On Fri, 2008-06-13 at 17:58 -0400, J. Bruce Fields wrote: > On Fri, Jun 13, 2008 at 05:34:54PM -0400, Trond Myklebust wrote: > > On Fri, 2008-06-13 at 17:31 -0400, J. Bruce Fields wrote: > > > No, that may be a problem too, but I mean: if we're worried about a > > > deadlock due to some allocations that gssd performs on the downcall at > > > the end of the context initiation, then shouldn't we also be worried > > > about all the other allocations that gssd performs itself? > > > > Definitely. > > Do you have any ideas (however speculative) about how we'll deal with > that? Not really: AFAIK there are no mechanisms for fully pre-provisioning the necessary resources that a userland app may need to ensure it can function. mlock() won't, for instance, allow you to pre-provision the socket buffer space that the kerberos stuff needs to communicate... -- Trond Myklebust Linux NFS client maintainer NetApp Trond.Myklebust@xxxxxxxxxx www.netapp.com -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
