Summary:
This patch reworks the CMake handling for building plugins. All this
does is pull a lot of shared and common logic into a single helper
function.
This also simplifies the OMPT libraries from being built separately
instead of just added.

The DW_AT_abstract_origin can be a cross-cu reference as a by-product of
LTO. On IR level for absolute references an address is stored, vs a DIE
for relative references. Added a map to keep track of cross-cu
referenced DIEs to use when we add an Entry.

For DWARF5 BOLT was not retreiving address and instead was setting an
index.
Changed so that an address is used, and added DWARF4 test because it was
missing.

The greendragon was recently moved and now it runs on somewhat newer
macOS
version - which breaks some sanitizers tests

rdar://125052915



Co-authored-by: Mariusz Borsa <m_borsa@apple.com>

This PR extracts `FIROpConversion` and `FIROpAndTypeConversion`
templated base patterns to a header file. All the functions from
FIROpConversion that do not require the template argument are moved to a
base class named `ConvertFIRToLLVMPattern`.
This move is done so the `FIROpConversion` pattern and all its utility
functions can be reused outside of the codegen pass.

For the most part the code is only moved to the new files and not
modified. The only update is that addition of the PatternBenefit
argument with a default value to the constructor so it can be forwarded
to the `ConversionPattern` ctor.

This split is done in a similar way for the `ConvertOpToLLVMPattern`
base pattern that is based on the `ConvertToLLVMPattern` base class in
`mlir/include/mlir/Conversion/LLVMCommon/Pattern.h`.

When adding fixups for RISCV_TLSDESC_ADD_LO and RISCV_TLSDESC_LOAD_LO,
the local label added for RISCV TLSDESC relocations have STT_TLS set,
which is incorrect. Instead, these labels should have `STT_NOTYPE`.

This patch stops adding such fixups and avoid setting the STT_TLS on
these symbols. Failing to do so can cause LLD to emit an error `has an
STT_TLS symbol but doesn't have an SHF_TLS section`. We additionally,
adjust how LLD services these relocations to avoid errors with
incompatible relocation and symbol types.

Reviewers: topperc, MaskRay

Reviewed By: MaskRay

Pull Request: https://github.com/llvm/llvm-project/pull/85817

This patch implements `affinity` for AIX, which is quite different from
platforms such as Linux.
- Setting CPU affinity through masks and related functions are not
supported. System call `bindprocessor()` is used to bind a thread to one
CPU per call.
- There are no system routines to get the affinity info of a thread. The
implementation of `get_system_affinity()` for AIX gets the mask of all
available CPUs, to be used as the full mask only.
- Topology is not available from the file system. It is obtained through
system SRAD (Scheduler Resource Allocation Domain).

This patch has run through the libomp LIT tests successfully with
`affinity` enabled.

![image](https://github.com/llvm/llvm-project/assets/20108837/94cb5718-3526-4bae-8a79-f5b1d19b352d)

The llvm.readcyclecounter intrinsic can be implemented via the
STORE CLOCK FAST (STCKF) instruction.

This reverts commit 1fe9c417.

This relands commit 6bbe8a29.

This was causing build failures on one of the ARMv8 builders. Still not
completely sure why, but relanding it to see if the failure pops up
again. If it does, the plan is to fix forward by disabling tests on ARM
temporarily as llvm-exegesis does not currently use SubprocessMemory
on ARM.

Reverting for
https://github.com/llvm/llvm-project/pull/86297#issuecomment-2015660662

This reverts commit ab8ace3b.

Actions triggered by pull_request_target events have access to all
repository secrets, so it is unsafe to use them when executing untrusted
code. The pr-code-format workflow does not execute any untrusted code,
but it passes untrused input into clang-format. An attacker could use
this to exploit a flaw in clang-format and potentially gain access to
the repository secrets.
    
By splitting the workflow, we can use the pull_request target which is
more secure and isolate the issue write permissions in a separate job.
The pull_request target also makes it easier to test changes to the
code-format-helepr.py script, because the version of the script from the
pull request will be used rather than the version of the script from
main.
    
Fixes #77142

Registers shouldn't depend on the scheduler, but a scheduler
predicate could depend on a register.

This would make it possible to move VLDSX0Pred out of the SiFive7
scheduler model to RISCVSchedule.td if another model needed it.

Follow up to #85438.

Implements the functions `strfromd()` and `strfroml()` introduced in
C23, and unifies the testing framework for `strfrom*()` functions.

Im working on the floating point fmaximum and fminimum functions right
now in a different pr and I ran the individual tests by doing

```
ninja libc.test.src.math.smoke.<func>_test.__unit__
```

@GeorgeHuyubo noticed an unchecked shared pointer result in
https://github.com/llvm/llvm-project/pull/85693/

. This is the fix for
that issue.

Co-authored-by: Kevin Frei <freik@meta.com>

I'm planning to add HwMode support to SubRegIdxRanges for RISC-V GPR
pairs. The MC layer is currently unaware of the HwMode for registers and
I'd like to keep it that way.

This information is not used by the MC layer so I think it is safe to
move it.

The answer to many requests in issues to be assigned to users is often
"just create a pull request". That's in contradiction to the
"introductory issue" instructions posted by the GitHub bot.

This change updates the instructions, mentioning the shortcut of "just
creating a PR". Moreover, it now explains linking PRs to issues in order
to close them automatically upon merge.

The previous implementation used a `notifyMatchFailure` to emit failure
message inappropriately and then used the
`emitDefaultSilenceableFailure`. This patch changes this to use the more
appropriate `emitSilenceableFailure` with error message. Additionally a
failure test has been added.

This reverts commit c3a41aac.

This relands commit bd493756.

Apparently I forgot to update a couple values, so this change failed on
every builder that builds those sections (should be every Linux
platform) rather than something architecture specific like originally
thought. I swore I updated the values and ran check-llvm before merging.
Wondering If I forgot to push those changes...

This reverts commit bd493756.

Causes build failures on non-X86 platforms.
https://lab.llvm.org/buildbot/#/changes/128363

This allows to avoid duplicating the code handling the instructions
outputted by `generateInstSeq` when emitting `MCInst`s.

This patch refactors the parent code to a separate function in the
subprocess executor to make the code more clear and easy to follow.

This bazel flag, that should be flipped in an upcoming release
https://github.com/bazelbuild/bazel/pull/15327, fails if globs have no
matches. This helps find libraries where you are accidentally not
including files because of typos. This change removes the various globs
that were not matching anything, and uncovered some targets that were
doing nothing because their source files were deleted. There are a few
cases where globs were intentionally optional in the case of loops that
expanded to different potential options, so those now use `allow_empty =
True`. This allows downstream consumers to also flip this flags for
their own builds, where previously this would fail in LLVM instead.

The downside to this change is that if files are added in these
relatively standard locations, manual work will have to be done to add
this patterns back. If folks prefer we could instead add `allow_empty =
True` to every glob.

This paper is about constant expression handling in the presence of
FLT_EVAL_METHOD, which we handle via insertion of implicit cast nodes
in the AST.

Fixed a small issue of matching pthread signature, which was causing the
build to break for the compiler-rt project after adding
-Wcast-function-type-mismatch to -Wextra dignostic group
(https://github.com/llvm/llvm-project/pull/77178 &
https://github.com/llvm/llvm-project/pull/86131).

Refactor managed variable handling in codegen so that the transformation
is done separately from registration.

This will allow the new driver to register the managed var in the linker
wrapper.

This reverts commit 2091c747.

Builtbot failure: https://lab.llvm.org/buildbot/#/builders/16/builds/63080

Follow-up for https://github.com/llvm/llvm-project/pull/81625

Fixs https://github.com/llvm/llvm-project/issues/83079

Make the name of a clang builtin as close to the mnemonic instruction
name as possible. The data type suffix may not be enough to tell what
instruction the builtin is going to produce.
  This patch also add the bf16 support for global_load_tr_b128 builtins.

Follow on from #84915 which adds the DbgRecord function variants.

Update the LLVMDIBuilderInsert... functions to insert DbgRecords instead
of debug intrinsics.

    LLVMDIBuilderInsertDeclareBefore
    LLVMDIBuilderInsertDeclareAtEnd
    LLVMDIBuilderInsertDbgValueBefore
    LLVMDIBuilderInsertDbgValueAtEnd

Calling these functions will now cause an assertion if the module is in the
wrong debug info format. They should only be used when the module is in "new
debug format".

Use LLVMIsNewDbgInfoFormat to query and LLVMSetIsNewDbgInfoFormat to change the
debug info format of a module.

Please see https://llvm.org/docs/RemoveDIsDebugInfo.html#c-api-change
(RemoveDIsDebugInfo.md) for more info.

YAML profile reader checks the number of basic blocks in regular,
no-stale-matching mode. Add it to BAT.

This increases the size of BAT section to:
- large binary: 39583080 bytes (1.02x of the original),
- medium binary: 3816492 bytes (0.64x),
- small binary: 920 bytes (0.64x, no change due to alignment).

Test Plan: Updated bolt-address-translation-yaml.test

Reviewers: rafaelauler, ayermolo, maksfb, dcci

Reviewed By: rafaelauler

Pull Request: https://github.com/llvm/llvm-project/pull/86045

This commit resolves a SROA bug caused by not properly checking if a
llvm store operation writes the pointer to memory or not. Now, we do no
longer consider stores that use a slot pointer as a value to store as
fixable.

Add input basic block index to BAT metadata. This addresses the case
where some basic blocks are eliminated, and output index is not equal
to the input block index. These indices are used in non-stale-matching
mode.

Increases BAT section size to:
- large binary: 39521512 bytes (1.02x original),
- medium binary: 3799988 bytes (0.64x),
- small binary: 920 bytes (0.64x).

Test Plan:
Updated bolt-address-translation{,-yaml}.test

Pull Request: https://github.com/llvm/llvm-project/pull/86044

This reverts commit 3b3de48f.

Validate `p_offset` in `dynamicEntries` before computing the entry offset.

Fixes: https://github.com/llvm/llvm-project/issues/85568.