1 // SPDX-License-Identifier: GPL-2.0-or-later
2 /*
3 drbd.c
4
5 This file is part of DRBD by Philipp Reisner and Lars Ellenberg.
6
7 Copyright (C) 2001-2008, LINBIT Information Technologies GmbH.
8 Copyright (C) 1999-2008, Philipp Reisner <philipp.reisner@linbit.com>.
9 Copyright (C) 2002-2008, Lars Ellenberg <lars.ellenberg@linbit.com>.
10
11 Thanks to Carter Burden, Bart Grantham and Gennadiy Nerubayev
12 from Logicworks, Inc. for making SDP replication support possible.
13
14
15 */
16
17 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
18
19 #include <linux/module.h>
20 #include <linux/jiffies.h>
21 #include <linux/drbd.h>
22 #include <linux/uaccess.h>
23 #include <asm/types.h>
24 #include <net/sock.h>
25 #include <linux/ctype.h>
26 #include <linux/mutex.h>
27 #include <linux/fs.h>
28 #include <linux/file.h>
29 #include <linux/proc_fs.h>
30 #include <linux/init.h>
31 #include <linux/mm.h>
32 #include <linux/memcontrol.h>
33 #include <linux/mm_inline.h>
34 #include <linux/slab.h>
35 #include <linux/random.h>
36 #include <linux/reboot.h>
37 #include <linux/notifier.h>
38 #include <linux/kthread.h>
39 #include <linux/workqueue.h>
40 #define __KERNEL_SYSCALLS__
41 #include <linux/unistd.h>
42 #include <linux/vmalloc.h>
43 #include <linux/sched/signal.h>
44
45 #include <linux/drbd_limits.h>
46 #include "drbd_int.h"
47 #include "drbd_protocol.h"
48 #include "drbd_req.h" /* only for _req_mod in tl_release and tl_clear */
49 #include "drbd_vli.h"
50 #include "drbd_debugfs.h"
51
52 static DEFINE_MUTEX(drbd_main_mutex);
53 static int drbd_open(struct block_device *bdev, fmode_t mode);
54 static void drbd_release(struct gendisk *gd, fmode_t mode);
55 static void md_sync_timer_fn(struct timer_list *t);
56 static int w_bitmap_io(struct drbd_work *w, int unused);
57
58 MODULE_AUTHOR("Philipp Reisner <phil@linbit.com>, "
59 "Lars Ellenberg <lars@linbit.com>");
60 MODULE_DESCRIPTION("drbd - Distributed Replicated Block Device v" REL_VERSION);
61 MODULE_VERSION(REL_VERSION);
62 MODULE_LICENSE("GPL");
63 MODULE_PARM_DESC(minor_count, "Approximate number of drbd devices ("
64 __stringify(DRBD_MINOR_COUNT_MIN) "-" __stringify(DRBD_MINOR_COUNT_MAX) ")");
65 MODULE_ALIAS_BLOCKDEV_MAJOR(DRBD_MAJOR);
66
67 #include <linux/moduleparam.h>
68 /* thanks to these macros, if compiled into the kernel (not-module),
69 * these become boot parameters (e.g., drbd.minor_count) */
70
71 #ifdef CONFIG_DRBD_FAULT_INJECTION
72 int drbd_enable_faults;
73 int drbd_fault_rate;
74 static int drbd_fault_count;
75 static int drbd_fault_devs;
76 /* bitmap of enabled faults */
77 module_param_named(enable_faults, drbd_enable_faults, int, 0664);
78 /* fault rate % value - applies to all enabled faults */
79 module_param_named(fault_rate, drbd_fault_rate, int, 0664);
80 /* count of faults inserted */
81 module_param_named(fault_count, drbd_fault_count, int, 0664);
82 /* bitmap of devices to insert faults on */
83 module_param_named(fault_devs, drbd_fault_devs, int, 0644);
84 #endif
85
86 /* module parameters we can keep static */
87 static bool drbd_allow_oos; /* allow_open_on_secondary */
88 static bool drbd_disable_sendpage;
89 MODULE_PARM_DESC(allow_oos, "DONT USE!");
90 module_param_named(allow_oos, drbd_allow_oos, bool, 0);
91 module_param_named(disable_sendpage, drbd_disable_sendpage, bool, 0644);
92
93 /* module parameters we share */
94 int drbd_proc_details; /* Detail level in proc drbd*/
95 module_param_named(proc_details, drbd_proc_details, int, 0644);
96 /* module parameters shared with defaults */
97 unsigned int drbd_minor_count = DRBD_MINOR_COUNT_DEF;
98 /* Module parameter for setting the user mode helper program
99 * to run. Default is /sbin/drbdadm */
100 char drbd_usermode_helper[80] = "/sbin/drbdadm";
101 module_param_named(minor_count, drbd_minor_count, uint, 0444);
102 module_param_string(usermode_helper, drbd_usermode_helper, sizeof(drbd_usermode_helper), 0644);
103
104 /* in 2.6.x, our device mapping and config info contains our virtual gendisks
105 * as member "struct gendisk *vdisk;"
106 */
107 struct idr drbd_devices;
108 struct list_head drbd_resources;
109 struct mutex resources_mutex;
110
111 struct kmem_cache *drbd_request_cache;
112 struct kmem_cache *drbd_ee_cache; /* peer requests */
113 struct kmem_cache *drbd_bm_ext_cache; /* bitmap extents */
114 struct kmem_cache *drbd_al_ext_cache; /* activity log extents */
115 mempool_t drbd_request_mempool;
116 mempool_t drbd_ee_mempool;
117 mempool_t drbd_md_io_page_pool;
118 struct bio_set drbd_md_io_bio_set;
119 struct bio_set drbd_io_bio_set;
120
121 /* I do not use a standard mempool, because:
122 1) I want to hand out the pre-allocated objects first.
123 2) I want to be able to interrupt sleeping allocation with a signal.
124 Note: This is a single linked list, the next pointer is the private
125 member of struct page.
126 */
127 struct page *drbd_pp_pool;
128 DEFINE_SPINLOCK(drbd_pp_lock);
129 int drbd_pp_vacant;
130 wait_queue_head_t drbd_pp_wait;
131
132 DEFINE_RATELIMIT_STATE(drbd_ratelimit_state, 5 * HZ, 5);
133
134 static const struct block_device_operations drbd_ops = {
135 .owner = THIS_MODULE,
136 .submit_bio = drbd_submit_bio,
137 .open = drbd_open,
138 .release = drbd_release,
139 };
140
141 #ifdef __CHECKER__
142 /* When checking with sparse, and this is an inline function, sparse will
143 give tons of false positives. When this is a real functions sparse works.
144 */
_get_ldev_if_state(struct drbd_device * device,enum drbd_disk_state mins)145 int _get_ldev_if_state(struct drbd_device *device, enum drbd_disk_state mins)
146 {
147 int io_allowed;
148
149 atomic_inc(&device->local_cnt);
150 io_allowed = (device->state.disk >= mins);
151 if (!io_allowed) {
152 if (atomic_dec_and_test(&device->local_cnt))
153 wake_up(&device->misc_wait);
154 }
155 return io_allowed;
156 }
157
158 #endif
159
160 /**
161 * tl_release() - mark as BARRIER_ACKED all requests in the corresponding transfer log epoch
162 * @connection: DRBD connection.
163 * @barrier_nr: Expected identifier of the DRBD write barrier packet.
164 * @set_size: Expected number of requests before that barrier.
165 *
166 * In case the passed barrier_nr or set_size does not match the oldest
167 * epoch of not yet barrier-acked requests, this function will cause a
168 * termination of the connection.
169 */
tl_release(struct drbd_connection * connection,unsigned int barrier_nr,unsigned int set_size)170 void tl_release(struct drbd_connection *connection, unsigned int barrier_nr,
171 unsigned int set_size)
172 {
173 struct drbd_request *r;
174 struct drbd_request *req = NULL, *tmp = NULL;
175 int expect_epoch = 0;
176 int expect_size = 0;
177
178 spin_lock_irq(&connection->resource->req_lock);
179
180 /* find oldest not yet barrier-acked write request,
181 * count writes in its epoch. */
182 list_for_each_entry(r, &connection->transfer_log, tl_requests) {
183 const unsigned s = r->rq_state;
184 if (!req) {
185 if (!(s & RQ_WRITE))
186 continue;
187 if (!(s & RQ_NET_MASK))
188 continue;
189 if (s & RQ_NET_DONE)
190 continue;
191 req = r;
192 expect_epoch = req->epoch;
193 expect_size ++;
194 } else {
195 if (r->epoch != expect_epoch)
196 break;
197 if (!(s & RQ_WRITE))
198 continue;
199 /* if (s & RQ_DONE): not expected */
200 /* if (!(s & RQ_NET_MASK)): not expected */
201 expect_size++;
202 }
203 }
204
205 /* first some paranoia code */
206 if (req == NULL) {
207 drbd_err(connection, "BAD! BarrierAck #%u received, but no epoch in tl!?\n",
208 barrier_nr);
209 goto bail;
210 }
211 if (expect_epoch != barrier_nr) {
212 drbd_err(connection, "BAD! BarrierAck #%u received, expected #%u!\n",
213 barrier_nr, expect_epoch);
214 goto bail;
215 }
216
217 if (expect_size != set_size) {
218 drbd_err(connection, "BAD! BarrierAck #%u received with n_writes=%u, expected n_writes=%u!\n",
219 barrier_nr, set_size, expect_size);
220 goto bail;
221 }
222
223 /* Clean up list of requests processed during current epoch. */
224 /* this extra list walk restart is paranoia,
225 * to catch requests being barrier-acked "unexpectedly".
226 * It usually should find the same req again, or some READ preceding it. */
227 list_for_each_entry(req, &connection->transfer_log, tl_requests)
228 if (req->epoch == expect_epoch) {
229 tmp = req;
230 break;
231 }
232 req = list_prepare_entry(tmp, &connection->transfer_log, tl_requests);
233 list_for_each_entry_safe_from(req, r, &connection->transfer_log, tl_requests) {
234 if (req->epoch != expect_epoch)
235 break;
236 _req_mod(req, BARRIER_ACKED);
237 }
238 spin_unlock_irq(&connection->resource->req_lock);
239
240 return;
241
242 bail:
243 spin_unlock_irq(&connection->resource->req_lock);
244 conn_request_state(connection, NS(conn, C_PROTOCOL_ERROR), CS_HARD);
245 }
246
247
248 /**
249 * _tl_restart() - Walks the transfer log, and applies an action to all requests
250 * @connection: DRBD connection to operate on.
251 * @what: The action/event to perform with all request objects
252 *
253 * @what might be one of CONNECTION_LOST_WHILE_PENDING, RESEND, FAIL_FROZEN_DISK_IO,
254 * RESTART_FROZEN_DISK_IO.
255 */
256 /* must hold resource->req_lock */
_tl_restart(struct drbd_connection * connection,enum drbd_req_event what)257 void _tl_restart(struct drbd_connection *connection, enum drbd_req_event what)
258 {
259 struct drbd_request *req, *r;
260
261 list_for_each_entry_safe(req, r, &connection->transfer_log, tl_requests)
262 _req_mod(req, what);
263 }
264
tl_restart(struct drbd_connection * connection,enum drbd_req_event what)265 void tl_restart(struct drbd_connection *connection, enum drbd_req_event what)
266 {
267 spin_lock_irq(&connection->resource->req_lock);
268 _tl_restart(connection, what);
269 spin_unlock_irq(&connection->resource->req_lock);
270 }
271
272 /**
273 * tl_clear() - Clears all requests and &struct drbd_tl_epoch objects out of the TL
274 * @connection: DRBD connection.
275 *
276 * This is called after the connection to the peer was lost. The storage covered
277 * by the requests on the transfer gets marked as our of sync. Called from the
278 * receiver thread and the worker thread.
279 */
tl_clear(struct drbd_connection * connection)280 void tl_clear(struct drbd_connection *connection)
281 {
282 tl_restart(connection, CONNECTION_LOST_WHILE_PENDING);
283 }
284
285 /**
286 * tl_abort_disk_io() - Abort disk I/O for all requests for a certain device in the TL
287 * @device: DRBD device.
288 */
tl_abort_disk_io(struct drbd_device * device)289 void tl_abort_disk_io(struct drbd_device *device)
290 {
291 struct drbd_connection *connection = first_peer_device(device)->connection;
292 struct drbd_request *req, *r;
293
294 spin_lock_irq(&connection->resource->req_lock);
295 list_for_each_entry_safe(req, r, &connection->transfer_log, tl_requests) {
296 if (!(req->rq_state & RQ_LOCAL_PENDING))
297 continue;
298 if (req->device != device)
299 continue;
300 _req_mod(req, ABORT_DISK_IO);
301 }
302 spin_unlock_irq(&connection->resource->req_lock);
303 }
304
drbd_thread_setup(void * arg)305 static int drbd_thread_setup(void *arg)
306 {
307 struct drbd_thread *thi = (struct drbd_thread *) arg;
308 struct drbd_resource *resource = thi->resource;
309 unsigned long flags;
310 int retval;
311
312 snprintf(current->comm, sizeof(current->comm), "drbd_%c_%s",
313 thi->name[0],
314 resource->name);
315
316 allow_kernel_signal(DRBD_SIGKILL);
317 allow_kernel_signal(SIGXCPU);
318 restart:
319 retval = thi->function(thi);
320
321 spin_lock_irqsave(&thi->t_lock, flags);
322
323 /* if the receiver has been "EXITING", the last thing it did
324 * was set the conn state to "StandAlone",
325 * if now a re-connect request comes in, conn state goes C_UNCONNECTED,
326 * and receiver thread will be "started".
327 * drbd_thread_start needs to set "RESTARTING" in that case.
328 * t_state check and assignment needs to be within the same spinlock,
329 * so either thread_start sees EXITING, and can remap to RESTARTING,
330 * or thread_start see NONE, and can proceed as normal.
331 */
332
333 if (thi->t_state == RESTARTING) {
334 drbd_info(resource, "Restarting %s thread\n", thi->name);
335 thi->t_state = RUNNING;
336 spin_unlock_irqrestore(&thi->t_lock, flags);
337 goto restart;
338 }
339
340 thi->task = NULL;
341 thi->t_state = NONE;
342 smp_mb();
343 complete_all(&thi->stop);
344 spin_unlock_irqrestore(&thi->t_lock, flags);
345
346 drbd_info(resource, "Terminating %s\n", current->comm);
347
348 /* Release mod reference taken when thread was started */
349
350 if (thi->connection)
351 kref_put(&thi->connection->kref, drbd_destroy_connection);
352 kref_put(&resource->kref, drbd_destroy_resource);
353 module_put(THIS_MODULE);
354 return retval;
355 }
356
drbd_thread_init(struct drbd_resource * resource,struct drbd_thread * thi,int (* func)(struct drbd_thread *),const char * name)357 static void drbd_thread_init(struct drbd_resource *resource, struct drbd_thread *thi,
358 int (*func) (struct drbd_thread *), const char *name)
359 {
360 spin_lock_init(&thi->t_lock);
361 thi->task = NULL;
362 thi->t_state = NONE;
363 thi->function = func;
364 thi->resource = resource;
365 thi->connection = NULL;
366 thi->name = name;
367 }
368
drbd_thread_start(struct drbd_thread * thi)369 int drbd_thread_start(struct drbd_thread *thi)
370 {
371 struct drbd_resource *resource = thi->resource;
372 struct task_struct *nt;
373 unsigned long flags;
374
375 /* is used from state engine doing drbd_thread_stop_nowait,
376 * while holding the req lock irqsave */
377 spin_lock_irqsave(&thi->t_lock, flags);
378
379 switch (thi->t_state) {
380 case NONE:
381 drbd_info(resource, "Starting %s thread (from %s [%d])\n",
382 thi->name, current->comm, current->pid);
383
384 /* Get ref on module for thread - this is released when thread exits */
385 if (!try_module_get(THIS_MODULE)) {
386 drbd_err(resource, "Failed to get module reference in drbd_thread_start\n");
387 spin_unlock_irqrestore(&thi->t_lock, flags);
388 return false;
389 }
390
391 kref_get(&resource->kref);
392 if (thi->connection)
393 kref_get(&thi->connection->kref);
394
395 init_completion(&thi->stop);
396 thi->reset_cpu_mask = 1;
397 thi->t_state = RUNNING;
398 spin_unlock_irqrestore(&thi->t_lock, flags);
399 flush_signals(current); /* otherw. may get -ERESTARTNOINTR */
400
401 nt = kthread_create(drbd_thread_setup, (void *) thi,
402 "drbd_%c_%s", thi->name[0], thi->resource->name);
403
404 if (IS_ERR(nt)) {
405 drbd_err(resource, "Couldn't start thread\n");
406
407 if (thi->connection)
408 kref_put(&thi->connection->kref, drbd_destroy_connection);
409 kref_put(&resource->kref, drbd_destroy_resource);
410 module_put(THIS_MODULE);
411 return false;
412 }
413 spin_lock_irqsave(&thi->t_lock, flags);
414 thi->task = nt;
415 thi->t_state = RUNNING;
416 spin_unlock_irqrestore(&thi->t_lock, flags);
417 wake_up_process(nt);
418 break;
419 case EXITING:
420 thi->t_state = RESTARTING;
421 drbd_info(resource, "Restarting %s thread (from %s [%d])\n",
422 thi->name, current->comm, current->pid);
423 fallthrough;
424 case RUNNING:
425 case RESTARTING:
426 default:
427 spin_unlock_irqrestore(&thi->t_lock, flags);
428 break;
429 }
430
431 return true;
432 }
433
434
_drbd_thread_stop(struct drbd_thread * thi,int restart,int wait)435 void _drbd_thread_stop(struct drbd_thread *thi, int restart, int wait)
436 {
437 unsigned long flags;
438
439 enum drbd_thread_state ns = restart ? RESTARTING : EXITING;
440
441 /* may be called from state engine, holding the req lock irqsave */
442 spin_lock_irqsave(&thi->t_lock, flags);
443
444 if (thi->t_state == NONE) {
445 spin_unlock_irqrestore(&thi->t_lock, flags);
446 if (restart)
447 drbd_thread_start(thi);
448 return;
449 }
450
451 if (thi->t_state != ns) {
452 if (thi->task == NULL) {
453 spin_unlock_irqrestore(&thi->t_lock, flags);
454 return;
455 }
456
457 thi->t_state = ns;
458 smp_mb();
459 init_completion(&thi->stop);
460 if (thi->task != current)
461 send_sig(DRBD_SIGKILL, thi->task, 1);
462 }
463
464 spin_unlock_irqrestore(&thi->t_lock, flags);
465
466 if (wait)
467 wait_for_completion(&thi->stop);
468 }
469
conn_lowest_minor(struct drbd_connection * connection)470 int conn_lowest_minor(struct drbd_connection *connection)
471 {
472 struct drbd_peer_device *peer_device;
473 int vnr = 0, minor = -1;
474
475 rcu_read_lock();
476 peer_device = idr_get_next(&connection->peer_devices, &vnr);
477 if (peer_device)
478 minor = device_to_minor(peer_device->device);
479 rcu_read_unlock();
480
481 return minor;
482 }
483
484 #ifdef CONFIG_SMP
485 /*
486 * drbd_calc_cpu_mask() - Generate CPU masks, spread over all CPUs
487 *
488 * Forces all threads of a resource onto the same CPU. This is beneficial for
489 * DRBD's performance. May be overwritten by user's configuration.
490 */
drbd_calc_cpu_mask(cpumask_var_t * cpu_mask)491 static void drbd_calc_cpu_mask(cpumask_var_t *cpu_mask)
492 {
493 unsigned int *resources_per_cpu, min_index = ~0;
494
495 resources_per_cpu = kcalloc(nr_cpu_ids, sizeof(*resources_per_cpu),
496 GFP_KERNEL);
497 if (resources_per_cpu) {
498 struct drbd_resource *resource;
499 unsigned int cpu, min = ~0;
500
501 rcu_read_lock();
502 for_each_resource_rcu(resource, &drbd_resources) {
503 for_each_cpu(cpu, resource->cpu_mask)
504 resources_per_cpu[cpu]++;
505 }
506 rcu_read_unlock();
507 for_each_online_cpu(cpu) {
508 if (resources_per_cpu[cpu] < min) {
509 min = resources_per_cpu[cpu];
510 min_index = cpu;
511 }
512 }
513 kfree(resources_per_cpu);
514 }
515 if (min_index == ~0) {
516 cpumask_setall(*cpu_mask);
517 return;
518 }
519 cpumask_set_cpu(min_index, *cpu_mask);
520 }
521
522 /**
523 * drbd_thread_current_set_cpu() - modifies the cpu mask of the _current_ thread
524 * @thi: drbd_thread object
525 *
526 * call in the "main loop" of _all_ threads, no need for any mutex, current won't die
527 * prematurely.
528 */
drbd_thread_current_set_cpu(struct drbd_thread * thi)529 void drbd_thread_current_set_cpu(struct drbd_thread *thi)
530 {
531 struct drbd_resource *resource = thi->resource;
532 struct task_struct *p = current;
533
534 if (!thi->reset_cpu_mask)
535 return;
536 thi->reset_cpu_mask = 0;
537 set_cpus_allowed_ptr(p, resource->cpu_mask);
538 }
539 #else
540 #define drbd_calc_cpu_mask(A) ({})
541 #endif
542
543 /*
544 * drbd_header_size - size of a packet header
545 *
546 * The header size is a multiple of 8, so any payload following the header is
547 * word aligned on 64-bit architectures. (The bitmap send and receive code
548 * relies on this.)
549 */
drbd_header_size(struct drbd_connection * connection)550 unsigned int drbd_header_size(struct drbd_connection *connection)
551 {
552 if (connection->agreed_pro_version >= 100) {
553 BUILD_BUG_ON(!IS_ALIGNED(sizeof(struct p_header100), 8));
554 return sizeof(struct p_header100);
555 } else {
556 BUILD_BUG_ON(sizeof(struct p_header80) !=
557 sizeof(struct p_header95));
558 BUILD_BUG_ON(!IS_ALIGNED(sizeof(struct p_header80), 8));
559 return sizeof(struct p_header80);
560 }
561 }
562
prepare_header80(struct p_header80 * h,enum drbd_packet cmd,int size)563 static unsigned int prepare_header80(struct p_header80 *h, enum drbd_packet cmd, int size)
564 {
565 h->magic = cpu_to_be32(DRBD_MAGIC);
566 h->command = cpu_to_be16(cmd);
567 h->length = cpu_to_be16(size);
568 return sizeof(struct p_header80);
569 }
570
prepare_header95(struct p_header95 * h,enum drbd_packet cmd,int size)571 static unsigned int prepare_header95(struct p_header95 *h, enum drbd_packet cmd, int size)
572 {
573 h->magic = cpu_to_be16(DRBD_MAGIC_BIG);
574 h->command = cpu_to_be16(cmd);
575 h->length = cpu_to_be32(size);
576 return sizeof(struct p_header95);
577 }
578
prepare_header100(struct p_header100 * h,enum drbd_packet cmd,int size,int vnr)579 static unsigned int prepare_header100(struct p_header100 *h, enum drbd_packet cmd,
580 int size, int vnr)
581 {
582 h->magic = cpu_to_be32(DRBD_MAGIC_100);
583 h->volume = cpu_to_be16(vnr);
584 h->command = cpu_to_be16(cmd);
585 h->length = cpu_to_be32(size);
586 h->pad = 0;
587 return sizeof(struct p_header100);
588 }
589
prepare_header(struct drbd_connection * connection,int vnr,void * buffer,enum drbd_packet cmd,int size)590 static unsigned int prepare_header(struct drbd_connection *connection, int vnr,
591 void *buffer, enum drbd_packet cmd, int size)
592 {
593 if (connection->agreed_pro_version >= 100)
594 return prepare_header100(buffer, cmd, size, vnr);
595 else if (connection->agreed_pro_version >= 95 &&
596 size > DRBD_MAX_SIZE_H80_PACKET)
597 return prepare_header95(buffer, cmd, size);
598 else
599 return prepare_header80(buffer, cmd, size);
600 }
601
__conn_prepare_command(struct drbd_connection * connection,struct drbd_socket * sock)602 static void *__conn_prepare_command(struct drbd_connection *connection,
603 struct drbd_socket *sock)
604 {
605 if (!sock->socket)
606 return NULL;
607 return sock->sbuf + drbd_header_size(connection);
608 }
609
conn_prepare_command(struct drbd_connection * connection,struct drbd_socket * sock)610 void *conn_prepare_command(struct drbd_connection *connection, struct drbd_socket *sock)
611 {
612 void *p;
613
614 mutex_lock(&sock->mutex);
615 p = __conn_prepare_command(connection, sock);
616 if (!p)
617 mutex_unlock(&sock->mutex);
618
619 return p;
620 }
621
drbd_prepare_command(struct drbd_peer_device * peer_device,struct drbd_socket * sock)622 void *drbd_prepare_command(struct drbd_peer_device *peer_device, struct drbd_socket *sock)
623 {
624 return conn_prepare_command(peer_device->connection, sock);
625 }
626
__send_command(struct drbd_connection * connection,int vnr,struct drbd_socket * sock,enum drbd_packet cmd,unsigned int header_size,void * data,unsigned int size)627 static int __send_command(struct drbd_connection *connection, int vnr,
628 struct drbd_socket *sock, enum drbd_packet cmd,
629 unsigned int header_size, void *data,
630 unsigned int size)
631 {
632 int msg_flags;
633 int err;
634
635 /*
636 * Called with @data == NULL and the size of the data blocks in @size
637 * for commands that send data blocks. For those commands, omit the
638 * MSG_MORE flag: this will increase the likelihood that data blocks
639 * which are page aligned on the sender will end up page aligned on the
640 * receiver.
641 */
642 msg_flags = data ? MSG_MORE : 0;
643
644 header_size += prepare_header(connection, vnr, sock->sbuf, cmd,
645 header_size + size);
646 err = drbd_send_all(connection, sock->socket, sock->sbuf, header_size,
647 msg_flags);
648 if (data && !err)
649 err = drbd_send_all(connection, sock->socket, data, size, 0);
650 /* DRBD protocol "pings" are latency critical.
651 * This is supposed to trigger tcp_push_pending_frames() */
652 if (!err && (cmd == P_PING || cmd == P_PING_ACK))
653 tcp_sock_set_nodelay(sock->socket->sk);
654
655 return err;
656 }
657
__conn_send_command(struct drbd_connection * connection,struct drbd_socket * sock,enum drbd_packet cmd,unsigned int header_size,void * data,unsigned int size)658 static int __conn_send_command(struct drbd_connection *connection, struct drbd_socket *sock,
659 enum drbd_packet cmd, unsigned int header_size,
660 void *data, unsigned int size)
661 {
662 return __send_command(connection, 0, sock, cmd, header_size, data, size);
663 }
664
conn_send_command(struct drbd_connection * connection,struct drbd_socket * sock,enum drbd_packet cmd,unsigned int header_size,void * data,unsigned int size)665 int conn_send_command(struct drbd_connection *connection, struct drbd_socket *sock,
666 enum drbd_packet cmd, unsigned int header_size,
667 void *data, unsigned int size)
668 {
669 int err;
670
671 err = __conn_send_command(connection, sock, cmd, header_size, data, size);
672 mutex_unlock(&sock->mutex);
673 return err;
674 }
675
drbd_send_command(struct drbd_peer_device * peer_device,struct drbd_socket * sock,enum drbd_packet cmd,unsigned int header_size,void * data,unsigned int size)676 int drbd_send_command(struct drbd_peer_device *peer_device, struct drbd_socket *sock,
677 enum drbd_packet cmd, unsigned int header_size,
678 void *data, unsigned int size)
679 {
680 int err;
681
682 err = __send_command(peer_device->connection, peer_device->device->vnr,
683 sock, cmd, header_size, data, size);
684 mutex_unlock(&sock->mutex);
685 return err;
686 }
687
drbd_send_ping(struct drbd_connection * connection)688 int drbd_send_ping(struct drbd_connection *connection)
689 {
690 struct drbd_socket *sock;
691
692 sock = &connection->meta;
693 if (!conn_prepare_command(connection, sock))
694 return -EIO;
695 return conn_send_command(connection, sock, P_PING, 0, NULL, 0);
696 }
697
drbd_send_ping_ack(struct drbd_connection * connection)698 int drbd_send_ping_ack(struct drbd_connection *connection)
699 {
700 struct drbd_socket *sock;
701
702 sock = &connection->meta;
703 if (!conn_prepare_command(connection, sock))
704 return -EIO;
705 return conn_send_command(connection, sock, P_PING_ACK, 0, NULL, 0);
706 }
707
drbd_send_sync_param(struct drbd_peer_device * peer_device)708 int drbd_send_sync_param(struct drbd_peer_device *peer_device)
709 {
710 struct drbd_socket *sock;
711 struct p_rs_param_95 *p;
712 int size;
713 const int apv = peer_device->connection->agreed_pro_version;
714 enum drbd_packet cmd;
715 struct net_conf *nc;
716 struct disk_conf *dc;
717
718 sock = &peer_device->connection->data;
719 p = drbd_prepare_command(peer_device, sock);
720 if (!p)
721 return -EIO;
722
723 rcu_read_lock();
724 nc = rcu_dereference(peer_device->connection->net_conf);
725
726 size = apv <= 87 ? sizeof(struct p_rs_param)
727 : apv == 88 ? sizeof(struct p_rs_param)
728 + strlen(nc->verify_alg) + 1
729 : apv <= 94 ? sizeof(struct p_rs_param_89)
730 : /* apv >= 95 */ sizeof(struct p_rs_param_95);
731
732 cmd = apv >= 89 ? P_SYNC_PARAM89 : P_SYNC_PARAM;
733
734 /* initialize verify_alg and csums_alg */
735 BUILD_BUG_ON(sizeof(p->algs) != 2 * SHARED_SECRET_MAX);
736 memset(&p->algs, 0, sizeof(p->algs));
737
738 if (get_ldev(peer_device->device)) {
739 dc = rcu_dereference(peer_device->device->ldev->disk_conf);
740 p->resync_rate = cpu_to_be32(dc->resync_rate);
741 p->c_plan_ahead = cpu_to_be32(dc->c_plan_ahead);
742 p->c_delay_target = cpu_to_be32(dc->c_delay_target);
743 p->c_fill_target = cpu_to_be32(dc->c_fill_target);
744 p->c_max_rate = cpu_to_be32(dc->c_max_rate);
745 put_ldev(peer_device->device);
746 } else {
747 p->resync_rate = cpu_to_be32(DRBD_RESYNC_RATE_DEF);
748 p->c_plan_ahead = cpu_to_be32(DRBD_C_PLAN_AHEAD_DEF);
749 p->c_delay_target = cpu_to_be32(DRBD_C_DELAY_TARGET_DEF);
750 p->c_fill_target = cpu_to_be32(DRBD_C_FILL_TARGET_DEF);
751 p->c_max_rate = cpu_to_be32(DRBD_C_MAX_RATE_DEF);
752 }
753
754 if (apv >= 88)
755 strcpy(p->verify_alg, nc->verify_alg);
756 if (apv >= 89)
757 strcpy(p->csums_alg, nc->csums_alg);
758 rcu_read_unlock();
759
760 return drbd_send_command(peer_device, sock, cmd, size, NULL, 0);
761 }
762
__drbd_send_protocol(struct drbd_connection * connection,enum drbd_packet cmd)763 int __drbd_send_protocol(struct drbd_connection *connection, enum drbd_packet cmd)
764 {
765 struct drbd_socket *sock;
766 struct p_protocol *p;
767 struct net_conf *nc;
768 int size, cf;
769
770 sock = &connection->data;
771 p = __conn_prepare_command(connection, sock);
772 if (!p)
773 return -EIO;
774
775 rcu_read_lock();
776 nc = rcu_dereference(connection->net_conf);
777
778 if (nc->tentative && connection->agreed_pro_version < 92) {
779 rcu_read_unlock();
780 drbd_err(connection, "--dry-run is not supported by peer");
781 return -EOPNOTSUPP;
782 }
783
784 size = sizeof(*p);
785 if (connection->agreed_pro_version >= 87)
786 size += strlen(nc->integrity_alg) + 1;
787
788 p->protocol = cpu_to_be32(nc->wire_protocol);
789 p->after_sb_0p = cpu_to_be32(nc->after_sb_0p);
790 p->after_sb_1p = cpu_to_be32(nc->after_sb_1p);
791 p->after_sb_2p = cpu_to_be32(nc->after_sb_2p);
792 p->two_primaries = cpu_to_be32(nc->two_primaries);
793 cf = 0;
794 if (nc->discard_my_data)
795 cf |= CF_DISCARD_MY_DATA;
796 if (nc->tentative)
797 cf |= CF_DRY_RUN;
798 p->conn_flags = cpu_to_be32(cf);
799
800 if (connection->agreed_pro_version >= 87)
801 strcpy(p->integrity_alg, nc->integrity_alg);
802 rcu_read_unlock();
803
804 return __conn_send_command(connection, sock, cmd, size, NULL, 0);
805 }
806
drbd_send_protocol(struct drbd_connection * connection)807 int drbd_send_protocol(struct drbd_connection *connection)
808 {
809 int err;
810
811 mutex_lock(&connection->data.mutex);
812 err = __drbd_send_protocol(connection, P_PROTOCOL);
813 mutex_unlock(&connection->data.mutex);
814
815 return err;
816 }
817
_drbd_send_uuids(struct drbd_peer_device * peer_device,u64 uuid_flags)818 static int _drbd_send_uuids(struct drbd_peer_device *peer_device, u64 uuid_flags)
819 {
820 struct drbd_device *device = peer_device->device;
821 struct drbd_socket *sock;
822 struct p_uuids *p;
823 int i;
824
825 if (!get_ldev_if_state(device, D_NEGOTIATING))
826 return 0;
827
828 sock = &peer_device->connection->data;
829 p = drbd_prepare_command(peer_device, sock);
830 if (!p) {
831 put_ldev(device);
832 return -EIO;
833 }
834 spin_lock_irq(&device->ldev->md.uuid_lock);
835 for (i = UI_CURRENT; i < UI_SIZE; i++)
836 p->uuid[i] = cpu_to_be64(device->ldev->md.uuid[i]);
837 spin_unlock_irq(&device->ldev->md.uuid_lock);
838
839 device->comm_bm_set = drbd_bm_total_weight(device);
840 p->uuid[UI_SIZE] = cpu_to_be64(device->comm_bm_set);
841 rcu_read_lock();
842 uuid_flags |= rcu_dereference(peer_device->connection->net_conf)->discard_my_data ? 1 : 0;
843 rcu_read_unlock();
844 uuid_flags |= test_bit(CRASHED_PRIMARY, &device->flags) ? 2 : 0;
845 uuid_flags |= device->new_state_tmp.disk == D_INCONSISTENT ? 4 : 0;
846 p->uuid[UI_FLAGS] = cpu_to_be64(uuid_flags);
847
848 put_ldev(device);
849 return drbd_send_command(peer_device, sock, P_UUIDS, sizeof(*p), NULL, 0);
850 }
851
drbd_send_uuids(struct drbd_peer_device * peer_device)852 int drbd_send_uuids(struct drbd_peer_device *peer_device)
853 {
854 return _drbd_send_uuids(peer_device, 0);
855 }
856
drbd_send_uuids_skip_initial_sync(struct drbd_peer_device * peer_device)857 int drbd_send_uuids_skip_initial_sync(struct drbd_peer_device *peer_device)
858 {
859 return _drbd_send_uuids(peer_device, 8);
860 }
861
drbd_print_uuids(struct drbd_device * device,const char * text)862 void drbd_print_uuids(struct drbd_device *device, const char *text)
863 {
864 if (get_ldev_if_state(device, D_NEGOTIATING)) {
865 u64 *uuid = device->ldev->md.uuid;
866 drbd_info(device, "%s %016llX:%016llX:%016llX:%016llX\n",
867 text,
868 (unsigned long long)uuid[UI_CURRENT],
869 (unsigned long long)uuid[UI_BITMAP],
870 (unsigned long long)uuid[UI_HISTORY_START],
871 (unsigned long long)uuid[UI_HISTORY_END]);
872 put_ldev(device);
873 } else {
874 drbd_info(device, "%s effective data uuid: %016llX\n",
875 text,
876 (unsigned long long)device->ed_uuid);
877 }
878 }
879
drbd_gen_and_send_sync_uuid(struct drbd_peer_device * peer_device)880 void drbd_gen_and_send_sync_uuid(struct drbd_peer_device *peer_device)
881 {
882 struct drbd_device *device = peer_device->device;
883 struct drbd_socket *sock;
884 struct p_rs_uuid *p;
885 u64 uuid;
886
887 D_ASSERT(device, device->state.disk == D_UP_TO_DATE);
888
889 uuid = device->ldev->md.uuid[UI_BITMAP];
890 if (uuid && uuid != UUID_JUST_CREATED)
891 uuid = uuid + UUID_NEW_BM_OFFSET;
892 else
893 get_random_bytes(&uuid, sizeof(u64));
894 drbd_uuid_set(device, UI_BITMAP, uuid);
895 drbd_print_uuids(device, "updated sync UUID");
896 drbd_md_sync(device);
897
898 sock = &peer_device->connection->data;
899 p = drbd_prepare_command(peer_device, sock);
900 if (p) {
901 p->uuid = cpu_to_be64(uuid);
902 drbd_send_command(peer_device, sock, P_SYNC_UUID, sizeof(*p), NULL, 0);
903 }
904 }
905
drbd_send_sizes(struct drbd_peer_device * peer_device,int trigger_reply,enum dds_flags flags)906 int drbd_send_sizes(struct drbd_peer_device *peer_device, int trigger_reply, enum dds_flags flags)
907 {
908 struct drbd_device *device = peer_device->device;
909 struct drbd_socket *sock;
910 struct p_sizes *p;
911 sector_t d_size, u_size;
912 int q_order_type;
913 unsigned int max_bio_size;
914 unsigned int packet_size;
915
916 sock = &peer_device->connection->data;
917 p = drbd_prepare_command(peer_device, sock);
918 if (!p)
919 return -EIO;
920
921 packet_size = sizeof(*p);
922 if (peer_device->connection->agreed_features & DRBD_FF_WSAME)
923 packet_size += sizeof(p->qlim[0]);
924
925 memset(p, 0, packet_size);
926 if (get_ldev_if_state(device, D_NEGOTIATING)) {
927 struct block_device *bdev = device->ldev->backing_bdev;
928 struct request_queue *q = bdev_get_queue(bdev);
929
930 d_size = drbd_get_max_capacity(device->ldev);
931 rcu_read_lock();
932 u_size = rcu_dereference(device->ldev->disk_conf)->disk_size;
933 rcu_read_unlock();
934 q_order_type = drbd_queue_order_type(device);
935 max_bio_size = queue_max_hw_sectors(q) << 9;
936 max_bio_size = min(max_bio_size, DRBD_MAX_BIO_SIZE);
937 p->qlim->physical_block_size =
938 cpu_to_be32(bdev_physical_block_size(bdev));
939 p->qlim->logical_block_size =
940 cpu_to_be32(bdev_logical_block_size(bdev));
941 p->qlim->alignment_offset =
942 cpu_to_be32(bdev_alignment_offset(bdev));
943 p->qlim->io_min = cpu_to_be32(bdev_io_min(bdev));
944 p->qlim->io_opt = cpu_to_be32(bdev_io_opt(bdev));
945 p->qlim->discard_enabled = !!bdev_max_discard_sectors(bdev);
946 put_ldev(device);
947 } else {
948 struct request_queue *q = device->rq_queue;
949
950 p->qlim->physical_block_size =
951 cpu_to_be32(queue_physical_block_size(q));
952 p->qlim->logical_block_size =
953 cpu_to_be32(queue_logical_block_size(q));
954 p->qlim->alignment_offset = 0;
955 p->qlim->io_min = cpu_to_be32(queue_io_min(q));
956 p->qlim->io_opt = cpu_to_be32(queue_io_opt(q));
957 p->qlim->discard_enabled = 0;
958
959 d_size = 0;
960 u_size = 0;
961 q_order_type = QUEUE_ORDERED_NONE;
962 max_bio_size = DRBD_MAX_BIO_SIZE; /* ... multiple BIOs per peer_request */
963 }
964
965 if (peer_device->connection->agreed_pro_version <= 94)
966 max_bio_size = min(max_bio_size, DRBD_MAX_SIZE_H80_PACKET);
967 else if (peer_device->connection->agreed_pro_version < 100)
968 max_bio_size = min(max_bio_size, DRBD_MAX_BIO_SIZE_P95);
969
970 p->d_size = cpu_to_be64(d_size);
971 p->u_size = cpu_to_be64(u_size);
972 if (trigger_reply)
973 p->c_size = 0;
974 else
975 p->c_size = cpu_to_be64(get_capacity(device->vdisk));
976 p->max_bio_size = cpu_to_be32(max_bio_size);
977 p->queue_order_type = cpu_to_be16(q_order_type);
978 p->dds_flags = cpu_to_be16(flags);
979
980 return drbd_send_command(peer_device, sock, P_SIZES, packet_size, NULL, 0);
981 }
982
983 /**
984 * drbd_send_current_state() - Sends the drbd state to the peer
985 * @peer_device: DRBD peer device.
986 */
drbd_send_current_state(struct drbd_peer_device * peer_device)987 int drbd_send_current_state(struct drbd_peer_device *peer_device)
988 {
989 struct drbd_socket *sock;
990 struct p_state *p;
991
992 sock = &peer_device->connection->data;
993 p = drbd_prepare_command(peer_device, sock);
994 if (!p)
995 return -EIO;
996 p->state = cpu_to_be32(peer_device->device->state.i); /* Within the send mutex */
997 return drbd_send_command(peer_device, sock, P_STATE, sizeof(*p), NULL, 0);
998 }
999
1000 /**
1001 * drbd_send_state() - After a state change, sends the new state to the peer
1002 * @peer_device: DRBD peer device.
1003 * @state: the state to send, not necessarily the current state.
1004 *
1005 * Each state change queues an "after_state_ch" work, which will eventually
1006 * send the resulting new state to the peer. If more state changes happen
1007 * between queuing and processing of the after_state_ch work, we still
1008 * want to send each intermediary state in the order it occurred.
1009 */
drbd_send_state(struct drbd_peer_device * peer_device,union drbd_state state)1010 int drbd_send_state(struct drbd_peer_device *peer_device, union drbd_state state)
1011 {
1012 struct drbd_socket *sock;
1013 struct p_state *p;
1014
1015 sock = &peer_device->connection->data;
1016 p = drbd_prepare_command(peer_device, sock);
1017 if (!p)
1018 return -EIO;
1019 p->state = cpu_to_be32(state.i); /* Within the send mutex */
1020 return drbd_send_command(peer_device, sock, P_STATE, sizeof(*p), NULL, 0);
1021 }
1022
drbd_send_state_req(struct drbd_peer_device * peer_device,union drbd_state mask,union drbd_state val)1023 int drbd_send_state_req(struct drbd_peer_device *peer_device, union drbd_state mask, union drbd_state val)
1024 {
1025 struct drbd_socket *sock;
1026 struct p_req_state *p;
1027
1028 sock = &peer_device->connection->data;
1029 p = drbd_prepare_command(peer_device, sock);
1030 if (!p)
1031 return -EIO;
1032 p->mask = cpu_to_be32(mask.i);
1033 p->val = cpu_to_be32(val.i);
1034 return drbd_send_command(peer_device, sock, P_STATE_CHG_REQ, sizeof(*p), NULL, 0);
1035 }
1036
conn_send_state_req(struct drbd_connection * connection,union drbd_state mask,union drbd_state val)1037 int conn_send_state_req(struct drbd_connection *connection, union drbd_state mask, union drbd_state val)
1038 {
1039 enum drbd_packet cmd;
1040 struct drbd_socket *sock;
1041 struct p_req_state *p;
1042
1043 cmd = connection->agreed_pro_version < 100 ? P_STATE_CHG_REQ : P_CONN_ST_CHG_REQ;
1044 sock = &connection->data;
1045 p = conn_prepare_command(connection, sock);
1046 if (!p)
1047 return -EIO;
1048 p->mask = cpu_to_be32(mask.i);
1049 p->val = cpu_to_be32(val.i);
1050 return conn_send_command(connection, sock, cmd, sizeof(*p), NULL, 0);
1051 }
1052
drbd_send_sr_reply(struct drbd_peer_device * peer_device,enum drbd_state_rv retcode)1053 void drbd_send_sr_reply(struct drbd_peer_device *peer_device, enum drbd_state_rv retcode)
1054 {
1055 struct drbd_socket *sock;
1056 struct p_req_state_reply *p;
1057
1058 sock = &peer_device->connection->meta;
1059 p = drbd_prepare_command(peer_device, sock);
1060 if (p) {
1061 p->retcode = cpu_to_be32(retcode);
1062 drbd_send_command(peer_device, sock, P_STATE_CHG_REPLY, sizeof(*p), NULL, 0);
1063 }
1064 }
1065
conn_send_sr_reply(struct drbd_connection * connection,enum drbd_state_rv retcode)1066 void conn_send_sr_reply(struct drbd_connection *connection, enum drbd_state_rv retcode)
1067 {
1068 struct drbd_socket *sock;
1069 struct p_req_state_reply *p;
1070 enum drbd_packet cmd = connection->agreed_pro_version < 100 ? P_STATE_CHG_REPLY : P_CONN_ST_CHG_REPLY;
1071
1072 sock = &connection->meta;
1073 p = conn_prepare_command(connection, sock);
1074 if (p) {
1075 p->retcode = cpu_to_be32(retcode);
1076 conn_send_command(connection, sock, cmd, sizeof(*p), NULL, 0);
1077 }
1078 }
1079
dcbp_set_code(struct p_compressed_bm * p,enum drbd_bitmap_code code)1080 static void dcbp_set_code(struct p_compressed_bm *p, enum drbd_bitmap_code code)
1081 {
1082 BUG_ON(code & ~0xf);
1083 p->encoding = (p->encoding & ~0xf) | code;
1084 }
1085
dcbp_set_start(struct p_compressed_bm * p,int set)1086 static void dcbp_set_start(struct p_compressed_bm *p, int set)
1087 {
1088 p->encoding = (p->encoding & ~0x80) | (set ? 0x80 : 0);
1089 }
1090
dcbp_set_pad_bits(struct p_compressed_bm * p,int n)1091 static void dcbp_set_pad_bits(struct p_compressed_bm *p, int n)
1092 {
1093 BUG_ON(n & ~0x7);
1094 p->encoding = (p->encoding & (~0x7 << 4)) | (n << 4);
1095 }
1096
fill_bitmap_rle_bits(struct drbd_device * device,struct p_compressed_bm * p,unsigned int size,struct bm_xfer_ctx * c)1097 static int fill_bitmap_rle_bits(struct drbd_device *device,
1098 struct p_compressed_bm *p,
1099 unsigned int size,
1100 struct bm_xfer_ctx *c)
1101 {
1102 struct bitstream bs;
1103 unsigned long plain_bits;
1104 unsigned long tmp;
1105 unsigned long rl;
1106 unsigned len;
1107 unsigned toggle;
1108 int bits, use_rle;
1109
1110 /* may we use this feature? */
1111 rcu_read_lock();
1112 use_rle = rcu_dereference(first_peer_device(device)->connection->net_conf)->use_rle;
1113 rcu_read_unlock();
1114 if (!use_rle || first_peer_device(device)->connection->agreed_pro_version < 90)
1115 return 0;
1116
1117 if (c->bit_offset >= c->bm_bits)
1118 return 0; /* nothing to do. */
1119
1120 /* use at most thus many bytes */
1121 bitstream_init(&bs, p->code, size, 0);
1122 memset(p->code, 0, size);
1123 /* plain bits covered in this code string */
1124 plain_bits = 0;
1125
1126 /* p->encoding & 0x80 stores whether the first run length is set.
1127 * bit offset is implicit.
1128 * start with toggle == 2 to be able to tell the first iteration */
1129 toggle = 2;
1130
1131 /* see how much plain bits we can stuff into one packet
1132 * using RLE and VLI. */
1133 do {
1134 tmp = (toggle == 0) ? _drbd_bm_find_next_zero(device, c->bit_offset)
1135 : _drbd_bm_find_next(device, c->bit_offset);
1136 if (tmp == -1UL)
1137 tmp = c->bm_bits;
1138 rl = tmp - c->bit_offset;
1139
1140 if (toggle == 2) { /* first iteration */
1141 if (rl == 0) {
1142 /* the first checked bit was set,
1143 * store start value, */
1144 dcbp_set_start(p, 1);
1145 /* but skip encoding of zero run length */
1146 toggle = !toggle;
1147 continue;
1148 }
1149 dcbp_set_start(p, 0);
1150 }
1151
1152 /* paranoia: catch zero runlength.
1153 * can only happen if bitmap is modified while we scan it. */
1154 if (rl == 0) {
1155 drbd_err(device, "unexpected zero runlength while encoding bitmap "
1156 "t:%u bo:%lu\n", toggle, c->bit_offset);
1157 return -1;
1158 }
1159
1160 bits = vli_encode_bits(&bs, rl);
1161 if (bits == -ENOBUFS) /* buffer full */
1162 break;
1163 if (bits <= 0) {
1164 drbd_err(device, "error while encoding bitmap: %d\n", bits);
1165 return 0;
1166 }
1167
1168 toggle = !toggle;
1169 plain_bits += rl;
1170 c->bit_offset = tmp;
1171 } while (c->bit_offset < c->bm_bits);
1172
1173 len = bs.cur.b - p->code + !!bs.cur.bit;
1174
1175 if (plain_bits < (len << 3)) {
1176 /* incompressible with this method.
1177 * we need to rewind both word and bit position. */
1178 c->bit_offset -= plain_bits;
1179 bm_xfer_ctx_bit_to_word_offset(c);
1180 c->bit_offset = c->word_offset * BITS_PER_LONG;
1181 return 0;
1182 }
1183
1184 /* RLE + VLI was able to compress it just fine.
1185 * update c->word_offset. */
1186 bm_xfer_ctx_bit_to_word_offset(c);
1187
1188 /* store pad_bits */
1189 dcbp_set_pad_bits(p, (8 - bs.cur.bit) & 0x7);
1190
1191 return len;
1192 }
1193
1194 /*
1195 * send_bitmap_rle_or_plain
1196 *
1197 * Return 0 when done, 1 when another iteration is needed, and a negative error
1198 * code upon failure.
1199 */
1200 static int
send_bitmap_rle_or_plain(struct drbd_device * device,struct bm_xfer_ctx * c)1201 send_bitmap_rle_or_plain(struct drbd_device *device, struct bm_xfer_ctx *c)
1202 {
1203 struct drbd_socket *sock = &first_peer_device(device)->connection->data;
1204 unsigned int header_size = drbd_header_size(first_peer_device(device)->connection);
1205 struct p_compressed_bm *p = sock->sbuf + header_size;
1206 int len, err;
1207
1208 len = fill_bitmap_rle_bits(device, p,
1209 DRBD_SOCKET_BUFFER_SIZE - header_size - sizeof(*p), c);
1210 if (len < 0)
1211 return -EIO;
1212
1213 if (len) {
1214 dcbp_set_code(p, RLE_VLI_Bits);
1215 err = __send_command(first_peer_device(device)->connection, device->vnr, sock,
1216 P_COMPRESSED_BITMAP, sizeof(*p) + len,
1217 NULL, 0);
1218 c->packets[0]++;
1219 c->bytes[0] += header_size + sizeof(*p) + len;
1220
1221 if (c->bit_offset >= c->bm_bits)
1222 len = 0; /* DONE */
1223 } else {
1224 /* was not compressible.
1225 * send a buffer full of plain text bits instead. */
1226 unsigned int data_size;
1227 unsigned long num_words;
1228 unsigned long *p = sock->sbuf + header_size;
1229
1230 data_size = DRBD_SOCKET_BUFFER_SIZE - header_size;
1231 num_words = min_t(size_t, data_size / sizeof(*p),
1232 c->bm_words - c->word_offset);
1233 len = num_words * sizeof(*p);
1234 if (len)
1235 drbd_bm_get_lel(device, c->word_offset, num_words, p);
1236 err = __send_command(first_peer_device(device)->connection, device->vnr, sock, P_BITMAP, len, NULL, 0);
1237 c->word_offset += num_words;
1238 c->bit_offset = c->word_offset * BITS_PER_LONG;
1239
1240 c->packets[1]++;
1241 c->bytes[1] += header_size + len;
1242
1243 if (c->bit_offset > c->bm_bits)
1244 c->bit_offset = c->bm_bits;
1245 }
1246 if (!err) {
1247 if (len == 0) {
1248 INFO_bm_xfer_stats(device, "send", c);
1249 return 0;
1250 } else
1251 return 1;
1252 }
1253 return -EIO;
1254 }
1255
1256 /* See the comment at receive_bitmap() */
_drbd_send_bitmap(struct drbd_device * device)1257 static int _drbd_send_bitmap(struct drbd_device *device)
1258 {
1259 struct bm_xfer_ctx c;
1260 int err;
1261
1262 if (!expect(device->bitmap))
1263 return false;
1264
1265 if (get_ldev(device)) {
1266 if (drbd_md_test_flag(device->ldev, MDF_FULL_SYNC)) {
1267 drbd_info(device, "Writing the whole bitmap, MDF_FullSync was set.\n");
1268 drbd_bm_set_all(device);
1269 if (drbd_bm_write(device)) {
1270 /* write_bm did fail! Leave full sync flag set in Meta P_DATA
1271 * but otherwise process as per normal - need to tell other
1272 * side that a full resync is required! */
1273 drbd_err(device, "Failed to write bitmap to disk!\n");
1274 } else {
1275 drbd_md_clear_flag(device, MDF_FULL_SYNC);
1276 drbd_md_sync(device);
1277 }
1278 }
1279 put_ldev(device);
1280 }
1281
1282 c = (struct bm_xfer_ctx) {
1283 .bm_bits = drbd_bm_bits(device),
1284 .bm_words = drbd_bm_words(device),
1285 };
1286
1287 do {
1288 err = send_bitmap_rle_or_plain(device, &c);
1289 } while (err > 0);
1290
1291 return err == 0;
1292 }
1293
drbd_send_bitmap(struct drbd_device * device)1294 int drbd_send_bitmap(struct drbd_device *device)
1295 {
1296 struct drbd_socket *sock = &first_peer_device(device)->connection->data;
1297 int err = -1;
1298
1299 mutex_lock(&sock->mutex);
1300 if (sock->socket)
1301 err = !_drbd_send_bitmap(device);
1302 mutex_unlock(&sock->mutex);
1303 return err;
1304 }
1305
drbd_send_b_ack(struct drbd_connection * connection,u32 barrier_nr,u32 set_size)1306 void drbd_send_b_ack(struct drbd_connection *connection, u32 barrier_nr, u32 set_size)
1307 {
1308 struct drbd_socket *sock;
1309 struct p_barrier_ack *p;
1310
1311 if (connection->cstate < C_WF_REPORT_PARAMS)
1312 return;
1313
1314 sock = &connection->meta;
1315 p = conn_prepare_command(connection, sock);
1316 if (!p)
1317 return;
1318 p->barrier = barrier_nr;
1319 p->set_size = cpu_to_be32(set_size);
1320 conn_send_command(connection, sock, P_BARRIER_ACK, sizeof(*p), NULL, 0);
1321 }
1322
1323 /**
1324 * _drbd_send_ack() - Sends an ack packet
1325 * @peer_device: DRBD peer device.
1326 * @cmd: Packet command code.
1327 * @sector: sector, needs to be in big endian byte order
1328 * @blksize: size in byte, needs to be in big endian byte order
1329 * @block_id: Id, big endian byte order
1330 */
_drbd_send_ack(struct drbd_peer_device * peer_device,enum drbd_packet cmd,u64 sector,u32 blksize,u64 block_id)1331 static int _drbd_send_ack(struct drbd_peer_device *peer_device, enum drbd_packet cmd,
1332 u64 sector, u32 blksize, u64 block_id)
1333 {
1334 struct drbd_socket *sock;
1335 struct p_block_ack *p;
1336
1337 if (peer_device->device->state.conn < C_CONNECTED)
1338 return -EIO;
1339
1340 sock = &peer_device->connection->meta;
1341 p = drbd_prepare_command(peer_device, sock);
1342 if (!p)
1343 return -EIO;
1344 p->sector = sector;
1345 p->block_id = block_id;
1346 p->blksize = blksize;
1347 p->seq_num = cpu_to_be32(atomic_inc_return(&peer_device->device->packet_seq));
1348 return drbd_send_command(peer_device, sock, cmd, sizeof(*p), NULL, 0);
1349 }
1350
1351 /* dp->sector and dp->block_id already/still in network byte order,
1352 * data_size is payload size according to dp->head,
1353 * and may need to be corrected for digest size. */
drbd_send_ack_dp(struct drbd_peer_device * peer_device,enum drbd_packet cmd,struct p_data * dp,int data_size)1354 void drbd_send_ack_dp(struct drbd_peer_device *peer_device, enum drbd_packet cmd,
1355 struct p_data *dp, int data_size)
1356 {
1357 if (peer_device->connection->peer_integrity_tfm)
1358 data_size -= crypto_shash_digestsize(peer_device->connection->peer_integrity_tfm);
1359 _drbd_send_ack(peer_device, cmd, dp->sector, cpu_to_be32(data_size),
1360 dp->block_id);
1361 }
1362
drbd_send_ack_rp(struct drbd_peer_device * peer_device,enum drbd_packet cmd,struct p_block_req * rp)1363 void drbd_send_ack_rp(struct drbd_peer_device *peer_device, enum drbd_packet cmd,
1364 struct p_block_req *rp)
1365 {
1366 _drbd_send_ack(peer_device, cmd, rp->sector, rp->blksize, rp->block_id);
1367 }
1368
1369 /**
1370 * drbd_send_ack() - Sends an ack packet
1371 * @peer_device: DRBD peer device
1372 * @cmd: packet command code
1373 * @peer_req: peer request
1374 */
drbd_send_ack(struct drbd_peer_device * peer_device,enum drbd_packet cmd,struct drbd_peer_request * peer_req)1375 int drbd_send_ack(struct drbd_peer_device *peer_device, enum drbd_packet cmd,
1376 struct drbd_peer_request *peer_req)
1377 {
1378 return _drbd_send_ack(peer_device, cmd,
1379 cpu_to_be64(peer_req->i.sector),
1380 cpu_to_be32(peer_req->i.size),
1381 peer_req->block_id);
1382 }
1383
1384 /* This function misuses the block_id field to signal if the blocks
1385 * are is sync or not. */
drbd_send_ack_ex(struct drbd_peer_device * peer_device,enum drbd_packet cmd,sector_t sector,int blksize,u64 block_id)1386 int drbd_send_ack_ex(struct drbd_peer_device *peer_device, enum drbd_packet cmd,
1387 sector_t sector, int blksize, u64 block_id)
1388 {
1389 return _drbd_send_ack(peer_device, cmd,
1390 cpu_to_be64(sector),
1391 cpu_to_be32(blksize),
1392 cpu_to_be64(block_id));
1393 }
1394
drbd_send_rs_deallocated(struct drbd_peer_device * peer_device,struct drbd_peer_request * peer_req)1395 int drbd_send_rs_deallocated(struct drbd_peer_device *peer_device,
1396 struct drbd_peer_request *peer_req)
1397 {
1398 struct drbd_socket *sock;
1399 struct p_block_desc *p;
1400
1401 sock = &peer_device->connection->data;
1402 p = drbd_prepare_command(peer_device, sock);
1403 if (!p)
1404 return -EIO;
1405 p->sector = cpu_to_be64(peer_req->i.sector);
1406 p->blksize = cpu_to_be32(peer_req->i.size);
1407 p->pad = 0;
1408 return drbd_send_command(peer_device, sock, P_RS_DEALLOCATED, sizeof(*p), NULL, 0);
1409 }
1410
drbd_send_drequest(struct drbd_peer_device * peer_device,int cmd,sector_t sector,int size,u64 block_id)1411 int drbd_send_drequest(struct drbd_peer_device *peer_device, int cmd,
1412 sector_t sector, int size, u64 block_id)
1413 {
1414 struct drbd_socket *sock;
1415 struct p_block_req *p;
1416
1417 sock = &peer_device->connection->data;
1418 p = drbd_prepare_command(peer_device, sock);
1419 if (!p)
1420 return -EIO;
1421 p->sector = cpu_to_be64(sector);
1422 p->block_id = block_id;
1423 p->blksize = cpu_to_be32(size);
1424 return drbd_send_command(peer_device, sock, cmd, sizeof(*p), NULL, 0);
1425 }
1426
drbd_send_drequest_csum(struct drbd_peer_device * peer_device,sector_t sector,int size,void * digest,int digest_size,enum drbd_packet cmd)1427 int drbd_send_drequest_csum(struct drbd_peer_device *peer_device, sector_t sector, int size,
1428 void *digest, int digest_size, enum drbd_packet cmd)
1429 {
1430 struct drbd_socket *sock;
1431 struct p_block_req *p;
1432
1433 /* FIXME: Put the digest into the preallocated socket buffer. */
1434
1435 sock = &peer_device->connection->data;
1436 p = drbd_prepare_command(peer_device, sock);
1437 if (!p)
1438 return -EIO;
1439 p->sector = cpu_to_be64(sector);
1440 p->block_id = ID_SYNCER /* unused */;
1441 p->blksize = cpu_to_be32(size);
1442 return drbd_send_command(peer_device, sock, cmd, sizeof(*p), digest, digest_size);
1443 }
1444
drbd_send_ov_request(struct drbd_peer_device * peer_device,sector_t sector,int size)1445 int drbd_send_ov_request(struct drbd_peer_device *peer_device, sector_t sector, int size)
1446 {
1447 struct drbd_socket *sock;
1448 struct p_block_req *p;
1449
1450 sock = &peer_device->connection->data;
1451 p = drbd_prepare_command(peer_device, sock);
1452 if (!p)
1453 return -EIO;
1454 p->sector = cpu_to_be64(sector);
1455 p->block_id = ID_SYNCER /* unused */;
1456 p->blksize = cpu_to_be32(size);
1457 return drbd_send_command(peer_device, sock, P_OV_REQUEST, sizeof(*p), NULL, 0);
1458 }
1459
1460 /* called on sndtimeo
1461 * returns false if we should retry,
1462 * true if we think connection is dead
1463 */
we_should_drop_the_connection(struct drbd_connection * connection,struct socket * sock)1464 static int we_should_drop_the_connection(struct drbd_connection *connection, struct socket *sock)
1465 {
1466 int drop_it;
1467 /* long elapsed = (long)(jiffies - device->last_received); */
1468
1469 drop_it = connection->meta.socket == sock
1470 || !connection->ack_receiver.task
1471 || get_t_state(&connection->ack_receiver) != RUNNING
1472 || connection->cstate < C_WF_REPORT_PARAMS;
1473
1474 if (drop_it)
1475 return true;
1476
1477 drop_it = !--connection->ko_count;
1478 if (!drop_it) {
1479 drbd_err(connection, "[%s/%d] sock_sendmsg time expired, ko = %u\n",
1480 current->comm, current->pid, connection->ko_count);
1481 request_ping(connection);
1482 }
1483
1484 return drop_it; /* && (device->state == R_PRIMARY) */;
1485 }
1486
drbd_update_congested(struct drbd_connection * connection)1487 static void drbd_update_congested(struct drbd_connection *connection)
1488 {
1489 struct sock *sk = connection->data.socket->sk;
1490 if (sk->sk_wmem_queued > sk->sk_sndbuf * 4 / 5)
1491 set_bit(NET_CONGESTED, &connection->flags);
1492 }
1493
1494 /* The idea of sendpage seems to be to put some kind of reference
1495 * to the page into the skb, and to hand it over to the NIC. In
1496 * this process get_page() gets called.
1497 *
1498 * As soon as the page was really sent over the network put_page()
1499 * gets called by some part of the network layer. [ NIC driver? ]
1500 *
1501 * [ get_page() / put_page() increment/decrement the count. If count
1502 * reaches 0 the page will be freed. ]
1503 *
1504 * This works nicely with pages from FSs.
1505 * But this means that in protocol A we might signal IO completion too early!
1506 *
1507 * In order not to corrupt data during a resync we must make sure
1508 * that we do not reuse our own buffer pages (EEs) to early, therefore
1509 * we have the net_ee list.
1510 *
1511 * XFS seems to have problems, still, it submits pages with page_count == 0!
1512 * As a workaround, we disable sendpage on pages
1513 * with page_count == 0 or PageSlab.
1514 */
_drbd_no_send_page(struct drbd_peer_device * peer_device,struct page * page,int offset,size_t size,unsigned msg_flags)1515 static int _drbd_no_send_page(struct drbd_peer_device *peer_device, struct page *page,
1516 int offset, size_t size, unsigned msg_flags)
1517 {
1518 struct socket *socket;
1519 void *addr;
1520 int err;
1521
1522 socket = peer_device->connection->data.socket;
1523 addr = kmap(page) + offset;
1524 err = drbd_send_all(peer_device->connection, socket, addr, size, msg_flags);
1525 kunmap(page);
1526 if (!err)
1527 peer_device->device->send_cnt += size >> 9;
1528 return err;
1529 }
1530
_drbd_send_page(struct drbd_peer_device * peer_device,struct page * page,int offset,size_t size,unsigned msg_flags)1531 static int _drbd_send_page(struct drbd_peer_device *peer_device, struct page *page,
1532 int offset, size_t size, unsigned msg_flags)
1533 {
1534 struct socket *socket = peer_device->connection->data.socket;
1535 int len = size;
1536 int err = -EIO;
1537
1538 /* e.g. XFS meta- & log-data is in slab pages, which have a
1539 * page_count of 0 and/or have PageSlab() set.
1540 * we cannot use send_page for those, as that does get_page();
1541 * put_page(); and would cause either a VM_BUG directly, or
1542 * __page_cache_release a page that would actually still be referenced
1543 * by someone, leading to some obscure delayed Oops somewhere else. */
1544 if (drbd_disable_sendpage || !sendpage_ok(page))
1545 return _drbd_no_send_page(peer_device, page, offset, size, msg_flags);
1546
1547 msg_flags |= MSG_NOSIGNAL;
1548 drbd_update_congested(peer_device->connection);
1549 do {
1550 int sent;
1551
1552 sent = socket->ops->sendpage(socket, page, offset, len, msg_flags);
1553 if (sent <= 0) {
1554 if (sent == -EAGAIN) {
1555 if (we_should_drop_the_connection(peer_device->connection, socket))
1556 break;
1557 continue;
1558 }
1559 drbd_warn(peer_device->device, "%s: size=%d len=%d sent=%d\n",
1560 __func__, (int)size, len, sent);
1561 if (sent < 0)
1562 err = sent;
1563 break;
1564 }
1565 len -= sent;
1566 offset += sent;
1567 } while (len > 0 /* THINK && device->cstate >= C_CONNECTED*/);
1568 clear_bit(NET_CONGESTED, &peer_device->connection->flags);
1569
1570 if (len == 0) {
1571 err = 0;
1572 peer_device->device->send_cnt += size >> 9;
1573 }
1574 return err;
1575 }
1576
_drbd_send_bio(struct drbd_peer_device * peer_device,struct bio * bio)1577 static int _drbd_send_bio(struct drbd_peer_device *peer_device, struct bio *bio)
1578 {
1579 struct bio_vec bvec;
1580 struct bvec_iter iter;
1581
1582 /* hint all but last page with MSG_MORE */
1583 bio_for_each_segment(bvec, bio, iter) {
1584 int err;
1585
1586 err = _drbd_no_send_page(peer_device, bvec.bv_page,
1587 bvec.bv_offset, bvec.bv_len,
1588 bio_iter_last(bvec, iter)
1589 ? 0 : MSG_MORE);
1590 if (err)
1591 return err;
1592 }
1593 return 0;
1594 }
1595
_drbd_send_zc_bio(struct drbd_peer_device * peer_device,struct bio * bio)1596 static int _drbd_send_zc_bio(struct drbd_peer_device *peer_device, struct bio *bio)
1597 {
1598 struct bio_vec bvec;
1599 struct bvec_iter iter;
1600
1601 /* hint all but last page with MSG_MORE */
1602 bio_for_each_segment(bvec, bio, iter) {
1603 int err;
1604
1605 err = _drbd_send_page(peer_device, bvec.bv_page,
1606 bvec.bv_offset, bvec.bv_len,
1607 bio_iter_last(bvec, iter) ? 0 : MSG_MORE);
1608 if (err)
1609 return err;
1610 }
1611 return 0;
1612 }
1613
_drbd_send_zc_ee(struct drbd_peer_device * peer_device,struct drbd_peer_request * peer_req)1614 static int _drbd_send_zc_ee(struct drbd_peer_device *peer_device,
1615 struct drbd_peer_request *peer_req)
1616 {
1617 struct page *page = peer_req->pages;
1618 unsigned len = peer_req->i.size;
1619 int err;
1620
1621 /* hint all but last page with MSG_MORE */
1622 page_chain_for_each(page) {
1623 unsigned l = min_t(unsigned, len, PAGE_SIZE);
1624
1625 err = _drbd_send_page(peer_device, page, 0, l,
1626 page_chain_next(page) ? MSG_MORE : 0);
1627 if (err)
1628 return err;
1629 len -= l;
1630 }
1631 return 0;
1632 }
1633
bio_flags_to_wire(struct drbd_connection * connection,struct bio * bio)1634 static u32 bio_flags_to_wire(struct drbd_connection *connection,
1635 struct bio *bio)
1636 {
1637 if (connection->agreed_pro_version >= 95)
1638 return (bio->bi_opf & REQ_SYNC ? DP_RW_SYNC : 0) |
1639 (bio->bi_opf & REQ_FUA ? DP_FUA : 0) |
1640 (bio->bi_opf & REQ_PREFLUSH ? DP_FLUSH : 0) |
1641 (bio_op(bio) == REQ_OP_DISCARD ? DP_DISCARD : 0) |
1642 (bio_op(bio) == REQ_OP_WRITE_ZEROES ?
1643 ((connection->agreed_features & DRBD_FF_WZEROES) ?
1644 (DP_ZEROES |(!(bio->bi_opf & REQ_NOUNMAP) ? DP_DISCARD : 0))
1645 : DP_DISCARD)
1646 : 0);
1647 else
1648 return bio->bi_opf & REQ_SYNC ? DP_RW_SYNC : 0;
1649 }
1650
1651 /* Used to send write or TRIM aka REQ_OP_DISCARD requests
1652 * R_PRIMARY -> Peer (P_DATA, P_TRIM)
1653 */
drbd_send_dblock(struct drbd_peer_device * peer_device,struct drbd_request * req)1654 int drbd_send_dblock(struct drbd_peer_device *peer_device, struct drbd_request *req)
1655 {
1656 struct drbd_device *device = peer_device->device;
1657 struct drbd_socket *sock;
1658 struct p_data *p;
1659 void *digest_out;
1660 unsigned int dp_flags = 0;
1661 int digest_size;
1662 int err;
1663
1664 sock = &peer_device->connection->data;
1665 p = drbd_prepare_command(peer_device, sock);
1666 digest_size = peer_device->connection->integrity_tfm ?
1667 crypto_shash_digestsize(peer_device->connection->integrity_tfm) : 0;
1668
1669 if (!p)
1670 return -EIO;
1671 p->sector = cpu_to_be64(req->i.sector);
1672 p->block_id = (unsigned long)req;
1673 p->seq_num = cpu_to_be32(atomic_inc_return(&device->packet_seq));
1674 dp_flags = bio_flags_to_wire(peer_device->connection, req->master_bio);
1675 if (device->state.conn >= C_SYNC_SOURCE &&
1676 device->state.conn <= C_PAUSED_SYNC_T)
1677 dp_flags |= DP_MAY_SET_IN_SYNC;
1678 if (peer_device->connection->agreed_pro_version >= 100) {
1679 if (req->rq_state & RQ_EXP_RECEIVE_ACK)
1680 dp_flags |= DP_SEND_RECEIVE_ACK;
1681 /* During resync, request an explicit write ack,
1682 * even in protocol != C */
1683 if (req->rq_state & RQ_EXP_WRITE_ACK
1684 || (dp_flags & DP_MAY_SET_IN_SYNC))
1685 dp_flags |= DP_SEND_WRITE_ACK;
1686 }
1687 p->dp_flags = cpu_to_be32(dp_flags);
1688
1689 if (dp_flags & (DP_DISCARD|DP_ZEROES)) {
1690 enum drbd_packet cmd = (dp_flags & DP_ZEROES) ? P_ZEROES : P_TRIM;
1691 struct p_trim *t = (struct p_trim*)p;
1692 t->size = cpu_to_be32(req->i.size);
1693 err = __send_command(peer_device->connection, device->vnr, sock, cmd, sizeof(*t), NULL, 0);
1694 goto out;
1695 }
1696 digest_out = p + 1;
1697
1698 /* our digest is still only over the payload.
1699 * TRIM does not carry any payload. */
1700 if (digest_size)
1701 drbd_csum_bio(peer_device->connection->integrity_tfm, req->master_bio, digest_out);
1702 err = __send_command(peer_device->connection, device->vnr, sock, P_DATA,
1703 sizeof(*p) + digest_size, NULL, req->i.size);
1704 if (!err) {
1705 /* For protocol A, we have to memcpy the payload into
1706 * socket buffers, as we may complete right away
1707 * as soon as we handed it over to tcp, at which point the data
1708 * pages may become invalid.
1709 *
1710 * For data-integrity enabled, we copy it as well, so we can be
1711 * sure that even if the bio pages may still be modified, it
1712 * won't change the data on the wire, thus if the digest checks
1713 * out ok after sending on this side, but does not fit on the
1714 * receiving side, we sure have detected corruption elsewhere.
1715 */
1716 if (!(req->rq_state & (RQ_EXP_RECEIVE_ACK | RQ_EXP_WRITE_ACK)) || digest_size)
1717 err = _drbd_send_bio(peer_device, req->master_bio);
1718 else
1719 err = _drbd_send_zc_bio(peer_device, req->master_bio);
1720
1721 /* double check digest, sometimes buffers have been modified in flight. */
1722 if (digest_size > 0 && digest_size <= 64) {
1723 /* 64 byte, 512 bit, is the largest digest size
1724 * currently supported in kernel crypto. */
1725 unsigned char digest[64];
1726 drbd_csum_bio(peer_device->connection->integrity_tfm, req->master_bio, digest);
1727 if (memcmp(p + 1, digest, digest_size)) {
1728 drbd_warn(device,
1729 "Digest mismatch, buffer modified by upper layers during write: %llus +%u\n",
1730 (unsigned long long)req->i.sector, req->i.size);
1731 }
1732 } /* else if (digest_size > 64) {
1733 ... Be noisy about digest too large ...
1734 } */
1735 }
1736 out:
1737 mutex_unlock(&sock->mutex); /* locked by drbd_prepare_command() */
1738
1739 return err;
1740 }
1741
1742 /* answer packet, used to send data back for read requests:
1743 * Peer -> (diskless) R_PRIMARY (P_DATA_REPLY)
1744 * C_SYNC_SOURCE -> C_SYNC_TARGET (P_RS_DATA_REPLY)
1745 */
drbd_send_block(struct drbd_peer_device * peer_device,enum drbd_packet cmd,struct drbd_peer_request * peer_req)1746 int drbd_send_block(struct drbd_peer_device *peer_device, enum drbd_packet cmd,
1747 struct drbd_peer_request *peer_req)
1748 {
1749 struct drbd_device *device = peer_device->device;
1750 struct drbd_socket *sock;
1751 struct p_data *p;
1752 int err;
1753 int digest_size;
1754
1755 sock = &peer_device->connection->data;
1756 p = drbd_prepare_command(peer_device, sock);
1757
1758 digest_size = peer_device->connection->integrity_tfm ?
1759 crypto_shash_digestsize(peer_device->connection->integrity_tfm) : 0;
1760
1761 if (!p)
1762 return -EIO;
1763 p->sector = cpu_to_be64(peer_req->i.sector);
1764 p->block_id = peer_req->block_id;
1765 p->seq_num = 0; /* unused */
1766 p->dp_flags = 0;
1767 if (digest_size)
1768 drbd_csum_ee(peer_device->connection->integrity_tfm, peer_req, p + 1);
1769 err = __send_command(peer_device->connection, device->vnr, sock, cmd, sizeof(*p) + digest_size, NULL, peer_req->i.size);
1770 if (!err)
1771 err = _drbd_send_zc_ee(peer_device, peer_req);
1772 mutex_unlock(&sock->mutex); /* locked by drbd_prepare_command() */
1773
1774 return err;
1775 }
1776
drbd_send_out_of_sync(struct drbd_peer_device * peer_device,struct drbd_request * req)1777 int drbd_send_out_of_sync(struct drbd_peer_device *peer_device, struct drbd_request *req)
1778 {
1779 struct drbd_socket *sock;
1780 struct p_block_desc *p;
1781
1782 sock = &peer_device->connection->data;
1783 p = drbd_prepare_command(peer_device, sock);
1784 if (!p)
1785 return -EIO;
1786 p->sector = cpu_to_be64(req->i.sector);
1787 p->blksize = cpu_to_be32(req->i.size);
1788 return drbd_send_command(peer_device, sock, P_OUT_OF_SYNC, sizeof(*p), NULL, 0);
1789 }
1790
1791 /*
1792 drbd_send distinguishes two cases:
1793
1794 Packets sent via the data socket "sock"
1795 and packets sent via the meta data socket "msock"
1796
1797 sock msock
1798 -----------------+-------------------------+------------------------------
1799 timeout conf.timeout / 2 conf.timeout / 2
1800 timeout action send a ping via msock Abort communication
1801 and close all sockets
1802 */
1803
1804 /*
1805 * you must have down()ed the appropriate [m]sock_mutex elsewhere!
1806 */
drbd_send(struct drbd_connection * connection,struct socket * sock,void * buf,size_t size,unsigned msg_flags)1807 int drbd_send(struct drbd_connection *connection, struct socket *sock,
1808 void *buf, size_t size, unsigned msg_flags)
1809 {
1810 struct kvec iov = {.iov_base = buf, .iov_len = size};
1811 struct msghdr msg = {.msg_flags = msg_flags | MSG_NOSIGNAL};
1812 int rv, sent = 0;
1813
1814 if (!sock)
1815 return -EBADR;
1816
1817 /* THINK if (signal_pending) return ... ? */
1818
1819 iov_iter_kvec(&msg.msg_iter, ITER_SOURCE, &iov, 1, size);
1820
1821 if (sock == connection->data.socket) {
1822 rcu_read_lock();
1823 connection->ko_count = rcu_dereference(connection->net_conf)->ko_count;
1824 rcu_read_unlock();
1825 drbd_update_congested(connection);
1826 }
1827 do {
1828 rv = sock_sendmsg(sock, &msg);
1829 if (rv == -EAGAIN) {
1830 if (we_should_drop_the_connection(connection, sock))
1831 break;
1832 else
1833 continue;
1834 }
1835 if (rv == -EINTR) {
1836 flush_signals(current);
1837 rv = 0;
1838 }
1839 if (rv < 0)
1840 break;
1841 sent += rv;
1842 } while (sent < size);
1843
1844 if (sock == connection->data.socket)
1845 clear_bit(NET_CONGESTED, &connection->flags);
1846
1847 if (rv <= 0) {
1848 if (rv != -EAGAIN) {
1849 drbd_err(connection, "%s_sendmsg returned %d\n",
1850 sock == connection->meta.socket ? "msock" : "sock",
1851 rv);
1852 conn_request_state(connection, NS(conn, C_BROKEN_PIPE), CS_HARD);
1853 } else
1854 conn_request_state(connection, NS(conn, C_TIMEOUT), CS_HARD);
1855 }
1856
1857 return sent;
1858 }
1859
1860 /*
1861 * drbd_send_all - Send an entire buffer
1862 *
1863 * Returns 0 upon success and a negative error value otherwise.
1864 */
drbd_send_all(struct drbd_connection * connection,struct socket * sock,void * buffer,size_t size,unsigned msg_flags)1865 int drbd_send_all(struct drbd_connection *connection, struct socket *sock, void *buffer,
1866 size_t size, unsigned msg_flags)
1867 {
1868 int err;
1869
1870 err = drbd_send(connection, sock, buffer, size, msg_flags);
1871 if (err < 0)
1872 return err;
1873 if (err != size)
1874 return -EIO;
1875 return 0;
1876 }
1877
drbd_open(struct block_device * bdev,fmode_t mode)1878 static int drbd_open(struct block_device *bdev, fmode_t mode)
1879 {
1880 struct drbd_device *device = bdev->bd_disk->private_data;
1881 unsigned long flags;
1882 int rv = 0;
1883
1884 mutex_lock(&drbd_main_mutex);
1885 spin_lock_irqsave(&device->resource->req_lock, flags);
1886 /* to have a stable device->state.role
1887 * and no race with updating open_cnt */
1888
1889 if (device->state.role != R_PRIMARY) {
1890 if (mode & FMODE_WRITE)
1891 rv = -EROFS;
1892 else if (!drbd_allow_oos)
1893 rv = -EMEDIUMTYPE;
1894 }
1895
1896 if (!rv)
1897 device->open_cnt++;
1898 spin_unlock_irqrestore(&device->resource->req_lock, flags);
1899 mutex_unlock(&drbd_main_mutex);
1900
1901 return rv;
1902 }
1903
drbd_release(struct gendisk * gd,fmode_t mode)1904 static void drbd_release(struct gendisk *gd, fmode_t mode)
1905 {
1906 struct drbd_device *device = gd->private_data;
1907 mutex_lock(&drbd_main_mutex);
1908 device->open_cnt--;
1909 mutex_unlock(&drbd_main_mutex);
1910 }
1911
1912 /* need to hold resource->req_lock */
drbd_queue_unplug(struct drbd_device * device)1913 void drbd_queue_unplug(struct drbd_device *device)
1914 {
1915 if (device->state.pdsk >= D_INCONSISTENT && device->state.conn >= C_CONNECTED) {
1916 D_ASSERT(device, device->state.role == R_PRIMARY);
1917 if (test_and_clear_bit(UNPLUG_REMOTE, &device->flags)) {
1918 drbd_queue_work_if_unqueued(
1919 &first_peer_device(device)->connection->sender_work,
1920 &device->unplug_work);
1921 }
1922 }
1923 }
1924
drbd_set_defaults(struct drbd_device * device)1925 static void drbd_set_defaults(struct drbd_device *device)
1926 {
1927 /* Beware! The actual layout differs
1928 * between big endian and little endian */
1929 device->state = (union drbd_dev_state) {
1930 { .role = R_SECONDARY,
1931 .peer = R_UNKNOWN,
1932 .conn = C_STANDALONE,
1933 .disk = D_DISKLESS,
1934 .pdsk = D_UNKNOWN,
1935 } };
1936 }
1937
drbd_init_set_defaults(struct drbd_device * device)1938 void drbd_init_set_defaults(struct drbd_device *device)
1939 {
1940 /* the memset(,0,) did most of this.
1941 * note: only assignments, no allocation in here */
1942
1943 drbd_set_defaults(device);
1944
1945 atomic_set(&device->ap_bio_cnt, 0);
1946 atomic_set(&device->ap_actlog_cnt, 0);
1947 atomic_set(&device->ap_pending_cnt, 0);
1948 atomic_set(&device->rs_pending_cnt, 0);
1949 atomic_set(&device->unacked_cnt, 0);
1950 atomic_set(&device->local_cnt, 0);
1951 atomic_set(&device->pp_in_use_by_net, 0);
1952 atomic_set(&device->rs_sect_in, 0);
1953 atomic_set(&device->rs_sect_ev, 0);
1954 atomic_set(&device->ap_in_flight, 0);
1955 atomic_set(&device->md_io.in_use, 0);
1956
1957 mutex_init(&device->own_state_mutex);
1958 device->state_mutex = &device->own_state_mutex;
1959
1960 spin_lock_init(&device->al_lock);
1961 spin_lock_init(&device->peer_seq_lock);
1962
1963 INIT_LIST_HEAD(&device->active_ee);
1964 INIT_LIST_HEAD(&device->sync_ee);
1965 INIT_LIST_HEAD(&device->done_ee);
1966 INIT_LIST_HEAD(&device->read_ee);
1967 INIT_LIST_HEAD(&device->net_ee);
1968 INIT_LIST_HEAD(&device->resync_reads);
1969 INIT_LIST_HEAD(&device->resync_work.list);
1970 INIT_LIST_HEAD(&device->unplug_work.list);
1971 INIT_LIST_HEAD(&device->bm_io_work.w.list);
1972 INIT_LIST_HEAD(&device->pending_master_completion[0]);
1973 INIT_LIST_HEAD(&device->pending_master_completion[1]);
1974 INIT_LIST_HEAD(&device->pending_completion[0]);
1975 INIT_LIST_HEAD(&device->pending_completion[1]);
1976
1977 device->resync_work.cb = w_resync_timer;
1978 device->unplug_work.cb = w_send_write_hint;
1979 device->bm_io_work.w.cb = w_bitmap_io;
1980
1981 timer_setup(&device->resync_timer, resync_timer_fn, 0);
1982 timer_setup(&device->md_sync_timer, md_sync_timer_fn, 0);
1983 timer_setup(&device->start_resync_timer, start_resync_timer_fn, 0);
1984 timer_setup(&device->request_timer, request_timer_fn, 0);
1985
1986 init_waitqueue_head(&device->misc_wait);
1987 init_waitqueue_head(&device->state_wait);
1988 init_waitqueue_head(&device->ee_wait);
1989 init_waitqueue_head(&device->al_wait);
1990 init_waitqueue_head(&device->seq_wait);
1991
1992 device->resync_wenr = LC_FREE;
1993 device->peer_max_bio_size = DRBD_MAX_BIO_SIZE_SAFE;
1994 device->local_max_bio_size = DRBD_MAX_BIO_SIZE_SAFE;
1995 }
1996
drbd_set_my_capacity(struct drbd_device * device,sector_t size)1997 void drbd_set_my_capacity(struct drbd_device *device, sector_t size)
1998 {
1999 char ppb[10];
2000
2001 set_capacity_and_notify(device->vdisk, size);
2002
2003 drbd_info(device, "size = %s (%llu KB)\n",
2004 ppsize(ppb, size>>1), (unsigned long long)size>>1);
2005 }
2006
drbd_device_cleanup(struct drbd_device * device)2007 void drbd_device_cleanup(struct drbd_device *device)
2008 {
2009 int i;
2010 if (first_peer_device(device)->connection->receiver.t_state != NONE)
2011 drbd_err(device, "ASSERT FAILED: receiver t_state == %d expected 0.\n",
2012 first_peer_device(device)->connection->receiver.t_state);
2013
2014 device->al_writ_cnt =
2015 device->bm_writ_cnt =
2016 device->read_cnt =
2017 device->recv_cnt =
2018 device->send_cnt =
2019 device->writ_cnt =
2020 device->p_size =
2021 device->rs_start =
2022 device->rs_total =
2023 device->rs_failed = 0;
2024 device->rs_last_events = 0;
2025 device->rs_last_sect_ev = 0;
2026 for (i = 0; i < DRBD_SYNC_MARKS; i++) {
2027 device->rs_mark_left[i] = 0;
2028 device->rs_mark_time[i] = 0;
2029 }
2030 D_ASSERT(device, first_peer_device(device)->connection->net_conf == NULL);
2031
2032 set_capacity_and_notify(device->vdisk, 0);
2033 if (device->bitmap) {
2034 /* maybe never allocated. */
2035 drbd_bm_resize(device, 0, 1);
2036 drbd_bm_cleanup(device);
2037 }
2038
2039 drbd_backing_dev_free(device, device->ldev);
2040 device->ldev = NULL;
2041
2042 clear_bit(AL_SUSPENDED, &device->flags);
2043
2044 D_ASSERT(device, list_empty(&device->active_ee));
2045 D_ASSERT(device, list_empty(&device->sync_ee));
2046 D_ASSERT(device, list_empty(&device->done_ee));
2047 D_ASSERT(device, list_empty(&device->read_ee));
2048 D_ASSERT(device, list_empty(&device->net_ee));
2049 D_ASSERT(device, list_empty(&device->resync_reads));
2050 D_ASSERT(device, list_empty(&first_peer_device(device)->connection->sender_work.q));
2051 D_ASSERT(device, list_empty(&device->resync_work.list));
2052 D_ASSERT(device, list_empty(&device->unplug_work.list));
2053
2054 drbd_set_defaults(device);
2055 }
2056
2057
drbd_destroy_mempools(void)2058 static void drbd_destroy_mempools(void)
2059 {
2060 struct page *page;
2061
2062 while (drbd_pp_pool) {
2063 page = drbd_pp_pool;
2064 drbd_pp_pool = (struct page *)page_private(page);
2065 __free_page(page);
2066 drbd_pp_vacant--;
2067 }
2068
2069 /* D_ASSERT(device, atomic_read(&drbd_pp_vacant)==0); */
2070
2071 bioset_exit(&drbd_io_bio_set);
2072 bioset_exit(&drbd_md_io_bio_set);
2073 mempool_exit(&drbd_md_io_page_pool);
2074 mempool_exit(&drbd_ee_mempool);
2075 mempool_exit(&drbd_request_mempool);
2076 kmem_cache_destroy(drbd_ee_cache);
2077 kmem_cache_destroy(drbd_request_cache);
2078 kmem_cache_destroy(drbd_bm_ext_cache);
2079 kmem_cache_destroy(drbd_al_ext_cache);
2080
2081 drbd_ee_cache = NULL;
2082 drbd_request_cache = NULL;
2083 drbd_bm_ext_cache = NULL;
2084 drbd_al_ext_cache = NULL;
2085
2086 return;
2087 }
2088
drbd_create_mempools(void)2089 static int drbd_create_mempools(void)
2090 {
2091 struct page *page;
2092 const int number = (DRBD_MAX_BIO_SIZE/PAGE_SIZE) * drbd_minor_count;
2093 int i, ret;
2094
2095 /* caches */
2096 drbd_request_cache = kmem_cache_create(
2097 "drbd_req", sizeof(struct drbd_request), 0, 0, NULL);
2098 if (drbd_request_cache == NULL)
2099 goto Enomem;
2100
2101 drbd_ee_cache = kmem_cache_create(
2102 "drbd_ee", sizeof(struct drbd_peer_request), 0, 0, NULL);
2103 if (drbd_ee_cache == NULL)
2104 goto Enomem;
2105
2106 drbd_bm_ext_cache = kmem_cache_create(
2107 "drbd_bm", sizeof(struct bm_extent), 0, 0, NULL);
2108 if (drbd_bm_ext_cache == NULL)
2109 goto Enomem;
2110
2111 drbd_al_ext_cache = kmem_cache_create(
2112 "drbd_al", sizeof(struct lc_element), 0, 0, NULL);
2113 if (drbd_al_ext_cache == NULL)
2114 goto Enomem;
2115
2116 /* mempools */
2117 ret = bioset_init(&drbd_io_bio_set, BIO_POOL_SIZE, 0, 0);
2118 if (ret)
2119 goto Enomem;
2120
2121 ret = bioset_init(&drbd_md_io_bio_set, DRBD_MIN_POOL_PAGES, 0,
2122 BIOSET_NEED_BVECS);
2123 if (ret)
2124 goto Enomem;
2125
2126 ret = mempool_init_page_pool(&drbd_md_io_page_pool, DRBD_MIN_POOL_PAGES, 0);
2127 if (ret)
2128 goto Enomem;
2129
2130 ret = mempool_init_slab_pool(&drbd_request_mempool, number,
2131 drbd_request_cache);
2132 if (ret)
2133 goto Enomem;
2134
2135 ret = mempool_init_slab_pool(&drbd_ee_mempool, number, drbd_ee_cache);
2136 if (ret)
2137 goto Enomem;
2138
2139 for (i = 0; i < number; i++) {
2140 page = alloc_page(GFP_HIGHUSER);
2141 if (!page)
2142 goto Enomem;
2143 set_page_private(page, (unsigned long)drbd_pp_pool);
2144 drbd_pp_pool = page;
2145 }
2146 drbd_pp_vacant = number;
2147
2148 return 0;
2149
2150 Enomem:
2151 drbd_destroy_mempools(); /* in case we allocated some */
2152 return -ENOMEM;
2153 }
2154
drbd_release_all_peer_reqs(struct drbd_device * device)2155 static void drbd_release_all_peer_reqs(struct drbd_device *device)
2156 {
2157 int rr;
2158
2159 rr = drbd_free_peer_reqs(device, &device->active_ee);
2160 if (rr)
2161 drbd_err(device, "%d EEs in active list found!\n", rr);
2162
2163 rr = drbd_free_peer_reqs(device, &device->sync_ee);
2164 if (rr)
2165 drbd_err(device, "%d EEs in sync list found!\n", rr);
2166
2167 rr = drbd_free_peer_reqs(device, &device->read_ee);
2168 if (rr)
2169 drbd_err(device, "%d EEs in read list found!\n", rr);
2170
2171 rr = drbd_free_peer_reqs(device, &device->done_ee);
2172 if (rr)
2173 drbd_err(device, "%d EEs in done list found!\n", rr);
2174
2175 rr = drbd_free_peer_reqs(device, &device->net_ee);
2176 if (rr)
2177 drbd_err(device, "%d EEs in net list found!\n", rr);
2178 }
2179
2180 /* caution. no locking. */
drbd_destroy_device(struct kref * kref)2181 void drbd_destroy_device(struct kref *kref)
2182 {
2183 struct drbd_device *device = container_of(kref, struct drbd_device, kref);
2184 struct drbd_resource *resource = device->resource;
2185 struct drbd_peer_device *peer_device, *tmp_peer_device;
2186
2187 del_timer_sync(&device->request_timer);
2188
2189 /* paranoia asserts */
2190 D_ASSERT(device, device->open_cnt == 0);
2191 /* end paranoia asserts */
2192
2193 /* cleanup stuff that may have been allocated during
2194 * device (re-)configuration or state changes */
2195
2196 drbd_backing_dev_free(device, device->ldev);
2197 device->ldev = NULL;
2198
2199 drbd_release_all_peer_reqs(device);
2200
2201 lc_destroy(device->act_log);
2202 lc_destroy(device->resync);
2203
2204 kfree(device->p_uuid);
2205 /* device->p_uuid = NULL; */
2206
2207 if (device->bitmap) /* should no longer be there. */
2208 drbd_bm_cleanup(device);
2209 __free_page(device->md_io.page);
2210 put_disk(device->vdisk);
2211 kfree(device->rs_plan_s);
2212
2213 /* not for_each_connection(connection, resource):
2214 * those may have been cleaned up and disassociated already.
2215 */
2216 for_each_peer_device_safe(peer_device, tmp_peer_device, device) {
2217 kref_put(&peer_device->connection->kref, drbd_destroy_connection);
2218 kfree(peer_device);
2219 }
2220 if (device->submit.wq)
2221 destroy_workqueue(device->submit.wq);
2222 kfree(device);
2223 kref_put(&resource->kref, drbd_destroy_resource);
2224 }
2225
2226 /* One global retry thread, if we need to push back some bio and have it
2227 * reinserted through our make request function.
2228 */
2229 static struct retry_worker {
2230 struct workqueue_struct *wq;
2231 struct work_struct worker;
2232
2233 spinlock_t lock;
2234 struct list_head writes;
2235 } retry;
2236
do_retry(struct work_struct * ws)2237 static void do_retry(struct work_struct *ws)
2238 {
2239 struct retry_worker *retry = container_of(ws, struct retry_worker, worker);
2240 LIST_HEAD(writes);
2241 struct drbd_request *req, *tmp;
2242
2243 spin_lock_irq(&retry->lock);
2244 list_splice_init(&retry->writes, &writes);
2245 spin_unlock_irq(&retry->lock);
2246
2247 list_for_each_entry_safe(req, tmp, &writes, tl_requests) {
2248 struct drbd_device *device = req->device;
2249 struct bio *bio = req->master_bio;
2250 bool expected;
2251
2252 expected =
2253 expect(atomic_read(&req->completion_ref) == 0) &&
2254 expect(req->rq_state & RQ_POSTPONED) &&
2255 expect((req->rq_state & RQ_LOCAL_PENDING) == 0 ||
2256 (req->rq_state & RQ_LOCAL_ABORTED) != 0);
2257
2258 if (!expected)
2259 drbd_err(device, "req=%p completion_ref=%d rq_state=%x\n",
2260 req, atomic_read(&req->completion_ref),
2261 req->rq_state);
2262
2263 /* We still need to put one kref associated with the
2264 * "completion_ref" going zero in the code path that queued it
2265 * here. The request object may still be referenced by a
2266 * frozen local req->private_bio, in case we force-detached.
2267 */
2268 kref_put(&req->kref, drbd_req_destroy);
2269
2270 /* A single suspended or otherwise blocking device may stall
2271 * all others as well. Fortunately, this code path is to
2272 * recover from a situation that "should not happen":
2273 * concurrent writes in multi-primary setup.
2274 * In a "normal" lifecycle, this workqueue is supposed to be
2275 * destroyed without ever doing anything.
2276 * If it turns out to be an issue anyways, we can do per
2277 * resource (replication group) or per device (minor) retry
2278 * workqueues instead.
2279 */
2280
2281 /* We are not just doing submit_bio_noacct(),
2282 * as we want to keep the start_time information. */
2283 inc_ap_bio(device);
2284 __drbd_make_request(device, bio);
2285 }
2286 }
2287
2288 /* called via drbd_req_put_completion_ref(),
2289 * holds resource->req_lock */
drbd_restart_request(struct drbd_request * req)2290 void drbd_restart_request(struct drbd_request *req)
2291 {
2292 unsigned long flags;
2293 spin_lock_irqsave(&retry.lock, flags);
2294 list_move_tail(&req->tl_requests, &retry.writes);
2295 spin_unlock_irqrestore(&retry.lock, flags);
2296
2297 /* Drop the extra reference that would otherwise
2298 * have been dropped by complete_master_bio.
2299 * do_retry() needs to grab a new one. */
2300 dec_ap_bio(req->device);
2301
2302 queue_work(retry.wq, &retry.worker);
2303 }
2304
drbd_destroy_resource(struct kref * kref)2305 void drbd_destroy_resource(struct kref *kref)
2306 {
2307 struct drbd_resource *resource =
2308 container_of(kref, struct drbd_resource, kref);
2309
2310 idr_destroy(&resource->devices);
2311 free_cpumask_var(resource->cpu_mask);
2312 kfree(resource->name);
2313 kfree(resource);
2314 }
2315
drbd_free_resource(struct drbd_resource * resource)2316 void drbd_free_resource(struct drbd_resource *resource)
2317 {
2318 struct drbd_connection *connection, *tmp;
2319
2320 for_each_connection_safe(connection, tmp, resource) {
2321 list_del(&connection->connections);
2322 drbd_debugfs_connection_cleanup(connection);
2323 kref_put(&connection->kref, drbd_destroy_connection);
2324 }
2325 drbd_debugfs_resource_cleanup(resource);
2326 kref_put(&resource->kref, drbd_destroy_resource);
2327 }
2328
drbd_cleanup(void)2329 static void drbd_cleanup(void)
2330 {
2331 unsigned int i;
2332 struct drbd_device *device;
2333 struct drbd_resource *resource, *tmp;
2334
2335 /* first remove proc,
2336 * drbdsetup uses it's presence to detect
2337 * whether DRBD is loaded.
2338 * If we would get stuck in proc removal,
2339 * but have netlink already deregistered,
2340 * some drbdsetup commands may wait forever
2341 * for an answer.
2342 */
2343 if (drbd_proc)
2344 remove_proc_entry("drbd", NULL);
2345
2346 if (retry.wq)
2347 destroy_workqueue(retry.wq);
2348
2349 drbd_genl_unregister();
2350
2351 idr_for_each_entry(&drbd_devices, device, i)
2352 drbd_delete_device(device);
2353
2354 /* not _rcu since, no other updater anymore. Genl already unregistered */
2355 for_each_resource_safe(resource, tmp, &drbd_resources) {
2356 list_del(&resource->resources);
2357 drbd_free_resource(resource);
2358 }
2359
2360 drbd_debugfs_cleanup();
2361
2362 drbd_destroy_mempools();
2363 unregister_blkdev(DRBD_MAJOR, "drbd");
2364
2365 idr_destroy(&drbd_devices);
2366
2367 pr_info("module cleanup done.\n");
2368 }
2369
drbd_init_workqueue(struct drbd_work_queue * wq)2370 static void drbd_init_workqueue(struct drbd_work_queue* wq)
2371 {
2372 spin_lock_init(&wq->q_lock);
2373 INIT_LIST_HEAD(&wq->q);
2374 init_waitqueue_head(&wq->q_wait);
2375 }
2376
2377 struct completion_work {
2378 struct drbd_work w;
2379 struct completion done;
2380 };
2381
w_complete(struct drbd_work * w,int cancel)2382 static int w_complete(struct drbd_work *w, int cancel)
2383 {
2384 struct completion_work *completion_work =
2385 container_of(w, struct completion_work, w);
2386
2387 complete(&completion_work->done);
2388 return 0;
2389 }
2390
drbd_flush_workqueue(struct drbd_work_queue * work_queue)2391 void drbd_flush_workqueue(struct drbd_work_queue *work_queue)
2392 {
2393 struct completion_work completion_work;
2394
2395 completion_work.w.cb = w_complete;
2396 init_completion(&completion_work.done);
2397 drbd_queue_work(work_queue, &completion_work.w);
2398 wait_for_completion(&completion_work.done);
2399 }
2400
drbd_find_resource(const char * name)2401 struct drbd_resource *drbd_find_resource(const char *name)
2402 {
2403 struct drbd_resource *resource;
2404
2405 if (!name || !name[0])
2406 return NULL;
2407
2408 rcu_read_lock();
2409 for_each_resource_rcu(resource, &drbd_resources) {
2410 if (!strcmp(resource->name, name)) {
2411 kref_get(&resource->kref);
2412 goto found;
2413 }
2414 }
2415 resource = NULL;
2416 found:
2417 rcu_read_unlock();
2418 return resource;
2419 }
2420
conn_get_by_addrs(void * my_addr,int my_addr_len,void * peer_addr,int peer_addr_len)2421 struct drbd_connection *conn_get_by_addrs(void *my_addr, int my_addr_len,
2422 void *peer_addr, int peer_addr_len)
2423 {
2424 struct drbd_resource *resource;
2425 struct drbd_connection *connection;
2426
2427 rcu_read_lock();
2428 for_each_resource_rcu(resource, &drbd_resources) {
2429 for_each_connection_rcu(connection, resource) {
2430 if (connection->my_addr_len == my_addr_len &&
2431 connection->peer_addr_len == peer_addr_len &&
2432 !memcmp(&connection->my_addr, my_addr, my_addr_len) &&
2433 !memcmp(&connection->peer_addr, peer_addr, peer_addr_len)) {
2434 kref_get(&connection->kref);
2435 goto found;
2436 }
2437 }
2438 }
2439 connection = NULL;
2440 found:
2441 rcu_read_unlock();
2442 return connection;
2443 }
2444
drbd_alloc_socket(struct drbd_socket * socket)2445 static int drbd_alloc_socket(struct drbd_socket *socket)
2446 {
2447 socket->rbuf = (void *) __get_free_page(GFP_KERNEL);
2448 if (!socket->rbuf)
2449 return -ENOMEM;
2450 socket->sbuf = (void *) __get_free_page(GFP_KERNEL);
2451 if (!socket->sbuf)
2452 return -ENOMEM;
2453 return 0;
2454 }
2455
drbd_free_socket(struct drbd_socket * socket)2456 static void drbd_free_socket(struct drbd_socket *socket)
2457 {
2458 free_page((unsigned long) socket->sbuf);
2459 free_page((unsigned long) socket->rbuf);
2460 }
2461
conn_free_crypto(struct drbd_connection * connection)2462 void conn_free_crypto(struct drbd_connection *connection)
2463 {
2464 drbd_free_sock(connection);
2465
2466 crypto_free_shash(connection->csums_tfm);
2467 crypto_free_shash(connection->verify_tfm);
2468 crypto_free_shash(connection->cram_hmac_tfm);
2469 crypto_free_shash(connection->integrity_tfm);
2470 crypto_free_shash(connection->peer_integrity_tfm);
2471 kfree(connection->int_dig_in);
2472 kfree(connection->int_dig_vv);
2473
2474 connection->csums_tfm = NULL;
2475 connection->verify_tfm = NULL;
2476 connection->cram_hmac_tfm = NULL;
2477 connection->integrity_tfm = NULL;
2478 connection->peer_integrity_tfm = NULL;
2479 connection->int_dig_in = NULL;
2480 connection->int_dig_vv = NULL;
2481 }
2482
set_resource_options(struct drbd_resource * resource,struct res_opts * res_opts)2483 int set_resource_options(struct drbd_resource *resource, struct res_opts *res_opts)
2484 {
2485 struct drbd_connection *connection;
2486 cpumask_var_t new_cpu_mask;
2487 int err;
2488
2489 if (!zalloc_cpumask_var(&new_cpu_mask, GFP_KERNEL))
2490 return -ENOMEM;
2491
2492 /* silently ignore cpu mask on UP kernel */
2493 if (nr_cpu_ids > 1 && res_opts->cpu_mask[0] != 0) {
2494 err = bitmap_parse(res_opts->cpu_mask, DRBD_CPU_MASK_SIZE,
2495 cpumask_bits(new_cpu_mask), nr_cpu_ids);
2496 if (err == -EOVERFLOW) {
2497 /* So what. mask it out. */
2498 cpumask_var_t tmp_cpu_mask;
2499 if (zalloc_cpumask_var(&tmp_cpu_mask, GFP_KERNEL)) {
2500 cpumask_setall(tmp_cpu_mask);
2501 cpumask_and(new_cpu_mask, new_cpu_mask, tmp_cpu_mask);
2502 drbd_warn(resource, "Overflow in bitmap_parse(%.12s%s), truncating to %u bits\n",
2503 res_opts->cpu_mask,
2504 strlen(res_opts->cpu_mask) > 12 ? "..." : "",
2505 nr_cpu_ids);
2506 free_cpumask_var(tmp_cpu_mask);
2507 err = 0;
2508 }
2509 }
2510 if (err) {
2511 drbd_warn(resource, "bitmap_parse() failed with %d\n", err);
2512 /* retcode = ERR_CPU_MASK_PARSE; */
2513 goto fail;
2514 }
2515 }
2516 resource->res_opts = *res_opts;
2517 if (cpumask_empty(new_cpu_mask))
2518 drbd_calc_cpu_mask(&new_cpu_mask);
2519 if (!cpumask_equal(resource->cpu_mask, new_cpu_mask)) {
2520 cpumask_copy(resource->cpu_mask, new_cpu_mask);
2521 for_each_connection_rcu(connection, resource) {
2522 connection->receiver.reset_cpu_mask = 1;
2523 connection->ack_receiver.reset_cpu_mask = 1;
2524 connection->worker.reset_cpu_mask = 1;
2525 }
2526 }
2527 err = 0;
2528
2529 fail:
2530 free_cpumask_var(new_cpu_mask);
2531 return err;
2532
2533 }
2534
drbd_create_resource(const char * name)2535 struct drbd_resource *drbd_create_resource(const char *name)
2536 {
2537 struct drbd_resource *resource;
2538
2539 resource = kzalloc(sizeof(struct drbd_resource), GFP_KERNEL);
2540 if (!resource)
2541 goto fail;
2542 resource->name = kstrdup(name, GFP_KERNEL);
2543 if (!resource->name)
2544 goto fail_free_resource;
2545 if (!zalloc_cpumask_var(&resource->cpu_mask, GFP_KERNEL))
2546 goto fail_free_name;
2547 kref_init(&resource->kref);
2548 idr_init(&resource->devices);
2549 INIT_LIST_HEAD(&resource->connections);
2550 resource->write_ordering = WO_BDEV_FLUSH;
2551 list_add_tail_rcu(&resource->resources, &drbd_resources);
2552 mutex_init(&resource->conf_update);
2553 mutex_init(&resource->adm_mutex);
2554 spin_lock_init(&resource->req_lock);
2555 drbd_debugfs_resource_add(resource);
2556 return resource;
2557
2558 fail_free_name:
2559 kfree(resource->name);
2560 fail_free_resource:
2561 kfree(resource);
2562 fail:
2563 return NULL;
2564 }
2565
2566 /* caller must be under adm_mutex */
conn_create(const char * name,struct res_opts * res_opts)2567 struct drbd_connection *conn_create(const char *name, struct res_opts *res_opts)
2568 {
2569 struct drbd_resource *resource;
2570 struct drbd_connection *connection;
2571
2572 connection = kzalloc(sizeof(struct drbd_connection), GFP_KERNEL);
2573 if (!connection)
2574 return NULL;
2575
2576 if (drbd_alloc_socket(&connection->data))
2577 goto fail;
2578 if (drbd_alloc_socket(&connection->meta))
2579 goto fail;
2580
2581 connection->current_epoch = kzalloc(sizeof(struct drbd_epoch), GFP_KERNEL);
2582 if (!connection->current_epoch)
2583 goto fail;
2584
2585 INIT_LIST_HEAD(&connection->transfer_log);
2586
2587 INIT_LIST_HEAD(&connection->current_epoch->list);
2588 connection->epochs = 1;
2589 spin_lock_init(&connection->epoch_lock);
2590
2591 connection->send.seen_any_write_yet = false;
2592 connection->send.current_epoch_nr = 0;
2593 connection->send.current_epoch_writes = 0;
2594
2595 resource = drbd_create_resource(name);
2596 if (!resource)
2597 goto fail;
2598
2599 connection->cstate = C_STANDALONE;
2600 mutex_init(&connection->cstate_mutex);
2601 init_waitqueue_head(&connection->ping_wait);
2602 idr_init(&connection->peer_devices);
2603
2604 drbd_init_workqueue(&connection->sender_work);
2605 mutex_init(&connection->data.mutex);
2606 mutex_init(&connection->meta.mutex);
2607
2608 drbd_thread_init(resource, &connection->receiver, drbd_receiver, "receiver");
2609 connection->receiver.connection = connection;
2610 drbd_thread_init(resource, &connection->worker, drbd_worker, "worker");
2611 connection->worker.connection = connection;
2612 drbd_thread_init(resource, &connection->ack_receiver, drbd_ack_receiver, "ack_recv");
2613 connection->ack_receiver.connection = connection;
2614
2615 kref_init(&connection->kref);
2616
2617 connection->resource = resource;
2618
2619 if (set_resource_options(resource, res_opts))
2620 goto fail_resource;
2621
2622 kref_get(&resource->kref);
2623 list_add_tail_rcu(&connection->connections, &resource->connections);
2624 drbd_debugfs_connection_add(connection);
2625 return connection;
2626
2627 fail_resource:
2628 list_del(&resource->resources);
2629 drbd_free_resource(resource);
2630 fail:
2631 kfree(connection->current_epoch);
2632 drbd_free_socket(&connection->meta);
2633 drbd_free_socket(&connection->data);
2634 kfree(connection);
2635 return NULL;
2636 }
2637
drbd_destroy_connection(struct kref * kref)2638 void drbd_destroy_connection(struct kref *kref)
2639 {
2640 struct drbd_connection *connection = container_of(kref, struct drbd_connection, kref);
2641 struct drbd_resource *resource = connection->resource;
2642
2643 if (atomic_read(&connection->current_epoch->epoch_size) != 0)
2644 drbd_err(connection, "epoch_size:%d\n", atomic_read(&connection->current_epoch->epoch_size));
2645 kfree(connection->current_epoch);
2646
2647 idr_destroy(&connection->peer_devices);
2648
2649 drbd_free_socket(&connection->meta);
2650 drbd_free_socket(&connection->data);
2651 kfree(connection->int_dig_in);
2652 kfree(connection->int_dig_vv);
2653 kfree(connection);
2654 kref_put(&resource->kref, drbd_destroy_resource);
2655 }
2656
init_submitter(struct drbd_device * device)2657 static int init_submitter(struct drbd_device *device)
2658 {
2659 /* opencoded create_singlethread_workqueue(),
2660 * to be able to say "drbd%d", ..., minor */
2661 device->submit.wq =
2662 alloc_ordered_workqueue("drbd%u_submit", WQ_MEM_RECLAIM, device->minor);
2663 if (!device->submit.wq)
2664 return -ENOMEM;
2665
2666 INIT_WORK(&device->submit.worker, do_submit);
2667 INIT_LIST_HEAD(&device->submit.writes);
2668 return 0;
2669 }
2670
drbd_create_device(struct drbd_config_context * adm_ctx,unsigned int minor)2671 enum drbd_ret_code drbd_create_device(struct drbd_config_context *adm_ctx, unsigned int minor)
2672 {
2673 struct drbd_resource *resource = adm_ctx->resource;
2674 struct drbd_connection *connection, *n;
2675 struct drbd_device *device;
2676 struct drbd_peer_device *peer_device, *tmp_peer_device;
2677 struct gendisk *disk;
2678 int id;
2679 int vnr = adm_ctx->volume;
2680 enum drbd_ret_code err = ERR_NOMEM;
2681
2682 device = minor_to_device(minor);
2683 if (device)
2684 return ERR_MINOR_OR_VOLUME_EXISTS;
2685
2686 /* GFP_KERNEL, we are outside of all write-out paths */
2687 device = kzalloc(sizeof(struct drbd_device), GFP_KERNEL);
2688 if (!device)
2689 return ERR_NOMEM;
2690 kref_init(&device->kref);
2691
2692 kref_get(&resource->kref);
2693 device->resource = resource;
2694 device->minor = minor;
2695 device->vnr = vnr;
2696
2697 drbd_init_set_defaults(device);
2698
2699 disk = blk_alloc_disk(NUMA_NO_NODE);
2700 if (!disk)
2701 goto out_no_disk;
2702
2703 device->vdisk = disk;
2704 device->rq_queue = disk->queue;
2705
2706 set_disk_ro(disk, true);
2707
2708 disk->major = DRBD_MAJOR;
2709 disk->first_minor = minor;
2710 disk->minors = 1;
2711 disk->fops = &drbd_ops;
2712 disk->flags |= GENHD_FL_NO_PART;
2713 sprintf(disk->disk_name, "drbd%d", minor);
2714 disk->private_data = device;
2715
2716 blk_queue_flag_set(QUEUE_FLAG_STABLE_WRITES, disk->queue);
2717 blk_queue_write_cache(disk->queue, true, true);
2718 /* Setting the max_hw_sectors to an odd value of 8kibyte here
2719 This triggers a max_bio_size message upon first attach or connect */
2720 blk_queue_max_hw_sectors(disk->queue, DRBD_MAX_BIO_SIZE_SAFE >> 8);
2721
2722 device->md_io.page = alloc_page(GFP_KERNEL);
2723 if (!device->md_io.page)
2724 goto out_no_io_page;
2725
2726 if (drbd_bm_init(device))
2727 goto out_no_bitmap;
2728 device->read_requests = RB_ROOT;
2729 device->write_requests = RB_ROOT;
2730
2731 id = idr_alloc(&drbd_devices, device, minor, minor + 1, GFP_KERNEL);
2732 if (id < 0) {
2733 if (id == -ENOSPC)
2734 err = ERR_MINOR_OR_VOLUME_EXISTS;
2735 goto out_no_minor_idr;
2736 }
2737 kref_get(&device->kref);
2738
2739 id = idr_alloc(&resource->devices, device, vnr, vnr + 1, GFP_KERNEL);
2740 if (id < 0) {
2741 if (id == -ENOSPC)
2742 err = ERR_MINOR_OR_VOLUME_EXISTS;
2743 goto out_idr_remove_minor;
2744 }
2745 kref_get(&device->kref);
2746
2747 INIT_LIST_HEAD(&device->peer_devices);
2748 INIT_LIST_HEAD(&device->pending_bitmap_io);
2749 for_each_connection(connection, resource) {
2750 peer_device = kzalloc(sizeof(struct drbd_peer_device), GFP_KERNEL);
2751 if (!peer_device)
2752 goto out_idr_remove_from_resource;
2753 peer_device->connection = connection;
2754 peer_device->device = device;
2755
2756 list_add(&peer_device->peer_devices, &device->peer_devices);
2757 kref_get(&device->kref);
2758
2759 id = idr_alloc(&connection->peer_devices, peer_device, vnr, vnr + 1, GFP_KERNEL);
2760 if (id < 0) {
2761 if (id == -ENOSPC)
2762 err = ERR_INVALID_REQUEST;
2763 goto out_idr_remove_from_resource;
2764 }
2765 kref_get(&connection->kref);
2766 INIT_WORK(&peer_device->send_acks_work, drbd_send_acks_wf);
2767 }
2768
2769 if (init_submitter(device)) {
2770 err = ERR_NOMEM;
2771 goto out_idr_remove_from_resource;
2772 }
2773
2774 err = add_disk(disk);
2775 if (err)
2776 goto out_destroy_workqueue;
2777
2778 /* inherit the connection state */
2779 device->state.conn = first_connection(resource)->cstate;
2780 if (device->state.conn == C_WF_REPORT_PARAMS) {
2781 for_each_peer_device(peer_device, device)
2782 drbd_connected(peer_device);
2783 }
2784 /* move to create_peer_device() */
2785 for_each_peer_device(peer_device, device)
2786 drbd_debugfs_peer_device_add(peer_device);
2787 drbd_debugfs_device_add(device);
2788 return NO_ERROR;
2789
2790 out_destroy_workqueue:
2791 destroy_workqueue(device->submit.wq);
2792 out_idr_remove_from_resource:
2793 for_each_connection_safe(connection, n, resource) {
2794 peer_device = idr_remove(&connection->peer_devices, vnr);
2795 if (peer_device)
2796 kref_put(&connection->kref, drbd_destroy_connection);
2797 }
2798 for_each_peer_device_safe(peer_device, tmp_peer_device, device) {
2799 list_del(&peer_device->peer_devices);
2800 kfree(peer_device);
2801 }
2802 idr_remove(&resource->devices, vnr);
2803 out_idr_remove_minor:
2804 idr_remove(&drbd_devices, minor);
2805 synchronize_rcu();
2806 out_no_minor_idr:
2807 drbd_bm_cleanup(device);
2808 out_no_bitmap:
2809 __free_page(device->md_io.page);
2810 out_no_io_page:
2811 put_disk(disk);
2812 out_no_disk:
2813 kref_put(&resource->kref, drbd_destroy_resource);
2814 kfree(device);
2815 return err;
2816 }
2817
drbd_delete_device(struct drbd_device * device)2818 void drbd_delete_device(struct drbd_device *device)
2819 {
2820 struct drbd_resource *resource = device->resource;
2821 struct drbd_connection *connection;
2822 struct drbd_peer_device *peer_device;
2823
2824 /* move to free_peer_device() */
2825 for_each_peer_device(peer_device, device)
2826 drbd_debugfs_peer_device_cleanup(peer_device);
2827 drbd_debugfs_device_cleanup(device);
2828 for_each_connection(connection, resource) {
2829 idr_remove(&connection->peer_devices, device->vnr);
2830 kref_put(&device->kref, drbd_destroy_device);
2831 }
2832 idr_remove(&resource->devices, device->vnr);
2833 kref_put(&device->kref, drbd_destroy_device);
2834 idr_remove(&drbd_devices, device_to_minor(device));
2835 kref_put(&device->kref, drbd_destroy_device);
2836 del_gendisk(device->vdisk);
2837 synchronize_rcu();
2838 kref_put(&device->kref, drbd_destroy_device);
2839 }
2840
drbd_init(void)2841 static int __init drbd_init(void)
2842 {
2843 int err;
2844
2845 if (drbd_minor_count < DRBD_MINOR_COUNT_MIN || drbd_minor_count > DRBD_MINOR_COUNT_MAX) {
2846 pr_err("invalid minor_count (%d)\n", drbd_minor_count);
2847 #ifdef MODULE
2848 return -EINVAL;
2849 #else
2850 drbd_minor_count = DRBD_MINOR_COUNT_DEF;
2851 #endif
2852 }
2853
2854 err = register_blkdev(DRBD_MAJOR, "drbd");
2855 if (err) {
2856 pr_err("unable to register block device major %d\n",
2857 DRBD_MAJOR);
2858 return err;
2859 }
2860
2861 /*
2862 * allocate all necessary structs
2863 */
2864 init_waitqueue_head(&drbd_pp_wait);
2865
2866 drbd_proc = NULL; /* play safe for drbd_cleanup */
2867 idr_init(&drbd_devices);
2868
2869 mutex_init(&resources_mutex);
2870 INIT_LIST_HEAD(&drbd_resources);
2871
2872 err = drbd_genl_register();
2873 if (err) {
2874 pr_err("unable to register generic netlink family\n");
2875 goto fail;
2876 }
2877
2878 err = drbd_create_mempools();
2879 if (err)
2880 goto fail;
2881
2882 err = -ENOMEM;
2883 drbd_proc = proc_create_single("drbd", S_IFREG | 0444 , NULL, drbd_seq_show);
2884 if (!drbd_proc) {
2885 pr_err("unable to register proc file\n");
2886 goto fail;
2887 }
2888
2889 retry.wq = create_singlethread_workqueue("drbd-reissue");
2890 if (!retry.wq) {
2891 pr_err("unable to create retry workqueue\n");
2892 goto fail;
2893 }
2894 INIT_WORK(&retry.worker, do_retry);
2895 spin_lock_init(&retry.lock);
2896 INIT_LIST_HEAD(&retry.writes);
2897
2898 drbd_debugfs_init();
2899
2900 pr_info("initialized. "
2901 "Version: " REL_VERSION " (api:%d/proto:%d-%d)\n",
2902 API_VERSION, PRO_VERSION_MIN, PRO_VERSION_MAX);
2903 pr_info("%s\n", drbd_buildtag());
2904 pr_info("registered as block device major %d\n", DRBD_MAJOR);
2905 return 0; /* Success! */
2906
2907 fail:
2908 drbd_cleanup();
2909 if (err == -ENOMEM)
2910 pr_err("ran out of memory\n");
2911 else
2912 pr_err("initialization failure\n");
2913 return err;
2914 }
2915
drbd_free_one_sock(struct drbd_socket * ds)2916 static void drbd_free_one_sock(struct drbd_socket *ds)
2917 {
2918 struct socket *s;
2919 mutex_lock(&ds->mutex);
2920 s = ds->socket;
2921 ds->socket = NULL;
2922 mutex_unlock(&ds->mutex);
2923 if (s) {
2924 /* so debugfs does not need to mutex_lock() */
2925 synchronize_rcu();
2926 kernel_sock_shutdown(s, SHUT_RDWR);
2927 sock_release(s);
2928 }
2929 }
2930
drbd_free_sock(struct drbd_connection * connection)2931 void drbd_free_sock(struct drbd_connection *connection)
2932 {
2933 if (connection->data.socket)
2934 drbd_free_one_sock(&connection->data);
2935 if (connection->meta.socket)
2936 drbd_free_one_sock(&connection->meta);
2937 }
2938
2939 /* meta data management */
2940
conn_md_sync(struct drbd_connection * connection)2941 void conn_md_sync(struct drbd_connection *connection)
2942 {
2943 struct drbd_peer_device *peer_device;
2944 int vnr;
2945
2946 rcu_read_lock();
2947 idr_for_each_entry(&connection->peer_devices, peer_device, vnr) {
2948 struct drbd_device *device = peer_device->device;
2949
2950 kref_get(&device->kref);
2951 rcu_read_unlock();
2952 drbd_md_sync(device);
2953 kref_put(&device->kref, drbd_destroy_device);
2954 rcu_read_lock();
2955 }
2956 rcu_read_unlock();
2957 }
2958
2959 /* aligned 4kByte */
2960 struct meta_data_on_disk {
2961 u64 la_size_sect; /* last agreed size. */
2962 u64 uuid[UI_SIZE]; /* UUIDs. */
2963 u64 device_uuid;
2964 u64 reserved_u64_1;
2965 u32 flags; /* MDF */
2966 u32 magic;
2967 u32 md_size_sect;
2968 u32 al_offset; /* offset to this block */
2969 u32 al_nr_extents; /* important for restoring the AL (userspace) */
2970 /* `-- act_log->nr_elements <-- ldev->dc.al_extents */
2971 u32 bm_offset; /* offset to the bitmap, from here */
2972 u32 bm_bytes_per_bit; /* BM_BLOCK_SIZE */
2973 u32 la_peer_max_bio_size; /* last peer max_bio_size */
2974
2975 /* see al_tr_number_to_on_disk_sector() */
2976 u32 al_stripes;
2977 u32 al_stripe_size_4k;
2978
2979 u8 reserved_u8[4096 - (7*8 + 10*4)];
2980 } __packed;
2981
2982
2983
drbd_md_write(struct drbd_device * device,void * b)2984 void drbd_md_write(struct drbd_device *device, void *b)
2985 {
2986 struct meta_data_on_disk *buffer = b;
2987 sector_t sector;
2988 int i;
2989
2990 memset(buffer, 0, sizeof(*buffer));
2991
2992 buffer->la_size_sect = cpu_to_be64(get_capacity(device->vdisk));
2993 for (i = UI_CURRENT; i < UI_SIZE; i++)
2994 buffer->uuid[i] = cpu_to_be64(device->ldev->md.uuid[i]);
2995 buffer->flags = cpu_to_be32(device->ldev->md.flags);
2996 buffer->magic = cpu_to_be32(DRBD_MD_MAGIC_84_UNCLEAN);
2997
2998 buffer->md_size_sect = cpu_to_be32(device->ldev->md.md_size_sect);
2999 buffer->al_offset = cpu_to_be32(device->ldev->md.al_offset);
3000 buffer->al_nr_extents = cpu_to_be32(device->act_log->nr_elements);
3001 buffer->bm_bytes_per_bit = cpu_to_be32(BM_BLOCK_SIZE);
3002 buffer->device_uuid = cpu_to_be64(device->ldev->md.device_uuid);
3003
3004 buffer->bm_offset = cpu_to_be32(device->ldev->md.bm_offset);
3005 buffer->la_peer_max_bio_size = cpu_to_be32(device->peer_max_bio_size);
3006
3007 buffer->al_stripes = cpu_to_be32(device->ldev->md.al_stripes);
3008 buffer->al_stripe_size_4k = cpu_to_be32(device->ldev->md.al_stripe_size_4k);
3009
3010 D_ASSERT(device, drbd_md_ss(device->ldev) == device->ldev->md.md_offset);
3011 sector = device->ldev->md.md_offset;
3012
3013 if (drbd_md_sync_page_io(device, device->ldev, sector, REQ_OP_WRITE)) {
3014 /* this was a try anyways ... */
3015 drbd_err(device, "meta data update failed!\n");
3016 drbd_chk_io_error(device, 1, DRBD_META_IO_ERROR);
3017 }
3018 }
3019
3020 /**
3021 * drbd_md_sync() - Writes the meta data super block if the MD_DIRTY flag bit is set
3022 * @device: DRBD device.
3023 */
drbd_md_sync(struct drbd_device * device)3024 void drbd_md_sync(struct drbd_device *device)
3025 {
3026 struct meta_data_on_disk *buffer;
3027
3028 /* Don't accidentally change the DRBD meta data layout. */
3029 BUILD_BUG_ON(UI_SIZE != 4);
3030 BUILD_BUG_ON(sizeof(struct meta_data_on_disk) != 4096);
3031
3032 del_timer(&device->md_sync_timer);
3033 /* timer may be rearmed by drbd_md_mark_dirty() now. */
3034 if (!test_and_clear_bit(MD_DIRTY, &device->flags))
3035 return;
3036
3037 /* We use here D_FAILED and not D_ATTACHING because we try to write
3038 * metadata even if we detach due to a disk failure! */
3039 if (!get_ldev_if_state(device, D_FAILED))
3040 return;
3041
3042 buffer = drbd_md_get_buffer(device, __func__);
3043 if (!buffer)
3044 goto out;
3045
3046 drbd_md_write(device, buffer);
3047
3048 /* Update device->ldev->md.la_size_sect,
3049 * since we updated it on metadata. */
3050 device->ldev->md.la_size_sect = get_capacity(device->vdisk);
3051
3052 drbd_md_put_buffer(device);
3053 out:
3054 put_ldev(device);
3055 }
3056
check_activity_log_stripe_size(struct drbd_device * device,struct meta_data_on_disk * on_disk,struct drbd_md * in_core)3057 static int check_activity_log_stripe_size(struct drbd_device *device,
3058 struct meta_data_on_disk *on_disk,
3059 struct drbd_md *in_core)
3060 {
3061 u32 al_stripes = be32_to_cpu(on_disk->al_stripes);
3062 u32 al_stripe_size_4k = be32_to_cpu(on_disk->al_stripe_size_4k);
3063 u64 al_size_4k;
3064
3065 /* both not set: default to old fixed size activity log */
3066 if (al_stripes == 0 && al_stripe_size_4k == 0) {
3067 al_stripes = 1;
3068 al_stripe_size_4k = MD_32kB_SECT/8;
3069 }
3070
3071 /* some paranoia plausibility checks */
3072
3073 /* we need both values to be set */
3074 if (al_stripes == 0 || al_stripe_size_4k == 0)
3075 goto err;
3076
3077 al_size_4k = (u64)al_stripes * al_stripe_size_4k;
3078
3079 /* Upper limit of activity log area, to avoid potential overflow
3080 * problems in al_tr_number_to_on_disk_sector(). As right now, more
3081 * than 72 * 4k blocks total only increases the amount of history,
3082 * limiting this arbitrarily to 16 GB is not a real limitation ;-) */
3083 if (al_size_4k > (16 * 1024 * 1024/4))
3084 goto err;
3085
3086 /* Lower limit: we need at least 8 transaction slots (32kB)
3087 * to not break existing setups */
3088 if (al_size_4k < MD_32kB_SECT/8)
3089 goto err;
3090
3091 in_core->al_stripe_size_4k = al_stripe_size_4k;
3092 in_core->al_stripes = al_stripes;
3093 in_core->al_size_4k = al_size_4k;
3094
3095 return 0;
3096 err:
3097 drbd_err(device, "invalid activity log striping: al_stripes=%u, al_stripe_size_4k=%u\n",
3098 al_stripes, al_stripe_size_4k);
3099 return -EINVAL;
3100 }
3101
check_offsets_and_sizes(struct drbd_device * device,struct drbd_backing_dev * bdev)3102 static int check_offsets_and_sizes(struct drbd_device *device, struct drbd_backing_dev *bdev)
3103 {
3104 sector_t capacity = drbd_get_capacity(bdev->md_bdev);
3105 struct drbd_md *in_core = &bdev->md;
3106 s32 on_disk_al_sect;
3107 s32 on_disk_bm_sect;
3108
3109 /* The on-disk size of the activity log, calculated from offsets, and
3110 * the size of the activity log calculated from the stripe settings,
3111 * should match.
3112 * Though we could relax this a bit: it is ok, if the striped activity log
3113 * fits in the available on-disk activity log size.
3114 * Right now, that would break how resize is implemented.
3115 * TODO: make drbd_determine_dev_size() (and the drbdmeta tool) aware
3116 * of possible unused padding space in the on disk layout. */
3117 if (in_core->al_offset < 0) {
3118 if (in_core->bm_offset > in_core->al_offset)
3119 goto err;
3120 on_disk_al_sect = -in_core->al_offset;
3121 on_disk_bm_sect = in_core->al_offset - in_core->bm_offset;
3122 } else {
3123 if (in_core->al_offset != MD_4kB_SECT)
3124 goto err;
3125 if (in_core->bm_offset < in_core->al_offset + in_core->al_size_4k * MD_4kB_SECT)
3126 goto err;
3127
3128 on_disk_al_sect = in_core->bm_offset - MD_4kB_SECT;
3129 on_disk_bm_sect = in_core->md_size_sect - in_core->bm_offset;
3130 }
3131
3132 /* old fixed size meta data is exactly that: fixed. */
3133 if (in_core->meta_dev_idx >= 0) {
3134 if (in_core->md_size_sect != MD_128MB_SECT
3135 || in_core->al_offset != MD_4kB_SECT
3136 || in_core->bm_offset != MD_4kB_SECT + MD_32kB_SECT
3137 || in_core->al_stripes != 1
3138 || in_core->al_stripe_size_4k != MD_32kB_SECT/8)
3139 goto err;
3140 }
3141
3142 if (capacity < in_core->md_size_sect)
3143 goto err;
3144 if (capacity - in_core->md_size_sect < drbd_md_first_sector(bdev))
3145 goto err;
3146
3147 /* should be aligned, and at least 32k */
3148 if ((on_disk_al_sect & 7) || (on_disk_al_sect < MD_32kB_SECT))
3149 goto err;
3150
3151 /* should fit (for now: exactly) into the available on-disk space;
3152 * overflow prevention is in check_activity_log_stripe_size() above. */
3153 if (on_disk_al_sect != in_core->al_size_4k * MD_4kB_SECT)
3154 goto err;
3155
3156 /* again, should be aligned */
3157 if (in_core->bm_offset & 7)
3158 goto err;
3159
3160 /* FIXME check for device grow with flex external meta data? */
3161
3162 /* can the available bitmap space cover the last agreed device size? */
3163 if (on_disk_bm_sect < (in_core->la_size_sect+7)/MD_4kB_SECT/8/512)
3164 goto err;
3165
3166 return 0;
3167
3168 err:
3169 drbd_err(device, "meta data offsets don't make sense: idx=%d "
3170 "al_s=%u, al_sz4k=%u, al_offset=%d, bm_offset=%d, "
3171 "md_size_sect=%u, la_size=%llu, md_capacity=%llu\n",
3172 in_core->meta_dev_idx,
3173 in_core->al_stripes, in_core->al_stripe_size_4k,
3174 in_core->al_offset, in_core->bm_offset, in_core->md_size_sect,
3175 (unsigned long long)in_core->la_size_sect,
3176 (unsigned long long)capacity);
3177
3178 return -EINVAL;
3179 }
3180
3181
3182 /**
3183 * drbd_md_read() - Reads in the meta data super block
3184 * @device: DRBD device.
3185 * @bdev: Device from which the meta data should be read in.
3186 *
3187 * Return NO_ERROR on success, and an enum drbd_ret_code in case
3188 * something goes wrong.
3189 *
3190 * Called exactly once during drbd_adm_attach(), while still being D_DISKLESS,
3191 * even before @bdev is assigned to @device->ldev.
3192 */
drbd_md_read(struct drbd_device * device,struct drbd_backing_dev * bdev)3193 int drbd_md_read(struct drbd_device *device, struct drbd_backing_dev *bdev)
3194 {
3195 struct meta_data_on_disk *buffer;
3196 u32 magic, flags;
3197 int i, rv = NO_ERROR;
3198
3199 if (device->state.disk != D_DISKLESS)
3200 return ERR_DISK_CONFIGURED;
3201
3202 buffer = drbd_md_get_buffer(device, __func__);
3203 if (!buffer)
3204 return ERR_NOMEM;
3205
3206 /* First, figure out where our meta data superblock is located,
3207 * and read it. */
3208 bdev->md.meta_dev_idx = bdev->disk_conf->meta_dev_idx;
3209 bdev->md.md_offset = drbd_md_ss(bdev);
3210 /* Even for (flexible or indexed) external meta data,
3211 * initially restrict us to the 4k superblock for now.
3212 * Affects the paranoia out-of-range access check in drbd_md_sync_page_io(). */
3213 bdev->md.md_size_sect = 8;
3214
3215 if (drbd_md_sync_page_io(device, bdev, bdev->md.md_offset,
3216 REQ_OP_READ)) {
3217 /* NOTE: can't do normal error processing here as this is
3218 called BEFORE disk is attached */
3219 drbd_err(device, "Error while reading metadata.\n");
3220 rv = ERR_IO_MD_DISK;
3221 goto err;
3222 }
3223
3224 magic = be32_to_cpu(buffer->magic);
3225 flags = be32_to_cpu(buffer->flags);
3226 if (magic == DRBD_MD_MAGIC_84_UNCLEAN ||
3227 (magic == DRBD_MD_MAGIC_08 && !(flags & MDF_AL_CLEAN))) {
3228 /* btw: that's Activity Log clean, not "all" clean. */
3229 drbd_err(device, "Found unclean meta data. Did you \"drbdadm apply-al\"?\n");
3230 rv = ERR_MD_UNCLEAN;
3231 goto err;
3232 }
3233
3234 rv = ERR_MD_INVALID;
3235 if (magic != DRBD_MD_MAGIC_08) {
3236 if (magic == DRBD_MD_MAGIC_07)
3237 drbd_err(device, "Found old (0.7) meta data magic. Did you \"drbdadm create-md\"?\n");
3238 else
3239 drbd_err(device, "Meta data magic not found. Did you \"drbdadm create-md\"?\n");
3240 goto err;
3241 }
3242
3243 if (be32_to_cpu(buffer->bm_bytes_per_bit) != BM_BLOCK_SIZE) {
3244 drbd_err(device, "unexpected bm_bytes_per_bit: %u (expected %u)\n",
3245 be32_to_cpu(buffer->bm_bytes_per_bit), BM_BLOCK_SIZE);
3246 goto err;
3247 }
3248
3249
3250 /* convert to in_core endian */
3251 bdev->md.la_size_sect = be64_to_cpu(buffer->la_size_sect);
3252 for (i = UI_CURRENT; i < UI_SIZE; i++)
3253 bdev->md.uuid[i] = be64_to_cpu(buffer->uuid[i]);
3254 bdev->md.flags = be32_to_cpu(buffer->flags);
3255 bdev->md.device_uuid = be64_to_cpu(buffer->device_uuid);
3256
3257 bdev->md.md_size_sect = be32_to_cpu(buffer->md_size_sect);
3258 bdev->md.al_offset = be32_to_cpu(buffer->al_offset);
3259 bdev->md.bm_offset = be32_to_cpu(buffer->bm_offset);
3260
3261 if (check_activity_log_stripe_size(device, buffer, &bdev->md))
3262 goto err;
3263 if (check_offsets_and_sizes(device, bdev))
3264 goto err;
3265
3266 if (be32_to_cpu(buffer->bm_offset) != bdev->md.bm_offset) {
3267 drbd_err(device, "unexpected bm_offset: %d (expected %d)\n",
3268 be32_to_cpu(buffer->bm_offset), bdev->md.bm_offset);
3269 goto err;
3270 }
3271 if (be32_to_cpu(buffer->md_size_sect) != bdev->md.md_size_sect) {
3272 drbd_err(device, "unexpected md_size: %u (expected %u)\n",
3273 be32_to_cpu(buffer->md_size_sect), bdev->md.md_size_sect);
3274 goto err;
3275 }
3276
3277 rv = NO_ERROR;
3278
3279 spin_lock_irq(&device->resource->req_lock);
3280 if (device->state.conn < C_CONNECTED) {
3281 unsigned int peer;
3282 peer = be32_to_cpu(buffer->la_peer_max_bio_size);
3283 peer = max(peer, DRBD_MAX_BIO_SIZE_SAFE);
3284 device->peer_max_bio_size = peer;
3285 }
3286 spin_unlock_irq(&device->resource->req_lock);
3287
3288 err:
3289 drbd_md_put_buffer(device);
3290
3291 return rv;
3292 }
3293
3294 /**
3295 * drbd_md_mark_dirty() - Mark meta data super block as dirty
3296 * @device: DRBD device.
3297 *
3298 * Call this function if you change anything that should be written to
3299 * the meta-data super block. This function sets MD_DIRTY, and starts a
3300 * timer that ensures that within five seconds you have to call drbd_md_sync().
3301 */
drbd_md_mark_dirty(struct drbd_device * device)3302 void drbd_md_mark_dirty(struct drbd_device *device)
3303 {
3304 if (!test_and_set_bit(MD_DIRTY, &device->flags))
3305 mod_timer(&device->md_sync_timer, jiffies + 5*HZ);
3306 }
3307
drbd_uuid_move_history(struct drbd_device * device)3308 void drbd_uuid_move_history(struct drbd_device *device) __must_hold(local)
3309 {
3310 int i;
3311
3312 for (i = UI_HISTORY_START; i < UI_HISTORY_END; i++)
3313 device->ldev->md.uuid[i+1] = device->ldev->md.uuid[i];
3314 }
3315
__drbd_uuid_set(struct drbd_device * device,int idx,u64 val)3316 void __drbd_uuid_set(struct drbd_device *device, int idx, u64 val) __must_hold(local)
3317 {
3318 if (idx == UI_CURRENT) {
3319 if (device->state.role == R_PRIMARY)
3320 val |= 1;
3321 else
3322 val &= ~((u64)1);
3323
3324 drbd_set_ed_uuid(device, val);
3325 }
3326
3327 device->ldev->md.uuid[idx] = val;
3328 drbd_md_mark_dirty(device);
3329 }
3330
_drbd_uuid_set(struct drbd_device * device,int idx,u64 val)3331 void _drbd_uuid_set(struct drbd_device *device, int idx, u64 val) __must_hold(local)
3332 {
3333 unsigned long flags;
3334 spin_lock_irqsave(&device->ldev->md.uuid_lock, flags);
3335 __drbd_uuid_set(device, idx, val);
3336 spin_unlock_irqrestore(&device->ldev->md.uuid_lock, flags);
3337 }
3338
drbd_uuid_set(struct drbd_device * device,int idx,u64 val)3339 void drbd_uuid_set(struct drbd_device *device, int idx, u64 val) __must_hold(local)
3340 {
3341 unsigned long flags;
3342 spin_lock_irqsave(&device->ldev->md.uuid_lock, flags);
3343 if (device->ldev->md.uuid[idx]) {
3344 drbd_uuid_move_history(device);
3345 device->ldev->md.uuid[UI_HISTORY_START] = device->ldev->md.uuid[idx];
3346 }
3347 __drbd_uuid_set(device, idx, val);
3348 spin_unlock_irqrestore(&device->ldev->md.uuid_lock, flags);
3349 }
3350
3351 /**
3352 * drbd_uuid_new_current() - Creates a new current UUID
3353 * @device: DRBD device.
3354 *
3355 * Creates a new current UUID, and rotates the old current UUID into
3356 * the bitmap slot. Causes an incremental resync upon next connect.
3357 */
drbd_uuid_new_current(struct drbd_device * device)3358 void drbd_uuid_new_current(struct drbd_device *device) __must_hold(local)
3359 {
3360 u64 val;
3361 unsigned long long bm_uuid;
3362
3363 get_random_bytes(&val, sizeof(u64));
3364
3365 spin_lock_irq(&device->ldev->md.uuid_lock);
3366 bm_uuid = device->ldev->md.uuid[UI_BITMAP];
3367
3368 if (bm_uuid)
3369 drbd_warn(device, "bm UUID was already set: %llX\n", bm_uuid);
3370
3371 device->ldev->md.uuid[UI_BITMAP] = device->ldev->md.uuid[UI_CURRENT];
3372 __drbd_uuid_set(device, UI_CURRENT, val);
3373 spin_unlock_irq(&device->ldev->md.uuid_lock);
3374
3375 drbd_print_uuids(device, "new current UUID");
3376 /* get it to stable storage _now_ */
3377 drbd_md_sync(device);
3378 }
3379
drbd_uuid_set_bm(struct drbd_device * device,u64 val)3380 void drbd_uuid_set_bm(struct drbd_device *device, u64 val) __must_hold(local)
3381 {
3382 unsigned long flags;
3383 if (device->ldev->md.uuid[UI_BITMAP] == 0 && val == 0)
3384 return;
3385
3386 spin_lock_irqsave(&device->ldev->md.uuid_lock, flags);
3387 if (val == 0) {
3388 drbd_uuid_move_history(device);
3389 device->ldev->md.uuid[UI_HISTORY_START] = device->ldev->md.uuid[UI_BITMAP];
3390 device->ldev->md.uuid[UI_BITMAP] = 0;
3391 } else {
3392 unsigned long long bm_uuid = device->ldev->md.uuid[UI_BITMAP];
3393 if (bm_uuid)
3394 drbd_warn(device, "bm UUID was already set: %llX\n", bm_uuid);
3395
3396 device->ldev->md.uuid[UI_BITMAP] = val & ~((u64)1);
3397 }
3398 spin_unlock_irqrestore(&device->ldev->md.uuid_lock, flags);
3399
3400 drbd_md_mark_dirty(device);
3401 }
3402
3403 /**
3404 * drbd_bmio_set_n_write() - io_fn for drbd_queue_bitmap_io() or drbd_bitmap_io()
3405 * @device: DRBD device.
3406 *
3407 * Sets all bits in the bitmap and writes the whole bitmap to stable storage.
3408 */
drbd_bmio_set_n_write(struct drbd_device * device)3409 int drbd_bmio_set_n_write(struct drbd_device *device) __must_hold(local)
3410 {
3411 int rv = -EIO;
3412
3413 drbd_md_set_flag(device, MDF_FULL_SYNC);
3414 drbd_md_sync(device);
3415 drbd_bm_set_all(device);
3416
3417 rv = drbd_bm_write(device);
3418
3419 if (!rv) {
3420 drbd_md_clear_flag(device, MDF_FULL_SYNC);
3421 drbd_md_sync(device);
3422 }
3423
3424 return rv;
3425 }
3426
3427 /**
3428 * drbd_bmio_clear_n_write() - io_fn for drbd_queue_bitmap_io() or drbd_bitmap_io()
3429 * @device: DRBD device.
3430 *
3431 * Clears all bits in the bitmap and writes the whole bitmap to stable storage.
3432 */
drbd_bmio_clear_n_write(struct drbd_device * device)3433 int drbd_bmio_clear_n_write(struct drbd_device *device) __must_hold(local)
3434 {
3435 drbd_resume_al(device);
3436 drbd_bm_clear_all(device);
3437 return drbd_bm_write(device);
3438 }
3439
w_bitmap_io(struct drbd_work * w,int unused)3440 static int w_bitmap_io(struct drbd_work *w, int unused)
3441 {
3442 struct drbd_device *device =
3443 container_of(w, struct drbd_device, bm_io_work.w);
3444 struct bm_io_work *work = &device->bm_io_work;
3445 int rv = -EIO;
3446
3447 if (work->flags != BM_LOCKED_CHANGE_ALLOWED) {
3448 int cnt = atomic_read(&device->ap_bio_cnt);
3449 if (cnt)
3450 drbd_err(device, "FIXME: ap_bio_cnt %d, expected 0; queued for '%s'\n",
3451 cnt, work->why);
3452 }
3453
3454 if (get_ldev(device)) {
3455 drbd_bm_lock(device, work->why, work->flags);
3456 rv = work->io_fn(device);
3457 drbd_bm_unlock(device);
3458 put_ldev(device);
3459 }
3460
3461 clear_bit_unlock(BITMAP_IO, &device->flags);
3462 wake_up(&device->misc_wait);
3463
3464 if (work->done)
3465 work->done(device, rv);
3466
3467 clear_bit(BITMAP_IO_QUEUED, &device->flags);
3468 work->why = NULL;
3469 work->flags = 0;
3470
3471 return 0;
3472 }
3473
3474 /**
3475 * drbd_queue_bitmap_io() - Queues an IO operation on the whole bitmap
3476 * @device: DRBD device.
3477 * @io_fn: IO callback to be called when bitmap IO is possible
3478 * @done: callback to be called after the bitmap IO was performed
3479 * @why: Descriptive text of the reason for doing the IO
3480 * @flags: Bitmap flags
3481 *
3482 * While IO on the bitmap happens we freeze application IO thus we ensure
3483 * that drbd_set_out_of_sync() can not be called. This function MAY ONLY be
3484 * called from worker context. It MUST NOT be used while a previous such
3485 * work is still pending!
3486 *
3487 * Its worker function encloses the call of io_fn() by get_ldev() and
3488 * put_ldev().
3489 */
drbd_queue_bitmap_io(struct drbd_device * device,int (* io_fn)(struct drbd_device *),void (* done)(struct drbd_device *,int),char * why,enum bm_flag flags)3490 void drbd_queue_bitmap_io(struct drbd_device *device,
3491 int (*io_fn)(struct drbd_device *),
3492 void (*done)(struct drbd_device *, int),
3493 char *why, enum bm_flag flags)
3494 {
3495 D_ASSERT(device, current == first_peer_device(device)->connection->worker.task);
3496
3497 D_ASSERT(device, !test_bit(BITMAP_IO_QUEUED, &device->flags));
3498 D_ASSERT(device, !test_bit(BITMAP_IO, &device->flags));
3499 D_ASSERT(device, list_empty(&device->bm_io_work.w.list));
3500 if (device->bm_io_work.why)
3501 drbd_err(device, "FIXME going to queue '%s' but '%s' still pending?\n",
3502 why, device->bm_io_work.why);
3503
3504 device->bm_io_work.io_fn = io_fn;
3505 device->bm_io_work.done = done;
3506 device->bm_io_work.why = why;
3507 device->bm_io_work.flags = flags;
3508
3509 spin_lock_irq(&device->resource->req_lock);
3510 set_bit(BITMAP_IO, &device->flags);
3511 /* don't wait for pending application IO if the caller indicates that
3512 * application IO does not conflict anyways. */
3513 if (flags == BM_LOCKED_CHANGE_ALLOWED || atomic_read(&device->ap_bio_cnt) == 0) {
3514 if (!test_and_set_bit(BITMAP_IO_QUEUED, &device->flags))
3515 drbd_queue_work(&first_peer_device(device)->connection->sender_work,
3516 &device->bm_io_work.w);
3517 }
3518 spin_unlock_irq(&device->resource->req_lock);
3519 }
3520
3521 /**
3522 * drbd_bitmap_io() - Does an IO operation on the whole bitmap
3523 * @device: DRBD device.
3524 * @io_fn: IO callback to be called when bitmap IO is possible
3525 * @why: Descriptive text of the reason for doing the IO
3526 * @flags: Bitmap flags
3527 *
3528 * freezes application IO while that the actual IO operations runs. This
3529 * functions MAY NOT be called from worker context.
3530 */
drbd_bitmap_io(struct drbd_device * device,int (* io_fn)(struct drbd_device *),char * why,enum bm_flag flags)3531 int drbd_bitmap_io(struct drbd_device *device, int (*io_fn)(struct drbd_device *),
3532 char *why, enum bm_flag flags)
3533 {
3534 /* Only suspend io, if some operation is supposed to be locked out */
3535 const bool do_suspend_io = flags & (BM_DONT_CLEAR|BM_DONT_SET|BM_DONT_TEST);
3536 int rv;
3537
3538 D_ASSERT(device, current != first_peer_device(device)->connection->worker.task);
3539
3540 if (do_suspend_io)
3541 drbd_suspend_io(device);
3542
3543 drbd_bm_lock(device, why, flags);
3544 rv = io_fn(device);
3545 drbd_bm_unlock(device);
3546
3547 if (do_suspend_io)
3548 drbd_resume_io(device);
3549
3550 return rv;
3551 }
3552
drbd_md_set_flag(struct drbd_device * device,int flag)3553 void drbd_md_set_flag(struct drbd_device *device, int flag) __must_hold(local)
3554 {
3555 if ((device->ldev->md.flags & flag) != flag) {
3556 drbd_md_mark_dirty(device);
3557 device->ldev->md.flags |= flag;
3558 }
3559 }
3560
drbd_md_clear_flag(struct drbd_device * device,int flag)3561 void drbd_md_clear_flag(struct drbd_device *device, int flag) __must_hold(local)
3562 {
3563 if ((device->ldev->md.flags & flag) != 0) {
3564 drbd_md_mark_dirty(device);
3565 device->ldev->md.flags &= ~flag;
3566 }
3567 }
drbd_md_test_flag(struct drbd_backing_dev * bdev,int flag)3568 int drbd_md_test_flag(struct drbd_backing_dev *bdev, int flag)
3569 {
3570 return (bdev->md.flags & flag) != 0;
3571 }
3572
md_sync_timer_fn(struct timer_list * t)3573 static void md_sync_timer_fn(struct timer_list *t)
3574 {
3575 struct drbd_device *device = from_timer(device, t, md_sync_timer);
3576 drbd_device_post_work(device, MD_SYNC);
3577 }
3578
cmdname(enum drbd_packet cmd)3579 const char *cmdname(enum drbd_packet cmd)
3580 {
3581 /* THINK may need to become several global tables
3582 * when we want to support more than
3583 * one PRO_VERSION */
3584 static const char *cmdnames[] = {
3585
3586 [P_DATA] = "Data",
3587 [P_DATA_REPLY] = "DataReply",
3588 [P_RS_DATA_REPLY] = "RSDataReply",
3589 [P_BARRIER] = "Barrier",
3590 [P_BITMAP] = "ReportBitMap",
3591 [P_BECOME_SYNC_TARGET] = "BecomeSyncTarget",
3592 [P_BECOME_SYNC_SOURCE] = "BecomeSyncSource",
3593 [P_UNPLUG_REMOTE] = "UnplugRemote",
3594 [P_DATA_REQUEST] = "DataRequest",
3595 [P_RS_DATA_REQUEST] = "RSDataRequest",
3596 [P_SYNC_PARAM] = "SyncParam",
3597 [P_PROTOCOL] = "ReportProtocol",
3598 [P_UUIDS] = "ReportUUIDs",
3599 [P_SIZES] = "ReportSizes",
3600 [P_STATE] = "ReportState",
3601 [P_SYNC_UUID] = "ReportSyncUUID",
3602 [P_AUTH_CHALLENGE] = "AuthChallenge",
3603 [P_AUTH_RESPONSE] = "AuthResponse",
3604 [P_STATE_CHG_REQ] = "StateChgRequest",
3605 [P_PING] = "Ping",
3606 [P_PING_ACK] = "PingAck",
3607 [P_RECV_ACK] = "RecvAck",
3608 [P_WRITE_ACK] = "WriteAck",
3609 [P_RS_WRITE_ACK] = "RSWriteAck",
3610 [P_SUPERSEDED] = "Superseded",
3611 [P_NEG_ACK] = "NegAck",
3612 [P_NEG_DREPLY] = "NegDReply",
3613 [P_NEG_RS_DREPLY] = "NegRSDReply",
3614 [P_BARRIER_ACK] = "BarrierAck",
3615 [P_STATE_CHG_REPLY] = "StateChgReply",
3616 [P_OV_REQUEST] = "OVRequest",
3617 [P_OV_REPLY] = "OVReply",
3618 [P_OV_RESULT] = "OVResult",
3619 [P_CSUM_RS_REQUEST] = "CsumRSRequest",
3620 [P_RS_IS_IN_SYNC] = "CsumRSIsInSync",
3621 [P_SYNC_PARAM89] = "SyncParam89",
3622 [P_COMPRESSED_BITMAP] = "CBitmap",
3623 [P_DELAY_PROBE] = "DelayProbe",
3624 [P_OUT_OF_SYNC] = "OutOfSync",
3625 [P_RS_CANCEL] = "RSCancel",
3626 [P_CONN_ST_CHG_REQ] = "conn_st_chg_req",
3627 [P_CONN_ST_CHG_REPLY] = "conn_st_chg_reply",
3628 [P_PROTOCOL_UPDATE] = "protocol_update",
3629 [P_TRIM] = "Trim",
3630 [P_RS_THIN_REQ] = "rs_thin_req",
3631 [P_RS_DEALLOCATED] = "rs_deallocated",
3632 [P_WSAME] = "WriteSame",
3633 [P_ZEROES] = "Zeroes",
3634
3635 /* enum drbd_packet, but not commands - obsoleted flags:
3636 * P_MAY_IGNORE
3637 * P_MAX_OPT_CMD
3638 */
3639 };
3640
3641 /* too big for the array: 0xfffX */
3642 if (cmd == P_INITIAL_META)
3643 return "InitialMeta";
3644 if (cmd == P_INITIAL_DATA)
3645 return "InitialData";
3646 if (cmd == P_CONNECTION_FEATURES)
3647 return "ConnectionFeatures";
3648 if (cmd >= ARRAY_SIZE(cmdnames))
3649 return "Unknown";
3650 return cmdnames[cmd];
3651 }
3652
3653 /**
3654 * drbd_wait_misc - wait for a request to make progress
3655 * @device: device associated with the request
3656 * @i: the struct drbd_interval embedded in struct drbd_request or
3657 * struct drbd_peer_request
3658 */
drbd_wait_misc(struct drbd_device * device,struct drbd_interval * i)3659 int drbd_wait_misc(struct drbd_device *device, struct drbd_interval *i)
3660 {
3661 struct net_conf *nc;
3662 DEFINE_WAIT(wait);
3663 long timeout;
3664
3665 rcu_read_lock();
3666 nc = rcu_dereference(first_peer_device(device)->connection->net_conf);
3667 if (!nc) {
3668 rcu_read_unlock();
3669 return -ETIMEDOUT;
3670 }
3671 timeout = nc->ko_count ? nc->timeout * HZ / 10 * nc->ko_count : MAX_SCHEDULE_TIMEOUT;
3672 rcu_read_unlock();
3673
3674 /* Indicate to wake up device->misc_wait on progress. */
3675 i->waiting = true;
3676 prepare_to_wait(&device->misc_wait, &wait, TASK_INTERRUPTIBLE);
3677 spin_unlock_irq(&device->resource->req_lock);
3678 timeout = schedule_timeout(timeout);
3679 finish_wait(&device->misc_wait, &wait);
3680 spin_lock_irq(&device->resource->req_lock);
3681 if (!timeout || device->state.conn < C_CONNECTED)
3682 return -ETIMEDOUT;
3683 if (signal_pending(current))
3684 return -ERESTARTSYS;
3685 return 0;
3686 }
3687
lock_all_resources(void)3688 void lock_all_resources(void)
3689 {
3690 struct drbd_resource *resource;
3691 int __maybe_unused i = 0;
3692
3693 mutex_lock(&resources_mutex);
3694 local_irq_disable();
3695 for_each_resource(resource, &drbd_resources)
3696 spin_lock_nested(&resource->req_lock, i++);
3697 }
3698
unlock_all_resources(void)3699 void unlock_all_resources(void)
3700 {
3701 struct drbd_resource *resource;
3702
3703 for_each_resource(resource, &drbd_resources)
3704 spin_unlock(&resource->req_lock);
3705 local_irq_enable();
3706 mutex_unlock(&resources_mutex);
3707 }
3708
3709 #ifdef CONFIG_DRBD_FAULT_INJECTION
3710 /* Fault insertion support including random number generator shamelessly
3711 * stolen from kernel/rcutorture.c */
3712 struct fault_random_state {
3713 unsigned long state;
3714 unsigned long count;
3715 };
3716
3717 #define FAULT_RANDOM_MULT 39916801 /* prime */
3718 #define FAULT_RANDOM_ADD 479001701 /* prime */
3719 #define FAULT_RANDOM_REFRESH 10000
3720
3721 /*
3722 * Crude but fast random-number generator. Uses a linear congruential
3723 * generator, with occasional help from get_random_bytes().
3724 */
3725 static unsigned long
_drbd_fault_random(struct fault_random_state * rsp)3726 _drbd_fault_random(struct fault_random_state *rsp)
3727 {
3728 long refresh;
3729
3730 if (!rsp->count--) {
3731 get_random_bytes(&refresh, sizeof(refresh));
3732 rsp->state += refresh;
3733 rsp->count = FAULT_RANDOM_REFRESH;
3734 }
3735 rsp->state = rsp->state * FAULT_RANDOM_MULT + FAULT_RANDOM_ADD;
3736 return swahw32(rsp->state);
3737 }
3738
3739 static char *
_drbd_fault_str(unsigned int type)3740 _drbd_fault_str(unsigned int type) {
3741 static char *_faults[] = {
3742 [DRBD_FAULT_MD_WR] = "Meta-data write",
3743 [DRBD_FAULT_MD_RD] = "Meta-data read",
3744 [DRBD_FAULT_RS_WR] = "Resync write",
3745 [DRBD_FAULT_RS_RD] = "Resync read",
3746 [DRBD_FAULT_DT_WR] = "Data write",
3747 [DRBD_FAULT_DT_RD] = "Data read",
3748 [DRBD_FAULT_DT_RA] = "Data read ahead",
3749 [DRBD_FAULT_BM_ALLOC] = "BM allocation",
3750 [DRBD_FAULT_AL_EE] = "EE allocation",
3751 [DRBD_FAULT_RECEIVE] = "receive data corruption",
3752 };
3753
3754 return (type < DRBD_FAULT_MAX) ? _faults[type] : "**Unknown**";
3755 }
3756
3757 unsigned int
_drbd_insert_fault(struct drbd_device * device,unsigned int type)3758 _drbd_insert_fault(struct drbd_device *device, unsigned int type)
3759 {
3760 static struct fault_random_state rrs = {0, 0};
3761
3762 unsigned int ret = (
3763 (drbd_fault_devs == 0 ||
3764 ((1 << device_to_minor(device)) & drbd_fault_devs) != 0) &&
3765 (((_drbd_fault_random(&rrs) % 100) + 1) <= drbd_fault_rate));
3766
3767 if (ret) {
3768 drbd_fault_count++;
3769
3770 if (__ratelimit(&drbd_ratelimit_state))
3771 drbd_warn(device, "***Simulating %s failure\n",
3772 _drbd_fault_str(type));
3773 }
3774
3775 return ret;
3776 }
3777 #endif
3778
drbd_buildtag(void)3779 const char *drbd_buildtag(void)
3780 {
3781 /* DRBD built from external sources has here a reference to the
3782 git hash of the source code. */
3783
3784 static char buildtag[38] = "\0uilt-in";
3785
3786 if (buildtag[0] == 0) {
3787 #ifdef MODULE
3788 sprintf(buildtag, "srcversion: %-24s", THIS_MODULE->srcversion);
3789 #else
3790 buildtag[0] = 'b';
3791 #endif
3792 }
3793
3794 return buildtag;
3795 }
3796
3797 module_init(drbd_init)
3798 module_exit(drbd_cleanup)
3799
3800 EXPORT_SYMBOL(drbd_conn_str);
3801 EXPORT_SYMBOL(drbd_role_str);
3802 EXPORT_SYMBOL(drbd_disk_str);
3803 EXPORT_SYMBOL(drbd_set_st_err_str);
3804