On Tue, Feb 20, 2018 at 10:49:36PM +0100, Moritz Mühlenhoff wrote: > On Tue, Feb 20, 2018 at 01:56:12PM -0600, Benjamin Kaduk wrote: > > > > It's probably also worth noting that this is not the first time that > > a linux security update caused an openafs regression, > > The only sane way to avoid such occasional breakage is to upstream > the OpenAFS kernel module into the Linux kernel. As long as this doesn't > happen, it'll inevitably happen again.
The IBM copyright and IBM Public License on a lot of the code would make that difficult, as I understand it. But there is an AFS client in the kernel, David Howells' "kafs", which is seeing active development (but is not really in a usable state in the stable kernel, let alone oldstable). -Ben
