Date: Thu, 07 Dec 2006 14:01:32 +0100 From: Andre Oppermann <andre@freebsd.org> To: Gleb Smirnoff <glebius@FreeBSD.org> Cc: Perforce Change Reviews <perforce@FreeBSD.org>, Paolo Pisati <piso@FreeBSD.org> Subject: Re: PERFORCE change 111230 for review Message-ID: <457810AC.9080406@freebsd.org> In-Reply-To: <20061207124112.GW32700@FreeBSD.org> References: <200612062319.kB6NJgsq031755@repoman.freebsd.org> <20061207110225.GU32700@FreeBSD.org> <4578070A.2030609@freebsd.org> <20061207124112.GW32700@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Gleb Smirnoff wrote: > On Thu, Dec 07, 2006 at 01:20:26PM +0100, Andre Oppermann wrote: > A> Gleb Smirnoff wrote: > A> >On Wed, Dec 06, 2006 at 11:19:42PM +0000, Paolo Pisati wrote: > A> >P> http://perforce.freebsd.org/chv.cgi?CH=111230 > A> >P> > A> >P> Change 111230 by piso@piso_newluxor on 2006/12/06 23:19:06 > A> >P> > A> >P> To workaround a problem when an mbuf is bigger than MCLBYTES > A> >P> (and thus not handled by m_megapullup()), introduce & use > A> >P> m_jumbo16pullup() that uses 16k jumbo size cluster. > A> >P> > A> >P> This problem only showed up on traffic generated on a box with > A> >P> a nic tso enabled. > A> > > A> >Paolo, > A> > > A> >this isn't a fix. Another application will do write(,, 16k + 1) and > A> >m_jumbo16pullup() will fail again. Please backout it, it is a hack. > A> > > A> >We need to fix TSO in such way that real packets, that will be > A> >transmitted to wire, will be passed to pfil handlers. > A> > A> That is not possible. > > ATM this should be at least documented behavior. And a solution should > be thought, because pfil must see real packets, not their precursors. I fully agree about the documentation. Fixing TSO pass 'packets' to pfil is not possible as that equals disabling TSO in the first place. The user has to chose what he needs more, pfil correctly counting packet, or the performance benefits of TSO. Can't have both at the same time. On routers and firewalls TSO is not much of an issues. -- Andre
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?457810AC.9080406>