I only received a few replies to my query about mcp (ALM 2) multiplexors
failing after we upgraded from 4.0.3 to 4.1.1, but they fixed the problem.
The solution was provided by Stefan Mochnacki of the University of Toronto
(stefan@ca.utoronto.astro.centaur) :
> The essential checklist is:
>
> (a) have you checked /etc/ttytab to make sure the ttyh* devices are
> turned on, have gettys running, and have "local" set correctly
> so that they will ignore the h/w carrier detect?
> (b) do you have the mcp* devices configured into the kernel?
> (c) do /dev/ttyh* exist?
> (d) make sure you understand what ttysoftcar(8) is doing!
>
> As it turns out, I had to do one extra little bit of "magic", rm'ing
> /dev/mcp* and /dev/ttyh* and re-running
>
> /dev/MAKEDEV mcp0
>
> in order to actually get things running after a kernal rebuild (which
> shouldn't have been necessary, I still don't understand why the rest
> of this wouldn't work right until after I did that).
>
> The important part, however, not just for ALM-2 (aka MCP) boards, but
> for ALL serial lines, is to understand something about how ttysoftcar(8)
> works. I'm still not sure I do, but here's what I think I know:
>
> In general under SunOS4.1.1 it's not necessary to fiddle with kernal
> rebuilds or even eeprom settings in order to turn on or turn off
> carrier detect. What IS necessary is to properly declare the ports
> in /etc/ttytab. If "local" is there, software carrier detection
> is turned ON (and hardware carrier detection disabled). If the
> entry reads anything else, software carrier detection is turned
> OFF, all of that via "/usr/etc/ttysoftcar -a > /dev/null 2>&1"
> command in /etc/rc.local.
>
> If the device is not listed in /etc/ttytab, apparently it defaults
> to the flags value in the kernal configuration and/or eeprom setup.
>
>
> Having said all of that, I should also add that--under some conditions
> still not completely clear to me--ttysoftcar may not always work.
> In particular, in using it directly (e.g., ttysoftcar [-y|-n]
> /dev/ttymumble)
> the command sometimes seems to hang. Also, having "remote" instead
> of "local" does not always seem to keep software carrier detection
> off. On several systems here with modems (e.g., on ttya), I've
> found myself having to add:
>
> /usr/etc/ttysoftcar -n /dev/ttya
>
> ..at the end of /etc/rc.local in order to actually get hardware
> carrier detect.
The problem was that under 4.1.1 our setup needs software carrier detection,
and I had just copied /etc/ttytab over from the 4.0.3 system. The solution was
just to specify all the ttyh's and ttyi's in /etc/ttytab as being 'local' and
rebooting.
Thanks also to :
Robert Evans : robert@cm.cf.ac.uk
and
Paul Humphreys : paul@hydres.uucp
+------------------------------------------------------------------------+
|Nick Murray <nam@uk.ac.aberdeen.computing-science> |
|Computer Officer phone (+44 224) 272303 |
|Dept.of Computing Science Fax (+44 224) 487048 |
|University of Aberdeen |
|King's College |
|Aberdeen AB9 2UB |
|Scotland UK |
+------------------------------------------------------------------------+
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:15 CDT