Commit Graph

220 Commits

Author SHA1 Message Date
Evan Cheng bdb43a9d99 Remat VLDRD from constpool. Clean up some instruction property specifications.
llvm-svn: 89478
2009-11-20 19:57:15 +00:00
Evan Cheng bbd50b0f78 Also CSE non-pic load from constant pools.
llvm-svn: 89440
2009-11-20 02:10:27 +00:00
Evan Cheng b18525937c More consistent thumb1 asm printing.
llvm-svn: 89328
2009-11-19 06:57:41 +00:00
Evan Cheng 2a6c92fcb6 Shrink ldr / str [sp, imm0-1024] to 16-bit instructions.
llvm-svn: 89326
2009-11-19 06:32:27 +00:00
Jim Grosbach cdde77c6a3 Enable arm jumpt table adjustment.
llvm-svn: 89143
2009-11-17 21:24:11 +00:00
Anton Korobeynikov a2873f4d59 Forgot to commit test fixes
llvm-svn: 89138
2009-11-17 20:38:36 +00:00
Jim Grosbach 0ad7efbace Convert to FileCheck
llvm-svn: 89007
2009-11-17 00:20:26 +00:00
Jim Grosbach 4781c3caf8 Convert to FileCheck
llvm-svn: 89002
2009-11-17 00:03:38 +00:00
Jim Grosbach 805d195649 Cleanup. Missed removing these when converting. Oops.
llvm-svn: 89001
2009-11-17 00:00:33 +00:00
Jim Grosbach 1deb0b9f53 Convert to FileCheck
llvm-svn: 88991
2009-11-16 23:19:29 +00:00
Jim Grosbach 9b32e22ad1 Convert to FileCheck
llvm-svn: 88947
2009-11-16 20:04:15 +00:00
Jim Grosbach 980d94164d Convert to FileCheck
llvm-svn: 88942
2009-11-16 19:46:46 +00:00
Jim Grosbach c670bdc311 tbb opt off by default
llvm-svn: 88921
2009-11-16 17:24:45 +00:00
Jim Grosbach 01c1cae34d Detect need for autoalignment of the stack earlier to catch spills more
conservatively. eliminateFrameIndex() machinery adjust to handle addr mode
6 (vld1/vst1) used for spills. Fix tests to expect aligned Q-reg spilling

llvm-svn: 88874
2009-11-15 21:45:34 +00:00
Jim Grosbach f16a3b7a9f remove xfail
llvm-svn: 88817
2009-11-14 21:57:35 +00:00
Evan Cheng 66401c90da When expanding t2STRDi8 r, r to two stores, add kill markers correctly.
llvm-svn: 88734
2009-11-14 01:50:00 +00:00
Jim Grosbach 1025a4998b Clean up testcase a bit. Simplify case blocks and adjust switch instruction to not take an undefined value as input.
llvm-svn: 86997
2009-11-12 17:19:09 +00:00
Benjamin Kramer 5218176bc6 Fix typo in run line.
llvm-svn: 86984
2009-11-12 12:35:27 +00:00
Evan Cheng 5d85a46f76 RegScavenger::enterBasicBlock should always reset register state.
llvm-svn: 86972
2009-11-12 07:49:10 +00:00
Evan Cheng 85a9f430e9 - Teach LSR to avoid changing cmp iv stride if it will create an immediate that
cannot be folded into target cmp instruction.
- Avoid a phase ordering issue where early cmp optimization would prevent the
  later count-to-zero optimization.
- Add missing checks which could cause LSR to reuse stride that does not have
  users.
- Fix a bug in count-to-zero optimization code which failed to find the pre-inc
  iv's phi node.
- Remove, tighten, loosen some incorrect checks disable valid transformations.
- Quite a bit of code clean up.

llvm-svn: 86969
2009-11-12 07:35:05 +00:00
Dan Gohman 64b5d0f468 Add support for tail duplication to BranchFolding, and extend
tail merging support to handle more cases.
 - Recognize several cases where tail merging is beneficial even when
   the tail size is smaller than the generic threshold.
 - Make use of MachineInstrDesc::isBarrier to help detect
   non-fallthrough blocks.
 - Check for and avoid disrupting fall-through edges in more cases.

llvm-svn: 86871
2009-11-11 19:48:59 +00:00
Jim Grosbach 47e3bcf180 Update test
llvm-svn: 86614
2009-11-09 22:59:01 +00:00
Jim Grosbach d7cf55cd0e Use Unified Assembly Syntax for the ARM backend.
llvm-svn: 86494
2009-11-09 00:11:35 +00:00
Evan Cheng a8e8a7c976 Refactor code. Fix a potential missing check. Teach isIdentical() about tLDRpci_pic.
llvm-svn: 86330
2009-11-07 04:04:34 +00:00
Evan Cheng 7ff831962a - Add TargetInstrInfo::isIdentical(). It's similar to MachineInstr::isIdentical
except it doesn't care if the definitions' virtual registers differ. This is
  used by machine LICM and other MI passes to perform CSE.
- Teach Thumb2InstrInfo::isIdentical() to check two t2LDRpci_pic are identical.
  Since pc relative constantpool entries are always different, this requires it
  it check if the values can actually the same.

llvm-svn: 86328
2009-11-07 03:52:02 +00:00
Evan Cheng 207b246650 - Add pseudo instructions tLDRpci_pic and t2LDRpci_pic which does a pc-relative
load of a GV from constantpool and then add pc. It allows the code sequence to
  be rematerializable so it would be hoisted by machine licm.
- Add a late pass to break these pseudo instructions into a number of real
  instructions. Also move the code in Thumb2 IT pass that breaks up t2MOVi32imm
  to this pass. This is done before post regalloc scheduling to allow the
  scheduler to proper schedule these instructions. It also allow them to be
  if-converted and shrunk by later passes.

llvm-svn: 86304
2009-11-06 23:52:48 +00:00
Bob Wilson db42ca663b Fix a broken test.
llvm-svn: 86298
2009-11-06 23:06:42 +00:00
Evan Cheng 8f4e3d99c9 Fix test.
llvm-svn: 85986
2009-11-04 00:42:33 +00:00
Evan Cheng 8d681f0471 Fix PR5367. QPR_8 is the super regclass of DPR_8 and SPR_8.
llvm-svn: 85871
2009-11-03 05:52:54 +00:00
Anton Korobeynikov 2c2dc9f64f Temporary xfail until PR5367 will be resolved
llvm-svn: 85848
2009-11-03 00:37:36 +00:00
Evan Cheng 1708b06c0e Unbreak ARMBaseRegisterInfo::copyRegToReg.
llvm-svn: 85787
2009-11-02 04:44:55 +00:00
Evan Cheng 43219997b6 Make use of imm12 version of Thumb2 ldr / str instructions more aggressively.
llvm-svn: 85743
2009-11-01 21:12:51 +00:00
Evan Cheng 50bc004b67 Fix tests.
llvm-svn: 85723
2009-11-01 18:13:29 +00:00
Evan Cheng 6f29ad9170 Use cbz and cbnz instructions.
llvm-svn: 85698
2009-10-31 23:46:45 +00:00
Jim Grosbach 5cba8de2c8 vml[as].f32 cause stalls in following advanced SIMD instructions. Avoid using
them for scalar floating point operations for now.

llvm-svn: 85697
2009-10-31 22:57:36 +00:00
Jim Grosbach 403202aef1 Consolidate test files
llvm-svn: 85696
2009-10-31 22:20:56 +00:00
Jim Grosbach c79fb530d4 Change to use FileCheck
llvm-svn: 85695
2009-10-31 22:16:14 +00:00
Jim Grosbach 69f364babc Make tests more explicit about which instructions are expected.
llvm-svn: 85694
2009-10-31 22:14:17 +00:00
Jim Grosbach 259c37cc55 Grammar tweak to comments
llvm-svn: 85693
2009-10-31 22:12:44 +00:00
Jim Grosbach 2c3e618a06 Update test to be more explicit about what instruction sequences are expected for each operation.
llvm-svn: 85691
2009-10-31 22:10:38 +00:00
Benjamin Kramer 6ef6fe1c31 Force triple; darwin's ASM syntax differs from linux's.
llvm-svn: 85676
2009-10-31 19:54:06 +00:00
Benjamin Kramer 7e06083a3a Add missing colons for FileCheck.
llvm-svn: 85674
2009-10-31 19:22:24 +00:00
Evan Cheng cdbb70c065 It's safe to remat t2LDRpci; Add PseudoSourceValue to load / store's to enable more machine licm. More changes coming.
llvm-svn: 85643
2009-10-31 03:39:36 +00:00
Bob Wilson 3d43b38f0f Fix Thumb2 failures by converting them to FileCheck.
llvm-svn: 85210
2009-10-27 06:31:02 +00:00
Evan Cheng 69140ec4fa Add a couple of ARM cross-rc coalescing tests.
llvm-svn: 85051
2009-10-25 08:01:41 +00:00
Evan Cheng b9f3520660 Update tests.
llvm-svn: 85050
2009-10-25 07:53:48 +00:00
Jim Grosbach a93ca3c637 Improve handling of immediates by splitting 32-bit immediates into two 16-bit
immediate operands when they will fit into the using instruction.

llvm-svn: 84778
2009-10-21 20:44:34 +00:00
Evan Cheng 786b15fe12 Match more patterns to movt.
llvm-svn: 84751
2009-10-21 08:15:52 +00:00
Dan Gohman 682a2d154a Revert r84658 and r84691. They were causing llvm-gcc bootstrap to fail.
llvm-svn: 84727
2009-10-21 01:44:44 +00:00
David Goodwin baf6dd26ea Checkpoint more aggressive anti-dependency breaking for post-ra scheduler.
llvm-svn: 84658
2009-10-20 19:54:44 +00:00
Sandeep Patel 3f23601b00 Branches must be the last instruction in a Thumb2 IT block. Approved by Evan Cheng.
llvm-svn: 84212
2009-10-15 22:25:32 +00:00
Evan Cheng 4ad726b4be Fix tests.
llvm-svn: 83241
2009-10-02 06:53:57 +00:00
David Goodwin 1cc6dd97da Remove neonfp attribute and instead set default based on CPU string. Add -arm-use-neon-fp to override the default.
llvm-svn: 83218
2009-10-01 22:19:57 +00:00
Evan Cheng 3ea1ba7739 Forgot this test earlier.
llvm-svn: 83143
2009-09-30 08:41:27 +00:00
Evan Cheng 83e0d481ae Make ARM and Thumb2 32-bit immediate materialization into a single 32-bit pseudo
instruction. This makes it re-materializable.

Thumb2 will split it back out into two instructions so IT pass will generate the
right mask. Also, this expose opportunies to optimize the movw to a 16-bit move.

llvm-svn: 82982
2009-09-28 09:14:39 +00:00
Evan Cheng a6b9cab822 Enable pre-regalloc load / store multiple pass for Thumb2.
llvm-svn: 82893
2009-09-27 09:46:04 +00:00
Daniel Dunbar ccde96e96b "Update" tests for -disable-if-conversion removal. I think branch.ll should just
be removed, but I XFAIL'd it for now.

llvm-svn: 82847
2009-09-26 05:29:36 +00:00
Dan Gohman a080159a7c Convert more tests to avoid llvm-as.
llvm-svn: 81545
2009-09-11 18:36:27 +00:00
Evan Cheng cf61d68eaf Cast MO.getImm() to unsigned before comparing with an unsigned limit.
llvm-svn: 81318
2009-09-09 06:05:16 +00:00
Dan Gohman c8054d90fb Eliminate more uses of llvm-as and llvm-dis.
llvm-svn: 81293
2009-09-09 00:09:15 +00:00
Chris Lattner f87421f088 update various tests for signedness changes in .s file.
llvm-svn: 81289
2009-09-08 23:51:06 +00:00
Chris Lattner 9ef94277f1 adjust for signedness change. I'd appreciate it if an ARM flavored person
could look at this: the top undefined bits of an immediate shouldn't affect
isel (cmp vs cmp.w)

llvm-svn: 81288
2009-09-08 23:44:53 +00:00
Chris Lattner 6837964819 merge thumb2-bic2.ll into thumb2-bic.ll and update for signedness changes.
llvm-svn: 81285
2009-09-08 23:41:06 +00:00
Evan Cheng 3d2fce01aa Run branch folding if if-converter make some transformations.
llvm-svn: 80994
2009-09-04 07:47:40 +00:00
Evan Cheng 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
Evan Cheng 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
Evan Cheng 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
Bob Wilson 9054d25808 Fix a typo. Somehow I thought this had passed before, but I guess not.
llvm-svn: 79937
2009-08-24 21:17:17 +00:00
Bob Wilson 5fe1d38607 Convert slow test to use FileCheck.
llvm-svn: 79935
2009-08-24 20:33:47 +00:00
Dan Gohman a41fa35992 Make tail merging handle blocks with repeated predecessors correctly, and
remove RemoveDuplicateSuccessor, as it is no longer necessary, and because
it breaks assumptions made in
MachineBasicBlock::isOnlyReachableByFallthrough.

Convert test/CodeGen/X86/omit-label.ll to FileCheck and add a testcase
for PR4732.

test/CodeGen/Thumb2/thumb2-ifcvt2.ll sees a diff with this commit due to
it being bugpoint-reduced to the point where it doesn't matter what the
condition for the branch is.

Add some more interesting code to
test/CodeGen/X86/2009-08-06-branchfolder-crash.ll, which is the testcase
that originally motivated the RemoveDuplicateSuccessor code, to help
verify that the original problem isn't being re-broken.

llvm-svn: 79338
2009-08-18 15:18:18 +00:00
Evan Cheng dd406177de Fix revsh pattern.
llvm-svn: 79318
2009-08-18 05:43:23 +00:00
Evan Cheng d7e1a79eea Fix tests.
llvm-svn: 79086
2009-08-15 08:23:11 +00:00
Evan Cheng 6ddd7bcdd1 Turn on if-conversion for thumb2.
llvm-svn: 79084
2009-08-15 07:59:10 +00:00
Evan Cheng 7dae88d2c9 Leaf functions which do not save CSRs can be frameless even with -disable-fp-elim.
llvm-svn: 79039
2009-08-14 20:48:13 +00:00
Evan Cheng e41903b10d Also shrink immediate branches; also more assembler workarounds.
llvm-svn: 79014
2009-08-14 18:31:44 +00:00
Evan Cheng db73d68cbe Shrink ADR and LDR from constantpool late during constantpool island pass.
llvm-svn: 78970
2009-08-14 00:32:16 +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
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
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
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
Evan Cheng f72c13bdf5 Handle the constantfp created during post-legalization dag combiner phase.
llvm-svn: 78594
2009-08-10 20:25:59 +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
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
Evan Cheng 4c3b1ca5a0 Fix support to use NEON for single precision fp math.
llvm-svn: 78397
2009-08-07 19:30:41 +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
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
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
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
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
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
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
Evan Cheng 093e124256 Fix a coaelescer bug. If a copy val# is extended to eliminate a non-trivially coalesced copy, and the copy kills its source register. Trim the source register's live range to the last use if possible. This fixes up kill marker to make the scavenger happy.
llvm-svn: 77967
2009-08-03 08:41:59 +00:00
Evan Cheng 8b9deebba3 Use the i12 variant of load / store opcodes if offset is zero. Now we pass all of multisource as well.
llvm-svn: 77939
2009-08-03 02:38:06 +00:00