1.1 Threads

A thread is a stream of control that can execute its instructions independently. In the past, threads have sometimes inaccurately been referred to as NetWare processes. Process is more properly a concept used in UNIX* or Windows* NT. For operating systems in which the kernel does not support threads, a process is the unit of execution. In contrast, for a multithreaded system in which the kernel supports threads, a thread-rather than a process-is the unit of execution.

The NetWare OS allows NLM™ applications to establish multiple threads, each representing a single path of execution. An NLM usually contains at least one thread to accommodate the main function. (This is not true if the NLM is a library, such as CLIB.NLM.)

Two or more threads can be running concurrently-simultaneously in the midst of code execution-although only one thread can have control of the CPU at any given time. Concurrent threads can be executing the same code or different code. In a multiprocessing evironment, two or more threads can also be running parallel-that is, simultaneously running on different processors. Parallel threads are of course also running concurrently.

Historically, the NetWare OS has been a nonpreemptive ("good guy") scheduling environment. (NLMs written for NetWare 5.x and 6.x can be marked preemtable, as explained below.) When a thread gains control of the CPU, the thread remains in control until it has run to the end of its execution or until it calls a function that ’blocks’-that is, relinquishes control of the CPU. (Blocking functions are identified in the function reference manuals.)

No other thread can interfere with an active thread, regardless of priority. Only a hardware interrupt can temporarily interrupt a currently running thread.