Date: Sun, 12 Jan 1997 14:54:53 -0800 (PST) From: John-Mark Gurney <jmg@nike.efn.org> To: J Wunsch <j@uriah.heep.sax.de> Cc: freebsd-bugs@freefall.freebsd.org, freebsd-gnats-submit@freebsd.org Subject: bin/2466 Re: bin/2466: telnet sleeps too much :-) Message-ID: <Pine.NEB.3.95.970112144317.16860i-100000@hydrogen.nike.efn.org> In-Reply-To: <199701121840.KAA10082@freefall.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, 12 Jan 1997, J Wunsch wrote:
oh... this is unrelated... but gnats-submit doesn't pass along the
reply-to: that you (Joerg) normally have.. just noticed this..
it will probably be missing my 'Reply-to: John-Mark Gurney
<gurney_j@efn.org>' also...
> The following reply was made to PR bin/2466; it has been noted by GNATS.
>
> From: j@uriah.heep.sax.de (J Wunsch)
> To: andrew@ugh.net.au
> Cc: freebsd-gnats-submit@FreeBSD.ORG
> Subject: Re: bin/2466: telnet sleeps too much :-)
> Date: Sun, 12 Jan 1997 19:10:12 +0100
>
> As andrew@ugh.net.au wrote:
>
> > If you leave telnet idle and then type something it will say "sleep(5)
> > from telnet, after select". It will continue to work but it messes up
> > the screen.
>
> > >How-To-Repeat:
> > telnet somewhere and leave it idle for a while, come back to it and type.
>
> No, this is not generally repeatable.
I can do it over here... just like he says... I sent a message to
hackers reciently about this... but didn't get a response... here is the
message:
-----<begin message>------
well.. I have a diskless machine (386/25sx 6meg ram, swap also nfs) that
regularly gets this message... I notice there was a comment about this...
that he (or whoever added it) doesn't like it... I just recompiled telnet
to give me the errno when this happens...
when it does it gives 14 == EFAULT... it only seems to happen after I've
let it be for a while... any body have any suggestions? should I just
remove the sleep and be gone with it? thanks for your suggesstions...
ttyl..
-------<end message>------
> I suspect something special with your network setup. Perhaps you add
> a debug printf to see what the actual error code is.
as the message states... I have... EFAULT... very wierd is all I can
say...
> The code in question (modulo some tn3720 hack) is:
>
> if ((c = select(16, &ibits, &obits, &xbits,
> (poll == 0)? (struct timeval *)0 : &TimeValue)) < 0) {
> if (c == -1) {
> /*
> * we can get EINTR if we are in line mode,
> * and the user does an escape (TSTP), or
> * some other signal generator.
> */
> if (errno == EINTR) {
> return 0;
> }
> /* I don't like this, does it ever happen? */
> printf("sleep(5) from telnet, after select\r\n");
> sleep(5);
> }
> return 0;
> }
>
> ...so the sleep only happens after select() returned -1 (nothing
> found) but without EINTR. This is a ``Cannot happen'' situation.
> Perhaps you can investigate a little why it happens in your
> environment. (Mind you, many people leave telnet sessions idle for
> days.)
yeh... I know... right now I'm trying to get a tcpdump of a failed select
to see if that would possibly help...
ok... I got it... the dumped file is 73k... I just managed to reproduce
it... it seems to happen when the process (telnet) gets swapped out... the
machine is a 386/25sx w/ 6megs ram netbooted with swap over nfs... so if
I do some stuff on another console the chances of telnet swapping out are
very high...
I hope this helps... ttyl...
John-Mark
gurney_j@efn.org
http://resnet.uoregon.edu/~gurney_j/
Modem/FAX: (541) 683-6954 (FreeBSD Box)
Live in Peace, destroy Micro$oft, support free software, run FreeBSD (unix)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.NEB.3.95.970112144317.16860i-100000>