I still don't see why we need to select to different Real instructions
on different targets, but at least this is less verbose.

Previously `report_fatal_error` is used for reporting something goes
wrong in the backend, but this is confusing because `report_fatal_error`
basically means there are something unexpected & crashed in the backend.

So, turn this "crash" into an elegant error reporting. After this patch,
clang can diagnose it:

    bpf-crash.c:4:30: error: Invalid usage of the XADD return value
4 | u32 next_event_id() { return __sync_fetch_and_add(&GLOBAL_EVENT_ID,
1); }
        |                              ^
    1 error generated.

[Clang][Sema] Don't build CXXDependentScopeMemberExprs for potentially implicit class member access expressions (#92318)

According to [expr.prim.id.general] p2:
> If an _id-expression_ `E` denotes a non-static non-type member of some
class `C` at a point where the current class is `X` and
> - `E` is potentially evaluated or `C` is `X` or a base class of `X`,
and
> - `E` is not the _id-expression_ of a class member access expression,
and
> - if `E` is a _qualified-id_, `E` is not the un-parenthesized operand
of the unary `&` operator,
>
> the _id-expression_ is transformed into a class member access
expression using `(*this)` as the object expression.

Consider the following:
```
struct A
{
    void f0();

    template<typename T>
    void f1();
};

template<typename T>
struct B : T
{
    auto g0() -> decltype(T::f0()); // ok

    auto g1() -> decltype(T::template f1<int>()); // error: call to non-static member function without an object argument
};

template struct B<A>;
```

Clang incorrectly rejects the call to `f1` in the _trailing-return-type_
of `g1`. Furthermore, the following snippet results in a crash during
codegen:
```
struct A
{
    void f();
};

template<typename T>
struct B : T
{
    template<typename U>
    static void g();
    
    template<>
    void g<int>()
    {
        return T::f(); // crash here
    }
};

template struct B<A>;
```
This happens because we unconditionally build a
`CXXDependentScopeMemberExpr` (with an implicit object expression) for
`T::f` when parsing the template definition, even though we don't know
whether `g` is an implicit object member function yet.

This patch fixes these issues by instead building
`DependentScopeDeclRefExpr`s for such expressions, and only transforming
them into implicit class member access expressions during instantiation.
Since we implemented the MS "unqualified lookup into dependent bases"
extension by building an implicit class member access (and relying on
the first component name of the _nested-name-specifier_ to be looked up
in the context of the object expression during instantiation), we
instead pre-append a fake _nested-name-specifier_ that refers to the
injected-class-name of the enclosing class. This patch also refactors
`Sema::BuildQualifiedDeclarationNameExpr` and
`Sema::BuildQualifiedTemplateIdExpr`, streamlining their implementation
and removing any redundant checks.

This is in effect a revert of f139ae3d, as we have since gained a
more sophisticated way of doing extra IRGen with the addition of
RawAddress in #86923.

Co-authored-by: Jeremy Kun <j2kun@users.noreply.github.com>

[SPIR-V] Ensure that internal intrinsic functions "ptrcast" for PHI's operand are inserted at the correct positions (#92536)

The goal of the PR is to ensure that newly inserted `ptrcast` internal
intrinsic functions are inserted at the correct positions, and don't
break rules of instruction domination and PHI nodes grouping at top of
basic block.

[SPIR-V] reqd_work_group_size and work_group_size_hint metadata are correctly converted to the LocalSize and LocalSizeHint execution mode (#92552)

The goal of this PR is to ensure that reqd_work_group_size and
work_group_size_hint metadata are correctly converted to the LocalSize
and LocalSizeHint execution mode.

reqd_work_group_size and work_group_size_hint require 3 operands (see
https://registry.khronos.org/SPIR-V/specs/unified1/SPIRV.html#Execution_Mode),
if metadata contains less operands, just add a default value (1).

In preparation for #92682.

This commit fixes the missing semicolon in the PTX codegen path where
barrier id is provided for the NVVM BarrierArriveOp. Also, updated
nvvm-to-llvm.mlir lit test to reflect the same

Co-authored-by: pradeepku <pradeepku@nvidia.com>

I may eventually add getVPZeroExtendInReg to SelectionDAG if there are
other cases, but for now just hardcode it.

We could try harder for nonsplat vectors but probably not worth the
effort.

LLVM is now using GitHub. This change makes revert_checker.py -u
generate commit links that go to GitHub, instead of the old Phabricator
URIs.

flang/lib/Optimizer/Dialect/CUF/Attributes/CUFAttr.cpp includes
CUFDialect.h.inc, but the target generating that isn't currently
depended on in CUF/Attributes. This patch adds that missing dependency.

Fixes #92635

Clang crashes when diagnosing the following invalid redeclaration in
C++11:
```
struct A {
  void f();
};

constexpr void A::f() { } // crash here
```
This happens because `DiagnoseInvalidRedeclaration` tries to create a
fix-it to remove `const` from the out-of-line declaration of `f`, but
there is no `SourceLocation` for the `const` qualifier (it's implicitly
`const` due to `constexpr`) and an assert in
`FunctionTypeInfo::getConstQualifierLoc` fails.

This patch changes `DiagnoseInvalidRedeclaration` to only suggest the removal of the
`const` qualifier when it was explicitly specified in the _cv-qualifier-seq_ of the declaration.

Reland "[ThinLTO] Populate declaration import status except for distributed ThinLTO under a default-off new option" (#92718)

The original PR is reviewed in
https://github.com/llvm/llvm-project/pull/88024, and this PR adds one
line (https://github.com/llvm/llvm-project/pull/92718/commits/b9f04d199dec4f3c221d981dcb91e55298d0693f)
to fix test

Limit to one thread for in-process ThinLTO to test `LLVM_DEBUG` log.
- This should fix build bot failure like
https://lab.llvm.org/buildbot/#/builders/259/builds/4727 and
https://lab.llvm.org/buildbot/#/builders/9/builds/43876
- I could repro the failure and see interleaved log messages by using
`-thinlto-threads=all`

**Original Commit Message:**

The goal is to populate `declaration` import status if a new flag
`-import-declaration` is on.

* For in-process ThinLTO, the `declaration` status is visible to backend
`function-import` pass, so `FunctionImporter::importFunctions` should
read the import status and be no-op for declaration summaries.
Basically, the postlink pipeline is updated to keep its current behavior
(import definitions), but not updated to handle `declaration` summaries.
Two use cases ([better call-graph
sort](https://discourse.llvm.org/t/rfc-for-better-call-graph-sort-build-a-more-complete-call-graph-by-adding-more-indirect-call-edges/74029#support-cross-module-function-declaration-import-5)
or [cross-module
auto-init](https://github.com/llvm/llvm-project/pull/87597#discussion_r1556067195))
would use this bit differently.

* For distributed ThinLTO, the `declaration` status is not serialized to
bitcode. As discussed, https://github.com/llvm/llvm-project/pull/87600
will do this.

Per comments in DivRemPairs, the rewrite from
```llvm
    %div = div %X, %Y
    %rem = rem %X, %Y
```
to
```llvm
    %div = div %X, %Y
    %.mul = mul %div, %Y
    %rem = sub %X, %mul
```
is unsound when %X or %Y are undef.

However, it is known to be sound if %X or %Y are poison but can't be
undef, since both the pre- and post-rewrite %rem are `poison`.

Additionally, proofs: https://alive2.llvm.org/ce/z/xtNQ8j

A comment in the pass listed a TODO for changing a usage of
isGuaranteedNotToBeUndefOrPoison() in the pass to something that only
detects undef. Such a function has been implemented since the time that
TODO was written, but has not been used.

Therefore, this commit updates DivRemPairs to use
isGuaranteedNotToBeUndef() instead.

The tests are added to a new AMDGPU/ subdirectory since I found the
missed optimization while hacking on AMDGPU code. Also, this ensures
that AMDGPU, which uses DivRemPass, is being checked for existing
expected behavior.

Summary:
The plugins are no longer linked to a share library, making this unused
and useless.

Fix linkage of `build_name`; it is not supposed to have external
linkage.

Fixes #92537

For PlayStation, make the propagation of DLL import/export attributes
for explicit template instantiations the same as MSVC and Windows
Itanium.

This predicate tells GlobalISelEmitter and DAGISelEmitter to check that
the instruction to emit has only one use of its result. This can be used
on a PatFrag instead of defining custom predicates for both emitters per
record that requires it.

clang rejects some valid code (see testcases) because of an incorrect
transformed deduction guides. This patch fixes it.

We miss the template argument packs during the transformation (`auto
(type-parameter-0-0...) -> Foo<>`). In
`TreeTransform::TransformTemplateArguments `, we have a logic of
handling template argument packs which were originally added to support
CTAD alias, it doesn't seem to be needed, we need to unpack them.

Extra tests for https://github.com/llvm/llvm-project/pull/92307

According to IEEE Std 754-2019, `sqrt` returns nan when the input is
negative (except for -0). In this case, we cannot make assumptions about
sign bit of the result.

Fixes https://github.com/llvm/llvm-project/issues/92217

We do not need to concern ourselves with CGSCC since all remaining CG
related updates went away in fa6ea7a4 as
pointed out by @nikic in
https://github.com/llvm/llvm-project/pull/87963#issuecomment-2113937182.

Split off from https://github.com/llvm/llvm-project/pull/89386, this
extends the binary matcher to support matching commuative operations.
This is used for a new m_c_BinaryOr matcher, used in simplifyRecipe.

PR: https://github.com/llvm/llvm-project/pull/92539

The parametric op was not checking the symbol it points to is a type or
attribute. This PR also fixes a small bug where an invalid IRDL file
would not end processing in mlir-opt. I also improved the error messages
for the already handled irdl.base invalid symbols.

This patch splits definitions for the OpenMP dialect into multiple files
to simplify the addition of new features, reduce merge conflicts, make
it easier to understand, etc. The split is based on the structure of the
more mature LLVMIR dialect. More specifically:

- The OpenMP dialect definition is located in OpenMPDialect.td.
- Base classes for OpenMP operations and types, as well as generic
OpenMP types are moved to OpenMPOpBase.td.
- OpenMP enumeration attributes, their case attributes and shared base
classes for these are placed in OpenMPEnums.td.
- Other OpenMP attributes are separated into OpenMPAttrDefs.td.
- OpenMPOps.td only contains operation definitions.

Even though this change should be useful on its own, it is intended as a
precursor to a follow-up PR in which operation arguments and attributes
are split into clause-specific classes which are then shared by all
operations to which they apply to. Without this prior change, that
approach would make the OpenMPOps.td file harder to navigate.

already.

This means that this pass will also run on hlfir intrinsics which are
not inside of functions.

See RFC:

https://discourse.llvm.org/t/rfc-add-an-interface-for-top-level-container-operations

Some of the changes are from moving the declaration and definition of
the constructor into tablegen (as requested during code review of
another pass).

Add examples of patterns that can be simplified, but are currently not.
This patch serves as a pre-commit test.

This patch canonicalizes constant expression GEPs to use i8 source
element type, aka ptradd. This is the ConstantFolding equivalent of the
InstCombine canonicalization introduced in #68882.

I believe all our optimizations working on constant expression GEPs
(like GlobalOpt etc) have already been switched to work on offsets, so I
don't expect any significant fallout from this change.

This is part of:
https://discourse.llvm.org/t/rfc-replacing-getelementptr-with-ptradd/68699