Recent versions of python3's multiprocessing module will blow up with
a Runtime error from this code, saying:

  An attempt has been made to start a new process before the
  current process has finished its bootstrapping phase

This is becuae the wrappers in bin/ are not using the  `__name__ == "__main__"`   idiom correctly.

Reviewed By: ldionne

Differential Revision: https://reviews.llvm.org/D87051

Currently we treat SHT_RISCV_ATTRIBUTES like a normal section and
concatenate all such input sections, yielding invalid output unless only
a single attributes section is present in the input. Instead, pick the
first as with SHT_ARM_ATTRIBUTES. We do not currently need to condition
our behaviour on the contents, unlike Arm. In future, we should both do
stricter validation of the input and merge all sections together to
ensure we have, for example, the full arch string requirement, but this
rudimentary implementation is good enough for most common cases.

Reviewed By: MaskRay

Differential Revision: https://reviews.llvm.org/D86309

A code block wasn't properly introduced.

After adding a field of one bit, the bitfield members would take
30+1+1+1 = 33 bits, causing the size of TemplateParameterList to
increase from 16 to 24 bytes on 64-bit systems.

With 29 bits for NumParams we can encode up to half a billion template
parameters, which is almost certainly still enough for anybody.

And duplicate tests for known non-negative from InstSimplify.

The constructor asserts that, use it in the ThreadSafetyAnalyzer.
Also note that the result of a cast<> cannot be null.

Instead of just mutex members we also consider mutex globals.
Unsurprisingly they are always in scope. Now the paper [1] says that

> The scope of a class member is assumed to be its enclosing class,
> while the scope of a global variable is the translation unit in
> which it is defined.

But I don't think we should limit this to TUs where a definition is
available - a declaration is enough to acquire the mutex, and if a mutex
is really limited in scope to a translation unit, it should probably be
only declared there.

[1] https://static.googleusercontent.com/media/research.google.com/en/us/pubs/archive/42958.pdf

Fixes PR46354.

Reviewed By: aaron.ballman

Differential Revision: https://reviews.llvm.org/D84604

Recognize umin/umax/smin/smax intrinsics and convert them to the
already existing SCEV nodes of the same name.

In the future we'll want SCEVExpander to also produce the intrinsics,
but we're not ready for that yet.

Differential Revision: https://reviews.llvm.org/D87160

Similar to D87168, but for abs. If we have a dominating x >= 0
condition, then we know that abs(x) is x. This fold is in
InstCombine, because we need to create a sub instruction for
the x < 0 case.

Differential Revision: https://reviews.llvm.org/D87184

If we have a dominating condition that x >= y, then umax(x, y) is x,
etc. I'm doing this in InstSimplify as the corresponding transform
for the select form is also done there.

Differential Revision: https://reviews.llvm.org/D87168

Following the same transform for the select version of abs.

Drop unused function arg from affine loop tiling validity check.

When parsing a C++17 binding declaration, we first create the
BindingDecls in Sema::ActOnDecompositionDeclarator, and then build the
DecompositionDecl in Sema::ActOnVariableDeclarator, so the contained
BindingDecls are never null. But when deserializing, we read the
DecompositionDecl with all properties before filling in the Bindings.
Among other things, reading a declaration reads whether it's invalid,
then calling setInvalidDecl which assumes that all bindings of the
DecompositionDecl are available, but that isn't the case.

Deserialization should just set all properties directly without invoking
subsequent functions, so we just set the flag without using the setter.

Fixes PR34960.

Reviewed By: rsmith

Differential Revision: https://reviews.llvm.org/D86207

I don't think this is obvious, since try-acquire seemingly contradicts
our usual requirements of "no conditional locking".

Reviewed By: aaron.ballman

Differential Revision: https://reviews.llvm.org/D87065

Same as the existing SPF_ABS fold. We don't need to explicitly
handle NABS, as the negs will get folded away first.

Same as the existing fold for SPF_ABS. We don't need to explicitly
handle the NABS variant, as we'll first fold away the neg in that
case.

Previously SDNodeFlags::instersectWith(Flags) would do nothing if Flags was
in an undefined state, which is very bad given that this is the default when
getNode() is called without passing an explicit SDNodeFlags argument.

This meant that if an already existing and reused node had a flag which the
second caller to getNode() did not set, that flag would remain uncleared.

This was exposed by https://bugs.llvm.org/show_bug.cgi?id=47092, where an NSW
flag was incorrectly set on an add instruction (which did in fact overflow in
one of the two original contexts), so when SystemZElimCompare removed the
compare with 0 trusting that flag, wrong-code resulted.

There is more that needs to be done in this area as discussed here:

Differential Revision: https://reviews.llvm.org/D86871

Review: Ulrich Weigand, Sanjay Patel

When a switch case is folded into default's case, that's an IR change that
should be reported, update ConstantFoldTerminator accordingly.

Differential Revision: https://reviews.llvm.org/D87142

Libcall __gcc_qtou is not available, which breaks some tests needing
it. On PowerPC, we have code to manually expand the operation, this
patch applies it to constrained conversion. To keep it strict-safe,
it's using the algorithm similar to expandFP_TO_UINT.

For constrained operations marking FP exception behavior as 'ignore',
we should set the NoFPExcept flag. However, in some custom lowering
the flag is missed. This should be fixed by future patches.

Reviewed By: uweigand

Differential Revision: https://reviews.llvm.org/D86605

These overloads are listed in appendix A of the ELFv2 ABI specification
without a requirement for ISA 3.0. So these need to be available on
all Altivec-capable architectures. The implementation in altivec.h
erroneously had them guarded for Power9 due to the availability of
the VCMPNE[BHW] instructions. However these need to be implemented
in terms of the VCMPEQ[BHW] instructions on older architectures.

Fixes: https://bugs.llvm.org/show_bug.cgi?id=47423

Copied from lldb/.clang-tidy (D75810).

Most compiler-rt code actually uses variableName or variable_name but not VariableName.
Lots of functions use `__function_name` and FunctionName instead of functionName.
Just exclude readability-identifier-naming.

One check was missed on the previous attempt

Just a shift, no other formatting changes.

With 6a754968, these two options are no longer
forwarded to GCC. This patch restores the original behavior.

Reviewed By: MaskRay

Differential Revision: https://reviews.llvm.org/D87162

Differential Revision: https://reviews.llvm.org/D86992

Differential Revision: https://reviews.llvm.org/D86992

This can cause an infinite loop if SimplifiedDemandedElts asks
for the node to replace itself.

A similar protection exists in other places in shuffle combining.

Fixes ISPC https://github.com/ispc/ispc/issues/1864

Differential Revision: https://reviews.llvm.org/D86991

TPCDynamicLibrarySearchGenerator was generating errors on missing
symbols, but that doesn't fit the DefinitionGenerator contract: A symbol
that isn't generated by a particular generator should not cause an
error.

This commit fixes the error by using SymbolLookupFlags::WeaklyReferencedSymbol
for all elements of the lookup, and switches llvm-jitlink to use
TPCDynamicLibrarySearchGenerator.