Merge tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma

Pull rdma updates from Jason Gunthorpe:
 "Nothing exciting this cycle, most of the diffstat is changing SPDX
  'or' to 'OR'.

  Summary:

   - Bugfixes for hns, mlx5, and hfi1

   - Hardening patches for size_*, counted_by, strscpy

   - rts fixes from static analysis

   - Dump SRQ objects in rdma netlink, with hns support

   - Fix a performance regression in mlx5 MR deregistration

   - New XDR (200Gb/lane) link speed

   - SRQ record doorbell latency optimization for hns

   - IPSEC support for mlx5 multi-port mode

   - ibv_rereg_mr() support for irdma

   - Affiliated event support for bnxt_re

   - Opt out for the spec compliant qkey security enforcement as we
     discovered SW that breaks under enforcement

   - Comment and trivial updates"

* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma: (50 commits)
  IB/mlx5: Fix init stage error handling to avoid double free of same QP and UAF
  RDMA/mlx5: Fix mkey cache WQ flush
  RDMA/hfi1: Workaround truncation compilation error
  IB/hfi1: Fix potential deadlock on &irq_src_lock and &dd->uctxt_lock
  RDMA/core: Remove NULL check before dev_{put, hold}
  RDMA/hfi1: Remove redundant assignment to pointer ppd
  RDMA/mlx5: Change the key being sent for MPV device affiliation
  RDMA/bnxt_re: Fix clang -Wimplicit-fallthrough in bnxt_re_handle_cq_async_error()
  RDMA/hns: Fix init failure of RoCE VF and HIP08
  RDMA/hns: Fix unnecessary port_num transition in HW stats allocation
  RDMA/hns: The UD mode can only be configured with DCQCN
  RDMA/hns: Add check for SL
  RDMA/hns: Fix signed-unsigned mixed comparisons
  RDMA/hns: Fix uninitialized ucmd in hns_roce_create_qp_common()
  RDMA/hns: Fix printing level of asynchronous events
  RDMA/core: Add support to set privileged QKEY parameter
  RDMA/bnxt_re: Do not report SRQ error in srq notification
  RDMA/bnxt_re: Report async events and errors
  RDMA/bnxt_re: Update HW interface headers
  IB/mlx5: Fix rdma counter binding for RAW QP
  ...

Author: torvalds

drivers/infiniband/core/cache.c

@@ -46,7 +46,7 @@
 
 struct ib_pkey_cache {
 	int             table_len;
-	u16             table[];
+	u16             table[] __counted_by(table_len);
 };
 
 struct ib_update_work {

drivers/infiniband/core/core_priv.h

@@ -373,4 +373,5 @@ void rdma_umap_priv_init(struct rdma_umap_priv *priv,
 
 void ib_cq_pool_cleanup(struct ib_device *dev);
 
+bool rdma_nl_get_privileged_qkey(void);
 #endif /* _CORE_PRIV_H */

drivers/infiniband/core/device.c

@@ -804,7 +804,7 @@ static int alloc_port_data(struct ib_device *device)
 	 * empty slots at the beginning.
 	 */
 	pdata_rcu = kzalloc(struct_size(pdata_rcu, pdata,
-					rdma_end_port(device) + 1),
+					size_add(rdma_end_port(device), 1)),
 			    GFP_KERNEL);
 	if (!pdata_rcu)
 		return -ENOMEM;
@@ -2651,6 +2651,8 @@ void ib_set_device_ops(struct ib_device *dev, const struct ib_device_ops *ops)
 	SET_DEVICE_OP(dev_ops, fill_res_mr_entry_raw);
 	SET_DEVICE_OP(dev_ops, fill_res_qp_entry);
 	SET_DEVICE_OP(dev_ops, fill_res_qp_entry_raw);
+	SET_DEVICE_OP(dev_ops, fill_res_srq_entry);
+	SET_DEVICE_OP(dev_ops, fill_res_srq_entry_raw);
 	SET_DEVICE_OP(dev_ops, fill_stat_mr_entry);
 	SET_DEVICE_OP(dev_ops, get_dev_fw_str);
 	SET_DEVICE_OP(dev_ops, get_dma_mr);

drivers/infiniband/core/lag.c

@@ -102,8 +102,7 @@ static struct net_device *rdma_get_xmit_slave_udp(struct ib_device *device,
 
 void rdma_lag_put_ah_roce_slave(struct net_device *xmit_slave)
 {
-	if (xmit_slave)
-		dev_put(xmit_slave);
+	dev_put(xmit_slave);
 }
 
 struct net_device *rdma_lag_get_ah_roce_slave(struct ib_device *device,

drivers/infiniband/core/nldev.c

@@ -43,6 +43,13 @@
 #include "restrack.h"
 #include "uverbs.h"
 
+/*
+ * This determines whether a non-privileged user is allowed to specify a
+ * controlled QKEY or not, when true non-privileged user is allowed to specify
+ * a controlled QKEY.
+ */
+static bool privileged_qkey;
+
 typedef int (*res_fill_func_t)(struct sk_buff*, bool,
 			       struct rdma_restrack_entry*, uint32_t);
 
@@ -156,6 +163,7 @@ static const struct nla_policy nldev_policy[RDMA_NLDEV_ATTR_MAX] = {
 	[RDMA_NLDEV_SYS_ATTR_COPY_ON_FORK]	= { .type = NLA_U8 },
 	[RDMA_NLDEV_ATTR_STAT_HWCOUNTER_INDEX]	= { .type = NLA_U32 },
 	[RDMA_NLDEV_ATTR_STAT_HWCOUNTER_DYNAMIC] = { .type = NLA_U8 },
+	[RDMA_NLDEV_SYS_ATTR_PRIVILEGED_QKEY_MODE] = { .type = NLA_U8 },
 };
 
 static int put_driver_name_print_type(struct sk_buff *msg, const char *name,
@@ -237,6 +245,12 @@ int rdma_nl_put_driver_u64_hex(struct sk_buff *msg, const char *name, u64 value)
 }
 EXPORT_SYMBOL(rdma_nl_put_driver_u64_hex);
 
+bool rdma_nl_get_privileged_qkey(void)
+{
+	return privileged_qkey || capable(CAP_NET_RAW);
+}
+EXPORT_SYMBOL(rdma_nl_get_privileged_qkey);
+
 static int fill_nldev_handle(struct sk_buff *msg, struct ib_device *device)
 {
 	if (nla_put_u32(msg, RDMA_NLDEV_ATTR_DEV_INDEX, device->index))
@@ -357,8 +371,7 @@ static int fill_port_info(struct sk_buff *msg,
 	}
 
 out:
-	if (netdev)
-		dev_put(netdev);
+	dev_put(netdev);
 	return ret;
 }
 
@@ -818,6 +831,7 @@ static int fill_res_srq_entry(struct sk_buff *msg, bool has_cap_net_admin,
 			      struct rdma_restrack_entry *res, uint32_t port)
 {
 	struct ib_srq *srq = container_of(res, struct ib_srq, res);
+	struct ib_device *dev = srq->device;
 
 	if (nla_put_u32(msg, RDMA_NLDEV_ATTR_RES_SRQN, srq->res.id))
 		goto err;
@@ -837,12 +851,29 @@ static int fill_res_srq_entry(struct sk_buff *msg, bool has_cap_net_admin,
 	if (fill_res_srq_qps(msg, srq))
 		goto err;
 
-	return fill_res_name_pid(msg, res);
+	if (fill_res_name_pid(msg, res))
+		goto err;
+
+	if (dev->ops.fill_res_srq_entry)
+		return dev->ops.fill_res_srq_entry(msg, srq);
+
+	return 0;
 
 err:
 	return -EMSGSIZE;
 }
 
+static int fill_res_srq_raw_entry(struct sk_buff *msg, bool has_cap_net_admin,
+				 struct rdma_restrack_entry *res, uint32_t port)
+{
+	struct ib_srq *srq = container_of(res, struct ib_srq, res);
+	struct ib_device *dev = srq->device;
+
+	if (!dev->ops.fill_res_srq_entry_raw)
+		return -EINVAL;
+	return dev->ops.fill_res_srq_entry_raw(msg, srq);
+}
+
 static int fill_stat_counter_mode(struct sk_buff *msg,
 				  struct rdma_counter *counter)
 {
@@ -1652,6 +1683,7 @@ RES_GET_FUNCS(mr_raw, RDMA_RESTRACK_MR);
 RES_GET_FUNCS(counter, RDMA_RESTRACK_COUNTER);
 RES_GET_FUNCS(ctx, RDMA_RESTRACK_CTX);
 RES_GET_FUNCS(srq, RDMA_RESTRACK_SRQ);
+RES_GET_FUNCS(srq_raw, RDMA_RESTRACK_SRQ);
 
 static LIST_HEAD(link_ops);
 static DECLARE_RWSEM(link_ops_rwsem);
@@ -1882,6 +1914,12 @@ static int nldev_sys_get_doit(struct sk_buff *skb, struct nlmsghdr *nlh,
 		return err;
 	}
 
+	err = nla_put_u8(msg, RDMA_NLDEV_SYS_ATTR_PRIVILEGED_QKEY_MODE,
+			 (u8)privileged_qkey);
+	if (err) {
+		nlmsg_free(msg);
+		return err;
+	}
 	/*
 	 * Copy-on-fork is supported.
 	 * See commits:
@@ -1898,18 +1936,11 @@ static int nldev_sys_get_doit(struct sk_buff *skb, struct nlmsghdr *nlh,
 	return rdma_nl_unicast(sock_net(skb->sk), msg, NETLINK_CB(skb).portid);
 }
 
-static int nldev_set_sys_set_doit(struct sk_buff *skb, struct nlmsghdr *nlh,
-				  struct netlink_ext_ack *extack)
+static int nldev_set_sys_set_netns_doit(struct nlattr *tb[])
 {
-	struct nlattr *tb[RDMA_NLDEV_ATTR_MAX];
 	u8 enable;
 	int err;
 
-	err = nlmsg_parse(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1,
-			  nldev_policy, extack);
-	if (err || !tb[RDMA_NLDEV_SYS_ATTR_NETNS_MODE])
-		return -EINVAL;
-
 	enable = nla_get_u8(tb[RDMA_NLDEV_SYS_ATTR_NETNS_MODE]);
 	/* Only 0 and 1 are supported */
 	if (enable > 1)
@@ -1919,6 +1950,40 @@ static int nldev_set_sys_set_doit(struct sk_buff *skb, struct nlmsghdr *nlh,
 	return err;
 }
 
+static int nldev_set_sys_set_pqkey_doit(struct nlattr *tb[])
+{
+	u8 enable;
+
+	enable = nla_get_u8(tb[RDMA_NLDEV_SYS_ATTR_PRIVILEGED_QKEY_MODE]);
+	/* Only 0 and 1 are supported */
+	if (enable > 1)
+		return -EINVAL;
+
+	privileged_qkey = enable;
+	return 0;
+}
+
+static int nldev_set_sys_set_doit(struct sk_buff *skb, struct nlmsghdr *nlh,
+				  struct netlink_ext_ack *extack)
+{
+	struct nlattr *tb[RDMA_NLDEV_ATTR_MAX];
+	int err;
+
+	err = nlmsg_parse(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1,
+			  nldev_policy, extack);
+	if (err)
+		return -EINVAL;
+
+	if (tb[RDMA_NLDEV_SYS_ATTR_NETNS_MODE])
+		return nldev_set_sys_set_netns_doit(tb);
+
+	if (tb[RDMA_NLDEV_SYS_ATTR_PRIVILEGED_QKEY_MODE])
+		return nldev_set_sys_set_pqkey_doit(tb);
+
+	return -EINVAL;
+}
+
+
 static int nldev_stat_set_mode_doit(struct sk_buff *msg,
 				    struct netlink_ext_ack *extack,
 				    struct nlattr *tb[],
@@ -2558,6 +2623,11 @@ static const struct rdma_nl_cbs nldev_cb_table[RDMA_NLDEV_NUM_OPS] = {
 		.dump = nldev_res_get_mr_raw_dumpit,
 		.flags = RDMA_NL_ADMIN_PERM,
 	},
+	[RDMA_NLDEV_CMD_RES_SRQ_GET_RAW] = {
+		.doit = nldev_res_get_srq_raw_doit,
+		.dump = nldev_res_get_srq_raw_dumpit,
+		.flags = RDMA_NL_ADMIN_PERM,
+	},
 	[RDMA_NLDEV_CMD_STAT_GET_STATUS] = {
 		.doit = nldev_stat_get_counter_status_doit,
 	},

drivers/infiniband/core/rw.c

@@ -666,7 +666,7 @@ void rdma_rw_init_qp(struct ib_device *dev, struct ib_qp_init_attr *attr)
 	factor = 1;
 
 	/*
-	 * If the devices needs MRs to perform RDMA READ or WRITE operations,
+	 * If the device needs MRs to perform RDMA READ or WRITE operations,
 	 * we'll need two additional MRs for the registrations and the
 	 * invalidation.
 	 */

drivers/infiniband/core/sa_query.c

@@ -2159,7 +2159,9 @@ static int ib_sa_add_one(struct ib_device *device)
 	s = rdma_start_port(device);
 	e = rdma_end_port(device);
 
-	sa_dev = kzalloc(struct_size(sa_dev, port, e - s + 1), GFP_KERNEL);
+	sa_dev = kzalloc(struct_size(sa_dev, port,
+				     size_add(size_sub(e, s), 1)),
+			 GFP_KERNEL);
 	if (!sa_dev)
 		return -ENOMEM;
 

drivers/infiniband/core/sysfs.c

@@ -342,6 +342,10 @@ static ssize_t rate_show(struct ib_device *ibdev, u32 port_num,
 		speed = " NDR";
 		rate = 1000;
 		break;
+	case IB_SPEED_XDR:
+		speed = " XDR";
+		rate = 2000;
+		break;
 	case IB_SPEED_SDR:
 	default:		/* default to SDR for invalid rates */
 		speed = " SDR";
@@ -903,7 +907,7 @@ alloc_hw_stats_device(struct ib_device *ibdev)
 	 * Two extra attribue elements here, one for the lifespan entry and
 	 * one to NULL terminate the list for the sysfs core code
 	 */
-	data = kzalloc(struct_size(data, attrs, stats->num_counters + 1),
+	data = kzalloc(struct_size(data, attrs, size_add(stats->num_counters, 1)),
 		       GFP_KERNEL);
 	if (!data)
 		goto err_free_stats;
@@ -1009,7 +1013,7 @@ alloc_hw_stats_port(struct ib_port *port, struct attribute_group *group)
 	 * Two extra attribue elements here, one for the lifespan entry and
 	 * one to NULL terminate the list for the sysfs core code
 	 */
-	data = kzalloc(struct_size(data, attrs, stats->num_counters + 1),
+	data = kzalloc(struct_size(data, attrs, size_add(stats->num_counters, 1)),
 		       GFP_KERNEL);
 	if (!data)
 		goto err_free_stats;
@@ -1140,7 +1144,7 @@ static int setup_gid_attrs(struct ib_port *port,
 	int ret;
 
 	gid_attr_group = kzalloc(struct_size(gid_attr_group, attrs_list,
-					     attr->gid_tbl_len * 2),
+					     size_mul(attr->gid_tbl_len, 2)),
 				 GFP_KERNEL);
 	if (!gid_attr_group)
 		return -ENOMEM;
@@ -1205,8 +1209,8 @@ static struct ib_port *setup_port(struct ib_core_device *coredev, int port_num,
 	int ret;
 
 	p = kvzalloc(struct_size(p, attrs_list,
-				attr->gid_tbl_len + attr->pkey_tbl_len),
-		    GFP_KERNEL);
+				size_add(attr->gid_tbl_len, attr->pkey_tbl_len)),
+		     GFP_KERNEL);
 	if (!p)
 		return ERR_PTR(-ENOMEM);
 	p->ibdev = device;

drivers/infiniband/core/user_mad.c

@@ -1378,7 +1378,9 @@ static int ib_umad_add_one(struct ib_device *device)
 	s = rdma_start_port(device);
 	e = rdma_end_port(device);
 
-	umad_dev = kzalloc(struct_size(umad_dev, ports, e - s + 1), GFP_KERNEL);
+	umad_dev = kzalloc(struct_size(umad_dev, ports,
+				       size_add(size_sub(e, s), 1)),
+			   GFP_KERNEL);
 	if (!umad_dev)
 		return -ENOMEM;
 

drivers/infiniband/core/uverbs_cmd.c

@@ -1851,7 +1851,8 @@ static int modify_qp(struct uverbs_attr_bundle *attrs,
 	if (cmd->base.attr_mask & IB_QP_PATH_MIG_STATE)
 		attr->path_mig_state = cmd->base.path_mig_state;
 	if (cmd->base.attr_mask & IB_QP_QKEY) {
-		if (cmd->base.qkey & IB_QP_SET_QKEY && !capable(CAP_NET_RAW)) {
+		if (cmd->base.qkey & IB_QP_SET_QKEY &&
+		    !rdma_nl_get_privileged_qkey()) {
 			ret = -EPERM;
 			goto release_qp;
 		}