diff mbox series

[v5,3/4] NFSD: add delegation reaper to react to low memory condition

Message ID 1668656688-22507-4-git-send-email-dai.ngo@oracle.com (mailing list archive)
State New, archived
Headers show
Series add support for CB_RECALL_ANY and the delegation reaper | expand

Commit Message

Dai Ngo Nov. 17, 2022, 3:44 a.m. UTC 
The delegation reaper is called by nfsd memory shrinker's on
the 'count' callback. It scans the client list and sends the
courtesy CB_RECALL_ANY to the clients that hold delegations.

To avoid flooding the clients with CB_RECALL_ANY requests, the
delegation reaper sends only one CB_RECALL_ANY request to each
client per 5 seconds.

Signed-off-by: Dai Ngo <dai.ngo@oracle.com>
---
 fs/nfsd/nfs4state.c | 89 ++++++++++++++++++++++++++++++++++++++++++++++++++---
 fs/nfsd/state.h     |  8 +++++
 2 files changed, 93 insertions(+), 4 deletions(-)

Comments

Chuck Lever III Nov. 17, 2022, 2:45 p.m. UTC | #1 
> On Nov 16, 2022, at 10:44 PM, Dai Ngo <dai.ngo@oracle.com> wrote:
> 
> The delegation reaper is called by nfsd memory shrinker's on
> the 'count' callback. It scans the client list and sends the
> courtesy CB_RECALL_ANY to the clients that hold delegations.
> 
> To avoid flooding the clients with CB_RECALL_ANY requests, the
> delegation reaper sends only one CB_RECALL_ANY request to each
> client per 5 seconds.
> 
> Signed-off-by: Dai Ngo <dai.ngo@oracle.com>
> ---
> fs/nfsd/nfs4state.c | 89 ++++++++++++++++++++++++++++++++++++++++++++++++++---
> fs/nfsd/state.h     |  8 +++++
> 2 files changed, 93 insertions(+), 4 deletions(-)
> 
> diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
> index 142481bc96de..13f326ae928c 100644
> --- a/fs/nfsd/nfs4state.c
> +++ b/fs/nfsd/nfs4state.c
> @@ -2131,6 +2131,7 @@ static void __free_client(struct kref *k)
> 	kfree(clp->cl_nii_domain.data);
> 	kfree(clp->cl_nii_name.data);
> 	idr_destroy(&clp->cl_stateids);
> +	kfree(clp->cl_ra);
> 	kmem_cache_free(client_slab, clp);
> }
> 
> @@ -2854,6 +2855,36 @@ static const struct tree_descr client_files[] = {
> 	[3] = {""},
> };
> 
> +static int
> +nfsd4_cb_recall_any_done(struct nfsd4_callback *cb,
> +				struct rpc_task *task)
> +{
> +	switch (task->tk_status) {
> +	case -NFS4ERR_DELAY:
> +		rpc_delay(task, 2 * HZ);
> +		return 0;
> +	default:
> +		return 1;
> +	}
> +}
> +
> +static void
> +nfsd4_cb_recall_any_release(struct nfsd4_callback *cb)
> +{
> +	struct nfs4_client *clp = cb->cb_clp;
> +	struct nfsd_net *nn = net_generic(clp->net, nfsd_net_id);
> +
> +	spin_lock(&nn->client_lock);
> +	clear_bit(NFSD4_CLIENT_CB_RECALL_ANY, &clp->cl_flags);
> +	put_client_renew_locked(clp);
> +	spin_unlock(&nn->client_lock);
> +}
> +
> +static const struct nfsd4_callback_ops nfsd4_cb_recall_any_ops = {
> +	.done		= nfsd4_cb_recall_any_done,
> +	.release	= nfsd4_cb_recall_any_release,
> +};
> +
> static struct nfs4_client *create_client(struct xdr_netobj name,
> 		struct svc_rqst *rqstp, nfs4_verifier *verf)
> {
> @@ -2891,6 +2922,14 @@ static struct nfs4_client *create_client(struct xdr_netobj name,
> 		free_client(clp);
> 		return NULL;
> 	}
> +	clp->cl_ra = kzalloc(sizeof(*clp->cl_ra), GFP_KERNEL);
> +	if (!clp->cl_ra) {
> +		free_client(clp);
> +		return NULL;
> +	}
> +	clp->cl_ra_time = 0;
> +	nfsd4_init_cb(&clp->cl_ra->ra_cb, clp, &nfsd4_cb_recall_any_ops,
> +			NFSPROC4_CLNT_CB_RECALL_ANY);
> 	return clp;
> }
> 
> @@ -4349,14 +4388,16 @@ nfsd4_init_slabs(void)
> static unsigned long
> nfsd_lowmem_shrinker_count(struct shrinker *shrink, struct shrink_control *sc)
> {
> -	int cnt;
> +	int count;
> 	struct nfsd_net *nn = container_of(shrink,
> 			struct nfsd_net, nfsd_client_shrinker);
> 
> -	cnt = atomic_read(&nn->nfsd_courtesy_clients);
> -	if (cnt > 0)
> +	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);
> -	return (unsigned long)cnt;
> +	return (unsigned long)count;
> }
> 
> static unsigned long
> @@ -6134,6 +6175,45 @@ courtesy_client_reaper(struct nfsd_net *nn)
> }
> 
> static void
> +deleg_reaper(struct nfsd_net *nn)
> +{
> +	struct list_head *pos, *next;
> +	struct nfs4_client *clp;
> +	struct list_head cblist;
> +
> +	INIT_LIST_HEAD(&cblist);
> +	spin_lock(&nn->client_lock);
> +	list_for_each_safe(pos, next, &nn->client_lru) {
> +		clp = list_entry(pos, struct nfs4_client, cl_lru);
> +		if (clp->cl_state != NFSD4_ACTIVE ||
> +			list_empty(&clp->cl_delegations) ||
> +			atomic_read(&clp->cl_delegs_in_recall) ||
> +			test_bit(NFSD4_CLIENT_CB_RECALL_ANY, &clp->cl_flags) ||
> +			(ktime_get_boottime_seconds() -
> +				clp->cl_ra_time < 5)) {
> +			continue;
> +		}
> +		list_add(&clp->cl_ra_cblist, &cblist);
> +
> +		/* release in nfsd4_cb_recall_any_release */
> +		atomic_inc(&clp->cl_rpc_users);
> +		set_bit(NFSD4_CLIENT_CB_RECALL_ANY, &clp->cl_flags);
> +		clp->cl_ra_time = ktime_get_boottime_seconds();
> +	}
> +	spin_unlock(&nn->client_lock);
> +
> +	while (!list_empty(&cblist)) {
> +		clp = list_first_entry(&cblist, struct nfs4_client,
> +					cl_ra_cblist);
> +		list_del_init(&clp->cl_ra_cblist);
> +		clp->cl_ra->ra_keep = 0;
> +		clp->cl_ra->ra_bmval[0] = BIT(RCA4_TYPE_MASK_RDATA_DLG) |
> +						BIT(RCA4_TYPE_MASK_WDATA_DLG);

Linux NFSD doesn't hand out write delegations. I don't think we
should set WDATA_DLG yet...?


> +		nfsd4_run_cb(&clp->cl_ra->ra_cb);
> +	}
> +}
> +
> +static void
> nfsd4_lowmem_shrinker(struct work_struct *work)
> {
> 	struct delayed_work *dwork = to_delayed_work(work);
> @@ -6141,6 +6221,7 @@ nfsd4_lowmem_shrinker(struct work_struct *work)
> 				nfsd_shrinker_work);
> 
> 	courtesy_client_reaper(nn);
> +	deleg_reaper(nn);
> }
> 
> static inline __be32 nfs4_check_fh(struct svc_fh *fhp, struct nfs4_stid *stp)
> diff --git a/fs/nfsd/state.h b/fs/nfsd/state.h
> index 6b33cbbe76d3..12ce9792c5b6 100644
> --- a/fs/nfsd/state.h
> +++ b/fs/nfsd/state.h
> @@ -368,6 +368,7 @@ struct nfs4_client {
> #define NFSD4_CLIENT_UPCALL_LOCK	(5)	/* upcall serialization */
> #define NFSD4_CLIENT_CB_FLAG_MASK	(1 << NFSD4_CLIENT_CB_UPDATE | \
> 					 1 << NFSD4_CLIENT_CB_KILL)
> +#define	NFSD4_CLIENT_CB_RECALL_ANY	(6)
> 	unsigned long		cl_flags;
> 	const struct cred	*cl_cb_cred;
> 	struct rpc_clnt		*cl_cb_client;
> @@ -411,6 +412,10 @@ struct nfs4_client {
> 
> 	unsigned int		cl_state;
> 	atomic_t		cl_delegs_in_recall;
> +
> +	struct nfsd4_cb_recall_any	*cl_ra;
> +	time64_t		cl_ra_time;
> +	struct list_head	cl_ra_cblist;
> };
> 
> /* struct nfs4_client_reset
> @@ -642,6 +647,9 @@ enum nfsd4_cb_op {
> 	NFSPROC4_CLNT_CB_RECALL_ANY,
> };
> 
> +#define	RCA4_TYPE_MASK_RDATA_DLG	0
> +#define	RCA4_TYPE_MASK_WDATA_DLG	1
> +
> /* Returns true iff a is later than b: */
> static inline bool nfsd4_stateid_generation_after(stateid_t *a, stateid_t *b)
> {
> -- 
> 2.9.5
> 

--
Chuck Lever
diff mbox series

Patch

diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
index 142481bc96de..13f326ae928c 100644
--- a/fs/nfsd/nfs4state.c
+++ b/fs/nfsd/nfs4state.c
@@ -2131,6 +2131,7 @@  static void __free_client(struct kref *k)
 	kfree(clp->cl_nii_domain.data);
 	kfree(clp->cl_nii_name.data);
 	idr_destroy(&clp->cl_stateids);
+	kfree(clp->cl_ra);
 	kmem_cache_free(client_slab, clp);
 }
 
@@ -2854,6 +2855,36 @@  static const struct tree_descr client_files[] = {
 	[3] = {""},
 };
 
+static int
+nfsd4_cb_recall_any_done(struct nfsd4_callback *cb,
+				struct rpc_task *task)
+{
+	switch (task->tk_status) {
+	case -NFS4ERR_DELAY:
+		rpc_delay(task, 2 * HZ);
+		return 0;
+	default:
+		return 1;
+	}
+}
+
+static void
+nfsd4_cb_recall_any_release(struct nfsd4_callback *cb)
+{
+	struct nfs4_client *clp = cb->cb_clp;
+	struct nfsd_net *nn = net_generic(clp->net, nfsd_net_id);
+
+	spin_lock(&nn->client_lock);
+	clear_bit(NFSD4_CLIENT_CB_RECALL_ANY, &clp->cl_flags);
+	put_client_renew_locked(clp);
+	spin_unlock(&nn->client_lock);
+}
+
+static const struct nfsd4_callback_ops nfsd4_cb_recall_any_ops = {
+	.done		= nfsd4_cb_recall_any_done,
+	.release	= nfsd4_cb_recall_any_release,
+};
+
 static struct nfs4_client *create_client(struct xdr_netobj name,
 		struct svc_rqst *rqstp, nfs4_verifier *verf)
 {
@@ -2891,6 +2922,14 @@  static struct nfs4_client *create_client(struct xdr_netobj name,
 		free_client(clp);
 		return NULL;
 	}
+	clp->cl_ra = kzalloc(sizeof(*clp->cl_ra), GFP_KERNEL);
+	if (!clp->cl_ra) {
+		free_client(clp);
+		return NULL;
+	}
+	clp->cl_ra_time = 0;
+	nfsd4_init_cb(&clp->cl_ra->ra_cb, clp, &nfsd4_cb_recall_any_ops,
+			NFSPROC4_CLNT_CB_RECALL_ANY);
 	return clp;
 }
 
@@ -4349,14 +4388,16 @@  nfsd4_init_slabs(void)
 static unsigned long
 nfsd_lowmem_shrinker_count(struct shrinker *shrink, struct shrink_control *sc)
 {
-	int cnt;
+	int count;
 	struct nfsd_net *nn = container_of(shrink,
 			struct nfsd_net, nfsd_client_shrinker);
 
-	cnt = atomic_read(&nn->nfsd_courtesy_clients);
-	if (cnt > 0)
+	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);
-	return (unsigned long)cnt;
+	return (unsigned long)count;
 }
 
 static unsigned long
@@ -6134,6 +6175,45 @@  courtesy_client_reaper(struct nfsd_net *nn)
 }
 
 static void
+deleg_reaper(struct nfsd_net *nn)
+{
+	struct list_head *pos, *next;
+	struct nfs4_client *clp;
+	struct list_head cblist;
+
+	INIT_LIST_HEAD(&cblist);
+	spin_lock(&nn->client_lock);
+	list_for_each_safe(pos, next, &nn->client_lru) {
+		clp = list_entry(pos, struct nfs4_client, cl_lru);
+		if (clp->cl_state != NFSD4_ACTIVE ||
+			list_empty(&clp->cl_delegations) ||
+			atomic_read(&clp->cl_delegs_in_recall) ||
+			test_bit(NFSD4_CLIENT_CB_RECALL_ANY, &clp->cl_flags) ||
+			(ktime_get_boottime_seconds() -
+				clp->cl_ra_time < 5)) {
+			continue;
+		}
+		list_add(&clp->cl_ra_cblist, &cblist);
+
+		/* release in nfsd4_cb_recall_any_release */
+		atomic_inc(&clp->cl_rpc_users);
+		set_bit(NFSD4_CLIENT_CB_RECALL_ANY, &clp->cl_flags);
+		clp->cl_ra_time = ktime_get_boottime_seconds();
+	}
+	spin_unlock(&nn->client_lock);
+
+	while (!list_empty(&cblist)) {
+		clp = list_first_entry(&cblist, struct nfs4_client,
+					cl_ra_cblist);
+		list_del_init(&clp->cl_ra_cblist);
+		clp->cl_ra->ra_keep = 0;
+		clp->cl_ra->ra_bmval[0] = BIT(RCA4_TYPE_MASK_RDATA_DLG) |
+						BIT(RCA4_TYPE_MASK_WDATA_DLG);
+		nfsd4_run_cb(&clp->cl_ra->ra_cb);
+	}
+}
+
+static void
 nfsd4_lowmem_shrinker(struct work_struct *work)
 {
 	struct delayed_work *dwork = to_delayed_work(work);
@@ -6141,6 +6221,7 @@  nfsd4_lowmem_shrinker(struct work_struct *work)
 				nfsd_shrinker_work);
 
 	courtesy_client_reaper(nn);
+	deleg_reaper(nn);
 }
 
 static inline __be32 nfs4_check_fh(struct svc_fh *fhp, struct nfs4_stid *stp)
diff --git a/fs/nfsd/state.h b/fs/nfsd/state.h
index 6b33cbbe76d3..12ce9792c5b6 100644
--- a/fs/nfsd/state.h
+++ b/fs/nfsd/state.h
@@ -368,6 +368,7 @@  struct nfs4_client {
 #define NFSD4_CLIENT_UPCALL_LOCK	(5)	/* upcall serialization */
 #define NFSD4_CLIENT_CB_FLAG_MASK	(1 << NFSD4_CLIENT_CB_UPDATE | \
 					 1 << NFSD4_CLIENT_CB_KILL)
+#define	NFSD4_CLIENT_CB_RECALL_ANY	(6)
 	unsigned long		cl_flags;
 	const struct cred	*cl_cb_cred;
 	struct rpc_clnt		*cl_cb_client;
@@ -411,6 +412,10 @@  struct nfs4_client {
 
 	unsigned int		cl_state;
 	atomic_t		cl_delegs_in_recall;
+
+	struct nfsd4_cb_recall_any	*cl_ra;
+	time64_t		cl_ra_time;
+	struct list_head	cl_ra_cblist;
 };
 
 /* struct nfs4_client_reset
@@ -642,6 +647,9 @@  enum nfsd4_cb_op {
 	NFSPROC4_CLNT_CB_RECALL_ANY,
 };
 
+#define	RCA4_TYPE_MASK_RDATA_DLG	0
+#define	RCA4_TYPE_MASK_WDATA_DLG	1
+
 /* Returns true iff a is later than b: */
 static inline bool nfsd4_stateid_generation_after(stateid_t *a, stateid_t *b)
 {