From owner-freebsd-current  Sun Jan 27 19:47:23 2002
Delivered-To: freebsd-current@freebsd.org
Received: from web11408.mail.yahoo.com (web11408.mail.yahoo.com [216.136.131.238])
	by hub.freebsd.org (Postfix) with SMTP id EAD6A37B402
	for <freebsd-current@freebsd.org>; Sun, 27 Jan 2002 19:47:16 -0800 (PST)
Message-ID: <20020128034716.90219.qmail@web11408.mail.yahoo.com>
Received: from [202.167.61.228] by web11408.mail.yahoo.com via HTTP; Sun, 27 Jan 2002 19:47:16 PST
Date: Sun, 27 Jan 2002 19:47:16 -0800 (PST)
From: Shizuka Kudo <shizukakudo_99@yahoo.com>
Subject: Re: No buffer space available
To: Terry Lambert <tlambert2@mindspring.com>
Cc: freebsd-current@freebsd.org
In-Reply-To: <3C51253C.6F59A4D4@mindspring.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: owner-freebsd-current@FreeBSD.ORG
Precedence: bulk
List-ID: <freebsd-current.FreeBSD.ORG>
List-Archive: <http://docs.freebsd.org/mail/> (Web Archive)
List-Help: <mailto:majordomo@FreeBSD.ORG?subject=help> (List Instructions)
List-Subscribe: <mailto:majordomo@FreeBSD.ORG?subject=subscribe%20freebsd-current>
List-Unsubscribe: <mailto:majordomo@FreeBSD.ORG?subject=unsubscribe%20freebsd-current>
X-Loop: FreeBSD.ORG

Terry,

Thanks for your response. I wonder if I misunderstand
your advice. When looking at the if_rl.c (dated Dec
14), there's already a timer attached to
ifp->if_watchdog. Is this the timer you referred to?
If so, it looks like this timer never called by the
driver in my case as I never saw "watchdog timeout"
error.

Any advice?

--- Terry Lambert <tlambert2@mindspring.com> wrote:
> Shizuka Kudo wrote:
> > Is anyone still seeing the "No buffer space
> availabe"
> > message in 5.0-CURRENT? I have checked the mail
> > archieve and saw several replies, but none worked
> in
> > my case.
> > 
> > I have a Thinkpad 600X with a Melco cardbus 10/100
> > ethernet card (a Realtek 8139B) running 5.0
> NEWCARD
> > kernel with NMBCLUSTERS=16384.
> > "No buffer space available" occurred when I tried
> to
> > ftp a file in my Thinkpad from other client.
> "ifconfig
> > rl0 down" and then "ifconfig rl0 up" resumed the
> > operation for awhile until the error happened
> again.
> > Setting media to 10baseT/UTP did not suffer from
> this
> > error and got about 900Kbytes/s throughput.
> > 
> > Would that be a bug in the driver that ftp server
> is
> > delivering too much traffic to the NIC? Any
> suggestion
> > that I can try?
> 
> That downing and reupping it worked indicates that
> you
> are losing a transmit interrupt draining the write
> queue.
> 
> I'm pretty sure that this has to be related to the
> use
> of "interrupt threads" in -current, as I do not have
> the
> problem with the 8139B's I have, with an older,
> stable
> version of FreeBSD.
> 
> The downing and reupping it resets the card (this is
> an
> intentional side effect, designed to make Tigon II
> cards
> suck^W^W^W^W^Wallow recovery from a hosed driver
> that
> people would rather hack around than
> fix^W^W^W^W^W^W^W^W,
> which should never happen in practice).
> 
> The "normal" workaround for this is to have a
> software
> "watchdog" timer that resets the card when it loses
> its
> mind like this (transmit interrupt pending, no
> transmit
> interrupt seen for timeout period).  You could add
> similar
> code the the RealTek driver pretty easily, using the
> Tigon
> II or another driver as a template, if you wanted to
> work
> around the problem instead of actually fixing
> it^W^W^W^W^W.
> 
> -- Terry


__________________________________________________
Do You Yahoo!?
Great stuff seeking new owners in Yahoo! Auctions! 
http://auctions.yahoo.com

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message