Changes

[GlobalISel] CombinerHelper: Fix a bug in matchCombineCopy
Summary: When combining COPY instructions, we were replacing the
destination registers with the source register without checking register
constraints. This patch adds a simple logic to check if the constraints
match before replacing registers.
Reviewers: qcolombet, aditya_nandakumar, aemerson, paquette, dsanders,
Petar.Avramovic
Reviewed By: aditya_nandakumar
Subscribers: rovka, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D70616

[lldb] Fix TestFormattersSBAPI test
Summary: This test was broken in two ways:
* Using the wrong API (e.g.: format = instead of SetFormat)
* The hex checker was only checking "01" which will pass with 0x0000001
Reviewers: clayborg, lanza, wallace
Reviewed By: clayborg
Subscribers: lldb-commits
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D70884

[LLDB] Set the right address size on output DataExtractors from
ObjectFile
If filling in a DataExtractor from an ObjectFile, e.g. via the 
ReadSectionData method, the output DataExtractor gets the address size
from the m_data member.
ObjectFile's m_data member is initialized without knowledge about the
address size (so the address size is set based on the host's 
sizeof(void*), and at that point within ObjectFile's constructor, 
virtual methods implemented in subclasses (like GetAddressByteSize()) 
can't be called, therefore fix it up when filling in external 
DataExtractors.
This makes sure that line tables from executables with a different 
address size are parsed properly; previously this tripped up 
DWARFDebugLine::LineTable::parse for 32 bit executables on a 64 bit 
host, as the address size in the line table (4) didn't match the one set
in the DWARFDataExtractor.
Differential Revision: https://reviews.llvm.org/D70848

[MIBundles] Move analyzePhysReg out of MIBundleOperands iterator (NFC).
analyzePhysReg does not really fit into the iterator and moving it makes
it easier to change the base iterator.
Reviewers: evandro, t.p.northover, paquette, MatzeB, arsenm, qcolombet
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D70559

[clangd] repair mac tests for 88bccded8fa1

[lldb/CMake] Simplify logic for adding example Python packages (NFC)
This simplifies the CMake logic for adding the Python examples to the 
Python package. It unifies the use of create_python_package by adding 
the NOINIT option and removes the `target` argument, which is always
`finish_swig`.

[lldb/CMake] Add in_call_stack to the utilities package
A subset of the examples are shipped as python packages. Include the 
in_call_stack utility.

[Remarks][ThinLTO] Use the correct file extension based on the format
Since we now have multiple formats, the ThinLTO remark files should also 
respect that.

Revert "[clangd] repair mac tests for 88bccded8fa1"
Revert "[clangd] Try harder to find a plausible `clang` as argv0,
particularly on Mac."

[libcxx{,abi}] Emit deplibs only when detected by CMake
This is a followup to 35bc5276ca3.  It fixes the dependent libs usage in
libcxx and libcxxabi to link pthread and rt libraries only if CMake 
detects them, rather than based on explicit platform blacklist.
Differential Revision: https://reviews.llvm.org/D70888

[LLDB] [test] Try to fix the test from 7d019d1a3b when run on Windows.

Remove redundant file.

[PGO][PGSO] Add an optional query type parameter to
shouldOptimizeForSize.
Summary: In case of a need to distinguish different query sites for
gradual commit or debugging of PGSO. NFC.
Reviewers: davidxl
Subscribers: hiraditya, zzheng, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D70510

Automatically generated arm64-abi-varargs.ll . NFC

Remove extraneous semicolon.

Temporarily revert "build: avoid hardcoding the libxml2 library name" as
it breaks uses of llvm-config --system-libs and the follow-on commit
"build: avoid cached literals being linked against"
This reverts commits 340e7c0b77a7037afefe7255503afe362967b577 and 
340e7c0b77a7037afefe7255503afe362967b577.

Fix comment to more accurately describe C++ language requirements around
tail padding.
Summary: As of C++ core issue 43 (http://wg21.link/cwg43), which was
voted into the C++ working draft in 1999, it is not permissible to
memcpy a base class subobject, even if it's of POD type, so there is no
problem with reusing the tail padding of a base class. That issue was
voted into the standard in DR status, so it applies retroactively to
C++98 (and is in any case part of C++03).
So stop suggesting that AlwaysUseTailPadding mode is non-conforming.
Reviewers: rjmccall
Reviewed By: rjmccall
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D70923

[clang test] Do not assume default target
Summary: clang test Driver/darwin-opt-record.c assumes the default
target is x86_64 by its uses of the -arch x86_64 and -arch x86_64h and
thus fail on systems where it is not the case. Adding a target 
x86_64-apple-darwin10 reveals another problem: the driver refuses 2
-arch for an assembly output so this commit also changes the output to 
be an object file.
Reviewers: thegameg
Reviewed By: thegameg
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D70748

FileCheck IR output for blockaddress in new test
Minor improvement to a test added in 1ac700cdef787383ad49a

Place the "cold" code piece into the same section as the original
function
Summary: This cropped up in the Linux kernel where cold code was placed
in an incompatible section.
Reviewers: compnerd, vsk, tejohnson
Reviewed By: vsk
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D70925

[clang-tidy] Rewrite modernize-avoid-bind check.
This represents largely a full re-write of modernize-avoid-bind, adding 
significant new functionality in the process. In particular:
* Both boost::bind and std::bind are now supported
* Function objects are supported in addition to functions
* Member functions are supported
* Nested calls are supported using capture-init syntax
* std::ref() and boost::ref() are now recognized, and will capture by
reference.
* Rather than capturing with a global =, we now build up an individual
 capture list that is both necessary and sufficient for the call.
* Fixits are supported in a much larger variety of scenarios than
before.
All previous tests pass under the re-write, but a large number of new 
tests have been added as well.
Differential Revision: https://reviews.llvm.org/D70368

Fix failing testcase to check for the correct output

[X86] Model MXCSR for AVX instructions other than AVX512
Summary: Model MXCSR for AVX instructions other than AVX512
Reviewers: craig.topper, RKSimon
Subscribers: hiraditya, llvm-commits, LuoYuanke, LiuChen3
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D70875

[clang-scan-deps] do not skip empty #if/#elif in the minimizer to avoid
missing `__has_include` dependencies
This patch makes the minimizer more conservative to avoid missing
dependency files that are brought in by __has_include PP expressions
that occur in a condition of an #if/#elif that was previously skipped.
The __has_include PP expressions can be used in an #if/#elif either
directly, or through macro expansion, so we can't detect them at the
time of minimization.
Differential Revision: https://reviews.llvm.org/D70936

[DebugInfo] Support for debug_macinfo.dwo section in llvm and
llvm-dwarfdump.
This patch adds support for debug_macinfo.dwo section[pre-standardized] 
to llvm and llvm-dwarfdump.
Reviewers: probinson, dblaikie, aprantl, jini.susan.george, alok
Differential Revision: https://reviews.llvm.org/D70705
Tags: #debug-info #llvm

Recommit "[DWARF5]Addition of alignment atrribute in typedef DIE."
This revision is revised to update Go-bindings and Release Notes.
The original commit message follows.
This patch, adds support for DW_AT_alignment[DWARF5] attribute, to be
emitted with typdef DIE. When explicit alignment is specified.
Patch by Awanish Pandey <Awanish.Pandey@amd.com>
Reviewers: aprantl, dblaikie, jini.susan.george, SouraVX, alok, 
deadalinx
Differential Revision: https://reviews.llvm.org/D70111

[NFC][PowerPC] Add the inheritable and additional features to make the
processor definition more clear
The old processor design assume that, all the old processor's feature
must be inherited into future processor. That is not true as instruction
fusion or some implementation defined features are not inheritable.
What this patch did:
 * Rename the old "specific features" to "additional features" that keep
the new added inheritable features.
 * Use the "specific features" to keep those features only for specific
processor.
 * Add the "inheritable features" to keep all the features that
inherited from early processor.
Differential Revision: https://reviews.llvm.org/D70768

[LegalizeDAG] When expanding vector SRA/SRL/SHL add the new BUILD_VECTOR
to the Results vector instead of just calling ReplaceNode
The code that processes the Results vector also calls ReplaceNode and
makes ExpandNode return true.
If we don't add it to the Results node, we end up returning false from
ExpandNode. This causes ConvertNodeToLibcall to be called next. But
ConvertNodeToLibcall doesn't do anything for shifts so they just pass
through unmodified. Except for printing a debug message.
Ultimately, I'd like to add more checks to ExpandNode and 
ConvertNodeToLibcall to make sure we don't have nodes marked as Expand
that don't have any Expand or libcall handling.

[LegalizeDAG] Return true from ExpandNode for some nodes that don't have
expand support.
These nodes have a FIXME that they only get here because a Custom 
handler returned SDValue() instead of the original Op.
Even though we aren't expanding them, we should return true here to 
prevent ConvertNodeToLibcall from also trying to process them until the
FIXME has been addressed.
I'm hoping to add checking to ConvertNodeToLibcall to make sure we don't
give it nodes it doesn't have support for.

AvoidBindCheck.cpp: Fix GCC 5.3 build errors
It was failing with:
clang-tools-extra/clang-tidy/modernize/AvoidBindCheck.cpp:61:29: error:
declaration of ‘clang::tidy::modernize::{anonymous}::CaptureMode
clang::tidy::modernize::{anonymous}::BindArgument::CaptureMode’
[-fpermissive]
  CaptureMode CaptureMode = CM_None;
                            ^ 
clang-tools-extra/clang-tidy/modernize/AvoidBindCheck.cpp:38:6: error:
changes meaning of ‘CaptureMode’ from ‘enum
clang::tidy::modernize::{anonymous}::CaptureMode’ [-fpermissive]
enum CaptureMode { CM_None, CM_ByRef, CM_ByValue, CM_InitExpression };
     ^

AvoidBindCheck.cpp: Fix unused variables warning

[lldb][NFC] Remove ThreadSafeSTLVector and ThreadSafeSTLMap and their
use in ValueObjectSynthetic
Summary: ThreadSafeSTLVector and ThreadSafeSTLMap are not useful for
achieving any degree of thread safety in LLDB and should be removed
before they are used in more places. They are only used (unsurprisingly
incorrectly) in
`ValueObjectSynthetic::GetChildAtIndex`, so this patch replaces their
use there with a simple mutex with which we guard the related data
structures. This doesn't make ValueObjectSynthetic::GetChildAtIndex any
more thread-safe, but on the other hand it at least allows us to get rid
of the ThreadSafeSTL* data structures without changing the observable
behaviour of ValueObjectSynthetic (beside that it is now a few bytes
smaller).
Reviewers: labath, JDevlieghere, jingham
Reviewed By: labath
Subscribers: lldb-commits
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D70845