linux-nfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Chuck Lever III <chuck.lever@oracle.com>
To: Jeff Layton <jlayton@kernel.org>
Cc: Linux NFS Mailing List <linux-nfs@vger.kernel.org>
Subject: Re: [PATCH 3/4] nfsd: make nfsd4_run_cb a bool return function
Date: Mon, 26 Sep 2022 18:31:13 +0000	[thread overview]
Message-ID: <5D93DF5C-B214-456A-9742-CCC90C8F6802@oracle.com> (raw)
In-Reply-To: <a751b036a0a46a42abe8809c340b05c093e3d148.camel@kernel.org>



> On Sep 26, 2022, at 2:24 PM, Jeff Layton <jlayton@kernel.org> wrote:
> 
> On Mon, 2022-09-26 at 17:51 +0000, Chuck Lever III wrote:
>> 
>>> On Sep 26, 2022, at 12:38 PM, Jeff Layton <jlayton@kernel.org> wrote:
>>> 
>>> ...and mirror the semantics of queue_work. Also, queueing a delegation
>>> recall should always succeed when queueing, so WARN if one ever doesn't.
>> 
>> The description is not especially clear. It seems like the point
>> of this patch is really the part after the "Also, ..." Otherwise,
>> I'm not getting why this change is really needed?
>> 
>> Maybe a tracepoint would be less alarming to users/administrators
>> than would a WARN with a full stack trace? The kdoc comment you
>> added (thank you!) suggests there are times when it is OK for the
>> nfsd4_queue_cb() to fail.
>> 
>> 
> 
> Yeah, that's pretty much the case with the "also". There may be some
> other cases where we ought to catch this sort of thing too. We'd likely
> never see a tracepoint for this. We'd just notice that there was a
> refcount leak.
> 
> queue_work can return false without queueing anything if the work is
> already queued. In this case, we will have taken an extra reference to
> the stid that will never get put. I think this should never happen
> because of the flc_lock, but it'd be good to catch it if it does.

I've pushed Dai's v3 SSC fix and the first two patches in this series
to the nfsd for-next tree. Can you send me 3/4 and 4/4 again but 

 - Include the above text in the description of 3/4, and

 - rebase 4/4 on top of the latest nfsd for-next

Thanks!


> That said, I don't feel that strongly about the patch, so if you think
> it's not worthwhile, I'm fine with holding off on it.
> 
>>> Signed-off-by: Jeff Layton <jlayton@kernel.org>
>>> ---
>>> fs/nfsd/nfs4callback.c | 14 ++++++++++++--
>>> fs/nfsd/nfs4state.c    |  5 ++---
>>> fs/nfsd/state.h        |  2 +-
>>> 3 files changed, 15 insertions(+), 6 deletions(-)
>>> 
>>> diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c
>>> index 4ce328209f61..ba904614ebf5 100644
>>> --- a/fs/nfsd/nfs4callback.c
>>> +++ b/fs/nfsd/nfs4callback.c
>>> @@ -1371,11 +1371,21 @@ void nfsd4_init_cb(struct nfsd4_callback *cb, struct nfs4_client *clp,
>>> 	cb->cb_holds_slot = false;
>>> }
>>> 
>>> -void nfsd4_run_cb(struct nfsd4_callback *cb)
>>> +/**
>>> + * nfsd4_run_cb - queue up a callback job to run
>>> + * @cb: callback to queue
>>> + *
>>> + * Kick off a callback to do its thing. Returns false if it was already
>>> + * queued or running, true otherwise.
>>> + */
>>> +bool nfsd4_run_cb(struct nfsd4_callback *cb)
>>> {
>>> +	bool queued;
>>> 	struct nfs4_client *clp = cb->cb_clp;
>> 
>> Reverse christmas tree, please.
>> 
>> 
>>> 
>>> 	nfsd41_cb_inflight_begin(clp);
>>> -	if (!nfsd4_queue_cb(cb))
>>> +	queued = nfsd4_queue_cb(cb);
>>> +	if (!queued)
>>> 		nfsd41_cb_inflight_end(clp);
>>> +	return queued;
>>> }
>>> diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
>>> index 211f1af1cfb3..90533f43fea9 100644
>>> --- a/fs/nfsd/nfs4state.c
>>> +++ b/fs/nfsd/nfs4state.c
>>> @@ -4861,14 +4861,13 @@ static void nfsd_break_one_deleg(struct nfs4_delegation *dp)
>>> 	 * we know it's safe to take a reference.
>>> 	 */
>>> 	refcount_inc(&dp->dl_stid.sc_count);
>>> -	nfsd4_run_cb(&dp->dl_recall);
>>> +	WARN_ON_ONCE(!nfsd4_run_cb(&dp->dl_recall));
>>> }
>>> 
>>> /* Called from break_lease() with flc_lock held. */
>>> static bool
>>> nfsd_break_deleg_cb(struct file_lock *fl)
>>> {
>>> -	bool ret = false;
>>> 	struct nfs4_delegation *dp = (struct nfs4_delegation *)fl->fl_owner;
>>> 	struct nfs4_file *fp = dp->dl_stid.sc_file;
>>> 	struct nfs4_client *clp = dp->dl_stid.sc_client;
>>> @@ -4894,7 +4893,7 @@ nfsd_break_deleg_cb(struct file_lock *fl)
>>> 	fp->fi_had_conflict = true;
>>> 	nfsd_break_one_deleg(dp);
>>> 	spin_unlock(&fp->fi_lock);
>>> -	return ret;
>>> +	return false;
>>> }
>>> 
>>> /**
>>> diff --git a/fs/nfsd/state.h b/fs/nfsd/state.h
>>> index b3477087a9fc..e2daef3cc003 100644
>>> --- a/fs/nfsd/state.h
>>> +++ b/fs/nfsd/state.h
>>> @@ -692,7 +692,7 @@ extern void nfsd4_probe_callback_sync(struct nfs4_client *clp);
>>> extern void nfsd4_change_callback(struct nfs4_client *clp, struct nfs4_cb_conn *);
>>> extern void nfsd4_init_cb(struct nfsd4_callback *cb, struct nfs4_client *clp,
>>> 		const struct nfsd4_callback_ops *ops, enum nfsd4_cb_op op);
>>> -extern void nfsd4_run_cb(struct nfsd4_callback *cb);
>>> +extern bool nfsd4_run_cb(struct nfsd4_callback *cb);
>>> extern int nfsd4_create_callback_queue(void);
>>> extern void nfsd4_destroy_callback_queue(void);
>>> extern void nfsd4_shutdown_callback(struct nfs4_client *);
>>> -- 
>>> 2.37.3
>>> 
>> 
>> --
>> Chuck Lever
>> 
>> 
>> 
> 
> -- 
> Jeff Layton <jlayton@kernel.org>

--
Chuck Lever




  reply	other threads:[~2022-09-26 18:31 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-09-26 16:38 [PATCH 0/4] nfsd: minor cleanup and sanity check patches for nfsd Jeff Layton
2022-09-26 16:38 ` [PATCH 1/4] nfsd: only fill out return pointer on success in nfsd4_lookup_stateid Jeff Layton
2022-09-26 16:38 ` [PATCH 2/4] nfsd: fix comments about spinlock handling with delegations Jeff Layton
2022-09-26 16:38 ` [PATCH 3/4] nfsd: make nfsd4_run_cb a bool return function Jeff Layton
2022-09-26 17:51   ` Chuck Lever III
2022-09-26 18:24     ` Jeff Layton
2022-09-26 18:31       ` Chuck Lever III [this message]
2022-09-26 16:38 ` [PATCH 4/4] nfsd: extra checks when freeing delegation stateids Jeff Layton

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=5D93DF5C-B214-456A-9742-CCC90C8F6802@oracle.com \
    --to=chuck.lever@oracle.com \
    --cc=jlayton@kernel.org \
    --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;
as well as URLs for NNTP newsgroup(s).