Changes

Solaris buildbot updates

Once D84029 <https://reviews.llvm.org/D84029> lands, it's no longer
necessary to set `CLANG_DEFAULT_LINKER`.  On the contrary, it causes test
failures since D84029 <https://reviews.llvm.org/D84029>.

The patch also changes the Solaris/amd64 `LLVM_HOST_TRIPLE` to match the
bot name and Solaris conventions and fixes a pasto in the Solaris/sparcv9
triple.

Differential Revision: https://reviews.llvm.org/D85462

[mlir][openacc] Introduce OpenACC dialect with parallel, data, loop operations

This patch introduces the OpenACC dialect with three operation defined
parallel, data and loop operations with custom parsing and printing.

OpenACC dialect RFC can be find here: https://llvm.discourse.group/t/rfc-openacc-dialect/546/2

Reviewed By: rriddle, kiranchandramohan

Differential Revision: https://reviews.llvm.org/D84268

clang: Fix minor grammar-o in diag added in c354b2e3b

rG9bd97d0363987b582 - Revert "[X86][SSE] Fold HOP(SHUFFLE(X),SHUFFLE(Y)) --> SHUFFLE(HOP(X,Y))"

This reverts commit 9bd97d0363987b582e4a92b354b02e86ac068407.

Seeing some codegen issues in internal testing.

Fix unused variable warning. NFC.

Reduce the dyn_cast<> to a isa<> as that's all non-assert builds require, and move the cast<> inside the assert.

[analyzer] Fix crash with pointer to members values

This fix unifies all of the different ways we handled pointer to
members into one.  The crash was caused by the fact that the type
of pointer-to-member values was `void *`, and while this works
for the vast majority of cases it breaks when we actually need
to explain the path for the report.

rdar://problem/64202361

Differential Revision: https://reviews.llvm.org/D85817

[clangd] Revert "[clangd] Fix crash-bug in preamble indexing when using modules."

This reverts commit 4061d9e42cff621462931ac7df9666806c77a237.
Tests are failing in some configuration, likely due to not cleaning up
module cache path before running the test.

Differential Revision: https://reviews.llvm.org/D85907

[libc++] Ensure the CI scripts can find Ninja

Ninja isn't necessarily installed on macOS out of the box, so make sure
we use the SDK Ninja instead of relying on one being available on the
system.

[llvm-objdump] Change symbol name/PLT decoding errors to warnings

If the referenced symbol of a J[U]MP_SLOT is invalid (e.g. symbol index 0), llvm-objdump -d will bail out:

```
error: 'a': st_name (0x326600) is past the end of the string table of size 0x7
```

where 0x326600 is the st_name field of the first entry past the end of .symtab

Change it to a warning to continue dumping.
`X86/plt.test` uses a prebuilt executable, so I pick `ELF/AArch64/plt.test`
which has a YAML input and can be easily modified.

Reviewed By: jhenderson

Differential Revision: https://reviews.llvm.org/D85623

[ELF] -r: allow SHT_X86_64_UNWIND to be merged into SHT_PROGBITS

* For .cfi_*, GCC/GNU as emits SHT_PROGBITS type .eh_frame sections.
* Since rL252300, clang emits SHT_X86_64_UNWIND type .eh_frame sections
  (originated from Solaris, documented in the x86-64 psABI).
* Some assembly use `.section .eh_frame,"a",@unwind` to generate
  SHT_X86_64_UNWIND .eh_frame sections.

In a non-relocatable link, input .eh_frame are combined and there is
only one SyntheticSection .eh_frame in the output section, so the
"section type mismatch" diagnostic does not fire.

In a relocatable link, there is no SyntheticSection .eh_frame. .eh_frame of
mixed types can trigger the diagnostic. This patch fixes it by adding another
special case 0x70000001 (= SHT_X86_64_UNWIND) to canMergeToProgbits().

    ld.lld -r gcc.o clang.o => error: section type mismatch for .eh_frame

There was a discussion "RFC: Usefulness of SHT_X86_64_UNWIND" on the x86-64-abi
mailing list. Folks are not wild about making the psABI value 0x70000001 into
gABI, but a few think defining 0x70000001 for .eh_frame may be a good idea for a
new architecture.

Reviewed By: grimar

Differential Revision: https://reviews.llvm.org/D85785

[AArch64][x86] add tests for x/sqrt(x); NFC

[clangd] Clean up old test fixture/names a little. NFC

[flang] Correct the default value for PAD= to PAD='YES'.

[flang] Ensure Preprocessor::Define saves macro names correctly

This fixes problems with macros defined with -D on the command line
and predefined macros defined in the throwaway driver program.

[flang] Correct manipulation of mixed complex expressions

Ensure that mixed complex expressions (one operand complex,
the other not) are properly manipulated; add test.

[flang] Fix assert on character literal substrings as arguments

Character literal substrings used as arguments were causing asserts.  This
happened when the code was trying to get the DynamicType of the substring.  We
were only recording the DynamicType of the Designator on which the substring
was based.  For character literal substrings, the Designator was a character
literal, and we weren't handling getting its type.

I fixed this by changing the `GetType()` method for `DynamicType` to check to
see if we were getting the type of a `Substring` and calculating the type of
the substring by getting the number of bytes in an element of the string.

I also changed the test `resolve49.f90` with some tests, one of which causes
the original crash.

Differential Revision: https://reviews.llvm.org/D85908

[ELF] --gdb-index: skip SHF_GROUP .debug_info

-gdwarf-5 -fdebug-types-section may produce multiple .debug_info sections.  All
except one are type units (.debug_types before DWARF v5). When constructing
.gdb_index, we should ignore these type units. We use a simple heuristic: the
compile unit does not have the SHF_GROUP flag. (This needs to be revisited if
people place compile unit .debug_info in COMDAT groups.)

This issue manifests as a data race: because an object file may have multiple
.debug_info sections, we may concurrently construct `LLDDwarfObj` for the same
file in multiple threads. The threads may access `InputSectionBase::data()`
concurrently on the same input section. `InputSectionBase::data()` does a lazy
uncompress() and rewrites the member variable `rawData`. A thread running zlib
`inflate()` (transitively called by uncompress()) on a buffer with `rawData`
tampered by another thread may fail with `uncompress failed: zlib error: Z_DATA_ERROR`.

Even if no data race occurred in an optimistic run, if there are N .debug_info,
one CU entry and its address ranges will be replicated N times. The result
.gdb_index can be much larger than a correct one.

The new test gdb-index-dwarf5-type-unit.s actually has two compile units. This
cannot be produced with regular approaches (it can be produced with -r
--unique). This is used to demonstrate that the .gdb_index construction code
only considers the last non-SHF_GROUP .debug_info

Reviewed By: grimar

Differential Revision: https://reviews.llvm.org/D85579

[clangd] Extract BackgroundIndex::Options struct. NFC

I've dropped the background context parameter, since we in practice just pass the
current context there, and we now have a different way to specify context too.
While here, clean up a couple of comments.

Reviewed By: kadircet

Differential Revision: https://reviews.llvm.org/D83157

[InstCombine] Handle gc.relocate(null) in one iteration

InstCombine adds users of transformed instruction to working list to
process on the same iteration. However gc.relocate may have a hidden
user (next gc.relocate) which is connected through gc.statepoint intrinsic and
there is no direct def-use chain between them.

In this case if the next gc.relocation is already processed it will not be added
to worklist and will not be able to be processed on the same iteration.
Let's we have the following case:
A = gc.relocate(null)
B = statepoint(A)
C = gc.relocate(B, hidden(A))
If C is already considered then after replacement of A with null, statepoint B
instruction will be added to the queue but not C.
C can be processed only on the next iteration.

If the chain of relocation is pretty long the many iteration may be required.
This change is to reduce the number of iteration to meet the latest changes
related to reducing infinite loop threshold.

This is a quick (not best) fix. In the follow up patches I plan to move gc relocation
handling into statepoint handler. This should also help to remove unused gc live
entries in statepoint bundle.

Reviewers: reames, dantrushin
Reviewed By: reames
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D75598

[VectorCombine] Fix for non-zero addrspace when creating vector load from scalar load

This is a fixup to commit 43bdac290663f4424f9fb, to make sure the
address space from the original load pointer is retained in the
vector pointer.

Resolves problem with
  Assertion `castIsValid(op, S, Ty) && "Invalid cast!"' failed.
due to address space mismatch.

Reviewed By: spatel

Differential Revision: https://reviews.llvm.org/D85912

Add cold attribute to one time construction APIs

_cxa_guard_acquire is used for only one purpose,
namely guarding local static variable initialization,
and since that purpose is definitionally cold,
it should be attributed as cold

Reviewed By: ldionne

Reviewers: mclow.lists, ldionne, jfb, yfeldblum

Differential Revision: https://reviews.llvm.org/D85873

Fix PR45442: Bail out when MemorySSA information is not available

[clangd] Send EOF before resetting diagnostics consumer

Summary:
Some clang-tidy checkers, e.g. llvm-include-order can emit diagnostics
at this callback (as mentioned in the comments).

Clangd was resetting diag consumer to IgnoreDiags before sending EOF, hence we
were unable to emit diagnostics for such checkers.

This patch changes the order of that reset and preprocosser event to make sure
we emit that diag.

Fixes https://github.com/clangd/clangd/issues/314.

Reviewers: sammccall

Subscribers: ilya-biryukov, MaskRay, jkorous, arphaman, usaxena95, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D83178

[clangd] Move clang-tidy check modifications into ClangdServer

Summary:
This enables sharing the logic between standalone clangd and embedders
of it. The new approach should be same performance-wise, as it is only called
once per addDocument call.

This patch also introduces a blacklisting code path for disabling crashy or
high-noise tests, until we figure out a way to make them work with clangd-setup.

The biggest difference is the way we make use of preambles, hence those checks
can't see directives coming from the preamble section of the file. The second
thing is the fact that code might-not be compiling while clangd is trying to
build an AST, hence some checks might choke on those incomplete ASTs.

Reviewers: sammccall

Subscribers: ilya-biryukov, MaskRay, jkorous, arphaman, usaxena95, aaron.ballman, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D83224

[SVE] Lower fixed length vector integer SMIN/SMAX

Differential Revision: https://reviews.llvm.org/D85855

[lldb] Skip TestStepScripted with reproducers

Some of the test methods were already skipped because of an unexpected
packet. The test started failing after it was expanded. Skip the whole
test with reproducers so we don't have to add the decorator for every
method.

[flang] Implement shape analysis of TRANSFER intrinsic function result

The shape (esp. the size) of the result of a call to TRANSFER
is implemented according to the definition in the standard.

Differential Revision: https://reviews.llvm.org/D85866

[ARM] Mark VMINNMA/VMAXNMA as commutative

These operations take Qda and Rn register operands, which are
commutative so long as the instruction is not predicated.

Differential Revision: https://reviews.llvm.org/D85813

[mlir] Test CallOp STD->LLVM conversion.

This exercises the corner case that was fixed in
https://reviews.llvm.org/rG8979a9cdf226066196f1710903d13492e6929563.

The bug can be reproduced when there is a @callee with a custom type argument and @caller has a producer of this argument passed to the @callee.

Example:
func @callee(!test.test_type) -> i32
func @caller() -> i32 {
  %arg = "test.type_producer"() : () -> !test.test_type
  %out = call @callee(%arg) : (!test.test_type) -> i32
  return %out : i32
}

Even though there is a type conversion for !test.test_type, the output IR (before the fix) contained a DialectCastOp:

module {
  llvm.func @callee(!llvm.ptr<i8>) -> !llvm.i32
  llvm.func @caller() -> !llvm.i32 {
    %0 = llvm.mlir.null : !llvm.ptr<i8>
    %1 = llvm.mlir.cast %0 : !llvm.ptr<i8> to !test.test_type
    %2 = llvm.call @callee(%1) : (!test.test_type) -> !llvm.i32
    llvm.return %2 : !llvm.i32
  }
}

instead of

module {
  llvm.func @callee(!llvm.ptr<i8>) -> !llvm.i32
  llvm.func @caller() -> !llvm.i32 {
    %0 = llvm.mlir.null : !llvm.ptr<i8>
    %1 = llvm.call @callee(%0) : (!llvm.ptr<i8>) -> !llvm.i32
    llvm.return %1 : !llvm.i32
  }
}

Differential Revision: https://reviews.llvm.org/D85914

[lldb] Set the launch flags to GetLaunchInfo().GetLaunchFlags()

Instead of clearing the launch flags, always pass the target's current
launch flags.

[X86][SSE] Add HADD combine regression case from rG9bd97d036398

rG9bd97d036398 caused a miscompile of this internal test case

[flang] Descriptor-based I/O using wrong size for contiguous unformatted I/O

The descriptor-based I/O routine was using the size of the descriptor
rather than the size of the described data for the transfer.  Fix,
and add a comment to the relevant API.

Differential Revision: https://reviews.llvm.org/D85863

[llvm-install-name-tool] Add more documentation

Add documentation for the remaining options of
`llvm-install-name-tool`.

Reviewed by jhenderson, smeenai

Differential Revision: https://reviews.llvm.org/D85655

[BPI] Improve static heuristics for integer comparisons

Similarly as for pointers, even for integers a == b is usually false.

GCC also uses this heuristic.

Reviewed By: ebrevnov

Differential Revision: https://reviews.llvm.org/D85781

[llvm-libtool-darwin] Refactor ArchiveWriter

Refactoring function `writeArchive` in ArchiveWriter. Added a new
function `writeArchiveBuffer` that returns the archive in a memory
buffer instead of writing it out to the disk. This refactor is necessary
so as to allow `llvm-libtool-darwin` to write universal files containing
archives.

Reviewed by jhenderson, MaskRay, smeenai

Differential Revision: https://reviews.llvm.org/D84858

[flang][openacc] Handle optional end directive in combined construct

OpenACC combined construct can have an optional end directive. This patch handle this
case in the parsing/unparsing with a canonicalization step. Unlike OmpEndLoopDirective,
this doesn't need a special treatment in the pre-fir tree as there is no clause attached to
a AccEndCombinedDirective.

Reviewed By: klausler

Differential Revision: https://reviews.llvm.org/D84481

[llvm-libtool-darwin] Add support for -arch_only

Add support for -arch_only option for llvm-libtool-darwin. This diff
also adds support for accepting universal files as input and flattening
them to create the required static library. Supports input universal
files contaning both Mach-O object files or archives.

Differences from cctools' libtool:
- `-arch_only` can be specified multiple times
- archives containing universal files are considered invalid (libtool
allows such archives)

Reviewed by jhenderson, smeenai

Differential Revision: https://reviews.llvm.org/D84770

[JITLink][MachO] Return an error when MachO TLV relocations are encountered.

MachO TLV relocations aren't supported yet. Error out rather than falling
through to llvm_unreachable.

[llvm-jitlink] Don't demote unreferenced definitions in -harness mode.

Demoting unreferenced externals is unsafe if multiple interdependent test
objects are used, including objects loaded from archives.

[Diagnostics] Skip var decl of structs for -Wstring-concatenation

[Tests] Removed debug copy command

Fix PR45442: Bail out when MemorySSA information is not available

Reviewers: sebpop, uabelho, fhahn
Reviewed by: fhahn

Differential Revision: https://reviews.llvm.org/D85881

[AMDGPU] Fix misleading SDWA verifier error. NFC.

The old error from GFX9 shall be updated to GFX9+.

[AMDGPU] Inhibit SDWA if target instruction has FI

Differential Revision: https://reviews.llvm.org/D85918

Revert "[llvm-jitlink] Don't demote unreferenced definitions in -harness mode."

This reverts commit e137b550587a85b0d9c9c539edc79de0122b6946.

This commit broke a test case. Reverting while I investigate.

[elfabi] Move llvm-elfabi related code to InterfaceStub library

This change moves elfabi related code to llvm/InterfaceStub library
so it can be shared by multiple llvm tools without causing cyclic
dependencies.

Differential Revision: https://reviews.llvm.org/D85678

[LLDB] Fix how ValueObjectChild handles bit-fields stored in a Scalar in UpdateValue()

When bit-field data was stored in a Scalar in ValueObjectChild during UpdateValue()
it was extracting the bit-field value. Later on in lldb_private::DumpDataExtractor(…)
we were again attempting to extract the bit-field. Which would then not obtain the
correct value. This will remove the extra extraction in UpdateValue().
We hit this specific case when values are passed in registers, which we could only
reproduce in an optimized build.

Differential Revision: https://reviews.llvm.org/D85376

[ORC] Fix SymbolLookupSet::containsDuplicates()

[ORC] cloneToNewContext() can work with a const-ref to ThreadSafeModule

[mlir][DialectConversion] Update the documentation for dialect conversion

This revision updates the documentation for dialect conversion, as many concepts have changed/evolved over time.

Differential Revision: https://reviews.llvm.org/D85167

[mlir][docs] Update/Add documentation for MLIRs Pattern Rewrite infrastructure

This infrastructure has evolved a lot over the course of MLIRs lifetime, and has never truly been documented outside of rationale or proposals. This revision aims to document the infrastructure and user facing API, with the rationale specific portions moved to the Rationale folder and updated.

Differential Revision: https://reviews.llvm.org/D85260

[darwin][driver] fix isMacosxVersionLT minimum supported OS version check

The previous Driver's triple check only worked for -target, but not for -arch -mmacosx-version-min invocations

[ORC][NFC] Fix typo in comment

[gn build] (manually) port d650cbc349ccc4

[gn build] (manually) port d650cbc349ccc4 better

[SVE] Lower fixed length vector integer UMIN/UMAX

Differential Revision: https://reviews.llvm.org/D85926

[libc] Add strtok_r implementation.

Reviewed By: sivachandra

Differential Revision: https://reviews.llvm.org/D85779

[tsan] Respect no_huge_pages_for_shadow.

Disable huge pages in the TSan shadow regions when no_huge_pages_for_shadow == true (default).

Reviewed By: eugenis

Differential Revision: https://reviews.llvm.org/D85841

[SystemZ][z/OS] enable trigraphs by default on z/OS

This patch enables trigraphs on z/OS.

Reviewed By: hubert.reinterpretcast, fanbo-meng

Differential Revision: https://reviews.llvm.org/D85722

Add "status" to the list of absl libraries.

The Abseil-NoInternalDependenciesCheck currently mistakenly triggers on any usage of internal helpers even if it is within absl/status.

Differential Revision: https://reviews.llvm.org/D85843

sanitizer_common: Introduce internal_madvise and start using it.

A recent change to sanitizer_common caused us to issue the syscall
madvise(MADV_HUGEPAGE) during HWASAN initialization. This may lead to a
problem if madvise is instrumented (e.g. because libc is instrumented
or the user intercepted it). For example, on Android the syscall may
fail if the kernel does not support transparent hugepages, which leads
to an attempt to set errno in a HWASAN instrumented function. Avoid
this problem by introducing a syscall wrapper and using it to issue
this syscall.

Tested only on Linux; includes untested updates for the other
platforms.

Differential Revision: https://reviews.llvm.org/D85870

[InstSimplify] Add tests for assume with min/max intrinsic (NFC)

If we assume one of the operands is smaller/greater, then min/max
may be simplified.

[SLC] Optimize strncpy(a, a, C) to memcpy(a, a000, C)
Solves PR47154

Re-apply "[llvm-jitlink] Don't demote unreferenced definitions in -harness mode"

This reapplies commit e137b550587a85b0d9c9c539edc79de0122b6946 with
fixes for the broken test case: Non-global symbols should only be
skipped after checking that they're not referenced by the harness.

[Tests] Update strncpy tests

[clang][Driver] Default to /usr/bin/ld on Solaris

`clang` currently requires the native linker on Solaris:

  - It passes `-C` to `ld` which GNU `ld` doesn't understand.

  - To use `gld`, one needs to pass the correct `-m EMU` option to select
    the right emulation.  Solaris `ld` cannot handle that option.

So far I've worked around this by passing `-DCLANG_DEFAULT_LINKER=/usr/bin/ld`
to `cmake`.  However, if someone forgets this, it depends on the user's
`PATH` whether or not `clang` finds the correct linker, which doesn't make
for a good user experience.

While it would be nice to detect the linker flavor at runtime, this is more
involved.  Instead, this patch defaults to `/usr/bin/ld` on Solaris.  This
doesn't work on its own, however: a link fails with

  clang-12: error: unable to execute command: Executable "x86_64-pc-solaris2.11-/usr/bin/ld" doesn't exist!

I avoid this by leaving absolute paths alone in `ToolChain::GetLinkerPath`.

Tested on `amd64-pc-solaris2.11`, `sparcv9-sun-solaris2.11`, and
`x86_64-pc-linux-gnu`.

Differential Revision: https://reviews.llvm.org/D84029

[WebAssembly] Allow inlining functions with different features

Allow inlining only when the Callee has a subset of the Caller's
features. In principle, we should be able to inline regardless of any
features because WebAssembly supports features at module granularity,
not function granularity, but without this restriction it would be
possible for a module to "forget" about features if all the functions
that used them were inlined.

Requested in PR46812.

Differential Revision: https://reviews.llvm.org/D85494

[AMDGPU] Fix FP/BP spills when MUBUF constant offset exceeded

If we need a scratch register for the spill don't use the same scratch
register that is being used for the MBUF offset.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D85772

[lldb] Improve diagnostics in lldb-repro when replay fails

 - Print the replay invocation.
 - Keep the reproducer around.
 - Return the "opposite" exit code so we don't have to rely on FileCheck
   to fail the test when the expected exit code is non-zero.

Revert "[ARM] Fix IT block generation after Thumb2SizeReduce with -Oz"

This reverts commit 18279a54b5d3382874924d6a3c7775b7e22598dc as it is
causing some chromium android test problems.

[NFC] Remove unused variables.

These were no longer needed following
96855125e77044b1a5d3c7f0ae90ea3a5cb035c0.

Reviewed By: rupprecht

Differential Revision: https://reviews.llvm.org/D85939

[SLC] sprintf(dst, "%s", str) -> strcpy(dst, str)

Solves 46489

[DebugInfo] Add -fuse-ctor-homing cc1 flag so we can turn on constructor homing only if limited debug info is already on.

This adds a cc1 flag to enable constructor homing but doesn't turn on debug
info if it wasn't enabled already (which is what using
-debug-info-kind=constructor does). This will be used for testing, and won't
be needed anymore once ctor homing is used as default / merged into =limited.

Bug to enable ctor homing: https://bugs.llvm.org/show_bug.cgi?id=46537

Differential Revision: https://reviews.llvm.org/D85799

Solaris buildbot updates

Once D84029 <https://reviews.llvm.org/D84029> lands, it's no longer
necessary to set `CLANG_DEFAULT_LINKER`.  On the contrary, it causes test
failures since D84029 <https://reviews.llvm.org/D84029>.

The patch also changes the Solaris/amd64 `LLVM_HOST_TRIPLE` to match the
bot name and Solaris conventions and fixes a pasto in the Solaris/sparcv9
triple.

Differential Revision: https://reviews.llvm.org/D85462