The taskloop construct specifies that the iterations of one or more associated loops will be executed in parallel using OpenMP tasks. The iterations are distributed across tasks created by the construct and scheduled to be executed.
The next code will be generated for the taskloop directive:
#pragma omp taskloop num_tasks(N) lastprivate(j)
for( i=0; i<N*GRAIN*STRIDE-1; i+=STRIDE ) {
int th = omp_get_thread_num();
#pragma omp atomic
counter++;
#pragma omp atomic
th_counter[th]++;
j = i;
}
Adrian Prantl [Sun, 24 Apr 2016 22:22:29 +0000 (22:22 +0000)]
Debug info: Apply an empty debug location for global OpenMP destructors.
LLVM really wants a debug location on every inlinable call in a function
with debug info, because it otherwise cannot set up inlining debug info.
This change applies an artificial line 0 debug location (which is how
DWARF marks automatically generated code that has no corresponding
source code) to the .__kmpc_global_dtor_. functions to avoid the
LLVM Verifier complaining.
Martin Probst [Sun, 24 Apr 2016 22:05:09 +0000 (22:05 +0000)]
clang-format: [JS] generator and async functions.
For generators, see:
https://developer.mozilla.org/en-US/docs/Web/JavaScript/Guide/Iterators_and_generators
async functions are not quite in the spec yet, but stage 3 and already widely used:
http://tc39.github.io/ecmascript-asyncawait/
Improve diagnostic checking for va_start to also warn on other instances of undefined behavior, such as a parameter declared with the register keyword in C, or a parameter of a type that undergoes default argument promotion.
This helps cover some more of the CERT secure coding rule EXP58-CPP. Pass an object of the correct type to va_start (https://www.securecoding.cert.org/confluence/display/cplusplus/EXP58-CPP.+Pass+an+object+of+the+correct+type+to+va_start).
DebugInfo: Adapt to loss of DITypeRef in LLVM r267296
LLVM stopped using MDString-based type references, and DIBuilder no
longer fills 'retainedTypes:' with every DICompositeType that has an
'identifier:' field. There are just minor changes to keep the same
behaviour in CFE.
Leaving 'retainedTypes:' unfilled has a dramatic impact on the output
order of the IR though. There are a huge number of testcase changes,
which were unfortunately not really scriptable.
Rong Xu [Fri, 22 Apr 2016 21:19:05 +0000 (21:19 +0000)]
PGO] PGOFuncName meta data if PGOFuncName is different from function's raw name
Write out the PGOFuncName meta data if PGOFuncName is different from function's
raw name. This should only apply to internal linkage functions. This is to be
consumed by indirect-call promotion when called in LTO optimization pass.
Fix a bug involving deferred decl emission and PCH
For various reasons, involving dllexport and class linkage compuations,
we have to wait until after the semicolon after a class declaration to
emit inline methods. These are "deferred" decls. Before this change,
finishing the tag decl would trigger us to deserialize some PCH so that
we could make a "pretty" IR-level type. Deserializing the PCH triggered
calls to HandleTopLevelDecl, which, when done, checked the deferred decl
list, and emitted some dllexported decls that weren't ready.
Avoid this re-entrancy. Deferred decls should not get emitted when a tag
is finished, they should only be emitted after a real top level decl in
the main file.
[OPENMP] Fix for PR27463: Privatizing struct fields with array type
causes code generation failure.
The codegen part of firstprivate clause for member decls used type of
original variable without skipping reference type from
OMPCapturedExprDecl. Patch fixes this problem.
[OPENMP] Fix for LCV in simd directives in explicit clauses.
If loop control variable for simd-based directives is explicitly marked
as linear/lastprivate in clauses, codegen for such construct would
crash. Patch fixes this problem.
Summary:
Adds a framework to enable the instrumentation pass for the new
EfficiencySanitizer ("esan") family of tools. Adds a flag for esan's
cache fragmentation tool via -fsanitize=efficiency-cache-frag.
Adds appropriate tests for the new flag.
Richard Trieu [Thu, 21 Apr 2016 21:04:55 +0000 (21:04 +0000)]
Split interesting warnings off from -Wfloat-conversion
Restructure the implict floating point to integer conversions so that
interesting sub-groups are under different flags. Breakdown of warnings:
No warning:
Exact conversions from floating point to integer:
int x = 10.0;
int x = 1e10;
-Wliteral-conversion - Floating point literal to integer with rounding:
int x = 5.5;
int x = -3.4;
-Wfloat-conversion - All conversions not covered by the above two:
int x = GetFloat();
int x = 5.5 + 3.5;
-Wfloat-zero-conversion - The expression converted has a non-zero floating
point value that gets converted to a zero integer value, excluded the cases
falling under -Wliteral-conversion. Subset of -Wfloat-conversion.
int x = 1.0 / 2.0;
-Wfloat-overflow-conversion - The floating point value is outside the range
of the integer type, exluding cases from -Wliteral conversion. Subset of
-Wfloat-conversion.
char x = 500;
char x = -1000;
-Wfloat-bool-conversion - Any conversion of a floating point type to bool.
Subset of -Wfloat-conversion.
if (GetFloat()) {}
bool x = 5.0;
-Wfloat-bool-constant-conversion - Conversion of a compile time evaluatable
floating point value to bool. Subset of -Wfloat-bool-conversion.
bool x = 1.0;
bool x = 4.0 / 20.0;
Also add EvaluateAsFloat to Sema, which is similar to EvaluateAsInt, but for
floating point values.
Richard Smith [Thu, 21 Apr 2016 01:46:37 +0000 (01:46 +0000)]
[modules] Make the tweak to avoid circular inclusion of emmintrin.h and
xmmintrin.h a bit more directed. If for whatever reason modules are enabled but
we textually include one of these headers, don't deploy the special case for
modules. To make this work cleanly, extend __building_module to be defined
even when modules is disabled.
Bob Wilson [Thu, 21 Apr 2016 00:11:24 +0000 (00:11 +0000)]
Remove the (ignored) -Wreceived-is-weak diagnostic.
We kept this around for a while since Xcode 6 and earlier had a build
setting for this warning. It was removed in Xcode 7 so there should be
no need for this warning now.
Adrian Prantl [Wed, 20 Apr 2016 23:59:32 +0000 (23:59 +0000)]
Module Debugging: Emit the canonical debug info for Objective-C classes
in the compile unit that contains their implementation even if their
interface is declared in a module.
The private @implementation of an @interface may have additional
hidden ivars so we should not defer to the public version of the
type that is found in the module.
If the untied clause is present on a task construct, any thread in the
team can resume the task region after a suspension. Patch adds proper
codegen for untied tasks.
This fixes an issue in response files where "\r\n" was being interpreted
as two EOL markers (i.e. we consumed the '\r' as terminating the
previous token, and then parsed the '\n' as a significant EOL). This
breaks response files where joined arguments get split across multiple
lines (like "-x\r\nc"). I also fixed an accidental issue in the
response-file.c test, where the response file is appended to, instead of
being overwritten.
Adam Nemet [Tue, 19 Apr 2016 22:17:45 +0000 (22:17 +0000)]
[Parse] Use StringSwitch to improve readability. NFC
A subsequent patch will propose a "distribute" loop hint. Similarly to
unroll, this does not have a "assume_safety" argument either so this
condition will get more complex.
The intrinsic is now called llvm.thread.pointer, not
llvm.aarch64.thread.pointer. Also, the code handling it in CGBuiltin.cpp
is dead - it's already covered by GCCBuiltin. Remove it.
Adrian Prantl [Tue, 19 Apr 2016 20:31:19 +0000 (20:31 +0000)]
cc1as: Don't crash when CIE is requested and no DWARF version is specified.
This patch changes the default DWARF version for cc1as from invalid 0 to 2,
which should be the lowest common denominator on all platforms.
Ahmed Bougacha [Tue, 19 Apr 2016 19:44:45 +0000 (19:44 +0000)]
[ARM NEON] Define vfms_f32 on ARM, and all vfms using vfma.
r259537 added vfma/vfms to armv7, but the builtin was only lowered
on the AArch64 side. Instead of supporting it on ARM, get rid of it.
The vfms builtin lowered to:
%nb = fsub float -0.0, %b
%r = @llvm.fma.f32(%a, %nb, %c)
Instead, define the operation in terms of vfma, and swap the
multiplicands. It now lowers to:
%na = fsub float -0.0, %a
%r = @llvm.fma.f32(%na, %b, %c)
This matches the instruction more closely, and lets current LLVM
generate the "natural" operand ordering:
fmls.2s v0, v1, v2
instead of the crooked (but equivalent):
fmls.2s v0, v2, v1
Except for theses changes, assembly is identical.
LLVM accepts both commutations, and the LLVM tests in:
test/CodeGen/AArch64/arm64-fmadd.ll
test/CodeGen/AArch64/fp-dp3.ll
test/CodeGen/AArch64/neon-fma.ll
test/CodeGen/ARM/fusedMAC.ll
already check either the new one only, or both.
Martin Probst [Tue, 19 Apr 2016 18:18:59 +0000 (18:18 +0000)]
clang-format: [JS] support `interface` as a free standing identifier.
Summary:
`interface` can be used as a fee standing identifier in JavaScript/TypeScript.
This change uses the heuristic of whether it's followed by another identifier
as an indication.
Ahmed Bougacha [Tue, 19 Apr 2016 17:54:29 +0000 (17:54 +0000)]
[CodeGen] Widen non-power-of-2 vector HFA base types.
Currently, for the ppc64--gnu and aarch64 ABIs, we recognize:
typedef __attribute__((__ext_vector_type__(3))) float v3f32;
typedef __attribute__((__ext_vector_type__(16))) char v16i8;
struct HFA {
v3f32 a;
v16i8 b;
};
as an HFA. Since the first type encountered is used as the base type,
we pass the HFA as:
[2 x <3 x float>]
Which leads to incorrect IR (relying on padding values) when the
second field is used.
Instead, explicitly widen the vector (after size rounding) in
isHomogeneousAggregate.
If the untied clause is present on a task construct, any thread in the
team can resume the task region after a suspension. Patch adds proper
codegen for untied tasks.
Samuel Benzaquen [Tue, 19 Apr 2016 15:52:56 +0000 (15:52 +0000)]
[ASTMatchers] Do not try to memoize nodes we can't compare.
Summary:
Prevent hasAncestor from comparing nodes that are not supported.
hasDescendant was fixed some time ago to avoid this problem.
I'm applying the same fix to hasAncestor: if any object in the Builder map is
not comparable, skip the cache.
Martin Probst [Tue, 19 Apr 2016 14:55:37 +0000 (14:55 +0000)]
clang-format: [JS] simplify import/export.
Summary:
Change `import` and `export` parsing to special case the renaming
syntax (`import x, {y as bar} ...`, `export {x}`) and otherwise just
parse a regular structural element.
This simplifies the code a bit and should be more correct - it's easier
to recognise the specific import syntax than to recognise arbitrary
expressions and declarations.
If the untied clause is present on a task construct, any thread in the team can resume the task region after a suspension. Patch adds proper codegen for untied tasks.
Warn if function or variable cannot be implicitly instantiated
With this patch compiler emits warning if it tries to make implicit instantiation
of a template but cannot find the template definition. The warning can be suppressed
by explicit instantiation declaration or by command line options
-Wundefined-var-template and -Wundefined-func-template. The implementation follows
the discussion of http://reviews.llvm.org/D12326.
Richard Smith [Mon, 18 Apr 2016 23:12:59 +0000 (23:12 +0000)]
Mark -Xclang as being a compilation-only option so that the clang driver
doesn't warn if it's passed to a link action. This matches the behavior for
most other compilation-only options (including things like -f flags), and is
necessary to suppress warnings on systems like cmake that always pass all
compile flags to the linker.
The builtins already exist in LLVM, but are not exposed to the C/C++
programmers. This patch adds all the information about the builtins
needed for clang, as well as a test for all available intrinsics.
When an explicit specialization of a function template, a member function template or a member function of a class template is declared, clang first implicitly instantiates the declaration of a specialization from the templated-entity being explicitly specialized (since their signatures must be the same) and then links the explicit specialization being declared as a redeclaration of the aforementioned specialization.
The problem was that when clang 'implicitly instantiates' the initial specialization, it marks the corresponding FunctionDecl as deleted if the corresponding templated-entity was deleted, rather than waiting to see whether the explicit specialization being declared provides a non-deleted body. (The eager marking of delete has advantages during overload resolution I suppose, where we don't have to try and instantiate a definition of the function to see if it is deleted).
The present fix entails recognizing that when clang knows that an explicit specialization is being declared (for whichever templated-entity), the prior implicit instantiation should not inherit the 'deleted' status, and so we reset it to false.
I suppose an alternative fix (amongst others) could consider creating a new context (ExplicitSpecializationDeclarationSubstitution or some such) that is checked during template-argument-deduction and final substitution, and avoid inheriting the deleted status during declaration substitution. But while conceptually cleaner, that would be a slightly more involved change (as could be some of the other alternatives: such as avoid tagging implicit specializations as deleted, and check their primary templates for the deleted status where needed), and so I chose a different path. Hopefully it'll prove to not be a bad choice.
DebugInfo: Make DICompositeTypes distinct most of the time
Since elements of most kinds of DICompositeType have back references,
most are involved in uniquing cycles. Except via the ODR 'identifier:'
field, which doesn't care about the storage type (see r266549),
they have no hope of being uniqued.
Distinct nodes are far more efficient, so use them for most kinds of
DICompositeType definitions (i.e., when DIType::isForwardDecl is false).
The exceptions:
- DW_TAG_array_type, since their elements never have back-references
and they never have ODR 'identifier:' fields;
- DW_TAG_enumeration_type when there is no ODR 'identifier:' field,
since their elements usually don't have back-references.
This breaks the last major uniquing cycle I'm aware of in the debug info
graph. The impact won't be enormous for C++ because references to
ODR-uniqued nodes still use string-based DITypeRefs; but this should
prevent a regression in C++ when we drop the string-based references.
This wouldn't have been reasonable until r266549, when composite types
stopped relying on being uniqued by structural equivalence to prevent
blow-ups at LTO time.
Richard Smith [Sat, 16 Apr 2016 00:46:26 +0000 (00:46 +0000)]
[modules] Don't expose *intrin.h headers that cannot be included standalone as
separate modules. These cause build breakage with -fmodules-local-submodule-visibility.
[CUDA] Raise an error if the CUDA install can't be found.
Summary:
Without this change, we silently proceed on without including
__clang_cuda_runtime_wrapper.h. This leads to very strange behavior --
you say you're compiling CUDA code, but e.g. __device__ is not defined!