llvm-svn: 200779

llvm-svn: 200778

This patch fixes the ldr-pseudo implementation to work when used in
inline assembly.  The fix is to move arm assembler constant pools
from the ARMAsmParser class to the ARMTargetStreamer class.

Previously we kept the assembler generated constant pools in the
ARMAsmParser object. This does not work for inline assembly because
a new parser object is created for each blob of inline assembly.
This patch moves the constant pools to the ARMTargetStreamer class
so that the constant pool will remain alive for the entire code
generation process.

An ARMTargetStreamer class is now required for the arm backend.
There was no existing implementation for MachO, only Asm and ELF.
Instead of creating an empty MachO subclass, we decided to make the
ARMTargetStreamer a non-abstract class and provide default
(llvm_unreachable) implementations for the non constant-pool related
methods.

Differential Revision: http://llvm-reviews.chandlerc.com/D2638

llvm-svn: 200777

This fixes a crashes in the OpenCV test suite and also the scrypt
kernel in bfgminer.

I was unable to come up with a reduced test case for this.

https://bugs.freedesktop.org/show_bug.cgi?id=72785

llvm-svn: 200776

There is no lit test for this, because it would be too big and
complicated, but it does fix a crash in the Arithm/Absdiff.* OpenCV test.

llvm-svn: 200775

llvm-svn: 200774

The OpenCL specs say: "The vector versions of the math functions operate
component-wise. The description is per-component."

Patch by: Jan Vesely

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 200773

Based on the following discussion:
http://llvm.1065342.n5.nabble.com/LLVM-3-4-stable-releases-td65005.html

llvm-svn: 200772

rdar://problem/13729466

llvm-svn: 200771

Now that the back-end intrinsics are more regular, there's no need for the
special handling these got in the front-end, so they can be moved to
EmitCommonNeonBuiltinExpr.

llvm-svn: 200769

There was an extremely confusing proliferation of LLVM intrinsics to implement
the vacge & vacgt instructions. This combines them all into two polymorphic
intrinsics, shared across both backends.

llvm-svn: 200768

Patch thanks to Stephan Tolksdorf!

llvm-svn: 200767

in preparation for https://codereview.appspot.com/55100044

llvm-svn: 200766

CMake's target_link_libraries() will manage dependencies.

Configuration time may be reduced by a few seconds.

llvm-svn: 200765

[CMake] LLVM-Config.cmake: Split explicit_map_components_to_libraries and introduce llvm_map_components_to_libnames and llvm_expand_dependencies.

llvm-svn: 200764

llvm-svn: 200763

In trunk, every users assume add_llvm_loadable_module as "loadable module" and no one sets neither SHARED, ... nor also MODULE!
Unfortunately, all loadable modules were linked as not "MODULE" but "SHARED".

If this change caused any regressions, I wish guys to fix it properly. ;)

llvm-svn: 200762

llvm-svn: 200761

llvm-svn: 200760

llvm-svn: 200759

llvm-svn: 200758

llvm-svn: 200757

[ASan] Use replace_str=0 and replace_intrin=0 in printf-{2,4}.c to avoid triggering a buffer overflow in strlen and memcpy.
This change fixes printf-2.c and printf-4.c under OSX.

llvm-svn: 200756

llvm-svn: 200755

Until now, when a path in a gcno file included a directory, we would
emit our .gcov file in that directory, whereas gcov always emits the
file in the current directory. In doing so, this implements gcov's
strange name-mangling -p flag, which is needed to avoid clobbering
files when two with the same name exist in different directories.

The path mangling is a bit ugly and only handles unix-like paths, but
it's simple, and it doesn't make any guesses as to how it should
behave outside of what gcov documents. If we decide this should be
cross platform later, we can consider the compatibility implications
then.

llvm-svn: 200754

If the -fno-function-sections flag isn't recognized, the compiler
emits a warning. This isn't enough to cause the check to fail, so we
enable -Werror as well.

llvm-svn: 200753

Missing braces on if meant we inserted both ARM and Thumb load for a litpool
entry. This didn't end well.

rdar://problem/15959157

llvm-svn: 200752

[ASan] Fix printf-5.c on OSX: fix a compiler warning, avoid intercepting memcpy (which triggers the error report before printf does).

llvm-svn: 200751

in preparation for https://codereview.appspot.com/55100044

llvm-svn: 200750

It's incorrect to call the function by its name in the interceptor, because that means WRAP(vname) on Linux and REAL(vname) on OSX.

This patch fixes asan/lit_tests/TestCases/printf-3.c on Mac.

llvm-svn: 200749

llvm-svn: 200748

so that it's possible to override them from a dynamic library.

llvm-svn: 200747

llvm-svn: 200746

llvm-svn: 200745

If LLVM_ENABLE_LIBCXX is specified, we should append -stdlib=libc++ to build
flags as early as possible, in particular, before we check for header presence
(as -stdlib=libc++ modifies header lookup rules). Otherwise we can find a header
at configure time (w/o -stdlib=libc++) but fail to find it at build time
(with -stdlib=libc++). See PR18569 for more details.

llvm-svn: 200744

V_ADD_F32 with source modifier does not produce -0.0 for this. Just
manipulate the sign bit directly instead.

Also add a pattern for (fneg (fabs ...)).

Fixes a bunch of bit encoding piglit tests with radeonsi.

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 200743

We do not detect errno spoiling for SIGTERM,
because some SIGTERM handlers do spoil errno but reraise SIGTERM,
tsan reports false positive in such case.
It's difficult to properly detect this situation (reraise),
because in async signal processing case (when handler is called directly
from rtl_generic_sighandler) we have not yet received the reraised
signal; and it looks too fragile to intercept all ways to reraise a signal.

llvm-svn: 200742

llvm-svn: 200741

When gcov is run without gcda data, it acts as if the counts are all
zero and labels the file as - to indicate that there was no data. We
should do the same.

llvm-svn: 200740

llvm-svn: 200739