if (drbd_submit_ee(mdev, e, WRITE, DRBD_FAULT_RS_WR) == 0)
return TRUE;
+ /* drbd_submit_ee currently fails for one reason only:
+ * not being able to allocate enough bios.
+ * Is dropping the connection going to help? */
+ spin_lock_irq(&mdev->req_lock);
+ list_del(&e->w.list);
+ spin_unlock_irq(&mdev->req_lock);
+
drbd_free_ee(mdev, e);
fail:
put_ldev(mdev);
if (drbd_submit_ee(mdev, e, rw, DRBD_FAULT_DT_WR) == 0)
return TRUE;
+ /* drbd_submit_ee currently fails for one reason only:
+ * not being able to allocate enough bios.
+ * Is dropping the connection going to help? */
+ spin_lock_irq(&mdev->req_lock);
+ list_del(&e->w.list);
+ hlist_del_init(&e->colision);
+ spin_unlock_irq(&mdev->req_lock);
+ if (e->flags & EE_CALL_AL_COMPLETE_IO)
+ drbd_al_complete_io(mdev, e->sector);
+
out_interrupted:
/* yes, the epoch_size now is imbalanced.
* but we drop the connection anyways, so we don't have a chance to
const sector_t capacity = drbd_get_capacity(mdev->this_bdev);
struct drbd_epoch_entry *e;
struct digest_info *di = NULL;
- int size;
+ int size, verb;
unsigned int fault_type;
struct p_block_req *p = &mdev->data.rbuf.block_req;
}
if (!get_ldev_if_state(mdev, D_UP_TO_DATE)) {
- if (__ratelimit(&drbd_ratelimit_state))
+ verb = 1;
+ switch (cmd) {
+ case P_DATA_REQUEST:
+ drbd_send_ack_rp(mdev, P_NEG_DREPLY, p);
+ break;
+ case P_RS_DATA_REQUEST:
+ case P_CSUM_RS_REQUEST:
+ case P_OV_REQUEST:
+ drbd_send_ack_rp(mdev, P_NEG_RS_DREPLY , p);
+ break;
+ case P_OV_REPLY:
+ verb = 0;
+ dec_rs_pending(mdev);
+ drbd_send_ack_ex(mdev, P_OV_RESULT, sector, size, ID_IN_SYNC);
+ break;
+ default:
+ dev_err(DEV, "unexpected command (%s) in receive_DataRequest\n",
+ cmdname(cmd));
+ }
+ if (verb && __ratelimit(&drbd_ratelimit_state))
dev_err(DEV, "Can not satisfy peer's read request, "
"no local data.\n");
- drbd_send_ack_rp(mdev, cmd == P_DATA_REQUEST ? P_NEG_DREPLY :
- P_NEG_RS_DREPLY , p);
+
/* drain possibly payload */
return drbd_drain_block(mdev, digest_size);
}
break;
case P_OV_REQUEST:
- if (mdev->state.conn >= C_CONNECTED &&
- mdev->state.conn != C_VERIFY_T)
- dev_warn(DEV, "ASSERT FAILED: got P_OV_REQUEST while being %s\n",
- drbd_conn_str(mdev->state.conn));
if (mdev->ov_start_sector == ~(sector_t)0 &&
mdev->agreed_pro_version >= 90) {
mdev->ov_start_sector = sector;
if (drbd_submit_ee(mdev, e, READ, fault_type) == 0)
return TRUE;
+ /* drbd_submit_ee currently fails for one reason only:
+ * not being able to allocate enough bios.
+ * Is dropping the connection going to help? */
+ spin_lock_irq(&mdev->req_lock);
+ list_del(&e->w.list);
+ spin_unlock_irq(&mdev->req_lock);
+ /* no drbd_rs_complete_io(), we are dropping the connection anyways */
+
out_free_e:
put_ldev(mdev);
drbd_free_ee(mdev, e);
static int receive_state(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned int data_size)
{
struct p_state *p = &mdev->data.rbuf.state;
- enum drbd_conns nconn, oconn;
- union drbd_state ns, peer_state;
+ union drbd_state os, ns, peer_state;
enum drbd_disk_state real_peer_disk;
enum chg_state_flags cs_flags;
int rv;
spin_lock_irq(&mdev->req_lock);
retry:
- oconn = nconn = mdev->state.conn;
+ os = ns = mdev->state;
spin_unlock_irq(&mdev->req_lock);
- if (nconn == C_WF_REPORT_PARAMS)
- nconn = C_CONNECTED;
+ /* 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 &&
+ 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
+ * will change to inconsistent once the peer reaches active
+ * syncing states.
+ * It may have changed syncer-paused flags, however, so we
+ * cannot ignore this completely. */
+ if (peer_state.conn > C_CONNECTED &&
+ peer_state.conn < C_SYNC_SOURCE)
+ real_peer_disk = D_INCONSISTENT;
+
+ /* if peer_state changes to connected at the same time,
+ * it explicitly notifies us that it finished resync.
+ * Maybe we should finish it up, too? */
+ else if (os.conn >= C_SYNC_SOURCE &&
+ peer_state.conn == C_CONNECTED) {
+ if (drbd_bm_total_weight(mdev) <= mdev->rs_failed)
+ drbd_resync_finished(mdev);
+ return TRUE;
+ }
+ }
+
+ /* 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.
+ * We ignore this to avoid flapping pdsk.
+ * This should not happen, if the peer is a recent version of drbd. */
+ if (os.pdsk == D_UP_TO_DATE && real_peer_disk == D_INCONSISTENT &&
+ os.conn == C_CONNECTED && peer_state.conn > C_SYNC_SOURCE)
+ real_peer_disk = D_UP_TO_DATE;
+
+ if (ns.conn == C_WF_REPORT_PARAMS)
+ ns.conn = C_CONNECTED;
if (mdev->p_uuid && peer_state.disk >= D_NEGOTIATING &&
get_ldev_if_state(mdev, D_NEGOTIATING)) {
int cr; /* consider resync */
/* if we established a new connection */
- cr = (oconn < C_CONNECTED);
+ cr = (os.conn < C_CONNECTED);
/* if we had an established connection
* and one of the nodes newly attaches a disk */
- cr |= (oconn == C_CONNECTED &&
+ cr |= (os.conn == C_CONNECTED &&
(peer_state.disk == D_NEGOTIATING ||
- mdev->state.disk == D_NEGOTIATING));
+ os.disk == D_NEGOTIATING));
/* if we have both been inconsistent, and the peer has been
* forced to be UpToDate with --overwrite-data */
cr |= test_bit(CONSIDER_RESYNC, &mdev->flags);
/* if we had been plain connected, and the admin requested to
* start a sync by "invalidate" or "invalidate-remote" */
- cr |= (oconn == C_CONNECTED &&
+ cr |= (os.conn == C_CONNECTED &&
(peer_state.conn >= C_STARTING_SYNC_S &&
peer_state.conn <= C_WF_BITMAP_T));
if (cr)
- nconn = drbd_sync_handshake(mdev, peer_state.role, real_peer_disk);
+ ns.conn = drbd_sync_handshake(mdev, peer_state.role, real_peer_disk);
put_ldev(mdev);
- if (nconn == C_MASK) {
- nconn = C_CONNECTED;
+ if (ns.conn == C_MASK) {
+ ns.conn = C_CONNECTED;
if (mdev->state.disk == D_NEGOTIATING) {
drbd_force_state(mdev, NS(disk, D_DISKLESS));
} else if (peer_state.disk == D_NEGOTIATING) {
} else {
if (test_and_clear_bit(CONN_DRY_RUN, &mdev->flags))
return FALSE;
- D_ASSERT(oconn == C_WF_REPORT_PARAMS);
+ D_ASSERT(os.conn == C_WF_REPORT_PARAMS);
drbd_force_state(mdev, NS(conn, C_DISCONNECTING));
return FALSE;
}
}
spin_lock_irq(&mdev->req_lock);
- if (mdev->state.conn != oconn)
+ if (mdev->state.i != os.i)
goto retry;
clear_bit(CONSIDER_RESYNC, &mdev->flags);
- ns.i = mdev->state.i;
- ns.conn = nconn;
ns.peer = peer_state.role;
ns.pdsk = real_peer_disk;
ns.peer_isp = (peer_state.aftr_isp | peer_state.user_isp);
- if ((nconn == C_CONNECTED || nconn == C_WF_BITMAP_S) && ns.disk == D_NEGOTIATING)
+ if ((ns.conn == C_CONNECTED || ns.conn == C_WF_BITMAP_S) && ns.disk == D_NEGOTIATING)
ns.disk = mdev->new_state_tmp.disk;
- cs_flags = CS_VERBOSE + (oconn < C_CONNECTED && nconn >= C_CONNECTED ? 0 : CS_HARD);
- if (ns.pdsk == D_CONSISTENT && is_susp(ns) && nconn == C_CONNECTED && oconn < C_CONNECTED &&
+ cs_flags = CS_VERBOSE + (os.conn < C_CONNECTED && ns.conn >= C_CONNECTED ? 0 : CS_HARD);
+ if (ns.pdsk == D_CONSISTENT && is_susp(ns) && ns.conn == C_CONNECTED && os.conn < C_CONNECTED &&
test_bit(NEW_CUR_UUID, &mdev->flags)) {
/* Do not allow tl_restart(resend) for a rebooted peer. We can only allow this
for temporal network outages! */
return FALSE;
}
- if (oconn > C_WF_REPORT_PARAMS) {
- if (nconn > C_CONNECTED && peer_state.conn <= C_CONNECTED &&
+ if (os.conn > C_WF_REPORT_PARAMS) {
+ if (ns.conn > C_CONNECTED && peer_state.conn <= C_CONNECTED &&
peer_state.disk != D_NEGOTIATING ) {
/* we want resync, peer has not yet decided to sync... */
/* Nowadays only used when forcing a node into primary role and
err_out:
drbd_force_state(mdev, NS(conn, C_PROTOCOL_ERROR));
}
+ /* If we leave here, we probably want to update at least the
+ * "Connected" indicator on stable storage. Do so explicitly here. */
+ drbd_md_sync(mdev);
}
void drbd_flush_workqueue(struct drbd_conf *mdev)
while (1) {
clear_bit(SIGNAL_ASENDER, &mdev->flags);
flush_signals(current);
- if (!drbd_process_done_ee(mdev)) {
- dev_err(DEV, "process_done_ee() = NOT_OK\n");
+ if (!drbd_process_done_ee(mdev))
goto reconnect;
- }
/* to avoid race with newly queued ACKs */
set_bit(SIGNAL_ASENDER, &mdev->flags);
spin_lock_irq(&mdev->req_lock);
if (0) {
reconnect:
drbd_force_state(mdev, NS(conn, C_NETWORK_FAILURE));
+ drbd_md_sync(mdev);
}
if (0) {
disconnect:
drbd_force_state(mdev, NS(conn, C_DISCONNECTING));
+ drbd_md_sync(mdev);
}
clear_bit(SIGNAL_ASENDER, &mdev->flags);
projects / deliverable / linux.git / blobdiff