Re-enable tests starting with a or b.

Fixes #93003

Libc++ has no separate C++98 support, it uses C++03 instead. This
removes some obsolete c++98 markers in the test.

Thanks to @StephanTLavavej for spotting this.

Additional test coverage for the VPlan-based cost model work.

This patch moves `Sema` functions that are specific for RISC-V into the
new `SemaRISCV` class. This continues previous efforts to split `Sema`
up. Additional context can be found in
https://github.com/llvm/llvm-project/pull/84184.

This PR is somewhat different from previous PRs on this topic:
1. Splitting out target-specific functions wasn't previously discussed.
It felt quite natural to do, though.
2. I had to make some static function in `SemaChecking.cpp` member
functions of `Sema` in order to use them in `SemaRISCV`.
3. I dropped "RISCV" from identifiers, but decided to leave "RVV"
(RISC-V "V" vector extensions) intact. I think it's an idiomatic
abbreviation at this point, but I'm open to input from contributors in
that area.
4. I repurposed `SemaRISCVVectorLookup.cpp` for `SemaRISCV`.

I think this was a successful experiment, which both helps the goal of
splitting `Sema` up, and shows a way to approach `SemaChecking.cpp`,
which I wasn't sure how to approach before. As we move more
target-specific function out of there, we'll gradually make the checking
"framework" inside `SemaChecking.cpp` public, which is currently a whole
bunch of static functions. This would enable us to move more functions
outside of `SemaChecking.cpp`.

Placing physically next to each other remaining constructors filling
vector with the same value
will make code better, since they all have nearly identical
implementation, which needs to be kept in sync.

Co-authored-by: Mark de Wever <koraq@xs4all.nl>

Recognize `cl_khr_kernel_clock` builtins and translate them to
`OpReadClockKHR` instructions. The `Scope` operand is deduced from the
builtin function name.

spirv-val does not pass yet due to OpReadClockKHR only supporting the
valid scopes for Vulkan (Device and Subgroup, but not Workgroup), so
leave validation disabled with a TODO.

This patch adds support for G_UADDSAT/G_SADDSAT/G_USUBSAT/G_SSUBSAT by
lowering it into add/sub with selects. When zbb is available,
min/max/minu/maxu will be selected.

This patch includes the "no signed wrap" and "no unsigned wrap" flags,
which can be used to annotate some Ops in the `arith` dialect and also
in LLVMIR, in the integer range inference.

The general approach is to use saturating arithmetic operations to infer
bounds which are assumed to not wrap and use overflowing arithmetic
operations in the normal case. If overflow is detected in the normal
case,
special handling makes sure that we don't underestimate the result
range.

The behaviour of the flag should be equivalent to
__arm_streaming_compatible.

At the moment, the name suggests that '-force-streaming-compatible-sve'
on its own (i.e. without specifying `+sve`) enables the compiler to use
the streaming-compatible subset of SVE instructions, but the semantics
merely are that the function can be called with either PSTATE.SM=0 or
PSTATE.SM=1.

This reverts commit c769079b.

This doesn't fix the build, it breaks it.

This PR attempts to consolidate the different topological sort utilities
into one place. It adds them to the analysis folder because the
`SliceAnalysis` uses some of these.

There are now two different sorting strategies: 
1. Sort only according to SSA use-def chains
2. Sort while taking regions into account. This requires a much more
elaborate traversal and cannot be applied on graph regions that easily.

This additionally reimplements the region aware topological sorting
because the previous implementation had an exponential space complexity.

I'm open to suggestions on how to combine this further or how to fuse
the test passes.

The `FileManager` might create a virtual directory that can be used
later as a search path. This is the case when we use remapping, as
demonstrated in the suggested LIT test.

We might encounter a 'false cache miss' and add the same directory
twice into `FileManager::SeenDirEntries` if the added record is a real
directory that is present on a disk:
- Once as a virtual directory 
- And once as a real one

This isn't a problem if the added directories have the same name, as in
this case, we will get a cache hit. However, it could lead to
compilation errors if the directory names are different but point to the
same folder. For example, one might use an absolute name and another a
relative one. For instance, the **implicit-module-remap.cpp** LIT test
will fail with the following message:
```
/.../implicit-module-remap.cpp.tmp/test.cpp:1:2: fatal error: module 'a' 
was built in directory '/.../implicit-module-remap.cpp.tmp' but now 
resides in directory '.'
    1 | #include "a.h"
      |  ^
1 error generated.
```

The suggested fix checks if the added virtual directory is present on
the disk and handles it as a real one if that is the case.

The codegen here differs substantially if UseARMMethodPtrABI is
enabled. Use a fixed x86 triple to avoid this.

https://bugs.llvm.org/show_bug.cgi?id=51735
https://github.com/llvm/llvm-project/issues/51077

In the given test case:
 ```
4 ...
 5 void bar() {
 6   int End = 777;
 7   int Index = 27;
 8   char Var = 1;
 9   for (; Index < End; ++Index)
10     ;
11   nop(Index);
12 }
13 ...
```
Missing local variable `Index` after loop `Induction Variable Elimination`.
When adding a breakpoint at line `11`, LLDB does not have information
on the variable. But it has info on `Var` and `End`.

Do not create icmp/fcmp constant expressions in IRBuilder etc anymore,
i.e. treat them as "undesirable". This is in preparation for removing
them entirely.

Part of:
https://discourse.llvm.org/t/rfc-remove-most-constant-expressions/63179

Revert "[llvm][ScheduleDAG] SUnit::biasCriticalPath() does not find the critical path consistently (#92368)" (#92992)

This reverts commit f0d1ae83.

This commit breaks tests: http://45.33.8.238/linux/138572/step_11.txt

Since C++14 has been released for about nine years and most standard
libraries have implemented sized deallocation functions, it's time to
make this feature default again.

This is another try of https://reviews.llvm.org/D112921.

The original commit cf5a8b48 was reverted by 2e5035ae due to some
failures (see #83774).

Fixes #60061

- Use `Header.Version` directly and remove Header::formatVersion

---------

Co-authored-by: Kazu Hirata <kazu@google.com>

Fixes: SWDEV-459212

We've decided to change `__USING_WASM_EXCEPTIONS__` preprocessor to
`__WASM_EXCEPTIONS__` given that it's more concise.

Disambiguate local functions using the containing file symbol in BAT
mode. Make local function naming consistent across BAT fdata and YAML
profiles.

Test Plan: updated register-fragments-bolt-symbols.s

To align YAML and fdata profiles produced in BAT mode, lift two
restrictions applied in non-relocation mode when BAT is present:
1) register secondary entry points from ignored functions,
2) treat functions with secondary entry points as simple.

This allows constructing CFG for non-simple functions in non-relocation
mode and emitting YAML profile for them, which can then be used for
optimizations in relocation mode.

Test Plan: added test ignored-interprocedural-reference.s

PR #90161 uncovered a bug that annotates C++ xor as UnaryOperator if
followed by a binary operator. This patch fixes that and all other C++
alternative operator keywords when followed by a binary operator in C.

Fixes #92688.

We already have `const uptr kMaxAllowedMallocSize = 1ULL << 40;` set for
ASAN, HWASAN, memprof, TSAN. This patch bumps the malloc limit for MSAN,
LSAN and DFSAN to 1TB as well. 8GB is simply not enough nowadays.

Fix https://github.com/llvm/llvm-project/issues/38037

This reverts commit 253c28fa.

This commit is causing failures on the lldb CI bots, e.g.
https://ci.swift.org/view/all/job/llvm.org/view/LLDB/job/as-lldb-cmake/4307/

On my local macOS desktop build,
```
bin/lldb-dotest -p TestImportBuiltinFileID.py
Assertion failed: (D->getLexicalDeclContext() == this && "Decl inserted into wrong lexical context"), function addHiddenDecl, file DeclBase.cpp, line 1692.

6  libsystem_c.dylib        0x0000000185f0b8d0 abort + 128
7  libsystem_c.dylib        0x0000000185f0abc8 err + 0
8  liblldb.19.0.0git.dylib  0x00000001311e5800 clang::DeclContext::addHiddenDecl(clang::Decl*) + 120
9  liblldb.19.0.0git.dylib  0x00000001311e5978 clang::DeclContext::addDecl(clang::Decl*) + 32
10 liblldb.19.0.0git.dylib  0x000000012f617b48 clang::Sema::ActOnStartTopLevelStmtDecl(clang::Scope*) + 64
11 liblldb.19.0.0git.dylib  0x000000012eaf76c8 clang::Parser::ParseTopLevelStmtDecl() + 208
12 liblldb.19.0.0git.dylib  0x000000012ec051fc clang::Parser::ParseExternalDeclaration(clang::ParsedAttributes&, clang::ParsedAttributes&, clang::ParsingDeclSpec*) + 3412
13 liblldb.19.0.0git.dylib  0x000000012ec03274 clang::Parser::ParseTopLevelDecl(clang::OpaquePtr<clang::DeclGroupRef>&, clang::Sema::ModuleImportState&) + 2020
14 liblldb.19.0.0git.dylib  0x000000012eaca860 clang::ParseAST(clang::Sema&, bool, bool) + 604
15 liblldb.19.0.0git.dylib  0x000000012e8554c0 clang::ASTFrontendAction::ExecuteAction() + 308
16 liblldb.19.0.0git.dylib  0x000000012e854c78 clang::FrontendAction::Execute() + 124
17 liblldb.19.0.0git.dylib  0x000000012e76dcfc clang::CompilerInstance::ExecuteAction(clang::FrontendAction&) + 984
18 liblldb.19.0.0git.dylib  0x000000012e784500 compileModuleImpl(clang::CompilerInstance&, clang::SourceLocation, llvm::StringRef, clang::FrontendInputFile, llvm::StringRef, llvm::StringRef, llvm::function_ref<void (clang::CompilerInstance&)>, llvm::function_ref<void (clang::CompilerInstance&)>)::$_1::operator()() const + 52
```
Reverting until Vassil has a chance to look int oit.

If `::poll` returns and `FD` equals -1, then `ListeningSocket::shutdown`
has been called. So, regardless of any other information that could be
gleaned from `FDs.revents` or `PollStatus`, it is appropriate to return
`std::errc::operation_canceled`. `ListeningSocket::shutdown` copies
`FD`'s value to `ObservedFD` then sets `FD` to -1 before canceling
`::poll` by calling `::close(ObservedFD)` and writing to the pipe.

Given the following metadata as, with `!6` as `ExistingCallbacks` and
`!8` as `NewCB`:
```
!6 = !{!7}
!7 = !{i64 0, i1 false}
!8 = !{i64 2, i64 3, i1 false}
```
The merge function should add `!8` to the list of `!6`, i.e. `!6 =
!{!7,!8}`. However, at the moment the check if this is legal, tries to
interpret `!7` as integer instead of it's operand.

HWASan instrumentation makes writeonly attribute on function parameters,
as well as most memory(*) attributes invalid. This causes
miscompilations with LTO, when more optimizations are run after the
HWASan pass.

Reverts llvm/llvm-project#92942

Addressing #81441

This is an enabler for a future patch.

This allows an type-parameter default argument to be set as an arbitrary
TemplateArgument, not just a type.
This allows template parameter packs to have default arguments in the
AST, even though the language proper doesn't support the syntax for it.

This will be used in a later patch which synthesizes template parameter
lists with arbitrary default arguments taken from template
specializations.

There are a few places we used SubsType, because we only had a type, now
we use SubstTemplateArgument.
SubstTemplateArgument was missing arguments for setting Instantiation
location and entity names.
Adding those is needed so we don't regress in diagnostics.

Addressing #91195

When scudo is built with LLVM-libc's headers, certain functions also
need to be linked from LLVM-libc. This patch adds those functions to the
list to be linked into the specific scudo test, which uses a minimal
subset of libc.

Fixes #92861 and #59453