James Y Knight [Sat, 2 Feb 2019 01:48:23 +0000 (01:48 +0000)]
Remove redundant FunctionDecl argument from a couple functions.
This argument was added in r254554 in order to support the
pass_object_size attribute. However, in r296076, the attribute's
presence is now also represented in FunctionProtoType's
ExtParameterInfo, and thus it's unnecessary to pass along a separate
FunctionDecl.
The functions modified are:
RequiredArgs::forPrototype{,Plus}, and
CodeGenTypes::ConvertFunctionType.
After this, it's also (again) unnecessary to have a separate
ConvertFunctionType function ConvertType, so convert callers back to
the latter, leaving the former as an internal helper function.
Further reviews (D57594, D57615) have revealed that this was not reviewed,
and that the differential's description was not read during the review,
thus rendering this commit invalid.
Dan Gohman [Fri, 1 Feb 2019 22:25:23 +0000 (22:25 +0000)]
[WebAssembly] Add an import_field function attribute
This is similar to import_module, but sets the import field name
instead.
By default, the import field name is the same as the C/asm/.o symbol
name. However, there are situations where it's useful to have it be
different. For example, suppose I have a wasm API with a module named
"pwsix" and a field named "read". There's no risk of namespace
collisions with user code at the wasm level because the generic name
"read" is qualified by the module name "pwsix". However in the C/asm/.o
namespaces, the module name is not used, so if I have a global function
named "read", it is intruding on the user's namespace.
With the import_field module, I can declare my function (in libc) to be
"__read", and then set the wasm import module to be "pwsix" and the wasm
import field to be "read". So at the C/asm/.o levels, my symbol is
outside the user namespace.
Michael Kruse [Fri, 1 Feb 2019 20:25:04 +0000 (20:25 +0000)]
[OpenMP 5.0] Parsing/sema support for "omp declare mapper" directive.
This patch implements parsing and sema for "omp declare mapper"
directive. User defined mapper, i.e., declare mapper directive, is a new
feature in OpenMP 5.0. It is introduced to extend existing map clauses
for the purpose of simplifying the copy of complex data structures
between host and device (i.e., deep copy). An example is shown below:
struct S { int len; int *d; };
#pragma omp declare mapper(struct S s) map(s, s.d[0:s.len]) // Memory region that d points to is also mapped using this mapper.
Contributed-by: Lingda Li <lildmh@gmail.com>
Differential Revision: https://reviews.llvm.org/D56326
Max Moroz [Fri, 1 Feb 2019 17:12:35 +0000 (17:12 +0000)]
Update SanitizerCoverage doc regarding the issue with pc-table and gc-sections.
Summary:
There is a bug for this: https://bugs.llvm.org/show_bug.cgi?id=34636
But it would be also helpful to leave a note in the docs to prevent users from
running into issues, e.g. https://crbug.com/926588.
Summary:
I'm working on a clang-tidy check, much like existing [[ http://clang.llvm.org/extra/clang-tidy/checks/bugprone-exception-escape.html | bugprone-exception-escape ]],
to detect when an exception might escape out of an OpenMP construct it isn't supposed to escape from.
For that i will be using the `nothrow` bit of `CapturedDecl`s.
While that bit is already correctly set for some constructs, e.g. `#pragma omp parallel`: https://godbolt.org/z/2La7pv
it isn't set for the `#pragma omp sections`, or `#pragma omp section`: https://godbolt.org/z/qZ-EbP
If i'm reading [[ https://www.openmp.org/wp-content/uploads/OpenMP-API-Specification-5.0.pdf | `OpenMP Application Programming Interface Version 5.0 November 2018` ]] correctly,
they should be, as per `2.8.1 sections Construct`, starting with page 86:
* The sections construct is a non-iterative worksharing construct that contains a set of **structured blocks**
that are to be distributed among and executed by the threads in a team. Each **structured block** is executed
once by one of the threads in the team in the context of its implicit task.
* The syntax of the sections construct is as follows:
#pragma omp sections [clause[ [,] clause] ... ] new-line
{
[#pragma omp section new-line]
**structured-block**
...
* Description
Each **structured block** in the sections construct is preceded by a section directive except
possibly **the first block**, for which a preceding section directive is optional.
* Restrictions
• The code enclosed in a sections construct must be a **structured block**.
* A throw executed inside a sections region must cause execution to resume within the same
section of the sections region, and the same thread that threw the exception must catch it.
Stefan Granitz [Fri, 1 Feb 2019 15:35:25 +0000 (15:35 +0000)]
[CMake] External compiler-rt-configure requires LLVMTestingSupport when including tests
Summary:
Apparently `LLVMTestingSupport` must be built before `llvm-config` can be asked for it. Symptom with `LLVM_INCLUDE_TESTS=ON` is:
```
$ ./path/to/llvm-build/bin/llvm-config --ldflags --libs testingsupport
-L/path/to/llvm-build/lib -Wl,-search_paths_first -Wl,-headerpad_max_install_names
llvm-config: error: component libraries and shared library
With `LLVMTestingSupport` as dependency of `compiler-rt-configure` we get the expected behavior:
```
$ ./path/to/llvm-build/bin/llvm-config --ldflags --libs testingsupport
-L/path/to/llvm-build/lib -Wl,-search_paths_first -Wl,-headerpad_max_install_names
-lLLVMTestingSupport -lLLVMSupport -lLLVMDemangle
```
Ilya Biryukov [Fri, 1 Feb 2019 11:20:13 +0000 (11:20 +0000)]
Disable tidy checks with too many hits
Summary:
Some tidy checks have too many hits in the codebase, making it hard to spot
other results from clang-tidy, therefore rendering the tool less useful.
Two checks were disabled:
- misc-non-private-member-variable-in-classes in the whole LLVM monorepo,
it is very common to have those in LLVM and the style guide does not forbid
them.
- readability-identifier-naming in the clang subtree. There are thousands of
violations in 'Sema.h' alone.
Before the change, 'Sema.h' had >1000 tidy warnings, after the change the number
dropped to 3 warnings (unterminated namespace comments).
Yevgeny Rouban [Fri, 1 Feb 2019 10:44:43 +0000 (10:44 +0000)]
Provide reason messages for unviable inlining
InlineCost's isInlineViable() is changed to return InlineResult
instead of bool. This provides messages for failure reasons and
allows to get more specific messages for cases where callsites
are not viable for inlining.
Serge Guelton [Fri, 1 Feb 2019 06:11:44 +0000 (06:11 +0000)]
Fix isInSystemMacro to handle pasted macros
Token pasted by the preprocessor (through ##) have a Spelling pointing to scratch buffer.
As a result they are not recognized at system macro, even though the pasting happened in
a system macro. Fix that by looking into the parent macro if the original lookup finds a
scratch buffer.
Brian Gesiak [Fri, 1 Feb 2019 03:30:29 +0000 (03:30 +0000)]
[SemaCXX] Param diagnostic matches overload logic
Summary:
Given the following test program:
```
class C {
public:
int A(int a, int& b);
};
int C::A(const int a, int b) {
return a * b;
}
```
Clang would produce an error message that correctly diagnosed the
redeclaration of `C::A` to not match the original declaration (the
parameters to the two declarations do not match -- the original takes an
`int &` as its 2nd parameter, but the redeclaration takes an `int`). However,
it also produced a note diagnostic that inaccurately pointed to the
first parameter, claiming that `const int` in the redeclaration did not
match the unqualified `int` in the original. The diagnostic is
misleading because it has nothing to do with why the program does not
compile.
The logic for checking for a function overload, in
`Sema::FunctionParamTypesAreEqual`, discards cv-qualifiers before
checking whether the types are equal. Do the same when producing the
overload diagnostic.
Julian Lettner [Fri, 1 Feb 2019 02:51:00 +0000 (02:51 +0000)]
[Sanitizers] UBSan unreachable incompatible with ASan in the presence of `noreturn` calls
Summary:
UBSan wants to detect when unreachable code is actually reached, so it
adds instrumentation before every unreachable instruction. However, the
optimizer will remove code after calls to functions marked with
noreturn. To avoid this UBSan removes noreturn from both the call
instruction as well as from the function itself. Unfortunately, ASan
relies on this annotation to unpoison the stack by inserting calls to
_asan_handle_no_return before noreturn functions. This is important for
functions that do not return but access the the stack memory, e.g.,
unwinder functions *like* longjmp (longjmp itself is actually
"double-proofed" via its interceptor). The result is that when ASan and
UBSan are combined, the noreturn attributes are missing and ASan cannot
unpoison the stack, so it has false positives when stack unwinding is
used.
Changes:
Clang-CodeGen now directly insert calls to `__asan_handle_no_return`
when a call to a noreturn function is encountered and both
UBsan-unreachable and ASan are enabled. This allows UBSan to continue
removing the noreturn attribute from functions without any changes to
the ASan pass.
James Y Knight [Fri, 1 Feb 2019 02:28:03 +0000 (02:28 +0000)]
[opaque pointer types] Add a FunctionCallee wrapper type, and use it.
Recommit r352791 after tweaking DerivedTypes.h slightly, so that gcc
doesn't choke on it, hopefully.
Original Message:
The FunctionCallee type is effectively a {FunctionType*,Value*} pair,
and is a useful convenience to enable code to continue passing the
result of getOrInsertFunction() through to EmitCall, even once pointer
types lose their pointee-type.
Then:
- update the CallInst/InvokeInst instruction creation functions to
take a Callee,
- modify getOrInsertFunction to return FunctionCallee, and
- update all callers appropriately.
One area of particular note is the change to the sanitizer
code. Previously, they had been casting the result of
`getOrInsertFunction` to a `Function*` via
`checkSanitizerInterfaceFunction`, and storing that. That would report
an error if someone had already inserted a function declaraction with
a mismatching signature.
However, in general, LLVM allows for such mismatches, as
`getOrInsertFunction` will automatically insert a bitcast if
needed. As part of this cleanup, cause the sanitizer code to do the
same. (It will call its functions using the expected signature,
however they may have been declared.)
Finally, in a small number of locations, callers of
`getOrInsertFunction` actually were expecting/requiring that a brand
new function was being created. In such cases, I've switched them to
Function::Create instead.
[analyzer] [RetainCountChecker] Fix object type for CF/Obj-C bridged casts
Having an incorrect type for a cast causes the checker to incorrectly
dismiss the operation under ARC, leading to a false positive
use-after-release on the test.
Akira Hatanaka [Fri, 1 Feb 2019 00:12:06 +0000 (00:12 +0000)]
Revert "[Sema] Make canPassInRegisters return true if the CXXRecordDecl passed"
This reverts commit r350920 as it is not clear whether we should force a
class to be returned in registers when copy and move constructors are
both deleted.
For more background, see the following discussion:
http://lists.llvm.org/pipermail/cfe-commits/Week-of-Mon-20190128/259907.html
[sanitizer-coverage] prune trace-cmp instrumentation for CMP isntructions that feed into the backedge branch. Instrumenting these CMP instructions is almost always useless (and harmful) for fuzzing
Nico Weber [Thu, 31 Jan 2019 22:15:32 +0000 (22:15 +0000)]
Make clang/test/Index/pch-from-libclang.c pass in more places
- fixes the test on macOS with LLVM_ENABLE_PIC=OFF
- together with D57343, gets the test to pass on Windows
- makes it run everywhere (it seems to just pass on Linux)
The main change is to pull out the resource directory computation into a
function shared by all 3 places that do it. In CIndexer.cpp, this now works no
matter if libclang is in lib/ or bin/ or statically linked to a binary in bin/.
Artem Belevich [Thu, 31 Jan 2019 21:34:03 +0000 (21:34 +0000)]
[CUDA] add support for the new kernel launch API in CUDA-9.2+.
Instead of calling CUDA runtime to arrange function arguments,
the new API constructs arguments in a local array and the kernels
are launched with __cudaLaunchKernel().
The old API has been deprecated and is expected to go away
in the next CUDA release.
Artem Belevich [Thu, 31 Jan 2019 21:32:24 +0000 (21:32 +0000)]
[CUDA] Propagate detected version of CUDA to cc1
..and use it to control that parts of CUDA compilation
that depend on the specific version of CUDA SDK.
This patch has a placeholder for a 'new launch API' support
which is in a separate patch. The list will be further
extended in the upcoming patch to support CUDA-10.1.
James Y Knight [Thu, 31 Jan 2019 20:35:56 +0000 (20:35 +0000)]
[opaque pointer types] Add a FunctionCallee wrapper type, and use it.
The FunctionCallee type is effectively a {FunctionType*,Value*} pair,
and is a useful convenience to enable code to continue passing the
result of getOrInsertFunction() through to EmitCall, even once pointer
types lose their pointee-type.
Then:
- update the CallInst/InvokeInst instruction creation functions to
take a Callee,
- modify getOrInsertFunction to return FunctionCallee, and
- update all callers appropriately.
One area of particular note is the change to the sanitizer
code. Previously, they had been casting the result of
`getOrInsertFunction` to a `Function*` via
`checkSanitizerInterfaceFunction`, and storing that. That would report
an error if someone had already inserted a function declaraction with
a mismatching signature.
However, in general, LLVM allows for such mismatches, as
`getOrInsertFunction` will automatically insert a bitcast if
needed. As part of this cleanup, cause the sanitizer code to do the
same. (It will call its functions using the expected signature,
however they may have been declared.)
Finally, in a small number of locations, callers of
`getOrInsertFunction` actually were expecting/requiring that a brand
new function was being created. In such cases, I've switched them to
Function::Create instead.
Ilya Biryukov [Thu, 31 Jan 2019 20:20:32 +0000 (20:20 +0000)]
[CodeComplete] Propagate preferred types through parser in more cases
Preferred types are used by code completion for ranking. This commit
considerably increases the number of points in code where those types
are propagated.
In order to avoid complicating signatures of Parser's methods, a
preferred type is kept as a member variable in the parser and updated
during parsing.
Rafael Auler [Thu, 31 Jan 2019 09:38:31 +0000 (09:38 +0000)]
Support attribute used in member funcs of class templates
Summary:
As PR17480 describes, clang does not support the used attribute
for member functions of class templates. This means that if the member
function is not used, its definition is never instantiated. This patch
changes clang to emit the definition if it has the used attribute.
Petr Hosek [Thu, 31 Jan 2019 06:21:01 +0000 (06:21 +0000)]
[CMake] Unify scripts for generating VCS headers
Previously, there were two different scripts for generating VCS headers:
one used by LLVM and one used by Clang. They were both similar, but
different. They were both broken in their own ways, for example the one
used by Clang didn't properly handle monorepo resulting in an incorrect
version information reported by Clang.
This change unifies two the scripts by introducing a new script that's
used from both LLVM and Clang, ensures that the new script supports both
monorepo and standalone SVN and Git setups, and removes the old scripts.
Julian Lettner [Wed, 30 Jan 2019 23:42:13 +0000 (23:42 +0000)]
[Sanitizers] UBSan unreachable incompatible with ASan in the presence of `noreturn` calls
Summary:
UBSan wants to detect when unreachable code is actually reached, so it
adds instrumentation before every unreachable instruction. However, the
optimizer will remove code after calls to functions marked with
noreturn. To avoid this UBSan removes noreturn from both the call
instruction as well as from the function itself. Unfortunately, ASan
relies on this annotation to unpoison the stack by inserting calls to
_asan_handle_no_return before noreturn functions. This is important for
functions that do not return but access the the stack memory, e.g.,
unwinder functions *like* longjmp (longjmp itself is actually
"double-proofed" via its interceptor). The result is that when ASan and
UBSan are combined, the noreturn attributes are missing and ASan cannot
unpoison the stack, so it has false positives when stack unwinding is
used.
Changes:
Clang-CodeGen now directly insert calls to `__asan_handle_no_return`
when a call to a noreturn function is encountered and both
UBsan-unreachable and ASan are enabled. This allows UBSan to continue
removing the noreturn attribute from functions without any changes to
the ASan pass.
Erik Pilkington [Wed, 30 Jan 2019 23:17:38 +0000 (23:17 +0000)]
[CodeGenObjC] Handle exceptions when calling objc_alloc or objc_allocWithZone
objc_alloc and objc_allocWithZone may throw exceptions if the
underlying method does. If we're in a @try block, then make sure we
emit an invoke instead of a call.
Alexey Bataev [Wed, 30 Jan 2019 20:49:52 +0000 (20:49 +0000)]
[OPENMP]Fix PR40536: Do not emit __kmpc_push_target_tripcount if not
required.
Function __kmpc_push_target_tripcount should be emitted only if the
offloading entry is going to be emitted (for use in tgt_target...
functions). Otherwise, it should not be emitted.
Erik Pilkington [Wed, 30 Jan 2019 20:34:53 +0000 (20:34 +0000)]
Add a new builtin: __builtin_dynamic_object_size
This builtin has the same UI as __builtin_object_size, but has the
potential to be evaluated dynamically. It is meant to be used as a
drop-in replacement for libraries that use __builtin_object_size when
a dynamic checking mode is enabled. For instance,
__builtin_object_size fails to provide any extra checking in the
following function:
void f(size_t alloc) {
char* p = malloc(alloc);
strcpy(p, "foobar"); // expands to __builtin___strcpy_chk(p, "foobar", __builtin_object_size(p, 0))
}
This is an overflow if alloc < 7, but because LLVM can't fold the
object size intrinsic statically, it folds __builtin_object_size to
-1. With __builtin_dynamic_object_size, alloc is passed through to
__builtin___strcpy_chk.
Erik Pilkington [Wed, 30 Jan 2019 20:34:35 +0000 (20:34 +0000)]
Add a 'dynamic' parameter to the objectsize intrinsic
This is meant to be used with clang's __builtin_dynamic_object_size.
When 'true' is passed to this parameter, the intrinsic has the
potential to be folded into instructions that will be evaluated
at run time. When 'false', the objectsize intrinsic behaviour is
unchanged.
Yaxun Liu [Wed, 30 Jan 2019 12:26:54 +0000 (12:26 +0000)]
[HIP] Fix size_t for MSVC environment
In 64 bit MSVC environment size_t is defined as unsigned long long.
In single source language like HIP, data layout should be consistent
in device and host compilation, therefore copy data layout controlling
fields from Aux target for AMDGPU target.
[OpenCL] Add generic addr space to the return of implicit assignment.
When creating the prototype of implicit assignment operators the
returned reference to the class should be qualified with the same
addr space as 'this' (i.e. __generic in OpenCL).
Michal Gorny [Wed, 30 Jan 2019 08:20:24 +0000 (08:20 +0000)]
[clang] [Driver] [NetBSD] Append -rpath for shared compiler-rt runtimes
Append appropriate -rpath when using shared compiler-rt runtimes,
e.g. '-fsanitize=address -shared-libasan'. There's already a similar
logic in CommonArgs.cpp but it uses non-standard arch-suffixed
installation directory while we want our driver to work with standard
installation paths.
We don't need to use the predetermined data-sharing attributes for the
loop counters if the user explicitly specified correct data-sharing
attributes for such variables.
Matt Arsenault [Tue, 29 Jan 2019 20:49:54 +0000 (20:49 +0000)]
OpenCL: Use length modifier for warning on vector printf arguments
Re-enable format string warnings on printf.
The warnings are still incomplete. Apparently it is undefined to use a
vector specifier without a length modifier, which is not currently
warned on. Additionally, type warnings appear to not be working with
the hh modifier, and aren't warning on all of the special restrictions
from c99 printf.
George Karpenkov [Tue, 29 Jan 2019 19:29:19 +0000 (19:29 +0000)]
Extend AnyCall to handle callable declarations without the call expressions
That weakens inner invariants, but allows the class to be more generic,
allowing usage in situations where the call expression is not known (or
should not matter).
Alexey Bataev [Tue, 29 Jan 2019 18:51:58 +0000 (18:51 +0000)]
[OPENMP]Make the loop with unsigned counter countable.
According to the report, better to keep the original strict compare
operation as the loop condition with unsigned loop counters to make the
loop countable. This allows further loop transformations.
James Y Knight [Tue, 29 Jan 2019 16:37:27 +0000 (16:37 +0000)]
Adjust documentation for git migration.
This fixes most references to the paths:
llvm.org/svn/
llvm.org/git/
llvm.org/viewvc/
github.com/llvm-mirror/
github.com/llvm-project/
reviews.llvm.org/diffusion/
to instead point to https://github.com/llvm/llvm-project.
This is *not* a trivial substitution, because additionally, all the
checkout instructions had to be migrated to instruct users on how to
use the monorepo layout, setting LLVM_ENABLE_PROJECTS instead of
checking out various projects into various subdirectories.
I've attempted to not change any scripts here, only documentation. The
scripts will have to be addressed separately.
Additionally, I've deleted one document which appeared to be outdated
and unneeded:
lldb/docs/building-with-debug-llvm.txt
Yaxun Liu [Tue, 29 Jan 2019 13:20:23 +0000 (13:20 +0000)]
[CUDA][HIP] Do not diagnose use of _Float16
r352221 caused regressions in CUDA/HIP since device function may use _Float16 whereas host does not support it.
In this case host compilation should not diagnose usage of _Float16 in device functions or variables.
For now just do not diagnose _Float16 for CUDA/HIP. In the future we should have more precise check.
Shafik Yaghmour [Mon, 28 Jan 2019 21:55:33 +0000 (21:55 +0000)]
[ASTImporter] Fix handling of overriden methods during ASTImport
Summary:
When importing classes we may add a CXXMethodDecl more than once to a CXXRecordDecl when handling overrides. This patch will fix the cases we currently know about and handle the case where we are only dealing with declarations.