Date: Wed, 6 Dec 2006 16:48:58 +0100 From: Marko Zec <zec@icir.org> To: Andre Oppermann <andre@freebsd.org> Cc: Perforce Change Reviews <perforce@freebsd.org>, Marko Zec <zec@freebsd.org> Subject: Re: PERFORCE change 111198 for review Message-ID: <200612061648.59255.zec@icir.org> In-Reply-To: <4576E406.4010108@freebsd.org> References: <200612061531.kB6FVR67021263@repoman.freebsd.org> <4576E406.4010108@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wednesday 06 December 2006 16:38, Andre Oppermann wrote: > Marko Zec wrote: > > http://perforce.freebsd.org/chv.cgi?CH=111198 > > > > Change 111198 by zec@zec_tpx32 on 2006/12/06 15:31:22 > > > > Store a pointer back to the owner vnet in struct inpcbinfo, > > so that whenever an inpcb is directly or indirectly available > > as an argument to a function, the function-local vnet ptr can > > be fetched from there, and not via the curvnetb macro. > > Ugh, this will become locking/refcounting hell. Hmm, why? This change doesn't introduce or need any additional refcounting. Each vnet has its own instance of inpcbinfo, hence inpcbinfo->vnet binding is permanent. A vnet can't be removed until its inpcbinfo->ipi_count tells us there are no inpcbs pointing back to it any more. Marko
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200612061648.59255.zec>