Describe in the BackEnd Developer's Guide. Instrument a few backends.

Remove an old unused timing facility. Add a null backend for timing
the parser.

Differential Revision: https://reviews.llvm.org/D91388

This was changed recently with D90554 / f7eac51b
...because we had a regression testing blindspot for intrinsics
that are expected to be lowered to libcalls.

In general, we want the *size* cost for a scalar call to be cheap
even if the other costs are expensive - we expect it to just be
a branch with some optional stack manipulation.

It is likely that we will want to carve out some
exceptions/overrides to this rule as follow-up patches for
calls that have some general and/or target-specific difference
to the expected lowering.

This was noticed as a regression in unrolling, so we have a test
for that now along with a couple of direct cost model tests.

If the assumed scalarization costs for the oversized vector
calls are not realistic, that would be another follow-up
refinement of the cost models.

This simplifies the code the allocator has to executed

Differential Revision: https://reviews.llvm.org/D91471

Clang supports compiling CUDA source files,
CUDA header files may contain CUDA specific code
that is why they have special extension, which
can be recognized by nvcc (CUDA compiler driver)
as CUDA source file.
Format them by default as well.

Reviewed By: MyDeveloperDay

Patch By: tomilov

Differential Revision: https://reviews.llvm.org/D90780

Differential Revision: https://reviews.llvm.org/D91431

See discussion in https://bugs.llvm.org/show_bug.cgi?id=45073 / https://reviews.llvm.org/D66324#2334485
the implementation is known-broken for certain inputs,
the bugreport was up for a significant amount of timer,
and there has been no activity to address it.
Therefore, just completely rip out all of misexpect handling.

I suspect, fixing it requires redesigning the internals of MD_misexpect.
Should anyone commit to fixing the implementation problem,
starting from clean slate may be better anyways.

This reverts commit 7bdad084,
and some of it's follow-ups, that don't stand on their own.

Tested 0.018s faster on my machine.

Make test values suitable for PC alignment arithmetic used
in StackTrace::Print().

These currently fail to select fsrw/fslw.

This unit test code was using malloc without a corresponding free.
When the system malloc is not being overridden by the code under
test, it might an asan/lsan allocator that notices leaks.

Reviewed By: phosek

Differential Revision: https://reviews.llvm.org/D91472

Bug: https://bugs.llvm.org/show_bug.cgi?id=48008

Differential Revision: https://reviews.llvm.org/D91037

This reverts commit e7ed2765.

Build is broken with  -DLLVM_LINK_LLVM_DYLIB=ON

Windows doesn't support remote connections.

motivated by a refactoring in the new sparse code (yet to be merged), this avoids some lengthy code dup

Reviewed By: mehdi_amini

Differential Revision: https://reviews.llvm.org/D91465

This library is only used in Flang at the moment and not tested withing LLVM.
Having it as a component is breaking llvm-config:

  $ bin/llvm-config --shared-mode
  llvm-config: error: component libraries and shared library

  llvm-config: error: missing: [...]/lib/libLLVMFrontendOpenACC.a

This will reverted when unit-tests are provided for it.

Reviewed By: clementval

Differential Revision: https://reviews.llvm.org/D91470

I think the check for whether the process is connected is totally bogus
in the first place, but on the off-chance that's it's not, we should
behave the same in synchronous and asynchronous mode.

This assertion ensures the input value isn't part of the vector when
growing is required. In such cases the vector will grow and the input
value is invalidated before being read from.

This found 14 failed Tests.

Reviewed By: bkramer

Differential Revision: https://reviews.llvm.org/D84293

The code has a few sequence that looked like:

    Ops.push_back(Ops[0]);
    Ops.erase(Ops.begin());

And are equivalent to:

    std::rotate(Ops.begin(), Ops.begin() + 1, Ops.end());

The latter has the advantage of never reallocating the vector, which
would be a bug in the original code as push_back would read from the
memory it deallocated.

That way, it runs in parallel across functions.

When comparing arrays whose shapes do not conform, the contant folding
code ran into problems trying to get the value of an extent that did not
exist.  There were actually two problems.  First, the routine
"CheckConformance()" was returning "true" when the compiler was unable
to get the extent of an array.  Second, the function
"ApplyElementwise()" was calling "CheckConformance()" prior to folding
the elements of two arrays, but it was ignoring the return value.

Differential Revision: https://reviews.llvm.org/D91440

Just like other sanitizers.
Fixes check-memprof under NPM.

Reviewed By: leonardchan

Differential Revision: https://reviews.llvm.org/D91389

When the load value is folded into the sin/cos operation, the
AMDGPU library calls simplifier could still mark the function
as unmodified. Instead ensure if there is an early return,
return whether the load was folded into the sin/cos call.

Authored by MJDSys

Differential Revision: https://reviews.llvm.org/D91401

Drop `IncludeBriefCommentsInCodeCompletion` since it is always `false`.

Differential Revision: https://reviews.llvm.org/D91295

I'm not why it was added to DAGToDAG oringally but it seems
to make sense alongside the non-TLS version:  LowerGlobalAddress

Differential Revision: https://reviews.llvm.org/D91432

Since RPATH initialization was disabled for the runtime libraries to
avoid overwriting RPATH unconditionally we need to explicity set up it
for the Win to Arm Linux cross builds.

See some details here: https://reviews.llvm.org/D91099

See discussion in D90554.

dependency. NFC

Use findSingleDependency in place of FindDependencies and stop passing a
set of Instructions around. Modify FindDependencies to return a boolean
flag which indicates whether the dependencies it has found are all
valid.

Add a convenience matcher which handles

```
G_XOR %not_reg, -1
```

And a convenience matcher which returns true if an integer constant is
all-ones.

Differential Revision: https://reviews.llvm.org/D91459

Make it required. Since it's a module pass, optnone won't test it, so
extend the clang test to also use opt-bisect now that it's supported.

14/16 check-dfsan tests failed with NPM enabled, now all pass.

Reviewed By: leonardchan

Differential Revision: https://reviews.llvm.org/D91385

ValueTracking was using a more powerful abs() implementation. Roll
it into KnownBits::abs(). Also add an exhaustive test for abs(),
in both the poisoning and non-poisoning variants.

Clean up the logic for `err_fe_{pch,module,ast}_file_modified` to use a
`select` like other ASTReader diagnostics. There should be no
functionality change here, just a cleanup.

Differential Revision: https://reviews.llvm.org/D91367

- Extract repeated code into helper function/lambdas.

Differential Revision: https://reviews.llvm.org/D91453