[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-id: <173818520688.22054.4135013466624893151@noble.neil.brown.name>
Date: Thu, 30 Jan 2025 08:13:26 +1100
From: "NeilBrown" <neilb@...e.de>
To: cel@...nel.org
Cc: "Olga Kornievskaia" <okorniev@...hat.com>, "Dai Ngo" <Dai.Ngo@...cle.com>,
"Tom Talpey" <tom@...pey.com>, "Jeff Layton" <jlayton@...nel.org>,
"Chuck Lever" <chuck.lever@...cle.com>,
"Salvatore Bonaccorso" <carnil@...ian.org>, linux-nfs@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject:
Re: [PATCH] nfsd: validate the nfsd_serv pointer before calling svc_wake_up
On Wed, 29 Jan 2025, cel@...nel.org wrote:
> From: Chuck Lever <chuck.lever@...cle.com>
>
> On Sat, 25 Jan 2025 20:13:18 -0500, Jeff Layton wrote:
> > nfsd_file_dispose_list_delayed can be called from the filecache
> > laundrette, which is shut down after the nfsd threads are shut down and
> > the nfsd_serv pointer is cleared. If nn->nfsd_serv is NULL then there
> > are no threads to wake.
> >
> > Ensure that the nn->nfsd_serv pointer is non-NULL before calling
> > svc_wake_up in nfsd_file_dispose_list_delayed. This is safe since the
> > svc_serv is not freed until after the filecache laundrette is cancelled.
> >
> > [...]
>
> Applied to nfsd-testing, thanks!
>
> Test experience should demonstrate whether more strict memory
> ordering semantics are needed.
No it won't. The need for READ_ONCE is theoretical, not - in this case
- practical.
“Program testing can be used to show the presence of bugs, but never to show their absence!”
― Edsger W. Dijkstra
NeilBrown
>
> [1/1] nfsd: validate the nfsd_serv pointer before calling svc_wake_up
> commit: 363683ced1718d66ad54e1bdf52d41d544f783b2
>
> --
> Chuck Lever
>
>
Powered by blists - more mailing lists