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

Pull infiniband updates from Roland Dreier:
 "First batch of InfiniBand/RDMA changes for the 3.7 merge window:
   - mlx4 IB support for SR-IOV
   - A couple of SRP initiator fixes
   - Batch of nes hardware driver fixes
   - Fix for long-standing use-after-free crash in IPoIB
   - Other miscellaneous fixes"

This merge also removes a new use of __cancel_delayed_work(), and
replaces it with the regular cancel_delayed_work() that is now irq-safe
thanks to the workqueue updates.

That said, I suspect the sequence in question should probably use
"mod_delayed_work()".  I just did the minimal "don't use deprecated
functions" fixup, though.

* tag 'rdma-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband: (45 commits)
  IB/qib: Fix local access validation for user MRs
  mlx4_core: Disable SENSE_PORT for multifunction devices
  mlx4_core: Clean up enabling of SENSE_PORT for older (ConnectX-1/-2) HCAs
  mlx4_core: Stash PCI ID driver_data in mlx4_priv structure
  IB/srp: Avoid having aborted requests hang
  IB/srp: Fix use-after-free in srp_reset_req()
  IB/qib: Add a qib driver version
  RDMA/nes: Fix compilation error when nes_debug is enabled
  RDMA/nes: Print hardware resource type
  RDMA/nes: Fix for crash when TX checksum offload is off
  RDMA/nes: Cosmetic changes
  RDMA/nes: Fix for incorrect MSS when TSO is on
  RDMA/nes: Fix incorrect resolving of the loopback MAC address
  mlx4_core: Fix crash on uninitialized priv->cmd.slave_sem
  mlx4_core: Trivial cleanups to driver log messages
  mlx4_core: Trivial readability fix: "0X30" -> "0x30"
  IB/mlx4: Create paravirt contexts for VFs when master IB driver initializes
  mlx4: Modify proxy/tunnel QP mechanism so that guests do no calculations
  mlx4: Paravirtualize Node Guids for slaves
  mlx4: Activate SR-IOV mode for IB
  ...

5 files changed, 87 insertions, 6 deletions

diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h
index 6e1b0f973a03..6d1acb04cd17 100644
--- a/include/linux/mlx4/device.h
+++ b/include/linux/mlx4/device.h
@@ -54,7 +54,13 @@ enum {
 };
 
 enum {
-	MLX4_MAX_PORTS		= 2
+	MLX4_PORT_CAP_IS_SM	= 1 << 1,
+	MLX4_PORT_CAP_DEV_MGMT_SUP = 1 << 19,
+};
+
+enum {
+	MLX4_MAX_PORTS		= 2,
+	MLX4_MAX_PORT_PKEYS	= 128
 };
 
 /* base qkey for use in sriov tunnel-qp/proxy-qp communication.
@@ -191,6 +197,25 @@ enum {
 	MLX4_FATAL_WARNING_SUBTYPE_WARMING = 0,
 };
 
+enum slave_port_state {
+	SLAVE_PORT_DOWN = 0,
+	SLAVE_PENDING_UP,
+	SLAVE_PORT_UP,
+};
+
+enum slave_port_gen_event {
+	SLAVE_PORT_GEN_EVENT_DOWN = 0,
+	SLAVE_PORT_GEN_EVENT_UP,
+	SLAVE_PORT_GEN_EVENT_NONE,
+};
+
+enum slave_port_state_event {
+	MLX4_PORT_STATE_DEV_EVENT_PORT_DOWN,
+	MLX4_PORT_STATE_DEV_EVENT_PORT_UP,
+	MLX4_PORT_STATE_IB_PORT_STATE_EVENT_GID_VALID,
+	MLX4_PORT_STATE_IB_EVENT_GID_INVALID,
+};
+
 enum {
 	MLX4_PERM_LOCAL_READ	= 1 << 10,
 	MLX4_PERM_LOCAL_WRITE	= 1 << 11,
@@ -303,6 +328,9 @@ struct mlx4_phys_caps {
 	u32			gid_phys_table_len[MLX4_MAX_PORTS + 1];
 	u32			pkey_phys_table_len[MLX4_MAX_PORTS + 1];
 	u32			num_phys_eqs;
+	u32			base_sqpn;
+	u32			base_proxy_sqpn;
+	u32			base_tunnel_sqpn;
 };
 
 struct mlx4_caps {
@@ -333,9 +361,10 @@ struct mlx4_caps {
 	int			max_rq_desc_sz;
 	int			max_qp_init_rdma;
 	int			max_qp_dest_rdma;
-	int			sqp_start;
-	u32			base_sqpn;
-	u32			base_tunnel_sqpn;
+	u32			*qp0_proxy;
+	u32			*qp1_proxy;
+	u32			*qp0_tunnel;
+	u32			*qp1_tunnel;
 	int			num_srqs;
 	int			max_srq_wqes;
 	int			max_srq_sge;
@@ -389,6 +418,7 @@ struct mlx4_caps {
 	enum mlx4_port_type	possible_type[MLX4_MAX_PORTS + 1];
 	u32			max_counters;
 	u8			port_ib_mtu[MLX4_MAX_PORTS + 1];
+	u16			sqp_demux;
 };
 
 struct mlx4_buf_list {
@@ -671,6 +701,10 @@ struct mlx4_init_port_param {
 	for ((port) = 1; (port) <= (dev)->caps.num_ports; (port)++)	\
 		if ((type) == (dev)->caps.port_mask[(port)])
 
+#define mlx4_foreach_non_ib_transport_port(port, dev)                     \
+	for ((port) = 1; (port) <= (dev)->caps.num_ports; (port)++)	  \
+		if (((dev)->caps.port_mask[port] != MLX4_PORT_TYPE_IB))
+
 #define mlx4_foreach_ib_transport_port(port, dev)                         \
 	for ((port) = 1; (port) <= (dev)->caps.num_ports; (port)++)	  \
 		if (((dev)->caps.port_mask[port] == MLX4_PORT_TYPE_IB) || \
@@ -692,7 +726,18 @@ static inline int mlx4_is_master(struct mlx4_dev *dev)
 
 static inline int mlx4_is_qp_reserved(struct mlx4_dev *dev, u32 qpn)
 {
-	return (qpn < dev->caps.sqp_start + 8);
+	return (qpn < dev->phys_caps.base_sqpn + 8 +
+		16 * MLX4_MFUNC_MAX * !!mlx4_is_master(dev));
+}
+
+static inline int mlx4_is_guest_proxy(struct mlx4_dev *dev, int slave, u32 qpn)
+{
+	int guest_proxy_base = dev->phys_caps.base_proxy_sqpn + slave * 8;
+
+	if (qpn >= guest_proxy_base && qpn < guest_proxy_base + 8)
+		return 1;
+
+	return 0;
 }
 
 static inline int mlx4_is_mfunc(struct mlx4_dev *dev)
@@ -927,6 +972,20 @@ int mlx4_flow_attach(struct mlx4_dev *dev,
 		     struct mlx4_net_trans_rule *rule, u64 *reg_id);
 int mlx4_flow_detach(struct mlx4_dev *dev, u64 reg_id);
 
+void mlx4_sync_pkey_table(struct mlx4_dev *dev, int slave, int port,
+			  int i, int val);
+
 int mlx4_get_parav_qkey(struct mlx4_dev *dev, u32 qpn, u32 *qkey);
 
+int mlx4_is_slave_active(struct mlx4_dev *dev, int slave);
+int mlx4_gen_pkey_eqe(struct mlx4_dev *dev, int slave, u8 port);
+int mlx4_gen_guid_change_eqe(struct mlx4_dev *dev, int slave, u8 port);
+int mlx4_gen_slaves_port_mgt_ev(struct mlx4_dev *dev, u8 port, int attr);
+int mlx4_gen_port_state_change_eqe(struct mlx4_dev *dev, int slave, u8 port, u8 port_subtype_change);
+enum slave_port_state mlx4_get_slave_port_state(struct mlx4_dev *dev, int slave, u8 port);
+int set_and_calc_slave_port_state(struct mlx4_dev *dev, int slave, u8 port, int event, enum slave_port_gen_event *gen_event);
+
+void mlx4_put_slave_node_guid(struct mlx4_dev *dev, int slave, __be64 guid);
+__be64 mlx4_get_slave_node_guid(struct mlx4_dev *dev, int slave);
+
 #endif /* MLX4_DEVICE_H */
diff --git a/include/linux/mlx4/driver.h b/include/linux/mlx4/driver.h
index d813704b963b..c257e1b211be 100644
--- a/include/linux/mlx4/driver.h
+++ b/include/linux/mlx4/driver.h
@@ -45,6 +45,8 @@ enum mlx4_dev_event {
 	MLX4_DEV_EVENT_PORT_DOWN,
 	MLX4_DEV_EVENT_PORT_REINIT,
 	MLX4_DEV_EVENT_PORT_MGMT_CHANGE,
+	MLX4_DEV_EVENT_SLAVE_INIT,
+	MLX4_DEV_EVENT_SLAVE_SHUTDOWN,
 };
 
 struct mlx4_interface {
diff --git a/include/linux/mlx4/qp.h b/include/linux/mlx4/qp.h
index 338388ba260a..4b4ad6ffef92 100644
--- a/include/linux/mlx4/qp.h
+++ b/include/linux/mlx4/qp.h
@@ -126,7 +126,8 @@ struct mlx4_rss_context {
 
 struct mlx4_qp_path {
 	u8			fl;
-	u8			reserved1[2];
+	u8			reserved1[1];
+	u8			disable_pkey_check;
 	u8			pkey_index;
 	u8			counter_index;
 	u8			grh_mylmc;
diff --git a/include/rdma/ib_cache.h b/include/rdma/ib_cache.h
index 00a2b8ec327f..ad9a3c280944 100644
--- a/include/rdma/ib_cache.h
+++ b/include/rdma/ib_cache.h
@@ -101,6 +101,22 @@ int ib_find_cached_pkey(struct ib_device    *device,
 			u16                 *index);
 
 /**
+ * ib_find_exact_cached_pkey - Returns the PKey table index where a specified
+ *   PKey value occurs. Comparison uses the FULL 16 bits (incl membership bit)
+ * @device: The device to query.
+ * @port_num: The port number of the device to search for the PKey.
+ * @pkey: The PKey value to search for.
+ * @index: The index into the cached PKey table where the PKey was found.
+ *
+ * ib_find_exact_cached_pkey() searches the specified PKey table in
+ * the local software cache.
+ */
+int ib_find_exact_cached_pkey(struct ib_device    *device,
+			      u8                   port_num,
+			      u16                  pkey,
+			      u16                 *index);
+
+/**
  * ib_get_cached_lmc - Returns a cached lmc table entry
  * @device: The device to query.
  * @port_num: The port number of the device to query.
diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index 07996af8265a..46bc045bbe15 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -614,6 +614,9 @@ enum ib_qp_type {
 enum ib_qp_create_flags {
 	IB_QP_CREATE_IPOIB_UD_LSO		= 1 << 0,
 	IB_QP_CREATE_BLOCK_MULTICAST_LOOPBACK	= 1 << 1,
+	/* reserve bits 26-31 for low level drivers' internal use */
+	IB_QP_CREATE_RESERVED_START		= 1 << 26,
+	IB_QP_CREATE_RESERVED_END		= 1 << 31,
 };
 
 struct ib_qp_init_attr {