A few months ago I did some patches to add an upcall to replace the legacy clientid tracking scheme in nfsd. At the time, several people including Boaz and Steve expressed that they would prefer an upcall that used call_usermodehelper instead of requiring a running daemon. While I'm loath to add yet another upcall to the kernel, I must admit that they have a good point. Yet another running daemon for something as infrequently called as this is less than ideal. The idiot-proofness of a usermodehelper upcall is hard to beat for the "just works" factor. This patchset adds a new set of client tracking ops to the kernel that use call_usermodehelper to exec a program in userland to do its bidding. I'll also be posting a set of nfs-utils patches soon to add the callout program for this. This seems to work as expected and is fairly simple. There are a couple of lingering issues. - Do we need to switch to a different mount namespace somehow based on the net namespace? Eventually we want to allow nfsd to run in a container. At some point we'll need a mechanism to ensure that the upcall runs within the correct container. Stanislav, I'd appreciate your input here... - What about nfsdcld? I don't believe any distros have deployed it yet, and there aren't a lot of compelling reasons to stick with it if this patchset seems reasonable. I suggest we plan to deprecate it in a couple of releases unless there are objections. Unless anyone sees major problems with this approach, I think these patches are reasonable for 3.7. Jeff Layton (2): nfsd: add a usermodehelper upcall for NFSv4 client ID tracking nfsd: change algorithm for selecting the client_tracking_ops fs/nfsd/nfs4recover.c | 174 +++++++++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 164 insertions(+), 10 deletions(-) -- 1.7.11.4 -- 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
