This is the patch that breaks recovery of opens upon server reboot. I have a test that open a file and gets a write delegation and tried to do a write. At this point, I reboot my server so that write fails with bad_session and then stale_clientid. Upon completing the exchange_id, create_session, and putrootfh, the client no longer sends the open to be recovered and instead resends the failed write. It would use all 0s stateid (this is 4.1) for the write and for the close that follows it. Reverting the patch fixes the problem. On Fri, Dec 19, 2014 at 3:44 PM, Trond Myklebust <trond.myklebust@xxxxxxxxxxxxxxx> wrote: > Remove an incorrect check for NFS_DELEGATION_NEED_RECLAIM in > can_open_delegated(). We are allowed to cache opens even in > a situation where we're doing reboot recovery. > > Signed-off-by: Trond Myklebust <trond.myklebust@xxxxxxxxxxxxxxx> > --- > fs/nfs/nfs4proc.c | 2 -- > 1 file changed, 2 deletions(-) > > diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c > index 8514b59a8c30..002c7dfedb08 100644 > --- a/fs/nfs/nfs4proc.c > +++ b/fs/nfs/nfs4proc.c > @@ -1119,8 +1119,6 @@ static int can_open_delegated(struct nfs_delegation *delegation, fmode_t fmode) > return 0; > if ((delegation->type & fmode) != fmode) > return 0; > - if (test_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags)) > - return 0; > if (test_bit(NFS_DELEGATION_RETURNING, &delegation->flags)) > return 0; > nfs_mark_delegation_referenced(delegation); > -- > 2.1.0 > -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
