platform/upstream/llvm.git
4 years ago[clangd] Print underlying type for decltypes in hover
Kadir Cetinkaya [Thu, 16 Jan 2020 14:20:13 +0000 (15:20 +0100)]
[clangd] Print underlying type for decltypes in hover

Summary: Fixes https://github.com/clangd/clangd/issues/249

Reviewers: sammccall

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

Tags: #clang

Differential Revision: https://reviews.llvm.org/D72498

4 years ago[clangd] Make define outline code action visible
Kadir Cetinkaya [Thu, 16 Jan 2020 14:36:12 +0000 (15:36 +0100)]
[clangd] Make define outline code action visible

Summary: This got forgotten during the process.

Reviewers: sammccall, usaxena95

Subscribers: ilya-biryukov, MaskRay, jkorous, arphaman, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D72840

4 years agoAMDGPU: Remove IR section from MIR test
Matt Arsenault [Thu, 16 Jan 2020 18:37:59 +0000 (13:37 -0500)]
AMDGPU: Remove IR section from MIR test

Also generate check lines so this isn't just testing the meaningless
block name.

4 years agoGlobalISel: Apply target MMO flags to atomics
Matt Arsenault [Sun, 12 Jan 2020 20:54:09 +0000 (15:54 -0500)]
GlobalISel: Apply target MMO flags to atomics

Unify MMO flag handling with SelectionDAG like with loads and stores.

4 years agoGlobalISel: Preserve load/store metadata in IRTranslator
Matt Arsenault [Sun, 12 Jan 2020 19:10:42 +0000 (14:10 -0500)]
GlobalISel: Preserve load/store metadata in IRTranslator

This was dropping the invariant metadata on dead argument loads, so
they weren't deleted.

Atomics still need to be fixed the same way. Also, apparently store
was never preserving dereferencable which should also be fixed.

4 years agoTableGen/GlobalISel: Fix srcvalue inputs
Matt Arsenault [Thu, 16 Jan 2020 15:47:13 +0000 (10:47 -0500)]
TableGen/GlobalISel: Fix srcvalue inputs

Allow using srcvalue for discarding pattern inputs.

4 years agoTableGen: Remove dead code
Matt Arsenault [Thu, 16 Jan 2020 03:50:37 +0000 (22:50 -0500)]
TableGen: Remove dead code

4 years agoAMDGPU: Update tests to use modern buffer intrinsics
Matt Arsenault [Thu, 16 Jan 2020 16:34:19 +0000 (11:34 -0500)]
AMDGPU: Update tests to use modern buffer intrinsics

4 years agoDelete control character from comment. (NFC)
Adrian Prantl [Thu, 16 Jan 2020 18:43:49 +0000 (10:43 -0800)]
Delete control character from comment. (NFC)

4 years ago[Hexagon] Add a target feature to disable compound instructions
Krzysztof Parzyszek [Thu, 16 Jan 2020 18:00:35 +0000 (12:00 -0600)]
[Hexagon] Add a target feature to disable compound instructions

This affects the following instructions:
Tag: M4_mpyrr_addr     Syntax: Ry32 = add(Ru32,mpyi(Ry32,Rs32))
Tag: M4_mpyri_addr_u2  Syntax: Rd32 = add(Ru32,mpyi(#u6:2,Rs32))
Tag: M4_mpyri_addr     Syntax: Rd32 = add(Ru32,mpyi(Rs32,#u6))
Tag: M4_mpyri_addi     Syntax: Rd32 = add(#u6,mpyi(Rs32,#U6))
Tag: M4_mpyrr_addi     Syntax: Rd32 = add(#u6,mpyi(Rs32,Rt32))
Tag: S4_addaddi        Syntax: Rd32 = add(Rs32,add(Ru32,#s6))
Tag: S4_subaddi        Syntax: Rd32 = add(Rs32,sub(#s6,Ru32))
Tag: S4_or_andix       Syntax: Rx32 = or(Ru32,and(Rx32,#s10))
Tag: S4_andi_asl_ri    Syntax: Rx32 = and(#u8,asl(Rx32,#U5))
Tag: S4_ori_asl_ri     Syntax: Rx32 = or(#u8,asl(Rx32,#U5))
Tag: S4_addi_asl_ri    Syntax: Rx32 = add(#u8,asl(Rx32,#U5))
Tag: S4_subi_asl_ri    Syntax: Rx32 = sub(#u8,asl(Rx32,#U5))
Tag: S4_andi_lsr_ri    Syntax: Rx32 = and(#u8,lsr(Rx32,#U5))
Tag: S4_ori_lsr_ri     Syntax: Rx32 = or(#u8,lsr(Rx32,#U5))
Tag: S4_addi_lsr_ri    Syntax: Rx32 = add(#u8,lsr(Rx32,#U5))
Tag: S4_subi_lsr_ri    Syntax: Rx32 = sub(#u8,lsr(Rx32,#U5))

4 years ago[ELF] Decrease alignment of ThunkSection on 64-bit targets from 8 to 4
Fangrui Song [Thu, 16 Jan 2020 01:44:14 +0000 (17:44 -0800)]
[ELF] Decrease alignment of ThunkSection on 64-bit targets from 8 to 4

ThunkSection contains 4-byte instructions on all targets that use
thunks. Thunks should not be used in any performance sensitive places,
and locality/cache line/instruction fetching arguments should not apply.

We use 16 bytes as preferred function alignments for modern PowerPC cores.
In any case, 8 is not optimal.

Differential Revision: https://reviews.llvm.org/D72819

4 years agoRevert "[Loop Peeling] Add possibility to enable peeling on loop nests."
Arkady Shlykov [Thu, 16 Jan 2020 16:35:00 +0000 (08:35 -0800)]
Revert "[Loop Peeling] Add possibility to enable peeling on loop nests."

This reverts commit 3f3017e because there's a failure on peel-loop-nests.ll
with LLVM_ENABLE_EXPENSIVE_CHECKS on.

Differential Revision: https://reviews.llvm.org/D70304

4 years ago[gn build] (manually) port bed7626f04f7
Nico Weber [Thu, 16 Jan 2020 18:19:09 +0000 (13:19 -0500)]
[gn build] (manually) port bed7626f04f7

4 years ago[gn build] include revision information in lld --version output
Nico Weber [Thu, 16 Jan 2020 18:10:27 +0000 (13:10 -0500)]
[gn build] include revision information in lld --version output

4 years ago[PowerPC][AIX] Make PIC the default relocation model for AIX
stevewan [Thu, 16 Jan 2020 16:42:11 +0000 (11:42 -0500)]
[PowerPC][AIX] Make PIC the default relocation model for AIX

Summary:
The `llc` tool currently defaults to Static relocation model and generates non-relocatable code for 32-bit Power.
This is not desirable on AIX where we always generate Position Independent Code (PIC). This patch makes PIC the default relocation model for AIX.

Reviewers: daltenty, hubert.reinterpretcast, DiggerLin, Xiangling_L, sfertile

Reviewed By: hubert.reinterpretcast

Subscribers: mgorny, wuzish, nemanjai, hiraditya, kbarton, jsji, shchenz, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D72479

4 years ago[SystemZ] Avoid unnecessary conversions in vecintrin.h
Ulrich Weigand [Thu, 16 Jan 2020 17:56:16 +0000 (18:56 +0100)]
[SystemZ] Avoid unnecessary conversions in vecintrin.h

Use floating-point instead of integer zero constants to avoid
creating implicit conversions, which currently cause suboptimal
code to be generated with -ffp-exception-behavior=strict.

NFC otherwise.

4 years ago[lldb/test] Revert changes to debug-names-compressed.cpp
Pavel Labath [Thu, 16 Jan 2020 17:55:59 +0000 (18:55 +0100)]
[lldb/test] Revert changes to debug-names-compressed.cpp

With the changes in 15a6df52efa, the test is failing in some
configurations. Reverting while I investigate

4 years agoremove an include that's unused after r347592
Nico Weber [Thu, 16 Jan 2020 17:47:56 +0000 (12:47 -0500)]
remove an include that's unused after r347592

4 years ago[LLDB] Convert Plugins/ObjectFile/wasm/ObjectFileWasm.h to unix line endings
Derek Schuff [Thu, 16 Jan 2020 17:38:37 +0000 (09:38 -0800)]
[LLDB] Convert Plugins/ObjectFile/wasm/ObjectFileWasm.h to unix line endings

4 years ago[LLDB] Add SymbolVendorWasm plugin for WebAssembly debugging
Paolo Severini [Thu, 16 Jan 2020 16:36:45 +0000 (08:36 -0800)]
[LLDB] Add SymbolVendorWasm plugin for WebAssembly debugging

Add plugin class SymbolVendorWasm, with the logic to manage debug symbols
for Wasm modules.

Reviewers: clayborg, labath, aprantl, sbc100, teemperor

Reviewed By: labath

Tags: #lldb

Differential Revision: https://reviews.llvm.org/D72650

4 years ago[GVN] introduce GVNOptions to control GVN pass behavior
Fedor Sergeev [Thu, 16 Jan 2020 16:02:44 +0000 (19:02 +0300)]
[GVN] introduce GVNOptions to control GVN pass behavior

There are a few global (cl::opt) controls that enable optional
behavior in GVN. Introduce GVNOptions that provide corresponding
per-pass instance controls.

That will allow to use GVN multiple times in pipeline each time
with different settings.

Reviewers: asbirlea, rnk, reames, skatkov, fhahn
Reviewed By: fhahn

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D72732

4 years ago[lldb/Scripts] Remove swig_bot_lib/__init__.py
Jonas Devlieghere [Thu, 16 Jan 2020 17:15:33 +0000 (09:15 -0800)]
[lldb/Scripts] Remove swig_bot_lib/__init__.py

4 years ago[llvm] Make new pass manager's OptimizationLevel a class
Mircea Trofin [Thu, 16 Jan 2020 16:51:50 +0000 (08:51 -0800)]
[llvm] Make new pass manager's OptimizationLevel a class

Summary:
The old pass manager separated speed optimization and size optimization
levels into two unsigned values. Coallescing both in an enum in the new
pass manager may lead to unintentional casts and comparisons.

In particular, taking a look at how the loop unroll passes were constructed
previously, the Os/Oz are now (==new pass manager) treated just like O3,
likely unintentionally.

This change disallows raw comparisons between optimization levels, to
avoid such unintended effects. As an effect, the O{s|z} behavior changes
for loop unrolling and loop unroll and jam, matching O2 rather than O3.

The change also parameterizes the threshold values used for loop
unrolling, primarily to aid testing.

Reviewers: tejohnson, davidxl

Reviewed By: tejohnson

Subscribers: zzheng, ychen, mehdi_amini, hiraditya, steven_wu, dexonsmith, dang, cfe-commits, llvm-commits

Tags: #clang, #llvm

Differential Revision: https://reviews.llvm.org/D72547

4 years ago[gn build] Port 6b357866496
LLVM GN Syncbot [Thu, 16 Jan 2020 16:56:26 +0000 (16:56 +0000)]
[gn build] Port 6b357866496

4 years ago[Hexagon] Fix alignment info for __builtin_circ_lduh
Krzysztof Parzyszek [Thu, 16 Jan 2020 16:30:09 +0000 (10:30 -0600)]
[Hexagon] Fix alignment info for __builtin_circ_lduh

4 years ago[llvm-exegesis][mips] Add RegisterAliasingTest unit test
Miloš Stojanović [Mon, 30 Dec 2019 15:18:40 +0000 (16:18 +0100)]
[llvm-exegesis][mips] Add RegisterAliasingTest unit test

Differential Revision: https://reviews.llvm.org/D72004

4 years ago[llvm-exegesis][NFC] Refactor Mips tests fixtures into a base class.
Miloš Stojanović [Mon, 30 Dec 2019 14:50:51 +0000 (15:50 +0100)]
[llvm-exegesis][NFC] Refactor Mips tests fixtures into a base class.

Differential Revision: https://reviews.llvm.org/D72003

4 years agoAMDGPU/GlobalISel: Don't handle legacy buffer intrinsic
Matt Arsenault [Thu, 16 Jan 2020 16:28:46 +0000 (11:28 -0500)]
AMDGPU/GlobalISel: Don't handle legacy buffer intrinsic

4 years ago[MC][test] Fix non-portable GNU diff option
Hubert Tong [Thu, 16 Jan 2020 16:28:37 +0000 (11:28 -0500)]
[MC][test] Fix non-portable GNU diff option

Summary: This patch replaces the non-portable GNU diff option `--strip-trailing-cr` with the POSIX `-b` option in two test files.

Reviewers: daltenty, jasonliu

Reviewed By: daltenty

Subscribers: llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D72745

4 years ago[lldb][NFC] Delete TestDataFormatterObjCNSDate.py
Raphael Isemann [Thu, 16 Jan 2020 16:26:34 +0000 (17:26 +0100)]
[lldb][NFC] Delete TestDataFormatterObjCNSDate.py

This test is just TestDataFormatterObjCNSData.py copied but without any changes
(and it therefore doesn't even test NSDate).

It's also failing as NSData has been changed by me in
4f244bba4f66b14382c446b62e122fa684b8db78.

4 years agoAMDGPU/GlobalISel: Select DS GWS intrinsics
Matt Arsenault [Mon, 13 Jan 2020 00:04:24 +0000 (19:04 -0500)]
AMDGPU/GlobalISel: Select DS GWS intrinsics

4 years ago[Hexagon] Add preprocessor test for hexagonv66
Krzysztof Parzyszek [Thu, 16 Jan 2020 16:17:20 +0000 (10:17 -0600)]
[Hexagon] Add preprocessor test for hexagonv66

4 years ago[GlobalISel] Don't arbitrarily limit a mask to 64 bits
Jay Foad [Thu, 16 Jan 2020 14:36:41 +0000 (14:36 +0000)]
[GlobalISel] Don't arbitrarily limit a mask to 64 bits

Reviewers: arsenm

Subscribers: wdng, rovka, hiraditya, volkan, Petar.Avramovic, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D72853

4 years ago[GlobalISel] Pass MachineOperands into MachineIRBuilder helper methods
Jay Foad [Thu, 16 Jan 2020 12:37:00 +0000 (12:37 +0000)]
[GlobalISel] Pass MachineOperands into MachineIRBuilder helper methods

Reviewers: arsenm, aditya_nandakumar, aemerson

Subscribers: wdng, rovka, hiraditya, volkan, Petar.Avramovic, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D72849

4 years agoRemove some SVN-specific code.
Nico Weber [Thu, 16 Jan 2020 15:27:25 +0000 (10:27 -0500)]
Remove some SVN-specific code.

$URL$ is an SVN keyword substitution enabled via
`svn propset svn:keywords "URL" tools/clang/lib/Basic/Version.cpp`.
Now that we no longer use SVN, it's no longer being replaced by
anything, and we no longer offer svn exports. So remove the
$URL$-specific logic.

The "cfe" path prefix removal also no longer makes sense now that
we're on git: Both CLANG_REPOSITORY and LLVM_REPOSITORY are usually
set to https://github.com/llvm/llvm-project.git

So remove that too, and remove the "llvm" prefix removal for symmetry.
With the github url, "llvm" _is_ found in the string, but not in
the place the function expected. Nobody noticed since the llvm
repository path is only used if CLANG_REVISION and LLVM_REVISION are
different, which in the git monorepo world they never should be.
(I might remove the "// Support LLVM in a separate repository"
block in a separate commit.)

Differential Revision: https://reviews.llvm.org/D72848

4 years ago[Hexagon] Remove unnecessary case in StringSwitch, NFC
Krzysztof Parzyszek [Thu, 16 Jan 2020 15:58:45 +0000 (09:58 -0600)]
[Hexagon] Remove unnecessary case in StringSwitch, NFC

4 years ago[ARM][LowOverheadLoops] Update liveness info
Sam Parker [Thu, 16 Jan 2020 15:42:41 +0000 (15:42 +0000)]
[ARM][LowOverheadLoops] Update liveness info

Recommitting e93e0d413f3a after reverting due to test failures, which
will hopefully now be fixed. Original commit message:

After expanding the pseudo instructions, update the liveness info.
We do this in a post-order traversal of the loop, including its
exit blocks and preheader(s).

Differential Revision: https://reviews.llvm.org/D72131

4 years ago[GlobalISel] Use more MachineIRBuilder helper methods
Jay Foad [Thu, 16 Jan 2020 12:09:48 +0000 (12:09 +0000)]
[GlobalISel] Use more MachineIRBuilder helper methods

Reviewers: arsenm, nhaehnle

Subscribers: wdng, rovka, hiraditya, volkan, Petar.Avramovic, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D72833

4 years ago[lldb/DWARF/test] Freshen up debug_names tests
Pavel Labath [Wed, 15 Jan 2020 11:17:35 +0000 (12:17 +0100)]
[lldb/DWARF/test] Freshen up debug_names tests

These tests used "clang -mllvm -accel-tables=Dwarf" as a way to
guarantee that clang will emit the debug_names table. Unfortunately,
a change it clang made that insufficient (-gpubnames is required now
too), which rendered these tests ineffective. Since lldb automatically
falls back to the manual index, the tests didn't fail and this change
went largely unnoticed.

This patch updates the tests to really use debug_names (-gdwarf-5
-gpubnames) is the combination that works now, and it adds additional
checks to ensure the section is actually emitted.

Fortunately, no regressions crept in while these tests were disabled.

4 years ago[ARM][MVE] Enable extending gathers
Anna Welker [Thu, 16 Jan 2020 13:48:18 +0000 (13:48 +0000)]
[ARM][MVE] Enable extending gathers

Enables the masked gather pass to
create extending masked gathers.

Differential Revision: https://reviews.llvm.org/D72451

4 years ago[mlir][Linalg] Fix Linalg EDSC builders
Nicolas Vasilache [Thu, 16 Jan 2020 14:30:17 +0000 (09:30 -0500)]
[mlir][Linalg] Fix Linalg EDSC builders

Summary:
This diff fixes the fact that the method `mlir::edsc::makeGenericLinalgOp`
incorrectly adds 2 blocks to Linalg ops.

Tests are updated accordingly.

Reviewers: ftynse, hanchung, herhut, pifon2a, asaadaldien

Reviewed By: asaadaldien

Subscribers: merge_guards_bot, mehdi_amini, rriddle, jpienaar, burmako, shauheen, antiagainst, arpith-jacob, mgester, lucyrfox, liufengdb, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D72780

4 years agotry to unbreak build after 4b6d9ac392613
Nico Weber [Thu, 16 Jan 2020 15:12:06 +0000 (10:12 -0500)]
try to unbreak build after 4b6d9ac392613

4 years ago[VectorUtils] Rework the Vector Function Database (VFDatabase).
Francesco Petrogalli [Fri, 13 Dec 2019 19:43:26 +0000 (19:43 +0000)]
[VectorUtils] Rework the Vector Function Database (VFDatabase).

Summary:
This commits is a rework of the patch in
https://reviews.llvm.org/D67572.

The rework was requested to prevent out-of-tree performance regression
when vectorizing out-of-tree IR intrinsics. The vectorization of such
intrinsics is enquired via the static function `isTLIScalarize`. For
detail see the discussion in https://reviews.llvm.org/D67572.

Reviewers: uabelho, fhahn, sdesmalen

Subscribers: hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D72734

4 years agoRemoved an unused include from TypeLocVisitor.h
Dmitri Gribenko [Thu, 16 Jan 2020 14:57:28 +0000 (15:57 +0100)]
Removed an unused include from TypeLocVisitor.h

4 years agoMake lld cmake not compute commit revision twice
Nico Weber [Thu, 16 Jan 2020 14:55:36 +0000 (09:55 -0500)]
Make lld cmake not compute commit revision twice

r354605 moved LLD to the unified revision handling introduced in
rL353268 / r352729 and removed uses of LLD_REPOSITORY_STRING and
LLD_REVISION_STRING.

After this change, we no longer compute the (now-unused) values
of these two variables.

Since this removes the only use of llvm/utils/GetRepositoryPath,
remove that too (it's redundant with the system added in r354605).

While here, also remove LLD_VERSION_MAJOR and LLD_VERSION_MINOR.
Their uses were removed in r285163.

Also remove LLD_VERSION from Version.inc which as far as I can
tell has been unused since the file was added in r219277.

No behavior change.

Differential Revision: https://reviews.llvm.org/D72803

4 years agoRevert "[PHIEliminate] Move dbg values after phi and label"
Jeremy Morse [Thu, 16 Jan 2020 14:01:27 +0000 (14:01 +0000)]
Revert "[PHIEliminate] Move dbg values after phi and label"

Testing compiler-rt, a new assertion failure occurs when building
the GwpAsanTestObjects object. I'm uploading a reproducer to D70597.

This reverts commit 75188b01e9af3a89639d84be912f84610d6885ba.

4 years ago[ELF] Optimization to LinkerScript::computeInputSections NFC
Andrew Ng [Wed, 15 Jan 2020 11:48:37 +0000 (11:48 +0000)]
[ELF] Optimization to LinkerScript::computeInputSections NFC

Moved the section name check ahead of any filename matching or
exclusion. Firstly, this reduces the need to retrieve the filename and
secondly, reduces the amount of potentially expensive filename pattern
matching if such rules are present in the linker script.

The impact of this change is particularly significant when linking
objects built with -ffunction-sections and -fstack-size-section, using a
linker script that includes non-trivial filename patterns. In a number
of such cases, the link time is halved.

Differential Revision: https://reviews.llvm.org/D72775

4 years ago[clangd] Make output order of allTargetDecls deterministic
Kadir Cetinkaya [Thu, 16 Jan 2020 10:37:58 +0000 (11:37 +0100)]
[clangd] Make output order of allTargetDecls deterministic

Summary:
Makes use of insertion order to stabilize output for multiple decls.

Fixes https://bugs.llvm.org/show_bug.cgi?id=44564

Reviewers: sammccall

Subscribers: ilya-biryukov, MaskRay, jkorous, mgrang, arphaman, usaxena95, cfe-commits, aemerson

Tags: #clang

Differential Revision: https://reviews.llvm.org/D72826

4 years agoFix unused variable warning. NFCI.
Simon Pilgrim [Thu, 16 Jan 2020 11:26:56 +0000 (11:26 +0000)]
Fix unused variable warning. NFCI.

4 years agoFix "pointer is null" static analyzer warnings. NFCI.
Simon Pilgrim [Wed, 15 Jan 2020 16:40:20 +0000 (16:40 +0000)]
Fix "pointer is null" static analyzer warnings. NFCI.

Use castAs<> instead of getAs<> since the pointer is dereferenced immediately in all cases and castAs will perform the null assertion for us.

4 years ago[lldb][NFC] Migrate several tests to expect_expr
Raphael Isemann [Wed, 15 Jan 2020 12:30:04 +0000 (13:30 +0100)]
[lldb][NFC] Migrate several tests to expect_expr

expect_expr is the stricter and safer way of testing these expressions.

4 years agoRemove release note about in-process-cc1
Hans Wennborg [Thu, 16 Jan 2020 12:24:21 +0000 (13:24 +0100)]
Remove release note about in-process-cc1

This feature landed before the 10.x branch, so it will be covered in the
clang 10 release notes instead.

4 years ago[PHIEliminate] Move dbg values after phi and label
Chris Ye [Thu, 16 Jan 2020 11:40:44 +0000 (11:40 +0000)]
[PHIEliminate] Move dbg values after phi and label

If there are DBG_VALUEs between phi and label (after phi and before label),
DBG_VALUE will block PHI lowering after the LABEL. Moving all DBG_VALUEs
after Labels in the function ScheduleDAGSDNodes::EmitSchedule to avoid
impacting PHI lowering.

  before:
     PHI
     DBG_VALUE
     LABEL
  after: (move DBG_VALUE after label)
     PHI
     LABEL
     DBG_VALUE
  then: (phi lowering after label)
     LABEL
     COPY
     DBG_VALUE

Fixes the issue: https://bugs.llvm.org/show_bug.cgi?id=43859

Differential Revision: https://reviews.llvm.org/D70597

4 years ago[Concepts] Fix ConceptSpecializationExpr profiling crash
Saar Raz [Thu, 16 Jan 2020 11:35:20 +0000 (13:35 +0200)]
[Concepts] Fix ConceptSpecializationExpr profiling crash

ConceptSpecializationExprs (CSEs) were being created with nullptr
TemplateArgsAsWritten during TemplateTemplateParmDecl canonicalization, and
we were relying on them during profiling which caused sporadic crashes
in test/CXX/.../temp.arg.template/p3-2a.cpp introduced in D44352.

Change profiling of CSEs to instead rely on the actual converted template
arguments and concept named.

4 years ago[clangd] Dont display `<unknown>` kinds in hover board
Kadir Cetinkaya [Wed, 15 Jan 2020 17:09:51 +0000 (18:09 +0100)]
[clangd] Dont display `<unknown>` kinds in hover board

Summary:
Currently when hovering over an `auto` or `decltype` that resolve to a
builtin-type, clangd would display `<unknown>` as the kind of the symbol.

Drop that to make rendering nicer.

Reviewers: usaxena95

Subscribers: ilya-biryukov, MaskRay, jkorous, arphaman, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D72777

4 years ago[IR] Mark memset.* intrinsics as IntrWriteMem.
Florian Hahn [Thu, 16 Jan 2020 10:25:29 +0000 (10:25 +0000)]
[IR] Mark memset.* intrinsics as IntrWriteMem.

llvm.memset intrinsics do only write memory, but are missing
IntrWriteMem, so they doesNotReadMemory() returns false for them.

The test change is due to the test checking the fn attribute ids at the
call sites, which got bumped up due to a new combination with writeonly
appearing in the test file.

Reviewers: jdoerfert, reames, efriedma, nlopes, lebedev.ri

Reviewed By: jdoerfert

Differential Revision: https://reviews.llvm.org/D72789

4 years agoclang-format: [JS] tests for async wrapping.
Martin Probst [Mon, 18 Nov 2019 09:07:32 +0000 (10:07 +0100)]
clang-format: [JS] tests for async wrapping.

Summary:
Adds tests to ensure that `async method() ...` does not wrap between async and
the method name, which would cause automatic semicolon insertion.

Reviewers: krasimir

Subscribers: cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D70377

4 years ago[LV] Allow assume calls in predicated blocks.
Florian Hahn [Thu, 16 Jan 2020 10:06:09 +0000 (10:06 +0000)]
[LV] Allow assume calls in predicated blocks.

The assume intrinsic is intentionally marked as may reading/writing
memory, to avoid passes moving them around. When flattening the CFG
for predicated blocks, we have to drop the assume calls, as they
are control-flow dependent.

There are some cases where we can do better (when control flow is
preserved), but that is follow-up work.

Fixes PR43620.

Reviewers: hsaito, rengolin, dcaballe, Ayal

Reviewed By: Ayal

Differential Revision: https://reviews.llvm.org/D68814

4 years ago[LV] Make X86/assume.ll X86 independent (NFC).
Florian Hahn [Thu, 16 Jan 2020 09:57:12 +0000 (09:57 +0000)]
[LV] Make X86/assume.ll X86 independent (NFC).

The test does not check anything X86 specific. This is a preparation for
the D68814.

4 years ago[gn build] Port ed181efa175
LLVM GN Syncbot [Thu, 16 Jan 2020 09:55:55 +0000 (09:55 +0000)]
[gn build] Port ed181efa175

4 years ago[HIP][AMDGPU] expand printf when compiling HIP to AMDGPU
Sameer Sahasrabuddhe [Thu, 22 Aug 2019 10:04:35 +0000 (15:34 +0530)]
[HIP][AMDGPU] expand printf when compiling HIP to AMDGPU

Summary:
This change implements the expansion in two parts:
- Add a utility function emitAMDGPUPrintfCall() in LLVM.
- Invoke the above function from Clang CodeGen, when processing a HIP
  program for the AMDGPU target.

The printf expansion has undefined behaviour if the format string is
not a compile-time constant. As a sufficient condition, the HIP
ToolChain now emits -Werror=format-nonliteral.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D71365

4 years ago[VE] i64 arguments, return values and constants
Kazushi (Jam) Marukawa [Thu, 16 Jan 2020 08:24:41 +0000 (09:24 +0100)]
[VE] i64 arguments, return values and constants

Summary: Support for i64 arguments (in register), return values and constants along with tests.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D72776

4 years ago[lldb] Fix asan failures in data-formatter-objc tests
Raphael Isemann [Thu, 16 Jan 2020 08:57:13 +0000 (09:57 +0100)]
[lldb] Fix asan failures in data-formatter-objc tests

The test is currently failing on some systems with ASAN enabled due to:
```
==22898==ERROR: AddressSanitizer: heap-buffer-overflow on address 0x603000003da4 at pc 0x00010951c33d bp 0x7ffee6709e00 sp 0x7ffee67095c0
READ of size 5 at 0x603000003da4 thread T0
    #0 0x10951c33c in wrap_memmove+0x16c (libclang_rt.asan_osx_dynamic.dylib:x86_64+0x1833c)
    #1 0x7fff4a327f57 in CFDataReplaceBytes+0x1ba (CoreFoundation:x86_64+0x13f57)
    #2 0x7fff4a415a44 in __CFDataInit+0x2db (CoreFoundation:x86_64+0x101a44)
    #3 0x1094f8490 in main main.m:424
    #4 0x7fff77482084 in start+0x0 (libdyld.dylib:x86_64+0x17084)
0x603000003da4 is located 0 bytes to the right of 20-byte region [0x603000003d90,0x603000003da4)
allocated by thread T0 here:
    #0 0x109547c02 in wrap_calloc+0xa2 (libclang_rt.asan_osx_dynamic.dylib:x86_64+0x43c02)
    #1 0x7fff763ad3ef in class_createInstance+0x52 (libobjc.A.dylib:x86_64+0x73ef)
    #2 0x7fff4c6b2d73 in NSAllocateObject+0x12 (Foundation:x86_64+0x1d73)
    #3 0x7fff4c6b5e5f in -[_NSPlaceholderData initWithBytes:length:copy:deallocator:]+0x40 (Foundation:x86_64+0x4e5f)
    #4 0x7fff4c6d4cf1 in -[NSData(NSData) initWithBytes:length:]+0x24 (Foundation:x86_64+0x23cf1)
    #5 0x1094f8245 in main main.m:404
    #6 0x7fff77482084 in start+0x0 (libdyld.dylib:x86_64+0x17084)
```

The reason is that we create a string "HELLO" but get the size wrong (it's 5 bytes instead
of 4). Later on we read the buffer and pretend it is 5 bytes long, causing an OOB read
which ASAN detects.

In general this test probably needs some cleanup as it produces on macOS 10.15 around
100 compiler warnings which isn't great, but let's first get the bot green.

4 years ago[DebugInfo] Simplify the constructor of DWARFDebugAranges::Range. NFC.
Igor Kudrin [Wed, 15 Jan 2020 11:17:07 +0000 (18:17 +0700)]
[DebugInfo] Simplify the constructor of DWARFDebugAranges::Range. NFC.

This removes the default values of the arguments. The only caller,
DWARFDebugAranges::construct(), provides all three parameters.

Differential Revision: https://reviews.llvm.org/D72757

4 years ago[LegalizeDAG][TargetLowering] Move vXi64/i64->vXf32/f32 uint_to_fp legalizing code...
Craig Topper [Thu, 16 Jan 2020 06:01:14 +0000 (22:01 -0800)]
[LegalizeDAG][TargetLowering] Move vXi64/i64->vXf32/f32 uint_to_fp legalizing code from TargetLowering::expandUINT_TO_FP back to LegalizeDAG.

This was moved in October 2018, but we don't appear to be using
this for vectors on any in tree target.

Moving it back simplifies D72794 so we can share the code for i32->f32.

4 years ago[mlir] fix broken links to Glossary
Hiroshi Inoue [Thu, 16 Jan 2020 05:15:34 +0000 (14:15 +0900)]
[mlir] fix broken links to Glossary

Differential Revision: https://reviews.llvm.org/D72697

4 years ago[lldb/Reproducers] Print more info for reproducer status
Jonas Devlieghere [Thu, 16 Jan 2020 04:22:20 +0000 (20:22 -0800)]
[lldb/Reproducers] Print more info for reproducer status

Reproducer status now prints the capture/replay path. It will also print
the status of auto generation when enabled.

4 years ago[gn build] Port 8fdafb7dced
LLVM GN Syncbot [Thu, 16 Jan 2020 04:13:31 +0000 (04:13 +0000)]
[gn build] Port 8fdafb7dced

4 years agoInsert wait instruction after X87 instructions which could raise
Liu, Chen3 [Thu, 16 Jan 2020 02:49:59 +0000 (10:49 +0800)]
Insert wait instruction after X87 instructions which could raise
float-point exception.

This patch also modify some mayRaiseFPException flag which set in D68854.

Differential Revision: https://reviews.llvm.org/D72750

4 years ago[lldb/Reproducers] Add a flag to always generating a reproducer
Jonas Devlieghere [Thu, 16 Jan 2020 03:44:46 +0000 (19:44 -0800)]
[lldb/Reproducers] Add a flag to always generating a reproducer

Add a flag which always generates a reproducer when normally it would be
discarded. This is meant for testing purposes to capture a debugger
session without modification the session itself.

4 years agoSet some fast math attributes in setFunctionAttributes
Matt Arsenault [Thu, 5 Dec 2019 09:37:27 +0000 (15:07 +0530)]
Set some fast math attributes in setFunctionAttributes

This will provide a more consistent view to codegen for these
attributes. The current system is somewhat awkward, and the fields in
TargetOptions are reset based on the command line flag if the
attribute isn't set. By forcing these attributes with the flag, there
can never be an inconsistency in the behavior if code directly
inspects the attribute on the function without considering the command
line flags.

4 years ago[lldb/Reproducers] Extract function for reading environment override (NFC)
Jonas Devlieghere [Thu, 16 Jan 2020 03:12:50 +0000 (19:12 -0800)]
[lldb/Reproducers] Extract function for reading environment override (NFC)

Create a helper function for reading reproducer overrides from
environment variables.

4 years agoPR42694 Support explicit(bool) in older language modes as an extension.
Richard Smith [Thu, 16 Jan 2020 02:37:32 +0000 (18:37 -0800)]
PR42694 Support explicit(bool) in older language modes as an extension.

This needs somewhat careful disambiguation, as C++2a explicit(bool) is a
breaking change. We only enable it in cases where the source construct
could not possibly be anything else.

4 years ago[SampleFDO] Fix invalid branch profile generated by indirect call promotion.
Wei Mi [Thu, 16 Jan 2020 02:23:36 +0000 (18:23 -0800)]
[SampleFDO] Fix invalid branch profile generated by indirect call promotion.

Suppose an inline instance has hot total sample count but 0 entry count, and
it is an indirect call target. If the indirect call has no other call target
and inline instance associated with it and it is promoted, currently the
conditional branch generated by indirect call promotion will have invalid
branch profile which is !{!"branch_weights", i32 0, i32 0} -- because the
entry count of the promoted target is 0 and the total entry count of all
targets is also 0. This caused a SEGV in Control Height Reduction and may
cause problem in other passes.

Function entry count of an inline instance is computed by a heuristic --
using either the sample of the starting line or starting inner inline
instance. The patch changes the heuristic a little bit so that when total
sample count is larger than 0, the computed entry count will be at least 1.
Then the new branch profile will be !{!"branch_weights", i32 1, i32 0}.

Differential Revision: https://reviews.llvm.org/D72790

4 years ago[X86] When handling i64->f32 sint_to_fp on 32-bit targets only bitcast to f64 if...
Craig Topper [Thu, 16 Jan 2020 02:07:15 +0000 (18:07 -0800)]
[X86] When handling i64->f32 sint_to_fp on 32-bit targets only bitcast to f64 if sse2 is enabled.

The code is trying to copy the i64 value to an xmm register to
use a 64-bit store so that the 64-bit fild can benefit from
store forwarding.

But this trick only works if f64 is going to be stored in an
XMM register. If we only have SSE1 then only float is in xmm
register. So this trick just causes 2 stores i32 stores, an f64
load into the x87, an f64 from x87, and a 64-bit fild. So we end
up with an extra stack temporary and still didn't get store forwarding.

We might be able to use v2f32 here instead, but I didn't check. I
just wanted the code to make sense.

Found by inspection as I continue to stare too hard at our
int_to_fp conversions.

4 years ago[X86] Add 32-bit mode sse1 command line to scalar-int-to-fp.ll. NFC
Craig Topper [Thu, 16 Jan 2020 02:04:46 +0000 (18:04 -0800)]
[X86] Add 32-bit mode sse1 command line to scalar-int-to-fp.ll. NFC

4 years agoFix the macos build after D71575.
Jim Ingham [Thu, 16 Jan 2020 02:10:31 +0000 (18:10 -0800)]
Fix the macos build after D71575.

size_t and uint64_t are spelled slightly differently on macOS, which was
causing the compiler to error out calling std::min - since the two types have
to be the same.

I fixed this by casting the uint64_t computation to a size_t.  That's probably
not the cleanest solution, but it gets us back to building.

4 years agollc: Don't overwrite frame-pointer attribute
Matt Arsenault [Mon, 9 Dec 2019 13:09:58 +0000 (18:39 +0530)]
llc: Don't overwrite frame-pointer attribute

Continue making command line flags with matching attribute behavior
consistent.

4 years agoRevert "[Support] make report_fatal_error `abort` instead of `exit`"
Yuanfang Chen [Thu, 16 Jan 2020 01:46:18 +0000 (17:46 -0800)]
Revert "[Support] make report_fatal_error `abort` instead of `exit`"

This reverts commit 647c3f4e47de8a850ffcaa897db68702d8d2459a.

Got bots failure from sanitizer-windows and maybe others.

4 years agodebugserver: Cut dependency on intrinsics_gen
Vedant Kumar [Thu, 16 Jan 2020 00:57:45 +0000 (16:57 -0800)]
debugserver: Cut dependency on intrinsics_gen

debugserver does not depend on intrinsics_gen or on llvm.

4 years ago[Support] make report_fatal_error `abort` instead of `exit`
Yuanfang Chen [Wed, 15 Jan 2020 06:43:51 +0000 (22:43 -0800)]
[Support] make report_fatal_error `abort` instead of `exit`

Summary:
This patch could be treated as a rebase of D33960. It also fixes PR35547.
A fix for `llvm/test/Other/close-stderr.ll` is proposed in D68164. Seems
the consensus is that the test is passing by chance and I'm not
sure how important it is for us. So it is removed like in D33960 for now.
The rest of the test fixes are just adding `--crash` flag to `not` tool.

** The reason it fixes PR35547 is

`exit` does cleanup including calling class destructor whereas `abort`
does not do any cleanup. In multithreading environment such as ThinLTO or JIT,
threads may share states which mostly are ManagedStatic<>. If faulting thread
tearing down a class when another thread is using it, there are chances of
memory corruption. This is bad 1. It will stop error reporting like pretty
stack printer; 2. The memory corruption is distracting and nondeterministic in
terms of error message, and corruption type (depending one the timing, it
could be double free, heap free after use, etc.).

Reviewers: rnk, chandlerc, zturner, sepavloff, MaskRay, espindola

Reviewed By: rnk, MaskRay

Subscribers: wuzish, jholewinski, qcolombet, dschuff, jyknight, emaste, sdardis, nemanjai, jvesely, nhaehnle, sbc100, arichardson, jgravelle-google, aheejin, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, apazos, sabuasal, niosHD, jrtc27, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, PkmX, jocewei, jsji, lenary, s.egerton, pzheng, cfe-commits, MaskRay, filcab, davide, MatzeB, mehdi_amini, hiraditya, steven_wu, dexonsmith, rupprecht, seiya, llvm-commits

Tags: #llvm, #clang

Differential Revision: https://reviews.llvm.org/D67847

4 years ago[lldb/Tools] Remove lldb-mi.exports
Jonas Devlieghere [Thu, 16 Jan 2020 00:49:33 +0000 (16:49 -0800)]
[lldb/Tools] Remove lldb-mi.exports

lldb-mi was removed from the repo a while ago.

4 years agollc: Change behavior of -mattr with existing attribute
Matt Arsenault [Mon, 9 Dec 2019 12:33:25 +0000 (18:03 +0530)]
llc: Change behavior of -mattr with existing attribute

Append this to the existing target-features attribute on the function.

Some flags ignore existing attributes, and some overwrite them. Move
towards consistently respecting existing attributes if present. Since
target features act as a state machine on their own, append to the
function attribute. The backend default added feature list, function
attributes, and -mattr will all be appended together, and the later
features can individually toggle the earlier settings.

4 years ago[LLDB] Add ObjectFileWasm plugin for WebAssembly debugging
Paolo Severini [Wed, 15 Jan 2020 23:29:24 +0000 (15:29 -0800)]
[LLDB] Add ObjectFileWasm plugin for WebAssembly debugging

Summary:
This is the first in a series of patches to enable LLDB debugging of
WebAssembly targets.

Current versions of Clang emit (partial) DWARF debug information in WebAssembly
modules and we can leverage this debug information to give LLDB the ability to
do source-level debugging of Wasm code that runs in a WebAssembly engine.

A way to do this could be to use the remote debugging functionalities provided
by LLDB via the GDB-remote protocol. Remote debugging can indeed be useful not
only to connect a debugger to a process running on a remote machine, but also to
connect the debugger to a managed VM or script engine that runs locally,
provided that the engine implements a GDB-remote stub that offers the ability to
access the engine runtime internal state.

To make this work, the GDB-remote protocol would need to be extended with a few
Wasm-specific custom query commands, used to access aspects of the Wasm engine
state (like the Wasm memory, Wasm local and global variables, and so on).
Furthermore, the DWARF format would need to be enriched with a few Wasm-specific
extensions, here detailed: https://yurydelendik.github.io/webassembly-dwarf.

This CL introduce classes **ObjectFileWasm**, a file plugin to represent a Wasm
module loaded in a debuggee process. It knows how to parse Wasm modules and
store the Code section and the DWARF-specific sections.

Reviewers: jasonmolenda, clayborg, labath

Tags: #lldb

Differential Revision: https://reviews.llvm.org/D71575

4 years agoFix pack deduction to only deduce the arity of packs that are actually
Richard Smith [Wed, 15 Jan 2020 21:14:13 +0000 (13:14 -0800)]
Fix pack deduction to only deduce the arity of packs that are actually
expanded by the deduced pack.

We recently started also deducing the arity of separately-expanded packs
that are merely mentioned within the pack in question, which is
incorrect.

4 years ago[lldb/Utils] Patch all variables used by lit (3/3)
Jonas Devlieghere [Thu, 16 Jan 2020 00:01:00 +0000 (16:01 -0800)]
[lldb/Utils] Patch all variables used by lit (3/3)

Instead of passing all the arguments for dotest.py as a single CMake
variable, lit now uses separate variables for the different test
binaries. Before this change they'd all get patched as part of the
LLDB_DOTEST_ARGS. We need to patch the new variables as well.

4 years ago[lldb/Utils] Patch all variables used by lldb-dotest (2/2)
Jonas Devlieghere [Wed, 15 Jan 2020 23:55:51 +0000 (15:55 -0800)]
[lldb/Utils] Patch all variables used by lldb-dotest (2/2)

Instead of passing all the arguments for dotest.py as a single CMake
variable, lldb-dotest now uses separate variables for the different test
binaries. Before this change they'd all get patched as part of the
LLDB_DOTEST_ARGS. We need to patch the new variables as well.

4 years agoProcess BUNDLE in tail duplication
Stanislav Mekhanoshin [Wed, 15 Jan 2020 17:38:08 +0000 (09:38 -0800)]
Process BUNDLE in tail duplication

When tail duplication estimates a size of tail it uses instruction
count. Account for a number of instrictions in a bundle too.

Differential Revision: https://reviews.llvm.org/D72783

4 years agoRevert "Further implement CWG 2292"
Amy Huang [Wed, 15 Jan 2020 23:27:58 +0000 (15:27 -0800)]
Revert "Further implement CWG 2292"

This reverts commit ee0f1f1edc3ec0d4e698d50cc3180217448802b7 because it
causes an error on valid code.
See https://reviews.llvm.org/rGee0f1f1edc3ec0d4e698d50cc3180217448802b7.

4 years ago[OPENMP]Use regular processing of vtable used when TU is a prefix.
Alexey Bataev [Wed, 15 Jan 2020 22:37:12 +0000 (17:37 -0500)]
[OPENMP]Use regular processing of vtable used when TU is a prefix.

If current kind of the translation unit is TU_Prefix and it is not
complete, cannot decide what to do with virtual members/table at that
time, need to delay it to later stages.

4 years ago[CodeExtractor] Transfer debug info to extracted function
Vedant Kumar [Wed, 15 Jan 2020 19:26:34 +0000 (11:26 -0800)]
[CodeExtractor] Transfer debug info to extracted function

After extracting, fix up debug info in both the old and new functions by

1) Pointing line locations and debug intrinsics to the new subprogram
   scope, and

2) Deleting intrinsics which point to values outside of the new
   function.

Depends on https://reviews.llvm.org/D72795.

Testing: check-llvm, check-clang, a build of LNT in the `-Os -g` config
with "-mllvm -hot-cold-split=1" set, and end-to-end debugging of a toy
program which undergoes splitting to verify that lldb can find
variables, single step, etc. in extracted code.

rdar://45507940

Differential Revision: https://reviews.llvm.org/D72801

4 years agoAMDGPU/GlobalISel: Select exp with patterns
Matt Arsenault [Sun, 5 Jan 2020 15:51:54 +0000 (10:51 -0500)]
AMDGPU/GlobalISel: Select exp with patterns

This does produce slightly different code. Now a unique IMPLICIT_DEF
is emitted for each of the implicit_def operands, rather than reusing
the same one.

4 years agoAMDGPU: Remove custom node for exports
Matt Arsenault [Sun, 5 Jan 2020 00:49:17 +0000 (19:49 -0500)]
AMDGPU: Remove custom node for exports

I'm mildly worried about potentially reordering exp/exp_done with
IntrWriteMem on the intrinsic.

Requires hacking out the illegal type on SI, so manually select that
case during lowering.

4 years agoGlobalISel: Handle more cases of G_SEXT narrowing
Matt Arsenault [Fri, 10 Jan 2020 15:07:24 +0000 (10:07 -0500)]
GlobalISel: Handle more cases of G_SEXT narrowing

This now develops the same problem G_ZEXT/G_ANYEXT have where the
requested type is assumed to be the source type. This will be fixed
separately by creating intermediate merges.

4 years ago[IR] Module's NamedMD table needn't be 'void *'
Brian Gesiak [Wed, 15 Jan 2020 21:39:39 +0000 (16:39 -0500)]
[IR] Module's NamedMD table needn't be 'void *'

Summary:
In July 21 2010 `llvm::NamedMDNode` was refactored such that it would no
longer subclass `llvm::Value`:
https://github.com/llvm/llvm-project/commit/2637cc1a38d7336ea30caf

As part of this change, a map type from metadata names to their named
metadata, `llvm::MDSymbolTable`, was deleted. In its place, the type
of member `llvm::Module::NamedMDSymTab` was changed, from
`llvm::MDSymbolTable` to `void *`. The underlying memory allocations
for this pointer were changed to `new StringMap<NamedMDNode *>()`.

However, as far as I can tell, there's no need for obscuring the
underlying type being pointed to by the `void *`, and no need for
static casts from `void *` to `StringMap`. In fact, I don't think
there's a need for explicit calls to `new` and `delete` at all.

This commit changes `NamedMDSymTab` from a pointer to a reference, which
automatically couples its lifetime with the lifetime of its owning
`llvm::Module` instance, thus removing the explicit calls to `new` and
`delete` in the `llvm::Module` constructor and destructor. It also
changes the type from `void *` to a newly defined `NamedMDSymTabType`,
and removes the static casts.

Test Plan:
An ASAN-enabled build and run of `check-all` succeeds with this change
(aside from some tests that always fail for me in ASAN for some reason,
such as `check-clang` `SemaTemplate/stack-exhaustion.cpp`).

Reviewers: aprantl, dblaikie, chandlerc, pcc, echristo

Reviewed By: dblaikie

Subscribers: hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D72812

4 years ago[lldb/Utils] Patch all variables used by lldb-dotest
Jonas Devlieghere [Wed, 15 Jan 2020 23:15:51 +0000 (15:15 -0800)]
[lldb/Utils] Patch all variables used by lldb-dotest

Instead of passing all the arguments for dotest.py as a single CMake
variable, lldb-dotest now uses separate variables for the different test
binaries. Before this change they'd all get patched as part of the
LLDB_DOTEST_ARGS. We need to patch the new variables as well.

4 years ago[libcxx] Use mtx_plain | mtx_recursive following C11 API
Petr Hosek [Wed, 15 Jan 2020 21:58:29 +0000 (13:58 -0800)]
[libcxx] Use mtx_plain | mtx_recursive following C11 API

The C11 API specifies that to initialize a recursive mutex,
mtx_plain | mtx_recursive should be used with mtx_init.

Differential Revision: https://reviews.llvm.org/D72809

4 years ago[COFF] Warn that LLD does not support /PDBSTRIPPED:
Reid Kleckner [Wed, 15 Jan 2020 23:09:25 +0000 (15:09 -0800)]
[COFF] Warn that LLD does not support /PDBSTRIPPED:

Doesn't really fix PR44491, but it avoids treating it as an input.

4 years agoRevert "Allow system header to provide their own implementation of some builtin"
Amy Huang [Wed, 15 Jan 2020 22:34:19 +0000 (14:34 -0800)]
Revert "Allow system header to provide their own implementation of some builtin"

This reverts commit 921f871ac438175ca8fcfcafdfcfac4d7ddf3905 because it
causes libc++ code to trigger __warn_memset_zero_len.

See https://reviews.llvm.org/D71082.

4 years ago[lldb/Debugger] Rename IO handler methods to be more meaningful (NFC)
Jonas Devlieghere [Wed, 15 Jan 2020 22:56:28 +0000 (14:56 -0800)]
[lldb/Debugger] Rename IO handler methods to be more meaningful (NFC)

Make it clear form the method names whether they are synchronous or
asynchronous.