Daniel Sanders [Tue, 31 Oct 2017 21:54:52 +0000 (21:54 +0000)]
Revert r317040: [globalisel][tablegen] Keep track of the insertion point while adding BuildMIAction's. NFC
The same bots fail but I believe I know what the issue is now. These bots are
missing the const_iterator versions of insert/emplace/etc. that were introduced
in C++11.
llvm-svn: 317042
Reid Kleckner [Tue, 31 Oct 2017 21:52:15 +0000 (21:52 +0000)]
[codeview] Merge file checksum entries for DIFiles with the same absolute path
Change the map key from DIFile* to the absolute path string. Computing
the absolute path isn't expensive because we already have a map that
caches the full path keyed on DIFile*.
llvm-svn: 317041
Daniel Sanders [Tue, 31 Oct 2017 21:34:53 +0000 (21:34 +0000)]
Re-commit: [globalisel][tablegen] Keep track of the insertion point while adding BuildMIAction's. NFC
Multi-instruction emission needs to ensure the the instructions are generated
a depth-first fashion. For example:
(ADDWrr (SUBWrr a, b), c)
needs to emit the SUBWrr before the ADDWrr. However, our walk over
TreePatternNode's is highly context sensitive which makes it difficult to append
BuildMIActions in the order we want. To fix this, we now keep track of the
insertion point as we add actions. This will allow multi-insn emission to insert
BuildMI's in the correct place.
The previous commit failed on the Ubuntu bots using GCC 4.8. These bots didn't
like a call to emplace(). I've replaced it with insert() to see if it's a quirk
of the C++11 support.
llvm-svn: 317040
Rui Ueyama [Tue, 31 Oct 2017 21:26:42 +0000 (21:26 +0000)]
Do not access beyond the end of local symbols.
This patch resurrects code that was removed in r317007 to not access
beyond allocated memory for a symbol.
Differential Revision: https://reviews.llvm.org/D39458
llvm-svn: 317039
Marek Olsak [Tue, 31 Oct 2017 21:06:42 +0000 (21:06 +0000)]
AMDGPU: Select s_buffer_load_dword with a non-constant SGPR offset
Summary:
Apps that benefit:
- alien isolation
- bioshock infinite
- civilization: beyond earth
- company of heroes 2
- dirt showdown
- dota 2
- F1 2015
- grid autosport
- hitman
- legend of grimrock
- serious sam 3: bfe
- shadow warrior
- talos principle
- total war: warhammer
- UE4 demos: effects cave, elemental, sun temple
Reviewers: arsenm, nhaehnle
Subscribers: kzhuravl, wdng, yaxunl, dstuttard, tpr, llvm-commits, t-tye
Differential Revision: https://reviews.llvm.org/D38914
llvm-svn: 317038
Adrian Prantl [Tue, 31 Oct 2017 21:03:22 +0000 (21:03 +0000)]
loop-rotate: simplify code by using llvm::findDbgValues(). (NFC)
llvm-svn: 317037
Vitaly Buka [Tue, 31 Oct 2017 20:50:07 +0000 (20:50 +0000)]
[fuzzer] Script to detect unbalanced allocation in -trace_malloc output
Reviewers: kcc
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D39466
llvm-svn: 317036
Vitaly Buka [Tue, 31 Oct 2017 20:49:57 +0000 (20:49 +0000)]
[clang-fuzzer] Fix incremental builds of the fuzzer
Summary:
Don't use BUILD_IN_SOURCE keep git checkout clean
Don't forward CMAKE_GENERATOR as ExternalProject_Add should do it already
Reset UPDATE_COMMAND to avoid git checkout updates on each build
Reviewers: kcc, morehouse
Subscribers: cfe-commits, mgorny
Differential Revision: https://reviews.llvm.org/D39445
llvm-svn: 317035
Vitaly Buka [Tue, 31 Oct 2017 20:49:48 +0000 (20:49 +0000)]
[fuzzer] Fix threaded stack printing and nested mallocs
Summary: Nested mallocs are possible with internal symbolizer.
Reviewers: kcc
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D39397
llvm-svn: 317034
Daniel Sanders [Tue, 31 Oct 2017 20:29:28 +0000 (20:29 +0000)]
Revert r317029: [globalisel][tablegen] Keep track of the insertion point while adding BuildMIAction's. NFC
The Linux bots don't seem to like this usage of emplace(). Reverting while I look into it.
llvm-svn: 317033
Richard Smith [Tue, 31 Oct 2017 20:29:22 +0000 (20:29 +0000)]
Fix usage of right shift operator in fold expressions
The right shift operator was not seen as a valid operator in a fold expression, which is PR32563.
Patch by Nicolas Lesser ("Blitz Rakete")!
llvm-svn: 317032
Sanjay Patel [Tue, 31 Oct 2017 20:19:39 +0000 (20:19 +0000)]
[CodeGen] map sqrt libcalls to llvm.sqrt when errno is not set
The LLVM sqrt intrinsic definition changed with:
D28797
...so we don't have to use any relaxed FP settings other than errno handling.
This patch sidesteps a question raised in PR27435:
https://bugs.llvm.org/show_bug.cgi?id=27435
Is a programmer using __builtin_sqrt() invoking the compiler's intrinsic definition of sqrt or the mathlib definition of sqrt?
But we have an answer now: the builtin should match the behavior of the libm function including errno handling.
Differential Revision: https://reviews.llvm.org/D39204
llvm-svn: 317031
Benjamin Kramer [Tue, 31 Oct 2017 19:55:08 +0000 (19:55 +0000)]
Revert "[DWARF] Now that Optional is standard layout, put it into an union instead of splatting it."
GCC doesn't like it. This reverts commit r317028.
llvm-svn: 317030
Daniel Sanders [Tue, 31 Oct 2017 19:54:05 +0000 (19:54 +0000)]
[globalisel][tablegen] Keep track of the insertion point while adding BuildMIAction's. NFC
Multi-instruction emission needs to ensure the the instructions are generated
a depth-first fashion. For example:
(ADDWrr (SUBWrr a, b), c)
needs to emit the SUBWrr before the ADDWrr. However, our walk over
TreePatternNode's is highly context sensitive which makes it difficult to append
BuildMIActions in the order we want. To fix this, we now keep track of the
insertion point as we add actions. This will allow multi-insn emission to insert
BuildMI's in the correct place.
llvm-svn: 317029
Benjamin Kramer [Tue, 31 Oct 2017 19:40:03 +0000 (19:40 +0000)]
[DWARF] Now that Optional is standard layout, put it into an union instead of splatting it.
No functionality change intended.
llvm-svn: 317028
Benjamin Kramer [Tue, 31 Oct 2017 19:22:55 +0000 (19:22 +0000)]
[coro] Make Spill a proper struct instead of deriving from pair.
No functionality change.
llvm-svn: 317027
Rui Ueyama [Tue, 31 Oct 2017 19:14:06 +0000 (19:14 +0000)]
Revert r316305: Remove a fast lookup table from MergeInputSection.
This reverts commit r316305 because performance regression was observed.
llvm-svn: 317026
Daniel Sanders [Tue, 31 Oct 2017 19:09:29 +0000 (19:09 +0000)]
[globalisel][tablegen] Factor out implicit def/use renderers from createAndImportInstructionRenderer(). NFC
Multi-instruction emission will require that we have separate handling for
the defs between the implicitly created temporaries and the rule outputs.
The former require new temporary vregs while the latter should copy existing
operands. Factor out the implicit def/use renderers to minimize the code
duplication when we implement that.
llvm-svn: 317025
Craig Topper [Tue, 31 Oct 2017 19:03:51 +0000 (19:03 +0000)]
[SimplifyCFG] Use a more generic name for the selects created by SpeculativelyExecuteBB to prevent long names from being created
Currently the selects are created with the names of their inputs concatenated together. It's possible to get cases that chain these selects together resulting in long names due to multiple levels of concatenation. Our internal branch of llvm managed to generate names over 100000 characters in length on a particular test due to an extreme compounding of the names.
This patch changes the name to a generic name that is not dependent on its inputs.
Differential Revision: https://reviews.llvm.org/D39440
llvm-svn: 317024
Craig Topper [Tue, 31 Oct 2017 19:03:49 +0000 (19:03 +0000)]
[SimplifyCFG] Regenerate some test cases using update_test_checks.py to prepare for an upcoming commit. NFC
A future commit will change how some of the value names in the IR are generated which causes these tests to break in their current form. The script generates checks with regular expressions so it should be immune.
llvm-svn: 317023
Daniel Sanders [Tue, 31 Oct 2017 18:50:24 +0000 (18:50 +0000)]
[globalisel][tablegen] Add infrastructure to potentially allow BuildMIAction to choose a mutatable instruction. NFC
Prepare for multiple instruction emission by allowing BuildMIAction to
search for a suitable matcher that will support mutation.
This patch deliberately neglects to add matchers aside from the root to
preserve NFC. That said, it should be noted that until we support mutations
other than just the opcode the chances of finding a non-root instruction
for which canMutate() is true, is essentially zero. Furthermore in the
presence of multi-instruction emission the chances of finding any
instruction for which canMutate() is true is also zero. Nevertheless, we
can't continue to require that all BuildMIAction's consider the root of the match
to be recyclable due to the risk of recycling it twice in the same rule.
llvm-svn: 317022
Simon Pilgrim [Tue, 31 Oct 2017 18:43:24 +0000 (18:43 +0000)]
[X86][AVX512] Regenerate tests to remove retl/retq regex
These are only testing 64-bit targets so we don't need the regex
llvm-svn: 317021
Simon Pilgrim [Tue, 31 Oct 2017 18:41:48 +0000 (18:41 +0000)]
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests
llvm-svn: 317020
Benjamin Kramer [Tue, 31 Oct 2017 18:35:54 +0000 (18:35 +0000)]
[ADT] Split optional to only include copy mechanics and dtor for non-trivial types.
This makes uses of Optional more transparent to the compiler (and
clang-tidy) and generates slightly smaller code.
llvm-svn: 317019
Wolfgang Pieb [Tue, 31 Oct 2017 18:25:28 +0000 (18:25 +0000)]
[Metadata][NFC] Make MDNode::resolve() public in preparation for the fix to PR33930.
Reviewers: aprantl
llvm-svn: 317018
Daniel Sanders [Tue, 31 Oct 2017 18:07:03 +0000 (18:07 +0000)]
[globalisel][tablegen] Allow any comment in DebugCommentAction. NFC
llvm-svn: 317017
Philip Reames [Tue, 31 Oct 2017 18:04:57 +0000 (18:04 +0000)]
[IndVarSimplify] Extract wrapper around SE-.isLoopInvariantPredicate [NFC]
This an intermediate state, the next patch will re-inline the markLoopInvariantPredicate function to reduce code duplication.
llvm-svn: 317016
Rui Ueyama [Tue, 31 Oct 2017 17:37:20 +0000 (17:37 +0000)]
[Support] Make the default chunk size of raw_fd_ostream to 1 GiB.
Previously, we call write(2) for each 32767 byte chunk. That is not
efficient because Linux can handle much larger write requests.
This patch changes the chunk size on Linux to 1 GiB.
This patch also changes the default chunks size to SSIZE_MAX. I think
that doesn't in practice change this function's behavior on any operating
system because SSIZE_MAX on 64-bit machine is unrealistically large,
and writing 2 GiB (SSIZE_MAX on 32-bit) on a 32-bit machine by a single
call of write(2) is also unrealistic, as the userspace is usually
limited to 2 GiB. That said, it is in general a good thing to do because
a write larger than SSIZE_MAX is implementation-defined in POSIX.
Differential Revision: https://reviews.llvm.org/D39444
llvm-svn: 317015
James Henderson [Tue, 31 Oct 2017 17:28:44 +0000 (17:28 +0000)]
[ELF] Remove unnecessary function call. NFC.
StringTableSection does not implement postThunkContents, so calling it on these
sections is pointless (it calls an empty virtual function), and we can remove it.
Reviewers: ruiu
Differential Revision: https://reviews.llvm.org/D39460
llvm-svn: 317014
Rui Ueyama [Tue, 31 Oct 2017 17:07:47 +0000 (17:07 +0000)]
Remove a redundant union member.
This removes DefinedCOFF from SymbolUnion because DefinedCOFF is
not a leaf class. Pointed out by pcc.
llvm-svn: 317013
Philip Reames [Tue, 31 Oct 2017 17:06:32 +0000 (17:06 +0000)]
[IndVarSimplify] Simplify code using a dictionary
Possibly very slightly slower, but this code is not performance critical and the readability benefit alone is huge.
llvm-svn: 317012
Reid Kleckner [Tue, 31 Oct 2017 16:47:38 +0000 (16:47 +0000)]
[X86][AsmParser] Treat '%' as the modulo operator under Intel syntax
It can't be a register prefix, anyway. This is consistent with the masm
docs on MSDN: https://msdn.microsoft.com/en-us/library/t4ax90d2.aspx
This is a straight-forward extension of our support for "MOD"
implemented in https://reviews.llvm.org/D33876 / r306425
llvm-svn: 317011
Nico Weber [Tue, 31 Oct 2017 16:39:47 +0000 (16:39 +0000)]
LTOModule::isBitcodeFile() shouldn't assert when returning false.
Fixes a bunch of assert-on-invalid-bitcode regressions after 315483.
Expected<> calls assertIsChecked() in its dtor, and operator bool() only calls
setChecked() if there's no error. So for functions that don't return an error
itself, the Expected<> version needs explicit code to disarm the error that the
ErrorOr<> code didn't need.
https://reviews.llvm.org/D39437
llvm-svn: 317010
Reid Kleckner [Tue, 31 Oct 2017 16:16:08 +0000 (16:16 +0000)]
[asan] Upgrade private linkage globals to internal linkage on COFF
COFF comdats require symbol table entries, which means the comdat leader
cannot have private linkage.
llvm-svn: 317009
Michael Kruse [Tue, 31 Oct 2017 16:11:46 +0000 (16:11 +0000)]
[ZoneAlgo/ForwardOpTree] Normalize PHIs to their known incoming values.
Represent PHIs by their incoming values instead of an opaque value of
themselves. This allows ForwardOpTree to "look through" the PHIs and
forward the incoming values since forwardings PHIs is currently not
supported.
This is particularly useful to cope with PHIs inserted by GVN LoadPRE.
The incoming values all resolve to a load from a single array element
which then can be forwarded.
It should in theory also reduce spurious conflicts in value mapping
(DeLICM), but I have not yet found a profitable case yet, so it is
not included here.
To avoid transitive closure and potentially necessary overapproximations
of those, PHIs that may reference themselves are excluded from
normalization and keep their opaque self-representation.
Differential Revision: https://reviews.llvm.org/D39333
llvm-svn: 317008
Rui Ueyama [Tue, 31 Oct 2017 16:10:24 +0000 (16:10 +0000)]
[COFF] Merge Symbol and SymbolBody.
llvm-svn: 317007
Rui Ueyama [Tue, 31 Oct 2017 16:07:41 +0000 (16:07 +0000)]
Merge SymbolBody and Symbol into one class, SymbolBody.
SymbolBody and Symbol were separated classes due to a historical reason.
Symbol used to be a pointer to a SymbolBody, and the relationship
between Symbol and SymbolBody was n:1.
r2681780 changed that. Since that patch, SymbolBody and Symbol are
allocated next to each other to improve memory locality, and they have
1:1 relationship now. So, the separation of Symbol and SymbolBody no
longer makes sense.
This patch merges them into one class. In order to avoid updating too
many places, I chose SymbolBody as a unified name. I'll rename it Symbol
in a follow-up patch.
Differential Revision: https://reviews.llvm.org/D39406
llvm-svn: 317006
Simon Pilgrim [Tue, 31 Oct 2017 16:06:21 +0000 (16:06 +0000)]
[X86][SSE] Add VSRLI/VSRAI/VSLLI demanded elts support to computeKnownBits/ComputeNumSignBits
Mainly a perf improvements as most combines will have occurred before we lower to these instructions
llvm-svn: 317005
Pavel Labath [Tue, 31 Oct 2017 15:27:19 +0000 (15:27 +0000)]
Add a "watchpoint" test category and annotate tests appropriately
Most of the watchpoint tests are organized into subtrees, so we can use the
file-based .categories approach to annotate them. The exception are the
concurrent_events tests, which needed to be annotated on a per-test basis.
The motivation behind this is to provide an easy way to disable watchpoint
tests on systems where the watchpoint functionality is not present/unreliable.
llvm-svn: 317004
Benjamin Kramer [Tue, 31 Oct 2017 14:58:22 +0000 (14:58 +0000)]
[LoopVectorize] Replace manual VPlan memory management with unique_ptr.
No functionality change intended.
llvm-svn: 317003
George Rimar [Tue, 31 Oct 2017 14:21:30 +0000 (14:21 +0000)]
[ELF] - Remove useless code. NFC.
We set Type and Flags inside OutputSection::addSection,
so this lines looks to be excessive.
llvm-svn: 317002
Jonas Devlieghere [Tue, 31 Oct 2017 14:19:02 +0000 (14:19 +0000)]
[test] Fix dsymutil/cmdline.test
This fixes dsymutil/cmdline.test on platforms where the dsymutil binary
has an extension.
llvm-svn: 317001
Florian Hahn [Tue, 31 Oct 2017 14:06:31 +0000 (14:06 +0000)]
[Reassociate] Remove FIXME from looptest.ll (NFC)
Summary: The loop invariant add (i+j) is reassoicated, I think the FIXME can be removed, because this is what the test case tries to check (AFAIK). I also changed the test to use FileCheck.
Reviewers: mcrosier, davide
Reviewed By: mcrosier, davide
Subscribers: davide, llvm-commits
Differential Revision: https://reviews.llvm.org/D39424
llvm-svn: 317000
Jonas Devlieghere [Tue, 31 Oct 2017 13:54:15 +0000 (13:54 +0000)]
[dsymutil] Implement the --threads option
This patch adds the --threads option to dsymutil to process
architectures in parallel. The feature is already present in the version
distributed with Xcode, but was not yet upstreamed.
This is NFC as far as the linking behavior is concerned. As threads are
used automatically, the current tests cover the change in
implementation.
Differential revision: https://reviews.llvm.org/D39355
llvm-svn: 316999
Oleg Ranevskyy [Tue, 31 Oct 2017 13:51:06 +0000 (13:51 +0000)]
[lld] Fix --exclude-libs broken when --whole-archive is used
Summary:
**Problem**
`--exclude-libs` does not work for static libraries affected by the `--whole-archive` option.
**Description**
`--exclude-libs` creates a list of static library paths and does library lookups in this list.
`--whole-archive` splits the static libraries that follow it into separate objects. As a result, lld no longer sees static libraries among linked files and does no `--exclude-libs` lookups.
**Solution**
The proposed solution is to make `--exclude-libs` consider object files too. When lld finds an object file it checks whether this file originates from an archive and, if so, looks the archive up in the `--exclude-libs` list.
Reviewers: ruiu, rafael
Reviewed By: ruiu
Subscribers: asl, ikudrin, llvm-commits, emaste
Tags: #lld
Differential Revision: https://reviews.llvm.org/D39353
llvm-svn: 316998
Pavel Labath [Tue, 31 Oct 2017 13:23:19 +0000 (13:23 +0000)]
Fix LLVM_LINK_LLVM_DYLIB build (pr35053)
Summary:
r316368 broke this build when it introduced a reference to a pthread
function to the Utility module. This caused cmake to generate an
incorrect link line (wrong order of libs) because it did not see the
dependency from Utility to the system libraries. Instead these libraries
were being manually added to each final target.
This changes moves the dependency management from the individual targets
to the lldbUtility module, which is consistent with how llvm does it.
The final targets will pick up these libraries as they will be a part of
the link interface of the module.
Technically, some of these dependencies could go into the host module,
as that's where most of the os-specific code is, but I did not try to
investigate which ones.
Reviewers: zturner, sylvestre.ledru
Subscribers: lldb-commits, mgorny
Differential Revision: https://reviews.llvm.org/D39246
llvm-svn: 316997
Teresa Johnson [Tue, 31 Oct 2017 12:56:09 +0000 (12:56 +0000)]
[ThinLTO] Double bits of module hash used for renaming
Summary:
Use 64 instead of 32 bits of the module hash as the suffix when renaming
after promotion to reduce the likelihood of a collision (which we
observed in a binary when using 32 bits).
Reviewers: pcc
Subscribers: llvm-commits, inglorion
Differential Revision: https://reviews.llvm.org/D39443
llvm-svn: 316996
Michael Kruse [Tue, 31 Oct 2017 12:50:25 +0000 (12:50 +0000)]
[DeLICM] Fix wrong assumed access execution order.
ForwardOpTree may already transform a scalar access to an array
accesses. The access remains implicit (isOriginalScalarKind(), meaning
that the access is always executed at the begin/end of a statement), but
targets an array (isLatestArrayKind(), which is unrelated to whether the
execution is implicit/explicit).
Fix by properly using isOriginalXXX() to determine execution order.
This fixes the buildbots on MultiSource/Benchmarks/DOE-ProxyApps-C/miniGMG.
llvm-svn: 316995
Matthew Simpson [Tue, 31 Oct 2017 12:34:02 +0000 (12:34 +0000)]
[InstCombine] Simplify selects that test cmpxchg instructions
If a select instruction tests the returned flag of a cmpxchg instruction and
selects between the returned value of the cmpxchg instruction and its compare
operand, the result of the select will always be equal to its false value.
Differential Revision: https://reviews.llvm.org/D39383
llvm-svn: 316994
Pavel Labath [Tue, 31 Oct 2017 12:27:46 +0000 (12:27 +0000)]
Increase AdbClient read timeout
The previous value was not sufficient for Pixel 2 phones. One would have
hoped that the newer phones are faster, but that does not seem to be the
case here.
llvm-svn: 316993
Pavel Labath [Tue, 31 Oct 2017 12:27:43 +0000 (12:27 +0000)]
Add data formatter for libc++'s forward_list
Summary:
This adds a data formatter for the implementation of forward_list in
libc++. I've refactored the existing std::list data formatter a bit to
enable more sharing of code (mainly the loop detection stuff).
Reviewers: jingham, EricWF
Subscribers: srhines, eugene, lldb-commits
Differential Revision: https://reviews.llvm.org/D35556
llvm-svn: 316992
Pavel Labath [Tue, 31 Oct 2017 11:48:33 +0000 (11:48 +0000)]
Fix mac build broken in r316987
Forgot one occurence of ArchSpec::SetTriple in mac-specific code.
llvm-svn: 316990
Ayman Musa [Tue, 31 Oct 2017 11:39:31 +0000 (11:39 +0000)]
Adding a shufflevector and select LLVM IR instructions fuzz tool
Based on similar python tool - utils/shuffle-fuzz.py - this tool extends the ability of it's previous by optionally attaching select instruction to the generated shufflevector instructions.
This was mainly developed to perform exhaustive testing of the X86 AVX512 masked shuffle instructions. But yet it can be used for various other targets.
The general design of the implementation is much modular than the original shuffle_fuzz.py tool, which makes it easier for anyone to extend it further.
Differential Revision: https://reviews.llvm.org/D38031
Change-Id: I0efc2aaa091b61a8a9552311c21cc77916a97111
llvm-svn: 316989
Ivan A. Kosarev [Tue, 31 Oct 2017 11:05:34 +0000 (11:05 +0000)]
[CodeGen] Propagate may-alias'ness of lvalues with TBAA info
This patch fixes various places in clang to propagate may-alias
TBAA access descriptors during construction of lvalues, thus
eliminating the need for the LValueBaseInfo::MayAlias flag.
This is part of D38126 reworked to be a separate patch to
simplify review.
Differential Revision: https://reviews.llvm.org/D39008
llvm-svn: 316988
Pavel Labath [Tue, 31 Oct 2017 10:56:03 +0000 (10:56 +0000)]
Invert ArchSpec<->Platform dependency
Summary:
ArchSpec::SetTriple was taking a Platform as an argument, and used it to
fill in missing pieces of the specified triple. I invert the dependency
by moving this code to other classes. For this purpose, I've created
three new functions.
- HostInfo::GetAugmentedArchSpec: fills in the triple using the host
platform (this used to be implemented by passing a null platform
pointer). By putting this code in the Host module, we can provide a
way to anyone who does not have a platform instance (lldb-server) an
easy way to get Host data.
- Platform::GetAugmentedArchSpec: if you have a platform instance, you
can call this to let it fill in the triple.
- static Platform::GetAugmentedArchSpec: implements the "if platform ==
0 then use_host() else use_platform()" part.
Reviewers: zturner, jingham, clayborg
Subscribers: mgorny, javed.absar, lldb-commits
Differential Revision: https://reviews.llvm.org/D39387
llvm-svn: 316987
David Green [Tue, 31 Oct 2017 10:47:46 +0000 (10:47 +0000)]
[LoopUnroll] Clean up remarks for unroll remainder
The optimisation remarks for loop unrolling with an unrolled remainder looks something like:
test.c:7:18: remark: completely unrolled loop with 3 iterations [-Rpass=loop-unroll]
C[i] += A[i*N+j];
^
test.c:6:9: remark: unrolled loop by a factor of 4 with run-time trip count [-Rpass=loop-unroll]
for(int j = 0; j < N; j++)
^
This removes the first of the two messages.
Differential revision: https://reviews.llvm.org/D38725
llvm-svn: 316986
Pavel Labath [Tue, 31 Oct 2017 10:33:03 +0000 (10:33 +0000)]
Android.rules: build with "unified android headers"
Unified headers will be the only way to build applications in NDK r16,
and it also works with NDK r15.
This also bumps the minimum supported android version to 16.
llvm-svn: 316985
George Rimar [Tue, 31 Oct 2017 10:31:58 +0000 (10:31 +0000)]
[ELF] - Simplify output section creation.
When there is no SECTION commands given, all sections are
technically orphans, but now we handle script orphans sections
and regular "orphans" sections for non-scripted case differently,
though we can handle them at one place.
Patch do that change.
Differential revision: https://reviews.llvm.org/D39045
llvm-svn: 316984
Michael Kruse [Tue, 31 Oct 2017 10:28:22 +0000 (10:28 +0000)]
[OpenMP] Fix reference collection of latest base ptrs.
When collecting base pointers that need to be made available in parallel
subfunctions, use the base pointer associated with the latest
ScopArrayInfo, instead of the original one.
llvm-svn: 316983
Pavel Labath [Tue, 31 Oct 2017 10:01:30 +0000 (10:01 +0000)]
Split makefile for TestTopLevelExprs
this test was using a single makefile to build two executables. This
setup, although not supported by Makefile.rules, happened to
work in most configurations, except when building with the android ndk
r16.
Here I move the building of the second executable to a separate
makefile, which is the solution other tests use for multiple targets.
llvm-svn: 316982
Michael Zuckerman [Tue, 31 Oct 2017 10:00:19 +0000 (10:00 +0000)]
[AVX512] Adding new patterns for extract_subvector of vXi1
extract subvector of vXi1 from vYi1 is poorly supported by LLVM and most of the time end with an assertion.
This patch fixes this issue by adding new patterns to the TD file.
Reviewers:
1. guyblank
2. igorb
3. zvi
4. ayman
5. craig.topper
Differential Revision: https://reviews.llvm.org/D39292
Change-Id: Ideb4d7e946c8d40cfce2920891f2d89fe64c58f8
llvm-svn: 316981
Serguei Katkov [Tue, 31 Oct 2017 07:01:35 +0000 (07:01 +0000)]
[CGP] Fix the detection of trivial case for addressing mode
The address can be presented as a bitcast of baseReg.
In this case it is still trivial but OriginalValue != baseReg.
llvm-svn: 316980
Max Kazantsev [Tue, 31 Oct 2017 06:19:05 +0000 (06:19 +0000)]
[IRCE][NFC] Rename fields of InductiveRangeCheck
Rename `Offset`, `Scale`, `Length` into `Begin`, `Step`, `End` respectively
to make naming of similar entities for Ranges and Range Checks more
consistent.
Differential Revision: https://reviews.llvm.org/D39414
llvm-svn: 316979
Craig Topper [Tue, 31 Oct 2017 06:01:04 +0000 (06:01 +0000)]
[X86] Make AVX512_512_SET0 XMM16-31 lower to 128-bit XOR when AVX512VL is enabled. Use 128-bit VLX instruction when VLX is enabled.
Unfortunately, this weakens our ability to do domain fixing when AVX512DQ is not enabled, but it is consistent with our 256-bit behavior.
Maybe we should add custom handling to domain fixing to allow EVEX integer XOR/AND/OR/ANDN to switch to VEX encoded fp instructions if the high registers aren't being used?
llvm-svn: 316978
Max Kazantsev [Tue, 31 Oct 2017 05:33:58 +0000 (05:33 +0000)]
[NFC] Get rid of variables used in assert only
llvm-svn: 316977
Philip Reames [Tue, 31 Oct 2017 05:16:46 +0000 (05:16 +0000)]
[IndVarSimplify] Simplify code using preheader assumption
As noted in the nice block comment, the previous code didn't actually handle multi-entry loops correctly, it just assumed SCEV didn't analyze such loops. Given SCEV has comments to the contrary, that seems a bit suspect. More importantly, the pass actually requires loopsimplify form which ensures a loop-preheader is available. Remove the excessive generaility and shorten the code greatly.
Note that we do successfully analyze many multi-entry loops, but we do so by converting them to single entry loops. See the added test case.
llvm-svn: 316976
Max Kazantsev [Tue, 31 Oct 2017 05:07:56 +0000 (05:07 +0000)]
Reapply "[GVN] Prevent LoadPRE from hoisting across instructions that don't pass control flow to successors"
This patch fixes the miscompile that happens when PRE hoists loads across guards and
other instructions that don't always pass control flow to their successors. PRE is now prohibited
to hoist across such instructions because there is no guarantee that the load standing after such
instruction is still valid before such instruction. For example, a load from under a guard may be
invalid before the guard in the following case:
int array[LEN];
...
guard(0 <= index && index < LEN);
use(array[index]);
Differential Revision: https://reviews.llvm.org/D37460
llvm-svn: 316975
Philip Reames [Tue, 31 Oct 2017 04:19:06 +0000 (04:19 +0000)]
[SimplifyIndVar] Extract out invariant expression handling
Previously, the code returned early from the *function* when it couldn't find a free expansion, it should be returning from the *transform*. I don't have a test case, noticed this via inspection.
As a follow up, I'm going to revisit the logic in the extract function. I think that essentially the whole helper routine can be replaced with SCEVExpander, but I wanted to do that in a series of separate commits.
llvm-svn: 316974
Craig Topper [Tue, 31 Oct 2017 02:34:29 +0000 (02:34 +0000)]
[X86] Clang-format some code. NFC
llvm-svn: 316973
Shoaib Meenai [Tue, 31 Oct 2017 01:30:46 +0000 (01:30 +0000)]
[cmake] Make check_linker_flags operate via linker flags
`check_linker_flags` currently sets the *compiler* flags (via
`CMAKE_REQUIRED_FLAGS`), and thus implicitly relies on cmake's default
behavior of passing the compiler flags to the linker. This breaks when
cmake's build rules have been altered to not pollute the link line with
compiler flags (which can be desirable for build cleanliness). Instead,
set `CMAKE_EXE_LINKER_FLAGS` explicitly and use `CMP0056` to ensure the
linker flags are passed along. Additionally, since we're inside a
function, we can just alter the variable directly (as the alteration
will be limited to the scope of the function) rather than saving and
restoring the old value.
Differential Revision: https://reviews.llvm.org/D39431
llvm-svn: 316972
Alex Lorenz [Tue, 31 Oct 2017 01:28:17 +0000 (01:28 +0000)]
[refactor] select the entire DeclStmt if one ifs decls is selected
llvm-svn: 316971
Marshall Clow [Tue, 31 Oct 2017 00:19:47 +0000 (00:19 +0000)]
Fix broken links; update more issues.
llvm-svn: 316970
Marshall Clow [Tue, 31 Oct 2017 00:05:17 +0000 (00:05 +0000)]
Add a fail test for aligned_union of an incomplete type. See LWG#2979. NFC
llvm-svn: 316969
Philip Reames [Tue, 31 Oct 2017 00:04:09 +0000 (00:04 +0000)]
Undo accidental commit
These files shouldn't have been submitted in 316967
llvm-svn: 316968
Philip Reames [Mon, 30 Oct 2017 23:59:51 +0000 (23:59 +0000)]
[CGP] Fix crash on i96 bit multiply
Issue found by llvm-isel-fuzzer on OSS fuzz, https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=3725
If anyone actually cares about > 64 bit arithmetic, there's a lot more to do in this area. There's a bunch of obviously wrong code in the same function. I don't have the time to fix all of them and am just using this to understand what the workflow for fixing fuzzer cases might look like.
llvm-svn: 316967
Alex Lorenz [Mon, 30 Oct 2017 22:55:11 +0000 (22:55 +0000)]
Typo correct the condition of 'do-while' before exiting its scope
rdar://
35172419
llvm-svn: 316966
Richard Smith [Mon, 30 Oct 2017 22:38:20 +0000 (22:38 +0000)]
[modules] Retain multiple using-directives in the same scope even if they name the same namespace.
They might have different visibility, and thus discarding all but one of them
can result in rejecting valid code. Also fix name lookup to cope with multiple
using-directives being found that denote the same namespace, where some are not
visible -- don't cache an "already visited" state for a using-directive that we
didn't visit because it was hidden.
llvm-svn: 316965
Simon Pilgrim [Mon, 30 Oct 2017 22:38:07 +0000 (22:38 +0000)]
Fix unused variable warnings. NFCI.
llvm-svn: 316964
George Karpenkov [Mon, 30 Oct 2017 22:31:57 +0000 (22:31 +0000)]
[analyzer] Use the same filename for the header and the implementation of BugReporterVisitor
Differential Revision: https://reviews.llvm.org/D37935
llvm-svn: 316963
Simon Pilgrim [Mon, 30 Oct 2017 22:23:57 +0000 (22:23 +0000)]
[SelectionDAG] Tidyup computeKnownBits extension/truncation cases. NFCI.
We don't need to extend/truncate the Known structure before calling computeKnownBits - it will reset at the start of the function.
llvm-svn: 316962
Jake Ehrlich [Mon, 30 Oct 2017 22:08:11 +0000 (22:08 +0000)]
Give .note.gnu.build-id section alignment 4
All SHT_NOTE sections should have minimum alignment 4.
Differential Revision: https://reviews.llvm.org/D38907
llvm-svn: 316961
Javed Absar [Mon, 30 Oct 2017 22:00:06 +0000 (22:00 +0000)]
[AArch64]: range loopify frame-lowering
llvm-svn: 316960
Petr Hosek [Mon, 30 Oct 2017 21:29:26 +0000 (21:29 +0000)]
[sanitizer][Fuchsia] Add a missing semicolon
Differential Revision: https://reviews.llvm.org/D39433
llvm-svn: 316959
Rui Ueyama [Mon, 30 Oct 2017 21:19:54 +0000 (21:19 +0000)]
Fix -fuse-ld feature detection error.
check_cxx_compiler_flag doesn't seem to try to link a program, so
the existing code doesn't correctly detect the availability of a given
linker. This patch uses check_cxx_source_compiles instead.
I confirmed that cmake now reports this error
Host compiler does not support '-fuse-ld=foo'
for -DLLVM_USE_LINKER=foo.
Differential Revision: https://reviews.llvm.org/D39274
llvm-svn: 316958
Yaxun Liu [Mon, 30 Oct 2017 21:19:41 +0000 (21:19 +0000)]
InferAddressSpaces: Fix bug about replacing addrspacecast
InferAddressSpaces assumes the pointee type of addrspacecast
is the same as the operand, which is not always true and causes
invalid IR.
This bug cause build failure in HCC.
This patch fixes that.
Differential Revision: https://reviews.llvm.org/D39432
llvm-svn: 316957
Tim Shen [Mon, 30 Oct 2017 21:12:14 +0000 (21:12 +0000)]
[CMake] Fix linker detection in AddLLVM.cmake
Fix linker not being correctly detected when a custom one is specified
through LLVM_USE_LINKER CMake variable.
In particular,
cmake -DCMAKE_BUILD_TYPE=Release -DLLVM_USE_LINKER=gold ../llvm
resulted into
Linker detection: GNU ld
instead of
Linker detection: GNU Gold
due to the construction not accounting for such variable. It led to the general
confusion and prevented setting linker-specific flags inside functions defined
in AddLLVM.cmake.
Thanks Oleksii Vilchanskyi for the patch!
llvm-svn: 316956
Craig Topper [Mon, 30 Oct 2017 21:09:19 +0000 (21:09 +0000)]
[X86] Add AVX512 support to fast isel's X86ChooseCmpOpcode.
llvm-svn: 316955
Jim Ingham [Mon, 30 Oct 2017 20:44:45 +0000 (20:44 +0000)]
Remove a stray space.
llvm-svn: 316954
Davide Italiano [Mon, 30 Oct 2017 20:20:16 +0000 (20:20 +0000)]
[NewGVN] Stop assuming PHI args ordering when looking at phi-of-ops.
It's not guaranteed. There's a bug open to sort them in predecessor
order, but it won't happen anytime soon. In the meanwhile, passes
will have to do an O(#preds) scan. Such is life.
llvm-svn: 316953
Stefan Pintilie [Mon, 30 Oct 2017 19:55:38 +0000 (19:55 +0000)]
Revert "[PowerPC] Try to simplify a Swap if it feeds a Splat"
Revert r316478.
A test case has failed.
Will recommit this change once we find and fix the failure.
This reverts commit
7c330fabaedaba3d02c58bc3cc1198896c895f34.
llvm-svn: 316952
Marshall Clow [Mon, 30 Oct 2017 19:51:58 +0000 (19:51 +0000)]
Add more fuzzing bits: partial_sort_copy, partition_copy, unique, unique_copy. No functional change to libc++; this is all test infastructure
llvm-svn: 316951
Daniel Neilson [Mon, 30 Oct 2017 19:51:48 +0000 (19:51 +0000)]
Create instruction classes for identifying any atomicity of memory intrinsic. (NFC)
Summary:
For reference, see: http://lists.llvm.org/pipermail/llvm-dev/2017-August/116589.html
This patch fleshes out the instruction class hierarchy with respect to atomic and
non-atomic memory intrinsics. With this change, the relevant part of the class
hierarchy becomes:
IntrinsicInst
-> MemIntrinsicBase (methods-only class)
-> MemIntrinsic (non-atomic intrinsics)
-> MemSetInst
-> MemTransferInst
-> MemCpyInst
-> MemMoveInst
-> AtomicMemIntrinsic (atomic intrinsics)
-> AtomicMemSetInst
-> AtomicMemTransferInst
-> AtomicMemCpyInst
-> AtomicMemMoveInst
-> AnyMemIntrinsic (both atomicities)
-> AnyMemSetInst
-> AnyMemTransferInst
-> AnyMemCpyInst
-> AnyMemMoveInst
This involves some class renaming:
ElementUnorderedAtomicMemCpyInst -> AtomicMemCpyInst
ElementUnorderedAtomicMemMoveInst -> AtomicMemMoveInst
ElementUnorderedAtomicMemSetInst -> AtomicMemSetInst
A script for doing this renaming in downstream trees is included below.
An example of where the Any* classes should be used in LLVM is when reasoning
about the effects of an instruction (ex: aliasing).
---
Script for renaming AtomicMem* classes:
PREFIXES="[<,([:space:]]"
CLASSES="MemIntrinsic|MemTransferInst|MemSetInst|MemMoveInst|MemCpyInst"
SUFFIXES="[;)>,[:space:]]"
REGEX="(${PREFIXES})ElementUnorderedAtomic(${CLASSES})(${SUFFIXES})"
REGEX2="visitElementUnorderedAtomic(${CLASSES})"
FILES=$( grep -E "(${REGEX}|${REGEX2})" -r . | tr ':' ' ' | awk '{print $1}' | sort | uniq )
SED_SCRIPT="s~${REGEX}~\1Atomic\2\3~g"
SED_SCRIPT2="s~${REGEX2}~visitAtomic\1~g"
for f in $FILES; do
echo "Processing: $f"
sed -i ".bak" -E "${SED_SCRIPT};${SED_SCRIPT2};${EA_SED_SCRIPT};${EA_SED_SCRIPT2}" $f
done
Reviewers: sanjoy, deadalnix, apilipenko, anna, skatkov, mkazantsev
Reviewed By: sanjoy
Subscribers: hfinkel, jholewinski, arsenm, sdardis, nhaehnle, JDevlieghere, javed.absar, llvm-commits
Differential Revision: https://reviews.llvm.org/D38419
llvm-svn: 316950
Mandeep Singh Grang [Mon, 30 Oct 2017 19:42:41 +0000 (19:42 +0000)]
[GVNHoist] Fix non-deterministic sort order of PHIs for identical instructions
Summary: This fixes failure in Transforms/GVNHoist/hoist.ll uncovered by D39245.
Reviewers: hiraditya, spop, dberlin
Reviewed By: dberlin
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D39410
llvm-svn: 316949
George Karpenkov [Mon, 30 Oct 2017 19:40:33 +0000 (19:40 +0000)]
[analyzer] [tests] Remove empty folders in reference results, do not store diffs.txt
Storing diffs.txt is now redundant, as we simply dump the CmpRuns output
to stdout (it is saved in CI and tends to be small).
Not generating those files enables us to remove empty folders, which
confuse git, as it would not add them with reference results.
llvm-svn: 316948
Simon Pilgrim [Mon, 30 Oct 2017 19:31:08 +0000 (19:31 +0000)]
[SelectionDAG] Add VSELECT demanded elts support to computeKnownBits
llvm-svn: 316947
Zvi Rackover [Mon, 30 Oct 2017 19:29:15 +0000 (19:29 +0000)]
X86 Tests: Update the variable-index permute tests with FP types. NFC.
These cases will be addressed in a future update to D39126.
llvm-svn: 316946
Simon Pilgrim [Mon, 30 Oct 2017 19:19:58 +0000 (19:19 +0000)]
[X86][SSE] Add another computeKnownBits test showing missing VSELECT demandedelts support
llvm-svn: 316945
Simon Pilgrim [Mon, 30 Oct 2017 19:08:21 +0000 (19:08 +0000)]
[SelectionDAG] Add VSELECT support to computeKnownBits
llvm-svn: 316944
Kostya Kortchinsky [Mon, 30 Oct 2017 19:06:59 +0000 (19:06 +0000)]
Fix warning + death test + failing test on Windows (D39072).
Summary: Fixes https://reviews.llvm.org/D39072
Reviewers: cryptoad
Reviewed By: cryptoad
Subscribers: kubamracek
Differential Revision: https://reviews.llvm.org/D39427
llvm-svn: 316943
Martin Storsjo [Mon, 30 Oct 2017 19:06:34 +0000 (19:06 +0000)]
Change unw_word_t to always have the same size as the pointer size
This matches the original libunwind API. This also unifies the
type between ARM EHABI and the other configurations, and allows
getting rid of a number of casts in log messages.
The cursor size updates for ppc and or1k are untested, but
unw_proc_info_t shrinks by 4 uint64_t units on i386 at least.
Differential Revision: https://reviews.llvm.org/D39365
llvm-svn: 316942