[mirror_ubuntu-artful-kernel.git] / drivers / block / drbd / drbd_req.h

/*
   drbd_req.h

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

   Copyright (C) 2006-2008, LINBIT Information Technologies GmbH.
   Copyright (C) 2006-2008, Lars Ellenberg <lars.ellenberg@linbit.com>.
   Copyright (C) 2006-2008, Philipp Reisner <philipp.reisner@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.
 */

#ifndef _DRBD_REQ_H
#define _DRBD_REQ_H

#include <linux/module.h>

#include <linux/slab.h>
#include <linux/drbd.h>
#include "drbd_int.h"
#include "drbd_wrappers.h"

/* The request callbacks will be called in irq context by the IDE drivers,
   and in Softirqs/Tasklets/BH context by the SCSI drivers,
   and by the receiver and worker in kernel-thread context.
   Try to get the locking right :) */

/*
 * Objects of type struct drbd_request do only exist on a R_PRIMARY node, and are
 * associated with IO requests originating from the block layer above us.
 *
 * There are quite a few things that may happen to a drbd request
 * during its lifetime.
 *
 *  It will be created.
 *  It will be marked with the intention to be
 *    submitted to local disk and/or
 *    send via the network.
 *
 *  It has to be placed on the transfer log and other housekeeping lists,
 *  In case we have a network connection.
 *
 *  It may be identified as a concurrent (write) request
 *    and be handled accordingly.
 *
 *  It may me handed over to the local disk subsystem.
 *  It may be completed by the local disk subsystem,
 *    either successfully or with io-error.
 *  In case it is a READ request, and it failed locally,
 *    it may be retried remotely.
 *
 *  It may be queued for sending.
 *  It may be handed over to the network stack,
 *    which may fail.
 *  It may be acknowledged by the "peer" according to the wire_protocol in use.
 *    this may be a negative ack.
 *  It may receive a faked ack when the network connection is lost and the
 *  transfer log is cleaned up.
 *  Sending may be canceled due to network connection loss.
 *  When it finally has outlived its time,
 *    corresponding dirty bits in the resync-bitmap may be cleared or set,
 *    it will be destroyed,
 *    and completion will be signalled to the originator,
 *      with or without "success".
 */

enum drbd_req_event {
	created,
	to_be_send,
	to_be_submitted,

	/* XXX yes, now I am inconsistent...
	 * 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,
	write_acked_by_peer,
	write_acked_by_peer_and_sis, /* and set_in_sync */
	conflict_discarded_by_peer,
	neg_acked,
	barrier_acked, /* in protocol A and B */
	data_received, /* (remote read) */

	read_completed_with_error,
	read_ahead_completed_with_error,
	write_completed_with_error,
	completed_ok,
	resend,
	fail_frozen_disk_io,
	restart_frozen_disk_io,
	nothing, /* for tracing only */
};

/* encoding of request states for now.  we don't actually need that many bits.
 * we don't need to do atomic bit operations either, since most of the time we
 * need to look at the connection state and/or manipulate some lists at the
 * same time, so we should hold the request lock anyways.
 */
enum drbd_req_state_bits {
	/* 210
	 * 000: no local possible
	 * 001: to be submitted
	 *    UNUSED, we could map: 011: submitted, completion still pending
	 * 110: completed ok
	 * 010: completed with error
	 */
	__RQ_LOCAL_PENDING,
	__RQ_LOCAL_COMPLETED,
	__RQ_LOCAL_OK,

	/* 76543
	 * 00000: no network possible
	 * 00001: to be send
	 * 00011: to be send, on worker queue
	 * 00101: sent, expecting recv_ack (B) or write_ack (C)
	 * 11101: sent,
	 *        recv_ack (B) or implicit "ack" (A),
	 *        still waiting for the barrier ack.
	 *        master_bio may already be completed and invalidated.
	 * 11100: write_acked (C),
	 *        data_received (for remote read, any protocol)
	 *        or finally the barrier ack has arrived (B,A)...
	 *        request can be freed
	 * 01100: neg-acked (write, protocol C)
	 *        or neg-d-acked (read, any protocol)
	 *        or killed from the transfer log
	 *        during cleanup after connection loss
	 *        request can be freed
	 * 01000: canceled or send failed...
	 *        request can be freed
	 */

	/* if "SENT" is not set, yet, this can still fail or be canceled.
	 * if "SENT" is set already, we still wait for an Ack packet.
	 * when cleared, the master_bio may be completed.
	 * in (B,A) the request object may still linger on the transaction log
	 * until the corresponding barrier ack comes in */
	__RQ_NET_PENDING,

	/* If it is QUEUED, and it is a WRITE, it is also registered in the
	 * transfer log. Currently we need this flag to avoid conflicts between
	 * worker canceling the request and tl_clear_barrier killing it from
	 * transfer log.  We should restructure the code so this conflict does
	 * no longer occur. */
	__RQ_NET_QUEUED,

	/* well, actually only "handed over to the network stack".
	 *
	 * TODO can potentially be dropped because of the similar meaning
	 * of RQ_NET_SENT and ~RQ_NET_QUEUED.
	 * however it is not exactly the same. before we drop it
	 * we must ensure that we can tell a request with network part
	 * from a request without, regardless of what happens to it. */
	__RQ_NET_SENT,

	/* when set, the request may be freed (if RQ_NET_QUEUED is clear).
	 * basically this means the corresponding P_BARRIER_ACK was received */
	__RQ_NET_DONE,

	/* whether or not we know (C) or pretend (B,A) that the write
	 * was successfully written on the peer.
	 */
	__RQ_NET_OK,

	/* peer called drbd_set_in_sync() for this write */
	__RQ_NET_SIS,

	/* keep this last, its for the RQ_NET_MASK */
	__RQ_NET_MAX,

	/* Set when this is a write, clear for a read */
	__RQ_WRITE,

	/* Should call drbd_al_complete_io() for this request... */
	__RQ_IN_ACT_LOG,
};

#define RQ_LOCAL_PENDING   (1UL << __RQ_LOCAL_PENDING)
#define RQ_LOCAL_COMPLETED (1UL << __RQ_LOCAL_COMPLETED)
#define RQ_LOCAL_OK        (1UL << __RQ_LOCAL_OK)

#define RQ_LOCAL_MASK      ((RQ_LOCAL_OK << 1)-1) /* 0x07 */

#define RQ_NET_PENDING     (1UL << __RQ_NET_PENDING)
#define RQ_NET_QUEUED      (1UL << __RQ_NET_QUEUED)
#define RQ_NET_SENT        (1UL << __RQ_NET_SENT)
#define RQ_NET_DONE        (1UL << __RQ_NET_DONE)
#define RQ_NET_OK          (1UL << __RQ_NET_OK)
#define RQ_NET_SIS         (1UL << __RQ_NET_SIS)

/* 0x1f8 */
#define RQ_NET_MASK        (((1UL << __RQ_NET_MAX)-1) & ~RQ_LOCAL_MASK)

#define RQ_WRITE           (1UL << __RQ_WRITE)
#define RQ_IN_ACT_LOG      (1UL << __RQ_IN_ACT_LOG)

/* For waking up the frozen transfer log mod_req() has to return if the request
   should be counted in the epoch object*/
#define MR_WRITE_SHIFT 0
#define MR_WRITE       (1 << MR_WRITE_SHIFT)
#define MR_READ_SHIFT  1
#define MR_READ        (1 << MR_READ_SHIFT)

/* epoch entries */
static inline
struct hlist_head *ee_hash_slot(struct drbd_conf *mdev, sector_t sector)
{
	BUG_ON(mdev->ee_hash_s == 0);
	return mdev->ee_hash +
		((unsigned int)(sector>>HT_SHIFT) % mdev->ee_hash_s);
}

/* transfer log (drbd_request objects) */
static inline
struct hlist_head *tl_hash_slot(struct drbd_conf *mdev, sector_t sector)
{
	BUG_ON(mdev->tl_hash_s == 0);
	return mdev->tl_hash +
		((unsigned int)(sector>>HT_SHIFT) % mdev->tl_hash_s);
}

/* application reads (drbd_request objects) */
static struct hlist_head *ar_hash_slot(struct drbd_conf *mdev, sector_t sector)
{
	return mdev->app_reads_hash
		+ ((unsigned int)(sector) % APP_R_HSIZE);
}

/* when we receive the answer for a read request,
 * verify that we actually know about it */
static inline struct drbd_request *_ar_id_to_req(struct drbd_conf *mdev,
	u64 id, sector_t sector)
{
	struct hlist_head *slot = ar_hash_slot(mdev, sector);
	struct hlist_node *n;
	struct drbd_request *req;

	hlist_for_each_entry(req, n, slot, collision) {
		if ((unsigned long)req == (unsigned long)id) {
			D_ASSERT(req->sector == sector);
			return req;
		}
	}
	return NULL;
}

static inline void drbd_req_make_private_bio(struct drbd_request *req, struct bio *bio_src)
{
	struct bio *bio;
	bio = bio_clone(bio_src, GFP_NOIO); /* XXX cannot fail?? */

	req->private_bio = bio;

	bio->bi_private  = req;
	bio->bi_end_io   = drbd_endio_pri;
	bio->bi_next     = NULL;
}

static inline struct drbd_request *drbd_req_new(struct drbd_conf *mdev,
	struct bio *bio_src)
{
	struct drbd_request *req =
		mempool_alloc(drbd_request_mempool, GFP_NOIO);
	if (likely(req)) {
		drbd_req_make_private_bio(req, bio_src);

		req->rq_state    = bio_data_dir(bio_src) == WRITE ? RQ_WRITE : 0;
		req->mdev        = mdev;
		req->master_bio  = bio_src;
		req->epoch       = 0;
		req->sector      = bio_src->bi_sector;
		req->size        = bio_src->bi_size;
		INIT_HLIST_NODE(&req->collision);
		INIT_LIST_HEAD(&req->tl_requests);
		INIT_LIST_HEAD(&req->w.list);
	}
	return req;
}

static inline void drbd_req_free(struct drbd_request *req)
{
	mempool_free(req, drbd_request_mempool);
}

static inline int overlaps(sector_t s1, int l1, sector_t s2, int l2)
{
	return !((s1 + (l1>>9) <= s2) || (s1 >= s2 + (l2>>9)));
}

/* Short lived temporary struct on the stack.
 * We could squirrel the error to be returned into
 * bio->bi_size, or similar. But that would be too ugly. */
struct bio_and_error {
	struct bio *bio;
	int error;
};

extern void _req_may_be_done(struct drbd_request *req,
		struct bio_and_error *m);
extern int __req_mod(struct drbd_request *req, enum drbd_req_event what,
		struct bio_and_error *m);
extern void complete_master_bio(struct drbd_conf *mdev,
		struct bio_and_error *m);
extern void request_timer_fn(unsigned long data);

/* use this if you don't want to deal with calling complete_master_bio()
 * outside the spinlock, e.g. when walking some list on cleanup. */
static inline int _req_mod(struct drbd_request *req, enum drbd_req_event what)
{
	struct drbd_conf *mdev = req->mdev;
	struct bio_and_error m;
	int rv;

	/* __req_mod possibly frees req, do not touch req after that! */
	rv = __req_mod(req, what, &m);
	if (m.bio)
		complete_master_bio(mdev, &m);

	return rv;
}

/* completion of master bio is outside of our spinlock.
 * We still may or may not be inside some irqs disabled section
 * of the lower level driver completion callback, so we need to
 * spin_lock_irqsave here. */
static inline int req_mod(struct drbd_request *req,
		enum drbd_req_event what)
{
	unsigned long flags;
	struct drbd_conf *mdev = req->mdev;
	struct bio_and_error m;
	int rv;

	spin_lock_irqsave(&mdev->req_lock, flags);
	rv = __req_mod(req, what, &m);
	spin_unlock_irqrestore(&mdev->req_lock, flags);

	if (m.bio)
		complete_master_bio(mdev, &m);

	return rv;
}

static inline bool drbd_should_do_remote(union drbd_state s)
{
	return s.pdsk == D_UP_TO_DATE ||
		(s.pdsk >= D_INCONSISTENT &&
		 s.conn >= C_WF_BITMAP_T &&
		 s.conn < C_AHEAD);
	/* Before proto 96 that was >= CONNECTED instead of >= C_WF_BITMAP_T.
	   That is equivalent since before 96 IO was frozen in the C_WF_BITMAP*
	   states. */
}
static inline bool drbd_should_send_oos(union drbd_state s)
{
	return s.conn == C_AHEAD || s.conn == C_WF_BITMAP_S;
	/* pdsk = D_INCONSISTENT as a consequence. Protocol 96 check not necessary
	   since we enter state C_AHEAD only if proto >= 96 */
}

#endif
Commit	Line	Data
b411b363 PR	1	/*
	2	drbd_req.h
	3
	4	This file is part of DRBD by Philipp Reisner and Lars Ellenberg.
	5
	6	Copyright (C) 2006-2008, LINBIT Information Technologies GmbH.
	7	Copyright (C) 2006-2008, Lars Ellenberg <lars.ellenberg@linbit.com>.
	8	Copyright (C) 2006-2008, Philipp Reisner <philipp.reisner@linbit.com>.
	9
	10	DRBD is free software; you can redistribute it and/or modify
	11	it under the terms of the GNU General Public License as published by
	12	the Free Software Foundation; either version 2, or (at your option)
	13	any later version.
	14
	15	DRBD is distributed in the hope that it will be useful,
	16	but WITHOUT ANY WARRANTY; without even the implied warranty of
	17	MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
	18	GNU General Public License for more details.
	19
	20	You should have received a copy of the GNU General Public License
	21	along with drbd; see the file COPYING. If not, write to
	22	the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA.
	23	*/
	24
	25	#ifndef _DRBD_REQ_H
	26	#define _DRBD_REQ_H
	27
b411b363 PR	28	#include <linux/module.h>
	29
	30	#include <linux/slab.h>
	31	#include <linux/drbd.h>
	32	#include "drbd_int.h"
	33	#include "drbd_wrappers.h"
	34
	35	/* The request callbacks will be called in irq context by the IDE drivers,
	36	and in Softirqs/Tasklets/BH context by the SCSI drivers,
	37	and by the receiver and worker in kernel-thread context.
	38	Try to get the locking right :) */
	39
	40	/*
	41	* Objects of type struct drbd_request do only exist on a R_PRIMARY node, and are
	42	* associated with IO requests originating from the block layer above us.
	43	*
	44	* There are quite a few things that may happen to a drbd request
	45	* during its lifetime.
	46	*
	47	* It will be created.
	48	* It will be marked with the intention to be
	49	* submitted to local disk and/or
	50	* send via the network.
	51	*
	52	* It has to be placed on the transfer log and other housekeeping lists,
	53	* In case we have a network connection.
	54	*
	55	* It may be identified as a concurrent (write) request
	56	* and be handled accordingly.
	57	*
	58	* It may me handed over to the local disk subsystem.
	59	* It may be completed by the local disk subsystem,
3ad2f3fb	60	* either successfully or with io-error.
b411b363 PR	61	* In case it is a READ request, and it failed locally,
	62	* it may be retried remotely.
	63	*
	64	* It may be queued for sending.
	65	* It may be handed over to the network stack,
	66	* which may fail.
	67	* It may be acknowledged by the "peer" according to the wire_protocol in use.
	68	* this may be a negative ack.
	69	* It may receive a faked ack when the network connection is lost and the
	70	* transfer log is cleaned up.
	71	* Sending may be canceled due to network connection loss.
	72	* When it finally has outlived its time,
	73	* corresponding dirty bits in the resync-bitmap may be cleared or set,
	74	* it will be destroyed,
	75	* and completion will be signalled to the originator,
	76	* with or without "success".
	77	*/
	78
	79	enum drbd_req_event {
	80	created,
	81	to_be_send,
	82	to_be_submitted,
	83
	84	/* XXX yes, now I am inconsistent...
73a01a18	85	* these are not "events" but "actions"
b411b363 PR	86	* oh, well... */
	87	queue_for_net_write,
	88	queue_for_net_read,
73a01a18	89	queue_for_send_oos,
b411b363 PR	90
	91	send_canceled,
	92	send_failed,
	93	handed_over_to_network,
73a01a18	94	oos_handed_to_network,
b411b363	95	connection_lost_while_pending,
d255e5ff	96	read_retry_remote_canceled,
b411b363 PR	97	recv_acked_by_peer,
	98	write_acked_by_peer,
	99	write_acked_by_peer_and_sis, /* and set_in_sync */
	100	conflict_discarded_by_peer,
	101	neg_acked,
	102	barrier_acked, /* in protocol A and B */
	103	data_received, /* (remote read) */
	104
	105	read_completed_with_error,
	106	read_ahead_completed_with_error,
	107	write_completed_with_error,
	108	completed_ok,
11b58e73	109	resend,
265be2d0 PR	110	fail_frozen_disk_io,
265be2d0 PR	111	restart_frozen_disk_io,
b411b363 PR	112	nothing, /* for tracing only */
	113	};
	114
	115	/* encoding of request states for now. we don't actually need that many bits.
	116	* we don't need to do atomic bit operations either, since most of the time we
	117	* need to look at the connection state and/or manipulate some lists at the
	118	* same time, so we should hold the request lock anyways.
	119	*/
	120	enum drbd_req_state_bits {
	121	/* 210
	122	* 000: no local possible
	123	* 001: to be submitted
	124	* UNUSED, we could map: 011: submitted, completion still pending
	125	* 110: completed ok
	126	* 010: completed with error
	127	*/
	128	__RQ_LOCAL_PENDING,
	129	__RQ_LOCAL_COMPLETED,
	130	__RQ_LOCAL_OK,
	131
	132	/* 76543
	133	* 00000: no network possible
	134	* 00001: to be send
	135	* 00011: to be send, on worker queue
	136	* 00101: sent, expecting recv_ack (B) or write_ack (C)
	137	* 11101: sent,
	138	* recv_ack (B) or implicit "ack" (A),
	139	* still waiting for the barrier ack.
	140	* master_bio may already be completed and invalidated.
	141	* 11100: write_acked (C),
	142	* data_received (for remote read, any protocol)
	143	* or finally the barrier ack has arrived (B,A)...
	144	* request can be freed
	145	* 01100: neg-acked (write, protocol C)
	146	* or neg-d-acked (read, any protocol)
	147	* or killed from the transfer log
	148	* during cleanup after connection loss
	149	* request can be freed
	150	* 01000: canceled or send failed...
	151	* request can be freed
	152	*/
	153
	154	/* if "SENT" is not set, yet, this can still fail or be canceled.
	155	* if "SENT" is set already, we still wait for an Ack packet.
	156	* when cleared, the master_bio may be completed.
	157	* in (B,A) the request object may still linger on the transaction log
	158	* until the corresponding barrier ack comes in */
	159	__RQ_NET_PENDING,
	160
	161	/* If it is QUEUED, and it is a WRITE, it is also registered in the
	162	* transfer log. Currently we need this flag to avoid conflicts between
	163	* worker canceling the request and tl_clear_barrier killing it from
	164	* transfer log. We should restructure the code so this conflict does
	165	* no longer occur. */
	166	__RQ_NET_QUEUED,
	167
	168	/* well, actually only "handed over to the network stack".
	169	*
	170	* TODO can potentially be dropped because of the similar meaning
	171	* of RQ_NET_SENT and ~RQ_NET_QUEUED.
	172	* however it is not exactly the same. before we drop it
	173	* we must ensure that we can tell a request with network part
	174	* from a request without, regardless of what happens to it. */
	175	__RQ_NET_SENT,
176
177	/* when set, the request may be freed (if RQ_NET_QUEUED is clear).
178	* basically this means the corresponding P_BARRIER_ACK was received */
179	__RQ_NET_DONE,
180
181	/* whether or not we know (C) or pretend (B,A) that the write
182	* was successfully written on the peer.
183	*/
184	__RQ_NET_OK,
185
186	/* peer called drbd_set_in_sync() for this write */
187	__RQ_NET_SIS,
188
189	/* keep this last, its for the RQ_NET_MASK */
190	__RQ_NET_MAX,
288f422e PR	191
	192	/* Set when this is a write, clear for a read */
	193	__RQ_WRITE,
0778286a PR	194
	195	/* Should call drbd_al_complete_io() for this request... */
	196	__RQ_IN_ACT_LOG,
b411b363 PR	197	};
	198
	199	#define RQ_LOCAL_PENDING (1UL << __RQ_LOCAL_PENDING)
	200	#define RQ_LOCAL_COMPLETED (1UL << __RQ_LOCAL_COMPLETED)
	201	#define RQ_LOCAL_OK (1UL << __RQ_LOCAL_OK)
	202
	203	#define RQ_LOCAL_MASK ((RQ_LOCAL_OK << 1)-1) /* 0x07 */
	204
	205	#define RQ_NET_PENDING (1UL << __RQ_NET_PENDING)
	206	#define RQ_NET_QUEUED (1UL << __RQ_NET_QUEUED)
	207	#define RQ_NET_SENT (1UL << __RQ_NET_SENT)
	208	#define RQ_NET_DONE (1UL << __RQ_NET_DONE)
	209	#define RQ_NET_OK (1UL << __RQ_NET_OK)
	210	#define RQ_NET_SIS (1UL << __RQ_NET_SIS)
	211
	212	/* 0x1f8 */
	213	#define RQ_NET_MASK (((1UL << __RQ_NET_MAX)-1) & ~RQ_LOCAL_MASK)
	214
288f422e	215	#define RQ_WRITE (1UL << __RQ_WRITE)
0778286a	216	#define RQ_IN_ACT_LOG (1UL << __RQ_IN_ACT_LOG)
288f422e	217
11b58e73 PR	218	/* For waking up the frozen transfer log mod_req() has to return if the request
	219	should be counted in the epoch object*/
	220	#define MR_WRITE_SHIFT 0
	221	#define MR_WRITE (1 << MR_WRITE_SHIFT)
	222	#define MR_READ_SHIFT 1
	223	#define MR_READ (1 << MR_READ_SHIFT)
	224
b411b363 PR	225	/* epoch entries */
	226	static inline
	227	struct hlist_head ee_hash_slot(struct drbd_conf mdev, sector_t sector)
	228	{
	229	BUG_ON(mdev->ee_hash_s == 0);
	230	return mdev->ee_hash +
	231	((unsigned int)(sector>>HT_SHIFT) % mdev->ee_hash_s);
	232	}
	233
	234	/* transfer log (drbd_request objects) */
	235	static inline
	236	struct hlist_head tl_hash_slot(struct drbd_conf mdev, sector_t sector)
	237	{
	238	BUG_ON(mdev->tl_hash_s == 0);
	239	return mdev->tl_hash +
	240	((unsigned int)(sector>>HT_SHIFT) % mdev->tl_hash_s);
	241	}
	242
	243	/* application reads (drbd_request objects) */
	244	static struct hlist_head ar_hash_slot(struct drbd_conf mdev, sector_t sector)
	245	{
	246	return mdev->app_reads_hash
	247	+ ((unsigned int)(sector) % APP_R_HSIZE);
	248	}
	249
	250	/* when we receive the answer for a read request,
	251	* verify that we actually know about it */
	252	static inline struct drbd_request _ar_id_to_req(struct drbd_conf mdev,
	253	u64 id, sector_t sector)
	254	{
	255	struct hlist_head *slot = ar_hash_slot(mdev, sector);
	256	struct hlist_node *n;
	257	struct drbd_request *req;
	258
24c4830c	259	hlist_for_each_entry(req, n, slot, collision) {
b411b363 PR	260	if ((unsigned long)req == (unsigned long)id) {
	261	D_ASSERT(req->sector == sector);
	262	return req;
	263	}
	264	}
	265	return NULL;
	266	}
	267
5ba82308 PR	268	static inline void drbd_req_make_private_bio(struct drbd_request req, struct bio bio_src)
	269	{
	270	struct bio *bio;
	271	bio = bio_clone(bio_src, GFP_NOIO); /* XXX cannot fail?? */
	272
	273	req->private_bio = bio;
	274
	275	bio->bi_private = req;
	276	bio->bi_end_io = drbd_endio_pri;
	277	bio->bi_next = NULL;
	278	}
	279
b411b363 PR	280	static inline struct drbd_request drbd_req_new(struct drbd_conf mdev,
	281	struct bio *bio_src)
	282	{
b411b363 PR	283	struct drbd_request *req =
	284	mempool_alloc(drbd_request_mempool, GFP_NOIO);
	285	if (likely(req)) {
5ba82308	286	drbd_req_make_private_bio(req, bio_src);
b411b363	287
288f422e	288	req->rq_state = bio_data_dir(bio_src) == WRITE ? RQ_WRITE : 0;
b411b363 PR	289	req->mdev = mdev;
b411b363 PR	290	req->master_bio = bio_src;
b411b363	291	req->epoch = 0;
5ba82308 PR	292	req->sector = bio_src->bi_sector;
5ba82308 PR	293	req->size = bio_src->bi_size;
24c4830c	294	INIT_HLIST_NODE(&req->collision);
b411b363 PR	295	INIT_LIST_HEAD(&req->tl_requests);
b411b363 PR	296	INIT_LIST_HEAD(&req->w.list);
b411b363 PR	297	}
	298	return req;
	299	}
	300
	301	static inline void drbd_req_free(struct drbd_request *req)
	302	{
	303	mempool_free(req, drbd_request_mempool);
	304	}
	305
	306	static inline int overlaps(sector_t s1, int l1, sector_t s2, int l2)
	307	{
	308	return !((s1 + (l1>>9) <= s2) \|\| (s1 >= s2 + (l2>>9)));
	309	}
	310
	311	/* Short lived temporary struct on the stack.
	312	* We could squirrel the error to be returned into
	313	* bio->bi_size, or similar. But that would be too ugly. */
	314	struct bio_and_error {
	315	struct bio *bio;
	316	int error;
	317	};
	318
	319	extern void _req_may_be_done(struct drbd_request *req,
	320	struct bio_and_error *m);
2a80699f	321	extern int __req_mod(struct drbd_request *req, enum drbd_req_event what,
b411b363 PR	322	struct bio_and_error *m);
	323	extern void complete_master_bio(struct drbd_conf *mdev,
	324	struct bio_and_error *m);
7fde2be9	325	extern void request_timer_fn(unsigned long data);
b411b363 PR	326
	327	/* use this if you don't want to deal with calling complete_master_bio()
	328	* outside the spinlock, e.g. when walking some list on cleanup. */
2a80699f	329	static inline int _req_mod(struct drbd_request *req, enum drbd_req_event what)
b411b363 PR	330	{
	331	struct drbd_conf *mdev = req->mdev;
	332	struct bio_and_error m;
2a80699f	333	int rv;
b411b363 PR	334
b411b363 PR	335	/* __req_mod possibly frees req, do not touch req after that! */
2a80699f	336	rv = __req_mod(req, what, &m);
b411b363 PR	337	if (m.bio)
b411b363 PR	338	complete_master_bio(mdev, &m);
2a80699f PR	339
2a80699f PR	340	return rv;
b411b363 PR	341	}
b411b363 PR	342
759fbdfb PR	343	/* completion of master bio is outside of our spinlock.
	344	* We still may or may not be inside some irqs disabled section
	345	* of the lower level driver completion callback, so we need to
	346	* spin_lock_irqsave here. */
2a80699f	347	static inline int req_mod(struct drbd_request *req,
b411b363 PR	348	enum drbd_req_event what)
b411b363 PR	349	{
759fbdfb	350	unsigned long flags;
b411b363 PR	351	struct drbd_conf *mdev = req->mdev;
b411b363 PR	352	struct bio_and_error m;
2a80699f PR	353	int rv;
2a80699f PR	354
759fbdfb	355	spin_lock_irqsave(&mdev->req_lock, flags);
2a80699f	356	rv = __req_mod(req, what, &m);
759fbdfb	357	spin_unlock_irqrestore(&mdev->req_lock, flags);
b411b363 PR	358
	359	if (m.bio)
	360	complete_master_bio(mdev, &m);
2a80699f PR	361
2a80699f PR	362	return rv;
b411b363	363	}
6a35c45f PR	364
	365	static inline bool drbd_should_do_remote(union drbd_state s)
	366	{
	367	return s.pdsk == D_UP_TO_DATE \|\|
	368	(s.pdsk >= D_INCONSISTENT &&
	369	s.conn >= C_WF_BITMAP_T &&
	370	s.conn < C_AHEAD);
	371	/* Before proto 96 that was >= CONNECTED instead of >= C_WF_BITMAP_T.
	372	That is equivalent since before 96 IO was frozen in the C_WF_BITMAP*
	373	states. */
	374	}
	375	static inline bool drbd_should_send_oos(union drbd_state s)
	376	{
	377	return s.conn == C_AHEAD \|\| s.conn == C_WF_BITMAP_S;
	378	/* pdsk = D_INCONSISTENT as a consequence. Protocol 96 check not necessary
	379	since we enter state C_AHEAD only if proto >= 96 */
	380	}
	381
b411b363	382	#endif