summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Misc cleanups to the FileSytem api.Rafael Espindola2014-09-117-86/+66
| | | | | | | | | | | | | | | | The main difference is the removal of std::error_code exists(const Twine &path, bool &result); It was an horribly redundant interface since a file not existing is also a valid error_code. Now we have an access function that returns just an error_code. This is the only function that has to be implemented for Unix and Windows. The functions can_write, exists and can_execute an now just wrappers. One still has to be very careful using these function to avoid introducing race conditions (Time of check to time of use). llvm-svn: 217625
* Fix Windows build.Zachary Turner2014-09-112-0/+3
| | | | | | | | * ssize_t isn't defined by default on Windows. * New public API files need to be defined in a different file for Windows. llvm-svn: 217624
* Add missing colon to RUN line...Bill Schmidt2014-09-111-1/+1
| | | | llvm-svn: 217623
* [PATCH, PowerPC] Accept 'U' and 'X' constraints in inline asmBill Schmidt2014-09-112-0/+42
| | | | | | | | | | | | | | | | | | Inline asm may specify 'U' and 'X' constraints to print a 'u' for an update-form memory reference, or an 'x' for an indexed-form memory reference. However, these are really only useful in GCC internal code generation. In inline asm the operand of the memory constraint is typically just a register containing the address, so 'U' and 'X' make no sense. This patch quietly accepts 'U' and 'X' in inline asm patterns, but otherwise does nothing. If we ever unexpectedly see a non-register, we'll assert and sort it out afterwards. I've added a new test for these constraints; the test case should be used for other asm-constraints changes down the road. llvm-svn: 217622
* Thread Safety Analysis: Avoid infinite recursion in an operator<<Justin Bogner2014-09-111-2/+2
| | | | | | | | | | | | | | | | | | r217556 introduced an operator<<(std::ostream &, StringRef) that seems to self recurse on some systems, because str.data(), which is a char *, was being implicitly converted back to StringRef in overload resolution. This manifested as SemaCXX/warn-thread-safety-analysis.cpp timing out in release builds and overflowing the stack in debug builds. One of the failing systems that saw this is here: http://lab.llvm.org:8013/builders/clang-x86_64-darwin11-nobootstrap-RAincremental/builds/4636 Using ostream's write method instead of operator<< should get the bots going again. llvm-svn: 217621
* [MCJIT] Add support for ARM HALF_DIFF relocations to MCJIT.Lang Hames2014-09-115-73/+202
| | | | | | Fixes <rdar://problem/18297804>. llvm-svn: 217620
* Objective-C. Under a special flag, -Wcstring-format-directive,Fariborz Jahanian2014-09-114-35/+89
| | | | | | | | | off by default, issue a warning if %s directive is used in formart argument of a function/method declared as __attribute__((format(CF/NSString, ...))) To complete rdar://18182443 llvm-svn: 217619
* Use simpler version of sys::fs::exists. NFC.Rafael Espindola2014-09-111-3/+2
| | | | llvm-svn: 217618
* Use the simpler sys::fs:;exists. NFC.Rafael Espindola2014-09-111-2/+1
| | | | llvm-svn: 217617
* [UBSan] Parse common flags from UBSAN_OPTIONS runtime variable even ifAlexey Samsonov2014-09-117-18/+21
| | | | | | UBSan is combined with ASan. llvm-svn: 217616
* clang-cl: Add support for the /o option for object files, executables, and ↵Ehsan Akhgari2014-09-113-8/+169
| | | | | | | | | | | | | | | | | | | | | | | | | preprocessor output Summary: cl.exe recognizes /o as a deprecated and undocumented option similar to /Fe. This patch adds support for this option to clang-cl for /Fe, /Fo and /Fi. It also ensures that the last option among /o and /F* wins, if both specified. This is required at least for building autoconf based software, since autoconf uses -o to specify the executable output. This fixes http://llvm.org/PR20894. Test Plan: The patch includes automated tests. Reviewers: rnk Subscribers: cfe-commits Differential Revision: http://reviews.llvm.org/D5308 llvm-svn: 217615
* Use the simpler version of llvm::sys::fs::exists.Rafael Espindola2014-09-115-20/+7
| | | | | | | In all these cases it looks like the intention was to handle error in a similar way to the file not existing. llvm-svn: 217614
* gdb-remote tests: added test to verify thread stop info when multi-threaded ↵Todd Fiala2014-09-111-9/+46
| | | | | | | | | | app is interrupted. The Linux version is marked XFAIL for the moment, fixing next. Related to http://llvm.org/bugs/show_bug.cgi?id=20908. llvm-svn: 217613
* Add triple to test to fix botsMatt Arsenault2014-09-111-1/+1
| | | | llvm-svn: 217612
* Provide an implementation of getNoopForMachoTarget for SPARC.Brad Smith2014-09-113-0/+15
| | | | llvm-svn: 217611
* Add DAG combine for shl + add of constants.Matt Arsenault2014-09-116-41/+157
| | | | | | | | | | | | | | Do (shl (add x, c1), c2) -> (add (shl x, c2), c1 << c2) This is already done for multiplies, but since multiplies by powers of two are turned into shifts, we also need to handle it here. This might want checks for isLegalAddImmediate to avoid transforming an add of a legal immediate with one that isn't. llvm-svn: 217610
* Update C++ status page to reflect that Clang 3.5 has released.Richard Smith2014-09-111-4/+4
| | | | llvm-svn: 217609
* Update DR status list to reflect that Clang 3.5 has been released.Richard Smith2014-09-112-18/+18
| | | | llvm-svn: 217608
* DWARF64 FixesTodd Fiala2014-09-1110-65/+145
| | | | | | | | | | | | | | | | 1. DW_FORM_strp and DW_FORM_sec_offset are 64bits for DWARF64 / 32bits for DWARF32 They are different from DW_FORM_addr, whose size is specified in .debug_info 2. Bump DWARF version support form [2,3] to [2,4] in DWARFDebugLine.cpp 3. Fix DWARFDebugLine to support DWARF64 See http://reviews.llvm.org/D5307 for more details. Reviewed by Greg Clayton and Jason Molenda. Change by Tong Shen. llvm-svn: 217607
* Tests for DR573-580.Richard Smith2014-09-112-8/+134
| | | | llvm-svn: 217606
* [MCJIT] Take the relocation addend into account when applying ARM MachO VANILLALang Hames2014-09-112-1/+5
| | | | | | | | and BR24 relocations. <rdar://problem/18296496> llvm-svn: 217605
* Adding ABI support for __cxa_throw_bad_array_new_length.Aaron Ballman2014-09-114-0/+48
| | | | llvm-svn: 217604
* Added missing LLVM_NOEXCEPT to the definition of _obj2yaml_error_category::nameIsmail Pazarbasi2014-09-111-1/+3
| | | | | | LLVM_NOEXCEPT was added in r210591. llvm-svn: 217603
* [AVX512] Fix miscompile for unpackAdam Nemet2014-09-112-60/+53
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | r189189 implemented AVX512 unpack by essentially performing a 256-bit unpack between the low and the high 256 bits of src1 into the low part of the destination and another unpack of the low and high 256 bits of src2 into the high part of the destination. I don't think that's how unpack works. AVX512 unpack simply has more 128-bit lanes but other than it works the same way as AVX. So in each 128-bit lane, we're always interleaving certain parts of both operands rather different parts of one of the operands. E.g. for this: __v16sf a = { 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 }; __v16sf b = { 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31 }; __v16sf c = __builtin_shufflevector(a, b, 0, 8, 1, 9, 4, 12, 5, 13, 16, 24, 17, 25, 20, 28, 21, 29); we generated punpcklps (notice how the elements of a and b are not interleaved in the shuffle). In turn, c was set to this: 0 16 1 17 4 20 5 21 8 24 9 25 12 28 13 29 Obviously this should have just returned the mask vector of the shuffle vector. I mostly reverted this change and made sure the original AVX code worked for 512-bit vectors as well. Also updated the tests because they matched the logic from the code. llvm-svn: 217602
* Add triple and remove hashes to account for buildbot differences in comment ↵Sanjay Patel2014-09-111-13/+13
| | | | | | strings. llvm-svn: 217601
* Move constant-sized bitvector to the stack.Benjamin Kramer2014-09-111-2/+2
| | | | llvm-svn: 217600
* Combine fmul vector FP constants when unsafe math is allowed.Sanjay Patel2014-09-112-6/+70
| | | | | | | | | | | | | | | | | | | This is an extension of the change made with r215820: http://llvm.org/viewvc/llvm-project?view=revision&revision=215820 That patch allowed combining of splatted vector FP constants that are multiplied. This patch allows combining non-uniform vector FP constants too by relaxing the check on the type of vector. Also, canonicalize a vector fmul in the same way that we already do for scalars - if only one operand of the fmul is a constant, make it operand 1. Otherwise, we miss potential folds. This fold is also done by -instcombine, but it's possible that extra fmuls may have been generated during lowering. Differential Revision: http://reviews.llvm.org/D5254 llvm-svn: 217599
* Merge GetAddrOfCXXConstructor and GetAddrOfCXXDonstructor. NFC.Rafael Espindola2014-09-119-50/+21
| | | | llvm-svn: 217598
* typoSanjay Patel2014-09-111-1/+1
| | | | llvm-svn: 217597
* R600: Add cmpxchg instruction for evergreenAaron Watry2014-09-112-5/+29
| | | | | | | | | | | | | | | | | Refactored the R600_LDS_1A2D class a bit to get it to actually work. It seemed to be previously unused and broken. We also have to disable the conversion to the noret variant for now in R600ISelLowering because the getLDSNoRetOp method only handles 1A1D LDS ops. Someone can feel free to modify the AMDGPU::getLDSNoRetOp method to work for more than 1A1D variants of LDS operations. It's being left as a future TODO for now. Signed-off-by: Aaron Watry <awatry at gmail.com> Reviewed-by: Matt Arsenault <matthew.arsenault@amd.com> llvm-svn: 217596
* R600: Test local atomics for evergreenAaron Watry2014-09-111-0/+27
| | | | | | | | Now that the operations are all implemented, we can test this sub-arch here. Signed-off-by: Aaron Watry <awatry@gmail.com> Reviewed-by: Matt Arsenault <matthew.arsenault@amd.com> llvm-svn: 217595
* R600: Add LDS_WRXCHG[_RET] instructions for Evergreen.Aaron Watry2014-09-111-0/+4
| | | | | | Signed-off-by: Aaron Watry <awatry@gmail.com> Reviewed-by: Matt Arsenault <matthew.arsenault@amd.com> llvm-svn: 217594
* R600: Add LDS_MIN_[U]INT[_RET] instructions for EvergreenAaron Watry2014-09-111-0/+8
| | | | | | Signed-off-by: Aaron Watry <awatry@gmail.com> Reviewed-by: Matt Arsenault <matthew.arsenault@amd.com> llvm-svn: 217593
* R600: Add LDS_XOR[_RET] instructions for EvergreenAaron Watry2014-09-111-0/+4
| | | | | | Signed-off-by: Aaron Watry <awatry@gmail.com> Reviewed-by: Matt Arsenault <matthew.arsenault@amd.com> llvm-svn: 217592
* R600: Add LDS_OR[_RET] instructions for EvergreenAaron Watry2014-09-111-0/+4
| | | | | | Signed-off-by: Aaron Watry <awatry@gmail.com> Reviewed-by: Matt Arsenault <matthew.arsenault@amd.com> llvm-svn: 217591
* R600: Add LDS_AND[_RET] instructions for EvergreenAaron Watry2014-09-111-0/+4
| | | | | | Signed-off-by: Aaron Watry <awatry@gmail.com> Reviewed-by: Matt Arsenault <matthew.arsenault@amd.com> llvm-svn: 217590
* R600: Add LDS_MAX_[U]INT[_RET] instructions for EvergreenAaron Watry2014-09-111-0/+8
| | | | | | | | | | | | | This was only present for SI before. Cayman may still be missing, but I am unable to test that currently. v2: Don't create atomicrmw max tests in separate file Signed-off-by: Aaron Watry <awatry@gmail.com> Reviewed-by: Matt Arsenault <matthew.arsenault@amd.com> CC: Tom Stellard <thomas.stellard@amd.com> llvm-svn: 217589
* Fix bug 20892 - clang-format does not handle C-style commentsRoman Kashitsyn2014-09-112-2/+19
| | | | | | | | | | | | | | | | | | | | | | Summary: http://llvm.org/bugs/show_bug.cgi?id=20892 Add support of C-style formatting enabling/disabling directives. Now the following two styles are supported: // clang-format on /* clang-format on */ The flexibility in comments (support of extra spaces and/or slashes, etc.) is deliberately avoided to simplify search in large code bases. Reviewers: djasper Reviewed By: djasper Subscribers: cfe-commits, curdeius, klimek Differential Revision: http://reviews.llvm.org/D5309 llvm-svn: 217588
* Remove executable bit on all header filesTobias Grosser2014-09-117-0/+0
| | | | | | Some header files had been marked executable by accident. llvm-svn: 217587
* Avoid some unnecessary SmallVector copies.Benjamin Kramer2014-09-115-13/+11
| | | | | | No functionality change. llvm-svn: 217586
* ARM Unwind syntaxRenato Golin2014-09-112-6/+30
| | | | | | | | | | | This patch fixes the bad argument that GAS accepted but the IAS didn't, ie. {#0x20}, moving it to {0x20} which both accept. It also makes the ARMv7+ save/restore correct by using VFP instructions rather than old co-processor ones. Fixes PR20529. llvm-svn: 217585
* [asan] Preserve existing LD_PRELOAD setting on Android.Evgeniy Stepanov2014-09-111-1/+1
| | | | llvm-svn: 217584
* [docs] Mention character array constants in docs/LangRef.rstDaniel Sanders2014-09-111-1/+3
| | | | | | | | | | | | | | | | | | Summary: They were used in the 'Module Structure' example but weren't otherwise documented. Credit to Reed Kotler for noticing. Reviewers: hans Reviewed By: hans Subscribers: hans, llvm-commits Differential Revision: http://reviews.llvm.org/D5191 llvm-svn: 217583
* [ARM] Add Thumb-2 code size optimization regression test for LSR (register).Tilmann Scheller2014-09-111-0/+9
| | | | llvm-svn: 217582
* [ARM] Add Thumb-2 code size optimization regression test for LSR (immediate).Tilmann Scheller2014-09-111-0/+9
| | | | llvm-svn: 217581
* [AArch64] Reenable the PBQP test now that the leak issue has been fixed.Arnaud A. de Grandmaison2014-09-111-0/+14
| | | | | | | | | | | David Blaikie's commits r217563 & r217564, which added shared_ptr to the CostPool have fixed some memory leak issues exposed by the PBQP with coalescing constraints. The sanitizer bot was failing because of those leaks. Now that the leaks are gone, we can reenable the aarch64/pbqp test. llvm-svn: 217580
* [ARM] Add Thumb-2 code size optimization regression test for LSL (register).Tilmann Scheller2014-09-111-0/+9
| | | | llvm-svn: 217579
* [mach-o]: support optional "0x" prefix for -image_baseTim Northover2014-09-112-2/+6
| | | | llvm-svn: 217578
* [mach-o]: tighten up diagnostics for -image_base optionTim Northover2014-09-112-2/+19
| | | | | | | | | The provided base must also be a multiple of the system's page size, which is a reasonable enough demand. Also check the other diagnostics more thoroughly. llvm-svn: 217577
* [ARM] Add Thumb2 code size optimization regression test for LSL (immediate).Tilmann Scheller2014-09-111-0/+9
| | | | llvm-svn: 217576
OpenPOWER on IntegriCloud