summaryrefslogtreecommitdiffstats
path: root/drivers/infiniband
diff options
context:
space:
mode:
authorLinus Torvalds <torvalds@linux-foundation.org>2014-06-12 14:27:40 -0700
committerLinus Torvalds <torvalds@linux-foundation.org>2014-06-12 14:27:40 -0700
commitf9da455b93f6ba076935b4ef4589f61e529ae046 (patch)
tree3c4e69ce1ba1d6bf65915b97a76ca2172105b278 /drivers/infiniband
parent0e04c641b199435f3779454055f6a7de258ecdfc (diff)
parente5eca6d41f53db48edd8cf88a3f59d2c30227f8e (diff)
downloadop-kernel-dev-f9da455b93f6ba076935b4ef4589f61e529ae046.zip
op-kernel-dev-f9da455b93f6ba076935b4ef4589f61e529ae046.tar.gz
Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next
Pull networking updates from David Miller: 1) Seccomp BPF filters can now be JIT'd, from Alexei Starovoitov. 2) Multiqueue support in xen-netback and xen-netfront, from Andrew J Benniston. 3) Allow tweaking of aggregation settings in cdc_ncm driver, from Bjørn Mork. 4) BPF now has a "random" opcode, from Chema Gonzalez. 5) Add more BPF documentation and improve test framework, from Daniel Borkmann. 6) Support TCP fastopen over ipv6, from Daniel Lee. 7) Add software TSO helper functions and use them to support software TSO in mvneta and mv643xx_eth drivers. From Ezequiel Garcia. 8) Support software TSO in fec driver too, from Nimrod Andy. 9) Add Broadcom SYSTEMPORT driver, from Florian Fainelli. 10) Handle broadcasts more gracefully over macvlan when there are large numbers of interfaces configured, from Herbert Xu. 11) Allow more control over fwmark used for non-socket based responses, from Lorenzo Colitti. 12) Do TCP congestion window limiting based upon measurements, from Neal Cardwell. 13) Support busy polling in SCTP, from Neal Horman. 14) Allow RSS key to be configured via ethtool, from Venkata Duvvuru. 15) Bridge promisc mode handling improvements from Vlad Yasevich. 16) Don't use inetpeer entries to implement ID generation any more, it performs poorly, from Eric Dumazet. * git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (1522 commits) rtnetlink: fix userspace API breakage for iproute2 < v3.9.0 tcp: fixing TLP's FIN recovery net: fec: Add software TSO support net: fec: Add Scatter/gather support net: fec: Increase buffer descriptor entry number net: fec: Factorize feature setting net: fec: Enable IP header hardware checksum net: fec: Factorize the .xmit transmit function bridge: fix compile error when compiling without IPv6 support bridge: fix smatch warning / potential null pointer dereference via-rhine: fix full-duplex with autoneg disable bnx2x: Enlarge the dorq threshold for VFs bnx2x: Check for UNDI in uncommon branch bnx2x: Fix 1G-baseT link bnx2x: Fix link for KR with swapped polarity lane sctp: Fix sk_ack_backlog wrap-around problem net/core: Add VF link state control policy net/fsl: xgmac_mdio is dependent on OF_MDIO net/fsl: Make xgmac_mdio read error message useful net_sched: drr: warn when qdisc is not work conserving ...
Diffstat (limited to 'drivers/infiniband')
-rw-r--r--drivers/infiniband/hw/cxgb4/cm.c127
-rw-r--r--drivers/infiniband/hw/cxgb4/cq.c7
-rw-r--r--drivers/infiniband/hw/cxgb4/iw_cxgb4.h2
-rw-r--r--drivers/infiniband/hw/cxgb4/provider.c2
-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/ulp/ipoib/ipoib_ethtool.c2
7 files changed, 124 insertions, 18 deletions
diff --git a/drivers/infiniband/hw/cxgb4/cm.c b/drivers/infiniband/hw/cxgb4/cm.c
index 96d7131..5e153f6 100644
--- a/drivers/infiniband/hw/cxgb4/cm.c
+++ b/drivers/infiniband/hw/cxgb4/cm.c
@@ -234,12 +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);
}
@@ -473,7 +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 */
@@ -565,6 +569,17 @@ static void c4iw_record_pm_msg(struct c4iw_ep *ep,
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;
@@ -591,6 +606,7 @@ static int send_connect(struct c4iw_ep *ep)
&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) :
@@ -606,8 +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) |
@@ -618,7 +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);
@@ -674,6 +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);
@@ -690,6 +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 *)
@@ -713,6 +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);
}
}
@@ -1186,6 +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);
@@ -1659,6 +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));
@@ -1681,8 +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) |
@@ -1694,7 +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) |
@@ -1731,6 +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,
@@ -1779,6 +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);
@@ -1797,6 +1863,7 @@ static int import_ep(struct c4iw_ep *ep, int iptype, __u8 *peer_ip,
cdev->rdev.lldi.nchan;
ep->rss_qid = cdev->rdev.lldi.rxq_ids[
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;
@@ -2027,13 +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) |
@@ -2044,7 +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);
@@ -2064,14 +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);
@@ -2136,6 +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) {
@@ -2193,8 +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;
diff --git a/drivers/infiniband/hw/cxgb4/cq.c b/drivers/infiniband/hw/cxgb4/cq.c
index 7151a02..c04292c 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);
diff --git a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
index 6f533fb..125bc5d 100644
--- a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
+++ b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
@@ -810,6 +810,8 @@ 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,
diff --git a/drivers/infiniband/hw/cxgb4/provider.c b/drivers/infiniband/hw/cxgb4/provider.c
index c777e22..b1d3053 100644
--- a/drivers/infiniband/hw/cxgb4/provider.c
+++ b/drivers/infiniband/hw/cxgb4/provider.c
@@ -500,7 +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 2178f31..68b0a6b 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 6121ca0..91289a0 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/ulp/ipoib/ipoib_ethtool.c b/drivers/infiniband/ulp/ipoib/ipoib_ethtool.c
index c4b3940..078cadd 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_ethtool.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_ethtool.c
@@ -105,5 +105,5 @@ static const struct ethtool_ops ipoib_ethtool_ops = {
void ipoib_set_ethtool_ops(struct net_device *dev)
{
- SET_ETHTOOL_OPS(dev, &ipoib_ethtool_ops);
+ dev->ethtool_ops = &ipoib_ethtool_ops;
}
OpenPOWER on IntegriCloud