Chandler Carruth [Fri, 5 Jun 2015 17:01:43 +0000 (17:01 +0000)]
[Unroll] Rework the naming and structure of the new unroll heuristics.
The new naming is (to me) much easier to understand. Here is a summary
of the new state of the world:
- '*Threshold' is the threshold for full unrolling. It is measured
against the estimated unrolled cost as computed by getUserCost in TTI
(or CodeMetrics, etc). We will exceed this threshold when unrolling
loops where unrolling exposes a significant degree of simplification
of the logic within the loop.
- '*PercentDynamicCostSavedThreshold' is the percentage of the loop's
estimated dynamic execution cost which needs to be saved by unrolling
to apply a discount to the estimated unrolled cost.
- '*DynamicCostSavingsDiscount' is the discount applied to the estimated
unrolling cost when the dynamic savings are expected to be high.
When actually analyzing the loop, we now produce both an estimated
unrolled cost, and an estimated rolled cost. The rolled cost is notably
a dynamic estimate based on our analysis of the expected execution of
each iteration.
While we're still working to build up the infrastructure for making
these estimates, to me it is much more clear *how* to make them better
when they have reasonably descriptive names. For example, we may want to
apply estimated (from heuristics or profiles) dynamic execution weights
to the *dynamic* cost estimates. If we start doing that, we would also
need to track the static unrolled cost and the dynamic unrolled cost, as
only the latter could reasonably be weighted by profile information.
This patch is sadly not without functionality change for the new unroll
analysis logic. Buried in the heuristic management were several things
that surprised me. For example, we never subtracted the optimized
instruction count off when comparing against the unroll heursistics!
I don't know if this just got lost somewhere along the way or what, but
with the new accounting of things, this is much easier to keep track of
and we use the post-simplification cost estimate to compare to the
thresholds, and use the dynamic cost reduction ratio to select whether
we can exceed the baseline threshold.
The old values of these flags also don't necessarily make sense. My
impression is that none of these thresholds or discounts have been tuned
yet, and so they're just arbitrary placehold numbers. As such, I've not
bothered to adjust for the fact that this is now a discount and not
a tow-tier threshold model. We need to tune all these values once the
logic is ready to be enabled.
Differential Revision: http://reviews.llvm.org/D9966
llvm-svn: 239164
Frederic Riss [Fri, 5 Jun 2015 16:35:44 +0000 (16:35 +0000)]
[dsymutil] Handle the -oso-prepend-path option when the input is a YAML debug map
All the tests using a YAML debug map will need this.
llvm-svn: 239163
Alexei Starovoitov [Fri, 5 Jun 2015 16:11:14 +0000 (16:11 +0000)]
[bpf] rename triple names bpf_be -> bpfeb
llvm-svn: 239162
Colin LeMahieu [Fri, 5 Jun 2015 16:00:11 +0000 (16:00 +0000)]
[Hexagon] Reapply r239097 with tests corrected for shuffling and duplexing.
llvm-svn: 239161
Joerg Sonnenberger [Fri, 5 Jun 2015 15:54:26 +0000 (15:54 +0000)]
Avoid C-style cast.
llvm-svn: 239160
Alexei Starovoitov [Fri, 5 Jun 2015 15:32:38 +0000 (15:32 +0000)]
Revert "Include BPF target in CMake builds."
This reverts commit r239035
llvm-svn: 239159
Benjamin Kramer [Fri, 5 Jun 2015 14:33:02 +0000 (14:33 +0000)]
[TargetParser] Properly attach functions of ARMTargetParser to the class
llvm-svn: 239158
Benjamin Kramer [Fri, 5 Jun 2015 14:32:54 +0000 (14:32 +0000)]
[ARM] Make helper function static.
This one had a declaration but it differed from the definition so the
declaration was actually dead.
llvm-svn: 239157
James Y Knight [Fri, 5 Jun 2015 14:16:39 +0000 (14:16 +0000)]
[SPARC] Fix windows test failure after r239154.
(Hopefully)
llvm-svn: 239156
Yaron Keren [Fri, 5 Jun 2015 14:15:07 +0000 (14:15 +0000)]
Rangify for loops in LegacyPassManager.cpp.
llvm-svn: 239155
James Y Knight [Fri, 5 Jun 2015 13:44:43 +0000 (13:44 +0000)]
[SPARC] Add multiarch include paths.
Adds tests verifying the proper dirs are found in the Debian 8/GCC4.9
layout for sparc (32bit), sparc (32bit) with lib64 multilib, and
sparc64.
The test cases added here also cover r239047, which fixed the linker
paths.
llvm-svn: 239154
Alexander Musman [Fri, 5 Jun 2015 13:40:59 +0000 (13:40 +0000)]
Fix for PR14269: Clang crashes when a bit field is used as inline assembler
input / output with memory constraint.
One generally can't get address of a bit field, so the general solution is to
error on such cases. GCC does the same.
Patch by Andrey Bokhanko
Differential Revision: http://reviews.llvm.org/D10086
llvm-svn: 239153
John Brawn [Fri, 5 Jun 2015 13:34:11 +0000 (13:34 +0000)]
[ARM] Use TargetParser to determine FPU subtarget features
The main effect of this is to fix anomalies where certain -mfpu options didn't
disable everything that they should causing strange behaviour when combined
with -mcpu or -march values that themselves enabled fpu subtarget features,
e.g. -mfpu=fpv5-dp-d16 with -march=armv7em previously behaved the same as
-mfpu=fpv5-sp-d16 due to fp-only-sp not being disabled.
Invalid -mfpu options now also give an error, which is consistent with the
handling of the .fpu directive.
Differential Revision: http://reviews.llvm.org/D10239
llvm-svn: 239152
John Brawn [Fri, 5 Jun 2015 13:31:19 +0000 (13:31 +0000)]
[ARM] Add support for -sp- FPUs and FPU none to TargetParser
These are added mainly for the benefit of clang, but this also means that they
are now allowed in .fpu directives and we emit the correct .fpu directive when
single-precision-only is used.
Differential Revision: http://reviews.llvm.org/D10238
llvm-svn: 239151
John Brawn [Fri, 5 Jun 2015 13:29:24 +0000 (13:29 +0000)]
[ARM] Add knowledge of FPU subtarget features to TargetParser
Add getFPUFeatures to TargetParser, which gets the list of subtarget features
that are enabled/disabled for each FPU, and use it when handling the .fpu
directive.
No functional change in this commit, though clang will start behaving
differently once it starts using this.
Differential Revision: http://reviews.llvm.org/D10237
llvm-svn: 239150
Daniel Jasper [Fri, 5 Jun 2015 13:18:09 +0000 (13:18 +0000)]
clang-format: More eagerly wrap trailing return types.
Before:
template <typename T>
auto
aaaaaaaaaaaaaaaaaaaaaa(T t) -> decltype(
eaaaaaaaaaaaaaaa<T>(t.a)
.
aaaaaaaa());
After:
template <typename T>
auto
aaaaaaaaaaaaaaaaaaaaaa(T t)
-> decltype(
eaaaaaaaaaaaaaaa<T>(t.a).
aaaaaaaa());
Also add a test case for a difficult template parsing case I stumbled accross.
Needs fixing.
llvm-svn: 239149
Ed Maste [Fri, 5 Jun 2015 13:03:08 +0000 (13:03 +0000)]
Improve OSType initialization in elf object file's arch_spec
Setting the OSType in the ArchSpec triple is needed to correctly setup
up the register context plugin. ArchSpec::SetArchitecture, for Mach-O
only, sets the OSType. For ELF it was left to the ObjectFileELF to fill
in the missing OSType.
This change moves the ObjectFileELF logic into ArchSpec.
A new optional 'os' parameter has been added to SetArchitecture.
For ELF, this value is the from the ELF header.e_ident[EI_OSABI].
The default value is 0 or ELFOSABI_NONE.
The real work of determining the OSType was done by the ObjectFileELF
helper function GetOsFromOSABI. This logic has been moved
SetArchitecture.
GetOsFromOSABI has been commented as being deprectated. It is left in
to support asserts.
For ELF the vendor value returned from SetArchitecture should be
UnknownVendor. An unneeded resetting in ObjectFileELF has been removed
and replaced with an assert.
This fixes a problem reading a core file on FreeBSD/ARM because the spec
triple was arm-unknown-unknown.
Patch by Tom Rix.
Differential Revision: http://reviews.llvm.org/D9292
llvm-svn: 239148
Artyom Skrobov [Fri, 5 Jun 2015 12:39:28 +0000 (12:39 +0000)]
[ARMTargetParser] Follow-up for r239099: one case was missed
llvm-svn: 239147
Simon Pilgrim [Fri, 5 Jun 2015 12:35:36 +0000 (12:35 +0000)]
[X86][AVX2] Added tests for v32i8 vector shifts
Currently still scalarized, but D9474 should remedy that.
llvm-svn: 239146
Toma Tabacu [Fri, 5 Jun 2015 12:19:27 +0000 (12:19 +0000)]
Revert "[mips] [IAS] Restore STI.FeatureBits in .set pop." (r239144).
This is breaking the Windows buildbots.
llvm-svn: 239145
Toma Tabacu [Fri, 5 Jun 2015 11:48:54 +0000 (11:48 +0000)]
[mips] [IAS] Restore STI.FeatureBits in .set pop.
Summary:
Only restoring AvailableFeatures is not enough and will lead to buggy behaviour.
For example, if we have a feature enabled and we ".set pop", the next time we try
to ".set" that feature nothing will happen because the "!(STI.getFeatureBits()[Feature])"
check will be false, because we didn't restore STI.FeatureBits.
In order to fix this, we need to make MipsAssemblerOptions remember the STI.FeatureBits
instead of the AvailableFeatures and then regenerate AvailableFeatures each time we ".set pop".
This is because, AFAIK, there is no way to convert from AvailableFeatures back to STI.FeatureBits,
but the reverse is possible by using ComputeAvailableFeatures(STI.FeatureBits).
I also moved the updating of AssemblerOptions inside the "if" statement in
setFeatureBits() and clearFeatureBits(), as there is no reason to update if
nothing changes.
Reviewers: dsanders, mkuper
Reviewed By: dsanders
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D9156
llvm-svn: 239144
David Majnemer [Fri, 5 Jun 2015 10:52:40 +0000 (10:52 +0000)]
[LoopVectorize] Don't crash on zero-sized types in isInductionPHI
isInductionPHI wants to calculate the stride based on the pointee size.
However, this is not possible when the pointee is zero sized.
This fixes PR23763.
llvm-svn: 239143
Andrea Di Biagio [Fri, 5 Jun 2015 10:29:55 +0000 (10:29 +0000)]
Simplify code; NFC.
Also, moved test cases from CodeGen/X86/fold-buildvector-bug.ll into
CodeGen/X86/buildvec-insertvec.ll and regenerated CHECK lines using
update_llc_test_checks.py.
llvm-svn: 239142
David Majnemer [Fri, 5 Jun 2015 09:57:57 +0000 (09:57 +0000)]
[InstCombine] Rephrase fix to SimplifyWithOpReplaced
I don't have the IR which is causing the build bot breakage but I can
postulate as to why they are timing out:
1. SimplifyWithOpReplaced was stripping flags from the simplified value.
2. visitSelectInstWithICmp was overriding SimplifyWithOpReplaced because
it's simplification wasn't correct.
3. InstCombine would revisit the add instruction and note that it can
rederive the flags.
4. By modifying the value, we chose to revisit instructions which reuse
the value. One of the instructions is the original select, causing
LLVM to never reach fixpoint.
Instead, strip the flags only when we are sure we are going to perform
the simplification.
llvm-svn: 239141
Yaron Keren [Fri, 5 Jun 2015 09:40:53 +0000 (09:40 +0000)]
C++ 11 rangify for loop.
llvm-svn: 239140
Daniel Jasper [Fri, 5 Jun 2015 09:31:20 +0000 (09:31 +0000)]
Revert "[InstCombine] Don't miscompile safe increment idiom"
This is breaking a lot of build bots and is causing very long-running
compiles (infinite loops)?
Likely, we shouldn't return nullptr?
llvm-svn: 239139
Daniel Jasper [Fri, 5 Jun 2015 08:25:37 +0000 (08:25 +0000)]
clang-format: [JS] Let fat arrows have 'Assignment' precedence.
This is a more correct representation than using "Equality" introduced
in r238942 which was a quick fix to solve an actual regression.
According to the typescript spec, arrows behave like "low-precedence"
assignments.
Before:
var a = a.aaaaaaa((a: a) =>
aaaaaaaaaaaaaaaaaaaaa(
bbbbbbbbb) &&
aaaaaaaaaaaaaaaaaaaaa(bbbbbbb));
After:
var a = a.aaaaaaa((a: a) =>
aaaaaaaaaaaaaaaaaaaaa(
bbbbbbbbb) &&
aaaaaaaaaaaaaaaaaaaaa(bbbbbbb));
llvm-svn: 239137
Simon Pilgrim [Fri, 5 Jun 2015 08:24:23 +0000 (08:24 +0000)]
[X86][SSE] Added tests for i8/i16 vector shifts
Currently still scalarized, but D9474 should remedy that.
llvm-svn: 239136
Chaoren Lin [Fri, 5 Jun 2015 06:28:43 +0000 (06:28 +0000)]
Check before using platform specific attributes.
Summary: `os.uname` in TestUniversal and `os.geteuid` in TestTerminal.
Reviewers: clayborg, zturner
Subscribers: lldb-commits
Differential Revision: http://reviews.llvm.org/D10202
llvm-svn: 239135
Yury Gribov [Fri, 5 Jun 2015 06:08:23 +0000 (06:08 +0000)]
[ASan] Add process basename to log name and error message to
simplify analysis of sanitized systems logs.
Differential Revision: http://reviews.llvm.org/D7333
llvm-svn: 239134
Tobias Grosser [Fri, 5 Jun 2015 05:52:15 +0000 (05:52 +0000)]
Avoid the use of std::map emplace
This functionality does not yet seem to exist on all buildbots.
llvm-svn: 239133
Sagar Thakur [Fri, 5 Jun 2015 05:47:54 +0000 (05:47 +0000)]
[lldb-server][MIPS] Read/Write FP registers in FR0 mode
Adding support for read/write FP registers in FR0 mode of mips.
Reviewers: clayborg, tberghammer, jaydeep
Subscribers: emaste, nitesh.jain, bhushan, mohit.bhakkad, lldb-commits
Differential Revision: http://reviews.llvm.org/D10242
llvm-svn: 239132
Tobias Grosser [Fri, 5 Jun 2015 05:31:46 +0000 (05:31 +0000)]
Free two strings produced by isl
With this commit 'make check-polly' is now address sanitizer clean.
llvm-svn: 239131
Oleksiy Vyalov [Fri, 5 Jun 2015 01:32:45 +0000 (01:32 +0000)]
Fix error handling in AdbClient - PushFile and PullFile.
http://reviews.llvm.org/D10176
llvm-svn: 239130
Justin Bogner [Fri, 5 Jun 2015 01:23:42 +0000 (01:23 +0000)]
InstrProf: Fix reading of consecutive 32 bit coverage maps
When we generate coverage data, we explicitly set each coverage map's
alignment to 8 (See InstrProfiling::lowerCoverageData), but when we
read the coverage data, we assume consecutive maps are exactly
adjacent. When we're dealing with 32 bit, maps can end on a 4 byte
boundary, causing us to think the padding is part of the next record.
Fix this by adjusting the buffer to an appropriately aligned address
between records.
This is pretty awkward to test, as it requires a binary with multiple
coverage maps to hit, so we'd need to check in multiple source files
and a binary blob as inputs.
llvm-svn: 239129
Hubert Tong [Fri, 5 Jun 2015 01:10:24 +0000 (01:10 +0000)]
[Concepts] lex keywords: concept and requires
Summary:
This patch enables lexing of `concept` and `requires` as keywords.
Further changes which add messages for future keyword compat are to
follow.
Test Plan:
Testing of C++14 + Concepts TS mode is added to
`test/Lexer/keywords_test.cpp`, which expects that the new keywords are
enabled under said mode.
Reviewers: faisalv, fraggamuffin, rsmith
Reviewed By: rsmith
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D10233
llvm-svn: 239128
Chaoren Lin [Fri, 5 Jun 2015 00:28:06 +0000 (00:28 +0000)]
Delegate path operations to FileSpec.
Summary:
- Added PrependPathComponent utility functions to FileSpec.
- Delegate path operations in ParseCompileUnit to FileSpec.
- Delegate path operations in ParseSupportFiles to FileSpec.
Reviewers: clayborg, vharron, ovyalov
Reviewed By: ovyalov
Subscribers: lldb-commits
Differential Revision: http://reviews.llvm.org/D10253
llvm-svn: 239127
Siva Chandra [Fri, 5 Jun 2015 00:22:49 +0000 (00:22 +0000)]
[TestChangeProcessGroup] Mark the test as xfail for Android API 16
Summary:
This change adds the infrastructure to mark tests as xfail for specific
Android API levels.
Test Plan: dotest.py TestChangeProcessGroup on an Android API 16 device.
Reviewers: chying, labath, chaoren
Reviewed By: labath, chaoren
Subscribers: tberghammer, lldb-commits
Differential Revision: http://reviews.llvm.org/D10261
llvm-svn: 239126
Alexey Samsonov [Thu, 4 Jun 2015 23:58:31 +0000 (23:58 +0000)]
Revert "[Object, ELF] Fix segmentation fault in ELFFile::getSectionName()."
This reverts commit r239124.
llvm-svn: 239125
Alexey Samsonov [Thu, 4 Jun 2015 23:40:23 +0000 (23:40 +0000)]
[Object, ELF] Fix segmentation fault in ELFFile::getSectionName().
Don't do a null dereference if .shstrtab section is missing.
llvm-svn: 239124
Sean Silva [Thu, 4 Jun 2015 23:38:11 +0000 (23:38 +0000)]
Remove unused defaulted argument `IncludeTextualHeaders`.
llvm-svn: 239123
John Thompson [Thu, 4 Jun 2015 23:35:19 +0000 (23:35 +0000)]
Fixed modularize to warn about missing headers referenced in a module map.
llvm-svn: 239122
Jim Grosbach [Thu, 4 Jun 2015 23:35:03 +0000 (23:35 +0000)]
MC: Clean up naming in MCObjectFileInfo.h.
Init*() methods to init*().
llvm-svn: 239121
Jim Grosbach [Thu, 4 Jun 2015 23:35:00 +0000 (23:35 +0000)]
MC: Tidy up formatting and doc comments. NFC.
llvm-svn: 239120
Jim Grosbach [Thu, 4 Jun 2015 23:25:54 +0000 (23:25 +0000)]
MC: Clean up the naming for MCMachObjectWriter. NFC.
s/ExecutePostLayoutBinding/executePostLayoutBinding/
s/ComputeSymbolTable/computeSymbolTable/
s/BindIndirectSymbols/bindIndirectSymbols/
s/RecordTLVPRelocation/recordTLVPRelocation/
s/RecordScatteredRelocation/recordScatteredRelocation/
s/WriteLinkerOptionsLoadCommand/writeLinkerOptionsLoadCommand/
s/WriteLinkeditLoadCommand/writeLinkeditLoadCommand/
s/WriteNlist/writeNlist/
s/WriteDysymtabLoadCommand/writeDysymtabLoadCommand/
s/WriteSymtabLoadCommand/writeSymtabLoadCommand/
s/WriteSection/writeSection/
s/WriteSegmentLoadCommand/writeSegmentLoadCommand/
s/WriteHeader/writeHeader/
llvm-svn: 239119
Jim Grosbach [Thu, 4 Jun 2015 23:25:19 +0000 (23:25 +0000)]
MC: Move MachObjectWriter::SectionAddress to private.
There's already a get() method everything that needs it is using.
llvm-svn: 239118
Jim Grosbach [Thu, 4 Jun 2015 23:25:05 +0000 (23:25 +0000)]
MC: Tidy up formatting and doc comments. NFC.
llvm-svn: 239117
Alexey Samsonov [Thu, 4 Jun 2015 23:14:43 +0000 (23:14 +0000)]
[Object, ELF] Don't assert on invalid magic in createELFObjectFile.
Instead, return a proper error code from factory.
llvm-svn: 239116
David Majnemer [Thu, 4 Jun 2015 23:11:30 +0000 (23:11 +0000)]
[InstCombine] Don't miscompile safe increment idiom
We cleverly handle cases where computation done in one argument of a select
instruction is suitable for the other operand, thus obviating the need
of the select and the comparison. However, the other operand cannot
have flags.
This fixes PR23757.
llvm-svn: 239115
Swaroop Sridhar [Thu, 4 Jun 2015 23:03:21 +0000 (23:03 +0000)]
Statepoint: Fix handling of Far Immediate calls
gc.statepoint intrinsics with a far immediate call target
were lowered incorrectly as pc-rel32 calls.
This change fixes the problem, and generates an indirect call
via a scratch register.
For example:
Intrinsic:
%safepoint_token = call i32 (i64, i32, void ()*, i32, i32, ...) @llvm.experimental.gc.statepoint.p0f_isVoidf(i64 0, i32 0, void ()* inttoptr (i64
140727162896504 to void ()*), i32 0, i32 0, i32 0, i32 0)
Old Incorrect Lowering:
callq
140727162896504
New Correct Lowering:
movabsq $
140727162896504, %rax
callq *%rax
In lowerCallFromStatepoint(), the callee-target was modified and
represented as a "TargetConstant" node, rather than a "Constant" node.
Undoing this modification enabled LowerCall() to generate the
correct CALL instruction.
llvm-svn: 239114
Alexey Samsonov [Thu, 4 Jun 2015 22:58:25 +0000 (22:58 +0000)]
[Object, ELF] Don't call llvm_unreachable() from createELFObjectFile.
Instead, return a proper error code from factory.
llvm-svn: 239113
Hubert Tong [Thu, 4 Jun 2015 22:53:21 +0000 (22:53 +0000)]
Test commit access.
Fixes trailing whitespace in lib/Sema/JumpDiagnostics.cpp.
llvm-svn: 239112
Charles Davis [Thu, 4 Jun 2015 22:50:05 +0000 (22:50 +0000)]
[Target/X86] Don't use callee-saved registers in a Win64 tail call on non-Windows.
Summary:
A small bit that I missed when I updated the X86 backend to account for
the Win64 calling convention on non-Windows. Now we don't use dead
non-volatile registers when emitting a Win64 indirect tail call on
non-Windows.
Should fix PR23710.
Test Plan: Added test for the correct behavior based on the case I posted to PR23710.
Reviewers: rnk
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D10258
llvm-svn: 239111
Alexey Samsonov [Thu, 4 Jun 2015 22:49:55 +0000 (22:49 +0000)]
[Object, MachO] Fixup for r239075: use union to store mach_header and mach_header_64.
llvm-svn: 239110
Alexey Samsonov [Thu, 4 Jun 2015 22:26:44 +0000 (22:26 +0000)]
[Object, MachO] Don't crash on incomplete MachO segment load commands.
Report proper error code from MachOObjectFile constructor if we
can't parse another segment load command (we already return a proper
error if segment load command contents is suspicious).
llvm-svn: 239109
Jim Grosbach [Thu, 4 Jun 2015 22:24:41 +0000 (22:24 +0000)]
MC: Clean up naming in MCObjectWriter. NFC.
s/WriteObject/writeObject/
s/RecordRelocation/recordRelocation/
s/IsSymbolRefDifferenceFullyResolved/isSymbolRefDifferenceFullyResolved/
s/Write8/write8/
s/WriteLE16/writeLE16/
s/WriteLE32/writeLE32/
s/WriteLE64/writeLE64/
s/WriteBE16/writeBE16/
s/WriteBE32/writeBE32/
s/WriteBE64/writeBE64/
s/Write16/write16/
s/Write32/write32/
s/Write64/write64/
s/WriteZeroes/writeZeroes/
s/WriteBytes/writeBytes/
llvm-svn: 239108
Jim Grosbach [Thu, 4 Jun 2015 22:24:29 +0000 (22:24 +0000)]
MC: Tidy up formatting and doc comments. NFC.
llvm-svn: 239107
Alexey Samsonov [Thu, 4 Jun 2015 22:08:37 +0000 (22:08 +0000)]
[Object, MachO] Simplify load segment parsing code. NFC.
llvm-svn: 239106
Benjamin Kramer [Thu, 4 Jun 2015 22:05:51 +0000 (22:05 +0000)]
[SDAG switch lowering] Fix switch case -> or merging for 0 and INT_MIN
The big/small ordering here is based on signed values so SmallValue will
be INT_MIN and BigValue 0. This shouldn't be a problem but the code
assumed that BigValue always had more bits set than SmallValue.
We used to just miss the transformation, but a recent refactoring of
mine turned this into an assertion failure.
llvm-svn: 239105
Siva Chandra [Thu, 4 Jun 2015 22:04:04 +0000 (22:04 +0000)]
[TestGdbRemoteRegisterState] Do not flip ARM *psr register in the test
Summary:
Not all of that register is readable/writable in user mode. This means
that even if the inferior is stopped, parts of the register could be
changing. So, do not flip this register to check if its value can be
restored.
Reviewers: tberghammer, chaoren
Reviewed By: tberghammer
Subscribers: rengolin, aemerson, lldb-commits
Differential Revision: http://reviews.llvm.org/D10204
llvm-svn: 239104
Colin LeMahieu [Thu, 4 Jun 2015 21:49:43 +0000 (21:49 +0000)]
Shouldn't be XFAIL'ed.
llvm-svn: 239103
Colin LeMahieu [Thu, 4 Jun 2015 21:32:42 +0000 (21:32 +0000)]
Revert r239095 incorrect test tree.
llvm-svn: 239102
Artyom Skrobov [Thu, 4 Jun 2015 21:31:41 +0000 (21:31 +0000)]
Simplify ARMTargetParser::parseArch(ARMTargetParser::getCanonical()), following r239099
Reviewers: rengolin
Reviewed By: rengolin
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D10256
llvm-svn: 239101
Jingyue Wu [Thu, 4 Jun 2015 21:28:26 +0000 (21:28 +0000)]
[NVPTX] roll forward r239082
NVPTXISelDAGToDAG translates "addrspacecast to param" to
NVPTX::nvvm_ptr_gen_to_param
Added an llc test in bug21465.
llvm-svn: 239100
Artyom Skrobov [Thu, 4 Jun 2015 21:26:58 +0000 (21:26 +0000)]
Simplify ARMTargetParser::getArchSynonym
Summary:
1) The only caller, ARMTargetParser::parseArch, uses the results for an "endswith" test; so, including the "arm" prefix into the result is unnecessary.
2) Most ARMTargetParser::parseArch callers pass it the output from ARMTargetParser::getCanonicalArchName; so, make this behaviour the default. Then, including the "arm" prefix into the cases is unnecessary.
Reviewers: rengolin
Reviewed By: rengolin
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D10249
llvm-svn: 239099
Chandler Carruth [Thu, 4 Jun 2015 21:23:07 +0000 (21:23 +0000)]
Fix terrible python goof in clang-format.py which broke my vim
integration.
Nothing is more important in life than clang-format integration with
vim. ;]
llvm-svn: 239098
Colin LeMahieu [Thu, 4 Jun 2015 21:22:12 +0000 (21:22 +0000)]
[Hexagon] Removing unused variable.
llvm-svn: 239097
Benjamin Kramer [Thu, 4 Jun 2015 21:17:27 +0000 (21:17 +0000)]
[MachOWriter] Use for-range and roll some duplicated code into loops.
No functionality change intended.
llvm-svn: 239096
Colin LeMahieu [Thu, 4 Jun 2015 21:16:16 +0000 (21:16 +0000)]
[Hexagon] Adding functionality for duplexing. Duplexing is a way to compress commonly used pairs of instructions in order to reduce code size. The test case duplex.ll normally would be 8 bytes, assign register to 0 and jump to link register. After duplexing this is only 4 bytes. This also tests the HexagonMCShuffler code path which is used to make sure duplexed instructions still follow slot requirements.
llvm-svn: 239095
Jingyue Wu [Thu, 4 Jun 2015 21:07:08 +0000 (21:07 +0000)]
Revert r239082
llc crashed for NVPTX backend
llvm-svn: 239094
Ahmed Bougacha [Thu, 4 Jun 2015 20:58:49 +0000 (20:58 +0000)]
[CodeGen] Add testcase for r239002.
Just realized I forgot to add it when committing.
llvm-svn: 239093
Rafael Espindola [Thu, 4 Jun 2015 20:55:49 +0000 (20:55 +0000)]
Remove unused argument. NFC.
llvm-svn: 239092
David Blaikie [Thu, 4 Jun 2015 20:54:32 +0000 (20:54 +0000)]
Re-unique_ptrify LoadedObjectInfo::clone after it was reverted due to some other changes that broke on GCC around the same time
Apparently this functionality isn't used in-tree (or I would go & make
the explicit unique_ptr constructions into implicit constructions to
make them more self documenting now that clone doesn't return a raw
owning pointer anymore) but only by the Julia frontend. This isn't
ideal.
llvm-svn: 239091
Diego Novillo [Thu, 4 Jun 2015 20:49:52 +0000 (20:49 +0000)]
Tidy comment. NFC.
llvm-svn: 239090
Sergey Dmitrouk [Thu, 4 Jun 2015 20:48:40 +0000 (20:48 +0000)]
Erase constant dbgloc on reuse in PHI node
Basic block selection involves checking successor BBs for PHI nodes
that depend on the current BB. In case such BBs are found, the value
being selected is a constant and such constant already exists in
current BB, it's value is reused.
This might lead to wrong locations in some situations, especially if
same constant value ends up being materialized twice in two different
ways, which discards that sharing and leaves us with wrong debug
location in the successor BB.
In code this involves the following sequence of calls:
SelectionDAGBuilder::HandlePHINodesInSuccessorBlocks ->
SelectionDAGBuilder::CopyValueToVirtualRegister ->
SelectionDAGBuilder::getNonRegisterValue
llvm-svn: 239089
David Blaikie [Thu, 4 Jun 2015 20:41:51 +0000 (20:41 +0000)]
Retry defaulting the virtual dtor in LoadedObjectInfo
Originally committed in r237975, GCC 4.7 gave a compilation error
regarding "looser throw specification" though it seemed to be pointing
to a virtual defaulted dtor in a base class and an override defaulted
dtor in a derived class - so I'm not quite sure why/how they could end
up with different throw specifications. To simplify and reduce the risk
of this, I've just removed the pointless override in the derived class,
the base class's should be sufficient. *fingers crossed*
llvm-svn: 239088
Ahmed Bougacha [Thu, 4 Jun 2015 20:39:23 +0000 (20:39 +0000)]
[GlobalMerge] Take into account minsize on Global users' parents.
Now that we can look at users, we can trivially do this: when we would
have otherwise disabled GlobalMerge (currently -O<3), we can just run
it for minsize functions, as it's usually a codesize win.
Differential Revision: http://reviews.llvm.org/D10054
llvm-svn: 239087
Sean Silva [Thu, 4 Jun 2015 20:28:09 +0000 (20:28 +0000)]
[docs] Document "LGTM" in the lexicon.
llvm-svn: 239085
Jim Grosbach [Thu, 4 Jun 2015 20:27:42 +0000 (20:27 +0000)]
MC: Remove obsolete MachO UseAggressiveSymbolFolding.
Fix the FIXME and remove this old as(1) compat option. It was useful for
bringup of the integrated assembler to diff object files, but now it's
just causing more relocations than strictly necessary to be generated.
rdar://
21201804
llvm-svn: 239084
David Blaikie [Thu, 4 Jun 2015 20:23:13 +0000 (20:23 +0000)]
Having another go at some simple cleanup from r237975/r237976
I made a few changes here in a couple of commits - breaking them out
into smaller ones in case I hit the GCC oddities again.
I'm still not /entirely/ sure what the issues were, so apologies if any
of these experiments break things again. Feel free to revert
immediately.
llvm-svn: 239083
Jingyue Wu [Thu, 4 Jun 2015 20:19:38 +0000 (20:19 +0000)]
[NVPTX] kernel pointer arguments point to the global address space
Summary:
With this patch, NVPTXLowerKernelArgs converts a kernel pointer argument to a
pointer in the global address space. This change, along with
NVPTXFavorNonGenericAddrSpaces, allows the NVPTX backend to emit ld.global.*
and st.global.* for accessing kernel pointer arguments.
Minor changes:
1. refactor: extract function convertToPointerInAddrSpace
2. fix a bug in the test case in bug21465.ll
Test Plan: lower-kernel-ptr-arg.ll
Reviewers: eliben, meheff, jholewinski
Reviewed By: jholewinski
Subscribers: wengxt, jholewinski, llvm-commits
Differential Revision: http://reviews.llvm.org/D10154
llvm-svn: 239082
Alexey Samsonov [Thu, 4 Jun 2015 20:08:52 +0000 (20:08 +0000)]
[Object, MachO] Don't crash on invalid MachO segment load commands.
Summary:
Properly report the error in segment load commands from MachOObjectFile
constructor instead of crashing the program.
Adjust the test case accordingly.
Test Plan: regression test suite
Reviewers: rafael, filcab
Subscribers: llvm-commits
llvm-svn: 239081
Alexey Samsonov [Thu, 4 Jun 2015 19:57:46 +0000 (19:57 +0000)]
[Object, MachO] Don't crash on invalid MachO load commands.
Summary:
Currently all load commands are parsed in MachOObjectFile constructor.
If the next load command cannot be parsed, or if command size is too
small, properly report it through the error code and fail to construct
the object, instead of crashing the program.
Test Plan: regression test suite
Reviewers: rafael, filcab
Subscribers: llvm-commits
llvm-svn: 239080
Colin LeMahieu [Thu, 4 Jun 2015 19:49:52 +0000 (19:49 +0000)]
[MC] Allowing operands to be erased from MCInst.
llvm-svn: 239079
Alexey Samsonov [Thu, 4 Jun 2015 19:45:22 +0000 (19:45 +0000)]
[Object, MachO] Don't crash on parsing invalid MachO header.
Summary: Instead, properly report this error from MachOObjectFile constructor.
Test Plan: regression test suite
Reviewers: rafael
Subscribers: llvm-commits
llvm-svn: 239078
Alexey Samsonov [Thu, 4 Jun 2015 19:34:14 +0000 (19:34 +0000)]
[Object, MachO] Remove some code duplication. NFC.
llvm-svn: 239077
Chris Bieneman [Thu, 4 Jun 2015 19:33:23 +0000 (19:33 +0000)]
[CMake] Fixing the check for Ninja
Enabling Ninja Job Pools needs to be dependent on the CMAKE_MAKE_PROGRAM not the CMAKE_GENERATOR. There are generators (like "Sublime Text 2 - Ninja") that also generate ninja build files. Basing of the CMAKE_MAKE_PROGRAM is the best future-proof way to handle this.
llvm-svn: 239076
Alexey Samsonov [Thu, 4 Jun 2015 19:22:03 +0000 (19:22 +0000)]
[Object, MachO] Cache parsed MachO header in MachOObjectFile. NFC.
Summary:
Avoid parsing object file each time MachOObjectFile::getHeader() is
called. Instead, cache the header in MachOObjectFile constructor, where
it's parsed anyway. In future, we must avoid constructing the object
at all if the header can't be parsed.
Test Plan: regression test suite.
Reviewers: rafael
Subscribers: llvm-commits
llvm-svn: 239075
Alexey Samsonov [Thu, 4 Jun 2015 19:22:00 +0000 (19:22 +0000)]
Fix buildbot failure on Windows by relaxing test expectations.
llvm-svn: 239074
Rui Ueyama [Thu, 4 Jun 2015 19:21:24 +0000 (19:21 +0000)]
COFF: Add /failifmismatch option.
llvm-svn: 239073
Rui Ueyama [Thu, 4 Jun 2015 19:21:22 +0000 (19:21 +0000)]
Add comments.
llvm-svn: 239072
Alexei Starovoitov [Thu, 4 Jun 2015 19:15:05 +0000 (19:15 +0000)]
[bpf] add big- and host- endian support
Summary:
-march=bpf -> host endian
-march=bpf_le -> little endian
-match=bpf_be -> big endian
Test Plan:
v1 was tested by IBM s390 guys and appears to be working there.
It bit rots too fast here.
Reviewers: chandlerc, tstellarAMD
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D10177
llvm-svn: 239071
Andrea Di Biagio [Thu, 4 Jun 2015 19:15:01 +0000 (19:15 +0000)]
[DAGCombiner] Fix wrong folding of a build_vector into a blend with zero.
Method 'visitBUILD_VECTOR' in the DAGCombiner knows how to combine a
build_vector of a bunch of extract_vector_elt nodes and constant zero nodes
into a shuffle blend with a zero vector.
However, method 'visitBUILD_VECTOR' forgot that a floating point
build_vector may contain negative zero as well as positive zero.
Example:
define <2 x double> @example(<2 x double> %A) {
entry:
%0 = extractelement <2 x double> %A, i32 0
%1 = insertelement <2 x double> undef, double %0, i32 0
%2 = insertelement <2 x double> %1, double -0.0, i32 1
ret <2 x double> %2
}
Before this patch, llc (with -mattr=+sse4.1) wrongly generated
movq %xmm0, %xmm0 # xmm0 = xmm0[0],zero
So, the sign bit of the negative zero was effectively lost.
This patch fixes the problem by adding explicit checks for positive zero.
With this patch, llc produces the following code for the example above:
movhpd .LCPI0_0(%rip), %xmm0
where .LCPI0_0 referes to a 'double -0'.
llvm-svn: 239070
Evgeniy Stepanov [Thu, 4 Jun 2015 18:54:16 +0000 (18:54 +0000)]
Fix the check for Ninja in the CMake build.
The current check never passes, because CMAKE_MAKE_PROGRAM, at least on Linux,
includes the full path to the "ninja" binary; this effectively disables
compile/link jobs pools.
Use CMAKE_GENERATOR STREQUAL "Ninja" as a more reliable check.
llvm-svn: 239069
Alexey Samsonov [Thu, 4 Jun 2015 18:50:04 +0000 (18:50 +0000)]
Make test case more readable: move CHECK-lines next to corresponding RUN-lines.
llvm-svn: 239068
Rafael Espindola [Thu, 4 Jun 2015 18:49:12 +0000 (18:49 +0000)]
Fix the test to use the correct size.
For some reason llvm's r239045 made lld propagate data_1's size. This indicates
a bug somewhere in lld.
I hesitated between changing the test or just checking in a .o produced with
the old llvm-mc. Since the size is now correct, it seemed better to update the
test.
llvm-svn: 239067
Bill Seurer [Thu, 4 Jun 2015 18:45:44 +0000 (18:45 +0000)]
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document.
I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h.
https://llvm.org/bugs/show_bug.cgi?id=23679
http://reviews.llvm.org/D10131
llvm-svn: 239066
Alexey Samsonov [Thu, 4 Jun 2015 18:34:11 +0000 (18:34 +0000)]
llvm-objdump: return non-zero exit code for certain cases of invalid input
* If the input file is missing;
* If the type of input object file can't be recognized;
* If the object file can't be parsed correctly.
llvm-svn: 239065
Yaron Keren [Thu, 4 Jun 2015 18:33:04 +0000 (18:33 +0000)]
C++11 rangify for loop.
llvm-svn: 239064
Evgeniy Stepanov [Thu, 4 Jun 2015 18:28:35 +0000 (18:28 +0000)]
[tsan] Disable a flaky test.
llvm-svn: 239063