1 files changed, 134 insertions, 55 deletions

diff --git a/fs/nfs/pnfs_dev.c b/fs/nfs/pnfs_dev.c
index 6da209bd9408..bf0f2d67e96c 100644
--- a/fs/nfs/pnfs_dev.c
+++ b/fs/nfs/pnfs_dev.c
@@ -29,8 +29,13 @@
  */
 
 #include <linux/export.h>
+#include <linux/nfs_fs.h>
+#include "nfs4session.h"
+#include "internal.h"
 #include "pnfs.h"
 
+#include "nfs4trace.h"
+
 #define NFSDBG_FACILITY		NFSDBG_PNFS
 
 /*
@@ -40,7 +45,6 @@
 #define NFS4_DEVICE_ID_HASH_SIZE	(1 << NFS4_DEVICE_ID_HASH_BITS)
 #define NFS4_DEVICE_ID_HASH_MASK	(NFS4_DEVICE_ID_HASH_SIZE - 1)
 
-#define PNFS_DEVICE_RETRY_TIMEOUT (120*HZ)
 
 static struct hlist_head nfs4_deviceid_cache[NFS4_DEVICE_ID_HASH_SIZE];
 static DEFINE_SPINLOCK(nfs4_deviceid_lock);
@@ -89,6 +93,73 @@ _lookup_deviceid(const struct pnfs_layoutdriver_type *ld,
 	return NULL;
 }
 
+static struct nfs4_deviceid_node *
+nfs4_get_device_info(struct nfs_server *server,
+		const struct nfs4_deviceid *dev_id,
+		const struct cred *cred, gfp_t gfp_flags)
+{
+	struct nfs4_deviceid_node *d = NULL;
+	struct pnfs_device *pdev = NULL;
+	struct page **pages = NULL;
+	u32 max_resp_sz;
+	int max_pages;
+	int rc, i;
+
+	/*
+	 * Use the session max response size as the basis for setting
+	 * GETDEVICEINFO's maxcount
+	 */
+	max_resp_sz = server->nfs_client->cl_session->fc_attrs.max_resp_sz;
+	max_pages = nfs_page_array_len(0, max_resp_sz);
+	dprintk("%s: server %p max_resp_sz %u max_pages %d\n",
+		__func__, server, max_resp_sz, max_pages);
+
+	pdev = kzalloc(sizeof(*pdev), gfp_flags);
+	if (!pdev)
+		return NULL;
+
+	pages = kcalloc(max_pages, sizeof(struct page *), gfp_flags);
+	if (!pages)
+		goto out_free_pdev;
+
+	for (i = 0; i < max_pages; i++) {
+		pages[i] = alloc_page(gfp_flags);
+		if (!pages[i])
+			goto out_free_pages;
+	}
+
+	memcpy(&pdev->dev_id, dev_id, sizeof(*dev_id));
+	pdev->layout_type = server->pnfs_curr_ld->id;
+	pdev->pages = pages;
+	pdev->pgbase = 0;
+	pdev->pglen = max_resp_sz;
+	pdev->mincount = 0;
+	pdev->maxcount = max_resp_sz - nfs41_maxgetdevinfo_overhead;
+
+	rc = nfs4_proc_getdeviceinfo(server, pdev, cred);
+	dprintk("%s getdevice info returns %d\n", __func__, rc);
+	if (rc)
+		goto out_free_pages;
+
+	/*
+	 * Found new device, need to decode it and then add it to the
+	 * list of known devices for this mountpoint.
+	 */
+	d = server->pnfs_curr_ld->alloc_deviceid_node(server, pdev,
+			gfp_flags);
+	if (d && pdev->nocache)
+		set_bit(NFS_DEVICEID_NOCACHE, &d->flags);
+
+out_free_pages:
+	while (--i >= 0)
+		__free_page(pages[i]);
+	kfree(pages);
+out_free_pdev:
+	kfree(pdev);
+	dprintk("<-- %s d %p\n", __func__, d);
+	return d;
+}
+
 /*
  * Lookup a deviceid in cache and get a reference count on it if found
  *
@@ -96,25 +167,52 @@ _lookup_deviceid(const struct pnfs_layoutdriver_type *ld,
  * @id deviceid to look up
  */
 static struct nfs4_deviceid_node *
-_find_get_deviceid(const struct pnfs_layoutdriver_type *ld,
-		   const struct nfs_client *clp, const struct nfs4_deviceid *id,
-		   long hash)
+__nfs4_find_get_deviceid(struct nfs_server *server,
+		const struct nfs4_deviceid *id, long hash)
 {
 	struct nfs4_deviceid_node *d;
 
 	rcu_read_lock();
-	d = _lookup_deviceid(ld, clp, id, hash);
-	if (d != NULL)
-		atomic_inc(&d->ref);
+	d = _lookup_deviceid(server->pnfs_curr_ld, server->nfs_client, id,
+			hash);
+	if (d != NULL && !atomic_inc_not_zero(&d->ref))
+		d = NULL;
 	rcu_read_unlock();
 	return d;
 }
 
 struct nfs4_deviceid_node *
-nfs4_find_get_deviceid(const struct pnfs_layoutdriver_type *ld,
-		       const struct nfs_client *clp, const struct nfs4_deviceid *id)
+nfs4_find_get_deviceid(struct nfs_server *server,
+		const struct nfs4_deviceid *id, const struct cred *cred,
+		gfp_t gfp_mask)
 {
-	return _find_get_deviceid(ld, clp, id, nfs4_deviceid_hash(id));
+	long hash = nfs4_deviceid_hash(id);
+	struct nfs4_deviceid_node *d, *new;
+
+	d = __nfs4_find_get_deviceid(server, id, hash);
+	if (d)
+		goto found;
+
+	new = nfs4_get_device_info(server, id, cred, gfp_mask);
+	if (!new) {
+		trace_nfs4_find_deviceid(server, id, -ENOENT);
+		return new;
+	}
+
+	spin_lock(&nfs4_deviceid_lock);
+	d = __nfs4_find_get_deviceid(server, id, hash);
+	if (d) {
+		spin_unlock(&nfs4_deviceid_lock);
+		server->pnfs_curr_ld->free_deviceid_node(new);
+	} else {
+		atomic_inc(&new->ref);
+		hlist_add_head_rcu(&new->node, &nfs4_deviceid_cache[hash]);
+		spin_unlock(&nfs4_deviceid_lock);
+		d = new;
+	}
+found:
+	trace_nfs4_find_deviceid(server, id, 0);
+	return d;
 }
 EXPORT_SYMBOL_GPL(nfs4_find_get_deviceid);
 
@@ -141,25 +239,22 @@ nfs4_delete_deviceid(const struct pnfs_layoutdriver_type *ld,
 		return;
 	}
 	hlist_del_init_rcu(&d->node);
+	clear_bit(NFS_DEVICEID_NOCACHE, &d->flags);
 	spin_unlock(&nfs4_deviceid_lock);
-	synchronize_rcu();
 
 	/* balance the initial ref set in pnfs_insert_deviceid */
-	if (atomic_dec_and_test(&d->ref))
-		d->ld->free_deviceid_node(d);
+	nfs4_put_deviceid_node(d);
 }
 EXPORT_SYMBOL_GPL(nfs4_delete_deviceid);
 
 void
-nfs4_init_deviceid_node(struct nfs4_deviceid_node *d,
-			const struct pnfs_layoutdriver_type *ld,
-			const struct nfs_client *nfs_client,
+nfs4_init_deviceid_node(struct nfs4_deviceid_node *d, struct nfs_server *server,
 			const struct nfs4_deviceid *id)
 {
 	INIT_HLIST_NODE(&d->node);
 	INIT_HLIST_NODE(&d->tmpnode);
-	d->ld = ld;
-	d->nfs_client = nfs_client;
+	d->ld = server->pnfs_curr_ld;
+	d->nfs_client = server->nfs_client;
 	d->flags = 0;
 	d->deviceid = *id;
 	atomic_set(&d->ref, 1);
@@ -167,39 +262,6 @@ nfs4_init_deviceid_node(struct nfs4_deviceid_node *d,
 EXPORT_SYMBOL_GPL(nfs4_init_deviceid_node);
 
 /*
- * Uniquely initialize and insert a deviceid node into cache
- *
- * @new new deviceid node
- *      Note that the caller must set up the following members:
- *        new->ld
- *        new->nfs_client
- *        new->deviceid
- *
- * @ret the inserted node, if none found, otherwise, the found entry.
- */
-struct nfs4_deviceid_node *
-nfs4_insert_deviceid_node(struct nfs4_deviceid_node *new)
-{
-	struct nfs4_deviceid_node *d;
-	long hash;
-
-	spin_lock(&nfs4_deviceid_lock);
-	hash = nfs4_deviceid_hash(&new->deviceid);
-	d = _find_get_deviceid(new->ld, new->nfs_client, &new->deviceid, hash);
-	if (d) {
-		spin_unlock(&nfs4_deviceid_lock);
-		return d;
-	}
-
-	hlist_add_head_rcu(&new->node, &nfs4_deviceid_cache[hash]);
-	spin_unlock(&nfs4_deviceid_lock);
-	atomic_inc(&new->ref);
-
-	return new;
-}
-EXPORT_SYMBOL_GPL(nfs4_insert_deviceid_node);
-
-/*
  * Dereference a deviceid node and delete it when its reference count drops
  * to zero.
  *
@@ -212,18 +274,36 @@ EXPORT_SYMBOL_GPL(nfs4_insert_deviceid_node);
 bool
 nfs4_put_deviceid_node(struct nfs4_deviceid_node *d)
 {
+	if (test_bit(NFS_DEVICEID_NOCACHE, &d->flags)) {
+		if (atomic_add_unless(&d->ref, -1, 2))
+			return false;
+		nfs4_delete_deviceid(d->ld, d->nfs_client, &d->deviceid);
+	}
 	if (!atomic_dec_and_test(&d->ref))
 		return false;
+	trace_nfs4_deviceid_free(d->nfs_client, &d->deviceid);
 	d->ld->free_deviceid_node(d);
 	return true;
 }
 EXPORT_SYMBOL_GPL(nfs4_put_deviceid_node);
 
 void
+nfs4_mark_deviceid_available(struct nfs4_deviceid_node *node)
+{
+	if (test_bit(NFS_DEVICEID_UNAVAILABLE, &node->flags)) {
+		clear_bit(NFS_DEVICEID_UNAVAILABLE, &node->flags);
+		smp_mb__after_atomic();
+	}
+}
+EXPORT_SYMBOL_GPL(nfs4_mark_deviceid_available);
+
+void
 nfs4_mark_deviceid_unavailable(struct nfs4_deviceid_node *node)
 {
 	node->timestamp_unavailable = jiffies;
+	smp_mb__before_atomic();
 	set_bit(NFS_DEVICEID_UNAVAILABLE, &node->flags);
+	smp_mb__after_atomic();
 }
 EXPORT_SYMBOL_GPL(nfs4_mark_deviceid_unavailable);
 
@@ -238,6 +318,7 @@ nfs4_test_deviceid_unavailable(struct nfs4_deviceid_node *node)
 		if (time_in_range(node->timestamp_unavailable, start, end))
 			return true;
 		clear_bit(NFS_DEVICEID_UNAVAILABLE, &node->flags);
+		smp_mb__after_atomic();
 	}
 	return false;
 }
@@ -255,6 +336,7 @@ _deviceid_purge_client(const struct nfs_client *clp, long hash)
 		if (d->nfs_client == clp && atomic_read(&d->ref)) {
 			hlist_del_init_rcu(&d->node);
 			hlist_add_head(&d->tmpnode, &tmp);
+			clear_bit(NFS_DEVICEID_NOCACHE, &d->flags);
 		}
 	rcu_read_unlock();
 	spin_unlock(&nfs4_deviceid_lock);
@@ -262,12 +344,10 @@ _deviceid_purge_client(const struct nfs_client *clp, long hash)
 	if (hlist_empty(&tmp))
 		return;
 
-	synchronize_rcu();
 	while (!hlist_empty(&tmp)) {
 		d = hlist_entry(tmp.first, struct nfs4_deviceid_node, tmpnode);
 		hlist_del(&d->tmpnode);
-		if (atomic_dec_and_test(&d->ref))
-			d->ld->free_deviceid_node(d);
+		nfs4_put_deviceid_node(d);
 	}
 }
 
@@ -299,4 +379,3 @@ nfs4_deviceid_mark_client_invalid(struct nfs_client *clp)
 	}
 	rcu_read_unlock();
 }
-