Philip Pfaffe [Wed, 16 Jan 2019 09:28:01 +0000 (09:28 +0000)]
[NewPM][TSan] Reiterate the TSan port
Summary:
Second iteration of D56433 which got reverted in rL350719. The problem
in the previous version was that we dropped the thunk calling the tsan init
function. The new version keeps the thunk which should appease dyld, but is not
actually OK wrt. the current semantics of function passes. Hence, add a
helper to insert the functions only on the first time. The helper
allows hooking into the insertion to be able to append them to the
global ctors list.
Reviewers: chandlerc, vitalybuka, fedor.sergeev, leonardchan
Subscribers: hiraditya, bollu, llvm-commits
Differential Revision: https://reviews.llvm.org/D56538
llvm-svn: 351314
George Rimar [Wed, 16 Jan 2019 09:27:04 +0000 (09:27 +0000)]
[lldb] - Fix crash when listing the history with the key up.
This is https://bugs.llvm.org/show_bug.cgi?id=40112,
Currently, lldb crashes after pressing the up arrow key when listing the history for expressions.
The patch fixes the mistype that was a reason.
Differential revision: https://reviews.llvm.org/D56014
llvm-svn: 351313
Hans Wennborg [Wed, 16 Jan 2019 09:13:47 +0000 (09:13 +0000)]
UsersManual.rst: Update the clang-cl flags section
llvm-svn: 351312
Joachim Protze [Wed, 16 Jan 2019 08:58:17 +0000 (08:58 +0000)]
[OMPT] Make sure that OMPT is enabled when accessing internals of the runtime
Make sure that OMPT is enabled in runtime entry points that access internals
of the runtime. Else, return an appropiate value indicating an error or that
the data is not available.
Patch provided by @sconvent
Reviewers: jlpeyton, omalyshe, hbae, Hahnfeld, joachim.protze
Reviewed By: joachim.protze
Tags: #openmp, #ompt
Differential Revision: https://reviews.llvm.org/D47717
llvm-svn: 351311
Sam Parker [Wed, 16 Jan 2019 08:40:12 +0000 (08:40 +0000)]
[DAGCombine] Fix ReduceLoadWidth for shifted offsets
ReduceLoadWidth can trigger using a shifted mask is used and this
requires that the function return a shl node to correct for the
offset. However, the way that this was implemented meant that the
returned result could be an existing node, which would be incorrect.
This fixes the method of inserting the new node and replacing uses.
Differential Revision: https://reviews.llvm.org/D50432
llvm-svn: 351310
Hans Wennborg [Wed, 16 Jan 2019 08:38:28 +0000 (08:38 +0000)]
Include lldb in Win snapshots again (PR37307)
llvm-svn: 351309
Clement Courbet [Wed, 16 Jan 2019 08:36:23 +0000 (08:36 +0000)]
[clang-tidy] bugprone-string-constructor: Catch string from nullptr.
Summary: Context: https://twitter.com/willkirkby/status/
1084219580799741953
Subscribers: xazax.hun, cfe-commits
Differential Revision: https://reviews.llvm.org/D56657
llvm-svn: 351308
Sam McCall [Wed, 16 Jan 2019 08:29:54 +0000 (08:29 +0000)]
[clang-tidy] Avoid fuchsia-overloaded-operator false positive with lambdas after r351047
llvm-svn: 351307
Fangrui Song [Wed, 16 Jan 2019 08:13:15 +0000 (08:13 +0000)]
[clangd] Fix XPC after rCTE351280
llvm-svn: 351306
Martin Storsjo [Wed, 16 Jan 2019 08:09:22 +0000 (08:09 +0000)]
[llvm-rc] Support '--' for delimiting options from input paths
This allows avoiding conflicts between paths that begin with the same
chars as some llvm-rc options (which can be used with either slashes
or dashes).
Differential Revision: https://reviews.llvm.org/D56743
llvm-svn: 351305
Michal Gorny [Wed, 16 Jan 2019 08:05:13 +0000 (08:05 +0000)]
[test] Disable Python binding tests w/ LLVM_ENABLE_PIC=OFF
Disable Python binding tests when LLVM_ENABLE_PIC is disabled,
as libclang.so is not being built in that case. Reported by Nico Weber.
Differential Revision: https://reviews.llvm.org/D56732
llvm-svn: 351304
Martin Bohme [Wed, 16 Jan 2019 07:53:25 +0000 (07:53 +0000)]
[clang-tidy] Treat references to smart pointers correctly in use-after-move.
Summary:
Previously, we weren't recognizing these as smart pointers and thus
weren't allowing non-dereference accesses as we should -- see new test
cases which fail without the fix.
Reviewers: alexfh, hokein, aaron.ballman, JonasToth
Reviewed By: JonasToth
Subscribers: xazax.hun, cfe-commits
Tags: #clang-tools-extra
Differential Revision: https://reviews.llvm.org/D56585
llvm-svn: 351303
Sanjin Sijaric [Wed, 16 Jan 2019 07:39:44 +0000 (07:39 +0000)]
[SEH] Pass the frame pointer from SEH finally to finally functions
Pass the frame pointer that the first finally block receives onto the nested
finally block, instead of generating it using localaddr.
Differential Revision: https://reviews.llvm.org/D56463
llvm-svn: 351302
Craig Topper [Wed, 16 Jan 2019 07:17:14 +0000 (07:17 +0000)]
[X86] Correct the type string for __builtin_ia32_gathersiv16sf to make the indices an integer type not an FP type.
The element count and width remain the same. This went unnoticed because default conversion from builtin to intrinsic will generate a bitcast if the types don't match.
llvm-svn: 351301
Dmitry Venikov [Wed, 16 Jan 2019 07:05:58 +0000 (07:05 +0000)]
[llvm-symbolizer] Add -C as a short alias to -demangle
Summary: Provides -C as alias to -demangle. Motivation: https://bugs.llvm.org/show_bug.cgi?id=40069.
Reviewers: jhenderson, ruiu, rnk, fjricci
Reviewed By: jhenderson, ruiu
Subscribers: rupprecht, erik.pilkington, llvm-commits
Differential Revision: https://reviews.llvm.org/D56591
llvm-svn: 351300
Eric Fiselier [Wed, 16 Jan 2019 05:43:02 +0000 (05:43 +0000)]
correct script name in generated tests
llvm-svn: 351299
Dan Gohman [Wed, 16 Jan 2019 05:23:57 +0000 (05:23 +0000)]
[WebAssembly] COWS has been renamed to WASI.
llvm-svn: 351298
Dan Gohman [Wed, 16 Jan 2019 05:23:52 +0000 (05:23 +0000)]
[WebAssembly] COWS has been renamed to WASI.
llvm-svn: 351297
Tom Stellard [Wed, 16 Jan 2019 05:15:31 +0000 (05:15 +0000)]
Only promote args when function attributes are compatible
Summary:
Check to make sure that the caller and the callee have compatible
function arguments before promoting arguments. This uses the same
TargetTransformInfo queries that are used to determine if attributes
are compatible for inlining.
The goal here is to avoid breaking ABI when a called function's ABI
depends on a target feature that is not enabled in the caller.
This is a very conservative fix for PR37358. Ideally we would have a more
sophisticated check for ABI compatiblity rather than checking if the
attributes are compatible for inlining.
Reviewers: echristo, chandlerc, eli.friedman, craig.topper
Reviewed By: echristo, chandlerc
Subscribers: nikic, xbolva00, rkruppe, alexcrichton, llvm-commits
Differential Revision: https://reviews.llvm.org/D53554
llvm-svn: 351296
Serguei Katkov [Wed, 16 Jan 2019 04:36:26 +0000 (04:36 +0000)]
[InstCombine]Avoid introduction of unaligned mem access
InstCombine is able to transform mem transfer instrinsic to alone store or store/load pair.
It might result in generation of unaligned atomic load/store which later in backend
will be transformed to libcall. It is not an evident gain and it is better to keep intrinsic as is
and handle it at backend.
Reviewers: reames, anna, apilipenko, mkazantsev
Reviewed By: reames
Subscribers: t.p.northover, jfb, llvm-commits
Differential Revision: https://reviews.llvm.org/D56582
llvm-svn: 351295
Eric Fiselier [Wed, 16 Jan 2019 02:34:36 +0000 (02:34 +0000)]
[SemaCXX] Unconfuse Clang when std::align_val_t is unscoped in C++03
When -faligned-allocation is specified in C++03 libc++ defines
std::align_val_t as an unscoped enumeration type (because Clang didn't
provide scoped enumerations as an extension until 8.0).
Unfortunately Clang confuses the `align_val_t` overloads of delete with
the sized deallocation overloads which aren't enabled. This caused Clang
to call the aligned deallocation function as if it were the sized
deallocation overload.
For example: https://godbolt.org/z/xXJELh
This patch fixes the confusion.
llvm-svn: 351294
Peter Collingbourne [Wed, 16 Jan 2019 02:27:12 +0000 (02:27 +0000)]
gn build: Merge r351283.
llvm-svn: 351293
Eric Fiselier [Wed, 16 Jan 2019 02:16:57 +0000 (02:16 +0000)]
Attempt to make test_macros.h even more minimal
llvm-svn: 351292
Eric Fiselier [Wed, 16 Jan 2019 02:10:28 +0000 (02:10 +0000)]
Fix feature test macros for atomics/mutexes without threading
llvm-svn: 351291
Eric Fiselier [Wed, 16 Jan 2019 01:54:34 +0000 (01:54 +0000)]
Fix PR40230 - std::pair may have padding on FreeBSD.
Summary:
FreeBSD ships a very old and deprecated ABI for std::pair where the copy and move constructors are not allowed to be trivial. D25389 change how this was implemented by introducing a non-trivial base class. This patch, introduced in October 2016, introduced an ABI bug that caused nested `std::pair` instantiations to have padding. For example:
```
using PairT = std::pair< std::pair<char, char>, char >;
static_assert(offsetof(PairT, first) == 0, "First member should exist at offset zero"); // Fails on FreeBSD!
```
The bug occurs because the base class for the first element (the nested pair) cannot be put at offset zero because the top-level pair already has the same base class laid out there.
This patch fixes that ABI bug by templating the dummy base class on the same parameters as the pair.
Technically this fix is an ABI break for users who depend on the "broken" ABI introduced in 2016. I'm putting this up for review so that the FreeBSD maintainers can sign off on fixing the ABI by breaking the ABI.
Another option, since we have to "break" the ABI to fix it, would be to move FreeBSD off the deprecated non-trivial pair ABI instead.
Also see:
* https://llvm.org/PR40230
* https://reviews.llvm.org/D21329
Reviewers: rsmith, dim, emaste
Reviewed By: rsmith
Subscribers: mclow.lists, krytarowski, christof, ldionne, libcxx-commits
Differential Revision: https://reviews.llvm.org/D56357
llvm-svn: 351290
Eric Fiselier [Wed, 16 Jan 2019 01:51:12 +0000 (01:51 +0000)]
Move internal usages of `alignof`/`__alignof` to use `_LIBCPP_ALIGNOF`.
Summary:
Starting in Clang 8.0 and GCC 8.0, `alignof` and `__alignof` return different values in same cases. Specifically `alignof` and `_Alignof` return the minimum alignment for a type, where as `__alignof` returns the preferred alignment. libc++ currently uses `__alignof` but means to use `alignof`. See llvm.org/PR39713
This patch introduces the macro `_LIBCPP_ALIGNOF` so we can control which spelling gets used.
This patch does not introduce any ABI guard to provide the old behavior with newer compilers. However, if we decide that is needed, this patch makes it trivial to implement.
I think we should commit this change immediately, and decide what we want to do about the ABI afterwards.
Reviewers: ldionne, EricWF
Reviewed By: ldionne, EricWF
Subscribers: jyknight, christof, libcxx-commits
Differential Revision: https://reviews.llvm.org/D54814
llvm-svn: 351289
Julian Lettner [Wed, 16 Jan 2019 01:45:12 +0000 (01:45 +0000)]
[TSan] Use switches when dealing with enums
Summary:
Small refactoring: replace some if-else cascades with switches so that the compiler warns us about missing cases.
Maybe found a small bug?
Reviewers: dcoughlin, kubamracek, dvyukov, delcypher, jfb
Reviewed By: dvyukov
Subscribers: llvm-commits, #sanitizers
Tags: #sanitizers
Differential Revision: https://reviews.llvm.org/D56295
llvm-svn: 351288
Sam Clegg [Wed, 16 Jan 2019 01:43:21 +0000 (01:43 +0000)]
[WebAssembly] Store section alignment as a power of 2
This change bumps for version number of the wasm object file
metadata.
See https://github.com/WebAssembly/tool-conventions/pull/92
Differential Revision: https://reviews.llvm.org/D56762
llvm-svn: 351287
Eric Fiselier [Wed, 16 Jan 2019 01:37:43 +0000 (01:37 +0000)]
Implement feature test macros using a script.
Summary:
This patch implements all the feature test macros libc++ currently supports, as specified by the standard or cppreference prior to C++2a.
The tests and `<version>` header are generated using a script. The script contains a table of each feature test macro, the headers it should be accessible from, and its values of each dialect of C++.
When a new feature test macro is added or needed, the table should be updated and the script re-run.
Reviewers: mclow.lists, jfb, serge-sans-paille
Reviewed By: mclow.lists
Subscribers: arphaman, jfb, ldionne, libcxx-commits
Differential Revision: https://reviews.llvm.org/D56750
llvm-svn: 351286
Sam Clegg [Wed, 16 Jan 2019 01:34:48 +0000 (01:34 +0000)]
[WebAssembly] Store section alignment as a power of 2
This change bumps for version number of the wasm object file
metadata.
See https://github.com/WebAssembly/tool-conventions/pull/92
Differential Revision: https://reviews.llvm.org/D56758
llvm-svn: 351285
Eli Friedman [Wed, 16 Jan 2019 00:50:44 +0000 (00:50 +0000)]
[EH] Rename llvm.x86.seh.recoverfp intrinsic to llvm.eh.recoverfp
This is the clang counterpart to D56747.
Patch by Mandeep Singh Grang.
Differential Revision: https://reviews.llvm.org/D56748
llvm-svn: 351284
Aditya Nandakumar [Wed, 16 Jan 2019 00:40:37 +0000 (00:40 +0000)]
[GISel]: Add support for CSEing continuously during GISel passes.
https://reviews.llvm.org/D52803
This patch adds support to continuously CSE instructions during
each of the GISel passes. It consists of a GISelCSEInfo analysis pass
that can be used by the CSEMIRBuilder.
llvm-svn: 351283
Vlad Tsyrklevich [Wed, 16 Jan 2019 00:37:39 +0000 (00:37 +0000)]
Revert "[Tooling] Make clang-tool find libc++ dir on mac when running on a file without compilation database."
This reverts commits r351222 and r351229, they were causing ASan/MSan failures
on the sanitizer bots.
llvm-svn: 351282
Mandeep Singh Grang [Wed, 16 Jan 2019 00:37:13 +0000 (00:37 +0000)]
[EH] Rename llvm.x86.seh.recoverfp intrinsic to llvm.eh.recoverfp
Summary:
Make recoverfp intrinsic target-independent so that it can be implemented for AArch64, etc.
Refer D53541 for the context. Clang counterpart D56748.
Reviewers: rnk, efriedma
Reviewed By: rnk, efriedma
Subscribers: javed.absar, kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D56747
llvm-svn: 351281
Jan Korous [Wed, 16 Jan 2019 00:24:22 +0000 (00:24 +0000)]
[clangd] XPC transport layer
- New transport layer for macOS.
- XPC Framework
- Test client
Framework and client were written by Alex Lorenz.
Differential Revision: https://reviews.llvm.org/D54428
llvm-svn: 351280
Craig Topper [Wed, 16 Jan 2019 00:21:59 +0000 (00:21 +0000)]
[LangRef] Fix typo adress->address. NFC
llvm-svn: 351279
Craig Topper [Wed, 16 Jan 2019 00:20:30 +0000 (00:20 +0000)]
[X86] Rename SHRUNKBLEND ISD node to BLENDV.
That's really what it is. If we didn't use intrinsics for BLENDVPS/BLENDVPD/PBLENDVB all the way to isel, this is the node we would use.
llvm-svn: 351278
Peter Collingbourne [Wed, 16 Jan 2019 00:15:25 +0000 (00:15 +0000)]
gn build: Add check-hwasan target.
The Android sanitizer tests are currently some of the most difficult
to run correctly, requiring at least 3 build directories which have
to be configured in just the right way and built in the correct order
(see e.g. [1] and the functions that it calls).
This patch adds a check-hwasan target which greatly simplifies running
the hwasan tests for gn users, taking advantage of its support for
multiple toolchains. With this the tests can be run simply by setting
an NDK path and running "ninja check-hwasan" with a compatible Android
device connected. The Linux/x86_64 and Android/aarch64 targets are
tested in parallel.
[1] https://github.com/llvm/llvm-zorg/blob/master/zorg/buildbot/builders/sanitizers/buildbot_android.sh
Differential Revision: https://reviews.llvm.org/D56713
llvm-svn: 351277
Alex Langford [Wed, 16 Jan 2019 00:09:50 +0000 (00:09 +0000)]
[lldb-mi] Remove use of dialog box
Summary:
This really is only implemented on Windows, and it requires us to pull
in User32. This was only useful when debugging on lldb-mi on Windows, and there
doesn't seem to be a good reason why using a dialog box is better than what
exists for other platforms.
Reviewers: zturner, jingham, compnerd
Subscribers: ki.stfu
Differential Revision: https://reviews.llvm.org/D56755
llvm-svn: 351276
Craig Topper [Tue, 15 Jan 2019 23:36:25 +0000 (23:36 +0000)]
[X86] Add avx512 scatter intrinsics that use a vXi1 mask instead of a scalar integer.
We're trying to have the vXi1 types in IR as much as possible. This prevents the need for bitcasts when the producer of the mask was already a vXi1 value like an icmp. The bitcasts can be subject to code motion and interfere with basic block at a time isel in bad ways.
llvm-svn: 351275
Adrian Prantl [Tue, 15 Jan 2019 23:33:26 +0000 (23:33 +0000)]
Remove redundant check.
llvm-svn: 351274
Changpeng Fang [Tue, 15 Jan 2019 23:12:36 +0000 (23:12 +0000)]
AMDGPU: Raise the priority of MAD24 in instruction selection.
Summary:
We have seen performance regression when v_add3 is generated. The major reason is that the v_mad pattern
is broken when v_add3 is generated. We also see the register pressure increased. While we could not properly
estimate register pressure during instruction selection, we can give mad a higher priority.
In this work, we raise the priority for mad24 in selection and resolve the performance regression.
Reviewers:
rampitec
Differential Revision:
https://reviews.llvm.org/D56745
llvm-svn: 351273
Stephen Kelly [Tue, 15 Jan 2019 23:07:30 +0000 (23:07 +0000)]
Re-order type param children of ObjC nodes
Reviewers: aaron.ballman
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D55394
llvm-svn: 351272
Stephen Kelly [Tue, 15 Jan 2019 23:05:11 +0000 (23:05 +0000)]
NFC: Some cleanups that I missed in the previous commit
llvm-svn: 351271
Peter Collingbourne [Tue, 15 Jan 2019 22:53:24 +0000 (22:53 +0000)]
compiler-rt/test: Bring back -pie on Android.
Looks like the sanitizer-x86_64-linux-android bot started failing
because -pie is still needed when targeting API levels < 16 (which
is the case by default for arm and i686).
llvm-svn: 351270
Stephen Kelly [Tue, 15 Jan 2019 22:50:37 +0000 (22:50 +0000)]
Re-order overrides in FunctionDecl dump
Output all content which is local to the FunctionDecl before traversing
to child AST nodes.
This is necessary so that all of the part which is local to the
FunctionDecl can be split into a different method.
Reviewers: aaron.ballman
Differential Revision: https://reviews.llvm.org/D55083
llvm-svn: 351269
Stephen Kelly [Tue, 15 Jan 2019 22:45:46 +0000 (22:45 +0000)]
NFC: Replace iterator loop with cxx_range_for
llvm-svn: 351268
Jonas Devlieghere [Tue, 15 Jan 2019 22:36:56 +0000 (22:36 +0000)]
[VFS] Add getter for mapping entries.
When generating a reproducer in LLDB we build up the mapping but don't
immediately copy over the files on the file system.
Rather than keeping a separate data structure with real and virtual
paths, we might as well reuse the entries already stored in the
YAMLVFSWriter to lazily copy over the files when needed.
llvm-svn: 351266
Jonas Devlieghere [Tue, 15 Jan 2019 22:36:41 +0000 (22:36 +0000)]
[VFS] Move RedirectingFileSystem interface into header (NFC)
This moves the RedirectingFileSystem into the header so it can be
extended. This is needed in LLDB we need a way to obtain the external
path to deal with FILE* and file descriptor APIs.
Discussion on the mailing list:
http://lists.llvm.org/pipermail/llvm-dev/2018-November/127755.html
Differential revision: https://reviews.llvm.org/D54277
llvm-svn: 351265
Adrian Prantl [Tue, 15 Jan 2019 22:30:01 +0000 (22:30 +0000)]
Simplify code by using Optional::getValueOr()
llvm-svn: 351264
Alex Langford [Tue, 15 Jan 2019 22:27:59 +0000 (22:27 +0000)]
[debugserver][CMake] Remove commented out line
This has been commented out since rL300111
(commit
d742d081f3a1e7412cc609765139ba32d597ac15). Looks like it was
committed as a commented out line, so I'm removing it.
llvm-svn: 351263
Jonathan Metzman [Tue, 15 Jan 2019 22:12:51 +0000 (22:12 +0000)]
[libFuzzer] Remove unstable edge handling
Summary:
Remove code for handling unstable edges from libFuzzer since
it has not been found useful.
Differential Revision: https://reviews.llvm.org/D56730
llvm-svn: 351262
Paul Hoad [Tue, 15 Jan 2019 22:06:49 +0000 (22:06 +0000)]
[clang-tidy] add options documentation to readability-identifier-naming checker
Summary:
The documentation for this clang-checker did not explain what the options are. But this checkers only works with at least some options defined.
To discover the options, you have to read the source code. This shouldn't be necessary for users who just have access to the clang-tidy binary.
This revision, explains the options and gives an example.
Patch by MyDeveloperDay.
Reviewers: JonasToth, Eugene.Zelenko
Reviewed By: JonasToth
Subscribers: xazax.hun, Eugene.Zelenko
Differential Revision: https://reviews.llvm.org/D56563
llvm-svn: 351261
Peter Collingbourne [Tue, 15 Jan 2019 22:06:48 +0000 (22:06 +0000)]
compiler-rt/test: Add a couple of convenience features for Android.
Add a ANDROID_SERIAL_FOR_TESTING CMake variable. This lets you
run the tests with multiple devices attached without having to set
ANDROID_SERIAL.
Add a mechanism for pushing files to the device. Currently most
sanitizers require llvm-symbolizer and the sanitizer runtime to
be pushed to the device. This lets the sanitizer make this happen
automatically before running the tests by specifying the paths in
the lit.site.cfg file.
Differential Revision: https://reviews.llvm.org/D56712
llvm-svn: 351260
Jordan Rupprecht [Tue, 15 Jan 2019 22:03:08 +0000 (22:03 +0000)]
[libObject] Tweak expected error output from llvm-ar
llvm-svn: 351259
Peter Collingbourne [Tue, 15 Jan 2019 22:02:12 +0000 (22:02 +0000)]
gn build: Add a stage2 host toolchain and make the hwasan runtime buildable on x86_64 Linux.
Differential Revision: https://reviews.llvm.org/D56711
llvm-svn: 351258
Rong Xu [Tue, 15 Jan 2019 21:59:17 +0000 (21:59 +0000)]
[profile] Sync up InstrProfData.inc with llvm copy /NFC
llvm-svn: 351257
Jordan Rupprecht [Tue, 15 Jan 2019 21:52:31 +0000 (21:52 +0000)]
[llvm-ar] Resubmit recursive thin archive test with fix for full path names and better error messages
llvm-svn: 351256
Peter Collingbourne [Tue, 15 Jan 2019 21:44:59 +0000 (21:44 +0000)]
gn build: Add a resource_dir.gni file.
The path to the resource directory will end up being used in several
more places once the support for running check-hwasan lands. This
moves the definition to a central location so that it can be used
from those places.
Differential Revision: https://reviews.llvm.org/D56700
llvm-svn: 351255
Craig Topper [Tue, 15 Jan 2019 21:41:31 +0000 (21:41 +0000)]
[X86] Add the GCCBuiltin name back to the deprecated avx512 gather intrinsics until the clang side patch for the new versions is approved.
llvm-svn: 351254
Roman Lebedev [Tue, 15 Jan 2019 21:31:18 +0000 (21:31 +0000)]
X86DAGToDAGISel::matchBitExtract() with truncation (PR36419)
Summary:
Previously in D54095 i have added support for extraction of `lshr` from `X` if we are to produce `BEXTR`.
That was good, but the fix was partial, there was still [[ https://bugs.llvm.org/show_bug.cgi?id=36419 | PR36419 ]].
That pattern can also appear, roughly, when you have a large (64-bit) storage, and the consume bits from it.
It will not be unexpected if you will be doing further computations in 32-bit width.
And then the current code breaks, as the tests show.
The basic idea/pattern here is following:
1. We have `i64` input
2. We perform `i64` right-shift on it.
3. We `trunc`ate that shifted value
4. We do all further work (masking) in `i32`
Since we see `trunc`ation and not `lshr`, we give up, and stop trying to extract that right-shift.
BUT. The mask is `i32`, therefore we can extend both of the operands of the masking (`and`) to `i64`
and truncate the result after masking: https://rise4fun.com/Alive/K4B
```
Name: @bextr64_32_b1 -> @bextr64_32_b0
%shiftedval = lshr i64 %val, %numskipbits
%truncshiftedval = trunc i64 %shiftedval to i32
%widenumlowbits1 = zext i8 %numlowbits to i32
%notmask1 = shl nsw i32 -1, %widenumlowbits1
%mask1 = xor i32 %notmask1, -1
%res = and i32 %truncshiftedval, %mask1
=>
%shiftedval = lshr i64 %val, %numskipbits
%widenumlowbits = zext i8 %numlowbits to i64
%notmask = shl nsw i64 -1, %widenumlowbits
%mask = xor i64 %notmask, -1
%wideres = and i64 %shiftedval, %mask
%res = trunc i64 %wideres to i32
```
Thus, we are again able to extract that `lshr` into `BEXTR`'s control.
Now, the perf (via `llvm-exegesis`) of the snippet suggests that it is not a good idea:
```
$ cat /tmp/old.s
# bextr64_32_b1
# LLVM-EXEGESIS-LIVEIN RSI
# LLVM-EXEGESIS-LIVEIN EDX
# LLVM-EXEGESIS-LIVEIN RDI
movq %rsi, %rcx
shrq %cl, %rdi
shll $8, %edx
bextrl %edx, %edi, %eax
$ cat /tmp/old.s | ./bin/llvm-exegesis -mode=latency -snippets-file=-
Check generated assembly with: /usr/bin/objdump -d /tmp/snippet-1e0082.o
---
mode: latency
key:
instructions:
- 'MOV64rr RCX RSI'
- 'SHR64rCL RDI RDI'
- 'SHL32ri EDX EDX i_0x8'
- 'BEXTR32rr EAX EDI EDX'
config: ''
register_initial_values: []
cpu_name: bdver2
llvm_triple: x86_64-unknown-linux-gnu
num_repetitions: 10000
measurements:
- { key: latency, value: 0.6638, per_snippet_value: 2.6552 }
error: ''
info: ''
assembled_snippet: 4889F148D3EFC1E208C4E268F7C74889F148D3EFC1E208C4E268F7C74889F148D3EFC1E208C4E268F7C74889F148D3EFC1E208C4E268F7C7C3
...
$ cat /tmp/old.s | ./bin/llvm-exegesis -mode=uops -snippets-file=-
Check generated assembly with: /usr/bin/objdump -d /tmp/snippet-43e346.o
---
mode: uops
key:
instructions:
- 'MOV64rr RCX RSI'
- 'SHR64rCL RDI RDI'
- 'SHL32ri EDX EDX i_0x8'
- 'BEXTR32rr EAX EDI EDX'
config: ''
register_initial_values: []
cpu_name: bdver2
llvm_triple: x86_64-unknown-linux-gnu
num_repetitions: 10000
measurements:
- { key: PdFPU0, value: 0, per_snippet_value: 0 }
- { key: PdFPU1, value: 0, per_snippet_value: 0 }
- { key: PdFPU2, value: 0, per_snippet_value: 0 }
- { key: PdFPU3, value: 0, per_snippet_value: 0 }
- { key: NumMicroOps, value: 1.2571, per_snippet_value: 5.0284 }
error: ''
info: ''
assembled_snippet: 4889F148D3EFC1E208C4E268F7C74889F148D3EFC1E208C4E268F7C74889F148D3EFC1E208C4E268F7C74889F148D3EFC1E208C4E268F7C7C3
...
```
vs
```
$ cat /tmp/new.s
# bextr64_32_b1
# LLVM-EXEGESIS-LIVEIN RDX
# LLVM-EXEGESIS-LIVEIN SIL
# LLVM-EXEGESIS-LIVEIN RDI
shlq $8, %rdx
movzbl %sil, %eax
orq %rdx, %rax
bextrq %rax, %rdi, %rax
$ cat /tmp/new.s | ./bin/llvm-exegesis -mode=latency -snippets-file=-
Check generated assembly with: /usr/bin/objdump -d /tmp/snippet-8944f1.o
---
mode: latency
key:
instructions:
- 'SHL64ri RDX RDX i_0x8'
- 'MOVZX32rr8 EAX SIL'
- 'OR64rr RAX RAX RDX'
- 'BEXTR64rr RAX RDI RAX'
config: ''
register_initial_values: []
cpu_name: bdver2
llvm_triple: x86_64-unknown-linux-gnu
num_repetitions: 10000
measurements:
- { key: latency, value: 0.7454, per_snippet_value: 2.9816 }
error: ''
info: ''
assembled_snippet: 48C1E208400FB6C64809D0C4E2F8F7C748C1E208400FB6C64809D0C4E2F8F7C748C1E208400FB6C64809D0C4E2F8F7C748C1E208400FB6C64809D0C4E2F8F7C7C3
...
$ cat /tmp/new.s | ./bin/llvm-exegesis -mode=uops -snippets-file=-
Check generated assembly with: /usr/bin/objdump -d /tmp/snippet-da403c.o
---
mode: uops
key:
instructions:
- 'SHL64ri RDX RDX i_0x8'
- 'MOVZX32rr8 EAX SIL'
- 'OR64rr RAX RAX RDX'
- 'BEXTR64rr RAX RDI RAX'
config: ''
register_initial_values: []
cpu_name: bdver2
llvm_triple: x86_64-unknown-linux-gnu
num_repetitions: 10000
measurements:
- { key: PdFPU0, value: 0, per_snippet_value: 0 }
- { key: PdFPU1, value: 0, per_snippet_value: 0 }
- { key: PdFPU2, value: 0, per_snippet_value: 0 }
- { key: PdFPU3, value: 0, per_snippet_value: 0 }
- { key: NumMicroOps, value: 1.2571, per_snippet_value: 5.0284 }
error: ''
info: ''
assembled_snippet: 48C1E208400FB6C64809D0C4E2F8F7C748C1E208400FB6C64809D0C4E2F8F7C748C1E208400FB6C64809D0C4E2F8F7C748C1E208400FB6C64809D0C4E2F8F7C7C3
...
```
^ latency increased (worse).
Except //maybe// not really.
Like with all synthetic benchmarks, they //may// be misleading.
Let's take a look on some actual real-world hotpath.
In this case it's 'my' [[ https://github.com/darktable-org/rawspeed | RawSpeed ]]'s `BitStream<>::peekBitsNoFill()`, in [[ https://github.com/darktable-org/rawspeed/blob/
e3316dc85127c2c29baa40f998f198a7b278bf36/src/librawspeed/decompressors/VC5Decompressor.cpp#L814 | GoPro VC5 decompressor ]]:
```
raw.pixls.us-unique/GoPro/HERO6 Black$ /usr/src/googlebenchmark/tools/compare.py -a benchmarks ~/rawspeed/build-clangs1-{old,new}/src/utilities/rsbench/rsbench --benchmark_counters_tabular=true --benchmark_min_time=0.
00000001 --benchmark_repetitions=128 GOPR9172.GPR
RUNNING: /home/lebedevri/rawspeed/build-clangs1-old/src/utilities/rsbench/rsbench --benchmark_counters_tabular=true --benchmark_min_time=0.
00000001 --benchmark_repetitions=128 GOPR9172.GPR --benchmark_display_aggregates_only=true --benchmark_out=/tmp/tmplwbKEM
2018-12-22 21:23:03
Running /home/lebedevri/rawspeed/build-clangs1-old/src/utilities/rsbench/rsbench
Run on (8 X 4012.81 MHz CPU s)
CPU Caches:
L1 Data 16K (x8)
L1 Instruction 64K (x4)
L2 Unified 2048K (x4)
L3 Unified 8192K (x1)
Load Average: 3.41, 2.41, 2.03
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Benchmark Time CPU Iterations CPUTime,s CPUTime/WallTime Pixels Pixels/CPUTime Pixels/WallTime Raws/CPUTime Raws/WallTime WallTime,s
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
GOPR9172.GPR/threads:8/real_time_mean 40 ms 40 ms 128 0.322244 7.96974 12M 37.4457M 298.534M 3.12047 24.8778 0.040465
GOPR9172.GPR/threads:8/real_time_median 39 ms 39 ms 128 0.312606 7.99155 12M 38.387M 306.788M 3.19891 25.5656 0.039115
GOPR9172.GPR/threads:8/real_time_stddev 4 ms 3 ms 128 0.0271557 0.130575 0 2.4941M 21.3909M 0.207842 1.78257 3.81081m
RUNNING: /home/lebedevri/rawspeed/build-clangs1-new/src/utilities/rsbench/rsbench --benchmark_counters_tabular=true --benchmark_min_time=0.
00000001 --benchmark_repetitions=128 GOPR9172.GPR --benchmark_display_aggregates_only=true --benchmark_out=/tmp/tmpWAkan9
2018-12-22 21:23:08
Running /home/lebedevri/rawspeed/build-clangs1-new/src/utilities/rsbench/rsbench
Run on (8 X 4013.1 MHz CPU s)
CPU Caches:
L1 Data 16K (x8)
L1 Instruction 64K (x4)
L2 Unified 2048K (x4)
L3 Unified 8192K (x1)
Load Average: 3.78, 2.50, 2.06
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Benchmark Time CPU Iterations CPUTime,s CPUTime/WallTime Pixels Pixels/CPUTime Pixels/WallTime Raws/CPUTime Raws/WallTime WallTime,s
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
GOPR9172.GPR/threads:8/real_time_mean 39 ms 39 ms 128 0.311533 7.97323 12M 38.6828M 308.471M 3.22356 25.706 0.0390928
GOPR9172.GPR/threads:8/real_time_median 38 ms 38 ms 128 0.304231 7.99005 12M 39.4437M 315.527M 3.28698 26.294 0.0380316
GOPR9172.GPR/threads:8/real_time_stddev 3 ms 3 ms 128 0.0229149 0.133814 0 2.26225M 19.1421M 0.188521 1.59517 3.13671m
Comparing /home/lebedevri/rawspeed/build-clangs1-old/src/utilities/rsbench/rsbench to /home/lebedevri/rawspeed/build-clangs1-new/src/utilities/rsbench/rsbench
Benchmark Time CPU Time Old Time New CPU Old CPU New
--------------------------------------------------------------------------------------------------------------------------------------
GOPR9172.GPR/threads:8/real_time_pvalue 0.0000 0.0000 U Test, Repetitions: 128 vs 128
GOPR9172.GPR/threads:8/real_time_mean -0.0339 -0.0316 40 39 40 39
GOPR9172.GPR/threads:8/real_time_median -0.0277 -0.0274 39 38 39 38
GOPR9172.GPR/threads:8/real_time_stddev -0.1769 -0.1267 4 3 3 3
```
I.e. this results in //roughly// -3% improvements in perf.
While this will help [[ https://bugs.llvm.org/show_bug.cgi?id=36419 | PR36419 ]], it won't address it fully.
Reviewers: RKSimon, craig.topper, andreadb, spatel
Reviewed By: craig.topper
Subscribers: courbet, llvm-commits
Differential Revision: https://reviews.llvm.org/D56052
llvm-svn: 351253
Peter Collingbourne [Tue, 15 Jan 2019 21:27:44 +0000 (21:27 +0000)]
compiler-rt/test: Clean up Android specific workarounds in lit.common.cfg.
-pie -Wl,--enable-new-dtags are no longer needed because
the driver passes them by default as of r316606.
Prepend -fuse-ld=gold instead of appending it so that the linker can
be overridden using COMPILER_RT_TEST_COMPILER_CFLAGS.
Differential Revision: https://reviews.llvm.org/D56697
llvm-svn: 351252
David Callahan [Tue, 15 Jan 2019 21:26:51 +0000 (21:26 +0000)]
treat invoke like call
Summary:
InvokeInst should be treated like CallInst and
assigned a separate discriminator. This is particularly
import when an Invoke is converted to a Call
during compilation and so can invalidate sample profile
data collected wtih different link time optimizations
Reviewers: twoh, Kader, danielcdh, wmi
Reviewed By: wmi
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D56491
llvm-svn: 351251
Adrian Prantl [Tue, 15 Jan 2019 21:26:03 +0000 (21:26 +0000)]
Simplify Value::GetValueByteSize()
llvm-svn: 351250
Reid Kleckner [Tue, 15 Jan 2019 21:24:55 +0000 (21:24 +0000)]
[clang-cl] Alias /Zc:alignedNew[-] to -f[no-]aligned-allocation
Implements PR40180.
clang-cl has one minor behavior difference with cl with this change.
Clang allows the user to enable the C++17 feature of aligned allocation
without enabling all of C++17, but MSVC will not call the aligned
allocation overloads unless -std:c++17 is passed. While our behavior is
technically incompatible, it would require making driver mode specific
changes to match MSVC precisely, and clang's behavior is useful because
it allows people to experiment with new C++17 features individually.
Therefore, I plan to leave it as is.
llvm-svn: 351249
Peter Collingbourne [Tue, 15 Jan 2019 21:24:00 +0000 (21:24 +0000)]
gn build: Move target flags from toolchain to a .gni file.
While here, add a use_lld flag and default it to true when using
clang on non-mac.
Differential Revision: https://reviews.llvm.org/D56710
llvm-svn: 351248
Matt Morehouse [Tue, 15 Jan 2019 21:21:01 +0000 (21:21 +0000)]
[SanitizerCoverage] Don't create comdat for interposable functions.
Summary:
Comdat groups override weak symbol behavior, allowing the linker to keep
the comdats for weak symbols in favor of comdats for strong symbols.
Fixes the issue described in:
https://bugs.chromium.org/p/chromium/issues/detail?id=918662
Reviewers: eugenis, pcc, rnk
Reviewed By: pcc, rnk
Subscribers: smeenai, rnk, bd1976llvm, hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D56516
llvm-svn: 351247
Peter Collingbourne [Tue, 15 Jan 2019 21:08:21 +0000 (21:08 +0000)]
gn build: Add build files for compiler-rt/lib/{hwasan,interception,sanitizer_common,ubsan}.
This allows the hwasan runtime to be built for Android aarch64.
Differential Revision: https://reviews.llvm.org/D56628
llvm-svn: 351246
Brad Smith [Tue, 15 Jan 2019 21:04:36 +0000 (21:04 +0000)]
Enable IAS for OpenBSD SPARC.
llvm-svn: 351245
Adrian Prantl [Tue, 15 Jan 2019 21:04:19 +0000 (21:04 +0000)]
Simplify code
llvm-svn: 351244
Adrian Prantl [Tue, 15 Jan 2019 21:04:18 +0000 (21:04 +0000)]
Add Doxygen comments.
llvm-svn: 351243
Peter Collingbourne [Tue, 15 Jan 2019 21:02:49 +0000 (21:02 +0000)]
gn build: Merge r351216, r351228.
llvm-svn: 351242
Peter Collingbourne [Tue, 15 Jan 2019 20:59:59 +0000 (20:59 +0000)]
CodeGen: Remove debug printf unintentionally added in r351228.
llvm-svn: 351241
Alexey Bataev [Tue, 15 Jan 2019 20:54:44 +0000 (20:54 +0000)]
[SLP] Added test for PR40310, NFC.
llvm-svn: 351240
Stephen Kelly [Tue, 15 Jan 2019 20:41:37 +0000 (20:41 +0000)]
Implement BlockDecl::Capture dump in terms of visitors
Reviewers: aaron.ballman
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D56709
llvm-svn: 351239
Michael Trent [Tue, 15 Jan 2019 20:41:30 +0000 (20:41 +0000)]
llvm-objdump -m -D should disassemble all text segments
Summary:
When running llvm-objdump with the -macho option objdump will by default
disassemble only the __TEXT,__text section (or __TEXT_EXEC,__text when
disassembling MH_KEXT_BUNDLE files). The -disassemble-all option is
treated no diferently than -disassemble.
This change upates llvm-objdump's MachO parsing code to disassemble all
__text sections found in a file when -disassemble-all is specified. This
is useful for disassembling files with more than one __text section, or
when disassembling files whose __text section is not present in __TEXT.
I added a lit test case that verifies "llvm-objdump -m -d" and
"llvm-objdump -m -D" produce the expected results on a reference binary.
I also updated the CommandGuide documentation for llvm-objdump.rst and
verified it renders correctly as man and html.
rdar://
42899338
Reviewers: ab, pete, lhames
Reviewed By: lhames
Subscribers: rupprecht, llvm-commits
Differential Revision: https://reviews.llvm.org/D56649
llvm-svn: 351238
Adrian Prantl [Tue, 15 Jan 2019 20:33:58 +0000 (20:33 +0000)]
Replace auto -> llvm::Optional<uint64_t>
This addresses post-commit feedback for https://reviews.llvm.org/D56688
llvm-svn: 351237
Stephen Kelly [Tue, 15 Jan 2019 20:31:31 +0000 (20:31 +0000)]
NFC: Implement OMPClause dump in terms of visitors
Reviewers: aaron.ballman
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D56708
llvm-svn: 351236
Stephen Kelly [Tue, 15 Jan 2019 20:17:33 +0000 (20:17 +0000)]
Implement CXXCtorInitializer dump in terms of Visitor
Reviewers: aaron.ballman
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D56707
llvm-svn: 351235
Craig Topper [Tue, 15 Jan 2019 20:12:33 +0000 (20:12 +0000)]
[X86] Add versions of the avx512 gather intrinsics that take the mask as a vXi1 vector instead of a scalar
In keeping with our general direction of having the vXi1 type present in IR, this patch converts the mask argument for avx512 gather to vXi1. This can avoid k-register to GPR to k-register transitions late in codegen.
I left the existing intrinsics behind because they have many out of tree users such as ISPC. They generate their own code and don't go through the autoupgrade path which only works for bitcode and ll parsing. Ideally we will get them to migrate to target independent intrinsics, but it might be easier for them to migrate to these new intrinsics.
I'll work on scatter and gatherpf/scatterpf next.
Differential Revision: https://reviews.llvm.org/D56527
llvm-svn: 351234
Anton Korobeynikov [Tue, 15 Jan 2019 20:10:46 +0000 (20:10 +0000)]
[MSP430] Recognize '{' as a line separator
msp430-as supports multiple assembly statements on the same line
separated by a '{' character.
llvm-svn: 351233
Volodymyr Sapsai [Tue, 15 Jan 2019 20:08:23 +0000 (20:08 +0000)]
[MSVC Compat] Fix typo correction for inclusion directives.
In MSVC compatibility mode we were checking not the typo corrected
filename but the original filename.
Reviewers: christylee, compnerd
Reviewed By: christylee
Subscribers: jkorous, dexonsmith, sammccall, hokein, cfe-commits
Differential Revision: https://reviews.llvm.org/D56631
llvm-svn: 351232
Craig Topper [Tue, 15 Jan 2019 19:59:19 +0000 (19:59 +0000)]
[Nios2] Remove Nios2 backend
As mentioned here http://lists.llvm.org/pipermail/llvm-dev/2019-January/129121.html This backend is incomplete and has not been maintained in several months.
Differential Revision: https://reviews.llvm.org/D56691
llvm-svn: 351231
Craig Topper [Tue, 15 Jan 2019 19:58:36 +0000 (19:58 +0000)]
[Nios2] Remove Nios2 backend
As mentioned here http://lists.llvm.org/pipermail/llvm-dev/2019-January/129121.html This backend is incomplete and has not been maintained in several months.
Differential Revision: https://reviews.llvm.org/D56690
llvm-svn: 351230
Haojian Wu [Tue, 15 Jan 2019 19:51:39 +0000 (19:51 +0000)]
[Tooling] Fix broken compliation databse tests.
I forgot to update the unittest in r351222.
llvm-svn: 351229
Anton Korobeynikov [Tue, 15 Jan 2019 19:44:05 +0000 (19:44 +0000)]
[MSP430] Provide a toolchain description
This is an initial implementation for msp430 toolchain including
-mmcu option support
-mhwmult options support
-integrated-as by default
The toolchain uses msp430-elf-as as a linker and supports msp430-gcc toolchain tree.
Patch by Kristina Bessonova!
Differential Revision: https://reviews.llvm.org/D56658
llvm-svn: 351228
Jonathan Peyton [Tue, 15 Jan 2019 19:39:32 +0000 (19:39 +0000)]
[OpenMP] Fix for nested proc_bind affinity bug
Using proc_bind clause on a nested #pragma omp parallel region
with KMP_AFFINITY set causes an assertion error. This assertion occurs because
the place-partition-var is not properly initialized in the nested master threads.
Trying to get an intuitive result with KMP_AFFINITY + proc_bind is difficult
because of how the KMP_AFFINITY gtid-to-place mapping occurs. This
patch creates an initial place list no matter what affinity mechanism is used.
For KMP_AFFINITY, the place-partition-var is initialized to all the places.
Differential Revision: https://reviews.llvm.org/D55795
llvm-svn: 351227
Dan Albert [Tue, 15 Jan 2019 19:16:25 +0000 (19:16 +0000)]
Fix size_t/off_t mixup in std::filesystem.
Summary: ftruncate takes an off_t, not a size_t.
Reviewers: EricWF, mclow.lists
Reviewed By: EricWF
Subscribers: christof, ldionne, libcxx-commits
Differential Revision: https://reviews.llvm.org/D56578
llvm-svn: 351226
Dan Albert [Tue, 15 Jan 2019 19:14:15 +0000 (19:14 +0000)]
Add large file support to create_file for 32-bit.
Summary:
The tests need to create files larger than 2GB, but size_t is 32-bit
on a 32-bit system. Make use of explicit off64_t APIs so we can still
use a default off_t for the tests while enabling 64-bit file offsets
for create_file.
Reviewers: mclow.lists, EricWF
Reviewed By: EricWF
Subscribers: christof, ldionne, libcxx-commits
Differential Revision: https://reviews.llvm.org/D56619
llvm-svn: 351225
Jonathan Peyton [Tue, 15 Jan 2019 19:14:00 +0000 (19:14 +0000)]
[OpenMP] Add lock function definitions to fix Bug 40042
This change fixes the sanity issue reported in Bug 40042.
Lock function definitions for the three lock kinds were added
to disambiguate calls to the lock functions done directly and indirectly.
Bugzilla: https://bugs.llvm.org/show_bug.cgi?id=40042
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D56103
llvm-svn: 351224
Jonathan Peyton [Tue, 15 Jan 2019 19:08:26 +0000 (19:08 +0000)]
[OpenMP][Cmake] Allowed OpenMP testing detect test compiler with same generator
Fix ninja build detect test compiler failed under windows.
Patch by Peiyuan Song
Differential Revision: https://reviews.llvm.org/D53479
llvm-svn: 351223
Haojian Wu [Tue, 15 Jan 2019 19:05:50 +0000 (19:05 +0000)]
[Tooling] Make clang-tool find libc++ dir on mac when running on a file without compilation database.
Summary:
This is a regression of r348365.
When clang-tools run on a file without a complation database (`clang-check /tmp/t.cc`),
we will use fixed compilation database as a fallback. However the actual compiler
path in the fallback complation command is just `clang-tool` which is
insufficient to detect the libc++ dir.
Reviewers: ilya-biryukov, EricWF
Reviewed By: ilya-biryukov
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D56680
llvm-svn: 351222
Jonathan Peyton [Tue, 15 Jan 2019 18:57:24 +0000 (18:57 +0000)]
[OpenMP] Fix performance regression in SPEC kdtree test
Make __ompt_implicit_task_end a static function and remove the inline part. Remove
pId variable that is unused. This fixes small regression in SPEC kdtree benchmark.
Also reformat some of __ompt_implicit_task_end.
Differential Revision: https://reviews.llvm.org/D55788
llvm-svn: 351221
Petr Hosek [Tue, 15 Jan 2019 18:55:55 +0000 (18:55 +0000)]
[libc++] Support different libc++ namespaces in the iterator test
libc++ allows changing the namespace, don't assume __1 in the test
to avoid the test failure if different namespace is being used.
Differential Revision: https://reviews.llvm.org/D56698
llvm-svn: 351220
Nikita Popov [Tue, 15 Jan 2019 18:43:41 +0000 (18:43 +0000)]
Reapply "[CodeGen][X86] Expand USUBSAT to UMAX+SUB, also for vectors"
Related to https://bugs.llvm.org/show_bug.cgi?id=40123.
Rather than scalarizing, expand a vector USUBSAT into UMAX+SUB,
which produces much better code for X86.
Reapplying with updated SLPVectorizer tests.
Differential Revision: https://reviews.llvm.org/D56636
llvm-svn: 351219
Rui Ueyama [Tue, 15 Jan 2019 18:30:23 +0000 (18:30 +0000)]
Do not emit a corrupt symbol table entry for .rela_iplt_{start,end}.
If .rela.iplt does not exist, we used to emit a corrupt symbol table
that contains two symbols, .rela_iplt_{start,end}, pointing to a
nonexisting section.
This patch fixes the issue by setting section index 0 to the symbols
if .rel.iplt section does not exist.
Differential Revision: https://reviews.llvm.org/D56623
llvm-svn: 351218
Brad Smith [Tue, 15 Jan 2019 18:24:03 +0000 (18:24 +0000)]
[Solaris] Move enabling IAS for SPARC from the Solaris toolchain to Generic_GCC.
llvm-svn: 351217
Yury Delendik [Tue, 15 Jan 2019 18:14:12 +0000 (18:14 +0000)]
[WebAssembly] Fix updating/moving DBG_VALUEs in RegStackify
Summary:
As described in PR40209, there can be issues in DBG_VALUEs handling when multiple defs present in a BB. This patch
adds logic for detection of related to def DBG_VALUEs and localizes register update and movement to found DBG_VALUEs.
Reviewers: aheejin
Subscribers: mgorny, dschuff, sbc100, jgravelle-google, sunfish, llvm-commits
Differential Revision: https://reviews.llvm.org/D56401
llvm-svn: 351216
Adrian Prantl [Tue, 15 Jan 2019 18:07:54 +0000 (18:07 +0000)]
Silence compiler warnings
llvm-svn: 351215
Adrian Prantl [Tue, 15 Jan 2019 18:07:52 +0000 (18:07 +0000)]
Make CompilerType::getBitSize() / getByteSize() return an optional result. NFC
The code in LLDB assumes that CompilerType and friends use the size 0
as a sentinel value to signal an error. This works for C++, where no
zero-sized type exists, but in many other programming languages
(including I believe C) types of size zero are possible and even
common. This is a particular pain point in swift-lldb, where extra
code exists to double-check that a type is *really* of size zero and
not an error at various locations.
To remedy this situation, this patch starts by converting
CompilerType::getBitSize() and getByteSize() to return an optional
result. To avoid wasting space, I hand-rolled my own optional data
type assuming that no type is larger than what fits into 63
bits. Follow-up patches would make similar changes to the ValueObject
hierarchy.
rdar://problem/
47178964
Differential Revision: https://reviews.llvm.org/D56688
llvm-svn: 351214