Gabor Buella [Fri, 13 Apr 2018 07:35:08 +0000 (07:35 +0000)]
[X86] Introduce cldemote instruction
Hint to hardware to move the cache line containing the
address to a more distant level of the cache without
writing back to memory.
Reviewers: craig.topper, zvi
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D45256
llvm-svn: 329992
Martin Storsjo [Fri, 13 Apr 2018 06:38:02 +0000 (06:38 +0000)]
[Support] Fix building for Windows on ARM
The commit in SVN r310001 that added support for this actually didn't
use the right struct field for the frame pointer - for ARM, there is
no register named Fp in the CONTEXT struct. On Windows, the R11
register is used as frame pointer.
Differential Revision: https://reviews.llvm.org/D45590
llvm-svn: 329991
Craig Topper [Fri, 13 Apr 2018 06:07:18 +0000 (06:07 +0000)]
[X86] Remove the pmuldq/pmuldq intrinsics and replace with native IR.
This completes the work started in r329604 and r329605 when we changed clang to no longer use the intrinsics.
We lost some InstCombine SimplifyDemandedBit optimizations through this change as we aren't able to fold 'and', bitcast, shuffle very well.
llvm-svn: 329990
Dean Michael Berris [Fri, 13 Apr 2018 05:59:57 +0000 (05:59 +0000)]
[XRay][clang] Make -fxray-modes= additive
Summary:
This allows us to do the following:
clang -fxray-modes=none ... -fxray-modes=xray-basic
It's important to be able to do this in cases where we'd like to
specialise the configuration for the invocation of the compiler, in
various scripting environments.
This is related to llvm.org/PR37066, a follow-up to D45474.
Reviewers: eizan, kpw, pelikan
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D45610
llvm-svn: 329989
Yunlian Jiang [Fri, 13 Apr 2018 05:03:28 +0000 (05:03 +0000)]
Enable debug fission for thinLTO linked via gold-plugin
Summary: This enables debug fission on implicit ThinLTO when linked with gold. It will put the .dwo files in a directory specified by user.
Reviewers: tejohnson, pcc, dblaikie
Reviewed By: pcc
Subscribers: JDevlieghere, mehdi_amini, inglorion
Differential Revision: https://reviews.llvm.org/D44792
llvm-svn: 329988
Xin Tong [Fri, 13 Apr 2018 04:35:38 +0000 (04:35 +0000)]
[CallSiteSplit] Fix comment. NFC
llvm-svn: 329987
Erik Pilkington [Fri, 13 Apr 2018 02:53:26 +0000 (02:53 +0000)]
Fix another bot failure from r329951.
llvm-svn: 329986
Dean Michael Berris [Fri, 13 Apr 2018 02:31:58 +0000 (02:31 +0000)]
[XRay][clang] Add flag to choose instrumentation bundles
Summary:
This change addresses http://llvm.org/PR36926 by allowing users to pick
which instrumentation bundles to use, when instrumenting with XRay. In
particular, the flag `-fxray-instrumentation-bundle=` has four valid
values:
- `all`: the default, emits all instrumentation kinds
- `none`: equivalent to -fnoxray-instrument
- `function`: emits the entry/exit instrumentation
- `custom`: emits the custom event instrumentation
These can be combined either as comma-separated values, or as
repeated flag values.
Reviewers: echristo, kpw, eizan, pelikan
Reviewed By: pelikan
Subscribers: mgorny, cfe-commits
Differential Revision: https://reviews.llvm.org/D44970
llvm-svn: 329985
Petr Hosek [Fri, 13 Apr 2018 02:02:13 +0000 (02:02 +0000)]
[CMake][Fuchsia] Don't specify libc++ ABI version for Fuchsia
This is now set automatically in libc++ config header.
Differential Revision: https://reviews.llvm.org/D45511
llvm-svn: 329984
Petr Hosek [Fri, 13 Apr 2018 01:59:15 +0000 (01:59 +0000)]
[CMake] Set the default ABI version for Fuchsia in CMake as well
This is neeeded since the CMake value is used for the SOVERSION.
Differential Revision: https://reviews.llvm.org/D45529
llvm-svn: 329983
Kuba Mracek [Fri, 13 Apr 2018 01:05:29 +0000 (01:05 +0000)]
[tsan] Add interceptors for objc_sync_enter and objc_sync_exit
Objective-C's @synchronize synchronization primitive uses calls to objc_sync_enter and objc_sync_exit runtime functions. In most cases, they end up just calling pthread_mutex_lock/pthread_mutex_unlock, but there are some cases where the synchronization from pthread_mutex_lock/pthread_mutex_unlock interceptors isn't enough. Let's add explicit interceptors for objc_sync_enter and objc_sync_exit to handle all cases.
Differential Revision: https://reviews.llvm.org/D45487
llvm-svn: 329982
Tony Tye [Fri, 13 Apr 2018 01:01:27 +0000 (01:01 +0000)]
[AMDGPU] Update relocation record description
Document which relocation records are static and dynamic.
Differential Revision: https://reviews.llvm.org/D45587
llvm-svn: 329981
Kuba Mracek [Fri, 13 Apr 2018 00:29:24 +0000 (00:29 +0000)]
[asan] Reduce flakiness in stack-overflow detection
IsStackOverflow only treats accesses within 512 bytes of SP as stack-overflow. This should really be the size of a page instead.
The scariness_score_test.cc triggers stack overflow with frames that are even larger than a page, which can also trigger a fault that will not be recognized as stack-overflow. Let's just use smaller frames.
llvm-svn: 329980
Stephan T. Lavavej [Fri, 13 Apr 2018 00:07:48 +0000 (00:07 +0000)]
[libcxx] [test] Use TEST_IGNORE_NODISCARD.
Fixes D45595.
llvm-svn: 329979
Stephan T. Lavavej [Thu, 12 Apr 2018 23:56:22 +0000 (23:56 +0000)]
[libcxx] [test] Fix whitespace, NFC.
test/std almost always uses spaces; now it is entirely tab-free.
llvm-svn: 329978
Stephan T. Lavavej [Thu, 12 Apr 2018 23:56:17 +0000 (23:56 +0000)]
[libcxx] [test] Use TEST_COMPILER_C1XX.
Also TEST_COMPILER_CLANG in one place. (More could be changed.)
llvm-svn: 329977
Stephan T. Lavavej [Thu, 12 Apr 2018 23:56:14 +0000 (23:56 +0000)]
[libcxx] [test] Silence MSVC warning C4146.
This test code triggers the MSVC warning:
"unary minus operator applied to unsigned type, result still unsigned"
Although it would be possible to change the test code to avoid
this warning, I have chosen to simply silence it.
Fixes D45594.
llvm-svn: 329976
Stephan T. Lavavej [Thu, 12 Apr 2018 23:56:12 +0000 (23:56 +0000)]
[libcxx] [test] Fix nodiscard warnings.
MSVC's STL has marked to_bytes/from_bytes as nodiscard.
Fixes D45595.
llvm-svn: 329975
Stephan T. Lavavej [Thu, 12 Apr 2018 23:56:10 +0000 (23:56 +0000)]
[libcxx] [test] Avoid unary_function.
Replace unary_function inheritance (which was never required,
even in C++98) with argument_type and result_type typedefs.
This increases portability, as unary_function was removed in C++17
and MSVC has implemented that removal.
Fixes D45596.
llvm-svn: 329974
Stephan T. Lavavej [Thu, 12 Apr 2018 23:56:07 +0000 (23:56 +0000)]
[libcxx] [test] Avoid MSVC truncation warnings.
MSVC emits "warning C4244: 'initializing': conversion from 'int'
to 'short', possible loss of data" when it sees pair<Whatever, short>
constructed from (whatever, 4), because int is being truncated to
short within pair's constructor. (The compiler doesn't take into
account the fact that 4 is a literal at the callsite; it generates
this warning when the constructor is instantiated, because it might
be called with a runtime-valued int that would actually truncate.)
Instead of static_cast<short>, we can simply change short to int
in these tests, without affecting the pair operations that they're
trying to test: move assignment, convert copy construction, and
convert move construction.
Fixes D45016.
llvm-svn: 329973
Aaron Smith [Thu, 12 Apr 2018 23:45:15 +0000 (23:45 +0000)]
[lit] Remove duplicate to_string method
There are two versions of to_string used by TestRunner.py. The one defined
in TestRunner.py and the one defined in utils/lit/lit/util.py. The util.py
version is superior to the TestRunner.py version.
This change removes the duplicate to_string in TestRunner.py in favor of
always using the version from util.py. Beside removing duplicate code, this
makes it easier to debug TestRunner.py since only one version of to_string
is used.
Patch by Stella Stamenova!
llvm-svn: 329972
Aaron Smith [Thu, 12 Apr 2018 23:29:49 +0000 (23:29 +0000)]
[docs] Add LLDB_TEST_USE_CUSTOM_C_COMPILER and LLDB_TEST_USE_CUSTOM_CXX_COMPILER to the build and test hmtl pages
llvm-svn: 329971
Simon Pilgrim [Thu, 12 Apr 2018 22:57:34 +0000 (22:57 +0000)]
[X86] Remove unused MoveLoadStoreItins/ShiftOpndItins schedule class wrappers.
Was being used to move around empty/unused itineraries...
llvm-svn: 329970
Erik Pilkington [Thu, 12 Apr 2018 22:54:47 +0000 (22:54 +0000)]
Free a pointer, fix a bot.
llvm-svn: 329969
Eli Friedman [Thu, 12 Apr 2018 22:50:50 +0000 (22:50 +0000)]
Fix test failure caused by r329965.
"-mllvm" options get parsed slightly earlier, and -arm-restrict-it is
only available if the ARM target is compiled in. Invoke "clang -cc1"
directly to avoid the issue.
llvm-svn: 329968
Simon Pilgrim [Thu, 12 Apr 2018 22:44:47 +0000 (22:44 +0000)]
[X86] Remove x86 InstrItinClass entries (PR37093)
This removes the last of the x86 schedule itineraries, I'm intending to cleanup the remaining uses of NoItinerary/OpndItins/etc. before resolving PR37093.
llvm-svn: 329967
Fangrui Song [Thu, 12 Apr 2018 22:27:38 +0000 (22:27 +0000)]
[CachePruning] Clarify the per-directory entry limit on Linux ext4.
Summary:
508 root node entries (root_limit)
510 internal node entries (node_limit)
For a filename with 40 bytes, its sizeof(ext4_dir_entry_2) = 48, a linear directory can contain at most floor(4096/48)=85 of them.
The real per-directory entry limit should be 508*510*85 =
22021800
The limit varies with the average length of filenames.
However, the Linux ext4 code does not try rebalancing the htree, so we will not be able to create filenames in a full leaf node. This is demonstrated with the following example, certain filenames cannot be used while others can:
% touch d/
0000000000000000000000000000000000816a6f
touch: cannot touch 'd/
0000000000000000000000000000000000816a6f': No
space left on device
% touch d/
0000000000000000000000000000000000816a70
# succeeded
Reviewers: pcc
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D45546
llvm-svn: 329966
Eli Friedman [Thu, 12 Apr 2018 22:21:36 +0000 (22:21 +0000)]
Remove -cc1 option "-backend-option".
It means the same thing as -mllvm; there isn't any reason to have two
options which do the same thing.
Differential Revision: https://reviews.llvm.org/D45109
llvm-svn: 329965
Sanjay Patel [Thu, 12 Apr 2018 22:14:58 +0000 (22:14 +0000)]
[DAGCombiner] simplify code; NFC
llvm-svn: 329964
Sanjay Patel [Thu, 12 Apr 2018 22:14:23 +0000 (22:14 +0000)]
[PowerPC] add fsub-fneg test; NFC
This is a test for a transform that was suggested in the post-commit
mailing list thread for rL329821. The target in question is not in
trunk, so PPC gets to stand in for it because it's the only in-tree
target that sets 'isFPExtFree()' to 'true'.
llvm-svn: 329963
Evgeniy Stepanov [Thu, 12 Apr 2018 22:06:18 +0000 (22:06 +0000)]
[asan] Remove malloc_context_size=0 from asan_device_setup.
This line was added in
r243679 - [asan] Support arm64 devices in asan_device_setup
without any good reason.
llvm-svn: 329962
Eli Friedman [Thu, 12 Apr 2018 22:04:11 +0000 (22:04 +0000)]
Don't call skipModule for CFI lowering passes.
opt-bisect shouldn't skip these passes; they lower intrinsics which
no other pass can handle.
llvm-svn: 329961
Rui Ueyama [Thu, 12 Apr 2018 21:57:04 +0000 (21:57 +0000)]
Do not keep shared symbols created from garbage-collected eliminated DSOs.
If all references to a DSO happen to be weak, and if the DSO is
specified with --as-needed, the DSO is not added to DT_NEEDED.
If that happens, we also need to eliminate shared symbols created
from the DSO. Otherwise, they become dangling references that point
to non-exsitent DSO.
Fixes https://bugs.llvm.org/show_bug.cgi?id=36991
Differential Revision: https://reviews.llvm.org/D45536
llvm-svn: 329960
George Burgess IV [Thu, 12 Apr 2018 21:38:43 +0000 (21:38 +0000)]
[ProfileSummary] Remove repeated cutoffs; NFCI
I'm told the repeat of "500000, 600000," is accidental, and should be
removed.
llvm-svn: 329959
Jan Korous [Thu, 12 Apr 2018 21:33:24 +0000 (21:33 +0000)]
[clangd][nfc] Simplify readDelimitedMessage()
istream::eof() is always false after successful getline()
llvm-svn: 329958
Brian Gesiak [Thu, 12 Apr 2018 21:28:04 +0000 (21:28 +0000)]
Let llvm-diff correctly deal with Undef/ConstantAggregateZero/ConstantVector/IndirectBr
Summary:
llvm-diff incorrectly reports that there's a diff when input IR contains undef/zeroinitializer/constantvector/indirectbr.
(This happens even if two identical files are given, e.g. `llvm-diff x.ll x.ll`)
This is fix to the bug report https://bugs.llvm.org/show_bug.cgi?id=33623 .
Reviewers: dexonsmith, rjmccall
Reviewed By: rjmccall
Subscribers: chenwj, mgrang, llvm-commits
Differential Revision: https://reviews.llvm.org/D34856
llvm-svn: 329957
Peter Collingbourne [Thu, 12 Apr 2018 21:23:55 +0000 (21:23 +0000)]
AArch64: Introduce a DAG combine for folding offsets into addresses.
This is a code size win in code that takes offseted addresses
frequently, such as C++ constructors that typically need to compute
an offseted address of a vtable. This reduces the size of Chromium
for Android's .text section by 108KB.
Differential Revision: https://reviews.llvm.org/D45199
llvm-svn: 329956
Vedant Kumar [Thu, 12 Apr 2018 20:58:24 +0000 (20:58 +0000)]
[DebugInfo] Create merged locations for instructions other than calls
This lifts a restriction on DILocation::getMergedLocation(), allowing it
to create merged locations for instructions other than calls.
Instruction::applyMergedLocation() now defaults to creating merged
locations for all instructions.
The default behavior of getMergedLocation() is unchanged: callers which
invoke it directly are unaffected.
This change will enable a follow-up Mem2Reg fix which improves crash
reporting.
Differential Revision: https://reviews.llvm.org/D45396
llvm-svn: 329955
George Burgess IV [Thu, 12 Apr 2018 20:54:05 +0000 (20:54 +0000)]
[ProfileSummary] Move a vector we're about to destroy anyway; NFC
llvm-svn: 329954
Simon Pilgrim [Thu, 12 Apr 2018 20:47:34 +0000 (20:47 +0000)]
[X86] Remove InstrItinClass entries from all x86 instruction defs (PR37093)
llvm-svn: 329953
Erich Keane [Thu, 12 Apr 2018 20:46:31 +0000 (20:46 +0000)]
[NFC] Fix terrible formatting of CGRecordLower constructor.
llvm-svn: 329952
Erik Pilkington [Thu, 12 Apr 2018 20:41:38 +0000 (20:41 +0000)]
[demangler] Add a partial demangling API for LLDB.
This parses a mangled name into an AST (typically an intermediate stage in
itaniumDemangle) and provides some functions to query certain properties or
print certain parts of the demangled name.
Differential revision: https://reviews.llvm.org/D44668
llvm-svn: 329951
Erik Pilkington [Thu, 12 Apr 2018 20:41:06 +0000 (20:41 +0000)]
[demangler] NFC: Some refactoring to support partial demangling.
I'm committing this to libcxxabi too so that the two demanglers remain as
simular as possible.
llvm-svn: 329950
Zinovy Nis [Thu, 12 Apr 2018 20:33:24 +0000 (20:33 +0000)]
Revert "[clang-tidy] [modernize-use-auto] Get only a length of token, not the token itself"
This reverts r329873 as getting only a single token length is wrong for multi-token type names,
like 'unsigned long int'.
llvm-svn: 329949
Sam Clegg [Thu, 12 Apr 2018 20:31:35 +0000 (20:31 +0000)]
[WebAssembly] Match llvm change to custom section size
Summary:
The content of custome sections no longer includes the
name itself.
See: https://reviews.llvm.org/D45579
Subscribers: jfb, dschuff, jgravelle-google, aheejin, sunfish, llvm-commits
Differential Revision: https://reviews.llvm.org/D45580
llvm-svn: 329948
Sam Clegg [Thu, 12 Apr 2018 20:31:12 +0000 (20:31 +0000)]
[WebAssembly] libObject: Don't include the name the size of custom sections
Differential Revision: https://reviews.llvm.org/D45579
llvm-svn: 329947
Martin Storsjo [Thu, 12 Apr 2018 20:07:38 +0000 (20:07 +0000)]
[MinGW] Look for libc++ headers in a triplet prefixed path as well
This makes it consistent with libstdc++ and the other default
include directories.
If these headers are found in both locations and one isn't a
symlink to the other, this will cause errors due to libc++ headers
having wrapper headers for some standard C headers, wrappers that
do #include_next the actual one.
If the same libc++ standard C wrapper header exists in more than one
include directory before the real system one, the header include
guard will stop it from doing another #include_next to pick up the
real one, breaking things.
As this is a rather uncommon situation, this should be acceptable
and toolchain maintainers can adapt accordingly if necessary.
Also simplify some of the existing code with a local variable.
Differential Revision: https://reviews.llvm.org/D45500
llvm-svn: 329946
Simon Pilgrim [Thu, 12 Apr 2018 19:59:35 +0000 (19:59 +0000)]
[X86] Remove InstrItinClass entries from SSE/AVX instructions defs (PR37093)
llvm-svn: 329945
Lang Hames [Thu, 12 Apr 2018 19:54:41 +0000 (19:54 +0000)]
[ORC] Use insert rather than emplace.
Hopefully this will fix the build failure at
http://lab.llvm.org:8011/builders/llvm-clang-x86_64-expensive-checks-win/builds/9028
llvm-svn: 329944
George Burgess IV [Thu, 12 Apr 2018 19:48:05 +0000 (19:48 +0000)]
Simplify; NFCI
llvm-svn: 329943
Vlad Tsyrklevich [Thu, 12 Apr 2018 19:35:39 +0000 (19:35 +0000)]
Fix doc typo
llvm-svn: 329942
Mandeep Singh Grang [Thu, 12 Apr 2018 19:31:37 +0000 (19:31 +0000)]
[RISCV] Fix logic to check if frame pointer should be used
Summary: The logic was broken for Linux triples as it returns true in the switch for Triple.isOSLinux().
Reviewers: asb, apazos
Reviewed By: asb
Subscribers: kito-cheng, shiva0217, cfe-commits
Differential Revision: https://reviews.llvm.org/D45237
llvm-svn: 329941
Simon Pilgrim [Thu, 12 Apr 2018 19:25:07 +0000 (19:25 +0000)]
[X86] Remove explicit SSE/AVX schedule itineraries from defs (PR37093)
llvm-svn: 329940
Sameer AbuAsal [Thu, 12 Apr 2018 19:22:40 +0000 (19:22 +0000)]
[RISCV] Add c.mv rs1, rs2 pattern for addi rs1, rs2, 0
Summary:
GCC compresses the pseudo instruction "mv rd, rs", which is an alias of
"addi rd, rs, 0", to "c.mv rd, rs".
In LLVM we rely on the canonical MC instruction (MCInst) to do our compression
checks and since there is no rule to compress "addi rd, rs, 0" --> "c.mv
rd, rs" we lose this compression opportunity to gcc.
In this patch we fix that by adding an addi to c.mv compression pattern, the
instruction "mv rd, rs" will be compressed to "c.mv rd, rs" just like
gcc does.
Patch by Zhaoshi Zheng (zzheng) and Sameer (sabuasal).
Reviewers: asb, apazos, zzheng, mgrang, shiva0217
Reviewed By: asb
Subscribers: rbar, johnrusso, simoncook, jordy.potman.lists, niosHD, kito-cheng, llvm-commits
Differential Revision: https://reviews.llvm.org/D45583
llvm-svn: 329939
Simon Pilgrim [Thu, 12 Apr 2018 18:46:15 +0000 (18:46 +0000)]
[X86] Remove remaining gpr schedule itineraries (PR37093)
llvm-svn: 329938
Gabor Buella [Thu, 12 Apr 2018 18:42:02 +0000 (18:42 +0000)]
[X86] Introduce wbinvd intrinsic
A previously missing intrinsic for an old instruction.
Reviewers: craig.topper, echristo
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D45311
llvm-svn: 329937
Gabor Buella [Thu, 12 Apr 2018 18:38:18 +0000 (18:38 +0000)]
[X86] Introduce LLVM wbinvd intrinsic
A previously missing intrinsic for an old instruction.
Reviewers: craig.topper, echristo
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D45312
llvm-svn: 329936
George Burgess IV [Thu, 12 Apr 2018 18:36:01 +0000 (18:36 +0000)]
Fix a typo in a comment; NFC
llvm-svn: 329935
Lang Hames [Thu, 12 Apr 2018 18:35:08 +0000 (18:35 +0000)]
[ORC] Plumb error notifications through the VSO interface.
This allows materializers to notify the VSO that they were unable to
resolve or finalize symbols.
llvm-svn: 329934
Simon Pilgrim [Thu, 12 Apr 2018 18:25:38 +0000 (18:25 +0000)]
[X86] Remove gpr shift/extension schedule itineraries (PR37093)
llvm-svn: 329933
Gabor Buella [Thu, 12 Apr 2018 18:15:39 +0000 (18:15 +0000)]
NFC - Indentation fixes in predefined-arch-macros.c
Consistently separating tests with empty lines.
Helps while navigating this file.
Reviewers: craig.topper
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D45561
llvm-svn: 329932
Lei Huang [Thu, 12 Apr 2018 18:00:14 +0000 (18:00 +0000)]
[Power9]Legalize and emit code for converting (Un)Signed DWord to Quad-Precision
Legalize and emit code for:
* xscvsdqp
* xscvudqp
Differential Revision: https://reviews.llvm.org/D45230
llvm-svn: 329931
Aaron Ballman [Thu, 12 Apr 2018 17:53:21 +0000 (17:53 +0000)]
Fix the try_acquire_capability attribute to behave like the other try-lock functions. Fixes PR32954.
llvm-svn: 329930
Benjamin Kramer [Thu, 12 Apr 2018 17:28:30 +0000 (17:28 +0000)]
[MCJIT] Remove the anchor from mcjit.
This is a layering violation. LTO shouldn't depend on MCJIT. The right
fix for this is moving the class somewhere else.
llvm-svn: 329929
Joachim Protze [Thu, 12 Apr 2018 17:23:26 +0000 (17:23 +0000)]
Set the license header for all OMPT files
llvm-svn: 329928
Eugene Zelenko [Thu, 12 Apr 2018 17:19:09 +0000 (17:19 +0000)]
[Documentation] Order of changes in existing checks in Release Notes.
llvm-svn: 329927
Petar Jovanovic [Thu, 12 Apr 2018 17:01:46 +0000 (17:01 +0000)]
[MIPS GlobalISel] remove superfluous #includes (NFC)
Remove superfluous #includes.
Minor code style change in MipsCallLowering::lowerFormalArguments().
llvm-svn: 329926
Kostya Kortchinsky [Thu, 12 Apr 2018 16:41:57 +0000 (16:41 +0000)]
[scudo] Remove duplicate cmake flags
Summary:
Now that common options are propagated again for runtimes build with D45507,
the -f{data,function}-sections flags are now duplicates, remove them.
Reviewers: alekseyshl
Reviewed By: alekseyshl
Subscribers: mgorny, delcypher, #sanitizers, llvm-commits
Differential Revision: https://reviews.llvm.org/D45575
llvm-svn: 329925
Aaron Ballman [Thu, 12 Apr 2018 16:41:55 +0000 (16:41 +0000)]
Correctly diagnose when a conversion function is declared with a type qualifier in the declaration specifiers rather than in the conversion type id. Fixes PR30595.
llvm-svn: 329924
Krzysztof Parzyszek [Thu, 12 Apr 2018 16:25:35 +0000 (16:25 +0000)]
[Hexagon] Enable auto-vectorization only when -fvectorize was given
llvm-svn: 329923
Jessica Paquette [Thu, 12 Apr 2018 16:16:18 +0000 (16:16 +0000)]
[AArch64] Move AFI->setRedZone(false) to top of emitPrologue
AFI->setRedZone(false) was put in the wrong place before, and so it only fired
on functions that didn't have stack frames. This moves that to the top of
emitPrologue to make sure that every function without a redzone has it set
correctly.
This also adds a function representing one of the early exit cases (GHC calling
convention) to the MachineOutliner noredzone test to ensure that we can outline
from functions like these, where we never use a redzone.
llvm-svn: 329922
Ben Hamilton [Thu, 12 Apr 2018 16:00:05 +0000 (16:00 +0000)]
[Test Fix] Fix broken test Index/comment-objc-parameterized-classes.m
I broke this test in D45498 when I changed the formatter to remove
spaces before Objective-C lightweight generics.
This fixes the test.
Test Plan:
% make -j16 check-llvm-tools-llvm-lit && ./bin/llvm-lit -sv ../llvm/tools/clang/test/Index/comment-objc-parameterized-classes.m
llvm-svn: 329921
Sanjay Patel [Thu, 12 Apr 2018 15:27:01 +0000 (15:27 +0000)]
revert r328921 - [DAGCombine] (float)((int) f) --> ftrunc (PR36617)
This change is exposing UB in source code - as was warned/predicted. :)
See D44909 for discussion. Reverting while we figure out how to fix things.
llvm-svn: 329920
Ben Hamilton [Thu, 12 Apr 2018 15:11:55 +0000 (15:11 +0000)]
[clang-format] Do not break after ObjC category open paren
Summary:
Previously, `clang-format` would break Objective-C
category extensions after the opening parenthesis to avoid
breaking the protocol list:
```
% echo "@interface
ccccccccccccc (
ccccccccccc) <
ccccccccccccc> { }" | \
clang-format -assume-filename=foo.h -style="{BasedOnStyle: llvm, \
ColumnLimit: 40}"
@interface
ccccccccccccc (
ccccccccccc) <
ccccccccccccc> {
}
```
This looks fairly odd, as we could have kept the category extension
on the previous line.
Category extensions are a single item, so they are generally very
short compared to protocol lists. We should prefer breaking after the
opening `<` of the protocol list over breaking after the opening `(`
of the category extension.
With this diff, we now avoid breaking after the category extension's
open paren, which causes us to break after the protocol list's
open angle bracket:
```
% echo "@interface
ccccccccccccc (
ccccccccccc) <
ccccccccccccc> { }" | \
./bin/clang-format -assume-filename=foo.h -style="{BasedOnStyle: llvm, \
ColumnLimit: 40}"
@interface
ccccccccccccc (
ccccccccccc) <
ccccccccccccc> {
}
```
Test Plan: New test added. Confirmed test failed before diff and
passed after diff by running:
% make -j16 FormatTests && ./tools/clang/unittests/Format/FormatTests
Reviewers: djasper, jolesiak
Reviewed By: djasper
Subscribers: klimek, cfe-commits
Differential Revision: https://reviews.llvm.org/D45526
llvm-svn: 329919
Ben Hamilton [Thu, 12 Apr 2018 15:11:53 +0000 (15:11 +0000)]
[clang-format] Improve ObjC guessing heuristic by supporting all @keywords
Summary:
This diff improves the Objective-C guessing heuristic by
replacing the hard-coded list of a subset of Objective-C @keywords
with a general check which supports all @keywords.
I also added a few more Foundation keywords which were missing from
the heuristic.
Test Plan: Unit tests updated. Ran tests with:
% make -j16 FormatTests && ./tools/clang/unittests/Format/FormatTests
Reviewers: djasper, jolesiak
Reviewed By: djasper
Subscribers: klimek, cfe-commits
Differential Revision: https://reviews.llvm.org/D45521
llvm-svn: 329918
Ben Hamilton [Thu, 12 Apr 2018 15:11:51 +0000 (15:11 +0000)]
[clang-format] Don't insert space between ObjC class and lightweight generic
Summary:
In D45185, I added clang-format parser support for Objective-C
generics. However, I didn't touch the whitespace logic, so they
got the same space logic as Objective-C protocol lists.
In every example in the Apple SDK and in the documentation,
there is no space between the class name and the opening `<`
for the lightweight generic specification, so this diff
removes the space and updates the tests.
Test Plan: Tests updated. Ran tests with:
% make -j16 FormatTests && ./tools/clang/unittests/Format/FormatTests
Reviewers: djasper, jolesiak
Reviewed By: djasper
Subscribers: klimek, cfe-commits
Differential Revision: https://reviews.llvm.org/D45498
llvm-svn: 329917
Ben Hamilton [Thu, 12 Apr 2018 15:11:48 +0000 (15:11 +0000)]
[clang-format] Always indent wrapped Objective-C selector names
Summary:
Currently, indentation of Objective-C method names which are wrapped
onto the next line due to a long return type is controlled by the
style option `IndentWrappedFunctionNames`.
This diff changes the behavior so we always indent wrapped Objective-C
selector names.
NOTE: I partially reverted https://github.com/llvm-mirror/clang/commit/
6159c0fbd1876c7f5f984b4830c664cc78f16e2e / rL242484, as it was causing wrapped selectors to be double-indented. Its tests in FormatTestObjC.cpp still pass.
Test Plan: Tests updated. Ran tests with:
% make -j12 FormatTests && ./tools/clang/unittests/Format/FormatTests
Reviewers: djasper, jolesiak, stephanemoore, thakis
Reviewed By: djasper
Subscribers: stephanemoore, klimek, cfe-commits
Differential Revision: https://reviews.llvm.org/D45004
llvm-svn: 329916
Krzysztof Parzyszek [Thu, 12 Apr 2018 15:11:11 +0000 (15:11 +0000)]
[Pipeliner] Use std::stable_sort when ordering NodeSets
There are cases when individual NodeSets can be equal with respect to
the ordering criteria. Since they are stored in an ordered container,
use stable_sort to preserve the relative order of equal NodeSets.
This should remove non-determinism discovered by shuffling done in
llvm::sort with expensive checks enabled.
llvm-svn: 329915
Malcolm Parsons [Thu, 12 Apr 2018 14:48:48 +0000 (14:48 +0000)]
Diagnose cases of "return x" that should be "return std::move(x)" for efficiency
Summary:
This patch adds two new diagnostics, which are off by default:
**-Wreturn-std-move**
This diagnostic is enabled by `-Wreturn-std-move`, `-Wmove`, or `-Wall`.
Diagnose cases of `return x` or `throw x`, where `x` is the name of a local variable or parameter, in which a copy operation is performed when a move operation would have been available. The user probably expected a move, but they're not getting a move, perhaps because the type of "x" is different from the return type of the function.
A place where this comes up in the wild is `stdext::inplace_function<Sig, N>` which implements conversion via a conversion operator rather than a converting constructor; see https://github.com/WG21-SG14/SG14/issues/125#issue-
297201412
Another place where this has come up in the wild, but where the fix ended up being different, was
try { ... } catch (ExceptionType ex) {
throw ex;
}
where the appropriate fix in that case was to replace `throw ex;` with `throw;`, and incidentally to catch by reference instead of by value. (But one could contrive a scenario where the slicing was intentional, in which case throw-by-move would have been the appropriate fix after all.)
Another example (intentional slicing to a base class) is dissected in https://github.com/accuBayArea/Slides/blob/master/slides/2018-03-07.pdf
**-Wreturn-std-move-in-c++11**
This diagnostic is enabled only by the exact spelling `-Wreturn-std-move-in-c++11`.
Diagnose cases of "return x;" or "throw x;" which in this version of Clang *do* produce moves, but which prior to Clang 3.9 / GCC 5.1 produced copies instead. This is useful in codebases which care about portability to those older compilers.
The name "-in-c++11" is not technically correct; what caused the version-to-version change in behavior here was actually CWG 1579, not C++14. I think it's likely that codebases that need portability to GCC 4.9-and-earlier may understand "C++11" as a colloquialism for "older compilers." The wording of this diagnostic is based on feedback from @rsmith.
**Discussion**
Notice that this patch is kind of a negative-space version of Richard Trieu's `-Wpessimizing-move`. That diagnostic warns about cases of `return std::move(x)` that should be `return x` for speed. These diagnostics warn about cases of `return x` that should be `return std::move(x)` for speed. (The two diagnostics' bailiwicks do not overlap: we don't have to worry about a `return` statement flipping between the two states indefinitely.)
I propose to write a paper for San Diego that would relax the implicit-move rules so that in C++2a the user //would// see the moves they expect, and the diagnostic could be re-worded in a later version of Clang to suggest explicit `std::move` only "in C++17 and earlier." But in the meantime (and/or forever if that proposal is not well received), this diagnostic will be useful to detect accidental copy operations.
Reviewers: rtrieu, rsmith
Reviewed By: rsmith
Subscribers: lebedev.ri, Rakete1111, rsmith, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D43322
Patch by Arthur O'Dwyer.
llvm-svn: 329914
Simon Dardis [Thu, 12 Apr 2018 14:41:38 +0000 (14:41 +0000)]
[mips] Correct the predicates of the load/store (double)word for coprocessor 3.
llvm-svn: 329913
Simon Pilgrim [Thu, 12 Apr 2018 14:31:42 +0000 (14:31 +0000)]
[X86] Remove AES/CLMUL/CRC32/LDDQU/MOVNT/POPCNT/SHA schedule itineraries (PR37093)
llvm-svn: 329912
Anastasia Stulova [Thu, 12 Apr 2018 14:17:04 +0000 (14:17 +0000)]
[OpenCL] Added -std/-cl-std=c++
This is std option for OpenCL C++ v1.0.
Differential Revision: https://reviews.llvm.org/D45363
llvm-svn: 329911
Benjamin Kramer [Thu, 12 Apr 2018 13:52:02 +0000 (13:52 +0000)]
Revert "Reapply "[PR16756] Use SSAUpdaterBulk in JumpThreading." one more time."
This reverts commit r329865. Causes stage2/stage3 miscompare.
llvm-svn: 329910
Sander de Smalen [Thu, 12 Apr 2018 13:19:32 +0000 (13:19 +0000)]
[AArch64][AsmParser] Unify 'addVectorListOperands' functions.
Summary:
Merged 'addVectorList64Operands' and 'addVectorList128Operands' into a
generic 'addVectorListOperands', which can be easily extended to work
for SVE vectors.
This is patch [4/6] in a series to add assembler/disassembler support for
SVE's contiguous ST1 (scalar+imm) instructions.
Reviewers: fhahn, rengolin, javed.absar, huntergr, SjoerdMeijer, t.p.northover, echristo, evandro
Reviewed By: rengolin
Subscribers: kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D45430
llvm-svn: 329909
Francis Visoiu Mistrih [Thu, 12 Apr 2018 12:59:50 +0000 (12:59 +0000)]
[CodeGen] Allow printing MachineMemOperands with less context in SDAGDumper
Don't assume SelectionDAG is non-null as the targets can use it with a
null pointer.
Differential Revision: https://reviews.llvm.org/D44611
llvm-svn: 329908
Sam Parker [Thu, 12 Apr 2018 12:49:40 +0000 (12:49 +0000)]
[IRCE] isKnownNonNegative helper function
Created a helper function to query for non negative SCEVs. Uses the
SGE predicate to catch constants that could be interpreted as
negative.
Differential Revision: https://reviews.llvm.org/D45481
llvm-svn: 329907
Simon Pilgrim [Thu, 12 Apr 2018 12:43:49 +0000 (12:43 +0000)]
[X86] Remove remaining system/special schedule itineraries (PR37093)
llvm-svn: 329906
Simon Dardis [Thu, 12 Apr 2018 12:37:02 +0000 (12:37 +0000)]
[mips] Correct the predicates for special nops, tlb ctrl instrs, software breakpoint and prefx.
Reviewers: atanasyan, abeserminji
Differential Revision: https://reviews.llvm.org/D44436
llvm-svn: 329905
Aaron Ballman [Thu, 12 Apr 2018 12:21:41 +0000 (12:21 +0000)]
Allow [[maybe_unused]] on static data members; these are considered variables and the attribute should appertain to them.
Patch by S. B. Tam.
llvm-svn: 329904
Simon Pilgrim [Thu, 12 Apr 2018 12:09:24 +0000 (12:09 +0000)]
[X86] Remove system/control schedule itineraries (PR37093)
llvm-svn: 329903
Roman Lebedev [Thu, 12 Apr 2018 12:06:42 +0000 (12:06 +0000)]
[clang-tidy] readability-function-size: add VariableThreshold param.
Summary:
Pretty straight-forward, just count all the variable declarations in the function's body, and if more than the configured threshold - do complain.
Note that this continues perverse practice of disabling the new option by default.
I'm not certain where is the balance point between not being too noisy, and actually enforcing the good practice.
If we really want to not disable this by default, but also to not cause too many new warnings, we could default to 50 or so.
But that is a lot of variables too...
I was able to find one coding style referencing variable count:
- https://www.kernel.org/doc/html/v4.15/process/coding-style.html#functions
> Another measure of the function is the number of local variables. They shouldn’t exceed 5-10, or you’re doing something wrong.
Reviewers: hokein, xazax.hun, JonasToth, aaron.ballman, alexfh
Reviewed By: aaron.ballman
Subscribers: kimgr, Eugene.Zelenko, rnkovacs, cfe-commits
Tags: #clang-tools-extra
Differential Revision: https://reviews.llvm.org/D44602
llvm-svn: 329902
Roman Lebedev [Thu, 12 Apr 2018 12:04:57 +0000 (12:04 +0000)]
[InstCombine][NFC]: Add tests: foldSelectICmpAndAnd(): and is commutative
Summary:
The fold added in D45108 did not account for the fact that
the and instruction is commutative, and if the mask is a variable,
the mask variable and the fold variable may be swapped.
I have noticed this by accident when looking into [[ https://bugs.llvm.org/show_bug.cgi?id=6773 | PR6773 ]]
Reviewers: spatel, craig.topper
Reviewed By: spatel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D45538
llvm-svn: 329901
Sander de Smalen [Thu, 12 Apr 2018 11:40:52 +0000 (11:40 +0000)]
[AArch64][AsmParser] Make parse function for VectorLists generic to other vector types.
Summary:
Added 'RegisterKind' to the VectorListOp structure, so that this operand
type can be reused for SVE vector lists in a later patch. It also
refactors the 'tryParseVectorList' function so it can be used directly
in the ParserMethod of an operand. The parsing can now parse multiple
kinds of vectors and recover if there is no match.
This is patch [3/6] in a series to add assembler/disassembler support for
SVE's contiguous ST1 (scalar+imm) instructions.
Reviewers: fhahn, rengolin, javed.absar, huntergr, SjoerdMeijer, t.p.northover, echristo, evandro
Reviewed By: rengolin
Subscribers: kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D45429
llvm-svn: 329900
Shiva Chen [Thu, 12 Apr 2018 11:30:59 +0000 (11:30 +0000)]
[RISCV] Change function alignment to 4 bytes, and 2 bytes for RVC
Summary:
According RISC-V ELF psABI specification, base RV32 and RV64 ISAs only
allow 32-bit instruction alignment, but instruction allow to be aligned
to 16-bit boundaries for C-extension.
So we just align to 4 bytes and 2 bytes for C-extension is enough.
Reviewers: asb, apazos
Differential Revision: https://reviews.llvm.org/D45560
Patch by Kito Cheng.
llvm-svn: 329899
Simon Pilgrim [Thu, 12 Apr 2018 11:01:40 +0000 (11:01 +0000)]
[X86] Remove CMOV/SETCC schedule itineraries (PR37093)
llvm-svn: 329898
Jonas Devlieghere [Thu, 12 Apr 2018 10:51:52 +0000 (10:51 +0000)]
Revert "Don't assume backing thread shares protocol ID."
This reverts r329891 because the test case is timing out on linux:
http://lab.llvm.org:8011/builders/lldb-x86_64-ubuntu-14.04-cmake/builds/21834
llvm-svn: 329897
Simon Pilgrim [Thu, 12 Apr 2018 10:49:57 +0000 (10:49 +0000)]
[X86] Remove MMX/3DNow schedule itineraries (PR37093)
llvm-svn: 329896
Andrea Di Biagio [Thu, 12 Apr 2018 10:49:40 +0000 (10:49 +0000)]
[llvm-mca] Removed unused argument from cycleEvent. NFC
llvm-svn: 329895
Benjamin Kramer [Thu, 12 Apr 2018 10:35:24 +0000 (10:35 +0000)]
[clang-apply-replacements] Always initialize FormatStyle.
The cleanup logic reads from this for cleanups even if reformatting is
not requested.
Found by msan.
llvm-svn: 329894
Simon Pilgrim [Thu, 12 Apr 2018 10:27:37 +0000 (10:27 +0000)]
[X86] Remove X87 schedule itineraries (PR37093)
First of a number of commits to remove x86 schedule itineraries entirely - approved off-line with @craig.topper
llvm-svn: 329893