platform/upstream/llvm.git
6 years ago[GISel]: Add Opcodes for a few LLVM Intrinsics
Aditya Nandakumar [Fri, 17 Aug 2018 01:41:56 +0000 (01:41 +0000)]
[GISel]: Add Opcodes for a few LLVM Intrinsics

https://reviews.llvm.org/D50401

Add opcodes for llvm.intrinsic.trunc, round, and update the IRTranslator
for the same.

Reviewed by: dsanders.

llvm-svn: 339977

6 years agoRevert "[libFuzzer] Use std::discrete_distribution for input selection."
Matt Morehouse [Fri, 17 Aug 2018 01:30:27 +0000 (01:30 +0000)]
Revert "[libFuzzer] Use std::discrete_distribution for input selection."

This reverts r339973 due to msan.test failing on
sanitizer-x86_64-linux-fuzzer bot.

llvm-svn: 339976

6 years ago[ADT] Replace a member initializer of a union with an explicit
Chandler Carruth [Fri, 17 Aug 2018 01:10:33 +0000 (01:10 +0000)]
[ADT] Replace a member initializer of a union with an explicit
constructor.

This breaking an old/weird host compiler is my best bet for the current
crashes I'm getting from bots since this functionality was added to this
ADT.

llvm-svn: 339975

6 years agoAdd a relocation for R_AARCH64_ABS32 in ObjectFileELF
Stephane Sezer [Fri, 17 Aug 2018 00:35:47 +0000 (00:35 +0000)]
Add a relocation for R_AARCH64_ABS32 in ObjectFileELF

Summary:
.rela.debug_info relocations are being done via
ObjectFileELF::ApplyRelocations for aarch64. Currently, the switch case
that iterates over the relocation type is only implemented for a few
different types and `assert(false)`es over the rest.

Implement the relocation for R_AARCH64_ABS32 in ApplyRelocations

Reviewers: sas, xiaobai, javed.absar, espindola

Reviewed By: sas

Subscribers: emaste, arichardson, kristof.beyls

Differential Revision: https://reviews.llvm.org/D50369

Change by Nathan Lanza <lanza@fb.com>

llvm-svn: 339974

6 years ago[libFuzzer] Use std::discrete_distribution for input selection.
Matt Morehouse [Fri, 17 Aug 2018 00:13:22 +0000 (00:13 +0000)]
[libFuzzer] Use std::discrete_distribution for input selection.

Summary:
Since we're casting from double to size_t during input selection, we
really want a discrete distribution over size_t rather than a piecewise
distribution over doubles.

Reviewers: kcc

Reviewed By: kcc

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D50356

llvm-svn: 339973

6 years ago[WebAssembly] Modify LateEHPrepare one-line description (NFC)
Heejin Ahn [Fri, 17 Aug 2018 00:12:04 +0000 (00:12 +0000)]
[WebAssembly] Modify LateEHPrepare one-line description (NFC)

llvm-svn: 339972

6 years agoRevert "Establish the <bit> header. NFC yet. Reviewed as https://reviews.llvm.org...
Vitaly Buka [Thu, 16 Aug 2018 23:57:16 +0000 (23:57 +0000)]
Revert "Establish the <bit> header. NFC yet. Reviewed as https://reviews.llvm.org/D50815"

Breaks build on sanitizer bots.

This reverts commit r339943.

llvm-svn: 339971

6 years agoDebugInfo: Remove command line (& target-based) disabling of pubnames in favor of...
David Blaikie [Thu, 16 Aug 2018 23:57:15 +0000 (23:57 +0000)]
DebugInfo: Remove command line (& target-based) disabling of pubnames in favor of metadata

Now that Clang disables NVPTX pubnames via metadata there's no need for
this fallback to target detection in the backend.

llvm-svn: 339970

6 years ago[libc++] Use correct rand.eng.mers all-zeroes seed sequence fallback
Hubert Tong [Thu, 16 Aug 2018 23:56:54 +0000 (23:56 +0000)]
[libc++] Use correct rand.eng.mers all-zeroes seed sequence fallback

Summary:
When a seed sequence would lead to having no non-zero significant bits
in the initial state of a `mersenne_twister_engine`, the fallback is to
flip the most significant bit of the first value that appears in the
textual representation of the initial state.

rand.eng.mers describes this as setting the value to be 2 to the power
of one less than w; the previous value encoded in the implementation,
namely one less than "2 to the power of w", is replaced by the correct
value in this patch.

Reviewers: mclow.lists, EricWF, jasonliu

Reviewed By: mclow.lists

Subscribers: mclow.lists, jasonliu, EricWF, christof, ldionne, cfe-commits

Differential Revision: https://reviews.llvm.org/D50736

llvm-svn: 339969

6 years agoDisable pubnames in NVPTX debug info using metadata
David Blaikie [Thu, 16 Aug 2018 23:56:32 +0000 (23:56 +0000)]
Disable pubnames in NVPTX debug info using metadata

llvm-svn: 339968

6 years ago[WebAssembly] CFG stackify support for exception handling
Heejin Ahn [Thu, 16 Aug 2018 23:50:59 +0000 (23:50 +0000)]
[WebAssembly] CFG stackify support for exception handling

Summary:
This adds support for exception handling to CFGStackify pass. This only
adds TRY / END_TRY markers and DOES NOT yet fix unwind mismatches that
can be created by the linearization of the CFG into the structural wasm
format. The mismatch fix will be added by following patches.

In detail, this patch
- Added support for TRY / END_TRY markers to support EH
- Changed many static functions into class member functions as they take
too many arguments now
- Added several more bookeeping data structures
- Refactored routines that decide where to insert markers, because
without refactoring this got too complicated as we added support for new
kinds of markers (TRY/END_TRY).
- Rewrote rethrow instructions' BB arguments to relative depths in EH
pad stack.

Reviewers: dschuff, sunfish

Subscribers: sbc100, jgravelle-google, llvm-commits

Differential Revision: https://reviews.llvm.org/D48273

llvm-svn: 339967

6 years agoRevert "Implement a (simple) Markdown generator"
Julie Hockett [Thu, 16 Aug 2018 23:50:51 +0000 (23:50 +0000)]
Revert "Implement a (simple) Markdown generator"

This reverts commit r339948, as it's breaking a few bots in ways that I
can't reproduce right now.

llvm-svn: 339966

6 years agoRemove commented out constructor from Scalar
Alex Langford [Thu, 16 Aug 2018 23:23:18 +0000 (23:23 +0000)]
Remove commented out constructor from Scalar

This appears to have been commented out since the initial checkin of
lldb.

llvm-svn: 339965

6 years agoRelax a CHECK line to allow for dso_local
Vedant Kumar [Thu, 16 Aug 2018 23:19:50 +0000 (23:19 +0000)]
Relax a CHECK line to allow for dso_local

Fixes a bot failure:
http://lab.llvm.org:8011/builders/llvm-clang-x86_64-expensive-checks-win/builds/11806

llvm-svn: 339964

6 years agoRevert "Revert "[hwasan] Add __hwasan_handle_longjmp.""
Evgeniy Stepanov [Thu, 16 Aug 2018 23:17:14 +0000 (23:17 +0000)]
Revert "Revert "[hwasan] Add __hwasan_handle_longjmp.""

This reapplies commit r339935 with the following changes:
* make longjmp test C, not C++, to avoid dependency on libc++/libstdc++
* untag pointer in memset interceptor

x86_64 does not have TBI, so hwasan barely works there. Tests must be carefully
written in a way that does not leak tagged pointer to system libraries.

llvm-svn: 339963

6 years ago[x86/MIR] Implement support for pre- and post-instruction symbols, as
Chandler Carruth [Thu, 16 Aug 2018 23:11:05 +0000 (23:11 +0000)]
[x86/MIR] Implement support for pre- and post-instruction symbols, as
well as MIR parsing support for `MCSymbol` `MachineOperand`s.

The only real way to test pre- and post-instruction symbol support is to
use them in operands, so I ended up implementing that within the patch
as well. I can split out the operand support if folks really want but it
doesn't really seem worth it.

The functional implementation of pre- and post-instruction symbols is
now *completely trivial*. Two tiny bits of code in the (misnamed)
AsmPrinter. It should be completely target independent as well. We emit
these exactly the same way as we emit basic block labels. Most of the
code here is to give full dumping, MIR printing, and MIR parsing support
so that we can write useful tests.

The MIR parsing of MC symbol operands still isn't 100%, as it forces the
symbols to be non-temporary and non-local symbols with names. However,
those names often can encode most (if not all) of the special semantics
desired, and unnamed symbols seem especially annoying to serialize and
de-serialize. While this isn't perfect or full support, it seems plenty
to write tests that exercise usage of these kinds of operands.

The MIR support for pre-and post-instruction symbols was quite
straightforward. I chose to print them out in an as-if-operand syntax
similar to debug locations as this seemed the cleanest way and let me
use nice introducer tokens rather than inventing more magic punctuation
like we use for memoperands.

However, supporting MIR-based parsing of these symbols caused me to
change the design of the symbol support to allow setting arbitrary
symbols. Without this, I don't see any reasonable way to test things
with MIR.

Differential Revision: https://reviews.llvm.org/D50833

llvm-svn: 339962

6 years agoAdd a basic integration test for C++ smart pointers
Vedant Kumar [Thu, 16 Aug 2018 23:05:15 +0000 (23:05 +0000)]
Add a basic integration test for C++ smart pointers

Check that the debugger can pretty-print unique_ptr and shared_ptr when
passed as a function argument.

rdar://42314305

llvm-svn: 339961

6 years agoRemove outdated TODOs in RegisterValue
Alex Langford [Thu, 16 Aug 2018 22:48:46 +0000 (22:48 +0000)]
Remove outdated TODOs in RegisterValue

These TODOs were for setting m_type in RegisterValue::SetValueFromString
in the case where reg_info's encoding was eEncodingUint or
eEncodingSint. m_type is set by SetUInt{8,16,32,64.128} during the
SetUInt call.

llvm-svn: 339959

6 years ago[InstCombine] add reflection fold for tan(-x)
Sanjay Patel [Thu, 16 Aug 2018 22:46:20 +0000 (22:46 +0000)]
[InstCombine] add reflection fold for tan(-x)

This is a follow-up suggested with rL339604.
For tan(), we don't have a corresponding LLVM
intrinsic -- unlike sin/cos -- so this is the
only way/place that we can do this fold currently.

llvm-svn: 339958

6 years agoRevert "[hwasan] Add __hwasan_handle_longjmp."
Evgeniy Stepanov [Thu, 16 Aug 2018 22:28:02 +0000 (22:28 +0000)]
Revert "[hwasan] Add __hwasan_handle_longjmp."

This reverts commit 339935 which breaks hwasan tests on x86_64.

llvm-svn: 339957

6 years ago[InstrProf] Use atomic profile counter updates for TSan
Vedant Kumar [Thu, 16 Aug 2018 22:24:47 +0000 (22:24 +0000)]
[InstrProf] Use atomic profile counter updates for TSan

Thread sanitizer instrumentation fails to skip all loads and stores to
profile counters. This can happen if profile counter updates are merged:

  %.sink = phi i64* ...
  %pgocount5 = load i64, i64* %.sink
  %27 = add i64 %pgocount5, 1
  %28 = bitcast i64* %.sink to i8*
  call void @__tsan_write8(i8* %28)
  store i64 %27, i64* %.sink

To suppress TSan diagnostics about racy counter updates, make the
counter updates atomic when TSan is enabled. If there's general interest
in this mode it can be surfaced as a clang/swift driver option.

Testing: check-{llvm,clang,profile}

rdar://40477803

Differential Revision: https://reviews.llvm.org/D50867

llvm-svn: 339955

6 years agoUpdate the Core file loading instructions so they keep the process stopped after...
Greg Clayton [Thu, 16 Aug 2018 22:13:01 +0000 (22:13 +0000)]
Update the Core file loading instructions so they keep the process stopped after attaching to a core file.

llvm-svn: 339954

6 years ago[InstCombine] add tests for tan with negated arg; NFC
Sanjay Patel [Thu, 16 Aug 2018 22:05:51 +0000 (22:05 +0000)]
[InstCombine] add tests for tan with negated arg; NFC

llvm-svn: 339953

6 years agoFactor Node creation out of the demangler. No functionality change intended.
Richard Smith [Thu, 16 Aug 2018 22:04:36 +0000 (22:04 +0000)]
Factor Node creation out of the demangler. No functionality change intended.

(This is a port of llvm r339944 to libcxxabi.)

llvm-svn: 339952

6 years agoUpdate MemorySSA in Local utils removing blocks.
Alina Sbirlea [Thu, 16 Aug 2018 21:58:44 +0000 (21:58 +0000)]
Update MemorySSA in Local utils removing blocks.

Summary: Extend Local utils to update MemorySSA.

Subscribers: sanjoy, jlebar, Prazek, george.burgess.iv, llvm-commits

Differential Revision: https://reviews.llvm.org/D48790

llvm-svn: 339951

6 years ago[hwasan] Enable Android logging.
Evgeniy Stepanov [Thu, 16 Aug 2018 21:56:04 +0000 (21:56 +0000)]
[hwasan] Enable Android logging.

Summary: Enable syslog as soon as the shadow is mapped.

Reviewers: vitalybuka, kcc

Subscribers: srhines, kubamracek, llvm-commits

Differential Revision: https://reviews.llvm.org/D50865

llvm-svn: 339950

6 years ago[docs] Try to clarify the FuzzingLLVM docs
Justin Bogner [Thu, 16 Aug 2018 21:55:09 +0000 (21:55 +0000)]
[docs] Try to clarify the FuzzingLLVM docs

Try to improve these docs based on some recent questions that were
sent to llvm-dev:

  http://lists.llvm.org/pipermail/llvm-dev/2018-August/125329.html

llvm-svn: 339949

6 years agoImplement a (simple) Markdown generator
Julie Hockett [Thu, 16 Aug 2018 21:54:34 +0000 (21:54 +0000)]
Implement a (simple) Markdown generator

Implementing a simple Markdown generator from the emitted bitcode
summary of declarations. Very primitive at this point, but will be
expanded. Currently emits an .md file for each class and namespace,
listing its contents.

For a more detailed overview of the tool, see the design document
on the mailing list:
http://lists.llvm.org/pipermail/cfe-dev/2017-December/056203.html

Differential Revision: https://reviews.llvm.org/D43424

llvm-svn: 339948

6 years ago[DomTree] Add constructor to create a new DT based on current DT/CFG and a set of...
Alina Sbirlea [Thu, 16 Aug 2018 21:54:33 +0000 (21:54 +0000)]
[DomTree] Add constructor to create a new DT based on current DT/CFG and a set of Updates.

Summary:
Add the posibility of creating a new DT using a set of Updates.
This will essentially create a DT based on a CFG snapshot/view.

Additional refactoring for either this patch or follow-ups:
- create an utility for building BUI.
- replace BUI with a GraphDiff.

Reviewers: kuhar

Subscribers: sanjoy, jlebar, llvm-commits

Differential Revision: https://reviews.llvm.org/D50671

llvm-svn: 339947

6 years ago[DAGCombiner] Don't reassociate operations that have the vector reduction flag set.
Craig Topper [Thu, 16 Aug 2018 21:54:05 +0000 (21:54 +0000)]
[DAGCombiner] Don't reassociate operations that have the vector reduction flag set.

When nodes are reassociated the vector-reduction flag gets lost.

The test case is here is what would happen if you had a sum of absolute differences loop that started with a non-zero but contant sum and that loop was unrolled. The vectorizer will generate a constant vector for the initial value. And DAGCombiner reassociate tries to move it down the addition tree erasing the vector-reduction flag. Interestingly this moves constants the opposite direction of the reassociate IR pass.

I've chosen to just punt on the reassociate, but I suppose we could maybe preserve the flag if both nodes have it set.

Differential Revision: https://reviews.llvm.org/D50827

llvm-svn: 339946

6 years ago[X86] In EFLAGS copy pass, don't emit EXTRACT_SUBREG instructions since we're after...
Craig Topper [Thu, 16 Aug 2018 21:54:02 +0000 (21:54 +0000)]
[X86] In EFLAGS copy pass, don't emit EXTRACT_SUBREG instructions since we're after peephole

Normally the peephole pass converts EXTRACT_SUBREG to COPY instructions. But we're after peephole so we can't rely on it to clean these up.

To fix this, the eflags pass now emits a COPY with a subreg input.

I also noticed that in 32-bit mode we need to constrain the input to the copy to ensure the subreg is valid. Otherwise we'll fail verify-machineinstrs

Differential Revision: https://reviews.llvm.org/D50656

llvm-svn: 339945

6 years agoFactor Node creation out of the demangler. No functionality change
Richard Smith [Thu, 16 Aug 2018 21:40:57 +0000 (21:40 +0000)]
Factor Node creation out of the demangler. No functionality change
intended.

llvm-svn: 339944

6 years agoEstablish the <bit> header. NFC yet. Reviewed as https://reviews.llvm.org/D50815
Marshall Clow [Thu, 16 Aug 2018 21:35:38 +0000 (21:35 +0000)]
Establish the <bit> header. NFC yet. Reviewed as https://reviews.llvm.org/D50815

llvm-svn: 339943

6 years ago[MC] Improve COFF associative section lookup
Reid Kleckner [Thu, 16 Aug 2018 21:34:41 +0000 (21:34 +0000)]
[MC] Improve COFF associative section lookup

Handle the case when the symbol is private. Private symbols are not in
the COFF object file symbol table, so they aren't inserted into
SymbolMap. We can't look up the section of the symbol that way. Instead,
get the MCSection from the MCSymbol and map that to the object file
section.

Print a better error message when the symbol has no section, like when
the symbol is undefined.

Fixes PR38607

llvm-svn: 339942

6 years agoUpdate for LLVM API change
David Blaikie [Thu, 16 Aug 2018 21:30:24 +0000 (21:30 +0000)]
Update for LLVM API change

llvm-svn: 339941

6 years ago[MI] Change the array of `MachineMemOperand` pointers to be
Chandler Carruth [Thu, 16 Aug 2018 21:30:05 +0000 (21:30 +0000)]
[MI] Change the array of `MachineMemOperand` pointers to be
a generically extensible collection of extra info attached to
a `MachineInstr`.

The primary change here is cleaning up the APIs used for setting and
manipulating the `MachineMemOperand` pointer arrays so chat we can
change how they are allocated.

Then we introduce an extra info object that using the trailing object
pattern to attach some number of MMOs but also other extra info. The
design of this is specifically so that this extra info has a fixed
necessary cost (the header tracking what extra info is included) and
everything else can be tail allocated. This pattern works especially
well with a `BumpPtrAllocator` which we use here.

I've also added the basic scaffolding for putting interesting pointers
into this, namely pre- and post-instruction symbols. These aren't used
anywhere yet, they're just there to ensure I've actually gotten the data
structure types correct. I'll flesh out support for these in
a subsequent patch (MIR dumping, parsing, the works).

Finally, I've included an optimization where we store any single pointer
inline in the `MachineInstr` to avoid the allocation overhead. This is
expected to be the overwhelmingly most common case and so should avoid
any memory usage growth due to slightly less clever / dense allocation
when dealing with >1 MMO. This did require several ergonomic
improvements to the `PointerSumType` to reasonably support the various
usage models.

This also has a side effect of freeing up 8 bits within the
`MachineInstr` which could be repurposed for something else.

The suggested direction here came largely from Hal Finkel. I hope it was
worth it. ;] It does hopefully clear a path for subsequent extensions
w/o nearly as much leg work. Lots of thanks to Reid and Justin for
careful reviews and ideas about how to do all of this.

Differential Revision: https://reviews.llvm.org/D50701

llvm-svn: 339940

6 years agoDebugInfo: Add metadata support for disabling DWARF pub sections
David Blaikie [Thu, 16 Aug 2018 21:29:55 +0000 (21:29 +0000)]
DebugInfo: Add metadata support for disabling DWARF pub sections

In cases where the debugger load time is a worthwhile tradeoff (or less
costly - such as loading from a DWP instead of a variety of DWOs
(possibly over a high-latency/distributed filesystem)) against object
file size, it can be reasonable to disable pubnames and corresponding
gdb-index creation in the linker.

A backend-flag version of this was implemented for NVPTX in
D44385/r327994 - which was fine for NVPTX which wouldn't mix-and-match
CUs. Now that it's going to be a user-facing option (likely powered by
"-gno-pubnames", the same as GCC) it should be encoded in the
DICompileUnit so it can vary per-CU.

After this, likely the NVPTX support should be migrated to the metadata
& the previous flag implementation should be removed.

Reviewers: aprantl

Differential Revision: https://reviews.llvm.org/D50213

llvm-svn: 339939

6 years agoadd a missed case for binary op FMF propagation under select folds
Michael Berg [Thu, 16 Aug 2018 20:59:45 +0000 (20:59 +0000)]
add a missed case for binary op FMF propagation under select folds

llvm-svn: 339938

6 years ago[AST] Speculative build fix for a polly buildbot
Philip Reames [Thu, 16 Aug 2018 20:58:48 +0000 (20:58 +0000)]
[AST] Speculative build fix for a polly buildbot

I don't have polly setup to bulld locally and don't plan to.  This should let the old API adapt to the new one.  Can someone from polly please migrate usage and then delete the wrapper?

llvm-svn: 339937

6 years ago[MemLoc] Fix a bug causing any use of invariant.end to crash in LICM
Philip Reames [Thu, 16 Aug 2018 20:48:55 +0000 (20:48 +0000)]
[MemLoc] Fix a bug causing any use of invariant.end to crash in LICM

The fix is fairly simple, but is says something unpleasant about the usage and testing of invariant.start/end scopes that this went undetected.  To put this in perspective, *any* invariant.end in a loop flowing through LICM crashed.  I haven't bothered to figure out just how far back this goes, but it's not caused by any of the recent changes.  We're probably talking months if not years.

llvm-svn: 339936

6 years ago[hwasan] Add __hwasan_handle_longjmp.
Evgeniy Stepanov [Thu, 16 Aug 2018 20:46:41 +0000 (20:46 +0000)]
[hwasan] Add __hwasan_handle_longjmp.

Summary:
A callback to annotate longjmp-like code.

Unlike __asan_handle_no_return, in hwasan we can not conservatively
"unpoison" the entire thread stack, because there is no such thing as
unpoisoned memory. Pointer and memory tags must always match.

Reviewers: vitalybuka, kcc

Subscribers: kubamracek, llvm-commits

Differential Revision: https://reviews.llvm.org/D50752

llvm-svn: 339935

6 years agoAMDGPU: Correct errors in device table
Matt Arsenault [Thu, 16 Aug 2018 20:19:47 +0000 (20:19 +0000)]
AMDGPU: Correct errors in device table

llvm-svn: 339934

6 years agoUpdate README and Dockerfile to include llvm-proto-fuzzer
Emmett Neyman [Thu, 16 Aug 2018 20:13:40 +0000 (20:13 +0000)]
Update README and Dockerfile to include llvm-proto-fuzzer

Summary: Added commands to Dockerfile to build llvm-proto-fuzzer and the other related tools. Also added a section to the bottom of the README describing what llvm-proto-fuzzer does and how to run it.

Reviewers: morehouse, kcc

Reviewed By: morehouse

Subscribers: cfe-commits, llvm-commits

Differential Revision: https://reviews.llvm.org/D50829

llvm-svn: 339933

6 years ago[hwasan] Add malloc_fill_byte and free_fill_byte flags.
Evgeniy Stepanov [Thu, 16 Aug 2018 20:13:09 +0000 (20:13 +0000)]
[hwasan] Add malloc_fill_byte and free_fill_byte flags.

Reviewers: vitalybuka, kcc

Subscribers: kubamracek, llvm-commits

Differential Revision: https://reviews.llvm.org/D50753

llvm-svn: 339932

6 years ago[SystemZ] Require asserts in subregliveness-06.mir
Krzysztof Parzyszek [Thu, 16 Aug 2018 20:12:15 +0000 (20:12 +0000)]
[SystemZ] Require asserts in subregliveness-06.mir

The option -misched=shuffle is only available with !NDEBUG builds.

llvm-svn: 339931

6 years ago[LICM][NFC] Restructure pointer invalidation API in terms of MemoryLocation
Philip Reames [Thu, 16 Aug 2018 20:11:15 +0000 (20:11 +0000)]
[LICM][NFC] Restructure pointer invalidation API in terms of MemoryLocation

Main value is just simplifying code.  I'll further simply the argument handling case in a bit, but that involved a slightly orthogonal change so I went with the mildy ugly intermediate for this patch.

Note that the isSized check in the old LICM code was not carried across.  It turns out that check was dead.  a) no test exercised it, and b) langref and verifier had been updated to disallow unsized types used in loads.

llvm-svn: 339930

6 years ago[dotest] Make --test-subdir work with --no-multiprocess
Vedant Kumar [Thu, 16 Aug 2018 19:56:38 +0000 (19:56 +0000)]
[dotest] Make --test-subdir work with --no-multiprocess

The single-process test runner is invoked in a number of different
scenarios, including when multiple test dirs are specified or (afaict)
when lit is used to drive the test suite.

Unfortunately the --test-subdir option did not work with the single
process test runner, breaking an important use case (using lit to run
swift-lldb Linux tests):

  Failure URL: https://ci.swift.org/job/swift-PR-Linux/6841

We won't be able to run lldb tests within swift PR testing without
filtering down the set of tests.

This change makes --test-subdir work with the single-process runner.

llvm-svn: 339929

6 years ago[llvm-mca] Fix -Wpessimizing-move warnings introduced by r339923.
Andrea Di Biagio [Thu, 16 Aug 2018 19:45:13 +0000 (19:45 +0000)]
[llvm-mca] Fix -Wpessimizing-move warnings introduced by r339923.

Reported by buildbot `clang-with-lto-ubuntu` ( build #9858 ).

llvm-svn: 339928

6 years agoAdd missing test file from r339799.
Peter Collingbourne [Thu, 16 Aug 2018 19:29:01 +0000 (19:29 +0000)]
Add missing test file from r339799.

llvm-svn: 339927

6 years ago[X86] Pre-commit test case for D50827.
Craig Topper [Thu, 16 Aug 2018 19:27:43 +0000 (19:27 +0000)]
[X86] Pre-commit test case for D50827.

llvm-svn: 339926

6 years ago[WebAssembly] Remove temporary workaround for function bitcasts
Jacob Gravelle [Thu, 16 Aug 2018 19:24:31 +0000 (19:24 +0000)]
[WebAssembly] Remove temporary workaround for function bitcasts

Summary:
EM_ASM no longer is lowered as varargs in C, so this workaround is
obsolete.

Reviewers: dschuff, sunfish

Subscribers: sbc100, aheejin, llvm-commits

Differential Revision: https://reviews.llvm.org/D50859

llvm-svn: 339925

6 years ago[MachineVerifier] Check if predecessor is jointly dominated by undefs
Krzysztof Parzyszek [Thu, 16 Aug 2018 19:13:28 +0000 (19:13 +0000)]
[MachineVerifier] Check if predecessor is jointly dominated by undefs

Each use of a value should be jointly dominated by the union of defs and
undefs. It can happen that it will only be jointly dominated by undefs,
and that is still legal. Make sure that the verifier is aware of that.

llvm-svn: 339924

6 years ago[llvm-mca] Refactor how execution is orchestrated by the Pipeline.
Andrea Di Biagio [Thu, 16 Aug 2018 19:00:48 +0000 (19:00 +0000)]
[llvm-mca] Refactor how execution is orchestrated by the Pipeline.

This patch changes how instruction execution is orchestrated by the Pipeline.
In particular, this patch makes it more explicit how instructions transition
through the various pipeline stages during execution.

The main goal is to simplify both the stage API and the Pipeline execution.  At
the same time, this patch fixes some design issues which are currently latent,
but that are likely to cause problems in future if people start defining custom
pipelines.

The new design assumes that each pipeline stage knows the "next-in-sequence".
The Stage API has gained three new methods:
 -   isAvailable(IR)
 -   checkNextStage(IR)
 -   moveToTheNextStage(IR).

An instruction IR can be executed by a Stage if method `Stage::isAvailable(IR)`
returns true.
Instructions can move to next stages using method moveToTheNextStage(IR).
An instruction cannot be moved to the next stage if method checkNextStage(IR)
(called on the current stage) returns false.
Stages are now responsible for moving instructions to the next stage in sequence
if necessary.

Instructions are allowed to transition through multiple stages during a single
cycle (as long as stages are available, and as long as all the calls to
`checkNextStage(IR)` returns true).

Methods `Stage::preExecute()` and `Stage::postExecute()` have now become
redundant, and those are removed by this patch.

Method Pipeline::runCycle() is now simpler, and it correctly visits stages
on every begin/end of cycle.

Other changes:
 - DispatchStage no longer requires a reference to the Scheduler.
 - ExecuteStage no longer needs to directly interact with the
   RetireControlUnit. Instead, executed instructions are now directly moved to the
   next stage (i.e. the retire stage).
 - RetireStage gained an execute method. This allowed us to remove the
   dependency with the RCU in ExecuteStage.
 - FecthStage now updates the "program counter" during cycleBegin() (i.e.
   before we start executing new instructions).
 - We no longer need Stage::Status to be returned by method execute(). It has
   been dropped in favor of a more lightweight llvm::Error.

Overally, I measured a ~11% performance gain w.r.t. the previous design.  I also
think that the Stage interface is probably easier to read now.  That being said,
code comments have to be improved, and I plan to do it in a follow-up patch.

Differential revision: https://reviews.llvm.org/D50849

llvm-svn: 339923

6 years ago[SelectionDAG] Improve the legalisation lowering of UMULO.
Eli Friedman [Thu, 16 Aug 2018 18:39:39 +0000 (18:39 +0000)]
[SelectionDAG] Improve the legalisation lowering of UMULO.

There is no way in the universe, that doing a full-width division in
software will be faster than doing overflowing multiplication in
software in the first place, especially given that this same full-width
multiplication needs to be done anyway.

This patch replaces the previous implementation with a direct lowering
into an overflowing multiplication algorithm based on half-width
operations.

Correctness of the algorithm was verified by exhaustively checking the
output of this algorithm for overflowing multiplication of 16 bit
integers against an obviously correct widening multiplication. Baring
any oversights introduced by porting the algorithm to DAG, confidence in
correctness of this algorithm is extremely high.

Following table shows the change in both t = runtime and s = space. The
change is expressed as a multiplier of original, so anything under 1 is
“better” and anything above 1 is worse.

+-------+-----------+-----------+-------------+-------------+
| Arch  | u64*u64 t | u64*u64 s | u128*u128 t | u128*u128 s |
+-------+-----------+-----------+-------------+-------------+
|   X64 |     -     |     -     |    ~0.5     |    ~0.64    |
|  i686 |   ~0.5    |   ~0.6666 |    ~0.05    |    ~0.9     |
| armv7 |     -     |   ~0.75   |      -      |    ~1.4     |
+-------+-----------+-----------+-------------+-------------+

Performance numbers have been collected by running overflowing
multiplication in a loop under `perf` on two x86_64 (one Intel Haswell,
other AMD Ryzen) based machines. Size numbers have been collected by
looking at the size of function containing an overflowing multiply in
a loop.

All in all, it can be seen that both performance and size has improved
except in the case of armv7 where code size has regressed for 128-bit
multiply. u128*u128 overflowing multiply on 32-bit platforms seem to
benefit from this change a lot, taking only 5% of the time compared to
original algorithm to calculate the same thing.

The final benefit of this change is that LLVM is now capable of lowering
the overflowing unsigned multiply for integers of any bit-width as long
as the target is capable of lowering regular multiplication for the same
bit-width. Previously, 128-bit overflowing multiply was the widest
possible.

Patch by Simonas Kazlauskas!

Differential Revision: https://reviews.llvm.org/D50310

llvm-svn: 339922

6 years ago[llvm-strip] Add support for -p/--preserve-dates
Jordan Rupprecht [Thu, 16 Aug 2018 18:29:40 +0000 (18:29 +0000)]
[llvm-strip] Add support for -p/--preserve-dates

Summary: [llvm-strip] Preserve access/modification timestamps when -p is used.

Reviewers: jakehehrlich, jhenderson, alexshap

Reviewed By: jhenderson

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D50744

llvm-svn: 339921

6 years agoFix lldb-vscode build on Windows
Reid Kleckner [Thu, 16 Aug 2018 18:24:59 +0000 (18:24 +0000)]
Fix lldb-vscode build on Windows

Include PosixAPI.h to get a PATH_MAX definition and replace CreateEvent
with CreateEventObject to avoid conflicts with the windows.h definition
of CreateEvent to CreateEventW.

llvm-svn: 339920

6 years ago[ASTImporter] Add test for member pointer types.
Raphael Isemann [Thu, 16 Aug 2018 18:22:21 +0000 (18:22 +0000)]
[ASTImporter] Add test for member pointer types.

Reviewers: a.sidorin, martong

Reviewed By: martong

Subscribers: rnkovacs, martong, cfe-commits

Differential Revision: https://reviews.llvm.org/D50792

llvm-svn: 339919

6 years ago[ASTImporter] Add test for importing CompoundAssignOperators
Raphael Isemann [Thu, 16 Aug 2018 18:21:33 +0000 (18:21 +0000)]
[ASTImporter] Add test for importing CompoundAssignOperators

Reviewers: a.sidorin, martong

Reviewed By: martong

Subscribers: rnkovacs, cfe-commits, martong

Differential Revision: https://reviews.llvm.org/D50793

llvm-svn: 339918

6 years ago[ASTImporter] Add test for DoStmt
Raphael Isemann [Thu, 16 Aug 2018 18:20:52 +0000 (18:20 +0000)]
[ASTImporter] Add test for DoStmt

Reviewers: a.sidorin, martong

Reviewed By: martong

Subscribers: rnkovacs, martong, cfe-commits

Differential Revision: https://reviews.llvm.org/D50810

llvm-svn: 339917

6 years ago[ASTImporter] Add test for WhileStmt
Raphael Isemann [Thu, 16 Aug 2018 18:20:05 +0000 (18:20 +0000)]
[ASTImporter] Add test for WhileStmt

Reviewers: a.sidorin, martong

Reviewed By: martong

Subscribers: rnkovacs, martong, cfe-commits

Differential Revision: https://reviews.llvm.org/D50811

llvm-svn: 339916

6 years ago[ASTImporter] Add test for IndirectGotoStmt
Raphael Isemann [Thu, 16 Aug 2018 18:19:21 +0000 (18:19 +0000)]
[ASTImporter] Add test for IndirectGotoStmt

Reviewers: a.sidorin, martong

Reviewed By: martong

Subscribers: rnkovacs, martong, cfe-commits

Differential Revision: https://reviews.llvm.org/D50813

llvm-svn: 339915

6 years agoDisable two flaky pexpect-backed tests on Darwin
Vedant Kumar [Thu, 16 Aug 2018 18:18:16 +0000 (18:18 +0000)]
Disable two flaky pexpect-backed tests on Darwin

These tests are sporadically timing out on our bots, e.g here:

  https://ci.swift.org/job/swift-PR-Linux/6841

llvm-svn: 339914

6 years agoUpdate LLD tests for CodeView dumper change in r339907
Reid Kleckner [Thu, 16 Aug 2018 18:03:06 +0000 (18:03 +0000)]
Update LLD tests for CodeView dumper change in r339907

llvm-svn: 339913

6 years ago[RegisterCoalescer] Shrink to uses if needed after removeCopyByCommutingDef
Krzysztof Parzyszek [Thu, 16 Aug 2018 18:02:59 +0000 (18:02 +0000)]
[RegisterCoalescer] Shrink to uses if needed after removeCopyByCommutingDef

llvm-svn: 339912

6 years agoAdd a new tool named "lldb-vscode" that implements the Visual Studio Code Debug Adapt...
Greg Clayton [Thu, 16 Aug 2018 17:59:38 +0000 (17:59 +0000)]
Add a new tool named "lldb-vscode" that implements the Visual Studio Code Debug Adaptor Protocol

This patch adds a new lldb-vscode tool that speaks the Microsoft Visual Studio Code debug adaptor protocol. It has full unit tests that test all packets.

This tool can be easily packaged up into a native extension and used with Visual Studio Code, and it can also be used by Nuclide

Differential Revision: https://reviews.llvm.org/D50365

llvm-svn: 339911

6 years agoUpdate LLDB for LLVM CodeView API change in r339907
Reid Kleckner [Thu, 16 Aug 2018 17:51:58 +0000 (17:51 +0000)]
Update LLDB for LLVM CodeView API change in r339907

llvm-svn: 339910

6 years agoFix memory leak in demangling of string literals.
Zachary Turner [Thu, 16 Aug 2018 17:48:32 +0000 (17:48 +0000)]
Fix memory leak in demangling of string literals.

llvm-svn: 339909

6 years ago[TargetLowering] Add support for non-uniform vectors to BuildSDIV
Simon Pilgrim [Thu, 16 Aug 2018 17:44:33 +0000 (17:44 +0000)]
[TargetLowering] Add support for non-uniform vectors to BuildSDIV

This patch refactors the existing TargetLowering::BuildSDIV base implementation to support non-uniform constant vector denominators.

This is the last patch necessary to close PR36545

Differential Revision: https://reviews.llvm.org/D50765

llvm-svn: 339908

6 years ago[codeview] Use push_macro to avoid conflicts instead of a prefix
Reid Kleckner [Thu, 16 Aug 2018 17:34:31 +0000 (17:34 +0000)]
[codeview] Use push_macro to avoid conflicts instead of a prefix

Summary:
This prefix was added in r333421, and it changed our dumper output to
say things like "CVRegEAX" instead of just "EAX". That's a functional
change that I'd rather avoid.

I tested GCC, Clang, and MSVC, and all of them support #pragma
push_macro. They don't issue warnings whem the macro is not defined
either.

I don't have a Mac so I can't test the real termios.h header, but I
looked at the termios.h sources online and looked for other conflicts.
I saw only the CR* macros, so those are the ones we work around.

Reviewers: zturner, JDevlieghere

Subscribers: hiraditya, llvm-commits

Differential Revision: https://reviews.llvm.org/D50851

llvm-svn: 339907

6 years ago[MC] Cleanup noop default case spelling. NFC.
Nirav Dave [Thu, 16 Aug 2018 17:22:31 +0000 (17:22 +0000)]
[MC] Cleanup noop default case spelling. NFC.

llvm-svn: 339906

6 years ago[ELF] mergeSections: remove non-alive MergeInputSection
Fangrui Song [Thu, 16 Aug 2018 17:22:02 +0000 (17:22 +0000)]
[ELF] mergeSections: remove non-alive MergeInputSection

Summary: This makes it conform to what the comment says. Otherwise when getErrPlace() is called afterwards, cast<InputSection>(D) will cause incompatible cast as MergeInputSection is not a subclass of InputSection.

Reviewers: ruiu, grimar, espindola, pcc

Reviewed By: grimar

Subscribers: emaste, arichardson, llvm-commits

Differential Revision: https://reviews.llvm.org/D50742

llvm-svn: 339904

6 years agoRevert "unittests: Don't install TestPlugin.so"
Tom Stellard [Thu, 16 Aug 2018 17:15:03 +0000 (17:15 +0000)]
Revert "unittests: Don't install TestPlugin.so"

This reverts commit r339897.

This breaks the build on Windows and platforms where loadable modules
aren't supported.

llvm-svn: 339903

6 years agoAMDGPU: Custom lower fexp
Matt Arsenault [Thu, 16 Aug 2018 17:07:52 +0000 (17:07 +0000)]
AMDGPU: Custom lower fexp

This will allow the library to just use __builtin_expf directly
without expanding this itself. Note f64 still won't work because
there is no exp instruction for it.

llvm-svn: 339902

6 years ago[X86][SSE] Add sdiv by nonuniform constant vector test containing -1/+1 and all-bits...
Simon Pilgrim [Thu, 16 Aug 2018 17:07:41 +0000 (17:07 +0000)]
[X86][SSE] Add sdiv by nonuniform constant vector test containing -1/+1 and all-bits style constants

llvm-svn: 339901

6 years ago[NFC] Fix typo in test cases
Evandro Menezes [Thu, 16 Aug 2018 17:03:22 +0000 (17:03 +0000)]
[NFC] Fix typo in test cases

llvm-svn: 339900

6 years agoNFC: Test commit access
Jordan Rupprecht [Thu, 16 Aug 2018 16:55:07 +0000 (16:55 +0000)]
NFC: Test commit access

Testing commit access from a new machine, so using this as an opportunity to revert my old test access commit (r336006) that I never cleaned up.

llvm-svn: 339899

6 years ago[TargetLowering] Refactor BuildSDIV in preparation for D50765. NFCI.
Simon Pilgrim [Thu, 16 Aug 2018 16:54:06 +0000 (16:54 +0000)]
[TargetLowering] Refactor BuildSDIV in preparation for D50765. NFCI.

Pull out magic factor calculators into a helper function, use 0/+1/-1 multiplication factor to (optionally) add/sub the numerator.

llvm-svn: 339898

6 years agounittests: Don't install TestPlugin.so
Tom Stellard [Thu, 16 Aug 2018 16:53:06 +0000 (16:53 +0000)]
unittests: Don't install TestPlugin.so

Summary:
add_llvm_loadable_module adds an install target by default, but this
module is only used for a unit test, so we don't need to install it.

Reviewers: philip.pfaffe, thakis

Subscribers: mgorny, llvm-commits

Differential Revision: https://reviews.llvm.org/D50668

llvm-svn: 339897

6 years ago[MC] Remove unused variable
Benjamin Kramer [Thu, 16 Aug 2018 16:50:23 +0000 (16:50 +0000)]
[MC] Remove unused variable

llvm-svn: 339896

6 years ago[MC][X86] Enhance X86 Register expression handling to more closely match GCC.
Nirav Dave [Thu, 16 Aug 2018 16:31:14 +0000 (16:31 +0000)]
[MC][X86] Enhance X86 Register expression handling to more closely match GCC.

Allow the comparison of x86 registers in the evaluation of assembler
directives. This generalizes and simplifies the extension from r334022
to catch another case found in the Linux kernel.

Reviewers: rnk, void

Reviewed By: rnk

Subscribers: hiraditya, nickdesaulniers, llvm-commits

Differential Revision: https://reviews.llvm.org/D50795

llvm-svn: 339895

6 years agoFix -Wmicrosoft-goto warnings.
Zachary Turner [Thu, 16 Aug 2018 16:30:27 +0000 (16:30 +0000)]
Fix -Wmicrosoft-goto warnings.

llvm-svn: 339894

6 years agoAdd support for AVX-512 CodeView registers.
Zachary Turner [Thu, 16 Aug 2018 16:17:55 +0000 (16:17 +0000)]
Add support for AVX-512 CodeView registers.

When compiling with /arch:AVX512 and optimizations turned on,
we could crash while emitting debug info because we did not
have CodeView register constants for the AVX 512 register
set defined.  This patch defines them.

Differential Revision: https://reviews.llvm.org/D50819

llvm-svn: 339893

6 years ago[MS Demangler] Demangle string literals.
Zachary Turner [Thu, 16 Aug 2018 16:17:36 +0000 (16:17 +0000)]
[MS Demangler] Demangle string literals.

When demangling string literals, Microsoft's undname
simply prints 'string'.  This patch implements string
literal demangling while doing a bit better than this
by decoding as much of the string as possible and
trying to faithfully reproduce the original string
literal definition.

This is a bit tricky because the different character
types char, char16_t, and char32_t are not uniquely
identified by the mangling, so we have to use a
heuristic to try to guess the character type.  But
it works pretty well, and many tests are added to
illustrate the behavior.

Differential Revision: https://reviews.llvm.org/D50806

llvm-svn: 339892

6 years ago[MS Demangler] Don't fail on MD5-mangled names.
Zachary Turner [Thu, 16 Aug 2018 16:17:17 +0000 (16:17 +0000)]
[MS Demangler] Don't fail on MD5-mangled names.

When we have an MD5 mangled name, we shouldn't choke and say
that it's an invalid name.  Even though it's impossible to demangle,
we should just output the original name.

llvm-svn: 339891

6 years ago[TableGen] TypeSetByHwMode::operator== optimization
Simon Pilgrim [Thu, 16 Aug 2018 16:16:28 +0000 (16:16 +0000)]
[TableGen] TypeSetByHwMode::operator== optimization

This operator is called a great deal, by checking for the cheap isSimple equality cases first (a common occurrence) we can improve performance as we avoid a lot of std::map find/iteration in hasDefault.

isSimple also means that a default value is present, so we can avoid some hasDefault calls.

This also avoids a rather dodgy piece of logic that was checking for isSimple() && !VTS.isSimple() but not the inverse - it now uses the general hasDefault mode comparison test instead.

Saves around 15secs in debug builds of x86 -gen-dag-isel.

Differential Revision: https://reviews.llvm.org/D50841

llvm-svn: 339890

6 years ago[ConstantFolding] add tests for funnel shift intrinsics; NFC
Sanjay Patel [Thu, 16 Aug 2018 16:10:42 +0000 (16:10 +0000)]
[ConstantFolding] add tests for funnel shift intrinsics; NFC

No functionality for this yet.

llvm-svn: 339889

6 years ago[TableGen] Avoid self getPredicates() != comparison. NFCI.
Simon Pilgrim [Thu, 16 Aug 2018 16:04:05 +0000 (16:04 +0000)]
[TableGen] Avoid self getPredicates() != comparison. NFCI.

We were performing a completely unnecessary full comparison of the same std::vector<Predicate>.

llvm-svn: 339888

6 years ago[InstCombine] Expand the simplification of pow(x, 0.5) to sqrt(x)
Evandro Menezes [Thu, 16 Aug 2018 15:58:08 +0000 (15:58 +0000)]
[InstCombine] Expand the simplification of pow(x, 0.5) to sqrt(x)

Expand the number of cases when `pow(x, 0.5)` is simplified into `sqrt(x)`
by considering the math semantics with more granularity.

Differential revision: https://reviews.llvm.org/D50036

llvm-svn: 339887

6 years ago[Sanitizer] Fix build openbsd
David Carlier [Thu, 16 Aug 2018 15:54:38 +0000 (15:54 +0000)]
[Sanitizer] Fix build openbsd

- The alternative syscall exists only on FreeBSD.
- Adding OpenBSD's equivalent and while at it other
remaining oses ones.

Reviewers: vitalybuka, krytarowsky, hiraditya

Reviewed By: hiraditya

Differential Revision: https://reviews.llvm.org/D50760

llvm-svn: 339886

6 years ago[llvm-mca] Small refactoring in preparation for another patch that will improve the...
Andrea Di Biagio [Thu, 16 Aug 2018 15:43:09 +0000 (15:43 +0000)]
[llvm-mca] Small refactoring in preparation for another patch that will improve the modularity of the Pipeline. NFCI

The main difference is that now `cycleStart()` and `cycleEnd()` return an
llvm::Error.

This patch implements a few minor style changes, and adds missing 'const' to
some methods.

llvm-svn: 339885

6 years ago[TableGen] Return ValueTypeByHwMode by const reference from CodeGenRegisterClass...
Simon Pilgrim [Thu, 16 Aug 2018 15:29:24 +0000 (15:29 +0000)]
[TableGen] Return ValueTypeByHwMode by const reference from CodeGenRegisterClass::getValueTypeNum

Avoids costly std::map copies inside ValueTypeByHwMode constructor

llvm-svn: 339884

6 years ago[cmake] Prevent LLVMgold.so from being unloaded on Linux
Hans Wennborg [Thu, 16 Aug 2018 15:12:12 +0000 (15:12 +0000)]
[cmake] Prevent LLVMgold.so from being unloaded on Linux

Extend the fix from D40459 to also apply to modules such as the LLVM
gold plugin. This is needed because current binutils master (and future
binutils 2.32) calls dlclose() on bfd plugins as part of a recent fix
for https://sourceware.org/bugzilla/show_bug.cgi?id=23460.

Patch by Evangelos Foutras!

Differential Revision: https://reviews.llvm.org/D50416

llvm-svn: 339883

6 years ago[LLD][ELF] - Add a test case for DT_SONAME entry reading.
George Rimar [Thu, 16 Aug 2018 14:07:29 +0000 (14:07 +0000)]
[LLD][ELF] - Add a test case for DT_SONAME entry reading.

This covers a following line with a test:
https://github.com/llvm-mirror/lld/blob/master/ELF/InputFiles.cpp#L899

llvm-svn: 339880

6 years agoFixed unused variable warning. NFCI.
Simon Pilgrim [Thu, 16 Aug 2018 13:55:10 +0000 (13:55 +0000)]
Fixed unused variable warning. NFCI.

llvm-svn: 339879

6 years ago[libunwind][mips] Include gcc_s for linkage
Stefan Maksimovic [Thu, 16 Aug 2018 13:40:16 +0000 (13:40 +0000)]
[libunwind][mips] Include gcc_s for linkage

When compiling with optimizations, mips requires various helper routines(__ashldi3 and the like) contained in libgcc_s.
Conditionally include libgcc_s in the set of libraries to be linked to.

Differential Revision: https://reviews.llvm.org/D50243

llvm-svn: 339878

6 years ago[clangd] NFC: Improve Dex Iterators debugging traits
Kirill Bobyrev [Thu, 16 Aug 2018 13:19:43 +0000 (13:19 +0000)]
[clangd] NFC: Improve Dex Iterators debugging traits

This patch improves `dex::Iterator` string representation by
incorporating the information about the element which is currently being
pointed to by the `DocumentIterator`.

Reviewed by: ioeric

Differential Revision: https://reviews.llvm.org/D50689

llvm-svn: 339877

6 years ago[LLD][ELF] - Add a test case.
George Rimar [Thu, 16 Aug 2018 13:02:50 +0000 (13:02 +0000)]
[LLD][ELF] - Add a test case.

This covers the following line with a test:
https://github.com/llvm-mirror/lld/blob/master/ELF/InputFiles.cpp#L487

llvm-svn: 339876

6 years ago[InstCombine] move vector compare before same-shuffled ops
Sanjay Patel [Thu, 16 Aug 2018 12:52:17 +0000 (12:52 +0000)]
[InstCombine] move vector compare before same-shuffled ops

This is a step towards fixing PR37463:
https://bugs.llvm.org/show_bug.cgi?id=37463

llvm-svn: 339875

6 years ago[libcxx] By default, do not use internal_linkage to hide symbols from the ABI
Louis Dionne [Thu, 16 Aug 2018 12:44:28 +0000 (12:44 +0000)]
[libcxx] By default, do not use internal_linkage to hide symbols from the ABI

Summary:
https://reviews.llvm.org/D49240 led to symbol size problems in Chromium, and
we expect this may be the case in other projects built in debug mode too.
Instead, unless users explicitly ask for internal_linkage, we use always_inline
like we used to.

In the future, when we have a solution that allows us to drop always_inline
without falling back on internal_linkage, we can replace always_inline by
that.

Note that this commit introduces a change in contract for existing libc++
users: by default, libc++ used to guarantee that TUs built with different
versions of libc++ could be linked together. With the introduction of the
_LIBCPP_HIDE_FROM_ABI_PER_TU macro, the default behavior is that TUs built
with different libc++ versions are not guaranteed to link. This is a change
in contract but not a change in behavior, since the current implementation
still allows linking TUs built with different libc++ versions together.

Reviewers: EricWF, mclow.lists, dexonsmith, hans, rnk

Subscribers: christof, cfe-commits

Differential Revision: https://reviews.llvm.org/D50652

llvm-svn: 339874

6 years ago[yaml2obj] - Allow to use numeric sh_link (Link) value for sections.
George Rimar [Thu, 16 Aug 2018 12:44:17 +0000 (12:44 +0000)]
[yaml2obj] - Allow to use numeric sh_link (Link) value for sections.

That change allows using numeric values for Link field.
It is consistent with the code for another fields in this method.

llvm-svn: 339873