drbd_main.c 92.8 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
/*
   drbd.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>.

   Thanks to Carter Burden, Bart Grantham and Gennadiy Nerubayev
   from Logicworks, Inc. for making SDP replication support possible.

   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 <asm/uaccess.h>
#include <asm/types.h>
#include <net/sock.h>
#include <linux/ctype.h>
35
#include <linux/mutex.h>
P
Philipp Reisner 已提交
36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58
#include <linux/fs.h>
#include <linux/file.h>
#include <linux/proc_fs.h>
#include <linux/init.h>
#include <linux/mm.h>
#include <linux/memcontrol.h>
#include <linux/mm_inline.h>
#include <linux/slab.h>
#include <linux/random.h>
#include <linux/reboot.h>
#include <linux/notifier.h>
#include <linux/kthread.h>

#define __KERNEL_SYSCALLS__
#include <linux/unistd.h>
#include <linux/vmalloc.h>

#include <linux/drbd_limits.h>
#include "drbd_int.h"
#include "drbd_req.h" /* only for _req_mod in tl_release and tl_clear */

#include "drbd_vli.h"

59
static DEFINE_MUTEX(drbd_main_mutex);
P
Philipp Reisner 已提交
60 61 62 63 64 65 66
int drbdd_init(struct drbd_thread *);
int drbd_worker(struct drbd_thread *);
int drbd_asender(struct drbd_thread *);

int drbd_init(void);
static int drbd_open(struct block_device *bdev, fmode_t mode);
static int drbd_release(struct gendisk *gd, fmode_t mode);
67
static int w_md_sync(struct drbd_work *w, int unused);
P
Philipp Reisner 已提交
68
static void md_sync_timer_fn(unsigned long data);
69 70
static int w_bitmap_io(struct drbd_work *w, int unused);
static int w_go_diskless(struct drbd_work *w, int unused);
P
Philipp Reisner 已提交
71 72 73 74 75 76

MODULE_AUTHOR("Philipp Reisner <phil@linbit.com>, "
	      "Lars Ellenberg <lars@linbit.com>");
MODULE_DESCRIPTION("drbd - Distributed Replicated Block Device v" REL_VERSION);
MODULE_VERSION(REL_VERSION);
MODULE_LICENSE("GPL");
77
MODULE_PARM_DESC(minor_count, "Approximate number of drbd devices ("
78
		 __stringify(DRBD_MINOR_COUNT_MIN) "-" __stringify(DRBD_MINOR_COUNT_MAX) ")");
P
Philipp Reisner 已提交
79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106
MODULE_ALIAS_BLOCKDEV_MAJOR(DRBD_MAJOR);

#include <linux/moduleparam.h>
/* allow_open_on_secondary */
MODULE_PARM_DESC(allow_oos, "DONT USE!");
/* thanks to these macros, if compiled into the kernel (not-module),
 * this becomes the boot parameter drbd.minor_count */
module_param(minor_count, uint, 0444);
module_param(disable_sendpage, bool, 0644);
module_param(allow_oos, bool, 0);
module_param(proc_details, int, 0644);

#ifdef CONFIG_DRBD_FAULT_INJECTION
int enable_faults;
int fault_rate;
static int fault_count;
int fault_devs;
/* bitmap of enabled faults */
module_param(enable_faults, int, 0664);
/* fault rate % value - applies to all enabled faults */
module_param(fault_rate, int, 0664);
/* count of faults inserted */
module_param(fault_count, int, 0664);
/* bitmap of devices to insert faults on */
module_param(fault_devs, int, 0644);
#endif

/* module parameter, defined */
107
unsigned int minor_count = DRBD_MINOR_COUNT_DEF;
P
Philipp Reisner 已提交
108 109 110 111 112 113 114 115 116 117 118 119 120
int disable_sendpage;
int allow_oos;
int proc_details;       /* Detail level in proc drbd*/

/* Module parameter for setting the user mode helper program
 * to run. Default is /sbin/drbdadm */
char usermode_helper[80] = "/sbin/drbdadm";

module_param_string(usermode_helper, usermode_helper, sizeof(usermode_helper), 0644);

/* in 2.6.x, our device mapping and config info contains our virtual gendisks
 * as member "struct gendisk *vdisk;"
 */
121
struct idr minors;
122
struct list_head drbd_tconns;  /* list of struct drbd_tconn */
P
Philipp Reisner 已提交
123 124

struct kmem_cache *drbd_request_cache;
A
Andreas Gruenbacher 已提交
125
struct kmem_cache *drbd_ee_cache;	/* peer requests */
P
Philipp Reisner 已提交
126 127 128 129
struct kmem_cache *drbd_bm_ext_cache;	/* bitmap extents */
struct kmem_cache *drbd_al_ext_cache;	/* activity log extents */
mempool_t *drbd_request_mempool;
mempool_t *drbd_ee_mempool;
130
mempool_t *drbd_md_io_page_pool;
131
struct bio_set *drbd_md_io_bio_set;
P
Philipp Reisner 已提交
132 133 134 135 136 137 138 139 140 141 142 143 144 145

/* I do not use a standard mempool, because:
   1) I want to hand out the pre-allocated objects first.
   2) I want to be able to interrupt sleeping allocation with a signal.
   Note: This is a single linked list, the next pointer is the private
	 member of struct page.
 */
struct page *drbd_pp_pool;
spinlock_t   drbd_pp_lock;
int          drbd_pp_vacant;
wait_queue_head_t drbd_pp_wait;

DEFINE_RATELIMIT_STATE(drbd_ratelimit_state, 5 * HZ, 5);

146
static const struct block_device_operations drbd_ops = {
P
Philipp Reisner 已提交
147 148 149 150 151
	.owner =   THIS_MODULE,
	.open =    drbd_open,
	.release = drbd_release,
};

152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170
static void bio_destructor_drbd(struct bio *bio)
{
	bio_free(bio, drbd_md_io_bio_set);
}

struct bio *bio_alloc_drbd(gfp_t gfp_mask)
{
	struct bio *bio;

	if (!drbd_md_io_bio_set)
		return bio_alloc(gfp_mask, 1);

	bio = bio_alloc_bioset(gfp_mask, 1, drbd_md_io_bio_set);
	if (!bio)
		return NULL;
	bio->bi_destructor = bio_destructor_drbd;
	return bio;
}

P
Philipp Reisner 已提交
171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190
#ifdef __CHECKER__
/* When checking with sparse, and this is an inline function, sparse will
   give tons of false positives. When this is a real functions sparse works.
 */
int _get_ldev_if_state(struct drbd_conf *mdev, enum drbd_disk_state mins)
{
	int io_allowed;

	atomic_inc(&mdev->local_cnt);
	io_allowed = (mdev->state.disk >= mins);
	if (!io_allowed) {
		if (atomic_dec_and_test(&mdev->local_cnt))
			wake_up(&mdev->misc_wait);
	}
	return io_allowed;
}

#endif

/**
191 192
 * tl_release() - mark as BARRIER_ACKED all requests in the corresponding transfer log epoch
 * @tconn:	DRBD connection.
P
Philipp Reisner 已提交
193 194 195 196
 * @barrier_nr:	Expected identifier of the DRBD write barrier packet.
 * @set_size:	Expected number of requests before that barrier.
 *
 * In case the passed barrier_nr or set_size does not match the oldest
197 198
 * epoch of not yet barrier-acked requests, this function will cause a
 * termination of the connection.
P
Philipp Reisner 已提交
199
 */
200 201
void tl_release(struct drbd_tconn *tconn, unsigned int barrier_nr,
		unsigned int set_size)
P
Philipp Reisner 已提交
202 203
{
	struct drbd_request *r;
204 205 206
	struct drbd_request *req = NULL;
	int expect_epoch = 0;
	int expect_size = 0;
P
Philipp Reisner 已提交
207

208
	spin_lock_irq(&tconn->req_lock);
P
Philipp Reisner 已提交
209

210 211 212
	/* find latest not yet barrier-acked write request,
	 * count writes in its epoch. */
	list_for_each_entry(r, &tconn->transfer_log, tl_requests) {
213
		const unsigned s = r->rq_state;
214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233
		if (!req) {
			if (!(s & RQ_WRITE))
				continue;
			if (!(s & RQ_NET_MASK))
				continue;
			if (s & RQ_NET_DONE)
				continue;
			req = r;
			expect_epoch = req->epoch;
			expect_size ++;
		} else {
			if (r->epoch != expect_epoch)
				break;
			if (!(s & RQ_WRITE))
				continue;
			/* if (s & RQ_DONE): not expected */
			/* if (!(s & RQ_NET_MASK)): not expected */
			expect_size++;
		}
	}
P
Philipp Reisner 已提交
234 235

	/* first some paranoia code */
236
	if (req == NULL) {
237 238
		conn_err(tconn, "BAD! BarrierAck #%u received, but no epoch in tl!?\n",
			 barrier_nr);
P
Philipp Reisner 已提交
239 240
		goto bail;
	}
241
	if (expect_epoch != barrier_nr) {
242
		conn_err(tconn, "BAD! BarrierAck #%u received, expected #%u!\n",
243
			 barrier_nr, expect_epoch);
P
Philipp Reisner 已提交
244 245
		goto bail;
	}
246 247

	if (expect_size != set_size) {
248
		conn_err(tconn, "BAD! BarrierAck #%u received with n_writes=%u, expected n_writes=%u!\n",
249
			 barrier_nr, set_size, expect_size);
P
Philipp Reisner 已提交
250 251 252 253
		goto bail;
	}

	/* Clean up list of requests processed during current epoch */
254 255 256 257
	list_for_each_entry_safe(req, r, &tconn->transfer_log, tl_requests) {
		if (req->epoch != expect_epoch)
			break;
		_req_mod(req, BARRIER_ACKED);
P
Philipp Reisner 已提交
258
	}
259
	spin_unlock_irq(&tconn->req_lock);
P
Philipp Reisner 已提交
260 261 262 263

	return;

bail:
264 265
	spin_unlock_irq(&tconn->req_lock);
	conn_request_state(tconn, NS(conn, C_PROTOCOL_ERROR), CS_HARD);
P
Philipp Reisner 已提交
266 267
}

268

P
Philipp Reisner 已提交
269
/**
270
 * _tl_restart() - Walks the transfer log, and applies an action to all requests
P
Philipp Reisner 已提交
271
 * @mdev:	DRBD device.
272
 * @what:       The action/event to perform with all request objects
P
Philipp Reisner 已提交
273
 *
274 275
 * @what might be one of CONNECTION_LOST_WHILE_PENDING, RESEND, FAIL_FROZEN_DISK_IO,
 * RESTART_FROZEN_DISK_IO.
P
Philipp Reisner 已提交
276
 */
277
/* must hold resource->req_lock */
278
void _tl_restart(struct drbd_tconn *tconn, enum drbd_req_event what)
P
Philipp Reisner 已提交
279
{
280
	struct drbd_request *req, *r;
281

282 283 284 285 286 287 288 289 290
	list_for_each_entry_safe(req, r, &tconn->transfer_log, tl_requests)
		_req_mod(req, what);
}

void tl_restart(struct drbd_tconn *tconn, enum drbd_req_event what)
{
	spin_lock_irq(&tconn->req_lock);
	_tl_restart(tconn, what);
	spin_unlock_irq(&tconn->req_lock);
291
}
P
Philipp Reisner 已提交
292 293 294 295 296 297 298 299 300

/**
 * tl_clear() - Clears all requests and &struct drbd_tl_epoch objects out of the TL
 * @mdev:	DRBD device.
 *
 * This is called after the connection to the peer was lost. The storage covered
 * by the requests on the transfer gets marked as our of sync. Called from the
 * receiver thread and the worker thread.
 */
301
void tl_clear(struct drbd_tconn *tconn)
P
Philipp Reisner 已提交
302
{
303
	tl_restart(tconn, CONNECTION_LOST_WHILE_PENDING);
P
Philipp Reisner 已提交
304 305
}

306
/**
307
 * tl_abort_disk_io() - Abort disk I/O for all requests for a certain mdev in the TL
308 309
 * @mdev:	DRBD device.
 */
310
void tl_abort_disk_io(struct drbd_conf *mdev)
311 312
{
	struct drbd_tconn *tconn = mdev->tconn;
313
	struct drbd_request *req, *r;
314 315

	spin_lock_irq(&tconn->req_lock);
316
	list_for_each_entry_safe(req, r, &tconn->transfer_log, tl_requests) {
317 318
		if (!(req->rq_state & RQ_LOCAL_PENDING))
			continue;
319 320 321
		if (req->w.mdev != mdev)
			continue;
		_req_mod(req, ABORT_DISK_IO);
322 323 324 325
	}
	spin_unlock_irq(&tconn->req_lock);
}

P
Philipp Reisner 已提交
326 327 328
static int drbd_thread_setup(void *arg)
{
	struct drbd_thread *thi = (struct drbd_thread *) arg;
329
	struct drbd_tconn *tconn = thi->tconn;
P
Philipp Reisner 已提交
330 331 332
	unsigned long flags;
	int retval;

333
	snprintf(current->comm, sizeof(current->comm), "drbd_%c_%s",
334
		 thi->name[0], thi->tconn->name);
335

P
Philipp Reisner 已提交
336 337 338 339 340
restart:
	retval = thi->function(thi);

	spin_lock_irqsave(&thi->t_lock, flags);

341
	/* if the receiver has been "EXITING", the last thing it did
P
Philipp Reisner 已提交
342 343 344
	 * was set the conn state to "StandAlone",
	 * if now a re-connect request comes in, conn state goes C_UNCONNECTED,
	 * and receiver thread will be "started".
345
	 * drbd_thread_start needs to set "RESTARTING" in that case.
P
Philipp Reisner 已提交
346
	 * t_state check and assignment needs to be within the same spinlock,
347 348
	 * so either thread_start sees EXITING, and can remap to RESTARTING,
	 * or thread_start see NONE, and can proceed as normal.
P
Philipp Reisner 已提交
349 350
	 */

351
	if (thi->t_state == RESTARTING) {
352
		conn_info(tconn, "Restarting %s thread\n", thi->name);
353
		thi->t_state = RUNNING;
P
Philipp Reisner 已提交
354 355 356 357 358
		spin_unlock_irqrestore(&thi->t_lock, flags);
		goto restart;
	}

	thi->task = NULL;
359
	thi->t_state = NONE;
P
Philipp Reisner 已提交
360
	smp_mb();
L
Lars Ellenberg 已提交
361
	complete_all(&thi->stop);
P
Philipp Reisner 已提交
362 363
	spin_unlock_irqrestore(&thi->t_lock, flags);

364
	conn_info(tconn, "Terminating %s\n", current->comm);
P
Philipp Reisner 已提交
365 366

	/* Release mod reference taken when thread was started */
367 368

	kref_put(&tconn->kref, &conn_destroy);
P
Philipp Reisner 已提交
369 370 371 372
	module_put(THIS_MODULE);
	return retval;
}

373
static void drbd_thread_init(struct drbd_tconn *tconn, struct drbd_thread *thi,
374
			     int (*func) (struct drbd_thread *), char *name)
P
Philipp Reisner 已提交
375 376 377
{
	spin_lock_init(&thi->t_lock);
	thi->task    = NULL;
378
	thi->t_state = NONE;
P
Philipp Reisner 已提交
379
	thi->function = func;
380
	thi->tconn = tconn;
381
	strncpy(thi->name, name, ARRAY_SIZE(thi->name));
P
Philipp Reisner 已提交
382 383 384 385
}

int drbd_thread_start(struct drbd_thread *thi)
{
386
	struct drbd_tconn *tconn = thi->tconn;
P
Philipp Reisner 已提交
387 388 389 390 391 392 393 394
	struct task_struct *nt;
	unsigned long flags;

	/* is used from state engine doing drbd_thread_stop_nowait,
	 * while holding the req lock irqsave */
	spin_lock_irqsave(&thi->t_lock, flags);

	switch (thi->t_state) {
395
	case NONE:
396
		conn_info(tconn, "Starting %s thread (from %s [%d])\n",
397
			 thi->name, current->comm, current->pid);
P
Philipp Reisner 已提交
398 399 400

		/* Get ref on module for thread - this is released when thread exits */
		if (!try_module_get(THIS_MODULE)) {
401
			conn_err(tconn, "Failed to get module reference in drbd_thread_start\n");
P
Philipp Reisner 已提交
402
			spin_unlock_irqrestore(&thi->t_lock, flags);
403
			return false;
P
Philipp Reisner 已提交
404 405
		}

406 407
		kref_get(&thi->tconn->kref);

P
Philipp Reisner 已提交
408 409
		init_completion(&thi->stop);
		thi->reset_cpu_mask = 1;
410
		thi->t_state = RUNNING;
P
Philipp Reisner 已提交
411 412 413 414
		spin_unlock_irqrestore(&thi->t_lock, flags);
		flush_signals(current); /* otherw. may get -ERESTARTNOINTR */

		nt = kthread_create(drbd_thread_setup, (void *) thi,
415
				    "drbd_%c_%s", thi->name[0], thi->tconn->name);
P
Philipp Reisner 已提交
416 417

		if (IS_ERR(nt)) {
418
			conn_err(tconn, "Couldn't start thread\n");
P
Philipp Reisner 已提交
419

420
			kref_put(&tconn->kref, &conn_destroy);
P
Philipp Reisner 已提交
421
			module_put(THIS_MODULE);
422
			return false;
P
Philipp Reisner 已提交
423 424 425
		}
		spin_lock_irqsave(&thi->t_lock, flags);
		thi->task = nt;
426
		thi->t_state = RUNNING;
P
Philipp Reisner 已提交
427 428 429
		spin_unlock_irqrestore(&thi->t_lock, flags);
		wake_up_process(nt);
		break;
430 431
	case EXITING:
		thi->t_state = RESTARTING;
432
		conn_info(tconn, "Restarting %s thread (from %s [%d])\n",
433
				thi->name, current->comm, current->pid);
P
Philipp Reisner 已提交
434
		/* fall through */
435 436
	case RUNNING:
	case RESTARTING:
P
Philipp Reisner 已提交
437 438 439 440 441
	default:
		spin_unlock_irqrestore(&thi->t_lock, flags);
		break;
	}

442
	return true;
P
Philipp Reisner 已提交
443 444 445 446 447 448 449
}


void _drbd_thread_stop(struct drbd_thread *thi, int restart, int wait)
{
	unsigned long flags;

450
	enum drbd_thread_state ns = restart ? RESTARTING : EXITING;
P
Philipp Reisner 已提交
451 452 453 454

	/* may be called from state engine, holding the req lock irqsave */
	spin_lock_irqsave(&thi->t_lock, flags);

455
	if (thi->t_state == NONE) {
P
Philipp Reisner 已提交
456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480
		spin_unlock_irqrestore(&thi->t_lock, flags);
		if (restart)
			drbd_thread_start(thi);
		return;
	}

	if (thi->t_state != ns) {
		if (thi->task == NULL) {
			spin_unlock_irqrestore(&thi->t_lock, flags);
			return;
		}

		thi->t_state = ns;
		smp_mb();
		init_completion(&thi->stop);
		if (thi->task != current)
			force_sig(DRBD_SIGKILL, thi->task);
	}

	spin_unlock_irqrestore(&thi->t_lock, flags);

	if (wait)
		wait_for_completion(&thi->stop);
}

481
static struct drbd_thread *drbd_task_to_thread(struct drbd_tconn *tconn, struct task_struct *task)
482 483 484 485 486 487 488 489 490
{
	struct drbd_thread *thi =
		task == tconn->receiver.task ? &tconn->receiver :
		task == tconn->asender.task  ? &tconn->asender :
		task == tconn->worker.task   ? &tconn->worker : NULL;

	return thi;
}

491
char *drbd_task_to_thread_name(struct drbd_tconn *tconn, struct task_struct *task)
492
{
493
	struct drbd_thread *thi = drbd_task_to_thread(tconn, task);
494 495 496
	return thi ? thi->name : task->comm;
}

497
int conn_lowest_minor(struct drbd_tconn *tconn)
498
{
P
Philipp Reisner 已提交
499
	struct drbd_conf *mdev;
500
	int vnr = 0, m;
501

502
	rcu_read_lock();
P
Philipp Reisner 已提交
503
	mdev = idr_get_next(&tconn->volumes, &vnr);
504 505 506 507
	m = mdev ? mdev_to_minor(mdev) : -1;
	rcu_read_unlock();

	return m;
508
}
509 510

#ifdef CONFIG_SMP
P
Philipp Reisner 已提交
511 512 513 514 515 516 517
/**
 * drbd_calc_cpu_mask() - Generate CPU masks, spread over all CPUs
 * @mdev:	DRBD device.
 *
 * Forces all threads of a device onto the same CPU. This is beneficial for
 * DRBD's performance. May be overwritten by user's configuration.
 */
518
void drbd_calc_cpu_mask(struct drbd_tconn *tconn)
P
Philipp Reisner 已提交
519 520 521 522
{
	int ord, cpu;

	/* user override. */
523
	if (cpumask_weight(tconn->cpu_mask))
P
Philipp Reisner 已提交
524 525
		return;

526
	ord = conn_lowest_minor(tconn) % cpumask_weight(cpu_online_mask);
P
Philipp Reisner 已提交
527 528
	for_each_online_cpu(cpu) {
		if (ord-- == 0) {
529
			cpumask_set_cpu(cpu, tconn->cpu_mask);
P
Philipp Reisner 已提交
530 531 532 533
			return;
		}
	}
	/* should not be reached */
534
	cpumask_setall(tconn->cpu_mask);
P
Philipp Reisner 已提交
535 536 537 538 539
}

/**
 * drbd_thread_current_set_cpu() - modifies the cpu mask of the _current_ thread
 * @mdev:	DRBD device.
540
 * @thi:	drbd_thread object
P
Philipp Reisner 已提交
541 542 543 544
 *
 * call in the "main loop" of _all_ threads, no need for any mutex, current won't die
 * prematurely.
 */
545
void drbd_thread_current_set_cpu(struct drbd_thread *thi)
P
Philipp Reisner 已提交
546 547
{
	struct task_struct *p = current;
548

P
Philipp Reisner 已提交
549 550 551
	if (!thi->reset_cpu_mask)
		return;
	thi->reset_cpu_mask = 0;
552
	set_cpus_allowed_ptr(p, thi->tconn->cpu_mask);
P
Philipp Reisner 已提交
553 554 555
}
#endif

556 557 558 559 560 561 562 563 564
/**
 * drbd_header_size  -  size of a packet header
 *
 * The header size is a multiple of 8, so any payload following the header is
 * word aligned on 64-bit architectures.  (The bitmap send and receive code
 * relies on this.)
 */
unsigned int drbd_header_size(struct drbd_tconn *tconn)
{
565 566 567 568 569 570 571 572 573
	if (tconn->agreed_pro_version >= 100) {
		BUILD_BUG_ON(!IS_ALIGNED(sizeof(struct p_header100), 8));
		return sizeof(struct p_header100);
	} else {
		BUILD_BUG_ON(sizeof(struct p_header80) !=
			     sizeof(struct p_header95));
		BUILD_BUG_ON(!IS_ALIGNED(sizeof(struct p_header80), 8));
		return sizeof(struct p_header80);
	}
574 575
}

576
static unsigned int prepare_header80(struct p_header80 *h, enum drbd_packet cmd, int size)
P
Philipp Reisner 已提交
577 578 579 580
{
	h->magic   = cpu_to_be32(DRBD_MAGIC);
	h->command = cpu_to_be16(cmd);
	h->length  = cpu_to_be16(size);
581
	return sizeof(struct p_header80);
P
Philipp Reisner 已提交
582 583
}

584
static unsigned int prepare_header95(struct p_header95 *h, enum drbd_packet cmd, int size)
P
Philipp Reisner 已提交
585 586 587
{
	h->magic   = cpu_to_be16(DRBD_MAGIC_BIG);
	h->command = cpu_to_be16(cmd);
588
	h->length = cpu_to_be32(size);
589
	return sizeof(struct p_header95);
P
Philipp Reisner 已提交
590 591
}

592 593 594 595 596 597 598 599 600 601 602 603 604
static unsigned int prepare_header100(struct p_header100 *h, enum drbd_packet cmd,
				      int size, int vnr)
{
	h->magic = cpu_to_be32(DRBD_MAGIC_100);
	h->volume = cpu_to_be16(vnr);
	h->command = cpu_to_be16(cmd);
	h->length = cpu_to_be32(size);
	h->pad = 0;
	return sizeof(struct p_header100);
}

static unsigned int prepare_header(struct drbd_tconn *tconn, int vnr,
				   void *buffer, enum drbd_packet cmd, int size)
605
{
606 607 608 609
	if (tconn->agreed_pro_version >= 100)
		return prepare_header100(buffer, cmd, size, vnr);
	else if (tconn->agreed_pro_version >= 95 &&
		 size > DRBD_MAX_SIZE_H80_PACKET)
610
		return prepare_header95(buffer, cmd, size);
611
	else
612
		return prepare_header80(buffer, cmd, size);
613 614
}

615 616 617 618 619 620 621 622
static void *__conn_prepare_command(struct drbd_tconn *tconn,
				    struct drbd_socket *sock)
{
	if (!sock->socket)
		return NULL;
	return sock->sbuf + drbd_header_size(tconn);
}

623 624
void *conn_prepare_command(struct drbd_tconn *tconn, struct drbd_socket *sock)
{
625 626
	void *p;

627
	mutex_lock(&sock->mutex);
628 629
	p = __conn_prepare_command(tconn, sock);
	if (!p)
630
		mutex_unlock(&sock->mutex);
631 632

	return p;
633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656
}

void *drbd_prepare_command(struct drbd_conf *mdev, struct drbd_socket *sock)
{
	return conn_prepare_command(mdev->tconn, sock);
}

static int __send_command(struct drbd_tconn *tconn, int vnr,
			  struct drbd_socket *sock, enum drbd_packet cmd,
			  unsigned int header_size, void *data,
			  unsigned int size)
{
	int msg_flags;
	int err;

	/*
	 * Called with @data == NULL and the size of the data blocks in @size
	 * for commands that send data blocks.  For those commands, omit the
	 * MSG_MORE flag: this will increase the likelihood that data blocks
	 * which are page aligned on the sender will end up page aligned on the
	 * receiver.
	 */
	msg_flags = data ? MSG_MORE : 0;

657 658
	header_size += prepare_header(tconn, vnr, sock->sbuf, cmd,
				      header_size + size);
659 660 661 662 663 664 665
	err = drbd_send_all(tconn, sock->socket, sock->sbuf, header_size,
			    msg_flags);
	if (data && !err)
		err = drbd_send_all(tconn, sock->socket, data, size, 0);
	return err;
}

666 667 668 669 670 671 672
static int __conn_send_command(struct drbd_tconn *tconn, struct drbd_socket *sock,
			       enum drbd_packet cmd, unsigned int header_size,
			       void *data, unsigned int size)
{
	return __send_command(tconn, 0, sock, cmd, header_size, data, size);
}

673 674 675 676 677 678
int conn_send_command(struct drbd_tconn *tconn, struct drbd_socket *sock,
		      enum drbd_packet cmd, unsigned int header_size,
		      void *data, unsigned int size)
{
	int err;

679
	err = __conn_send_command(tconn, sock, cmd, header_size, data, size);
680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695
	mutex_unlock(&sock->mutex);
	return err;
}

int drbd_send_command(struct drbd_conf *mdev, struct drbd_socket *sock,
		      enum drbd_packet cmd, unsigned int header_size,
		      void *data, unsigned int size)
{
	int err;

	err = __send_command(mdev->tconn, mdev->vnr, sock, cmd, header_size,
			     data, size);
	mutex_unlock(&sock->mutex);
	return err;
}

696 697
int drbd_send_ping(struct drbd_tconn *tconn)
{
698 699 700 701 702
	struct drbd_socket *sock;

	sock = &tconn->meta;
	if (!conn_prepare_command(tconn, sock))
		return -EIO;
703
	return conn_send_command(tconn, sock, P_PING, 0, NULL, 0);
704 705 706 707
}

int drbd_send_ping_ack(struct drbd_tconn *tconn)
{
708 709 710 711 712
	struct drbd_socket *sock;

	sock = &tconn->meta;
	if (!conn_prepare_command(tconn, sock))
		return -EIO;
713
	return conn_send_command(tconn, sock, P_PING_ACK, 0, NULL, 0);
714 715
}

716
int drbd_send_sync_param(struct drbd_conf *mdev)
P
Philipp Reisner 已提交
717
{
718
	struct drbd_socket *sock;
719 720
	struct p_rs_param_95 *p;
	int size;
721
	const int apv = mdev->tconn->agreed_pro_version;
722
	enum drbd_packet cmd;
723
	struct net_conf *nc;
P
Philipp Reisner 已提交
724
	struct disk_conf *dc;
725 726 727 728 729

	sock = &mdev->tconn->data;
	p = drbd_prepare_command(mdev, sock);
	if (!p)
		return -EIO;
P
Philipp Reisner 已提交
730

731 732 733
	rcu_read_lock();
	nc = rcu_dereference(mdev->tconn->net_conf);

P
Philipp Reisner 已提交
734 735
	size = apv <= 87 ? sizeof(struct p_rs_param)
		: apv == 88 ? sizeof(struct p_rs_param)
736
			+ strlen(nc->verify_alg) + 1
737 738
		: apv <= 94 ? sizeof(struct p_rs_param_89)
		: /* apv >= 95 */ sizeof(struct p_rs_param_95);
P
Philipp Reisner 已提交
739

740
	cmd = apv >= 89 ? P_SYNC_PARAM89 : P_SYNC_PARAM;
P
Philipp Reisner 已提交
741

742 743
	/* initialize verify_alg and csums_alg */
	memset(p->verify_alg, 0, 2 * SHARED_SECRET_MAX);
P
Philipp Reisner 已提交
744

745
	if (get_ldev(mdev)) {
P
Philipp Reisner 已提交
746
		dc = rcu_dereference(mdev->ldev->disk_conf);
747
		p->resync_rate = cpu_to_be32(dc->resync_rate);
P
Philipp Reisner 已提交
748 749 750 751
		p->c_plan_ahead = cpu_to_be32(dc->c_plan_ahead);
		p->c_delay_target = cpu_to_be32(dc->c_delay_target);
		p->c_fill_target = cpu_to_be32(dc->c_fill_target);
		p->c_max_rate = cpu_to_be32(dc->c_max_rate);
752 753
		put_ldev(mdev);
	} else {
754
		p->resync_rate = cpu_to_be32(DRBD_RESYNC_RATE_DEF);
755 756 757 758 759
		p->c_plan_ahead = cpu_to_be32(DRBD_C_PLAN_AHEAD_DEF);
		p->c_delay_target = cpu_to_be32(DRBD_C_DELAY_TARGET_DEF);
		p->c_fill_target = cpu_to_be32(DRBD_C_FILL_TARGET_DEF);
		p->c_max_rate = cpu_to_be32(DRBD_C_MAX_RATE_DEF);
	}
P
Philipp Reisner 已提交
760

761
	if (apv >= 88)
762
		strcpy(p->verify_alg, nc->verify_alg);
763
	if (apv >= 89)
764 765
		strcpy(p->csums_alg, nc->csums_alg);
	rcu_read_unlock();
P
Philipp Reisner 已提交
766

767
	return drbd_send_command(mdev, sock, cmd, size, NULL, 0);
P
Philipp Reisner 已提交
768 769
}

770
int __drbd_send_protocol(struct drbd_tconn *tconn, enum drbd_packet cmd)
P
Philipp Reisner 已提交
771
{
772
	struct drbd_socket *sock;
P
Philipp Reisner 已提交
773
	struct p_protocol *p;
774
	struct net_conf *nc;
775
	int size, cf;
P
Philipp Reisner 已提交
776

777
	sock = &tconn->data;
778
	p = __conn_prepare_command(tconn, sock);
779 780 781
	if (!p)
		return -EIO;

782 783 784
	rcu_read_lock();
	nc = rcu_dereference(tconn->net_conf);

785
	if (nc->tentative && tconn->agreed_pro_version < 92) {
786 787 788 789 790 791
		rcu_read_unlock();
		mutex_unlock(&sock->mutex);
		conn_err(tconn, "--dry-run is not supported by peer");
		return -EOPNOTSUPP;
	}

792
	size = sizeof(*p);
793
	if (tconn->agreed_pro_version >= 87)
794
		size += strlen(nc->integrity_alg) + 1;
P
Philipp Reisner 已提交
795

796 797 798 799 800
	p->protocol      = cpu_to_be32(nc->wire_protocol);
	p->after_sb_0p   = cpu_to_be32(nc->after_sb_0p);
	p->after_sb_1p   = cpu_to_be32(nc->after_sb_1p);
	p->after_sb_2p   = cpu_to_be32(nc->after_sb_2p);
	p->two_primaries = cpu_to_be32(nc->two_primaries);
801
	cf = 0;
802 803
	if (nc->discard_my_data)
		cf |= CF_DISCARD_MY_DATA;
804
	if (nc->tentative)
805
		cf |= CF_DRY_RUN;
806 807
	p->conn_flags    = cpu_to_be32(cf);

808
	if (tconn->agreed_pro_version >= 87)
809 810 811
		strcpy(p->integrity_alg, nc->integrity_alg);
	rcu_read_unlock();

812
	return __conn_send_command(tconn, sock, cmd, size, NULL, 0);
813 814 815 816 817 818 819
}

int drbd_send_protocol(struct drbd_tconn *tconn)
{
	int err;

	mutex_lock(&tconn->data.mutex);
820
	err = __drbd_send_protocol(tconn, P_PROTOCOL);
821 822 823
	mutex_unlock(&tconn->data.mutex);

	return err;
P
Philipp Reisner 已提交
824 825 826 827
}

int _drbd_send_uuids(struct drbd_conf *mdev, u64 uuid_flags)
{
828 829
	struct drbd_socket *sock;
	struct p_uuids *p;
P
Philipp Reisner 已提交
830 831 832
	int i;

	if (!get_ldev_if_state(mdev, D_NEGOTIATING))
833
		return 0;
P
Philipp Reisner 已提交
834

835 836 837 838 839 840
	sock = &mdev->tconn->data;
	p = drbd_prepare_command(mdev, sock);
	if (!p) {
		put_ldev(mdev);
		return -EIO;
	}
P
Philipp Reisner 已提交
841
	for (i = UI_CURRENT; i < UI_SIZE; i++)
842
		p->uuid[i] = mdev->ldev ? cpu_to_be64(mdev->ldev->md.uuid[i]) : 0;
P
Philipp Reisner 已提交
843 844

	mdev->comm_bm_set = drbd_bm_total_weight(mdev);
845
	p->uuid[UI_SIZE] = cpu_to_be64(mdev->comm_bm_set);
846
	rcu_read_lock();
847
	uuid_flags |= rcu_dereference(mdev->tconn->net_conf)->discard_my_data ? 1 : 0;
848
	rcu_read_unlock();
P
Philipp Reisner 已提交
849 850
	uuid_flags |= test_bit(CRASHED_PRIMARY, &mdev->flags) ? 2 : 0;
	uuid_flags |= mdev->new_state_tmp.disk == D_INCONSISTENT ? 4 : 0;
851
	p->uuid[UI_FLAGS] = cpu_to_be64(uuid_flags);
P
Philipp Reisner 已提交
852 853

	put_ldev(mdev);
854
	return drbd_send_command(mdev, sock, P_UUIDS, sizeof(*p), NULL, 0);
P
Philipp Reisner 已提交
855 856 857 858 859 860 861 862 863 864 865 866
}

int drbd_send_uuids(struct drbd_conf *mdev)
{
	return _drbd_send_uuids(mdev, 0);
}

int drbd_send_uuids_skip_initial_sync(struct drbd_conf *mdev)
{
	return _drbd_send_uuids(mdev, 8);
}

867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884
void drbd_print_uuids(struct drbd_conf *mdev, const char *text)
{
	if (get_ldev_if_state(mdev, D_NEGOTIATING)) {
		u64 *uuid = mdev->ldev->md.uuid;
		dev_info(DEV, "%s %016llX:%016llX:%016llX:%016llX\n",
		     text,
		     (unsigned long long)uuid[UI_CURRENT],
		     (unsigned long long)uuid[UI_BITMAP],
		     (unsigned long long)uuid[UI_HISTORY_START],
		     (unsigned long long)uuid[UI_HISTORY_END]);
		put_ldev(mdev);
	} else {
		dev_info(DEV, "%s effective data uuid: %016llX\n",
				text,
				(unsigned long long)mdev->ed_uuid);
	}
}

885
void drbd_gen_and_send_sync_uuid(struct drbd_conf *mdev)
P
Philipp Reisner 已提交
886
{
887 888
	struct drbd_socket *sock;
	struct p_rs_uuid *p;
889 890 891
	u64 uuid;

	D_ASSERT(mdev->state.disk == D_UP_TO_DATE);
P
Philipp Reisner 已提交
892

893 894 895 896 897
	uuid = mdev->ldev->md.uuid[UI_BITMAP];
	if (uuid && uuid != UUID_JUST_CREATED)
		uuid = uuid + UUID_NEW_BM_OFFSET;
	else
		get_random_bytes(&uuid, sizeof(u64));
898
	drbd_uuid_set(mdev, UI_BITMAP, uuid);
899
	drbd_print_uuids(mdev, "updated sync UUID");
900
	drbd_md_sync(mdev);
P
Philipp Reisner 已提交
901

902 903 904 905 906 907
	sock = &mdev->tconn->data;
	p = drbd_prepare_command(mdev, sock);
	if (p) {
		p->uuid = cpu_to_be64(uuid);
		drbd_send_command(mdev, sock, P_SYNC_UUID, sizeof(*p), NULL, 0);
	}
P
Philipp Reisner 已提交
908 909
}

910
int drbd_send_sizes(struct drbd_conf *mdev, int trigger_reply, enum dds_flags flags)
P
Philipp Reisner 已提交
911
{
912 913
	struct drbd_socket *sock;
	struct p_sizes *p;
P
Philipp Reisner 已提交
914
	sector_t d_size, u_size;
915
	int q_order_type, max_bio_size;
P
Philipp Reisner 已提交
916 917 918 919

	if (get_ldev_if_state(mdev, D_NEGOTIATING)) {
		D_ASSERT(mdev->ldev->backing_bdev);
		d_size = drbd_get_max_capacity(mdev->ldev);
P
Philipp Reisner 已提交
920 921 922
		rcu_read_lock();
		u_size = rcu_dereference(mdev->ldev->disk_conf)->disk_size;
		rcu_read_unlock();
P
Philipp Reisner 已提交
923
		q_order_type = drbd_queue_order_type(mdev);
924 925
		max_bio_size = queue_max_hw_sectors(mdev->ldev->backing_bdev->bd_disk->queue) << 9;
		max_bio_size = min_t(int, max_bio_size, DRBD_MAX_BIO_SIZE);
P
Philipp Reisner 已提交
926 927 928 929 930
		put_ldev(mdev);
	} else {
		d_size = 0;
		u_size = 0;
		q_order_type = QUEUE_ORDERED_NONE;
931
		max_bio_size = DRBD_MAX_BIO_SIZE; /* ... multiple BIOs per peer_request */
P
Philipp Reisner 已提交
932 933
	}

934 935 936 937
	sock = &mdev->tconn->data;
	p = drbd_prepare_command(mdev, sock);
	if (!p)
		return -EIO;
938 939 940 941 942 943

	if (mdev->tconn->agreed_pro_version <= 94)
		max_bio_size = min_t(int, max_bio_size, DRBD_MAX_SIZE_H80_PACKET);
	else if (mdev->tconn->agreed_pro_version < 100)
		max_bio_size = min_t(int, max_bio_size, DRBD_MAX_BIO_SIZE_P95);

944 945 946 947 948 949 950
	p->d_size = cpu_to_be64(d_size);
	p->u_size = cpu_to_be64(u_size);
	p->c_size = cpu_to_be64(trigger_reply ? 0 : drbd_get_capacity(mdev->this_bdev));
	p->max_bio_size = cpu_to_be32(max_bio_size);
	p->queue_order_type = cpu_to_be16(q_order_type);
	p->dds_flags = cpu_to_be16(flags);
	return drbd_send_command(mdev, sock, P_SIZES, sizeof(*p), NULL, 0);
P
Philipp Reisner 已提交
951 952 953
}

/**
954
 * drbd_send_current_state() - Sends the drbd state to the peer
P
Philipp Reisner 已提交
955 956
 * @mdev:	DRBD device.
 */
957
int drbd_send_current_state(struct drbd_conf *mdev)
P
Philipp Reisner 已提交
958
{
959
	struct drbd_socket *sock;
960
	struct p_state *p;
P
Philipp Reisner 已提交
961

962
	sock = &mdev->tconn->data;
963 964 965 966 967 968
	p = drbd_prepare_command(mdev, sock);
	if (!p)
		return -EIO;
	p->state = cpu_to_be32(mdev->state.i); /* Within the send mutex */
	return drbd_send_command(mdev, sock, P_STATE, sizeof(*p), NULL, 0);
}
P
Philipp Reisner 已提交
969

970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992
/**
 * drbd_send_state() - After a state change, sends the new state to the peer
 * @mdev:      DRBD device.
 * @state:     the state to send, not necessarily the current state.
 *
 * Each state change queues an "after_state_ch" work, which will eventually
 * send the resulting new state to the peer. If more state changes happen
 * between queuing and processing of the after_state_ch work, we still
 * want to send each intermediary state in the order it occurred.
 */
int drbd_send_state(struct drbd_conf *mdev, union drbd_state state)
{
	struct drbd_socket *sock;
	struct p_state *p;

	sock = &mdev->tconn->data;
	p = drbd_prepare_command(mdev, sock);
	if (!p)
		return -EIO;
	p->state = cpu_to_be32(state.i); /* Within the send mutex */
	return drbd_send_command(mdev, sock, P_STATE, sizeof(*p), NULL, 0);
}

993 994 995 996
int drbd_send_state_req(struct drbd_conf *mdev, union drbd_state mask, union drbd_state val)
{
	struct drbd_socket *sock;
	struct p_req_state *p;
P
Philipp Reisner 已提交
997

998 999 1000 1001 1002 1003 1004
	sock = &mdev->tconn->data;
	p = drbd_prepare_command(mdev, sock);
	if (!p)
		return -EIO;
	p->mask = cpu_to_be32(mask.i);
	p->val = cpu_to_be32(val.i);
	return drbd_send_command(mdev, sock, P_STATE_CHG_REQ, sizeof(*p), NULL, 0);
P
Philipp Reisner 已提交
1005 1006
}

1007
int conn_send_state_req(struct drbd_tconn *tconn, union drbd_state mask, union drbd_state val)
P
Philipp Reisner 已提交
1008
{
1009 1010 1011
	enum drbd_packet cmd;
	struct drbd_socket *sock;
	struct p_req_state *p;
P
Philipp Reisner 已提交
1012

1013 1014 1015 1016 1017 1018 1019 1020
	cmd = tconn->agreed_pro_version < 100 ? P_STATE_CHG_REQ : P_CONN_ST_CHG_REQ;
	sock = &tconn->data;
	p = conn_prepare_command(tconn, sock);
	if (!p)
		return -EIO;
	p->mask = cpu_to_be32(mask.i);
	p->val = cpu_to_be32(val.i);
	return conn_send_command(tconn, sock, cmd, sizeof(*p), NULL, 0);
P
Philipp Reisner 已提交
1021 1022
}

1023
void drbd_send_sr_reply(struct drbd_conf *mdev, enum drbd_state_rv retcode)
P
Philipp Reisner 已提交
1024
{
1025 1026
	struct drbd_socket *sock;
	struct p_req_state_reply *p;
P
Philipp Reisner 已提交
1027

1028 1029 1030 1031 1032 1033
	sock = &mdev->tconn->meta;
	p = drbd_prepare_command(mdev, sock);
	if (p) {
		p->retcode = cpu_to_be32(retcode);
		drbd_send_command(mdev, sock, P_STATE_CHG_REPLY, sizeof(*p), NULL, 0);
	}
P
Philipp Reisner 已提交
1034 1035
}

1036
void conn_send_sr_reply(struct drbd_tconn *tconn, enum drbd_state_rv retcode)
1037
{
1038 1039
	struct drbd_socket *sock;
	struct p_req_state_reply *p;
1040 1041
	enum drbd_packet cmd = tconn->agreed_pro_version < 100 ? P_STATE_CHG_REPLY : P_CONN_ST_CHG_REPLY;

1042 1043 1044 1045 1046 1047
	sock = &tconn->meta;
	p = conn_prepare_command(tconn, sock);
	if (p) {
		p->retcode = cpu_to_be32(retcode);
		conn_send_command(tconn, sock, cmd, sizeof(*p), NULL, 0);
	}
1048 1049
}

1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066
static void dcbp_set_code(struct p_compressed_bm *p, enum drbd_bitmap_code code)
{
	BUG_ON(code & ~0xf);
	p->encoding = (p->encoding & ~0xf) | code;
}

static void dcbp_set_start(struct p_compressed_bm *p, int set)
{
	p->encoding = (p->encoding & ~0x80) | (set ? 0x80 : 0);
}

static void dcbp_set_pad_bits(struct p_compressed_bm *p, int n)
{
	BUG_ON(n & ~0x7);
	p->encoding = (p->encoding & (~0x7 << 4)) | (n << 4);
}

P
Philipp Reisner 已提交
1067
int fill_bitmap_rle_bits(struct drbd_conf *mdev,
1068 1069 1070
			 struct p_compressed_bm *p,
			 unsigned int size,
			 struct bm_xfer_ctx *c)
P
Philipp Reisner 已提交
1071 1072 1073 1074 1075 1076 1077
{
	struct bitstream bs;
	unsigned long plain_bits;
	unsigned long tmp;
	unsigned long rl;
	unsigned len;
	unsigned toggle;
1078
	int bits, use_rle;
P
Philipp Reisner 已提交
1079 1080

	/* may we use this feature? */
1081 1082 1083 1084 1085
	rcu_read_lock();
	use_rle = rcu_dereference(mdev->tconn->net_conf)->use_rle;
	rcu_read_unlock();
	if (!use_rle || mdev->tconn->agreed_pro_version < 90)
		return 0;
P
Philipp Reisner 已提交
1086 1087 1088 1089 1090

	if (c->bit_offset >= c->bm_bits)
		return 0; /* nothing to do. */

	/* use at most thus many bytes */
1091 1092
	bitstream_init(&bs, p->code, size, 0);
	memset(p->code, 0, size);
P
Philipp Reisner 已提交
1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113
	/* plain bits covered in this code string */
	plain_bits = 0;

	/* p->encoding & 0x80 stores whether the first run length is set.
	 * bit offset is implicit.
	 * start with toggle == 2 to be able to tell the first iteration */
	toggle = 2;

	/* see how much plain bits we can stuff into one packet
	 * using RLE and VLI. */
	do {
		tmp = (toggle == 0) ? _drbd_bm_find_next_zero(mdev, c->bit_offset)
				    : _drbd_bm_find_next(mdev, c->bit_offset);
		if (tmp == -1UL)
			tmp = c->bm_bits;
		rl = tmp - c->bit_offset;

		if (toggle == 2) { /* first iteration */
			if (rl == 0) {
				/* the first checked bit was set,
				 * store start value, */
1114
				dcbp_set_start(p, 1);
P
Philipp Reisner 已提交
1115 1116 1117 1118
				/* but skip encoding of zero run length */
				toggle = !toggle;
				continue;
			}
1119
			dcbp_set_start(p, 0);
P
Philipp Reisner 已提交
1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158
		}

		/* paranoia: catch zero runlength.
		 * can only happen if bitmap is modified while we scan it. */
		if (rl == 0) {
			dev_err(DEV, "unexpected zero runlength while encoding bitmap "
			    "t:%u bo:%lu\n", toggle, c->bit_offset);
			return -1;
		}

		bits = vli_encode_bits(&bs, rl);
		if (bits == -ENOBUFS) /* buffer full */
			break;
		if (bits <= 0) {
			dev_err(DEV, "error while encoding bitmap: %d\n", bits);
			return 0;
		}

		toggle = !toggle;
		plain_bits += rl;
		c->bit_offset = tmp;
	} while (c->bit_offset < c->bm_bits);

	len = bs.cur.b - p->code + !!bs.cur.bit;

	if (plain_bits < (len << 3)) {
		/* incompressible with this method.
		 * we need to rewind both word and bit position. */
		c->bit_offset -= plain_bits;
		bm_xfer_ctx_bit_to_word_offset(c);
		c->bit_offset = c->word_offset * BITS_PER_LONG;
		return 0;
	}

	/* RLE + VLI was able to compress it just fine.
	 * update c->word_offset. */
	bm_xfer_ctx_bit_to_word_offset(c);

	/* store pad_bits */
1159
	dcbp_set_pad_bits(p, (8 - bs.cur.bit) & 0x7);
P
Philipp Reisner 已提交
1160 1161 1162 1163

	return len;
}

1164 1165 1166 1167 1168 1169 1170
/**
 * send_bitmap_rle_or_plain
 *
 * Return 0 when done, 1 when another iteration is needed, and a negative error
 * code upon failure.
 */
static int
1171
send_bitmap_rle_or_plain(struct drbd_conf *mdev, struct bm_xfer_ctx *c)
P
Philipp Reisner 已提交
1172
{
1173
	struct drbd_socket *sock = &mdev->tconn->data;
1174
	unsigned int header_size = drbd_header_size(mdev->tconn);
1175
	struct p_compressed_bm *p = sock->sbuf + header_size;
1176
	int len, err;
P
Philipp Reisner 已提交
1177

1178 1179
	len = fill_bitmap_rle_bits(mdev, p,
			DRBD_SOCKET_BUFFER_SIZE - header_size - sizeof(*p), c);
P
Philipp Reisner 已提交
1180
	if (len < 0)
1181
		return -EIO;
P
Philipp Reisner 已提交
1182 1183

	if (len) {
1184
		dcbp_set_code(p, RLE_VLI_Bits);
1185 1186 1187
		err = __send_command(mdev->tconn, mdev->vnr, sock,
				     P_COMPRESSED_BITMAP, sizeof(*p) + len,
				     NULL, 0);
P
Philipp Reisner 已提交
1188
		c->packets[0]++;
1189
		c->bytes[0] += header_size + sizeof(*p) + len;
P
Philipp Reisner 已提交
1190 1191 1192 1193 1194 1195

		if (c->bit_offset >= c->bm_bits)
			len = 0; /* DONE */
	} else {
		/* was not compressible.
		 * send a buffer full of plain text bits instead. */
1196 1197
		unsigned int data_size;
		unsigned long num_words;
1198
		unsigned long *p = sock->sbuf + header_size;
1199 1200

		data_size = DRBD_SOCKET_BUFFER_SIZE - header_size;
1201
		num_words = min_t(size_t, data_size / sizeof(*p),
1202
				  c->bm_words - c->word_offset);
1203
		len = num_words * sizeof(*p);
P
Philipp Reisner 已提交
1204
		if (len)
1205 1206
			drbd_bm_get_lel(mdev, c->word_offset, num_words, p);
		err = __send_command(mdev->tconn, mdev->vnr, sock, P_BITMAP, len, NULL, 0);
P
Philipp Reisner 已提交
1207 1208 1209 1210
		c->word_offset += num_words;
		c->bit_offset = c->word_offset * BITS_PER_LONG;

		c->packets[1]++;
1211
		c->bytes[1] += header_size + len;
P
Philipp Reisner 已提交
1212 1213 1214 1215

		if (c->bit_offset > c->bm_bits)
			c->bit_offset = c->bm_bits;
	}
1216
	if (!err) {
1217 1218 1219 1220 1221 1222 1223
		if (len == 0) {
			INFO_bm_xfer_stats(mdev, "send", c);
			return 0;
		} else
			return 1;
	}
	return -EIO;
P
Philipp Reisner 已提交
1224 1225 1226
}

/* See the comment at receive_bitmap() */
1227
static int _drbd_send_bitmap(struct drbd_conf *mdev)
P
Philipp Reisner 已提交
1228 1229
{
	struct bm_xfer_ctx c;
1230
	int err;
P
Philipp Reisner 已提交
1231

1232 1233
	if (!expect(mdev->bitmap))
		return false;
P
Philipp Reisner 已提交
1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257

	if (get_ldev(mdev)) {
		if (drbd_md_test_flag(mdev->ldev, MDF_FULL_SYNC)) {
			dev_info(DEV, "Writing the whole bitmap, MDF_FullSync was set.\n");
			drbd_bm_set_all(mdev);
			if (drbd_bm_write(mdev)) {
				/* write_bm did fail! Leave full sync flag set in Meta P_DATA
				 * but otherwise process as per normal - need to tell other
				 * side that a full resync is required! */
				dev_err(DEV, "Failed to write bitmap to disk!\n");
			} else {
				drbd_md_clear_flag(mdev, MDF_FULL_SYNC);
				drbd_md_sync(mdev);
			}
		}
		put_ldev(mdev);
	}

	c = (struct bm_xfer_ctx) {
		.bm_bits = drbd_bm_bits(mdev),
		.bm_words = drbd_bm_words(mdev),
	};

	do {
1258
		err = send_bitmap_rle_or_plain(mdev, &c);
1259
	} while (err > 0);
P
Philipp Reisner 已提交
1260

1261
	return err == 0;
P
Philipp Reisner 已提交
1262 1263 1264 1265
}

int drbd_send_bitmap(struct drbd_conf *mdev)
{
1266 1267
	struct drbd_socket *sock = &mdev->tconn->data;
	int err = -1;
P
Philipp Reisner 已提交
1268

1269 1270 1271 1272
	mutex_lock(&sock->mutex);
	if (sock->socket)
		err = !_drbd_send_bitmap(mdev);
	mutex_unlock(&sock->mutex);
P
Philipp Reisner 已提交
1273 1274
	return err;
}
1275

1276
void drbd_send_b_ack(struct drbd_tconn *tconn, u32 barrier_nr, u32 set_size)
P
Philipp Reisner 已提交
1277
{
1278 1279
	struct drbd_socket *sock;
	struct p_barrier_ack *p;
P
Philipp Reisner 已提交
1280

1281
	if (tconn->cstate < C_WF_REPORT_PARAMS)
1282
		return;
P
Philipp Reisner 已提交
1283

1284 1285
	sock = &tconn->meta;
	p = conn_prepare_command(tconn, sock);
1286 1287 1288 1289
	if (!p)
		return;
	p->barrier = barrier_nr;
	p->set_size = cpu_to_be32(set_size);
1290
	conn_send_command(tconn, sock, P_BARRIER_ACK, sizeof(*p), NULL, 0);
P
Philipp Reisner 已提交
1291 1292 1293 1294 1295 1296 1297 1298 1299 1300
}

/**
 * _drbd_send_ack() - Sends an ack packet
 * @mdev:	DRBD device.
 * @cmd:	Packet command code.
 * @sector:	sector, needs to be in big endian byte order
 * @blksize:	size in byte, needs to be in big endian byte order
 * @block_id:	Id, big endian byte order
 */
1301 1302
static int _drbd_send_ack(struct drbd_conf *mdev, enum drbd_packet cmd,
			  u64 sector, u32 blksize, u64 block_id)
P
Philipp Reisner 已提交
1303
{
1304 1305
	struct drbd_socket *sock;
	struct p_block_ack *p;
P
Philipp Reisner 已提交
1306

1307 1308
	if (mdev->state.conn < C_CONNECTED)
		return -EIO;
P
Philipp Reisner 已提交
1309

1310 1311 1312
	sock = &mdev->tconn->meta;
	p = drbd_prepare_command(mdev, sock);
	if (!p)
1313
		return -EIO;
1314 1315 1316 1317 1318
	p->sector = sector;
	p->block_id = block_id;
	p->blksize = blksize;
	p->seq_num = cpu_to_be32(atomic_inc_return(&mdev->packet_seq));
	return drbd_send_command(mdev, sock, cmd, sizeof(*p), NULL, 0);
P
Philipp Reisner 已提交
1319 1320
}

1321 1322 1323
/* dp->sector and dp->block_id already/still in network byte order,
 * data_size is payload size according to dp->head,
 * and may need to be corrected for digest size. */
1324 1325
void drbd_send_ack_dp(struct drbd_conf *mdev, enum drbd_packet cmd,
		      struct p_data *dp, int data_size)
P
Philipp Reisner 已提交
1326
{
1327 1328
	if (mdev->tconn->peer_integrity_tfm)
		data_size -= crypto_hash_digestsize(mdev->tconn->peer_integrity_tfm);
1329 1330
	_drbd_send_ack(mdev, cmd, dp->sector, cpu_to_be32(data_size),
		       dp->block_id);
P
Philipp Reisner 已提交
1331 1332
}

1333 1334
void drbd_send_ack_rp(struct drbd_conf *mdev, enum drbd_packet cmd,
		      struct p_block_req *rp)
P
Philipp Reisner 已提交
1335
{
1336
	_drbd_send_ack(mdev, cmd, rp->sector, rp->blksize, rp->block_id);
P
Philipp Reisner 已提交
1337 1338 1339 1340
}

/**
 * drbd_send_ack() - Sends an ack packet
1341 1342 1343
 * @mdev:	DRBD device
 * @cmd:	packet command code
 * @peer_req:	peer request
P
Philipp Reisner 已提交
1344
 */
1345
int drbd_send_ack(struct drbd_conf *mdev, enum drbd_packet cmd,
1346
		  struct drbd_peer_request *peer_req)
P
Philipp Reisner 已提交
1347
{
1348 1349 1350 1351
	return _drbd_send_ack(mdev, cmd,
			      cpu_to_be64(peer_req->i.sector),
			      cpu_to_be32(peer_req->i.size),
			      peer_req->block_id);
P
Philipp Reisner 已提交
1352 1353 1354 1355
}

/* This function misuses the block_id field to signal if the blocks
 * are is sync or not. */
1356
int drbd_send_ack_ex(struct drbd_conf *mdev, enum drbd_packet cmd,
P
Philipp Reisner 已提交
1357 1358
		     sector_t sector, int blksize, u64 block_id)
{
1359 1360 1361 1362
	return _drbd_send_ack(mdev, cmd,
			      cpu_to_be64(sector),
			      cpu_to_be32(blksize),
			      cpu_to_be64(block_id));
P
Philipp Reisner 已提交
1363 1364 1365 1366 1367
}

int drbd_send_drequest(struct drbd_conf *mdev, int cmd,
		       sector_t sector, int size, u64 block_id)
{
1368 1369
	struct drbd_socket *sock;
	struct p_block_req *p;
P
Philipp Reisner 已提交
1370

1371 1372 1373 1374 1375 1376 1377 1378
	sock = &mdev->tconn->data;
	p = drbd_prepare_command(mdev, sock);
	if (!p)
		return -EIO;
	p->sector = cpu_to_be64(sector);
	p->block_id = block_id;
	p->blksize = cpu_to_be32(size);
	return drbd_send_command(mdev, sock, cmd, sizeof(*p), NULL, 0);
P
Philipp Reisner 已提交
1379 1380
}

1381 1382
int drbd_send_drequest_csum(struct drbd_conf *mdev, sector_t sector, int size,
			    void *digest, int digest_size, enum drbd_packet cmd)
P
Philipp Reisner 已提交
1383
{
1384 1385
	struct drbd_socket *sock;
	struct p_block_req *p;
P
Philipp Reisner 已提交
1386

1387
	/* FIXME: Put the digest into the preallocated socket buffer.  */
P
Philipp Reisner 已提交
1388

1389 1390 1391 1392 1393 1394 1395 1396 1397
	sock = &mdev->tconn->data;
	p = drbd_prepare_command(mdev, sock);
	if (!p)
		return -EIO;
	p->sector = cpu_to_be64(sector);
	p->block_id = ID_SYNCER /* unused */;
	p->blksize = cpu_to_be32(size);
	return drbd_send_command(mdev, sock, cmd, sizeof(*p),
				 digest, digest_size);
P
Philipp Reisner 已提交
1398 1399 1400 1401
}

int drbd_send_ov_request(struct drbd_conf *mdev, sector_t sector, int size)
{
1402 1403
	struct drbd_socket *sock;
	struct p_block_req *p;
P
Philipp Reisner 已提交
1404

1405 1406 1407 1408 1409 1410 1411 1412
	sock = &mdev->tconn->data;
	p = drbd_prepare_command(mdev, sock);
	if (!p)
		return -EIO;
	p->sector = cpu_to_be64(sector);
	p->block_id = ID_SYNCER /* unused */;
	p->blksize = cpu_to_be32(size);
	return drbd_send_command(mdev, sock, P_OV_REQUEST, sizeof(*p), NULL, 0);
P
Philipp Reisner 已提交
1413 1414 1415
}

/* called on sndtimeo
1416 1417
 * returns false if we should retry,
 * true if we think connection is dead
P
Philipp Reisner 已提交
1418
 */
1419
static int we_should_drop_the_connection(struct drbd_tconn *tconn, struct socket *sock)
P
Philipp Reisner 已提交
1420 1421 1422 1423
{
	int drop_it;
	/* long elapsed = (long)(jiffies - mdev->last_received); */

1424 1425 1426
	drop_it =   tconn->meta.socket == sock
		|| !tconn->asender.task
		|| get_t_state(&tconn->asender) != RUNNING
1427
		|| tconn->cstate < C_WF_REPORT_PARAMS;
P
Philipp Reisner 已提交
1428 1429

	if (drop_it)
1430
		return true;
P
Philipp Reisner 已提交
1431

1432
	drop_it = !--tconn->ko_count;
P
Philipp Reisner 已提交
1433
	if (!drop_it) {
1434 1435 1436
		conn_err(tconn, "[%s/%d] sock_sendmsg time expired, ko = %u\n",
			 current->comm, current->pid, tconn->ko_count);
		request_ping(tconn);
P
Philipp Reisner 已提交
1437 1438 1439 1440 1441
	}

	return drop_it; /* && (mdev->state == R_PRIMARY) */;
}

1442
static void drbd_update_congested(struct drbd_tconn *tconn)
1443
{
1444
	struct sock *sk = tconn->data.socket->sk;
1445
	if (sk->sk_wmem_queued > sk->sk_sndbuf * 4 / 5)
1446
		set_bit(NET_CONGESTED, &tconn->flags);
1447 1448
}

P
Philipp Reisner 已提交
1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470
/* The idea of sendpage seems to be to put some kind of reference
 * to the page into the skb, and to hand it over to the NIC. In
 * this process get_page() gets called.
 *
 * As soon as the page was really sent over the network put_page()
 * gets called by some part of the network layer. [ NIC driver? ]
 *
 * [ get_page() / put_page() increment/decrement the count. If count
 *   reaches 0 the page will be freed. ]
 *
 * This works nicely with pages from FSs.
 * But this means that in protocol A we might signal IO completion too early!
 *
 * In order not to corrupt data during a resync we must make sure
 * that we do not reuse our own buffer pages (EEs) to early, therefore
 * we have the net_ee list.
 *
 * XFS seems to have problems, still, it submits pages with page_count == 0!
 * As a workaround, we disable sendpage on pages
 * with page_count == 0 or PageSlab.
 */
static int _drbd_no_send_page(struct drbd_conf *mdev, struct page *page,
1471
			      int offset, size_t size, unsigned msg_flags)
P
Philipp Reisner 已提交
1472
{
1473 1474 1475 1476 1477 1478 1479
	struct socket *socket;
	void *addr;
	int err;

	socket = mdev->tconn->data.socket;
	addr = kmap(page) + offset;
	err = drbd_send_all(mdev->tconn, socket, addr, size, msg_flags);
P
Philipp Reisner 已提交
1480
	kunmap(page);
1481 1482 1483
	if (!err)
		mdev->send_cnt += size >> 9;
	return err;
P
Philipp Reisner 已提交
1484 1485 1486
}

static int _drbd_send_page(struct drbd_conf *mdev, struct page *page,
L
Lars Ellenberg 已提交
1487
		    int offset, size_t size, unsigned msg_flags)
P
Philipp Reisner 已提交
1488
{
1489
	struct socket *socket = mdev->tconn->data.socket;
P
Philipp Reisner 已提交
1490 1491
	mm_segment_t oldfs = get_fs();
	int len = size;
1492
	int err = -EIO;
P
Philipp Reisner 已提交
1493 1494 1495 1496 1497 1498 1499 1500

	/* e.g. XFS meta- & log-data is in slab pages, which have a
	 * page_count of 0 and/or have PageSlab() set.
	 * we cannot use send_page for those, as that does get_page();
	 * put_page(); and would cause either a VM_BUG directly, or
	 * __page_cache_release a page that would actually still be referenced
	 * by someone, leading to some obscure delayed Oops somewhere else. */
	if (disable_sendpage || (page_count(page) < 1) || PageSlab(page))
1501
		return _drbd_no_send_page(mdev, page, offset, size, msg_flags);
P
Philipp Reisner 已提交
1502

L
Lars Ellenberg 已提交
1503
	msg_flags |= MSG_NOSIGNAL;
1504
	drbd_update_congested(mdev->tconn);
P
Philipp Reisner 已提交
1505 1506
	set_fs(KERNEL_DS);
	do {
1507 1508 1509
		int sent;

		sent = socket->ops->sendpage(socket, page, offset, len, msg_flags);
P
Philipp Reisner 已提交
1510
		if (sent <= 0) {
1511 1512 1513 1514 1515
			if (sent == -EAGAIN) {
				if (we_should_drop_the_connection(mdev->tconn, socket))
					break;
				continue;
			}
P
Philipp Reisner 已提交
1516 1517
			dev_warn(DEV, "%s: size=%d len=%d sent=%d\n",
			     __func__, (int)size, len, sent);
1518 1519
			if (sent < 0)
				err = sent;
P
Philipp Reisner 已提交
1520 1521 1522 1523 1524 1525
			break;
		}
		len    -= sent;
		offset += sent;
	} while (len > 0 /* THINK && mdev->cstate >= C_CONNECTED*/);
	set_fs(oldfs);
1526
	clear_bit(NET_CONGESTED, &mdev->tconn->flags);
P
Philipp Reisner 已提交
1527

1528 1529 1530 1531 1532
	if (len == 0) {
		err = 0;
		mdev->send_cnt += size >> 9;
	}
	return err;
P
Philipp Reisner 已提交
1533 1534 1535 1536 1537 1538
}

static int _drbd_send_bio(struct drbd_conf *mdev, struct bio *bio)
{
	struct bio_vec *bvec;
	int i;
L
Lars Ellenberg 已提交
1539
	/* hint all but last page with MSG_MORE */
1540
	bio_for_each_segment(bvec, bio, i) {
1541 1542 1543 1544 1545 1546 1547
		int err;

		err = _drbd_no_send_page(mdev, bvec->bv_page,
					 bvec->bv_offset, bvec->bv_len,
					 i == bio->bi_vcnt - 1 ? 0 : MSG_MORE);
		if (err)
			return err;
P
Philipp Reisner 已提交
1548
	}
1549
	return 0;
P
Philipp Reisner 已提交
1550 1551 1552 1553 1554 1555
}

static int _drbd_send_zc_bio(struct drbd_conf *mdev, struct bio *bio)
{
	struct bio_vec *bvec;
	int i;
L
Lars Ellenberg 已提交
1556
	/* hint all but last page with MSG_MORE */
1557
	bio_for_each_segment(bvec, bio, i) {
1558 1559 1560 1561 1562 1563 1564
		int err;

		err = _drbd_send_page(mdev, bvec->bv_page,
				      bvec->bv_offset, bvec->bv_len,
				      i == bio->bi_vcnt - 1 ? 0 : MSG_MORE);
		if (err)
			return err;
P
Philipp Reisner 已提交
1565
	}
1566
	return 0;
P
Philipp Reisner 已提交
1567 1568
}

1569 1570
static int _drbd_send_zc_ee(struct drbd_conf *mdev,
			    struct drbd_peer_request *peer_req)
1571
{
1572 1573
	struct page *page = peer_req->pages;
	unsigned len = peer_req->i.size;
1574
	int err;
1575

L
Lars Ellenberg 已提交
1576
	/* hint all but last page with MSG_MORE */
1577 1578
	page_chain_for_each(page) {
		unsigned l = min_t(unsigned, len, PAGE_SIZE);
1579 1580 1581 1582 1583

		err = _drbd_send_page(mdev, page, 0, l,
				      page_chain_next(page) ? MSG_MORE : 0);
		if (err)
			return err;
1584 1585
		len -= l;
	}
1586
	return 0;
1587 1588
}

1589 1590
static u32 bio_flags_to_wire(struct drbd_conf *mdev, unsigned long bi_rw)
{
1591
	if (mdev->tconn->agreed_pro_version >= 95)
1592 1593 1594 1595 1596
		return  (bi_rw & REQ_SYNC ? DP_RW_SYNC : 0) |
			(bi_rw & REQ_FUA ? DP_FUA : 0) |
			(bi_rw & REQ_FLUSH ? DP_FLUSH : 0) |
			(bi_rw & REQ_DISCARD ? DP_DISCARD : 0);
	else
J
Jens Axboe 已提交
1597
		return bi_rw & REQ_SYNC ? DP_RW_SYNC : 0;
1598 1599
}

P
Philipp Reisner 已提交
1600 1601 1602 1603 1604
/* Used to send write requests
 * R_PRIMARY -> Peer	(P_DATA)
 */
int drbd_send_dblock(struct drbd_conf *mdev, struct drbd_request *req)
{
1605 1606
	struct drbd_socket *sock;
	struct p_data *p;
P
Philipp Reisner 已提交
1607 1608
	unsigned int dp_flags = 0;
	int dgs;
1609
	int err;
P
Philipp Reisner 已提交
1610

1611 1612
	sock = &mdev->tconn->data;
	p = drbd_prepare_command(mdev, sock);
1613
	dgs = mdev->tconn->integrity_tfm ? crypto_hash_digestsize(mdev->tconn->integrity_tfm) : 0;
P
Philipp Reisner 已提交
1614

1615 1616 1617 1618
	if (!p)
		return -EIO;
	p->sector = cpu_to_be64(req->i.sector);
	p->block_id = (unsigned long)req;
1619
	p->seq_num = cpu_to_be32(atomic_inc_return(&mdev->packet_seq));
1620
	dp_flags = bio_flags_to_wire(mdev, req->master_bio->bi_rw);
P
Philipp Reisner 已提交
1621 1622 1623
	if (mdev->state.conn >= C_SYNC_SOURCE &&
	    mdev->state.conn <= C_PAUSED_SYNC_T)
		dp_flags |= DP_MAY_SET_IN_SYNC;
1624 1625 1626 1627 1628 1629
	if (mdev->tconn->agreed_pro_version >= 100) {
		if (req->rq_state & RQ_EXP_RECEIVE_ACK)
			dp_flags |= DP_SEND_RECEIVE_ACK;
		if (req->rq_state & RQ_EXP_WRITE_ACK)
			dp_flags |= DP_SEND_WRITE_ACK;
	}
1630 1631
	p->dp_flags = cpu_to_be32(dp_flags);
	if (dgs)
1632
		drbd_csum_bio(mdev, mdev->tconn->integrity_tfm, req->master_bio, p + 1);
1633
	err = __send_command(mdev->tconn, mdev->vnr, sock, P_DATA, sizeof(*p) + dgs, NULL, req->i.size);
1634
	if (!err) {
1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645
		/* For protocol A, we have to memcpy the payload into
		 * socket buffers, as we may complete right away
		 * as soon as we handed it over to tcp, at which point the data
		 * pages may become invalid.
		 *
		 * For data-integrity enabled, we copy it as well, so we can be
		 * sure that even if the bio pages may still be modified, it
		 * won't change the data on the wire, thus if the digest checks
		 * out ok after sending on this side, but does not fit on the
		 * receiving side, we sure have detected corruption elsewhere.
		 */
1646
		if (!(req->rq_state & (RQ_EXP_RECEIVE_ACK | RQ_EXP_WRITE_ACK)) || dgs)
1647
			err = _drbd_send_bio(mdev, req->master_bio);
P
Philipp Reisner 已提交
1648
		else
1649
			err = _drbd_send_zc_bio(mdev, req->master_bio);
1650 1651 1652

		/* double check digest, sometimes buffers have been modified in flight. */
		if (dgs > 0 && dgs <= 64) {
B
Bart Van Assche 已提交
1653
			/* 64 byte, 512 bit, is the largest digest size
1654 1655
			 * currently supported in kernel crypto. */
			unsigned char digest[64];
1656
			drbd_csum_bio(mdev, mdev->tconn->integrity_tfm, req->master_bio, digest);
1657
			if (memcmp(p + 1, digest, dgs)) {
1658 1659
				dev_warn(DEV,
					"Digest mismatch, buffer modified by upper layers during write: %llus +%u\n",
1660
					(unsigned long long)req->i.sector, req->i.size);
1661 1662 1663 1664
			}
		} /* else if (dgs > 64) {
		     ... Be noisy about digest too large ...
		} */
P
Philipp Reisner 已提交
1665
	}
1666
	mutex_unlock(&sock->mutex);  /* locked by drbd_prepare_command() */
1667

1668
	return err;
P
Philipp Reisner 已提交
1669 1670 1671 1672 1673 1674
}

/* answer packet, used to send data back for read requests:
 *  Peer       -> (diskless) R_PRIMARY   (P_DATA_REPLY)
 *  C_SYNC_SOURCE -> C_SYNC_TARGET         (P_RS_DATA_REPLY)
 */
1675
int drbd_send_block(struct drbd_conf *mdev, enum drbd_packet cmd,
1676
		    struct drbd_peer_request *peer_req)
P
Philipp Reisner 已提交
1677
{
1678 1679
	struct drbd_socket *sock;
	struct p_data *p;
1680
	int err;
P
Philipp Reisner 已提交
1681 1682
	int dgs;

1683 1684 1685
	sock = &mdev->tconn->data;
	p = drbd_prepare_command(mdev, sock);

1686
	dgs = mdev->tconn->integrity_tfm ? crypto_hash_digestsize(mdev->tconn->integrity_tfm) : 0;
P
Philipp Reisner 已提交
1687

1688 1689 1690 1691 1692
	if (!p)
		return -EIO;
	p->sector = cpu_to_be64(peer_req->i.sector);
	p->block_id = peer_req->block_id;
	p->seq_num = 0;  /* unused */
1693
	p->dp_flags = 0;
1694
	if (dgs)
1695
		drbd_csum_ee(mdev, mdev->tconn->integrity_tfm, peer_req, p + 1);
1696
	err = __send_command(mdev->tconn, mdev->vnr, sock, cmd, sizeof(*p) + dgs, NULL, peer_req->i.size);
1697 1698
	if (!err)
		err = _drbd_send_zc_ee(mdev, peer_req);
1699
	mutex_unlock(&sock->mutex);  /* locked by drbd_prepare_command() */
1700

1701
	return err;
P
Philipp Reisner 已提交
1702 1703
}

1704
int drbd_send_out_of_sync(struct drbd_conf *mdev, struct drbd_request *req)
1705
{
1706 1707
	struct drbd_socket *sock;
	struct p_block_desc *p;
1708

1709 1710 1711 1712 1713 1714 1715
	sock = &mdev->tconn->data;
	p = drbd_prepare_command(mdev, sock);
	if (!p)
		return -EIO;
	p->sector = cpu_to_be64(req->i.sector);
	p->blksize = cpu_to_be32(req->i.size);
	return drbd_send_command(mdev, sock, P_OUT_OF_SYNC, sizeof(*p), NULL, 0);
1716 1717
}

P
Philipp Reisner 已提交
1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733
/*
  drbd_send distinguishes two cases:

  Packets sent via the data socket "sock"
  and packets sent via the meta data socket "msock"

		    sock                      msock
  -----------------+-------------------------+------------------------------
  timeout           conf.timeout / 2          conf.timeout / 2
  timeout action    send a ping via msock     Abort communication
					      and close all sockets
*/

/*
 * you must have down()ed the appropriate [m]sock_mutex elsewhere!
 */
1734
int drbd_send(struct drbd_tconn *tconn, struct socket *sock,
P
Philipp Reisner 已提交
1735 1736 1737 1738 1739 1740 1741
	      void *buf, size_t size, unsigned msg_flags)
{
	struct kvec iov;
	struct msghdr msg;
	int rv, sent = 0;

	if (!sock)
1742
		return -EBADR;
P
Philipp Reisner 已提交
1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754

	/* THINK  if (signal_pending) return ... ? */

	iov.iov_base = buf;
	iov.iov_len  = size;

	msg.msg_name       = NULL;
	msg.msg_namelen    = 0;
	msg.msg_control    = NULL;
	msg.msg_controllen = 0;
	msg.msg_flags      = msg_flags | MSG_NOSIGNAL;

1755
	if (sock == tconn->data.socket) {
1756 1757 1758
		rcu_read_lock();
		tconn->ko_count = rcu_dereference(tconn->net_conf)->ko_count;
		rcu_read_unlock();
1759
		drbd_update_congested(tconn);
P
Philipp Reisner 已提交
1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772
	}
	do {
		/* STRANGE
		 * tcp_sendmsg does _not_ use its size parameter at all ?
		 *
		 * -EAGAIN on timeout, -EINTR on signal.
		 */
/* THINK
 * do we need to block DRBD_SIG if sock == &meta.socket ??
 * otherwise wake_asender() might interrupt some send_*Ack !
 */
		rv = kernel_sendmsg(sock, &msg, &iov, 1, size);
		if (rv == -EAGAIN) {
1773
			if (we_should_drop_the_connection(tconn, sock))
P
Philipp Reisner 已提交
1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788
				break;
			else
				continue;
		}
		if (rv == -EINTR) {
			flush_signals(current);
			rv = 0;
		}
		if (rv < 0)
			break;
		sent += rv;
		iov.iov_base += rv;
		iov.iov_len  -= rv;
	} while (sent < size);

1789 1790
	if (sock == tconn->data.socket)
		clear_bit(NET_CONGESTED, &tconn->flags);
P
Philipp Reisner 已提交
1791 1792 1793

	if (rv <= 0) {
		if (rv != -EAGAIN) {
1794 1795 1796
			conn_err(tconn, "%s_sendmsg returned %d\n",
				 sock == tconn->meta.socket ? "msock" : "sock",
				 rv);
1797
			conn_request_state(tconn, NS(conn, C_BROKEN_PIPE), CS_HARD);
P
Philipp Reisner 已提交
1798
		} else
1799
			conn_request_state(tconn, NS(conn, C_TIMEOUT), CS_HARD);
P
Philipp Reisner 已提交
1800 1801 1802 1803 1804
	}

	return sent;
}

1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822
/**
 * drbd_send_all  -  Send an entire buffer
 *
 * Returns 0 upon success and a negative error value otherwise.
 */
int drbd_send_all(struct drbd_tconn *tconn, struct socket *sock, void *buffer,
		  size_t size, unsigned msg_flags)
{
	int err;

	err = drbd_send(tconn, sock, buffer, size, msg_flags);
	if (err < 0)
		return err;
	if (err != size)
		return -EIO;
	return 0;
}

P
Philipp Reisner 已提交
1823 1824 1825 1826 1827 1828
static int drbd_open(struct block_device *bdev, fmode_t mode)
{
	struct drbd_conf *mdev = bdev->bd_disk->private_data;
	unsigned long flags;
	int rv = 0;

1829
	mutex_lock(&drbd_main_mutex);
1830
	spin_lock_irqsave(&mdev->tconn->req_lock, flags);
P
Philipp Reisner 已提交
1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842
	/* to have a stable mdev->state.role
	 * and no race with updating open_cnt */

	if (mdev->state.role != R_PRIMARY) {
		if (mode & FMODE_WRITE)
			rv = -EROFS;
		else if (!allow_oos)
			rv = -EMEDIUMTYPE;
	}

	if (!rv)
		mdev->open_cnt++;
1843
	spin_unlock_irqrestore(&mdev->tconn->req_lock, flags);
1844
	mutex_unlock(&drbd_main_mutex);
P
Philipp Reisner 已提交
1845 1846 1847 1848 1849 1850 1851

	return rv;
}

static int drbd_release(struct gendisk *gd, fmode_t mode)
{
	struct drbd_conf *mdev = gd->private_data;
1852
	mutex_lock(&drbd_main_mutex);
P
Philipp Reisner 已提交
1853
	mdev->open_cnt--;
1854
	mutex_unlock(&drbd_main_mutex);
P
Philipp Reisner 已提交
1855 1856 1857 1858 1859
	return 0;
}

static void drbd_set_defaults(struct drbd_conf *mdev)
{
1860 1861
	/* Beware! The actual layout differs
	 * between big endian and little endian */
1862
	mdev->state = (union drbd_dev_state) {
P
Philipp Reisner 已提交
1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882
		{ .role = R_SECONDARY,
		  .peer = R_UNKNOWN,
		  .conn = C_STANDALONE,
		  .disk = D_DISKLESS,
		  .pdsk = D_UNKNOWN,
		} };
}

void drbd_init_set_defaults(struct drbd_conf *mdev)
{
	/* the memset(,0,) did most of this.
	 * note: only assignments, no allocation in here */

	drbd_set_defaults(mdev);

	atomic_set(&mdev->ap_bio_cnt, 0);
	atomic_set(&mdev->ap_pending_cnt, 0);
	atomic_set(&mdev->rs_pending_cnt, 0);
	atomic_set(&mdev->unacked_cnt, 0);
	atomic_set(&mdev->local_cnt, 0);
1883
	atomic_set(&mdev->pp_in_use_by_net, 0);
1884
	atomic_set(&mdev->rs_sect_in, 0);
1885
	atomic_set(&mdev->rs_sect_ev, 0);
1886
	atomic_set(&mdev->ap_in_flight, 0);
1887
	atomic_set(&mdev->md_io_in_use, 0);
P
Philipp Reisner 已提交
1888

1889 1890
	mutex_init(&mdev->own_state_mutex);
	mdev->state_mutex = &mdev->own_state_mutex;
P
Philipp Reisner 已提交
1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902

	spin_lock_init(&mdev->al_lock);
	spin_lock_init(&mdev->peer_seq_lock);

	INIT_LIST_HEAD(&mdev->active_ee);
	INIT_LIST_HEAD(&mdev->sync_ee);
	INIT_LIST_HEAD(&mdev->done_ee);
	INIT_LIST_HEAD(&mdev->read_ee);
	INIT_LIST_HEAD(&mdev->net_ee);
	INIT_LIST_HEAD(&mdev->resync_reads);
	INIT_LIST_HEAD(&mdev->resync_work.list);
	INIT_LIST_HEAD(&mdev->unplug_work.list);
1903
	INIT_LIST_HEAD(&mdev->go_diskless.list);
P
Philipp Reisner 已提交
1904
	INIT_LIST_HEAD(&mdev->md_sync_work.list);
1905
	INIT_LIST_HEAD(&mdev->start_resync_work.list);
P
Philipp Reisner 已提交
1906
	INIT_LIST_HEAD(&mdev->bm_io_work.w.list);
1907

1908
	mdev->resync_work.cb  = w_resync_timer;
P
Philipp Reisner 已提交
1909
	mdev->unplug_work.cb  = w_send_write_hint;
1910
	mdev->go_diskless.cb  = w_go_diskless;
P
Philipp Reisner 已提交
1911 1912
	mdev->md_sync_work.cb = w_md_sync;
	mdev->bm_io_work.w.cb = w_bitmap_io;
1913
	mdev->start_resync_work.cb = w_start_resync;
1914 1915 1916 1917 1918 1919 1920 1921

	mdev->resync_work.mdev  = mdev;
	mdev->unplug_work.mdev  = mdev;
	mdev->go_diskless.mdev  = mdev;
	mdev->md_sync_work.mdev = mdev;
	mdev->bm_io_work.w.mdev = mdev;
	mdev->start_resync_work.mdev = mdev;

P
Philipp Reisner 已提交
1922 1923
	init_timer(&mdev->resync_timer);
	init_timer(&mdev->md_sync_timer);
1924
	init_timer(&mdev->start_resync_timer);
1925
	init_timer(&mdev->request_timer);
P
Philipp Reisner 已提交
1926 1927 1928 1929
	mdev->resync_timer.function = resync_timer_fn;
	mdev->resync_timer.data = (unsigned long) mdev;
	mdev->md_sync_timer.function = md_sync_timer_fn;
	mdev->md_sync_timer.data = (unsigned long) mdev;
1930 1931
	mdev->start_resync_timer.function = start_resync_timer_fn;
	mdev->start_resync_timer.data = (unsigned long) mdev;
1932 1933
	mdev->request_timer.function = request_timer_fn;
	mdev->request_timer.data = (unsigned long) mdev;
P
Philipp Reisner 已提交
1934 1935 1936 1937 1938 1939 1940 1941

	init_waitqueue_head(&mdev->misc_wait);
	init_waitqueue_head(&mdev->state_wait);
	init_waitqueue_head(&mdev->ee_wait);
	init_waitqueue_head(&mdev->al_wait);
	init_waitqueue_head(&mdev->seq_wait);

	mdev->resync_wenr = LC_FREE;
1942 1943
	mdev->peer_max_bio_size = DRBD_MAX_BIO_SIZE_SAFE;
	mdev->local_max_bio_size = DRBD_MAX_BIO_SIZE_SAFE;
P
Philipp Reisner 已提交
1944 1945 1946 1947
}

void drbd_mdev_cleanup(struct drbd_conf *mdev)
{
1948
	int i;
1949
	if (mdev->tconn->receiver.t_state != NONE)
P
Philipp Reisner 已提交
1950
		dev_err(DEV, "ASSERT FAILED: receiver t_state == %d expected 0.\n",
1951
				mdev->tconn->receiver.t_state);
P
Philipp Reisner 已提交
1952 1953 1954 1955 1956 1957 1958 1959 1960 1961

	mdev->al_writ_cnt  =
	mdev->bm_writ_cnt  =
	mdev->read_cnt     =
	mdev->recv_cnt     =
	mdev->send_cnt     =
	mdev->writ_cnt     =
	mdev->p_size       =
	mdev->rs_start     =
	mdev->rs_total     =
1962 1963
	mdev->rs_failed    = 0;
	mdev->rs_last_events = 0;
1964
	mdev->rs_last_sect_ev = 0;
1965 1966 1967 1968
	for (i = 0; i < DRBD_SYNC_MARKS; i++) {
		mdev->rs_mark_left[i] = 0;
		mdev->rs_mark_time[i] = 0;
	}
1969
	D_ASSERT(mdev->tconn->net_conf == NULL);
P
Philipp Reisner 已提交
1970 1971 1972 1973

	drbd_set_my_capacity(mdev, 0);
	if (mdev->bitmap) {
		/* maybe never allocated. */
1974
		drbd_bm_resize(mdev, 0, 1);
P
Philipp Reisner 已提交
1975 1976 1977
		drbd_bm_cleanup(mdev);
	}

1978 1979 1980
	drbd_free_bc(mdev->ldev);
	mdev->ldev = NULL;

1981
	clear_bit(AL_SUSPENDED, &mdev->flags);
P
Philipp Reisner 已提交
1982 1983 1984 1985 1986 1987 1988

	D_ASSERT(list_empty(&mdev->active_ee));
	D_ASSERT(list_empty(&mdev->sync_ee));
	D_ASSERT(list_empty(&mdev->done_ee));
	D_ASSERT(list_empty(&mdev->read_ee));
	D_ASSERT(list_empty(&mdev->net_ee));
	D_ASSERT(list_empty(&mdev->resync_reads));
1989
	D_ASSERT(list_empty(&mdev->tconn->sender_work.q));
P
Philipp Reisner 已提交
1990 1991
	D_ASSERT(list_empty(&mdev->resync_work.list));
	D_ASSERT(list_empty(&mdev->unplug_work.list));
1992
	D_ASSERT(list_empty(&mdev->go_diskless.list));
1993 1994

	drbd_set_defaults(mdev);
P
Philipp Reisner 已提交
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
}


static void drbd_destroy_mempools(void)
{
	struct page *page;

	while (drbd_pp_pool) {
		page = drbd_pp_pool;
		drbd_pp_pool = (struct page *)page_private(page);
		__free_page(page);
		drbd_pp_vacant--;
	}

	/* D_ASSERT(atomic_read(&drbd_pp_vacant)==0); */

2011 2012
	if (drbd_md_io_bio_set)
		bioset_free(drbd_md_io_bio_set);
2013 2014
	if (drbd_md_io_page_pool)
		mempool_destroy(drbd_md_io_page_pool);
P
Philipp Reisner 已提交
2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027
	if (drbd_ee_mempool)
		mempool_destroy(drbd_ee_mempool);
	if (drbd_request_mempool)
		mempool_destroy(drbd_request_mempool);
	if (drbd_ee_cache)
		kmem_cache_destroy(drbd_ee_cache);
	if (drbd_request_cache)
		kmem_cache_destroy(drbd_request_cache);
	if (drbd_bm_ext_cache)
		kmem_cache_destroy(drbd_bm_ext_cache);
	if (drbd_al_ext_cache)
		kmem_cache_destroy(drbd_al_ext_cache);

2028
	drbd_md_io_bio_set   = NULL;
2029
	drbd_md_io_page_pool = NULL;
P
Philipp Reisner 已提交
2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042
	drbd_ee_mempool      = NULL;
	drbd_request_mempool = NULL;
	drbd_ee_cache        = NULL;
	drbd_request_cache   = NULL;
	drbd_bm_ext_cache    = NULL;
	drbd_al_ext_cache    = NULL;

	return;
}

static int drbd_create_mempools(void)
{
	struct page *page;
2043
	const int number = (DRBD_MAX_BIO_SIZE/PAGE_SIZE) * minor_count;
P
Philipp Reisner 已提交
2044 2045 2046 2047 2048 2049 2050 2051 2052
	int i;

	/* prepare our caches and mempools */
	drbd_request_mempool = NULL;
	drbd_ee_cache        = NULL;
	drbd_request_cache   = NULL;
	drbd_bm_ext_cache    = NULL;
	drbd_al_ext_cache    = NULL;
	drbd_pp_pool         = NULL;
2053
	drbd_md_io_page_pool = NULL;
2054
	drbd_md_io_bio_set   = NULL;
P
Philipp Reisner 已提交
2055 2056 2057 2058 2059 2060 2061 2062

	/* caches */
	drbd_request_cache = kmem_cache_create(
		"drbd_req", sizeof(struct drbd_request), 0, 0, NULL);
	if (drbd_request_cache == NULL)
		goto Enomem;

	drbd_ee_cache = kmem_cache_create(
2063
		"drbd_ee", sizeof(struct drbd_peer_request), 0, 0, NULL);
P
Philipp Reisner 已提交
2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077
	if (drbd_ee_cache == NULL)
		goto Enomem;

	drbd_bm_ext_cache = kmem_cache_create(
		"drbd_bm", sizeof(struct bm_extent), 0, 0, NULL);
	if (drbd_bm_ext_cache == NULL)
		goto Enomem;

	drbd_al_ext_cache = kmem_cache_create(
		"drbd_al", sizeof(struct lc_element), 0, 0, NULL);
	if (drbd_al_ext_cache == NULL)
		goto Enomem;

	/* mempools */
2078 2079 2080 2081
	drbd_md_io_bio_set = bioset_create(DRBD_MIN_POOL_PAGES, 0);
	if (drbd_md_io_bio_set == NULL)
		goto Enomem;

2082 2083 2084 2085
	drbd_md_io_page_pool = mempool_create_page_pool(DRBD_MIN_POOL_PAGES, 0);
	if (drbd_md_io_page_pool == NULL)
		goto Enomem;

P
Philipp Reisner 已提交
2086 2087 2088 2089 2090 2091 2092
	drbd_request_mempool = mempool_create(number,
		mempool_alloc_slab, mempool_free_slab, drbd_request_cache);
	if (drbd_request_mempool == NULL)
		goto Enomem;

	drbd_ee_mempool = mempool_create(number,
		mempool_alloc_slab, mempool_free_slab, drbd_ee_cache);
2093
	if (drbd_ee_mempool == NULL)
P
Philipp Reisner 已提交
2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128
		goto Enomem;

	/* drbd's page pool */
	spin_lock_init(&drbd_pp_lock);

	for (i = 0; i < number; i++) {
		page = alloc_page(GFP_HIGHUSER);
		if (!page)
			goto Enomem;
		set_page_private(page, (unsigned long)drbd_pp_pool);
		drbd_pp_pool = page;
	}
	drbd_pp_vacant = number;

	return 0;

Enomem:
	drbd_destroy_mempools(); /* in case we allocated some */
	return -ENOMEM;
}

static int drbd_notify_sys(struct notifier_block *this, unsigned long code,
	void *unused)
{
	/* just so we have it.  you never know what interesting things we
	 * might want to do here some day...
	 */

	return NOTIFY_DONE;
}

static struct notifier_block drbd_notifier = {
	.notifier_call = drbd_notify_sys,
};

2129
static void drbd_release_all_peer_reqs(struct drbd_conf *mdev)
P
Philipp Reisner 已提交
2130 2131 2132
{
	int rr;

2133
	rr = drbd_free_peer_reqs(mdev, &mdev->active_ee);
P
Philipp Reisner 已提交
2134 2135 2136
	if (rr)
		dev_err(DEV, "%d EEs in active list found!\n", rr);

2137
	rr = drbd_free_peer_reqs(mdev, &mdev->sync_ee);
P
Philipp Reisner 已提交
2138 2139 2140
	if (rr)
		dev_err(DEV, "%d EEs in sync list found!\n", rr);

2141
	rr = drbd_free_peer_reqs(mdev, &mdev->read_ee);
P
Philipp Reisner 已提交
2142 2143 2144
	if (rr)
		dev_err(DEV, "%d EEs in read list found!\n", rr);

2145
	rr = drbd_free_peer_reqs(mdev, &mdev->done_ee);
P
Philipp Reisner 已提交
2146 2147 2148
	if (rr)
		dev_err(DEV, "%d EEs in done list found!\n", rr);

2149
	rr = drbd_free_peer_reqs(mdev, &mdev->net_ee);
P
Philipp Reisner 已提交
2150 2151 2152 2153
	if (rr)
		dev_err(DEV, "%d EEs in net list found!\n", rr);
}

2154
/* caution. no locking. */
2155
void drbd_minor_destroy(struct kref *kref)
P
Philipp Reisner 已提交
2156
{
2157
	struct drbd_conf *mdev = container_of(kref, struct drbd_conf, kref);
2158 2159
	struct drbd_tconn *tconn = mdev->tconn;

2160 2161
	del_timer_sync(&mdev->request_timer);

P
Philipp Reisner 已提交
2162
	/* paranoia asserts */
2163
	D_ASSERT(mdev->open_cnt == 0);
P
Philipp Reisner 已提交
2164 2165 2166 2167 2168 2169 2170 2171
	/* end paranoia asserts */

	/* cleanup stuff that may have been allocated during
	 * device (re-)configuration or state changes */

	if (mdev->this_bdev)
		bdput(mdev->this_bdev);

2172 2173
	drbd_free_bc(mdev->ldev);
	mdev->ldev = NULL;
P
Philipp Reisner 已提交
2174

2175
	drbd_release_all_peer_reqs(mdev);
P
Philipp Reisner 已提交
2176 2177 2178 2179 2180 2181 2182

	lc_destroy(mdev->act_log);
	lc_destroy(mdev->resync);

	kfree(mdev->p_uuid);
	/* mdev->p_uuid = NULL; */

2183 2184 2185 2186 2187
	if (mdev->bitmap) /* should no longer be there. */
		drbd_bm_cleanup(mdev);
	__free_page(mdev->md_io_page);
	put_disk(mdev->vdisk);
	blk_cleanup_queue(mdev->rq_queue);
2188
	kfree(mdev->rs_plan_s);
2189
	kfree(mdev);
2190 2191

	kref_put(&tconn->kref, &conn_destroy);
P
Philipp Reisner 已提交
2192 2193
}

2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238
/* One global retry thread, if we need to push back some bio and have it
 * reinserted through our make request function.
 */
static struct retry_worker {
	struct workqueue_struct *wq;
	struct work_struct worker;

	spinlock_t lock;
	struct list_head writes;
} retry;

static void do_retry(struct work_struct *ws)
{
	struct retry_worker *retry = container_of(ws, struct retry_worker, worker);
	LIST_HEAD(writes);
	struct drbd_request *req, *tmp;

	spin_lock_irq(&retry->lock);
	list_splice_init(&retry->writes, &writes);
	spin_unlock_irq(&retry->lock);

	list_for_each_entry_safe(req, tmp, &writes, tl_requests) {
		struct drbd_conf *mdev = req->w.mdev;
		struct bio *bio = req->master_bio;
		unsigned long start_time = req->start_time;

		/* We have exclusive access to this request object.
		 * If it had not been RQ_POSTPONED, the code path which queued
		 * it here would have completed and freed it already.
		 */
		mempool_free(req, drbd_request_mempool);

		/* A single suspended or otherwise blocking device may stall
		 * all others as well.  Fortunately, this code path is to
		 * recover from a situation that "should not happen":
		 * concurrent writes in multi-primary setup.
		 * In a "normal" lifecycle, this workqueue is supposed to be
		 * destroyed without ever doing anything.
		 * If it turns out to be an issue anyways, we can do per
		 * resource (replication group) or per device (minor) retry
		 * workqueues instead.
		 */

		/* We are not just doing generic_make_request(),
		 * as we want to keep the start_time information. */
2239 2240
		inc_ap_bio(mdev);
		__drbd_make_request(mdev, bio, start_time);
2241 2242 2243
	}
}

2244
void drbd_restart_request(struct drbd_request *req)
2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259
{
	unsigned long flags;
	spin_lock_irqsave(&retry.lock, flags);
	list_move_tail(&req->tl_requests, &retry.writes);
	spin_unlock_irqrestore(&retry.lock, flags);

	/* Drop the extra reference that would otherwise
	 * have been dropped by complete_master_bio.
	 * do_retry() needs to grab a new one. */
	dec_ap_bio(req->w.mdev);

	queue_work(retry.wq, &retry.worker);
}


P
Philipp Reisner 已提交
2260 2261 2262
static void drbd_cleanup(void)
{
	unsigned int i;
2263
	struct drbd_conf *mdev;
2264
	struct drbd_tconn *tconn, *tmp;
P
Philipp Reisner 已提交
2265 2266 2267

	unregister_reboot_notifier(&drbd_notifier);

2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278
	/* first remove proc,
	 * drbdsetup uses it's presence to detect
	 * whether DRBD is loaded.
	 * If we would get stuck in proc removal,
	 * but have netlink already deregistered,
	 * some drbdsetup commands may wait forever
	 * for an answer.
	 */
	if (drbd_proc)
		remove_proc_entry("drbd", NULL);

2279 2280 2281
	if (retry.wq)
		destroy_workqueue(retry.wq);

2282
	drbd_genl_unregister();
P
Philipp Reisner 已提交
2283

2284 2285 2286 2287
	idr_for_each_entry(&minors, mdev, i) {
		idr_remove(&minors, mdev_to_minor(mdev));
		idr_remove(&mdev->tconn->volumes, mdev->vnr);
		del_gendisk(mdev->vdisk);
P
Philipp Reisner 已提交
2288
		/* synchronize_rcu(); No other threads running at this point */
2289 2290 2291
		kref_put(&mdev->kref, &drbd_minor_destroy);
	}

P
Philipp Reisner 已提交
2292
	/* not _rcu since, no other updater anymore. Genl already unregistered */
2293
	list_for_each_entry_safe(tconn, tmp, &drbd_tconns, all_tconn) {
P
Philipp Reisner 已提交
2294 2295
		list_del(&tconn->all_tconn); /* not _rcu no proc, not other threads */
		/* synchronize_rcu(); */
2296 2297
		kref_put(&tconn->kref, &conn_destroy);
	}
2298

2299
	drbd_destroy_mempools();
P
Philipp Reisner 已提交
2300 2301
	unregister_blkdev(DRBD_MAJOR, "drbd");

2302 2303
	idr_destroy(&minors);

P
Philipp Reisner 已提交
2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320
	printk(KERN_INFO "drbd: module cleanup done.\n");
}

/**
 * drbd_congested() - Callback for pdflush
 * @congested_data:	User data
 * @bdi_bits:		Bits pdflush is currently interested in
 *
 * Returns 1<<BDI_async_congested and/or 1<<BDI_sync_congested if we are congested.
 */
static int drbd_congested(void *congested_data, int bdi_bits)
{
	struct drbd_conf *mdev = congested_data;
	struct request_queue *q;
	char reason = '-';
	int r = 0;

2321
	if (!may_inc_ap_bio(mdev)) {
P
Philipp Reisner 已提交
2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335
		/* DRBD has frozen IO */
		r = bdi_bits;
		reason = 'd';
		goto out;
	}

	if (get_ldev(mdev)) {
		q = bdev_get_queue(mdev->ldev->backing_bdev);
		r = bdi_congested(&q->backing_dev_info, bdi_bits);
		put_ldev(mdev);
		if (r)
			reason = 'b';
	}

2336
	if (bdi_bits & (1 << BDI_async_congested) && test_bit(NET_CONGESTED, &mdev->tconn->flags)) {
P
Philipp Reisner 已提交
2337 2338 2339 2340 2341 2342 2343 2344 2345
		r |= (1 << BDI_async_congested);
		reason = reason == 'b' ? 'a' : 'n';
	}

out:
	mdev->congestion_reason = reason;
	return r;
}

2346 2347 2348 2349
static void drbd_init_workqueue(struct drbd_work_queue* wq)
{
	spin_lock_init(&wq->q_lock);
	INIT_LIST_HEAD(&wq->q);
2350
	init_waitqueue_head(&wq->q_wait);
2351 2352
}

2353
struct drbd_tconn *conn_get_by_name(const char *name)
2354 2355 2356
{
	struct drbd_tconn *tconn;

2357 2358 2359
	if (!name || !name[0])
		return NULL;

P
Philipp Reisner 已提交
2360
	rcu_read_lock();
2361
	list_for_each_entry_rcu(tconn, &drbd_tconns, all_tconn) {
2362 2363
		if (!strcmp(tconn->name, name)) {
			kref_get(&tconn->kref);
2364
			goto found;
2365
		}
2366 2367 2368
	}
	tconn = NULL;
found:
P
Philipp Reisner 已提交
2369
	rcu_read_unlock();
2370 2371 2372
	return tconn;
}

2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393
struct drbd_tconn *conn_get_by_addrs(void *my_addr, int my_addr_len,
				     void *peer_addr, int peer_addr_len)
{
	struct drbd_tconn *tconn;

	rcu_read_lock();
	list_for_each_entry_rcu(tconn, &drbd_tconns, all_tconn) {
		if (tconn->my_addr_len == my_addr_len &&
		    tconn->peer_addr_len == peer_addr_len &&
		    !memcmp(&tconn->my_addr, my_addr, my_addr_len) &&
		    !memcmp(&tconn->peer_addr, peer_addr, peer_addr_len)) {
			kref_get(&tconn->kref);
			goto found;
		}
	}
	tconn = NULL;
found:
	rcu_read_unlock();
	return tconn;
}

2394 2395 2396 2397 2398
static int drbd_alloc_socket(struct drbd_socket *socket)
{
	socket->rbuf = (void *) __get_free_page(GFP_KERNEL);
	if (!socket->rbuf)
		return -ENOMEM;
2399 2400 2401
	socket->sbuf = (void *) __get_free_page(GFP_KERNEL);
	if (!socket->sbuf)
		return -ENOMEM;
2402 2403 2404 2405 2406
	return 0;
}

static void drbd_free_socket(struct drbd_socket *socket)
{
2407
	free_page((unsigned long) socket->sbuf);
2408 2409 2410
	free_page((unsigned long) socket->rbuf);
}

2411 2412
void conn_free_crypto(struct drbd_tconn *tconn)
{
2413 2414 2415 2416
	drbd_free_sock(tconn);

	crypto_free_hash(tconn->csums_tfm);
	crypto_free_hash(tconn->verify_tfm);
2417
	crypto_free_hash(tconn->cram_hmac_tfm);
2418
	crypto_free_hash(tconn->integrity_tfm);
2419
	crypto_free_hash(tconn->peer_integrity_tfm);
2420 2421
	kfree(tconn->int_dig_in);
	kfree(tconn->int_dig_vv);
2422 2423 2424

	tconn->csums_tfm = NULL;
	tconn->verify_tfm = NULL;
2425
	tconn->cram_hmac_tfm = NULL;
2426
	tconn->integrity_tfm = NULL;
2427
	tconn->peer_integrity_tfm = NULL;
2428 2429 2430 2431
	tconn->int_dig_in = NULL;
	tconn->int_dig_vv = NULL;
}

2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446
int set_resource_options(struct drbd_tconn *tconn, struct res_opts *res_opts)
{
	cpumask_var_t new_cpu_mask;
	int err;

	if (!zalloc_cpumask_var(&new_cpu_mask, GFP_KERNEL))
		return -ENOMEM;
		/*
		retcode = ERR_NOMEM;
		drbd_msg_put_info("unable to allocate cpumask");
		*/

	/* silently ignore cpu mask on UP kernel */
	if (nr_cpu_ids > 1 && res_opts->cpu_mask[0] != 0) {
		/* FIXME: Get rid of constant 32 here */
2447 2448
		err = bitmap_parse(res_opts->cpu_mask, 32,
				   cpumask_bits(new_cpu_mask), nr_cpu_ids);
2449
		if (err) {
2450
			conn_warn(tconn, "bitmap_parse() failed with %d\n", err);
2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470
			/* retcode = ERR_CPU_MASK_PARSE; */
			goto fail;
		}
	}
	tconn->res_opts = *res_opts;
	if (!cpumask_equal(tconn->cpu_mask, new_cpu_mask)) {
		cpumask_copy(tconn->cpu_mask, new_cpu_mask);
		drbd_calc_cpu_mask(tconn);
		tconn->receiver.reset_cpu_mask = 1;
		tconn->asender.reset_cpu_mask = 1;
		tconn->worker.reset_cpu_mask = 1;
	}
	err = 0;

fail:
	free_cpumask_var(new_cpu_mask);
	return err;

}

2471
/* caller must be under genl_lock() */
2472
struct drbd_tconn *conn_create(const char *name, struct res_opts *res_opts)
2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483
{
	struct drbd_tconn *tconn;

	tconn = kzalloc(sizeof(struct drbd_tconn), GFP_KERNEL);
	if (!tconn)
		return NULL;

	tconn->name = kstrdup(name, GFP_KERNEL);
	if (!tconn->name)
		goto fail;

2484 2485 2486 2487 2488
	if (drbd_alloc_socket(&tconn->data))
		goto fail;
	if (drbd_alloc_socket(&tconn->meta))
		goto fail;

2489 2490 2491
	if (!zalloc_cpumask_var(&tconn->cpu_mask, GFP_KERNEL))
		goto fail;

2492 2493 2494
	if (set_resource_options(tconn, res_opts))
		goto fail;

2495 2496 2497
	tconn->current_epoch = kzalloc(sizeof(struct drbd_epoch), GFP_KERNEL);
	if (!tconn->current_epoch)
		goto fail;
2498 2499 2500

	INIT_LIST_HEAD(&tconn->transfer_log);

2501 2502 2503
	INIT_LIST_HEAD(&tconn->current_epoch->list);
	tconn->epochs = 1;
	spin_lock_init(&tconn->epoch_lock);
2504 2505
	tconn->write_ordering = WO_bdev_flush;

2506 2507 2508 2509
	tconn->send.seen_any_write_yet = false;
	tconn->send.current_epoch_nr = 0;
	tconn->send.current_epoch_writes = 0;

2510
	tconn->cstate = C_STANDALONE;
2511
	mutex_init(&tconn->cstate_mutex);
2512
	spin_lock_init(&tconn->req_lock);
2513
	mutex_init(&tconn->conf_update);
2514
	init_waitqueue_head(&tconn->ping_wait);
2515
	idr_init(&tconn->volumes);
2516

2517
	drbd_init_workqueue(&tconn->sender_work);
2518 2519 2520
	mutex_init(&tconn->data.mutex);
	mutex_init(&tconn->meta.mutex);

2521 2522 2523 2524
	drbd_thread_init(tconn, &tconn->receiver, drbdd_init, "receiver");
	drbd_thread_init(tconn, &tconn->worker, drbd_worker, "worker");
	drbd_thread_init(tconn, &tconn->asender, drbd_asender, "asender");

2525
	kref_init(&tconn->kref);
2526
	list_add_tail_rcu(&tconn->all_tconn, &drbd_tconns);
2527 2528 2529 2530

	return tconn;

fail:
2531
	kfree(tconn->current_epoch);
2532
	free_cpumask_var(tconn->cpu_mask);
2533 2534
	drbd_free_socket(&tconn->meta);
	drbd_free_socket(&tconn->data);
2535 2536 2537 2538 2539 2540
	kfree(tconn->name);
	kfree(tconn);

	return NULL;
}

2541
void conn_destroy(struct kref *kref)
2542
{
2543 2544
	struct drbd_tconn *tconn = container_of(kref, struct drbd_tconn, kref);

2545 2546 2547 2548
	if (atomic_read(&tconn->current_epoch->epoch_size) !=  0)
		conn_err(tconn, "epoch_size:%d\n", atomic_read(&tconn->current_epoch->epoch_size));
	kfree(tconn->current_epoch);

2549
	idr_destroy(&tconn->volumes);
2550

2551
	free_cpumask_var(tconn->cpu_mask);
2552 2553
	drbd_free_socket(&tconn->meta);
	drbd_free_socket(&tconn->data);
2554
	kfree(tconn->name);
2555 2556
	kfree(tconn->int_dig_in);
	kfree(tconn->int_dig_vv);
2557 2558 2559
	kfree(tconn);
}

2560
enum drbd_ret_code conn_new_minor(struct drbd_tconn *tconn, unsigned int minor, int vnr)
P
Philipp Reisner 已提交
2561 2562 2563 2564
{
	struct drbd_conf *mdev;
	struct gendisk *disk;
	struct request_queue *q;
2565
	int vnr_got = vnr;
2566
	int minor_got = minor;
2567
	enum drbd_ret_code err = ERR_NOMEM;
2568 2569 2570 2571

	mdev = minor_to_mdev(minor);
	if (mdev)
		return ERR_MINOR_EXISTS;
P
Philipp Reisner 已提交
2572 2573 2574 2575

	/* GFP_KERNEL, we are outside of all write-out paths */
	mdev = kzalloc(sizeof(struct drbd_conf), GFP_KERNEL);
	if (!mdev)
2576 2577
		return ERR_NOMEM;

2578
	kref_get(&tconn->kref);
2579
	mdev->tconn = tconn;
2580

P
Philipp Reisner 已提交
2581
	mdev->minor = minor;
2582
	mdev->vnr = vnr;
P
Philipp Reisner 已提交
2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596

	drbd_init_set_defaults(mdev);

	q = blk_alloc_queue(GFP_KERNEL);
	if (!q)
		goto out_no_q;
	mdev->rq_queue = q;
	q->queuedata   = mdev;

	disk = alloc_disk(1);
	if (!disk)
		goto out_no_disk;
	mdev->vdisk = disk;

2597
	set_disk_ro(disk, true);
P
Philipp Reisner 已提交
2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612

	disk->queue = q;
	disk->major = DRBD_MAJOR;
	disk->first_minor = minor;
	disk->fops = &drbd_ops;
	sprintf(disk->disk_name, "drbd%d", minor);
	disk->private_data = mdev;

	mdev->this_bdev = bdget(MKDEV(DRBD_MAJOR, minor));
	/* we have no partitions. we contain only ourselves. */
	mdev->this_bdev->bd_contains = mdev->this_bdev;

	q->backing_dev_info.congested_fn = drbd_congested;
	q->backing_dev_info.congested_data = mdev;

2613
	blk_queue_make_request(q, drbd_make_request);
2614 2615 2616
	/* Setting the max_hw_sectors to an odd value of 8kibyte here
	   This triggers a max_bio_size message upon first attach or connect */
	blk_queue_max_hw_sectors(q, DRBD_MAX_BIO_SIZE_SAFE >> 8);
P
Philipp Reisner 已提交
2617 2618
	blk_queue_bounce_limit(q, BLK_BOUNCE_ANY);
	blk_queue_merge_bvec(q, drbd_merge_bvec);
2619
	q->queue_lock = &mdev->tconn->req_lock; /* needed since we use */
P
Philipp Reisner 已提交
2620 2621 2622 2623 2624 2625 2626

	mdev->md_io_page = alloc_page(GFP_KERNEL);
	if (!mdev->md_io_page)
		goto out_no_io_page;

	if (drbd_bm_init(mdev))
		goto out_no_bitmap;
2627
	mdev->read_requests = RB_ROOT;
2628
	mdev->write_requests = RB_ROOT;
P
Philipp Reisner 已提交
2629

2630
	if (!idr_pre_get(&minors, GFP_KERNEL))
2631 2632 2633
		goto out_no_minor_idr;
	if (idr_get_new_above(&minors, mdev, minor, &minor_got))
		goto out_no_minor_idr;
2634
	if (minor_got != minor) {
2635 2636
		err = ERR_MINOR_EXISTS;
		drbd_msg_put_info("requested minor exists already");
2637
		goto out_idr_remove_minor;
2638
	}
2639 2640 2641 2642 2643 2644 2645 2646 2647 2648

	if (!idr_pre_get(&tconn->volumes, GFP_KERNEL))
		goto out_idr_remove_minor;
	if (idr_get_new_above(&tconn->volumes, mdev, vnr, &vnr_got))
		goto out_idr_remove_minor;
	if (vnr_got != vnr) {
		err = ERR_INVALID_REQUEST;
		drbd_msg_put_info("requested volume exists already");
		goto out_idr_remove_vol;
	}
2649
	add_disk(disk);
2650
	kref_init(&mdev->kref); /* one ref for both idrs and the the add_disk */
2651

2652 2653 2654
	/* inherit the connection state */
	mdev->state.conn = tconn->cstate;
	if (mdev->state.conn == C_WF_REPORT_PARAMS)
P
Philipp Reisner 已提交
2655
		drbd_connected(mdev);
2656

2657
	return NO_ERROR;
P
Philipp Reisner 已提交
2658

2659 2660
out_idr_remove_vol:
	idr_remove(&tconn->volumes, vnr_got);
2661 2662
out_idr_remove_minor:
	idr_remove(&minors, minor_got);
2663
	synchronize_rcu();
2664
out_no_minor_idr:
P
Philipp Reisner 已提交
2665 2666 2667 2668 2669 2670 2671 2672 2673
	drbd_bm_cleanup(mdev);
out_no_bitmap:
	__free_page(mdev->md_io_page);
out_no_io_page:
	put_disk(disk);
out_no_disk:
	blk_cleanup_queue(q);
out_no_q:
	kfree(mdev);
2674
	kref_put(&tconn->kref, &conn_destroy);
2675
	return err;
P
Philipp Reisner 已提交
2676 2677 2678 2679 2680 2681
}

int __init drbd_init(void)
{
	int err;

2682
	if (minor_count < DRBD_MINOR_COUNT_MIN || minor_count > DRBD_MINOR_COUNT_MAX) {
P
Philipp Reisner 已提交
2683
		printk(KERN_ERR
2684
		       "drbd: invalid minor_count (%d)\n", minor_count);
P
Philipp Reisner 已提交
2685 2686 2687
#ifdef MODULE
		return -EINVAL;
#else
2688
		minor_count = DRBD_MINOR_COUNT_DEF;
P
Philipp Reisner 已提交
2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699
#endif
	}

	err = register_blkdev(DRBD_MAJOR, "drbd");
	if (err) {
		printk(KERN_ERR
		       "drbd: unable to register block device major %d\n",
		       DRBD_MAJOR);
		return err;
	}

2700 2701 2702 2703 2704 2705 2706
	err = drbd_genl_register();
	if (err) {
		printk(KERN_ERR "drbd: unable to register generic netlink family\n");
		goto fail;
	}


P
Philipp Reisner 已提交
2707 2708 2709 2710 2711 2712 2713 2714 2715 2716
	register_reboot_notifier(&drbd_notifier);

	/*
	 * allocate all necessary structs
	 */
	err = -ENOMEM;

	init_waitqueue_head(&drbd_pp_wait);

	drbd_proc = NULL; /* play safe for drbd_cleanup */
2717
	idr_init(&minors);
P
Philipp Reisner 已提交
2718 2719 2720

	err = drbd_create_mempools();
	if (err)
2721
		goto fail;
P
Philipp Reisner 已提交
2722

2723
	drbd_proc = proc_create_data("drbd", S_IFREG | S_IRUGO , NULL, &drbd_proc_fops, NULL);
P
Philipp Reisner 已提交
2724 2725
	if (!drbd_proc)	{
		printk(KERN_ERR "drbd: unable to register proc file\n");
2726
		goto fail;
P
Philipp Reisner 已提交
2727 2728 2729
	}

	rwlock_init(&global_state_lock);
2730
	INIT_LIST_HEAD(&drbd_tconns);
P
Philipp Reisner 已提交
2731

2732 2733 2734 2735 2736 2737 2738 2739 2740
	retry.wq = create_singlethread_workqueue("drbd-reissue");
	if (!retry.wq) {
		printk(KERN_ERR "drbd: unable to create retry workqueue\n");
		goto fail;
	}
	INIT_WORK(&retry.worker, do_retry);
	spin_lock_init(&retry.lock);
	INIT_LIST_HEAD(&retry.writes);

P
Philipp Reisner 已提交
2741 2742 2743 2744 2745 2746 2747 2748 2749
	printk(KERN_INFO "drbd: initialized. "
	       "Version: " REL_VERSION " (api:%d/proto:%d-%d)\n",
	       API_VERSION, PRO_VERSION_MIN, PRO_VERSION_MAX);
	printk(KERN_INFO "drbd: %s\n", drbd_buildtag());
	printk(KERN_INFO "drbd: registered as block device major %d\n",
		DRBD_MAJOR);

	return 0; /* Success! */

2750
fail:
P
Philipp Reisner 已提交
2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764
	drbd_cleanup();
	if (err == -ENOMEM)
		/* currently always the case */
		printk(KERN_ERR "drbd: ran out of memory\n");
	else
		printk(KERN_ERR "drbd: initialization failure\n");
	return err;
}

void drbd_free_bc(struct drbd_backing_dev *ldev)
{
	if (ldev == NULL)
		return;

2765 2766
	blkdev_put(ldev->backing_bdev, FMODE_READ | FMODE_WRITE | FMODE_EXCL);
	blkdev_put(ldev->md_bdev, FMODE_READ | FMODE_WRITE | FMODE_EXCL);
P
Philipp Reisner 已提交
2767 2768 2769 2770

	kfree(ldev);
}

2771 2772 2773 2774 2775 2776 2777 2778
void drbd_free_sock(struct drbd_tconn *tconn)
{
	if (tconn->data.socket) {
		mutex_lock(&tconn->data.mutex);
		kernel_sock_shutdown(tconn->data.socket, SHUT_RDWR);
		sock_release(tconn->data.socket);
		tconn->data.socket = NULL;
		mutex_unlock(&tconn->data.mutex);
P
Philipp Reisner 已提交
2779
	}
2780 2781 2782 2783 2784 2785
	if (tconn->meta.socket) {
		mutex_lock(&tconn->meta.mutex);
		kernel_sock_shutdown(tconn->meta.socket, SHUT_RDWR);
		sock_release(tconn->meta.socket);
		tconn->meta.socket = NULL;
		mutex_unlock(&tconn->meta.mutex);
P
Philipp Reisner 已提交
2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800
	}
}

/* meta data management */

struct meta_data_on_disk {
	u64 la_size;           /* last agreed size. */
	u64 uuid[UI_SIZE];   /* UUIDs. */
	u64 device_uuid;
	u64 reserved_u64_1;
	u32 flags;             /* MDF */
	u32 magic;
	u32 md_size_sect;
	u32 al_offset;         /* offset to this block */
	u32 al_nr_extents;     /* important for restoring the AL */
2801
	      /* `-- act_log->nr_elements <-- ldev->dc.al_extents */
P
Philipp Reisner 已提交
2802 2803
	u32 bm_offset;         /* offset to the bitmap, from here */
	u32 bm_bytes_per_bit;  /* BM_BLOCK_SIZE */
2804 2805
	u32 la_peer_max_bio_size;   /* last peer max_bio_size */
	u32 reserved_u32[3];
P
Philipp Reisner 已提交
2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818

} __packed;

/**
 * drbd_md_sync() - Writes the meta data super block if the MD_DIRTY flag bit is set
 * @mdev:	DRBD device.
 */
void drbd_md_sync(struct drbd_conf *mdev)
{
	struct meta_data_on_disk *buffer;
	sector_t sector;
	int i;

2819 2820
	del_timer(&mdev->md_sync_timer);
	/* timer may be rearmed by drbd_md_mark_dirty() now. */
P
Philipp Reisner 已提交
2821 2822 2823 2824 2825 2826 2827 2828
	if (!test_and_clear_bit(MD_DIRTY, &mdev->flags))
		return;

	/* We use here D_FAILED and not D_ATTACHING because we try to write
	 * metadata even if we detach due to a disk failure! */
	if (!get_ldev_if_state(mdev, D_FAILED))
		return;

2829 2830 2831 2832
	buffer = drbd_md_get_buffer(mdev);
	if (!buffer)
		goto out;

P
Philipp Reisner 已提交
2833 2834 2835 2836 2837 2838
	memset(buffer, 0, 512);

	buffer->la_size = cpu_to_be64(drbd_get_capacity(mdev->this_bdev));
	for (i = UI_CURRENT; i < UI_SIZE; i++)
		buffer->uuid[i] = cpu_to_be64(mdev->ldev->md.uuid[i]);
	buffer->flags = cpu_to_be32(mdev->ldev->md.flags);
2839
	buffer->magic = cpu_to_be32(DRBD_MD_MAGIC_84_UNCLEAN);
P
Philipp Reisner 已提交
2840 2841 2842 2843 2844 2845 2846 2847

	buffer->md_size_sect  = cpu_to_be32(mdev->ldev->md.md_size_sect);
	buffer->al_offset     = cpu_to_be32(mdev->ldev->md.al_offset);
	buffer->al_nr_extents = cpu_to_be32(mdev->act_log->nr_elements);
	buffer->bm_bytes_per_bit = cpu_to_be32(BM_BLOCK_SIZE);
	buffer->device_uuid = cpu_to_be64(mdev->ldev->md.device_uuid);

	buffer->bm_offset = cpu_to_be32(mdev->ldev->md.bm_offset);
2848
	buffer->la_peer_max_bio_size = cpu_to_be32(mdev->peer_max_bio_size);
P
Philipp Reisner 已提交
2849 2850 2851 2852

	D_ASSERT(drbd_md_ss__(mdev, mdev->ldev) == mdev->ldev->md.md_offset);
	sector = mdev->ldev->md.md_offset;

2853
	if (drbd_md_sync_page_io(mdev, mdev->ldev, sector, WRITE)) {
P
Philipp Reisner 已提交
2854 2855
		/* this was a try anyways ... */
		dev_err(DEV, "meta data update failed!\n");
2856
		drbd_chk_io_error(mdev, 1, true);
P
Philipp Reisner 已提交
2857 2858 2859 2860 2861 2862
	}

	/* Update mdev->ldev->md.la_size_sect,
	 * since we updated it on metadata. */
	mdev->ldev->md.la_size_sect = drbd_get_capacity(mdev->this_bdev);

2863 2864
	drbd_md_put_buffer(mdev);
out:
P
Philipp Reisner 已提交
2865 2866 2867 2868 2869 2870 2871 2872
	put_ldev(mdev);
}

/**
 * drbd_md_read() - Reads in the meta data super block
 * @mdev:	DRBD device.
 * @bdev:	Device from which the meta data should be read in.
 *
2873
 * Return 0 (NO_ERROR) on success, and an enum drbd_ret_code in case
2874
 * something goes wrong.
P
Philipp Reisner 已提交
2875 2876 2877 2878
 */
int drbd_md_read(struct drbd_conf *mdev, struct drbd_backing_dev *bdev)
{
	struct meta_data_on_disk *buffer;
2879
	u32 magic, flags;
P
Philipp Reisner 已提交
2880 2881 2882 2883 2884
	int i, rv = NO_ERROR;

	if (!get_ldev_if_state(mdev, D_ATTACHING))
		return ERR_IO_MD_DISK;

2885 2886 2887
	buffer = drbd_md_get_buffer(mdev);
	if (!buffer)
		goto out;
P
Philipp Reisner 已提交
2888

2889
	if (drbd_md_sync_page_io(mdev, bdev, bdev->md.md_offset, READ)) {
L
Lucas De Marchi 已提交
2890
		/* NOTE: can't do normal error processing here as this is
P
Philipp Reisner 已提交
2891 2892 2893 2894 2895 2896
		   called BEFORE disk is attached */
		dev_err(DEV, "Error while reading metadata.\n");
		rv = ERR_IO_MD_DISK;
		goto err;
	}

2897 2898 2899 2900 2901 2902 2903 2904 2905 2906
	magic = be32_to_cpu(buffer->magic);
	flags = be32_to_cpu(buffer->flags);
	if (magic == DRBD_MD_MAGIC_84_UNCLEAN ||
	    (magic == DRBD_MD_MAGIC_08 && !(flags & MDF_AL_CLEAN))) {
			/* btw: that's Activity Log clean, not "all" clean. */
		dev_err(DEV, "Found unclean meta data. Did you \"drbdadm apply-al\"?\n");
		rv = ERR_MD_UNCLEAN;
		goto err;
	}
	if (magic != DRBD_MD_MAGIC_08) {
2907
		if (magic == DRBD_MD_MAGIC_07)
2908 2909 2910
			dev_err(DEV, "Found old (0.7) meta data magic. Did you \"drbdadm create-md\"?\n");
		else
			dev_err(DEV, "Meta data magic not found. Did you \"drbdadm create-md\"?\n");
P
Philipp Reisner 已提交
2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945
		rv = ERR_MD_INVALID;
		goto err;
	}
	if (be32_to_cpu(buffer->al_offset) != bdev->md.al_offset) {
		dev_err(DEV, "unexpected al_offset: %d (expected %d)\n",
		    be32_to_cpu(buffer->al_offset), bdev->md.al_offset);
		rv = ERR_MD_INVALID;
		goto err;
	}
	if (be32_to_cpu(buffer->bm_offset) != bdev->md.bm_offset) {
		dev_err(DEV, "unexpected bm_offset: %d (expected %d)\n",
		    be32_to_cpu(buffer->bm_offset), bdev->md.bm_offset);
		rv = ERR_MD_INVALID;
		goto err;
	}
	if (be32_to_cpu(buffer->md_size_sect) != bdev->md.md_size_sect) {
		dev_err(DEV, "unexpected md_size: %u (expected %u)\n",
		    be32_to_cpu(buffer->md_size_sect), bdev->md.md_size_sect);
		rv = ERR_MD_INVALID;
		goto err;
	}

	if (be32_to_cpu(buffer->bm_bytes_per_bit) != BM_BLOCK_SIZE) {
		dev_err(DEV, "unexpected bm_bytes_per_bit: %u (expected %u)\n",
		    be32_to_cpu(buffer->bm_bytes_per_bit), BM_BLOCK_SIZE);
		rv = ERR_MD_INVALID;
		goto err;
	}

	bdev->md.la_size_sect = be64_to_cpu(buffer->la_size);
	for (i = UI_CURRENT; i < UI_SIZE; i++)
		bdev->md.uuid[i] = be64_to_cpu(buffer->uuid[i]);
	bdev->md.flags = be32_to_cpu(buffer->flags);
	bdev->md.device_uuid = be64_to_cpu(buffer->device_uuid);

2946
	spin_lock_irq(&mdev->tconn->req_lock);
2947 2948 2949 2950 2951 2952
	if (mdev->state.conn < C_CONNECTED) {
		int peer;
		peer = be32_to_cpu(buffer->la_peer_max_bio_size);
		peer = max_t(int, peer, DRBD_MAX_BIO_SIZE_SAFE);
		mdev->peer_max_bio_size = peer;
	}
2953
	spin_unlock_irq(&mdev->tconn->req_lock);
2954

P
Philipp Reisner 已提交
2955
 err:
2956 2957
	drbd_md_put_buffer(mdev);
 out:
P
Philipp Reisner 已提交
2958 2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970
	put_ldev(mdev);

	return rv;
}

/**
 * drbd_md_mark_dirty() - Mark meta data super block as dirty
 * @mdev:	DRBD device.
 *
 * Call this function if you change anything that should be written to
 * the meta-data super block. This function sets MD_DIRTY, and starts a
 * timer that ensures that within five seconds you have to call drbd_md_sync().
 */
2971
#ifdef DEBUG
2972 2973 2974 2975 2976 2977 2978 2979 2980
void drbd_md_mark_dirty_(struct drbd_conf *mdev, unsigned int line, const char *func)
{
	if (!test_and_set_bit(MD_DIRTY, &mdev->flags)) {
		mod_timer(&mdev->md_sync_timer, jiffies + HZ);
		mdev->last_md_mark_dirty.line = line;
		mdev->last_md_mark_dirty.func = func;
	}
}
#else
P
Philipp Reisner 已提交
2981 2982
void drbd_md_mark_dirty(struct drbd_conf *mdev)
{
2983
	if (!test_and_set_bit(MD_DIRTY, &mdev->flags))
2984
		mod_timer(&mdev->md_sync_timer, jiffies + 5*HZ);
P
Philipp Reisner 已提交
2985
}
2986
#endif
P
Philipp Reisner 已提交
2987 2988 2989 2990 2991

static void drbd_uuid_move_history(struct drbd_conf *mdev) __must_hold(local)
{
	int i;

2992
	for (i = UI_HISTORY_START; i < UI_HISTORY_END; i++)
P
Philipp Reisner 已提交
2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030
		mdev->ldev->md.uuid[i+1] = mdev->ldev->md.uuid[i];
}

void _drbd_uuid_set(struct drbd_conf *mdev, int idx, u64 val) __must_hold(local)
{
	if (idx == UI_CURRENT) {
		if (mdev->state.role == R_PRIMARY)
			val |= 1;
		else
			val &= ~((u64)1);

		drbd_set_ed_uuid(mdev, val);
	}

	mdev->ldev->md.uuid[idx] = val;
	drbd_md_mark_dirty(mdev);
}


void drbd_uuid_set(struct drbd_conf *mdev, int idx, u64 val) __must_hold(local)
{
	if (mdev->ldev->md.uuid[idx]) {
		drbd_uuid_move_history(mdev);
		mdev->ldev->md.uuid[UI_HISTORY_START] = mdev->ldev->md.uuid[idx];
	}
	_drbd_uuid_set(mdev, idx, val);
}

/**
 * drbd_uuid_new_current() - Creates a new current UUID
 * @mdev:	DRBD device.
 *
 * Creates a new current UUID, and rotates the old current UUID into
 * the bitmap slot. Causes an incremental resync upon next connect.
 */
void drbd_uuid_new_current(struct drbd_conf *mdev) __must_hold(local)
{
	u64 val;
3031 3032 3033 3034
	unsigned long long bm_uuid = mdev->ldev->md.uuid[UI_BITMAP];

	if (bm_uuid)
		dev_warn(DEV, "bm UUID was already set: %llX\n", bm_uuid);
P
Philipp Reisner 已提交
3035 3036 3037 3038 3039

	mdev->ldev->md.uuid[UI_BITMAP] = mdev->ldev->md.uuid[UI_CURRENT];

	get_random_bytes(&val, sizeof(u64));
	_drbd_uuid_set(mdev, UI_CURRENT, val);
3040
	drbd_print_uuids(mdev, "new current UUID");
3041 3042
	/* get it to stable storage _now_ */
	drbd_md_sync(mdev);
P
Philipp Reisner 已提交
3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054
}

void drbd_uuid_set_bm(struct drbd_conf *mdev, u64 val) __must_hold(local)
{
	if (mdev->ldev->md.uuid[UI_BITMAP] == 0 && val == 0)
		return;

	if (val == 0) {
		drbd_uuid_move_history(mdev);
		mdev->ldev->md.uuid[UI_HISTORY_START] = mdev->ldev->md.uuid[UI_BITMAP];
		mdev->ldev->md.uuid[UI_BITMAP] = 0;
	} else {
3055 3056 3057
		unsigned long long bm_uuid = mdev->ldev->md.uuid[UI_BITMAP];
		if (bm_uuid)
			dev_warn(DEV, "bm UUID was already set: %llX\n", bm_uuid);
P
Philipp Reisner 已提交
3058

3059
		mdev->ldev->md.uuid[UI_BITMAP] = val & ~((u64)1);
P
Philipp Reisner 已提交
3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101
	}
	drbd_md_mark_dirty(mdev);
}

/**
 * drbd_bmio_set_n_write() - io_fn for drbd_queue_bitmap_io() or drbd_bitmap_io()
 * @mdev:	DRBD device.
 *
 * Sets all bits in the bitmap and writes the whole bitmap to stable storage.
 */
int drbd_bmio_set_n_write(struct drbd_conf *mdev)
{
	int rv = -EIO;

	if (get_ldev_if_state(mdev, D_ATTACHING)) {
		drbd_md_set_flag(mdev, MDF_FULL_SYNC);
		drbd_md_sync(mdev);
		drbd_bm_set_all(mdev);

		rv = drbd_bm_write(mdev);

		if (!rv) {
			drbd_md_clear_flag(mdev, MDF_FULL_SYNC);
			drbd_md_sync(mdev);
		}

		put_ldev(mdev);
	}

	return rv;
}

/**
 * drbd_bmio_clear_n_write() - io_fn for drbd_queue_bitmap_io() or drbd_bitmap_io()
 * @mdev:	DRBD device.
 *
 * Clears all bits in the bitmap and writes the whole bitmap to stable storage.
 */
int drbd_bmio_clear_n_write(struct drbd_conf *mdev)
{
	int rv = -EIO;

3102
	drbd_resume_al(mdev);
P
Philipp Reisner 已提交
3103 3104 3105 3106 3107 3108 3109 3110 3111
	if (get_ldev_if_state(mdev, D_ATTACHING)) {
		drbd_bm_clear_all(mdev);
		rv = drbd_bm_write(mdev);
		put_ldev(mdev);
	}

	return rv;
}

3112
static int w_bitmap_io(struct drbd_work *w, int unused)
P
Philipp Reisner 已提交
3113 3114
{
	struct bm_io_work *work = container_of(w, struct bm_io_work, w);
3115
	struct drbd_conf *mdev = w->mdev;
3116
	int rv = -EIO;
P
Philipp Reisner 已提交
3117 3118 3119

	D_ASSERT(atomic_read(&mdev->ap_bio_cnt) == 0);

3120
	if (get_ldev(mdev)) {
3121
		drbd_bm_lock(mdev, work->why, work->flags);
3122 3123 3124 3125
		rv = work->io_fn(mdev);
		drbd_bm_unlock(mdev);
		put_ldev(mdev);
	}
P
Philipp Reisner 已提交
3126

3127
	clear_bit_unlock(BITMAP_IO, &mdev->flags);
P
Philipp Reisner 已提交
3128 3129 3130 3131 3132 3133 3134
	wake_up(&mdev->misc_wait);

	if (work->done)
		work->done(mdev, rv);

	clear_bit(BITMAP_IO_QUEUED, &mdev->flags);
	work->why = NULL;
3135
	work->flags = 0;
P
Philipp Reisner 已提交
3136

3137
	return 0;
P
Philipp Reisner 已提交
3138 3139
}

3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152
void drbd_ldev_destroy(struct drbd_conf *mdev)
{
	lc_destroy(mdev->resync);
	mdev->resync = NULL;
	lc_destroy(mdev->act_log);
	mdev->act_log = NULL;
	__no_warn(local,
		drbd_free_bc(mdev->ldev);
		mdev->ldev = NULL;);

	clear_bit(GO_DISKLESS, &mdev->flags);
}

3153
static int w_go_diskless(struct drbd_work *w, int unused)
3154
{
3155 3156
	struct drbd_conf *mdev = w->mdev;

3157
	D_ASSERT(mdev->state.disk == D_FAILED);
3158 3159
	/* we cannot assert local_cnt == 0 here, as get_ldev_if_state will
	 * inc/dec it frequently. Once we are D_DISKLESS, no one will touch
3160 3161
	 * the protected members anymore, though, so once put_ldev reaches zero
	 * again, it will be safe to free them. */
3162
	drbd_force_state(mdev, NS(disk, D_DISKLESS));
3163
	return 0;
3164 3165 3166 3167 3168 3169
}

void drbd_go_diskless(struct drbd_conf *mdev)
{
	D_ASSERT(mdev->state.disk == D_FAILED);
	if (!test_and_set_bit(GO_DISKLESS, &mdev->flags))
3170
		drbd_queue_work(&mdev->tconn->sender_work, &mdev->go_diskless);
3171 3172
}

P
Philipp Reisner 已提交
3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187
/**
 * drbd_queue_bitmap_io() - Queues an IO operation on the whole bitmap
 * @mdev:	DRBD device.
 * @io_fn:	IO callback to be called when bitmap IO is possible
 * @done:	callback to be called after the bitmap IO was performed
 * @why:	Descriptive text of the reason for doing the IO
 *
 * While IO on the bitmap happens we freeze application IO thus we ensure
 * that drbd_set_out_of_sync() can not be called. This function MAY ONLY be
 * called from worker context. It MUST NOT be used while a previous such
 * work is still pending!
 */
void drbd_queue_bitmap_io(struct drbd_conf *mdev,
			  int (*io_fn)(struct drbd_conf *),
			  void (*done)(struct drbd_conf *, int),
3188
			  char *why, enum bm_flag flags)
P
Philipp Reisner 已提交
3189
{
3190
	D_ASSERT(current == mdev->tconn->worker.task);
P
Philipp Reisner 已提交
3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201

	D_ASSERT(!test_bit(BITMAP_IO_QUEUED, &mdev->flags));
	D_ASSERT(!test_bit(BITMAP_IO, &mdev->flags));
	D_ASSERT(list_empty(&mdev->bm_io_work.w.list));
	if (mdev->bm_io_work.why)
		dev_err(DEV, "FIXME going to queue '%s' but '%s' still pending?\n",
			why, mdev->bm_io_work.why);

	mdev->bm_io_work.io_fn = io_fn;
	mdev->bm_io_work.done = done;
	mdev->bm_io_work.why = why;
3202
	mdev->bm_io_work.flags = flags;
P
Philipp Reisner 已提交
3203

3204
	spin_lock_irq(&mdev->tconn->req_lock);
P
Philipp Reisner 已提交
3205 3206
	set_bit(BITMAP_IO, &mdev->flags);
	if (atomic_read(&mdev->ap_bio_cnt) == 0) {
3207
		if (!test_and_set_bit(BITMAP_IO_QUEUED, &mdev->flags))
3208
			drbd_queue_work(&mdev->tconn->sender_work, &mdev->bm_io_work.w);
P
Philipp Reisner 已提交
3209
	}
3210
	spin_unlock_irq(&mdev->tconn->req_lock);
P
Philipp Reisner 已提交
3211 3212 3213 3214 3215 3216 3217 3218 3219 3220 3221
}

/**
 * drbd_bitmap_io() -  Does an IO operation on the whole bitmap
 * @mdev:	DRBD device.
 * @io_fn:	IO callback to be called when bitmap IO is possible
 * @why:	Descriptive text of the reason for doing the IO
 *
 * freezes application IO while that the actual IO operations runs. This
 * functions MAY NOT be called from worker context.
 */
3222 3223
int drbd_bitmap_io(struct drbd_conf *mdev, int (*io_fn)(struct drbd_conf *),
		char *why, enum bm_flag flags)
P
Philipp Reisner 已提交
3224 3225 3226
{
	int rv;

3227
	D_ASSERT(current != mdev->tconn->worker.task);
P
Philipp Reisner 已提交
3228

3229 3230
	if ((flags & BM_LOCKED_SET_ALLOWED) == 0)
		drbd_suspend_io(mdev);
P
Philipp Reisner 已提交
3231

3232
	drbd_bm_lock(mdev, why, flags);
P
Philipp Reisner 已提交
3233 3234 3235
	rv = io_fn(mdev);
	drbd_bm_unlock(mdev);

3236 3237
	if ((flags & BM_LOCKED_SET_ALLOWED) == 0)
		drbd_resume_io(mdev);
P
Philipp Reisner 已提交
3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265

	return rv;
}

void drbd_md_set_flag(struct drbd_conf *mdev, int flag) __must_hold(local)
{
	if ((mdev->ldev->md.flags & flag) != flag) {
		drbd_md_mark_dirty(mdev);
		mdev->ldev->md.flags |= flag;
	}
}

void drbd_md_clear_flag(struct drbd_conf *mdev, int flag) __must_hold(local)
{
	if ((mdev->ldev->md.flags & flag) != 0) {
		drbd_md_mark_dirty(mdev);
		mdev->ldev->md.flags &= ~flag;
	}
}
int drbd_md_test_flag(struct drbd_backing_dev *bdev, int flag)
{
	return (bdev->md.flags & flag) != 0;
}

static void md_sync_timer_fn(unsigned long data)
{
	struct drbd_conf *mdev = (struct drbd_conf *) data;

3266
	drbd_queue_work_front(&mdev->tconn->sender_work, &mdev->md_sync_work);
P
Philipp Reisner 已提交
3267 3268
}

3269
static int w_md_sync(struct drbd_work *w, int unused)
P
Philipp Reisner 已提交
3270
{
3271 3272
	struct drbd_conf *mdev = w->mdev;

P
Philipp Reisner 已提交
3273
	dev_warn(DEV, "md_sync_timer expired! Worker calls drbd_md_sync().\n");
3274 3275 3276 3277
#ifdef DEBUG
	dev_warn(DEV, "last md_mark_dirty: %s:%u\n",
		mdev->last_md_mark_dirty.func, mdev->last_md_mark_dirty.line);
#endif
P
Philipp Reisner 已提交
3278
	drbd_md_sync(mdev);
3279
	return 0;
P
Philipp Reisner 已提交
3280 3281
}

3282
const char *cmdname(enum drbd_packet cmd)
3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308 3309 3310 3311
{
	/* THINK may need to become several global tables
	 * when we want to support more than
	 * one PRO_VERSION */
	static const char *cmdnames[] = {
		[P_DATA]	        = "Data",
		[P_DATA_REPLY]	        = "DataReply",
		[P_RS_DATA_REPLY]	= "RSDataReply",
		[P_BARRIER]	        = "Barrier",
		[P_BITMAP]	        = "ReportBitMap",
		[P_BECOME_SYNC_TARGET]  = "BecomeSyncTarget",
		[P_BECOME_SYNC_SOURCE]  = "BecomeSyncSource",
		[P_UNPLUG_REMOTE]	= "UnplugRemote",
		[P_DATA_REQUEST]	= "DataRequest",
		[P_RS_DATA_REQUEST]     = "RSDataRequest",
		[P_SYNC_PARAM]	        = "SyncParam",
		[P_SYNC_PARAM89]	= "SyncParam89",
		[P_PROTOCOL]            = "ReportProtocol",
		[P_UUIDS]	        = "ReportUUIDs",
		[P_SIZES]	        = "ReportSizes",
		[P_STATE]	        = "ReportState",
		[P_SYNC_UUID]           = "ReportSyncUUID",
		[P_AUTH_CHALLENGE]      = "AuthChallenge",
		[P_AUTH_RESPONSE]	= "AuthResponse",
		[P_PING]		= "Ping",
		[P_PING_ACK]	        = "PingAck",
		[P_RECV_ACK]	        = "RecvAck",
		[P_WRITE_ACK]	        = "WriteAck",
		[P_RS_WRITE_ACK]	= "RSWriteAck",
3312
		[P_DISCARD_WRITE]        = "DiscardWrite",
3313 3314 3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326
		[P_NEG_ACK]	        = "NegAck",
		[P_NEG_DREPLY]	        = "NegDReply",
		[P_NEG_RS_DREPLY]	= "NegRSDReply",
		[P_BARRIER_ACK]	        = "BarrierAck",
		[P_STATE_CHG_REQ]       = "StateChgRequest",
		[P_STATE_CHG_REPLY]     = "StateChgReply",
		[P_OV_REQUEST]          = "OVRequest",
		[P_OV_REPLY]            = "OVReply",
		[P_OV_RESULT]           = "OVResult",
		[P_CSUM_RS_REQUEST]     = "CsumRSRequest",
		[P_RS_IS_IN_SYNC]	= "CsumRSIsInSync",
		[P_COMPRESSED_BITMAP]   = "CBitmap",
		[P_DELAY_PROBE]         = "DelayProbe",
		[P_OUT_OF_SYNC]		= "OutOfSync",
3327
		[P_RETRY_WRITE]		= "RetryWrite",
3328 3329 3330
		[P_RS_CANCEL]		= "RSCancel",
		[P_CONN_ST_CHG_REQ]	= "conn_st_chg_req",
		[P_CONN_ST_CHG_REPLY]	= "conn_st_chg_reply",
3331 3332
		[P_RETRY_WRITE]		= "retry_write",
		[P_PROTOCOL_UPDATE]	= "protocol_update",
3333 3334 3335 3336 3337

		/* enum drbd_packet, but not commands - obsoleted flags:
		 *	P_MAY_IGNORE
		 *	P_MAX_OPT_CMD
		 */
3338 3339
	};

3340
	/* too big for the array: 0xfffX */
3341 3342 3343 3344
	if (cmd == P_INITIAL_META)
		return "InitialMeta";
	if (cmd == P_INITIAL_DATA)
		return "InitialData";
3345 3346
	if (cmd == P_CONNECTION_FEATURES)
		return "ConnectionFeatures";
A
Andreas Gruenbacher 已提交
3347
	if (cmd >= ARRAY_SIZE(cmdnames))
3348 3349 3350 3351
		return "Unknown";
	return cmdnames[cmd];
}

3352 3353 3354 3355 3356 3357 3358 3359
/**
 * drbd_wait_misc  -  wait for a request to make progress
 * @mdev:	device associated with the request
 * @i:		the struct drbd_interval embedded in struct drbd_request or
 *		struct drbd_peer_request
 */
int drbd_wait_misc(struct drbd_conf *mdev, struct drbd_interval *i)
{
3360
	struct net_conf *nc;
3361 3362 3363
	DEFINE_WAIT(wait);
	long timeout;

3364 3365 3366 3367
	rcu_read_lock();
	nc = rcu_dereference(mdev->tconn->net_conf);
	if (!nc) {
		rcu_read_unlock();
3368
		return -ETIMEDOUT;
3369 3370 3371
	}
	timeout = nc->ko_count ? nc->timeout * HZ / 10 * nc->ko_count : MAX_SCHEDULE_TIMEOUT;
	rcu_read_unlock();
3372 3373 3374 3375 3376 3377 3378 3379 3380 3381 3382 3383 3384 3385 3386

	/* Indicate to wake up mdev->misc_wait on progress.  */
	i->waiting = true;
	prepare_to_wait(&mdev->misc_wait, &wait, TASK_INTERRUPTIBLE);
	spin_unlock_irq(&mdev->tconn->req_lock);
	timeout = schedule_timeout(timeout);
	finish_wait(&mdev->misc_wait, &wait);
	spin_lock_irq(&mdev->tconn->req_lock);
	if (!timeout || mdev->state.conn < C_CONNECTED)
		return -ETIMEDOUT;
	if (signal_pending(current))
		return -ERESTARTSYS;
	return 0;
}

P
Philipp Reisner 已提交
3387 3388 3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407
#ifdef CONFIG_DRBD_FAULT_INJECTION
/* Fault insertion support including random number generator shamelessly
 * stolen from kernel/rcutorture.c */
struct fault_random_state {
	unsigned long state;
	unsigned long count;
};

#define FAULT_RANDOM_MULT 39916801  /* prime */
#define FAULT_RANDOM_ADD	479001701 /* prime */
#define FAULT_RANDOM_REFRESH 10000

/*
 * Crude but fast random-number generator.  Uses a linear congruential
 * generator, with occasional help from get_random_bytes().
 */
static unsigned long
_drbd_fault_random(struct fault_random_state *rsp)
{
	long refresh;

3408
	if (!rsp->count--) {
P
Philipp Reisner 已提交
3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427
		get_random_bytes(&refresh, sizeof(refresh));
		rsp->state += refresh;
		rsp->count = FAULT_RANDOM_REFRESH;
	}
	rsp->state = rsp->state * FAULT_RANDOM_MULT + FAULT_RANDOM_ADD;
	return swahw32(rsp->state);
}

static char *
_drbd_fault_str(unsigned int type) {
	static char *_faults[] = {
		[DRBD_FAULT_MD_WR] = "Meta-data write",
		[DRBD_FAULT_MD_RD] = "Meta-data read",
		[DRBD_FAULT_RS_WR] = "Resync write",
		[DRBD_FAULT_RS_RD] = "Resync read",
		[DRBD_FAULT_DT_WR] = "Data write",
		[DRBD_FAULT_DT_RD] = "Data read",
		[DRBD_FAULT_DT_RA] = "Data read ahead",
		[DRBD_FAULT_BM_ALLOC] = "BM allocation",
3428 3429
		[DRBD_FAULT_AL_EE] = "EE allocation",
		[DRBD_FAULT_RECEIVE] = "receive data corruption",
P
Philipp Reisner 已提交
3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447
	};

	return (type < DRBD_FAULT_MAX) ? _faults[type] : "**Unknown**";
}

unsigned int
_drbd_insert_fault(struct drbd_conf *mdev, unsigned int type)
{
	static struct fault_random_state rrs = {0, 0};

	unsigned int ret = (
		(fault_devs == 0 ||
			((1 << mdev_to_minor(mdev)) & fault_devs) != 0) &&
		(((_drbd_fault_random(&rrs) % 100) + 1) <= fault_rate));

	if (ret) {
		fault_count++;

3448
		if (__ratelimit(&drbd_ratelimit_state))
P
Philipp Reisner 已提交
3449 3450 3451 3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468 3469 3470 3471 3472 3473 3474 3475 3476 3477 3478 3479 3480 3481 3482
			dev_warn(DEV, "***Simulating %s failure\n",
				_drbd_fault_str(type));
	}

	return ret;
}
#endif

const char *drbd_buildtag(void)
{
	/* DRBD built from external sources has here a reference to the
	   git hash of the source code. */

	static char buildtag[38] = "\0uilt-in";

	if (buildtag[0] == 0) {
#ifdef CONFIG_MODULES
		if (THIS_MODULE != NULL)
			sprintf(buildtag, "srcversion: %-24s", THIS_MODULE->srcversion);
		else
#endif
			buildtag[0] = 'b';
	}

	return buildtag;
}

module_init(drbd_init)
module_exit(drbd_cleanup)

EXPORT_SYMBOL(drbd_conn_str);
EXPORT_SYMBOL(drbd_role_str);
EXPORT_SYMBOL(drbd_disk_str);
EXPORT_SYMBOL(drbd_set_st_err_str);