Re: RE: keyboard freeze with one percent CPU load

[Marvin Germain <megermain@earthlink.net>]

Hi, Rich.

    I calculated the CPU load by timing the ISR (50 us, not ms),

and comparing this with the time between interrupts.  The computations

are actually commented out for now, so all the second module is doing

is pthread_suspend_np().  The operating system is definately not frozen,

because I can regain control by pulling the plug on the external

data clock, thereby stopping the interrupts.  I think the problem is

with the scheduler, based on the results in my follow-up post. To re-

iterate, I put a call to rtl_schedule() in a periodic task, which

seemed to fix the problem.  This implies that user-space tasks are

simply not being scheduled when the interrupts arrive once every five

mill-seconds.

Marvin
-
*********************************************
*         Dr. Marvin E. Germain             *
*         Zygo Corporation - Tucson AZ      *
*         megermain@earthlink.net           *
*********************************************