On Fri, Apr 10, 2009 at 05:56:36PM -0500, Serge E. Hallyn wrote: > Quoting Alexey Dobriyan (adobriyan@xxxxxxxxx): > > net_create() will be used by C/R code to create fresh netns on restart. > > > > Signed-off-by: Alexey Dobriyan <adobriyan@xxxxxxxxx> > > Acked-by: Serge Hallyn <serue@xxxxxxxxxx> > > Although this ends up making a new assumption that setup_net(), and > therefore the pernet_ops->init() functions, do not return error > 0. > There's nothing actually stopping that (outside of code review, but > there appears to be no comment anywhere saying that's uncooth). > > If it returns > 0, then we return the ns as though it were valid, but > don't add it to the list of namespaces... Do we care? I think, no. Recalling all the code put under netns changes, ->init hooks were put under 0/-E convention. For sure, my changes in netns xtables, conntracking and xfrm were definitely under it. > > -struct net *copy_net_ns(unsigned long flags, struct net *old_net) > > +struct net *net_create(void) _______________________________________________ Containers mailing list Containers@xxxxxxxxxxxxxxxxxxxxxxxxxx https://lists.linux-foundation.org/mailman/listinfo/containers
