This PR fixes validity of atomic instructions and improves type
inference. More tests are able now to be accepted by `spirv-val`.

Fixes #85209

This patch removes the truncation from `hash_code` aka `size_t` down to
`unsigned`, that currently happens on DenseMap accesses in
RegisterBankInfo. This reduces the likelihood of hash collisions, as
well as the likelihood of hitting EmptyKey or TombstoneKey, the special
key values of DenseMap. This is not the ultimate solution to the
problem, but we can do it in any case.

The ExternalNameConversion pass can be surprisingly slow on big
programs. On an example with a 50kloc Fortran file with about 10000
calls to external procedures, the pass alone took 25s on my machine.
This patch reduces this to 0.16s.

The root cause is that using `replaceAllSymbolUses` on each modified
FuncOp is very expensive: it is walking all operations and attribute
every time.

An alternative would be to use mlir::SymbolUserMap to avoid walking the
module again and again, but this is still much more expensive than what
is needed because it is essentially caching all symbol uses of the
module, and there is no need to such caching here.

Instead:
- Do a shallow walk of the module (only top level operation) to detect
FuncOp/GlobalOp that needs to be updated. Update them and place the name
remapping in a DenseMap.
- If any remapping were done, do a single deep walk of the module
operation, and update any SymbolRefAttr that matches a name that was
remapped.

This patch fixes the crash caused by the pull request:
https://github.com/llvm/llvm-project/pull/85592

The recent change to split the PR code formatting job accidentally
misspelled the repository field when specifying the repository to fetch
the code formatting utils from. This patch fixes the spelling so that
the job does not throw a warning and clones the tools from the specified
repository.

Handling of the `BIenqueue_kernel` builtin must not fallthrough to the
`BIget_kernel_work_group_size` builtin, as these builtins have no common
functionality.

preliminary steps

Until now function `fseek` returned nonzero on error, this is changed to
-1 only. And it does not produce EOF error any more.
This complies better with the POSIX standard.

Argument-register pairs in CallSiteInfo is only needed when
EmitCallSiteInfo
is on. Currently, the pairs are always pushed to the vector but only
used
when EmitCallSiteInfo is on.

Don't fill the CallSiteInfo vector unless used.

Differential Revision: https://reviews.llvm.org/D107108?id=362887



Co-authored-by: Necip Fazil Yildiran <necip@google.com>

We are missing patterns for fixed vectors, where the sexts and zexts are
legalized to _vl nodes.

Dialect conversion sometimes can have a hanging use of an argument.
Ensured that argument uses are dropped before removing the block.

Turn `RewriterBase::replaceAllUsesWith` into a non-templatized
implementation, so that it can be made virtual and be overridden in the
`ConversionPatternRewriter` in a subsequent change.

This change is in preparation of adding dialect conversion support for
`replaceAllUsesWith`.

Required in the review process of
https://github.com/llvm/llvm-project/pull/85050.

Before this change: `notifyOperationReplaced` was triggered when calling
`RewriteBase::replaceOp`.
After this change: `notifyOperationReplaced` is triggered when
`RewriterBase::replaceAllOpUsesWith` or `RewriterBase::replaceOp` is
called.

Until now, every `notifyOperationReplaced` was always sent together with
a `notifyOperationErased`, which made that `notifyOperationErased`
callback irrelevant. More importantly, when a user called
`RewriterBase::replaceAllOpUsesWith`+`RewriterBase::eraseOp` instead of
`RewriterBase::replaceOp`, no `notifyOperationReplaced` callback was
sent, even though the two notations are semantically equivalent. As an
example, this can be a problem when applying patterns with the transform
dialect because the `TrackingListener` will only see the
`notifyOperationErased` callback and the payload op is dropped from the
mappings.

Note: It is still possible to write semantically equivalent code that
does not trigger a `notifyOperationReplaced` (e.g., when op results are
replaced one-by-one), but this commit already improves the situation a
lot.

Use the destination data type to measure the LMUL size for
latency/throughput cost

Update the test to run when the compiler is built to support
arm64-darwin targets.

Use the return type to measure the LMUL size for throughput/latency cost

GatherAllAliases is time consuming. Add an debug option on PPC to
control the complexity of the function. This is useful when debuging
compile time related issues.

After #87191 we had to add
8b135a7d, which
makes symbolizer to calls a global constructor
with `realloc`.

This change checks if there is already a `deps` key in `kwargs` and
concatenate it to avoid multiple values for `deps` key argument.

background:
https://github.com/llvm/llvm-project/pull/87022 recently added explicit
`deps` to the lit_test. This is causing StableHLO bazel build failures
at
https://github.com/openxla/stablehlo/actions/runs/8511888283/job/23312383380?pr=2147

Tested: local build run is successful

The code does some (overly simple?) checks on file syntax. These checks
assume unix line endings and fail on windows. This commit updates the
code to strip extra whitespace, making the checks more robust,
particularly in the presence of windows line endings.

Fixes #86023

There is no good way to report detailed errors from inside
`Pass::initializeOptions` function as context may not be available at
this point and writing directly to `llvm::errs()` is not composable.

See
https://github.com/llvm/llvm-project/pull/87166#discussion_r1546426763

* Add error handler callback to `Pass::initializeOptions`
* Update `PassOptions::parseFromString` to support custom error stream
instead of using `llvm::errs()` directly.
* Update default `Pass::initializeOptions` implementation to propagate
error string from `parseFromString` to new error handler.
* Update `MapMemRefStorageClassPass` to report error details using new
API.

These are probably actually unreachable - perhaps an lldb developer
would be interested in rephrasing this change to move the new cases into
some unreachable/unsupported bucket, rather than my half-hearted guess
at what the desired behavior would be (completely untested, because
they're probably untestable/unreachable - maybe debugging from modules?)

Summary:
We want to pass these GPU libraries by default if a certain offloading
toolchain is loaded for OpenMP. Previously I parsed this from the
arguments because it's only available in the compilation. This doesn't
really work for `native` and it's extra effort, so this patch just
passes in the `Compilation` as an extr argument and uses that. Tests
should be unaffected.

Fixes #87102 .

Add annotated vtable GUID as referenced variables in per function
summary, and update bitcode writer to create value-ids for these
referenced vtables.

- This is the part3 of type profiling work, and described in the "Virtual Table Definition Import" [1] section of the
RFC.

[1] https://github.com/llvm/llvm-project/pull/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

The docstring for this function incorrectly specified when a widening is
not performed. This patch adds the additional specification for what
happens when the type size is a power of two but it is less than MinSize.

Finish the rename by #86604

This will eliminate some unnecessary REST API calls.

Use a seek offset that fits within the file size.

This was missed in presubmit because the FILE based stdio tests aren't
run in
overlay mode; fullbuild is not tested in presubmit.

WRITE_SIZE == 11, so using a value of 42 for offseto would cause the
expression
`WRITE_SIZE - offseto` to evaluate to -31 as an unsigned 64b integer
(18446744073709551585ULL).

Fixes #86928

We check if the next character after `N.` is `*` before we check its
length. Using `split` on the string is cleaner and less error prone than
using indices with `find` and `substr`.

Note: this does not make `N.` mean anything, it just prevents assertion
failures. `N.` is treated the same as an unrecognized breakpoint name:

```
(lldb) breakpoint enable 1
1 breakpoints enabled.
(lldb) breakpoint enable 1.*
1 breakpoints enabled.
(lldb) breakpoint enable 1.
0 breakpoints enabled.
(lldb) breakpoint enable xyz
0 breakpoints enabled.
```

Found via LLDB fuzzers.

I was profiling our compiler and noticed that `llvm::get_threadid` was
at the top of the hotlist, taking up a surprising 5% (7 seconds) in the
profile trace. It seems that computing this on MacOS systems is
non-trivial, so cache the result in a thread_local.

Co-authored-by: Mehdi Amini <joker.eph@gmail.com>

As before we only skip instrumenting probe of `unwind`(`KnownColdBlock`)
block, this PR extends to skip the both EH flow from `invoke`, i.e. also
skip the `normal` dest. For more contexts: when doing call-to-invoke
conversion, the block is split by the `invoke` and two extra
blocks(`normal` and `unwind`) are added. With this PR, the
instrumentation is the same as the one before the call-to-invoke
conversion.
One significant benefit is this can help mitigate the "unstable IR"
issue(https://discourse.llvm.org/t/ipo-for-linkonce-odr-functions/69404),
the two versions now are on the same probe instrumentation, expected to
be the same checksum.
To achieve the same checksum, some tweaks is needed:
-  Now it also skips incrementing the probe ID for the skipped probe.
- The checksum is also computed based on the CFG that skips the EH
edges.

We observed this fixes ~5% mismatched samples.

Extracted from #84858

Clean-up all of the calls and remove the redundant == 0 checks.

There is only one small visible change. For non-Android, the memalign
function will now fail if alignment is zero. Before this would have
passed.

This fixes #87227, a vbase can be placed below nvsize when empty members and/or bases are in play. We must account for that.

The ``dataclasses`` package makes sense for Python 3.6, becauses
``dataclasses`` is only included in the standard library with 3.7
version. Now, 3.6 has reached EOL, so all current supported versions of
Python (3.8, 3.9, 3.10, 3.11, 3.12) have this feature in their standard
libraries.

Therefore there's no need to install the ``dataclasses`` package now.