Date: Wed, 10 Jul 2013 10:34:50 -0500 From: Pedro Giffuni <pfg@FreeBSD.org> To: Claude Buisson <clbuisson@orange.fr> Cc: rmacklem@freebsd.org, FreeBSD Current <freebsd-current@freebsd.org> Subject: Re: (follow-up) "Stale NFS file handle" for NFS exported UFS from r252435 Message-ID: <51DD7F1A.4010907@FreeBSD.org> In-Reply-To: <51DD7AB1.4090202@orange.fr> References: <51DD5451.2010801@orange.fr> <51DD6777.90803@orange.fr> <51DD782E.7090503@FreeBSD.org> <51DD7AB1.4090202@orange.fr>
next in thread | previous in thread | raw e-mail | index | archive | help
On 10.07.2013 10:16, Claude Buisson wrote: > On 07/10/2013 17:05, Pedro Giffuni wrote: >> Hello guys; >> >> Thank for finding this, however ... >> > > <snip> > >> >> While I understand this change caused the issue and I am willing to >> revert it, >> I think the problem is actually in NFS. At least ext2/3/4 and fuse (so I >> presume >> glusterfs) have unsigned i_gen. >> > > I have the same thinking (and was rather astonished by the success of > my try at > reverting it): there is something somewhere in the NFS code which have > not been > synced with the UFS change. > > It is the reason I CC'ed rmacklem@ > There is an ongoing port of glusterfs, and glusterfs is known to use both NFS and fuse so I think we would have this problem sooner or later. I hope it's easy to find: I did a search for i_gen with opengrok before making this change and couldn't find any user. Fortunately I had no plans to merge the change for 9.2. Pedro.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?51DD7F1A.4010907>