George Karpenkov [Fri, 21 Sep 2018 20:36:01 +0000 (20:36 +0000)]
[analyzer] Fix bug in isInevitablySinking
If the non-sink report is generated at the exit node, it will be
suppressed by the current functionality in isInevitablySinking, as it
only checks the successors of the block, but not the block itself.
Add necessary support for storing code-model to module IR.
Currently the code-model does not get saved in the module IR, so if a
code model is specified when compiling with LTO, it gets lost and is
not propagated properly to LTO. This patch does what is necessary in
the front end to pass the code-model to the module, so that the back
end can store it in the Module .
Bruno Ricci [Fri, 21 Sep 2018 14:03:32 +0000 (14:03 +0000)]
[AST][NFC] DeclarationName.h : add missing parentheses to silence warnings
Some bots are complaining about missing parentheses in assertions added in
r342729: [AST] Various optimizations + refactoring in DeclarationName(Table)
NFC: deduplicate isRepeatedBytePattern from clang to LLVM's isBytewiseValue
Summary:
This code was in CGDecl.cpp and really belongs in LLVM. It happened to have isBytewiseValue which served a very similar purpose but wasn't as powerful as clang's version. Remove the clang version, and augment isBytewiseValue to be as powerful so that clang does the same thing it used to.
Bruno Ricci [Fri, 21 Sep 2018 13:11:39 +0000 (13:11 +0000)]
[AST][NFC] Remove a superfluous enum in ObjCMethodDeclBitfields added in r338641
I originally added this enum to avoid including Basic/IdentifierTable.h
in AST/DeclBase.h. However I did not realise it is already included
transitively by AST/DeclarationName.h. Therefore remove this enum and
explicitly include Basic/IdentifierTable.h
Bruno Ricci [Fri, 21 Sep 2018 12:53:22 +0000 (12:53 +0000)]
[AST] Various optimizations + refactoring in DeclarationName(Table)
Introduce the following optimizations in DeclarationName(Table):
1. Store common kinds inline in DeclarationName instead of
DeclarationNameExtra. Currently the kind of C++ constructor, destructor,
conversion function and overloaded operator names is stored in
DeclarationNameExtra. Instead store it inline in DeclarationName.
To do this align IdentifierInfo, CXXSpecialName, DeclarationNameExtra
and CXXOperatorIdName to 8 bytes so that we can use the lower 3 bits of
DeclarationName::Ptr. This is already the case on 64 bits archs anyway.
This also allow us to remove DeclarationNameExtra from CXXSpecialName
and CXXOperatorIdName, which shave off a pointer from CXXSpecialName.
2. Synchronize the enumerations DeclarationName::NameKind,
DeclarationName::StoredNameKind and Selector::IdentifierInfoFlag.
This makes DeclarationName::getNameKind much more efficient since we can
replace the switch table by a single comparison and an addition.
3. Put the overloaded operator names inline in DeclarationNameTable to remove
an indirection. This increase the size of DeclarationNameTable a little
bit but this is not important since it is only used in ASTContext, and
never copied nor moved from. This also get rid of the last dynamic
allocation in DeclarationNameTable.
Altogether these optimizations cut the run time of parsing all of Boost by
about 0.8%. While we are at it, do the following NFC modifications:
1. Put the internal classes CXXSpecialName, CXXDeductionGuideNameExtra,
CXXOperatorIdName, CXXLiteralOperatorIdName and DeclarationNameExtra
in a namespace detail since these classes are only meant to be used by
DeclarationName and DeclarationNameTable. Make this more explicit by making
the members of these classes private and friending DeclarationName(Table).
2. Make DeclarationName::getFETokenInfo a non-template since every users are
using it to get a void *. It was supposed to be used with a type to avoid
a subsequent static_cast.
3. Change the internal functions DeclarationName::getAs* to castAs* since when
we use them we already know the correct kind. This has no external impact
since all of these are private.
[CodeGen] Add to emitted DebugLoc information about coverage when it's required
Summary:
Some lines have a hit counter where they should not have one.
Cleanup stuff is located to the last line of the body which is most of the time a '}'.
And Exception stuff is added at the beginning of a function and at the end (represented by '{' and '}').
So in such cases, the DebugLoc used in GCOVProfiling.cpp must be marked as not covered.
This patch is a followup of https://reviews.llvm.org/D49915.
Tests in projects/compiler_rt are fixed by: https://reviews.llvm.org/D49917
Removing on behalf of Jorge Moya. This test is broken on
Windows due to it actually being able to resolve the path. There
is an actual Windows-specific bug somewhere, but we already have
sufficient test coverage of this with a different test, so removing
this was the approach suggested by Jorge.
Eric Christopher [Thu, 20 Sep 2018 17:21:56 +0000 (17:21 +0000)]
r342177 introduced a hint in cases where an #included file is not found. It tries to find a suggestion by removing leading or trailing non-alphanumeric characters and checking if a matching file exists, then it reports an error like:
include-likely-typo.c:3:10: error: '<empty_file_to_include.h>' file not found, did you mean 'empty_file_to_include.h'?
^~~~~~~~~~~~~~~~~~~~~~~~~~~
"empty_file_to_include.h"
1 error generated.
However, if a hint is not found, the error message will show only the trimmed name we use to look for a hint, so:
will result in:
include-leading-nonalpha-no-suggest.c:3:10: fatal error: 'non_existing_file_to_include.h' file not found
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1 error generated.
where the name reported after "fatal error:" doesn't match what the user wrote.
unsigned long long builtin_unpack_vector_int128 (vector int128_t, int);
vector int128_t builtin_pack_vector_int128 (unsigned long long, unsigned long long);
Thread Safety Analysis: warnings for attributes without arguments
Summary:
When thread safety annotations are used without capability arguments,
they are assumed to apply to `this` instead. So we warn when either
`this` doesn't exist, or the class is not a capability type.
This is based on earlier work by Josh Gao that was committed in r310403,
but reverted in r310698 because it didn't properly work in template
classes. See also D36237.
The solution is not to go via the QualType of `this`, which is then a
template type, hence the attributes are not known because it could be
specialized. Instead we look directly at the class in which we are
contained.
Additionally I grouped two of the warnings together. There are two
issues here: the existence of `this`, which requires us to be a
non-static member function, and the appropriate annotation on the class
we are contained in. So we don't distinguish between not being in a
class and being static, because in both cases we don't have `this`.
I had locally changed the test to add an explicit triple to figure out the issue
with the SCEI buildbots, and that hid the error. This now works with and
without the explicit triple.
Sema: handle `wint_t` more carefully for printf checking
In the case that `win_t` is an `unsigned short` (e.g. on Windows), we would
previously incorrectly diagnose the conversion because we would immediately
promote the argument type from `wint_t` (aka `unsigned short`) to `int` before
checking if the type matched. This should repair the Windows hosted bots.
Windows uses `unsigned short` for `wint_t`. Correct the type definition as
vended by the compiler. This type is defined in corecrt.h and is
unconditionally typedef'ed. cl does not have an equivalent to `__WINT_TYPE__`
which is why this was never detected.
Thread safety analysis: Fix crash for function pointers
For function pointers, the FunctionDecl of the callee is unknown, so
getDirectCallee will return nullptr. We have to catch that case to avoid
crashing. We assume there is no attribute then.
Add a callback for `__has_include` and use it for dependency scanning.
This adds a preprocessor callback for the `__has_include` and
`__has_include_next` directives.
Successful checking for the presence of a header should add it to the list of
header dependencies so this overrides the callback in the dependency scanner.
[MS] Defer dllexport inline friend functions like other inline methods
This special case was added in r264841, but the code breaks our
invariants by calling EmitTopLevelDecl without first creating a
HandlingTopLevelDeclRAII scope.
This fixes the PCH crash in https://crbug.com/884427. I was never able
to make a satisfactory reduction, unfortunately. I'm not very worried
about this regressing since this change makes the code simpler while
passing the existing test that shows we do emit dllexported friend
function definitions. Now we just defer their emission until the tag is
fully complete, which is generally good.
Move the floating point argument handling into the RenderFloatingPointOptions
helper. This relocation just puts the floating point related options into a
single location.
Stephen Hines [Tue, 18 Sep 2018 18:34:33 +0000 (18:34 +0000)]
Fix logic around determining use of frame pointer with -pg.
Summary:
As part of r342165, I rewrote the logic to check whether
-fno-omit-frame-pointer was passed after a -fomit-frame-pointer
argument. This CL switches that logic to use the consolidated
shouldUseFramePointer() function. This fixes a potential issue where -pg
gets used with -fomit-frame-pointer on a platform that must always retain
frame pointers.
[Modules] Add platform and environment features to requires clause
Allows module map writers to add build requirements based on
platform/os. This helps when target features and language dialects
aren't enough to conditionalize building a module, among other things,
it allow module maps for different platforms to live in the same file.
Sam McCall [Tue, 18 Sep 2018 08:40:41 +0000 (08:40 +0000)]
[CodeComplete] Add completions for filenames in #include directives.
Summary:
The dir component ("somedir" in #include <somedir/fo...>) is considered fixed.
We append "foo" to each directory on the include path, and then list its files.
Completions are of the forms:
#include <somedir/fo^
foo.h>
fox/
The filter is set to the filename part ("fo"), so fuzzy matching can be
applied to the filename only.
No fancy scoring/priorities are set, and no information is added to
CodeCompleteResult to make smart scoring possible. Could be in future.
Thread safety analysis: Run more tests with capability attributes [NFC]
Summary:
We run the tests for -Wthread-safety-{negative,verbose} with the new
attributes as well as the old ones. Also put the macros in a header so
that we don't have to copy them all around.
The warn-thread-safety-parsing.cpp test checks for warnings depending on
the actual attribute name, so it can't undergo the same treatment.
George Karpenkov [Mon, 17 Sep 2018 20:46:53 +0000 (20:46 +0000)]
[analyzer] ExplodedGraph printing fixes
Fixes a number of issues:
- Global variables are not used for communication
- Trait should be defined on a graph, not on a node
- Defining the trait on a graph allows us to use a correct allocator,
no longer crashing while printing trimmed graphs
Shuai Wang [Mon, 17 Sep 2018 20:10:56 +0000 (20:10 +0000)]
[analyzer] Treat std::{move,forward} as casts in ExprMutationAnalyzer.
Summary:
This is a follow up of D52008 and should make the analyzer being able to handle perfect forwardings in real world cases where forwardings are done through multiple layers of function calls with `std::forward`.
Summary:
The init expression of a VarDecl is overwritten in the "To" context if we
import a VarDecl without an init expression (and with a definition). Please
refer to the added tests, especially InitAndDefinitionAreInDifferentTUs. This
patch fixes the malfunction by importing the whole Decl chain similarly as we
did that in case of FunctionDecls. We handle the init expression similarly to
a definition, alas only one init expression will be in the merged ast.
Summary:
Merged the recently added `err_attribute_argument_negative` diagnostic
with existing `err_attribute_requires_positive_integer` diagnostic:
the former allows only strictly positive integer, while the latter
also allows zero.
scan-build: Using '/usr/lib/llvm-7/bin/clang' for static analysis
scan-build: Emitting reports for this run to '/tmp/scan-build-2018-09-16-214531-8410-1'.
foo.c:3:3: warning: Value stored to 'x' is never read
x = 1; // warn
^ ~
1 warning generated.
scan-build: File '/tmp/foo.c' deleted: part of an ignored directory.
scan-build: 0 bugs found.
```
Kelvin Li [Sat, 15 Sep 2018 13:54:15 +0000 (13:54 +0000)]
[OPENMP] Move OMPClauseReader/Writer classes to ASTReader/Writer (NFC)
Move declarations for OMPClauseReader, OMPClauseWriter to ASTReader.h
and ASTWriter.h and move implementation to ASTReader.cpp and
ASTWriter.cpp. This change helps generalize the serialization of
OpenMP clauses and will be used in the future implementation of new
OpenMP directives (e.g. requires).
Shuai Wang [Fri, 14 Sep 2018 20:07:18 +0000 (20:07 +0000)]
[analyzer] Handle forwarding reference better in ExprMutationAnalyzer.
Summary:
We used to treat an `Expr` mutated whenever it's passed as non-const
reference argument to a function. This results in false positives in
cases like this:
```
int x;
std::vector<int> v;
v.emplace_back(x); // `x` is passed as non-const reference to `emplace_back`
```
In theory the false positives can be suppressed with
`v.emplace_back(std::as_const(x))` but that's considered overly verbose,
inconsistent with existing code and spammy as diags.
This diff handles such cases by following into the function definition
and see whether the argument is mutated inside.
Hans Wennborg [Fri, 14 Sep 2018 15:18:30 +0000 (15:18 +0000)]
[clang-cl] Fix PR38934: failing to dllexport class template member w/ explicit instantiation and PCH
The code in ASTContext::DeclMustBeEmitted was supposed to handle this,
but didn't take into account that synthesized members such as operator=
might not get marked as template specializations, because they're
synthesized on the instantiation directly when handling the class-level
dllexport attribute.
Louis Dionne [Fri, 14 Sep 2018 14:07:16 +0000 (14:07 +0000)]
[clang] Make sure attributes on member classes are applied properly
Summary:
Attributes on member classes of class templates and member class templates
of class templates are not currently instantiated. This was discovered by
Richard Smith here:
This commit makes sure that attributes are instantiated properly. This
commit does not fix the broken behavior for member partial and explicit
specializations of class templates.
Sam McCall [Fri, 14 Sep 2018 12:47:38 +0000 (12:47 +0000)]
[VFS] vfs::directory_iterator yields path and file type instead of full Status
Summary:
Most callers I can find are using only `getName()`. Type is used by the
recursive iterator.
Now we don't have to call stat() on every listed file (on most platforms).
Exceptions are e.g. Solaris where readdir() doesn't include type information.
On those platforms we'll still stat() - see D51918.
The result is significantly faster (stat() can be slow).
My motivation: this may allow us to improve clang IO on large TUs with long
include search paths. Caching readdir() results may allow us to skip many stat()
and open() operations on nonexistent files.
[Driver] Fix missing MultiArch include dir on powerpcspe
On powerpc-linux-gnuspe, the header files are located in their
own include directory named /usr/lib/powerpc-linux-gnuspe,
so add this directory to PPCMultiarchIncludeDirs.
Sam McCall [Fri, 14 Sep 2018 12:24:09 +0000 (12:24 +0000)]
[Tooling] JSONCompilationDatabasePlugin infers compile commands for missing files
Summary:
See the existing InterpolatingCompilationDatabase for details on how this works.
We've been using this in clangd for a while, the heuristics seem to work well.
[analyzer][UninitializedObjectChecker] New flag to ignore records based on it's fields
Based on a suggestion from @george.karpenkov.
In some cases, structs are used as unions with a help of a tag/kind field.
This patch adds a new string flag (a pattern), that is matched against the
fields of a record, and should a match be found, the entire record is ignored.
For more info refer to http://lists.llvm.org/pipermail/cfe-dev/2018-August/058906.html
and to the responses to that, especially http://lists.llvm.org/pipermail/cfe-dev/2018-August/059215.html.
Since I plan to add a number of new flags, it made sense to encapsulate
them in a new struct, in order not to pollute FindUninitializedFields's
constructor with new boolean options with super long names.
This revision practically reverts D50508, since FindUninitializedFields
now accesses the pedantic flag anyways.
Some of the comments are incorrect, imprecise, or simply nonexistent.
Since I have a better grasp on how the analyzer works, it makes sense
to update most of them in a single swoop.
I tried not to flood the code with comments too much, this amount
feels just right to me.
projects / clang / log