14.2.1 Mapping Ada Tasks onto the Underlying Kernel Threads

GNAT's run-time support comprises two layers:

• GNARL (GNAT Run-time Layer)
• GNULL (GNAT Low-level Library)

In GNAT, Ada's tasking services rely on a platform and OS independent layer known as GNARL. This code is responsible for implementing the correct semantics of Ada's task creation, rendezvous, protected operations etc.

GNARL decomposes Ada's tasking semantics into simpler lower level operations such as create a thread, set the priority of a thread, yield, create a lock, lock/unlock, etc. The spec for these low-level operations constitutes GNULLI, the GNULL Interface. This interface is directly inspired from the POSIX real-time API.

If the underlying executive or OS implements the POSIX standard faithfully, the GNULL Interface maps as is to the services offered by the underlying kernel. Otherwise, some target dependent glue code maps the services offered by the underlying kernel to the semantics expected by GNARL.

Whatever the underlying OS (VxWorks, UNIX, Windows, etc.) the key point is that each Ada task is mapped on a thread in the underlying kernel. For example, in the case of VxWorks, one Ada task = one VxWorks task.

In addition Ada task priorities map onto the underlying thread priorities. Mapping Ada tasks onto the underlying kernel threads has several advantages:

• The underlying scheduler is used to schedule the Ada tasks. This makes Ada tasks as efficient as kernel threads from a scheduling standpoint.
• Interaction with code written in C containing threads is eased since at the lowest level Ada tasks and C threads map onto the same underlying kernel concept.
• When an Ada task is blocked during I/O the remaining Ada tasks are able to proceed.
• On multiprocessor systems Ada tasks can execute in parallel.

Some threads libraries offer a mechanism to fork a new process, with the child process duplicating the threads from the parent. GNAT does not support this functionality when the parent contains more than one task.