Now when we are checking if the expression `co_await
promise.final_suspend()` is not throw, we would check unconditionally
for its child expressions recursively. It takes unnecessary time. And
the compiler would complains if the implementation in final_suspend()
may throw even if the higher level function signature marked noexcept
already.

This fixes bug48453 too.

Do not build llc and mlir-cpu-runner to run MLIR tests with the native target isn't configured (NFC)

The relevant tests are already disabled in this case.

Differential Revision: https://reviews.llvm.org/D117283

This makes `ninja check-mlir` work without the host targets configured.

Differential Revision: https://reviews.llvm.org/D117241

Summary:
Follow the clang-tidy suggestion to replace reset-release with move assignment.

Move assignment's effect for unique_ptr:
> Effects: Transfers ownership from `u` to `*this` as if by calling `reset(u.release())`
followed by an assignment from `std::forward<D>(u.get_deleter())`.

Summary:
LLVM Programmer’s Manual strongly discourages the use of `std::vector<bool>`
and suggests `llvm::BitVector` as a possible replacement.

Partially revert cac7aabb: some APIs don't have in-tree users and some build breaks weren't caught

This is important because *Objects targets need to only depend on other *Objects targets, not on the unsuffixed CAPI rules. Depending on how it is linked in the current setup, it can cause duplicate symbols.

Differential Revision: https://reviews.llvm.org/D117176

We could use vmv.v.i/vmv.v.x whose eew is 32 to lower the i64 splat vector if the i64 constant scalar could be splitted into two same i32 scalar.

Differential Revision: https://reviews.llvm.org/D117079

Precommit test for D117079.

Differential Revision: https://reviews.llvm.org/D117081

The abi name test shows incorrect fcvt.d.h, fcvt.h.d, fcvt.f.h,
and fcvt.h.f instructions around the inline assembly.

Using named registers as input or output constraints creates fcvt.d.s
and fcvt.s.d instructions around the inline assembly.

This makes the data unusable by the inline assembly and corrupts
the results of the inline assembly.

Reviewed By: #sanitizers, vitalybuka

Differential Revision: https://reviews.llvm.org/D116376

This builds on D116521 to add this functionality to Polly without more copy-pasted code.

Reviewed By: beanz

Differential Revision: https://reviews.llvm.org/D116555

TLS initializers, for example constructors of thread-local variables, don't necessarily get called. If a thread was created before a module is loaded, the module's TLS initializers are not executed for this particular thread.

This is why Microsoft added support for dynamic TLS initialization. Before every use of thread-local variables, a check is added that runs the module's TLS initializers on-demand.

To do this, the method `__dyn_tls_on_demand_init` gets called. Internally, it simply calls `__dyn_tls_init`.

No additional TLS initializer that sets the guard needs to be emitted, as the guard always gets set by `__dyn_tls_init`.
The guard is also checked again within `__dyn_tls_init`. This makes our check redundant, however, as Microsoft's compiler also emits this check, the behaviour is adopted here.

Reviewed By: majnemer

Differential Revision: https://reviews.llvm.org/D115456

When constructing an OperationName, the overwhelming majority of
cases are from registered operations. This revision adds a non-locked
lookup into the currently registered operations, which prevents locking
in the common case. This revision also optimizes several uses of
RegisteredOperationName that expect the operation to be registered,
e.g. such as in OpBuilder.

These changes provides a reasonable speedup (5-10%) in some
compilations, especially on platforms where locking is expensive.

Differential Revision: https://reviews.llvm.org/D117187

This commit sometimes causes a crash when compiling a vtable thunk. E.g.:

clang '--target=aarch64-grtev4-linux-gnu' -xc++ - -c -o /dev/null <<EOF
struct a {
  virtual int f();
};
struct c {
  virtual int &g() const;
};
struct d : a, c {
  int &g() const;
};
int &d::g() const {}
EOF

Some follow-up commits have been reverted as well:
Revert "IR: Make getRetAlign check callee function attributes"
Revert "Fix MSVC "32-bit shift implicitly converted to 64 bits" warning. NFC."
Revert "Fix MSVC "32-bit shift implicitly converted to 64 bits" warning. NFC."

This reverts commit 4f414af6.
This reverts commit a5507d2e.
This reverts commit 3d2d208f.
This reverts commit 07ddfa95.

Differential Revision: https://reviews.llvm.org/D116642

This guarantees the preconditions of fromCOO; whereas prior to this, one could call the constructor directly with an unsorted tensor, which would cause fromCOO to misbehave.

Reviewed By: aartbik

Differential Revision: https://reviews.llvm.org/D117167

Differential Revision: https://reviews.llvm.org/D116300

This patch fixes a crash due to following simple program:
> export struct Unit {
>    bool operator<(const Unit&);
> };

It would crash since the compiler would set the module ownership for
Unit. And the declaration with a module ownership is assumed to own a
module. But here isn't one. So here is the crash.

This patch fixes this by exiting early if it finds the export decl is
already invalid.

Reviewed By: aaron.ballman

Differential Revision: https://reviews.llvm.org/D117093

Add `std::ranges::in_in_result`

Reviewed By: Quuxplusone, Mordante, #libc

Spies: ldionne, libcxx-commits, mgorny

Differential Revision: https://reviews.llvm.org/D116278