Matt Arsenault [Wed, 9 May 2018 21:11:35 +0000 (21:11 +0000)]
AMDGPU: Ignore any_extend in mul24 combine
If a multiply is truncated, SimplifyDemandedBits
sometimes turns a zero_extend of the inputs into an
any_extend, which makes the known bits computation unhelpful.
Ignore these and compute known bits for the underlying value,
since we insert the correct extend type after.
llvm-svn: 331919
Krzysztof Parzyszek [Wed, 9 May 2018 21:10:41 +0000 (21:10 +0000)]
[Hexagon] Add patterns for vector shift-and-accumulate
llvm-svn: 331918
Matt Arsenault [Wed, 9 May 2018 20:52:54 +0000 (20:52 +0000)]
AMDGPU: Handle partial shift reduction for variable shifts
If the variable shift amount has known bits, we can still reduce
the shift.
llvm-svn: 331917
Matt Arsenault [Wed, 9 May 2018 20:52:43 +0000 (20:52 +0000)]
AMDGPU: Partially shrink 64-bit shifts if reduced to 16-bit
This is an extension of an existing combine to reduce wider
shls if the result fits in the final result type. This
introduces the same combine, but reduces the shift to a middle
sized type to avoid the slow 64-bit shift.
llvm-svn: 331916
Vitaly Buka [Wed, 9 May 2018 20:42:11 +0000 (20:42 +0000)]
[sanitizer] Cleanup sorting functions
llvm-svn: 331915
Stella Stamenova [Wed, 9 May 2018 19:58:51 +0000 (19:58 +0000)]
[cmake, unittests] Fix the CMake file for the LLDB unittests to support multiple configurations
Summary: The current setup for the unit tests only works correctly when the generator does not support multiple configurations. When the generator supports multiple configurations, the inputs are not copied to the correct per-configuration directory. This change sets up the build to copy the inputs in each configuration directory.
Reviewers: labath, asmith, zturner
Reviewed By: labath
Subscribers: mgorny, llvm-commits
Differential Revision: https://reviews.llvm.org/D46642
llvm-svn: 331914
Simon Pilgrim [Wed, 9 May 2018 19:27:48 +0000 (19:27 +0000)]
[X86] Fix Broadwell's Shuffle256 schedule classes load latency values.
Allows us to remove some unnecessary InstRW overrides.
llvm-svn: 331913
Martin Storsjo [Wed, 9 May 2018 19:07:10 +0000 (19:07 +0000)]
[COFF] Fix dangling StringRefs from SVN 331900
llvm-svn: 331912
Simon Pilgrim [Wed, 9 May 2018 19:04:15 +0000 (19:04 +0000)]
[X86] Merge instregex patterns to reduce InstrRW compile time.
llvm-svn: 331911
David Bolvansky [Wed, 9 May 2018 18:57:17 +0000 (18:57 +0000)]
Allow copy elision in path concatenation
Summary:
Just port of libstdc++'s fix to libc++ fs: https://github.com/gcc-mirror/gcc/commit/
e6ac4004fe49d785c63bf87aec4b095b5ce1d19f
Author of fix: Jonathan Wakely
Reviewers: EricWF, mclow.lists
Reviewed By: EricWF
Subscribers: smeenai, christof, cfe-commits, llvm-commits
Differential Revision: https://reviews.llvm.org/D46593
llvm-svn: 331910
Matt Arsenault [Wed, 9 May 2018 18:37:39 +0000 (18:37 +0000)]
AMDGPU: Add combine for trunc of bitcast from build_vector
If the truncate is only accessing the first element of the vector,
we can use the original source value.
This helps with some combine ordering issues after operations are
lowered to integer operations between bitcasts of build_vector.
In particular it stops unnecessarily materializing the unused
top half of a vector in some cases.
llvm-svn: 331909
Kostya Serebryany [Wed, 9 May 2018 18:35:09 +0000 (18:35 +0000)]
[dfsan] add one more sanitizer-coverage hook to the whitelist
llvm-svn: 331908
Krzysztof Parzyszek [Wed, 9 May 2018 18:33:59 +0000 (18:33 +0000)]
[Hexagon] Check the end of the correct container (fix typo)
llvm-svn: 331907
Matt Arsenault [Wed, 9 May 2018 18:29:26 +0000 (18:29 +0000)]
AMDGPU: Stop special casing constant indexes of extract_vector_elt
The same result folds out of the dynamic expansion logic if the
index is constant.
llvm-svn: 331906
Julie Hockett [Wed, 9 May 2018 18:27:37 +0000 (18:27 +0000)]
[tools] Updating PPCallbacks::InclusionDirective calls
[revision] added SrcMgr::CharacteristicKind to the InclusionDirective
callback, this revision updates instances of it in clang-tools-extra.
Differential Revision: https://reviews.llvm.org/D46615
llvm-svn: 331905
Julie Hockett [Wed, 9 May 2018 18:27:33 +0000 (18:27 +0000)]
[clang] Adding CharacteristicKind to PPCallbacks::InclusionDirective
Adding a SrcMgr::CharacteristicKind parameter to the InclusionDirective
in PPCallbacks, and updating calls to that function. This will be useful
in https://reviews.llvm.org/D43778 to determine which includes are system
headers.
Differential Revision: https://reviews.llvm.org/D46614
llvm-svn: 331904
Martin Storsjo [Wed, 9 May 2018 18:21:03 +0000 (18:21 +0000)]
[llvm-rc] Handle C preprocessor output
When preprocessing resource scripts (which can easily be done outside
of llvm-rc), included headers can leave behind C declarations (despite
preprocessing with -DRC_INVOKED), that can't be parsed by a resource
compiler.
This is handled in all of rc.exe, by parsing the preprocessor output
line markers and ignoring content from files named *.h and *.c,
documented at [1].
In addition to this filtering, strip out any other preprocessor directive
that is left behind (like pragmas) which also can't be handled by the
tokenizer.
The added test uses both standard #line markers (supported by rc.exe) and
GNU style extended line markers, thus this test doesn't pass with rc.exe,
but passes with GNU windres. (Windres on the other hand doesn't filter
out files named *.c, only *.h.)
Differential Revision: https://reviews.llvm.org/D46579
[1] https://msdn.microsoft.com/en-us/library/windows/desktop/
aa381033(v=vs.85).aspx
llvm-svn: 331903
Martin Storsjo [Wed, 9 May 2018 18:20:56 +0000 (18:20 +0000)]
[llvm-rc] Add support for the RCDATA resource type
This is the same as any other user defined resource, but with
a specific allocated resource type number.
Differential Revision: https://reviews.llvm.org/D46636
llvm-svn: 331902
Martin Storsjo [Wed, 9 May 2018 18:20:49 +0000 (18:20 +0000)]
[llvm-rc] Allow -1 for control IDs in old style dialogs with 16 bit fields
-1 is commonly used as ID for controls that one don't want to
refer to later. For DIALOG resources, the IDs are 16 bit numbers,
and -1 gets interpreted as UINT32_MAX earlier, which then later is
too large to write into a uint16_t.
Differential Revision: https://reviews.llvm.org/D46506
llvm-svn: 331901
Martin Storsjo [Wed, 9 May 2018 18:19:41 +0000 (18:19 +0000)]
[COFF] Allow specifying export forwarding in a def file
Previously this was only supported when specified on the command line
or in directives.
Differential Revision: https://reviews.llvm.org/D46244
llvm-svn: 331900
Alexey Bataev [Wed, 9 May 2018 18:02:37 +0000 (18:02 +0000)]
[OPENMP] Generate unique names for offloading regions id.
It is required to emit unique names for offloading regions ids. Required
to support compilation and linking of several compilation units.
llvm-svn: 331899
Roman Tereshin [Wed, 9 May 2018 17:28:18 +0000 (17:28 +0000)]
Reapplying r331819 [GlobalISel][Legalizer] More concise and faster widenScalar, NFC
The commit was a suspect for clang-cmake-aarch64-global-isel and
clang-cmake-aarch64-quick bot failures, proved to be innocent.
llvm-svn: 331898
David Blaikie [Wed, 9 May 2018 17:28:10 +0000 (17:28 +0000)]
llvm-mca: Add missing includes
Move the header include in the primary source file to the top to
validate that it doesn't depend on any other inclusions.
llvm-svn: 331897
Craig Topper [Wed, 9 May 2018 17:14:27 +0000 (17:14 +0000)]
[DAGCombiner] In visitBITCAST when trying to constant fold the bitcast, only call getBitcast if its an fp->int or int->fp conversion even when before legalize ops.
Previously if !LegalOperations we would blindly call getBitcast and hope that getNode would constant fold it. But if the conversion is between a vector and a scalar, getNode has no simplification.
This means we would just get back the original N. We would then return that N which would make the caller of visitBITCAST think that we used CombineTo and did our own worklist management. This prevents target specific optimizations from being called for vector/scalar bitcasts until after legal operations.
llvm-svn: 331896
Yaxun Liu [Wed, 9 May 2018 17:07:06 +0000 (17:07 +0000)]
[OpenCL] Fix typos in emitted enqueue kernel function names
Two typos:
vaarg => vararg
get_kernel_preferred_work_group_multiple => get_kernel_preferred_work_group_size_multiple
Differential Revision: https://reviews.llvm.org/D46601
llvm-svn: 331895
Matt Morehouse [Wed, 9 May 2018 17:05:07 +0000 (17:05 +0000)]
[libFuzzer] Disable print-func.test on Darwin.
The try-lock guard change seems to be making this test fail on Mac, but
I haven't been able to reproduce the failure. Disabling the test on Mac
to fix build bot.
llvm-svn: 331894
Craig Topper [Wed, 9 May 2018 16:57:48 +0000 (16:57 +0000)]
[X86] Only enable the __ud2 and __int2c builtins if intrin.h has been included.
Differential Revision: https://reviews.llvm.org/D46332
llvm-svn: 331893
Greg Clayton [Wed, 9 May 2018 16:42:53 +0000 (16:42 +0000)]
General cleanup to minimize the .debug_types patch
This cleanup is designed to make the https://reviews.llvm.org/D32167 patch smaller and easier to read.
Cleanup in this patch:
Allow DWARFUnit subclasses to hand out the data that should be used when decoding data for a DIE. The information might be in .debug_info or could be in .debug_types. There is a new virtual function on DWARFUnit that each subclass must override:
virtual const lldb_private::DWARFDataExtractor &DWARFUnit::GetData() const;
This allows DWARFCompileUnit and eventually DWARFTypeUnit to hand out different data to be used when decoding the DIE information.
Add a new pure virtual function to get the size of the DWARF unit header:
virtual uint32_t DWARFUnit::GetHeaderByteSize() const = 0;
This allows DWARFCompileUnit and eventually DWARFTypeUnit to hand out different offsets where the first DIE starts when decoding DIE information from the unit.
Added a new function to DWARFDataExtractor to get the size of an offset:
size_t DWARFDataExtractor::GetDWARFSizeOfOffset() const;
Removed dead dumping and parsing code in the DWARFDebugInfo class.
Inlined a bunch of calls in DWARFUnit for accessors that were just returning integer member variables.
Renamed DWARFUnit::Size() to DWARFUnit::GetHeaderByteSize() as it clearly states what it is doing and makes more sense.
Differential Revision: https://reviews.llvm.org/D46606
llvm-svn: 331892
Michael Kruse [Wed, 9 May 2018 16:23:56 +0000 (16:23 +0000)]
[ScopInfo] Remove bail out condition in buildMinMaxAccess().
The condition was introduced in r267142 to mitigate a long compile-time
case. In r306087, a max-computation limit was introduced that should
handle the same case while leaving the max disjuncts heuristic it
should have replaced intact.
Today, the max disjuncts bail-out causes problems in that it prematurely
stops SCoPs from being detected, e.g. in SPEC's lbm. This would hit less
like if isl_set_coalesce would be called after isl_set_remove_divs
(which makes more basic_set likely to be coalescable) instead of before.
This patch tries to remove the premature max-disjuncts bail-out
condition by using simple_hull() to reduce the computational overhead,
instead of directly invalidating that SCoP.
Differential Revision: https://reviews.llvm.org/D45066
Contributed-by: Sahil Girish Yerawar <cs15btech11044@iith.ac.in>
llvm-svn: 331891
Kostya Kortchinsky [Wed, 9 May 2018 16:20:52 +0000 (16:20 +0000)]
[sanitizer] Correct 64-bit atomic_store on 32-bit "other" platforms
Summary:
I think there might be something to optimize in `atomic_store`.
Currently, if everything goes well (and we have a different new value), we
always iterate 3 times.
For example, `with a = 0`, `oldval = a`, `newval = 42`, we get:
```
oldval = 0, newval = 42, curval = 0
oldval = 0, newval = 42, curval = 42
oldval = 42, newval = 42, curval = 42
```
and then it breaks.
Unless I am not seeing something, I don't see a point to the third iteration.
If the current value is the one we want, we should just break.
This means that 2 iterations (with a different newval) should be sufficient to
achieve what we want.
Reviewers: dvyukov, alekseyshl
Reviewed By: dvyukov
Subscribers: kubamracek, delcypher, #sanitizers, llvm-commits
Differential Revision: https://reviews.llvm.org/D46597
llvm-svn: 331890
David Bolvansky [Wed, 9 May 2018 16:09:31 +0000 (16:09 +0000)]
[InstCombine] snprintf optimizations
Reviewers: spatel, efriedma, majnemer, rja, bkramer
Reviewed By: rja, bkramer
Subscribers: rja, llvm-commits
Differential Revision: https://reviews.llvm.org/D46285
llvm-svn: 331889
Amara Emerson [Wed, 9 May 2018 15:53:06 +0000 (15:53 +0000)]
[DAGCombine] Change store merge candidates check cut off to 1024.
The previous value of 8192 resulted in severe compile time hits in
some pathological cases.
rdar://
39781410
Differential Revision: https://reviews.llvm.org/D46581
llvm-svn: 331888
Krzysztof Parzyszek [Wed, 9 May 2018 15:44:40 +0000 (15:44 +0000)]
[Hexagon] Fix sanitizer error about using -1u in variable of enum type
llvm-svn: 331887
Stella Stamenova [Wed, 9 May 2018 15:35:19 +0000 (15:35 +0000)]
[lit, lldbsuite] Add a bug reference to the failing TestLinuxCore and fix an undefined property in dotest.py
Summary:
1) In TestLinuxCore rather than skipping the tests on Windows, mark them as expected failures and add a bug reference
2) In dotest.py replace the undefined property in the exceptions with the actual property causing the exception
Reviewers: asmith, labath, zturner
Reviewed By: labath, zturner
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D46590
llvm-svn: 331886
Krzysztof Parzyszek [Wed, 9 May 2018 15:18:12 +0000 (15:18 +0000)]
[LV] Change MaxVectorSize bound to 256 in assertion, NFC otherwise
It's possible to have a vector of 256 bytes in HVX code on Hexagon
(vector pair in 128-byte mode).
llvm-svn: 331885
Pavel Labath [Wed, 9 May 2018 15:13:45 +0000 (15:13 +0000)]
APFloat/x87: Fix string conversion for "unnormal" values (pr35860)
Summary:
Unnormal values are a feature of some very old x87 processors. We handle
them correctly for the most part -- the only exception was an unnormal
value whose significand happened to be zero. In this case the APFloat
was still initialized as normal number (category = fcNormal), but a
subsequent toString operation would assert because the math would
produce nonsensical values for the zero significand.
During review, it was decided that the correct way to fix this is to
treat all unnormal values as NaNs (as that is what any >=386 processor
will do).
The issue was discovered because LLDB would crash when trying to print
some "long double" values.
Reviewers: skatkov, scanon, gottesmm
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D41868
llvm-svn: 331884
Krzysztof Parzyszek [Wed, 9 May 2018 15:02:04 +0000 (15:02 +0000)]
[Hexagon] Simplify MCCodeEmitter, move data to tables
llvm-svn: 331883
Pavel Labath [Wed, 9 May 2018 14:49:43 +0000 (14:49 +0000)]
Fix Windows build for the Predicate.h refactor in r331880
llvm-svn: 331882
Sid Manning [Wed, 9 May 2018 14:44:54 +0000 (14:44 +0000)]
Add basic compiler-rt builtins support for hexagon.
Differential Revision: https://reviews.llvm.org/D46364
llvm-svn: 331881
Pavel Labath [Wed, 9 May 2018 14:29:30 +0000 (14:29 +0000)]
Modernize and clean-up the Predicate class
Summary:
The comments on this class were out of date with the implementation, and
the implementation itself was inconsistent with our usage of the Timeout
class (I started converting everything to use this class back in D27136,
but I missed this one). I avoid duplicating the waiting logic by
introducing a templated WaitFor function, and make other functions
delegate to that. This function can be also used as a replacement for
the unused WaitForBitToBeSet functions I removed, if it turns out to be
necessary.
As this changes the meaning of a "zero" timeout, I tracked down all the
callers of these functions and updated them accordingly. Propagating the
changes to all the callers of RunShellCommand was a bit too much for
this patch, so I stopped there and will continue that in a follow-up
patch.
I also add some basic unittests for the functions I modified.
Reviewers: jingham, clayborg
Subscribers: mgorny, lldb-commits
Differential Revision: https://reviews.llvm.org/D46580
llvm-svn: 331880
Alexey Bataev [Wed, 9 May 2018 14:15:18 +0000 (14:15 +0000)]
[OPENMP] Mark global tors/dtors as used.
If the global variables are marked as declare target and they need
ctors/dtors, these ctors/dtors are emitted and then invoked by the
offloading runtime library. They are not explicitly used in the emitted
code and thus can be optimized out. Patch marks these functions as used,
so the optimizer cannot remove these function during the optimization
phase.
llvm-svn: 331879
Karl-Johan Karlsson [Wed, 9 May 2018 13:34:57 +0000 (13:34 +0000)]
[LV] Add lit testcase for bitcast problem. NFC
llvm-svn: 331878
Anastasia Stulova [Wed, 9 May 2018 13:23:26 +0000 (13:23 +0000)]
[OpenCL] Add constant address space to __func__ in AST.
Added string literal helper function to obtain the type
attributed by a constant address space.
Also fixed predefind __func__ expr to use the helper
to constract the string literal correctly.
Differential Revision: https://reviews.llvm.org/D46049
llvm-svn: 331877
Pavel Labath [Wed, 9 May 2018 13:21:16 +0000 (13:21 +0000)]
[Support/Path] Make handling of paths like "///" consistent
Summary:
Various path functions were not treating paths consisting of slashes
alone consistently. For example, the iterator-based accessors decomposed the
path "///" into two elements: "/" and ".". This is not too bad, but it
is different from the behavior specified by posix:
```
A pathname that contains ***at least one non-slash character*** and that
ends with one or more trailing slashes shall be resolved as if a single
dot character ( '.' ) were appended to the pathname.
```
More importantly, this was different from how we treated the same path
in the filename+parent_path functions, which decomposed this path into
"." and "". This was completely wrong as it lost the information that
this was an absolute path which referred to the root directory.
This patch fixes this behavior by making sure all functions treat paths
consisting of (back)slashes alone the same way as "/". I.e., the
iterator-based functions will just report one component ("/"), and the
filename+parent_path will decompose them into "/" and "".
A slightly controversial topic here may be the treatment of "//". Posix
says that paths beginning with "//" may have special meaning and indeed
we have code which parses paths like "//net/foo/bar" specially. However,
as we were already not being consistent in parsing the "//" string
alone, and any special parsing for it would complicate the code further,
I chose to treat it the same way as longer sequences of slashes (which
are guaranteed to be the same as "/").
Another slight change of behavior is in the parsing of paths like
"//net//". Previously the last component of this path was ".". However,
as in our parsing the "//net" part in this path was the same as the
"drive" part in "c:\" and the next slash was the "root directory", it
made sense to treat "//net//" the same way as "//net/" (i.e., not to add
the extra "." component at the end).
Reviewers: zturner, rnk, dblaikie, Bigcheese
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D45942
llvm-svn: 331876
Manuel Klimek [Wed, 9 May 2018 13:20:03 +0000 (13:20 +0000)]
Do not warn on unused parameters for functions with empty bodies.
If a function has an empty body, all parameters are trivially unused.
llvm-svn: 331875
Sven van Haastregt [Wed, 9 May 2018 13:16:17 +0000 (13:16 +0000)]
[OpenCL] Restrict various keywords in OpenCL C++ mode
Restrict the following keywords in the OpenCL C++ language mode,
according to Sections 2.2 & 2.9 of the OpenCL C++ 1.0 Specification.
- dynamic_cast
- typeid
- register (already restricted in OpenCL C, update the diagnostic)
- thread_local
- exceptions (try/catch/throw)
- access qualifiers read_only, write_only, read_write
Support the `__global`, `__local`, `__constant`, `__private`, and
`__generic` keywords in OpenCL C++. Leave the unprefixed address
space qualifiers such as global available, i.e., do not mark them as
reserved keywords in OpenCL C++. libclcxx provides explicit address
space pointer classes such as `global_ptr` and `global<T>` that are
implemented using the `__`-prefixed qualifiers.
Differential Revision: https://reviews.llvm.org/D46022
llvm-svn: 331874
Adhemerval Zanella [Wed, 9 May 2018 12:48:22 +0000 (12:48 +0000)]
[AArch64] Improve cost of vector division by constant
With custom lowering for vector MULLH{S,U}, it is now profitable to
vectorize a divide by constant loop for the custom types (v16i8, v8i16,
and v4i32). The cost if based on TargetLowering::Build{S,U}DIV which
uses a multiply by constant plus adjustment to express a divide by
constant.
Both {u,s}mull{2} are expressed as Instruction::Mul and shifts by
Instruction::AShr.
llvm-svn: 331873
Nico Weber [Wed, 9 May 2018 12:39:39 +0000 (12:39 +0000)]
Remove 'abi-breaking-checks' lit feature.
Its only two uses were removed in r311730.
Effectively reverts r304851 (but that code has removed around a bit since then).
https://reviews.llvm.org/D46619
clang side done in r331871.
llvm-svn: 331872
Nico Weber [Wed, 9 May 2018 12:38:51 +0000 (12:38 +0000)]
Remove unused lit setting, see https://reviews.llvm.org/D46619
llvm-svn: 331871
Alexander Kornienko [Wed, 9 May 2018 12:27:21 +0000 (12:27 +0000)]
Fixes issue introduced by r331556.
Closes bug: https://bugs.llvm.org/show_bug.cgi?id=37357
Patch by Rafael Stahl!
Differential revision: https://reviews.llvm.org/D46633
llvm-svn: 331870
Pavel Labath [Wed, 9 May 2018 12:26:19 +0000 (12:26 +0000)]
Revert "DWARFVerifier: Check "completeness" of .debug_names section"
The new verifier check has found an error in the
debug-names-name-collisions.ll test on the PS4 bot:
error: Name Index @ 0x0: Entry @ 0xdc: mismatched Name of DIE @ 0x23: index - _ZN3foo3fooE; debug_info - foo.
Reverting while I investigate whether this is a bug in the verifier or
the generator.
This reverts commit r331868.
llvm-svn: 331869
Pavel Labath [Wed, 9 May 2018 12:06:17 +0000 (12:06 +0000)]
DWARFVerifier: Check "completeness" of .debug_names section
Summary:
This patch implements a check which makes sure all entries required by
the DWARF v5 specification are present in the Name Index. The algorithm
tries to follow the wording of Section 6.1.1.1 of the spec as closely as
possible.
The main deviation from it is that instead of a whitelist-based approach
in the spec "The name index must contain an entry for each debugging
information entry that defines a named subprogram, label, variable,
type, or namespace" I chose a blacklist-based one, where I consider
everything to be "in" and then remove the entries that don't make sense.
I did this because it has more potential for catching interesting cases
and the above is a bit vague (it uses plain words like "variable" and
"subprogram", but the rest of the section speaks about specific TAGs).
This approach has raised some interesting questions, the main one being
whether enumerator values should be indexed. The consensus seems to be
that they should, although it does not follow from section 6.1.1.1.
For the time being I made the verifier ignore these, as LLVM does not do
this yet, and I wanted to get a clean run when verifying generated debug
info.
Another interesting case was the DW_TAG_imported_declaration. It was not
immediately clear to me whether this should go in or not, but currently
it is not indexed, and (unlike the enumerators) in does not seem to cause
problems for LLDB, so I've also ignored it.
Reviewers: JDevlieghere, aprantl, dblaikie
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D46583
llvm-svn: 331868
Simon Pilgrim [Wed, 9 May 2018 11:42:34 +0000 (11:42 +0000)]
[CostModel][X86] Split off SLM checks
A future patch will require this and the diff is much better if we perform the split separately.
llvm-svn: 331867
Benjamin Kramer [Wed, 9 May 2018 11:38:57 +0000 (11:38 +0000)]
Revert "[InstCombine] snprintf optimizations"
This reverts commit r331849. It miscompiles
snprintf(buf, sizeof(buf), "%s", "any constant string); into
memcpy(buf, "%s", sizeof("any constant string"));
llvm-svn: 331866
Benjamin Kramer [Wed, 9 May 2018 11:17:30 +0000 (11:17 +0000)]
[DebugInfo] Mark tests using -debug-only as REQUIRES: asserts
llvm-svn: 331865
Simon Pilgrim [Wed, 9 May 2018 11:01:16 +0000 (11:01 +0000)]
[X86] Cleanup WriteFStore/WriteVecStore schedules
MOVNTPD/MOVNTPS should be WriteFStore
Standardized BDW/HSW/SKL/SKX WriteFStore/WriteVecStore - fixes some missed instregex patterns. (V)MASKMOVDQU was already using the default, its costs gets increased but is still nowhere near the real cost of that nasty instruction....
llvm-svn: 331864
Simon Dardis [Wed, 9 May 2018 10:33:21 +0000 (10:33 +0000)]
[mips] Move conditional moves out of isCodeGenOnly
Reviewers: atanasyan, smaksimovic, abeserminji
Differential Revision: https://reviews.llvm.org/D46389
llvm-svn: 331863
Benjamin Kramer [Wed, 9 May 2018 10:09:23 +0000 (10:09 +0000)]
[wasm] Update test for variables->retainedNodes rename.
This fixes the test after LLVM r331841.
llvm-svn: 331862
Hans Wennborg [Wed, 9 May 2018 09:29:58 +0000 (09:29 +0000)]
Revert r331843 "[DebugInfo] Generate debug information for labels."
It broke the Chromium build (see reply on the review).
> Generate DILabel metadata and call llvm.dbg.label after label
> statement to associate the metadata with the label.
>
> Differential Revision: https://reviews.llvm.org/D45045
>
> Patch by Hsiangkai Wang.
This doesn't revert the change to backend-unsupported-error.ll
that seems to correspond to an llvm-side change.
llvm-svn: 331861
Martin Storsjo [Wed, 9 May 2018 09:22:03 +0000 (09:22 +0000)]
[COFF] Improve correctness of def parsing for GNU features
The operator == used for exporting a function with a different
name in the DLL compared to the name in the import library
(which is useful for adding linker level aliases for function
in the import library) is a feature distinct and different from
the operator = used for exporting a function with a different
name (both in import library and DLL) than in the implementation
producing the DLL.
When creating an import library using dlltool, from a def file that
contains forwards (Func = OtherDll.Func), this shouldn't affect the
produced import library, which should still behave just as if it
was a normal exported function.
This clears a lot of confusion and subtle misunderstandings, and
avoids a parameter that was used to avoid creating weak aliases
when invoked from lld. (This parameter was added previously due to
the existing conflation of the two features.)
Differential Revision: https://reviews.llvm.org/D46245
llvm-svn: 331860
Martin Storsjo [Wed, 9 May 2018 09:21:53 +0000 (09:21 +0000)]
[COFF] Improve correctness of def parsing for GNU features
The operator == used for exporting a function with a different
name in the DLL compared to the name in the import library
(which is useful for adding linker level aliases for function
in the import library) is a feature distinct and different from
the operator = used for exporting a function with a different
name (both in import library and DLL) than in the implementation
producing the DLL.
When creating an import library using dlltool, from a def file that
contains forwards (Func = OtherDll.Func), this shouldn't affect the
produced import library, which should still behave just as if it
was a normal exported function.
This clears a lot of confusion and subtle misunderstandings, and
avoids a parameter that was used to avoid creating weak aliases
when invoked from lld. (This parameter was added previously due to
the existing conflation of the two features.)
Differential Revision: https://reviews.llvm.org/D46245
llvm-svn: 331859
Martin Storsjo [Wed, 9 May 2018 09:11:01 +0000 (09:11 +0000)]
Revert "[Driver] Use -fuse-line-directives by default in MSVC mode"
This reverts commit SVN r331666.
It was afterwards pointed out in https://reviews.llvm.org/D46520
that #line directives lose information about what parts come from a
system header. That means the result of -E usually won't compile,
since Windows headers are typically full of warnings and
default-error warnings.
llvm-svn: 331858
Krasimir Georgiev [Wed, 9 May 2018 09:02:11 +0000 (09:02 +0000)]
[clang-format] Respect BreakBeforeClosingBrace while calculating length
Summary:
This patch makes `getLengthToMatchingParen` respect the `BreakBeforeClosingBrace`
ParenState for matching scope closers. In order to distinguish between paren states
introduced by real vs. fake parens, I've added the token opening the ParensState
to that struct.
Reviewers: djasper
Reviewed By: djasper
Subscribers: klimek, cfe-commits
Differential Revision: https://reviews.llvm.org/D46519
llvm-svn: 331857
Simon Atanasyan [Wed, 9 May 2018 08:42:30 +0000 (08:42 +0000)]
[driver] Add mips_Features_Group to Options to improve documentation sorting
Move all of the MIPS-only options into a new m_mips_Features_Group.
Nearly all other targets have most target-specific options grouped,
but MIPS does not.
The primary benefits are that the options will be listed together (and
thus identifiable as MIPS-specific even if they have no help string) in
the ClangCommandLineReference, and that Options.td is a bit more organized.
A secondary benefit is that a custom version of clang can more easily
hide/disable groups of options for unsupported targets.
Patch by Vince Del Vecchio
Differential Revision: https://reviews.llvm.org/D46450
llvm-svn: 331856
Pavel Labath [Wed, 9 May 2018 08:21:25 +0000 (08:21 +0000)]
[DWARF] Align non-accelerated function fullname searching with the apple-tables path
Summary:
Before this patch the two paths were doing very different things
- the apple path searched the .apple_names section, which contained
mangled names, as well as basenames of all functions. It returned any
name it found.
- the non-accelerated path looked in the "full name" index we built
ourselves, which contained mangled as well as demangled names of all
functions (but no basenames). Then however, if it did not find a match
it did an extra search in the basename index, with some special
handling for anonymous namespaces.
This aligns the two paths by changing the non-accelerated path to return
the same results as in the apple-tables one. In pratice, this means we
will search in both the "basename", "method" and "fullname" indexes (in
the manual indexes these are separate indexes. This means the function
will return some slightly inappropriate results (e.g. bar::baz::foo when
one asks for a "full name" foo), but this can be handled by additional
filtering, independently indexing method. I've also stopped inserting
demangled names into the "fullname" index, as that is inconsistent with
the apple path.
Reviewers: clayborg, JDevlieghere
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D46576
llvm-svn: 331855
Hans Wennborg [Wed, 9 May 2018 08:20:14 +0000 (08:20 +0000)]
Add a test for r331746.
Thanks to pcc for creating the test file!
llvm-svn: 331854
Shiva Chen [Wed, 9 May 2018 07:09:28 +0000 (07:09 +0000)]
[DebugInfo] Fix test failed due to debug-label-mi.ll and debug-label-opt.ll
Make these two test cases more generic for other architectures.
Please refer to '[DebugInfo] Convert intrinsic llvm.dbg.label to
MachineInstr.'
Patch by Hsiangkai Wang
llvm-svn: 331853
Bjorn Pettersson [Wed, 9 May 2018 06:52:12 +0000 (06:52 +0000)]
[MergedLoadStoreMotion] Fix a debug invariant bug in mergeStores
Summary:
MergedLoadStoreMotion::mergeStores is using some heuristics
to limit the amount of stores that it tries to sink (see
MagicCompileTimeControl in MergedLoadStoreMotion.cpp). The
heuristic involves counting the number of instructions in
one of the basic blocks that is part of the transformation.
We now ignore dbg intrinsics when counting instruction for
the MagicCompileTimeControl heuristic. This to make sure that
the amount of stores that are sunk doesn't depend on the amount
of debug information (if -g is used or not).
Reviewers: Gerolf, davide, majnemer
Reviewed By: davide
Subscribers: dberlin, bjope, aprantl, JDevlieghere, llvm-commits
Differential Revision: https://reviews.llvm.org/D46600
llvm-svn: 331852
Simon Atanasyan [Wed, 9 May 2018 06:51:58 +0000 (06:51 +0000)]
[ELF][MIPS] Add test case to cover LA25 r6 thunks generation. NFC
llvm-svn: 331851
Robert Widmann [Wed, 9 May 2018 06:45:28 +0000 (06:45 +0000)]
[LLVM-C] Correct types in Go bindings
Summary: Fixes a test failure introduced in rL331114.
Reviewers: whitequark
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D46631
llvm-svn: 331850
David Bolvansky [Wed, 9 May 2018 06:34:20 +0000 (06:34 +0000)]
[InstCombine] snprintf optimizations
Reviewers: spatel, efriedma, majnemer, rja
Reviewed By: rja
Subscribers: rja, llvm-commits
Differential Revision: https://reviews.llvm.org/D46285
llvm-svn: 331849
Shiva Chen [Wed, 9 May 2018 06:22:39 +0000 (06:22 +0000)]
[DebugInfo] Fix test failed due to new DISubprogram attributes.
Please refer to '[DebugInfo] Add DILabel metadata and intrinsic
llvm.dbg.label'. I have renamed the 'variables' attributes to
'retainedNodes' to include local variables and local labels for the
function.
Patch by Hsiangkai Wang.
llvm-svn: 331848
Craig Topper [Wed, 9 May 2018 06:07:20 +0000 (06:07 +0000)]
[X86] Combine (vXi1 (bitcast (-1)))) and (vXi1 (bitcast (0))) to all ones or all zeros vXi1 vector.
llvm-svn: 331847
Daniel Sanders [Wed, 9 May 2018 05:00:17 +0000 (05:00 +0000)]
Revert r331816 and r331820 - [globalisel] Add a combiner helpers for extending loads and use them in a pre-legalize combiner for AArch64
Reverting this to see if the clang-cmake-aarch64-global-isel and
clang-cmake-aarch64-quick bots are failing because of this commit.
We know it wasn't r331819.
llvm-svn: 331846
JF Bastien [Wed, 9 May 2018 03:51:12 +0000 (03:51 +0000)]
_Atomic of empty struct shouldn't assert
Summary:
An _Atomic of an empty struct is pretty silly. In general we just widen empty
structs to hold a byte's worth of storage, and we represent size and alignment
as 0 internally and let LLVM figure out what to do. For _Atomic it's a bit
different: the memory model mandates concrete effects occur when atomic
operations occur, so in most cases actual instructions need to get emitted. It's
really not worth trying to optimize empty struct atomics by figuring out e.g.
that a fence would do, even though sane compilers should do optimize atomics.
Further, wg21.link/p0528 will fix C++20 atomics with padding bits so that
cmpxchg on them works, which means that we'll likely need to do the zero-init
song and dance for empty atomic structs anyways (and I think we shouldn't
special-case this behavior to C++20 because prior standards are just broken).
This patch therefore makes a minor change to r176658 "Promote atomic type sizes
up to a power of two": if the width of the atomic's value type is 0, just use 1
byte for width and leave alignment as-is (since it should never be zero, and
over-aligned zero-width structs are weird but fine).
This fixes an assertion:
(NumBits >= MIN_INT_BITS && "bitwidth too small"), function get, file ../lib/IR/Type.cpp, line 241.
It seems like this has run into other assertions before (namely the unreachable
Kind check in ImpCastExprToType), but I haven't reproduced that issue with
tip-of-tree.
<rdar://problem/
39678063>
Reviewers: arphaman, rjmccall
Subscribers: aheejin, cfe-commits
Differential Revision: https://reviews.llvm.org/D46613
llvm-svn: 331845
Shiva Chen [Wed, 9 May 2018 02:42:00 +0000 (02:42 +0000)]
[DebugInfo] Examine all uses of isDebugValue() for debug instructions.
Because we create a new kind of debug instruction, DBG_LABEL, we need to
check all passes which use isDebugValue() to check MachineInstr is debug
instruction or not. When expelling debug instructions, we should expel
both DBG_VALUE and DBG_LABEL. So, I create a new function,
isDebugInstr(), in MachineInstr to check whether the MachineInstr is
debug instruction or not.
This patch has no new test case. I have run regression test and there is
no difference in regression test.
Differential Revision: https://reviews.llvm.org/D45342
Patch by Hsiangkai Wang.
llvm-svn: 331844
Shiva Chen [Wed, 9 May 2018 02:41:56 +0000 (02:41 +0000)]
[DebugInfo] Generate debug information for labels.
Generate DILabel metadata and call llvm.dbg.label after label
statement to associate the metadata with the label.
Differential Revision: https://reviews.llvm.org/D45045
Patch by Hsiangkai Wang.
llvm-svn: 331843
Shiva Chen [Wed, 9 May 2018 02:41:08 +0000 (02:41 +0000)]
[DebugInfo] Convert intrinsic llvm.dbg.label to MachineInstr.
In order to convert LLVM IR to MachineInstr, we need a new TargetOpcode,
DBG_LABEL, to ‘lower’ intrinsic llvm.dbg.label. The patch
creates this new TargetOpcode and convert intrinsic llvm.dbg.label to
MachineInstr through SelectionDAG.
In SelectionDAG, debug information is stored in SDDbgInfo. We create a
new data member of SDDbgInfo for labels and use the new data member,
SDDbgLabel, to create DBG_LABEL MachineInstr.
The new DBG_LABEL MachineInstr uses label metadata from LLVM IR as its
parameter. So, the backend could get metadata information of labels from
DBG_LABEL MachineInstr.
Differential Revision: https://reviews.llvm.org/D45341
Patch by Hsiangkai Wang.
llvm-svn: 331842
Shiva Chen [Wed, 9 May 2018 02:40:45 +0000 (02:40 +0000)]
[DebugInfo] Add DILabel metadata and intrinsic llvm.dbg.label.
In order to set breakpoints on labels and list source code around
labels, we need collect debug information for labels, i.e., label
name, the function label belong, line number in the file, and the
address label located. In order to keep these information in LLVM
IR and to allow backend to generate debug information correctly.
We create a new kind of metadata for labels, DILabel. The format
of DILabel is
!DILabel(scope: !1, name: "foo", file: !2, line: 3)
We hope to keep debug information as much as possible even the
code is optimized. So, we create a new kind of intrinsic for label
metadata to avoid the metadata is eliminated with basic block.
The intrinsic will keep existing if we keep it from optimized out.
The format of the intrinsic is
llvm.dbg.label(metadata !1)
It has only one argument, that is the DILabel metadata. The
intrinsic will follow the label immediately. Backend could get the
label metadata through the intrinsic's parameter.
We also create DIBuilder API for labels to be used by Frontend.
Frontend could use createLabel() to allocate DILabel objects, and use
insertLabel() to insert llvm.dbg.label intrinsic in LLVM IR.
Differential Revision: https://reviews.llvm.org/D45024
Patch by Hsiangkai Wang.
llvm-svn: 331841
Sean Fertile [Wed, 9 May 2018 02:07:53 +0000 (02:07 +0000)]
[PPC64] Add lazy symbol resolution stubs.
Adds support for .glink resolver stubs from the example implementation in the V2
ABI (Section 4.2.5.3. Procedure Linkage Table). The stubs are written to the
PltSection, and the sections are renamed to match the PPC64 ABI:
.got.plt --> .plt Type = SHT_NOBITS
.plt --> .glink
And adds the DT_PPC64_GLINK dynamic tag to the dynamic section when the plt is
not empty.
Differential Revision: https://reviews.llvm.org/D45642
llvm-svn: 331840
Roman Tereshin [Wed, 9 May 2018 01:43:12 +0000 (01:43 +0000)]
Revert r331819 [GlobalISel][Legalizer] More concise and faster widenScalar, NFC
Reverting this to see if the clang-cmake-aarch64-global-isel and
clang-cmake-aarch64-quick bots are failing because of this commit
llvm-svn: 331839
Lang Hames [Wed, 9 May 2018 01:38:13 +0000 (01:38 +0000)]
[RuntimeDyld][MachO] Properly handle thumb to thumb calls within a section.
Previously thumb bits were only checked for external relocations (thumb to arm
code and vice-versa). This patch adds detection for thumb callees in the same
section asthe (also thumb) caller.
The MachO/Thumb test case is updated to cover this, and redundant checks
(handled by the MachO/ARM test) are removed.
llvm-svn: 331838
Nico Weber [Wed, 9 May 2018 01:15:38 +0000 (01:15 +0000)]
Make CMakeLists.txt formatting more consistent with the rest of LLVM.
llvm-svn: 331837
Nico Weber [Wed, 9 May 2018 01:15:06 +0000 (01:15 +0000)]
fix path to llvm-cfi-verify unittests in docs
llvm-svn: 331836
Nico Weber [Wed, 9 May 2018 01:07:02 +0000 (01:07 +0000)]
Make llvm-cfi-verify CMakeLists.txt formatting more consistent with the rest of LLVM.
llvm-svn: 331835
Adrian Prantl [Wed, 9 May 2018 01:00:01 +0000 (01:00 +0000)]
Remove \brief commands from doxygen comments.
This is similar to the LLVM change https://reviews.llvm.org/D46290.
We've been running doxygen with the autobrief option for a couple of
years now. This makes the \brief markers into our comments
redundant. Since they are a visual distraction and we don't want to
encourage more \brief markers in new code either, this patch removes
them all.
Patch produced by
for i in $(git grep -l '\@brief'); do perl -pi -e 's/\@brief //g' $i & done
for i in $(git grep -l '\\brief'); do perl -pi -e 's/\\brief //g' $i & done
Differential Revision: https://reviews.llvm.org/D46320
llvm-svn: 331834
Petr Hosek [Wed, 9 May 2018 00:58:12 +0000 (00:58 +0000)]
Set CMAKE_BUILD_WITH_INSTALL_RPATH for Fuchsia runtimes
This doesn't make any difference since we don't use RPATH/RUNPATH
on Fuchsia but it avoids the CMake error when re-linking libraries
while building with Ninja.
Differntial Revision: https://reviews.llvm.org/D46610
llvm-svn: 331833
Heejin Ahn [Wed, 9 May 2018 00:53:50 +0000 (00:53 +0000)]
Support a funclet operand bundle in LowerInvoke
Summary:
The current LowerInvoke pass cannot handle invoke instructions with a
funclet bundle operand. The order of operands for an invoke instruction
is {call arguments, callee, funclet operand (if any), normal dest,
unwind dest}. The current code assumes there is no funclet operand and
incorrectly includes a funclet operand into call arguments.
Reviewers: rnk
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D46242
llvm-svn: 331832
Vitaly Buka [Wed, 9 May 2018 00:44:26 +0000 (00:44 +0000)]
[sanitizer] Remove unneeded blank lines
llvm-svn: 331831
Nico Weber [Wed, 9 May 2018 00:42:17 +0000 (00:42 +0000)]
Inline contents of LLVM_XRAY_TOOLS variable into its only use.
No behavior change.
https://reviews.llvm.org/D46402
llvm-svn: 331830
Vitaly Buka [Wed, 9 May 2018 00:41:53 +0000 (00:41 +0000)]
[sanitizer] Update .clang-format in compiler-rt
Historically style is Google, but we never used AllowShortIfStatementsOnASingleLine.
llvm-svn: 331829
Sriraman Tallam [Wed, 9 May 2018 00:20:45 +0000 (00:20 +0000)]
Fix test by adding -triple=x86_64-unknown-linux
llvm-svn: 331828
Petr Hosek [Wed, 9 May 2018 00:07:42 +0000 (00:07 +0000)]
[CMake] Use CMAKE_OBJCOPY and CMAKE_STRIP to externalize debug info
Don't hardcode objcopy and strip names, rather use CMAKE_OBJCOPY and
CMAKE_STRIP variables which allows users to override the tools used
such as using llvm-objcopy and llvm-strip instead of binutils versions.
Differential Revision: https://reviews.llvm.org/D46611
llvm-svn: 331827
Petr Hosek [Wed, 9 May 2018 00:05:28 +0000 (00:05 +0000)]
[CMake] Include llvm-strip in Fuchsia toolchain distribution
Now that llvm-strip is available, include it in the Fuchsia toolchain.
Differential Revision: https://reviews.llvm.org/D46612
llvm-svn: 331826
Matt Morehouse [Tue, 8 May 2018 23:45:05 +0000 (23:45 +0000)]
[libFuzzer] Guard symbolization with try-lock.
Summary:
When out-of-memory or timeout occurs, threads can be stopped during
symbolization, thereby causing a deadlock when the OOM/TO handlers
attempt symbolization. We avoid this deadlock by skipping symbolization
if another thread is symbolizing.
Reviewers: kcc
Reviewed By: kcc
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D46605
llvm-svn: 331825
Davide Italiano [Tue, 8 May 2018 23:28:15 +0000 (23:28 +0000)]
[SimplifyCFG] Fix a crash when folding PHIs.
We enter MergeBlockIntoPredecessor with a block looking like this:
for.inc.us-lcssa: ; preds = %cond.end
%k.1.lcssa.ph = phi i32 [ %conv15, %cond.end ]
%t.3.lcssa.ph = phi i32 [ %k.1.lcssa.ph, %cond.end ]
br label %for.inc, !dbg !66
[note the first arg of the PHI being a PHI].
FoldSingleEntryPHINodes gets rid of both PHIs (calling, eraseFromParent).
But right before we call the function, we push into IncomingValues the
only argument of the PHIs, and shortly after we try to iterate over
something which has been invalidated before :(
The fix its not trying to remove PHIs which have an incoming value
coming from the same BB we're looking at.
Fixes PR37300 and rdar://problem/
39910460
Differential Revision: https://reviews.llvm.org/D46568
llvm-svn: 331824
Sriraman Tallam [Tue, 8 May 2018 23:19:50 +0000 (23:19 +0000)]
New option -z keep-text-section-prefix to keep text sections with prefixes separate.
Separate output sections for selected text section prefixes to enable TLB optimizations and for readablilty.
Differential Revision: https://reviews.llvm.org/D45841
llvm-svn: 331823
Roman Lebedev [Tue, 8 May 2018 23:15:58 +0000 (23:15 +0000)]
Partially revert r331456: [clang-tidy] Remove AnalyzeTemporaryDtors option.
That broke every single .clang-tidy config out there
which happened to specify AnalyzeTemporaryDtors option:
YAML:5:24: error: unknown key 'AnalyzeTemporaryDtors'
AnalyzeTemporaryDtors: false
^~~~~
Error parsing <...>/.clang-tidy: Invalid argument
More so, that error isn't actually a error, the
clang-tidy does not exit with $? != 0, it continues
with the default config.
Surely this breakage isn't the intended behavior.
But if it is, feel free to revert this commit.
llvm-svn: 331822
Peter Collingbourne [Tue, 8 May 2018 22:59:05 +0000 (22:59 +0000)]
MC: Remove dead code. NFCI.
We should never emit an SHT_DYNSYM into an object file.
llvm-svn: 331821
Daniel Sanders [Tue, 8 May 2018 22:58:35 +0000 (22:58 +0000)]
[globalisel] Correct r331816 to check the opcode before calling getOperand().
Fix a silly mistake in my pre-commit changes for r331816. It should check what
opcode the insn is before extracting the operands.
NFC at the moment since the caller already checked the opcode.
llvm-svn: 331820