This fixes a crash with instructions where multiple operands are
first-order-recurrences.

Summary:
I found the above named method hard to read because it had

a) many nested blocks,
b) one return statement at the end with some logic involved,
c) a duplicated while-loop with just small differences in it.

I decided to refactor this function by employing an early exit strategy.
In order to capture the logic in the return statement and to not have it
repeated more than once I chose to implement a very small lamda function
that captures all the variables it needs.
I also replaced the two while-loops with just one.

This is a non-functional change (NFC).

Reviewers: jdoerfert, teemperor

Reviewed By: teemperor

Subscribers: labath, teemperor, lldb-commits

Tags: #lldb

Differential Revision: https://reviews.llvm.org/D70774

This fixes the Utility/StatusTest.ErrorWin32 unit test on non-English locales.

Differential Revision: https://reviews.llvm.org/D70442

Differential Revision: https://reviews.llvm.org/D70448

Summary:
Pre gfx9 we need to scavenge a 64-bit SGPR to use as the carry out for an Add.
If only one SGPR was available this crashed when trying to scavenge another
32bit SGPR to materialize the offset.

Instead, reuse a 32-bit SGPR from the carry out as the offset register.

Also prefer to use vcc for the unused carry out when it is available.

Reviewers: arsenm, rampitec

Subscribers: kzhuravl, jvesely, wdng, nhaehnle, yaxunl, dstuttard, tpr, t-tye, hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D70614

Summary:
In rG643ac6c0, the syntax `ldraa x1, [x0]!` was added as an alias
for `ldraa x1, [x0, #0]!`. That syntax is less obvious in meaning, and
also will not be accepted by assemblers that haven't been updated yet.
So it would be better not to emit it as the preferred disassembly for
that instruction.

This change lowers the EmitPriority of the new alias so that the more
explicit syntax `[x0, #0]!` is preferred by the disassembler. The new
syntax is still accepted by the assembler.

Reviewers: ab, ostannard

Reviewed By: ostannard

Subscribers: kristof.beyls, hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D70813

On MIPS `zeroext` or `signext` can appear in the output.

Differential Revision: https://reviews.llvm.org/D70820

This method is only used in ClangASTContext.

Also removes the includes we only needed for the ClangASTContext RTTI check
in the CompilerDecl[Context].cpp files.

Now that CompilerDeclContext is a trivial class, Clang started warning
that this unused variable is in fact unused. Let's remove it.

The logic of this function was quite hard to follow. Replace it with a
much simpler, equivalent, implementation.

this is in preparation of a refactor of this method.

Summary:
A most of these tests create FileSpecs with a hardcoded style. Add
utility functions which create a file spec of a given style to simplify
things.

While in there add SCOPED_TRACE messages to tests which loop over
multiple inputs to ensure it's clear which of the inputs failed.

Reviewers: teemperor

Subscribers: lldb-commits

Tags: #lldb

Differential Revision: https://reviews.llvm.org/D70814

This reverts commit 373e2a4f.

This broke breakpoint setting.

Patch by Yannick Brehon!

Summary:
I found the above named method hard to read because it had

a) many nested blocks and
b) one return statement at the end with some logic involved.

I decided to refactor this function by employing an early exit strategy.
In order to capture the logic in the return statement and to not have it
repeated more than once I chose to implement a very small lamda function
that captures all the variables it needs.

This is a non-functional change (NFC).

Reviewers: jdoerfert

Subscribers: lldb-commits

Tags: #lldb

Differential Revision: https://reviews.llvm.org/D70774

Summary:
Also do an early return if the number of affected files > limit to save
some unnecessary FileURI computations.

Reviewers: ilya-biryukov

Subscribers: MaskRay, jkorous, arphaman, kadircet, usaxena95, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D70811

Summary: This would make go-to-def works on the cases like int A = abc^;

Reviewers: sammccall

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

Tags: #clang

Differential Revision: https://reviews.llvm.org/D70807

Summary: The current error message doesn't fit well for cross-file rename.

Reviewers: ilya-biryukov

Subscribers: MaskRay, jkorous, arphaman, kadircet, usaxena95, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D70809

That's declared directly above the actual definition, so it serves no use.

All other casting functions there are static, so this should
be too.

Summary:
I was seeing some failures on a test with slightly different instruction
ordering. Adding in some DAG directives solved the issue.

Change-Id: If5a3d3969055fb19279943bd45161bb70a3dabce

Subscribers: kzhuravl, jvesely, wdng, nhaehnle, yaxunl, tpr, t-tye, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D70531

Moving the different parts into their own functions without any additional
cleanup/refactoring, so this is NFC.

When we dump SHT_GNU_verdef section that has sh_link that references a non-existent section,
llvm-readobj reports a warning and continues dump, but llvm-readelf fails with a error.

This patch fixes the issue and opens road for futher follow-ups for
improving the printGNUVersionSectionProlog().

Differential revision: https://reviews.llvm.org/D70776

Windows on ARM always uses thumb mode, and doesn't have most of the
mechanisms that are used in e.g. ELF for distinguishing between arm
and thumb.

Differential Revision: https://reviews.llvm.org/D70796

COFF section names can either be stored truncated to 8 chars, in the
section header, or as a longer section name, stored separately in the
string table.

libunwind locates the .eh_frame section by runtime introspection,
which only works for section names stored in the section header (as
the string table isn't mapped at runtime). To support this behaviour,
lld always truncates the section names for sections that will be
mapped, like .eh_frame.

Differential Revision: https://reviews.llvm.org/D70745

Keep the existing special cases based on combinations of section name,
flags and sizes/offsets.

Differential Revision: https://reviews.llvm.org/D70778

Reviewers: labath, clayborg, shafik

Reviewed By: labath

Subscribers: JDevlieghere, lldb-commits

Tags: #lldb

Differential Revision: https://reviews.llvm.org/D70802

Currently we have 2 tests for testing versioning sections:
1) elf-versioninfo.test
2) elf-invalid-versioning.test

The first one currently checks how versioning sections are dumped +
how tools dump invalid SHT_GNU_verdef section.

The second despite of its name contains only tests for invalid SHT_GNU_verneed section.

In this patch I`ve renamed elf-invalid-versioning.test->elf-verneed-invalid.test,
and moved a few tests from elf-versioninfo.test to a new elf-verdef-invalid.test.

It will help to maintain these and a new tests for broken versioning sections.

Differential revision:

These need to be sign extended when loading into Edge addends.

Currently every time we encounter an indirect call of a known function,
we try to evaluate the inline cost of that function. In case of a
recursion, that evaluation never stops.

The solution I propose is to evaluate only the indirect call of the
function, while any further indirect calls (of a known function) will be
treated just as direct function calls, which, actually, never tries to
evaluate the call.

Fixes PR35469.

Differential Revision: https://reviews.llvm.org/D69349

f16 is promoted to f32 if it is not legal on the target.

Found while reviewing what else needed to be done for strict FP in
the softening code.

Update clang-tidy to use InitLLVM, like several other llvm tools that
were previously updated. On Windows, this allows clang-tidy to operate
on arguments containing characters which cannot be represented in the
system's ANSI code page such as filenames with Unicode characters.

Fixes bugzilla bug 43751.

Patch by Tristan Labelle.

Differential Revision: https://reviews.llvm.org/D70694