summaryrefslogtreecommitdiffstats
path: root/drivers/infiniband/hw
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/infiniband/hw')
-rw-r--r--drivers/infiniband/hw/Makefile12
-rw-r--r--drivers/infiniband/hw/cxgb3/cxio_hal.c6
-rw-r--r--drivers/infiniband/hw/cxgb3/iwch_cm.c1
-rw-r--r--drivers/infiniband/hw/cxgb4/cm.c324
-rw-r--r--drivers/infiniband/hw/cxgb4/cq.c11
-rw-r--r--drivers/infiniband/hw/cxgb4/device.c83
-rw-r--r--drivers/infiniband/hw/cxgb4/iw_cxgb4.h48
-rw-r--r--drivers/infiniband/hw/cxgb4/provider.c7
-rw-r--r--drivers/infiniband/hw/cxgb4/t4.h1
-rw-r--r--drivers/infiniband/hw/cxgb4/t4fw_ri_api.h1
-rw-r--r--drivers/infiniband/hw/cxgb4/user.h2
-rw-r--r--drivers/infiniband/hw/ipath/ipath_diag.c4
-rw-r--r--drivers/infiniband/hw/ipath/ipath_intr.c4
-rw-r--r--drivers/infiniband/hw/ipath/ipath_sdma.c4
-rw-r--r--drivers/infiniband/hw/mlx4/ah.c2
-rw-r--r--drivers/infiniband/hw/mlx4/cq.c6
-rw-r--r--drivers/infiniband/hw/mlx4/mad.c40
-rw-r--r--drivers/infiniband/hw/mlx4/main.c93
-rw-r--r--drivers/infiniband/hw/mlx4/mlx4_ib.h4
-rw-r--r--drivers/infiniband/hw/mlx4/qp.c110
-rw-r--r--drivers/infiniband/hw/mlx4/srq.c7
-rw-r--r--drivers/infiniband/hw/mlx4/sysfs.c105
-rw-r--r--drivers/infiniband/hw/mlx5/cq.c13
-rw-r--r--drivers/infiniband/hw/mlx5/mlx5_ib.h13
-rw-r--r--drivers/infiniband/hw/mlx5/mr.c76
-rw-r--r--drivers/infiniband/hw/mlx5/qp.c42
-rw-r--r--drivers/infiniband/hw/mlx5/srq.c14
-rw-r--r--drivers/infiniband/hw/mlx5/user.h2
-rw-r--r--drivers/infiniband/hw/nes/nes.c25
-rw-r--r--drivers/infiniband/hw/nes/nes.h3
-rw-r--r--drivers/infiniband/hw/nes/nes_cm.c320
-rw-r--r--drivers/infiniband/hw/nes/nes_cm.h12
-rw-r--r--drivers/infiniband/hw/ocrdma/ocrdma_stats.c9
-rw-r--r--drivers/infiniband/hw/qib/qib_init.c8
-rw-r--r--drivers/infiniband/hw/qib/qib_mad.c2
-rw-r--r--drivers/infiniband/hw/qib/qib_qp.c3
-rw-r--r--drivers/infiniband/hw/usnic/usnic_ib_verbs.c3
-rw-r--r--drivers/infiniband/hw/usnic/usnic_uiom_interval_tree.c18
38 files changed, 1153 insertions, 285 deletions
diff --git a/drivers/infiniband/hw/Makefile b/drivers/infiniband/hw/Makefile
new file mode 100644
index 00000000000..e900b03531a
--- /dev/null
+++ b/drivers/infiniband/hw/Makefile
@@ -0,0 +1,12 @@
+obj-$(CONFIG_INFINIBAND_MTHCA) += mthca/
+obj-$(CONFIG_INFINIBAND_IPATH) += ipath/
+obj-$(CONFIG_INFINIBAND_QIB) += qib/
+obj-$(CONFIG_INFINIBAND_EHCA) += ehca/
+obj-$(CONFIG_INFINIBAND_AMSO1100) += amso1100/
+obj-$(CONFIG_INFINIBAND_CXGB3) += cxgb3/
+obj-$(CONFIG_INFINIBAND_CXGB4) += cxgb4/
+obj-$(CONFIG_MLX4_INFINIBAND) += mlx4/
+obj-$(CONFIG_MLX5_INFINIBAND) += mlx5/
+obj-$(CONFIG_INFINIBAND_NES) += nes/
+obj-$(CONFIG_INFINIBAND_OCRDMA) += ocrdma/
+obj-$(CONFIG_INFINIBAND_USNIC) += usnic/
diff --git a/drivers/infiniband/hw/cxgb3/cxio_hal.c b/drivers/infiniband/hw/cxgb3/cxio_hal.c
index c3f5aca4ef0..de1c61b417d 100644
--- a/drivers/infiniband/hw/cxgb3/cxio_hal.c
+++ b/drivers/infiniband/hw/cxgb3/cxio_hal.c
@@ -735,14 +735,12 @@ static int __cxio_tpt_op(struct cxio_rdev *rdev_p, u32 reset_tpt_entry,
((perm & TPT_MW_BIND) ? F_TPT_MW_BIND_ENABLE : 0) |
V_TPT_ADDR_TYPE((zbva ? TPT_ZBTO : TPT_VATO)) |
V_TPT_PAGE_SIZE(page_size));
- tpt.rsvd_pbl_addr = reset_tpt_entry ? 0 :
- cpu_to_be32(V_TPT_PBL_ADDR(PBL_OFF(rdev_p, pbl_addr)>>3));
+ tpt.rsvd_pbl_addr = cpu_to_be32(V_TPT_PBL_ADDR(PBL_OFF(rdev_p, pbl_addr)>>3));
tpt.len = cpu_to_be32(len);
tpt.va_hi = cpu_to_be32((u32) (to >> 32));
tpt.va_low_or_fbo = cpu_to_be32((u32) (to & 0xFFFFFFFFULL));
tpt.rsvd_bind_cnt_or_pstag = 0;
- tpt.rsvd_pbl_size = reset_tpt_entry ? 0 :
- cpu_to_be32(V_TPT_PBL_SIZE(pbl_size >> 2));
+ tpt.rsvd_pbl_size = cpu_to_be32(V_TPT_PBL_SIZE(pbl_size >> 2));
}
err = cxio_hal_ctrl_qp_write_mem(rdev_p,
stag_idx +
diff --git a/drivers/infiniband/hw/cxgb3/iwch_cm.c b/drivers/infiniband/hw/cxgb3/iwch_cm.c
index 095bb046e2c..cb78b1e9bcd 100644
--- a/drivers/infiniband/hw/cxgb3/iwch_cm.c
+++ b/drivers/infiniband/hw/cxgb3/iwch_cm.c
@@ -418,6 +418,7 @@ static int send_abort(struct iwch_ep *ep, struct sk_buff *skb, gfp_t gfp)
skb->priority = CPL_PRIORITY_DATA;
set_arp_failure_handler(skb, abort_arp_failure);
req = (struct cpl_abort_req *) skb_put(skb, sizeof(*req));
+ memset(req, 0, sizeof(*req));
req->wr.wr_hi = htonl(V_WR_OP(FW_WROPCODE_OFLD_HOST_ABORT_CON_REQ));
req->wr.wr_lo = htonl(V_WR_TID(ep->hwtid));
OPCODE_TID(req) = htonl(MK_OPCODE_TID(CPL_ABORT_REQ, ep->hwtid));
diff --git a/drivers/infiniband/hw/cxgb4/cm.c b/drivers/infiniband/hw/cxgb4/cm.c
index 1f863a96a48..5e153f6d4b4 100644
--- a/drivers/infiniband/hw/cxgb4/cm.c
+++ b/drivers/infiniband/hw/cxgb4/cm.c
@@ -1,5 +1,5 @@
/*
- * Copyright (c) 2009-2010 Chelsio, Inc. All rights reserved.
+ * Copyright (c) 2009-2014 Chelsio, Inc. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
@@ -47,6 +47,8 @@
#include <net/ip6_route.h>
#include <net/addrconf.h>
+#include <rdma/ib_addr.h>
+
#include "iw_cxgb4.h"
static char *states[] = {
@@ -232,12 +234,16 @@ static void release_tid(struct c4iw_rdev *rdev, u32 hwtid, struct sk_buff *skb)
static void set_emss(struct c4iw_ep *ep, u16 opt)
{
- ep->emss = ep->com.dev->rdev.lldi.mtus[GET_TCPOPT_MSS(opt)] - 40;
+ ep->emss = ep->com.dev->rdev.lldi.mtus[GET_TCPOPT_MSS(opt)] -
+ sizeof(struct iphdr) - sizeof(struct tcphdr);
ep->mss = ep->emss;
if (GET_TCPOPT_TSTAMP(opt))
ep->emss -= 12;
if (ep->emss < 128)
ep->emss = 128;
+ if (ep->emss & 7)
+ PDBG("Warning: misaligned mtu idx %u mss %u emss=%u\n",
+ GET_TCPOPT_MSS(opt), ep->mss, ep->emss);
PDBG("%s mss_idx %u mss %u emss=%u\n", __func__, GET_TCPOPT_MSS(opt),
ep->mss, ep->emss);
}
@@ -294,6 +300,12 @@ void _c4iw_free_ep(struct kref *kref)
dst_release(ep->dst);
cxgb4_l2t_release(ep->l2t);
}
+ if (test_bit(RELEASE_MAPINFO, &ep->com.flags)) {
+ print_addr(&ep->com, __func__, "remove_mapinfo/mapping");
+ iwpm_remove_mapinfo(&ep->com.local_addr,
+ &ep->com.mapped_local_addr);
+ iwpm_remove_mapping(&ep->com.local_addr, RDMA_NL_C4IW);
+ }
kfree(ep);
}
@@ -341,10 +353,7 @@ static struct sk_buff *get_skb(struct sk_buff *skb, int len, gfp_t gfp)
static struct net_device *get_real_dev(struct net_device *egress_dev)
{
- struct net_device *phys_dev = egress_dev;
- if (egress_dev->priv_flags & IFF_802_1Q_VLAN)
- phys_dev = vlan_dev_real_dev(egress_dev);
- return phys_dev;
+ return rdma_vlan_dev_real_dev(egress_dev) ? : egress_dev;
}
static int our_interface(struct c4iw_dev *dev, struct net_device *egress_dev)
@@ -468,7 +477,7 @@ static void send_flowc(struct c4iw_ep *ep, struct sk_buff *skb)
flowc->mnemval[5].mnemonic = FW_FLOWC_MNEM_RCVNXT;
flowc->mnemval[5].val = cpu_to_be32(ep->rcv_seq);
flowc->mnemval[6].mnemonic = FW_FLOWC_MNEM_SNDBUF;
- flowc->mnemval[6].val = cpu_to_be32(snd_win);
+ flowc->mnemval[6].val = cpu_to_be32(ep->snd_win);
flowc->mnemval[7].mnemonic = FW_FLOWC_MNEM_MSS;
flowc->mnemval[7].val = cpu_to_be32(ep->emss);
/* Pad WR to 16 byte boundary */
@@ -528,6 +537,49 @@ static int send_abort(struct c4iw_ep *ep, struct sk_buff *skb, gfp_t gfp)
return c4iw_l2t_send(&ep->com.dev->rdev, skb, ep->l2t);
}
+/*
+ * c4iw_form_pm_msg - Form a port mapper message with mapping info
+ */
+static void c4iw_form_pm_msg(struct c4iw_ep *ep,
+ struct iwpm_sa_data *pm_msg)
+{
+ memcpy(&pm_msg->loc_addr, &ep->com.local_addr,
+ sizeof(ep->com.local_addr));
+ memcpy(&pm_msg->rem_addr, &ep->com.remote_addr,
+ sizeof(ep->com.remote_addr));
+}
+
+/*
+ * c4iw_form_reg_msg - Form a port mapper message with dev info
+ */
+static void c4iw_form_reg_msg(struct c4iw_dev *dev,
+ struct iwpm_dev_data *pm_msg)
+{
+ memcpy(pm_msg->dev_name, dev->ibdev.name, IWPM_DEVNAME_SIZE);
+ memcpy(pm_msg->if_name, dev->rdev.lldi.ports[0]->name,
+ IWPM_IFNAME_SIZE);
+}
+
+static void c4iw_record_pm_msg(struct c4iw_ep *ep,
+ struct iwpm_sa_data *pm_msg)
+{
+ memcpy(&ep->com.mapped_local_addr, &pm_msg->mapped_loc_addr,
+ sizeof(ep->com.mapped_local_addr));
+ memcpy(&ep->com.mapped_remote_addr, &pm_msg->mapped_rem_addr,
+ sizeof(ep->com.mapped_remote_addr));
+}
+
+static void best_mtu(const unsigned short *mtus, unsigned short mtu,
+ unsigned int *idx, int use_ts)
+{
+ unsigned short hdr_size = sizeof(struct iphdr) +
+ sizeof(struct tcphdr) +
+ (use_ts ? 12 : 0);
+ unsigned short data_size = mtu - hdr_size;
+
+ cxgb4_best_aligned_mtu(mtus, hdr_size, data_size, 8, idx);
+}
+
static int send_connect(struct c4iw_ep *ep)
{
struct cpl_act_open_req *req;
@@ -546,10 +598,15 @@ static int send_connect(struct c4iw_ep *ep)
int sizev6 = is_t4(ep->com.dev->rdev.lldi.adapter_type) ?
sizeof(struct cpl_act_open_req6) :
sizeof(struct cpl_t5_act_open_req6);
- struct sockaddr_in *la = (struct sockaddr_in *)&ep->com.local_addr;
- struct sockaddr_in *ra = (struct sockaddr_in *)&ep->com.remote_addr;
- struct sockaddr_in6 *la6 = (struct sockaddr_in6 *)&ep->com.local_addr;
- struct sockaddr_in6 *ra6 = (struct sockaddr_in6 *)&ep->com.remote_addr;
+ struct sockaddr_in *la = (struct sockaddr_in *)
+ &ep->com.mapped_local_addr;
+ struct sockaddr_in *ra = (struct sockaddr_in *)
+ &ep->com.mapped_remote_addr;
+ struct sockaddr_in6 *la6 = (struct sockaddr_in6 *)
+ &ep->com.mapped_local_addr;
+ struct sockaddr_in6 *ra6 = (struct sockaddr_in6 *)
+ &ep->com.mapped_remote_addr;
+ int win;
wrlen = (ep->com.remote_addr.ss_family == AF_INET) ?
roundup(sizev4, 16) :
@@ -565,8 +622,18 @@ static int send_connect(struct c4iw_ep *ep)
}
set_wr_txq(skb, CPL_PRIORITY_SETUP, ep->ctrlq_idx);
- cxgb4_best_mtu(ep->com.dev->rdev.lldi.mtus, ep->mtu, &mtu_idx);
+ best_mtu(ep->com.dev->rdev.lldi.mtus, ep->mtu, &mtu_idx,
+ enable_tcp_timestamps);
wscale = compute_wscale(rcv_win);
+
+ /*
+ * Specify the largest window that will fit in opt0. The
+ * remainder will be specified in the rx_data_ack.
+ */
+ win = ep->rcv_win >> 10;
+ if (win > RCV_BUFSIZ_MASK)
+ win = RCV_BUFSIZ_MASK;
+
opt0 = (nocong ? NO_CONG(1) : 0) |
KEEP_ALIVE(1) |
DELACK(1) |
@@ -577,7 +644,7 @@ static int send_connect(struct c4iw_ep *ep)
SMAC_SEL(ep->smac_idx) |
DSCP(ep->tos) |
ULP_MODE(ULP_MODE_TCPDDP) |
- RCV_BUFSIZ(rcv_win>>10);
+ RCV_BUFSIZ(win);
opt2 = RX_CHANNEL(0) |
CCTRL_ECN(enable_ecn) |
RSS_QUEUE_VALID | RSS_QUEUE(ep->rss_qid);
@@ -633,6 +700,13 @@ static int send_connect(struct c4iw_ep *ep)
req6->opt2 = cpu_to_be32(opt2);
}
} else {
+ u32 isn = (prandom_u32() & ~7UL) - 1;
+
+ opt2 |= T5_OPT_2_VALID;
+ opt2 |= CONG_CNTRL_VALID; /* OPT_2_ISS for T5 */
+ if (peer2peer)
+ isn += 4;
+
if (ep->com.remote_addr.ss_family == AF_INET) {
t5_req = (struct cpl_t5_act_open_req *)
skb_put(skb, wrlen);
@@ -649,6 +723,9 @@ static int send_connect(struct c4iw_ep *ep)
cxgb4_select_ntuple(
ep->com.dev->rdev.lldi.ports[0],
ep->l2t)));
+ t5_req->rsvd = cpu_to_be32(isn);
+ PDBG("%s snd_isn %u\n", __func__,
+ be32_to_cpu(t5_req->rsvd));
t5_req->opt2 = cpu_to_be32(opt2);
} else {
t5_req6 = (struct cpl_t5_act_open_req6 *)
@@ -672,6 +749,9 @@ static int send_connect(struct c4iw_ep *ep)
cxgb4_select_ntuple(
ep->com.dev->rdev.lldi.ports[0],
ep->l2t));
+ t5_req6->rsvd = cpu_to_be32(isn);
+ PDBG("%s snd_isn %u\n", __func__,
+ be32_to_cpu(t5_req6->rsvd));
t5_req6->opt2 = cpu_to_be32(opt2);
}
}
@@ -1145,6 +1225,14 @@ static int update_rx_credits(struct c4iw_ep *ep, u32 credits)
return 0;
}
+ /*
+ * If we couldn't specify the entire rcv window at connection setup
+ * due to the limit in the number of bits in the RCV_BUFSIZ field,
+ * then add the overage in to the credits returned.
+ */
+ if (ep->rcv_win > RCV_BUFSIZ_MASK * 1024)
+ credits += ep->rcv_win - RCV_BUFSIZ_MASK * 1024;
+
req = (struct cpl_rx_data_ack *) skb_put(skb, wrlen);
memset(req, 0, wrlen);
INIT_TP_WR(req, ep->hwtid);
@@ -1618,6 +1706,7 @@ static void send_fw_act_open_req(struct c4iw_ep *ep, unsigned int atid)
unsigned int mtu_idx;
int wscale;
struct sockaddr_in *sin;
+ int win;
skb = get_skb(NULL, sizeof(*req), GFP_KERNEL);
req = (struct fw_ofld_connection_wr *)__skb_put(skb, sizeof(*req));
@@ -1627,10 +1716,10 @@ static void send_fw_act_open_req(struct c4iw_ep *ep, unsigned int atid)
req->le.filter = cpu_to_be32(cxgb4_select_ntuple(
ep->com.dev->rdev.lldi.ports[0],
ep->l2t));
- sin = (struct sockaddr_in *)&ep->com.local_addr;
+ sin = (struct sockaddr_in *)&ep->com.mapped_local_addr;
req->le.lport = sin->sin_port;
req->le.u.ipv4.lip = sin->sin_addr.s_addr;
- sin = (struct sockaddr_in *)&ep->com.remote_addr;
+ sin = (struct sockaddr_in *)&ep->com.mapped_remote_addr;
req->le.pport = sin->sin_port;
req->le.u.ipv4.pip = sin->sin_addr.s_addr;
req->tcb.t_state_to_astid =
@@ -1640,8 +1729,18 @@ static void send_fw_act_open_req(struct c4iw_ep *ep, unsigned int atid)
htons(F_FW_OFLD_CONNECTION_WR_CPLRXDATAACK);
req->tcb.tx_max = (__force __be32) jiffies;
req->tcb.rcv_adv = htons(1);
- cxgb4_best_mtu(ep->com.dev->rdev.lldi.mtus, ep->mtu, &mtu_idx);
+ best_mtu(ep->com.dev->rdev.lldi.mtus, ep->mtu, &mtu_idx,
+ enable_tcp_timestamps);
wscale = compute_wscale(rcv_win);
+
+ /*
+ * Specify the largest window that will fit in opt0. The
+ * remainder will be specified in the rx_data_ack.
+ */
+ win = ep->rcv_win >> 10;
+ if (win > RCV_BUFSIZ_MASK)
+ win = RCV_BUFSIZ_MASK;
+
req->tcb.opt0 = (__force __be64) (TCAM_BYPASS(1) |
(nocong ? NO_CONG(1) : 0) |
KEEP_ALIVE(1) |
@@ -1653,7 +1752,7 @@ static void send_fw_act_open_req(struct c4iw_ep *ep, unsigned int atid)
SMAC_SEL(ep->smac_idx) |
DSCP(ep->tos) |
ULP_MODE(ULP_MODE_TCPDDP) |
- RCV_BUFSIZ(rcv_win >> 10));
+ RCV_BUFSIZ(win));
req->tcb.opt2 = (__force __be32) (PACE(1) |
TX_QUEUE(ep->com.dev->rdev.lldi.tx_modq[ep->tx_chan]) |
RX_CHANNEL(0) |
@@ -1690,6 +1789,13 @@ static int is_neg_adv(unsigned int status)
status == CPL_ERR_KEEPALV_NEG_ADVICE;
}
+static void set_tcp_window(struct c4iw_ep *ep, struct port_info *pi)
+{
+ ep->snd_win = snd_win;
+ ep->rcv_win = rcv_win;
+ PDBG("%s snd_win %d rcv_win %d\n", __func__, ep->snd_win, ep->rcv_win);
+}
+
#define ACT_OPEN_RETRY_COUNT 2
static int import_ep(struct c4iw_ep *ep, int iptype, __u8 *peer_ip,
@@ -1738,6 +1844,7 @@ static int import_ep(struct c4iw_ep *ep, int iptype, __u8 *peer_ip,
ep->ctrlq_idx = cxgb4_port_idx(pdev);
ep->rss_qid = cdev->rdev.lldi.rxq_ids[
cxgb4_port_idx(pdev) * step];
+ set_tcp_window(ep, (struct port_info *)netdev_priv(pdev));
dev_put(pdev);
} else {
pdev = get_real_dev(n->dev);
@@ -1746,16 +1853,17 @@ static int import_ep(struct c4iw_ep *ep, int iptype, __u8 *peer_ip,
if (!ep->l2t)
goto out;
ep->mtu = dst_mtu(dst);
- ep->tx_chan = cxgb4_port_chan(n->dev);
- ep->smac_idx = (cxgb4_port_viid(n->dev) & 0x7F) << 1;
+ ep->tx_chan = cxgb4_port_chan(pdev);
+ ep->smac_idx = (cxgb4_port_viid(pdev) & 0x7F) << 1;
step = cdev->rdev.lldi.ntxq /
cdev->rdev.lldi.nchan;
- ep->txq_idx = cxgb4_port_idx(n->dev) * step;
- ep->ctrlq_idx = cxgb4_port_idx(n->dev);
+ ep->txq_idx = cxgb4_port_idx(pdev) * step;
+ ep->ctrlq_idx = cxgb4_port_idx(pdev);
step = cdev->rdev.lldi.nrxq /
cdev->rdev.lldi.nchan;
ep->rss_qid = cdev->rdev.lldi.rxq_ids[
- cxgb4_port_idx(n->dev) * step];
+ cxgb4_port_idx(pdev) * step];
+ set_tcp_window(ep, (struct port_info *)netdev_priv(pdev));
if (clear_mpa_v1) {
ep->retry_with_mpa_v1 = 0;
@@ -1870,10 +1978,10 @@ static int act_open_rpl(struct c4iw_dev *dev, struct sk_buff *skb)
struct sockaddr_in6 *ra6;
ep = lookup_atid(t, atid);
- la = (struct sockaddr_in *)&ep->com.local_addr;
- ra = (struct sockaddr_in *)&ep->com.remote_addr;
- la6 = (struct sockaddr_in6 *)&ep->com.local_addr;
- ra6 = (struct sockaddr_in6 *)&ep->com.remote_addr;
+ la = (struct sockaddr_in *)&ep->com.mapped_local_addr;
+ ra = (struct sockaddr_in *)&ep->com.mapped_remote_addr;
+ la6 = (struct sockaddr_in6 *)&ep->com.mapped_local_addr;
+ ra6 = (struct sockaddr_in6 *)&ep->com.mapped_remote_addr;
PDBG("%s ep %p atid %u status %u errno %d\n", __func__, ep, atid,
status, status2errno(status));
@@ -1986,13 +2094,36 @@ static void accept_cr(struct c4iw_ep *ep, struct sk_buff *skb,
u64 opt0;
u32 opt2;
int wscale;
+ struct cpl_t5_pass_accept_rpl *rpl5 = NULL;
+ int win;
PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
BUG_ON(skb_cloned(skb));
- skb_trim(skb, sizeof(*rpl));
+
skb_get(skb);
- cxgb4_best_mtu(ep->com.dev->rdev.lldi.mtus, ep->mtu, &mtu_idx);
+ rpl = cplhdr(skb);
+ if (is_t5(ep->com.dev->rdev.lldi.adapter_type)) {
+ skb_trim(skb, roundup(sizeof(*rpl5), 16));
+ rpl5 = (void *)rpl;
+ INIT_TP_WR(rpl5, ep->hwtid);
+ } else {
+ skb_trim(skb, sizeof(*rpl));
+ INIT_TP_WR(rpl, ep->hwtid);
+ }
+ OPCODE_TID(rpl) = cpu_to_be32(MK_OPCODE_TID(CPL_PASS_ACCEPT_RPL,
+ ep->hwtid));
+
+ best_mtu(ep->com.dev->rdev.lldi.mtus, ep->mtu, &mtu_idx,
+ enable_tcp_timestamps && req->tcpopt.tstamp);
wscale = compute_wscale(rcv_win);
+
+ /*
+ * Specify the largest window that will fit in opt0. The
+ * remainder will be specified in the rx_data_ack.
+ */
+ win = ep->rcv_win >> 10;
+ if (win > RCV_BUFSIZ_MASK)
+ win = RCV_BUFSIZ_MASK;
opt0 = (nocong ? NO_CONG(1) : 0) |
KEEP_ALIVE(1) |
DELACK(1) |
@@ -2003,7 +2134,7 @@ static void accept_cr(struct c4iw_ep *ep, struct sk_buff *skb,
SMAC_SEL(ep->smac_idx) |
DSCP(ep->tos >> 2) |
ULP_MODE(ULP_MODE_TCPDDP) |
- RCV_BUFSIZ(rcv_win>>10);
+ RCV_BUFSIZ(win);
opt2 = RX_CHANNEL(0) |
RSS_QUEUE_VALID | RSS_QUEUE(ep->rss_qid);
@@ -2023,14 +2154,18 @@ static void accept_cr(struct c4iw_ep *ep, struct sk_buff *skb,
opt2 |= CCTRL_ECN(1);
}
if (is_t5(ep->com.dev->rdev.lldi.adapter_type)) {
+ u32 isn = (prandom_u32() & ~7UL) - 1;
opt2 |= T5_OPT_2_VALID;
opt2 |= V_CONG_CNTRL(CONG_ALG_TAHOE);
+ opt2 |= CONG_CNTRL_VALID; /* OPT_2_ISS for T5 */
+ rpl5 = (void *)rpl;
+ memset(&rpl5->iss, 0, roundup(sizeof(*rpl5)-sizeof(*rpl), 16));
+ if (peer2peer)
+ isn += 4;
+ rpl5->iss = cpu_to_be32(isn);
+ PDBG("%s iss %u\n", __func__, be32_to_cpu(rpl5->iss));
}
- rpl = cplhdr(skb);
- INIT_TP_WR(rpl, ep->hwtid);
- OPCODE_TID(rpl) = cpu_to_be32(MK_OPCODE_TID(CPL_PASS_ACCEPT_RPL,
- ep->hwtid));
rpl->opt0 = cpu_to_be64(opt0);
rpl->opt2 = cpu_to_be32(opt2);
set_wr_txq(skb, CPL_PRIORITY_SETUP, ep->ctrlq_idx);
@@ -2095,6 +2230,7 @@ static int pass_accept_req(struct c4iw_dev *dev, struct sk_buff *skb)
int err;
u16 peer_mss = ntohs(req->tcpopt.mss);
int iptype;
+ unsigned short hdrs;
parent_ep = lookup_stid(t, stid);
if (!parent_ep) {
@@ -2152,8 +2288,10 @@ static int pass_accept_req(struct c4iw_dev *dev, struct sk_buff *skb)
goto reject;
}
- if (peer_mss && child_ep->mtu > (peer_mss + 40))
- child_ep->mtu = peer_mss + 40;
+ hdrs = sizeof(struct iphdr) + sizeof(struct tcphdr) +
+ ((enable_tcp_timestamps && req->tcpopt.tstamp) ? 12 : 0);
+ if (peer_mss && child_ep->mtu > (peer_mss + hdrs))
+ child_ep->mtu = peer_mss + hdrs;
state_set(&child_ep->com, CONNECTING);
child_ep->com.dev = dev;
@@ -2730,13 +2868,15 @@ int c4iw_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
struct c4iw_dev *dev = to_c4iw_dev(cm_id->device);
struct c4iw_ep *ep;
int err = 0;
- struct sockaddr_in *laddr = (struct sockaddr_in *)&cm_id->local_addr;
- struct sockaddr_in *raddr = (struct sockaddr_in *)&cm_id->remote_addr;
- struct sockaddr_in6 *laddr6 = (struct sockaddr_in6 *)&cm_id->local_addr;
- struct sockaddr_in6 *raddr6 = (struct sockaddr_in6 *)
- &cm_id->remote_addr;
+ struct sockaddr_in *laddr;
+ struct sockaddr_in *raddr;
+ struct sockaddr_in6 *laddr6;
+ struct sockaddr_in6 *raddr6;
+ struct iwpm_dev_data pm_reg_msg;
+ struct iwpm_sa_data pm_msg;
__u8 *ra;
int iptype;
+ int iwpm_err = 0;
if ((conn_param->ord > c4iw_max_read_depth) ||
(conn_param->ird > c4iw_max_read_depth)) {
@@ -2767,7 +2907,7 @@ int c4iw_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
if (!ep->com.qp) {
PDBG("%s qpn 0x%x not found!\n", __func__, conn_param->qpn);
err = -EINVAL;
- goto fail2;
+ goto fail1;
}
ref_qp(ep);
PDBG("%s qpn 0x%x qp %p cm_id %p\n", __func__, conn_param->qpn,
@@ -2780,10 +2920,50 @@ int c4iw_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
if (ep->atid == -1) {
printk(KERN_ERR MOD "%s - cannot alloc atid.\n", __func__);
err = -ENOMEM;
- goto fail2;
+ goto fail1;
}
insert_handle(dev, &dev->atid_idr, ep, ep->atid);
+ memcpy(&ep->com.local_addr, &cm_id->local_addr,
+ sizeof(ep->com.local_addr));
+ memcpy(&ep->com.remote_addr, &cm_id->remote_addr,
+ sizeof(ep->com.remote_addr));
+
+ /* No port mapper available, go with the specified peer information */
+ memcpy(&ep->com.mapped_local_addr, &cm_id->local_addr,
+ sizeof(ep->com.mapped_local_addr));
+ memcpy(&ep->com.mapped_remote_addr, &cm_id->remote_addr,
+ sizeof(ep->com.mapped_remote_addr));
+
+ c4iw_form_reg_msg(dev, &pm_reg_msg);
+ iwpm_err = iwpm_register_pid(&pm_reg_msg, RDMA_NL_C4IW);
+ if (iwpm_err) {
+ PDBG("%s: Port Mapper reg pid fail (err = %d).\n",
+ __func__, iwpm_err);
+ }
+ if (iwpm_valid_pid() && !iwpm_err) {
+ c4iw_form_pm_msg(ep, &pm_msg);
+ iwpm_err = iwpm_add_and_query_mapping(&pm_msg, RDMA_NL_C4IW);
+ if (iwpm_err)
+ PDBG("%s: Port Mapper query fail (err = %d).\n",
+ __func__, iwpm_err);
+ else
+ c4iw_record_pm_msg(ep, &pm_msg);
+ }
+ if (iwpm_create_mapinfo(&ep->com.local_addr,
+ &ep->com.mapped_local_addr, RDMA_NL_C4IW)) {
+ iwpm_remove_mapping(&ep->com.local_addr, RDMA_NL_C4IW);
+ err = -ENOMEM;
+ goto fail1;
+ }
+ print_addr(&ep->com, __func__, "add_query/create_mapinfo");
+ set_bit(RELEASE_MAPINFO, &ep->com.flags);
+
+ laddr = (struct sockaddr_in *)&ep->com.mapped_local_addr;
+ raddr = (struct sockaddr_in *)&ep->com.mapped_remote_addr;
+ laddr6 = (struct sockaddr_in6 *)&ep->com.mapped_local_addr;
+ raddr6 = (struct sockaddr_in6 *) &ep->com.mapped_remote_addr;
+
if (cm_id->remote_addr.ss_family == AF_INET) {
iptype = 4;
ra = (__u8 *)&raddr->sin_addr;
@@ -2794,7 +2974,7 @@ int c4iw_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
if ((__force int)raddr->sin_addr.s_addr == INADDR_ANY) {
err = pick_local_ipaddrs(dev, cm_id);
if (err)
- goto fail2;
+ goto fail1;
}
/* find a route */
@@ -2814,7 +2994,7 @@ int c4iw_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
if (ipv6_addr_type(&raddr6->sin6_addr) == IPV6_ADDR_ANY) {
err = pick_local_ip6addrs(dev, cm_id);
if (err)
- goto fail2;
+ goto fail1;
}
/* find a route */
@@ -2830,13 +3010,13 @@ int c4iw_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
if (!ep->dst) {
printk(KERN_ERR MOD "%s - cannot find route.\n", __func__);
err = -EHOSTUNREACH;
- goto fail3;
+ goto fail2;
}
err = import_ep(ep, iptype, ra, ep->dst, ep->com.dev, true);
if (err) {
printk(KERN_ERR MOD "%s - cannot alloc l2e.\n", __func__);
- goto fail4;
+ goto fail3;
}
PDBG("%s txq_idx %u tx_chan %u smac_idx %u rss_qid %u l2t_idx %u\n",
@@ -2845,10 +3025,6 @@ int c4iw_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
state_set(&ep->com, CONNECTING);
ep->tos = 0;
- memcpy(&ep->com.local_addr, &cm_id->local_addr,
- sizeof(ep->com.local_addr));
- memcpy(&ep->com.remote_addr, &cm_id->remote_addr,
- sizeof(ep->com.remote_addr));
/* send connect request to rnic */
err = send_connect(ep);
@@ -2856,12 +3032,12 @@ int c4iw_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
goto out;
cxgb4_l2t_release(ep->l2t);
-fail4:
- dst_release(ep->dst);
fail3:
+ dst_release(ep->dst);
+fail2:
remove_handle(ep->com.dev, &ep->com.dev->atid_idr, ep->atid);
cxgb4_free_atid(ep->com.dev->rdev.lldi.tids, ep->atid);
-fail2:
+fail1:
cm_id->rem_ref(cm_id);
c4iw_put_ep(&ep->com);
out:
@@ -2871,7 +3047,8 @@ out:
static int create_server6(struct c4iw_dev *dev, struct c4iw_listen_ep *ep)
{
int err;
- struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)&ep->com.local_addr;
+ struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)
+ &ep->com.mapped_local_addr;
c4iw_init_wr_wait(&ep->com.wr_wait);
err = cxgb4_create_server6(ep->com.dev->rdev.lldi.ports[0],
@@ -2892,7 +3069,8 @@ static int create_server6(struct c4iw_dev *dev, struct c4iw_listen_ep *ep)
static int create_server4(struct c4iw_dev *dev, struct c4iw_listen_ep *ep)
{
int err;
- struct sockaddr_in *sin = (struct sockaddr_in *)&ep->com.local_addr;
+ struct sockaddr_in *sin = (struct sockaddr_in *)
+ &ep->com.mapped_local_addr;
if (dev->rdev.lldi.enable_fw_ofld_conn) {
do {
@@ -2927,6 +3105,9 @@ int c4iw_create_listen(struct iw_cm_id *cm_id, int backlog)
int err = 0;
struct c4iw_dev *dev = to_c4iw_dev(cm_id->device);
struct c4iw_listen_ep *ep;
+ struct iwpm_dev_data pm_reg_msg;
+ struct iwpm_sa_data pm_msg;
+ int iwpm_err = 0;
might_sleep();
@@ -2961,6 +3142,37 @@ int c4iw_create_listen(struct iw_cm_id *cm_id, int backlog)
goto fail2;
}
insert_handle(dev, &dev->stid_idr, ep, ep->stid);
+
+ /* No port mapper available, go with the specified info */
+ memcpy(&ep->com.mapped_local_addr, &cm_id->local_addr,
+ sizeof(ep->com.mapped_local_addr));
+
+ c4iw_form_reg_msg(dev, &pm_reg_msg);
+ iwpm_err = iwpm_register_pid(&pm_reg_msg, RDMA_NL_C4IW);
+ if (iwpm_err) {
+ PDBG("%s: Port Mapper reg pid fail (err = %d).\n",
+ __func__, iwpm_err);
+ }
+ if (iwpm_valid_pid() && !iwpm_err) {
+ memcpy(&pm_msg.loc_addr, &ep->com.local_addr,
+ sizeof(ep->com.local_addr));
+ iwpm_err = iwpm_add_mapping(&pm_msg, RDMA_NL_C4IW);
+ if (iwpm_err)
+ PDBG("%s: Port Mapper query fail (err = %d).\n",
+ __func__, iwpm_err);
+ else
+ memcpy(&ep->com.mapped_local_addr,
+ &pm_msg.mapped_loc_addr,
+ sizeof(ep->com.mapped_local_addr));
+ }
+ if (iwpm_create_mapinfo(&ep->com.local_addr,
+ &ep->com.mapped_local_addr, RDMA_NL_C4IW)) {
+ err = -ENOMEM;
+ goto fail3;
+ }
+ print_addr(&ep->com, __func__, "add_mapping/create_mapinfo");
+
+ set_bit(RELEASE_MAPINFO, &ep->com.flags);
state_set(&ep->com, LISTEN);
if (ep->com.local_addr.ss_family == AF_INET)
err = create_server4(dev, ep);
@@ -2970,6 +3182,8 @@ int c4iw_create_listen(struct iw_cm_id *cm_id, int backlog)
cm_id->provider_data = ep;
goto out;
}
+
+fail3:
cxgb4_free_stid(ep->com.dev->rdev.lldi.tids, ep->stid,
ep->com.local_addr.ss_family);
fail2:
diff --git a/drivers/infiniband/hw/cxgb4/cq.c b/drivers/infiniband/hw/cxgb4/cq.c
index cfaa56ada18..c04292c950f 100644
--- a/drivers/infiniband/hw/cxgb4/cq.c
+++ b/drivers/infiniband/hw/cxgb4/cq.c
@@ -134,7 +134,8 @@ static int create_cq(struct c4iw_rdev *rdev, struct t4_cq *cq,
V_FW_RI_RES_WR_IQANUS(0) |
V_FW_RI_RES_WR_IQANUD(1) |
F_FW_RI_RES_WR_IQANDST |
- V_FW_RI_RES_WR_IQANDSTINDEX(*rdev->lldi.rxq_ids));
+ V_FW_RI_RES_WR_IQANDSTINDEX(
+ rdev->lldi.ciq_ids[cq->vector]));
res->u.cq.iqdroprss_to_iqesize = cpu_to_be16(
F_FW_RI_RES_WR_IQDROPRSS |
V_FW_RI_RES_WR_IQPCIECH(2) |
@@ -870,6 +871,9 @@ struct ib_cq *c4iw_create_cq(struct ib_device *ibdev, int entries,
rhp = to_c4iw_dev(ibdev);
+ if (vector >= rhp->rdev.lldi.nciq)
+ return ERR_PTR(-EINVAL);
+
chp = kzalloc(sizeof(*chp), GFP_KERNEL);
if (!chp)
return ERR_PTR(-ENOMEM);
@@ -915,6 +919,7 @@ struct ib_cq *c4iw_create_cq(struct ib_device *ibdev, int entries,
}
chp->cq.size = hwentries;
chp->cq.memsize = memsize;
+ chp->cq.vector = vector;
ret = create_cq(&rhp->rdev, &chp->cq,
ucontext ? &ucontext->uctx : &rhp->rdev.uctx);
@@ -940,7 +945,6 @@ struct ib_cq *c4iw_create_cq(struct ib_device *ibdev, int entries,
if (!mm2)
goto err4;
- memset(&uresp, 0, sizeof(uresp));
uresp.qid_mask = rhp->rdev.cqmask;
uresp.cqid = chp->cq.cqid;
uresp.size = chp->cq.size;
@@ -951,7 +955,8 @@ struct ib_cq *c4iw_create_cq(struct ib_device *ibdev, int entries,
uresp.gts_key = ucontext->key;
ucontext->key += PAGE_SIZE;
spin_unlock(&ucontext->mmap_lock);
- ret = ib_copy_to_udata(udata, &uresp, sizeof uresp);
+ ret = ib_copy_to_udata(udata, &uresp,
+ sizeof(uresp) - sizeof(uresp.reserved));
if (ret)
goto err5;
diff --git a/drivers/infiniband/hw/cxgb4/device.c b/drivers/infiniband/hw/cxgb4/device.c
index f4fa50a609e..dd93aadc996 100644
--- a/drivers/infiniband/hw/cxgb4/device.c
+++ b/drivers/infiniband/hw/cxgb4/device.c
@@ -77,6 +77,16 @@ struct c4iw_debugfs_data {
int pos;
};
+/* registered cxgb4 netlink callbacks */
+static struct ibnl_client_cbs c4iw_nl_cb_table[] = {
+ [RDMA_NL_IWPM_REG_PID] = {.dump = iwpm_register_pid_cb},
+ [RDMA_NL_IWPM_ADD_MAPPING] = {.dump = iwpm_add_mapping_cb},
+ [RDMA_NL_IWPM_QUERY_MAPPING] = {.dump = iwpm_add_and_query_mapping_cb},
+ [RDMA_NL_IWPM_HANDLE_ERR] = {.dump = iwpm_mapping_error_cb},
+ [RDMA_NL_IWPM_MAPINFO] = {.dump = iwpm_mapping_info_cb},
+ [RDMA_NL_IWPM_MAPINFO_NUM] = {.dump = iwpm_ack_mapping_info_cb}
+};
+
static int count_idrs(int id, void *p, void *data)
{
int *countp = data;
@@ -113,35 +123,49 @@ static int dump_qp(int id, void *p, void *data)
&qp->ep->com.local_addr;
struct sockaddr_in *rsin = (struct sockaddr_in *)
&qp->ep->com.remote_addr;
+ struct sockaddr_in *mapped_lsin = (struct sockaddr_in *)
+ &qp->ep->com.mapped_local_addr;
+ struct sockaddr_in *mapped_rsin = (struct sockaddr_in *)
+ &qp->ep->com.mapped_remote_addr;
cc = snprintf(qpd->buf + qpd->pos, space,
"rc qp sq id %u rq id %u state %u "
"onchip %u ep tid %u state %u "
- "%pI4:%u->%pI4:%u\n",
+ "%pI4:%u/%u->%pI4:%u/%u\n",
qp->wq.sq.qid, qp->wq.rq.qid,
(int)qp->attr.state,
qp->wq.sq.flags & T4_SQ_ONCHIP,
qp->ep->hwtid, (int)qp->ep->com.state,
&lsin->sin_addr, ntohs(lsin->sin_port),
- &rsin->sin_addr, ntohs(rsin->sin_port));
+ ntohs(mapped_lsin->sin_port),
+ &rsin->sin_addr, ntohs(rsin->sin_port),
+ ntohs(mapped_rsin->sin_port));
} else {
struct sockaddr_in6 *lsin6 = (struct sockaddr_in6 *)
&qp->ep->com.local_addr;
struct sockaddr_in6 *rsin6 = (struct sockaddr_in6 *)
&qp->ep->com.remote_addr;
+ struct sockaddr_in6 *mapped_lsin6 =
+ (struct sockaddr_in6 *)
+ &qp->ep->com.mapped_local_addr;
+ struct sockaddr_in6 *mapped_rsin6 =
+ (struct sockaddr_in6 *)
+ &qp->ep->com.mapped_remote_addr;
cc = snprintf(qpd->buf + qpd->pos, space,
"rc qp sq id %u rq id %u state %u "
"onchip %u ep tid %u state %u "
- "%pI6:%u->%pI6:%u\n",
+ "%pI6:%u/%u->%pI6:%u/%u\n",
qp->wq.sq.qid, qp->wq.rq.qid,
(int)qp->attr.state,
qp->wq.sq.flags & T4_SQ_ONCHIP,
qp->ep->hwtid, (int)qp->ep->com.state,
&lsin6->sin6_addr,
ntohs(lsin6->sin6_port),
+ ntohs(mapped_lsin6->sin6_port),
&rsin6->sin6_addr,
- ntohs(rsin6->sin6_port));
+ ntohs(rsin6->sin6_port),
+ ntohs(mapped_rsin6->sin6_port));
}
} else
cc = snprintf(qpd->buf + qpd->pos, space,
@@ -386,31 +410,43 @@ static int dump_ep(int id, void *p, void *data)
&ep->com.local_addr;
struct sockaddr_in *rsin = (struct sockaddr_in *)
&ep->com.remote_addr;
+ struct sockaddr_in *mapped_lsin = (struct sockaddr_in *)
+ &ep->com.mapped_local_addr;
+ struct sockaddr_in *mapped_rsin = (struct sockaddr_in *)
+ &ep->com.mapped_remote_addr;
cc = snprintf(epd->buf + epd->pos, space,
"ep %p cm_id %p qp %p state %d flags 0x%lx "
"history 0x%lx hwtid %d atid %d "
- "%pI4:%d <-> %pI4:%d\n",
+ "%pI4:%d/%d <-> %pI4:%d/%d\n",
ep, ep->com.cm_id, ep->com.qp,
(int)ep->com.state, ep->com.flags,
ep->com.history, ep->hwtid, ep->atid,
&lsin->sin_addr, ntohs(lsin->sin_port),
- &rsin->sin_addr, ntohs(rsin->sin_port));
+ ntohs(mapped_lsin->sin_port),
+ &rsin->sin_addr, ntohs(rsin->sin_port),
+ ntohs(mapped_rsin->sin_port));
} else {
struct sockaddr_in6 *lsin6 = (struct sockaddr_in6 *)
&ep->com.local_addr;
struct sockaddr_in6 *rsin6 = (struct sockaddr_in6 *)
&ep->com.remote_addr;
+ struct sockaddr_in6 *mapped_lsin6 = (struct sockaddr_in6 *)
+ &ep->com.mapped_local_addr;
+ struct sockaddr_in6 *mapped_rsin6 = (struct sockaddr_in6 *)
+ &ep->com.mapped_remote_addr;
cc = snprintf(epd->buf + epd->pos, space,
"ep %p cm_id %p qp %p state %d flags 0x%lx "
"history 0x%lx hwtid %d atid %d "
- "%pI6:%d <-> %pI6:%d\n",
+ "%pI6:%d/%d <-> %pI6:%d/%d\n",
ep, ep->com.cm_id, ep->com.qp,
(int)ep->com.state, ep->com.flags,
ep->com.history, ep->hwtid, ep->atid,
&lsin6->sin6_addr, ntohs(lsin6->sin6_port),
- &rsin6->sin6_addr, ntohs(rsin6->sin6_port));
+ ntohs(mapped_lsin6->sin6_port),
+ &rsin6->sin6_addr, ntohs(rsin6->sin6_port),
+ ntohs(mapped_rsin6->sin6_port));
}
if (cc < space)
epd->pos += cc;
@@ -431,23 +467,29 @@ static int dump_listen_ep(int id, void *p, void *data)
if (ep->com.local_addr.ss_family == AF_INET) {
struct sockaddr_in *lsin = (struct sockaddr_in *)
&ep->com.local_addr;
+ struct sockaddr_in *mapped_lsin = (struct sockaddr_in *)
+ &ep->com.mapped_local_addr;
cc = snprintf(epd->buf + epd->pos, space,
"ep %p cm_id %p state %d flags 0x%lx stid %d "
- "backlog %d %pI4:%d\n",
+ "backlog %d %pI4:%d/%d\n",
ep, ep->com.cm_id, (int)ep->com.state,
ep->com.flags, ep->stid, ep->backlog,
- &lsin->sin_addr, ntohs(lsin->sin_port));
+ &lsin->sin_addr, ntohs(lsin->sin_port),
+ ntohs(mapped_lsin->sin_port));
} else {
struct sockaddr_in6 *lsin6 = (struct sockaddr_in6 *)
&ep->com.local_addr;
+ struct sockaddr_in6 *mapped_lsin6 = (struct sockaddr_in6 *)
+ &ep->com.mapped_local_addr;
cc = snprintf(epd->buf + epd->pos, space,
"ep %p cm_id %p state %d flags 0x%lx stid %d "
- "backlog %d %pI6:%d\n",
+ "backlog %d %pI6:%d/%d\n",
ep, ep->com.cm_id, (int)ep->com.state,
ep->com.flags, ep->stid, ep->backlog,
- &lsin6->sin6_addr, ntohs(lsin6->sin6_port));
+ &lsin6->sin6_addr, ntohs(lsin6->sin6_port),
+ ntohs(mapped_lsin6->sin6_port));
}
if (cc < space)
epd->pos += cc;
@@ -687,6 +729,7 @@ static void c4iw_dealloc(struct uld_ctx *ctx)
if (ctx->dev->rdev.oc_mw_kva)
iounmap(ctx->dev->rdev.oc_mw_kva);
ib_dealloc_device(&ctx->dev->ibdev);
+ iwpm_exit(RDMA_NL_C4IW);
ctx->dev = NULL;
}
@@ -736,6 +779,7 @@ static struct c4iw_dev *c4iw_alloc(const struct cxgb4_lld_info *infop)
pci_resource_len(devp->rdev.lldi.pdev, 2));
if (!devp->rdev.bar2_kva) {
pr_err(MOD "Unable to ioremap BAR2\n");
+ ib_dealloc_device(&devp->ibdev);
return ERR_PTR(-EINVAL);
}
} else if (ocqp_supported(infop)) {
@@ -747,6 +791,7 @@ static struct c4iw_dev *c4iw_alloc(const struct cxgb4_lld_info *infop)
devp->rdev.lldi.vr->ocq.size);
if (!devp->rdev.oc_mw_kva) {
pr_err(MOD "Unable to ioremap onchip mem\n");
+ ib_dealloc_device(&devp->ibdev);
return ERR_PTR(-EINVAL);
}
}
@@ -780,6 +825,14 @@ static struct c4iw_dev *c4iw_alloc(const struct cxgb4_lld_info *infop)
c4iw_debugfs_root);
setup_debugfs(devp);
}
+
+ ret = iwpm_init(RDMA_NL_C4IW);
+ if (ret) {
+ pr_err("port mapper initialization failed with %d\n", ret);
+ ib_dealloc_device(&devp->ibdev);
+ return ERR_PTR(ret);
+ }
+
return devp;
}
@@ -1274,6 +1327,11 @@ static int __init c4iw_init_module(void)
printk(KERN_WARNING MOD
"could not create debugfs entry, continuing\n");
+ if (ibnl_add_client(RDMA_NL_C4IW, RDMA_NL_IWPM_NUM_OPS,
+ c4iw_nl_cb_table))
+ pr_err("%s[%u]: Failed to add netlink callback\n"
+ , __func__, __LINE__);
+
cxgb4_register_uld(CXGB4_ULD_RDMA, &c4iw_uld_info);
return 0;
@@ -1291,6 +1349,7 @@ static void __exit c4iw_exit_module(void)
}
mutex_unlock(&dev_mutex);
cxgb4_unregister_uld(CXGB4_ULD_RDMA);
+ ibnl_remove_client(RDMA_NL_C4IW);
c4iw_cm_term();
debugfs_remove_recursive(c4iw_debugfs_root);
}
diff --git a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
index 7474b490760..125bc5d1e17 100644
--- a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
+++ b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
@@ -52,6 +52,8 @@
#include <rdma/ib_verbs.h>
#include <rdma/iw_cm.h>
+#include <rdma/rdma_netlink.h>
+#include <rdma/iw_portmap.h>
#include "cxgb4.h"
#include "cxgb4_uld.h"
@@ -728,6 +730,7 @@ enum c4iw_ep_flags {
CLOSE_SENT = 3,
TIMEOUT = 4,
QP_REFERENCED = 5,
+ RELEASE_MAPINFO = 6,
};
enum c4iw_ep_history {
@@ -764,6 +767,8 @@ struct c4iw_ep_common {
struct mutex mutex;
struct sockaddr_storage local_addr;
struct sockaddr_storage remote_addr;
+ struct sockaddr_storage mapped_local_addr;
+ struct sockaddr_storage mapped_remote_addr;
struct c4iw_wr_wait wr_wait;
unsigned long flags;
unsigned long history;
@@ -805,7 +810,48 @@ struct c4iw_ep {
u8 retry_with_mpa_v1;
u8 tried_with_mpa_v1;
unsigned int retry_count;
-};
+ int snd_win;
+ int rcv_win;
+};
+
+static inline void print_addr(struct c4iw_ep_common *epc, const char *func,
+ const char *msg)
+{
+
+#define SINA(a) (&(((struct sockaddr_in *)(a))->sin_addr.s_addr))
+#define SINP(a) ntohs(((struct sockaddr_in *)(a))->sin_port)
+#define SIN6A(a) (&(((struct sockaddr_in6 *)(a))->sin6_addr))
+#define SIN6P(a) ntohs(((struct sockaddr_in6 *)(a))->sin6_port)
+
+ if (c4iw_debug) {
+ switch (epc->local_addr.ss_family) {
+ case AF_INET:
+ PDBG("%s %s %pI4:%u/%u <-> %pI4:%u/%u\n",
+ func, msg, SINA(&epc->local_addr),
+ SINP(&epc->local_addr),
+ SINP(&epc->mapped_local_addr),
+ SINA(&epc->remote_addr),
+ SINP(&epc->remote_addr),
+ SINP(&epc->mapped_remote_addr));
+ break;
+ case AF_INET6:
+ PDBG("%s %s %pI6:%u/%u <-> %pI6:%u/%u\n",
+ func, msg, SIN6A(&epc->local_addr),
+ SIN6P(&epc->local_addr),
+ SIN6P(&epc->mapped_local_addr),
+ SIN6A(&epc->remote_addr),
+ SIN6P(&epc->remote_addr),
+ SIN6P(&epc->mapped_remote_addr));
+ break;
+ default:
+ break;
+ }
+ }
+#undef SINA
+#undef SINP
+#undef SIN6A
+#undef SIN6P
+}
static inline struct c4iw_ep *to_ep(struct iw_cm_id *cm_id)
{
diff --git a/drivers/infiniband/hw/cxgb4/provider.c b/drivers/infiniband/hw/cxgb4/provider.c
index a94a3e12c34..b1d305338de 100644
--- a/drivers/infiniband/hw/cxgb4/provider.c
+++ b/drivers/infiniband/hw/cxgb4/provider.c
@@ -122,7 +122,7 @@ static struct ib_ucontext *c4iw_alloc_ucontext(struct ib_device *ibdev,
INIT_LIST_HEAD(&context->mmaps);
spin_lock_init(&context->mmap_lock);
- if (udata->outlen < sizeof(uresp)) {
+ if (udata->outlen < sizeof(uresp) - sizeof(uresp.reserved)) {
if (!warned++)
pr_err(MOD "Warning - downlevel libcxgb4 (non-fatal), device status page disabled.");
rhp->rdev.flags |= T4_STATUS_PAGE_DISABLED;
@@ -140,7 +140,8 @@ static struct ib_ucontext *c4iw_alloc_ucontext(struct ib_device *ibdev,
context->key += PAGE_SIZE;
spin_unlock(&context->mmap_lock);
- ret = ib_copy_to_udata(udata, &uresp, sizeof(uresp));
+ ret = ib_copy_to_udata(udata, &uresp,
+ sizeof(uresp) - sizeof(uresp.reserved));
if (ret)
goto err_mm;
@@ -499,7 +500,7 @@ int c4iw_register_device(struct c4iw_dev *dev)
dev->ibdev.node_type = RDMA_NODE_RNIC;
memcpy(dev->ibdev.node_desc, C4IW_NODE_DESC, sizeof(C4IW_NODE_DESC));
dev->ibdev.phys_port_cnt = dev->rdev.lldi.nports;
- dev->ibdev.num_comp_vectors = 1;
+ dev->ibdev.num_comp_vectors = dev->rdev.lldi.nciq;
dev->ibdev.dma_device = &(dev->rdev.lldi.pdev->dev);
dev->ibdev.query_device = c4iw_query_device;
dev->ibdev.query_port = c4iw_query_port;
diff --git a/drivers/infiniband/hw/cxgb4/t4.h b/drivers/infiniband/hw/cxgb4/t4.h
index 2178f319841..68b0a6bf4eb 100644
--- a/drivers/infiniband/hw/cxgb4/t4.h
+++ b/drivers/infiniband/hw/cxgb4/t4.h
@@ -542,6 +542,7 @@ struct t4_cq {
size_t memsize;
__be64 bits_type_ts;
u32 cqid;
+ int vector;
u16 size; /* including status page */
u16 cidx;
u16 sw_pidx;
diff --git a/drivers/infiniband/hw/cxgb4/t4fw_ri_api.h b/drivers/infiniband/hw/cxgb4/t4fw_ri_api.h
index 6121ca08fe5..91289a051af 100644
--- a/drivers/infiniband/hw/cxgb4/t4fw_ri_api.h
+++ b/drivers/infiniband/hw/cxgb4/t4fw_ri_api.h
@@ -848,6 +848,7 @@ enum { /* TCP congestion control algorithms */
#define V_CONG_CNTRL(x) ((x) << S_CONG_CNTRL)
#define G_CONG_CNTRL(x) (((x) >> S_CONG_CNTRL) & M_CONG_CNTRL)
+#define CONG_CNTRL_VALID (1 << 18)
#define T5_OPT_2_VALID (1 << 31)
#endif /* _T4FW_RI_API_H_ */
diff --git a/drivers/infiniband/hw/cxgb4/user.h b/drivers/infiniband/hw/cxgb4/user.h
index 11ccd276e5d..cbd0ce17072 100644
--- a/drivers/infiniband/hw/cxgb4/user.h
+++ b/drivers/infiniband/hw/cxgb4/user.h
@@ -48,6 +48,7 @@ struct c4iw_create_cq_resp {
__u32 cqid;
__u32 size;
__u32 qid_mask;
+ __u32 reserved; /* explicit padding (optional for i386) */
};
@@ -74,5 +75,6 @@ struct c4iw_create_qp_resp {
struct c4iw_alloc_ucontext_resp {
__u64 status_page_key;
__u32 status_page_size;
+ __u32 reserved; /* explicit padding (optional for i386) */
};
#endif
diff --git a/drivers/infiniband/hw/ipath/ipath_diag.c b/drivers/infiniband/hw/ipath/ipath_diag.c
index e2f9a51f4a3..45802e97332 100644
--- a/drivers/infiniband/hw/ipath/ipath_diag.c
+++ b/drivers/infiniband/hw/ipath/ipath_diag.c
@@ -346,6 +346,10 @@ static ssize_t ipath_diagpkt_write(struct file *fp,
ret = -EFAULT;
goto bail;
}
+ dp.len = odp.len;
+ dp.unit = odp.unit;
+ dp.data = odp.data;
+ dp.pbc_wd = 0;
} else {
ret = -EINVAL;
goto bail;
diff --git a/drivers/infiniband/hw/ipath/ipath_intr.c b/drivers/infiniband/hw/ipath/ipath_intr.c
index 26dfbc8ee0f..01ba792791a 100644
--- a/drivers/infiniband/hw/ipath/ipath_intr.c
+++ b/drivers/infiniband/hw/ipath/ipath_intr.c
@@ -70,7 +70,7 @@ void ipath_disarm_senderrbufs(struct ipath_devdata *dd)
if (sbuf[0] || sbuf[1] || (piobcnt > 128 && (sbuf[2] || sbuf[3]))) {
int i;
if (ipath_debug & (__IPATH_PKTDBG|__IPATH_DBG) &&
- dd->ipath_lastcancel > jiffies) {
+ time_after(dd->ipath_lastcancel, jiffies)) {
__IPATH_DBG_WHICH(__IPATH_PKTDBG|__IPATH_DBG,
"SendbufErrs %lx %lx", sbuf[0],
sbuf[1]);
@@ -755,7 +755,7 @@ static int handle_errors(struct ipath_devdata *dd, ipath_err_t errs)
/* likely due to cancel; so suppress message unless verbose */
if ((errs & (INFINIPATH_E_SPKTLEN | INFINIPATH_E_SPIOARMLAUNCH)) &&
- dd->ipath_lastcancel > jiffies) {
+ time_after(dd->ipath_lastcancel, jiffies)) {
/* armlaunch takes precedence; it often causes both. */
ipath_cdbg(VERBOSE,
"Suppressed %s error (%llx) after sendbuf cancel\n",
diff --git a/drivers/infiniband/hw/ipath/ipath_sdma.c b/drivers/infiniband/hw/ipath/ipath_sdma.c
index 98ac18ec977..17a517766ad 100644
--- a/drivers/infiniband/hw/ipath/ipath_sdma.c
+++ b/drivers/infiniband/hw/ipath/ipath_sdma.c
@@ -247,7 +247,7 @@ static void sdma_abort_task(unsigned long opaque)
/* ipath_sdma_abort() is done, waiting for interrupt */
if (status == IPATH_SDMA_ABORT_DISARMED) {
- if (jiffies < dd->ipath_sdma_abort_intr_timeout)
+ if (time_before(jiffies, dd->ipath_sdma_abort_intr_timeout))
goto resched_noprint;
/* give up, intr got lost somewhere */
ipath_dbg("give up waiting for SDMADISABLED intr\n");
@@ -341,7 +341,7 @@ resched:
* JAG - this is bad to just have default be a loop without
* state change
*/
- if (jiffies > dd->ipath_sdma_abort_jiffies) {
+ if (time_after(jiffies, dd->ipath_sdma_abort_jiffies)) {
ipath_dbg("looping with status 0x%08lx\n",
dd->ipath_sdma_status);
dd->ipath_sdma_abort_jiffies = jiffies + 5 * HZ;
diff --git a/drivers/infiniband/hw/mlx4/ah.c b/drivers/infiniband/hw/mlx4/ah.c
index 170dca60804..2d8c3397774 100644
--- a/drivers/infiniband/hw/mlx4/ah.c
+++ b/drivers/infiniband/hw/mlx4/ah.c
@@ -73,7 +73,7 @@ static struct ib_ah *create_iboe_ah(struct ib_pd *pd, struct ib_ah_attr *ah_attr
{
struct mlx4_ib_dev *ibdev = to_mdev(pd->device);
struct mlx4_dev *dev = ibdev->dev;
- int is_mcast;
+ int is_mcast = 0;
struct in6_addr in6;
u16 vlan_tag;
diff --git a/drivers/infiniband/hw/mlx4/cq.c b/drivers/infiniband/hw/mlx4/cq.c
index 5f640814cc8..1066eec854a 100644
--- a/drivers/infiniband/hw/mlx4/cq.c
+++ b/drivers/infiniband/hw/mlx4/cq.c
@@ -102,7 +102,7 @@ static int mlx4_ib_alloc_cq_buf(struct mlx4_ib_dev *dev, struct mlx4_ib_cq_buf *
int err;
err = mlx4_buf_alloc(dev->dev, nent * dev->dev->caps.cqe_size,
- PAGE_SIZE * 2, &buf->buf);
+ PAGE_SIZE * 2, &buf->buf, GFP_KERNEL);
if (err)
goto out;
@@ -113,7 +113,7 @@ static int mlx4_ib_alloc_cq_buf(struct mlx4_ib_dev *dev, struct mlx4_ib_cq_buf *
if (err)
goto err_buf;
- err = mlx4_buf_write_mtt(dev->dev, &buf->mtt, &buf->buf);
+ err = mlx4_buf_write_mtt(dev->dev, &buf->mtt, &buf->buf, GFP_KERNEL);
if (err)
goto err_mtt;
@@ -209,7 +209,7 @@ struct ib_cq *mlx4_ib_create_cq(struct ib_device *ibdev, int entries, int vector
uar = &to_mucontext(context)->uar;
} else {
- err = mlx4_db_alloc(dev->dev, &cq->db, 1);
+ err = mlx4_db_alloc(dev->dev, &cq->db, 1, GFP_KERNEL);
if (err)
goto err_cq;
diff --git a/drivers/infiniband/hw/mlx4/mad.c b/drivers/infiniband/hw/mlx4/mad.c
index fd36ec67263..287ad0564ac 100644
--- a/drivers/infiniband/hw/mlx4/mad.c
+++ b/drivers/infiniband/hw/mlx4/mad.c
@@ -478,10 +478,6 @@ int mlx4_ib_send_to_slave(struct mlx4_ib_dev *dev, int slave, u8 port,
if (!tun_ctx || tun_ctx->state != DEMUX_PV_STATE_ACTIVE)
return -EAGAIN;
- /* QP0 forwarding only for Dom0 */
- if (!dest_qpt && (mlx4_master_func_num(dev->dev) != slave))
- return -EINVAL;
-
if (!dest_qpt)
tun_qp = &tun_ctx->qp[0];
else
@@ -667,6 +663,21 @@ static int mlx4_ib_demux_mad(struct ib_device *ibdev, u8 port,
}
/* Class-specific handling */
switch (mad->mad_hdr.mgmt_class) {
+ case IB_MGMT_CLASS_SUBN_LID_ROUTED:
+ case IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE:
+ /* 255 indicates the dom0 */
+ if (slave != 255 && slave != mlx4_master_func_num(dev->dev)) {
+ if (!mlx4_vf_smi_enabled(dev->dev, slave, port))
+ return -EPERM;
+ /* for a VF. drop unsolicited MADs */
+ if (!(mad->mad_hdr.method & IB_MGMT_METHOD_RESP)) {
+ mlx4_ib_warn(ibdev, "demux QP0. rejecting unsolicited mad for slave %d class 0x%x, method 0x%x\n",
+ slave, mad->mad_hdr.mgmt_class,
+ mad->mad_hdr.method);
+ return -EINVAL;
+ }
+ }
+ break;
case IB_MGMT_CLASS_SUBN_ADM:
if (mlx4_ib_demux_sa_handler(ibdev, port, slave,
(struct ib_sa_mad *) mad))
@@ -1165,10 +1176,6 @@ int mlx4_ib_send_to_wire(struct mlx4_ib_dev *dev, int slave, u8 port,
if (!sqp_ctx || sqp_ctx->state != DEMUX_PV_STATE_ACTIVE)
return -EAGAIN;
- /* QP0 forwarding only for Dom0 */
- if (dest_qpt == IB_QPT_SMI && (mlx4_master_func_num(dev->dev) != slave))
- return -EINVAL;
-
if (dest_qpt == IB_QPT_SMI) {
src_qpnum = 0;
sqp = &sqp_ctx->qp[0];
@@ -1285,11 +1292,6 @@ static void mlx4_ib_multiplex_mad(struct mlx4_ib_demux_pv_ctx *ctx, struct ib_wc
"belongs to another slave\n", wc->src_qp);
return;
}
- if (slave != mlx4_master_func_num(dev->dev) && !(wc->src_qp & 0x2)) {
- mlx4_ib_warn(ctx->ib_dev, "can't multiplex bad sqp:%d: "
- "non-master trying to send QP0 packets\n", wc->src_qp);
- return;
- }
/* Map transaction ID */
ib_dma_sync_single_for_cpu(ctx->ib_dev, tun_qp->ring[wr_ix].map,
@@ -1317,6 +1319,12 @@ static void mlx4_ib_multiplex_mad(struct mlx4_ib_demux_pv_ctx *ctx, struct ib_wc
/* Class-specific handling */
switch (tunnel->mad.mad_hdr.mgmt_class) {
+ case IB_MGMT_CLASS_SUBN_LID_ROUTED:
+ case IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE:
+ if (slave != mlx4_master_func_num(dev->dev) &&
+ !mlx4_vf_smi_enabled(dev->dev, slave, ctx->port))
+ return;
+ break;
case IB_MGMT_CLASS_SUBN_ADM:
if (mlx4_ib_multiplex_sa_handler(ctx->ib_dev, ctx->port, slave,
(struct ib_sa_mad *) &tunnel->mad))
@@ -1749,9 +1757,9 @@ static int create_pv_resources(struct ib_device *ibdev, int slave, int port,
return -EEXIST;
ctx->state = DEMUX_PV_STATE_STARTING;
- /* have QP0 only on port owner, and only if link layer is IB */
- if (ctx->slave == mlx4_master_func_num(to_mdev(ctx->ib_dev)->dev) &&
- rdma_port_get_link_layer(ibdev, ctx->port) == IB_LINK_LAYER_INFINIBAND)
+ /* have QP0 only if link layer is IB */
+ if (rdma_port_get_link_layer(ibdev, ctx->port) ==
+ IB_LINK_LAYER_INFINIBAND)
ctx->has_smi = 1;
if (ctx->has_smi) {
diff --git a/drivers/infiniband/hw/mlx4/main.c b/drivers/infiniband/hw/mlx4/main.c
index 1b6dbe156a3..0f7027e7db1 100644
--- a/drivers/infiniband/hw/mlx4/main.c
+++ b/drivers/infiniband/hw/mlx4/main.c
@@ -48,6 +48,7 @@
#include <linux/mlx4/driver.h>
#include <linux/mlx4/cmd.h>
+#include <linux/mlx4/qp.h>
#include "mlx4_ib.h"
#include "user.h"
@@ -544,12 +545,11 @@ static int mlx4_ib_modify_device(struct ib_device *ibdev, int mask,
return 0;
}
-static int mlx4_SET_PORT(struct mlx4_ib_dev *dev, u8 port, int reset_qkey_viols,
- u32 cap_mask)
+static int mlx4_ib_SET_PORT(struct mlx4_ib_dev *dev, u8 port, int reset_qkey_viols,
+ u32 cap_mask)
{
struct mlx4_cmd_mailbox *mailbox;
int err;
- u8 is_eth = dev->dev->caps.port_type[port] == MLX4_PORT_TYPE_ETH;
mailbox = mlx4_alloc_cmd_mailbox(dev->dev);
if (IS_ERR(mailbox))
@@ -563,8 +563,8 @@ static int mlx4_SET_PORT(struct mlx4_ib_dev *dev, u8 port, int reset_qkey_viols,
((__be32 *) mailbox->buf)[1] = cpu_to_be32(cap_mask);
}
- err = mlx4_cmd(dev->dev, mailbox->dma, port, is_eth, MLX4_CMD_SET_PORT,
- MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+ err = mlx4_cmd(dev->dev, mailbox->dma, port, 0, MLX4_CMD_SET_PORT,
+ MLX4_CMD_TIME_CLASS_B, MLX4_CMD_WRAPPED);
mlx4_free_cmd_mailbox(dev->dev, mailbox);
return err;
@@ -573,11 +573,20 @@ static int mlx4_SET_PORT(struct mlx4_ib_dev *dev, u8 port, int reset_qkey_viols,
static int mlx4_ib_modify_port(struct ib_device *ibdev, u8 port, int mask,
struct ib_port_modify *props)
{
+ struct mlx4_ib_dev *mdev = to_mdev(ibdev);
+ u8 is_eth = mdev->dev->caps.port_type[port] == MLX4_PORT_TYPE_ETH;
struct ib_port_attr attr;
u32 cap_mask;
int err;
- mutex_lock(&to_mdev(ibdev)->cap_mask_mutex);
+ /* return OK if this is RoCE. CM calls ib_modify_port() regardless
+ * of whether port link layer is ETH or IB. For ETH ports, qkey
+ * violations and port capabilities are not meaningful.
+ */
+ if (is_eth)
+ return 0;
+
+ mutex_lock(&mdev->cap_mask_mutex);
err = mlx4_ib_query_port(ibdev, port, &attr);
if (err)
@@ -586,9 +595,9 @@ static int mlx4_ib_modify_port(struct ib_device *ibdev, u8 port, int mask,
cap_mask = (attr.port_cap_flags | props->set_port_cap_mask) &
~props->clr_port_cap_mask;
- err = mlx4_SET_PORT(to_mdev(ibdev), port,
- !!(mask & IB_PORT_RESET_QKEY_CNTR),
- cap_mask);
+ err = mlx4_ib_SET_PORT(mdev, port,
+ !!(mask & IB_PORT_RESET_QKEY_CNTR),
+ cap_mask);
out:
mutex_unlock(&to_mdev(ibdev)->cap_mask_mutex);
@@ -1614,6 +1623,53 @@ static int mlx4_ib_inet6_event(struct notifier_block *this, unsigned long event,
}
#endif
+#define MLX4_IB_INVALID_MAC ((u64)-1)
+static void mlx4_ib_update_qps(struct mlx4_ib_dev *ibdev,
+ struct net_device *dev,
+ int port)
+{
+ u64 new_smac = 0;
+ u64 release_mac = MLX4_IB_INVALID_MAC;
+ struct mlx4_ib_qp *qp;
+
+ read_lock(&dev_base_lock);
+ new_smac = mlx4_mac_to_u64(dev->dev_addr);
+ read_unlock(&dev_base_lock);
+
+ mutex_lock(&ibdev->qp1_proxy_lock[port - 1]);
+ qp = ibdev->qp1_proxy[port - 1];
+ if (qp) {
+ int new_smac_index;
+ u64 old_smac = qp->pri.smac;
+ struct mlx4_update_qp_params update_params;
+
+ if (new_smac == old_smac)
+ goto unlock;
+
+ new_smac_index = mlx4_register_mac(ibdev->dev, port, new_smac);
+
+ if (new_smac_index < 0)
+ goto unlock;
+
+ update_params.smac_index = new_smac_index;
+ if (mlx4_update_qp(ibdev->dev, &qp->mqp, MLX4_UPDATE_QP_SMAC,
+ &update_params)) {
+ release_mac = new_smac;
+ goto unlock;
+ }
+
+ qp->pri.smac = new_smac;
+ qp->pri.smac_index = new_smac_index;
+
+ release_mac = old_smac;
+ }
+
+unlock:
+ mutex_unlock(&ibdev->qp1_proxy_lock[port - 1]);
+ if (release_mac != MLX4_IB_INVALID_MAC)
+ mlx4_unregister_mac(ibdev->dev, port, release_mac);
+}
+
static void mlx4_ib_get_dev_addr(struct net_device *dev,
struct mlx4_ib_dev *ibdev, u8 port)
{
@@ -1689,9 +1745,13 @@ static int mlx4_ib_init_gid_table(struct mlx4_ib_dev *ibdev)
return 0;
}
-static void mlx4_ib_scan_netdevs(struct mlx4_ib_dev *ibdev)
+static void mlx4_ib_scan_netdevs(struct mlx4_ib_dev *ibdev,
+ struct net_device *dev,
+ unsigned long event)
+
{
struct mlx4_ib_iboe *iboe;
+ int update_qps_port = -1;
int port;
iboe = &ibdev->iboe;
@@ -1719,6 +1779,11 @@ static void mlx4_ib_scan_netdevs(struct mlx4_ib_dev *ibdev)
}
curr_master = iboe->masters[port - 1];
+ if (dev == iboe->netdevs[port - 1] &&
+ (event == NETDEV_CHANGEADDR || event == NETDEV_REGISTER ||
+ event == NETDEV_UP || event == NETDEV_CHANGE))
+ update_qps_port = port;
+
if (curr_netdev) {
port_state = (netif_running(curr_netdev) && netif_carrier_ok(curr_netdev)) ?
IB_PORT_ACTIVE : IB_PORT_DOWN;
@@ -1752,6 +1817,9 @@ static void mlx4_ib_scan_netdevs(struct mlx4_ib_dev *ibdev)
}
spin_unlock(&iboe->lock);
+
+ if (update_qps_port > 0)
+ mlx4_ib_update_qps(ibdev, dev, update_qps_port);
}
static int mlx4_ib_netdev_event(struct notifier_block *this,
@@ -1764,7 +1832,7 @@ static int mlx4_ib_netdev_event(struct notifier_block *this,
return NOTIFY_DONE;
ibdev = container_of(this, struct mlx4_ib_dev, iboe.nb);
- mlx4_ib_scan_netdevs(ibdev);
+ mlx4_ib_scan_netdevs(ibdev, dev, event);
return NOTIFY_DONE;
}
@@ -2043,6 +2111,7 @@ static void *mlx4_ib_add(struct mlx4_dev *dev)
goto err_map;
for (i = 0; i < ibdev->num_ports; ++i) {
+ mutex_init(&ibdev->qp1_proxy_lock[i]);
if (mlx4_ib_port_link_layer(&ibdev->ib_dev, i + 1) ==
IB_LINK_LAYER_ETHERNET) {
err = mlx4_counter_alloc(ibdev->dev, &ibdev->counters[i]);
@@ -2126,7 +2195,7 @@ static void *mlx4_ib_add(struct mlx4_dev *dev)
for (i = 1 ; i <= ibdev->num_ports ; ++i)
reset_gid_table(ibdev, i);
rtnl_lock();
- mlx4_ib_scan_netdevs(ibdev);
+ mlx4_ib_scan_netdevs(ibdev, NULL, 0);
rtnl_unlock();
mlx4_ib_init_gid_table(ibdev);
}
diff --git a/drivers/infiniband/hw/mlx4/mlx4_ib.h b/drivers/infiniband/hw/mlx4/mlx4_ib.h
index f589522fddf..369da3ca5d6 100644
--- a/drivers/infiniband/hw/mlx4/mlx4_ib.h
+++ b/drivers/infiniband/hw/mlx4/mlx4_ib.h
@@ -156,6 +156,7 @@ enum mlx4_ib_qp_flags {
MLX4_IB_QP_LSO = IB_QP_CREATE_IPOIB_UD_LSO,
MLX4_IB_QP_BLOCK_MULTICAST_LOOPBACK = IB_QP_CREATE_BLOCK_MULTICAST_LOOPBACK,
MLX4_IB_QP_NETIF = IB_QP_CREATE_NETIF_QP,
+ MLX4_IB_QP_CREATE_USE_GFP_NOIO = IB_QP_CREATE_USE_GFP_NOIO,
MLX4_IB_SRIOV_TUNNEL_QP = 1 << 30,
MLX4_IB_SRIOV_SQP = 1 << 31,
};
@@ -522,6 +523,9 @@ struct mlx4_ib_dev {
int steer_qpn_count;
int steer_qpn_base;
int steering_support;
+ struct mlx4_ib_qp *qp1_proxy[MLX4_MAX_PORTS];
+ /* lock when destroying qp1_proxy and getting netdev events */
+ struct mutex qp1_proxy_lock[MLX4_MAX_PORTS];
};
struct ib_event_work {
diff --git a/drivers/infiniband/hw/mlx4/qp.c b/drivers/infiniband/hw/mlx4/qp.c
index 41308af4163..67780452f0c 100644
--- a/drivers/infiniband/hw/mlx4/qp.c
+++ b/drivers/infiniband/hw/mlx4/qp.c
@@ -608,9 +608,20 @@ static int qp_has_rq(struct ib_qp_init_attr *attr)
return !attr->srq;
}
+static int qp0_enabled_vf(struct mlx4_dev *dev, int qpn)
+{
+ int i;
+ for (i = 0; i < dev->caps.num_ports; i++) {
+ if (qpn == dev->caps.qp0_proxy[i])
+ return !!dev->caps.qp0_qkey[i];
+ }
+ return 0;
+}
+
static int create_qp_common(struct mlx4_ib_dev *dev, struct ib_pd *pd,
struct ib_qp_init_attr *init_attr,
- struct ib_udata *udata, int sqpn, struct mlx4_ib_qp **caller_qp)
+ struct ib_udata *udata, int sqpn, struct mlx4_ib_qp **caller_qp,
+ gfp_t gfp)
{
int qpn;
int err;
@@ -625,10 +636,13 @@ static int create_qp_common(struct mlx4_ib_dev *dev, struct ib_pd *pd,
!(init_attr->create_flags & MLX4_IB_SRIOV_SQP))) {
if (init_attr->qp_type == IB_QPT_GSI)
qp_type = MLX4_IB_QPT_PROXY_GSI;
- else if (mlx4_is_master(dev->dev))
- qp_type = MLX4_IB_QPT_PROXY_SMI_OWNER;
- else
- qp_type = MLX4_IB_QPT_PROXY_SMI;
+ else {
+ if (mlx4_is_master(dev->dev) ||
+ qp0_enabled_vf(dev->dev, sqpn))
+ qp_type = MLX4_IB_QPT_PROXY_SMI_OWNER;
+ else
+ qp_type = MLX4_IB_QPT_PROXY_SMI;
+ }
}
qpn = sqpn;
/* add extra sg entry for tunneling */
@@ -643,7 +657,9 @@ static int create_qp_common(struct mlx4_ib_dev *dev, struct ib_pd *pd,
return -EINVAL;
if (tnl_init->proxy_qp_type == IB_QPT_GSI)
qp_type = MLX4_IB_QPT_TUN_GSI;
- else if (tnl_init->slave == mlx4_master_func_num(dev->dev))
+ else if (tnl_init->slave == mlx4_master_func_num(dev->dev) ||
+ mlx4_vf_smi_enabled(dev->dev, tnl_init->slave,
+ tnl_init->port))
qp_type = MLX4_IB_QPT_TUN_SMI_OWNER;
else
qp_type = MLX4_IB_QPT_TUN_SMI;
@@ -658,14 +674,14 @@ static int create_qp_common(struct mlx4_ib_dev *dev, struct ib_pd *pd,
if (qp_type == MLX4_IB_QPT_SMI || qp_type == MLX4_IB_QPT_GSI ||
(qp_type & (MLX4_IB_QPT_PROXY_SMI | MLX4_IB_QPT_PROXY_SMI_OWNER |
MLX4_IB_QPT_PROXY_GSI | MLX4_IB_QPT_TUN_SMI_OWNER))) {
- sqp = kzalloc(sizeof (struct mlx4_ib_sqp), GFP_KERNEL);
+ sqp = kzalloc(sizeof (struct mlx4_ib_sqp), gfp);
if (!sqp)
return -ENOMEM;
qp = &sqp->qp;
qp->pri.vid = 0xFFFF;
qp->alt.vid = 0xFFFF;
} else {
- qp = kzalloc(sizeof (struct mlx4_ib_qp), GFP_KERNEL);
+ qp = kzalloc(sizeof (struct mlx4_ib_qp), gfp);
if (!qp)
return -ENOMEM;
qp->pri.vid = 0xFFFF;
@@ -748,14 +764,14 @@ static int create_qp_common(struct mlx4_ib_dev *dev, struct ib_pd *pd,
goto err;
if (qp_has_rq(init_attr)) {
- err = mlx4_db_alloc(dev->dev, &qp->db, 0);
+ err = mlx4_db_alloc(dev->dev, &qp->db, 0, gfp);
if (err)
goto err;
*qp->db.db = 0;
}
- if (mlx4_buf_alloc(dev->dev, qp->buf_size, PAGE_SIZE * 2, &qp->buf)) {
+ if (mlx4_buf_alloc(dev->dev, qp->buf_size, PAGE_SIZE * 2, &qp->buf, gfp)) {
err = -ENOMEM;
goto err_db;
}
@@ -765,13 +781,12 @@ static int create_qp_common(struct mlx4_ib_dev *dev, struct ib_pd *pd,
if (err)
goto err_buf;
- err = mlx4_buf_write_mtt(dev->dev, &qp->mtt, &qp->buf);
+ err = mlx4_buf_write_mtt(dev->dev, &qp->mtt, &qp->buf, gfp);
if (err)
goto err_mtt;
- qp->sq.wrid = kmalloc(qp->sq.wqe_cnt * sizeof (u64), GFP_KERNEL);
- qp->rq.wrid = kmalloc(qp->rq.wqe_cnt * sizeof (u64), GFP_KERNEL);
-
+ qp->sq.wrid = kmalloc(qp->sq.wqe_cnt * sizeof (u64), gfp);
+ qp->rq.wrid = kmalloc(qp->rq.wqe_cnt * sizeof (u64), gfp);
if (!qp->sq.wrid || !qp->rq.wrid) {
err = -ENOMEM;
goto err_wrid;
@@ -801,7 +816,7 @@ static int create_qp_common(struct mlx4_ib_dev *dev, struct ib_pd *pd,
goto err_proxy;
}
- err = mlx4_qp_alloc(dev->dev, qpn, &qp->mqp);
+ err = mlx4_qp_alloc(dev->dev, qpn, &qp->mqp, gfp);
if (err)
goto err_qpn;
@@ -1040,7 +1055,10 @@ struct ib_qp *mlx4_ib_create_qp(struct ib_pd *pd,
struct mlx4_ib_qp *qp = NULL;
int err;
u16 xrcdn = 0;
+ gfp_t gfp;
+ gfp = (init_attr->create_flags & MLX4_IB_QP_CREATE_USE_GFP_NOIO) ?
+ GFP_NOIO : GFP_KERNEL;
/*
* We only support LSO, vendor flag1, and multicast loopback blocking,
* and only for kernel UD QPs.
@@ -1049,7 +1067,8 @@ struct ib_qp *mlx4_ib_create_qp(struct ib_pd *pd,
MLX4_IB_QP_BLOCK_MULTICAST_LOOPBACK |
MLX4_IB_SRIOV_TUNNEL_QP |
MLX4_IB_SRIOV_SQP |
- MLX4_IB_QP_NETIF))
+ MLX4_IB_QP_NETIF |
+ MLX4_IB_QP_CREATE_USE_GFP_NOIO))
return ERR_PTR(-EINVAL);
if (init_attr->create_flags & IB_QP_CREATE_NETIF_QP) {
@@ -1059,7 +1078,7 @@ struct ib_qp *mlx4_ib_create_qp(struct ib_pd *pd,
if (init_attr->create_flags &&
(udata ||
- ((init_attr->create_flags & ~MLX4_IB_SRIOV_SQP) &&
+ ((init_attr->create_flags & ~(MLX4_IB_SRIOV_SQP | MLX4_IB_QP_CREATE_USE_GFP_NOIO)) &&
init_attr->qp_type != IB_QPT_UD) ||
((init_attr->create_flags & MLX4_IB_SRIOV_SQP) &&
init_attr->qp_type > IB_QPT_GSI)))
@@ -1079,7 +1098,7 @@ struct ib_qp *mlx4_ib_create_qp(struct ib_pd *pd,
case IB_QPT_RC:
case IB_QPT_UC:
case IB_QPT_RAW_PACKET:
- qp = kzalloc(sizeof *qp, GFP_KERNEL);
+ qp = kzalloc(sizeof *qp, gfp);
if (!qp)
return ERR_PTR(-ENOMEM);
qp->pri.vid = 0xFFFF;
@@ -1088,7 +1107,7 @@ struct ib_qp *mlx4_ib_create_qp(struct ib_pd *pd,
case IB_QPT_UD:
{
err = create_qp_common(to_mdev(pd->device), pd, init_attr,
- udata, 0, &qp);
+ udata, 0, &qp, gfp);
if (err)
return ERR_PTR(err);
@@ -1106,7 +1125,7 @@ struct ib_qp *mlx4_ib_create_qp(struct ib_pd *pd,
err = create_qp_common(to_mdev(pd->device), pd, init_attr, udata,
get_sqp_num(to_mdev(pd->device), init_attr),
- &qp);
+ &qp, gfp);
if (err)
return ERR_PTR(err);
@@ -1132,6 +1151,12 @@ int mlx4_ib_destroy_qp(struct ib_qp *qp)
if (is_qp0(dev, mqp))
mlx4_CLOSE_PORT(dev->dev, mqp->port);
+ if (dev->qp1_proxy[mqp->port - 1] == mqp) {
+ mutex_lock(&dev->qp1_proxy_lock[mqp->port - 1]);
+ dev->qp1_proxy[mqp->port - 1] = NULL;
+ mutex_unlock(&dev->qp1_proxy_lock[mqp->port - 1]);
+ }
+
pd = get_pd(mqp);
destroy_qp_common(dev, mqp, !!pd->ibpd.uobject);
@@ -1646,6 +1671,8 @@ static int __mlx4_ib_modify_qp(struct ib_qp *ibqp,
err = handle_eth_ud_smac_index(dev, qp, (u8 *)attr->smac, context);
if (err)
return -EINVAL;
+ if (qp->mlx4_ib_qp_type == MLX4_IB_QPT_PROXY_GSI)
+ dev->qp1_proxy[qp->port - 1] = qp;
}
}
}
@@ -1930,6 +1957,19 @@ out:
return err;
}
+static int vf_get_qp0_qkey(struct mlx4_dev *dev, int qpn, u32 *qkey)
+{
+ int i;
+ for (i = 0; i < dev->caps.num_ports; i++) {
+ if (qpn == dev->caps.qp0_proxy[i] ||
+ qpn == dev->caps.qp0_tunnel[i]) {
+ *qkey = dev->caps.qp0_qkey[i];
+ return 0;
+ }
+ }
+ return -EINVAL;
+}
+
static int build_sriov_qp0_header(struct mlx4_ib_sqp *sqp,
struct ib_send_wr *wr,
void *wqe, unsigned *mlx_seg_len)
@@ -1987,8 +2027,13 @@ static int build_sriov_qp0_header(struct mlx4_ib_sqp *sqp,
cpu_to_be32(mdev->dev->caps.qp0_tunnel[sqp->qp.port - 1]);
sqp->ud_header.bth.psn = cpu_to_be32((sqp->send_psn++) & ((1 << 24) - 1));
- if (mlx4_get_parav_qkey(mdev->dev, sqp->qp.mqp.qpn, &qkey))
- return -EINVAL;
+ if (mlx4_is_master(mdev->dev)) {
+ if (mlx4_get_parav_qkey(mdev->dev, sqp->qp.mqp.qpn, &qkey))
+ return -EINVAL;
+ } else {
+ if (vf_get_qp0_qkey(mdev->dev, sqp->qp.mqp.qpn, &qkey))
+ return -EINVAL;
+ }
sqp->ud_header.deth.qkey = cpu_to_be32(qkey);
sqp->ud_header.deth.source_qpn = cpu_to_be32(sqp->qp.mqp.qpn);
@@ -2370,7 +2415,8 @@ static void set_datagram_seg(struct mlx4_wqe_datagram_seg *dseg,
static void set_tunnel_datagram_seg(struct mlx4_ib_dev *dev,
struct mlx4_wqe_datagram_seg *dseg,
- struct ib_send_wr *wr, enum ib_qp_type qpt)
+ struct ib_send_wr *wr,
+ enum mlx4_ib_qp_type qpt)
{
union mlx4_ext_av *av = &to_mah(wr->wr.ud.ah)->av;
struct mlx4_av sqp_av = {0};
@@ -2383,8 +2429,10 @@ static void set_tunnel_datagram_seg(struct mlx4_ib_dev *dev,
cpu_to_be32(0xf0000000);
memcpy(dseg->av, &sqp_av, sizeof (struct mlx4_av));
- /* This function used only for sending on QP1 proxies */
- dseg->dqpn = cpu_to_be32(dev->dev->caps.qp1_tunnel[port - 1]);
+ if (qpt == MLX4_IB_QPT_PROXY_GSI)
+ dseg->dqpn = cpu_to_be32(dev->dev->caps.qp1_tunnel[port - 1]);
+ else
+ dseg->dqpn = cpu_to_be32(dev->dev->caps.qp0_tunnel[port - 1]);
/* Use QKEY from the QP context, which is set by master */
dseg->qkey = cpu_to_be32(IB_QP_SET_QKEY);
}
@@ -2679,11 +2727,6 @@ int mlx4_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
break;
case MLX4_IB_QPT_PROXY_SMI_OWNER:
- if (unlikely(!mlx4_is_master(to_mdev(ibqp->device)->dev))) {
- err = -ENOSYS;
- *bad_wr = wr;
- goto out;
- }
err = build_sriov_qp0_header(to_msqp(qp), wr, ctrl, &seglen);
if (unlikely(err)) {
*bad_wr = wr;
@@ -2700,16 +2743,13 @@ int mlx4_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
size += seglen / 16;
break;
case MLX4_IB_QPT_PROXY_SMI:
- /* don't allow QP0 sends on guests */
- err = -ENOSYS;
- *bad_wr = wr;
- goto out;
case MLX4_IB_QPT_PROXY_GSI:
/* If we are tunneling special qps, this is a UD qp.
* In this case we first add a UD segment targeting
* the tunnel qp, and then add a header with address
* information */
- set_tunnel_datagram_seg(to_mdev(ibqp->device), wqe, wr, ibqp->qp_type);
+ set_tunnel_datagram_seg(to_mdev(ibqp->device), wqe, wr,
+ qp->mlx4_ib_qp_type);
wqe += sizeof (struct mlx4_wqe_datagram_seg);
size += sizeof (struct mlx4_wqe_datagram_seg) / 16;
build_tunnel_header(wr, wqe, &seglen);
diff --git a/drivers/infiniband/hw/mlx4/srq.c b/drivers/infiniband/hw/mlx4/srq.c
index 60c5fb025fc..62d9285300a 100644
--- a/drivers/infiniband/hw/mlx4/srq.c
+++ b/drivers/infiniband/hw/mlx4/srq.c
@@ -134,13 +134,14 @@ struct ib_srq *mlx4_ib_create_srq(struct ib_pd *pd,
if (err)
goto err_mtt;
} else {
- err = mlx4_db_alloc(dev->dev, &srq->db, 0);
+ err = mlx4_db_alloc(dev->dev, &srq->db, 0, GFP_KERNEL);
if (err)
goto err_srq;
*srq->db.db = 0;
- if (mlx4_buf_alloc(dev->dev, buf_size, PAGE_SIZE * 2, &srq->buf)) {
+ if (mlx4_buf_alloc(dev->dev, buf_size, PAGE_SIZE * 2, &srq->buf,
+ GFP_KERNEL)) {
err = -ENOMEM;
goto err_db;
}
@@ -165,7 +166,7 @@ struct ib_srq *mlx4_ib_create_srq(struct ib_pd *pd,
if (err)
goto err_buf;
- err = mlx4_buf_write_mtt(dev->dev, &srq->mtt, &srq->buf);
+ err = mlx4_buf_write_mtt(dev->dev, &srq->mtt, &srq->buf, GFP_KERNEL);
if (err)
goto err_mtt;
diff --git a/drivers/infiniband/hw/mlx4/sysfs.c b/drivers/infiniband/hw/mlx4/sysfs.c
index 5a38e43eca6..cb4c66e723b 100644
--- a/drivers/infiniband/hw/mlx4/sysfs.c
+++ b/drivers/infiniband/hw/mlx4/sysfs.c
@@ -389,8 +389,10 @@ struct mlx4_port {
struct mlx4_ib_dev *dev;
struct attribute_group pkey_group;
struct attribute_group gid_group;
- u8 port_num;
+ struct device_attribute enable_smi_admin;
+ struct device_attribute smi_enabled;
int slave;
+ u8 port_num;
};
@@ -558,6 +560,101 @@ err:
return NULL;
}
+static ssize_t sysfs_show_smi_enabled(struct device *dev,
+ struct device_attribute *attr, char *buf)
+{
+ struct mlx4_port *p =
+ container_of(attr, struct mlx4_port, smi_enabled);
+ ssize_t len = 0;
+
+ if (mlx4_vf_smi_enabled(p->dev->dev, p->slave, p->port_num))
+ len = sprintf(buf, "%d\n", 1);
+ else
+ len = sprintf(buf, "%d\n", 0);
+
+ return len;
+}
+
+static ssize_t sysfs_show_enable_smi_admin(struct device *dev,
+ struct device_attribute *attr,
+ char *buf)
+{
+ struct mlx4_port *p =
+ container_of(attr, struct mlx4_port, enable_smi_admin);
+ ssize_t len = 0;
+
+ if (mlx4_vf_get_enable_smi_admin(p->dev->dev, p->slave, p->port_num))
+ len = sprintf(buf, "%d\n", 1);
+ else
+ len = sprintf(buf, "%d\n", 0);
+
+ return len;
+}
+
+static ssize_t sysfs_store_enable_smi_admin(struct device *dev,
+ struct device_attribute *attr,
+ const char *buf, size_t count)
+{
+ struct mlx4_port *p =
+ container_of(attr, struct mlx4_port, enable_smi_admin);
+ int enable;
+
+ if (sscanf(buf, "%i", &enable) != 1 ||
+ enable < 0 || enable > 1)
+ return -EINVAL;
+
+ if (mlx4_vf_set_enable_smi_admin(p->dev->dev, p->slave, p->port_num, enable))
+ return -EINVAL;
+ return count;
+}
+
+static int add_vf_smi_entries(struct mlx4_port *p)
+{
+ int is_eth = rdma_port_get_link_layer(&p->dev->ib_dev, p->port_num) ==
+ IB_LINK_LAYER_ETHERNET;
+ int ret;
+
+ /* do not display entries if eth transport, or if master */
+ if (is_eth || p->slave == mlx4_master_func_num(p->dev->dev))
+ return 0;
+
+ sysfs_attr_init(&p->smi_enabled.attr);
+ p->smi_enabled.show = sysfs_show_smi_enabled;
+ p->smi_enabled.store = NULL;
+ p->smi_enabled.attr.name = "smi_enabled";
+ p->smi_enabled.attr.mode = 0444;
+ ret = sysfs_create_file(&p->kobj, &p->smi_enabled.attr);
+ if (ret) {
+ pr_err("failed to create smi_enabled\n");
+ return ret;
+ }
+
+ sysfs_attr_init(&p->enable_smi_admin.attr);
+ p->enable_smi_admin.show = sysfs_show_enable_smi_admin;
+ p->enable_smi_admin.store = sysfs_store_enable_smi_admin;
+ p->enable_smi_admin.attr.name = "enable_smi_admin";
+ p->enable_smi_admin.attr.mode = 0644;
+ ret = sysfs_create_file(&p->kobj, &p->enable_smi_admin.attr);
+ if (ret) {
+ pr_err("failed to create enable_smi_admin\n");
+ sysfs_remove_file(&p->kobj, &p->smi_enabled.attr);
+ return ret;
+ }
+ return 0;
+}
+
+static void remove_vf_smi_entries(struct mlx4_port *p)
+{
+ int is_eth = rdma_port_get_link_layer(&p->dev->ib_dev, p->port_num) ==
+ IB_LINK_LAYER_ETHERNET;
+
+ if (is_eth || p->slave == mlx4_master_func_num(p->dev->dev))
+ return;
+
+ sysfs_remove_file(&p->kobj, &p->smi_enabled.attr);
+ sysfs_remove_file(&p->kobj, &p->enable_smi_admin.attr);
+}
+
static int add_port(struct mlx4_ib_dev *dev, int port_num, int slave)
{
struct mlx4_port *p;
@@ -602,6 +699,10 @@ static int add_port(struct mlx4_ib_dev *dev, int port_num, int slave)
if (ret)
goto err_free_gid;
+ ret = add_vf_smi_entries(p);
+ if (ret)
+ goto err_free_gid;
+
list_add_tail(&p->kobj.entry, &dev->pkeys.pkey_port_list[slave]);
return 0;
@@ -669,6 +770,7 @@ err_add:
mport = container_of(p, struct mlx4_port, kobj);
sysfs_remove_group(p, &mport->pkey_group);
sysfs_remove_group(p, &mport->gid_group);
+ remove_vf_smi_entries(mport);
kobject_put(p);
}
kobject_put(dev->dev_ports_parent[slave]);
@@ -713,6 +815,7 @@ static void unregister_pkey_tree(struct mlx4_ib_dev *device)
port = container_of(p, struct mlx4_port, kobj);
sysfs_remove_group(p, &port->pkey_group);
sysfs_remove_group(p, &port->gid_group);
+ remove_vf_smi_entries(port);
kobject_put(p);
kobject_put(device->dev_ports_parent[slave]);
}
diff --git a/drivers/infiniband/hw/mlx5/cq.c b/drivers/infiniband/hw/mlx5/cq.c
index 62bb6b49dc1..8ae4f896cb4 100644
--- a/drivers/infiniband/hw/mlx5/cq.c
+++ b/drivers/infiniband/hw/mlx5/cq.c
@@ -32,6 +32,7 @@
#include <linux/kref.h>
#include <rdma/ib_umem.h>
+#include <rdma/ib_user_verbs.h>
#include "mlx5_ib.h"
#include "user.h"
@@ -602,14 +603,24 @@ static int create_cq_user(struct mlx5_ib_dev *dev, struct ib_udata *udata,
int *cqe_size, int *index, int *inlen)
{
struct mlx5_ib_create_cq ucmd;
+ size_t ucmdlen;
int page_shift;
int npages;
int ncont;
int err;
- if (ib_copy_from_udata(&ucmd, udata, sizeof(ucmd)))
+ ucmdlen =
+ (udata->inlen - sizeof(struct ib_uverbs_cmd_hdr) <
+ sizeof(ucmd)) ? (sizeof(ucmd) -
+ sizeof(ucmd.reserved)) : sizeof(ucmd);
+
+ if (ib_copy_from_udata(&ucmd, udata, ucmdlen))
return -EFAULT;
+ if (ucmdlen == sizeof(ucmd) &&
+ ucmd.reserved != 0)
+ return -EINVAL;
+
if (ucmd.cqe_size != 64 && ucmd.cqe_size != 128)
return -EINVAL;
diff --git a/drivers/infiniband/hw/mlx5/mlx5_ib.h b/drivers/infiniband/hw/mlx5/mlx5_ib.h
index 50541586e0a..f2ccf1a5a29 100644
--- a/drivers/infiniband/hw/mlx5/mlx5_ib.h
+++ b/drivers/infiniband/hw/mlx5/mlx5_ib.h
@@ -264,8 +264,6 @@ struct mlx5_ib_mr {
__be64 *pas;
dma_addr_t dma;
int npages;
- struct completion done;
- enum ib_wc_status status;
struct mlx5_ib_dev *dev;
struct mlx5_create_mkey_mbox_out out;
struct mlx5_core_sig_ctx *sig;
@@ -277,6 +275,17 @@ struct mlx5_ib_fast_reg_page_list {
dma_addr_t map;
};
+struct mlx5_ib_umr_context {
+ enum ib_wc_status status;
+ struct completion done;
+};
+
+static inline void mlx5_ib_init_umr_context(struct mlx5_ib_umr_context *context)
+{
+ context->status = -1;
+ init_completion(&context->done);
+}
+
struct umr_common {
struct ib_pd *pd;
struct ib_cq *cq;
diff --git a/drivers/infiniband/hw/mlx5/mr.c b/drivers/infiniband/hw/mlx5/mr.c
index 81392b26d07..afa873bd028 100644
--- a/drivers/infiniband/hw/mlx5/mr.c
+++ b/drivers/infiniband/hw/mlx5/mr.c
@@ -73,6 +73,8 @@ static void reg_mr_callback(int status, void *context)
struct mlx5_cache_ent *ent = &cache->ent[c];
u8 key;
unsigned long flags;
+ struct mlx5_mr_table *table = &dev->mdev.priv.mr_table;
+ int err;
spin_lock_irqsave(&ent->lock, flags);
ent->pending--;
@@ -107,6 +109,13 @@ static void reg_mr_callback(int status, void *context)
ent->cur++;
ent->size++;
spin_unlock_irqrestore(&ent->lock, flags);
+
+ write_lock_irqsave(&table->lock, flags);
+ err = radix_tree_insert(&table->tree, mlx5_base_mkey(mr->mmr.key),
+ &mr->mmr);
+ if (err)
+ pr_err("Error inserting to mr tree. 0x%x\n", -err);
+ write_unlock_irqrestore(&table->lock, flags);
}
static int add_keys(struct mlx5_ib_dev *dev, int c, int num)
@@ -699,7 +708,7 @@ static void prep_umr_unreg_wqe(struct mlx5_ib_dev *dev,
void mlx5_umr_cq_handler(struct ib_cq *cq, void *cq_context)
{
- struct mlx5_ib_mr *mr;
+ struct mlx5_ib_umr_context *context;
struct ib_wc wc;
int err;
@@ -712,9 +721,9 @@ void mlx5_umr_cq_handler(struct ib_cq *cq, void *cq_context)
if (err == 0)
break;
- mr = (struct mlx5_ib_mr *)(unsigned long)wc.wr_id;
- mr->status = wc.status;
- complete(&mr->done);
+ context = (struct mlx5_ib_umr_context *) (unsigned long) wc.wr_id;
+ context->status = wc.status;
+ complete(&context->done);
}
ib_req_notify_cq(cq, IB_CQ_NEXT_COMP);
}
@@ -726,11 +735,12 @@ static struct mlx5_ib_mr *reg_umr(struct ib_pd *pd, struct ib_umem *umem,
struct mlx5_ib_dev *dev = to_mdev(pd->device);
struct device *ddev = dev->ib_dev.dma_device;
struct umr_common *umrc = &dev->umrc;
+ struct mlx5_ib_umr_context umr_context;
struct ib_send_wr wr, *bad;
struct mlx5_ib_mr *mr;
struct ib_sge sg;
int size = sizeof(u64) * npages;
- int err;
+ int err = 0;
int i;
for (i = 0; i < 1; i++) {
@@ -751,7 +761,7 @@ static struct mlx5_ib_mr *reg_umr(struct ib_pd *pd, struct ib_umem *umem,
mr->pas = kmalloc(size + MLX5_UMR_ALIGN - 1, GFP_KERNEL);
if (!mr->pas) {
err = -ENOMEM;
- goto error;
+ goto free_mr;
}
mlx5_ib_populate_pas(dev, umem, page_shift,
@@ -760,44 +770,46 @@ static struct mlx5_ib_mr *reg_umr(struct ib_pd *pd, struct ib_umem *umem,
mr->dma = dma_map_single(ddev, mr_align(mr->pas, MLX5_UMR_ALIGN), size,
DMA_TO_DEVICE);
if (dma_mapping_error(ddev, mr->dma)) {
- kfree(mr->pas);
err = -ENOMEM;
- goto error;
+ goto free_pas;
}
memset(&wr, 0, sizeof(wr));
- wr.wr_id = (u64)(unsigned long)mr;
+ wr.wr_id = (u64)(unsigned long)&umr_context;
prep_umr_reg_wqe(pd, &wr, &sg, mr->dma, npages, mr->mmr.key, page_shift, virt_addr, len, access_flags);
- /* We serialize polls so one process does not kidnap another's
- * completion. This is not a problem since wr is completed in
- * around 1 usec
- */
+ mlx5_ib_init_umr_context(&umr_context);
down(&umrc->sem);
- init_completion(&mr->done);
err = ib_post_send(umrc->qp, &wr, &bad);
if (err) {
mlx5_ib_warn(dev, "post send failed, err %d\n", err);
- up(&umrc->sem);
- goto error;
+ goto unmap_dma;
+ } else {
+ wait_for_completion(&umr_context.done);
+ if (umr_context.status != IB_WC_SUCCESS) {
+ mlx5_ib_warn(dev, "reg umr failed\n");
+ err = -EFAULT;
+ }
}
- wait_for_completion(&mr->done);
- up(&umrc->sem);
+ mr->mmr.iova = virt_addr;
+ mr->mmr.size = len;
+ mr->mmr.pd = to_mpd(pd)->pdn;
+
+unmap_dma:
+ up(&umrc->sem);
dma_unmap_single(ddev, mr->dma, size, DMA_TO_DEVICE);
+
+free_pas:
kfree(mr->pas);
- if (mr->status != IB_WC_SUCCESS) {
- mlx5_ib_warn(dev, "reg umr failed\n");
- err = -EFAULT;
- goto error;
+free_mr:
+ if (err) {
+ free_cached_mr(dev, mr);
+ return ERR_PTR(err);
}
return mr;
-
-error:
- free_cached_mr(dev, mr);
- return ERR_PTR(err);
}
static struct mlx5_ib_mr *reg_create(struct ib_pd *pd, u64 virt_addr,
@@ -926,24 +938,26 @@ error:
static int unreg_umr(struct mlx5_ib_dev *dev, struct mlx5_ib_mr *mr)
{
struct umr_common *umrc = &dev->umrc;
+ struct mlx5_ib_umr_context umr_context;
struct ib_send_wr wr, *bad;
int err;
memset(&wr, 0, sizeof(wr));
- wr.wr_id = (u64)(unsigned long)mr;
+ wr.wr_id = (u64)(unsigned long)&umr_context;
prep_umr_unreg_wqe(dev, &wr, mr->mmr.key);
+ mlx5_ib_init_umr_context(&umr_context);
down(&umrc->sem);
- init_completion(&mr->done);
err = ib_post_send(umrc->qp, &wr, &bad);
if (err) {
up(&umrc->sem);
mlx5_ib_dbg(dev, "err %d\n", err);
goto error;
+ } else {
+ wait_for_completion(&umr_context.done);
+ up(&umrc->sem);
}
- wait_for_completion(&mr->done);
- up(&umrc->sem);
- if (mr->status != IB_WC_SUCCESS) {
+ if (umr_context.status != IB_WC_SUCCESS) {
mlx5_ib_warn(dev, "unreg umr failed\n");
err = -EFAULT;
goto error;
diff --git a/drivers/infiniband/hw/mlx5/qp.c b/drivers/infiniband/hw/mlx5/qp.c
index dc930ed21ec..d13ddf1c003 100644
--- a/drivers/infiniband/hw/mlx5/qp.c
+++ b/drivers/infiniband/hw/mlx5/qp.c
@@ -574,6 +574,10 @@ static int create_user_qp(struct mlx5_ib_dev *dev, struct ib_pd *pd,
uar_index = uuarn_to_uar_index(&context->uuari, uuarn);
mlx5_ib_dbg(dev, "uuarn 0x%x, uar_index 0x%x\n", uuarn, uar_index);
+ qp->rq.offset = 0;
+ qp->sq.wqe_shift = ilog2(MLX5_SEND_WQE_BB);
+ qp->sq.offset = qp->rq.wqe_cnt << qp->rq.wqe_shift;
+
err = set_user_buf_size(dev, qp, &ucmd);
if (err)
goto err_uuar;
@@ -2078,6 +2082,7 @@ static int mlx5_set_bsf(struct ib_mr *sig_mr,
struct ib_sig_domain *wire = &sig_attrs->wire;
int ret, selector;
+ memset(bsf, 0, sizeof(*bsf));
switch (sig_attrs->mem.sig_type) {
case IB_SIG_TYPE_T10_DIF:
if (sig_attrs->wire.sig_type != IB_SIG_TYPE_T10_DIF)
@@ -2090,9 +2095,11 @@ static int mlx5_set_bsf(struct ib_mr *sig_mr,
/* Same block structure */
basic->bsf_size_sbs = 1 << 4;
if (mem->sig.dif.bg_type == wire->sig.dif.bg_type)
- basic->wire.copy_byte_mask = 0xff;
- else
- basic->wire.copy_byte_mask = 0x3f;
+ basic->wire.copy_byte_mask |= 0xc0;
+ if (mem->sig.dif.app_tag == wire->sig.dif.app_tag)
+ basic->wire.copy_byte_mask |= 0x30;
+ if (mem->sig.dif.ref_tag == wire->sig.dif.ref_tag)
+ basic->wire.copy_byte_mask |= 0x0f;
} else
basic->wire.bs_selector = bs_selector(wire->sig.dif.pi_interval);
@@ -2131,9 +2138,13 @@ static int set_sig_data_segment(struct ib_send_wr *wr, struct mlx5_ib_qp *qp,
int ret;
int wqe_size;
- if (!wr->wr.sig_handover.prot) {
+ if (!wr->wr.sig_handover.prot ||
+ (data_key == wr->wr.sig_handover.prot->lkey &&
+ data_va == wr->wr.sig_handover.prot->addr &&
+ data_len == wr->wr.sig_handover.prot->length)) {
/**
* Source domain doesn't contain signature information
+ * or data and protection are interleaved in memory.
* So need construct:
* ------------------
* | data_klm |
@@ -2187,23 +2198,13 @@ static int set_sig_data_segment(struct ib_send_wr *wr, struct mlx5_ib_qp *qp,
data_sentry->bcount = cpu_to_be16(block_size);
data_sentry->key = cpu_to_be32(data_key);
data_sentry->va = cpu_to_be64(data_va);
+ data_sentry->stride = cpu_to_be16(block_size);
+
prot_sentry->bcount = cpu_to_be16(prot_size);
prot_sentry->key = cpu_to_be32(prot_key);
+ prot_sentry->va = cpu_to_be64(prot_va);
+ prot_sentry->stride = cpu_to_be16(prot_size);
- if (prot_key == data_key && prot_va == data_va) {
- /**
- * The data and protection are interleaved
- * in a single memory region
- **/
- prot_sentry->va = cpu_to_be64(data_va + block_size);
- prot_sentry->stride = cpu_to_be16(block_size + prot_size);
- data_sentry->stride = prot_sentry->stride;
- } else {
- /* The data and protection are two different buffers */
- prot_sentry->va = cpu_to_be64(prot_va);
- data_sentry->stride = cpu_to_be16(block_size);
- prot_sentry->stride = cpu_to_be16(prot_size);
- }
wqe_size = ALIGN(sizeof(*sblock_ctrl) + sizeof(*data_sentry) +
sizeof(*prot_sentry), 64);
}
@@ -2275,7 +2276,10 @@ static int set_sig_umr_wr(struct ib_send_wr *wr, struct mlx5_ib_qp *qp,
/* length of the protected region, data + protection */
region_len = wr->sg_list->length;
- if (wr->wr.sig_handover.prot)
+ if (wr->wr.sig_handover.prot &&
+ (wr->wr.sig_handover.prot->lkey != wr->sg_list->lkey ||
+ wr->wr.sig_handover.prot->addr != wr->sg_list->addr ||
+ wr->wr.sig_handover.prot->length != wr->sg_list->length))
region_len += wr->wr.sig_handover.prot->length;
/**
diff --git a/drivers/infiniband/hw/mlx5/srq.c b/drivers/infiniband/hw/mlx5/srq.c
index 210b3eaf188..384af6dec5e 100644
--- a/drivers/infiniband/hw/mlx5/srq.c
+++ b/drivers/infiniband/hw/mlx5/srq.c
@@ -35,6 +35,7 @@
#include <linux/mlx5/srq.h>
#include <linux/slab.h>
#include <rdma/ib_umem.h>
+#include <rdma/ib_user_verbs.h>
#include "mlx5_ib.h"
#include "user.h"
@@ -78,16 +79,27 @@ static int create_srq_user(struct ib_pd *pd, struct mlx5_ib_srq *srq,
{
struct mlx5_ib_dev *dev = to_mdev(pd->device);
struct mlx5_ib_create_srq ucmd;
+ size_t ucmdlen;
int err;
int npages;
int page_shift;
int ncont;
u32 offset;
- if (ib_copy_from_udata(&ucmd, udata, sizeof(ucmd))) {
+ ucmdlen =
+ (udata->inlen - sizeof(struct ib_uverbs_cmd_hdr) <
+ sizeof(ucmd)) ? (sizeof(ucmd) -
+ sizeof(ucmd.reserved)) : sizeof(ucmd);
+
+ if (ib_copy_from_udata(&ucmd, udata, ucmdlen)) {
mlx5_ib_dbg(dev, "failed copy udata\n");
return -EFAULT;
}
+
+ if (ucmdlen == sizeof(ucmd) &&
+ ucmd.reserved != 0)
+ return -EINVAL;
+
srq->wq_sig = !!(ucmd.flags & MLX5_SRQ_FLAG_SIGNATURE);
srq->umem = ib_umem_get(pd->uobject->context, ucmd.buf_addr, buf_size,
diff --git a/drivers/infiniband/hw/mlx5/user.h b/drivers/infiniband/hw/mlx5/user.h
index 0f4f8e42a17..d0ba264ac1e 100644
--- a/drivers/infiniband/hw/mlx5/user.h
+++ b/drivers/infiniband/hw/mlx5/user.h
@@ -91,6 +91,7 @@ struct mlx5_ib_create_cq {
__u64 buf_addr;
__u64 db_addr;
__u32 cqe_size;
+ __u32 reserved; /* explicit padding (optional on i386) */
};
struct mlx5_ib_create_cq_resp {
@@ -109,6 +110,7 @@ struct mlx5_ib_create_srq {
__u64 buf_addr;
__u64 db_addr;
__u32 flags;
+ __u32 reserved; /* explicit padding (optional on i386) */
};
struct mlx5_ib_create_srq_resp {
diff --git a/drivers/infiniband/hw/nes/nes.c b/drivers/infiniband/hw/nes/nes.c
index 353c7b05a90..3b2a6dc8ea9 100644
--- a/drivers/infiniband/hw/nes/nes.c
+++ b/drivers/infiniband/hw/nes/nes.c
@@ -68,7 +68,6 @@ MODULE_VERSION(DRV_VERSION);
int max_mtu = 9000;
int interrupt_mod_interval = 0;
-
/* Interoperability */
int mpa_version = 1;
module_param(mpa_version, int, 0644);
@@ -112,6 +111,16 @@ static struct pci_device_id nes_pci_table[] = {
MODULE_DEVICE_TABLE(pci, nes_pci_table);
+/* registered nes netlink callbacks */
+static struct ibnl_client_cbs nes_nl_cb_table[] = {
+ [RDMA_NL_IWPM_REG_PID] = {.dump = iwpm_register_pid_cb},
+ [RDMA_NL_IWPM_ADD_MAPPING] = {.dump = iwpm_add_mapping_cb},
+ [RDMA_NL_IWPM_QUERY_MAPPING] = {.dump = iwpm_add_and_query_mapping_cb},
+ [RDMA_NL_IWPM_HANDLE_ERR] = {.dump = iwpm_mapping_error_cb},
+ [RDMA_NL_IWPM_MAPINFO] = {.dump = iwpm_mapping_info_cb},
+ [RDMA_NL_IWPM_MAPINFO_NUM] = {.dump = iwpm_ack_mapping_info_cb}
+};
+
static int nes_inetaddr_event(struct notifier_block *, unsigned long, void *);
static int nes_net_event(struct notifier_block *, unsigned long, void *);
static int nes_notifiers_registered;
@@ -672,6 +681,17 @@ static int nes_probe(struct pci_dev *pcidev, const struct pci_device_id *ent)
}
nes_notifiers_registered++;
+ if (ibnl_add_client(RDMA_NL_NES, RDMA_NL_IWPM_NUM_OPS, nes_nl_cb_table))
+ printk(KERN_ERR PFX "%s[%u]: Failed to add netlink callback\n",
+ __func__, __LINE__);
+
+ ret = iwpm_init(RDMA_NL_NES);
+ if (ret) {
+ printk(KERN_ERR PFX "%s: port mapper initialization failed\n",
+ pci_name(pcidev));
+ goto bail7;
+ }
+
INIT_DELAYED_WORK(&nesdev->work, nes_recheck_link_status);
/* Initialize network devices */
@@ -710,6 +730,7 @@ static int nes_probe(struct pci_dev *pcidev, const struct pci_device_id *ent)
nes_debug(NES_DBG_INIT, "netdev_count=%d, nesadapter->netdev_count=%d\n",
nesdev->netdev_count, nesdev->nesadapter->netdev_count);
+ ibnl_remove_client(RDMA_NL_NES);
nes_notifiers_registered--;
if (nes_notifiers_registered == 0) {
@@ -773,6 +794,8 @@ static void nes_remove(struct pci_dev *pcidev)
nesdev->nesadapter->netdev_count--;
}
}
+ ibnl_remove_client(RDMA_NL_NES);
+ iwpm_exit(RDMA_NL_NES);
nes_notifiers_registered--;
if (nes_notifiers_registered == 0) {
diff --git a/drivers/infiniband/hw/nes/nes.h b/drivers/infiniband/hw/nes/nes.h
index 33cc58941a3..bd9d132f11c 100644
--- a/drivers/infiniband/hw/nes/nes.h
+++ b/drivers/infiniband/hw/nes/nes.h
@@ -51,6 +51,8 @@
#include <rdma/ib_pack.h>
#include <rdma/rdma_cm.h>
#include <rdma/iw_cm.h>
+#include <rdma/rdma_netlink.h>
+#include <rdma/iw_portmap.h>
#define NES_SEND_FIRST_WRITE
@@ -130,6 +132,7 @@
#define NES_DBG_IW_TX 0x00040000
#define NES_DBG_SHUTDOWN 0x00080000
#define NES_DBG_PAU 0x00100000
+#define NES_DBG_NLMSG 0x00200000
#define NES_DBG_RSVD1 0x10000000
#define NES_DBG_RSVD2 0x20000000
#define NES_DBG_RSVD3 0x40000000
diff --git a/drivers/infiniband/hw/nes/nes_cm.c b/drivers/infiniband/hw/nes/nes_cm.c
index dfa9df48450..6f09a72e78d 100644
--- a/drivers/infiniband/hw/nes/nes_cm.c
+++ b/drivers/infiniband/hw/nes/nes_cm.c
@@ -1,5 +1,5 @@
/*
- * Copyright (c) 2006 - 2011 Intel Corporation. All rights reserved.
+ * Copyright (c) 2006 - 2014 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
@@ -59,6 +59,7 @@
#include <net/route.h>
#include <net/ip_fib.h>
#include <net/tcp.h>
+#include <linux/fcntl.h>
#include "nes.h"
@@ -166,7 +167,6 @@ int nes_rem_ref_cm_node(struct nes_cm_node *cm_node)
{
return rem_ref_cm_node(cm_node->cm_core, cm_node);
}
-
/**
* create_event
*/
@@ -482,11 +482,11 @@ static void form_cm_frame(struct sk_buff *skb,
iph->ttl = 0x40;
iph->protocol = 0x06; /* IPPROTO_TCP */
- iph->saddr = htonl(cm_node->loc_addr);
- iph->daddr = htonl(cm_node->rem_addr);
+ iph->saddr = htonl(cm_node->mapped_loc_addr);
+ iph->daddr = htonl(cm_node->mapped_rem_addr);
- tcph->source = htons(cm_node->loc_port);
- tcph->dest = htons(cm_node->rem_port);
+ tcph->source = htons(cm_node->mapped_loc_port);
+ tcph->dest = htons(cm_node->mapped_rem_port);
tcph->seq = htonl(cm_node->tcp_cntxt.loc_seq_num);
if (flags & SET_ACK) {
@@ -525,6 +525,100 @@ static void form_cm_frame(struct sk_buff *skb,
cm_packets_created++;
}
+/*
+ * nes_create_sockaddr - Record ip addr and tcp port in a sockaddr struct
+ */
+static void nes_create_sockaddr(__be32 ip_addr, __be16 port,
+ struct sockaddr_storage *addr)
+{
+ struct sockaddr_in *nes_sockaddr = (struct sockaddr_in *)addr;
+ nes_sockaddr->sin_family = AF_INET;
+ memcpy(&nes_sockaddr->sin_addr.s_addr, &ip_addr, sizeof(__be32));
+ nes_sockaddr->sin_port = port;
+}
+
+/*
+ * nes_create_mapinfo - Create a mapinfo object in the port mapper data base
+ */
+static int nes_create_mapinfo(struct nes_cm_info *cm_info)
+{
+ struct sockaddr_storage local_sockaddr;
+ struct sockaddr_storage mapped_sockaddr;
+
+ nes_create_sockaddr(htonl(cm_info->loc_addr), htons(cm_info->loc_port),
+ &local_sockaddr);
+ nes_create_sockaddr(htonl(cm_info->mapped_loc_addr),
+ htons(cm_info->mapped_loc_port), &mapped_sockaddr);
+
+ return iwpm_create_mapinfo(&local_sockaddr,
+ &mapped_sockaddr, RDMA_NL_NES);
+}
+
+/*
+ * nes_remove_mapinfo - Remove a mapinfo object from the port mapper data base
+ * and send a remove mapping op message to
+ * the userspace port mapper
+ */
+static int nes_remove_mapinfo(u32 loc_addr, u16 loc_port,
+ u32 mapped_loc_addr, u16 mapped_loc_port)
+{
+ struct sockaddr_storage local_sockaddr;
+ struct sockaddr_storage mapped_sockaddr;
+
+ nes_create_sockaddr(htonl(loc_addr), htons(loc_port), &local_sockaddr);
+ nes_create_sockaddr(htonl(mapped_loc_addr), htons(mapped_loc_port),
+ &mapped_sockaddr);
+
+ iwpm_remove_mapinfo(&local_sockaddr, &mapped_sockaddr);
+ return iwpm_remove_mapping(&local_sockaddr, RDMA_NL_NES);
+}
+
+/*
+ * nes_form_pm_msg - Form a port mapper message with mapping info
+ */
+static void nes_form_pm_msg(struct nes_cm_info *cm_info,
+ struct iwpm_sa_data *pm_msg)
+{
+ nes_create_sockaddr(htonl(cm_info->loc_addr), htons(cm_info->loc_port),
+ &pm_msg->loc_addr);
+ nes_create_sockaddr(htonl(cm_info->rem_addr), htons(cm_info->rem_port),
+ &pm_msg->rem_addr);
+}
+
+/*
+ * nes_form_reg_msg - Form a port mapper message with dev info
+ */
+static void nes_form_reg_msg(struct nes_vnic *nesvnic,
+ struct iwpm_dev_data *pm_msg)
+{
+ memcpy(pm_msg->dev_name, nesvnic->nesibdev->ibdev.name,
+ IWPM_DEVNAME_SIZE);
+ memcpy(pm_msg->if_name, nesvnic->netdev->name, IWPM_IFNAME_SIZE);
+}
+
+/*
+ * nes_record_pm_msg - Save the received mapping info
+ */
+static void nes_record_pm_msg(struct nes_cm_info *cm_info,
+ struct iwpm_sa_data *pm_msg)
+{
+ struct sockaddr_in *mapped_loc_addr =
+ (struct sockaddr_in *)&pm_msg->mapped_loc_addr;
+ struct sockaddr_in *mapped_rem_addr =
+ (struct sockaddr_in *)&pm_msg->mapped_rem_addr;
+
+ if (mapped_loc_addr->sin_family == AF_INET) {
+ cm_info->mapped_loc_addr =
+ ntohl(mapped_loc_addr->sin_addr.s_addr);
+ cm_info->mapped_loc_port = ntohs(mapped_loc_addr->sin_port);
+ }
+ if (mapped_rem_addr->sin_family == AF_INET) {
+ cm_info->mapped_rem_addr =
+ ntohl(mapped_rem_addr->sin_addr.s_addr);
+ cm_info->mapped_rem_port = ntohs(mapped_rem_addr->sin_port);
+ }
+}
+
/**
* print_core - dump a cm core
*/
@@ -1147,8 +1241,11 @@ static struct nes_cm_node *find_node(struct nes_cm_core *cm_core,
loc_addr, loc_port,
cm_node->rem_addr, cm_node->rem_port,
rem_addr, rem_port);
- if ((cm_node->loc_addr == loc_addr) && (cm_node->loc_port == loc_port) &&
- (cm_node->rem_addr == rem_addr) && (cm_node->rem_port == rem_port)) {
+ if ((cm_node->mapped_loc_addr == loc_addr) &&
+ (cm_node->mapped_loc_port == loc_port) &&
+ (cm_node->mapped_rem_addr == rem_addr) &&
+ (cm_node->mapped_rem_port == rem_port)) {
+
add_ref_cm_node(cm_node);
spin_unlock_irqrestore(&cm_core->ht_lock, flags);
return cm_node;
@@ -1165,18 +1262,28 @@ static struct nes_cm_node *find_node(struct nes_cm_core *cm_core,
* find_listener - find a cm node listening on this addr-port pair
*/
static struct nes_cm_listener *find_listener(struct nes_cm_core *cm_core,
- nes_addr_t dst_addr, u16 dst_port, enum nes_cm_listener_state listener_state)
+ nes_addr_t dst_addr, u16 dst_port,
+ enum nes_cm_listener_state listener_state, int local)
{
unsigned long flags;
struct nes_cm_listener *listen_node;
+ nes_addr_t listen_addr;
+ u16 listen_port;
/* walk list and find cm_node associated with this session ID */
spin_lock_irqsave(&cm_core->listen_list_lock, flags);
list_for_each_entry(listen_node, &cm_core->listen_list.list, list) {
+ if (local) {
+ listen_addr = listen_node->loc_addr;
+ listen_port = listen_node->loc_port;
+ } else {
+ listen_addr = listen_node->mapped_loc_addr;
+ listen_port = listen_node->mapped_loc_port;
+ }
/* compare node pair, return node handle if a match */
- if (((listen_node->loc_addr == dst_addr) ||
- listen_node->loc_addr == 0x00000000) &&
- (listen_node->loc_port == dst_port) &&
+ if (((listen_addr == dst_addr) ||
+ listen_addr == 0x00000000) &&
+ (listen_port == dst_port) &&
(listener_state & listen_node->listener_state)) {
atomic_inc(&listen_node->ref_count);
spin_unlock_irqrestore(&cm_core->listen_list_lock, flags);
@@ -1189,7 +1296,6 @@ static struct nes_cm_listener *find_listener(struct nes_cm_core *cm_core,
return NULL;
}
-
/**
* add_hte_node - add a cm node to the hash table
*/
@@ -1310,9 +1416,20 @@ static int mini_cm_dec_refcnt_listen(struct nes_cm_core *cm_core,
spin_unlock_irqrestore(&cm_core->listen_list_lock, flags);
- if (listener->nesvnic)
- nes_manage_apbvt(listener->nesvnic, listener->loc_port,
- PCI_FUNC(listener->nesvnic->nesdev->pcidev->devfn), NES_MANAGE_APBVT_DEL);
+ if (listener->nesvnic) {
+ nes_manage_apbvt(listener->nesvnic,
+ listener->mapped_loc_port,
+ PCI_FUNC(listener->nesvnic->nesdev->pcidev->devfn),
+ NES_MANAGE_APBVT_DEL);
+
+ nes_remove_mapinfo(listener->loc_addr,
+ listener->loc_port,
+ listener->mapped_loc_addr,
+ listener->mapped_loc_port);
+ nes_debug(NES_DBG_NLMSG,
+ "Delete APBVT mapped_loc_port = %04X\n",
+ listener->mapped_loc_port);
+ }
nes_debug(NES_DBG_CM, "destroying listener (%p)\n", listener);
@@ -1454,6 +1571,11 @@ static struct nes_cm_node *make_cm_node(struct nes_cm_core *cm_core,
cm_node->loc_port = cm_info->loc_port;
cm_node->rem_port = cm_info->rem_port;
+ cm_node->mapped_loc_addr = cm_info->mapped_loc_addr;
+ cm_node->mapped_rem_addr = cm_info->mapped_rem_addr;
+ cm_node->mapped_loc_port = cm_info->mapped_loc_port;
+ cm_node->mapped_rem_port = cm_info->mapped_rem_port;
+
cm_node->mpa_frame_rev = mpa_version;
cm_node->send_rdma0_op = SEND_RDMA_READ_ZERO;
cm_node->mpav2_ird_ord = 0;
@@ -1500,8 +1622,10 @@ static struct nes_cm_node *make_cm_node(struct nes_cm_core *cm_core,
cm_node->loopbackpartner = NULL;
/* get the mac addr for the remote node */
- oldarpindex = nes_arp_table(nesdev, cm_node->rem_addr, NULL, NES_ARP_RESOLVE);
- arpindex = nes_addr_resolve_neigh(nesvnic, cm_info->rem_addr, oldarpindex);
+ oldarpindex = nes_arp_table(nesdev, cm_node->mapped_rem_addr,
+ NULL, NES_ARP_RESOLVE);
+ arpindex = nes_addr_resolve_neigh(nesvnic,
+ cm_node->mapped_rem_addr, oldarpindex);
if (arpindex < 0) {
kfree(cm_node);
return NULL;
@@ -1563,11 +1687,14 @@ static int rem_ref_cm_node(struct nes_cm_core *cm_core,
mini_cm_dec_refcnt_listen(cm_core, cm_node->listener, 0);
} else {
if (cm_node->apbvt_set && cm_node->nesvnic) {
- nes_manage_apbvt(cm_node->nesvnic, cm_node->loc_port,
- PCI_FUNC(
- cm_node->nesvnic->nesdev->pcidev->devfn),
+ nes_manage_apbvt(cm_node->nesvnic, cm_node->mapped_loc_port,
+ PCI_FUNC(cm_node->nesvnic->nesdev->pcidev->devfn),
NES_MANAGE_APBVT_DEL);
}
+ nes_debug(NES_DBG_NLMSG, "Delete APBVT mapped_loc_port = %04X\n",
+ cm_node->mapped_loc_port);
+ nes_remove_mapinfo(cm_node->loc_addr, cm_node->loc_port,
+ cm_node->mapped_loc_addr, cm_node->mapped_loc_port);
}
atomic_dec(&cm_core->node_cnt);
@@ -2235,17 +2362,21 @@ static void process_packet(struct nes_cm_node *cm_node, struct sk_buff *skb,
* mini_cm_listen - create a listen node with params
*/
static struct nes_cm_listener *mini_cm_listen(struct nes_cm_core *cm_core,
- struct nes_vnic *nesvnic, struct nes_cm_info *cm_info)
+ struct nes_vnic *nesvnic, struct nes_cm_info *cm_info)
{
struct nes_cm_listener *listener;
+ struct iwpm_dev_data pm_reg_msg;
+ struct iwpm_sa_data pm_msg;
unsigned long flags;
+ int iwpm_err = 0;
nes_debug(NES_DBG_CM, "Search for 0x%08x : 0x%04x\n",
cm_info->loc_addr, cm_info->loc_port);
/* cannot have multiple matching listeners */
- listener = find_listener(cm_core, htonl(cm_info->loc_addr),
- htons(cm_info->loc_port), NES_CM_LISTENER_EITHER_STATE);
+ listener = find_listener(cm_core, cm_info->loc_addr, cm_info->loc_port,
+ NES_CM_LISTENER_EITHER_STATE, 1);
+
if (listener && listener->listener_state == NES_CM_LISTENER_ACTIVE_STATE) {
/* find automatically incs ref count ??? */
atomic_dec(&listener->ref_count);
@@ -2254,6 +2385,22 @@ static struct nes_cm_listener *mini_cm_listen(struct nes_cm_core *cm_core,
}
if (!listener) {
+ nes_form_reg_msg(nesvnic, &pm_reg_msg);
+ iwpm_err = iwpm_register_pid(&pm_reg_msg, RDMA_NL_NES);
+ if (iwpm_err) {
+ nes_debug(NES_DBG_NLMSG,
+ "Port Mapper reg pid fail (err = %d).\n", iwpm_err);
+ }
+ if (iwpm_valid_pid() && !iwpm_err) {
+ nes_form_pm_msg(cm_info, &pm_msg);
+ iwpm_err = iwpm_add_mapping(&pm_msg, RDMA_NL_NES);
+ if (iwpm_err)
+ nes_debug(NES_DBG_NLMSG,
+ "Port Mapper query fail (err = %d).\n", iwpm_err);
+ else
+ nes_record_pm_msg(cm_info, &pm_msg);
+ }
+
/* create a CM listen node (1/2 node to compare incoming traffic to) */
listener = kzalloc(sizeof(*listener), GFP_ATOMIC);
if (!listener) {
@@ -2261,8 +2408,10 @@ static struct nes_cm_listener *mini_cm_listen(struct nes_cm_core *cm_core,
return NULL;
}
- listener->loc_addr = htonl(cm_info->loc_addr);
- listener->loc_port = htons(cm_info->loc_port);
+ listener->loc_addr = cm_info->loc_addr;
+ listener->loc_port = cm_info->loc_port;
+ listener->mapped_loc_addr = cm_info->mapped_loc_addr;
+ listener->mapped_loc_port = cm_info->mapped_loc_port;
listener->reused_node = 0;
atomic_set(&listener->ref_count, 1);
@@ -2324,14 +2473,18 @@ static struct nes_cm_node *mini_cm_connect(struct nes_cm_core *cm_core,
if (cm_info->loc_addr == cm_info->rem_addr) {
loopbackremotelistener = find_listener(cm_core,
- ntohl(nesvnic->local_ipaddr), cm_node->rem_port,
- NES_CM_LISTENER_ACTIVE_STATE);
+ cm_node->mapped_loc_addr, cm_node->mapped_rem_port,
+ NES_CM_LISTENER_ACTIVE_STATE, 0);
if (loopbackremotelistener == NULL) {
create_event(cm_node, NES_CM_EVENT_ABORTED);
} else {
loopback_cm_info = *cm_info;
loopback_cm_info.loc_port = cm_info->rem_port;
loopback_cm_info.rem_port = cm_info->loc_port;
+ loopback_cm_info.mapped_loc_port =
+ cm_info->mapped_rem_port;
+ loopback_cm_info.mapped_rem_port =
+ cm_info->mapped_loc_port;
loopback_cm_info.cm_id = loopbackremotelistener->cm_id;
loopbackremotenode = make_cm_node(cm_core, nesvnic,
&loopback_cm_info, loopbackremotelistener);
@@ -2560,6 +2713,12 @@ static int mini_cm_recv_pkt(struct nes_cm_core *cm_core,
nfo.rem_addr = ntohl(iph->saddr);
nfo.rem_port = ntohs(tcph->source);
+ /* If port mapper is available these should be mapped address info */
+ nfo.mapped_loc_addr = ntohl(iph->daddr);
+ nfo.mapped_loc_port = ntohs(tcph->dest);
+ nfo.mapped_rem_addr = ntohl(iph->saddr);
+ nfo.mapped_rem_port = ntohs(tcph->source);
+
tmp_daddr = cpu_to_be32(iph->daddr);
tmp_saddr = cpu_to_be32(iph->saddr);
@@ -2568,8 +2727,8 @@ static int mini_cm_recv_pkt(struct nes_cm_core *cm_core,
do {
cm_node = find_node(cm_core,
- nfo.rem_port, nfo.rem_addr,
- nfo.loc_port, nfo.loc_addr);
+ nfo.mapped_rem_port, nfo.mapped_rem_addr,
+ nfo.mapped_loc_port, nfo.mapped_loc_addr);
if (!cm_node) {
/* Only type of packet accepted are for */
@@ -2578,9 +2737,9 @@ static int mini_cm_recv_pkt(struct nes_cm_core *cm_core,
skb_handled = 0;
break;
}
- listener = find_listener(cm_core, nfo.loc_addr,
- nfo.loc_port,
- NES_CM_LISTENER_ACTIVE_STATE);
+ listener = find_listener(cm_core, nfo.mapped_loc_addr,
+ nfo.mapped_loc_port,
+ NES_CM_LISTENER_ACTIVE_STATE, 0);
if (!listener) {
nfo.cm_id = NULL;
nfo.conn_type = 0;
@@ -3184,10 +3343,12 @@ int nes_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
nes_cm_init_tsa_conn(nesqp, cm_node);
- nesqp->nesqp_context->tcpPorts[0] = cpu_to_le16(ntohs(laddr->sin_port));
- nesqp->nesqp_context->tcpPorts[1] = cpu_to_le16(ntohs(raddr->sin_port));
+ nesqp->nesqp_context->tcpPorts[0] =
+ cpu_to_le16(cm_node->mapped_loc_port);
+ nesqp->nesqp_context->tcpPorts[1] =
+ cpu_to_le16(cm_node->mapped_rem_port);
- nesqp->nesqp_context->ip0 = cpu_to_le32(ntohl(raddr->sin_addr.s_addr));
+ nesqp->nesqp_context->ip0 = cpu_to_le32(cm_node->mapped_rem_addr);
nesqp->nesqp_context->misc2 |= cpu_to_le32(
(u32)PCI_FUNC(nesdev->pcidev->devfn) <<
@@ -3211,9 +3372,9 @@ int nes_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
memset(&nes_quad, 0, sizeof(nes_quad));
nes_quad.DstIpAdrIndex =
cpu_to_le32((u32)PCI_FUNC(nesdev->pcidev->devfn) << 24);
- nes_quad.SrcIpadr = raddr->sin_addr.s_addr;
- nes_quad.TcpPorts[0] = raddr->sin_port;
- nes_quad.TcpPorts[1] = laddr->sin_port;
+ nes_quad.SrcIpadr = htonl(cm_node->mapped_rem_addr);
+ nes_quad.TcpPorts[0] = htons(cm_node->mapped_rem_port);
+ nes_quad.TcpPorts[1] = htons(cm_node->mapped_loc_port);
/* Produce hash key */
crc_value = get_crc_value(&nes_quad);
@@ -3315,6 +3476,9 @@ int nes_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
int apbvt_set = 0;
struct sockaddr_in *laddr = (struct sockaddr_in *)&cm_id->local_addr;
struct sockaddr_in *raddr = (struct sockaddr_in *)&cm_id->remote_addr;
+ struct iwpm_dev_data pm_reg_msg;
+ struct iwpm_sa_data pm_msg;
+ int iwpm_err = 0;
if (cm_id->remote_addr.ss_family != AF_INET)
return -ENOSYS;
@@ -3352,20 +3516,44 @@ int nes_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
nes_debug(NES_DBG_CM, "mpa private data len =%u\n",
conn_param->private_data_len);
+ /* set up the connection params for the node */
+ cm_info.loc_addr = ntohl(laddr->sin_addr.s_addr);
+ cm_info.loc_port = ntohs(laddr->sin_port);
+ cm_info.rem_addr = ntohl(raddr->sin_addr.s_addr);
+ cm_info.rem_port = ntohs(raddr->sin_port);
+ cm_info.cm_id = cm_id;
+ cm_info.conn_type = NES_CM_IWARP_CONN_TYPE;
+
+ /* No port mapper available, go with the specified peer information */
+ cm_info.mapped_loc_addr = cm_info.loc_addr;
+ cm_info.mapped_loc_port = cm_info.loc_port;
+ cm_info.mapped_rem_addr = cm_info.rem_addr;
+ cm_info.mapped_rem_port = cm_info.rem_port;
+
+ nes_form_reg_msg(nesvnic, &pm_reg_msg);
+ iwpm_err = iwpm_register_pid(&pm_reg_msg, RDMA_NL_NES);
+ if (iwpm_err) {
+ nes_debug(NES_DBG_NLMSG,
+ "Port Mapper reg pid fail (err = %d).\n", iwpm_err);
+ }
+ if (iwpm_valid_pid() && !iwpm_err) {
+ nes_form_pm_msg(&cm_info, &pm_msg);
+ iwpm_err = iwpm_add_and_query_mapping(&pm_msg, RDMA_NL_NES);
+ if (iwpm_err)
+ nes_debug(NES_DBG_NLMSG,
+ "Port Mapper query fail (err = %d).\n", iwpm_err);
+ else
+ nes_record_pm_msg(&cm_info, &pm_msg);
+ }
+
if (laddr->sin_addr.s_addr != raddr->sin_addr.s_addr) {
- nes_manage_apbvt(nesvnic, ntohs(laddr->sin_port),
- PCI_FUNC(nesdev->pcidev->devfn),
- NES_MANAGE_APBVT_ADD);
+ nes_manage_apbvt(nesvnic, cm_info.mapped_loc_port,
+ PCI_FUNC(nesdev->pcidev->devfn), NES_MANAGE_APBVT_ADD);
apbvt_set = 1;
}
- /* set up the connection params for the node */
- cm_info.loc_addr = htonl(laddr->sin_addr.s_addr);
- cm_info.loc_port = htons(laddr->sin_port);
- cm_info.rem_addr = htonl(raddr->sin_addr.s_addr);
- cm_info.rem_port = htons(raddr->sin_port);
- cm_info.cm_id = cm_id;
- cm_info.conn_type = NES_CM_IWARP_CONN_TYPE;
+ if (nes_create_mapinfo(&cm_info))
+ return -ENOMEM;
cm_id->add_ref(cm_id);
@@ -3375,10 +3563,14 @@ int nes_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
&cm_info);
if (!cm_node) {
if (apbvt_set)
- nes_manage_apbvt(nesvnic, ntohs(laddr->sin_port),
+ nes_manage_apbvt(nesvnic, cm_info.mapped_loc_port,
PCI_FUNC(nesdev->pcidev->devfn),
NES_MANAGE_APBVT_DEL);
+ nes_debug(NES_DBG_NLMSG, "Delete mapped_loc_port = %04X\n",
+ cm_info.mapped_loc_port);
+ nes_remove_mapinfo(cm_info.loc_addr, cm_info.loc_port,
+ cm_info.mapped_loc_addr, cm_info.mapped_loc_port);
cm_id->rem_ref(cm_id);
return -ENOMEM;
}
@@ -3424,13 +3616,16 @@ int nes_create_listen(struct iw_cm_id *cm_id, int backlog)
nesvnic->local_ipaddr, laddr->sin_addr.s_addr);
/* setup listen params in our api call struct */
- cm_info.loc_addr = nesvnic->local_ipaddr;
- cm_info.loc_port = laddr->sin_port;
+ cm_info.loc_addr = ntohl(nesvnic->local_ipaddr);
+ cm_info.loc_port = ntohs(laddr->sin_port);
cm_info.backlog = backlog;
cm_info.cm_id = cm_id;
cm_info.conn_type = NES_CM_IWARP_CONN_TYPE;
+ /* No port mapper available, go with the specified info */
+ cm_info.mapped_loc_addr = cm_info.loc_addr;
+ cm_info.mapped_loc_port = cm_info.loc_port;
cm_node = g_cm_core->api->listen(g_cm_core, nesvnic, &cm_info);
if (!cm_node) {
@@ -3442,7 +3637,10 @@ int nes_create_listen(struct iw_cm_id *cm_id, int backlog)
cm_id->provider_data = cm_node;
if (!cm_node->reused_node) {
- err = nes_manage_apbvt(nesvnic, ntohs(laddr->sin_port),
+ if (nes_create_mapinfo(&cm_info))
+ return -ENOMEM;
+
+ err = nes_manage_apbvt(nesvnic, cm_node->mapped_loc_port,
PCI_FUNC(nesvnic->nesdev->pcidev->devfn),
NES_MANAGE_APBVT_ADD);
if (err) {
@@ -3567,9 +3765,11 @@ static void cm_event_connected(struct nes_cm_event *event)
nes_cm_init_tsa_conn(nesqp, cm_node);
/* set the QP tsa context */
- nesqp->nesqp_context->tcpPorts[0] = cpu_to_le16(ntohs(laddr->sin_port));
- nesqp->nesqp_context->tcpPorts[1] = cpu_to_le16(ntohs(raddr->sin_port));
- nesqp->nesqp_context->ip0 = cpu_to_le32(ntohl(raddr->sin_addr.s_addr));
+ nesqp->nesqp_context->tcpPorts[0] =
+ cpu_to_le16(cm_node->mapped_loc_port);
+ nesqp->nesqp_context->tcpPorts[1] =
+ cpu_to_le16(cm_node->mapped_rem_port);
+ nesqp->nesqp_context->ip0 = cpu_to_le32(cm_node->mapped_rem_addr);
nesqp->nesqp_context->misc2 |= cpu_to_le32(
(u32)PCI_FUNC(nesdev->pcidev->devfn) <<
@@ -3599,9 +3799,9 @@ static void cm_event_connected(struct nes_cm_event *event)
nes_quad.DstIpAdrIndex =
cpu_to_le32((u32)PCI_FUNC(nesdev->pcidev->devfn) << 24);
- nes_quad.SrcIpadr = raddr->sin_addr.s_addr;
- nes_quad.TcpPorts[0] = raddr->sin_port;
- nes_quad.TcpPorts[1] = laddr->sin_port;
+ nes_quad.SrcIpadr = htonl(cm_node->mapped_rem_addr);
+ nes_quad.TcpPorts[0] = htons(cm_node->mapped_rem_port);
+ nes_quad.TcpPorts[1] = htons(cm_node->mapped_loc_port);
/* Produce hash key */
crc_value = get_crc_value(&nes_quad);
@@ -3629,7 +3829,7 @@ static void cm_event_connected(struct nes_cm_event *event)
cm_event.ird = cm_node->ird_size;
cm_event.ord = cm_node->ord_size;
- cm_event_laddr->sin_addr.s_addr = event->cm_info.rem_addr;
+ cm_event_laddr->sin_addr.s_addr = htonl(event->cm_info.rem_addr);
ret = cm_id->event_handler(cm_id, &cm_event);
nes_debug(NES_DBG_CM, "OFA CM event_handler returned, ret=%d\n", ret);
diff --git a/drivers/infiniband/hw/nes/nes_cm.h b/drivers/infiniband/hw/nes/nes_cm.h
index 522c99cd07c..f522cf63978 100644
--- a/drivers/infiniband/hw/nes/nes_cm.h
+++ b/drivers/infiniband/hw/nes/nes_cm.h
@@ -1,5 +1,5 @@
/*
- * Copyright (c) 2006 - 2011 Intel Corporation. All rights reserved.
+ * Copyright (c) 2006 - 2014 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
@@ -293,8 +293,8 @@ struct nes_cm_listener {
struct list_head list;
struct nes_cm_core *cm_core;
u8 loc_mac[ETH_ALEN];
- nes_addr_t loc_addr;
- u16 loc_port;
+ nes_addr_t loc_addr, mapped_loc_addr;
+ u16 loc_port, mapped_loc_port;
struct iw_cm_id *cm_id;
enum nes_cm_conn_type conn_type;
atomic_t ref_count;
@@ -308,7 +308,9 @@ struct nes_cm_listener {
/* per connection node and node state information */
struct nes_cm_node {
nes_addr_t loc_addr, rem_addr;
+ nes_addr_t mapped_loc_addr, mapped_rem_addr;
u16 loc_port, rem_port;
+ u16 mapped_loc_port, mapped_rem_port;
u8 loc_mac[ETH_ALEN];
u8 rem_mac[ETH_ALEN];
@@ -364,6 +366,10 @@ struct nes_cm_info {
u16 rem_port;
nes_addr_t loc_addr;
nes_addr_t rem_addr;
+ u16 mapped_loc_port;
+ u16 mapped_rem_port;
+ nes_addr_t mapped_loc_addr;
+ nes_addr_t mapped_rem_addr;
enum nes_cm_conn_type conn_type;
int backlog;
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_stats.c b/drivers/infiniband/hw/ocrdma/ocrdma_stats.c
index 6c54106f5e6..41a9aec9998 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma_stats.c
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_stats.c
@@ -510,16 +510,9 @@ exit:
return status;
}
-static int ocrdma_debugfs_open(struct inode *inode, struct file *file)
-{
- if (inode->i_private)
- file->private_data = inode->i_private;
- return 0;
-}
-
static const struct file_operations ocrdma_dbg_ops = {
.owner = THIS_MODULE,
- .open = ocrdma_debugfs_open,
+ .open = simple_open,
.read = ocrdma_dbgfs_ops_read,
};
diff --git a/drivers/infiniband/hw/qib/qib_init.c b/drivers/infiniband/hw/qib/qib_init.c
index 5b7aeb224a3..8d3c78ddc90 100644
--- a/drivers/infiniband/hw/qib/qib_init.c
+++ b/drivers/infiniband/hw/qib/qib_init.c
@@ -1272,7 +1272,7 @@ static int qib_notify_dca(struct notifier_block *nb, unsigned long event,
* Do all the generic driver unit- and chip-independent memory
* allocation and initialization.
*/
-static int __init qlogic_ib_init(void)
+static int __init qib_ib_init(void)
{
int ret;
@@ -1316,12 +1316,12 @@ bail:
return ret;
}
-module_init(qlogic_ib_init);
+module_init(qib_ib_init);
/*
* Do the non-unit driver cleanup, memory free, etc. at unload.
*/
-static void __exit qlogic_ib_cleanup(void)
+static void __exit qib_ib_cleanup(void)
{
int ret;
@@ -1346,7 +1346,7 @@ static void __exit qlogic_ib_cleanup(void)
qib_dev_cleanup();
}
-module_exit(qlogic_ib_cleanup);
+module_exit(qib_ib_cleanup);
/* this can only be called after a successful initialization */
static void cleanup_device_data(struct qib_devdata *dd)
diff --git a/drivers/infiniband/hw/qib/qib_mad.c b/drivers/infiniband/hw/qib/qib_mad.c
index edad991d60e..22c720e5740 100644
--- a/drivers/infiniband/hw/qib/qib_mad.c
+++ b/drivers/infiniband/hw/qib/qib_mad.c
@@ -1028,7 +1028,7 @@ static int set_pkeys(struct qib_devdata *dd, u8 port, u16 *pkeys)
event.event = IB_EVENT_PKEY_CHANGE;
event.device = &dd->verbs_dev.ibdev;
- event.element.port_num = 1;
+ event.element.port_num = port;
ib_dispatch_event(&event);
}
return 0;
diff --git a/drivers/infiniband/hw/qib/qib_qp.c b/drivers/infiniband/hw/qib/qib_qp.c
index 0cad0c40d74..7fcc150d603 100644
--- a/drivers/infiniband/hw/qib/qib_qp.c
+++ b/drivers/infiniband/hw/qib/qib_qp.c
@@ -985,7 +985,8 @@ struct ib_qp *qib_create_qp(struct ib_pd *ibpd,
struct ib_qp *ret;
if (init_attr->cap.max_send_sge > ib_qib_max_sges ||
- init_attr->cap.max_send_wr > ib_qib_max_qp_wrs) {
+ init_attr->cap.max_send_wr > ib_qib_max_qp_wrs ||
+ init_attr->create_flags) {
ret = ERR_PTR(-EINVAL);
goto bail;
}
diff --git a/drivers/infiniband/hw/usnic/usnic_ib_verbs.c b/drivers/infiniband/hw/usnic/usnic_ib_verbs.c
index d48d2c0a2e3..53bd6a2d9cd 100644
--- a/drivers/infiniband/hw/usnic/usnic_ib_verbs.c
+++ b/drivers/infiniband/hw/usnic/usnic_ib_verbs.c
@@ -466,6 +466,9 @@ struct ib_qp *usnic_ib_create_qp(struct ib_pd *pd,
ucontext = to_uucontext(pd->uobject->context);
us_ibdev = to_usdev(pd->device);
+ if (init_attr->create_flags)
+ return ERR_PTR(-EINVAL);
+
err = ib_copy_from_udata(&cmd, udata, sizeof(cmd));
if (err) {
usnic_err("%s: cannot copy udata for create_qp\n",
diff --git a/drivers/infiniband/hw/usnic/usnic_uiom_interval_tree.c b/drivers/infiniband/hw/usnic/usnic_uiom_interval_tree.c
index d135ad90d91..3a4288e0fba 100644
--- a/drivers/infiniband/hw/usnic/usnic_uiom_interval_tree.c
+++ b/drivers/infiniband/hw/usnic/usnic_uiom_interval_tree.c
@@ -1,3 +1,21 @@
+/*
+ * Copyright (c) 2014, Cisco Systems, Inc. All rights reserved.
+ *
+ * This program is free software; you may redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; version 2 of the License.
+ *
+ * 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/init.h>
#include <linux/list.h>
#include <linux/slab.h>