Date: Fri, 31 Aug 2012 12:10:43 +0400 From: Lev Serebryakov <lev@FreeBSD.org> To: Adam Vande More <amvandemore@gmail.com> Cc: freebsd-net@freebsd.org, Ian Smith <smithi@nimnet.asn.au> Subject: Re: Bad routing performance on 500Mhz Geode LX with CURRENT, ipfw and mpd5 (was: ipfw, "ip|all" proto and PPPoE -- does PPPoE packets passed to ipfw?) Message-ID: <643110717.20120831121043@serebryakov.spb.ru> In-Reply-To: <CA%2BtpaK1P6Y-8tDYrP%2B4ca3FG_QwHjPpoZNzUQBCZFjf1LbyEZg@mail.gmail.com> References: <1865271844.20120829131610@serebryakov.spb.ru> <CAHu1Y70MynCMQTrJUMwTZ0%2BLrM1JiZFt_B77028XHfoiRgzmaA@mail.gmail.com> <1807373989.20120829223125@serebryakov.spb.ru> <20120830152726.A33776@sola.nimnet.asn.au> <534292400.20120830131158@serebryakov.spb.ru> <CA%2BtpaK1P6Y-8tDYrP%2B4ca3FG_QwHjPpoZNzUQBCZFjf1LbyEZg@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Hello, Adam. You wrote 31 =C1=D7=C7=D5=D3=D4=C1 2012 =C7., 0:32:25: AVM> BUGS AVM> The vr driver always copies transmit mbuf chains into longword-al= igned AVM> buffers prior to transmission in order to pacify the Rhine chips.= If AVM> buffers are not aligned correctly, the chip will round the suppli= ed AVM> buffer address and begin DMAing from the wrong location. This bu= ffer AVM> copying impairs transmit performance on slower systems but cannot= be AVM> avoided. On faster machines (e.g. a Pentium II), the performance AVM> impact AVM> is much less noticeable. Really, this is not truth. Or, to be more specific, it is not complete truth. Now if_vr copies data only for some chips (quirk VR_Q_NEEDALIGN in sources). My chips (VT6105M) don't need it, and driver doesn't copy data for them. --=20 // Black Lion AKA Lev Serebryakov <lev@FreeBSD.org>
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?643110717.20120831121043>