[X86] Use native shuffle vector for the perm2f128 intrinsics
This patch replaces the perm2f128 intrinsics with native shuffle vectors.
This uses a pretty simple approach to allocate source 0 to the lower half input and source 1 to the upper half input. Then its just a matter of filling in the indices to use either the lower or upper half of that specific source. This can result in the same source being used by both operands. InstCombine or SelectionDAGBuilder should be able to clean that up.
Resubmit "[lit] Force site configs to run before source-tree configs"
This is a resubmission of r313270. It broke standalone builds of
compiler-rt because we were not correctly generating the llvm-lit
script in the standalone build directory.
The fixes incorporated here attempt to find llvm/utils/llvm-lit
from the source tree returned by llvm-config. If present, it
will generate llvm-lit into the output directory. Regardless,
the user can specify -DLLVM_EXTERNAL_LIT to point to a specific
lit.py on their file system. This supports the use case of
someone installing lit via a package manager. If it cannot find
a source tree, and -DLLVM_EXTERNAL_LIT is either unspecified or
invalid, then we print a warning that tests will not be able
to run.
Summary:
Remove `__builtin_wasm_rethrow` builtin. I thought it was required to implement
`__cxa_rethrow` function in libcxxabi, but it turned out it will be using
`__builtin_wasm_throw` instead.
[X86] Disable _mm512_maskz_set1_epi64 intrinsic on 32-bit targets to prevent a backend isel failure.
The __builtin_ia32_pbroadcastq512_mem_mask we were previously trying to use in 32-bit mode is not implemented in the x86 backend and causes isel to fail in release builds. In debug builds it fails even earlier during legalization with an llvm_unreachable.
While there add the missing test case for this intrinsic for this for 64-bit mode.
This fixes PR34631. D37668 should be able to recover this for 32-bit mode soon. But I wanted to fix the crash ahead of that.
[Sema] Error out early for tags defined inside an enumeration.
This fixes PR28903 by avoiding access check for inner enum constant. We
are performing access check because one enum constant references another
and because enum is defined in CXXRecordDecl. But access check doesn't
work because FindDeclaringClass doesn't expect more than one EnumDecl
and because inner enum has access AS_none due to not being an immediate
child of a record.
The change detects an enum is defined in wrong place and allows to skip
parsing its body. Access check is skipped together with body parsing.
There was no crash in C, added test case to cover the new error.
In CUDA-9 some of device-side math functions that we need are conditionally
defined within '#if _GLIBCXX_MATH_H'. We need to temporarily undo the guard
around inclusion of math_functions.hpp.
[clang-format] New flag - BraceWrapping.AfterExternBlock
Summary:
Bug: https://bugs.llvm.org/show_bug.cgi?id=34016 - **"extern C part"**
**Problem:**
Due to the lack of "brace wrapping extern" flag, clang format does parse the block after **extern** keyword moving the opening bracket to the header line always!
**Patch description:**
A new style added, new configuration flag - **BraceWrapping.AfterExternBlock** that allows us to decide whether we want a break before brace or not.
Revert "[lit] Force site configs to run before source-tree configs"
This patch is still breaking several multi-stage compiler-rt bots.
I already know what the fix is, but I want to get the bots green
for now and then try re-applying in the morning.
[Sema] Correct typos in LHS, RHS before building a binop expression.
Specifically, typo correction should be done before dispatching between
different kinds of binary operations like pseudo-object assignment,
overloaded binary operation, etc.
Without this change we hit an assertion
Assertion failed: (!LHSExpr->hasPlaceholderType(BuiltinType::PseudoObject)), function CheckAssignmentOperands
when in Objective-C we reference a property without `self` and there are
2 equally good typo correction candidates: ivar and a class name. In
this case LHS expression in `BuildBinOp` is
CXXDependentScopeMemberExpr
`-TypoExpr
and instead of handling Obj-C property assignment as pseudo-object
assignment, we call `CreateBuiltinBinOp` which corrects typo to
ObjCPropertyRefExpr '<pseudo-object type>'
but cannot handle pseudo-objects and asserts about it (indirectly,
through `CheckAssignmentOperands`).
Douglas Gregor [Thu, 14 Sep 2017 23:38:44 +0000 (23:38 +0000)]
[Module map] Introduce a private module re-export directive.
Introduce a new "export_as" directive for top-level modules, which
indicates that the current module is a "private" module whose symbols
will eventually be exported through the named "public" module. This is
in support of a common pattern in the Darwin ecosystem where a single
public framework is constructed of several private frameworks, with
(currently) header duplication and some support from the linker.
The recent lit refactor changed the location of the lit script
run by check targets from <source>/utils/lit/lit.py to
<bin>/llvm-lit.py. In some 2-stage build scenarios, the location
of <bin> was not properly passed through to the second stage,
and it was looking for /llvm-lit.py instead, causing failures.
Fix suggested by Mike Edwards and Chris Bieneman @apple
[lit] Force site configs to be run before source-tree configs
This patch simplifies LLVM's lit infrastructure by enforcing an ordering
that a site config is always run before a source-tree config.
A significant amount of the complexity from lit config files arises from
the fact that inside of a source-tree config file, we don't yet know if
the site config has been run. However it is *always* required to run
a site config first, because it passes various variables down through
CMake that the main config depends on. As a result, every config
file has to do a bunch of magic to try to reverse-engineer the location
of the site config file if they detect (heuristically) that the site
config file has not yet been run.
This patch solves the problem by emitting a mapping from source tree
config file to binary tree site config file in llvm-lit.py. Then, during
discovery when we find a config file, we check to see if we have a
target mapping for it, and if so we use that instead.
This mechanism is generic enough that it does not affect external users
of lit. They will just not have a config mapping defined, and everything
will work as normal.
On the other hand, for us it allows us to make many simplifications:
* We are guaranteed that a site config will be executed first
* Inside of a main config, we no longer have to assume that attributes
might not be present and use getattr everywhere.
* We no longer have to pass parameters such as --param llvm_site_config=<path>
on the command line.
* It is future-proof, meaning you don't have to edit llvm-lit.in to add
support for new projects.
* All of the duplicated logic of trying various fallback mechanisms of
finding a site config from the main config are now gone.
One potentially noteworthy thing that was required to implement this
change is that whereas the ninja check targets previously used the first
method to spawn lit, they now use the second. In particular, you can no
longer run lit.py against the source tree while specifying the various
`foo_site_config=<path>` parameters. Instead, you need to run
llvm-lit.py.
Alex Lorenz [Thu, 14 Sep 2017 10:06:52 +0000 (10:06 +0000)]
[refactor] add clang-refactor tool with initial testing support and
local-rename action
This commit introduces the clang-refactor tool alongside the local-rename action
which uses the existing renaming engine used by clang-rename. The tool
doesn't actually perform the source transformations yet, it just provides
testing support. This commit also moves only one test from clang-rename over to
test/Refactor. I will continue to move the other tests throughout
development of clang-refactor.
The following options are supported by clang-refactor:
-v: use verbose output
-selection: The source range that corresponds to the portion of the source
that's selected (currently only special command test:<file> is supported).
Please note that a follow-up commit will migrate clang-refactor to
libTooling's common option parser, so clang-refactor will be able to use
the common interface with compilation database and options like -p, -extra-arg,
etc.
The testing support provided by clang-refactor is described below:
When -selection=test:<file> is given, clang-refactor will parse the selection
commands from that file. The selection commands are grouped and the specified
refactoring action invoked by the tool. Each command in a group is expected to
produce an identical result. The precise syntax for the selection commands is
described in a comment in TestSupport.h.
Adrian McCarthy [Wed, 13 Sep 2017 20:53:55 +0000 (20:53 +0000)]
Mark static member functions as static in CodeViewDebug
Summary:
To improve CodeView quality for static member functions, we need to make the
static explicit. In addition to a small change in LLVM's CodeViewDebug to
return the appropriate MethodKind, this requires a small change in Clang to
note the staticness in the debug info metadata.
SplitEmptyFunction should be true in the Mozilla coding style
Summary:
As defined here: https://developer.mozilla.org/en-US/docs/Mozilla/Developer_guide/Coding_Style#Classes
See for the downstream bug report: https://bugzilla.mozilla.org/show_bug.cgi?id=1399359
Introduce the 'TLS Kind' property of variable declarations through
libclang. Additionally, provide a Python accessor for it, and test that
functionality.
This reverts rL313102 because it still fails some build bot tests.
On many linux bots it fails with the following error.
error: invalid linker name in argument '-fuse-ld=lld'
and on some windows bots also because there is no ld.lld.exe
lld-link.exe: warning: ignoring unknown argument: -fuse-ld=lld
In rL289668 the ability to specify the default linker at compile time
was added but because the MinGW driver used custom detection we could
not take advantage of this new CMAKE flag CLANG_DEFAULT_LINKER.
This is a re-apply of rL313082 which was reverted in rL313088
due to failing buildbot tests.
[ubsan] Function Sanitizer: Don't require writable text segments
This change will make it possible to use -fsanitize=function on Darwin and
possibly on other platforms. It fixes an issue with the way RTTI is stored into
function prologue data.
On Darwin, addresses stored in prologue data can't require run-time fixups and
must be PC-relative. Run-time fixups are undesirable because they necessitate
writable text segments, which can lead to security issues. And absolute
addresses are undesirable because they break PIE mode.
The fix is to create a private global which points to the RTTI, and then to
encode a PC-relative reference to the global into prologue data.
[Driver] Disable uwtable by default in -ffreestanding mode
We make the same decision when compiling the kernel or kexts -- we
should do this in -ffreestanding mode as well to avoid size regressions
in a potentially large set of firmware projects.
It's still possible to get uwtable information in -ffreestanding mode by
compiling with -funwind-tables (I expect this to be a rare case: I
certainly haven't seen any projects like that).
Context: -munwind-tables was enabled by default for some arm targets in
r310006.
In rL289668 the ability to specify the default linker at compile time
was added but because the MinGW driver used custom detection we could
not take advantage of this new CMAKE flag CLANG_DEFAULT_LINKER.
[Driver] Darwin: Link in the profile runtime archive first
While building a project with code coverage enabled, we can link in
dependencies which export a weak definition of __llvm_profile_filename.
After r306710, linking in the profiling runtime could pull in a weak
definition of this symbol from a dependency, instead of from within the
runtime's archive.
This inconsistency causes issues during API verification, and is also a
practical problem (the symbol would go missing were the dependent dylib
to be switched out). Introduce a LinkFirst runtime link option to make
sure we always search the profiling runtime for this symbol first.
Hans Wennborg [Tue, 12 Sep 2017 18:38:34 +0000 (18:38 +0000)]
clang-format-vs: Update the VSIX ID.
We're moving the extension to a new account on the VS Marketplace, and
apparently it's not possible to re-upload an extension with an existing
ID on a new account.
Use the VFS from the CompilerInvocation by default
Summary:
The CompilerInstance should create its default VFS from its CompilerInvocation. Right now the
user has to manually create the VFS before creating the FileManager even though
`-ivfsoverlay file.yaml` was passed via the CompilerInvocation (which is exactly how we worked
around this issue in `FrontendAction.cpp` so far).
This patch uses the invocation's VFS by default and also tests this behavior now from the
point of view of a program that uses the clang API.
Fix recording preamble's conditional stack in skipped PP branches.
Summary:
This fixes PR34547.
`Lexer::LexEndOfFile` handles recording of ConditionalStack for
preamble and reporting errors about unmatched conditionalal PP
directives.
However, SkipExcludedConditionalBlock contianed duplicated logic for
reporting errors and clearing ConditionalStack, but not for preamble
recording.
This fix removes error reporting logic from
`SkipExcludedConditionalBlock`, unmatched PP conditionals are now
reported inside `Lexer::LexEndOfFile`.
Petr Hosek [Tue, 12 Sep 2017 01:28:37 +0000 (01:28 +0000)]
[Driver] Fuchsia targets default to -fasynchronous-unwind-tables
This regressed for x86-64 in r307856 because it's no longer inherited
from Generic_GCC. We'd never noticed that it was missing other
targets (i.e. aarch64), but Fuchsia is uniform across all machines.
Bob Haarman [Mon, 11 Sep 2017 22:11:57 +0000 (22:11 +0000)]
[codeview] omit debug locations for nested exprs unless column info enabled
Summary:
Microsoft Visual Studio expects debug locations to correspond to
statements. We used to emit locations for expressions nested inside statements.
This would confuse the debugger, causing it to stop multiple times on the
same line and breaking the "step into specific" feature. This change inhibits
the emission of debug locations for nested expressions when emitting CodeView
debug information, unless column information is enabled.
[Lexer] Report more precise skipped regions (PR34166)
This patch teaches the preprocessor to report more precise source ranges for
code that is skipped due to conditional directives.
The new behavior includes the '#' from the opening directive and the full text
of the line containing the closing directive in the skipped area. This matches
up clang's behavior (we don't IRGen the code between the closing "endif" and
the end of a line).
This also affects the code coverage implementation. See llvm.org/PR34166 (this
also happens to be rdar://problem/23224058).
The old behavior (report the end of the skipped range as the end
location of the 'endif' token) is preserved for indexing clients.
Summary:
The use case is that renaming multiple symbols in a large enough codebase is
much faster if all of these can be done with a single invocation, but
there will be multiple translation units where one or more symbols are
not found.
Old behavior was to exit with an error (default) or exit without
reporting an error (-force). New behavior is that -force results in a
best-effort rename: rename symbols which are found and just ignore the
rest.
The existing help for -force sort of already implies this behavior.
This primarily impacts the Windows MSVC and Windows itanium
environments. Windows MSVC does not use `__cxa_atexit` and Itanium
follows suit. Simplify the logic for the default value calculation and
blanket the Windows environments to default to off for use of
`__cxa_atexit`.
Martin Probst [Mon, 11 Sep 2017 15:22:52 +0000 (15:22 +0000)]
clang-format: [JS] wrap and indent `goog.setTestOnly` calls.
Summary:
While `goog.setTestOnly` usually appears in the imports section of a file, it is
not actually an import, and also usually doesn't take long parameters (nor
namespaces as a parameter, it's a description/message that should be wrapped).
This fixes a regression where a `goog.setTestOnly` call nested in a function was
not wrapped.
[PCH] Allow VFS to be used for tests that generate PCH files
When using a virtual file-system (VFS) and a preamble file (PCH) is generated,
it is generated on-disk in the real file-system instead of in the VFS (which
makes sense, since the VFS is read-only). However, when subsequently reading
the generated PCH, the frontend passes through the VFS it has been given --
resulting in an error and a failed parse (since the VFS doesn't contain the
PCH; the real filesystem does).
This patch fixes that by detecting when a VFS is being used for a parse that
needs to work with a PCH file, and creating an overlay VFS that includes the
PCH file from the real file-system.
This allows tests to be written which make use of both PCH files and a VFS.
[AST] Make RecursiveASTVisitor visit TemplateDecls in source order
Summary:
This causes template arguments to be traversed before the templated
declaration, which is useful for clients that expect the nodes in
the same order as they are in the source code. Additionally, there
seems to be no good reason not to do so.
This was moved here from LexicallyOrderedRecursiveASTVisitor. The tests
still reside in LexicallyOrderedRecursiveASTVisitorTest.cpp under
VisitTemplateDecls.
Fixed bug: https://bugs.llvm.org/show_bug.cgi?id=34001
Clang-format bug resulting in a strange behavior of control statements short blocks. Different flags combinations do not guarantee expected result. Turned on option AllowShortBlocksOnASingleLine does not work as intended.
**Description of the problem:**
Cpp source file UnwrappedLineFormatter does not handle AllowShortBlocksOnASingleLine flag as it should. Putting a single-line control statement without any braces, clang-format works as expected (depending on AllowShortIfStatementOnASingleLine or AllowShortLoopsOnASingleLine value). Putting a single-line control statement in braces, we can observe strange and incorrect behavior.
Our short block is intercepted by tryFitMultipleLinesInOne function. The function returns a number of lines to be merged. Unfortunately, our control statement block is not covered properly. There are several if-return statements, but none of them handles our block. A block is identified by the line first token and by left and right braces. A function block works as expected, there is such an if-return statement doing proper job. A control statement block, from the other hand, falls into strange conditional construct, which depends on BraceWrapping.AfterFunction flag (with condition that the line’s last token is left brace, what is possible in our case) or goes even further. That should definitely not happen.
**Description of the patch:**
By adding three different if statements, we guarantee that our short control statement block, however it looks like (different brace wrapping flags may be turned on), is handled properly and does not fall into wrong conditional construct. Depending on appropriate options we return either 0 (when something disturbs our merging attempt) or let another function (tryMergeSimpleBlock) take the responsibility of returned result (number of merged lines). Nevertheless, one more correction is required in mentioned tryMergeSimpleBlock function. The function, previously, returned either 0 or 2. The problem was that this did not handle the case when our block had the left brace in a separate line, not the header one. After change, after adding condition, we return the result compatible with block’s structure. In case of left brace in the header’s line we do everything as before the patch. In case of left brace in a separate line we do the job similar to the one we do in case of a “non-header left brace” function short block. To be precise, we try to merge the block ignoring the header line. Then, if success, we increment our returned result.
**After fix:**
**CONFIG:**
```
AllowShortBlocksOnASingleLine: true
AllowShortIfStatementsOnASingleLine: true
BreakBeforeBraces: Custom
BraceWrapping: {
AfterClass: true, AfterControlStatement: true, AfterEnum: true, AfterFunction: true, AfterNamespace: false, AfterStruct: true, AfterUnion: true, BeforeCatch: true, BeforeElse: true
}
```
**BEFORE:**
```
if (statement) doSomething();
if (statement) { doSomething(); }
if (statement) {
doSomething();
}
if (statement)
{
doSomething();
}
if (statement)
doSomething();
if (statement) {
doSomething1();
doSomething2();
}
```
**AFTER:**
```
if (statement) doSomething();
if (statement) { doSomething(); }
if (statement) { doSomething(); }
if (statement) { doSomething(); }
if (statement) doSomething();
if (statement)
{
doSomething1();
doSomething2();
}
```
MinSeong Kim [Sat, 9 Sep 2017 14:18:53 +0000 (14:18 +0000)]
[Basic] Update CMakeLists.txt to handle repo
Summary:
The find_first_existing_file and find_first_existing_vc_file macros in
lib/Basic/CMakeLists.txt are removed. The macros are also defined in
{LLVM}/cmake/modules/AddLLVM.cmake for the same purpose. This change
serves the following 2 objectives:
1. To remove the redundant code in clang to use the same
macros in llvm,
2. The macros in AddLLVM.cmake can also handle repo for
displaying correct version information.
Richard Smith [Sat, 9 Sep 2017 01:14:04 +0000 (01:14 +0000)]
Fix ownership of the MemoryBuffer in a FrontendInputFile.
This fixes a possible crash on certain kinds of corrupted AST file, but
checking in an AST file corrupted in just the right way will be a maintenance
nightmare because the format changes frequently.
Richard Trieu [Sat, 9 Sep 2017 00:25:05 +0000 (00:25 +0000)]
Catch more cases with -Wenum-compare
Treat typedef enum as named enums instead of anonymous enums. Anonymous enums
are ignored by the warning, so previously, typedef enums were ignored as well.
When performing a NSFastEnumeration, the compiler synthesizes a call to
`countByEnumeratingWithState:objects:count:` where the `count` parameter
is of type `NSUInteger` and the return type is a `NSUInteger`. We would
previously always use a `UnsignedLongTy` for the `NSUInteger` type. On
32-bit targets, `long` is 32-bits which is the same as `unsigned int`.
Most 64-bit targets are LP64, where `long` is 64-bits. However, on
LLP64 targets, such as Windows, `long` is 32-bits. Introduce new
`getNSUIntegerType` and `getNSIntegerType` helpers to allow us to
determine the correct type for the `NSUInteger` type. Wire those
through into the generation of the message dispatch to the selector.
Wei Mi [Fri, 8 Sep 2017 21:58:18 +0000 (21:58 +0000)]
Reinstall the patch "Use EmitPointerWithAlignment to get alignment information of the pointer used in atomic expr".
This is to fix PR34347. EmitAtomicExpr now only uses alignment information from
Type, instead of Decl, so when the declaration of an atomic variable is marked
to have the alignment equal as its size, EmitAtomicExpr doesn't know about it and
will generate libcall instead of atomic op. The patch uses EmitPointerWithAlignment
to get the precise alignment information.
[Coverage] Precise region termination with deferred regions (reapply)
The current coverage implementation doesn't handle region termination
very precisely. Take for example an `if' statement with a `return':
void f() {
if (true) {
return; // The `if' body's region is terminated here.
}
// This line gets the same coverage as the `if' condition.
}
If the function `f' is called, the line containing the comment will be
marked as having executed once, which is not correct.
The solution here is to create a deferred region after terminating a
region. The deferred region is completed once the start location of the
next statement is known, and is then pushed onto the region stack.
In the cases where it's not possible to complete a deferred region, it
can safely be dropped.
Testing: lit test updates, a stage2 coverage-enabled build of clang
This is a reapplication but there are no changes from the original commit.
With D36813, the segment builder in llvm will be able to handle deferred
regions correctly.
Wei Mi [Fri, 8 Sep 2017 17:07:32 +0000 (17:07 +0000)]
Use EmitPointerWithAlignment to get alignment information of the pointer used in atomic expr.
This is to fix PR34347. EmitAtomicExpr now only uses alignment information from
Type, instead of Decl, so when the declaration of an atomic variable is marked
to have the alignment equal as its size, EmitAtomicExpr doesn't know about it and
will generate libcall instead of atomic op. The patch uses EmitPointerWithAlignment
to get the precise alignment information.
projects / clang / log