drbd_req.c 35.8 KB
Newer Older
P
Philipp Reisner 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41
/*
   drbd_req.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/slab.h>
#include <linux/drbd.h>
#include "drbd_int.h"
#include "drbd_req.h"


/* Update disk stats at start of I/O request */
static void _drbd_start_io_acct(struct drbd_conf *mdev, struct drbd_request *req, struct bio *bio)
{
	const int rw = bio_data_dir(bio);
	int cpu;
	cpu = part_stat_lock();
	part_stat_inc(cpu, &mdev->vdisk->part0, ios[rw]);
	part_stat_add(cpu, &mdev->vdisk->part0, sectors[rw], bio_sectors(bio));
42
	part_inc_in_flight(&mdev->vdisk->part0, rw);
P
Philipp Reisner 已提交
43 44 45 46 47 48 49 50 51 52 53 54
	part_stat_unlock();
}

/* Update disk stats when completing request upwards */
static void _drbd_end_io_acct(struct drbd_conf *mdev, struct drbd_request *req)
{
	int rw = bio_data_dir(req->master_bio);
	unsigned long duration = jiffies - req->start_time;
	int cpu;
	cpu = part_stat_lock();
	part_stat_add(cpu, &mdev->vdisk->part0, ticks[rw], duration);
	part_round_stats(cpu, &mdev->vdisk->part0);
55
	part_dec_in_flight(&mdev->vdisk->part0, rw);
P
Philipp Reisner 已提交
56 57 58
	part_stat_unlock();
}

59 60 61 62 63 64 65 66 67 68 69
static struct drbd_request *drbd_req_new(struct drbd_conf *mdev,
					       struct bio *bio_src)
{
	struct drbd_request *req;

	req = mempool_alloc(drbd_request_mempool, GFP_NOIO);
	if (!req)
		return NULL;

	drbd_req_make_private_bio(req, bio_src);
	req->rq_state    = bio_data_dir(bio_src) == WRITE ? RQ_WRITE : 0;
70
	req->w.mdev      = mdev;
71 72
	req->master_bio  = bio_src;
	req->epoch       = 0;
73

74 75 76
	drbd_clear_interval(&req->i);
	req->i.sector     = bio_src->bi_sector;
	req->i.size      = bio_src->bi_size;
77
	req->i.local = true;
78 79
	req->i.waiting = false;

80 81 82 83 84 85 86 87 88 89 90 91
	INIT_LIST_HEAD(&req->tl_requests);
	INIT_LIST_HEAD(&req->w.list);

	return req;
}

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

/* rw is bio_data_dir(), only READ or WRITE */
P
Philipp Reisner 已提交
92 93 94
static void _req_is_done(struct drbd_conf *mdev, struct drbd_request *req, const int rw)
{
	const unsigned long s = req->rq_state;
95 96 97 98 99 100 101 102 103

	/* remove it from the transfer log.
	 * well, only if it had been there in the first
	 * place... if it had not (local only or conflicting
	 * and never sent), it should still be "empty" as
	 * initialized in drbd_req_new(), so we can list_del() it
	 * here unconditionally */
	list_del(&req->tl_requests);

P
Philipp Reisner 已提交
104 105 106 107 108 109 110 111 112
	/* if it was a write, we may have to set the corresponding
	 * bit(s) out-of-sync first. If it had a local part, we need to
	 * release the reference to the activity log. */
	if (rw == WRITE) {
		/* Set out-of-sync unless both OK flags are set
		 * (local only or remote failed).
		 * Other places where we set out-of-sync:
		 * READ with local io-error */
		if (!(s & RQ_NET_OK) || !(s & RQ_LOCAL_OK))
113
			drbd_set_out_of_sync(mdev, req->i.sector, req->i.size);
P
Philipp Reisner 已提交
114 115

		if ((s & RQ_NET_OK) && (s & RQ_LOCAL_OK) && (s & RQ_NET_SIS))
116
			drbd_set_in_sync(mdev, req->i.sector, req->i.size);
P
Philipp Reisner 已提交
117 118

		/* one might be tempted to move the drbd_al_complete_io
119
		 * to the local io completion callback drbd_request_endio.
P
Philipp Reisner 已提交
120 121 122 123 124 125 126 127 128 129
		 * but, if this was a mirror write, we may only
		 * drbd_al_complete_io after this is RQ_NET_DONE,
		 * otherwise the extent could be dropped from the al
		 * before it has actually been written on the peer.
		 * if we crash before our peer knows about the request,
		 * but after the extent has been dropped from the al,
		 * we would forget to resync the corresponding extent.
		 */
		if (s & RQ_LOCAL_MASK) {
			if (get_ldev_if_state(mdev, D_FAILED)) {
130
				if (s & RQ_IN_ACT_LOG)
131
					drbd_al_complete_io(mdev, req->i.sector);
P
Philipp Reisner 已提交
132 133 134 135
				put_ldev(mdev);
			} else if (__ratelimit(&drbd_ratelimit_state)) {
				dev_warn(DEV, "Should have called drbd_al_complete_io(, %llu), "
				     "but my Disk seems to have failed :(\n",
136
				     (unsigned long long) req->i.sector);
P
Philipp Reisner 已提交
137 138 139 140
			}
		}
	}

141
	drbd_req_free(req);
P
Philipp Reisner 已提交
142 143 144 145 146 147 148 149 150 151 152 153 154 155
}

static void queue_barrier(struct drbd_conf *mdev)
{
	struct drbd_tl_epoch *b;

	/* We are within the req_lock. Once we queued the barrier for sending,
	 * we set the CREATE_BARRIER bit. It is cleared as soon as a new
	 * barrier/epoch object is added. This is the only place this bit is
	 * set. It indicates that the barrier for this epoch is already queued,
	 * and no new epoch has been created yet. */
	if (test_bit(CREATE_BARRIER, &mdev->flags))
		return;

156
	b = mdev->tconn->newest_tle;
P
Philipp Reisner 已提交
157
	b->w.cb = w_send_barrier;
158
	b->w.mdev = mdev;
P
Philipp Reisner 已提交
159 160 161 162 163
	/* inc_ap_pending done here, so we won't
	 * get imbalanced on connection loss.
	 * dec_ap_pending will be done in got_BarrierAck
	 * or (on connection loss) in tl_clear.  */
	inc_ap_pending(mdev);
164
	drbd_queue_work(&mdev->tconn->data.work, &b->w);
P
Philipp Reisner 已提交
165 166 167 168 169 170 171 172
	set_bit(CREATE_BARRIER, &mdev->flags);
}

static void _about_to_complete_local_write(struct drbd_conf *mdev,
	struct drbd_request *req)
{
	const unsigned long s = req->rq_state;

173 174 175 176 177 178 179 180
	/* Before we can signal completion to the upper layers,
	 * we may need to close the current epoch.
	 * We can skip this, if this request has not even been sent, because we
	 * did not have a fully established connection yet/anymore, during
	 * bitmap exchange, or while we are C_AHEAD due to congestion policy.
	 */
	if (mdev->state.conn >= C_CONNECTED &&
	    (s & RQ_NET_SENT) != 0 &&
181
	    req->epoch == mdev->tconn->newest_tle->br_number)
P
Philipp Reisner 已提交
182 183 184 185 186 187 188 189 190 191
		queue_barrier(mdev);
}

void complete_master_bio(struct drbd_conf *mdev,
		struct bio_and_error *m)
{
	bio_endio(m->bio, m->error);
	dec_ap_bio(mdev);
}

192 193 194 195

static void drbd_remove_request_interval(struct rb_root *root,
					 struct drbd_request *req)
{
196
	struct drbd_conf *mdev = req->w.mdev;
197 198 199 200 201 202 203 204 205
	struct drbd_interval *i = &req->i;

	drbd_remove_interval(root, i);

	/* Wake up any processes waiting for this request to complete.  */
	if (i->waiting)
		wake_up(&mdev->misc_wait);
}

P
Philipp Reisner 已提交
206 207 208 209 210 211 212 213 214
/* Helper for __req_mod().
 * Set m->bio to the master bio, if it is fit to be completed,
 * or leave it alone (it is initialized to NULL in __req_mod),
 * if it has already been completed, or cannot be completed yet.
 * If m->bio is set, the error status to be returned is placed in m->error.
 */
void _req_may_be_done(struct drbd_request *req, struct bio_and_error *m)
{
	const unsigned long s = req->rq_state;
215
	struct drbd_conf *mdev = req->w.mdev;
P
Philipp Reisner 已提交
216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235
	/* only WRITES may end up here without a master bio (on barrier ack) */
	int rw = req->master_bio ? bio_data_dir(req->master_bio) : WRITE;

	/* we must not complete the master bio, while it is
	 *	still being processed by _drbd_send_zc_bio (drbd_send_dblock)
	 *	not yet acknowledged by the peer
	 *	not yet completed by the local io subsystem
	 * these flags may get cleared in any order by
	 *	the worker,
	 *	the receiver,
	 *	the bio_endio completion callbacks.
	 */
	if (s & RQ_NET_QUEUED)
		return;
	if (s & RQ_NET_PENDING)
		return;
	if (s & RQ_LOCAL_PENDING)
		return;

	if (req->master_bio) {
236
		/* this is DATA_RECEIVED (remote read)
P
Philipp Reisner 已提交
237 238
		 * or protocol C P_WRITE_ACK
		 * or protocol B P_RECV_ACK
239
		 * or protocol A "HANDED_OVER_TO_NETWORK" (SendAck)
P
Philipp Reisner 已提交
240 241 242 243 244 245 246 247 248 249 250 251 252 253 254
		 * or canceled or failed,
		 * or killed from the transfer log due to connection loss.
		 */

		/*
		 * figure out whether to report success or failure.
		 *
		 * report success when at least one of the operations succeeded.
		 * or, to put the other way,
		 * only report failure, when both operations failed.
		 *
		 * what to do about the failures is handled elsewhere.
		 * what we need to do here is just: complete the master_bio.
		 *
		 * local completion error, if any, has been stored as ERR_PTR
255
		 * in private_bio within drbd_request_endio.
P
Philipp Reisner 已提交
256 257 258 259 260 261
		 */
		int ok = (s & RQ_LOCAL_OK) || (s & RQ_NET_OK);
		int error = PTR_ERR(req->private_bio);

		/* remove the request from the conflict detection
		 * respective block_id verification hash */
262 263 264 265 266 267 268
		if (!drbd_interval_empty(&req->i)) {
			struct rb_root *root;

			if (rw == WRITE)
				root = &mdev->write_requests;
			else
				root = &mdev->read_requests;
269
			drbd_remove_request_interval(root, req);
270
		} else
P
Philipp Reisner 已提交
271
			D_ASSERT((s & (RQ_NET_MASK & ~RQ_NET_DONE)) == 0);
P
Philipp Reisner 已提交
272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295

		/* for writes we need to do some extra housekeeping */
		if (rw == WRITE)
			_about_to_complete_local_write(mdev, req);

		/* Update disk stats */
		_drbd_end_io_acct(mdev, req);

		m->error = ok ? 0 : (error ?: -EIO);
		m->bio = req->master_bio;
		req->master_bio = NULL;
	}

	if ((s & RQ_NET_MASK) == 0 || (s & RQ_NET_DONE)) {
		/* this is disconnected (local only) operation,
		 * or protocol C P_WRITE_ACK,
		 * or protocol A or B P_BARRIER_ACK,
		 * or killed from the transfer log due to connection loss. */
		_req_is_done(mdev, req, rw);
	}
	/* else: network part and not DONE yet. that is
	 * protocol A or B, barrier ack still pending... */
}

296 297
static void _req_may_be_done_not_susp(struct drbd_request *req, struct bio_and_error *m)
{
298
	struct drbd_conf *mdev = req->w.mdev;
299

300
	if (!is_susp(mdev->state))
301 302 303
		_req_may_be_done(req, m);
}

P
Philipp Reisner 已提交
304 305 306 307 308 309 310 311 312 313 314 315
/* obviously this could be coded as many single functions
 * instead of one huge switch,
 * or by putting the code directly in the respective locations
 * (as it has been before).
 *
 * but having it this way
 *  enforces that it is all in this one place, where it is easier to audit,
 *  it makes it obvious that whatever "event" "happens" to a request should
 *  happen "atomically" within the req_lock,
 *  and it enforces that we have to think in a very structured manner
 *  about the "events" that may happen to a request during its life time ...
 */
316
int __req_mod(struct drbd_request *req, enum drbd_req_event what,
P
Philipp Reisner 已提交
317 318
		struct bio_and_error *m)
{
319
	struct drbd_conf *mdev = req->w.mdev;
320
	int rv = 0;
P
Philipp Reisner 已提交
321 322 323 324 325 326 327 328 329
	m->bio = NULL;

	switch (what) {
	default:
		dev_err(DEV, "LOGIC BUG in %s:%u\n", __FILE__ , __LINE__);
		break;

	/* does not happen...
	 * initialization done in drbd_req_new
330
	case CREATED:
P
Philipp Reisner 已提交
331 332 333
		break;
		*/

334
	case TO_BE_SENT: /* via network */
P
Philipp Reisner 已提交
335 336 337 338 339 340 341
		/* reached via drbd_make_request_common
		 * and from w_read_retry_remote */
		D_ASSERT(!(req->rq_state & RQ_NET_MASK));
		req->rq_state |= RQ_NET_PENDING;
		inc_ap_pending(mdev);
		break;

342
	case TO_BE_SUBMITTED: /* locally */
P
Philipp Reisner 已提交
343 344 345 346 347
		/* reached via drbd_make_request_common */
		D_ASSERT(!(req->rq_state & RQ_LOCAL_MASK));
		req->rq_state |= RQ_LOCAL_PENDING;
		break;

348
	case COMPLETED_OK:
P
Philipp Reisner 已提交
349
		if (bio_data_dir(req->master_bio) == WRITE)
350
			mdev->writ_cnt += req->i.size >> 9;
P
Philipp Reisner 已提交
351
		else
352
			mdev->read_cnt += req->i.size >> 9;
P
Philipp Reisner 已提交
353 354 355 356

		req->rq_state |= (RQ_LOCAL_COMPLETED|RQ_LOCAL_OK);
		req->rq_state &= ~RQ_LOCAL_PENDING;

357
		_req_may_be_done_not_susp(req, m);
P
Philipp Reisner 已提交
358 359 360
		put_ldev(mdev);
		break;

361
	case WRITE_COMPLETED_WITH_ERROR:
P
Philipp Reisner 已提交
362 363 364
		req->rq_state |= RQ_LOCAL_COMPLETED;
		req->rq_state &= ~RQ_LOCAL_PENDING;

365
		__drbd_chk_io_error(mdev, false);
366
		_req_may_be_done_not_susp(req, m);
P
Philipp Reisner 已提交
367 368 369
		put_ldev(mdev);
		break;

370
	case READ_AHEAD_COMPLETED_WITH_ERROR:
P
Philipp Reisner 已提交
371 372 373
		/* it is legal to fail READA */
		req->rq_state |= RQ_LOCAL_COMPLETED;
		req->rq_state &= ~RQ_LOCAL_PENDING;
374
		_req_may_be_done_not_susp(req, m);
P
Philipp Reisner 已提交
375 376 377
		put_ldev(mdev);
		break;

378
	case READ_COMPLETED_WITH_ERROR:
379
		drbd_set_out_of_sync(mdev, req->i.sector, req->i.size);
P
Philipp Reisner 已提交
380 381 382 383 384 385

		req->rq_state |= RQ_LOCAL_COMPLETED;
		req->rq_state &= ~RQ_LOCAL_PENDING;

		D_ASSERT(!(req->rq_state & RQ_NET_MASK));

386
		__drbd_chk_io_error(mdev, false);
P
Philipp Reisner 已提交
387 388
		put_ldev(mdev);

389 390 391
		/* no point in retrying if there is no good remote data,
		 * or we have no connection. */
		if (mdev->state.pdsk != D_UP_TO_DATE) {
392
			_req_may_be_done_not_susp(req, m);
393 394 395
			break;
		}

396
		/* _req_mod(req,TO_BE_SENT); oops, recursion... */
397 398
		req->rq_state |= RQ_NET_PENDING;
		inc_ap_pending(mdev);
399
		/* fall through: _req_mod(req,QUEUE_FOR_NET_READ); */
P
Philipp Reisner 已提交
400

401
	case QUEUE_FOR_NET_READ:
P
Philipp Reisner 已提交
402 403 404 405 406 407 408 409 410
		/* READ or READA, and
		 * no local disk,
		 * or target area marked as invalid,
		 * or just got an io-error. */
		/* from drbd_make_request_common
		 * or from bio_endio during read io-error recovery */

		/* so we can verify the handle in the answer packet
		 * corresponding hlist_del is in _req_may_be_done() */
411
		drbd_insert_interval(&mdev->read_requests, &req->i);
P
Philipp Reisner 已提交
412

413
		set_bit(UNPLUG_REMOTE, &mdev->flags);
P
Philipp Reisner 已提交
414 415 416 417 418 419

		D_ASSERT(req->rq_state & RQ_NET_PENDING);
		req->rq_state |= RQ_NET_QUEUED;
		req->w.cb = (req->rq_state & RQ_LOCAL_MASK)
			? w_read_retry_remote
			: w_send_read_req;
420
		drbd_queue_work(&mdev->tconn->data.work, &req->w);
P
Philipp Reisner 已提交
421 422
		break;

423
	case QUEUE_FOR_NET_WRITE:
P
Philipp Reisner 已提交
424 425 426 427
		/* assert something? */
		/* from drbd_make_request_common only */

		/* corresponding hlist_del is in _req_may_be_done() */
428
		drbd_insert_interval(&mdev->write_requests, &req->i);
P
Philipp Reisner 已提交
429 430 431 432 433 434 435 436 437 438 439 440 441 442 443

		/* NOTE
		 * In case the req ended up on the transfer log before being
		 * queued on the worker, it could lead to this request being
		 * missed during cleanup after connection loss.
		 * So we have to do both operations here,
		 * within the same lock that protects the transfer log.
		 *
		 * _req_add_to_epoch(req); this has to be after the
		 * _maybe_start_new_epoch(req); which happened in
		 * drbd_make_request_common, because we now may set the bit
		 * again ourselves to close the current epoch.
		 *
		 * Add req to the (now) current epoch (barrier). */

444 445 446 447 448
		/* otherwise we may lose an unplug, which may cause some remote
		 * io-scheduler timeout to expire, increasing maximum latency,
		 * hurting performance. */
		set_bit(UNPLUG_REMOTE, &mdev->flags);

P
Philipp Reisner 已提交
449 450 451 452
		/* see drbd_make_request_common,
		 * just after it grabs the req_lock */
		D_ASSERT(test_bit(CREATE_BARRIER, &mdev->flags) == 0);

453
		req->epoch = mdev->tconn->newest_tle->br_number;
P
Philipp Reisner 已提交
454 455

		/* increment size of current epoch */
456
		mdev->tconn->newest_tle->n_writes++;
P
Philipp Reisner 已提交
457 458 459 460 461

		/* queue work item to send data */
		D_ASSERT(req->rq_state & RQ_NET_PENDING);
		req->rq_state |= RQ_NET_QUEUED;
		req->w.cb =  w_send_dblock;
462
		drbd_queue_work(&mdev->tconn->data.work, &req->w);
P
Philipp Reisner 已提交
463 464

		/* close the epoch, in case it outgrew the limit */
465
		if (mdev->tconn->newest_tle->n_writes >= mdev->tconn->net_conf->max_epoch_size)
P
Philipp Reisner 已提交
466 467 468 469
			queue_barrier(mdev);

		break;

470
	case QUEUE_FOR_SEND_OOS:
471 472
		req->rq_state |= RQ_NET_QUEUED;
		req->w.cb =  w_send_oos;
473
		drbd_queue_work(&mdev->tconn->data.work, &req->w);
474 475
		break;

476
	case OOS_HANDED_TO_NETWORK:
477
		/* actually the same */
478
	case SEND_CANCELED:
P
Philipp Reisner 已提交
479
		/* treat it the same */
480
	case SEND_FAILED:
P
Philipp Reisner 已提交
481 482 483 484 485
		/* real cleanup will be done from tl_clear.  just update flags
		 * so it is no longer marked as on the worker queue */
		req->rq_state &= ~RQ_NET_QUEUED;
		/* if we did it right, tl_clear should be scheduled only after
		 * this, so this should not be necessary! */
486
		_req_may_be_done_not_susp(req, m);
P
Philipp Reisner 已提交
487 488
		break;

489
	case HANDED_OVER_TO_NETWORK:
P
Philipp Reisner 已提交
490
		/* assert something? */
491
		if (bio_data_dir(req->master_bio) == WRITE)
492
			atomic_add(req->i.size >> 9, &mdev->ap_in_flight);
493

P
Philipp Reisner 已提交
494
		if (bio_data_dir(req->master_bio) == WRITE &&
495
		    mdev->tconn->net_conf->wire_protocol == DRBD_PROT_A) {
P
Philipp Reisner 已提交
496 497 498 499 500 501 502 503 504 505 506 507 508 509
			/* this is what is dangerous about protocol A:
			 * pretend it was successfully written on the peer. */
			if (req->rq_state & RQ_NET_PENDING) {
				dec_ap_pending(mdev);
				req->rq_state &= ~RQ_NET_PENDING;
				req->rq_state |= RQ_NET_OK;
			} /* else: neg-ack was faster... */
			/* it is still not yet RQ_NET_DONE until the
			 * corresponding epoch barrier got acked as well,
			 * so we know what to dirty on connection loss */
		}
		req->rq_state &= ~RQ_NET_QUEUED;
		req->rq_state |= RQ_NET_SENT;
		/* because _drbd_send_zc_bio could sleep, and may want to
510 511
		 * dereference the bio even after the "WRITE_ACKED_BY_PEER" and
		 * "COMPLETED_OK" events came in, once we return from
P
Philipp Reisner 已提交
512 513
		 * _drbd_send_zc_bio (drbd_send_dblock), we have to check
		 * whether it is done already, and end it.  */
514
		_req_may_be_done_not_susp(req, m);
P
Philipp Reisner 已提交
515 516
		break;

517
	case READ_RETRY_REMOTE_CANCELED:
518 519
		req->rq_state &= ~RQ_NET_QUEUED;
		/* fall through, in case we raced with drbd_disconnect */
520
	case CONNECTION_LOST_WHILE_PENDING:
P
Philipp Reisner 已提交
521 522 523 524 525 526
		/* transfer log cleanup after connection loss */
		/* assert something? */
		if (req->rq_state & RQ_NET_PENDING)
			dec_ap_pending(mdev);
		req->rq_state &= ~(RQ_NET_OK|RQ_NET_PENDING);
		req->rq_state |= RQ_NET_DONE;
527
		if (req->rq_state & RQ_NET_SENT && req->rq_state & RQ_WRITE)
528
			atomic_sub(req->i.size >> 9, &mdev->ap_in_flight);
529

P
Philipp Reisner 已提交
530 531 532
		/* if it is still queued, we may not complete it here.
		 * it will be canceled soon. */
		if (!(req->rq_state & RQ_NET_QUEUED))
533
			_req_may_be_done(req, m); /* Allowed while state.susp */
P
Philipp Reisner 已提交
534 535
		break;

536
	case WRITE_ACKED_BY_PEER_AND_SIS:
P
Philipp Reisner 已提交
537
		req->rq_state |= RQ_NET_SIS;
538
	case CONFLICT_DISCARDED_BY_PEER:
P
Philipp Reisner 已提交
539 540 541
		/* for discarded conflicting writes of multiple primaries,
		 * there is no need to keep anything in the tl, potential
		 * node crashes are covered by the activity log. */
542
		if (what == CONFLICT_DISCARDED_BY_PEER)
P
Philipp Reisner 已提交
543 544
			dev_alert(DEV, "Got DiscardAck packet %llus +%u!"
			      " DRBD is not a random data generator!\n",
545
			      (unsigned long long)req->i.sector, req->i.size);
P
Philipp Reisner 已提交
546 547
		req->rq_state |= RQ_NET_DONE;
		/* fall through */
548
	case WRITE_ACKED_BY_PEER:
P
Philipp Reisner 已提交
549 550 551 552 553 554 555 556 557 558 559
		/* protocol C; successfully written on peer.
		 * Nothing to do here.
		 * We want to keep the tl in place for all protocols, to cater
		 * for volatile write-back caches on lower level devices.
		 *
		 * A barrier request is expected to have forced all prior
		 * requests onto stable storage, so completion of a barrier
		 * request could set NET_DONE right here, and not wait for the
		 * P_BARRIER_ACK, but that is an unnecessary optimization. */

		/* this makes it effectively the same as for: */
560
	case RECV_ACKED_BY_PEER:
P
Philipp Reisner 已提交
561
		/* protocol B; pretends to be successfully written on peer.
562
		 * see also notes above in HANDED_OVER_TO_NETWORK about
P
Philipp Reisner 已提交
563 564 565 566
		 * protocol != C */
		req->rq_state |= RQ_NET_OK;
		D_ASSERT(req->rq_state & RQ_NET_PENDING);
		dec_ap_pending(mdev);
567
		atomic_sub(req->i.size >> 9, &mdev->ap_in_flight);
P
Philipp Reisner 已提交
568
		req->rq_state &= ~RQ_NET_PENDING;
569
		_req_may_be_done_not_susp(req, m);
P
Philipp Reisner 已提交
570 571
		break;

572
	case NEG_ACKED:
P
Philipp Reisner 已提交
573
		/* assert something? */
574
		if (req->rq_state & RQ_NET_PENDING) {
P
Philipp Reisner 已提交
575
			dec_ap_pending(mdev);
576
			atomic_sub(req->i.size >> 9, &mdev->ap_in_flight);
577
		}
P
Philipp Reisner 已提交
578 579 580
		req->rq_state &= ~(RQ_NET_OK|RQ_NET_PENDING);

		req->rq_state |= RQ_NET_DONE;
581
		_req_may_be_done_not_susp(req, m);
582
		/* else: done by HANDED_OVER_TO_NETWORK */
P
Philipp Reisner 已提交
583 584
		break;

585
	case FAIL_FROZEN_DISK_IO:
586 587 588
		if (!(req->rq_state & RQ_LOCAL_COMPLETED))
			break;

589
		_req_may_be_done(req, m); /* Allowed while state.susp */
590 591
		break;

592
	case RESTART_FROZEN_DISK_IO:
593 594 595 596 597 598 599 600 601 602 603
		if (!(req->rq_state & RQ_LOCAL_COMPLETED))
			break;

		req->rq_state &= ~RQ_LOCAL_COMPLETED;

		rv = MR_READ;
		if (bio_data_dir(req->master_bio) == WRITE)
			rv = MR_WRITE;

		get_ldev(mdev);
		req->w.cb = w_restart_disk_io;
604
		drbd_queue_work(&mdev->tconn->data.work, &req->w);
605 606
		break;

607
	case RESEND:
608
		/* If RQ_NET_OK is already set, we got a P_WRITE_ACK or P_RECV_ACK
609
		   before the connection loss (B&C only); only P_BARRIER_ACK was missing.
610
		   Trowing them out of the TL here by pretending we got a BARRIER_ACK
611
		   We ensure that the peer was not rebooted */
612 613
		if (!(req->rq_state & RQ_NET_OK)) {
			if (req->w.cb) {
614
				drbd_queue_work(&mdev->tconn->data.work, &req->w);
615 616 617 618
				rv = req->rq_state & RQ_WRITE ? MR_WRITE : MR_READ;
			}
			break;
		}
619
		/* else, fall through to BARRIER_ACKED */
620

621
	case BARRIER_ACKED:
622 623 624
		if (!(req->rq_state & RQ_WRITE))
			break;

P
Philipp Reisner 已提交
625 626 627 628
		if (req->rq_state & RQ_NET_PENDING) {
			/* barrier came in before all requests have been acked.
			 * this is bad, because if the connection is lost now,
			 * we won't be able to clean them up... */
629
			dev_err(DEV, "FIXME (BARRIER_ACKED but pending)\n");
630
			list_move(&req->tl_requests, &mdev->tconn->out_of_sequence_requests);
P
Philipp Reisner 已提交
631
		}
632 633
		if ((req->rq_state & RQ_NET_MASK) != 0) {
			req->rq_state |= RQ_NET_DONE;
634 635
			if (mdev->tconn->net_conf->wire_protocol == DRBD_PROT_A)
				atomic_sub(req->i.size>>9, &mdev->ap_in_flight);
636
		}
637
		_req_may_be_done(req, m); /* Allowed while state.susp */
P
Philipp Reisner 已提交
638 639
		break;

640
	case DATA_RECEIVED:
P
Philipp Reisner 已提交
641 642 643 644
		D_ASSERT(req->rq_state & RQ_NET_PENDING);
		dec_ap_pending(mdev);
		req->rq_state &= ~RQ_NET_PENDING;
		req->rq_state |= (RQ_NET_OK|RQ_NET_DONE);
645
		_req_may_be_done_not_susp(req, m);
P
Philipp Reisner 已提交
646 647
		break;
	};
648 649

	return rv;
P
Philipp Reisner 已提交
650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682
}

/* we may do a local read if:
 * - we are consistent (of course),
 * - or we are generally inconsistent,
 *   BUT we are still/already IN SYNC for this area.
 *   since size may be bigger than BM_BLOCK_SIZE,
 *   we may need to check several bits.
 */
static int drbd_may_do_local_read(struct drbd_conf *mdev, sector_t sector, int size)
{
	unsigned long sbnr, ebnr;
	sector_t esector, nr_sectors;

	if (mdev->state.disk == D_UP_TO_DATE)
		return 1;
	if (mdev->state.disk >= D_OUTDATED)
		return 0;
	if (mdev->state.disk <  D_INCONSISTENT)
		return 0;
	/* state.disk == D_INCONSISTENT   We will have a look at the BitMap */
	nr_sectors = drbd_get_capacity(mdev->this_bdev);
	esector = sector + (size >> 9) - 1;

	D_ASSERT(sector  < nr_sectors);
	D_ASSERT(esector < nr_sectors);

	sbnr = BM_SECT_TO_BIT(sector);
	ebnr = BM_SECT_TO_BIT(esector);

	return 0 == drbd_bm_count_bits(mdev, sbnr, ebnr);
}

683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710
/*
 * complete_conflicting_writes  -  wait for any conflicting write requests
 *
 * The write_requests tree contains all active write requests which we
 * currently know about.  Wait for any requests to complete which conflict with
 * the new one.
 */
static int complete_conflicting_writes(struct drbd_conf *mdev,
				       sector_t sector, int size)
{
	for(;;) {
		DEFINE_WAIT(wait);
		struct drbd_interval *i;

		i = drbd_find_overlap(&mdev->write_requests, sector, size);
		if (!i)
			return 0;
		i->waiting = true;
		prepare_to_wait(&mdev->misc_wait, &wait, TASK_INTERRUPTIBLE);
		spin_unlock_irq(&mdev->tconn->req_lock);
		schedule();
		finish_wait(&mdev->misc_wait, &wait);
		spin_lock_irq(&mdev->tconn->req_lock);
		if (signal_pending(current))
			return -ERESTARTSYS;
	}
}

711
static int drbd_make_request_common(struct drbd_conf *mdev, struct bio *bio, unsigned long start_time)
P
Philipp Reisner 已提交
712 713 714 715 716 717
{
	const int rw = bio_rw(bio);
	const int size = bio->bi_size;
	const sector_t sector = bio->bi_sector;
	struct drbd_tl_epoch *b = NULL;
	struct drbd_request *req;
718
	int local, remote, send_oos = 0;
719
	int err;
720
	int ret = 0;
P
Philipp Reisner 已提交
721 722 723 724 725 726 727 728 729 730 731

	/* allocate outside of all locks; */
	req = drbd_req_new(mdev, bio);
	if (!req) {
		dec_ap_bio(mdev);
		/* only pass the error to the upper layers.
		 * if user cannot handle io errors, that's not our business. */
		dev_err(DEV, "could not kmalloc() req\n");
		bio_endio(bio, -ENOMEM);
		return 0;
	}
732
	req->start_time = start_time;
P
Philipp Reisner 已提交
733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776

	local = get_ldev(mdev);
	if (!local) {
		bio_put(req->private_bio); /* or we get a bio leak */
		req->private_bio = NULL;
	}
	if (rw == WRITE) {
		remote = 1;
	} else {
		/* READ || READA */
		if (local) {
			if (!drbd_may_do_local_read(mdev, sector, size)) {
				/* we could kick the syncer to
				 * sync this extent asap, wait for
				 * it, then continue locally.
				 * Or just issue the request remotely.
				 */
				local = 0;
				bio_put(req->private_bio);
				req->private_bio = NULL;
				put_ldev(mdev);
			}
		}
		remote = !local && mdev->state.pdsk >= D_UP_TO_DATE;
	}

	/* If we have a disk, but a READA request is mapped to remote,
	 * we are R_PRIMARY, D_INCONSISTENT, SyncTarget.
	 * Just fail that READA request right here.
	 *
	 * THINK: maybe fail all READA when not local?
	 *        or make this configurable...
	 *        if network is slow, READA won't do any good.
	 */
	if (rw == READA && mdev->state.disk >= D_INCONSISTENT && !local) {
		err = -EWOULDBLOCK;
		goto fail_and_free_req;
	}

	/* For WRITES going to the local disk, grab a reference on the target
	 * extent.  This waits for any resync activity in the corresponding
	 * resync extent to finish, and, if necessary, pulls in the target
	 * extent into the activity log, which involves further disk io because
	 * of transactional on-disk meta data updates. */
777 778
	if (rw == WRITE && local && !test_bit(AL_SUSPENDED, &mdev->flags)) {
		req->rq_state |= RQ_IN_ACT_LOG;
P
Philipp Reisner 已提交
779
		drbd_al_begin_io(mdev, sector);
780
	}
P
Philipp Reisner 已提交
781

782 783
	remote = remote && drbd_should_do_remote(mdev->state);
	send_oos = rw == WRITE && drbd_should_send_oos(mdev->state);
784
	D_ASSERT(!(remote && send_oos));
P
Philipp Reisner 已提交
785

786
	if (!(local || remote) && !is_susp(mdev->state)) {
L
Lars Ellenberg 已提交
787 788
		if (__ratelimit(&drbd_ratelimit_state))
			dev_err(DEV, "IO ERROR: neither local nor remote disk\n");
789
		err = -EIO;
P
Philipp Reisner 已提交
790 791 792 793 794 795 796 797 798
		goto fail_free_complete;
	}

	/* For WRITE request, we have to make sure that we have an
	 * unused_spare_tle, in case we need to start a new epoch.
	 * I try to be smart and avoid to pre-allocate always "just in case",
	 * 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.  */
799
	if (rw == WRITE && (remote || send_oos) &&
800
	    mdev->tconn->unused_spare_tle == NULL &&
P
Philipp Reisner 已提交
801 802 803 804 805 806 807 808 809 810 811
	    test_bit(CREATE_BARRIER, &mdev->flags)) {
allocate_barrier:
		b = kmalloc(sizeof(struct drbd_tl_epoch), GFP_NOIO);
		if (!b) {
			dev_err(DEV, "Failed to alloc barrier.\n");
			err = -ENOMEM;
			goto fail_free_complete;
		}
	}

	/* GOOD, everything prepared, grab the spin_lock */
812
	spin_lock_irq(&mdev->tconn->req_lock);
P
Philipp Reisner 已提交
813

814 815 816 817 818 819 820 821
	if (rw == WRITE) {
		err = complete_conflicting_writes(mdev, sector, size);
		if (err) {
			spin_unlock_irq(&mdev->tconn->req_lock);
			goto fail_free_complete;
		}
	}

822
	if (is_susp(mdev->state)) {
823 824
		/* If we got suspended, use the retry mechanism of
		   generic_make_request() to restart processing of this
825
		   bio. In the next call to drbd_make_request
826 827
		   we sleep in inc_ap_bio() */
		ret = 1;
828
		spin_unlock_irq(&mdev->tconn->req_lock);
829 830 831
		goto fail_free_complete;
	}

832
	if (remote || send_oos) {
833 834
		remote = drbd_should_do_remote(mdev->state);
		send_oos = rw == WRITE && drbd_should_send_oos(mdev->state);
835
		D_ASSERT(!(remote && send_oos));
836 837

		if (!(remote || send_oos))
P
Philipp Reisner 已提交
838 839 840
			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");
841
			spin_unlock_irq(&mdev->tconn->req_lock);
842
			err = -EIO;
P
Philipp Reisner 已提交
843 844 845 846
			goto fail_free_complete;
		}
	}

847 848
	if (b && mdev->tconn->unused_spare_tle == NULL) {
		mdev->tconn->unused_spare_tle = b;
P
Philipp Reisner 已提交
849 850
		b = NULL;
	}
851
	if (rw == WRITE && (remote || send_oos) &&
852
	    mdev->tconn->unused_spare_tle == NULL &&
P
Philipp Reisner 已提交
853 854 855
	    test_bit(CREATE_BARRIER, &mdev->flags)) {
		/* someone closed the current epoch
		 * while we were grabbing the spinlock */
856
		spin_unlock_irq(&mdev->tconn->req_lock);
P
Philipp Reisner 已提交
857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873
		goto allocate_barrier;
	}


	/* Update disk stats */
	_drbd_start_io_acct(mdev, req, bio);

	/* _maybe_start_new_epoch(mdev);
	 * If we need to generate a write barrier packet, we have to add the
	 * new epoch (barrier) object, and queue the barrier packet for sending,
	 * and queue the req's data after it _within the same lock_, otherwise
	 * we have race conditions were the reorder domains could be mixed up.
	 *
	 * Even read requests may start a new epoch and queue the corresponding
	 * 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. */
874
	if ((remote || send_oos) && mdev->tconn->unused_spare_tle &&
P
Philipp Reisner 已提交
875
	    test_and_clear_bit(CREATE_BARRIER, &mdev->flags)) {
876 877
		_tl_add_barrier(mdev, mdev->tconn->unused_spare_tle);
		mdev->tconn->unused_spare_tle = NULL;
P
Philipp Reisner 已提交
878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898
	} else {
		D_ASSERT(!(remote && rw == WRITE &&
			   test_bit(CREATE_BARRIER, &mdev->flags)));
	}

	/* NOTE
	 * Actually, 'local' may be wrong here already, since we may have failed
	 * to write to the meta data, and may become wrong anytime because of
	 * local io-error for some other request, which would lead to us
	 * "detaching" the local disk.
	 *
	 * 'remote' may become wrong any time because the network could fail.
	 *
	 * This is a harmless race condition, though, since it is handled
	 * correctly at the appropriate places; so it just defers the failure
	 * of the respective operation.
	 */

	/* mark them early for readability.
	 * this just sets some state flags. */
	if (remote)
899
		_req_mod(req, TO_BE_SENT);
P
Philipp Reisner 已提交
900
	if (local)
901
		_req_mod(req, TO_BE_SUBMITTED);
P
Philipp Reisner 已提交
902

903
	list_add_tail(&req->tl_requests, &mdev->tconn->newest_tle->requests);
904

P
Philipp Reisner 已提交
905 906 907 908 909 910 911 912
	/* NOTE remote first: to get the concurrent write detection right,
	 * we must register the request before start of local IO.  */
	if (remote) {
		/* either WRITE and C_CONNECTED,
		 * or READ, and no local disk,
		 * or READ, but not in sync.
		 */
		_req_mod(req, (rw == WRITE)
913 914
				? QUEUE_FOR_NET_WRITE
				: QUEUE_FOR_NET_READ);
P
Philipp Reisner 已提交
915
	}
916
	if (send_oos && drbd_set_out_of_sync(mdev, sector, size))
917
		_req_mod(req, QUEUE_FOR_SEND_OOS);
918

919
	if (remote &&
920
	    mdev->tconn->net_conf->on_congestion != OC_BLOCK && mdev->tconn->agreed_pro_version >= 96) {
921 922
		int congested = 0;

923 924
		if (mdev->tconn->net_conf->cong_fill &&
		    atomic_read(&mdev->ap_in_flight) >= mdev->tconn->net_conf->cong_fill) {
925 926 927 928
			dev_info(DEV, "Congestion-fill threshold reached\n");
			congested = 1;
		}

929
		if (mdev->act_log->used >= mdev->tconn->net_conf->cong_extents) {
930 931 932 933
			dev_info(DEV, "Congestion-extents threshold reached\n");
			congested = 1;
		}

934
		if (congested) {
935
			queue_barrier(mdev); /* last barrier, after mirrored writes */
936

937
			if (mdev->tconn->net_conf->on_congestion == OC_PULL_AHEAD)
938
				_drbd_set_state(_NS(mdev, conn, C_AHEAD), 0, NULL);
939
			else  /*mdev->tconn->net_conf->on_congestion == OC_DISCONNECT */
940 941 942 943
				_drbd_set_state(_NS(mdev, conn, C_DISCONNECTING), 0, NULL);
		}
	}

944
	spin_unlock_irq(&mdev->tconn->req_lock);
P
Philipp Reisner 已提交
945 946 947 948 949
	kfree(b); /* if someone else has beaten us to it... */

	if (local) {
		req->private_bio->bi_bdev = mdev->ldev->backing_bdev;

950 951 952 953 954 955
		/* State may have changed since we grabbed our reference on the
		 * mdev->ldev member. Double check, and short-circuit to endio.
		 * In case the last activity log transaction failed to get on
		 * stable storage, and this is a WRITE, we may not even submit
		 * this bio. */
		if (get_ldev(mdev)) {
956 957 958
			if (drbd_insert_fault(mdev,   rw == WRITE ? DRBD_FAULT_DT_WR
						    : rw == READ  ? DRBD_FAULT_DT_RD
						    :               DRBD_FAULT_DT_RA))
959 960 961 962 963
				bio_endio(req->private_bio, -EIO);
			else
				generic_make_request(req->private_bio);
			put_ldev(mdev);
		} else
P
Philipp Reisner 已提交
964 965 966 967 968 969
			bio_endio(req->private_bio, -EIO);
	}

	return 0;

fail_free_complete:
970
	if (req->rq_state & RQ_IN_ACT_LOG)
P
Philipp Reisner 已提交
971 972 973 974 975 976 977
		drbd_al_complete_io(mdev, sector);
fail_and_free_req:
	if (local) {
		bio_put(req->private_bio);
		req->private_bio = NULL;
		put_ldev(mdev);
	}
978 979 980
	if (!ret)
		bio_endio(bio, err);

P
Philipp Reisner 已提交
981 982 983 984
	drbd_req_free(req);
	dec_ap_bio(mdev);
	kfree(b);

985
	return ret;
P
Philipp Reisner 已提交
986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009
}

/* helper function for drbd_make_request
 * if we can determine just by the mdev (state) that this request will fail,
 * return 1
 * otherwise return 0
 */
static int drbd_fail_request_early(struct drbd_conf *mdev, int is_write)
{
	if (mdev->state.role != R_PRIMARY &&
		(!allow_oos || is_write)) {
		if (__ratelimit(&drbd_ratelimit_state)) {
			dev_err(DEV, "Process %s[%u] tried to %s; "
			    "since we are not in Primary state, "
			    "we cannot allow this\n",
			    current->comm, current->pid,
			    is_write ? "WRITE" : "READ");
		}
		return 1;
	}

	return 0;
}

1010
int drbd_make_request(struct request_queue *q, struct bio *bio)
P
Philipp Reisner 已提交
1011 1012 1013
{
	unsigned int s_enr, e_enr;
	struct drbd_conf *mdev = (struct drbd_conf *) q->queuedata;
1014
	unsigned long start_time;
P
Philipp Reisner 已提交
1015 1016 1017 1018 1019 1020

	if (drbd_fail_request_early(mdev, bio_data_dir(bio) & WRITE)) {
		bio_endio(bio, -EPERM);
		return 0;
	}

1021 1022
	start_time = jiffies;

P
Philipp Reisner 已提交
1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036
	/*
	 * what we "blindly" assume:
	 */
	D_ASSERT(bio->bi_size > 0);
	D_ASSERT((bio->bi_size & 0x1ff) == 0);
	D_ASSERT(bio->bi_idx == 0);

	/* to make some things easier, force alignment of requests within the
	 * granularity of our hash tables */
	s_enr = bio->bi_sector >> HT_SHIFT;
	e_enr = (bio->bi_sector+(bio->bi_size>>9)-1) >> HT_SHIFT;

	if (likely(s_enr == e_enr)) {
		inc_ap_bio(mdev, 1);
1037
		return drbd_make_request_common(mdev, bio, start_time);
P
Philipp Reisner 已提交
1038 1039 1040 1041
	}

	/* can this bio be split generically?
	 * Maybe add our own split-arbitrary-bios function. */
1042
	if (bio->bi_vcnt != 1 || bio->bi_idx != 0 || bio->bi_size > DRBD_MAX_BIO_SIZE) {
P
Philipp Reisner 已提交
1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063
		/* rather error out here than BUG in bio_split */
		dev_err(DEV, "bio would need to, but cannot, be split: "
		    "(vcnt=%u,idx=%u,size=%u,sector=%llu)\n",
		    bio->bi_vcnt, bio->bi_idx, bio->bi_size,
		    (unsigned long long)bio->bi_sector);
		bio_endio(bio, -EINVAL);
	} else {
		/* This bio crosses some boundary, so we have to split it. */
		struct bio_pair *bp;
		/* works for the "do not cross hash slot boundaries" case
		 * e.g. sector 262269, size 4096
		 * s_enr = 262269 >> 6 = 4097
		 * e_enr = (262269+8-1) >> 6 = 4098
		 * HT_SHIFT = 6
		 * sps = 64, mask = 63
		 * first_sectors = 64 - (262269 & 63) = 3
		 */
		const sector_t sect = bio->bi_sector;
		const int sps = 1 << HT_SHIFT; /* sectors per slot */
		const int mask = sps - 1;
		const sector_t first_sectors = sps - (sect & mask);
1064
		bp = bio_split(bio, first_sectors);
P
Philipp Reisner 已提交
1065 1066 1067

		/* we need to get a "reference count" (ap_bio_cnt)
		 * to avoid races with the disconnect/reconnect/suspend code.
1068
		 * In case we need to split the bio here, we need to get three references
P
Philipp Reisner 已提交
1069 1070
		 * atomically, otherwise we might deadlock when trying to submit the
		 * second one! */
1071
		inc_ap_bio(mdev, 3);
P
Philipp Reisner 已提交
1072 1073 1074

		D_ASSERT(e_enr == s_enr + 1);

1075
		while (drbd_make_request_common(mdev, &bp->bio1, start_time))
1076 1077
			inc_ap_bio(mdev, 1);

1078
		while (drbd_make_request_common(mdev, &bp->bio2, start_time))
1079 1080 1081 1082
			inc_ap_bio(mdev, 1);

		dec_ap_bio(mdev);

P
Philipp Reisner 已提交
1083 1084 1085 1086 1087 1088
		bio_pair_release(bp);
	}
	return 0;
}

/* This is called by bio_add_page().  With this function we reduce
1089
 * the number of BIOs that span over multiple DRBD_MAX_BIO_SIZEs
P
Philipp Reisner 已提交
1090 1091 1092 1093 1094 1095 1096 1097 1098
 * units (was AL_EXTENTs).
 *
 * we do the calculation within the lower 32bit of the byte offsets,
 * since we don't care for actual offset, but only check whether it
 * would cross "activity log extent" boundaries.
 *
 * As long as the BIO is empty we have to allow at least one bvec,
 * regardless of size and offset.  so the resulting bio may still
 * cross extent boundaries.  those are dealt with (bio_split) in
1099
 * drbd_make_request.
P
Philipp Reisner 已提交
1100 1101 1102 1103 1104 1105 1106 1107 1108
 */
int drbd_merge_bvec(struct request_queue *q, struct bvec_merge_data *bvm, struct bio_vec *bvec)
{
	struct drbd_conf *mdev = (struct drbd_conf *) q->queuedata;
	unsigned int bio_offset =
		(unsigned int)bvm->bi_sector << 9; /* 32 bit */
	unsigned int bio_size = bvm->bi_size;
	int limit, backing_limit;

1109 1110
	limit = DRBD_MAX_BIO_SIZE
	      - ((bio_offset & (DRBD_MAX_BIO_SIZE-1)) + bio_size);
P
Philipp Reisner 已提交
1111 1112 1113 1114 1115 1116 1117 1118
	if (limit < 0)
		limit = 0;
	if (bio_size == 0) {
		if (limit <= bvec->bv_len)
			limit = bvec->bv_len;
	} else if (limit && get_ldev(mdev)) {
		struct request_queue * const b =
			mdev->ldev->backing_bdev->bd_disk->queue;
1119
		if (b->merge_bvec_fn) {
P
Philipp Reisner 已提交
1120 1121 1122 1123 1124 1125 1126
			backing_limit = b->merge_bvec_fn(b, bvm, bvec);
			limit = min(limit, backing_limit);
		}
		put_ldev(mdev);
	}
	return limit;
}
1127 1128 1129 1130 1131 1132 1133 1134

void request_timer_fn(unsigned long data)
{
	struct drbd_conf *mdev = (struct drbd_conf *) data;
	struct drbd_request *req; /* oldest request */
	struct list_head *le;
	unsigned long et = 0; /* effective timeout = ko_count * timeout */

1135
	if (get_net_conf(mdev->tconn)) {
1136
		et = mdev->tconn->net_conf->timeout*HZ/10 * mdev->tconn->net_conf->ko_count;
1137
		put_net_conf(mdev->tconn);
1138 1139 1140 1141
	}
	if (!et || mdev->state.conn < C_WF_REPORT_PARAMS)
		return; /* Recurring timer stopped */

1142 1143
	spin_lock_irq(&mdev->tconn->req_lock);
	le = &mdev->tconn->oldest_tle->requests;
1144
	if (list_empty(le)) {
1145
		spin_unlock_irq(&mdev->tconn->req_lock);
1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163
		mod_timer(&mdev->request_timer, jiffies + et);
		return;
	}

	le = le->prev;
	req = list_entry(le, struct drbd_request, tl_requests);
	if (time_is_before_eq_jiffies(req->start_time + et)) {
		if (req->rq_state & RQ_NET_PENDING) {
			dev_warn(DEV, "Remote failed to finish a request within ko-count * timeout\n");
			_drbd_set_state(_NS(mdev, conn, C_TIMEOUT), CS_VERBOSE, NULL);
		} else {
			dev_warn(DEV, "Local backing block device frozen?\n");
			mod_timer(&mdev->request_timer, jiffies + et);
		}
	} else {
		mod_timer(&mdev->request_timer, req->start_time + et);
	}

1164
	spin_unlock_irq(&mdev->tconn->req_lock);
1165
}