int rc;
int err = -EIO;
- LASSERT (net != NULL);
- LASSERT (rx->rx_nob < 0); /* was posted */
- rx->rx_nob = 0; /* isn't now */
+ LASSERT(net);
+ LASSERT(rx->rx_nob < 0); /* was posted */
+ rx->rx_nob = 0; /* isn't now */
- if (conn->ibc_state > IBLND_CONN_ESTABLISHED)
- goto ignore;
+ if (conn->ibc_state > IBLND_CONN_ESTABLISHED)
+ goto ignore;
- if (status != IB_WC_SUCCESS) {
- CNETERR("Rx from %s failed: %d\n",
- libcfs_nid2str(conn->ibc_peer->ibp_nid), status);
- goto failed;
- }
+ if (status != IB_WC_SUCCESS) {
+ CNETERR("Rx from %s failed: %d\n",
+ libcfs_nid2str(conn->ibc_peer->ibp_nid), status);
+ goto failed;
+ }
- LASSERT (nob >= 0);
- rx->rx_nob = nob;
+ LASSERT(nob >= 0);
+ rx->rx_nob = nob;
- rc = kiblnd_unpack_msg(msg, rx->rx_nob);
- if (rc != 0) {
- CERROR ("Error %d unpacking rx from %s\n",
- rc, libcfs_nid2str(conn->ibc_peer->ibp_nid));
- goto failed;
- }
+ rc = kiblnd_unpack_msg(msg, rx->rx_nob);
+ if (rc != 0) {
+ CERROR("Error %d unpacking rx from %s\n",
+ rc, libcfs_nid2str(conn->ibc_peer->ibp_nid));
+ goto failed;
+ }
- if (msg->ibm_srcnid != conn->ibc_peer->ibp_nid ||
- msg->ibm_dstnid != ni->ni_nid ||
- msg->ibm_srcstamp != conn->ibc_incarnation ||
- msg->ibm_dststamp != net->ibn_incarnation) {
- CERROR ("Stale rx from %s\n",
- libcfs_nid2str(conn->ibc_peer->ibp_nid));
- err = -ESTALE;
- goto failed;
- }
+ if (msg->ibm_srcnid != conn->ibc_peer->ibp_nid ||
+ msg->ibm_dstnid != lnet_nid_to_nid4(&ni->ni_nid) ||
+ msg->ibm_srcstamp != conn->ibc_incarnation ||
+ msg->ibm_dststamp != net->ibn_incarnation) {
+ CERROR("Stale rx from %s\n",
+ libcfs_nid2str(conn->ibc_peer->ibp_nid));
+ err = -ESTALE;
+ goto failed;
+ }
- /* set time last known alive */
- kiblnd_peer_alive(conn->ibc_peer);
+ /* set time last known alive */
+ kiblnd_peer_alive(conn->ibc_peer);
- /* racing with connection establishment/teardown! */
+ /* racing with connection establishment/teardown! */
- if (conn->ibc_state < IBLND_CONN_ESTABLISHED) {
+ if (conn->ibc_state < IBLND_CONN_ESTABLISHED) {
rwlock_t *g_lock = &kiblnd_data.kib_global_lock;
unsigned long flags;
return;
}
write_unlock_irqrestore(g_lock, flags);
- }
- kiblnd_handle_rx(rx);
- return;
+ }
+ kiblnd_handle_rx(rx);
+ return;
- failed:
- CDEBUG(D_NET, "rx %p conn %p\n", rx, conn);
- kiblnd_close_conn(conn, err);
- ignore:
- kiblnd_drop_rx(rx); /* Don't re-post rx. */
+failed:
+ CDEBUG(D_NET, "rx %p conn %p\n", rx, conn);
+ kiblnd_close_conn(conn, err);
+ignore:
+ kiblnd_drop_rx(rx); /* Don't re-post rx. */
}
static int
kiblnd_fmr_pool_unmap(&tx->tx_fmr, tx->tx_status);
if (tx->tx_nfrags != 0) {
- kiblnd_dma_unmap_sg(tx->tx_pool->tpo_hdev->ibh_ibdev,
+ kiblnd_dma_unmap_sg(tx->tx_pool->tpo_hdev,
tx->tx_frags, tx->tx_nfrags, tx->tx_dmadir);
tx->tx_nfrags = 0;
}
tx->tx_dmadir = (rd != tx->tx_rd) ? DMA_FROM_DEVICE : DMA_TO_DEVICE;
tx->tx_nfrags = nfrags;
- rd->rd_nfrags = kiblnd_dma_map_sg(hdev->ibh_ibdev, tx->tx_frags,
+ rd->rd_nfrags = kiblnd_dma_map_sg(hdev, tx->tx_frags,
tx->tx_nfrags, tx->tx_dmadir);
for (i = 0, nob = 0; i < rd->rd_nfrags; i++) {
{
struct kib_net *net = ni->ni_data;
struct scatterlist *sg;
- int fragnob;
- int max_nkiov;
+ int fragnob;
+ int max_nkiov;
+ int sg_count = 0;
CDEBUG(D_NET, "niov %d offset %d nob %d\n", nkiov, offset, nob);
do {
LASSERT(nkiov > 0);
+ if (!sg) {
+ CERROR("lacking enough sg entries to map tx\n");
+ return -EFAULT;
+ }
+ sg_count++;
+
fragnob = min((int)(kiov->bv_len - offset), nob);
/*
sg_set_page(sg, kiov->bv_page, fragnob,
kiov->bv_offset + offset);
sg = sg_next(sg);
- if (!sg) {
- CERROR("lacking enough sg entries to map tx\n");
- return -EFAULT;
- }
offset = 0;
kiov++;
nob -= fragnob;
} while (nob > 0);
- return kiblnd_map_tx(ni, tx, rd, sg - tx->tx_frags);
+ return kiblnd_map_tx(ni, tx, rd, sg_count);
}
static int
struct kib_msg *msg = tx->tx_msg;
struct kib_peer_ni *peer_ni = conn->ibc_peer;
struct lnet_ni *ni = peer_ni->ibp_ni;
+ struct kib_fast_reg_descriptor *frd = tx->tx_fmr.fmr_frd;
int ver = conn->ibc_version;
int rc;
int done;
/* close_conn will launch failover */
rc = -ENETDOWN;
} else {
- struct kib_fast_reg_descriptor *frd = tx->tx_fmr.fmr_frd;
struct ib_send_wr *bad = &tx->tx_wrq[tx->tx_nwrq - 1].wr;
struct ib_send_wr *wr = &tx->tx_wrq[0].wr;
- if (frd != NULL) {
+ if (frd != NULL && !frd->frd_posted) {
if (!frd->frd_valid) {
wr = &frd->frd_inv_wr.wr;
wr->next = &frd->frd_fastreg_wr.wr;
conn->ibc_last_send = ktime_get();
- if (rc == 0)
- return 0;
+ if (rc == 0) {
+ if (frd != NULL)
+ frd->frd_posted = true;
+ return 0;
+ }
/* NB credits are transferred in the actual
* message, which can only be the last work item */
#endif
LASSERT(tx->tx_nwrq >= 0);
- LASSERT(tx->tx_nwrq < IBLND_MAX_RDMA_FRAGS + 1);
+ LASSERT(tx->tx_nwrq <= IBLND_MAX_RDMA_FRAGS);
LASSERT(nob <= IBLND_MSG_SIZE);
#ifdef HAVE_IB_GET_DMA_MR
LASSERT(mr != NULL);
/* Thread context */
LASSERT (!in_interrupt());
+ tx = kiblnd_get_idle_tx(ni, target.nid);
+ if (tx == NULL) {
+ CERROR("Can't allocate %s txd for %s\n",
+ lnet_msgtyp2str(type),
+ libcfs_nid2str(target.nid));
+ return -ENOMEM;
+ }
+ ibmsg = tx->tx_msg;
+
switch (type) {
default:
LBUG();
/* is the REPLY message too small for RDMA? */
nob = offsetof(struct kib_msg, ibm_u.immediate.ibim_payload[lntmsg->msg_md->md_length]);
- if (nob <= IBLND_MSG_SIZE && !lntmsg->msg_rdma_force)
- break; /* send IMMEDIATE */
+ if (nob <= IBLND_MSG_SIZE && !lntmsg->msg_rdma_force)
+ break; /* send IMMEDIATE */
- tx = kiblnd_get_idle_tx(ni, target.nid);
- if (tx == NULL) {
- CERROR("Can't allocate txd for GET to %s\n",
- libcfs_nid2str(target.nid));
- return -ENOMEM;
- }
-
- ibmsg = tx->tx_msg;
rd = &ibmsg->ibm_u.get.ibgm_rd;
rc = kiblnd_setup_rd_kiov(ni, tx, rd,
lntmsg->msg_md->md_niov,
return -EIO;
}
- tx->tx_lntmsg[0] = lntmsg; /* finalise lntmsg[0,1] on completion */
- tx->tx_waiting = 1; /* waiting for GET_DONE */
- kiblnd_launch_tx(ni, tx, target.nid);
- return 0;
+ /* finalise lntmsg[0,1] on completion */
+ tx->tx_lntmsg[0] = lntmsg;
+ tx->tx_waiting = 1; /* waiting for GET_DONE */
+ kiblnd_launch_tx(ni, tx, target.nid);
+ return 0;
- case LNET_MSG_REPLY:
- case LNET_MSG_PUT:
- /* Is the payload small enough not to need RDMA? */
+ case LNET_MSG_REPLY:
+ case LNET_MSG_PUT:
+ /* Is the payload small enough not to need RDMA? */
nob = offsetof(struct kib_msg, ibm_u.immediate.ibim_payload[payload_nob]);
- if (nob <= IBLND_MSG_SIZE && !lntmsg->msg_rdma_force)
- break; /* send IMMEDIATE */
-
- tx = kiblnd_get_idle_tx(ni, target.nid);
- if (tx == NULL) {
- CERROR("Can't allocate %s txd for %s\n",
- type == LNET_MSG_PUT ? "PUT" : "REPLY",
- libcfs_nid2str(target.nid));
- return -ENOMEM;
- }
+ if (nob <= IBLND_MSG_SIZE && !lntmsg->msg_rdma_force)
+ break; /* send IMMEDIATE */
rc = kiblnd_setup_rd_kiov(ni, tx, tx->tx_rd,
payload_niov, payload_kiov,
return -EIO;
}
- ibmsg = tx->tx_msg;
- ibmsg->ibm_u.putreq.ibprm_hdr = *hdr;
- ibmsg->ibm_u.putreq.ibprm_cookie = tx->tx_cookie;
+ ibmsg->ibm_u.putreq.ibprm_hdr = *hdr;
+ ibmsg->ibm_u.putreq.ibprm_cookie = tx->tx_cookie;
kiblnd_init_tx_msg(ni, tx, IBLND_MSG_PUT_REQ,
sizeof(struct kib_putreq_msg));
- tx->tx_lntmsg[0] = lntmsg; /* finalise lntmsg on completion */
- tx->tx_waiting = 1; /* waiting for PUT_{ACK,NAK} */
- kiblnd_launch_tx(ni, tx, target.nid);
- return 0;
- }
+ /* finalise lntmsg[0,1] on completion */
+ tx->tx_lntmsg[0] = lntmsg;
+ tx->tx_waiting = 1; /* waiting for PUT_{ACK,NAK} */
+ kiblnd_launch_tx(ni, tx, target.nid);
+ return 0;
+ }
/* send IMMEDIATE */
LASSERT(offsetof(struct kib_msg, ibm_u.immediate.ibim_payload[payload_nob])
<= IBLND_MSG_SIZE);
- tx = kiblnd_get_idle_tx(ni, target.nid);
- if (tx == NULL) {
- CERROR ("Can't send %d to %s: tx descs exhausted\n",
- type, libcfs_nid2str(target.nid));
- return -ENOMEM;
- }
-
- ibmsg = tx->tx_msg;
- ibmsg->ibm_u.immediate.ibim_hdr = *hdr;
+ ibmsg = tx->tx_msg;
+ ibmsg->ibm_u.immediate.ibim_hdr = *hdr;
lnet_copy_kiov2flat(IBLND_MSG_SIZE, ibmsg,
offsetof(struct kib_msg,
payload_offset, payload_nob);
nob = offsetof(struct kib_immediate_msg, ibim_payload[payload_nob]);
- kiblnd_init_tx_msg(ni, tx, IBLND_MSG_IMMEDIATE, nob);
- tx->tx_lntmsg[0] = lntmsg; /* finalise lntmsg on completion */
- kiblnd_launch_tx(ni, tx, target.nid);
- return 0;
+ kiblnd_init_tx_msg(ni, tx, IBLND_MSG_IMMEDIATE, nob);
+
+ /* finalise lntmsg on completion */
+ tx->tx_lntmsg[0] = lntmsg;
+
+ kiblnd_launch_tx(ni, tx, target.nid);
+ return 0;
}
static void
lnet_finalize(lntmsg, -EIO);
}
+unsigned int
+kiblnd_get_dev_prio(struct lnet_ni *ni, unsigned int dev_idx)
+{
+ struct kib_net *net = ni->ni_data;
+ struct device *dev = NULL;
+
+ if (net)
+ dev = net->ibn_dev->ibd_hdev->ibh_ibdev->dma_device;
+
+ return lnet_get_dev_prio(dev, dev_idx);
+
+}
+
int
kiblnd_recv(struct lnet_ni *ni, void *private, struct lnet_msg *lntmsg,
int delayed, unsigned int niov, struct bio_vec *kiov,
return rc;
}
-int
-kiblnd_thread_start(int (*fn)(void *arg), void *arg, char *name)
-{
- struct task_struct *task = kthread_run(fn, arg, "%s", name);
-
- if (IS_ERR(task))
- return PTR_ERR(task);
-
- atomic_inc(&kiblnd_data.kib_nthreads);
- return 0;
-}
-
static void
kiblnd_thread_fini (void)
{
rej.ibr_incarnation = net->ibn_incarnation;
}
- if (ni == NULL || /* no matching net */
- ni->ni_nid != reqmsg->ibm_dstnid || /* right NET, wrong NID! */
- net->ibn_dev != ibdev) { /* wrong device */
+ if (ni == NULL || /* no matching net */
+ lnet_nid_to_nid4(&ni->ni_nid) !=
+ reqmsg->ibm_dstnid || /* right NET, wrong NID! */
+ net->ibn_dev != ibdev) { /* wrong device */
CERROR("Can't accept conn from %s on %s (%s:%d:%pI4h): bad dst nid %s\n", libcfs_nid2str(nid),
- ni ? libcfs_nid2str(ni->ni_nid) : "NA",
+ ni ? libcfs_nidstr(&ni->ni_nid) : "NA",
ibdev->ibd_ifname, ibdev->ibd_nnets,
&ibdev->ibd_ifip,
libcfs_nid2str(reqmsg->ibm_dstnid));
* the lower NID connection win so we can move forward.
*/
if (peer2->ibp_connecting != 0 &&
- nid < ni->ni_nid && peer2->ibp_races <
- MAX_CONN_RACES_BEFORE_ABORT) {
+ nid < lnet_nid_to_nid4(&ni->ni_nid) &&
+ peer2->ibp_races < MAX_CONN_RACES_BEFORE_ABORT) {
peer2->ibp_races++;
write_unlock_irqrestore(g_lock, flags);
break;
case IBLND_REJECT_RDMA_FRAGS: {
- struct lnet_ioctl_config_o2iblnd_tunables *tunables;
-
if (!cp) {
reason = "can't negotiate max frags";
goto out;
}
- tunables = &peer_ni->ibp_ni->ni_lnd_tunables.lnd_tun_u.lnd_o2ib;
-#ifdef HAVE_IB_GET_DMA_MR
- /*
- * This check only makes sense if the kernel supports global
- * memory registration. Otherwise, map_on_demand will never == 0
- */
- if (!tunables->lnd_map_on_demand) {
- reason = "map_on_demand must be enabled";
- goto out;
- }
-#endif
+
if (conn->ibc_max_frags <= frag_num) {
reason = "unsupported max frags";
goto out;
}
read_lock_irqsave(&kiblnd_data.kib_global_lock, flags);
- if (msg->ibm_dstnid == ni->ni_nid &&
+ if (msg->ibm_dstnid == lnet_nid_to_nid4(&ni->ni_nid) &&
msg->ibm_dststamp == net->ibn_incarnation)
rc = 0;
else
}
conn = list_first_entry_or_null(&kiblnd_data.kib_connd_waits,
- struct kib_conn,
- ibc_sched_list);
+ struct kib_conn, ibc_list);
if (conn) {
list_del(&conn->ibc_list);
spin_unlock_irqrestore(lock, flags);
case IB_EVENT_PORT_ERR:
case IB_EVENT_DEVICE_FATAL:
CERROR("Fatal device error for NI %s\n",
- libcfs_nid2str(conn->ibc_peer->ibp_ni->ni_nid));
+ libcfs_nidstr(&conn->ibc_peer->ibp_ni->ni_nid));
atomic_set(&conn->ibc_peer->ibp_ni->ni_fatal_error_on, 1);
return;
case IB_EVENT_PORT_ACTIVE:
CERROR("Port reactivated for NI %s\n",
- libcfs_nid2str(conn->ibc_peer->ibp_ni->ni_nid));
+ libcfs_nidstr(&conn->ibc_peer->ibp_ni->ni_nid));
atomic_set(&conn->ibc_peer->ibp_ni->ni_fatal_error_on, 0);
return;