nathan/Chromebook-Device-Nyan-NVIDIA-Tegra-K1-T124-Upstream-Linux-3.10-LTS-Kernel
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Initial commit; kernel source import

Browse Source
This commit is contained in:
Nathan
2025-04-06 23:50:55 -05:00
commit 25c6d769f4
45093 changed files with 18199410 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

49
drivers/infiniband/ulp/ipoib/Kconfig Normal file
View File

@@ -0,0 +1,49 @@
config INFINIBAND_IPOIB
tristate "IP-over-InfiniBand"
depends on NETDEVICES && INET && (IPV6 || IPV6=n)
---help---
Support for the IP-over-InfiniBand protocol (IPoIB). This
transports IP packets over InfiniBand so you can use your IB
device as a fancy NIC.
See Documentation/infiniband/ipoib.txt for more information
config INFINIBAND_IPOIB_CM
bool "IP-over-InfiniBand Connected Mode support"
depends on INFINIBAND_IPOIB
default n
---help---
This option enables support for IPoIB connected mode. After
enabling this option, you need to switch to connected mode
through /sys/class/net/ibXXX/mode to actually create
connections, and then increase the interface MTU with
e.g. ifconfig ib0 mtu 65520.
WARNING: Enabling connected mode will trigger some packet
drops for multicast and UD mode traffic from this interface,
unless you limit mtu for these destinations to 2044.
config INFINIBAND_IPOIB_DEBUG
bool "IP-over-InfiniBand debugging" if EXPERT
depends on INFINIBAND_IPOIB
default y
---help---
This option causes debugging code to be compiled into the
IPoIB driver. The output can be turned on via the
debug_level and mcast_debug_level module parameters (which
can also be set after the driver is loaded through sysfs).
This option also creates a directory tree under ipoib/ in
debugfs, which contains files that expose debugging
information about IB multicast groups used by the IPoIB
driver.
config INFINIBAND_IPOIB_DEBUG_DATA
bool "IP-over-InfiniBand data path debugging"
depends on INFINIBAND_IPOIB_DEBUG
---help---
This option compiles debugging code into the data path
of the IPoIB driver. The output can be turned on via the
data_debug_level module parameter; however, even with output
turned off, this debugging code will have some performance
impact.

12
drivers/infiniband/ulp/ipoib/Makefile Normal file
View File

@@ -0,0 +1,12 @@
obj-$(CONFIG_INFINIBAND_IPOIB) += ib_ipoib.o
ib_ipoib-y := ipoib_main.o \
ipoib_ib.o \
ipoib_multicast.o \
ipoib_verbs.o \
ipoib_vlan.o \
ipoib_ethtool.o \
ipoib_netlink.o
ib_ipoib-$(CONFIG_INFINIBAND_IPOIB_CM) += ipoib_cm.o
ib_ipoib-$(CONFIG_INFINIBAND_IPOIB_DEBUG) += ipoib_fs.o

767
drivers/infiniband/ulp/ipoib/ipoib.h Normal file
View File

@@ -0,0 +1,767 @@
/*
* Copyright (c) 2004, 2005 Topspin Communications. All rights reserved.
* Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved.
* Copyright (c) 2004 Voltaire, Inc. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#ifndef _IPOIB_H
#define _IPOIB_H
#include <linux/list.h>
#include <linux/skbuff.h>
#include <linux/netdevice.h>
#include <linux/workqueue.h>
#include <linux/kref.h>
#include <linux/if_infiniband.h>
#include <linux/mutex.h>
#include <net/neighbour.h>
#include <net/sch_generic.h>
#include <linux/atomic.h>
#include <rdma/ib_verbs.h>
#include <rdma/ib_pack.h>
#include <rdma/ib_sa.h>
#include <linux/sched.h>
/* constants */
enum ipoib_flush_level {
IPOIB_FLUSH_LIGHT,
IPOIB_FLUSH_NORMAL,
IPOIB_FLUSH_HEAVY
};
enum {
IPOIB_ENCAP_LEN = 4,
IPOIB_UD_HEAD_SIZE = IB_GRH_BYTES + IPOIB_ENCAP_LEN,
IPOIB_UD_RX_SG = 2, /* max buffer needed for 4K mtu */
IPOIB_CM_MTU = 0x10000 - 0x10, /* padding to align header to 16 */
IPOIB_CM_BUF_SIZE = IPOIB_CM_MTU + IPOIB_ENCAP_LEN,
IPOIB_CM_HEAD_SIZE = IPOIB_CM_BUF_SIZE % PAGE_SIZE,
IPOIB_CM_RX_SG = ALIGN(IPOIB_CM_BUF_SIZE, PAGE_SIZE) / PAGE_SIZE,
IPOIB_RX_RING_SIZE = 256,
IPOIB_TX_RING_SIZE = 128,
IPOIB_MAX_QUEUE_SIZE = 8192,
IPOIB_MIN_QUEUE_SIZE = 2,
IPOIB_CM_MAX_CONN_QP = 4096,
IPOIB_NUM_WC = 4,
IPOIB_MAX_PATH_REC_QUEUE = 3,
IPOIB_MAX_MCAST_QUEUE = 3,
IPOIB_FLAG_OPER_UP = 0,
IPOIB_FLAG_INITIALIZED = 1,
IPOIB_FLAG_ADMIN_UP = 2,
IPOIB_PKEY_ASSIGNED = 3,
IPOIB_PKEY_STOP = 4,
IPOIB_FLAG_SUBINTERFACE = 5,
IPOIB_MCAST_RUN = 6,
IPOIB_STOP_REAPER = 7,
IPOIB_FLAG_ADMIN_CM = 9,
IPOIB_FLAG_UMCAST = 10,
IPOIB_STOP_NEIGH_GC = 11,
IPOIB_NEIGH_TBL_FLUSH = 12,
IPOIB_MAX_BACKOFF_SECONDS = 16,
IPOIB_MCAST_FLAG_FOUND = 0, /* used in set_multicast_list */
IPOIB_MCAST_FLAG_SENDONLY = 1,
IPOIB_MCAST_FLAG_BUSY = 2, /* joining or already joined */
IPOIB_MCAST_FLAG_ATTACHED = 3,
MAX_SEND_CQE = 16,
IPOIB_CM_COPYBREAK = 256,
IPOIB_NON_CHILD = 0,
IPOIB_LEGACY_CHILD = 1,
IPOIB_RTNL_CHILD = 2,
};
#define IPOIB_OP_RECV (1ul << 31)
#ifdef CONFIG_INFINIBAND_IPOIB_CM
#define IPOIB_OP_CM (1ul << 30)
#else
#define IPOIB_OP_CM (0)
#endif
#define IPOIB_QPN_MASK ((__force u32) cpu_to_be32(0xFFFFFF))
/* structs */
struct ipoib_header {
__be16 proto;
u16 reserved;
};
struct ipoib_cb {
struct qdisc_skb_cb qdisc_cb;
u8 hwaddr[INFINIBAND_ALEN];
};
/* Used for all multicast joins (broadcast, IPv4 mcast and IPv6 mcast) */
struct ipoib_mcast {
struct ib_sa_mcmember_rec mcmember;
struct ib_sa_multicast *mc;
struct ipoib_ah *ah;
struct rb_node rb_node;
struct list_head list;
unsigned long created;
unsigned long backoff;
unsigned long flags;
unsigned char logcount;
struct list_head neigh_list;
struct sk_buff_head pkt_queue;
struct net_device *dev;
};
struct ipoib_rx_buf {
struct sk_buff *skb;
u64 mapping[IPOIB_UD_RX_SG];
};
struct ipoib_tx_buf {
struct sk_buff *skb;
u64 mapping[MAX_SKB_FRAGS + 1];
};
struct ipoib_cm_tx_buf {
struct sk_buff *skb;
u64 mapping;
};
struct ib_cm_id;
struct ipoib_cm_data {
__be32 qpn; /* High byte MUST be ignored on receive */
__be32 mtu;
};
/*
* Quoting 10.3.1 Queue Pair and EE Context States:
*
* Note, for QPs that are associated with an SRQ, the Consumer should take the
* QP through the Error State before invoking a Destroy QP or a Modify QP to the
* Reset State. The Consumer may invoke the Destroy QP without first performing
* a Modify QP to the Error State and waiting for the Affiliated Asynchronous
* Last WQE Reached Event. However, if the Consumer does not wait for the
* Affiliated Asynchronous Last WQE Reached Event, then WQE and Data Segment
* leakage may occur. Therefore, it is good programming practice to tear down a
* QP that is associated with an SRQ by using the following process:
*
* - Put the QP in the Error State
* - Wait for the Affiliated Asynchronous Last WQE Reached Event;
* - either:
* drain the CQ by invoking the Poll CQ verb and either wait for CQ
* to be empty or the number of Poll CQ operations has exceeded
* CQ capacity size;
* - or
* post another WR that completes on the same CQ and wait for this
* WR to return as a WC;
* - and then invoke a Destroy QP or Reset QP.
*
* We use the second option and wait for a completion on the
* same CQ before destroying QPs attached to our SRQ.
*/
enum ipoib_cm_state {
IPOIB_CM_RX_LIVE,
IPOIB_CM_RX_ERROR, /* Ignored by stale task */
IPOIB_CM_RX_FLUSH /* Last WQE Reached event observed */
};
struct ipoib_cm_rx {
struct ib_cm_id *id;
struct ib_qp *qp;
struct ipoib_cm_rx_buf *rx_ring;
struct list_head list;
struct net_device *dev;
unsigned long jiffies;
enum ipoib_cm_state state;
int recv_count;
};
struct ipoib_cm_tx {
struct ib_cm_id *id;
struct ib_qp *qp;
struct list_head list;
struct net_device *dev;
struct ipoib_neigh *neigh;
struct ipoib_path *path;
struct ipoib_cm_tx_buf *tx_ring;
unsigned tx_head;
unsigned tx_tail;
unsigned long flags;
u32 mtu;
};
struct ipoib_cm_rx_buf {
struct sk_buff *skb;
u64 mapping[IPOIB_CM_RX_SG];
};
struct ipoib_cm_dev_priv {
struct ib_srq *srq;
struct ipoib_cm_rx_buf *srq_ring;
struct ib_cm_id *id;
struct list_head passive_ids; /* state: LIVE */
struct list_head rx_error_list; /* state: ERROR */
struct list_head rx_flush_list; /* state: FLUSH, drain not started */
struct list_head rx_drain_list; /* state: FLUSH, drain started */
struct list_head rx_reap_list; /* state: FLUSH, drain done */
struct work_struct start_task;
struct work_struct reap_task;
struct work_struct skb_task;
struct work_struct rx_reap_task;
struct delayed_work stale_task;
struct sk_buff_head skb_queue;
struct list_head start_list;
struct list_head reap_list;
struct ib_wc ibwc[IPOIB_NUM_WC];
struct ib_sge rx_sge[IPOIB_CM_RX_SG];
struct ib_recv_wr rx_wr;
int nonsrq_conn_qp;
int max_cm_mtu;
int num_frags;
};
struct ipoib_ethtool_st {
u16 coalesce_usecs;
u16 max_coalesced_frames;
};
struct ipoib_neigh_table;
struct ipoib_neigh_hash {
struct ipoib_neigh_table *ntbl;
struct ipoib_neigh __rcu **buckets;
struct rcu_head rcu;
u32 mask;
u32 size;
};
struct ipoib_neigh_table {
struct ipoib_neigh_hash __rcu *htbl;
atomic_t entries;
struct completion flushed;
struct completion deleted;
};
/*
* Device private locking: network stack tx_lock protects members used
* in TX fast path, lock protects everything else. lock nests inside
* of tx_lock (ie tx_lock must be acquired first if needed).
*/
struct ipoib_dev_priv {
spinlock_t lock;
struct net_device *dev;
struct napi_struct napi;
unsigned long flags;
struct mutex vlan_mutex;
struct rb_root path_tree;
struct list_head path_list;
struct ipoib_neigh_table ntbl;
struct ipoib_mcast *broadcast;
struct list_head multicast_list;
struct rb_root multicast_tree;
struct delayed_work pkey_poll_task;
struct delayed_work mcast_task;
struct work_struct carrier_on_task;
struct work_struct flush_light;
struct work_struct flush_normal;
struct work_struct flush_heavy;
struct work_struct restart_task;
struct delayed_work ah_reap_task;
struct delayed_work neigh_reap_task;
struct ib_device *ca;
u8 port;
u16 pkey;
u16 pkey_index;
struct ib_pd *pd;
struct ib_mr *mr;
struct ib_cq *recv_cq;
struct ib_cq *send_cq;
struct ib_qp *qp;
u32 qkey;
union ib_gid local_gid;
u16 local_lid;
unsigned int admin_mtu;
unsigned int mcast_mtu;
unsigned int max_ib_mtu;
struct ipoib_rx_buf *rx_ring;
struct ipoib_tx_buf *tx_ring;
unsigned tx_head;
unsigned tx_tail;
struct ib_sge tx_sge[MAX_SKB_FRAGS + 1];
struct ib_send_wr tx_wr;
unsigned tx_outstanding;
struct ib_wc send_wc[MAX_SEND_CQE];
struct ib_recv_wr rx_wr;
struct ib_sge rx_sge[IPOIB_UD_RX_SG];
struct ib_wc ibwc[IPOIB_NUM_WC];
struct list_head dead_ahs;
struct ib_event_handler event_handler;
struct net_device *parent;
struct list_head child_intfs;
struct list_head list;
int child_type;
#ifdef CONFIG_INFINIBAND_IPOIB_CM
struct ipoib_cm_dev_priv cm;
#endif
#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG
struct list_head fs_list;
struct dentry *mcg_dentry;
struct dentry *path_dentry;
#endif
int hca_caps;
struct ipoib_ethtool_st ethtool;
struct timer_list poll_timer;
};
struct ipoib_ah {
struct net_device *dev;
struct ib_ah *ah;
struct list_head list;
struct kref ref;
unsigned last_send;
};
struct ipoib_path {
struct net_device *dev;
struct ib_sa_path_rec pathrec;
struct ipoib_ah *ah;
struct sk_buff_head queue;
struct list_head neigh_list;
int query_id;
struct ib_sa_query *query;
struct completion done;
struct rb_node rb_node;
struct list_head list;
int valid;
};
struct ipoib_neigh {
struct ipoib_ah *ah;
#ifdef CONFIG_INFINIBAND_IPOIB_CM
struct ipoib_cm_tx *cm;
#endif
u8 daddr[INFINIBAND_ALEN];
struct sk_buff_head queue;
struct net_device *dev;
struct list_head list;
struct ipoib_neigh __rcu *hnext;
struct rcu_head rcu;
atomic_t refcnt;
unsigned long alive;
};
#define IPOIB_UD_MTU(ib_mtu) (ib_mtu - IPOIB_ENCAP_LEN)
#define IPOIB_UD_BUF_SIZE(ib_mtu) (ib_mtu + IB_GRH_BYTES)
static inline int ipoib_ud_need_sg(unsigned int ib_mtu)
{
return IPOIB_UD_BUF_SIZE(ib_mtu) > PAGE_SIZE;
}
void ipoib_neigh_dtor(struct ipoib_neigh *neigh);
static inline void ipoib_neigh_put(struct ipoib_neigh *neigh)
{
if (atomic_dec_and_test(&neigh->refcnt))
ipoib_neigh_dtor(neigh);
}
struct ipoib_neigh *ipoib_neigh_get(struct net_device *dev, u8 *daddr);
struct ipoib_neigh *ipoib_neigh_alloc(u8 *daddr,
struct net_device *dev);
void ipoib_neigh_free(struct ipoib_neigh *neigh);
void ipoib_del_neighs_by_gid(struct net_device *dev, u8 *gid);
extern struct workqueue_struct *ipoib_workqueue;
/* functions */
int ipoib_poll(struct napi_struct *napi, int budget);
void ipoib_ib_completion(struct ib_cq *cq, void *dev_ptr);
void ipoib_send_comp_handler(struct ib_cq *cq, void *dev_ptr);
struct ipoib_ah *ipoib_create_ah(struct net_device *dev,
struct ib_pd *pd, struct ib_ah_attr *attr);
void ipoib_free_ah(struct kref *kref);
static inline void ipoib_put_ah(struct ipoib_ah *ah)
{
kref_put(&ah->ref, ipoib_free_ah);
}
int ipoib_open(struct net_device *dev);
int ipoib_add_pkey_attr(struct net_device *dev);
int ipoib_add_umcast_attr(struct net_device *dev);
void ipoib_send(struct net_device *dev, struct sk_buff *skb,
struct ipoib_ah *address, u32 qpn);
void ipoib_reap_ah(struct work_struct *work);
void ipoib_mark_paths_invalid(struct net_device *dev);
void ipoib_flush_paths(struct net_device *dev);
struct ipoib_dev_priv *ipoib_intf_alloc(const char *format);
int ipoib_ib_dev_init(struct net_device *dev, struct ib_device *ca, int port);
void ipoib_ib_dev_flush_light(struct work_struct *work);
void ipoib_ib_dev_flush_normal(struct work_struct *work);
void ipoib_ib_dev_flush_heavy(struct work_struct *work);
void ipoib_pkey_event(struct work_struct *work);
void ipoib_ib_dev_cleanup(struct net_device *dev);
int ipoib_ib_dev_open(struct net_device *dev);
int ipoib_ib_dev_up(struct net_device *dev);
int ipoib_ib_dev_down(struct net_device *dev, int flush);
int ipoib_ib_dev_stop(struct net_device *dev, int flush);
int ipoib_dev_init(struct net_device *dev, struct ib_device *ca, int port);
void ipoib_dev_cleanup(struct net_device *dev);
void ipoib_mcast_join_task(struct work_struct *work);
void ipoib_mcast_carrier_on_task(struct work_struct *work);
void ipoib_mcast_send(struct net_device *dev, u8 *daddr, struct sk_buff *skb);
void ipoib_mcast_restart_task(struct work_struct *work);
int ipoib_mcast_start_thread(struct net_device *dev);
int ipoib_mcast_stop_thread(struct net_device *dev, int flush);
void ipoib_mcast_dev_down(struct net_device *dev);
void ipoib_mcast_dev_flush(struct net_device *dev);
#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG
struct ipoib_mcast_iter *ipoib_mcast_iter_init(struct net_device *dev);
int ipoib_mcast_iter_next(struct ipoib_mcast_iter *iter);
void ipoib_mcast_iter_read(struct ipoib_mcast_iter *iter,
union ib_gid *gid,
unsigned long *created,
unsigned int *queuelen,
unsigned int *complete,
unsigned int *send_only);
struct ipoib_path_iter *ipoib_path_iter_init(struct net_device *dev);
int ipoib_path_iter_next(struct ipoib_path_iter *iter);
void ipoib_path_iter_read(struct ipoib_path_iter *iter,
struct ipoib_path *path);
#endif
int ipoib_mcast_attach(struct net_device *dev, u16 mlid,
union ib_gid *mgid, int set_qkey);
int ipoib_init_qp(struct net_device *dev);
int ipoib_transport_dev_init(struct net_device *dev, struct ib_device *ca);
void ipoib_transport_dev_cleanup(struct net_device *dev);
void ipoib_event(struct ib_event_handler *handler,
struct ib_event *record);
int ipoib_vlan_add(struct net_device *pdev, unsigned short pkey);
int ipoib_vlan_delete(struct net_device *pdev, unsigned short pkey);
int __ipoib_vlan_add(struct ipoib_dev_priv *ppriv, struct ipoib_dev_priv *priv,
u16 pkey, int child_type);
int __init ipoib_netlink_init(void);
void __exit ipoib_netlink_fini(void);
void ipoib_set_umcast(struct net_device *ndev, int umcast_val);
int ipoib_set_mode(struct net_device *dev, const char *buf);
void ipoib_setup(struct net_device *dev);
void ipoib_pkey_poll(struct work_struct *work);
int ipoib_pkey_dev_delay_open(struct net_device *dev);
void ipoib_drain_cq(struct net_device *dev);
void ipoib_set_ethtool_ops(struct net_device *dev);
int ipoib_set_dev_features(struct ipoib_dev_priv *priv, struct ib_device *hca);
#define IPOIB_FLAGS_RC 0x80
#define IPOIB_FLAGS_UC 0x40
/* We don't support UC connections at the moment */
#define IPOIB_CM_SUPPORTED(ha) (ha[0] & (IPOIB_FLAGS_RC))
#ifdef CONFIG_INFINIBAND_IPOIB_CM
extern int ipoib_max_conn_qp;
static inline int ipoib_cm_admin_enabled(struct net_device *dev)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
return IPOIB_CM_SUPPORTED(dev->dev_addr) &&
test_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags);
}
static inline int ipoib_cm_enabled(struct net_device *dev, u8 *hwaddr)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
return IPOIB_CM_SUPPORTED(hwaddr) &&
test_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags);
}
static inline int ipoib_cm_up(struct ipoib_neigh *neigh)
{
return test_bit(IPOIB_FLAG_OPER_UP, &neigh->cm->flags);
}
static inline struct ipoib_cm_tx *ipoib_cm_get(struct ipoib_neigh *neigh)
{
return neigh->cm;
}
static inline void ipoib_cm_set(struct ipoib_neigh *neigh, struct ipoib_cm_tx *tx)
{
neigh->cm = tx;
}
static inline int ipoib_cm_has_srq(struct net_device *dev)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
return !!priv->cm.srq;
}
static inline unsigned int ipoib_cm_max_mtu(struct net_device *dev)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
return priv->cm.max_cm_mtu;
}
void ipoib_cm_send(struct net_device *dev, struct sk_buff *skb, struct ipoib_cm_tx *tx);
int ipoib_cm_dev_open(struct net_device *dev);
void ipoib_cm_dev_stop(struct net_device *dev);
int ipoib_cm_dev_init(struct net_device *dev);
int ipoib_cm_add_mode_attr(struct net_device *dev);
void ipoib_cm_dev_cleanup(struct net_device *dev);
struct ipoib_cm_tx *ipoib_cm_create_tx(struct net_device *dev, struct ipoib_path *path,
struct ipoib_neigh *neigh);
void ipoib_cm_destroy_tx(struct ipoib_cm_tx *tx);
void ipoib_cm_skb_too_long(struct net_device *dev, struct sk_buff *skb,
unsigned int mtu);
void ipoib_cm_handle_rx_wc(struct net_device *dev, struct ib_wc *wc);
void ipoib_cm_handle_tx_wc(struct net_device *dev, struct ib_wc *wc);
#else
struct ipoib_cm_tx;
#define ipoib_max_conn_qp 0
static inline int ipoib_cm_admin_enabled(struct net_device *dev)
{
return 0;
}
static inline int ipoib_cm_enabled(struct net_device *dev, u8 *hwaddr)
{
return 0;
}
static inline int ipoib_cm_up(struct ipoib_neigh *neigh)
{
return 0;
}
static inline struct ipoib_cm_tx *ipoib_cm_get(struct ipoib_neigh *neigh)
{
return NULL;
}
static inline void ipoib_cm_set(struct ipoib_neigh *neigh, struct ipoib_cm_tx *tx)
{
}
static inline int ipoib_cm_has_srq(struct net_device *dev)
{
return 0;
}
static inline unsigned int ipoib_cm_max_mtu(struct net_device *dev)
{
return 0;
}
static inline
void ipoib_cm_send(struct net_device *dev, struct sk_buff *skb, struct ipoib_cm_tx *tx)
{
return;
}
static inline
int ipoib_cm_dev_open(struct net_device *dev)
{
return 0;
}
static inline
void ipoib_cm_dev_stop(struct net_device *dev)
{
return;
}
static inline
int ipoib_cm_dev_init(struct net_device *dev)
{
return -ENOSYS;
}
static inline
void ipoib_cm_dev_cleanup(struct net_device *dev)
{
return;
}
static inline
struct ipoib_cm_tx *ipoib_cm_create_tx(struct net_device *dev, struct ipoib_path *path,
struct ipoib_neigh *neigh)
{
return NULL;
}
static inline
void ipoib_cm_destroy_tx(struct ipoib_cm_tx *tx)
{
return;
}
static inline
int ipoib_cm_add_mode_attr(struct net_device *dev)
{
return 0;
}
static inline void ipoib_cm_skb_too_long(struct net_device *dev, struct sk_buff *skb,
unsigned int mtu)
{
dev_kfree_skb_any(skb);
}
static inline void ipoib_cm_handle_rx_wc(struct net_device *dev, struct ib_wc *wc)
{
}
static inline void ipoib_cm_handle_tx_wc(struct net_device *dev, struct ib_wc *wc)
{
}
#endif
#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG
void ipoib_create_debug_files(struct net_device *dev);
void ipoib_delete_debug_files(struct net_device *dev);
int ipoib_register_debugfs(void);
void ipoib_unregister_debugfs(void);
#else
static inline void ipoib_create_debug_files(struct net_device *dev) { }
static inline void ipoib_delete_debug_files(struct net_device *dev) { }
static inline int ipoib_register_debugfs(void) { return 0; }
static inline void ipoib_unregister_debugfs(void) { }
#endif
#define ipoib_printk(level, priv, format, arg...) \
printk(level "%s: " format, ((struct ipoib_dev_priv *) priv)->dev->name , ## arg)
#define ipoib_warn(priv, format, arg...) \
ipoib_printk(KERN_WARNING, priv, format , ## arg)
extern int ipoib_sendq_size;
extern int ipoib_recvq_size;
extern struct ib_sa_client ipoib_sa_client;
#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG
extern int ipoib_debug_level;
#define ipoib_dbg(priv, format, arg...) \
do { \
if (ipoib_debug_level > 0) \
ipoib_printk(KERN_DEBUG, priv, format , ## arg); \
} while (0)
#define ipoib_dbg_mcast(priv, format, arg...) \
do { \
if (mcast_debug_level > 0) \
ipoib_printk(KERN_DEBUG, priv, format , ## arg); \
} while (0)
#else /* CONFIG_INFINIBAND_IPOIB_DEBUG */
#define ipoib_dbg(priv, format, arg...) \
do { (void) (priv); } while (0)
#define ipoib_dbg_mcast(priv, format, arg...) \
do { (void) (priv); } while (0)
#endif /* CONFIG_INFINIBAND_IPOIB_DEBUG */
#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG_DATA
#define ipoib_dbg_data(priv, format, arg...) \
do { \
if (data_debug_level > 0) \
ipoib_printk(KERN_DEBUG, priv, format , ## arg); \
} while (0)
#else /* CONFIG_INFINIBAND_IPOIB_DEBUG_DATA */
#define ipoib_dbg_data(priv, format, arg...) \
do { (void) (priv); } while (0)
#endif /* CONFIG_INFINIBAND_IPOIB_DEBUG_DATA */
#define IPOIB_QPN(ha) (be32_to_cpup((__be32 *) ha) & 0xffffff)
extern const char ipoib_driver_version[];
#endif /* _IPOIB_H */

1602
drivers/infiniband/ulp/ipoib/ipoib_cm.c Normal file
View File

File diff suppressed because it is too large Load Diff

109
drivers/infiniband/ulp/ipoib/ipoib_ethtool.c Normal file
View File

@@ -0,0 +1,109 @@
/*
* Copyright (c) 2007 Mellanox Technologies. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#include <linux/kernel.h>
#include <linux/ethtool.h>
#include <linux/netdevice.h>
#include "ipoib.h"
static void ipoib_get_drvinfo(struct net_device *netdev,
struct ethtool_drvinfo *drvinfo)
{
struct ipoib_dev_priv *priv = netdev_priv(netdev);
struct ib_device_attr *attr;
attr = kmalloc(sizeof(*attr), GFP_KERNEL);
if (attr && !ib_query_device(priv->ca, attr))
snprintf(drvinfo->fw_version, sizeof(drvinfo->fw_version),
"%d.%d.%d", (int)(attr->fw_ver >> 32),
(int)(attr->fw_ver >> 16) & 0xffff,
(int)attr->fw_ver & 0xffff);
kfree(attr);
strlcpy(drvinfo->bus_info, dev_name(priv->ca->dma_device),
sizeof(drvinfo->bus_info));
strlcpy(drvinfo->version, ipoib_driver_version,
sizeof(drvinfo->version));
strlcpy(drvinfo->driver, "ib_ipoib", sizeof(drvinfo->driver));
}
static int ipoib_get_coalesce(struct net_device *dev,
struct ethtool_coalesce *coal)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
coal->rx_coalesce_usecs = priv->ethtool.coalesce_usecs;
coal->rx_max_coalesced_frames = priv->ethtool.max_coalesced_frames;
return 0;
}
static int ipoib_set_coalesce(struct net_device *dev,
struct ethtool_coalesce *coal)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
int ret;
/*
* These values are saved in the private data and returned
* when ipoib_get_coalesce() is called
*/
if (coal->rx_coalesce_usecs > 0xffff ||
coal->rx_max_coalesced_frames > 0xffff)
return -EINVAL;
ret = ib_modify_cq(priv->recv_cq, coal->rx_max_coalesced_frames,
coal->rx_coalesce_usecs);
if (ret && ret != -ENOSYS) {
ipoib_warn(priv, "failed modifying CQ (%d)\n", ret);
return ret;
}
priv->ethtool.coalesce_usecs = coal->rx_coalesce_usecs;
priv->ethtool.max_coalesced_frames = coal->rx_max_coalesced_frames;
return 0;
}
static const struct ethtool_ops ipoib_ethtool_ops = {
.get_drvinfo = ipoib_get_drvinfo,
.get_coalesce = ipoib_get_coalesce,
.set_coalesce = ipoib_set_coalesce,
};
void ipoib_set_ethtool_ops(struct net_device *dev)
{
SET_ETHTOOL_OPS(dev, &ipoib_ethtool_ops);
}

299
drivers/infiniband/ulp/ipoib/ipoib_fs.c Normal file
View File

@@ -0,0 +1,299 @@
/*
* Copyright (c) 2004 Topspin Communications. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#include <linux/err.h>
#include <linux/seq_file.h>
#include <linux/slab.h>
struct file_operations;
#include <linux/debugfs.h>
#include <linux/export.h>
#include "ipoib.h"
static struct dentry *ipoib_root;
static void format_gid(union ib_gid *gid, char *buf)
{
int i, n;
for (n = 0, i = 0; i < 8; ++i) {
n += sprintf(buf + n, "%x",
be16_to_cpu(((__be16 *) gid->raw)[i]));
if (i < 7)
buf[n++] = ':';
}
}
static void *ipoib_mcg_seq_start(struct seq_file *file, loff_t *pos)
{
struct ipoib_mcast_iter *iter;
loff_t n = *pos;
iter = ipoib_mcast_iter_init(file->private);
if (!iter)
return NULL;
while (n--) {
if (ipoib_mcast_iter_next(iter)) {
kfree(iter);
return NULL;
}
}
return iter;
}
static void *ipoib_mcg_seq_next(struct seq_file *file, void *iter_ptr,
loff_t *pos)
{
struct ipoib_mcast_iter *iter = iter_ptr;
(*pos)++;
if (ipoib_mcast_iter_next(iter)) {
kfree(iter);
return NULL;
}
return iter;
}
static void ipoib_mcg_seq_stop(struct seq_file *file, void *iter_ptr)
{
/* nothing for now */
}
static int ipoib_mcg_seq_show(struct seq_file *file, void *iter_ptr)
{
struct ipoib_mcast_iter *iter = iter_ptr;
char gid_buf[sizeof "ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff"];
union ib_gid mgid;
unsigned long created;
unsigned int queuelen, complete, send_only;
if (!iter)
return 0;
ipoib_mcast_iter_read(iter, &mgid, &created, &queuelen,
&complete, &send_only);
format_gid(&mgid, gid_buf);
seq_printf(file,
"GID: %s\n"
" created: %10ld\n"
" queuelen: %9d\n"
" complete: %9s\n"
" send_only: %8s\n"
"\n",
gid_buf, created, queuelen,
complete ? "yes" : "no",
send_only ? "yes" : "no");
return 0;
}
static const struct seq_operations ipoib_mcg_seq_ops = {
.start = ipoib_mcg_seq_start,
.next = ipoib_mcg_seq_next,
.stop = ipoib_mcg_seq_stop,
.show = ipoib_mcg_seq_show,
};
static int ipoib_mcg_open(struct inode *inode, struct file *file)
{
struct seq_file *seq;
int ret;
ret = seq_open(file, &ipoib_mcg_seq_ops);
if (ret)
return ret;
seq = file->private_data;
seq->private = inode->i_private;
return 0;
}
static const struct file_operations ipoib_mcg_fops = {
.owner = THIS_MODULE,
.open = ipoib_mcg_open,
.read = seq_read,
.llseek = seq_lseek,
.release = seq_release
};
static void *ipoib_path_seq_start(struct seq_file *file, loff_t *pos)
{
struct ipoib_path_iter *iter;
loff_t n = *pos;
iter = ipoib_path_iter_init(file->private);
if (!iter)
return NULL;
while (n--) {
if (ipoib_path_iter_next(iter)) {
kfree(iter);
return NULL;
}
}
return iter;
}
static void *ipoib_path_seq_next(struct seq_file *file, void *iter_ptr,
loff_t *pos)
{
struct ipoib_path_iter *iter = iter_ptr;
(*pos)++;
if (ipoib_path_iter_next(iter)) {
kfree(iter);
return NULL;
}
return iter;
}
static void ipoib_path_seq_stop(struct seq_file *file, void *iter_ptr)
{
/* nothing for now */
}
static int ipoib_path_seq_show(struct seq_file *file, void *iter_ptr)
{
struct ipoib_path_iter *iter = iter_ptr;
char gid_buf[sizeof "ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff"];
struct ipoib_path path;
int rate;
if (!iter)
return 0;
ipoib_path_iter_read(iter, &path);
format_gid(&path.pathrec.dgid, gid_buf);
seq_printf(file,
"GID: %s\n"
" complete: %6s\n",
gid_buf, path.pathrec.dlid ? "yes" : "no");
if (path.pathrec.dlid) {
rate = ib_rate_to_mbps(path.pathrec.rate);
seq_printf(file,
" DLID: 0x%04x\n"
" SL: %12d\n"
" rate: %8d.%d Gb/sec\n",
be16_to_cpu(path.pathrec.dlid),
path.pathrec.sl,
rate / 1000, rate % 1000);
}
seq_putc(file, '\n');
return 0;
}
static const struct seq_operations ipoib_path_seq_ops = {
.start = ipoib_path_seq_start,
.next = ipoib_path_seq_next,
.stop = ipoib_path_seq_stop,
.show = ipoib_path_seq_show,
};
static int ipoib_path_open(struct inode *inode, struct file *file)
{
struct seq_file *seq;
int ret;
ret = seq_open(file, &ipoib_path_seq_ops);
if (ret)
return ret;
seq = file->private_data;
seq->private = inode->i_private;
return 0;
}
static const struct file_operations ipoib_path_fops = {
.owner = THIS_MODULE,
.open = ipoib_path_open,
.read = seq_read,
.llseek = seq_lseek,
.release = seq_release
};
void ipoib_create_debug_files(struct net_device *dev)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
char name[IFNAMSIZ + sizeof "_path"];
snprintf(name, sizeof name, "%s_mcg", dev->name);
priv->mcg_dentry = debugfs_create_file(name, S_IFREG | S_IRUGO,
ipoib_root, dev, &ipoib_mcg_fops);
if (!priv->mcg_dentry)
ipoib_warn(priv, "failed to create mcg debug file\n");
snprintf(name, sizeof name, "%s_path", dev->name);
priv->path_dentry = debugfs_create_file(name, S_IFREG | S_IRUGO,
ipoib_root, dev, &ipoib_path_fops);
if (!priv->path_dentry)
ipoib_warn(priv, "failed to create path debug file\n");
}
void ipoib_delete_debug_files(struct net_device *dev)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
if (priv->mcg_dentry)
debugfs_remove(priv->mcg_dentry);
if (priv->path_dentry)
debugfs_remove(priv->path_dentry);
}
int ipoib_register_debugfs(void)
{
ipoib_root = debugfs_create_dir("ipoib", NULL);
return ipoib_root ? 0 : -ENOMEM;
}
void ipoib_unregister_debugfs(void)
{
debugfs_remove(ipoib_root);
}

1091
drivers/infiniband/ulp/ipoib/ipoib_ib.c Normal file
View File

File diff suppressed because it is too large Load Diff

1797
drivers/infiniband/ulp/ipoib/ipoib_main.c Normal file
View File

File diff suppressed because it is too large Load Diff

950
drivers/infiniband/ulp/ipoib/ipoib_multicast.c Normal file
View File

@@ -0,0 +1,950 @@
/*
* Copyright (c) 2004, 2005 Topspin Communications. All rights reserved.
* Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved.
* Copyright (c) 2004 Voltaire, Inc. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#include <linux/skbuff.h>
#include <linux/rtnetlink.h>
#include <linux/moduleparam.h>
#include <linux/ip.h>
#include <linux/in.h>
#include <linux/igmp.h>
#include <linux/inetdevice.h>
#include <linux/delay.h>
#include <linux/completion.h>
#include <linux/slab.h>
#include <net/dst.h>
#include "ipoib.h"
#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG
static int mcast_debug_level;
module_param(mcast_debug_level, int, 0644);
MODULE_PARM_DESC(mcast_debug_level,
"Enable multicast debug tracing if > 0");
#endif
static DEFINE_MUTEX(mcast_mutex);
struct ipoib_mcast_iter {
struct net_device *dev;
union ib_gid mgid;
unsigned long created;
unsigned int queuelen;
unsigned int complete;
unsigned int send_only;
};
static void ipoib_mcast_free(struct ipoib_mcast *mcast)
{
struct net_device *dev = mcast->dev;
int tx_dropped = 0;
ipoib_dbg_mcast(netdev_priv(dev), "deleting multicast group %pI6\n",
mcast->mcmember.mgid.raw);
/* remove all neigh connected to this mcast */
ipoib_del_neighs_by_gid(dev, mcast->mcmember.mgid.raw);
if (mcast->ah)
ipoib_put_ah(mcast->ah);
while (!skb_queue_empty(&mcast->pkt_queue)) {
++tx_dropped;
dev_kfree_skb_any(skb_dequeue(&mcast->pkt_queue));
}
netif_tx_lock_bh(dev);
dev->stats.tx_dropped += tx_dropped;
netif_tx_unlock_bh(dev);
kfree(mcast);
}
static struct ipoib_mcast *ipoib_mcast_alloc(struct net_device *dev,
int can_sleep)
{
struct ipoib_mcast *mcast;
mcast = kzalloc(sizeof *mcast, can_sleep ? GFP_KERNEL : GFP_ATOMIC);
if (!mcast)
return NULL;
mcast->dev = dev;
mcast->created = jiffies;
mcast->backoff = 1;
INIT_LIST_HEAD(&mcast->list);
INIT_LIST_HEAD(&mcast->neigh_list);
skb_queue_head_init(&mcast->pkt_queue);
return mcast;
}
static struct ipoib_mcast *__ipoib_mcast_find(struct net_device *dev, void *mgid)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
struct rb_node *n = priv->multicast_tree.rb_node;
while (n) {
struct ipoib_mcast *mcast;
int ret;
mcast = rb_entry(n, struct ipoib_mcast, rb_node);
ret = memcmp(mgid, mcast->mcmember.mgid.raw,
sizeof (union ib_gid));
if (ret < 0)
n = n->rb_left;
else if (ret > 0)
n = n->rb_right;
else
return mcast;
}
return NULL;
}
static int __ipoib_mcast_add(struct net_device *dev, struct ipoib_mcast *mcast)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
struct rb_node **n = &priv->multicast_tree.rb_node, *pn = NULL;
while (*n) {
struct ipoib_mcast *tmcast;
int ret;
pn = *n;
tmcast = rb_entry(pn, struct ipoib_mcast, rb_node);
ret = memcmp(mcast->mcmember.mgid.raw, tmcast->mcmember.mgid.raw,
sizeof (union ib_gid));
if (ret < 0)
n = &pn->rb_left;
else if (ret > 0)
n = &pn->rb_right;
else
return -EEXIST;
}
rb_link_node(&mcast->rb_node, pn, n);
rb_insert_color(&mcast->rb_node, &priv->multicast_tree);
return 0;
}
static int ipoib_mcast_join_finish(struct ipoib_mcast *mcast,
struct ib_sa_mcmember_rec *mcmember)
{
struct net_device *dev = mcast->dev;
struct ipoib_dev_priv *priv = netdev_priv(dev);
struct ipoib_ah *ah;
int ret;
int set_qkey = 0;
mcast->mcmember = *mcmember;
/* Set the multicast MTU and cached Q_Key before we attach if it's
* the broadcast group.
*/
if (!memcmp(mcast->mcmember.mgid.raw, priv->dev->broadcast + 4,
sizeof (union ib_gid))) {
spin_lock_irq(&priv->lock);
if (!priv->broadcast) {
spin_unlock_irq(&priv->lock);
return -EAGAIN;
}
priv->mcast_mtu = IPOIB_UD_MTU(ib_mtu_enum_to_int(priv->broadcast->mcmember.mtu));
priv->qkey = be32_to_cpu(priv->broadcast->mcmember.qkey);
spin_unlock_irq(&priv->lock);
priv->tx_wr.wr.ud.remote_qkey = priv->qkey;
set_qkey = 1;
if (!ipoib_cm_admin_enabled(dev)) {
rtnl_lock();
dev_set_mtu(dev, min(priv->mcast_mtu, priv->admin_mtu));
rtnl_unlock();
}
}
if (!test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags)) {
if (test_and_set_bit(IPOIB_MCAST_FLAG_ATTACHED, &mcast->flags)) {
ipoib_warn(priv, "multicast group %pI6 already attached\n",
mcast->mcmember.mgid.raw);
return 0;
}
ret = ipoib_mcast_attach(dev, be16_to_cpu(mcast->mcmember.mlid),
&mcast->mcmember.mgid, set_qkey);
if (ret < 0) {
ipoib_warn(priv, "couldn't attach QP to multicast group %pI6\n",
mcast->mcmember.mgid.raw);
clear_bit(IPOIB_MCAST_FLAG_ATTACHED, &mcast->flags);
return ret;
}
}
{
struct ib_ah_attr av = {
.dlid = be16_to_cpu(mcast->mcmember.mlid),
.port_num = priv->port,
.sl = mcast->mcmember.sl,
.ah_flags = IB_AH_GRH,
.static_rate = mcast->mcmember.rate,
.grh = {
.flow_label = be32_to_cpu(mcast->mcmember.flow_label),
.hop_limit = mcast->mcmember.hop_limit,
.sgid_index = 0,
.traffic_class = mcast->mcmember.traffic_class
}
};
av.grh.dgid = mcast->mcmember.mgid;
ah = ipoib_create_ah(dev, priv->pd, &av);
if (IS_ERR(ah)) {
ipoib_warn(priv, "ib_address_create failed %ld\n",
-PTR_ERR(ah));
/* use original error */
return PTR_ERR(ah);
} else {
spin_lock_irq(&priv->lock);
mcast->ah = ah;
spin_unlock_irq(&priv->lock);
ipoib_dbg_mcast(priv, "MGID %pI6 AV %p, LID 0x%04x, SL %d\n",
mcast->mcmember.mgid.raw,
mcast->ah->ah,
be16_to_cpu(mcast->mcmember.mlid),
mcast->mcmember.sl);
}
}
/* actually send any queued packets */
netif_tx_lock_bh(dev);
while (!skb_queue_empty(&mcast->pkt_queue)) {
struct sk_buff *skb = skb_dequeue(&mcast->pkt_queue);
netif_tx_unlock_bh(dev);
skb->dev = dev;
if (dev_queue_xmit(skb))
ipoib_warn(priv, "dev_queue_xmit failed to requeue packet\n");
netif_tx_lock_bh(dev);
}
netif_tx_unlock_bh(dev);
return 0;
}
static int
ipoib_mcast_sendonly_join_complete(int status,
struct ib_sa_multicast *multicast)
{
struct ipoib_mcast *mcast = multicast->context;
struct net_device *dev = mcast->dev;
/* We trap for port events ourselves. */
if (status == -ENETRESET)
return 0;
if (!status)
status = ipoib_mcast_join_finish(mcast, &multicast->rec);
if (status) {
if (mcast->logcount++ < 20)
ipoib_dbg_mcast(netdev_priv(dev), "multicast join failed for %pI6, status %d\n",
mcast->mcmember.mgid.raw, status);
/* Flush out any queued packets */
netif_tx_lock_bh(dev);
while (!skb_queue_empty(&mcast->pkt_queue)) {
++dev->stats.tx_dropped;
dev_kfree_skb_any(skb_dequeue(&mcast->pkt_queue));
}
netif_tx_unlock_bh(dev);
/* Clear the busy flag so we try again */
status = test_and_clear_bit(IPOIB_MCAST_FLAG_BUSY,
&mcast->flags);
}
return status;
}
static int ipoib_mcast_sendonly_join(struct ipoib_mcast *mcast)
{
struct net_device *dev = mcast->dev;
struct ipoib_dev_priv *priv = netdev_priv(dev);
struct ib_sa_mcmember_rec rec = {
#if 0 /* Some SMs don't support send-only yet */
.join_state = 4
#else
.join_state = 1
#endif
};
int ret = 0;
if (!test_bit(IPOIB_FLAG_OPER_UP, &priv->flags)) {
ipoib_dbg_mcast(priv, "device shutting down, no multicast joins\n");
return -ENODEV;
}
if (test_and_set_bit(IPOIB_MCAST_FLAG_BUSY, &mcast->flags)) {
ipoib_dbg_mcast(priv, "multicast entry busy, skipping\n");
return -EBUSY;
}
rec.mgid = mcast->mcmember.mgid;
rec.port_gid = priv->local_gid;
rec.pkey = cpu_to_be16(priv->pkey);
mcast->mc = ib_sa_join_multicast(&ipoib_sa_client, priv->ca,
priv->port, &rec,
IB_SA_MCMEMBER_REC_MGID |
IB_SA_MCMEMBER_REC_PORT_GID |
IB_SA_MCMEMBER_REC_PKEY |
IB_SA_MCMEMBER_REC_JOIN_STATE,
GFP_ATOMIC,
ipoib_mcast_sendonly_join_complete,
mcast);
if (IS_ERR(mcast->mc)) {
ret = PTR_ERR(mcast->mc);
clear_bit(IPOIB_MCAST_FLAG_BUSY, &mcast->flags);
ipoib_warn(priv, "ib_sa_join_multicast failed (ret = %d)\n",
ret);
} else {
ipoib_dbg_mcast(priv, "no multicast record for %pI6, starting join\n",
mcast->mcmember.mgid.raw);
}
return ret;
}
void ipoib_mcast_carrier_on_task(struct work_struct *work)
{
struct ipoib_dev_priv *priv = container_of(work, struct ipoib_dev_priv,
carrier_on_task);
struct ib_port_attr attr;
/*
* Take rtnl_lock to avoid racing with ipoib_stop() and
* turning the carrier back on while a device is being
* removed.
*/
if (ib_query_port(priv->ca, priv->port, &attr) ||
attr.state != IB_PORT_ACTIVE) {
ipoib_dbg(priv, "Keeping carrier off until IB port is active\n");
return;
}
rtnl_lock();
netif_carrier_on(priv->dev);
rtnl_unlock();
}
static int ipoib_mcast_join_complete(int status,
struct ib_sa_multicast *multicast)
{
struct ipoib_mcast *mcast = multicast->context;
struct net_device *dev = mcast->dev;
struct ipoib_dev_priv *priv = netdev_priv(dev);
ipoib_dbg_mcast(priv, "join completion for %pI6 (status %d)\n",
mcast->mcmember.mgid.raw, status);
/* We trap for port events ourselves. */
if (status == -ENETRESET)
return 0;
if (!status)
status = ipoib_mcast_join_finish(mcast, &multicast->rec);
if (!status) {
mcast->backoff = 1;
mutex_lock(&mcast_mutex);
if (test_bit(IPOIB_MCAST_RUN, &priv->flags))
queue_delayed_work(ipoib_workqueue,
&priv->mcast_task, 0);
mutex_unlock(&mcast_mutex);
/*
* Defer carrier on work to ipoib_workqueue to avoid a
* deadlock on rtnl_lock here.
*/
if (mcast == priv->broadcast)
queue_work(ipoib_workqueue, &priv->carrier_on_task);
return 0;
}
if (mcast->logcount++ < 20) {
if (status == -ETIMEDOUT || status == -EAGAIN) {
ipoib_dbg_mcast(priv, "multicast join failed for %pI6, status %d\n",
mcast->mcmember.mgid.raw, status);
} else {
ipoib_warn(priv, "multicast join failed for %pI6, status %d\n",
mcast->mcmember.mgid.raw, status);
}
}
mcast->backoff *= 2;
if (mcast->backoff > IPOIB_MAX_BACKOFF_SECONDS)
mcast->backoff = IPOIB_MAX_BACKOFF_SECONDS;
/* Clear the busy flag so we try again */
status = test_and_clear_bit(IPOIB_MCAST_FLAG_BUSY, &mcast->flags);
mutex_lock(&mcast_mutex);
spin_lock_irq(&priv->lock);
if (test_bit(IPOIB_MCAST_RUN, &priv->flags))
queue_delayed_work(ipoib_workqueue, &priv->mcast_task,
mcast->backoff * HZ);
spin_unlock_irq(&priv->lock);
mutex_unlock(&mcast_mutex);
return status;
}
static void ipoib_mcast_join(struct net_device *dev, struct ipoib_mcast *mcast,
int create)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
struct ib_sa_mcmember_rec rec = {
.join_state = 1
};
ib_sa_comp_mask comp_mask;
int ret = 0;
ipoib_dbg_mcast(priv, "joining MGID %pI6\n", mcast->mcmember.mgid.raw);
rec.mgid = mcast->mcmember.mgid;
rec.port_gid = priv->local_gid;
rec.pkey = cpu_to_be16(priv->pkey);
comp_mask =
IB_SA_MCMEMBER_REC_MGID |
IB_SA_MCMEMBER_REC_PORT_GID |
IB_SA_MCMEMBER_REC_PKEY |
IB_SA_MCMEMBER_REC_JOIN_STATE;
if (create) {
comp_mask |=
IB_SA_MCMEMBER_REC_QKEY |
IB_SA_MCMEMBER_REC_MTU_SELECTOR |
IB_SA_MCMEMBER_REC_MTU |
IB_SA_MCMEMBER_REC_TRAFFIC_CLASS |
IB_SA_MCMEMBER_REC_RATE_SELECTOR |
IB_SA_MCMEMBER_REC_RATE |
IB_SA_MCMEMBER_REC_SL |
IB_SA_MCMEMBER_REC_FLOW_LABEL |
IB_SA_MCMEMBER_REC_HOP_LIMIT;
rec.qkey = priv->broadcast->mcmember.qkey;
rec.mtu_selector = IB_SA_EQ;
rec.mtu = priv->broadcast->mcmember.mtu;
rec.traffic_class = priv->broadcast->mcmember.traffic_class;
rec.rate_selector = IB_SA_EQ;
rec.rate = priv->broadcast->mcmember.rate;
rec.sl = priv->broadcast->mcmember.sl;
rec.flow_label = priv->broadcast->mcmember.flow_label;
rec.hop_limit = priv->broadcast->mcmember.hop_limit;
}
set_bit(IPOIB_MCAST_FLAG_BUSY, &mcast->flags);
mcast->mc = ib_sa_join_multicast(&ipoib_sa_client, priv->ca, priv->port,
&rec, comp_mask, GFP_KERNEL,
ipoib_mcast_join_complete, mcast);
if (IS_ERR(mcast->mc)) {
clear_bit(IPOIB_MCAST_FLAG_BUSY, &mcast->flags);
ret = PTR_ERR(mcast->mc);
ipoib_warn(priv, "ib_sa_join_multicast failed, status %d\n", ret);
mcast->backoff *= 2;
if (mcast->backoff > IPOIB_MAX_BACKOFF_SECONDS)
mcast->backoff = IPOIB_MAX_BACKOFF_SECONDS;
mutex_lock(&mcast_mutex);
if (test_bit(IPOIB_MCAST_RUN, &priv->flags))
queue_delayed_work(ipoib_workqueue,
&priv->mcast_task,
mcast->backoff * HZ);
mutex_unlock(&mcast_mutex);
}
}
void ipoib_mcast_join_task(struct work_struct *work)
{
struct ipoib_dev_priv *priv =
container_of(work, struct ipoib_dev_priv, mcast_task.work);
struct net_device *dev = priv->dev;
if (!test_bit(IPOIB_MCAST_RUN, &priv->flags))
return;
if (ib_query_gid(priv->ca, priv->port, 0, &priv->local_gid))
ipoib_warn(priv, "ib_query_gid() failed\n");
else
memcpy(priv->dev->dev_addr + 4, priv->local_gid.raw, sizeof (union ib_gid));
{
struct ib_port_attr attr;
if (!ib_query_port(priv->ca, priv->port, &attr))
priv->local_lid = attr.lid;
else
ipoib_warn(priv, "ib_query_port failed\n");
}
if (!priv->broadcast) {
struct ipoib_mcast *broadcast;
if (!test_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags))
return;
broadcast = ipoib_mcast_alloc(dev, 1);
if (!broadcast) {
ipoib_warn(priv, "failed to allocate broadcast group\n");
mutex_lock(&mcast_mutex);
if (test_bit(IPOIB_MCAST_RUN, &priv->flags))
queue_delayed_work(ipoib_workqueue,
&priv->mcast_task, HZ);
mutex_unlock(&mcast_mutex);
return;
}
spin_lock_irq(&priv->lock);
memcpy(broadcast->mcmember.mgid.raw, priv->dev->broadcast + 4,
sizeof (union ib_gid));
priv->broadcast = broadcast;
__ipoib_mcast_add(dev, priv->broadcast);
spin_unlock_irq(&priv->lock);
}
if (!test_bit(IPOIB_MCAST_FLAG_ATTACHED, &priv->broadcast->flags)) {
if (!test_bit(IPOIB_MCAST_FLAG_BUSY, &priv->broadcast->flags))
ipoib_mcast_join(dev, priv->broadcast, 0);
return;
}
while (1) {
struct ipoib_mcast *mcast = NULL;
spin_lock_irq(&priv->lock);
list_for_each_entry(mcast, &priv->multicast_list, list) {
if (!test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags)
&& !test_bit(IPOIB_MCAST_FLAG_BUSY, &mcast->flags)
&& !test_bit(IPOIB_MCAST_FLAG_ATTACHED, &mcast->flags)) {
/* Found the next unjoined group */
break;
}
}
spin_unlock_irq(&priv->lock);
if (&mcast->list == &priv->multicast_list) {
/* All done */
break;
}
ipoib_mcast_join(dev, mcast, 1);
return;
}
ipoib_dbg_mcast(priv, "successfully joined all multicast groups\n");
clear_bit(IPOIB_MCAST_RUN, &priv->flags);
}
int ipoib_mcast_start_thread(struct net_device *dev)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
ipoib_dbg_mcast(priv, "starting multicast thread\n");
mutex_lock(&mcast_mutex);
if (!test_and_set_bit(IPOIB_MCAST_RUN, &priv->flags))
queue_delayed_work(ipoib_workqueue, &priv->mcast_task, 0);
mutex_unlock(&mcast_mutex);
return 0;
}
int ipoib_mcast_stop_thread(struct net_device *dev, int flush)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
ipoib_dbg_mcast(priv, "stopping multicast thread\n");
mutex_lock(&mcast_mutex);
clear_bit(IPOIB_MCAST_RUN, &priv->flags);
cancel_delayed_work(&priv->mcast_task);
mutex_unlock(&mcast_mutex);
if (flush)
flush_workqueue(ipoib_workqueue);
return 0;
}
static int ipoib_mcast_leave(struct net_device *dev, struct ipoib_mcast *mcast)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
int ret = 0;
if (test_and_clear_bit(IPOIB_MCAST_FLAG_BUSY, &mcast->flags))
ib_sa_free_multicast(mcast->mc);
if (test_and_clear_bit(IPOIB_MCAST_FLAG_ATTACHED, &mcast->flags)) {
ipoib_dbg_mcast(priv, "leaving MGID %pI6\n",
mcast->mcmember.mgid.raw);
/* Remove ourselves from the multicast group */
ret = ib_detach_mcast(priv->qp, &mcast->mcmember.mgid,
be16_to_cpu(mcast->mcmember.mlid));
if (ret)
ipoib_warn(priv, "ib_detach_mcast failed (result = %d)\n", ret);
}
return 0;
}
void ipoib_mcast_send(struct net_device *dev, u8 *daddr, struct sk_buff *skb)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
struct ipoib_mcast *mcast;
unsigned long flags;
void *mgid = daddr + 4;
spin_lock_irqsave(&priv->lock, flags);
if (!test_bit(IPOIB_FLAG_OPER_UP, &priv->flags) ||
!priv->broadcast ||
!test_bit(IPOIB_MCAST_FLAG_ATTACHED, &priv->broadcast->flags)) {
++dev->stats.tx_dropped;
dev_kfree_skb_any(skb);
goto unlock;
}
mcast = __ipoib_mcast_find(dev, mgid);
if (!mcast) {
/* Let's create a new send only group now */
ipoib_dbg_mcast(priv, "setting up send only multicast group for %pI6\n",
mgid);
mcast = ipoib_mcast_alloc(dev, 0);
if (!mcast) {
ipoib_warn(priv, "unable to allocate memory for "
"multicast structure\n");
++dev->stats.tx_dropped;
dev_kfree_skb_any(skb);
goto out;
}
set_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags);
memcpy(mcast->mcmember.mgid.raw, mgid, sizeof (union ib_gid));
__ipoib_mcast_add(dev, mcast);
list_add_tail(&mcast->list, &priv->multicast_list);
}
if (!mcast->ah) {
if (skb_queue_len(&mcast->pkt_queue) < IPOIB_MAX_MCAST_QUEUE)
skb_queue_tail(&mcast->pkt_queue, skb);
else {
++dev->stats.tx_dropped;
dev_kfree_skb_any(skb);
}
if (test_bit(IPOIB_MCAST_FLAG_BUSY, &mcast->flags))
ipoib_dbg_mcast(priv, "no address vector, "
"but multicast join already started\n");
else if (test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags))
ipoib_mcast_sendonly_join(mcast);
/*
* If lookup completes between here and out:, don't
* want to send packet twice.
*/
mcast = NULL;
}
out:
if (mcast && mcast->ah) {
struct ipoib_neigh *neigh;
spin_unlock_irqrestore(&priv->lock, flags);
neigh = ipoib_neigh_get(dev, daddr);
spin_lock_irqsave(&priv->lock, flags);
if (!neigh) {
neigh = ipoib_neigh_alloc(daddr, dev);
if (neigh) {
kref_get(&mcast->ah->ref);
neigh->ah = mcast->ah;
list_add_tail(&neigh->list, &mcast->neigh_list);
}
}
spin_unlock_irqrestore(&priv->lock, flags);
ipoib_send(dev, skb, mcast->ah, IB_MULTICAST_QPN);
if (neigh)
ipoib_neigh_put(neigh);
return;
}
unlock:
spin_unlock_irqrestore(&priv->lock, flags);
}
void ipoib_mcast_dev_flush(struct net_device *dev)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
LIST_HEAD(remove_list);
struct ipoib_mcast *mcast, *tmcast;
unsigned long flags;
ipoib_dbg_mcast(priv, "flushing multicast list\n");
spin_lock_irqsave(&priv->lock, flags);
list_for_each_entry_safe(mcast, tmcast, &priv->multicast_list, list) {
list_del(&mcast->list);
rb_erase(&mcast->rb_node, &priv->multicast_tree);
list_add_tail(&mcast->list, &remove_list);
}
if (priv->broadcast) {
rb_erase(&priv->broadcast->rb_node, &priv->multicast_tree);
list_add_tail(&priv->broadcast->list, &remove_list);
priv->broadcast = NULL;
}
spin_unlock_irqrestore(&priv->lock, flags);
list_for_each_entry_safe(mcast, tmcast, &remove_list, list) {
ipoib_mcast_leave(dev, mcast);
ipoib_mcast_free(mcast);
}
}
static int ipoib_mcast_addr_is_valid(const u8 *addr, const u8 *broadcast)
{
/* reserved QPN, prefix, scope */
if (memcmp(addr, broadcast, 6))
return 0;
/* signature lower, pkey */
if (memcmp(addr + 7, broadcast + 7, 3))
return 0;
return 1;
}
void ipoib_mcast_restart_task(struct work_struct *work)
{
struct ipoib_dev_priv *priv =
container_of(work, struct ipoib_dev_priv, restart_task);
struct net_device *dev = priv->dev;
struct netdev_hw_addr *ha;
struct ipoib_mcast *mcast, *tmcast;
LIST_HEAD(remove_list);
unsigned long flags;
struct ib_sa_mcmember_rec rec;
ipoib_dbg_mcast(priv, "restarting multicast task\n");
ipoib_mcast_stop_thread(dev, 0);
local_irq_save(flags);
netif_addr_lock(dev);
spin_lock(&priv->lock);
/*
* Unfortunately, the networking core only gives us a list of all of
* the multicast hardware addresses. We need to figure out which ones
* are new and which ones have been removed
*/
/* Clear out the found flag */
list_for_each_entry(mcast, &priv->multicast_list, list)
clear_bit(IPOIB_MCAST_FLAG_FOUND, &mcast->flags);
/* Mark all of the entries that are found or don't exist */
netdev_for_each_mc_addr(ha, dev) {
union ib_gid mgid;
if (!ipoib_mcast_addr_is_valid(ha->addr, dev->broadcast))
continue;
memcpy(mgid.raw, ha->addr + 4, sizeof mgid);
mcast = __ipoib_mcast_find(dev, &mgid);
if (!mcast || test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags)) {
struct ipoib_mcast *nmcast;
/* ignore group which is directly joined by userspace */
if (test_bit(IPOIB_FLAG_UMCAST, &priv->flags) &&
!ib_sa_get_mcmember_rec(priv->ca, priv->port, &mgid, &rec)) {
ipoib_dbg_mcast(priv, "ignoring multicast entry for mgid %pI6\n",
mgid.raw);
continue;
}
/* Not found or send-only group, let's add a new entry */
ipoib_dbg_mcast(priv, "adding multicast entry for mgid %pI6\n",
mgid.raw);
nmcast = ipoib_mcast_alloc(dev, 0);
if (!nmcast) {
ipoib_warn(priv, "unable to allocate memory for multicast structure\n");
continue;
}
set_bit(IPOIB_MCAST_FLAG_FOUND, &nmcast->flags);
nmcast->mcmember.mgid = mgid;
if (mcast) {
/* Destroy the send only entry */
list_move_tail(&mcast->list, &remove_list);
rb_replace_node(&mcast->rb_node,
&nmcast->rb_node,
&priv->multicast_tree);
} else
__ipoib_mcast_add(dev, nmcast);
list_add_tail(&nmcast->list, &priv->multicast_list);
}
if (mcast)
set_bit(IPOIB_MCAST_FLAG_FOUND, &mcast->flags);
}
/* Remove all of the entries don't exist anymore */
list_for_each_entry_safe(mcast, tmcast, &priv->multicast_list, list) {
if (!test_bit(IPOIB_MCAST_FLAG_FOUND, &mcast->flags) &&
!test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags)) {
ipoib_dbg_mcast(priv, "deleting multicast group %pI6\n",
mcast->mcmember.mgid.raw);
rb_erase(&mcast->rb_node, &priv->multicast_tree);
/* Move to the remove list */
list_move_tail(&mcast->list, &remove_list);
}
}
spin_unlock(&priv->lock);
netif_addr_unlock(dev);
local_irq_restore(flags);
/* We have to cancel outside of the spinlock */
list_for_each_entry_safe(mcast, tmcast, &remove_list, list) {
ipoib_mcast_leave(mcast->dev, mcast);
ipoib_mcast_free(mcast);
}
if (test_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags))
ipoib_mcast_start_thread(dev);
}
#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG
struct ipoib_mcast_iter *ipoib_mcast_iter_init(struct net_device *dev)
{
struct ipoib_mcast_iter *iter;
iter = kmalloc(sizeof *iter, GFP_KERNEL);
if (!iter)
return NULL;
iter->dev = dev;
memset(iter->mgid.raw, 0, 16);
if (ipoib_mcast_iter_next(iter)) {
kfree(iter);
return NULL;
}
return iter;
}
int ipoib_mcast_iter_next(struct ipoib_mcast_iter *iter)
{
struct ipoib_dev_priv *priv = netdev_priv(iter->dev);
struct rb_node *n;
struct ipoib_mcast *mcast;
int ret = 1;
spin_lock_irq(&priv->lock);
n = rb_first(&priv->multicast_tree);
while (n) {
mcast = rb_entry(n, struct ipoib_mcast, rb_node);
if (memcmp(iter->mgid.raw, mcast->mcmember.mgid.raw,
sizeof (union ib_gid)) < 0) {
iter->mgid = mcast->mcmember.mgid;
iter->created = mcast->created;
iter->queuelen = skb_queue_len(&mcast->pkt_queue);
iter->complete = !!mcast->ah;
iter->send_only = !!(mcast->flags & (1 << IPOIB_MCAST_FLAG_SENDONLY));
ret = 0;
break;
}
n = rb_next(n);
}
spin_unlock_irq(&priv->lock);
return ret;
}
void ipoib_mcast_iter_read(struct ipoib_mcast_iter *iter,
union ib_gid *mgid,
unsigned long *created,
unsigned int *queuelen,
unsigned int *complete,
unsigned int *send_only)
{
*mgid = iter->mgid;
*created = iter->created;
*queuelen = iter->queuelen;
*complete = iter->complete;
*send_only = iter->send_only;
}
#endif /* CONFIG_INFINIBAND_IPOIB_DEBUG */

172
drivers/infiniband/ulp/ipoib/ipoib_netlink.c Normal file
View File

@@ -0,0 +1,172 @@
/*
* Copyright (c) 2012 Mellanox Technologies. - All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#include <linux/netdevice.h>
#include <linux/module.h>
#include <net/rtnetlink.h>
#include "ipoib.h"
static const struct nla_policy ipoib_policy[IFLA_IPOIB_MAX + 1] = {
[IFLA_IPOIB_PKEY] = { .type = NLA_U16 },
[IFLA_IPOIB_MODE] = { .type = NLA_U16 },
[IFLA_IPOIB_UMCAST] = { .type = NLA_U16 },
};
static int ipoib_fill_info(struct sk_buff *skb, const struct net_device *dev)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
u16 val;
if (nla_put_u16(skb, IFLA_IPOIB_PKEY, priv->pkey))
goto nla_put_failure;
val = test_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags);
if (nla_put_u16(skb, IFLA_IPOIB_MODE, val))
goto nla_put_failure;
val = test_bit(IPOIB_FLAG_UMCAST, &priv->flags);
if (nla_put_u16(skb, IFLA_IPOIB_UMCAST, val))
goto nla_put_failure;
return 0;
nla_put_failure:
return -EMSGSIZE;
}
static int ipoib_changelink(struct net_device *dev,
struct nlattr *tb[], struct nlattr *data[])
{
u16 mode, umcast;
int ret = 0;
if (data[IFLA_IPOIB_MODE]) {
mode = nla_get_u16(data[IFLA_IPOIB_MODE]);
if (mode == IPOIB_MODE_DATAGRAM)
ret = ipoib_set_mode(dev, "datagram\n");
else if (mode == IPOIB_MODE_CONNECTED)
ret = ipoib_set_mode(dev, "connected\n");
else
ret = -EINVAL;
if (ret < 0)
goto out_err;
}
if (data[IFLA_IPOIB_UMCAST]) {
umcast = nla_get_u16(data[IFLA_IPOIB_UMCAST]);
ipoib_set_umcast(dev, umcast);
}
out_err:
return ret;
}
static int ipoib_new_child_link(struct net *src_net, struct net_device *dev,
struct nlattr *tb[], struct nlattr *data[])
{
struct net_device *pdev;
struct ipoib_dev_priv *ppriv;
u16 child_pkey;
int err;
if (!tb[IFLA_LINK])
return -EINVAL;
pdev = __dev_get_by_index(src_net, nla_get_u32(tb[IFLA_LINK]));
if (!pdev)
return -ENODEV;
ppriv = netdev_priv(pdev);
if (test_bit(IPOIB_FLAG_SUBINTERFACE, &ppriv->flags)) {
ipoib_warn(ppriv, "child creation disallowed for child devices\n");
return -EINVAL;
}
if (!data || !data[IFLA_IPOIB_PKEY]) {
ipoib_dbg(ppriv, "no pkey specified, using parent pkey\n");
child_pkey = ppriv->pkey;
} else
child_pkey = nla_get_u16(data[IFLA_IPOIB_PKEY]);
err = __ipoib_vlan_add(ppriv, netdev_priv(dev), child_pkey, IPOIB_RTNL_CHILD);
if (!err && data)
err = ipoib_changelink(dev, tb, data);
return err;
}
static void ipoib_unregister_child_dev(struct net_device *dev, struct list_head *head)
{
struct ipoib_dev_priv *priv, *ppriv;
priv = netdev_priv(dev);
ppriv = netdev_priv(priv->parent);
mutex_lock(&ppriv->vlan_mutex);
unregister_netdevice_queue(dev, head);
list_del(&priv->list);
mutex_unlock(&ppriv->vlan_mutex);
}
static size_t ipoib_get_size(const struct net_device *dev)
{
return nla_total_size(2) + /* IFLA_IPOIB_PKEY */
nla_total_size(2) + /* IFLA_IPOIB_MODE */
nla_total_size(2); /* IFLA_IPOIB_UMCAST */
}
static struct rtnl_link_ops ipoib_link_ops __read_mostly = {
.kind = "ipoib",
.maxtype = IFLA_IPOIB_MAX,
.policy = ipoib_policy,
.priv_size = sizeof(struct ipoib_dev_priv),
.setup = ipoib_setup,
.newlink = ipoib_new_child_link,
.changelink = ipoib_changelink,
.dellink = ipoib_unregister_child_dev,
.get_size = ipoib_get_size,
.fill_info = ipoib_fill_info,
};
int __init ipoib_netlink_init(void)
{
return rtnl_link_register(&ipoib_link_ops);
}
void __exit ipoib_netlink_fini(void)
{
rtnl_link_unregister(&ipoib_link_ops);
}
MODULE_ALIAS_RTNL_LINK("ipoib");

294
drivers/infiniband/ulp/ipoib/ipoib_verbs.c Normal file
View File

@@ -0,0 +1,294 @@
/*
* Copyright (c) 2004, 2005 Topspin Communications. All rights reserved.
* Copyright (c) 2005 Mellanox Technologies. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#include <linux/slab.h>
#include "ipoib.h"
int ipoib_mcast_attach(struct net_device *dev, u16 mlid, union ib_gid *mgid, int set_qkey)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
struct ib_qp_attr *qp_attr = NULL;
int ret;
u16 pkey_index;
if (ib_find_pkey(priv->ca, priv->port, priv->pkey, &pkey_index)) {
clear_bit(IPOIB_PKEY_ASSIGNED, &priv->flags);
ret = -ENXIO;
goto out;
}
set_bit(IPOIB_PKEY_ASSIGNED, &priv->flags);
if (set_qkey) {
ret = -ENOMEM;
qp_attr = kmalloc(sizeof *qp_attr, GFP_KERNEL);
if (!qp_attr)
goto out;
/* set correct QKey for QP */
qp_attr->qkey = priv->qkey;
ret = ib_modify_qp(priv->qp, qp_attr, IB_QP_QKEY);
if (ret) {
ipoib_warn(priv, "failed to modify QP, ret = %d\n", ret);
goto out;
}
}
/* attach QP to multicast group */
ret = ib_attach_mcast(priv->qp, mgid, mlid);
if (ret)
ipoib_warn(priv, "failed to attach to multicast group, ret = %d\n", ret);
out:
kfree(qp_attr);
return ret;
}
int ipoib_init_qp(struct net_device *dev)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
int ret;
struct ib_qp_attr qp_attr;
int attr_mask;
if (!test_bit(IPOIB_PKEY_ASSIGNED, &priv->flags))
return -1;
qp_attr.qp_state = IB_QPS_INIT;
qp_attr.qkey = 0;
qp_attr.port_num = priv->port;
qp_attr.pkey_index = priv->pkey_index;
attr_mask =
IB_QP_QKEY |
IB_QP_PORT |
IB_QP_PKEY_INDEX |
IB_QP_STATE;
ret = ib_modify_qp(priv->qp, &qp_attr, attr_mask);
if (ret) {
ipoib_warn(priv, "failed to modify QP to init, ret = %d\n", ret);
goto out_fail;
}
qp_attr.qp_state = IB_QPS_RTR;
/* Can't set this in a INIT->RTR transition */
attr_mask &= ~IB_QP_PORT;
ret = ib_modify_qp(priv->qp, &qp_attr, attr_mask);
if (ret) {
ipoib_warn(priv, "failed to modify QP to RTR, ret = %d\n", ret);
goto out_fail;
}
qp_attr.qp_state = IB_QPS_RTS;
qp_attr.sq_psn = 0;
attr_mask |= IB_QP_SQ_PSN;
attr_mask &= ~IB_QP_PKEY_INDEX;
ret = ib_modify_qp(priv->qp, &qp_attr, attr_mask);
if (ret) {
ipoib_warn(priv, "failed to modify QP to RTS, ret = %d\n", ret);
goto out_fail;
}
return 0;
out_fail:
qp_attr.qp_state = IB_QPS_RESET;
if (ib_modify_qp(priv->qp, &qp_attr, IB_QP_STATE))
ipoib_warn(priv, "Failed to modify QP to RESET state\n");
return ret;
}
int ipoib_transport_dev_init(struct net_device *dev, struct ib_device *ca)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
struct ib_qp_init_attr init_attr = {
.cap = {
.max_send_wr = ipoib_sendq_size,
.max_recv_wr = ipoib_recvq_size,
.max_send_sge = 1,
.max_recv_sge = IPOIB_UD_RX_SG
},
.sq_sig_type = IB_SIGNAL_ALL_WR,
.qp_type = IB_QPT_UD
};
int ret, size;
int i;
priv->pd = ib_alloc_pd(priv->ca);
if (IS_ERR(priv->pd)) {
printk(KERN_WARNING "%s: failed to allocate PD\n", ca->name);
return -ENODEV;
}
priv->mr = ib_get_dma_mr(priv->pd, IB_ACCESS_LOCAL_WRITE);
if (IS_ERR(priv->mr)) {
printk(KERN_WARNING "%s: ib_get_dma_mr failed\n", ca->name);
goto out_free_pd;
}
size = ipoib_recvq_size + 1;
ret = ipoib_cm_dev_init(dev);
if (!ret) {
size += ipoib_sendq_size;
if (ipoib_cm_has_srq(dev))
size += ipoib_recvq_size + 1; /* 1 extra for rx_drain_qp */
else
size += ipoib_recvq_size * ipoib_max_conn_qp;
}
priv->recv_cq = ib_create_cq(priv->ca, ipoib_ib_completion, NULL, dev, size, 0);
if (IS_ERR(priv->recv_cq)) {
printk(KERN_WARNING "%s: failed to create receive CQ\n", ca->name);
goto out_free_mr;
}
priv->send_cq = ib_create_cq(priv->ca, ipoib_send_comp_handler, NULL,
dev, ipoib_sendq_size, 0);
if (IS_ERR(priv->send_cq)) {
printk(KERN_WARNING "%s: failed to create send CQ\n", ca->name);
goto out_free_recv_cq;
}
if (ib_req_notify_cq(priv->recv_cq, IB_CQ_NEXT_COMP))
goto out_free_send_cq;
init_attr.send_cq = priv->send_cq;
init_attr.recv_cq = priv->recv_cq;
if (priv->hca_caps & IB_DEVICE_UD_TSO)
init_attr.create_flags |= IB_QP_CREATE_IPOIB_UD_LSO;
if (priv->hca_caps & IB_DEVICE_BLOCK_MULTICAST_LOOPBACK)
init_attr.create_flags |= IB_QP_CREATE_BLOCK_MULTICAST_LOOPBACK;
if (dev->features & NETIF_F_SG)
init_attr.cap.max_send_sge = MAX_SKB_FRAGS + 1;
priv->qp = ib_create_qp(priv->pd, &init_attr);
if (IS_ERR(priv->qp)) {
printk(KERN_WARNING "%s: failed to create QP\n", ca->name);
goto out_free_send_cq;
}
priv->dev->dev_addr[1] = (priv->qp->qp_num >> 16) & 0xff;
priv->dev->dev_addr[2] = (priv->qp->qp_num >> 8) & 0xff;
priv->dev->dev_addr[3] = (priv->qp->qp_num ) & 0xff;
for (i = 0; i < MAX_SKB_FRAGS + 1; ++i)
priv->tx_sge[i].lkey = priv->mr->lkey;
priv->tx_wr.opcode = IB_WR_SEND;
priv->tx_wr.sg_list = priv->tx_sge;
priv->tx_wr.send_flags = IB_SEND_SIGNALED;
priv->rx_sge[0].lkey = priv->mr->lkey;
if (ipoib_ud_need_sg(priv->max_ib_mtu)) {
priv->rx_sge[0].length = IPOIB_UD_HEAD_SIZE;
priv->rx_sge[1].length = PAGE_SIZE;
priv->rx_sge[1].lkey = priv->mr->lkey;
priv->rx_wr.num_sge = IPOIB_UD_RX_SG;
} else {
priv->rx_sge[0].length = IPOIB_UD_BUF_SIZE(priv->max_ib_mtu);
priv->rx_wr.num_sge = 1;
}
priv->rx_wr.next = NULL;
priv->rx_wr.sg_list = priv->rx_sge;
return 0;
out_free_send_cq:
ib_destroy_cq(priv->send_cq);
out_free_recv_cq:
ib_destroy_cq(priv->recv_cq);
out_free_mr:
ib_dereg_mr(priv->mr);
ipoib_cm_dev_cleanup(dev);
out_free_pd:
ib_dealloc_pd(priv->pd);
return -ENODEV;
}
void ipoib_transport_dev_cleanup(struct net_device *dev)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
if (priv->qp) {
if (ib_destroy_qp(priv->qp))
ipoib_warn(priv, "ib_qp_destroy failed\n");
priv->qp = NULL;
clear_bit(IPOIB_PKEY_ASSIGNED, &priv->flags);
}
if (ib_destroy_cq(priv->send_cq))
ipoib_warn(priv, "ib_cq_destroy (send) failed\n");
if (ib_destroy_cq(priv->recv_cq))
ipoib_warn(priv, "ib_cq_destroy (recv) failed\n");
ipoib_cm_dev_cleanup(dev);
if (ib_dereg_mr(priv->mr))
ipoib_warn(priv, "ib_dereg_mr failed\n");
if (ib_dealloc_pd(priv->pd))
ipoib_warn(priv, "ib_dealloc_pd failed\n");
}
void ipoib_event(struct ib_event_handler *handler,
struct ib_event *record)
{
struct ipoib_dev_priv *priv =
container_of(handler, struct ipoib_dev_priv, event_handler);
if (record->element.port_num != priv->port)
return;
ipoib_dbg(priv, "Event %d on device %s port %d\n", record->event,
record->device->name, record->element.port_num);
if (record->event == IB_EVENT_SM_CHANGE ||
record->event == IB_EVENT_CLIENT_REREGISTER) {
queue_work(ipoib_workqueue, &priv->flush_light);
} else if (record->event == IB_EVENT_PORT_ERR ||
record->event == IB_EVENT_PORT_ACTIVE ||
record->event == IB_EVENT_LID_CHANGE) {
queue_work(ipoib_workqueue, &priv->flush_normal);
} else if (record->event == IB_EVENT_PKEY_CHANGE) {
queue_work(ipoib_workqueue, &priv->flush_heavy);
}
}

207
drivers/infiniband/ulp/ipoib/ipoib_vlan.c Normal file
View File

@@ -0,0 +1,207 @@
/*
* Copyright (c) 2004 Topspin Communications. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#include <linux/module.h>
#include <linux/init.h>
#include <linux/seq_file.h>
#include <asm/uaccess.h>
#include "ipoib.h"
static ssize_t show_parent(struct device *d, struct device_attribute *attr,
char *buf)
{
struct net_device *dev = to_net_dev(d);
struct ipoib_dev_priv *priv = netdev_priv(dev);
return sprintf(buf, "%s\n", priv->parent->name);
}
static DEVICE_ATTR(parent, S_IRUGO, show_parent, NULL);
int __ipoib_vlan_add(struct ipoib_dev_priv *ppriv, struct ipoib_dev_priv *priv,
u16 pkey, int type)
{
int result;
priv->max_ib_mtu = ppriv->max_ib_mtu;
/* MTU will be reset when mcast join happens */
priv->dev->mtu = IPOIB_UD_MTU(priv->max_ib_mtu);
priv->mcast_mtu = priv->admin_mtu = priv->dev->mtu;
set_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags);
result = ipoib_set_dev_features(priv, ppriv->ca);
if (result)
goto err;
priv->pkey = pkey;
memcpy(priv->dev->dev_addr, ppriv->dev->dev_addr, INFINIBAND_ALEN);
priv->dev->broadcast[8] = pkey >> 8;
priv->dev->broadcast[9] = pkey & 0xff;
result = ipoib_dev_init(priv->dev, ppriv->ca, ppriv->port);
if (result < 0) {
ipoib_warn(ppriv, "failed to initialize subinterface: "
"device %s, port %d",
ppriv->ca->name, ppriv->port);
goto err;
}
result = register_netdevice(priv->dev);
if (result) {
ipoib_warn(priv, "failed to initialize; error %i", result);
goto register_failed;
}
priv->parent = ppriv->dev;
ipoib_create_debug_files(priv->dev);
/* RTNL childs don't need proprietary sysfs entries */
if (type == IPOIB_LEGACY_CHILD) {
if (ipoib_cm_add_mode_attr(priv->dev))
goto sysfs_failed;
if (ipoib_add_pkey_attr(priv->dev))
goto sysfs_failed;
if (ipoib_add_umcast_attr(priv->dev))
goto sysfs_failed;
if (device_create_file(&priv->dev->dev, &dev_attr_parent))
goto sysfs_failed;
}
priv->child_type = type;
priv->dev->iflink = ppriv->dev->ifindex;
list_add_tail(&priv->list, &ppriv->child_intfs);
return 0;
sysfs_failed:
result = -ENOMEM;
ipoib_delete_debug_files(priv->dev);
unregister_netdevice(priv->dev);
register_failed:
ipoib_dev_cleanup(priv->dev);
err:
return result;
}
int ipoib_vlan_add(struct net_device *pdev, unsigned short pkey)
{
struct ipoib_dev_priv *ppriv, *priv;
char intf_name[IFNAMSIZ];
struct ipoib_dev_priv *tpriv;
int result;
if (!capable(CAP_NET_ADMIN))
return -EPERM;
ppriv = netdev_priv(pdev);
snprintf(intf_name, sizeof intf_name, "%s.%04x",
ppriv->dev->name, pkey);
priv = ipoib_intf_alloc(intf_name);
if (!priv)
return -ENOMEM;
if (!rtnl_trylock())
return restart_syscall();
mutex_lock(&ppriv->vlan_mutex);
/*
* First ensure this isn't a duplicate. We check the parent device and
* then all of the legacy child interfaces to make sure the Pkey
* doesn't match.
*/
if (ppriv->pkey == pkey) {
result = -ENOTUNIQ;
goto out;
}
list_for_each_entry(tpriv, &ppriv->child_intfs, list) {
if (tpriv->pkey == pkey &&
tpriv->child_type == IPOIB_LEGACY_CHILD) {
result = -ENOTUNIQ;
goto out;
}
}
result = __ipoib_vlan_add(ppriv, priv, pkey, IPOIB_LEGACY_CHILD);
out:
mutex_unlock(&ppriv->vlan_mutex);
if (result)
free_netdev(priv->dev);
rtnl_unlock();
return result;
}
int ipoib_vlan_delete(struct net_device *pdev, unsigned short pkey)
{
struct ipoib_dev_priv *ppriv, *priv, *tpriv;
struct net_device *dev = NULL;
if (!capable(CAP_NET_ADMIN))
return -EPERM;
ppriv = netdev_priv(pdev);
if (!rtnl_trylock())
return restart_syscall();
mutex_lock(&ppriv->vlan_mutex);
list_for_each_entry_safe(priv, tpriv, &ppriv->child_intfs, list) {
if (priv->pkey == pkey &&
priv->child_type == IPOIB_LEGACY_CHILD) {
unregister_netdevice(priv->dev);
list_del(&priv->list);
dev = priv->dev;
break;
}
}
mutex_unlock(&ppriv->vlan_mutex);
rtnl_unlock();
if (dev) {
free_netdev(dev);
return 0;
}
return -ENODEV;
}

12
drivers/infiniband/ulp/iser/Kconfig Normal file
View File

@@ -0,0 +1,12 @@
config INFINIBAND_ISER
tristate "iSCSI Extensions for RDMA (iSER)"
depends on SCSI && INET && INFINIBAND_ADDR_TRANS
select SCSI_ISCSI_ATTRS
---help---
Support for the iSCSI Extensions for RDMA (iSER) Protocol
over InfiniBand. This allows you to access storage devices
that speak iSCSI over iSER over InfiniBand.
The iSER protocol is defined by IETF.
See <http://www.ietf.org/rfc/rfc5046.txt>
and <http://members.infinibandta.org/kwspub/spec/Annex_iSER.PDF>

4
drivers/infiniband/ulp/iser/Makefile Normal file
View File

@@ -0,0 +1,4 @@
obj-$(CONFIG_INFINIBAND_ISER) += ib_iser.o
ib_iser-y := iser_verbs.o iser_initiator.o iser_memory.o \
iscsi_iser.o

787
drivers/infiniband/ulp/iser/iscsi_iser.c Normal file
View File

@@ -0,0 +1,787 @@
/*
* iSCSI Initiator over iSER Data-Path
*
* Copyright (C) 2004 Dmitry Yusupov
* Copyright (C) 2004 Alex Aizman
* Copyright (C) 2005 Mike Christie
* Copyright (c) 2005, 2006 Voltaire, Inc. All rights reserved.
* Copyright (c) 2013 Mellanox Technologies. All rights reserved.
* maintained by openib-general@openib.org
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*
* Credits:
* Christoph Hellwig
* FUJITA Tomonori
* Arne Redlich
* Zhenyu Wang
* Modified by:
* Erez Zilber
*/
#include <linux/types.h>
#include <linux/list.h>
#include <linux/hardirq.h>
#include <linux/kfifo.h>
#include <linux/blkdev.h>
#include <linux/init.h>
#include <linux/ioctl.h>
#include <linux/cdev.h>
#include <linux/in.h>
#include <linux/net.h>
#include <linux/scatterlist.h>
#include <linux/delay.h>
#include <linux/slab.h>
#include <linux/module.h>
#include <net/sock.h>
#include <asm/uaccess.h>
#include <scsi/scsi_cmnd.h>
#include <scsi/scsi_device.h>
#include <scsi/scsi_eh.h>
#include <scsi/scsi_tcq.h>
#include <scsi/scsi_host.h>
#include <scsi/scsi.h>
#include <scsi/scsi_transport_iscsi.h>
#include "iscsi_iser.h"
static struct scsi_host_template iscsi_iser_sht;
static struct iscsi_transport iscsi_iser_transport;
static struct scsi_transport_template *iscsi_iser_scsi_transport;
static unsigned int iscsi_max_lun = 512;
module_param_named(max_lun, iscsi_max_lun, uint, S_IRUGO);
int iser_debug_level = 0;
MODULE_DESCRIPTION("iSER (iSCSI Extensions for RDMA) Datamover");
MODULE_LICENSE("Dual BSD/GPL");
MODULE_AUTHOR("Alex Nezhinsky, Dan Bar Dov, Or Gerlitz");
MODULE_VERSION(DRV_VER);
module_param_named(debug_level, iser_debug_level, int, 0644);
MODULE_PARM_DESC(debug_level, "Enable debug tracing if > 0 (default:disabled)");
struct iser_global ig;
void
iscsi_iser_recv(struct iscsi_conn *conn,
struct iscsi_hdr *hdr, char *rx_data, int rx_data_len)
{
int rc = 0;
int datalen;
int ahslen;
/* verify PDU length */
datalen = ntoh24(hdr->dlength);
if (datalen > rx_data_len || (datalen + 4) < rx_data_len) {
iser_err("wrong datalen %d (hdr), %d (IB)\n",
datalen, rx_data_len);
rc = ISCSI_ERR_DATALEN;
goto error;
}
if (datalen != rx_data_len)
iser_dbg("aligned datalen (%d) hdr, %d (IB)\n",
datalen, rx_data_len);
/* read AHS */
ahslen = hdr->hlength * 4;
rc = iscsi_complete_pdu(conn, hdr, rx_data, rx_data_len);
if (rc && rc != ISCSI_ERR_NO_SCSI_CMD)
goto error;
return;
error:
iscsi_conn_failure(conn, rc);
}
static int iscsi_iser_pdu_alloc(struct iscsi_task *task, uint8_t opcode)
{
struct iscsi_iser_task *iser_task = task->dd_data;
task->hdr = (struct iscsi_hdr *)&iser_task->desc.iscsi_header;
task->hdr_max = sizeof(iser_task->desc.iscsi_header);
return 0;
}
int iser_initialize_task_headers(struct iscsi_task *task,
struct iser_tx_desc *tx_desc)
{
struct iscsi_iser_conn *iser_conn = task->conn->dd_data;
struct iser_device *device = iser_conn->ib_conn->device;
struct iscsi_iser_task *iser_task = task->dd_data;
u64 dma_addr;
dma_addr = ib_dma_map_single(device->ib_device, (void *)tx_desc,
ISER_HEADERS_LEN, DMA_TO_DEVICE);
if (ib_dma_mapping_error(device->ib_device, dma_addr))
return -ENOMEM;
tx_desc->dma_addr = dma_addr;
tx_desc->tx_sg[0].addr = tx_desc->dma_addr;
tx_desc->tx_sg[0].length = ISER_HEADERS_LEN;
tx_desc->tx_sg[0].lkey = device->mr->lkey;
iser_task->iser_conn = iser_conn;
return 0;
}
/**
* iscsi_iser_task_init - Initialize task
* @task: iscsi task
*
* Initialize the task for the scsi command or mgmt command.
*/
static int
iscsi_iser_task_init(struct iscsi_task *task)
{
struct iscsi_iser_task *iser_task = task->dd_data;
if (iser_initialize_task_headers(task, &iser_task->desc))
return -ENOMEM;
/* mgmt task */
if (!task->sc)
return 0;
iser_task->command_sent = 0;
iser_task_rdma_init(iser_task);
return 0;
}
/**
* iscsi_iser_mtask_xmit - xmit management(immediate) task
* @conn: iscsi connection
* @task: task management task
*
* Notes:
* The function can return -EAGAIN in which case caller must
* call it again later, or recover. '0' return code means successful
* xmit.
*
**/
static int
iscsi_iser_mtask_xmit(struct iscsi_conn *conn, struct iscsi_task *task)
{
int error = 0;
iser_dbg("mtask xmit [cid %d itt 0x%x]\n", conn->id, task->itt);
error = iser_send_control(conn, task);
/* since iser xmits control with zero copy, tasks can not be recycled
* right after sending them.
* The recycling scheme is based on whether a response is expected
* - if yes, the task is recycled at iscsi_complete_pdu
* - if no, the task is recycled at iser_snd_completion
*/
return error;
}
static int
iscsi_iser_task_xmit_unsol_data(struct iscsi_conn *conn,
struct iscsi_task *task)
{
struct iscsi_r2t_info *r2t = &task->unsol_r2t;
struct iscsi_data hdr;
int error = 0;
/* Send data-out PDUs while there's still unsolicited data to send */
while (iscsi_task_has_unsol_data(task)) {
iscsi_prep_data_out_pdu(task, r2t, &hdr);
iser_dbg("Sending data-out: itt 0x%x, data count %d\n",
hdr.itt, r2t->data_count);
/* the buffer description has been passed with the command */
/* Send the command */
error = iser_send_data_out(conn, task, &hdr);
if (error) {
r2t->datasn--;
goto iscsi_iser_task_xmit_unsol_data_exit;
}
r2t->sent += r2t->data_count;
iser_dbg("Need to send %d more as data-out PDUs\n",
r2t->data_length - r2t->sent);
}
iscsi_iser_task_xmit_unsol_data_exit:
return error;
}
static int
iscsi_iser_task_xmit(struct iscsi_task *task)
{
struct iscsi_conn *conn = task->conn;
struct iscsi_iser_task *iser_task = task->dd_data;
int error = 0;
if (!task->sc)
return iscsi_iser_mtask_xmit(conn, task);
if (task->sc->sc_data_direction == DMA_TO_DEVICE) {
BUG_ON(scsi_bufflen(task->sc) == 0);
iser_dbg("cmd [itt %x total %d imm %d unsol_data %d\n",
task->itt, scsi_bufflen(task->sc),
task->imm_count, task->unsol_r2t.data_length);
}
iser_dbg("ctask xmit [cid %d itt 0x%x]\n",
conn->id, task->itt);
/* Send the cmd PDU */
if (!iser_task->command_sent) {
error = iser_send_command(conn, task);
if (error)
goto iscsi_iser_task_xmit_exit;
iser_task->command_sent = 1;
}
/* Send unsolicited data-out PDU(s) if necessary */
if (iscsi_task_has_unsol_data(task))
error = iscsi_iser_task_xmit_unsol_data(conn, task);
iscsi_iser_task_xmit_exit:
return error;
}
static void iscsi_iser_cleanup_task(struct iscsi_task *task)
{
struct iscsi_iser_task *iser_task = task->dd_data;
struct iser_tx_desc *tx_desc = &iser_task->desc;
struct iscsi_iser_conn *iser_conn = task->conn->dd_data;
struct iser_device *device = iser_conn->ib_conn->device;
ib_dma_unmap_single(device->ib_device,
tx_desc->dma_addr, ISER_HEADERS_LEN, DMA_TO_DEVICE);
/* mgmt tasks do not need special cleanup */
if (!task->sc)
return;
if (iser_task->status == ISER_TASK_STATUS_STARTED) {
iser_task->status = ISER_TASK_STATUS_COMPLETED;
iser_task_rdma_finalize(iser_task);
}
}
static struct iscsi_cls_conn *
iscsi_iser_conn_create(struct iscsi_cls_session *cls_session, uint32_t conn_idx)
{
struct iscsi_conn *conn;
struct iscsi_cls_conn *cls_conn;
struct iscsi_iser_conn *iser_conn;
cls_conn = iscsi_conn_setup(cls_session, sizeof(*iser_conn), conn_idx);
if (!cls_conn)
return NULL;
conn = cls_conn->dd_data;
/*
* due to issues with the login code re iser sematics
* this not set in iscsi_conn_setup - FIXME
*/
conn->max_recv_dlength = ISER_RECV_DATA_SEG_LEN;
iser_conn = conn->dd_data;
conn->dd_data = iser_conn;
iser_conn->iscsi_conn = conn;
return cls_conn;
}
static void
iscsi_iser_conn_destroy(struct iscsi_cls_conn *cls_conn)
{
struct iscsi_conn *conn = cls_conn->dd_data;
struct iscsi_iser_conn *iser_conn = conn->dd_data;
struct iser_conn *ib_conn = iser_conn->ib_conn;
iscsi_conn_teardown(cls_conn);
/*
* Userspace will normally call the stop callback and
* already have freed the ib_conn, but if it goofed up then
* we free it here.
*/
if (ib_conn) {
ib_conn->iser_conn = NULL;
iser_conn_put(ib_conn, 1); /* deref iscsi/ib conn unbinding */
}
}
static int
iscsi_iser_conn_bind(struct iscsi_cls_session *cls_session,
struct iscsi_cls_conn *cls_conn, uint64_t transport_eph,
int is_leading)
{
struct iscsi_conn *conn = cls_conn->dd_data;
struct iscsi_iser_conn *iser_conn;
struct iser_conn *ib_conn;
struct iscsi_endpoint *ep;
int error;
error = iscsi_conn_bind(cls_session, cls_conn, is_leading);
if (error)
return error;
/* the transport ep handle comes from user space so it must be
* verified against the global ib connections list */
ep = iscsi_lookup_endpoint(transport_eph);
if (!ep) {
iser_err("can't bind eph %llx\n",
(unsigned long long)transport_eph);
return -EINVAL;
}
ib_conn = ep->dd_data;
if (iser_alloc_rx_descriptors(ib_conn))
return -ENOMEM;
/* binds the iSER connection retrieved from the previously
* connected ep_handle to the iSCSI layer connection. exchanges
* connection pointers */
iser_info("binding iscsi/iser conn %p %p to ib_conn %p\n",
conn, conn->dd_data, ib_conn);
iser_conn = conn->dd_data;
ib_conn->iser_conn = iser_conn;
iser_conn->ib_conn = ib_conn;
iser_conn_get(ib_conn); /* ref iscsi/ib conn binding */
return 0;
}
static void
iscsi_iser_conn_stop(struct iscsi_cls_conn *cls_conn, int flag)
{
struct iscsi_conn *conn = cls_conn->dd_data;
struct iscsi_iser_conn *iser_conn = conn->dd_data;
struct iser_conn *ib_conn = iser_conn->ib_conn;
/*
* Userspace may have goofed up and not bound the connection or
* might have only partially setup the connection.
*/
if (ib_conn) {
iscsi_conn_stop(cls_conn, flag);
/*
* There is no unbind event so the stop callback
* must release the ref from the bind.
*/
iser_conn_put(ib_conn, 1); /* deref iscsi/ib conn unbinding */
}
iser_conn->ib_conn = NULL;
}
static void iscsi_iser_session_destroy(struct iscsi_cls_session *cls_session)
{
struct Scsi_Host *shost = iscsi_session_to_shost(cls_session);
iscsi_session_teardown(cls_session);
iscsi_host_remove(shost);
iscsi_host_free(shost);
}
static struct iscsi_cls_session *
iscsi_iser_session_create(struct iscsi_endpoint *ep,
uint16_t cmds_max, uint16_t qdepth,
uint32_t initial_cmdsn)
{
struct iscsi_cls_session *cls_session;
struct iscsi_session *session;
struct Scsi_Host *shost;
struct iser_conn *ib_conn;
shost = iscsi_host_alloc(&iscsi_iser_sht, 0, 0);
if (!shost)
return NULL;
shost->transportt = iscsi_iser_scsi_transport;
shost->max_lun = iscsi_max_lun;
shost->max_id = 0;
shost->max_channel = 0;
shost->max_cmd_len = 16;
/*
* older userspace tools (before 2.0-870) did not pass us
* the leading conn's ep so this will be NULL;
*/
if (ep)
ib_conn = ep->dd_data;
if (iscsi_host_add(shost,
ep ? ib_conn->device->ib_device->dma_device : NULL))
goto free_host;
/*
* we do not support setting can_queue cmd_per_lun from userspace yet
* because we preallocate so many resources
*/
cls_session = iscsi_session_setup(&iscsi_iser_transport, shost,
ISCSI_DEF_XMIT_CMDS_MAX, 0,
sizeof(struct iscsi_iser_task),
initial_cmdsn, 0);
if (!cls_session)
goto remove_host;
session = cls_session->dd_data;
shost->can_queue = session->scsi_cmds_max;
return cls_session;
remove_host:
iscsi_host_remove(shost);
free_host:
iscsi_host_free(shost);
return NULL;
}
static int
iscsi_iser_set_param(struct iscsi_cls_conn *cls_conn,
enum iscsi_param param, char *buf, int buflen)
{
int value;
switch (param) {
case ISCSI_PARAM_MAX_RECV_DLENGTH:
/* TBD */
break;
case ISCSI_PARAM_HDRDGST_EN:
sscanf(buf, "%d", &value);
if (value) {
iser_err("DataDigest wasn't negotiated to None");
return -EPROTO;
}
break;
case ISCSI_PARAM_DATADGST_EN:
sscanf(buf, "%d", &value);
if (value) {
iser_err("DataDigest wasn't negotiated to None");
return -EPROTO;
}
break;
case ISCSI_PARAM_IFMARKER_EN:
sscanf(buf, "%d", &value);
if (value) {
iser_err("IFMarker wasn't negotiated to No");
return -EPROTO;
}
break;
case ISCSI_PARAM_OFMARKER_EN:
sscanf(buf, "%d", &value);
if (value) {
iser_err("OFMarker wasn't negotiated to No");
return -EPROTO;
}
break;
default:
return iscsi_set_param(cls_conn, param, buf, buflen);
}
return 0;
}
static void
iscsi_iser_conn_get_stats(struct iscsi_cls_conn *cls_conn, struct iscsi_stats *stats)
{
struct iscsi_conn *conn = cls_conn->dd_data;
stats->txdata_octets = conn->txdata_octets;
stats->rxdata_octets = conn->rxdata_octets;
stats->scsicmd_pdus = conn->scsicmd_pdus_cnt;
stats->dataout_pdus = conn->dataout_pdus_cnt;
stats->scsirsp_pdus = conn->scsirsp_pdus_cnt;
stats->datain_pdus = conn->datain_pdus_cnt; /* always 0 */
stats->r2t_pdus = conn->r2t_pdus_cnt; /* always 0 */
stats->tmfcmd_pdus = conn->tmfcmd_pdus_cnt;
stats->tmfrsp_pdus = conn->tmfrsp_pdus_cnt;
stats->custom_length = 4;
strcpy(stats->custom[0].desc, "qp_tx_queue_full");
stats->custom[0].value = 0; /* TB iser_conn->qp_tx_queue_full; */
strcpy(stats->custom[1].desc, "fmr_map_not_avail");
stats->custom[1].value = 0; /* TB iser_conn->fmr_map_not_avail */;
strcpy(stats->custom[2].desc, "eh_abort_cnt");
stats->custom[2].value = conn->eh_abort_cnt;
strcpy(stats->custom[3].desc, "fmr_unalign_cnt");
stats->custom[3].value = conn->fmr_unalign_cnt;
}
static int iscsi_iser_get_ep_param(struct iscsi_endpoint *ep,
enum iscsi_param param, char *buf)
{
struct iser_conn *ib_conn = ep->dd_data;
int len;
switch (param) {
case ISCSI_PARAM_CONN_PORT:
case ISCSI_PARAM_CONN_ADDRESS:
if (!ib_conn || !ib_conn->cma_id)
return -ENOTCONN;
return iscsi_conn_get_addr_param((struct sockaddr_storage *)
&ib_conn->cma_id->route.addr.dst_addr,
param, buf);
break;
default:
return -ENOSYS;
}
return len;
}
static struct iscsi_endpoint *
iscsi_iser_ep_connect(struct Scsi_Host *shost, struct sockaddr *dst_addr,
int non_blocking)
{
int err;
struct iser_conn *ib_conn;
struct iscsi_endpoint *ep;
ep = iscsi_create_endpoint(sizeof(*ib_conn));
if (!ep)
return ERR_PTR(-ENOMEM);
ib_conn = ep->dd_data;
ib_conn->ep = ep;
iser_conn_init(ib_conn);
err = iser_connect(ib_conn, NULL, (struct sockaddr_in *)dst_addr,
non_blocking);
if (err)
return ERR_PTR(err);
return ep;
}
static int
iscsi_iser_ep_poll(struct iscsi_endpoint *ep, int timeout_ms)
{
struct iser_conn *ib_conn;
int rc;
ib_conn = ep->dd_data;
rc = wait_event_interruptible_timeout(ib_conn->wait,
ib_conn->state == ISER_CONN_UP,
msecs_to_jiffies(timeout_ms));
/* if conn establishment failed, return error code to iscsi */
if (!rc &&
(ib_conn->state == ISER_CONN_TERMINATING ||
ib_conn->state == ISER_CONN_DOWN))
rc = -1;
iser_info("ib conn %p rc = %d\n", ib_conn, rc);
if (rc > 0)
return 1; /* success, this is the equivalent of POLLOUT */
else if (!rc)
return 0; /* timeout */
else
return rc; /* signal */
}
static void
iscsi_iser_ep_disconnect(struct iscsi_endpoint *ep)
{
struct iser_conn *ib_conn;
ib_conn = ep->dd_data;
if (ib_conn->iser_conn)
/*
* Must suspend xmit path if the ep is bound to the
* iscsi_conn, so we know we are not accessing the ib_conn
* when we free it.
*
* This may not be bound if the ep poll failed.
*/
iscsi_suspend_tx(ib_conn->iser_conn->iscsi_conn);
iser_info("ib conn %p state %d\n", ib_conn, ib_conn->state);
iser_conn_terminate(ib_conn);
}
static umode_t iser_attr_is_visible(int param_type, int param)
{
switch (param_type) {
case ISCSI_HOST_PARAM:
switch (param) {
case ISCSI_HOST_PARAM_NETDEV_NAME:
case ISCSI_HOST_PARAM_HWADDRESS:
case ISCSI_HOST_PARAM_INITIATOR_NAME:
return S_IRUGO;
default:
return 0;
}
case ISCSI_PARAM:
switch (param) {
case ISCSI_PARAM_MAX_RECV_DLENGTH:
case ISCSI_PARAM_MAX_XMIT_DLENGTH:
case ISCSI_PARAM_HDRDGST_EN:
case ISCSI_PARAM_DATADGST_EN:
case ISCSI_PARAM_CONN_ADDRESS:
case ISCSI_PARAM_CONN_PORT:
case ISCSI_PARAM_EXP_STATSN:
case ISCSI_PARAM_PERSISTENT_ADDRESS:
case ISCSI_PARAM_PERSISTENT_PORT:
case ISCSI_PARAM_PING_TMO:
case ISCSI_PARAM_RECV_TMO:
case ISCSI_PARAM_INITIAL_R2T_EN:
case ISCSI_PARAM_MAX_R2T:
case ISCSI_PARAM_IMM_DATA_EN:
case ISCSI_PARAM_FIRST_BURST:
case ISCSI_PARAM_MAX_BURST:
case ISCSI_PARAM_PDU_INORDER_EN:
case ISCSI_PARAM_DATASEQ_INORDER_EN:
case ISCSI_PARAM_TARGET_NAME:
case ISCSI_PARAM_TPGT:
case ISCSI_PARAM_USERNAME:
case ISCSI_PARAM_PASSWORD:
case ISCSI_PARAM_USERNAME_IN:
case ISCSI_PARAM_PASSWORD_IN:
case ISCSI_PARAM_FAST_ABORT:
case ISCSI_PARAM_ABORT_TMO:
case ISCSI_PARAM_LU_RESET_TMO:
case ISCSI_PARAM_TGT_RESET_TMO:
case ISCSI_PARAM_IFACE_NAME:
case ISCSI_PARAM_INITIATOR_NAME:
return S_IRUGO;
default:
return 0;
}
}
return 0;
}
static struct scsi_host_template iscsi_iser_sht = {
.module = THIS_MODULE,
.name = "iSCSI Initiator over iSER",
.queuecommand = iscsi_queuecommand,
.change_queue_depth = iscsi_change_queue_depth,
.sg_tablesize = ISCSI_ISER_SG_TABLESIZE,
.max_sectors = 1024,
.cmd_per_lun = ISER_DEF_CMD_PER_LUN,
.eh_abort_handler = iscsi_eh_abort,
.eh_device_reset_handler= iscsi_eh_device_reset,
.eh_target_reset_handler = iscsi_eh_recover_target,
.target_alloc = iscsi_target_alloc,
.use_clustering = DISABLE_CLUSTERING,
.proc_name = "iscsi_iser",
.this_id = -1,
};
static struct iscsi_transport iscsi_iser_transport = {
.owner = THIS_MODULE,
.name = "iser",
.caps = CAP_RECOVERY_L0 | CAP_MULTI_R2T,
/* session management */
.create_session = iscsi_iser_session_create,
.destroy_session = iscsi_iser_session_destroy,
/* connection management */
.create_conn = iscsi_iser_conn_create,
.bind_conn = iscsi_iser_conn_bind,
.destroy_conn = iscsi_iser_conn_destroy,
.attr_is_visible = iser_attr_is_visible,
.set_param = iscsi_iser_set_param,
.get_conn_param = iscsi_conn_get_param,
.get_ep_param = iscsi_iser_get_ep_param,
.get_session_param = iscsi_session_get_param,
.start_conn = iscsi_conn_start,
.stop_conn = iscsi_iser_conn_stop,
/* iscsi host params */
.get_host_param = iscsi_host_get_param,
.set_host_param = iscsi_host_set_param,
/* IO */
.send_pdu = iscsi_conn_send_pdu,
.get_stats = iscsi_iser_conn_get_stats,
.init_task = iscsi_iser_task_init,
.xmit_task = iscsi_iser_task_xmit,
.cleanup_task = iscsi_iser_cleanup_task,
.alloc_pdu = iscsi_iser_pdu_alloc,
/* recovery */
.session_recovery_timedout = iscsi_session_recovery_timedout,
.ep_connect = iscsi_iser_ep_connect,
.ep_poll = iscsi_iser_ep_poll,
.ep_disconnect = iscsi_iser_ep_disconnect
};
static int __init iser_init(void)
{
int err;
iser_dbg("Starting iSER datamover...\n");
if (iscsi_max_lun < 1) {
iser_err("Invalid max_lun value of %u\n", iscsi_max_lun);
return -EINVAL;
}
memset(&ig, 0, sizeof(struct iser_global));
ig.desc_cache = kmem_cache_create("iser_descriptors",
sizeof(struct iser_tx_desc),
0, SLAB_HWCACHE_ALIGN,
NULL);
if (ig.desc_cache == NULL)
return -ENOMEM;
/* device init is called only after the first addr resolution */
mutex_init(&ig.device_list_mutex);
INIT_LIST_HEAD(&ig.device_list);
mutex_init(&ig.connlist_mutex);
INIT_LIST_HEAD(&ig.connlist);
iscsi_iser_scsi_transport = iscsi_register_transport(
&iscsi_iser_transport);
if (!iscsi_iser_scsi_transport) {
iser_err("iscsi_register_transport failed\n");
err = -EINVAL;
goto register_transport_failure;
}
return 0;
register_transport_failure:
kmem_cache_destroy(ig.desc_cache);
return err;
}
static void __exit iser_exit(void)
{
iser_dbg("Removing iSER datamover...\n");
iscsi_unregister_transport(&iscsi_iser_transport);
kmem_cache_destroy(ig.desc_cache);
}
module_init(iser_init);
module_exit(iser_exit);

398
drivers/infiniband/ulp/iser/iscsi_iser.h Normal file
View File

@@ -0,0 +1,398 @@
/*
* iSER transport for the Open iSCSI Initiator & iSER transport internals
*
* Copyright (C) 2004 Dmitry Yusupov
* Copyright (C) 2004 Alex Aizman
* Copyright (C) 2005 Mike Christie
* based on code maintained by open-iscsi@googlegroups.com
*
* Copyright (c) 2004, 2005, 2006 Voltaire, Inc. All rights reserved.
* Copyright (c) 2005, 2006 Cisco Systems. All rights reserved.
* Copyright (c) 2013 Mellanox Technologies. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#ifndef __ISCSI_ISER_H__
#define __ISCSI_ISER_H__
#include <linux/types.h>
#include <linux/net.h>
#include <linux/printk.h>
#include <scsi/libiscsi.h>
#include <scsi/scsi_transport_iscsi.h>
#include <linux/interrupt.h>
#include <linux/wait.h>
#include <linux/sched.h>
#include <linux/list.h>
#include <linux/slab.h>
#include <linux/dma-mapping.h>
#include <linux/mutex.h>
#include <linux/mempool.h>
#include <linux/uio.h>
#include <linux/socket.h>
#include <linux/in.h>
#include <linux/in6.h>
#include <rdma/ib_verbs.h>
#include <rdma/ib_fmr_pool.h>
#include <rdma/rdma_cm.h>
#define DRV_NAME "iser"
#define PFX DRV_NAME ": "
#define DRV_VER "1.1"
#define iser_dbg(fmt, arg...) \
do { \
if (iser_debug_level > 2) \
printk(KERN_DEBUG PFX "%s:" fmt,\
__func__ , ## arg); \
} while (0)
#define iser_warn(fmt, arg...) \
do { \
if (iser_debug_level > 1) \
pr_warn(PFX "%s:" fmt, \
__func__ , ## arg); \
} while (0)
#define iser_info(fmt, arg...) \
do { \
if (iser_debug_level > 0) \
pr_info(PFX "%s:" fmt, \
__func__ , ## arg); \
} while (0)
#define iser_err(fmt, arg...) \
do { \
printk(KERN_ERR PFX "%s:" fmt, \
__func__ , ## arg); \
} while (0)
#define SHIFT_4K 12
#define SIZE_4K (1ULL << SHIFT_4K)
#define MASK_4K (~(SIZE_4K-1))
/* support up to 512KB in one RDMA */
#define ISCSI_ISER_SG_TABLESIZE (0x80000 >> SHIFT_4K)
#define ISER_DEF_CMD_PER_LUN ISCSI_DEF_XMIT_CMDS_MAX
/* QP settings */
/* Maximal bounds on received asynchronous PDUs */
#define ISER_MAX_RX_MISC_PDUS 4 /* NOOP_IN(2) , ASYNC_EVENT(2) */
#define ISER_MAX_TX_MISC_PDUS 6 /* NOOP_OUT(2), TEXT(1), *
* SCSI_TMFUNC(2), LOGOUT(1) */
#define ISER_QP_MAX_RECV_DTOS (ISCSI_DEF_XMIT_CMDS_MAX)
#define ISER_MIN_POSTED_RX (ISCSI_DEF_XMIT_CMDS_MAX >> 2)
/* the max TX (send) WR supported by the iSER QP is defined by *
* max_send_wr = T * (1 + D) + C ; D is how many inflight dataouts we expect *
* to have at max for SCSI command. The tx posting & completion handling code *
* supports -EAGAIN scheme where tx is suspended till the QP has room for more *
* send WR. D=8 comes from 64K/8K */
#define ISER_INFLIGHT_DATAOUTS 8
#define ISER_QP_MAX_REQ_DTOS (ISCSI_DEF_XMIT_CMDS_MAX * \
(1 + ISER_INFLIGHT_DATAOUTS) + \
ISER_MAX_TX_MISC_PDUS + \
ISER_MAX_RX_MISC_PDUS)
#define ISER_VER 0x10
#define ISER_WSV 0x08
#define ISER_RSV 0x04
struct iser_hdr {
u8 flags;
u8 rsvd[3];
__be32 write_stag; /* write rkey */
__be64 write_va;
__be32 read_stag; /* read rkey */
__be64 read_va;
} __attribute__((packed));
#define ISER_ZBVA_NOT_SUPPORTED 0x80
#define ISER_SEND_W_INV_NOT_SUPPORTED 0x40
struct iser_cm_hdr {
u8 flags;
u8 rsvd[3];
} __packed;
/* Constant PDU lengths calculations */
#define ISER_HEADERS_LEN (sizeof(struct iser_hdr) + sizeof(struct iscsi_hdr))
#define ISER_RECV_DATA_SEG_LEN 128
#define ISER_RX_PAYLOAD_SIZE (ISER_HEADERS_LEN + ISER_RECV_DATA_SEG_LEN)
#define ISER_RX_LOGIN_SIZE (ISER_HEADERS_LEN + ISCSI_DEF_MAX_RECV_SEG_LEN)
/* Length of an object name string */
#define ISER_OBJECT_NAME_SIZE 64
enum iser_ib_conn_state {
ISER_CONN_INIT, /* descriptor allocd, no conn */
ISER_CONN_PENDING, /* in the process of being established */
ISER_CONN_UP, /* up and running */
ISER_CONN_TERMINATING, /* in the process of being terminated */
ISER_CONN_DOWN, /* shut down */
ISER_CONN_STATES_NUM
};
enum iser_task_status {
ISER_TASK_STATUS_INIT = 0,
ISER_TASK_STATUS_STARTED,
ISER_TASK_STATUS_COMPLETED
};
enum iser_data_dir {
ISER_DIR_IN = 0, /* to initiator */
ISER_DIR_OUT, /* from initiator */
ISER_DIRS_NUM
};
struct iser_data_buf {
void *buf; /* pointer to the sg list */
unsigned int size; /* num entries of this sg */
unsigned long data_len; /* total data len */
unsigned int dma_nents; /* returned by dma_map_sg */
char *copy_buf; /* allocated copy buf for SGs unaligned *
* for rdma which are copied */
struct scatterlist sg_single; /* SG-ified clone of a non SG SC or *
* unaligned SG */
};
/* fwd declarations */
struct iser_device;
struct iser_cq_desc;
struct iscsi_iser_conn;
struct iscsi_iser_task;
struct iscsi_endpoint;
struct iser_mem_reg {
u32 lkey;
u32 rkey;
u64 va;
u64 len;
void *mem_h;
int is_fmr;
};
struct iser_regd_buf {
struct iser_mem_reg reg; /* memory registration info */
void *virt_addr;
struct iser_device *device; /* device->device for dma_unmap */
enum dma_data_direction direction; /* direction for dma_unmap */
unsigned int data_size;
};
enum iser_desc_type {
ISCSI_TX_CONTROL ,
ISCSI_TX_SCSI_COMMAND,
ISCSI_TX_DATAOUT
};
struct iser_tx_desc {
struct iser_hdr iser_header;
struct iscsi_hdr iscsi_header;
enum iser_desc_type type;
u64 dma_addr;
/* sg[0] points to iser/iscsi headers, sg[1] optionally points to either
of immediate data, unsolicited data-out or control (login,text) */
struct ib_sge tx_sg[2];
int num_sge;
};
#define ISER_RX_PAD_SIZE (256 - (ISER_RX_PAYLOAD_SIZE + \
sizeof(u64) + sizeof(struct ib_sge)))
struct iser_rx_desc {
struct iser_hdr iser_header;
struct iscsi_hdr iscsi_header;
char data[ISER_RECV_DATA_SEG_LEN];
u64 dma_addr;
struct ib_sge rx_sg;
char pad[ISER_RX_PAD_SIZE];
} __attribute__((packed));
#define ISER_MAX_CQ 4
struct iser_device {
struct ib_device *ib_device;
struct ib_pd *pd;
struct ib_cq *rx_cq[ISER_MAX_CQ];
struct ib_cq *tx_cq[ISER_MAX_CQ];
struct ib_mr *mr;
struct tasklet_struct cq_tasklet[ISER_MAX_CQ];
struct ib_event_handler event_handler;
struct list_head ig_list; /* entry in ig devices list */
int refcount;
int cq_active_qps[ISER_MAX_CQ];
int cqs_used;
struct iser_cq_desc *cq_desc;
};
struct iser_conn {
struct iscsi_iser_conn *iser_conn; /* iser conn for upcalls */
struct iscsi_endpoint *ep;
enum iser_ib_conn_state state; /* rdma connection state */
atomic_t refcount;
spinlock_t lock; /* used for state changes */
struct iser_device *device; /* device context */
struct rdma_cm_id *cma_id; /* CMA ID */
struct ib_qp *qp; /* QP */
struct ib_fmr_pool *fmr_pool; /* pool of IB FMRs */
wait_queue_head_t wait; /* waitq for conn/disconn */
int post_recv_buf_count; /* posted rx count */
atomic_t post_send_buf_count; /* posted tx count */
char name[ISER_OBJECT_NAME_SIZE];
struct iser_page_vec *page_vec; /* represents SG to fmr maps*
* maps serialized as tx is*/
struct list_head conn_list; /* entry in ig conn list */
char *login_buf;
char *login_req_buf, *login_resp_buf;
u64 login_req_dma, login_resp_dma;
unsigned int rx_desc_head;
struct iser_rx_desc *rx_descs;
struct ib_recv_wr rx_wr[ISER_MIN_POSTED_RX];
};
struct iscsi_iser_conn {
struct iscsi_conn *iscsi_conn;/* ptr to iscsi conn */
struct iser_conn *ib_conn; /* iSER IB conn */
};
struct iscsi_iser_task {
struct iser_tx_desc desc;
struct iscsi_iser_conn *iser_conn;
enum iser_task_status status;
int command_sent; /* set if command sent */
int dir[ISER_DIRS_NUM]; /* set if dir use*/
struct iser_regd_buf rdma_regd[ISER_DIRS_NUM];/* regd rdma buf */
struct iser_data_buf data[ISER_DIRS_NUM]; /* orig. data des*/
struct iser_data_buf data_copy[ISER_DIRS_NUM];/* contig. copy */
};
struct iser_page_vec {
u64 *pages;
int length;
int offset;
int data_size;
};
struct iser_cq_desc {
struct iser_device *device;
int cq_index;
};
struct iser_global {
struct mutex device_list_mutex;/* */
struct list_head device_list; /* all iSER devices */
struct mutex connlist_mutex;
struct list_head connlist; /* all iSER IB connections */
struct kmem_cache *desc_cache;
};
extern struct iser_global ig;
extern int iser_debug_level;
/* allocate connection resources needed for rdma functionality */
int iser_conn_set_full_featured_mode(struct iscsi_conn *conn);
int iser_send_control(struct iscsi_conn *conn,
struct iscsi_task *task);
int iser_send_command(struct iscsi_conn *conn,
struct iscsi_task *task);
int iser_send_data_out(struct iscsi_conn *conn,
struct iscsi_task *task,
struct iscsi_data *hdr);
void iscsi_iser_recv(struct iscsi_conn *conn,
struct iscsi_hdr *hdr,
char *rx_data,
int rx_data_len);
void iser_conn_init(struct iser_conn *ib_conn);
void iser_conn_get(struct iser_conn *ib_conn);
int iser_conn_put(struct iser_conn *ib_conn, int destroy_cma_id_allowed);
void iser_conn_terminate(struct iser_conn *ib_conn);
void iser_rcv_completion(struct iser_rx_desc *desc,
unsigned long dto_xfer_len,
struct iser_conn *ib_conn);
void iser_snd_completion(struct iser_tx_desc *desc, struct iser_conn *ib_conn);
void iser_task_rdma_init(struct iscsi_iser_task *task);
void iser_task_rdma_finalize(struct iscsi_iser_task *task);
void iser_free_rx_descriptors(struct iser_conn *ib_conn);
void iser_finalize_rdma_unaligned_sg(struct iscsi_iser_task *task,
enum iser_data_dir cmd_dir);
int iser_reg_rdma_mem(struct iscsi_iser_task *task,
enum iser_data_dir cmd_dir);
int iser_connect(struct iser_conn *ib_conn,
struct sockaddr_in *src_addr,
struct sockaddr_in *dst_addr,
int non_blocking);
int iser_reg_page_vec(struct iser_conn *ib_conn,
struct iser_page_vec *page_vec,
struct iser_mem_reg *mem_reg);
void iser_unreg_mem(struct iser_mem_reg *mem_reg);
int iser_post_recvl(struct iser_conn *ib_conn);
int iser_post_recvm(struct iser_conn *ib_conn, int count);
int iser_post_send(struct iser_conn *ib_conn, struct iser_tx_desc *tx_desc);
int iser_dma_map_task_data(struct iscsi_iser_task *iser_task,
struct iser_data_buf *data,
enum iser_data_dir iser_dir,
enum dma_data_direction dma_dir);
void iser_dma_unmap_task_data(struct iscsi_iser_task *iser_task);
int iser_initialize_task_headers(struct iscsi_task *task,
struct iser_tx_desc *tx_desc);
int iser_alloc_rx_descriptors(struct iser_conn *ib_conn);
#endif

571
drivers/infiniband/ulp/iser/iser_initiator.c Normal file
View File

@@ -0,0 +1,571 @@
/*
* Copyright (c) 2004, 2005, 2006 Voltaire, Inc. All rights reserved.
* Copyright (c) 2013 Mellanox Technologies. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#include <linux/kernel.h>
#include <linux/slab.h>
#include <linux/mm.h>
#include <linux/scatterlist.h>
#include <linux/kfifo.h>
#include <scsi/scsi_cmnd.h>
#include <scsi/scsi_host.h>
#include "iscsi_iser.h"
/* Register user buffer memory and initialize passive rdma
* dto descriptor. Total data size is stored in
* iser_task->data[ISER_DIR_IN].data_len
*/
static int iser_prepare_read_cmd(struct iscsi_task *task,
unsigned int edtl)
{
struct iscsi_iser_task *iser_task = task->dd_data;
struct iser_regd_buf *regd_buf;
int err;
struct iser_hdr *hdr = &iser_task->desc.iser_header;
struct iser_data_buf *buf_in = &iser_task->data[ISER_DIR_IN];
err = iser_dma_map_task_data(iser_task,
buf_in,
ISER_DIR_IN,
DMA_FROM_DEVICE);
if (err)
return err;
if (edtl > iser_task->data[ISER_DIR_IN].data_len) {
iser_err("Total data length: %ld, less than EDTL: "
"%d, in READ cmd BHS itt: %d, conn: 0x%p\n",
iser_task->data[ISER_DIR_IN].data_len, edtl,
task->itt, iser_task->iser_conn);
return -EINVAL;
}
err = iser_reg_rdma_mem(iser_task,ISER_DIR_IN);
if (err) {
iser_err("Failed to set up Data-IN RDMA\n");
return err;
}
regd_buf = &iser_task->rdma_regd[ISER_DIR_IN];
hdr->flags |= ISER_RSV;
hdr->read_stag = cpu_to_be32(regd_buf->reg.rkey);
hdr->read_va = cpu_to_be64(regd_buf->reg.va);
iser_dbg("Cmd itt:%d READ tags RKEY:%#.4X VA:%#llX\n",
task->itt, regd_buf->reg.rkey,
(unsigned long long)regd_buf->reg.va);
return 0;
}
/* Register user buffer memory and initialize passive rdma
* dto descriptor. Total data size is stored in
* task->data[ISER_DIR_OUT].data_len
*/
static int
iser_prepare_write_cmd(struct iscsi_task *task,
unsigned int imm_sz,
unsigned int unsol_sz,
unsigned int edtl)
{
struct iscsi_iser_task *iser_task = task->dd_data;
struct iser_regd_buf *regd_buf;
int err;
struct iser_hdr *hdr = &iser_task->desc.iser_header;
struct iser_data_buf *buf_out = &iser_task->data[ISER_DIR_OUT];
struct ib_sge *tx_dsg = &iser_task->desc.tx_sg[1];
err = iser_dma_map_task_data(iser_task,
buf_out,
ISER_DIR_OUT,
DMA_TO_DEVICE);
if (err)
return err;
if (edtl > iser_task->data[ISER_DIR_OUT].data_len) {
iser_err("Total data length: %ld, less than EDTL: %d, "
"in WRITE cmd BHS itt: %d, conn: 0x%p\n",
iser_task->data[ISER_DIR_OUT].data_len,
edtl, task->itt, task->conn);
return -EINVAL;
}
err = iser_reg_rdma_mem(iser_task,ISER_DIR_OUT);
if (err != 0) {
iser_err("Failed to register write cmd RDMA mem\n");
return err;
}
regd_buf = &iser_task->rdma_regd[ISER_DIR_OUT];
if (unsol_sz < edtl) {
hdr->flags |= ISER_WSV;
hdr->write_stag = cpu_to_be32(regd_buf->reg.rkey);
hdr->write_va = cpu_to_be64(regd_buf->reg.va + unsol_sz);
iser_dbg("Cmd itt:%d, WRITE tags, RKEY:%#.4X "
"VA:%#llX + unsol:%d\n",
task->itt, regd_buf->reg.rkey,
(unsigned long long)regd_buf->reg.va, unsol_sz);
}
if (imm_sz > 0) {
iser_dbg("Cmd itt:%d, WRITE, adding imm.data sz: %d\n",
task->itt, imm_sz);
tx_dsg->addr = regd_buf->reg.va;
tx_dsg->length = imm_sz;
tx_dsg->lkey = regd_buf->reg.lkey;
iser_task->desc.num_sge = 2;
}
return 0;
}
/* creates a new tx descriptor and adds header regd buffer */
static void iser_create_send_desc(struct iser_conn *ib_conn,
struct iser_tx_desc *tx_desc)
{
struct iser_device *device = ib_conn->device;
ib_dma_sync_single_for_cpu(device->ib_device,
tx_desc->dma_addr, ISER_HEADERS_LEN, DMA_TO_DEVICE);
memset(&tx_desc->iser_header, 0, sizeof(struct iser_hdr));
tx_desc->iser_header.flags = ISER_VER;
tx_desc->num_sge = 1;
if (tx_desc->tx_sg[0].lkey != device->mr->lkey) {
tx_desc->tx_sg[0].lkey = device->mr->lkey;
iser_dbg("sdesc %p lkey mismatch, fixing\n", tx_desc);
}
}
int iser_alloc_rx_descriptors(struct iser_conn *ib_conn)
{
int i, j;
u64 dma_addr;
struct iser_rx_desc *rx_desc;
struct ib_sge *rx_sg;
struct iser_device *device = ib_conn->device;
ib_conn->rx_descs = kmalloc(ISER_QP_MAX_RECV_DTOS *
sizeof(struct iser_rx_desc), GFP_KERNEL);
if (!ib_conn->rx_descs)
goto rx_desc_alloc_fail;
rx_desc = ib_conn->rx_descs;
for (i = 0; i < ISER_QP_MAX_RECV_DTOS; i++, rx_desc++) {
dma_addr = ib_dma_map_single(device->ib_device, (void *)rx_desc,
ISER_RX_PAYLOAD_SIZE, DMA_FROM_DEVICE);
if (ib_dma_mapping_error(device->ib_device, dma_addr))
goto rx_desc_dma_map_failed;
rx_desc->dma_addr = dma_addr;
rx_sg = &rx_desc->rx_sg;
rx_sg->addr = rx_desc->dma_addr;
rx_sg->length = ISER_RX_PAYLOAD_SIZE;
rx_sg->lkey = device->mr->lkey;
}
ib_conn->rx_desc_head = 0;
return 0;
rx_desc_dma_map_failed:
rx_desc = ib_conn->rx_descs;
for (j = 0; j < i; j++, rx_desc++)
ib_dma_unmap_single(device->ib_device, rx_desc->dma_addr,
ISER_RX_PAYLOAD_SIZE, DMA_FROM_DEVICE);
kfree(ib_conn->rx_descs);
ib_conn->rx_descs = NULL;
rx_desc_alloc_fail:
iser_err("failed allocating rx descriptors / data buffers\n");
return -ENOMEM;
}
void iser_free_rx_descriptors(struct iser_conn *ib_conn)
{
int i;
struct iser_rx_desc *rx_desc;
struct iser_device *device = ib_conn->device;
if (!ib_conn->rx_descs)
return;
rx_desc = ib_conn->rx_descs;
for (i = 0; i < ISER_QP_MAX_RECV_DTOS; i++, rx_desc++)
ib_dma_unmap_single(device->ib_device, rx_desc->dma_addr,
ISER_RX_PAYLOAD_SIZE, DMA_FROM_DEVICE);
kfree(ib_conn->rx_descs);
}
static int iser_post_rx_bufs(struct iscsi_conn *conn, struct iscsi_hdr *req)
{
struct iscsi_iser_conn *iser_conn = conn->dd_data;
iser_dbg("req op %x flags %x\n", req->opcode, req->flags);
/* check if this is the last login - going to full feature phase */
if ((req->flags & ISCSI_FULL_FEATURE_PHASE) != ISCSI_FULL_FEATURE_PHASE)
return 0;
/*
* Check that there is one posted recv buffer (for the last login
* response) and no posted send buffers left - they must have been
* consumed during previous login phases.
*/
WARN_ON(iser_conn->ib_conn->post_recv_buf_count != 1);
WARN_ON(atomic_read(&iser_conn->ib_conn->post_send_buf_count) != 0);
iser_dbg("Initially post: %d\n", ISER_MIN_POSTED_RX);
/* Initial post receive buffers */
if (iser_post_recvm(iser_conn->ib_conn, ISER_MIN_POSTED_RX))
return -ENOMEM;
return 0;
}
/**
* iser_send_command - send command PDU
*/
int iser_send_command(struct iscsi_conn *conn,
struct iscsi_task *task)
{
struct iscsi_iser_conn *iser_conn = conn->dd_data;
struct iscsi_iser_task *iser_task = task->dd_data;
unsigned long edtl;
int err;
struct iser_data_buf *data_buf;
struct iscsi_scsi_req *hdr = (struct iscsi_scsi_req *)task->hdr;
struct scsi_cmnd *sc = task->sc;
struct iser_tx_desc *tx_desc = &iser_task->desc;
edtl = ntohl(hdr->data_length);
/* build the tx desc regd header and add it to the tx desc dto */
tx_desc->type = ISCSI_TX_SCSI_COMMAND;
iser_create_send_desc(iser_conn->ib_conn, tx_desc);
if (hdr->flags & ISCSI_FLAG_CMD_READ)
data_buf = &iser_task->data[ISER_DIR_IN];
else
data_buf = &iser_task->data[ISER_DIR_OUT];
if (scsi_sg_count(sc)) { /* using a scatter list */
data_buf->buf = scsi_sglist(sc);
data_buf->size = scsi_sg_count(sc);
}
data_buf->data_len = scsi_bufflen(sc);
if (hdr->flags & ISCSI_FLAG_CMD_READ) {
err = iser_prepare_read_cmd(task, edtl);
if (err)
goto send_command_error;
}
if (hdr->flags & ISCSI_FLAG_CMD_WRITE) {
err = iser_prepare_write_cmd(task,
task->imm_count,
task->imm_count +
task->unsol_r2t.data_length,
edtl);
if (err)
goto send_command_error;
}
iser_task->status = ISER_TASK_STATUS_STARTED;
err = iser_post_send(iser_conn->ib_conn, tx_desc);
if (!err)
return 0;
send_command_error:
iser_err("conn %p failed task->itt %d err %d\n",conn, task->itt, err);
return err;
}
/**
* iser_send_data_out - send data out PDU
*/
int iser_send_data_out(struct iscsi_conn *conn,
struct iscsi_task *task,
struct iscsi_data *hdr)
{
struct iscsi_iser_conn *iser_conn = conn->dd_data;
struct iscsi_iser_task *iser_task = task->dd_data;
struct iser_tx_desc *tx_desc = NULL;
struct iser_regd_buf *regd_buf;
unsigned long buf_offset;
unsigned long data_seg_len;
uint32_t itt;
int err = 0;
struct ib_sge *tx_dsg;
itt = (__force uint32_t)hdr->itt;
data_seg_len = ntoh24(hdr->dlength);
buf_offset = ntohl(hdr->offset);
iser_dbg("%s itt %d dseg_len %d offset %d\n",
__func__,(int)itt,(int)data_seg_len,(int)buf_offset);
tx_desc = kmem_cache_zalloc(ig.desc_cache, GFP_ATOMIC);
if (tx_desc == NULL) {
iser_err("Failed to alloc desc for post dataout\n");
return -ENOMEM;
}
tx_desc->type = ISCSI_TX_DATAOUT;
tx_desc->iser_header.flags = ISER_VER;
memcpy(&tx_desc->iscsi_header, hdr, sizeof(struct iscsi_hdr));
/* build the tx desc */
iser_initialize_task_headers(task, tx_desc);
regd_buf = &iser_task->rdma_regd[ISER_DIR_OUT];
tx_dsg = &tx_desc->tx_sg[1];
tx_dsg->addr = regd_buf->reg.va + buf_offset;
tx_dsg->length = data_seg_len;
tx_dsg->lkey = regd_buf->reg.lkey;
tx_desc->num_sge = 2;
if (buf_offset + data_seg_len > iser_task->data[ISER_DIR_OUT].data_len) {
iser_err("Offset:%ld & DSL:%ld in Data-Out "
"inconsistent with total len:%ld, itt:%d\n",
buf_offset, data_seg_len,
iser_task->data[ISER_DIR_OUT].data_len, itt);
err = -EINVAL;
goto send_data_out_error;
}
iser_dbg("data-out itt: %d, offset: %ld, sz: %ld\n",
itt, buf_offset, data_seg_len);
err = iser_post_send(iser_conn->ib_conn, tx_desc);
if (!err)
return 0;
send_data_out_error:
kmem_cache_free(ig.desc_cache, tx_desc);
iser_err("conn %p failed err %d\n",conn, err);
return err;
}
int iser_send_control(struct iscsi_conn *conn,
struct iscsi_task *task)
{
struct iscsi_iser_conn *iser_conn = conn->dd_data;
struct iscsi_iser_task *iser_task = task->dd_data;
struct iser_tx_desc *mdesc = &iser_task->desc;
unsigned long data_seg_len;
int err = 0;
struct iser_device *device;
struct iser_conn *ib_conn = iser_conn->ib_conn;
/* build the tx desc regd header and add it to the tx desc dto */
mdesc->type = ISCSI_TX_CONTROL;
iser_create_send_desc(iser_conn->ib_conn, mdesc);
device = iser_conn->ib_conn->device;
data_seg_len = ntoh24(task->hdr->dlength);
if (data_seg_len > 0) {
struct ib_sge *tx_dsg = &mdesc->tx_sg[1];
if (task != conn->login_task) {
iser_err("data present on non login task!!!\n");
goto send_control_error;
}
ib_dma_sync_single_for_cpu(device->ib_device,
ib_conn->login_req_dma, task->data_count,
DMA_TO_DEVICE);
memcpy(iser_conn->ib_conn->login_req_buf, task->data,
task->data_count);
ib_dma_sync_single_for_device(device->ib_device,
ib_conn->login_req_dma, task->data_count,
DMA_TO_DEVICE);
tx_dsg->addr = iser_conn->ib_conn->login_req_dma;
tx_dsg->length = task->data_count;
tx_dsg->lkey = device->mr->lkey;
mdesc->num_sge = 2;
}
if (task == conn->login_task) {
err = iser_post_recvl(iser_conn->ib_conn);
if (err)
goto send_control_error;
err = iser_post_rx_bufs(conn, task->hdr);
if (err)
goto send_control_error;
}
err = iser_post_send(iser_conn->ib_conn, mdesc);
if (!err)
return 0;
send_control_error:
iser_err("conn %p failed err %d\n",conn, err);
return err;
}
/**
* iser_rcv_dto_completion - recv DTO completion
*/
void iser_rcv_completion(struct iser_rx_desc *rx_desc,
unsigned long rx_xfer_len,
struct iser_conn *ib_conn)
{
struct iscsi_iser_conn *conn = ib_conn->iser_conn;
struct iscsi_hdr *hdr;
u64 rx_dma;
int rx_buflen, outstanding, count, err;
/* differentiate between login to all other PDUs */
if ((char *)rx_desc == ib_conn->login_resp_buf) {
rx_dma = ib_conn->login_resp_dma;
rx_buflen = ISER_RX_LOGIN_SIZE;
} else {
rx_dma = rx_desc->dma_addr;
rx_buflen = ISER_RX_PAYLOAD_SIZE;
}
ib_dma_sync_single_for_cpu(ib_conn->device->ib_device, rx_dma,
rx_buflen, DMA_FROM_DEVICE);
hdr = &rx_desc->iscsi_header;
iser_dbg("op 0x%x itt 0x%x dlen %d\n", hdr->opcode,
hdr->itt, (int)(rx_xfer_len - ISER_HEADERS_LEN));
iscsi_iser_recv(conn->iscsi_conn, hdr,
rx_desc->data, rx_xfer_len - ISER_HEADERS_LEN);
ib_dma_sync_single_for_device(ib_conn->device->ib_device, rx_dma,
rx_buflen, DMA_FROM_DEVICE);
/* decrementing conn->post_recv_buf_count only --after-- freeing the *
* task eliminates the need to worry on tasks which are completed in *
* parallel to the execution of iser_conn_term. So the code that waits *
* for the posted rx bufs refcount to become zero handles everything */
conn->ib_conn->post_recv_buf_count--;
if (rx_dma == ib_conn->login_resp_dma)
return;
outstanding = ib_conn->post_recv_buf_count;
if (outstanding + ISER_MIN_POSTED_RX <= ISER_QP_MAX_RECV_DTOS) {
count = min(ISER_QP_MAX_RECV_DTOS - outstanding,
ISER_MIN_POSTED_RX);
err = iser_post_recvm(ib_conn, count);
if (err)
iser_err("posting %d rx bufs err %d\n", count, err);
}
}
void iser_snd_completion(struct iser_tx_desc *tx_desc,
struct iser_conn *ib_conn)
{
struct iscsi_task *task;
struct iser_device *device = ib_conn->device;
if (tx_desc->type == ISCSI_TX_DATAOUT) {
ib_dma_unmap_single(device->ib_device, tx_desc->dma_addr,
ISER_HEADERS_LEN, DMA_TO_DEVICE);
kmem_cache_free(ig.desc_cache, tx_desc);
}
atomic_dec(&ib_conn->post_send_buf_count);
if (tx_desc->type == ISCSI_TX_CONTROL) {
/* this arithmetic is legal by libiscsi dd_data allocation */
task = (void *) ((long)(void *)tx_desc -
sizeof(struct iscsi_task));
if (task->hdr->itt == RESERVED_ITT)
iscsi_put_task(task);
}
}
void iser_task_rdma_init(struct iscsi_iser_task *iser_task)
{
iser_task->status = ISER_TASK_STATUS_INIT;
iser_task->dir[ISER_DIR_IN] = 0;
iser_task->dir[ISER_DIR_OUT] = 0;
iser_task->data[ISER_DIR_IN].data_len = 0;
iser_task->data[ISER_DIR_OUT].data_len = 0;
memset(&iser_task->rdma_regd[ISER_DIR_IN], 0,
sizeof(struct iser_regd_buf));
memset(&iser_task->rdma_regd[ISER_DIR_OUT], 0,
sizeof(struct iser_regd_buf));
}
void iser_task_rdma_finalize(struct iscsi_iser_task *iser_task)
{
int is_rdma_aligned = 1;
struct iser_regd_buf *regd;
/* if we were reading, copy back to unaligned sglist,
* anyway dma_unmap and free the copy
*/
if (iser_task->data_copy[ISER_DIR_IN].copy_buf != NULL) {
is_rdma_aligned = 0;
iser_finalize_rdma_unaligned_sg(iser_task, ISER_DIR_IN);
}
if (iser_task->data_copy[ISER_DIR_OUT].copy_buf != NULL) {
is_rdma_aligned = 0;
iser_finalize_rdma_unaligned_sg(iser_task, ISER_DIR_OUT);
}
if (iser_task->dir[ISER_DIR_IN]) {
regd = &iser_task->rdma_regd[ISER_DIR_IN];
if (regd->reg.is_fmr)
iser_unreg_mem(&regd->reg);
}
if (iser_task->dir[ISER_DIR_OUT]) {
regd = &iser_task->rdma_regd[ISER_DIR_OUT];
if (regd->reg.is_fmr)
iser_unreg_mem(&regd->reg);
}
/* if the data was unaligned, it was already unmapped and then copied */
if (is_rdma_aligned)
iser_dma_unmap_task_data(iser_task);
}

425
drivers/infiniband/ulp/iser/iser_memory.c Normal file
View File

@@ -0,0 +1,425 @@
/*
* Copyright (c) 2004, 2005, 2006 Voltaire, Inc. All rights reserved.
* Copyright (c) 2013 Mellanox Technologies. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/slab.h>
#include <linux/mm.h>
#include <linux/highmem.h>
#include <linux/scatterlist.h>
#include "iscsi_iser.h"
#define ISER_KMALLOC_THRESHOLD 0x20000 /* 128K - kmalloc limit */
/**
* iser_start_rdma_unaligned_sg
*/
static int iser_start_rdma_unaligned_sg(struct iscsi_iser_task *iser_task,
enum iser_data_dir cmd_dir)
{
int dma_nents;
struct ib_device *dev;
char *mem = NULL;
struct iser_data_buf *data = &iser_task->data[cmd_dir];
unsigned long cmd_data_len = data->data_len;
if (cmd_data_len > ISER_KMALLOC_THRESHOLD)
mem = (void *)__get_free_pages(GFP_ATOMIC,
ilog2(roundup_pow_of_two(cmd_data_len)) - PAGE_SHIFT);
else
mem = kmalloc(cmd_data_len, GFP_ATOMIC);
if (mem == NULL) {
iser_err("Failed to allocate mem size %d %d for copying sglist\n",
data->size,(int)cmd_data_len);
return -ENOMEM;
}
if (cmd_dir == ISER_DIR_OUT) {
/* copy the unaligned sg the buffer which is used for RDMA */
struct scatterlist *sgl = (struct scatterlist *)data->buf;
struct scatterlist *sg;
int i;
char *p, *from;
p = mem;
for_each_sg(sgl, sg, data->size, i) {
from = kmap_atomic(sg_page(sg));
memcpy(p,
from + sg->offset,
sg->length);
kunmap_atomic(from);
p += sg->length;
}
}
sg_init_one(&iser_task->data_copy[cmd_dir].sg_single, mem, cmd_data_len);
iser_task->data_copy[cmd_dir].buf =
&iser_task->data_copy[cmd_dir].sg_single;
iser_task->data_copy[cmd_dir].size = 1;
iser_task->data_copy[cmd_dir].copy_buf = mem;
dev = iser_task->iser_conn->ib_conn->device->ib_device;
dma_nents = ib_dma_map_sg(dev,
&iser_task->data_copy[cmd_dir].sg_single,
1,
(cmd_dir == ISER_DIR_OUT) ?
DMA_TO_DEVICE : DMA_FROM_DEVICE);
BUG_ON(dma_nents == 0);
iser_task->data_copy[cmd_dir].dma_nents = dma_nents;
return 0;
}
/**
* iser_finalize_rdma_unaligned_sg
*/
void iser_finalize_rdma_unaligned_sg(struct iscsi_iser_task *iser_task,
enum iser_data_dir cmd_dir)
{
struct ib_device *dev;
struct iser_data_buf *mem_copy;
unsigned long cmd_data_len;
dev = iser_task->iser_conn->ib_conn->device->ib_device;
mem_copy = &iser_task->data_copy[cmd_dir];
ib_dma_unmap_sg(dev, &mem_copy->sg_single, 1,
(cmd_dir == ISER_DIR_OUT) ?
DMA_TO_DEVICE : DMA_FROM_DEVICE);
if (cmd_dir == ISER_DIR_IN) {
char *mem;
struct scatterlist *sgl, *sg;
unsigned char *p, *to;
unsigned int sg_size;
int i;
/* copy back read RDMA to unaligned sg */
mem = mem_copy->copy_buf;
sgl = (struct scatterlist *)iser_task->data[ISER_DIR_IN].buf;
sg_size = iser_task->data[ISER_DIR_IN].size;
p = mem;
for_each_sg(sgl, sg, sg_size, i) {
to = kmap_atomic(sg_page(sg));
memcpy(to + sg->offset,
p,
sg->length);
kunmap_atomic(to);
p += sg->length;
}
}
cmd_data_len = iser_task->data[cmd_dir].data_len;
if (cmd_data_len > ISER_KMALLOC_THRESHOLD)
free_pages((unsigned long)mem_copy->copy_buf,
ilog2(roundup_pow_of_two(cmd_data_len)) - PAGE_SHIFT);
else
kfree(mem_copy->copy_buf);
mem_copy->copy_buf = NULL;
}
#define IS_4K_ALIGNED(addr) ((((unsigned long)addr) & ~MASK_4K) == 0)
/**
* iser_sg_to_page_vec - Translates scatterlist entries to physical addresses
* and returns the length of resulting physical address array (may be less than
* the original due to possible compaction).
*
* we build a "page vec" under the assumption that the SG meets the RDMA
* alignment requirements. Other then the first and last SG elements, all
* the "internal" elements can be compacted into a list whose elements are
* dma addresses of physical pages. The code supports also the weird case
* where --few fragments of the same page-- are present in the SG as
* consecutive elements. Also, it handles one entry SG.
*/
static int iser_sg_to_page_vec(struct iser_data_buf *data,
struct iser_page_vec *page_vec,
struct ib_device *ibdev)
{
struct scatterlist *sg, *sgl = (struct scatterlist *)data->buf;
u64 start_addr, end_addr, page, chunk_start = 0;
unsigned long total_sz = 0;
unsigned int dma_len;
int i, new_chunk, cur_page, last_ent = data->dma_nents - 1;
/* compute the offset of first element */
page_vec->offset = (u64) sgl[0].offset & ~MASK_4K;
new_chunk = 1;
cur_page = 0;
for_each_sg(sgl, sg, data->dma_nents, i) {
start_addr = ib_sg_dma_address(ibdev, sg);
if (new_chunk)
chunk_start = start_addr;
dma_len = ib_sg_dma_len(ibdev, sg);
end_addr = start_addr + dma_len;
total_sz += dma_len;
/* collect page fragments until aligned or end of SG list */
if (!IS_4K_ALIGNED(end_addr) && i < last_ent) {
new_chunk = 0;
continue;
}
new_chunk = 1;
/* address of the first page in the contiguous chunk;
masking relevant for the very first SG entry,
which might be unaligned */
page = chunk_start & MASK_4K;
do {
page_vec->pages[cur_page++] = page;
page += SIZE_4K;
} while (page < end_addr);
}
page_vec->data_size = total_sz;
iser_dbg("page_vec->data_size:%d cur_page %d\n", page_vec->data_size,cur_page);
return cur_page;
}
/**
* iser_data_buf_aligned_len - Tries to determine the maximal correctly aligned
* for RDMA sub-list of a scatter-gather list of memory buffers, and returns
* the number of entries which are aligned correctly. Supports the case where
* consecutive SG elements are actually fragments of the same physcial page.
*/
static int iser_data_buf_aligned_len(struct iser_data_buf *data,
struct ib_device *ibdev)
{
struct scatterlist *sgl, *sg, *next_sg = NULL;
u64 start_addr, end_addr;
int i, ret_len, start_check = 0;
if (data->dma_nents == 1)
return 1;
sgl = (struct scatterlist *)data->buf;
start_addr = ib_sg_dma_address(ibdev, sgl);
for_each_sg(sgl, sg, data->dma_nents, i) {
if (start_check && !IS_4K_ALIGNED(start_addr))
break;
next_sg = sg_next(sg);
if (!next_sg)
break;
end_addr = start_addr + ib_sg_dma_len(ibdev, sg);
start_addr = ib_sg_dma_address(ibdev, next_sg);
if (end_addr == start_addr) {
start_check = 0;
continue;
} else
start_check = 1;
if (!IS_4K_ALIGNED(end_addr))
break;
}
ret_len = (next_sg) ? i : i+1;
iser_dbg("Found %d aligned entries out of %d in sg:0x%p\n",
ret_len, data->dma_nents, data);
return ret_len;
}
static void iser_data_buf_dump(struct iser_data_buf *data,
struct ib_device *ibdev)
{
struct scatterlist *sgl = (struct scatterlist *)data->buf;
struct scatterlist *sg;
int i;
if (iser_debug_level == 0)
return;
for_each_sg(sgl, sg, data->dma_nents, i)
iser_warn("sg[%d] dma_addr:0x%lX page:0x%p "
"off:0x%x sz:0x%x dma_len:0x%x\n",
i, (unsigned long)ib_sg_dma_address(ibdev, sg),
sg_page(sg), sg->offset,
sg->length, ib_sg_dma_len(ibdev, sg));
}
static void iser_dump_page_vec(struct iser_page_vec *page_vec)
{
int i;
iser_err("page vec length %d data size %d\n",
page_vec->length, page_vec->data_size);
for (i = 0; i < page_vec->length; i++)
iser_err("%d %lx\n",i,(unsigned long)page_vec->pages[i]);
}
static void iser_page_vec_build(struct iser_data_buf *data,
struct iser_page_vec *page_vec,
struct ib_device *ibdev)
{
int page_vec_len = 0;
page_vec->length = 0;
page_vec->offset = 0;
iser_dbg("Translating sg sz: %d\n", data->dma_nents);
page_vec_len = iser_sg_to_page_vec(data, page_vec, ibdev);
iser_dbg("sg len %d page_vec_len %d\n", data->dma_nents,page_vec_len);
page_vec->length = page_vec_len;
if (page_vec_len * SIZE_4K < page_vec->data_size) {
iser_err("page_vec too short to hold this SG\n");
iser_data_buf_dump(data, ibdev);
iser_dump_page_vec(page_vec);
BUG();
}
}
int iser_dma_map_task_data(struct iscsi_iser_task *iser_task,
struct iser_data_buf *data,
enum iser_data_dir iser_dir,
enum dma_data_direction dma_dir)
{
struct ib_device *dev;
iser_task->dir[iser_dir] = 1;
dev = iser_task->iser_conn->ib_conn->device->ib_device;
data->dma_nents = ib_dma_map_sg(dev, data->buf, data->size, dma_dir);
if (data->dma_nents == 0) {
iser_err("dma_map_sg failed!!!\n");
return -EINVAL;
}
return 0;
}
void iser_dma_unmap_task_data(struct iscsi_iser_task *iser_task)
{
struct ib_device *dev;
struct iser_data_buf *data;
dev = iser_task->iser_conn->ib_conn->device->ib_device;
if (iser_task->dir[ISER_DIR_IN]) {
data = &iser_task->data[ISER_DIR_IN];
ib_dma_unmap_sg(dev, data->buf, data->size, DMA_FROM_DEVICE);
}
if (iser_task->dir[ISER_DIR_OUT]) {
data = &iser_task->data[ISER_DIR_OUT];
ib_dma_unmap_sg(dev, data->buf, data->size, DMA_TO_DEVICE);
}
}
/**
* iser_reg_rdma_mem - Registers memory intended for RDMA,
* obtaining rkey and va
*
* returns 0 on success, errno code on failure
*/
int iser_reg_rdma_mem(struct iscsi_iser_task *iser_task,
enum iser_data_dir cmd_dir)
{
struct iscsi_conn *iscsi_conn = iser_task->iser_conn->iscsi_conn;
struct iser_conn *ib_conn = iser_task->iser_conn->ib_conn;
struct iser_device *device = ib_conn->device;
struct ib_device *ibdev = device->ib_device;
struct iser_data_buf *mem = &iser_task->data[cmd_dir];
struct iser_regd_buf *regd_buf;
int aligned_len;
int err;
int i;
struct scatterlist *sg;
regd_buf = &iser_task->rdma_regd[cmd_dir];
aligned_len = iser_data_buf_aligned_len(mem, ibdev);
if (aligned_len != mem->dma_nents ||
(!ib_conn->fmr_pool && mem->dma_nents > 1)) {
iscsi_conn->fmr_unalign_cnt++;
iser_warn("rdma alignment violation (%d/%d aligned) or FMR not supported\n",
aligned_len, mem->size);
iser_data_buf_dump(mem, ibdev);
/* unmap the command data before accessing it */
iser_dma_unmap_task_data(iser_task);
/* allocate copy buf, if we are writing, copy the */
/* unaligned scatterlist, dma map the copy */
if (iser_start_rdma_unaligned_sg(iser_task, cmd_dir) != 0)
return -ENOMEM;
mem = &iser_task->data_copy[cmd_dir];
}
/* if there a single dma entry, FMR is not needed */
if (mem->dma_nents == 1) {
sg = (struct scatterlist *)mem->buf;
regd_buf->reg.lkey = device->mr->lkey;
regd_buf->reg.rkey = device->mr->rkey;
regd_buf->reg.len = ib_sg_dma_len(ibdev, &sg[0]);
regd_buf->reg.va = ib_sg_dma_address(ibdev, &sg[0]);
regd_buf->reg.is_fmr = 0;
iser_dbg("PHYSICAL Mem.register: lkey: 0x%08X rkey: 0x%08X "
"va: 0x%08lX sz: %ld]\n",
(unsigned int)regd_buf->reg.lkey,
(unsigned int)regd_buf->reg.rkey,
(unsigned long)regd_buf->reg.va,
(unsigned long)regd_buf->reg.len);
} else { /* use FMR for multiple dma entries */
iser_page_vec_build(mem, ib_conn->page_vec, ibdev);
err = iser_reg_page_vec(ib_conn, ib_conn->page_vec, &regd_buf->reg);
if (err && err != -EAGAIN) {
iser_data_buf_dump(mem, ibdev);
iser_err("mem->dma_nents = %d (dlength = 0x%x)\n",
mem->dma_nents,
ntoh24(iser_task->desc.iscsi_header.dlength));
iser_err("page_vec: data_size = 0x%x, length = %d, offset = 0x%x\n",
ib_conn->page_vec->data_size, ib_conn->page_vec->length,
ib_conn->page_vec->offset);
for (i=0 ; i<ib_conn->page_vec->length ; i++)
iser_err("page_vec[%d] = 0x%llx\n", i,
(unsigned long long) ib_conn->page_vec->pages[i]);
}
if (err)
return err;
}
return 0;
}

931
drivers/infiniband/ulp/iser/iser_verbs.c Normal file
View File

@@ -0,0 +1,931 @@
/*
* Copyright (c) 2004, 2005, 2006 Voltaire, Inc. All rights reserved.
* Copyright (c) 2005, 2006 Cisco Systems. All rights reserved.
* Copyright (c) 2013 Mellanox Technologies. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/slab.h>
#include <linux/delay.h>
#include "iscsi_iser.h"
#define ISCSI_ISER_MAX_CONN 8
#define ISER_MAX_RX_CQ_LEN (ISER_QP_MAX_RECV_DTOS * ISCSI_ISER_MAX_CONN)
#define ISER_MAX_TX_CQ_LEN (ISER_QP_MAX_REQ_DTOS * ISCSI_ISER_MAX_CONN)
static void iser_cq_tasklet_fn(unsigned long data);
static void iser_cq_callback(struct ib_cq *cq, void *cq_context);
static void iser_cq_event_callback(struct ib_event *cause, void *context)
{
iser_err("got cq event %d \n", cause->event);
}
static void iser_qp_event_callback(struct ib_event *cause, void *context)
{
iser_err("got qp event %d\n",cause->event);
}
static void iser_event_handler(struct ib_event_handler *handler,
struct ib_event *event)
{
iser_err("async event %d on device %s port %d\n", event->event,
event->device->name, event->element.port_num);
}
/**
* iser_create_device_ib_res - creates Protection Domain (PD), Completion
* Queue (CQ), DMA Memory Region (DMA MR) with the device associated with
* the adapator.
*
* returns 0 on success, -1 on failure
*/
static int iser_create_device_ib_res(struct iser_device *device)
{
int i, j;
struct iser_cq_desc *cq_desc;
device->cqs_used = min(ISER_MAX_CQ, device->ib_device->num_comp_vectors);
iser_info("using %d CQs, device %s supports %d vectors\n",
device->cqs_used, device->ib_device->name,
device->ib_device->num_comp_vectors);
device->cq_desc = kmalloc(sizeof(struct iser_cq_desc) * device->cqs_used,
GFP_KERNEL);
if (device->cq_desc == NULL)
goto cq_desc_err;
cq_desc = device->cq_desc;
device->pd = ib_alloc_pd(device->ib_device);
if (IS_ERR(device->pd))
goto pd_err;
for (i = 0; i < device->cqs_used; i++) {
cq_desc[i].device = device;
cq_desc[i].cq_index = i;
device->rx_cq[i] = ib_create_cq(device->ib_device,
iser_cq_callback,
iser_cq_event_callback,
(void *)&cq_desc[i],
ISER_MAX_RX_CQ_LEN, i);
if (IS_ERR(device->rx_cq[i]))
goto cq_err;
device->tx_cq[i] = ib_create_cq(device->ib_device,
NULL, iser_cq_event_callback,
(void *)&cq_desc[i],
ISER_MAX_TX_CQ_LEN, i);
if (IS_ERR(device->tx_cq[i]))
goto cq_err;
if (ib_req_notify_cq(device->rx_cq[i], IB_CQ_NEXT_COMP))
goto cq_err;
tasklet_init(&device->cq_tasklet[i],
iser_cq_tasklet_fn,
(unsigned long)&cq_desc[i]);
}
device->mr = ib_get_dma_mr(device->pd, IB_ACCESS_LOCAL_WRITE |
IB_ACCESS_REMOTE_WRITE |
IB_ACCESS_REMOTE_READ);
if (IS_ERR(device->mr))
goto dma_mr_err;
INIT_IB_EVENT_HANDLER(&device->event_handler, device->ib_device,
iser_event_handler);
if (ib_register_event_handler(&device->event_handler))
goto handler_err;
return 0;
handler_err:
ib_dereg_mr(device->mr);
dma_mr_err:
for (j = 0; j < device->cqs_used; j++)
tasklet_kill(&device->cq_tasklet[j]);
cq_err:
for (j = 0; j < i; j++) {
if (device->tx_cq[j])
ib_destroy_cq(device->tx_cq[j]);
if (device->rx_cq[j])
ib_destroy_cq(device->rx_cq[j]);
}
ib_dealloc_pd(device->pd);
pd_err:
kfree(device->cq_desc);
cq_desc_err:
iser_err("failed to allocate an IB resource\n");
return -1;
}
/**
* iser_free_device_ib_res - destroy/dealloc/dereg the DMA MR,
* CQ and PD created with the device associated with the adapator.
*/
static void iser_free_device_ib_res(struct iser_device *device)
{
int i;
BUG_ON(device->mr == NULL);
for (i = 0; i < device->cqs_used; i++) {
tasklet_kill(&device->cq_tasklet[i]);
(void)ib_destroy_cq(device->tx_cq[i]);
(void)ib_destroy_cq(device->rx_cq[i]);
device->tx_cq[i] = NULL;
device->rx_cq[i] = NULL;
}
(void)ib_unregister_event_handler(&device->event_handler);
(void)ib_dereg_mr(device->mr);
(void)ib_dealloc_pd(device->pd);
kfree(device->cq_desc);
device->mr = NULL;
device->pd = NULL;
}
/**
* iser_create_ib_conn_res - Creates FMR pool and Queue-Pair (QP)
*
* returns 0 on success, -1 on failure
*/
static int iser_create_ib_conn_res(struct iser_conn *ib_conn)
{
struct iser_device *device;
struct ib_qp_init_attr init_attr;
int req_err, resp_err, ret = -ENOMEM;
struct ib_fmr_pool_param params;
int index, min_index = 0;
BUG_ON(ib_conn->device == NULL);
device = ib_conn->device;
ib_conn->login_buf = kmalloc(ISCSI_DEF_MAX_RECV_SEG_LEN +
ISER_RX_LOGIN_SIZE, GFP_KERNEL);
if (!ib_conn->login_buf)
goto out_err;
ib_conn->login_req_buf = ib_conn->login_buf;
ib_conn->login_resp_buf = ib_conn->login_buf + ISCSI_DEF_MAX_RECV_SEG_LEN;
ib_conn->login_req_dma = ib_dma_map_single(ib_conn->device->ib_device,
(void *)ib_conn->login_req_buf,
ISCSI_DEF_MAX_RECV_SEG_LEN, DMA_TO_DEVICE);
ib_conn->login_resp_dma = ib_dma_map_single(ib_conn->device->ib_device,
(void *)ib_conn->login_resp_buf,
ISER_RX_LOGIN_SIZE, DMA_FROM_DEVICE);
req_err = ib_dma_mapping_error(device->ib_device, ib_conn->login_req_dma);
resp_err = ib_dma_mapping_error(device->ib_device, ib_conn->login_resp_dma);
if (req_err || resp_err) {
if (req_err)
ib_conn->login_req_dma = 0;
if (resp_err)
ib_conn->login_resp_dma = 0;
goto out_err;
}
ib_conn->page_vec = kmalloc(sizeof(struct iser_page_vec) +
(sizeof(u64) * (ISCSI_ISER_SG_TABLESIZE +1)),
GFP_KERNEL);
if (!ib_conn->page_vec)
goto out_err;
ib_conn->page_vec->pages = (u64 *) (ib_conn->page_vec + 1);
params.page_shift = SHIFT_4K;
/* when the first/last SG element are not start/end *
* page aligned, the map whould be of N+1 pages */
params.max_pages_per_fmr = ISCSI_ISER_SG_TABLESIZE + 1;
/* make the pool size twice the max number of SCSI commands *
* the ML is expected to queue, watermark for unmap at 50% */
params.pool_size = ISCSI_DEF_XMIT_CMDS_MAX * 2;
params.dirty_watermark = ISCSI_DEF_XMIT_CMDS_MAX;
params.cache = 0;
params.flush_function = NULL;
params.access = (IB_ACCESS_LOCAL_WRITE |
IB_ACCESS_REMOTE_WRITE |
IB_ACCESS_REMOTE_READ);
ib_conn->fmr_pool = ib_create_fmr_pool(device->pd, &params);
ret = PTR_ERR(ib_conn->fmr_pool);
if (IS_ERR(ib_conn->fmr_pool) && ret != -ENOSYS) {
ib_conn->fmr_pool = NULL;
goto out_err;
} else if (ret == -ENOSYS) {
ib_conn->fmr_pool = NULL;
iser_warn("FMRs are not supported, using unaligned mode\n");
ret = 0;
}
memset(&init_attr, 0, sizeof init_attr);
mutex_lock(&ig.connlist_mutex);
/* select the CQ with the minimal number of usages */
for (index = 0; index < device->cqs_used; index++)
if (device->cq_active_qps[index] <
device->cq_active_qps[min_index])
min_index = index;
device->cq_active_qps[min_index]++;
mutex_unlock(&ig.connlist_mutex);
iser_info("cq index %d used for ib_conn %p\n", min_index, ib_conn);
init_attr.event_handler = iser_qp_event_callback;
init_attr.qp_context = (void *)ib_conn;
init_attr.send_cq = device->tx_cq[min_index];
init_attr.recv_cq = device->rx_cq[min_index];
init_attr.cap.max_send_wr = ISER_QP_MAX_REQ_DTOS;
init_attr.cap.max_recv_wr = ISER_QP_MAX_RECV_DTOS;
init_attr.cap.max_send_sge = 2;
init_attr.cap.max_recv_sge = 1;
init_attr.sq_sig_type = IB_SIGNAL_REQ_WR;
init_attr.qp_type = IB_QPT_RC;
ret = rdma_create_qp(ib_conn->cma_id, device->pd, &init_attr);
if (ret)
goto out_err;
ib_conn->qp = ib_conn->cma_id->qp;
iser_info("setting conn %p cma_id %p: fmr_pool %p qp %p\n",
ib_conn, ib_conn->cma_id,
ib_conn->fmr_pool, ib_conn->cma_id->qp);
return ret;
out_err:
iser_err("unable to alloc mem or create resource, err %d\n", ret);
return ret;
}
/**
* releases the FMR pool and QP objects, returns 0 on success,
* -1 on failure
*/
static int iser_free_ib_conn_res(struct iser_conn *ib_conn)
{
int cq_index;
BUG_ON(ib_conn == NULL);
iser_info("freeing conn %p cma_id %p fmr pool %p qp %p\n",
ib_conn, ib_conn->cma_id,
ib_conn->fmr_pool, ib_conn->qp);
/* qp is created only once both addr & route are resolved */
if (ib_conn->fmr_pool != NULL)
ib_destroy_fmr_pool(ib_conn->fmr_pool);
if (ib_conn->qp != NULL) {
cq_index = ((struct iser_cq_desc *)ib_conn->qp->recv_cq->cq_context)->cq_index;
ib_conn->device->cq_active_qps[cq_index]--;
rdma_destroy_qp(ib_conn->cma_id);
}
ib_conn->fmr_pool = NULL;
ib_conn->qp = NULL;
kfree(ib_conn->page_vec);
if (ib_conn->login_buf) {
if (ib_conn->login_req_dma)
ib_dma_unmap_single(ib_conn->device->ib_device,
ib_conn->login_req_dma,
ISCSI_DEF_MAX_RECV_SEG_LEN, DMA_TO_DEVICE);
if (ib_conn->login_resp_dma)
ib_dma_unmap_single(ib_conn->device->ib_device,
ib_conn->login_resp_dma,
ISER_RX_LOGIN_SIZE, DMA_FROM_DEVICE);
kfree(ib_conn->login_buf);
}
return 0;
}
/**
* based on the resolved device node GUID see if there already allocated
* device for this device. If there's no such, create one.
*/
static
struct iser_device *iser_device_find_by_ib_device(struct rdma_cm_id *cma_id)
{
struct iser_device *device;
mutex_lock(&ig.device_list_mutex);
list_for_each_entry(device, &ig.device_list, ig_list)
/* find if there's a match using the node GUID */
if (device->ib_device->node_guid == cma_id->device->node_guid)
goto inc_refcnt;
device = kzalloc(sizeof *device, GFP_KERNEL);
if (device == NULL)
goto out;
/* assign this device to the device */
device->ib_device = cma_id->device;
/* init the device and link it into ig device list */
if (iser_create_device_ib_res(device)) {
kfree(device);
device = NULL;
goto out;
}
list_add(&device->ig_list, &ig.device_list);
inc_refcnt:
device->refcount++;
out:
mutex_unlock(&ig.device_list_mutex);
return device;
}
/* if there's no demand for this device, release it */
static void iser_device_try_release(struct iser_device *device)
{
mutex_lock(&ig.device_list_mutex);
device->refcount--;
iser_info("device %p refcount %d\n", device, device->refcount);
if (!device->refcount) {
iser_free_device_ib_res(device);
list_del(&device->ig_list);
kfree(device);
}
mutex_unlock(&ig.device_list_mutex);
}
static int iser_conn_state_comp_exch(struct iser_conn *ib_conn,
enum iser_ib_conn_state comp,
enum iser_ib_conn_state exch)
{
int ret;
spin_lock_bh(&ib_conn->lock);
if ((ret = (ib_conn->state == comp)))
ib_conn->state = exch;
spin_unlock_bh(&ib_conn->lock);
return ret;
}
/**
* Frees all conn objects and deallocs conn descriptor
*/
static void iser_conn_release(struct iser_conn *ib_conn, int can_destroy_id)
{
struct iser_device *device = ib_conn->device;
BUG_ON(ib_conn->state != ISER_CONN_DOWN);
mutex_lock(&ig.connlist_mutex);
list_del(&ib_conn->conn_list);
mutex_unlock(&ig.connlist_mutex);
iser_free_rx_descriptors(ib_conn);
iser_free_ib_conn_res(ib_conn);
ib_conn->device = NULL;
/* on EVENT_ADDR_ERROR there's no device yet for this conn */
if (device != NULL)
iser_device_try_release(device);
/* if cma handler context, the caller actually destroy the id */
if (ib_conn->cma_id != NULL && can_destroy_id) {
rdma_destroy_id(ib_conn->cma_id);
ib_conn->cma_id = NULL;
}
iscsi_destroy_endpoint(ib_conn->ep);
}
void iser_conn_get(struct iser_conn *ib_conn)
{
atomic_inc(&ib_conn->refcount);
}
int iser_conn_put(struct iser_conn *ib_conn, int can_destroy_id)
{
if (atomic_dec_and_test(&ib_conn->refcount)) {
iser_conn_release(ib_conn, can_destroy_id);
return 1;
}
return 0;
}
/**
* triggers start of the disconnect procedures and wait for them to be done
*/
void iser_conn_terminate(struct iser_conn *ib_conn)
{
int err = 0;
/* change the ib conn state only if the conn is UP, however always call
* rdma_disconnect since this is the only way to cause the CMA to change
* the QP state to ERROR
*/
iser_conn_state_comp_exch(ib_conn, ISER_CONN_UP, ISER_CONN_TERMINATING);
err = rdma_disconnect(ib_conn->cma_id);
if (err)
iser_err("Failed to disconnect, conn: 0x%p err %d\n",
ib_conn,err);
wait_event_interruptible(ib_conn->wait,
ib_conn->state == ISER_CONN_DOWN);
iser_conn_put(ib_conn, 1); /* deref ib conn deallocate */
}
static int iser_connect_error(struct rdma_cm_id *cma_id)
{
struct iser_conn *ib_conn;
ib_conn = (struct iser_conn *)cma_id->context;
ib_conn->state = ISER_CONN_DOWN;
wake_up_interruptible(&ib_conn->wait);
return iser_conn_put(ib_conn, 0); /* deref ib conn's cma id */
}
static int iser_addr_handler(struct rdma_cm_id *cma_id)
{
struct iser_device *device;
struct iser_conn *ib_conn;
int ret;
device = iser_device_find_by_ib_device(cma_id);
if (!device) {
iser_err("device lookup/creation failed\n");
return iser_connect_error(cma_id);
}
ib_conn = (struct iser_conn *)cma_id->context;
ib_conn->device = device;
ret = rdma_resolve_route(cma_id, 1000);
if (ret) {
iser_err("resolve route failed: %d\n", ret);
return iser_connect_error(cma_id);
}
return 0;
}
static int iser_route_handler(struct rdma_cm_id *cma_id)
{
struct rdma_conn_param conn_param;
int ret;
struct iser_cm_hdr req_hdr;
ret = iser_create_ib_conn_res((struct iser_conn *)cma_id->context);
if (ret)
goto failure;
memset(&conn_param, 0, sizeof conn_param);
conn_param.responder_resources = 4;
conn_param.initiator_depth = 1;
conn_param.retry_count = 7;
conn_param.rnr_retry_count = 6;
memset(&req_hdr, 0, sizeof(req_hdr));
req_hdr.flags = (ISER_ZBVA_NOT_SUPPORTED |
ISER_SEND_W_INV_NOT_SUPPORTED);
conn_param.private_data = (void *)&req_hdr;
conn_param.private_data_len = sizeof(struct iser_cm_hdr);
ret = rdma_connect(cma_id, &conn_param);
if (ret) {
iser_err("failure connecting: %d\n", ret);
goto failure;
}
return 0;
failure:
return iser_connect_error(cma_id);
}
static void iser_connected_handler(struct rdma_cm_id *cma_id)
{
struct iser_conn *ib_conn;
ib_conn = (struct iser_conn *)cma_id->context;
ib_conn->state = ISER_CONN_UP;
wake_up_interruptible(&ib_conn->wait);
}
static int iser_disconnected_handler(struct rdma_cm_id *cma_id)
{
struct iser_conn *ib_conn;
int ret;
ib_conn = (struct iser_conn *)cma_id->context;
/* getting here when the state is UP means that the conn is being *
* terminated asynchronously from the iSCSI layer's perspective. */
if (iser_conn_state_comp_exch(ib_conn, ISER_CONN_UP,
ISER_CONN_TERMINATING))
iscsi_conn_failure(ib_conn->iser_conn->iscsi_conn,
ISCSI_ERR_CONN_FAILED);
/* Complete the termination process if no posts are pending */
if (ib_conn->post_recv_buf_count == 0 &&
(atomic_read(&ib_conn->post_send_buf_count) == 0)) {
ib_conn->state = ISER_CONN_DOWN;
wake_up_interruptible(&ib_conn->wait);
}
ret = iser_conn_put(ib_conn, 0); /* deref ib conn's cma id */
return ret;
}
static int iser_cma_handler(struct rdma_cm_id *cma_id, struct rdma_cm_event *event)
{
int ret = 0;
iser_info("event %d status %d conn %p id %p\n",
event->event, event->status, cma_id->context, cma_id);
switch (event->event) {
case RDMA_CM_EVENT_ADDR_RESOLVED:
ret = iser_addr_handler(cma_id);
break;
case RDMA_CM_EVENT_ROUTE_RESOLVED:
ret = iser_route_handler(cma_id);
break;
case RDMA_CM_EVENT_ESTABLISHED:
iser_connected_handler(cma_id);
break;
case RDMA_CM_EVENT_ADDR_ERROR:
case RDMA_CM_EVENT_ROUTE_ERROR:
case RDMA_CM_EVENT_CONNECT_ERROR:
case RDMA_CM_EVENT_UNREACHABLE:
case RDMA_CM_EVENT_REJECTED:
ret = iser_connect_error(cma_id);
break;
case RDMA_CM_EVENT_DISCONNECTED:
case RDMA_CM_EVENT_DEVICE_REMOVAL:
case RDMA_CM_EVENT_ADDR_CHANGE:
ret = iser_disconnected_handler(cma_id);
break;
default:
iser_err("Unexpected RDMA CM event (%d)\n", event->event);
break;
}
return ret;
}
void iser_conn_init(struct iser_conn *ib_conn)
{
ib_conn->state = ISER_CONN_INIT;
init_waitqueue_head(&ib_conn->wait);
ib_conn->post_recv_buf_count = 0;
atomic_set(&ib_conn->post_send_buf_count, 0);
atomic_set(&ib_conn->refcount, 1); /* ref ib conn allocation */
INIT_LIST_HEAD(&ib_conn->conn_list);
spin_lock_init(&ib_conn->lock);
}
/**
* starts the process of connecting to the target
* sleeps until the connection is established or rejected
*/
int iser_connect(struct iser_conn *ib_conn,
struct sockaddr_in *src_addr,
struct sockaddr_in *dst_addr,
int non_blocking)
{
struct sockaddr *src, *dst;
int err = 0;
sprintf(ib_conn->name, "%pI4:%d",
&dst_addr->sin_addr.s_addr, dst_addr->sin_port);
/* the device is known only --after-- address resolution */
ib_conn->device = NULL;
iser_info("connecting to: %pI4, port 0x%x\n",
&dst_addr->sin_addr, dst_addr->sin_port);
ib_conn->state = ISER_CONN_PENDING;
iser_conn_get(ib_conn); /* ref ib conn's cma id */
ib_conn->cma_id = rdma_create_id(iser_cma_handler,
(void *)ib_conn,
RDMA_PS_TCP, IB_QPT_RC);
if (IS_ERR(ib_conn->cma_id)) {
err = PTR_ERR(ib_conn->cma_id);
iser_err("rdma_create_id failed: %d\n", err);
goto id_failure;
}
src = (struct sockaddr *)src_addr;
dst = (struct sockaddr *)dst_addr;
err = rdma_resolve_addr(ib_conn->cma_id, src, dst, 1000);
if (err) {
iser_err("rdma_resolve_addr failed: %d\n", err);
goto addr_failure;
}
if (!non_blocking) {
wait_event_interruptible(ib_conn->wait,
(ib_conn->state != ISER_CONN_PENDING));
if (ib_conn->state != ISER_CONN_UP) {
err = -EIO;
goto connect_failure;
}
}
mutex_lock(&ig.connlist_mutex);
list_add(&ib_conn->conn_list, &ig.connlist);
mutex_unlock(&ig.connlist_mutex);
return 0;
id_failure:
ib_conn->cma_id = NULL;
addr_failure:
ib_conn->state = ISER_CONN_DOWN;
iser_conn_put(ib_conn, 1); /* deref ib conn's cma id */
connect_failure:
iser_conn_put(ib_conn, 1); /* deref ib conn deallocate */
return err;
}
/**
* iser_reg_page_vec - Register physical memory
*
* returns: 0 on success, errno code on failure
*/
int iser_reg_page_vec(struct iser_conn *ib_conn,
struct iser_page_vec *page_vec,
struct iser_mem_reg *mem_reg)
{
struct ib_pool_fmr *mem;
u64 io_addr;
u64 *page_list;
int status;
page_list = page_vec->pages;
io_addr = page_list[0];
mem = ib_fmr_pool_map_phys(ib_conn->fmr_pool,
page_list,
page_vec->length,
io_addr);
if (IS_ERR(mem)) {
status = (int)PTR_ERR(mem);
iser_err("ib_fmr_pool_map_phys failed: %d\n", status);
return status;
}
mem_reg->lkey = mem->fmr->lkey;
mem_reg->rkey = mem->fmr->rkey;
mem_reg->len = page_vec->length * SIZE_4K;
mem_reg->va = io_addr;
mem_reg->is_fmr = 1;
mem_reg->mem_h = (void *)mem;
mem_reg->va += page_vec->offset;
mem_reg->len = page_vec->data_size;
iser_dbg("PHYSICAL Mem.register, [PHYS p_array: 0x%p, sz: %d, "
"entry[0]: (0x%08lx,%ld)] -> "
"[lkey: 0x%08X mem_h: 0x%p va: 0x%08lX sz: %ld]\n",
page_vec, page_vec->length,
(unsigned long)page_vec->pages[0],
(unsigned long)page_vec->data_size,
(unsigned int)mem_reg->lkey, mem_reg->mem_h,
(unsigned long)mem_reg->va, (unsigned long)mem_reg->len);
return 0;
}
/**
* Unregister (previosuly registered) memory.
*/
void iser_unreg_mem(struct iser_mem_reg *reg)
{
int ret;
iser_dbg("PHYSICAL Mem.Unregister mem_h %p\n",reg->mem_h);
ret = ib_fmr_pool_unmap((struct ib_pool_fmr *)reg->mem_h);
if (ret)
iser_err("ib_fmr_pool_unmap failed %d\n", ret);
reg->mem_h = NULL;
}
int iser_post_recvl(struct iser_conn *ib_conn)
{
struct ib_recv_wr rx_wr, *rx_wr_failed;
struct ib_sge sge;
int ib_ret;
sge.addr = ib_conn->login_resp_dma;
sge.length = ISER_RX_LOGIN_SIZE;
sge.lkey = ib_conn->device->mr->lkey;
rx_wr.wr_id = (unsigned long)ib_conn->login_resp_buf;
rx_wr.sg_list = &sge;
rx_wr.num_sge = 1;
rx_wr.next = NULL;
ib_conn->post_recv_buf_count++;
ib_ret = ib_post_recv(ib_conn->qp, &rx_wr, &rx_wr_failed);
if (ib_ret) {
iser_err("ib_post_recv failed ret=%d\n", ib_ret);
ib_conn->post_recv_buf_count--;
}
return ib_ret;
}
int iser_post_recvm(struct iser_conn *ib_conn, int count)
{
struct ib_recv_wr *rx_wr, *rx_wr_failed;
int i, ib_ret;
unsigned int my_rx_head = ib_conn->rx_desc_head;
struct iser_rx_desc *rx_desc;
for (rx_wr = ib_conn->rx_wr, i = 0; i < count; i++, rx_wr++) {
rx_desc = &ib_conn->rx_descs[my_rx_head];
rx_wr->wr_id = (unsigned long)rx_desc;
rx_wr->sg_list = &rx_desc->rx_sg;
rx_wr->num_sge = 1;
rx_wr->next = rx_wr + 1;
my_rx_head = (my_rx_head + 1) & (ISER_QP_MAX_RECV_DTOS - 1);
}
rx_wr--;
rx_wr->next = NULL; /* mark end of work requests list */
ib_conn->post_recv_buf_count += count;
ib_ret = ib_post_recv(ib_conn->qp, ib_conn->rx_wr, &rx_wr_failed);
if (ib_ret) {
iser_err("ib_post_recv failed ret=%d\n", ib_ret);
ib_conn->post_recv_buf_count -= count;
} else
ib_conn->rx_desc_head = my_rx_head;
return ib_ret;
}
/**
* iser_start_send - Initiate a Send DTO operation
*
* returns 0 on success, -1 on failure
*/
int iser_post_send(struct iser_conn *ib_conn, struct iser_tx_desc *tx_desc)
{
int ib_ret;
struct ib_send_wr send_wr, *send_wr_failed;
ib_dma_sync_single_for_device(ib_conn->device->ib_device,
tx_desc->dma_addr, ISER_HEADERS_LEN, DMA_TO_DEVICE);
send_wr.next = NULL;
send_wr.wr_id = (unsigned long)tx_desc;
send_wr.sg_list = tx_desc->tx_sg;
send_wr.num_sge = tx_desc->num_sge;
send_wr.opcode = IB_WR_SEND;
send_wr.send_flags = IB_SEND_SIGNALED;
atomic_inc(&ib_conn->post_send_buf_count);
ib_ret = ib_post_send(ib_conn->qp, &send_wr, &send_wr_failed);
if (ib_ret) {
iser_err("ib_post_send failed, ret:%d\n", ib_ret);
atomic_dec(&ib_conn->post_send_buf_count);
}
return ib_ret;
}
static void iser_handle_comp_error(struct iser_tx_desc *desc,
struct iser_conn *ib_conn)
{
if (desc && desc->type == ISCSI_TX_DATAOUT)
kmem_cache_free(ig.desc_cache, desc);
if (ib_conn->post_recv_buf_count == 0 &&
atomic_read(&ib_conn->post_send_buf_count) == 0) {
/* getting here when the state is UP means that the conn is *
* being terminated asynchronously from the iSCSI layer's *
* perspective. */
if (iser_conn_state_comp_exch(ib_conn, ISER_CONN_UP,
ISER_CONN_TERMINATING))
iscsi_conn_failure(ib_conn->iser_conn->iscsi_conn,
ISCSI_ERR_CONN_FAILED);
/* no more non completed posts to the QP, complete the
* termination process w.o worrying on disconnect event */
ib_conn->state = ISER_CONN_DOWN;
wake_up_interruptible(&ib_conn->wait);
}
}
static int iser_drain_tx_cq(struct iser_device *device, int cq_index)
{
struct ib_cq *cq = device->tx_cq[cq_index];
struct ib_wc wc;
struct iser_tx_desc *tx_desc;
struct iser_conn *ib_conn;
int completed_tx = 0;
while (ib_poll_cq(cq, 1, &wc) == 1) {
tx_desc = (struct iser_tx_desc *) (unsigned long) wc.wr_id;
ib_conn = wc.qp->qp_context;
if (wc.status == IB_WC_SUCCESS) {
if (wc.opcode == IB_WC_SEND)
iser_snd_completion(tx_desc, ib_conn);
else
iser_err("expected opcode %d got %d\n",
IB_WC_SEND, wc.opcode);
} else {
iser_err("tx id %llx status %d vend_err %x\n",
wc.wr_id, wc.status, wc.vendor_err);
atomic_dec(&ib_conn->post_send_buf_count);
iser_handle_comp_error(tx_desc, ib_conn);
}
completed_tx++;
}
return completed_tx;
}
static void iser_cq_tasklet_fn(unsigned long data)
{
struct iser_cq_desc *cq_desc = (struct iser_cq_desc *)data;
struct iser_device *device = cq_desc->device;
int cq_index = cq_desc->cq_index;
struct ib_cq *cq = device->rx_cq[cq_index];
struct ib_wc wc;
struct iser_rx_desc *desc;
unsigned long xfer_len;
struct iser_conn *ib_conn;
int completed_tx, completed_rx;
completed_tx = completed_rx = 0;
while (ib_poll_cq(cq, 1, &wc) == 1) {
desc = (struct iser_rx_desc *) (unsigned long) wc.wr_id;
BUG_ON(desc == NULL);
ib_conn = wc.qp->qp_context;
if (wc.status == IB_WC_SUCCESS) {
if (wc.opcode == IB_WC_RECV) {
xfer_len = (unsigned long)wc.byte_len;
iser_rcv_completion(desc, xfer_len, ib_conn);
} else
iser_err("expected opcode %d got %d\n",
IB_WC_RECV, wc.opcode);
} else {
if (wc.status != IB_WC_WR_FLUSH_ERR)
iser_err("rx id %llx status %d vend_err %x\n",
wc.wr_id, wc.status, wc.vendor_err);
ib_conn->post_recv_buf_count--;
iser_handle_comp_error(NULL, ib_conn);
}
completed_rx++;
if (!(completed_rx & 63))
completed_tx += iser_drain_tx_cq(device, cq_index);
}
/* #warning "it is assumed here that arming CQ only once its empty" *
* " would not cause interrupts to be missed" */
ib_req_notify_cq(cq, IB_CQ_NEXT_COMP);
completed_tx += iser_drain_tx_cq(device, cq_index);
iser_dbg("got %d rx %d tx completions\n", completed_rx, completed_tx);
}
static void iser_cq_callback(struct ib_cq *cq, void *cq_context)
{
struct iser_cq_desc *cq_desc = (struct iser_cq_desc *)cq_context;
struct iser_device *device = cq_desc->device;
int cq_index = cq_desc->cq_index;
tasklet_schedule(&device->cq_tasklet[cq_index]);
}

5
drivers/infiniband/ulp/isert/Kconfig Normal file
View File

@@ -0,0 +1,5 @@
config INFINIBAND_ISERT
tristate "iSCSI Extentions for RDMA (iSER) target support"
depends on INET && INFINIBAND_ADDR_TRANS && TARGET_CORE && ISCSI_TARGET
---help---
Support for iSCSI Extentions for RDMA (iSER) Target on Infiniband fabrics.

2
drivers/infiniband/ulp/isert/Makefile Normal file
View File

@@ -0,0 +1,2 @@
ccflags-y := -Idrivers/target -Idrivers/target/iscsi
obj-$(CONFIG_INFINIBAND_ISERT) += ib_isert.o

2344
drivers/infiniband/ulp/isert/ib_isert.c Normal file
View File

File diff suppressed because it is too large Load Diff

139
drivers/infiniband/ulp/isert/ib_isert.h Normal file
View File

@@ -0,0 +1,139 @@
#include <linux/socket.h>
#include <linux/in.h>
#include <linux/in6.h>
#include <rdma/ib_verbs.h>
#include <rdma/rdma_cm.h>
#define ISERT_RDMA_LISTEN_BACKLOG 10
enum isert_desc_type {
ISCSI_TX_CONTROL,
ISCSI_TX_DATAIN
};
enum iser_ib_op_code {
ISER_IB_RECV,
ISER_IB_SEND,
ISER_IB_RDMA_WRITE,
ISER_IB_RDMA_READ,
};
enum iser_conn_state {
ISER_CONN_INIT,
ISER_CONN_UP,
ISER_CONN_TERMINATING,
ISER_CONN_DOWN,
};
struct iser_rx_desc {
struct iser_hdr iser_header;
struct iscsi_hdr iscsi_header;
char data[ISER_RECV_DATA_SEG_LEN];
u64 dma_addr;
struct ib_sge rx_sg;
char pad[ISER_RX_PAD_SIZE];
} __packed;
struct iser_tx_desc {
struct iser_hdr iser_header;
struct iscsi_hdr iscsi_header;
enum isert_desc_type type;
u64 dma_addr;
struct ib_sge tx_sg[2];
int num_sge;
struct isert_cmd *isert_cmd;
struct ib_send_wr send_wr;
} __packed;
struct isert_rdma_wr {
struct list_head wr_list;
struct isert_cmd *isert_cmd;
enum iser_ib_op_code iser_ib_op;
struct ib_sge *ib_sge;
int num_sge;
struct scatterlist *sge;
int send_wr_num;
struct ib_send_wr *send_wr;
};
struct isert_cmd {
uint32_t read_stag;
uint32_t write_stag;
uint64_t read_va;
uint64_t write_va;
u64 sense_buf_dma;
u32 sense_buf_len;
u32 read_va_off;
u32 write_va_off;
u32 rdma_wr_num;
struct isert_conn *conn;
struct iscsi_cmd iscsi_cmd;
struct ib_sge *ib_sge;
struct iser_tx_desc tx_desc;
struct isert_rdma_wr rdma_wr;
struct work_struct comp_work;
};
struct isert_device;
struct isert_conn {
enum iser_conn_state state;
bool logout_posted;
int post_recv_buf_count;
atomic_t post_send_buf_count;
u32 responder_resources;
u32 initiator_depth;
u32 max_sge;
char *login_buf;
char *login_req_buf;
char *login_rsp_buf;
u64 login_req_dma;
u64 login_rsp_dma;
unsigned int conn_rx_desc_head;
struct iser_rx_desc *conn_rx_descs;
struct ib_recv_wr conn_rx_wr[ISERT_MIN_POSTED_RX];
struct iscsi_conn *conn;
struct list_head conn_accept_node;
struct completion conn_login_comp;
struct iser_tx_desc conn_login_tx_desc;
struct rdma_cm_id *conn_cm_id;
struct ib_pd *conn_pd;
struct ib_mr *conn_mr;
struct ib_qp *conn_qp;
struct isert_device *conn_device;
struct work_struct conn_logout_work;
struct mutex conn_mutex;
wait_queue_head_t conn_wait;
wait_queue_head_t conn_wait_comp_err;
struct kref conn_kref;
};
#define ISERT_MAX_CQ 64
struct isert_cq_desc {
struct isert_device *device;
int cq_index;
struct work_struct cq_rx_work;
struct work_struct cq_tx_work;
};
struct isert_device {
int cqs_used;
int refcount;
int cq_active_qps[ISERT_MAX_CQ];
struct ib_device *ib_device;
struct ib_pd *dev_pd;
struct ib_mr *dev_mr;
struct ib_cq *dev_rx_cq[ISERT_MAX_CQ];
struct ib_cq *dev_tx_cq[ISERT_MAX_CQ];
struct isert_cq_desc *cq_desc;
struct list_head dev_node;
};
struct isert_np {
wait_queue_head_t np_accept_wq;
struct rdma_cm_id *np_cm_id;
struct mutex np_accept_mutex;
struct list_head np_accept_list;
struct completion np_login_comp;
};

47
drivers/infiniband/ulp/isert/isert_proto.h Normal file
View File

@@ -0,0 +1,47 @@
/* From iscsi_iser.h */
struct iser_hdr {
u8 flags;
u8 rsvd[3];
__be32 write_stag; /* write rkey */
__be64 write_va;
__be32 read_stag; /* read rkey */
__be64 read_va;
} __packed;
/*Constant PDU lengths calculations */
#define ISER_HEADERS_LEN (sizeof(struct iser_hdr) + sizeof(struct iscsi_hdr))
#define ISER_RECV_DATA_SEG_LEN 8192
#define ISER_RX_PAYLOAD_SIZE (ISER_HEADERS_LEN + ISER_RECV_DATA_SEG_LEN)
#define ISER_RX_LOGIN_SIZE (ISER_HEADERS_LEN + ISCSI_DEF_MAX_RECV_SEG_LEN)
/* QP settings */
/* Maximal bounds on received asynchronous PDUs */
#define ISERT_MAX_TX_MISC_PDUS 4 /* NOOP_IN(2) , ASYNC_EVENT(2) */
#define ISERT_MAX_RX_MISC_PDUS 6 /* NOOP_OUT(2), TEXT(1), *
* SCSI_TMFUNC(2), LOGOUT(1) */
#define ISCSI_DEF_XMIT_CMDS_MAX 128 /* from libiscsi.h, must be power of 2 */
#define ISERT_QP_MAX_RECV_DTOS (ISCSI_DEF_XMIT_CMDS_MAX)
#define ISERT_MIN_POSTED_RX (ISCSI_DEF_XMIT_CMDS_MAX >> 2)
#define ISERT_INFLIGHT_DATAOUTS 8
#define ISERT_QP_MAX_REQ_DTOS (ISCSI_DEF_XMIT_CMDS_MAX * \
(1 + ISERT_INFLIGHT_DATAOUTS) + \
ISERT_MAX_TX_MISC_PDUS + \
ISERT_MAX_RX_MISC_PDUS)
#define ISER_RX_PAD_SIZE (ISER_RECV_DATA_SEG_LEN + 4096 - \
(ISER_RX_PAYLOAD_SIZE + sizeof(u64) + sizeof(struct ib_sge)))
#define ISER_VER 0x10
#define ISER_WSV 0x08
#define ISER_RSV 0x04
#define ISCSI_CTRL 0x10
#define ISER_HELLO 0x20
#define ISER_HELLORPLY 0x30

1
drivers/infiniband/ulp/srp/Kbuild Normal file
View File

@@ -0,0 +1 @@
obj-$(CONFIG_INFINIBAND_SRP) += ib_srp.o

12
drivers/infiniband/ulp/srp/Kconfig Normal file
View File

@@ -0,0 +1,12 @@
config INFINIBAND_SRP
tristate "InfiniBand SCSI RDMA Protocol"
depends on SCSI
select SCSI_SRP_ATTRS
---help---
Support for the SCSI RDMA Protocol over InfiniBand. This
allows you to access storage devices that speak SRP over
InfiniBand.
The SRP protocol is defined by the INCITS T10 technical
committee. See <http://www.t10.org/>.

2610
drivers/infiniband/ulp/srp/ib_srp.c Normal file
View File

File diff suppressed because it is too large Load Diff

212
drivers/infiniband/ulp/srp/ib_srp.h Normal file
View File

@@ -0,0 +1,212 @@
/*
* Copyright (c) 2005 Cisco Systems. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#ifndef IB_SRP_H
#define IB_SRP_H
#include <linux/types.h>
#include <linux/list.h>
#include <linux/mutex.h>
#include <linux/scatterlist.h>
#include <scsi/scsi_host.h>
#include <scsi/scsi_cmnd.h>
#include <rdma/ib_verbs.h>
#include <rdma/ib_sa.h>
#include <rdma/ib_cm.h>
#include <rdma/ib_fmr_pool.h>
enum {
SRP_PATH_REC_TIMEOUT_MS = 1000,
SRP_ABORT_TIMEOUT_MS = 5000,
SRP_PORT_REDIRECT = 1,
SRP_DLID_REDIRECT = 2,
SRP_STALE_CONN = 3,
SRP_MAX_LUN = 512,
SRP_DEF_SG_TABLESIZE = 12,
SRP_RQ_SHIFT = 6,
SRP_RQ_SIZE = 1 << SRP_RQ_SHIFT,
SRP_SQ_SIZE = SRP_RQ_SIZE,
SRP_RSP_SQ_SIZE = 1,
SRP_REQ_SQ_SIZE = SRP_SQ_SIZE - SRP_RSP_SQ_SIZE,
SRP_TSK_MGMT_SQ_SIZE = 1,
SRP_CMD_SQ_SIZE = SRP_REQ_SQ_SIZE - SRP_TSK_MGMT_SQ_SIZE,
SRP_TAG_NO_REQ = ~0U,
SRP_TAG_TSK_MGMT = 1U << 31,
SRP_FMR_SIZE = 512,
SRP_FMR_MIN_SIZE = 128,
SRP_FMR_POOL_SIZE = 1024,
SRP_FMR_DIRTY_SIZE = SRP_FMR_POOL_SIZE / 4,
SRP_MAP_ALLOW_FMR = 0,
SRP_MAP_NO_FMR = 1,
};
enum srp_target_state {
SRP_TARGET_LIVE,
SRP_TARGET_REMOVED,
};
enum srp_iu_type {
SRP_IU_CMD,
SRP_IU_TSK_MGMT,
SRP_IU_RSP,
};
struct srp_device {
struct list_head dev_list;
struct ib_device *dev;
struct ib_pd *pd;
struct ib_mr *mr;
struct ib_fmr_pool *fmr_pool;
u64 fmr_page_mask;
int fmr_page_size;
int fmr_max_size;
};
struct srp_host {
struct srp_device *srp_dev;
u8 port;
struct device dev;
struct list_head target_list;
spinlock_t target_lock;
struct completion released;
struct list_head list;
};
struct srp_request {
struct list_head list;
struct scsi_cmnd *scmnd;
struct srp_iu *cmd;
struct ib_pool_fmr **fmr_list;
u64 *map_page;
struct srp_direct_buf *indirect_desc;
dma_addr_t indirect_dma_addr;
short nfmr;
short index;
};
struct srp_target_port {
/* These are RW in the hot path, and commonly used together */
struct list_head free_tx;
struct list_head free_reqs;
spinlock_t lock;
s32 req_lim;
/* These are read-only in the hot path */
struct ib_cq *send_cq ____cacheline_aligned_in_smp;
struct ib_cq *recv_cq;
struct ib_qp *qp;
u32 lkey;
u32 rkey;
enum srp_target_state state;
unsigned int max_iu_len;
unsigned int cmd_sg_cnt;
unsigned int indirect_size;
bool allow_ext_sg;
bool transport_offline;
/* Everything above this point is used in the hot path of
* command processing. Try to keep them packed into cachelines.
*/
__be64 id_ext;
__be64 ioc_guid;
__be64 service_id;
__be64 initiator_ext;
u16 io_class;
struct srp_host *srp_host;
struct Scsi_Host *scsi_host;
char target_name[32];
unsigned int scsi_id;
unsigned int sg_tablesize;
struct ib_sa_path_rec path;
__be16 orig_dgid[8];
struct ib_sa_query *path_query;
int path_query_id;
u32 rq_tmo_jiffies;
bool connected;
struct ib_cm_id *cm_id;
int max_ti_iu_len;
int zero_req_lim;
struct srp_iu *tx_ring[SRP_SQ_SIZE];
struct srp_iu *rx_ring[SRP_RQ_SIZE];
struct srp_request req_ring[SRP_CMD_SQ_SIZE];
struct work_struct remove_work;
struct list_head list;
struct completion done;
int status;
bool qp_in_error;
struct completion tsk_mgmt_done;
u8 tsk_mgmt_status;
};
struct srp_iu {
struct list_head list;
u64 dma;
void *buf;
size_t size;
enum dma_data_direction direction;
};
struct srp_map_state {
struct ib_pool_fmr **next_fmr;
struct srp_direct_buf *desc;
u64 *pages;
dma_addr_t base_dma_addr;
u32 fmr_len;
u32 total_len;
unsigned int npages;
unsigned int nfmr;
unsigned int ndesc;
struct scatterlist *unmapped_sg;
int unmapped_index;
dma_addr_t unmapped_addr;
};
#endif /* IB_SRP_H */

12
drivers/infiniband/ulp/srpt/Kconfig Normal file
View File

@@ -0,0 +1,12 @@
config INFINIBAND_SRPT
tristate "InfiniBand SCSI RDMA Protocol target support"
depends on INFINIBAND && TARGET_CORE
---help---
Support for the SCSI RDMA Protocol (SRP) Target driver. The
SRP protocol is a protocol that allows an initiator to access
a block storage device on another host (target) over a network
that supports the RDMA protocol. Currently the RDMA protocol is
supported by InfiniBand and by iWarp network hardware. More
information about the SRP protocol can be found on the website
of the INCITS T10 technical committee (http://www.t10.org/).

2
drivers/infiniband/ulp/srpt/Makefile Normal file
View File

@@ -0,0 +1,2 @@
ccflags-y := -Idrivers/target
obj-$(CONFIG_INFINIBAND_SRPT) += ib_srpt.o

139
drivers/infiniband/ulp/srpt/ib_dm_mad.h Normal file
View File

@@ -0,0 +1,139 @@
/*
* Copyright (c) 2006 - 2009 Mellanox Technology Inc. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*
*/
#ifndef IB_DM_MAD_H
#define IB_DM_MAD_H
#include <linux/types.h>
#include <rdma/ib_mad.h>
enum {
/*
* See also section 13.4.7 Status Field, table 115 MAD Common Status
* Field Bit Values and also section 16.3.1.1 Status Field in the
* InfiniBand Architecture Specification.
*/
DM_MAD_STATUS_UNSUP_METHOD = 0x0008,
DM_MAD_STATUS_UNSUP_METHOD_ATTR = 0x000c,
DM_MAD_STATUS_INVALID_FIELD = 0x001c,
DM_MAD_STATUS_NO_IOC = 0x0100,
/*
* See also the Device Management chapter, section 16.3.3 Attributes,
* table 279 Device Management Attributes in the InfiniBand
* Architecture Specification.
*/
DM_ATTR_CLASS_PORT_INFO = 0x01,
DM_ATTR_IOU_INFO = 0x10,
DM_ATTR_IOC_PROFILE = 0x11,
DM_ATTR_SVC_ENTRIES = 0x12
};
struct ib_dm_hdr {
u8 reserved[28];
};
/*
* Structure of management datagram sent by the SRP target implementation.
* Contains a management datagram header, reliable multi-packet transaction
* protocol (RMPP) header and ib_dm_hdr. Notes:
* - The SRP target implementation does not use RMPP or ib_dm_hdr when sending
* management datagrams.
* - The header size must be exactly 64 bytes (IB_MGMT_DEVICE_HDR), since this
* is the header size that is passed to ib_create_send_mad() in ib_srpt.c.
* - The maximum supported size for a management datagram when not using RMPP
* is 256 bytes -- 64 bytes header and 192 (IB_MGMT_DEVICE_DATA) bytes data.
*/
struct ib_dm_mad {
struct ib_mad_hdr mad_hdr;
struct ib_rmpp_hdr rmpp_hdr;
struct ib_dm_hdr dm_hdr;
u8 data[IB_MGMT_DEVICE_DATA];
};
/*
* IOUnitInfo as defined in section 16.3.3.3 IOUnitInfo of the InfiniBand
* Architecture Specification.
*/
struct ib_dm_iou_info {
__be16 change_id;
u8 max_controllers;
u8 op_rom;
u8 controller_list[128];
};
/*
* IOControllerprofile as defined in section 16.3.3.4 IOControllerProfile of
* the InfiniBand Architecture Specification.
*/
struct ib_dm_ioc_profile {
__be64 guid;
__be32 vendor_id;
__be32 device_id;
__be16 device_version;
__be16 reserved1;
__be32 subsys_vendor_id;
__be32 subsys_device_id;
__be16 io_class;
__be16 io_subclass;
__be16 protocol;
__be16 protocol_version;
__be16 service_conn;
__be16 initiators_supported;
__be16 send_queue_depth;
u8 reserved2;
u8 rdma_read_depth;
__be32 send_size;
__be32 rdma_size;
u8 op_cap_mask;
u8 svc_cap_mask;
u8 num_svc_entries;
u8 reserved3[9];
u8 id_string[64];
};
struct ib_dm_svc_entry {
u8 name[40];
__be64 id;
};
/*
* See also section 16.3.3.5 ServiceEntries in the InfiniBand Architecture
* Specification. See also section B.7, table B.8 in the T10 SRP r16a document.
*/
struct ib_dm_svc_entries {
struct ib_dm_svc_entry service_entries[4];
};
#endif

4032
drivers/infiniband/ulp/srpt/ib_srpt.c Normal file
View File

File diff suppressed because it is too large Load Diff

443
drivers/infiniband/ulp/srpt/ib_srpt.h Normal file
View File

@@ -0,0 +1,443 @@
/*
* Copyright (c) 2006 - 2009 Mellanox Technology Inc. All rights reserved.
* Copyright (C) 2009 - 2010 Bart Van Assche <bvanassche@acm.org>.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*
*/
#ifndef IB_SRPT_H
#define IB_SRPT_H
#include <linux/types.h>
#include <linux/list.h>
#include <linux/wait.h>
#include <rdma/ib_verbs.h>
#include <rdma/ib_sa.h>
#include <rdma/ib_cm.h>
#include <scsi/srp.h>
#include "ib_dm_mad.h"
/*
* The prefix the ServiceName field must start with in the device management
* ServiceEntries attribute pair. See also the SRP specification.
*/
#define SRP_SERVICE_NAME_PREFIX "SRP.T10:"
enum {
/*
* SRP IOControllerProfile attributes for SRP target ports that have
* not been defined in <scsi/srp.h>. Source: section B.7, table B.7
* in the SRP specification.
*/
SRP_PROTOCOL = 0x0108,
SRP_PROTOCOL_VERSION = 0x0001,
SRP_IO_SUBCLASS = 0x609e,
SRP_SEND_TO_IOC = 0x01,
SRP_SEND_FROM_IOC = 0x02,
SRP_RDMA_READ_FROM_IOC = 0x08,
SRP_RDMA_WRITE_FROM_IOC = 0x20,
/*
* srp_login_cmd.req_flags bitmasks. See also table 9 in the SRP
* specification.
*/
SRP_MTCH_ACTION = 0x03, /* MULTI-CHANNEL ACTION */
SRP_LOSOLNT = 0x10, /* logout solicited notification */
SRP_CRSOLNT = 0x20, /* credit request solicited notification */
SRP_AESOLNT = 0x40, /* asynchronous event solicited notification */
/*
* srp_cmd.sol_nt / srp_tsk_mgmt.sol_not bitmasks. See also tables
* 18 and 20 in the SRP specification.
*/
SRP_SCSOLNT = 0x02, /* SCSOLNT = successful solicited notification */
SRP_UCSOLNT = 0x04, /* UCSOLNT = unsuccessful solicited notification */
/*
* srp_rsp.sol_not / srp_t_logout.sol_not bitmasks. See also tables
* 16 and 22 in the SRP specification.
*/
SRP_SOLNT = 0x01, /* SOLNT = solicited notification */
/* See also table 24 in the SRP specification. */
SRP_TSK_MGMT_SUCCESS = 0x00,
SRP_TSK_MGMT_FUNC_NOT_SUPP = 0x04,
SRP_TSK_MGMT_FAILED = 0x05,
/* See also table 21 in the SRP specification. */
SRP_CMD_SIMPLE_Q = 0x0,
SRP_CMD_HEAD_OF_Q = 0x1,
SRP_CMD_ORDERED_Q = 0x2,
SRP_CMD_ACA = 0x4,
SRP_LOGIN_RSP_MULTICHAN_NO_CHAN = 0x0,
SRP_LOGIN_RSP_MULTICHAN_TERMINATED = 0x1,
SRP_LOGIN_RSP_MULTICHAN_MAINTAINED = 0x2,
SRPT_DEF_SG_TABLESIZE = 128,
SRPT_DEF_SG_PER_WQE = 16,
MIN_SRPT_SQ_SIZE = 16,
DEF_SRPT_SQ_SIZE = 4096,
SRPT_RQ_SIZE = 128,
MIN_SRPT_SRQ_SIZE = 4,
DEFAULT_SRPT_SRQ_SIZE = 4095,
MAX_SRPT_SRQ_SIZE = 65535,
MAX_SRPT_RDMA_SIZE = 1U << 24,
MAX_SRPT_RSP_SIZE = 1024,
MIN_MAX_REQ_SIZE = 996,
DEFAULT_MAX_REQ_SIZE
= sizeof(struct srp_cmd)/*48*/
+ sizeof(struct srp_indirect_buf)/*20*/
+ 128 * sizeof(struct srp_direct_buf)/*16*/,
MIN_MAX_RSP_SIZE = sizeof(struct srp_rsp)/*36*/ + 4,
DEFAULT_MAX_RSP_SIZE = 256, /* leaves 220 bytes for sense data */
DEFAULT_MAX_RDMA_SIZE = 65536,
};
enum srpt_opcode {
SRPT_RECV,
SRPT_SEND,
SRPT_RDMA_MID,
SRPT_RDMA_ABORT,
SRPT_RDMA_READ_LAST,
SRPT_RDMA_WRITE_LAST,
};
static inline u64 encode_wr_id(u8 opcode, u32 idx)
{
return ((u64)opcode << 32) | idx;
}
static inline enum srpt_opcode opcode_from_wr_id(u64 wr_id)
{
return wr_id >> 32;
}
static inline u32 idx_from_wr_id(u64 wr_id)
{
return (u32)wr_id;
}
struct rdma_iu {
u64 raddr;
u32 rkey;
struct ib_sge *sge;
u32 sge_cnt;
int mem_id;
};
/**
* enum srpt_command_state - SCSI command state managed by SRPT.
* @SRPT_STATE_NEW: New command arrived and is being processed.
* @SRPT_STATE_NEED_DATA: Processing a write or bidir command and waiting
* for data arrival.
* @SRPT_STATE_DATA_IN: Data for the write or bidir command arrived and is
* being processed.
* @SRPT_STATE_CMD_RSP_SENT: SRP_RSP for SRP_CMD has been sent.
* @SRPT_STATE_MGMT: Processing a SCSI task management command.
* @SRPT_STATE_MGMT_RSP_SENT: SRP_RSP for SRP_TSK_MGMT has been sent.
* @SRPT_STATE_DONE: Command processing finished successfully, command
* processing has been aborted or command processing
* failed.
*/
enum srpt_command_state {
SRPT_STATE_NEW = 0,
SRPT_STATE_NEED_DATA = 1,
SRPT_STATE_DATA_IN = 2,
SRPT_STATE_CMD_RSP_SENT = 3,
SRPT_STATE_MGMT = 4,
SRPT_STATE_MGMT_RSP_SENT = 5,
SRPT_STATE_DONE = 6,
};
/**
* struct srpt_ioctx - Shared SRPT I/O context information.
* @buf: Pointer to the buffer.
* @dma: DMA address of the buffer.
* @index: Index of the I/O context in its ioctx_ring array.
*/
struct srpt_ioctx {
void *buf;
dma_addr_t dma;
uint32_t index;
};
/**
* struct srpt_recv_ioctx - SRPT receive I/O context.
* @ioctx: See above.
* @wait_list: Node for insertion in srpt_rdma_ch.cmd_wait_list.
*/
struct srpt_recv_ioctx {
struct srpt_ioctx ioctx;
struct list_head wait_list;
};
/**
* struct srpt_send_ioctx - SRPT send I/O context.
* @ioctx: See above.
* @ch: Channel pointer.
* @free_list: Node in srpt_rdma_ch.free_list.
* @n_rbuf: Number of data buffers in the received SRP command.
* @rbufs: Pointer to SRP data buffer array.
* @single_rbuf: SRP data buffer if the command has only a single buffer.
* @sg: Pointer to sg-list associated with this I/O context.
* @sg_cnt: SG-list size.
* @mapped_sg_count: ib_dma_map_sg() return value.
* @n_rdma_ius: Number of elements in the rdma_ius array.
* @rdma_ius: Array with information about the RDMA mapping.
* @tag: Tag of the received SRP information unit.
* @spinlock: Protects 'state'.
* @state: I/O context state.
* @rdma_aborted: If initiating a multipart RDMA transfer failed, whether
* the already initiated transfers have finished.
* @cmd: Target core command data structure.
* @sense_data: SCSI sense data.
*/
struct srpt_send_ioctx {
struct srpt_ioctx ioctx;
struct srpt_rdma_ch *ch;
struct rdma_iu *rdma_ius;
struct srp_direct_buf *rbufs;
struct srp_direct_buf single_rbuf;
struct scatterlist *sg;
struct list_head free_list;
spinlock_t spinlock;
enum srpt_command_state state;
bool rdma_aborted;
struct se_cmd cmd;
struct completion tx_done;
u64 tag;
int sg_cnt;
int mapped_sg_count;
u16 n_rdma_ius;
u8 n_rdma;
u8 n_rbuf;
bool queue_status_only;
u8 sense_data[SCSI_SENSE_BUFFERSIZE];
};
/**
* enum rdma_ch_state - SRP channel state.
* @CH_CONNECTING: QP is in RTR state; waiting for RTU.
* @CH_LIVE: QP is in RTS state.
* @CH_DISCONNECTING: DREQ has been received; waiting for DREP
* or DREQ has been send and waiting for DREP
* or .
* @CH_DRAINING: QP is in ERR state; waiting for last WQE event.
* @CH_RELEASING: Last WQE event has been received; releasing resources.
*/
enum rdma_ch_state {
CH_CONNECTING,
CH_LIVE,
CH_DISCONNECTING,
CH_DRAINING,
CH_RELEASING
};
/**
* struct srpt_rdma_ch - RDMA channel.
* @wait_queue: Allows the kernel thread to wait for more work.
* @thread: Kernel thread that processes the IB queues associated with
* the channel.
* @cm_id: IB CM ID associated with the channel.
* @qp: IB queue pair used for communicating over this channel.
* @cq: IB completion queue for this channel.
* @rq_size: IB receive queue size.
* @rsp_size IB response message size in bytes.
* @sq_wr_avail: number of work requests available in the send queue.
* @sport: pointer to the information of the HCA port used by this
* channel.
* @i_port_id: 128-bit initiator port identifier copied from SRP_LOGIN_REQ.
* @t_port_id: 128-bit target port identifier copied from SRP_LOGIN_REQ.
* @max_ti_iu_len: maximum target-to-initiator information unit length.
* @req_lim: request limit: maximum number of requests that may be sent
* by the initiator without having received a response.
* @req_lim_delta: Number of credits not yet sent back to the initiator.
* @spinlock: Protects free_list and state.
* @free_list: Head of list with free send I/O contexts.
* @state: channel state. See also enum rdma_ch_state.
* @ioctx_ring: Send ring.
* @wc: IB work completion array for srpt_process_completion().
* @list: Node for insertion in the srpt_device.rch_list list.
* @cmd_wait_list: List of SCSI commands that arrived before the RTU event. This
* list contains struct srpt_ioctx elements and is protected
* against concurrent modification by the cm_id spinlock.
* @sess: Session information associated with this SRP channel.
* @sess_name: Session name.
* @release_work: Allows scheduling of srpt_release_channel().
* @release_done: Enables waiting for srpt_release_channel() completion.
*/
struct srpt_rdma_ch {
wait_queue_head_t wait_queue;
struct task_struct *thread;
struct ib_cm_id *cm_id;
struct ib_qp *qp;
struct ib_cq *cq;
int rq_size;
u32 rsp_size;
atomic_t sq_wr_avail;
struct srpt_port *sport;
u8 i_port_id[16];
u8 t_port_id[16];
int max_ti_iu_len;
atomic_t req_lim;
atomic_t req_lim_delta;
spinlock_t spinlock;
struct list_head free_list;
enum rdma_ch_state state;
struct srpt_send_ioctx **ioctx_ring;
struct ib_wc wc[16];
struct list_head list;
struct list_head cmd_wait_list;
struct se_session *sess;
u8 sess_name[36];
struct work_struct release_work;
struct completion *release_done;
bool in_shutdown;
};
/**
* struct srpt_port_attib - Attributes for SRPT port
* @srp_max_rdma_size: Maximum size of SRP RDMA transfers for new connections.
* @srp_max_rsp_size: Maximum size of SRP response messages in bytes.
* @srp_sq_size: Shared receive queue (SRQ) size.
*/
struct srpt_port_attrib {
u32 srp_max_rdma_size;
u32 srp_max_rsp_size;
u32 srp_sq_size;
};
/**
* struct srpt_port - Information associated by SRPT with a single IB port.
* @sdev: backpointer to the HCA information.
* @mad_agent: per-port management datagram processing information.
* @enabled: Whether or not this target port is enabled.
* @port_guid: ASCII representation of Port GUID
* @port: one-based port number.
* @sm_lid: cached value of the port's sm_lid.
* @lid: cached value of the port's lid.
* @gid: cached value of the port's gid.
* @port_acl_lock spinlock for port_acl_list:
* @work: work structure for refreshing the aforementioned cached values.
* @port_tpg_1 Target portal group = 1 data.
* @port_wwn: Target core WWN data.
* @port_acl_list: Head of the list with all node ACLs for this port.
*/
struct srpt_port {
struct srpt_device *sdev;
struct ib_mad_agent *mad_agent;
bool enabled;
u8 port_guid[64];
u8 port;
u16 sm_lid;
u16 lid;
union ib_gid gid;
spinlock_t port_acl_lock;
struct work_struct work;
struct se_portal_group port_tpg_1;
struct se_wwn port_wwn;
struct list_head port_acl_list;
struct srpt_port_attrib port_attrib;
};
/**
* struct srpt_device - Information associated by SRPT with a single HCA.
* @device: Backpointer to the struct ib_device managed by the IB core.
* @pd: IB protection domain.
* @mr: L_Key (local key) with write access to all local memory.
* @srq: Per-HCA SRQ (shared receive queue).
* @cm_id: Connection identifier.
* @dev_attr: Attributes of the InfiniBand device as obtained during the
* ib_client.add() callback.
* @srq_size: SRQ size.
* @ioctx_ring: Per-HCA SRQ.
* @rch_list: Per-device channel list -- see also srpt_rdma_ch.list.
* @ch_releaseQ: Enables waiting for removal from rch_list.
* @spinlock: Protects rch_list and tpg.
* @port: Information about the ports owned by this HCA.
* @event_handler: Per-HCA asynchronous IB event handler.
* @list: Node in srpt_dev_list.
*/
struct srpt_device {
struct ib_device *device;
struct ib_pd *pd;
struct ib_mr *mr;
struct ib_srq *srq;
struct ib_cm_id *cm_id;
struct ib_device_attr dev_attr;
int srq_size;
struct srpt_recv_ioctx **ioctx_ring;
struct list_head rch_list;
wait_queue_head_t ch_releaseQ;
spinlock_t spinlock;
struct srpt_port port[2];
struct ib_event_handler event_handler;
struct list_head list;
};
/**
* struct srpt_node_acl - Per-initiator ACL data (managed via configfs).
* @i_port_id: 128-bit SRP initiator port ID.
* @sport: port information.
* @nacl: Target core node ACL information.
* @list: Element of the per-HCA ACL list.
*/
struct srpt_node_acl {
u8 i_port_id[16];
struct srpt_port *sport;
struct se_node_acl nacl;
struct list_head list;
};
/*
* SRP-releated SCSI persistent reservation definitions.
*
* See also SPC4r28, section 7.6.1 (Protocol specific parameters introduction).
* See also SPC4r28, section 7.6.4.5 (TransportID for initiator ports using
* SCSI over an RDMA interface).
*/
enum {
SCSI_TRANSPORTID_PROTOCOLID_SRP = 4,
};
struct spc_rdma_transport_id {
uint8_t protocol_identifier;
uint8_t reserved[7];
uint8_t i_port_id[16];
};
#endif /* IB_SRPT_H */