Thsi patch add the infrastructure to lower the random related
intrinsics:

- `random_init`
- `random_number`
- `random_seed`

This patch is part of the upstreaming effort from fir-dev branch.

Reviewed By: PeteSteinfeld, schweitz

Differential Revision: https://reviews.llvm.org/D121704



Co-authored-by: V Donaldson <vdonaldson@nvidia.com>
Co-authored-by: Jean Perier <jperier@nvidia.com>

When regcomp() fails, the same regex_t* should be passed to regerror()
for potentially better error messages.  But doing that with msan would
report a use-of-uninitialized-value.

Fixes https://github.com/google/sanitizers/issues/1496

Reviewed By: vitalybuka

Differential Revision: https://reviews.llvm.org/D120591

This patch adds lowering for the following intrinsics:
- `max`
- `maxloc`
- `maxval`
- `minloc`
- `minval`

This patch is part of the upstreaming effort from fir-dev branch.

Reviewed By: PeteSteinfeld

Differential Revision: https://reviews.llvm.org/D121701



Co-authored-by: Jean Perier <jperier@nvidia.com>
Co-authored-by: Eric Schweitz <eschweitz@nvidia.com>
Co-authored-by: mleair <leairmark@gmail.com>

Optimize a pattern where a sequence of 8/16 or 32 bits is tested for
zero: LLVM normalizes this towards and `AND` with mask which is usually
good, but does not work well on X86 when the mask does not fit into a
64bit register. This DagToDAG peephole transforms sequences like:

```
movabsq $562941363486720, %rax # imm = 0x1FFFE00000000
testq %rax, %rdi
```

to

```
shrq $33, %rdi
testw %di, %di
```

The result has a shorter encoding and saves a register if the tested
value isn't used otherwise.

Differential Revision: https://reviews.llvm.org/D121320

Differential Revision: https://reviews.llvm.org/D121319

This should make CI consistent on all the compilers we support. Most of
this patch is working around various warnings emitted by GCC in our code
base, which are now being shown when we compile the tests.

After this patch, the whole test suite should be warning free on all
compilers we support and test, except for a few warnings on GCC that
we silence explicitly until we figure out the proper fix for them.

Differential Revision: https://reviews.llvm.org/D120684

This patch adds lowering for the following character related intrinsics:
- `len`
- `len_trim`
- `lge`, `lgt`, `lle` and `llt`

This patch is part of the upstreaming effort from fir-dev branch.

Reviewed By: PeteSteinfeld

Differential Revision: https://reviews.llvm.org/D121703



Co-authored-by: V Donaldson <vdonaldson@nvidia.com>
Co-authored-by: Jean Perier <jperier@nvidia.com>
Co-authored-by: Eric Schweitz <eschweitz@nvidia.com>

This replaces a bunch of duplicate logic to set up environment variables
and a few substitutions.

It does a little more than we were doing previously:
 - searching for clang and setting up substitutions for it
 - setting up some substitutions for target triples, which are
   potentially useful but not actually used

clangd has been happily using this for its tests for a while.

Differential Revision: https://reviews.llvm.org/D121422

This patch adds lowering for the `allocated`
intrinsic.

This patch is part of the upstreaming effort from fir-dev branch.

Reviewed By: jeanPerier, PeteSteinfeld

Differential Revision: https://reviews.llvm.org/D121702



Co-authored-by: Jean Perier <jperier@nvidia.com>

A version of this logic appears in ~every lit.site.cfg.in (28 copies total).
This patch just removes two, but I'll update the rest of llvm-project next.

Differential Revision: https://reviews.llvm.org/D121427

`UdtRecordCompleter` shouldn't complete static members' types. static members' types are going to be completed when the types are called in `SymbolFile::CompleteType`.

Reviewed By: labath

Differential Revision: https://reviews.llvm.org/D121030

This patch adds the lowering infrastructure for the lowering of
alternat returns.

This patch is part of the upstreaming effort from fir-dev branch.

Depends on D121698

Reviewed By: PeteSteinfeld

Differential Revision: https://reviews.llvm.org/D121699



Co-authored-by: V Donaldson <vdonaldson@nvidia.com>
Co-authored-by: Jean Perier <jperier@nvidia.com>
Co-authored-by: Eric Schweitz <eschweitz@nvidia.com>

This patch add the lowering for the entry statement.

This patch is part of the upstreaming effort from fir-dev branch.

Depends on D121697

Reviewed By: PeteSteinfeld

Differential Revision: https://reviews.llvm.org/D121698



Co-authored-by: V Donaldson <vdonaldson@nvidia.com>
Co-authored-by: Jean Perier <jperier@nvidia.com>
Co-authored-by: Eric Schweitz <eschweitz@nvidia.com>

This patch lowers more cases of pointer assignments and
disassociations.

This patch is part of the upstreaming effort from fir-dev branch.

Reviewed By: PeteSteinfeld, schweitz

Differential Revision: https://reviews.llvm.org/D121697



Co-authored-by: V Donaldson <vdonaldson@nvidia.com>
Co-authored-by: Jean Perier <jperier@nvidia.com>
Co-authored-by: mleair <leairmark@gmail.com>
Co-authored-by: Eric Schweitz <eschweitz@nvidia.com>

This fixes a bug that happens when using -fdebug-prefix-map to remap an
absolute path to a relative path. Since the path was absolute before
remapping, it is safe to assume that concatenating the remapped working
directory would be wrong.

This was originally submitted as https://reviews.llvm.org/D113718, but
reverted because when testing with dwarf 5 enabled, the tests were too
strict.

Differential Revision: https://reviews.llvm.org/D121663

Fixes https://github.com/llvm/llvm-project/issues/54367
Fixes https://github.com/llvm/llvm-project/issues/54368

Differential Revision: https://reviews.llvm.org/D121596

Back in 3a208c68, we implemented the range-based constructor for <span>.
However, in doing so, we removed a previous non-standard constructor that
we provided before shipping <ranges>. Unfortunately, that breaks code that
was relying on a range-based constructor until we ship all of <ranges>.

This patch reintroduces the old non-conforming constructors and tests
that were removed in 3a208c68 and uses them whenever <ranges> is
not provided (e.g. in LLVM 14). This is only a temporary workaround
until we enable <ranges> by default in C++20, which should hopefully
happen by LLVM 15.

The goal is to cherry-pick this workaround back to the LLVM 14 release
branch, since I suspect the constructor removal may otherwise cause
breakage out there, like the breakage I saw internally.

We could have avoided this situation by waiting for C++20 to be finalized
before shipping std::span. For example, we could have guarded it with
something like _LIBCPP_HAS_NO_INCOMPLETE_RANGES to prevent users from
accidentally starting to depend on it before it is stable. We did not
have these mechanisms when std::span was first implemented, though.

Differential Revision: https://reviews.llvm.org/D121626

If sanitizer cannot determine name of the module it
will use "<unknown module>". Then it can be suppressed
if needed.

Reviewed By: kda

Differential Revision: https://reviews.llvm.org/D121674

function candidates

See: https://github.com/llvm/llvm-project/issues/54379

I tried to see if I can reuse ResolveAddressOfOverloadedFunction for
explicit function instantiation and so I managed to hit this ICE.

Bug was the diagnostic required an argument (%0) and specific code path
didn't pass an argument.

Differential Revision: https://reviews.llvm.org/D121646

Differential Revision: https://reviews.llvm.org/D121732

Add synchronization to the IOHandler to prevent multiple threads from
writing concurrently to the output or error stream.

A scenario where this could happen is when a thread (the default event
thread for example) is using the debugger's asynchronous stream. We
would delegate this operation to the IOHandler which might be running on
another thread. Until this patch there was nothing to synchronize the
two at the IOHandler level.

Differential revision: https://reviews.llvm.org/D121500

Reviewed By: yota9

Differential Revision: https://reviews.llvm.org/D121729

Update FunctionAttrs to use FunctionModRefBehavior instead
MemoryAccessKind.

This allows for adding support for inferring argmemonly and others,
see D121415.

Reviewed By: nikic

Differential Revision: https://reviews.llvm.org/D121460

GCC supports:
- `namespace <gnu attributes> identifier`
- `namespace identifier <gnu attributes>`

But clang supports only `namespace identifier <gnu attributes>` and diagnostics for `namespace <gnu attributes> identifier` case looks unclear:
Code:
```
namespace __attribute__((visibility("hidden"))) A
{
}
```
Diags:
```
test.cpp:1:49: error: expected identifier or '{'
namespace __attribute__((visibility("hidden"))) A
                                                ^
test.cpp:1:49: error: C++ requires a type specifier for all declarations
test.cpp:3:2: error: expected ';' after top level declarator
}
```

This patch adds support for `namespace <gnu attributes> identifier` and also forbids gnu attributes for nested namespaces (this already done for C++ attributes).

Reviewed By: aaron.ballman

Differential Revision: https://reviews.llvm.org/D121245

This patch follows the same reasoning as D114481. The PCH reader looks for `__clangast` section in the precompiled module file, which is not present in the file on AIX and not supported in XCOFF yet.

Reviewed By: daltenty

Differential Revision: https://reviews.llvm.org/D121709

The cold text section alignment is set using the maximum alignment value
passed to the emitCodeAlignment. In order to calculate tentetive layout
right we will set the minimum alignment of such sections to the maximum
possible function alignment explicitly.

Differential Revision: https://reviews.llvm.org/D121392

Add the calling convention for the vector pair registers.
These registers overlap with the vector registers.

Part of an original patch by: Lei Huang

Reviewed By: nemanjai, #powerpc

Differential Revision: https://reviews.llvm.org/D117225

List derived from https://docs.microsoft.com/en-us/cpp/cpp/keywords-cpp
. Not that this is something we really want to encourage, but some of
these show up in practice, so I figured I should just complete the list.

Differential Revision: https://reviews.llvm.org/D121412

Fold linalg.fill into linalg.generic.
Remove dead arguments used in linalg.generic.

Reviewed By: ThomasRaoux

Differential Revision: https://reviews.llvm.org/D121535

With 229d576b the class EHFrameSplitter was renamed to DWARFRecordSectionSplitter. This change merely moves it to it's own .cpp/.h file

Differential Revision: https://reviews.llvm.org/D121721

The class `__dir_stream` is currently declared in two places: as a
top-level forward declaration in `directory_iterator.h`, and as a friend
declaration in class `directory_entry`, in `directory_entry.h`.

The former has a `_LIBCPP_HIDDEN` attribute, but the latter does not,
causing the Firefox build to complain about the visibility not matching
the previous declaration. This is because Firefox plays games with
pushing and popping visibility.

Work around this by making both `__dir_stream` declarations consistently
use `_LIBCPP_HIDDEN`.

Reviewed By: ldionne, philnik, #libc

Differential Revision: https://reviews.llvm.org/D121639

This is a fix for #53737. In addition to commas, newline characters are
considered as separators of checks.

If libcurl was built with CMake, CMake's FindCURL module defers
completely to the included config file. This config file doesn't set any
of the variables that the current check script depends on; it just sets
up an imported CMake target. Accordingly, the smoke test fails, since it
can't find the libcurl (or its static dependencies).

This changes the compile smoke test to refer to the imported library
instead; this should in turn bring in the necessary include and library
directories via the interface properties set up by CMake. This better
mirrors the way libcurl is referred to elsewhere in the CMakeLists.

Reviewed By: MaskRay

Differential Revision: https://reviews.llvm.org/D121718

NFC. Update script does not behave right since the run lines have
identical output. Delete the duplicated check prefix added in
22cfbf7e

NFC. Hasn't been updated since script added check-next

NFC. Hasn't been updated since the update script started adding
check-next.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D121719