[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: More Ports for Debian's



wb2oyc@cyberenet.net  says:
>>I've got an intel based Debian box, and an m68k (Atari Falcon), also running
>>Linux. I'd liked to connect them using ppp or plip, but the bigger box has
>>all serial and paralel ports ocupied (mouse, modem, printer). So I decided
>>to purchase a card giving me an additional paralel and two serial ports. This
>>was the only card I could find after looking for a month, it's no-name, w/o
>>any kind of manual (just a jumper-charts regarding the i/o addresses),
>>and the vendor hasn't got an idea.
>>
>Ciccio,
>	One thing you might try, is to use one of the new serial ports as
>cua2 (DOS COM3) and disable the fourth one (at least for now).  The new
>parallel should be OK as LPT2.

You're right; the new parallel port works OK. The problem is that the
Triton II motherboard, it's bios , plug and pray facilities, or something else, is
changing the I/O and IRQ's everytime I set up something. My Soundblaster
was on IRQ 5, and suddently I had to recompile the kernel because I found it
again in IRQ10. Also, when I uninstalled the new card, I expected to get the
old configuration back; but at the first moment, I couldn't find neither the
mouse nor the modem port. I played a while, cat'ing things to the device, and
using chat and `dip -t', and at a certain moment I got the modem port where
it was. Then I started X, but there was no mouse. I kept playing the same way,
and finally found the mouse too on the original port. I didn't touch any of the
device files, setserial, which eventually could explain to me this behaviour
(because I wouldn't know how to do it).

> I use that very same setup on my machine
>here.

You also have an Atari Falcon connected?

>  COM4 is sometimes troublesome because of interference caused by the
>video cards, especially if your I/O cards are a mix of 8 and 16 bit variety.
>The reason is this: the older 8 (and many 16 bit for that matter) only decode
>the lower 12 bits of the I/O address!  Ever notice how the doc's will sometimes
>show them as x3e8!  That 'x' represents the top 4 bits of the I/O address, that
>for most older cards is "a don't care"!  Well, the newer cards DO care!  As 
>they fully decode the I/O address (all sixteen bits are significant).  This mix
>of cards that fully decode the address, and those that don't, create the issue.
>For example; one card may be set at 03e8, and another at 13e8, and guess what?
>The old card will see both addresses as the same!  Since it doesn't look at the
>upper four bits, it doesn't have any way of knowing that it is NOT the correct
>address (it thinks it is its own address anyway).  In my case, with my COM3, I
>set it to IRQ 5 to avoid conflict with the "shared" IRQ, which is often possible
>when running DOS, but not likely to be successful with any other intelligent 
>OS.

The tiny sheet doesn't mention any IRQ's, and IO addresses are specified
with only 3 hex-digits (so, there is no `x' nor `0'). I'm a newcomer to
hardware issues, but I thought, one device in IRQ4 and 0x03F8 shouldn't
conflict with another on IRQ4 and 0x?2F8. I'll try it right now, but anyway,
it would be a pitty to loose the second (forth) serial port. Isn't there a way
to avoid this?

TIA

--
Ciccio C. Simon
ciccio@arrakis.es



--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-user-REQUEST@lists.debian.org . Trouble? e-mail to Bruce@Pixar.com


Reply to: