87f77b5f9a
Add hook in MCSection to decide when to use "optimized nops", for each
...
section kind. Previously, optimized nops were only used for MachO.
Also added tests for ELF and COFF.
llvm-svn: 115523
2010-10-04 17:32:41 +00:00
d3593c3a8e
the immediate field of pshufw is actually an 8-bit field, not a 8-bit field that is sign extended. This fixes PR8288
...
llvm-svn: 115473
2010-10-03 19:09:13 +00:00
66e08d43d2
Jim Asked us to move DataLayout on ARM back to the most specialized classes. Do
...
so and also change X86 for consistency.
Investigating if this can be improved a bit.
llvm-svn: 115469
2010-10-03 18:59:45 +00:00
5db3884d94
GCC extensions are no longer used here - update the comment.
...
llvm-svn: 115463
2010-10-03 15:31:50 +00:00
6d80481b68
Remove two uses of the gcc specific 'aligned' attribute. This
...
is partly because this attribute caused trouble in the past (the
SmallVector one had to be changed from aligned to aligned(8) due
to causing crashes on i386 for example; in theory the same might
be needed in the Allocator case...). But it's mostly because
there seems to be no point in special casing gcc here. Using the
same implementation for all compilers results in better testing.
llvm-svn: 115462
2010-10-03 15:15:19 +00:00
28792c4a3d
When RemoveCopyByCommutingDef is creating additional identity copies, just use
...
LiveInterval::MergeValueNumberInto instead of trying to extend LiveRanges and
getting it wrong.
This fixed PR8249 where a valno with a multi-segment live range was defined by
an identity copy created by RemoveCopyByCommutingDef. Some of the live
segments disappeared.
llvm-svn: 115385
2010-10-01 23:52:25 +00:00
e1c714647c
Add support to let FE mark explict methods as explict in debug info.
...
llvm-svn: 115378
2010-10-01 23:31:40 +00:00
f31f33ea89
Thread the determination of branch prediction hit rates back through the if-conversion heuristic APIs. For now,
...
stick with a constant estimate of 90% (branch predictors are good!), but we might find that we want to provide
more nuanced estimates in the future.
llvm-svn: 115364
2010-10-01 22:45:50 +00:00
dd224d2333
Massive rewrite of MMX:
...
The x86_mmx type is used for MMX intrinsics, parameters and
return values where these use MMX registers, and is also
supported in load, store, and bitcast.
Only the above operations generate MMX instructions, and optimizations
do not operate on or produce MMX intrinsics.
MMX-sized vectors <2 x i32> etc. are lowered to XMM or split into
smaller pieces. Optimizations may occur on these forms and the
result casted back to x86_mmx, provided the result feeds into a
previous existing x86_mmx operation.
The point of all this is prevent optimizations from introducing
MMX operations, which is unsafe due to the EMMS problem.
llvm-svn: 115243
2010-09-30 23:57:10 +00:00
a4a83283d5
Comments about operand cycles and pipeline forwarding pathes.
...
llvm-svn: 115214
2010-09-30 22:01:50 +00:00
ddb503e5d1
Did my commit for the last patch for the .loc directory from the wrong place and
...
missed a bunch of files. Here the rest. Sorry about that.
llvm-svn: 115173
2010-09-30 17:16:09 +00:00
bad267fa05
Adds getPointerSize() to the AsmBackend which will be needed by the final patch
...
for the dwarf .loc support to emit dwarf line number tables.
llvm-svn: 115153
2010-09-30 16:38:07 +00:00
efbdbe5565
Move logic of determining ELF entsize from the .s printer to initialization
...
time. That way, the EntrySize field is initialized for other code paths,
namely, the .ll -> .o code path.
llvm-svn: 115141
2010-09-30 05:59:22 +00:00
645f6c2bef
Tiny patch for proof-of-concept cleanup of ARMAsmPrinter::EmitStartOfAsmFile()
...
Small test for sanity check of resulting ARM .s file.
Tested against -r115129.
llvm-svn: 115133
2010-09-30 02:45:56 +00:00
2ebaee9c75
Make it possible for the MCObjectWriter to decide if a given fixup is fully
...
resolved or not. Different object files have different restrictions and
different native assemblers have different idiosyncrasies we want to emulate
for now.
Move the existing MachO logic to the new place and implement an ELF one that
gets fixups to globals right.
llvm-svn: 115131
2010-09-30 02:22:20 +00:00
4a57b76eea
Let a target specify whether it wants an assembly printer to be the MC version
...
or not. TableGen needs to generate the printInstruction() function as taking
an MCInstr* or a MachineInstr*, depending. Default to the old non-MC
version so that everything not yet using MC continues to just work without
fidding.
llvm-svn: 115126
2010-09-30 01:29:54 +00:00
74c06b202f
Revert r115099 (adding early jump threading). It's not clear if the benefits are worth the compile time cost.
...
llvm-svn: 115106
2010-09-29 23:31:09 +00:00
bea08d1c85
Let FE mark a variable as artificial variable.
...
llvm-svn: 115102
2010-09-29 23:07:21 +00:00
6d33f59952
Early CFG simplification can fold conditionals down to selects, which is often a good thing, but it can also
...
hide jump threading opportunities by turning control flow into data flow. Run an early JumpThreading pass
(adds approximately an additional 1% to optimization time on SPEC), allowing it to get a shot at these cases
first. Fixes <rdar://problem/8447345>.
llvm-svn: 115099
2010-09-29 22:57:02 +00:00
4a010fd1ea
Model Cortex-a9 load to SUB, RSB, ADD, ADC, SBC, RSC, CMN, MVN, or CMP
...
pipeline forwarding path.
llvm-svn: 115098
2010-09-29 22:42:35 +00:00
cb03b14089
Add support to let FE encode method access specifier.
...
llvm-svn: 115089
2010-09-29 21:44:16 +00:00
95ae73c394
Generalize DISubprogram element to encode various flags instead of just one boolean for isArtificial.
...
This is a backword compatible change.
llvm-svn: 115084
2010-09-29 21:04:46 +00:00
af995f0ee5
remove PointerTracking from mainline, Edwin is going to move it out to ClamAV
...
for LLVM 2.9
llvm-svn: 115062
2010-09-29 18:43:27 +00:00
a63292a3ca
implement rdar://8456378 and PR7557 - support for the fstsw,
...
an instruction that requires a WHOLE NEW wonderful kind of alias.
llvm-svn: 115015
2010-09-29 01:50:45 +00:00
b44fd24fc1
change the protocol TargetAsmPArser::MatchInstruction method to take an
...
MCStreamer to emit into instead of an MCInst to fill in. This allows the
matcher extra flexibility and is more convenient.
llvm-svn: 115014
2010-09-29 01:42:58 +00:00
0097dd0d5a
Add support to model pipeline bypass / forwarding.
...
llvm-svn: 115005
2010-09-28 23:50:49 +00:00
88af7d00fc
Part one of switching to using a more sane heuristic for determining if-conversion profitability.
...
Rather than having arbitrary cutoffs, actually try to cost model the conversion.
For now, the constants are tuned to more or less match our existing behavior, but these will be
changed to reflect realistic values as this work proceeds.
llvm-svn: 114973
2010-09-28 18:32:13 +00:00
7a55481fa4
Provide an interface to let FEs anchor debug info for types.
...
llvm-svn: 114969
2010-09-28 18:08:20 +00:00
dcd7c2b833
Add a new scope type "LTO_SYMBOL_SCOPE_DEFAULT_CAN_BE_HIDDEN" for the
...
"linker_private_weak_auto_def" linkage type for LTO.
llvm-svn: 114868
2010-09-27 20:17:45 +00:00
d116d8a4e9
MC/AsmParser: Handle exponents in floating point literals.
...
llvm-svn: 114861
2010-09-27 20:12:52 +00:00
cceded493f
Remove unused argument.
...
llvm-svn: 114852
2010-09-27 18:13:03 +00:00
c758311025
Push twines deeper into SourceMgr's error handling methods.
...
llvm-svn: 114847
2010-09-27 17:42:11 +00:00
cf935acf53
writeGraph doesn't need its ShortNames argument.
...
llvm-svn: 114842
2010-09-27 16:59:51 +00:00
68fe970732
Don't construct a redundant GraphWriter object.
...
llvm-svn: 114838
2010-09-27 16:54:13 +00:00
06ccde5f40
Factor out code from the standalone WriteGraph function into a helper
...
function on GraphWriter.
llvm-svn: 114837
2010-09-27 16:44:11 +00:00
ce752acfae
Constify properlyDominates in the same manner as dominates.
...
Add constified overloads for findNearestCommonDominator.
llvm-svn: 114834
2010-09-27 16:33:31 +00:00
a0da88931c
Add support for viewing graphviz graphs with xdot.py.
...
llvm-svn: 114832
2010-09-27 16:28:34 +00:00
b179cb2ce7
Add an all() method to BitVector, for testing whether all bits are set.
...
llvm-svn: 114830
2010-09-27 15:48:37 +00:00
24697d3b8d
Remove WriteGraph's Name argument, which it didn't use, and
...
rename writeHeader's Name argument to Title, to be consistent
with WriteGraph.
llvm-svn: 114829
2010-09-27 15:34:19 +00:00
58a9d28a68
Fix indentation.
...
llvm-svn: 114827
2010-09-27 14:44:14 +00:00
9f06f911d1
the latest assembler that runs on powerpc 10.4 machines doesn't
...
support aligned comm. Detect when compiling for 10.4 and don't
emit an alignment for comm. THis will hopefully fix PR8198.
llvm-svn: 114817
2010-09-27 06:44:54 +00:00
0c7eca334a
Avoid warnings about implicit conversions to `bool' in MSVC. This time
...
for real.
Patch by Nathan Jeffords!
llvm-svn: 114796
2010-09-25 20:27:36 +00:00
10117c762a
Avoid using VNInfo::getCopy as much as possible. I want to get rid of it.
...
llvm-svn: 114794
2010-09-25 18:10:38 +00:00
564956867e
Removed VNInfo::isDefAccurate(). Def "accuracy" can be checked by testing whether LiveIntervals::getInstructionFromIndex(def) returns NULL.
...
llvm-svn: 114791
2010-09-25 12:04:16 +00:00
299479020a
Add ret instruction to PTX backend
...
llvm-svn: 114788
2010-09-25 07:46:17 +00:00
b38a05572a
Remove trailing spaces of CallingConv.h
...
llvm-svn: 114787
2010-09-25 07:02:30 +00:00
bc71af341e
Remove SlotIndex::PHI_BIT. It is no longer used by anything.
...
llvm-svn: 114779
2010-09-25 00:45:18 +00:00
335b9a8ea9
Terminator gaps were unused. Might as well delete them.
...
llvm-svn: 114776
2010-09-24 23:58:56 +00:00
cbb421887d
Attach a DebugLoc to a GC point in order to get precise information in the JIT of a GC point.
...
llvm-svn: 114736
2010-09-24 17:27:50 +00:00
3068a93dc1
MC/Lexer: Add 'Real' token type for floating point literals.
...
llvm-svn: 114718
2010-09-24 01:59:31 +00:00
2c5df619c4
Revert r114703 and r114702, removing the isConditionalMove flag from instructions. After further
...
reflection, this isn't going to achieve the purpose I intended it for. Back to the drawing board!
llvm-svn: 114710
2010-09-23 23:45:25 +00:00
6e0e8d7d64
Add an TargetInstrDesc bit to indicate that a given instruction is a conditional move.
...
Not intended functionality change, as nothing uses this yet.
llvm-svn: 114702
2010-09-23 22:44:10 +00:00
9876dd3040
trailing whitespace
...
llvm-svn: 114680
2010-09-23 18:07:04 +00:00
d5f4130ffb
Avoid warnings about conversions to `bool' in MS compilers.
...
Patch by Nathan Jeffords!
llvm-svn: 114662
2010-09-23 16:59:44 +00:00
fd1bc42230
Moved the PBQP allocator class out of the header and back in to the cpp file to hide the gory details.
...
Allocator instances can now be created by calling createPBQPRegisterAllocator.
Tidied up use of CoalescerPair as per Jakob's suggestions.
Made the new PBQPBuilder based construction process the default. The internal construction process
remains in-place and available via -pbqp-builder=false for now. It will be removed shortly if the new
process doesn't cause any regressions.
llvm-svn: 114626
2010-09-23 04:28:54 +00:00
85dcd3d0f4
Add support for ELF PLT references for ARM MC asm printing. Adding a
...
new VariantKind to the MCSymbolExpr seems like overkill, but I'm not sure
there's a more straightforward way to get the printing difference captured.
(i.e., x86 uses @PLT, ARM uses (PLT)).
llvm-svn: 114613
2010-09-22 23:27:36 +00:00
98d500b024
Make SetVector's remove indicate whether it actually removed something.
...
llvm-svn: 114612
2010-09-22 23:20:04 +00:00
dee12d8110
Avoid some Mach-O specific alignment being done on ELF.
...
llvm-svn: 114594
2010-09-22 22:27:05 +00:00
7bce0596a7
allow target-specific label suffixes, patch by Yuri Gribov!
...
llvm-svn: 114592
2010-09-22 22:19:53 +00:00
71954b8d54
Fix uninitialized TBAAFlag field values.
...
llvm-svn: 114591
2010-09-22 22:18:07 +00:00
2348393cf5
Teach memdep about TBAA tags.
...
llvm-svn: 114588
2010-09-22 21:41:02 +00:00
ba28207dcd
Constify.
...
llvm-svn: 114574
2010-09-22 20:11:43 +00:00
a9e57e0eff
Rework passing parent pointers into complexpatterns, I forgot
...
that complex patterns are matched after the entire pattern has
a structural match, therefore the NodeStack isn't in a useful
state when the actual call to the matcher happens.
llvm-svn: 114489
2010-09-21 22:00:25 +00:00
b0c0eaf579
Add some utility routines.
...
llvm-svn: 114483
2010-09-21 21:20:13 +00:00
c3e4b141a3
Add insertAfter. This should have accompanied previous check-in.
...
llvm-svn: 114481
2010-09-21 21:10:42 +00:00
dd83548fea
just like they can opt into getting the root of the pattern being
...
matched, allow ComplexPatterns to opt into getting the parent node
of the operand being matched.
llvm-svn: 114472
2010-09-21 20:37:12 +00:00
0e023ea02a
fix a long standing wart: all the ComplexPattern's were being
...
passed the root of the match, even though only a few patterns
actually needed this (one in X86, several in ARM [which should
be refactored anyway], and some in CellSPU that I don't feel
like detangling). Instead of requiring all ComplexPatterns to
take the dead root, have targets opt into getting the root by
putting SDNPWantRoot on the ComplexPattern.
llvm-svn: 114471
2010-09-21 20:31:19 +00:00
a4f199720d
finish pushing MachinePointerInfo through selectiondags. At this point,
...
I think I've audited all uses, so it should be dependable for address spaces,
and the pointer+offset info should also be accurate when there.
llvm-svn: 114464
2010-09-21 18:58:22 +00:00
886250c8f0
convert a couple more places to use the new getStore()
...
llvm-svn: 114463
2010-09-21 18:51:21 +00:00
6963c1f789
eliminate an old SelectionDAG::getTruncStore method, propagating
...
MachinePointerInfo around more.
llvm-svn: 114452
2010-09-21 17:42:31 +00:00
5e39ffd02f
eliminate last SelectionDAG::getLoad old entrypoint, on to stores.
...
llvm-svn: 114450
2010-09-21 17:28:52 +00:00
ea952f05a5
fix the code that infers SV info to be correct when dealing
...
with an indexed load/store that has an offset in the index.
llvm-svn: 114449
2010-09-21 17:24:05 +00:00
1ccded77c0
Add LiveInterval::find and use it for most LiveRange searching operations
...
instead of calling lower_bound or upper_bound directly.
This cleans up the search logic a bit because {lower,upper}_bound compare
LR->start by default, and it is usually simpler to search LR->end.
Funnelling all searches through one function also makes it possible to replace
the search algorithm with something faster than binary search.
llvm-svn: 114448
2010-09-21 17:12:18 +00:00
04610c63cb
Remove dead method.
...
llvm-svn: 114447
2010-09-21 17:12:15 +00:00
3d178ed4d4
propagate MachinePointerInfo through various uses of the old
...
SelectionDAG::getExtLoad overload, and eliminate it.
llvm-svn: 114446
2010-09-21 17:04:51 +00:00
1ffcf527c7
continue MachinePointerInfo'izing, eliminating use of one of the old
...
getLoad overloads.
llvm-svn: 114443
2010-09-21 16:36:31 +00:00
4021d906f1
Make CreateComplexVariable independent of SmallVector.
...
llvm-svn: 114439
2010-09-21 16:00:03 +00:00
ed79d5f24d
llvmc: Allow multiple output languages.
...
llvm-svn: 114433
2010-09-21 14:59:42 +00:00
a4862ae43f
Trailing whitespace.
...
llvm-svn: 114432
2010-09-21 14:59:34 +00:00
0937fc4b7f
Added an additional PBQP problem builder which adds coalescing costs (both between pairs of virtuals, and between virtuals and physicals).
...
llvm-svn: 114429
2010-09-21 13:19:36 +00:00
adbbb93d3d
Move the search for the appropriate AND instruction
...
into OptimizeCompareInstr.
This necessitates the passing of CmpValue around,
so widen the virtual functions to accomodate.
No functionality changes.
llvm-svn: 114428
2010-09-21 12:01:15 +00:00
50287ea65a
add some accessors
...
llvm-svn: 114409
2010-09-21 06:43:24 +00:00
82fd06d3ce
it's more elegant to put the "getConstantPool" and
...
"getFixedStack" on the MachinePointerInfo class. While
this isn't the problem I'm setting out to solve, it is the
right way to eliminate PseudoSourceValue, so lets go with it.
llvm-svn: 114406
2010-09-21 06:22:23 +00:00
8363087ba8
ugh, missed a file.
...
llvm-svn: 114405
2010-09-21 06:16:40 +00:00
2510de2bea
reimplement memcpy/memmove/memset lowering to use MachinePointerInfo
...
instead of srcvalue/offset pairs. This corrects SV info for mem
operations whose size is > 32-bits.
llvm-svn: 114401
2010-09-21 05:40:29 +00:00
de93bb065d
add some helpful accessors.
...
llvm-svn: 114400
2010-09-21 05:39:30 +00:00
bc419ba98f
add overloads for SelectionDAG::getLoad, getStore, getTruncStore that take a
...
MachinePointerInfo. Among other virtues, this doesn't silently truncate the
svoffset to 32-bits.
llvm-svn: 114399
2010-09-21 05:10:45 +00:00
d2d58ada70
simplify interface to SelectionDAG::getMemIntrinsicNode, making it take a MachinePointerInfo
...
llvm-svn: 114397
2010-09-21 04:57:15 +00:00
15d84c460a
chagne interface to SelectionDAG::getAtomic to take a MachinePointerInfo,
...
eliminating some weird "infer a frame address" logic which was dead.
llvm-svn: 114396
2010-09-21 04:53:42 +00:00
b5f4920979
force clients of MachineFunction::getMachineMemOperand to provide a
...
MachinePointerInfo, propagating the type out a level of API. Remove
the old MachineFunction::getMachineMemOperand impl.
llvm-svn: 114393
2010-09-21 04:46:39 +00:00
00ca0b8e98
start pushing MachinePointerInfo out through the MachineMemOperand interface
...
to the MachineFunction construction methods.
llvm-svn: 114390
2010-09-21 04:32:08 +00:00
187f653418
refactor the Value*/offset pair from MachineMemOperand out to a new
...
MachinePointerInfo struct, no functionality change.
This also adds an assert to MachineMemOperand::MachineMemOperand
that verifies that the Value* is either null or is an IR pointer type.
llvm-svn: 114389
2010-09-21 04:23:39 +00:00
497b3cea1d
Relax this check to silently swallow FE_INEXACT, following directions
...
from rdar://8452472. This unbreaks gcc.dg/builtins-17.c.
llvm-svn: 114368
2010-09-20 22:32:25 +00:00
a005f9400d
fix a bug I introduced back in the hayday of version #2 .
...
llvm-svn: 114319
2010-09-20 03:58:32 +00:00
aaa6badd05
restrict dyn_cast_or_null to pointer types, just like cast_or_null; re-commit of r114279, backed out in r114280
...
llvm-svn: 114282
2010-09-18 13:03:32 +00:00
239a07c297
back out r114279 as some darwin buildbots get errors compiling clang:
...
svn merge -c -114279 llvm/include/llvm/Support/Casting.h
llvm-svn: 114280
2010-09-18 12:56:47 +00:00
d1d99c8483
restrict dyn_cast_or_null to pointer types, just like cast_or_null
...
llvm-svn: 114279
2010-09-18 12:30:15 +00:00
c44201c91e
remove CallSite::get; it is still present (as protected) in the baseclass, use one of the constructors intead
...
llvm-svn: 114275
2010-09-18 11:48:36 +00:00
cb1e1017dd
Added a separate class (PBQPBuilder) for PBQP Problem construction. This class can be extended to support custom constraints.
...
For now the allocator still uses the old (internal) construction mechanism by default. This will be phased out soon assuming
no issues with the builder system come up.
To invoke the new construction mechanism just pass '-regalloc=pbqp -pbqp-builder' to llc. To provide custom constraints a
Target just needs to extend PBQPBuilder and pass an instance of their derived builder to the RegAllocPBQP constructor.
llvm-svn: 114272
2010-09-18 09:07:10 +00:00
757732c25e
trailing whitespace
...
llvm-svn: 114211
2010-09-17 21:23:56 +00:00
Jan Wen Voung