Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 10 Mar 1995 08:57:43 -0800 (PST)
From:      pete@pelican.com (Pete Carah)
To:        bde@zeta.org.au (Bruce Evans)
Cc:        current@FreeBSD.org
Subject:   Re: bad outgoing serial coms
Message-ID:  <m0rn80V-000K3WC@pelican.com>
In-Reply-To: <199503101115.VAA30652@godzilla.zeta.org.au> from "Bruce Evans" at Mar 10, 95 09:15:10 pm

next in thread | previous in thread | raw e-mail | index | archive | help
Bruce Evans writes:

> >Make TTYHOG big - one friend of mine needs 10240 and another 8192.
> >It could need to be as big as Taylor's i-protocol window which is
> >16384 (if your problem is with uucp).
 
> RTS flow control is supposed to work now, so the big buffers should
> be unnecessary (but advisable for speed) if whatever you are connecting
> to supports CTS flow control.
That is not (all of) the problem; my friend here who gets this message
all the time uses rts/cts and at least it works going out.  As far as
I know his modems support rts/cts both ways and are config'd right. 
I'll try to look further at his configuration.  He only sees this
message with uucp, though, and also only sees it when there are lots
of file turnarounds in 'i' protocol (like with the mail for freebsd-*
with short messages).

My modems here are on a 1.1.5 system and it isn't convenient to put either
one on the 2.x system yet; I know that both directions of flow control
work on the 1.1.5....

> >If you are using another protocol, TTYHOG may need to be as big
> >as the TCP window plus the overhead; this probably defaults to 4096 or
> >8192.
> 
> SLIP and PPP don't use clists for input so they don't depend on TTYHOG.

The cited message comes from a place that involves the "raw" clist queue.
There is only one place in the driver that increments the counter
involved...
--------------------------------------------------
                        com->delta_error_counts[CE_TTY_BUF_OVERFLOW]
                                += b_to_q((char *)buf, incc, &tp->t_rawq);
---------------------------------------------------

I don't know why it might come in slip/ppp if they don't use clists;
are you *sure* they don't use the raw clist queue?

-- Pete



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?m0rn80V-000K3WC>