Dan Albert [Thu, 18 Oct 2018 18:04:28 +0000 (18:04 +0000)]
[compiler-rt] Don't explicitly set CMAKE_CXX_FLAGS.
Summary:
C++ flags should not be used for not-C++ files as it may trigger
-Werror=unused-command-line-argument. CMake will use CMAKE_C_FLAGS,
CMAKE_CXX_FLAGS, and CMAKE_ASM_FLAGS as appropriate implicitly, so
this does not need to be explicitly handled here.
This change depends on https://reviews.llvm.org/D53301, since one of
the builders depended on this behavior because it was not configuring
CMAKE_ASM_FLAGS.
Reviewers: eugenis, vitalybuka
Reviewed By: eugenis, vitalybuka
Subscribers: dberris, mgorny, delcypher, #sanitizers, llvm-commits
Differential Revision: https://reviews.llvm.org/D53335
llvm-svn: 344751
Aaron Ballman [Thu, 18 Oct 2018 17:47:18 +0000 (17:47 +0000)]
Add check-clang-python to the Clang tests directory in IDEs; NFC.
llvm-svn: 344750
Aaron Ballman [Thu, 18 Oct 2018 17:42:41 +0000 (17:42 +0000)]
Add language standard aliases for -std=c18, -std=gnu18, and -std=iso9899:2018.
As described in D40225, the C17 standard was balloted and approved in 2017, but the ISO publication process delayed the actual publication until 2018. WG14 considers the release to be C17 and describes it as such, but users can still be confused by the publication year which is why -std=c18 adds value. These aliases map to c17 and are all supported by GCC 8.x with the same behavior. Note that the value of __STDC_VERSION__ remains at 201710L.
llvm-svn: 344749
Sumanth Gundapaneni [Thu, 18 Oct 2018 15:51:16 +0000 (15:51 +0000)]
[Pipeliner] copyToPhi DAG Mutation to improve scheduling.
In a loop, create artificial dependences between the source of a
COPY/REG_SEQUENCE to the use in next iteration.
Eg:
SRC ----Data Dep--> COPY
COPY ---Anti Dep--> PHI (implies, to be used in next iteration)
PHI ----Data Dep--> USE
This patches creates
USE ----Artificial Dep---> SRC
This will effectively schedule the COPY late to eliminate additional copies.
Before this patch, the schedule can be
SRC, COPY, USE : The COPY is used in next iteration and it needs to be
preserved.
After this patch, the schedule can be
USE, SRC, COPY : The COPY is used in next iteration and the live interval is
reduced.
Differential Revision: https://reviews.llvm.org/D53303
llvm-svn: 344748
Sean Fertile [Thu, 18 Oct 2018 15:43:41 +0000 (15:43 +0000)]
[PPC64] Fix offset checks on rel24 call relocations.
Adjusted the range check on a call instruction from 24 bits signed to
26 bits signed. While the instruction only encodes 24 bits, the target is
assumed to be 4 byte aligned, and the value that is encoded in the instruction
gets shifted left by 2 to form the offset. Also added a check that the offset is
indeed at least 4 byte aligned.
Differential Revision: https://reviews.llvm.org/D53401
llvm-svn: 344747
Michal Gorny [Thu, 18 Oct 2018 15:39:22 +0000 (15:39 +0000)]
[SymbolFileNativePDB] Fix missing linkage to DebugInfoCodeView
Differential Revision: https://reviews.llvm.org/D53402
llvm-svn: 344746
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
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
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
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
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
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
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
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
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
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
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
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
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
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
Guillaume Chatelet [Thu, 18 Oct 2018 08:20:50 +0000 (08:20 +0000)]
[llvm-exegesis] Fix off by one error
llvm-svn: 344731
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
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
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
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
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
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
Mikael Holmen [Thu, 18 Oct 2018 06:00:39 +0000 (06:00 +0000)]
Fix warning about unused variable [NFC]
llvm-svn: 344724
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
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
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
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
Chandler Carruth [Thu, 18 Oct 2018 00:40:26 +0000 (00:40 +0000)]
[TI removal] Switch simple loop unswitch to `Instruction`.
llvm-svn: 344719
Chandler Carruth [Thu, 18 Oct 2018 00:39:46 +0000 (00:39 +0000)]
[TI removal] Switch NewGVN to directly use `Instruction`.
llvm-svn: 344718
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
Chandler Carruth [Thu, 18 Oct 2018 00:38:54 +0000 (00:38 +0000)]
[TI removal] Update CodeExtractor to use Instruction directly.
llvm-svn: 344716
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
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
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
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
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
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
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
Lang Hames [Wed, 17 Oct 2018 19:35:38 +0000 (19:35 +0000)]
[BuildingAJIT] Fix a function signature in the documentation.
llvm-svn: 344705
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
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
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
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
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
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
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
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
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
Theodoros Theodoridis [Wed, 17 Oct 2018 15:01:18 +0000 (15:01 +0000)]
Fix broken formatting caused by test commit
llvm-svn: 344694
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
Guillaume Chatelet [Wed, 17 Oct 2018 12:27:46 +0000 (12:27 +0000)]
Fix uninitialized variable
llvm-svn: 344692
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
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
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
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
Max Kazantsev [Wed, 17 Oct 2018 11:16:25 +0000 (11:16 +0000)]
[NFC] Remove GOTO from SCEV
llvm-svn: 344687
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
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
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
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
Theodoros Theodoridis [Wed, 17 Oct 2018 09:59:41 +0000 (09:59 +0000)]
Test commit
llvm-svn: 344682
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
Haojian Wu [Wed, 17 Oct 2018 08:54:48 +0000 (08:54 +0000)]
[clangd] Fix buildbot failure.
llvm-svn: 344680
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
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
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
Sam McCall [Wed, 17 Oct 2018 07:41:53 +0000 (07:41 +0000)]
[clangd] Hide unused function. NFC
llvm-svn: 344676
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
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
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
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
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
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
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
Takuto Ikuta [Wed, 17 Oct 2018 04:29:56 +0000 (04:29 +0000)]
NFC: Remove trailing space from CodeGenModule.cpp
llvm-svn: 344668
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
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
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
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
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
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
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
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
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
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
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
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
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
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
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
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
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
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
Adrian Prantl [Tue, 16 Oct 2018 22:01:49 +0000 (22:01 +0000)]
Delete commented-out code.
llvm-svn: 344648
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