Rui Ueyama [Sun, 20 Nov 2016 02:43:44 +0000 (02:43 +0000)]
Use auto for obvious types.
llvm-svn: 287481
Rui Ueyama [Sun, 20 Nov 2016 02:39:59 +0000 (02:39 +0000)]
Do not expose ICF class from the file.
Also this patch uses file-scope functions instead of class member function.
Now that ICF class is not visible from outside, InputSection class
can no longer be "friend" of it. So I removed the friend relation
and just make it expose the features to public.
llvm-svn: 287480
Saleem Abdulrasool [Sun, 20 Nov 2016 02:36:38 +0000 (02:36 +0000)]
ExceptionDemo: remove some undefined behaviour
The casting based reading of the LSDA could attempt to read unsuitably aligned
data. Avoid that case by explicitly using a memcpy. A similar approach is used
in libc++abi to address the same UB.
llvm-svn: 287479
Saleem Abdulrasool [Sun, 20 Nov 2016 02:36:36 +0000 (02:36 +0000)]
ExceptionDemo: prefer headers over redeclarations
Rather than redeclaring the interfaces for exceptions, prefer using the
`unwind.h` header. This is vended by at least gcc and clang, and can also be
found by an external unwinding library (e.g. libunwind). Doing this simplifies
the example to the exception handling itself. Minor tweaks are the result of
_Unwind_Context_t not being defined, which is just a typedef for struct
_Unwind_Context *. NFC.
llvm-svn: 287478
Alexei Starovoitov [Sun, 20 Nov 2016 02:25:00 +0000 (02:25 +0000)]
[bpf] add BPF disassembler
add BPF disassembler, so tools like llvm-objdump can be used:
$ llvm-objdump -d -no-show-raw-insn ./sockex1_kern.o
./sockex1_kern.o: file format ELF64-BPF
Disassembly of section socket1:
bpf_prog1:
0: r6 = r1
8: r0 = *(u8 *)skb[23]
10: *(u32 *)(r10 - 4) = r0
18: r1 = *(u32 *)(r6 + 4)
20: if r1 != 4 goto 8
28: r2 = r10
30: r2 += -4
ld_imm64 (the only 16-byte insn) and special ld_abs/ld_ind instructions
had to be treated in a special way. The decoders for the rest of the insns
are automatically generated.
Add tests to cover new functionality.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
llvm-svn: 287477
Rui Ueyama [Sun, 20 Nov 2016 01:41:28 +0000 (01:41 +0000)]
Attempt to fix big-endian buildbots.
llvm-svn: 287476
Rui Ueyama [Sun, 20 Nov 2016 01:15:56 +0000 (01:15 +0000)]
Style fix. NFC.
llvm-svn: 287475
Rui Ueyama [Sun, 20 Nov 2016 01:13:22 +0000 (01:13 +0000)]
Fix buildbot.
llvm-svn: 287474
Rui Ueyama [Sun, 20 Nov 2016 01:03:22 +0000 (01:03 +0000)]
SHA1: unroll loop in hashBlock.
This code is taken from public domain.
https://github.com/jsonn/src/blob/trunk/common/lib/libc/hash/sha1/sha1.c
I wrote a sha1 command and ran it on my Xeon E5-2680 v2 2.80GHz machine.
Here is a result. The new hash function is 37% faster than before.
Performance counter stats for './llvm-sha1-old /ssd/build/bin/lld' (10 runs):
6640.503687 task-clock (msec) # 1.001 CPUs utilized ( +- 0.03% )
54 context-switches # 0.008 K/sec ( +- 5.03% )
5 cpu-migrations # 0.001 K/sec ( +- 31.73% )
183,803 page-faults # 0.028 M/sec ( +- 0.00% )
18,527,954,113 cycles # 2.790 GHz ( +- 0.03% )
4,993,237,485 stalled-cycles-frontend # 26.95% frontend cycles idle ( +- 0.11% )
<not supported> stalled-cycles-backend
50,217,149,423 instructions # 2.71 insns per cycle
# 0.10 stalled cycles per insn ( +- 0.00% )
6,094,322,337 branches # 917.750 M/sec ( +- 0.00% )
11,778,239 branch-misses # 0.19% of all branches ( +- 0.01% )
6.
634017401 seconds time elapsed ( +- 0.03% )
Performance counter stats for './llvm-sha1-new /ssd/build/bin/lld' (10 runs):
4167.062720 task-clock (msec) # 1.001 CPUs utilized ( +- 0.02% )
52 context-switches # 0.012 K/sec ( +- 16.45% )
7 cpu-migrations # 0.002 K/sec ( +- 32.20% )
183,804 page-faults # 0.044 M/sec ( +- 0.00% )
11,626,611,958 cycles # 2.790 GHz ( +- 0.02% )
4,491,897,976 stalled-cycles-frontend # 38.63% frontend cycles idle ( +- 0.05% )
<not supported> stalled-cycles-backend
24,320,180,617 instructions # 2.09 insns per cycle
# 0.18 stalled cycles per insn ( +- 0.00% )
1,574,674,576 branches # 377.886 M/sec ( +- 0.00% )
11,769,693 branch-misses # 0.75% of all branches ( +- 0.00% )
4.
163251552 seconds time elapsed ( +- 0.02% )
Differential Revision: https://reviews.llvm.org/D26890
llvm-svn: 287473
Saleem Abdulrasool [Sun, 20 Nov 2016 00:20:27 +0000 (00:20 +0000)]
Demangle: remove references to allocator for default allocator
The demangler had stopped using a custom allocator but had not been updated to
remove the use of the explicit allocator passing. This removes that as we do
not need to do anything special here anymore. This just makes the code more
compact. NFC.
llvm-svn: 287472
Saleem Abdulrasool [Sun, 20 Nov 2016 00:20:25 +0000 (00:20 +0000)]
Demangle: remove unnecessary typedef for std::vector
We could create a local typedef for std::vector called Vector. Inline the use
of std::vector rather than use the typedef. NFC.
llvm-svn: 287471
Saleem Abdulrasool [Sun, 20 Nov 2016 00:20:23 +0000 (00:20 +0000)]
Demangle: replace custom typedef for std::string with std::string
We created a local typedef for `std::basic_string<char, std::char_traits<char>>`
which is just `std::string`. Remove the local typedef and propagate the type
information through the rest of the demangler. NFC.
llvm-svn: 287470
Saleem Abdulrasool [Sun, 20 Nov 2016 00:20:20 +0000 (00:20 +0000)]
Demangle: use direct member initialization (NFC)
Prefer direct member initialization over the explicit out-of-line initialization
for the construction of the local type. NFC.
llvm-svn: 287469
Rui Ueyama [Sat, 19 Nov 2016 23:26:41 +0000 (23:26 +0000)]
Remove a file that is too short to be an independent file.
We have a .cpp and a .h for parseDynamicList(). This patch
moves the function to DriverUtil.cpp.
llvm-svn: 287468
Rui Ueyama [Sat, 19 Nov 2016 23:18:43 +0000 (23:18 +0000)]
Remove unused #include.
llvm-svn: 287467
Rui Ueyama [Sat, 19 Nov 2016 23:14:23 +0000 (23:14 +0000)]
Refactor ICF.
In order to use forEachGroup in the final loop in ICF::run,
I changed some function parameter types.
llvm-svn: 287466
Martin Storsjo [Sat, 19 Nov 2016 21:22:38 +0000 (21:22 +0000)]
builtins: Allow building windows arm functions for mingw
When building with clang/LLVM in MSVC mode, the msvcrt libraries contain
these functions.
When building in a mingw environment, we need to provide them somehow,
e.g. via compiler-rt.
The aeabi divmod functions work in the same way as the corresponding
__rt_*div* functions for windows, but their parameters are swapped.
The functions for converting float to integer and vice versa are the
same as their aeabi equivalents, only with different function names.
Differential Revision: https://reviews.llvm.org/D26183
llvm-svn: 287465
Renato Golin [Sat, 19 Nov 2016 21:19:19 +0000 (21:19 +0000)]
[ARM|ASAN] Disabling more unstable ASAN tests
We're seeying these errors with GCC and Clang on different systems, while
some other identical OSs on different boards fail. Like many other ASAN
tests, there seem to be no easy way to investigate this other than someone
familiar with the sanitizer code and the ARM libraries.
At least, for now, we'll silence the bots. I'll create a bugzilla entry.
llvm-svn: 287464
Chris Bieneman [Sat, 19 Nov 2016 21:14:59 +0000 (21:14 +0000)]
[CMake] Fixing clang standalone build
I broke this in r287406 and r287407.
llvm-svn: 287463
Benjamin Kramer [Sat, 19 Nov 2016 20:44:26 +0000 (20:44 +0000)]
Give some helper classes/functions internal linkage. NFC.
llvm-svn: 287462
Simon Pilgrim [Sat, 19 Nov 2016 20:41:48 +0000 (20:41 +0000)]
[X86][SSE] Improve PSHUFB lowering from either input
Canonicalization may leave the zeroable vector in the first input.
llvm-svn: 287461
Rui Ueyama [Sat, 19 Nov 2016 20:15:55 +0000 (20:15 +0000)]
Use std::equal instead of hand-written loops.
llvm-svn: 287460
Simon Pilgrim [Sat, 19 Nov 2016 20:12:34 +0000 (20:12 +0000)]
[X86][AVX512] Add VPERMV/VPERMV3 v64i8 byte shuffles on avx512vbmi targets
llvm-svn: 287459
John McCall [Sat, 19 Nov 2016 20:12:25 +0000 (20:12 +0000)]
Remove alignment from ctors/dtors lists in an attempt to placate LTO.
llvm-svn: 287458
Rui Ueyama [Sat, 19 Nov 2016 19:26:52 +0000 (19:26 +0000)]
Restore a comment that was accidentally changed.
llvm-svn: 287457
Rui Ueyama [Sat, 19 Nov 2016 19:23:58 +0000 (19:23 +0000)]
Use Optional<std::string> instead of "" to represent a failure.
llvm-svn: 287456
Rui Ueyama [Sat, 19 Nov 2016 19:23:56 +0000 (19:23 +0000)]
Make buildSysrootedPath file-scoped.
This patch also changes its return type to simplify callers.
llvm-svn: 287455
Rui Ueyama [Sat, 19 Nov 2016 18:49:38 +0000 (18:49 +0000)]
Simplify "missing argument" error message.
We do not have an option taking more than one arguments,
so we can just say "missing argument" instead of "missing argument(s)".
llvm-svn: 287454
Mehdi Amini [Sat, 19 Nov 2016 18:44:16 +0000 (18:44 +0000)]
[ThinLTO] Fix crash when importing an opaque type
It seems that because ThinLTO does not import the full module,
some invariant of the type mapper are broken.
In Monolithic LTO, we import every globals: when calling
IRLinker::copyFunctionProto() on @foo(), we end-up calling
TypeMapTy::get(FTy) on the type of @foo(), which will map
%0 and record the destination as opaque.
ThinLTO skips this because @foo is not imported and goes directly
to the next stage.
Next we call computeTypeMapping() that map the types for each
globals, and ends up checking for type isomorphism, and may add
type mapping. However it doesn't record if there was an opaque
destination type that was resolved.
Instead of lazily "discovering" opaque type in the destination
module on the go, we change the TypeFinder to eagerly record all
types and not only the named ones.
Differential Revision: https://reviews.llvm.org/D26840
llvm-svn: 287453
Rui Ueyama [Sat, 19 Nov 2016 18:44:09 +0000 (18:44 +0000)]
Split getFdeEncoding.
llvm-svn: 287452
Rui Ueyama [Sat, 19 Nov 2016 18:34:55 +0000 (18:34 +0000)]
Fix typo in error message.
llvm-svn: 287451
Mehdi Amini [Sat, 19 Nov 2016 18:20:05 +0000 (18:20 +0000)]
[ThinLTO] Implement -pass-remarks-output in ThinLTOCodeGenerator
Summary:
This will also be added to the LTO API, right now this will
bring ThinLTO on par with Monolithic LTO on Darwin.
Reviewers: anemet
Subscribers: tejohnson, llvm-commits
Differential Revision: https://reviews.llvm.org/D26886
llvm-svn: 287450
Mehdi Amini [Sat, 19 Nov 2016 18:19:41 +0000 (18:19 +0000)]
Change setDiagnosticsOutputFile to take a unique_ptr from a raw pointer (NFC)
Summary:
This makes it explicit that ownership is taken. Also replace all `new`
with make_unique<> at call sites.
Reviewers: anemet
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D26884
llvm-svn: 287449
George Rimar [Sat, 19 Nov 2016 18:14:24 +0000 (18:14 +0000)]
[ELF] - Exit on --version call.
GNU linkers disagree here.
Though both -version and -v are mentioned
in help to print the version information, GNU ld just normally exits,
while gold can continue linking. We are compatible with ld.bfd here.
This fixes PR31057.
Differential revision: https://reviews.llvm.org/D26865
llvm-svn: 287448
Simon Pilgrim [Sat, 19 Nov 2016 18:12:48 +0000 (18:12 +0000)]
[X86][AVX512] Add avx512vbmi tests
llvm-svn: 287447
Rui Ueyama [Sat, 19 Nov 2016 18:05:58 +0000 (18:05 +0000)]
Simplify. NFC.
llvm-svn: 287446
Rui Ueyama [Sat, 19 Nov 2016 18:05:56 +0000 (18:05 +0000)]
Change filler type from ArrayRef<uint8_t> to uint32_t.
Filler expressions in linker script "=fillexp" are always handled
as 32-bit integers. Thus the new type is more natural.
llvm-svn: 287445
Simon Pilgrim [Sat, 19 Nov 2016 17:50:14 +0000 (17:50 +0000)]
[X86][AVX512] Added some more complex v64i8 shuffles
llvm-svn: 287444
Craig Topper [Sat, 19 Nov 2016 17:33:17 +0000 (17:33 +0000)]
[X86] Simplify some code a little by removing a dulicate variable and combinining two if statements. NFCI
llvm-svn: 287443
Daniel Sanders [Sat, 19 Nov 2016 14:47:41 +0000 (14:47 +0000)]
Try again to fix unused variable warning on lld-x86_64-darwin13 after r287439.
The previous attempt didn't work. I assume LLVM_ATTRIBUTE_UNUSED isn't
available on that machine.
llvm-svn: 287442
Daniel Sanders [Sat, 19 Nov 2016 13:50:32 +0000 (13:50 +0000)]
Try to fix unused variable warning on lld-x86_64-darwin13 after r287439.
Whether the variable is used or not depends on NDEBUG.
llvm-svn: 287440
Daniel Sanders [Sat, 19 Nov 2016 13:05:44 +0000 (13:05 +0000)]
Check that emitted instructions meet their predicates on all targets except ARM, Mips, and X86.
Summary:
* ARM is omitted from this patch because this check appears to expose bugs in this target.
* Mips is omitted from this patch because this check either detects bugs or deliberate
emission of instructions that don't satisfy their predicates. One deliberate
use is the SYNC instruction where the version with an operand is correctly
defined as requiring MIPS32 while the version without an operand is defined
as an alias of 'SYNC 0' and requires MIPS2.
* X86 is omitted from this patch because it doesn't use the tablegen-erated
MCCodeEmitter infrastructure.
Patches for ARM and Mips will follow.
Depends on D25617
Reviewers: tstellarAMD, jmolloy
Subscribers: wdng, jmolloy, aemerson, rengolin, arsenm, jyknight, nemanjai, nhaehnle, tstellarAMD, llvm-commits
Differential Revision: https://reviews.llvm.org/D25618
llvm-svn: 287439
Daniel Sanders [Sat, 19 Nov 2016 12:21:34 +0000 (12:21 +0000)]
[tablegen] Merge duplicate definitions of getMinimalTypeForRange. NFC.
Summary: Depends on D25614
Reviewers: qcolombet
Subscribers: qcolombet, beanz, mgorny, llvm-commits
Differential Revision: https://reviews.llvm.org/D25617
llvm-svn: 287438
John McCall [Sat, 19 Nov 2016 08:17:24 +0000 (08:17 +0000)]
Introduce a helper class for building complex constant initializers. NFC.
I've adopted this in most of the places it makes sense, but v-tables
and CGObjCMac will need a second pass.
llvm-svn: 287437
Ekaterina Romanova [Sat, 19 Nov 2016 04:59:08 +0000 (04:59 +0000)]
Doxygen comments for avxintrin.h.
Added doxygen comments to avxintrin.h's intrinsics. As of now, all the intrinsics in this file that were documented by Sony's intrinsics guide should have corresponding doxygen comments.
Note: The doxygen comments are automatically generated based on Sony's intrinsic
s document.
I got an OK from Eric Christopher to commit doxygen comments without prior code
review upstream.
Reviewed by Wolfgang Pieb.
llvm-svn: 287436
Eric Fiselier [Sat, 19 Nov 2016 03:29:03 +0000 (03:29 +0000)]
Fix stdint/cstdint modules
llvm-svn: 287435
Chris Bieneman [Sat, 19 Nov 2016 03:19:58 +0000 (03:19 +0000)]
[CMake] llvm-lto2 depends on intrinsics_gen
llvm-lto2.cpp has the following include chain:
llvm/LTO/Caching.h
llvm/LTO/LTO.h
llvm/CodeGen/Analysis.h
llvm/IR/CallSite.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means llvm-lto2 needs to depend on intrinsics_gen.
llvm-svn: 287434
Chris Bieneman [Sat, 19 Nov 2016 03:18:50 +0000 (03:18 +0000)]
[CMake] opt depends on intrinsics_gen
AnalysisWrappers.cpp has the following include chain:
llvm/Analysis/CallGraph.h
llvm/IR/CallSite.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means opt needs to depend on intrinsics_gen.
llvm-svn: 287433
Chris Bieneman [Sat, 19 Nov 2016 03:16:33 +0000 (03:16 +0000)]
[CMake] llvm-nm depends on intrinsics_gen
llvm-nm.cpp has the following include chain:
llvm/IR/Function.h
llvm/IR/Argument.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means llvm-nm needs to depend on intrinsics_gen.
llvm-svn: 287432
Chris Bieneman [Sat, 19 Nov 2016 02:36:28 +0000 (02:36 +0000)]
[CMake] llvm-link depends on intrinsics_gen
llvm-link.cpp has the following include chain:
llvm/Bitcode/BitcodeWriter.h
llvm/IR/ModuleSummaryIndex.h
llvm/IR/Module.h
llvm/IR/Function.h
llvm/IR/Argument.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means llvm-link needs to depend on intrinsics_gen.
llvm-svn: 287431
Chris Bieneman [Sat, 19 Nov 2016 02:33:57 +0000 (02:33 +0000)]
[CMake] llvm-extract depends on intrinsics_gen
llvm-extract.cpp has the following include chain:
llvm/Bitcode/BitcodeWriterPass.h
llvm/IR/PassManager.h
llvm/IR/Function.h
llvm/IR/Argument.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means llvm-extract needs to depend on intrinsics_gen.
llvm-svn: 287430
Chris Bieneman [Sat, 19 Nov 2016 02:33:42 +0000 (02:33 +0000)]
[CMake] llvm-dwp depends on intrinsics_gen
llvm-dwp.cpp has the following include chain:
llvm/CodeGen/AsmPrinter.h
llvm/CodeGen/MachineFunctionPass.h
llvm/CodeGen/MachineFunction.h
llvm/CodeGen/MachineBasicBlock.h
llvm/CodeGen/MachineInstr.h
llvm/Analysis/AliasAnalysis.h
llvm/IR/CallSite.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means llvm-dwp needs to depend on intrinsics_gen.
llvm-svn: 287429
Chris Bieneman [Sat, 19 Nov 2016 02:31:14 +0000 (02:31 +0000)]
[CMake] llvm-dis depends on intrinsics_gen
llvm-dis.cpp has the following include chain:
llvm/Bitcode/BitcodeReader.h
llvm/IR/ModuleSummaryIndex.h
llvm/IR/Module.h
llvm/IR/Function.h
llvm/IR/Argument.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means llvm-dis needs to depend on intrinsics_gen.
llvm-svn: 287428
Chris Bieneman [Sat, 19 Nov 2016 02:28:18 +0000 (02:28 +0000)]
[CMake] llvm-diff depends on intrinsics_gen
llvm-diff.cpp has the following include chain:
llvm/IR/Module.h
llvm/IR/Function.h
llvm/IR/Argument.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means llvm-diff needs to depend on intrinsics_gen.
llvm-svn: 287427
Chris Bieneman [Sat, 19 Nov 2016 02:25:54 +0000 (02:25 +0000)]
[CMake] llvm-stress depends on intrinsics_gen
llvm-stress.cpp has the following include chain:
llvm/Analysis/CallGraphSCCPass.h
llvm/Analysis/CallGraph.h
llvm/IR/CallSite.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means llvm-stress needs to depend on intrinsics_gen.
llvm-svn: 287426
Chris Bieneman [Sat, 19 Nov 2016 02:20:59 +0000 (02:20 +0000)]
[CMake] bugpoint-passes depends on intrinsics_gen
TestPasses.cpp has the following include chain:
llvm/IR/InstVisitor.h
llvm/IR/CallSite.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means bugpoint-passes needs to depend on intrinsics_gen.
llvm-svn: 287425
Chris Bieneman [Sat, 19 Nov 2016 02:17:12 +0000 (02:17 +0000)]
[CMake] llvm-bcanalyzer depends on intrinsics_gen
llvm-bcanalyzer.cpp has the following include chain:
llvm/Bitcode/BitcodeReader.h
llvm/IR/ModuleSummaryIndex.h
llvm/IR/Module.h
llvm/IR/Function.h
llvm/IR/Argument.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means llvm-bcanalyzer needs to depend on intrinsics_gen.
llvm-svn: 287424
Chris Bieneman [Sat, 19 Nov 2016 02:15:04 +0000 (02:15 +0000)]
[CMake] llvm-as depends on intrinsics_gen
llvm-as.cpp has the following include chain:
llvm/Bitcode/BitcodeWriter.h
llvm/IR/ModuleSummaryIndex.h
llvm/IR/Module.h
llvm/IR/Function.h
llvm/IR/Argument.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means llvm-as needs to depend on intrinsics_gen.
llvm-svn: 287423
Chris Bieneman [Sat, 19 Nov 2016 02:12:03 +0000 (02:12 +0000)]
[CMake] llc depends on intrinsics_gen
llc.cpp has the following include chain:
llvm/Analysis/TargetLibraryInfo.h
llvm/IR/Function.h
llvm/IR/Argument.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means llc needs to depend on intrinsics_gen.
llvm-svn: 287422
Chris Bieneman [Sat, 19 Nov 2016 02:09:51 +0000 (02:09 +0000)]
[CMake] lli-child-target depends on intrinsics gen
Messed up in r287420, it isn't just lli, but also but lli-child-target that need to depend on intrinsics_gen.
llvm-svn: 287421
Chris Bieneman [Sat, 19 Nov 2016 02:05:19 +0000 (02:05 +0000)]
[CMake] lli depends on intrinsics_gen
ChildTarget.cpp has the following include chain:
llvm/ExecutionEngine/Orc/OrcABISupport.h
llvm/ExecutionEngine/Orc/IndirectionUtils.h
llvm/IR/IRBuilder.h
llvm/IR/ConstantFolder.h
llvm/IR/InstrTypes.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means lli needs to depend on intrinsics_gen.
llvm-svn: 287420
Chris Bieneman [Sat, 19 Nov 2016 02:02:46 +0000 (02:02 +0000)]
[CMake] llvm-dsymutil depends on intrinsics_gen
DwarfLinker.cpp has the following include chain:
llvm/CodeGen/AsmPrinter.h
llvm/CodeGen/MachineFunctionPass.h
llvm/CodeGen/MachineFunction.h
llvm/CodeGen/MachineBasicBlock.h
llvm/CodeGen/MachineInstr.h
llvm/Analysis/AliasAnalysis.h
llvm/IR/CallSite.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means llvm-dsymutil needs to depend on intrinsics_gen.
llvm-svn: 287419
Eric Fiselier [Sat, 19 Nov 2016 01:38:00 +0000 (01:38 +0000)]
Mark test as unsupported in C++03
llvm-svn: 287417
Dylan McKay [Sat, 19 Nov 2016 01:33:42 +0000 (01:33 +0000)]
[AVR] Remove a bunch of unused variables
llvm-svn: 287416
Chris Bieneman [Sat, 19 Nov 2016 01:32:09 +0000 (01:32 +0000)]
[CMake] Apply sandbox profile to target not directory
When LLVM_DEPENDENCY_DEBUGGING=On we should apply the sandbox only on the target, not the directory. This is important for directories that create more than one target, or for nested directories.
llvm-svn: 287415
Dylan McKay [Sat, 19 Nov 2016 01:14:44 +0000 (01:14 +0000)]
[AVR] Remove a variable which was unused in release mode
In release mode where assertions are not enabled, this caused an 'unused
variable' warning.
llvm-svn: 287414
Eric Fiselier [Sat, 19 Nov 2016 01:14:15 +0000 (01:14 +0000)]
[libcxx] Implement locale.h to fix modules build
Summary:
Because `locale.h` isn't part of the libc++ modules the class definitions it provides are exported as part of `__locale` (since it happens to be build first). This breaks `<clocale>` which exports `std::lconv` without including `<__locale>`.
This patch implements `locale.h` to fix this issue, it also adds support for testing libc++ with modules.
Reviewers: mclow.lists, rsmith, EricWF
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D26826
llvm-svn: 287413
Zachary Turner [Sat, 19 Nov 2016 00:50:29 +0000 (00:50 +0000)]
Fix some accidental Prints of StringRefs that snuck in.
llvm-svn: 287412
Peter Collingbourne [Sat, 19 Nov 2016 00:30:56 +0000 (00:30 +0000)]
Sema: As of MSVC 2015, a user-declared move operation causes the deletion of both copy operations.
Differential Revision: https://reviews.llvm.org/D26868
llvm-svn: 287411
Akira Hatanaka [Sat, 19 Nov 2016 00:13:03 +0000 (00:13 +0000)]
[Sema] Don't allow applying address-of operator to a call to a function
with __unknown_anytype return type.
When the following code is compiled, Sema infers that the type of
__unknown_anytype is double:
extern __unknown_anytype func();
double *d = (double*)&func();
This triggers an assert in CodeGenFunction::EmitCallExprLValue because
it doesn't expect to see a call to a function with a non-reference
scalar return type.
This commit prevents the assert by making VisitUnaryAddrOf error out if
the address-of operator is applied to a call to a function with
__unknown_anytype return type.
rdar://problem/
20287610
Differential revision: https://reviews.llvm.org/D26808
llvm-svn: 287410
Zachary Turner [Fri, 18 Nov 2016 23:32:37 +0000 (23:32 +0000)]
Fix some build errors.
llvm-svn: 287409
Chris Bieneman [Fri, 18 Nov 2016 23:31:53 +0000 (23:31 +0000)]
[CMake] NFC. Updating CMake dependency specifications
This patch updates a bunch of places where add_dependencies was being explicitly called to add dependencies on intrinsics_gen to instead use the DEPENDS named parameter. This cleanup is needed for a patch I'm working on to add a dependency debugging mode to the build system.
llvm-svn: 287408
Chris Bieneman [Fri, 18 Nov 2016 23:31:16 +0000 (23:31 +0000)]
[CMake] clang depends on intrinsics_gen
cc1_main.cpp has the following include chain:
llvm/LinkAllPasses.h
llvm/Analysis/AliasSetTracker.h
llvm/Analysis/AliasAnalysis.h
llvm/IR/CallSite.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means clang needs to depend on intrinsics_gen.
llvm-svn: 287407
Chris Bieneman [Fri, 18 Nov 2016 23:31:14 +0000 (23:31 +0000)]
[CMake] clang-offload-bundler depends on intrinsics_gen
ClangOffloadBundler.cpp has the following include chain:
llvm/Bitcode/BitcodeWriter.h
llvm/IR/ModuleSummaryIndex.h
llvm/IR/Module.h
llvm/IR/Function.h
llvm/IR/Argument.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means clang-offload-bundler needs to depend on intrinsics_gen.
llvm-svn: 287406
Chris Bieneman [Fri, 18 Nov 2016 23:30:58 +0000 (23:30 +0000)]
[CMake] verify-uselistorder depends on intrinsics_gen
verify-uselistorder.cpp has the following include chain:
llvm/Bitcode/BitcodeReader.h
llvm/IR/ModuleSummaryIndex.h
llvm/IR/Module.h
llvm/IR/Function.h
llvm/IR/Argument.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means verify-uselistorder needs to depend on intrinsics_gen.
llvm-svn: 287405
Chris Bieneman [Fri, 18 Nov 2016 23:30:39 +0000 (23:30 +0000)]
[CMake] sanstats depends on intrinsics_gen
sanstats.cpp has the following include chain:
llvm/Transforms/Utils/SanitizerStats.h
llvm/IR/IRBuilder.h
llvm/IR/ConstantFolder.h
llvm/IR/InstrTypes.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means sanstats needs to depend on intrinsics_gen.
llvm-svn: 287404
Kuba Mracek [Fri, 18 Nov 2016 23:25:57 +0000 (23:25 +0000)]
[lit] When setting SDKROOT on Darwin, use '--sdk macosx' to find the right SDK path.
This will make sure that we find an actual path in case you have Command Line Tools installed.
llvm-svn: 287403
Chris Bieneman [Fri, 18 Nov 2016 23:25:30 +0000 (23:25 +0000)]
[CMake] bugpoint depends on intrinsics_gen
CrashDebugger.cpp has the following include chain:
llvm/Analysis/TargetTransformInfo.h
llvm/IR/IntrinsicInst.h
llvm/IR/Function.h
llvm/IR/Argument.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means bugpoint needs to depend on intrinsics_gen.
llvm-svn: 287402
Zachary Turner [Fri, 18 Nov 2016 23:22:42 +0000 (23:22 +0000)]
Convert CommandHistory functions to StringRef.
llvm-svn: 287401
Sanjay Patel [Fri, 18 Nov 2016 23:22:00 +0000 (23:22 +0000)]
[InstCombine] add tests to show likely unwanted select widening; NFC
This is a prerequisite patch for D26556:
https://reviews.llvm.org/D26556
...because there was no direct coverage for these folds (which in some cases are adding instructions).
llvm-svn: 287400
Chris Bieneman [Fri, 18 Nov 2016 23:20:38 +0000 (23:20 +0000)]
[CMake] llvm-split depends on intrinsics_gen
llvm-split.cpp has the following include chain:
llvm/Bitcode/BitcodeWriter.h
llvm/IR/ModuleSummaryIndex.h
llvm/IR/Module.h
llvm/IR/Function.h
llvm/IR/Argument.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means llvm-split needs to depend on intrinsics_gen.
llvm-svn: 287399
Chris Bieneman [Fri, 18 Nov 2016 23:20:35 +0000 (23:20 +0000)]
[CMake] llvm-lto depends on intrinsics_gen
llvm-lto.cpp has the following include chain:
llvm/Bitcode/BitcodeReader.h
llvm/IR/ModuleSummaryIndex.h
llvm/IR/Module.h
llvm/IR/Function.h
llvm/IR/Argument.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means llvm-lto needs to depend on intrinsics_gen.
llvm-svn: 287398
Enrico Granata [Fri, 18 Nov 2016 23:18:11 +0000 (23:18 +0000)]
Removing myself from CODE_OWNERS, and distributing those duties among other members of the community
That's All, Folks
llvm-svn: 287397
Chris Bieneman [Fri, 18 Nov 2016 23:04:27 +0000 (23:04 +0000)]
[CMake] llvm-ar depends on intrinsics_gen
llvm-ar.cpp has the following include chain:
llvm/IR/Module.h
llvm/IR/Function.h
llvm/IR/Argument.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means llvm-ar needs to depend on intrinsics_gen.
llvm-svn: 287395
Chris Bieneman [Fri, 18 Nov 2016 23:04:15 +0000 (23:04 +0000)]
[CMake] llvm-profdata depends on intrinsics_gen
llvm-profdata.cpp has the following include chain:
llvm/ProfileData/SampleProfReader.h
llvm/IR/Function.h
llvm/IR/Argument.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means llvm-profdata needs to depend on intrinsics_gen.
llvm-svn: 287394
Chris Bieneman [Fri, 18 Nov 2016 23:03:51 +0000 (23:03 +0000)]
[CMake] LTO depends on intrinsics_gen
lto.cpp has the following include chain:
llvm/Bitcode/BitcodeReader.h
llvm/IR/ModuleSummaryIndex.h
llvm/IR/Module.h
llvm/IR/Function.h
llvm/IR/Argument.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen
This means LTO needs to depend on intrinsics_gen.
llvm-svn: 287393
Stephan T. Lavavej [Fri, 18 Nov 2016 22:45:32 +0000 (22:45 +0000)]
[libcxx] [test] D26812: In random tests, use real static_asserts and silence a warning.
In C++11 mode and newer, use real static_asserts.
In C++03 mode, min() and max() aren't constexpr, so use plain asserts.
One test triggers MSVC's warning C4310 "cast truncates constant value".
The code is valid, and yet the warning is valid, so I'm silencing it
through push-disable-pop.
llvm-svn: 287391
Konstantin Zhuravlyov [Fri, 18 Nov 2016 22:31:51 +0000 (22:31 +0000)]
[AMDGPU] Change frexp.exp builtin to return i16 for f16 input
Differential Revision: https://reviews.llvm.org/D26863
llvm-svn: 287390
Konstantin Zhuravlyov [Fri, 18 Nov 2016 22:31:08 +0000 (22:31 +0000)]
[AMDGPU] Change frexp.exp intrinsic to return i16 for f16 input
Differential Revision: https://reviews.llvm.org/D26862
llvm-svn: 287389
Eric Fiselier [Fri, 18 Nov 2016 22:25:41 +0000 (22:25 +0000)]
Allow using libsupc++ with LIBCXX_ENABLE_STATIC_ABI_LIBRARY. Patch from Michael Daniels.
The code cannot currently link when using libsupc++ with the
LIBCXX_ENABLE_STATIC_ABI_LIBRARY option.
This change ifdef's out the the destructor and 'what' function for
bad_array_length and bad_array_new_length when GLIBCXX is defined.
The constructors that are left in are the only functions not being provided by
libsupc++ itself, and follows the same pattern that was used to ifdef bad_alloc.
Testing was done on a Linux x86_64 host using GCC 5.4 and libc++ from ToT.
I see no change to the test results when using libsup++ or libstdc++ without
LIBCXX_ENABLE_STATIC_ABI_LIBRARY. When using libsupc++ with
LIBCXX_ENABLE_STATIC_ABI_LIBRARY it will now build and test results are the
same as those without the option specified.
Reviewed as https://reviews.llvm.org/D26186
llvm-svn: 287388
Simon Pilgrim [Fri, 18 Nov 2016 22:21:22 +0000 (22:21 +0000)]
[SelectionDAG] Add knowbits support for CONCAT_VECTOR opcode
llvm-svn: 287387
Jim Ingham [Fri, 18 Nov 2016 22:06:10 +0000 (22:06 +0000)]
Fix "thread step until" handling of multiple line inputs.
Also document that it handles same, and add some tests.
llvm-svn: 287386
Simon Pilgrim [Fri, 18 Nov 2016 21:59:38 +0000 (21:59 +0000)]
[X86] Add knownbits concat_vector test
Support coming in a future patch
llvm-svn: 287385
Eugene Zelenko [Fri, 18 Nov 2016 21:57:58 +0000 (21:57 +0000)]
[Examples] Fix some Clang-tidy modernize-use-default and Include What You Use warnings; other minor fixes.
Differential revision: https://reviews.llvm.org/D26433
llvm-svn: 287384
Stephan T. Lavavej [Fri, 18 Nov 2016 21:54:43 +0000 (21:54 +0000)]
[libcxx] [test] D26816: Fix non-Standard assumptions when testing sample().
sample() isn't specified with a reproducible algorithm, so expecting
exact output is non-Standard. Mark those tests with LIBCPP_ASSERT.
In test_small_population(), we're guaranteed to get all of the elements,
but not necessarily in their original order. When PopulationCategory is
forward, we're guaranteed stability (and can therefore test equal()).
Otherwise, we can only test is_permutation(). (As it happens, both libcxx
and MSVC's STL provide stability in this scenario for input-only iterators.)
llvm-svn: 287383
Stephan T. Lavavej [Fri, 18 Nov 2016 21:54:38 +0000 (21:54 +0000)]
[libcxx] [test] D26815: Fix an assumption about the state of moved-from std::functions.
The Standard doesn't provide any guarantees beyond "valid but unspecified" for
moved-from std::functions. libcxx moves from small targets and leaves them
there, while MSVC's STL empties out the source. Mark these assertions as
libcxx-specific.
llvm-svn: 287382
Stephan T. Lavavej [Fri, 18 Nov 2016 21:54:31 +0000 (21:54 +0000)]
[libcxx] [test] D26813: allocator<const T> is non-Standard.
N4582 17.6.3.5 [allocator.requirements] says that allocators are given
cv-unqualified object types, and N4582 20.9.9 [default.allocator]
implies that allocator<const T> is ill-formed (due to colliding
address() overloads). Therefore, tests for allocator<const T>
should be marked as libcxx-specific (if not removed outright).
llvm-svn: 287381
Dominic Chen [Fri, 18 Nov 2016 21:07:03 +0000 (21:07 +0000)]
[analyzer] Refactor recursive symbol reachability check to use symbol_iterator
Reviewers: zaks.anna, dcoughlin
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D26773
llvm-svn: 287380
Michael Zolotukhin [Fri, 18 Nov 2016 21:01:12 +0000 (21:01 +0000)]
[LoopSimplify] Preserve LCSSA when removing edges from unreachable blocks.
This fixes PR30454.
llvm-svn: 287379
Oleg Ranevskyy [Fri, 18 Nov 2016 21:00:08 +0000 (21:00 +0000)]
[ARM] Fix sema check of ARM special register names
Summary:
This is a simple sema check patch for arguments of `__builtin_arm_rsr` and the related builtins, which currently do not allow special registers with indexes >7.
Some of the possible register name formats these builtins accept are:
```
{c}p<coprocessor>:<op1>:c<CRn>:c<CRm>:<op2>
```
```
o0:op1:CRn:CRm:op2
```
where `op1` / `op2` are integers in the range [0, 7] and `CRn` / `CRm` are integers in the range [0, 15].
The current sema check does not allow `CRn` > 7 and accepts `op2` up to 15.
Reviewers: LukeCheeseman, rengolin
Subscribers: asl, aemerson, rengolin, cfe-commits
Differential Revision: https://reviews.llvm.org/D26464
llvm-svn: 287378