llvm-svn: 243639

Summary:
This hidden option would disable code generation through FastISel by
default. It was removed from the available options and from the
Fast-ISel tests that required it in order to run the tests.

Reviewers: dsanders

Subscribers: qcolombet, llvm-commits

Differential Revision: http://reviews.llvm.org/D11610

llvm-svn: 243638

Differential revision: http://reviews.llvm.org/D11595

llvm-svn: 243637

Summary:
Previously, we would sign-extend non-boolean negative constants and
zero-extend otherwise. This was problematic for PHI instructions with
negative values that had a type with bitwidth less than that of the
register used for materialization.

More specifically, ComputePHILiveOutRegInfo() assumes the constants
present in a PHI node are zero extended in their container and
afterwards deduces the known bits.

For example, previously we would materialize an i16 -4 with the
following instruction:

  addiu $r, $zero, -4

The register would end-up with the 32-bit 2's complement representation
of -4. However, ComputePHILiveOutRegInfo() would generate a constant
with the upper 16-bits set to zero. The SelectionDAG builder would use
that information to generate an AssertZero node that would remove any
subsequent trunc & zero_extend nodes.

In theory, we should modify ComputePHILiveOutRegInfo() to consult
target-specific hooks about the way they prefer to materialize the
given constants. However, git-blame reports that this specific code
has not been touched since 2011 and it seems to be working well for every
target so far.

Reviewers: dsanders

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D11592

llvm-svn: 243636

OpenMP 4.1 introduces optional argument '(n)' for 'ordered' clause, where 'n' is a number of loops that immediately follow the directive.
'n' must be constant positive integer expressions and it must be less or equal than the number of the loops in the resulting loop nest.
Patch adds parsing and semantic analysis for this optional argument.

llvm-svn: 243635

Summary:
Mips doesn't implement unw_getcontext() or libunwind::Registers_*::jumpto() yet
so we must disable libunwind for this release.

Reviewers: hans

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D11563

llvm-svn: 243633

This contains the test-case for r243630.

Patch by: marina.yatsina@intel.com
Differential Revision: http://reviews.llvm.org/D11513

llvm-svn: 243632

Patch by: marina.yatsina@intel.com
Differential Revision: http://reviews.llvm.org/D11512

llvm-svn: 243630

llvm-svn: 243628

This is used by the AMD x86 OpenCL implementation
to change some ABI details on Windows and Linux.

llvm-svn: 243627

llvm-svn: 243626

llvm-svn: 243625

I think I was isolating the failure in the last RUN: line during
development, and forgot to re-enable the others.

llvm-svn: 243624

llvm-svn: 243623

Patch by Nitesh Jain

Reviewers: clayborg, ovyalov.
Subscribers: jaydeep, bhushan, mohit.bhakkad, sagar, emaste, lldb-commits.
Differential Revision: http://reviews.llvm.org/D11562

llvm-svn: 243622

llvm-svn: 243621

Patch by Nitesh Jain

Reviewers: clayborg, ovyalov.
Subscribers: jaydeep, bhushan, mohit.bhakkad, sagar, emaste, lldb-commits.
Differential Revision: http://reviews.llvm.org/D11176

llvm-svn: 243620

Summary:
Currently if the "first child" of the pointer is a char type then the pointer is displayed as a string. This test succeeds incorrectly when the pointer is to a structured type with a char type as its first field. Fix this by switching the test to retrieve the pointee type and checking that it is a char type.


Reviewers: abidh, ChuckR, ki.stfu

Subscribers: greggm, lldb-commits

Differential Revision: http://reviews.llvm.org/D11488

llvm-svn: 243619

    SUMMARY:
    The patch creates Unix Signals based on target architecture. For MIPS it creates MipsLinuxSignals.
    
    Reviewers: clayborg
    Subscribers: mohit.bhakkad, sagar, lldb-commits
    Differential Revision: http://reviews.llvm.org/D11455

llvm-svn: 243618

The reason I was passing this vector by value in the constructor so that
I wouldn't have to copy when initializing the corresponding member but
then I forgot the std::move.

The use-case is LoopDistribution which filters the checks then
std::moves it to LoopVersioning's constructor.  With this interface we
can avoid any copies.

llvm-svn: 243616

llvm-svn: 243615

llvm-svn: 243614

Before, we were passing the pointer partitions to LAA.  Now, we get all
the checks from LAA and filter out the checks within partitions in
LoopDistribution.

This effectively concludes the steps to move filtering memchecks from
LAA into its clients.  There is still some cleanup left to remove the
unused interfaces in LAA that still take PtrPartition.

(Moving this functionality to LoopDistribution also requires
needsChecking on pointers to be made public.)

llvm-svn: 243613

[modules] Remove redundant information written into DeclContext name lookup tables. We don't need to store the data length twice.

llvm-svn: 243612

llvm-svn: 243611

[sanitizer] add a weak hook for strncmp interceptor, both to dfsan and other sanitizers. Hide the declaration and the calls in better macros

llvm-svn: 243610

llvm-svn: 243609

Clang will not define __i686__, even when the target triple is i686,
without -march=i686.

With this patch, the compiler-rt build will successfully detect that
Clang can target i686.

The open_memstream.cc test is a little funny. Before my patch, it
was invoked with "-m32 -m64". To make it work after my -march
change, I had to add '-march=x86-64'.

Differential Revision: http://reviews.llvm.org/D11618

llvm-svn: 243604

[libFuzzer] implement memcmp hook for data-flow-guided fuzzing (w/o dfsan), extend the memcmp fuzzer test

llvm-svn: 243603

We currently don't canonicalize paths in the preprocessed files.
But we do when writing to PCH.
This causes a discrepancy on Windows with the test below.
This test fails even on unix if you change the test to use
`%S//preprocess.h`.

I am led to conclude that the invariant that this test was intending to
test has not been upheld for a while (and may never have been).

llvm-svn: 243602

[sanitizer] add a weak hook for memcmp interceptor, to be used primarily for fuzzing. More hooks will be added later. So far this is a Linux-only feature

llvm-svn: 243601

It looks like we were somehow relying somewhere on removing 'foo/./bar'
but *not* 'foo/../foo/bar'. Currently investigating.

llvm-svn: 243600

Let me tell you a story. Suppose you want to build your project (e.g. LLVM)
with CMake 2.8, Clang and AddressSanitizer. You also want to ensure that
Clang is fresh enough and check that CMAKE_CXX_COMPILER_VERSION is 3.1+.
This check would fail - CMake would fail to correctly calculate compiler
version if you pass CMAKE_CXX_FLAGS=-fsanitize=address.

The problem is funky compiler version calculation in
CMakeDetermineCompilerId.cmake module: it compiles the sample source
file with provided compiler and compile flags, runs "strings" and greps
for "INFO:" ASCII strings contained on the executable to fetch
"INFO:compiler", "INFO:compiler_version" etc. It limits the output of
grep to just 4 lines.

Unfortunately, if your executable was built with ASan, it would also contain
an ASCII string
  INFO: %s ignores mlock/mlockall/munlock/munlockall
and INFO:compiler_version string would never be parsed.

All of the above actually happened after r243574 when we tried to
configure libcxx with just-built Clang with TSan/MSan, and the version
check mentioned above failed in HandleLLVMOptions.cmake

(╯°□°)╯.~.┻━┻

llvm-svn: 243599

This was fallout from r243581.  Turns out C++14 has make_reverse_iterator.

Thanks to Filipe and David for the quick fix suggestion.

llvm-svn: 243598

Also fix completely broken and untested code which was hiding the
primary bug. The !LLVM_ON_UNIX branch of the ifdef was actually a no-op.

I ran into this in the wild. It was causing failures in our SDK build.

Ideally we'd have a perfect llvm::sys::fs::canonical, but at least this
is a step in the right direction, and fixes an obviously broken case.
In some sense the test case I've added here is an integration test. We
should have these routines thoroughly unit tested in llvm::sys::fs.

llvm-svn: 243597

This makes it simpler to add instruction types that don't depend on fast-math.

llvm-svn: 243596

[asan,tsan,msan] move the memcmp interceptor from asan/tsan to sanitizer_common. This may potentially lead to more reports from msan as it now sees the reads inside memcmp. To disable, use the flag intercept_memcmp=0. Likewise, it may potentially cause new races to appear due to more strict memcmp checking (flag strict_memcmp=1)

llvm-svn: 243595

Without DR1579 implemented, the only case for -Wredundant-move is for a
parameter being returned with the same type as the function return type.  Also
include a check to verify that the move constructor will be used by matching
nodes in the AST dump.

llvm-svn: 243594

llvm-svn: 243593

UsingShadowDecls over other declarations of the same entity in the lookup
results. This ensures that we build correct redeclaration chains for the
UsingShadowDecls (otherwise we could see assertions and other misbehavior in
modules builds, when merging combines multiple redeclaration chains for the
same entity from the same module into one chain).

llvm-svn: 243592