Date: Fri, 20 Oct 2006 00:12:58 +0200 From: "Conrad Burger" <conrad.burger@gmail.com> To: "Bill Moran" <wmoran@collaborativefusion.com> Cc: Jason Thomson <jason.thomson@mintel.com>, freebsd-stable@freebsd.org Subject: Re: When will the new BCE driver in HEAD be incorporated into RELENG_6? Message-ID: <d0c4d1610610191512n2d43b6deu966ed5d316b4b599@mail.gmail.com> In-Reply-To: <20061019130040.5f06baf7.wmoran@collaborativefusion.com> References: <d0c4d1610610180258l1fb94e96m349022fcb2d0330c@mail.gmail.com> <45364782.8090209@samsco.org> <45366122.6090704@mintel.com> <20061019130040.5f06baf7.wmoran@collaborativefusion.com>
next in thread | previous in thread | raw e-mail | index | archive | help
------=_Part_128523_11302630.1161295978155 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline On 19/10/06, Bill Moran <wmoran@collaborativefusion.com> wrote: > In response to Jason Thomson <jason.thomson@mintel.com>: > > > Scott Long wrote: > > > Conrad Burger wrote: > > > > > >> Hi > > >> > > >> It looks like there is a "new" version of the bce driver in HEAD. > > >> When will it be incorporated into Releng_6? > > > > > > > > > It will be merged when someone, preferably 2-3 people, tell me that > > > the changes in HEAD work for them. So far, no one has. > > > > > > Scott > > > _______________________________________________ > > > > Using the driver from HEAD* in the latest RELENG_6 didn't fix our > > problems. > > > > We could still trigger the Watchdog timeout when copying a local file to > > an NFS mounted filesystem (UDP mount, GigE speeds). > > Same here, although it seemed to require a lot more effort to produce the > problem. > > I see there have been additional updates to the driver in the past 10 > hours. I'll grab those and try again to see if they help any. > > -- > Bill Moran > Collaborative Fusion Inc. > Would it be possible for someone to provide instructions on how to compile the new bce driver in HEAD on RELENG_6? Just to give you an idea of the problem we are sitting with. Attached is an image of a comparison between a FreeBSD 5.4 with em and a FreeBSD 6.2-pre with bce. The blue line is the 5.4 system. You will notice how smooth the graph is for the 5.4 system. The image can also be downloaded at "http://www.mxit.co.za/mxa_227430610303000_sess.png_bytesinout.png" Both systems run the same java app and both receive the same load. On the 6.2-pre network I/Os sometime takes up to 7 seconds to complete. We are starting to think that the performance difference might have something to do with synchronization within the bce dirver/SMP network stack... we don't know enough so any help would be much appreciated! Does anyone know how to disable interrupt coalescing on the bce driver? Regards Conrad ------=_Part_128523_11302630.1161295978155--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?d0c4d1610610191512n2d43b6deu966ed5d316b4b599>