killing a login on Zilog

[George Robbins]

In article <818 at hropus.UUCP> jrw at hropus.UUCP (Jim Webb) writes:
>> Every once in a while, a "login" process runs forever on our Zilog
>> 8000/31 computer.  I cannot kill this process, and nobody can use
>> that port.
>
>  vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
>> I know the cause of the problem (ie. what causes the login to get
>> stuck).  But, assuming that I can't make that go away, I need to
>> find out how to kill this process.
>
>Care to let us in on it?  Usually, non-killable processes are sleeping
>on an event at a priority that does not recognize signals, so you have
>to cause the wakeup.  If on a terminal, sometimes the brute force method
>of plugging and unplugging the cable causes the hardware to awaken the
>process.
>-- 
>Jim Webb             "Out of phase--get help"          ...!ihnp4!hropus!jrw
>    "Make sure comments and code agree.  If not, write a man page..."

I assume that this is the XON/XOFF race problem with Zilogs's ICP "intelligent"
communications processor board.  Call it firmware confusion.  It was added to
Zilog's known problem list maybe a year ago, but unfortunatly Zilog never caught
on to the idea that bugs, especially gross, embarrassing/must reboot type bugs
are supposed to be fixed!  Perhaps that's one reason the systems division is
all but down the tubes...

People interested in Zilog S8000/ZEUS issues are invited to join the mail-zilog
mailing list - {seismo|ihnp4|rutgers|allegra}!cbmvax!mail-zilog-request

-- 
George Robbins - now working for,	uucp: {ihnp4|seismo|rutgers}!cbmvax!grr
but no way officially representing	arpa: cbmvax!grr at seismo.css.GOV
Commodore, Engineering Department	fone: 215-431-9255 (only by moonlite)