Justin Lebar [Fri, 30 Sep 2016 19:55:55 +0000 (19:55 +0000)]
[CUDA] Emit a warning if a CUDA host/device/global attribute is placed after '(...)'.
Summary:
This is probably the sane place for the attribute to go, but nvcc
specifically rejects it. Other GNU-style attributes are allowed in this
position (although judging from the warning it emits for
host/device/global, those attributes are applied to the lambda's
anonymous struct, not to the function itself).
It would be nice to have a FixIt message here, but doing so, or even
just getting the correct range for the attribute, including its '((' and
'))'s, is apparently Hard.
Reviewers: rnk
Subscribers: cfe-commits, tra
Differential Revision: https://reviews.llvm.org/D25115
llvm-svn: 282911
Justin Lebar [Fri, 30 Sep 2016 19:55:48 +0000 (19:55 +0000)]
[CUDA] Fix up MaybeParseGNUAttributes call used for out-of-place attributes on CUDA lambdas.
Summary: There's an overload that we can use to make this a bit cleaner.
Reviewers: rnk
Subscribers: cfe-commits, tra
Differential Revision: https://reviews.llvm.org/D25114
llvm-svn: 282910
Joerg Sonnenberger [Fri, 30 Sep 2016 19:55:37 +0000 (19:55 +0000)]
Turn LLVM_USE_OPROFILE into a 0/1 definition.
llvm-svn: 282909
Joerg Sonnenberger [Fri, 30 Sep 2016 19:54:25 +0000 (19:54 +0000)]
Turn LLVM_USE_INTEL_JITEVENTS into a 0/1 definition.
llvm-svn: 282908
Joerg Sonnenberger [Fri, 30 Sep 2016 19:52:27 +0000 (19:52 +0000)]
Turn LLVM_ENABLE_ABI_BREAKING_CHECKS into a 0/1 definition like
LLVM_ENABLE_THREADS. Include llvm-config.h explicitly in headers to make
sure that the definition is available.
llvm-svn: 282907
Sanjay Patel [Fri, 30 Sep 2016 19:49:22 +0000 (19:49 +0000)]
[InstCombine] allow non-splat folds of select cond (ext X), C
llvm-svn: 282906
Etienne Bergeron [Fri, 30 Sep 2016 19:37:11 +0000 (19:37 +0000)]
[compiler-rt] Fix interception of crt atoll on win10 CRT.
Summary:
The check-asan-dynamic tests were broken on win10 because the interception
library was not able to hook on some functions.
credits: thanks sebastian marchand to help debugging this on win10.
Reviewers: rnk
Subscribers: chrisha, llvm-commits, dberris
Differential Revision: https://reviews.llvm.org/D25120
llvm-svn: 282904
Dehao Chen [Fri, 30 Sep 2016 19:25:23 +0000 (19:25 +0000)]
Revert test change in r282894 as it's broken in some platforms.
llvm-svn: 282903
Gor Nishanov [Fri, 30 Sep 2016 19:24:19 +0000 (19:24 +0000)]
[Coroutines] Part15c: Fix coro-split to correctly handle definitions between coro.save and coro.suspend
Summary:
In the case below, %Result.i19 is defined between coro.save and coro.suspend and used after coro.suspend. We need to correctly place such a value into the coroutine frame.
```
%save = call token @llvm.coro.save(i8* null)
%Result.i19 = getelementptr inbounds %"struct.lean_future<int>::Awaiter", %"struct.lean_future<int>::Awaiter"* %ref.tmp7, i64 0, i32 0
%suspend = call i8 @llvm.coro.suspend(token %save, i1 false)
switch i8 %suspend, label %exit [
i8 0, label %await.ready
i8 1, label %exit
]
await.ready:
%val = load i32, i32* %Result.i19
```
Reviewers: majnemer
Subscribers: llvm-commits, mehdi_amini
Differential Revision: https://reviews.llvm.org/D24418
llvm-svn: 282902
Sanjay Patel [Fri, 30 Sep 2016 19:15:41 +0000 (19:15 +0000)]
[InstCombine] add tests for non-splat select(ext)
llvm-svn: 282901
Martin Storsjo [Fri, 30 Sep 2016 19:13:46 +0000 (19:13 +0000)]
[MS] Implement __iso_volatile loads/stores as builtins
These are supposed to produce the same as normal volatile
pointer loads/stores. When -volatile:ms is specified,
normal volatile pointers are forced to have atomic semantics
(as is the default on x86 in MSVC mode). In that case,
these builtins should still produce non-atomic volatile
loads/stores without acquire/release semantics, which
the new test verifies.
These are only available on ARM (and on AArch64,
although clang doesn't support AArch64/Windows yet).
This implements what is missing for PR30394, making it possible
to compile C++ for ARM in MSVC mode with MSVC headers.
Differential Revision: https://reviews.llvm.org/D24986
llvm-svn: 282900
Gor Nishanov [Fri, 30 Sep 2016 19:05:06 +0000 (19:05 +0000)]
[Coroutines] Part15b: Fix dbg information handling in coro-split.
Summary:
Without the fix, if there was a function inlined into the coroutine with debug information, CloneFunctionInto(NewF, &F, VMap, /*ModuleLevelChanges=*/true, Returns); would duplicate all of the debug information including the DICompileUnit.
We know use VMap to indicate that debug metadata for a File, Unit and FunctionType should not be duplicated when we creating clones that will become f.resume, f.destroy and f.cleanup.
Reviewers: majnemer
Subscribers: mehdi_amini, llvm-commits
Differential Revision: https://reviews.llvm.org/D24417
llvm-svn: 282899
Sean Callanan [Fri, 30 Sep 2016 18:44:43 +0000 (18:44 +0000)]
Fixed several i386 Objective-C tests by completing objects, not their pointers.
llvm-svn: 282898
Gor Nishanov [Fri, 30 Sep 2016 18:41:35 +0000 (18:41 +0000)]
[Coroutines] Part 15a: Lower coro.subfn.addr in CoroCleanup
Summary: Not all coro.subfn.addr intrinsics can be eliminated in CoroElide through devirtualization. Those that remain need to be lowered in CoroCleanup.
Reviewers: majnemer
Subscribers: llvm-commits, mehdi_amini
Differential Revision: https://reviews.llvm.org/D24412
llvm-svn: 282897
Sanjay Patel [Fri, 30 Sep 2016 18:37:34 +0000 (18:37 +0000)]
clean up tests and auto-generate checks
llvm-svn: 282896
Michal Gorny [Fri, 30 Sep 2016 18:34:23 +0000 (18:34 +0000)]
cmake: Install the OCaml libraries into a more correct path
Add a OCAML_INSTALL_PATH variable that can be used to control
the install path for OCaml libraries. The new variable defaults to
${OCAML_STDLIB_PATH}, i.e. the OCaml library path obtained from
the OCaml compiler. Install libraries into "llvm" subdirectory.
This fixes two issues:
1. OCaml library directories differ between systems, and 'lib/ocaml' is
incorrect e.g. on amd64 Gentoo where OCaml is installed
in 'lib64/ocaml'. Therefore, obtain the library path from the OCaml
compiler using 'ocamlc -where' (which is already used to set
OCAML_STDLIB_PATH), which is the method used commonly in OCaml packages.
2. The top-level directory is reserved for the standard library, and has
precedence over local directory in search path. As a result, OCaml
preferred the files installed along with previous LLVM version over the
source tree when building a new version, resulting in two versions being
mixed during the build. The new layout is used commonly by other OCaml
packages, and findlib is able to find the LLVM libraries successfully.
Bug: https://bugs.gentoo.org/559134
Bug: https://bugs.gentoo.org/559624
Differential Revision: https://reviews.llvm.org/D24354
llvm-svn: 282895
Dehao Chen [Fri, 30 Sep 2016 18:30:04 +0000 (18:30 +0000)]
Update loop unroller cost model to make sure debug info does not affect optimization decisions.
Summary: Debug info should *not* affect optimization decisions. This patch updates loop unroller cost model to make it not affected by debug info.
Reviewers: davidxl, mzolotukhin
Subscribers: haicheng, llvm-commits, mzolotukhin
Differential Revision: https://reviews.llvm.org/D25098
llvm-svn: 282894
Michael Kruse [Fri, 30 Sep 2016 18:29:37 +0000 (18:29 +0000)]
[CodeGen] Add assertion for indirect array index expression generation. NFC.
Currently Polly cannot generate code for index expressions if the base pointer
is computed within the scop. The base pointer must be generated as well, but
there is no code that triggers that.
Add an assertion to detect when this would occur and miscompile. The IR verifier
should catch it as well.
llvm-svn: 282893
Kostya Serebryany [Fri, 30 Sep 2016 18:16:16 +0000 (18:16 +0000)]
[libFuzzer] add a fuzzer test that finds CVE-2015-3193
llvm-svn: 282892
Sanjay Patel [Fri, 30 Sep 2016 18:10:14 +0000 (18:10 +0000)]
[InstCombine] add tests for select X, (ext X), C
llvm-svn: 282891
Derek Schuff [Fri, 30 Sep 2016 18:02:54 +0000 (18:02 +0000)]
[WebAssembly] Make register stackification more conservative
Register stackification currently checks VNInfo for changes. Make that
more accurate by testing each intervening instruction for any other defs
to the same virtual register.
Patch by Jacob Gravelle
Differential Revision: https://reviews.llvm.org/D24942
llvm-svn: 282886
Rui Ueyama [Fri, 30 Sep 2016 17:56:20 +0000 (17:56 +0000)]
Improve error check for an empty archive.
Previously, it warned on any archive file that has no symbol.
It turned out that that is too noisy.
With this patch, it warns on such archive file that contains no file.
Differential Revision: https://reviews.llvm.org/D25111
llvm-svn: 282885
Rui Ueyama [Fri, 30 Sep 2016 17:54:31 +0000 (17:54 +0000)]
[Object] Define Archive::isEmpty().
llvm-svn: 282884
Michael Kruse [Fri, 30 Sep 2016 17:47:39 +0000 (17:47 +0000)]
[Support] Complete ISL annotations to IslPtr<>. NFC.
Add missing __isl_(give/take/keep) annotations to IslPtr<> and NonowningIslPtr<>
methods.
Because IslPtr's constructor's annotation would depend on the TakeOwnership
parameter, the parameter has been removed. Caller must copy the object
themselves if the do not want to take ownership.
llvm-svn: 282883
Etienne Bergeron [Fri, 30 Sep 2016 17:47:34 +0000 (17:47 +0000)]
[compiler-rt] Add support for the dynamic shadow allocation
Summary:
This patch is adding support for dynamic shadow allocation.
This is a merge and re-commit of the following patches.
```
[compiler-rt] Fix Asan build on Android
https://reviews.llvm.org/D24768
[compiler-rt] Add support for the dynamic shadow allocation
https://reviews.llvm.org/D23363
```
This patch needed to re-land at the same time:
```
[asan] Support dynamic shadow address instrumentation
https://reviews.llvm.org/D23354
```
Reviewers: rnk, zaks.anna
Subscribers: tberghammer, danalbert, kubabrecka, dberris, chrisha, llvm-commits
Differential Revision: https://reviews.llvm.org/D25104
llvm-svn: 282882
Etienne Bergeron [Fri, 30 Sep 2016 17:46:32 +0000 (17:46 +0000)]
[asan] Support dynamic shadow address instrumentation
Summary:
This patch is adding the support for a shadow memory with
dynamically allocated address range.
The compiler-rt needs to export a symbol containing the shadow
memory range.
This is required to support ASAN on windows 64-bits.
Reviewers: kcc, rnk, vitalybuka
Subscribers: zaks.anna, kubabrecka, dberris, llvm-commits, chrisha
Differential Revision: https://reviews.llvm.org/D23354
llvm-svn: 282881
Justin Lebar [Fri, 30 Sep 2016 17:14:53 +0000 (17:14 +0000)]
[CUDA] Make lambdas inherit __host__ and __device__ attributes from the scope in which they're created.
Summary: NVCC compat. Fixes bug 30567.
Reviewers: tra
Subscribers: cfe-commits, rnk
Differential Revision: https://reviews.llvm.org/D25105
llvm-svn: 282880
Justin Lebar [Fri, 30 Sep 2016 17:14:48 +0000 (17:14 +0000)]
[CUDA] Handle attributes on CUDA lambdas appearing between [...] and (...).
Summary: This is ugh, but it makes us compatible with NVCC. Fixes bug 26341.
Reviewers: rnk
Subscribers: cfe-commits, tra
Differential Revision: https://reviews.llvm.org/D25103
llvm-svn: 282879
Justin Lebar [Fri, 30 Sep 2016 17:14:44 +0000 (17:14 +0000)]
[CUDA] Add missing comment on Sema::CheckCUDAVLA.
llvm-svn: 282878
Konstantin Zhuravlyov [Fri, 30 Sep 2016 17:01:40 +0000 (17:01 +0000)]
[AMDGPU] Choose VMCNT, EXPCNT, LGKMCNT masks and shifts based on the isa version
Differential Revision: https://reviews.llvm.org/D24973
llvm-svn: 282877
Michal Gorny [Fri, 30 Sep 2016 16:56:16 +0000 (16:56 +0000)]
[test] Support 'lit' executable name
Support finding lit as plain 'lit', which is the name used by setup.py
in LLVM's utils/lit.
Differential Revision: https://reviews.llvm.org/D25072
llvm-svn: 282876
Konstantin Zhuravlyov [Fri, 30 Sep 2016 16:50:36 +0000 (16:50 +0000)]
[AMDGPU] Ask subtarget if waitcnt instruction is needed before barrier instruction
Differential Revision: https://reviews.llvm.org/D24985
llvm-svn: 282875
Michael Kruse [Fri, 30 Sep 2016 16:47:43 +0000 (16:47 +0000)]
[Support] Compile fix for gcc. NFC.
gcc 5.4 insists on template specialization to be in a namespace polly { ... }
block, instead of being prefixed with 'polly::'. Error message:
root/src/llvm/tools/polly/lib/Support/GICHelper.cpp:203:54: error: specialization of ‘template<class T> void polly::IslPtr<T>::dump() const’ in different namespace [-fpermissive]
template <> void polly::IslPtr<isl_##TYPE>::dump() const { \
^
msvc14 and clang 3.8 did not complain.
llvm-svn: 282874
Konstantin Zhuravlyov [Fri, 30 Sep 2016 16:39:24 +0000 (16:39 +0000)]
[AMDGPU] Do not run scalar optimization passes at "-O0"
Differential Revision: https://reviews.llvm.org/D25055
llvm-svn: 282873
Artur Pilipenko [Fri, 30 Sep 2016 16:20:08 +0000 (16:20 +0000)]
CVP. Turn marking adds as no wrap on by default (was turned off by 279082)
With 282650 in tree extra no wrap on adds doesn't cause regressions anymore. Reenable the optimzation.
llvm-svn: 282872
Adrian McCarthy [Fri, 30 Sep 2016 16:11:42 +0000 (16:11 +0000)]
Add namespace qualifiers for UTF functions that just moved.
llvm-svn: 282871
Michael Kruse [Fri, 30 Sep 2016 16:10:19 +0000 (16:10 +0000)]
[Support] Add (Nonowning-)IslPtr::dump(). NFC.
The dump() methods can be called from a debugger instead of e.g.
isl_*_dump(Var.Obj)
where Var is a variable of type IslPtr/NonowningIslPtr. To ensure that the
existence of the function pointers do not depdend on whether the methods are
used somwhere, they are declared with external linkage.
llvm-svn: 282870
Sean Callanan [Fri, 30 Sep 2016 16:02:28 +0000 (16:02 +0000)]
Skip TestRuntimeIvars on i386; the Objective-C V1 runtime doesn't list ivars.
llvm-svn: 282869
Jonathan Peyton [Fri, 30 Sep 2016 15:50:14 +0000 (15:50 +0000)]
Fix incorrect OpenMP version in Fortran module.
Add check for "45" version to use "201511" string for OpenMP 4.5,
otherwise "200505" is used in Fortran module. Also, fix kmp_openmp_version
variable (used for the debugger, e.g.) and kmp_version_omp_api that is used
in KMP_VERSION=1 output.
Patch by Olga Malysheva
Differential Revision: https://reviews.llvm.org/D24761
llvm-svn: 282868
Zachary Turner [Fri, 30 Sep 2016 15:43:59 +0000 (15:43 +0000)]
Resubmit "Add llvm::enumerate() to STLExtras."
The CL was originally failing due to the use of some C++14
specific features, so I've removed those. Hopefully this will
satisfy the bots.
llvm-svn: 282867
Dimitar Vlahovski [Fri, 30 Sep 2016 15:41:33 +0000 (15:41 +0000)]
Again fixing windows build breakage like in rL282862
llvm-svn: 282866
Samuel Antao [Fri, 30 Sep 2016 15:34:19 +0000 (15:34 +0000)]
[CUDA][OpenMP] Add a generic offload action builder
Summary:
This patch proposes a new class to generate and record action dependences related with offloading. The builder provides three main functionalities:
- Add device dependences to host actions.
- Add host dependence to device actions.
- Register device top-level actions.
The constructor of the builder detect the programming models that should be supported, and generates a specialized builder for each. If a new programming model is to be added in the future, only a new specialized builder has to be implemented.
When the specialized builder is generated, it produces programming-model-specific diagnostics.
A CUDA specialized builder is proposed in the patch that mostly consists of the partition of the current `buildCudaAction` by the three different functionalities.
Reviewers: tra, echristo, ABataev, jlebar, hfinkel
Subscribers: Hahnfeld, whchung, guansong, jlebar, mehdi_amini, andreybokhanko, tcramer, mkuron, cfe-commits, arpith-jacob, carlo.bertolli, caomhin
Differential Revision: https://reviews.llvm.org/D18172
llvm-svn: 282865
Michael Kruse [Fri, 30 Sep 2016 15:29:43 +0000 (15:29 +0000)]
[Support] Call isl_*_free() only on non-null pointers. NFC.
Add a non-NULL check before calling the free function into functions that are
supposed to be inlined. First, this is a form of partial inlining of the free
function, namely the nullptr test that free has to do. Secondly, and more
importantly, it allows the compiler to remove the call to isl_*_free() when it
knows that the object is nullptr, for instance because the last call is a
take(). "Consuming" the last use of an ISL object using take()
(instead of copy()) is a common pattern.
llvm-svn: 282864
Matthew Simpson [Fri, 30 Sep 2016 15:13:52 +0000 (15:13 +0000)]
[LV] Build all scalar steps for non-uniform induction variables
When building the steps for scalar induction variables, we previously attempted
to determine if all the scalar users of the induction variable were uniform. If
they were, we would only emit the step corresponding to vector lane zero. This
optimization was too aggressive. We generally don't know the entire set of
induction variable users that will be scalar. We have
isScalarAfterVectorization, but this is only a conservative estimate of the
instructions that will be scalarized. Thus, an induction variable may have
scalar users that aren't already known to be scalar. To avoid emitting unused
steps, we can only check that the induction variable is uniform. This should
fix PR30542.
Reference: https://llvm.org/bugs/show_bug.cgi?id=30542
llvm-svn: 282863
Dimitar Vlahovski [Fri, 30 Sep 2016 14:36:17 +0000 (14:36 +0000)]
Fixing windows build breakage caused by rL282822
The breakage was because of the moving of the UTF functions to the llvm
namespace
llvm-svn: 282862
Michael Kruse [Fri, 30 Sep 2016 14:34:05 +0000 (14:34 +0000)]
[CodeGen] Change 'Scalar' to 'Array' in method names. NFC.
generateScalarLoad() and generateScalarStore() are used for explicit (MK_Array)
memory accesses, therefore the method names were misleading. The names also
were similar to generateScalarLoads() and generateScalarStores() (plural forms)
which indeed handle scalar accesses. Presumbly, they were originally named to
contrast VectorBlockGenerator::generateLoad().
Rename the two methods to generateArrayLoad(),
respectively generateArrayStore().
llvm-svn: 282861
Etienne Bergeron [Fri, 30 Sep 2016 14:31:39 +0000 (14:31 +0000)]
[compiler-rt][asan] Disable a broken test on windows 64-bits
Summary:
This test is broken on wndows 64-bit.
The interception library is not able to hook on the memchr functions.
Snippet of the function that is not hookable:
```
--- No source file -------------------------------------------------------------
000007FEFA1A18CD CC int 3
000007FEFA1A18CE CC int 3
000007FEFA1A18CF CC int 3
--- f:\dd\vctools\crt\vcruntime\src\string\amd64_arm_arm64\memchr.c ------------
while ( cnt && (*(unsigned char *)buf != (unsigned char)chr) ) {
000007FEFA1A18D0 4D 85 C0 test r8,r8
000007FEFA1A18D3 74 0D je memchr+12h (07FEFA1A18E2h)
000007FEFA1A18D5 38 11 cmp byte ptr [rcx],dl
000007FEFA1A18D7 74 09 je memchr+12h (07FEFA1A18E2h)
buf = (unsigned char *)buf + 1;
000007FEFA1A18D9 48 FF C1 inc rcx
cnt--;
000007FEFA1A18DC 49 83 E8 01 sub r8,1
000007FEFA1A18E0 75 F3 jne memchr+5h (07FEFA1A18D5h)
}
```
Reviewers: rnk
Subscribers: kubabrecka, dberris, llvm-commits, chrisha
Differential Revision: https://reviews.llvm.org/D25037
llvm-svn: 282860
Etienne Bergeron [Fri, 30 Sep 2016 14:29:43 +0000 (14:29 +0000)]
[compiler-rt][asan] Fix incorrect regexp breaking an asan unittest on win64
Summary:
On windows, the memcpy and memmove function can be the same.
This is correcly detected when hooking, but it's not possible
to report the right function name when doing symbolisation.
The same fix was applied for the static asan unittest.
We forgot to apply the fix for the dynamic asan tests.
```
lvm\projects\compiler-rt\test\asan/TestCases/Windows/.svn/text-base/intercept_memcpy.cc.svn-base:// CHECK-NEXT: __asan_{{.*}}mem{{.*}}
```
This patch is fixing this test (win64):
```
ddressSanitizer-x86_64-windows-dynamic :: TestCases/Windows/dll_intercept_memcpy_indirect.cc
```
Reviewers: rnk, vitalybuka
Subscribers: llvm-commits, kubabrecka, chrisha, dberris
Differential Revision: https://reviews.llvm.org/D25038
llvm-svn: 282859
Simon Pilgrim [Fri, 30 Sep 2016 14:25:09 +0000 (14:25 +0000)]
Strip trailing whitespace (NFCI)
llvm-svn: 282858
Simon Pilgrim [Fri, 30 Sep 2016 14:18:06 +0000 (14:18 +0000)]
Fix int <= bool comparison warning on MSVC
llvm-svn: 282857
Dylan McKay [Fri, 30 Sep 2016 14:09:20 +0000 (14:09 +0000)]
[AVR] Add the ELF object file writer
Summary: This adds the ELF32 writer for AVR.
Reviewers: kparzysz
Subscribers: beanz, mgorny
Differential Revision: https://reviews.llvm.org/D25031
llvm-svn: 282856
Dylan McKay [Fri, 30 Sep 2016 14:05:15 +0000 (14:05 +0000)]
Revert "[RegAllocGreedy] Attempt to split unspillable live intervals"
It was accidentally committed.
llvm-svn: 282855
Dylan McKay [Fri, 30 Sep 2016 14:01:50 +0000 (14:01 +0000)]
[AVR] Add the assembly instruction printer
Summary:
This change adds the AVR assembly instruction printer.
No tests are included in this patch. I have left them downstream so we can
add them once `llc` successfully runs (there's very few components left
to upstream until this).
Reviewers: arsenm, kparzysz
Subscribers: wdng, beanz, mgorny
Differential Revision: https://reviews.llvm.org/D25028
llvm-svn: 282854
Michael Kruse [Fri, 30 Sep 2016 14:01:46 +0000 (14:01 +0000)]
[CodeGen] Add assertion for partial scalar accesses. NFC.
The code generator always adds unconditional LoadInst and StoreInst, hence the
MemoryAccess must be defined over all statement instances.
llvm-svn: 282853
Dylan McKay [Fri, 30 Sep 2016 13:59:20 +0000 (13:59 +0000)]
[RegAllocGreedy] Attempt to split unspillable live intervals
Summary:
Previously, when allocating unspillable live ranges, we would never
attempt to split. We would always bail out and try last ditch graph
recoloring.
This patch changes this by attempting to split all live intervals before
performing recoloring.
This fixes LLVM bug PR14879.
I can't add test cases for any backends other than AVR because none of
them have small enough register classes to trigger the bug.
Reviewers: qcolombet
Subscribers: MatzeB
Differential Revision: https://reviews.llvm.org/D25070
llvm-svn: 282852
George Rimar [Fri, 30 Sep 2016 13:46:34 +0000 (13:46 +0000)]
[ELF] - Inlined template argument variable. NFC.
llvm-svn: 282851
George Rimar [Fri, 30 Sep 2016 13:44:02 +0000 (13:44 +0000)]
[ELF] - Removed excessive variables. NFC.
llvm-svn: 282850
Pavel Labath [Fri, 30 Sep 2016 12:12:15 +0000 (12:12 +0000)]
XFAIL TestSBData for gcc-4.9 i386
test broken in r282659.
llvm-svn: 282849
Pavel Labath [Fri, 30 Sep 2016 11:47:54 +0000 (11:47 +0000)]
Fix PDB unittests on non-windows platforms
llvm r282788 changed how the presence of windows DIA SDK is signalled. Adjust to
that.
llvm-svn: 282848
Rafael Espindola [Fri, 30 Sep 2016 06:48:09 +0000 (06:48 +0000)]
Fix another --gc-sections crash.
The missing case was when a merge section was only referenced from
non-alloca sections.
llvm-svn: 282847
Victor Leschuk [Fri, 30 Sep 2016 06:39:48 +0000 (06:39 +0000)]
Cosmetic fix: deleted unnecessary line break in comment.
llvm-svn: 282846
Rafael Espindola [Fri, 30 Sep 2016 06:37:29 +0000 (06:37 +0000)]
Fix --gc-sections crash.
We would crash when a non-alloca section pointed to a gced part of a
merge section.
That can happen when a C/c++ constant in put in a merge section and
debug info is present.
llvm-svn: 282845
Petr Hosek [Fri, 30 Sep 2016 06:29:28 +0000 (06:29 +0000)]
[CMake] Support symlinks even with LLVM_INSTALL_TOOLCHAIN_ONLY
When LLVM_INSTALL_TOOLCHAIN_ONLY is used and LLVM_TOOLCHAIN_TOOLS
contains a tool which is a symlink, it would be ignored. This already
worked before but got broken in r282510.
Differential Revision: https://reviews.llvm.org/D25067
llvm-svn: 282844
Craig Topper [Fri, 30 Sep 2016 05:35:47 +0000 (05:35 +0000)]
[AVX-512] Store address operand should be an input operand for the special stack spilling pseudos for XMM16-31 and YMM16-31 without VLX.
llvm-svn: 282843
Craig Topper [Fri, 30 Sep 2016 05:35:45 +0000 (05:35 +0000)]
[AVX-512] Add the special stack spilling pseudos for XMM16-31 and YMM16-31 without VLX to teh isFrameLoadOpcode and isFrameStoreOpcode.
llvm-svn: 282842
Craig Topper [Fri, 30 Sep 2016 05:35:42 +0000 (05:35 +0000)]
Revert r282835 "[AVX-512] Always use the full 32 register vector classes for addRegisterClass regardless of whether AVX512/VLX is enabled or not."
Turns out this doesn't pass verify-machineinstrs.
llvm-svn: 282841
Kostya Serebryany [Fri, 30 Sep 2016 05:15:45 +0000 (05:15 +0000)]
[libfuzzer] test for c-ares CVE-2016-5180
llvm-svn: 282839
Adam Nemet [Fri, 30 Sep 2016 04:56:25 +0000 (04:56 +0000)]
[LDist] Port to new streaming API for opt remarks
llvm-svn: 282838
Eric Liu [Fri, 30 Sep 2016 04:32:39 +0000 (04:32 +0000)]
[change-namespace] fix namespace specifier of global variables.
Reviewers: hokein
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D24963
llvm-svn: 282837
Craig Topper [Fri, 30 Sep 2016 04:31:37 +0000 (04:31 +0000)]
[X86] Add AVX-512 VTs to findRepresentativeClass as well as v16i16 which was also missing. Change register class to include the extra 16 AVX512 registers.
I'm not completely sure what this method does or why all the 256-bit VTs returned VR128RegClass when the comments on the method definiton say it should return the largest super register class. I just figured AVX-512 should be similar.
llvm-svn: 282836
Craig Topper [Fri, 30 Sep 2016 04:31:33 +0000 (04:31 +0000)]
[AVX-512] Always use the full 32 register vector classes for addRegisterClass regardless of whether AVX512/VLX is enabled or not.
If AVX512 is disabled, the registers should already be marked reserved. Pattern predicates and register classes on instructions should take care of most of the rest. Loads/stores and physical register copies for XMM16-31 and YMM16-31 without VLX have already been taken care of.
I'm a little unclear why this changed the register allocation of the SSE2 run of the sad.ll test, but the registers selected appear to be valid after this change.
llvm-svn: 282835
Adam Nemet [Fri, 30 Sep 2016 03:44:16 +0000 (03:44 +0000)]
[LoopUnroll] Port to the new streaming interface for opt remarks.
llvm-svn: 282834
Piotr Padlewski [Fri, 30 Sep 2016 03:01:17 +0000 (03:01 +0000)]
[thinlto] Don't decay threshold for hot callsites
Summary:
We don't want to decay hot callsites to import chains of hot
callsites. The same mechanism is used in LIPO.
Reviewers: tejohnson, eraman, mehdi_amini
Subscribers: llvm-commits, mehdi_amini
Differential Revision: https://reviews.llvm.org/D24976
llvm-svn: 282833
Matt Arsenault [Fri, 30 Sep 2016 01:50:20 +0000 (01:50 +0000)]
AMDGPU: Use unsigned compare for eq/ne
For some reason there are both of these available, except
for scalar 64-bit compares which only has u64. I'm not sure
why there are both (I'm guessing it's for the one bit inputs we
don't use), but for consistency always using the
unsigned one.
llvm-svn: 282832
Kostya Serebryany [Fri, 30 Sep 2016 01:24:57 +0000 (01:24 +0000)]
[libFuzzer] remove the code for -print_pcs=1 with the old coverage. It still works with the new one (trace-pc-guard)
llvm-svn: 282831
Jim Ingham [Fri, 30 Sep 2016 01:23:46 +0000 (01:23 +0000)]
Add the tracking radar on our end.
llvm-svn: 282830
Kostya Serebryany [Fri, 30 Sep 2016 01:19:56 +0000 (01:19 +0000)]
[libFuzzer] more the feature set to InputCorpus; on feature update, change the feature counter of the old best input
llvm-svn: 282829
Adam Nemet [Fri, 30 Sep 2016 00:42:43 +0000 (00:42 +0000)]
[LoopDataPrefetch] Port to new streaming API for opt remarks
llvm-svn: 282826
Jason Molenda [Fri, 30 Sep 2016 00:41:15 +0000 (00:41 +0000)]
Add unit tests for simple frameless i386 and x86_64 function
instruction inspection to UnwindPlans.
llvm-svn: 282825
Enrico Granata [Fri, 30 Sep 2016 00:41:07 +0000 (00:41 +0000)]
Prefer skipping over x-failing
llvm-svn: 282824
Enrico Granata [Fri, 30 Sep 2016 00:40:04 +0000 (00:40 +0000)]
Do not assume we will be able to discover the return type of this selector call, for that is not true in i386 mode
llvm-svn: 282823
Justin Lebar [Fri, 30 Sep 2016 00:38:45 +0000 (00:38 +0000)]
Move UTF functions into namespace llvm.
Summary:
This lets people link against LLVM and their own version of the UTF
library.
I determined this only affects llvm, clang, lld, and lldb by running
$ git grep -wl 'UTF[0-9]\+\|\bConvertUTF\bisLegalUTF\|getNumBytesFor' | cut -f 1 -d '/' | sort | uniq
clang
lld
lldb
llvm
Tested with
ninja lldb
ninja check-clang check-llvm check-lld
(ninja check-lldb doesn't complete for me with or without this patch.)
Reviewers: rnk
Subscribers: klimek, beanz, mgorny, llvm-commits
Differential Revision: https://reviews.llvm.org/D24996
llvm-svn: 282822
Adam Nemet [Fri, 30 Sep 2016 00:29:30 +0000 (00:29 +0000)]
[LV] Port the remarks in processLoop to the new streaming API
This completes LV.
llvm-svn: 282821
Adam Nemet [Fri, 30 Sep 2016 00:29:25 +0000 (00:29 +0000)]
[LV] Port the last opt remark in Hints to the new streaming interface
llvm-svn: 282820
Reid Kleckner [Fri, 30 Sep 2016 00:17:49 +0000 (00:17 +0000)]
[X86] Don't preserve Win64 SSE CSRs when SSE is disabled
Code that doesn't use floating point and doesn't use SSE (kernel code)
shouldn't save and restore SSE registers.
Fixes PR30503
llvm-svn: 282819
Rafael Espindola [Fri, 30 Sep 2016 00:16:11 +0000 (00:16 +0000)]
Start linker scripts at 0.
We were implicitly creating space for the headers. That is not the
behaviour of bfd, which requires the script to use SIZEOF_HEADERS. The
difference is important for scripts that don't use SIZEOF_HEADERS and
expect the first section to be at 0.
llvm-svn: 282818
Quentin Colombet [Fri, 30 Sep 2016 00:10:00 +0000 (00:10 +0000)]
[AArch64][RegisterBankInfo] Use static mapping for 3-operands instrs.
This uses a TableGen'ed like structure for all 3-operands instrs.
The output of the RegBankSelect pass should be identical but the
RegisterBankInfo will do less dynamic allocations.
llvm-svn: 282817
Quentin Colombet [Fri, 30 Sep 2016 00:09:58 +0000 (00:09 +0000)]
[AArch64][RegisterBankInfo] Add static value mapping for 3-op instrs.
This is the kind of input TableGen should generate at some point.
NFC.
llvm-svn: 282816
Quentin Colombet [Fri, 30 Sep 2016 00:09:43 +0000 (00:09 +0000)]
[AArch64][RegisterBankInfo] Check the statically created ValueMapping.
Make sure that the ValueMappings contain the value we expect at the
indices we expect.
NFC.
llvm-svn: 282815
Rafael Espindola [Fri, 30 Sep 2016 00:06:24 +0000 (00:06 +0000)]
Don't depend on lld creating space for the headers.
Currently lld will implicitly reserve space for the headers. This is
not the case is bfd, where it is the script responsibility to use
SIZEOF_HEADERS. This means that a script not using SIZEOF_HEADERS and
expecting the address of the first section to be 0 would fail with lld.
I am fixing that is the next commit. This one just makes the tests
explicitly use SIZEOF_HEADERS to avoid the dependency on the current
behaviour.
llvm-svn: 282814
Adam Nemet [Fri, 30 Sep 2016 00:01:30 +0000 (00:01 +0000)]
[LAA, LV] Port to new streaming interface for opt remarks. Update LV
(Recommit after making sure IsVerbose gets properly initialized in
DiagnosticInfoOptimizationBase. See previous commit that takes care of
this.)
OptimizationRemarkAnalysis directly takes the role of the report that is
generated by LAA.
Then we need the magic to be able to turn an LAA remark into an LV
remark. This is done via a new OptimizationRemark ctor.
llvm-svn: 282813
Adam Nemet [Fri, 30 Sep 2016 00:01:27 +0000 (00:01 +0000)]
[Diag] Use non-static member initializer for IsVerbose. NFC
llvm-svn: 282812
Jason Molenda [Thu, 29 Sep 2016 23:57:33 +0000 (23:57 +0000)]
Switch to using TEST_EQ, TEST_FALSE where appropriate.
llvm-svn: 282811
Jim Ingham [Thu, 29 Sep 2016 23:48:21 +0000 (23:48 +0000)]
Add some logging when trace is on. We're getting a bot failure on i386 that doesn't
I can't reproduce locally. Hopefully this will help us catch the reason.
llvm-svn: 282810
Zachary Turner [Thu, 29 Sep 2016 23:05:41 +0000 (23:05 +0000)]
Revert "Add llvm::enumerate() to STLExtras."
This reverts commit r282804 as it seems to use some C++ features
that not all compilers support.
llvm-svn: 282809
Vitaly Buka [Thu, 29 Sep 2016 23:05:35 +0000 (23:05 +0000)]
Remove Google specific code.
llvm-svn: 282808
Rui Ueyama [Thu, 29 Sep 2016 23:04:50 +0000 (23:04 +0000)]
Add comments.
llvm-svn: 282807
Rafael Espindola [Thu, 29 Sep 2016 23:02:57 +0000 (23:02 +0000)]
Break long line. NFC.
llvm-svn: 282806
Vitaly Buka [Thu, 29 Sep 2016 23:00:54 +0000 (23:00 +0000)]
Don't use internal symbolizer if we are in process of reporting Out-of-Memory.
Reviewed by eugenis offline, as reviews.llvm.org is down.
llvm-svn: 282805
Zachary Turner [Thu, 29 Sep 2016 22:59:30 +0000 (22:59 +0000)]
Add llvm::enumerate() to STLExtras.
enumerate allows you to iterate over a range by pairing the
iterator's value with its index in the enumeration. This gives
you most of the benefits of using a for loop while still allowing
the range syntax.
llvm-svn: 282804