We are only adding this in `flang` - the bash wrapper script for the
driver. We don't have any specific plans to support `-M{fixed|free}` in
`flang-new` (i.e. the actual driver).

That feature was requested by one of our users.

Differential Revision: https://reviews.llvm.org/D107081

Android provides ashmem/ASharedMemory support on newer releases, which
we can use if requested by openmp users on Android.

Also refactor the preprocessor check for using shared memory to
kmp_config.h.cmake.

Differential Revision: https://reviews.llvm.org/D107181

Define an API for, and implement, runtime support for arbitrary
assignment of one descriptor's data to another, with full support for
(re)allocation of allocatables with finalization when necessary,
user-defined derived type assignment TBP calls, and intrinsic (default)
componentwise assignment of derived type instances with allocation of
automatic components.  Also clean up API and implementation of
finalization/destruction using knowledge gained while studying
edge cases for assignment in the 2018 standard.

The look-up procedure for special procedure bindings in derived
types has been optimized from O(N) to O(1) since it will probably
matter more.  This required some analysis in runtime derived type
description table construction in semantics and some changes to the
table schemata.

Executable Fortran tests have been developed; they'll be added
to the test base once they can be lowered and run by f18.

Differential Revision: https://reviews.llvm.org/D107678

Shorty before branching  LLVM 13 a new CMake option was added. This
option `LIBCXX_ENABLE_INCOMPLETE_FEATURES` lacks the contributor
documentation. This patch rectifies that issue.

Differential Revision: https://reviews.llvm.org/D107596

Switching `__builtin_clzll`  to `__libcpp_clz` should work on all
platforms and no longer require MSVC specific code.

Reviewed By: #libc, ldionne

Differential Revision: https://reviews.llvm.org/D107709

The last read access to this variable was removed in 2015 in
4446487d .

The `File::OpenOptions` were renamed; this fixes up a callsite that breaks for
macOS builds. (See
https://github.com/llvm/llvm-project/commit/14735cab655441ba45c4b88ad82f11267e5fe916)

Differential Revision: https://reviews.llvm.org/D107767

Due to unavailability of Flang testing on Windows, the shell scripts
are being ported to Python. The following changes are being made in
this patch: removed test_symbols.sh and common.sh, and ported them
to Python. Changes to the tests themselves so that they use the
python scripts instead of the shell script.

Reviewed By: Meinersbur, awarzynski, kiranchandramohan

Differential Revision: https://reviews.llvm.org/D107041

Reviewed By: Whitney

Differential Revision: https://reviews.llvm.org/D106815

The last use was removed on Apr 15, 2019 in commit
f56a4368.

All distributions have been updated using 7 days worth of data.

The patterns for fixed length gather/scatter with 32-bit offsets and
64-bit memory type are slightly different that the rest of the patterns,
as such the lowering needs to be slightly different to ensure the
correct types are used.

Differential Revision: https://reviews.llvm.org/D107576

This reverts commit 3229c971.

With a2acac68 in place this should provide enough info to work out
any repeat of the failure in cross_ovver_uniform_dist.test.

Followup after f9ffe61f ("[OpenCL] Factor out
OpenCLBuiltinFileEmitterBase; NFC", 2021-08-09) introduced a
-Wnon-virtual-dtor warning.

This patch is a revert of e08f205f. In that patch, DW_TAG_subprograms
were permitted to be referenced across CU boundaries, to improve stack
trace construction using call site information. Unfortunately, as
documented in PR48790, the way that subprograms are "owned" by dwarf units
is sufficiently complicated that subprograms end up in unexpected units,
invalidating cross-unit references.

There's no obvious way to easily fix this, and several attempts have
failed. Revert this to ensure correct DWARF is always emitted.

Three tests change in addition to the reversion, but they're all very
light alterations.

Differential Revision: https://reviews.llvm.org/D107076

When going through the CU entries in the name index,
make sure to compare the name entry's CU
offset against the skeleton CU's offset.

Previously there would be a mismatch, since the
wrong offset was compared, and thus no suitable
entry was found.

Reviewed By: jankratochvil

Differential Revision: https://reviews.llvm.org/D106270

Factor out functionality that can be shared with a header file emitter
that is to be added in the future.

Use hexadecimal numbers rather than decimal in various vFile packets
in order to fix compatibility with gdbserver.  This also changes the few
custom LLDB packets -- while technically they do not have to be changed,
it is easier to use the same syntax consistently across LLDB.

Differential Revision: https://reviews.llvm.org/D107475

Sync the mode constants used to drive vFile:open requests with these
used by GDB and defined for the gdb remote protocol.  This makes it
possible to use 'platform file open' after connecting to gdbremote
server (and to some degree to operate on the open file modulo other
incompatibilities).

Differential Revision: https://reviews.llvm.org/D106985

Modify OpenOptions enum to open the future path into synchronizing
vFile:open bits with GDB.  Currently, LLDB and GDB use different flag
models effectively making it impossible to match bits.  Notably, LLDB
uses two bits to indicate read and write status, and uses union of both
for read/write.  GDB uses a value of 0 for read-only, 1 for write-only
and 2 for read/write.

In order to future-proof the code for the GDB variant:

1. Add a distinct eOpenOptionReadWrite constant to be used instead
   of (eOpenOptionRead | eOpenOptionWrite) when R/W access is required.

2. Rename eOpenOptionRead and eOpenOptionWrite to eOpenOptionReadOnly
   and eOpenOptionWriteOnly respectively, to make it clear that they
   do not mean to be combined and require update to all call sites.

3. Use the intersection of all three flags when matching against
   the three possible values.

This commit does not change the actual bits used by LLDB.

Differential Revision: https://reviews.llvm.org/D106984

These ops were not ported to the nD vector conversion when it was introduced
and nobody needed them so far.

Reviewed By: gysit

Differential Revision: https://reviews.llvm.org/D107750

Shuffles which are broken into separate halves reveal splats in which
a half is accessed via one index; such operations can be optimized to
use "vrgather.vi".

This optimization could be achieved by adding extra patterns to match
`vrgather_vv_vl` which uses a splat as an index operand, but this patch
instead identifies splat earlier. This way, future optimizations can
build on top of the data gathered here, e.g., to splat-gather dominant
indices and insert any leftovers.

Reviewed By: craig.topper

Differential Revision: https://reviews.llvm.org/D107449

Differential Revision: https://reviews.llvm.org/D107433

These shuffles all take the form of a "splat" of the LHS and/or RHS to
some degree, with one or two elements needing patched up afterwards. We
currently lower all of these to full LHS/RHS vector-index shuffles with
vrgather.vv.

Reviewed By: craig.topper

Differential Revision: https://reviews.llvm.org/D107447

Following tests fail on Arm/AArch64 randomly with timeouts:

TestMultilineNavigation.py
TestBatchMode.py
TestUnicode.py
TestGdbRemote_vContThreads.py

I am marking them as skipped until we find a away make to pass reliably.

The patch removes mutable accessor methods for sections and segments.
As for now, const variants of them are not used because all callers have
mutable access to an instance of Object. On the other hand, they do not
actually modify the sets, so it looks better to keep only const ones.

Differential Revision: https://reviews.llvm.org/D107652

54902e00 added a use of
/alternatename via a #pragma comment(linker); in MinGW mode, this
requires building with -fms-extensions. (This flag is added to
SANITIZER_COMMON_CFLAGS in the toplevel CMakeLists.txt.)

This avoids a warning when building in MinGW mode (about an unknown
pragma being ignored), and presumably also makes the code work as
intended.

Differential Revision: https://reviews.llvm.org/D107620

The Headers.CountersDelta field is an uint64_t, not a pointer,
so just cast to uint32_t to truncate it.

Differential Revision: https://reviews.llvm.org/D107619

llvm/tools/llvm-readobj/XCOFFDumper.cpp:464:5: error:
  call to member function 'printNumber' is ambiguous
    W.printNumber("Length", StrTabSize);

In MSVC mode, CMake provides CMAKE_C_COMPILER_ARCHITECTURE_ID,
which should be accurate for what the compiler produces. Keep
the old CMAKE_C_SIZEOF_VOID_P based fallback in case this
doesn't provide matches.

In GCC/MinGW mode, CMake doesn't provide that. As we're trying to
guess the default host architecture, use CMAKE_HOST_SYSTEM_PROCESSOR
in combination with the existing CMAKE_C_SIZEOF_VOID_P check to
estimate whether it's x86/x86_64/arm/aarch64.

Differential Revision: https://reviews.llvm.org/D107626

Differential Revision: https://reviews.llvm.org/D107713

Added a new section on generating the html documentation
from the rst/md sources to our documentation.

Background: I wanted to check what my documenation
changes would look like on the website and had a hard
time finding how to do that. So I wanted to save other
folks the effort.

Differential Revision: https://reviews.llvm.org/D107460

This test didn't include all test check lines, thanks to .'s in function
names. It also changed the triple to hard float to make a more
interesting test for NEON code generation.