Return-Path: Received: from mail-gx0-f174.google.com ([209.85.161.174]:56310 "EHLO mail-gx0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755220Ab1HaU2M (ORCPT ); Wed, 31 Aug 2011 16:28:12 -0400 Received: by gxk21 with SMTP id 21so887436gxk.19 for ; Wed, 31 Aug 2011 13:28:12 -0700 (PDT) Message-ID: <4E5E994E.6040107@tonian.com> Date: Wed, 31 Aug 2011 13:27:58 -0700 From: Benny Halevy To: "J. Bruce Fields" CC: linux-nfs@vger.kernel.org, Casey Bodley Subject: Re: [PATCH] nfsd4: in-stateid seqid should start with 1 References: <20110831195329.GD19223@fieldses.org> In-Reply-To: <20110831195329.GD19223@fieldses.org> Content-Type: text/plain; charset=ISO-8859-1 Sender: linux-nfs-owner@vger.kernel.org List-ID: MIME-Version: 1.0 Don't we also need the following? As per RFC5661, skip zero seqid when wrapping around while updating a stateid. Signed-off-by: Benny Halevy git diff --stat -p -M fs/nfsd/state.h | 5 ++++- 1 files changed, 4 insertions(+), 1 deletions(-) diff --git a/fs/nfsd/state.h b/fs/nfsd/state.h index 77f98b8..7d18d3e 100644 --- a/fs/nfsd/state.h +++ b/fs/nfsd/state.h @@ -289,7 +289,10 @@ struct nfs4_client_reclaim { static inline void update_stateid(stateid_t *stateid) { - stateid->si_generation++; + u32 new = stateid->si_generation + 1; + if (new == 0) + new = 1; + stateid->si_generation = new; } /* A reasonable value for REPLAY_ISIZE was estimated as follows: On 2011-08-31 12:53, J. Bruce Fields wrote: > From: J. Bruce Fields > > Thanks to Casey for reminding me that 5661 gives a special meaning to a > value of 0 in the stateid's seqid field, so we should instead be > starting out stateid's with si_generation 1. > > Reported-by: Casey Bodley > Signed-off-by: J. Bruce Fields > --- > fs/nfsd/nfs4state.c | 6 +++--- > 1 files changed, 3 insertions(+), 3 deletions(-) > > diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c > index 49c3dd1..7caa812 100644 > --- a/fs/nfsd/nfs4state.c > +++ b/fs/nfsd/nfs4state.c > @@ -251,7 +251,7 @@ alloc_init_deleg(struct nfs4_client *clp, struct nfs4_stateid *stp, struct svc_f > dp->dl_stateid.si_boot = boot_time; > dp->dl_stateid.si_stateownerid = current_delegid++; > dp->dl_stateid.si_fileid = 0; > - dp->dl_stateid.si_generation = 0; > + dp->dl_stateid.si_generation = 1; > fh_copy_shallow(&dp->dl_fh, ¤t_fh->fh_handle); > dp->dl_time = 0; > atomic_set(&dp->dl_count, 1); > @@ -2303,7 +2303,7 @@ init_stateid(struct nfs4_stateid *stp, struct nfs4_file *fp, struct nfsd4_open * > stp->st_stateid.si_boot = boot_time; > stp->st_stateid.si_stateownerid = oo->oo_owner.so_id; > stp->st_stateid.si_fileid = fp->fi_id; > - stp->st_stateid.si_generation = 0; > + stp->st_stateid.si_generation = 1; > stp->st_access_bmap = 0; > stp->st_deny_bmap = 0; > __set_bit(open->op_share_access & ~NFS4_SHARE_WANT_MASK, > @@ -3898,7 +3898,7 @@ alloc_init_lock_stateid(struct nfs4_lockowner *lo, struct nfs4_file *fp, struct > stp->st_stateid.si_boot = boot_time; > stp->st_stateid.si_stateownerid = lo->lo_owner.so_id; > stp->st_stateid.si_fileid = fp->fi_id; > - stp->st_stateid.si_generation = 0; > + stp->st_stateid.si_generation = 1; > stp->st_access_bmap = 0; > stp->st_deny_bmap = open_stp->st_deny_bmap; > stp->st_openstp = open_stp;