Appeal 2007-2127
Reexamination Control No. 90/006,621
There are two ways in which a running thread gives up control
of the CPU. It can give up control of the CPU voluntarily, for
example when it finishes executing, or while waiting for the return of
a function call or the completion of a device I/O request. It would not
be wise, however, to design an operating system under the
presumption that all the active threads within the operating system
will cooperate among themselves to share processor cycles equally.
OS/2 will automatically preempt a running thread after a certain
number of clock cycles has passed (this value is called the time slice),
and go on to execute another. This makes OS/2 a preemptive
scheduling system.
OS/2 utilized both "priority" and "timesliced" preemptive scheduling
in 1988 as described in Young, Programmer's Guide to OS/2, page 180:
Specifically, a thread is preempted when either of the following
events occurs:
♦ Another thread, which has a higher priority, becomes ready to
run. (This can occur, for example, if a system call or hardware
interrupt routine frees some resource for which the higher-
priority thread has been waiting.)
♦ The thread's time slice expires. The system receives control
with each clock-tick interrupt, which gives the scheduler
36
Page: Previous 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 Next
Last modified: September 9, 2013