1 files changed, 55 insertions, 34 deletions

diff --git a/net/smc/smc_pnet.c b/net/smc/smc_pnet.c
index 67e9d9fde085..25fb2fd186e2 100644
--- a/net/smc/smc_pnet.c
+++ b/net/smc/smc_pnet.c
@@ -64,6 +64,7 @@ struct smc_pnetentry {
 		struct {
 			char eth_name[IFNAMSIZ + 1];
 			struct net_device *ndev;
+			netdevice_tracker dev_tracker;
 		};
 		struct {
 			char ib_name[IB_DEVICE_NAME_MAX + 1];
@@ -112,14 +113,15 @@ static int smc_pnet_remove_by_pnetid(struct net *net, char *pnet_name)
 	pnettable = &sn->pnettable;
 
 	/* remove table entry */
-	write_lock(&pnettable->lock);
+	mutex_lock(&pnettable->lock);
 	list_for_each_entry_safe(pnetelem, tmp_pe, &pnettable->pnetlist,
 				 list) {
 		if (!pnet_name ||
 		    smc_pnet_match(pnetelem->pnet_name, pnet_name)) {
 			list_del(&pnetelem->list);
 			if (pnetelem->type == SMC_PNET_ETH && pnetelem->ndev) {
-				dev_put(pnetelem->ndev);
+				netdev_put(pnetelem->ndev,
+					   &pnetelem->dev_tracker);
 				pr_warn_ratelimited("smc: net device %s "
 						    "erased user defined "
 						    "pnetid %.16s\n",
@@ -130,7 +132,7 @@ static int smc_pnet_remove_by_pnetid(struct net *net, char *pnet_name)
 			rc = 0;
 		}
 	}
-	write_unlock(&pnettable->lock);
+	mutex_unlock(&pnettable->lock);
 
 	/* if this is not the initial namespace, stop here */
 	if (net != &init_net)
@@ -191,11 +193,11 @@ static int smc_pnet_add_by_ndev(struct net_device *ndev)
 	sn = net_generic(net, smc_net_id);
 	pnettable = &sn->pnettable;
 
-	write_lock(&pnettable->lock);
+	mutex_lock(&pnettable->lock);
 	list_for_each_entry_safe(pnetelem, tmp_pe, &pnettable->pnetlist, list) {
 		if (pnetelem->type == SMC_PNET_ETH && !pnetelem->ndev &&
 		    !strncmp(pnetelem->eth_name, ndev->name, IFNAMSIZ)) {
-			dev_hold(ndev);
+			netdev_hold(ndev, &pnetelem->dev_tracker, GFP_ATOMIC);
 			pnetelem->ndev = ndev;
 			rc = 0;
 			pr_warn_ratelimited("smc: adding net device %s with "
@@ -205,7 +207,7 @@ static int smc_pnet_add_by_ndev(struct net_device *ndev)
 			break;
 		}
 	}
-	write_unlock(&pnettable->lock);
+	mutex_unlock(&pnettable->lock);
 	return rc;
 }
 
@@ -223,10 +225,10 @@ static int smc_pnet_remove_by_ndev(struct net_device *ndev)
 	sn = net_generic(net, smc_net_id);
 	pnettable = &sn->pnettable;
 
-	write_lock(&pnettable->lock);
+	mutex_lock(&pnettable->lock);
 	list_for_each_entry_safe(pnetelem, tmp_pe, &pnettable->pnetlist, list) {
 		if (pnetelem->type == SMC_PNET_ETH && pnetelem->ndev == ndev) {
-			dev_put(pnetelem->ndev);
+			netdev_put(pnetelem->ndev, &pnetelem->dev_tracker);
 			pnetelem->ndev = NULL;
 			rc = 0;
 			pr_warn_ratelimited("smc: removing net device %s with "
@@ -236,7 +238,7 @@ static int smc_pnet_remove_by_ndev(struct net_device *ndev)
 			break;
 		}
 	}
-	write_unlock(&pnettable->lock);
+	mutex_unlock(&pnettable->lock);
 	return rc;
 }
 
@@ -310,8 +312,9 @@ static struct smc_ib_device *smc_pnet_find_ib(char *ib_name)
 	list_for_each_entry(ibdev, &smc_ib_devices.list, list) {
 		if (!strncmp(ibdev->ibdev->name, ib_name,
 			     sizeof(ibdev->ibdev->name)) ||
-		    !strncmp(dev_name(ibdev->ibdev->dev.parent), ib_name,
-			     IB_DEVICE_NAME_MAX - 1)) {
+		    (ibdev->ibdev->dev.parent &&
+		     !strncmp(dev_name(ibdev->ibdev->dev.parent), ib_name,
+			     IB_DEVICE_NAME_MAX - 1))) {
 			goto out;
 		}
 	}
@@ -367,11 +370,9 @@ static int smc_pnet_add_eth(struct smc_pnettable *pnettable, struct net *net,
 	new_pe->type = SMC_PNET_ETH;
 	memcpy(new_pe->pnet_name, pnet_name, SMC_MAX_PNETID_LEN);
 	strncpy(new_pe->eth_name, eth_name, IFNAMSIZ);
-	new_pe->ndev = ndev;
-
 	rc = -EEXIST;
 	new_netdev = true;
-	write_lock(&pnettable->lock);
+	mutex_lock(&pnettable->lock);
 	list_for_each_entry(tmp_pe, &pnettable->pnetlist, list) {
 		if (tmp_pe->type == SMC_PNET_ETH &&
 		    !strncmp(tmp_pe->eth_name, eth_name, IFNAMSIZ)) {
@@ -380,10 +381,15 @@ static int smc_pnet_add_eth(struct smc_pnettable *pnettable, struct net *net,
 		}
 	}
 	if (new_netdev) {
+		if (ndev) {
+			new_pe->ndev = ndev;
+			netdev_tracker_alloc(ndev, &new_pe->dev_tracker,
+					     GFP_ATOMIC);
+		}
 		list_add_tail(&new_pe->list, &pnettable->pnetlist);
-		write_unlock(&pnettable->lock);
+		mutex_unlock(&pnettable->lock);
 	} else {
-		write_unlock(&pnettable->lock);
+		mutex_unlock(&pnettable->lock);
 		kfree(new_pe);
 		goto out_put;
 	}
@@ -444,7 +450,7 @@ static int smc_pnet_add_ib(struct smc_pnettable *pnettable, char *ib_name,
 	new_pe->ib_port = ib_port;
 
 	new_ibdev = true;
-	write_lock(&pnettable->lock);
+	mutex_lock(&pnettable->lock);
 	list_for_each_entry(tmp_pe, &pnettable->pnetlist, list) {
 		if (tmp_pe->type == SMC_PNET_IB &&
 		    !strncmp(tmp_pe->ib_name, ib_name, IB_DEVICE_NAME_MAX)) {
@@ -454,9 +460,9 @@ static int smc_pnet_add_ib(struct smc_pnettable *pnettable, char *ib_name,
 	}
 	if (new_ibdev) {
 		list_add_tail(&new_pe->list, &pnettable->pnetlist);
-		write_unlock(&pnettable->lock);
+		mutex_unlock(&pnettable->lock);
 	} else {
-		write_unlock(&pnettable->lock);
+		mutex_unlock(&pnettable->lock);
 		kfree(new_pe);
 	}
 	return (new_ibdev) ? 0 : -EEXIST;
@@ -601,7 +607,7 @@ static int _smc_pnet_dump(struct net *net, struct sk_buff *skb, u32 portid,
 	pnettable = &sn->pnettable;
 
 	/* dump pnettable entries */
-	read_lock(&pnettable->lock);
+	mutex_lock(&pnettable->lock);
 	list_for_each_entry(pnetelem, &pnettable->pnetlist, list) {
 		if (pnetid && !smc_pnet_match(pnetelem->pnet_name, pnetid))
 			continue;
@@ -616,7 +622,7 @@ static int _smc_pnet_dump(struct net *net, struct sk_buff *skb, u32 portid,
 			break;
 		}
 	}
-	read_unlock(&pnettable->lock);
+	mutex_unlock(&pnettable->lock);
 	return idx;
 }
 
@@ -709,7 +715,8 @@ static struct genl_family smc_pnet_nl_family __ro_after_init = {
 	.netnsok = true,
 	.module = THIS_MODULE,
 	.ops = smc_pnet_ops,
-	.n_ops =  ARRAY_SIZE(smc_pnet_ops)
+	.n_ops =  ARRAY_SIZE(smc_pnet_ops),
+	.resv_start_op = SMC_PNETID_FLUSH + 1,
 };
 
 bool smc_pnet_is_ndev_pnetid(struct net *net, u8 *pnetid)
@@ -860,12 +867,15 @@ int smc_pnet_net_init(struct net *net)
 	struct smc_pnetids_ndev *pnetids_ndev = &sn->pnetids_ndev;
 
 	INIT_LIST_HEAD(&pnettable->pnetlist);
-	rwlock_init(&pnettable->lock);
+	mutex_init(&pnettable->lock);
 	INIT_LIST_HEAD(&pnetids_ndev->list);
 	rwlock_init(&pnetids_ndev->lock);
 
 	smc_pnet_create_pnetids_list(net);
 
+	/* disable handshake limitation by default */
+	net->smc.limit_smc_hs = 0;
+
 	return 0;
 }
 
@@ -940,7 +950,7 @@ static int smc_pnet_find_ndev_pnetid_by_table(struct net_device *ndev,
 	sn = net_generic(net, smc_net_id);
 	pnettable = &sn->pnettable;
 
-	read_lock(&pnettable->lock);
+	mutex_lock(&pnettable->lock);
 	list_for_each_entry(pnetelem, &pnettable->pnetlist, list) {
 		if (pnetelem->type == SMC_PNET_ETH && ndev == pnetelem->ndev) {
 			/* get pnetid of netdev device */
@@ -949,7 +959,7 @@ static int smc_pnet_find_ndev_pnetid_by_table(struct net_device *ndev,
 			break;
 		}
 	}
-	read_unlock(&pnettable->lock);
+	mutex_unlock(&pnettable->lock);
 	return rc;
 }
 
@@ -976,14 +986,16 @@ static int smc_pnet_determine_gid(struct smc_ib_device *ibdev, int i,
 /* find a roce device for the given pnetid */
 static void _smc_pnet_find_roce_by_pnetid(u8 *pnet_id,
 					  struct smc_init_info *ini,
-					  struct smc_ib_device *known_dev)
+					  struct smc_ib_device *known_dev,
+					  struct net *net)
 {
 	struct smc_ib_device *ibdev;
 	int i;
 
 	mutex_lock(&smc_ib_devices.mutex);
 	list_for_each_entry(ibdev, &smc_ib_devices.list, list) {
-		if (ibdev == known_dev)
+		if (ibdev == known_dev ||
+		    !rdma_dev_access_netns(ibdev->ibdev, net))
 			continue;
 		for (i = 1; i <= SMC_MAX_PORTS; i++) {
 			if (!rdma_is_port_valid(ibdev->ibdev, i))
@@ -1000,12 +1012,14 @@ out:
 	mutex_unlock(&smc_ib_devices.mutex);
 }
 
-/* find alternate roce device with same pnet_id and vlan_id */
+/* find alternate roce device with same pnet_id, vlan_id and net namespace */
 void smc_pnet_find_alt_roce(struct smc_link_group *lgr,
 			    struct smc_init_info *ini,
 			    struct smc_ib_device *known_dev)
 {
-	_smc_pnet_find_roce_by_pnetid(lgr->pnet_id, ini, known_dev);
+	struct net *net = lgr->net;
+
+	_smc_pnet_find_roce_by_pnetid(lgr->pnet_id, ini, known_dev, net);
 }
 
 /* if handshake network device belongs to a roce device, return its
@@ -1014,6 +1028,7 @@ void smc_pnet_find_alt_roce(struct smc_link_group *lgr,
 static void smc_pnet_find_rdma_dev(struct net_device *netdev,
 				   struct smc_init_info *ini)
 {
+	struct net *net = dev_net(netdev);
 	struct smc_ib_device *ibdev;
 
 	mutex_lock(&smc_ib_devices.mutex);
@@ -1021,6 +1036,10 @@ static void smc_pnet_find_rdma_dev(struct net_device *netdev,
 		struct net_device *ndev;
 		int i;
 
+		/* check rdma net namespace */
+		if (!rdma_dev_access_netns(ibdev->ibdev, net))
+			continue;
+
 		for (i = 1; i <= SMC_MAX_PORTS; i++) {
 			if (!rdma_is_port_valid(ibdev->ibdev, i))
 				continue;
@@ -1051,15 +1070,17 @@ static void smc_pnet_find_roce_by_pnetid(struct net_device *ndev,
 					 struct smc_init_info *ini)
 {
 	u8 ndev_pnetid[SMC_MAX_PNETID_LEN];
+	struct net *net;
 
 	ndev = pnet_find_base_ndev(ndev);
+	net = dev_net(ndev);
 	if (smc_pnetid_by_dev_port(ndev->dev.parent, ndev->dev_port,
 				   ndev_pnetid) &&
 	    smc_pnet_find_ndev_pnetid_by_table(ndev, ndev_pnetid)) {
 		smc_pnet_find_rdma_dev(ndev, ini);
 		return; /* pnetid could not be determined */
 	}
-	_smc_pnet_find_roce_by_pnetid(ndev_pnetid, ini, NULL);
+	_smc_pnet_find_roce_by_pnetid(ndev_pnetid, ini, NULL, net);
 }
 
 static void smc_pnet_find_ism_by_pnetid(struct net_device *ndev,
@@ -1141,7 +1162,7 @@ int smc_pnetid_by_table_ib(struct smc_ib_device *smcibdev, u8 ib_port)
 	sn = net_generic(&init_net, smc_net_id);
 	pnettable = &sn->pnettable;
 
-	read_lock(&pnettable->lock);
+	mutex_lock(&pnettable->lock);
 	list_for_each_entry(tmp_pe, &pnettable->pnetlist, list) {
 		if (tmp_pe->type == SMC_PNET_IB &&
 		    !strncmp(tmp_pe->ib_name, ib_name, IB_DEVICE_NAME_MAX) &&
@@ -1151,7 +1172,7 @@ int smc_pnetid_by_table_ib(struct smc_ib_device *smcibdev, u8 ib_port)
 			break;
 		}
 	}
-	read_unlock(&pnettable->lock);
+	mutex_unlock(&pnettable->lock);
 
 	return rc;
 }
@@ -1170,7 +1191,7 @@ int smc_pnetid_by_table_smcd(struct smcd_dev *smcddev)
 	sn = net_generic(&init_net, smc_net_id);
 	pnettable = &sn->pnettable;
 
-	read_lock(&pnettable->lock);
+	mutex_lock(&pnettable->lock);
 	list_for_each_entry(tmp_pe, &pnettable->pnetlist, list) {
 		if (tmp_pe->type == SMC_PNET_IB &&
 		    !strncmp(tmp_pe->ib_name, ib_name, IB_DEVICE_NAME_MAX)) {
@@ -1179,7 +1200,7 @@ int smc_pnetid_by_table_smcd(struct smcd_dev *smcddev)
 			break;
 		}
 	}
-	read_unlock(&pnettable->lock);
+	mutex_unlock(&pnettable->lock);
 
 	return rc;
 }