This recovers more of the max-rss regression introduced by
https://reviews.llvm.org/D157657.

This is an about 1.2% max-rss improvement for the LTO link stage with
debuginfo on CTMark.

Reverts llvm/llvm-project#75196

GCC complains about change of meaning for `FPBits`
```
/home/llvm-libc-buildbot/buildbot-worker/libc-x86_64-debian-fullbuild/libc-x86_64-debian-gcc-fullbuild-dbg/llvm-project/libc/src/__support/FPUtil/generic/FMod.h:188:9: error: declaration of ‘using FPBits = struct __llvm_libc_18_0_0_git::fputil::FPBits<T>’ changes meaning of ‘FPBits’ [-fpermissive]
  188 |   using FPBits = FPBits<T>;
      |         ^~~~~~
```

I'll reland with a different name.

Implement Gram-Schmidt orthogonalisation for the FracMatrix class.
This requires dotProduct, which has been added as a util.

Also make type naming consistent:
 - `UIntType` instead of `intU_t`
 - `FPBits` instead of `FPBits_t`, `FPB`

If the rematerialize was placing a subregister into a super register,
and implicit operands referenced the original register, we need to add
undef flags to the now-subregister indexed implicit operands.

Depends #75152

This patch fixes cmake configuration when building with
LLVM_LIBC_INCLUDE_SCUDO. In libc, LIBC_TARGET_ARCHITECTURE is renamed
from `riscv64` to `riscv`. However, `compiler-rt`, hence `scudo`,
distinguishes `riscv32` and `riscv64` in the support list. As a result,
we need to translate the architecture name accordingly.

Most users of AddImm and CheckConstantInt only use 1 byte immediates, so
I added an opcode variants for those. That way all those instructions
save 7 bytes.
Also added an opcode for AddTempRegister for the cases where there are
no register flags.

Space savings:
    - AMDGPUGenGlobalISel: 470180 bytes to 422564 (-10%)
    - AArch64GenGlobalISel.inc: 383893 bytes to 374046

If the copy being hoisted was undef, we have the same problems that
eliminateUndefCopy needs to solve. We would effectively be introducing a
new live out implicit_def. We need to add an undef flag to avoid
artificially introducing a live through undef value. Previously, the
verifier would fail due to the dead def inside the loop providing the
live in value for the %1 use.

Found in #75288

- `CollectorMetadataAnalysis` provides `GCStrategyMap`.
- `GCFunctionAnalysis` provides `GCFunctionInfo`.

`GCStrategyMap` owns `GCStrategy` pointers and this
pass is used by `AsmPrinter` to iterate all GC strategies.

Most passes that require `GCModuleInfo` actually require the
`GCFunctionInfo`,
so add `GCFunctionAnalysis` for convenience.

Current context set to where alias was met, not where it is declared
caused incorrect access check in case alias referenced private members
of the parent class.
This is a recommit of 6b1aa319 with a slight modification in order to fix
reported regression.

Fixes https://github.com/llvm/llvm-project/issues/41693

See
https://discourse.llvm.org/t/rfc-make-globalisel-match-table-entries-1-byte-instead-of-8/75411

This helps reduce llc's binary size, at the cost of some added
complexity to the MatchTable machinery.

Normally warning is not reported when a field has default initializer.
Do so for anonymous unions with default initializers as well. No release
note since it is a regression in clang 18.

Fixes https://github.com/llvm/llvm-project/issues/70384

It's required for hwasan `aliasing` mode.

Bump to https://github.com/riscv/riscv-cfi/releases/tag/v0.4.0. Actually
there is no functional change here.

This avoids internal libstdc++ leaks.

```
CommutativeBinOp(select(V, A, B), select(V, B, A) --> CommutativeBinOp(A, B)
CommutativeIntrinsicCall(select(V, A, B), select(V, B, A), ...) --> CommutativeIntrinsicCall(A, B, ...)
```

https://alive2.llvm.org/ce/z/8CDUZ4

Closes #73904

Also add links to github actions workflow.

We only allow for assembly code in naked function, and PGO
instrumentation (esp. temporal instrumentation that introduces a
function call) can wreak havoc in this.

Fix #74573

Fixes sanitizer-ppc64le-linux bot.

Currently on mcpu=v3 we do not support sdiv, srem instructions. And the
backend crashes with stacktrace & coredump, which is misleading for end
users, as this is not a "bug"

Add llvm bug reporting for sdiv/srem on ISel legalize-op phase.

For clang frontend we can get detailed location & bug report.

    $ build/bin/clang -g -target bpf -c local/sdiv.c
local/sdiv.c:1:35: error: unsupported signed division, please convert to
unsigned div/mod.
        1 | int sdiv(int a, int b) { return a / b; }
          |                                   ^
    1 error generated.

Fixes: #70433
Fixes: #48647

This also improves error handling for dynamic stack allocation:

    local/vla.c:2:3: error: unsupported dynamic stack allocation
        2 |   int b[n];
          |   ^
    1 error generated.

Fixes: https://github.com/llvm/llvm-project/issues/57171

Reverts llvm/llvm-project#75104

Original commit causes clang to generate invalid IR:
```
mismatched subprogram between llvm.dbg.declare variable and !dbg attachment
  call void @llvm.dbg.declare(metadata ptr %4, metadata !34468, metadata !DIExpression(DW_OP_plus_uconst, 176)), !dbg !34467
```

This commit introduces basic annotations for `std::basic_string`,
mirroring the approach used in `std::vector` and `std::deque`.
Initially, only long strings with the default allocator will be
annotated. Short strings (_SSO - short string optimization_) and strings
with non-default allocators will be annotated in the near future, with
separate commits dedicated to enabling them. The process will be similar
to the workflow employed for enabling annotations in `std::deque`.

**Please note**: these annotations function effectively only when libc++
and libc++abi dylibs are instrumented (with ASan). This aligns with the
prevailing behavior of Memory Sanitizer.

To avoid breaking everything, this commit also appends
`_LIBCPP_INSTRUMENTED_WITH_ASAN` to `__config_site` whenever libc++ is
compiled with ASan. If this macro is not defined, string annotations are
not enabled. However, linking a binary that does **not** annotate
strings with a dynamic library that annotates strings, is not permitted.

Originally proposed here: https://reviews.llvm.org/D132769

Related patches on Phabricator:
- Turning on annotations for short strings:
https://reviews.llvm.org/D147680
- Turning on annotations for all allocators:
https://reviews.llvm.org/D146214

This PR is a part of a series of patches extending AddressSanitizer C++
container overflow detection capabilities by adding annotations, similar
to those existing in `std::vector` and `std::deque` collections. These
enhancements empower ASan to effectively detect instances where the
instrumented program attempts to access memory within a collection's
internal allocation that remains unused. This includes cases where
access occurs before or after the stored elements in `std::deque`, or
between the `std::basic_string`'s size (including the null terminator)
and capacity bounds.

The introduction of these annotations was spurred by a real-world
software bug discovered by Trail of Bits, involving an out-of-bounds
memory access during the comparison of two strings using the
`std::equals` function. This function was taking iterators
(`iter1_begin`, `iter1_end`, `iter2_begin`) to perform the comparison,
using a custom comparison function. When the `iter1` object exceeded the
length of `iter2`, an out-of-bounds read could occur on the `iter2`
object. Container sanitization, upon enabling these annotations, would
effectively identify and flag this potential vulnerability.

This Pull Request introduces basic annotations for `std::basic_string`.
Long strings exhibit structural similarities to `std::vector` and will
be annotated accordingly. Short strings are already implemented, but
will be turned on separately in a forthcoming commit. Look at [a
comment](https://github.com/llvm/llvm-project/pull/72677#issuecomment-1850554465)
below to read about SSO issues at current moment.

Due to the functionality introduced in
[D132522](https://github.com/llvm/llvm-project/commit/dd1b7b797a116eed588fd752fbe61d34deeb24e4),
the `__sanitizer_annotate_contiguous_container` function now offers
compatibility with all allocators. However, enabling this support will
be done in a subsequent commit. For the time being, only strings with
the default allocator will be annotated.

If you have any questions, please email:
- advenam.tacet@trailofbits.com
- disconnect3d@trailofbits.com

Add test for https://github.com/llvm/llvm-project/issues/75115

They already include workarounds.

This reverts commit aa217eb0.
This reverts commit 634feddc.

This breaks buildbots by introducing cycle dependency between libObject
and TextAPI and breaks gcc compiles on buildbots.

* Reported in buildbot failures