike me :-)
...
llvm-svn: 80224
2009-08-27 06:41:46 +00:00
d3374e8dfd
Implement a new optimization in the inliner: if inlining multiple
...
calls into a function and if the calls bring in arrays, try to merge
them together to reduce stack size. For example, in the testcase
we'd previously end up with 4 allocas, now we end up with 2 allocas.
As described in the comments, this is not really the ideal solution
to this problem, but it is surprisingly effective. For example, on
176.gcc, we end up eliminating 67 arrays at "gccas" time and another
24 at "llvm-ld" time.
One piece of concern that I didn't look into: at -O0 -g with
forced inlining this will almost certainly result in worse debug
info. I think this is acceptable though given that this is a case
of "debugging optimized code", and we don't want debug info to
prevent the optimizer from doing things anyway.
llvm-svn: 80215
2009-08-27 06:29:33 +00:00
6802a2056a
the inliner shouldn't crash on this.
...
llvm-svn: 80214
2009-08-27 06:20:45 +00:00
a1cb9ba541
For now, only run MC tests if X86 is configured.
...
llvm-svn: 80213
2009-08-27 06:11:15 +00:00
10cdbd3b73
This is passing for PPC on Mac OS X.
...
llvm-svn: 80210
2009-08-27 05:49:50 +00:00
51c58b8131
Revert accidental commit.
...
llvm-svn: 80208
2009-08-27 05:35:28 +00:00
1ce61b82ac
unbreak the build, yay for symlinks + makefiles. :(
...
llvm-svn: 80205
2009-08-27 04:43:05 +00:00
b9d0a961f9
reduce header #include'age
...
llvm-svn: 80204
2009-08-27 04:32:07 +00:00
b1cba3f91e
enhance InlineFunction to be able to optionally return
...
a the list of static allocas that it inlined.
llvm-svn: 80203
2009-08-27 04:20:52 +00:00
d84dbb3443
smallvectorize the list of returns built by CloneAndPruneFunctionInto.
...
llvm-svn: 80202
2009-08-27 04:02:30 +00:00
9d0235dc6b
remove CloneTrace, which appears to be dead since 2004.
...
llvm-svn: 80201
2009-08-27 03:56:43 +00:00
5eef6ad6a9
reduce inlining factor some stuff out to a static helper function,
...
and other code cleanups. No functionality change.
llvm-svn: 80199
2009-08-27 03:51:50 +00:00
e5b5d293e9
Revert accidental commit.
...
llvm-svn: 80198
2009-08-27 03:32:50 +00:00
c7d230f736
--- Reverse-merging r80147 into '.':
...
A include/llvm/ADT/iterator.cmake
U autoconf/configure.ac
--- Reverse-merging r80161 into '.':
U cmake/config-ix.cmake
--- Reverse-merging r80171 into '.':
U Makefile
--- Reverse-merging r80173 into '.':
U configure
U include/llvm/Config/config.h.in
--- Reverse-merging r80180 into '.':
A include/llvm/ADT/iterator.h.in
Despite common miscomceptions, iterator.h is alive and well. It broke the build
bots for several hours. And yet no one bothered to look at them.
Gabor and Doug, please review your changes and make sure that they actually
build before resubmitting them.
llvm-svn: 80197
2009-08-27 03:29:26 +00:00
352a49586d
Add some checks for MachineCFG consistency. Use AnalyzeBranch and
...
do extra checking when it succeeds, as those are cases where
CodeGen will be doing particularly interesting CFG modifications.
llvm-svn: 80196
2009-08-27 02:43:49 +00:00
9cefb135a5
Update CMake
...
llvm-svn: 80195
2009-08-27 02:08:37 +00:00
81cb753298
llvm-mc: Tweak MCCodeEmitter skeleton.
...
llvm-svn: 80193
2009-08-27 01:34:22 +00:00
fb3d84b524
Initialize the PoisonMemory member before initializing
...
members that call methods that read the PoisonMemory member.
This fixes potential spurious (though probably otherwise
harmless) poising of unused memory, and fixes the
associated valgrind error.
llvm-svn: 80192
2009-08-27 01:25:57 +00:00
7a37b1a2ca
Fix PR4789. Teach eliminateFrameIndex how to handle VLDRQ and VSTRQ which cannot fold any immediate offset.
...
llvm-svn: 80191
2009-08-27 01:23:50 +00:00
ab3912f55b
Add missing declarations.
...
llvm-svn: 80190
2009-08-27 01:15:54 +00:00
65105174a6
Sketch TargetRegistry support for MCCodeEmitter abstract interface.
...
- Of course, nothing actually can provide this interface yet.
llvm-svn: 80188
2009-08-27 00:51:57 +00:00
47f11ac5b5
Try to make MSVC just a little happier.
...
llvm-svn: 80187
2009-08-27 00:38:04 +00:00
7f0ca9a34c
X86FastISel support for loading and storing values of type i1.
...
llvm-svn: 80186
2009-08-27 00:31:47 +00:00
f1abb5511b
Expand i8 selects into control flow instead of 16-bit conditional
...
moves. This avoids the need to promote the operands (or implicitly
extend them, a partial register update condition), and can reduce
i8 register pressure. This substantially speeds up code such as
write_hex in lib/Support/raw_ostream.cpp.
subclass-coalesce.ll is too trivial and no longer tests what it was
originally intended to test.
llvm-svn: 80184
2009-08-27 00:14:12 +00:00
9bd34b3c40
now that include/llvm/ADT/iterator.h.in is not attempted to be built, its prerequisite can go away too
...
llvm-svn: 80180
2009-08-26 23:35:33 +00:00
b039caa1c3
Simplify.
...
llvm-svn: 80176
2009-08-26 23:12:33 +00:00
985c9f418a
Regenerate configure
...
llvm-svn: 80173
2009-08-26 22:59:05 +00:00
a9ced1d3ae
eliminate references to ADT/iterator.h
...
llvm-svn: 80171
2009-08-26 22:55:19 +00:00
109ca5a14a
Make this into a static method.
...
llvm-svn: 80170
2009-08-26 22:55:11 +00:00
904d4c71d5
Fix bad length argument to substr calls. Apparently I'm the first one to
...
attempt more than 2 constraints on an instruction.
llvm-svn: 80169
2009-08-26 22:50:39 +00:00
4d7b2e3307
llvm-mc/Mach-O: Unique sections properly, so we don't get duplicate text
...
sections, etc.
- The quick and dirty way, just clone the TargetLoweringObjectFile
code. Eventually this should be shared... somehow.
llvm-svn: 80168
2009-08-26 22:49:51 +00:00
e73b267301
llvm-mc/Mach-O: Don't put assembler temporary labels in the symbol table.
...
- I moved section creation back into AsmParser. I think policy decisions like
this should be pushed higher, not lower, when possible (in addition the
assembler has flags which change this behavior, for example).
llvm-svn: 80162
2009-08-26 22:13:22 +00:00
7e3778948a
De-bork CMake build
...
llvm-svn: 80161
2009-08-26 22:12:02 +00:00
9233565a0a
Rework getPersonalityIndex slightly - 0 is now a valid and not-NULL
...
personality function.
llvm-svn: 80153
2009-08-26 21:44:57 +00:00
36595c8653
nobody includes llvm/ADT/iterator.h any more,
...
so get rid of this monstrosity. iterator.h.in is scheduled for deletion in my working copy,
but I wait till I see that configure gets regenerated, as it depends on it. I'll commit
then.
There are still some AC_* tests in the configure.ac dealing with iterators, those can
be zapped probably too.
llvm-svn: 80147
2009-08-26 21:36:59 +00:00
a258c62ea1
If we're emitting additional CIEs due to personality functions
...
don't emit the default one. Explicitly check for the NULL
CIE later.
llvm-svn: 80146
2009-08-26 21:30:49 +00:00
c2645717ca
Nuke trailing whitespace.
...
llvm-svn: 80145
2009-08-26 21:27:09 +00:00
e2697738a4
llvm-mc/Mach-O: Set .subsections_via_symbols flag properly.
...
llvm-svn: 80144
2009-08-26 21:22:22 +00:00
25eeddf112
Don't submit test directory.
...
llvm-svn: 80139
2009-08-26 21:00:34 +00:00
ab7700479f
Allocate the module provider in the Kaleidoscope code on the heap, not the stack, so that it can be properly deleted. Also update the tutorial with the new code. This fixes PR4762, hopefully better than the last time.
...
llvm-svn: 80138
2009-08-26 20:58:25 +00:00
bdabe5e5ea
Remove all the LLVM_COMPACTIFY_SENTINELS-related macro magic as discussed with Chris on IRC. Anybody wanting to debug sentinel dereferencing problems must revert this patch and perform the indicated modifications.
...
llvm-svn: 80128
2009-08-26 19:16:32 +00:00
3d3fc043f2
CMake: Removed outdated TODO.
...
llvm-svn: 80124
2009-08-26 18:37:05 +00:00
bc2190fe50
Generate section for bss and enable weak symbols
...
llvm-svn: 80121
2009-08-26 18:24:12 +00:00
faebdee4dd
Convert some more Neon tests to FileCheck.
...
llvm-svn: 80120
2009-08-26 18:11:50 +00:00
f582ac7c11
Alter 79292 to produce output that actually assembles.
...
llvm-svn: 80119
2009-08-26 18:10:32 +00:00
6cc5c93ce4
Allow unsigned long long DenseMapInfo.
...
llvm-svn: 80118
2009-08-26 18:02:19 +00:00
e0636a7aed
Remove unneeded ARM-specific DAG nodes for VLD* and VST* Neon operations.
...
The instructions can be selected directly from the intrinsics. We will need
to add some ARM-specific nodes for VLD/VST of 3 and 4 128-bit vectors, but
those are not yet implemented.
llvm-svn: 80117
2009-08-26 17:39:53 +00:00
50509fc2cb
Add extload expansion for f128
...
llvm-svn: 80116
2009-08-26 17:39:40 +00:00
5f6fc970c4
Unbreak FP128 stuff in cbe
...
llvm-svn: 80115
2009-08-26 17:39:23 +00:00
175a6e4a3b
Ignore -fPIC test on Windows. Suggested by Yonggang Luo.
...
llvm-svn: 80111
2009-08-26 17:05:06 +00:00
1bb5d32bdd
Unbreak CMake build
...
llvm-svn: 80109
2009-08-26 16:33:57 +00:00
0f756b27ae
Expand scalar_to_vector - we don't have any isel logic for it now
...
llvm-svn: 80107
2009-08-26 16:26:09 +00:00
45a6b06a45
Add comments detailing a known bug, so that people writing other
...
backends don't use it as an example.
llvm-svn: 80105
2009-08-26 16:06:11 +00:00
b4720905ff
-fast is now -O0. -fast-isel is no longer experimental.
...
llvm-svn: 80104
2009-08-26 15:57:57 +00:00
6942375aa4
Move ProfileInfo::Edge's operator<< out of line. Among other benefits,
...
this eliminates the ATTRIBUTE_USED, which wasn't being used in a manner
acceptable to some GCC versions, according to the buildbots.
llvm-svn: 80103
2009-08-26 15:56:38 +00:00
e685cba880
Bugfix for r80100, forgot include. Sorry.
...
llvm-svn: 80101
2009-08-26 15:18:38 +00:00
73730f9390
Implemented comments from Daniel Dunbar.
...
(See http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090817/084958.html )
llvm-svn: 80100
2009-08-26 15:13:44 +00:00
6c23fa2442
Don't use INSERT_SUBREG to model anyext operations on x86-64, as it
...
leads to partial-register definitions. To help avoid redundant
zero-extensions, also teach the h-register matching patterns that
use movzbl to match anyext as well as zext.
llvm-svn: 80099
2009-08-26 14:59:13 +00:00
d926b985df
Create a ScalarEvolution-based AliasAnalysis implementation.
...
This is a simple AliasAnalysis implementation which works by making
ScalarEvolution queries. ScalarEvolution has a more complete understanding
of arithmetic than BasicAA's collection of ad-hoc checks, so it handles
some cases that BasicAA misses, for example p[i] and p[i+1] within the
same iteration of a loop.
This is currently experimental. It may be that the main use for this pass
will be to help find cases where BasicAA can be profitably extended, or
to help in the development of the overall AliasAnalysis infrastructure,
however it's also possible that it could grow up to become a directly
useful pass.
llvm-svn: 80098
2009-08-26 14:53:06 +00:00
331b99888e
Fix a missing newline, now that Value's operator<< doesn't add one of its own.
...
llvm-svn: 80096
2009-08-26 14:34:12 +00:00
0672e92e92
Use SetVector instead of std::set so that alias relations are tested and
...
printed in a deterministic order.
llvm-svn: 80095
2009-08-26 14:32:17 +00:00
c4581eaca8
llvm-mc/Mach-O: Add support for relocations.
...
- I haven't really tried to find the "right" way to store the fixups or apply
them, yet. This works, but isn't particularly elegant or fast.
- Still no evaluation support, so we don't actually ever not turn a fixup into
a relocation entry.
llvm-svn: 80089
2009-08-26 13:58:10 +00:00
a9eeb78ed5
llvm-mc/Mach-O: Move symbol indices into the MCSymbolData structure.
...
llvm-svn: 80088
2009-08-26 13:57:54 +00:00
b40044444b
llvm-mc/Mach-O: Dump relocations and section data (optionally) in my Mach-O dumper.
...
llvm-svn: 80087
2009-08-26 13:57:44 +00:00
c1710a04ba
llvm-mc: Add symbol entries for undefined symbols used in .fill and .org.
...
llvm-svn: 80086
2009-08-26 13:57:37 +00:00
a0e01bec87
Add dummy inline asm handling for 'r' constraint. This fixes PR4778
...
llvm-svn: 80085
2009-08-26 13:44:29 +00:00
83489058d0
Moved isDeclaration() check further down to allow for function counts for
...
declarations if necessary.
llvm-svn: 80084
2009-08-26 13:33:09 +00:00
ffda9eb31a
llvm-mc: Change MCContext value table to take const MCSymbol*s.
...
llvm-svn: 80079
2009-08-26 09:16:57 +00:00
90398bd222
llvm-mc: Make MCValue take const MCSymbol*s.
...
llvm-svn: 80078
2009-08-26 09:16:46 +00:00
Douglas Gregor