Re: [rtl] 8254 programming

[Nallur Ramesh <nalluri@ladron.cs.nmt.edu>]

On Mon, 25 May 1998, Paolo Mantegazza wrote:

> Ciao,
> - what is the need of using outb_p, in place of the simpler and faster
> outb, in programming the 8254?

IO is horrible and if we dont give the proper delays sometimes it just 
hangs the system. Since we reprogram the counter everytime we might need
to put delays at the appropriate places. We will look into this and see if
we can get away with non-delayed versions then surely outb s' will be used
instead of outb_ps'. 

> - why in rtl the read-back approach is preferred to the simple
> latched-read of 8254 ports, even if just one port is latched and there
> is no need to verify the status?

True! We need not use the read-back approach, but I dont think it is
slower in any way than the latch-read approach. But u are right, there is
no need to use the read-back approach.

> - when oneshot mode is programmed for a count of N does the interrupt
> appear after N or after N+1 counts?

True, N+1 counts. Need to make the changes accordingly. 

Thanks a lot for u'r input. 

Hoping to see some more mails like these.

Bye.






--- [rtl] ---
For more information on Real-Time Linux see:
http://rtlinux.cs.nmt.edu/