Let the std.compat module use the std module instead of duplicating the
exports.

Based on @ChuanqiXu9's suggestion in #71438.

As suggested in #71438 we should use
  export import std;
in the std.compat module.

Using this exports some named declarations from functions and records,
adding them to the global namespace. Clang correctly, does not export
these it's and issue in the declaration filtering. Declarations in
function or record context are not considered a global named
declaration.

As suggested in #71438 we should use
  export import std;
in the std.compat module.

Testing this locally failed when building with the clang-tidy-17
plugin. The std module was considered corrupt in the test
  libcxx/test/libcxx/module_std_compat.gen.py
however the test
  libcxx/test/libcxx/module_std.gen.py
passed. Both test generated identical std.pcm files. Using the
clang-tidy-18 plugin solves the issue.

Co-authored-by: Will Hawkins <hawkinsw@obs.cr>

Co-authored-by: Will Hawkins <hawkinsw@obs.cr>

This adds a new module test infrastructure. This requires tagging tests
using modules. The test runner uses this information to determine the
compiler flags needed to build and use the module.

Currently modules are build per test, which allows testing them for
tests with ADDITIONAL_COMPILE_FLAGS. At the moment only 4 tests use
modules. Therefore the performance penalty is not measurable. If in the
future more tests use modules it would be good to measure the overhead
and determine whether it's acceptable.

This removes the entire modules testing infrastructure.

The current infrastructure uses CMake to generate the std and std.compat
module. This requires quite a bit of plumbing and uses CMake. Since
CMake introduced module support in CMake 3.26, modules have a higher
CMake requirement than the rest of the LLVM project. (The LLVM project
requires 3.20.) The main motivation for this approach was how libc++
generated its modules. Every header had its own module partition. This
was changed to improve performance and now only two modules remain.
The code to build these can be manually crafted.

A followup patch will reenable testing modules, using a different
approach.

Extend the logic check if an operand of an AND is implied by the other
to also support OR. This is done by checking if !op1 implies op2 or vice
versa.

Introduce RecordVisitor. This is used for different clients that want to
extract information out of RecordSlice types.
The first and immediate use case is for serializing symbol information
into TBD files.

As part of startup refactoring, this patch adds a function to merge
multiple objects into a single relocatable object:
                     cc -r obj1.o obj2.o -o obj.o

A relocatable object is an object file that is not fully linked into an
executable or a shared library. It is an intermediate file format that
can be passed into the linker.

A crt object can have arch-specific code and arch-agnostic code. To
reduce code cohesion, the implementation is splitted into multiple
units. As a result, we need to merge them into a single relocatable
object.

Summary:
Fix some linting warnings present in the plugins.

emitPopInst checks a single function exit MBB. If other paths also exit
the function and any of there terminators uses LR implicitly, it is not
save to clear the Restored bit.

Check all terminators for the function before clearing Restored.

This fixes a mis-compile in outlined-fn-may-clobber-lr-in-caller.ll
where the machine-outliner previously introduced BLs that clobbered LR
which in turn is used by the tail call return.

Alternative to #73553

This introduces assembly support for the Checked Pointer Arithmetic
Extension (FEAT_CPA), annouced as part of the Armv9.5-A architecture
version.

The changes include:
* New subtarget feature for FEAT_CPA
* New scalar instruction for pointer arithmetic
  * ADDPT, SUBPT, MADDPT, and MSUBPT
* New SVE instructions for pointer arithmetic
  * ADDPT (vectors, predicated), ADDPT (vectors, unpredicated)
  * SUBPT (vectors, predicated), SUBPT (vectors, unpredicated)
  * MADPT and MLAPT
* New ID_AA64ISAR3_EL1 system register

Mode details about the extension can be found at:
* https://community.arm.com/arm-community-blogs/b/architectures-and-processors-blog/posts/arm-a-profile-architecture-developments-2023
* https://developer.arm.com/documentation/ddi0602/2023-09/



Co-authored-by: Rodolfo Wottrich <rodolfo.wottrich@arm.com>

#74652 adds `__buildid` symbol which allows us to dump it at runtime.

[X86] SimplifyDemandedVectorEltsForTargetShuffle - don't simplify constant mask if it has multiple uses

Avoid generating extra constant vectors

The `acc` dialect operations now implement MemoryEffects interfaces in
the following ways:
- Data entry operations which may read host memory via `varPtr` are now
marked as so. The majority of them do NOT actually read the host memory.
For example, `acc.present` works on the basis of presence of pointer and
not necessarily what the data points to - so they are not marked as
reading the host memory. They still use `varPtr` though but this
dependency is reflected through ssa.
- Data clause operations which may mutate the data pointed to by
`accPtr` are marked as doing so.
- Data clause operations which update required structured or dynamic
runtime counters are marked as reading and writing the newly defined
`RuntimeCounters` resource. Some operations, like `acc.getdeviceptr` do
not actually use the runtime counters - but are marked as reading them
since the address obtained depends on the mapping operations which do
update the runtime counters. Namely, `acc.getdeviceptr` cannot be moved
across other mapping operations.
- Constructs are marked as writing to the `ConstructResource`. This may
be too strict but is needed for the following reasons: 1) Structured
constructs may not use `accPtr` and instead use `varPtr` - when this is
the case, data actions may be removed even when used. 2) Unstructured
constructs are currently used to aggregate multiple data actions. We do
not want such constructs removed or moved for now.
- Terminators are marked as `Pure` as in other dialects.

The current approach has the following limitations which may require
further improvements:
- Subsequent `acc.copyin` operations on same data do not actually read
host memory pointed to by `varPtr` but are still marked as so.
- Two `acc.delete` operations on same data may not mutate `accPtr` until
the runtime counters are zero (but are still marked as mutating).
- The `varPtrPtr` argument, when present, points to the address of
location of `varPtr`. When mapping to target device, an `accPtrPtr`
needs computed and this memory is mutated. This effect is not captured
since the current operations do not produce `accPtrPtr`.
- Runtime counter effects are imprecise since two operations with
differing `varPtr` increment/decrement different counters. Additionally,
operations with `varPtrPtr` mutate attachment counters.
- The `ConstructResource` is too strict and likely can be relaxed with
better modeling.

Functions using different constant expressions were incorrectly
merged, because a lot of state was missing from the comparison,
including the opcode, the comparison predicate, the GEP element
type, as well as the inbounds, inrange and nowrap poison flags.

Also set HAVE_DLADDR to 1 on non-Win instead of just on macOS.
That looked like an oversight.

If decided to resize the instruction, need to drop wrapping flags from
the resulting vector instructions to avoid incorrect
optimizations/assumptions later.
Fixes PR75995.

The way this code was updated in
dd958779 meant that if the first module
did not have the symbol, the iteration stopped as returning true means
stop. So only if every module had the symbol would we find it, in the
last module.

Invert the condition to break when we find the first instance, which is
what the previous code did.

Add isel pattern to replase min(max(v1,v2),v3) by clamp
Add tests for uclamp, sclamp, bfclamp, fclamp.

minbitwidth types.

When the sub arguments are ptr2int it is not possible to determine
computeKnownBits() of its arguments.
For scalar case generally sub of 2 ptr2int are converted to sub of
indexes.
However a loop with recursive GEP/PHI where the arguments to sub is of
type ptr2int, if it is possible to determine that a sub of this GEP and
another pointer with the same base is KnownNonZero we can return this.
This helps subsequent passes to optimize the loop further.

Otherwise inbounds will be inferred, and we don't actually end
up testing the case of one gep without inbounds and one with.

This reverts commit 01c4ecb7,
d14d5215 and
a756dc47.

This removes the logging and workaround I added earlier,
and puts back the skip for Arm/AArch64 Linux.

I've not seen it fail on AArch64 since, but let's not create
more noise if it does.

I've written up the issue as https://github.com/llvm/llvm-project/issues/76057.
It's something to do with trying to destroy a process while
a thread is doing a single sep. So my workaround wouldn't have
worked in any case. It needs a more involved fix.

Looks like we don't even check the opcode :(

Use llvm-config.h in CIndexer.cpp instead of private header

Add an emitc.expression operation that models C expressions, and provide
transforms to form and fold expressions. The translator emits the body
of
emitc.expression ops as a single C expression.
This expression is emitted by default as the RHS of an EmitC SSA value,
but if
possible, expressions with a single use that is not another expression
are
instead inlined. Specific expression's inlining can be fine tuned by
lowering
passes and transforms.

Fix bitcast test, which was splitting apart phis intended to force
bitcasts that survive all the way to selection.

Disable the amdgpu-codegenprepare phi splitting, which defeats the technique
of using a phi to ensure a bitcast reaches all the way to selection. Also
add a variety of bfloat tests. These probably need revisiting to avoid the
cast folding into argument loads. Also round out set of bfloat bitcast and
ABI tests.

Add codegen tests for more bf16 operations The promotion of these works
contrary to the comment.

Also fix assorted vector extload failures for AMDGPU.

This makes an adjustment to the existing fir minloc/maxloc generation
code to handle functions with a dim=1 that produce a scalar result. This
should allow us to get the same benefits as the existing generated
minmax reductions.