static int drbd_do_auth(struct drbd_tconn *tconn);
static int drbd_disconnected(struct drbd_conf *mdev);
-static enum finish_epoch drbd_may_finish_epoch(struct drbd_conf *, struct drbd_epoch *, enum epoch_event);
+static enum finish_epoch drbd_may_finish_epoch(struct drbd_tconn *, struct drbd_epoch *, enum epoch_event);
static int e_end_block(struct drbd_work *, int);
atomic_t *a = is_net ? &mdev->pp_in_use_by_net : &mdev->pp_in_use;
int i;
+ if (page == NULL)
+ return;
+
if (drbd_pp_vacant > (DRBD_MAX_BIO_SIZE/PAGE_SIZE) * minor_count)
i = page_chain_free(page);
else {
unsigned int data_size, gfp_t gfp_mask) __must_hold(local)
{
struct drbd_peer_request *peer_req;
- struct page *page;
+ struct page *page = NULL;
unsigned nr_pages = (data_size + PAGE_SIZE -1) >> PAGE_SHIFT;
if (drbd_insert_fault(mdev, DRBD_FAULT_AL_EE))
return NULL;
}
- page = drbd_alloc_pages(mdev, nr_pages, (gfp_mask & __GFP_WAIT));
- if (!page)
- goto fail;
+ if (data_size) {
+ page = drbd_alloc_pages(mdev, nr_pages, (gfp_mask & __GFP_WAIT));
+ if (!page)
+ goto fail;
+ }
drbd_clear_interval(&peer_req->i);
peer_req->i.size = data_size;
drbd_free_net_peer_req(mdev, peer_req);
/* possible callbacks here:
- * e_end_block, and e_end_resync_block, e_send_discard_write.
+ * e_end_block, and e_end_resync_block, e_send_superseded.
* all ignore the last argument.
*/
list_for_each_entry_safe(peer_req, t, &work_list, w.list) {
spin_unlock_irq(&mdev->tconn->req_lock);
}
-/* see also kernel_accept; which is only present since 2.6.18.
- * also we want to log which part of it failed, exactly */
-static int drbd_accept(const char **what, struct socket *sock, struct socket **newsock)
-{
- struct sock *sk = sock->sk;
- int err = 0;
-
- *what = "listen";
- err = sock->ops->listen(sock, 5);
- if (err < 0)
- goto out;
-
- *what = "sock_create_lite";
- err = sock_create_lite(sk->sk_family, sk->sk_type, sk->sk_protocol,
- newsock);
- if (err < 0)
- goto out;
-
- *what = "accept";
- err = sock->ops->accept(sock, *newsock, 0);
- if (err < 0) {
- sock_release(*newsock);
- *newsock = NULL;
- goto out;
- }
- (*newsock)->ops = sock->ops;
-
-out:
- return err;
-}
-
static int drbd_recv_short(struct socket *sock, void *buf, size_t size, int flags)
{
mm_segment_t oldfs;
static int drbd_recv(struct drbd_tconn *tconn, void *buf, size_t size)
{
- mm_segment_t oldfs;
- struct kvec iov = {
- .iov_base = buf,
- .iov_len = size,
- };
- struct msghdr msg = {
- .msg_iovlen = 1,
- .msg_iov = (struct iovec *)&iov,
- .msg_flags = MSG_WAITALL | MSG_NOSIGNAL
- };
int rv;
- oldfs = get_fs();
- set_fs(KERNEL_DS);
+ rv = drbd_recv_short(tconn->data.socket, buf, size, 0);
- for (;;) {
- rv = sock_recvmsg(tconn->data.socket, &msg, size, msg.msg_flags);
- if (rv == size)
- break;
+ if (rv < 0) {
+ if (rv == -ECONNRESET)
+ conn_info(tconn, "sock was reset by peer\n");
+ else if (rv != -ERESTARTSYS)
+ conn_err(tconn, "sock_recvmsg returned %d\n", rv);
+ } else if (rv == 0) {
+ if (test_bit(DISCONNECT_SENT, &tconn->flags)) {
+ long t;
+ rcu_read_lock();
+ t = rcu_dereference(tconn->net_conf)->ping_timeo * HZ/10;
+ rcu_read_unlock();
- /* Note:
- * ECONNRESET other side closed the connection
- * ERESTARTSYS (on sock) we got a signal
- */
+ t = wait_event_timeout(tconn->ping_wait, tconn->cstate < C_WF_REPORT_PARAMS, t);
- if (rv < 0) {
- if (rv == -ECONNRESET)
- conn_info(tconn, "sock was reset by peer\n");
- else if (rv != -ERESTARTSYS)
- conn_err(tconn, "sock_recvmsg returned %d\n", rv);
- break;
- } else if (rv == 0) {
- conn_info(tconn, "sock was shut down by peer\n");
- break;
- } else {
- /* signal came in, or peer/link went down,
- * after we read a partial message
- */
- /* D_ASSERT(signal_pending(current)); */
- break;
+ if (t)
+ goto out;
}
- };
-
- set_fs(oldfs);
+ conn_info(tconn, "sock was shut down by peer\n");
+ }
if (rv != size)
conn_request_state(tconn, NS(conn, C_BROKEN_PIPE), CS_HARD);
+out:
return rv;
}
return sock;
}
-static struct socket *drbd_wait_for_connect(struct drbd_tconn *tconn)
+struct accept_wait_data {
+ struct drbd_tconn *tconn;
+ struct socket *s_listen;
+ struct completion door_bell;
+ void (*original_sk_state_change)(struct sock *sk);
+
+};
+
+static void drbd_incoming_connection(struct sock *sk)
{
- int timeo, err, my_addr_len;
- int sndbuf_size, rcvbuf_size, connect_int;
- struct socket *s_estab = NULL, *s_listen;
+ struct accept_wait_data *ad = sk->sk_user_data;
+ void (*state_change)(struct sock *sk);
+
+ state_change = ad->original_sk_state_change;
+ if (sk->sk_state == TCP_ESTABLISHED)
+ complete(&ad->door_bell);
+ state_change(sk);
+}
+
+static int prepare_listen_socket(struct drbd_tconn *tconn, struct accept_wait_data *ad)
+{
+ int err, sndbuf_size, rcvbuf_size, my_addr_len;
struct sockaddr_in6 my_addr;
+ struct socket *s_listen;
struct net_conf *nc;
const char *what;
nc = rcu_dereference(tconn->net_conf);
if (!nc) {
rcu_read_unlock();
- return NULL;
+ return -EIO;
}
sndbuf_size = nc->sndbuf_size;
rcvbuf_size = nc->rcvbuf_size;
- connect_int = nc->connect_int;
rcu_read_unlock();
my_addr_len = min_t(int, tconn->my_addr_len, sizeof(struct sockaddr_in6));
what = "sock_create_kern";
err = sock_create_kern(((struct sockaddr *)&my_addr)->sa_family,
- SOCK_STREAM, IPPROTO_TCP, &s_listen);
+ SOCK_STREAM, IPPROTO_TCP, &s_listen);
if (err) {
s_listen = NULL;
goto out;
}
- timeo = connect_int * HZ;
- timeo += (random32() & 1) ? timeo / 7 : -timeo / 7; /* 28.5% random jitter */
-
- s_listen->sk->sk_reuse = 1; /* SO_REUSEADDR */
- s_listen->sk->sk_rcvtimeo = timeo;
- s_listen->sk->sk_sndtimeo = timeo;
+ s_listen->sk->sk_reuse = SK_CAN_REUSE; /* SO_REUSEADDR */
drbd_setbufsize(s_listen, sndbuf_size, rcvbuf_size);
what = "bind before listen";
if (err < 0)
goto out;
- err = drbd_accept(&what, s_listen, &s_estab);
+ ad->s_listen = s_listen;
+ write_lock_bh(&s_listen->sk->sk_callback_lock);
+ ad->original_sk_state_change = s_listen->sk->sk_state_change;
+ s_listen->sk->sk_state_change = drbd_incoming_connection;
+ s_listen->sk->sk_user_data = ad;
+ write_unlock_bh(&s_listen->sk->sk_callback_lock);
+
+ what = "listen";
+ err = s_listen->ops->listen(s_listen, 5);
+ if (err < 0)
+ goto out;
+ return 0;
out:
if (s_listen)
sock_release(s_listen);
}
}
+ return -EIO;
+}
+
+static void unregister_state_change(struct sock *sk, struct accept_wait_data *ad)
+{
+ write_lock_bh(&sk->sk_callback_lock);
+ sk->sk_state_change = ad->original_sk_state_change;
+ sk->sk_user_data = NULL;
+ write_unlock_bh(&sk->sk_callback_lock);
+}
+
+static struct socket *drbd_wait_for_connect(struct drbd_tconn *tconn, struct accept_wait_data *ad)
+{
+ int timeo, connect_int, err = 0;
+ struct socket *s_estab = NULL;
+ struct net_conf *nc;
+
+ rcu_read_lock();
+ nc = rcu_dereference(tconn->net_conf);
+ if (!nc) {
+ rcu_read_unlock();
+ return NULL;
+ }
+ connect_int = nc->connect_int;
+ rcu_read_unlock();
+
+ timeo = connect_int * HZ;
+ timeo += (random32() & 1) ? timeo / 7 : -timeo / 7; /* 28.5% random jitter */
+
+ err = wait_for_completion_interruptible_timeout(&ad->door_bell, timeo);
+ if (err <= 0)
+ return NULL;
+
+ err = kernel_accept(ad->s_listen, &s_estab, 0);
+ if (err < 0) {
+ if (err != -EAGAIN && err != -EINTR && err != -ERESTARTSYS) {
+ conn_err(tconn, "accept failed, err = %d\n", err);
+ conn_request_state(tconn, NS(conn, C_DISCONNECTING), CS_HARD);
+ }
+ }
+
+ if (s_estab)
+ unregister_state_change(s_estab->sk, ad);
+
return s_estab;
}
if (!err)
err = drbd_send_uuids(mdev);
if (!err)
- err = drbd_send_state(mdev);
+ err = drbd_send_current_state(mdev);
clear_bit(USE_DEGR_WFC_T, &mdev->flags);
clear_bit(RESIZE_PENDING, &mdev->flags);
mod_timer(&mdev->request_timer, jiffies + HZ); /* just start it here. */
*/
static int conn_connect(struct drbd_tconn *tconn)
{
- struct socket *sock, *msock;
+ struct drbd_socket sock, msock;
struct drbd_conf *mdev;
struct net_conf *nc;
- int vnr, timeout, try, h, ok;
+ int vnr, timeout, h, ok;
+ bool discard_my_data;
+ enum drbd_state_rv rv;
+ struct accept_wait_data ad = {
+ .tconn = tconn,
+ .door_bell = COMPLETION_INITIALIZER_ONSTACK(ad.door_bell),
+ };
+ clear_bit(DISCONNECT_SENT, &tconn->flags);
if (conn_request_state(tconn, NS(conn, C_WF_CONNECTION), CS_VERBOSE) < SS_SUCCESS)
return -2;
- clear_bit(DISCARD_CONCURRENT, &tconn->flags);
+ mutex_init(&sock.mutex);
+ sock.sbuf = tconn->data.sbuf;
+ sock.rbuf = tconn->data.rbuf;
+ sock.socket = NULL;
+ mutex_init(&msock.mutex);
+ msock.sbuf = tconn->meta.sbuf;
+ msock.rbuf = tconn->meta.rbuf;
+ msock.socket = NULL;
/* Assume that the peer only understands protocol 80 until we know better. */
tconn->agreed_pro_version = 80;
+ if (prepare_listen_socket(tconn, &ad))
+ return 0;
+
do {
struct socket *s;
- for (try = 0;;) {
- /* 3 tries, this should take less than a second! */
- s = drbd_try_connect(tconn);
- if (s || ++try >= 3)
- break;
- /* give the other side time to call bind() & listen() */
- schedule_timeout_interruptible(HZ / 10);
- }
-
+ s = drbd_try_connect(tconn);
if (s) {
- if (!tconn->data.socket) {
- tconn->data.socket = s;
- send_first_packet(tconn, &tconn->data, P_INITIAL_DATA);
- } else if (!tconn->meta.socket) {
- tconn->meta.socket = s;
- send_first_packet(tconn, &tconn->meta, P_INITIAL_META);
+ if (!sock.socket) {
+ sock.socket = s;
+ send_first_packet(tconn, &sock, P_INITIAL_DATA);
+ } else if (!msock.socket) {
+ clear_bit(RESOLVE_CONFLICTS, &tconn->flags);
+ msock.socket = s;
+ send_first_packet(tconn, &msock, P_INITIAL_META);
} else {
conn_err(tconn, "Logic error in conn_connect()\n");
goto out_release_sockets;
}
}
- if (tconn->data.socket && tconn->meta.socket) {
- schedule_timeout_interruptible(tconn->net_conf->ping_timeo*HZ/10);
- ok = drbd_socket_okay(&tconn->data.socket);
- ok = drbd_socket_okay(&tconn->meta.socket) && ok;
+ if (sock.socket && msock.socket) {
+ rcu_read_lock();
+ nc = rcu_dereference(tconn->net_conf);
+ timeout = nc->ping_timeo * HZ / 10;
+ rcu_read_unlock();
+ schedule_timeout_interruptible(timeout);
+ ok = drbd_socket_okay(&sock.socket);
+ ok = drbd_socket_okay(&msock.socket) && ok;
if (ok)
break;
}
retry:
- s = drbd_wait_for_connect(tconn);
+ s = drbd_wait_for_connect(tconn, &ad);
if (s) {
- try = receive_first_packet(tconn, s);
- drbd_socket_okay(&tconn->data.socket);
- drbd_socket_okay(&tconn->meta.socket);
- switch (try) {
+ int fp = receive_first_packet(tconn, s);
+ drbd_socket_okay(&sock.socket);
+ drbd_socket_okay(&msock.socket);
+ switch (fp) {
case P_INITIAL_DATA:
- if (tconn->data.socket) {
+ if (sock.socket) {
conn_warn(tconn, "initial packet S crossed\n");
- sock_release(tconn->data.socket);
+ sock_release(sock.socket);
+ sock.socket = s;
+ goto randomize;
}
- tconn->data.socket = s;
+ sock.socket = s;
break;
case P_INITIAL_META:
- if (tconn->meta.socket) {
+ set_bit(RESOLVE_CONFLICTS, &tconn->flags);
+ if (msock.socket) {
conn_warn(tconn, "initial packet M crossed\n");
- sock_release(tconn->meta.socket);
+ sock_release(msock.socket);
+ msock.socket = s;
+ goto randomize;
}
- tconn->meta.socket = s;
- set_bit(DISCARD_CONCURRENT, &tconn->flags);
+ msock.socket = s;
break;
default:
conn_warn(tconn, "Error receiving initial packet\n");
sock_release(s);
+randomize:
if (random32() & 1)
goto retry;
}
goto out_release_sockets;
}
- if (tconn->data.socket && &tconn->meta.socket) {
- ok = drbd_socket_okay(&tconn->data.socket);
- ok = drbd_socket_okay(&tconn->meta.socket) && ok;
- if (ok)
- break;
- }
- } while (1);
+ ok = drbd_socket_okay(&sock.socket);
+ ok = drbd_socket_okay(&msock.socket) && ok;
+ } while (!ok);
- sock = tconn->data.socket;
- msock = tconn->meta.socket;
+ if (ad.s_listen)
+ sock_release(ad.s_listen);
- msock->sk->sk_reuse = 1; /* SO_REUSEADDR */
- sock->sk->sk_reuse = 1; /* SO_REUSEADDR */
+ sock.socket->sk->sk_reuse = SK_CAN_REUSE; /* SO_REUSEADDR */
+ msock.socket->sk->sk_reuse = SK_CAN_REUSE; /* SO_REUSEADDR */
- sock->sk->sk_allocation = GFP_NOIO;
- msock->sk->sk_allocation = GFP_NOIO;
+ sock.socket->sk->sk_allocation = GFP_NOIO;
+ msock.socket->sk->sk_allocation = GFP_NOIO;
- sock->sk->sk_priority = TC_PRIO_INTERACTIVE_BULK;
- msock->sk->sk_priority = TC_PRIO_INTERACTIVE;
+ sock.socket->sk->sk_priority = TC_PRIO_INTERACTIVE_BULK;
+ msock.socket->sk->sk_priority = TC_PRIO_INTERACTIVE;
/* NOT YET ...
- * sock->sk->sk_sndtimeo = tconn->net_conf->timeout*HZ/10;
- * sock->sk->sk_rcvtimeo = MAX_SCHEDULE_TIMEOUT;
+ * sock.socket->sk->sk_sndtimeo = tconn->net_conf->timeout*HZ/10;
+ * sock.socket->sk->sk_rcvtimeo = MAX_SCHEDULE_TIMEOUT;
* first set it to the P_CONNECTION_FEATURES timeout,
* which we set to 4x the configured ping_timeout. */
rcu_read_lock();
nc = rcu_dereference(tconn->net_conf);
- sock->sk->sk_sndtimeo =
- sock->sk->sk_rcvtimeo = nc->ping_timeo*4*HZ/10;
+ sock.socket->sk->sk_sndtimeo =
+ sock.socket->sk->sk_rcvtimeo = nc->ping_timeo*4*HZ/10;
- msock->sk->sk_rcvtimeo = nc->ping_int*HZ;
+ msock.socket->sk->sk_rcvtimeo = nc->ping_int*HZ;
timeout = nc->timeout * HZ / 10;
+ discard_my_data = nc->discard_my_data;
rcu_read_unlock();
- msock->sk->sk_sndtimeo = timeout;
+ msock.socket->sk->sk_sndtimeo = timeout;
/* we don't want delays.
* we use TCP_CORK where appropriate, though */
- drbd_tcp_nodelay(sock);
- drbd_tcp_nodelay(msock);
+ drbd_tcp_nodelay(sock.socket);
+ drbd_tcp_nodelay(msock.socket);
+ tconn->data.socket = sock.socket;
+ tconn->meta.socket = msock.socket;
tconn->last_received = jiffies;
h = drbd_do_features(tconn);
}
}
- if (conn_request_state(tconn, NS(conn, C_WF_REPORT_PARAMS), CS_VERBOSE) < SS_SUCCESS)
- return 0;
-
- sock->sk->sk_sndtimeo = timeout;
- sock->sk->sk_rcvtimeo = MAX_SCHEDULE_TIMEOUT;
-
- drbd_thread_start(&tconn->asender);
+ tconn->data.socket->sk->sk_sndtimeo = timeout;
+ tconn->data.socket->sk->sk_rcvtimeo = MAX_SCHEDULE_TIMEOUT;
if (drbd_send_protocol(tconn) == -EOPNOTSUPP)
return -1;
+ set_bit(STATE_SENT, &tconn->flags);
+
rcu_read_lock();
idr_for_each_entry(&tconn->volumes, mdev, vnr) {
kref_get(&mdev->kref);
rcu_read_unlock();
+
+ if (discard_my_data)
+ set_bit(DISCARD_MY_DATA, &mdev->flags);
+ else
+ clear_bit(DISCARD_MY_DATA, &mdev->flags);
+
drbd_connected(mdev);
kref_put(&mdev->kref, &drbd_minor_destroy);
rcu_read_lock();
}
rcu_read_unlock();
+ rv = conn_request_state(tconn, NS(conn, C_WF_REPORT_PARAMS), CS_VERBOSE);
+ if (rv < SS_SUCCESS) {
+ clear_bit(STATE_SENT, &tconn->flags);
+ return 0;
+ }
+
+ drbd_thread_start(&tconn->asender);
+
+ mutex_lock(&tconn->conf_update);
+ /* The discard_my_data flag is a single-shot modifier to the next
+ * connection attempt, the handshake of which is now well underway.
+ * No need for rcu style copying of the whole struct
+ * just to clear a single value. */
+ tconn->net_conf->discard_my_data = 0;
+ mutex_unlock(&tconn->conf_update);
+
return h;
out_release_sockets:
- if (tconn->data.socket) {
- sock_release(tconn->data.socket);
- tconn->data.socket = NULL;
- }
- if (tconn->meta.socket) {
- sock_release(tconn->meta.socket);
- tconn->meta.socket = NULL;
- }
+ if (ad.s_listen)
+ sock_release(ad.s_listen);
+ if (sock.socket)
+ sock_release(sock.socket);
+ if (msock.socket)
+ sock_release(msock.socket);
return -1;
}
return err;
}
-static void drbd_flush(struct drbd_conf *mdev)
+static void drbd_flush(struct drbd_tconn *tconn)
{
int rv;
+ struct drbd_conf *mdev;
+ int vnr;
+
+ if (tconn->write_ordering >= WO_bdev_flush) {
+ rcu_read_lock();
+ idr_for_each_entry(&tconn->volumes, mdev, vnr) {
+ if (!get_ldev(mdev))
+ continue;
+ kref_get(&mdev->kref);
+ rcu_read_unlock();
+
+ rv = blkdev_issue_flush(mdev->ldev->backing_bdev,
+ GFP_NOIO, NULL);
+ if (rv) {
+ dev_info(DEV, "local disk flush failed with status %d\n", rv);
+ /* would rather check on EOPNOTSUPP, but that is not reliable.
+ * don't try again for ANY return value != 0
+ * if (rv == -EOPNOTSUPP) */
+ drbd_bump_write_ordering(tconn, WO_drain_io);
+ }
+ put_ldev(mdev);
+ kref_put(&mdev->kref, &drbd_minor_destroy);
- if (mdev->write_ordering >= WO_bdev_flush && get_ldev(mdev)) {
- rv = blkdev_issue_flush(mdev->ldev->backing_bdev, GFP_KERNEL,
- NULL);
- if (rv) {
- dev_info(DEV, "local disk flush failed with status %d\n", rv);
- /* would rather check on EOPNOTSUPP, but that is not reliable.
- * don't try again for ANY return value != 0
- * if (rv == -EOPNOTSUPP) */
- drbd_bump_write_ordering(mdev, WO_drain_io);
+ rcu_read_lock();
+ if (rv)
+ break;
}
- put_ldev(mdev);
+ rcu_read_unlock();
}
}
* @epoch: Epoch object.
* @ev: Epoch event.
*/
-static enum finish_epoch drbd_may_finish_epoch(struct drbd_conf *mdev,
+static enum finish_epoch drbd_may_finish_epoch(struct drbd_tconn *tconn,
struct drbd_epoch *epoch,
enum epoch_event ev)
{
struct drbd_epoch *next_epoch;
enum finish_epoch rv = FE_STILL_LIVE;
- spin_lock(&mdev->epoch_lock);
+ spin_lock(&tconn->epoch_lock);
do {
next_epoch = NULL;
if (epoch_size != 0 &&
atomic_read(&epoch->active) == 0 &&
- test_bit(DE_HAVE_BARRIER_NUMBER, &epoch->flags)) {
+ (test_bit(DE_HAVE_BARRIER_NUMBER, &epoch->flags) || ev & EV_CLEANUP)) {
if (!(ev & EV_CLEANUP)) {
- spin_unlock(&mdev->epoch_lock);
- drbd_send_b_ack(mdev, epoch->barrier_nr, epoch_size);
- spin_lock(&mdev->epoch_lock);
+ spin_unlock(&tconn->epoch_lock);
+ drbd_send_b_ack(epoch->tconn, epoch->barrier_nr, epoch_size);
+ spin_lock(&tconn->epoch_lock);
}
- dec_unacked(mdev);
+#if 0
+ /* FIXME: dec unacked on connection, once we have
+ * something to count pending connection packets in. */
+ if (test_bit(DE_HAVE_BARRIER_NUMBER, &epoch->flags))
+ dec_unacked(epoch->tconn);
+#endif
- if (mdev->current_epoch != epoch) {
+ if (tconn->current_epoch != epoch) {
next_epoch = list_entry(epoch->list.next, struct drbd_epoch, list);
list_del(&epoch->list);
ev = EV_BECAME_LAST | (ev & EV_CLEANUP);
- mdev->epochs--;
+ tconn->epochs--;
kfree(epoch);
if (rv == FE_STILL_LIVE)
/* atomic_set(&epoch->active, 0); is already zero */
if (rv == FE_STILL_LIVE)
rv = FE_RECYCLED;
- wake_up(&mdev->ee_wait);
}
}
epoch = next_epoch;
} while (1);
- spin_unlock(&mdev->epoch_lock);
+ spin_unlock(&tconn->epoch_lock);
return rv;
}
/**
* drbd_bump_write_ordering() - Fall back to an other write ordering method
- * @mdev: DRBD device.
+ * @tconn: DRBD connection.
* @wo: Write ordering method to try.
*/
-void drbd_bump_write_ordering(struct drbd_conf *mdev, enum write_ordering_e wo) __must_hold(local)
+void drbd_bump_write_ordering(struct drbd_tconn *tconn, enum write_ordering_e wo)
{
struct disk_conf *dc;
+ struct drbd_conf *mdev;
enum write_ordering_e pwo;
+ int vnr;
static char *write_ordering_str[] = {
[WO_none] = "none",
[WO_drain_io] = "drain",
[WO_bdev_flush] = "flush",
};
- pwo = mdev->write_ordering;
+ pwo = tconn->write_ordering;
wo = min(pwo, wo);
rcu_read_lock();
- dc = rcu_dereference(mdev->ldev->disk_conf);
+ idr_for_each_entry(&tconn->volumes, mdev, vnr) {
+ if (!get_ldev_if_state(mdev, D_ATTACHING))
+ continue;
+ dc = rcu_dereference(mdev->ldev->disk_conf);
- if (wo == WO_bdev_flush && !dc->disk_flushes)
- wo = WO_drain_io;
- if (wo == WO_drain_io && !dc->disk_drain)
- wo = WO_none;
+ if (wo == WO_bdev_flush && !dc->disk_flushes)
+ wo = WO_drain_io;
+ if (wo == WO_drain_io && !dc->disk_drain)
+ wo = WO_none;
+ put_ldev(mdev);
+ }
rcu_read_unlock();
- mdev->write_ordering = wo;
- if (pwo != mdev->write_ordering || wo == WO_bdev_flush)
- dev_info(DEV, "Method to ensure write ordering: %s\n", write_ordering_str[mdev->write_ordering]);
+ tconn->write_ordering = wo;
+ if (pwo != tconn->write_ordering || wo == WO_bdev_flush)
+ conn_info(tconn, "Method to ensure write ordering: %s\n", write_ordering_str[tconn->write_ordering]);
}
/**
wake_up(&mdev->misc_wait);
}
-static int receive_Barrier(struct drbd_tconn *tconn, struct packet_info *pi)
+void conn_wait_active_ee_empty(struct drbd_tconn *tconn)
{
struct drbd_conf *mdev;
+ int vnr;
+
+ rcu_read_lock();
+ idr_for_each_entry(&tconn->volumes, mdev, vnr) {
+ kref_get(&mdev->kref);
+ rcu_read_unlock();
+ drbd_wait_ee_list_empty(mdev, &mdev->active_ee);
+ kref_put(&mdev->kref, &drbd_minor_destroy);
+ rcu_read_lock();
+ }
+ rcu_read_unlock();
+}
+
+static int receive_Barrier(struct drbd_tconn *tconn, struct packet_info *pi)
+{
int rv;
struct p_barrier *p = pi->data;
struct drbd_epoch *epoch;
- mdev = vnr_to_mdev(tconn, pi->vnr);
- if (!mdev)
- return -EIO;
-
- inc_unacked(mdev);
-
- mdev->current_epoch->barrier_nr = p->barrier;
- rv = drbd_may_finish_epoch(mdev, mdev->current_epoch, EV_GOT_BARRIER_NR);
+ /* FIXME these are unacked on connection,
+ * not a specific (peer)device.
+ */
+ tconn->current_epoch->barrier_nr = p->barrier;
+ tconn->current_epoch->tconn = tconn;
+ rv = drbd_may_finish_epoch(tconn, tconn->current_epoch, EV_GOT_BARRIER_NR);
/* P_BARRIER_ACK may imply that the corresponding extent is dropped from
* the activity log, which means it would not be resynced in case the
* R_PRIMARY crashes now.
* Therefore we must send the barrier_ack after the barrier request was
* completed. */
- switch (mdev->write_ordering) {
+ switch (tconn->write_ordering) {
case WO_none:
if (rv == FE_RECYCLED)
return 0;
if (epoch)
break;
else
- dev_warn(DEV, "Allocation of an epoch failed, slowing down\n");
+ conn_warn(tconn, "Allocation of an epoch failed, slowing down\n");
/* Fall through */
case WO_bdev_flush:
case WO_drain_io:
- drbd_wait_ee_list_empty(mdev, &mdev->active_ee);
- drbd_flush(mdev);
+ conn_wait_active_ee_empty(tconn);
+ drbd_flush(tconn);
- if (atomic_read(&mdev->current_epoch->epoch_size)) {
+ if (atomic_read(&tconn->current_epoch->epoch_size)) {
epoch = kmalloc(sizeof(struct drbd_epoch), GFP_NOIO);
if (epoch)
break;
}
- epoch = mdev->current_epoch;
- wait_event(mdev->ee_wait, atomic_read(&epoch->epoch_size) == 0);
-
- D_ASSERT(atomic_read(&epoch->active) == 0);
- D_ASSERT(epoch->flags == 0);
-
return 0;
default:
- dev_err(DEV, "Strangeness in mdev->write_ordering %d\n", mdev->write_ordering);
+ conn_err(tconn, "Strangeness in tconn->write_ordering %d\n", tconn->write_ordering);
return -EIO;
}
atomic_set(&epoch->epoch_size, 0);
atomic_set(&epoch->active, 0);
- spin_lock(&mdev->epoch_lock);
- if (atomic_read(&mdev->current_epoch->epoch_size)) {
- list_add(&epoch->list, &mdev->current_epoch->list);
- mdev->current_epoch = epoch;
- mdev->epochs++;
+ spin_lock(&tconn->epoch_lock);
+ if (atomic_read(&tconn->current_epoch->epoch_size)) {
+ list_add(&epoch->list, &tconn->current_epoch->list);
+ tconn->current_epoch = epoch;
+ tconn->epochs++;
} else {
/* The current_epoch got recycled while we allocated this one... */
kfree(epoch);
}
- spin_unlock(&mdev->epoch_lock);
+ spin_unlock(&tconn->epoch_lock);
return 0;
}
data_size -= dgs;
}
- if (!expect(data_size != 0))
- return NULL;
if (!expect(IS_ALIGNED(data_size, 512)))
return NULL;
if (!expect(data_size <= DRBD_MAX_BIO_SIZE))
if (!peer_req)
return NULL;
+ if (!data_size)
+ return peer_req;
+
ds = data_size;
page = peer_req->pages;
page_chain_for_each(page) {
return err;
}
-static int w_restart_write(struct drbd_work *w, int cancel)
-{
- struct drbd_request *req = container_of(w, struct drbd_request, w);
- struct drbd_conf *mdev = w->mdev;
- struct bio *bio;
- unsigned long start_time;
- unsigned long flags;
-
- spin_lock_irqsave(&mdev->tconn->req_lock, flags);
- if (!expect(req->rq_state & RQ_POSTPONED)) {
- spin_unlock_irqrestore(&mdev->tconn->req_lock, flags);
- return -EIO;
- }
- bio = req->master_bio;
- start_time = req->start_time;
- /* Postponed requests will not have their master_bio completed! */
- __req_mod(req, DISCARD_WRITE, NULL);
- spin_unlock_irqrestore(&mdev->tconn->req_lock, flags);
-
- while (__drbd_make_request(mdev, bio, start_time))
- /* retry */ ;
- return 0;
-}
-
static void restart_conflicting_writes(struct drbd_conf *mdev,
sector_t sector, int size)
{
if (req->rq_state & RQ_LOCAL_PENDING ||
!(req->rq_state & RQ_POSTPONED))
continue;
- if (expect(list_empty(&req->w.list))) {
- req->w.mdev = mdev;
- req->w.cb = w_restart_write;
- drbd_queue_work(&mdev->tconn->data.work, &req->w);
- }
+ /* as it is RQ_POSTPONED, this will cause it to
+ * be queued on the retry workqueue. */
+ __req_mod(req, CONFLICT_RESOLVED, NULL);
}
}
} else
D_ASSERT(drbd_interval_empty(&peer_req->i));
- drbd_may_finish_epoch(mdev, peer_req->epoch, EV_PUT + (cancel ? EV_CLEANUP : 0));
+ drbd_may_finish_epoch(mdev->tconn, peer_req->epoch, EV_PUT + (cancel ? EV_CLEANUP : 0));
return err;
}
return err;
}
-static int e_send_discard_write(struct drbd_work *w, int unused)
+static int e_send_superseded(struct drbd_work *w, int unused)
{
- return e_send_ack(w, P_DISCARD_WRITE);
+ return e_send_ack(w, P_SUPERSEDED);
}
static int e_send_retry_write(struct drbd_work *w, int unused)
struct drbd_tconn *tconn = w->mdev->tconn;
return e_send_ack(w, tconn->agreed_pro_version >= 100 ?
- P_RETRY_WRITE : P_DISCARD_WRITE);
+ P_RETRY_WRITE : P_SUPERSEDED);
}
static bool seq_greater(u32 a, u32 b)
/*
* We only need to keep track of the last packet_seq number of our peer
- * if we are in dual-primary mode and we have the discard flag set; see
+ * if we are in dual-primary mode and we have the resolve-conflicts flag set; see
* handle_write_conflicts().
*/
tp = rcu_dereference(mdev->tconn->net_conf)->two_primaries;
rcu_read_unlock();
- return tp && test_bit(DISCARD_CONCURRENT, &tconn->flags);
+ return tp && test_bit(RESOLVE_CONFLICTS, &tconn->flags);
}
static void update_peer_seq(struct drbd_conf *mdev, unsigned int peer_seq)
}
}
+static inline int overlaps(sector_t s1, int l1, sector_t s2, int l2)
+{
+ return !((s1 + (l1>>9) <= s2) || (s1 >= s2 + (l2>>9)));
+}
+
+/* maybe change sync_ee into interval trees as well? */
+static bool overlapping_resync_write(struct drbd_conf *mdev, struct drbd_peer_request *peer_req)
+{
+ struct drbd_peer_request *rs_req;
+ bool rv = 0;
+
+ spin_lock_irq(&mdev->tconn->req_lock);
+ list_for_each_entry(rs_req, &mdev->sync_ee, w.list) {
+ if (overlaps(peer_req->i.sector, peer_req->i.size,
+ rs_req->i.sector, rs_req->i.size)) {
+ rv = 1;
+ break;
+ }
+ }
+ spin_unlock_irq(&mdev->tconn->req_lock);
+
+ return rv;
+}
+
/* Called from receive_Data.
* Synchronize packets on sock with packets on msock.
*
struct drbd_peer_request *peer_req)
{
struct drbd_tconn *tconn = mdev->tconn;
- bool resolve_conflicts = test_bit(DISCARD_CONCURRENT, &tconn->flags);
+ bool resolve_conflicts = test_bit(RESOLVE_CONFLICTS, &tconn->flags);
sector_t sector = peer_req->i.sector;
const unsigned int size = peer_req->i.size;
struct drbd_interval *i;
if (resolve_conflicts) {
/*
* If the peer request is fully contained within the
- * overlapping request, it can be discarded; otherwise,
- * it will be retried once all overlapping requests
- * have completed.
+ * overlapping request, it can be considered overwritten
+ * and thus superseded; otherwise, it will be retried
+ * once all overlapping requests have completed.
*/
- bool discard = i->sector <= sector && i->sector +
+ bool superseded = i->sector <= sector && i->sector +
(i->size >> 9) >= sector + (size >> 9);
if (!equal)
"assuming %s came first\n",
(unsigned long long)i->sector, i->size,
(unsigned long long)sector, size,
- discard ? "local" : "remote");
+ superseded ? "local" : "remote");
inc_unacked(mdev);
- peer_req->w.cb = discard ? e_send_discard_write :
+ peer_req->w.cb = superseded ? e_send_superseded :
e_send_retry_write;
list_add_tail(&peer_req->w.list, &mdev->done_ee);
wake_asender(mdev->tconn);
!(req->rq_state & RQ_POSTPONED)) {
/*
* Wait for the node with the discard flag to
- * decide if this request will be discarded or
- * retried. Requests that are discarded will
+ * decide if this request has been superseded
+ * or needs to be retried.
+ * Requests that have been superseded will
* disappear from the write_requests tree.
*
* In addition, wait for the conflicting
err = wait_for_and_update_peer_seq(mdev, peer_seq);
drbd_send_ack_dp(mdev, P_NEG_ACK, p, pi->size);
- atomic_inc(&mdev->current_epoch->epoch_size);
+ atomic_inc(&tconn->current_epoch->epoch_size);
err2 = drbd_drain_block(mdev, pi->size);
if (!err)
err = err2;
dp_flags = be32_to_cpu(p->dp_flags);
rw |= wire_flags_to_bio(mdev, dp_flags);
+ if (peer_req->pages == NULL) {
+ D_ASSERT(peer_req->i.size == 0);
+ D_ASSERT(dp_flags & DP_FLUSH);
+ }
if (dp_flags & DP_MAY_SET_IN_SYNC)
peer_req->flags |= EE_MAY_SET_IN_SYNC;
- spin_lock(&mdev->epoch_lock);
- peer_req->epoch = mdev->current_epoch;
+ spin_lock(&tconn->epoch_lock);
+ peer_req->epoch = tconn->current_epoch;
atomic_inc(&peer_req->epoch->epoch_size);
atomic_inc(&peer_req->epoch->active);
- spin_unlock(&mdev->epoch_lock);
+ spin_unlock(&tconn->epoch_lock);
rcu_read_lock();
tp = rcu_dereference(mdev->tconn->net_conf)->two_primaries;
list_add(&peer_req->w.list, &mdev->active_ee);
spin_unlock_irq(&mdev->tconn->req_lock);
+ if (mdev->state.conn == C_SYNC_TARGET)
+ wait_event(mdev->ee_wait, !overlapping_resync_write(mdev, peer_req));
+
if (mdev->tconn->agreed_pro_version < 100) {
rcu_read_lock();
switch (rcu_dereference(mdev->tconn->net_conf)->wire_protocol) {
drbd_al_complete_io(mdev, &peer_req->i);
out_interrupted:
- drbd_may_finish_epoch(mdev, peer_req->epoch, EV_PUT + EV_CLEANUP);
+ drbd_may_finish_epoch(tconn, peer_req->epoch, EV_PUT + EV_CLEANUP);
put_ldev(mdev);
drbd_free_peer_req(mdev, peer_req);
return err;
"Using discard-least-changes instead\n");
case ASB_DISCARD_ZERO_CHG:
if (ch_peer == 0 && ch_self == 0) {
- rv = test_bit(DISCARD_CONCURRENT, &mdev->tconn->flags)
+ rv = test_bit(RESOLVE_CONFLICTS, &mdev->tconn->flags)
? -1 : 1;
break;
} else {
rv = 1;
else /* ( ch_self == ch_peer ) */
/* Well, then use something else. */
- rv = test_bit(DISCARD_CONCURRENT, &mdev->tconn->flags)
+ rv = test_bit(RESOLVE_CONFLICTS, &mdev->tconn->flags)
? -1 : 1;
break;
case ASB_DISCARD_LOCAL:
if ((mdev->ldev->md.uuid[UI_BITMAP] & ~((u64)1)) == (mdev->p_uuid[UI_HISTORY_START] & ~((u64)1)) &&
(mdev->ldev->md.uuid[UI_HISTORY_START] & ~((u64)1)) == (mdev->p_uuid[UI_HISTORY_START + 1] & ~((u64)1))) {
dev_info(DEV, "was SyncSource, missed the resync finished event, corrected myself:\n");
- drbd_uuid_set_bm(mdev, 0UL);
+ drbd_uuid_move_history(mdev);
+ mdev->ldev->md.uuid[UI_HISTORY_START] = mdev->ldev->md.uuid[UI_BITMAP];
+ mdev->ldev->md.uuid[UI_BITMAP] = 0;
drbd_uuid_dump(mdev, "self", mdev->ldev->md.uuid,
mdev->state.disk >= D_NEGOTIATING ? drbd_bm_total_weight(mdev) : 0, 0);
case 1: /* self_pri && !peer_pri */ return 1;
case 2: /* !self_pri && peer_pri */ return -1;
case 3: /* self_pri && peer_pri */
- dc = test_bit(DISCARD_CONCURRENT, &mdev->tconn->flags);
+ dc = test_bit(RESOLVE_CONFLICTS, &mdev->tconn->flags);
return dc ? -1 : 1;
}
}
mdev->p_uuid[UI_BITMAP] = mdev->p_uuid[UI_HISTORY_START];
mdev->p_uuid[UI_HISTORY_START] = mdev->p_uuid[UI_HISTORY_START + 1];
- dev_info(DEV, "Did not got last syncUUID packet, corrected:\n");
+ dev_info(DEV, "Lost last syncUUID packet, corrected:\n");
drbd_uuid_dump(mdev, "peer", mdev->p_uuid, mdev->p_uuid[UI_SIZE], mdev->p_uuid[UI_FLAGS]);
return -1;
if (mdev->tconn->agreed_pro_version < 91)
return -1091;
- _drbd_uuid_set(mdev, UI_BITMAP, mdev->ldev->md.uuid[UI_HISTORY_START]);
- _drbd_uuid_set(mdev, UI_HISTORY_START, mdev->ldev->md.uuid[UI_HISTORY_START + 1]);
+ __drbd_uuid_set(mdev, UI_BITMAP, mdev->ldev->md.uuid[UI_HISTORY_START]);
+ __drbd_uuid_set(mdev, UI_HISTORY_START, mdev->ldev->md.uuid[UI_HISTORY_START + 1]);
dev_info(DEV, "Last syncUUID did not get through, corrected:\n");
drbd_uuid_dump(mdev, "self", mdev->ldev->md.uuid,
mydisk = mdev->new_state_tmp.disk;
dev_info(DEV, "drbd_sync_handshake:\n");
+
+ spin_lock_irq(&mdev->ldev->md.uuid_lock);
drbd_uuid_dump(mdev, "self", mdev->ldev->md.uuid, mdev->comm_bm_set, 0);
drbd_uuid_dump(mdev, "peer", mdev->p_uuid,
mdev->p_uuid[UI_SIZE], mdev->p_uuid[UI_FLAGS]);
hg = drbd_uuid_compare(mdev, &rule_nr);
+ spin_unlock_irq(&mdev->ldev->md.uuid_lock);
dev_info(DEV, "uuid_compare()=%d by rule %d\n", hg, rule_nr);
}
if (hg == -100) {
- if (nc->discard_my_data && !(mdev->p_uuid[UI_FLAGS]&1))
+ if (test_bit(DISCARD_MY_DATA, &mdev->flags) && !(mdev->p_uuid[UI_FLAGS]&1))
hg = -1;
- if (!nc->discard_my_data && (mdev->p_uuid[UI_FLAGS]&1))
+ if (!test_bit(DISCARD_MY_DATA, &mdev->flags) && (mdev->p_uuid[UI_FLAGS]&1))
hg = 1;
if (abs(hg) < 100)
int p_proto, p_discard_my_data, p_two_primaries, cf;
struct net_conf *nc, *old_net_conf, *new_net_conf = NULL;
char integrity_alg[SHARED_SECRET_MAX] = "";
- struct crypto_hash *peer_integrity_tfm = NULL, *integrity_tfm = NULL;
+ struct crypto_hash *peer_integrity_tfm = NULL;
void *int_dig_in = NULL, *int_dig_vv = NULL;
p_proto = be32_to_cpu(p->protocol);
integrity_alg[SHARED_SECRET_MAX - 1] = 0;
}
- if (pi->cmd == P_PROTOCOL_UPDATE) {
- if (integrity_alg[0]) {
- int hash_size;
-
- peer_integrity_tfm = crypto_alloc_hash(integrity_alg, 0, CRYPTO_ALG_ASYNC);
- integrity_tfm = crypto_alloc_hash(integrity_alg, 0, CRYPTO_ALG_ASYNC);
- if (!(peer_integrity_tfm && integrity_tfm)) {
- conn_err(tconn, "peer data-integrity-alg %s not supported\n",
- integrity_alg);
- goto disconnect;
- }
-
- hash_size = crypto_hash_digestsize(integrity_tfm);
- int_dig_in = kmalloc(hash_size, GFP_KERNEL);
- int_dig_vv = kmalloc(hash_size, GFP_KERNEL);
- if (!(int_dig_in && int_dig_vv)) {
- conn_err(tconn, "Allocation of buffers for data integrity checking failed\n");
- goto disconnect;
- }
- }
-
- new_net_conf = kmalloc(sizeof(struct net_conf), GFP_KERNEL);
- if (!new_net_conf) {
- conn_err(tconn, "Allocation of new net_conf failed\n");
- goto disconnect;
- }
-
- mutex_lock(&tconn->data.mutex);
- mutex_lock(&tconn->conf_update);
- old_net_conf = tconn->net_conf;
- *new_net_conf = *old_net_conf;
-
- new_net_conf->wire_protocol = p_proto;
- new_net_conf->after_sb_0p = convert_after_sb(p_after_sb_0p);
- new_net_conf->after_sb_1p = convert_after_sb(p_after_sb_1p);
- new_net_conf->after_sb_2p = convert_after_sb(p_after_sb_2p);
- new_net_conf->two_primaries = p_two_primaries;
- strcpy(new_net_conf->integrity_alg, integrity_alg);
- new_net_conf->integrity_alg_len = strlen(integrity_alg) + 1;
-
- crypto_free_hash(tconn->integrity_tfm);
- tconn->integrity_tfm = integrity_tfm;
-
- rcu_assign_pointer(tconn->net_conf, new_net_conf);
- mutex_unlock(&tconn->conf_update);
- mutex_unlock(&tconn->data.mutex);
-
- crypto_free_hash(tconn->peer_integrity_tfm);
- kfree(tconn->int_dig_in);
- kfree(tconn->int_dig_vv);
- tconn->peer_integrity_tfm = peer_integrity_tfm;
- tconn->int_dig_in = int_dig_in;
- tconn->int_dig_vv = int_dig_vv;
-
- if (strcmp(old_net_conf->integrity_alg, integrity_alg))
- conn_info(tconn, "peer data-integrity-alg: %s\n", integrity_alg);
-
- synchronize_rcu();
- kfree(old_net_conf);
- } else {
+ if (pi->cmd != P_PROTOCOL_UPDATE) {
clear_bit(CONN_DRY_RUN, &tconn->flags);
if (cf & CF_DRY_RUN)
rcu_read_unlock();
}
+
+ if (integrity_alg[0]) {
+ int hash_size;
+
+ /*
+ * We can only change the peer data integrity algorithm
+ * here. Changing our own data integrity algorithm
+ * requires that we send a P_PROTOCOL_UPDATE packet at
+ * the same time; otherwise, the peer has no way to
+ * tell between which packets the algorithm should
+ * change.
+ */
+
+ peer_integrity_tfm = crypto_alloc_hash(integrity_alg, 0, CRYPTO_ALG_ASYNC);
+ if (!peer_integrity_tfm) {
+ conn_err(tconn, "peer data-integrity-alg %s not supported\n",
+ integrity_alg);
+ goto disconnect;
+ }
+
+ hash_size = crypto_hash_digestsize(peer_integrity_tfm);
+ int_dig_in = kmalloc(hash_size, GFP_KERNEL);
+ int_dig_vv = kmalloc(hash_size, GFP_KERNEL);
+ if (!(int_dig_in && int_dig_vv)) {
+ conn_err(tconn, "Allocation of buffers for data integrity checking failed\n");
+ goto disconnect;
+ }
+ }
+
+ new_net_conf = kmalloc(sizeof(struct net_conf), GFP_KERNEL);
+ if (!new_net_conf) {
+ conn_err(tconn, "Allocation of new net_conf failed\n");
+ goto disconnect;
+ }
+
+ mutex_lock(&tconn->data.mutex);
+ mutex_lock(&tconn->conf_update);
+ old_net_conf = tconn->net_conf;
+ *new_net_conf = *old_net_conf;
+
+ new_net_conf->wire_protocol = p_proto;
+ new_net_conf->after_sb_0p = convert_after_sb(p_after_sb_0p);
+ new_net_conf->after_sb_1p = convert_after_sb(p_after_sb_1p);
+ new_net_conf->after_sb_2p = convert_after_sb(p_after_sb_2p);
+ new_net_conf->two_primaries = p_two_primaries;
+
+ rcu_assign_pointer(tconn->net_conf, new_net_conf);
+ mutex_unlock(&tconn->conf_update);
+ mutex_unlock(&tconn->data.mutex);
+
+ crypto_free_hash(tconn->peer_integrity_tfm);
+ kfree(tconn->int_dig_in);
+ kfree(tconn->int_dig_vv);
+ tconn->peer_integrity_tfm = peer_integrity_tfm;
+ tconn->int_dig_in = int_dig_in;
+ tconn->int_dig_vv = int_dig_vv;
+
+ if (strcmp(old_net_conf->integrity_alg, integrity_alg))
+ conn_info(tconn, "peer data-integrity-alg: %s\n",
+ integrity_alg[0] ? integrity_alg : "(none)");
+
+ synchronize_rcu();
+ kfree(old_net_conf);
return 0;
disconnect_rcu_unlock:
rcu_read_unlock();
disconnect:
crypto_free_hash(peer_integrity_tfm);
- crypto_free_hash(integrity_tfm);
kfree(int_dig_in);
kfree(int_dig_vv);
conn_request_state(tconn, NS(conn, C_DISCONNECTING), CS_HARD);
if (apv >= 88) {
if (apv == 88) {
- if (data_size > SHARED_SECRET_MAX) {
- dev_err(DEV, "verify-alg too long, "
- "peer wants %u, accepting only %u byte\n",
- data_size, SHARED_SECRET_MAX);
+ if (data_size > SHARED_SECRET_MAX || data_size == 0) {
+ dev_err(DEV, "verify-alg of wrong size, "
+ "peer wants %u, accepting only up to %u byte\n",
+ data_size, SHARED_SECRET_MAX);
err = -EIO;
goto reconnect;
}
mask.i = be32_to_cpu(p->mask);
val.i = be32_to_cpu(p->val);
- if (test_bit(DISCARD_CONCURRENT, &mdev->tconn->flags) &&
+ if (test_bit(RESOLVE_CONFLICTS, &mdev->tconn->flags) &&
mutex_is_locked(mdev->state_mutex)) {
drbd_send_sr_reply(mdev, SS_CONCURRENT_ST_CHG);
return 0;
mask.i = be32_to_cpu(p->mask);
val.i = be32_to_cpu(p->val);
- if (test_bit(DISCARD_CONCURRENT, &tconn->flags) &&
+ if (test_bit(RESOLVE_CONFLICTS, &tconn->flags) &&
mutex_is_locked(&tconn->cstate_mutex)) {
conn_send_sr_reply(tconn, SS_CONCURRENT_ST_CHG);
return 0;
os = ns = drbd_read_state(mdev);
spin_unlock_irq(&mdev->tconn->req_lock);
- /* peer says his disk is uptodate, while we think it is inconsistent,
- * and this happens while we think we have a sync going on. */
- if (os.pdsk == D_INCONSISTENT && real_peer_disk == D_UP_TO_DATE &&
+ /* If some other part of the code (asender thread, timeout)
+ * already decided to close the connection again,
+ * we must not "re-establish" it here. */
+ if (os.conn <= C_TEAR_DOWN)
+ return -ECONNRESET;
+
+ /* If this is the "end of sync" confirmation, usually the peer disk
+ * transitions from D_INCONSISTENT to D_UP_TO_DATE. For empty (0 bits
+ * set) resync started in PausedSyncT, or if the timing of pause-/
+ * unpause-sync events has been "just right", the peer disk may
+ * transition from D_CONSISTENT to D_UP_TO_DATE as well.
+ */
+ if ((os.pdsk == D_INCONSISTENT || os.pdsk == D_CONSISTENT) &&
+ real_peer_disk == D_UP_TO_DATE &&
os.conn > C_CONNECTED && os.disk == D_UP_TO_DATE) {
/* If we are (becoming) SyncSource, but peer is still in sync
* preparation, ignore its uptodate-ness to avoid flapping, it
}
}
+ /* explicit verify finished notification, stop sector reached. */
+ if (os.conn == C_VERIFY_T && os.disk == D_UP_TO_DATE &&
+ peer_state.conn == C_CONNECTED && real_peer_disk == D_UP_TO_DATE) {
+ ov_out_of_sync_print(mdev);
+ drbd_resync_finished(mdev);
+ return 0;
+ }
+
/* peer says his disk is inconsistent, while we think it is uptodate,
* and this happens while the peer still thinks we have a sync going on,
* but we think we are already done with the sync.
/* Nowadays only used when forcing a node into primary role and
setting its disk to UpToDate with that */
drbd_send_uuids(mdev);
- drbd_send_state(mdev);
+ drbd_send_current_state(mdev);
}
}
- mutex_lock(&mdev->tconn->conf_update);
- mdev->tconn->net_conf->discard_my_data = 0; /* without copy; single bit op is atomic */
- mutex_unlock(&mdev->tconn->conf_update);
+ clear_bit(DISCARD_MY_DATA, &mdev->flags);
drbd_md_sync(mdev); /* update connected indicator, la_size, ... */
barr.w.cb = w_prev_work_done;
barr.w.tconn = tconn;
init_completion(&barr.done);
- drbd_queue_work(&tconn->data.work, &barr.w);
+ drbd_queue_work(&tconn->sender_work, &barr.w);
wait_for_completion(&barr.done);
}
{
struct drbd_conf *mdev;
enum drbd_conns oc;
- int vnr, rv = SS_UNKNOWN_ERROR;
+ int vnr;
if (tconn->cstate == C_STANDALONE)
return;
+ /* We are about to start the cleanup after connection loss.
+ * Make sure drbd_make_request knows about that.
+ * Usually we should be in some network failure state already,
+ * but just in case we are not, we fix it up here.
+ */
+ conn_request_state(tconn, NS(conn, C_NETWORK_FAILURE), CS_HARD);
+
/* asender does not clean up anything. it must not interfere, either */
drbd_thread_stop(&tconn->asender);
drbd_free_sock(tconn);
}
rcu_read_unlock();
+ if (!list_empty(&tconn->current_epoch->list))
+ conn_err(tconn, "ASSERTION FAILED: tconn->current_epoch->list not empty\n");
+ /* ok, no more ee's on the fly, it is safe to reset the epoch_size */
+ atomic_set(&tconn->current_epoch->epoch_size, 0);
+ tconn->send.seen_any_write_yet = false;
+
conn_info(tconn, "Connection closed\n");
if (conn_highest_role(tconn) == R_PRIMARY && conn_highest_pdsk(tconn) >= D_UNKNOWN)
spin_lock_irq(&tconn->req_lock);
oc = tconn->cstate;
if (oc >= C_UNCONNECTED)
- rv = _conn_request_state(tconn, NS(conn, C_UNCONNECTED), CS_VERBOSE);
+ _conn_request_state(tconn, NS(conn, C_UNCONNECTED), CS_VERBOSE);
spin_unlock_irq(&tconn->req_lock);
static int drbd_disconnected(struct drbd_conf *mdev)
{
- enum drbd_fencing_p fp;
unsigned int i;
/* wait for current activity to cease. */
drbd_finish_peer_reqs(mdev);
+ /* This second workqueue flush is necessary, since drbd_finish_peer_reqs()
+ might have issued a work again. The one before drbd_finish_peer_reqs() is
+ necessary to reclain net_ee in drbd_finish_peer_reqs(). */
+ drbd_flush_workqueue(mdev);
+
+ /* need to do it again, drbd_finish_peer_reqs() may have populated it
+ * again via drbd_try_clear_on_disk_bm(). */
+ drbd_rs_cancel_all(mdev);
+
kfree(mdev->p_uuid);
mdev->p_uuid = NULL;
drbd_md_sync(mdev);
- fp = FP_DONT_CARE;
- if (get_ldev(mdev)) {
- rcu_read_lock();
- fp = rcu_dereference(mdev->ldev->disk_conf)->fencing;
- rcu_read_unlock();
- put_ldev(mdev);
- }
-
/* serialize with bitmap writeout triggered by the state change,
* if any. */
wait_event(mdev->misc_wait, !test_bit(BITMAP_IO, &mdev->flags));
D_ASSERT(list_empty(&mdev->sync_ee));
D_ASSERT(list_empty(&mdev->done_ee));
- /* ok, no more ee's on the fly, it is safe to reset the epoch_size */
- atomic_set(&mdev->current_epoch->epoch_size, 0);
- D_ASSERT(list_empty(&mdev->current_epoch->list));
-
return 0;
}
if (!mdev)
return -EIO;
+ if (test_bit(CONN_WD_ST_CHG_REQ, &tconn->flags)) {
+ D_ASSERT(tconn->agreed_pro_version < 100);
+ return got_conn_RqSReply(tconn, pi);
+ }
+
if (retcode >= SS_SUCCESS) {
set_bit(CL_ST_CHG_SUCCESS, &mdev->flags);
} else {
case P_RECV_ACK:
what = RECV_ACKED_BY_PEER;
break;
- case P_DISCARD_WRITE:
- what = DISCARD_WRITE;
+ case P_SUPERSEDED:
+ what = CONFLICT_RESOLVED;
break;
case P_RETRY_WRITE:
what = POSTPONE_WRITE;
update_peer_seq(mdev, be32_to_cpu(p->seq_num));
- dev_err(DEV, "Got NegDReply; Sector %llus, len %u; Fail original request.\n",
+ dev_err(DEV, "Got NegDReply; Sector %llus, len %u.\n",
(unsigned long long)sector, be32_to_cpu(p->blksize));
return validate_req_change_req_state(mdev, p->block_id, sector,
static int got_BarrierAck(struct drbd_tconn *tconn, struct packet_info *pi)
{
- struct drbd_conf *mdev;
struct p_barrier_ack *p = pi->data;
+ struct drbd_conf *mdev;
+ int vnr;
- mdev = vnr_to_mdev(tconn, pi->vnr);
- if (!mdev)
- return -EIO;
-
- tl_release(mdev->tconn, p->barrier, be32_to_cpu(p->set_size));
+ tl_release(tconn, p->barrier, be32_to_cpu(p->set_size));
- if (mdev->state.conn == C_AHEAD &&
- atomic_read(&mdev->ap_in_flight) == 0 &&
- !test_and_set_bit(AHEAD_TO_SYNC_SOURCE, &mdev->current_epoch->flags)) {
- mdev->start_resync_timer.expires = jiffies + HZ;
- add_timer(&mdev->start_resync_timer);
+ rcu_read_lock();
+ idr_for_each_entry(&tconn->volumes, mdev, vnr) {
+ if (mdev->state.conn == C_AHEAD &&
+ atomic_read(&mdev->ap_in_flight) == 0 &&
+ !test_and_set_bit(AHEAD_TO_SYNC_SOURCE, &mdev->flags)) {
+ mdev->start_resync_timer.expires = jiffies + HZ;
+ add_timer(&mdev->start_resync_timer);
+ }
}
+ rcu_read_unlock();
return 0;
}
if (w) {
w->cb = w_ov_finished;
w->mdev = mdev;
- drbd_queue_work_front(&mdev->tconn->data.work, w);
+ drbd_queue_work(&mdev->tconn->sender_work, w);
} else {
dev_err(DEV, "kmalloc(w) failed.");
ov_out_of_sync_print(mdev);
[P_RECV_ACK] = { sizeof(struct p_block_ack), got_BlockAck },
[P_WRITE_ACK] = { sizeof(struct p_block_ack), got_BlockAck },
[P_RS_WRITE_ACK] = { sizeof(struct p_block_ack), got_BlockAck },
- [P_DISCARD_WRITE] = { sizeof(struct p_block_ack), got_BlockAck },
+ [P_SUPERSEDED] = { sizeof(struct p_block_ack), got_BlockAck },
[P_NEG_ACK] = { sizeof(struct p_block_ack), got_NegAck },
[P_NEG_DREPLY] = { sizeof(struct p_block_ack), got_NegDReply },
[P_NEG_RS_DREPLY] = { sizeof(struct p_block_ack), got_NegRSDReply },
received += rv;
buf += rv;
} else if (rv == 0) {
+ if (test_bit(DISCONNECT_SENT, &tconn->flags)) {
+ long t;
+ rcu_read_lock();
+ t = rcu_dereference(tconn->net_conf)->ping_timeo * HZ/10;
+ rcu_read_unlock();
+
+ t = wait_event_timeout(tconn->ping_wait,
+ tconn->cstate < C_WF_REPORT_PARAMS,
+ t);
+ if (t)
+ break;
+ }
conn_err(tconn, "meta connection shut down by peer.\n");
goto reconnect;
} else if (rv == -EAGAIN) {
if (0) {
reconnect:
conn_request_state(tconn, NS(conn, C_NETWORK_FAILURE), CS_HARD);
+ conn_md_sync(tconn);
}
if (0) {
disconnect: