NFS Server Scalability Improvement & Other NFS Enhancements For Linux 6.13

There are a few Network File System (NFS) enhancements worth pointing out with the in-development Linux 6.13 kernel.

Sent out on Tuesday were the NFSD changes for Linux 6.13 which include a scalability improvement to the NFS server's NFSv4 back-channel session implementation. That pull request explains:

"Jeff Layton contributed a scalability improvement to NFSD's NFSv4 backchannel session implementation. This improvement is intended to increase the rate at which NFSD can safely recall NFSv4 delegations from clients, to avoid the need to revoke them. Revoking requires a slow state recovery process."

Among those changes by Jeff Layton are for allowing up to 32 callback session slots compared to NFSD currently using just a single slot and was becoming a bottleneck in some cases during a lot of callback activity.

The NFSD updates also include bug fixes and other incremental improvements.

On Thursday the NFS client changes were sent out for Linux 6.13. This round there are a number of NFS client fixes, local I/O clean-ups, making processes that are waiting for an I/O lock killable, and other changes.

Also of related interest is the VFS exportfs pull request. This work allows for NFS connectable file handles for user-space servers. From that pull request:

"This contains work to bring NFS connectable file handles to userspace servers.

The name_to_handle_at() system call is extended to encode connectable file handles. Such file handles can be resolved to an open file with a connected path. So far userspace NFS servers couldn't make use of this functionality even though the kernel does already support it. This is achieved by introducing a new flag for name_to_handle_at().

Similarly, the open_by_handle_at() system call is tought to understand connectable file handles explicitly created via name_to_handle_at()."

The Linux 6.13 merge window is set to wrap up on Sunday with the Linux 6.13-rc1 release.