Re: strange hangs in 2.0.32 (WCHAN = do_down)

Olaf Titz (olaf@bigred.inka.de)
Mon, 28 Sep 1998 20:38:36 +0200


> server because of over 28 processes stacked up over a three day period.
> All the same:
> FLAG UID PID PPID PRI NI SIZE RSS WCHAN STA TTY TIME COMMAND
> 40 0 5713 1 0 0 920 284 do_down D ? 0:00 syslogd
> 40 0 5725 1 0 0 920 284 do_down D ? 0:00 syslogd

I had this problem, or one which looks similar at least, on 2.1.122
yesterday. The machine uses a serial console; the terminal is an old
laptop PC which I keep switched off most of the time (because of the
horrible noise it makes). When I turned on the terminal, the processes
resumed their normal operation.

On my system it was "innd" and several "ctlinnd" processes which were
hanging. ctlinnd uses AF_UNIX sockets, as does syslogd. I see a
pattern here.

So is there any way how a serial interface can lock up unix domain
sockets? Any semaphore indirectly shared by those paths? And while
we're at it, is there any way I can set the serial interface for the
console to ignore modem control lines and stay that way?

olaf

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/