This brings the LLI orc-lazy JIT's behavior more closely in-line with LLI's
mcjit bahavior.

llvm-svn: 285413

Previously we were asserting that this declaration doesn't have a body
*and* won't have a body after we continue parsing.  This is too strong
and breaks the go-bindings test during codegen.

llvm-svn: 285412

Implement the -dI as supported by GCC: Output ‘#include’ directives in addition
to the result of preprocessing.

This change aims to add this option, pass it through to the preprocessor via
the options class, and when inclusions occur we output some information (+ test
cases).

Patch by Steve O'Brien!

Differential Revision: https://reviews.llvm.org/D25153

llvm-svn: 285411

Summary:
In CUDA compilation, we call isInlineDefinitionExternallyVisible (via
getGVALinkageForFunction) on functions while parsing their definitions.

At the point in time when we call getGVALinkageForFunction, we haven't
yet added the body to the function, so we trip this assert.  But as far
as I can tell, this is harmless.

To work around this, we add a new flag to FunctionDecl, "WillHaveBody".

There was other code that was working around the existing assert with a
really awful hack -- this change lets us get rid of that hack.

Reviewers: rsmith, tra

Subscribers: aemerson, cfe-commits

Differential Revision: https://reviews.llvm.org/D25640

llvm-svn: 285410

Do not use LiveIntervals to recalculate kills, because that cannot be
done accurately without implicit uses on predicated instructions.

llvm-svn: 285409

llvm-svn: 285408

Summary:
We were trying to add APInt values with different bit sizes after
visiting an addrspacecast instruction which changed the bit width
of the pointer.

Reviewers: majnemer, hfinkel

Subscribers: hfinkel, wdng, llvm-commits

Differential Revision: https://reviews.llvm.org/D24774

llvm-svn: 285407

Summary:
This is temporary, until bot that builds public facing LLVM
documentation is upgraded. It reverts only the cmake change in r284497,
but leaves the doc changes in place to preserve intent.

Reviewers: aaron.ballman

Subscribers: mgorny, llvm-commits

Differential Revision: https://reviews.llvm.org/D26078

llvm-svn: 285406

2. Cosmetic changes

llvm-svn: 285405

This was causing failures on windows bots. 

llvm-svn: 285404

llvm-svn: 285403

llvm-svn: 285402

s/disable_core/disable_coredump/
Add missing space in text.

llvm-svn: 285401

Currently fails with:

C:\workdir\compiler-rt\lib\tsan\go>gcc -c -o race_windows_amd64.syso gotsan.cc -I..\rtl -I..\.. -I..\..\sanitizer_common -I..\..\..\include -m64 -Wall -fno-exceptions -fno-rtti -DSANITIZER_GO -Wno-error=at
tributes -Wno-attributes -Wno-format -Wno-maybe-uninitialized -DSANITIZER_DEBUG=0 -O3 -fomit-frame-pointer -std=c++11 
gotsan.cc:9229:0: warning: ignoring #pragma comment  [-Wunknown-pragmas]
 #pragma comment(linker, "/alternatename:__sanitizer_print_memory_profile=__sanitizer_default_print_memory_profile") // NOLINT
 ^
gotsan.cc: In function 'void __sanitizer_print_memory_profile(int)':
gotsan.cc:9226:17: error: redefinition of 'void __sanitizer_print_memory_profile(int)'
 extern "C" void __sanitizer_print_memory_profile(int top_percent) {}
                 ^
gotsan.cc:6898:6: note: 'void __sanitizer_print_memory_profile(int)' previously defined here
 void __sanitizer_print_memory_profile(int top_percent) {


Remove __sanitizer_print_memory_profile from Go build, it is not needed.

llvm-svn: 285400

No need to check the remaining elements - no common known bits are available.

llvm-svn: 285399

No need to clear KnownOne2/KnownZero2 bits as the next call to computeKnownBits will overwrite them anyway

llvm-svn: 285398

[SelectionDAG] Increment computeKnownBits recursion depth for SMIN/SMAX/UMIN/UMAX like all other ops

llvm-svn: 285397

We suppress all Clang diagnostics (because they would be wrong,
include-fixer does custom recovery) but still want to give some feedback
in case there was a compiler error we couldn't recover from. The most
common case for this is a #include in the file that couldn't be found.

llvm-svn: 285396

OpenCL disallows using variadic arguments (s6.9.e and s6.12.5 OpenCL v2.0)
apart from some exceptions:
- printf
- enqueue_kernel

This change adds error diagnostic for variadic functions but accepts printf
and any compiler internal function (which should cover __enqueue_kernel_XXX cases).

It also unifies diagnostic with block prototype and adds missing uncaught cases for blocks.

llvm-svn: 285395

Now LPPassManager will run LCSSA verification only for the top-level loop
which was processed on the current iteration.

Differential Revision: https://reviews.llvm.org/D25873

llvm-svn: 285394

Most of them fail right now and are commented out. The main problem is handling
of backslashes on windows, but also the posix path code has a couple of issues.

llvm-svn: 285393

llvm-svn: 285392

This patch adds an objc_subclassing_restricted attribute into clang. This
attribute acts similarly to 'final' - Objective-C classes with this attribute
can't be subclassed. However, @interface declarations that have
objc_subclassing_restricted but don't have @implementation are allowed to
inherit other @interface declarations with objc_subclassing_restricted. This is
needed to describe the Swift class hierarchy in clang while making sure that
the Objective-C classes cannot subclass the Swift classes.

This attribute is already implemented in a fork of clang that's used for Swift
(https://github.com/apple/swift-clang) and this patch moves that code to the
upstream clang repository.

rdar://28937548

Differential Revision: https://reviews.llvm.org/D25993

llvm-svn: 285391

Add unreachable after enum switch statement

llvm-svn: 285390

llvm-svn: 285389

Patch by Sam McCall!

Differential Revision: https://reviews.llvm.org/D26067

llvm-svn: 285388

llvm-svn: 285387

If we pass a header to libclang, e.g. because it's open in an editor in
an IDE, warnings about unused const vars are not useful: other files
that include the header might use those constants. So when -x *-header
is passed as command-line option, suppress this warning.

llvm-svn: 285386

There is possible deadlock in dynamic ASan runtime when we dlopen() shared lib
which creates a thread at the global initialization stage. The scenario:

1) dlopen grabs a GI_pthread_mutex_lock in main thread.
2) main thread calls pthread_create, ASan intercepts it, calls real pthread_create
   and waits for the second thread to be "fully initialized".
3) Newly created thread tries to access a thread local disable_counter in LSan
   (to complete its "full initialization") and hangs in tls_get_addr_tail, because
   it also tries to acquire GI_pthread_mutex_lock.

The issue is reproducible on relative recent Glibc versions e.g. 2.23.

Differential Revision: https://reviews.llvm.org/D26028

llvm-svn: 285385

llvm-svn: 285384

llvm-svn: 285383

This patch does two seperate things. First it adds a file called
"__libcpp_version" which only contains the current libc++ version
(currently 4000). This file is not intended for use as a header. This file
is used by Clang in order to easily determine the installed libc++ version.
This allows Clang to enable/disable certain language features only when the
library supports them.

The second change is the addition of _LIBCPP_LIBRARY_VERSION macro, which
returns the version of the installed dylib since it may be different than
the headers.

llvm-svn: 285382

This seems to have increased LTO compile time bejond 2x of previous builds.
See http://lab.llvm.org:8080/green/job/clang-stage2-configure-Rlto/10676/

llvm-svn: 285381

Fixes PR 30784. Discussed with Justin, who pointed out that
in the new PassManager infrastructure we can have more fine-grained
control on which analyses we want to preserve, but this is the
best we can do with the current infrastructure.

llvm-svn: 285380

Summary:
Previously we were creating the alias summary on the fly while writing
the summary to bitcode. This moves the creation of these summaries to
the module summary index builder where we build the rest of the summary
index.

This is going to be necessary for setting the NoRename flag for values
possibly used in inline asm or module level asm.

Reviewers: mehdi_amini

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D26049

llvm-svn: 285379

Appease buildbots:
http://bb.pgr.jp/builders/cmake-clang-x86_64-linux/builds/55876

llvm-svn: 285378

This used to work before r284797 + r285152, which exposed something
interesting; some users include builtins from umbrella headers.

Clang should emit a warning to warn users this is not a good practice
and umbrella headers shouldn't get the
implicitly-add-the-builtin-version behavior for builtin header names.

While we're not there, add the testcase to represent the way it
currently works.

llvm-svn: 285377

Summary:
This is in preparation for a change to utilize this flag for symbols
referenced/defined in either inline or module level assembly.

Reviewers: mehdi_amini

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D26048

llvm-svn: 285376

llvm-svn: 285375

Summary:
After experimenting with CUDA, I realized that we really only need to
set the active context right before creating an object such as a stream
or a device memory allocation. When we go on to use these objects later,
it is fine if the context that created them is no longer active,
operations with those objects will succeed anyway.

Since it turns out that we don't have to check the active context for
every operation, it makes sense to hide this active context from users
(by removing the "ActiveDeviceForThread" setter and getter) and to
change the Acxxel API to explicitly pass in the device ID to create
objects.

This change improves the Acxxel API and greatly simplifies the CUDA and
OpenCL implementations because they no longer require thread_local data.

Reviewers: jlebar, jprice

Subscribers: mgorny, parallel_libs-commits

Differential Revision: https://reviews.llvm.org/D26050

llvm-svn: 285372