Commit Graph

7761 Commits

Author SHA1 Message Date
Daniel Dunbar 86c065dd68 Revert 78892 and 78895, these break generating working executables on
x86_64-apple-darwin10.

--- Reverse-merging r78895 into '.':
U    test/CodeGen/PowerPC/2008-12-12-EH.ll
U    lib/Target/DarwinTargetAsmInfo.cpp
--- Reverse-merging r78892 into '.':
U    include/llvm/Target/DarwinTargetAsmInfo.h
U    lib/Target/X86/X86TargetAsmInfo.cpp
U    lib/Target/X86/X86TargetAsmInfo.h
U    lib/Target/ARM/ARMTargetAsmInfo.h
U    lib/Target/ARM/ARMTargetMachine.cpp
U    lib/Target/ARM/ARMTargetAsmInfo.cpp
U    lib/Target/PowerPC/PPCTargetAsmInfo.cpp
U    lib/Target/PowerPC/PPCTargetAsmInfo.h
U    lib/Target/PowerPC/PPCTargetMachine.cpp
G    lib/Target/DarwinTargetAsmInfo.cpp

llvm-svn: 78919
2009-08-13 17:03:38 +00:00
Chris Lattner 68535f7603 reintroduce support for Mips "small" section handling. This is
implemented somewhat differently than before, but it should have
the same functionality and the previous testcase passes again.

llvm-svn: 78900
2009-08-13 06:28:06 +00:00
Evan Cheng f59e9f4288 tPOP_RET now has predicate operands.
llvm-svn: 78898
2009-08-13 06:05:07 +00:00
Chris Lattner 00a8de054c fix typo, add 10.6 version of test for my previous patch.
llvm-svn: 78895
2009-08-13 05:43:33 +00:00
Evan Cheng e5801bd220 It's ok to spill a tGPR register as long as it's still allocated a low register.
llvm-svn: 78893
2009-08-13 05:40:51 +00:00
Mon P Wang a95379d165 When InstCombine simplifies a load -> extract element to gep -> load, place
the new load by the old load instead of by the extract element because
a store could have occurred between the load and extract element.

llvm-svn: 78891
2009-08-13 05:12:13 +00:00
Bruno Cardoso Lopes 607cd3b63a Change MCSectionELF to represent a section semantically instead of
syntactically as a string, very similiar to what Chris did with MachO.
The parsing support and validation is not introduced yet.

llvm-svn: 78890
2009-08-13 05:07:35 +00:00
Dan Gohman ef3d457126 Various AsmWriter output cleanups. Use WriteAsOperand instead of
PrintUnmangledNameSafely.

llvm-svn: 78878
2009-08-13 01:36:44 +00:00
Dan Gohman 69273e6474 Now that numbered types have their number printed, it's no longer
interesting to print the number in a comment. Numbered instructions
don't need their number in a comment either.

Also, tidy up newline printing.

llvm-svn: 78865
2009-08-12 23:54:22 +00:00
Dan Gohman 466876b0a6 Extend the AsmWriter to print unnamed numbered types as "%0 = type ..."
and unnamed numbered global variables as "@0 = global ...". Extend the
AsmParser to recognize these forms.

llvm-svn: 78859
2009-08-12 23:32:33 +00:00
Dan Gohman 1c0e13fe66 Use WriteAsOperand to print BasicBlock names.
llvm-svn: 78838
2009-08-12 20:56:56 +00:00
Bob Wilson 4b35448360 Generate Neon VTBL and VTBX instructions from the corresponding intrinsics.
llvm-svn: 78835
2009-08-12 20:51:55 +00:00
Dale Johannesen 58043874ce Test for 78821, sort of. While that bug is nondeterministic,
this test failed consistently on a Darwin build.

llvm-svn: 78822
2009-08-12 17:43:47 +00:00
Dan Gohman de51b4ebf4 Fix whitespace expectations to match the new AsmWriter behavior.
llvm-svn: 78819
2009-08-12 17:39:04 +00:00
Dan Gohman e274526d78 Make LLVM Assembly dramatically easier to read by aligning the comments,
using formatted_raw_ostream's PadToColumn.

Before:

bb1:            ; preds = %bb
  %2 = sext i32 %i.01 to i64            ; <i64> [#uses=1]
  %3 = getelementptr double* %p, i64 %2         ; <double*> [#uses=1]
  %4 = load double* %3, align 8         ; <double> [#uses=1]
  %5 = fmul double %4, 1.100000e+00             ; <double> [#uses=1]
  %6 = sext i32 %i.01 to i64            ; <i64> [#uses=1]
  %7 = getelementptr double* %p, i64 %6         ; <double*> [#uses=1]

After:

bb1:                                        ; preds = %bb
  %2 = sext i32 %i.01 to i64                ; <i64> [#uses=1]
  %3 = getelementptr double* %p, i64 %2     ; <double*> [#uses=1]
  %4 = load double* %3, align 8             ; <double> [#uses=1]
  %5 = fmul double %4, 1.100000e+00         ; <double> [#uses=1]
  %6 = sext i32 %i.01 to i64                ; <i64> [#uses=1]
  %7 = getelementptr double* %p, i64 %6     ; <double*> [#uses=1]

Several tests required whitespace adjustments.

llvm-svn: 78816
2009-08-12 17:23:50 +00:00
Dan Gohman 4ac2f639cd Transform -X/C to X/-C, implementing a README.txt entry.
llvm-svn: 78812
2009-08-12 16:37:02 +00:00
Dan Gohman 908da3d97e Optimize (x/C)*C to x if the division is exact.
llvm-svn: 78811
2009-08-12 16:33:09 +00:00
Chris Lattner 1235f20744 one last (?) bad x86 triple test.
llvm-svn: 78801
2009-08-12 06:49:44 +00:00
Chris Lattner 6628e17344 fix some pastos in triple lines.
llvm-svn: 78800
2009-08-12 06:49:12 +00:00
Chris Lattner 0ea6e4cc7f another bogus triple
llvm-svn: 78798
2009-08-12 06:36:52 +00:00
Chris Lattner 8c2d846a42 fix another broken target triple.
llvm-svn: 78796
2009-08-12 06:29:18 +00:00
Chris Lattner d4a70aedb0 fix an incorrect target triple.
llvm-svn: 78795
2009-08-12 06:28:51 +00:00
Chris Lattner 774a88aa77 add nounwind
llvm-svn: 78791
2009-08-12 05:44:03 +00:00
Evan Cheng 608d92c943 Remove an Darwin assembler workaround.
llvm-svn: 78777
2009-08-12 01:56:42 +00:00
Evan Cheng 1e6c2a1c17 Shrink ADDS, ADC, RSB, and SUBS.
llvm-svn: 78776
2009-08-12 01:49:45 +00:00
Bill Wendling c1aa4edbde Testcase for r78770.
llvm-svn: 78774
2009-08-12 01:47:51 +00:00
Dale Johannesen 419a5b98ec Test for llvm-gcc patch 78762.
llvm-svn: 78763
2009-08-12 01:14:30 +00:00
Dan Gohman a3be3d7a89 llvm-gcc now emits inbounds for this getelementptr.
llvm-svn: 78758
2009-08-12 00:35:55 +00:00
Daniel Dunbar 3fb754af06 llvm-mc/AsmParser: Match hard coded registers (e.g. 'shldl %cl, %eax, %eax')
We now match all of 403.gcc (as emitted by clang). :)

llvm-svn: 78750
2009-08-11 23:23:44 +00:00
Evan Cheng f6a9d06241 Shrinkify Thumb2 r = add sp, imm.
llvm-svn: 78745
2009-08-11 23:00:31 +00:00
Evan Cheng cc9ca3500d Shrinkify Thumb2 load / store multiple instructions.
llvm-svn: 78717
2009-08-11 21:11:32 +00:00
Dan Gohman dbae4db67a Optimize exact sdiv by a constant power of 2 to ashr.
llvm-svn: 78714
2009-08-11 20:47:47 +00:00
Dan Gohman f9d12f23cb Use wchar.h to get wchar_t, not ctype.h.
llvm-svn: 78703
2009-08-11 19:59:21 +00:00
Devang Patel 5c310be0ff Link NamedMDNodes.
llvm-svn: 78696
2009-08-11 18:01:24 +00:00
Daniel Dunbar 7f8a9ebc8d llvm-mc: Fix a crash on invalid due to a typo in relocatable expression
evaluation.

llvm-svn: 78692
2009-08-11 17:47:52 +00:00
Chris Lattner 0c533d909a now that these are in file-check format, we can merge them together
into one bigger test (which runs faster)

llvm-svn: 78672
2009-08-11 15:54:17 +00:00
Dan Gohman a72f856aa4 Don't assume that external global variables are aligned at their preferred
alignment. Only the minimum alignment guaranteed by the ABI may be assumed.

llvm-svn: 78668
2009-08-11 15:50:03 +00:00
Evan Cheng 806845daec Fix the previous accidental commit. Now shrinking common Thumb2 load / store instructions.
llvm-svn: 78659
2009-08-11 09:37:40 +00:00
Devang Patel 4d4a2e09a9 Link metadata.
llvm-svn: 78652
2009-08-11 06:46:31 +00:00
Jakob Stoklund Olesen b39a5aa794 Rebuild RegScavenger::DistanceMap each time it is needed.
The register scavenger maintains a DistanceMap that maps MI pointers to their
distance from the top of the current MBB. The DistanceMap is built
incrementally in forward() and in bulk in findFirstUse(). It is used by
scavengeRegister() to determine which candidate register has the longest
unused interval.

Unfortunately the DistanceMap contents can become outdated. The first time
scavengeRegister() is called, the DistanceMap is filled to cover the MBB. If
then instructions are inserted in the MBB (as they always are following
scavengeRegister()), the recorded distances are too short. This causes bad
behaviour in the included test case where a register use /after/ the current
position is ignored because findFirstUse() thinks is is /before/ the current
position. A "using an undefined register" assertion follows promptly.

The fix is to build a fresh DistanceMap at the top of scavengeRegister(), and
discard it after use. This means that DistanceMap is no longer needed as a
RegScavenger member variable, and forward() doesn't need to update it.

The fix then discloses issue number two in the same test case: The candidate
search in scavengeRegister() finds a CSR that has been saved in the prologue,
but is currently unused. It would be both inefficient and wrong to spill such
a register in the emergency spill slot. In the present case, the emergency
slot restore is placed immediately before the normal epilogue restore, leading
to a "Redefining a live register" assertion.

Fix number two: When scavengerRegister() stumbles upon an unused register that
is overwritten later in the MBB, return that register early. It is important
to verify that the register is defined later in the MBB, otherwise it might be
an unspilled CSR.

llvm-svn: 78650
2009-08-11 06:25:12 +00:00
Bob Wilson 8f5c447bfa Convert more Neon tests to use FileCheck.
llvm-svn: 78648
2009-08-11 05:51:19 +00:00
Bob Wilson 12842f9865 Use vAny type to get rid of Neon intrinsics that differed only in whether
the overloaded vector types allowed floating-point or integer vector elements.
Most of these operations actually depend on the element type, so bitcasting
was not an option.

If you include the vpadd intrinsics that I updated earlier, this gets rid
of 20 intrinsics.

llvm-svn: 78646
2009-08-11 05:39:44 +00:00
Daniel Dunbar 71527c1493 llvm-mc/X86: Parse '*' correctly (in the way the matcher expects).
llvm-svn: 78642
2009-08-11 05:00:25 +00:00
Daniel Dunbar ba95a7cd8d llvm-mc: Accept .word as a synonym for .short
llvm-svn: 78641
2009-08-11 04:44:00 +00:00
Daniel Dunbar a4b069ce00 llvm-mc: Sketch parsing for .file, .line, and .loc. No streamer hooks for these
yet (I'm not even sure what they do).

llvm-svn: 78639
2009-08-11 04:24:50 +00:00
Daniel Dunbar 9af747b421 llvm-mc: Fix darwin .section parsing. It was skipping the section name and a ','
(and outputting a diagnostic pointing at the wrong place), all of which lead to
much confusion.

llvm-svn: 78637
2009-08-11 03:42:33 +00:00
Daniel Dunbar bb98db2d21 llvm-mc/AsmParser: Implement automatic classification of RegisterClass operands.
- This drops us to 123 ambiguous instructions (previously ~500) on X86.

llvm-svn: 78636
2009-08-11 02:59:53 +00:00
Bob Wilson 741a9c7bf6 Use new EVT::vAny type to combine Neon intrinsics for VPADD.
llvm-svn: 78632
2009-08-11 01:15:26 +00:00
Evan Cheng 475f8a4fa2 Enable Thumb2 instruction shrinking (32-bit to 16-bit) pass. Convert a bunch of thumb2 tests to FileCheck.
llvm-svn: 78622
2009-08-10 23:56:04 +00:00
Dan Gohman 9d26c85bdc Fix a bug in the DAGCombiner's handling of multiple linked
MERGE_VALUES nodes. Replacing the result values with the
operands in one MERGE_VALUES node may cause another
MERGE_VALUES node be CSE'd with the first one, and bring
its uses along, so that the first one isn't dead, as this
code expects. Fix this by iterating until the node is
really dead. This fixes PR4699.

llvm-svn: 78619
2009-08-10 23:43:19 +00:00
David Goodwin d9aedcae23 Use FileCheck.
llvm-svn: 78614
2009-08-10 23:14:14 +00:00
David Goodwin bdf1a1d1a2 Use FileCheck... its good for you...
llvm-svn: 78613
2009-08-10 23:06:57 +00:00
David Goodwin 9e7c7e748f Fix test.
llvm-svn: 78611
2009-08-10 22:58:08 +00:00
David Goodwin 108b522912 Fix test.
llvm-svn: 78606
2009-08-10 22:31:04 +00:00
David Goodwin 85b5b027f7 Use NEON for single-precision int<->FP conversions.
llvm-svn: 78604
2009-08-10 22:17:39 +00:00
Daniel Dunbar e4f79d140d llvm-mc/AsmParser: Disambiguate i64i8imm.
llvm-svn: 78598
2009-08-10 21:06:41 +00:00
Evan Cheng f72c13bdf5 Handle the constantfp created during post-legalization dag combiner phase.
llvm-svn: 78594
2009-08-10 20:25:59 +00:00
Erick Tryzelaar a1025faf84 Fix ocaml "make check" tests, that wasn't finding the proper c++ compiler.
llvm-svn: 78592
2009-08-10 19:45:05 +00:00
Dan Gohman 676d115ce5 Add nounwind keywords.
llvm-svn: 78568
2009-08-10 16:48:40 +00:00
Daniel Dunbar 2587b61533 llvm-mc/AsmParser: Check for matches with super classes when matching
instruction operands.

llvm-svn: 78565
2009-08-10 16:05:47 +00:00
Chris Lattner cb307a27bf Make the big switch: Change MCSectionMachO to represent a section *semantically*
instead of syntactically as a string.  This means that it keeps track of the 
segment, section, flags, etc directly and asmprints them in the right format.
This also includes parsing and validation support for llvm-mc and 
"attribute(section)", so we should now start getting errors about invalid 
section attributes from the compiler instead of the assembler on darwin.

Still todo: 
1) Uniquing of darwin mcsections
2) Move all the Darwin stuff out to MCSectionMachO.[cpp|h]
3) there are a few FIXMEs, for example what is the syntax to get the
   S_GB_ZEROFILL segment type?

llvm-svn: 78547
2009-08-10 01:39:42 +00:00
Daniel Dunbar 5877782962 Disable this test for now, we don't check for super classes when matching yet.
llvm-svn: 78531
2009-08-09 07:35:56 +00:00
Daniel Dunbar 8e33cb2de1 llvm-mc/AsmParser: Implement user defined super classes.
- We can now discriminate SUB32ri8 from SUB32ri, for example.

llvm-svn: 78530
2009-08-09 07:20:21 +00:00
Bob Wilson 7fc63417d7 Add tests for Neon VZIP and VUZP instructions.
llvm-svn: 78529
2009-08-09 06:48:29 +00:00
Bob Wilson f60c8807e8 Add a test for Neon VTRN instructions.
llvm-svn: 78528
2009-08-09 06:30:46 +00:00
Chris Lattner 591105c540 sink the 'name' and 'isdirective' state out of MCSection into its derived classes.
This totally optimizes PIC16 sections by not having an 'isdirective' bit anymore!! ;-)

llvm-svn: 78517
2009-08-08 23:39:42 +00:00
Chris Lattner c51023b0bc remove nondeterminstic test.
llvm-svn: 78512
2009-08-08 22:56:37 +00:00
Eric Christopher 7dfa9f2e56 Add crc32 instruction and intrinsics. Add a new class of prefix
bytes for F2 0F 38 and propagate. Add a FIXME for a set
of possibilities which correspond to intrinsics already used.

New test.

llvm-svn: 78508
2009-08-08 21:55:08 +00:00
Jakob Stoklund Olesen e2dc8a46e9 Add support for READCYCLECOUNTER in Blackfin back-end.
llvm-svn: 78506
2009-08-08 21:42:22 +00:00
Jakob Stoklund Olesen dc6bccbaa6 Don't build illegal ops in DAGCombiner::SimplifyBinOpWithSameOpcodeHands().
Blackfin supports and/or/xor on i32 but not on i16. Teach
DAGCombiner::SimplifyBinOpWithSameOpcodeHands to not produce illegal nodes
after legalize ops.

llvm-svn: 78497
2009-08-08 20:42:17 +00:00
Jakob Stoklund Olesen ac51533b8a Simplify RegScavenger::forward a bit more.
Verify that early clobber registers and their aliases are not used.

All changes to RegsAvailable are now done as a transaction so the order of
operands makes no difference.

The included test case is from PR4686. It has behaviour that was dependent on the order of operands.

llvm-svn: 78465
2009-08-08 13:18:47 +00:00
Daniel Dunbar 541efcc5c4 llvm-mc/AsmMatcher: Improve match code.
- This doesn't actually improve the algorithm (its still linear), but the
   generated (match) code is now fairly compact and table driven. Still need a
   generic string matcher.

 - The table still needs to be compressed, this is quite simple to do and should
   shrink it to under 16k.

 - This also simplifies and restructures the code to make the match classes more
   explicit, in anticipation of resolving ambiguities.

llvm-svn: 78461
2009-08-08 07:50:56 +00:00
Andreas Bolka 3becda83ef Add another Strong-SIV testcase.
llvm-svn: 78446
2009-08-08 00:21:49 +00:00
Anton Korobeynikov 674ffc1e59 Do not generate 32-bit call on win64 when imm does not fit
llvm-svn: 78443
2009-08-07 23:59:21 +00:00
Chris Lattner 6eceb7c85d rename test
llvm-svn: 78441
2009-08-07 23:57:30 +00:00
Chris Lattner ff8d04e815 merge a bunch of tests together into one, convert to filecheck which
is more tolerant of whitespace differences.

llvm-svn: 78439
2009-08-07 23:56:42 +00:00
Bob Wilson 97262e01d5 Convert more Neon tests to use FileCheck.
llvm-svn: 78433
2009-08-07 23:45:02 +00:00
David Goodwin 742db6a6d4 Make NEON single-precision FP support the default for cortex-a8 (again).
llvm-svn: 78430
2009-08-07 23:32:33 +00:00
Kevin Enderby d9f952948e Added Mac OS X assembler style conditional assembly. I may come back and see if
I can clean this up a bit more and do way with the TheCondState and just use
the top element on the TheCondStack if not empty.  Also may tweak the code
around ParseConditionalAssemblyDirectives() to simplify the AsmParser code.

llvm-svn: 78423
2009-08-07 22:46:00 +00:00
Anton Korobeynikov 23b28cb824 2 more vdup.32 cases
llvm-svn: 78419
2009-08-07 22:36:50 +00:00
Evan Cheng 6e130db3b7 Thumb2 32-bit ldm / stm needs .w suffix if submode is ia.
llvm-svn: 78410
2009-08-07 21:19:10 +00:00
Daniel Dunbar bfdd58e103 XFAIL 2006-11-06-StackTrace.cpp on powerpc-apple-darwin9 until someone feels
motivated to fix it.

llvm-svn: 78406
2009-08-07 20:50:58 +00:00
Evan Cheng 4c3b1ca5a0 Fix support to use NEON for single precision fp math.
llvm-svn: 78397
2009-08-07 19:30:41 +00:00
Andreas Bolka 787591a594 Fix Strong-SIV testcase.
llvm-svn: 78384
2009-08-07 15:42:32 +00:00
Daniel Dunbar 32d0bb4c4e Disable X86 AsmMatcher for now, it is causing gcc-4.0 to run out of memory on
i386-apple-darwin9. This presumably will get fixed once the generated code
improves.

llvm-svn: 78379
2009-08-07 08:45:03 +00:00
Daniel Dunbar e10787e710 llvm-mc/AsmMatcher: Move to a slightly more sane matching design.
- Still not very sane, but a least its not 60k lines on X86. :)

 - In terms of correctness, currently some things are hard wired for X86, and we
   still don't properly resolve ambiguities (this is ignoring the instructions
   we don't even match due to funny .td stuff or other corner cases).

The high level changes:
 1. Represent tokens which are significant for matching explicitly as separate
    operands. This uniformly handles not only the instruction mnemonic, but
    also 'signficiant' syntax like the '*' in "call * ...".

 2. Separate the matching of operands to an instruction from the construction of
    the MCInst. In theory this can be done during matching, but since the number
    of variations is small I think it makes sense to decompose the problems.

 3. Improved a few of the mechanisms to at least successfully flatten / tokenize
    the assembly strings for PowerPC and ARM.

 4. The comment at the top of AsmMatcherEmitter.cpp explains the approach I'm
    moving towards for handling ambiguous instructions. The high-bit is to infer
    a partial ordering of the operand classes (and force the user to specify one
    if we can't) and use that to resolve ambiguities.

llvm-svn: 78378
2009-08-07 08:26:05 +00:00
Evan Cheng b1aeeed03e Another coalescer bug. When a dead copy is eliminated, transfer the kill to a def of the exact register rather than a super-register.
llvm-svn: 78376
2009-08-07 07:14:14 +00:00
Evan Cheng b972e5633f It turns out most of the thumb2 instructions are not allowed to touch SP. The semantics of such instructions are unpredictable. We have just been lucky that tests have been passing.
This patch takes pain to ensure all the PEI lowering code does the right thing when lowering frame indices, insert code to manipulate stack pointers, etc. It's also custom lowering dynamic stack alloc into pseudo instructions so we can insert the right instructions at scheduling time.

This fixes PR4659 and PR4682.

llvm-svn: 78361
2009-08-07 00:34:42 +00:00
Dale Johannesen 3a127ce1d8 Add the testcase from PR 4668. This works at the
moment, but it's a fragile area.

llvm-svn: 78358
2009-08-07 00:04:42 +00:00
Dale Johannesen 15a5fad94b Fix PR 4626, a crash in branch folding after OptimizeBlock
produced a CFG it wasn't prepared for.

llvm-svn: 78351
2009-08-06 22:56:40 +00:00
Bob Wilson 0127031c20 Implement Neon VST[234] operations.
llvm-svn: 78330
2009-08-06 18:47:44 +00:00
Bob Wilson e3ec5b6d76 Fix incorrect intrinsic declarations.
llvm-svn: 78329
2009-08-06 18:46:26 +00:00
Dale Johannesen 8d7f6f83fd Testcase for llvm-gcc 78324 (although in normal
testing mode it won't test much).

llvm-svn: 78325
2009-08-06 17:28:23 +00:00
Dan Gohman b4764e5b7f Tidy up this testcase.
llvm-svn: 78322
2009-08-06 17:11:55 +00:00
Chris Lattner a7e2662770 reduce testcase.
llvm-svn: 78315
2009-08-06 16:14:33 +00:00
Dan Gohman ee902509a8 Remove an over-aggressive assert. Functions with empty struct return
types don't have any return values, from CodeGen's perspective.
This fixes PR4688.

llvm-svn: 78311
2009-08-06 15:07:58 +00:00
Anton Korobeynikov 644caa0cdb Add tests for X86-64 code model handling. Small and kernel for now.
llvm-svn: 78300
2009-08-06 12:25:20 +00:00
Dan Gohman 130e2c7aed Fix a bug in x86's PreprocessForRMW logic that was exposed
by aggressive chain operand optimization. UpdateNodeOperands
does not modify the node in place if it would result in
a node identical to an existing node.

llvm-svn: 78297
2009-08-06 09:22:57 +00:00
Dan Gohman 5758e1e92a Fix a few places in DAGCombiner that were creating all-ones-bits
and high-bits values in ways that weren't correct for integer
types wider than 64 bits. This fixes a miscompile in
PPMacroExpansion.cpp in clang on x86-64.

llvm-svn: 78295
2009-08-06 09:18:59 +00:00
Chris Lattner 5ccb3af0aa add a testcase for my llvm-gcc hack.
llvm-svn: 78289
2009-08-06 06:18:52 +00:00
Chris Lattner 87754848ab this passes.
llvm-svn: 78281
2009-08-06 03:55:49 +00:00
Sanjiv Gupta aee88e46c1 XFAIL it while it is being worked on.
llvm-svn: 78275
2009-08-06 02:19:20 +00:00
Bob Wilson 3389c2f7d0 Add tests for new NEON vld instructions.
llvm-svn: 78264
2009-08-06 00:38:31 +00:00
Bob Wilson dd611f44cb Convert more Neon tests to FileCheck.
llvm-svn: 78261
2009-08-05 23:51:20 +00:00
Dale Johannesen dc3416b7b3 Adjust test for llvm-gcc checkin 78249.
llvm-svn: 78251
2009-08-05 22:18:47 +00:00
Dale Johannesen 1de1897bf0 Test for llvm-gcc checkin 78223.
llvm-svn: 78229
2009-08-05 19:10:41 +00:00
Anton Korobeynikov 22ef75155e Missed pieces for ARM HardFP ABI.
Patch by Sandeep Patel!

llvm-svn: 78225
2009-08-05 19:04:42 +00:00
Dan Gohman df7ea32af7 Enable the new no-SP register classes by default. This is to address
PR4572. A few tests have some minor code regressions due to different
coalescing.

llvm-svn: 78217
2009-08-05 17:40:24 +00:00
Anton Korobeynikov f6e25b3039 Add testcases for reg-mem arithemtics added recently
llvm-svn: 78214
2009-08-05 17:04:32 +00:00
Anton Korobeynikov be47ccffef Convert bswap test to filecheck, add more test entries & convert stuff to filecheck
llvm-svn: 78212
2009-08-05 16:50:53 +00:00
Dan Gohman 477fd55c9a Fix a bug in the PIC16 backend.
llvm-svn: 78211
2009-08-05 16:46:43 +00:00
Dan Gohman 2bebfc38af Change these tests to use function attributes rather than special llc
command-line options.

llvm-svn: 78204
2009-08-05 16:37:27 +00:00
Chris Lattner abde7f9d27 checking in broken testcases is not such a good idea.
llvm-svn: 78201
2009-08-05 16:04:18 +00:00
Sanjiv Gupta 63c5ede173 Quite a few tests crashed in llc after 78142. This is just one of them. I hope to add a few more.
llvm-svn: 78198
2009-08-05 15:52:14 +00:00
Evan Cheng ea2b82b8fc Disable stack coloring with register for now. It's not able to set kill markers.
llvm-svn: 78179
2009-08-05 07:26:17 +00:00
Evan Cheng a2ce665f60 Another nasty coalescer bug (is there another kind):
After coalescing reg1027's def and kill are both at the same point:
 %reg1027,0.000000e+00 = [56,814:0)  0@70-(814)

bb5:
60   %reg1027<def> = t2MOVr %reg1027, 14, %reg0, %reg0
68   %reg1027<def> = t2LDRi12 %reg1027<kill>, 8, 14, %reg0
76   t2CMPzri %reg1038<kill,undef>, 0, 14, %reg0, %CPSR<imp-def>
84   %reg1027<def> = t2MOVr %reg1027, 14, %reg0, %reg0
96   t2Bcc mbb<bb5,0x2030910>, 1, %CPSR<kill>

Do not remove the kill marker on t2LDRi12.

llvm-svn: 78178
2009-08-05 07:05:41 +00:00
Dan Gohman e32c0fe584 Revert changes accidentally committed along with r78163.
llvm-svn: 78165
2009-08-05 05:38:13 +00:00
Dan Gohman 8c79569853 Teach X86FastISel how to handle CCValAssign::BCvt, which is used for
MMX arguments. This fixes PR4684.

llvm-svn: 78163
2009-08-05 05:33:42 +00:00
Andreas Bolka 13b860992a ZIV tester for LDA.
llvm-svn: 78157
2009-08-05 04:26:05 +00:00
Andreas Bolka 2979eb8f35 Fix LDA testcases.
llvm-svn: 78153
2009-08-05 04:03:29 +00:00
Evan Cheng 379429200e Turn some insert_subreg, extract_subreg, subreg_to_reg into implicit_defs.
llvm-svn: 78151
2009-08-05 03:53:14 +00:00
Evan Cheng 1f7b549c79 One more. Transfer kill of the larger register when lowering an EXTRACT_SUBREG.
llvm-svn: 78145
2009-08-05 02:25:11 +00:00
Evan Cheng 6376367356 One more place where subreg lowering forgot to transfer undefness.
llvm-svn: 78144
2009-08-05 01:57:22 +00:00
Dan Gohman f9bbcd1afd Major calling convention code refactoring.
Instead of awkwardly encoding calling-convention information with ISD::CALL,
ISD::FORMAL_ARGUMENTS, ISD::RET, and ISD::ARG_FLAGS nodes, TargetLowering
provides three virtual functions for targets to override:
LowerFormalArguments, LowerCall, and LowerRet, which replace the custom
lowering done on the special nodes. They provide the same information, but
in a more immediately usable format.

This also reworks much of the target-independent tail call logic. The
decision of whether or not to perform a tail call is now cleanly split
between target-independent portions, and the target dependent portion
in IsEligibleForTailCallOptimization.

This also synchronizes all in-tree targets, to help enable future
refactoring and feature work.

llvm-svn: 78142
2009-08-05 01:29:28 +00:00
Evan Cheng cdb125ce66 If the insert_subreg source is <undef>, insert an implicit_def instead of a copy.
llvm-svn: 78141
2009-08-05 01:29:24 +00:00
Evan Cheng 7cc6aca1e6 Fix part 1 of pr4682. PICADD is a 16-bit instruction even in thumb2 mode.
llvm-svn: 78126
2009-08-04 23:47:55 +00:00
Evan Cheng 28c2d9809d Fix test.
llvm-svn: 78113
2009-08-04 22:22:58 +00:00
Bob Wilson 6092c8d231 Convert more Neon tests to use FileCheck.
llvm-svn: 78111
2009-08-04 22:01:41 +00:00
Bob Wilson 2b60721464 Convert a few Neon tests to use FileCheck.
llvm-svn: 78108
2009-08-04 21:33:22 +00:00
Jakob Stoklund Olesen 0e0b5405f5 Clean up the handling of two-address operands in RegScavenger.
This fixes PR4528.

llvm-svn: 78107
2009-08-04 21:30:30 +00:00
Evan Cheng 783b65b546 Enable load / store multiple pass for Thumb2. It's not using ldrd / strd yet.
llvm-svn: 78104
2009-08-04 21:12:13 +00:00
David Goodwin 30bf625ac2 Add NEON single-precision FP support for fabs and fneg.
llvm-svn: 78101
2009-08-04 20:39:05 +00:00
Jakob Stoklund Olesen 6304369c4e LowerSubregsInstructionPass::LowerExtract should not extend the live range of registers.
When LowerExtract eliminates an EXTRACT_SUBREG with a kill flag, it moves the
kill flag to the place where the sub-register is killed. This can accidentally
overlap with the use of a sibling sub-register, and we have trouble.

In the test case we have this code:

Live Ins: %R0 %R1 %R2
	%R2L<def> = EXTRACT_SUBREG %R2<kill>, 1
	%R2H<def> = LOAD16fi <fi#-1>, 0, Mem:LD(2,4) [FixedStack-1 + 0]
	%R1L<def> = EXTRACT_SUBREG %R1<kill>, 1
	%R0L<def> = EXTRACT_SUBREG %R0<kill>, 1
	%R0H<def> = ADD16 %R2H<kill>, %R2L<kill>, %AZ<imp-def>, %AN<imp-def>, %AC0<imp-def>, %V<imp-def>, %VS<imp-def>

subreg: CONVERTING: %R2L<def> = EXTRACT_SUBREG %R2<kill>, 1
subreg: eliminated!
subreg: killed here: %R0H<def> = ADD16 %R2H, %R2L, %R2<imp-use,kill>, %AZ<imp-def>, %AN<imp-def>, %AC0<imp-def>, %V<imp-def>, %VS<imp-def>

The kill flag on %R2 is moved to the last instruction, and the live range overlaps with the definition of %R2H:

*** Bad machine code: Redefining a live physical register ***
- function:    f
- basic block:  0x18358c0 (#0)
- instruction: %R2H<def> = LOAD16fi <fi#-1>, 0, Mem:LD(2,4) [FixedStack-1 + 0]
Register R2H was defined but already live.

The fix is to replace EXTRACT_SUBREG with IMPLICIT_DEF instead of eliminating
it completely:

subreg: CONVERTING: %R2L<def> = EXTRACT_SUBREG %R2<kill>, 1
subreg: replace by: %R2L<def> = IMPLICIT_DEF %R2<kill>

Note that these IMPLICIT_DEF instructions survive to the asm output. It is
necessary to fix the stack-color-with-reg test case because of that.

llvm-svn: 78093
2009-08-04 20:01:11 +00:00
Evan Cheng a3abe2a7ce In thumb mode, r7 is used as frame register. This fixes pr4681.
llvm-svn: 78086
2009-08-04 18:46:17 +00:00
David Goodwin a3839bc6c0 Match common pattern for FNMAC. Add NEON SP support.
llvm-svn: 78085
2009-08-04 18:44:29 +00:00
David Goodwin a2824d5700 Improve tests.
llvm-svn: 78083
2009-08-04 18:11:59 +00:00
David Goodwin 3b9c52c5c1 Initial support for single-precision FP using NEON. Added "neonfp" attribute to enable. Added patterns for some binary FP operations.
llvm-svn: 78081
2009-08-04 17:53:06 +00:00
Evan Cheng 206ee96bd6 Fix PR4528. This scavenger assertion is too strict. The two-address value is
killed by another operand.

There is probably a better fix. Either 1) scavenger can look at other operands, or
2) livevariables can be smarter about kill markers. Patches welcome.

llvm-svn: 78072
2009-08-04 16:52:44 +00:00
Chris Lattner f222054df7 enhance codegen to put 16-bit character strings into the
__TEXT,__ustring section on darwin.

llvm-svn: 78068
2009-08-04 16:27:13 +00:00
Chris Lattner 81bbf443fe Add support emiting for 2/4 byte mergable strings to the ".rodata.str*"
section on ELF targets.

llvm-svn: 78066
2009-08-04 16:13:09 +00:00
Devang Patel 05eb617da5 Use separate ValueList for metadata.
This fixes PR4666.

llvm-svn: 78056
2009-08-04 06:00:18 +00:00
Devang Patel 8cca7b4abe Revert recent bitcode writer patches.
llvm-svn: 78053
2009-08-04 05:01:35 +00:00
Bill Wendling fd5d735e2b llvm-gcc checks the static asm variable is valid in ValidateRegisterVariable. Make this work for PPC.
llvm-svn: 78051
2009-08-04 04:48:31 +00:00
Bill Wendling 0549ae4181 Fixed now.
llvm-svn: 78042
2009-08-04 03:46:54 +00:00
Rafael Espindola dde8b5848a Add test for PR4678
llvm-svn: 78040
2009-08-04 03:44:37 +00:00
Bill Wendling 0c685e1055 XFAIL for the moment.
llvm-svn: 78038
2009-08-04 03:40:47 +00:00
Devang Patel 8abe6bc291 Constants and Metadata share ValueList. This means they must be emitted interleaved (using appropriate BLOCK_IDs) otherwise ValuePtrs index gets out of sync.
llvm-svn: 78033
2009-08-04 02:26:56 +00:00
Evan Cheng 03eb0e3c33 Emit sub r, #c instead of transforming it to add r, #-c if c fits in 8-bit. This is a bit of pre-mature optimization. 8-bit variant makes it likely it will be narrowed to a 16-bit instruction.
llvm-svn: 78030
2009-08-04 01:41:15 +00:00
Dan Gohman eaed26eb0e Add -disable-output. Thanks Bill!
llvm-svn: 78009
2009-08-03 22:24:22 +00:00
Dan Gohman f011f5a8a2 Add a new Constant::getIntegerValue helper function, and convert a
few places in InstCombine to use it, to fix problems handling pointer
types. This fixes the recent llvm-gcc bootstrap error.

llvm-svn: 78005
2009-08-03 22:07:33 +00:00