*BSD News Article 1848


Return to BSD News archive

Path: sserve!manuel!munnari.oz.au!hp9000.csc.cuhk.hk!uakari.primate.wisc.edu!sdd.hp.com!elroy.jpl.nasa.gov!usc!sol.ctr.columbia.edu!destroyer!uunet!decwrl!netcomsv!mork!hasty
From: hasty@netcom.com (Amancio Hasty Jr)
Newsgroups: comp.unix.bsd
Subject: Re: COM driver and X11
Keywords: COM X11
Message-ID: <9d8lvff.hasty@netcom.com>
Date: 9 Jul 92 04:38:43 GMT
References: <cproto.710554678@marsh>
Organization: Netcom - Online Communication Services  (408 241-9760 guest)
Lines: 42

In article <cproto.710554678@marsh> cproto@marsh.cs.curtin.edu.au (Computer Protocol) writes:
>
>I'm running 386BSD (17/03/92) and experience silo overflows when
>running a kermit filetransfer (2400 b/s) in background and anything
>(even just a ls) in foreground. I get the silo overflows both with
>the released com driver as well as with the latest post which is
>supposedly fixing this problem. My COM port is implemented with a
>NS16450 so I can't use the 16 byte FIFO supported in the latest
>driver. But even with a NS16550 with the FIFO I doubt whether the
>driver would keep up due to the TTY stuff it has to go through for
>every received character (at 9600 b/s or higher).
>
>BTW I also experience the problem running X and moving the serial
>mouse while just launching some X application.
>
Do you have my ast patches?
Also, bear in mind that disk i/o is done at a high IPL level. So if you
do a lot of disk i/o you will get com overflow for sure.

To make matters worse there are couple areas in the kernel which don't
restore back the ipl!

>I think the driver should be really split into two parts. A lower 
>level driver which is handling the receive interrupts at a high
>priority and a very short execution path. This part should write
>received characters plus status into a circular buffer and then
>cause some software interrupt to invoke the higher level driver
>to do all the fancy TTY bits.
>
I am sick and tire of trying to make the com driver work, I arrived
at the same conclusion as you. It will be nice if the rest of the
set ipl and restore ipls in the system be balanced; also to lower
the ipl of the disk i/o.

>Before I start on this please let me know what you guys think ?!?
>
Let me know if I could help you.

>Regards - Tibor Sashegyi (cproto@abel.cs.curtin.edu.au)
>
Cheers,
Amancio