Merge branches 'cxgb4-2', 'i40iw-2', 'ipoib', 'misc-4.7' and 'mlx5-fcs' into k.o/for-4.7
diff --git a/drivers/infiniband/core/cma.c b/drivers/infiniband/core/cma.c
index 6ebaf20..f0c91ba 100644
--- a/drivers/infiniband/core/cma.c
+++ b/drivers/infiniband/core/cma.c
@@ -4295,7 +4295,8 @@
if (ret)
goto err;
- if (ibnl_add_client(RDMA_NL_RDMA_CM, RDMA_NL_RDMA_CM_NUM_OPS, cma_cb_table))
+ if (ibnl_add_client(RDMA_NL_RDMA_CM, ARRAY_SIZE(cma_cb_table),
+ cma_cb_table))
pr_warn("RDMA CMA: failed to add netlink callback\n");
cma_configfs_init();
diff --git a/drivers/infiniband/core/iwcm.c b/drivers/infiniband/core/iwcm.c
index e28a160..f057204 100644
--- a/drivers/infiniband/core/iwcm.c
+++ b/drivers/infiniband/core/iwcm.c
@@ -459,7 +459,7 @@
if (pm_addr->ss_family == AF_INET) {
struct sockaddr_in *pm4_addr = (struct sockaddr_in *)pm_addr;
- if (pm4_addr->sin_addr.s_addr == INADDR_ANY) {
+ if (pm4_addr->sin_addr.s_addr == htonl(INADDR_ANY)) {
struct sockaddr_in *cm4_addr =
(struct sockaddr_in *)cm_addr;
struct sockaddr_in *cm4_outaddr =
@@ -1175,7 +1175,7 @@
if (ret)
pr_err("iw_cm: couldn't init iwpm\n");
- ret = ibnl_add_client(RDMA_NL_IWCM, RDMA_NL_IWPM_NUM_OPS,
+ ret = ibnl_add_client(RDMA_NL_IWCM, ARRAY_SIZE(iwcm_nl_cb_table),
iwcm_nl_cb_table);
if (ret)
pr_err("iw_cm: couldn't register netlink callbacks\n");
diff --git a/drivers/infiniband/core/iwpm_util.c b/drivers/infiniband/core/iwpm_util.c
index 9b2bf2f..b65e06c 100644
--- a/drivers/infiniband/core/iwpm_util.c
+++ b/drivers/infiniband/core/iwpm_util.c
@@ -634,6 +634,7 @@
if (!(ibnl_put_msg(skb, &nlh, 0, 0, nl_client,
RDMA_NL_IWPM_MAPINFO, NLM_F_MULTI))) {
pr_warn("%s Unable to put NLMSG_DONE\n", __func__);
+ dev_kfree_skb(skb);
return -ENOMEM;
}
nlh->nlmsg_type = NLMSG_DONE;
diff --git a/drivers/infiniband/core/netlink.c b/drivers/infiniband/core/netlink.c
index d47df93..9b8c20c 100644
--- a/drivers/infiniband/core/netlink.c
+++ b/drivers/infiniband/core/netlink.c
@@ -151,12 +151,11 @@
struct ibnl_client *client;
int type = nlh->nlmsg_type;
int index = RDMA_NL_GET_CLIENT(type);
- int op = RDMA_NL_GET_OP(type);
+ unsigned int op = RDMA_NL_GET_OP(type);
list_for_each_entry(client, &client_list, list) {
if (client->index == index) {
- if (op < 0 || op >= client->nops ||
- !client->cb_table[op].dump)
+ if (op >= client->nops || !client->cb_table[op].dump)
return -EINVAL;
/*
diff --git a/drivers/infiniband/core/sa_query.c b/drivers/infiniband/core/sa_query.c
index 8a09c0f..3ebd108 100644
--- a/drivers/infiniband/core/sa_query.c
+++ b/drivers/infiniband/core/sa_query.c
@@ -536,7 +536,7 @@
data = ibnl_put_msg(skb, &nlh, query->seq, 0, RDMA_NL_LS,
RDMA_NL_LS_OP_RESOLVE, NLM_F_REQUEST);
if (!data) {
- kfree_skb(skb);
+ nlmsg_free(skb);
return -EMSGSIZE;
}
@@ -1820,7 +1820,7 @@
goto err3;
}
- if (ibnl_add_client(RDMA_NL_LS, RDMA_NL_LS_NUM_OPS,
+ if (ibnl_add_client(RDMA_NL_LS, ARRAY_SIZE(ib_sa_cb_table),
ib_sa_cb_table)) {
pr_err("Failed to add netlink callback\n");
ret = -EINVAL;
diff --git a/drivers/infiniband/core/uverbs_cmd.c b/drivers/infiniband/core/uverbs_cmd.c
index 6fdc7ec..03e39c2 100644
--- a/drivers/infiniband/core/uverbs_cmd.c
+++ b/drivers/infiniband/core/uverbs_cmd.c
@@ -1833,7 +1833,8 @@
if (attr.create_flags & ~(IB_QP_CREATE_BLOCK_MULTICAST_LOOPBACK |
IB_QP_CREATE_CROSS_CHANNEL |
IB_QP_CREATE_MANAGED_SEND |
- IB_QP_CREATE_MANAGED_RECV)) {
+ IB_QP_CREATE_MANAGED_RECV |
+ IB_QP_CREATE_SCATTER_FCS)) {
ret = -EINVAL;
goto err_put;
}
@@ -3655,6 +3656,11 @@
resp.hca_core_clock = attr.hca_core_clock;
resp.response_length += sizeof(resp.hca_core_clock);
+ if (ucore->outlen < resp.response_length + sizeof(resp.device_cap_flags_ex))
+ goto end;
+
+ resp.device_cap_flags_ex = attr.device_cap_flags;
+ resp.response_length += sizeof(resp.device_cap_flags_ex);
end:
err = ib_copy_to_udata(ucore, &resp, resp.response_length);
return err;
diff --git a/drivers/infiniband/hw/cxgb4/cm.c b/drivers/infiniband/hw/cxgb4/cm.c
index d7f7ab3..a3a6721 100644
--- a/drivers/infiniband/hw/cxgb4/cm.c
+++ b/drivers/infiniband/hw/cxgb4/cm.c
@@ -119,7 +119,7 @@
static int mpa_rev = 2;
module_param(mpa_rev, int, 0644);
MODULE_PARM_DESC(mpa_rev, "MPA Revision, 0 supports amso1100, "
- "1 is RFC0544 spec compliant, 2 is IETF MPA Peer Connect Draft"
+ "1 is RFC5044 spec compliant, 2 is IETF MPA Peer Connect Draft"
" compliant (default=2)");
static int markers_enabled;
@@ -150,15 +150,30 @@
static LIST_HEAD(timeout_list);
static spinlock_t timeout_lock;
+static void deref_cm_id(struct c4iw_ep_common *epc)
+{
+ epc->cm_id->rem_ref(epc->cm_id);
+ epc->cm_id = NULL;
+ set_bit(CM_ID_DEREFED, &epc->history);
+}
+
+static void ref_cm_id(struct c4iw_ep_common *epc)
+{
+ set_bit(CM_ID_REFED, &epc->history);
+ epc->cm_id->add_ref(epc->cm_id);
+}
+
static void deref_qp(struct c4iw_ep *ep)
{
c4iw_qp_rem_ref(&ep->com.qp->ibqp);
clear_bit(QP_REFERENCED, &ep->com.flags);
+ set_bit(QP_DEREFED, &ep->com.history);
}
static void ref_qp(struct c4iw_ep *ep)
{
set_bit(QP_REFERENCED, &ep->com.flags);
+ set_bit(QP_REFED, &ep->com.history);
c4iw_qp_add_ref(&ep->com.qp->ibqp);
}
@@ -202,6 +217,8 @@
error = cxgb4_l2t_send(rdev->lldi.ports[0], skb, l2e);
if (error < 0)
kfree_skb(skb);
+ else if (error == NET_XMIT_DROP)
+ return -ENOMEM;
return error < 0 ? error : 0;
}
@@ -291,6 +308,57 @@
return epc;
}
+static void remove_ep_tid(struct c4iw_ep *ep)
+{
+ unsigned long flags;
+
+ spin_lock_irqsave(&ep->com.dev->lock, flags);
+ _remove_handle(ep->com.dev, &ep->com.dev->hwtid_idr, ep->hwtid, 0);
+ spin_unlock_irqrestore(&ep->com.dev->lock, flags);
+}
+
+static void insert_ep_tid(struct c4iw_ep *ep)
+{
+ unsigned long flags;
+
+ spin_lock_irqsave(&ep->com.dev->lock, flags);
+ _insert_handle(ep->com.dev, &ep->com.dev->hwtid_idr, ep, ep->hwtid, 0);
+ spin_unlock_irqrestore(&ep->com.dev->lock, flags);
+}
+
+/*
+ * Atomically lookup the ep ptr given the tid and grab a reference on the ep.
+ */
+static struct c4iw_ep *get_ep_from_tid(struct c4iw_dev *dev, unsigned int tid)
+{
+ struct c4iw_ep *ep;
+ unsigned long flags;
+
+ spin_lock_irqsave(&dev->lock, flags);
+ ep = idr_find(&dev->hwtid_idr, tid);
+ if (ep)
+ c4iw_get_ep(&ep->com);
+ spin_unlock_irqrestore(&dev->lock, flags);
+ return ep;
+}
+
+/*
+ * Atomically lookup the ep ptr given the stid and grab a reference on the ep.
+ */
+static struct c4iw_listen_ep *get_ep_from_stid(struct c4iw_dev *dev,
+ unsigned int stid)
+{
+ struct c4iw_listen_ep *ep;
+ unsigned long flags;
+
+ spin_lock_irqsave(&dev->lock, flags);
+ ep = idr_find(&dev->stid_idr, stid);
+ if (ep)
+ c4iw_get_ep(&ep->com);
+ spin_unlock_irqrestore(&dev->lock, flags);
+ return ep;
+}
+
void _c4iw_free_ep(struct kref *kref)
{
struct c4iw_ep *ep;
@@ -310,10 +378,11 @@
(const u32 *)&sin6->sin6_addr.s6_addr,
1);
}
- remove_handle(ep->com.dev, &ep->com.dev->hwtid_idr, ep->hwtid);
cxgb4_remove_tid(ep->com.dev->rdev.lldi.tids, 0, ep->hwtid);
dst_release(ep->dst);
cxgb4_l2t_release(ep->l2t);
+ if (ep->mpa_skb)
+ kfree_skb(ep->mpa_skb);
}
kfree(ep);
}
@@ -321,6 +390,15 @@
static void release_ep_resources(struct c4iw_ep *ep)
{
set_bit(RELEASE_RESOURCES, &ep->com.flags);
+
+ /*
+ * If we have a hwtid, then remove it from the idr table
+ * so lookups will no longer find this endpoint. Otherwise
+ * we have a race where one thread finds the ep ptr just
+ * before the other thread is freeing the ep memory.
+ */
+ if (ep->hwtid != -1)
+ remove_ep_tid(ep);
c4iw_put_ep(&ep->com);
}
@@ -437,9 +515,15 @@
kfree_skb(skb);
}
+static void mpa_start_arp_failure(void *handle, struct sk_buff *skb)
+{
+ pr_err("ARP failure during MPA Negotiation - Closing Connection\n");
+}
+
enum {
- NUM_FAKE_CPLS = 1,
+ NUM_FAKE_CPLS = 2,
FAKE_CPL_PUT_EP_SAFE = NUM_CPL_CMDS + 0,
+ FAKE_CPL_PASS_PUT_EP_SAFE = NUM_CPL_CMDS + 1,
};
static int _put_ep_safe(struct c4iw_dev *dev, struct sk_buff *skb)
@@ -451,18 +535,29 @@
return 0;
}
+static int _put_pass_ep_safe(struct c4iw_dev *dev, struct sk_buff *skb)
+{
+ struct c4iw_ep *ep;
+
+ ep = *((struct c4iw_ep **)(skb->cb + 2 * sizeof(void *)));
+ c4iw_put_ep(&ep->parent_ep->com);
+ release_ep_resources(ep);
+ return 0;
+}
+
/*
* Fake up a special CPL opcode and call sched() so process_work() will call
* _put_ep_safe() in a safe context to free the ep resources. This is needed
* because ARP error handlers are called in an ATOMIC context, and
* _c4iw_free_ep() needs to block.
*/
-static void queue_arp_failure_cpl(struct c4iw_ep *ep, struct sk_buff *skb)
+static void queue_arp_failure_cpl(struct c4iw_ep *ep, struct sk_buff *skb,
+ int cpl)
{
struct cpl_act_establish *rpl = cplhdr(skb);
/* Set our special ARP_FAILURE opcode */
- rpl->ot.opcode = FAKE_CPL_PUT_EP_SAFE;
+ rpl->ot.opcode = cpl;
/*
* Save ep in the skb->cb area, after where sched() will save the dev
@@ -481,7 +576,7 @@
ep->hwtid);
__state_set(&ep->com, DEAD);
- queue_arp_failure_cpl(ep, skb);
+ queue_arp_failure_cpl(ep, skb, FAKE_CPL_PASS_PUT_EP_SAFE);
}
/*
@@ -502,7 +597,7 @@
}
remove_handle(ep->com.dev, &ep->com.dev->atid_idr, ep->atid);
cxgb4_free_atid(ep->com.dev->rdev.lldi.tids, ep->atid);
- queue_arp_failure_cpl(ep, skb);
+ queue_arp_failure_cpl(ep, skb, FAKE_CPL_PUT_EP_SAFE);
}
/*
@@ -511,12 +606,18 @@
*/
static void abort_arp_failure(void *handle, struct sk_buff *skb)
{
- struct c4iw_rdev *rdev = handle;
+ int ret;
+ struct c4iw_ep *ep = handle;
+ struct c4iw_rdev *rdev = &ep->com.dev->rdev;
struct cpl_abort_req *req = cplhdr(skb);
PDBG("%s rdev %p\n", __func__, rdev);
req->cmd = CPL_ABORT_NO_RST;
- c4iw_ofld_send(rdev, skb);
+ ret = c4iw_ofld_send(rdev, skb);
+ if (ret) {
+ __state_set(&ep->com, DEAD);
+ queue_arp_failure_cpl(ep, skb, FAKE_CPL_PUT_EP_SAFE);
+ }
}
static int send_flowc(struct c4iw_ep *ep, struct sk_buff *skb)
@@ -613,7 +714,7 @@
return -ENOMEM;
}
set_wr_txq(skb, CPL_PRIORITY_DATA, ep->txq_idx);
- t4_set_arp_err_handler(skb, &ep->com.dev->rdev, abort_arp_failure);
+ t4_set_arp_err_handler(skb, ep, abort_arp_failure);
req = (struct cpl_abort_req *) skb_put(skb, wrlen);
memset(req, 0, wrlen);
INIT_TP_WR(req, ep->hwtid);
@@ -852,10 +953,10 @@
return ret;
}
-static void send_mpa_req(struct c4iw_ep *ep, struct sk_buff *skb,
- u8 mpa_rev_to_use)
+static int send_mpa_req(struct c4iw_ep *ep, struct sk_buff *skb,
+ u8 mpa_rev_to_use)
{
- int mpalen, wrlen;
+ int mpalen, wrlen, ret;
struct fw_ofld_tx_data_wr *req;
struct mpa_message *mpa;
struct mpa_v2_conn_params mpa_v2_params;
@@ -871,7 +972,7 @@
skb = get_skb(skb, wrlen, GFP_KERNEL);
if (!skb) {
connect_reply_upcall(ep, -ENOMEM);
- return;
+ return -ENOMEM;
}
set_wr_txq(skb, CPL_PRIORITY_DATA, ep->txq_idx);
@@ -939,12 +1040,14 @@
t4_set_arp_err_handler(skb, NULL, arp_failure_discard);
BUG_ON(ep->mpa_skb);
ep->mpa_skb = skb;
- c4iw_l2t_send(&ep->com.dev->rdev, skb, ep->l2t);
+ ret = c4iw_l2t_send(&ep->com.dev->rdev, skb, ep->l2t);
+ if (ret)
+ return ret;
start_ep_timer(ep);
__state_set(&ep->com, MPA_REQ_SENT);
ep->mpa_attr.initiator = 1;
ep->snd_seq += mpalen;
- return;
+ return ret;
}
static int send_mpa_reject(struct c4iw_ep *ep, const void *pdata, u8 plen)
@@ -1020,7 +1123,7 @@
*/
skb_get(skb);
set_wr_txq(skb, CPL_PRIORITY_DATA, ep->txq_idx);
- t4_set_arp_err_handler(skb, NULL, arp_failure_discard);
+ t4_set_arp_err_handler(skb, NULL, mpa_start_arp_failure);
BUG_ON(ep->mpa_skb);
ep->mpa_skb = skb;
ep->snd_seq += mpalen;
@@ -1105,7 +1208,7 @@
* Function fw4_ack() will deref it.
*/
skb_get(skb);
- t4_set_arp_err_handler(skb, NULL, arp_failure_discard);
+ t4_set_arp_err_handler(skb, NULL, mpa_start_arp_failure);
ep->mpa_skb = skb;
__state_set(&ep->com, MPA_REP_SENT);
ep->snd_seq += mpalen;
@@ -1132,7 +1235,7 @@
/* setup the hwtid for this connection */
ep->hwtid = tid;
cxgb4_insert_tid(t, ep, tid);
- insert_handle(dev, &dev->hwtid_idr, ep, ep->hwtid);
+ insert_ep_tid(ep);
ep->snd_seq = be32_to_cpu(req->snd_isn);
ep->rcv_seq = be32_to_cpu(req->rcv_isn);
@@ -1149,9 +1252,11 @@
if (ret)
goto err;
if (ep->retry_with_mpa_v1)
- send_mpa_req(ep, skb, 1);
+ ret = send_mpa_req(ep, skb, 1);
else
- send_mpa_req(ep, skb, mpa_rev);
+ ret = send_mpa_req(ep, skb, mpa_rev);
+ if (ret)
+ goto err;
mutex_unlock(&ep->com.mutex);
return 0;
err:
@@ -1173,8 +1278,7 @@
PDBG("close complete delivered ep %p cm_id %p tid %u\n",
ep, ep->com.cm_id, ep->hwtid);
ep->com.cm_id->event_handler(ep->com.cm_id, &event);
- ep->com.cm_id->rem_ref(ep->com.cm_id);
- ep->com.cm_id = NULL;
+ deref_cm_id(&ep->com);
set_bit(CLOSE_UPCALL, &ep->com.history);
}
}
@@ -1206,8 +1310,7 @@
PDBG("abort delivered ep %p cm_id %p tid %u\n", ep,
ep->com.cm_id, ep->hwtid);
ep->com.cm_id->event_handler(ep->com.cm_id, &event);
- ep->com.cm_id->rem_ref(ep->com.cm_id);
- ep->com.cm_id = NULL;
+ deref_cm_id(&ep->com);
set_bit(ABORT_UPCALL, &ep->com.history);
}
}
@@ -1250,10 +1353,8 @@
set_bit(CONN_RPL_UPCALL, &ep->com.history);
ep->com.cm_id->event_handler(ep->com.cm_id, &event);
- if (status < 0) {
- ep->com.cm_id->rem_ref(ep->com.cm_id);
- ep->com.cm_id = NULL;
- }
+ if (status < 0)
+ deref_cm_id(&ep->com);
}
static int connect_request_upcall(struct c4iw_ep *ep)
@@ -1373,20 +1474,12 @@
PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
/*
- * Stop mpa timer. If it expired, then
- * we ignore the MPA reply. process_timeout()
- * will abort the connection.
- */
- if (stop_ep_timer(ep))
- return 0;
-
- /*
* If we get more than the supported amount of private data
* then we must fail this connection.
*/
if (ep->mpa_pkt_len + skb->len > sizeof(ep->mpa_pkt)) {
err = -EINVAL;
- goto err;
+ goto err_stop_timer;
}
/*
@@ -1408,11 +1501,11 @@
printk(KERN_ERR MOD "%s MPA version mismatch. Local = %d,"
" Received = %d\n", __func__, mpa_rev, mpa->revision);
err = -EPROTO;
- goto err;
+ goto err_stop_timer;
}
if (memcmp(mpa->key, MPA_KEY_REP, sizeof(mpa->key))) {
err = -EPROTO;
- goto err;
+ goto err_stop_timer;
}
plen = ntohs(mpa->private_data_size);
@@ -1422,7 +1515,7 @@
*/
if (plen > MPA_MAX_PRIVATE_DATA) {
err = -EPROTO;
- goto err;
+ goto err_stop_timer;
}
/*
@@ -1430,7 +1523,7 @@
*/
if (ep->mpa_pkt_len > (sizeof(*mpa) + plen)) {
err = -EPROTO;
- goto err;
+ goto err_stop_timer;
}
ep->plen = (u8) plen;
@@ -1444,10 +1537,18 @@
if (mpa->flags & MPA_REJECT) {
err = -ECONNREFUSED;
- goto err;
+ goto err_stop_timer;
}
/*
+ * Stop mpa timer. If it expired, then
+ * we ignore the MPA reply. process_timeout()
+ * will abort the connection.
+ */
+ if (stop_ep_timer(ep))
+ return 0;
+
+ /*
* If we get here we have accumulated the entire mpa
* start reply message including private data. And
* the MPA header is valid.
@@ -1586,6 +1687,8 @@
goto out;
}
goto out;
+err_stop_timer:
+ stop_ep_timer(ep);
err:
disconnect = 2;
out:
@@ -1719,25 +1822,17 @@
ep->mpa_attr.xmit_marker_enabled, ep->mpa_attr.version,
ep->mpa_attr.p2p_type);
- /*
- * If the endpoint timer already expired, then we ignore
- * the start request. process_timeout() will abort
- * the connection.
- */
- if (!stop_ep_timer(ep)) {
- __state_set(&ep->com, MPA_REQ_RCVD);
+ __state_set(&ep->com, MPA_REQ_RCVD);
- /* drive upcall */
- mutex_lock_nested(&ep->parent_ep->com.mutex,
- SINGLE_DEPTH_NESTING);
- if (ep->parent_ep->com.state != DEAD) {
- if (connect_request_upcall(ep))
- goto err_unlock_parent;
- } else {
+ /* drive upcall */
+ mutex_lock_nested(&ep->parent_ep->com.mutex, SINGLE_DEPTH_NESTING);
+ if (ep->parent_ep->com.state != DEAD) {
+ if (connect_request_upcall(ep))
goto err_unlock_parent;
- }
- mutex_unlock(&ep->parent_ep->com.mutex);
+ } else {
+ goto err_unlock_parent;
}
+ mutex_unlock(&ep->parent_ep->com.mutex);
return 0;
err_unlock_parent:
@@ -1755,11 +1850,10 @@
struct cpl_rx_data *hdr = cplhdr(skb);
unsigned int dlen = ntohs(hdr->len);
unsigned int tid = GET_TID(hdr);
- struct tid_info *t = dev->rdev.lldi.tids;
__u8 status = hdr->status;
int disconnect = 0;
- ep = lookup_tid(t, tid);
+ ep = get_ep_from_tid(dev, tid);
if (!ep)
return 0;
PDBG("%s ep %p tid %u dlen %u\n", __func__, ep, ep->hwtid, dlen);
@@ -1777,7 +1871,7 @@
break;
case MPA_REQ_WAIT:
ep->rcv_seq += dlen;
- process_mpa_request(ep, skb);
+ disconnect = process_mpa_request(ep, skb);
break;
case FPDU_MODE: {
struct c4iw_qp_attributes attrs;
@@ -1798,7 +1892,8 @@
}
mutex_unlock(&ep->com.mutex);
if (disconnect)
- c4iw_ep_disconnect(ep, 0, GFP_KERNEL);
+ c4iw_ep_disconnect(ep, disconnect == 2, GFP_KERNEL);
+ c4iw_put_ep(&ep->com);
return 0;
}
@@ -1808,9 +1903,8 @@
struct cpl_abort_rpl_rss *rpl = cplhdr(skb);
int release = 0;
unsigned int tid = GET_TID(rpl);
- struct tid_info *t = dev->rdev.lldi.tids;
- ep = lookup_tid(t, tid);
+ ep = get_ep_from_tid(dev, tid);
if (!ep) {
printk(KERN_WARNING MOD "Abort rpl to freed endpoint\n");
return 0;
@@ -1832,10 +1926,11 @@
if (release)
release_ep_resources(ep);
+ c4iw_put_ep(&ep->com);
return 0;
}
-static void send_fw_act_open_req(struct c4iw_ep *ep, unsigned int atid)
+static int send_fw_act_open_req(struct c4iw_ep *ep, unsigned int atid)
{
struct sk_buff *skb;
struct fw_ofld_connection_wr *req;
@@ -1905,7 +2000,7 @@
req->tcb.opt2 = cpu_to_be32((__force u32)req->tcb.opt2);
set_wr_txq(skb, CPL_PRIORITY_CONTROL, ep->ctrlq_idx);
set_bit(ACT_OFLD_CONN, &ep->com.history);
- c4iw_l2t_send(&ep->com.dev->rdev, skb, ep->l2t);
+ return c4iw_l2t_send(&ep->com.dev->rdev, skb, ep->l2t);
}
/*
@@ -2048,6 +2143,7 @@
PDBG("%s qp %p cm_id %p\n", __func__, ep->com.qp, ep->com.cm_id);
init_timer(&ep->timer);
+ c4iw_init_wr_wait(&ep->com.wr_wait);
/*
* Allocate an active TID to initiate a TCP connection.
@@ -2131,6 +2227,7 @@
struct sockaddr_in *ra;
struct sockaddr_in6 *la6;
struct sockaddr_in6 *ra6;
+ int ret = 0;
ep = lookup_atid(t, atid);
la = (struct sockaddr_in *)&ep->com.local_addr;
@@ -2166,9 +2263,10 @@
mutex_unlock(&dev->rdev.stats.lock);
if (ep->com.local_addr.ss_family == AF_INET &&
dev->rdev.lldi.enable_fw_ofld_conn) {
- send_fw_act_open_req(ep,
- TID_TID_G(AOPEN_ATID_G(
- ntohl(rpl->atid_status))));
+ ret = send_fw_act_open_req(ep, TID_TID_G(AOPEN_ATID_G(
+ ntohl(rpl->atid_status))));
+ if (ret)
+ goto fail;
return 0;
}
break;
@@ -2208,6 +2306,7 @@
break;
}
+fail:
connect_reply_upcall(ep, status2errno(status));
state_set(&ep->com, DEAD);
@@ -2232,9 +2331,8 @@
static int pass_open_rpl(struct c4iw_dev *dev, struct sk_buff *skb)
{
struct cpl_pass_open_rpl *rpl = cplhdr(skb);
- struct tid_info *t = dev->rdev.lldi.tids;
unsigned int stid = GET_TID(rpl);
- struct c4iw_listen_ep *ep = lookup_stid(t, stid);
+ struct c4iw_listen_ep *ep = get_ep_from_stid(dev, stid);
if (!ep) {
PDBG("%s stid %d lookup failure!\n", __func__, stid);
@@ -2243,7 +2341,7 @@
PDBG("%s ep %p status %d error %d\n", __func__, ep,
rpl->status, status2errno(rpl->status));
c4iw_wake_up(&ep->com.wr_wait, status2errno(rpl->status));
-
+ c4iw_put_ep(&ep->com);
out:
return 0;
}
@@ -2251,12 +2349,12 @@
static int close_listsrv_rpl(struct c4iw_dev *dev, struct sk_buff *skb)
{
struct cpl_close_listsvr_rpl *rpl = cplhdr(skb);
- struct tid_info *t = dev->rdev.lldi.tids;
unsigned int stid = GET_TID(rpl);
- struct c4iw_listen_ep *ep = lookup_stid(t, stid);
+ struct c4iw_listen_ep *ep = get_ep_from_stid(dev, stid);
PDBG("%s ep %p\n", __func__, ep);
c4iw_wake_up(&ep->com.wr_wait, status2errno(rpl->status));
+ c4iw_put_ep(&ep->com);
return 0;
}
@@ -2416,7 +2514,7 @@
unsigned short hdrs;
u8 tos = PASS_OPEN_TOS_G(ntohl(req->tos_stid));
- parent_ep = lookup_stid(t, stid);
+ parent_ep = (struct c4iw_ep *)get_ep_from_stid(dev, stid);
if (!parent_ep) {
PDBG("%s connect request on invalid stid %d\n", __func__, stid);
goto reject;
@@ -2529,7 +2627,7 @@
init_timer(&child_ep->timer);
cxgb4_insert_tid(t, child_ep, hwtid);
- insert_handle(dev, &dev->hwtid_idr, child_ep, child_ep->hwtid);
+ insert_ep_tid(child_ep);
if (accept_cr(child_ep, skb, req)) {
c4iw_put_ep(&parent_ep->com);
release_ep_resources(child_ep);
@@ -2544,6 +2642,8 @@
goto out;
reject:
reject_cr(dev, hwtid, skb);
+ if (parent_ep)
+ c4iw_put_ep(&parent_ep->com);
out:
return 0;
}
@@ -2552,11 +2652,10 @@
{
struct c4iw_ep *ep;
struct cpl_pass_establish *req = cplhdr(skb);
- struct tid_info *t = dev->rdev.lldi.tids;
unsigned int tid = GET_TID(req);
int ret;
- ep = lookup_tid(t, tid);
+ ep = get_ep_from_tid(dev, tid);
PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
ep->snd_seq = be32_to_cpu(req->snd_isn);
ep->rcv_seq = be32_to_cpu(req->rcv_isn);
@@ -2575,6 +2674,7 @@
mutex_unlock(&ep->com.mutex);
if (ret)
c4iw_ep_disconnect(ep, 1, GFP_KERNEL);
+ c4iw_put_ep(&ep->com);
return 0;
}
@@ -2586,11 +2686,13 @@
struct c4iw_qp_attributes attrs;
int disconnect = 1;
int release = 0;
- struct tid_info *t = dev->rdev.lldi.tids;
unsigned int tid = GET_TID(hdr);
int ret;
- ep = lookup_tid(t, tid);
+ ep = get_ep_from_tid(dev, tid);
+ if (!ep)
+ return 0;
+
PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
dst_confirm(ep->dst);
@@ -2662,6 +2764,7 @@
c4iw_ep_disconnect(ep, 0, GFP_KERNEL);
if (release)
release_ep_resources(ep);
+ c4iw_put_ep(&ep->com);
return 0;
}
@@ -2674,10 +2777,12 @@
struct c4iw_qp_attributes attrs;
int ret;
int release = 0;
- struct tid_info *t = dev->rdev.lldi.tids;
unsigned int tid = GET_TID(req);
- ep = lookup_tid(t, tid);
+ ep = get_ep_from_tid(dev, tid);
+ if (!ep)
+ return 0;
+
if (is_neg_adv(req->status)) {
PDBG("%s Negative advice on abort- tid %u status %d (%s)\n",
__func__, ep->hwtid, req->status,
@@ -2686,7 +2791,7 @@
mutex_lock(&dev->rdev.stats.lock);
dev->rdev.stats.neg_adv++;
mutex_unlock(&dev->rdev.stats.lock);
- return 0;
+ goto deref_ep;
}
PDBG("%s ep %p tid %u state %u\n", __func__, ep, ep->hwtid,
ep->com.state);
@@ -2752,7 +2857,7 @@
case DEAD:
PDBG("%s PEER_ABORT IN DEAD STATE!!!!\n", __func__);
mutex_unlock(&ep->com.mutex);
- return 0;
+ goto deref_ep;
default:
BUG_ON(1);
break;
@@ -2799,6 +2904,10 @@
c4iw_reconnect(ep);
}
+deref_ep:
+ c4iw_put_ep(&ep->com);
+ /* Dereferencing ep, referenced in peer_abort_intr() */
+ c4iw_put_ep(&ep->com);
return 0;
}
@@ -2808,16 +2917,18 @@
struct c4iw_qp_attributes attrs;
struct cpl_close_con_rpl *rpl = cplhdr(skb);
int release = 0;
- struct tid_info *t = dev->rdev.lldi.tids;
unsigned int tid = GET_TID(rpl);
- ep = lookup_tid(t, tid);
+ ep = get_ep_from_tid(dev, tid);
+ if (!ep)
+ return 0;
PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
BUG_ON(!ep);
/* The cm_id may be null if we failed to connect */
mutex_lock(&ep->com.mutex);
+ set_bit(CLOSE_CON_RPL, &ep->com.history);
switch (ep->com.state) {
case CLOSING:
__state_set(&ep->com, MORIBUND);
@@ -2845,18 +2956,18 @@
mutex_unlock(&ep->com.mutex);
if (release)
release_ep_resources(ep);
+ c4iw_put_ep(&ep->com);
return 0;
}
static int terminate(struct c4iw_dev *dev, struct sk_buff *skb)
{
struct cpl_rdma_terminate *rpl = cplhdr(skb);
- struct tid_info *t = dev->rdev.lldi.tids;
unsigned int tid = GET_TID(rpl);
struct c4iw_ep *ep;
struct c4iw_qp_attributes attrs;
- ep = lookup_tid(t, tid);
+ ep = get_ep_from_tid(dev, tid);
BUG_ON(!ep);
if (ep && ep->com.qp) {
@@ -2867,6 +2978,7 @@
C4IW_QP_ATTR_NEXT_STATE, &attrs, 1);
} else
printk(KERN_WARNING MOD "TERM received tid %u no ep/qp\n", tid);
+ c4iw_put_ep(&ep->com);
return 0;
}
@@ -2882,15 +2994,16 @@
struct cpl_fw4_ack *hdr = cplhdr(skb);
u8 credits = hdr->credits;
unsigned int tid = GET_TID(hdr);
- struct tid_info *t = dev->rdev.lldi.tids;
- ep = lookup_tid(t, tid);
+ ep = get_ep_from_tid(dev, tid);
+ if (!ep)
+ return 0;
PDBG("%s ep %p tid %u credits %u\n", __func__, ep, ep->hwtid, credits);
if (credits == 0) {
PDBG("%s 0 credit ack ep %p tid %u state %u\n",
__func__, ep, ep->hwtid, state_read(&ep->com));
- return 0;
+ goto out;
}
dst_confirm(ep->dst);
@@ -2900,7 +3013,13 @@
state_read(&ep->com), ep->mpa_attr.initiator ? 1 : 0);
kfree_skb(ep->mpa_skb);
ep->mpa_skb = NULL;
+ mutex_lock(&ep->com.mutex);
+ if (test_bit(STOP_MPA_TIMER, &ep->com.flags))
+ stop_ep_timer(ep);
+ mutex_unlock(&ep->com.mutex);
}
+out:
+ c4iw_put_ep(&ep->com);
return 0;
}
@@ -2912,13 +3031,12 @@
PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
mutex_lock(&ep->com.mutex);
- if (ep->com.state == DEAD) {
+ if (ep->com.state != MPA_REQ_RCVD) {
mutex_unlock(&ep->com.mutex);
c4iw_put_ep(&ep->com);
return -ECONNRESET;
}
set_bit(ULP_REJECT, &ep->com.history);
- BUG_ON(ep->com.state != MPA_REQ_RCVD);
if (mpa_rev == 0)
disconnect = 2;
else {
@@ -2926,8 +3044,10 @@
disconnect = 1;
}
mutex_unlock(&ep->com.mutex);
- if (disconnect)
+ if (disconnect) {
+ stop_ep_timer(ep);
err = c4iw_ep_disconnect(ep, disconnect == 2, GFP_KERNEL);
+ }
c4iw_put_ep(&ep->com);
return 0;
}
@@ -2945,12 +3065,11 @@
PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
mutex_lock(&ep->com.mutex);
- if (ep->com.state == DEAD) {
+ if (ep->com.state != MPA_REQ_RCVD) {
err = -ECONNRESET;
goto err_out;
}
- BUG_ON(ep->com.state != MPA_REQ_RCVD);
BUG_ON(!qp);
set_bit(ULP_ACCEPT, &ep->com.history);
@@ -2998,8 +3117,8 @@
PDBG("%s %d ird %d ord %d\n", __func__, __LINE__, ep->ird, ep->ord);
- cm_id->add_ref(cm_id);
ep->com.cm_id = cm_id;
+ ref_cm_id(&ep->com);
ep->com.qp = qp;
ref_qp(ep);
@@ -3021,6 +3140,8 @@
ep->com.qp, mask, &attrs, 1);
if (err)
goto err_deref_cm_id;
+
+ set_bit(STOP_MPA_TIMER, &ep->com.flags);
err = send_mpa_reply(ep, conn_param->private_data,
conn_param->private_data_len);
if (err)
@@ -3032,8 +3153,7 @@
c4iw_put_ep(&ep->com);
return 0;
err_deref_cm_id:
- ep->com.cm_id = NULL;
- cm_id->rem_ref(cm_id);
+ deref_cm_id(&ep->com);
err_abort:
abort = 1;
err_out:
@@ -3139,9 +3259,9 @@
if (peer2peer && ep->ord == 0)
ep->ord = 1;
- cm_id->add_ref(cm_id);
- ep->com.dev = dev;
ep->com.cm_id = cm_id;
+ ref_cm_id(&ep->com);
+ ep->com.dev = dev;
ep->com.qp = get_qhp(dev, conn_param->qpn);
if (!ep->com.qp) {
PDBG("%s qpn 0x%x not found!\n", __func__, conn_param->qpn);
@@ -3180,7 +3300,7 @@
/*
* Handle loopback requests to INADDR_ANY.
*/
- if ((__force int)raddr->sin_addr.s_addr == INADDR_ANY) {
+ if (raddr->sin_addr.s_addr == htonl(INADDR_ANY)) {
err = pick_local_ipaddrs(dev, cm_id);
if (err)
goto fail1;
@@ -3248,7 +3368,7 @@
remove_handle(ep->com.dev, &ep->com.dev->atid_idr, ep->atid);
cxgb4_free_atid(ep->com.dev->rdev.lldi.tids, ep->atid);
fail1:
- cm_id->rem_ref(cm_id);
+ deref_cm_id(&ep->com);
c4iw_put_ep(&ep->com);
out:
return err;
@@ -3342,8 +3462,8 @@
goto fail1;
}
PDBG("%s ep %p\n", __func__, ep);
- cm_id->add_ref(cm_id);
ep->com.cm_id = cm_id;
+ ref_cm_id(&ep->com);
ep->com.dev = dev;
ep->backlog = backlog;
memcpy(&ep->com.local_addr, &cm_id->m_local_addr,
@@ -3383,7 +3503,7 @@
cxgb4_free_stid(ep->com.dev->rdev.lldi.tids, ep->stid,
ep->com.local_addr.ss_family);
fail2:
- cm_id->rem_ref(cm_id);
+ deref_cm_id(&ep->com);
c4iw_put_ep(&ep->com);
fail1:
out:
@@ -3422,7 +3542,7 @@
cxgb4_free_stid(ep->com.dev->rdev.lldi.tids, ep->stid,
ep->com.local_addr.ss_family);
done:
- cm_id->rem_ref(cm_id);
+ deref_cm_id(&ep->com);
c4iw_put_ep(&ep->com);
return err;
}
@@ -3497,6 +3617,7 @@
ret = send_halfclose(ep, gfp);
}
if (ret) {
+ set_bit(EP_DISC_FAIL, &ep->com.history);
if (!abrupt) {
stop_ep_timer(ep);
close_complete_upcall(ep, -EIO);
@@ -3773,7 +3894,7 @@
struct cpl_pass_accept_req *req = (void *)(rss + 1);
struct l2t_entry *e;
struct dst_entry *dst;
- struct c4iw_ep *lep;
+ struct c4iw_ep *lep = NULL;
u16 window;
struct port_info *pi;
struct net_device *pdev;
@@ -3798,7 +3919,7 @@
*/
stid = (__force int) cpu_to_be32((__force u32) rss->hash_val);
- lep = (struct c4iw_ep *)lookup_stid(dev->rdev.lldi.tids, stid);
+ lep = (struct c4iw_ep *)get_ep_from_stid(dev, stid);
if (!lep) {
PDBG("%s connect request on invalid stid %d\n", __func__, stid);
goto reject;
@@ -3899,6 +4020,8 @@
free_dst:
dst_release(dst);
reject:
+ if (lep)
+ c4iw_put_ep(&lep->com);
return 0;
}
@@ -3923,7 +4046,8 @@
[CPL_FW4_ACK] = fw4_ack,
[CPL_FW6_MSG] = deferred_fw6_msg,
[CPL_RX_PKT] = rx_pkt,
- [FAKE_CPL_PUT_EP_SAFE] = _put_ep_safe
+ [FAKE_CPL_PUT_EP_SAFE] = _put_ep_safe,
+ [FAKE_CPL_PASS_PUT_EP_SAFE] = _put_pass_ep_safe
};
static void process_timeout(struct c4iw_ep *ep)
@@ -3937,11 +4061,12 @@
set_bit(TIMEDOUT, &ep->com.history);
switch (ep->com.state) {
case MPA_REQ_SENT:
- __state_set(&ep->com, ABORTING);
connect_reply_upcall(ep, -ETIMEDOUT);
break;
case MPA_REQ_WAIT:
- __state_set(&ep->com, ABORTING);
+ case MPA_REQ_RCVD:
+ case MPA_REP_SENT:
+ case FPDU_MODE:
break;
case CLOSING:
case MORIBUND:
@@ -3951,7 +4076,6 @@
ep->com.qp, C4IW_QP_ATTR_NEXT_STATE,
&attrs, 1);
}
- __state_set(&ep->com, ABORTING);
close_complete_upcall(ep, -ETIMEDOUT);
break;
case ABORTING:
@@ -4104,10 +4228,10 @@
{
struct cpl_abort_req_rss *req = cplhdr(skb);
struct c4iw_ep *ep;
- struct tid_info *t = dev->rdev.lldi.tids;
unsigned int tid = GET_TID(req);
- ep = lookup_tid(t, tid);
+ ep = get_ep_from_tid(dev, tid);
+ /* This EP will be dereferenced in peer_abort() */
if (!ep) {
printk(KERN_WARNING MOD
"Abort on non-existent endpoint, tid %d\n", tid);
@@ -4118,24 +4242,13 @@
PDBG("%s Negative advice on abort- tid %u status %d (%s)\n",
__func__, ep->hwtid, req->status,
neg_adv_str(req->status));
- ep->stats.abort_neg_adv++;
- dev->rdev.stats.neg_adv++;
- kfree_skb(skb);
- return 0;
+ goto out;
}
PDBG("%s ep %p tid %u state %u\n", __func__, ep, ep->hwtid,
ep->com.state);
- /*
- * Wake up any threads in rdma_init() or rdma_fini().
- * However, if we are on MPAv2 and want to retry with MPAv1
- * then, don't wake up yet.
- */
- if (mpa_rev == 2 && !ep->tried_with_mpa_v1) {
- if (ep->com.state != MPA_REQ_SENT)
- c4iw_wake_up(&ep->com.wr_wait, -ECONNRESET);
- } else
- c4iw_wake_up(&ep->com.wr_wait, -ECONNRESET);
+ c4iw_wake_up(&ep->com.wr_wait, -ECONNRESET);
+out:
sched(dev, skb);
return 0;
}
diff --git a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
index 1ff3ba8..f6f34a7 100644
--- a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
+++ b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
@@ -755,6 +755,7 @@
CLOSE_SENT = 3,
TIMEOUT = 4,
QP_REFERENCED = 5,
+ STOP_MPA_TIMER = 7,
};
enum c4iw_ep_history {
@@ -779,7 +780,13 @@
EP_DISC_ABORT = 18,
CONN_RPL_UPCALL = 19,
ACT_RETRY_NOMEM = 20,
- ACT_RETRY_INUSE = 21
+ ACT_RETRY_INUSE = 21,
+ CLOSE_CON_RPL = 22,
+ EP_DISC_FAIL = 24,
+ QP_REFED = 25,
+ QP_DEREFED = 26,
+ CM_ID_REFED = 27,
+ CM_ID_DEREFED = 28,
};
struct c4iw_ep_common {
diff --git a/drivers/infiniband/hw/cxgb4/mem.c b/drivers/infiniband/hw/cxgb4/mem.c
index 83960df..55d0651 100644
--- a/drivers/infiniband/hw/cxgb4/mem.c
+++ b/drivers/infiniband/hw/cxgb4/mem.c
@@ -86,8 +86,9 @@
(wait ? FW_WR_COMPL_F : 0));
req->wr.wr_lo = wait ? (__force __be64)(unsigned long) &wr_wait : 0L;
req->wr.wr_mid = cpu_to_be32(FW_WR_LEN16_V(DIV_ROUND_UP(wr_len, 16)));
- req->cmd = cpu_to_be32(ULPTX_CMD_V(ULP_TX_MEM_WRITE));
- req->cmd |= cpu_to_be32(T5_ULP_MEMIO_ORDER_V(1));
+ req->cmd = cpu_to_be32(ULPTX_CMD_V(ULP_TX_MEM_WRITE) |
+ T5_ULP_MEMIO_ORDER_V(1) |
+ T5_ULP_MEMIO_FID_V(rdev->lldi.rxq_ids[0]));
req->dlen = cpu_to_be32(ULP_MEMIO_DATA_LEN_V(len>>5));
req->len16 = cpu_to_be32(DIV_ROUND_UP(wr_len-sizeof(req->wr), 16));
req->lock_addr = cpu_to_be32(ULP_MEMIO_ADDR_V(addr));
diff --git a/drivers/infiniband/hw/mlx4/main.c b/drivers/infiniband/hw/mlx4/main.c
index f014eaf..b01ef6e 100644
--- a/drivers/infiniband/hw/mlx4/main.c
+++ b/drivers/infiniband/hw/mlx4/main.c
@@ -1601,7 +1601,7 @@
else if (ret == -ENXIO)
pr_err("Device managed flow steering is disabled. Fail to register network rule.\n");
else if (ret)
- pr_err("Invalid argumant. Fail to register network rule.\n");
+ pr_err("Invalid argument. Fail to register network rule.\n");
mlx4_free_cmd_mailbox(mdev->dev, mailbox);
return ret;
diff --git a/drivers/infiniband/hw/mlx4/mcg.c b/drivers/infiniband/hw/mlx4/mcg.c
index 99451d8..ebdca2b 100644
--- a/drivers/infiniband/hw/mlx4/mcg.c
+++ b/drivers/infiniband/hw/mlx4/mcg.c
@@ -747,14 +747,11 @@
__be64 tid,
union ib_gid *new_mgid)
{
- struct mcast_group *group = NULL, *cur_group;
+ struct mcast_group *group = NULL, *cur_group, *n;
struct mcast_req *req;
- struct list_head *pos;
- struct list_head *n;
mutex_lock(&ctx->mcg_table_lock);
- list_for_each_safe(pos, n, &ctx->mcg_mgid0_list) {
- group = list_entry(pos, struct mcast_group, mgid0_list);
+ list_for_each_entry_safe(group, n, &ctx->mcg_mgid0_list, mgid0_list) {
mutex_lock(&group->lock);
if (group->last_req_tid == tid) {
if (memcmp(new_mgid, &mgid0, sizeof mgid0)) {
diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c
index 6ad0489..1b04df7 100644
--- a/drivers/infiniband/hw/mlx5/main.c
+++ b/drivers/infiniband/hw/mlx5/main.c
@@ -38,6 +38,9 @@
#include <linux/dma-mapping.h>
#include <linux/slab.h>
#include <linux/io-mapping.h>
+#if defined(CONFIG_X86)
+#include <asm/pat.h>
+#endif
#include <linux/sched.h>
#include <rdma/ib_user_verbs.h>
#include <rdma/ib_addr.h>
@@ -517,6 +520,10 @@
props->device_cap_flags |= IB_DEVICE_UD_TSO;
}
+ if (MLX5_CAP_GEN(dev->mdev, eth_net_offloads) &&
+ MLX5_CAP_ETH(dev->mdev, scatter_fcs))
+ props->device_cap_flags |= IB_DEVICE_RAW_SCATTER_FCS;
+
props->vendor_part_id = mdev->pdev->device;
props->hw_ver = mdev->pdev->revision;
@@ -1068,38 +1075,89 @@
return get_arg(offset);
}
+static inline char *mmap_cmd2str(enum mlx5_ib_mmap_cmd cmd)
+{
+ switch (cmd) {
+ case MLX5_IB_MMAP_WC_PAGE:
+ return "WC";
+ case MLX5_IB_MMAP_REGULAR_PAGE:
+ return "best effort WC";
+ case MLX5_IB_MMAP_NC_PAGE:
+ return "NC";
+ default:
+ return NULL;
+ }
+}
+
+static int uar_mmap(struct mlx5_ib_dev *dev, enum mlx5_ib_mmap_cmd cmd,
+ struct vm_area_struct *vma, struct mlx5_uuar_info *uuari)
+{
+ int err;
+ unsigned long idx;
+ phys_addr_t pfn, pa;
+ pgprot_t prot;
+
+ switch (cmd) {
+ case MLX5_IB_MMAP_WC_PAGE:
+/* Some architectures don't support WC memory */
+#if defined(CONFIG_X86)
+ if (!pat_enabled())
+ return -EPERM;
+#elif !(defined(CONFIG_PPC) || (defined(CONFIG_ARM) && defined(CONFIG_MMU)))
+ return -EPERM;
+#endif
+ /* fall through */
+ case MLX5_IB_MMAP_REGULAR_PAGE:
+ /* For MLX5_IB_MMAP_REGULAR_PAGE do the best effort to get WC */
+ prot = pgprot_writecombine(vma->vm_page_prot);
+ break;
+ case MLX5_IB_MMAP_NC_PAGE:
+ prot = pgprot_noncached(vma->vm_page_prot);
+ break;
+ default:
+ return -EINVAL;
+ }
+
+ if (vma->vm_end - vma->vm_start != PAGE_SIZE)
+ return -EINVAL;
+
+ idx = get_index(vma->vm_pgoff);
+ if (idx >= uuari->num_uars)
+ return -EINVAL;
+
+ pfn = uar_index2pfn(dev, uuari->uars[idx].index);
+ mlx5_ib_dbg(dev, "uar idx 0x%lx, pfn %pa\n", idx, &pfn);
+
+ vma->vm_page_prot = prot;
+ err = io_remap_pfn_range(vma, vma->vm_start, pfn,
+ PAGE_SIZE, vma->vm_page_prot);
+ if (err) {
+ mlx5_ib_err(dev, "io_remap_pfn_range failed with error=%d, vm_start=0x%lx, pfn=%pa, mmap_cmd=%s\n",
+ err, vma->vm_start, &pfn, mmap_cmd2str(cmd));
+ return -EAGAIN;
+ }
+
+ pa = pfn << PAGE_SHIFT;
+ mlx5_ib_dbg(dev, "mapped %s at 0x%lx, PA %pa\n", mmap_cmd2str(cmd),
+ vma->vm_start, &pa);
+
+ return 0;
+}
+
static int mlx5_ib_mmap(struct ib_ucontext *ibcontext, struct vm_area_struct *vma)
{
struct mlx5_ib_ucontext *context = to_mucontext(ibcontext);
struct mlx5_ib_dev *dev = to_mdev(ibcontext->device);
struct mlx5_uuar_info *uuari = &context->uuari;
unsigned long command;
- unsigned long idx;
phys_addr_t pfn;
command = get_command(vma->vm_pgoff);
switch (command) {
+ case MLX5_IB_MMAP_WC_PAGE:
+ case MLX5_IB_MMAP_NC_PAGE:
case MLX5_IB_MMAP_REGULAR_PAGE:
- if (vma->vm_end - vma->vm_start != PAGE_SIZE)
- return -EINVAL;
-
- idx = get_index(vma->vm_pgoff);
- if (idx >= uuari->num_uars)
- return -EINVAL;
-
- pfn = uar_index2pfn(dev, uuari->uars[idx].index);
- mlx5_ib_dbg(dev, "uar idx 0x%lx, pfn 0x%llx\n", idx,
- (unsigned long long)pfn);
-
- vma->vm_page_prot = pgprot_writecombine(vma->vm_page_prot);
- if (io_remap_pfn_range(vma, vma->vm_start, pfn,
- PAGE_SIZE, vma->vm_page_prot))
- return -EAGAIN;
-
- mlx5_ib_dbg(dev, "mapped WC at 0x%lx, PA 0x%llx\n",
- vma->vm_start,
- (unsigned long long)pfn << PAGE_SHIFT);
- break;
+ return uar_mmap(dev, command, vma, uuari);
case MLX5_IB_MMAP_GET_CONTIGUOUS_PAGES:
return -ENOSYS;
@@ -1108,7 +1166,7 @@
if (vma->vm_end - vma->vm_start != PAGE_SIZE)
return -EINVAL;
- if (vma->vm_flags & (VM_WRITE | VM_EXEC))
+ if (vma->vm_flags & VM_WRITE)
return -EPERM;
/* Don't expose to user-space information it shouldn't have */
diff --git a/drivers/infiniband/hw/mlx5/mlx5_ib.h b/drivers/infiniband/hw/mlx5/mlx5_ib.h
index f05cf57..c4a9825 100644
--- a/drivers/infiniband/hw/mlx5/mlx5_ib.h
+++ b/drivers/infiniband/hw/mlx5/mlx5_ib.h
@@ -70,6 +70,8 @@
enum mlx5_ib_mmap_cmd {
MLX5_IB_MMAP_REGULAR_PAGE = 0,
MLX5_IB_MMAP_GET_CONTIGUOUS_PAGES = 1,
+ MLX5_IB_MMAP_WC_PAGE = 2,
+ MLX5_IB_MMAP_NC_PAGE = 3,
/* 5 is chosen in order to be compatible with old versions of libmlx5 */
MLX5_IB_MMAP_CORE_CLOCK = 5,
};
@@ -356,6 +358,7 @@
MLX5_IB_QP_SIGNATURE_HANDLING = 1 << 5,
/* QP uses 1 as its source QP number */
MLX5_IB_QP_SQPN_QP1 = 1 << 6,
+ MLX5_IB_QP_CAP_SCATTER_FCS = 1 << 7,
};
struct mlx5_umr_wr {
diff --git a/drivers/infiniband/hw/mlx5/qp.c b/drivers/infiniband/hw/mlx5/qp.c
index 8dee8bc..5041176 100644
--- a/drivers/infiniband/hw/mlx5/qp.c
+++ b/drivers/infiniband/hw/mlx5/qp.c
@@ -1028,6 +1028,7 @@
static int create_raw_packet_qp_rq(struct mlx5_ib_dev *dev,
struct mlx5_ib_rq *rq, void *qpin)
{
+ struct mlx5_ib_qp *mqp = rq->base.container_mibqp;
__be64 *pas;
__be64 *qp_pas;
void *in;
@@ -1051,6 +1052,9 @@
MLX5_SET(rqc, rqc, user_index, MLX5_GET(qpc, qpc, user_index));
MLX5_SET(rqc, rqc, cqn, MLX5_GET(qpc, qpc, cqn_rcv));
+ if (mqp->flags & MLX5_IB_QP_CAP_SCATTER_FCS)
+ MLX5_SET(rqc, rqc, scatter_fcs, 1);
+
wq = MLX5_ADDR_OF(rqc, rqc, wq);
MLX5_SET(wq, wq, wq_type, MLX5_WQ_TYPE_CYCLIC);
MLX5_SET(wq, wq, end_padding_mode,
@@ -1136,11 +1140,12 @@
}
if (qp->rq.wqe_cnt) {
+ rq->base.container_mibqp = qp;
+
err = create_raw_packet_qp_rq(dev, rq, in);
if (err)
goto err_destroy_sq;
- rq->base.container_mibqp = qp;
err = create_raw_packet_qp_tir(dev, rq, tdn);
if (err)
@@ -1252,6 +1257,19 @@
return -EOPNOTSUPP;
}
+ if (init_attr->create_flags & IB_QP_CREATE_SCATTER_FCS) {
+ if (init_attr->qp_type != IB_QPT_RAW_PACKET) {
+ mlx5_ib_dbg(dev, "Scatter FCS is supported only for Raw Packet QPs");
+ return -EOPNOTSUPP;
+ }
+ if (!MLX5_CAP_GEN(dev->mdev, eth_net_offloads) ||
+ !MLX5_CAP_ETH(dev->mdev, scatter_fcs)) {
+ mlx5_ib_dbg(dev, "Scatter FCS isn't supported\n");
+ return -EOPNOTSUPP;
+ }
+ qp->flags |= MLX5_IB_QP_CAP_SCATTER_FCS;
+ }
+
if (init_attr->sq_sig_type == IB_SIGNAL_ALL_WR)
qp->sq_signal_bits = MLX5_WQE_CTRL_CQ_UPDATE;
diff --git a/drivers/infiniband/hw/nes/nes_utils.c b/drivers/infiniband/hw/nes/nes_utils.c
index 6d3a169..37331e2 100644
--- a/drivers/infiniband/hw/nes/nes_utils.c
+++ b/drivers/infiniband/hw/nes/nes_utils.c
@@ -44,6 +44,7 @@
#include <linux/ip.h>
#include <linux/tcp.h>
#include <linux/init.h>
+#include <linux/kernel.h>
#include <asm/io.h>
#include <asm/irq.h>
@@ -903,70 +904,15 @@
*/
void nes_dump_mem(unsigned int dump_debug_level, void *addr, int length)
{
- char xlate[] = {'0', '1', '2', '3', '4', '5', '6', '7', '8', '9',
- 'a', 'b', 'c', 'd', 'e', 'f'};
- char *ptr;
- char hex_buf[80];
- char ascii_buf[20];
- int num_char;
- int num_ascii;
- int num_hex;
-
if (!(nes_debug_level & dump_debug_level)) {
return;
}
- ptr = addr;
if (length > 0x100) {
nes_debug(dump_debug_level, "Length truncated from %x to %x\n", length, 0x100);
length = 0x100;
}
- nes_debug(dump_debug_level, "Address=0x%p, length=0x%x (%d)\n", ptr, length, length);
+ nes_debug(dump_debug_level, "Address=0x%p, length=0x%x (%d)\n", addr, length, length);
- memset(ascii_buf, 0, 20);
- memset(hex_buf, 0, 80);
-
- num_ascii = 0;
- num_hex = 0;
- for (num_char = 0; num_char < length; num_char++) {
- if (num_ascii == 8) {
- ascii_buf[num_ascii++] = ' ';
- hex_buf[num_hex++] = '-';
- hex_buf[num_hex++] = ' ';
- }
-
- if (*ptr < 0x20 || *ptr > 0x7e)
- ascii_buf[num_ascii++] = '.';
- else
- ascii_buf[num_ascii++] = *ptr;
- hex_buf[num_hex++] = xlate[((*ptr & 0xf0) >> 4)];
- hex_buf[num_hex++] = xlate[*ptr & 0x0f];
- hex_buf[num_hex++] = ' ';
- ptr++;
-
- if (num_ascii >= 17) {
- /* output line and reset */
- nes_debug(dump_debug_level, " %s | %s\n", hex_buf, ascii_buf);
- memset(ascii_buf, 0, 20);
- memset(hex_buf, 0, 80);
- num_ascii = 0;
- num_hex = 0;
- }
- }
-
- /* output the rest */
- if (num_ascii) {
- while (num_ascii < 17) {
- if (num_ascii == 8) {
- hex_buf[num_hex++] = ' ';
- hex_buf[num_hex++] = ' ';
- }
- hex_buf[num_hex++] = ' ';
- hex_buf[num_hex++] = ' ';
- hex_buf[num_hex++] = ' ';
- num_ascii++;
- }
-
- nes_debug(dump_debug_level, " %s | %s\n", hex_buf, ascii_buf);
- }
+ print_hex_dump(KERN_ERR, PFX, DUMP_PREFIX_NONE, 16, 1, addr, length, true);
}
diff --git a/drivers/infiniband/hw/nes/nes_verbs.c b/drivers/infiniband/hw/nes/nes_verbs.c
index 4ebea4c..464d6da 100644
--- a/drivers/infiniband/hw/nes/nes_verbs.c
+++ b/drivers/infiniband/hw/nes/nes_verbs.c
@@ -980,7 +980,7 @@
/**
* nes_free_qp_mem() is to free up the qp's pci_alloc_consistent() memory.
*/
-static inline void nes_free_qp_mem(struct nes_device *nesdev,
+static void nes_free_qp_mem(struct nes_device *nesdev,
struct nes_qp *nesqp, int virt_wqs)
{
unsigned long flags;
@@ -1314,6 +1314,8 @@
nes_debug(NES_DBG_QP, "Invalid QP type: %d\n", init_attr->qp_type);
return ERR_PTR(-EINVAL);
}
+ init_completion(&nesqp->sq_drained);
+ init_completion(&nesqp->rq_drained);
nesqp->sig_all = (init_attr->sq_sig_type == IB_SIGNAL_ALL_WR);
init_timer(&nesqp->terminate_timer);
@@ -3451,6 +3453,29 @@
return err;
}
+/**
+ * nes_drain_sq - drain sq
+ * @ibqp: pointer to ibqp
+ */
+static void nes_drain_sq(struct ib_qp *ibqp)
+{
+ struct nes_qp *nesqp = to_nesqp(ibqp);
+
+ if (nesqp->hwqp.sq_tail != nesqp->hwqp.sq_head)
+ wait_for_completion(&nesqp->sq_drained);
+}
+
+/**
+ * nes_drain_rq - drain rq
+ * @ibqp: pointer to ibqp
+ */
+static void nes_drain_rq(struct ib_qp *ibqp)
+{
+ struct nes_qp *nesqp = to_nesqp(ibqp);
+
+ if (nesqp->hwqp.rq_tail != nesqp->hwqp.rq_head)
+ wait_for_completion(&nesqp->rq_drained);
+}
/**
* nes_poll_cq
@@ -3581,6 +3606,13 @@
}
}
+ if (nesqp->iwarp_state > NES_CQP_QP_IWARP_STATE_RTS) {
+ if (nesqp->hwqp.sq_tail == nesqp->hwqp.sq_head)
+ complete(&nesqp->sq_drained);
+ if (nesqp->hwqp.rq_tail == nesqp->hwqp.rq_head)
+ complete(&nesqp->rq_drained);
+ }
+
entry->wr_id = wrid;
entry++;
cqe_count++;
@@ -3753,6 +3785,8 @@
nesibdev->ibdev.req_notify_cq = nes_req_notify_cq;
nesibdev->ibdev.post_send = nes_post_send;
nesibdev->ibdev.post_recv = nes_post_recv;
+ nesibdev->ibdev.drain_sq = nes_drain_sq;
+ nesibdev->ibdev.drain_rq = nes_drain_rq;
nesibdev->ibdev.iwcm = kzalloc(sizeof(*nesibdev->ibdev.iwcm), GFP_KERNEL);
if (nesibdev->ibdev.iwcm == NULL) {
diff --git a/drivers/infiniband/hw/nes/nes_verbs.h b/drivers/infiniband/hw/nes/nes_verbs.h
index 7029088..e02a566 100644
--- a/drivers/infiniband/hw/nes/nes_verbs.h
+++ b/drivers/infiniband/hw/nes/nes_verbs.h
@@ -189,6 +189,8 @@
u8 pau_pending;
u8 pau_state;
__u64 nesuqp_addr;
+ struct completion sq_drained;
+ struct completion rq_drained;
};
struct ib_mr *nes_reg_phys_mr(struct ib_pd *ib_pd,
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ethtool.c b/drivers/infiniband/ulp/ipoib/ipoib_ethtool.c
index a53fa5f..1502199 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_ethtool.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_ethtool.c
@@ -36,6 +36,27 @@
#include "ipoib.h"
+struct ipoib_stats {
+ char stat_string[ETH_GSTRING_LEN];
+ int stat_offset;
+};
+
+#define IPOIB_NETDEV_STAT(m) { \
+ .stat_string = #m, \
+ .stat_offset = offsetof(struct rtnl_link_stats64, m) }
+
+static const struct ipoib_stats ipoib_gstrings_stats[] = {
+ IPOIB_NETDEV_STAT(rx_packets),
+ IPOIB_NETDEV_STAT(tx_packets),
+ IPOIB_NETDEV_STAT(rx_bytes),
+ IPOIB_NETDEV_STAT(tx_bytes),
+ IPOIB_NETDEV_STAT(tx_errors),
+ IPOIB_NETDEV_STAT(rx_dropped),
+ IPOIB_NETDEV_STAT(tx_dropped)
+};
+
+#define IPOIB_GLOBAL_STATS_LEN ARRAY_SIZE(ipoib_gstrings_stats)
+
static void ipoib_get_drvinfo(struct net_device *netdev,
struct ethtool_drvinfo *drvinfo)
{
@@ -92,11 +113,57 @@
return 0;
}
+static void ipoib_get_ethtool_stats(struct net_device *dev,
+ struct ethtool_stats __always_unused *stats,
+ u64 *data)
+{
+ int i;
+ struct net_device_stats *net_stats = &dev->stats;
+ u8 *p = (u8 *)net_stats;
+
+ for (i = 0; i < IPOIB_GLOBAL_STATS_LEN; i++)
+ data[i] = *(u64 *)(p + ipoib_gstrings_stats[i].stat_offset);
+
+}
+static void ipoib_get_strings(struct net_device __always_unused *dev,
+ u32 stringset, u8 *data)
+{
+ u8 *p = data;
+ int i;
+
+ switch (stringset) {
+ case ETH_SS_STATS:
+ for (i = 0; i < IPOIB_GLOBAL_STATS_LEN; i++) {
+ memcpy(p, ipoib_gstrings_stats[i].stat_string,
+ ETH_GSTRING_LEN);
+ p += ETH_GSTRING_LEN;
+ }
+ break;
+ case ETH_SS_TEST:
+ default:
+ break;
+ }
+}
+static int ipoib_get_sset_count(struct net_device __always_unused *dev,
+ int sset)
+{
+ switch (sset) {
+ case ETH_SS_STATS:
+ return IPOIB_GLOBAL_STATS_LEN;
+ case ETH_SS_TEST:
+ default:
+ break;
+ }
+ return -EOPNOTSUPP;
+}
static const struct ethtool_ops ipoib_ethtool_ops = {
.get_drvinfo = ipoib_get_drvinfo,
.get_coalesce = ipoib_get_coalesce,
.set_coalesce = ipoib_set_coalesce,
+ .get_strings = ipoib_get_strings,
+ .get_ethtool_stats = ipoib_get_ethtool_stats,
+ .get_sset_count = ipoib_get_sset_count,
};
void ipoib_set_ethtool_ops(struct net_device *dev)
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ib.c b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
index f0e55e4..da5f28c 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_ib.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
@@ -51,8 +51,6 @@
"Enable data path debug tracing if > 0");
#endif
-static DEFINE_MUTEX(pkey_mutex);
-
struct ipoib_ah *ipoib_create_ah(struct net_device *dev,
struct ib_pd *pd, struct ib_ah_attr *attr)
{
diff --git a/drivers/infiniband/ulp/srp/ib_srp.c b/drivers/infiniband/ulp/srp/ib_srp.c
index 0b576c5..7ecb197 100644
--- a/drivers/infiniband/ulp/srp/ib_srp.c
+++ b/drivers/infiniband/ulp/srp/ib_srp.c
@@ -1528,7 +1528,7 @@
if (dev->use_fast_reg) {
state.sg = idb_sg;
- sg_set_buf(idb_sg, req->indirect_desc, idb_len);
+ sg_init_one(idb_sg, req->indirect_desc, idb_len);
idb_sg->dma_address = req->indirect_dma_addr; /* hack! */
#ifdef CONFIG_NEED_SG_DMA_LENGTH
idb_sg->dma_length = idb_sg->length; /* hack^2 */
diff --git a/drivers/net/ethernet/chelsio/cxgb4/t4_msg.h b/drivers/net/ethernet/chelsio/cxgb4/t4_msg.h
index 80417fc..4705e2d 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/t4_msg.h
+++ b/drivers/net/ethernet/chelsio/cxgb4/t4_msg.h
@@ -1392,6 +1392,10 @@
#define T5_ULP_MEMIO_ORDER_V(x) ((x) << T5_ULP_MEMIO_ORDER_S)
#define T5_ULP_MEMIO_ORDER_F T5_ULP_MEMIO_ORDER_V(1U)
+#define T5_ULP_MEMIO_FID_S 4
+#define T5_ULP_MEMIO_FID_M 0x7ff
+#define T5_ULP_MEMIO_FID_V(x) ((x) << T5_ULP_MEMIO_FID_S)
+
/* ulp_mem_io.lock_addr fields */
#define ULP_MEMIO_ADDR_S 0
#define ULP_MEMIO_ADDR_V(x) ((x) << ULP_MEMIO_ADDR_S)
diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index 56bb0f3..fc0320c 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -220,6 +220,7 @@
IB_DEVICE_ON_DEMAND_PAGING = (1 << 31),
IB_DEVICE_SG_GAPS_REG = (1ULL << 32),
IB_DEVICE_VIRTUAL_FUNCTION = ((u64)1 << 33),
+ IB_DEVICE_RAW_SCATTER_FCS = ((u64)1 << 34),
};
enum ib_signature_prot_cap {
@@ -988,6 +989,7 @@
IB_QP_CREATE_NETIF_QP = 1 << 5,
IB_QP_CREATE_SIGNATURE_EN = 1 << 6,
IB_QP_CREATE_USE_GFP_NOIO = 1 << 7,
+ IB_QP_CREATE_SCATTER_FCS = 1 << 8,
/* reserve bits 26-31 for low level drivers' internal use */
IB_QP_CREATE_RESERVED_START = 1 << 26,
IB_QP_CREATE_RESERVED_END = 1 << 31,
diff --git a/include/uapi/rdma/ib_user_verbs.h b/include/uapi/rdma/ib_user_verbs.h
index 8126c14..b6543d7 100644
--- a/include/uapi/rdma/ib_user_verbs.h
+++ b/include/uapi/rdma/ib_user_verbs.h
@@ -226,6 +226,7 @@
struct ib_uverbs_odp_caps odp_caps;
__u64 timestamp_mask;
__u64 hca_core_clock; /* in KHZ */
+ __u64 device_cap_flags_ex;
};
struct ib_uverbs_query_port {