Translating between JSON objects and C++ strutctures is common.
From experience in clangd, fromJSON/ObjectMapper work well and save a lot of
code, but aren't adopted elsewhere at least partly due to total lack of error
reporting beyond "ok"/"bad".

The recently-added error model should be rich enough for most applications.
It requires tracking the path within the root object and reporting local
errors at appropriate places.
To do this, we exploit the fact that the call graph of recursive
parse functions mirror the structure of the JSON itself.
The current path is represented as a linked list of segments, each of which is
on the stack as a parameter. Concretely, fromJSON now looks like:
  bool fromJSON(const Value&, T&, Path);

Beyond the signature change, this is reasonably unobtrusive: building
the path segments is mostly handled by ObjectMapper and the vector<T> fromJSON.
However the root caller of fromJSON must now create a Root object to
store the errors, which is a little clunky.

I've added high-level parse<T>(StringRef) -> Expected<T>, but it's not
general enough to be the primary interface I think (at least, not usable in
clangd).

All existing users (mostly just clangd) are updated in this patch,
making this change backwards-compatible is a bit hairy.

Differential Revision: https://reviews.llvm.org/D88103

Currently, findUnwindSectionsByPhdr is slightly micro-optimized for the
case where the first callback has the target address, and is otherwise
very inefficient -- it decodes .eh_frame_hdr even when no PT_LOAD
matches the PC. (If the FrameHeaderCache is enabled, then the
micro-optimization only helps the first time unwind info is looked up.)

Instead, it makes more sense to optimize for the case where the
callback *doesn't* find the target address, so search for a PT_LOAD
segment first, and only look for the unwind info section if a matching
PT_LOAD is found.

This change helps on an Android benchmark with 100 shared objects,
where the DSO at the end of the dl_iterate_phdr list throws 10000
exceptions. Assuming the frame cache is disabled, this change cuts
about 30-40% off the benchmark's runtime.

Reviewed By: compnerd, saugustine, #libunwind

Differential Revision: https://reviews.llvm.org/D87881

dl_iterate_phdr is used to search for unwind info provided by either
PT_GNU_EH_FRAME or PT_ARM_EXIDX. Most of the code between the two is
the same, so combine them, and factor out what's different into
checkForUnwindInfoSegment.

Details:
 - The FrameHeaderCache can now be enabled for ARM EHABI.
 - findUnwindSectionsByPhdr now finds the last PT_ARM_EXIDX rather than
   the first. There should only be one segment.
 - The dso_base and text_segment_length fields of UnwindInfoSections
   are now needed for dl_iterate_phdr when using EHABI, to hold the
   low and high PC values for a cache entry.

Reviewed By: compnerd, danielkiss, #libunwind, saugustine

Differential Revision: https://reviews.llvm.org/D87880

When an error occurs processing a JSON object, seeing the actual
surrounding data helps. Dumping just the node where the problem
was identified can be too much or too little information.

printErrorContext() shows the error message in its context, as a comment.
JSON values along the path to the broken place are shown in some detail,
the rest of the document is elided. For example:

```
{
  "credentials": [
    {
      "username": /* error: expected string */ 42,
      "password": "secret"
    },
    { ... }
  ]
  "backups": { ... }
}
```

Differential Revision: https://reviews.llvm.org/D88103

This is in preparation for supporting -debugify-each, which adds a debug
info pass before and after each pass.

Switch VerifyEach to use this.

Reviewed By: ychen

Differential Revision: https://reviews.llvm.org/D88107

This seems to fit the CGSCC updates model better than calling
addNewFunctionInto{Ref,}SCC() on newly created/outlined functions.
Now addNewFunctionInto{Ref,}SCC() are no longer necessary.

However, this doesn't work on newly outlined functions that aren't
referenced by the original function. e.g. if a() was outlined into b()
and c(), but c() is only referenced by b() and not by a(), this will
trigger an assert.

This also fixes an issue I was seeing with newly created functions not
having passes run on them.

Ran check-llvm with expensive checks.

Reviewed By: asbirlea

Differential Revision: https://reviews.llvm.org/D87798

Reviewed By: asbirlea

Differential Revision: https://reviews.llvm.org/D88128

This error model should be rich enough for most applications. It comprises:

- a name for the root object, so the user knows what we're parsing
- a path from the root object to the JSON node most associated with the error
- a local error message

This can be presented as an llvm::Error e.g.
  "expected string at ConfigFile.credentials[0].username"

It's designed to be cheap: Paths are a linked list of lightweight
objects on the stack. No heap allocations unless errors are encountered.

A subsequent commit will make use of this in the JSON-to-object
translation facilities: fromJSON and ObjectMapper.
However it's independent of these and can be used for e.g. validation alone.

Another subsequent commit will support showing the error in its context
within the parsed value.

Differential Revision: https://reviews.llvm.org/D88103

Differential Revision: https://reviews.llvm.org/D88054

[X86] Add a memory clobber to the bittest intrinsic inline asm. Get default clobbers from the target

I believe the inline asm emitted here should have a memory clobber since it writes to memory.

It was also missing the dirflag clobber that we use by default along with flags and fpsr. To avoid missing defaults in the future, get the default list from the target

Differential Revision: https://reviews.llvm.org/D88121

Remove all spurious `HelpHidden` flags from  `lld/MachO/Options.td`. Add test for `HelpHidden` to `warnIfUnimplementedOption()` so that the empty `// handled elsewhere` case is unnecessary.

Reviewed By: #lld-macho, int3, smeenai

Differential Revision: https://reviews.llvm.org/D88160

This isn't standard JSON, but is a popular extension.
It will be used to show errors in context, rendering pseudo-json for humans.

Differential Revision: https://reviews.llvm.org/D88103

Previously, if a floating-point type was legal, but FNEG wasn't legal,
we would use FSUB.  Instead, we should use integer ops, to preserve the
semantics.  (Alternatively, there's a compiler-rt call we could use, but
there isn't much reason to use that.)

It turns out we actually are still using this obscure codepath in a few
cases: on some targets, we have "legal" floating-point types that don't
actually support any floating-point operations.  In particular, ARM and
AArch64 are using this path.

The implementation for SelectionDAG is pretty simple because we can
reuse the infrastructure from FCOPYSIGN.

See also 9a3dc3e6, the corresponding change to type legalization.

Also includes a "bonus" change to STRICT_FSUB legalization, so we can
lower a STRICT_FSUB to a float libcall.

Includes the changes to both LegalizeDAG and GlobalISel so we don't have
inconsistent results in the future.

Fixes https://bugs.llvm.org/show_bug.cgi?id=46792 .

Differential Revision: https://reviews.llvm.org/D84287

With the exception of VECREDUCE_ADD, there are no NEON instructions to support vector of i64 reductions. This patch removes the Custom lowerings for those and adds some test coverage to confirm.

Differential Revision: https://reviews.llvm.org/D88161

This recommits 829d14ee.

The patch was reverted due to a regression in some CUDA app
which was thought to be caused by this patch. However, investigation
showed that the regression was due to some other issues, therefore
recommit this patch.

This patch implements the vec_[all|any]_[eq | ne | lt | gt | le | ge] builtins for vector signed/unsigned __int128.

Differential Revision: https://reviews.llvm.org/D87910

This patch implements Vector signed/unsigned __int128 overloads for the comparison builtins.

Differential Revision: https://reviews.llvm.org/D87804

The minidump-sysroot test I added in commit 20f84257 compares two paths
using a string comparison.  This causes the Windows buildbot to fail
because of mismatched forward slashes and backslashes.  Use
os.path.normcase to normalize before comparing.

This also adds some bare-bones documentation for the attribute rather
than leaving it undocumented.

Add the functionality to lower frecpx for passthru variant

Reviewed By: paulwalker-arm

Differential Revision: https://reviews.llvm.org/D88032

This reverts commit 0caad9fe.
This reverts commit c96d0cce.

Causes linker errors which were not fixed by the subsequent commit
either:

/home/nikic/llvm-project/compiler-rt/lib/asan/asan_rtl.cpp:503: error: undefined reference to '__asan::InstallAtExitCheckLeaks()'

Fix a potential UB in `appendSignedDecimal` (with -INT64_MIN) by making
it a special case.

Fix the terrible test cases for `isOwned`: I was pretty sloppy on those
and used some stack & static variables, but since `isOwned` accesses
memory prior to the pointer to check for the validity of the Scudo
header, it ended up being detected as some global and stack buffer out
of bounds accesses. So not I am using buffers with enough room so that
the test will not access memory prior to the variables.

With those fixes, the tests pass on the ASan+UBSan Fuchsia build.

Thanks to Roland for pointing those out!

Differential Revision: https://reviews.llvm.org/D88170

The `if (0)` isn't necessarily optimized out so as not to create
a link-time reference to LSan runtime functions that might not
exist.  So use explicit conditional compilation instead.

Reviewed By: phosek

Differential Revision: https://reviews.llvm.org/D88173

This patch is the initial support for the Local Dynamic Thread Local Storage
model to produce code sequence and relocation correct to the ABI for the model
when using PC relative memory operations.

Differential Revision: https://reviews.llvm.org/D87721

The IRSimilarityCandidate is a container to hold a region of
IRInstructions and offer interfaces for the starting instruction, ending
instruction, parent function, length.  It also assigns a global value
number for each unique instance of a value in the region.

It also contains an interface to compare two IRSimilarity as to whether
they have the same sequence of similar instructions.

Tests for whether the instructions are similar are found in
unittests/Analysis/IRSimilarityIdentifierTest.cpp.

Recommit of: 4944bb19

Differential Revision: https://reviews.llvm.org/D86970

Differential Revision: https://reviews.llvm.org/D87947

Differential Revision: https://reviews.llvm.org/D88129

If d8 is saved, the fp is not actually adjacent to the SVE
spills/allocations.  Fix the offset calculation to account for this.

Differential Revision: https://reviews.llvm.org/D88117

This tweaks the generated code for parsing attributes with a custom
directive to call `addAttribute` on the `OperationState` directly,
and adds a newline after this call. Previously, the generated code
would call `addAttribute` on the `OperationState` field `attributes`,
which has no such method and fails to compile. Furthermore, the lack
of newline would generate code with incorrectly formatted single line
`if` statements. Added tests for parsing and printing attributes with
a custom directive.

Reviewed By: mehdi_amini

Differential Revision: https://reviews.llvm.org/D87860

Reviewed By: thakis

Differential Revision: https://reviews.llvm.org/D88056

Fuchsia's system libraries are instrumented and use the lsan
allocator for internal purposes.  So leak checking needs to run
after all atexit hooks and after the system libraries' internal
exit-time hooks.  The <zircon/sanitizer.h> hook API calls the
__sanitizer_process_exit_hook function at exactly the right time.

Reviewed By: vitalybuka, phosek

Differential Revision: https://reviews.llvm.org/D86171

This reverts commit 5dd1b6d6.

All of the callers already have an Instruction *. Many of them
from a dyn_cast.

Also update the OperationData constructor to use a Instruction&
to remove a dyn_cast and make it clear that the pointer is non-null.

Differential Revision: https://reviews.llvm.org/D88132

If the control is constant we can figure out exactly which bits
of the input are demanded.

Differential Revision: https://reviews.llvm.org/D88072

I'm not sure what this means, but the order in which we try
the matches makes a difference on at least 1 regression test...