Commit Graph

245019 Commits

Author SHA1 Message Date
Michal Gorny e2f4b25d04 [cmake] Use LLVM_CMAKE_PATH for GetSVN script
Use the LLVM_CMAKE_PATH variable to locate the GetSVN.cmake script.
The variable was already available in stand-alone builds, and is also
set by LLVM since r284581.

llvm-svn: 284584
2016-10-19 12:54:28 +00:00
Michal Gorny 7c3e4ebd92 [cmake] Allow testing against installed LLVM with no sources
Adjust the stand-alone build files to accept either CMake files from
LLVM_CMAKE_PATH or from LLVM_MAIN_SRC_DIR instead of requiring both.
This makes it possible to run libcxx tests on top of installed LLVM
and lit, without having to unpack a copy of LLVM sources. Furthermore,
it avoids adding duplicate paths.

Differential Revision: https://reviews.llvm.org/D25093

llvm-svn: 284583
2016-10-19 12:34:17 +00:00
Michal Gorny 0ed2f4e98b [cmake] Use LLVM_CMAKE_PATH for GetSVN script
Use the LLVM_CMAKE_PATH variable to locate the GetSVN.cmake script.
The variable was already available in stand-alone builds, and is also
set by LLVM since r284581.

llvm-svn: 284582
2016-10-19 12:21:39 +00:00
Michal Gorny e22b7f5aea [cmake] Declare LLVM_CMAKE_PATH for use in subprojects
Declare the LLVM_CMAKE_PATH to the source directory location of CMake
files, in order to make it possible to easily use them in subprojects.
Such a variable is already declared in most of LLVM projects
(and inconsistently mixed with direct source tree references), including
Clang, LLDB, compiler-rt, libcxx... Declaring it inside main LLVM tree
makes it possible to avoid having to declare fallback values or use
conditionals in those projects.

It should be noted that in some of the subprojects LLVM_CMAKE_PATH is
used to reference generated LLVMConfig.cmake file. However, these
references are conditional to stand-alone builds and explicitly
including this file is unnecessary in combined builds.

Differential Revision: https://reviews.llvm.org/D25724

llvm-svn: 284581
2016-10-19 12:18:34 +00:00
James Molloy fbfd173447 [Thumb-1] Synthesize TBB/TBH instructions to make use of compressed jump tables
The TBB and TBH instructions in Thumb-2 allow jump tables to be compressed into sequences of bytes or shorts respectively. These instructions do not exist in Thumb-1, however it is possible to synthesize them out of a sequence of other instructions.

It turns out this sequence is so short that it's almost never a lose for performance and is ALWAYS a significant win for code size.

TBB example:
Before: lsls r0, r0, #2    After: add  r0, pc
        adr  r1, .LJTI0_0         ldrb r0, [r0, #6]
        ldr  r0, [r0, r1]         lsls r0, r0, #1
        mov  pc, r0               add  pc, r0
  => No change in prologue code size or dynamic instruction count. Jump table shrunk by a factor of 4.

The only case that can increase dynamic instruction count is the TBH case:

Before: lsls r0, r4, #2    After: lsls r4, r4, #1
        adr  r1, .LJTI0_0         add  r4, pc
        ldr  r0, [r0, r1]         ldrh r4, [r4, #6]
        mov  pc, r0               lsls r4, r4, #1
                                  add  pc, r4
  => 1 more instruction in prologue. Jump table shrunk by a factor of 2.

So there is an argument that this should be disabled when optimizing for performance (and a TBH needs to be generated). I'm not so sure about that in practice, because on small cores with Thumb-1 performance is often tied to code size. But I'm willing to turn it off when optimizing for performance if people want (also note that TBHs are fairly rare in practice!)

llvm-svn: 284580
2016-10-19 12:06:49 +00:00
Andrey Bokhanko 9941ca8af6 [Sema] Gcc compatibility of vector shift
Gcc prints error if elements of left and right parts of a shift have different
sizes. This patch is provided the GCC compatibility.

Patch by Vladimir Yakovlev.

Differential Revision: https://reviews.llvm.org/D24669

llvm-svn: 284579
2016-10-19 12:06:10 +00:00
Simon Pilgrim 7dcb6e572e [DAGCombiner] Just call isConstOrConstSplat directly. NFCI.
This will get the same ConstantSDNode scalar or vector splat value as the current separate dyn_cast<ConstantSDNode> / isVector() approach.

llvm-svn: 284578
2016-10-19 11:28:15 +00:00
Vassil Vassilev c774a23bd9 [modules] Do not report missing definitions of demoted constexpr variable templates.
This is a followup to regression introduced in r284284.

This should fix our libstdc++ modules builds.

https://reviews.llvm.org/D25678

Reviewed by Richard Smith!

llvm-svn: 284577
2016-10-19 11:19:30 +00:00
Simon Pilgrim 9122ac994b Fix line endings
llvm-svn: 284576
2016-10-19 11:16:58 +00:00
Eugene Leviant 74c1b6c548 [ELF] Add "error" and "warning" prefixes to linker messages
Differential revision: https://reviews.llvm.org/D25729

llvm-svn: 284575
2016-10-19 09:35:19 +00:00
Simon Pilgrim b2ca2505cc [DAGCombine] Generalize distributeTruncateThroughAnd to work with any non-opaque constant or constant vector
llvm-svn: 284574
2016-10-19 08:57:37 +00:00
Eric Liu 99eeab7ff3 [clang-format] Add comment manipulation header
Summary:
Introduces a separate target for comment manipulation.
Currently, comment manipulation is in BreakableComment.cpp.
Towards implementing comment reflowing, we want to factor out the
comment-related functionality, so it can be reused.
Start simple by just moving out getLineCommentIndentPrefix.

Patch by Krasimir Georgiev!

Reviewers: djasper

Subscribers: klimek, beanz, mgorny, modocache

Differential Revision: https://reviews.llvm.org/D25725

llvm-svn: 284573
2016-10-19 08:19:46 +00:00
Sjoerd Meijer 3f5111d363 Revert of r284571 because of failing tests.
llvm-svn: 284572
2016-10-19 07:45:48 +00:00
Sjoerd Meijer a318779263 Checking FP function attribute values and adding more build attribute tests.
This renames the function for checking FP function attribute values and also
adds more build attribute tests (which are in separate files because build
attributes are set per file).

Differential Revision: https://reviews.llvm.org/D25625

llvm-svn: 284571
2016-10-19 07:25:06 +00:00
NAKAMURA Takumi da9dc6ad84 DenseSet: Appease msc18 to define derived constructors explicitly.
msc18 doesn't recognize "using BaseT::BaseT;"

  llvm\include\llvm/ADT/DenseSet.h(213) : error C2875: using-declaration causes a multiple declaration of 'BaseT'
          llvm\include\llvm/ADT/DenseSet.h(214) : see reference to class template instantiation 'llvm::DenseSet<ValueT,ValueInfoT>' being compiled
  llvm\include\llvm/ADT/DenseSet.h(231) : error C2875: using-declaration causes a multiple declaration of 'BaseT'
          llvm\include\llvm/ADT/DenseSet.h(232) : see reference to class template instantiation 'llvm::SmallDenseSet<ValueT,InlineBuckets,ValueInfoT>' being compiled

llvm-svn: 284570
2016-10-19 05:43:17 +00:00
Craig Topper a4dc340cf2 [AVX-512] Teach isel lowering that a subvector broadcast being inserted into both halves of a 512-bit vector can be combined into a larger subvector broadcast.
Summary:
This allows us to create broadcasts of 128-bit vector loads into 512-bit vectors.

New patterns added to support 8-bit and 16-bit vector types and v2f64/v2i64->v8f64/v8i64 without DQI instructions.

There also fallback patterns when the load can't be folded. These patterns are a little complex as we first need to insert the lower 128-bits into the second 128-bits using a zmm subvector insert instruction. We need to use a zmm insert in case VLX isn't available. Then use another zmm sub vector insert to take those 256-bits and insert them into the upper bits. Since we used a zmm insert to create the 256-bits we also need to do a extract_subreg to get just the lower 256-bits to pass to the second insert.

The outer insert for the fallback patterns should have its type correct because eventually we should also supported masked operations here too. So we need a DQI and a NoDQI version of the v16f32/v16i32 patterns.

Reviewers: RKSimon, delena, igorb

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D25651

llvm-svn: 284567
2016-10-19 04:44:17 +00:00
Dehao Chen 4b41571d24 Update the section.ll to fix non-x86 failure.
llvm-svn: 284566
2016-10-19 03:53:41 +00:00
Chris Bieneman e0dcd6d7f4 Revert back to the state before r284550
This patch is causing a lot of issues on bots that I didn't see in local testing. I'm going to have to work on this. Reverting for now while I sort it out.

llvm-svn: 284565
2016-10-19 02:44:20 +00:00
Chris Bieneman 716ac7684c Trying to fix a few more missing LDFLAGS.
llvm-svn: 284564
2016-10-19 02:31:31 +00:00
Dehao Chen 95fc43143d Revert r284545 again as the regression in ppc still exists. There is bug in MBPI exposed by th patch.
Also update the section.ll to fix non-x86 failure.

llvm-svn: 284563
2016-10-19 01:18:25 +00:00
Vitaly Buka 490fda3366 [asan] Replace std::to_string with llvm::to_string
llvm-svn: 284557
2016-10-19 00:16:56 +00:00
Richard Smith 5f4b388442 Resolve exception specifications when selecting an overloaded operator.
llvm-svn: 284556
2016-10-19 00:14:23 +00:00
Chris Bieneman 87378bd21e Fixing the linux bots I broke in r284550
Need to gate cxx linker adding driver-mode flag based on the linker being clang.

llvm-svn: 284555
2016-10-19 00:13:56 +00:00
Kostya Serebryany 95b1a434d2 [libFuzzer] extend -print_coverage to also print uncovered lines, functions, and files.
Example of output:
COVERAGE:
COVERED: in DSO2(int) /pathto/DSO2.cpp:6
COVERED: in DSO2(int) /pathto/DSO2.cpp:8
COVERED: in DSO1(int) /pathto/DSO1.cpp:6
COVERED: in DSO1(int) /pathto/DSO1.cpp:8
COVERED: in LLVMFuzzerTestOneInput /pathto/DSOTestMain.cpp:16
COVERED: in LLVMFuzzerTestOneInput /pathto/DSOTestMain.cpp:19
COVERED: in LLVMFuzzerTestOneInput /pathto/DSOTestMain.cpp:25
COVERED: in LLVMFuzzerTestOneInput /pathto/DSOTestMain.cpp:26
MODULE_WITH_COVERAGE: /pathto/libLLVMFuzzer-DSO1.so
UNCOVERED_LINE: in DSO1(int) /pathto/DSO1.cpp:9
UNCOVERED_FUNC: in Uncovered1()
MODULE_WITH_COVERAGE: /pathto/libLLVMFuzzer-DSO2.so
UNCOVERED_LINE: in DSO2(int) /pathto/DSO2.cpp:9
UNCOVERED_FUNC: in Uncovered2()
MODULE_WITH_COVERAGE: /pathto/LLVMFuzzer-DSOTest
UNCOVERED_LINE: in LLVMFuzzerTestOneInput /pathto/DSOTestMain.cpp:21
UNCOVERED_LINE: in LLVMFuzzerTestOneInput /pathto/DSOTestMain.cpp:27
UNCOVERED_FILE: /pathto/DSOTestExtra.cpp

Several things are not perfect here:
* we are using objdump+awk instead of sancov because sancov does not support DSOs yet.
* this breaks in the presence of ASAN_OPTIONS=strip_path_prefix=...
  (need to implement another API to get the module name by PC)

llvm-svn: 284554
2016-10-19 00:12:03 +00:00
Justin Lebar d3fd70dedd [CUDA] Rework tests now that we emit deferred diagnostics during sema. Test-only change.
Summary:
Previously we had to split out a lot of our tests into a test that
checked only immediate errors and a test that checked only deferred
errors.  This was because, if you emitted any immediate errors, we
wouldn't run codegen, where the deferred errors were emitted.

We've fixed this, and now emit deferred errors during sema.  This lets
us merge a bunch of tests, and lets us convert some other tests to
-fsyntax-only.

Reviewers: tra

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D25755

llvm-svn: 284553
2016-10-19 00:06:49 +00:00
Chris Bieneman b47711ab27 When invoking Terminal, don't assume the default shell
Summary:
If a user has their shell set to a non-POSIX conferment shell the TestTerminal.py tests fail because the shell blurb constructed here may not work in their shell.

In my specific case fish-shell (The Friendly Interactive Shell - http://fishshell.com) does not support $?, it instead uses $status (because it is friendly).

This patch removes the assumption of your default shell by running the constructed bash command via "/bin/bash -c ...". This should be safer for users mutating their shell environment.

Reviewers: tfiala

Subscribers: joerg, lldb-commits

Differential Revision: https://reviews.llvm.org/D25750

llvm-svn: 284552
2016-10-18 23:55:34 +00:00
Chris Bieneman f50761f128 [CMake] Don't include LLDB_TEST_COMPILER in cached variable
Summary:
CMake has no builtin mechanism for cache invalidation. As a general convention you want to not expand user-specified variables in other cached variables because they will not get updated when the user changes their specified value.

This patch moves the "-C" option for dotest.py into the LLDB_TEST_COMMON_ARGS and out of the CMake cache. In order to prevent issues with out-of-date cache files on builders I've added code to scrub "-C ${LLDB_TEST_COMPILER}" out of the CMake caches, by Force writing the variable. This code can be removed in a few days once the change has trickled through CI systems.

Reviewers: tfiala, labath, zturner

Subscribers: lldb-commits, mgorny

Differential Revision: https://reviews.llvm.org/D25751

llvm-svn: 284551
2016-10-18 23:54:28 +00:00
Chris Bieneman e80f007dfb Use clang --driver-mode instead of guessing c++ compiler path
Summary:
When building the LLDB test programs, if your CC is clang it actually isn't safe to make CXX a string replace of "clang -> clang++". This falls down on unix configurations if your compiler is clang-${version}.

A safer approach is to use the "--driver-mode=g++" option to tell clang to act like clang++.

Reviewers: tfiala, zturner, labath

Subscribers: lldb-commits

Differential Revision: https://reviews.llvm.org/D25753

llvm-svn: 284550
2016-10-18 23:53:24 +00:00
Richard Smith 84a0b6dba1 DR1330: instantiate exception-specifications when "needed". We previously did
not instantiate exception specifications of functions if they were only used in
unevaluated contexts (other than 'noexcept' expressions).

In C++17 onwards, this becomes essential since the exception specification is
now part of the function's type.

Note that this means that constructs like the following no longer work:

  struct A {
    static T f() noexcept(...);
    decltype(f()) *p;
  };

... because the decltype expression now needs the exception specification of
'f', which has not yet been parsed.

llvm-svn: 284549
2016-10-18 23:39:12 +00:00
Vitaly Buka 832c383b25 [asan] Update test for D25715
Reviewers: eugenis

Subscribers: kubabrecka, llvm-commits

Differential Revision: https://reviews.llvm.org/D25716

llvm-svn: 284548
2016-10-18 23:30:07 +00:00
Vitaly Buka 5910a92560 [asan] Simplify calculation of stack frame layout extraction calculation of stack description into separate function.
Reviewers: eugenis

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D25754

llvm-svn: 284547
2016-10-18 23:29:52 +00:00
Vitaly Buka d88e52012b [asan] Append line number to variable name if line is available and in the same file as the function.
PR30498

Reviewers: eugenis

Differential Revision: https://reviews.llvm.org/D25715

llvm-svn: 284546
2016-10-18 23:29:41 +00:00
Dehao Chen f8ac3d26d5 Using branch probability to guide critical edge splitting.
Summary:
The original heuristic to break critical edge during machine sink is relatively conservertive: when there is only one instruction sinkable to the critical edge, it is likely that the machine sink pass will not break the critical edge. This leads to many speculative instructions executed at runtime. However, with profile info, we could model the splitting benefits: if the critical edge has 50% taken rate, it would always be beneficial to split the critical edge to avoid the speculated runtime instructions. This patch uses profile to guide critical edge splitting in machine sink pass.

The performance impact on speccpu2006 on Intel sandybridge machines:

spec/2006/fp/C++/444.namd                  25.3  +0.26%
spec/2006/fp/C++/447.dealII               45.96  -0.10%
spec/2006/fp/C++/450.soplex               41.97  +1.49%
spec/2006/fp/C++/453.povray               36.83  -0.96%
spec/2006/fp/C/433.milc                   23.81  +0.32%
spec/2006/fp/C/470.lbm                    41.17  +0.34%
spec/2006/fp/C/482.sphinx3                48.13  +0.69%
spec/2006/int/C++/471.omnetpp             22.45  +3.25%
spec/2006/int/C++/473.astar               21.35  -2.06%
spec/2006/int/C++/483.xalancbmk           36.02  -2.39%
spec/2006/int/C/400.perlbench              33.7  -0.17%
spec/2006/int/C/401.bzip2                  22.9  +0.52%
spec/2006/int/C/403.gcc                   32.42  -0.54%
spec/2006/int/C/429.mcf                   39.59  +0.19%
spec/2006/int/C/445.gobmk                 26.98  -0.00%
spec/2006/int/C/456.hmmer                 24.52  -0.18%
spec/2006/int/C/458.sjeng                 28.26  +0.02%
spec/2006/int/C/462.libquantum            55.44  +3.74%
spec/2006/int/C/464.h264ref               46.67  -0.39%

geometric mean                                   +0.20%

Manually checked 473 and 471 to verify the diff is in the noise range.

Reviewers: rengolin, davidxl

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D24818

llvm-svn: 284545
2016-10-18 23:24:02 +00:00
Dehao Chen 62d0e64e9e revert r284541.
llvm-svn: 284544
2016-10-18 23:11:20 +00:00
Michael Kruse 6b87504973 [test] Fix buildbot after SCEV change.
Update test after commit r284501:
[SCEV] Make CompareValueComplexity a little bit smarter

Contributed-by: Sanjoy Das <sanjoy@playingwithpointers.com>
llvm-svn: 284543
2016-10-18 22:58:09 +00:00
Rong Xu 1c0e9b97d2 Conditionally eliminate library calls where the result value is not used
Summary:
This pass shrink-wraps a condition to some library calls where the call
result is not used. For example:
   sqrt(val);
 is transformed to
   if (val < 0)
     sqrt(val);
Even if the result of library call is not being used, the compiler cannot
safely delete the call because the function can set errno on error
conditions.
Note in many functions, the error condition solely depends on the incoming
parameter. In this optimization, we can generate the condition can lead to
the errno to shrink-wrap the call. Since the chances of hitting the error
condition is low, the runtime call is effectively eliminated.

These partially dead calls are usually results of C++ abstraction penalty
exposed by inlining. This optimization hits 108 times in 19 C/C++ programs
in SPEC2006.

Reviewers: hfinkel, mehdi_amini, davidxl

Subscribers: modocache, mgorny, mehdi_amini, xur, llvm-commits, beanz

Differential Revision: https://reviews.llvm.org/D24414

llvm-svn: 284542
2016-10-18 21:36:27 +00:00
Dehao Chen ea62ae9844 Using branch probability to guide critical edge splitting.
Summary:
The original heuristic to break critical edge during machine sink is relatively conservertive: when there is only one instruction sinkable to the critical edge, it is likely that the machine sink pass will not break the critical edge. This leads to many speculative instructions executed at runtime. However, with profile info, we could model the splitting benefits: if the critical edge has 50% taken rate, it would always be beneficial to split the critical edge to avoid the speculated runtime instructions. This patch uses profile to guide critical edge splitting in machine sink pass.

The performance impact on speccpu2006 on Intel sandybridge machines:

spec/2006/fp/C++/444.namd                  25.3  +0.26%
spec/2006/fp/C++/447.dealII               45.96  -0.10%
spec/2006/fp/C++/450.soplex               41.97  +1.49%
spec/2006/fp/C++/453.povray               36.83  -0.96%
spec/2006/fp/C/433.milc                   23.81  +0.32%
spec/2006/fp/C/470.lbm                    41.17  +0.34%
spec/2006/fp/C/482.sphinx3                48.13  +0.69%
spec/2006/int/C++/471.omnetpp             22.45  +3.25%
spec/2006/int/C++/473.astar               21.35  -2.06%
spec/2006/int/C++/483.xalancbmk           36.02  -2.39%
spec/2006/int/C/400.perlbench              33.7  -0.17%
spec/2006/int/C/401.bzip2                  22.9  +0.52%
spec/2006/int/C/403.gcc                   32.42  -0.54%
spec/2006/int/C/429.mcf                   39.59  +0.19%
spec/2006/int/C/445.gobmk                 26.98  -0.00%
spec/2006/int/C/456.hmmer                 24.52  -0.18%
spec/2006/int/C/458.sjeng                 28.26  +0.02%
spec/2006/int/C/462.libquantum            55.44  +3.74%
spec/2006/int/C/464.h264ref               46.67  -0.39%

geometric mean                                   +0.20%

Manually checked 473 and 471 to verify the diff is in the noise range.

Reviewers: rengolin, davidxl

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D24818

llvm-svn: 284541
2016-10-18 21:36:11 +00:00
David Blaikie 69494a9805 dwarfdump: add space missing from the type unit header description
llvm-svn: 284540
2016-10-18 21:18:43 +00:00
David Blaikie e4c3915a5a dwarfdump: Include the name in the unit description, even in non-summarized mode
(accidentally removed this from my previous change when I was rejecting
some clang-format formatting... )

llvm-svn: 284539
2016-10-18 21:16:45 +00:00
Dehao Chen 83033e0b9a Add target for test to fix regression introduced by r284533.
llvm-svn: 284538
2016-10-18 21:13:31 +00:00
David Blaikie 50cc27ecb9 dwarfdump: -summarize-types: print a short summary (unqualified type name, hash, length) of type units rather than dumping contents
This is just a quick utility handy for getting rough summaries of types
in a given object or dwo file. I've been using it to investigate the
amount of type info redundancy across a project build, for example.

llvm-svn: 284537
2016-10-18 21:09:48 +00:00
Eli Friedman c0a717ba5b Improve ARM lowering for "icmp <2 x i64> eq".
The custom lowering is pretty straightforward: basically, just AND
together the two halves of a <4 x i32> compare.

Differential Revision: https://reviews.llvm.org/D25713

llvm-svn: 284536
2016-10-18 21:03:40 +00:00
Davide Italiano 36efa68463 [GVN] Consistently use division instead of shift. NFCI.
This is in line with other places of GVN (e.g. load coercion
logic).

llvm-svn: 284535
2016-10-18 21:02:27 +00:00
Davide Italiano 64cd985e44 [GVN] Remove dead code. NFC.
llvm-svn: 284534
2016-10-18 21:00:26 +00:00
Dehao Chen 302b69c940 Use profile info to set function section prefix to group hot/cold functions.
Summary:
The original implementation is in r261607, which was reverted in r269726 to accomendate the ProfileSummaryInfo analysis pass. The new implementation:
1. add a new metadata for function section prefix
2. query against ProfileSummaryInfo in CGP to set the correct section prefix for each function
3. output the section prefix set by CGP

Reviewers: davidxl, eraman

Subscribers: vsk, llvm-commits

Differential Revision: https://reviews.llvm.org/D24989

llvm-svn: 284533
2016-10-18 20:42:47 +00:00
Evandro Menezes 4dd6c68d67 [AArch64] Fix test triplet
llvm-svn: 284532
2016-10-18 20:41:30 +00:00
Evandro Menezes ce8d60156c [AArch64] Avoid materializing 0.0 when generating FP SELECT
Transform `a == 0.0 ? 0.0 : x` to `a == 0.0 ? a : x` and `a != 0.0 ? x : 0.0`
to `a != 0.0 ? x : a` to avoid materializing 0.0 for FCSEL, since it does not
have to be materialized beforehand for FCMP, as it has a form that has 0.0
as an implicit operand.

Differential Revision: https://reviews.llvm.org/D24808

llvm-svn: 284531
2016-10-18 20:37:35 +00:00
Richard Smith fccb5128b2 Add missing warning for use of C++1z init-statements in C++14 and before.
llvm-svn: 284530
2016-10-18 20:27:16 +00:00
Kevin Enderby 89baf99c92 One more additional error check for invalid Mach-O files for a
load command that use the MachO:: linkedit_data_command
type but is not used in llvm libObject code but used in llvm tool code.

This is for the LC_CODE_SIGNATURE load command.

llvm-svn: 284529
2016-10-18 20:24:12 +00:00
Richard Smith 304b124a11 [c++1z] Fix corner case where we could create a function type whose canonical type is not actually canonical.
llvm-svn: 284528
2016-10-18 20:13:25 +00:00