static int l2cap_sock_accept(struct socket *sock, struct socket *newsock,
int flags)
{
DEFINE_WAIT_FUNC(wait, woken_wake_function);
struct sock *sk = sock->sk, *nsk;
long timeo;
int err = 0;
lock_sock_nested(sk, L2CAP_NESTING_PARENT);
timeo = sock_rcvtimeo(sk, flags & O_NONBLOCK);
BT_DBG("sk %p timeo %ld", sk, timeo);
/* Wait for an incoming connection. (wake-one). */
add_wait_queue_exclusive(sk_sleep(sk), &wait);
while (1) {
if (sk->sk_state != BT_LISTEN) {
err = -EBADFD;
break;
}
nsk = bt_accept_dequeue(sk, newsock);
if (nsk)
break;
if (!timeo) {
err = -EAGAIN;
break;
}
if (signal_pending(current)) {
err = sock_intr_errno(timeo);
break;
}
release_sock(sk);
timeo = wait_woken(&wait, TASK_INTERRUPTIBLE, timeo);
lock_sock_nested(sk, L2CAP_NESTING_PARENT);
}
remove_wait_queue(sk_sleep(sk), &wait);
if (err)
goto done;
newsock->state = SS_CONNECTED;
BT_DBG("new socket %p", nsk);
done:
release_sock(sk);
return err;
}
static int l2cap_sock_accept(struct socket *sock, struct socket *newsock, int flags)
{
DECLARE_WAITQUEUE(wait, current);
struct sock *sk = sock->sk, *nsk;
long timeo;
int err = 0;
lock_sock_nested(sk, SINGLE_DEPTH_NESTING);
if (sk->sk_state != BT_LISTEN) {
err = -EBADFD;
goto done;
}
timeo = sock_rcvtimeo(sk, flags & O_NONBLOCK);
BT_DBG("sk %p timeo %ld", sk, timeo);
/* Wait for an incoming connection. (wake-one). */
add_wait_queue_exclusive(sk_sleep(sk), &wait);
while (!(nsk = bt_accept_dequeue(sk, newsock))) {
set_current_state(TASK_INTERRUPTIBLE);
if (!timeo) {
err = -EAGAIN;
break;
}
release_sock(sk);
timeo = schedule_timeout(timeo);
lock_sock_nested(sk, SINGLE_DEPTH_NESTING);
if (sk->sk_state != BT_LISTEN) {
err = -EBADFD;
break;
}
if (signal_pending(current)) {
err = sock_intr_errno(timeo);
break;
}
}
set_current_state(TASK_RUNNING);
remove_wait_queue(sk_sleep(sk), &wait);
if (err)
goto done;
newsock->state = SS_CONNECTED;
BT_DBG("new socket %p", nsk);
done:
release_sock(sk);
return err;
}
void vsock_pending_work(struct work_struct *work)
{
struct sock *sk;
struct sock *listener;
struct vsock_sock *vsk;
bool cleanup;
vsk = container_of(work, struct vsock_sock, dwork.work);
sk = sk_vsock(vsk);
listener = vsk->listener;
cleanup = true;
lock_sock(listener);
lock_sock_nested(sk, SINGLE_DEPTH_NESTING);
if (vsock_is_pending(sk)) {
vsock_remove_pending(listener, sk);
listener->sk_ack_backlog--;
} else if (!vsk->rejected) {
/* We are not on the pending list and accept() did not reject
* us, so we must have been accepted by our user process. We
* just need to drop our references to the sockets and be on
* our way.
*/
cleanup = false;
goto out;
}
/* We need to remove ourself from the global connected sockets list so
* incoming packets can't find this socket, and to reduce the reference
* count.
*/
if (vsock_in_connected_table(vsk))
vsock_remove_connected(vsk);
sk->sk_state = TCP_CLOSE;
out:
release_sock(sk);
release_sock(listener);
if (cleanup)
sock_put(sk);
sock_put(sk);
sock_put(listener);
}
static int vsock_accept(struct socket *sock, struct socket *newsock, int flags,
bool kern)
{
struct sock *listener;
int err;
struct sock *connected;
struct vsock_sock *vconnected;
long timeout;
DEFINE_WAIT(wait);
err = 0;
listener = sock->sk;
lock_sock(listener);
if (sock->type != SOCK_STREAM) {
err = -EOPNOTSUPP;
goto out;
}
if (listener->sk_state != TCP_LISTEN) {
err = -EINVAL;
goto out;
}
/* Wait for children sockets to appear; these are the new sockets
* created upon connection establishment.
*/
timeout = sock_sndtimeo(listener, flags & O_NONBLOCK);
prepare_to_wait(sk_sleep(listener), &wait, TASK_INTERRUPTIBLE);
while ((connected = vsock_dequeue_accept(listener)) == NULL &&
listener->sk_err == 0) {
release_sock(listener);
timeout = schedule_timeout(timeout);
finish_wait(sk_sleep(listener), &wait);
lock_sock(listener);
if (signal_pending(current)) {
err = sock_intr_errno(timeout);
goto out;
} else if (timeout == 0) {
err = -EAGAIN;
goto out;
}
prepare_to_wait(sk_sleep(listener), &wait, TASK_INTERRUPTIBLE);
}
finish_wait(sk_sleep(listener), &wait);
if (listener->sk_err)
err = -listener->sk_err;
if (connected) {
listener->sk_ack_backlog--;
lock_sock_nested(connected, SINGLE_DEPTH_NESTING);
vconnected = vsock_sk(connected);
/* If the listener socket has received an error, then we should
* reject this socket and return. Note that we simply mark the
* socket rejected, drop our reference, and let the cleanup
* function handle the cleanup; the fact that we found it in
* the listener's accept queue guarantees that the cleanup
* function hasn't run yet.
*/
if (err) {
vconnected->rejected = true;
} else {
newsock->state = SS_CONNECTED;
sock_graft(connected, newsock);
}
release_sock(connected);
sock_put(connected);
}
out:
release_sock(listener);
return err;
}
static int sdp_init_qp(struct sock *sk, struct rdma_cm_id *id)
{
struct ib_qp_init_attr qp_init_attr = {
.event_handler = sdp_qp_event_handler,
.cap.max_send_wr = SDP_TX_SIZE,
.cap.max_recv_wr = sdp_rx_size,
.cap.max_inline_data = sdp_inline_thresh,
.sq_sig_type = IB_SIGNAL_REQ_WR,
.qp_type = IB_QPT_RC,
};
struct ib_device *device = id->device;
int rc;
sdp_dbg(sk, "%s\n", __func__);
sdp_sk(sk)->max_sge = sdp_get_max_dev_sge(device);
sdp_dbg(sk, "Max sges: %d\n", sdp_sk(sk)->max_sge);
qp_init_attr.cap.max_send_sge = MIN(sdp_sk(sk)->max_sge, SDP_MAX_SEND_SGES);
sdp_dbg(sk, "Setting max send sge to: %d\n", qp_init_attr.cap.max_send_sge);
qp_init_attr.cap.max_recv_sge = MIN(sdp_sk(sk)->max_sge, SDP_MAX_RECV_SGES);
sdp_dbg(sk, "Setting max recv sge to: %d\n", qp_init_attr.cap.max_recv_sge);
sdp_sk(sk)->sdp_dev = ib_get_client_data(device, &sdp_client);
if (!sdp_sk(sk)->sdp_dev) {
sdp_warn(sk, "SDP not available on device %s\n", device->name);
rc = -ENODEV;
goto err_rx;
}
rc = sdp_rx_ring_create(sdp_sk(sk), device);
if (rc)
goto err_rx;
rc = sdp_tx_ring_create(sdp_sk(sk), device);
if (rc)
goto err_tx;
qp_init_attr.recv_cq = sdp_sk(sk)->rx_ring.cq;
qp_init_attr.send_cq = sdp_sk(sk)->tx_ring.cq;
rc = rdma_create_qp(id, sdp_sk(sk)->sdp_dev->pd, &qp_init_attr);
if (rc) {
sdp_warn(sk, "Unable to create QP: %d.\n", rc);
goto err_qp;
}
sdp_sk(sk)->qp = id->qp;
sdp_sk(sk)->ib_device = device;
sdp_sk(sk)->qp_active = 1;
sdp_sk(sk)->context.device = device;
sdp_sk(sk)->inline_thresh = qp_init_attr.cap.max_inline_data;
sdp_dbg(sk, "%s done\n", __func__);
return 0;
err_qp:
sdp_tx_ring_destroy(sdp_sk(sk));
err_tx:
sdp_rx_ring_destroy(sdp_sk(sk));
err_rx:
return rc;
}
static int sdp_get_max_send_frags(u32 buf_size)
{
return MIN(
/* +1 to conpensate on not aligned buffers */
(PAGE_ALIGN(buf_size) >> PAGE_SHIFT) + 1,
SDP_MAX_SEND_SGES - 1);
}
static int sdp_connect_handler(struct sock *sk, struct rdma_cm_id *id,
struct rdma_cm_event *event)
{
struct sockaddr_in *dst_addr;
struct sock *child;
const struct sdp_hh *h;
int rc = 0;
sdp_dbg(sk, "%s %p -> %p\n", __func__, sdp_sk(sk)->id, id);
h = event->param.conn.private_data;
SDP_DUMP_PACKET(sk, "RX", NULL, &h->bsdh);
if (h->ipv_cap & HH_IPV_MASK & ~(HH_IPV4 | HH_IPV6)) {
sdp_warn(sk, "Bad IPV field in SDP Hello header: 0x%x\n",
h->ipv_cap & HH_IPV_MASK);
return -EINVAL;
}
if (!h->max_adverts)
return -EINVAL;
#if (LINUX_VERSION_CODE < KERNEL_VERSION(3, 3, 0))
child = sk_clone(sk, GFP_KERNEL);
#else
child = sk_clone_lock(sk, GFP_KERNEL);
#endif
if (!child)
return -ENOMEM;
sdp_init_sock(child);
dst_addr = (struct sockaddr_in *)&id->route.addr.dst_addr;
sdp_inet_dport(child) = dst_addr->sin_port;
#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
if (inet6_sk(sk)) {
struct ipv6_pinfo *newnp;
newnp = inet_sk(child)->pinet6 = sdp_inet6_sk_generic(child);
memcpy(newnp, inet6_sk(sk), sizeof(struct ipv6_pinfo));
if ((h->ipv_cap & HH_IPV_MASK) == HH_IPV4) {
/* V6 mapped */
sdp_inet_daddr(child) = dst_addr->sin_addr.s_addr;
ipv6_addr_set(&child->sk_v6_daddr, 0, 0, htonl(0x0000FFFF),
h->src_addr.ip4.addr);
ipv6_addr_set(&child->sk_v6_rcv_saddr, 0, 0, htonl(0x0000FFFF),
h->dst_addr.ip4.addr);
ipv6_addr_copy(&child->sk_v6_rcv_saddr, &child->sk_v6_daddr);
} else if ((h->ipv_cap & HH_IPV_MASK) == HH_IPV6) {
struct sockaddr_in6 *dst_addr6 = (struct sockaddr_in6 *)dst_addr;
struct sockaddr_in6 *src_addr6 =
(struct sockaddr_in6 *)&id->route.addr.src_addr;
ipv6_addr_copy(&child->sk_v6_daddr, &dst_addr6->sin6_addr);
ipv6_addr_copy(&child->sk_v6_rcv_saddr, &src_addr6->sin6_addr);
ipv6_addr_copy(&newnp->saddr, &src_addr6->sin6_addr);
} else {
sdp_warn(child, "Bad IPV field: 0x%x\n", h->ipv_cap & HH_IPV_MASK);
}
sdp_inet_daddr(child) = sdp_inet_saddr(child) =
sdp_inet_rcv_saddr(child) = LOOPBACK4_IPV6;
} else
#endif
{
sdp_inet_daddr(child) = dst_addr->sin_addr.s_addr;
}
#ifdef SDP_SOCK_HISTORY
sdp_ssk_hist_rename(sk);
#endif
__sock_put(child, SOCK_REF_CLONE);
down_read(&device_removal_lock);
rc = sdp_init_qp(child, id);
if (rc) {
bh_unlock_sock(child);
up_read(&device_removal_lock);
sdp_sk(child)->destructed_already = 1;
#ifdef SDP_SOCK_HISTORY
sdp_ssk_hist_close(child);
#endif
sk_free(child);
return rc;
}
sdp_sk(child)->max_bufs = ntohs(h->bsdh.bufs);
atomic_set(&sdp_sk(child)->tx_ring.credits, sdp_sk(child)->max_bufs);
sdp_sk(child)->min_bufs = tx_credits(sdp_sk(child)) / 4;
sdp_sk(child)->xmit_size_goal = ntohl(h->localrcvsz) -
sizeof(struct sdp_bsdh);
sdp_sk(child)->send_frags = sdp_get_max_send_frags(sdp_sk(child)->xmit_size_goal);
sdp_init_buffers(sdp_sk(child), rcvbuf_initial_size);
id->context = child;
sdp_sk(child)->id = id;
list_add_tail(&sdp_sk(child)->backlog_queue,
&sdp_sk(sk)->backlog_queue);
sdp_sk(child)->parent = sk;
bh_unlock_sock(child);
sdp_add_sock(sdp_sk(child));
up_read(&device_removal_lock);
sdp_exch_state(child, TCPF_LISTEN | TCPF_CLOSE, TCP_SYN_RECV);
/* child->sk_write_space(child); */
/* child->sk_data_ready(child, 0); */
sk->sk_data_ready(sk);
return 0;
}
static int sdp_response_handler(struct sock *sk, struct rdma_cm_id *id,
struct rdma_cm_event *event)
{
const struct sdp_hah *h;
struct sockaddr_in *dst_addr;
sdp_dbg(sk, "%s\n", __func__);
sdp_exch_state(sk, TCPF_SYN_SENT, TCP_ESTABLISHED);
sdp_set_default_moderation(sdp_sk(sk));
if (sock_flag(sk, SOCK_KEEPOPEN))
sdp_start_keepalive_timer(sk);
if (sock_flag(sk, SOCK_DEAD))
return 0;
h = event->param.conn.private_data;
SDP_DUMP_PACKET(sk, "RX", NULL, &h->bsdh);
sdp_sk(sk)->max_bufs = ntohs(h->bsdh.bufs);
atomic_set(&sdp_sk(sk)->tx_ring.credits, sdp_sk(sk)->max_bufs);
sdp_sk(sk)->min_bufs = tx_credits(sdp_sk(sk)) / 4;
sdp_sk(sk)->xmit_size_goal =
ntohl(h->actrcvsz) - sizeof(struct sdp_bsdh);
sdp_sk(sk)->send_frags = sdp_get_max_send_frags(sdp_sk(sk)->xmit_size_goal);
sdp_sk(sk)->xmit_size_goal = MIN(sdp_sk(sk)->xmit_size_goal,
sdp_sk(sk)->send_frags * PAGE_SIZE);
sdp_sk(sk)->poll_cq = 1;
sk->sk_state_change(sk);
sk_wake_async(sk, 0, POLL_OUT);
dst_addr = (struct sockaddr_in *)&id->route.addr.dst_addr;
sdp_inet_dport(sk) = dst_addr->sin_port;
sdp_inet_daddr(sk) = dst_addr->sin_addr.s_addr;
#ifdef SDP_SOCK_HISTORY
sdp_ssk_hist_rename(sk);
#endif
return 0;
}
static int sdp_connected_handler(struct sock *sk)
{
struct sock *parent;
sdp_dbg(sk, "%s\n", __func__);
parent = sdp_sk(sk)->parent;
BUG_ON(!parent);
sdp_exch_state(sk, TCPF_SYN_RECV, TCP_ESTABLISHED);
#ifdef SDP_SOCK_HISTORY
sdp_ssk_hist_rename(sk);
#endif
sdp_set_default_moderation(sdp_sk(sk));
if (sock_flag(sk, SOCK_KEEPOPEN))
sdp_start_keepalive_timer(sk);
if (sock_flag(sk, SOCK_DEAD))
return 0;
lock_sock(parent);
if (!sdp_sk(parent)->id) { /* TODO: look at SOCK_DEAD? */
sdp_dbg(sk, "parent is going away.\n");
goto done;
}
sk_acceptq_added(parent);
sdp_dbg(parent, "%s child connection established\n", __func__);
list_del_init(&sdp_sk(sk)->backlog_queue);
list_add_tail(&sdp_sk(sk)->accept_queue,
&sdp_sk(parent)->accept_queue);
parent->sk_state_change(parent);
sk_wake_async(parent, 0, POLL_OUT);
done:
release_sock(parent);
return 0;
}
static int sdp_disconnected_handler(struct sock *sk)
{
struct sdp_sock *ssk = sdp_sk(sk);
sdp_dbg(sk, "%s\n", __func__);
if (ssk->tx_ring.cq)
if (sdp_xmit_poll(ssk, 1))
sdp_post_sends(ssk, 0);
if (sk->sk_state == TCP_SYN_RECV) {
sdp_connected_handler(sk);
if (rcv_nxt(ssk))
return 0;
}
return -ECONNRESET;
}
int sdp_cma_handler(struct rdma_cm_id *id, struct rdma_cm_event *event)
{
struct rdma_conn_param conn_param;
struct sock *parent = NULL;
struct sock *child = NULL;
struct sock *sk;
struct sdp_hah hah;
struct sdp_hh hh;
int rc = 0, rc2;
sk = id->context;
if (!sk) {
sdp_dbg(NULL, "cm_id is being torn down, event %s\n",
rdma_cm_event_str(event->event));
return event->event == RDMA_CM_EVENT_CONNECT_REQUEST ?
-EINVAL : 0;
}
sdp_add_to_history(sk, rdma_cm_event_str(event->event));
lock_sock_nested(sk, SINGLE_DEPTH_NESTING);
sdp_dbg(sk, "event: %s\n", rdma_cm_event_str(event->event));
if (!sdp_sk(sk)->id) {
sdp_dbg(sk, "socket is being torn down\n");
rc = event->event == RDMA_CM_EVENT_CONNECT_REQUEST ?
-EINVAL : 0;
release_sock(sk);
return rc;
}
switch (event->event) {
case RDMA_CM_EVENT_ADDR_RESOLVED:
if (sdp_link_layer_ib_only &&
rdma_node_get_transport(id->device->node_type) ==
RDMA_TRANSPORT_IB &&
rdma_port_get_link_layer(id->device, id->port_num) !=
IB_LINK_LAYER_INFINIBAND) {
sdp_dbg(sk, "Link layer is: %d. Only IB link layer "
"is allowed\n",
rdma_port_get_link_layer(id->device,
id->port_num));
rc = -ENETUNREACH;
break;
}
rc = rdma_resolve_route(id, SDP_ROUTE_TIMEOUT);
break;
case RDMA_CM_EVENT_ADDR_ERROR:
rc = -ENETUNREACH;
break;
case RDMA_CM_EVENT_ROUTE_RESOLVED:
rc = sdp_init_qp(sk, id);
if (rc)
break;
memset(&hh, 0, sizeof hh);
hh.bsdh.mid = SDP_MID_HELLO;
hh.bsdh.len = htonl(sizeof(struct sdp_hh));
hh.max_adverts = 1;
hh.majv_minv = SDP_MAJV_MINV;
sdp_init_buffers(sdp_sk(sk), rcvbuf_initial_size);
hh.bsdh.bufs = htons(rx_ring_posted(sdp_sk(sk)));
atomic_set(&sdp_sk(sk)->remote_credits,
rx_ring_posted(sdp_sk(sk)));
hh.localrcvsz = hh.desremrcvsz = htonl(sdp_sk(sk)->recv_frags *
PAGE_SIZE + sizeof(struct sdp_bsdh));
#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
if (inet6_sk(sk)) {
struct sockaddr *src_addr = (struct sockaddr *)&id->route.addr.src_addr;
struct sockaddr_in *addr4 = (struct sockaddr_in *)src_addr;
struct sockaddr_in6 *addr6 = (struct sockaddr_in6 *)src_addr;
if (src_addr->sa_family == AF_INET) {
/* IPv4 over IPv6 */
ipv6_addr_set(&sk->sk_v6_rcv_saddr, 0, 0, htonl(0xFFFF),
addr4->sin_addr.s_addr);
} else {
sk->sk_v6_rcv_saddr = addr6->sin6_addr;
}
inet6_sk(sk)->saddr = sk->sk_v6_rcv_saddr;
}
else
#endif
{
sdp_inet_saddr(sk) = sdp_inet_rcv_saddr(sk) =
((struct sockaddr_in *)&id->route.addr.src_addr)->sin_addr.s_addr;
}
memset(&conn_param, 0, sizeof conn_param);
conn_param.private_data_len = sizeof hh;
conn_param.private_data = &hh;
conn_param.responder_resources = 4 /* TODO */;
conn_param.initiator_depth = 4 /* TODO */;
conn_param.retry_count = sdp_retry_count;
SDP_DUMP_PACKET(sk, "TX", NULL, &hh.bsdh);
if (sdp_apm_enable) {
rc = rdma_enable_apm(id, RDMA_ALT_PATH_BEST);
if (rc)
sdp_warn(sk, "APM couldn't be enabled for active side: %d\n", rc);
}
rc = rdma_connect(id, &conn_param);
break;
case RDMA_CM_EVENT_ALT_ROUTE_RESOLVED:
sdp_dbg(sk, "alt route was resolved slid=%d, dlid=%d\n",
id->route.path_rec[1].slid, id->route.path_rec[1].dlid);
break;
case RDMA_CM_EVENT_ALT_PATH_LOADED:
sdp_dbg(sk, "alt route path loaded\n");
break;
case RDMA_CM_EVENT_ALT_ROUTE_ERROR:
sdp_warn(sk, "alt route resolve error\n");
break;
case RDMA_CM_EVENT_ROUTE_ERROR:
rc = -ETIMEDOUT;
break;
case RDMA_CM_EVENT_CONNECT_REQUEST:
rc = sdp_connect_handler(sk, id, event);
if (rc) {
sdp_dbg(sk, "Destroying qp\n");
rdma_reject(id, NULL, 0);
break;
}
child = id->context;
atomic_set(&sdp_sk(child)->remote_credits,
rx_ring_posted(sdp_sk(child)));
memset(&hah, 0, sizeof hah);
hah.bsdh.mid = SDP_MID_HELLO_ACK;
hah.bsdh.bufs = htons(rx_ring_posted(sdp_sk(child)));
hah.bsdh.len = htonl(sizeof(struct sdp_hah));
hah.majv_minv = SDP_MAJV_MINV;
hah.ext_max_adverts = 1; /* Doesn't seem to be mandated by spec,
but just in case */
hah.actrcvsz = htonl(sdp_sk(child)->recv_frags * PAGE_SIZE +
sizeof(struct sdp_bsdh));
memset(&conn_param, 0, sizeof conn_param);
conn_param.private_data_len = sizeof hah;
conn_param.private_data = &hah;
conn_param.responder_resources = 4 /* TODO */;
conn_param.initiator_depth = 4 /* TODO */;
conn_param.retry_count = sdp_retry_count;
SDP_DUMP_PACKET(sk, "TX", NULL, &hah.bsdh);
rc = rdma_accept(id, &conn_param);
if (rc) {
sdp_sk(child)->id = NULL;
id->qp = NULL;
id->context = NULL;
parent = sdp_sk(child)->parent; /* TODO: hold ? */
} else if (sdp_apm_enable) {
rc2 = rdma_enable_apm(id, RDMA_ALT_PATH_BEST);
if (rc2)
sdp_warn(sk, "APM couldn't be enabled for passive side: %d\n", rc2);
}
break;
case RDMA_CM_EVENT_CONNECT_RESPONSE:
rc = sdp_response_handler(sk, id, event);
if (rc) {
sdp_dbg(sk, "Destroying qp\n");
rdma_reject(id, NULL, 0);
} else {
rc = rdma_accept(id, NULL);
if (!rc && sdp_apm_enable) {
rc2 = rdma_enable_apm(id, RDMA_ALT_PATH_BEST);
if (rc2)
sdp_warn(sk, "APM couldn't be enabled for passive side:%d \n", rc2);
}
}
break;
case RDMA_CM_EVENT_CONNECT_ERROR:
rc = -ETIMEDOUT;
break;
case RDMA_CM_EVENT_UNREACHABLE:
rc = -ENETUNREACH;
break;
case RDMA_CM_EVENT_REJECTED:
rc = -ECONNREFUSED;
break;
case RDMA_CM_EVENT_ESTABLISHED:
sdp_inet_saddr(sk) = sdp_inet_rcv_saddr(sk) =
((struct sockaddr_in *)&id->route.addr.src_addr)->sin_addr.s_addr;
rc = sdp_connected_handler(sk);
break;
case RDMA_CM_EVENT_DISCONNECTED: /* This means DREQ/DREP received */
if (sk->sk_state == TCP_LAST_ACK) {
sdp_cancel_dreq_wait_timeout(sdp_sk(sk));
sdp_exch_state(sk, TCPF_LAST_ACK, TCP_TIME_WAIT);
sdp_dbg(sk, "%s: waiting for Infiniband tear down\n",
__func__);
}
sdp_sk(sk)->qp_active = 0;
rdma_disconnect(id);
if (sk->sk_state != TCP_TIME_WAIT) {
if (sk->sk_state == TCP_CLOSE_WAIT) {
sdp_dbg(sk, "IB teardown while in "
"TCP_CLOSE_WAIT taking reference to "
"let close() finish the work\n");
sock_hold(sk, SOCK_REF_CMA);
sdp_start_cma_timewait_timeout(sdp_sk(sk),
SDP_CMA_TIMEWAIT_TIMEOUT);
}
sdp_set_error(sk, -EPIPE);
rc = sdp_disconnected_handler(sk);
}
break;
case RDMA_CM_EVENT_TIMEWAIT_EXIT:
rc = sdp_disconnected_handler(sk);
break;
case RDMA_CM_EVENT_DEVICE_REMOVAL:
rc = -ENETRESET;
break;
case RDMA_CM_EVENT_ADDR_CHANGE:
sdp_dbg(sk, "Got Address change event\n");
rc = 0;
break;
default:
printk(KERN_ERR "SDP: Unexpected CMA event: %d\n",
event->event);
rc = -ECONNABORTED;
break;
}
sdp_dbg(sk, "event: %s handled\n", rdma_cm_event_str(event->event));
if (rc && sdp_sk(sk)->id == id) {
child = sk;
sdp_sk(sk)->id = NULL;
id->qp = NULL;
id->context = NULL;
parent = sdp_sk(sk)->parent;
sdp_reset_sk(sk, rc);
}
release_sock(sk);
sdp_dbg(sk, "event: %s done. status %d\n",
rdma_cm_event_str(event->event), rc);
if (parent) {
lock_sock(parent);
if (!sdp_sk(parent)->id) { /* TODO: look at SOCK_DEAD? */
sdp_dbg(sk, "parent is going away.\n");
child = NULL;
goto done;
}
if (!list_empty(&sdp_sk(child)->backlog_queue))
list_del_init(&sdp_sk(child)->backlog_queue);
else
child = NULL;
done:
release_sock(parent);
if (child)
sdp_common_release(child);
}
return rc;
}
/**
* accept - wait for connection request
* @sock: listening socket
* @newsock: new socket that is to be connected
* @flags: file-related flags associated with socket
*
* Returns 0 on success, errno otherwise
*/
static int accept(struct socket *sock, struct socket *new_sock, int flags)
{
struct sock *new_sk, *sk = sock->sk;
struct sk_buff *buf;
struct tipc_sock *new_tsock;
struct tipc_port *new_tport;
struct tipc_msg *msg;
u32 new_ref;
int res;
lock_sock(sk);
if (sock->state != SS_LISTENING) {
res = -EINVAL;
goto exit;
}
while (skb_queue_empty(&sk->sk_receive_queue)) {
if (flags & O_NONBLOCK) {
res = -EWOULDBLOCK;
goto exit;
}
release_sock(sk);
res = wait_event_interruptible(*sk_sleep(sk),
(!skb_queue_empty(&sk->sk_receive_queue)));
lock_sock(sk);
if (res)
goto exit;
}
buf = skb_peek(&sk->sk_receive_queue);
res = tipc_create(sock_net(sock->sk), new_sock, 0, 0);
if (res)
goto exit;
new_sk = new_sock->sk;
new_tsock = tipc_sk(new_sk);
new_tport = new_tsock->p;
new_ref = new_tport->ref;
msg = buf_msg(buf);
/* we lock on new_sk; but lockdep sees the lock on sk */
lock_sock_nested(new_sk, SINGLE_DEPTH_NESTING);
/*
* Reject any stray messages received by new socket
* before the socket lock was taken (very, very unlikely)
*/
reject_rx_queue(new_sk);
/* Connect new socket to it's peer */
new_tsock->peer_name.ref = msg_origport(msg);
new_tsock->peer_name.node = msg_orignode(msg);
tipc_connect(new_ref, &new_tsock->peer_name);
new_sock->state = SS_CONNECTED;
tipc_set_portimportance(new_ref, msg_importance(msg));
if (msg_named(msg)) {
new_tport->conn_type = msg_nametype(msg);
new_tport->conn_instance = msg_nameinst(msg);
}
/*
* Respond to 'SYN-' by discarding it & returning 'ACK'-.
* Respond to 'SYN+' by queuing it on new socket.
*/
if (!msg_data_sz(msg)) {
struct msghdr m = {NULL,};
advance_rx_queue(sk);
send_packet(NULL, new_sock, &m, 0);
} else {
__skb_dequeue(&sk->sk_receive_queue);
__skb_queue_head(&new_sk->sk_receive_queue, buf);
skb_set_owner_r(buf, new_sk);
}
release_sock(new_sk);
exit:
release_sock(sk);
return res;
}
int virtio_transport_shutdown(struct vsock_sock *vsk, int mode)
{
struct virtio_vsock_pkt_info info = {
.op = VIRTIO_VSOCK_OP_SHUTDOWN,
.type = VIRTIO_VSOCK_TYPE_STREAM,
.flags = (mode & RCV_SHUTDOWN ?
VIRTIO_VSOCK_SHUTDOWN_RCV : 0) |
(mode & SEND_SHUTDOWN ?
VIRTIO_VSOCK_SHUTDOWN_SEND : 0),
.vsk = vsk,
};
return virtio_transport_send_pkt_info(vsk, &info);
}
EXPORT_SYMBOL_GPL(virtio_transport_shutdown);
int
virtio_transport_dgram_enqueue(struct vsock_sock *vsk,
struct sockaddr_vm *remote_addr,
struct msghdr *msg,
size_t dgram_len)
{
return -EOPNOTSUPP;
}
EXPORT_SYMBOL_GPL(virtio_transport_dgram_enqueue);
ssize_t
virtio_transport_stream_enqueue(struct vsock_sock *vsk,
struct msghdr *msg,
size_t len)
{
struct virtio_vsock_pkt_info info = {
.op = VIRTIO_VSOCK_OP_RW,
.type = VIRTIO_VSOCK_TYPE_STREAM,
.msg = msg,
.pkt_len = len,
.vsk = vsk,
};
return virtio_transport_send_pkt_info(vsk, &info);
}
EXPORT_SYMBOL_GPL(virtio_transport_stream_enqueue);
void virtio_transport_destruct(struct vsock_sock *vsk)
{
struct virtio_vsock_sock *vvs = vsk->trans;
kfree(vvs);
}
EXPORT_SYMBOL_GPL(virtio_transport_destruct);
static int virtio_transport_reset(struct vsock_sock *vsk,
struct virtio_vsock_pkt *pkt)
{
struct virtio_vsock_pkt_info info = {
.op = VIRTIO_VSOCK_OP_RST,
.type = VIRTIO_VSOCK_TYPE_STREAM,
.reply = !!pkt,
.vsk = vsk,
};
/* Send RST only if the original pkt is not a RST pkt */
if (pkt && le16_to_cpu(pkt->hdr.op) == VIRTIO_VSOCK_OP_RST)
return 0;
return virtio_transport_send_pkt_info(vsk, &info);
}
/* Normally packets are associated with a socket. There may be no socket if an
* attempt was made to connect to a socket that does not exist.
*/
static int virtio_transport_reset_no_sock(struct virtio_vsock_pkt *pkt)
{
struct virtio_vsock_pkt_info info = {
.op = VIRTIO_VSOCK_OP_RST,
.type = le16_to_cpu(pkt->hdr.type),
.reply = true,
};
/* Send RST only if the original pkt is not a RST pkt */
if (le16_to_cpu(pkt->hdr.op) == VIRTIO_VSOCK_OP_RST)
return 0;
pkt = virtio_transport_alloc_pkt(&info, 0,
le64_to_cpu(pkt->hdr.dst_cid),
le32_to_cpu(pkt->hdr.dst_port),
le64_to_cpu(pkt->hdr.src_cid),
le32_to_cpu(pkt->hdr.src_port));
if (!pkt)
return -ENOMEM;
return virtio_transport_get_ops()->send_pkt(pkt);
}
static void virtio_transport_wait_close(struct sock *sk, long timeout)
{
if (timeout) {
DEFINE_WAIT_FUNC(wait, woken_wake_function);
add_wait_queue(sk_sleep(sk), &wait);
do {
if (sk_wait_event(sk, &timeout,
sock_flag(sk, SOCK_DONE), &wait))
break;
} while (!signal_pending(current) && timeout);
remove_wait_queue(sk_sleep(sk), &wait);
}
}
static void virtio_transport_do_close(struct vsock_sock *vsk,
bool cancel_timeout)
{
struct sock *sk = sk_vsock(vsk);
sock_set_flag(sk, SOCK_DONE);
vsk->peer_shutdown = SHUTDOWN_MASK;
if (vsock_stream_has_data(vsk) <= 0)
sk->sk_state = SS_DISCONNECTING;
sk->sk_state_change(sk);
if (vsk->close_work_scheduled &&
(!cancel_timeout || cancel_delayed_work(&vsk->close_work))) {
vsk->close_work_scheduled = false;
vsock_remove_sock(vsk);
/* Release refcnt obtained when we scheduled the timeout */
sock_put(sk);
}
}
static void virtio_transport_close_timeout(struct work_struct *work)
{
struct vsock_sock *vsk =
container_of(work, struct vsock_sock, close_work.work);
struct sock *sk = sk_vsock(vsk);
sock_hold(sk);
lock_sock(sk);
if (!sock_flag(sk, SOCK_DONE)) {
(void)virtio_transport_reset(vsk, NULL);
virtio_transport_do_close(vsk, false);
}
vsk->close_work_scheduled = false;
release_sock(sk);
sock_put(sk);
}
/* User context, vsk->sk is locked */
static bool virtio_transport_close(struct vsock_sock *vsk)
{
struct sock *sk = &vsk->sk;
if (!(sk->sk_state == SS_CONNECTED ||
sk->sk_state == SS_DISCONNECTING))
return true;
/* Already received SHUTDOWN from peer, reply with RST */
if ((vsk->peer_shutdown & SHUTDOWN_MASK) == SHUTDOWN_MASK) {
(void)virtio_transport_reset(vsk, NULL);
return true;
}
if ((sk->sk_shutdown & SHUTDOWN_MASK) != SHUTDOWN_MASK)
(void)virtio_transport_shutdown(vsk, SHUTDOWN_MASK);
if (sock_flag(sk, SOCK_LINGER) && !(current->flags & PF_EXITING))
virtio_transport_wait_close(sk, sk->sk_lingertime);
if (sock_flag(sk, SOCK_DONE)) {
return true;
}
sock_hold(sk);
INIT_DELAYED_WORK(&vsk->close_work,
virtio_transport_close_timeout);
vsk->close_work_scheduled = true;
schedule_delayed_work(&vsk->close_work, VSOCK_CLOSE_TIMEOUT);
return false;
}
void virtio_transport_release(struct vsock_sock *vsk)
{
struct sock *sk = &vsk->sk;
bool remove_sock = true;
lock_sock(sk);
if (sk->sk_type == SOCK_STREAM)
remove_sock = virtio_transport_close(vsk);
release_sock(sk);
if (remove_sock)
vsock_remove_sock(vsk);
}
EXPORT_SYMBOL_GPL(virtio_transport_release);
static int
virtio_transport_recv_connecting(struct sock *sk,
struct virtio_vsock_pkt *pkt)
{
struct vsock_sock *vsk = vsock_sk(sk);
int err;
int skerr;
switch (le16_to_cpu(pkt->hdr.op)) {
case VIRTIO_VSOCK_OP_RESPONSE:
sk->sk_state = SS_CONNECTED;
sk->sk_socket->state = SS_CONNECTED;
vsock_insert_connected(vsk);
sk->sk_state_change(sk);
break;
case VIRTIO_VSOCK_OP_INVALID:
break;
case VIRTIO_VSOCK_OP_RST:
skerr = ECONNRESET;
err = 0;
goto destroy;
default:
skerr = EPROTO;
err = -EINVAL;
goto destroy;
}
return 0;
destroy:
virtio_transport_reset(vsk, pkt);
sk->sk_state = SS_UNCONNECTED;
sk->sk_err = skerr;
sk->sk_error_report(sk);
return err;
}
static int
virtio_transport_recv_connected(struct sock *sk,
struct virtio_vsock_pkt *pkt)
{
struct vsock_sock *vsk = vsock_sk(sk);
struct virtio_vsock_sock *vvs = vsk->trans;
int err = 0;
switch (le16_to_cpu(pkt->hdr.op)) {
case VIRTIO_VSOCK_OP_RW:
pkt->len = le32_to_cpu(pkt->hdr.len);
pkt->off = 0;
spin_lock_bh(&vvs->rx_lock);
virtio_transport_inc_rx_pkt(vvs, pkt);
list_add_tail(&pkt->list, &vvs->rx_queue);
spin_unlock_bh(&vvs->rx_lock);
sk->sk_data_ready(sk);
return err;
case VIRTIO_VSOCK_OP_CREDIT_UPDATE:
sk->sk_write_space(sk);
break;
case VIRTIO_VSOCK_OP_SHUTDOWN:
if (le32_to_cpu(pkt->hdr.flags) & VIRTIO_VSOCK_SHUTDOWN_RCV)
vsk->peer_shutdown |= RCV_SHUTDOWN;
if (le32_to_cpu(pkt->hdr.flags) & VIRTIO_VSOCK_SHUTDOWN_SEND)
vsk->peer_shutdown |= SEND_SHUTDOWN;
if (vsk->peer_shutdown == SHUTDOWN_MASK &&
vsock_stream_has_data(vsk) <= 0)
sk->sk_state = SS_DISCONNECTING;
if (le32_to_cpu(pkt->hdr.flags))
sk->sk_state_change(sk);
break;
case VIRTIO_VSOCK_OP_RST:
virtio_transport_do_close(vsk, true);
break;
default:
err = -EINVAL;
break;
}
virtio_transport_free_pkt(pkt);
return err;
}
static void
virtio_transport_recv_disconnecting(struct sock *sk,
struct virtio_vsock_pkt *pkt)
{
struct vsock_sock *vsk = vsock_sk(sk);
if (le16_to_cpu(pkt->hdr.op) == VIRTIO_VSOCK_OP_RST)
virtio_transport_do_close(vsk, true);
}
static int
virtio_transport_send_response(struct vsock_sock *vsk,
struct virtio_vsock_pkt *pkt)
{
struct virtio_vsock_pkt_info info = {
.op = VIRTIO_VSOCK_OP_RESPONSE,
.type = VIRTIO_VSOCK_TYPE_STREAM,
.remote_cid = le64_to_cpu(pkt->hdr.src_cid),
.remote_port = le32_to_cpu(pkt->hdr.src_port),
.reply = true,
.vsk = vsk,
};
return virtio_transport_send_pkt_info(vsk, &info);
}
/* Handle server socket */
static int
virtio_transport_recv_listen(struct sock *sk, struct virtio_vsock_pkt *pkt)
{
struct vsock_sock *vsk = vsock_sk(sk);
struct vsock_sock *vchild;
struct sock *child;
if (le16_to_cpu(pkt->hdr.op) != VIRTIO_VSOCK_OP_REQUEST) {
virtio_transport_reset(vsk, pkt);
return -EINVAL;
}
if (sk_acceptq_is_full(sk)) {
virtio_transport_reset(vsk, pkt);
return -ENOMEM;
}
child = __vsock_create(sock_net(sk), NULL, sk, GFP_KERNEL,
sk->sk_type, 0);
if (!child) {
virtio_transport_reset(vsk, pkt);
return -ENOMEM;
}
sk->sk_ack_backlog++;
lock_sock_nested(child, SINGLE_DEPTH_NESTING);
child->sk_state = SS_CONNECTED;
vchild = vsock_sk(child);
vsock_addr_init(&vchild->local_addr, le64_to_cpu(pkt->hdr.dst_cid),
le32_to_cpu(pkt->hdr.dst_port));
vsock_addr_init(&vchild->remote_addr, le64_to_cpu(pkt->hdr.src_cid),
le32_to_cpu(pkt->hdr.src_port));
vsock_insert_connected(vchild);
vsock_enqueue_accept(sk, child);
virtio_transport_send_response(vchild, pkt);
release_sock(child);
sk->sk_data_ready(sk);
return 0;
}
static bool virtio_transport_space_update(struct sock *sk,
struct virtio_vsock_pkt *pkt)
{
struct vsock_sock *vsk = vsock_sk(sk);
struct virtio_vsock_sock *vvs = vsk->trans;
bool space_available;
/* buf_alloc and fwd_cnt is always included in the hdr */
spin_lock_bh(&vvs->tx_lock);
vvs->peer_buf_alloc = le32_to_cpu(pkt->hdr.buf_alloc);
vvs->peer_fwd_cnt = le32_to_cpu(pkt->hdr.fwd_cnt);
space_available = virtio_transport_has_space(vsk);
spin_unlock_bh(&vvs->tx_lock);
return space_available;
}
/* We are under the virtio-vsock's vsock->rx_lock or vhost-vsock's vq->mutex
* lock.
*/
void virtio_transport_recv_pkt(struct virtio_vsock_pkt *pkt)
{
struct sockaddr_vm src, dst;
struct vsock_sock *vsk;
struct sock *sk;
bool space_available;
vsock_addr_init(&src, le64_to_cpu(pkt->hdr.src_cid),
le32_to_cpu(pkt->hdr.src_port));
vsock_addr_init(&dst, le64_to_cpu(pkt->hdr.dst_cid),
le32_to_cpu(pkt->hdr.dst_port));
trace_virtio_transport_recv_pkt(src.svm_cid, src.svm_port,
dst.svm_cid, dst.svm_port,
le32_to_cpu(pkt->hdr.len),
le16_to_cpu(pkt->hdr.type),
le16_to_cpu(pkt->hdr.op),
le32_to_cpu(pkt->hdr.flags),
le32_to_cpu(pkt->hdr.buf_alloc),
le32_to_cpu(pkt->hdr.fwd_cnt));
if (le16_to_cpu(pkt->hdr.type) != VIRTIO_VSOCK_TYPE_STREAM) {
(void)virtio_transport_reset_no_sock(pkt);
goto free_pkt;
}
/* The socket must be in connected or bound table
* otherwise send reset back
*/
sk = vsock_find_connected_socket(&src, &dst);
if (!sk) {
sk = vsock_find_bound_socket(&dst);
if (!sk) {
(void)virtio_transport_reset_no_sock(pkt);
goto free_pkt;
}
}
vsk = vsock_sk(sk);
space_available = virtio_transport_space_update(sk, pkt);
lock_sock(sk);
/* Update CID in case it has changed after a transport reset event */
vsk->local_addr.svm_cid = dst.svm_cid;
if (space_available)
sk->sk_write_space(sk);
switch (sk->sk_state) {
case VSOCK_SS_LISTEN:
virtio_transport_recv_listen(sk, pkt);
virtio_transport_free_pkt(pkt);
break;
case SS_CONNECTING:
virtio_transport_recv_connecting(sk, pkt);
virtio_transport_free_pkt(pkt);
break;
case SS_CONNECTED:
virtio_transport_recv_connected(sk, pkt);
break;
case SS_DISCONNECTING:
virtio_transport_recv_disconnecting(sk, pkt);
virtio_transport_free_pkt(pkt);
break;
default:
virtio_transport_free_pkt(pkt);
break;
}
release_sock(sk);
/* Release refcnt obtained when we fetched this socket out of the
* bound or connected list.
*/
sock_put(sk);
return;
free_pkt:
virtio_transport_free_pkt(pkt);
}
EXPORT_SYMBOL_GPL(virtio_transport_recv_pkt);
void virtio_transport_free_pkt(struct virtio_vsock_pkt *pkt)
{
kfree(pkt->buf);
kfree(pkt);
}
EXPORT_SYMBOL_GPL(virtio_transport_free_pkt);
MODULE_LICENSE("GPL v2");
MODULE_AUTHOR("Asias He");
MODULE_DESCRIPTION("common code for virtio vsock");
