They already include workarounds.

This reverts commit aa217eb0.
This reverts commit 634feddc.

This breaks buildbots by introducing cycle dependency between libObject
and TextAPI and breaks gcc compiles on buildbots.

* Reported in buildbot failures

Add support for reading binary Mach-o dynamic libraries. It uses
libObject APIs for extracting information relavant to TAPI and tbd
files. This includes but is not limited to load commands encode data
like install names, current/compat versions and symbols.

Rather than using size_t, use uint32_t. We don't have more than 4
billion intrinsics.

This patch revives the effort to get this Phabricator patch into
upstream:

https://reviews.llvm.org/D137900

This patch was accepted before in Phabricator but I found some
-gsimple-template-names issues that are fixed in this patch.

A fixed up version of the description from the original patch starts
now.

This patch started off trying to fix Module::FindFirstType() as it
sometimes didn't work. The issue was the SymbolFile plug-ins didn't do
any filtering of the matching types they produced, and they only looked
up types using the type basename. This means if you have two types with
the same basename, your type lookup can fail when only looking up a
single type. We would ask the Module::FindFirstType to lookup "Foo::Bar"
and it would ask the symbol file to find only 1 type matching the
basename "Bar", and then we would filter out any matches that didn't
match "Foo::Bar". So if the SymbolFile f...

Port `CFGuard` to new pass manager, add a pass parameter to choose guard
mechanism.

Revert "[X86] Set SHF_X86_64_LARGE for globals with explicit well-known large section name (#74381)"

This reverts commit 323451ab.

Code with these section names in the wild doesn't compile because
support for large globals in the small code model is not complete yet.

Separates actual transformation files from supporting utility files in
the transforms directory. Includes a bazel overlay fix for the build (as
well as a bit of cleanup of that file to be less verbose and more
flexible).

I was not able to fully triage why this just started failing on one of
our bots as it seems that the use was added 4 months ago. I would assume
that it was accidentally coming in transitively in some way as the dep
was definitely missing.

For context, this started failing in [our
byo_llvm](https://github.com/openxla/iree/blob/main/build_tools/llvm/byo_llvm.sh)
build on a stock build of MLIR on top of an existing LLVM. We were
getting:

```
ld.lld: error: undefined symbol: mlir::registerSPIRVDialectTranslation(mlir::DialectRegistry&)                                                        >>> referenced by mlir-opt.cpp
>>>               tools/mlir-opt/CMakeFiles/mlir-opt.dir/mlir-opt.cpp.o:(main)
```

This allows to remove `asan_init_is_running` which likely had a data
race.

Simplifies https://github.com/llvm/llvm-project/pull/74086 and reduces a difference between platforms.

Reviewers: zacklj89, eugenis, dvyukov

Reviewed By: zacklj89, dvyukov

Pull Request: https://github.com/llvm/llvm-project/pull/74387

This patch adds the bind function to go with the socket function. It
also cleans up a lot of socket related data structures.

`importUndefined` is only used a couple of places and both of those
already handle `isPIC` separately.

The fix changes the way the validation script determines the qualified
name of a declaration. Inline namespaces without a reserved name are now
always part of the name. The Clang code only does this when the names
are ambigious. This method is often used for the operator""foo for UDLs.

Adjusted the newly flagged issue and removed a work-around in the test
code that is no longer required.

Fixes https://github.com/llvm/llvm-project/issues/72427

[PR 74670](https://github.com/llvm/llvm-project/pull/74670) added
support for merging locations at constant folding time. We have
discovered that in some cases, the number of locations grows so big as
to cause a compilation process to OOM. In that case, many of the
locations end up appearing several times in nested fused locations.

We add here a helper that always flattens fused locations in order to
eliminate duplicates in the case of nested fused locations.

DeviceTy provides an abstraction for "middle-level" operations that can
be done with a offload device. Mapping was tied into it but is not
strictly necessary. Other languages do not track mapping, and even
OpenMP can be used completely without mapping. This simply moves the
relevant code into the OpenMP/Mapping.cpp as part of a new class
MappingInfoTy. Each device still has one, but it does not clutter the
device.cpp anymore.

This patch resolves the following errors on z/OS:

error: no member named 'wait4' in the global namespace
error: no member named 'ru_maxrss' in 'rusage'
error: use of undeclared identifier 'strsignal'
error: Cannot get usage times on this platform
error: Cannot get malloc info on this platform

To avoid crashes with explicitly large objects.

I will clean up fast-isel with large objects/medium code model soon.

Extend `std::basic_string` tests to cover more buffer situations and
length in general, particularly non-SSO cases after SSO test cases
(changing buffers). This commit is a side effect of working on tests for
ASan annotations.

Related PR: https://github.com/llvm/llvm-project/pull/72677

If a function is declared with stack-protector, the compiler may try to
load the TLS. However, inside certain runtime functions, TLS may not be
available. This patch disables stack protectors for such routines to fix
the problem.

Closes #74487.

When you debug a binary and the change & rebuild and then rerun in lldb
w/o quitting lldb, the Modules in the Global Module Cache for the old
binary & .o files if used are now "unreachable". Nothing in lldb is
holding them alive, and they've already been unlinked. lldb will
properly discard them if there's not another Target referencing them.

However, this only works in simple cases at present. If you have several
Targets that reference the same modules, it's pretty easy to end up
stranding Modules that are no longer reachable, and if you use a
sequence of SBDebuggers unreachable modules can also get stranded. If
you run a long-lived lldb process and are iteratively developing on a
large code base, lldb's memory gets filled with useless Modules.

This patch adds a test for the mode that currently works:

(lldb) target create foo
(lldb) run
<rebuild foo outside lldb>
(lldb) run

In that case, we do delete the unreachable Modules.

The next step will be to add tests for the cases where we fail to do
this, then see how to safely/efficiently evict unreachable modules in
those cases as well.

Captures its responsibility a bit better.

1. A single dimension can either be blocked (with floordiv and mod pair)
or non-blocked. Mixing them would be invalid.
2. Block size should be non-zero value.

The AVX512F/AVX512BW checks had been removed despite still being used

Does transformations like
all_reduce(x) + all_reduce(y) -> all_reduce(x + y)

max(all_reduce(x), all_reduce(y)) -> all_reduce(max(x, y))
when the all_reduce element-wise op is max.

Added general rewrite pattern HomomorphismSimplification and
EndomorphismSimplification that encapsulate the general algorithm.
Made specialization for all-reduce with respect to
addf, addi, minsi, maxsi, minimumf and maximumf
in the Arithmetic dialect.

Add a configuration option to allow vector distribution with multiple
elements written by a single lane.

This is so that we can perform vector multi-reduction with multiple
results per workgroup.

Copy relocations and canonical PLT entries are for symbols defined in a
DSO. Currently we create them even for a `Defined`, possibly leading to
an output that won't work at run-time (e.g. R_X86_64_JUMP_SLOT
referencing a null symbol).
```
% cat a.s
.globl _start, main
.type main, @function
_start: main: ret

.rodata
.quad main
% clang -fuse-ld=lld -pie -nostdlib a.s
% readelf -Wr a.out

Relocation section '.rela.plt' at offset 0x290 contains 1 entry:
    Offset             Info             Type               Symbol's Value  Symbol's Name + Addend
00000000000033b8  0000000000000007 R_X86_64_JUMP_SLOT                        12b0
```

Report an error instead for the default `-z text` mode. GNU ld reports
an error in `-z text` mode as well.

Adding months to a year_month should wrap the year when the month
becomes greater than twelve or less than one.

This fixes the issue for year_month. Other classes with a year and month
do not have this issue. This has been verified and tests are added to
avoid possible regressions.

Also fixes some variable copy-paste errors in the tests.

Fixes https://github.com/llvm/llvm-project/issues/73162

When the main file is preprocessed and we change `MainFileName` to the
original source file name (e.g. `a.i => a.c`), the source manager does
not contain `a.c`, but we incorrectly associate the DIFile(a.c) with
md5(a.i). This causes CGDebugInfo::emitFunctionStart to create a
duplicate DIFile and leads to a spurious "inconsistent use of MD5
checksums" warning.

```
% cat a.c
void f() {}
% clang -c -g a.c  # no warning
% clang -E a.c -o a.i && clang -g -S a.i && clang -g -c a.s
a.s:9:2: warning: inconsistent use of MD5 checksums
        .file   1 "a.c"
        ^
% grep DIFile a.ll
!1 = !DIFile(filename: "a.c", directory: "/tmp/c", checksumkind: CSK_MD5, checksum: "c5b2e246df7d5f53e176b097a0641c3d")
!11 = !DIFile(filename: "a.c", directory: "/tmp/c")
% grep 'file.*a.c' a.s
        .file   "a.c"
        .file   0 "/tmp/c" "a.c" md5 0x2d14ea70fee15102033eb8d899914cce
        .file   1 "a.c"
```

Fix #56378 by disas...

Adds the std.compat module and add support for CMake 3.28.