Linux NFS development
 help / color / mirror / Atom feed
From: Jeff Layton <jlayton@kernel.org>
To: Dai Ngo <dai.ngo@oracle.com>, chuck.lever@oracle.com
Cc: efault@gmx.de, linux-nfs@vger.kernel.org
Subject: Re: [PATCH v3 1/1] NFSD: fix WARN_ON_ONCE in __queue_delayed_work
Date: Wed, 11 Jan 2023 05:47:52 -0500	[thread overview]
Message-ID: <956fe608b456141b1b43f9dfe65581f168247cb8.camel@kernel.org> (raw)
In-Reply-To: <1673432658-4140-1-git-send-email-dai.ngo@oracle.com>

On Wed, 2023-01-11 at 02:24 -0800, Dai Ngo wrote:
> Currently nfsd4_state_shrinker_worker can be schduled multiple times
> from nfsd4_state_shrinker_count when memory is low. This causes
> the WARN_ON_ONCE in __queue_delayed_work to trigger.
> 
> This patch allows only one instance of nfsd4_state_shrinker_worker
> at a time using the nfsd_shrinker_active flag, protected by the
> client_lock.
> 
> Change nfsd_shrinker_work from delayed_work to work_struct since we
> don't use the delay.
> 
> Replace mod_delayed_work in nfsd4_state_shrinker_count with queue_work.
> 
> Cancel work_struct nfsd_shrinker_work after unregistering shrinker
> in nfs4_state_shutdown_net
> 
> Fixes: 44df6f439a17 ("NFSD: add delegation reaper to react to low memory condition")
> Reported-by: Mike Galbraith <efault@gmx.de>
> Signed-off-by: Dai Ngo <dai.ngo@oracle.com>
> ---
> v2:
>   . Change nfsd_shrinker_work from delayed_work to work_struct
>   . Replace mod_delayed_work in nfsd4_state_shrinker_count with queue_work
>   . Cancel work_struct nfsd_shrinker_work after unregistering shrinker
> v3:
>   . set nfsd_shrinker_active earlier in nfsd4_state_shrinker_count
> 
>  fs/nfsd/netns.h     |  3 ++-
>  fs/nfsd/nfs4state.c | 24 +++++++++++++++++++-----
>  2 files changed, 21 insertions(+), 6 deletions(-)
> 
> diff --git a/fs/nfsd/netns.h b/fs/nfsd/netns.h
> index 8c854ba3285b..b0c7b657324b 100644
> --- a/fs/nfsd/netns.h
> +++ b/fs/nfsd/netns.h
> @@ -195,7 +195,8 @@ struct nfsd_net {
>  
>  	atomic_t		nfsd_courtesy_clients;
>  	struct shrinker		nfsd_client_shrinker;
> -	struct delayed_work	nfsd_shrinker_work;
> +	struct work_struct	nfsd_shrinker_work;
> +	bool			nfsd_shrinker_active;
>  };
>  
>  /* Simple check to find out if a given net was properly initialized */
> diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
> index a7cfefd7c205..35ec4cba88b3 100644
> --- a/fs/nfsd/nfs4state.c
> +++ b/fs/nfsd/nfs4state.c
> @@ -4407,11 +4407,22 @@ nfsd4_state_shrinker_count(struct shrinker *shrink, struct shrink_control *sc)
>  	struct nfsd_net *nn = container_of(shrink,
>  			struct nfsd_net, nfsd_client_shrinker);
>  
> +	spin_lock(&nn->client_lock);
> +	if (nn->nfsd_shrinker_active) {
> +		spin_unlock(&nn->client_lock);
> +		return 0;
> +	}
> +	nn->nfsd_shrinker_active = true;
>  	count = atomic_read(&nn->nfsd_courtesy_clients);
>  	if (!count)
>  		count = atomic_long_read(&num_delegations);
> -	if (count)
> -		mod_delayed_work(laundry_wq, &nn->nfsd_shrinker_work, 0);
> +	if (count) {
> +		spin_unlock(&nn->client_lock);
> +		queue_work(laundry_wq, &nn->nfsd_shrinker_work);
> +	} else {
> +		nn->nfsd_shrinker_active = false;
> +		spin_unlock(&nn->client_lock);
> +	}

The change to normal work_struct is an improvement, but NAK on this
patch. The spinlocking and flag are not needed here. I seriously doubt
that we have a clear understanding of this problem.

>  	return (unsigned long)count;
>  }
>  
> @@ -6233,12 +6244,14 @@ deleg_reaper(struct nfsd_net *nn)
>  static void
>  nfsd4_state_shrinker_worker(struct work_struct *work)
>  {
> -	struct delayed_work *dwork = to_delayed_work(work);
> -	struct nfsd_net *nn = container_of(dwork, struct nfsd_net,
> +	struct nfsd_net *nn = container_of(work, struct nfsd_net,
>  				nfsd_shrinker_work);
>  
>  	courtesy_client_reaper(nn);
>  	deleg_reaper(nn);
> +	spin_lock(&nn->client_lock);
> +	nn->nfsd_shrinker_active = false;
> +	spin_unlock(&nn->client_lock);
>  }
>  
>  static inline __be32 nfs4_check_fh(struct svc_fh *fhp, struct nfs4_stid *stp)
> @@ -8064,7 +8077,7 @@ static int nfs4_state_create_net(struct net *net)
>  	INIT_LIST_HEAD(&nn->blocked_locks_lru);
>  
>  	INIT_DELAYED_WORK(&nn->laundromat_work, laundromat_main);
> -	INIT_DELAYED_WORK(&nn->nfsd_shrinker_work, nfsd4_state_shrinker_worker);
> +	INIT_WORK(&nn->nfsd_shrinker_work, nfsd4_state_shrinker_worker);
>  	get_net(net);
>  
>  	nn->nfsd_client_shrinker.scan_objects = nfsd4_state_shrinker_scan;
> @@ -8171,6 +8184,7 @@ nfs4_state_shutdown_net(struct net *net)
>  	struct nfsd_net *nn = net_generic(net, nfsd_net_id);
>  
>  	unregister_shrinker(&nn->nfsd_client_shrinker);
> +	cancel_work(&nn->nfsd_shrinker_work);
>  	cancel_delayed_work_sync(&nn->laundromat_work);
>  	locks_end_grace(&nn->nfsd4_manager);
>  

-- 
Jeff Layton <jlayton@kernel.org>

  reply	other threads:[~2023-01-11 10:48 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-01-11 10:24 [PATCH v3 1/1] NFSD: fix WARN_ON_ONCE in __queue_delayed_work Dai Ngo
2023-01-11 10:47 ` Jeff Layton [this message]
2023-01-11 11:36   ` dai.ngo
2023-01-11 11:01 ` Mike Galbraith

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=956fe608b456141b1b43f9dfe65581f168247cb8.camel@kernel.org \
    --to=jlayton@kernel.org \
    --cc=chuck.lever@oracle.com \
    --cc=dai.ngo@oracle.com \
    --cc=efault@gmx.de \
    --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