Richard Smith [Thu, 11 Jun 2015 23:46:11 +0000 (23:46 +0000)]
[modules] Fix crash with multiple levels of default template argument merging.
llvm-svn: 239575
Reid Kleckner [Thu, 11 Jun 2015 23:37:18 +0000 (23:37 +0000)]
[WinEH] Put finally pointers in the handler scope table field
We were putting them in the filter field, which is correct for 64-bit
but wrong for 32-bit.
Also switch the order of scope table entry emission so outermost entries
are emitted first, and fix an obvious state assignment bug.
llvm-svn: 239574
Sean Silva [Thu, 11 Jun 2015 23:34:13 +0000 (23:34 +0000)]
[cleanup] Remove some unused #ifdef's
This is all going through the VFS layer now, so there's nothing
platform-specific here.
llvm-svn: 239573
Tyler Nowicki [Thu, 11 Jun 2015 23:23:17 +0000 (23:23 +0000)]
Add assume_safety option for pragma loop vectorize and interleave.
Specifying #pragma clang loop vectorize(assume_safety) on a loop adds the
mem.parallel_loop_access metadata to each load/store operation in the loop. This
metadata tells loop access analysis (LAA) to skip memory dependency checking.
llvm-svn: 239572
NAKAMURA Takumi [Thu, 11 Jun 2015 23:04:56 +0000 (23:04 +0000)]
MC: Prune \return corresponding to r239552. [-Wdocumentation]
llvm-svn: 239571
Lang Hames [Thu, 11 Jun 2015 22:51:01 +0000 (22:51 +0000)]
[Orc] Attempted fix for GCC ICE on Polly builder.
Along the same lines as the fix in r228568.
llvm-svn: 239570
Richard Smith [Thu, 11 Jun 2015 22:48:25 +0000 (22:48 +0000)]
[modules] Fix assert/crash when parsing and merging a definition of a class with a base-specifier inside a namespace.
llvm-svn: 239569
Juergen Ributzka [Thu, 11 Jun 2015 22:40:04 +0000 (22:40 +0000)]
[Stackmaps][X86] Remove EFLAGS and IP registers from the live-out mask.
Remove the EFLAGS from the stackmap live-out mask. The EFLAGS register is not
supposed to be part of that set, because the X86 calling conventions mark the
register as NOT preserved.
Also remove the IP registers, since spilling and restoring those doesn't really
make any sense.
Related to rdar://problem/
21019635.
llvm-svn: 239568
Reid Kleckner [Thu, 11 Jun 2015 22:32:23 +0000 (22:32 +0000)]
[WinEH] Create an llvm.x86.seh.exceptioninfo intrinsic
This intrinsic is like framerecover plus a load. It recovers the EH
registration stack allocation from the parent frame and loads the
exception information field out of it, giving back a pointer to an
EXCEPTION_POINTERS struct. It's designed for clang to use in SEH filter
expressions instead of accessing the EXCEPTION_POINTERS parameter that
is available on x64.
This required a minor change to MC to allow defining a label variable to
another absolute framerecover label variable.
llvm-svn: 239567
Reid Kleckner [Thu, 11 Jun 2015 22:22:45 +0000 (22:22 +0000)]
[Support] Fix a race initializing a static local in MSVC
static local initialization isn't thread safe with MSVC and a race was
reported in PR23817. We can't use std::atomic because it's not trivially
constructible, so instead do some lame volatile global integer
manipulation.
llvm-svn: 239566
Michael Zolotukhin [Thu, 11 Jun 2015 22:17:39 +0000 (22:17 +0000)]
Update stale comment before analyzeLoopUnrollCost. NFC.
llvm-svn: 239565
Lang Hames [Thu, 11 Jun 2015 22:12:24 +0000 (22:12 +0000)]
[Orc] Remove some unnecesary includes and whitespace that slipped in to r239561.
NFC.
llvm-svn: 239564
Peter Collingbourne [Thu, 11 Jun 2015 21:49:54 +0000 (21:49 +0000)]
COFF: Symbol resolution for common and comdat symbols defined in bitcode.
In the case where either a bitcode file and a regular file or two bitcode
files export a common or comdat symbol with the same name, the linker needs
to pick one of them following COFF semantics. This patch implements a design
for resolving such symbols that pushes most of the work onto either LLD's
regular mechanism for resolving common or comdat symbols or the IR linker's
mechanism for doing the same.
We modify SymbolBody::compare to always prefer non-bitcode symbols, so that
during the initial phase of symbol resolution, the symbol table always contains
a regular symbol in any case where we need to choose between a regular and
a bitcode symbol. In SymbolTable::addCombinedLTOObject, we force export
any bitcode symbols that were initially pre-empted by a regular symbol,
and later use SymbolBody::compare to choose between the regular symbol in
the symbol table and the regular symbol from the combined LTO object file.
This design seems to be sound, so long as the resolution mechanism is defined
to be commutative and associative modulo arbitrary choices between symbols
(which seems to be the case for COFF).
Differential Revision: http://reviews.llvm.org/D10329
llvm-svn: 239563
Marshall Clow [Thu, 11 Jun 2015 21:47:39 +0000 (21:47 +0000)]
Change #ifdefs in test to UNSUPPORTED. No functionality change in the tests
llvm-svn: 239562
Lang Hames [Thu, 11 Jun 2015 21:45:19 +0000 (21:45 +0000)]
[Orc] Make partition identification in the CompileOnDemand layer lazy.
This also breaks out the logical dylib symbol resolution logic.
llvm-svn: 239561
Peter Collingbourne [Thu, 11 Jun 2015 21:42:18 +0000 (21:42 +0000)]
Object: Prepend __imp_ when mangling a dllimport symbol in IRObjectFile.
We cannot prepend __imp_ in the IR mangler because a function reference may
be emitted unmangled in a constant initializer. The linker is expected to
resolve such references to thunks. This is covered by the new test case.
Strictly speaking we ought to emit two undefined symbols, one with __imp_ and
one without, as we cannot know which symbol the final object file will refer
to. However, this would require rather intrusive changes to IRObjectFile,
and lld works fine without it for now.
This reimplements r239437, which was reverted in r239502.
Differential Revision: http://reviews.llvm.org/D10400
llvm-svn: 239560
Peter Collingbourne [Thu, 11 Jun 2015 21:41:27 +0000 (21:41 +0000)]
LTO: expose LTO_SYMBOL_COMDAT flag, which indicates that the definition is part of a comdat group.
Reviewers: rafael
Subscribers: llvm-commits, ruiu
Differential Revision: http://reviews.llvm.org/D10330
llvm-svn: 239559
Hans Wennborg [Thu, 11 Jun 2015 21:21:57 +0000 (21:21 +0000)]
[ms] Do lookup in dependent base classes also when overload resolution fails (PR23810)
This patch does two things in order to enable compilation of the problematic code in PR23810:
1. In Sema::buildOverloadedCallSet, it postpones lookup for MS mode when no
viable candidate is found in the overload set. Previously, lookup would only
be postponed here if the overload set was empty.
2. Make BuildRecoveryCallExpr call Sema::DiagnoseEmptyLookup under more circumstances.
There is a comment in DiagnoseTwoPhaseLookup that says "Don't diagnose names we find in
classes; we get much better diagnostics for these from DiagnoseEmptyLookup." The problem
was that DiagnoseEmptyLookup might not get called later, and we failed to recover.
Differential Revision: http://reviews.llvm.org/D10369
llvm-svn: 239558
Douglas Katzman [Thu, 11 Jun 2015 20:03:23 +0000 (20:03 +0000)]
Fix English usage in command line flag help string.
llvm-svn: 239556
Davide Italiano [Thu, 11 Jun 2015 19:59:04 +0000 (19:59 +0000)]
[ELF] Introduce getValue() for ELF Symbols.
Differential Revision: http://reviews.llvm.org/D10328
Reviewed by: rafael
llvm-svn: 239555
Daniel Sanders [Thu, 11 Jun 2015 19:41:26 +0000 (19:41 +0000)]
Replace string GNU Triples with llvm::Triple in TargetMachine. NFC.
Summary:
For the moment, TargetMachine::getTargetTriple() still returns a StringRef.
This continues the patch series to eliminate StringRef forms of GNU triples
from the internals of LLVM that began in r239036.
Reviewers: rengolin
Reviewed By: rengolin
Subscribers: ted, llvm-commits, rengolin, jholewinski
Differential Revision: http://reviews.llvm.org/D10362
llvm-svn: 239554
Ahmed Bougacha [Thu, 11 Jun 2015 19:30:37 +0000 (19:30 +0000)]
[CodeGen] ArrayRef'ize cond/pred in various TII APIs. NFC.
llvm-svn: 239553
Rafael Espindola [Thu, 11 Jun 2015 18:58:08 +0000 (18:58 +0000)]
Generalize emitAbsoluteSymbolDiff.
This makes emitAbsoluteSymbolDiff always succeed and moves logic from the asm
printer to it.
The object one now also works on ELF. If two symbols are in the same fragment,
we will never move them apart.
llvm-svn: 239552
Alexey Samsonov [Thu, 11 Jun 2015 18:25:54 +0000 (18:25 +0000)]
Set proper debug location for branch added in BasicBlock::splitBasicBlock().
This improves debug locations in passes that do a lot of basic block
transformations. Important case is LoopUnroll pass, the test for correct
debug locations accompanies this change.
Test Plan: regression test suite
Reviewers: dblaikie, sanjoy
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D10367
llvm-svn: 239551
Alexey Samsonov [Thu, 11 Jun 2015 18:25:44 +0000 (18:25 +0000)]
[LoopUnroll] Use IRBuilder to create branch instructions.
Use IRBuilder::Create(Cond)?Br instead of constructing instructions
manually with BranchInst::Create(). It's consistent with other
uses of IRBuilder in this pass, and has an additional important
benefit:
Using IRBuilder will ensure that new branch instruction will get
the same debug location as original terminator instruction it will
eventually replace.
For now I'm not adding a testcase, as currently original terminator
instruction also lack debug location due to missing debug location
propagation in BasicBlock::splitBasicBlock. That is, the testcase
will accompany the fix for the latter I'm going to mail soon.
llvm-svn: 239550
Ahmed Bougacha [Thu, 11 Jun 2015 18:19:34 +0000 (18:19 +0000)]
[CodeGen] Emit Constants for immediate inlineasm arguments.
For inline assembly immediate constraints, we currently always use
EmitScalarExpr, instead of directly emitting the constant. When the
overflow sanitizer is enabled, this generates overflow intrinsics
instead of constants.
Instead, emit a constant for constraints that either require an
immediate (e.g. 'I' on X86), or only accepts constants (immediate
or symbolic; i.e., don't accept registers or memory).
Fixes PR19763.
Differential Revision: http://reviews.llvm.org/D10255
llvm-svn: 239549
Hans Wennborg [Thu, 11 Jun 2015 18:14:11 +0000 (18:14 +0000)]
clang-cl: Add an alias for /wd4910
llvm-svn: 239548
Hans Wennborg [Thu, 11 Jun 2015 17:51:58 +0000 (17:51 +0000)]
Spell -Wdllexport-explicit-instantiation-decl correctly.
llvm-svn: 239547
Jonathan Peyton [Thu, 11 Jun 2015 17:36:16 +0000 (17:36 +0000)]
Remove unused variables '__kmp_build_check_*' for non assert builds.
Add new LIBOMP_ENABLE_ASSERTIONS macro which can be set in a standalone build
or takes the value of LLVM_ENABLE_ASSERTIONS when inside llvm/projects. This
change also defines the KMP_BUILD_ASSERT() macro to do nothing when ENABLE_ASSERTIONS
is off. This means the __kmp_build_check_* types won't be defined and thus, no warnings.
http://lists.cs.uiuc.edu/pipermail/openmp-dev/2015-June/000719.html
Patch by Jack Howarth and Jonathan Peyton
llvm-svn: 239546
Benjamin Kramer [Thu, 11 Jun 2015 17:30:34 +0000 (17:30 +0000)]
Replace an instance of custom atomics with standard ones.
Eventually I want to get rid of them entirely, but Statistic.h is still blocked
on MSVC bugs. No functionality change.
llvm-svn: 239545
Rafael Espindola [Thu, 11 Jun 2015 17:30:33 +0000 (17:30 +0000)]
This reverts commit r239529 and r239514.
Revert "[AArch64] Match interleaved memory accesses into ldN/stN instructions."
Revert "Fixing MSVC 2013 build error."
The test/CodeGen/AArch64/aarch64-interleaved-accesses.ll test was failing on OS X.
llvm-svn: 239544
Reid Kleckner [Thu, 11 Jun 2015 17:25:24 +0000 (17:25 +0000)]
Revert "Fix merges of non-zero vector stores"
This reverts commit r239539.
It was causing SDAG assertions while building freetype.
llvm-svn: 239543
Jonathan Peyton [Thu, 11 Jun 2015 17:23:57 +0000 (17:23 +0000)]
Implement recursive CMake.
Most CMake build systems put CMakeLists.txt files inside source directories where
items need to get built. This change follows that convention by adding a new
runtime/src/CMakeLists.txt file. An additional benefit is this helps logically
seperate configuring with building as well. This change is mostly just copying and
pasting the bottom half of runtime/CMakeLists.txt into runtime/src/CMakeLists.txt,
but a few changes had to be made to get it to work. Most of those changes were to
directory prefixes.
Differential Revision: http://reviews.llvm.org/D10344
llvm-svn: 239542
Douglas Katzman [Thu, 11 Jun 2015 16:46:27 +0000 (16:46 +0000)]
Fix comment typos.
llvm-svn: 239541
Matt Arsenault [Thu, 11 Jun 2015 16:13:39 +0000 (16:13 +0000)]
SLSR: Pass address space to isLegalAddressingMode
This only updates one of the uses. The other is used in cases
that may never touch memory, so I'm not sure why this is even
calling it. Perhaps there should be a new, similar hook for such
cases or pass -1 for unknown address space.
llvm-svn: 239540
Matt Arsenault [Thu, 11 Jun 2015 16:03:52 +0000 (16:03 +0000)]
Fix merges of non-zero vector stores
Now actually stores the non-zero constant instead of 0.
I somehow forgot to include this part of r238108.
The test change was just an independent instruction order swap,
so just add another check line to satisfy CHECK-NEXT.
llvm-svn: 239539
Daniel Sanders [Thu, 11 Jun 2015 15:34:59 +0000 (15:34 +0000)]
Replace string GNU Triples with llvm::Triple in computeDataLayout(). NFC.
Summary:
This continues the patch series to eliminate StringRef forms of GNU triples
from the internals of LLVM that began in r239036.
Reviewers: rengolin
Reviewed By: rengolin
Subscribers: llvm-commits, jfb, rengolin
Differential Revision: http://reviews.llvm.org/D10361
llvm-svn: 239538
Douglas Katzman [Thu, 11 Jun 2015 15:05:22 +0000 (15:05 +0000)]
Add comments to PrintActions1 and Driver::PrintActions.
Differential Revision: http://reviews.llvm.org/D10214
llvm-svn: 239537
Sanjay Patel [Thu, 11 Jun 2015 14:53:41 +0000 (14:53 +0000)]
add the -mrecip driver flag and process its options (3rd try)
The 1st and 2nd tries to land this (r238055, r238851) were reverted due to
bot failures caused by the LLVM part of the patch. That was hopefully fixed
after r239001.
This is the front-end counterpart to D8982.
The -mrecip option interface is based on maintaining compatibility with gcc:
https://gcc.gnu.org/onlinedocs/gcc-4.9.2/gcc/i386-and-x86-64-Options.html#index-mrecip_003dopt-1627
https://gcc.gnu.org/onlinedocs/gcc-4.9.2/gcc/RS_002f6000-and-PowerPC-Options.html#index-mrecip-2289
...while adding more functionality (allowing users to specify the number of refinement steps for each
estimate type).
Differential Revision: http://reviews.llvm.org/D8989
llvm-svn: 239536
Tom Stellard [Thu, 11 Jun 2015 14:51:50 +0000 (14:51 +0000)]
R600/SI: Define latency for flat instructions
llvm-svn: 239535
Tom Stellard [Thu, 11 Jun 2015 14:51:49 +0000 (14:51 +0000)]
R600/SI: Move flat instruction defs to CIInstructions.td
llvm-svn: 239534
Tom Stellard [Thu, 11 Jun 2015 14:51:46 +0000 (14:51 +0000)]
R600/SI: Add -mcpu=bonaire to a test that uses flat address space
Flat instructions don't exist on SI, but there is a bug in the backend that
allows them to be selected.
llvm-svn: 239533
Sanjay Patel [Thu, 11 Jun 2015 14:26:49 +0000 (14:26 +0000)]
remove function names from comments; NFC
llvm-svn: 239532
Daniel Jasper [Thu, 11 Jun 2015 13:31:45 +0000 (13:31 +0000)]
clang-format: Make SFS_Inline imply SFS_Empty.
In the long run, these two might be independent or we might to only
allow specific combinations. Until we have a corresponding request,
however, it is hard to do the right thing and choose the right
configuration options. Thus, just don't touch the options yet and
just modify the behavior slightly.
llvm-svn: 239531
Daniel Jasper [Thu, 11 Jun 2015 13:29:20 +0000 (13:29 +0000)]
clang-format: [JS] Ensure that formatting actually takes place in tests.
And fix formatting issue discovered by that :-).
llvm-svn: 239530
Aaron Ballman [Thu, 11 Jun 2015 13:06:02 +0000 (13:06 +0000)]
Fixing MSVC 2013 build error.
llvm-svn: 239529
Yaron Keren [Thu, 11 Jun 2015 12:33:25 +0000 (12:33 +0000)]
C++11 rangify several loops.
llvm-svn: 239528
Gabor Ballabas [Thu, 11 Jun 2015 12:29:56 +0000 (12:29 +0000)]
Allow case-insensitive values for -march for ARM in line with GCC.
GCC allows case-insensitive values for -mcpu, -march and -mtune options.
This patch implements the same behaviour for the -march option for ARM.
llvm-svn: 239527
Daniel Marjamaki [Thu, 11 Jun 2015 12:28:14 +0000 (12:28 +0000)]
Token: complement is() method with isOneOf() to allow easier usage
llvm-svn: 239526
Toma Tabacu [Thu, 11 Jun 2015 12:13:18 +0000 (12:13 +0000)]
[mips] Pass on -m{single,double}-float to GAS.
Summary: We already pass these to the IAS, but not to GAS.
Reviewers: dsanders, atanasyan
Reviewed By: atanasyan
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D10358
llvm-svn: 239525
Alexey Bataev [Thu, 11 Jun 2015 10:53:56 +0000 (10:53 +0000)]
[OPENMP] Fox for llvm.org/PR23663: OpenMP crash
Destroy RuntimeCleanupScope before generation of termination instruction in parallel loop precondition.
llvm-svn: 239524
Toma Tabacu [Thu, 11 Jun 2015 10:36:10 +0000 (10:36 +0000)]
Recommit "[mips] [IAS] Add support for BNE and BEQ with an immediate operand." (r239396).
Apparently, Arcanist didn't include some of my local changes in my previous
commit attempt.
llvm-svn: 239523
Zoran Jovanovic [Thu, 11 Jun 2015 10:22:46 +0000 (10:22 +0000)]
[mips][microMIPS] Implement ERET and ERETNC instructions
http://reviews.llvm.org/D10091
llvm-svn: 239522
Manuel Klimek [Thu, 11 Jun 2015 10:14:13 +0000 (10:14 +0000)]
Fix crash in clang-format.
The following example used to crash clang-format.
#define a\
/**/}
Adjusting the indentation level cache for the line starting with the
comment would lead to an out-of-bounds array read.
llvm-svn: 239521
Zoran Jovanovic [Thu, 11 Jun 2015 09:51:58 +0000 (09:51 +0000)]
[mips] Change existing uimm10 operand to restrict the accepted immediates
http://reviews.llvm.org/D10312
llvm-svn: 239520
Zoran Jovanovic [Thu, 11 Jun 2015 09:42:10 +0000 (09:42 +0000)]
[mips][microMIPSr6] Change disassembler tests to one line format
llvm-svn: 239519
Hao Liu [Thu, 11 Jun 2015 09:18:07 +0000 (09:18 +0000)]
[LoopVectorize] Revert the enabling of interleaved memory access in Loop Vectorizor, which was wrongly committed in r239514.
llvm-svn: 239515
Hao Liu [Thu, 11 Jun 2015 09:05:02 +0000 (09:05 +0000)]
[AArch64] Match interleaved memory accesses into ldN/stN instructions.
Add a pass AArch64InterleavedAccess to identify and match interleaved memory accesses. This pass transforms an interleaved load/store into ldN/stN intrinsic. As Loop Vectorizor disables optimization on interleaved accesses by default, this optimization is also disabled by default. To enable it by "-aarch64-interleaved-access-opt=true"
E.g. Transform an interleaved load (Factor = 2):
%wide.vec = load <8 x i32>, <8 x i32>* %ptr
%v0 = shuffle %wide.vec, undef, <0, 2, 4, 6> ; Extract even elements
%v1 = shuffle %wide.vec, undef, <1, 3, 5, 7> ; Extract odd elements
Into:
%ld2 = { <4 x i32>, <4 x i32> } call aarch64.neon.ld2(%ptr)
%v0 = extractelement { <4 x i32>, <4 x i32> } %ld2, i32 0
%v1 = extractelement { <4 x i32>, <4 x i32> } %ld2, i32 1
E.g. Transform an interleaved store (Factor = 2):
%i.vec = shuffle %v0, %v1, <0, 4, 1, 5, 2, 6, 3, 7> ; Interleaved vec
store <8 x i32> %i.vec, <8 x i32>* %ptr
Into:
%v0 = shuffle %i.vec, undef, <0, 1, 2, 3>
%v1 = shuffle %i.vec, undef, <4, 5, 6, 7>
call void aarch64.neon.st2(%v0, %v1, %ptr)
llvm-svn: 239514
Daniel Jasper [Thu, 11 Jun 2015 08:38:19 +0000 (08:38 +0000)]
clang-format: Don't add spaces in foreach macro definition.
Before clang-format would e.g. add a space into
#define Q_FOREACH(x, y)
which turns this into a non-function-like macro.
Patch by Strager Neds, thank you!
llvm-svn: 239513
David Majnemer [Thu, 11 Jun 2015 08:12:44 +0000 (08:12 +0000)]
Reinstate r239499 and r239503
They were reverted because the FileCheck patterns didn't match on
release builds.
llvm-svn: 239512
Manuel Klimek [Thu, 11 Jun 2015 07:54:35 +0000 (07:54 +0000)]
Revert "[MS ABI] Allow fastcall member function pointers to get CodeGen'd"
Revert "[MS ABI] Allow memfn pointers with unconvertible types to be formed"
This reverts r239499 and r239503; the former breaks tests [1] and the
latter is based on the former.
[1]
http://lab.llvm.org:8080/green/job/clang-stage2-configure-Rlto_check/4473/testReport/Clang/CodeGenCXX/microsoft_abi_virtual_member_pointers_cpp/
llvm-svn: 239511
Arnaud A. de Grandmaison [Thu, 11 Jun 2015 07:50:21 +0000 (07:50 +0000)]
[LiveVariables] Improve isLiveOut runtime performances. NFC.
On large goto table based interpreters, where phi nodes can have (very) large
fan-ins, isLiveOut exhibited poor performances: about 40% of the full
codegen time was spent in PHIElim, sorting MachineBasicBlock addresses.
This patch improve the performances for such cases, and does not show
compile time regressions on the LNT, at bootstrap (llvm+clang+lldb) or
any other benchmarks we have in-house.
llvm-svn: 239510
Simon Pilgrim [Thu, 11 Jun 2015 07:46:37 +0000 (07:46 +0000)]
[X86][SSE] Vectorized i8 and i16 shift operators
This patch ensures that SHL/SRL/SRA shifts for i8 and i16 vectors avoid scalarization. It builds on the existing i8 SHL vectorized implementation of moving the shift bits up to the sign bit position and separating the 4, 2 & 1 bit shifts with several improvements:
1 - SSE41 targets can use (v)pblendvb directly with the sign bit instead of performing a comparison to feed into a VSELECT node.
2 - pre-SSE41 targets were masking + comparing with an 0x80 constant - we avoid this by using the fact that a set sign bit means a negative integer which can be compared against zero to then feed into VSELECT, avoiding the need for a constant mask (zero generation is much cheaper).
3 - SRA i8 needs to be unpacked to the upper byte of a i16 so that the i16 psraw instruction can be correctly used for sign extension - we have to do more work than for SHL/SRL but perf tests indicate that this is still beneficial.
The i16 implementation is similar but simpler than for i8 - we have to do 8, 4, 2 & 1 bit shifts but less shift masking is involved. SSE41 use of (v)pblendvb requires that the i16 shift amount is splatted to both bytes however.
Tested on SSE2, SSE41 and AVX machines.
Differential Revision: http://reviews.llvm.org/D9474
llvm-svn: 239509
Arnaud A. de Grandmaison [Thu, 11 Jun 2015 07:45:05 +0000 (07:45 +0000)]
[PHIElim] Use ranges and const-ify, NFC.
llvm-svn: 239508
Nemanja Ivanovic [Thu, 11 Jun 2015 06:25:36 +0000 (06:25 +0000)]
Clang support for vector quad bit permute and gather instructions through builtins
This patch corresponds to review:
http://reviews.llvm.org/D10095
This is for just two instructions and related builtins:
vbpermq
vgbbd
llvm-svn: 239506
Nemanja Ivanovic [Thu, 11 Jun 2015 06:21:25 +0000 (06:21 +0000)]
LLVM support for vector quad bit permute and gather instructions through builtins
This patch corresponds to review:
http://reviews.llvm.org/D10096
This is the back end portion of the patch related to D10095.
The patch adds the instructions and back end intrinsics for:
vbpermq
vgbbd
llvm-svn: 239505
Richard Smith [Thu, 11 Jun 2015 03:05:39 +0000 (03:05 +0000)]
[modules] Fix a few places where merging wasn't performed if modules was disabled but local module visibilty was enabled.
llvm-svn: 239504
David Majnemer [Thu, 11 Jun 2015 02:38:06 +0000 (02:38 +0000)]
[MS Compatibility] Handle cleanups we create for a ctor closure
This fixes PR23801.
llvm-svn: 239503
Reid Kleckner [Thu, 11 Jun 2015 01:31:48 +0000 (01:31 +0000)]
Revert "Move dllimport name mangling to IR mangler."
This reverts commit r239437.
This broke clang-cl self-hosts. We'd end up calling the __imp_ symbol
directly instead of using it to do an indirect function call.
llvm-svn: 239502
Pete Cooper [Thu, 11 Jun 2015 01:04:56 +0000 (01:04 +0000)]
Remove MachineModuleInfo::UsedFunctions as it has no users.
It hasn't been used since r130964.
This also removes MachineModuleInfo::isUsedFunction and
MachineModuleInfo::AnalyzeModule, both of which were only
there to support UsedFunctions.
llvm-svn: 239501
David Majnemer [Thu, 11 Jun 2015 00:45:44 +0000 (00:45 +0000)]
[MS ABI] Allow fastcall member function pointers to get CodeGen'd
This restriction appears unnecessary and most likely came about during
early work for musttail.
llvm-svn: 239500
David Majnemer [Thu, 11 Jun 2015 00:20:57 +0000 (00:20 +0000)]
[MS ABI] Allow memfn pointers with unconvertible types to be formed
Remove the restriction which forbade forming pointers to member
functions which had parameter types or return types which were not
convertible.
llvm-svn: 239499
Chris Bieneman [Wed, 10 Jun 2015 23:55:07 +0000 (23:55 +0000)]
[CMake] Cleanup add_compiler_rt_object_library to be platform-agnostic
Summary:
This change takes darwin-specific goop that was scattered around CMakeLists files and spread between add_compiler_rt_object_library and add_compiler_rt_darwin_object_library and moves it all under add_compiler_rt_object_library.
The goal of this is to try to push platform handling as low in the utility functions as possible.
Reviewers: rnk, samsonov
Reviewed By: rnk, samsonov
Subscribers: rnk, rsmith, llvm-commits
Differential Revision: http://reviews.llvm.org/D10250
llvm-svn: 239498
Sanjay Patel [Wed, 10 Jun 2015 23:27:33 +0000 (23:27 +0000)]
change assert that will never fire to llvm_unreachable
llvm-svn: 239497
Alexei Starovoitov [Wed, 10 Jun 2015 22:59:13 +0000 (22:59 +0000)]
[bpf] add support for BPF backend
add support for bpfel/bpfeb targets
llvm-svn: 239496
Jingyue Wu [Wed, 10 Jun 2015 22:54:02 +0000 (22:54 +0000)]
[NFC] added a missing space
llvm-svn: 239495
Richard Smith [Wed, 10 Jun 2015 22:49:14 +0000 (22:49 +0000)]
Work around MSVC miscompilation.
llvm-svn: 239494
Pete Cooper [Wed, 10 Jun 2015 22:38:46 +0000 (22:38 +0000)]
Stop returning a Use* from allocHungOffUses.
This always just set the User::OperandList which is now set
in that method instead of being returned.
Reviewed by Duncan Exon Smith.
llvm-svn: 239493
Pete Cooper [Wed, 10 Jun 2015 22:38:41 +0000 (22:38 +0000)]
Add User::growHungoffUses and use it to grow the hung off uses. NFC.
PhiNode, SwitchInst, LandingPad and IndirectBr all had virtually identical
logic for growing the hung off uses.
Move it to User so that they can all call a single shared implementation.
Their destructors were all empty after this change and were deleted. They all
have virtual clone_impl methods which can be used as vtable anchors.
Reviewed by Duncan Exon Smith.
llvm-svn: 239492
Pete Cooper [Wed, 10 Jun 2015 22:38:38 +0000 (22:38 +0000)]
Delete User::dropHungOffUses and move it in to ~User which is the only caller. NFC.
Now that the subclasses which care about hung off uses let ~User clean it up,
there's no need for a separate method. Just inline it to ~User and delete it.
Reviewed by Duncan Exon Smith.
llvm-svn: 239491
Pete Cooper [Wed, 10 Jun 2015 22:38:34 +0000 (22:38 +0000)]
Make User track whether a class has 'hung off uses' and delete them in its destructor.
Currently all of the logic for deleting hung off uses, which PHI/switch/etc use,
is in their classes.
This adds a bit to Value which tracks whether that user had hung off uses,
then User can be responsible for clearing them instead of the sub classes.
Note, the bit used here was taken from NumOperands which was 30-bits.
Given the reduction to 29 bits, and the average User being just over 100 bytes,
a single User with 29-bits of num operands would need 50GB of RAM for itself
so its reasonable to assume that 29-bits is enough for now.
This is a step towards hiding all the hung off uses logic in the User.
Reviewed by Duncan Exon Smith.
llvm-svn: 239490
Pete Cooper [Wed, 10 Jun 2015 22:38:30 +0000 (22:38 +0000)]
Move the special Phi logic for hung off uses in to User::allocHungOffUses. NFC.
PhiNode's need to allocate space for an array of Use[N] and then BasicBlock*[N].
They had their own allocHungOffUses to handle all of this. This moves the logic
in to User::allocHungOffUses and PhiNode passes in a bool to say to allocate
the BB* space too.
Reviewed by Duncan Exon Smith.
llvm-svn: 239489
Peter Collingbourne [Wed, 10 Jun 2015 21:14:34 +0000 (21:14 +0000)]
ArgumentPromotion: Drop sret attribute on functions that are only called directly.
If the first argument to a function is a 'this' argument and the second
has the sret attribute, the ArgumentPromotion pass may promote the 'this'
argument to more than one argument, violating the IR constraint that 'sret'
may only be applied to the first or second argument.
Although this IR constraint is arguably unnecessary, it highlighted the fact
that ArgPromotion does not need to preserve this attribute. Dropping the
attribute reduces register pressure in the backend by avoiding the register
copy required by sret. Because sret implies noalias, we also replace the
former with the latter.
Differential Revision: http://reviews.llvm.org/D10353
llvm-svn: 239488
Richard Smith [Wed, 10 Jun 2015 20:36:34 +0000 (20:36 +0000)]
[modules] Don't allow use of non-visible (inherited) default template arguments.
llvm-svn: 239487
Sanjay Patel [Wed, 10 Jun 2015 20:32:21 +0000 (20:32 +0000)]
[x86] Add a reassociation optimization to increase ILP via the MachineCombiner pass
This is a reimplementation of D9780 at the machine instruction level rather than the DAG.
Use the MachineCombiner pass to reassociate scalar single-precision AVX additions (just a
starting point; see the TODO comments) to increase ILP when it's safe to do so.
The code is closely based on the existing MachineCombiner optimization that is implemented
for AArch64.
This patch should not cause the kind of spilling tragedy that led to the reversion of r236031.
Differential Revision: http://reviews.llvm.org/D10321
llvm-svn: 239486
Richard Smith [Wed, 10 Jun 2015 20:30:23 +0000 (20:30 +0000)]
[modules] Track all default template arguments for a given parameter across
modules, and allow use of a default template argument if any of the parameters
providing it is visible.
llvm-svn: 239485
Sanjay Patel [Wed, 10 Jun 2015 19:52:58 +0000 (19:52 +0000)]
punctuation policing; NFC
llvm-svn: 239484
Serge Pavlov [Wed, 10 Jun 2015 19:06:59 +0000 (19:06 +0000)]
Do not parse members of incomplete class.
If definition of a class is unknown and out-of-line definition of its
member is encountered, do not parse the member declaration.
This change fixes PR18542.
Differential Revision: http://reviews.llvm.org/D8010
llvm-svn: 239483
Reid Kleckner [Wed, 10 Jun 2015 18:14:07 +0000 (18:14 +0000)]
[WinEH] _except_handlerN uses 0 instead of 1 to indicate catch-all
Our usage of 1 was a holdover from __C_specific_handler.
llvm-svn: 239482
Teresa Johnson [Wed, 10 Jun 2015 17:49:45 +0000 (17:49 +0000)]
Pass down the -flto option to the -cc1 job, and from there into the
CodeGenOptions and onto the PassManagerBuilder. This enables gating
the new EliminateAvailableExternally module pass on whether we are
preparing for LTO.
If we are preparing for LTO (e.g. a -flto -c compile), the new pass is not
included as we want to preserve available externally functions for possible
link time inlining.
llvm-svn: 239481
Teresa Johnson [Wed, 10 Jun 2015 17:49:28 +0000 (17:49 +0000)]
Add new EliminateAvailableExternally module pass, which is performed in
O2 compiles just before GlobalDCE, unless we are preparing for LTO.
This pass eliminates available externally globals (turning them into
declarations), regardless of whether they are dead/unreferenced, since
we are guaranteed to have a copy available elsewhere at link time.
This enables additional opportunities for GlobalDCE.
If we are preparing for LTO (e.g. a -flto -c compile), the pass is not
included as we want to preserve available externally functions for possible
link time inlining. The FE indicates whether we are doing an -flto compile
via the new PrepareForLTO flag on the PassManagerBuilder.
llvm-svn: 239480
Alexey Samsonov [Wed, 10 Jun 2015 17:37:38 +0000 (17:37 +0000)]
[GVN] Set proper debug locations for some instructions created by GVN.
Determining proper debug locations for instructions created in
PHITransAddr is tricky. We use a simple approach here and simply copy
debug locations from instructions computing load address to
"corresponding" instructions re-creating the address computation
in predecessor basic blocks.
This may not always be correct, given all the rearrangement and
simplification going on, and debug locations may jump around a lot,
as the basic blocks we copy locations between may be very far from
each other.
Still, this would work good in most simple cases (e.g. when chain
of address computing instruction is short, or our mapping turns out
to be 1-to-1), and we desire to have *some* reasonable debug locations
associated with newly inserted instructions.
See http://reviews.llvm.org/D10351 review thread for more details.
Test Plan: regression test suite
Reviewers: spatel, dblaikie
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D10351
llvm-svn: 239479
Sanjay Patel [Wed, 10 Jun 2015 17:08:12 +0000 (17:08 +0000)]
fix typo in comment; NFC
llvm-svn: 239478
Colin LeMahieu [Wed, 10 Jun 2015 16:52:32 +0000 (16:52 +0000)]
[Hexagon] Adding decoders for signed operands and ensuring all signed operand types disassemble correctly.
llvm-svn: 239477
Scott Douglass [Wed, 10 Jun 2015 15:18:23 +0000 (15:18 +0000)]
some StmtExprs do not have side-effects
Differential Revision: http://reviews.llvm.org/D10211
llvm-svn: 239476
Benjamin Kramer [Wed, 10 Jun 2015 14:43:59 +0000 (14:43 +0000)]
[Hexagon] Make global arrays 'static const'. NFC.
llvm-svn: 239475
Scott Douglass [Wed, 10 Jun 2015 13:53:15 +0000 (13:53 +0000)]
add ConstEvaluatedExprVisitor
Differential Revision: http://reviews.llvm.org/D10210
llvm-svn: 239474
Igor Laevsky [Wed, 10 Jun 2015 13:24:00 +0000 (13:24 +0000)]
[Statepoints] Add test case to check that statepoint is marked with Throwable attribute.
Differential Revision: http://reviews.llvm.org/D10215
llvm-svn: 239473
Igor Laevsky [Wed, 10 Jun 2015 12:31:53 +0000 (12:31 +0000)]
[StatepointLowering] Reuse stack slots across basic blocks
During statepoint lowering we can sometimes avoid spilling of the value if we know that it was already spilled for previous statepoint.
We were doing this by checking if incoming statepoint value was lowered into load from stack slot. This was working only in boundaries of one basic block.
But instead of looking at the lowered node we can look directly at the llvm-ir value and if it was gc.relocate (or some simple modification of it) look up stack slot for it's derived pointer and reuse stack slot from it. This allows us to look across basic block boundaries.
Differential Revision: http://reviews.llvm.org/D10251
llvm-svn: 239472
Daniel Sanders [Wed, 10 Jun 2015 12:11:26 +0000 (12:11 +0000)]
Replace string GNU Triples with llvm::Triple in MCSubtargetInfo and create*MCSubtargetInfo(). NFC.
Summary:
This continues the patch series to eliminate StringRef forms of GNU triples
from the internals of LLVM that began in r239036.
Reviewers: rafael
Reviewed By: rafael
Subscribers: rafael, ted, jfb, llvm-commits, rengolin, jholewinski
Differential Revision: http://reviews.llvm.org/D10311
llvm-svn: 239467