summaryrefslogtreecommitdiff
path: root/drivers/infiniband/ulp
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/infiniband/ulp')
-rw-r--r--drivers/infiniband/ulp/ipoib/ipoib.h6
-rw-r--r--drivers/infiniband/ulp/ipoib/ipoib_cm.c6
-rw-r--r--drivers/infiniband/ulp/ipoib/ipoib_ib.c2
-rw-r--r--drivers/infiniband/ulp/ipoib/ipoib_main.c26
-rw-r--r--drivers/infiniband/ulp/iser/iscsi_iser.h1
-rw-r--r--drivers/infiniband/ulp/isert/ib_isert.c16
-rw-r--r--drivers/infiniband/ulp/rtrs/rtrs-clt-sysfs.c48
-rw-r--r--drivers/infiniband/ulp/rtrs/rtrs-clt.c122
-rw-r--r--drivers/infiniband/ulp/rtrs/rtrs-clt.h1
-rw-r--r--drivers/infiniband/ulp/rtrs/rtrs-pri.h3
-rw-r--r--drivers/infiniband/ulp/rtrs/rtrs-srv-sysfs.c20
-rw-r--r--drivers/infiniband/ulp/rtrs/rtrs-srv.c36
-rw-r--r--drivers/infiniband/ulp/rtrs/rtrs.c35
-rw-r--r--drivers/infiniband/ulp/rtrs/rtrs.h3
-rw-r--r--drivers/infiniband/ulp/srpt/ib_srpt.c4
15 files changed, 239 insertions, 90 deletions
diff --git a/drivers/infiniband/ulp/ipoib/ipoib.h b/drivers/infiniband/ulp/ipoib/ipoib.h
index 179ff1d068e5..75cd44789661 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib.h
+++ b/drivers/infiniband/ulp/ipoib/ipoib.h
@@ -501,9 +501,9 @@ void ipoib_reap_ah(struct work_struct *work);
struct ipoib_path *__path_find(struct net_device *dev, void *gid);
void ipoib_mark_paths_invalid(struct net_device *dev);
void ipoib_flush_paths(struct net_device *dev);
-struct net_device *ipoib_intf_alloc(struct ib_device *hca, u8 port,
+struct net_device *ipoib_intf_alloc(struct ib_device *hca, u32 port,
const char *format);
-int ipoib_intf_init(struct ib_device *hca, u8 port, const char *format,
+int ipoib_intf_init(struct ib_device *hca, u32 port, const char *format,
struct net_device *dev);
void ipoib_ib_tx_timer_func(struct timer_list *t);
void ipoib_ib_dev_flush_light(struct work_struct *work);
@@ -677,8 +677,6 @@ void ipoib_cm_handle_rx_wc(struct net_device *dev, struct ib_wc *wc);
void ipoib_cm_handle_tx_wc(struct net_device *dev, struct ib_wc *wc);
#else
-struct ipoib_cm_tx;
-
#define ipoib_max_conn_qp 0
static inline int ipoib_cm_admin_enabled(struct net_device *dev)
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_cm.c b/drivers/infiniband/ulp/ipoib/ipoib_cm.c
index d5d592bdab35..9dbc85a6b702 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_cm.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_cm.c
@@ -1122,12 +1122,8 @@ static int ipoib_cm_modify_tx_init(struct net_device *dev,
struct ipoib_dev_priv *priv = ipoib_priv(dev);
struct ib_qp_attr qp_attr;
int qp_attr_mask, ret;
- ret = ib_find_pkey(priv->ca, priv->port, priv->pkey, &qp_attr.pkey_index);
- if (ret) {
- ipoib_warn(priv, "pkey 0x%x not found: %d\n", priv->pkey, ret);
- return ret;
- }
+ qp_attr.pkey_index = priv->pkey_index;
qp_attr.qp_state = IB_QPS_INIT;
qp_attr.qp_access_flags = IB_ACCESS_LOCAL_WRITE;
qp_attr.port_num = priv->port;
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ib.c b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
index 494f413dc3c6..ceabfb0b0a83 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_ib.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
@@ -1060,7 +1060,7 @@ static bool ipoib_dev_addr_changed_valid(struct ipoib_dev_priv *priv)
union ib_gid *netdev_gid;
int err;
u16 index;
- u8 port;
+ u32 port;
bool ret = false;
netdev_gid = (union ib_gid *)(priv->dev->dev_addr + 4);
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c
index e16b40c09f82..bbb18087fdab 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_main.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c
@@ -90,7 +90,7 @@ static int ipoib_add_one(struct ib_device *device);
static void ipoib_remove_one(struct ib_device *device, void *client_data);
static void ipoib_neigh_reclaim(struct rcu_head *rp);
static struct net_device *ipoib_get_net_dev_by_params(
- struct ib_device *dev, u8 port, u16 pkey,
+ struct ib_device *dev, u32 port, u16 pkey,
const union ib_gid *gid, const struct sockaddr *addr,
void *client_data);
static int ipoib_set_mac(struct net_device *dev, void *addr);
@@ -164,8 +164,13 @@ int ipoib_open(struct net_device *dev)
dev_change_flags(cpriv->dev, flags | IFF_UP, NULL);
}
up_read(&priv->vlan_rwsem);
- }
+ } else if (priv->parent) {
+ struct ipoib_dev_priv *ppriv = ipoib_priv(priv->parent);
+ if (!test_bit(IPOIB_FLAG_ADMIN_UP, &ppriv->flags))
+ ipoib_dbg(priv, "parent device %s is not up, so child device may be not functioning.\n",
+ ppriv->dev->name);
+ }
netif_start_queue(dev);
return 0;
@@ -438,7 +443,7 @@ static int ipoib_match_gid_pkey_addr(struct ipoib_dev_priv *priv,
/* Returns the number of matching net_devs found (between 0 and 2). Also
* return the matching net_device in the @net_dev parameter, holding a
* reference to the net_device, if the number of matches >= 1 */
-static int __ipoib_get_net_dev_by_params(struct list_head *dev_list, u8 port,
+static int __ipoib_get_net_dev_by_params(struct list_head *dev_list, u32 port,
u16 pkey_index,
const union ib_gid *gid,
const struct sockaddr *addr,
@@ -463,7 +468,7 @@ static int __ipoib_get_net_dev_by_params(struct list_head *dev_list, u8 port,
}
static struct net_device *ipoib_get_net_dev_by_params(
- struct ib_device *dev, u8 port, u16 pkey,
+ struct ib_device *dev, u32 port, u16 pkey,
const union ib_gid *gid, const struct sockaddr *addr,
void *client_data)
{
@@ -1181,7 +1186,12 @@ unref:
static void ipoib_timeout(struct net_device *dev, unsigned int txqueue)
{
struct ipoib_dev_priv *priv = ipoib_priv(dev);
+ struct rdma_netdev *rn = netdev_priv(dev);
+ if (rn->tx_timeout) {
+ rn->tx_timeout(dev, txqueue);
+ return;
+ }
ipoib_warn(priv, "transmit timeout: latency %d msecs\n",
jiffies_to_msecs(jiffies - dev_trans_start(dev)));
ipoib_warn(priv,
@@ -2145,7 +2155,7 @@ static void ipoib_build_priv(struct net_device *dev)
INIT_DELAYED_WORK(&priv->neigh_reap_task, ipoib_reap_neigh);
}
-static struct net_device *ipoib_alloc_netdev(struct ib_device *hca, u8 port,
+static struct net_device *ipoib_alloc_netdev(struct ib_device *hca, u32 port,
const char *name)
{
struct net_device *dev;
@@ -2162,7 +2172,7 @@ static struct net_device *ipoib_alloc_netdev(struct ib_device *hca, u8 port,
return dev;
}
-int ipoib_intf_init(struct ib_device *hca, u8 port, const char *name,
+int ipoib_intf_init(struct ib_device *hca, u32 port, const char *name,
struct net_device *dev)
{
struct rdma_netdev *rn = netdev_priv(dev);
@@ -2213,7 +2223,7 @@ out:
return rc;
}
-struct net_device *ipoib_intf_alloc(struct ib_device *hca, u8 port,
+struct net_device *ipoib_intf_alloc(struct ib_device *hca, u32 port,
const char *name)
{
struct net_device *dev;
@@ -2456,7 +2466,7 @@ static int ipoib_intercept_dev_id_attr(struct net_device *dev)
}
static struct net_device *ipoib_add_port(const char *format,
- struct ib_device *hca, u8 port)
+ struct ib_device *hca, u32 port)
{
struct rtnl_link_ops *ops = ipoib_get_link_ops();
struct rdma_netdev_alloc_params params;
diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.h b/drivers/infiniband/ulp/iser/iscsi_iser.h
index 78ee9445f801..9f6ac0a09a78 100644
--- a/drivers/infiniband/ulp/iser/iscsi_iser.h
+++ b/drivers/infiniband/ulp/iser/iscsi_iser.h
@@ -297,7 +297,6 @@ struct iser_login_desc {
struct iser_conn;
struct ib_conn;
-struct iscsi_iser_task;
/**
* struct iser_device - iSER device handle
diff --git a/drivers/infiniband/ulp/isert/ib_isert.c b/drivers/infiniband/ulp/isert/ib_isert.c
index 7305ed8976c2..18266f07c58d 100644
--- a/drivers/infiniband/ulp/isert/ib_isert.c
+++ b/drivers/infiniband/ulp/isert/ib_isert.c
@@ -438,23 +438,23 @@ isert_connect_request(struct rdma_cm_id *cma_id, struct rdma_cm_event *event)
isert_init_conn(isert_conn);
isert_conn->cm_id = cma_id;
- ret = isert_alloc_login_buf(isert_conn, cma_id->device);
- if (ret)
- goto out;
-
device = isert_device_get(cma_id);
if (IS_ERR(device)) {
ret = PTR_ERR(device);
- goto out_rsp_dma_map;
+ goto out;
}
isert_conn->device = device;
+ ret = isert_alloc_login_buf(isert_conn, cma_id->device);
+ if (ret)
+ goto out_conn_dev;
+
isert_set_nego_params(isert_conn, &event->param.conn);
isert_conn->qp = isert_create_qp(isert_conn, cma_id);
if (IS_ERR(isert_conn->qp)) {
ret = PTR_ERR(isert_conn->qp);
- goto out_conn_dev;
+ goto out_rsp_dma_map;
}
ret = isert_login_post_recv(isert_conn);
@@ -473,10 +473,10 @@ isert_connect_request(struct rdma_cm_id *cma_id, struct rdma_cm_event *event)
out_destroy_qp:
isert_destroy_qp(isert_conn);
-out_conn_dev:
- isert_device_put(device);
out_rsp_dma_map:
isert_free_login_buf(isert_conn);
+out_conn_dev:
+ isert_device_put(device);
out:
kfree(isert_conn);
rdma_reject(cma_id, NULL, 0, IB_CM_REJ_CONSUMER_DEFINED);
diff --git a/drivers/infiniband/ulp/rtrs/rtrs-clt-sysfs.c b/drivers/infiniband/ulp/rtrs/rtrs-clt-sysfs.c
index b6a0abf40589..7d53d18a5004 100644
--- a/drivers/infiniband/ulp/rtrs/rtrs-clt-sysfs.c
+++ b/drivers/infiniband/ulp/rtrs/rtrs-clt-sysfs.c
@@ -101,6 +101,9 @@ static ssize_t mpath_policy_show(struct device *dev,
case MP_POLICY_MIN_INFLIGHT:
return sysfs_emit(page, "min-inflight (MI: %d)\n",
clt->mp_policy);
+ case MP_POLICY_MIN_LATENCY:
+ return sysfs_emit(page, "min-latency (ML: %d)\n",
+ clt->mp_policy);
default:
return sysfs_emit(page, "Unknown (%d)\n", clt->mp_policy);
}
@@ -114,22 +117,32 @@ static ssize_t mpath_policy_store(struct device *dev,
struct rtrs_clt *clt;
int value;
int ret;
+ size_t len = 0;
clt = container_of(dev, struct rtrs_clt, dev);
ret = kstrtoint(buf, 10, &value);
if (!ret && (value == MP_POLICY_RR ||
- value == MP_POLICY_MIN_INFLIGHT)) {
+ value == MP_POLICY_MIN_INFLIGHT ||
+ value == MP_POLICY_MIN_LATENCY)) {
clt->mp_policy = value;
return count;
}
+ /* distinguish "mi" and "min-latency" with length */
+ len = strnlen(buf, NAME_MAX);
+ if (buf[len - 1] == '\n')
+ len--;
+
if (!strncasecmp(buf, "round-robin", 11) ||
- !strncasecmp(buf, "rr", 2))
+ (len == 2 && !strncasecmp(buf, "rr", 2)))
clt->mp_policy = MP_POLICY_RR;
else if (!strncasecmp(buf, "min-inflight", 12) ||
- !strncasecmp(buf, "mi", 2))
+ (len == 2 && !strncasecmp(buf, "mi", 2)))
clt->mp_policy = MP_POLICY_MIN_INFLIGHT;
+ else if (!strncasecmp(buf, "min-latency", 11) ||
+ (len == 2 && !strncasecmp(buf, "ml", 2)))
+ clt->mp_policy = MP_POLICY_MIN_LATENCY;
else
return -EINVAL;
@@ -342,6 +355,21 @@ static ssize_t rtrs_clt_hca_name_show(struct kobject *kobj,
static struct kobj_attribute rtrs_clt_hca_name_attr =
__ATTR(hca_name, 0444, rtrs_clt_hca_name_show, NULL);
+static ssize_t rtrs_clt_cur_latency_show(struct kobject *kobj,
+ struct kobj_attribute *attr,
+ char *page)
+{
+ struct rtrs_clt_sess *sess;
+
+ sess = container_of(kobj, struct rtrs_clt_sess, kobj);
+
+ return sysfs_emit(page, "%lld ns\n",
+ ktime_to_ns(sess->s.hb_cur_latency));
+}
+
+static struct kobj_attribute rtrs_clt_cur_latency_attr =
+ __ATTR(cur_latency, 0444, rtrs_clt_cur_latency_show, NULL);
+
static ssize_t rtrs_clt_src_addr_show(struct kobject *kobj,
struct kobj_attribute *attr,
char *page)
@@ -385,6 +413,7 @@ static struct attribute *rtrs_clt_sess_attrs[] = {
&rtrs_clt_reconnect_attr.attr,
&rtrs_clt_disconnect_attr.attr,
&rtrs_clt_remove_path_attr.attr,
+ &rtrs_clt_cur_latency_attr.attr,
NULL,
};
@@ -396,14 +425,13 @@ int rtrs_clt_create_sess_files(struct rtrs_clt_sess *sess)
{
struct rtrs_clt *clt = sess->clt;
char str[NAME_MAX];
- int err, cnt;
-
- cnt = sockaddr_to_str((struct sockaddr *)&sess->s.src_addr,
- str, sizeof(str));
- cnt += scnprintf(str + cnt, sizeof(str) - cnt, "@");
- sockaddr_to_str((struct sockaddr *)&sess->s.dst_addr,
- str + cnt, sizeof(str) - cnt);
+ int err;
+ struct rtrs_addr path = {
+ .src = &sess->s.src_addr,
+ .dst = &sess->s.dst_addr,
+ };
+ rtrs_addr_to_str(&path, str, sizeof(str));
err = kobject_init_and_add(&sess->kobj, &ktype_sess, clt->kobj_paths,
"%s", str);
if (err) {
diff --git a/drivers/infiniband/ulp/rtrs/rtrs-clt.c b/drivers/infiniband/ulp/rtrs/rtrs-clt.c
index b74a872387c4..40f4383764de 100644
--- a/drivers/infiniband/ulp/rtrs/rtrs-clt.c
+++ b/drivers/infiniband/ulp/rtrs/rtrs-clt.c
@@ -325,7 +325,7 @@ static void rtrs_rdma_error_recovery(struct rtrs_clt_con *con)
static void rtrs_clt_fast_reg_done(struct ib_cq *cq, struct ib_wc *wc)
{
- struct rtrs_clt_con *con = cq->cq_context;
+ struct rtrs_clt_con *con = to_clt_con(wc->qp->qp_context);
if (unlikely(wc->status != IB_WC_SUCCESS)) {
rtrs_err(con->c.sess, "Failed IB_WR_REG_MR: %s\n",
@@ -345,7 +345,7 @@ static void rtrs_clt_inv_rkey_done(struct ib_cq *cq, struct ib_wc *wc)
{
struct rtrs_clt_io_req *req =
container_of(wc->wr_cqe, typeof(*req), inv_cqe);
- struct rtrs_clt_con *con = cq->cq_context;
+ struct rtrs_clt_con *con = to_clt_con(wc->qp->qp_context);
if (unlikely(wc->status != IB_WC_SUCCESS)) {
rtrs_err(con->c.sess, "Failed IB_WR_LOCAL_INV: %s\n",
@@ -437,6 +437,13 @@ static void complete_rdma_req(struct rtrs_clt_io_req *req, int errno,
req->in_use = false;
req->con = NULL;
+ if (errno) {
+ rtrs_err_rl(con->c.sess,
+ "IO request failed: error=%d path=%s [%s:%u]\n",
+ errno, kobject_name(&sess->kobj), sess->hca_name,
+ sess->hca_port);
+ }
+
if (notify)
req->conf(req->priv, errno);
}
@@ -586,7 +593,7 @@ static int rtrs_post_recv_empty_x2(struct rtrs_con *con, struct ib_cqe *cqe)
static void rtrs_clt_rdma_done(struct ib_cq *cq, struct ib_wc *wc)
{
- struct rtrs_clt_con *con = cq->cq_context;
+ struct rtrs_clt_con *con = to_clt_con(wc->qp->qp_context);
struct rtrs_clt_sess *sess = to_clt_sess(con->c.sess);
u32 imm_type, imm_payload;
bool w_inval = false;
@@ -628,6 +635,8 @@ static void rtrs_clt_rdma_done(struct ib_cq *cq, struct ib_wc *wc)
} else if (imm_type == RTRS_HB_ACK_IMM) {
WARN_ON(con->c.cid);
sess->s.hb_missed_cnt = 0;
+ sess->s.hb_cur_latency =
+ ktime_sub(ktime_get(), sess->s.hb_last_sent);
if (sess->flags & RTRS_MSG_NEW_RKEY_F)
return rtrs_clt_recv_done(con, wc);
} else {
@@ -826,6 +835,57 @@ static struct rtrs_clt_sess *get_next_path_min_inflight(struct path_it *it)
return min_path;
}
+/**
+ * get_next_path_min_latency() - Returns path with minimal latency.
+ * @it: the path pointer
+ *
+ * Return: a path with the lowest latency or NULL if all paths are tried
+ *
+ * Locks:
+ * rcu_read_lock() must be hold.
+ *
+ * Related to @MP_POLICY_MIN_LATENCY
+ *
+ * This DOES skip an already-tried path.
+ * There is a skip-list to skip a path if the path has tried but failed.
+ * It will try the minimum latency path and then the second minimum latency
+ * path and so on. Finally it will return NULL if all paths are tried.
+ * Therefore the caller MUST check the returned
+ * path is NULL and trigger the IO error.
+ */
+static struct rtrs_clt_sess *get_next_path_min_latency(struct path_it *it)
+{
+ struct rtrs_clt_sess *min_path = NULL;
+ struct rtrs_clt *clt = it->clt;
+ struct rtrs_clt_sess *sess;
+ ktime_t min_latency = INT_MAX;
+ ktime_t latency;
+
+ list_for_each_entry_rcu(sess, &clt->paths_list, s.entry) {
+ if (unlikely(READ_ONCE(sess->state) != RTRS_CLT_CONNECTED))
+ continue;
+
+ if (unlikely(!list_empty(raw_cpu_ptr(sess->mp_skip_entry))))
+ continue;
+
+ latency = sess->s.hb_cur_latency;
+
+ if (latency < min_latency) {
+ min_latency = latency;
+ min_path = sess;
+ }
+ }
+
+ /*
+ * add the path to the skip list, so that next time we can get
+ * a different one
+ */
+ if (min_path)
+ list_add(raw_cpu_ptr(min_path->mp_skip_entry), &it->skip_list);
+
+ return min_path;
+}
+
static inline void path_it_init(struct path_it *it, struct rtrs_clt *clt)
{
INIT_LIST_HEAD(&it->skip_list);
@@ -834,8 +894,10 @@ static inline void path_it_init(struct path_it *it, struct rtrs_clt *clt)
if (clt->mp_policy == MP_POLICY_RR)
it->next_path = get_next_path_rr;
- else
+ else if (clt->mp_policy == MP_POLICY_MIN_INFLIGHT)
it->next_path = get_next_path_min_inflight;
+ else
+ it->next_path = get_next_path_min_latency;
}
static inline void path_it_deinit(struct path_it *it)
@@ -1020,7 +1082,10 @@ static int rtrs_clt_write_req(struct rtrs_clt_io_req *req)
req->usr_len + sizeof(*msg),
imm);
if (unlikely(ret)) {
- rtrs_err(s, "Write request failed: %d\n", ret);
+ rtrs_err_rl(s,
+ "Write request failed: error=%d path=%s [%s:%u]\n",
+ ret, kobject_name(&sess->kobj), sess->hca_name,
+ sess->hca_port);
if (sess->clt->mp_policy == MP_POLICY_MIN_INFLIGHT)
atomic_dec(&sess->stats->inflight);
if (req->sg_cnt)
@@ -1052,7 +1117,7 @@ static int rtrs_clt_read_req(struct rtrs_clt_io_req *req)
struct rtrs_sess *s = con->c.sess;
struct rtrs_clt_sess *sess = to_clt_sess(s);
struct rtrs_msg_rdma_read *msg;
- struct rtrs_ib_dev *dev;
+ struct rtrs_ib_dev *dev = sess->s.dev;
struct ib_reg_wr rwr;
struct ib_send_wr *wr = NULL;
@@ -1062,9 +1127,6 @@ static int rtrs_clt_read_req(struct rtrs_clt_io_req *req)
const size_t tsize = sizeof(*msg) + req->data_len + req->usr_len;
- s = &sess->s;
- dev = sess->s.dev;
-
if (unlikely(tsize > sess->chunk_size)) {
rtrs_wrn(s,
"Read request failed, message size is %zu, bigger than CHUNK_SIZE %d\n",
@@ -1141,7 +1203,10 @@ static int rtrs_clt_read_req(struct rtrs_clt_io_req *req)
ret = rtrs_post_send_rdma(req->con, req, &sess->rbufs[buf_id],
req->data_len, imm, wr);
if (unlikely(ret)) {
- rtrs_err(s, "Read request failed: %d\n", ret);
+ rtrs_err_rl(s,
+ "Read request failed: error=%d path=%s [%s:%u]\n",
+ ret, kobject_name(&sess->kobj), sess->hca_name,
+ sess->hca_port);
if (sess->clt->mp_policy == MP_POLICY_MIN_INFLIGHT)
atomic_dec(&sess->stats->inflight);
req->need_inv = false;
@@ -1863,12 +1928,14 @@ static int rtrs_clt_rdma_cm_handler(struct rdma_cm_id *cm_id,
case RDMA_CM_EVENT_UNREACHABLE:
case RDMA_CM_EVENT_ADDR_CHANGE:
case RDMA_CM_EVENT_TIMEWAIT_EXIT:
- rtrs_wrn(s, "CM error event %d\n", ev->event);
+ rtrs_wrn(s, "CM error (CM event: %s, err: %d)\n",
+ rdma_event_msg(ev->event), ev->status);
cm_err = -ECONNRESET;
break;
case RDMA_CM_EVENT_ADDR_ERROR:
case RDMA_CM_EVENT_ROUTE_ERROR:
- rtrs_wrn(s, "CM error event %d\n", ev->event);
+ rtrs_wrn(s, "CM error (CM event: %s, err: %d)\n",
+ rdma_event_msg(ev->event), ev->status);
cm_err = -EHOSTUNREACH;
break;
case RDMA_CM_EVENT_DEVICE_REMOVAL:
@@ -1878,7 +1945,8 @@ static int rtrs_clt_rdma_cm_handler(struct rdma_cm_id *cm_id,
rtrs_clt_close_conns(sess, false);
return 0;
default:
- rtrs_err(s, "Unexpected RDMA CM event (%d)\n", ev->event);
+ rtrs_err(s, "Unexpected RDMA CM error (CM event: %s, err: %d)\n",
+ rdma_event_msg(ev->event), ev->status);
cm_err = -ECONNRESET;
break;
}
@@ -2251,7 +2319,7 @@ destroy:
static void rtrs_clt_info_req_done(struct ib_cq *cq, struct ib_wc *wc)
{
- struct rtrs_clt_con *con = cq->cq_context;
+ struct rtrs_clt_con *con = to_clt_con(wc->qp->qp_context);
struct rtrs_clt_sess *sess = to_clt_sess(con->c.sess);
struct rtrs_iu *iu;
@@ -2333,7 +2401,7 @@ static int process_info_rsp(struct rtrs_clt_sess *sess,
static void rtrs_clt_info_rsp_done(struct ib_cq *cq, struct ib_wc *wc)
{
- struct rtrs_clt_con *con = cq->cq_context;
+ struct rtrs_clt_con *con = to_clt_con(wc->qp->qp_context);
struct rtrs_clt_sess *sess = to_clt_sess(con->c.sess);
struct rtrs_msg_info_rsp *msg;
enum rtrs_clt_state state;
@@ -2464,16 +2532,28 @@ out:
static int init_sess(struct rtrs_clt_sess *sess)
{
int err;
+ char str[NAME_MAX];
+ struct rtrs_addr path = {
+ .src = &sess->s.src_addr,
+ .dst = &sess->s.dst_addr,
+ };
+
+ rtrs_addr_to_str(&path, str, sizeof(str));
mutex_lock(&sess->init_mutex);
err = init_conns(sess);
if (err) {
- rtrs_err(sess->clt, "init_conns(), err: %d\n", err);
+ rtrs_err(sess->clt,
+ "init_conns() failed: err=%d path=%s [%s:%u]\n", err,
+ str, sess->hca_name, sess->hca_port);
goto out;
}
err = rtrs_send_sess_info(sess);
if (err) {
- rtrs_err(sess->clt, "rtrs_send_sess_info(), err: %d\n", err);
+ rtrs_err(
+ sess->clt,
+ "rtrs_send_sess_info() failed: err=%d path=%s [%s:%u]\n",
+ err, str, sess->hca_name, sess->hca_port);
goto out;
}
rtrs_clt_sess_up(sess);
@@ -2791,8 +2871,8 @@ int rtrs_clt_remove_path_from_sysfs(struct rtrs_clt_sess *sess,
} while (!changed && old_state != RTRS_CLT_DEAD);
if (likely(changed)) {
- rtrs_clt_destroy_sess_files(sess, sysfs_self);
rtrs_clt_remove_path_from_arr(sess);
+ rtrs_clt_destroy_sess_files(sess, sysfs_self);
kobject_put(&sess->kobj);
}
@@ -2933,9 +3013,9 @@ int rtrs_clt_query(struct rtrs_clt *clt, struct rtrs_attrs *attr)
return -ECOMM;
attr->queue_depth = clt->queue_depth;
- attr->max_io_size = clt->max_io_size;
- attr->sess_kobj = &clt->dev.kobj;
- strlcpy(attr->sessname, clt->sessname, sizeof(attr->sessname));
+ /* Cap max_io_size to min of remote buffer size and the fr pages */
+ attr->max_io_size = min_t(int, clt->max_io_size,
+ clt->max_segments * SZ_4K);
return 0;
}
diff --git a/drivers/infiniband/ulp/rtrs/rtrs-clt.h b/drivers/infiniband/ulp/rtrs/rtrs-clt.h
index 98ba5d0a48b8..4c52f30e4da1 100644
--- a/drivers/infiniband/ulp/rtrs/rtrs-clt.h
+++ b/drivers/infiniband/ulp/rtrs/rtrs-clt.h
@@ -29,6 +29,7 @@ enum rtrs_clt_state {
enum rtrs_mp_policy {
MP_POLICY_RR,
MP_POLICY_MIN_INFLIGHT,
+ MP_POLICY_MIN_LATENCY,
};
/* see Documentation/ABI/testing/sysfs-class-rtrs-client for details */
diff --git a/drivers/infiniband/ulp/rtrs/rtrs-pri.h b/drivers/infiniband/ulp/rtrs/rtrs-pri.h
index 00eb45053339..86e65cf30cab 100644
--- a/drivers/infiniband/ulp/rtrs/rtrs-pri.h
+++ b/drivers/infiniband/ulp/rtrs/rtrs-pri.h
@@ -91,6 +91,7 @@ struct rtrs_con {
struct ib_cq *cq;
struct rdma_cm_id *cm_id;
unsigned int cid;
+ u16 cq_size;
};
struct rtrs_sess {
@@ -112,6 +113,8 @@ struct rtrs_sess {
unsigned int hb_interval_ms;
unsigned int hb_missed_cnt;
unsigned int hb_missed_max;
+ ktime_t hb_last_sent;
+ ktime_t hb_cur_latency;
};
/* rtrs information unit */
diff --git a/drivers/infiniband/ulp/rtrs/rtrs-srv-sysfs.c b/drivers/infiniband/ulp/rtrs/rtrs-srv-sysfs.c
index 126a96e75c62..a9288175fbb5 100644
--- a/drivers/infiniband/ulp/rtrs/rtrs-srv-sysfs.c
+++ b/drivers/infiniband/ulp/rtrs/rtrs-srv-sysfs.c
@@ -176,7 +176,8 @@ static int rtrs_srv_create_once_sysfs_root_folders(struct rtrs_srv_sess *sess)
err = device_add(&srv->dev);
if (err) {
pr_err("device_add(): %d\n", err);
- goto put;
+ put_device(&srv->dev);
+ goto unlock;
}
srv->kobj_paths = kobject_create_and_add("paths", &srv->dev.kobj);
if (!srv->kobj_paths) {
@@ -188,10 +189,6 @@ static int rtrs_srv_create_once_sysfs_root_folders(struct rtrs_srv_sess *sess)
}
dev_set_uevent_suppress(&srv->dev, false);
kobject_uevent(&srv->dev.kobj, KOBJ_ADD);
- goto unlock;
-
-put:
- put_device(&srv->dev);
unlock:
mutex_unlock(&srv->paths_mutex);
@@ -262,14 +259,13 @@ int rtrs_srv_create_sess_files(struct rtrs_srv_sess *sess)
struct rtrs_srv *srv = sess->srv;
struct rtrs_sess *s = &sess->s;
char str[NAME_MAX];
- int err, cnt;
-
- cnt = sockaddr_to_str((struct sockaddr *)&sess->s.dst_addr,
- str, sizeof(str));
- cnt += scnprintf(str + cnt, sizeof(str) - cnt, "@");
- sockaddr_to_str((struct sockaddr *)&sess->s.src_addr,
- str + cnt, sizeof(str) - cnt);
+ int err;
+ struct rtrs_addr path = {
+ .src = &sess->s.dst_addr,
+ .dst = &sess->s.src_addr,
+ };
+ rtrs_addr_to_str(&path, str, sizeof(str));
err = rtrs_srv_create_once_sysfs_root_folders(sess);
if (err)
return err;
diff --git a/drivers/infiniband/ulp/rtrs/rtrs-srv.c b/drivers/infiniband/ulp/rtrs/rtrs-srv.c
index f7aa2a7e7442..0fa116cabc44 100644
--- a/drivers/infiniband/ulp/rtrs/rtrs-srv.c
+++ b/drivers/infiniband/ulp/rtrs/rtrs-srv.c
@@ -199,7 +199,7 @@ static void rtrs_srv_wait_ops_ids(struct rtrs_srv_sess *sess)
static void rtrs_srv_reg_mr_done(struct ib_cq *cq, struct ib_wc *wc)
{
- struct rtrs_srv_con *con = cq->cq_context;
+ struct rtrs_srv_con *con = to_srv_con(wc->qp->qp_context);
struct rtrs_sess *s = con->c.sess;
struct rtrs_srv_sess *sess = to_srv_sess(s);
@@ -518,8 +518,9 @@ bool rtrs_srv_resp_rdma(struct rtrs_srv_op *id, int status)
if (unlikely(sess->state != RTRS_SRV_CONNECTED)) {
rtrs_err_rl(s,
- "Sending I/O response failed, session is disconnected, sess state %s\n",
- rtrs_srv_state_str(sess->state));
+ "Sending I/O response failed, session %s is disconnected, sess state %s\n",
+ kobject_name(&sess->kobj),
+ rtrs_srv_state_str(sess->state));
goto out;
}
if (always_invalidate) {
@@ -529,7 +530,9 @@ bool rtrs_srv_resp_rdma(struct rtrs_srv_op *id, int status)
}
if (unlikely(atomic_sub_return(1,
&con->sq_wr_avail) < 0)) {
- pr_err("IB send queue full\n");
+ rtrs_err(s, "IB send queue full: sess=%s cid=%d\n",
+ kobject_name(&sess->kobj),
+ con->c.cid);
atomic_add(1, &con->sq_wr_avail);
spin_lock(&con->rsp_wr_wait_lock);
list_add_tail(&id->wait_list, &con->rsp_wr_wait_list);
@@ -543,7 +546,8 @@ bool rtrs_srv_resp_rdma(struct rtrs_srv_op *id, int status)
err = rdma_write_sg(id);
if (unlikely(err)) {
- rtrs_err_rl(s, "IO response failed: %d\n", err);
+ rtrs_err_rl(s, "IO response failed: %d: sess=%s\n", err,
+ kobject_name(&sess->kobj));
close_sess(sess);
}
out:
@@ -720,7 +724,7 @@ static void rtrs_srv_stop_hb(struct rtrs_srv_sess *sess)
static void rtrs_srv_info_rsp_done(struct ib_cq *cq, struct ib_wc *wc)
{
- struct rtrs_srv_con *con = cq->cq_context;
+ struct rtrs_srv_con *con = to_srv_con(wc->qp->qp_context);
struct rtrs_sess *s = con->c.sess;
struct rtrs_srv_sess *sess = to_srv_sess(s);
struct rtrs_iu *iu;
@@ -862,7 +866,7 @@ rwr_free:
static void rtrs_srv_info_req_done(struct ib_cq *cq, struct ib_wc *wc)
{
- struct rtrs_srv_con *con = cq->cq_context;
+ struct rtrs_srv_con *con = to_srv_con(wc->qp->qp_context);
struct rtrs_sess *s = con->c.sess;
struct rtrs_srv_sess *sess = to_srv_sess(s);
struct rtrs_msg_info_req *msg;
@@ -1110,7 +1114,7 @@ static void rtrs_srv_inv_rkey_done(struct ib_cq *cq, struct ib_wc *wc)
{
struct rtrs_srv_mr *mr =
container_of(wc->wr_cqe, typeof(*mr), inv_cqe);
- struct rtrs_srv_con *con = cq->cq_context;
+ struct rtrs_srv_con *con = to_srv_con(wc->qp->qp_context);
struct rtrs_sess *s = con->c.sess;
struct rtrs_srv_sess *sess = to_srv_sess(s);
struct rtrs_srv *srv = sess->srv;
@@ -1167,7 +1171,7 @@ static void rtrs_rdma_process_wr_wait_list(struct rtrs_srv_con *con)
static void rtrs_srv_rdma_done(struct ib_cq *cq, struct ib_wc *wc)
{
- struct rtrs_srv_con *con = cq->cq_context;
+ struct rtrs_srv_con *con = to_srv_con(wc->qp->qp_context);
struct rtrs_sess *s = con->c.sess;
struct rtrs_srv_sess *sess = to_srv_sess(s);
struct rtrs_srv *srv = sess->srv;
@@ -1683,6 +1687,8 @@ static struct rtrs_srv_sess *__alloc_sess(struct rtrs_srv *srv,
{
struct rtrs_srv_sess *sess;
int err = -ENOMEM;
+ char str[NAME_MAX];
+ struct rtrs_addr path;
if (srv->paths_num >= MAX_PATHS_NUM) {
err = -ECONNRESET;
@@ -1717,6 +1723,13 @@ static struct rtrs_srv_sess *__alloc_sess(struct rtrs_srv *srv,
sess->cur_cq_vector = -1;
sess->s.dst_addr = cm_id->route.addr.dst_addr;
sess->s.src_addr = cm_id->route.addr.src_addr;
+
+ /* temporary until receiving session-name from client */
+ path.src = &sess->s.src_addr;
+ path.dst = &sess->s.dst_addr;
+ rtrs_addr_to_str(&path, str, sizeof(str));
+ strlcpy(sess->s.sessname, str, sizeof(sess->s.sessname));
+
sess->s.con_num = con_num;
sess->s.recon_cnt = recon_cnt;
uuid_copy(&sess->s.uuid, uuid);
@@ -1908,13 +1921,10 @@ static int rtrs_srv_rdma_cm_handler(struct rdma_cm_id *cm_id,
case RDMA_CM_EVENT_UNREACHABLE:
rtrs_err(s, "CM error (CM event: %s, err: %d)\n",
rdma_event_msg(ev->event), ev->status);
- close_sess(sess);
- break;
+ fallthrough;
case RDMA_CM_EVENT_DISCONNECTED:
case RDMA_CM_EVENT_ADDR_CHANGE:
case RDMA_CM_EVENT_TIMEWAIT_EXIT:
- close_sess(sess);
- break;
case RDMA_CM_EVENT_DEVICE_REMOVAL:
close_sess(sess);
break;
diff --git a/drivers/infiniband/ulp/rtrs/rtrs.c b/drivers/infiniband/ulp/rtrs/rtrs.c
index d13aff0aa816..a7847282a2eb 100644
--- a/drivers/infiniband/ulp/rtrs/rtrs.c
+++ b/drivers/infiniband/ulp/rtrs/rtrs.c
@@ -218,14 +218,14 @@ static int create_cq(struct rtrs_con *con, int cq_vector, u16 cq_size,
struct rdma_cm_id *cm_id = con->cm_id;
struct ib_cq *cq;
- cq = ib_alloc_cq(cm_id->device, con, cq_size,
- cq_vector, poll_ctx);
+ cq = ib_cq_pool_get(cm_id->device, cq_size, cq_vector, poll_ctx);
if (IS_ERR(cq)) {
rtrs_err(con->sess, "Creating completion queue failed, errno: %ld\n",
PTR_ERR(cq));
return PTR_ERR(cq);
}
con->cq = cq;
+ con->cq_size = cq_size;
return 0;
}
@@ -273,7 +273,7 @@ int rtrs_cq_qp_create(struct rtrs_sess *sess, struct rtrs_con *con,
err = create_qp(con, sess->dev->ib_pd, max_send_wr, max_recv_wr,
max_send_sge);
if (err) {
- ib_free_cq(con->cq);
+ ib_cq_pool_put(con->cq, con->cq_size);
con->cq = NULL;
return err;
}
@@ -290,7 +290,7 @@ void rtrs_cq_qp_destroy(struct rtrs_con *con)
con->qp = NULL;
}
if (con->cq) {
- ib_free_cq(con->cq);
+ ib_cq_pool_put(con->cq, con->cq_size);
con->cq = NULL;
}
}
@@ -337,6 +337,9 @@ static void hb_work(struct work_struct *work)
schedule_hb(sess);
return;
}
+
+ sess->hb_last_sent = ktime_get();
+
imm = rtrs_to_imm(RTRS_HB_MSG_IMM, 0);
err = rtrs_post_rdma_write_imm_empty(usr_con, sess->hb_cqe, imm,
0, NULL);
@@ -464,6 +467,30 @@ int sockaddr_to_str(const struct sockaddr *addr, char *buf, size_t len)
EXPORT_SYMBOL(sockaddr_to_str);
/**
+ * rtrs_addr_to_str() - convert rtrs_addr to a string "src@dst"
+ * @addr: the rtrs_addr structure to be converted
+ * @buf: string containing source and destination addr of a path
+ * separated by '@' I.e. "ip:1.1.1.1@ip:1.1.1.2"
+ * "ip:1.1.1.1@ip:1.1.1.2".
+ * @len: string length
+ *
+ * The return value is the number of characters written into buf not
+ * including the trailing '\0'.
+ */
+int rtrs_addr_to_str(const struct rtrs_addr *addr, char *buf, size_t len)
+{
+ int cnt;
+
+ cnt = sockaddr_to_str((struct sockaddr *)addr->src,
+ buf, len);
+ cnt += scnprintf(buf + cnt, len - cnt, "@");
+ sockaddr_to_str((struct sockaddr *)addr->dst,
+ buf + cnt, len - cnt);
+ return cnt;
+}
+EXPORT_SYMBOL(rtrs_addr_to_str);
+
+/**
* rtrs_addr_to_sockaddr() - convert path string "src,dst" or "src@dst"
* to sockaddreses
* @str: string containing source and destination addr of a path
diff --git a/drivers/infiniband/ulp/rtrs/rtrs.h b/drivers/infiniband/ulp/rtrs/rtrs.h
index bebaa94c4728..dc3e1af1a85b 100644
--- a/drivers/infiniband/ulp/rtrs/rtrs.h
+++ b/drivers/infiniband/ulp/rtrs/rtrs.h
@@ -110,8 +110,6 @@ int rtrs_clt_rdma_cq_direct(struct rtrs_clt *clt, unsigned int index);
struct rtrs_attrs {
u32 queue_depth;
u32 max_io_size;
- u8 sessname[NAME_MAX];
- struct kobject *sess_kobj;
};
int rtrs_clt_query(struct rtrs_clt *sess, struct rtrs_attrs *attr);
@@ -185,4 +183,5 @@ int rtrs_addr_to_sockaddr(const char *str, size_t len, u16 port,
struct rtrs_addr *addr);
int sockaddr_to_str(const struct sockaddr *addr, char *buf, size_t len);
+int rtrs_addr_to_str(const struct rtrs_addr *addr, char *buf, size_t len);
#endif
diff --git a/drivers/infiniband/ulp/srpt/ib_srpt.c b/drivers/infiniband/ulp/srpt/ib_srpt.c
index 51c386a215f5..ea447805d4ea 100644
--- a/drivers/infiniband/ulp/srpt/ib_srpt.c
+++ b/drivers/infiniband/ulp/srpt/ib_srpt.c
@@ -2382,6 +2382,7 @@ static int srpt_cm_req_recv(struct srpt_device *const sdev,
pr_info("rejected SRP_LOGIN_REQ because target %s_%d is not enabled\n",
dev_name(&sdev->device->dev), port_num);
mutex_unlock(&sport->mutex);
+ ret = -EINVAL;
goto reject;
}
@@ -3109,7 +3110,8 @@ static int srpt_add_one(struct ib_device *device)
{
struct srpt_device *sdev;
struct srpt_port *sport;
- int i, ret;
+ int ret;
+ u32 i;
pr_debug("device = %p\n", device);