Date: Sat, 8 Mar 2003 13:08:10 +0000 From: Bruce Cran <bruce@cran.org.uk> To: Terry Lambert <tlambert2@mindspring.com> Cc: hackers@freebsd.org Subject: Re: High CPU usage when forwarding packets Message-ID: <20030308130810.GA5234@fourtytwo.brucec.backnet> In-Reply-To: <3E68EBCF.E19242C1@mindspring.com> References: <20030307131038.GA919@fourtytwo.brucec.backnet> <3E68EBCF.E19242C1@mindspring.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Mar 07, 2003 at 10:58:23AM -0800, Terry Lambert wrote: > Bruce Cran wrote: > > Also, I'm getting > > several thousand 'lnc0: Missed packet -- no receive buffer' messages. > > Could this be the problem, or is the system just not powerful enough do > > nat? The sis0 card is 100MBit PCI, while the lcn0 is 10MBit ISA. > > The "no receive buffers available" message happens when the > system runs out of mbufs. > > There are a lot of reasons this could happen, but the proximal > cause is you didn't tune the number NMBCLUSTERS, et. al. high > enough. You should try rebuilding your kernel with a larger > number. > > If the problem still happens, you need to do a "netstat -a > x", > and then look for large numbers in the "Recv-Q" and "Send-Q" > columns, and then figure out what's causing them. > Thanks, I added kern.ipc.nmbclusters=8192 to /boot/loader.conf and the messages have stopped. I have also learnt that the high CPU usage is simply because I'm trying to push 600KB/sec over an ISA bus, and lots of copying is going on. I'd like to get a PCI card and stop using the onboard lnc, but unfortunately the single PCI slot is already taken up by other other NIC. Bruce Cran To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030308130810.GA5234>