Date: Wed, 20 Dec 1995 19:18:19 -0800 From: Stephen Melvin <melvin@zytek.com> To: imb@scgt.oz.au, mnewell@lupine.nsi.nasa.gov Cc: hackers@FreeBSD.org Subject: Re: User space PPP & leased line Message-ID: <199512210318.TAA14161@syzygy.zytek.com>
next in thread | raw e-mail | index | archive | help
>> We tried using the user mode PPP on our systems for a while, but under
>> very heavy load after 2-3 minutes the line ALWAYS dropped. No indication
>> of what caused it. We use the identical setup with the kernel PPP (pppd)
>> and run for days with no line drops.
>
>This is most likely the result of dropping LCP echo packets on a busy line.
>Try adding ..
>
> disable lqr
> deny lqr
>
Actually, it may be due to *delayed* rather than *dropped* LCP echo
packets and in that case, it would be nice to keep lqr enabled since it
will cause the line to be dropped when there really is nobody on the
other end.
I reported on this problem a few months back. The code in lqr.c
looks for a separation between sent and received packets of 5. I
find that under heavy load, I am experiencing a separation of 6
(*but all packets are eventually echoed*).
I applied the patch shown below to /usr/src/usr.sbin/ppp/lqr.c
and everything works fine. I don't know if this violates some PPP
protocol specification though. Regards,
Stephen Melvin
melvin@zytek.com
--------
*** lqr.c Thu Sep 7 21:49:54 1995
--- lqr.c.orig Thu Sep 7 21:48:49 1995
***************
*** 116,122 ****
lqrsendcnt++;
}
} else if (lqmmethod & LQM_ECHO) {
! if (echoseq - gotseq > 10) {
LogPrintf(LOG_PHASE, "** Too many ECHO packets are lost. **\n");
LcpClose();
Cleanup(EX_ERRDEAD);
--- 116,122 ----
lqrsendcnt++;
}
} else if (lqmmethod & LQM_ECHO) {
! if (echoseq - gotseq > 5) {
LogPrintf(LOG_PHASE, "** Too many ECHO packets are lost. **\n");
LcpClose();
Cleanup(EX_ERRDEAD);
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199512210318.TAA14161>