Ed Schouten [Mon, 23 Feb 2015 09:12:31 +0000 (09:12 +0000)]
Add C11 *_DECIMAL_DIG.
Before C11 there was only the DECIMAL_DIG definition. As of C11, we now
have one definition per floating point type (e.g. DBL_DECIMAL_DIG).
Change the existing code to define the new versions. To remain backward
compatible, define __DECIMAL_DIG__ as __LDBL_DECIMAL_DIG__.
Also update the tests. It seems that some of the existing test vectors
were incorrect. Change all tests for __DECIMAL_DIG__ to expect
__LDBL_DECIMAL_DIG__. Add tests for *_DECIMAL_DIG for FreeBSD/amd64, as
I happen to have such a system laying around. I've validated that the
values are in sync with <float.h>.
llvm-svn: 230207
David Majnemer [Mon, 23 Feb 2015 07:13:52 +0000 (07:13 +0000)]
AsmParser: Check ConstantExpr insertvalue operands for type correctness
llvm-svn: 230206
Zachary Turner [Mon, 23 Feb 2015 06:13:27 +0000 (06:13 +0000)]
[llvm-pdbdump] Fix builders again.
llvm-svn: 230205
Zachary Turner [Mon, 23 Feb 2015 05:59:14 +0000 (05:59 +0000)]
[llvm-pdbdump] Very minor code cleanup.
This just removes some dead enums as well as some debug flushes
of stdout.
llvm-svn: 230204
Zachary Turner [Mon, 23 Feb 2015 05:58:34 +0000 (05:58 +0000)]
[llvm-pdbdump] Add an option to dump full class definitions.
This adds the --class-definitions flag. If specified, when dumping
types, instead of "class Foo" you will see the full class definition,
with member functions, constructors, access specifiers.
NOTE: Using this option can be very slow, as generating a full class
definition requires accessing many different parts of the PDB.
llvm-svn: 230203
Lang Hames [Mon, 23 Feb 2015 04:45:05 +0000 (04:45 +0000)]
[Orc][Kaleidoscope] Tidy up the lazy_irgen tutorial, touch up a couple of
comments in the fully_lazy tutorial to minimize the diff between the two.
llvm-svn: 230202
Lang Hames [Mon, 23 Feb 2015 04:34:43 +0000 (04:34 +0000)]
[Orc][Kaleidoscope] Remove dead AST map in SessionContext.
llvm-svn: 230201
Nico Weber [Mon, 23 Feb 2015 03:31:29 +0000 (03:31 +0000)]
Try to fix reST markup for an external link.
llvm-svn: 230200
Nico Weber [Mon, 23 Feb 2015 02:23:19 +0000 (02:23 +0000)]
Remove comment addressed by d0k in r229327.
llvm-svn: 230199
Alexander Kornienko [Mon, 23 Feb 2015 01:12:41 +0000 (01:12 +0000)]
Fixed script name in the clang-tidy documentation.
llvm-svn: 230198
David Blaikie [Mon, 23 Feb 2015 00:53:35 +0000 (00:53 +0000)]
Fix Makefile build
llvm-svn: 230197
David Blaikie [Mon, 23 Feb 2015 00:36:25 +0000 (00:36 +0000)]
[orc] Add a trivial unit test to get the ball rolling
I made my best guess at the Makefile, since I don't have a make build.
I'm not sure if it should be valid to add an empty list of things, but
it seemed the sort of degenerate case.
llvm-svn: 230196
Shankar Easwaran [Mon, 23 Feb 2015 00:30:00 +0000 (00:30 +0000)]
[ELF] Add section group/COMDAT support.
SHF_GROUP: Group Member Sections
----------------------------------
A section which is part of a group, and is to be retained or discarded with the
group as a whole, is identified by a new section header attribute: SHF_GROUP
This section is a member (perhaps the only one) of a group of sections, and the
linker should retain or discard all or none of the members. This section must be
referenced in a SHT_GROUP section. This attribute flag may be set in any section
header, and no other modification or indication is made in the grouped sections.
All additional information is contained in the associated SHT_GROUP section.
SHT_GROUP: Section Group Definition
-------------------------------------
Represents a group section.
The section group's sh_link field identifies a symbol table section, and its
sh_info field the index of a symbol in that section. The name of that symbol is
treated as the identifier of the section group.
More information: https://mentorembedded.github.io/cxx-abi/abi/prop-72-comdat.html
Added a lot of extensive tests, that tests functionality.
llvm-svn: 230195
Shankar Easwaran [Mon, 23 Feb 2015 00:04:49 +0000 (00:04 +0000)]
[ELF] Add .gnu.linkonce support.
When the GNU linker sees two input sections with the same name, and the name
starts with ".gnu.linkonce.", the linker will only keep one copy and discard the
other. Any section whose name starts with “.gnu.linkonce.” is a COMDAT section.
Some architectures like Hexagon use this section to store floating point constants,
that need be deduped.
This patch adds gnu.linkonce functionality to the ELFReader.
llvm-svn: 230194
David Majnemer [Mon, 23 Feb 2015 00:01:32 +0000 (00:01 +0000)]
AsmParser: Call instructions can't have an alignment
llvm-svn: 230193
Shankar Easwaran [Sun, 22 Feb 2015 23:54:38 +0000 (23:54 +0000)]
[Core,MachO,Test] Remove trailing whitespace.
llvm-svn: 230192
Shankar Easwaran [Sun, 22 Feb 2015 23:46:21 +0000 (23:46 +0000)]
[ELF] Add symbol to ELFReference.
Relocation handling need more information about the Symbol that we are creating
references for.
No change in functionality.
llvm-svn: 230191
Shankar Easwaran [Sun, 22 Feb 2015 23:40:58 +0000 (23:40 +0000)]
[Core] Fix handling of Section Groups.
There is code(added by me) in the YAMLReader which isn't correct when it handles references
for section groups. The test case was also checking for wrong outputs.
This fixes the bug and the testcase so that they check for proper outputs.
llvm-svn: 230190
Shankar Easwaran [Sun, 22 Feb 2015 23:32:34 +0000 (23:32 +0000)]
[ELF][X86_64] R_X86_64_16 relocation support
llvm-svn: 230189
David Majnemer [Sun, 22 Feb 2015 23:14:52 +0000 (23:14 +0000)]
AsmParser: Check ConstantExpr GEP operands for validity
llvm-svn: 230188
David Majnemer [Sun, 22 Feb 2015 22:33:57 +0000 (22:33 +0000)]
llvm-pdbdump: Fix gcc/clang build
llvm-svn: 230187
Zachary Turner [Sun, 22 Feb 2015 22:20:26 +0000 (22:20 +0000)]
[llvm-pdbdump] Fix a few compilation failures.
llvm-svn: 230186
David Blaikie [Sun, 22 Feb 2015 22:18:55 +0000 (22:18 +0000)]
Add missing header
llvm-svn: 230185
Zachary Turner [Sun, 22 Feb 2015 22:03:38 +0000 (22:03 +0000)]
[llvm-pdbdump] Rewrite dumper using visitor pattern.
This increases the flexibility of how to dump different
symbol types -- necessary for context-sensitive formatting of
symbol types -- and also improves the modularity by allowing
the dumping to be implemented in the actual dumper, as opposed
to in the PDB library.
llvm-svn: 230184
Ismail Pazarbasi [Sun, 22 Feb 2015 22:01:09 +0000 (22:01 +0000)]
Unix/BSD system calls are prefixed with SYS_ on FreeBSD and Mac
Also, __syscall form should be used when one or more of the
parameters is a 64-bit argument to ensure that argument alignment
is correct.
llvm-svn: 230183
Zachary Turner [Sun, 22 Feb 2015 21:45:38 +0000 (21:45 +0000)]
[llvm-pdbdump] Simplify options and output.
This removes a wealth of options, and instead now only provides
three options. -symbols, -types, and -compilands. This greatly
simplifies use of the tool, and makes it easier to understand
what you're going to see when you run the tool.
llvm-svn: 230182
David Blaikie [Sun, 22 Feb 2015 20:58:38 +0000 (20:58 +0000)]
Roll condition into an assert then wrap it 'ifndef NDEBUG' to protect from the inevitable "unused variable" warning in a non-asserts build.
llvm-svn: 230181
JF Bastien [Sun, 22 Feb 2015 19:32:03 +0000 (19:32 +0000)]
Use common parse routine to read alignment values from bitcode
While fuzzing LLVM bitcode files, I discovered that (1) the bitcode reader doesn't check that alignments are no larger than 2**29; (2) downstream code doesn't check the range; and (3) for values out of range, corresponding large memory requests (based on alignment size) will fail. This code fixes the bitcode reader to check for valid alignments, fixing this problem.
This CL fixes alignment value on global variables, functions, and instructions: alloca, load, load atomic, store, store atomic.
Patch by Karl Schimpf (kschimpf@google.com).
llvm-svn: 230180
Renato Golin [Sun, 22 Feb 2015 18:58:18 +0000 (18:58 +0000)]
Revert "[ASan] Make the argument of '__sanitizer_annotate_contiguous_container' is not aligned error message easier to understand"
This reverts commit r230019, as it was breaking the ARM sanitizer buildbot
and let other errors be introduced since it wasn't fixed/reverted in
time.
llvm-svn: 230179
Hal Finkel [Sun, 22 Feb 2015 18:35:32 +0000 (18:35 +0000)]
[LICM] Refactor to expose functionality as utility functions
This refactors the core functionality of LICM: HoistRegion, SinkRegion and
PromoteAliasSet (renamed to promoteLoopAccessesToScalars) as utility functions
in LoopUtils. This will enable other transformations to make use of them
directly.
Patch by Ashutosh Nema.
llvm-svn: 230178
Simon Pilgrim [Sun, 22 Feb 2015 18:17:28 +0000 (18:17 +0000)]
[DagCombiner] Generalized BuildVector Vector Concatenation
The CONCAT_VECTORS combiner pass can transform the concat of two BUILD_VECTOR nodes into a single BUILD_VECTOR node.
This patch generalises this to support any number of BUILD_VECTOR nodes, and also permits UNDEF nodes to be included as well.
This was noticed as AVX vec128 -> vec256 canonicalization sometimes creates a CONCAT_VECTOR with a real vec128 lower and an vec128 UNDEF upper.
Differential Revision: http://reviews.llvm.org/D7816
llvm-svn: 230177
Hal Finkel [Sun, 22 Feb 2015 16:10:22 +0000 (16:10 +0000)]
[DAGCombine] Don't assume integer-type legailty in reduceBuildVecConvertToConvertBuildVec
DAGCombine will rewrite an BUILD_VECTOR where all non-undef inputs some from
[US]INT_TO_FP, as a BUILD_VECTOR of integers with the conversion applied as a
vector operation. We check operation legality of the conversion, but fail to
check legality of the integer vector type itself. Because targets don't
normally override operation legality defaults for illegal types, we need to
check this also.
This came up in the context of the QPX vector entensions for PowerPC (which can
have legal floating-point vector types without corresponding legal integer
vector types). No in-tree test case for this yes, but one can be added once
the QPX support has been committed.
llvm-svn: 230176
Hal Finkel [Sun, 22 Feb 2015 15:58:04 +0000 (15:58 +0000)]
[SDAG] Use correct alignments on expanded vector trunc-store/ext-loads
When expanding a truncating store or extending load using vector extracts or
inserts and scalar stores and loads, we were giving each of these scalar stores
or loads the same alignment as the original vector operation. While this will
often be right (most vector operations, especially those produced by
autovectorization, have the alignment of the underlying scalar type), the
vector operation could certainly have a larger alignment.
No test case (yet); noticed by inspection.
llvm-svn: 230175
Dmitri Gribenko [Sun, 22 Feb 2015 12:20:05 +0000 (12:20 +0000)]
Revert r230171 "TSan runtime: unbreak the build with ccache"
It broke a buildbot with a non-clang (?) compiler.
llvm-svn: 230174
Kuba Brecka [Sun, 22 Feb 2015 11:49:21 +0000 (11:49 +0000)]
Whitespace fix for r230172.
llvm-svn: 230173
Kuba Brecka [Sun, 22 Feb 2015 11:12:17 +0000 (11:12 +0000)]
Fix gc-test.cc to work under higher -O levels
The gc-test.cc tries underflows of a variable up to -32 bytes, but on i386, the left redzone is not 32 bytes, it’s only 16 bytes and therefore the access to var[-32] is completely off. The reason why this test didn’t fail before is that we’ve been lucky and there was another variable before the var array, which was also instrumented. This fix uses “-32” for 64-bit systems and “-16” for 32-bit.
Reviewed at http://reviews.llvm.org/D7809
llvm-svn: 230172
Dmitri Gribenko [Sun, 22 Feb 2015 11:04:59 +0000 (11:04 +0000)]
TSan runtime: unbreak the build with ccache
llvm-svn: 230171
NAKAMURA Takumi [Sun, 22 Feb 2015 09:58:29 +0000 (09:58 +0000)]
Fix a warning on HexagonMCCodeEmitter::MCII. [-Wunused-private-field]
llvm-svn: 230170
NAKAMURA Takumi [Sun, 22 Feb 2015 09:58:19 +0000 (09:58 +0000)]
RewriteStatepointsForGC.cpp: Fix for -Asserts to mark isNullConstant() as LLVM_ATTRIBUTE_UNUSED. [-Wunused-function]
llvm-svn: 230169
NAKAMURA Takumi [Sun, 22 Feb 2015 09:58:13 +0000 (09:58 +0000)]
RewriteStatepointsForGC.cpp: Fix for -Asserts. [-Wunused-variable]
llvm-svn: 230168
NAKAMURA Takumi [Sun, 22 Feb 2015 09:51:42 +0000 (09:51 +0000)]
LowerBitSets.cpp: Prune incorrect \param(s). [-Wdocumentation]
\param should be used as itemized.
llvm-svn: 230167
Chandler Carruth [Sun, 22 Feb 2015 08:34:47 +0000 (08:34 +0000)]
Switch to use override, fixes -Winconsistent-missing-override on LLD.
llvm-svn: 230166
Craig Topper [Sun, 22 Feb 2015 07:50:41 +0000 (07:50 +0000)]
[X86] Add some missing redundant MMX and SSE encodings for disassembler.
llvm-svn: 230165
Zachary Turner [Sun, 22 Feb 2015 07:31:42 +0000 (07:31 +0000)]
Remove log statements from config scripts.
The bots seem to be happy now.
llvm-svn: 230164
Zachary Turner [Sun, 22 Feb 2015 07:13:52 +0000 (07:13 +0000)]
Really fix the build this time.
I was setting the python variable to "@HAVE_DIA_SDK@", which will
always be a string, and will always evaluate to True.
llvm-svn: 230163
Zachary Turner [Sun, 22 Feb 2015 07:01:41 +0000 (07:01 +0000)]
Attempt to fix the builders.
The issue was that the test Makefile had not been updated to
provide a value for HAVE_DIA_SDK, so it was being initialized
incorrectly. Hopefully this brings everything back to green.
llvm-svn: 230162
Zachary Turner [Sun, 22 Feb 2015 06:47:32 +0000 (06:47 +0000)]
[llvm-pdbdump] Resubmit "Add some tests for llvm-pdbdump".
NOTE: This patch intentionally breaks the build. It attempts
to resubmit r230083, but with some debug logging in the CMake
and lit config files to determine why certain bots do not
correctly disable the DIA tests when DIA is not available.
After a sufficient number of bots fail, this patch will either
be reverted or, if the cause of the failure becomes obvious,
a fix submitted with the log statements removed.
llvm-svn: 230161
Larisse Voufo [Sun, 22 Feb 2015 06:36:53 +0000 (06:36 +0000)]
Relax the requirement on sized deallocation a bit: Default on unsized delete if sized delete is not provided in global scope, and -fdefine-sized-deallocation option is disabled.
llvm-svn: 230160
Larisse Voufo [Sun, 22 Feb 2015 06:36:49 +0000 (06:36 +0000)]
Fix typo.
llvm-svn: 230159
Larisse Voufo [Sun, 22 Feb 2015 06:36:46 +0000 (06:36 +0000)]
Move -fdefine-sized-deallocation and -fno-sized-deallocation options from driver into CC1 for now.
llvm-svn: 230158
Davide Italiano [Sun, 22 Feb 2015 03:12:21 +0000 (03:12 +0000)]
[ELF] Teach GNU Driver about --stats.
This is mainly for back-compatibility with GNU ld.
Ideally --stats should be a general option in LinkingContext, providing
individual stats for every pass in the linking process.
In the GNU driver, a better wording could be used, but there's no need
to change it for now.
Differential Revision: D7657
Reviewed by: ruiu
llvm-svn: 230157
David Majnemer [Sun, 22 Feb 2015 02:35:27 +0000 (02:35 +0000)]
COFF: Add 'IMAGE_SCN_CNT_INITIALIZED_DATA' to all DWARF sections
The CodeView debug info section, .debug$S, also has this set. MinGW
sets this bit for their DWARF sections as well.
llvm-svn: 230156
David Majnemer [Sun, 22 Feb 2015 02:35:22 +0000 (02:35 +0000)]
COFF: Consistently format the DWARF sections
llvm-svn: 230155
Lang Hames [Sun, 22 Feb 2015 01:48:23 +0000 (01:48 +0000)]
[Orc] Remove redundant using directive.
llvm-svn: 230154
Lang Hames [Sun, 22 Feb 2015 01:45:31 +0000 (01:45 +0000)]
[Orc] Add header comment to IndirectionUtils.cpp.
llvm-svn: 230153
Nico Weber [Sun, 22 Feb 2015 00:27:32 +0000 (00:27 +0000)]
Remove two unused methods. No behavior change.
llvm-svn: 230152
Sanjoy Das [Sat, 21 Feb 2015 22:20:22 +0000 (22:20 +0000)]
IRCE: generalize InductiveRangeCheck::computeSafeIterationSpace to
work with a non-canonical induction variable.
This is currently a non-functional change because we only ever call
computeSafeIterationSpace on a canonical induction variable; but the
generalization will be useful in a later commit.
llvm-svn: 230151
Sanjoy Das [Sat, 21 Feb 2015 22:07:32 +0000 (22:07 +0000)]
IRCE: use SCEVs instead of llvm::Value's for intermediate
calculations. Semantically non-functional change.
This gets rid of some of the SCEV -> Value -> SCEV round tripping and
the Construct(SMin|SMax)Of and MaybeSimplify helper routines.
llvm-svn: 230150
Matt Arsenault [Sat, 21 Feb 2015 21:29:10 +0000 (21:29 +0000)]
R600/SI: Use v_madmk_f32
llvm-svn: 230149
Matt Arsenault [Sat, 21 Feb 2015 21:29:07 +0000 (21:29 +0000)]
R600/SI: Try to use v_madak_f32
This is a code size optimization when the constant
only has one use.
llvm-svn: 230148
Matt Arsenault [Sat, 21 Feb 2015 21:29:04 +0000 (21:29 +0000)]
R600/SI: Don't crash when getting immediate operand size
llvm-svn: 230147
Matt Arsenault [Sat, 21 Feb 2015 21:29:00 +0000 (21:29 +0000)]
R600/SI: Fix mad*k definitions
llvm-svn: 230146
Simon Pilgrim [Sat, 21 Feb 2015 21:25:16 +0000 (21:25 +0000)]
[X86][SSE] Added shuffle based integer zero extension tests.
llvm-svn: 230145
Justin Bogner [Sat, 21 Feb 2015 20:53:36 +0000 (20:53 +0000)]
docs: Mention that assertions must be enabled to use the -stats flag
Patch by Rob Stewart. Thanks!
llvm-svn: 230144
Lang Hames [Sat, 21 Feb 2015 20:44:36 +0000 (20:44 +0000)]
[Orc] Move Orc code into a namespace (llvm::orc), update Kaleidoscope code.
NFC.
llvm-svn: 230143
Benjamin Kramer [Sat, 21 Feb 2015 17:08:08 +0000 (17:08 +0000)]
MachineInstr: Use range-based for loops. NFC.
llvm-svn: 230142
Benjamin Kramer [Sat, 21 Feb 2015 16:22:48 +0000 (16:22 +0000)]
Calling memmove on a MachineOperand is totally safe.
While it's not POD due to the user-defined constructor, it's still a trivially
copyable type. No functional change.
llvm-svn: 230141
Johannes Doerfert [Sat, 21 Feb 2015 16:18:28 +0000 (16:18 +0000)]
Remove leftover code
llvm-svn: 230140
Shankar Easwaran [Sat, 21 Feb 2015 15:51:54 +0000 (15:51 +0000)]
[ELF] Remove FIXME(s) that are already fixed.
FIXME code was left around in few places where its already been taken care of.
This removes the FIXME's that are not needed.
llvm-svn: 230139
Shankar Easwaran [Sat, 21 Feb 2015 15:49:34 +0000 (15:49 +0000)]
[ELF][Writer] Use Path to create AtomSection.
Now since the correct file path for atoms is available and not clobbered,
commit r222309 which was reverted previously can be added back.
No change in functionality.
llvm-svn: 230138
Benjamin Kramer [Sat, 21 Feb 2015 14:35:00 +0000 (14:35 +0000)]
Remove dead prototype.
llvm-svn: 230137
Benjamin Kramer [Sat, 21 Feb 2015 14:31:29 +0000 (14:31 +0000)]
X86: Remove custom lowering of SIGN_EXTEND_INREG
This was just replicating logic from the legalizer. Covered by existing
tests.
llvm-svn: 230136
Eric Christopher [Sat, 21 Feb 2015 09:09:15 +0000 (09:09 +0000)]
Unconditionally create a new MCInstrInfo in the asm printer for
asm parsing since it's not subtarget dependent and we can't depend
upon the one hanging off the MachineFunction's subtarget still
being around.
llvm-svn: 230135
Eric Christopher [Sat, 21 Feb 2015 08:48:23 +0000 (08:48 +0000)]
Remove obsolete comment.
llvm-svn: 230134
Eric Christopher [Sat, 21 Feb 2015 08:48:22 +0000 (08:48 +0000)]
Have the MipsAsmPrinter fp stub emission code take a custom
MCSubtargetInfo as the MachineFunction has gone away and we need
to emit code at the module level.
llvm-svn: 230133
Eric Christopher [Sat, 21 Feb 2015 08:32:38 +0000 (08:32 +0000)]
Turn an if+llvm_unreachable into an assert and reword comment.
llvm-svn: 230132
Eric Christopher [Sat, 21 Feb 2015 08:32:22 +0000 (08:32 +0000)]
Endianness can be gotten from the DataLayout which we already
have. Also, the subtarget is invalid at this point.
llvm-svn: 230131
Jason Molenda [Sat, 21 Feb 2015 06:13:51 +0000 (06:13 +0000)]
If constructed llvm_build_dir doesn't exist, retry as an iphoneos path.
llvm-svn: 230130
David Majnemer [Sat, 21 Feb 2015 05:49:45 +0000 (05:49 +0000)]
X86: Call __main using the SelectionDAG
Synthesizing a call directly using the MI layer would confuse the frame
lowering code. This is problematic as frame lowering is highly
sensitive the particularities of calls, etc.
llvm-svn: 230129
Zachary Turner [Sat, 21 Feb 2015 04:45:12 +0000 (04:45 +0000)]
Add missing #include. This should fix the Windows build.
llvm-svn: 230128
Shankar Easwaran [Sat, 21 Feb 2015 04:42:43 +0000 (04:42 +0000)]
[ELF] Fix References being ignored.
The ELFReader was skipping references for sections that contained relocations.
This fixes the bug.
llvm-svn: 230127
Chandler Carruth [Sat, 21 Feb 2015 04:31:13 +0000 (04:31 +0000)]
Update for Clang API change in r230123 -- lookup_result was always
const, there was never a need for lookup_const_result. Now that vestigal
type is gone, so switch LLDB to lookup_result and to use the
DeclContextLookupResult rather than the Const variant.
llvm-svn: 230126
Eric Fiselier [Sat, 21 Feb 2015 04:30:36 +0000 (04:30 +0000)]
Add temporary workaround for missing symbol __cxa_throw_bad_new_array_length on OS X.
llvm-svn: 230125
Shankar Easwaran [Sat, 21 Feb 2015 04:28:26 +0000 (04:28 +0000)]
[obj2yaml/yaml2obj] Add SHT_GROUP support.
This adds section group support to the tools obj2yaml and yaml2obj.
llvm-svn: 230124
Richard Smith [Sat, 21 Feb 2015 02:45:19 +0000 (02:45 +0000)]
Cleanup: remove artificial division between lookup results and const lookup
results. No-one was ever modifying a lookup result, and it would not be
reasonable to do so.
llvm-svn: 230123
Davide Italiano [Sat, 21 Feb 2015 02:36:54 +0000 (02:36 +0000)]
Small cleanup. Don't use else when not needed.
Pointed out by David Majnemer.
llvm-svn: 230122
Richard Smith [Sat, 21 Feb 2015 02:31:57 +0000 (02:31 +0000)]
Revert r167816 and replace it with a proper fix for the issue: do not
invalidate lookup_iterators and lookup_results for some name within a
DeclContext if the lookup results for a *different* name change.
llvm-svn: 230121
Eric Fiselier [Sat, 21 Feb 2015 02:30:41 +0000 (02:30 +0000)]
[libc++] Try and prevent evaluation of `is_default_constructible` on tuples default constructor if it is not needed.
Summary:
Currently parts of the SFINAE on tuples default constructor always gets evaluated even when the default constructor is never called or instantiated. This can cause a hard compile error when a tuple is created with types that do not have a default constructor. Below is a self contained example using a pair like class. This code will not compile but probably should.
```
#include <type_traits>
template <class T>
struct IllFormedDefaultImp {
IllFormedDefaultImp(T x) : value(x) {}
constexpr IllFormedDefaultImp() {}
T value;
};
typedef IllFormedDefaultImp<int &> IllFormedDefault;
template <class T, class U>
struct pair
{
template <bool Dummy = true,
class = typename std::enable_if<
std::is_default_constructible<T>::value
&& std::is_default_constructible<U>::value
&& Dummy>::type
>
constexpr pair() : first(), second() {}
pair(T const & t, U const & u) : first(t), second(u) {}
T first;
U second;
};
int main()
{
int x = 1;
IllFormedDefault v(x);
pair<IllFormedDefault, IllFormedDefault> p(v, v);
}
```
One way to fix this is to use `Dummy` in a more involved way in the constructor SFINAE. The following patch fixes these sorts of hard compile errors for tuple.
Reviewers: mclow.lists, rsmith, K-ballo, EricWF
Reviewed By: EricWF
Subscribers: ldionne, cfe-commits
Differential Revision: http://reviews.llvm.org/D7569
llvm-svn: 230120
Eric Fiselier [Sat, 21 Feb 2015 02:26:24 +0000 (02:26 +0000)]
[libcxx] Move to using libc++abi2.exp as the default symbol list for libc++
Summary:
libc++abi2.exp should be used whenever `cxxabi.h` defines `_LIBCPPABI_VERSION`. This macro was added to libc++abi in 2012 in r149632. For this reason we should use libc++abi2.exp as default unless otherwise specified.
Also when building against an in-tree libc++abi we definitely want to use libc++abi2.exp.
I would love to know what OSX was the last to use libc++abi.exp but I can only test on 10.9.
Reviewers: danalbert, mclow.lists, EricWF
Reviewed By: EricWF
Subscribers: meadori, cfe-commits
Differential Revision: http://reviews.llvm.org/D7773
llvm-svn: 230119
Tim Northover [Sat, 21 Feb 2015 02:11:17 +0000 (02:11 +0000)]
CodeGen: convert CCState interface to using ArrayRefs
Everyone except R600 was manually passing the length of a static array
at each callsite, calculated in a variety of interesting ways. Far
easier to let ArrayRef handle that.
There should be no functional change, but out of tree targets may have
to tweak their calls as with these examples.
llvm-svn: 230118
Peter Collingbourne [Sat, 21 Feb 2015 01:48:51 +0000 (01:48 +0000)]
Run simple-fail.cpp at -O{1,2,3}.
It's probably overkill to run the other tests at -O* given the increasing
combinatorial explosion.
llvm-svn: 230117
Peter Collingbourne [Sat, 21 Feb 2015 01:36:08 +0000 (01:36 +0000)]
CFI: Add tests for 32-bit, 64-bit and memory bitsets. Break optimization in more places.
llvm-svn: 230116
Hans Wennborg [Sat, 21 Feb 2015 01:07:24 +0000 (01:07 +0000)]
Improve diagnostic when failing to synthesize implicit member due to dllexport (PR22591)
This is only a problem in C++03 mode targeting MS ABI (MinGW doesn't
export inline methods, and C++11 marks these methods implicitly
deleted).
Since targeting the MS ABI in pre-C++11 mode is a rare configuration,
this will probably not get fixed, but we can at least have a better
error message.
llvm-svn: 230115
Larisse Voufo [Sat, 21 Feb 2015 01:05:29 +0000 (01:05 +0000)]
Fix typo.
llvm-svn: 230114
David Majnemer [Sat, 21 Feb 2015 01:04:47 +0000 (01:04 +0000)]
Win64: Stack alignment constraints aren't applied during SET_FPREG
Stack realignment occurs after the prolog, not during, for Win64.
Because of this, don't factor in the maximum stack alignment when
establishing a frame pointer.
This fixes PR22572.
llvm-svn: 230113
Larisse Voufo [Sat, 21 Feb 2015 01:03:48 +0000 (01:03 +0000)]
Move -fdefine-sized-deallocation and -fno-sized-deallocation options from driver into CC1 for now.
llvm-svn: 230112
Duncan P. N. Exon Smith [Sat, 21 Feb 2015 01:02:18 +0000 (01:02 +0000)]
AsmParser/Writer: Handle symbolic constants in DI 'flags:'
Parse (and write) symbolic constants in debug info `flags:` fields.
This prevents a readability (and CHECK-ability) regression with the new
debug info hierarchy.
Old (well, current) assembly, with pretty-printing:
!{!"...\\0016387", ...} ; ... [public] [rvalue reference]
Flags field without this change:
!MDDerivedType(flags: 16387, ...)
Flags field with this change:
!MDDerivedType(flags: DIFlagPublic | DIFlagRValueReference, ...)
As discussed in the review thread, this isn't a final state. Most of
these flags correspond to `DW_AT_` symbolic constants, and we might
eventually want to support arbitrary attributes in some form. However,
as it stands now, some of the flags correspond to other concepts (like
`FlagStaticMember`); until things are refactored this is the simplest
way to move forward without regressing assembly.
llvm-svn: 230111
Greg Clayton [Sat, 21 Feb 2015 00:57:59 +0000 (00:57 +0000)]
Add missing Xcode project files for MICmdCmdSymbol.cpp and MICmdCmdSymbol.h.
llvm-svn: 230110
Fariborz Jahanian [Sat, 21 Feb 2015 00:45:58 +0000 (00:45 +0000)]
[Objective-C]. Provide a new formatting kind, "os_trace" which
can take a "const char*" format but supports standard printf
and CF/NS types . rdar://
19904147
llvm-svn: 230109
Duncan P. N. Exon Smith [Sat, 21 Feb 2015 00:45:26 +0000 (00:45 +0000)]
IR: Add helper to split debug info flags bitfield
Split debug info 'flags' bitfield over a vector so the current flags can
be iterated over. This API (in combination with r230107) will be used
for assembly support for symbolic constants.
llvm-svn: 230108