xref: /openbmc/linux/drivers/block/drbd/drbd_worker.c (revision 22cc37a9)

/*
   drbd_worker.c

   This file is part of DRBD by Philipp Reisner and Lars Ellenberg.

   Copyright (C) 2001-2008, LINBIT Information Technologies GmbH.
   Copyright (C) 1999-2008, Philipp Reisner <philipp.reisner@linbit.com>.
   Copyright (C) 2002-2008, Lars Ellenberg <lars.ellenberg@linbit.com>.

   drbd is free software; you can redistribute it and/or modify
   it under the terms of the GNU General Public License as published by
   the Free Software Foundation; either version 2, or (at your option)
   any later version.

   drbd is distributed in the hope that it will be useful,
   but WITHOUT ANY WARRANTY; without even the implied warranty of
   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
   GNU General Public License for more details.

   You should have received a copy of the GNU General Public License
   along with drbd; see the file COPYING.  If not, write to
   the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA.

 */

#include <linux/module.h>
#include <linux/drbd.h>
#include <linux/sched.h>
#include <linux/smp_lock.h>
#include <linux/wait.h>
#include <linux/mm.h>
#include <linux/memcontrol.h>
#include <linux/mm_inline.h>
#include <linux/slab.h>
#include <linux/random.h>
#include <linux/string.h>
#include <linux/scatterlist.h>

#include "drbd_int.h"
#include "drbd_req.h"

static int w_make_ov_request(struct drbd_conf *mdev, struct drbd_work *w, int cancel);



/* 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

 * 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.
 * Try to get the locking right :)
 *
 */


/* About the global_state_lock
   Each state transition on an device holds a read lock. In case we have
   to evaluate the sync after dependencies, we grab a write lock, because
   we need stable states on all devices for that.  */
rwlock_t global_state_lock;

/* used for synchronous meta data and bitmap IO
 * submitted by drbd_md_sync_page_io()
 */
void drbd_md_io_complete(struct bio *bio, int error)
{
	struct drbd_md_io *md_io;

	md_io = (struct drbd_md_io *)bio->bi_private;
	md_io->error = error;

	complete(&md_io->event);
}

/* reads on behalf of the partner,
 * "submitted" by the receiver
 */
void drbd_endio_read_sec_final(struct drbd_epoch_entry *e) __releases(local)
{
	unsigned long flags = 0;
	struct drbd_conf *mdev = e->mdev;

	D_ASSERT(e->block_id != ID_VACANT);

	spin_lock_irqsave(&mdev->req_lock, flags);
	mdev->read_cnt += e->size >> 9;
	list_del(&e->w.list);
	if (list_empty(&mdev->read_ee))
		wake_up(&mdev->ee_wait);
	if (test_bit(__EE_WAS_ERROR, &e->flags))
		__drbd_chk_io_error(mdev, FALSE);
	spin_unlock_irqrestore(&mdev->req_lock, flags);

	drbd_queue_work(&mdev->data.work, &e->w);
	put_ldev(mdev);
}

static int is_failed_barrier(int ee_flags)
{
	return (ee_flags & (EE_IS_BARRIER|EE_WAS_ERROR|EE_RESUBMITTED))
			== (EE_IS_BARRIER|EE_WAS_ERROR);
}

/* writes on behalf of the partner, or resync writes,
 * "submitted" by the receiver, final stage.  */
static void drbd_endio_write_sec_final(struct drbd_epoch_entry *e) __releases(local)
{
	unsigned long flags = 0;
	struct drbd_conf *mdev = e->mdev;
	sector_t e_sector;
	int do_wake;
	int is_syncer_req;
	int do_al_complete_io;

	/* if this is a failed barrier request, disable use of barriers,
	 * and schedule for resubmission */
	if (is_failed_barrier(e->flags)) {
		drbd_bump_write_ordering(mdev, WO_bdev_flush);
		spin_lock_irqsave(&mdev->req_lock, flags);
		list_del(&e->w.list);
		e->flags = (e->flags & ~EE_WAS_ERROR) | EE_RESUBMITTED;
		e->w.cb = w_e_reissue;
		/* put_ldev actually happens below, once we come here again. */
		__release(local);
		spin_unlock_irqrestore(&mdev->req_lock, flags);
		drbd_queue_work(&mdev->data.work, &e->w);
		return;
	}

	D_ASSERT(e->block_id != ID_VACANT);

	/* after we moved e to done_ee,
	 * we may no longer access it,
	 * it may be freed/reused already!
	 * (as soon as we release the req_lock) */
	e_sector = e->sector;
	do_al_complete_io = e->flags & EE_CALL_AL_COMPLETE_IO;
	is_syncer_req = is_syncer_block_id(e->block_id);

	spin_lock_irqsave(&mdev->req_lock, flags);
	mdev->writ_cnt += e->size >> 9;
	list_del(&e->w.list); /* has been on active_ee or sync_ee */
	list_add_tail(&e->w.list, &mdev->done_ee);

	/* No hlist_del_init(&e->colision) here, we did not send the Ack yet,
	 * neither did we wake possibly waiting conflicting requests.
	 * done from "drbd_process_done_ee" within the appropriate w.cb
	 * (e_end_block/e_end_resync_block) or from _drbd_clear_done_ee */

	do_wake = is_syncer_req
		? list_empty(&mdev->sync_ee)
		: list_empty(&mdev->active_ee);

	if (test_bit(__EE_WAS_ERROR, &e->flags))
		__drbd_chk_io_error(mdev, FALSE);
	spin_unlock_irqrestore(&mdev->req_lock, flags);

	if (is_syncer_req)
		drbd_rs_complete_io(mdev, e_sector);

	if (do_wake)
		wake_up(&mdev->ee_wait);

	if (do_al_complete_io)
		drbd_al_complete_io(mdev, e_sector);

	wake_asender(mdev);
	put_ldev(mdev);
}

/* writes on behalf of the partner, or resync writes,
 * "submitted" by the receiver.
 */
void drbd_endio_sec(struct bio *bio, int error)
{
	struct drbd_epoch_entry *e = bio->bi_private;
	struct drbd_conf *mdev = e->mdev;
	int uptodate = bio_flagged(bio, BIO_UPTODATE);
	int is_write = bio_data_dir(bio) == WRITE;

	if (error)
		dev_warn(DEV, "%s: error=%d s=%llus\n",
				is_write ? "write" : "read", error,
				(unsigned long long)e->sector);
	if (!error && !uptodate) {
		dev_warn(DEV, "%s: setting error to -EIO s=%llus\n",
				is_write ? "write" : "read",
				(unsigned long long)e->sector);
		/* strange behavior of some lower level drivers...
		 * fail the request by clearing the uptodate flag,
		 * but do not return any error?! */
		error = -EIO;
	}

	if (error)
		set_bit(__EE_WAS_ERROR, &e->flags);

	bio_put(bio); /* no need for the bio anymore */
	if (atomic_dec_and_test(&e->pending_bios)) {
		if (is_write)
			drbd_endio_write_sec_final(e);
		else
			drbd_endio_read_sec_final(e);
	}
}

/* read, readA or write requests on R_PRIMARY coming from drbd_make_request
 */
void drbd_endio_pri(struct bio *bio, int error)
{
	struct drbd_request *req = bio->bi_private;
	struct drbd_conf *mdev = req->mdev;
	enum drbd_req_event what;
	int uptodate = bio_flagged(bio, BIO_UPTODATE);

	if (!error && !uptodate) {
		dev_warn(DEV, "p %s: setting error to -EIO\n",
			 bio_data_dir(bio) == WRITE ? "write" : "read");
		/* strange behavior of some lower level drivers...
		 * fail the request by clearing the uptodate flag,
		 * but do not return any error?! */
		error = -EIO;
	}

	/* to avoid recursion in __req_mod */
	if (unlikely(error)) {
		what = (bio_data_dir(bio) == WRITE)
			? write_completed_with_error
			: (bio_rw(bio) == READ)
			  ? read_completed_with_error
			  : read_ahead_completed_with_error;
	} else
		what = completed_ok;

	bio_put(req->private_bio);
	req->private_bio = ERR_PTR(error);

	req_mod(req, what);
}

int w_read_retry_remote(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
{
	struct drbd_request *req = container_of(w, struct drbd_request, w);

	/* We should not detach for read io-error,
	 * but try to WRITE the P_DATA_REPLY to the failed location,
	 * to give the disk the chance to relocate that block */

	spin_lock_irq(&mdev->req_lock);
	if (cancel || mdev->state.pdsk != D_UP_TO_DATE) {
		_req_mod(req, read_retry_remote_canceled);
		spin_unlock_irq(&mdev->req_lock);
		return 1;
	}
	spin_unlock_irq(&mdev->req_lock);

	return w_send_read_req(mdev, w, 0);
}

int w_resync_inactive(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
{
	ERR_IF(cancel) return 1;
	dev_err(DEV, "resync inactive, but callback triggered??\n");
	return 1; /* Simply ignore this! */
}

void drbd_csum_ee(struct drbd_conf *mdev, struct crypto_hash *tfm, struct drbd_epoch_entry *e, void *digest)
{
	struct hash_desc desc;
	struct scatterlist sg;
	struct page *page = e->pages;
	struct page *tmp;
	unsigned len;

	desc.tfm = tfm;
	desc.flags = 0;

	sg_init_table(&sg, 1);
	crypto_hash_init(&desc);

	while ((tmp = page_chain_next(page))) {
		/* all but the last page will be fully used */
		sg_set_page(&sg, page, PAGE_SIZE, 0);
		crypto_hash_update(&desc, &sg, sg.length);
		page = tmp;
	}
	/* and now the last, possibly only partially used page */
	len = e->size & (PAGE_SIZE - 1);
	sg_set_page(&sg, page, len ?: PAGE_SIZE, 0);
	crypto_hash_update(&desc, &sg, sg.length);
	crypto_hash_final(&desc, digest);
}

void drbd_csum_bio(struct drbd_conf *mdev, struct crypto_hash *tfm, struct bio *bio, void *digest)
{
	struct hash_desc desc;
	struct scatterlist sg;
	struct bio_vec *bvec;
	int i;

	desc.tfm = tfm;
	desc.flags = 0;

	sg_init_table(&sg, 1);
	crypto_hash_init(&desc);

	__bio_for_each_segment(bvec, bio, i, 0) {
		sg_set_page(&sg, bvec->bv_page, bvec->bv_len, bvec->bv_offset);
		crypto_hash_update(&desc, &sg, sg.length);
	}
	crypto_hash_final(&desc, digest);
}

static int w_e_send_csum(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
{
	struct drbd_epoch_entry *e = container_of(w, struct drbd_epoch_entry, w);
	int digest_size;
	void *digest;
	int ok;

	D_ASSERT(e->block_id == DRBD_MAGIC + 0xbeef);

	if (unlikely(cancel)) {
		drbd_free_ee(mdev, e);
		return 1;
	}

	if (likely((e->flags & EE_WAS_ERROR) == 0)) {
		digest_size = crypto_hash_digestsize(mdev->csums_tfm);
		digest = kmalloc(digest_size, GFP_NOIO);
		if (digest) {
			drbd_csum_ee(mdev, mdev->csums_tfm, e, digest);

			inc_rs_pending(mdev);
			ok = drbd_send_drequest_csum(mdev,
						     e->sector,
						     e->size,
						     digest,
						     digest_size,
						     P_CSUM_RS_REQUEST);
			kfree(digest);
		} else {
			dev_err(DEV, "kmalloc() of digest failed.\n");
			ok = 0;
		}
	} else
		ok = 1;

	drbd_free_ee(mdev, e);

	if (unlikely(!ok))
		dev_err(DEV, "drbd_send_drequest(..., csum) failed\n");
	return ok;
}

#define GFP_TRY	(__GFP_HIGHMEM | __GFP_NOWARN)

static int read_for_csum(struct drbd_conf *mdev, sector_t sector, int size)
{
	struct drbd_epoch_entry *e;

	if (!get_ldev(mdev))
		return -EIO;

	if (drbd_rs_should_slow_down(mdev))
		goto defer;

	/* GFP_TRY, because if there is no memory available right now, this may
	 * be rescheduled for later. It is "only" background resync, after all. */
	e = drbd_alloc_ee(mdev, DRBD_MAGIC+0xbeef, sector, size, GFP_TRY);
	if (!e)
		goto defer;

	e->w.cb = w_e_send_csum;
	spin_lock_irq(&mdev->req_lock);
	list_add(&e->w.list, &mdev->read_ee);
	spin_unlock_irq(&mdev->req_lock);

	atomic_add(size >> 9, &mdev->rs_sect_ev);
	if (drbd_submit_ee(mdev, e, READ, DRBD_FAULT_RS_RD) == 0)
		return 0;

	/* 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);
defer:
	put_ldev(mdev);
	return -EAGAIN;
}

void resync_timer_fn(unsigned long data)
{
	struct drbd_conf *mdev = (struct drbd_conf *) data;
	int queue;

	queue = 1;
	switch (mdev->state.conn) {
	case C_VERIFY_S:
		mdev->resync_work.cb = w_make_ov_request;
		break;
	case C_SYNC_TARGET:
		mdev->resync_work.cb = w_make_resync_request;
		break;
	default:
		queue = 0;
		mdev->resync_work.cb = w_resync_inactive;
	}

	/* harmless race: list_empty outside data.work.q_lock */
	if (list_empty(&mdev->resync_work.list) && queue)
		drbd_queue_work(&mdev->data.work, &mdev->resync_work);
}

static void fifo_set(struct fifo_buffer *fb, int value)
{
	int i;

	for (i = 0; i < fb->size; i++)
		fb->values[i] = value;
}

static int fifo_push(struct fifo_buffer *fb, int value)
{
	int ov;

	ov = fb->values[fb->head_index];
	fb->values[fb->head_index++] = value;

	if (fb->head_index >= fb->size)
		fb->head_index = 0;

	return ov;
}

static void fifo_add_val(struct fifo_buffer *fb, int value)
{
	int i;

	for (i = 0; i < fb->size; i++)
		fb->values[i] += value;
}

int drbd_rs_controller(struct drbd_conf *mdev)
{
	unsigned int sect_in;  /* Number of sectors that came in since the last turn */
	unsigned int want;     /* The number of sectors we want in the proxy */
	int req_sect; /* Number of sectors to request in this turn */
	int correction; /* Number of sectors more we need in the proxy*/
	int cps; /* correction per invocation of drbd_rs_controller() */
	int steps; /* Number of time steps to plan ahead */
	int curr_corr;
	int max_sect;

	sect_in = atomic_xchg(&mdev->rs_sect_in, 0); /* Number of sectors that came in */
	mdev->rs_in_flight -= sect_in;

	spin_lock(&mdev->peer_seq_lock); /* get an atomic view on mdev->rs_plan_s */

	steps = mdev->rs_plan_s.size; /* (mdev->sync_conf.c_plan_ahead * 10 * SLEEP_TIME) / HZ; */

	if (mdev->rs_in_flight + sect_in == 0) { /* At start of resync */
		want = ((mdev->sync_conf.rate * 2 * SLEEP_TIME) / HZ) * steps;
	} else { /* normal path */
		want = mdev->sync_conf.c_fill_target ? mdev->sync_conf.c_fill_target :
			sect_in * mdev->sync_conf.c_delay_target * HZ / (SLEEP_TIME * 10);
	}

	correction = want - mdev->rs_in_flight - mdev->rs_planed;

	/* Plan ahead */
	cps = correction / steps;
	fifo_add_val(&mdev->rs_plan_s, cps);
	mdev->rs_planed += cps * steps;

	/* What we do in this step */
	curr_corr = fifo_push(&mdev->rs_plan_s, 0);
	spin_unlock(&mdev->peer_seq_lock);
	mdev->rs_planed -= curr_corr;

	req_sect = sect_in + curr_corr;
	if (req_sect < 0)
		req_sect = 0;

	max_sect = (mdev->sync_conf.c_max_rate * 2 * SLEEP_TIME) / HZ;
	if (req_sect > max_sect)
		req_sect = max_sect;

	/*
	dev_warn(DEV, "si=%u if=%d wa=%u co=%d st=%d cps=%d pl=%d cc=%d rs=%d\n",
		 sect_in, mdev->rs_in_flight, want, correction,
		 steps, cps, mdev->rs_planed, curr_corr, req_sect);
	*/

	return req_sect;
}

int w_make_resync_request(struct drbd_conf *mdev,
		struct drbd_work *w, int cancel)
{
	unsigned long bit;
	sector_t sector;
	const sector_t capacity = drbd_get_capacity(mdev->this_bdev);
	int max_segment_size;
	int number, rollback_i, size, pe, mx;
	int align, queued, sndbuf;
	int i = 0;

	if (unlikely(cancel))
		return 1;

	if (unlikely(mdev->state.conn < C_CONNECTED)) {
		dev_err(DEV, "Confused in w_make_resync_request()! cstate < Connected");
		return 0;
	}

	if (mdev->state.conn != C_SYNC_TARGET)
		dev_err(DEV, "%s in w_make_resync_request\n",
			drbd_conn_str(mdev->state.conn));

	if (mdev->rs_total == 0) {
		/* empty resync? */
		drbd_resync_finished(mdev);
		return 1;
	}

	if (!get_ldev(mdev)) {
		/* Since we only need to access mdev->rsync a
		   get_ldev_if_state(mdev,D_FAILED) would be sufficient, but
		   to continue resync with a broken disk makes no sense at
		   all */
		dev_err(DEV, "Disk broke down during resync!\n");
		mdev->resync_work.cb = w_resync_inactive;
		return 1;
	}

	/* starting with drbd 8.3.8, we can handle multi-bio EEs,
	 * if it should be necessary */
	max_segment_size =
		mdev->agreed_pro_version < 94 ? queue_max_segment_size(mdev->rq_queue) :
		mdev->agreed_pro_version < 95 ?	DRBD_MAX_SIZE_H80_PACKET : DRBD_MAX_SEGMENT_SIZE;

	if (mdev->rs_plan_s.size) { /* mdev->sync_conf.c_plan_ahead */
		number = drbd_rs_controller(mdev) >> (BM_BLOCK_SHIFT - 9);
		mdev->c_sync_rate = number * HZ * (BM_BLOCK_SIZE / 1024) / SLEEP_TIME;
	} else {
		mdev->c_sync_rate = mdev->sync_conf.rate;
		number = SLEEP_TIME * mdev->c_sync_rate  / ((BM_BLOCK_SIZE / 1024) * HZ);
	}

	/* Throttle resync on lower level disk activity, which may also be
	 * caused by application IO on Primary/SyncTarget.
	 * Keep this after the call to drbd_rs_controller, as that assumes
	 * to be called as precisely as possible every SLEEP_TIME,
	 * and would be confused otherwise. */
	if (drbd_rs_should_slow_down(mdev))
		goto requeue;

	mutex_lock(&mdev->data.mutex);
	if (mdev->data.socket)
		mx = mdev->data.socket->sk->sk_rcvbuf / sizeof(struct p_block_req);
	else
		mx = 1;
	mutex_unlock(&mdev->data.mutex);

	/* For resync rates >160MB/sec, allow more pending RS requests */
	if (number > mx)
		mx = number;

	/* Limit the number of pending RS requests to no more than the peer's receive buffer */
	pe = atomic_read(&mdev->rs_pending_cnt);
	if ((pe + number) > mx) {
		number = mx - pe;
	}

	for (i = 0; i < number; i++) {
		/* Stop generating RS requests, when half of the send buffer is filled */
		mutex_lock(&mdev->data.mutex);
		if (mdev->data.socket) {
			queued = mdev->data.socket->sk->sk_wmem_queued;
			sndbuf = mdev->data.socket->sk->sk_sndbuf;
		} else {
			queued = 1;
			sndbuf = 0;
		}
		mutex_unlock(&mdev->data.mutex);
		if (queued > sndbuf / 2)
			goto requeue;

next_sector:
		size = BM_BLOCK_SIZE;
		bit  = drbd_bm_find_next(mdev, mdev->bm_resync_fo);

		if (bit == -1UL) {
			mdev->bm_resync_fo = drbd_bm_bits(mdev);
			mdev->resync_work.cb = w_resync_inactive;
			put_ldev(mdev);
			return 1;
		}

		sector = BM_BIT_TO_SECT(bit);

		if (drbd_try_rs_begin_io(mdev, sector)) {
			mdev->bm_resync_fo = bit;
			goto requeue;
		}
		mdev->bm_resync_fo = bit + 1;

		if (unlikely(drbd_bm_test_bit(mdev, bit) == 0)) {
			drbd_rs_complete_io(mdev, sector);
			goto next_sector;
		}

#if DRBD_MAX_SEGMENT_SIZE > BM_BLOCK_SIZE
		/* try to find some adjacent bits.
		 * we stop if we have already the maximum req size.
		 *
		 * Additionally always align bigger requests, in order to
		 * be prepared for all stripe sizes of software RAIDs.
		 */
		align = 1;
		rollback_i = i;
		for (;;) {
			if (size + BM_BLOCK_SIZE > max_segment_size)
				break;

			/* Be always aligned */
			if (sector & ((1<<(align+3))-1))
				break;

			/* do not cross extent boundaries */
			if (((bit+1) & BM_BLOCKS_PER_BM_EXT_MASK) == 0)
				break;
			/* now, is it actually dirty, after all?
			 * caution, drbd_bm_test_bit is tri-state for some
			 * obscure reason; ( b == 0 ) would get the out-of-band
			 * only accidentally right because of the "oddly sized"
			 * adjustment below */
			if (drbd_bm_test_bit(mdev, bit+1) != 1)
				break;
			bit++;
			size += BM_BLOCK_SIZE;
			if ((BM_BLOCK_SIZE << align) <= size)
				align++;
			i++;
		}
		/* if we merged some,
		 * reset the offset to start the next drbd_bm_find_next from */
		if (size > BM_BLOCK_SIZE)
			mdev->bm_resync_fo = bit + 1;
#endif

		/* adjust very last sectors, in case we are oddly sized */
		if (sector + (size>>9) > capacity)
			size = (capacity-sector)<<9;
		if (mdev->agreed_pro_version >= 89 && mdev->csums_tfm) {
			switch (read_for_csum(mdev, sector, size)) {
			case -EIO: /* Disk failure */
				put_ldev(mdev);
				return 0;
			case -EAGAIN: /* allocation failed, or ldev busy */
				drbd_rs_complete_io(mdev, sector);
				mdev->bm_resync_fo = BM_SECT_TO_BIT(sector);
				i = rollback_i;
				goto requeue;
			case 0:
				/* everything ok */
				break;
			default:
				BUG();
			}
		} else {
			inc_rs_pending(mdev);
			if (!drbd_send_drequest(mdev, P_RS_DATA_REQUEST,
					       sector, size, ID_SYNCER)) {
				dev_err(DEV, "drbd_send_drequest() failed, aborting...\n");
				dec_rs_pending(mdev);
				put_ldev(mdev);
				return 0;
			}
		}
	}

	if (mdev->bm_resync_fo >= drbd_bm_bits(mdev)) {
		/* last syncer _request_ was sent,
		 * but the P_RS_DATA_REPLY not yet received.  sync will end (and
		 * next sync group will resume), as soon as we receive the last
		 * resync data block, and the last bit is cleared.
		 * until then resync "work" is "inactive" ...
		 */
		mdev->resync_work.cb = w_resync_inactive;
		put_ldev(mdev);
		return 1;
	}

 requeue:
	mdev->rs_in_flight += (i << (BM_BLOCK_SHIFT - 9));
	mod_timer(&mdev->resync_timer, jiffies + SLEEP_TIME);
	put_ldev(mdev);
	return 1;
}

static int w_make_ov_request(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
{
	int number, i, size;
	sector_t sector;
	const sector_t capacity = drbd_get_capacity(mdev->this_bdev);

	if (unlikely(cancel))
		return 1;

	if (unlikely(mdev->state.conn < C_CONNECTED)) {
		dev_err(DEV, "Confused in w_make_ov_request()! cstate < Connected");
		return 0;
	}

	number = SLEEP_TIME*mdev->sync_conf.rate / ((BM_BLOCK_SIZE/1024)*HZ);
	if (atomic_read(&mdev->rs_pending_cnt) > number)
		goto requeue;

	number -= atomic_read(&mdev->rs_pending_cnt);

	sector = mdev->ov_position;
	for (i = 0; i < number; i++) {
		if (sector >= capacity) {
			mdev->resync_work.cb = w_resync_inactive;
			return 1;
		}

		size = BM_BLOCK_SIZE;

		if (drbd_try_rs_begin_io(mdev, sector)) {
			mdev->ov_position = sector;
			goto requeue;
		}

		if (sector + (size>>9) > capacity)
			size = (capacity-sector)<<9;

		inc_rs_pending(mdev);
		if (!drbd_send_ov_request(mdev, sector, size)) {
			dec_rs_pending(mdev);
			return 0;
		}
		sector += BM_SECT_PER_BIT;
	}
	mdev->ov_position = sector;

 requeue:
	mod_timer(&mdev->resync_timer, jiffies + SLEEP_TIME);
	return 1;
}


int w_ov_finished(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
{
	kfree(w);
	ov_oos_print(mdev);
	drbd_resync_finished(mdev);

	return 1;
}

static int w_resync_finished(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
{
	kfree(w);

	drbd_resync_finished(mdev);

	return 1;
}

static void ping_peer(struct drbd_conf *mdev)
{
	clear_bit(GOT_PING_ACK, &mdev->flags);
	request_ping(mdev);
	wait_event(mdev->misc_wait,
		   test_bit(GOT_PING_ACK, &mdev->flags) || mdev->state.conn < C_CONNECTED);
}

int drbd_resync_finished(struct drbd_conf *mdev)
{
	unsigned long db, dt, dbdt;
	unsigned long n_oos;
	union drbd_state os, ns;
	struct drbd_work *w;
	char *khelper_cmd = NULL;

	/* Remove all elements from the resync LRU. Since future actions
	 * might set bits in the (main) bitmap, then the entries in the
	 * resync LRU would be wrong. */
	if (drbd_rs_del_all(mdev)) {
		/* In case this is not possible now, most probably because
		 * there are P_RS_DATA_REPLY Packets lingering on the worker's
		 * queue (or even the read operations for those packets
		 * is not finished by now).   Retry in 100ms. */

		drbd_kick_lo(mdev);
		__set_current_state(TASK_INTERRUPTIBLE);
		schedule_timeout(HZ / 10);
		w = kmalloc(sizeof(struct drbd_work), GFP_ATOMIC);
		if (w) {
			w->cb = w_resync_finished;
			drbd_queue_work(&mdev->data.work, w);
			return 1;
		}
		dev_err(DEV, "Warn failed to drbd_rs_del_all() and to kmalloc(w).\n");
	}

	dt = (jiffies - mdev->rs_start - mdev->rs_paused) / HZ;
	if (dt <= 0)
		dt = 1;
	db = mdev->rs_total;
	dbdt = Bit2KB(db/dt);
	mdev->rs_paused /= HZ;

	if (!get_ldev(mdev))
		goto out;

	ping_peer(mdev);

	spin_lock_irq(&mdev->req_lock);
	os = mdev->state;

	/* This protects us against multiple calls (that can happen in the presence
	   of application IO), and against connectivity loss just before we arrive here. */
	if (os.conn <= C_CONNECTED)
		goto out_unlock;

	ns = os;
	ns.conn = C_CONNECTED;

	dev_info(DEV, "%s done (total %lu sec; paused %lu sec; %lu K/sec)\n",
	     (os.conn == C_VERIFY_S || os.conn == C_VERIFY_T) ?
	     "Online verify " : "Resync",
	     dt + mdev->rs_paused, mdev->rs_paused, dbdt);

	n_oos = drbd_bm_total_weight(mdev);

	if (os.conn == C_VERIFY_S || os.conn == C_VERIFY_T) {
		if (n_oos) {
			dev_alert(DEV, "Online verify found %lu %dk block out of sync!\n",
			      n_oos, Bit2KB(1));
			khelper_cmd = "out-of-sync";
		}
	} else {
		D_ASSERT((n_oos - mdev->rs_failed) == 0);

		if (os.conn == C_SYNC_TARGET || os.conn == C_PAUSED_SYNC_T)
			khelper_cmd = "after-resync-target";

		if (mdev->csums_tfm && mdev->rs_total) {
			const unsigned long s = mdev->rs_same_csum;
			const unsigned long t = mdev->rs_total;
			const int ratio =
				(t == 0)     ? 0 :
			(t < 100000) ? ((s*100)/t) : (s/(t/100));
			dev_info(DEV, "%u %% had equal check sums, eliminated: %luK; "
			     "transferred %luK total %luK\n",
			     ratio,
			     Bit2KB(mdev->rs_same_csum),
			     Bit2KB(mdev->rs_total - mdev->rs_same_csum),
			     Bit2KB(mdev->rs_total));
		}
	}

	if (mdev->rs_failed) {
		dev_info(DEV, "            %lu failed blocks\n", mdev->rs_failed);

		if (os.conn == C_SYNC_TARGET || os.conn == C_PAUSED_SYNC_T) {
			ns.disk = D_INCONSISTENT;
			ns.pdsk = D_UP_TO_DATE;
		} else {
			ns.disk = D_UP_TO_DATE;
			ns.pdsk = D_INCONSISTENT;
		}
	} else {
		ns.disk = D_UP_TO_DATE;
		ns.pdsk = D_UP_TO_DATE;

		if (os.conn == C_SYNC_TARGET || os.conn == C_PAUSED_SYNC_T) {
			if (mdev->p_uuid) {
				int i;
				for (i = UI_BITMAP ; i <= UI_HISTORY_END ; i++)
					_drbd_uuid_set(mdev, i, mdev->p_uuid[i]);
				drbd_uuid_set(mdev, UI_BITMAP, mdev->ldev->md.uuid[UI_CURRENT]);
				_drbd_uuid_set(mdev, UI_CURRENT, mdev->p_uuid[UI_CURRENT]);
			} else {
				dev_err(DEV, "mdev->p_uuid is NULL! BUG\n");
			}
		}

		drbd_uuid_set_bm(mdev, 0UL);

		if (mdev->p_uuid) {
			/* Now the two UUID sets are equal, update what we
			 * know of the peer. */
			int i;
			for (i = UI_CURRENT ; i <= UI_HISTORY_END ; i++)
				mdev->p_uuid[i] = mdev->ldev->md.uuid[i];
		}
	}

	_drbd_set_state(mdev, ns, CS_VERBOSE, NULL);
out_unlock:
	spin_unlock_irq(&mdev->req_lock);
	put_ldev(mdev);
out:
	mdev->rs_total  = 0;
	mdev->rs_failed = 0;
	mdev->rs_paused = 0;
	mdev->ov_start_sector = 0;

	if (test_and_clear_bit(WRITE_BM_AFTER_RESYNC, &mdev->flags)) {
		dev_warn(DEV, "Writing the whole bitmap, due to failed kmalloc\n");
		drbd_queue_bitmap_io(mdev, &drbd_bm_write, NULL, "write from resync_finished");
	}

	if (khelper_cmd)
		drbd_khelper(mdev, khelper_cmd);

	return 1;
}

/* helper */
static void move_to_net_ee_or_free(struct drbd_conf *mdev, struct drbd_epoch_entry *e)
{
	if (drbd_ee_has_active_page(e)) {
		/* This might happen if sendpage() has not finished */
		int i = (e->size + PAGE_SIZE -1) >> PAGE_SHIFT;
		atomic_add(i, &mdev->pp_in_use_by_net);
		atomic_sub(i, &mdev->pp_in_use);
		spin_lock_irq(&mdev->req_lock);
		list_add_tail(&e->w.list, &mdev->net_ee);
		spin_unlock_irq(&mdev->req_lock);
		wake_up(&drbd_pp_wait);
	} else
		drbd_free_ee(mdev, e);
}

/**
 * w_e_end_data_req() - Worker callback, to send a P_DATA_REPLY packet in response to a P_DATA_REQUEST
 * @mdev:	DRBD device.
 * @w:		work object.
 * @cancel:	The connection will be closed anyways
 */
int w_e_end_data_req(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
{
	struct drbd_epoch_entry *e = container_of(w, struct drbd_epoch_entry, w);
	int ok;

	if (unlikely(cancel)) {
		drbd_free_ee(mdev, e);
		dec_unacked(mdev);
		return 1;
	}

	if (likely((e->flags & EE_WAS_ERROR) == 0)) {
		ok = drbd_send_block(mdev, P_DATA_REPLY, e);
	} else {
		if (__ratelimit(&drbd_ratelimit_state))
			dev_err(DEV, "Sending NegDReply. sector=%llus.\n",
			    (unsigned long long)e->sector);

		ok = drbd_send_ack(mdev, P_NEG_DREPLY, e);
	}

	dec_unacked(mdev);

	move_to_net_ee_or_free(mdev, e);

	if (unlikely(!ok))
		dev_err(DEV, "drbd_send_block() failed\n");
	return ok;
}

/**
 * w_e_end_rsdata_req() - Worker callback to send a P_RS_DATA_REPLY packet in response to a P_RS_DATA_REQUESTRS
 * @mdev:	DRBD device.
 * @w:		work object.
 * @cancel:	The connection will be closed anyways
 */
int w_e_end_rsdata_req(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
{
	struct drbd_epoch_entry *e = container_of(w, struct drbd_epoch_entry, w);
	int ok;

	if (unlikely(cancel)) {
		drbd_free_ee(mdev, e);
		dec_unacked(mdev);
		return 1;
	}

	if (get_ldev_if_state(mdev, D_FAILED)) {
		drbd_rs_complete_io(mdev, e->sector);
		put_ldev(mdev);
	}

	if (likely((e->flags & EE_WAS_ERROR) == 0)) {
		if (likely(mdev->state.pdsk >= D_INCONSISTENT)) {
			inc_rs_pending(mdev);
			ok = drbd_send_block(mdev, P_RS_DATA_REPLY, e);
		} else {
			if (__ratelimit(&drbd_ratelimit_state))
				dev_err(DEV, "Not sending RSDataReply, "
				    "partner DISKLESS!\n");
			ok = 1;
		}
	} else {
		if (__ratelimit(&drbd_ratelimit_state))
			dev_err(DEV, "Sending NegRSDReply. sector %llus.\n",
			    (unsigned long long)e->sector);

		ok = drbd_send_ack(mdev, P_NEG_RS_DREPLY, e);

		/* update resync data with failure */
		drbd_rs_failed_io(mdev, e->sector, e->size);
	}

	dec_unacked(mdev);

	move_to_net_ee_or_free(mdev, e);

	if (unlikely(!ok))
		dev_err(DEV, "drbd_send_block() failed\n");
	return ok;
}

int w_e_end_csum_rs_req(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
{
	struct drbd_epoch_entry *e = container_of(w, struct drbd_epoch_entry, w);
	struct digest_info *di;
	int digest_size;
	void *digest = NULL;
	int ok, eq = 0;

	if (unlikely(cancel)) {
		drbd_free_ee(mdev, e);
		dec_unacked(mdev);
		return 1;
	}

	if (get_ldev(mdev)) {
		drbd_rs_complete_io(mdev, e->sector);
		put_ldev(mdev);
	}

	di = e->digest;

	if (likely((e->flags & EE_WAS_ERROR) == 0)) {
		/* quick hack to try to avoid a race against reconfiguration.
		 * a real fix would be much more involved,
		 * introducing more locking mechanisms */
		if (mdev->csums_tfm) {
			digest_size = crypto_hash_digestsize(mdev->csums_tfm);
			D_ASSERT(digest_size == di->digest_size);
			digest = kmalloc(digest_size, GFP_NOIO);
		}
		if (digest) {
			drbd_csum_ee(mdev, mdev->csums_tfm, e, digest);
			eq = !memcmp(digest, di->digest, digest_size);
			kfree(digest);
		}

		if (eq) {
			drbd_set_in_sync(mdev, e->sector, e->size);
			/* rs_same_csums unit is BM_BLOCK_SIZE */
			mdev->rs_same_csum += e->size >> BM_BLOCK_SHIFT;
			ok = drbd_send_ack(mdev, P_RS_IS_IN_SYNC, e);
		} else {
			inc_rs_pending(mdev);
			e->block_id = ID_SYNCER; /* By setting block_id, digest pointer becomes invalid! */
			e->flags &= ~EE_HAS_DIGEST; /* This e no longer has a digest pointer */
			kfree(di);
			ok = drbd_send_block(mdev, P_RS_DATA_REPLY, e);
		}
	} else {
		ok = drbd_send_ack(mdev, P_NEG_RS_DREPLY, e);
		if (__ratelimit(&drbd_ratelimit_state))
			dev_err(DEV, "Sending NegDReply. I guess it gets messy.\n");
	}

	dec_unacked(mdev);
	move_to_net_ee_or_free(mdev, e);

	if (unlikely(!ok))
		dev_err(DEV, "drbd_send_block/ack() failed\n");
	return ok;
}

int w_e_end_ov_req(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
{
	struct drbd_epoch_entry *e = container_of(w, struct drbd_epoch_entry, w);
	int digest_size;
	void *digest;
	int ok = 1;

	if (unlikely(cancel))
		goto out;

	if (unlikely((e->flags & EE_WAS_ERROR) != 0))
		goto out;

	digest_size = crypto_hash_digestsize(mdev->verify_tfm);
	/* FIXME if this allocation fails, online verify will not terminate! */
	digest = kmalloc(digest_size, GFP_NOIO);
	if (digest) {
		drbd_csum_ee(mdev, mdev->verify_tfm, e, digest);
		inc_rs_pending(mdev);
		ok = drbd_send_drequest_csum(mdev, e->sector, e->size,
					     digest, digest_size, P_OV_REPLY);
		if (!ok)
			dec_rs_pending(mdev);
		kfree(digest);
	}

out:
	drbd_free_ee(mdev, e);

	dec_unacked(mdev);

	return ok;
}

void drbd_ov_oos_found(struct drbd_conf *mdev, sector_t sector, int size)
{
	if (mdev->ov_last_oos_start + mdev->ov_last_oos_size == sector) {
		mdev->ov_last_oos_size += size>>9;
	} else {
		mdev->ov_last_oos_start = sector;
		mdev->ov_last_oos_size = size>>9;
	}
	drbd_set_out_of_sync(mdev, sector, size);
	set_bit(WRITE_BM_AFTER_RESYNC, &mdev->flags);
}

int w_e_end_ov_reply(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
{
	struct drbd_epoch_entry *e = container_of(w, struct drbd_epoch_entry, w);
	struct digest_info *di;
	int digest_size;
	void *digest;
	int ok, eq = 0;

	if (unlikely(cancel)) {
		drbd_free_ee(mdev, e);
		dec_unacked(mdev);
		return 1;
	}

	/* after "cancel", because after drbd_disconnect/drbd_rs_cancel_all
	 * the resync lru has been cleaned up already */
	if (get_ldev(mdev)) {
		drbd_rs_complete_io(mdev, e->sector);
		put_ldev(mdev);
	}

	di = e->digest;

	if (likely((e->flags & EE_WAS_ERROR) == 0)) {
		digest_size = crypto_hash_digestsize(mdev->verify_tfm);
		digest = kmalloc(digest_size, GFP_NOIO);
		if (digest) {
			drbd_csum_ee(mdev, mdev->verify_tfm, e, digest);

			D_ASSERT(digest_size == di->digest_size);
			eq = !memcmp(digest, di->digest, digest_size);
			kfree(digest);
		}
	} else {
		ok = drbd_send_ack(mdev, P_NEG_RS_DREPLY, e);
		if (__ratelimit(&drbd_ratelimit_state))
			dev_err(DEV, "Sending NegDReply. I guess it gets messy.\n");
	}

	dec_unacked(mdev);
	if (!eq)
		drbd_ov_oos_found(mdev, e->sector, e->size);
	else
		ov_oos_print(mdev);

	ok = drbd_send_ack_ex(mdev, P_OV_RESULT, e->sector, e->size,
			      eq ? ID_IN_SYNC : ID_OUT_OF_SYNC);

	drbd_free_ee(mdev, e);

	if (--mdev->ov_left == 0) {
		ov_oos_print(mdev);
		drbd_resync_finished(mdev);
	}

	return ok;
}

int w_prev_work_done(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
{
	struct drbd_wq_barrier *b = container_of(w, struct drbd_wq_barrier, w);
	complete(&b->done);
	return 1;
}

int w_send_barrier(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
{
	struct drbd_tl_epoch *b = container_of(w, struct drbd_tl_epoch, w);
	struct p_barrier *p = &mdev->data.sbuf.barrier;
	int ok = 1;

	/* really avoid racing with tl_clear.  w.cb may have been referenced
	 * just before it was reassigned and re-queued, so double check that.
	 * actually, this race was harmless, since we only try to send the
	 * barrier packet here, and otherwise do nothing with the object.
	 * but compare with the head of w_clear_epoch */
	spin_lock_irq(&mdev->req_lock);
	if (w->cb != w_send_barrier || mdev->state.conn < C_CONNECTED)
		cancel = 1;
	spin_unlock_irq(&mdev->req_lock);
	if (cancel)
		return 1;

	if (!drbd_get_data_sock(mdev))
		return 0;
	p->barrier = b->br_number;
	/* inc_ap_pending was done where this was queued.
	 * dec_ap_pending will be done in got_BarrierAck
	 * or (on connection loss) in w_clear_epoch.  */
	ok = _drbd_send_cmd(mdev, mdev->data.socket, P_BARRIER,
				(struct p_header80 *)p, sizeof(*p), 0);
	drbd_put_data_sock(mdev);

	return ok;
}

int w_send_write_hint(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
{
	if (cancel)
		return 1;
	return drbd_send_short_cmd(mdev, P_UNPLUG_REMOTE);
}

/**
 * w_send_dblock() - Worker callback to send a P_DATA packet in order to mirror a write request
 * @mdev:	DRBD device.
 * @w:		work object.
 * @cancel:	The connection will be closed anyways
 */
int w_send_dblock(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_dblock(mdev, req);
	req_mod(req, ok ? handed_over_to_network : send_failed);

	return ok;
}

/**
 * w_send_read_req() - Worker callback to send a read request (P_DATA_REQUEST) packet
 * @mdev:	DRBD device.
 * @w:		work object.
 * @cancel:	The connection will be closed anyways
 */
int w_send_read_req(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_drequest(mdev, P_DATA_REQUEST, req->sector, req->size,
				(unsigned long)req);

	if (!ok) {
		/* ?? we set C_TIMEOUT or C_BROKEN_PIPE in drbd_send();
		 * so this is probably redundant */
		if (mdev->state.conn >= C_CONNECTED)
			drbd_force_state(mdev, NS(conn, C_NETWORK_FAILURE));
	}
	req_mod(req, ok ? handed_over_to_network : send_failed);

	return ok;
}

int w_restart_disk_io(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
{
	struct drbd_request *req = container_of(w, struct drbd_request, w);

	if (bio_data_dir(req->master_bio) == WRITE && req->rq_state & RQ_IN_ACT_LOG)
		drbd_al_begin_io(mdev, req->sector);
	/* Calling drbd_al_begin_io() out of the worker might deadlocks
	   theoretically. Practically it can not deadlock, since this is
	   only used when unfreezing IOs. All the extents of the requests
	   that made it into the TL are already active */

	drbd_req_make_private_bio(req, req->master_bio);
	req->private_bio->bi_bdev = mdev->ldev->backing_bdev;
	generic_make_request(req->private_bio);

	return 1;
}

static int _drbd_may_sync_now(struct drbd_conf *mdev)
{
	struct drbd_conf *odev = mdev;

	while (1) {
		if (odev->sync_conf.after == -1)
			return 1;
		odev = minor_to_mdev(odev->sync_conf.after);
		ERR_IF(!odev) return 1;
		if ((odev->state.conn >= C_SYNC_SOURCE &&
		     odev->state.conn <= C_PAUSED_SYNC_T) ||
		    odev->state.aftr_isp || odev->state.peer_isp ||
		    odev->state.user_isp)
			return 0;
	}
}

/**
 * _drbd_pause_after() - Pause resync on all devices that may not resync now
 * @mdev:	DRBD device.
 *
 * Called from process context only (admin command and after_state_ch).
 */
static int _drbd_pause_after(struct drbd_conf *mdev)
{
	struct drbd_conf *odev;
	int i, rv = 0;

	for (i = 0; i < minor_count; i++) {
		odev = minor_to_mdev(i);
		if (!odev)
			continue;
		if (odev->state.conn == C_STANDALONE && odev->state.disk == D_DISKLESS)
			continue;
		if (!_drbd_may_sync_now(odev))
			rv |= (__drbd_set_state(_NS(odev, aftr_isp, 1), CS_HARD, NULL)
			       != SS_NOTHING_TO_DO);
	}

	return rv;
}

/**
 * _drbd_resume_next() - Resume resync on all devices that may resync now
 * @mdev:	DRBD device.
 *
 * Called from process context only (admin command and worker).
 */
static int _drbd_resume_next(struct drbd_conf *mdev)
{
	struct drbd_conf *odev;
	int i, rv = 0;

	for (i = 0; i < minor_count; i++) {
		odev = minor_to_mdev(i);
		if (!odev)
			continue;
		if (odev->state.conn == C_STANDALONE && odev->state.disk == D_DISKLESS)
			continue;
		if (odev->state.aftr_isp) {
			if (_drbd_may_sync_now(odev))
				rv |= (__drbd_set_state(_NS(odev, aftr_isp, 0),
							CS_HARD, NULL)
				       != SS_NOTHING_TO_DO) ;
		}
	}
	return rv;
}

void resume_next_sg(struct drbd_conf *mdev)
{
	write_lock_irq(&global_state_lock);
	_drbd_resume_next(mdev);
	write_unlock_irq(&global_state_lock);
}

void suspend_other_sg(struct drbd_conf *mdev)
{
	write_lock_irq(&global_state_lock);
	_drbd_pause_after(mdev);
	write_unlock_irq(&global_state_lock);
}

static int sync_after_error(struct drbd_conf *mdev, int o_minor)
{
	struct drbd_conf *odev;

	if (o_minor == -1)
		return NO_ERROR;
	if (o_minor < -1 || minor_to_mdev(o_minor) == NULL)
		return ERR_SYNC_AFTER;

	/* check for loops */
	odev = minor_to_mdev(o_minor);
	while (1) {
		if (odev == mdev)
			return ERR_SYNC_AFTER_CYCLE;

		/* dependency chain ends here, no cycles. */
		if (odev->sync_conf.after == -1)
			return NO_ERROR;

		/* follow the dependency chain */
		odev = minor_to_mdev(odev->sync_conf.after);
	}
}

int drbd_alter_sa(struct drbd_conf *mdev, int na)
{
	int changes;
	int retcode;

	write_lock_irq(&global_state_lock);
	retcode = sync_after_error(mdev, na);
	if (retcode == NO_ERROR) {
		mdev->sync_conf.after = na;
		do {
			changes  = _drbd_pause_after(mdev);
			changes |= _drbd_resume_next(mdev);
		} while (changes);
	}
	write_unlock_irq(&global_state_lock);
	return retcode;
}

/**
 * drbd_start_resync() - Start the resync process
 * @mdev:	DRBD device.
 * @side:	Either C_SYNC_SOURCE or C_SYNC_TARGET
 *
 * This function might bring you directly into one of the
 * C_PAUSED_SYNC_* states.
 */
void drbd_start_resync(struct drbd_conf *mdev, enum drbd_conns side)
{
	union drbd_state ns;
	int r;

	if (mdev->state.conn >= C_SYNC_SOURCE) {
		dev_err(DEV, "Resync already running!\n");
		return;
	}

	/* In case a previous resync run was aborted by an IO error/detach on the peer. */
	drbd_rs_cancel_all(mdev);

	if (side == C_SYNC_TARGET) {
		/* Since application IO was locked out during C_WF_BITMAP_T and
		   C_WF_SYNC_UUID we are still unmodified. Before going to C_SYNC_TARGET
		   we check that we might make the data inconsistent. */
		r = drbd_khelper(mdev, "before-resync-target");
		r = (r >> 8) & 0xff;
		if (r > 0) {
			dev_info(DEV, "before-resync-target handler returned %d, "
			     "dropping connection.\n", r);
			drbd_force_state(mdev, NS(conn, C_DISCONNECTING));
			return;
		}
	}

	drbd_state_lock(mdev);

	if (!get_ldev_if_state(mdev, D_NEGOTIATING)) {
		drbd_state_unlock(mdev);
		return;
	}

	if (side == C_SYNC_TARGET) {
		mdev->bm_resync_fo = 0;
	} else /* side == C_SYNC_SOURCE */ {
		u64 uuid;

		get_random_bytes(&uuid, sizeof(u64));
		drbd_uuid_set(mdev, UI_BITMAP, uuid);
		drbd_send_sync_uuid(mdev, uuid);

		D_ASSERT(mdev->state.disk == D_UP_TO_DATE);
	}

	write_lock_irq(&global_state_lock);
	ns = mdev->state;

	ns.aftr_isp = !_drbd_may_sync_now(mdev);

	ns.conn = side;

	if (side == C_SYNC_TARGET)
		ns.disk = D_INCONSISTENT;
	else /* side == C_SYNC_SOURCE */
		ns.pdsk = D_INCONSISTENT;

	r = __drbd_set_state(mdev, ns, CS_VERBOSE, NULL);
	ns = mdev->state;

	if (ns.conn < C_CONNECTED)
		r = SS_UNKNOWN_ERROR;

	if (r == SS_SUCCESS) {
		unsigned long tw = drbd_bm_total_weight(mdev);
		unsigned long now = jiffies;
		int i;

		mdev->rs_failed    = 0;
		mdev->rs_paused    = 0;
		mdev->rs_same_csum = 0;
		mdev->rs_last_events = 0;
		mdev->rs_last_sect_ev = 0;
		mdev->rs_total     = tw;
		mdev->rs_start     = now;
		for (i = 0; i < DRBD_SYNC_MARKS; i++) {
			mdev->rs_mark_left[i] = tw;
			mdev->rs_mark_time[i] = now;
		}
		_drbd_pause_after(mdev);
	}
	write_unlock_irq(&global_state_lock);
	put_ldev(mdev);

	if (r == SS_SUCCESS) {
		dev_info(DEV, "Began resync as %s (will sync %lu KB [%lu bits set]).\n",
		     drbd_conn_str(ns.conn),
		     (unsigned long) mdev->rs_total << (BM_BLOCK_SHIFT-10),
		     (unsigned long) mdev->rs_total);

		if (mdev->agreed_pro_version < 95 && mdev->rs_total == 0) {
			/* This still has a race (about when exactly the peers
			 * detect connection loss) that can lead to a full sync
			 * on next handshake. In 8.3.9 we fixed this with explicit
			 * resync-finished notifications, but the fix
			 * introduces a protocol change.  Sleeping for some
			 * time longer than the ping interval + timeout on the
			 * SyncSource, to give the SyncTarget the chance to
			 * detect connection loss, then waiting for a ping
			 * response (implicit in drbd_resync_finished) reduces
			 * the race considerably, but does not solve it. */
			if (side == C_SYNC_SOURCE)
				schedule_timeout_interruptible(
					mdev->net_conf->ping_int * HZ +
					mdev->net_conf->ping_timeo*HZ/9);
			drbd_resync_finished(mdev);
		}

		atomic_set(&mdev->rs_sect_in, 0);
		atomic_set(&mdev->rs_sect_ev, 0);
		mdev->rs_in_flight = 0;
		mdev->rs_planed = 0;
		spin_lock(&mdev->peer_seq_lock);
		fifo_set(&mdev->rs_plan_s, 0);
		spin_unlock(&mdev->peer_seq_lock);
		/* ns.conn may already be != mdev->state.conn,
		 * we may have been paused in between, or become paused until
		 * the timer triggers.
		 * No matter, that is handled in resync_timer_fn() */
		if (ns.conn == C_SYNC_TARGET)
			mod_timer(&mdev->resync_timer, jiffies);

		drbd_md_sync(mdev);
	}
	drbd_state_unlock(mdev);
}

int drbd_worker(struct drbd_thread *thi)
{
	struct drbd_conf *mdev = thi->mdev;
	struct drbd_work *w = NULL;
	LIST_HEAD(work_list);
	int intr = 0, i;

	sprintf(current->comm, "drbd%d_worker", mdev_to_minor(mdev));

	while (get_t_state(thi) == Running) {
		drbd_thread_current_set_cpu(mdev);

		if (down_trylock(&mdev->data.work.s)) {
			mutex_lock(&mdev->data.mutex);
			if (mdev->data.socket && !mdev->net_conf->no_cork)
				drbd_tcp_uncork(mdev->data.socket);
			mutex_unlock(&mdev->data.mutex);

			intr = down_interruptible(&mdev->data.work.s);

			mutex_lock(&mdev->data.mutex);
			if (mdev->data.socket  && !mdev->net_conf->no_cork)
				drbd_tcp_cork(mdev->data.socket);
			mutex_unlock(&mdev->data.mutex);
		}

		if (intr) {
			D_ASSERT(intr == -EINTR);
			flush_signals(current);
			ERR_IF (get_t_state(thi) == Running)
				continue;
			break;
		}

		if (get_t_state(thi) != Running)
			break;
		/* With this break, we have done a down() but not consumed
		   the entry from the list. The cleanup code takes care of
		   this...   */

		w = NULL;
		spin_lock_irq(&mdev->data.work.q_lock);
		ERR_IF(list_empty(&mdev->data.work.q)) {
			/* something terribly wrong in our logic.
			 * we were able to down() the semaphore,
			 * but the list is empty... doh.
			 *
			 * what is the best thing to do now?
			 * try again from scratch, restarting the receiver,
			 * asender, whatnot? could break even more ugly,
			 * e.g. when we are primary, but no good local data.
			 *
			 * I'll try to get away just starting over this loop.
			 */
			spin_unlock_irq(&mdev->data.work.q_lock);
			continue;
		}
		w = list_entry(mdev->data.work.q.next, struct drbd_work, list);
		list_del_init(&w->list);
		spin_unlock_irq(&mdev->data.work.q_lock);

		if (!w->cb(mdev, w, mdev->state.conn < C_CONNECTED)) {
			/* dev_warn(DEV, "worker: a callback failed! \n"); */
			if (mdev->state.conn >= C_CONNECTED)
				drbd_force_state(mdev,
						NS(conn, C_NETWORK_FAILURE));
		}
	}
	D_ASSERT(test_bit(DEVICE_DYING, &mdev->flags));
	D_ASSERT(test_bit(CONFIG_PENDING, &mdev->flags));

	spin_lock_irq(&mdev->data.work.q_lock);
	i = 0;
	while (!list_empty(&mdev->data.work.q)) {
		list_splice_init(&mdev->data.work.q, &work_list);
		spin_unlock_irq(&mdev->data.work.q_lock);

		while (!list_empty(&work_list)) {
			w = list_entry(work_list.next, struct drbd_work, list);
			list_del_init(&w->list);
			w->cb(mdev, w, 1);
			i++; /* dead debugging code */
		}

		spin_lock_irq(&mdev->data.work.q_lock);
	}
	sema_init(&mdev->data.work.s, 0);
	/* DANGEROUS race: if someone did queue his work within the spinlock,
	 * but up() ed outside the spinlock, we could get an up() on the
	 * semaphore without corresponding list entry.
	 * So don't do that.
	 */
	spin_unlock_irq(&mdev->data.work.q_lock);

	D_ASSERT(mdev->state.disk == D_DISKLESS && mdev->state.conn == C_STANDALONE);
	/* _drbd_set_state only uses stop_nowait.
	 * wait here for the Exiting receiver. */
	drbd_thread_stop(&mdev->receiver);
	drbd_mdev_cleanup(mdev);

	dev_info(DEV, "worker terminated\n");

	clear_bit(DEVICE_DYING, &mdev->flags);
	clear_bit(CONFIG_PENDING, &mdev->flags);
	wake_up(&mdev->state_wait);

	return 0;
}