public inbox for linux-nfs@vger.kernel.org
 help / color / mirror / Atom feed
From: "J. Bruce Fields" <bfields@fieldses.org>
To: Kinglong Mee <kinglongmee@gmail.com>
Cc: "linux-nfs@vger.kernel.org" <linux-nfs@vger.kernel.org>
Subject: Re: [PATCH 01/14] nfsd: Add layouts checking for state resources
Date: Fri, 17 Jul 2015 11:54:23 -0400	[thread overview]
Message-ID: <20150717155423.GB6263@fieldses.org> (raw)
In-Reply-To: <55A7174D.3070007@gmail.com>

On Thu, Jul 16, 2015 at 10:30:37AM +0800, Kinglong Mee wrote:
> On 7/15/2015 23:03, J. Bruce Fields wrote:
> > On Mon, Jul 13, 2015 at 05:29:01PM +0800, Kinglong Mee wrote:
> >> Layout is a state resource, nfsd should check it too.
> >>
> >> Signed-off-by: Kinglong Mee <kinglongmee@gmail.com>
> >> ---
> >>  fs/nfsd/nfs4state.c | 6 ++++--
> >>  1 file changed, 4 insertions(+), 2 deletions(-)
> >>
> >> diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
> >> index 61dfb33..e0a4556 100644
> >> --- a/fs/nfsd/nfs4state.c
> >> +++ b/fs/nfsd/nfs4state.c
> >> @@ -2241,6 +2241,7 @@ static bool client_has_state(struct nfs4_client *clp)
> >>  	 * Also note we should probably be using this in 4.0 case too.
> >>  	 */
> >>  	return !list_empty(&clp->cl_openowners)
> >> +		|| !list_empty(&clp->cl_lo_states)
> >>  		|| !list_empty(&clp->cl_delegations)
> >>  		|| !list_empty(&clp->cl_sessions);
> >>  }
> >> @@ -4257,8 +4258,9 @@ nfsd4_renew(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
> >>  		goto out;
> >>  	clp = cstate->clp;
> >>  	status = nfserr_cb_path_down;
> >> -	if (!list_empty(&clp->cl_delegations)
> >> -			&& clp->cl_cb_state != NFSD4_CB_UP)
> >> +	if ((!list_empty(&clp->cl_delegations)
> >> +	     || !list_empty(&clp->cl_lo_states))
> > 
> > nfsd4_renew is NFSv4.0 only, so I don't think this part is necessary.
> > 
> > That said, it's harmless--why don't we just call client_has_state() here
> > instead?
> 
> I don't think so.
> 
> NFSv4.0's callback path is used for delegation recalling only,
> without delegation, callback path is useless.
> 
> Also, describing in rfc3530,
> 
>    When the client holds delegations, it needs to use RENEW to detect
>    when the server has determined that the callback path is down.  When
>    the server has made such a determination, only the RENEW operation
>    will renew the lease on delegations.  If the server determines the
>    callback path is down, it returns NFS4ERR_CB_PATH_DOWN.  Even though
>    it returns NFS4ERR_CB_PATH_DOWN, the server MUST renew the lease on
>    the record locks and share reservations that the client has
>    established on the server.
> 
> I will drop the second modify in version 2.

Oh, you're right, the cl_openowners check would be wrong for the renew
case.

Applying v2, thanks.--b.

> 
> thanks,
> Kinglong Mee

  parent reply	other threads:[~2015-07-17 15:54 UTC|newest]

Thread overview: 38+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-07-13  9:28 [PATCH 00/14] nfsd: some updates and cleanups Kinglong Mee
2015-07-13  9:29 ` [PATCH 01/14] nfsd: Add layouts checking for state resources Kinglong Mee
2015-07-15 15:03   ` J. Bruce Fields
2015-07-16  2:30     ` Kinglong Mee
2015-07-16  2:33       ` [PATCH v2] nfsd: Add layouts checking in client_has_state() Kinglong Mee
2015-07-17 15:54       ` J. Bruce Fields [this message]
2015-07-13  9:29 ` [PATCH 02/14] nfsd: Add missing gen_confirm in nfsd4_setclientid() Kinglong Mee
2015-07-15 20:47   ` J. Bruce Fields
2015-07-15 20:49     ` J. Bruce Fields
2015-07-16  3:36       ` Kinglong Mee
2015-07-16  3:50         ` Kinglong Mee
2015-07-17 15:58           ` J. Bruce Fields
2015-07-17 17:42             ` Jeff Layton
2015-07-17 23:33               ` [PATCH] nfsd: New counter for generating client confirm verifier Kinglong Mee
2015-07-18 12:16                 ` Jeff Layton
2015-07-20 20:44                   ` J. Bruce Fields
2015-07-23  1:16   ` [PATCH 02/14] nfsd: Add missing gen_confirm in nfsd4_setclientid() Kinglong Mee
2015-07-23 15:53     ` J. Bruce Fields
2015-07-13  9:30 ` [PATCH 03/14] nfsd: Fix memory leak of so_owner.data in nfs4_stateowner Kinglong Mee
2015-07-15 20:57   ` J. Bruce Fields
2015-07-16  4:05     ` [PATCH v2] " Kinglong Mee
2015-07-17 15:59       ` J. Bruce Fields
2015-07-13  9:30 ` [PATCH 04/14] nfsd: Fix a memory leak of struct file_lock Kinglong Mee
2015-07-15 20:59   ` J. Bruce Fields
2015-07-13  9:31 ` [PATCH 05/14] nfsd: Use check_stateid_generation() for generation checking Kinglong Mee
2015-07-22 18:22   ` J. Bruce Fields
2015-07-23  1:09     ` Kinglong Mee
2015-07-13  9:31 ` [PATCH 06/14] nfsd: Drop duplicate locks_init_lock() Kinglong Mee
2015-07-22 18:24   ` J. Bruce Fields
2015-07-13  9:32 ` [PATCH 07/14] nfsd: Remove unneeded values in nfsd4_open() Kinglong Mee
2015-07-13  9:32 ` [PATCH 08/14] nfsd: Drop duplicate checking of seqid in nfsd4_create_session() Kinglong Mee
2015-07-13  9:32 ` [PATCH 09/14] nfsd: Remove nfs4_set_claim_prev() Kinglong Mee
2015-07-13  9:33 ` [PATCH 10/14] nfsd: Remove unused values in nfs4_setlease() Kinglong Mee
2015-07-13  9:33 ` [PATCH 11/14] nfsd: Remove duplicate checking of nfsd_net in nfs4_laundromat() Kinglong Mee
2015-07-13  9:34 ` [PATCH 12/14] nfsd: Remove macro LOFF_OVERFLOW Kinglong Mee
2015-07-13  9:35 ` [PATCH 13/14] nfsd: Use lk_new_xxx instead of v.new.xxx for nfs4_lockowner Kinglong Mee
2015-07-13  9:35 ` [PATCH 14/14] nfsd: Remove unused clientid arguments from, find_lockowner_str{_locked} Kinglong Mee
2015-07-22 19:22   ` J. Bruce Fields

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20150717155423.GB6263@fieldses.org \
    --to=bfields@fieldses.org \
    --cc=kinglongmee@gmail.com \
    --cc=linux-nfs@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox