The changes made in D88594 caused the test OpenMP/driver.c to fail on a 32-bit host becuase it was offloading to a 64-bit architecture by default. The offloading test was moved to a new file and a feature was added to the lit config to check for a 64-bit host.

Reviewed By: daltenty

Differential Revision: https://reviews.llvm.org/D89904

This commit should really be named "Workaround external projects depending
on libc++ build system implementation details". It seems that the compiler-rt
build (and perhaps other projects) is relying on the fact that we copy libc++
and libc++abi headers to `<build-root>/include/c++/v1`. This was changed
by 5d796645, which moved the headers to `<build-root>/projects/libcxx/include/c++/v1`
and broke the compiler-rt build.

I'm committing this workaround to fix the compiler-rt build, but we should
remove reliance on implementation details like that. The correct way to
setup the compiler-rt build would be to "link" against the `cxx-headers`
target in CMake, or to run `install-cxx-headers` using an appropriate
installation prefix, and then manually add a `-I` path to that location.

Forward missing attributes when creating the new transfer op otherwise the
builder would use default values.

Differential Revision: https://reviews.llvm.org/D89907

emit-obj is unnecessary here and further wasn't redirected to /dev/null.

MLIRTransforms is needed to provide BufferizeTypeConverter
definitions.

fix a few exposed bugs.

non-type template parameters.

Create a unique TemplateParamObjectDecl instance for each such value,
representing the globally unique template parameter object to which the
template parameter refers.

No IR generation support yet; that will follow in a separate patch.

While running this test on a bare metal target, I got an error as 'sleep' was not available on that system. As 'sleep' call is not doing anything useful for cases when _LIBCXXABI_HAS_NO_THREADS is defined. This patch puts it under this check.

Reviewed By: ldionne

Differential Revision: https://reviews.llvm.org/D89871

Reviewed By: thakis

Differential Revision: https://reviews.llvm.org/D89015

As discussed in the review for D87120 (specifically at
https://reviews.llvm.org/D87120#inline-831939), clean up PrintModuleMap
and DumpProcessMap usage differences. The former is only implemented for
Mac OSX, whereas the latter is implemented for all OSes. The former is
called by asan and tsan, and the latter by hwasan and now memprof, under
the same option. Simply rename the PrintModuleMap implementation for Mac
to DumpProcessMap, remove other empty PrintModuleMap implementations,
and convert asan/tsan to new name. The existing posix DumpProcessMap is
disabled for SANITIZER_MAC.

Differential Revision: https://reviews.llvm.org/D89630

PreservedCFGCheckerInstrumentation was saying that LowerMatrixIntrinsics
didn't properly preserve CFG even though it claimed to. The legacy pass
says it doesn't. Match the legacy pass's preserved analyses.

Reviewed By: thakis

Differential Revision: https://reviews.llvm.org/D89175

For GC parseable element atomic memcpy/memmove we'll need to
shuffle statepoint arguments. Make it possible by storing the
arguments as Value *, not Use *.

* Adds a new MlirOpPrintingFlags type and supporting accessors.
* Adds a new mlirOperationPrintWithFlags function.
* Adds a full featured python Operation.print method with all options and the ability to print directly to files/stdout in text or binary.
* Adds an Operation.get_asm which delegates to print and returns a str or bytes.
* Reworks Operation.__str__ to be based on get_asm.

Differential Revision: https://reviews.llvm.org/D89848

A "structural" type conversion is one where the underlying ops are
completely agnostic to the actual types involved and simply need to update
their types. An example of this is shape.assuming -- the shape.assuming op
and the corresponding shape.assuming_yield op need to update their types
accordingly to the TypeConverter, but otherwise don't care what type
conversions are happening.

Also, the previous conversion code would not correctly materialize
conversions for the shape.assuming_yield op. This should have caused a
verification failure, but shape.assuming's verifier wasn't calling
RegionBranchOpInterface::verifyTypes (which for reasons can't be called
automatically as part of the trait verification, and requires being
called manually). This patch also adds that verification.

Differential Revision: https://reviews.llvm.org/D89833

A "structural" type conversion is one where the underlying ops are
completely agnostic to the actual types involved and simply need to update
their types. An example of this is scf.if -- the scf.if op and the
corresponding scf.yield ops need to update their types accordingly to the
TypeConverter, but otherwise don't care what type conversions are happening.

To test the structural type conversions, it is convenient to define a
bufferize pass for a dialect, which exercises them nicely.

Differential Revision: https://reviews.llvm.org/D89757

This is similar in spirit to 01ea93d8 (memcpy) except that
here the underlying caller assumptions were created for vectorizer
use (throughput) rather than other passes.

That meant ARM could have an enormous throughput cost with no
corresponding size, latency, or blended cost increase. X86 has
the same throughput restriction as the basic implementation, so
it is still unchanged.

Paraphrasing from the previous commit:
This may not make sense for some callers, but at least now the
costs will be consistently wrong instead of mysteriously wrong.

Targets should provide better overrides if the current modeling
is not accurate.

This is a copy of the x86 file to provide better coverage;
x86 may have strange overrides that mask changes in the
generic model.

Update comments describing how OMPGridValues enums will be used in
clang, deviceRTLs, and hsa and cuda plugins.

Reviewed By: jdoerfert

Differential Revision: https://reviews.llvm.org/D86232

37c030f8 made it so that depending on //libcxx/include
automatically added the copied header dir to the include search path.
For some reason, clang can't build against the copied libcxx headers
(it complains about ldiv_t not being a type). I don't have a mac
to debug right now, but for the clang target this change was
unintentional anyways -- only depend on the copies target, instead
of on the target that also adjusts the include path.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D89023

[libomptarget] Require LLVM source tree to build libomptarget

This is to permit reliably #including files from the LLVM tree in libomptarget,
as an improvement on the copy and paste that is currently in use. See D87841
for the first example of removing duplication given this new requirement.

The weekly openmp dev call reached consensus on this approach. See also D87841
for some alternatives that were considered. In the future, we may want to
introduce a new top level repo for shared constants, or start using the ADT
library within openmp.

This will break sufficiently exotic build systems, trivial fixes as below.

Building libomptarget as part of the monorepo will continue to work.
If openmp is built separately, it now requires a cmake macro indicating
where to find the LLVM source tree.

If openmp is built separately, without the llvm source tree already on disk,
the build machine will need a copy of a subset of the llvm source tree and
the cmake macro indicating where it is.

Reviewed By: protze.joachim

Differential Revision: https://reviews.llvm.org/D89426

Differential revision: https://reviews.llvm.org/D89876

In my previous implementation of the semantic checks for ASSOCIATED(), I
had neglected to check the TARGET= argument for objects to ensure that
it has either the POINTER or TARGET attributes.

I added an implementation and a test.

Differential Revision: https://reviews.llvm.org/D89717

1. Fixed liveness issue with implicit kills.
2. Fixed potential problem with an indirect mov.

Fixes: SWDEV-256848

Differential Revision: https://reviews.llvm.org/D89599

Fix http://lists.llvm.org/pipermail/llvm-dev/2020-October/145908.html

Currently non-SHF_ALLOC SHT_REL[A] (due to --emit-relocs) and SHF_LINK_ORDER are not
marked live.

Reviewed By: grimar, psmith

Differential Revision: https://reviews.llvm.org/D89841

Add tests for insertion into non-undef vectors

Add SSE2/AVX512F/AVX512BW coverage

Changes:
 - initializer expressions of constexpr variable are now wraped in a ConstantExpr. this is mainly used for testing purposes. the old caching system has not yet been removed.
 - Add all the missing Serialization and Importing for APValue.
 - Improve dumping of APValue when ASTContext isn't available.
 - Cleanup leftover from last patch.
 - Add Tests for Import and serialization.

Differential Revision: https://reviews.llvm.org/D63640

This is needed to prevent the test from failing when llvm is
configured so that the arm target is not present, which is the case
for some buildbots.

We use the Real vs Pseudo instruction abstraction for other
types of instructions to facilitate changes in opcode
between gpu generations.
This patch introduces that abstraction to SOPC and SOPP.

Reviewed By: rampitec

Differential Revision: https://reviews.llvm.org/D89738

Change-Id: I59d53c2c7058b49d05b60350f4062a9b542d3138