Move the implementation of the `toString` function from
`llvm/Support/Error.h` to the source file, which allows us to move
`#include "llvm/ADT/StringExtras.h"` to the source file as well.

As `Error.h` is present in a large number of translation units this
means we are unnecessarily bringing in the contents of
`StringExtras.h` - itself a large file with lots of includes - and
slowing down compilation.

Also move the `#include "llvm/ADT/SmallVector.h"` directive to the
source file as it's no longer needed, but this does not give as much of
a benefit.

This reduces the total number of preprocessing tokens across the LLVM
source files in lib from (roughly) 1,920,413,050 to 1,903,629,230 - a
reduction of ~0.87%. This should result in a small improvement in
compilation time.

Differential Revision: https://reviews.llvm.org/D153229

In preparation for removing the `#include "llvm/ADT/StringExtras.h"`
from the header to source file of `llvm/Support/Error.h`, first add in
all the missing includes that were previously included transitively
through this header.

Reviewed By: wangpc

Differential Revision: https://reviews.llvm.org/D153669

SiFive7's scalar execution consists of 4 stages AG, M1, M2, WB.
Most simple arithmetic and branch instructions can execute in
either AG or M2.

If the operands are ready, the instruction will execute in the AG
stage. Otherwise, it executes in the M2 stage. Everything is fully
bypassed, so dependent instructions should only see 1 cycle latency.

This patch adds ReadAdvances to pretend that these instructions
execute in the M2 ALU and reads their operands then. This allows
the scheduler to schedule dependent instructions back to back.

I've increased branch latency to 3 since they are also executed in both
stages. Still need to fix JALR, but I want to cleanup some scheduler
classes first.

Multiply, cpop and division instructions can only start in the AG stage.

Still need to do some work for FP instructions that produce integer results.

I've added an llvm-mca test that creates a long dependency chain.
The timeline view can show that things are bypassed. I
didn't check all permutations, but we have some variety.

Reviewed By: wangpc

Differential Revision: https://reviews.llvm.org/D153666

Now that the SVE parts are in, we can fill in the double reduction tests
without them causing problems.

The following test fails to compile TU b.cpp because we are not making the transitively imported modules visible (per [module.import]/p7)

```
a.cppm:
export module a;

export int foo() {
   return 42;
}

b.cppm:
export module b;
import a;

export int bar();

b.cpp:
module b;

int bar() {
   return foo();
}

clang++ -c -std=c++2b -fmodule-output a.cppm
clang++ -c -std=c++2b -fmodule-output -fprebuilt-module-path=. b.cppm
clang++ -c -std=c++2b -fprebuilt-module-path=. b.cpp
b.cpp:4:12: error: declaration of 'foo' must be imported from module 'a' before it is required
   return foo();
```

This is fixed by the following patch (which also addresses a FIXME in basic.def.odr/p6.cppm).

Differential Revision: https://reviews.llvm.org/D152746

D137796 made these passes unused.

`opt --bpf-ir-peephole` is specified in one test. Add a `registerPipelineParsingCallback`
so that we can use change the test to use `opt --passes=bpf-ir-peephole` instead.

D152340 has split WinCOFFObjectWriter to WinCOFFWriter. This patch adds
another WinCOFFWriter as DwoWriter to write Dwo sections to dwo file.
Driver options are also updated accordingly to support -gsplit-dwarf in
CL mode.

e.g. $ clang-cl  -c -gdwarf -gsplit-dwarf foo.c

Like what -gsplit-dwarf did in ELF, using this option will create DWARF object
(.dwo) file. DWARF debug info is split between COFF object and DWARF object
file. It can reduce the executable file size especially for large project.

Reviewed By: skan, MaskRay

Differential Revision: https://reviews.llvm.org/D152785

Reviewed By: MaskRay

Differential Revision: https://reviews.llvm.org/D153699

Reviewed By: kkwli0

Differential Revision: https://reviews.llvm.org/D153703

Implement P2494R2 `Relaxing range adaptors to allow for move only types`

https://open-std.org/jtc1/sc22/wg21/docs/papers/2022/p2494r2.html#wording-ftm

According to the words in P2494R2, I haven't add new test for `drop_while_view`, `take_while_view` and `filter_view`, because these views has the requirement that the predicate is an `indirect_unary_predicate`, which requires that the predicate is `copy_constructible`, so they still can't accept move only types as predicate.

```
[P2483R0] also suggests future work to relax the requirements on the predicate types stored by standard views. This paper does not perform this relaxation, as the copy constructibility requirement is enshrined in the indirect callable concepts ([indirectcallable.indirectinvocable]). Thus, while this paper modifies the views that currently use copyable-box for user provided predicates, it only does so to apply the rename of the exposition-only type to movable-box; it does not change any of the constraints on those views. It does, however, relax the requirements on invocables accepted by the transform family of views, because those are not constrained using the indirect callable concepts.
```

Reviewed By: #libc, var-const

Differential Revision: https://reviews.llvm.org/D151629

 Fixes windows build.

There is a minor behavior difference that is not worth testing for the obsoleted
format. Previously, llvm-dwarfdump considers .zdebug_info as a debug section but
does not decompress it, leading to a warning when the content cannot be parsed.
Now llvm-dwarfdump just ignores the section without a warning.

type_traits is currently unable to include __type_traits/noexcept_move_assign_container.h, because it would cause several include cycles.

type_traits -> __type_traits/noexcept_move_assign_container.h -> __memory/allocator_traits.h -> __memory/construct_at.h -> new -> exception -> type_traits

type_traits -> __type_traits/noexcept_move_assign_container.h -> __memory/allocator_traits.h -> __memory/construct_at.h -> new -> type_traits

type_traits -> __type_traits/noexcept_move_assign_container.h -> __memory/allocator_traits.h -> limits -> type_traits

This is a problem for clang modules after the std mega module is broken up (D144322), because it becomes a module cycle which is a hard error.

Unconditionally remove the type_traits includes from limits and new in all versions, and also remove the exception include from new. (These are already removed in C++23.)

Reviewed By: ldionne, Mordante, #libc

Differential Revision: https://reviews.llvm.org/D153214

D106624 added a .dwo warning (when there are relocations) that may fire for
non-debug sections, e.g. `.rodata.dwo` when there is a data symbol foo in
-fdata-sections mode. Adjust it to only warn for .debug sections.

While here, change the diagnostic to be more conventional
https://llvm.org/docs/CodingStandards.html#error-and-warning-messages and use
the relocated section name instead of the relocation section name.

This change does not handle `.zdebug` (support was mostly removed from LLVM) or
`__debug` (Mach-O, no DWO support).

Reviewed By: ayermolo, HaohaiWen

Differential Revision: https://reviews.llvm.org/D153602

Fix build error caused by D153679

Reviewed By: nikic

Differential Revision: https://reviews.llvm.org/D153704

Reviewed By: #libc, Mordante

Differential Revision: https://reviews.llvm.org/D153427

Reviewed By: RKSimon

Differential Revision: https://reviews.llvm.org/D153684

Addresses:  LWG3719 - Directory iterators should be usable with default sentinel
https://wg21.link/LWG3719

Reviewed By: #libc, Mordante

Differential Revision: https://reviews.llvm.org/D153299

A nested class in a member function can erroneously confuse the check into
thinking that any CXXThisExpr found relate to the outer class, preventing any warnings.
Fix this by not traversing any nested classes.

Fixes https://github.com/llvm/llvm-project/issues/56765

Reviewed By: PiotrZSL

Differential Revision: https://reviews.llvm.org/D130665

The last use was removed by:

  commit 7d894374
  Author: OCHyams <orlando.hyams@sony.com>
  Date:   Thu Mar 16 08:48:36 2023 +0000

The last uses were removed by:

  commit 79b3f082
  Author: Tim Northover <tnorthover@apple.com>
  Date:   Tue Feb 7 11:28:00 2023 +0000

Differential Revision: https://reviews.llvm.org/D153679

Differential Revision: https://reviews.llvm.org/D153678

The last use was removed by:

  commit 73523021
  Author: Chandler Carruth <chandlerc@gmail.com>
  Date:   Mon Jan 13 13:07:17 2014 +0000

The last use was removed by:

  commit 243acd5d
  Author: Nikita Popov <npopov@redhat.com>
  Date:   Fri Dec 9 15:29:35 2022 +0100

The last use was removed by:

  commit 2d79b713
  Author: Philip Reames <preames@rivosinc.com>
  Date:   Thu Apr 6 08:23:08 2023 -0700

The last use was removed by:

  commit f9b523eb
  Author: Kazu Hirata <kazu@google.com>
  Date:   Wed May 31 07:30:14 2023 -0700

Check detects cyclic #include dependencies between user-defined headers.

Reviewed By: njames93

Differential Revision: https://reviews.llvm.org/D144828

Following what is already performed for vector.reduce.fmin/fmax, this adds
lowering for the new vector.reduce.fminimum/fmaximum nodes to the SVE fminv
and fmaxv instructions via the existing FMINV_PRED/FMAXV_PRED nodes.

Differential Revision: https://reviews.llvm.org/D153288

This reverts commit f3aa6cc0.