platform/upstream/llvm.git
5 years ago[clangd] Clear the semantic of RefSlab::size.
Haojian Wu [Thu, 18 Oct 2018 15:33:20 +0000 (15:33 +0000)]
[clangd] Clear the semantic of RefSlab::size.

Summary:
The RefSlab::size can easily cause confusions, it returns the number of
different symbols, rahter than the number of all references.

- add numRefs() method and cache it, since calculating it everytime is nontrivial.
- clear misused places.

Reviewers: sammccall

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

Differential Revision: https://reviews.llvm.org/D53389

llvm-svn: 344745

5 years ago[lldbsuite] Mark the TestScriptedResolver tests as XFAIL on Windows
Stella Stamenova [Thu, 18 Oct 2018 15:30:31 +0000 (15:30 +0000)]
[lldbsuite] Mark the TestScriptedResolver tests as XFAIL on Windows

Summary: They fail similarly to some of the other breakpoint tests on Windows, so I suspect the cause is the same. I've linked to the same bug.

Reviewers: asmith, zturner, jingham

Subscribers: abidh, lldb-commits

Differential Revision: https://reviews.llvm.org/D53331

llvm-svn: 344744

5 years ago[LV] Fold tail by masking to vectorize loops of arbitrary trip count under opt for...
Ayal Zaks [Thu, 18 Oct 2018 15:03:15 +0000 (15:03 +0000)]
[LV] Fold tail by masking to vectorize loops of arbitrary trip count under opt for size

When optimizing for size, a loop is vectorized only if the resulting vector loop
completely replaces the original scalar loop. This holds if no runtime guards
are needed, if the original trip-count TC does not overflow, and if TC is a
known constant that is a multiple of the VF. The last two TC-related conditions
can be overcome by
1. rounding the trip-count of the vector loop up from TC to a multiple of VF;
2. masking the vector body under a newly introduced "if (i <= TC-1)" condition.

The patch allows loops with arbitrary trip counts to be vectorized under -Os,
subject to the existing cost model considerations. It also applies to loops with
small trip counts (under -O2) which are currently handled as if under -Os.

The patch does not handle loops with reductions, live-outs, or w/o a primary
induction variable, and disallows interleave groups.

(Third, final and main part of -)
Differential Revision: https://reviews.llvm.org/D50480

llvm-svn: 344743

5 years ago[X86][Tests] Make sure tls-direct-seg-refs tests only run where supported
Kristina Brooks [Thu, 18 Oct 2018 14:44:25 +0000 (14:44 +0000)]
[X86][Tests] Make sure tls-direct-seg-refs tests only run where supported

This flag is only supported for x86 targets, make sure the tests only run
for those.

llvm-svn: 344742

5 years ago[clangd] Enforce rules around "initialize" request, and create ClangdServer lazily.
Sam McCall [Thu, 18 Oct 2018 14:41:50 +0000 (14:41 +0000)]
[clangd] Enforce rules around "initialize" request, and create ClangdServer lazily.

Summary:
LSP is a slightly awkward map to C++ object lifetimes: the initialize request
is part of the protocol and provides information that doesn't change over the
lifetime of the server.

Until now, we handled this by initializing ClangdServer and ClangdLSPServer
right away, and making anything that can be set in the "initialize" request
mutable.
With this patch, we create ClangdLSPServer immediately, but defer creating
ClangdServer until "initialize". This opens the door to passing the relevant
initialize params in the constructor and storing them immutably.
(That change isn't actually done in this patch).

To make this safe, we have the MessageDispatcher enforce that the "initialize"
method is called before any other (as required by LSP). That way each method
handler can assume Server is initialized, as today.

As usual, while implementing this I found places where our test cases violated
the protocol.

Reviewers: ioeric

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

Differential Revision: https://reviews.llvm.org/D53398

llvm-svn: 344741

5 years ago[OPENMP] Move OMPClausePrinter to OpenMPClause.h/OpenMPClause.cpp - NFC. Differential...
Patrick Lyster [Thu, 18 Oct 2018 14:28:23 +0000 (14:28 +0000)]
[OPENMP] Move OMPClausePrinter to OpenMPClause.h/OpenMPClause.cpp - NFC. Differential Revision: https://reviews.llvm.org/D53102

llvm-svn: 344740

5 years agoAdd support for -mno-tls-direct-seg-refs to Clang
Kristina Brooks [Thu, 18 Oct 2018 14:07:02 +0000 (14:07 +0000)]
Add support for -mno-tls-direct-seg-refs to Clang

This patch exposes functionality added in rL344723 to the Clang driver/frontend
as a flag and adds appropriate metadata.

Driver tests pass:
```
ninja check-clang-driver
-snip-
  Expected Passes    : 472
  Expected Failures  : 3
  Unsupported Tests  : 65
```

Odd failure in CodeGen tests but unrelated to this:
```
ninja check-clang-codegen
-snip-
/SourceCache/llvm-trunk-8.0/tools/clang/test/CodeGen/builtins-wasm.c:87:10:
error: cannot compile this builtin function yet
-snip-
Failing Tests (1):
    Clang :: CodeGen/builtins-wasm.c

  Expected Passes    : 1250
  Expected Failures  : 2
  Unsupported Tests  : 120
  Unexpected Failures: 1
```

Original commit:
[X86] Support for the mno-tls-direct-seg-refs flag
Allows to disable direct TLS segment access (%fs or %gs). GCC supports a
similar flag, it can be useful in some circumstances, e.g. when a thread
context block needs to be updated directly from user space. More info and
specific use cases: https://bugs.llvm.org/show_bug.cgi?id=16145

Patch by nruslan (Ruslan Nikolaev).

Differential Revision: https://reviews.llvm.org/D53102

llvm-svn: 344739

5 years agoDivergenceAnalysisTest: fix use of uninitialized memory
Nicolai Haehnle [Thu, 18 Oct 2018 12:54:39 +0000 (12:54 +0000)]
DivergenceAnalysisTest: fix use of uninitialized memory

Thanks to Simon Moll for chasing it down.

Change-Id: If188f07c4aaec217f40a7a2ca029818f9202f1cb
llvm-svn: 344738

5 years ago[clangd] Lay JSONRPCDispatcher to rest.
Sam McCall [Thu, 18 Oct 2018 12:32:04 +0000 (12:32 +0000)]
[clangd] Lay JSONRPCDispatcher to rest.

Summary:
Most of its functionality is moved into ClangdLSPServer.
The decoupling between JSONRPCDispatcher, ProtocolCallbacks, ClangdLSPServer
was never real, and only served to obfuscate.

Some previous implicit/magic stuff is now explicit:
 - the return type of LSP method calls are now in the signature
 - no more reply() that gets the ID using global context magic
 - arg tracing no longer relies on RequestArgs::stash context magic either

This is mostly refactoring, but some deliberate fixes while here:
 - LSP method params are now by const reference
 - notifications and calls are now distinct namespaces.
   (some tests had protocol errors and needed updating)
 - we now reply to calls we failed to decode
 - outgoing calls use distinct IDs
A few error codes and message IDs changed in unimportant ways (see tests).

Reviewers: ioeric

Subscribers: mgorny, ilya-biryukov, javed.absar, MaskRay, jkorous, arphaman, jfb, kadircet, cfe-commits

Differential Revision: https://reviews.llvm.org/D53387

llvm-svn: 344737

5 years ago[clangd] Names that are not spelled in source code are reserved.
Eric Liu [Thu, 18 Oct 2018 12:23:05 +0000 (12:23 +0000)]
[clangd] Names that are not spelled in source code are reserved.

Summary:
These are often not expected to be used directly e.g.
```
TEST_F(Fixture, X) {
  ^  // "Fixture_X_Test" expanded in the macro should be down ranked.
}
```

Only doing this for sema for now, as such symbols are mostly coming from sema
e.g. gtest macros expanded in the main file. We could also add a similar field
for the index symbol.

Reviewers: sammccall

Reviewed By: sammccall

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

Differential Revision: https://reviews.llvm.org/D53374

llvm-svn: 344736

5 years ago[clangd] Encode Line/Column as a 32-bits integer.
Haojian Wu [Thu, 18 Oct 2018 10:43:50 +0000 (10:43 +0000)]
[clangd] Encode Line/Column as a 32-bits integer.

Summary:
This would buy us more memory. Using a 32-bits integer is enough for
most human-readable source code (up to 4M lines and 4K columns).

Previsouly, we used 8 bytes for a position, now 4 bytes, it would save
us 8 bytes for each Ref and each Symbol instance.

For LLVM-project binary index file, we save ~13% memory.

| Before | After |
| 412MB  | 355MB |

Reviewers: sammccall

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

Differential Revision: https://reviews.llvm.org/D53363

llvm-svn: 344735

5 years ago[DA] DivergenceAnalysis for unstructured, reducible CFGs
Nicolai Haehnle [Thu, 18 Oct 2018 09:38:44 +0000 (09:38 +0000)]
[DA] DivergenceAnalysis for unstructured, reducible CFGs

Summary:
This is patch 2 of the new DivergenceAnalysis (https://reviews.llvm.org/D50433).

This patch contains a generic divergence analysis implementation for
unstructured, reducible Control-Flow Graphs. It contains two new classes.
The `SyncDependenceAnalysis` class lazily computes sync dependences, which
relate divergent branches to points of joining divergent control. The
`DivergenceAnalysis` class contains the generic divergence analysis
implementation.

Reviewers: nhaehnle

Reviewed By: nhaehnle

Subscribers: sameerds, kristina, nhaehnle, xbolva00, tschuett, mgorny, llvm-commits

Differential Revision: https://reviews.llvm.org/D51491

llvm-svn: 344734

5 years ago[clang-tidy] Ignore a case where the fix of make_unique check introduces side effect.
Haojian Wu [Thu, 18 Oct 2018 09:13:34 +0000 (09:13 +0000)]
[clang-tidy] Ignore a case where the fix of make_unique check introduces side effect.

Summary:
Previously, ptr.reset(new char[5]) will be replaced with `p =
make_unique<char[]>(5)`, the fix has side effect -- doing
default initialization, it may cause performace regression (we are
bitten by this rececntly)

The check should be conservative for these cases.

Reviewers: alexfh

Subscribers: xazax.hun, cfe-commits

Differential Revision: https://reviews.llvm.org/D53377

llvm-svn: 344733

5 years ago[Support] json::Value construction from std::vector<T> and std::map<string,T>.
Sam McCall [Thu, 18 Oct 2018 08:47:24 +0000 (08:47 +0000)]
[Support] json::Value construction from std::vector<T> and std::map<string,T>.

Summary: Previously this required a conversion to json::Array/json::Object first.

Reviewers: ioeric

Subscribers: kristina, llvm-commits

Differential Revision: https://reviews.llvm.org/D53385

llvm-svn: 344732

5 years ago[llvm-exegesis] Fix off by one error
Guillaume Chatelet [Thu, 18 Oct 2018 08:20:50 +0000 (08:20 +0000)]
[llvm-exegesis] Fix off by one error

llvm-svn: 344731

5 years ago[TI removal] Test predicate rather than casting to detect a terminator
Chandler Carruth [Thu, 18 Oct 2018 08:16:20 +0000 (08:16 +0000)]
[TI removal] Test predicate rather than casting to detect a terminator
and use the range based successor API.

llvm-svn: 344730

5 years ago[Windows] Fix threads comparison on Windows
Aleksandr Urakov [Thu, 18 Oct 2018 07:52:56 +0000 (07:52 +0000)]
[Windows] Fix threads comparison on Windows

Summary:
This patch makes Windows threads to compare by a thread ID, not by a handle.
It's because the same thread can have different handles on Windows
(for example, `GetCurrentThread` always returns the fake handle `-2`).
This leads to some incorrect behavior. For example, in `Process::GetRunLock`
always `m_public_run_lock` is returned without this patch.

Reviewers: zturner, clayborg, stella.stamenova

Reviewed By: stella.stamenova

Subscribers: stella.stamenova, lldb-commits

Tags: #lldb

Differential Revision: https://reviews.llvm.org/D53357

llvm-svn: 344729

5 years ago[TI removal] Generically discuss terminators rather than use the soon to
Chandler Carruth [Thu, 18 Oct 2018 07:43:30 +0000 (07:43 +0000)]
[TI removal] Generically discuss terminators rather than use the soon to
vanish subclass name.

llvm-svn: 344728

5 years ago[TI removal] Remove discussion of `TerminatorInst` from the LLVM
Chandler Carruth [Thu, 18 Oct 2018 07:40:24 +0000 (07:40 +0000)]
[TI removal] Remove discussion of `TerminatorInst` from the LLVM
documentation.

llvm-svn: 344727

5 years ago[TI removal] Remove TerminatorInst references from bindings.
Chandler Carruth [Thu, 18 Oct 2018 07:40:03 +0000 (07:40 +0000)]
[TI removal] Remove TerminatorInst references from bindings.

For the Go bindings, this just removes the no longer useful "isa"-style
wrapper. If there is a user that is interested, they can add a wrapper
for `Instruction::isTerminator`.

For the OCaml bindings, this is just a documentation update.

llvm-svn: 344726

5 years agoAdd a emitUnaryFloatFnCall version that fetches the function name from TLI
Mikael Holmen [Thu, 18 Oct 2018 06:27:53 +0000 (06:27 +0000)]
Add a emitUnaryFloatFnCall version that fetches the function name from TLI

Summary:
In several places in the code we use the following pattern:

  if (hasUnaryFloatFn(&TLI, Ty, LibFunc_tan, LibFunc_tanf, LibFunc_tanl)) {
    [...]
    Value *Res = emitUnaryFloatFnCall(X, TLI.getName(LibFunc_tan), B, Attrs);
    [...]
  }

In short, we check if there is a lib-function for a certain type, and then
we _always_ fetch the name of the "double" version of the lib function and
construct a call to the appropriate function, that we just checked exists,
using that "double" name as a basis.

This is of course a problem in cases where the target doesn't support the
"double" version, but e.g. only the "float" version.

In that case TLI.getName(LibFunc_tan) returns "", and
emitUnaryFloatFnCall happily appends an "f" to "", and we erroneously end
up with a call to a function called "f".

To solve this, the above pattern is changed to

  if (hasUnaryFloatFn(&TLI, Ty, LibFunc_tan, LibFunc_tanf, LibFunc_tanl)) {
    [...]
    Value *Res = emitUnaryFloatFnCall(X, &TLI, LibFunc_tan, LibFunc_tanf,
                                      LibFunc_tanl, B, Attrs);
    [...]
  }

I.e instead of first fetching the name of the "double" version and then
letting emitUnaryFloatFnCall() add the final "f" or "l", we let
emitUnaryFloatFnCall() fetch the right name from TLI.

Reviewers: eli.friedman, efriedma

Reviewed By: efriedma

Subscribers: efriedma, bjope, llvm-commits

Differential Revision: https://reviews.llvm.org/D53370

llvm-svn: 344725

5 years agoFix warning about unused variable [NFC]
Mikael Holmen [Thu, 18 Oct 2018 06:00:39 +0000 (06:00 +0000)]
Fix warning about unused variable [NFC]

llvm-svn: 344724

5 years ago[X86] Support for the mno-tls-direct-seg-refs flag
Kristina Brooks [Thu, 18 Oct 2018 03:14:37 +0000 (03:14 +0000)]
[X86] Support for the mno-tls-direct-seg-refs flag

Allows to disable direct TLS segment access (%fs or %gs). GCC supports
a similar flag, it can be useful in some circumstances, e.g. when a thread
context block needs to be updated directly from user space. More info
and specific use cases: https://bugs.llvm.org/show_bug.cgi?id=16145

There is another revision for clang as well.
Related: D53102

All X86 CodeGen tests appear to pass:
```
[46/47] Running lit suite /SourceCache/llvm-trunk-8.0/test/CodeGen
Testing Time: 23.17s
  Expected Passes    : 3801
  Expected Failures  : 15
  Unsupported Tests  : 8021
```

Reviewed by: Craig Topper.

Patch by nruslan (Ruslan Nikolaev).

Differential Revision: https://reviews.llvm.org/D53103

llvm-svn: 344723

5 years agoRevert "Return a named error in the result object of an expression with no result"
Krasimir Georgiev [Thu, 18 Oct 2018 03:10:43 +0000 (03:10 +0000)]
Revert "Return a named error in the result object of an expression with no result"

This reverts commit r344647.
This causes build failures with [-Werror, -Wswitch]. Some cases where the newly
introduced enum value is not handled in particular are in:
  lldb/source/Expression/REPL.cpp:350
  lldb/source/Interpreter/CommandInterpreter.cpp:1529
(maybe there could be more)

As I don't understand lldb to make sure the likely trivial fixes are
correct and also as they might need additional tests, leaving to the
author to resolve.

llvm-svn: 344722

5 years ago[llvm-exegesis] Mark destructor virtual after r344695
Krasimir Georgiev [Thu, 18 Oct 2018 02:06:16 +0000 (02:06 +0000)]
[llvm-exegesis] Mark destructor virtual after r344695

This was causing a -Wnon-virtual-dtor warning.

llvm-svn: 344721

5 years ago[BuildingAJIT] Update the Ch1 KaleidoscopeJIT class to expose errors to clients.
Lang Hames [Thu, 18 Oct 2018 00:51:38 +0000 (00:51 +0000)]
[BuildingAJIT] Update the Ch1 KaleidoscopeJIT class to expose errors to clients.

Returning the error to clients provides an opportunity to introduce readers to
the Expected and Error APIs and makes the tutorial more useful as a starting
point for a real JIT class, while only slightly complicating the code.

llvm-svn: 344720

5 years ago[TI removal] Switch simple loop unswitch to `Instruction`.
Chandler Carruth [Thu, 18 Oct 2018 00:40:26 +0000 (00:40 +0000)]
[TI removal] Switch simple loop unswitch to `Instruction`.

llvm-svn: 344719

5 years ago[TI removal] Switch NewGVN to directly use `Instruction`.
Chandler Carruth [Thu, 18 Oct 2018 00:39:46 +0000 (00:39 +0000)]
[TI removal] Switch NewGVN to directly use `Instruction`.

llvm-svn: 344718

5 years ago[TI removal] Use `Instruction` instead of `TerminatorInst` for
Chandler Carruth [Thu, 18 Oct 2018 00:39:18 +0000 (00:39 +0000)]
[TI removal] Use `Instruction` instead of `TerminatorInst` for
a variable's type.

llvm-svn: 344717

5 years ago[TI removal] Update CodeExtractor to use Instruction directly.
Chandler Carruth [Thu, 18 Oct 2018 00:38:54 +0000 (00:38 +0000)]
[TI removal] Update CodeExtractor to use Instruction directly.

llvm-svn: 344716

5 years ago[TI removal] Switch ObjCARC code to directly use the nice range-based
Chandler Carruth [Thu, 18 Oct 2018 00:38:34 +0000 (00:38 +0000)]
[TI removal] Switch ObjCARC code to directly use the nice range-based
successors API or directly build the iterators out of the terminator
instruction and avoid requiring a TerminatorInst variable.

llvm-svn: 344715

5 years ago[TI removal] Switch MergeFunctions to directly use Instruction API.
Chandler Carruth [Thu, 18 Oct 2018 00:37:37 +0000 (00:37 +0000)]
[TI removal] Switch MergeFunctions to directly use Instruction API.

llvm-svn: 344714

5 years ago[TI removal] Switch an analysis to just use Instruction.
Chandler Carruth [Thu, 18 Oct 2018 00:36:15 +0000 (00:36 +0000)]
[TI removal] Switch an analysis to just use Instruction.

llvm-svn: 344713

5 years ago[BuildingAJIT] Simplify a tutorial example and fix a syntax error.
Lang Hames [Wed, 17 Oct 2018 22:27:09 +0000 (22:27 +0000)]
[BuildingAJIT] Simplify a tutorial example and fix a syntax error.

llvm-svn: 344712

5 years agoAMDGPU: Add options to enable/disable code object v3
Konstantin Zhuravlyov [Wed, 17 Oct 2018 21:39:12 +0000 (21:39 +0000)]
AMDGPU: Add options to enable/disable code object v3

Differential Revision: https://reviews.llvm.org/D53386

llvm-svn: 344711

5 years ago[AArch64] Define __ELF__ for aarch64-none-elf and other similar triples.
Eli Friedman [Wed, 17 Oct 2018 21:07:11 +0000 (21:07 +0000)]
[AArch64] Define __ELF__ for aarch64-none-elf and other similar triples.

"aarch64-none-elf" is commonly used for AArch64 baremetal toolchains.

Differential Revision: https://reviews.llvm.org/D53348

llvm-svn: 344710

5 years ago[clang-doc] Bringing bitcode tests in line
Julie Hockett [Wed, 17 Oct 2018 20:16:05 +0000 (20:16 +0000)]
[clang-doc] Bringing bitcode tests in line

Makes bitcode tests line up with what's actually called in the tool.
Should fix the failing bot.

Also fixes a warning that was being thrown about initialization braces.

Differential Revision: https://reviews.llvm.org/D53381

llvm-svn: 344707

5 years ago[BuildingAJIT] Fix a function signature in the documentation.
Lang Hames [Wed, 17 Oct 2018 19:35:38 +0000 (19:35 +0000)]
[BuildingAJIT] Fix a function signature in the documentation.

llvm-svn: 344705

5 years agoPort libcxxabi r344607 into llvm
Pavel Labath [Wed, 17 Oct 2018 18:50:25 +0000 (18:50 +0000)]
Port libcxxabi r344607 into llvm

Summary:
The original commit message was:
    This uses CRTP (for performance reasons) to allow a user the override
    demangler functions to implement custom parsing logic. The motivation
    for this is LLDB, which needs to occasionaly modify the mangled names.
    One such instance is already implemented via the TypeCallback member,
    but this is very specific functionality which does not help with any
    other use case. Currently we have a use case for modifying the
    constructor flavours, which would require adding another callback. This
    approach does not scale.

    With CRTP, the user (LLDB) can override any function it needs without
    any special support from the demangler library. After LLDB is ported to
    use this instead of the TypeCallback mechanism, the callback can be
    removed.

The only difference here is the addition of a unit test which exercises
the CRTP mechanism to override a function in the parser.

Reviewers: erik.pilkington, rsmith, EricWF

Subscribers: mgorny, kristina, llvm-commits

Differential Revision: https://reviews.llvm.org/D53300

llvm-svn: 344703

5 years agoFix for arm bots afternew PM pass port. Prevent cross compiling on arm.
Leonard Chan [Wed, 17 Oct 2018 18:12:18 +0000 (18:12 +0000)]
Fix for arm bots afternew PM pass port. Prevent cross compiling on arm.

llvm-svn: 344702

5 years agoFix for failing unit tests on some bots after r344696.
Leonard Chan [Wed, 17 Oct 2018 16:21:19 +0000 (16:21 +0000)]
Fix for failing unit tests on some bots after r344696.

llvm-svn: 344701

5 years ago[libcxx] Improve reporting when running the lit test suite
Louis Dionne [Wed, 17 Oct 2018 16:12:04 +0000 (16:12 +0000)]
[libcxx] Improve reporting when running the lit test suite

Summary:
Running the test suite with -a will now properly show all the executed
commands. The reports also include the environment under which the test
is being executed, which is helpful for reproducing issues.

Reviewers: EricWF

Subscribers: christof, dexonsmith, libcxx-commits

Differential Revision: https://reviews.llvm.org/D53215

llvm-svn: 344700

5 years ago[PassManager/Sanitizer] Enable usage of ported AddressSanitizer passes with -fsanitiz...
Leonard Chan [Wed, 17 Oct 2018 15:38:22 +0000 (15:38 +0000)]
[PassManager/Sanitizer] Enable usage of ported AddressSanitizer passes with -fsanitize=address

Enable usage of `AddressSanitizer` and `AddressModuleSanitizer` ported from the
legacy to the new PassManager.

This patch depends on https://reviews.llvm.org/D52739.

Differential Revision: https://reviews.llvm.org/D52814

llvm-svn: 344699

5 years agoAMDGPU: Avoid selecting ds_{read,write}2_b32 on SI
Nicolai Haehnle [Wed, 17 Oct 2018 15:37:48 +0000 (15:37 +0000)]
AMDGPU: Avoid selecting ds_{read,write}2_b32 on SI

Summary:
To workaround a hardware issue in the (base + offset) calculation
when base is negative. The impact on code quality should be limited
since SILoadStoreOptimizer still runs afterwards and is able to
combine loads/stores based on known sign information.

This fixes visible corruption in Hitman on SI (easily reproducible
by running benchmark mode).

Change-Id: Ia178d207a5e2ac38ae7cd98b532ea2ae74704e5f
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=99923

Reviewers: arsenm, mareko

Subscribers: jholewinski, kzhuravl, jvesely, wdng, yaxunl, dstuttard, tpr, t-tye, llvm-commits

Differential Revision: https://reviews.llvm.org/D53160

llvm-svn: 344698

5 years agoStructurizeCFG: Simplify inserted PHI nodes
Nicolai Haehnle [Wed, 17 Oct 2018 15:37:41 +0000 (15:37 +0000)]
StructurizeCFG: Simplify inserted PHI nodes

Summary:
This improves subsequent divergence analysis in some cases.

Change-Id: I5e95e7ec7fd3fa80d414d1a53a02fea23e3d67d3

Reviewers: arsenm, rampitec

Subscribers: jvesely, wdng, llvm-commits

Differential Revision: https://reviews.llvm.org/D53316

llvm-svn: 344697

5 years agoAMDGPU: Divergence-driven selection of scalar buffer load intrinsics
Nicolai Haehnle [Wed, 17 Oct 2018 15:37:30 +0000 (15:37 +0000)]
AMDGPU: Divergence-driven selection of scalar buffer load intrinsics

Summary:
Moving SMRD to VMEM in SIFixSGPRCopies is rather bad for performance if
the load is really uniform. So select the scalar load intrinsics directly
to either VMEM or SMRD buffer loads based on divergence analysis.

If an offset happens to end up in a VGPR -- either because a floating
point calculation was involved, or due to other remaining deficiencies
in SIFixSGPRCopies -- we use v_readfirstlane.

There is some unrelated churn in tests since we now select MUBUF offsets
in a unified way with non-scalar buffer loads.

Change-Id: I170e6816323beb1348677b358c9d380865cd1a19

Reviewers: arsenm, alex-t, rampitec, tpr

Subscribers: kzhuravl, jvesely, wdng, yaxunl, dstuttard, t-tye, llvm-commits

Differential Revision: https://reviews.llvm.org/D53283

llvm-svn: 344696

5 years ago[llvm-exegesis] Allow measuring several instructions in a single run.
Clement Courbet [Wed, 17 Oct 2018 15:04:15 +0000 (15:04 +0000)]
[llvm-exegesis] Allow measuring several instructions in a single run.

Summary:
We try to recover gracefully on instructions that would crash the
program.

This includes some refactoring of runMeasurement() implementations.

Reviewers: gchatelet

Subscribers: tschuett, llvm-commits

Differential Revision: https://reviews.llvm.org/D53371

llvm-svn: 344695

5 years agoFix broken formatting caused by test commit
Theodoros Theodoridis [Wed, 17 Oct 2018 15:01:18 +0000 (15:01 +0000)]
Fix broken formatting caused by test commit

llvm-svn: 344694

5 years ago[ARM] bottom-top mul support in ARMParallelDSP
Sam Parker [Wed, 17 Oct 2018 13:02:48 +0000 (13:02 +0000)]
[ARM] bottom-top mul support in ARMParallelDSP

Previously reverted in rL343082.

Original commit message:

On failing to find sequences that can be converted into dual macs,
try to find sequential 16-bit loads that are used by muls which we
can then use smultb, smulbt, smultt with a wide load.

Differential Revision: https://reviews.llvm.org/D51983

llvm-svn: 344693

5 years agoFix uninitialized variable
Guillaume Chatelet [Wed, 17 Oct 2018 12:27:46 +0000 (12:27 +0000)]
Fix uninitialized variable

llvm-svn: 344692

5 years agoAMDGPU: Remove dead TableGen code
Nicolai Haehnle [Wed, 17 Oct 2018 12:14:26 +0000 (12:14 +0000)]
AMDGPU: Remove dead TableGen code

Summary: Change-Id: Ic1f2c1d0cf9e90a0baa9fc6bacd0d3c386069fb0

Reviewers: tpr

Subscribers: arsenm, kzhuravl, jvesely, wdng, yaxunl, dstuttard, t-tye, llvm-commits

Differential Revision: https://reviews.llvm.org/D53318

Change-Id: Ib4d143c898801e5cf6cb9999a495d62c91ae77fb
llvm-svn: 344691

5 years agoBuildBot fix, compiler complains about array decay to pointer
Guillaume Chatelet [Wed, 17 Oct 2018 12:09:21 +0000 (12:09 +0000)]
BuildBot fix, compiler complains about array decay to pointer

llvm-svn: 344690

5 years ago[llvm-exegeis] Computing Latency configuration upfront so we can generate many CodeTe...
Guillaume Chatelet [Wed, 17 Oct 2018 11:37:28 +0000 (11:37 +0000)]
[llvm-exegeis] Computing Latency configuration upfront so we can generate many CodeTemplates at once.

Summary: LatencyGenerator now computes all possible mode of serial execution for an Instruction upfront and generates CodeTemplate for the ones that give the best results (e.g. no need to generate a two instructions snippet when repeating a single one would do). The next step is to generate even more configurations for cases (e.g. for XOR we should generate "XOR EAX, EAX, EAX" and "XOR EAX, EAX, EBX")

Reviewers: courbet

Reviewed By: courbet

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D53320

llvm-svn: 344689

5 years ago[clangd] Support scope proximity in code completion.
Eric Liu [Wed, 17 Oct 2018 11:19:02 +0000 (11:19 +0000)]
[clangd] Support scope proximity in code completion.

Summary:
This should make all-scope completion more usable. Scope proximity for
indexes will be added in followup patch.

Reviewers: sammccall

Reviewed By: sammccall

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

Differential Revision: https://reviews.llvm.org/D53131

llvm-svn: 344688

5 years ago[NFC] Remove GOTO from SCEV
Max Kazantsev [Wed, 17 Oct 2018 11:16:25 +0000 (11:16 +0000)]
[NFC] Remove GOTO from SCEV

llvm-svn: 344687

5 years ago[NewPM] Fixing test failure on Windows - removed opt binary name from pattern
Fedor Sergeev [Wed, 17 Oct 2018 11:01:15 +0000 (11:01 +0000)]
[NewPM] Fixing test failure on Windows - removed opt binary name from pattern

llvm-svn: 344686

5 years ago[NewPM] teach -passes= to emit meaningful error messages
Fedor Sergeev [Wed, 17 Oct 2018 10:36:23 +0000 (10:36 +0000)]
[NewPM] teach -passes= to emit meaningful error messages

All the PassBuilder::parse interfaces now return descriptive StringError
instead of a plain bool. It allows to make -passes/aa-pipeline parsing
errors context-specific and thus less confusing.

TODO: ideally we should also make suggestions for misspelled pass names,
but that requires some extensions to PassBuilder.

Reviewed By: philip.pfaffe, chandlerc
Differential Revision: https://reviews.llvm.org/D53246

llvm-svn: 344685

5 years ago[MIPS GlobalISel] Legalize constants
Petar Jovanovic [Wed, 17 Oct 2018 10:30:03 +0000 (10:30 +0000)]
[MIPS GlobalISel] Legalize constants

Legalize s1, s8, s16 and s64 G_CONSTANT for MIPS32.

Patch by Petar Avramovic.

Differential Revision: https://reviews.llvm.org/D53077

llvm-svn: 344684

5 years ago[ARM] Do not fuse VADD and VMUL, continued (2/2)
Sjoerd Meijer [Wed, 17 Oct 2018 10:05:44 +0000 (10:05 +0000)]
[ARM] Do not fuse VADD and VMUL, continued (2/2)

This is patch 2/2, following up on D53314, and is the functional change
to prevent fusing mul + add sequences into VFMAs.

Differential revision: https://reviews.llvm.org/D53315

llvm-svn: 344683

5 years agoTest commit
Theodoros Theodoridis [Wed, 17 Oct 2018 09:59:41 +0000 (09:59 +0000)]
Test commit

llvm-svn: 344682

5 years ago[LoopPredication] add some simple stats
Fedor Sergeev [Wed, 17 Oct 2018 09:02:54 +0000 (09:02 +0000)]
[LoopPredication] add some simple stats

Just adding some useful statistics to LoopPredication pass
which was lacking any of these.

llvm-svn: 344681

5 years ago[clangd] Fix buildbot failure.
Haojian Wu [Wed, 17 Oct 2018 08:54:48 +0000 (08:54 +0000)]
[clangd] Fix buildbot failure.

llvm-svn: 344680

5 years ago[clangd] Print numbers of symbols and refs as well when loading the
Haojian Wu [Wed, 17 Oct 2018 08:48:04 +0000 (08:48 +0000)]
[clangd] Print numbers of symbols and refs as well when loading the
index.

llvm-svn: 344679

5 years ago[clangd] Collect refs from headers.
Haojian Wu [Wed, 17 Oct 2018 08:38:36 +0000 (08:38 +0000)]
[clangd] Collect refs from headers.

Summary:
Add a flag to SymbolCollector to collect refs fdrom headers.

Note that we collect refs from headers in static index, and we don't do it for
dynamic index because of the preamble (we skip function body in preamble,
collecting it will result incomplete results).

Reviewers: sammccall

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

Differential Revision: https://reviews.llvm.org/D53322

llvm-svn: 344678

5 years ago[ARM] Follow up of rL344671, attempt to pacify a buildbot
Sjoerd Meijer [Wed, 17 Oct 2018 07:51:24 +0000 (07:51 +0000)]
[ARM] Follow up of rL344671, attempt to pacify a buildbot

It was rightfully complaining about an unpretty logical expression.

llvm-svn: 344677

5 years ago[clangd] Hide unused function. NFC
Sam McCall [Wed, 17 Oct 2018 07:41:53 +0000 (07:41 +0000)]
[clangd] Hide unused function. NFC

llvm-svn: 344676

5 years ago[clangd] Rename and move trivial logger to Logger.cpp. NFC
Sam McCall [Wed, 17 Oct 2018 07:39:32 +0000 (07:39 +0000)]
[clangd] Rename and move trivial logger to Logger.cpp. NFC

llvm-svn: 344675

5 years ago[LLDB] - Add support for DW_RLE_start_end entries (.debug_rnglists)
George Rimar [Wed, 17 Oct 2018 07:37:26 +0000 (07:37 +0000)]
[LLDB] - Add support for DW_RLE_start_end entries (.debug_rnglists)

DWARF5 describes DW_RLE_start_end as:

This is a form of bounded range entry that has two target address operands.
Each operand is the same size as used in DW_FORM_addr. These indicate
the starting and ending addresses, respectively, that define the address range
for which the following location is valid.

The patch implements the support.

Differential revision: https://reviews.llvm.org/D53193

llvm-svn: 344674

5 years ago[clangd] Simplify client capabilities parsing.
Sam McCall [Wed, 17 Oct 2018 07:33:42 +0000 (07:33 +0000)]
[clangd] Simplify client capabilities parsing.

Summary:
Instead of parsing into structs that mirror LSP, simply parse into a flat struct
that contains the info we need.
This is an exception to our strategy with Protocol.h, which seems justified:
 - the structure here is very large and deeply nested
 - we care about almost none of it
 - we should never have to serialize client capabilities

Reviewers: kadircet

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

Differential Revision: https://reviews.llvm.org/D53266

llvm-svn: 344673

5 years ago[clangd] Refactor JSON-over-stdin/stdout code into Transport abstraction. (re-land...
Sam McCall [Wed, 17 Oct 2018 07:32:05 +0000 (07:32 +0000)]
[clangd] Refactor JSON-over-stdin/stdout code into Transport abstraction. (re-land r344620)

Summary:
This paves the way for alternative transports (mac XPC, maybe messagepack?),
and also generally improves layering: testing ClangdLSPServer becomes less of
a pipe dream, we split up the JSONOutput monolith, etc.

This isn't a final state, much of what remains in JSONRPCDispatcher can go away,
handlers can call reply() on the transport directly, JSONOutput can be renamed
to StreamLogger and removed, etc. But this patch is sprawling already.

The main observable change (see tests) is that hitting EOF on input is now an
error: the client should send the 'exit' notification.
This is defensible: the protocol doesn't spell this case out. Reproducing the
current behavior for all combinations of shutdown/exit/EOF clutters interfaces.
We can iterate on this if desired.

Reviewers: jkorous, ioeric, hokein

Subscribers: mgorny, ilya-biryukov, MaskRay, arphaman, kadircet, cfe-commits

Differential Revision: https://reviews.llvm.org/D53286

llvm-svn: 344672

5 years ago[ARM][NFCI] Do not fuse VADD and VMUL, continued (1/2)
Sjoerd Meijer [Wed, 17 Oct 2018 07:26:35 +0000 (07:26 +0000)]
[ARM][NFCI] Do not fuse VADD and VMUL, continued (1/2)

This is a follow up of rL342874, which stopped fusing muls and adds into VMLAs
for performance reasons on the Cortex-M4 and Cortex-M33.  This is a serie of 2
patches, that is trying to achieve the same for VFMA.  The second column in the
table below shows what we were generating before rL342874, the third column
what changed with rL342874, and the last column what we want to achieve with
these 2 patches:

 --------------------------------------------------------
 | Opt   |  < rL342874   |  >= rL342874   |             |
 |------------------------------------------------------|
 |-O3    |     vmla      |      vmul      |     vmul    |
 |       |               |      vadd      |     vadd    |
 |------------------------------------------------------|
 |-Ofast |     vfma      |      vfma      |     vmul    |
 |       |               |                |     vadd    |
 |------------------------------------------------------|
 |-Oz    |     vmla      |      vmla      |     vmla    |
 --------------------------------------------------------

This patch 1/2, is a cleanup of the spaghetti predicate logic on the different
VMLA and VFMA codegen rules, so that we can make the final functional change in
patch 2/2.  This also fixes a typo in the regression test added in rL342874.

Differential revision: https://reviews.llvm.org/D53314

llvm-svn: 344671

5 years ago[XRay][compiler-rt] Generational Buffer Management
Dean Michael Berris [Wed, 17 Oct 2018 06:57:50 +0000 (06:57 +0000)]
[XRay][compiler-rt] Generational Buffer Management

Summary:
This change updates the buffer queue implementation to support using a
generation number to identify the lifetime of buffers. This first part
introduces the notion of the generation number, without changing the way
we handle the buffers yet.

What's missing here is the cleanup of the buffers. Ideally we'll keep
the two most recent generations. We need to ensure that before we do any
writes to the buffers, that we check the generation number(s) first.

Those changes will follow-on from this change.

Depends on D52588.

Reviewers: mboerger, eizan

Subscribers: llvm-commits, jfb

Differential Revision: https://reviews.llvm.org/D52974

llvm-svn: 344670

5 years agoDocument the behavior of option passing when using -DCLANG_ENABLE_BOOTSTRAP=On
Sylvestre Ledru [Wed, 17 Oct 2018 06:35:10 +0000 (06:35 +0000)]
Document the behavior of option passing when using -DCLANG_ENABLE_BOOTSTRAP=On
Also document -DCLANG_BOOTSTRAP_PASSTHROUGH

Reviewers: ecbeckmann

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D53018

llvm-svn: 344669

5 years agoNFC: Remove trailing space from CodeGenModule.cpp
Takuto Ikuta [Wed, 17 Oct 2018 04:29:56 +0000 (04:29 +0000)]
NFC: Remove trailing space from CodeGenModule.cpp

llvm-svn: 344668

5 years ago[BuildingAJIT] Update chapter 1 to use the ORCv2 APIs.
Lang Hames [Wed, 17 Oct 2018 03:34:09 +0000 (03:34 +0000)]
[BuildingAJIT] Update chapter 1 to use the ORCv2 APIs.

llvm-svn: 344667

5 years ago[python] [tests] Disable on known-broken arches
Michal Gorny [Wed, 17 Oct 2018 03:05:39 +0000 (03:05 +0000)]
[python] [tests] Disable on known-broken arches

Disable the Python binding tests on AArch64, Hexagon and SystemZ
following reports on test failures.  The first two yield different
results, possibly indicating test case problems.  The last one seems
to have broken FFI in Python.

While at it, refactor the code to make adding future test restrictions
easier.

Differential Revision: https://reviews.llvm.org/D53326

llvm-svn: 344666

5 years agoAMDGPU: add __builtin_amdgcn_update_dpp
Yaxun Liu [Wed, 17 Oct 2018 02:32:26 +0000 (02:32 +0000)]
AMDGPU: add __builtin_amdgcn_update_dpp

Emit llvm.amdgcn.update.dpp for both __builtin_amdgcn_mov_dpp and
__builtin_amdgcn_update_dpp. The first argument to
llvm.amdgcn.update.dpp will be undef for __builtin_amdgcn_mov_dpp.

Differential Revision: https://reviews.llvm.org/D52320

llvm-svn: 344665

5 years ago[analyzer] [www] Minor improvements to the text in open_projects
George Karpenkov [Wed, 17 Oct 2018 01:15:53 +0000 (01:15 +0000)]
[analyzer] [www] Minor improvements to the text in open_projects

llvm-svn: 344664

5 years ago[analyzer] [www] Updated a list of open projects
George Karpenkov [Wed, 17 Oct 2018 01:06:20 +0000 (01:06 +0000)]
[analyzer] [www] Updated a list of open projects

Differential Revision: https://reviews.llvm.org/D53024

llvm-svn: 344663

5 years agoNew test requires x86-registered-target
Teresa Johnson [Wed, 17 Oct 2018 00:59:14 +0000 (00:59 +0000)]
New test requires x86-registered-target

New test added in r344658 also requires x86-registered-target.

llvm-svn: 344662

5 years ago[libcxx] Mark chrono literal unit tests as being unsupported on AppleClang 10
Louis Dionne [Wed, 17 Oct 2018 00:34:13 +0000 (00:34 +0000)]
[libcxx] Mark chrono literal unit tests as being unsupported on AppleClang 10

llvm-svn: 344661

5 years ago[ThinLTO] Fix test to require asserts
Teresa Johnson [Wed, 17 Oct 2018 00:19:21 +0000 (00:19 +0000)]
[ThinLTO] Fix test to require asserts

New test added in r344658 requires asserts due to -stats.

While here, augment it to test new global variable importing
message as well.

llvm-svn: 344660

5 years ago[Sanitizer][PassManager] Fix for failing ASan tests on arm-linux-gnueabihf
Leonard Chan [Wed, 17 Oct 2018 00:16:07 +0000 (00:16 +0000)]
[Sanitizer][PassManager] Fix for failing ASan tests on arm-linux-gnueabihf

Forgot to initialize the legacy pass in it's constructor.

Differential Revision: https://reviews.llvm.org/D53350

llvm-svn: 344659

5 years ago[ThinLTO] Add importing stats to thin link
Teresa Johnson [Tue, 16 Oct 2018 23:49:50 +0000 (23:49 +0000)]
[ThinLTO] Add importing stats to thin link

Summary:
Previously we could only get the number of imported functions and
variables from the backend. This adds stats to the thin link where the
importing is decided.

Reviewers: wmi

Subscribers: inglorion, dexonsmith, llvm-commits

Differential Revision: https://reviews.llvm.org/D53337

llvm-svn: 344658

5 years ago[SanitizerCoverage] Don't duplicate code to get section pointers
Jonathan Metzman [Tue, 16 Oct 2018 23:43:57 +0000 (23:43 +0000)]
[SanitizerCoverage] Don't duplicate code to get section pointers

Summary:
Merge code used to get section start and section end pointers
for SanitizerCoverage constructors. This includes code that handles
getting the start pointers when targeting MSVC.

Reviewers: kcc, morehouse

Reviewed By: morehouse

Subscribers: kcc, hiraditya

Differential Revision: https://reviews.llvm.org/D53211

llvm-svn: 344657

5 years ago[clang-doc] Limit integration tests
Julie Hockett [Tue, 16 Oct 2018 23:07:37 +0000 (23:07 +0000)]
[clang-doc] Limit integration tests

Now that the clang-doc libraries are covered by unit tests, we don't
need to have extensive (and unmaintainable) integration tests. This
replaces the integration test suite with a smaller one that just tests
the tool itself and removes extraneous dumping logic from the tool
itself.

Includes tests that cover the parse->serialize->merge->generate
pipeline, as well as tests for the --public, --format, --doxygen, and
--output flags.

Differential Revision: https://reviews.llvm.org/D53150

llvm-svn: 344655

5 years ago[clang-doc] Add unit tests for Markdown generation
Julie Hockett [Tue, 16 Oct 2018 23:07:26 +0000 (23:07 +0000)]
[clang-doc] Add unit tests for Markdown generation

Add unit tests for Markdown generation.

This is part of a move to convert clang-doc's tests to a more
maintainable unit test framework, with a smaller number of integration
tests to maintain and more granular failure feedback.

Differential Revision: https://reviews.llvm.org/D53085

llvm-svn: 344654

5 years ago[clang-doc] Add unit tests for YAML generation
Julie Hockett [Tue, 16 Oct 2018 23:07:16 +0000 (23:07 +0000)]
[clang-doc] Add unit tests for YAML generation

Adds unit tests for the YAML generator library.

This is part of a move to convert clang-doc's tests to a more
maintainable unit test framework, with a smaller number of integration
tests to maintain and more granular failure feedback.

Differential Revision: https://reviews.llvm.org/D53084

llvm-svn: 344653

5 years ago[clang-doc] Add unit tests for merging
Julie Hockett [Tue, 16 Oct 2018 23:07:04 +0000 (23:07 +0000)]
[clang-doc] Add unit tests for merging

Adds unit tests for the merging logic in Respresentation.cpp.

This is part of a move to convert clang-doc's tests to a more
maintainable unit test framework, with a smaller number of integration
tests to maintain and more granular failure feedback.

Differential Revision: https://reviews.llvm.org/D53083

llvm-svn: 344652

5 years ago[clang-doc] Add unit tests for bitcode
Julie Hockett [Tue, 16 Oct 2018 23:06:53 +0000 (23:06 +0000)]
[clang-doc] Add unit tests for bitcode

Adds unit tests for the BitcodeWriter and BitcodeReader libraries.

This is part of a move to convert clang-doc's tests to a more
maintainable unit test framework, with a smaller number of integration
tests to maintain and more granular failure feedback.

Differential Revision: https://reviews.llvm.org/D53082

llvm-svn: 344651

5 years ago[clang-doc] Add unit tests for serialization
Julie Hockett [Tue, 16 Oct 2018 23:06:42 +0000 (23:06 +0000)]
[clang-doc] Add unit tests for serialization

Adds unit tests for the Serialize library.

This is part of a move to convert clang-doc's tests to a more
maintainable unit test framework, with a smaller number of integration
tests to maintain and more granular failure feedback.

Differential Revision: https://reviews.llvm.org/D53081

llvm-svn: 344650

5 years ago[X86] Match (cmp (and (shr X, C), mask), 0) to BEXTR+TEST.
Craig Topper [Tue, 16 Oct 2018 22:29:36 +0000 (22:29 +0000)]
[X86] Match (cmp (and (shr X, C), mask), 0) to BEXTR+TEST.

Without this we match the CMP+AND to a TEST and then match the SHR separately. I'm trusting analyzeCompare to remove the TEST during the peephole pass. Otherwise we need to check the flag users to see if they only use the Z flag.

This recovers a case lost by r344270.

Differential Revision: https://reviews.llvm.org/D53310

llvm-svn: 344649

5 years agoDelete commented-out code.
Adrian Prantl [Tue, 16 Oct 2018 22:01:49 +0000 (22:01 +0000)]
Delete commented-out code.

llvm-svn: 344648

5 years agoReturn a named error in the result object of an expression with no result
Jim Ingham [Tue, 16 Oct 2018 21:58:40 +0000 (21:58 +0000)]
Return a named error in the result object of an expression with no result

Before we returned an error that was not exposed in the SB API and no useful
error message.  This change returns eExpressionProducedNoResult and an
appropriate error string.

<rdar://problem/44539514>

Differential Revision: https://reviews.llvm.org/D53309

llvm-svn: 344647

5 years agoFixed an issue that a bot found with my changes
Jason Molenda [Tue, 16 Oct 2018 21:49:31 +0000 (21:49 +0000)]
Fixed an issue that a bot found with my changes
in r344626 & recommitting.  Original commit msg:

Simplify LocateDSYMInVincinityOfExecutable by moving
some redundant code into a separate function,
LookForDsymNextToExecutablePath, and having that function
also look for .dSYM.yaa files in addition to .dSYM
bundles.

Differential Revision: https://reviews.llvm.org/D53305

<rdar://problem/40406580>

llvm-svn: 344646

5 years ago[InstCombine] Cleanup libfunc attribute inferring
David Bolvansky [Tue, 16 Oct 2018 21:18:31 +0000 (21:18 +0000)]
[InstCombine] Cleanup libfunc attribute inferring

Reviewers: efriedma

Reviewed By: efriedma

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D53338

llvm-svn: 344645

5 years agoCode cleanup: Remove DWARFDebugInfoEntry::m_empty_children
Jan Kratochvil [Tue, 16 Oct 2018 20:49:15 +0000 (20:49 +0000)]
Code cleanup: Remove DWARFDebugInfoEntry::m_empty_children

It merges DWARFDebugInfoEntry's m_empty_children into m_has_children.
m_empty_children was implemented by rL144983.

As Greg confirmed m_has_children was used to represent what was in the DWARF in
the byte that follows the DW_TAG. m_empty_children was used for DIEs that said
they had children but actually only contain a single NULL tag. It is fine to
not differentiate between the two.

Also changed assert()->lldbassert() for m_abbr_idx 16-bit overflow check as
that could be a tough bug to catch if it ever happens.

I have checked all calls of HasChildren() that this change should not matter to
them. The code even wants to know if there are any children - it does not
matter how the children presence is coded in the binary.

Patch written based on suggestions by Greg Clayton.

Differential Revision: https://reviews.llvm.org/D53321

llvm-svn: 344644

5 years ago[ORC] Make the VModuleKey optional, propagate it via MaterializationUnit and
Lang Hames [Tue, 16 Oct 2018 20:13:06 +0000 (20:13 +0000)]
[ORC] Make the VModuleKey optional, propagate it via MaterializationUnit and
MaterializationResponsibility.

VModuleKeys are intended to enable selective removal of modules from a JIT
session, however for a wide variety of use cases selective removal is not
needed and introduces unnecessary overhead. As of this commit, the default
constructed VModuleKey value is reserved as a "do not track" value, and
becomes the default when adding a new module to the JIT.

This commit also changes the propagation of VModuleKeys. They were passed
alongside the MaterializationResponsibity instance in XXLayer::emit methods,
but are now propagated as part of the MaterializationResponsibility instance
itself (and as part of MaterializationUnit when stored in a JITDylib).
Associating VModuleKeys with MaterializationUnits in this way should allow
for a thread-safe module removal mechanism in the future, even when a module
is in the process of being compiled, by having the
MaterializationResponsibility object check in on its VModuleKey's state
before commiting its results to the JITDylib.

llvm-svn: 344643

5 years ago[libcxx] Avoid repeating the definition of std:: namespaces
Louis Dionne [Tue, 16 Oct 2018 20:02:59 +0000 (20:02 +0000)]
[libcxx] Avoid repeating the definition of std:: namespaces

This reduces code duplication a tiny bit.

llvm-svn: 344642

5 years ago[libcxx] Remove _LIBCPP_EXTERN_TEMPLATE_INLINE_VISIBILITY
Louis Dionne [Tue, 16 Oct 2018 19:26:23 +0000 (19:26 +0000)]
[libcxx] Remove _LIBCPP_EXTERN_TEMPLATE_INLINE_VISIBILITY

That macro has been defined to _LIBCPP_HIDE_FROM_ABI_AFTER_V1 for many
weeks now, so we're actually replacing uses of it for uses of
_LIBCPP_HIDE_FROM_ABI_AFTER_V1 directly.

This should not change or break anything since the two macros are
100% equivalent, unless somebody is (incorrectly!) relying on
_LIBCPP_EXTERN_TEMPLATE_INLINE_VISIBILITY being defined.

llvm-svn: 344641