This reverts commit e9d9a612.

This patch was previously revert by 04879086
with the reapplication being done after breaking the assert used to
ensure SP is always 16-byte aligned, which is a requirement of the AAPCS.

For extra context the latest patch caused runtime failures when
building with "-march=armv8-a+sve -mllvm -aarch64-sve-vector-bits-min=256".

Unmerges have the same fundamental problem as G_TRUNC, and G_TRUNC
could be implemented in terms of G_UNMERGE_VALUES. Reducing the number
of elements in unmerge results ends up producing the original unmerge
type profile, so the artifact combiner needs to eliminate the
intermediate illegal registers. This avoids infinite looping in the
legalizer in a future change.

Assuming an unmerge has each result unmerged the same way, this ends
up producing a new unmerge of the source for every definition. I'm not
sure if the artifact combiner should either insert temporary merges
here and erase the original merge, or if the combiner should look at
uses from defs rather than defs from uses for unmerges.

In a few cases this regresses from using 16-bit shifts for 8-bit
values to using 32-bit shifts, but I think these can be legalized
later (the other legalization rules don't try very hard to use 16-bit
shifts either).

This will make updating the checks easier in a future change.

Currently the dbg_value ends up in the relaxed branch block. A future
commit will push the dbg_value out of this block, and I'm not sure how
to coax the IR into producing the same MIR at the relevant point.

Differential Revision: https://reviews.llvm.org/D86707

This patch adds the initial toolchain for z/OS that will set some defaults. In subsequent patches, we plan to add support to use the system linker and assembler.

Reviewed By: hubert.reinterpretcast

Differential Revision: https://reviews.llvm.org/D86707

Loop Idiom Recognize Pass (LIRP) attempts to transform loops with subscripted arrays
into memcpy/memset function calls. In some particular situation, this transformation
introduces negative impacts. For example: https://bugs.llvm.org/show_bug.cgi?id=47300

This patch will enable users to disable a particular part of the transformation, while
he/she can still enjoy the benefit brought about by the rest of LIRP. The default
behavior stays unchanged: no part of LIRP is disabled by default.

Reviewed By: etiotto (Ettore Tiotto)

Differential Revision: https://reviews.llvm.org/D86262

Differential Revision: https://reviews.llvm.org/D85324

We have Error.cpp/.h which contains some code for working with error codes.
In fact we use Error/Expected<> almost everywhere already and we can get rid
of these files.

Note: a few places in the code used readobj specific error codes,
e.g. `return readobj_error::unknown_symbol`. But these codes are never really used,
i.e. the code checks the fact of a success/error call only.
So I've changes them to `return inconvertibleErrorCode()` for now.
It seems that these places probably should be converted to use `Error`/`Expected<>`.

Differential revision: https://reviews.llvm.org/D86772

Goes with D86726

This relands e9a3d1a4 which was originally
missing linking LLVMSupport into LLMVFileCheck which broke the SHARED_LIBS build.

Original summary:

The actual FileCheck logic seems to be implemented in LLVMSupport. I don't see a
good reason for having FileCheck implemented there as it has a very specific use
while LLVMSupport is a dependency of pretty much every LLVM tool there is. In
fact, the only use of FileCheck I could find (outside the FileCheck tool and the
FileCheck unit test) is a single call in GISelMITest.h.

This moves the FileCheck logic to its own LLVMFileCheck library. This way only
FileCheck and the GlobalISelTests now have a dependency on this code.

Reviewed By: jhenderson

Differential Revision: https://reviews.llvm.org/D86344

Function Sema::isOpenMPGlobalCapturedDecl() has a parameter `unsigned Level`,
but use `Level >= 0` as the condition of `while`, thus cause an infinite loop.
Fix by changing the loop condition to `Level > 0`.

Reviewed By: ABataev

Differential Revision: https://reviews.llvm.org/D86858

This patch adds tests for dumping the .debug_line_str section. The
dumper of the debug_line_str section is able to write nonprintable
strings as escaped chars while the dumper of the debug_str and
debug_str.dwo sections cannot. I will fix them and add tests for them
in a follow-up patch.

Reviewed By: jhenderson

Differential Revision: https://reviews.llvm.org/D86916

Fix compilation of libcxx when using -DLIBCXX_BUILD_EXTERNAL_THREAD_LIBRARY. Target `cxx_external_threads` gets linked to `cxx-headers` to include all needed headers and flags.

Reviewed By: #libc, ldionne

Differential Revision: https://reviews.llvm.org/D86773

`amdgpu-elf-definitions.test` is a file that tests
`Format`, 'OS/ABI' and 'Machine' fields printed for EM_AMDGPU.
It uses a precompiled object.

It is not needed to have it, because:
1) We already test formats in `file-header-format.test` + `llvm/unittests/Object/ELFObjectFileTest.cpp`.
2) We test 'OS/ABI' in `file-header-os-abi.test` for all possible values.
3) We test machine types in `file-header-machine-types.test`

In addition we have `amdgpu-elf-headers.test` which tests EM_AMDGPU specific flags. I.e.
it also tests the fact that llvm-readobj/elf are able to dump such objects.

This patch removes the test case and the corresponding binary.

Differential revision: https://reviews.llvm.org/D86774

Addresses post commit review comment for D86771.

This replaces `reportError` calls with `reportUniqueWarning` and improves testing
for the code that is related to stack sizes dumping.

Differential revision: https://reviews.llvm.org/D86783

This reverts commit 297f69af. It broke
the Fedora 33 x86-64 bot. See the review for more info.

By using optionals, we no longer have to check the validity of types that we
get from a lookup. This way, the definition of the summaries have a declarative
form, there are no superflous conditions in the source code.

Differential Revision: https://reviews.llvm.org/D86531

I have fixed up a number of warnings resulting from TypeSize -> uint64_t
casts and calling getVectorNumElements() on scalable vector types. I
think most of the changes are fairly trivial except for those in
DAGTypeLegalizer::SplitVecRes_MLOAD I've tried to ensure we create
the MachineMemoryOperands in a sensible way for scalable vectors.

I have added a CHECK line to the following test:

  CodeGen/AArch64/sve-split-load.ll

that ensures no new warnings are added.

Differential Revision: https://reviews.llvm.org/D86697

Expensive checks are failing, complaining about additional MMO operands
added to the branch.

-Wtautological-unsigned-zero-compare not under
-Wtautological-value-range-compare.

Continue to heuristically pick the wider of the two operands for
narrowing conversion warnings so that some_char + 1 isn't treated as
being wider than a char, but use the more accurate computation for
tautological comparison warnings.

Differential Revision: https://reviews.llvm.org/D85778

-frewrite-includes.

Remove the special-case (and highly implausible) diagnostic for a
compound token that crosses a file boundary, and instead model that case
the same as a compound token separated by whitespace, so that file
transitions and presumed file transitions behave the same way.

Use of a linebreak between the `(` and `{` in a GNU statement-expression
appears to be too common to include this warning in -Wall -- this occurs
in some Linux kernel headers, for example.

Currently it is hard to avoid having LLVM link to the system install of
ncurses, since it uses check_library_exists to find e.g. libtinfo and
not find_library or find_package.

With this change the ncurses lib is found with find_library, which also
considers CMAKE_PREFIX_PATH. This solves an issue for the spack package
manager, where we want to use the zlib installed by spack, and spack
provides the CMAKE_PREFIX_PATH for it.

This is a similar change as https://reviews.llvm.org/D79219, which just
landed in master.

Patch By: haampie

Differential Revision: https://reviews.llvm.org/D85820

MemoryPhis with a single value are correct, but can lead to errors when
updating. Clean up single entry Phis newly added when cloning blocks.
Resolves PR46574.

This patch makes the debug_str section optional. When the debug_str
section exists but doesn't contain anything, yaml2obj will emit a
section header for it.

Reviewed By: grimar

Differential Revision: https://reviews.llvm.org/D86860

This patch add the missing operands to the acc.loop operation. Only the device_type
information is not part of the operation for now.

Reviewed By: rriddle, kiranchandramohan

Differential Revision: https://reviews.llvm.org/D86753

Specifically, ensure that initializers conform with their objects
according to 8.2 para 4.

Differential Revision: https://reviews.llvm.org/D86886

Clients who rely on the Context loading dialects from the global
registry can call `mlir::enableGlobalDialectRegistry(true);` before
creating an MLIRContext

Differential Revision: https://reviews.llvm.org/D86897

Differential Revision: https://reviews.llvm.org/D86898

Change the expression representation TypeParamInquiry from being
a class that's templatized on the integer KIND of its result into
a monomorphic representation that results in a SubscriptInteger
that can then be converted.

This is a minor simplification, but it's worth doing because
it is believed to also be a work-around for bugs in the MSVC
compiler with overload resolution that affect the expression
traversal framework.

Differential Revision: https://reviews.llvm.org/D86551