DRBD updates for the 3.19 merge window

PATCH Applied Philipp Reisner 2 participants 9 messages Nov 10, 2014

Subsystems (2)

Block layer DRBD driver

lore

Share & tools

lore

Philipp Reisner

author maintainer

2014-11-10 16:21 UTC #0 lore raw

cover DRBD updates for the 3.19 merge window

Hi Jens,

please take these patches for you for-3.19/drivers branch. They where posted for review on 2014-10-29.

Some cleanups and 3 fixes. Two of them marked for stable:
0005-drbd-merge_bvec_fn-properly-remap-bvm-bi_bdev.patch and 0006-drbd-Fix-state-change-in-case-of-connection-timeout.patch (marked for stable)

Thanks!

Andreas Gruenbacher (2):
drbd: Minor cleanups
drbd: Only use drbd_msg_put_info() in drbd_nl.c

Lars Ellenberg (2):
drbd: fix resync throttling initialization
drbd: merge_bvec_fn: properly remap bvm->bi_bdev

Philipp Reisner (3):
drbd: fix race between role change and handshake
drbd: Fix state change in case of connection timeout
drbd: Remove an useless copy of kernel_setsockopt()

 drivers/block/drbd/drbd_actlog.c   |  3 +-
 drivers/block/drbd/drbd_int.h      | 39 +++++------------------
 drivers/block/drbd/drbd_main.c     | 23 ++++----------
 drivers/block/drbd/drbd_nl.c       | 64 +++++++++++++++++---------------------
 drivers/block/drbd/drbd_receiver.c |  2 +-
 drivers/block/drbd/drbd_req.c      |  3 +-
 drivers/block/drbd/drbd_state.c    | 42 ++++++++++++++++++++-----
 drivers/block/drbd/drbd_state.h    |  5 +++
 drivers/block/drbd/drbd_worker.c   |  5 ++-
 include/linux/drbd.h               |  2 +-
 10 files changed, 90 insertions(+), 98 deletions(-)

Signature · 1 line

1.9.1

Philipp Reisner

author maintainer

2014-11-10 16:21 UTC #1 lore raw

PATCH 1/7 drbd: Minor cleanups

From: Andreas Gruenbacher <[email protected]>

. Update comments
. drbd_set_{in,out_of}_sync(): Remove unused parameters . Move common code into adm_del_resource()
. Redefine ERR_MINOR_EXISTS -> ERR_MINOR_OR_VOLUME_EXISTS

Signed-off-by Philipp Reisner <[email protected]>

Signed-off-by Lars Ellenberg <[email protected]>

---
 drivers/block/drbd/drbd_actlog.c |  3 +-
 drivers/block/drbd/drbd_int.h    |  9 +++---
 drivers/block/drbd/drbd_main.c   |  6 ++--
 drivers/block/drbd/drbd_nl.c     | 60 +++++++++++++++++-----------------------
 include/linux/drbd.h             |  2 +-
 5 files changed, 35 insertions(+), 45 deletions(-)

drivers/block/drbd/drbd_actlog.c +1 -2

index a2dfa16..1318e32 100644
@@ -827,8 +827,7 @@ static int update_sync_bits(struct drbd_device *device,
  *
  */
 int __drbd_change_sync(struct drbd_device *device, sector_t sector, int size,
-		enum update_sync_bits_mode mode,
-		const char *file, const unsigned int line)
+		enum update_sync_bits_mode mode)
 {
 	/* Is called from worker and receiver context _only_ */
 	unsigned long sbnr, ebnr, lbnr;

drivers/block/drbd/drbd_int.h +4 -5

index 9b22f8f..c14b718 100644
@@ -1662,14 +1662,13 @@ extern void drbd_advance_rs_marks(struct drbd_device *device, unsigned long stil
 
 enum update_sync_bits_mode { RECORD_RS_FAILED, SET_OUT_OF_SYNC, SET_IN_SYNC };
 extern int __drbd_change_sync(struct drbd_device *device, sector_t sector, int size,
-		enum update_sync_bits_mode mode,
-		const char *file, const unsigned int line);
+		enum update_sync_bits_mode mode);
 #define drbd_set_in_sync(device, sector, size) \
-	__drbd_change_sync(device, sector, size, SET_IN_SYNC, __FILE__, __LINE__)
+	__drbd_change_sync(device, sector, size, SET_IN_SYNC)
 #define drbd_set_out_of_sync(device, sector, size) \
-	__drbd_change_sync(device, sector, size, SET_OUT_OF_SYNC, __FILE__, __LINE__)
+	__drbd_change_sync(device, sector, size, SET_OUT_OF_SYNC)
 #define drbd_rs_failed_io(device, sector, size) \
-	__drbd_change_sync(device, sector, size, RECORD_RS_FAILED, __FILE__, __LINE__)
+	__drbd_change_sync(device, sector, size, RECORD_RS_FAILED)
 extern void drbd_al_shrink(struct drbd_device *device);
 extern int drbd_initialize_al(struct drbd_device *, void *);

drivers/block/drbd/drbd_main.c +3 -3

index 973c185..cce25a5 100644
@@ -2731,7 +2731,7 @@ enum drbd_ret_code drbd_create_device(struct drbd_config_context *adm_ctx, unsig
 
 	device = minor_to_device(minor);
 	if (device)
-		return ERR_MINOR_EXISTS;
+		return ERR_MINOR_OR_VOLUME_EXISTS;
 
 	/* GFP_KERNEL, we are outside of all write-out paths */
 	device = kzalloc(sizeof(struct drbd_device), GFP_KERNEL);

@@ -2794,7 +2794,7 @@ enum drbd_ret_code drbd_create_device(struct drbd_config_context *adm_ctx, unsig
 	id = idr_alloc(&drbd_devices, device, minor, minor + 1, GFP_KERNEL);
 	if (id < 0) {
 		if (id == -ENOSPC) {
-			err = ERR_MINOR_EXISTS;
+			err = ERR_MINOR_OR_VOLUME_EXISTS;
 			drbd_msg_put_info(adm_ctx->reply_skb, "requested minor exists already");
 		}
 		goto out_no_minor_idr;

@@ -2804,7 +2804,7 @@ enum drbd_ret_code drbd_create_device(struct drbd_config_context *adm_ctx, unsig
 	id = idr_alloc(&resource->devices, device, vnr, vnr + 1, GFP_KERNEL);
 	if (id < 0) {
 		if (id == -ENOSPC) {
-			err = ERR_MINOR_EXISTS;
+			err = ERR_MINOR_OR_VOLUME_EXISTS;
 			drbd_msg_put_info(adm_ctx->reply_skb, "requested minor exists already");
 		}
 		goto out_idr_remove_minor;

drivers/block/drbd/drbd_nl.c +26 ~~-34~~

104 lines not on this page · raw message

include/linux/drbd.h +1 -1

index debb70d..8723f2a 100644
@@ -172,7 +172,7 @@ enum drbd_ret_code {
 	ERR_RES_NOT_KNOWN	= 158,
 	ERR_RES_IN_USE		= 159,
 	ERR_MINOR_CONFIGURED    = 160,
-	ERR_MINOR_EXISTS	= 161,
+	ERR_MINOR_OR_VOLUME_EXISTS = 161,
 	ERR_INVALID_REQUEST	= 162,
 	ERR_NEED_APV_100	= 163,
 	ERR_NEED_ALLOW_TWO_PRI  = 164,

Signature · 1 line

1.9.1

Philipp Reisner

author maintainer

2014-11-10 16:21 UTC #2 lore raw

PATCH 2/7 drbd: Only use drbd_msg_put_info() in drbd_nl.c

In reply to Philipp Reisner

From: Andreas Gruenbacher <[email protected]>

Avoid generic netlink calls in other parts of the code base.

Signed-off-by Philipp Reisner <[email protected]>

Signed-off-by Lars Ellenberg <[email protected]>

---
 drivers/block/drbd/drbd_int.h  |  1 -
 drivers/block/drbd/drbd_main.c | 17 +++--------------
 drivers/block/drbd/drbd_nl.c   |  2 +-
 3 files changed, 4 insertions(+), 16 deletions(-)

drivers/block/drbd/drbd_int.h +0 -1

index c14b718..d2a7589 100644
@@ -1454,7 +1454,6 @@ extern int is_valid_ar_handle(struct drbd_request *, sector_t);
 
 
 /* drbd_nl.c */
-extern int drbd_msg_put_info(struct sk_buff *skb, const char *info);
 extern void drbd_suspend_io(struct drbd_device *device);
 extern void drbd_resume_io(struct drbd_device *device);
 extern char *ppsize(char *buf, unsigned long long size);

drivers/block/drbd/drbd_main.c +3 ~~-14~~

index cce25a5..1fc8342 100644
@@ -2532,10 +2532,6 @@ int set_resource_options(struct drbd_resource *resource, struct res_opts *res_op
 
 	if (!zalloc_cpumask_var(&new_cpu_mask, GFP_KERNEL))
 		return -ENOMEM;
-		/*
-		retcode = ERR_NOMEM;
-		drbd_msg_put_info("unable to allocate cpumask");
-		*/
 
 	/* silently ignore cpu mask on UP kernel */
 	if (nr_cpu_ids > 1 && res_opts->cpu_mask[0] != 0) {

@@ -2793,20 +2789,16 @@ enum drbd_ret_code drbd_create_device(struct drbd_config_context *adm_ctx, unsig
 
 	id = idr_alloc(&drbd_devices, device, minor, minor + 1, GFP_KERNEL);
 	if (id < 0) {
-		if (id == -ENOSPC) {
+		if (id == -ENOSPC)
 			err = ERR_MINOR_OR_VOLUME_EXISTS;
-			drbd_msg_put_info(adm_ctx->reply_skb, "requested minor exists already");
-		}
 		goto out_no_minor_idr;
 	}
 	kref_get(&device->kref);
 
 	id = idr_alloc(&resource->devices, device, vnr, vnr + 1, GFP_KERNEL);
 	if (id < 0) {
-		if (id == -ENOSPC) {
+		if (id == -ENOSPC)
 			err = ERR_MINOR_OR_VOLUME_EXISTS;
-			drbd_msg_put_info(adm_ctx->reply_skb, "requested minor exists already");
-		}
 		goto out_idr_remove_minor;
 	}
 	kref_get(&device->kref);

@@ -2825,10 +2817,8 @@ enum drbd_ret_code drbd_create_device(struct drbd_config_context *adm_ctx, unsig
 
 		id = idr_alloc(&connection->peer_devices, peer_device, vnr, vnr + 1, GFP_KERNEL);
 		if (id < 0) {
-			if (id == -ENOSPC) {
+			if (id == -ENOSPC)
 				err = ERR_INVALID_REQUEST;
-				drbd_msg_put_info(adm_ctx->reply_skb, "requested volume exists already");
-			}
 			goto out_idr_remove_from_resource;
 		}
 		kref_get(&connection->kref);

@@ -2836,7 +2826,6 @@ enum drbd_ret_code drbd_create_device(struct drbd_config_context *adm_ctx, unsig
 
 	if (init_submitter(device)) {
 		err = ERR_NOMEM;
-		drbd_msg_put_info(adm_ctx->reply_skb, "unable to create submit workqueue");
 		goto out_idr_remove_vol;
 	}

drivers/block/drbd/drbd_nl.c +1 -1

index c145619..4782d07 100644
@@ -92,7 +92,7 @@ static void drbd_adm_send_reply(struct sk_buff *skb, struct genl_info *info)
 
 /* Used on a fresh "drbd_adm_prepare"d reply_skb, this cannot fail: The only
  * reason it could fail was no space in skb, and there are 4k available. */
-int drbd_msg_put_info(struct sk_buff *skb, const char *info)
+static int drbd_msg_put_info(struct sk_buff *skb, const char *info)
 {
 	struct nlattr *nla;
 	int err = -EMSGSIZE;

Signature · 1 line

1.9.1

Philipp Reisner

author maintainer

2014-11-10 16:21 UTC #3 lore raw

PATCH 3/7 drbd: fix race between role change and handshake

In reply to Philipp Reisner

Symptoms:
If DRBD was "cleanly shut down" (all in sync, both Secondary before disconnect, identical data generation uuids), and then one side was promoted *during* the next connection handshake, the role change could confuse the handshake.

The Primary would get stuck in WFBitmapS, the Secondary would log unexpected cstate (Connected) in receive_bitmap
and get stuck in WFBitmapT.

Fix:
The test in is_valid_soft_transition wrong. It works because the not allowed actions (promote/attach) do not touch the cstate. The previous condition failed to demand a cstate change in one clause.

In order to avoid deadlocks give up the state_mutex while waiting for the transient state to go away.

Conflicts:
drbd/drbd_state.c
drbd/drbd_state.h
drbd/drbd_wrappers.h

Signed-off-by Philipp Reisner <[email protected]>

Signed-off-by Lars Ellenberg <[email protected]>

---
 drivers/block/drbd/drbd_nl.c    |  2 +-
 drivers/block/drbd/drbd_state.c | 41 ++++++++++++++++++++++++++++++++++-------
 drivers/block/drbd/drbd_state.h |  5 +++++
 3 files changed, 40 insertions(+), 8 deletions(-)

drivers/block/drbd/drbd_nl.c +1 -1

index 4782d07..74df8cf 100644
@@ -588,7 +588,7 @@ drbd_set_role(struct drbd_device *const device, enum drbd_role new_role, int for
 	val.i  = 0; val.role  = new_role;
 
 	while (try++ < max_tries) {
-		rv = _drbd_request_state(device, mask, val, CS_WAIT_COMPLETE);
+		rv = _drbd_request_state_holding_state_mutex(device, mask, val, CS_WAIT_COMPLETE);
 
 		/* in case we first succeeded to outdate,
 		 * but now suddenly could establish a connection */

drivers/block/drbd/drbd_state.c +34 -7

68 lines not on this page · raw message

drivers/block/drbd/drbd_state.h +5 -0

index cc41605..7f53c40 100644
@@ -117,6 +117,11 @@ extern enum drbd_state_rv _drbd_request_state(struct drbd_device *,
 					      union drbd_state,
 					      union drbd_state,
 					      enum chg_state_flags);
+
+extern enum drbd_state_rv
+_drbd_request_state_holding_state_mutex(struct drbd_device *, union drbd_state,
+					union drbd_state, enum chg_state_flags);
+
 extern enum drbd_state_rv __drbd_set_state(struct drbd_device *, union drbd_state,
 					   enum chg_state_flags,
 					   struct completion *done);

Signature · 1 line

1.9.1

Philipp Reisner

author maintainer

2014-11-10 16:21 UTC #4 lore raw

PATCH 4/7 drbd: fix resync throttling initialization

In reply to Philipp Reisner

From: Lars Ellenberg <[email protected]>

If for some reason DRBD resync was the only activity on a backend device, drbd_rs_c_min_rate_throttle() would mistakenly decide that it is still initialization time, and keep throttling the resync.

This patch explicitly initializes ->rs_last_events to the current backend event counters, and drops the rs_last_events == 0 from the throttle condition.

Reported-by Mikhail Sugakov <[email protected]>

Signed-off-by Philipp Reisner <[email protected]>

Signed-off-by Lars Ellenberg <[email protected]>

---
 drivers/block/drbd/drbd_receiver.c | 2 +-
 drivers/block/drbd/drbd_state.c    | 1 -
 drivers/block/drbd/drbd_worker.c   | 5 ++++-
 3 files changed, 5 insertions(+), 3 deletions(-)

drivers/block/drbd/drbd_receiver.c +1 -1

index 6960fb0..d169b4a 100644
@@ -2482,7 +2482,7 @@ bool drbd_rs_c_min_rate_throttle(struct drbd_device *device)
 			atomic_read(&device->rs_sect_ev);
 
 	if (atomic_read(&device->ap_actlog_cnt)
-	    || !device->rs_last_events || curr_events - device->rs_last_events > 64) {
+	    || curr_events - device->rs_last_events > 64) {
 		unsigned long rs_left;
 		int i;

drivers/block/drbd/drbd_state.c +0 -1

index 4529d92..2d7dd26 100644
@@ -1099,7 +1099,6 @@ __drbd_set_state(struct drbd_device *device, union drbd_state ns,
 
 		set_ov_position(device, ns.conn);
 		device->rs_start = now;
-		device->rs_last_events = 0;
 		device->rs_last_sect_ev = 0;
 		device->ov_last_oos_size = 0;
 		device->ov_last_oos_start = 0;

drivers/block/drbd/drbd_worker.c +4 -1

index d2d1f97..d0fae55 100644
@@ -1592,11 +1592,15 @@ void drbd_resync_after_changed(struct drbd_device *device)
 
 void drbd_rs_controller_reset(struct drbd_device *device)
 {
+	struct gendisk *disk = device->ldev->backing_bdev->bd_contains->bd_disk;
 	struct fifo_buffer *plan;
 
 	atomic_set(&device->rs_sect_in, 0);
 	atomic_set(&device->rs_sect_ev, 0);
 	device->rs_in_flight = 0;
+	device->rs_last_events =
+		(int)part_stat_read(&disk->part0, sectors[0]) +
+		(int)part_stat_read(&disk->part0, sectors[1]);
 
 	/* Updating the RCU protected object in place is necessary since
 	   this function gets called from atomic context.

@@ -1743,7 +1747,6 @@ void drbd_start_resync(struct drbd_device *device, enum drbd_conns side)
 		device->rs_failed    = 0;
 		device->rs_paused    = 0;
 		device->rs_same_csum = 0;
-		device->rs_last_events = 0;
 		device->rs_last_sect_ev = 0;
 		device->rs_total     = tw;
 		device->rs_start     = now;

Signature · 1 line

1.9.1

Philipp Reisner

author maintainer

2014-11-10 16:21 UTC #5 lore raw

PATCH 5/7 drbd: merge_bvec_fn: properly remap bvm->bi_bdev

In reply to Philipp Reisner

From: Lars Ellenberg <[email protected]>

This was not noticed for many years. Affects operation if md raid is used a backing device for DRBD.

CC [email protected] # v3.2+

Signed-off-by Philipp Reisner <[email protected]>

Signed-off-by Lars Ellenberg <[email protected]>

---
 drivers/block/drbd/drbd_req.c | 1 +
 1 file changed, 1 insertion(+)

drivers/block/drbd/drbd_req.c +1 -0

index 5a01c53..90319b1 100644
@@ -1545,6 +1545,7 @@ int drbd_merge_bvec(struct request_queue *q, struct bvec_merge_data *bvm, struct
 		struct request_queue * const b =
 			device->ldev->backing_bdev->bd_disk->queue;
 		if (b->merge_bvec_fn) {
+			bvm->bi_bdev = device->ldev->backing_bdev;
 			backing_limit = b->merge_bvec_fn(b, bvm, bvec);
 			limit = min(limit, backing_limit);
 		}

Signature · 1 line

1.9.1

Philipp Reisner

author maintainer

2014-11-10 16:21 UTC #6 lore raw

PATCH 6/7 drbd: Fix state change in case of connection timeout

In reply to Philipp Reisner

A connection timeout affects all volumes of a resource! Under the following conditions:

A resource with multiple volumes
AND
ko-count >=1
AND
a write request triggers the timeout (ko-count * timeout)

DRBD's internal state gets confused. That in turn may lead to very miss leading follow up failures. E.g. "BUG: scheduling while atomic"

CC [email protected] # v3.17

Signed-off-by Philipp Reisner <[email protected]>

Signed-off-by Lars Ellenberg <[email protected]>

---
 drivers/block/drbd/drbd_req.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

drivers/block/drbd/drbd_req.c +1 -1

index 90319b1..3b797cd 100644
@@ -1629,7 +1629,7 @@ void request_timer_fn(unsigned long data)
 		 time_after(now, req_peer->pre_send_jif + ent) &&
 		!time_in_range(now, connection->last_reconnect_jif, connection->last_reconnect_jif + ent)) {
 		drbd_warn(device, "Remote failed to finish a request within ko-count * timeout\n");
-		_drbd_set_state(_NS(device, conn, C_TIMEOUT), CS_VERBOSE | CS_HARD, NULL);
+		_conn_request_state(connection, NS(conn, C_TIMEOUT), CS_VERBOSE | CS_HARD);
 	}
 	if (dt && oldest_submit_jif != now &&
 		 time_after(now, oldest_submit_jif + dt) &&

Signature · 1 line

1.9.1

Philipp Reisner

author maintainer

2014-11-10 16:21 UTC #7 lore raw

PATCH 7/7 drbd: Remove an useless copy of kernel_setsockopt()

In reply to Philipp Reisner

Old backward-compat cruft

Signed-off-by Philipp Reisner <[email protected]>

Signed-off-by Lars Ellenberg <[email protected]>

---
 drivers/block/drbd/drbd_int.h | 29 ++++-------------------------
 1 file changed, 4 insertions(+), 25 deletions(-)

drivers/block/drbd/drbd_int.h +4 ~~-25~~

index d2a7589..b905e98 100644
@@ -1557,52 +1557,31 @@ extern void drbd_set_recv_tcq(struct drbd_device *device, int tcq_enabled);
 extern void _drbd_clear_done_ee(struct drbd_device *device, struct list_head *to_be_freed);
 extern int drbd_connected(struct drbd_peer_device *);
 
-/* Yes, there is kernel_setsockopt, but only since 2.6.18.
- * So we have our own copy of it here. */
-static inline int drbd_setsockopt(struct socket *sock, int level, int optname,
-				  char *optval, int optlen)
-{
-	mm_segment_t oldfs = get_fs();
-	char __user *uoptval;
-	int err;
-
-	uoptval = (char __user __force *)optval;
-
-	set_fs(KERNEL_DS);
-	if (level == SOL_SOCKET)
-		err = sock_setsockopt(sock, level, optname, uoptval, optlen);
-	else
-		err = sock->ops->setsockopt(sock, level, optname, uoptval,
-					    optlen);
-	set_fs(oldfs);
-	return err;
-}
-
 static inline void drbd_tcp_cork(struct socket *sock)
 {
 	int val = 1;
-	(void) drbd_setsockopt(sock, SOL_TCP, TCP_CORK,
+	(void) kernel_setsockopt(sock, SOL_TCP, TCP_CORK,
 			(char*)&val, sizeof(val));
 }
 
 static inline void drbd_tcp_uncork(struct socket *sock)
 {
 	int val = 0;
-	(void) drbd_setsockopt(sock, SOL_TCP, TCP_CORK,
+	(void) kernel_setsockopt(sock, SOL_TCP, TCP_CORK,
 			(char*)&val, sizeof(val));
 }
 
 static inline void drbd_tcp_nodelay(struct socket *sock)
 {
 	int val = 1;
-	(void) drbd_setsockopt(sock, SOL_TCP, TCP_NODELAY,
+	(void) kernel_setsockopt(sock, SOL_TCP, TCP_NODELAY,
 			(char*)&val, sizeof(val));
 }
 
 static inline void drbd_tcp_quickack(struct socket *sock)
 {
 	int val = 2;
-	(void) drbd_setsockopt(sock, SOL_TCP, TCP_QUICKACK,
+	(void) kernel_setsockopt(sock, SOL_TCP, TCP_QUICKACK,
 			(char*)&val, sizeof(val));
 }

Signature · 1 line

1.9.1

Jens Axboe

maintainer Applied

2014-11-10 16:38 UTC #8 lore raw

In reply to Philipp Reisner

Quoted 9 lines · Philipp Reisner

Hi Jens,
please take these patches for you for-3.19/drivers branch. They where posted for review on 2014-10-29.
Some cleanups and 3 fixes. Two of them marked for stable:
0005-drbd-merge_bvec_fn-properly-remap-bvm-bi_bdev.patch and 0006-drbd-Fix-state-change-in-case-of-connection-timeout.patch (marked for stable)

Applied, thanks.

Signature · 1 line

Jens Axboe