Date: Mon, 30 Jul 2007 23:30:22 +0200 From: Marko Zec <zec@icir.org> To: Julian Elischer <julian@elischer.org> Cc: Perforce Change Reviews <perforce@freebsd.org> Subject: Re: PERFORCE change 124354 for review Message-ID: <200707302330.23247.zec@icir.org> In-Reply-To: <46AE1EB0.1020909@elischer.org> References: <200707301337.l6UDbq0H044330@repoman.freebsd.org> <46AE1EB0.1020909@elischer.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Monday 30 July 2007 19:24, Julian Elischer wrote: > Marko Zec wrote: > > http://perforce.freebsd.org/chv.cgi?CH=124354 > > > > Change 124354 by zec@zec_tpx32 on 2007/07/30 13:37:17 > > > > > > > > While here, deprecate backpointers from vnet modules to > > the corresponding parent vnets, given that so far they have > > never been used. > > hmmmm > I was thinking about that.. The idea seems to make sense initially.. > I am surprised that they are not used but wonder if that means > they would never have been used? > would they not be used in the case of hierarchical schedulers? Yes we (will) need one form or another of hierarchy among vnets, but this deleted backpointer used to be something else... It was included in each per-module structure (i.e. NET, INET, INET6, IPFW etc.) and was pointing back to the corresponding vnet. It turns out that we always know in which vnet we are operating (by consulting the curvnet macro or by looking at some other appropriate struct such as socket or ifnet), hence maintaining this information in each vnet-module seems to be redundant... Marko
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200707302330.23247.zec>