Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucbvax!pasteur!ames!amelia!izen
From: izen@amelia.nas.nasa.gov (Steven H. Izen)
Newsgroups: comp.unix.i386
Subject: Re: Third serial port
Message-ID: <4332@amelia.nas.nasa.gov>
Date: 22 Dec 89 05:52:58 GMT
References: <12978@cit-vax.Caltech.Edu> <%??HY=@b-tech.uucp>
Reply-To: izen@cwru.cwru.edu (Steven H. Izen)
Organization: Case Western Reserve University
Lines: 40

In article <%??HY=@b-tech.uucp> zeeff@b-tech.ann-arbor.mi.us (Jon Zeeff) writes:
>>I am having trouble installing a third serial port in my 386 running
>>ISC Unix/386.  I have the X5 update (new serial driver).  I have the
>>card set to IO address 3E8 (COM3), IRQ 2, and I have added appropriate
>>lines to /etc/conf/sdevice.d/asy and remade the kernel.

>>Unfortunately, I don't seem to be getting any interrupts from that
>>device.  All output seems to be at about one character per second,

>>Does anybody know if it is possible to use IRQ 2 for a serial port?

>The AT bus generally has many unused interrupt lines.  All you have to do is
>run a jumper wire from your serial card to something that has a 16 bit
>connector.  Try INT 10 or 15 (you need an AT tech ref manual to show you
>what pin).

Whoa.  I tried that with my second serial port under 386/ix.  I changed 
*everything* that needed to be changed to point to the new interrupt 
including some stuff in the appropriate header files.  It didn't work! The
kernel built o.k. and I saw characters going out to the modem, but incoming
characters were lost, and a message which was something like "stray channel
interrupt" (I forget the exact message) was sent to the console.  Nothing I
tried would work with the 386/ix asy driver.  Admittedly this was 2.0.1, before
i installed the x5 update and 2.0.2.  

My solution: I freed up an an interrupt by doing the interrupt remapping trick
suggested above to my second parallel port.  All that I needed to do was put
the appropriate irq # into the device file and rebuild the kernel.  It works
quite well.  The lp device driver on 386/ix is robust and seems to work quite
well! (as opposed to the infamous asy drivers).  I know the author of the lp
drivers reads the net, and he is to be congratulated for a job well done.
(sorry, I forgot your name, otherwise I'd mention it.)



-- 
Steve Izen: {sun,uunet}!cwjcc!skybridge!izen386!steve  / Quote corner:
or steve%izen386.uucp@skybridge.scl.cwru.edu          / 
or izen@cwru.cwru.edu	   /-------------------------/ My second bike is a car.
                           | The problem is that I *was* paying attention.