summaryrefslogtreecommitdiffstats
path: root/drivers/infiniband/core
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/infiniband/core')
-rw-r--r--drivers/infiniband/core/Makefile6
-rw-r--r--drivers/infiniband/core/addr.c25
-rw-r--r--drivers/infiniband/core/cache.c7
-rw-r--r--drivers/infiniband/core/cm.c144
-rw-r--r--drivers/infiniband/core/cma.c490
-rw-r--r--drivers/infiniband/core/fmr_pool.c12
-rw-r--r--drivers/infiniband/core/iwcm.c47
-rw-r--r--drivers/infiniband/core/mad.c117
-rw-r--r--drivers/infiniband/core/mad_priv.h8
-rw-r--r--drivers/infiniband/core/mad_rmpp.c18
-rw-r--r--drivers/infiniband/core/sa_query.c10
-rw-r--r--drivers/infiniband/core/ucm.c20
-rw-r--r--drivers/infiniband/core/ucma.c885
-rw-r--r--drivers/infiniband/core/uverbs_main.c6
-rw-r--r--drivers/infiniband/core/uverbs_marshall.c5
-rw-r--r--drivers/infiniband/core/uverbs_mem.c19
16 files changed, 1500 insertions, 319 deletions
diff --git a/drivers/infiniband/core/Makefile b/drivers/infiniband/core/Makefile
index 163d991..50fb1cd 100644
--- a/drivers/infiniband/core/Makefile
+++ b/drivers/infiniband/core/Makefile
@@ -1,9 +1,11 @@
infiniband-$(CONFIG_INFINIBAND_ADDR_TRANS) := ib_addr.o rdma_cm.o
+user_access-$(CONFIG_INFINIBAND_ADDR_TRANS) := rdma_ucm.o
obj-$(CONFIG_INFINIBAND) += ib_core.o ib_mad.o ib_sa.o \
ib_cm.o iw_cm.o $(infiniband-y)
obj-$(CONFIG_INFINIBAND_USER_MAD) += ib_umad.o
-obj-$(CONFIG_INFINIBAND_USER_ACCESS) += ib_uverbs.o ib_ucm.o
+obj-$(CONFIG_INFINIBAND_USER_ACCESS) += ib_uverbs.o ib_ucm.o \
+ $(user_access-y)
ib_core-y := packer.o ud_header.o verbs.o sysfs.o \
device.o fmr_pool.o cache.o
@@ -18,6 +20,8 @@ iw_cm-y := iwcm.o
rdma_cm-y := cma.o
+rdma_ucm-y := ucma.o
+
ib_addr-y := addr.o
ib_umad-y := user_mad.o
diff --git a/drivers/infiniband/core/addr.c b/drivers/infiniband/core/addr.c
index e11187e..af93979 100644
--- a/drivers/infiniband/core/addr.c
+++ b/drivers/infiniband/core/addr.c
@@ -55,11 +55,11 @@ struct addr_req {
int status;
};
-static void process_req(void *data);
+static void process_req(struct work_struct *work);
static DEFINE_MUTEX(lock);
static LIST_HEAD(req_list);
-static DECLARE_WORK(work, process_req, NULL);
+static DECLARE_DELAYED_WORK(work, process_req);
static struct workqueue_struct *addr_wq;
void rdma_addr_register_client(struct rdma_addr_client *client)
@@ -139,7 +139,7 @@ static void queue_req(struct addr_req *req)
mutex_lock(&lock);
list_for_each_entry_reverse(temp_req, &req_list, list) {
- if (time_after(req->timeout, temp_req->timeout))
+ if (time_after_eq(req->timeout, temp_req->timeout))
break;
}
@@ -215,7 +215,7 @@ out:
return ret;
}
-static void process_req(void *data)
+static void process_req(struct work_struct *work)
{
struct addr_req *req, *temp_req;
struct sockaddr_in *src_in, *dst_in;
@@ -225,19 +225,17 @@ static void process_req(void *data)
mutex_lock(&lock);
list_for_each_entry_safe(req, temp_req, &req_list, list) {
- if (req->status) {
+ if (req->status == -ENODATA) {
src_in = (struct sockaddr_in *) &req->src_addr;
dst_in = (struct sockaddr_in *) &req->dst_addr;
req->status = addr_resolve_remote(src_in, dst_in,
req->addr);
+ if (req->status && time_after_eq(jiffies, req->timeout))
+ req->status = -ETIMEDOUT;
+ else if (req->status == -ENODATA)
+ continue;
}
- if (req->status && time_after(jiffies, req->timeout))
- req->status = -ETIMEDOUT;
- else if (req->status == -ENODATA)
- continue;
-
- list_del(&req->list);
- list_add_tail(&req->list, &done_list);
+ list_move_tail(&req->list, &done_list);
}
if (!list_empty(&req_list)) {
@@ -347,8 +345,7 @@ void rdma_addr_cancel(struct rdma_dev_addr *addr)
if (req->addr == addr) {
req->status = -ECANCELED;
req->timeout = jiffies;
- list_del(&req->list);
- list_add(&req->list, &req_list);
+ list_move(&req->list, &req_list);
set_timeout(req->timeout);
break;
}
diff --git a/drivers/infiniband/core/cache.c b/drivers/infiniband/core/cache.c
index 20e9f64..98272fb 100644
--- a/drivers/infiniband/core/cache.c
+++ b/drivers/infiniband/core/cache.c
@@ -285,9 +285,10 @@ err:
kfree(tprops);
}
-static void ib_cache_task(void *work_ptr)
+static void ib_cache_task(struct work_struct *_work)
{
- struct ib_update_work *work = work_ptr;
+ struct ib_update_work *work =
+ container_of(_work, struct ib_update_work, work);
ib_cache_update(work->device, work->port_num);
kfree(work);
@@ -306,7 +307,7 @@ static void ib_cache_event(struct ib_event_handler *handler,
event->event == IB_EVENT_CLIENT_REREGISTER) {
work = kmalloc(sizeof *work, GFP_ATOMIC);
if (work) {
- INIT_WORK(&work->work, ib_cache_task, work);
+ INIT_WORK(&work->work, ib_cache_task);
work->device = event->device;
work->port_num = event->element.port_num;
schedule_work(&work->work);
diff --git a/drivers/infiniband/core/cm.c b/drivers/infiniband/core/cm.c
index 25b1018..d446998 100644
--- a/drivers/infiniband/core/cm.c
+++ b/drivers/infiniband/core/cm.c
@@ -101,7 +101,7 @@ struct cm_av {
};
struct cm_work {
- struct work_struct work;
+ struct delayed_work work;
struct list_head list;
struct cm_port *port;
struct ib_mad_recv_wc *mad_recv_wc; /* Received MADs */
@@ -147,12 +147,12 @@ struct cm_id_private {
__be32 rq_psn;
int timeout_ms;
enum ib_mtu path_mtu;
+ __be16 pkey;
u8 private_data_len;
u8 max_cm_retries;
u8 peer_to_peer;
u8 responder_resources;
u8 initiator_depth;
- u8 local_ack_timeout;
u8 retry_count;
u8 rnr_retry_count;
u8 service_timeout;
@@ -161,7 +161,7 @@ struct cm_id_private {
atomic_t work_count;
};
-static void cm_work_handler(void *data);
+static void cm_work_handler(struct work_struct *work);
static inline void cm_deref_id(struct cm_id_private *cm_id_priv)
{
@@ -240,11 +240,10 @@ static void * cm_copy_private_data(const void *private_data,
if (!private_data || !private_data_len)
return NULL;
- data = kmalloc(private_data_len, GFP_KERNEL);
+ data = kmemdup(private_data, private_data_len, GFP_KERNEL);
if (!data)
return ERR_PTR(-ENOMEM);
- memcpy(data, private_data, private_data_len);
return data;
}
@@ -669,8 +668,7 @@ static struct cm_timewait_info * cm_create_timewait_info(__be32 local_id)
return ERR_PTR(-ENOMEM);
timewait_info->work.local_id = local_id;
- INIT_WORK(&timewait_info->work.work, cm_work_handler,
- &timewait_info->work);
+ INIT_DELAYED_WORK(&timewait_info->work.work, cm_work_handler);
timewait_info->work.cm_event.event = IB_CM_TIMEWAIT_EXIT;
return timewait_info;
}
@@ -691,7 +689,7 @@ static void cm_enter_timewait(struct cm_id_private *cm_id_priv)
* timewait before notifying the user that we've exited timewait.
*/
cm_id_priv->id.state = IB_CM_TIMEWAIT;
- wait_time = cm_convert_to_ms(cm_id_priv->local_ack_timeout);
+ wait_time = cm_convert_to_ms(cm_id_priv->av.packet_life_time + 1);
queue_delayed_work(cm.wq, &cm_id_priv->timewait_info->work.work,
msecs_to_jiffies(wait_time));
cm_id_priv->timewait_info = NULL;
@@ -1010,6 +1008,7 @@ int ib_send_cm_req(struct ib_cm_id *cm_id,
cm_id_priv->responder_resources = param->responder_resources;
cm_id_priv->retry_count = param->retry_count;
cm_id_priv->path_mtu = param->primary_path->mtu;
+ cm_id_priv->pkey = param->primary_path->pkey;
cm_id_priv->qp_type = param->qp_type;
ret = cm_alloc_msg(cm_id_priv, &cm_id_priv->msg);
@@ -1024,8 +1023,6 @@ int ib_send_cm_req(struct ib_cm_id *cm_id,
cm_id_priv->local_qpn = cm_req_get_local_qpn(req_msg);
cm_id_priv->rq_psn = cm_req_get_starting_psn(req_msg);
- cm_id_priv->local_ack_timeout =
- cm_req_get_primary_local_ack_timeout(req_msg);
spin_lock_irqsave(&cm_id_priv->lock, flags);
ret = ib_post_send_mad(cm_id_priv->msg, NULL);
@@ -1410,9 +1407,8 @@ static int cm_req_handler(struct cm_work *work)
cm_id_priv->initiator_depth = cm_req_get_resp_res(req_msg);
cm_id_priv->responder_resources = cm_req_get_init_depth(req_msg);
cm_id_priv->path_mtu = cm_req_get_path_mtu(req_msg);
+ cm_id_priv->pkey = req_msg->pkey;
cm_id_priv->sq_psn = cm_req_get_starting_psn(req_msg);
- cm_id_priv->local_ack_timeout =
- cm_req_get_primary_local_ack_timeout(req_msg);
cm_id_priv->retry_count = cm_req_get_retry_count(req_msg);
cm_id_priv->rnr_retry_count = cm_req_get_rnr_retry_count(req_msg);
cm_id_priv->qp_type = cm_req_get_qp_type(req_msg);
@@ -1716,7 +1712,7 @@ static int cm_establish_handler(struct cm_work *work)
unsigned long flags;
int ret;
- /* See comment in ib_cm_establish about lookup. */
+ /* See comment in cm_establish about lookup. */
cm_id_priv = cm_acquire_id(work->local_id, work->remote_id);
if (!cm_id_priv)
return -EINVAL;
@@ -2402,11 +2398,16 @@ int ib_send_cm_lap(struct ib_cm_id *cm_id,
cm_id_priv = container_of(cm_id, struct cm_id_private, id);
spin_lock_irqsave(&cm_id_priv->lock, flags);
if (cm_id->state != IB_CM_ESTABLISHED ||
- cm_id->lap_state != IB_CM_LAP_IDLE) {
+ (cm_id->lap_state != IB_CM_LAP_UNINIT &&
+ cm_id->lap_state != IB_CM_LAP_IDLE)) {
ret = -EINVAL;
goto out;
}
+ ret = cm_init_av_by_path(alternate_path, &cm_id_priv->alt_av);
+ if (ret)
+ goto out;
+
ret = cm_alloc_msg(cm_id_priv, &msg);
if (ret)
goto out;
@@ -2431,7 +2432,8 @@ out: spin_unlock_irqrestore(&cm_id_priv->lock, flags);
}
EXPORT_SYMBOL(ib_send_cm_lap);
-static void cm_format_path_from_lap(struct ib_sa_path_rec *path,
+static void cm_format_path_from_lap(struct cm_id_private *cm_id_priv,
+ struct ib_sa_path_rec *path,
struct cm_lap_msg *lap_msg)
{
memset(path, 0, sizeof *path);
@@ -2443,10 +2445,10 @@ static void cm_format_path_from_lap(struct ib_sa_path_rec *path,
path->hop_limit = lap_msg->alt_hop_limit;
path->traffic_class = cm_lap_get_traffic_class(lap_msg);
path->reversible = 1;
- /* pkey is same as in REQ */
+ path->pkey = cm_id_priv->pkey;
path->sl = cm_lap_get_sl(lap_msg);
path->mtu_selector = IB_SA_EQ;
- /* mtu is same as in REQ */
+ path->mtu = cm_id_priv->path_mtu;
path->rate_selector = IB_SA_EQ;
path->rate = cm_lap_get_packet_rate(lap_msg);
path->packet_life_time_selector = IB_SA_EQ;
@@ -2472,7 +2474,7 @@ static int cm_lap_handler(struct cm_work *work)
param = &work->cm_event.param.lap_rcvd;
param->alternate_path = &work->path[0];
- cm_format_path_from_lap(param->alternate_path, lap_msg);
+ cm_format_path_from_lap(cm_id_priv, param->alternate_path, lap_msg);
work->cm_event.private_data = &lap_msg->private_data;
spin_lock_irqsave(&cm_id_priv->lock, flags);
@@ -2480,6 +2482,7 @@ static int cm_lap_handler(struct cm_work *work)
goto unlock;
switch (cm_id_priv->id.lap_state) {
+ case IB_CM_LAP_UNINIT:
case IB_CM_LAP_IDLE:
break;
case IB_CM_MRA_LAP_SENT:
@@ -2502,6 +2505,10 @@ static int cm_lap_handler(struct cm_work *work)
cm_id_priv->id.lap_state = IB_CM_LAP_RCVD;
cm_id_priv->tid = lap_msg->hdr.tid;
+ cm_init_av_for_response(work->port, work->mad_recv_wc->wc,
+ work->mad_recv_wc->recv_buf.grh,
+ &cm_id_priv->av);
+ cm_init_av_by_path(param->alternate_path, &cm_id_priv->alt_av);
ret = atomic_inc_and_test(&cm_id_priv->work_count);
if (!ret)
list_add_tail(&work->list, &cm_id_priv->work_list);
@@ -2987,9 +2994,9 @@ static void cm_send_handler(struct ib_mad_agent *mad_agent,
}
}
-static void cm_work_handler(void *data)
+static void cm_work_handler(struct work_struct *_work)
{
- struct cm_work *work = data;
+ struct cm_work *work = container_of(_work, struct cm_work, work.work);
int ret;
switch (work->cm_event.event) {
@@ -3040,7 +3047,7 @@ static void cm_work_handler(void *data)
cm_free_work(work);
}
-int ib_cm_establish(struct ib_cm_id *cm_id)
+static int cm_establish(struct ib_cm_id *cm_id)
{
struct cm_id_private *cm_id_priv;
struct cm_work *work;
@@ -3079,16 +3086,53 @@ int ib_cm_establish(struct ib_cm_id *cm_id)
* we need to find the cm_id once we're in the context of the
* worker thread, rather than holding a reference on it.
*/
- INIT_WORK(&work->work, cm_work_handler, work);
+ INIT_DELAYED_WORK(&work->work, cm_work_handler);
work->local_id = cm_id->local_id;
work->remote_id = cm_id->remote_id;
work->mad_recv_wc = NULL;
work->cm_event.event = IB_CM_USER_ESTABLISHED;
- queue_work(cm.wq, &work->work);
+ queue_delayed_work(cm.wq, &work->work, 0);
out:
return ret;
}
-EXPORT_SYMBOL(ib_cm_establish);
+
+static int cm_migrate(struct ib_cm_id *cm_id)
+{
+ struct cm_id_private *cm_id_priv;
+ unsigned long flags;
+ int ret = 0;
+
+ cm_id_priv = container_of(cm_id, struct cm_id_private, id);
+ spin_lock_irqsave(&cm_id_priv->lock, flags);
+ if (cm_id->state == IB_CM_ESTABLISHED &&
+ (cm_id->lap_state == IB_CM_LAP_UNINIT ||
+ cm_id->lap_state == IB_CM_LAP_IDLE)) {
+ cm_id->lap_state = IB_CM_LAP_IDLE;
+ cm_id_priv->av = cm_id_priv->alt_av;
+ } else
+ ret = -EINVAL;
+ spin_unlock_irqrestore(&cm_id_priv->lock, flags);
+
+ return ret;
+}
+
+int ib_cm_notify(struct ib_cm_id *cm_id, enum ib_event_type event)
+{
+ int ret;
+
+ switch (event) {
+ case IB_EVENT_COMM_EST:
+ ret = cm_establish(cm_id);
+ break;
+ case IB_EVENT_PATH_MIG:
+ ret = cm_migrate(cm_id);
+ break;
+ default:
+ ret = -EINVAL;
+ }
+ return ret;
+}
+EXPORT_SYMBOL(ib_cm_notify);
static void cm_recv_handler(struct ib_mad_agent *mad_agent,
struct ib_mad_recv_wc *mad_recv_wc)
@@ -3146,11 +3190,11 @@ static void cm_recv_handler(struct ib_mad_agent *mad_agent,
return;
}
- INIT_WORK(&work->work, cm_work_handler, work);
+ INIT_DELAYED_WORK(&work->work, cm_work_handler);
work->cm_event.event = event;
work->mad_recv_wc = mad_recv_wc;
work->port = (struct cm_port *)mad_agent->context;
- queue_work(cm.wq, &work->work);
+ queue_delayed_work(cm.wq, &work->work, 0);
}
static int cm_init_qp_init_attr(struct cm_id_private *cm_id_priv,
@@ -3173,8 +3217,7 @@ static int cm_init_qp_init_attr(struct cm_id_private *cm_id_priv,
case IB_CM_ESTABLISHED:
*qp_attr_mask = IB_QP_STATE | IB_QP_ACCESS_FLAGS |
IB_QP_PKEY_INDEX | IB_QP_PORT;
- qp_attr->qp_access_flags = IB_ACCESS_LOCAL_WRITE |
- IB_ACCESS_REMOTE_WRITE;
+ qp_attr->qp_access_flags = IB_ACCESS_REMOTE_WRITE;
if (cm_id_priv->responder_resources)
qp_attr->qp_access_flags |= IB_ACCESS_REMOTE_READ |
IB_ACCESS_REMOTE_ATOMIC;
@@ -3222,6 +3265,9 @@ static int cm_init_qp_rtr_attr(struct cm_id_private *cm_id_priv,
if (cm_id_priv->alt_av.ah_attr.dlid) {
*qp_attr_mask |= IB_QP_ALT_PATH;
qp_attr->alt_port_num = cm_id_priv->alt_av.port->port_num;
+ qp_attr->alt_pkey_index = cm_id_priv->alt_av.pkey_index;
+ qp_attr->alt_timeout =
+ cm_id_priv->alt_av.packet_life_time + 1;
qp_attr->alt_ah_attr = cm_id_priv->alt_av.ah_attr;
}
ret = 0;
@@ -3243,24 +3289,40 @@ static int cm_init_qp_rts_attr(struct cm_id_private *cm_id_priv,
spin_lock_irqsave(&cm_id_priv->lock, flags);
switch (cm_id_priv->id.state) {
+ /* Allow transition to RTS before sending REP */
+ case IB_CM_REQ_RCVD:
+ case IB_CM_MRA_REQ_SENT:
+
case IB_CM_REP_RCVD:
case IB_CM_MRA_REP_SENT:
case IB_CM_REP_SENT:
case IB_CM_MRA_REP_RCVD:
case IB_CM_ESTABLISHED:
- *qp_attr_mask = IB_QP_STATE | IB_QP_SQ_PSN;
- qp_attr->sq_psn = be32_to_cpu(cm_id_priv->sq_psn);
- if (cm_id_priv->qp_type == IB_QPT_RC) {
- *qp_attr_mask |= IB_QP_TIMEOUT | IB_QP_RETRY_CNT |
- IB_QP_RNR_RETRY |
- IB_QP_MAX_QP_RD_ATOMIC;
- qp_attr->timeout = cm_id_priv->local_ack_timeout;
- qp_attr->retry_cnt = cm_id_priv->retry_count;
- qp_attr->rnr_retry = cm_id_priv->rnr_retry_count;
- qp_attr->max_rd_atomic = cm_id_priv->initiator_depth;
- }
- if (cm_id_priv->alt_av.ah_attr.dlid) {
- *qp_attr_mask |= IB_QP_PATH_MIG_STATE;
+ if (cm_id_priv->id.lap_state == IB_CM_LAP_UNINIT) {
+ *qp_attr_mask = IB_QP_STATE | IB_QP_SQ_PSN;
+ qp_attr->sq_psn = be32_to_cpu(cm_id_priv->sq_psn);
+ if (cm_id_priv->qp_type == IB_QPT_RC) {
+ *qp_attr_mask |= IB_QP_TIMEOUT | IB_QP_RETRY_CNT |
+ IB_QP_RNR_RETRY |
+ IB_QP_MAX_QP_RD_ATOMIC;
+ qp_attr->timeout =
+ cm_id_priv->av.packet_life_time + 1;
+ qp_attr->retry_cnt = cm_id_priv->retry_count;
+ qp_attr->rnr_retry = cm_id_priv->rnr_retry_count;
+ qp_attr->max_rd_atomic =
+ cm_id_priv->initiator_depth;
+ }
+ if (cm_id_priv->alt_av.ah_attr.dlid) {
+ *qp_attr_mask |= IB_QP_PATH_MIG_STATE;
+ qp_attr->path_mig_state = IB_MIG_REARM;
+ }
+ } else {
+ *qp_attr_mask = IB_QP_ALT_PATH | IB_QP_PATH_MIG_STATE;
+ qp_attr->alt_port_num = cm_id_priv->alt_av.port->port_num;
+ qp_attr->alt_pkey_index = cm_id_priv->alt_av.pkey_index;
+ qp_attr->alt_timeout =
+ cm_id_priv->alt_av.packet_life_time + 1;
+ qp_attr->alt_ah_attr = cm_id_priv->alt_av.ah_attr;
qp_attr->path_mig_state = IB_MIG_REARM;
}
ret = 0;
diff --git a/drivers/infiniband/core/cma.c b/drivers/infiniband/core/cma.c
index 845090b..9e0ab04 100644
--- a/drivers/infiniband/core/cma.c
+++ b/drivers/infiniband/core/cma.c
@@ -70,6 +70,7 @@ static DEFINE_MUTEX(lock);
static struct workqueue_struct *cma_wq;
static DEFINE_IDR(sdp_ps);
static DEFINE_IDR(tcp_ps);
+static DEFINE_IDR(udp_ps);
struct cma_device {
struct list_head list;
@@ -133,7 +134,6 @@ struct rdma_id_private {
u32 seq_num;
u32 qp_num;
- enum ib_qp_type qp_type;
u8 srq;
};
@@ -344,7 +344,7 @@ static int cma_init_ib_qp(struct rdma_id_private *id_priv, struct ib_qp *qp)
return ret;
qp_attr.qp_state = IB_QPS_INIT;
- qp_attr.qp_access_flags = IB_ACCESS_LOCAL_WRITE;
+ qp_attr.qp_access_flags = 0;
qp_attr.port_num = id_priv->id.port_num;
return ib_modify_qp(qp, &qp_attr, IB_QP_STATE | IB_QP_ACCESS_FLAGS |
IB_QP_PKEY_INDEX | IB_QP_PORT);
@@ -392,7 +392,6 @@ int rdma_create_qp(struct rdma_cm_id *id, struct ib_pd *pd,
id->qp = qp;
id_priv->qp_num = qp->qp_num;
- id_priv->qp_type = qp->qp_type;
id_priv->srq = (qp->srq != NULL);
return 0;
err:
@@ -510,9 +509,17 @@ static inline int cma_any_addr(struct sockaddr *addr)
return cma_zero_addr(addr) || cma_loopback_addr(addr);
}
+static inline __be16 cma_port(struct sockaddr *addr)
+{
+ if (addr->sa_family == AF_INET)
+ return ((struct sockaddr_in *) addr)->sin_port;
+ else
+ return ((struct sockaddr_in6 *) addr)->sin6_port;
+}
+
static inline int cma_any_port(struct sockaddr *addr)
{
- return !((struct sockaddr_in *) addr)->sin_port;
+ return !cma_port(addr);
}
static int cma_get_net_info(void *hdr, enum rdma_port_space ps,
@@ -594,20 +601,6 @@ static inline int cma_user_data_offset(enum rdma_port_space ps)
}
}
-static int cma_notify_user(struct rdma_id_private *id_priv,
- enum rdma_cm_event_type type, int status,
- void *data, u8 data_len)
-{
- struct rdma_cm_event event;
-
- event.event = type;
- event.status = status;
- event.private_data = data;
- event.private_data_len = data_len;
-
- return id_priv->id.event_handler(&id_priv->id, &event);
-}
-
static void cma_cancel_route(struct rdma_id_private *id_priv)
{
switch (rdma_node_get_transport(id_priv->id.device->node_type)) {
@@ -776,63 +769,61 @@ static int cma_verify_rep(struct rdma_id_private *id_priv, void *data)
return 0;
}
-static int cma_rtu_recv(struct rdma_id_private *id_priv)
+static void cma_set_rep_event_data(struct rdma_cm_event *event,
+ struct ib_cm_rep_event_param *rep_data,
+ void *private_data)
{
- int ret;
-
- ret = cma_modify_qp_rts(&id_priv->id);
- if (ret)
- goto reject;
-
- return 0;
-reject:
- cma_modify_qp_err(&id_priv->id);
- ib_send_cm_rej(id_priv->cm_id.ib, IB_CM_REJ_CONSUMER_DEFINED,
- NULL, 0, NULL, 0);
- return ret;
+ event->param.conn.private_data = private_data;
+ event->param.conn.private_data_len = IB_CM_REP_PRIVATE_DATA_SIZE;
+ event->param.conn.responder_resources = rep_data->responder_resources;
+ event->param.conn.initiator_depth = rep_data->initiator_depth;
+ event->param.conn.flow_control = rep_data->flow_control;
+ event->param.conn.rnr_retry_count = rep_data->rnr_retry_count;
+ event->param.conn.srq = rep_data->srq;
+ event->param.conn.qp_num = rep_data->remote_qpn;
}
static int cma_ib_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event)
{
struct rdma_id_private *id_priv = cm_id->context;
- enum rdma_cm_event_type event;
- u8 private_data_len = 0;
- int ret = 0, status = 0;
+ struct rdma_cm_event event;
+ int ret = 0;
atomic_inc(&id_priv->dev_remove);
if (!cma_comp(id_priv, CMA_CONNECT))
goto out;
+ memset(&event, 0, sizeof event);
switch (ib_event->event) {
case IB_CM_REQ_ERROR:
case IB_CM_REP_ERROR:
- event = RDMA_CM_EVENT_UNREACHABLE;
- status = -ETIMEDOUT;
+ event.event = RDMA_CM_EVENT_UNREACHABLE;
+ event.status = -ETIMEDOUT;
break;
case IB_CM_REP_RECEIVED:
- status = cma_verify_rep(id_priv, ib_event->private_data);
- if (status)
- event = RDMA_CM_EVENT_CONNECT_ERROR;
+ event.status = cma_verify_rep(id_priv, ib_event->private_data);
+ if (event.status)
+ event.event = RDMA_CM_EVENT_CONNECT_ERROR;
else if (id_priv->id.qp && id_priv->id.ps != RDMA_PS_SDP) {
- status = cma_rep_recv(id_priv);
- event = status ? RDMA_CM_EVENT_CONNECT_ERROR :
- RDMA_CM_EVENT_ESTABLISHED;
+ event.status = cma_rep_recv(id_priv);
+ event.event = event.status ? RDMA_CM_EVENT_CONNECT_ERROR :
+ RDMA_CM_EVENT_ESTABLISHED;
} else
- event = RDMA_CM_EVENT_CONNECT_RESPONSE;
- private_data_len = IB_CM_REP_PRIVATE_DATA_SIZE;
+ event.event = RDMA_CM_EVENT_CONNECT_RESPONSE;
+ cma_set_rep_event_data(&event, &ib_event->param.rep_rcvd,
+ ib_event->private_data);
break;
case IB_CM_RTU_RECEIVED:
- status = cma_rtu_recv(id_priv);
- event = status ? RDMA_CM_EVENT_CONNECT_ERROR :
- RDMA_CM_EVENT_ESTABLISHED;
+ case IB_CM_USER_ESTABLISHED:
+ event.event = RDMA_CM_EVENT_ESTABLISHED;
break;
case IB_CM_DREQ_ERROR:
- status = -ETIMEDOUT; /* fall through */
+ event.status = -ETIMEDOUT; /* fall through */
case IB_CM_DREQ_RECEIVED:
case IB_CM_DREP_RECEIVED:
if (!cma_comp_exch(id_priv, CMA_CONNECT, CMA_DISCONNECT))
goto out;
- event = RDMA_CM_EVENT_DISCONNECTED;
+ event.event = RDMA_CM_EVENT_DISCONNECTED;
break;
case IB_CM_TIMEWAIT_EXIT:
case IB_CM_MRA_RECEIVED:
@@ -840,9 +831,10 @@ static int cma_ib_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event)
goto out;
case IB_CM_REJ_RECEIVED:
cma_modify_qp_err(&id_priv->id);
- status = ib_event->param.rej_rcvd.reason;
- event = RDMA_CM_EVENT_REJECTED;
- private_data_len = IB_CM_REJ_PRIVATE_DATA_SIZE;
+ event.status = ib_event->param.rej_rcvd.reason;
+ event.event = RDMA_CM_EVENT_REJECTED;
+ event.param.conn.private_data = ib_event->private_data;
+ event.param.conn.private_data_len = IB_CM_REJ_PRIVATE_DATA_SIZE;
break;
default:
printk(KERN_ERR "RDMA CMA: unexpected IB CM event: %d",
@@ -850,8 +842,7 @@ static int cma_ib_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event)
goto out;
}
- ret = cma_notify_user(id_priv, event, status, ib_event->private_data,
- private_data_len);
+ ret = id_priv->id.event_handler(&id_priv->id, &event);
if (ret) {
/* Destroy the CM ID by returning a non-zero value. */
id_priv->cm_id.ib = NULL;
@@ -865,8 +856,8 @@ out:
return ret;
}
-static struct rdma_id_private *cma_new_id(struct rdma_cm_id *listen_id,
- struct ib_cm_event *ib_event)
+static struct rdma_id_private *cma_new_conn_id(struct rdma_cm_id *listen_id,
+ struct ib_cm_event *ib_event)
{
struct rdma_id_private *id_priv;
struct rdma_cm_id *id;
@@ -913,9 +904,61 @@ err:
return NULL;
}
+static struct rdma_id_private *cma_new_udp_id(struct rdma_cm_id *listen_id,
+ struct ib_cm_event *ib_event)
+{
+ struct rdma_id_private *id_priv;
+ struct rdma_cm_id *id;
+ union cma_ip_addr *src, *dst;
+ __u16 port;
+ u8 ip_ver;
+ int ret;
+
+ id = rdma_create_id(listen_id->event_handler, listen_id->context,
+ listen_id->ps);
+ if (IS_ERR(id))
+ return NULL;
+
+
+ if (cma_get_net_info(ib_event->private_data, listen_id->ps,
+ &ip_ver, &port, &src, &dst))
+ goto err;
+
+ cma_save_net_info(&id->route.addr, &listen_id->route.addr,
+ ip_ver, port, src, dst);
+
+ ret = rdma_translate_ip(&id->route.addr.src_addr,
+ &id->route.addr.dev_addr);
+ if (ret)
+ goto err;
+
+ id_priv = container_of(id, struct rdma_id_private, id);
+ id_priv->state = CMA_CONNECT;
+ return id_priv;
+err:
+ rdma_destroy_id(id);
+ return NULL;
+}
+
+static void cma_set_req_event_data(struct rdma_cm_event *event,
+ struct ib_cm_req_event_param *req_data,
+ void *private_data, int offset)
+{
+ event->param.conn.private_data = private_data + offset;
+ event->param.conn.private_data_len = IB_CM_REQ_PRIVATE_DATA_SIZE - offset;
+ event->param.conn.responder_resources = req_data->responder_resources;
+ event->param.conn.initiator_depth = req_data->initiator_depth;
+ event->param.conn.flow_control = req_data->flow_control;
+ event->param.conn.retry_count = req_data->retry_count;
+ event->param.conn.rnr_retry_count = req_data->rnr_retry_count;
+ event->param.conn.srq = req_data->srq;
+ event->param.conn.qp_num = req_data->remote_qpn;
+}
+
static int cma_req_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event)
{
struct rdma_id_private *listen_id, *conn_id;
+ struct rdma_cm_event event;
int offset, ret;
listen_id = cm_id->context;
@@ -925,7 +968,19 @@ static int cma_req_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event)
goto out;
}
- conn_id = cma_new_id(&listen_id->id, ib_event);
+ memset(&event, 0, sizeof event);
+ offset = cma_user_data_offset(listen_id->id.ps);
+ event.event = RDMA_CM_EVENT_CONNECT_REQUEST;
+ if (listen_id->id.ps == RDMA_PS_UDP) {
+ conn_id = cma_new_udp_id(&listen_id->id, ib_event);
+ event.param.ud.private_data = ib_event->private_data + offset;
+ event.param.ud.private_data_len =
+ IB_CM_SIDR_REQ_PRIVATE_DATA_SIZE - offset;
+ } else {
+ conn_id = cma_new_conn_id(&listen_id->id, ib_event);
+ cma_set_req_event_data(&event, &ib_event->param.req_rcvd,
+ ib_event->private_data, offset);
+ }
if (!conn_id) {
ret = -ENOMEM;
goto out;
@@ -935,29 +990,25 @@ static int cma_req_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event)
mutex_lock(&lock);
ret = cma_acquire_dev(conn_id);
mutex_unlock(&lock);
- if (ret) {
- ret = -ENODEV;
- cma_exch(conn_id, CMA_DESTROYING);
- cma_release_remove(conn_id);
- rdma_destroy_id(&conn_id->id);
- goto out;
- }
+ if (ret)
+ goto release_conn_id;
conn_id->cm_id.ib = cm_id;
cm_id->context = conn_id;
cm_id->cm_handler = cma_ib_handler;
- offset = cma_user_data_offset(listen_id->id.ps);
- ret = cma_notify_user(conn_id, RDMA_CM_EVENT_CONNECT_REQUEST, 0,
- ib_event->private_data + offset,
- IB_CM_REQ_PRIVATE_DATA_SIZE - offset);
- if (ret) {
- /* Destroy the CM ID by returning a non-zero value. */
- conn_id->cm_id.ib = NULL;
- cma_exch(conn_id, CMA_DESTROYING);
- cma_release_remove(conn_id);
- rdma_destroy_id(&conn_id->id);
- }
+ ret = conn_id->id.event_handler(&conn_id->id, &event);
+ if (!ret)
+ goto out;
+
+ /* Destroy the CM ID by returning a non-zero value. */
+ conn_id->cm_id.ib = NULL;
+
+release_conn_id:
+ cma_exch(conn_id, CMA_DESTROYING);
+ cma_release_remove(conn_id);
+ rdma_destroy_id(&conn_id->id);
+
out:
cma_release_remove(listen_id);
return ret;
@@ -965,8 +1016,7 @@ out:
static __be64 cma_get_service_id(enum rdma_port_space ps, struct sockaddr *addr)
{
- return cpu_to_be64(((u64)ps << 16) +
- be16_to_cpu(((struct sockaddr_in *) addr)->sin_port));
+ return cpu_to_be64(((u64)ps << 16) + be16_to_cpu(cma_port(addr)));
}
static void cma_set_compare_data(enum rdma_port_space ps, struct sockaddr *addr,
@@ -1022,36 +1072,49 @@ static void cma_set_compare_data(enum rdma_port_space ps, struct sockaddr *addr,
static int cma_iw_handler(struct iw_cm_id *iw_id, struct iw_cm_event *iw_event)
{
struct rdma_id_private *id_priv = iw_id->context;
- enum rdma_cm_event_type event = 0;
+ struct rdma_cm_event event;
struct sockaddr_in *sin;
int ret = 0;
+ memset(&event, 0, sizeof event);
atomic_inc(&id_priv->dev_remove);
switch (iw_event->event) {
case IW_CM_EVENT_CLOSE:
- event = RDMA_CM_EVENT_DISCONNECTED;
+ event.event = RDMA_CM_EVENT_DISCONNECTED;
break;
case IW_CM_EVENT_CONNECT_REPLY:
sin = (struct sockaddr_in *) &id_priv->id.route.addr.src_addr;
*sin = iw_event->local_addr;
sin = (struct sockaddr_in *) &id_priv->id.route.addr.dst_addr;
*sin = iw_event->remote_addr;
- if (iw_event->status)
- event = RDMA_CM_EVENT_REJECTED;
- else
- event = RDMA_CM_EVENT_ESTABLISHED;
+ switch (iw_event->status) {
+ case 0:
+ event.event = RDMA_CM_EVENT_ESTABLISHED;
+ break;
+ case -ECONNRESET:
+ case -ECONNREFUSED:
+ event.event = RDMA_CM_EVENT_REJECTED;
+ break;
+ case -ETIMEDOUT:
+ event.event = RDMA_CM_EVENT_UNREACHABLE;
+ break;
+ default:
+ event.event = RDMA_CM_EVENT_CONNECT_ERROR;
+ break;
+ }
break;
case IW_CM_EVENT_ESTABLISHED:
- event = RDMA_CM_EVENT_ESTABLISHED;
+ event.event = RDMA_CM_EVENT_ESTABLISHED;
break;
default:
BUG_ON(1);
}
- ret = cma_notify_user(id_priv, event, iw_event->status,
- iw_event->private_data,
- iw_event->private_data_len);
+ event.status = iw_event->status;
+ event.param.conn.private_data = iw_event->private_data;
+ event.param.conn.private_data_len = iw_event->private_data_len;
+ ret = id_priv->id.event_handler(&id_priv->id, &event);
if (ret) {
/* Destroy the CM ID by returning a non-zero value. */
id_priv->cm_id.iw = NULL;
@@ -1072,6 +1135,7 @@ static int iw_conn_req_handler(struct iw_cm_id *cm_id,
struct rdma_id_private *listen_id, *conn_id;
struct sockaddr_in *sin;
struct net_device *dev = NULL;
+ struct rdma_cm_event event;
int ret;
listen_id = cm_id->context;
@@ -1125,9 +1189,11 @@ static int iw_conn_req_handler(struct iw_cm_id *cm_id,
sin = (struct sockaddr_in *) &new_cm_id->route.addr.dst_addr;
*sin = iw_event->remote_addr;
- ret = cma_notify_user(conn_id, RDMA_CM_EVENT_CONNECT_REQUEST, 0,
- iw_event->private_data,
- iw_event->private_data_len);
+ memset(&event, 0, sizeof event);
+ event.event = RDMA_CM_EVENT_CONNECT_REQUEST;
+ event.param.conn.private_data = iw_event->private_data;
+ event.param.conn.private_data_len = iw_event->private_data_len;
+ ret = conn_id->id.event_handler(&conn_id->id, &event);
if (ret) {
/* User wants to destroy the CM ID */
conn_id->cm_id.iw = NULL;
@@ -1341,9 +1407,9 @@ static int cma_query_ib_route(struct rdma_id_private *id_priv, int timeout_ms,
return (id_priv->query_id < 0) ? id_priv->query_id : 0;
}
-static void cma_work_handler(void *data)
+static void cma_work_handler(struct work_struct *_work)
{
- struct cma_work *work = data;
+ struct cma_work *work = container_of(_work, struct cma_work, work);
struct rdma_id_private *id_priv = work->id;
int destroy = 0;
@@ -1374,7 +1440,7 @@ static int cma_resolve_ib_route(struct rdma_id_private *id_priv, int timeout_ms)
return -ENOMEM;
work->id = id_priv;
- INIT_WORK(&work->work, cma_work_handler, work);
+ INIT_WORK(&work->work, cma_work_handler);
work->old_state = CMA_ROUTE_QUERY;
work->new_state = CMA_ROUTE_RESOLVED;
work->event.event = RDMA_CM_EVENT_ROUTE_RESOLVED;
@@ -1431,7 +1497,7 @@ static int cma_resolve_iw_route(struct rdma_id_private *id_priv, int timeout_ms)
return -ENOMEM;
work->id = id_priv;
- INIT_WORK(&work->work, cma_work_handler, work);
+ INIT_WORK(&work->work, cma_work_handler);
work->old_state = CMA_ROUTE_QUERY;
work->new_state = CMA_ROUTE_RESOLVED;
work->event.event = RDMA_CM_EVENT_ROUTE_RESOLVED;
@@ -1481,19 +1547,18 @@ static int cma_bind_loopback(struct rdma_id_private *id_priv)
u8 p;
mutex_lock(&lock);
+ if (list_empty(&dev_list)) {
+ ret = -ENODEV;
+ goto out;
+ }
list_for_each_entry(cma_dev, &dev_list, list)
for (p = 1; p <= cma_dev->device->phys_port_cnt; ++p)
- if (!ib_query_port (cma_dev->device, p, &port_attr) &&
+ if (!ib_query_port(cma_dev->device, p, &port_attr) &&
port_attr.state == IB_PORT_ACTIVE)
goto port_found;
- if (!list_empty(&dev_list)) {
- p = 1;
- cma_dev = list_entry(dev_list.next, struct cma_device, list);
- } else {
- ret = -ENODEV;
- goto out;
- }
+ p = 1;
+ cma_dev = list_entry(dev_list.next, struct cma_device, list);
port_found:
ret = ib_get_cached_gid(cma_dev->device, p, 0, &gid);
@@ -1517,8 +1582,9 @@ static void addr_handler(int status, struct sockaddr *src_addr,
struct rdma_dev_addr *dev_addr, void *context)
{
struct rdma_id_private *id_priv = context;
- enum rdma_cm_event_type event;
+ struct rdma_cm_event event;
+ memset(&event, 0, sizeof event);
atomic_inc(&id_priv->dev_remove);
/*
@@ -1538,14 +1604,15 @@ static void addr_handler(int status, struct sockaddr *src_addr,
if (status) {
if (!cma_comp_exch(id_priv, CMA_ADDR_RESOLVED, CMA_ADDR_BOUND))
goto out;
- event = RDMA_CM_EVENT_ADDR_ERROR;
+ event.event = RDMA_CM_EVENT_ADDR_ERROR;
+ event.status = status;
} else {
memcpy(&id_priv->id.route.addr.src_addr, src_addr,
ip_addr_size(src_addr));
- event = RDMA_CM_EVENT_ADDR_RESOLVED;
+ event.event = RDMA_CM_EVENT_ADDR_RESOLVED;
}
- if (cma_notify_user(id_priv, event, status, NULL, 0)) {
+ if (id_priv->id.event_handler(&id_priv->id, &event)) {
cma_exch(id_priv, CMA_DESTROYING);
cma_release_remove(id_priv);
cma_deref_id(id_priv);
@@ -1585,7 +1652,7 @@ static int cma_resolve_loopback(struct rdma_id_private *id_priv)
}
work->id = id_priv;
- INIT_WORK(&work->work, cma_work_handler, work);
+ INIT_WORK(&work->work, cma_work_handler);
work->old_state = CMA_ADDR_QUERY;
work->new_state = CMA_ADDR_RESOLVED;
work->event.event = RDMA_CM_EVENT_ADDR_RESOLVED;
@@ -1735,6 +1802,9 @@ static int cma_get_port(struct rdma_id_private *id_priv)
case RDMA_PS_TCP:
ps = &tcp_ps;
break;
+ case RDMA_PS_UDP:
+ ps = &udp_ps;
+ break;
default:
return -EPROTONOSUPPORT;
}
@@ -1823,6 +1893,110 @@ static int cma_format_hdr(void *hdr, enum rdma_port_space ps,
return 0;
}
+static int cma_sidr_rep_handler(struct ib_cm_id *cm_id,
+ struct ib_cm_event *ib_event)
+{
+ struct rdma_id_private *id_priv = cm_id->context;
+ struct rdma_cm_event event;
+ struct ib_cm_sidr_rep_event_param *rep = &ib_event->param.sidr_rep_rcvd;
+ int ret = 0;
+
+ memset(&event, 0, sizeof event);
+ atomic_inc(&id_priv->dev_remove);
+ if (!cma_comp(id_priv, CMA_CONNECT))
+ goto out;
+
+ switch (ib_event->event) {
+ case IB_CM_SIDR_REQ_ERROR:
+ event.event = RDMA_CM_EVENT_UNREACHABLE;
+ event.status = -ETIMEDOUT;
+ break;
+ case IB_CM_SIDR_REP_RECEIVED:
+ event.param.ud.private_data = ib_event->private_data;
+ event.param.ud.private_data_len = IB_CM_SIDR_REP_PRIVATE_DATA_SIZE;
+ if (rep->status != IB_SIDR_SUCCESS) {
+ event.event = RDMA_CM_EVENT_UNREACHABLE;
+ event.status = ib_event->param.sidr_rep_rcvd.status;
+ break;
+ }
+ if (rep->qkey != RDMA_UD_QKEY) {
+ event.event = RDMA_CM_EVENT_UNREACHABLE;
+ event.status = -EINVAL;
+ break;
+ }
+ ib_init_ah_from_path(id_priv->id.device, id_priv->id.port_num,
+ id_priv->id.route.path_rec,
+ &event.param.ud.ah_attr);
+ event.param.ud.qp_num = rep->qpn;
+ event.param.ud.qkey = rep->qkey;
+ event.event = RDMA_CM_EVENT_ESTABLISHED;
+ event.status = 0;
+ break;
+ default:
+ printk(KERN_ERR "RDMA CMA: unexpected IB CM event: %d",
+ ib_event->event);
+ goto out;
+ }
+
+ ret = id_priv->id.event_handler(&id_priv->id, &event);
+ if (ret) {
+ /* Destroy the CM ID by returning a non-zero value. */
+ id_priv->cm_id.ib = NULL;
+ cma_exch(id_priv, CMA_DESTROYING);
+ cma_release_remove(id_priv);
+ rdma_destroy_id(&id_priv->id);
+ return ret;
+ }
+out:
+ cma_release_remove(id_priv);
+ return ret;
+}
+
+static int cma_resolve_ib_udp(struct rdma_id_private *id_priv,
+ struct rdma_conn_param *conn_param)
+{
+ struct ib_cm_sidr_req_param req;
+ struct rdma_route *route;
+ int ret;
+
+ req.private_data_len = sizeof(struct cma_hdr) +
+ conn_param->private_data_len;
+ req.private_data = kzalloc(req.private_data_len, GFP_ATOMIC);
+ if (!req.private_data)
+ return -ENOMEM;
+
+ if (conn_param->private_data && conn_param->private_data_len)
+ memcpy((void *) req.private_data + sizeof(struct cma_hdr),
+ conn_param->private_data, conn_param->private_data_len);
+
+ route = &id_priv->id.route;
+ ret = cma_format_hdr((void *) req.private_data, id_priv->id.ps, route);
+ if (ret)
+ goto out;
+
+ id_priv->cm_id.ib = ib_create_cm_id(id_priv->id.device,
+ cma_sidr_rep_handler, id_priv);
+ if (IS_ERR(id_priv->cm_id.ib)) {
+ ret = PTR_ERR(id_priv->cm_id.ib);
+ goto out;
+ }
+
+ req.path = route->path_rec;
+ req.service_id = cma_get_service_id(id_priv->id.ps,
+ &route->addr.dst_addr);
+ req.timeout_ms = 1 << (CMA_CM_RESPONSE_TIMEOUT - 8);
+ req.max_cm_retries = CMA_MAX_CM_RETRIES;
+
+ ret = ib_send_cm_sidr_req(id_priv->cm_id.ib, &req);
+ if (ret) {
+ ib_destroy_cm_id(id_priv->cm_id.ib);
+ id_priv->cm_id.ib = NULL;
+ }
+out:
+ kfree(req.private_data);
+ return ret;
+}
+
static int cma_connect_ib(struct rdma_id_private *id_priv,
struct rdma_conn_param *conn_param)
{
@@ -1862,7 +2036,7 @@ static int cma_connect_ib(struct rdma_id_private *id_priv,
req.service_id = cma_get_service_id(id_priv->id.ps,
&route->addr.dst_addr);
req.qp_num = id_priv->qp_num;
- req.qp_type = id_priv->qp_type;
+ req.qp_type = IB_QPT_RC;
req.starting_psn = id_priv->seq_num;
req.responder_resources = conn_param->responder_resources;
req.initiator_depth = conn_param->initiator_depth;
@@ -1939,13 +2113,15 @@ int rdma_connect(struct rdma_cm_id *id, struct rdma_conn_param *conn_param)
if (!id->qp) {
id_priv->qp_num = conn_param->qp_num;
- id_priv->qp_type = conn_param->qp_type;
id_priv->srq = conn_param->srq;
}
switch (rdma_node_get_transport(id->device->node_type)) {
case RDMA_TRANSPORT_IB:
- ret = cma_connect_ib(id_priv, conn_param);
+ if (id->ps == RDMA_PS_UDP)
+ ret = cma_resolve_ib_udp(id_priv, conn_param);
+ else
+ ret = cma_connect_ib(id_priv, conn_param);
break;
case RDMA_TRANSPORT_IWARP:
ret = cma_connect_iw(id_priv, conn_param);
@@ -1968,11 +2144,25 @@ static int cma_accept_ib(struct rdma_id_private *id_priv,
struct rdma_conn_param *conn_param)
{
struct ib_cm_rep_param rep;
- int ret;
+ struct ib_qp_attr qp_attr;
+ int qp_attr_mask, ret;
- ret = cma_modify_qp_rtr(&id_priv->id);
- if (ret)
- return ret;
+ if (id_priv->id.qp) {
+ ret = cma_modify_qp_rtr(&id_priv->id);
+ if (ret)
+ goto out;
+
+ qp_attr.qp_state = IB_QPS_RTS;
+ ret = ib_cm_init_qp_attr(id_priv->cm_id.ib, &qp_attr,
+ &qp_attr_mask);
+ if (ret)
+ goto out;
+
+ qp_attr.max_rd_atomic = conn_param->initiator_depth;
+ ret = ib_modify_qp(id_priv->id.qp, &qp_attr, qp_attr_mask);
+ if (ret)
+ goto out;
+ }
memset(&rep, 0, sizeof rep);
rep.qp_num = id_priv->qp_num;
@@ -1987,7 +2177,9 @@ static int cma_accept_ib(struct rdma_id_private *id_priv,
rep.rnr_retry_count = conn_param->rnr_retry_count;
rep.srq = id_priv->srq ? 1 : 0;
- return ib_send_cm_rep(id_priv->cm_id.ib, &rep);
+ ret = ib_send_cm_rep(id_priv->cm_id.ib, &rep);
+out:
+ return ret;
}
static int cma_accept_iw(struct rdma_id_private *id_priv,
@@ -2012,6 +2204,24 @@ static int cma_accept_iw(struct rdma_id_private *id_priv,
return iw_cm_accept(id_priv->cm_id.iw, &iw_param);
}
+static int cma_send_sidr_rep(struct rdma_id_private *id_priv,
+ enum ib_cm_sidr_status status,
+ const void *private_data, int private_data_len)
+{
+ struct ib_cm_sidr_rep_param rep;
+
+ memset(&rep, 0, sizeof rep);
+ rep.status = status;
+ if (status == IB_SIDR_SUCCESS) {
+ rep.qp_num = id_priv->qp_num;
+ rep.qkey = RDMA_UD_QKEY;
+ }
+ rep.private_data = private_data;
+ rep.private_data_len = private_data_len;
+
+ return ib_send_cm_sidr_rep(id_priv->cm_id.ib, &rep);
+}
+
int rdma_accept(struct rdma_cm_id *id, struct rdma_conn_param *conn_param)
{
struct rdma_id_private *id_priv;
@@ -2023,13 +2233,16 @@ int rdma_accept(struct rdma_cm_id *id, struct rdma_conn_param *conn_param)
if (!id->qp && conn_param) {
id_priv->qp_num = conn_param->qp_num;
- id_priv->qp_type = conn_param->qp_type;
id_priv->srq = conn_param->srq;
}
switch (rdma_node_get_transport(id->device->node_type)) {
case RDMA_TRANSPORT_IB:
- if (conn_param)
+ if (id->ps == RDMA_PS_UDP)
+ ret = cma_send_sidr_rep(id_priv, IB_SIDR_SUCCESS,
+ conn_param->private_data,
+ conn_param->private_data_len);
+ else if (conn_param)
ret = cma_accept_ib(id_priv, conn_param);
else
ret = cma_rep_recv(id_priv);
@@ -2053,6 +2266,27 @@ reject:
}
EXPORT_SYMBOL(rdma_accept);
+int rdma_notify(struct rdma_cm_id *id, enum ib_event_type event)
+{
+ struct rdma_id_private *id_priv;
+ int ret;
+
+ id_priv = container_of(id, struct rdma_id_private, id);
+ if (!cma_comp(id_priv, CMA_CONNECT))
+ return -EINVAL;
+
+ switch (id->device->node_type) {
+ case RDMA_NODE_IB_CA:
+ ret = ib_cm_notify(id_priv->cm_id.ib, event);
+ break;
+ default:
+ ret = 0;
+ break;
+ }
+ return ret;
+}
+EXPORT_SYMBOL(rdma_notify);
+
int rdma_reject(struct rdma_cm_id *id, const void *private_data,
u8 private_data_len)
{
@@ -2065,9 +2299,13 @@ int rdma_reject(struct rdma_cm_id *id, const void *private_data,
switch (rdma_node_get_transport(id->device->node_type)) {
case RDMA_TRANSPORT_IB:
- ret = ib_send_cm_rej(id_priv->cm_id.ib,
- IB_CM_REJ_CONSUMER_DEFINED, NULL, 0,
- private_data, private_data_len);
+ if (id->ps == RDMA_PS_UDP)
+ ret = cma_send_sidr_rep(id_priv, IB_SIDR_REJECT,
+ private_data, private_data_len);
+ else
+ ret = ib_send_cm_rej(id_priv->cm_id.ib,
+ IB_CM_REJ_CONSUMER_DEFINED, NULL,
+ 0, private_data, private_data_len);
break;
case RDMA_TRANSPORT_IWARP:
ret = iw_cm_reject(id_priv->cm_id.iw,
@@ -2123,8 +2361,6 @@ static void cma_add_one(struct ib_device *device)
cma_dev->device = device;
cma_dev->node_guid = device->node_guid;
- if (!cma_dev->node_guid)
- goto err;
init_completion(&cma_dev->comp);
atomic_set(&cma_dev->refcount, 1);
@@ -2136,13 +2372,11 @@ static void cma_add_one(struct ib_device *device)
list_for_each_entry(id_priv, &listen_any_list, list)
cma_listen_on_dev(id_priv, cma_dev);
mutex_unlock(&lock);
- return;
-err:
- kfree(cma_dev);
}
static int cma_remove_id_dev(struct rdma_id_private *id_priv)
{
+ struct rdma_cm_event event;
enum cma_state state;
/* Record that we want to remove the device */
@@ -2157,8 +2391,9 @@ static int cma_remove_id_dev(struct rdma_id_private *id_priv)
if (!cma_comp(id_priv, CMA_DEVICE_REMOVAL))
return 0;
- return cma_notify_user(id_priv, RDMA_CM_EVENT_DEVICE_REMOVAL,
- 0, NULL, 0);
+ memset(&event, 0, sizeof event);
+ event.event = RDMA_CM_EVENT_DEVICE_REMOVAL;
+ return id_priv->id.event_handler(&id_priv->id, &event);
}
static void cma_process_remove(struct cma_device *cma_dev)
@@ -2240,6 +2475,7 @@ static void cma_cleanup(void)
destroy_workqueue(cma_wq);
idr_destroy(&sdp_ps);
idr_destroy(&tcp_ps);
+ idr_destroy(&udp_ps);
}
module_init(cma_init);
diff --git a/drivers/infiniband/core/fmr_pool.c b/drivers/infiniband/core/fmr_pool.c
index 86a3b2d..8926a2b 100644
--- a/drivers/infiniband/core/fmr_pool.c
+++ b/drivers/infiniband/core/fmr_pool.c
@@ -394,20 +394,12 @@ EXPORT_SYMBOL(ib_destroy_fmr_pool);
*/
int ib_flush_fmr_pool(struct ib_fmr_pool *pool)
{
- int serial;
-
- atomic_inc(&pool->req_ser);
- /*
- * It's OK if someone else bumps req_ser again here -- we'll
- * just wait a little longer.
- */
- serial = atomic_read(&pool->req_ser);
+ int serial = atomic_inc_return(&pool->req_ser);
wake_up_process(pool->thread);
if (wait_event_interruptible(pool->force_wait,
- atomic_read(&pool->flush_ser) -
- atomic_read(&pool->req_ser) >= 0))
+ atomic_read(&pool->flush_ser) - serial >= 0))
return -EINTR;
return 0;
diff --git a/drivers/infiniband/core/iwcm.c b/drivers/infiniband/core/iwcm.c
index c3fb304..1039ad5 100644
--- a/drivers/infiniband/core/iwcm.c
+++ b/drivers/infiniband/core/iwcm.c
@@ -80,7 +80,7 @@ struct iwcm_work {
* 1) in the event upcall, cm_event_handler(), for a listening cm_id. If
* the backlog is exceeded, then no more connection request events will
* be processed. cm_event_handler() returns -ENOMEM in this case. Its up
- * to the provider to reject the connectino request.
+ * to the provider to reject the connection request.
* 2) in the connection request workqueue handler, cm_conn_req_handler().
* If work elements cannot be allocated for the new connect request cm_id,
* then IWCM will call the provider reject method. This is ok since
@@ -131,26 +131,25 @@ static int alloc_work_entries(struct iwcm_id_private *cm_id_priv, int count)
}
/*
- * Save private data from incoming connection requests in the
- * cm_id_priv so the low level driver doesn't have to. Adjust
+ * Save private data from incoming connection requests to
+ * iw_cm_event, so the low level driver doesn't have to. Adjust
* the event ptr to point to the local copy.
*/
-static int copy_private_data(struct iwcm_id_private *cm_id_priv,
- struct iw_cm_event *event)
+static int copy_private_data(struct iw_cm_event *event)
{
void *p;
- p = kmalloc(event->private_data_len, GFP_ATOMIC);
+ p = kmemdup(event->private_data, event->private_data_len, GFP_ATOMIC);
if (!p)
return -ENOMEM;
- memcpy(p, event->private_data, event->private_data_len);
event->private_data = p;
return 0;
}
/*
- * Release a reference on cm_id. If the last reference is being removed
- * and iw_destroy_cm_id is waiting, wake up the waiting thread.
+ * Release a reference on cm_id. If the last reference is being
+ * released, enable the waiting thread (in iw_destroy_cm_id) to
+ * get woken up, and return 1 if a thread is already waiting.
*/
static int iwcm_deref_id(struct iwcm_id_private *cm_id_priv)
{
@@ -243,7 +242,7 @@ static int iwcm_modify_qp_sqd(struct ib_qp *qp)
/*
* CM_ID <-- CLOSING
*
- * Block if a passive or active connection is currenlty being processed. Then
+ * Block if a passive or active connection is currently being processed. Then
* process the event as follows:
* - If we are ESTABLISHED, move to CLOSING and modify the QP state
* based on the abrupt flag
@@ -408,7 +407,7 @@ int iw_cm_listen(struct iw_cm_id *cm_id, int backlog)
{
struct iwcm_id_private *cm_id_priv;
unsigned long flags;
- int ret = 0;
+ int ret;
cm_id_priv = container_of(cm_id, struct iwcm_id_private, id);
@@ -535,7 +534,7 @@ EXPORT_SYMBOL(iw_cm_accept);
int iw_cm_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *iw_param)
{
struct iwcm_id_private *cm_id_priv;
- int ret = 0;
+ int ret;
unsigned long flags;
struct ib_qp *qp;
@@ -620,7 +619,7 @@ static void cm_conn_req_handler(struct iwcm_id_private *listen_id_priv,
spin_lock_irqsave(&listen_id_priv->lock, flags);
if (listen_id_priv->state != IW_CM_STATE_LISTEN) {
spin_unlock_irqrestore(&listen_id_priv->lock, flags);
- return;
+ goto out;
}
spin_unlock_irqrestore(&listen_id_priv->lock, flags);
@@ -629,7 +628,7 @@ static void cm_conn_req_handler(struct iwcm_id_private *listen_id_priv,
listen_id_priv->id.context);
/* If the cm_id could not be created, ignore the request */
if (IS_ERR(cm_id))
- return;
+ goto out;
cm_id->provider_data = iw_event->provider_data;
cm_id->local_addr = iw_event->local_addr;
@@ -642,7 +641,7 @@ static void cm_conn_req_handler(struct iwcm_id_private *listen_id_priv,
if (ret) {
iw_cm_reject(cm_id, NULL, 0);
iw_destroy_cm_id(cm_id);
- return;
+ goto out;
}
/* Call the client CM handler */
@@ -654,6 +653,7 @@ static void cm_conn_req_handler(struct iwcm_id_private *listen_id_priv,
kfree(cm_id);
}
+out:
if (iw_event->private_data_len)
kfree(iw_event->private_data);
}
@@ -674,7 +674,7 @@ static int cm_conn_est_handler(struct iwcm_id_private *cm_id_priv,
struct iw_cm_event *iw_event)
{
unsigned long flags;
- int ret = 0;
+ int ret;
spin_lock_irqsave(&cm_id_priv->lock, flags);
@@ -704,7 +704,7 @@ static int cm_conn_rep_handler(struct iwcm_id_private *cm_id_priv,
struct iw_cm_event *iw_event)
{
unsigned long flags;
- int ret = 0;
+ int ret;
spin_lock_irqsave(&cm_id_priv->lock, flags);
/*
@@ -828,9 +828,10 @@ static int process_event(struct iwcm_id_private *cm_id_priv,
* thread asleep on the destroy_comp list vs. an object destroyed
* here synchronously when the last reference is removed.
*/
-static void cm_work_handler(void *arg)
+static void cm_work_handler(struct work_struct *_work)
{
- struct iwcm_work *work = arg, lwork;
+ struct iwcm_work *work = container_of(_work, struct iwcm_work, work);
+ struct iw_cm_event levent;
struct iwcm_id_private *cm_id_priv = work->cm_id;
unsigned long flags;
int empty;
@@ -843,11 +844,11 @@ static void cm_work_handler(void *arg)
struct iwcm_work, list);
list_del_init(&work->list);
empty = list_empty(&cm_id_priv->work_list);
- lwork = *work;
+ levent = work->event;
put_work(work);
spin_unlock_irqrestore(&cm_id_priv->lock, flags);
- ret = process_event(cm_id_priv, &work->event);
+ ret = process_event(cm_id_priv, &levent);
if (ret) {
set_bit(IWCM_F_CALLBACK_DESTROY, &cm_id_priv->flags);
destroy_cm_id(&cm_id_priv->id);
@@ -899,14 +900,14 @@ static int cm_event_handler(struct iw_cm_id *cm_id,
goto out;
}
- INIT_WORK(&work->work, cm_work_handler, work);
+ INIT_WORK(&work->work, cm_work_handler);
work->cm_id = cm_id_priv;
work->event = *iw_event;
if ((work->event.event == IW_CM_EVENT_CONNECT_REQUEST ||
work->event.event == IW_CM_EVENT_CONNECT_REPLY) &&
work->event.private_data_len) {
- ret = copy_private_data(cm_id_priv, &work->event);
+ ret = copy_private_data(&work->event);
if (ret) {
put_work(work);
goto out;
diff --git a/drivers/infiniband/core/mad.c b/drivers/infiniband/core/mad.c
index a72bcea..5ed141e 100644
--- a/drivers/infiniband/core/mad.c
+++ b/drivers/infiniband/core/mad.c
@@ -46,7 +46,7 @@ MODULE_DESCRIPTION("kernel IB MAD API");
MODULE_AUTHOR("Hal Rosenstock");
MODULE_AUTHOR("Sean Hefty");
-static kmem_cache_t *ib_mad_cache;
+static struct kmem_cache *ib_mad_cache;
static struct list_head ib_mad_port_list;
static u32 ib_mad_client_id = 0;
@@ -65,8 +65,8 @@ static struct ib_mad_agent_private *find_mad_agent(
static int ib_mad_post_receive_mads(struct ib_mad_qp_info *qp_info,
struct ib_mad_private *mad);
static void cancel_mads(struct ib_mad_agent_private *mad_agent_priv);
-static void timeout_sends(void *data);
-static void local_completions(void *data);
+static void timeout_sends(struct work_struct *work);
+static void local_completions(struct work_struct *work);
static int add_nonoui_reg_req(struct ib_mad_reg_req *mad_reg_req,
struct ib_mad_agent_private *agent_priv,
u8 mgmt_class);
@@ -356,10 +356,9 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device,
INIT_LIST_HEAD(&mad_agent_priv->wait_list);
INIT_LIST_HEAD(&mad_agent_priv->done_list);
INIT_LIST_HEAD(&mad_agent_priv->rmpp_list);
- INIT_WORK(&mad_agent_priv->timed_work, timeout_sends, mad_agent_priv);
+ INIT_DELAYED_WORK(&mad_agent_priv->timed_work, timeout_sends);
INIT_LIST_HEAD(&mad_agent_priv->local_list);
- INIT_WORK(&mad_agent_priv->local_work, local_completions,
- mad_agent_priv);
+ INIT_WORK(&mad_agent_priv->local_work, local_completions);
atomic_set(&mad_agent_priv->refcount, 1);
init_completion(&mad_agent_priv->comp);
@@ -999,17 +998,17 @@ int ib_send_mad(struct ib_mad_send_wr_private *mad_send_wr)
mad_agent = mad_send_wr->send_buf.mad_agent;
sge = mad_send_wr->sg_list;
- sge[0].addr = dma_map_single(mad_agent->device->dma_device,
- mad_send_wr->send_buf.mad,
- sge[0].length,
- DMA_TO_DEVICE);
- pci_unmap_addr_set(mad_send_wr, header_mapping, sge[0].addr);
-
- sge[1].addr = dma_map_single(mad_agent->device->dma_device,
- ib_get_payload(mad_send_wr),
- sge[1].length,
- DMA_TO_DEVICE);
- pci_unmap_addr_set(mad_send_wr, payload_mapping, sge[1].addr);
+ sge[0].addr = ib_dma_map_single(mad_agent->device,
+ mad_send_wr->send_buf.mad,
+ sge[0].length,
+ DMA_TO_DEVICE);
+ mad_send_wr->header_mapping = sge[0].addr;
+
+ sge[1].addr = ib_dma_map_single(mad_agent->device,
+ ib_get_payload(mad_send_wr),
+ sge[1].length,
+ DMA_TO_DEVICE);
+ mad_send_wr->payload_mapping = sge[1].addr;
spin_lock_irqsave(&qp_info->send_queue.lock, flags);
if (qp_info->send_queue.count < qp_info->send_queue.max_active) {
@@ -1027,12 +1026,12 @@ int ib_send_mad(struct ib_mad_send_wr_private *mad_send_wr)
}
spin_unlock_irqrestore(&qp_info->send_queue.lock, flags);
if (ret) {
- dma_unmap_single(mad_agent->device->dma_device,
- pci_unmap_addr(mad_send_wr, header_mapping),
- sge[0].length, DMA_TO_DEVICE);
- dma_unmap_single(mad_agent->device->dma_device,
- pci_unmap_addr(mad_send_wr, payload_mapping),
- sge[1].length, DMA_TO_DEVICE);
+ ib_dma_unmap_single(mad_agent->device,
+ mad_send_wr->header_mapping,
+ sge[0].length, DMA_TO_DEVICE);
+ ib_dma_unmap_single(mad_agent->device,
+ mad_send_wr->payload_mapping,
+ sge[1].length, DMA_TO_DEVICE);
}
return ret;
}
@@ -1851,11 +1850,11 @@ static void ib_mad_recv_done_handler(struct ib_mad_port_private *port_priv,
mad_priv_hdr = container_of(mad_list, struct ib_mad_private_header,
mad_list);
recv = container_of(mad_priv_hdr, struct ib_mad_private, header);
- dma_unmap_single(port_priv->device->dma_device,
- pci_unmap_addr(&recv->header, mapping),
- sizeof(struct ib_mad_private) -
- sizeof(struct ib_mad_private_header),
- DMA_FROM_DEVICE);
+ ib_dma_unmap_single(port_priv->device,
+ recv->header.mapping,
+ sizeof(struct ib_mad_private) -
+ sizeof(struct ib_mad_private_header),
+ DMA_FROM_DEVICE);
/* Setup MAD receive work completion from "normal" work completion */
recv->header.wc = *wc;
@@ -2081,12 +2080,12 @@ static void ib_mad_send_done_handler(struct ib_mad_port_private *port_priv,
qp_info = send_queue->qp_info;
retry:
- dma_unmap_single(mad_send_wr->send_buf.mad_agent->device->dma_device,
- pci_unmap_addr(mad_send_wr, header_mapping),
- mad_send_wr->sg_list[0].length, DMA_TO_DEVICE);
- dma_unmap_single(mad_send_wr->send_buf.mad_agent->device->dma_device,
- pci_unmap_addr(mad_send_wr, payload_mapping),
- mad_send_wr->sg_list[1].length, DMA_TO_DEVICE);
+ ib_dma_unmap_single(mad_send_wr->send_buf.mad_agent->device,
+ mad_send_wr->header_mapping,
+ mad_send_wr->sg_list[0].length, DMA_TO_DEVICE);
+ ib_dma_unmap_single(mad_send_wr->send_buf.mad_agent->device,
+ mad_send_wr->payload_mapping,
+ mad_send_wr->sg_list[1].length, DMA_TO_DEVICE);
queued_send_wr = NULL;
spin_lock_irqsave(&send_queue->lock, flags);
list_del(&mad_list->list);
@@ -2198,12 +2197,12 @@ static void mad_error_handler(struct ib_mad_port_private *port_priv,
/*
* IB MAD completion callback
*/
-static void ib_mad_completion_handler(void *data)
+static void ib_mad_completion_handler(struct work_struct *work)
{
struct ib_mad_port_private *port_priv;
struct ib_wc wc;
- port_priv = (struct ib_mad_port_private *)data;
+ port_priv = container_of(work, struct ib_mad_port_private, work);
ib_req_notify_cq(port_priv->cq, IB_CQ_NEXT_COMP);
while (ib_poll_cq(port_priv->cq, 1, &wc) == 1) {
@@ -2324,7 +2323,7 @@ void ib_cancel_mad(struct ib_mad_agent *mad_agent,
}
EXPORT_SYMBOL(ib_cancel_mad);
-static void local_completions(void *data)
+static void local_completions(struct work_struct *work)
{
struct ib_mad_agent_private *mad_agent_priv;
struct ib_mad_local_private *local;
@@ -2334,7 +2333,8 @@ static void local_completions(void *data)
struct ib_wc wc;
struct ib_mad_send_wc mad_send_wc;
- mad_agent_priv = (struct ib_mad_agent_private *)data;
+ mad_agent_priv =
+ container_of(work, struct ib_mad_agent_private, local_work);
spin_lock_irqsave(&mad_agent_priv->lock, flags);
while (!list_empty(&mad_agent_priv->local_list)) {
@@ -2434,14 +2434,15 @@ static int retry_send(struct ib_mad_send_wr_private *mad_send_wr)
return ret;
}
-static void timeout_sends(void *data)
+static void timeout_sends(struct work_struct *work)
{
struct ib_mad_agent_private *mad_agent_priv;
struct ib_mad_send_wr_private *mad_send_wr;
struct ib_mad_send_wc mad_send_wc;
unsigned long flags, delay;
- mad_agent_priv = (struct ib_mad_agent_private *)data;
+ mad_agent_priv = container_of(work, struct ib_mad_agent_private,
+ timed_work.work);
mad_send_wc.vendor_err = 0;
spin_lock_irqsave(&mad_agent_priv->lock, flags);
@@ -2527,13 +2528,12 @@ static int ib_mad_post_receive_mads(struct ib_mad_qp_info *qp_info,
break;
}
}
- sg_list.addr = dma_map_single(qp_info->port_priv->
- device->dma_device,
- &mad_priv->grh,
- sizeof *mad_priv -
- sizeof mad_priv->header,
- DMA_FROM_DEVICE);
- pci_unmap_addr_set(&mad_priv->header, mapping, sg_list.addr);
+ sg_list.addr = ib_dma_map_single(qp_info->port_priv->device,
+ &mad_priv->grh,
+ sizeof *mad_priv -
+ sizeof mad_priv->header,
+ DMA_FROM_DEVICE);
+ mad_priv->header.mapping = sg_list.addr;
recv_wr.wr_id = (unsigned long)&mad_priv->header.mad_list;
mad_priv->header.mad_list.mad_queue = recv_queue;
@@ -2548,12 +2548,11 @@ static int ib_mad_post_receive_mads(struct ib_mad_qp_info *qp_info,
list_del(&mad_priv->header.mad_list.list);
recv_queue->count--;
spin_unlock_irqrestore(&recv_queue->lock, flags);
- dma_unmap_single(qp_info->port_priv->device->dma_device,
- pci_unmap_addr(&mad_priv->header,
- mapping),
- sizeof *mad_priv -
- sizeof mad_priv->header,
- DMA_FROM_DEVICE);
+ ib_dma_unmap_single(qp_info->port_priv->device,
+ mad_priv->header.mapping,
+ sizeof *mad_priv -
+ sizeof mad_priv->header,
+ DMA_FROM_DEVICE);
kmem_cache_free(ib_mad_cache, mad_priv);
printk(KERN_ERR PFX "ib_post_recv failed: %d\n", ret);
break;
@@ -2585,11 +2584,11 @@ static void cleanup_recv_queue(struct ib_mad_qp_info *qp_info)
/* Remove from posted receive MAD list */
list_del(&mad_list->list);
- dma_unmap_single(qp_info->port_priv->device->dma_device,
- pci_unmap_addr(&recv->header, mapping),
- sizeof(struct ib_mad_private) -
- sizeof(struct ib_mad_private_header),
- DMA_FROM_DEVICE);
+ ib_dma_unmap_single(qp_info->port_priv->device,
+ recv->header.mapping,
+ sizeof(struct ib_mad_private) -
+ sizeof(struct ib_mad_private_header),
+ DMA_FROM_DEVICE);
kmem_cache_free(ib_mad_cache, recv);
}
@@ -2799,7 +2798,7 @@ static int ib_mad_port_open(struct ib_device *device,
ret = -ENOMEM;
goto error8;
}
- INIT_WORK(&port_priv->work, ib_mad_completion_handler, port_priv);
+ INIT_WORK(&port_priv->work, ib_mad_completion_handler);
spin_lock_irqsave(&ib_mad_port_list_lock, flags);
list_add_tail(&port_priv->port_list, &ib_mad_port_list);
diff --git a/drivers/infiniband/core/mad_priv.h b/drivers/infiniband/core/mad_priv.h
index d06b590..de89717 100644
--- a/drivers/infiniband/core/mad_priv.h
+++ b/drivers/infiniband/core/mad_priv.h
@@ -73,7 +73,7 @@ struct ib_mad_private_header {
struct ib_mad_list_head mad_list;
struct ib_mad_recv_wc recv_wc;
struct ib_wc wc;
- DECLARE_PCI_UNMAP_ADDR(mapping)
+ u64 mapping;
} __attribute__ ((packed));
struct ib_mad_private {
@@ -102,7 +102,7 @@ struct ib_mad_agent_private {
struct list_head send_list;
struct list_head wait_list;
struct list_head done_list;
- struct work_struct timed_work;
+ struct delayed_work timed_work;
unsigned long timeout;
struct list_head local_list;
struct work_struct local_work;
@@ -126,8 +126,8 @@ struct ib_mad_send_wr_private {
struct list_head agent_list;
struct ib_mad_agent_private *mad_agent_priv;
struct ib_mad_send_buf send_buf;
- DECLARE_PCI_UNMAP_ADDR(header_mapping)
- DECLARE_PCI_UNMAP_ADDR(payload_mapping)
+ u64 header_mapping;
+ u64 payload_mapping;
struct ib_send_wr send_wr;
struct ib_sge sg_list[IB_MAD_SEND_REQ_MAX_SG];
__be64 tid;
diff --git a/drivers/infiniband/core/mad_rmpp.c b/drivers/infiniband/core/mad_rmpp.c
index 1ef79d0..3663fd7 100644
--- a/drivers/infiniband/core/mad_rmpp.c
+++ b/drivers/infiniband/core/mad_rmpp.c
@@ -45,8 +45,8 @@ enum rmpp_state {
struct mad_rmpp_recv {
struct ib_mad_agent_private *agent;
struct list_head list;
- struct work_struct timeout_work;
- struct work_struct cleanup_work;
+ struct delayed_work timeout_work;
+ struct delayed_work cleanup_work;
struct completion comp;
enum rmpp_state state;
spinlock_t lock;
@@ -233,9 +233,10 @@ static void nack_recv(struct ib_mad_agent_private *agent,
}
}
-static void recv_timeout_handler(void *data)
+static void recv_timeout_handler(struct work_struct *work)
{
- struct mad_rmpp_recv *rmpp_recv = data;
+ struct mad_rmpp_recv *rmpp_recv =
+ container_of(work, struct mad_rmpp_recv, timeout_work.work);
struct ib_mad_recv_wc *rmpp_wc;
unsigned long flags;
@@ -254,9 +255,10 @@ static void recv_timeout_handler(void *data)
ib_free_recv_mad(rmpp_wc);
}
-static void recv_cleanup_handler(void *data)
+static void recv_cleanup_handler(struct work_struct *work)
{
- struct mad_rmpp_recv *rmpp_recv = data;
+ struct mad_rmpp_recv *rmpp_recv =
+ container_of(work, struct mad_rmpp_recv, cleanup_work.work);
unsigned long flags;
spin_lock_irqsave(&rmpp_recv->agent->lock, flags);
@@ -285,8 +287,8 @@ create_rmpp_recv(struct ib_mad_agent_private *agent,
rmpp_recv->agent = agent;
init_completion(&rmpp_recv->comp);
- INIT_WORK(&rmpp_recv->timeout_work, recv_timeout_handler, rmpp_recv);
- INIT_WORK(&rmpp_recv->cleanup_work, recv_cleanup_handler, rmpp_recv);
+ INIT_DELAYED_WORK(&rmpp_recv->timeout_work, recv_timeout_handler);
+ INIT_DELAYED_WORK(&rmpp_recv->cleanup_work, recv_cleanup_handler);
spin_lock_init(&rmpp_recv->lock);
rmpp_recv->state = RMPP_STATE_ACTIVE;
atomic_set(&rmpp_recv->refcount, 1);
diff --git a/drivers/infiniband/core/sa_query.c b/drivers/infiniband/core/sa_query.c
index 1706d3c..e45afba 100644
--- a/drivers/infiniband/core/sa_query.c
+++ b/drivers/infiniband/core/sa_query.c
@@ -360,9 +360,10 @@ static void free_sm_ah(struct kref *kref)
kfree(sm_ah);
}
-static void update_sm_ah(void *port_ptr)
+static void update_sm_ah(struct work_struct *work)
{
- struct ib_sa_port *port = port_ptr;
+ struct ib_sa_port *port =
+ container_of(work, struct ib_sa_port, update_task);
struct ib_sa_sm_ah *new_ah, *old_ah;
struct ib_port_attr port_attr;
struct ib_ah_attr ah_attr;
@@ -992,8 +993,7 @@ static void ib_sa_add_one(struct ib_device *device)
if (IS_ERR(sa_dev->port[i].agent))
goto err;
- INIT_WORK(&sa_dev->port[i].update_task,
- update_sm_ah, &sa_dev->port[i]);
+ INIT_WORK(&sa_dev->port[i].update_task, update_sm_ah);
}
ib_set_client_data(device, &sa_client, sa_dev);
@@ -1010,7 +1010,7 @@ static void ib_sa_add_one(struct ib_device *device)
goto err;
for (i = 0; i <= e - s; ++i)
- update_sm_ah(&sa_dev->port[i]);
+ update_sm_ah(&sa_dev->port[i].update_task);
return;
diff --git a/drivers/infiniband/core/ucm.c b/drivers/infiniband/core/ucm.c
index ad4f4d5..f15220a 100644
--- a/drivers/infiniband/core/ucm.c
+++ b/drivers/infiniband/core/ucm.c
@@ -161,12 +161,14 @@ static void ib_ucm_cleanup_events(struct ib_ucm_context *ctx)
struct ib_ucm_event, ctx_list);
list_del(&uevent->file_list);
list_del(&uevent->ctx_list);
+ mutex_unlock(&ctx->file->file_mutex);
/* clear incoming connections. */
if (ib_ucm_new_cm_id(uevent->resp.event))
ib_destroy_cm_id(uevent->cm_id);
kfree(uevent);
+ mutex_lock(&ctx->file->file_mutex);
}
mutex_unlock(&ctx->file->file_mutex);
}
@@ -328,20 +330,18 @@ static int ib_ucm_event_process(struct ib_cm_event *evt,
}
if (uvt->data_len) {
- uvt->data = kmalloc(uvt->data_len, GFP_KERNEL);
+ uvt->data = kmemdup(evt->private_data, uvt->data_len, GFP_KERNEL);
if (!uvt->data)
goto err1;
- memcpy(uvt->data, evt->private_data, uvt->data_len);
uvt->resp.present |= IB_UCM_PRES_DATA;
}
if (uvt->info_len) {
- uvt->info = kmalloc(uvt->info_len, GFP_KERNEL);
+ uvt->info = kmemdup(info, uvt->info_len, GFP_KERNEL);
if (!uvt->info)
goto err2;
- memcpy(uvt->info, info, uvt->info_len);
uvt->resp.present |= IB_UCM_PRES_INFO;
}
return 0;
@@ -685,11 +685,11 @@ out:
return result;
}
-static ssize_t ib_ucm_establish(struct ib_ucm_file *file,
- const char __user *inbuf,
- int in_len, int out_len)
+static ssize_t ib_ucm_notify(struct ib_ucm_file *file,
+ const char __user *inbuf,
+ int in_len, int out_len)
{
- struct ib_ucm_establish cmd;
+ struct ib_ucm_notify cmd;
struct ib_ucm_context *ctx;
int result;
@@ -700,7 +700,7 @@ static ssize_t ib_ucm_establish(struct ib_ucm_file *file,
if (IS_ERR(ctx))
return PTR_ERR(ctx);
- result = ib_cm_establish(ctx->cm_id);
+ result = ib_cm_notify(ctx->cm_id, (enum ib_event_type) cmd.event);
ib_ucm_ctx_put(ctx);
return result;
}
@@ -1107,7 +1107,7 @@ static ssize_t (*ucm_cmd_table[])(struct ib_ucm_file *file,
[IB_USER_CM_CMD_DESTROY_ID] = ib_ucm_destroy_id,
[IB_USER_CM_CMD_ATTR_ID] = ib_ucm_attr_id,
[IB_USER_CM_CMD_LISTEN] = ib_ucm_listen,
- [IB_USER_CM_CMD_ESTABLISH] = ib_ucm_establish,
+ [IB_USER_CM_CMD_NOTIFY] = ib_ucm_notify,
[IB_USER_CM_CMD_SEND_REQ] = ib_ucm_send_req,
[IB_USER_CM_CMD_SEND_REP] = ib_ucm_send_rep,
[IB_USER_CM_CMD_SEND_RTU] = ib_ucm_send_rtu,
diff --git a/drivers/infiniband/core/ucma.c b/drivers/infiniband/core/ucma.c
new file mode 100644
index 0000000..e2e8d32
--- /dev/null
+++ b/drivers/infiniband/core/ucma.c
@@ -0,0 +1,885 @@
+/*
+ * Copyright (c) 2005-2006 Intel Corporation. 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/completion.h>
+#include <linux/mutex.h>
+#include <linux/poll.h>
+#include <linux/idr.h>
+#include <linux/in.h>
+#include <linux/in6.h>
+#include <linux/miscdevice.h>
+
+#include <rdma/rdma_user_cm.h>
+#include <rdma/ib_marshall.h>
+#include <rdma/rdma_cm.h>
+
+MODULE_AUTHOR("Sean Hefty");
+MODULE_DESCRIPTION("RDMA Userspace Connection Manager Access");
+MODULE_LICENSE("Dual BSD/GPL");
+
+enum {
+ UCMA_MAX_BACKLOG = 128
+};
+
+struct ucma_file {
+ struct mutex mut;
+ struct file *filp;
+ struct list_head ctx_list;
+ struct list_head event_list;
+ wait_queue_head_t poll_wait;
+};
+
+struct ucma_context {
+ int id;
+ struct completion comp;
+ atomic_t ref;
+ int events_reported;
+ int backlog;
+
+ struct ucma_file *file;
+ struct rdma_cm_id *cm_id;
+ u64 uid;
+
+ struct list_head list;
+};
+
+struct ucma_event {
+ struct ucma_context *ctx;
+ struct list_head list;
+ struct rdma_cm_id *cm_id;
+ struct rdma_ucm_event_resp resp;
+};
+
+static DEFINE_MUTEX(mut);
+static DEFINE_IDR(ctx_idr);
+
+static inline struct ucma_context *_ucma_find_context(int id,
+ struct ucma_file *file)
+{
+ struct ucma_context *ctx;
+
+ ctx = idr_find(&ctx_idr, id);
+ if (!ctx)
+ ctx = ERR_PTR(-ENOENT);
+ else if (ctx->file != file)
+ ctx = ERR_PTR(-EINVAL);
+ return ctx;
+}
+
+static struct ucma_context *ucma_get_ctx(struct ucma_file *file, int id)
+{
+ struct ucma_context *ctx;
+
+ mutex_lock(&mut);
+ ctx = _ucma_find_context(id, file);
+ if (!IS_ERR(ctx))
+ atomic_inc(&ctx->ref);
+ mutex_unlock(&mut);
+ return ctx;
+}
+
+static void ucma_put_ctx(struct ucma_context *ctx)
+{
+ if (atomic_dec_and_test(&ctx->ref))
+ complete(&ctx->comp);
+}
+
+static struct ucma_context *ucma_alloc_ctx(struct ucma_file *file)
+{
+ struct ucma_context *ctx;
+ int ret;
+
+ ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
+ if (!ctx)
+ return NULL;
+
+ atomic_set(&ctx->ref, 1);
+ init_completion(&ctx->comp);
+ ctx->file = file;
+
+ do {
+ ret = idr_pre_get(&ctx_idr, GFP_KERNEL);
+ if (!ret)
+ goto error;
+
+ mutex_lock(&mut);
+ ret = idr_get_new(&ctx_idr, ctx, &ctx->id);
+ mutex_unlock(&mut);
+ } while (ret == -EAGAIN);
+
+ if (ret)
+ goto error;
+
+ list_add_tail(&ctx->list, &file->ctx_list);
+ return ctx;
+
+error:
+ kfree(ctx);
+ return NULL;
+}
+
+static void ucma_copy_conn_event(struct rdma_ucm_conn_param *dst,
+ struct rdma_conn_param *src)
+{
+ if (src->private_data_len)
+ memcpy(dst->private_data, src->private_data,
+ src->private_data_len);
+ dst->private_data_len = src->private_data_len;
+ dst->responder_resources =src->responder_resources;
+ dst->initiator_depth = src->initiator_depth;
+ dst->flow_control = src->flow_control;
+ dst->retry_count = src->retry_count;
+ dst->rnr_retry_count = src->rnr_retry_count;
+ dst->srq = src->srq;
+ dst->qp_num = src->qp_num;
+}
+
+static void ucma_copy_ud_event(struct rdma_ucm_ud_param *dst,
+ struct rdma_ud_param *src)
+{
+ if (src->private_data_len)
+ memcpy(dst->private_data, src->private_data,
+ src->private_data_len);
+ dst->private_data_len = src->private_data_len;
+ ib_copy_ah_attr_to_user(&dst->ah_attr, &src->ah_attr);
+ dst->qp_num = src->qp_num;
+ dst->qkey = src->qkey;
+}
+
+static void ucma_set_event_context(struct ucma_context *ctx,
+ struct rdma_cm_event *event,
+ struct ucma_event *uevent)
+{
+ uevent->ctx = ctx;
+ uevent->resp.uid = ctx->uid;
+ uevent->resp.id = ctx->id;
+}
+
+static int ucma_event_handler(struct rdma_cm_id *cm_id,
+ struct rdma_cm_event *event)
+{
+ struct ucma_event *uevent;
+ struct ucma_context *ctx = cm_id->context;
+ int ret = 0;
+
+ uevent = kzalloc(sizeof(*uevent), GFP_KERNEL);
+ if (!uevent)
+ return event->event == RDMA_CM_EVENT_CONNECT_REQUEST;
+
+ uevent->cm_id = cm_id;
+ ucma_set_event_context(ctx, event, uevent);
+ uevent->resp.event = event->event;
+ uevent->resp.status = event->status;
+ if (cm_id->ps == RDMA_PS_UDP)
+ ucma_copy_ud_event(&uevent->resp.param.ud, &event->param.ud);
+ else
+ ucma_copy_conn_event(&uevent->resp.param.conn,
+ &event->param.conn);
+
+ mutex_lock(&ctx->file->mut);
+ if (event->event == RDMA_CM_EVENT_CONNECT_REQUEST) {
+ if (!ctx->backlog) {
+ ret = -EDQUOT;
+ kfree(uevent);
+ goto out;
+ }
+ ctx->backlog--;
+ } else if (!ctx->uid) {
+ /*
+ * We ignore events for new connections until userspace has set
+ * their context. This can only happen if an error occurs on a
+ * new connection before the user accepts it. This is okay,
+ * since the accept will just fail later.
+ */
+ kfree(uevent);
+ goto out;
+ }
+
+ list_add_tail(&uevent->list, &ctx->file->event_list);
+ wake_up_interruptible(&ctx->file->poll_wait);
+out:
+ mutex_unlock(&ctx->file->mut);
+ return ret;
+}
+
+static ssize_t ucma_get_event(struct ucma_file *file, const char __user *inbuf,
+ int in_len, int out_len)
+{
+ struct ucma_context *ctx;
+ struct rdma_ucm_get_event cmd;
+ struct ucma_event *uevent;
+ int ret = 0;
+ DEFINE_WAIT(wait);
+
+ if (out_len < sizeof uevent->resp)
+ return -ENOSPC;
+
+ if (copy_from_user(&cmd, inbuf, sizeof(cmd)))
+ return -EFAULT;
+
+ mutex_lock(&file->mut);
+ while (list_empty(&file->event_list)) {
+ if (file->filp->f_flags & O_NONBLOCK) {
+ ret = -EAGAIN;
+ break;
+ }
+
+ if (signal_pending(current)) {
+ ret = -ERESTARTSYS;
+ break;
+ }
+
+ prepare_to_wait(&file->poll_wait, &wait, TASK_INTERRUPTIBLE);
+ mutex_unlock(&file->mut);
+ schedule();
+ mutex_lock(&file->mut);
+ finish_wait(&file->poll_wait, &wait);
+ }
+
+ if (ret)
+ goto done;
+
+ uevent = list_entry(file->event_list.next, struct ucma_event, list);
+
+ if (uevent->resp.event == RDMA_CM_EVENT_CONNECT_REQUEST) {
+ ctx = ucma_alloc_ctx(file);
+ if (!ctx) {
+ ret = -ENOMEM;
+ goto done;
+ }
+ uevent->ctx->backlog++;
+ ctx->cm_id = uevent->cm_id;
+ ctx->cm_id->context = ctx;
+ uevent->resp.id = ctx->id;
+ }
+
+ if (copy_to_user((void __user *)(unsigned long)cmd.response,
+ &uevent->resp, sizeof uevent->resp)) {
+ ret = -EFAULT;
+ goto done;
+ }
+
+ list_del(&uevent->list);
+ uevent->ctx->events_reported++;
+ kfree(uevent);
+done:
+ mutex_unlock(&file->mut);
+ return ret;
+}
+
+static ssize_t ucma_create_id(struct ucma_file *file,
+ const char __user *inbuf,
+ int in_len, int out_len)
+{
+ struct rdma_ucm_create_id cmd;
+ struct rdma_ucm_create_id_resp resp;
+ struct ucma_context *ctx;
+ int ret;
+
+ if (out_len < sizeof(resp))
+ return -ENOSPC;
+
+ if (copy_from_user(&cmd, inbuf, sizeof(cmd)))
+ return -EFAULT;
+
+ mutex_lock(&file->mut);
+ ctx = ucma_alloc_ctx(file);
+ mutex_unlock(&file->mut);
+ if (!ctx)
+ return -ENOMEM;
+
+ ctx->uid = cmd.uid;
+ ctx->cm_id = rdma_create_id(ucma_event_handler, ctx, cmd.ps);
+ if (IS_ERR(ctx->cm_id)) {
+ ret = PTR_ERR(ctx->cm_id);
+ goto err1;
+ }
+
+ resp.id = ctx->id;
+ if (copy_to_user((void __user *)(unsigned long)cmd.response,
+ &resp, sizeof(resp))) {
+ ret = -EFAULT;
+ goto err2;
+ }
+ return 0;
+
+err2:
+ rdma_destroy_id(ctx->cm_id);
+err1:
+ mutex_lock(&mut);
+ idr_remove(&ctx_idr, ctx->id);
+ mutex_unlock(&mut);
+ kfree(ctx);
+ return ret;
+}
+
+static void ucma_cleanup_events(struct ucma_context *ctx)
+{
+ struct ucma_event *uevent, *tmp;
+
+ list_for_each_entry_safe(uevent, tmp, &ctx->file->event_list, list) {
+ if (uevent->ctx != ctx)
+ continue;
+
+ list_del(&uevent->list);
+
+ /* clear incoming connections. */
+ if (uevent->resp.event == RDMA_CM_EVENT_CONNECT_REQUEST)
+ rdma_destroy_id(uevent->cm_id);
+
+ kfree(uevent);
+ }
+}
+
+static int ucma_free_ctx(struct ucma_context *ctx)
+{
+ int events_reported;
+
+ /* No new events will be generated after destroying the id. */
+ rdma_destroy_id(ctx->cm_id);
+
+ /* Cleanup events not yet reported to the user. */
+ mutex_lock(&ctx->file->mut);
+ ucma_cleanup_events(ctx);
+ list_del(&ctx->list);
+ mutex_unlock(&ctx->file->mut);
+
+ events_reported = ctx->events_reported;
+ kfree(ctx);
+ return events_reported;
+}
+
+static ssize_t ucma_destroy_id(struct ucma_file *file, const char __user *inbuf,
+ int in_len, int out_len)
+{
+ struct rdma_ucm_destroy_id cmd;
+ struct rdma_ucm_destroy_id_resp resp;
+ struct ucma_context *ctx;
+ int ret = 0;
+
+ if (out_len < sizeof(resp))
+ return -ENOSPC;
+
+ if (copy_from_user(&cmd, inbuf, sizeof(cmd)))
+ return -EFAULT;
+
+ mutex_lock(&mut);
+ ctx = _ucma_find_context(cmd.id, file);
+ if (!IS_ERR(ctx))
+ idr_remove(&ctx_idr, ctx->id);
+ mutex_unlock(&mut);
+
+ if (IS_ERR(ctx))
+ return PTR_ERR(ctx);
+
+ ucma_put_ctx(ctx);
+ wait_for_completion(&ctx->comp);
+ resp.events_reported = ucma_free_ctx(ctx);
+
+ if (copy_to_user((void __user *)(unsigned long)cmd.response,
+ &resp, sizeof(resp)))
+ ret = -EFAULT;
+
+ return ret;
+}
+
+static ssize_t ucma_bind_addr(struct ucma_file *file, const char __user *inbuf,
+ int in_len, int out_len)
+{
+ struct rdma_ucm_bind_addr cmd;
+ struct ucma_context *ctx;
+ int ret;
+
+ if (copy_from_user(&cmd, inbuf, sizeof(cmd)))
+ return -EFAULT;
+
+ ctx = ucma_get_ctx(file, cmd.id);
+ if (IS_ERR(ctx))
+ return PTR_ERR(ctx);
+
+ ret = rdma_bind_addr(ctx->cm_id, (struct sockaddr *) &cmd.addr);
+ ucma_put_ctx(ctx);
+ return ret;
+}
+
+static ssize_t ucma_resolve_addr(struct ucma_file *file,
+ const char __user *inbuf,
+ int in_len, int out_len)
+{
+ struct rdma_ucm_resolve_addr cmd;
+ struct ucma_context *ctx;
+ int ret;
+
+ if (copy_from_user(&cmd, inbuf, sizeof(cmd)))
+ return -EFAULT;
+
+ ctx = ucma_get_ctx(file, cmd.id);
+ if (IS_ERR(ctx))
+ return PTR_ERR(ctx);
+
+ ret = rdma_resolve_addr(ctx->cm_id, (struct sockaddr *) &cmd.src_addr,
+ (struct sockaddr *) &cmd.dst_addr,
+ cmd.timeout_ms);
+ ucma_put_ctx(ctx);
+ return ret;
+}
+
+static ssize_t ucma_resolve_route(struct ucma_file *file,
+ const char __user *inbuf,
+ int in_len, int out_len)
+{
+ struct rdma_ucm_resolve_route cmd;
+ struct ucma_context *ctx;
+ int ret;
+
+ if (copy_from_user(&cmd, inbuf, sizeof(cmd)))
+ return -EFAULT;
+
+ ctx = ucma_get_ctx(file, cmd.id);
+ if (IS_ERR(ctx))
+ return PTR_ERR(ctx);
+
+ ret = rdma_resolve_route(ctx->cm_id, cmd.timeout_ms);
+ ucma_put_ctx(ctx);
+ return ret;
+}
+
+static void ucma_copy_ib_route(struct rdma_ucm_query_route_resp *resp,
+ struct rdma_route *route)
+{
+ struct rdma_dev_addr *dev_addr;
+
+ resp->num_paths = route->num_paths;
+ switch (route->num_paths) {
+ case 0:
+ dev_addr = &route->addr.dev_addr;
+ ib_addr_get_dgid(dev_addr,
+ (union ib_gid *) &resp->ib_route[0].dgid);
+ ib_addr_get_sgid(dev_addr,
+ (union ib_gid *) &resp->ib_route[0].sgid);
+ resp->ib_route[0].pkey = cpu_to_be16(ib_addr_get_pkey(dev_addr));
+ break;
+ case 2:
+ ib_copy_path_rec_to_user(&resp->ib_route[1],
+ &route->path_rec[1]);
+ /* fall through */
+ case 1:
+ ib_copy_path_rec_to_user(&resp->ib_route[0],
+ &route->path_rec[0]);
+ break;
+ default:
+ break;
+ }
+}
+
+static ssize_t ucma_query_route(struct ucma_file *file,
+ const char __user *inbuf,
+ int in_len, int out_len)
+{
+ struct rdma_ucm_query_route cmd;
+ struct rdma_ucm_query_route_resp resp;
+ struct ucma_context *ctx;
+ struct sockaddr *addr;
+ int ret = 0;
+
+ if (out_len < sizeof(resp))
+ return -ENOSPC;
+
+ if (copy_from_user(&cmd, inbuf, sizeof(cmd)))
+ return -EFAULT;
+
+ ctx = ucma_get_ctx(file, cmd.id);
+ if (IS_ERR(ctx))
+ return PTR_ERR(ctx);
+
+ memset(&resp, 0, sizeof resp);
+ addr = &ctx->cm_id->route.addr.src_addr;
+ memcpy(&resp.src_addr, addr, addr->sa_family == AF_INET ?
+ sizeof(struct sockaddr_in) :
+ sizeof(struct sockaddr_in6));
+ addr = &ctx->cm_id->route.addr.dst_addr;
+ memcpy(&resp.dst_addr, addr, addr->sa_family == AF_INET ?
+ sizeof(struct sockaddr_in) :
+ sizeof(struct sockaddr_in6));
+ if (!ctx->cm_id->device)
+ goto out;
+
+ resp.node_guid = ctx->cm_id->device->node_guid;
+ resp.port_num = ctx->cm_id->port_num;
+ switch (rdma_node_get_transport(ctx->cm_id->device->node_type)) {
+ case RDMA_TRANSPORT_IB:
+ ucma_copy_ib_route(&resp, &ctx->cm_id->route);
+ break;
+ default:
+ break;
+ }
+
+out:
+ if (copy_to_user((void __user *)(unsigned long)cmd.response,
+ &resp, sizeof(resp)))
+ ret = -EFAULT;
+
+ ucma_put_ctx(ctx);
+ return ret;
+}
+
+static void ucma_copy_conn_param(struct rdma_conn_param *dst,
+ struct rdma_ucm_conn_param *src)
+{
+ dst->private_data = src->private_data;
+ dst->private_data_len = src->private_data_len;
+ dst->responder_resources =src->responder_resources;
+ dst->initiator_depth = src->initiator_depth;
+ dst->flow_control = src->flow_control;
+ dst->retry_count = src->retry_count;
+ dst->rnr_retry_count = src->rnr_retry_count;
+ dst->srq = src->srq;
+ dst->qp_num = src->qp_num;
+}
+
+static ssize_t ucma_connect(struct ucma_file *file, const char __user *inbuf,
+ int in_len, int out_len)
+{
+ struct rdma_ucm_connect cmd;
+ struct rdma_conn_param conn_param;
+ struct ucma_context *ctx;
+ int ret;
+
+ if (copy_from_user(&cmd, inbuf, sizeof(cmd)))
+ return -EFAULT;
+
+ if (!cmd.conn_param.valid)
+ return -EINVAL;
+
+ ctx = ucma_get_ctx(file, cmd.id);
+ if (IS_ERR(ctx))
+ return PTR_ERR(ctx);
+
+ ucma_copy_conn_param(&conn_param, &cmd.conn_param);
+ ret = rdma_connect(ctx->cm_id, &conn_param);
+ ucma_put_ctx(ctx);
+ return ret;
+}
+
+static ssize_t ucma_listen(struct ucma_file *file, const char __user *inbuf,
+ int in_len, int out_len)
+{
+ struct rdma_ucm_listen cmd;
+ struct ucma_context *ctx;
+ int ret;
+
+ if (copy_from_user(&cmd, inbuf, sizeof(cmd)))
+ return -EFAULT;
+
+ ctx = ucma_get_ctx(file, cmd.id);
+ if (IS_ERR(ctx))
+ return PTR_ERR(ctx);
+
+ ctx->backlog = cmd.backlog > 0 && cmd.backlog < UCMA_MAX_BACKLOG ?
+ cmd.backlog : UCMA_MAX_BACKLOG;
+ ret = rdma_listen(ctx->cm_id, ctx->backlog);
+ ucma_put_ctx(ctx);
+ return ret;
+}
+
+static ssize_t ucma_accept(struct ucma_file *file, const char __user *inbuf,
+ int in_len, int out_len)
+{
+ struct rdma_ucm_accept cmd;
+ struct rdma_conn_param conn_param;
+ struct ucma_context *ctx;
+ int ret;
+
+ if (copy_from_user(&cmd, inbuf, sizeof(cmd)))
+ return -EFAULT;
+
+ ctx = ucma_get_ctx(file, cmd.id);
+ if (IS_ERR(ctx))
+ return PTR_ERR(ctx);
+
+ if (cmd.conn_param.valid) {
+ ctx->uid = cmd.uid;
+ ucma_copy_conn_param(&conn_param, &cmd.conn_param);
+ ret = rdma_accept(ctx->cm_id, &conn_param);
+ } else
+ ret = rdma_accept(ctx->cm_id, NULL);
+
+ ucma_put_ctx(ctx);
+ return ret;
+}
+
+static ssize_t ucma_reject(struct ucma_file *file, const char __user *inbuf,
+ int in_len, int out_len)
+{
+ struct rdma_ucm_reject cmd;
+ struct ucma_context *ctx;
+ int ret;
+
+ if (copy_from_user(&cmd, inbuf, sizeof(cmd)))
+ return -EFAULT;
+
+ ctx = ucma_get_ctx(file, cmd.id);
+ if (IS_ERR(ctx))
+ return PTR_ERR(ctx);
+
+ ret = rdma_reject(ctx->cm_id, cmd.private_data, cmd.private_data_len);
+ ucma_put_ctx(ctx);
+ return ret;
+}
+
+static ssize_t ucma_disconnect(struct ucma_file *file, const char __user *inbuf,
+ int in_len, int out_len)
+{
+ struct rdma_ucm_disconnect cmd;
+ struct ucma_context *ctx;
+ int ret;
+
+ if (copy_from_user(&cmd, inbuf, sizeof(cmd)))
+ return -EFAULT;
+
+ ctx = ucma_get_ctx(file, cmd.id);
+ if (IS_ERR(ctx))
+ return PTR_ERR(ctx);
+
+ ret = rdma_disconnect(ctx->cm_id);
+ ucma_put_ctx(ctx);
+ return ret;
+}
+
+static ssize_t ucma_init_qp_attr(struct ucma_file *file,
+ const char __user *inbuf,
+ int in_len, int out_len)
+{
+ struct rdma_ucm_init_qp_attr cmd;
+ struct ib_uverbs_qp_attr resp;
+ struct ucma_context *ctx;
+ struct ib_qp_attr qp_attr;
+ int ret;
+
+ if (out_len < sizeof(resp))
+ return -ENOSPC;
+
+ if (copy_from_user(&cmd, inbuf, sizeof(cmd)))
+ return -EFAULT;
+
+ ctx = ucma_get_ctx(file, cmd.id);
+ if (IS_ERR(ctx))
+ return PTR_ERR(ctx);
+
+ resp.qp_attr_mask = 0;
+ memset(&qp_attr, 0, sizeof qp_attr);
+ qp_attr.qp_state = cmd.qp_state;
+ ret = rdma_init_qp_attr(ctx->cm_id, &qp_attr, &resp.qp_attr_mask);
+ if (ret)
+ goto out;
+
+ ib_copy_qp_attr_to_user(&resp, &qp_attr);
+ if (copy_to_user((void __user *)(unsigned long)cmd.response,
+ &resp, sizeof(resp)))
+ ret = -EFAULT;
+
+out:
+ ucma_put_ctx(ctx);
+ return ret;
+}
+
+static ssize_t ucma_notify(struct ucma_file *file, const char __user *inbuf,
+ int in_len, int out_len)
+{
+ struct rdma_ucm_notify cmd;
+ struct ucma_context *ctx;
+ int ret;
+
+ if (copy_from_user(&cmd, inbuf, sizeof(cmd)))
+ return -EFAULT;
+
+ ctx = ucma_get_ctx(file, cmd.id);
+ if (IS_ERR(ctx))
+ return PTR_ERR(ctx);
+
+ ret = rdma_notify(ctx->cm_id, (enum ib_event_type) cmd.event);
+ ucma_put_ctx(ctx);
+ return ret;
+}
+
+static ssize_t (*ucma_cmd_table[])(struct ucma_file *file,
+ const char __user *inbuf,
+ int in_len, int out_len) = {
+ [RDMA_USER_CM_CMD_CREATE_ID] = ucma_create_id,
+ [RDMA_USER_CM_CMD_DESTROY_ID] = ucma_destroy_id,
+ [RDMA_USER_CM_CMD_BIND_ADDR] = ucma_bind_addr,
+ [RDMA_USER_CM_CMD_RESOLVE_ADDR] = ucma_resolve_addr,
+ [RDMA_USER_CM_CMD_RESOLVE_ROUTE]= ucma_resolve_route,
+ [RDMA_USER_CM_CMD_QUERY_ROUTE] = ucma_query_route,
+ [RDMA_USER_CM_CMD_CONNECT] = ucma_connect,
+ [RDMA_USER_CM_CMD_LISTEN] = ucma_listen,
+ [RDMA_USER_CM_CMD_ACCEPT] = ucma_accept,
+ [RDMA_USER_CM_CMD_REJECT] = ucma_reject,
+ [RDMA_USER_CM_CMD_DISCONNECT] = ucma_disconnect,
+ [RDMA_USER_CM_CMD_INIT_QP_ATTR] = ucma_init_qp_attr,
+ [RDMA_USER_CM_CMD_GET_EVENT] = ucma_get_event,
+ [RDMA_USER_CM_CMD_GET_OPTION] = NULL,
+ [RDMA_USER_CM_CMD_SET_OPTION] = NULL,
+ [RDMA_USER_CM_CMD_NOTIFY] = ucma_notify,
+};
+
+static ssize_t ucma_write(struct file *filp, const char __user *buf,
+ size_t len, loff_t *pos)
+{
+ struct ucma_file *file = filp->private_data;
+ struct rdma_ucm_cmd_hdr hdr;
+ ssize_t ret;
+
+ if (len < sizeof(hdr))
+ return -EINVAL;
+
+ if (copy_from_user(&hdr, buf, sizeof(hdr)))
+ return -EFAULT;
+
+ if (hdr.cmd < 0 || hdr.cmd >= ARRAY_SIZE(ucma_cmd_table))
+ return -EINVAL;
+
+ if (hdr.in + sizeof(hdr) > len)
+ return -EINVAL;
+
+ if (!ucma_cmd_table[hdr.cmd])
+ return -ENOSYS;
+
+ ret = ucma_cmd_table[hdr.cmd](file, buf + sizeof(hdr), hdr.in, hdr.out);
+ if (!ret)
+ ret = len;
+
+ return ret;
+}
+
+static unsigned int ucma_poll(struct file *filp, struct poll_table_struct *wait)
+{
+ struct ucma_file *file = filp->private_data;
+ unsigned int mask = 0;
+
+ poll_wait(filp, &file->poll_wait, wait);
+
+ if (!list_empty(&file->event_list))
+ mask = POLLIN | POLLRDNORM;
+
+ return mask;
+}
+
+static int ucma_open(struct inode *inode, struct file *filp)
+{
+ struct ucma_file *file;
+
+ file = kmalloc(sizeof *file, GFP_KERNEL);
+ if (!file)
+ return -ENOMEM;
+
+ INIT_LIST_HEAD(&file->event_list);
+ INIT_LIST_HEAD(&file->ctx_list);
+ init_waitqueue_head(&file->poll_wait);
+ mutex_init(&file->mut);
+
+ filp->private_data = file;
+ file->filp = filp;
+ return 0;
+}
+
+static int ucma_close(struct inode *inode, struct file *filp)
+{
+ struct ucma_file *file = filp->private_data;
+ struct ucma_context *ctx, *tmp;
+
+ mutex_lock(&file->mut);
+ list_for_each_entry_safe(ctx, tmp, &file->ctx_list, list) {
+ mutex_unlock(&file->mut);
+
+ mutex_lock(&mut);
+ idr_remove(&ctx_idr, ctx->id);
+ mutex_unlock(&mut);
+
+ ucma_free_ctx(ctx);
+ mutex_lock(&file->mut);
+ }
+ mutex_unlock(&file->mut);
+ kfree(file);
+ return 0;
+}
+
+static struct file_operations ucma_fops = {
+ .owner = THIS_MODULE,
+ .open = ucma_open,
+ .release = ucma_close,
+ .write = ucma_write,
+ .poll = ucma_poll,
+};
+
+static struct miscdevice ucma_misc = {
+ .minor = MISC_DYNAMIC_MINOR,
+ .name = "rdma_cm",
+ .fops = &ucma_fops,
+};
+
+static ssize_t show_abi_version(struct device *dev,
+ struct device_attribute *attr,
+ char *buf)
+{
+ return sprintf(buf, "%d\n", RDMA_USER_CM_ABI_VERSION);
+}
+static DEVICE_ATTR(abi_version, S_IRUGO, show_abi_version, NULL);
+
+static int __init ucma_init(void)
+{
+ int ret;
+
+ ret = misc_register(&ucma_misc);
+ if (ret)
+ return ret;
+
+ ret = device_create_file(ucma_misc.this_device, &dev_attr_abi_version);
+ if (ret) {
+ printk(KERN_ERR "rdma_ucm: couldn't create abi_version attr\n");
+ goto err;
+ }
+ return 0;
+err:
+ misc_deregister(&ucma_misc);
+ return ret;
+}
+
+static void __exit ucma_cleanup(void)
+{
+ device_remove_file(ucma_misc.this_device, &dev_attr_abi_version);
+ misc_deregister(&ucma_misc);
+ idr_destroy(&ctx_idr);
+}
+
+module_init(ucma_init);
+module_exit(ucma_cleanup);
diff --git a/drivers/infiniband/core/uverbs_main.c b/drivers/infiniband/core/uverbs_main.c
index 4e16314..a617ca7 100644
--- a/drivers/infiniband/core/uverbs_main.c
+++ b/drivers/infiniband/core/uverbs_main.c
@@ -534,9 +534,9 @@ struct file *ib_uverbs_alloc_event_file(struct ib_uverbs_file *uverbs_file,
* module reference.
*/
filp->f_op = fops_get(&uverbs_event_fops);
- filp->f_vfsmnt = mntget(uverbs_event_mnt);
- filp->f_dentry = dget(uverbs_event_mnt->mnt_root);
- filp->f_mapping = filp->f_dentry->d_inode->i_mapping;
+ filp->f_path.mnt = mntget(uverbs_event_mnt);
+ filp->f_path.dentry = dget(uverbs_event_mnt->mnt_root);
+ filp->f_mapping = filp->f_path.dentry->d_inode->i_mapping;
filp->f_flags = O_RDONLY;
filp->f_mode = FMODE_READ;
filp->private_data = ev_file;
diff --git a/drivers/infiniband/core/uverbs_marshall.c b/drivers/infiniband/core/uverbs_marshall.c
index ce46b13..5440da0 100644
--- a/drivers/infiniband/core/uverbs_marshall.c
+++ b/drivers/infiniband/core/uverbs_marshall.c
@@ -32,8 +32,8 @@
#include <rdma/ib_marshall.h>
-static void ib_copy_ah_attr_to_user(struct ib_uverbs_ah_attr *dst,
- struct ib_ah_attr *src)
+void ib_copy_ah_attr_to_user(struct ib_uverbs_ah_attr *dst,
+ struct ib_ah_attr *src)
{
memcpy(dst->grh.dgid, src->grh.dgid.raw, sizeof src->grh.dgid);
dst->grh.flow_label = src->grh.flow_label;
@@ -47,6 +47,7 @@ static void ib_copy_ah_attr_to_user(struct ib_uverbs_ah_attr *dst,
dst->is_global = src->ah_flags & IB_AH_GRH ? 1 : 0;
dst->port_num = src->port_num;
}
+EXPORT_SYMBOL(ib_copy_ah_attr_to_user);
void ib_copy_qp_attr_to_user(struct ib_uverbs_qp_attr *dst,
struct ib_qp_attr *src)
diff --git a/drivers/infiniband/core/uverbs_mem.c b/drivers/infiniband/core/uverbs_mem.c
index efe147d..c95fe952 100644
--- a/drivers/infiniband/core/uverbs_mem.c
+++ b/drivers/infiniband/core/uverbs_mem.c
@@ -52,8 +52,8 @@ static void __ib_umem_release(struct ib_device *dev, struct ib_umem *umem, int d
int i;
list_for_each_entry_safe(chunk, tmp, &umem->chunk_list, list) {
- dma_unmap_sg(dev->dma_device, chunk->page_list,
- chunk->nents, DMA_BIDIRECTIONAL);
+ ib_dma_unmap_sg(dev, chunk->page_list,
+ chunk->nents, DMA_BIDIRECTIONAL);
for (i = 0; i < chunk->nents; ++i) {
if (umem->writable && dirty)
set_page_dirty_lock(chunk->page_list[i].page);
@@ -136,10 +136,10 @@ int ib_umem_get(struct ib_device *dev, struct ib_umem *mem,
chunk->page_list[i].length = PAGE_SIZE;
}
- chunk->nmap = dma_map_sg(dev->dma_device,
- &chunk->page_list[0],
- chunk->nents,
- DMA_BIDIRECTIONAL);
+ chunk->nmap = ib_dma_map_sg(dev,
+ &chunk->page_list[0],
+ chunk->nents,
+ DMA_BIDIRECTIONAL);
if (chunk->nmap <= 0) {
for (i = 0; i < chunk->nents; ++i)
put_page(chunk->page_list[i].page);
@@ -179,9 +179,10 @@ void ib_umem_release(struct ib_device *dev, struct ib_umem *umem)
up_write(&current->mm->mmap_sem);
}
-static void ib_umem_account(void *work_ptr)
+static void ib_umem_account(struct work_struct *_work)
{
- struct ib_umem_account_work *work = work_ptr;
+ struct ib_umem_account_work *work =
+ container_of(_work, struct ib_umem_account_work, work);
down_write(&work->mm->mmap_sem);
work->mm->locked_vm -= work->diff;
@@ -216,7 +217,7 @@ void ib_umem_release_on_close(struct ib_device *dev, struct ib_umem *umem)
return;
}
- INIT_WORK(&work->work, ib_umem_account, work);
+ INIT_WORK(&work->work, ib_umem_account);
work->mm = mm;
work->diff = PAGE_ALIGN(umem->length + umem->offset) >> PAGE_SHIFT;
OpenPOWER on IntegriCloud