platform/upstream/llvm.git
5 years ago[MachineOutliner][NFC] Remove IntegerInstructionMap from InstructionMapper
Jessica Paquette [Thu, 6 Dec 2018 00:01:51 +0000 (00:01 +0000)]
[MachineOutliner][NFC] Remove IntegerInstructionMap from InstructionMapper

Refactoring.

This map was only used when we used a string of integers to output the outlined
sequence. Since it's no longer used for anything, there's no reason to keep it
around.

llvm-svn: 348432

5 years agoFix title underlines being too short after r348429
Pete Cooper [Thu, 6 Dec 2018 00:01:44 +0000 (00:01 +0000)]
Fix title underlines being too short after r348429

llvm-svn: 348431

5 years ago[GlobalISel] Introduce G_BUILD_VECTOR, G_BUILD_VECTOR_TRUNC and G_CONCAT_VECTOR opcodes.
Amara Emerson [Wed, 5 Dec 2018 23:53:30 +0000 (23:53 +0000)]
[GlobalISel] Introduce G_BUILD_VECTOR, G_BUILD_VECTOR_TRUNC and G_CONCAT_VECTOR opcodes.

These opcodes are intended to subsume some of the capability of G_MERGE_VALUES,
as it was too powerful and thus complex to add deal with throughout the GISel
pipeline.

G_BUILD_VECTOR creates a vector value from a sequence of uniformly typed
scalar values. G_BUILD_VECTOR_TRUNC is a special opcode for handling scalar
operands which are larger than the destination vector element type, and
therefore does an implicit truncate.

G_CONCAT_VECTOR creates a vector by concatenating smaller, uniformly typed,
vectors together.

These will be used in a subsequent commit. This commit just adds the initial
infrastructure.

Differential Revision: https://reviews.llvm.org/D53594

llvm-svn: 348430

5 years agoUpdate ARC docs as objc_storeStrong returns void not id
Pete Cooper [Wed, 5 Dec 2018 23:49:52 +0000 (23:49 +0000)]
Update ARC docs as objc_storeStrong returns void not id

llvm-svn: 348429

5 years ago[MachineOutliner][NFC] Remove buildCandidateList and replace with findCandidates
Jessica Paquette [Wed, 5 Dec 2018 23:39:07 +0000 (23:39 +0000)]
[MachineOutliner][NFC] Remove buildCandidateList and replace with findCandidates

More refactoring.

Since the pruning logic has changed, and the candidate list is gone,
everything can be sunk into findCandidates.

We no longer need to keep track of the length of the longest substring, so we
can drop all of that logic as well.

After this, we just find all of the candidates and move to outlining.

llvm-svn: 348428

5 years ago[MachineOutliner][NFC] Candidates don't need to be shared_ptrs anymore
Jessica Paquette [Wed, 5 Dec 2018 23:24:22 +0000 (23:24 +0000)]
[MachineOutliner][NFC] Candidates don't need to be shared_ptrs anymore

More refactoring.

After the changes to the pruning logic, and removing CandidateList, there's
no reason for Candiates to be shared_ptrs (or pointers at all).

std::shared_ptr<Candidate> -> Candidate.

llvm-svn: 348427

5 years agoRevert r347934 "[SCEV] Guard movement of insertion point for loop-invariants"
David L. Jones [Wed, 5 Dec 2018 23:13:50 +0000 (23:13 +0000)]
Revert r347934 "[SCEV] Guard movement of insertion point for loop-invariants"

This change caused SEGVs in instcombine. (The r347934 change seems to me to be a
precipitating cause, not a root cause. Details are on the llvm-commits thread
for r347934.)

llvm-svn: 348426

5 years agoFix test change from r348365 to deal with Windows paths correctly.
Douglas Yung [Wed, 5 Dec 2018 23:10:14 +0000 (23:10 +0000)]
Fix test change from r348365 to deal with Windows paths correctly.

llvm-svn: 348425

5 years ago[WebAssembly] Change event section code to 13
Heejin Ahn [Wed, 5 Dec 2018 23:10:09 +0000 (23:10 +0000)]
[WebAssembly] Change event section code to 13

Summary:
We decided to change the event section code from 12 to 13 as new
`DataCount` section in the bulk memory operations proposal will take the
code 12 instead.

Reviewers: sbc100

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

Differential Revision: https://reviews.llvm.org/D55343

llvm-svn: 348424

5 years ago[InstCombine] remove dead code from visitExtractElement
Sanjay Patel [Wed, 5 Dec 2018 23:09:33 +0000 (23:09 +0000)]
[InstCombine] remove dead code from visitExtractElement

Extracting from a splat constant is always handled by InstSimplify.
Move the test for this from InstCombine to InstSimplify to make
sure that stays true.

llvm-svn: 348423

5 years ago[MachineOutliner][NFC] Remove CandidateList, since it's now unused.
Jessica Paquette [Wed, 5 Dec 2018 22:50:26 +0000 (22:50 +0000)]
[MachineOutliner][NFC] Remove CandidateList, since it's now unused.

After removing the pruning logic, there's no reason to populate a list of
Candidates. Remove CandidateList and update comments.

llvm-svn: 348422

5 years agoFix buildbot capture warning
Jessica Paquette [Wed, 5 Dec 2018 22:47:25 +0000 (22:47 +0000)]
Fix buildbot capture warning

A bot didn't like my lambda. This ought to fix it.

Example:

http://lab.llvm.org:8011/builders/lld-x86_64-win7/builds/30139/steps/build%20lld/logs/stdio

error C3493: 'AlreadyRemoved' cannot be implicitly captured because no default
capture mode has been specified

llvm-svn: 348421

5 years ago[MachineOutliner][NFC] Simplify and unify pruning/outlining logic
Jessica Paquette [Wed, 5 Dec 2018 22:27:38 +0000 (22:27 +0000)]
[MachineOutliner][NFC] Simplify and unify pruning/outlining logic

Since we're now performing outlining per OutlinedFunction rather than per
Candidate, we can simply outline each candidate as it shows up.

Instead of having a pruning phase, instead, we'll outline entire functions.
Then we'll update the UnsignedVec we mapped to reflect the deletion. If any
candidate is in a space that's marked dirty, then we'll drop it.

This lets us remove the pruning logic entirely, and greatly simplifies the
code.

llvm-svn: 348420

5 years ago[Hexagon] Add intrinsics for Hexagon V66
Krzysztof Parzyszek [Wed, 5 Dec 2018 22:03:04 +0000 (22:03 +0000)]
[Hexagon] Add intrinsics for Hexagon V66

llvm-svn: 348419

5 years ago[InstCombine] reduce duplication in visitExtractElementInst; NFC
Sanjay Patel [Wed, 5 Dec 2018 21:57:51 +0000 (21:57 +0000)]
[InstCombine] reduce duplication in visitExtractElementInst; NFC

llvm-svn: 348418

5 years ago[InstCombine] add/move tests for extractelement; NFC
Sanjay Patel [Wed, 5 Dec 2018 21:56:13 +0000 (21:56 +0000)]
[InstCombine] add/move tests for extractelement; NFC

llvm-svn: 348417

5 years agoThinLTO: Do not import debug info for imported global constants
David Blaikie [Wed, 5 Dec 2018 21:42:17 +0000 (21:42 +0000)]
ThinLTO: Do not import debug info for imported global constants

It looks like this isn't necessary (in any tests I've done, it results
in the global being described with no location or value in the imported
side - while it's still fully described in the place it's imported from)
& results in significant/pathological debug info growth to home these
location-less global variable descriptions on the import side.

This is a rather pressing/important issue to address - this regressed
executable size for one example I'm looking at by 15%, object size is probably
similar though I haven't measured it, and a 22x increase in the number of CUs
in the cu_index in split DWARF DWP files, creating a similarly large regression
in the time it takes llvm-symbolizer to run on such binaries.

Reviewers: tejohnson, evgeny777

Differential Revision: https://reviews.llvm.org/D55309

llvm-svn: 348416

5 years ago[Hexagon] Add support for Hexagon V66
Krzysztof Parzyszek [Wed, 5 Dec 2018 21:38:35 +0000 (21:38 +0000)]
[Hexagon] Add support for Hexagon V66

llvm-svn: 348415

5 years ago[MachineOutliner] Outline functions by order of benefit
Jessica Paquette [Wed, 5 Dec 2018 21:36:04 +0000 (21:36 +0000)]
[MachineOutliner] Outline functions by order of benefit

Mostly NFC, only change is the order of outlined function names.

Loop over the outlined functions instead of walking the candidate list.

This is a bit easier to understand. It's far more natural to create a function,
then replace all of its occurrences with calls than the other way around.

The functions outlined after this do not change, but their names will be
decided by their benefit. E.g, OUTLINED_FUNCTION_0 will now always be the
most beneficial function, rather than the first one seen.

This makes it easier to enforce an ordering on the outlined functions. So,
this also adds a test to make sure that the ordering works as expected.

llvm-svn: 348414

5 years ago[Hexagon] Add intrinsics for Hexagon V66
Krzysztof Parzyszek [Wed, 5 Dec 2018 21:14:51 +0000 (21:14 +0000)]
[Hexagon] Add intrinsics for Hexagon V66

llvm-svn: 348413

5 years agoNFC: Extract TextNodeDumper class
Stephen Kelly [Wed, 5 Dec 2018 21:12:39 +0000 (21:12 +0000)]
NFC: Extract TextNodeDumper class

Summary:
Start by moving some utilities to it. It will eventually house dumping
of individual nodes (after indentation etc has already been accounted
for).

Reviewers: aaron.ballman

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D55189

llvm-svn: 348412

5 years ago[Hexagon] Add instruction definitions for Hexagon V66
Krzysztof Parzyszek [Wed, 5 Dec 2018 21:01:07 +0000 (21:01 +0000)]
[Hexagon] Add instruction definitions for Hexagon V66

llvm-svn: 348411

5 years agoNFC: Extract TextTreeStructure class
Stephen Kelly [Wed, 5 Dec 2018 20:53:14 +0000 (20:53 +0000)]
NFC: Extract TextTreeStructure class

Reviewers: aaron.ballman

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D55188

llvm-svn: 348410

5 years agoNFC: Inline handling of DependentSizedArrayType
Stephen Kelly [Wed, 5 Dec 2018 20:34:07 +0000 (20:34 +0000)]
NFC: Inline handling of DependentSizedArrayType

Summary:
Re-order handling of getElementType and getBracketsRange. It is
necessary to perform all printing before any traversal to child nodes.

This causes no change in the output of ast-dump-array.cpp due to the way
child nodes are printed with a delay.  This new order of the code is
also the order that produces the expected output anyway.

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D55257

llvm-svn: 348409

5 years agoAdd dump tests for inherited default template parameters
Stephen Kelly [Wed, 5 Dec 2018 20:21:13 +0000 (20:21 +0000)]
Add dump tests for inherited default template parameters

llvm-svn: 348408

5 years ago[Hexagon] Foundation of support for Hexagon V66
Krzysztof Parzyszek [Wed, 5 Dec 2018 20:18:09 +0000 (20:18 +0000)]
[Hexagon] Foundation of support for Hexagon V66

llvm-svn: 348407

5 years ago[GISel]: Provide standard interface to observe changes in GISel passes
Aditya Nandakumar [Wed, 5 Dec 2018 20:14:52 +0000 (20:14 +0000)]
[GISel]: Provide standard interface to observe changes in GISel passes

https://reviews.llvm.org/D54980

This provides a standard API across GISel passes to observe and notify
passes about changes (insertions/deletions/mutations) to MachineInstrs.
This patch also removes the recordInsertion method in MachineIRBuilder
and instead provides method to setObserver.

Reviewed by: vkeles.

llvm-svn: 348406

5 years ago[CodeExtractor] Do not marked outlined calls which may resume EH as noreturn
Vedant Kumar [Wed, 5 Dec 2018 19:35:37 +0000 (19:35 +0000)]
[CodeExtractor] Do not marked outlined calls which may resume EH as noreturn

Treat terminators which resume exception propagation as returning instructions
(at least, for the purposes of marking outlined functions `noreturn`). This is
to avoid inserting traps after calls to outlined functions which unwind.

rdar://46129950

llvm-svn: 348404

5 years ago[X86][SSE] Fix a copy+paste typo that was folding the sext/zext of partial vectors
Simon Pilgrim [Wed, 5 Dec 2018 19:32:19 +0000 (19:32 +0000)]
[X86][SSE] Fix a copy+paste typo that was folding the sext/zext of partial vectors

llvm-svn: 348403

5 years agoRevert "[RISCV] Mark unit tests as "requires: riscv-registered-target""
Mandeep Singh Grang [Wed, 5 Dec 2018 19:19:38 +0000 (19:19 +0000)]
Revert "[RISCV] Mark unit tests as "requires: riscv-registered-target""

This reverts commit 8908dd12e7bbfc74e264233e900206ad31e285f0.

llvm-svn: 348402

5 years agoDo not use a hash table to uniquify mergeable strings.
Rui Ueyama [Wed, 5 Dec 2018 19:13:31 +0000 (19:13 +0000)]
Do not use a hash table to uniquify mergeable strings.

Previously, we have a hash table containing strings and their offsets
to manage mergeable strings. Technically we can live without that, because
we can do binary search on a vector of mergeable strings to find a mergeable
strings.

We did have both the hash table and the binary search because we thought
that that is faster.

We recently observed that lld tend to consume more memory than gold when
building an output with debug info. A few percent of memory is consumed by
the hash table. So, we needed to reevaluate whether or not having the extra
hash table is a good CPU/memory tradeoff. I run a few benchmarks with and
without the hash table.

I got a mixed result for the benchmark. We observed a regression for some
programs by removing the hash table (that's what we expected), but we also
observed that performance imrpovements for some programs. This is perhaps
due to reduced memory usage.

Differential Revision: https://reviews.llvm.org/D55234

llvm-svn: 348401

5 years agoDo not check for parameters shadowing fields in function declarations.
Aaron Ballman [Wed, 5 Dec 2018 18:56:57 +0000 (18:56 +0000)]
Do not check for parameters shadowing fields in function declarations.

We would issue a false-positive diagnostic for parameters in function declarations shadowing fields; we now only issue the diagnostic on a function definition instead.

llvm-svn: 348400

5 years agoAdding tests for -ast-dump; NFC.
Aaron Ballman [Wed, 5 Dec 2018 18:53:57 +0000 (18:53 +0000)]
Adding tests for -ast-dump; NFC.

This adds tests for various function and class template declarations.

llvm-svn: 348399

5 years ago[AArch64] Reword description of feature (NFC)
Evandro Menezes [Wed, 5 Dec 2018 18:42:57 +0000 (18:42 +0000)]
[AArch64] Reword description of feature (NFC)

Reword the description of the feature that enables custom handling of cheap
instructions.

llvm-svn: 348398

5 years agoHonor -fdebug-prefix-map when creating function names for the debug info.
Adrian Prantl [Wed, 5 Dec 2018 18:37:44 +0000 (18:37 +0000)]
Honor -fdebug-prefix-map when creating function names for the debug info.

This adds a callback to PrintingPolicy to allow CGDebugInfo to remap
file paths according to -fdebug-prefix-map. Otherwise the debug info
(particularly function names for C++ lambdas) may contain paths that
should have been remapped in the debug info.

<rdar://problem/46128056>

Differential Revision: https://reviews.llvm.org/D55137

llvm-svn: 348397

5 years ago[analyzer] Attribute for RetainCountChecker for OSObject should propagate with inheri...
George Karpenkov [Wed, 5 Dec 2018 18:34:54 +0000 (18:34 +0000)]
[analyzer] Attribute for RetainCountChecker for OSObject should propagate with inheritance

rdar://46388388

Differential Revision: https://reviews.llvm.org/D55154

llvm-svn: 348396

5 years ago[llvm-mca] Simplify test (NFC)
Evandro Menezes [Wed, 5 Dec 2018 18:34:51 +0000 (18:34 +0000)]
[llvm-mca] Simplify test (NFC)

llvm-svn: 348395

5 years agoMention changes to libc++ include dir lookup in release notes.
Ilya Biryukov [Wed, 5 Dec 2018 18:32:05 +0000 (18:32 +0000)]
Mention changes to libc++ include dir lookup in release notes.

Summary: The change itself landed as r348365, see the comment for more details.

Reviewers: arphaman, EricWF

Reviewed By: arphaman

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D55322

llvm-svn: 348394

5 years ago[llvm-mca] Sort test run lines (NFC)
Evandro Menezes [Wed, 5 Dec 2018 18:30:06 +0000 (18:30 +0000)]
[llvm-mca] Sort test run lines (NFC)

llvm-svn: 348393

5 years ago[MachineOutliner][NFC] Use getOccurrenceCount() in getNotOutlinedCost()
Jessica Paquette [Wed, 5 Dec 2018 18:17:40 +0000 (18:17 +0000)]
[MachineOutliner][NFC] Use getOccurrenceCount() in getNotOutlinedCost()

Some more gardening.

llvm-svn: 348392

5 years ago[MachineOutliner][NFC] Make getters in MachineOutliner.h const
Jessica Paquette [Wed, 5 Dec 2018 18:12:52 +0000 (18:12 +0000)]
[MachineOutliner][NFC] Make getters in MachineOutliner.h const

Just some refactoring. A few of the getters in OutlinedFunction weren't const.

llvm-svn: 348391

5 years ago[MachineOutliner][NFC] Don't create outlined sequence from integer mapping
Jessica Paquette [Wed, 5 Dec 2018 17:57:33 +0000 (17:57 +0000)]
[MachineOutliner][NFC] Don't create outlined sequence from integer mapping

Some gardening/refactoring.

It's cleaner to copy the instructions into the MachineFunction using the first
candidate instead of going to the mapper.

Also, by doing this we can remove the Seq member from OutlinedFunction entirely.

llvm-svn: 348390

5 years ago[gold-plugin] allow function/data sections to be toggleable
Nick Desaulniers [Wed, 5 Dec 2018 17:46:24 +0000 (17:46 +0000)]
[gold-plugin] allow function/data sections to be toggleable

Summary:
r336838 allowed these to be toggleable.
r336858 reverted r336838.
r336943 made the generation of these sections conditional on LDPO_REL.

This commit brings back the toggle-ability.  You can specify:
-plugin-opt=-function-sections
-plugin-opt=-data-sections
For your linker flags to disable the changes made in r336943.

Without toggling r336943 off, arm64 linux kernels linked with gold-plugin
see significant boot time regressions, but with r336943 outright reverted
x86_64 linux kernels linked with gold-plugin fail to boot.

Reviewers: pcc, void

Reviewed By: pcc

Subscribers: javed.absar, kristof.beyls, llvm-commits, srhines

Differential Revision: https://reviews.llvm.org/D55291

llvm-svn: 348389

5 years agoAddress a post-commit review comment on r348325.
Erik Pilkington [Wed, 5 Dec 2018 17:43:26 +0000 (17:43 +0000)]
Address a post-commit review comment on r348325.

llvm-svn: 348388

5 years ago[CodeComplete] Fix a crash in access checks of inner classes
Ilya Biryukov [Wed, 5 Dec 2018 17:38:39 +0000 (17:38 +0000)]
[CodeComplete] Fix a crash in access checks of inner classes

Summary: The crash was introduced in r348135.

Reviewers: kadircet

Reviewed By: kadircet

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D55260

llvm-svn: 348387

5 years ago[PDB] Make PDB lit tests use the new builder
Aleksandr Urakov [Wed, 5 Dec 2018 17:35:30 +0000 (17:35 +0000)]
[PDB] Make PDB lit tests use the new builder

Reviewers: zturner, stella.stamenova

Reviewed By: zturner

Tags: #lldb

Differential Revision: https://reviews.llvm.org/D54942

llvm-svn: 348386

5 years agoAMDGPU: Fix using old address spaces in some tests
Matt Arsenault [Wed, 5 Dec 2018 17:34:59 +0000 (17:34 +0000)]
AMDGPU: Fix using old address spaces in some tests

llvm-svn: 348385

5 years ago[Basic] Cleanups in IdentifierInfo following the removal of PTH
Bruno Ricci [Wed, 5 Dec 2018 17:16:55 +0000 (17:16 +0000)]
[Basic] Cleanups in IdentifierInfo following the removal of PTH

The Entry pointer in IdentifierInfo was only null for IdentifierInfo
created from a PTH. Now that PTH support has been removed we can remove
some PTH specific code in IdentifierInfo::getLength and
IdentifierInfo::getNameStart.

Also make the constructor of IdentifierInfo private to make sure that
they are only created by IdentifierTable, and move it to the header so
that it can be inlined in IdentifierTable::get and IdentifierTable::getOwn.

Differential Revision: https://reviews.llvm.org/D54866

Reviewed By: erichkeane

llvm-svn: 348384

5 years ago[DAGCombiner] don't try to extract a fraction of a vector binop and crash (PR39893)
Sanjay Patel [Wed, 5 Dec 2018 17:10:30 +0000 (17:10 +0000)]
[DAGCombiner] don't try to extract a fraction of a vector binop and crash (PR39893)

Because we're potentially peeking through a bitcast in this transform,
we need to use overall bitwidths rather than number of elements to
determine when it's safe to proceed.

Should fix:
https://bugs.llvm.org/show_bug.cgi?id=39893

llvm-svn: 348383

5 years ago[OpenCL] Diagnose conflicting address spaces in templates.
Anastasia Stulova [Wed, 5 Dec 2018 17:02:22 +0000 (17:02 +0000)]
[OpenCL] Diagnose conflicting address spaces in templates.

Added new diagnostic when templates are instantiated with
different address space from the one provided in its definition.

This also prevents deducing generic address space in pointer
type of templates to allow giving them concrete address space
during instantiation.

Differential Revision: https://reviews.llvm.org/D55127

llvm-svn: 348382

5 years agoAllow norecurse attribute on functions that have debug infos.
Christian Bruel [Wed, 5 Dec 2018 16:48:00 +0000 (16:48 +0000)]
Allow norecurse attribute on functions that have debug infos.

Summary: debug intrinsics might be marked norecurse to enable the caller function to be norecurse and optimized if needed. This avoids code gen optimisation differences when -g is used, as in globalOpt.cpp:processInternalGlobal checks.

Reviewers: chandlerc, jmolloy, aprantl

Reviewed By: aprantl

Subscribers: aprantl, llvm-commits

Differential Revision: https://reviews.llvm.org/D55187

llvm-svn: 348381

5 years ago[X86] Add test case to show missed opportunity to combine a concat_vector into a...
Andrea Di Biagio [Wed, 5 Dec 2018 16:23:27 +0000 (16:23 +0000)]
[X86] Add test case to show missed opportunity to combine a concat_vector into a scalar_to_vector. NFC

This is a test for D55274.

llvm-svn: 348380

5 years ago[NFC] Use clang-format on PrintingPolicy::PrintingPolicy() after fd5c386f743
Jan Korous [Wed, 5 Dec 2018 16:19:49 +0000 (16:19 +0000)]
[NFC] Use clang-format on PrintingPolicy::PrintingPolicy() after fd5c386f743

The white-space change was causing conflicts downstream.

rdar://problem/46486841

llvm-svn: 348379

5 years agoRemove XFAIL in get_module_and_offset_for_pc.cc for NetBSD-MSan
Kamil Rytarowski [Wed, 5 Dec 2018 16:05:25 +0000 (16:05 +0000)]
Remove XFAIL in get_module_and_offset_for_pc.cc for NetBSD-MSan

After updating GET_LINK_MAP_BY_DLOPEN_HANDLE() for recent NetBSD
this test no longer fails.

llvm-svn: 348378

5 years ago[Sanitizer] nl_langinfo forgotten bit.
David Carlier [Wed, 5 Dec 2018 16:02:26 +0000 (16:02 +0000)]
[Sanitizer] nl_langinfo forgotten bit.

M    lib/sanitizer_common/sanitizer_platform_interceptors.h

llvm-svn: 348377

5 years ago[Sanitizer] expand nl_langinfo interception to FreeBSD
David Carlier [Wed, 5 Dec 2018 16:01:22 +0000 (16:01 +0000)]
[Sanitizer] expand nl_langinfo interception to FreeBSD

Reviewers: krytarowski

Reviewed By: krytarowski

Differential Revision: https://reviews.llvm.org/D55327

llvm-svn: 348376

5 years agoRevert "[IR] Add NODISCARD to attribute functions"
Brian Gesiak [Wed, 5 Dec 2018 15:56:09 +0000 (15:56 +0000)]
Revert "[IR] Add NODISCARD to attribute functions"

Revert https://reviews.llvm.org/D55217 due to warnings-turned-into-errors in
AMGPU targets. I'll fix the warnings first, then re-commit this patch.

llvm-svn: 348375

5 years ago[SLH] Fix a nasty bug in SLH.
Chandler Carruth [Wed, 5 Dec 2018 15:42:11 +0000 (15:42 +0000)]
[SLH] Fix a nasty bug in SLH.

Whenever we effectively take the address of a basic block we need to
manually update that basic block to reflect that fact or later passes
such as tail duplication and tail merging can break the invariants of
the code. =/ Sadly, there doesn't appear to be any good way of
automating this or even writing a reasonable assert to catch it early.

The change seems trivially and obviously correct, but sadly the only
really good test case I have is 1000s of basic blocks. I've tried
directly writing a test case that happens to make tail duplication do
something that crashes later on, but this appears to require an
*amazingly* complex set of conditions that I've not yet reproduced.

The change is technically covered by the tests because we mark the
blocks as having their address taken, but that doesn't really count as
properly testing the functionality.

llvm-svn: 348374

5 years ago[SLH] Regenerate tests with --no_x86_scrub_rip to restore the higher
Chandler Carruth [Wed, 5 Dec 2018 15:41:13 +0000 (15:41 +0000)]
[SLH] Regenerate tests with --no_x86_scrub_rip to restore the higher
fidelity checking of RIP-based references to basic blocks and other
labels.

These labels are super important for SLH tests so we should keep them
readable in the test cases.

llvm-svn: 348373

5 years ago[IR] Add NODISCARD to attribute functions
Brian Gesiak [Wed, 5 Dec 2018 15:33:55 +0000 (15:33 +0000)]
[IR] Add NODISCARD to attribute functions

Summary:
Many functions on `llvm::AttributeList` and `llvm::AttributeSet` are
documented with "returns a new {list,set} because attribute
{lists,sets} are immutable." This documentation can be aided by the
addition of an attribute, `LLVM_NODISCARD`. Adding this prevents
unsuspecting users of the API from expecting
`AttributeList::setAttributes` from modifying the underlying list.

At the very least, it would have saved me a few hours of debugging, since I
had been doing just that! I had a bug in my program where I was calling
`setAttributes` but then passing in the unmutated `AttributeList`.
I tried adding LLVM_NODISCARD and confirmed that it would have made my bug
immediately obvious.

Reviewers: rnk, javed.absar

Reviewed By: rnk

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D55217

llvm-svn: 348372

5 years ago[AMDGPU]: Turn on the DPP combiner by default
Valery Pykhtin [Wed, 5 Dec 2018 15:21:17 +0000 (15:21 +0000)]
[AMDGPU]: Turn on the DPP combiner by default

Differential revision: https://reviews.llvm.org/D55314

llvm-svn: 348371

5 years agoAdd a new interceptor for modctl(2) from NetBSD
Kamil Rytarowski [Wed, 5 Dec 2018 15:13:20 +0000 (15:13 +0000)]
Add a new interceptor for modctl(2) from NetBSD

Summary:
modctl - controls loadable kernel modules.

Skip tests as this call uses privileged operations.

Reviewers: vitalybuka, joerg

Reviewed By: vitalybuka

Subscribers: kubamracek, llvm-commits, mgorny, #sanitizers

Tags: #sanitizers

Differential Revision: https://reviews.llvm.org/D55025

llvm-svn: 348370

5 years agoAdd a new interceptor for nl_langinfo(3) from NetBSD
Kamil Rytarowski [Wed, 5 Dec 2018 15:06:53 +0000 (15:06 +0000)]
Add a new interceptor for nl_langinfo(3) from NetBSD

Summary:
nl_langinfo - gets locale information.

Add a dedicated test.

Reviewers: vitalybuka, joerg

Reviewed By: vitalybuka

Subscribers: kubamracek, llvm-commits, mgorny, #sanitizers

Tags: #sanitizers

Differential Revision: https://reviews.llvm.org/D55015

llvm-svn: 348369

5 years ago[Haiku] Support __float128 for x86 and x86_64
Kristina Brooks [Wed, 5 Dec 2018 15:05:06 +0000 (15:05 +0000)]
[Haiku] Support __float128 for x86 and x86_64

This patch addresses a compilation error with clang when
running in Haiku being unable to compile code using
float128 (throws compilation error such as 'float128 is
not supported on this target').

Patch by kallisti5 (Alexander von Gluck IV)

Differential Revision: https://reviews.llvm.org/D54901

llvm-svn: 348368

5 years ago[InstCombine] simplify icmps with same operands based on dominating cmp
Sanjay Patel [Wed, 5 Dec 2018 15:04:00 +0000 (15:04 +0000)]
[InstCombine] simplify icmps with same operands based on dominating cmp

The tests here are based on the motivating cases from D54827.

More background:
1. We don't get these cases in general with SimplifyCFG because the root
   of the pattern match is an icmp, not a branch. I'm not sure how often
   we encounter this pattern vs. the seemingly more likely case with
   branches, but I don't see evidence to leave the minimal pattern
   unoptimized.

2. This has a chance of increasing compile-time because we're using a
   ValueTracking call to handle the match. The motivating cases could be
   handled with a simpler pair of calls to isImpliedTrueByMatchingCmp/
   isImpliedFalseByMatchingCmp, but I saw that we have a more
   comprehensive wrapper around those, so we might as well use it here
   unless there's evidence that it's significantly slower.

3. Ideally, we'd handle the fold to constants in InstSimplify, but as
   with the existing code here, we could extend this to handle cases
   where the result is not a constant, but a new combined predicate.
   That would mean splitting the logic across the 2 passes and possibly
   duplicating the pattern-matching cost.

4. As mentioned in D54827, this seems like the kind of thing that should
   be handled in Correlated Value Propagation, but that pass is currently
   limited to dealing with instructions with constant operands, so extending
   this bit of InstCombine is the smallest/easiest way to get these patterns
   optimized.

llvm-svn: 348367

5 years ago[X86][SSE] Begun adding modulo rotate support to LowerRotate
Simon Pilgrim [Wed, 5 Dec 2018 14:46:37 +0000 (14:46 +0000)]
[X86][SSE] Begun adding modulo rotate support to LowerRotate

Prep work for PR38243 - mainly adding comments on where we need to add modulo support (doing so at the moment causes massive codegen regressions).

I've also consistently added support for modulo folding for uniform constants (although at the moment we have no way to trigger this) and removed the old assertions.

llvm-svn: 348366

5 years agoMove detection of libc++ include dirs to Driver on MacOS
Ilya Biryukov [Wed, 5 Dec 2018 14:24:14 +0000 (14:24 +0000)]
Move detection of libc++ include dirs to Driver on MacOS

Summary:
The intention is to make the tools replaying compilations from 'compile_commands.json'
(clang-tidy, clangd, etc.) find the same standard library as the original compiler
specified in 'compile_commands.json'.

Previously, the library detection logic was in the frontend (InitHeaderSearch.cpp) and relied
on the value of resource dir as an approximation of the compiler install dir. The new logic
uses the actual compiler install dir and is performed in the driver. This is consistent with
the C++ standard library detection on other platforms and allows to override the resource dir
in the tools using the compile_commands.json without altering the
standard library detection mechanism. The tools have to override the resource dir to make sure
they use a consistent version of the builtin headers.

There is still logic in InitHeaderSearch that attemps to add the absolute includes for the
the C++ standard library, so we keep passing the -stdlib=libc++ from the driver to the frontend
via cc1 args to avoid breaking that. In the long run, we should move this logic to the driver too,
but it could potentially break the library detection on other systems, so we don't tackle it in this
patch to keep its scope manageable.

This is a second attempt to fix the issue, first one was commited in r346652 and reverted in r346675.
The original fix relied on an ad-hoc propagation (bypassing the cc1 flags) of the install dir from the
driver to the frontend's HeaderSearchOptions. Unsurpisingly, the propagation was incomplete, it broke
the libc++ detection in clang itself, which caused LLDB tests to break.

The LLDB tests pass with new fix.

Reviewers: JDevlieghere, arphaman, EricWF

Reviewed By: arphaman

Subscribers: mclow.lists, ldionne, dexonsmith, ioeric, christof, kadircet, cfe-commits

Differential Revision: https://reviews.llvm.org/D54630

llvm-svn: 348365

5 years agoRevert: Honor -fdebug-prefix-map when creating function names for the debug info.
Renato Golin [Wed, 5 Dec 2018 13:56:26 +0000 (13:56 +0000)]
Revert: Honor -fdebug-prefix-map when creating function names for the debug info.

This commit reverts r348060 and r348062 due to it breaking the AArch64 Full
buildbot: https://bugs.llvm.org/show_bug.cgi?id=39892

llvm-svn: 348364

5 years ago[llvm-rc] Support not expressions.
Martin Storsjo [Wed, 5 Dec 2018 13:22:56 +0000 (13:22 +0000)]
[llvm-rc] Support not expressions.

Patch by Jacek Caban!

Differential Revision: https://reviews.llvm.org/D55242

llvm-svn: 348363

5 years ago[TargetLowering] Remove ISD::ANY_EXTEND/ANY_EXTEND_VECTOR_INREG opcodes from Simplify...
Simon Pilgrim [Wed, 5 Dec 2018 12:20:05 +0000 (12:20 +0000)]
[TargetLowering] Remove ISD::ANY_EXTEND/ANY_EXTEND_VECTOR_INREG opcodes from SimplifyDemandedVectorElts

These have no test coverage and the KnownZero flags can't be guaranteed unlike SIGN/ZERO_EXTEND cases.

llvm-svn: 348361

5 years ago[CMake] Fix side-effect from LLDB_VERSION change in r346668 for framework-enabled...
Stefan Granitz [Wed, 5 Dec 2018 12:18:44 +0000 (12:18 +0000)]
[CMake] Fix side-effect from LLDB_VERSION change in r346668 for framework-enabled builds

llvm-svn: 348360

5 years ago[clangd] Dont provide locations for non-existent files.
Kadir Cetinkaya [Wed, 5 Dec 2018 11:57:15 +0000 (11:57 +0000)]
[clangd] Dont provide locations for non-existent files.

Summary:
We were getting assertion errors when we had bad file names, instead we
should skip those.

Reviewers: hokein

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

Differential Revision: https://reviews.llvm.org/D55275

llvm-svn: 348359

5 years ago[DAG] Add fshl/fshr tblgen opcodes
Simon Pilgrim [Wed, 5 Dec 2018 11:55:33 +0000 (11:55 +0000)]
[DAG] Add fshl/fshr tblgen opcodes

Missed off from https://reviews.llvm.org/D54698

llvm-svn: 348358

5 years agoFix compilation error when using clang 3.6.0
Mikael Holmen [Wed, 5 Dec 2018 11:29:27 +0000 (11:29 +0000)]
Fix compilation error when using clang 3.6.0

llvm-svn: 348357

5 years ago[test] Disable Modules/prune.m on NetBSD as it requires 'touch -a'
Michal Gorny [Wed, 5 Dec 2018 11:17:50 +0000 (11:17 +0000)]
[test] Disable Modules/prune.m on NetBSD as it requires 'touch -a'

llvm-svn: 348356

5 years ago[test] Skip ThinLTO cache tests requiring atime setting on NetBSD
Michal Gorny [Wed, 5 Dec 2018 11:15:50 +0000 (11:15 +0000)]
[test] Skip ThinLTO cache tests requiring atime setting on NetBSD

Skip the ThinLTO cache tests on NetBSD.  They require 'touch' being
able to alter atime of files, while NetBSD inhibits atime updates
when filesystem is mounted noatime.

Differential Revision: https://reviews.llvm.org/D55273

llvm-svn: 348355

5 years ago[test] Split strip-preserve-time.test, and skip atime test on NetBSD
Michal Gorny [Wed, 5 Dec 2018 11:15:46 +0000 (11:15 +0000)]
[test] Split strip-preserve-time.test, and skip atime test on NetBSD

Split timestamp preservation tests into atime and mtime test, and skip
the former on NetBSD.  When the filesystem is mounted noatime, NetBSD
not only inhibits implicit atime updates but also prevents setting atime
via utime(), causing the test to fail.

Differential Revision: https://reviews.llvm.org/D55271

llvm-svn: 348354

5 years ago[SelectionDAG] Initial support for FSHL/FSHR funnel shift opcodes (PR39467)
Simon Pilgrim [Wed, 5 Dec 2018 11:12:12 +0000 (11:12 +0000)]
[SelectionDAG] Initial support for FSHL/FSHR funnel shift opcodes (PR39467)

This is an initial patch to add a minimum level of support for funnel shifts to the SelectionDAG and to begin wiring it up to the X86 SHLD/SHRD instructions.

Some partial legalization code has been added to handle the case for 'SlowSHLD' where we want to expand instead and I've added a few DAG combines so we don't get regressions from the existing DAG builder expansion code.

Differential Revision: https://reviews.llvm.org/D54698

llvm-svn: 348353

5 years ago[clang] - Simplify tools::SplitDebugName.
George Rimar [Wed, 5 Dec 2018 11:09:10 +0000 (11:09 +0000)]
[clang] - Simplify tools::SplitDebugName.

This is an updated version of the D54576, which was reverted.

Problem was that SplitDebugName calls the InputInfo::getFilename
which asserts if InputInfo given is not of type Filename:

const char *getFilename() const {
  assert(isFilename() && "Invalid accessor.");
  return Data.Filename;
}
At the same time at that point, it can be of type Nothing and
we need to use getBaseInput(), like original code did.

Differential revision: https://reviews.llvm.org/D55006

llvm-svn: 348352

5 years ago[MC] - Fix build bot.
George Rimar [Wed, 5 Dec 2018 11:06:29 +0000 (11:06 +0000)]
[MC] - Fix build bot.

Error was:
/home/buildslave/slave_as-bldslv8/lld-perf-testsuite/llvm/lib/MC/MCFragment.cpp:241:22: error: field 'Offset' will be initialized after field 'LayoutOrder' [-Werror,-Wreorder]
      Atom(nullptr), Offset(~UINT64_C(0)), LayoutOrder(0) {

http://lab.llvm.org:8011/builders/lld-perf-testsuite/builds/9628/steps/build-bin%2Flld/logs/stdio

llvm-svn: 348351

5 years agoRemove superfluous comments. NFCI.
Simon Pilgrim [Wed, 5 Dec 2018 10:45:44 +0000 (10:45 +0000)]
Remove superfluous comments. NFCI.

As requested in D54698.

llvm-svn: 348350

5 years agoRecommit r348243 - "[llvm-mc] - Do not crash when referencing undefined debug sections."
George Rimar [Wed, 5 Dec 2018 10:43:58 +0000 (10:43 +0000)]
Recommit r348243 - "[llvm-mc] - Do not crash when referencing undefined debug sections."

The patch triggered an unrelated msan issue: LayoutOrder variable was not initialized.
(http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-fast/builds/26794/steps/check-llvm%20msan/logs/stdio)
It was fixed.

Original commit message:
MC has code that pre-creates few debug sections:
https://github.com/llvm-mirror/llvm/blob/master/lib/MC/MCObjectFileInfo.cpp#L396

If users code has a reference to such section but does not redefine it,
MC code currently asserts, because still thinks they are normally defined.

The patch fixes the issue.

Differential revision: https://reviews.llvm.org/D55173
----
Modified : /llvm/trunk/lib/MC/ELFObjectWriter.cpp
Added : /llvm/trunk/test/MC/ELF/undefined-debug.s

llvm-svn: 348349

5 years ago[TargetLowering] SimplifyDemandedVectorElts - don't alter DemandedElts mask
Simon Pilgrim [Wed, 5 Dec 2018 10:37:45 +0000 (10:37 +0000)]
[TargetLowering] SimplifyDemandedVectorElts - don't alter DemandedElts mask

Fix potential issue with the ISD::INSERT_VECTOR_ELT case tweaking the DemandedElts mask instead of using a local copy - so later uses of the mask use the tweaked version.....

Noticed while investigating adding zero/undef folding to SimplifyDemandedVectorElts and the altered DemandedElts mask was causing mismatches.

llvm-svn: 348348

5 years ago[ARM GlobalISel] Implement call lowering for Thumb2
Diana Picus [Wed, 5 Dec 2018 10:35:28 +0000 (10:35 +0000)]
[ARM GlobalISel] Implement call lowering for Thumb2

The only things that are different from arm are:
* different opcodes for calls and returns
* Thumb calls take predicate operands

llvm-svn: 348347

5 years agoRevert r348335 "[XRay] Move-only Allocator, FunctionCallTrie, and Array"
Hans Wennborg [Wed, 5 Dec 2018 10:19:55 +0000 (10:19 +0000)]
Revert r348335 "[XRay] Move-only Allocator, FunctionCallTrie, and Array"

.. and also the follow-ups r348336 r348338.

It broke stand-alone compiler-rt builds with GCC 4.8:

In file included from /work/llvm/projects/compiler-rt/lib/xray/xray_function_call_trie.h:20:0,
                 from /work/llvm/projects/compiler-rt/lib/xray/xray_profile_collector.h:21,
                 from /work/llvm/projects/compiler-rt/lib/xray/xray_profile_collector.cc:15:
/work/llvm/projects/compiler-rt/lib/xray/xray_segmented_array.h: In instantiation of ‘T* __xray::Array<T>::AppendEmplace(Args&& ...) [with Args = {const __xray::FunctionCallTrie::mergeInto(__xray::FunctionCallTrie&) const::NodeAndTarget&}; T = __xray::FunctionCallTrie::mergeInto(__xray::FunctionCallTrie&) const::NodeAndTarget]’:
/work/llvm/projects/compiler-rt/lib/xray/xray_segmented_array.h:383:71:   required from ‘T* __xray::Array<T>::Append(const T&) [with T = __xray::FunctionCallTrie::mergeInto(__xray::FunctionCallTrie&) const::NodeAndTarget]’
/work/llvm/projects/compiler-rt/lib/xray/xray_function_call_trie.h:517:54:   required from here
/work/llvm/projects/compiler-rt/lib/xray/xray_segmented_array.h:378:5: error: could not convert ‘{std::forward<const __xray::FunctionCallTrie::mergeInto(__xray::FunctionCallTrie&) const::NodeAndTarget&>((* & args#0))}’ from ‘<brace-enclosed initializer list>’ to ‘__xray::FunctionCallTrie::mergeInto(__xray::FunctionCallTrie&) const::NodeAndTarget’
     new (AlignedOffset) T{std::forward<Args>(args)...};
     ^
/work/llvm/projects/compiler-rt/lib/xray/xray_segmented_array.h: In instantiation of ‘T* __xray::Array<T>::AppendEmplace(Args&& ...) [with Args = {const __xray::profileCollectorService::{anonymous}::ThreadTrie&}; T = __xray::profileCollectorService::{anonymous}::ThreadTrie]’:
/work/llvm/projects/compiler-rt/lib/xray/xray_segmented_array.h:383:71:   required from ‘T* __xray::Array<T>::Append(const T&) [with T = __xray::profileCollectorService::{anonymous}::ThreadTrie]’
/work/llvm/projects/compiler-rt/lib/xray/xray_profile_collector.cc:98:34:   required from here
/work/llvm/projects/compiler-rt/lib/xray/xray_segmented_array.h:378:5: error: could not convert ‘{std::forward<const __xray::profileCollectorService::{anonymous}::ThreadTrie&>((* & args#0))}’ from
‘<brace-enclosed initializer list>’ to ‘__xray::profileCollectorService::{anonymous}::ThreadTrie’
/work/llvm/projects/compiler-rt/lib/xray/xray_segmented_array.h: In instantiation of ‘T* __xray::Array<T>::AppendEmplace(Args&& ...) [with Args = {const __xray::profileCollectorService::{anonymous}::ProfileBuffer&}; T = __xray::profileCollectorService::{anonymous}::ProfileBuffer]’:
/work/llvm/projects/compiler-rt/lib/xray/xray_segmented_array.h:383:71:   required from ‘T* __xray::Array<T>::Append(const T&) [with T = __xray::profileCollectorService::{anonymous}::ProfileBuffer]

/work/llvm/projects/compiler-rt/lib/xray/xray_profile_collector.cc:244:44:   required from here
/work/llvm/projects/compiler-rt/lib/xray/xray_segmented_array.h:378:5: error: could not convert ‘{std::forward<const __xray::profileCollectorService::{anonymous}::ProfileBuffer&>((* & args#0))}’ from ‘<brace-enclosed initializer list>’ to ‘__xray::profileCollectorService::{anonymous}::ProfileBuffer’

> Summary:
> This change makes the allocator and function call trie implementations
> move-aware and remove the FunctionCallTrie's reliance on a
> heap-allocated set of allocators.
>
> The change makes it possible to always have storage associated with
> Allocator instances, not necessarily having heap-allocated memory
> obtainable from these allocator instances. We also use thread-local
> uninitialised storage.
>
> We've also re-worked the segmented array implementation to have more
> precondition and post-condition checks when built in debug mode. This
> enables us to better implement some of the operations with surrounding
> documentation as well. The `trim` algorithm now has more documentation
> on the implementation, reducing the requirement to handle special
> conditions, and being more rigorous on the computations involved.
>
> In this change we also introduce an initialisation guard, through which
> we prevent an initialisation operation from racing with a cleanup
> operation.
>
> We also ensure that the ThreadTries array is not destroyed while copies
> into the elements are still being performed by other threads submitting
> profiles.
>
> Note that this change still has an issue with accessing thread-local
> storage from signal handlers that are instrumented with XRay. We also
> learn that with the testing of this patch, that there will be cases
> where calls to mmap(...) (through internal_mmap(...)) might be called in
> signal handlers, but are not async-signal-safe. Subsequent patches will
> address this, by re-using the `BufferQueue` type used in the FDR mode
> implementation for pre-allocated memory segments per active, tracing
> thread.
>
> We still want to land this change despite the known issues, with fixes
> forthcoming.
>
> Reviewers: mboerger, jfb
>
> Subscribers: jfb, llvm-commits
>
> Differential Revision: https://reviews.llvm.org/D54989

llvm-svn: 348346

5 years ago[LICM] *Actually* disable ControlFlowHoisting.
Alina Sbirlea [Wed, 5 Dec 2018 10:16:21 +0000 (10:16 +0000)]
[LICM] *Actually* disable ControlFlowHoisting.

Summary:
The remaining code paths that ControlFlowHoisting introduced that were
not disabled, increased compile time by 3x for some benchmarks.
The time is spent in DominatorTree updates.

Reviewers: john.brawn, mkazantsev

Subscribers: sanjoy, jlebar, llvm-commits

Differential Revision: https://reviews.llvm.org/D55313

llvm-svn: 348345

5 years agoRevert "[clang-tidy] new check: bugprone-branch-clone"
Jonas Toth [Wed, 5 Dec 2018 09:34:18 +0000 (09:34 +0000)]
Revert "[clang-tidy] new check: bugprone-branch-clone"

The patch broke on buildbot with assertion-failure. Revert until this
is figured out.

llvm-svn: 348344

5 years ago[clang-tidy] new check: bugprone-branch-clone
Jonas Toth [Wed, 5 Dec 2018 09:16:25 +0000 (09:16 +0000)]
[clang-tidy] new check: bugprone-branch-clone

Summary:
Implement a check for detecting if/else if/else chains where two or more
branches are Type I clones of each other (that is, they contain identical code)
and for detecting switch statements where two or more consecutive branches are
Type I clones of each other.

Patch by donat.nagy.

Reviewers: alexfh, hokein, aaron.ballman, JonasToth

Reviewed By: JonasToth

Subscribers: MTC, lebedev.ri, whisperity, xazax.hun, Eugene.Zelenko, mgorny, rnkovacs, dkrupp, Szelethus, gamesh411, cfe-commits

Tags: #clang-tools-extra

Differential Revision: https://reviews.llvm.org/D54757

llvm-svn: 348343

5 years agoHowToBuildWithPGO.rst: Fix a few details in the manual steps
Hans Wennborg [Wed, 5 Dec 2018 08:35:30 +0000 (08:35 +0000)]
HowToBuildWithPGO.rst: Fix a few details in the manual steps

Differential revision: https://reviews.llvm.org/D55268

llvm-svn: 348342

5 years agoFix a false positive in misplaced-widening-cast
Jonas Toth [Wed, 5 Dec 2018 08:29:56 +0000 (08:29 +0000)]
Fix a false positive in misplaced-widening-cast

Summary:
bugprone-misplaced-widening-cast check
used to give a false warning to the
following example.

 enum DaysEnum{
    MON = 0,
    TUE = 1
    };

 day = (DaysEnum)(day + 1);
 //warning: either cast from 'int' to 'DaysEnum' is ineffective...

But i think int to enum cast is not widening neither ineffective.

Patch by dkrupp.

Reviewers: JonasToth, alexfh

Reviewed By: alexfh

Subscribers: rnkovacs, Szelethus, gamesh411, cfe-commits

Tags: #clang-tools-extra

Differential Revision: https://reviews.llvm.org/D55255

llvm-svn: 348341

5 years ago[X86] Remove -costmodel-reduxcost=true from the experimental vector reduction intrins...
Craig Topper [Wed, 5 Dec 2018 07:56:50 +0000 (07:56 +0000)]
[X86] Remove -costmodel-reduxcost=true from the experimental vector reduction intrinsic tests as it appears to be unnecessary. NFC

I think this has something to do with matching reductions from extractelement, binops, and shuffles. But we're not matching here.

llvm-svn: 348340

5 years ago[X86] Add more cost model tests for vector reductions with narrow vector types. NFC
Craig Topper [Wed, 5 Dec 2018 07:26:57 +0000 (07:26 +0000)]
[X86] Add more cost model tests for vector reductions with narrow vector types. NFC

llvm-svn: 348339

5 years ago[XRay] Use uptr instead of uintptr_t
Dean Michael Berris [Wed, 5 Dec 2018 07:14:06 +0000 (07:14 +0000)]
[XRay] Use uptr instead of uintptr_t

Follow-up to D54989.

llvm-svn: 348338

5 years agoAArch64: support funclets in fastcall and swift_call
Saleem Abdulrasool [Wed, 5 Dec 2018 07:09:20 +0000 (07:09 +0000)]
AArch64: support funclets in fastcall and swift_call

Functions annotated with `__fastcall` or `__attribute__((__fastcall__))`
or `__attribute__((__swiftcall__))` may contain SEH handlers even on
Win64.  This matches the behaviour of cl which allows for
`__try`/`__except` inside a `__fastcall` function.  This was detected
while trying to self-host clang on Windows ARM64.

llvm-svn: 348337

5 years ago[XRay] Use deallocateBuffer instead of deallocate
Dean Michael Berris [Wed, 5 Dec 2018 07:05:44 +0000 (07:05 +0000)]
[XRay] Use deallocateBuffer instead of deallocate

Follow-up to D54989.

llvm-svn: 348336

5 years ago[XRay] Move-only Allocator, FunctionCallTrie, and Array
Dean Michael Berris [Wed, 5 Dec 2018 06:44:34 +0000 (06:44 +0000)]
[XRay] Move-only Allocator, FunctionCallTrie, and Array

Summary:
This change makes the allocator and function call trie implementations
move-aware and remove the FunctionCallTrie's reliance on a
heap-allocated set of allocators.

The change makes it possible to always have storage associated with
Allocator instances, not necessarily having heap-allocated memory
obtainable from these allocator instances. We also use thread-local
uninitialised storage.

We've also re-worked the segmented array implementation to have more
precondition and post-condition checks when built in debug mode. This
enables us to better implement some of the operations with surrounding
documentation as well. The `trim` algorithm now has more documentation
on the implementation, reducing the requirement to handle special
conditions, and being more rigorous on the computations involved.

In this change we also introduce an initialisation guard, through which
we prevent an initialisation operation from racing with a cleanup
operation.

We also ensure that the ThreadTries array is not destroyed while copies
into the elements are still being performed by other threads submitting
profiles.

Note that this change still has an issue with accessing thread-local
storage from signal handlers that are instrumented with XRay. We also
learn that with the testing of this patch, that there will be cases
where calls to mmap(...) (through internal_mmap(...)) might be called in
signal handlers, but are not async-signal-safe. Subsequent patches will
address this, by re-using the `BufferQueue` type used in the FDR mode
implementation for pre-allocated memory segments per active, tracing
thread.

We still want to land this change despite the known issues, with fixes
forthcoming.

Reviewers: mboerger, jfb

Subscribers: jfb, llvm-commits

Differential Revision: https://reviews.llvm.org/D54989

llvm-svn: 348335

5 years ago[X86] Add narrow vector test cases to vector-reduce* tests. Add copies of the tests...
Craig Topper [Wed, 5 Dec 2018 06:29:44 +0000 (06:29 +0000)]
[X86] Add narrow vector test cases to vector-reduce* tests. Add copies of the tests with -x86-experimental-vector-widening-legalization

llvm-svn: 348334

5 years ago[NFC] Verify memoryssa in test for PR39783
Max Kazantsev [Wed, 5 Dec 2018 05:20:08 +0000 (05:20 +0000)]
[NFC] Verify memoryssa in test for PR39783

llvm-svn: 348333

5 years agogdb-remote: use elaborated type specifier for `Module`
Saleem Abdulrasool [Wed, 5 Dec 2018 04:04:14 +0000 (04:04 +0000)]
gdb-remote: use elaborated type specifier for `Module`

When building with MSVC, the type `Module` is ambiguous due to both the
lldb_private and llvm namespaces being used.  Use the elaborated type
instead to resolve the ambiguity.

llvm-svn: 348332

5 years ago[clang-tidy/checks] Update objc-property-declaration check to allow arbitrary acronym...
Stephane Moore [Wed, 5 Dec 2018 03:44:03 +0000 (03:44 +0000)]
[clang-tidy/checks] Update objc-property-declaration check to allow arbitrary acronyms and initialisms 🔧

Summary:
§1 Description

This changes the objc-property-declaration check to allow arbitrary acronyms and initialisms instead of using whitelisted acronyms. In Objective-C it is relatively common to use project prefixes in property names for the purposes of disambiguation. For example, the CIColor¹ and CGColor² properties on UIColor both represent symbol prefixes being used in proeprty names outside of Apple's accepted acronyms³. The union of Apple's accepted acronyms and all symbol prefixes that might be used for disambiguation in property declarations effectively allows for any arbitrary sequence of capital alphanumeric characters to be acceptable in property declarations. This change updates the check accordingly.

The test variants with custom configurations are deleted as part of this change because their configurations no longer impact behavior. The acronym configurations are currently preserved for backwards compatibility of check configuration.

[1] https://developer.apple.com/documentation/uikit/uicolor/1621951-cicolor?language=objc
[2] https://developer.apple.com/documentation/uikit/uicolor/1621954-cgcolor?language=objc
[3] https://developer.apple.com/library/archive/documentation/Cocoa/Conceptual/CodingGuidelines/Articles/APIAbbreviations.html#//apple_ref/doc/uid/20001285-BCIHCGAE

§2 Test Notes

Changes verified by:
• Running clang-tidy unit tests.
• Used check_clang_tidy.py to verify expected output of processing objc-property-declaration.m

Reviewers: benhamilton, Wizard

Reviewed By: benhamilton

Subscribers: jfb, cfe-commits

Differential Revision: https://reviews.llvm.org/D51832

llvm-svn: 348331