summaryrefslogtreecommitdiffstats
path: root/drivers/infiniband
diff options
context:
space:
mode:
authorDmitry Torokhov <dtor@insightbb.com>2007-02-10 01:26:32 -0500
committerDmitry Torokhov <dtor@insightbb.com>2007-02-10 01:26:32 -0500
commitb22364c8eec89e6b0c081a237f3b6348df87796f (patch)
tree233a923281fb640106465d076997ff511efb6edf /drivers/infiniband
parent2c8dc071517ec2843869024dc82be2e246f41064 (diff)
parent66efc5a7e3061c3597ac43a8bb1026488d57e66b (diff)
Merge rsync://rsync.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6
Diffstat (limited to 'drivers/infiniband')
-rw-r--r--drivers/infiniband/core/Makefile6
-rw-r--r--drivers/infiniband/core/addr.c3
-rw-r--r--drivers/infiniband/core/cm.c4
-rw-r--r--drivers/infiniband/core/cma.c431
-rw-r--r--drivers/infiniband/core/fmr_pool.c12
-rw-r--r--drivers/infiniband/core/mad.c101
-rw-r--r--drivers/infiniband/core/mad_priv.h6
-rw-r--r--drivers/infiniband/core/ucma.c885
-rw-r--r--drivers/infiniband/core/uverbs_cmd.c2
-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.c12
-rw-r--r--drivers/infiniband/hw/amso1100/c2_cq.c2
-rw-r--r--drivers/infiniband/hw/amso1100/c2_qp.c13
-rw-r--r--drivers/infiniband/hw/ehca/ehca_classes.h29
-rw-r--r--drivers/infiniband/hw/ehca/ehca_cq.c70
-rw-r--r--drivers/infiniband/hw/ehca/ehca_hca.c8
-rw-r--r--drivers/infiniband/hw/ehca/ehca_irq.c5
-rw-r--r--drivers/infiniband/hw/ehca/ehca_iverbs.h12
-rw-r--r--drivers/infiniband/hw/ehca/ehca_main.c16
-rw-r--r--drivers/infiniband/hw/ehca/ehca_mrmw.c4
-rw-r--r--drivers/infiniband/hw/ehca/ehca_qp.c82
-rw-r--r--drivers/infiniband/hw/ehca/ehca_reqs.c2
-rw-r--r--drivers/infiniband/hw/ehca/ehca_uverbs.c395
-rw-r--r--drivers/infiniband/hw/ipath/Makefile1
-rw-r--r--drivers/infiniband/hw/ipath/ipath_dma.c189
-rw-r--r--drivers/infiniband/hw/ipath/ipath_driver.c4
-rw-r--r--drivers/infiniband/hw/ipath/ipath_file_ops.c9
-rw-r--r--drivers/infiniband/hw/ipath/ipath_fs.c10
-rw-r--r--drivers/infiniband/hw/ipath/ipath_iba6110.c3
-rw-r--r--drivers/infiniband/hw/ipath/ipath_iba6120.c8
-rw-r--r--drivers/infiniband/hw/ipath/ipath_init_chip.c3
-rw-r--r--drivers/infiniband/hw/ipath/ipath_intr.c3
-rw-r--r--drivers/infiniband/hw/ipath/ipath_keys.c8
-rw-r--r--drivers/infiniband/hw/ipath/ipath_mr.c7
-rw-r--r--drivers/infiniband/hw/ipath/ipath_qp.c2
-rw-r--r--drivers/infiniband/hw/ipath/ipath_rc.c8
-rw-r--r--drivers/infiniband/hw/ipath/ipath_ruc.c8
-rw-r--r--drivers/infiniband/hw/ipath/ipath_sysfs.c3
-rw-r--r--drivers/infiniband/hw/ipath/ipath_uc.c4
-rw-r--r--drivers/infiniband/hw/ipath/ipath_ud.c8
-rw-r--r--drivers/infiniband/hw/ipath/ipath_verbs.c1
-rw-r--r--drivers/infiniband/hw/ipath/ipath_verbs.h2
-rw-r--r--drivers/infiniband/hw/mthca/mthca_cmd.c2
-rw-r--r--drivers/infiniband/hw/mthca/mthca_cq.c10
-rw-r--r--drivers/infiniband/hw/mthca/mthca_main.c113
-rw-r--r--drivers/infiniband/hw/mthca/mthca_memfree.c2
-rw-r--r--drivers/infiniband/hw/mthca/mthca_provider.c7
-rw-r--r--drivers/infiniband/hw/mthca/mthca_qp.c30
-rw-r--r--drivers/infiniband/hw/mthca/mthca_srq.c4
-rw-r--r--drivers/infiniband/ulp/ipoib/ipoib.h4
-rw-r--r--drivers/infiniband/ulp/ipoib/ipoib_ib.c75
-rw-r--r--drivers/infiniband/ulp/ipoib/ipoib_main.c36
-rw-r--r--drivers/infiniband/ulp/ipoib/ipoib_vlan.c11
-rw-r--r--drivers/infiniband/ulp/iser/iscsi_iser.c4
-rw-r--r--drivers/infiniband/ulp/iser/iscsi_iser.h2
-rw-r--r--drivers/infiniband/ulp/iser/iser_initiator.c34
-rw-r--r--drivers/infiniband/ulp/iser/iser_memory.c129
-rw-r--r--drivers/infiniband/ulp/srp/ib_srp.c110
-rw-r--r--drivers/infiniband/ulp/srp/ib_srp.h5
60 files changed, 2119 insertions, 851 deletions
diff --git a/drivers/infiniband/core/Makefile b/drivers/infiniband/core/Makefile
index 163d991eb8c..50fb1cd447b 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 af939796750..d2bb5a9a303 100644
--- a/drivers/infiniband/core/addr.c
+++ b/drivers/infiniband/core/addr.c
@@ -360,8 +360,7 @@ static int netevent_callback(struct notifier_block *self, unsigned long event,
if (event == NETEVENT_NEIGH_UPDATE) {
struct neighbour *neigh = ctx;
- if (neigh->dev->type == ARPHRD_INFINIBAND &&
- (neigh->nud_state & NUD_VALID)) {
+ if (neigh->nud_state & NUD_VALID) {
set_timeout(jiffies);
}
}
diff --git a/drivers/infiniband/core/cm.c b/drivers/infiniband/core/cm.c
index 79c937bf696..d446998b12a 100644
--- a/drivers/infiniband/core/cm.c
+++ b/drivers/infiniband/core/cm.c
@@ -3289,6 +3289,10 @@ 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:
diff --git a/drivers/infiniband/core/cma.c b/drivers/infiniband/core/cma.c
index 985a6b564d8..9e0ab048c87 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;
};
@@ -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;
@@ -942,10 +997,7 @@ static int cma_req_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event)
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);
+ ret = conn_id->id.event_handler(&conn_id->id, &event);
if (!ret)
goto out;
@@ -964,8 +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,
@@ -1021,36 +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;
@@ -1071,6 +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;
@@ -1124,9 +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;
@@ -1515,8 +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);
/*
@@ -1536,14 +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);
@@ -1733,6 +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;
}
@@ -1821,6 +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)
{
@@ -1860,7 +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;
@@ -1937,13 +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);
@@ -1966,11 +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;
@@ -1985,7 +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,
@@ -2010,6 +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;
@@ -2021,13 +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);
@@ -2051,6 +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)
{
@@ -2063,9 +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,
@@ -2136,6 +2376,7 @@ static void cma_add_one(struct ib_device *device)
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 */
@@ -2150,8 +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)
@@ -2233,6 +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 86a3b2d401d..8926a2bd4a8 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/mad.c b/drivers/infiniband/core/mad.c
index 15f38d94b3a..13efd417034 100644
--- a/drivers/infiniband/core/mad.c
+++ b/drivers/infiniband/core/mad.c
@@ -642,7 +642,8 @@ static void snoop_recv(struct ib_mad_qp_info *qp_info,
spin_unlock_irqrestore(&qp_info->snoop_lock, flags);
}
-static void build_smp_wc(u64 wr_id, u16 slid, u16 pkey_index, u8 port_num,
+static void build_smp_wc(struct ib_qp *qp,
+ u64 wr_id, u16 slid, u16 pkey_index, u8 port_num,
struct ib_wc *wc)
{
memset(wc, 0, sizeof *wc);
@@ -652,7 +653,7 @@ static void build_smp_wc(u64 wr_id, u16 slid, u16 pkey_index, u8 port_num,
wc->pkey_index = pkey_index;
wc->byte_len = sizeof(struct ib_mad) + sizeof(struct ib_grh);
wc->src_qp = IB_QP0;
- wc->qp_num = IB_QP0;
+ wc->qp = qp;
wc->slid = slid;
wc->sl = 0;
wc->dlid_path_bits = 0;
@@ -713,7 +714,8 @@ static int handle_outgoing_dr_smp(struct ib_mad_agent_private *mad_agent_priv,
goto out;
}
- build_smp_wc(send_wr->wr_id, be16_to_cpu(smp->dr_slid),
+ build_smp_wc(mad_agent_priv->agent.qp,
+ send_wr->wr_id, be16_to_cpu(smp->dr_slid),
send_wr->wr.ud.pkey_index,
send_wr->wr.ud.port_num, &mad_wc);
@@ -998,17 +1000,17 @@ 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) {
@@ -1026,12 +1028,12 @@ 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;
}
@@ -1850,11 +1852,11 @@ 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;
@@ -2080,12 +2082,12 @@ 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);
@@ -2355,7 +2357,8 @@ static void local_completions(struct work_struct *work)
* Defined behavior is to complete response
* before request
*/
- build_smp_wc((unsigned long) local->mad_send_wr,
+ build_smp_wc(recv_mad_agent->agent.qp,
+ (unsigned long) local->mad_send_wr,
be16_to_cpu(IB_LID_PERMISSIVE),
0, recv_mad_agent->agent.port_num, &wc);
@@ -2528,13 +2531,12 @@ 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;
@@ -2549,12 +2551,11 @@ 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;
@@ -2586,11 +2587,11 @@ 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);
}
diff --git a/drivers/infiniband/core/mad_priv.h b/drivers/infiniband/core/mad_priv.h
index d5548e73e06..de89717f49f 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 {
@@ -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/ucma.c b/drivers/infiniband/core/ucma.c
new file mode 100644
index 00000000000..e2e8d329b44
--- /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_cmd.c b/drivers/infiniband/core/uverbs_cmd.c
index 743247ec065..df1efbc1088 100644
--- a/drivers/infiniband/core/uverbs_cmd.c
+++ b/drivers/infiniband/core/uverbs_cmd.c
@@ -933,7 +933,7 @@ ssize_t ib_uverbs_poll_cq(struct ib_uverbs_file *file,
resp->wc[i].vendor_err = wc[i].vendor_err;
resp->wc[i].byte_len = wc[i].byte_len;
resp->wc[i].imm_data = (__u32 __force) wc[i].imm_data;
- resp->wc[i].qp_num = wc[i].qp_num;
+ resp->wc[i].qp_num = wc[i].qp->qp_num;
resp->wc[i].src_qp = wc[i].src_qp;
resp->wc[i].wc_flags = wc[i].wc_flags;
resp->wc[i].pkey_index = wc[i].pkey_index;
diff --git a/drivers/infiniband/core/uverbs_main.c b/drivers/infiniband/core/uverbs_main.c
index 4e16314e8e6..a617ca7b692 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 ce46b13ae02..5440da0e59b 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 db12cc0841d..c95fe952abd 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);
diff --git a/drivers/infiniband/hw/amso1100/c2_cq.c b/drivers/infiniband/hw/amso1100/c2_cq.c
index 05c9154d46f..5175c99ee58 100644
--- a/drivers/infiniband/hw/amso1100/c2_cq.c
+++ b/drivers/infiniband/hw/amso1100/c2_cq.c
@@ -153,7 +153,7 @@ static inline int c2_poll_one(struct c2_dev *c2dev,
entry->status = c2_cqe_status_to_openib(c2_wr_get_result(ce));
entry->wr_id = ce->hdr.context;
- entry->qp_num = ce->handle;
+ entry->qp = &qp->ibqp;
entry->wc_flags = 0;
entry->slid = 0;
entry->sl = 0;
diff --git a/drivers/infiniband/hw/amso1100/c2_qp.c b/drivers/infiniband/hw/amso1100/c2_qp.c
index 179d005ed4a..420c1380f5c 100644
--- a/drivers/infiniband/hw/amso1100/c2_qp.c
+++ b/drivers/infiniband/hw/amso1100/c2_qp.c
@@ -161,8 +161,10 @@ int c2_qp_modify(struct c2_dev *c2dev, struct c2_qp *qp,
if (attr_mask & IB_QP_STATE) {
/* Ensure the state is valid */
- if (attr->qp_state < 0 || attr->qp_state > IB_QPS_ERR)
- return -EINVAL;
+ if (attr->qp_state < 0 || attr->qp_state > IB_QPS_ERR) {
+ err = -EINVAL;
+ goto bail0;
+ }
wr.next_qp_state = cpu_to_be32(to_c2_state(attr->qp_state));
@@ -184,9 +186,10 @@ int c2_qp_modify(struct c2_dev *c2dev, struct c2_qp *qp,
if (attr->cur_qp_state != IB_QPS_RTR &&
attr->cur_qp_state != IB_QPS_RTS &&
attr->cur_qp_state != IB_QPS_SQD &&
- attr->cur_qp_state != IB_QPS_SQE)
- return -EINVAL;
- else
+ attr->cur_qp_state != IB_QPS_SQE) {
+ err = -EINVAL;
+ goto bail0;
+ } else
wr.next_qp_state =
cpu_to_be32(to_c2_state(attr->cur_qp_state));
diff --git a/drivers/infiniband/hw/ehca/ehca_classes.h b/drivers/infiniband/hw/ehca/ehca_classes.h
index 1c722032319..cf95ee474b0 100644
--- a/drivers/infiniband/hw/ehca/ehca_classes.h
+++ b/drivers/infiniband/hw/ehca/ehca_classes.h
@@ -119,13 +119,14 @@ struct ehca_qp {
struct ipz_qp_handle ipz_qp_handle;
struct ehca_pfqp pf;
struct ib_qp_init_attr init_attr;
- u64 uspace_squeue;
- u64 uspace_rqueue;
- u64 uspace_fwh;
struct ehca_cq *send_cq;
struct ehca_cq *recv_cq;
unsigned int sqerr_purgeflag;
struct hlist_node list_entries;
+ /* mmap counter for resources mapped into user space */
+ u32 mm_count_squeue;
+ u32 mm_count_rqueue;
+ u32 mm_count_galpa;
};
/* must be power of 2 */
@@ -142,13 +143,14 @@ struct ehca_cq {
struct ipz_cq_handle ipz_cq_handle;
struct ehca_pfcq pf;
spinlock_t cb_lock;
- u64 uspace_queue;
- u64 uspace_fwh;
struct hlist_head qp_hashtab[QP_HASHTAB_LEN];
struct list_head entry;
u32 nr_callbacks;
spinlock_t task_lock;
u32 ownpid;
+ /* mmap counter for resources mapped into user space */
+ u32 mm_count_queue;
+ u32 mm_count_galpa;
};
enum ehca_mr_flag {
@@ -248,20 +250,6 @@ struct ehca_ucontext {
struct ib_ucontext ib_ucontext;
};
-struct ehca_module *ehca_module_new(void);
-
-int ehca_module_delete(struct ehca_module *me);
-
-int ehca_eq_ctor(struct ehca_eq *eq);
-
-int ehca_eq_dtor(struct ehca_eq *eq);
-
-struct ehca_shca *ehca_shca_new(void);
-
-int ehca_shca_delete(struct ehca_shca *me);
-
-struct ehca_sport *ehca_sport_new(struct ehca_shca *anchor);
-
int ehca_init_pd_cache(void);
void ehca_cleanup_pd_cache(void);
int ehca_init_cq_cache(void);
@@ -283,7 +271,6 @@ extern int ehca_port_act_time;
extern int ehca_use_hp_mr;
struct ipzu_queue_resp {
- u64 queue; /* points to first queue entry */
u32 qe_size; /* queue entry size */
u32 act_nr_of_sg;
u32 queue_length; /* queue length allocated in bytes */
@@ -296,7 +283,6 @@ struct ehca_create_cq_resp {
u32 cq_number;
u32 token;
struct ipzu_queue_resp ipz_queue;
- struct h_galpas galpas;
};
struct ehca_create_qp_resp {
@@ -309,7 +295,6 @@ struct ehca_create_qp_resp {
u32 dummy; /* padding for 8 byte alignment */
struct ipzu_queue_resp ipz_squeue;
struct ipzu_queue_resp ipz_rqueue;
- struct h_galpas galpas;
};
struct ehca_alloc_cq_parms {
diff --git a/drivers/infiniband/hw/ehca/ehca_cq.c b/drivers/infiniband/hw/ehca/ehca_cq.c
index 93995b658d9..9291a86ca05 100644
--- a/drivers/infiniband/hw/ehca/ehca_cq.c
+++ b/drivers/infiniband/hw/ehca/ehca_cq.c
@@ -267,7 +267,6 @@ struct ib_cq *ehca_create_cq(struct ib_device *device, int cqe,
if (context) {
struct ipz_queue *ipz_queue = &my_cq->ipz_queue;
struct ehca_create_cq_resp resp;
- struct vm_area_struct *vma;
memset(&resp, 0, sizeof(resp));
resp.cq_number = my_cq->cq_number;
resp.token = my_cq->token;
@@ -276,40 +275,14 @@ struct ib_cq *ehca_create_cq(struct ib_device *device, int cqe,
resp.ipz_queue.queue_length = ipz_queue->queue_length;
resp.ipz_queue.pagesize = ipz_queue->pagesize;
resp.ipz_queue.toggle_state = ipz_queue->toggle_state;
- ret = ehca_mmap_nopage(((u64)(my_cq->token) << 32) | 0x12000000,
- ipz_queue->queue_length,
- (void**)&resp.ipz_queue.queue,
- &vma);
- if (ret) {
- ehca_err(device, "Could not mmap queue pages");
- cq = ERR_PTR(ret);
- goto create_cq_exit4;
- }
- my_cq->uspace_queue = resp.ipz_queue.queue;
- resp.galpas = my_cq->galpas;
- ret = ehca_mmap_register(my_cq->galpas.user.fw_handle,
- (void**)&resp.galpas.kernel.fw_handle,
- &vma);
- if (ret) {
- ehca_err(device, "Could not mmap fw_handle");
- cq = ERR_PTR(ret);
- goto create_cq_exit5;
- }
- my_cq->uspace_fwh = (u64)resp.galpas.kernel.fw_handle;
if (ib_copy_to_udata(udata, &resp, sizeof(resp))) {
ehca_err(device, "Copy to udata failed.");
- goto create_cq_exit6;
+ goto create_cq_exit4;
}
}
return cq;
-create_cq_exit6:
- ehca_munmap(my_cq->uspace_fwh, EHCA_PAGESIZE);
-
-create_cq_exit5:
- ehca_munmap(my_cq->uspace_queue, my_cq->ipz_queue.queue_length);
-
create_cq_exit4:
ipz_queue_dtor(&my_cq->ipz_queue);
@@ -333,7 +306,6 @@ create_cq_exit1:
int ehca_destroy_cq(struct ib_cq *cq)
{
u64 h_ret;
- int ret;
struct ehca_cq *my_cq = container_of(cq, struct ehca_cq, ib_cq);
int cq_num = my_cq->cq_number;
struct ib_device *device = cq->device;
@@ -343,32 +315,30 @@ int ehca_destroy_cq(struct ib_cq *cq)
u32 cur_pid = current->tgid;
unsigned long flags;
+ if (cq->uobject) {
+ if (my_cq->mm_count_galpa || my_cq->mm_count_queue) {
+ ehca_err(device, "Resources still referenced in "
+ "user space cq_num=%x", my_cq->cq_number);
+ return -EINVAL;
+ }
+ if (my_cq->ownpid != cur_pid) {
+ ehca_err(device, "Invalid caller pid=%x ownpid=%x "
+ "cq_num=%x",
+ cur_pid, my_cq->ownpid, my_cq->cq_number);
+ return -EINVAL;
+ }
+ }
+
spin_lock_irqsave(&ehca_cq_idr_lock, flags);
- while (my_cq->nr_callbacks)
+ while (my_cq->nr_callbacks) {
+ spin_unlock_irqrestore(&ehca_cq_idr_lock, flags);
yield();
+ spin_lock_irqsave(&ehca_cq_idr_lock, flags);
+ }
idr_remove(&ehca_cq_idr, my_cq->token);
spin_unlock_irqrestore(&ehca_cq_idr_lock, flags);
- if (my_cq->uspace_queue && my_cq->ownpid != cur_pid) {
- ehca_err(device, "Invalid caller pid=%x ownpid=%x",
- cur_pid, my_cq->ownpid);
- return -EINVAL;
- }
-
- /* un-mmap if vma alloc */
- if (my_cq->uspace_queue ) {
- ret = ehca_munmap(my_cq->uspace_queue,
- my_cq->ipz_queue.queue_length);
- if (ret)
- ehca_err(device, "Could not munmap queue ehca_cq=%p "
- "cq_num=%x", my_cq, cq_num);
- ret = ehca_munmap(my_cq->uspace_fwh, EHCA_PAGESIZE);
- if (ret)
- ehca_err(device, "Could not munmap fwh ehca_cq=%p "
- "cq_num=%x", my_cq, cq_num);
- }
-
h_ret = hipz_h_destroy_cq(adapter_handle, my_cq, 0);
if (h_ret == H_R_STATE) {
/* cq in err: read err data and destroy it forcibly */
@@ -397,7 +367,7 @@ int ehca_resize_cq(struct ib_cq *cq, int cqe, struct ib_udata *udata)
struct ehca_cq *my_cq = container_of(cq, struct ehca_cq, ib_cq);
u32 cur_pid = current->tgid;
- if (my_cq->uspace_queue && my_cq->ownpid != cur_pid) {
+ if (cq->uobject && my_cq->ownpid != cur_pid) {
ehca_err(cq->device, "Invalid caller pid=%x ownpid=%x",
cur_pid, my_cq->ownpid);
return -EINVAL;
diff --git a/drivers/infiniband/hw/ehca/ehca_hca.c b/drivers/infiniband/hw/ehca/ehca_hca.c
index e1b618c5f68..b7be950ab47 100644
--- a/drivers/infiniband/hw/ehca/ehca_hca.c
+++ b/drivers/infiniband/hw/ehca/ehca_hca.c
@@ -50,7 +50,7 @@ int ehca_query_device(struct ib_device *ibdev, struct ib_device_attr *props)
ib_device);
struct hipz_query_hca *rblock;
- rblock = ehca_alloc_fw_ctrlblock();
+ rblock = ehca_alloc_fw_ctrlblock(GFP_KERNEL);
if (!rblock) {
ehca_err(&shca->ib_device, "Can't allocate rblock memory.");
return -ENOMEM;
@@ -110,7 +110,7 @@ int ehca_query_port(struct ib_device *ibdev,
ib_device);
struct hipz_query_port *rblock;
- rblock = ehca_alloc_fw_ctrlblock();
+ rblock = ehca_alloc_fw_ctrlblock(GFP_KERNEL);
if (!rblock) {
ehca_err(&shca->ib_device, "Can't allocate rblock memory.");
return -ENOMEM;
@@ -179,7 +179,7 @@ int ehca_query_pkey(struct ib_device *ibdev, u8 port, u16 index, u16 *pkey)
return -EINVAL;
}
- rblock = ehca_alloc_fw_ctrlblock();
+ rblock = ehca_alloc_fw_ctrlblock(GFP_KERNEL);
if (!rblock) {
ehca_err(&shca->ib_device, "Can't allocate rblock memory.");
return -ENOMEM;
@@ -212,7 +212,7 @@ int ehca_query_gid(struct ib_device *ibdev, u8 port,
return -EINVAL;
}
- rblock = ehca_alloc_fw_ctrlblock();
+ rblock = ehca_alloc_fw_ctrlblock(GFP_KERNEL);
if (!rblock) {
ehca_err(&shca->ib_device, "Can't allocate rblock memory.");
return -ENOMEM;
diff --git a/drivers/infiniband/hw/ehca/ehca_irq.c b/drivers/infiniband/hw/ehca/ehca_irq.c
index c3ea746e904..c069be8cbcb 100644
--- a/drivers/infiniband/hw/ehca/ehca_irq.c
+++ b/drivers/infiniband/hw/ehca/ehca_irq.c
@@ -138,7 +138,7 @@ int ehca_error_data(struct ehca_shca *shca, void *data,
u64 *rblock;
unsigned long block_count;
- rblock = ehca_alloc_fw_ctrlblock();
+ rblock = ehca_alloc_fw_ctrlblock(GFP_ATOMIC);
if (!rblock) {
ehca_err(&shca->ib_device, "Cannot allocate rblock memory.");
ret = -ENOMEM;
@@ -440,7 +440,8 @@ void ehca_tasklet_eq(unsigned long data)
cq = idr_find(&ehca_cq_idr, token);
if (cq == NULL) {
- spin_unlock(&ehca_cq_idr_lock);
+ spin_unlock_irqrestore(&ehca_cq_idr_lock,
+ flags);
break;
}
diff --git a/drivers/infiniband/hw/ehca/ehca_iverbs.h b/drivers/infiniband/hw/ehca/ehca_iverbs.h
index 3720e3032cc..95fd59fb452 100644
--- a/drivers/infiniband/hw/ehca/ehca_iverbs.h
+++ b/drivers/infiniband/hw/ehca/ehca_iverbs.h
@@ -171,19 +171,11 @@ int ehca_mmap(struct ib_ucontext *context, struct vm_area_struct *vma);
void ehca_poll_eqs(unsigned long data);
-int ehca_mmap_nopage(u64 foffset,u64 length,void **mapped,
- struct vm_area_struct **vma);
-
-int ehca_mmap_register(u64 physical,void **mapped,
- struct vm_area_struct **vma);
-
-int ehca_munmap(unsigned long addr, size_t len);
-
#ifdef CONFIG_PPC_64K_PAGES
-void *ehca_alloc_fw_ctrlblock(void);
+void *ehca_alloc_fw_ctrlblock(gfp_t flags);
void ehca_free_fw_ctrlblock(void *ptr);
#else
-#define ehca_alloc_fw_ctrlblock() ((void *) get_zeroed_page(GFP_KERNEL))
+#define ehca_alloc_fw_ctrlblock(flags) ((void *) get_zeroed_page(flags))
#define ehca_free_fw_ctrlblock(ptr) free_page((unsigned long)(ptr))
#endif
diff --git a/drivers/infiniband/hw/ehca/ehca_main.c b/drivers/infiniband/hw/ehca/ehca_main.c
index cc47e4c13a1..1155bcf4821 100644
--- a/drivers/infiniband/hw/ehca/ehca_main.c
+++ b/drivers/infiniband/hw/ehca/ehca_main.c
@@ -52,7 +52,7 @@
MODULE_LICENSE("Dual BSD/GPL");
MODULE_AUTHOR("Christoph Raisch <raisch@de.ibm.com>");
MODULE_DESCRIPTION("IBM eServer HCA InfiniBand Device Driver");
-MODULE_VERSION("SVNEHCA_0019");
+MODULE_VERSION("SVNEHCA_0020");
int ehca_open_aqp1 = 0;
int ehca_debug_level = 0;
@@ -106,9 +106,9 @@ static struct timer_list poll_eqs_timer;
#ifdef CONFIG_PPC_64K_PAGES
static struct kmem_cache *ctblk_cache = NULL;
-void *ehca_alloc_fw_ctrlblock(void)
+void *ehca_alloc_fw_ctrlblock(gfp_t flags)
{
- void *ret = kmem_cache_zalloc(ctblk_cache, GFP_KERNEL);
+ void *ret = kmem_cache_zalloc(ctblk_cache, flags);
if (!ret)
ehca_gen_err("Out of memory for ctblk");
return ret;
@@ -206,7 +206,7 @@ int ehca_sense_attributes(struct ehca_shca *shca)
u64 h_ret;
struct hipz_query_hca *rblock;
- rblock = ehca_alloc_fw_ctrlblock();
+ rblock = ehca_alloc_fw_ctrlblock(GFP_KERNEL);
if (!rblock) {
ehca_gen_err("Cannot allocate rblock memory.");
return -ENOMEM;
@@ -258,7 +258,7 @@ static int init_node_guid(struct ehca_shca *shca)
int ret = 0;
struct hipz_query_hca *rblock;
- rblock = ehca_alloc_fw_ctrlblock();
+ rblock = ehca_alloc_fw_ctrlblock(GFP_KERNEL);
if (!rblock) {
ehca_err(&shca->ib_device, "Can't allocate rblock memory.");
return -ENOMEM;
@@ -288,7 +288,7 @@ int ehca_init_device(struct ehca_shca *shca)
strlcpy(shca->ib_device.name, "ehca%d", IB_DEVICE_NAME_MAX);
shca->ib_device.owner = THIS_MODULE;
- shca->ib_device.uverbs_abi_ver = 5;
+ shca->ib_device.uverbs_abi_ver = 6;
shca->ib_device.uverbs_cmd_mask =
(1ull << IB_USER_VERBS_CMD_GET_CONTEXT) |
(1ull << IB_USER_VERBS_CMD_QUERY_DEVICE) |
@@ -469,7 +469,7 @@ static ssize_t ehca_show_##name(struct device *dev, \
\
shca = dev->driver_data; \
\
- rblock = ehca_alloc_fw_ctrlblock(); \
+ rblock = ehca_alloc_fw_ctrlblock(GFP_KERNEL); \
if (!rblock) { \
dev_err(dev, "Can't allocate rblock memory."); \
return 0; \
@@ -790,7 +790,7 @@ int __init ehca_module_init(void)
int ret;
printk(KERN_INFO "eHCA Infiniband Device Driver "
- "(Rel.: SVNEHCA_0019)\n");
+ "(Rel.: SVNEHCA_0020)\n");
idr_init(&ehca_qp_idr);
idr_init(&ehca_cq_idr);
spin_lock_init(&ehca_qp_idr_lock);
diff --git a/drivers/infiniband/hw/ehca/ehca_mrmw.c b/drivers/infiniband/hw/ehca/ehca_mrmw.c
index 0a5e2214cc5..cfb362a1029 100644
--- a/drivers/infiniband/hw/ehca/ehca_mrmw.c
+++ b/drivers/infiniband/hw/ehca/ehca_mrmw.c
@@ -1013,7 +1013,7 @@ int ehca_reg_mr_rpages(struct ehca_shca *shca,
u32 i;
u64 *kpage;
- kpage = ehca_alloc_fw_ctrlblock();
+ kpage = ehca_alloc_fw_ctrlblock(GFP_KERNEL);
if (!kpage) {
ehca_err(&shca->ib_device, "kpage alloc failed");
ret = -ENOMEM;
@@ -1124,7 +1124,7 @@ inline int ehca_rereg_mr_rereg1(struct ehca_shca *shca,
ehca_mrmw_map_acl(acl, &hipz_acl);
ehca_mrmw_set_pgsize_hipz_acl(&hipz_acl);
- kpage = ehca_alloc_fw_ctrlblock();
+ kpage = ehca_alloc_fw_ctrlblock(GFP_KERNEL);
if (!kpage) {
ehca_err(&shca->ib_device, "kpage alloc failed");
ret = -ENOMEM;
diff --git a/drivers/infiniband/hw/ehca/ehca_qp.c b/drivers/infiniband/hw/ehca/ehca_qp.c
index c6c9cef203e..95efef921f1 100644
--- a/drivers/infiniband/hw/ehca/ehca_qp.c
+++ b/drivers/infiniband/hw/ehca/ehca_qp.c
@@ -637,7 +637,6 @@ struct ib_qp *ehca_create_qp(struct ib_pd *pd,
struct ipz_queue *ipz_rqueue = &my_qp->ipz_rqueue;
struct ipz_queue *ipz_squeue = &my_qp->ipz_squeue;
struct ehca_create_qp_resp resp;
- struct vm_area_struct * vma;
memset(&resp, 0, sizeof(resp));
resp.qp_num = my_qp->real_qp_num;
@@ -651,59 +650,21 @@ struct ib_qp *ehca_create_qp(struct ib_pd *pd,
resp.ipz_rqueue.queue_length = ipz_rqueue->queue_length;
resp.ipz_rqueue.pagesize = ipz_rqueue->pagesize;
resp.ipz_rqueue.toggle_state = ipz_rqueue->toggle_state;
- ret = ehca_mmap_nopage(((u64)(my_qp->token) << 32) | 0x22000000,
- ipz_rqueue->queue_length,
- (void**)&resp.ipz_rqueue.queue,
- &vma);
- if (ret) {
- ehca_err(pd->device, "Could not mmap rqueue pages");
- goto create_qp_exit3;
- }
- my_qp->uspace_rqueue = resp.ipz_rqueue.queue;
/* squeue properties */
resp.ipz_squeue.qe_size = ipz_squeue->qe_size;
resp.ipz_squeue.act_nr_of_sg = ipz_squeue->act_nr_of_sg;
resp.ipz_squeue.queue_length = ipz_squeue->queue_length;
resp.ipz_squeue.pagesize = ipz_squeue->pagesize;
resp.ipz_squeue.toggle_state = ipz_squeue->toggle_state;
- ret = ehca_mmap_nopage(((u64)(my_qp->token) << 32) | 0x23000000,
- ipz_squeue->queue_length,
- (void**)&resp.ipz_squeue.queue,
- &vma);
- if (ret) {
- ehca_err(pd->device, "Could not mmap squeue pages");
- goto create_qp_exit4;
- }
- my_qp->uspace_squeue = resp.ipz_squeue.queue;
- /* fw_handle */
- resp.galpas = my_qp->galpas;
- ret = ehca_mmap_register(my_qp->galpas.user.fw_handle,
- (void**)&resp.galpas.kernel.fw_handle,
- &vma);
- if (ret) {
- ehca_err(pd->device, "Could not mmap fw_handle");
- goto create_qp_exit5;
- }
- my_qp->uspace_fwh = (u64)resp.galpas.kernel.fw_handle;
-
if (ib_copy_to_udata(udata, &resp, sizeof resp)) {
ehca_err(pd->device, "Copy to udata failed");
ret = -EINVAL;
- goto create_qp_exit6;
+ goto create_qp_exit3;
}
}
return &my_qp->ib_qp;
-create_qp_exit6:
- ehca_munmap(my_qp->uspace_fwh, EHCA_PAGESIZE);
-
-create_qp_exit5:
- ehca_munmap(my_qp->uspace_squeue, my_qp->ipz_squeue.queue_length);
-
-create_qp_exit4:
- ehca_munmap(my_qp->uspace_rqueue, my_qp->ipz_rqueue.queue_length);
-
create_qp_exit3:
ipz_queue_dtor(&my_qp->ipz_rqueue);
ipz_queue_dtor(&my_qp->ipz_squeue);
@@ -807,7 +768,7 @@ static int internal_modify_qp(struct ib_qp *ibqp,
unsigned long spl_flags = 0;
/* do query_qp to obtain current attr values */
- mqpcb = ehca_alloc_fw_ctrlblock();
+ mqpcb = ehca_alloc_fw_ctrlblock(GFP_KERNEL);
if (!mqpcb) {
ehca_err(ibqp->device, "Could not get zeroed page for mqpcb "
"ehca_qp=%p qp_num=%x ", my_qp, ibqp->qp_num);
@@ -931,7 +892,7 @@ static int internal_modify_qp(struct ib_qp *ibqp,
my_qp->qp_type == IB_QPT_SMI) &&
statetrans == IB_QPST_SQE2RTS) {
/* mark next free wqe if kernel */
- if (my_qp->uspace_squeue == 0) {
+ if (!ibqp->uobject) {
struct ehca_wqe *wqe;
/* lock send queue */
spin_lock_irqsave(&my_qp->spinlock_s, spl_flags);
@@ -1273,7 +1234,7 @@ int ehca_query_qp(struct ib_qp *qp,
return -EINVAL;
}
- qpcb = ehca_alloc_fw_ctrlblock();
+ qpcb = ehca_alloc_fw_ctrlblock(GFP_KERNEL);
if (!qpcb) {
ehca_err(qp->device,"Out of memory for qpcb "
"ehca_qp=%p qp_num=%x", my_qp, qp->qp_num);
@@ -1417,11 +1378,18 @@ int ehca_destroy_qp(struct ib_qp *ibqp)
enum ib_qp_type qp_type;
unsigned long flags;
- if (my_pd->ib_pd.uobject && my_pd->ib_pd.uobject->context &&
- my_pd->ownpid != cur_pid) {
- ehca_err(ibqp->device, "Invalid caller pid=%x ownpid=%x",
- cur_pid, my_pd->ownpid);
- return -EINVAL;
+ if (ibqp->uobject) {
+ if (my_qp->mm_count_galpa ||
+ my_qp->mm_count_rqueue || my_qp->mm_count_squeue) {
+ ehca_err(ibqp->device, "Resources still referenced in "
+ "user space qp_num=%x", ibqp->qp_num);
+ return -EINVAL;
+ }
+ if (my_pd->ownpid != cur_pid) {
+ ehca_err(ibqp->device, "Invalid caller pid=%x ownpid=%x",
+ cur_pid, my_pd->ownpid);
+ return -EINVAL;
+ }
}
if (my_qp->send_cq) {
@@ -1439,24 +1407,6 @@ int ehca_destroy_qp(struct ib_qp *ibqp)
idr_remove(&ehca_qp_idr, my_qp->token);
spin_unlock_irqrestore(&ehca_qp_idr_lock, flags);
- /* un-mmap if vma alloc */
- if (my_qp->uspace_rqueue) {
- ret = ehca_munmap(my_qp->uspace_rqueue,
- my_qp->ipz_rqueue.queue_length);
- if (ret)
- ehca_err(ibqp->device, "Could not munmap rqueue "
- "qp_num=%x", qp_num);
- ret = ehca_munmap(my_qp->uspace_squeue,
- my_qp->ipz_squeue.queue_length);
- if (ret)
- ehca_err(ibqp->device, "Could not munmap squeue "
- "qp_num=%x", qp_num);
- ret = ehca_munmap(my_qp->uspace_fwh, EHCA_PAGESIZE);
- if (ret)
- ehca_err(ibqp->device, "Could not munmap fwh qp_num=%x",
- qp_num);
- }
-
h_ret = hipz_h_destroy_qp(shca->ipz_hca_handle, my_qp);
if (h_ret != H_SUCCESS) {
ehca_err(ibqp->device, "hipz_h_destroy_qp() failed rc=%lx "
diff --git a/drivers/infiniband/hw/ehca/ehca_reqs.c b/drivers/infiniband/hw/ehca/ehca_reqs.c
index b46bda1bf85..08d3f892d9f 100644
--- a/drivers/infiniband/hw/ehca/ehca_reqs.c
+++ b/drivers/infiniband/hw/ehca/ehca_reqs.c
@@ -579,7 +579,7 @@ poll_cq_one_read_cqe:
} else
wc->status = IB_WC_SUCCESS;
- wc->qp_num = cqe->local_qp_number;
+ wc->qp = NULL;
wc->byte_len = cqe->nr_bytes_transferred;
wc->pkey_index = cqe->pkey_index;
wc->slid = cqe->rlid;
diff --git a/drivers/infiniband/hw/ehca/ehca_uverbs.c b/drivers/infiniband/hw/ehca/ehca_uverbs.c
index e08764e4aef..73db920b694 100644
--- a/drivers/infiniband/hw/ehca/ehca_uverbs.c
+++ b/drivers/infiniband/hw/ehca/ehca_uverbs.c
@@ -68,105 +68,183 @@ int ehca_dealloc_ucontext(struct ib_ucontext *context)
return 0;
}
-struct page *ehca_nopage(struct vm_area_struct *vma,
- unsigned long address, int *type)
+static void ehca_mm_open(struct vm_area_struct *vma)
{
- struct page *mypage = NULL;
- u64 fileoffset = vma->vm_pgoff << PAGE_SHIFT;
- u32 idr_handle = fileoffset >> 32;
- u32 q_type = (fileoffset >> 28) & 0xF; /* CQ, QP,... */
- u32 rsrc_type = (fileoffset >> 24) & 0xF; /* sq,rq,cmnd_window */
- u32 cur_pid = current->tgid;
- unsigned long flags;
- struct ehca_cq *cq;
- struct ehca_qp *qp;
- struct ehca_pd *pd;
- u64 offset;
- void *vaddr;
+ u32 *count = (u32*)vma->vm_private_data;
+ if (!count) {
+ ehca_gen_err("Invalid vma struct vm_start=%lx vm_end=%lx",
+ vma->vm_start, vma->vm_end);
+ return;
+ }
+ (*count)++;
+ if (!(*count))
+ ehca_gen_err("Use count overflow vm_start=%lx vm_end=%lx",
+ vma->vm_start, vma->vm_end);
+ ehca_gen_dbg("vm_start=%lx vm_end=%lx count=%x",
+ vma->vm_start, vma->vm_end, *count);
+}
- switch (q_type) {
- case 1: /* CQ */
- spin_lock_irqsave(&ehca_cq_idr_lock, flags);
- cq = idr_find(&ehca_cq_idr, idr_handle);
- spin_unlock_irqrestore(&ehca_cq_idr_lock, flags);
+static void ehca_mm_close(struct vm_area_struct *vma)
+{
+ u32 *count = (u32*)vma->vm_private_data;
+ if (!count) {
+ ehca_gen_err("Invalid vma struct vm_start=%lx vm_end=%lx",
+ vma->vm_start, vma->vm_end);
+ return;
+ }
+ (*count)--;
+ ehca_gen_dbg("vm_start=%lx vm_end=%lx count=%x",
+ vma->vm_start, vma->vm_end, *count);
+}
- /* make sure this mmap really belongs to the authorized user */
- if (!cq) {
- ehca_gen_err("cq is NULL ret=NOPAGE_SIGBUS");
- return NOPAGE_SIGBUS;
+static struct vm_operations_struct vm_ops = {
+ .open = ehca_mm_open,
+ .close = ehca_mm_close,
+};
+
+static int ehca_mmap_fw(struct vm_area_struct *vma, struct h_galpas *galpas,
+ u32 *mm_count)
+{
+ int ret;
+ u64 vsize, physical;
+
+ vsize = vma->vm_end - vma->vm_start;
+ if (vsize != EHCA_PAGESIZE) {
+ ehca_gen_err("invalid vsize=%lx", vma->vm_end - vma->vm_start);
+ return -EINVAL;
+ }
+
+ physical = galpas->user.fw_handle;
+ vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
+ ehca_gen_dbg("vsize=%lx physical=%lx", vsize, physical);
+ /* VM_IO | VM_RESERVED are set by remap_pfn_range() */
+ ret = remap_pfn_range(vma, vma->vm_start, physical >> PAGE_SHIFT,
+ vsize, vma->vm_page_prot);
+ if (unlikely(ret)) {
+ ehca_gen_err("remap_pfn_range() failed ret=%x", ret);
+ return -ENOMEM;
+ }
+
+ vma->vm_private_data = mm_count;
+ (*mm_count)++;
+ vma->vm_ops = &vm_ops;
+
+ return 0;
+}
+
+static int ehca_mmap_queue(struct vm_area_struct *vma, struct ipz_queue *queue,
+ u32 *mm_count)
+{
+ int ret;
+ u64 start, ofs;
+ struct page *page;
+
+ vma->vm_flags |= VM_RESERVED;
+ start = vma->vm_start;
+ for (ofs = 0; ofs < queue->queue_length; ofs += PAGE_SIZE) {
+ u64 virt_addr = (u64)ipz_qeit_calc(queue, ofs);
+ page = virt_to_page(virt_addr);
+ ret = vm_insert_page(vma, start, page);
+ if (unlikely(ret)) {
+ ehca_gen_err("vm_insert_page() failed rc=%x", ret);
+ return ret;
}
+ start += PAGE_SIZE;
+ }
+ vma->vm_private_data = mm_count;
+ (*mm_count)++;
+ vma->vm_ops = &vm_ops;
- if (cq->ownpid != cur_pid) {
+ return 0;
+}
+
+static int ehca_mmap_cq(struct vm_area_struct *vma, struct ehca_cq *cq,
+ u32 rsrc_type)
+{
+ int ret;
+
+ switch (rsrc_type) {
+ case 1: /* galpa fw handle */
+ ehca_dbg(cq->ib_cq.device, "cq_num=%x fw", cq->cq_number);
+ ret = ehca_mmap_fw(vma, &cq->galpas, &cq->mm_count_galpa);
+ if (unlikely(ret)) {
ehca_err(cq->ib_cq.device,
- "Invalid caller pid=%x ownpid=%x",
- cur_pid, cq->ownpid);
- return NOPAGE_SIGBUS;
+ "ehca_mmap_fw() failed rc=%x cq_num=%x",
+ ret, cq->cq_number);
+ return ret;
}
+ break;
- if (rsrc_type == 2) {
- ehca_dbg(cq->ib_cq.device, "cq=%p cq queuearea", cq);
- offset = address - vma->vm_start;
- vaddr = ipz_qeit_calc(&cq->ipz_queue, offset);
- ehca_dbg(cq->ib_cq.device, "offset=%lx vaddr=%p",
- offset, vaddr);
- mypage = virt_to_page(vaddr);
+ case 2: /* cq queue_addr */
+ ehca_dbg(cq->ib_cq.device, "cq_num=%x queue", cq->cq_number);
+ ret = ehca_mmap_queue(vma, &cq->ipz_queue, &cq->mm_count_queue);
+ if (unlikely(ret)) {
+ ehca_err(cq->ib_cq.device,
+ "ehca_mmap_queue() failed rc=%x cq_num=%x",
+ ret, cq->cq_number);
+ return ret;
}
break;
- case 2: /* QP */
- spin_lock_irqsave(&ehca_qp_idr_lock, flags);
- qp = idr_find(&ehca_qp_idr, idr_handle);
- spin_unlock_irqrestore(&ehca_qp_idr_lock, flags);
+ default:
+ ehca_err(cq->ib_cq.device, "bad resource type=%x cq_num=%x",
+ rsrc_type, cq->cq_number);
+ return -EINVAL;
+ }
- /* make sure this mmap really belongs to the authorized user */
- if (!qp) {
- ehca_gen_err("qp is NULL ret=NOPAGE_SIGBUS");
- return NOPAGE_SIGBUS;
+ return 0;
+}
+
+static int ehca_mmap_qp(struct vm_area_struct *vma, struct ehca_qp *qp,
+ u32 rsrc_type)
+{
+ int ret;
+
+ switch (rsrc_type) {
+ case 1: /* galpa fw handle */
+ ehca_dbg(qp->ib_qp.device, "qp_num=%x fw", qp->ib_qp.qp_num);
+ ret = ehca_mmap_fw(vma, &qp->galpas, &qp->mm_count_galpa);
+ if (unlikely(ret)) {
+ ehca_err(qp->ib_qp.device,
+ "remap_pfn_range() failed ret=%x qp_num=%x",
+ ret, qp->ib_qp.qp_num);
+ return -ENOMEM;
}
+ break;
- pd = container_of(qp->ib_qp.pd, struct ehca_pd, ib_pd);
- if (pd->ownpid != cur_pid) {
+ case 2: /* qp rqueue_addr */
+ ehca_dbg(qp->ib_qp.device, "qp_num=%x rqueue",
+ qp->ib_qp.qp_num);
+ ret = ehca_mmap_queue(vma, &qp->ipz_rqueue, &qp->mm_count_rqueue);
+ if (unlikely(ret)) {
ehca_err(qp->ib_qp.device,
- "Invalid caller pid=%x ownpid=%x",
- cur_pid, pd->ownpid);
- return NOPAGE_SIGBUS;
+ "ehca_mmap_queue(rq) failed rc=%x qp_num=%x",
+ ret, qp->ib_qp.qp_num);
+ return ret;
}
+ break;
- if (rsrc_type == 2) { /* rqueue */
- ehca_dbg(qp->ib_qp.device, "qp=%p qp rqueuearea", qp);
- offset = address - vma->vm_start;
- vaddr = ipz_qeit_calc(&qp->ipz_rqueue, offset);
- ehca_dbg(qp->ib_qp.device, "offset=%lx vaddr=%p",
- offset, vaddr);
- mypage = virt_to_page(vaddr);
- } else if (rsrc_type == 3) { /* squeue */
- ehca_dbg(qp->ib_qp.device, "qp=%p qp squeuearea", qp);
- offset = address - vma->vm_start;
- vaddr = ipz_qeit_calc(&qp->ipz_squeue, offset);
- ehca_dbg(qp->ib_qp.device, "offset=%lx vaddr=%p",
- offset, vaddr);
- mypage = virt_to_page(vaddr);
+ case 3: /* qp squeue_addr */
+ ehca_dbg(qp->ib_qp.device, "qp_num=%x squeue",
+ qp->ib_qp.qp_num);
+ ret = ehca_mmap_queue(vma, &qp->ipz_squeue, &qp->mm_count_squeue);
+ if (unlikely(ret)) {
+ ehca_err(qp->ib_qp.device,
+ "ehca_mmap_queue(sq) failed rc=%x qp_num=%x",
+ ret, qp->ib_qp.qp_num);
+ return ret;
}
break;
default:
- ehca_gen_err("bad queue type %x", q_type);
- return NOPAGE_SIGBUS;
- }
-
- if (!mypage) {
- ehca_gen_err("Invalid page adr==NULL ret=NOPAGE_SIGBUS");
- return NOPAGE_SIGBUS;
+ ehca_err(qp->ib_qp.device, "bad resource type=%x qp=num=%x",
+ rsrc_type, qp->ib_qp.qp_num);
+ return -EINVAL;
}
- get_page(mypage);
- return mypage;
+ return 0;
}
-static struct vm_operations_struct ehcau_vm_ops = {
- .nopage = ehca_nopage,
-};
-
int ehca_mmap(struct ib_ucontext *context, struct vm_area_struct *vma)
{
u64 fileoffset = vma->vm_pgoff << PAGE_SHIFT;
@@ -175,7 +253,6 @@ int ehca_mmap(struct ib_ucontext *context, struct vm_area_struct *vma)
u32 rsrc_type = (fileoffset >> 24) & 0xF; /* sq,rq,cmnd_window */
u32 cur_pid = current->tgid;
u32 ret;
- u64 vsize, physical;
unsigned long flags;
struct ehca_cq *cq;
struct ehca_qp *qp;
@@ -201,44 +278,12 @@ int ehca_mmap(struct ib_ucontext *context, struct vm_area_struct *vma)
if (!cq->ib_cq.uobject || cq->ib_cq.uobject->context != context)
return -EINVAL;
- switch (rsrc_type) {
- case 1: /* galpa fw handle */
- ehca_dbg(cq->ib_cq.device, "cq=%p cq triggerarea", cq);
- vma->vm_flags |= VM_RESERVED;
- vsize = vma->vm_end - vma->vm_start;
- if (vsize != EHCA_PAGESIZE) {
- ehca_err(cq->ib_cq.device, "invalid vsize=%lx",
- vma->vm_end - vma->vm_start);
- return -EINVAL;
- }
-
- physical = cq->galpas.user.fw_handle;
- vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
- vma->vm_flags |= VM_IO | VM_RESERVED;
-
- ehca_dbg(cq->ib_cq.device,
- "vsize=%lx physical=%lx", vsize, physical);
- ret = remap_pfn_range(vma, vma->vm_start,
- physical >> PAGE_SHIFT, vsize,
- vma->vm_page_prot);
- if (ret) {
- ehca_err(cq->ib_cq.device,
- "remap_pfn_range() failed ret=%x",
- ret);
- return -ENOMEM;
- }
- break;
-
- case 2: /* cq queue_addr */
- ehca_dbg(cq->ib_cq.device, "cq=%p cq q_addr", cq);
- vma->vm_flags |= VM_RESERVED;
- vma->vm_ops = &ehcau_vm_ops;
- break;
-
- default:
- ehca_err(cq->ib_cq.device, "bad resource type %x",
- rsrc_type);
- return -EINVAL;
+ ret = ehca_mmap_cq(vma, cq, rsrc_type);
+ if (unlikely(ret)) {
+ ehca_err(cq->ib_cq.device,
+ "ehca_mmap_cq() failed rc=%x cq_num=%x",
+ ret, cq->cq_number);
+ return ret;
}
break;
@@ -262,50 +307,12 @@ int ehca_mmap(struct ib_ucontext *context, struct vm_area_struct *vma)
if (!qp->ib_qp.uobject || qp->ib_qp.uobject->context != context)
return -EINVAL;
- switch (rsrc_type) {
- case 1: /* galpa fw handle */
- ehca_dbg(qp->ib_qp.device, "qp=%p qp triggerarea", qp);
- vma->vm_flags |= VM_RESERVED;
- vsize = vma->vm_end - vma->vm_start;
- if (vsize != EHCA_PAGESIZE) {
- ehca_err(qp->ib_qp.device, "invalid vsize=%lx",
- vma->vm_end - vma->vm_start);
- return -EINVAL;
- }
-
- physical = qp->galpas.user.fw_handle;
- vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
- vma->vm_flags |= VM_IO | VM_RESERVED;
-
- ehca_dbg(qp->ib_qp.device, "vsize=%lx physical=%lx",
- vsize, physical);
- ret = remap_pfn_range(vma, vma->vm_start,
- physical >> PAGE_SHIFT, vsize,
- vma->vm_page_prot);
- if (ret) {
- ehca_err(qp->ib_qp.device,
- "remap_pfn_range() failed ret=%x",
- ret);
- return -ENOMEM;
- }
- break;
-
- case 2: /* qp rqueue_addr */
- ehca_dbg(qp->ib_qp.device, "qp=%p qp rqueue_addr", qp);
- vma->vm_flags |= VM_RESERVED;
- vma->vm_ops = &ehcau_vm_ops;
- break;
-
- case 3: /* qp squeue_addr */
- ehca_dbg(qp->ib_qp.device, "qp=%p qp squeue_addr", qp);
- vma->vm_flags |= VM_RESERVED;
- vma->vm_ops = &ehcau_vm_ops;
- break;
-
- default:
- ehca_err(qp->ib_qp.device, "bad resource type %x",
- rsrc_type);
- return -EINVAL;
+ ret = ehca_mmap_qp(vma, qp, rsrc_type);
+ if (unlikely(ret)) {
+ ehca_err(qp->ib_qp.device,
+ "ehca_mmap_qp() failed rc=%x qp_num=%x",
+ ret, qp->ib_qp.qp_num);
+ return ret;
}
break;
@@ -316,77 +323,3 @@ int ehca_mmap(struct ib_ucontext *context, struct vm_area_struct *vma)
return 0;
}
-
-int ehca_mmap_nopage(u64 foffset, u64 length, void **mapped,
- struct vm_area_struct **vma)
-{
- down_write(&current->mm->mmap_sem);
- *mapped = (void*)do_mmap(NULL,0, length, PROT_WRITE,
- MAP_SHARED | MAP_ANONYMOUS,
- foffset);
- up_write(&current->mm->mmap_sem);
- if (!(*mapped)) {
- ehca_gen_err("couldn't mmap foffset=%lx length=%lx",
- foffset, length);
- return -EINVAL;
- }
-
- *vma = find_vma(current->mm, (u64)*mapped);
- if (!(*vma)) {
- down_write(&current->mm->mmap_sem);
- do_munmap(current->mm, 0, length);
- up_write(&current->mm->mmap_sem);
- ehca_gen_err("couldn't find vma queue=%p", *mapped);
- return -EINVAL;
- }
- (*vma)->vm_flags |= VM_RESERVED;
- (*vma)->vm_ops = &ehcau_vm_ops;
-
- return 0;
-}
-
-int ehca_mmap_register(u64 physical, void **mapped,
- struct vm_area_struct **vma)
-{
- int ret;
- unsigned long vsize;
- /* ehca hw supports only 4k page */
- ret = ehca_mmap_nopage(0, EHCA_PAGESIZE, mapped, vma);
- if (ret) {
- ehca_gen_err("could'nt mmap physical=%lx", physical);
- return ret;
- }
-
- (*vma)->vm_flags |= VM_RESERVED;
- vsize = (*vma)->vm_end - (*vma)->vm_start;
- if (vsize != EHCA_PAGESIZE) {
- ehca_gen_err("invalid vsize=%lx",
- (*vma)->vm_end - (*vma)->vm_start);
- return -EINVAL;
- }
-
- (*vma)->vm_page_prot = pgprot_noncached((*vma)->vm_page_prot);
- (*vma)->vm_flags |= VM_IO | VM_RESERVED;
-
- ret = remap_pfn_range((*vma), (*vma)->vm_start,
- physical >> PAGE_SHIFT, vsize,
- (*vma)->vm_page_prot);
- if (ret) {
- ehca_gen_err("remap_pfn_range() failed ret=%x", ret);
- return -ENOMEM;
- }
-
- return 0;
-
-}
-
-int ehca_munmap(unsigned long addr, size_t len) {
- int ret = 0;
- struct mm_struct *mm = current->mm;
- if (mm) {
- down_write(&mm->mmap_sem);
- ret = do_munmap(mm, addr, len);
- up_write(&mm->mmap_sem);
- }
- return ret;
-}
diff --git a/drivers/infiniband/hw/ipath/Makefile b/drivers/infiniband/hw/ipath/Makefile
index 7dc10551cf1..ec2e603ea24 100644
--- a/drivers/infiniband/hw/ipath/Makefile
+++ b/drivers/infiniband/hw/ipath/Makefile
@@ -6,6 +6,7 @@ obj-$(CONFIG_INFINIBAND_IPATH) += ib_ipath.o
ib_ipath-y := \
ipath_cq.o \
ipath_diag.o \
+ ipath_dma.o \
ipath_driver.o \
ipath_eeprom.o \
ipath_file_ops.o \
diff --git a/drivers/infiniband/hw/ipath/ipath_dma.c b/drivers/infiniband/hw/ipath/ipath_dma.c
new file mode 100644
index 00000000000..6e0f2b8918c
--- /dev/null
+++ b/drivers/infiniband/hw/ipath/ipath_dma.c
@@ -0,0 +1,189 @@
+/*
+ * Copyright (c) 2006 QLogic, 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 <rdma/ib_verbs.h>
+
+#include "ipath_verbs.h"
+
+#define BAD_DMA_ADDRESS ((u64) 0)
+
+/*
+ * The following functions implement driver specific replacements
+ * for the ib_dma_*() functions.
+ *
+ * These functions return kernel virtual addresses instead of
+ * device bus addresses since the driver uses the CPU to copy
+ * data instead of using hardware DMA.
+ */
+
+static int ipath_mapping_error(struct ib_device *dev, u64 dma_addr)
+{
+ return dma_addr == BAD_DMA_ADDRESS;
+}
+
+static u64 ipath_dma_map_single(struct ib_device *dev,
+ void *cpu_addr, size_t size,
+ enum dma_data_direction direction)
+{
+ BUG_ON(!valid_dma_direction(direction));
+ return (u64) cpu_addr;
+}
+
+static void ipath_dma_unmap_single(struct ib_device *dev,
+ u64 addr, size_t size,
+ enum dma_data_direction direction)
+{
+ BUG_ON(!valid_dma_direction(direction));
+}
+
+static u64 ipath_dma_map_page(struct ib_device *dev,
+ struct page *page,
+ unsigned long offset,
+ size_t size,
+ enum dma_data_direction direction)
+{
+ u64 addr;
+
+ BUG_ON(!valid_dma_direction(direction));
+
+ if (offset + size > PAGE_SIZE) {
+ addr = BAD_DMA_ADDRESS;
+ goto done;
+ }
+
+ addr = (u64) page_address(page);
+ if (addr)
+ addr += offset;
+ /* TODO: handle highmem pages */
+
+done:
+ return addr;
+}
+
+static void ipath_dma_unmap_page(struct ib_device *dev,
+ u64 addr, size_t size,
+ enum dma_data_direction direction)
+{
+ BUG_ON(!valid_dma_direction(direction));
+}
+
+int ipath_map_sg(struct ib_device *dev, struct scatterlist *sg, int nents,
+ enum dma_data_direction direction)
+{
+ u64 addr;
+ int i;
+ int ret = nents;
+
+ BUG_ON(!valid_dma_direction(direction));
+
+ for (i = 0; i < nents; i++) {
+ addr = (u64) page_address(sg[i].page);
+ /* TODO: handle highmem pages */
+ if (!addr) {
+ ret = 0;
+ break;
+ }
+ }
+ return ret;
+}
+
+static void ipath_unmap_sg(struct ib_device *dev,
+ struct scatterlist *sg, int nents,
+ enum dma_data_direction direction)
+{
+ BUG_ON(!valid_dma_direction(direction));
+}
+
+static u64 ipath_sg_dma_address(struct ib_device *dev, struct scatterlist *sg)
+{
+ u64 addr = (u64) page_address(sg->page);
+
+ if (addr)
+ addr += sg->offset;
+ return addr;
+}
+
+static unsigned int ipath_sg_dma_len(struct ib_device *dev,
+ struct scatterlist *sg)
+{
+ return sg->length;
+}
+
+static void ipath_sync_single_for_cpu(struct ib_device *dev,
+ u64 addr,
+ size_t size,
+ enum dma_data_direction dir)
+{
+}
+
+static void ipath_sync_single_for_device(struct ib_device *dev,
+ u64 addr,
+ size_t size,
+ enum dma_data_direction dir)
+{
+}
+
+static void *ipath_dma_alloc_coherent(struct ib_device *dev, size_t size,
+ u64 *dma_handle, gfp_t flag)
+{
+ struct page *p;
+ void *addr = NULL;
+
+ p = alloc_pages(flag, get_order(size));
+ if (p)
+ addr = page_address(p);
+ if (dma_handle)
+ *dma_handle = (u64) addr;
+ return addr;
+}
+
+static void ipath_dma_free_coherent(struct ib_device *dev, size_t size,
+ void *cpu_addr, dma_addr_t dma_handle)
+{
+ free_pages((unsigned long) cpu_addr, get_order(size));
+}
+
+struct ib_dma_mapping_ops ipath_dma_mapping_ops = {
+ ipath_mapping_error,
+ ipath_dma_map_single,
+ ipath_dma_unmap_single,
+ ipath_dma_map_page,
+ ipath_dma_unmap_page,
+ ipath_map_sg,
+ ipath_unmap_sg,
+ ipath_sg_dma_address,
+ ipath_sg_dma_len,
+ ipath_sync_single_for_cpu,
+ ipath_sync_single_for_device,
+ ipath_dma_alloc_coherent,
+ ipath_dma_free_coherent
+};
diff --git a/drivers/infiniband/hw/ipath/ipath_driver.c b/drivers/infiniband/hw/ipath/ipath_driver.c
index 1aeddb48e35..ae7f21a0cdc 100644
--- a/drivers/infiniband/hw/ipath/ipath_driver.c
+++ b/drivers/infiniband/hw/ipath/ipath_driver.c
@@ -1825,8 +1825,6 @@ void ipath_write_kreg_port(const struct ipath_devdata *dd, ipath_kreg regno,
*/
void ipath_shutdown_device(struct ipath_devdata *dd)
{
- u64 val;
-
ipath_dbg("Shutting down the device\n");
dd->ipath_flags |= IPATH_LINKUNK;
@@ -1849,7 +1847,7 @@ void ipath_shutdown_device(struct ipath_devdata *dd)
*/
ipath_write_kreg(dd, dd->ipath_kregs->kr_sendctrl, 0ULL);
/* flush it */
- val = ipath_read_kreg64(dd, dd->ipath_kregs->kr_scratch);
+ ipath_read_kreg64(dd, dd->ipath_kregs->kr_scratch);
/*
* enough for anything that's going to trickle out to have actually
* done so.
diff --git a/drivers/infiniband/hw/ipath/ipath_file_ops.c b/drivers/infiniband/hw/ipath/ipath_file_ops.c
index a9ddc6911f6..b932bcb67a5 100644
--- a/drivers/infiniband/hw/ipath/ipath_file_ops.c
+++ b/drivers/infiniband/hw/ipath/ipath_file_ops.c
@@ -699,7 +699,6 @@ static int ipath_manage_rcvq(struct ipath_portdata *pd, unsigned subport,
int start_stop)
{
struct ipath_devdata *dd = pd->port_dd;
- u64 tval;
ipath_cdbg(PROC, "%sabling rcv for unit %u port %u:%u\n",
start_stop ? "en" : "dis", dd->ipath_unit,
@@ -729,7 +728,7 @@ static int ipath_manage_rcvq(struct ipath_portdata *pd, unsigned subport,
ipath_write_kreg(dd, dd->ipath_kregs->kr_rcvctrl,
dd->ipath_rcvctrl);
/* now be sure chip saw it before we return */
- tval = ipath_read_kreg64(dd, dd->ipath_kregs->kr_scratch);
+ ipath_read_kreg64(dd, dd->ipath_kregs->kr_scratch);
if (start_stop) {
/*
* And try to be sure that tail reg update has happened too.
@@ -738,7 +737,7 @@ static int ipath_manage_rcvq(struct ipath_portdata *pd, unsigned subport,
* in memory copy, since we could overwrite an update by the
* chip if we did.
*/
- tval = ipath_read_ureg32(dd, ur_rcvhdrtail, pd->port_port);
+ ipath_read_ureg32(dd, ur_rcvhdrtail, pd->port_port);
}
/* always; new head should be equal to new tail; see above */
bail:
@@ -1745,9 +1744,9 @@ static int ipath_assign_port(struct file *fp,
goto done;
}
- i_minor = iminor(fp->f_dentry->d_inode) - IPATH_USER_MINOR_BASE;
+ i_minor = iminor(fp->f_path.dentry->d_inode) - IPATH_USER_MINOR_BASE;
ipath_cdbg(VERBOSE, "open on dev %lx (minor %d)\n",
- (long)fp->f_dentry->d_inode->i_rdev, i_minor);
+ (long)fp->f_path.dentry->d_inode->i_rdev, i_minor);
if (i_minor)
ret = find_free_port(i_minor - 1, fp, uinfo);
diff --git a/drivers/infiniband/hw/ipath/ipath_fs.c b/drivers/infiniband/hw/ipath/ipath_fs.c
index d9ff283f725..79a60f020a2 100644
--- a/drivers/infiniband/hw/ipath/ipath_fs.c
+++ b/drivers/infiniband/hw/ipath/ipath_fs.c
@@ -118,7 +118,7 @@ static ssize_t atomic_counters_read(struct file *file, char __user *buf,
u16 i;
struct ipath_devdata *dd;
- dd = file->f_dentry->d_inode->i_private;
+ dd = file->f_path.dentry->d_inode->i_private;
for (i = 0; i < NUM_COUNTERS; i++)
counters[i] = ipath_snap_cntr(dd, i);
@@ -138,7 +138,7 @@ static ssize_t atomic_node_info_read(struct file *file, char __user *buf,
struct ipath_devdata *dd;
u64 guid;
- dd = file->f_dentry->d_inode->i_private;
+ dd = file->f_path.dentry->d_inode->i_private;
guid = be64_to_cpu(dd->ipath_guid);
@@ -177,7 +177,7 @@ static ssize_t atomic_port_info_read(struct file *file, char __user *buf,
u32 tmp, tmp2;
struct ipath_devdata *dd;
- dd = file->f_dentry->d_inode->i_private;
+ dd = file->f_path.dentry->d_inode->i_private;
/* so we only initialize non-zero fields. */
memset(portinfo, 0, sizeof portinfo);
@@ -324,7 +324,7 @@ static ssize_t flash_read(struct file *file, char __user *buf,
goto bail;
}
- dd = file->f_dentry->d_inode->i_private;
+ dd = file->f_path.dentry->d_inode->i_private;
if (ipath_eeprom_read(dd, pos, tmp, count)) {
ipath_dev_err(dd, "failed to read from flash\n");
ret = -ENXIO;
@@ -377,7 +377,7 @@ static ssize_t flash_write(struct file *file, const char __user *buf,
goto bail_tmp;
}
- dd = file->f_dentry->d_inode->i_private;
+ dd = file->f_path.dentry->d_inode->i_private;
if (ipath_eeprom_write(dd, pos, tmp, count)) {
ret = -ENXIO;
ipath_dev_err(dd, "failed to write to flash\n");
diff --git a/drivers/infiniband/hw/ipath/ipath_iba6110.c b/drivers/infiniband/hw/ipath/ipath_iba6110.c
index e57c7a351cb..7468477ba83 100644
--- a/drivers/infiniband/hw/ipath/ipath_iba6110.c
+++ b/drivers/infiniband/hw/ipath/ipath_iba6110.c
@@ -1447,7 +1447,7 @@ static void ipath_ht_tidtemplate(struct ipath_devdata *dd)
static int ipath_ht_early_init(struct ipath_devdata *dd)
{
u32 __iomem *piobuf;
- u32 pioincr, val32, egrsize;
+ u32 pioincr, val32;
int i;
/*
@@ -1467,7 +1467,6 @@ static int ipath_ht_early_init(struct ipath_devdata *dd)
* errors interrupts if we ever see one).
*/
dd->ipath_rcvegrbufsize = dd->ipath_piosize2k;
- egrsize = dd->ipath_rcvegrbufsize;
/*
* the min() check here is currently a nop, but it may not
diff --git a/drivers/infiniband/hw/ipath/ipath_iba6120.c b/drivers/infiniband/hw/ipath/ipath_iba6120.c
index 6af89683f71..ae8bf9950c6 100644
--- a/drivers/infiniband/hw/ipath/ipath_iba6120.c
+++ b/drivers/infiniband/hw/ipath/ipath_iba6120.c
@@ -602,7 +602,7 @@ static void ipath_pe_init_hwerrors(struct ipath_devdata *dd)
*/
static int ipath_pe_bringup_serdes(struct ipath_devdata *dd)
{
- u64 val, tmp, config1, prev_val;
+ u64 val, config1, prev_val;
int ret = 0;
ipath_dbg("Trying to bringup serdes\n");
@@ -633,7 +633,7 @@ static int ipath_pe_bringup_serdes(struct ipath_devdata *dd)
| INFINIPATH_SERDC0_L1PWR_DN;
ipath_write_kreg(dd, dd->ipath_kregs->kr_serdesconfig0, val);
/* be sure chip saw it */
- tmp = ipath_read_kreg64(dd, dd->ipath_kregs->kr_scratch);
+ ipath_read_kreg64(dd, dd->ipath_kregs->kr_scratch);
udelay(5); /* need pll reset set at least for a bit */
/*
* after PLL is reset, set the per-lane Resets and TxIdle and
@@ -647,7 +647,7 @@ static int ipath_pe_bringup_serdes(struct ipath_devdata *dd)
"and txidle (%llx)\n", (unsigned long long) val);
ipath_write_kreg(dd, dd->ipath_kregs->kr_serdesconfig0, val);
/* be sure chip saw it */
- tmp = ipath_read_kreg64(dd, dd->ipath_kregs->kr_scratch);
+ ipath_read_kreg64(dd, dd->ipath_kregs->kr_scratch);
/* need PLL reset clear for at least 11 usec before lane
* resets cleared; give it a few more to be sure */
udelay(15);
@@ -851,12 +851,12 @@ static int ipath_setup_pe_config(struct ipath_devdata *dd,
int pos, ret;
dd->ipath_msi_lo = 0; /* used as a flag during reset processing */
- dd->ipath_irq = pdev->irq;
ret = pci_enable_msi(dd->pcidev);
if (ret)
ipath_dev_err(dd, "pci_enable_msi failed: %d, "
"interrupts may not work\n", ret);
/* continue even if it fails, we may still be OK... */
+ dd->ipath_irq = pdev->irq;
if ((pos = pci_find_capability(dd->pcidev, PCI_CAP_ID_MSI))) {
u16 control;
diff --git a/drivers/infiniband/hw/ipath/ipath_init_chip.c b/drivers/infiniband/hw/ipath/ipath_init_chip.c
index d819cca524c..d4f6b5239ef 100644
--- a/drivers/infiniband/hw/ipath/ipath_init_chip.c
+++ b/drivers/infiniband/hw/ipath/ipath_init_chip.c
@@ -347,10 +347,9 @@ done:
static int init_chip_reset(struct ipath_devdata *dd,
struct ipath_portdata **pdp)
{
- struct ipath_portdata *pd;
u32 rtmp;
- *pdp = pd = dd->ipath_pd[0];
+ *pdp = dd->ipath_pd[0];
/* ensure chip does no sends or receives while we re-initialize */
dd->ipath_control = dd->ipath_sendctrl = dd->ipath_rcvctrl = 0U;
ipath_write_kreg(dd, dd->ipath_kregs->kr_rcvctrl, 0);
diff --git a/drivers/infiniband/hw/ipath/ipath_intr.c b/drivers/infiniband/hw/ipath/ipath_intr.c
index 5652a550d44..72b9e279d19 100644
--- a/drivers/infiniband/hw/ipath/ipath_intr.c
+++ b/drivers/infiniband/hw/ipath/ipath_intr.c
@@ -598,10 +598,9 @@ static int handle_errors(struct ipath_devdata *dd, ipath_err_t errs)
* on close
*/
if (errs & INFINIPATH_E_RRCVHDRFULL) {
- int any;
u32 hd, tl;
ipath_stats.sps_hdrqfull++;
- for (any = i = 0; i < dd->ipath_cfgports; i++) {
+ for (i = 0; i < dd->ipath_cfgports; i++) {
struct ipath_portdata *pd = dd->ipath_pd[i];
if (i == 0) {
hd = dd->ipath_port0head;
diff --git a/drivers/infiniband/hw/ipath/ipath_keys.c b/drivers/infiniband/hw/ipath/ipath_keys.c
index 9a6cbd05adc..851763d7d2d 100644
--- a/drivers/infiniband/hw/ipath/ipath_keys.c
+++ b/drivers/infiniband/hw/ipath/ipath_keys.c
@@ -134,7 +134,7 @@ int ipath_lkey_ok(struct ipath_qp *qp, struct ipath_sge *isge,
*/
if (sge->lkey == 0) {
isge->mr = NULL;
- isge->vaddr = bus_to_virt(sge->addr);
+ isge->vaddr = (void *) sge->addr;
isge->length = sge->length;
isge->sge_length = sge->length;
ret = 1;
@@ -202,12 +202,12 @@ int ipath_rkey_ok(struct ipath_qp *qp, struct ipath_sge_state *ss,
int ret;
/*
- * We use RKEY == zero for physical addresses
- * (see ipath_get_dma_mr).
+ * We use RKEY == zero for kernel virtual addresses
+ * (see ipath_get_dma_mr and ipath_dma.c).
*/
if (rkey == 0) {
sge->mr = NULL;
- sge->vaddr = phys_to_virt(vaddr);
+ sge->vaddr = (void *) vaddr;
sge->length = len;
sge->sge_length = len;
ss->sg_list = NULL;
diff --git a/drivers/infiniband/hw/ipath/ipath_mr.c b/drivers/infiniband/hw/ipath/ipath_mr.c
index a0673c1eef7..8cc8598d6c6 100644
--- a/drivers/infiniband/hw/ipath/ipath_mr.c
+++ b/drivers/infiniband/hw/ipath/ipath_mr.c
@@ -54,6 +54,8 @@ static inline struct ipath_fmr *to_ifmr(struct ib_fmr *ibfmr)
* @acc: access flags
*
* Returns the memory region on success, otherwise returns an errno.
+ * Note that all DMA addresses should be created via the
+ * struct ib_dma_mapping_ops functions (see ipath_dma.c).
*/
struct ib_mr *ipath_get_dma_mr(struct ib_pd *pd, int acc)
{
@@ -149,8 +151,7 @@ struct ib_mr *ipath_reg_phys_mr(struct ib_pd *pd,
m = 0;
n = 0;
for (i = 0; i < num_phys_buf; i++) {
- mr->mr.map[m]->segs[n].vaddr =
- phys_to_virt(buffer_list[i].addr);
+ mr->mr.map[m]->segs[n].vaddr = (void *) buffer_list[i].addr;
mr->mr.map[m]->segs[n].length = buffer_list[i].size;
mr->mr.length += buffer_list[i].size;
n++;
@@ -347,7 +348,7 @@ int ipath_map_phys_fmr(struct ib_fmr *ibfmr, u64 * page_list,
n = 0;
ps = 1 << fmr->page_shift;
for (i = 0; i < list_len; i++) {
- fmr->mr.map[m]->segs[n].vaddr = phys_to_virt(page_list[i]);
+ fmr->mr.map[m]->segs[n].vaddr = (void *) page_list[i];
fmr->mr.map[m]->segs[n].length = ps;
if (++n == IPATH_SEGSZ) {
m++;
diff --git a/drivers/infiniband/hw/ipath/ipath_qp.c b/drivers/infiniband/hw/ipath/ipath_qp.c
index 46c1c89bf6a..64f07b19349 100644
--- a/drivers/infiniband/hw/ipath/ipath_qp.c
+++ b/drivers/infiniband/hw/ipath/ipath_qp.c
@@ -379,7 +379,7 @@ void ipath_error_qp(struct ipath_qp *qp, enum ib_wc_status err)
wc.vendor_err = 0;
wc.byte_len = 0;
wc.imm_data = 0;
- wc.qp_num = qp->ibqp.qp_num;
+ wc.qp = &qp->ibqp;
wc.src_qp = 0;
wc.wc_flags = 0;
wc.pkey_index = 0;
diff --git a/drivers/infiniband/hw/ipath/ipath_rc.c b/drivers/infiniband/hw/ipath/ipath_rc.c
index ce6038743c5..5ff20cb0449 100644
--- a/drivers/infiniband/hw/ipath/ipath_rc.c
+++ b/drivers/infiniband/hw/ipath/ipath_rc.c
@@ -702,7 +702,7 @@ void ipath_restart_rc(struct ipath_qp *qp, u32 psn, struct ib_wc *wc)
wc->opcode = ib_ipath_wc_opcode[wqe->wr.opcode];
wc->vendor_err = 0;
wc->byte_len = 0;
- wc->qp_num = qp->ibqp.qp_num;
+ wc->qp = &qp->ibqp;
wc->src_qp = qp->remote_qpn;
wc->pkey_index = 0;
wc->slid = qp->remote_ah_attr.dlid;
@@ -836,7 +836,7 @@ static int do_rc_ack(struct ipath_qp *qp, u32 aeth, u32 psn, int opcode)
wc.opcode = ib_ipath_wc_opcode[wqe->wr.opcode];
wc.vendor_err = 0;
wc.byte_len = wqe->length;
- wc.qp_num = qp->ibqp.qp_num;
+ wc.qp = &qp->ibqp;
wc.src_qp = qp->remote_qpn;
wc.pkey_index = 0;
wc.slid = qp->remote_ah_attr.dlid;
@@ -951,7 +951,7 @@ static int do_rc_ack(struct ipath_qp *qp, u32 aeth, u32 psn, int opcode)
wc.opcode = ib_ipath_wc_opcode[wqe->wr.opcode];
wc.vendor_err = 0;
wc.byte_len = 0;
- wc.qp_num = qp->ibqp.qp_num;
+ wc.qp = &qp->ibqp;
wc.src_qp = qp->remote_qpn;
wc.pkey_index = 0;
wc.slid = qp->remote_ah_attr.dlid;
@@ -1511,7 +1511,7 @@ void ipath_rc_rcv(struct ipath_ibdev *dev, struct ipath_ib_header *hdr,
wc.status = IB_WC_SUCCESS;
wc.opcode = IB_WC_RECV;
wc.vendor_err = 0;
- wc.qp_num = qp->ibqp.qp_num;
+ wc.qp = &qp->ibqp;
wc.src_qp = qp->remote_qpn;
wc.pkey_index = 0;
wc.slid = qp->remote_ah_attr.dlid;
diff --git a/drivers/infiniband/hw/ipath/ipath_ruc.c b/drivers/infiniband/hw/ipath/ipath_ruc.c
index f7530512045..e86cb171872 100644
--- a/drivers/infiniband/hw/ipath/ipath_ruc.c
+++ b/drivers/infiniband/hw/ipath/ipath_ruc.c
@@ -137,7 +137,7 @@ bad_lkey:
wc.vendor_err = 0;
wc.byte_len = 0;
wc.imm_data = 0;
- wc.qp_num = qp->ibqp.qp_num;
+ wc.qp = &qp->ibqp;
wc.src_qp = 0;
wc.wc_flags = 0;
wc.pkey_index = 0;
@@ -336,7 +336,7 @@ again:
wc.opcode = ib_ipath_wc_opcode[wqe->wr.opcode];
wc.vendor_err = 0;
wc.byte_len = 0;
- wc.qp_num = sqp->ibqp.qp_num;
+ wc.qp = &sqp->ibqp;
wc.src_qp = sqp->remote_qpn;
wc.pkey_index = 0;
wc.slid = sqp->remote_ah_attr.dlid;
@@ -426,7 +426,7 @@ again:
wc.status = IB_WC_SUCCESS;
wc.vendor_err = 0;
wc.byte_len = wqe->length;
- wc.qp_num = qp->ibqp.qp_num;
+ wc.qp = &qp->ibqp;
wc.src_qp = qp->remote_qpn;
/* XXX do we know which pkey matched? Only needed for GSI. */
wc.pkey_index = 0;
@@ -447,7 +447,7 @@ send_comp:
wc.opcode = ib_ipath_wc_opcode[wqe->wr.opcode];
wc.vendor_err = 0;
wc.byte_len = wqe->length;
- wc.qp_num = sqp->ibqp.qp_num;
+ wc.qp = &sqp->ibqp;
wc.src_qp = 0;
wc.pkey_index = 0;
wc.slid = 0;
diff --git a/drivers/infiniband/hw/ipath/ipath_sysfs.c b/drivers/infiniband/hw/ipath/ipath_sysfs.c
index 182de34f9f4..ffa6318ad0c 100644
--- a/drivers/infiniband/hw/ipath/ipath_sysfs.c
+++ b/drivers/infiniband/hw/ipath/ipath_sysfs.c
@@ -215,7 +215,6 @@ static ssize_t store_mlid(struct device *dev,
size_t count)
{
struct ipath_devdata *dd = dev_get_drvdata(dev);
- int unit;
u16 mlid;
int ret;
@@ -223,8 +222,6 @@ static ssize_t store_mlid(struct device *dev,
if (ret < 0 || mlid < IPATH_MULTICAST_LID_BASE)
goto invalid;
- unit = dd->ipath_unit;
-
dd->ipath_mlid = mlid;
goto bail;
diff --git a/drivers/infiniband/hw/ipath/ipath_uc.c b/drivers/infiniband/hw/ipath/ipath_uc.c
index e636cfd67a8..325d6634ff5 100644
--- a/drivers/infiniband/hw/ipath/ipath_uc.c
+++ b/drivers/infiniband/hw/ipath/ipath_uc.c
@@ -49,7 +49,7 @@ static void complete_last_send(struct ipath_qp *qp, struct ipath_swqe *wqe,
wc->opcode = ib_ipath_wc_opcode[wqe->wr.opcode];
wc->vendor_err = 0;
wc->byte_len = wqe->length;
- wc->qp_num = qp->ibqp.qp_num;
+ wc->qp = &qp->ibqp;
wc->src_qp = qp->remote_qpn;
wc->pkey_index = 0;
wc->slid = qp->remote_ah_attr.dlid;
@@ -411,7 +411,7 @@ void ipath_uc_rcv(struct ipath_ibdev *dev, struct ipath_ib_header *hdr,
wc.status = IB_WC_SUCCESS;
wc.opcode = IB_WC_RECV;
wc.vendor_err = 0;
- wc.qp_num = qp->ibqp.qp_num;
+ wc.qp = &qp->ibqp;
wc.src_qp = qp->remote_qpn;
wc.pkey_index = 0;
wc.slid = qp->remote_ah_attr.dlid;
diff --git a/drivers/infiniband/hw/ipath/ipath_ud.c b/drivers/infiniband/hw/ipath/ipath_ud.c
index 49f1102af8b..9a3e54664ee 100644
--- a/drivers/infiniband/hw/ipath/ipath_ud.c
+++ b/drivers/infiniband/hw/ipath/ipath_ud.c
@@ -66,7 +66,7 @@ bad_lkey:
wc.vendor_err = 0;
wc.byte_len = 0;
wc.imm_data = 0;
- wc.qp_num = qp->ibqp.qp_num;
+ wc.qp = &qp->ibqp;
wc.src_qp = 0;
wc.wc_flags = 0;
wc.pkey_index = 0;
@@ -255,7 +255,7 @@ static void ipath_ud_loopback(struct ipath_qp *sqp,
wc->status = IB_WC_SUCCESS;
wc->opcode = IB_WC_RECV;
wc->vendor_err = 0;
- wc->qp_num = qp->ibqp.qp_num;
+ wc->qp = &qp->ibqp;
wc->src_qp = sqp->ibqp.qp_num;
/* XXX do we know which pkey matched? Only needed for GSI. */
wc->pkey_index = 0;
@@ -474,7 +474,7 @@ done:
wc.vendor_err = 0;
wc.opcode = IB_WC_SEND;
wc.byte_len = len;
- wc.qp_num = qp->ibqp.qp_num;
+ wc.qp = &qp->ibqp;
wc.src_qp = 0;
wc.wc_flags = 0;
/* XXX initialize other fields? */
@@ -651,7 +651,7 @@ void ipath_ud_rcv(struct ipath_ibdev *dev, struct ipath_ib_header *hdr,
wc.status = IB_WC_SUCCESS;
wc.opcode = IB_WC_RECV;
wc.vendor_err = 0;
- wc.qp_num = qp->ibqp.qp_num;
+ wc.qp = &qp->ibqp;
wc.src_qp = src_qp;
/* XXX do we know which pkey matched? Only needed for GSI. */
wc.pkey_index = 0;
diff --git a/drivers/infiniband/hw/ipath/ipath_verbs.c b/drivers/infiniband/hw/ipath/ipath_verbs.c
index acdee33ee1f..2aaacdb7e52 100644
--- a/drivers/infiniband/hw/ipath/ipath_verbs.c
+++ b/drivers/infiniband/hw/ipath/ipath_verbs.c
@@ -1599,6 +1599,7 @@ int ipath_register_ib_device(struct ipath_devdata *dd)
dev->detach_mcast = ipath_multicast_detach;
dev->process_mad = ipath_process_mad;
dev->mmap = ipath_mmap;
+ dev->dma_ops = &ipath_dma_mapping_ops;
snprintf(dev->node_desc, sizeof(dev->node_desc),
IPATH_IDSTR " %s", init_utsname()->nodename);
diff --git a/drivers/infiniband/hw/ipath/ipath_verbs.h b/drivers/infiniband/hw/ipath/ipath_verbs.h
index 8039f6e5f0c..c0c8d5b24a7 100644
--- a/drivers/infiniband/hw/ipath/ipath_verbs.h
+++ b/drivers/infiniband/hw/ipath/ipath_verbs.h
@@ -812,4 +812,6 @@ extern unsigned int ib_ipath_max_srq_wrs;
extern const u32 ib_ipath_rnr_table[];
+extern struct ib_dma_mapping_ops ipath_dma_mapping_ops;
+
#endif /* IPATH_VERBS_H */
diff --git a/drivers/infiniband/hw/mthca/mthca_cmd.c b/drivers/infiniband/hw/mthca/mthca_cmd.c
index 768df7265b8..968d1519761 100644
--- a/drivers/infiniband/hw/mthca/mthca_cmd.c
+++ b/drivers/infiniband/hw/mthca/mthca_cmd.c
@@ -1854,7 +1854,7 @@ int mthca_MAD_IFC(struct mthca_dev *dev, int ignore_mkey, int ignore_bkey,
memset(inbox + 256, 0, 256);
- MTHCA_PUT(inbox, in_wc->qp_num, MAD_IFC_MY_QPN_OFFSET);
+ MTHCA_PUT(inbox, in_wc->qp->qp_num, MAD_IFC_MY_QPN_OFFSET);
MTHCA_PUT(inbox, in_wc->src_qp, MAD_IFC_RQPN_OFFSET);
val = in_wc->sl << 4;
diff --git a/drivers/infiniband/hw/mthca/mthca_cq.c b/drivers/infiniband/hw/mthca/mthca_cq.c
index 283d50b76c3..efd79ef109a 100644
--- a/drivers/infiniband/hw/mthca/mthca_cq.c
+++ b/drivers/infiniband/hw/mthca/mthca_cq.c
@@ -54,6 +54,10 @@ enum {
MTHCA_CQ_ENTRY_SIZE = 0x20
};
+enum {
+ MTHCA_ATOMIC_BYTE_LEN = 8
+};
+
/*
* Must be packed because start is 64 bits but only aligned to 32 bits.
*/
@@ -530,7 +534,7 @@ static inline int mthca_poll_one(struct mthca_dev *dev,
}
}
- entry->qp_num = (*cur_qp)->qpn;
+ entry->qp = &(*cur_qp)->ibqp;
if (is_send) {
wq = &(*cur_qp)->sq;
@@ -599,11 +603,11 @@ static inline int mthca_poll_one(struct mthca_dev *dev,
break;
case MTHCA_OPCODE_ATOMIC_CS:
entry->opcode = IB_WC_COMP_SWAP;
- entry->byte_len = be32_to_cpu(cqe->byte_cnt);
+ entry->byte_len = MTHCA_ATOMIC_BYTE_LEN;
break;
case MTHCA_OPCODE_ATOMIC_FA:
entry->opcode = IB_WC_FETCH_ADD;
- entry->byte_len = be32_to_cpu(cqe->byte_cnt);
+ entry->byte_len = MTHCA_ATOMIC_BYTE_LEN;
break;
case MTHCA_OPCODE_BIND_MW:
entry->opcode = IB_WC_BIND_MW;
diff --git a/drivers/infiniband/hw/mthca/mthca_main.c b/drivers/infiniband/hw/mthca/mthca_main.c
index 0491ec7a7c0..44bc6cc734a 100644
--- a/drivers/infiniband/hw/mthca/mthca_main.c
+++ b/drivers/infiniband/hw/mthca/mthca_main.c
@@ -80,24 +80,61 @@ static int tune_pci = 0;
module_param(tune_pci, int, 0444);
MODULE_PARM_DESC(tune_pci, "increase PCI burst from the default set by BIOS if nonzero");
-struct mutex mthca_device_mutex;
+DEFINE_MUTEX(mthca_device_mutex);
+
+#define MTHCA_DEFAULT_NUM_QP (1 << 16)
+#define MTHCA_DEFAULT_RDB_PER_QP (1 << 2)
+#define MTHCA_DEFAULT_NUM_CQ (1 << 16)
+#define MTHCA_DEFAULT_NUM_MCG (1 << 13)
+#define MTHCA_DEFAULT_NUM_MPT (1 << 17)
+#define MTHCA_DEFAULT_NUM_MTT (1 << 20)
+#define MTHCA_DEFAULT_NUM_UDAV (1 << 15)
+#define MTHCA_DEFAULT_NUM_RESERVED_MTTS (1 << 18)
+#define MTHCA_DEFAULT_NUM_UARC_SIZE (1 << 18)
+
+static struct mthca_profile hca_profile = {
+ .num_qp = MTHCA_DEFAULT_NUM_QP,
+ .rdb_per_qp = MTHCA_DEFAULT_RDB_PER_QP,
+ .num_cq = MTHCA_DEFAULT_NUM_CQ,
+ .num_mcg = MTHCA_DEFAULT_NUM_MCG,
+ .num_mpt = MTHCA_DEFAULT_NUM_MPT,
+ .num_mtt = MTHCA_DEFAULT_NUM_MTT,
+ .num_udav = MTHCA_DEFAULT_NUM_UDAV, /* Tavor only */
+ .fmr_reserved_mtts = MTHCA_DEFAULT_NUM_RESERVED_MTTS, /* Tavor only */
+ .uarc_size = MTHCA_DEFAULT_NUM_UARC_SIZE, /* Arbel only */
+};
+
+module_param_named(num_qp, hca_profile.num_qp, int, 0444);
+MODULE_PARM_DESC(num_qp, "maximum number of QPs per HCA");
+
+module_param_named(rdb_per_qp, hca_profile.rdb_per_qp, int, 0444);
+MODULE_PARM_DESC(rdb_per_qp, "number of RDB buffers per QP");
+
+module_param_named(num_cq, hca_profile.num_cq, int, 0444);
+MODULE_PARM_DESC(num_cq, "maximum number of CQs per HCA");
+
+module_param_named(num_mcg, hca_profile.num_mcg, int, 0444);
+MODULE_PARM_DESC(num_mcg, "maximum number of multicast groups per HCA");
+
+module_param_named(num_mpt, hca_profile.num_mpt, int, 0444);
+MODULE_PARM_DESC(num_mpt,
+ "maximum number of memory protection table entries per HCA");
+
+module_param_named(num_mtt, hca_profile.num_mtt, int, 0444);
+MODULE_PARM_DESC(num_mtt,
+ "maximum number of memory translation table segments per HCA");
+
+module_param_named(num_udav, hca_profile.num_udav, int, 0444);
+MODULE_PARM_DESC(num_udav, "maximum number of UD address vectors per HCA");
+
+module_param_named(fmr_reserved_mtts, hca_profile.fmr_reserved_mtts, int, 0444);
+MODULE_PARM_DESC(fmr_reserved_mtts,
+ "number of memory translation table segments reserved for FMR");
static const char mthca_version[] __devinitdata =
DRV_NAME ": Mellanox InfiniBand HCA driver v"
DRV_VERSION " (" DRV_RELDATE ")\n";
-static struct mthca_profile default_profile = {
- .num_qp = 1 << 16,
- .rdb_per_qp = 4,
- .num_cq = 1 << 16,
- .num_mcg = 1 << 13,
- .num_mpt = 1 << 17,
- .num_mtt = 1 << 20,
- .num_udav = 1 << 15, /* Tavor only */
- .fmr_reserved_mtts = 1 << 18, /* Tavor only */
- .uarc_size = 1 << 18, /* Arbel only */
-};
-
static int mthca_tune_pci(struct mthca_dev *mdev)
{
int cap;
@@ -303,7 +340,7 @@ static int mthca_init_tavor(struct mthca_dev *mdev)
goto err_disable;
}
- profile = default_profile;
+ profile = hca_profile;
profile.num_uar = dev_lim.uar_size / PAGE_SIZE;
profile.uarc_size = 0;
if (mdev->mthca_flags & MTHCA_FLAG_SRQ)
@@ -621,7 +658,7 @@ static int mthca_init_arbel(struct mthca_dev *mdev)
goto err_stop_fw;
}
- profile = default_profile;
+ profile = hca_profile;
profile.num_uar = dev_lim.uar_size / PAGE_SIZE;
profile.num_udav = 0;
if (mdev->mthca_flags & MTHCA_FLAG_SRQ)
@@ -1278,11 +1315,55 @@ static struct pci_driver mthca_driver = {
.remove = __devexit_p(mthca_remove_one)
};
+static void __init __mthca_check_profile_val(const char *name, int *pval,
+ int pval_default)
+{
+ /* value must be positive and power of 2 */
+ int old_pval = *pval;
+
+ if (old_pval <= 0)
+ *pval = pval_default;
+ else
+ *pval = roundup_pow_of_two(old_pval);
+
+ if (old_pval != *pval) {
+ printk(KERN_WARNING PFX "Invalid value %d for %s in module parameter.\n",
+ old_pval, name);
+ printk(KERN_WARNING PFX "Corrected %s to %d.\n", name, *pval);
+ }
+}
+
+#define mthca_check_profile_val(name, default) \
+ __mthca_check_profile_val(#name, &hca_profile.name, default)
+
+static void __init mthca_validate_profile(void)
+{
+ mthca_check_profile_val(num_qp, MTHCA_DEFAULT_NUM_QP);
+ mthca_check_profile_val(rdb_per_qp, MTHCA_DEFAULT_RDB_PER_QP);
+ mthca_check_profile_val(num_cq, MTHCA_DEFAULT_NUM_CQ);
+ mthca_check_profile_val(num_mcg, MTHCA_DEFAULT_NUM_MCG);
+ mthca_check_profile_val(num_mpt, MTHCA_DEFAULT_NUM_MPT);
+ mthca_check_profile_val(num_mtt, MTHCA_DEFAULT_NUM_MTT);
+ mthca_check_profile_val(num_udav, MTHCA_DEFAULT_NUM_UDAV);
+ mthca_check_profile_val(fmr_reserved_mtts, MTHCA_DEFAULT_NUM_RESERVED_MTTS);
+
+ if (hca_profile.fmr_reserved_mtts >= hca_profile.num_mtt) {
+ printk(KERN_WARNING PFX "Invalid fmr_reserved_mtts module parameter %d.\n",
+ hca_profile.fmr_reserved_mtts);
+ printk(KERN_WARNING PFX "(Must be smaller than num_mtt %d)\n",
+ hca_profile.num_mtt);
+ hca_profile.fmr_reserved_mtts = hca_profile.num_mtt / 2;
+ printk(KERN_WARNING PFX "Corrected fmr_reserved_mtts to %d.\n",
+ hca_profile.fmr_reserved_mtts);
+ }
+}
+
static int __init mthca_init(void)
{
int ret;
- mutex_init(&mthca_device_mutex);
+ mthca_validate_profile();
+
ret = mthca_catas_init();
if (ret)
return ret;
diff --git a/drivers/infiniband/hw/mthca/mthca_memfree.c b/drivers/infiniband/hw/mthca/mthca_memfree.c
index 15cc2f6eb47..6b19645d946 100644
--- a/drivers/infiniband/hw/mthca/mthca_memfree.c
+++ b/drivers/infiniband/hw/mthca/mthca_memfree.c
@@ -232,7 +232,7 @@ void *mthca_table_find(struct mthca_icm_table *table, int obj)
list_for_each_entry(chunk, &icm->chunk_list, list) {
for (i = 0; i < chunk->npages; ++i) {
- if (chunk->mem[i].length >= offset) {
+ if (chunk->mem[i].length > offset) {
page = chunk->mem[i].page;
goto out;
}
diff --git a/drivers/infiniband/hw/mthca/mthca_provider.c b/drivers/infiniband/hw/mthca/mthca_provider.c
index 21422a3336a..7b96751695e 100644
--- a/drivers/infiniband/hw/mthca/mthca_provider.c
+++ b/drivers/infiniband/hw/mthca/mthca_provider.c
@@ -124,7 +124,7 @@ static int mthca_query_device(struct ib_device *ibdev,
props->max_map_per_fmr = 255;
else
props->max_map_per_fmr =
- (1 << (32 - long_log2(mdev->limits.num_mpts))) - 1;
+ (1 << (32 - ilog2(mdev->limits.num_mpts))) - 1;
err = 0;
out:
@@ -816,7 +816,7 @@ static int mthca_resize_cq(struct ib_cq *ibcq, int entries, struct ib_udata *uda
lkey = ucmd.lkey;
}
- ret = mthca_RESIZE_CQ(dev, cq->cqn, lkey, long_log2(entries), &status);
+ ret = mthca_RESIZE_CQ(dev, cq->cqn, lkey, ilog2(entries), &status);
if (status)
ret = -EINVAL;
@@ -1100,10 +1100,11 @@ static struct ib_fmr *mthca_alloc_fmr(struct ib_pd *pd, int mr_access_flags,
struct mthca_fmr *fmr;
int err;
- fmr = kmemdup(fmr_attr, sizeof *fmr, GFP_KERNEL);
+ fmr = kmalloc(sizeof *fmr, GFP_KERNEL);
if (!fmr)
return ERR_PTR(-ENOMEM);
+ memcpy(&fmr->attr, fmr_attr, sizeof *fmr_attr);
err = mthca_fmr_alloc(to_mdev(pd->device), to_mpd(pd)->pd_num,
convert_access(mr_access_flags), fmr);
diff --git a/drivers/infiniband/hw/mthca/mthca_qp.c b/drivers/infiniband/hw/mthca/mthca_qp.c
index 33e3ba7937f..5f5214c0337 100644
--- a/drivers/infiniband/hw/mthca/mthca_qp.c
+++ b/drivers/infiniband/hw/mthca/mthca_qp.c
@@ -429,13 +429,18 @@ int mthca_query_qp(struct ib_qp *ibqp, struct ib_qp_attr *qp_attr, int qp_attr_m
{
struct mthca_dev *dev = to_mdev(ibqp->device);
struct mthca_qp *qp = to_mqp(ibqp);
- int err;
- struct mthca_mailbox *mailbox;
+ int err = 0;
+ struct mthca_mailbox *mailbox = NULL;
struct mthca_qp_param *qp_param;
struct mthca_qp_context *context;
int mthca_state;
u8 status;
+ if (qp->state == IB_QPS_RESET) {
+ qp_attr->qp_state = IB_QPS_RESET;
+ goto done;
+ }
+
mailbox = mthca_alloc_mailbox(dev, GFP_KERNEL);
if (IS_ERR(mailbox))
return PTR_ERR(mailbox);
@@ -454,7 +459,6 @@ int mthca_query_qp(struct ib_qp *ibqp, struct ib_qp_attr *qp_attr, int qp_attr_m
mthca_state = be32_to_cpu(context->flags) >> 28;
qp_attr->qp_state = to_ib_qp_state(mthca_state);
- qp_attr->cur_qp_state = qp_attr->qp_state;
qp_attr->path_mtu = context->mtu_msgmax >> 5;
qp_attr->path_mig_state =
to_ib_mig_state((be32_to_cpu(context->flags) >> 11) & 0x3);
@@ -464,11 +468,6 @@ int mthca_query_qp(struct ib_qp *ibqp, struct ib_qp_attr *qp_attr, int qp_attr_m
qp_attr->dest_qp_num = be32_to_cpu(context->remote_qpn) & 0xffffff;
qp_attr->qp_access_flags =
to_ib_qp_access_flags(be32_to_cpu(context->params2));
- qp_attr->cap.max_send_wr = qp->sq.max;
- qp_attr->cap.max_recv_wr = qp->rq.max;
- qp_attr->cap.max_send_sge = qp->sq.max_gs;
- qp_attr->cap.max_recv_sge = qp->rq.max_gs;
- qp_attr->cap.max_inline_data = qp->max_inline_data;
if (qp->transport == RC || qp->transport == UC) {
to_ib_ah_attr(dev, &qp_attr->ah_attr, &context->pri_path);
@@ -495,7 +494,16 @@ int mthca_query_qp(struct ib_qp *ibqp, struct ib_qp_attr *qp_attr, int qp_attr_m
qp_attr->retry_cnt = (be32_to_cpu(context->params1) >> 16) & 0x7;
qp_attr->rnr_retry = context->pri_path.rnr_retry >> 5;
qp_attr->alt_timeout = context->alt_path.ackto >> 3;
- qp_init_attr->cap = qp_attr->cap;
+
+done:
+ qp_attr->cur_qp_state = qp_attr->qp_state;
+ qp_attr->cap.max_send_wr = qp->sq.max;
+ qp_attr->cap.max_recv_wr = qp->rq.max;
+ qp_attr->cap.max_send_sge = qp->sq.max_gs;
+ qp_attr->cap.max_recv_sge = qp->rq.max_gs;
+ qp_attr->cap.max_inline_data = qp->max_inline_data;
+
+ qp_init_attr->cap = qp_attr->cap;
out:
mthca_free_mailbox(dev, mailbox);
@@ -636,11 +644,11 @@ int mthca_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr, int attr_mask,
if (mthca_is_memfree(dev)) {
if (qp->rq.max)
- qp_context->rq_size_stride = long_log2(qp->rq.max) << 3;
+ qp_context->rq_size_stride = ilog2(qp->rq.max) << 3;
qp_context->rq_size_stride |= qp->rq.wqe_shift - 4;
if (qp->sq.max)
- qp_context->sq_size_stride = long_log2(qp->sq.max) << 3;
+ qp_context->sq_size_stride = ilog2(qp->sq.max) << 3;
qp_context->sq_size_stride |= qp->sq.wqe_shift - 4;
}
diff --git a/drivers/infiniband/hw/mthca/mthca_srq.c b/drivers/infiniband/hw/mthca/mthca_srq.c
index 34d2c476896..10684da33d5 100644
--- a/drivers/infiniband/hw/mthca/mthca_srq.c
+++ b/drivers/infiniband/hw/mthca/mthca_srq.c
@@ -120,7 +120,7 @@ static void mthca_arbel_init_srq_context(struct mthca_dev *dev,
memset(context, 0, sizeof *context);
- logsize = long_log2(srq->max);
+ logsize = ilog2(srq->max);
context->state_logsize_srqn = cpu_to_be32(logsize << 24 | srq->srqn);
context->lkey = cpu_to_be32(srq->mr.ibmr.lkey);
context->db_index = cpu_to_be32(srq->db_index);
@@ -213,7 +213,7 @@ int mthca_alloc_srq(struct mthca_dev *dev, struct mthca_pd *pd,
if (!mthca_is_memfree(dev) && (ds > dev->limits.max_desc_sz))
return -EINVAL;
- srq->wqe_shift = long_log2(ds);
+ srq->wqe_shift = ilog2(ds);
srq->srqn = mthca_alloc(&dev->srq_table.alloc);
if (srq->srqn == -1)
diff --git a/drivers/infiniband/ulp/ipoib/ipoib.h b/drivers/infiniband/ulp/ipoib/ipoib.h
index 99547996aba..07deee8f81c 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib.h
+++ b/drivers/infiniband/ulp/ipoib/ipoib.h
@@ -105,12 +105,12 @@ struct ipoib_mcast;
struct ipoib_rx_buf {
struct sk_buff *skb;
- dma_addr_t mapping;
+ u64 mapping;
};
struct ipoib_tx_buf {
struct sk_buff *skb;
- DECLARE_PCI_UNMAP_ADDR(mapping)
+ u64 mapping;
};
/*
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ib.c b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
index f10fba5d326..59d9594ed6d 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_ib.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
@@ -109,9 +109,8 @@ static int ipoib_ib_post_receive(struct net_device *dev, int id)
ret = ib_post_recv(priv->qp, &param, &bad_wr);
if (unlikely(ret)) {
ipoib_warn(priv, "receive failed for buf %d (%d)\n", id, ret);
- dma_unmap_single(priv->ca->dma_device,
- priv->rx_ring[id].mapping,
- IPOIB_BUF_SIZE, DMA_FROM_DEVICE);
+ ib_dma_unmap_single(priv->ca, priv->rx_ring[id].mapping,
+ IPOIB_BUF_SIZE, DMA_FROM_DEVICE);
dev_kfree_skb_any(priv->rx_ring[id].skb);
priv->rx_ring[id].skb = NULL;
}
@@ -123,7 +122,7 @@ static int ipoib_alloc_rx_skb(struct net_device *dev, int id)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
struct sk_buff *skb;
- dma_addr_t addr;
+ u64 addr;
skb = dev_alloc_skb(IPOIB_BUF_SIZE + 4);
if (!skb)
@@ -136,10 +135,9 @@ static int ipoib_alloc_rx_skb(struct net_device *dev, int id)
*/
skb_reserve(skb, 4);
- addr = dma_map_single(priv->ca->dma_device,
- skb->data, IPOIB_BUF_SIZE,
- DMA_FROM_DEVICE);
- if (unlikely(dma_mapping_error(addr))) {
+ addr = ib_dma_map_single(priv->ca, skb->data, IPOIB_BUF_SIZE,
+ DMA_FROM_DEVICE);
+ if (unlikely(ib_dma_mapping_error(priv->ca, addr))) {
dev_kfree_skb_any(skb);
return -EIO;
}
@@ -174,7 +172,7 @@ static void ipoib_ib_handle_rx_wc(struct net_device *dev, struct ib_wc *wc)
struct ipoib_dev_priv *priv = netdev_priv(dev);
unsigned int wr_id = wc->wr_id & ~IPOIB_OP_RECV;
struct sk_buff *skb;
- dma_addr_t addr;
+ u64 addr;
ipoib_dbg_data(priv, "recv completion: id %d, op %d, status: %d\n",
wr_id, wc->opcode, wc->status);
@@ -193,8 +191,8 @@ static void ipoib_ib_handle_rx_wc(struct net_device *dev, struct ib_wc *wc)
ipoib_warn(priv, "failed recv event "
"(status=%d, wrid=%d vend_err %x)\n",
wc->status, wr_id, wc->vendor_err);
- dma_unmap_single(priv->ca->dma_device, addr,
- IPOIB_BUF_SIZE, DMA_FROM_DEVICE);
+ ib_dma_unmap_single(priv->ca, addr,
+ IPOIB_BUF_SIZE, DMA_FROM_DEVICE);
dev_kfree_skb_any(skb);
priv->rx_ring[wr_id].skb = NULL;
return;
@@ -212,8 +210,7 @@ static void ipoib_ib_handle_rx_wc(struct net_device *dev, struct ib_wc *wc)
ipoib_dbg_data(priv, "received %d bytes, SLID 0x%04x\n",
wc->byte_len, wc->slid);
- dma_unmap_single(priv->ca->dma_device, addr,
- IPOIB_BUF_SIZE, DMA_FROM_DEVICE);
+ ib_dma_unmap_single(priv->ca, addr, IPOIB_BUF_SIZE, DMA_FROM_DEVICE);
skb_put(skb, wc->byte_len);
skb_pull(skb, IB_GRH_BYTES);
@@ -261,10 +258,8 @@ static void ipoib_ib_handle_tx_wc(struct net_device *dev, struct ib_wc *wc)
tx_req = &priv->tx_ring[wr_id];
- dma_unmap_single(priv->ca->dma_device,
- pci_unmap_addr(tx_req, mapping),
- tx_req->skb->len,
- DMA_TO_DEVICE);
+ ib_dma_unmap_single(priv->ca, tx_req->mapping,
+ tx_req->skb->len, DMA_TO_DEVICE);
++priv->stats.tx_packets;
priv->stats.tx_bytes += tx_req->skb->len;
@@ -311,7 +306,7 @@ void ipoib_ib_completion(struct ib_cq *cq, void *dev_ptr)
static inline int post_send(struct ipoib_dev_priv *priv,
unsigned int wr_id,
struct ib_ah *address, u32 qpn,
- dma_addr_t addr, int len)
+ u64 addr, int len)
{
struct ib_send_wr *bad_wr;
@@ -330,7 +325,7 @@ void ipoib_send(struct net_device *dev, struct sk_buff *skb,
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
struct ipoib_tx_buf *tx_req;
- dma_addr_t addr;
+ u64 addr;
if (unlikely(skb->len > dev->mtu + INFINIBAND_ALEN)) {
ipoib_warn(priv, "packet len %d (> %d) too long to send, dropping\n",
@@ -353,21 +348,20 @@ void ipoib_send(struct net_device *dev, struct sk_buff *skb,
*/
tx_req = &priv->tx_ring[priv->tx_head & (ipoib_sendq_size - 1)];
tx_req->skb = skb;
- addr = dma_map_single(priv->ca->dma_device, skb->data, skb->len,
- DMA_TO_DEVICE);
- if (unlikely(dma_mapping_error(addr))) {
+ addr = ib_dma_map_single(priv->ca, skb->data, skb->len,
+ DMA_TO_DEVICE);
+ if (unlikely(ib_dma_mapping_error(priv->ca, addr))) {
++priv->stats.tx_errors;
dev_kfree_skb_any(skb);
return;
}
- pci_unmap_addr_set(tx_req, mapping, addr);
+ tx_req->mapping = addr;
if (unlikely(post_send(priv, priv->tx_head & (ipoib_sendq_size - 1),
address->ah, qpn, addr, skb->len))) {
ipoib_warn(priv, "post_send failed\n");
++priv->stats.tx_errors;
- dma_unmap_single(priv->ca->dma_device, addr, skb->len,
- DMA_TO_DEVICE);
+ ib_dma_unmap_single(priv->ca, addr, skb->len, DMA_TO_DEVICE);
dev_kfree_skb_any(skb);
} else {
dev->trans_start = jiffies;
@@ -538,24 +532,27 @@ int ipoib_ib_dev_stop(struct net_device *dev)
while ((int) priv->tx_tail - (int) priv->tx_head < 0) {
tx_req = &priv->tx_ring[priv->tx_tail &
(ipoib_sendq_size - 1)];
- dma_unmap_single(priv->ca->dma_device,
- pci_unmap_addr(tx_req, mapping),
- tx_req->skb->len,
- DMA_TO_DEVICE);
+ ib_dma_unmap_single(priv->ca,
+ tx_req->mapping,
+ tx_req->skb->len,
+ DMA_TO_DEVICE);
dev_kfree_skb_any(tx_req->skb);
++priv->tx_tail;
}
- for (i = 0; i < ipoib_recvq_size; ++i)
- if (priv->rx_ring[i].skb) {
- dma_unmap_single(priv->ca->dma_device,
- pci_unmap_addr(&priv->rx_ring[i],
- mapping),
- IPOIB_BUF_SIZE,
- DMA_FROM_DEVICE);
- dev_kfree_skb_any(priv->rx_ring[i].skb);
- priv->rx_ring[i].skb = NULL;
- }
+ for (i = 0; i < ipoib_recvq_size; ++i) {
+ struct ipoib_rx_buf *rx_req;
+
+ rx_req = &priv->rx_ring[i];
+ if (!rx_req->skb)
+ continue;
+ ib_dma_unmap_single(priv->ca,
+ rx_req->mapping,
+ IPOIB_BUF_SIZE,
+ DMA_FROM_DEVICE);
+ dev_kfree_skb_any(rx_req->skb);
+ rx_req->skb = NULL;
+ }
goto timeout;
}
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c
index c0928024372..af5ee2ec449 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_main.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c
@@ -497,8 +497,6 @@ static void neigh_add_path(struct sk_buff *skb, struct net_device *dev)
return;
}
- skb_queue_head_init(&neigh->queue);
-
/*
* We can only be called from ipoib_start_xmit, so we're
* inside tx_lock -- no need to save/restore flags.
@@ -806,6 +804,7 @@ struct ipoib_neigh *ipoib_neigh_alloc(struct neighbour *neighbour)
neigh->neighbour = neighbour;
*to_ipoib_neigh(neighbour) = neigh;
+ skb_queue_head_init(&neigh->queue);
return neigh;
}
@@ -959,16 +958,17 @@ struct ipoib_dev_priv *ipoib_intf_alloc(const char *name)
return netdev_priv(dev);
}
-static ssize_t show_pkey(struct class_device *cdev, char *buf)
+static ssize_t show_pkey(struct device *dev,
+ struct device_attribute *attr, char *buf)
{
- struct ipoib_dev_priv *priv =
- netdev_priv(container_of(cdev, struct net_device, class_dev));
+ struct ipoib_dev_priv *priv = netdev_priv(to_net_dev(dev));
return sprintf(buf, "0x%04x\n", priv->pkey);
}
-static CLASS_DEVICE_ATTR(pkey, S_IRUGO, show_pkey, NULL);
+static DEVICE_ATTR(pkey, S_IRUGO, show_pkey, NULL);
-static ssize_t create_child(struct class_device *cdev,
+static ssize_t create_child(struct device *dev,
+ struct device_attribute *attr,
const char *buf, size_t count)
{
int pkey;
@@ -986,14 +986,14 @@ static ssize_t create_child(struct class_device *cdev,
*/
pkey |= 0x8000;
- ret = ipoib_vlan_add(container_of(cdev, struct net_device, class_dev),
- pkey);
+ ret = ipoib_vlan_add(to_net_dev(dev), pkey);
return ret ? ret : count;
}
-static CLASS_DEVICE_ATTR(create_child, S_IWUGO, NULL, create_child);
+static DEVICE_ATTR(create_child, S_IWUGO, NULL, create_child);
-static ssize_t delete_child(struct class_device *cdev,
+static ssize_t delete_child(struct device *dev,
+ struct device_attribute *attr,
const char *buf, size_t count)
{
int pkey;
@@ -1005,18 +1005,16 @@ static ssize_t delete_child(struct class_device *cdev,
if (pkey < 0 || pkey > 0xffff)
return -EINVAL;
- ret = ipoib_vlan_delete(container_of(cdev, struct net_device, class_dev),
- pkey);
+ ret = ipoib_vlan_delete(to_net_dev(dev), pkey);
return ret ? ret : count;
}
-static CLASS_DEVICE_ATTR(delete_child, S_IWUGO, NULL, delete_child);
+static DEVICE_ATTR(delete_child, S_IWUGO, NULL, delete_child);
int ipoib_add_pkey_attr(struct net_device *dev)
{
- return class_device_create_file(&dev->class_dev,
- &class_device_attr_pkey);
+ return device_create_file(&dev->dev, &dev_attr_pkey);
}
static struct net_device *ipoib_add_port(const char *format,
@@ -1084,11 +1082,9 @@ static struct net_device *ipoib_add_port(const char *format,
if (ipoib_add_pkey_attr(priv->dev))
goto sysfs_failed;
- if (class_device_create_file(&priv->dev->class_dev,
- &class_device_attr_create_child))
+ if (device_create_file(&priv->dev->dev, &dev_attr_create_child))
goto sysfs_failed;
- if (class_device_create_file(&priv->dev->class_dev,
- &class_device_attr_delete_child))
+ if (device_create_file(&priv->dev->dev, &dev_attr_delete_child))
goto sysfs_failed;
return priv->dev;
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_vlan.c b/drivers/infiniband/ulp/ipoib/ipoib_vlan.c
index f887780e809..085eafe6667 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_vlan.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_vlan.c
@@ -42,15 +42,15 @@
#include "ipoib.h"
-static ssize_t show_parent(struct class_device *class_dev, char *buf)
+static ssize_t show_parent(struct device *d, struct device_attribute *attr,
+ char *buf)
{
- struct net_device *dev =
- container_of(class_dev, struct net_device, class_dev);
+ 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 CLASS_DEVICE_ATTR(parent, S_IRUGO, show_parent, NULL);
+static DEVICE_ATTR(parent, S_IRUGO, show_parent, NULL);
int ipoib_vlan_add(struct net_device *pdev, unsigned short pkey)
{
@@ -118,8 +118,7 @@ int ipoib_vlan_add(struct net_device *pdev, unsigned short pkey)
if (ipoib_add_pkey_attr(priv->dev))
goto sysfs_failed;
- if (class_device_create_file(&priv->dev->class_dev,
- &class_device_attr_parent))
+ if (device_create_file(&priv->dev->dev, &dev_attr_parent))
goto sysfs_failed;
list_add_tail(&priv->list, &ppriv->child_intfs);
diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.c b/drivers/infiniband/ulp/iser/iscsi_iser.c
index 9b2041e25d5..dd221eda3ea 100644
--- a/drivers/infiniband/ulp/iser/iscsi_iser.c
+++ b/drivers/infiniband/ulp/iser/iscsi_iser.c
@@ -177,7 +177,7 @@ iscsi_iser_mtask_xmit(struct iscsi_conn *conn,
* - if yes, the mtask is recycled at iscsi_complete_pdu
* - if no, the mtask is recycled at iser_snd_completion
*/
- if (error && error != -EAGAIN)
+ if (error && error != -ENOBUFS)
iscsi_conn_failure(conn, ISCSI_ERR_CONN_FAILED);
return error;
@@ -241,7 +241,7 @@ iscsi_iser_ctask_xmit(struct iscsi_conn *conn,
error = iscsi_iser_ctask_xmit_unsol_data(conn, ctask);
iscsi_iser_ctask_xmit_exit:
- if (error && error != -EAGAIN)
+ if (error && error != -ENOBUFS)
iscsi_conn_failure(conn, ISCSI_ERR_CONN_FAILED);
return error;
}
diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.h b/drivers/infiniband/ulp/iser/iscsi_iser.h
index 234e5b061a7..cae8c96a55f 100644
--- a/drivers/infiniband/ulp/iser/iscsi_iser.h
+++ b/drivers/infiniband/ulp/iser/iscsi_iser.h
@@ -182,7 +182,7 @@ struct iser_regd_buf {
struct iser_mem_reg reg; /* memory registration info */
void *virt_addr;
struct iser_device *device; /* device->device for dma_unmap */
- dma_addr_t dma_addr; /* if non zero, addr for dma_unmap */
+ u64 dma_addr; /* if non zero, addr for dma_unmap */
enum dma_data_direction direction; /* direction for dma_unmap */
unsigned int data_size;
atomic_t ref_count; /* refcount, freed when dec to 0 */
diff --git a/drivers/infiniband/ulp/iser/iser_initiator.c b/drivers/infiniband/ulp/iser/iser_initiator.c
index 9b3d79c796c..89e37283c83 100644
--- a/drivers/infiniband/ulp/iser/iser_initiator.c
+++ b/drivers/infiniband/ulp/iser/iser_initiator.c
@@ -304,18 +304,14 @@ int iser_conn_set_full_featured_mode(struct iscsi_conn *conn)
static int
iser_check_xmit(struct iscsi_conn *conn, void *task)
{
- int rc = 0;
struct iscsi_iser_conn *iser_conn = conn->dd_data;
- write_lock_bh(conn->recv_lock);
if (atomic_read(&iser_conn->ib_conn->post_send_buf_count) ==
ISER_QP_MAX_REQ_DTOS) {
- iser_dbg("%ld can't xmit task %p, suspending tx\n",jiffies,task);
- set_bit(ISCSI_SUSPEND_BIT, &conn->suspend_tx);
- rc = -EAGAIN;
+ iser_dbg("%ld can't xmit task %p\n",jiffies,task);
+ return -ENOBUFS;
}
- write_unlock_bh(conn->recv_lock);
- return rc;
+ return 0;
}
@@ -340,7 +336,7 @@ int iser_send_command(struct iscsi_conn *conn,
return -EPERM;
}
if (iser_check_xmit(conn, ctask))
- return -EAGAIN;
+ return -ENOBUFS;
edtl = ntohl(hdr->data_length);
@@ -426,7 +422,7 @@ int iser_send_data_out(struct iscsi_conn *conn,
}
if (iser_check_xmit(conn, ctask))
- return -EAGAIN;
+ return -ENOBUFS;
itt = ntohl(hdr->itt);
data_seg_len = ntoh24(hdr->dlength);
@@ -487,10 +483,8 @@ int iser_send_control(struct iscsi_conn *conn,
struct iscsi_iser_conn *iser_conn = conn->dd_data;
struct iser_desc *mdesc = mtask->dd_data;
struct iser_dto *send_dto = NULL;
- unsigned int itt;
unsigned long data_seg_len;
int err = 0;
- unsigned char opcode;
struct iser_regd_buf *regd_buf;
struct iser_device *device;
@@ -500,7 +494,7 @@ int iser_send_control(struct iscsi_conn *conn,
}
if (iser_check_xmit(conn,mtask))
- return -EAGAIN;
+ return -ENOBUFS;
/* build the tx desc regd header and add it to the tx desc dto */
mdesc->type = ISCSI_TX_CONTROL;
@@ -512,8 +506,6 @@ int iser_send_control(struct iscsi_conn *conn,
iser_reg_single(device, send_dto->regd[0], DMA_TO_DEVICE);
- itt = ntohl(mtask->hdr->itt);
- opcode = mtask->hdr->opcode & ISCSI_OPCODE_MASK;
data_seg_len = ntoh24(mtask->hdr->dlength);
if (data_seg_len > 0) {
@@ -575,7 +567,7 @@ void iser_rcv_completion(struct iser_desc *rx_desc,
opcode = hdr->opcode & ISCSI_OPCODE_MASK;
if (opcode == ISCSI_OP_SCSI_CMD_RSP) {
- itt = hdr->itt & ISCSI_ITT_MASK; /* mask out cid and age bits */
+ itt = get_itt(hdr->itt); /* mask out cid and age bits */
if (!(itt < session->cmds_max))
iser_err("itt can't be matched to task!!!"
"conn %p opcode %d cmds_max %d itt %d\n",
@@ -609,6 +601,7 @@ void iser_snd_completion(struct iser_desc *tx_desc)
struct iscsi_iser_conn *iser_conn = ib_conn->iser_conn;
struct iscsi_conn *conn = iser_conn->iscsi_conn;
struct iscsi_mgmt_task *mtask;
+ int resume_tx = 0;
iser_dbg("Initiator, Data sent dto=0x%p\n", dto);
@@ -617,21 +610,22 @@ void iser_snd_completion(struct iser_desc *tx_desc)
if (tx_desc->type == ISCSI_TX_DATAOUT)
kmem_cache_free(ig.desc_cache, tx_desc);
+ if (atomic_read(&iser_conn->ib_conn->post_send_buf_count) ==
+ ISER_QP_MAX_REQ_DTOS)
+ resume_tx = 1;
+
atomic_dec(&ib_conn->post_send_buf_count);
- write_lock(conn->recv_lock);
- if (conn->suspend_tx) {
+ if (resume_tx) {
iser_dbg("%ld resuming tx\n",jiffies);
- clear_bit(ISCSI_SUSPEND_BIT, &conn->suspend_tx);
scsi_queue_work(conn->session->host, &conn->xmitwork);
}
- write_unlock(conn->recv_lock);
if (tx_desc->type == ISCSI_TX_CONTROL) {
/* this arithmetic is legal by libiscsi dd_data allocation */
mtask = (void *) ((long)(void *)tx_desc -
sizeof(struct iscsi_mgmt_task));
- if (mtask->hdr->itt == cpu_to_be32(ISCSI_RESERVED_TAG)) {
+ if (mtask->hdr->itt == RESERVED_ITT) {
struct iscsi_session *session = conn->session;
spin_lock(&conn->session->lock);
diff --git a/drivers/infiniband/ulp/iser/iser_memory.c b/drivers/infiniband/ulp/iser/iser_memory.c
index 5e122501fd8..fc9f1fd0ae5 100644
--- a/drivers/infiniband/ulp/iser/iser_memory.c
+++ b/drivers/infiniband/ulp/iser/iser_memory.c
@@ -52,7 +52,7 @@
*/
int iser_regd_buff_release(struct iser_regd_buf *regd_buf)
{
- struct device *dma_device;
+ struct ib_device *dev;
if ((atomic_read(&regd_buf->ref_count) == 0) ||
atomic_dec_and_test(&regd_buf->ref_count)) {
@@ -61,8 +61,8 @@ int iser_regd_buff_release(struct iser_regd_buf *regd_buf)
iser_unreg_mem(&regd_buf->reg);
if (regd_buf->dma_addr) {
- dma_device = regd_buf->device->ib_device->dma_device;
- dma_unmap_single(dma_device,
+ dev = regd_buf->device->ib_device;
+ ib_dma_unmap_single(dev,
regd_buf->dma_addr,
regd_buf->data_size,
regd_buf->direction);
@@ -84,12 +84,12 @@ void iser_reg_single(struct iser_device *device,
struct iser_regd_buf *regd_buf,
enum dma_data_direction direction)
{
- dma_addr_t dma_addr;
+ u64 dma_addr;
- dma_addr = dma_map_single(device->ib_device->dma_device,
- regd_buf->virt_addr,
- regd_buf->data_size, direction);
- BUG_ON(dma_mapping_error(dma_addr));
+ dma_addr = ib_dma_map_single(device->ib_device,
+ regd_buf->virt_addr,
+ regd_buf->data_size, direction);
+ BUG_ON(ib_dma_mapping_error(device->ib_device, dma_addr));
regd_buf->reg.lkey = device->mr->lkey;
regd_buf->reg.len = regd_buf->data_size;
@@ -107,14 +107,14 @@ int iser_start_rdma_unaligned_sg(struct iscsi_iser_cmd_task *iser_ctask,
enum iser_data_dir cmd_dir)
{
int dma_nents;
- struct device *dma_device;
+ struct ib_device *dev;
char *mem = NULL;
struct iser_data_buf *data = &iser_ctask->data[cmd_dir];
unsigned long cmd_data_len = data->data_len;
if (cmd_data_len > ISER_KMALLOC_THRESHOLD)
mem = (void *)__get_free_pages(GFP_NOIO,
- long_log2(roundup_pow_of_two(cmd_data_len)) - PAGE_SHIFT);
+ ilog2(roundup_pow_of_two(cmd_data_len)) - PAGE_SHIFT);
else
mem = kmalloc(cmd_data_len, GFP_NOIO);
@@ -147,17 +147,12 @@ int iser_start_rdma_unaligned_sg(struct iscsi_iser_cmd_task *iser_ctask,
iser_ctask->data_copy[cmd_dir].copy_buf = mem;
- dma_device = iser_ctask->iser_conn->ib_conn->device->ib_device->dma_device;
-
- if (cmd_dir == ISER_DIR_OUT)
- dma_nents = dma_map_sg(dma_device,
- &iser_ctask->data_copy[cmd_dir].sg_single,
- 1, DMA_TO_DEVICE);
- else
- dma_nents = dma_map_sg(dma_device,
- &iser_ctask->data_copy[cmd_dir].sg_single,
- 1, DMA_FROM_DEVICE);
-
+ dev = iser_ctask->iser_conn->ib_conn->device->ib_device;
+ dma_nents = ib_dma_map_sg(dev,
+ &iser_ctask->data_copy[cmd_dir].sg_single,
+ 1,
+ (cmd_dir == ISER_DIR_OUT) ?
+ DMA_TO_DEVICE : DMA_FROM_DEVICE);
BUG_ON(dma_nents == 0);
iser_ctask->data_copy[cmd_dir].dma_nents = dma_nents;
@@ -170,19 +165,16 @@ int iser_start_rdma_unaligned_sg(struct iscsi_iser_cmd_task *iser_ctask,
void iser_finalize_rdma_unaligned_sg(struct iscsi_iser_cmd_task *iser_ctask,
enum iser_data_dir cmd_dir)
{
- struct device *dma_device;
+ struct ib_device *dev;
struct iser_data_buf *mem_copy;
unsigned long cmd_data_len;
- dma_device = iser_ctask->iser_conn->ib_conn->device->ib_device->dma_device;
- mem_copy = &iser_ctask->data_copy[cmd_dir];
+ dev = iser_ctask->iser_conn->ib_conn->device->ib_device;
+ mem_copy = &iser_ctask->data_copy[cmd_dir];
- if (cmd_dir == ISER_DIR_OUT)
- dma_unmap_sg(dma_device, &mem_copy->sg_single, 1,
- DMA_TO_DEVICE);
- else
- dma_unmap_sg(dma_device, &mem_copy->sg_single, 1,
- DMA_FROM_DEVICE);
+ 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;
@@ -211,7 +203,7 @@ void iser_finalize_rdma_unaligned_sg(struct iscsi_iser_cmd_task *iser_ctask,
if (cmd_data_len > ISER_KMALLOC_THRESHOLD)
free_pages((unsigned long)mem_copy->copy_buf,
- long_log2(roundup_pow_of_two(cmd_data_len)) - PAGE_SHIFT);
+ ilog2(roundup_pow_of_two(cmd_data_len)) - PAGE_SHIFT);
else
kfree(mem_copy->copy_buf);
@@ -231,11 +223,12 @@ void iser_finalize_rdma_unaligned_sg(struct iscsi_iser_cmd_task *iser_ctask,
* 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 iser_page_vec *page_vec,
+ struct ib_device *ibdev)
{
struct scatterlist *sg = (struct scatterlist *)data->buf;
- dma_addr_t first_addr, last_addr, page;
- int start_aligned, end_aligned;
+ u64 first_addr, last_addr, page;
+ int end_aligned;
unsigned int cur_page = 0;
unsigned long total_sz = 0;
int i;
@@ -244,19 +237,21 @@ static int iser_sg_to_page_vec(struct iser_data_buf *data,
page_vec->offset = (u64) sg[0].offset & ~MASK_4K;
for (i = 0; i < data->dma_nents; i++) {
- total_sz += sg_dma_len(&sg[i]);
+ unsigned int dma_len = ib_sg_dma_len(ibdev, &sg[i]);
+
+ total_sz += dma_len;
- first_addr = sg_dma_address(&sg[i]);
- last_addr = first_addr + sg_dma_len(&sg[i]);
+ first_addr = ib_sg_dma_address(ibdev, &sg[i]);
+ last_addr = first_addr + dma_len;
- start_aligned = !(first_addr & ~MASK_4K);
end_aligned = !(last_addr & ~MASK_4K);
/* continue to collect page fragments till aligned or SG ends */
while (!end_aligned && (i + 1 < data->dma_nents)) {
i++;
- total_sz += sg_dma_len(&sg[i]);
- last_addr = sg_dma_address(&sg[i]) + sg_dma_len(&sg[i]);
+ dma_len = ib_sg_dma_len(ibdev, &sg[i]);
+ total_sz += dma_len;
+ last_addr = ib_sg_dma_address(ibdev, &sg[i]) + dma_len;
end_aligned = !(last_addr & ~MASK_4K);
}
@@ -288,10 +283,11 @@ static int iser_sg_to_page_vec(struct iser_data_buf *data,
* the number of entries which are aligned correctly. Supports the case where
* consecutive SG elements are actually fragments of the same physcial page.
*/
-static unsigned int iser_data_buf_aligned_len(struct iser_data_buf *data)
+static unsigned int iser_data_buf_aligned_len(struct iser_data_buf *data,
+ struct ib_device *ibdev)
{
struct scatterlist *sg;
- dma_addr_t end_addr, next_addr;
+ u64 end_addr, next_addr;
int i, cnt;
unsigned int ret_len = 0;
@@ -303,12 +299,12 @@ static unsigned int iser_data_buf_aligned_len(struct iser_data_buf *data)
(unsigned long)page_to_phys(sg[i].page),
(unsigned long)sg[i].offset,
(unsigned long)sg[i].length); */
- end_addr = sg_dma_address(&sg[i]) +
- sg_dma_len(&sg[i]);
+ end_addr = ib_sg_dma_address(ibdev, &sg[i]) +
+ ib_sg_dma_len(ibdev, &sg[i]);
/* iser_dbg("Checking sg iobuf end address "
"0x%08lX\n", end_addr); */
if (i + 1 < data->dma_nents) {
- next_addr = sg_dma_address(&sg[i+1]);
+ next_addr = ib_sg_dma_address(ibdev, &sg[i+1]);
/* are i, i+1 fragments of the same page? */
if (end_addr == next_addr)
continue;
@@ -325,7 +321,8 @@ static unsigned int iser_data_buf_aligned_len(struct iser_data_buf *data)
return ret_len;
}
-static void iser_data_buf_dump(struct iser_data_buf *data)
+static void iser_data_buf_dump(struct iser_data_buf *data,
+ struct ib_device *ibdev)
{
struct scatterlist *sg = (struct scatterlist *)data->buf;
int i;
@@ -333,9 +330,9 @@ static void iser_data_buf_dump(struct iser_data_buf *data)
for (i = 0; i < data->dma_nents; i++)
iser_err("sg[%d] dma_addr:0x%lX page:0x%p "
"off:0x%x sz:0x%x dma_len:0x%x\n",
- i, (unsigned long)sg_dma_address(&sg[i]),
+ i, (unsigned long)ib_sg_dma_address(ibdev, &sg[i]),
sg[i].page, sg[i].offset,
- sg[i].length,sg_dma_len(&sg[i]));
+ sg[i].length, ib_sg_dma_len(ibdev, &sg[i]));
}
static void iser_dump_page_vec(struct iser_page_vec *page_vec)
@@ -349,7 +346,8 @@ static void iser_dump_page_vec(struct iser_page_vec *page_vec)
}
static void iser_page_vec_build(struct iser_data_buf *data,
- struct iser_page_vec *page_vec)
+ struct iser_page_vec *page_vec,
+ struct ib_device *ibdev)
{
int page_vec_len = 0;
@@ -357,14 +355,14 @@ static void iser_page_vec_build(struct iser_data_buf *data,
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);
+ 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);
+ iser_data_buf_dump(data, ibdev);
iser_dump_page_vec(page_vec);
BUG();
}
@@ -375,13 +373,12 @@ int iser_dma_map_task_data(struct iscsi_iser_cmd_task *iser_ctask,
enum iser_data_dir iser_dir,
enum dma_data_direction dma_dir)
{
- struct device *dma_device;
+ struct ib_device *dev;
iser_ctask->dir[iser_dir] = 1;
- dma_device =
- iser_ctask->iser_conn->ib_conn->device->ib_device->dma_device;
+ dev = iser_ctask->iser_conn->ib_conn->device->ib_device;
- data->dma_nents = dma_map_sg(dma_device, data->buf, data->size, dma_dir);
+ 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;
@@ -391,20 +388,19 @@ int iser_dma_map_task_data(struct iscsi_iser_cmd_task *iser_ctask,
void iser_dma_unmap_task_data(struct iscsi_iser_cmd_task *iser_ctask)
{
- struct device *dma_device;
+ struct ib_device *dev;
struct iser_data_buf *data;
- dma_device =
- iser_ctask->iser_conn->ib_conn->device->ib_device->dma_device;
+ dev = iser_ctask->iser_conn->ib_conn->device->ib_device;
if (iser_ctask->dir[ISER_DIR_IN]) {
data = &iser_ctask->data[ISER_DIR_IN];
- dma_unmap_sg(dma_device, data->buf, data->size, DMA_FROM_DEVICE);
+ ib_dma_unmap_sg(dev, data->buf, data->size, DMA_FROM_DEVICE);
}
if (iser_ctask->dir[ISER_DIR_OUT]) {
data = &iser_ctask->data[ISER_DIR_OUT];
- dma_unmap_sg(dma_device, data->buf, data->size, DMA_TO_DEVICE);
+ ib_dma_unmap_sg(dev, data->buf, data->size, DMA_TO_DEVICE);
}
}
@@ -419,6 +415,7 @@ int iser_reg_rdma_mem(struct iscsi_iser_cmd_task *iser_ctask,
{
struct iser_conn *ib_conn = iser_ctask->iser_conn->ib_conn;
struct iser_device *device = ib_conn->device;
+ struct ib_device *ibdev = device->ib_device;
struct iser_data_buf *mem = &iser_ctask->data[cmd_dir];
struct iser_regd_buf *regd_buf;
int aligned_len;
@@ -428,11 +425,11 @@ int iser_reg_rdma_mem(struct iscsi_iser_cmd_task *iser_ctask,
regd_buf = &iser_ctask->rdma_regd[cmd_dir];
- aligned_len = iser_data_buf_aligned_len(mem);
+ aligned_len = iser_data_buf_aligned_len(mem, ibdev);
if (aligned_len != mem->dma_nents) {
iser_err("rdma alignment violation %d/%d aligned\n",
aligned_len, mem->size);
- iser_data_buf_dump(mem);
+ iser_data_buf_dump(mem, ibdev);
/* unmap the command data before accessing it */
iser_dma_unmap_task_data(iser_ctask);
@@ -450,8 +447,8 @@ int iser_reg_rdma_mem(struct iscsi_iser_cmd_task *iser_ctask,
regd_buf->reg.lkey = device->mr->lkey;
regd_buf->reg.rkey = device->mr->rkey;
- regd_buf->reg.len = sg_dma_len(&sg[0]);
- regd_buf->reg.va = sg_dma_address(&sg[0]);
+ 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 "
@@ -461,10 +458,10 @@ int iser_reg_rdma_mem(struct iscsi_iser_cmd_task *iser_ctask,
(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);
+ 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) {
- iser_data_buf_dump(mem);
+ iser_data_buf_dump(mem, ibdev);
iser_err("mem->dma_nents = %d (dlength = 0x%x)\n", mem->dma_nents,
ntoh24(iser_ctask->desc.iscsi_header.dlength));
iser_err("page_vec: data_size = 0x%x, length = %d, offset = 0x%x\n",
diff --git a/drivers/infiniband/ulp/srp/ib_srp.c b/drivers/infiniband/ulp/srp/ib_srp.c
index a6289595557..5e8ac577f0a 100644
--- a/drivers/infiniband/ulp/srp/ib_srp.c
+++ b/drivers/infiniband/ulp/srp/ib_srp.c
@@ -122,9 +122,8 @@ static struct srp_iu *srp_alloc_iu(struct srp_host *host, size_t size,
if (!iu->buf)
goto out_free_iu;
- iu->dma = dma_map_single(host->dev->dev->dma_device,
- iu->buf, size, direction);
- if (dma_mapping_error(iu->dma))
+ iu->dma = ib_dma_map_single(host->dev->dev, iu->buf, size, direction);
+ if (ib_dma_mapping_error(host->dev->dev, iu->dma))
goto out_free_buf;
iu->size = size;
@@ -145,8 +144,7 @@ static void srp_free_iu(struct srp_host *host, struct srp_iu *iu)
if (!iu)
return;
- dma_unmap_single(host->dev->dev->dma_device,
- iu->dma, iu->size, iu->direction);
+ ib_dma_unmap_single(host->dev->dev, iu->dma, iu->size, iu->direction);
kfree(iu->buf);
kfree(iu);
}
@@ -482,8 +480,8 @@ static void srp_unmap_data(struct scsi_cmnd *scmnd,
scat = &req->fake_sg;
}
- dma_unmap_sg(target->srp_host->dev->dev->dma_device, scat, nents,
- scmnd->sc_data_direction);
+ ib_dma_unmap_sg(target->srp_host->dev->dev, scat, nents,
+ scmnd->sc_data_direction);
}
static void srp_remove_req(struct srp_target_port *target, struct srp_request *req)
@@ -550,6 +548,7 @@ static int srp_reconnect_target(struct srp_target_port *target)
target->tx_head = 0;
target->tx_tail = 0;
+ target->qp_in_error = 0;
ret = srp_connect_target(target);
if (ret)
goto err;
@@ -595,23 +594,26 @@ static int srp_map_fmr(struct srp_target_port *target, struct scatterlist *scat,
int i, j;
int ret;
struct srp_device *dev = target->srp_host->dev;
+ struct ib_device *ibdev = dev->dev;
if (!dev->fmr_pool)
return -ENODEV;
- if ((sg_dma_address(&scat[0]) & ~dev->fmr_page_mask) &&
+ if ((ib_sg_dma_address(ibdev, &scat[0]) & ~dev->fmr_page_mask) &&
mellanox_workarounds && !memcmp(&target->ioc_guid, mellanox_oui, 3))
return -EINVAL;
len = page_cnt = 0;
for (i = 0; i < sg_cnt; ++i) {
- if (sg_dma_address(&scat[i]) & ~dev->fmr_page_mask) {
+ unsigned int dma_len = ib_sg_dma_len(ibdev, &scat[i]);
+
+ if (ib_sg_dma_address(ibdev, &scat[i]) & ~dev->fmr_page_mask) {
if (i > 0)
return -EINVAL;
else
++page_cnt;
}
- if ((sg_dma_address(&scat[i]) + sg_dma_len(&scat[i])) &
+ if ((ib_sg_dma_address(ibdev, &scat[i]) + dma_len) &
~dev->fmr_page_mask) {
if (i < sg_cnt - 1)
return -EINVAL;
@@ -619,7 +621,7 @@ static int srp_map_fmr(struct srp_target_port *target, struct scatterlist *scat,
++page_cnt;
}
- len += sg_dma_len(&scat[i]);
+ len += dma_len;
}
page_cnt += len >> dev->fmr_page_shift;
@@ -631,10 +633,14 @@ static int srp_map_fmr(struct srp_target_port *target, struct scatterlist *scat,
return -ENOMEM;
page_cnt = 0;
- for (i = 0; i < sg_cnt; ++i)
- for (j = 0; j < sg_dma_len(&scat[i]); j += dev->fmr_page_size)
+ for (i = 0; i < sg_cnt; ++i) {
+ unsigned int dma_len = ib_sg_dma_len(ibdev, &scat[i]);
+
+ for (j = 0; j < dma_len; j += dev->fmr_page_size)
dma_pages[page_cnt++] =
- (sg_dma_address(&scat[i]) & dev->fmr_page_mask) + j;
+ (ib_sg_dma_address(ibdev, &scat[i]) &
+ dev->fmr_page_mask) + j;
+ }
req->fmr = ib_fmr_pool_map_phys(dev->fmr_pool,
dma_pages, page_cnt, io_addr);
@@ -644,7 +650,8 @@ static int srp_map_fmr(struct srp_target_port *target, struct scatterlist *scat,
goto out;
}
- buf->va = cpu_to_be64(sg_dma_address(&scat[0]) & ~dev->fmr_page_mask);
+ buf->va = cpu_to_be64(ib_sg_dma_address(ibdev, &scat[0]) &
+ ~dev->fmr_page_mask);
buf->key = cpu_to_be32(req->fmr->fmr->rkey);
buf->len = cpu_to_be32(len);
@@ -663,6 +670,8 @@ static int srp_map_data(struct scsi_cmnd *scmnd, struct srp_target_port *target,
struct srp_cmd *cmd = req->cmd->buf;
int len, nents, count;
u8 fmt = SRP_DATA_DESC_DIRECT;
+ struct srp_device *dev;
+ struct ib_device *ibdev;
if (!scmnd->request_buffer || scmnd->sc_data_direction == DMA_NONE)
return sizeof (struct srp_cmd);
@@ -687,8 +696,10 @@ static int srp_map_data(struct scsi_cmnd *scmnd, struct srp_target_port *target,
sg_init_one(scat, scmnd->request_buffer, scmnd->request_bufflen);
}
- count = dma_map_sg(target->srp_host->dev->dev->dma_device,
- scat, nents, scmnd->sc_data_direction);
+ dev = target->srp_host->dev;
+ ibdev = dev->dev;
+
+ count = ib_dma_map_sg(ibdev, scat, nents, scmnd->sc_data_direction);
fmt = SRP_DATA_DESC_DIRECT;
len = sizeof (struct srp_cmd) + sizeof (struct srp_direct_buf);
@@ -702,9 +713,9 @@ static int srp_map_data(struct scsi_cmnd *scmnd, struct srp_target_port *target,
*/
struct srp_direct_buf *buf = (void *) cmd->add_data;
- buf->va = cpu_to_be64(sg_dma_address(scat));
- buf->key = cpu_to_be32(target->srp_host->dev->mr->rkey);
- buf->len = cpu_to_be32(sg_dma_len(scat));
+ buf->va = cpu_to_be64(ib_sg_dma_address(ibdev, scat));
+ buf->key = cpu_to_be32(dev->mr->rkey);
+ buf->len = cpu_to_be32(ib_sg_dma_len(ibdev, scat));
} else if (srp_map_fmr(target, scat, count, req,
(void *) cmd->add_data)) {
/*
@@ -722,13 +733,14 @@ static int srp_map_data(struct scsi_cmnd *scmnd, struct srp_target_port *target,
count * sizeof (struct srp_direct_buf);
for (i = 0; i < count; ++i) {
+ unsigned int dma_len = ib_sg_dma_len(ibdev, &scat[i]);
+
buf->desc_list[i].va =
- cpu_to_be64(sg_dma_address(&scat[i]));
+ cpu_to_be64(ib_sg_dma_address(ibdev, &scat[i]));
buf->desc_list[i].key =
- cpu_to_be32(target->srp_host->dev->mr->rkey);
- buf->desc_list[i].len =
- cpu_to_be32(sg_dma_len(&scat[i]));
- datalen += sg_dma_len(&scat[i]);
+ cpu_to_be32(dev->mr->rkey);
+ buf->desc_list[i].len = cpu_to_be32(dma_len);
+ datalen += dma_len;
}
if (scmnd->sc_data_direction == DMA_TO_DEVICE)
@@ -808,13 +820,15 @@ static void srp_process_rsp(struct srp_target_port *target, struct srp_rsp *rsp)
static void srp_handle_recv(struct srp_target_port *target, struct ib_wc *wc)
{
+ struct ib_device *dev;
struct srp_iu *iu;
u8 opcode;
iu = target->rx_ring[wc->wr_id & ~SRP_OP_RECV];
- dma_sync_single_for_cpu(target->srp_host->dev->dev->dma_device, iu->dma,
- target->max_ti_iu_len, DMA_FROM_DEVICE);
+ dev = target->srp_host->dev->dev;
+ ib_dma_sync_single_for_cpu(dev, iu->dma, target->max_ti_iu_len,
+ DMA_FROM_DEVICE);
opcode = *(u8 *) iu->buf;
@@ -850,8 +864,8 @@ static void srp_handle_recv(struct srp_target_port *target, struct ib_wc *wc)
break;
}
- dma_sync_single_for_device(target->srp_host->dev->dev->dma_device, iu->dma,
- target->max_ti_iu_len, DMA_FROM_DEVICE);
+ ib_dma_sync_single_for_device(dev, iu->dma, target->max_ti_iu_len,
+ DMA_FROM_DEVICE);
}
static void srp_completion(struct ib_cq *cq, void *target_ptr)
@@ -865,6 +879,7 @@ static void srp_completion(struct ib_cq *cq, void *target_ptr)
printk(KERN_ERR PFX "failed %s status %d\n",
wc.wr_id & SRP_OP_RECV ? "receive" : "send",
wc.status);
+ target->qp_in_error = 1;
break;
}
@@ -969,6 +984,7 @@ static int srp_queuecommand(struct scsi_cmnd *scmnd,
struct srp_request *req;
struct srp_iu *iu;
struct srp_cmd *cmd;
+ struct ib_device *dev;
int len;
if (target->state == SRP_TARGET_CONNECTING)
@@ -985,8 +1001,9 @@ static int srp_queuecommand(struct scsi_cmnd *scmnd,
if (!iu)
goto err;
- dma_sync_single_for_cpu(target->srp_host->dev->dev->dma_device, iu->dma,
- srp_max_iu_len, DMA_TO_DEVICE);
+ dev = target->srp_host->dev->dev;
+ ib_dma_sync_single_for_cpu(dev, iu->dma, srp_max_iu_len,
+ DMA_TO_DEVICE);
req = list_entry(target->free_reqs.next, struct srp_request, list);
@@ -1018,8 +1035,8 @@ static int srp_queuecommand(struct scsi_cmnd *scmnd,
goto err_unmap;
}
- dma_sync_single_for_device(target->srp_host->dev->dev->dma_device, iu->dma,
- srp_max_iu_len, DMA_TO_DEVICE);
+ ib_dma_sync_single_for_device(dev, iu->dma, srp_max_iu_len,
+ DMA_TO_DEVICE);
if (__srp_post_send(target, iu, len)) {
printk(KERN_ERR PFX "Send failed\n");
@@ -1322,6 +1339,8 @@ static int srp_abort(struct scsi_cmnd *scmnd)
printk(KERN_ERR "SRP abort called\n");
+ if (target->qp_in_error)
+ return FAILED;
if (srp_find_req(target, scmnd, &req))
return FAILED;
if (srp_send_tsk_mgmt(target, req, SRP_TSK_ABORT_TASK))
@@ -1350,6 +1369,8 @@ static int srp_reset_device(struct scsi_cmnd *scmnd)
printk(KERN_ERR "SRP reset_device called\n");
+ if (target->qp_in_error)
+ return FAILED;
if (srp_find_req(target, scmnd, &req))
return FAILED;
if (srp_send_tsk_mgmt(target, req, SRP_TSK_LUN_RESET))
@@ -1606,18 +1627,30 @@ static int srp_parse_options(const char *buf, struct srp_target_port *target)
switch (token) {
case SRP_OPT_ID_EXT:
p = match_strdup(args);
+ if (!p) {
+ ret = -ENOMEM;
+ goto out;
+ }
target->id_ext = cpu_to_be64(simple_strtoull(p, NULL, 16));
kfree(p);
break;
case SRP_OPT_IOC_GUID:
p = match_strdup(args);
+ if (!p) {
+ ret = -ENOMEM;
+ goto out;
+ }
target->ioc_guid = cpu_to_be64(simple_strtoull(p, NULL, 16));
kfree(p);
break;
case SRP_OPT_DGID:
p = match_strdup(args);
+ if (!p) {
+ ret = -ENOMEM;
+ goto out;
+ }
if (strlen(p) != 32) {
printk(KERN_WARNING PFX "bad dest GID parameter '%s'\n", p);
kfree(p);
@@ -1641,6 +1674,10 @@ static int srp_parse_options(const char *buf, struct srp_target_port *target)
case SRP_OPT_SERVICE_ID:
p = match_strdup(args);
+ if (!p) {
+ ret = -ENOMEM;
+ goto out;
+ }
target->service_id = cpu_to_be64(simple_strtoull(p, NULL, 16));
kfree(p);
break;
@@ -1678,6 +1715,10 @@ static int srp_parse_options(const char *buf, struct srp_target_port *target)
case SRP_OPT_INITIATOR_EXT:
p = match_strdup(args);
+ if (!p) {
+ ret = -ENOMEM;
+ goto out;
+ }
target->initiator_ext = cpu_to_be64(simple_strtoull(p, NULL, 16));
kfree(p);
break;
@@ -1766,6 +1807,7 @@ static ssize_t srp_create_target(struct class_device *class_dev,
goto err_free;
}
+ target->qp_in_error = 0;
ret = srp_connect_target(target);
if (ret) {
printk(KERN_ERR PFX "Connection failed\n");
@@ -1883,7 +1925,7 @@ static void srp_add_one(struct ib_device *device)
*/
srp_dev->fmr_page_shift = max(9, ffs(dev_attr->page_size_cap) - 1);
srp_dev->fmr_page_size = 1 << srp_dev->fmr_page_shift;
- srp_dev->fmr_page_mask = ~((unsigned long) srp_dev->fmr_page_size - 1);
+ srp_dev->fmr_page_mask = ~((u64) srp_dev->fmr_page_size - 1);
INIT_LIST_HEAD(&srp_dev->dev_list);
diff --git a/drivers/infiniband/ulp/srp/ib_srp.h b/drivers/infiniband/ulp/srp/ib_srp.h
index d4e35ef5137..2f3319c719a 100644
--- a/drivers/infiniband/ulp/srp/ib_srp.h
+++ b/drivers/infiniband/ulp/srp/ib_srp.h
@@ -87,7 +87,7 @@ struct srp_device {
struct ib_fmr_pool *fmr_pool;
int fmr_page_shift;
int fmr_page_size;
- unsigned long fmr_page_mask;
+ u64 fmr_page_mask;
};
struct srp_host {
@@ -158,10 +158,11 @@ struct srp_target_port {
struct completion done;
int status;
enum srp_target_state state;
+ int qp_in_error;
};
struct srp_iu {
- dma_addr_t dma;
+ u64 dma;
void *buf;
size_t size;
enum dma_data_direction direction;