Commit Graph

7619 Commits

Author SHA1 Message Date
Chris Lattner
d3fe75e2bf now that these are in file-check format, we can merge them together
into one bigger test (which runs faster)


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78672 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-11 15:54:17 +00:00
Dan Gohman
004072508b Don't assume that external global variables are aligned at their preferred
alignment. Only the minimum alignment guaranteed by the ABI may be assumed.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78668 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-11 15:50:03 +00:00
Evan Cheng
195c71b472 Fix the previous accidental commit. Now shrinking common Thumb2 load / store instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78659 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-11 09:37:40 +00:00
Devang Patel
ae709eaeb1 Link metadata.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78652 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-11 06:46:31 +00:00
Jakob Stoklund Olesen
8c54a62061 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Unfortunately the DistanceMap contents can become outdated.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.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78650 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-11 06:25:12 +00:00
Bob Wilson
d1fa3f58a0 Convert more Neon tests to use FileCheck.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78648 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-11 05:51:19 +00:00
Bob Wilson
b0abb4dc42 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78646 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-11 05:39:44 +00:00
Daniel Dunbar
0db68f4e3a llvm-mc/X86: Parse '*' correctly (in the way the matcher expects).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78642 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-11 05:00:25 +00:00
Daniel Dunbar
1e840b2c61 llvm-mc: Accept .word as a synonym for .short
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78641 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-11 04:44:00 +00:00
Daniel Dunbar
d0c14d6913 llvm-mc: Sketch parsing for .file, .line, and .loc. No streamer hooks for these
yet (I'm not even sure what they do).


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78639 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-11 04:24:50 +00:00
Daniel Dunbar
ace63127bc 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78637 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-11 03:42:33 +00:00
Daniel Dunbar
ea6408f8cd llvm-mc/AsmParser: Implement automatic classification of RegisterClass operands.
- This drops us to 123 ambiguous instructions (previously ~500) on X86.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78636 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-11 02:59:53 +00:00
Bob Wilson
f24bd401eb Use new EVT::vAny type to combine Neon intrinsics for VPADD.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78632 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-11 01:15:26 +00:00
Evan Cheng
3a1f0f6785 Enable Thumb2 instruction shrinking (32-bit to 16-bit) pass. Convert a bunch of thumb2 tests to FileCheck.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78622 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-10 23:56:04 +00:00
Dan Gohman
00edf39b34 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78619 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-10 23:43:19 +00:00
David Goodwin
c5cfea3c02 Use FileCheck.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78614 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-10 23:14:14 +00:00
David Goodwin
4980187dd9 Use FileCheck... its good for you...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78613 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-10 23:06:57 +00:00
David Goodwin
af37f341df Fix test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78611 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-10 22:58:08 +00:00
David Goodwin
aeb66fe54e Fix test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78606 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-10 22:31:04 +00:00
David Goodwin
338268c67f Use NEON for single-precision int<->FP conversions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78604 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-10 22:17:39 +00:00
Daniel Dunbar
44f63f9335 llvm-mc/AsmParser: Disambiguate i64i8imm.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78598 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-10 21:06:41 +00:00
Evan Cheng
e2b861f7d9 Handle the constantfp created during post-legalization dag combiner phase.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78594 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-10 20:25:59 +00:00
Erick Tryzelaar
1afd3e67d1 Fix ocaml "make check" tests, that wasn't finding the proper c++ compiler.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78592 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-10 19:45:05 +00:00
Dan Gohman
5743a3f201 Add nounwind keywords.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78568 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-10 16:48:40 +00:00
Daniel Dunbar
fdb1f493ab llvm-mc/AsmParser: Check for matches with super classes when matching
instruction operands.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78565 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-10 16:05:47 +00:00
Chris Lattner
ff4bc460c5 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?



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78547 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-10 01:39:42 +00:00
Daniel Dunbar
a62b02a7ef Disable this test for now, we don't check for super classes when matching yet.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78531 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-09 07:35:56 +00:00
Daniel Dunbar
5fe6338ac8 llvm-mc/AsmParser: Implement user defined super classes.
- We can now discriminate SUB32ri8 from SUB32ri, for example.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78530 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-09 07:20:21 +00:00
Bob Wilson
22cd2547fe Add tests for Neon VZIP and VUZP instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78529 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-09 06:48:29 +00:00
Bob Wilson
4861fd5d0c Add a test for Neon VTRN instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78528 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-09 06:30:46 +00:00
Chris Lattner
93b6db3de9 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!! ;-)


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78517 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-08 23:39:42 +00:00
Chris Lattner
76eb6f5685 remove nondeterminstic test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78512 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-08 22:56:37 +00:00
Eric Christopher
b4dc13cab7 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78508 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-08 21:55:08 +00:00
Jakob Stoklund Olesen
46151de6a0 Add support for READCYCLECOUNTER in Blackfin back-end.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78506 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-08 21:42:22 +00:00
Jakob Stoklund Olesen
17421d81fd 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.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78497 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-08 20:42:17 +00:00
Jakob Stoklund Olesen
dffb051c21 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.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78465 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-08 13:18:47 +00:00
Daniel Dunbar
a3741fa28b 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78461 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-08 07:50:56 +00:00
Andreas Bolka
e2652f2cd3 Add another Strong-SIV testcase.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78446 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-08 00:21:49 +00:00
Anton Korobeynikov
941222eea0 Do not generate 32-bit call on win64 when imm does not fit
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78443 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-07 23:59:21 +00:00
Chris Lattner
ce0e9abef3 rename test
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78441 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-07 23:57:30 +00:00
Chris Lattner
eed4657594 merge a bunch of tests together into one, convert to filecheck which
is more tolerant of whitespace differences.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78439 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-07 23:56:42 +00:00
Bob Wilson
e9ce5d5ef9 Convert more Neon tests to use FileCheck.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78433 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-07 23:45:02 +00:00
David Goodwin
767a952a6f Make NEON single-precision FP support the default for cortex-a8 (again).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78430 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-07 23:32:33 +00:00
Kevin Enderby
c114ed711e 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78423 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-07 22:46:00 +00:00
Anton Korobeynikov
32a1b25781 2 more vdup.32 cases
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78419 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-07 22:36:50 +00:00
Evan Cheng
d77c7aba83 Thumb2 32-bit ldm / stm needs .w suffix if submode is ia.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78410 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-07 21:19:10 +00:00
Daniel Dunbar
ef80fe1025 XFAIL 2006-11-06-StackTrace.cpp on powerpc-apple-darwin9 until someone feels
motivated to fix it.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78406 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-07 20:50:58 +00:00
Evan Cheng
1d2426c470 Fix support to use NEON for single precision fp math.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78397 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-07 19:30:41 +00:00
Andreas Bolka
bdab0e9695 Fix Strong-SIV testcase.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78384 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-07 15:42:32 +00:00
Daniel Dunbar
7ddca30f31 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78379 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-07 08:45:03 +00:00