And use it to model LLVM IR's `ptrtoint` cast.

This is essentially an alternative to D88806, but with no chance for
all the problems it caused due to having the cast as implicit there.
(see rG7ee6c402)

As we've established by now, there are at least two reasons why we want this:
* It will allow SCEV to actually model the `ptrtoint` casts
  and their operands, instead of treating them as `SCEVUnknown`
* It should help with initial problem of PR46786 - this should eventually allow us
  to not loose pointer-ness of an expression in more cases

As discussed in [[ https://bugs.llvm.org/show_bug.cgi?id=46786 | PR46786 ]], in principle,
we could just extend `SCEVUnknown` with a `is ptrtoint` cast, because `ScalarEvolution::getPtrToIntExpr()`
should sink the cast as far down into the expression as possible,
so in the end we should always end up with `SCEVPtrToIntExpr` of `SCEVUnknown`.

But i think that it isn't the best solution, because it doesn't really matter
from memory consumption side - there probably won't be *that* many `SCEVPtrToIntExpr`s
for it to matter, and it allows for much better discoverability.

Reviewed By: mkazantsev

Differential Revision: https://reviews.llvm.org/D89456

Differential revision: https://reviews.llvm.org/D90321

From D89768

Previously they were separated into "instance" and "kind" aliases, and also required that the dialect know ahead of time all of the instances that would have a corresponding alias. This approach was very clunky and not ergonomic to interact with. The new approach is to provide the dialect with an instance  of an attribute/type to provide an alias for, fully replacing the original split approach.

Differential Revision: https://reviews.llvm.org/D89354

The existing logic in determining whether an alloca should live on the frame only looks explicit def-use relationships. However a value defined by an alloca may be implicitly needed across suspension points, either because an alias has across-suspension-point def-use relationship, or escaped by store/call/memory intrinsics. To properly handle all these cases, we have to properly visit the alloca pointer up-front. Thie patch extends the exisiting alloca use visitor to determine whether an alloca should live on the frame.

Differential Revision: https://reviews.llvm.org/D89768

We have the `--check-prefixes` invocation for a single suffix.
It is cleaner to use `--check-prefix` in this case.

Almost any sanitizer can try to install interceptors.

This adds some initial content as well as structure to the new OpenMP
Sphinx documentation hosted at http://openmp.llvm.org/docs/ .

The content contains some useful links but most pages are still empty.

This uses a "custom" theme which is a copy of the default "agogo" one
with minor modifications to get a nicer table of content in the sidebar.
This way we can also adjust the theme as we go.

Reviewed By: jhuber6, JonChesterfield

Differential Revision: https://reviews.llvm.org/D90256

ParseOpenMP.cpp was pretty much clang-formatted except a few minor
locations. Let's make it a clang formatted file.

Reviewed By: JonChesterfield

Differential Revision: https://reviews.llvm.org/D90440

If MCContext has an error, MCAssembler::layout may stop early
and some MCFragment's may not finalize.

In the Linux kernel, arch/x86/lib/memcpy_64.S could trigger the assert before
"x86_64: Change .weak to SYM_FUNC_START_WEAK for arch/x86/lib/mem*_64.S"

From below mentioned standard references
OpenACC 3.0 Standards document
840 • A program may not branch into or out of an OpenACC parallel construct

OpenMP 5.0 Standards document
A program that branches into or out of a parallel region is non-conforming.

This patch
Resolves the issue of exit out of a parallel region, other branching out issues like goto statements are not handled with this patch.
Moves code from D87906 to be reused by other OpenMP/OpenACC to check-directive-structure.h.
Adds support in OpenMP parallel construct and a test case to verify.

Reviewed By: clementval

Differential Revision: https://reviews.llvm.org/D88655

Fixes check-all with LLVM_USE_SANITIZER=Memory

It is semantically equivalent, but the intent was really lost there.
This fixes a warning/error from MSVC as well, see PR48013

Allow standalone builds to inherit the LLVM_LIT_ARGS.

[MC] Add SMLoc to MCStreamer::emitSymbolAttribute and report changed binding warnings/errors for ELF

This patch is mainly doing two things:

1. Adding support for parentheses, making the combination of target features
   more diverse;
2. Making the priority of ’,‘ is higher than that of '|' by default. So I need
   to make some change with PTX Builtin function.

Differential Revision: https://reviews.llvm.org/D89184

Temporarily XFAIL'ing TestTypeGetModule.py while the DWO failure is
being investigated.

Differential revision: https://reviews.llvm.org/D88483

prune-eh will not be ported to the NPM. Instead, a combination of
function-attrs and simplifycfg should be used (as described in
https://reviews.llvm.org/D44415).

This pins most tests using -prune-eh to the legacy PM. Some of these
were testing legacy PM infra (mostly the CGPassManager). Some of these
can be tested in the NPM using function-attrs and simplifycfg.

One interesting case is simplenoreturntest.ll. function-attrs +
simplifycfg does not yet make a caller of a caller of a noreturn
function end with unreachable like prune-eh does. That can be added in
the future.

Reviewed By: asbirlea

Differential Revision: https://reviews.llvm.org/D90012

Pass AAQI in places where it was missed.

Part of D89991.
Author: haoranxu510 (Haoran Xu)

Recognize the __apple_ sections as debug info sections and make sure
they're included in the --show-sections-sizes output.

Differential revision: https://reviews.llvm.org/D90433

This is to enable --allow-unused-duplicates=false. This prefix appears
to be outdated and intentionally unused.

Reviewed By: rupprecht

Differential Revision: https://reviews.llvm.org/D90427

* Removes index based insertion. All insertion now happens through the insertion point.
* Introduces thread local context managers for implicit creation relative to an insertion point.
* Introduces (but does not yet use) binding the Context to the thread local context stack. Intent is to refactor all methods to take context optionally and have them use the default if available.
* Adds C APIs for mlirOperationGetParentOperation(), mlirOperationGetBlock() and mlirBlockGetTerminator().
* Removes an assert in PyOperation creation that was incorrectly constraining. There is already a TODO to rework the keepAlive field that it was guarding and without the assert, it is no worse than the current state.

Differential Revision: https://reviews.llvm.org/D90368

Differential Revision: https://reviews.llvm.org/D90428

Make check_clang_tidy.py not just pass -format-style=none by default
but a full -config={}. Without this, with a build dir outside of
the llvm root dir and a .clang-tidy config further up that contains

  CheckOptions:
    - key:          modernize-use-default-member-init.UseAssignment
      value:        1

these tests would fail:

   Clang Tools :: clang-tidy/checkers/cppcoreguidelines-prefer-member-initializer-modernize-use-default-member-init.cpp
   Clang Tools :: clang-tidy/checkers/modernize-use-default-member-init-bitfield.cpp
   Clang Tools :: clang-tidy/checkers/modernize-use-default-member-init.cpp

After this change, they pass fine, despite the unrelated
.clang-tidy file further up.

The test can be cleaned up a bit, but this should be good to see why the
Debian bot is failing...

This fixes the issue pointed out in D84604#2363134. For now we exclude
static members completely, we'll take them into account later.

Make all of the "AMDGPU Machine Code GFX*" columns in the Memory Model
table a consistent width of 32-characters.

Best viewed with something like --word-diff

Differential Revision: https://reviews.llvm.org/D89977

Mostly NFC, but some changes are "bug fixes" rather than just e.g.
formatting changes or typo corrections.

- Fix typo "competing" -> "completing".
- Document why waintcnt is added to stores and not loads for
  sequentially consistent ordering.
- Lowercase some mentions of `buffer_gl{0,1}_inv`.
- Make mentions of `*cnt(0)` consistently include the `(0)` count.
- Remove some mentions of instructions for incorrect address spaces. For
  example, remove mention of `flat_load` from
  `load atomic acquire workgroup global`.
- Re-flow some text to get all the target columns to fit in a
  32-character wide column. Makes a future NFC patch to make these columns
  both 32-character wide more straightforward.

Modified cherry-pick of patch by Tony Tye

Reviewed By: t-tye

Differential Revision: https://reviews.llvm.org/D89596

Mitch expressed a preference to not have `#ifdef`s in platform agnostic
code, this change tries to accomodate this.

I am not attached to the method this CL proposes, so if anyone has a
suggestion, I am open.

We move the platform specific member of the mutex into its own platform
specific class that the main `Mutex` class inherits from. Functions are
implemented in their respective platform specific compilation units.

For Fuchsia, we use the sync APIs, as those are also the ones being
used in Scudo.

Differential Revision: https://reviews.llvm.org/D90351

This is very similar to 7f1e6fcf, just fixing a left-over.
With this, it should be possible to use both, -x cuda and -fopenmp in the same invocation,
enabling to use both OpenMP, targeting CPU, and CUDA, targeting the GPU.

Reviewed By: jdoerfert

Differential Revision: https://reviews.llvm.org/D90415

Fix semantic in the distribute integration test based on offline feedback. This
exposed a bug in block distribution, we need to make sure the id is multiplied
by the stride of the vector. Fix the transformation and unit test.

Differential Revision: https://reviews.llvm.org/D89291