MIME-Version: 1.0
In-Reply-To: <CAHQdGtTqO02BWwTOsXn2s641_KBdawBexG5+kGwmJKWaAmewtA@mail.gmail.com>
References: <CAHQdGtQ_aHRknHzebarn8NUos_wBf4BZOCeg627FsLzcK3erOA@mail.gmail.com>
	<1419015154-91037-1-git-send-email-trond.myklebust@primarydata.com>
	<CAHQdGtTqO02BWwTOsXn2s641_KBdawBexG5+kGwmJKWaAmewtA@mail.gmail.com>
Date: Fri, 19 Dec 2014 15:39:33 -0500
Message-ID: <CAN-5tyFBgkFQUhH9aNsmYmfZn2RBtD8_7qYbMcLHBJr4OZ8m3A@mail.gmail.com>
Subject: Re: [PATCH v2 1/2] NFSv4: Deal with atomic upgrades of an existing delegation
From: Olga Kornievskaia <aglo@umich.edu>
To: Trond Myklebust <trond.myklebust@primarydata.com>
Cc: linux-nfs <linux-nfs@vger.kernel.org>
Content-Type: text/plain; charset=UTF-8
Sender: linux-nfs-owner@vger.kernel.org

On Fri, Dec 19, 2014 at 3:31 PM, Trond Myklebust
<trond.myklebust@primarydata.com> wrote:
> On Fri, Dec 19, 2014 at 1:52 PM, Trond Myklebust
> <trond.myklebust@primarydata.com> wrote:
>> Ensure that we deal correctly with the case where the server sends us a
>> newer instance of the same delegation. If the stateids match, but the
>> sequence numbers differ, then treat the new delegation as if it were
>> an atomic upgrade.
>>
>> Signed-off-by: Trond Myklebust <Trond.Myklebust@primarydata.com>
>> ---
>>  fs/nfs/delegation.c | 21 ++++++++++++++++++---
>>  1 file changed, 18 insertions(+), 3 deletions(-)
>>
>> diff --git a/fs/nfs/delegation.c b/fs/nfs/delegation.c
>> index 7f3f60641344..02b5b2a6d557 100644
>> --- a/fs/nfs/delegation.c
>> +++ b/fs/nfs/delegation.c
>> @@ -301,6 +301,19 @@ nfs_inode_detach_delegation(struct inode *inode)
>>         return nfs_detach_delegation(nfsi, delegation, server);
>>  }
>>
>> +static void
>> +nfs_update_inplace_delegation(struct nfs_inode *nfsi,
>> +               struct nfs_delegation *delegation,
>> +               struct nfs_delegation *update)
>> +{
>> +       if (nfs4_stateid_is_newer(&update->stateid, &delegation->stateid))
>
> ...and the above comparison should be reversed.

do you mean: if(!nfs4_stateid_is_newer())

but if we received a delegation stateid with sequence number lower
than what we have, shouldn't that be some kind of an error?

>
>> +               return;
>> +       delegation->stateid.seqid = update->stateid.seqid;
>> +       smp_wmb();
>> +       delegation->type = update->type;
>> +       nfsi->delegation_state = update->type;
>> +}
>> +
>>  /**
>>   * nfs_inode_set_delegation - set up a delegation on an inode
>>   * @inode: inode to which delegation applies
>> @@ -334,9 +347,11 @@ int nfs_inode_set_delegation(struct inode *inode, struct rpc_cred *cred, struct
>>         old_delegation = rcu_dereference_protected(nfsi->delegation,
>>                                         lockdep_is_held(&clp->cl_lock));
>>         if (old_delegation != NULL) {
>> -               if (nfs4_stateid_match(&delegation->stateid,
>> -                                       &old_delegation->stateid) &&
>> -                               delegation->type == old_delegation->type) {
>> +               /* Is this an update of the existing delegation? */
>> +               if (nfs4_stateid_match_other(&old_delegation->stateid,
>> +                                       &delegation->stateid)) {
>> +                       nfs_update_inplace_delegation(nfsi, old_delegation,
>> +                                       delegation);
>>                         goto out;
>>                 }
>>                 /*
>> --
>> 2.1.0
>>
>
>
>
> --
> Trond Myklebust
>
> Linux NFS client maintainer, PrimaryData
>
> trond.myklebust@primarydata.com