2014-10-18 21:59:07

by Toralf Förster

[permalink] [raw]
Subject: fuzz testing gave today : "kernel: NFS: nfs4_reclaim_open_state: unhandled error -13"

happens within a 32 bit user mode linux guest running latest git tree v3.17-9981-g929254d
It occurred few times about half an hour after the test case was started.

Well, it is very early in the current release cycle, so just asking, if this is already known / worth to investigate ?
--
Toralf
pgp key: 0076 E94E



2014-10-19 20:34:15

by Toralf Förster

[permalink] [raw]
Subject: Re: fuzz testing gave today : "kernel: NFS: nfs4_reclaim_open_state: unhandled error -13"

On 10/19/2014 10:03 PM, Trond Myklebust wrote:
> Errno 13 is EACCES. Might the test case involve changing the mode
> bits/acls on an open file, then putting the client in situation where
> an expired lease causes it to attempt open state recovery?

right - that's likely possible due to 2 trinity childs running in parallel


--
Toralf
pgp key: 0076 E94E


2014-10-19 20:03:01

by Trond Myklebust

[permalink] [raw]
Subject: Re: fuzz testing gave today : "kernel: NFS: nfs4_reclaim_open_state: unhandled error -13"

On Sun, Oct 19, 2014 at 12:59 AM, Toralf Förster <[email protected]> wrote:
> happens within a 32 bit user mode linux guest running latest git tree v3.17-9981-g929254d
> It occurred few times about half an hour after the test case was started.
>
> Well, it is very early in the current release cycle, so just asking, if this is already known / worth to investigate ?

Errno 13 is EACCES. Might the test case involve changing the mode
bits/acls on an open file, then putting the client in situation where
an expired lease causes it to attempt open state recovery?

Cheers
Trond