Hi Doug, thanks very much for your patch and efforts. I manually backported the patch to nfs-utils 1.3.4-2.5 source in Debian Buster. I am now testing the modified build on one of my NFSv4 file servers. Looks promising. One additional question: Which nfs-utils branch are your working on - steved/nfs-utils.git ? Best Sebastian __________________ Sebastian Kraus Team IT am Institut für Chemie Gebäude C, Straße des 17. Juni 115, Raum C7 Technische Universität Berlin Fakultät II Institut für Chemie Sekretariat C3 Straße des 17. Juni 135 10623 Berlin Email: sebastian.kraus@xxxxxxxxxxxx ________________________________________ From: linux-nfs-owner@xxxxxxxxxxxxxxx <linux-nfs-owner@xxxxxxxxxxxxxxx> on behalf of Doug Nazar <nazard@xxxxxxxx> Sent: Friday, June 26, 2020 23:30 To: J. Bruce Fields Cc: Kraus, Sebastian; linux-nfs@xxxxxxxxxxxxxxx; Steve Dickson; Olga Kornievskaia Subject: [PATCH v2] Re: Strange segmentation violations of rpc.gssd in Debian Buster On 2020-06-26 17:02, J. Bruce Fields wrote: > Unless I'm missing something--an upcall thread could still be using this > file descriptor. > > If we're particularly unlucky, we could do a new open in a moment and > reuse this file descriptor number, and then then writes in do_downcall() > could end up going to some other random file. > > I think we want these closes done by gssd_free_client() in the !refcnt > case? Makes sense. I was thinking more that it was an abort situation and we shouldn't be sending any data to the kernel but re-use is definitely a concern. I've split it so that we are removed from the event loop in destroy() but the close happens in free(). Doug
