Date: Tue, 14 Dec 1999 08:26:02 +1000 From: George Michaelson <ggm@dstc.edu.au> To: "Kees Verstoep" <versto@cs.vu.nl>, freebsd-atm@FreeBSD.ORG Subject: Re: Bandwidth Capping fore ATM circuits? Message-ID: <19673.945123962@dstc.edu.au> In-Reply-To: Your message of "Fri, 19 Nov 1999 14:21:25 %2B1000." <6046.942985285@dstc.edu.au>
next in thread | previous in thread | raw e-mail | index | archive | help
/* ... bunch of code justifying a numeric value in this u_long
which is a fusion of the active/free cell count */
xdp->xd_rate = 3473610;
#endif
I have been running under this simple mod for a couple of weeks with
low level traffic and it works fine as in "it doesn't crash FreeBSD"
But I don't see evidence this is actually stopping TCP/IP flooding
the back-end ATM bandwidth.
Can somebody who knows the Fore card confirm if the microcode actually
does honour the free/data cells setting when passed in?
the Linux code maintains state inside the driver and asserts
ENOBUF type errors back on write attempts if the queue of data exceeds
the rate assignment, so its not just using card-resident methods to
try and cap the cell usage.
I have also just noticed the ALTQ codebase seems to say "we don't support
the Fore NIC" so my alternatives for controlling b/w usage are not looking
good right now...
cheers
-George
--
George Michaelson | DSTC Pty Ltd
Email: ggm@dstc.edu.au | University of Qld 4072
Phone: +61 7 3365 4310 | Australia
Fax: +61 7 3365 4311 | http://www.dstc.edu.au
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-atm" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19673.945123962>