Dylan McKay [Mon, 10 Oct 2016 01:28:45 +0000 (01:28 +0000)]
[AVR] Enable generation of the TableGen assembly writer tables
This also changes the order of the statements in CMakeLists.txt to be
alphabetical.
llvm-svn: 283711
Brian Gesiak [Mon, 10 Oct 2016 01:22:06 +0000 (01:22 +0000)]
[lit] Remove (or allow specific) unused imports
Summary:
Using Python linter flake8 on the utils/lit reveals several linter
warnings designated "F401: Unused import". Fix or silence these
warnings.
Some of these unused imports are legitimate, while some are part of lit's API.
For example, users of lit expect to be able to access `lit.formats.ShTest` in
their `lit.cfg`, despite the module hierarchy for that symbol actually being
`lit.formats.shtest.ShTest`. To silence linter errors for these lines,
include a "noqa" directive.
Reviewers: echristo, delcypher, beanz, ddunbar
Subscribers: mehdi_amini, llvm-commits
Differential Revision: https://reviews.llvm.org/D25407
llvm-svn: 283710
Brian Gesiak [Mon, 10 Oct 2016 01:20:43 +0000 (01:20 +0000)]
[lit] Remove unused TestingProgressDisplay attr
Summary:
`TestingProgressDisplay` initializes its `current` attribute to `None`, but
never reads or writes the value again. Remove it.
Reviewers: echristo, delcypher, beanz, ddunbar
Subscribers: llvm-commits, mehdi_amini
Differential Revision: https://reviews.llvm.org/D25415
llvm-svn: 283709
Brian Gesiak [Mon, 10 Oct 2016 01:19:27 +0000 (01:19 +0000)]
[lit] Fix undefined symbol ArgumentError
Summary:
`ArgumentError` is not defined by the Python standard library.
Executing this line of code would throw a exception, but not the
intended one. It would throw a `NameError` exception, since `ArgumentError`
is undefined.
Use `ValueError` instead, which is defined by the Python standard
library.
Reviewers: echristo, delcypher, beanz, ddunbar
Subscribers: llvm-commits, mehdi_amini
Differential Revision: https://reviews.llvm.org/D25410
llvm-svn: 283708
Brian Gesiak [Mon, 10 Oct 2016 01:18:14 +0000 (01:18 +0000)]
[lit] Remove semicolons in Python code
Summary:
Semicolons aren't necessary as statement terminators in Python, and
each of these uses are superfluous as they appear at the end of a line.
The convention is to not use semicolons where not needed, so remove them.
Reviewers: echristo, delcypher, beanz, ddunbar
Subscribers: llvm-commits, mehdi_amini
Differential Revision: https://reviews.llvm.org/D25409
llvm-svn: 283707
Brian Gesiak [Mon, 10 Oct 2016 01:15:33 +0000 (01:15 +0000)]
[lit] Remove unused variable in googletest format
Summary: `prefix` is written to but never read.
Reviewers: echristo, delcypher, beanz, ddunbar
Subscribers: llvm-commits, mehdi_amini
Differential Revision: https://reviews.llvm.org/D25408
llvm-svn: 283706
Brian Gesiak [Mon, 10 Oct 2016 01:11:52 +0000 (01:11 +0000)]
[lit] Remove Python 2.6 and below exec workaround
Summary:
The minimum version of Python required to run LLVM's test suite is 2.7.
Remove a workaround for older Python versions.
Reviewers: echristo, delcypher, beanz, ddunbar
Subscribers: llvm-commits, mehdi_amini
Differential Revision: https://reviews.llvm.org/D25400
llvm-svn: 283705
Craig Topper [Sun, 9 Oct 2016 23:08:39 +0000 (23:08 +0000)]
[AVX-512] Port 128 and 256-bit memory->register sign/zero extend patterns from SSE file. Also add a minimal set for 512-bit.
llvm-svn: 283704
Craig Topper [Sun, 9 Oct 2016 23:08:33 +0000 (23:08 +0000)]
[X86] Remove redundant patterns. The same pattern appears a few lines up.
llvm-svn: 283703
Mehdi Amini [Sun, 9 Oct 2016 23:00:34 +0000 (23:00 +0000)]
Move the global variables representing each Target behind accessor function
This avoids "static initialization order fiasco"
Differential Revision: https://reviews.llvm.org/D25412
llvm-svn: 283702
Eric Fiselier [Sun, 9 Oct 2016 21:43:30 +0000 (21:43 +0000)]
Remove debug output from gen_linker_script.py
llvm-svn: 283701
Eric Fiselier [Sun, 9 Oct 2016 21:34:03 +0000 (21:34 +0000)]
Fix linker script generation for in-tree builds
llvm-svn: 283700
Eric Fiselier [Sun, 9 Oct 2016 20:44:48 +0000 (20:44 +0000)]
[libc++abi] Fix bug which cased the static libunwind to always be chosen
llvm-svn: 283699
Eric Fiselier [Sun, 9 Oct 2016 20:38:29 +0000 (20:38 +0000)]
[CMake] Correct configuration order of the sub-projects based on ther dependancies
llvm-svn: 283698
Davide Italiano [Sun, 9 Oct 2016 17:15:04 +0000 (17:15 +0000)]
[llvm-link] Fix description of -disable-lazy-loading option
Patch by Will Dietz!
llvm-svn: 283697
Zvi Rackover [Sun, 9 Oct 2016 13:33:51 +0000 (13:33 +0000)]
[X86] Adding the 'nounwind' attribute to test functions for cleaner generated code
Thanks to RKSimon for the suggestion.
llvm-svn: 283696
Zvi Rackover [Sun, 9 Oct 2016 13:07:25 +0000 (13:07 +0000)]
[X86] Improve the rotate ISel test
Summary:
- Added 64-bit target testing.
- Added 64-bit operand test cases.
- Added cases that demonstrate pr30644
Reviewers: RKSimon, craig.topper, igorb
Differential Revision: https://reviews.llvm.org/D25401
llvm-svn: 283695
Elena Demikhovsky [Sun, 9 Oct 2016 10:48:52 +0000 (10:48 +0000)]
DAG: Setting Masked-Expand-Load as a variant of Masked-Load node
Masked-expand-load node represents load operation that loads a variable amount of elements from memory according to amount of "true" bits in the mask and expands the loaded elements according to their position in the mask vector.
Right now, the node is used in intrinsics for VEXPAND* instructions.
The work is done towards implementation of masked.expandload and masked.compressstore intrinsics.
Differential Revision: https://reviews.llvm.org/D25322
llvm-svn: 283694
Tobias Grosser [Sun, 9 Oct 2016 08:31:35 +0000 (08:31 +0000)]
Fix formatting after recent cl:: changes
This fixes 'make check-polly'
llvm-svn: 283693
Craig Topper [Sun, 9 Oct 2016 06:41:47 +0000 (06:41 +0000)]
[AVX-512] Fix execution domain for EVEX encoded VINSERTPS.
llvm-svn: 283692
Peter Collingbourne [Sun, 9 Oct 2016 04:39:13 +0000 (04:39 +0000)]
MC: Remove unused entities.
llvm-svn: 283691
Peter Collingbourne [Sun, 9 Oct 2016 04:38:57 +0000 (04:38 +0000)]
Target: Remove unused entities.
llvm-svn: 283690
Craig Topper [Sun, 9 Oct 2016 03:54:13 +0000 (03:54 +0000)]
[AVX-512] Add subvector insert and extract to load/store folding tables.
llvm-svn: 283689
Craig Topper [Sun, 9 Oct 2016 03:54:09 +0000 (03:54 +0000)]
[AVX-512] Add avx512dq to the fp stack folding test.
llvm-svn: 283688
Craig Topper [Sun, 9 Oct 2016 03:54:05 +0000 (03:54 +0000)]
[AVX-512] Add the vector down convert instructions to the store folding tables.
llvm-svn: 283687
Kostya Serebryany [Sun, 9 Oct 2016 03:45:38 +0000 (03:45 +0000)]
[libFuzzer] make a test less flaky
llvm-svn: 283686
Hal Finkel [Sun, 9 Oct 2016 03:06:31 +0000 (03:06 +0000)]
When optimizing for size, enable loop rerolling by default
We have a loop-rerolling optimization which can be enabled by using
-freroll-loops. While sometimes loops are hand-unrolled for performance
reasons, when optimizing for size, we should always undo this manual
optimization to produce smaller code (our optimizer's unroller will still
unroll the rerolled loops if it thinks that is a good idea).
llvm-svn: 283685
Hal Finkel [Sun, 9 Oct 2016 02:49:31 +0000 (02:49 +0000)]
[CMake] Fix in-tree libcxxabi build support after r283659
r283659 changed the argument to gen_link_script.py from SCRIPT_ABI_LIBNAME to
LIBCXX_LIBRARIES_PUBLIC, assuming that all of the items in the
LIBCXX_LIBRARIES_PUBLIC list were library names. This is not right, however,
for in-tree libcxxabi builds, we might have the target name in this list. There
was special logic to fixup SCRIPT_ABI_LIBNAME for this situation; change it to
apply a similar fixup for LIBCXX_LIBRARIES_PUBLIC.
llvm-svn: 283684
Justin Lebar [Sun, 9 Oct 2016 00:27:39 +0000 (00:27 +0000)]
[CUDA] Don't install cuda_wrappers/{algorithm,complex} into the main include dir.
This is obviously wrong -- if we do this, then all compiles will pick up
these wrappers, which is not what we want.
llvm-svn: 283683
Kostya Serebryany [Sat, 8 Oct 2016 23:24:45 +0000 (23:24 +0000)]
[libFuzzer] when shrinking the corpus, delete evicted files previously created by the current process
llvm-svn: 283682
Mehdi Amini [Sat, 8 Oct 2016 22:49:28 +0000 (22:49 +0000)]
ThinLTO: Fix Gold test after caching fix in r283655
(I don't have Gold available, so this is speculative)
llvm-svn: 283681
Justin Lebar [Sat, 8 Oct 2016 22:16:12 +0000 (22:16 +0000)]
[CUDA] Support <complex> and std::min/max on the device.
Summary:
We do this by wrapping <complex> and <algorithm>.
Tests are in the test-suite.
Reviewers: tra
Subscribers: jhen, beanz, cfe-commits, mgorny
Differential Revision: https://reviews.llvm.org/D24979
llvm-svn: 283680
Justin Lebar [Sat, 8 Oct 2016 22:16:08 +0000 (22:16 +0000)]
[CUDA] Rename cuda_builtin_vars.h to __clang_cuda_builtin_vars.h.
Summary: This matches the idiom we use for our other CUDA wrapper headers.
Reviewers: tra
Subscribers: beanz, mgorny, cfe-commits
Differential Revision: https://reviews.llvm.org/D24978
llvm-svn: 283679
Justin Lebar [Sat, 8 Oct 2016 22:16:03 +0000 (22:16 +0000)]
[CUDA] Declare our __device__ math functions in the same inline namespace as our standard library.
Summary:
Currently we declare our inline __device__ math functions in namespace
std. But libstdc++ and libc++ declare these functions in an inline
namespace inside namespace std. We need to match this because, in a
later patch, we want to get e.g. <complex> to use our device overloads,
and it only will if those overloads are in the right inline namespace.
Reviewers: tra
Subscribers: cfe-commits, jhen
Differential Revision: https://reviews.llvm.org/D24977
llvm-svn: 283678
Justin Lebar [Sat, 8 Oct 2016 22:15:58 +0000 (22:15 +0000)]
[CUDA] Add #pragma clang force_cuda_host_device_{begin,end} pragmas.
Summary:
These cause us to consider all functions in-between to be __host__
__device__.
You can nest these pragmas; you just can't have more 'end's than
'begin's.
Reviewers: rsmith
Subscribers: tra, jhen, cfe-commits
Differential Revision: https://reviews.llvm.org/D24975
llvm-svn: 283677
Kostya Serebryany [Sat, 8 Oct 2016 22:12:14 +0000 (22:12 +0000)]
[libFuzzer] control the reload interval by a flag, make it 10 seconds by default
llvm-svn: 283676
Kostya Serebryany [Sat, 8 Oct 2016 21:57:48 +0000 (21:57 +0000)]
[libFuzzer] fix use-after-free in libFuzzer found by ... fuzzing.
llvm-svn: 283675
Simon Pilgrim [Sat, 8 Oct 2016 21:17:44 +0000 (21:17 +0000)]
[X86][SSE] Regenerate select tests
llvm-svn: 283674
Zvi Rackover [Sat, 8 Oct 2016 20:54:20 +0000 (20:54 +0000)]
Revert "[X86] Apply the Update LLC Test Checks tool on the rotate tests."
This reverts commit 283667.
llvm-svn: 283673
Simon Pilgrim [Sat, 8 Oct 2016 19:54:28 +0000 (19:54 +0000)]
[X86][SSE] Regenerate and add 32-bit tests to widening tests
llvm-svn: 283672
Mehdi Amini [Sat, 8 Oct 2016 19:41:06 +0000 (19:41 +0000)]
Turn cl::values() (for enum) from a vararg function to using C++ variadic template
The core of the change is supposed to be NFC, however it also fixes
what I believe was an undefined behavior when calling:
va_start(ValueArgs, Desc);
with Desc being a StringRef.
Differential Revision: https://reviews.llvm.org/D25342
llvm-svn: 283671
Simon Pilgrim [Sat, 8 Oct 2016 18:51:55 +0000 (18:51 +0000)]
Fix comment typos - full update script path in assertions note
llvm-svn: 283670
Craig Topper [Sat, 8 Oct 2016 18:50:00 +0000 (18:50 +0000)]
[AVX-512] Add test case for PR30430 that I should have added in r281959.
llvm-svn: 283669
Craig Topper [Sat, 8 Oct 2016 18:49:57 +0000 (18:49 +0000)]
[AVX-512] Fix a bug in getLargestLegalSuperClass where we inflated to VR128X/VR256X even when VLX isn't supported.
This seems to have been responsible for the XMM16-31 spills observed in PR29112. With this fixed the test case has been modified to no longer have a spill of XMM16.
llvm-svn: 283668
Zvi Rackover [Sat, 8 Oct 2016 18:44:47 +0000 (18:44 +0000)]
[X86] Apply the Update LLC Test Checks tool on the rotate tests.
Also added cases demonstrating pr30644.
llvm-svn: 283667
Simon Pilgrim [Sat, 8 Oct 2016 18:36:57 +0000 (18:36 +0000)]
[X86][AVX2] Regenerate and add 32-bit tests to core tests
llvm-svn: 283666
Colin LeMahieu [Sat, 8 Oct 2016 17:18:51 +0000 (17:18 +0000)]
[Hexagon] Adding change of flow max 1 (cofMax1) TS flag for marking this restriction rather than implying it from TypeJR.
llvm-svn: 283665
Teresa Johnson [Sat, 8 Oct 2016 16:11:42 +0000 (16:11 +0000)]
[ThinLTO] Record calls to aliases
Summary:
When there is a call to an alias in the same module, we were not
adding a call edge. So we could incorrectly think that the alias
was dead if it was inlined in that function, despite having a
reference imported elsewhere. This resulted in unsats at link time.
Add a call edge when the call is to an alias.
Reviewers: davide, mehdi_amini
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D25384
llvm-svn: 283664
Sebastian Pop [Sat, 8 Oct 2016 12:30:07 +0000 (12:30 +0000)]
[AArch64] Avoid generating indexed vector instructions for Exynos
Avoid generating indexed vector instructions for Exynos. This is needed for
fmla/fmls/fmul/fmulx. For example, the instruction
fmla v0.4s, v1.4s, v2.s[1]
is less efficient than the instructions
dup v2.4s, v2.s[1]
fmla v0.4s, v1.4s, v2.4s
Patch written by Abderrazek Zaafrani.
Differential Revision: https://reviews.llvm.org/D21571
llvm-svn: 283663
Artem Dergachev [Sat, 8 Oct 2016 11:07:21 +0000 (11:07 +0000)]
[analyzer] Fix build error after r283660 - remove constexpr strings.
llvm-svn: 283662
Artem Dergachev [Sat, 8 Oct 2016 10:54:30 +0000 (10:54 +0000)]
[analyzer] Re-apply r283094 "Improve CloneChecker diagnostics"
The parent commit (r283092) was reverted before and now finally landed.
llvm-svn: 283661
Artem Dergachev [Sat, 8 Oct 2016 10:53:00 +0000 (10:53 +0000)]
[analyzer] Re-apply r283093 "Add extra notes to ObjCDeallocChecker"
The parent commit (r283092) was reverted before and now finally landed.
llvm-svn: 283660
Michal Gorny [Sat, 8 Oct 2016 10:27:45 +0000 (10:27 +0000)]
[cmake] Split linked libraries into private & public, for linker script
Introduce LIBCXX_LIBRARIES_PUBLIC in addition to LIBCXX_LIBRARIES that
holds 'public' interface libraries -- that is, libraries that both
libc++ links to and programs linked against it need to link to.
Currently this includes the ABI library and optionally -lunwind (when
LIBCXXABI_USE_LLVM_UNWINDER is on). The libraries are included in the
linker script, in order to make it possible to link C++ programs using
clang with compiler-rt runtime out-of-the-box.
Differential Revision: https://reviews.llvm.org/D25008
llvm-svn: 283659
Kuba Brecka [Sat, 8 Oct 2016 09:01:27 +0000 (09:01 +0000)]
[compiler-rt] Exclude compiler-rt from using -fmodules when LLVM_ENABLE_MODULES=On
Sanitizers are intentionally not including system headers and often declare slightly different function prototypes, which is incompatible with -fmodules and -fcxx-modules. Let’s simply exclude compiler-rt from using -fmodules.
Differential Revision: https://reviews.llvm.org/D25230
llvm-svn: 283658
Yaron Keren [Sat, 8 Oct 2016 06:45:10 +0000 (06:45 +0000)]
Un-tabify source files, NFC.
llvm-svn: 283657
Adam Nemet [Sat, 8 Oct 2016 04:47:20 +0000 (04:47 +0000)]
[OptRemarks] Remove non-printable chars from function name
Value names may be prefixed with a binary '1' to indicate that the
backend should not modify the symbols due to any platform naming
convention.
This should not show up in the YAML opt record file because it breaks
the YAML parser.
llvm-svn: 283656
Mehdi Amini [Sat, 8 Oct 2016 04:44:23 +0000 (04:44 +0000)]
ThinLTO: don't perform incremental LTO on module without a hash
Clang always emit a hash for ThinLTO, but as other frontend are
starting to use ThinLTO, this could be a serious bug.
Differential Revision: https://reviews.llvm.org/D25379
llvm-svn: 283655
Mehdi Amini [Sat, 8 Oct 2016 04:44:18 +0000 (04:44 +0000)]
ThinLTO: handles modules with empty summaries
We need to add an entry in the combined-index for modules that have
a hash but otherwise empty summary, this is needed so that we can
get the hash for the module.
Also, if no entry is present in the combined index for a module, we
need to skip it when trying to compute a cache entry.
Differential Revision: https://reviews.llvm.org/D25300
llvm-svn: 283654
Mehdi Amini [Sat, 8 Oct 2016 04:39:34 +0000 (04:39 +0000)]
Requires the AVR backend for running test/CodeGen/AVR
llvm-svn: 283653
Pavel Labath [Sat, 8 Oct 2016 04:03:19 +0000 (04:03 +0000)]
Fix compiler warnings in TestClangASTContext.cpp
llvm-svn: 283652
Pavel Labath [Sat, 8 Oct 2016 04:03:15 +0000 (04:03 +0000)]
Fix compiler warnings in PlatformDarwinTest.cpp
llvm-svn: 283651
Kyle Butt [Sat, 8 Oct 2016 01:47:05 +0000 (01:47 +0000)]
Revert "Codegen: Tail-duplicate during placement."
This reverts commit
71c312652c10f1855b28d06697c08d47e7a243e4.
llvm-svn: 283647
Mehdi Amini [Sat, 8 Oct 2016 01:38:43 +0000 (01:38 +0000)]
Use StringRef in Command::printArg() instead of raw pointer (NFC)
llvm-svn: 283645
Eric Fiselier [Sat, 8 Oct 2016 01:32:56 +0000 (01:32 +0000)]
Remove all _LIBCPP_VERSION tests from under test/std
llvm-svn: 283644
Eric Fiselier [Sat, 8 Oct 2016 01:25:23 +0000 (01:25 +0000)]
Purge all usages of _LIBCPP_STD_VER under test/std/algorithm
llvm-svn: 283643
Dylan McKay [Sat, 8 Oct 2016 01:14:23 +0000 (01:14 +0000)]
[AVR] Add backend dependencies to MCTargetDesc/LLVMBuild.txt
llvm-svn: 283642
Zachary Turner [Sat, 8 Oct 2016 01:12:01 +0000 (01:12 +0000)]
[pdb] Dump Module Symbols to Yaml.
This is the first step towards round-tripping symbol information,
and thusly being able to write symbol information to a PDB.
This patch writes the symbol information for each compiland to
the Yaml when running in pdb2yaml mode. There's still some loose
ends, such as what to do about relocations (necessary in order to
print linkage names), how to print enums with friendly names, and
how to give the dumper access to the StringTable, but this is a
good first start.
llvm-svn: 283641
Dylan McKay [Sat, 8 Oct 2016 01:10:36 +0000 (01:10 +0000)]
Fix incorrect assertion in AVRFrameLowering.cpp
This wasn't looking at the right instruction, and would always fail.
llvm-svn: 283640
Dylan McKay [Sat, 8 Oct 2016 01:10:31 +0000 (01:10 +0000)]
[AVR] Don't worry about call frame size when initializing frame pointer
We previously only used the frame pointer if the frame pointer was too
big. This was to work around a bug (described in this old commit)
https://sourceforge.net/p/avr-llvm/code/204/tree//llvm/trunk/AVR/AVRFrameLowering.cpp?diff=
50d64d912718465cb887d17a:203
I mistakenly invered the condition assuming it was a typo. I am now
removing it because it doesn't seem to be a problem anymore (plus it's a
dirty hack).
llvm-svn: 283639
Dylan McKay [Sat, 8 Oct 2016 01:09:06 +0000 (01:09 +0000)]
[AVR] Don't shadow container while iterating in range-based loop
This works on clang, but fails on GCC 4.6
llvm-svn: 283638
Justin Lebar [Sat, 8 Oct 2016 01:07:11 +0000 (01:07 +0000)]
[CUDA] Do a better job at detecting wrong-side calls.
Summary:
Move CheckCUDACall from ActOnCallExpr and BuildDeclRefExpr to
DiagnoseUseOfDecl. This lets us catch some edge cases we were missing,
specifically around class operators.
This necessitates a few other changes:
- Avoid emitting duplicate deferred diags in CheckCUDACall.
Previously we'd carefully placed our call to CheckCUDACall such that
it would only ever run once for a particular callsite. But now this
isn't the case.
- Emit deferred diagnostics from a template
specialization/instantiation's primary template, in addition to from
the specialization/instantiation itself. DiagnoseUseOfDecl ends up
putting the deferred diagnostics on the template, rather than the
specialization, so we need to check both.
Reviewers: rsmith
Subscribers: cfe-commits, tra
Differential Revision: https://reviews.llvm.org/D24573
llvm-svn: 283637
Dylan McKay [Sat, 8 Oct 2016 01:06:21 +0000 (01:06 +0000)]
[AVR] Use references rather than pointers in AVRISelLowering
llvm-svn: 283636
Dylan McKay [Sat, 8 Oct 2016 01:05:09 +0000 (01:05 +0000)]
Allow a maximum of 64 bits to be returned in registers
The rest spills to the stack
Authored by Jake Goulding
llvm-svn: 283635
Dylan McKay [Sat, 8 Oct 2016 01:01:49 +0000 (01:01 +0000)]
[AVR] Expand MULHS for all types
Once MULHS was expanded, this exposed an issue where the condition
register was thought to be 16-bit. This caused an attempt to copy a
16-bit register to an 8-bit register.
Authored by Jake Goulding
llvm-svn: 283634
Eric Fiselier [Sat, 8 Oct 2016 00:59:16 +0000 (00:59 +0000)]
Add missing <memory> include in test
llvm-svn: 283633
Eric Fiselier [Sat, 8 Oct 2016 00:57:56 +0000 (00:57 +0000)]
Add missing include in test_allocator.h
llvm-svn: 283632
Eric Fiselier [Sat, 8 Oct 2016 00:56:22 +0000 (00:56 +0000)]
[libc++] Fix stack_allocator
Summary:
To quote STL the problems with stack allocator are"
>"stack_allocator<T, N> is seriously nonconformant to N4582 17.6.3.5 [allocator.requirements].
> First, it lacks a rebinding constructor. (The nested "struct rebind" isn't sufficient.)
> Second, it lacks templated equality/inequality.
> Third, it completely ignores alignment.
> Finally, and most severely, the Standard forbids its existence. Allocators are forbidden from returning memory "inside themselves". This requirement is implied by the Standard's requirements for rebinding and equality. It's permitted to return memory from a separate buffer object on the stack, though."
This patch attempts to address all of those issues.
First, instead of storing the buffer inside the allocator I've change `stack_allocator` to accept the buffer as an argument.
Second, in order to fix rebinding I changed the parameter list from `<class T, size_t NumElements>` to `<class T, size_t NumBytes>`. This allows allocator rebinding
between types that have different sizes.
Third, I added copy and rebinding constructors and assignment operators.
And finally I fixed the allocation logic to always return properly aligned storage.
Reviewers: mclow.lists, howard.hinnant, STL_MSFT
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D25154
llvm-svn: 283631
Dylan McKay [Sat, 8 Oct 2016 00:55:46 +0000 (00:55 +0000)]
[AVR] Add the 'SoftFail' field to all instruction formats
This will be used in the future for disassembly.
llvm-svn: 283630
Dylan McKay [Sat, 8 Oct 2016 00:50:11 +0000 (00:50 +0000)]
[AVR] Set up the instruction printer and the assembly backend
llvm-svn: 283629
Dylan McKay [Sat, 8 Oct 2016 00:45:24 +0000 (00:45 +0000)]
[AVR] Add dependencies to AVR libraries in AVRCodeGen
llvm-svn: 283628
Dylan McKay [Sat, 8 Oct 2016 00:42:58 +0000 (00:42 +0000)]
[AVR] Add missing subdirectories to LLVMBuild
llvm-svn: 283627
Hal Finkel [Sat, 8 Oct 2016 00:26:54 +0000 (00:26 +0000)]
[llvm-opt-report] Don't leave space for opts that never happen
Because screen space is precious, if an optimization (vectorization, for
example) never happens, don't leave empty space for the associated markers on
every line of the output. This makes the output much more compact, and allows
for the later inclusion of markers for more (although perhaps rare)
optimizations.
llvm-svn: 283626
Gor Nishanov [Sat, 8 Oct 2016 00:22:50 +0000 (00:22 +0000)]
[coroutines] Store an address of destroy OR cleanup part in the coroutine frame.
Summary:
If heap allocation of a coroutine is elided, we need to make sure that we will update an address stored in the coroutine frame from f.destroy to f.cleanup.
Before this change, CoroSplit synthesized these stores after coro.begin:
```
store void (%f.Frame*)* @f.resume, void (%f.Frame*)** %resume.addr
store void (%f.Frame*)* @f.destroy, void (%f.Frame*)** %destroy.addr
```
In those cases where we did heap elision, but were not able to devirtualize all indirect calls, destroy call will attempt to "free" the coroutine frame stored on the stack. Oops.
Now we use select to put an appropriate coroutine subfunction in the destroy slot. As bellow:
```
store void (%f.Frame*)* @f.resume, void (%f.Frame*)** %resume.addr
%0 = select i1 %need.alloc, void (%f.Frame*)* @f.destroy, void (%f.Frame*)* @f.cleanup
store void (%f.Frame*)* %0, void (%f.Frame*)** %destroy.addr
```
Reviewers: majnemer
Subscribers: mehdi_amini, llvm-commits
Differential Revision: https://reviews.llvm.org/D25377
llvm-svn: 283625
Lang Hames [Sat, 8 Oct 2016 00:20:42 +0000 (00:20 +0000)]
[docs] Fix indentation bug in LangRef.
llvm-svn: 283624
Dylan McKay [Sat, 8 Oct 2016 00:02:36 +0000 (00:02 +0000)]
[AVR] Add the assembly printer
Summary: This adds the AVRAsmPrinter class.
Reviewers: arsenm, kparzysz
Subscribers: llvm-commits, wdng, beanz, japaric, mgorny
Differential Revision: https://reviews.llvm.org/D25271
llvm-svn: 283623
Tom Stellard [Fri, 7 Oct 2016 23:42:48 +0000 (23:42 +0000)]
AMDGPU/SI: Handle div_fmas hazard in GCNHazardRecognizer
Reviewers: arsenm
Subscribers: kzhuravl, wdng, nhaehnle, yaxunl, tony-tye, llvm-commits
Differential Revision: https://reviews.llvm.org/D25250
llvm-svn: 283622
Marshall Clow [Fri, 7 Oct 2016 23:19:04 +0000 (23:19 +0000)]
Disable alignment support of 0x4000 for Win32. https://reviews.llvm.org/D25053
llvm-svn: 283621
Eric Fiselier [Fri, 7 Oct 2016 23:07:59 +0000 (23:07 +0000)]
Fix PR30642 - libc++ leaks always-visible symbols into programs
This was caused by r281673, specifically changing `_LIBCPP_EXTERN_TEMPLATE_TYPE_VIS`
from `__attribute__((__type_visibility__("default")))` to
`__attribute__((__visibility("default")))`.
I made that change because I thought the external instantiations needed
their members to have default visibility. However since libc++ never builds
with -fvisibility=hidden this appears not to be needed. Instead this change
caused previously hidden inline methods to become un-hidden, which is a regression.
This patch reverts the problematic change and fixes PR30642.
llvm-svn: 283620
Kyle Butt [Fri, 7 Oct 2016 22:33:20 +0000 (22:33 +0000)]
Codegen: Tail-duplicate during placement.
The tail duplication pass uses an assumed layout when making duplication
decisions. This is fine, but passes up duplication opportunities that
may arise when blocks are outlined. Because we want the updated CFG to
affect subsequent placement decisions, this change must occur during
placement.
In order to achieve this goal, TailDuplicationPass is split into a
utility class, TailDuplicator, and the pass itself. The pass delegates
nearly everything to the TailDuplicator object, except for looping over
the blocks in a function. This allows the same code to be used for tail
duplication in both places.
This change, in concert with outlining optional branches, allows
triangle shaped code to perform much better, esepecially when the
taken/untaken branches are correlated, as it creates a second spine when
the tests are small enough.
Issue from previous rollback fixed, and a new test was added for that
case as well. Issue was worklist/scheduling/taildup issue in layout.
Issue from 2nd rollback fixed, with 2 additional tests. Issue was
tail merging/loop info/tail-duplication causing issue with loops that share
a header block.
Differential revision: https://reviews.llvm.org/D18226
llvm-svn: 283619
Eric Fiselier [Fri, 7 Oct 2016 22:10:35 +0000 (22:10 +0000)]
Fix shadow warnings. Patch from STL@microsoft.com
llvm-svn: 283618
Arnold Schwaighofer [Fri, 7 Oct 2016 22:06:55 +0000 (22:06 +0000)]
swifterror: Don't compute swifterror vregs during instruction selection
The code used llvm basic block predecessors to decided where to insert phi
nodes. Instruction selection can and will liberally insert new machine basic
block predecessors. There is not a guaranteed one-to-one mapping from pred.
llvm basic blocks and machine basic blocks.
Therefore the current approach does not work as it assumes we can mark
predecessor machine basic block as needing a copy, and needs to know the set of
all predecessor machine basic blocks to decide when to insert phis.
Instead of computing the swifterror vregs as we select instructions, propagate
them at the end of instruction selection when the MBB CFG is complete.
When an instruction needs a swifterror vreg and we don't know the value yet,
generate a new vreg and remember this "upward exposed" use, and reconcile this
at the end of instruction selection.
This will only happen if the target supports promoting swifterror parameters to
registers and the swifterror attribute is used.
rdar://
28300923
llvm-svn: 283617
Mehdi Amini [Fri, 7 Oct 2016 22:03:03 +0000 (22:03 +0000)]
Rename variable to not use C++ reserved name (leading underscore + capital) (NFC)
llvm-svn: 283616
Mehdi Amini [Fri, 7 Oct 2016 22:02:59 +0000 (22:02 +0000)]
Fix MSVC build: requires namespace in front of StringRef (NFC)
llvm-svn: 283615
Sanjay Patel [Fri, 7 Oct 2016 21:55:42 +0000 (21:55 +0000)]
[DAG] clean up foldSelectOfConstants(); NFCI
Rename variables, simplify logic.
Not clear yet why we don't handle a target with ZeroOrNegativeOneBooleanContent too.
llvm-svn: 283613
Davide Italiano [Fri, 7 Oct 2016 21:53:09 +0000 (21:53 +0000)]
[InstCombine] Don't unpack arrays that are too large (part 2).
This is similar to r283599, but for store instructions.
Thanks to David for pointing out!
llvm-svn: 283612
Mehdi Amini [Fri, 7 Oct 2016 21:41:00 +0000 (21:41 +0000)]
Driver: use StringRef instead of raw pointer in lookupTypeForExtension() (NFC)
llvm-svn: 283611
Zachary Turner [Fri, 7 Oct 2016 21:40:06 +0000 (21:40 +0000)]
Add missing include.
llvm-svn: 283610
Zachary Turner [Fri, 7 Oct 2016 21:34:46 +0000 (21:34 +0000)]
Refactor Symbol visitor code.
Type visitor code had already been refactored previously to
decouple the visitor and the visitor callback interface. This
was necessary for having the flexibility to visit in different
ways (for example, dumping to yaml, reading from yaml, dumping
to ScopedPrinter, etc).
This patch merely implements the same visitation pattern for
symbol records that has already been implemented for type records.
llvm-svn: 283609
Hongbin Zheng [Fri, 7 Oct 2016 21:32:47 +0000 (21:32 +0000)]
[cmake] Treat polly as "in tree" if LLVM_EXTERNAL_POLLY_SOURCE_DIR is provided
Differential Revision: https://reviews.llvm.org/D25354
llvm-svn: 283608
Zachary Turner [Fri, 7 Oct 2016 21:32:16 +0000 (21:32 +0000)]
Fix a few warnings caught by clang.
llvm-svn: 283607