Atomics: Cleanup, move defines closer to usage

There is a special defines block needed for ARM on Linux. Move it from
public header to an implementation file.

No functional changes.

This is a non-functional part of original fix done by Brecht in D9577.

Atomics: Fix missing 64bit implementation for 32bit platforms

Blender uses 64bit atomics to manipulate SessionUUID, and these atomics
were not defined on any of 32bit platforms.

While official support is limited to 64bit platforms only, the code
should not make assumptions about bitness or endianess, in terms that
there should be codepaths and fallback (or provision of them) for 32bit
platforms.

This change makes 64bit atomic functions defined for all platforms.
The atomic_test was compiled and successfully tested on i686 and armv7l
platforms. The rest of compilation process of Blender will be very
tedious, so that was not done.

This change is essential, but not necessarily enough to make Blender
compilable on i686 (ability to compile Blender on 32bit platforms was
lost during the 2.91 development).

This is a functional part of original fix done by Brecht in D9577.

Tracking: Cleanup, de-duplicate implementation of marker lookup

The logic was duplicated.

Should be no functional changes. The modified function is expected
to give same exact results for all inputs.

On the "caching last-used track" topic. The code was using last_marker
to allow faster lookup of marker closest to the frame. With this
change it is still the case since the BKE_tracking_marker_get() does
cache last used marker.

Fix T77261: Multires creates spikes when subdividing ngons

The spikes were caused by non-initialized tangent matrix used during
smoothing process. The reason tangent matrix was not initialized was
because wrong usage of API: n-gons should pass corner of 0 to the
matrix construction function.

Corrected usage of the API and added assert() to help catching such
kind of non-initialized issues easier.

Tracking: Cleanup, typos and spelling in DNA comments

Tracking: Refactor, move track allocation to helper function

The caller is still responsible for allocating list of markers,
but the track allocation and initialization can now be reused.

Currently no functional changes, preparing for an upcoming
development.

Tracking: Cleanup, clarify comments in tracking settings

Make unit more explicit.

Ideally would be "embedded" into the field name itself, but this will be
more involved change.

Tracking: Cleanup, make units explicit in new track creation

Tracking: Cover markers access with basic tests

Just a beginning of tests coverage, related on the planned upcoming
development.

Atomic: Cover with unit tests

Cover all atomic functions with unit tests.

The tests are quite simple, nothing special so far. The goal is to:

- Make sure implementation exists.
- Implementation behaves the same way on all platforms
(We had issue when MSVC and GCC were behaving differently in the
past).
- Proper bitness is used for implementation and non-fixed-size
function implementation.

The tests can be extended further to make sure, for example, that
CAS operations do not cast arguments to a more narrow type for
comparison. Considering it a possible further improvement, as it is
better be done being focused on that specific task.

There is an annoying ifdef around 64bit implementation, which uses
same internal ifdef as the header does. This check is aimed to be
removed, so is easier to simply accept such duplication for now.

The tests seems somewhat duplicate for signed/unsigned variants and
things like this. The reason for that is to keep test code as simple
as possible: attempting to do something smart/tricky in the test code
often causes the test code to be a subject of being covered with its
own unit tests.

Differential Revision: https://developer.blender.org/D9590