This reverts commit f79676a1.

`atomic` is required to be followed by a special `atomic clause`, so
this patch manages the parsing of that. We are representing each of the
variants of the atomic construct as separate kinds, because they have
distinct rules/application/etc, and this should make it easier to check
rules in the future.

Update tests in "vector-contract-matvec-transforms.mlir" so that they
are consistent with similar tests in:
 * "vector-contract-to-outerproduct-transforms.mlir".

This is to enable further refactoring in a follow-up patch, namely to:
  * remove duplication (this will be much easier once consistent naming
    is used),
  * extend tests in "vector-contract-matvec-transforms.mlir" with cases
    for scalable vectors,
  * merge "vector-contract-matvec-transforms.mlir" and
    "vector-contract-to-outerproduct-transforms.mlir"  (there's no need
    for 2 different files testing identical transformations).

Overview of changes in this patch:
  1. Simplify the test by removing MemRef wrappers - this test verifies
     Vector -> Vector transformations and  MemRefs are not needed.
  2. Use (m, k) indices instead of (i, j).
  3. Rename function names.

This is part of a larger effort to improve test coverage for scalable
vectors i...

We can use short forward branch to conditionally negate if the
value is negative.

This is part-2 change to improve codegen for vec_fabs. In this patch,
v16f16 and v132f16 fabs are improved.

There will be at least two followups patches after this one.
1) fixing the ISEL crash when fabs.v32f16 uses custom lowering with
AVX512
2) better expansion for v16f16, v32f16 types on AVX1 subtargets.

This showed up when simplifying some large testcase, where the cfi directives
became out of sync with the proc's they enclose.

Now restricted to platforms that support .subsections_via_symbols.

This reverts commit 797b68c0.

Fixes: #72802

Differential revision: https://reviews.llvm.org/D153167

rdar://111459507

[X86] X86FixupVectorConstantsPass - attempt to match VEX logic ops back to EVEX if we can create a broadcast fold

On non-DQI AVX512 targets, X86InstrInfo::setExecutionDomainCustom will convert EVEX int-domain instructions to VEX fp-domain instructions. But, if we have the chance to use a broadcast fold we're better off using a EVEX instruction, so handle a reverse fold.

Sort x86 intrinsics for better readability and use common prefixes to reduce number of comparisons.

From "3.1 Reducing the number of edges" of this [[ https://hal.science/hal-04136674v1/ | paper ]] - Optimization (b)

Task (dependency) nodes have a `successors` list built upon passed dependency.
Given the following code, B will be added to A's successors list building the graph `A` -> `B`
```
// A
 # pragma omp task depend(out: x)
{}

// B
 # pragma omp task depend(in: x)
{}
```

In the following code, B is currently added twice to A's successor list
```
// A
 # pragma omp task depend(out: x, y)
{}

// B
 # pragma omp task depend(in: x, y)
{}
```

This patch removes such dupplicates by checking lastly inserted task in `A` successor list.

Authored by: Romain Pereira (rpereira-dev)
Differential Revision: https://reviews.llvm.org/D158544

Don't use unsigned for sizes as they may be larger than that type can
hold.

Fixes #72678.

3! = 6

Since goma is going away soon, we need to experiment with other
wrappers.

We can clean up the goma stuff once it's fully non-functional.

HLFIR lowering as been set by default now and FIR lowering support will
be removed in the near future. This patch removes the specific FIR check
lines on declare tests.

DPValues are the non-intrinsic replacements for dbg.values, and when an
IR function is converted by SelectionDAG we need to convert the variable
location information in the same way. Happily all the information is in
the same format, it's just stored in a slightly different object,
therefore this patch refactors a few things to store the set of
{Variable,Expr,DILocation,Location} instead of just a pointer to a
DbgValueInst.

This also adds a hook in llc that's much like the one I've added to opt
in PR #71937, allowing tests to optionally ask for the use RemoveDIs
mode if support for it is built into the compiler.

I've added that flag to a variety of SelectionDAG debug-info tests to
ensure that we get some coverage on the RemoveDIs / debug-info-iterator
buildbot.

This reverts commit 42204c94.

It was accidentally backed out.

#20571
#70743

  This patch adds the quadword gather load intrinsics of the form

    sv<type>_t svld1q_gather_u64index_<typ>(svbool_t, const <type>_t *, svuint64_t);
    sv<type>_t svld1q_gather_u64base_index_<typ>(svbool_t, svuint64_t, int64_t);

and the quadword scatter store intrinsics of the form

    void svst1q_scatter_u64index_<typ>(svbool_t, <type>_t *, svuint64_t, sv<type>_t);
    void svst1q_scatter_u64base_index_<typ>(svbool, svuint64_t, int64_t, sv<type>_t);

ACLE spec: https://github.com/ARM-software/acle/pull/257

Opaque ptr cleanup effort.

These two constructs, 'enter data' and 'exit data', are novel compared
to what is currently in the parser, as this is the first set implemented
where the first token is itself not a valid construct. Because of this,
it requires some additional work to do the first keyword parsing.

For a label difference like `.uleb128 A-B`, MC generates a pair of
R_RISCV_SET_ULEB128/R_RISCV_SUB_ULEB128 if A-B cannot be folded as a
constant. GNU assembler generates a pair of relocations in more cases
(when A or B is in a code section with linker relaxation).

`.uleb128 A-B` is primarily used by DWARF v5
.debug_loclists/.debug_rnglists (DW_LLE_offset_pair/DW_RLE_offset_pair
entry kinds) implemented in Clang and GCC.

`.uleb128 A-B` can be used in SHF_ALLOC sections as well (e.g.
`.gcc_except_table`). This patch does not handle SHF_ALLOC.

`-z dead-reloc-in-nonalloc=` can be used to change the relocated value,
if the R_RISCV_SET_ULEB128 symbol is in a discarded section. We don't
check the R_RISCV_SUB_ULEB128 symbol since for the expected cases A and
B should be defined in the same input section.

tan
atan
atanh
sinh
cosh
asin
asinh
acos
acosh
atan2

Fold any constant input to the `s_wqm` intrinsic.

This patch adds a set of SVE2.1 quadword load/store intrisics:

  * Contiguous zero-extending load to quadword (single vector)

    sv<type>_t svld1uwq[_<typ>](svbool_t, const <type>_t *ptr);
    sv<type>_t svld1uwq_vnum[_<typ>](svbool_t, const <type> *ptr, int64_t vnum);
 
    sv<type>_t svld1udq[_<typ>](svbool_t, const <type>_t *ptr);
    sv<type>_t svld1udq_vnum[_<typ>](svbool_t, const <type>_t *ptr, int64_t vnum);

  * Contiguous truncating store of single vector operand

    void svst1uwq[_<typ>](svbool_t, const <type>_t *ptr, sv<type>_t data);
    void svst1uwq_vnum[_<typ>](svbool_t, const <type>_t *ptr, int64_t vnum, sv<type>_t data);

    void svst1udq[_<typ>](svbool_t, const <type>_t *ptr, sv<type>_t data);
    void svst1udq_vnum[_<typ>](svbool_t, const <type>_t *ptr, int64_t vnum, sv<type>_t data);

  * Gather load quadword

    sv<type>_t svld1q_gather[_u64base]_<typ>(svbool_t pg, svuint64_t zn);
    sv<type>_t svld1q_gather[_u64base]_offset_<typ>(svbool_t pg, svuint64_t zn, int64_t offset);

  * Scatter store quadword

    void svst1q_scatter[_u64base][_<typ>](svbool_t pg, svuint64_t zn, sv<type>_t data);
    void svst1q_scatter[_u64base]_offset[_<typ>](svbool_t pg, svuint64_t zn, int64_t offset, sv<type>_t data);

  * Contiguous load two, three or four quadword structures.

    sv<type>x2_t svld2q[_<typ>](svbool_t pg, const <type>_t *rn);
    sv<type>x2_t svld2q_vnum[_<typ>](svbool_t pg, const <type>_t *rn, uint64_t vnum);
    sv<type>x3_t svld3q[_<typ>](svbool_t pg, const <type>_t *rn);
    sv<type>x3_t svld3q_vnum[_<typ>](svbool_t pg, const <type>_t *rn, uint64_t vnum);
    sv<type>x4_t svld4q[_<typ>](svbool_t pg, const <type>_t *rn);
    sv<type>x4_t svld4q_vnum[_<typ>](svbool_t pg, const <type>_t *rn, uint64_t vnum);

  * Contiguous store two, three or four quadword structures.

    void svst2q[_<typ>](svbool_t pg, <type>_t *rn, sv<type>x2_t zt);
    void svst2q_vnum[_<typ>](svbool_t pg, <type>_t *rn, int64_t vnum, sv<type>x2_t zt);
    void svst3q[_<typ>](svbool_t pg, <type>_t *rn, sv<type>x3_t zt);
    void svst3q_vnum[_<typ>](svbool_t pg, <type>_t *rn, int64_t vnum, sv<type>x3_t zt);
    void svst4q[_<typ>](svbool_t pg, <type>_t *rn, sv<type>x4_t zt);
    void svst4q_vnum[_<typ>](svbool_t pg, <type>_t *rn, int64_t vnum, sv<type>x4_t zt);

ACLE spec: https://github.com/ARM-software/acle/pull/257



Co-authored-by: Caroline Concatto <caroline.concatto@arm.com>
Co-authored-by: Hassnaa Hamdi <hassnaa.hamdi@arm.com>

In most cases we do not actually have to perform an index size
adjustment. Don't perform any APInt operations in that case.

Data layout queries may be issued for types whose size exceeds the range
of 32-bit integer as well as for types that don't have a size known at
compile time, such as scalable vectors. Use best practices from LLVM IR
and adopt `llvm::TypeSize` for size-related queries and `uint64_t` for
alignment-related queries.

See #72678.

Add all-gather, all-reduce, all-to-all and reduce-scatter. These
operations have device mesh semantics.

Whether this is valid depends on the bit widths of the involved
integers.

Fixes https://github.com/llvm/llvm-project/issues/72927.

The second test is a miscompile.

Add a test case where the AddRec for the pointers in the inner loop
have the AddRec of the outer loop as start value.

It is sufficient to subtract the start values (%dst, %src) of the outer
AddRecs. This simplification will be done in a follow-up commit.

Lit test generates different outputs for usage of __int128_t in
clang-armv8-quick environment. This patch adds triple to fix the lit
failure.

```
Step 5 (ninja check 1) failure: 1 unexpected failures 38623 expected passes 71 expected failures 36752 unsupported tests (failure)
******************** TEST 'Clang :: Sema/code_align.c' FAILED ******************** Exit Code: 1

Command Output (stderr):
--
RUN: at line 1: /home/tcwg-buildbot/worker/clang-armv8-quick/stage1/bin/clang -cc1 -internal-isystem /home/tcwg-buildbot/worker/clang-armv8-quick/stage1/lib/clang/18/include -nostdsysteminc -fsyntax-only -verify=expected,c-local -x c /home/tcwg-buildbot/worker/clang-armv8-quick/llvm/clang/test/Sema/code_align.c
+ /home/tcwg-buildbot/worker/clang-armv8-quick/stage1/bin/clang -cc1 
+ -internal-isystem 
+ /home/tcwg-buildbot/worker/clang-armv8-quick/stage1/lib/clang/18/inclu
+ de -nostdsysteminc -fsyntax-only -verify=expected,c-local -x c 
+ /home/tcwg-buildbot/worker/clang-armv8-quick/llvm/clang/test/Sema/code
+ _align.c
error: 'c-local-error' diagnostics expected but not seen: 
  File /home/tcwg-buildbot/worker/clang-armv8-quick/llvm/clang/test/Sema/code_align.c Line 79 (directive at /home/tcwg-buildbot/worker/clang-armv8-quick/llvm/clang/test/Sema/code_align.c:78): 'code_align' attribute requires an integer argument which is a constant power of two between 1 and 4096 inclusive; provided argument was (__int128_t)1311768467294899680ULL << 64
  File /home/tcwg-buildbot/worker/clang-armv8-quick/llvm/clang/test/Sema/code_align.c Line 89 (directive at /home/tcwg-buildbot/worker/clang-armv8-quick/llvm/clang/test/Sema/code_align.c:88): 'code_align' attribute requires an integer argument which is a constant power of two between 1 and 4096 inclusive; provided argument was -(__int128_t)1311768467294899680ULL << 64
error: 'c-local-error' diagnostics seen but not expected: 
  File /home/tcwg-buildbot/worker/clang-armv8-quick/llvm/clang/test/Sema/code_align.c Line 79: use of undeclared identifier '__int128_t'
  File /home/tcwg-buildbot/worker/clang-armv8-quick/llvm/clang/test/Sema/code_align.c Line 89: use of undeclared identifier '__int128_t'
4 errors generated.

```

This block of code was here to create pseudo handling of implicit
captures in target regions to prevent gfortran test regressions and
allow certain pieces of code to function, however, with the introduction
of the IFA patch which adds proper handling of implicits by adding them
to the map operands list alongside explicit mappings at the initial
Fortran -> MLIR generation phase this should no longer be required and
may cause some adverse affects at worse in the future.

The current code incorrectly assumed that the absolute variable index
needs to be at least 1, if the variable is != 0. This is incorrect, in
case multiplying with Scale wraps.

The code below already checks for wrapping properly, so just remove the
incorrect assignment.

Fixes https://github.com/llvm/llvm-project/issues/72831.