OSDN Git Service
Rafael Espindola [Mon, 2 Dec 2013 23:39:26 +0000 (23:39 +0000)]
Refactor the setting of PrivateGlobalPrefix.
No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196170
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Mon, 2 Dec 2013 23:26:31 +0000 (23:26 +0000)]
Don't set PrivateGlobalPrefix twice in the same function.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196169
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Mon, 2 Dec 2013 23:04:51 +0000 (23:04 +0000)]
Convert two char* that are only ever used as booleans to bool.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196168
91177308-0d34-0410-b5e6-
96231b3b80d8
Kay Tiong Khoo [Mon, 2 Dec 2013 22:23:32 +0000 (22:23 +0000)]
Use local variable for repeated use rather than 'get' method. No functional change intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196164
91177308-0d34-0410-b5e6-
96231b3b80d8
Kay Tiong Khoo [Mon, 2 Dec 2013 22:20:40 +0000 (22:20 +0000)]
Move variables to where they are used and give them better names. No functional change intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196163
91177308-0d34-0410-b5e6-
96231b3b80d8
Kay Tiong Khoo [Mon, 2 Dec 2013 22:11:56 +0000 (22:11 +0000)]
Rename variables to be consistent (CST -> Cst). No functional change intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196161
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Mon, 2 Dec 2013 22:11:08 +0000 (22:11 +0000)]
Remove unnecessary/commented-out header inclusion.
Review feedback from Eric Christopher on r196140
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196160
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Mon, 2 Dec 2013 22:09:48 +0000 (22:09 +0000)]
DebugInfo: Rename generic unit references to "TheU" instead of TheCU now that they might be type units instead of compile units.
CR feedback from Eric Christopher on r196139.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196159
91177308-0d34-0410-b5e6-
96231b3b80d8
Manman Ren [Mon, 2 Dec 2013 21:29:56 +0000 (21:29 +0000)]
Debug Info: drop debug info via upgrading path if version number does not match.
Add a helper function getDebugInfoVersionFromModule to return the debug info
version number for a module.
"Verifier/module-flags-1.ll" checks for verification errors.
It will seg fault when calling getDebugInfoVersionFromModule because of the
incorrect format for module flags in the testing case. We make
getModuleFlagsMetadata more robust by checking for error conditions.
PR17982
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196158
91177308-0d34-0410-b5e6-
96231b3b80d8
Manman Ren [Mon, 2 Dec 2013 21:25:56 +0000 (21:25 +0000)]
Update Ocaml/vmcore.ml to emit a "Debug Info Version" module flag.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196156
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Mon, 2 Dec 2013 21:05:16 +0000 (21:05 +0000)]
[AArch64] Implemented vcopy_lane patterns using scalar DUP instruction.
Patch by Ana Pazos!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196151
91177308-0d34-0410-b5e6-
96231b3b80d8
Mark Seaborn [Mon, 2 Dec 2013 20:50:59 +0000 (20:50 +0000)]
InlineFunction.cpp: Remove a return value that is always false
Remove some associated dead code.
This cleanup is associated with PR17872.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196147
91177308-0d34-0410-b5e6-
96231b3b80d8
Manman Ren [Mon, 2 Dec 2013 20:09:52 +0000 (20:09 +0000)]
Debug Info: Move the constant for Debug Info Version from Dwarf.h to Metadata.h.
Suggested by Eric.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196144
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Mon, 2 Dec 2013 19:33:15 +0000 (19:33 +0000)]
DebugInfo: Rename DwarfCompileUnit.* to DwarfUnit.* to match their contents.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196140
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Mon, 2 Dec 2013 19:33:10 +0000 (19:33 +0000)]
DebugInfo: Refactor CompileUnit into a Unit baseclass and CompileUnit/TypeUnit derived classes.
Header/cpp file rename to follow immediately - just splitting out the
commits for ease of review/reading to demonstrate that the renaming
changes are entirely mechanical.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196139
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Mon, 2 Dec 2013 18:44:29 +0000 (18:44 +0000)]
DebugInfo: Type Units: Propagate the correct DW_AT_language into type units.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196130
91177308-0d34-0410-b5e6-
96231b3b80d8
Kay Tiong Khoo [Mon, 2 Dec 2013 18:43:59 +0000 (18:43 +0000)]
Conservative fix for PR17827 - don't optimize a shift + and + compare sequence where the shift is logical unless the comparison is unsigned
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196129
91177308-0d34-0410-b5e6-
96231b3b80d8
Vincent Lejeune [Mon, 2 Dec 2013 17:29:37 +0000 (17:29 +0000)]
R600: Workaround for cayman loop bug
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196121
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Mon, 2 Dec 2013 16:25:47 +0000 (16:25 +0000)]
Move getSymbolWithGlobalValueBase to TargetLoweringObjectFile.
This allows it to be used in TargetLoweringObjectFileImpl.cpp.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196117
91177308-0d34-0410-b5e6-
96231b3b80d8
Alp Toker [Mon, 2 Dec 2013 16:06:06 +0000 (16:06 +0000)]
Introduce poor man's consumeToken() in X86AsmParser
This makes the code a little more idiomatic.
No change in behaviour.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196113
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Mon, 2 Dec 2013 15:36:37 +0000 (15:36 +0000)]
Remove dead code.
MO_JumpTableIndex and MO_ExternalSymbol don't show up on inline asm.
Keeping parts of the old asm printer just to print inline asm to a string that
we then parse back looks like a hack.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196111
91177308-0d34-0410-b5e6-
96231b3b80d8
Diego Novillo [Mon, 2 Dec 2013 15:12:50 +0000 (15:12 +0000)]
Add tests for profile sample file parsing.
The profile file parser needed some tests for its parsing actions.
This adds tests for each of the error messages emitted by the parser.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196106
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Mon, 2 Dec 2013 14:59:34 +0000 (14:59 +0000)]
Output .eh_frames on COFF too now that the integrated as is used on mingw.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196104
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Mon, 2 Dec 2013 14:46:26 +0000 (14:46 +0000)]
ARM: decide whether to use movw/movt based on "minsize" attribute.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196102
91177308-0d34-0410-b5e6-
96231b3b80d8
Alp Toker [Mon, 2 Dec 2013 14:17:47 +0000 (14:17 +0000)]
Cut the gold plugin README down to size
This file hasn't been updated in years. Remove old information and point to
the current documentation at GoldPlugin.rst.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196100
91177308-0d34-0410-b5e6-
96231b3b80d8
Diego Novillo [Mon, 2 Dec 2013 14:08:27 +0000 (14:08 +0000)]
Fix dominator descendants for unreachable blocks.
When a block is unreachable, asking its dom tree descendants should
return the empty set. However, the computation of the descendants
was causing a segmentation fault because the dom tree node we get
from the basic block is initially NULL.
Fixed by adding a test for a valid dom tree node before we iterate.
The patch also adds some unit tests to the existing dom tree tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196099
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Mon, 2 Dec 2013 12:35:56 +0000 (12:35 +0000)]
[PM] [cleanup] Rearrange the public and private sections of this class
to be a bit more sensible. The public interface now is first followed by
the implementation details.
This also resolves a FIXME to make something private -- it was already
possible as the one special caller was already a friend.
No functionality changed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196095
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Mon, 2 Dec 2013 11:31:25 +0000 (11:31 +0000)]
XCoreFrameLowering.cpp: Use [in,out] instead of [in] [out]. [-Wdocumentation]
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196094
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Mon, 2 Dec 2013 11:31:19 +0000 (11:31 +0000)]
[CMake] add_lit_target: Tests should be excluded from "Build Solution".
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196093
91177308-0d34-0410-b5e6-
96231b3b80d8
Robert Lytton [Mon, 2 Dec 2013 11:05:28 +0000 (11:05 +0000)]
XCore target: Make handling of large frames not dependent upon an FP.
eliminateFrameIndex() has been reworked to handle both small & large frames
with either a FP or SP.
An additional Slot is required for Scavenging spills when not using FP for large frames.
Reworked the handling of Register Scavenging.
Whether we are using an FP or not, whether it is a large frame or not,
and whether we are using a large code model or not are now independent.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196091
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Mon, 2 Dec 2013 10:35:41 +0000 (10:35 +0000)]
ARM: add pseudo-instructions for lit-pool global materialisation
These are used by MachO only at the moment, and (much like the existing
MOVW/MOVT set) work around the fact that the labels used in the actual
instructions often contain PC-dependent components, which means that repeatedly
materialising the same global can't be CSEed.
With small modifications, it could be adapted to how ELF finds the address of
_GLOBAL_OFFSET_TABLE_, which would give similar benefits in PIC mode there.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196090
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Mon, 2 Dec 2013 10:29:26 +0000 (10:29 +0000)]
XCore: Unbreak C++11 build.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196089
91177308-0d34-0410-b5e6-
96231b3b80d8
Robert Lytton [Mon, 2 Dec 2013 10:18:37 +0000 (10:18 +0000)]
XCore target: fix large code model 'select' indirect address handling.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196088
91177308-0d34-0410-b5e6-
96231b3b80d8
Robert Lytton [Mon, 2 Dec 2013 10:18:31 +0000 (10:18 +0000)]
XCore target: Add large code model
When using large code model:
Global objects larger than 'CodeModelLargeSize' bytes are placed in sections named with a trailing ".large"
The folded global address of such objects are lowered into the const pool.
During inspection it was noted that LowerConstantPool() was using a default offset of zero.
A fix was made, but due to only offsets of zero being generated, testing only verifies the change is not detrimental.
Correct the flags emitted for explicitly specified sections.
We assume the size of the object queried by getSectionForConstant() is never greater than CodeModelLargeSize.
To handle greater than CodeModelLargeSize, changes to AsmPrinter would be required.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196087
91177308-0d34-0410-b5e6-
96231b3b80d8
Robert Lytton [Mon, 2 Dec 2013 10:18:24 +0000 (10:18 +0000)]
XCore target: extend tests in preparation
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196086
91177308-0d34-0410-b5e6-
96231b3b80d8
Robert Lytton [Mon, 2 Dec 2013 10:18:19 +0000 (10:18 +0000)]
XCore target: Fix eliminateFrameIndex() to handle large frames
Large frame offsets are loaded from the ConstantPool.
Where possible, offsets are encoded using the smaller MKMSK instruction.
Large frame offsets can only be used when there is a frame-pointer.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196085
91177308-0d34-0410-b5e6-
96231b3b80d8
Robert Lytton [Mon, 2 Dec 2013 10:18:14 +0000 (10:18 +0000)]
XCore target: Enable frames larger than 65535 to be lowered
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196084
91177308-0d34-0410-b5e6-
96231b3b80d8
Alexander Kornienko [Mon, 2 Dec 2013 09:28:49 +0000 (09:28 +0000)]
Remove an empty directory left after r194304, which incompletely reverted r194218
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196081
91177308-0d34-0410-b5e6-
96231b3b80d8
Kostya Serebryany [Mon, 2 Dec 2013 08:07:15 +0000 (08:07 +0000)]
[tsan] fix instrumentation of vector vptr updates (https://code.google.com/p/thread-sanitizer/issues/detail?id=43)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196079
91177308-0d34-0410-b5e6-
96231b3b80d8
Alp Toker [Mon, 2 Dec 2013 07:15:33 +0000 (07:15 +0000)]
Update the LTO GoldPlugin documentation
* Update build instructions to reflect the current source tree layout.
* Don't inflict CVS on readers; there's a perfectly good git mirror.
* configure with --disable-werror making it possible to build using clang.
* ar and nm-new now support the -plugin option.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196069
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Mon, 2 Dec 2013 05:42:16 +0000 (05:42 +0000)]
Remove leftovers from a non-MC asm printer.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196068
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Mon, 2 Dec 2013 05:24:28 +0000 (05:24 +0000)]
Remove #if 0 declarations.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196067
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Mon, 2 Dec 2013 05:10:04 +0000 (05:10 +0000)]
Remove dead code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196066
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Mon, 2 Dec 2013 04:55:42 +0000 (04:55 +0000)]
Change the default of AsmWriterClassName and isMCAsmWriter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196065
91177308-0d34-0410-b5e6-
96231b3b80d8
Alp Toker [Mon, 2 Dec 2013 04:31:36 +0000 (04:31 +0000)]
Rename test with misspelt filename
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196064
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Mon, 2 Dec 2013 04:18:19 +0000 (04:18 +0000)]
Remove dead declarations.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196063
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Mon, 2 Dec 2013 03:26:43 +0000 (03:26 +0000)]
Refactor for clarity and efficiency.
The PPC GetSymbolFromOperand already prefixed stubs of MO_ExternalSymbol, so
this should be a nop.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196059
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Sun, 1 Dec 2013 21:24:30 +0000 (21:24 +0000)]
Also test the created stubs on 32 bits.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196052
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Sun, 1 Dec 2013 18:17:05 +0000 (18:17 +0000)]
Add -mcpu to stackmap.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196051
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Sun, 1 Dec 2013 14:16:24 +0000 (14:16 +0000)]
ARM: fix bug in -Oz stack adjustment folding
Previously, we clobbered callee-saved registers when folding an "add
sp, #N" into a "pop {rD, ...}" instruction. This change checks whether
a register we're going to add to the "pop" could actually be live
outside the function before doing so and should fix the issue.
This should fix PR18081.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196046
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sun, 1 Dec 2013 11:47:42 +0000 (11:47 +0000)]
Revamp error checking in the ms inline asm parser.
- Actually abort when an error occurred.
- Check that the frontend lookup worked when parsing length/size/type operators.
Tested by a clang test. PR18096.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196044
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Kuperstein [Sun, 1 Dec 2013 10:16:35 +0000 (10:16 +0000)]
Ensure bitcode encoding of linkage types stays stable. Patch by Boaz Ouriel
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196042
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sun, 1 Dec 2013 03:40:42 +0000 (03:40 +0000)]
Use accessor methods instead.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196006
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sun, 1 Dec 2013 03:36:07 +0000 (03:36 +0000)]
Use 'unsigned char' to get this past gcc error message:
error: invalid conversion from 'unsigned char' to '{anonymous}::Sequence'
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@196004
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Sat, 30 Nov 2013 20:55:12 +0000 (20:55 +0000)]
Add a scheduling model (with itinerary) for the PPC POWER7
This adds a scheduling model for the POWER7 (P7) core, and enables the
machine-instruction scheduler when targeting the P7. Scheduling for the P7,
like earlier ooo PPC cores, requires considering both dispatch group hazards,
and functional unit resources and latencies. These are both modeled in a
combined itinerary. Dispatch group formation is still handled by the post-RA
scheduler (which still needs to be updated for the P7, but nevertheless does a
pretty good job).
One interesting aspect of this change is that I've also enabled to use of AA
duing CodeGen for the P7 (just as it is for the embedded cores). The benchmark
results seem to support this decision (see below), and while this is normally
useful for in-order cores, and not for ooo cores like the P7, I think that the
dispatch slot hazards are enough like in-order resources to make the AA useful.
Test suite significant performance differences (where negative is a speedup,
and positive is a regression) vs. the current situation:
MultiSource/Benchmarks/BitBench/drop3/drop3
with AA: N/A
without AA: -28.7614% +/- 19.8356%
(significantly against AA)
MultiSource/Benchmarks/FreeBench/neural/neural
with AA: -17.7406% +/- 11.2712%
without AA: N/A
(significantly in favor of AA)
MultiSource/Benchmarks/SciMark2-C/scimark2
with AA: -11.2079% +/- 1.80543%
without AA: -11.3263% +/- 2.79651%
MultiSource/Benchmarks/TSVC/Symbolics-flt/Symbolics-flt
with AA: -41.8649% +/- 17.0053%
without AA: -34.5256% +/- 23.7072%
MultiSource/Benchmarks/mafft/pairlocalalign
with AA: 25.3016% +/- 17.8614%
without AA: 38.6629% +/- 14.9391%
(significantly in favor of AA)
MultiSource/Benchmarks/sim/sim
with AA: N/A
without AA: 13.4844% +/- 7.18195%
(significantly in favor of AA)
SingleSource/Benchmarks/BenchmarkGame/Large/fasta
with AA: 15.0664% +/- 6.70216%
without AA: 12.7747% +/- 8.43043%
SingleSource/Benchmarks/BenchmarkGame/puzzle
with AA: 82.2713% +/- 26.3567%
without AA: 75.7525% +/- 41.1842%
SingleSource/Benchmarks/Misc/flops-2
with AA: -37.1621% +/- 20.7964%
without AA: -35.2342% +/- 20.2999%
(significantly in favor of AA)
These are 99.5% confidence intervals from 5 runs per configuration. Regarding
the choice to turn on AA during CodeGen, of these results, four seem
significantly in favor of using AA, and one seems significantly against. I'm
not making this decision based on these numbers alone, but these results
seem consistent with results I have from other tests, and so I think that, on
balance, using AA is a win.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195981
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Sat, 30 Nov 2013 20:41:13 +0000 (20:41 +0000)]
Split some PPC itinerary classes
In preparation for adding scheduling definitions for the POWER7, split some PPC
itinerary classes so that the P7's latencies and hazards can be better
described. For the most part, this means differentiating indexed from non-index
pre-increment loads and stores. Also, differentiate single from
double-precision sqrt.
No functionality change intended (except for a more-specific latency for
single-precision sqrt on the A2).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195980
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Sat, 30 Nov 2013 20:04:33 +0000 (20:04 +0000)]
Convert a PPC test from grep to FileCheck
Convert this test to FileCheck, and improve it to check for the instructions it
is trying to exclude instead of checking for register use (especially because
grepping for r1 can be thrown off, for example, by a use of r12).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195979
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Sat, 30 Nov 2013 19:52:28 +0000 (19:52 +0000)]
Desensitize a couple of PPC regression tests
Use CHECK-DAG to make these regression tests more resilient against changes in
instruction scheduling.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195978
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Sat, 30 Nov 2013 19:39:27 +0000 (19:39 +0000)]
Update the cpu specified on some PPC regression tests
Some of these tests did not specify a cpu but were also sensitive to
instruction scheduling and/or register assignment choices. A few others
similarly-sensitive tests specified a cpu (often the POWER7), and while the P7
currently uses the default model for PPC64, this will soon change. For those
tests which should not really be cpu-dependent anyway, the cpu is set to the
generic 'ppc64'.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195977
91177308-0d34-0410-b5e6-
96231b3b80d8
Zoran Jovanovic [Sat, 30 Nov 2013 19:13:15 +0000 (19:13 +0000)]
Test case for issue with microMIPS long branch.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195976
91177308-0d34-0410-b5e6-
96231b3b80d8
Zoran Jovanovic [Sat, 30 Nov 2013 19:12:28 +0000 (19:12 +0000)]
Fixed issue with microMIPS long branch.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195975
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Sanders [Sat, 30 Nov 2013 13:47:57 +0000 (13:47 +0000)]
[mips][msa] MSA loads and stores have a 10-bit offset. Account for this when lowering FrameIndex.
This prevents the compiler from emitting invalid ld.[bhwd]'s and st.[bhwd]'s
when the stack frame is between 512 and 32,768 bytes in size.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195973
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Sanders [Sat, 30 Nov 2013 13:15:21 +0000 (13:15 +0000)]
[mips][msa] A small refactor to reduce patch noise in my next commit
No functional change. An if-statement has been split into two nested if-statements.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195972
91177308-0d34-0410-b5e6-
96231b3b80d8
Juergen Ributzka [Sat, 30 Nov 2013 03:07:16 +0000 (03:07 +0000)]
Force CPU type to unbreak unit tests on Haswell machines.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195971
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Fri, 29 Nov 2013 23:49:38 +0000 (23:49 +0000)]
Reverse the order of eviction checks for possible compile time savings. No functionality.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195969
91177308-0d34-0410-b5e6-
96231b3b80d8
Reed Kotler [Fri, 29 Nov 2013 22:32:56 +0000 (22:32 +0000)]
Part 1 of 3 patches that completes very long conditional branches
in constant islands for Mips16. We introdcuce JalB16 as a synomnym
for Jal16. It makes it easier to read and is also necessary because
Jal16 is a call instruction but JalB16 is being used as a branch.
Various parts of LLVM will not work properly even in this late stage of
the backend if we use what was declared as a call instruction to function
as a branch. For one, basic block labels may not get emitted in some
situations.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195968
91177308-0d34-0410-b5e6-
96231b3b80d8
Zoran Jovanovic [Fri, 29 Nov 2013 22:10:02 +0000 (22:10 +0000)]
Revert revision 195965.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195967
91177308-0d34-0410-b5e6-
96231b3b80d8
Petar Jovanovic [Fri, 29 Nov 2013 21:59:09 +0000 (21:59 +0000)]
mips: XFAIL llvm-cov test
XFAIL llvm-cov.test for MIPS until big-endian issues are fixed for llvm-cov.
The test does pass on MIPS little-endian.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195966
91177308-0d34-0410-b5e6-
96231b3b80d8
Zoran Jovanovic [Fri, 29 Nov 2013 21:41:24 +0000 (21:41 +0000)]
Fixed issue with microMIPS long branch.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195965
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Fri, 29 Nov 2013 07:04:59 +0000 (07:04 +0000)]
Adjust PPC A2 input operand latencies
On the PPC A2, instructions are only issued after their input operands are
ready. Model this by specifying that input operands are read at dispatch (0
cycles after issue). This changes all input operand latencies from 1 to 0.
Significant test-suite performance changes (these are 99.5% confidence
intervals on 6 runs for both before and after):
speedups:
MultiSource/Benchmarks/sim/sim
-1.21915% +/- 0.175063%
MultiSource/Benchmarks/TSVC/LinearDependence-flt/LinearDependence-flt
-1.23946% +/- 1.05133%
SingleSource/Benchmarks/Misc/flops-2
-1.24237% +/- 0.681362%
MultiSource/Applications/JM/lencod/lencod
-1.33992% +/- 0.757498%
MultiSource/Benchmarks/TSVC/InductionVariable-flt/InductionVariable-flt
-1.51802% +/- 1.21468%
MultiSource/Benchmarks/TSVC/GlobalDataFlow-flt/GlobalDataFlow-flt
-2.18818% +/- 1.28605%
MultiSource/Benchmarks/TSVC/Packing-flt/Packing-flt
-2.21977% +/- 1.19499%
SingleSource/Benchmarks/BenchmarkGame/spectral-norm
-2.29822% +/- 0.671871%
MultiSource/Benchmarks/TSVC/Packing-dbl/Packing-dbl
-2.40975% +/- 0.355931%
SingleSource/Benchmarks/Misc/fp-convert
-2.41899% +/- 1.04751%
MultiSource/Benchmarks/TSVC/Searching-dbl/Searching-dbl
-2.50349% +/- 0.126765%
SingleSource/Benchmarks/Misc/flops-3
-3.00214% +/- 0.700795%
MultiSource/Benchmarks/TSVC/LoopRestructuring-flt/LoopRestructuring-flt
-3.56995% +/- 3.2929%
MultiSource/Applications/sgefa/sgefa
-4.24908% +/- 2.00413%
MultiSource/Benchmarks/ASC_Sequoia/IRSmk/IRSmk
-18.1294% +/- 3.96489%
regressions:
MultiSource/Benchmarks/TSVC/Reductions-dbl/Reductions-dbl
1.03249% +/- 0.178547%
MultiSource/Applications/hexxagon/hexxagon
1.16597% +/- 0.285235%
MultiSource/Benchmarks/TSVC/IndirectAddressing-flt/IndirectAddressing-flt
1.39576% +/- 1.07855%
SingleSource/Benchmarks/Misc-C++/stepanov_v1p2
1.71539% +/- 0.173182%
MultiSource/Benchmarks/Fhourstones-3.1/fhourstones3.1
1.90013% +/- 0.866472%
MultiSource/Benchmarks/TSVC/Recurrences-dbl/Recurrences-dbl
2.39854% +/- 1.05914%
MultiSource/Benchmarks/TSVC/ControlFlow-dbl/ControlFlow-dbl
2.4402% +/- 0.817904%
MultiSource/Benchmarks/TSVC/LoopRestructuring-dbl/LoopRestructuring-dbl
5.87997% +/- 3.3172%
MultiSource/Benchmarks/Trimaran/netbench-crc/netbench-crc
9.02643% +/- 5.79591%
MultiSource/Benchmarks/VersaBench/bmm/bmm
10.3517% +/- 1.227%
Obviously, there are data points on both sides of this; but I think, overall,
this supports making the change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195951
91177308-0d34-0410-b5e6-
96231b3b80d8
Lang Hames [Fri, 29 Nov 2013 06:35:30 +0000 (06:35 +0000)]
Teach LocalStackSlotAllocation that stackmaps/patchpoints don't have range
constraints on their frame offsets.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195950
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Fri, 29 Nov 2013 06:32:17 +0000 (06:32 +0000)]
Create a PPC440 SchedMachineModel
Some of the older PPC processor definitions don't have associated
SchedMachineModels; correct this for the PPC440.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195949
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Fri, 29 Nov 2013 06:19:43 +0000 (06:19 +0000)]
Fixup PPC440 load/store operand latencies
The operand latencies for loads and stores in the PPC440 itinerary were wrong
(the store operands are all inputs, and the "with update" (pre-increment)
instructions need a latency for the additional output).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195948
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Fri, 29 Nov 2013 05:59:00 +0000 (05:59 +0000)]
Adjust PPC440 operand latencies
The operand latencies for the PPC440 should be specified relative to dispatch,
not relative to the initial fetch-and-decode stages. Because most instructions
(ignoring bypass) wait in dispatch until their operands are ready, this is
modeled as reading input operands "at dispatch" (0 cycles after issue), and so
every input and output operand has 4 cycles subtracted from it.
This could alter scheduling slightly, but I don't expect a large effect.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195947
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Fri, 29 Nov 2013 05:58:38 +0000 (05:58 +0000)]
Don't model the fetch and decode units for the PPC440
Modeling the fetch and decode units in the PPC440 itinerary does not add
anything to the hazard detection capability (and so modeling them just wastes
compile time).
No functionality change intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195946
91177308-0d34-0410-b5e6-
96231b3b80d8
Lang Hames [Fri, 29 Nov 2013 03:36:53 +0000 (03:36 +0000)]
Remove unused variable from r195944.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195945
91177308-0d34-0410-b5e6-
96231b3b80d8
Lang Hames [Fri, 29 Nov 2013 03:07:54 +0000 (03:07 +0000)]
Refactor a lot of patchpoint/stackmap related code to simplify and make it
target independent.
Most of the x86 specific stackmap/patchpoint handling was necessitated by the
use of the native address-mode format for frame index operands. PEI has now
been modified to treat stackmap/patchpoint similarly to DEBUG_INFO, allowing
us to use a simple, platform independent register/offset pair for frame
indexes on stackmap/patchpoints.
Notes:
- Folding is now platform independent and automatically supported.
- Emiting patchpoints with direct memory references now just involves calling
the TargetLoweringBase::emitPatchPoint utility method from the target's
XXXTargetLowering::EmitInstrWithCustomInserter method. (See
X86TargetLowering for an example).
- No more ugly platform-specific operand parsers.
This patch shouldn't change the generated output for X86.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195944
91177308-0d34-0410-b5e6-
96231b3b80d8
Hao Liu [Fri, 29 Nov 2013 02:11:22 +0000 (02:11 +0000)]
AArch64: The pattern match should check the range of the immediate value.
Or we can generate some illegal instructions.
E.g. shrn2 v0.4s, v1.2d, #35. The legal range should be in [1, 16].
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195941
91177308-0d34-0410-b5e6-
96231b3b80d8
Jiangning Liu [Fri, 29 Nov 2013 01:38:08 +0000 (01:38 +0000)]
Add missing test case for bsl_f64 support of AArch64 NEON.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195939
91177308-0d34-0410-b5e6-
96231b3b80d8
Jiangning Liu [Fri, 29 Nov 2013 01:37:15 +0000 (01:37 +0000)]
Add missing pattern for supporting intrinsic function vbsl_f64 with
argument double floating point.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195938
91177308-0d34-0410-b5e6-
96231b3b80d8
Kevin Qin [Fri, 29 Nov 2013 01:29:16 +0000 (01:29 +0000)]
[AArch64 NEON]Fix a assertion failure when disassemble SHLL instruction.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195936
91177308-0d34-0410-b5e6-
96231b3b80d8
Stephen Canon [Thu, 28 Nov 2013 21:38:05 +0000 (21:38 +0000)]
Rein in overzealous InstCombine of fptrunc(OP(fpextend, fpextend)).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195934
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 28 Nov 2013 20:12:44 +0000 (20:12 +0000)]
Refactor to remove a bit of duplication. No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195933
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Thu, 28 Nov 2013 19:58:56 +0000 (19:58 +0000)]
Silence sign-compare warning and reduce nesting.
No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195932
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 28 Nov 2013 19:35:07 +0000 (19:35 +0000)]
Remove an always true parameter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195931
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Thu, 28 Nov 2013 17:04:31 +0000 (17:04 +0000)]
[CMake] Let add_public_tablegen_target() provide intrinsics_gen, too.
I think, in principle, intrinsics_gen may be added explicitly.
That said, it can be added incidentally, since each target already has dependencies to llvm-tblgen.
Almost all source files depend on both CommonTaleGen and intrinsics_gen.
Explicit add_dependencies() have been pruned under lib/Target.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195929
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Thu, 28 Nov 2013 17:04:13 +0000 (17:04 +0000)]
[CMake] Also OptionTests can be free from add_dependencies() with add_public_tablegen_target().
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195928
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Thu, 28 Nov 2013 17:04:04 +0000 (17:04 +0000)]
[CMake] Let add_public_tablegen_target responsible to provide dependency to CommonTableGen.
add_public_tablegen_target adds *CommonTableGen to LLVM_COMMON_DEPENDS.
LLVM_COMMON_DEPENDS affects add_llvm_library (and other add_target stuff) within its scope.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195927
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 28 Nov 2013 17:00:49 +0000 (17:00 +0000)]
The global prefix is always one char. Don't use a string for it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195926
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Thu, 28 Nov 2013 15:30:37 +0000 (15:30 +0000)]
[CMake] Prune include_directories() in llvm/lib/Target, take #2.
I forgot to commit them. They were staging in my local repo.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195924
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Sanders [Thu, 28 Nov 2013 15:26:33 +0000 (15:26 +0000)]
[mips] Revert test commit r195922.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195923
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Sanders [Thu, 28 Nov 2013 15:25:43 +0000 (15:25 +0000)]
[mips] A test commit to test my Herald and Audit workflow
Will be reverted in the next commit
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195922
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Thu, 28 Nov 2013 14:53:30 +0000 (14:53 +0000)]
[CMake] Prune include_directories() in llvm/lib/Target. add_llvm_target() sets them.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195921
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Thu, 28 Nov 2013 14:52:52 +0000 (14:52 +0000)]
Add newline at eof.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195920
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Sanders [Thu, 28 Nov 2013 09:36:44 +0000 (09:36 +0000)]
As myself as code-owner of the MIPS backend (lib/Target/Mips/*)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195915
91177308-0d34-0410-b5e6-
96231b3b80d8
Peter Zotov [Thu, 28 Nov 2013 09:03:28 +0000 (09:03 +0000)]
[OCaml] Add a slash accidentally omitted from Makefile
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195912
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 28 Nov 2013 08:59:52 +0000 (08:59 +0000)]
Use the mangler consistently instead of using getGlobalPrefix directly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195911
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Thu, 28 Nov 2013 06:05:59 +0000 (06:05 +0000)]
Don't share functional units among the PPC itineraries
Instead of sharing functional unit names between the various PPC itineraries,
give each core its own unit names prefixed with the core name. This follows
the convention used by other backends (such as ARM), and removes a non-obvious
ordering dependency between the various PPCSchedule*.td files.
No functionality change intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195908
91177308-0d34-0410-b5e6-
96231b3b80d8
Jiangning Liu [Thu, 28 Nov 2013 01:34:55 +0000 (01:34 +0000)]
Remove the variable only used by assert to avoid the build failure
caused by build options [-Werror,-Wunused-variable].
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195905
91177308-0d34-0410-b5e6-
96231b3b80d8
Hao Liu [Thu, 28 Nov 2013 01:07:45 +0000 (01:07 +0000)]
AArch64: Fix a bug about disassembling post-index load single element to 4 vectors
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@195903
91177308-0d34-0410-b5e6-
96231b3b80d8