Check results of all type conversions in `--convert-arith-to-spirv`.

Fixes: https://github.com/llvm/llvm-project/issues/59496

Reviewed By: antiagainst

Differential Revision: https://reviews.llvm.org/D140033

This is part of a series of patches that aim at making Function::getBasicBlockList() private.

Differential Revision: https://reviews.llvm.org/D140064

Reviewed By: vitalybuka

Differential Revision: https://reviews.llvm.org/D140065

As far as I can tell m_riscv_Features_Group is for options that
can be passed to cc1's -target-feature. These mcmodel aliases aren't
that.

I think they get treated as aliases and disappear before we get
to processing m_riscv_Features_Group so it didn't cause any problems.

I've left them in m_Group to match mcmodel_EQ.

This is part of a series of patches that aim at making Function::getBasicBlockList() private.

Differential Revision: https://reviews.llvm.org/D139984

Exponentiation is lowered to either math::FPowI or Fortran runtime
call (in case of --math-runtime=precise).

MathToFuncs convertor will convert math::FPowI operations with
exponent width >32 to calls of outlined implementations and otherwise
will leave the operation to MathToLLVM convertor.

Reviewed By: jeanPerier

Differential Revision: https://reviews.llvm.org/D139806

We don’t have W versions of AND/OR/XOR/ANDN/ORN/XNOR so we should recursively check their users. We should limit the recursion to SelectionDAG::MaxRecursionDepth levels.

We need to add a Depth argument, all existing callers should pass 0 to the Depth. The new recursive calls should increment it by 1. At the top of the function we should give up and return false if Depth >= SelectionDAG::MaxRecursionDepth.

Reviewed By: craig.topper

Differential Revision: https://reviews.llvm.org/D139462

Reviewed By: aartbik

Differential Revision: https://reviews.llvm.org/D140057

[mlir][sparse] Add another call to ConvertVectorToLLVMPass, to lower the vector operations added by ConvertMathToLLVMPass.

Run sparse_tanh with vectorization.

Reviewed By: aartbik

Differential Revision: https://reviews.llvm.org/D139958

Differential Revision: https://reviews.llvm.org/D140052

llvm::Optional<T> has transform, which is equivalent to
std::optional<T>::transform.  The problem is that
std::optional<T>::transform won't be available until C++23, implying
that we probably cannot use it in our codebase untli 2028 or so.  We
certainly don't want to keep llvm::Optional just for transform.

This patch adds llvm::transformOptional to STLForwardCompat.h so that
we can use transform during the migration to std::optional and beyond.

I've shamelessly borrowed the implementation and test from
llvm/include/llvm/ADT/Optional.h and
llvm/unittests/ADT/OptionalTest.cpp respectively.

This is part of an effort to migrate from llvm::Optional to
std::optional:

https://discourse.llvm.org/t/deprecating-llvm-optional-x-hasvalue-getvalue-getvalueor/63716

Differential Revision: https://reviews.llvm.org/D139779

Reviewed By: nicolasvasilache, springerm

Differential Revision: https://reviews.llvm.org/D139978

This adds some SPARC feature flags to clang, for those that we have in common with GCC:

-m[no-]fpu
-m[no-]fsmuld
-m[no-]popc
-m[no-]vis
-m[no-]vis2
-m[no-]vis3
-m[hard/soft]-quad-float

All have the same meanings as GCC's options (https://gcc.gnu.org/onlinedocs/gcc/SPARC-Options.html).

This fixes, among other things, the -mno-fpu part of bug #40792

Reviewed By: nickdesaulniers

Differential Revision: https://reviews.llvm.org/D139768

**Summary**

Starting with DWARFv5, DW_AT_default_value can be used to indicate
that a template argument has a default value. With this patch LLVM
will emit the this attribute earlier versions of DWARF, unless
compiling with -gstrict-dwarf.

Differential Revision: https://reviews.llvm.org/D139953

**Summary**

Starting with DWARFv5, DW_AT_default_value can be used to indicate
that a template argument has a default value. With this patch Clang
will attach the attribute to the debug metadata regardless of version.

In a follow-up patch we will change llvm to emit this attribute in
earlier versions of DWARF, unless compiling with -gstrict-dwarf.

**Details**

* Previously the DwarfVersion check in CGDebugInfo was inconsistent:
  For non-type template arguments we attached the attribute to the debug
  metadata in DWARFv5 only. Whereas for type template arguments we didn't
  have such a version restriction. With this patch we attach the attribute
  regardless of DWARF version (and instead offload the check to the AsmPrinter
  in a future patch).

This allows allocaBuffer to be used outside of SparseTensorConversion.cpp, which will be helpful for a some future commits.

Reviewed By: aartbik, Peiming

Differential Revision: https://reviews.llvm.org/D140047

Reviewed By: aartbik

Differential Revision: https://reviews.llvm.org/D139961

Differential Revision: https://reviews.llvm.org/D140054

This reverts commit 7734053f
because it broke powerpc64 bot:
https://lab.llvm.org/buildbot#builders/231/builds/6229

Adds a target hook canPredicatePredicatedInstr(const MachineInstr&) that
assumes an instruction is already predicated and returns true if it can
be predicated again, used by the early if-conversion pass in order to
iterate multiple times on architectures supporting predicate logic.

No test added since there is no upstream target that can take advantage.

Differential Revision: https://reviews.llvm.org/D139981

Changes all unqualified calls to __throw_format_error to use a qualified
call.

Reviewed By: #libc, philnik

Differential Revision: https://reviews.llvm.org/D140038

Reviewed By: bixia

Differential Revision: https://reviews.llvm.org/D139983

Reviewed By: python3kgae

Differential Revision: https://reviews.llvm.org/D139742

The other operands in this function had it and every other test
in the file has it. So I assume this was a mistake.

This patch adds support for including binary ids in an indexed profile.
It adds a new field into the header that points to the offset of the
binary id section. The binary id section consists of a size of the
section, and a list of binary ids (if they are present) that consist
of two parts: length and data.

This patch guarantees that indexed profile is backwards compatible
after adding binary ids.

Differential Revision: https://reviews.llvm.org/D135929

Fixes issue 59413.

On RV64 these instructions produce a 32-bit value and sign extend
to 64-bits.

Reviewed By: reames

Differential Revision: https://reviews.llvm.org/D140036

Avoid calling separate isa<> and cast<> if we can - dyn_cast<> can more efficiently check for a safe cast and give the casted pointer.

This is a clean up following the revision D139868 (https://reviews.llvm.org/D139868).

Reviewed By: gribozavr2, ymandel

Differential Revision: https://reviews.llvm.org/D140037

[GVN] hasUsersIn - merge isa<> and cast<> into single dyn_cast<> and convert for-range loop to any_of() test. NFCI.

Avoid running isa<> and cast<> if we can - dyn_cast<> can more efficiently check for a safe cast and give the casted pointer.

Shadow memory was not marked as non-dumpable after a reset resulting in process taking a very long time to generate a coredump.

Reviewed By: vitalybuka

Differential Revision: https://reviews.llvm.org/D139425

These tests covers mulf, ori, and subi.

NFC

Differential Revision: https://reviews.llvm.org/D139625

This is a follow up to patch discussion on D139656. As noted there, M2/M4/M8 versions of these instructions don't actually exist, and using them results in overly constrained register allocation.

In that review, we'd talked about moving towards a variant of the instructions which ignored LMUL. I decided to see what happened if we just stopped generating the high LMUL variants, and the results are surprisingly neutral. I only see one minor thing which looks like a real regression among all the churn. I think this is worth doing now to loosen register allocation constraints, and avoid digging our hole around these instructions deeper while thinking about the right model change.

Differential Revision: https://reviews.llvm.org/D140027