It's been pointed out that arbitrarily spraying raw profiles into a
build directory is insane. Doing this wastes a tremendous amount of
space and is also very lossy, since the test harness tends to wipe away
temporary sub-directories (which usually contain relevant profile data).

The new default is a `profiles` directory inside of the build dir.

llvm-svn: 276504

This is in preparation of
s/getConstantDifference/computeConstantDifference/ in a later change.

llvm-svn: 276503

llvm-svn: 276502

llvm-svn: 276501

llvm-svn: 276500

Only around 50% of the intrinsics in this file are documented now. The patches for the rest of the intrisics in this file will be send out later.

The doxygen comments are automatically generated based on Sony's intrinsics docu
ment.

I got an OK from Eric Christopher to commit doxygen comments without prior code
review upstream. This patch was internally reviewed by Paul Robinson.

llvm-svn: 276499

This reverts commit r276298.

Data stored in .rodata can have a negative offset from .text, but we
don't support negative values in relocations yet.

This caused a regression in one of the amp conformance tests:
5_Data_Cont/5_2_a_v/5_2_3_m/Assignment/Test.02.01

llvm-svn: 276498

This implementation was ported from the AMD builtin library
and has been tested with piglit, OpenCV, and the ocl conformance tests.

llvm-svn: 276497

This implementation was ported from the AMD builtin library
and has been tested with piglit, OpenCV, and the ocl conformance tests.

llvm-svn: 276496

llvm-svn: 276495

llvm-svn: 276494

To test that online merging is enabled by default.

llvm-svn: 276493

decomposition declarations.

There are a couple of things in the wording that seem strange here:
decomposition declarations are permitted at namespace scope (which we partially
support here) and they are permitted as the declaration in a template (which we
reject).

llvm-svn: 276492

llvm-svn: 276491

Revert "fix https://reviews.llvm.org/D22610" and "[compiler-rt] Fix memmove/memcpy overlap detection on windows"

This currently fails ~500 tests on Darwin:
http://lab.llvm.org:8080/green/job/clang-stage1-configure-RA_check/20456/

This reverts commit 4cfee0dff1facb8fa2827d25c5943bfef96d1a8f and
dbd91205d578cb61ab77be06087e9f65ba8a7ec8.

llvm-svn: 276490

llvm-svn: 276488

llvm-svn: 276487

This patch teaches FunctionInfo about offsets.

Like the last patch, this one doesn't introduce any visible
functionality change (the core algorithm knows nothing about offsets;
they're just plumbed through). Tests will come when we start acting
differently because of the offsets.

Patch by Jia Chen.

(N.B. I made a tiny change to Jia's patch to avoid warnings by GCC: I
put DenseMapInfo specializations in the `llvm` namespace. Only realized
that those appeared when compiling locally. :) )

Differential Revision: https://reviews.llvm.org/D22634

llvm-svn: 276486

llvm-svn: 276485

This patch enables raw profile merging for this option which is the
new intended behavior.

llvm-svn: 276484

llvm-svn: 276482

llvm-svn: 276481

Summary:
Clang inserts cleanup code before resume similar way as before return instruction.
This makes asan poison local variables causing false use-after-scope reports.

__asan_handle_no_return does not help here as it was executed before
llvm.lifetime.end inserted into resume block.

To avoid false report we need to unpoison stack for resume same way as for return.

PR27453

Reviewers: kcc, eugenis

Differential Revision: https://reviews.llvm.org/D22661

llvm-svn: 276480

Summary:
Adding a flag to diable GVN Hoisting by default.
Note: The GVN Hoist Pass causes some Halide tests to hang. Halide will disable the pass while investigating.

Reviewers: llvm-commits, chandlerc, spop, dberlin

Subscribers: mehdi_amini

Differential Revision: https://reviews.llvm.org/D22639

llvm-svn: 276479

This change breaks up the monolithic TestConcurrentEvents.py into a
separate subdir per test method. This allows them to run concurrently,
reduces the chance of a timeout occurring during normal operation, and
allows us to home in on any test methods that may be locking up.

This is step one in the process of squashing timeouts in these test
methods.

The reason for breaking each test method into its own file is to make it
very clear to us if there are a subset of the tests that do in fact lock
up frequently. This will limit how much hunting we need to do to
recreate it.

The reason for putting each file in a separate subdirectory is so that
our concurrent test runner can run multiple test files at the same time.
The unit of serialization in the LLDB test suite is the test directory,
so moving them into separate directories enables the test runner to do
more at the same time.

This change introduces usage of VPATH from gnu make. I use that to
facilitate keeping a single copy of the main.cpp in the parent
concurrent_events directory. Initially I had tried specifying the source
file as ../main.cpp, but our current makefile rules get confused by that
and then also build the output into the parent directory, which defeats
the ability to run each of the test methods concurrently. In the event
that not all systems support VPATH, I can do a bit of surgery on the
Makefile rules and attempt to make it smarter with regards to relative
paths to source files used in the build.

llvm-svn: 276478

When vectorizing a tree rooted at a store bundle, we currently try to sort the
stores before building the tree, so that the stores can be vectorized. For other
trees, the order of the root bundle - which determines the order of all other
bundles - is arbitrary. That is bad, since if a leaf bundle of consecutive loads
happens to appear in the wrong order, we will not vectorize it.

This is partially mitigated when the root is a binary operator, by trying to
build a "reversed" tree when that's considered profitable. This patch extends the
workaround we have for binops to trees rooted in a horizontal reduction.

This fixes PR28474.

Differential Revision: https://reviews.llvm.org/D22554

llvm-svn: 276477

llvm-svn: 276476

llvm-svn: 276475

llvm-svn: 276474

Processing update records (and loading a module, in general) might trigger
unexpected calls to the ASTWriter (being a mutation listener). Now we have a
mechanism to suppress those calls to the ASTWriter but notify other possible
mutation listeners.

Fixes https://llvm.org/bugs/show_bug.cgi?id=28332

Patch by Cristina Cristescu and me.

Reviewed by Richard Smith (D21800).

llvm-svn: 276473

llvm-svn: 276472

Differential Revision: https://reviews.llvm.org/D22700

llvm-svn: 276471

This adds versions of operator + and - which are optimized for the LHS/RHS of the
operator being RValue's.  When an RValue is available, we can use its storage space
instead of allocating new space.

On code such as ConstantRange which makes heavy use of APInt's over 64-bits in size,
this results in significant numbers of saved allocations.

Thanks to David Blaikie for all the review and most of the code here.

llvm-svn: 276470

llvm-svn: 276469

Just link in the stylesheet from the toplevel dir of the report.

llvm-svn: 276468

The helper will get smarter in a later change, but right now this is
just code reorganization.

llvm-svn: 276467

llvm-svn: 276466

llvm-svn: 276465

llvm-svn: 276464

For some reason this doesn't cause linker errors with MSVC or
clang-cl, but the bots seem to be failing with other compilers.

llvm-svn: 276463