Aaron Ballman [Wed, 11 Jun 2014 19:11:24 +0000 (19:11 +0000)]
The automated server-side process isn't quite right yet, so this is a newly-generated attribute reference. Tanya is looking into the server process when she gets the chance.
llvm-svn: 210691
Fariborz Jahanian [Wed, 11 Jun 2014 19:10:46 +0000 (19:10 +0000)]
Objective-C. More tests for both bridging attributes and
a fix to make it work when CFStructs have no definition.
// rdar://
17238954.
llvm-svn: 210690
Rafael Espindola [Wed, 11 Jun 2014 19:05:58 +0000 (19:05 +0000)]
Use std::error_code instead of llvm::error_code.
This is an update for a llvm api change.
llvm-svn: 210689
Rafael Espindola [Wed, 11 Jun 2014 19:05:55 +0000 (19:05 +0000)]
Use std::error_code instead of llvm::error_code.
This is an update for a llvm api change.
llvm-svn: 210688
Rafael Espindola [Wed, 11 Jun 2014 19:05:50 +0000 (19:05 +0000)]
Use std::error_code instead of llvm::error_code.
The idea of this patch is to turn llvm/Support/system_error.h into a
transitional header that just brings in the erorr_code api to the llvm
namespace. I will remove it shortly afterwards.
The cases where the general idea needed some tweaking:
* std::errc is a namespace in msvc, so we cannot use "using std::errc". I could
add an #ifdef, but there were not that many uses, so I just added std:: to
them in this patch.
* Template specialization had to be moved to the std namespace in this
patch set already.
* The msvc implementation of default_error_condition doesn't seem to
provide the same transformations as we need. Not too surprising since
the standard doesn't actually say what "equivalent" means. I fixed the
problem by keeping our old mapping and using it at error_code
construction time.
Despite these shortcomings I think this is still a good thing. Some reasons:
* The different implementations of system_error might improve over time.
* It removes 925 lines of code from llvm already.
* It removes 6313 bytes from the text segment of the clang binary when
it is built with gcc and 2816 bytes when building with clang and
libstdc++.
llvm-svn: 210687
Kaelyn Takata [Wed, 11 Jun 2014 18:33:46 +0000 (18:33 +0000)]
Add some comments to TypoCorrectionConsumer.
Also move the constructor for NamespaceSpecifierSet out of line to
improve the class' readability. I meant to do these two things while
cleaning up the previous TypoCorrectionConsumer changes and have them
folded into those changes.
No functionality changed.
llvm-svn: 210686
Chad Rosier [Wed, 11 Jun 2014 18:28:45 +0000 (18:28 +0000)]
[Reassociate] FileCheckize and cleanup a few testcases. No functional change
intended.
llvm-svn: 210685
Chad Rosier [Wed, 11 Jun 2014 18:26:29 +0000 (18:26 +0000)]
Fix assert comments in Instruction.cpp.
llvm-svn: 210684
Matt Arsenault [Wed, 11 Jun 2014 18:11:34 +0000 (18:11 +0000)]
R600/SI: Update place using old subtarget predicate
llvm-svn: 210683
Ed Maste [Wed, 11 Jun 2014 18:10:41 +0000 (18:10 +0000)]
Restore select limitations comment for non-Apple platforms
llvm-svn: 210682
Todd Fiala [Wed, 11 Jun 2014 18:09:01 +0000 (18:09 +0000)]
Add gdb-remote P register write test.
Added test that attempts to write a value to each general
purpose register that is bit-flipped from the initial read
value. It then reads it back to see if it takes.
Right now I just assert that at least one register bit flip
write succeeds. I added a note that on the MacOSX x86_64
debugserver case, the only writes that succeed from the GPR
set are rax thru rdx, rdi, rsi and rbp. The failures are E32
failure-at-write-attempt issues on the debugserver end.
I'll revisit this after implementing in the llgs linux-x86_64 branch.
The packet log looks good but I may have a subtle mistake in the code.
llvm-svn: 210681
Matt Arsenault [Wed, 11 Jun 2014 18:08:54 +0000 (18:08 +0000)]
R600/SI: Add common 64-bit LDS atomics
llvm-svn: 210680
Matt Arsenault [Wed, 11 Jun 2014 18:08:50 +0000 (18:08 +0000)]
R600/SI: Add instruction definitions for 64-bit LDS atomics
llvm-svn: 210679
Matt Arsenault [Wed, 11 Jun 2014 18:08:48 +0000 (18:08 +0000)]
R600/SI: Add 32-bit LDS atomic cmpxchg
llvm-svn: 210678
Matt Arsenault [Wed, 11 Jun 2014 18:08:45 +0000 (18:08 +0000)]
R600/SI: Use LDS atomic inc / dec
llvm-svn: 210677
Matt Arsenault [Wed, 11 Jun 2014 18:08:42 +0000 (18:08 +0000)]
R600/SI: Add other LDS atomic operations
llvm-svn: 210676
Matt Arsenault [Wed, 11 Jun 2014 18:08:39 +0000 (18:08 +0000)]
R600/SI: Add instruction definitions for more LDS ops
llvm-svn: 210675
Matt Arsenault [Wed, 11 Jun 2014 18:08:37 +0000 (18:08 +0000)]
R600/SI: Fix backwards names for local atomic instructions.
The manual lists them as *_RTN_U32, not *_U32_RTN, which is more
consistent with how every other sized instruction is named.
llvm-svn: 210674
Matt Arsenault [Wed, 11 Jun 2014 18:08:34 +0000 (18:08 +0000)]
R600/SI: Refactor local atomics.
Use patterns that will also match the immediate offset to
match the normal read / writes.
llvm-svn: 210673
Kaelyn Takata [Wed, 11 Jun 2014 18:07:08 +0000 (18:07 +0000)]
Sink SpecifierInfo into the only class that uses it.
SpecifierInfo is not used outside of NamespaceSpecifierSet except
indirectly through NamespaceSpecifierSet's iterator, so clean up the
code a bit by moving SpecifierInfo into NamespaceSpecifierSet. Also drop
SpecifierInfo's trivial yet verbose constructor since brace
initiialization is sufficient in the only two places the constructor was
being explicitly called.
No functionality changed.
llvm-svn: 210672
Kaelyn Takata [Wed, 11 Jun 2014 18:07:05 +0000 (18:07 +0000)]
Make a couple of helper classes private to TypoCorrectionConsumer.
The SpecifierInfo and NamespaceSpecifierSet are now only used by
TypoCorrectionConsumer, so treat them as the implementation details of
TypoCorrectionConsumer that they are. Also make NamespaceSpecifierSet's
method names more style guide compliant.
No functionality changed.
llvm-svn: 210671
Kaelyn Takata [Wed, 11 Jun 2014 18:07:03 +0000 (18:07 +0000)]
Clean up the TypoCorrectionConsumer's interface a bit.
No functionality changed.
llvm-svn: 210670
Kaelyn Takata [Wed, 11 Jun 2014 18:07:01 +0000 (18:07 +0000)]
Move TypoCorrection filtering into the TypoCorrectionConsumer and make it incremental.
The only external/visible functional change that fell out of this
refactoring is that there was one less case where the typo caching
and/or counting didn't work properly. The result is that a test case
had to be moved from typo-correction.cpp to typo-correction-pt2.cpp
to avoid the hard-coded limit on per file/TU typo correction attempts.
llvm-svn: 210669
Kaelyn Takata [Wed, 11 Jun 2014 18:06:56 +0000 (18:06 +0000)]
Move a few classes to be before TypoCorrectionConsumer.
This is in preparation for moving TypoCorrection filtering
into the TypoCorrectionConsumer, to separate out some of the purely
mechanical churn. It also makes some of the method names in
NamespaceSpecifierSet be more style guide compliant.
No functionality changed.
llvm-svn: 210668
Eli Bendersky [Wed, 11 Jun 2014 17:56:26 +0000 (17:56 +0000)]
Add loop unroll pragma support
http://reviews.llvm.org/D4089
Patch by Mark Heffernan.
llvm-svn: 210667
Matt Arsenault [Wed, 11 Jun 2014 17:50:44 +0000 (17:50 +0000)]
R600/SI: Use v_cvt_f32_ubyte* instructions
This eliminates extra extract instructions when loading an i8 vector to
a float vector.
llvm-svn: 210666
David Blaikie [Wed, 11 Jun 2014 17:50:14 +0000 (17:50 +0000)]
SmallVectorTest: Make the deleted member functions private to help MSVC users.
llvm-svn: 210665
Matt Arsenault [Wed, 11 Jun 2014 17:40:32 +0000 (17:40 +0000)]
R600/SI: Fix selection failure on scalar_to_vector
There seem to be only 2 places that produce these,
and it's kind of tricky to hit them.
Also fixes failure to bitcast between i64 and v2f32,
although this for some reason wasn't actually broken in the
simple bitcast testcase, but did in the scalar_to_vector one.
llvm-svn: 210664
Tim Northover [Wed, 11 Jun 2014 17:04:08 +0000 (17:04 +0000)]
X86: add stringy name for X86ISD::LCMPXCHG16_DAG
I don't know what "target specific node #383" is, and I don't want to
have to.
llvm-svn: 210663
Eric Christopher [Wed, 11 Jun 2014 16:59:33 +0000 (16:59 +0000)]
Revert r210613 to conform to coding standards.
Thanks Duncan for noticing.
llvm-svn: 210662
Marshall Clow [Wed, 11 Jun 2014 16:54:09 +0000 (16:54 +0000)]
Add a FAQ section, with a question about why the std::exception class destructors live in libc++abi
llvm-svn: 210661
Fariborz Jahanian [Wed, 11 Jun 2014 16:52:44 +0000 (16:52 +0000)]
Objective-C. Patch to handle bridge attribute warnings
correctly when both NSAttributedString and
NSMutableAttributedString are specified on the same
CFStruct via different typedefs. // rdar://
17238954
llvm-svn: 210660
Marshall Clow [Wed, 11 Jun 2014 16:44:55 +0000 (16:44 +0000)]
Implement string_view from the library fundamentals TS (n4023). Also works in C++11 and 03, with reduced functionality (mostly in the area of constexpr)
llvm-svn: 210659
Bill Schmidt [Wed, 11 Jun 2014 15:48:46 +0000 (15:48 +0000)]
[PPC64LE] Fix vec_sld and vec_vsldoi for little endian
The vec_sld and vec_vsldoi interfaces perform a left-shift on vector
arguments for both big and little endian. However, because they rely
on the vec_perm interface which is endian-dependent, the permutation
vector needs to be reversed for LE to get the proper shift direction.
I've added some extra testing for these interfaces for LE in the
builtins-ppc-altivec.c.
llvm-svn: 210657
Daniel Sanders [Wed, 11 Jun 2014 15:48:00 +0000 (15:48 +0000)]
[mips][mips64r6] Improve tests affected by the changes to multiplies and divides
Summary:
MIPS32r6/MIPS64r6 support has not been added yet.
inlineasm-cnstrnt-reg.ll:
Explicitly specify the CPU since it will not work on MIPS32r6/MIPS64r6
when -integrated-as is the default. We can't change the mnemonic since the
LO register is an implicit def of mtlo and MIPS32r6/MIPS64r6 has no
instructions that use LO.
2008-08-01-AsmInline.ll:
Explicitly specify the CPU since MIPS32r6/MIPS64r6 will correctly emit
different code and this is a regression test.
mips64instrs.ll and mips64muldiv.ll
Check registers and the way the multiply is used in m1
divrem.ll
Check registers and use multiple filecheck prefixes to limit redundancy
Reviewers: vmedic, jkolek, zoran.jovanovic, matheusalmeida
Reviewed By: matheusalmeida
Subscribers: matheusalmeida
Differential Revision: http://reviews.llvm.org/D3894
llvm-svn: 210656
Evgeniy Stepanov [Wed, 11 Jun 2014 15:11:26 +0000 (15:11 +0000)]
[asan] Fix mmap-ed coverage with coverage=0.
llvm-svn: 210655
Matheus Almeida [Wed, 11 Jun 2014 15:05:56 +0000 (15:05 +0000)]
[mips] Implement jr.hb and jalr.hb (Jump Register and Jump and Link Register with Hazard Barrier).
Summary: These instructions are available in ISAs >= mips32/mips64. For mips32r6/mips64r6, jr.hb has a new encoding format.
Reviewers: dsanders
Reviewed By: dsanders
Differential Revision: http://reviews.llvm.org/D4019
llvm-svn: 210654
Cameron McInally [Wed, 11 Jun 2014 12:54:45 +0000 (12:54 +0000)]
Add AVX512 masked leadz instrinsic support.
llvm-svn: 210652
Daniel Jasper [Wed, 11 Jun 2014 12:18:24 +0000 (12:18 +0000)]
clang-tidy: [use-override] Remove 'override' if 'final' is also present.
Also, make warning more precise by distinguishing different cases.
llvm-svn: 210651
Evgeniy Stepanov [Wed, 11 Jun 2014 10:37:06 +0000 (10:37 +0000)]
[asan] Run Android test with asanwrapper.
This better replicates the target environment.
llvm-svn: 210650
Evgeniy Stepanov [Wed, 11 Jun 2014 10:11:51 +0000 (10:11 +0000)]
[asan] Exclude non-executable mappings from coverage.
llvm-svn: 210649
Simon Atanasyan [Wed, 11 Jun 2014 10:08:08 +0000 (10:08 +0000)]
[Mips] Fix the bug -- symbol referred by the R_MIPS_TLS_GD relocation
does not get an entry in the dynamic symbol table.
llvm-svn: 210648
Evgeniy Stepanov [Wed, 11 Jun 2014 09:23:25 +0000 (09:23 +0000)]
[asan] One more coverage test.
Test that coverage is dumped correctly (and completely) even when the processed
gets SIGKILL-ed.
llvm-svn: 210647
Evgeniy Stepanov [Wed, 11 Jun 2014 09:09:09 +0000 (09:09 +0000)]
[asan] Fix coverage-android test to avoid name conflict.
And re-enable it.
llvm-svn: 210646
Evgeniy Stepanov [Wed, 11 Jun 2014 08:46:45 +0000 (08:46 +0000)]
Improve the test for inlining of __no_debug__ functions.
llvm-svn: 210645
Andrea Di Biagio [Wed, 11 Jun 2014 07:57:50 +0000 (07:57 +0000)]
[X86] Refactor the logic to select horizontal adds/subs to a helper function.
This patch moves part of the logic implemented by the target specific
combine rules added at r210477 to a separate helper function.
This should make easier to add more rules for matching AVX/AVX2 horizontal
adds/subs.
This patch also fixes a problem caused by a wrong check performed on indices
of extract_vector_elt dag nodes in input to the scalar adds/subs.
New tests have been added to verify that we correctly check indices of
extract_vector_elt dag nodes when selecting a horizontal operation.
llvm-svn: 210644
Daniel Jasper [Wed, 11 Jun 2014 07:35:16 +0000 (07:35 +0000)]
clang-format: Fix pointer/reference detection after decltype.
Before:
[](const decltype(*a) & value) {}
After:
[](const decltype(*a)& value) {}
llvm-svn: 210643
David Majnemer [Wed, 11 Jun 2014 07:08:37 +0000 (07:08 +0000)]
MS ABI: Only class templates get complex nullptr pointer-to-member-functions
r210637 regressed CodeGenCXX/mangle-ms-templates-memptrs.cpp because it
did not believe that there is a distinction between class templates and
function templates.
Sadly, there is. Function templates should behave in a compatible
manner with MSVC.
llvm-svn: 210642
Jiangning Liu [Wed, 11 Jun 2014 07:04:37 +0000 (07:04 +0000)]
Create macro INITIALIZE_TM_PASS.
Pass initialization requires to initialize TargetMachine for back-end
specific passes. This commit creates a new macro INITIALIZE_TM_PASS to
simplify this kind of initialization.
llvm-svn: 210641
Jiangning Liu [Wed, 11 Jun 2014 06:44:53 +0000 (06:44 +0000)]
Global merge for global symbols.
This commit is to improve global merge pass and support global symbol merge.
The global symbol merge is not enabled by default. For aarch64, we need some
more back-end fix to make it really benifit ADRP CSE.
llvm-svn: 210640
Jiangning Liu [Wed, 11 Jun 2014 06:35:26 +0000 (06:35 +0000)]
Rename global-merge to enable-global-merge.
llvm-svn: 210639
Craig Topper [Wed, 11 Jun 2014 05:35:56 +0000 (05:35 +0000)]
Convert StringMapEntry::Create to use StringRef instead of start/end pointers. Simpliies all in tree call sites. No functional change.
llvm-svn: 210638
David Majnemer [Wed, 11 Jun 2014 04:55:08 +0000 (04:55 +0000)]
MS ABI: Mangle null pointer-to-member-functions compatibly
Summary:
Previously, we would mangle nullptr pointer-to-member-functions in class
templates with a mangling we invented because contemporary versions of
MSVC would crash when trying to compile such code.
However, VS "14" can successfully compile these sorts of template
instantiations. This commit updates our mangling to be compatible with
theirs.
Reviewers: rnk
Reviewed By: rnk
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D4059
llvm-svn: 210637
Rafael Espindola [Wed, 11 Jun 2014 04:41:37 +0000 (04:41 +0000)]
Try to fix the msvc build.
llvm-svn: 210636
Rafael Espindola [Wed, 11 Jun 2014 04:34:41 +0000 (04:34 +0000)]
Uses generic_category instead of system_category.
Some c++ libraries (libstdc++ at least) don't seem to map to the generic
category in in the system_category's default_error_condition.
llvm-svn: 210635
Saleem Abdulrasool [Wed, 11 Jun 2014 04:19:25 +0000 (04:19 +0000)]
MC: add enumeration of WinEH data encoding
Most Windows platforms use auxiliary data for unwinding. This information is
stored in the .pdata section. The encoding format for the data differs between
architectures and Windows variants. Windows MIPS and Alpha use identical
formats; Alpha64 is the same with different widths. Windows x86_64 and Itanium
share the representation. All Windows CE entries are identical irrespective of
the architecture. ARMv7 (Windows [NT] on ARM) has its own format.
This enumeration will become the differentiator once the windows EH emission
infrastructure is generalised, allowing us to emit the necessary unwinding
information for Windows on ARM.
llvm-svn: 210634
Saleem Abdulrasool [Wed, 11 Jun 2014 04:19:19 +0000 (04:19 +0000)]
MC: clang-format MCAsmInfo
Apply clang-format over the header. Reformat the docs to current LLVM style.
NFC.
llvm-svn: 210633
David Majnemer [Wed, 11 Jun 2014 04:08:55 +0000 (04:08 +0000)]
CodeGen: Correct linkage of thread_local for OS X
The backing store of thread local variables is internal for OS X and all
accesses must go through the thread wrapper.
However, individual TUs may have inlined through the thread wrapper.
To fix this, give the thread wrapper functions WeakAnyLinkage. This
prevents them from getting inlined into call-sites.
This fixes PR19989.
llvm-svn: 210632
Rafael Espindola [Wed, 11 Jun 2014 03:58:34 +0000 (03:58 +0000)]
Remove windows_error.
MSVC doesn't seem to provide any is_error_code_enum enumeration for the
windows errors.
Fortunately very few places in llvm have to handle raw windows errors, so
we can just construct the corresponding error_code directly.
llvm-svn: 210631
Rafael Espindola [Wed, 11 Jun 2014 03:49:13 +0000 (03:49 +0000)]
There is no posix_category in std, use generic_category.
llvm-svn: 210630
Rafael Espindola [Wed, 11 Jun 2014 03:35:50 +0000 (03:35 +0000)]
Missed these from the previous commit.
llvm-svn: 210629
Matt Arsenault [Wed, 11 Jun 2014 03:30:06 +0000 (03:30 +0000)]
Use cast instead of assert + dyn_cast
llvm-svn: 210628
Matt Arsenault [Wed, 11 Jun 2014 03:29:54 +0000 (03:29 +0000)]
R600: Add helper functions.
Extract these from some of my other patches, since this
is the only thing really making them dependent on each other.
llvm-svn: 210627
Rafael Espindola [Wed, 11 Jun 2014 03:28:13 +0000 (03:28 +0000)]
Inherit from error_category instead of _do_message.
There is no std::_do_message and these classes override message anyway.
llvm-svn: 210626
Kevin Qin [Wed, 11 Jun 2014 01:42:16 +0000 (01:42 +0000)]
[AArch64] Add default features for CPUs on AArch64 target.
For ARM target, we can use CRYPTO and CRC features if we select
cortex-a57 by '-mcpu', but for AArch64 target, it doesn't work
unless adding with '-mfpu=crypto-neon-fp-armv8'. To keep consistency
between front-end and back-end and get end-users more easier to use,
we'd better add default feature for CPUs on AArch64 target as well.
llvm-svn: 210625
Nick Kledzik [Wed, 11 Jun 2014 01:30:55 +0000 (01:30 +0000)]
[mach-o] fix use of resolver functions to not cause duplicate sections.
The previous commit uncovered a bug in the mach-o writer whereby two __text
sections were created. But the test case did not catch that. So I updated
the test case to run the linker a second time, reading the output of the
first pass.
llvm-svn: 210624
Rafael Espindola [Wed, 11 Jun 2014 01:22:20 +0000 (01:22 +0000)]
Use an enum class.
llvm-svn: 210623
Saleem Abdulrasool [Wed, 11 Jun 2014 01:19:03 +0000 (01:19 +0000)]
CodeGen: refactor DwarfException
DwarfException served as a base class for exception handling directive emission.
However, this is also used by other exception models (e.g. Win64EH). Rename
this class to EHStreamer and split it out of DwarfException.h. NFC.
Use the opportunity to fix up some of the documentation comments to match
current LLVM style. Also rename some functions to conform better with current
LLVM coding style.
llvm-svn: 210622
Rui Ueyama [Wed, 11 Jun 2014 01:18:05 +0000 (01:18 +0000)]
Fix pecoff/export.test on Windows.
llvm-svn: 210621
Rafael Espindola [Wed, 11 Jun 2014 01:09:09 +0000 (01:09 +0000)]
Use an enum class.
llvm-svn: 210620
Eric Christopher [Wed, 11 Jun 2014 00:53:17 +0000 (00:53 +0000)]
Remove duplicate copy of InstrItineraryData from the TargetMachine,
it's already on the subtarget.
llvm-svn: 210619
David Blaikie [Wed, 11 Jun 2014 00:50:07 +0000 (00:50 +0000)]
Ensure test/Misc/win32-macho.c doesn't try to write to the source tree
llvm-svn: 210618
Richard Smith [Wed, 11 Jun 2014 00:49:52 +0000 (00:49 +0000)]
PR19993: don't assert/crash if a static data member is initialized by '= delete;'.
llvm-svn: 210617
Eric Christopher [Wed, 11 Jun 2014 00:46:34 +0000 (00:46 +0000)]
Move to a private function to initialize the subtarget dependencies
so that we can use initializer lists for the AArch64 Subtarget.
llvm-svn: 210616
Richard Smith [Wed, 11 Jun 2014 00:31:00 +0000 (00:31 +0000)]
PR19996: don't crash if -Wuninitialized sees a c++1y lambda init-capture.
llvm-svn: 210615
Eric Christopher [Wed, 11 Jun 2014 00:25:19 +0000 (00:25 +0000)]
Move to a private function to initialize the subtarget dependencies
so that we can use initializer lists for the X86Subtarget.
llvm-svn: 210614
Eric Christopher [Wed, 11 Jun 2014 00:25:16 +0000 (00:25 +0000)]
Sort includes.
llvm-svn: 210613
Nick Kledzik [Wed, 11 Jun 2014 00:24:16 +0000 (00:24 +0000)]
[mach-o] add support for parsing resolver functions
llvm-svn: 210612
Reid Kleckner [Wed, 11 Jun 2014 00:01:28 +0000 (00:01 +0000)]
Allow lookup into dependent bases in more places under -fms-compatibility
We currently allow unqualified lookup for instance methods but not
static methods because we can't recover with a semantic 'this->'
insertion.
ATL headers have static methods that do unqualified lookup into
dependent base classes. The pattern looks like:
template <typename T> struct Foo : T {
static int *getBarFromT() { return Bar; }
};
Now we recover as if the user had written:
template <typename T> struct Foo : T {
static int *getBarFromT() { return Foo::Bar; }
};
... which will eventually look up Bar in T at instantiation time.
Now we emit a diagnostic in both cases, and delay lookup in other
contexts where 'this' is available and refers to a class with dependent
bases.
Reviewed by: rsmith
Differential Revision: http://reviews.llvm.org/D4079
llvm-svn: 210611
Juergen Ributzka [Tue, 10 Jun 2014 23:52:44 +0000 (23:52 +0000)]
[FastISel][X86] Extend support for {s|u}{add|sub|mul}.with.overflow intrinsics.
llvm-svn: 210610
Richard Smith [Tue, 10 Jun 2014 23:43:44 +0000 (23:43 +0000)]
Teach __alignof__ to look through arrays before performing the
preferred-alignment transformations. Corrects alignof(T[]) to return
alignof(T) in all cases, as required by relevant standards.
llvm-svn: 210609
Richard Smith [Tue, 10 Jun 2014 23:34:28 +0000 (23:34 +0000)]
Related to PR19992: when the GNU alignof-expression extension is applied to an
expression of array-of-unknown-bound type, don't try to complete the array
bound, and return the alignment of the element type rather than 1.
llvm-svn: 210608
Reid Kleckner [Tue, 10 Jun 2014 23:29:48 +0000 (23:29 +0000)]
Recover from missing typenames on template args for MSVC compatibility
While matching a non-type template argument against a known template
type parameter we now modify the AST's TemplateArgumentLoc to assume the
user wrote typename. Under -fms-compatibility, we downgrade our
diagnostic from an error to an extwarn.
Reviewed by: rsmith
Differential Revision: http://reviews.llvm.org/D4049
llvm-svn: 210607
Eric Christopher [Tue, 10 Jun 2014 23:26:47 +0000 (23:26 +0000)]
Use unique_ptr for X86Subtarget pointer members.
llvm-svn: 210606
Eric Christopher [Tue, 10 Jun 2014 23:26:45 +0000 (23:26 +0000)]
Move AArch64TargetLowering to AArch64Subtarget.
This currently necessitates a TargetMachine for the TargetLowering
constructor and TLOF.
llvm-svn: 210605
Zachary Turner [Tue, 10 Jun 2014 23:17:11 +0000 (23:17 +0000)]
Revert "Clang changes to support LLVM removal of runtime multithreading"
This reverts r210601.
llvm-svn: 210604
Zachary Turner [Tue, 10 Jun 2014 23:15:43 +0000 (23:15 +0000)]
Revert "Remove support for runtime multi-threading."
This reverts revision r210600.
llvm-svn: 210603
Todd Fiala [Tue, 10 Jun 2014 23:12:28 +0000 (23:12 +0000)]
Fixed gdb-remote M test length parameter, added overwrite to test.
Initial check-in provided a nibble count instead of byte count for
the memory to write. Fixed that.
Enhanced test to check for overwrite past the expected range of
writing to verify the correct amount is written.
llvm-svn: 210602
Zachary Turner [Tue, 10 Jun 2014 23:02:10 +0000 (23:02 +0000)]
Clang changes to support LLVM removal of runtime multithreading
support (llvm revision r210600).
Reviewers: rnk
Differential Revision: http://reviews.llvm.org/D4077
llvm-svn: 210601
Zachary Turner [Tue, 10 Jun 2014 23:01:20 +0000 (23:01 +0000)]
Remove support for runtime multi-threading.
This patch removes the functions llvm_start_multithreaded() and
llvm_stop_multithreaded(), and changes llvm_is_multithreaded()
to return a constant value based on the value of the compile-time
definition LLVM_ENABLE_THREADS.
Previously, it was possible to have compile-time support for
threads on, and runtime support for threads off, in which case
certain mutexes were not allocated or ever acquired. Now, if the
build is created with threads enabled, mutexes are always acquired.
A test before/after patch of compiling a very large TU showed no
noticeable performance impact of this change.
Reviewers: rnk
Differential Revision: http://reviews.llvm.org/D4076
llvm-svn: 210600
Eric Christopher [Tue, 10 Jun 2014 22:57:25 +0000 (22:57 +0000)]
Move AArch64InstrInfo to AArch64Subtarget.
llvm-svn: 210599
Eric Christopher [Tue, 10 Jun 2014 22:57:21 +0000 (22:57 +0000)]
Remove a method that was just replacing direct access to a member.
llvm-svn: 210598
Hans Wennborg [Tue, 10 Jun 2014 22:51:58 +0000 (22:51 +0000)]
lit: warn when passed invalid pathname
It would previously say things like
warning: input 'test/Frontend/foo.c' contained no tests
and have the user pull their hair trying to figure out what's wrong with that
file. This patch changes the message to the much clearer:
warning: no such file or directory: 'test/Frontend/foo.c'
Differential Revision: http://reviews.llvm.org/D4097
llvm-svn: 210597
Eric Christopher [Tue, 10 Jun 2014 22:34:31 +0000 (22:34 +0000)]
Remove the use of TargetMachine from X86InstrInfo.
llvm-svn: 210596
Eric Christopher [Tue, 10 Jun 2014 22:34:28 +0000 (22:34 +0000)]
Move X86RegisterInfo away from using the TargetMachine and only
using the subtarget.
llvm-svn: 210595
Todd Fiala [Tue, 10 Jun 2014 22:15:56 +0000 (22:15 +0000)]
Added gdb-remote test to verify M memory write.
llvm-svn: 210594
Reid Kleckner [Tue, 10 Jun 2014 21:35:24 +0000 (21:35 +0000)]
*Really* fix DOS newlines introduced in r210330
r210369 didn't quite catch all of them.
llvm-svn: 210593
Ed Maste [Tue, 10 Jun 2014 21:33:43 +0000 (21:33 +0000)]
Remove duplicated code
We preivously had two copies of ::BytesAvailable with only trivial
differences between them, and fixes have been applied to only one of
them.
Instead of duplicating the whole function, hide the FD_SET differences
behind a macro. This leaves only one small __APPLE__-specific #if
block, and fixes ^C on non-__APPLE__ platforms.
llvm-svn: 210592
Rafael Espindola [Tue, 10 Jun 2014 21:26:47 +0000 (21:26 +0000)]
Mark a few functions noexcept.
This reduces the difference between std::error_code and llvm::error_code.
llvm-svn: 210591
Rafael Espindola [Tue, 10 Jun 2014 21:26:18 +0000 (21:26 +0000)]
Mark a few functions noexcept.
llvm-svn: 210590