Re: [BUG] NFSv4.2: futimens() returns success without sending SETATTR

Holger Hoffstätte <[email protected]> Thu, 23 Jul 2026 21:22:30 +0200
Newsgroups gmane.linux.nfs
Organization Applied Asynchrony, Inc.
Message-ID <[email protected]>
On 2026-07-20 05:18, Michael wrote:
> I am seeing a reproducible NFSv4.2 timestamp issue where an fd-based
> timestamp update returns success but does not appear to send an NFS
> SETATTR operation or update the server-side inode.

Confirmed (also with 7.1.x) and thank you for tracking down the details.
This has been driving me nuts for quite a while, but I could not figure out
whether it's my very historic NFS config, server- or client-side, autofs or
whatever. Thanks for saving me a ton of time wiresharking and stracing.

I got curious to find out when this started, so I went and tried 7.0.0
and voila: reproduces. So it probably happened in the 7.0 merge window,
which matches with my hazy recollection.

Two commits caught my eye:
b8b3002fbfef nfs: split nfs_update_timestamps
761475268fa8 fs: refactor ->update_time handling

The latter is part of a larger vfs series.

Hope that helps someone to find the cause.

cheers
Holger