Tamas Berghammer [Wed, 10 Feb 2016 10:43:34 +0000 (10:43 +0000)]
Improve the handling of missing elf symtab and missing symbol sizes
* Generate artificial symbol names from eh_fame during symbol parsing
so these symbols are already present when we calcualte the size of
the symbols where 0 is specified.
* Fix symbol size calculation for the last symbol in the file where
it have to last until the end of the parent section.
Differential revision: http://reviews.llvm.org/D16996
llvm-svn: 260369
Tamas Berghammer [Wed, 10 Feb 2016 10:42:13 +0000 (10:42 +0000)]
Fix handling of the arm IT instruction in the unwinder
The IT instruction can specify condition code for up to 4 consecutive
instruction and it is used quite often by clang in epilogues causing
an issue when trying to unwind from locations covered by the IT
instruction and for locatins inmediately after the IT instruction.
Changes made to fix it:
* Introduce the concept of conditional instruction block what is a list
of consecutive instructions with the same condition. We update the
unwind information during the conditional instruction block and when
we reach the end of it (first instruction with a differemt condition)
then we restore the unwind information we had before the condition.
* Fix a bug in the ARM instruction emulator where neither PC nor the
ITSTATE was advanced when we reached an instruction what we can't
decode.
After the change we have no regression on android-arm running the
regular test suit and TestStandardUnwind also passes when running it
with clang as the compiler (previously it failed on an IT instruction).
Differential revision: http://reviews.llvm.org/D16814
llvm-svn: 260368
Tamas Berghammer [Wed, 10 Feb 2016 10:42:00 +0000 (10:42 +0000)]
Use BKPT instead of UDF for arm/thumb breakpoints
The UDF instruction is deprecated in armv7 and in case of thumb2
instructions set it don't work well together with the IT instruction.
Differential revision: http://reviews.llvm.org/D16853
llvm-svn: 260367
Tamas Berghammer [Wed, 10 Feb 2016 10:35:48 +0000 (10:35 +0000)]
Add new option to lldb-server to display its version
llvm-svn: 260366
Simon Atanasyan [Wed, 10 Feb 2016 10:08:39 +0000 (10:08 +0000)]
[ELF][MIPS] Group and sort initialization of MipsTargetInfo class fields. NFC
llvm-svn: 260365
Simon Atanasyan [Wed, 10 Feb 2016 10:08:35 +0000 (10:08 +0000)]
[ELF][MIPS] Rename MipsTargetInfo::relocateOne argument SA to S
MIPS 32-bit ABI uses REL relocation record format. We read addends from
relocation destinations right in the MipsTargetInfo::relocateOne
function.
llvm-svn: 260364
Dylan McKay [Wed, 10 Feb 2016 08:55:23 +0000 (08:55 +0000)]
[AVR] Add instruction definitions
Summary: Add the AVR instruction tablegen definitions.
Reviewers: stoklund, hfinkel, dsanders, arsenm, vkalintiris
Subscribers: dylanmckay, agnat, rjordans, llvm-commits
Differential Revision: http://reviews.llvm.org/D15703
llvm-svn: 260363
Mohit K. Bhakkad [Wed, 10 Feb 2016 06:58:13 +0000 (06:58 +0000)]
[LLDB][MIPS] Generalise MIPS arch names
Patch by Nitesh Jain
Reviewers: clayborg, jaydeep.
Subscribers: zturner, bhushan, mohit.bhakkad, sagar, lldb-commits.
Differential Revision: http://reviews.llvm.org/D16840
llvm-svn: 260362
Xinliang David Li [Wed, 10 Feb 2016 06:36:55 +0000 (06:36 +0000)]
[PGO] fix prof symbol lookup bug
Patch by Rong Xu
The problem is exposed by intra-module indirect call promotion where
prof symtab is created from module which does not contain all symbols
from the program. With partial symtab, the result needs to be checked
more strictly.
llvm-svn: 260361
Akira Hatanaka [Wed, 10 Feb 2016 06:06:06 +0000 (06:06 +0000)]
[Sema] Issue a warning for integer overflow in nested struct initializer
r257357 fixed clang to warn on integer overflow in struct initializers.
However, it didn't warn when a struct had a nested initializer. This
commit makes changes in Sema::CheckForIntOverflow to handle nested
initializers.
For example:
struct s {
struct t {
unsigned x;
} t;
} s = {
{
.x = 4 * 1024 * 1024 * 1024
}
};
rdar://problem/
23526454
llvm-svn: 260360
JF Bastien [Wed, 10 Feb 2016 04:04:12 +0000 (04:04 +0000)]
X86: Remove useless semicolon
llvm-svn: 260359
Jim Ingham [Wed, 10 Feb 2016 03:25:24 +0000 (03:25 +0000)]
This is an idea to make "thread step-in --target" work for the common
case where you have:
1 -> foo (bar(),
2 baz(),
3 lala());
4
You are sitting on line 1, and want to step into foo, but not bar, baz & lala. Unfortunately
there are line table entries for lines 1-3, and lldb doesn't know anything about the nesting
of statement in these lines. So we'll have to use the user's intelligence... This patch adds:
(lldb) thread step-in -t foo --end-line 4
That tells lldb to keep stepping in till line 4, but stop if you step into foo. I think I would
remember to use this when faced with some of the long gnarly call sequences in lldb. But there
might be ways I haven't thought of to make it more convenient. Jason suggests having "end" as a
special token for --end-line which just means keep going to the end of the function, I really want
to get into this thing...
There should be an SB API and tests, which will come if this seems useful.
llvm-svn: 260352
NAKAMURA Takumi [Wed, 10 Feb 2016 03:09:13 +0000 (03:09 +0000)]
llvm-config: Fix DLL prefix on cygming.
For example, LLVMSupport takes the form of LLVMSupport.dll, not libLLVMSupport.dll.
llvm-svn: 260351
Chris Bieneman [Wed, 10 Feb 2016 02:17:21 +0000 (02:17 +0000)]
[CMake] For multi-stage builds to be deterministic we need to disable timestamps.
Duh! With this change I've verified -O3 builds are deterministic.
llvm-svn: 260350
Enrico Granata [Wed, 10 Feb 2016 02:12:42 +0000 (02:12 +0000)]
Change lldb.value.__int__() so that it takes into account the signedness of the value being cast to return a Python number with the proper value
The explicit APIs on SBValue obviously remain if one wants to be explicit in intent, or override this guess, but since __int__() has to pick one, an educated guess is definitely better than than always going to signed regardless
Fixes rdar://
24556976
llvm-svn: 260349
Paul Robinson [Wed, 10 Feb 2016 02:08:24 +0000 (02:08 +0000)]
Get rid of another SAME-NOT. FileCheck does not have this suffix.
Differential Revision: http://reviews.llvm.org/D17062
llvm-svn: 260348
NAKAMURA Takumi [Wed, 10 Feb 2016 01:45:51 +0000 (01:45 +0000)]
clang/test/Index/skip-parsed-bodies/compile_commands.json: Appease targeting msvc.
llvm-svn: 260347
Justin Lebar [Wed, 10 Feb 2016 01:34:17 +0000 (01:34 +0000)]
Remove some unnecessary CHECK-SAMEs in ModuleDebugInfo.{cpp,m}.
Suggested by Paul Robinson.
llvm-svn: 260346
Jim Ingham [Wed, 10 Feb 2016 01:33:58 +0000 (01:33 +0000)]
Add a missing break.
llvm-svn: 260345
NAKAMURA Takumi [Wed, 10 Feb 2016 01:29:57 +0000 (01:29 +0000)]
libclang: Enable skip-parsed-bodies on win32.
I guess it would be working since Rafael's r187619.
llvm-svn: 260344
NAKAMURA Takumi [Wed, 10 Feb 2016 01:12:55 +0000 (01:12 +0000)]
llvm-config: Quick fix for cross compilation. Don't be confsed between host and target. It has been there since r252532.
FIXME: The clause may use conditions of host compiler, not HOST_TRIPLE.
llvm-svn: 260343
Tom Stellard [Wed, 10 Feb 2016 01:10:09 +0000 (01:10 +0000)]
StructurizeCFG: Initialize SkipUniformRegions in the default constructor
This should fix some random bot failures caused by r260336.
llvm-svn: 260342
Chris Bieneman [Wed, 10 Feb 2016 01:09:59 +0000 (01:09 +0000)]
[CMake] Pass LLVM_EXTERNAL_*_SOURCE_DIR variables to subsequent stages
For multi-stage builds we need to pass any overridden source directory variables. Without passing these the subsequent stages won't find the project sources.
llvm-svn: 260341
Chris Bieneman [Wed, 10 Feb 2016 01:09:56 +0000 (01:09 +0000)]
[CMake] Fixing the 3-stage cmake cache.
I had hoped this would work from a single cache file, but turns out there is a bug I can't quite figure out relating to passing list arguments to recursive CMake invocations.
This change works around that.
llvm-svn: 260340
Reid Kleckner [Wed, 10 Feb 2016 01:06:37 +0000 (01:06 +0000)]
Revert "Turn on LLVM_ENABLE_PLUGINS by default on Windows"
This reverts r260262, it broke some LLVM tests (bugpoint).
llvm-svn: 260339
Lang Hames [Wed, 10 Feb 2016 01:02:33 +0000 (01:02 +0000)]
[Orc] Add lazy-JITting support for i386.
This patch adds a new class, OrcI386, which contains the hooks needed to
support lazy-JITing on i386 (currently only for Pentium 2 or above, as the JIT
re-entry code uses the FXSAVE/FXRSTOR instructions).
Support for i386 is enabled in the LLI lazy JIT and the Orc C API, and
regression and unit tests are enabled for this architecture.
llvm-svn: 260338
Richard Smith [Wed, 10 Feb 2016 00:59:02 +0000 (00:59 +0000)]
Fix overload sets of strchr, strpbrk, strrchr, memchr and strstr from
<string.h> and wcschr, wcspbrk, wcsrchr, wmemchr, and wcsstr from <wchar.h> to
provide a const-correct overload set even when the underlying C library does
not.
This change adds a new macro, _LIBCPP_PREFERRED_OVERLOAD, which (if defined)
specifies that a given overload is a better match than an otherwise equally
good function declaration without the overload. This is implemented in modern
versions of Clang via __attribute__((enable_if)), and not elsewhere.
We use this new macro to define overloads in the global namespace for these
functions that displace the overloads provided by the C library, unless we
believe the C library is already providing the correct signatures.
llvm-svn: 260337
Tom Stellard [Wed, 10 Feb 2016 00:39:37 +0000 (00:39 +0000)]
StructurizeCFG: Add an option for skipping regions with only uniform branches
Summary:
Tests for this will be added once the AMDGPU backend enables this
option.
Reviewers: arsenm
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D16602
llvm-svn: 260336
Amaury Sechet [Wed, 10 Feb 2016 00:38:50 +0000 (00:38 +0000)]
Add support for struct in C API test
Summary: As per title. This also include extra support for insertvalue and extracvalue.
Reviewers: bogner, chandlerc, echristo, dblaikie, joker.eph, Wallbraker
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D17055
llvm-svn: 260335
Justin Lebar [Wed, 10 Feb 2016 00:38:15 +0000 (00:38 +0000)]
Get rid of CHECK-SAME-NOT in tests.
Summary: This isn't a FileCheck directive; it does nothing.
Reviewers: jroelofs
Subscribers: cfe-commits, majnemer
Differential Revision: http://reviews.llvm.org/D17051
llvm-svn: 260334
Ekaterina Romanova [Wed, 10 Feb 2016 00:12:24 +0000 (00:12 +0000)]
This patch adds doxygen comments for all the intrinsincs in the header file f16cintrin.h. The doxygen comments are automatically generated based on Sony's intrinsics document.
Differential Revision: http://reviews.llvm.org/D17021
llvm-svn: 260333
Amaury Sechet [Wed, 10 Feb 2016 00:09:37 +0000 (00:09 +0000)]
Add binding for counting argument and find called value of call and invoke instruction from the C API.
Summary: As per title. This remove the need to rely on internal knowledge of call and invoke instruction to find called value and argument count.
Reviewers: bogner, chandlerc, echristo, dblaikie, joker.eph, Wallbraker
Differential Revision: http://reviews.llvm.org/D17054
llvm-svn: 260332
Adrian McCarthy [Wed, 10 Feb 2016 00:06:50 +0000 (00:06 +0000)]
Don't dereference the first element of an empty container.
llvm-svn: 260331
Ed Maste [Wed, 10 Feb 2016 00:03:52 +0000 (00:03 +0000)]
Remove FreeBSD failure decorator from TestCppIncompleteTypes
CFLAGS is now being set correctly to pass -flimit-debug-info or
-fno-limit-debug-info on FreeBSD. I'm not sure which change is
responsible for the fix, though.
llvm.org/pr25626
llvm-svn: 260330
Tim Northover [Tue, 9 Feb 2016 23:55:48 +0000 (23:55 +0000)]
ARM: improve documentation slightly.
It seems the ARMv8 instruction set overview is no longer provided by ARM, so
I've removed it. Since most of the other documents were the same I unified the
two sections.
llvm-svn: 260329
Adrian McCarthy [Tue, 9 Feb 2016 23:52:57 +0000 (23:52 +0000)]
Add status for LLDB on Windows to status.html.
Differential Revision: http://reviews.llvm.org/D16902
llvm-svn: 260328
Kostya Serebryany [Tue, 9 Feb 2016 23:46:43 +0000 (23:46 +0000)]
[asan] update the scariness score: tweak a few weights and add tests
llvm-svn: 260327
Zachary Turner [Tue, 9 Feb 2016 23:45:21 +0000 (23:45 +0000)]
Change `CoreTests` to LLDBCoreTests to avoid name clash.
lld was already using a target named CoreTests so CMake
was erroring due to this conflict.
llvm-svn: 260326
Sanjay Patel [Tue, 9 Feb 2016 23:44:35 +0000 (23:44 +0000)]
[x86] convert masked load of exactly one element to scalar load
This is the load counterpart to the store optimization that was added in:
http://reviews.llvm.org/rL260145
llvm-svn: 260325
Amaury Sechet [Tue, 9 Feb 2016 23:41:20 +0000 (23:41 +0000)]
Split clone value and instruction in the echo C API test
Summary: This is a bit of refactoring required to be able to generate instruction in forward basic block. This, for instance, is a requirement for phi in loops.
Reviewers: bogner, chandlerc, echristo, dblaikie, joker.eph, Wallbraker
Differential Revision: http://reviews.llvm.org/D17050
llvm-svn: 260324
Vasileios Kalintiris [Tue, 9 Feb 2016 23:38:28 +0000 (23:38 +0000)]
Revert "Introduce a cmake module to figure out whether we need to link with libatomic."
This reverts commit r260235. It breaks LLVM's bootstrap when building
with a -gcc-toolchain and the system's gcc installation does not provide
the libatomic library and its headers. We should check whether
LIBCXX_GCC_TOOLCHAIN is set and adjust the flags accordingly.
llvm-svn: 260323
Greg Clayton [Tue, 9 Feb 2016 23:25:54 +0000 (23:25 +0000)]
Added code that was commented out during testing to stops template member functions from being added to class definitions (see revision 260308 for details).
<rdar://problem/
24483905>
<rdar://problem/
24508374>
llvm-svn: 260322
Amaury Sechet [Tue, 9 Feb 2016 23:15:02 +0000 (23:15 +0000)]
Add icmp and conditional branches in the C API echo test.
Summary:
Improving coverage.
Depends on D16912 .
Reviewers: bogner, chandlerc, echristo, dblaikie, joker.eph, Wallbraker
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D16937
llvm-svn: 260321
Rui Ueyama [Tue, 9 Feb 2016 23:11:21 +0000 (23:11 +0000)]
Rangefy, and replace a switch with `if`s. NFC.
llvm-svn: 260320
Justin Lebar [Tue, 9 Feb 2016 23:03:22 +0000 (23:03 +0000)]
Add convergent-removing bits to FunctionAttrs pass.
Summary:
Remove the convergent attribute on any functions which provably do not
contain or invoke any convergent functions.
After this change, we'll be able to modify clang to conservatively add
'convergent' to all functions when compiling CUDA.
Reviewers: jingyue, joker.eph
Subscribers: llvm-commits, tra, jhen, hfinkel, resistor, chandlerc, arsenm
Differential Revision: http://reviews.llvm.org/D17013
llvm-svn: 260319
Justin Lebar [Tue, 9 Feb 2016 23:03:17 +0000 (23:03 +0000)]
Update documentation about convergent attribute.
Summary:
Be more explicit about what 'convergent' means, and indicate that the
compiler may remove the attribute from a function if it can prove that
the function doesn't in fact execute any convergent ops.
Reviewers: resistor, jingyue, joker.eph
Subscribers: hfinkel, chandlerc, arsenm, jhen, tra, llvm-commits
Differential Revision: http://reviews.llvm.org/D17012
llvm-svn: 260318
Peter Collingbourne [Tue, 9 Feb 2016 23:01:38 +0000 (23:01 +0000)]
Fix GCC build.
llvm-svn: 260317
Ahmed Bougacha [Tue, 9 Feb 2016 22:54:12 +0000 (22:54 +0000)]
[CodeGen] Prefer "if (SDValue R = ...)" to "if (R.getNode())". NFCI.
llvm-svn: 260316
Ahmed Bougacha [Tue, 9 Feb 2016 22:54:05 +0000 (22:54 +0000)]
[X86] Don't reuse an unrelated variable, create a new one. NFC.
Using Op makes it look like we're doing something with it.
We're really not.
llvm-svn: 260315
Ahmed Bougacha [Tue, 9 Feb 2016 22:53:58 +0000 (22:53 +0000)]
[X86] Remove unnecessary assignment. NFC.
llvm-svn: 260314
Amaury Sechet [Tue, 9 Feb 2016 22:50:53 +0000 (22:50 +0000)]
Add C binding for AllocaInst::getAllocatedType
Summary:
Comes with an awesome test.
Depends on D16912
Reviewers: bogner, chandlerc, echristo, dblaikie, joker.eph, Wallbraker
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D16942
llvm-svn: 260313
Peter Collingbourne [Tue, 9 Feb 2016 22:50:34 +0000 (22:50 +0000)]
WholeProgramDevirt: introduce.
This pass implements whole program optimization of virtual calls in cases
where we know (via bitset information) that the list of callees is fixed. This
includes the following:
- Single implementation devirtualization: if a virtual call has a single
possible callee, replace all calls with a direct call to that callee.
- Virtual constant propagation: if the virtual function's return type is an
integer <=64 bits and all possible callees are readnone, for each class and
each list of constant arguments: evaluate the function, store the return
value alongside the virtual table, and rewrite each virtual call as a load
from the virtual table.
- Uniform return value optimization: if the conditions for virtual constant
propagation hold and each function returns the same constant value, replace
each virtual call with that constant.
- Unique return value optimization for i1 return values: if the conditions
for virtual constant propagation hold and a single vtable's function
returns 0, or a single vtable's function returns 1, replace each virtual
call with a comparison of the vptr against that vtable's address.
Differential Revision: http://reviews.llvm.org/D16795
llvm-svn: 260312
Richard Smith [Tue, 9 Feb 2016 22:48:14 +0000 (22:48 +0000)]
Update documentation to reflect that libc++abi provides __cxa_thread_atexit (and has for quite a while). Also document that we have not yet implemented the new inheriting constructor rules.
llvm-svn: 260311
Rafael Espindola [Tue, 9 Feb 2016 22:47:34 +0000 (22:47 +0000)]
Fix inverted comparison.
llvm-svn: 260310
Amaury Sechet [Tue, 9 Feb 2016 22:36:41 +0000 (22:36 +0000)]
Improve the C API echo test tool to emit basic block is the right order.
Summary: As per title. Also add a facility method to get the name of a basic block from the C API.
Reviewers: bogner, chandlerc, echristo, dblaikie, joker.eph, Wallbraker
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D16912
llvm-svn: 260309
Greg Clayton [Tue, 9 Feb 2016 22:36:24 +0000 (22:36 +0000)]
Fixed many issues that were causing differing type definition issues to show up when parsing expressions.
1) Turns out we weren't correctly uniquing types for C++. We would search our repository for "lldb_private::Process", but yet store just "Process" in the unique type map. Now we store things correctly and correctly unique types.
2) SymbolFileDWARF::CompleteType() can be called at any time in order to complete a C++ or Objective C class. All public inquiries into the SymbolFile go through SymbolVendor, and SymbolVendor correctly takes the module lock before it call the SymbolFile API call, but when we let CompilerType objects out in the wild, they can complete themselves at any time from the expression parser, so the ValueObjects or (SBValue objects in the public API), and many more places. So we now take the module lock when completing a type to avoid two threads being in the SymbolFileDWARF at the same time.
3) If a class has a template member function like:
class A
{
<template T>
void Foo(T t);
};
The DWARF will _only_ contain a DW_TAG_subprogram for "Foo" if anyone specialized it. This would cause a class definition for A inside a.cpp that used a "int" and "float" overload to look like:
class A
{
void Foo(int t);
void Foo(double t);
};
And a version from b.cpp that used a "float" overload to look like:
class A
{
void Foo(float t);
};
And a version from c.cpp that use no overloads to look like:
class A
{
};
Then in an expression if you have two variables, one name "a" from a.cpp in liba.dylib, and one named "b" from b.cpp in libb.dylib, you will get conflicting definitions for "A" and your expression will fail. This all stems from the fact that DWARF _only_ emits template specializations, not generic definitions, and they are only emitted if they are used. There are two solutions to this:
a) When ever you run into ANY class, you must say "just because this class doesn't have templatized member functions, it doesn't mean that any other instances might not have any, so when ever I run into ANY class, I must parse all compile units and parse all instances of class "A" just in case it has member functions that are templatized.". That is really bad because it means you always pull in ALL DWARF that contains most likely exact duplicate definitions of the class "A" and you bloat the memory that the SymbolFileDWARF plug-in uses in LLDB (since you pull in all DIEs from all compile units that contain a "A" definition) uses for little value most of the time.
b) Modify DWARF to emit generic template member function definitions so that you know from looking at any instance of class "A" wether it has template member functions or not. In order to do this, we would have to have the ability to correctly parse a member function template, but there is a compiler bug:
<rdar://problem/
24515533> [PR 26553] C++ Debug info should reference DW_TAG_template_type_parameter
This bugs means that not all of the info needed to correctly make a template member function is in the DWARF. The main source of the problem is if we have DWARF for a template instantiation for "int" like: "void A::Foo<int>(T)" the DWARF comes out as "void A::Foo<int>(int)" (it doesn't mention type "T", it resolves the type to the specialized type to "int"). But if you actually have your function defined as "<template T> void Foo(int t)" and you only use T for local variables inside the function call, we can't correctly make the function prototype up in the clang::ASTContext.
So the best we can do for now we just omit all member functions that are templatized from the class definition so that "A" never has any template member functions. This means all defintions of "A" look like:
class A
{
};
And our expressions will work. You won't be able to call template member fucntions in expressions (not a regression, we weren't able to do this before) and if you are stopped in a templatized member function, we won't know that are are in a method of class "A". All things we should fix, but we need <rdar://problem/
24515533> fixed first, followed by:
<rdar://problem/
24515624> Classes should always include a template subprogram definition, even when no template member functions are used
before we can do anything about it in LLDB.
This bug mainly fixed the following Apple radar:
<rdar://problem/
24483905>
llvm-svn: 260308
Richard Smith [Tue, 9 Feb 2016 22:34:35 +0000 (22:34 +0000)]
PR26349: correctly check whether a digit sequence is empty in the presence of digit separators.
llvm-svn: 260307
Jonathan Peyton [Tue, 9 Feb 2016 22:32:41 +0000 (22:32 +0000)]
Proxy task fix: task_state stack push condition on fork
The problem is that the master's thread state was not saved before entering a
parallel region so it does not remember tasks when it returns.
llvm-svn: 260306
Simon Atanasyan [Tue, 9 Feb 2016 22:31:49 +0000 (22:31 +0000)]
[mips] Extend MipsAsmParser class to handle %got(sym + const) expressions
Now the parser supports `%got(sym)` expressions only but `%got(sym + const)`
variant is also valid and accepted by GAS.
Differential Revision: http://reviews.llvm.org/D16885
llvm-svn: 260305
Jan Vesely [Tue, 9 Feb 2016 22:17:48 +0000 (22:17 +0000)]
configure: Remove llvm 3.6 defines
we require llvm 3.7
reviewer: tstellard
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 260304
Jan Vesely [Tue, 9 Feb 2016 22:17:46 +0000 (22:17 +0000)]
configure: Remove cl_khr_fp64 for device that don't support doubles
Also remove definitions if provided by clang (3.7+)
This halves the size of builtin.opt.{cedar,barts}.bc
reviewer: tstellard
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 260303
Jan Vesely [Tue, 9 Feb 2016 22:17:45 +0000 (22:17 +0000)]
configure: Introduce per device defines
Make cl_khr_fp64 define per-device.
This patch does not change the generated Makefile (for llvm 3.6, 3.7)
v2: Make the device defines per LLVM version, 'all' for all versions
reviewer: tstellard
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 260302
Jan Vesely [Tue, 9 Feb 2016 22:17:42 +0000 (22:17 +0000)]
math: Fix log2 vectorization on non-fp64 hw
reviewer: tstellard
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 260301
Jonathan Peyton [Tue, 9 Feb 2016 22:15:30 +0000 (22:15 +0000)]
Have Mac builds use @rpath when supported in CMake
The -install_name linker flag will use "@rpath/" when supported in CMake
which is the recommended usage for dynamic libraries on Mac OSX.
llvm-svn: 260300
Greg Clayton [Tue, 9 Feb 2016 22:09:26 +0000 (22:09 +0000)]
Don't complete a class type just to test if it is a class. Code in CompilerType and in clang::QualType knows how to complete a type if it needs to.
llvm-svn: 260299
Sanjoy Das [Tue, 9 Feb 2016 21:54:18 +0000 (21:54 +0000)]
[MemoryBuiltins] Fix an issue with hasNoAliasAttr
Summary:
`hasNoAliasAttr` is buggy: it checks to see if the called function has
a `noalias` attribute, which is incorrect since functions are not even
allowed to have the `noalias` attribute. The comment on its only
caller, `llvm::isNoAliasFn`, makes it pretty clear that the intention
to do the `noalias` check on the return value, and not the callee.
Unfortunately I couldn't find a way to test this upstream -- fixing
this does not change the observable behavior of any of the passes that
use this. This is not very surprising, since `noalias` does not tell
anything about the contents of the allocated memory (so, e.g., you
still cannot fold loads). I'll be happy to be proven wrong though.
Reviewers: chandlerc, reames
Subscribers: mcrosier, llvm-commits
Differential Revision: http://reviews.llvm.org/D17037
llvm-svn: 260298
Rui Ueyama [Tue, 9 Feb 2016 21:46:11 +0000 (21:46 +0000)]
ELF: Add 'using namespace llvm::dwarf'.
Most constants defined in llvm::dwarf namespace start with "DW_".
We don't have to worry about name conflicts.
llvm-svn: 260297
Philip Reames [Tue, 9 Feb 2016 21:43:12 +0000 (21:43 +0000)]
[Verifier] Add checks for masked.load and masked.store intrinsics
While trying to track down what appears to be a LoopVectorizer bug, I noticed that we had no validation of the correctness of calls emitted to @llvm.masked.load and @llvm.masked.store. This meant malformed IR was showing up much much later than it should. Hopefully, having Verifier rules in place will make this easier to isolate.
llvm-svn: 260296
Sanjay Patel [Tue, 9 Feb 2016 21:42:04 +0000 (21:42 +0000)]
[SelectionDAG] make getMemBasePlusOffset() accessible; NFCI
I reinvented this functionality in http://reviews.llvm.org/D16828 because it was
hidden away as a static function. The changes in x86 are not based on a complete
audit. I suspect there are other possible uses there, and there are almost certainly
more potential users in other targets.
llvm-svn: 260295
Rui Ueyama [Tue, 9 Feb 2016 21:41:01 +0000 (21:41 +0000)]
ELF: Add .eh_frame 'P', 'R' and 'L' augmentation characters.
This basically reverts commit r260073 because it is found that
augmentation strings don't always start with "zR". It is reported
as https://llvm.org/bugs/show_bug.cgi?id=26541.
llvm-svn: 260294
Rafael Espindola [Tue, 9 Feb 2016 21:39:49 +0000 (21:39 +0000)]
This brings back commit r259578.
But now it follows the llvm style, uses an early return and doesn't
include a file named 1.o.
llvm-svn: 260293
Zachary Turner [Tue, 9 Feb 2016 21:36:38 +0000 (21:36 +0000)]
Fix incorrect usage of `str` in _match_decorator_property.
llvm-svn: 260292
Zachary Turner [Tue, 9 Feb 2016 21:36:32 +0000 (21:36 +0000)]
Remove skipIf<compiler> decorators.
These were supposed to have been removed in a previous patch,
but I missed them.
llvm-svn: 260291
Zachary Turner [Tue, 9 Feb 2016 21:36:23 +0000 (21:36 +0000)]
Remove decorators related to debug info types.
All existing usages were ported over to the common decorators.
llvm-svn: 260290
Zachary Turner [Tue, 9 Feb 2016 21:36:15 +0000 (21:36 +0000)]
Delete all the xfail / skip decorators for specific compilers.
Ported everything over to using expectedFailureAll.
llvm-svn: 260289
Zachary Turner [Tue, 9 Feb 2016 21:36:02 +0000 (21:36 +0000)]
Remove expected failure and skip decorators for host os.
All decorator invocations are updated to use the generic
expectedFailureAll and skipIf decorators.
llvm-svn: 260288
Rafael Espindola [Tue, 9 Feb 2016 21:35:14 +0000 (21:35 +0000)]
Fix name style.
Functions start with a lower case, variables with an upper case.
llvm-svn: 260287
Rafael Espindola [Tue, 9 Feb 2016 21:32:56 +0000 (21:32 +0000)]
Update comment style.
llvm-svn: 260286
Rafael Espindola [Tue, 9 Feb 2016 21:30:10 +0000 (21:30 +0000)]
Revert "Correct size calculations for ELF files"
This reverts commit r259578.
There are enough issues with this small patch that it is better to
revert and then commit a fixed version (will be committed shortly).
llvm-svn: 260285
Greg Clayton [Tue, 9 Feb 2016 21:20:17 +0000 (21:20 +0000)]
If we set the DYLD_INSERT_LIBRARIES environment variable when launching debugserver, for use with /usr/lib/libgmalloc.dylib, then make sure we don't pass this environment variable on to any child processes.
llvm-svn: 260284
Chad Rosier [Tue, 9 Feb 2016 21:20:12 +0000 (21:20 +0000)]
[AArch64] This check is specific to merging instructions. NFC.
llvm-svn: 260283
Greg Clayton [Tue, 9 Feb 2016 21:18:44 +0000 (21:18 +0000)]
Don't enable -flimit-debug-info by default for the LLDB Xcode builds. This was put in back in 2011 when -flimit-debug-info did a lot less than it does now. With llvm-dsymutil doing ODR type unique stuff we still have reasonably sized dSYM files.
llvm-svn: 260282
Ehsan Akhgari [Tue, 9 Feb 2016 21:17:58 +0000 (21:17 +0000)]
Remove system-libs.test for now
This test fails in the ninja-x64-msvc-RA-centos6 builder, so the
UNSUPPORTED: system-windows condition is insufficient. Removing it
for now; I will investigate how this can be fixed later.
llvm-svn: 260281
Ehsan Akhgari [Tue, 9 Feb 2016 21:10:22 +0000 (21:10 +0000)]
Fix a test added in r260263
llvm-svn: 260280
Philip Reames [Tue, 9 Feb 2016 21:09:22 +0000 (21:09 +0000)]
[InstCombine][GC] Handle gc.relocations of vector type
We introduced gc.relocates of vector-of-pointer types a couple of weeks back. Somehow, I missed updating the InstCombine rule to account for this. If we hit this code path with a vector-of-pointers gc.relocate, we'd crash on a cast<PointerType>.
I also took the chance to do a bit of code style cleanup.
llvm-svn: 260279
Richard Smith [Tue, 9 Feb 2016 21:04:04 +0000 (21:04 +0000)]
Add Tooling functionality to get a name for a QualType that can be used to name
that type from the global scope.
Patch by Sterling Augustine, derived (with permission) from code from Cling by
Vassil Vassilev and Philippe Canal.
llvm-svn: 260278
Richard Smith [Tue, 9 Feb 2016 20:59:05 +0000 (20:59 +0000)]
Simplify and rename ASTMatchFinder's getCXXRecordDecl to make it more obvious
what it's actually trying to do.
llvm-svn: 260277
Ehsan Akhgari [Tue, 9 Feb 2016 20:49:24 +0000 (20:49 +0000)]
Fix the test added in r260266
llvm-svn: 260276
Geoff Berry [Tue, 9 Feb 2016 20:47:21 +0000 (20:47 +0000)]
[AArch64] AArch64LoadStoreOptimizer: fix bug in pre-inc check iterator
Summary:
Fix case where a pre-inc/dec load/store would not be formed if the
add/sub that forms the inc/dec part of the operation was the first
instruction in the block being examined.
Reviewers: mcrosier, jmolloy, t.p.northover, junbuml
Subscribers: aemerson, rengolin, mcrosier, llvm-commits
Differential Revision: http://reviews.llvm.org/D16785
llvm-svn: 260275
Chad Rosier [Tue, 9 Feb 2016 20:44:41 +0000 (20:44 +0000)]
[AArch64] Bail even earlier if the instructions modifieds the base register. NFC.
llvm-svn: 260274
Chad Rosier [Tue, 9 Feb 2016 20:27:45 +0000 (20:27 +0000)]
[AArch64] Simplify. NFC.
llvm-svn: 260273
Chad Rosier [Tue, 9 Feb 2016 20:18:07 +0000 (20:18 +0000)]
[AArch64] Add an assert to ensure we don't scale an offset that can't be scaled.
llvm-svn: 260272
Pete Cooper [Tue, 9 Feb 2016 20:11:17 +0000 (20:11 +0000)]
Use __nl_symbol_ptr instead of __got in the stubs pass on x86 archs.
The non lazy atoms generated in the stubs pass use an image cache to
hold all of the pointers. On arm archs, this is the __got section,
but on x86 archs it should be __nl_symbol_ptr.
rdar://problem/
24572729
llvm-svn: 260271
Xinliang David Li [Tue, 9 Feb 2016 20:02:59 +0000 (20:02 +0000)]
[PGO] Fix issue: explicitly defaulted assignop is not profiled
Differential Revision: http://reviews.llvm.org/D16947
llvm-svn: 260270
Ehsan Akhgari [Tue, 9 Feb 2016 20:02:35 +0000 (20:02 +0000)]
Fix a typo in r260263
llvm-svn: 260269
Reid Kleckner [Tue, 9 Feb 2016 19:53:30 +0000 (19:53 +0000)]
Pass /bigobj when building lib/ASTMatchers/Dynamic/Registry.cpp
This is the third time it has crossed the 2^16 section limit. We've
already spent time optimizing this file to reduce template
instantiations, and it's not clear that there is anymore low hanging
fruit.
llvm-svn: 260267
Ehsan Akhgari [Tue, 9 Feb 2016 19:43:13 +0000 (19:43 +0000)]
clang-cl: Enable plugins on Windows
llvm-svn: 260266
Ehsan Akhgari [Tue, 9 Feb 2016 19:43:11 +0000 (19:43 +0000)]
clang-cl: Support loading plugins on Windows
This builds on the support being added to LLVM to import and export
registries from DLLs. This will allow us to pick up the registry
entries added in the DLL's copy of FrontendPluginRegistry.
This will allow us to use plugins on Windows using:
$ clang-cl -Xclang -load -Xclang plugin.dll \
-Xclang -add-plugin -Xclang foo
llvm-svn: 260265
Chad Rosier [Tue, 9 Feb 2016 19:42:19 +0000 (19:42 +0000)]
[AArch64] Add a FIXME about invalid KILL markers after the ld/st opt pass.
llvm-svn: 260264
Ehsan Akhgari [Tue, 9 Feb 2016 19:41:14 +0000 (19:41 +0000)]
llvm-config: Add preliminary Windows support
Summary:
This patch adds Windows support for a few of the llvm-config commands,
including cflags, ldflags, libs, and system-libs.
Currently llvm-config is untested, so this patch adds tests for the
commands that it fixes as well.
Reviewers: rnk
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D16762
llvm-svn: 260263