Identified by misc-unused-using-decls.

This is in advance of a following change which needs to the non-icmp API.

Line 1140 is a duplicate of line 1119; it tests the two-argument version
of std::hypot, whereas all the lines in this section are supposed to be
testing the C++17 three-argument version. Remove the erroneous duplicated line.

Split out of D116295.

First, let's check we get a TemplateDecl, before complaining about
where it might have been found.

Second, if it came from an unexpected place, show where that location is.

Reviewed By: ChuanqiXu

Differential Revision: https://reviews.llvm.org/D116164

VPWidenCanonicalIVRecipe does not create PHI instructions, so it does
not need to be placed in the phi section of a VPBasicBlock.

Also tidies the code so the WidenCanonicalIV recipe and the
compare/lane-masks are created in the header.

Discussed D113223.

Reviewed By: Ayal

Differential Revision: https://reviews.llvm.org/D116473

GCC on Windows ignores this flag completely [0] which some build systems sadly rely on when compiling for Windows using MinGW. The current behaviour of clang however is to error out as -fPIC & friends has no effect on Windows.

This patch instead changes the behaviour for MinGW to ignore the option for the sake of compatibility

Fixes https://github.com/llvm/llvm-project/issues/52947

[0] https://gcc.gnu.org/legacy-ml/gcc-patches/2015-08/msg00836.html

Differential Revision: https://reviews.llvm.org/D116485

These flags are specific to ELF, but are still accepted by GNU ld, even if it does not do anything. This patch adds them as ignored option for the sake of compatibility.

Part of https://github.com/llvm/llvm-project/issues/52947

Differential Revision: https://reviews.llvm.org/D116484

hasInterestingContent was introduced without a use on Sep 15, 2015 in
commit e5162dba.

The last use was removed on May 31, 2016 in commit
f9acacaa.

These functions seem to be unused for at least 1 year.

For .vf instructions, we don't need MF8 pseudos for f16. We don't
need MF8 or MF4 pseudos for f32. Or MF8, MF4, MF2 for f64.

Reviewed By: khchen

Differential Revision: https://reviews.llvm.org/D116437

Reviewed By: rriddle, Mogball

Differential Revision: https://reviews.llvm.org/D116253

Reviewed By: rriddle, Mogball

Differential Revision: https://reviews.llvm.org/D116252

Reviewed By: rriddle, Mogball

Differential Revision: https://reviews.llvm.org/D116251

Reviewed By: Mogball

Differential Revision: https://reviews.llvm.org/D116250

Reviewed By: rriddle, Mogball

Differential Revision: https://reviews.llvm.org/D116357

Reviewed By: rriddle, jpienaar

Differential Revision: https://reviews.llvm.org/D116356

Reviewed By: rriddle, Mogball

Differential Revision: https://reviews.llvm.org/D116355

Reviewed By: rriddle, Mogball

Differential Revision: https://reviews.llvm.org/D116354

Reviewed By: rriddle, Mogball

Differential Revision: https://reviews.llvm.org/D116249

Differential Revision: https://reviews.llvm.org/D116248

Differential Revision: https://reviews.llvm.org/D116247

Reviewed By: rriddle, Mogball

Differential Revision: https://reviews.llvm.org/D116245

Differential Revision: https://reviews.llvm.org/D116244

The dry-run of clang-tidy on the codebase with these enable were
well receive, and the codebase is "clean" (or almost) with respect
to these right now.

Identified with readability-redundant-member-init.

This isn't a "small" struct, flagged by Coverity.

thread_result_t is defined as unsigned on Windows.

This patch prevents clang-tidy from replacing 0 with nullptr.

Identified by readability-redundant-string-init.

This adds at extra check into ARMBaseInstrInfo::verifyInstruction to
verify the offsets used in addressing mode immediates using
isLegalAddressImm. Some tests needed fixing up as a result, adjusting
the opcode created from CMSE stack adjustments.

Differential Revision: https://reviews.llvm.org/D114939

This is a re-submission of 24d24055
without the hunks in HostNativeThreadBase.{h,cpp}, which break builds
on Windows.

Identified with modernize-use-nullptr.

This reverts commit 913457ac.

It again broke builds on Windows:

  lldb/source/Host/common/HostNativeThreadBase.cpp(37,14): error:
  assigning to 'lldb::thread_result_t' (aka 'unsigned int') from
  incompatible type 'std::nullptr_t'

This is a re-submission of 24d24055
without the hunk in HostNativeThreadBase.h, which breaks builds on
Windows.

Identified with modernize-use-nullptr.