From 422028b1ca4c07995af82a18abced022ff4c296c Mon Sep 17 00:00:00 2001
From: Philipp Reisner <philipp.reisner@linbit.com>
Date: Wed, 27 Oct 2010 11:12:07 +0200
Subject: drbd: New configuration parameters for dealing with network
 congestion

net {
    on_congestion {block|pull-ahead|disconnect};
    congestion-fill {sectors};
    congestion-extents {al-extents};
}

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 include/linux/drbd.h | 7 +++++++
 1 file changed, 7 insertions(+)

(limited to 'include/linux/drbd.h')

diff --git a/include/linux/drbd.h b/include/linux/drbd.h
index ef44c7a0638..03a08baabf1 100644
--- a/include/linux/drbd.h
+++ b/include/linux/drbd.h
@@ -96,6 +96,12 @@ enum drbd_on_no_data {
 	OND_SUSPEND_IO
 };
 
+enum drbd_on_congestion {
+	OC_BLOCK,
+	OC_PULL_AHEAD,
+	OC_DISCONNECT,
+};
+
 /* KEEP the order, do not delete or insert. Only append. */
 enum drbd_ret_codes {
 	ERR_CODE_BASE		= 100,
@@ -146,6 +152,7 @@ enum drbd_ret_codes {
 	ERR_PERM		= 152,
 	ERR_NEED_APV_93		= 153,
 	ERR_STONITH_AND_PROT_A  = 154,
+	ERR_CONG_NOT_PROTO_A	= 155,
 
 	/* insert new ones above this line */
 	AFTER_LAST_ERR_CODE
-- 
cgit v1.2.3-70-g09d2


From 67531718d8f1259f01ab84c2aa25f7b03c7afd46 Mon Sep 17 00:00:00 2001
From: Philipp Reisner <philipp.reisner@linbit.com>
Date: Wed, 27 Oct 2010 12:21:30 +0200
Subject: drbd: Implemented two new connection states Ahead/Behind

In this connection mode, the ahead node no longer replicates
application IO. The behind's disk becomes out dated.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 drivers/block/drbd/drbd_int.h      |  2 ++
 drivers/block/drbd/drbd_main.c     | 12 ++++++++++--
 drivers/block/drbd/drbd_receiver.c |  3 +++
 drivers/block/drbd/drbd_req.c      | 23 +++++++++++++++++++++++
 drivers/block/drbd/drbd_strings.c  |  4 +++-
 include/linux/drbd.h               |  4 ++++
 6 files changed, 45 insertions(+), 3 deletions(-)

(limited to 'include/linux/drbd.h')

diff --git a/drivers/block/drbd/drbd_int.h b/drivers/block/drbd/drbd_int.h
index c804e44b945..21b7439438c 100644
--- a/drivers/block/drbd/drbd_int.h
+++ b/drivers/block/drbd/drbd_int.h
@@ -2217,6 +2217,8 @@ static inline int drbd_state_is_stable(union drbd_state s)
 	case C_VERIFY_T:
 	case C_PAUSED_SYNC_S:
 	case C_PAUSED_SYNC_T:
+	case C_AHEAD:
+	case C_BEHIND:
 		/* maybe stable, look at the disk state */
 		break;
 
diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c
index e81d009dd06..46f27d6c0b2 100644
--- a/drivers/block/drbd/drbd_main.c
+++ b/drivers/block/drbd/drbd_main.c
@@ -871,16 +871,19 @@ static union drbd_state sanitize_state(struct drbd_conf *mdev, union drbd_state
 
 	if (ns.conn >= C_CONNECTED &&
 	    ((ns.disk == D_CONSISTENT || ns.disk == D_OUTDATED) ||
-	     (ns.disk == D_NEGOTIATING && ns.conn == C_WF_BITMAP_T))) {
+	     (ns.disk == D_NEGOTIATING && ns.conn == C_WF_BITMAP_T) ||
+	     ns.conn >= C_AHEAD)) {
 		switch (ns.conn) {
 		case C_WF_BITMAP_T:
 		case C_PAUSED_SYNC_T:
+		case C_BEHIND:
 			ns.disk = D_OUTDATED;
 			break;
 		case C_CONNECTED:
 		case C_WF_BITMAP_S:
 		case C_SYNC_SOURCE:
 		case C_PAUSED_SYNC_S:
+		case C_AHEAD:
 			ns.disk = D_UP_TO_DATE;
 			break;
 		case C_SYNC_TARGET:
@@ -893,16 +896,18 @@ static union drbd_state sanitize_state(struct drbd_conf *mdev, union drbd_state
 	}
 
 	if (ns.conn >= C_CONNECTED &&
-	    (ns.pdsk == D_CONSISTENT || ns.pdsk == D_OUTDATED)) {
+	    (ns.pdsk == D_CONSISTENT || ns.pdsk == D_OUTDATED || ns.conn >= C_AHEAD)) {
 		switch (ns.conn) {
 		case C_CONNECTED:
 		case C_WF_BITMAP_T:
 		case C_PAUSED_SYNC_T:
 		case C_SYNC_TARGET:
+		case C_BEHIND:
 			ns.pdsk = D_UP_TO_DATE;
 			break;
 		case C_WF_BITMAP_S:
 		case C_PAUSED_SYNC_S:
+		case C_AHEAD:
 			/* remap any consistent state to D_OUTDATED,
 			 * but disallow "upgrade" of not even consistent states.
 			 */
@@ -1374,6 +1379,9 @@ static void after_state_ch(struct drbd_conf *mdev, union drbd_state os,
 	if (os.conn == C_WF_REPORT_PARAMS && ns.conn >= C_CONNECTED)
 		drbd_send_state(mdev);
 
+	if (os.conn != C_AHEAD && ns.conn == C_AHEAD)
+		drbd_send_state(mdev);
+
 	/* We are in the progress to start a full sync... */
 	if ((os.conn != C_STARTING_SYNC_T && ns.conn == C_STARTING_SYNC_T) ||
 	    (os.conn != C_STARTING_SYNC_S && ns.conn == C_STARTING_SYNC_S))
diff --git a/drivers/block/drbd/drbd_receiver.c b/drivers/block/drbd/drbd_receiver.c
index f3052d871d3..b19e8b2c4ce 100644
--- a/drivers/block/drbd/drbd_receiver.c
+++ b/drivers/block/drbd/drbd_receiver.c
@@ -3179,6 +3179,9 @@ static int receive_state(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned
 	if (ns.conn == C_WF_REPORT_PARAMS)
 		ns.conn = C_CONNECTED;
 
+	if (peer_state.conn == C_AHEAD)
+		ns.conn = C_BEHIND;
+
 	if (mdev->p_uuid && peer_state.disk >= D_NEGOTIATING &&
 	    get_ldev_if_state(mdev, D_NEGOTIATING)) {
 		int cr; /* consider resync */
diff --git a/drivers/block/drbd/drbd_req.c b/drivers/block/drbd/drbd_req.c
index 5c60d77d447..60288fb3c4d 100644
--- a/drivers/block/drbd/drbd_req.c
+++ b/drivers/block/drbd/drbd_req.c
@@ -948,6 +948,29 @@ allocate_barrier:
 				? queue_for_net_write
 				: queue_for_net_read);
 	}
+
+	if (remote && mdev->net_conf->on_congestion != OC_BLOCK) {
+		int congested = 0;
+
+		if (mdev->net_conf->cong_fill &&
+		    atomic_read(&mdev->ap_in_flight) >= mdev->net_conf->cong_fill) {
+			dev_info(DEV, "Congestion-fill threshold reached\n");
+			congested = 1;
+		}
+
+		if (mdev->act_log->used >= mdev->net_conf->cong_extents) {
+			dev_info(DEV, "Congestion-extents threshold reached\n");
+			congested = 1;
+		}
+
+		if (congested) {
+			if (mdev->net_conf->on_congestion == OC_PULL_AHEAD)
+				_drbd_set_state(_NS(mdev, conn, C_AHEAD), 0, NULL);
+			else  /*mdev->net_conf->on_congestion == OC_DISCONNECT */
+				_drbd_set_state(_NS(mdev, conn, C_DISCONNECTING), 0, NULL);
+		}
+	}
+
 	spin_unlock_irq(&mdev->req_lock);
 	kfree(b); /* if someone else has beaten us to it... */
 
diff --git a/drivers/block/drbd/drbd_strings.c b/drivers/block/drbd/drbd_strings.c
index 85179e1fb50..5b970adc3b6 100644
--- a/drivers/block/drbd/drbd_strings.c
+++ b/drivers/block/drbd/drbd_strings.c
@@ -48,6 +48,8 @@ static const char *drbd_conn_s_names[] = {
 	[C_PAUSED_SYNC_T]    = "PausedSyncT",
 	[C_VERIFY_S]         = "VerifyS",
 	[C_VERIFY_T]         = "VerifyT",
+	[C_AHEAD]            = "Ahead",
+	[C_BEHIND]           = "Behind",
 };
 
 static const char *drbd_role_s_names[] = {
@@ -92,7 +94,7 @@ static const char *drbd_state_sw_errors[] = {
 const char *drbd_conn_str(enum drbd_conns s)
 {
 	/* enums are unsigned... */
-	return s > C_PAUSED_SYNC_T ? "TOO_LARGE" : drbd_conn_s_names[s];
+	return s > C_BEHIND ? "TOO_LARGE" : drbd_conn_s_names[s];
 }
 
 const char *drbd_role_str(enum drbd_role s)
diff --git a/include/linux/drbd.h b/include/linux/drbd.h
index 03a08baabf1..23f31be6f00 100644
--- a/include/linux/drbd.h
+++ b/include/linux/drbd.h
@@ -206,6 +206,10 @@ enum drbd_conns {
 	C_VERIFY_T,
 	C_PAUSED_SYNC_S,
 	C_PAUSED_SYNC_T,
+
+	C_AHEAD,
+	C_BEHIND,
+
 	C_MASK = 31
 };
 
-- 
cgit v1.2.3-70-g09d2


From 73a01a18b9c28a0fab1131ece5b0a9bc00a879b8 Mon Sep 17 00:00:00 2001
From: Philipp Reisner <philipp.reisner@linbit.com>
Date: Wed, 27 Oct 2010 14:33:00 +0200
Subject: drbd: New packet for Ahead/Behind mode: P_OUT_OF_SYNC

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 drivers/block/drbd/drbd_actlog.c   | 10 +++++----
 drivers/block/drbd/drbd_int.h      | 14 +++++++++++-
 drivers/block/drbd/drbd_main.c     | 10 +++++++++
 drivers/block/drbd/drbd_receiver.c | 10 +++++++++
 drivers/block/drbd/drbd_req.c      | 44 +++++++++++++++++++++++++++-----------
 drivers/block/drbd/drbd_req.h      |  4 +++-
 drivers/block/drbd/drbd_worker.c   | 16 ++++++++++++++
 include/linux/drbd.h               |  2 +-
 8 files changed, 91 insertions(+), 19 deletions(-)

(limited to 'include/linux/drbd.h')

diff --git a/drivers/block/drbd/drbd_actlog.c b/drivers/block/drbd/drbd_actlog.c
index b4adb58c747..33f6cc537d0 100644
--- a/drivers/block/drbd/drbd_actlog.c
+++ b/drivers/block/drbd/drbd_actlog.c
@@ -1007,22 +1007,22 @@ void __drbd_set_in_sync(struct drbd_conf *mdev, sector_t sector, int size,
  * called by tl_clear and drbd_send_dblock (==drbd_make_request).
  * so this can be _any_ process.
  */
-void __drbd_set_out_of_sync(struct drbd_conf *mdev, sector_t sector, int size,
+int __drbd_set_out_of_sync(struct drbd_conf *mdev, sector_t sector, int size,
 			    const char *file, const unsigned int line)
 {
 	unsigned long sbnr, ebnr, lbnr, flags;
 	sector_t esector, nr_sectors;
-	unsigned int enr, count;
+	unsigned int enr, count = 0;
 	struct lc_element *e;
 
 	if (size <= 0 || (size & 0x1ff) != 0 || size > DRBD_MAX_BIO_SIZE) {
 		dev_err(DEV, "sector: %llus, size: %d\n",
 			(unsigned long long)sector, size);
-		return;
+		return 0;
 	}
 
 	if (!get_ldev(mdev))
-		return; /* no disk, no metadata, no bitmap to set bits in */
+		return 0; /* no disk, no metadata, no bitmap to set bits in */
 
 	nr_sectors = drbd_get_capacity(mdev->this_bdev);
 	esector = sector + (size >> 9) - 1;
@@ -1052,6 +1052,8 @@ void __drbd_set_out_of_sync(struct drbd_conf *mdev, sector_t sector, int size,
 
 out:
 	put_ldev(mdev);
+
+	return count;
 }
 
 static
diff --git a/drivers/block/drbd/drbd_int.h b/drivers/block/drbd/drbd_int.h
index 21b7439438c..47133123682 100644
--- a/drivers/block/drbd/drbd_int.h
+++ b/drivers/block/drbd/drbd_int.h
@@ -212,6 +212,7 @@ enum drbd_packets {
 	/* P_CKPT_FENCE_REQ      = 0x25, * currently reserved for protocol D */
 	/* P_CKPT_DISABLE_REQ    = 0x26, * currently reserved for protocol D */
 	P_DELAY_PROBE         = 0x27, /* is used on BOTH sockets */
+	P_OUT_OF_SYNC         = 0x28, /* Mark as out of sync (Outrunning), data socket */
 
 	P_MAX_CMD	      = 0x28,
 	P_MAY_IGNORE	      = 0x100, /* Flag to test if (cmd > P_MAY_IGNORE) ... */
@@ -269,6 +270,7 @@ static inline const char *cmdname(enum drbd_packets cmd)
 		[P_RS_IS_IN_SYNC]	= "CsumRSIsInSync",
 		[P_COMPRESSED_BITMAP]   = "CBitmap",
 		[P_DELAY_PROBE]         = "DelayProbe",
+		[P_OUT_OF_SYNC]		= "OutOfSync",
 		[P_MAX_CMD]	        = NULL,
 	};
 
@@ -550,6 +552,13 @@ struct p_discard {
 	u32	    pad;
 } __packed;
 
+struct p_block_desc {
+	struct p_header80 head;
+	u64 sector;
+	u32 blksize;
+	u32 pad;	/* to multiple of 8 Byte */
+} __packed;
+
 /* Valid values for the encoding field.
  * Bump proto version when changing this. */
 enum drbd_bitmap_code {
@@ -647,6 +656,7 @@ union p_polymorph {
         struct p_block_req       block_req;
 	struct p_delay_probe93   delay_probe93;
 	struct p_rs_uuid         rs_uuid;
+	struct p_block_desc      block_desc;
 } __packed;
 
 /**********************************************************************/
@@ -1221,6 +1231,7 @@ extern int drbd_send_ack_dp(struct drbd_conf *mdev, enum drbd_packets cmd,
 			struct p_data *dp, int data_size);
 extern int drbd_send_ack_ex(struct drbd_conf *mdev, enum drbd_packets cmd,
 			    sector_t sector, int blksize, u64 block_id);
+extern int drbd_send_oos(struct drbd_conf *mdev, struct drbd_request *req);
 extern int drbd_send_block(struct drbd_conf *mdev, enum drbd_packets cmd,
 			   struct drbd_epoch_entry *e);
 extern int drbd_send_dblock(struct drbd_conf *mdev, struct drbd_request *req);
@@ -1534,6 +1545,7 @@ extern int w_send_read_req(struct drbd_conf *, struct drbd_work *, int);
 extern int w_prev_work_done(struct drbd_conf *, struct drbd_work *, int);
 extern int w_e_reissue(struct drbd_conf *, struct drbd_work *, int);
 extern int w_restart_disk_io(struct drbd_conf *, struct drbd_work *, int);
+extern int w_send_oos(struct drbd_conf *, struct drbd_work *, int);
 
 extern void resync_timer_fn(unsigned long data);
 
@@ -1626,7 +1638,7 @@ extern void __drbd_set_in_sync(struct drbd_conf *mdev, sector_t sector,
 		int size, const char *file, const unsigned int line);
 #define drbd_set_in_sync(mdev, sector, size) \
 	__drbd_set_in_sync(mdev, sector, size, __FILE__, __LINE__)
-extern void __drbd_set_out_of_sync(struct drbd_conf *mdev, sector_t sector,
+extern int __drbd_set_out_of_sync(struct drbd_conf *mdev, sector_t sector,
 		int size, const char *file, const unsigned int line);
 #define drbd_set_out_of_sync(mdev, sector, size) \
 	__drbd_set_out_of_sync(mdev, sector, size, __FILE__, __LINE__)
diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c
index 46f27d6c0b2..0dc93f43a47 100644
--- a/drivers/block/drbd/drbd_main.c
+++ b/drivers/block/drbd/drbd_main.c
@@ -2634,6 +2634,16 @@ int drbd_send_block(struct drbd_conf *mdev, enum drbd_packets cmd,
 	return ok;
 }
 
+int drbd_send_oos(struct drbd_conf *mdev, struct drbd_request *req)
+{
+	struct p_block_desc p;
+
+	p.sector  = cpu_to_be64(req->sector);
+	p.blksize = cpu_to_be32(req->size);
+
+	return drbd_send_cmd(mdev, USE_DATA_SOCKET, P_OUT_OF_SYNC, &p.head, sizeof(p));
+}
+
 /*
   drbd_send distinguishes two cases:
 
diff --git a/drivers/block/drbd/drbd_receiver.c b/drivers/block/drbd/drbd_receiver.c
index b19e8b2c4ce..04a08e7541c 100644
--- a/drivers/block/drbd/drbd_receiver.c
+++ b/drivers/block/drbd/drbd_receiver.c
@@ -3562,6 +3562,15 @@ static int receive_UnplugRemote(struct drbd_conf *mdev, enum drbd_packets cmd, u
 	return TRUE;
 }
 
+static int receive_out_of_sync(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned int data_size)
+{
+	struct p_block_desc *p = &mdev->data.rbuf.block_desc;
+
+	drbd_set_out_of_sync(mdev, be64_to_cpu(p->sector), be32_to_cpu(p->blksize));
+
+	return TRUE;
+}
+
 typedef int (*drbd_cmd_handler_f)(struct drbd_conf *, enum drbd_packets cmd, unsigned int to_receive);
 
 struct data_cmd {
@@ -3592,6 +3601,7 @@ static struct data_cmd drbd_cmd_handler[] = {
 	[P_OV_REPLY]        = { 1, sizeof(struct p_block_req), receive_DataRequest },
 	[P_CSUM_RS_REQUEST] = { 1, sizeof(struct p_block_req), receive_DataRequest },
 	[P_DELAY_PROBE]     = { 0, sizeof(struct p_delay_probe93), receive_skip },
+	[P_OUT_OF_SYNC]     = { 0, sizeof(struct p_block_desc), receive_out_of_sync },
 	/* anything missing from this table is in
 	 * the asender_tbl, see get_asender_cmd */
 	[P_MAX_CMD]	    = { 0, 0, NULL },
diff --git a/drivers/block/drbd/drbd_req.c b/drivers/block/drbd/drbd_req.c
index 60288fb3c4d..a8d1ff2bda2 100644
--- a/drivers/block/drbd/drbd_req.c
+++ b/drivers/block/drbd/drbd_req.c
@@ -142,7 +142,7 @@ static void _about_to_complete_local_write(struct drbd_conf *mdev,
 
 	/* before we can signal completion to the upper layers,
 	 * we may need to close the current epoch */
-	if (mdev->state.conn >= C_CONNECTED &&
+	if (mdev->state.conn >= C_CONNECTED && mdev->state.conn < C_AHEAD &&
 	    req->epoch == mdev->newest_tle->br_number)
 		queue_barrier(mdev);
 
@@ -545,6 +545,14 @@ int __req_mod(struct drbd_request *req, enum drbd_req_event what,
 
 		break;
 
+	case queue_for_send_oos:
+		req->rq_state |= RQ_NET_QUEUED;
+		req->w.cb =  w_send_oos;
+		drbd_queue_work(&mdev->data.work, &req->w);
+		break;
+
+	case oos_handed_to_network:
+		/* actually the same */
 	case send_canceled:
 		/* treat it the same */
 	case send_failed:
@@ -756,7 +764,7 @@ static int drbd_make_request_common(struct drbd_conf *mdev, struct bio *bio)
 	const sector_t sector = bio->bi_sector;
 	struct drbd_tl_epoch *b = NULL;
 	struct drbd_request *req;
-	int local, remote;
+	int local, remote, send_oos = 0;
 	int err = -EIO;
 	int ret = 0;
 
@@ -820,8 +828,11 @@ static int drbd_make_request_common(struct drbd_conf *mdev, struct bio *bio)
 	}
 
 	remote = remote && (mdev->state.pdsk == D_UP_TO_DATE ||
-			    (mdev->state.pdsk == D_INCONSISTENT &&
-			     mdev->state.conn >= C_CONNECTED));
+			    (mdev->state.pdsk >= D_INCONSISTENT &&
+			     mdev->state.conn >= C_CONNECTED &&
+			     mdev->state.conn < C_AHEAD));
+	send_oos = (rw == WRITE && mdev->state.conn == C_AHEAD &&
+		    mdev->state.pdsk >= D_INCONSISTENT);
 
 	if (!(local || remote) && !is_susp(mdev->state)) {
 		if (__ratelimit(&drbd_ratelimit_state))
@@ -835,7 +846,7 @@ static int drbd_make_request_common(struct drbd_conf *mdev, struct bio *bio)
 	 * but there is a race between testing the bit and pointer outside the
 	 * spinlock, and grabbing the spinlock.
 	 * if we lost that race, we retry.  */
-	if (rw == WRITE && remote &&
+	if (rw == WRITE && (remote || send_oos) &&
 	    mdev->unused_spare_tle == NULL &&
 	    test_bit(CREATE_BARRIER, &mdev->flags)) {
 allocate_barrier:
@@ -860,11 +871,15 @@ allocate_barrier:
 		goto fail_free_complete;
 	}
 
-	if (remote) {
+	if (remote || send_oos) {
 		remote = (mdev->state.pdsk == D_UP_TO_DATE ||
-			    (mdev->state.pdsk == D_INCONSISTENT &&
-			     mdev->state.conn >= C_CONNECTED));
-		if (!remote)
+			    (mdev->state.pdsk >= D_INCONSISTENT &&
+			     mdev->state.conn >= C_CONNECTED &&
+			     mdev->state.conn < C_AHEAD));
+		send_oos = (rw == WRITE && mdev->state.conn == C_AHEAD &&
+			    mdev->state.pdsk >= D_INCONSISTENT);
+
+		if (!(remote || send_oos))
 			dev_warn(DEV, "lost connection while grabbing the req_lock!\n");
 		if (!(local || remote)) {
 			dev_err(DEV, "IO ERROR: neither local nor remote disk\n");
@@ -877,7 +892,7 @@ allocate_barrier:
 		mdev->unused_spare_tle = b;
 		b = NULL;
 	}
-	if (rw == WRITE && remote &&
+	if (rw == WRITE && (remote || send_oos) &&
 	    mdev->unused_spare_tle == NULL &&
 	    test_bit(CREATE_BARRIER, &mdev->flags)) {
 		/* someone closed the current epoch
@@ -900,7 +915,7 @@ allocate_barrier:
 	 * barrier packet.  To get the write ordering right, we only have to
 	 * make sure that, if this is a write request and it triggered a
 	 * barrier packet, this request is queued within the same spinlock. */
-	if (remote && mdev->unused_spare_tle &&
+	if ((remote || send_oos) && mdev->unused_spare_tle &&
 	    test_and_clear_bit(CREATE_BARRIER, &mdev->flags)) {
 		_tl_add_barrier(mdev, mdev->unused_spare_tle);
 		mdev->unused_spare_tle = NULL;
@@ -948,8 +963,11 @@ allocate_barrier:
 				? queue_for_net_write
 				: queue_for_net_read);
 	}
+	if (send_oos && drbd_set_out_of_sync(mdev, sector, size))
+		_req_mod(req, queue_for_send_oos);
 
-	if (remote && mdev->net_conf->on_congestion != OC_BLOCK) {
+	if (remote &&
+	    mdev->net_conf->on_congestion != OC_BLOCK && mdev->agreed_pro_version >= 96) {
 		int congested = 0;
 
 		if (mdev->net_conf->cong_fill &&
@@ -964,6 +982,8 @@ allocate_barrier:
 		}
 
 		if (congested) {
+			queue_barrier(mdev);
+
 			if (mdev->net_conf->on_congestion == OC_PULL_AHEAD)
 				_drbd_set_state(_NS(mdev, conn, C_AHEAD), 0, NULL);
 			else  /*mdev->net_conf->on_congestion == OC_DISCONNECT */
diff --git a/drivers/block/drbd/drbd_req.h b/drivers/block/drbd/drbd_req.h
index 69d350fe7c1..40d3dcd8fc8 100644
--- a/drivers/block/drbd/drbd_req.h
+++ b/drivers/block/drbd/drbd_req.h
@@ -82,14 +82,16 @@ enum drbd_req_event {
 	to_be_submitted,
 
 	/* XXX yes, now I am inconsistent...
-	 * these two are not "events" but "actions"
+	 * these are not "events" but "actions"
 	 * oh, well... */
 	queue_for_net_write,
 	queue_for_net_read,
+	queue_for_send_oos,
 
 	send_canceled,
 	send_failed,
 	handed_over_to_network,
+	oos_handed_to_network,
 	connection_lost_while_pending,
 	read_retry_remote_canceled,
 	recv_acked_by_peer,
diff --git a/drivers/block/drbd/drbd_worker.c b/drivers/block/drbd/drbd_worker.c
index 782d87237cb..67499077c48 100644
--- a/drivers/block/drbd/drbd_worker.c
+++ b/drivers/block/drbd/drbd_worker.c
@@ -1237,6 +1237,22 @@ int w_send_write_hint(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
 	return drbd_send_short_cmd(mdev, P_UNPLUG_REMOTE);
 }
 
+int w_send_oos(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
+{
+	struct drbd_request *req = container_of(w, struct drbd_request, w);
+	int ok;
+
+	if (unlikely(cancel)) {
+		req_mod(req, send_canceled);
+		return 1;
+	}
+
+	ok = drbd_send_oos(mdev, req);
+	req_mod(req, oos_handed_to_network);
+
+	return ok;
+}
+
 /**
  * w_send_dblock() - Worker callback to send a P_DATA packet in order to mirror a write request
  * @mdev:	DRBD device.
diff --git a/include/linux/drbd.h b/include/linux/drbd.h
index 23f31be6f00..41da654cc0b 100644
--- a/include/linux/drbd.h
+++ b/include/linux/drbd.h
@@ -56,7 +56,7 @@ extern const char *drbd_buildtag(void);
 #define REL_VERSION "8.3.9"
 #define API_VERSION 88
 #define PRO_VERSION_MIN 86
-#define PRO_VERSION_MAX 95
+#define PRO_VERSION_MAX 96
 
 
 enum drbd_io_error_p {
-- 
cgit v1.2.3-70-g09d2


From 116676ca621a862a8124969772f4dd61c8b40eee Mon Sep 17 00:00:00 2001
From: Andreas Gruenbacher <agruen@linbit.com>
Date: Wed, 8 Dec 2010 13:33:11 +0100
Subject: drbd: Rename enum drbd_ret_codes to enum drbd_ret_code

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 drivers/block/drbd/drbd_main.c | 2 +-
 drivers/block/drbd/drbd_nl.c   | 4 ++--
 include/linux/drbd.h           | 2 +-
 3 files changed, 4 insertions(+), 4 deletions(-)

(limited to 'include/linux/drbd.h')

diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c
index f43e2aa354a..8d69e3a1b3c 100644
--- a/drivers/block/drbd/drbd_main.c
+++ b/drivers/block/drbd/drbd_main.c
@@ -3570,7 +3570,7 @@ void drbd_md_sync(struct drbd_conf *mdev)
  * @mdev:	DRBD device.
  * @bdev:	Device from which the meta data should be read in.
  *
- * Return 0 (NO_ERROR) on success, and an enum drbd_ret_codes in case
+ * Return 0 (NO_ERROR) on success, and an enum drbd_ret_code in case
  * something goes wrong.  Currently only: ERR_IO_MD_DISK, ERR_MD_INVALID.
  */
 int drbd_md_read(struct drbd_conf *mdev, struct drbd_backing_dev *bdev)
diff --git a/drivers/block/drbd/drbd_nl.c b/drivers/block/drbd/drbd_nl.c
index cd0459f0403..fe336592e53 100644
--- a/drivers/block/drbd/drbd_nl.c
+++ b/drivers/block/drbd/drbd_nl.c
@@ -849,7 +849,7 @@ static void drbd_suspend_al(struct drbd_conf *mdev)
 static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
 			     struct drbd_nl_cfg_reply *reply)
 {
-	enum drbd_ret_codes retcode;
+	enum drbd_ret_code retcode;
 	enum determine_dev_size dd;
 	sector_t max_possible_sectors;
 	sector_t min_md_device_sectors;
@@ -1278,7 +1278,7 @@ static int drbd_nl_net_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
 			    struct drbd_nl_cfg_reply *reply)
 {
 	int i, ns;
-	enum drbd_ret_codes retcode;
+	enum drbd_ret_code retcode;
 	struct net_conf *new_conf = NULL;
 	struct crypto_hash *tfm = NULL;
 	struct crypto_hash *integrity_w_tfm = NULL;
diff --git a/include/linux/drbd.h b/include/linux/drbd.h
index 41da654cc0b..d92f989036c 100644
--- a/include/linux/drbd.h
+++ b/include/linux/drbd.h
@@ -103,7 +103,7 @@ enum drbd_on_congestion {
 };
 
 /* KEEP the order, do not delete or insert. Only append. */
-enum drbd_ret_codes {
+enum drbd_ret_code {
 	ERR_CODE_BASE		= 100,
 	NO_ERROR		= 101,
 	ERR_LOCAL_ADDR		= 102,
-- 
cgit v1.2.3-70-g09d2


From c8b325632f0e5ffdaeca3d1f3be77c9399316a40 Mon Sep 17 00:00:00 2001
From: Andreas Gruenbacher <agruen@linbit.com>
Date: Wed, 8 Dec 2010 01:06:16 +0100
Subject: drbd: Rename enum drbd_state_ret_codes to enum drbd_state_rv

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 drivers/block/drbd/drbd_main.c    | 5 +++--
 drivers/block/drbd/drbd_strings.c | 2 +-
 include/linux/drbd.h              | 4 ++--
 3 files changed, 6 insertions(+), 5 deletions(-)

(limited to 'include/linux/drbd.h')

diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c
index 8d69e3a1b3c..cddf311b742 100644
--- a/drivers/block/drbd/drbd_main.c
+++ b/drivers/block/drbd/drbd_main.c
@@ -510,8 +510,9 @@ static union drbd_state sanitize_state(struct drbd_conf *mdev, union drbd_state
 int drbd_send_state_req(struct drbd_conf *,
 			union drbd_state, union drbd_state);
 
-static enum drbd_state_ret_codes _req_st_cond(struct drbd_conf *mdev,
-				    union drbd_state mask, union drbd_state val)
+static enum drbd_state_rv
+_req_st_cond(struct drbd_conf *mdev, union drbd_state mask,
+	     union drbd_state val)
 {
 	union drbd_state os, ns;
 	unsigned long flags;
diff --git a/drivers/block/drbd/drbd_strings.c b/drivers/block/drbd/drbd_strings.c
index 5b970adc3b6..c44a2a60277 100644
--- a/drivers/block/drbd/drbd_strings.c
+++ b/drivers/block/drbd/drbd_strings.c
@@ -107,7 +107,7 @@ const char *drbd_disk_str(enum drbd_disk_state s)
 	return s > D_UP_TO_DATE    ? "TOO_LARGE" : drbd_disk_s_names[s];
 }
 
-const char *drbd_set_st_err_str(enum drbd_state_ret_codes err)
+const char *drbd_set_st_err_str(enum drbd_state_rv err)
 {
 	return err <= SS_AFTER_LAST_ERROR ? "TOO_SMALL" :
 	       err > SS_TWO_PRIMARIES ? "TOO_LARGE"
diff --git a/include/linux/drbd.h b/include/linux/drbd.h
index d92f989036c..d10431fab00 100644
--- a/include/linux/drbd.h
+++ b/include/linux/drbd.h
@@ -270,7 +270,7 @@ union drbd_state {
 	unsigned int i;
 };
 
-enum drbd_state_ret_codes {
+enum drbd_state_rv {
 	SS_CW_NO_NEED = 4,
 	SS_CW_SUCCESS = 3,
 	SS_NOTHING_TO_DO = 2,
@@ -301,7 +301,7 @@ enum drbd_state_ret_codes {
 extern const char *drbd_conn_str(enum drbd_conns);
 extern const char *drbd_role_str(enum drbd_role);
 extern const char *drbd_disk_str(enum drbd_disk_state);
-extern const char *drbd_set_st_err_str(enum drbd_state_ret_codes);
+extern const char *drbd_set_st_err_str(enum drbd_state_rv);
 
 #define SHARED_SECRET_MAX 64
 
-- 
cgit v1.2.3-70-g09d2


From cd88d030d41a9b0100fd5fee872024e6ebc8b276 Mon Sep 17 00:00:00 2001
From: Philipp Reisner <philipp.reisner@linbit.com>
Date: Thu, 20 Jan 2011 11:46:41 +0100
Subject: drbd: Provide hints with the error message when clearing the sync
 pause flag

When the user clears the sync-pause flag, and sync stays in pause
state, give hints to the user, why it still is in pause state.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 drivers/block/drbd/drbd_nl.c | 12 ++++++++++--
 include/linux/drbd.h         |  2 ++
 2 files changed, 12 insertions(+), 2 deletions(-)

(limited to 'include/linux/drbd.h')

diff --git a/drivers/block/drbd/drbd_nl.c b/drivers/block/drbd/drbd_nl.c
index 434b621f76a..ffe3a97fef9 100644
--- a/drivers/block/drbd/drbd_nl.c
+++ b/drivers/block/drbd/drbd_nl.c
@@ -1952,9 +1952,17 @@ static int drbd_nl_resume_sync(struct drbd_conf *mdev, struct drbd_nl_cfg_req *n
 			       struct drbd_nl_cfg_reply *reply)
 {
 	int retcode = NO_ERROR;
+	union drbd_state s;
 
-	if (drbd_request_state(mdev, NS(user_isp, 0)) == SS_NOTHING_TO_DO)
-		retcode = ERR_PAUSE_IS_CLEAR;
+	if (drbd_request_state(mdev, NS(user_isp, 0)) == SS_NOTHING_TO_DO) {
+		s = mdev->state;
+		if (s.conn == C_PAUSED_SYNC_S || s.conn == C_PAUSED_SYNC_T) {
+			retcode = s.aftr_isp ? ERR_PIC_AFTER_DEP :
+				  s.peer_isp ? ERR_PIC_PEER_DEP : ERR_PAUSE_IS_CLEAR;
+		} else {
+			retcode = ERR_PAUSE_IS_CLEAR;
+		}
+	}
 
 	reply->ret_code = retcode;
 	return 0;
diff --git a/include/linux/drbd.h b/include/linux/drbd.h
index d10431fab00..ba5c785d3f7 100644
--- a/include/linux/drbd.h
+++ b/include/linux/drbd.h
@@ -153,6 +153,8 @@ enum drbd_ret_code {
 	ERR_NEED_APV_93		= 153,
 	ERR_STONITH_AND_PROT_A  = 154,
 	ERR_CONG_NOT_PROTO_A	= 155,
+	ERR_PIC_AFTER_DEP	= 156,
+	ERR_PIC_PEER_DEP	= 157,
 
 	/* insert new ones above this line */
 	AFTER_LAST_ERR_CODE
-- 
cgit v1.2.3-70-g09d2


From c5a91619793d444e5103ec5841045bf878718398 Mon Sep 17 00:00:00 2001
From: Andreas Gruenbacher <agruen@linbit.com>
Date: Tue, 25 Jan 2011 17:33:38 +0100
Subject: drbd: Remove unused function atodb_endio()

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 drivers/block/drbd/drbd_actlog.c | 27 ---------------------------
 drivers/block/drbd/drbd_worker.c | 15 ++++++---------
 include/linux/drbd.h             |  2 +-
 3 files changed, 7 insertions(+), 37 deletions(-)

(limited to 'include/linux/drbd.h')

diff --git a/drivers/block/drbd/drbd_actlog.c b/drivers/block/drbd/drbd_actlog.c
index a6050791401..2a1642bc451 100644
--- a/drivers/block/drbd/drbd_actlog.c
+++ b/drivers/block/drbd/drbd_actlog.c
@@ -542,33 +542,6 @@ cancel:
 	return 1;
 }
 
-static void atodb_endio(struct bio *bio, int error)
-{
-	struct drbd_atodb_wait *wc = bio->bi_private;
-	struct drbd_conf *mdev = wc->mdev;
-	struct page *page;
-	int uptodate = bio_flagged(bio, BIO_UPTODATE);
-
-	/* strange behavior of some lower level drivers...
-	 * fail the request by clearing the uptodate flag,
-	 * but do not return any error?! */
-	if (!error && !uptodate)
-		error = -EIO;
-
-	drbd_chk_io_error(mdev, error, true);
-	if (error && wc->error == 0)
-		wc->error = error;
-
-	if (atomic_dec_and_test(&wc->count))
-		complete(&wc->io_done);
-
-	page = bio->bi_io_vec[0].bv_page;
-	put_page(page);
-	bio_put(bio);
-	mdev->bm_writ_cnt++;
-	put_ldev(mdev);
-}
-
 /**
  * drbd_al_apply_to_bm() - Sets the bitmap to diry(1) where covered ba active AL extents
  * @mdev:	DRBD device.
diff --git a/drivers/block/drbd/drbd_worker.c b/drivers/block/drbd/drbd_worker.c
index cfd324b9f95..3d70d8d015d 100644
--- a/drivers/block/drbd/drbd_worker.c
+++ b/drivers/block/drbd/drbd_worker.c
@@ -44,15 +44,12 @@ static int w_make_resync_request(struct drbd_conf *mdev,
 
 
 
-/* defined here:
-   drbd_md_io_complete
-   drbd_endio_sec
-   drbd_endio_pri
-
- * more endio handlers:
-   atodb_endio in drbd_actlog.c
-   drbd_bm_async_io_complete in drbd_bitmap.c
-
+/* endio handlers:
+ *   drbd_md_io_complete (defined here)
+ *   drbd_endio_pri (defined here)
+ *   drbd_endio_sec (defined here)
+ *   bm_async_io_complete (defined in drbd_bitmap.c)
+ *
  * For all these callbacks, note the following:
  * The callbacks will be called in irq context by the IDE drivers,
  * and in Softirqs/Tasklets/BH context by the SCSI drivers.
diff --git a/include/linux/drbd.h b/include/linux/drbd.h
index ba5c785d3f7..d18d673ebc7 100644
--- a/include/linux/drbd.h
+++ b/include/linux/drbd.h
@@ -53,7 +53,7 @@
 
 
 extern const char *drbd_buildtag(void);
-#define REL_VERSION "8.3.9"
+#define REL_VERSION "8.3.10"
 #define API_VERSION 88
 #define PRO_VERSION_MIN 86
 #define PRO_VERSION_MAX 96
-- 
cgit v1.2.3-70-g09d2