From owner-freebsd-bugs  Tue May  7 20:15:30 1996
Return-Path: owner-bugs
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id UAA15692
          for bugs-outgoing; Tue, 7 May 1996 20:15:30 -0700 (PDT)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id UAA15684
          for <freebsd-bugs@freebsd.org>; Tue, 7 May 1996 20:15:23 -0700 (PDT)
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.12/8.6.9) id NAA20126; Wed, 8 May 1996 13:08:44 +1000
Date: Wed, 8 May 1996 13:08:44 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199605080308.NAA20126@godzilla.zeta.org.au>
To: cattelan@thebarn.com, freebsd-bugs@freebsd.org
Subject: Re: ppp-2.2 misperceives carrier loss under heavy loads
Sender: owner-bugs@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

>Upon ugrading to 2.2-960323-SNAP and thusly ppp 2.2 pppd is getting
>carrier loss when ever a heavy load is applied to the link.
>(note upgraded to 2.2-960501 last night, problem persists)

>...
>May  5 11:09:09 lupo pppd[8908]: sent [CCP TermReq id=0x2]
>May  5 11:09:09 lupo pppd[8908]: rcvd [CCP TermAck id=0x2]
>May  5 11:09:24 lupo pppd[8908]: Hangup (SIGHUP)
>May  5 11:09:24 lupo /kernel: ppp0: no carrier
>May  5 11:09:24 lupo last message repeated 14 times
>May  5 11:09:24 lupo pppd[8908]: Modem hangup
>May  5 11:09:24 lupo pppd[8908]: Exit.

>I've tracked this down a bit and the no carrier is showing up in
>/sys/net/ppp_tty.c

>And no it's not the modem, I commented out the section that cycles the
>DTR modem stays connected but now I have a dead link.

The SIGHUP messages means that carrier loss was detected.  Perhaps the
cable is noisy.  The 14 repetitions of the "ppp0: no carrier" message
is consistent with the carrier being lost.  The connection doesn't
come back even if the carrier comes back and the driver apparently
delivers some input after carrier is lost.

>One thing I found out after digging into the problem is that my serial
>chips are 8250's uhgggg! I going to get a different board with 16550's
>to see if that has any affect.

A 16550 probably won't help, but a different board might.

>All of this ran quite well for weeks at a time under 2.1 and
>ppp-2.1.2.

Neither the driver nor ppp seems to have changed significantly in ths
area.

Bruce