platform/upstream/llvm.git
6 years ago[MS Demangler] Demangle pointers to member functions.
Zachary Turner [Thu, 26 Jul 2018 20:20:10 +0000 (20:20 +0000)]
[MS Demangler] Demangle pointers to member functions.

After this patch, we can now properly demangle pointers to member
functions.  The calling convention is located in the wrong place,
but this will be fixed in a followup since it also affects non
member function pointers.

Differential Revision: https://reviews.llvm.org/D49639

llvm-svn: 338065

6 years ago[COFF] Handle comdat sections without leader symbols
Martin Storsjo [Thu, 26 Jul 2018 20:14:50 +0000 (20:14 +0000)]
[COFF] Handle comdat sections without leader symbols

Discard them unless they have been associated by other means (yet
uimplemented).

According to MS link.exe, such sections are illegal, but MinGW setups
use them in their take on associative comdats.

This avoids leaving references to the bogus SectionChunk* PendingComdat,
which cannot be dereferenced.

This fixes PR38183.

Differential Revision: https://reviews.llvm.org/D49653

llvm-svn: 338064

6 years ago[MC] Add support for the .rva assembler directive for COFF targets
Martin Storsjo [Thu, 26 Jul 2018 20:11:26 +0000 (20:11 +0000)]
[MC] Add support for the .rva assembler directive for COFF targets

Even though gas doesn't document it, it has been supported there for
a very long time.

This produces the 32 bit relative virtual address (aka image relative
address) for a given symbol. ".rva foo" is essentially equal to
".long foo@imgrel".

Differential Revision: https://reviews.llvm.org/D49821

llvm-svn: 338063

6 years agoHandle the lack of a symbol table correctly.
Stephen Hines [Thu, 26 Jul 2018 20:05:31 +0000 (20:05 +0000)]
Handle the lack of a symbol table correctly.

Summary:
These two cases will trigger a dereference on a nullptr, since the
SymbolTable can be nonexistent for a given library, in addition to just
being empty.

Reviewers: alexshap

Reviewed By: alexshap

Subscribers: meikeb, kongyi, chh, jakehehrlich, llvm-commits, pirama

Differential Revision: https://reviews.llvm.org/D49534

llvm-svn: 338062

6 years ago[MS Demangler] Demangle data member pointers.
Zachary Turner [Thu, 26 Jul 2018 19:56:09 +0000 (19:56 +0000)]
[MS Demangler] Demangle data member pointers.

Differential Revision: https://reviews.llvm.org/D49630

llvm-svn: 338061

6 years ago[AMDGPU] Fix VGPR spills where offset doesn't fit in 12 bits
Scott Linder [Thu, 26 Jul 2018 19:47:51 +0000 (19:47 +0000)]
[AMDGPU] Fix VGPR spills where offset doesn't fit in 12 bits

Scale the offset of VGPR spills by the wave size when it cannot fit in the
12-bit offset immediate field and so is added to the soffset SGPR. This
accounts for hardware swizzling of scratch memory.

Differential Revision: https://reviews.llvm.org/D49448

llvm-svn: 338060

6 years ago[InstCombine] fold udiv with common factor from muls with nuw
Sanjay Patel [Thu, 26 Jul 2018 19:22:41 +0000 (19:22 +0000)]
[InstCombine] fold udiv with common factor from muls with nuw

Unfortunately, sdiv isn't as simple because of UB due to overflow.

This fold is mentioned in PR38239:
https://bugs.llvm.org/show_bug.cgi?id=38239

llvm-svn: 338059

6 years agoMake framework-header-fix process copied headers
Alex Langford [Thu, 26 Jul 2018 19:04:46 +0000 (19:04 +0000)]
Make framework-header-fix process copied headers

Summary:
Previously the framework-header-fix script would change the sources
before they were copied, leading to unnecessary rebuilds on repeat
`ninja lldb` invocations. This runs the script on the headers after
they're copied into the produced LLDB.framework, meaning it doesn't
affect any files being built.

Patch by Keith Smiley <keithbsmiley@gmail.com>!

Differential Revision: https://reviews.llvm.org/D49779

llvm-svn: 338058

6 years ago[VirtualFileSystem] InMemoryFileSystem::status: Return a Status with the requested...
Simon Marchi [Thu, 26 Jul 2018 18:55:02 +0000 (18:55 +0000)]
[VirtualFileSystem] InMemoryFileSystem::status: Return a Status with the requested name

Summary:

InMemoryFileSystem::status behaves differently than
RealFileSystem::status.  The Name contained in the Status returned by
RealFileSystem::status will be the path as requested by the caller,
whereas InMemoryFileSystem::status returns the normalized path.

For example, when requested the status for "../src/first.h",
RealFileSystem returns a Status with "../src/first.h" as the Name.
InMemoryFileSystem returns "/absolute/path/to/src/first.h".

The reason for this change is that I want to make a unit test in the
clangd testsuite (where we use an InMemoryFileSystem) to reproduce a
bug I get with the clangd program (where a RealFileSystem is used).
This difference in behavior "hides" the bug in the unit test version.

Reviewers: malaperle, ilya-biryukov, bkramer

Subscribers: cfe-commits, ioeric, ilya-biryukov, bkramer, hokein, omtcyfz

Differential Revision: https://reviews.llvm.org/D48903

llvm-svn: 338057

6 years agoRefactor checking of switch conditions and case values.
Richard Smith [Thu, 26 Jul 2018 18:41:30 +0000 (18:41 +0000)]
Refactor checking of switch conditions and case values.

Check each case value in turn while parsing it, performing the
conversion to the switch type within the context of the expression
itself. This will become necessary in order to properly handle cleanups
for temporaries created as part of the case label (in an upcoming
patch). For now it's just good hygiene.

This necessitates moving the checking for the switch condition itself to
earlier, so that the destination type is available when checking the
case labels.

As a nice side-effect, we get slightly improved diagnostic quality and
error recovery by separating the case expression checking from the case
statement checking and from tracking whether there are discarded case
labels.

llvm-svn: 338056

6 years ago[OPENMP, DOCS] Fixed typo, NFC.
Alexey Bataev [Thu, 26 Jul 2018 18:40:41 +0000 (18:40 +0000)]
[OPENMP, DOCS] Fixed typo, NFC.

llvm-svn: 338055

6 years agoRevert Clangd Dex Iterators patch
Kirill Bobyrev [Thu, 26 Jul 2018 18:25:48 +0000 (18:25 +0000)]
Revert Clangd Dex Iterators patch

This reverts two revisions:

* https://reviews.llvm.org/rL338017
* https://reviews.llvm.org/rL338028

They caused crash for Clang 3.6 & Clang 3.7 buildbots, it was
reported by Jeremy Morse.

llvm-svn: 338054

6 years ago[test] Use printf instead of C++ iostream, NFC.
Jonas Hahnfeld [Thu, 26 Jul 2018 18:23:40 +0000 (18:23 +0000)]
[test] Use printf instead of C++ iostream, NFC.

This test fails with libc++ when built with MemorySanitizer. This
is because we link to an uninstrumented version of the library
so msan detects a nested error when calling std::cout << "...".
This can be easily avoided by using good old printf.

Differential Revision: https://reviews.llvm.org/D49867

llvm-svn: 338053

6 years agoPR30734: Remove __kmp_ft_page_allocate()
Jonas Hahnfeld [Thu, 26 Jul 2018 18:15:02 +0000 (18:15 +0000)]
PR30734: Remove __kmp_ft_page_allocate()

This function was not enabled by default and not exported when manually
tweaking the build flags. Additionally it was hard to use since there
is no corresponding __kmp_ft_page_free().
The code itself is questionable because the returned memory address
is padded by an extra pointer which stores the unpadded start of the
allocated region (this would need to be freed).

Differential Revision: https://reviews.llvm.org/D49802

llvm-svn: 338052

6 years ago[test] Remove XFAIL of omp_for_bigbounds.c for Intel Compiler
Jonas Hahnfeld [Thu, 26 Jul 2018 18:14:57 +0000 (18:14 +0000)]
[test] Remove XFAIL of omp_for_bigbounds.c for Intel Compiler

The initial commit said that the test passes with Intel Compiler,
so change XFAIL to only list clang and gcc.

Differential Revision: https://reviews.llvm.org/D49801

llvm-svn: 338051

6 years ago[COFF, ARM64] Decide when to mark struct returns as SRet
Mandeep Singh Grang [Thu, 26 Jul 2018 18:07:59 +0000 (18:07 +0000)]
[COFF, ARM64] Decide when to mark struct returns as SRet

Summary:
Refer the MS ARM64 ABI Convention for the behavior for struct returns:
https://docs.microsoft.com/en-us/cpp/build/arm64-windows-abi-conventions#return-values

Reviewers: mstorsjo, compnerd, rnk, javed.absar, yinma, efriedma

Reviewed By: rnk, efriedma

Subscribers: haripul, TomTan, yinma, efriedma, kristof.beyls, chrib, llvm-commits

Differential Revision: https://reviews.llvm.org/D49464

llvm-svn: 338050

6 years ago[OPENMP] What's new for OpenMP in clang.
Alexey Bataev [Thu, 26 Jul 2018 17:53:45 +0000 (17:53 +0000)]
[OPENMP] What's new for OpenMP in clang.

Updated ReleaseNotes + Status of the OpenMP support in clang.

llvm-svn: 338049

6 years ago[Sema][ObjC] Do not propagate the nullability specifier on the receiver
Akira Hatanaka [Thu, 26 Jul 2018 17:51:13 +0000 (17:51 +0000)]
[Sema][ObjC] Do not propagate the nullability specifier on the receiver
to the result type of a message send if the result type cannot have a
nullability specifier.

Previously, clang would print the following message when the code in
nullability.m was compiled:

"incompatible integer to pointer conversion initializing 'int *' with
an expression of type 'int _Nullable'"

This is wrong as 'int' isn't supposed to have any nullability
specifiers.

rdar://problem/40830514

llvm-svn: 338048

6 years ago[RISCV] Add support for _interrupt attribute
Ana Pazos [Thu, 26 Jul 2018 17:49:43 +0000 (17:49 +0000)]
[RISCV] Add support for _interrupt attribute

- Save/restore only registers that are used.
This includes Callee saved registers and Caller saved registers
(arguments and temporaries) for integer and FP registers.
- If there is a call in the interrupt handler, save/restore all
Caller saved registers (arguments and temporaries) and all FP registers.
- Emit special return instructions depending on "interrupt"
attribute type.
Based on initial patch by Zhaoshi Zheng.

Reviewers: asb

Reviewed By: asb

Subscribers: rkruppe, the_o, MartinMosbeck, brucehoult, rbar, johnrusso, simoncook, sabuasal, niosHD, kito-cheng, shiva0217, zzheng, edward-jones, mgrang, rogfer01, llvm-commits

Differential Revision: https://reviews.llvm.org/D48411

llvm-svn: 338047

6 years agoMacroFusion: Fix macro fusion with ExitSU failing in top-down scheduling
Matthias Braun [Thu, 26 Jul 2018 17:43:56 +0000 (17:43 +0000)]
MacroFusion: Fix macro fusion with ExitSU failing in top-down scheduling

When fusing instructions A and B, we must add all predecessors of B as
predecessors of A to avoid instructions getting scheduling in between.

There is a special case involving ExitSU: Every other node must be
scheduled before it by design and we don't need to make this explicit in
the graph, however when fusing with a different node we need to schedule
every othere node before the fused node too and we need to make this
explicit now: This patch adds a dependency from the fused node to all
roots in the graph.

Differential Revision: https://reviews.llvm.org/D49830

llvm-svn: 338046

6 years ago [RISCV] Add support for interrupt attribute
Ana Pazos [Thu, 26 Jul 2018 17:37:45 +0000 (17:37 +0000)]
 [RISCV] Add support for interrupt attribute

Summary:
Clang supports the GNU style ``__attribute__((interrupt))`` attribute  on RISCV targets.
Permissible values for this parameter are user, supervisor, and machine.
If there is no parameter, then it defaults to machine.
Reference: https://gcc.gnu.org/onlinedocs/gcc/RISC-V-Function-Attributes.html
Based on initial patch by Zhaoshi Zheng.

Reviewers: asb, aaron.ballman

Reviewed By: asb, aaron.ballman

Subscribers: rkruppe, the_o, aaron.ballman, MartinMosbeck, brucehoult, rbar, johnrusso, simoncook, sabuasal, niosHD, kito-cheng, shiva0217, zzheng, edward-jones, mgrang, rogfer01, cfe-commits

Differential Revision: https://reviews.llvm.org/D48412

llvm-svn: 338045

6 years ago[DAGCombine] optimizeSetCCOfSignedTruncationCheck(): handle ule,ugt CondCodes.
Roman Lebedev [Thu, 26 Jul 2018 17:34:28 +0000 (17:34 +0000)]
[DAGCombine] optimizeSetCCOfSignedTruncationCheck(): handle ule,ugt CondCodes.

Summary:
A follow-up for D49266 / rL337166.

At least one of these cases is more canonical,
so we really do have to handle it.
https://godbolt.org/g/pkzP3X
https://rise4fun.com/Alive/pQyhZZ

We won't get to these cases with I1 being -1,
as that will be constant-folded to true or false.

I'm also not sure we actually hit the 'ule' case,
but i think the worst think that could happen is that being dead code.

Reviewers: spatel, craig.topper, RKSimon, javed.absar, efriedma

Reviewed By: spatel

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D49497

llvm-svn: 338044

6 years agoFix duplicate suggestions after an ambiguous command
Raphael Isemann [Thu, 26 Jul 2018 17:14:18 +0000 (17:14 +0000)]
Fix duplicate suggestions after an ambiguous command

Summary:
So far lldb is printing this when it finds an ambiguous command:
```
(lldb) g
Ambiguous command 'g'. Possible matches:
        gdb-remote
        gui
        gdb-remote
        gui
```
The duplicates come from the fact that we call the same query twice with the same parameters
and add it to the same list. This patch just removes the second query call to `GetCommandObject`.

As `GetCommandObject` is const and the name parameter is also not modified, this shouldn't break
anything else. I didn't merge the remaining if statement into the else as I think otherwise the
`if obj==nullptr do X else Y` pattern in there becomes hard to recognize.

Reviewers: davide

Reviewed By: davide

Subscribers: lldb-commits

Differential Revision: https://reviews.llvm.org/D49866

llvm-svn: 338043

6 years ago[COFF] clean up global resources after completion
Rui Ueyama [Thu, 26 Jul 2018 17:11:24 +0000 (17:11 +0000)]
[COFF] clean up global resources after completion

Patch by Andrew Kelley.

Previously, running lld::coff::link() twice in the same process would
access stale pointers because of these global variables not being reset.
After this patch, lld::coff::link() can be called any number of times,
just like its ELF and MACH-O counterparts.

Differential Revision: https://reviews.llvm.org/D49856

llvm-svn: 338042

6 years ago[CodeGen][ObjC] Make block copy/dispose helper functions exception-safe.
Akira Hatanaka [Thu, 26 Jul 2018 16:51:21 +0000 (16:51 +0000)]
[CodeGen][ObjC] Make block copy/dispose helper functions exception-safe.

When an exception is thrown in a block copy helper function, captured
objects that have previously been copied should be destructed or
released. Similarly, captured objects that are yet to be released should
be released when an exception is thrown in a dispose helper function.

rdar://problem/42410255

Differential Revision: https://reviews.llvm.org/D49718

llvm-svn: 338041

6 years agoDon't print two errors for unknown commands.
Raphael Isemann [Thu, 26 Jul 2018 16:32:05 +0000 (16:32 +0000)]
Don't print two errors for unknown commands.

Summary:
We always print two error messages when we hit an unknown command. As the function
`CommandInterpreter::HandleCommand` that prints the second error message unconditionally called the `CommandInterpreter::ResolveCommandImpl` before (which prints the first error message), we can just remove
that second error message.

Fixes https://bugs.llvm.org/show_bug.cgi?id=38312

Reviewers: labath

Reviewed By: labath

Subscribers: labath, lldb-commits

Differential Revision: https://reviews.llvm.org/D49831

llvm-svn: 338040

6 years ago[DEBUGINFO, NVPTX] Emit correct debug information for local variables.
Alexey Bataev [Thu, 26 Jul 2018 16:29:52 +0000 (16:29 +0000)]
[DEBUGINFO, NVPTX] Emit correct debug information for local variables.

Summary:
NVPTX target dos not use register-based frame information. Instead it
relies on the artificial local_depot that is used instead of the frame
and the data for variables must be emitted relatively to this
local_depot.

Reviewers: tra, jlebar, echristo

Subscribers: jholewinski, aprantl, JDevlieghere, llvm-commits

Differential Revision: https://reviews.llvm.org/D45963

llvm-svn: 338039

6 years ago[InstCombine] add tests for udiv with common factor; NFC
Sanjay Patel [Thu, 26 Jul 2018 16:14:53 +0000 (16:14 +0000)]
[InstCombine] add tests for udiv with common factor; NFC

This fold is mentioned in PR38239:
https://bugs.llvm.org/show_bug.cgi?id=38239

The general case probably belongs in -reassociate, but given that we do
basic reassociation optimizations similar to this in instcombine already,
we might as well be consistent within instcombine and handle this pattern?

llvm-svn: 338038

6 years ago[clangd] Use 'const Twine&' instead of 'Twine'. NFC
Ilya Biryukov [Thu, 26 Jul 2018 16:13:52 +0000 (16:13 +0000)]
[clangd] Use 'const Twine&' instead of 'Twine'. NFC

To fix clang-tidy warning

llvm-svn: 338037

6 years ago[DEBUGINFO, NVPTX] Set `DW_AT_frame_base` to `DW_OP_call_frame_cfa`.
Alexey Bataev [Thu, 26 Jul 2018 16:10:05 +0000 (16:10 +0000)]
[DEBUGINFO, NVPTX] Set `DW_AT_frame_base` to `DW_OP_call_frame_cfa`.

Summary:
For NVPTX target the value of `DW_AT_frame_base` attribute must be set
to `DW_OP_call_frame_cfa`.

Reviewers: tra, jlebar, echristo

Subscribers: jholewinski, JDevlieghere, llvm-commits

Differential Revision: https://reviews.llvm.org/D45785

llvm-svn: 338036

6 years agoRevert r338027 to pacify build bot
James Henderson [Thu, 26 Jul 2018 15:54:53 +0000 (15:54 +0000)]
Revert r338027 to pacify build bot

llvm-svn: 338035

6 years ago[ADT] Replace std::isprint by llvm::isPrint.
Michael Kruse [Thu, 26 Jul 2018 15:31:41 +0000 (15:31 +0000)]
[ADT] Replace std::isprint by llvm::isPrint.

The standard library functions ::isprint/std::isprint have platform-
and locale-dependent behavior which makes LLVM's output less
predictable. In particular, regression tests my fail depending on the
implementation of these functions.

Implement llvm::isPrint in StringExtras.h with a standard behavior and
replace all uses of ::isprint/std::isprint by a call it llvm::isPrint.
The function is inlined and does not look up language settings so it
should perform better than the standard library's version.

Such a replacement has already been done for isdigit, isalpha, isxdigit
in r314883. gtest does the same in gtest-printers.cc using the following
justification:

    // Returns true if c is a printable ASCII character.  We test the
    // value of c directly instead of calling isprint(), which is buggy on
    // Windows Mobile.
    inline bool IsPrintableAscii(wchar_t c) {
      return 0x20 <= c && c <= 0x7E;
    }

Similar issues have also been encountered by Julia:
https://github.com/JuliaLang/julia/issues/7416

I noticed the problem myself when on Windows isprint('\t') started to
evaluate to true (see https://stackoverflow.com/questions/51435249) and
thus caused several unit tests to fail. The result of isprint doesn't
seem to be well-defined even for ASCII characters. Therefore I suggest
to replace isprint by a platform-independent version.

Differential Revision: https://reviews.llvm.org/D49680

llvm-svn: 338034

6 years ago[UnJ] Common some code. NFC
David Green [Thu, 26 Jul 2018 15:19:07 +0000 (15:19 +0000)]
[UnJ] Common some code. NFC

Create a processHeaderPhiOperands for analysing the instructions
in the aft blocks that must be moved before the loop.

Differential Revision: https://reviews.llvm.org/D49061

llvm-svn: 338033

6 years ago[OPENMP] Force OpenMP 4.5 when compiling for offloading.
Alexey Bataev [Thu, 26 Jul 2018 15:17:38 +0000 (15:17 +0000)]
[OPENMP] Force OpenMP 4.5 when compiling for offloading.

If the user requested compilation for OpenMP with the offloading
support, force the version of the OpenMP standard to 4.5 by default.

llvm-svn: 338032

6 years agodwarfgen: Add support for generating the debug_str_offsets section, take 3
Pavel Labath [Thu, 26 Jul 2018 14:36:07 +0000 (14:36 +0000)]
dwarfgen: Add support for generating the debug_str_offsets section, take 3

Previous version of this patch failed on darwin targets because of
different handling of cross-debug-section relocations. This fixes the
tests to emit the DW_AT_str_offsets_base attribute correctly in both
cases. Since doing this is a non-trivial amount of code, and I'm going
to need it in more than one test, I've added a helper function to the
dwarfgen DIE class to do it.

Original commit message follows:

The motivation for this is D49493, where we'd like to test details of
debug_str_offsets behavior which is difficult to trigger from a
traditional test.

This adds the plubming necessary for dwarfgen to generate this section.
The more interesting changes are:
- I've moved emitStringOffsetsTableHeader function from DwarfFile to
  DwarfStringPool, so I can generate the section header more easily from
  the unit test.
- added a new addAttribute overload taking an MCExpr*. This is used to
  generate the DW_AT_str_offsets_base, which links a compile unit to the
  offset table.

I've also added a basic test for reading and writing DW_form_strx forms.

Reviewers: dblaikie, JDevlieghere, probinson

Subscribers: llvm-commits, aprantl

Differential Revision: https://reviews.llvm.org/D49670

llvm-svn: 338031

6 years ago[test] Disable dsymutil update test on windows
Jonas Devlieghere [Thu, 26 Jul 2018 14:16:19 +0000 (14:16 +0000)]
[test] Disable dsymutil update test on windows

Apparently, the issue with dsymutil update functionality on Windows was
that Windows doesn't like dsymutil renaming files that have open handles
to them. This disables the new accelerator test and updates the comment
in the other two test.

We should be able to enable the tests again once we updated the
implementation to use TempFile::keep() to keep the temporary files in
MachOUtils.

A big thank you to Jeremy Morse from Sony for figuring this out and
bringing it to my attention.

llvm-svn: 338030

6 years agoEnable some pointer authentication instructions for aarch64 v8a targets
Luke Cheeseman [Thu, 26 Jul 2018 14:00:50 +0000 (14:00 +0000)]
Enable some pointer authentication instructions for aarch64 v8a targets

- Some of the v8.3 pointer authentication instruction inhabit the Hint space
- These instructions can be assembled to hint instructions which act as NOP instructions prior to v8.3
- This patch permits using the hint instructions for all v8a targets
- Also, correct the RETA{A,B} instructions to match the instruction attributes of RET (set isTerminator and isBarrier)

Differential Revision: https://reviews.llvm.org/D49786

llvm-svn: 338029

6 years ago[clangd] Fix unit tests for Dex
Kirill Bobyrev [Thu, 26 Jul 2018 14:00:00 +0000 (14:00 +0000)]
[clangd] Fix unit tests for Dex

Iterators took temporary objects in constructors, objects were
invalidated when built with recent Clang which resulted in crashes.

llvm-svn: 338028

6 years agoFix raw_fd_ostream::write_impl hang with large output
James Henderson [Thu, 26 Jul 2018 13:22:07 +0000 (13:22 +0000)]
Fix raw_fd_ostream::write_impl hang with large output

On Windows when raw_fd_ostream::write_impl calls write, a 32 bit input
is required for character count. As a variable with size_t is used for
this argument on x64 integral demotion occurs. In the case of large
files an infinite loop follows.

See PR37926.

This fix allows the output of files larger than previous int32 limit.

Differential Revision: https://reviews.llvm.org/D48948

Patch by Owen Reynolds

Reviewed by: zturner

llvm-svn: 338027

6 years agodwarfgen: Don't create an AsmPrinter with an invalid ObjFile lowering
Pavel Labath [Thu, 26 Jul 2018 13:16:06 +0000 (13:16 +0000)]
dwarfgen: Don't create an AsmPrinter with an invalid ObjFile lowering

The AsmPrinter created in the tests contained an uninitialized
TargetLoweringObjectFile. Things mostly worked regardless, because we
used a separate instance of that class to specify sections to emit.

This rearanges the object construction order so that we can avoid
creating two lowering objects. Instead, we properly initialize the
object in the AsmPrinter, and have the DWARF generator store a pointer
to it.

llvm-svn: 338026

6 years ago[clang-tidy] Fix llvm.org/PR38315 (support type aliases in modernize-shrink-to-fit)
Alexander Kornienko [Thu, 26 Jul 2018 13:13:54 +0000 (13:13 +0000)]
[clang-tidy] Fix llvm.org/PR38315 (support type aliases in modernize-shrink-to-fit)

llvm-svn: 338025

6 years agoAllow thread safety annotation lock upgrading and downgrading.
Aaron Ballman [Thu, 26 Jul 2018 13:03:16 +0000 (13:03 +0000)]
Allow thread safety annotation lock upgrading and downgrading.

Patch thanks to Aaron Puchert!

llvm-svn: 338024

6 years ago[tsan] Fix gcc pedantic warning
Dmitry Vyukov [Thu, 26 Jul 2018 13:02:54 +0000 (13:02 +0000)]
[tsan] Fix gcc pedantic warning

Fix gcc (7.2.0) pedantic warning
warning: extra â€˜;’ [-Wpedantic]

Author: jasonl220 (Jason Lovett)
Review: https://reviews.llvm.org/D49817
llvm-svn: 338023

6 years ago[ASTMatchers] fix the missing documentation for new decltypeType matcher
Jonas Toth [Thu, 26 Jul 2018 13:02:05 +0000 (13:02 +0000)]
[ASTMatchers] fix the missing documentation for new decltypeType matcher

Summary: Regenerate the Matchers documentation, forgotten in the original patch.

Reviewers: alexfh, aaron.ballman

Reviewed By: aaron.ballman

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D49850

llvm-svn: 338022

6 years ago[clangd] Fix (most) naming warnings from clang-tidy. NFC
Ilya Biryukov [Thu, 26 Jul 2018 12:05:31 +0000 (12:05 +0000)]
[clangd] Fix (most) naming warnings from clang-tidy. NFC

llvm-svn: 338021

6 years ago[analyzer] Fixed method to get APSInt model
Mikhail R. Gadelha [Thu, 26 Jul 2018 11:17:13 +0000 (11:17 +0000)]
[analyzer] Fixed method to get APSInt model

Summary:
This patch replaces the current method of getting an `APSInt` from Z3's model by calling generic API method `getBitvector` instead of `Z3_get_numeral_uint64`.

By calling `getBitvector`, there's no need to handle bitvectors with bit width == 128 separately.

And, as a bonus, clang now compiles correctly with Z3 4.7.1.

Reviewers: NoQ, george.karpenkov

Reviewed By: george.karpenkov

Subscribers: xazax.hun, szepet, a.sidorin

Differential Revision: https://reviews.llvm.org/D49818

llvm-svn: 338020

6 years ago[mips] Sign extend i32 return values on MIPS64
Stefan Maksimovic [Thu, 26 Jul 2018 10:59:35 +0000 (10:59 +0000)]
[mips] Sign extend i32 return values on MIPS64

Override getTypeForExtReturn so that functions returning
an i32 typed value have it sign extended on MIPS64.

Also provide patterns to get rid of unneeded sign extensions
for arithmetic instructions which implicitly sign extend
their results.

Differential Revision: https://reviews.llvm.org/D48374

llvm-svn: 338019

6 years agoRevert "[COFF] Use comdat shared constants for MinGW as well"
Martin Storsjo [Thu, 26 Jul 2018 10:48:20 +0000 (10:48 +0000)]
Revert "[COFF] Use comdat shared constants for MinGW as well"

This reverts commit r337951.

While that kind of shared constant generally works fine in a MinGW
setting, it broke some cases of inline assembly that worked before:

$ cat const-asm.c
int MULH(int a, int b) {
    int rt, dummy;
    __asm__ (
        "imull %3"
        :"=d"(rt), "=a"(dummy)
        :"a"(a), "rm"(b)
    );
    return rt;
}
int func(int a) {
    return MULH(a, 1);
}
$ clang -target x86_64-win32-gnu -c const-asm.c -O2
const-asm.c:4:9: error: invalid variant '00000001'
        "imull %3"
        ^
<inline asm>:1:15: note: instantiated into assembly here
        imull __real@00000001(%rip)
                     ^

A similar error is produced for i686 as well. The same test with a
target of x86_64-win32-msvc or i686-win32-msvc works fine.

llvm-svn: 338018

6 years ago[clangd] Proof-of-concept query iterators for Dex symbol index
Kirill Bobyrev [Thu, 26 Jul 2018 10:42:31 +0000 (10:42 +0000)]
[clangd] Proof-of-concept query iterators for Dex symbol index

This patch introduces three essential types of query iterators:
`DocumentIterator`, `AndIterator`, `OrIterator`. It provides a
convenient API for query tree generation and serves as a building block
for the next generation symbol index - Dex. Currently, many
optimizations are missed to improve code readability and to serve as the
reference implementation. Potential improvements are briefly mentioned
in `FIXME`s and will be addressed in the following patches.

Dex RFC in the mailing list:
http://lists.llvm.org/pipermail/clangd-dev/2018-July/000022.html

Iterators, their applications and potential extensions are explained in
detail in the design proposal:
https://docs.google.com/document/d/1C-A6PGT6TynyaX4PXyExNMiGmJ2jL1UwV91Kyx11gOI/edit#heading=h.903u1zon9nkj

Reviewers: ioeric, sammccall, ilya-biryukov

Subscribers: cfe-commits, klimek, jfb, mgrang, mgorny, MaskRay, jkorous,
arphaman

Differential Revision: https://reviews.llvm.org/D49546

llvm-svn: 338017

6 years ago[x86/SLH] Extract the logic to trace predicate state through calls to
Chandler Carruth [Thu, 26 Jul 2018 09:42:57 +0000 (09:42 +0000)]
[x86/SLH] Extract the logic to trace predicate state through calls to
a helper function with a nice overview comment. NFC.

This is a preperatory refactoring to implementing another component of
mitigation here that was descibed in the design document but hadn't been
implemented yet.

llvm-svn: 338016

6 years ago[clangd] Give an example for symbol-builder usage
Kirill Bobyrev [Thu, 26 Jul 2018 09:41:24 +0000 (09:41 +0000)]
[clangd] Give an example for symbol-builder usage

`global-symbol-builder` help message mentions `-executor=<string>`
option, but doesn't give any example of what the value could be

Assuming the most popular use case to be building the whole project
index, help message should probably give an example of such usage.

Reviewers: ioeric

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D49785

llvm-svn: 338015

6 years ago[test] Do dsymutil update in place
Jonas Devlieghere [Thu, 26 Jul 2018 09:23:10 +0000 (09:23 +0000)]
[test] Do dsymutil update in place

Update the dSYM bundle in place when swapping out the accelerator
tables. This should unbreak the windows bot that have been failing with
an access denied.

llvm-svn: 338014

6 years agoAllow users of the GCOV API to extend the FileInfo class to implement custom output...
Marco Castelluccio [Thu, 26 Jul 2018 09:21:56 +0000 (09:21 +0000)]
Allow users of the GCOV API to extend the FileInfo class to implement custom output formats

Summary:
The GCOV API can be used to parse gcda/gcno files, but in order to implement custom output formats, users need to reimplement everything.
If the FileInfo members were protected instead of private, they'd be able to reuse the code.

Reviewers: bogner, davide, scott.smith

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D41802

llvm-svn: 338013

6 years ago[clangd] Do not rebuild AST if inputs have not changed
Ilya Biryukov [Thu, 26 Jul 2018 09:21:07 +0000 (09:21 +0000)]
[clangd] Do not rebuild AST if inputs have not changed

Summary:
If the contents are the same, the update most likely comes from the
fact that compile commands were invalidated. In that case we want to
avoid rebuilds in case the compile commands are actually the same.

Reviewers: ioeric

Reviewed By: ioeric

Subscribers: simark, javed.absar, MaskRay, jkorous, arphaman, jfb, cfe-commits

Differential Revision: https://reviews.llvm.org/D49783

llvm-svn: 338012

6 years ago[AArch64][NFC] Removed tab characters from test files.
Sjoerd Meijer [Thu, 26 Jul 2018 07:59:39 +0000 (07:59 +0000)]
[AArch64][NFC] Removed tab characters from test files.

llvm-svn: 338011

6 years ago[AArch64] Armv8.2-A: add the crypto extensions
Sjoerd Meijer [Thu, 26 Jul 2018 07:13:59 +0000 (07:13 +0000)]
[AArch64] Armv8.2-A: add the crypto extensions

This adds MC support for the crypto instructions that were made optional
extensions in Armv8.2-A (AArch64 only).

Differential Revision: https://reviews.llvm.org/D49370

llvm-svn: 338010

6 years ago[ConstProp] Fix calls-math-finite.ll on FreeBSD
Fangrui Song [Thu, 26 Jul 2018 06:24:11 +0000 (06:24 +0000)]
[ConstProp] Fix calls-math-finite.ll on FreeBSD

FreeBSD's log(3.0) is less precise than glibc and musl.
Let's forgive its rounding error of more than half an ulp.

llvm-svn: 338009

6 years ago[AsmParser] Fix preserve-comments-crlf.s on FreeBSD
Fangrui Song [Thu, 26 Jul 2018 06:07:03 +0000 (06:07 +0000)]
[AsmParser] Fix preserve-comments-crlf.s on FreeBSD

--strip-trailing-cr is a diffutils option which is also available on
BSD-licensed diff introduced in FreeBSD 11.2, however, it has a bug
comparing files mixing \r and \r\n. Use -b (POSIX) instead.

llvm-svn: 338008

6 years ago[X86] Don't use CombineTo to skip adding new nodes to the DAGCombiner worklist in...
Craig Topper [Thu, 26 Jul 2018 05:40:10 +0000 (05:40 +0000)]
[X86] Don't use CombineTo to skip adding new nodes to the DAGCombiner worklist in combineMul.

I'm not sure if this was trying to avoid optimizing the new nodes further or what. Or maybe to prevent a cycle if something tried to reform the multiply? But I don't think its a reliable way to do that. If the user of the expanded multiply is visited by the DAGCombiner after this conversion happens, the DAGCombiner will check its operands, see that they haven't been visited by the DAGCombiner before and it will then add the first node to the worklist. This process will repeat until all the new nodes are visited.

So this seems like an unreliable prevention at best. So this patch just returns the new nodes like any other combine. If this starts causing problems we can try to add target specific nodes or something to more directly prevent optimizations.

Now that we handle the combine normally, we can combine any negates the mul expansion creates into their users since those will be visited now.

llvm-svn: 338007

6 years ago[CMake] Don't generate linker script only when shared library isn't statically linked
Petr Hosek [Thu, 26 Jul 2018 05:10:24 +0000 (05:10 +0000)]
[CMake] Don't generate linker script only when shared library isn't statically linked

Since r337668, we support statically linking dependencies only to shared
or static library. However, that change hasn't updated the check whether
to generate a linker script. We shouldn't generate linker script only in
the case when we aren't statically linked ABI into the shared library.

Differential Revision: https://reviews.llvm.org/D49834

llvm-svn: 338006

6 years agoCopy LLVM CMake configuration for CMake Policy CMP0068
Eric Fiselier [Thu, 26 Jul 2018 05:08:30 +0000 (05:08 +0000)]
Copy LLVM CMake configuration for CMake Policy CMP0068

llvm-svn: 338005

6 years ago[sanitizer] Include signal.h instead of sys/signal.h
Fangrui Song [Thu, 26 Jul 2018 04:50:33 +0000 (04:50 +0000)]
[sanitizer] Include signal.h instead of sys/signal.h

llvm-svn: 338004

6 years ago[Docs] Update of Xray page
David Carlier [Thu, 26 Jul 2018 04:44:31 +0000 (04:44 +0000)]
[Docs] Update of Xray page

Update of supported oses.

Reviewers: dberris

Reviewed By: dberris

Differential Revision: https://reviews.llvm.org/D49789

llvm-svn: 338003

6 years agoBe more consistent about which bool value means an error occurred
Eric Fiselier [Thu, 26 Jul 2018 04:02:06 +0000 (04:02 +0000)]
Be more consistent about which bool value means an error occurred

llvm-svn: 338002

6 years agoCleanup the last_write_time internals
Eric Fiselier [Thu, 26 Jul 2018 03:57:26 +0000 (03:57 +0000)]
Cleanup the last_write_time internals

llvm-svn: 338001

6 years agoCorrect comment about stat truncating st_mtimespec to seconds
Eric Fiselier [Thu, 26 Jul 2018 03:42:25 +0000 (03:42 +0000)]
Correct comment about stat truncating st_mtimespec to seconds

llvm-svn: 338000

6 years agoFix attribute placement WRT extern C
Eric Fiselier [Thu, 26 Jul 2018 03:36:37 +0000 (03:36 +0000)]
Fix attribute placement WRT extern C

llvm-svn: 337999

6 years agoWorkaround OS X 10.11 behavior where stat truncates st_mtimespec to seconds.
Eric Fiselier [Thu, 26 Jul 2018 03:28:48 +0000 (03:28 +0000)]
Workaround OS X 10.11 behavior where stat truncates st_mtimespec to seconds.

llvm-svn: 337998

6 years agoRevert r337981: it breaks the debuginfo-tests
Alex Lorenz [Thu, 26 Jul 2018 03:21:40 +0000 (03:21 +0000)]
Revert r337981: it breaks the debuginfo-tests

This commit caused a regression in the debuginfo-tests:

FAIL: debuginfo-tests :: apple-accel.cpp (40748 of 46595)
llvm-svn: 337997

6 years ago[X86] Remove some unnecessary explicit calls to DCI.AddToWorkList.
Craig Topper [Thu, 26 Jul 2018 03:20:27 +0000 (03:20 +0000)]
[X86] Remove some unnecessary explicit calls to DCI.AddToWorkList.

These calls were making sure some newly created nodes were added to worklist, but the DAGCombiner has internal support for ensuring it has visited all nodes. Any time it visits a node it ensures the operands have been queued to be visited as well. This means if we only need to return the last new node. The DAGCombiner will take care of adding its inputs thus walking backwards through all the new nodes.

llvm-svn: 337996

6 years ago[Support] Introduce createStringError helper function
Victor Leschuk [Thu, 26 Jul 2018 02:21:40 +0000 (02:21 +0000)]
[Support] Introduce createStringError helper function

The function in question is copy-pasted lots of times in DWARF-related classes.
Thus it will make sense to place its implementation into the Support library.

Reviewed by: lhames

Differential Revision: https://reviews.llvm.org/D49824

llvm-svn: 337995

6 years ago[GlobalISel] Fall back to SDISel for swifterror/swiftself attributes.
Amara Emerson [Thu, 26 Jul 2018 01:25:58 +0000 (01:25 +0000)]
[GlobalISel] Fall back to SDISel for swifterror/swiftself attributes.

We don't currently support these, fall back until we do.

llvm-svn: 337994

6 years ago[DWARF v5] Don't report an error when the .debug_rnglists section is empty or non...
Wolfgang Pieb [Thu, 26 Jul 2018 01:12:41 +0000 (01:12 +0000)]
[DWARF v5] Don't report an error when the .debug_rnglists section is empty or non-existent. Fixes PR38297.

Reviewer: JDevlieghere

Differential Revision: https://reviews.llvm.org/D49815

llvm-svn: 337993

6 years ago[LoadStoreVectorizer] Use const reference
Fangrui Song [Thu, 26 Jul 2018 01:11:36 +0000 (01:11 +0000)]
[LoadStoreVectorizer] Use const reference

llvm-svn: 337992

6 years agoAdd print statements to help debugging
Eric Fiselier [Thu, 26 Jul 2018 01:10:50 +0000 (01:10 +0000)]
Add print statements to help debugging

llvm-svn: 337991

6 years ago[libc++] Add hack to allow ubsan to work w/o compiler-rt (__muloti4 is undefined)
Eric Fiselier [Thu, 26 Jul 2018 00:34:50 +0000 (00:34 +0000)]
[libc++] Add hack to allow ubsan to work w/o compiler-rt (__muloti4 is undefined)

Summary:
Using int128_t with UBSAN causes link errors unless compiler-rt is providing the runtime library.
Specifically ubsan generates calls to __muloti4 but libgcc doesn't provide a definition.

In order to avoid this, and allow users to continue using sanitized versions of libc++, this patch introduces a hack.
It adds a cribbed version of  the compiler-rt builtin to the libc++ filesystem sources.

I don't think this approach will work in the long run, but it seems OK for now.

Also see:

https://bugs.llvm.org/show_bug.cgi?id=30643
https://bugs.llvm.org/show_bug.cgi?id=16404

Reviewers: mclow.lists, ldionne, rsmith, jyknight, echristo

Reviewed By: echristo

Subscribers: dberris, cfe-commits

Differential Revision: https://reviews.llvm.org/D49828

llvm-svn: 337990

6 years agoRegUsageInfo: Cleanup; NFC
Matthias Braun [Thu, 26 Jul 2018 00:27:51 +0000 (00:27 +0000)]
RegUsageInfo: Cleanup; NFC

- Remove unnecessary anchor function
- Remove unnecessary override of getAnalysisUsage
- Use reference instead of pointers where things cannot be nullptr
- Use ArrayRef instead of std::vector where possible

llvm-svn: 337989

6 years agoCodeGen.cpp: Sort initializers; NFC
Matthias Braun [Thu, 26 Jul 2018 00:27:49 +0000 (00:27 +0000)]
CodeGen.cpp: Sort initializers; NFC

llvm-svn: 337988

6 years agoInitializePasses: Sort declarations; NFC
Matthias Braun [Thu, 26 Jul 2018 00:27:48 +0000 (00:27 +0000)]
InitializePasses: Sort declarations; NFC

llvm-svn: 337987

6 years agoCodeGen: Cleanup regmask construction; NFC
Matthias Braun [Thu, 26 Jul 2018 00:27:47 +0000 (00:27 +0000)]
CodeGen: Cleanup regmask construction; NFC

- Avoid duplication of regmask size calculation.
- Simplify allocateRegisterMask() call.
- Rename allocateRegisterMask() to allocateRegMask() to be consistent
  with naming in MachineOperand.

llvm-svn: 337986

6 years ago[MCA] Avoid an InstrDesc copy in mca::LSUnit::reserve.
Dean Michael Berris [Thu, 26 Jul 2018 00:02:54 +0000 (00:02 +0000)]
[MCA] Avoid an InstrDesc copy in mca::LSUnit::reserve.

Summary:
InstrDesc contains 4 vectors (as well as some other data), so it's
expensive to copy.

Authored By: orodley

Reviewers: andreadb, mattd, dberris

Reviewed By: mattd, dberris

Subscribers: dberris, gbedwell, llvm-commits

Differential Revision: https://reviews.llvm.org/D49775

llvm-svn: 337985

6 years ago[libc++] Follow-up to r337968: use an explicit cast as suggested by Eric
Alex Lorenz [Wed, 25 Jul 2018 23:59:54 +0000 (23:59 +0000)]
[libc++] Follow-up to r337968: use an explicit cast as suggested by Eric

llvm-svn: 337984

6 years ago[ELF] Fix linker-script-in-search-path.s
Fangrui Song [Wed, 25 Jul 2018 23:29:15 +0000 (23:29 +0000)]
[ELF] Fix linker-script-in-search-path.s

llvm-svn: 337983

6 years ago[CMake] Don't use LIBCXXABI_ENABLE_STATIC option before its declared
Sam Clegg [Wed, 25 Jul 2018 23:13:00 +0000 (23:13 +0000)]
[CMake] Don't use LIBCXXABI_ENABLE_STATIC option before its declared

Summary:
rL337867 introduced two new cmake_dependent_option options:
- LIBCXXABI_INSTALL_STATIC_LIBRARY
- LIBCXXABI_INSTALL_SHARED_LIBRARY

They depend on LIBCXXABI_ENABLE_STATIC and LIBCXXABI_ENABLE_SHARED
and so therefore need to (it seems) come after the declaration of
these two options.

Subscribers: mgorny, aheejin, christof, ldionne, cfe-commits

Differential Revision: https://reviews.llvm.org/D49825

llvm-svn: 337982

6 years ago[DWARF v5] Don't emit multiple DW_AT_rnglists_base attributes. Some refactoring of
Wolfgang Pieb [Wed, 25 Jul 2018 23:03:22 +0000 (23:03 +0000)]
[DWARF v5] Don't emit multiple DW_AT_rnglists_base attributes. Some refactoring of
range lists emissions and added test cases.

Reviewer: dblaikie

Differential Revision: https://reviews.llvm.org/D49522

llvm-svn: 337981

6 years ago[dsymutil] Add support for generating DWARF5 accelerator tables.
Jonas Devlieghere [Wed, 25 Jul 2018 23:01:38 +0000 (23:01 +0000)]
[dsymutil] Add support for generating DWARF5 accelerator tables.

This patch add support for emitting DWARF5 accelerator tables
(.debug_names) from dsymutil. Just as with the Apple style accelerator
tables, it's possible to update existing dSYMs. This patch includes a
test that show how you can convert back and forth between the two types.

If no kind of table is specified, dsymutil will default to generating
Apple-style accelerator tables whenever it finds those in its input. The
same is true when there are no accelerator tables at all. Finally, in
the remaining case, where there's at least one DWARF v5 table and no
Apple ones, the output will contains a DWARF accelerator tables
(.debug_names).

Differential revision: https://reviews.llvm.org/D49137

llvm-svn: 337980

6 years ago[CMake] Don't install c++abi headers in standalone libc++ build
Petr Hosek [Wed, 25 Jul 2018 22:57:39 +0000 (22:57 +0000)]
[CMake] Don't install c++abi headers in standalone libc++ build

This is a refinement on r337833. Previously we were installing two
copies of c++abi headers in libc++ build directory, one in
include/c++build and another one in include/c++/v1. However, the
second copy is unnecessary when building libc++ standalone.

Differential Revision: https://reviews.llvm.org/D49752

llvm-svn: 337979

6 years ago[ODRHash] Support hashing enums.
Richard Trieu [Wed, 25 Jul 2018 22:52:05 +0000 (22:52 +0000)]
[ODRHash] Support hashing enums.

llvm-svn: 337978

6 years agobpf: new option -bpf-expand-memcpy-in-order to expand memcpy in order
Yonghong Song [Wed, 25 Jul 2018 22:40:02 +0000 (22:40 +0000)]
bpf: new option -bpf-expand-memcpy-in-order to expand memcpy in order

Some BPF JIT backends would want to optimize memcpy in their own
architecture specific way.

However, at the moment, there is no way for JIT backends to see memcpy
semantics in a reliable way. This is due to LLVM BPF backend is expanding
memcpy into load/store sequences and could possibly schedule them apart from
each other further. So, BPF JIT backends inside kernel can't reliably
recognize memcpy semantics by peephole BPF sequence.

This patch introduce new intrinsic expand infrastructure to memcpy.

To get stable in-order load/store sequence from memcpy, we first lower
memcpy into BPF::MEMCPY node which then expanded into in-order load/store
sequences in expandPostRAPseudo pass which will happen after instruction
scheduling. By this way, kernel JIT backends could reliably recognize
memcpy through scanning BPF sequence.

This new memcpy expand infrastructure is gated by a new option:

  -bpf-expand-memcpy-in-order

Acked-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Jiong Wang <jiong.wang@netronome.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
llvm-svn: 337977

6 years agoWork around GCC bug in constexpr function
Eric Fiselier [Wed, 25 Jul 2018 22:21:47 +0000 (22:21 +0000)]
Work around GCC bug in constexpr function

llvm-svn: 337976

6 years ago[ELF] Make linker-script-in-search-path.s robust per r337972
Fangrui Song [Wed, 25 Jul 2018 22:11:00 +0000 (22:11 +0000)]
[ELF] Make linker-script-in-search-path.s robust per r337972

llvm-svn: 337975

6 years agoFix GCC build in C++14 w/o c++14 constexpr
Eric Fiselier [Wed, 25 Jul 2018 22:07:36 +0000 (22:07 +0000)]
Fix GCC build in C++14 w/o c++14 constexpr

llvm-svn: 337974

6 years ago[GlobalMerge] Handle llvm.compiler.used correctly.
Eli Friedman [Wed, 25 Jul 2018 22:03:35 +0000 (22:03 +0000)]
[GlobalMerge] Handle llvm.compiler.used correctly.

Reuse the handling for llvm.used, and don't transform such globals.

Fixes a failure on the asan buildbot caused by my previous commit.

llvm-svn: 337973

6 years agoMake a test more robust.
Rui Ueyama [Wed, 25 Jul 2018 22:01:54 +0000 (22:01 +0000)]
Make a test more robust.

Previously, this test fails if there's other test that happens to
create "t.script" in the current directory.

llvm-svn: 337972

6 years agoRemove test which shouldn't have been committed
Eric Fiselier [Wed, 25 Jul 2018 21:58:37 +0000 (21:58 +0000)]
Remove test which shouldn't have been committed

llvm-svn: 337971

6 years agoFix failing test under C++14
Eric Fiselier [Wed, 25 Jul 2018 21:53:43 +0000 (21:53 +0000)]
Fix failing test under C++14

llvm-svn: 337970

6 years ago[ELF] Use search paths for --version-script=
Fangrui Song [Wed, 25 Jul 2018 21:53:18 +0000 (21:53 +0000)]
[ELF] Use search paths for --version-script=

Summary: This behavior matches ld.bfd -Ld --version-script=t.script a.o

Reviewers: ruiu, espindola

Subscribers: emaste, arichardson, llvm-commits

Differential Revision: https://reviews.llvm.org/D49820

llvm-svn: 337969

6 years ago[libc++] Follow-up to r337960: specify lambda's return type to avoid
Alex Lorenz [Wed, 25 Jul 2018 21:50:44 +0000 (21:50 +0000)]
[libc++] Follow-up to r337960: specify lambda's return type to avoid
-Wc++11-narrowing warning on Darwin

The internal CI produced the following diagnostic:
error: non-constant-expression cannot be narrowed from type 'long long' to '__darwin_suseconds_t' (aka 'int') in initializer list [-Wc++11-narrowing]
  struct ::timeval ConvertedTS[2] = {{TS[0].tv_sec, Convert(TS[0].tv_nsec)},
                                                    ^~~~~~~~~~~~~~~~~~~~~~

llvm-svn: 337968

6 years agoELF: Do not ICF SHF_LINK_ORDER sections.
Peter Collingbourne [Wed, 25 Jul 2018 21:40:54 +0000 (21:40 +0000)]
ELF: Do not ICF SHF_LINK_ORDER sections.

We are already ICF'ing these sections as a unit with their dependent
sections, so they don't need to be considered for ICF individually.

This change also "fixes" slowness caused by our quadratic-in-group-size
relocation segregation algorithm on 32-bit ARM platforms with unwind
data and ICF on rodata. In this scenario almost every function's
.ARM.exidx is identical except for the targets of the relocations
that refer to the function and its .ARM.extab, which causes almost
all of the program's .ARM.exidx sections to be initially added to the
same class, which causes us to compare every such section with every
other such section.

Differential Revision: https://reviews.llvm.org/D49716

llvm-svn: 337967

6 years ago[SelectionDAG] try to convert funnel shift directly to rotate if legal
Sanjay Patel [Wed, 25 Jul 2018 21:38:30 +0000 (21:38 +0000)]
[SelectionDAG] try to convert funnel shift directly to rotate if legal

If the DAGCombiner's rotate matching was working as expected,
I don't think we'd see any test diffs here.

This sidesteps the issue of custom lowering for rotates raised in PR38243:
https://bugs.llvm.org/show_bug.cgi?id=38243
...by only dealing with legal operations.

llvm-svn: 337966