platform/upstream/llvm.git
4 years agoAdd way to omit debug-location from MIR output
Daniel Sanders [Mon, 6 Apr 2020 17:57:21 +0000 (10:57 -0700)]
Add way to omit debug-location from MIR output

Summary:
In lieu of a proper pass that strips debug info, add a way
to omit debug-locations from the MIR output so that
instructions with MMO's continue to match CHECK's when
mir-debugify is used

Reviewers: aprantl, bogner, vsk

Subscribers: hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D77575

4 years ago[CallSite Removal] a CallBase is never an IndirectCall for isInlineAsm
Nick Desaulniers [Mon, 6 Apr 2020 23:14:40 +0000 (16:14 -0700)]
[CallSite Removal] a CallBase is never an IndirectCall for isInlineAsm

Summary:
Thanks to Bill Wendling (void) for the report and steps to reproduce.  It looks
like this was missed during r350508's cleanup of the CallSite split into
CallBase, CallInst, and CallBrInst.

This was exposed by running pgo on a callbr, which was creating a ptrtoint to
the inline asm thinking it was an indirect call. The relevant callchain looks
like:

    IndirectCallPromotionPlugin::run()
    -> PGOIndirectCallVisitor::findIndirectCalls()
      -> PGOIndirectCallVisitor::visitCallBase()
        -> CallBase::isIndirectCall()

Reviewers: void, chandlerc

Reviewed By: void

Subscribers: hiraditya, llvm-commits, craig.topper, srhines

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D77600

4 years ago[lldb/API] Add missing LLDB_REGISTER_METHOD macros
Jonas Devlieghere [Mon, 6 Apr 2020 23:09:09 +0000 (16:09 -0700)]
[lldb/API] Add missing LLDB_REGISTER_METHOD macros

Add LLDB_REGISTER_METHOD macros for GetRetriesWithFixIts and
SetRetriesWithFixIts.

4 years agoRemove case in rewritergen unmatched opdefgen side
Jacques Pienaar [Mon, 6 Apr 2020 22:55:25 +0000 (15:55 -0700)]
Remove case in rewritergen unmatched opdefgen side

The rewriter generates a call to build that is not handled by opdef generator
and so will fail to compile. Also if this is a root node being replaced
(depth 0) then using the more generic build method in the rewrite suffices.

4 years ago[AddressSanitizer] Fix for wrong argument values appearing in backtraces
Vedant Kumar [Tue, 31 Mar 2020 22:27:06 +0000 (15:27 -0700)]
[AddressSanitizer] Fix for wrong argument values appearing in backtraces

Summary:
In some cases, ASan may insert instrumentation before function arguments
have been stored into their allocas. This causes two issues:

1) The argument value must be spilled until it can be stored into the
   reserved alloca, wasting a stack slot.

2) Until the store occurs in a later basic block, the debug location
   will point to the wrong frame offset, and backtraces will show an
   uninitialized value.

The proposed solution is to move instructions which initialize allocas
for arguments up into the entry block, before the position where ASan
starts inserting its instrumentation.

For the motivating test case, before the patch we see:

```
 | 0033: movq %rdi, 0x68(%rbx)  |   | DW_TAG_formal_parameter     |
 | ...                          |   |   DW_AT_name ("a")          |
 | 00d1: movq 0x68(%rbx), %rsi  |   |   DW_AT_location (RBX+0x90) |
 | 00d5: movq %rsi, 0x90(%rbx)  |   |       ^ not correct ...     |
```

and after the patch we see:

```
 | 002f: movq %rdi, 0x70(%rbx)  |   | DW_TAG_formal_parameter     |
 |                              |   |   DW_AT_name ("a")          |
 |                              |   |   DW_AT_location (RBX+0x70) |
```

rdar://61122691

Reviewers: aprantl, eugenis

Subscribers: hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D77182

4 years agofix comment typo to cycle bots
Nico Weber [Mon, 6 Apr 2020 22:58:16 +0000 (18:58 -0400)]
fix comment typo to cycle bots

4 years agoRemap the target (Xcode) SDK directory to the host SDK directory.
Adrian Prantl [Fri, 20 Mar 2020 01:51:36 +0000 (18:51 -0700)]
Remap the target (Xcode) SDK directory to the host SDK directory.

This is mostly useful for Swift support; it allows LLDB to substitute
a matching SDK it shipped with instead of the sysroot path that was
used at compile time.

The goal of this is to make the Xcode SDK something that behaves more
like the compiler's resource directory, as in that it ships with LLDB
rather than with the debugged program. This important primarily for
importing Swift and Clang modules in the expression evaluator, and
getting at the APINotes from the SDK in Swift.

For a cross-debugging scenario, this means you have to have an SDK for
your target installed alongside LLDB. In Xcode this will always be the
case.

rdar://problem/60640017

Differential Revision: https://reviews.llvm.org/D76471

4 years ago[libc++] 2/N: Enable the new libc++ testing format by default
Louis Dionne [Mon, 6 Apr 2020 22:30:19 +0000 (18:30 -0400)]
[libc++] 2/N: Enable the new libc++ testing format by default

Both test formats are equivalent, so this *should* not be a problem.
We've fixed a couple of failures uncovered by the first time we tried
making the switch, so this new attempt should go even farther.

If failures are noticed, it should be fine to revert this commit, but
please give a heads up afterwards so we know to address the issues!

Also note that it is still possible to use the old format by passing
`--param=use_old_format=True` when running Lit for the time being.

4 years agoSupport bfdname "elf32-hexagon".
Sid Manning [Mon, 6 Apr 2020 17:40:19 +0000 (12:40 -0500)]
Support bfdname "elf32-hexagon".

Add support and update testcases.

Differential Revision: https://reviews.llvm.org/D77579

4 years agoAllow MachineFunction to obtain non-const Function (to enable MIR-level debugify)
Daniel Sanders [Fri, 3 Apr 2020 22:55:15 +0000 (15:55 -0700)]
Allow MachineFunction to obtain non-const Function (to enable MIR-level debugify)

Summary:
To debugify MIR, we need to be able to create metadata and to do that, we
need a non-const Module. However, MachineFunction only had a const reference
to the Function preventing this.

Reviewers: aprantl, bogner

Subscribers: hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D77439

4 years ago[libc++] Only support std::to_chars availability test when back-deploying
Louis Dionne [Mon, 6 Apr 2020 22:09:34 +0000 (18:09 -0400)]
[libc++] Only support std::to_chars availability test when back-deploying

4 years ago[libc++] Add availability markup for the C++20 Synchronization Library on Apple
Louis Dionne [Mon, 24 Feb 2020 15:09:29 +0000 (10:09 -0500)]
[libc++] Add availability markup for the C++20 Synchronization Library on Apple

4 years agoAdd option to limit Debugify to locations (omitting variables)
Daniel Sanders [Fri, 3 Apr 2020 22:50:11 +0000 (15:50 -0700)]
Add option to limit Debugify to locations (omitting variables)

Summary:
It can be helpful to test behaviour w.r.t locations without having DEBUG_VALUE
around. In particular, because DEBUG_VALUE has the potential to change CodeGen
behaviour (e.g. hasOneUse() vs hasOneNonDbgUse()) while locations generally
don't.

Reviewers: aprantl, bogner

Subscribers: hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D77438

4 years ago[gdb-remote] Moving prevents copy elision. Found by clang.
Davide Italiano [Mon, 6 Apr 2020 21:59:01 +0000 (14:59 -0700)]
[gdb-remote] Moving prevents copy elision. Found by clang.

4 years agoRevert "Make llvm_source_root in llvm-lit relative too."
Hubert Tong [Mon, 6 Apr 2020 21:29:10 +0000 (17:29 -0400)]
Revert "Make llvm_source_root in llvm-lit relative too."

This reverts commit 6c1a9fb174ac3cf66bfdbab5a327957a0a35ad85.

Commit causes failures:
http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-autoconf/builds/49206

4 years ago[lldb/Test] Enable TestGdbRemoteThreadsInfoMemory.py on Windows.
Jonas Devlieghere [Mon, 6 Apr 2020 21:28:27 +0000 (14:28 -0700)]
[lldb/Test] Enable TestGdbRemoteThreadsInfoMemory.py on Windows.

This test is currently XFAILed but is passing on the Windows bot.

4 years agoX86ISelLowering: Minor refactor to avoid redundant initialization while ensuring...
David Blaikie [Mon, 6 Apr 2020 21:24:31 +0000 (14:24 -0700)]
X86ISelLowering: Minor refactor to avoid redundant initialization while ensuring compiler warnings can hopefully still prove initialization

Based on post-commit review/discussion in fabe52a7412b

4 years ago[clang-format] Don't break multi block parameters on ObjCBreakBeforeNestedBlockParam
Kanglei Fang [Mon, 6 Apr 2020 19:45:38 +0000 (12:45 -0700)]
[clang-format] Don't break multi block parameters on ObjCBreakBeforeNestedBlockParam

Summary:
While [the original diff](https://reviews.llvm.org/D42493) makes a lot of sense, and multiple inline block parameter/trailing paramemter after inline block paramemter should be discouraged, the formatting result is different than what xcode does by default
For the exact same example provided in the original diff:
```
[object
  blockArgument:^{
    a = 42;
  }
     anotherArg:42];
```
The code is hard to read and not very visually pleasing

This diff uses `ObjCBreakBeforeNestedBlockParam` to shield from the formatting
When it's set to false, don't allign the inline block paramemters.

Reviewers: jolesiak, benhamilton, jinlin

Reviewed By: jolesiak

Subscribers: cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D77039

4 years agoAdd more detailed symbol type categorization, based on a swift patch by
Jason Molenda [Fri, 3 Apr 2020 05:10:00 +0000 (22:10 -0700)]
Add more detailed symbol type categorization, based on a swift patch by
Greg Clayton a few years ago.

My patch to augment the symbol table in Mach-O files with the
dyld trie exports data structure only categorized symbols as code
or data, but Greg Clayton had a patch to do something similar to
swift a few years ago that had a more extensive categorization of
symbols, as well as extracting some objc class/ivar names from the
entries. This patch is basically just Greg's, updated a bit and
with a test case added to it.

<rdar://problem/50791451>

Differential Revision: https://reviews.llvm.org/D77369

4 years ago[AMDGPU] Disable 'Skip Uniform Regions' optimization by default for AMDGPU.
Konstantin Pyzhov [Mon, 6 Apr 2020 13:05:58 +0000 (09:05 -0400)]
[AMDGPU] Disable 'Skip Uniform Regions' optimization by default for AMDGPU.

Reviewers: sameerds, dstuttard

Differential Revision: https://reviews.llvm.org/D77228

4 years ago[llvm-objdump][NFC] Declare command-line externs in headers with namespace
Hubert Tong [Mon, 6 Apr 2020 20:56:13 +0000 (16:56 -0400)]
[llvm-objdump][NFC] Declare command-line externs in headers with namespace

Summary:
This patch moves the forward declarations of command-line `cl::*`
externs in `MachODump.cpp` and `llvm-objdump.cpp` into the headers
corresponding to the file that defines the variable. At the same time,
these externs are moved into the `llvm::objdump` namespace. The externs
that are not referenced outside their defining translation unit are made
static.

This does not factor out uses of the Mach-O options from
`llvm-objdump.cpp`.

Reviewers: jhenderson, MaskRay, DiggerLin, jasonliu, daltenty

Reviewed By: jhenderson, MaskRay

Subscribers: rupprecht, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D77388

4 years ago[gn build] (manually) port 6c1a9fb174ac3c
Nico Weber [Mon, 6 Apr 2020 20:50:02 +0000 (16:50 -0400)]
[gn build] (manually) port 6c1a9fb174ac3c

4 years ago[AsmPrinter] Do not define local aliases for global objects in a comdat
Leonard Chan [Fri, 3 Apr 2020 20:29:14 +0000 (13:29 -0700)]
[AsmPrinter] Do not define local aliases for global objects in a comdat

A global symbol that is defined in a comdat should not generate an alias since
call sites that would've referred to that symbol will refer to their own
independent local aliases rather than the surviving global comdat one. This
could result in something that looks like:

```
ld.lld: error: relocation refers to a discarded section: .text._ZN3fbl8internal18NullFunctionTargetIvJjjPjEED1Ev.stub
>>> defined in user-x64-clang/obj/system/ulib/minfs/libminfs.a(minfs._sources.file.cc.o)
>>> section group signature: _ZN3fbl8internal18NullFunctionTargetIvJjjPjEED1Ev.stub
>>> prevailing definition is in user-x64-clang/obj/system/ulib/minfs/libminfs.a(minfs._sources.vnode.cc.o)
>>> referenced by function.h:169 (../../zircon/system/ulib/fbl/include/fbl/function.h:169)
>>>               minfs._sources.file.cc.o:(minfs::File::AllocateAndCommitData(std::__2::unique_ptr<minfs::Transaction, std::__2::default_delete<minfs::Transaction> >)) in archive user-x64-clang/obj/system/ulib/minfs/libminfs.a
```

We ran into this when experimenting with a new C++ ABI for fuchsia
(refer to D72959) which takes relative offsets between comdat'd functions
which is why the normal C++ user wouldn't run into this.

Differential Revision: https://reviews.llvm.org/D77429

4 years agoMake llvm_source_root in llvm-lit relative too.
Nico Weber [Mon, 6 Apr 2020 20:46:24 +0000 (16:46 -0400)]
Make llvm_source_root in llvm-lit relative too.

No intended behavior change.

4 years ago[SelectionDAG] fix predecessor list for INLINEASM_BRs' parent
Nick Desaulniers [Mon, 6 Apr 2020 20:34:06 +0000 (13:34 -0700)]
[SelectionDAG] fix predecessor list for INLINEASM_BRs' parent

Summary:
A bug report mentioned that LLVM was producing jumps off the end of a
function when using "asm goto with outputs". Further digging pointed to
MachineBasicBlocks that had their address taken and were indirect
targets of INLINEASM_BR being removed by BranchFolder, because their
 predecessor list was empty, so they appeared to have no entry.

This was a cascading failure caused earlier, during Pre-RA instruction
scheduling. We have a few special cases in Pre-RA instruction scheduling
where we split a MachineBasicBlock in two.  This requires careful
handing of predecessor and successor lists for a MachineBasicBlock that
was split, and careful handing of PHI MachineInstrs that referred to the
MachineBasicBlock before it was split.

The clue that led to this fix was the observation that many callers of
MachineBasicBlock::splice() frequently call
MachineBasicBlock::transferSuccessorsAndUpdatePHIs() to update their PHI
nodes after a splice. We don't want to reuse that method, as we have
custom successor transferring logic for this block split.

This patch fixes 2 pre-existing bugs, and adds tests.

The first bug was that MachineBasicBlock::splice() correctly handles
updating most successors and predecessors; we don't need to do anything
more than removing the previous fallthrough block from the first half of
the split block post splice. Previously, we were updating the successor
list incorrectly (updating successors updates predecessors).

The second bug was that PHI nodes that needed registers from the first
half of the split block were not having entries populated.  The register
live out information was correct, and the FuncInfo->PHINodesToUpdate was
correct. Specifically, the check in SelectionDAGISel::FinishBasicBlock:

    for (unsigned i = 0, e = FuncInfo->PHINodesToUpdate.size(); i != e; ++i) {
      MachineInstrBuilder PHI(*MF, FuncInfo->PHINodesToUpdate[i].first);
      if (!FuncInfo->MBB->isSuccessor(PHI->getParent()))
        continue;
      PHI.addReg(FuncInfo->PHINodesToUpdate[i].second).addMBB(FuncInfo->MBB);

was `continue`ing because FuncInfo->MBB tracks the second half of
the post-split block; no one was updating PHI entries for the first half
of the post-split block.

SelectionDAGBuilder::UpdateSplitBlock() already expects to perform
special handling for MachineBasicBlocks that were split post calls to
ScheduleDAGSDNodes::EmitSchedule(), so I'm confident that it's both
correct for ScheduleDAGSDNodes::EmitSchedule() to return the second half
of the split block `CopyBB` which updates `FuncInfo->MBB` (ie. the
current MachineBasicBlock being processed), and perform special handling
for this in SelectionDAGBuilder::UpdateSplitBlock().

Reviewers: void, craig.topper, efriedma

Reviewed By: void, efriedma

Subscribers: hfinkel, fhahn, MatzeB, efriedma, hiraditya, llvm-commits, srhines

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D76961

4 years agoUpstream Bionic definitions of ctype_base/regex.
Dan Albert [Mon, 6 Apr 2020 20:34:27 +0000 (13:34 -0700)]
Upstream Bionic definitions of ctype_base/regex.

Summary:
This is a patch that Android has been carrying in its tree for several
years. This patch upstreams the existing ABI.

There's some historical cruft here. __regex_word used to be a part of
regex_traits rather than ctype_base. Bionic also used to use its own
ctype implementation because the libc++ builtin one wasn't available
yet. Bionic's ctype masks were 8 bits wide and already saturated, so a
wider type needed to be used for the regex mask, and the existing
value was already used so Android needed to specify its own.

Since then Android has migrated to the builtin ctype implementation
and this patch probably should have been dropped then. Unfortunately
that was not noticed at the time, so now we need to keep this to
maintain the current ABI.

Reviewers: EricWF, #libc, ldionne

Reviewed By: #libc, ldionne

Subscribers: dexonsmith, ldionne, libcxx-commits

Tags: #libc

Differential Revision: https://reviews.llvm.org/D76171

4 years ago[debugserver] Get rid of `else` after `return`. NFC.
Davide Italiano [Thu, 2 Apr 2020 20:22:52 +0000 (13:22 -0700)]
[debugserver] Get rid of `else` after `return`. NFC.

4 years ago[libunwind] Support the new libc++ test format
Sergej Jaskiewicz [Mon, 6 Apr 2020 20:26:37 +0000 (23:26 +0300)]
[libunwind] Support the new libc++ test format

Reviewers: ldionne, #libunwind, mstorsjo

Reviewed By: ldionne, #libunwind, mstorsjo

Subscribers: mstorsjo, dexonsmith, llvm-commits, libcxx-commits

Tags: #llvm, #libunwind

Differential Revision: https://reviews.llvm.org/D77501

4 years agoMinor typo improvements in documentation, NFC.
Chris Lattner [Mon, 6 Apr 2020 20:18:49 +0000 (13:18 -0700)]
Minor typo improvements in documentation, NFC.

4 years agoAMDGPU: Remove dead paths for requiresUniformRegister
Matt Arsenault [Thu, 6 Feb 2020 17:02:51 +0000 (12:02 -0500)]
AMDGPU: Remove dead paths for requiresUniformRegister

The extracts from control flow intrinsics are already properly handled
by divergence analysis. The inline asm case isn't dead, but has also
never really worked correctly so leave it as-is for now.

4 years ago[CodeGenObjC] Fix a crash when attempting to copy a zero-sized bit-field in a non...
Erik Pilkington [Mon, 6 Apr 2020 14:17:30 +0000 (10:17 -0400)]
[CodeGenObjC] Fix a crash when attempting to copy a zero-sized bit-field in a non-trivial C struct

Zero sized bit-fields aren't included in the CGRecordLayout, so we shouldn't be
calling EmitLValueForField for them. rdar://60695105

Differential revision: https://reviews.llvm.org/D76782

4 years ago[clangd] Fix bad include
Kadir Cetinkaya [Mon, 6 Apr 2020 19:36:43 +0000 (21:36 +0200)]
[clangd] Fix bad include

4 years agoDelete a variable that's been unused since r313407.
Nico Weber [Mon, 6 Apr 2020 19:25:19 +0000 (15:25 -0400)]
Delete a variable that's been unused since r313407.

No intended behavior change.

4 years ago[gn build] Port 427c1dc4f42
LLVM GN Syncbot [Mon, 6 Apr 2020 19:22:43 +0000 (19:22 +0000)]
[gn build] Port 427c1dc4f42

4 years ago[clangd] Build ASTs only with fresh preambles or after building a new preamble
Kadir Cetinkaya [Sun, 15 Mar 2020 20:43:00 +0000 (21:43 +0100)]
[clangd] Build ASTs only with fresh preambles or after building a new preamble

Summary:
This is another step for out-of-order preamble builds. To keep the
diagnostic behavior same, we only build ASTs either with "usable" preambles,
the ones that are fully applicable to a given ParseInput, or after building a
new preamble. Which is the same behaviour as what we do today. ASTs
built in the latter is called golden ASTs.

Reviewers: sammccall

Subscribers: ilya-biryukov, javed.absar, MaskRay, jkorous, arphaman, usaxena95, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D76725

4 years ago[clangd] Update TUStatus api to accommodate preamble thread
Kadir Cetinkaya [Tue, 17 Mar 2020 18:08:23 +0000 (19:08 +0100)]
[clangd] Update TUStatus api to accommodate preamble thread

Summary:
TUStatus api had a single thread in mind. This introudces a section
action to represent state of the preamble thread. In the file status extension,
we keep old behavior almost the same. We only prepend current task with a
`parsing includes` if preamble thread is working. We omit the idle thread in the
output unless both threads are idle.

Reviewers: sammccall

Subscribers: ilya-biryukov, javed.absar, MaskRay, jkorous, arphaman, usaxena95, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D76304

4 years ago[clangd] Decouple preambleworker from astworker, NFCI
Kadir Cetinkaya [Fri, 13 Mar 2020 10:52:19 +0000 (11:52 +0100)]
[clangd] Decouple preambleworker from astworker, NFCI

Summary:
First step to enable deferred preamble builds. Not intending to land it
alone, will have follow-ups that will implement full deferred build
functionality and will land after all of them are ready.

Reviewers: sammccall

Subscribers: ilya-biryukov, javed.absar, MaskRay, jkorous, arphaman, usaxena95, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D76125

4 years agoFix mismatch from D77112 and D77499
Nathan James [Mon, 6 Apr 2020 19:19:12 +0000 (20:19 +0100)]
Fix mismatch from D77112 and D77499

4 years ago[ASTMatchers] Matchers that take enumerations args provide hints with invalid arguments
Nathan James [Sun, 5 Apr 2020 14:33:34 +0000 (15:33 +0100)]
[ASTMatchers] Matchers that take enumerations args provide hints with invalid arguments

Summary:
This adds support for giving hints when using dynamic matchers with enum args. Take this query, I couldn't figure out why the matcher wasn't working:
(Turns out it needed to be "IntegralToBoolean", but thats another bug itself)
```
clang-query> match implicitCastExpr(hasCastKind("CK_IntegralToBoolean"))
1:1: Error parsing argument 1 for matcher implicitCastExpr.
1:18: Error building matcher hasCastKind.
1:30: Incorrect type for arg 1. (Expected = string) != (Actual = String)
```
With this patch the new behaviour looks like this:
```
clang-query> match implicitCastExpr(hasCastKind("CK_IntegralToBoolean"))
1:1: Error parsing argument 1 for matcher implicitCastExpr.
1:18: Error building matcher hasCastKind.
1:30: Unknown value 'CK_IntegralToBoolean' for arg 1; did you mean 'IntegralToBoolean'
```

There are no test cases for this yet as there simply isn't any infrastructure for testing errors reported when parsing args that I can see.

Reviewers: klimek, jdoerfert, aaron.ballman

Reviewed By: aaron.ballman

Subscribers: aaron.ballman, dexonsmith, mgorny, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D77499

4 years ago[llvm][CodeGen] Avoid implicit cast of TypeSize to integer in `initActions`.
Francesco Petrogalli [Mon, 6 Apr 2020 18:25:17 +0000 (19:25 +0100)]
[llvm][CodeGen] Avoid implicit cast of TypeSize to integer in `initActions`.

Reviewers: sdesmalen, efriedma

Reviewed By: efriedma

Subscribers: hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D77317

4 years ago[WebAssembly] Handle event exports
Heejin Ahn [Sat, 4 Apr 2020 13:56:35 +0000 (06:56 -0700)]
[WebAssembly] Handle event exports

Summary: This handles exports of events, which was missing.

Reviewers: sbc100

Subscribers: dschuff, jgravelle-google, sunfish, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D77566

4 years agoRemove math.h/cmath include from DataTypes.h
Reid Kleckner [Mon, 6 Apr 2020 18:14:59 +0000 (11:14 -0700)]
Remove math.h/cmath include from DataTypes.h

DataTypes.h is meant to wrap the integer type and limits headers, which
have some unfortunate variance. The FP math functions declared by math.h
are not unnecessary. math.h took a noticeable amount of time to parse
(~40ms), but that could be startup costs.

Anyway, we don't need to include it, so skipping it can't hurt.

This has been present since the initial CMake build was added in 2008.

4 years agoAdd InjectTLIMappings pass to new pass manager
Masoud Ataei jaliseh [Mon, 6 Apr 2020 18:16:48 +0000 (13:16 -0500)]
Add InjectTLIMappings pass to new pass manager

This pass is created in d6de5f12d485a85504bc99d384a85634574a27e2 and tested
for new and legacy pass manager but never added to new pass manager pipeline.
I am adding it to new pass manager pipeline.

This pass is get used in Vector Function Database (VFDatabase) and without
this pass in new pass manager pipeline, none of the vector libraries are work
ing with new pass manager.

Related passes:
66c120f02560ef528a60924104ead66f330190f1
https://reviews.llvm.org/D74944

Differential revision: https://reviews.llvm.org/D75354

4 years ago[mlir][AsmPrinter] Change value numbering for local scope to be the next isolated...
River Riddle [Mon, 6 Apr 2020 18:04:22 +0000 (11:04 -0700)]
[mlir][AsmPrinter] Change value numbering for local scope to be the next isolated operation.

Summary: This revision updates the value numbering when printing to number from the next parent operation that is isolated from above. This is the highest level to number from that still ensures thread-safety. This revision also changes the behavior of Operator::operator<< to use local scope to avoid thread races when numbering operations.

Differential Revision: https://reviews.llvm.org/D77525

4 years ago[clang][CodeComplete] Dont perform fallback completion for incomplete member ref
Kadir Cetinkaya [Mon, 6 Apr 2020 17:13:08 +0000 (19:13 +0200)]
[clang][CodeComplete] Dont perform fallback completion for incomplete member ref

Summary:
Clang performs expression based completion whenever it can't figure out
base of a member reference expression. It might be quite confusing in cases like
incomplete types. This patch disables that fallback.

Unfortunately `ParsePostfixExpressionSuffix` is quite tangled and this patch
adds more to it.

Reviewers: sammccall

Subscribers: cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D77570

4 years ago[OpenMP] Fix layering problem with FrontendOpenMP
Johannes Doerfert [Mon, 6 Apr 2020 17:39:40 +0000 (12:39 -0500)]
[OpenMP] Fix layering problem with FrontendOpenMP

Summary:
ASTMatchers is used in various places and it now exposes the
LLVMFrontendOpenMP library to its users without them needing to depend
on it explicitly.

Reviewers: lebedev.ri

Subscribers: mgorny, yaxunl, bollu, guansong, martong, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D77574

4 years ago[SelectionDAGBuilder] Fix ISD::FREEZE creation for structs with fields of different...
Craig Topper [Mon, 6 Apr 2020 17:22:50 +0000 (10:22 -0700)]
[SelectionDAGBuilder] Fix ISD::FREEZE creation for structs with fields of different types.

The previous code used the type of the first field for the VT
passed to getNode for every field.

I've based the implementation here off what is done in visitSelect
as it removes the need to special case aggregates.

Differential Revision: https://reviews.llvm.org/D77093

4 years ago[NFC] Minor fix for deferred diagnostics emitter
Yaxun (Sam) Liu [Mon, 6 Apr 2020 17:52:47 +0000 (13:52 -0400)]
[NFC] Minor fix for deferred diagnostics emitter

Differential Revision: https://reviews.llvm.org/D77028

4 years ago[Hexagon] Make lld be the default linker for linux/musl
Sid Manning [Sun, 5 Apr 2020 00:36:00 +0000 (19:36 -0500)]
[Hexagon] Make lld be the default linker for linux/musl

When the target is hexagon-unknown-linux-musl select lld as the default
linker.

Differential Revision: https://reviews.llvm.org/D77498

4 years ago[lit] Move the recursiveExpansionLimit setting to TestingConfig
Louis Dionne [Fri, 3 Apr 2020 16:49:09 +0000 (12:49 -0400)]
[lit] Move the recursiveExpansionLimit setting to TestingConfig

The LitConfig is shared across the whole test suite. However, since
enabling recursive expansion can be a breaking change for some test
suites, it's important to confine the setting to test suites that
enable it explicitly.

Note that other issues were raised with the way recursiveExpansionLimit
operates. However, this commit simply moves the setting to the right
place -- the mechanism by which it works can be improved independently.

Differential Revision: https://reviews.llvm.org/D77415

4 years ago[gn build] Port 00a57558978
LLVM GN Syncbot [Mon, 6 Apr 2020 17:54:02 +0000 (17:54 +0000)]
[gn build] Port 00a57558978

4 years agoRevert e1730cfeb3588f20dcf4a96b181ad52761666e52
Konstantin Pyzhov [Mon, 6 Apr 2020 09:56:11 +0000 (05:56 -0400)]
Revert e1730cfeb3588f20dcf4a96b181ad52761666e52

4 years ago[clang-tidy] Add check llvmlibc-implementation-in-namespace.
Paula Toth [Mon, 6 Apr 2020 17:40:50 +0000 (10:40 -0700)]
[clang-tidy] Add check llvmlibc-implementation-in-namespace.

Summary:
This check makes sure all llvm-libc implementations falls within the `__llvm_libc` namespace.

Reviewers: alexfh, aaron.ballman, hokein, njames93

Reviewed By: aaron.ballman

Subscribers: Eugene.Zelenko, libc-commits, mgorny, xazax.hun, MaskRay, cfe-commits, sivachandra

Tags: #clang-tools-extra, #clang

Differential Revision: https://reviews.llvm.org/D76818

4 years ago[CFGPrinter][CallPrinter][polly] Adding distinct structure for CFGDOTInfo
Kirill Naumov [Mon, 6 Apr 2020 16:53:10 +0000 (16:53 +0000)]
[CFGPrinter][CallPrinter][polly] Adding distinct structure for CFGDOTInfo

The patch introduces the system to distinctively store the information
needed for the Control Flow Graph as well as the instrumentary needed for
the follow-up changes: BlockFrequencyInfo and BranchProbabilityInfo.
The patch is a part of sequence of three patches, related to graphs Heat Coloring.

Reviewers: rcorcs, apilipenko, davidxl, sfertile, fedor.sergeev, eraman, bollu

Differential Revision: https://reviews.llvm.org/D76820

4 years agoRemove llvm::Error include form Diagnostic.h
Reid Kleckner [Mon, 6 Apr 2020 17:32:16 +0000 (10:32 -0700)]
Remove llvm::Error include form Diagnostic.h

Saves ~400 related LLVM ADT. llvm/ADT/Error.h takes 90ms to parse.

$ diff -u <(sort thedeps-before.txt) <(sort thedeps-after.txt) \
    | grep '^[-+] ' | sort | uniq -c | sort -nr
    403 -    /usr/local/google/home/rnk/llvm-project/llvm/include/llvm/Support/Error.h
    403 -    /usr/local/google/home/rnk/llvm-project/llvm/include/llvm-c/Error.h
    397 -    /usr/local/google/home/rnk/llvm-project/llvm/include/llvm/Support/Format.h
    397 -    /usr/local/google/home/rnk/llvm-project/llvm/include/llvm/Support/Debug.h
    377 -    /usr/local/google/home/rnk/llvm-project/llvm/include/llvm/ADT/StringExtras.h
    158 -    /usr/local/google/home/rnk/llvm-project/llvm/include/llvm-c/ExternC.h
    138 -    /usr/local/google/home/rnk/llvm-project/llvm/include/llvm/Support/ErrorOr.h
     13 -    /usr/local/google/home/rnk/llvm-project/llvm/include/llvm/Support/raw_ostream.h
     13 -    /usr/local/google/home/rnk/llvm-project/llvm/include/llvm/ADT/SmallString.h
      5 -    /usr/local/google/home/rnk/llvm-project/llvm/include/llvm/ADT/Twine.h

4 years ago[mlir] Fix typo in docs/DefiningAttributesAndTypes.md
Jean-Michel Gorius [Mon, 6 Apr 2020 17:38:00 +0000 (19:38 +0200)]
[mlir] Fix typo in docs/DefiningAttributesAndTypes.md

4 years ago[libc++] Make sure we execute tests with the current environment
Louis Dionne [Mon, 6 Apr 2020 16:43:55 +0000 (12:43 -0400)]
[libc++] Make sure we execute tests with the current environment

The new libc++ test format doesn't automatically do this (cause it
would be the wrong place to do it).

4 years ago[analyzer] Fix NSErrorChecker false positives on constructors.
Artem Dergachev [Mon, 6 Apr 2020 17:27:40 +0000 (20:27 +0300)]
[analyzer] Fix NSErrorChecker false positives on constructors.

Constructors and delete operators cannot return a boolean value.
Therefore they cannot possibly follow the NS/CFError-related coding
conventions.

Patch by Valeriy Savchenko!

Differential Revision: https://reviews.llvm.org/D77551

4 years ago[analyzer] Fix scan-build and exploded-graph-rewriter tests on Windows.
Artem Dergachev [Mon, 6 Apr 2020 17:18:13 +0000 (20:18 +0300)]
[analyzer] Fix scan-build and exploded-graph-rewriter tests on Windows.

Detect script locations in a more straightforward way: we don't need to
search for them because we know exactly where they are anyway.

Fix a file path escaping issue in exploded-graph-rewriter with Windows
backslashes in the path.

'REQUIRES: shell' remains in scan-build tests for now, so that to
observe the buildbot reaction on removing it in a cleaner experiment.

Patch by Denys Petrov!

Differential Revision: https://reviews.llvm.org/D76768

4 years ago[AST] Remove DeclCXX.h dep on ASTContext.h
Reid Kleckner [Mon, 6 Apr 2020 16:55:33 +0000 (09:55 -0700)]
[AST] Remove DeclCXX.h dep on ASTContext.h

Saves only 36 includes of ASTContext.h and related headers.

There are two deps on ASTContext.h:
- C++ method overrides iterator types (TinyPtrVector)
- getting LangOptions

For #1, duplicate the iterator type, which is
TinyPtrVector<>::const_iterator.

For #2, add an out-of-line accessor to get the language options. Getting
the ASTContext from a Decl is already an out of line method that loops
over the parent DeclContexts, so if it is ever performance critical, the
proper fix is to pass the context (or LangOpts) into the predicate in
question.

Other changes are just header fixups.

4 years agoSpeed up deferred diagnostic emitter
Yaxun (Sam) Liu [Thu, 2 Apr 2020 03:26:56 +0000 (23:26 -0400)]
Speed up deferred diagnostic emitter

Move function emitDeferredDiags from Sema to DeferredDiagsEmitter since it
is only used by DeferredDiagsEmitter.

Also skip visited functions to avoid exponential compile time.

Differential Revision: https://reviews.llvm.org/D77028

4 years ago[AMDGPU] Disable 'Skip Uniform Regions' optimization by default for AMDGPU.
Konstantin Pyzhov [Mon, 6 Apr 2020 09:10:37 +0000 (05:10 -0400)]
[AMDGPU] Disable 'Skip Uniform Regions' optimization by default for AMDGPU.

Reviewers: sameerds, dstuttard

Differential Revision: https://reviews.llvm.org/D77228

4 years ago[sanitizer_common] Update global_symbols.txt.
Matt Morehouse [Mon, 6 Apr 2020 16:55:56 +0000 (09:55 -0700)]
[sanitizer_common] Update global_symbols.txt.

4 years agoRevert "[sanitizer_common] Show new unexpected symbols."
Matt Morehouse [Mon, 6 Apr 2020 16:55:03 +0000 (09:55 -0700)]
Revert "[sanitizer_common] Show new unexpected symbols."

This reverts commit 68d73dfa1a422fd045e25d06c3e8ff01621c377a.

4 years agoRevert "[sanitizer_common] Update global_symbols.txt."
Matt Morehouse [Mon, 6 Apr 2020 16:54:56 +0000 (09:54 -0700)]
Revert "[sanitizer_common] Update global_symbols.txt."

This reverts commit 98bb7fdbe586e7f95588c99c5363e94870da75f5.

4 years ago[AArch64] Allow logical immediates to have all-1 in top bits
Fangrui Song [Tue, 10 Mar 2020 00:35:50 +0000 (17:35 -0700)]
[AArch64] Allow logical immediates to have all-1 in top bits

So that constant expressions like the following are permitted:

and w0, w0, #~(0xfe<<24)
and w1, w1, #~(0xff<<24)

The behavior matches GNU as (opcodes/aarch64-opc.c:aarch64_logical_immediate_p).

Reviewed By: sdesmalen

Differential Revision: https://reviews.llvm.org/D75885

4 years ago[LV] Fix value that could be read uninitialized.
Florian Hahn [Mon, 6 Apr 2020 16:53:38 +0000 (17:53 +0100)]
[LV] Fix value that could be read uninitialized.

This should fix
http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-bootstrap-msan/builds/18569

4 years ago[DebugInfo] Change to constructor homing debug info mode: skip literal types
Amy Huang [Tue, 31 Mar 2020 20:31:42 +0000 (13:31 -0700)]
[DebugInfo] Change to constructor homing debug info mode: skip literal types

Summary:
In constructor type homing mode sometimes complete debug info for constexpr
types was missing, because there was not a constructor emitted. This change
makes constructor type homing ignore constexpr types.

Reviewers: rnk, dblaikie

Subscribers: aprantl, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D77432

4 years ago[ELF] Make --version-script/--dynamic-list work for lazy symbols fetched by LTO libcalls
Fangrui Song [Wed, 1 Apr 2020 23:11:31 +0000 (16:11 -0700)]
[ELF] Make --version-script/--dynamic-list work for lazy symbols fetched by LTO libcalls

Fixes https://bugs.llvm.org/show_bug.cgi?id=45391

The LTO code generator happens after version script scanning and may
create references which will fetch some lazy symbols.

Currently a version script does not assign VER_NDX_LOCAL to lazy symbols
and such symbols will be made global after they are fetched.

Change findByVersion and findAllByVersion to work on lazy symbols.
For unfetched lazy symbols, we should keep them non-local (D35263).
Check isDefined() in computeBinding() as a compensation.

This patch fixes a companion bug that --dynamic-list does not export
libcall fetched symbols.

Reviewed By: grimar

Differential Revision: https://reviews.llvm.org/D77280

4 years ago[sanitizer_common] Update global_symbols.txt.
Matt Morehouse [Mon, 6 Apr 2020 16:42:32 +0000 (09:42 -0700)]
[sanitizer_common] Update global_symbols.txt.

4 years ago[sanitizer_common] Show new unexpected symbols.
Matt Morehouse [Mon, 6 Apr 2020 16:39:53 +0000 (09:39 -0700)]
[sanitizer_common] Show new unexpected symbols.

4 years ago[libc++] Mark two std::timed_mutex tests as flaky
Louis Dionne [Mon, 6 Apr 2020 16:40:54 +0000 (12:40 -0400)]
[libc++] Mark two std::timed_mutex tests as flaky

4 years ago[RDA] Only store most recent reaching def from predecessors (NFCI)
Nikita Popov [Sat, 4 Apr 2020 22:46:20 +0000 (00:46 +0200)]
[RDA] Only store most recent reaching def from predecessors (NFCI)

When entering a basic block, RDA inserts reaching definitions coming
from predecessor blocks (which will be negative numbers) in a rather
peculiar way. If you have incoming reaching definitions -4, -3, -2, -1,
it will insert those. If you have incoming reaching definitions
-1, -2, -3, -4, it will insert -1, -1, -1, -1, as the max is taken
at each step. That's probably not what was intended...

However, RDA only actually cares about the most recent reaching
definition from a predecessor (to calculate clearance), so this
ends up working fine as far as behavior is concerned. It does
waste memory on unnecessary reaching definitions though.

This patch changes the implementation to first compute the most
recent reaching definition in one loop, and then insert only that
one in a separate loop.

Differential Revision: https://reviews.llvm.org/D77508

4 years ago[x86] add size cost tests for casts and binops; NFC
Sanjay Patel [Mon, 6 Apr 2020 16:34:59 +0000 (12:34 -0400)]
[x86] add size cost tests for casts and binops; NFC

Shows bugs for div/rem/fdiv and possibly others.

4 years ago[RDA] Don't adjust ReachingDefDefaultVal (NFCI)
Nikita Popov [Sun, 5 Apr 2020 17:16:56 +0000 (19:16 +0200)]
[RDA] Don't adjust ReachingDefDefaultVal (NFCI)

At the end of a basic block, RDA adjusts all the reaching defs it
found to be relative to the end of the basic block, rather than the
start of it. However, it also does this to registers which don't
have a reaching def, indicated by ReachingDefDefaultVal. This means
that code checking against ReachingDefDefaultVal will not skip them,
and may insert them into the reaching definition list. This is
ultimately harmless, but causes unnecessary work and is logically
not right.

Differential Revision: https://reviews.llvm.org/D77506

4 years ago[OPENMP]Fix PR45439: `omp for collapse(2) ordered(2)` generates invalid
Alexey Bataev [Mon, 6 Apr 2020 15:49:17 +0000 (11:49 -0400)]
[OPENMP]Fix PR45439: `omp for collapse(2) ordered(2)` generates invalid
IR.

Fixed a crash because of the not quite correct casting of the value of
iterations.

4 years ago[lldb] NFC: Fix trivial typo in comments, documents, and messages
Kazuaki Ishizaki [Mon, 6 Apr 2020 16:06:02 +0000 (01:06 +0900)]
[lldb] NFC: Fix trivial typo in comments, documents, and messages

Differential Revision: https://reviews.llvm.org/D77460

4 years agoUse in-tree clang-format-diff.py as Arcanist linter
Scott Linder [Fri, 3 Apr 2020 19:48:28 +0000 (15:48 -0400)]
Use in-tree clang-format-diff.py as Arcanist linter

Summary:
The only guarantee there seems to be in the clang-format packaging is
that an executable called `clang-format` is in the PATH. Use the
in-tree `clang-format-diff.py` to avoid assuming anything else.

Also remove dead code for SVN repo and switch to `git diff-index` which
is the git plumbing equivalent of `git diff` in this case.

Reviewers: starsid, mehdi_amini, vitalybuka, fhahn, kadircet

Subscribers: llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D77428

4 years ago[llvm-objdump] Fix case of -Wmismatched-tags
Jonathan Roelofs [Mon, 6 Apr 2020 15:37:14 +0000 (09:37 -0600)]
[llvm-objdump] Fix case of -Wmismatched-tags

4 years ago[llvm] Fix missing FileCheck directive colons
Jonathan Roelofs [Thu, 2 Apr 2020 22:28:32 +0000 (16:28 -0600)]
[llvm] Fix missing FileCheck directive colons

https://reviews.llvm.org/D77352

4 years ago[ValueTracking] enhance matching of umin/umax with 'not' operands
Sanjay Patel [Mon, 6 Apr 2020 15:40:45 +0000 (11:40 -0400)]
[ValueTracking] enhance matching of umin/umax with 'not' operands

The cmyk test is based on the known regression that resulted from:
rGf2fbdf76d8d0

This improves on the equivalent signed min/max change:
rG867f0c3c4d8c

The underlying icmp equivalence is:
  ~X pred ~Y --> Y pred X

For an icmp with constant, canonicalization results in a swapped pred:
  ~X < C -->  X > ~C

4 years agoFix a typo in an assert message; NFC.
BoYao Zhang [Mon, 6 Apr 2020 15:49:51 +0000 (11:49 -0400)]
Fix a typo in an assert message; NFC.

4 years agoAMDGPU/GlobalISel: Select llvm.amdgcn.div.scale
Matt Arsenault [Mon, 6 Apr 2020 01:16:00 +0000 (21:16 -0400)]
AMDGPU/GlobalISel: Select llvm.amdgcn.div.scale

4 years agoAMDGPU/GlobalISel: Fix llvm.amdgcn.div.fmas.ll
Matt Arsenault [Sun, 5 Apr 2020 23:49:09 +0000 (19:49 -0400)]
AMDGPU/GlobalISel: Fix llvm.amdgcn.div.fmas.ll

4 years ago[AMDGPU] Fix inaccurate comments
Jay Foad [Mon, 6 Apr 2020 15:44:08 +0000 (16:44 +0100)]
[AMDGPU] Fix inaccurate comments

4 years ago[libc++] Always use -fsyntax-only in .fail.cpp tests
Louis Dionne [Mon, 6 Apr 2020 15:04:13 +0000 (11:04 -0400)]
[libc++] Always use -fsyntax-only in .fail.cpp tests

We had a workaround because GCC 5 does not evaluate static assertions
that are dependent on template parameters. This commit removes the
workaround and marks the corresponding tests as unsupported with GCC 5.
This has the benefit of bringing the new and the old test formats closer
without having to carry a workaround for an old compiler in the new
test format.

4 years ago[analyzer] StdLibraryFunctionsChecker: match signature based on FunctionDecl
Gabor Marton [Fri, 3 Apr 2020 15:01:00 +0000 (17:01 +0200)]
[analyzer] StdLibraryFunctionsChecker: match signature based on FunctionDecl

Summary:
Currently we match the summary signature based on the arguments in the CallExpr.
There are a few problems with this approach.
1) Variadic arguments are handled badly. Consider the below code:
     int foo(void *stream, const char *format, ...);
     void test_arg_constraint_on_variadic_fun() {
        foo(0, "%d%d", 1, 2); // CallExpr
     }
   Here the call expression holds 4 arguments, whereas the function declaration
   has only 2 `ParmVarDecl`s. So there is no way to create a summary that
   matches the call expression, because the discrepancy in the number of
   arguments causes a mismatch.
2) The call expression does not handle the `restrict` type qualifier.
   In C99, fwrite's signature is the following:
     size_t fwrite(const void *restrict, size_t, size_t, FILE *restrict);
   However, in a call expression, like below, the type of the argument does not
   have the restrict qualifier.
    void test_fread_fwrite(FILE *fp, int *buf) {
      size_t x = fwrite(buf, sizeof(int), 10, fp);
    }
   This can result in an unmatches signature, so the summary is not applied.
The solution is to match the summary against the referened callee
`FunctionDecl` that we can query from the `CallExpr`.

Further patches will continue with additional refactoring where I am going to
do a lookup during the checker initialization and the signature match will
happen there. That way, we will not check the signature during every call,
rather we will compare only two `FunctionDecl` pointers.

Reviewers: NoQ, Szelethus, gamesh411, baloghadamsoftware

Subscribers: whisperity, xazax.hun, kristof.beyls, szepet, rnkovacs, a.sidorin, mikhail.ramalho, donat.nagy, dkrupp, Charusso, steakhal, danielkiss, ASDenysPetrov, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D77410

4 years ago[libc++] Make sure we include %{flags} when building with the new format
Louis Dionne [Mon, 6 Apr 2020 14:25:51 +0000 (10:25 -0400)]
[libc++] Make sure we include %{flags} when building with the new format

Otherwise, we're missing some flags like the flags that are used by
sanitizer builds and the 32-bit builds. In the long term, I think it
would be better to have only %{compile_flags} and %{link_flags}, but
for the benefit of adopting the new format by default, I think it's OK
to add %{flags} to it.

4 years ago[libc++] Mark is_scalar test as unsupported in C++11 and C++14
Louis Dionne [Mon, 6 Apr 2020 14:46:14 +0000 (10:46 -0400)]
[libc++] Mark is_scalar test as unsupported in C++11 and C++14

That test requires std::optional. We never noticed that because our
test format was skipping Objective-C++ tests altogether.

4 years ago[mlir][Linalg] Add a linalg.tensor_reshape to operate on tensors
Nicolas Vasilache [Mon, 6 Apr 2020 15:18:28 +0000 (11:18 -0400)]
[mlir][Linalg] Add a linalg.tensor_reshape to operate on tensors

Summary:
This revision adds a tensor_reshape operation that operates on tensors.
In the tensor world the constraints are less stringent and we can allow more
arbitrary dynamic reshapes, as long as they are contractions.

The expansion of a dynamic dimension into multiple dynamic dimensions is under-specified and is punted on for now.

Differential Revision: https://reviews.llvm.org/D77360

4 years ago[ValueTracking] add/adjust tests for min/max; NFC
Sanjay Patel [Mon, 6 Apr 2020 15:12:16 +0000 (11:12 -0400)]
[ValueTracking] add/adjust tests for min/max; NFC

4 years agoclang-format: [JS] handle pseudo-keywords.
Martin Probst [Mon, 6 Apr 2020 13:20:54 +0000 (15:20 +0200)]
clang-format: [JS] handle pseudo-keywords.

Summary:
The previous change in https://reviews.llvm.org/D77311 attempted to
detect more C++ keywords. However it also precisely detected all
JavaScript keywords. That's generally correct, but many JavaScripy
keywords, e.g. `get`, are so-called pseudo-keywords. They can be used in
positions where a keyword would never be legal, e.g. in a dotted
expression:

    x.type;  // type is a pseudo-keyword, but can be used here.
    x.get;   // same for get etc.

This change introduces an additional parameter to
`IsJavaScriptIdentifier`, allowing clients to toggle whether they want
to allow `IdentifierName` tokens, i.e. pseudo-keywords.

Reviewers: krasimir

Subscribers: cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D77548

4 years ago[analyzer] StdLibraryFunctionsChecker: Add test for function with default parameter
Gabor Marton [Fri, 3 Apr 2020 15:56:28 +0000 (17:56 +0200)]
[analyzer] StdLibraryFunctionsChecker: Add test for function with default parameter

Reviewers: Szelethus, baloghadamsoftware, gamesh411, steakhal, balazske

Subscribers: whisperity, xazax.hun, szepet, rnkovacs, a.sidorin, mikhail.ramalho, donat.nagy, dkrupp, Charusso, ASDenysPetrov, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D77411

4 years ago[VPlan] Introduce new VPWidenCallRecipe (NFC).
Florian Hahn [Mon, 6 Apr 2020 14:59:45 +0000 (15:59 +0100)]
[VPlan] Introduce new VPWidenCallRecipe (NFC).

This patch moves calls to their own recipe, to simplify the transition
to VPUser for operands of VPWidenRecipe, as discussed in D76992.

Subsequently additional information can be added to the recipe rather
than computing it during the execute step.

Reviewers: rengolin, Ayal, gilr, hsaito

Reviewed By: gilr

Differential Revision: https://reviews.llvm.org/D77467

4 years ago[AIX][PPC] Implement by-val caller arguments in multiple registers
Chris Bowler [Mon, 6 Apr 2020 14:41:48 +0000 (10:41 -0400)]
[AIX][PPC] Implement by-val caller arguments in multiple registers

Differential Revision: https://reviews.llvm.org/D76380

4 years agoAMDGPU/GlobalISel: Add unmerge of concat tests
Matt Arsenault [Thu, 6 Feb 2020 19:44:56 +0000 (14:44 -0500)]
AMDGPU/GlobalISel: Add unmerge of concat tests

4 years ago[Alignment][NFC] Assume AlignmentFromAssumptions::getNewAlignment is always set.
Guillaume Chatelet [Mon, 6 Apr 2020 10:53:56 +0000 (10:53 +0000)]
[Alignment][NFC] Assume AlignmentFromAssumptions::getNewAlignment is always set.

Summary:
In D77454 we explain that `LoadInst` and `StoreInst` always have their alignment defined.
This allows to work backward here and to infer that `getNewAlignment` does not need to return `0` in case of failure.
Returning `1` also works since it needs to be greater than the Load/Store alignment which is a least `1`.

This is patch is part of a series to introduce an Alignment type.
See this thread for context: http://lists.llvm.org/pipermail/llvm-dev/2019-July/133851.html
See this patch for the introduction of the type: https://reviews.llvm.org/D64790

Reviewers: courbet

Subscribers: hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D77538

4 years agotry to fix tsan bot
Nico Weber [Mon, 6 Apr 2020 14:47:41 +0000 (10:47 -0400)]
try to fix tsan bot

4 years ago[NFC][PPC][AIX] Test updates for byval args that fit in a single register
Chris Bowler [Mon, 6 Apr 2020 13:51:21 +0000 (09:51 -0400)]
[NFC][PPC][AIX] Test updates for byval args that fit in a single register

Differential Revision: https://reviews.llvm.org/D76875

4 years ago[llvm-objdump][XCOFF] Use symbol index+symbol name + storage mapping class as label...
diggerlin [Mon, 6 Apr 2020 14:09:12 +0000 (10:09 -0400)]
[llvm-objdump][XCOFF] Use symbol index+symbol name + storage mapping class as label for -D

SUMMARY:

For the llvm-objdump -D, the symbol name is used as a label in the disassembly for the specific address (when a symbol address is equal to the virtual address in the dump).

In XCOFF, multiple symbols may have the same name, being differentiated by their storage mapping class. It is helpful to print the QualName and not just the name when forming the output label for a csect symbol. The symbol index further removes any ambiguity caused by duplicate names.

To maintain compatibility with the binutils objdump, the XCOFF-specific --symbol-description option is added to enable the enhanced format.

Reviewers: hubert.reinterpretcast, James Henderson, Jason Liu ,daltenty
Subscribers: wuzish, nemanjai, hiraditya

Differential Revision: https://reviews.llvm.org/D72973