Reentrancy and Thread-safety

This documents the state of the thread support in the ML and the toolbox libraries.

Within the ML and its libraries, the following definitions are used:

Reentrancy

A reentrant function can be called simultaneously by multiple threads provided that each invocation of the function references unique data. A class is said to be reentrant if each and every one of its functions can be called simultaneously by multiple threads on different instances of the class

Thread-Safety

A thread-safe function can be called simultaneously by multiple threads when each invocation references shared data. All access to the shared data is serialized. A class is said to be thread-safe if the functions can be called by different threads on the same instance.

Thread-safe Classes and Functions in the ML

The following classes/functions are reentrant:

All classes in the MLLinearAlgebra library are reentrant (e.g. Vector2, Vector3, ..., Matrix2, Matrix3, ...)
Reentrant Classes

The following classes/functions are thread-safe:

ML_PRINT_ERROR, ML_PRINT_WARNING etc. may be used in any thread (as long as the error output handler that is registered with the ML supports this)
Thread-safe Classes and Functions

The following classes/functions are not thread-safe:

ML_TRACE_IN* macros, see TraceBuffer for details (although it is crash-safe to use them from other threads)
The Host, which may only be accessed from the main thread, unless stated differently in individual methods
The Field class and its notification mechanism, which may be used from the main thread only
The Module class, which may be used from the main thread only, with the exception of some methods that are documented to be thread-safe.
Non-thread-safe Classes and Functions