Commit Graph

291585 Commits

Author SHA1 Message Date
Ryan Prichard 1e216a59f6 [builtins] Delay emutls deallocation for one round
Summary:
With Android/Bionic, delay deallocation to round 2 of 4. It must run after
C++ thread_local destructors have been called, but before the final 2
rounds, because emutls calls free, and jemalloc then needs another 2
rounds to free its thread-specific data.

Fixes https://github.com/android-ndk/ndk/issues/687

Reviewers: cmtice, srhines, jyknight, chh, echristo

Reviewed By: srhines, chh, echristo

Subscribers: echristo, delcypher, llvm-commits, #sanitizers

Differential Revision: https://reviews.llvm.org/D46978

llvm-svn: 334463
2018-06-12 01:32:26 +00:00
Ryan Prichard 31e4a550a3 [builtins] emutls cleanup: determine header size using sizeof
Summary: Also add a few post-#else/#endif comments

Reviewers: echristo, srhines

Reviewed By: echristo

Subscribers: chh, delcypher, llvm-commits, #sanitizers

Differential Revision: https://reviews.llvm.org/D47861

llvm-svn: 334462
2018-06-12 01:32:25 +00:00
Michael Berg 95f3a430a8 NFC, some additional tests added and some renaming for planned fma support changes
llvm-svn: 334461
2018-06-12 00:52:43 +00:00
Craig Topper 957b738432 [X86] Add isel patterns for folding loads when creating ROUND instructions from ffloor/fnearbyint/fceil/frint/ftrunc.
We were missing packed isel folding patterns for all of sse41, avx, and avx512.

For some reason avx512 had scalar load folding patterns under optsize(due to partial/undef reg update), but we didn't have the equivalent sse41 and avx patterns.

Sometimes we would get load folding due to peephole pass anyway, but we're also missing avx512 instructions from the load folding table. I'll try to fix that in another patch.

Some of this was spotted in the review for D47993.

This patch adds all the folds to isel, adds a few spot tests, and disables the peephole pass on a few tests to ensure we're testing some of these patterns.

llvm-svn: 334460
2018-06-12 00:48:57 +00:00
Mark Searles 987f292c56 [AMDGPU] prevent hitting Assertion `isReg() && "Wrong MachineOperand accessor"'
The use iterator, used within findMaskOperands(), can return anything which is
not a def. isUse() requires a register, so check isReg() before calling isUse().

Differential Revision: https://reviews.llvm.org/D48047

llvm-svn: 334459
2018-06-12 00:41:26 +00:00
Matt Morehouse a4374ab222 [CMake] Remove -Wno-maybe-uninitialized from passthrough vars.
Fixes http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux/
buildbot.

llvm-svn: 334458
2018-06-12 00:37:06 +00:00
Yaxun Liu 6c10a66ec7 [CUDA][HIP] Set kernel calling convention before arrange function
Currently clang set kernel calling convention for CUDA/HIP after
arranging function, which causes incorrect kernel function type since
it depends on calling convention.

This patch moves setting kernel convention before arranging
function.

Differential Revision: https://reviews.llvm.org/D47733

llvm-svn: 334457
2018-06-12 00:16:33 +00:00
Jonas Devlieghere 9deaf68ed1 Exempt some compilers from new static variable test.
Apparently some compilers generate incomplete debug information which
caused the updated test to fail. Therefore I've extracted the new check
into a separate test case with the necessary decorators.

llvm-svn: 334456
2018-06-12 00:15:59 +00:00
Wei Mi 864ae8be95 Fix a buildbot error reported by sanitizer-x86_64-linux-fast:
Function::getGUID may complain if its input is an empty StringRef.

llvm-svn: 334455
2018-06-11 23:39:02 +00:00
Jonas Devlieghere 48f0c26902 [Test] Update static variable test.
Before Pavel's change in r334181, we were printing too many global
variables. This patch updates the test suite to ensure we don't regress
again in the future.

rdar://problem/29180927

llvm-svn: 334454
2018-06-11 23:26:15 +00:00
Wei Mi c6b96c8db2 Fix a warning issued by clang.
llvm-svn: 334453
2018-06-11 23:09:04 +00:00
Justin Bogner 0af77dae16 [CMake] Fix dropped dependency in install-llvm-headers
This dependency was accidentally dropped in r319480, causing
install-distribution and install-llvm-headers to install an incomplete
set of headers (the generated Intrinsics and Attributes would be
missing).

llvm-svn: 334452
2018-06-11 23:05:28 +00:00
George Burgess IV c72204d5b5 Simplify; NFC
Not shown in the diff: AQ is a `vector<SUnit *>`, and SU is a `SUnit *`

llvm-svn: 334451
2018-06-11 22:58:32 +00:00
Peter Wu 1c05c95739 [sanitizer] Add fgets, fputs and puts into sanitizer_common
Summary:
Add fgets, fputs and puts to sanitizer_common. This adds ASAN coverage
for these functions, extends MSAN support from fgets to fputs/puts and
extends TSAN support from puts to fputs.

Fixes: https://github.com/google/sanitizers/issues/952

Reviewed By: vitalybuka

Differential Revision: https://reviews.llvm.org/D46545

llvm-svn: 334450
2018-06-11 22:58:04 +00:00
Wei Mi 61db138bfe Fix a warning reported by clang but not by gcc.
llvm-svn: 334449
2018-06-11 22:51:28 +00:00
Yan Zhang 72cecf9a63 - Add "AV" as new default acronym. - Add support for "I" and "A" in lowerCamelCase pattern
Summary: Now we can support property names like "hasADog" correctly.

Reviewers: benhamilton, hokein

Reviewed By: benhamilton

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D48039

llvm-svn: 334448
2018-06-11 22:44:06 +00:00
Wei Mi a0c0857e7a [SampleFDO] Add a new compact binary format for sample profile.
Name table occupies a big chunk of size in current binary format sample profile.
In order to reduce its size, the patch changes the sample writer/reader to
save/restore MD5Hash of names in the name table. Sample annotation phase will
also use MD5Hash of name to query samples accordingly.

Experiment shows compact binary format can reduce the size of sample profile by
2/3 compared with binary format generally.

Differential Revision: https://reviews.llvm.org/D47955

llvm-svn: 334447
2018-06-11 22:40:43 +00:00
David Blaikie 19b22d406d TableGen: Change some pointer parameters to references since they're never null anyway
llvm-svn: 334446
2018-06-11 22:14:43 +00:00
Petr Hosek 5e1f49e89b [CMake] Use libc++ and compiler-rt for bootstrap Fuchsia Clang
We want to build the second stage compiler with libc++ and compiler-rt,
also include builtins and runtimes into extra bootstrap components to
ensure these get built.

Differential Revision: https://reviews.llvm.org/D47356

llvm-svn: 334445
2018-06-11 22:06:44 +00:00
Konstantin Zhuravlyov d70c367267 AMDGPU/LLD: Handle R_AMDGPU_REL64 relocation
Requires r334443 from llvm

Differential Revision: https://reviews.llvm.org/D47734

llvm-svn: 334444
2018-06-11 21:42:53 +00:00
Konstantin Zhuravlyov 3e5d66ac66 AMDGPU: Add 64-bit relative variant kind
Differential Revision: https://reviews.llvm.org/D47601

llvm-svn: 334443
2018-06-11 21:37:57 +00:00
Jason Molenda 808490138b Document how lldb uses the DBGSourcePathRemapping
source path remapping src/dest path pairs with
respect to the DBGVersion number in the plist.

llvm-svn: 334442
2018-06-11 21:36:40 +00:00
Jason Molenda c710f5d40a Add DebugNamesDWARFIndex.cpp.
llvm-svn: 334441
2018-06-11 21:35:36 +00:00
Matt Arsenault 5615fa0a87 DAG: Fix extract_subvector combine for a single element
This would fail before because 1x vectors aren't legal,
so instead just use the scalar type.

Avoids regressions in a future AMDGPU commit to add
v4i16/v4f16 as legal types.

Test update is just the one test that this triggers
on in tree now. It wasn't checking anything before.
The result is completely  changed since the selects
are eliminated. Not sure if it's considered better
or not.

llvm-svn: 334440
2018-06-11 21:27:41 +00:00
Leonard Mosescu e1bb51789d Add a new SBTarget::LoadCore() overload which surfaces errors if the load fails
There was no way to find out what's wrong if SBProcess SBTarget::LoadCore(const char *core_file) failed. 
Additionally, the implementation was unconditionally setting sb_process, so it wasn't even possible to check if the return SBProcess is valid.

This change adds a new overload which surfaces the errors and also returns a valid SBProcess only if the core load succeeds:

SBProcess SBTarget::LoadCore(const char *core_file, SBError &error);

Differential Revision: https://reviews.llvm.org/D48049

llvm-svn: 334439
2018-06-11 21:19:26 +00:00
Alex Shlyapnikov ed32baa84b [Sanitizers] %tool_options -> %env_tool_opts in allocator_returns_null.cc
Use proper substitution in the common allocator_returns_null.cc test.

llvm-svn: 334438
2018-06-11 21:04:24 +00:00
Petr Hosek 534a104317 [CMake] Allow specifying extra dependencies of bootstrap stage
This allows adding additional bootstrap dependencies to the bootstrap
compiler that may be needed by later stages.

Differential Revision: https://reviews.llvm.org/D47355

llvm-svn: 334437
2018-06-11 20:59:31 +00:00
Craig Topper 3efdb7ce19 [X86] Push some variable declarations down into the individual switch cases that need them. NFC
All of the cases are already wrapped in curly braces so declaring a variable there isn't an issue. And the variables aren't assigned or used in the larger scope.

llvm-svn: 334436
2018-06-11 20:50:58 +00:00
Farhana Aleen 078cd48a39 [SLP] Add testcases of min/max reduction pattern for AMDGPU.
Author: FarhanaAleen
llvm-svn: 334435
2018-06-11 20:29:31 +00:00
Tim Shen df2d6652c1 Fix incorrect CHECK-LABEL
llvm-svn: 334434
2018-06-11 19:56:12 +00:00
Alex Shlyapnikov 406385505d [Sanitizers] Move allocator_returns_null.cc test to common.
Summary:
Add allocator_returns_null.cc test to sanitizer_common and
remove all sanitizer-specific ones except:
- HWASan is not covered by sanitizer_common
- TSan allocator does not have comprehensive error reporting yet

Reviewers: vitalybuka

Subscribers: kubamracek, delcypher, #sanitizers, llvm-commits

Differential Revision: https://reviews.llvm.org/D47971

llvm-svn: 334433
2018-06-11 19:45:59 +00:00
Fangrui Song 2389a240d0 [ELF] Fix copy relocation when two symbols share the same Symbol instance.
In glibc libc.so.6, the multiple versions of sys_errlist share the same Symbol instance. When sys_errlist is copy relocated, we would replace SharedSymbol with Defined in the first iteration of the following loop:

  for (SharedSymbol *Sym : getSymbolsAt<ELFT>(SS))

Then in the second iteration, we think the symbol (which has been changed to Defined) is still SharedSymbol and screw up (the address ends up in the `Size` field).

llvm-svn: 334432
2018-06-11 19:42:57 +00:00
Volodymyr Sapsai d1df7b7ee7 Mark the test using <experimental/memory_resource> to require c++experimental.
When built against the old libc++ version the test was causing linker error

Undefined symbols for architecture x86_64:
  "std::experimental::fundamentals_v1::pmr::new_delete_resource()", referenced from:
      void test_evil<WidgetV0, WidgetV0>() in construct_piecewise_pair_evil.pass.cpp.o
      void test_evil<WidgetV0, WidgetV1>() in construct_piecewise_pair_evil.pass.cpp.o
      void test_evil<WidgetV0, WidgetV2>() in construct_piecewise_pair_evil.pass.cpp.o
      void test_evil<WidgetV0, WidgetV3>() in construct_piecewise_pair_evil.pass.cpp.o
      void test_evil<WidgetV1, WidgetV0>() in construct_piecewise_pair_evil.pass.cpp.o
      void test_evil<WidgetV1, WidgetV1>() in construct_piecewise_pair_evil.pass.cpp.o
      void test_evil<WidgetV1, WidgetV2>() in construct_piecewise_pair_evil.pass.cpp.o
      ...

llvm-svn: 334431
2018-06-11 19:42:27 +00:00
Craig Topper ceed99baf0 [X86] Reorder some type constraints to force things to be vectors and integer/fp before forcing them to be the same size.
This may be needed by another patch that I'm working on. It should have no effect on any of the generated outputs.

llvm-svn: 334430
2018-06-11 19:20:15 +00:00
Justin Lebar 4da41c13a5 [SCEV] Add transform zext((A * B * ...)<nuw>) --> (zext(A) * zext(B) * ...)<nuw>.
Reviewers: sanjoy

Subscribers: hiraditya, llvm-commits

Differential Revision: https://reviews.llvm.org/D48041

llvm-svn: 334429
2018-06-11 18:57:58 +00:00
Justin Lebar aa4fec94d8 [SCEV] Add nuw/nsw to mul ops in StrengthenNoWrapFlags where safe.
Summary:
Previously we would add them for adds, but not multiplies.

Reviewers: sanjoy

Subscribers: llvm-commits, hiraditya

Differential Revision: https://reviews.llvm.org/D48038

llvm-svn: 334428
2018-06-11 18:57:42 +00:00
Justin Lebar 7b4656c1d3 Fix indentation in ScalarEvolution.cpp.
Whitespace-only change.  (clang-formatted the whole block.)

llvm-svn: 334427
2018-06-11 18:57:27 +00:00
Krzysztof Parzyszek dd9415d550 [Hexagon] Late predicate producers cannot be used as dot-new sources
llvm-svn: 334426
2018-06-11 18:45:52 +00:00
Tim Shen cc63761720 [SCEV] Canonicalize "A /u C1 /u C2" to "A /u (C1*C2)".
Summary: FWIW InstCombine already folds this. Also avoid the case where C1*C2 overflows.

Reviewers: sunfish, sanjoy

Subscribers: hiraditya, bixia, llvm-commits

Differential Revision: https://reviews.llvm.org/D47965

llvm-svn: 334425
2018-06-11 18:44:58 +00:00
Alex Shlyapnikov 789494026e [Sanitizers] Move pvalloc overflow tests to common.
Summary:
Now all sanitizers with improved allocator error reporting are covered
by these common tests.

Also, add pvalloc-specific checks to LSan.

HWASan is not covered by sanitizer_common, hence its own pvalloc
and other allocator tests.

Reviewers: vitalybuka

Subscribers: srhines, kubamracek, delcypher, #sanitizers, llvm-commits

Differential Revision: https://reviews.llvm.org/D47970

llvm-svn: 334424
2018-06-11 17:33:53 +00:00
Simon Pilgrim 14ee66ef37 [X86][AVX512] Tag AVX5124FMAPS/AVX5124VNNIW with missing scheduler classes
Necessary for D46276 as even though btver2 doesn't use these instructions, its now flagged as complete so complains if ANY instruction isn't tagged.....

UnsupportedFeatures wouldn't help here as these instructions don't appear to have a feature predicate (like a lot of AVX512).

llvm-svn: 334423
2018-06-11 17:28:00 +00:00
Craig Topper 201b9dd334 [X86] Fix operand order in the shuffle created for blend builtins.
This was broken when the builtin was added in r334249.

llvm-svn: 334422
2018-06-11 17:06:01 +00:00
Matt Morehouse 3416773cb1 [clang-fuzzer] Modified protobuf and converter to add new signature, remove conditionals.
Changed the function signature and removed conditionals from loop body.

Patch By:  emmettneyman

Differential Revision: https://reviews.llvm.org/D47964

llvm-svn: 334421
2018-06-11 17:05:45 +00:00
Stanislav Mekhanoshin 7ba3fc730c [AMDGPU] Do not consider indirect acces through phi for wave limiter
Rational: if there is indirect access that is usually an issue
because load is not ready by the use. However, if use is inside a
loop and load is outside that is potentially an issue for a first
iteration only.

Differential Revision: https://reviews.llvm.org/D47740

llvm-svn: 334420
2018-06-11 16:50:49 +00:00
Aleksandar Beserminji 62cf9d21ab [mips] Fix spill slot for mips3, n64 abi
When program is compiled for mips3 with n64 abi, wrong register class
is used for creating an emergency spill slot. This patch fixes the
correct register class to be chosen.

This patch resolves PR35859.

Thanks to John Baldwin for reporting the issue!

Differential Revision: https://reviews.llvm.org/D47938

llvm-svn: 334419
2018-06-11 16:50:28 +00:00
Reid Kleckner 03ab0fccc1 Enable crash recovery tests on Windows, globs work in the lit internal shell now
llvm-svn: 334418
2018-06-11 16:50:07 +00:00
Reid Kleckner 3513fdcc0f [MS] Use mangled names and comdats for string merging with ASan
This should reduce the binary size penalty of ASan on Windows. After
r334313, ASan will add red zones to globals in comdats, so we will still
find OOB accesses to string literals.

llvm-svn: 334417
2018-06-11 16:49:43 +00:00
Craig Topper ae7179414d [X86] Properly account for the immediate being multiplied by 8 in the immediate range checking for BI__builtin_ia32_psrldqi128 and friends.
The limit was set to 1023 which only up to 127*8. It needs to be 2047 to allow 255*8.

llvm-svn: 334416
2018-06-11 16:34:10 +00:00
Martin Probst c8b7a41a00 clang-format: [JS] strict prop init annotation.
Summary:
TypeScript uses the `!` token for strict property initialization
assertions, as in:

    class X {
      strictPropAsserted!: string;
    }

Previously, clang-format would wrap between the `!` and the `:` for
overly long lines. This patch fixes that by generally preventing the
wrap in that location.

Reviewers: krasimir

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D48030

llvm-svn: 334415
2018-06-11 16:20:13 +00:00
Mikhail Maltsev 5434047862 [Driver] Add aliases for -Qn/-Qy
This patch adds aliases for -Qn (-fno-ident) and -Qy (-fident) which
look less cryptic than -Qn/-Qy. The aliases are compatible with GCC.

Differential Revision: https://reviews.llvm.org/D48021

llvm-svn: 334414
2018-06-11 16:10:06 +00:00