drbd: Fixes from the drbd-8.3 branch

* drbd-8.3:
  drbd: O_SYNC gives EIO on ramdisks for some kernels (eg. RHEL6).
  drbd: send intermediate state change results to the peer

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>

5 files changed, 40 insertions, 18 deletions

diff --git a/drivers/block/drbd/drbd_int.h b/drivers/block/drbd/drbd_int.h
index ccf28e986d4d..3177e3e60610 100644
--- a/drivers/block/drbd/drbd_int.h
+++ b/drivers/block/drbd/drbd_int.h
@@ -1062,8 +1062,8 @@ extern void drbd_gen_and_send_sync_uuid(struct drbd_conf *mdev);
 extern int drbd_send_sizes(struct drbd_conf *mdev, int trigger_reply, enum dds_flags flags);
 extern int _conn_send_state_req(struct drbd_tconn *, int vnr, enum drbd_packet cmd,
 				union drbd_state, union drbd_state);
-extern int _drbd_send_state(struct drbd_conf *mdev);
-extern int drbd_send_state(struct drbd_conf *mdev);
+extern int drbd_send_state(struct drbd_conf *mdev, union drbd_state s);
+extern int drbd_send_current_state(struct drbd_conf *mdev);
 extern int drbd_send_sync_param(struct drbd_conf *mdev);
 extern void drbd_send_b_ack(struct drbd_conf *mdev, u32 barrier_nr,
 			    u32 set_size);
diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c
index 822fb3d42356..e3186f2dad0b 100644
--- a/drivers/block/drbd/drbd_main.c
+++ b/drivers/block/drbd/drbd_main.c
@@ -1148,10 +1148,10 @@ int drbd_send_sizes(struct drbd_conf *mdev, int trigger_reply, enum dds_flags fl
 }
 
 /**
- * drbd_send_state() - Sends the drbd state to the peer
+ * drbd_send_current_state() - Sends the drbd state to the peer
  * @mdev:	DRBD device.
  */
-int drbd_send_state(struct drbd_conf *mdev)
+int drbd_send_current_state(struct drbd_conf *mdev)
 {
 	struct drbd_socket *sock;
 	struct p_state *p;
@@ -1164,6 +1164,29 @@ int drbd_send_state(struct drbd_conf *mdev)
 	return drbd_send_command(mdev, sock, P_STATE, sizeof(*p), NULL, 0);
 }
 
+/**
+ * drbd_send_state() - After a state change, sends the new state to the peer
+ * @mdev:      DRBD device.
+ * @state:     the state to send, not necessarily the current state.
+ *
+ * Each state change queues an "after_state_ch" work, which will eventually
+ * send the resulting new state to the peer. If more state changes happen
+ * between queuing and processing of the after_state_ch work, we still
+ * want to send each intermediary state in the order it occurred.
+ */
+int drbd_send_state(struct drbd_conf *mdev, union drbd_state state)
+{
+	struct drbd_socket *sock;
+	struct p_state *p;
+
+	sock = &mdev->tconn->data;
+	p = drbd_prepare_command(mdev, sock);
+	if (!p)
+		return -EIO;
+	p->state = cpu_to_be32(state.i); /* Within the send mutex */
+	return drbd_send_command(mdev, sock, P_STATE, sizeof(*p), NULL, 0);
+}
+
 int drbd_send_state_req(struct drbd_conf *mdev, union drbd_state mask, union drbd_state val)
 {
 	struct drbd_socket *sock;
@@ -1176,7 +1199,6 @@ int drbd_send_state_req(struct drbd_conf *mdev, union drbd_state mask, union drb
 	p->mask = cpu_to_be32(mask.i);
 	p->val = cpu_to_be32(val.i);
 	return drbd_send_command(mdev, sock, P_STATE_CHG_REQ, sizeof(*p), NULL, 0);
-
 }
 
 int conn_send_state_req(struct drbd_tconn *tconn, union drbd_state mask, union drbd_state val)
@@ -3003,7 +3025,7 @@ int drbd_md_read(struct drbd_conf *mdev, struct drbd_backing_dev *bdev)
 		goto err;
 	}
 	if (magic != DRBD_MD_MAGIC_08) {
-		if (magic == DRBD_MD_MAGIC_07) 
+		if (magic == DRBD_MD_MAGIC_07)
 			dev_err(DEV, "Found old (0.7) meta data magic. Did you \"drbdadm create-md\"?\n");
 		else
 			dev_err(DEV, "Meta data magic not found. Did you \"drbdadm create-md\"?\n");
diff --git a/drivers/block/drbd/drbd_nl.c b/drivers/block/drbd/drbd_nl.c
index c50c1753aa9a..b342c95b724b 100644
--- a/drivers/block/drbd/drbd_nl.c
+++ b/drivers/block/drbd/drbd_nl.c
@@ -654,7 +654,7 @@ drbd_set_role(struct drbd_conf *mdev, enum drbd_role new_role, int force)
 		/* if this was forced, we should consider sync */
 		if (forced)
 			drbd_send_uuids(mdev);
-		drbd_send_state(mdev);
+		drbd_send_current_state(mdev);
 	}
 
 	drbd_md_sync(mdev);
diff --git a/drivers/block/drbd/drbd_receiver.c b/drivers/block/drbd/drbd_receiver.c
index e823968f1cd2..ed1fa3f7cc4e 100644
--- a/drivers/block/drbd/drbd_receiver.c
+++ b/drivers/block/drbd/drbd_receiver.c
@@ -825,7 +825,7 @@ int drbd_connected(struct drbd_conf *mdev)
 	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. */
@@ -3873,7 +3873,7 @@ static int receive_state(struct drbd_tconn *tconn, struct packet_info *pi)
 			/* 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);
 		}
 	}
 
diff --git a/drivers/block/drbd/drbd_state.c b/drivers/block/drbd/drbd_state.c
index dffc6973e2b5..58c77bbae445 100644
--- a/drivers/block/drbd/drbd_state.c
+++ b/drivers/block/drbd/drbd_state.c
@@ -1198,7 +1198,7 @@ static void after_state_ch(struct drbd_conf *mdev, union drbd_state os,
 	if (os.pdsk == D_DISKLESS &&
 	    ns.pdsk > D_DISKLESS && ns.pdsk != D_UNKNOWN) {      /* attach on the peer */
 		drbd_send_uuids(mdev);
-		drbd_send_state(mdev);
+		drbd_send_state(mdev, ns);
 	}
 	/* No point in queuing send_bitmap if we don't have a connection
 	 * anymore, so check also the _current_ state, not only the new state
@@ -1263,14 +1263,14 @@ static void after_state_ch(struct drbd_conf *mdev, union drbd_state os,
 	    os.disk == D_ATTACHING && ns.disk == D_NEGOTIATING) {
 		drbd_send_sizes(mdev, 0, 0);  /* to start sync... */
 		drbd_send_uuids(mdev);
-		drbd_send_state(mdev);
+		drbd_send_state(mdev, ns);
 	}
 
 	/* We want to pause/continue resync, tell peer. */
 	if (ns.conn >= C_CONNECTED &&
 	     ((os.aftr_isp != ns.aftr_isp) ||
 	      (os.user_isp != ns.user_isp)))
-		drbd_send_state(mdev);
+		drbd_send_state(mdev, ns);
 
 	/* In case one of the isp bits got set, suspend other devices. */
 	if ((!os.aftr_isp && !os.peer_isp && !os.user_isp) &&
@@ -1280,10 +1280,10 @@ static void after_state_ch(struct drbd_conf *mdev, union drbd_state os,
 	/* Make sure the peer gets informed about eventual state
 	   changes (ISP bits) while we were in WFReportParams. */
 	if (os.conn == C_WF_REPORT_PARAMS && ns.conn >= C_CONNECTED)
-		drbd_send_state(mdev);
+		drbd_send_state(mdev, ns);
 
 	if (os.conn != C_AHEAD && ns.conn == C_AHEAD)
-		drbd_send_state(mdev);
+		drbd_send_state(mdev, ns);
 
 	/* We are in the progress to start a full sync... */
 	if ((os.conn != C_STARTING_SYNC_T && ns.conn == C_STARTING_SYNC_T) ||
@@ -1325,7 +1325,7 @@ static void after_state_ch(struct drbd_conf *mdev, union drbd_state os,
 				"ASSERT FAILED: disk is %s during detach\n",
 				drbd_disk_str(mdev->state.disk));
 
-		drbd_send_state(mdev);
+		drbd_send_state(mdev, ns);
 		drbd_rs_cancel_all(mdev);
 
 		/* In case we want to get something to stable storage still,
@@ -1353,7 +1353,7 @@ static void after_state_ch(struct drbd_conf *mdev, union drbd_state os,
                 mdev->rs_failed = 0;
                 atomic_set(&mdev->rs_pending_cnt, 0);
 
-		drbd_send_state(mdev);
+		drbd_send_state(mdev, ns);
 		/* corresponding get_ldev in __drbd_set_state
 		 * this may finally trigger drbd_ldev_destroy. */
 		put_ldev(mdev);
@@ -1361,7 +1361,7 @@ static void after_state_ch(struct drbd_conf *mdev, union drbd_state os,
 
 	/* Notify peer that I had a local IO error, and did not detached.. */
 	if (os.disk == D_UP_TO_DATE && ns.disk == D_INCONSISTENT)
-		drbd_send_state(mdev);
+		drbd_send_state(mdev, ns);
 
 	/* Disks got bigger while they were detached */
 	if (ns.disk > D_NEGOTIATING && ns.pdsk > D_NEGOTIATING &&
@@ -1379,7 +1379,7 @@ static void after_state_ch(struct drbd_conf *mdev, union drbd_state os,
 	/* sync target done with resync.  Explicitly notify peer, even though
 	 * it should (at least for non-empty resyncs) already know itself. */
 	if (os.disk < D_UP_TO_DATE && os.conn >= C_SYNC_SOURCE && ns.conn == C_CONNECTED)
-		drbd_send_state(mdev);
+		drbd_send_state(mdev, ns);
 
 	/* This triggers bitmap writeout of potentially still unwritten pages
 	 * if the resync finished cleanly, or aborted because of peer disk