]>
Commit | Line | Data |
---|---|---|
b411b363 PR |
1 | /* |
2 | drbd_worker.c | |
3 | ||
4 | This file is part of DRBD by Philipp Reisner and Lars Ellenberg. | |
5 | ||
6 | Copyright (C) 2001-2008, LINBIT Information Technologies GmbH. | |
7 | Copyright (C) 1999-2008, Philipp Reisner <philipp.reisner@linbit.com>. | |
8 | Copyright (C) 2002-2008, Lars Ellenberg <lars.ellenberg@linbit.com>. | |
9 | ||
10 | drbd is free software; you can redistribute it and/or modify | |
11 | it under the terms of the GNU General Public License as published by | |
12 | the Free Software Foundation; either version 2, or (at your option) | |
13 | any later version. | |
14 | ||
15 | drbd is distributed in the hope that it will be useful, | |
16 | but WITHOUT ANY WARRANTY; without even the implied warranty of | |
17 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
18 | GNU General Public License for more details. | |
19 | ||
20 | You should have received a copy of the GNU General Public License | |
21 | along with drbd; see the file COPYING. If not, write to | |
22 | the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. | |
23 | ||
84b8c06b | 24 | */ |
b411b363 | 25 | |
b411b363 | 26 | #include <linux/module.h> |
b411b363 PR |
27 | #include <linux/drbd.h> |
28 | #include <linux/sched.h> | |
b411b363 PR |
29 | #include <linux/wait.h> |
30 | #include <linux/mm.h> | |
31 | #include <linux/memcontrol.h> | |
32 | #include <linux/mm_inline.h> | |
33 | #include <linux/slab.h> | |
34 | #include <linux/random.h> | |
b411b363 PR |
35 | #include <linux/string.h> |
36 | #include <linux/scatterlist.h> | |
37 | ||
38 | #include "drbd_int.h" | |
a3603a6e | 39 | #include "drbd_protocol.h" |
b411b363 | 40 | #include "drbd_req.h" |
b411b363 | 41 | |
d448a2e1 AG |
42 | static int make_ov_request(struct drbd_device *, int); |
43 | static int make_resync_request(struct drbd_device *, int); | |
b411b363 | 44 | |
c5a91619 | 45 | /* endio handlers: |
ed15b795 | 46 | * drbd_md_endio (defined here) |
fcefa62e AG |
47 | * drbd_request_endio (defined here) |
48 | * drbd_peer_request_endio (defined here) | |
ed15b795 | 49 | * drbd_bm_endio (defined in drbd_bitmap.c) |
c5a91619 | 50 | * |
b411b363 PR |
51 | * For all these callbacks, note the following: |
52 | * The callbacks will be called in irq context by the IDE drivers, | |
53 | * and in Softirqs/Tasklets/BH context by the SCSI drivers. | |
54 | * Try to get the locking right :) | |
55 | * | |
56 | */ | |
57 | ||
58 | ||
59 | /* About the global_state_lock | |
60 | Each state transition on an device holds a read lock. In case we have | |
95f8efd0 | 61 | to evaluate the resync after dependencies, we grab a write lock, because |
b411b363 PR |
62 | we need stable states on all devices for that. */ |
63 | rwlock_t global_state_lock; | |
64 | ||
65 | /* used for synchronous meta data and bitmap IO | |
66 | * submitted by drbd_md_sync_page_io() | |
67 | */ | |
ed15b795 | 68 | void drbd_md_endio(struct bio *bio, int error) |
b411b363 | 69 | { |
b30ab791 | 70 | struct drbd_device *device; |
b411b363 | 71 | |
e37d2438 LE |
72 | device = bio->bi_private; |
73 | device->md_io.error = error; | |
b411b363 | 74 | |
0cfac5dd PR |
75 | /* We grabbed an extra reference in _drbd_md_sync_page_io() to be able |
76 | * to timeout on the lower level device, and eventually detach from it. | |
77 | * If this io completion runs after that timeout expired, this | |
78 | * drbd_md_put_buffer() may allow us to finally try and re-attach. | |
79 | * During normal operation, this only puts that extra reference | |
80 | * down to 1 again. | |
81 | * Make sure we first drop the reference, and only then signal | |
82 | * completion, or we may (in drbd_al_read_log()) cycle so fast into the | |
83 | * next drbd_md_sync_page_io(), that we trigger the | |
b30ab791 | 84 | * ASSERT(atomic_read(&device->md_io_in_use) == 1) there. |
0cfac5dd | 85 | */ |
b30ab791 | 86 | drbd_md_put_buffer(device); |
e37d2438 | 87 | device->md_io.done = 1; |
b30ab791 | 88 | wake_up(&device->misc_wait); |
cdfda633 | 89 | bio_put(bio); |
b30ab791 AG |
90 | if (device->ldev) /* special case: drbd_md_read() during drbd_adm_attach() */ |
91 | put_ldev(device); | |
b411b363 PR |
92 | } |
93 | ||
94 | /* reads on behalf of the partner, | |
95 | * "submitted" by the receiver | |
96 | */ | |
a186e478 | 97 | static void drbd_endio_read_sec_final(struct drbd_peer_request *peer_req) __releases(local) |
b411b363 PR |
98 | { |
99 | unsigned long flags = 0; | |
6780139c AG |
100 | struct drbd_peer_device *peer_device = peer_req->peer_device; |
101 | struct drbd_device *device = peer_device->device; | |
b411b363 | 102 | |
0500813f | 103 | spin_lock_irqsave(&device->resource->req_lock, flags); |
b30ab791 | 104 | device->read_cnt += peer_req->i.size >> 9; |
a8cd15ba | 105 | list_del(&peer_req->w.list); |
b30ab791 AG |
106 | if (list_empty(&device->read_ee)) |
107 | wake_up(&device->ee_wait); | |
db830c46 | 108 | if (test_bit(__EE_WAS_ERROR, &peer_req->flags)) |
b30ab791 | 109 | __drbd_chk_io_error(device, DRBD_READ_ERROR); |
0500813f | 110 | spin_unlock_irqrestore(&device->resource->req_lock, flags); |
b411b363 | 111 | |
6780139c | 112 | drbd_queue_work(&peer_device->connection->sender_work, &peer_req->w); |
b30ab791 | 113 | put_ldev(device); |
b411b363 PR |
114 | } |
115 | ||
116 | /* writes on behalf of the partner, or resync writes, | |
45bb912b | 117 | * "submitted" by the receiver, final stage. */ |
a0fb3c47 | 118 | void drbd_endio_write_sec_final(struct drbd_peer_request *peer_req) __releases(local) |
b411b363 PR |
119 | { |
120 | unsigned long flags = 0; | |
6780139c AG |
121 | struct drbd_peer_device *peer_device = peer_req->peer_device; |
122 | struct drbd_device *device = peer_device->device; | |
181286ad | 123 | struct drbd_interval i; |
b411b363 | 124 | int do_wake; |
579b57ed | 125 | u64 block_id; |
b411b363 | 126 | int do_al_complete_io; |
b411b363 | 127 | |
db830c46 | 128 | /* after we moved peer_req to done_ee, |
b411b363 PR |
129 | * we may no longer access it, |
130 | * it may be freed/reused already! | |
131 | * (as soon as we release the req_lock) */ | |
181286ad | 132 | i = peer_req->i; |
db830c46 AG |
133 | do_al_complete_io = peer_req->flags & EE_CALL_AL_COMPLETE_IO; |
134 | block_id = peer_req->block_id; | |
21ae5d7f | 135 | peer_req->flags &= ~EE_CALL_AL_COMPLETE_IO; |
b411b363 | 136 | |
0500813f | 137 | spin_lock_irqsave(&device->resource->req_lock, flags); |
b30ab791 | 138 | device->writ_cnt += peer_req->i.size >> 9; |
a8cd15ba | 139 | list_move_tail(&peer_req->w.list, &device->done_ee); |
b411b363 | 140 | |
bb3bfe96 | 141 | /* |
5e472264 | 142 | * Do not remove from the write_requests tree here: we did not send the |
bb3bfe96 AG |
143 | * Ack yet and did not wake possibly waiting conflicting requests. |
144 | * Removed from the tree from "drbd_process_done_ee" within the | |
84b8c06b | 145 | * appropriate dw.cb (e_end_block/e_end_resync_block) or from |
bb3bfe96 AG |
146 | * _drbd_clear_done_ee. |
147 | */ | |
b411b363 | 148 | |
b30ab791 | 149 | do_wake = list_empty(block_id == ID_SYNCER ? &device->sync_ee : &device->active_ee); |
b411b363 | 150 | |
a0fb3c47 LE |
151 | /* FIXME do we want to detach for failed REQ_DISCARD? |
152 | * ((peer_req->flags & (EE_WAS_ERROR|EE_IS_TRIM)) == EE_WAS_ERROR) */ | |
153 | if (peer_req->flags & EE_WAS_ERROR) | |
b30ab791 | 154 | __drbd_chk_io_error(device, DRBD_WRITE_ERROR); |
0500813f | 155 | spin_unlock_irqrestore(&device->resource->req_lock, flags); |
b411b363 | 156 | |
579b57ed | 157 | if (block_id == ID_SYNCER) |
b30ab791 | 158 | drbd_rs_complete_io(device, i.sector); |
b411b363 PR |
159 | |
160 | if (do_wake) | |
b30ab791 | 161 | wake_up(&device->ee_wait); |
b411b363 PR |
162 | |
163 | if (do_al_complete_io) | |
b30ab791 | 164 | drbd_al_complete_io(device, &i); |
b411b363 | 165 | |
6780139c | 166 | wake_asender(peer_device->connection); |
b30ab791 | 167 | put_ldev(device); |
45bb912b | 168 | } |
b411b363 | 169 | |
45bb912b LE |
170 | /* writes on behalf of the partner, or resync writes, |
171 | * "submitted" by the receiver. | |
172 | */ | |
fcefa62e | 173 | void drbd_peer_request_endio(struct bio *bio, int error) |
45bb912b | 174 | { |
db830c46 | 175 | struct drbd_peer_request *peer_req = bio->bi_private; |
a8cd15ba | 176 | struct drbd_device *device = peer_req->peer_device->device; |
45bb912b LE |
177 | int uptodate = bio_flagged(bio, BIO_UPTODATE); |
178 | int is_write = bio_data_dir(bio) == WRITE; | |
a0fb3c47 | 179 | int is_discard = !!(bio->bi_rw & REQ_DISCARD); |
45bb912b | 180 | |
07194272 | 181 | if (error && __ratelimit(&drbd_ratelimit_state)) |
d0180171 | 182 | drbd_warn(device, "%s: error=%d s=%llus\n", |
a0fb3c47 LE |
183 | is_write ? (is_discard ? "discard" : "write") |
184 | : "read", error, | |
db830c46 | 185 | (unsigned long long)peer_req->i.sector); |
45bb912b | 186 | if (!error && !uptodate) { |
07194272 | 187 | if (__ratelimit(&drbd_ratelimit_state)) |
d0180171 | 188 | drbd_warn(device, "%s: setting error to -EIO s=%llus\n", |
07194272 | 189 | is_write ? "write" : "read", |
db830c46 | 190 | (unsigned long long)peer_req->i.sector); |
45bb912b LE |
191 | /* strange behavior of some lower level drivers... |
192 | * fail the request by clearing the uptodate flag, | |
193 | * but do not return any error?! */ | |
194 | error = -EIO; | |
195 | } | |
196 | ||
197 | if (error) | |
db830c46 | 198 | set_bit(__EE_WAS_ERROR, &peer_req->flags); |
45bb912b LE |
199 | |
200 | bio_put(bio); /* no need for the bio anymore */ | |
db830c46 | 201 | if (atomic_dec_and_test(&peer_req->pending_bios)) { |
45bb912b | 202 | if (is_write) |
db830c46 | 203 | drbd_endio_write_sec_final(peer_req); |
45bb912b | 204 | else |
db830c46 | 205 | drbd_endio_read_sec_final(peer_req); |
45bb912b | 206 | } |
b411b363 PR |
207 | } |
208 | ||
209 | /* read, readA or write requests on R_PRIMARY coming from drbd_make_request | |
210 | */ | |
fcefa62e | 211 | void drbd_request_endio(struct bio *bio, int error) |
b411b363 | 212 | { |
a115413d | 213 | unsigned long flags; |
b411b363 | 214 | struct drbd_request *req = bio->bi_private; |
84b8c06b | 215 | struct drbd_device *device = req->device; |
a115413d | 216 | struct bio_and_error m; |
b411b363 PR |
217 | enum drbd_req_event what; |
218 | int uptodate = bio_flagged(bio, BIO_UPTODATE); | |
219 | ||
b411b363 | 220 | if (!error && !uptodate) { |
d0180171 | 221 | drbd_warn(device, "p %s: setting error to -EIO\n", |
b411b363 PR |
222 | bio_data_dir(bio) == WRITE ? "write" : "read"); |
223 | /* strange behavior of some lower level drivers... | |
224 | * fail the request by clearing the uptodate flag, | |
225 | * but do not return any error?! */ | |
226 | error = -EIO; | |
227 | } | |
228 | ||
1b6dd252 PR |
229 | |
230 | /* If this request was aborted locally before, | |
231 | * but now was completed "successfully", | |
232 | * chances are that this caused arbitrary data corruption. | |
233 | * | |
234 | * "aborting" requests, or force-detaching the disk, is intended for | |
235 | * completely blocked/hung local backing devices which do no longer | |
236 | * complete requests at all, not even do error completions. In this | |
237 | * situation, usually a hard-reset and failover is the only way out. | |
238 | * | |
239 | * By "aborting", basically faking a local error-completion, | |
240 | * we allow for a more graceful swichover by cleanly migrating services. | |
241 | * Still the affected node has to be rebooted "soon". | |
242 | * | |
243 | * By completing these requests, we allow the upper layers to re-use | |
244 | * the associated data pages. | |
245 | * | |
246 | * If later the local backing device "recovers", and now DMAs some data | |
247 | * from disk into the original request pages, in the best case it will | |
248 | * just put random data into unused pages; but typically it will corrupt | |
249 | * meanwhile completely unrelated data, causing all sorts of damage. | |
250 | * | |
251 | * Which means delayed successful completion, | |
252 | * especially for READ requests, | |
253 | * is a reason to panic(). | |
254 | * | |
255 | * We assume that a delayed *error* completion is OK, | |
256 | * though we still will complain noisily about it. | |
257 | */ | |
258 | if (unlikely(req->rq_state & RQ_LOCAL_ABORTED)) { | |
259 | if (__ratelimit(&drbd_ratelimit_state)) | |
d0180171 | 260 | drbd_emerg(device, "delayed completion of aborted local request; disk-timeout may be too aggressive\n"); |
1b6dd252 PR |
261 | |
262 | if (!error) | |
263 | panic("possible random memory corruption caused by delayed completion of aborted local request\n"); | |
264 | } | |
265 | ||
b411b363 PR |
266 | /* to avoid recursion in __req_mod */ |
267 | if (unlikely(error)) { | |
2f632aeb LE |
268 | if (bio->bi_rw & REQ_DISCARD) |
269 | what = (error == -EOPNOTSUPP) | |
270 | ? DISCARD_COMPLETED_NOTSUPP | |
271 | : DISCARD_COMPLETED_WITH_ERROR; | |
272 | else | |
273 | what = (bio_data_dir(bio) == WRITE) | |
8554df1c | 274 | ? WRITE_COMPLETED_WITH_ERROR |
5c3c7e64 | 275 | : (bio_rw(bio) == READ) |
8554df1c AG |
276 | ? READ_COMPLETED_WITH_ERROR |
277 | : READ_AHEAD_COMPLETED_WITH_ERROR; | |
b411b363 | 278 | } else |
8554df1c | 279 | what = COMPLETED_OK; |
b411b363 PR |
280 | |
281 | bio_put(req->private_bio); | |
282 | req->private_bio = ERR_PTR(error); | |
283 | ||
a115413d | 284 | /* not req_mod(), we need irqsave here! */ |
0500813f | 285 | spin_lock_irqsave(&device->resource->req_lock, flags); |
a115413d | 286 | __req_mod(req, what, &m); |
0500813f | 287 | spin_unlock_irqrestore(&device->resource->req_lock, flags); |
b30ab791 | 288 | put_ldev(device); |
a115413d LE |
289 | |
290 | if (m.bio) | |
b30ab791 | 291 | complete_master_bio(device, &m); |
b411b363 PR |
292 | } |
293 | ||
79a3c8d3 | 294 | void drbd_csum_ee(struct crypto_hash *tfm, struct drbd_peer_request *peer_req, void *digest) |
45bb912b LE |
295 | { |
296 | struct hash_desc desc; | |
297 | struct scatterlist sg; | |
db830c46 | 298 | struct page *page = peer_req->pages; |
45bb912b LE |
299 | struct page *tmp; |
300 | unsigned len; | |
301 | ||
302 | desc.tfm = tfm; | |
303 | desc.flags = 0; | |
304 | ||
305 | sg_init_table(&sg, 1); | |
306 | crypto_hash_init(&desc); | |
307 | ||
308 | while ((tmp = page_chain_next(page))) { | |
309 | /* all but the last page will be fully used */ | |
310 | sg_set_page(&sg, page, PAGE_SIZE, 0); | |
311 | crypto_hash_update(&desc, &sg, sg.length); | |
312 | page = tmp; | |
313 | } | |
314 | /* and now the last, possibly only partially used page */ | |
db830c46 | 315 | len = peer_req->i.size & (PAGE_SIZE - 1); |
45bb912b LE |
316 | sg_set_page(&sg, page, len ?: PAGE_SIZE, 0); |
317 | crypto_hash_update(&desc, &sg, sg.length); | |
318 | crypto_hash_final(&desc, digest); | |
319 | } | |
320 | ||
79a3c8d3 | 321 | void drbd_csum_bio(struct crypto_hash *tfm, struct bio *bio, void *digest) |
b411b363 PR |
322 | { |
323 | struct hash_desc desc; | |
324 | struct scatterlist sg; | |
7988613b KO |
325 | struct bio_vec bvec; |
326 | struct bvec_iter iter; | |
b411b363 PR |
327 | |
328 | desc.tfm = tfm; | |
329 | desc.flags = 0; | |
330 | ||
331 | sg_init_table(&sg, 1); | |
332 | crypto_hash_init(&desc); | |
333 | ||
7988613b KO |
334 | bio_for_each_segment(bvec, bio, iter) { |
335 | sg_set_page(&sg, bvec.bv_page, bvec.bv_len, bvec.bv_offset); | |
b411b363 PR |
336 | crypto_hash_update(&desc, &sg, sg.length); |
337 | } | |
338 | crypto_hash_final(&desc, digest); | |
339 | } | |
340 | ||
9676c760 | 341 | /* MAYBE merge common code with w_e_end_ov_req */ |
99920dc5 | 342 | static int w_e_send_csum(struct drbd_work *w, int cancel) |
b411b363 | 343 | { |
a8cd15ba | 344 | struct drbd_peer_request *peer_req = container_of(w, struct drbd_peer_request, w); |
6780139c AG |
345 | struct drbd_peer_device *peer_device = peer_req->peer_device; |
346 | struct drbd_device *device = peer_device->device; | |
b411b363 PR |
347 | int digest_size; |
348 | void *digest; | |
99920dc5 | 349 | int err = 0; |
b411b363 | 350 | |
53ea4331 LE |
351 | if (unlikely(cancel)) |
352 | goto out; | |
b411b363 | 353 | |
9676c760 | 354 | if (unlikely((peer_req->flags & EE_WAS_ERROR) != 0)) |
53ea4331 | 355 | goto out; |
b411b363 | 356 | |
6780139c | 357 | digest_size = crypto_hash_digestsize(peer_device->connection->csums_tfm); |
53ea4331 LE |
358 | digest = kmalloc(digest_size, GFP_NOIO); |
359 | if (digest) { | |
db830c46 AG |
360 | sector_t sector = peer_req->i.sector; |
361 | unsigned int size = peer_req->i.size; | |
6780139c | 362 | drbd_csum_ee(peer_device->connection->csums_tfm, peer_req, digest); |
9676c760 | 363 | /* Free peer_req and pages before send. |
53ea4331 LE |
364 | * In case we block on congestion, we could otherwise run into |
365 | * some distributed deadlock, if the other side blocks on | |
366 | * congestion as well, because our receiver blocks in | |
c37c8ecf | 367 | * drbd_alloc_pages due to pp_in_use > max_buffers. */ |
b30ab791 | 368 | drbd_free_peer_req(device, peer_req); |
db830c46 | 369 | peer_req = NULL; |
b30ab791 | 370 | inc_rs_pending(device); |
6780139c | 371 | err = drbd_send_drequest_csum(peer_device, sector, size, |
db1b0b72 AG |
372 | digest, digest_size, |
373 | P_CSUM_RS_REQUEST); | |
53ea4331 LE |
374 | kfree(digest); |
375 | } else { | |
d0180171 | 376 | drbd_err(device, "kmalloc() of digest failed.\n"); |
99920dc5 | 377 | err = -ENOMEM; |
53ea4331 | 378 | } |
b411b363 | 379 | |
53ea4331 | 380 | out: |
db830c46 | 381 | if (peer_req) |
b30ab791 | 382 | drbd_free_peer_req(device, peer_req); |
b411b363 | 383 | |
99920dc5 | 384 | if (unlikely(err)) |
d0180171 | 385 | drbd_err(device, "drbd_send_drequest(..., csum) failed\n"); |
99920dc5 | 386 | return err; |
b411b363 PR |
387 | } |
388 | ||
389 | #define GFP_TRY (__GFP_HIGHMEM | __GFP_NOWARN) | |
390 | ||
69a22773 | 391 | static int read_for_csum(struct drbd_peer_device *peer_device, sector_t sector, int size) |
b411b363 | 392 | { |
69a22773 | 393 | struct drbd_device *device = peer_device->device; |
db830c46 | 394 | struct drbd_peer_request *peer_req; |
b411b363 | 395 | |
b30ab791 | 396 | if (!get_ldev(device)) |
80a40e43 | 397 | return -EIO; |
b411b363 PR |
398 | |
399 | /* GFP_TRY, because if there is no memory available right now, this may | |
400 | * be rescheduled for later. It is "only" background resync, after all. */ | |
69a22773 | 401 | peer_req = drbd_alloc_peer_req(peer_device, ID_SYNCER /* unused */, sector, |
a0fb3c47 | 402 | size, true /* has real payload */, GFP_TRY); |
db830c46 | 403 | if (!peer_req) |
80a40e43 | 404 | goto defer; |
b411b363 | 405 | |
a8cd15ba | 406 | peer_req->w.cb = w_e_send_csum; |
0500813f | 407 | spin_lock_irq(&device->resource->req_lock); |
b9ed7080 | 408 | list_add_tail(&peer_req->w.list, &device->read_ee); |
0500813f | 409 | spin_unlock_irq(&device->resource->req_lock); |
b411b363 | 410 | |
b30ab791 AG |
411 | atomic_add(size >> 9, &device->rs_sect_ev); |
412 | if (drbd_submit_peer_request(device, peer_req, READ, DRBD_FAULT_RS_RD) == 0) | |
80a40e43 | 413 | return 0; |
b411b363 | 414 | |
10f6d992 LE |
415 | /* If it failed because of ENOMEM, retry should help. If it failed |
416 | * because bio_add_page failed (probably broken lower level driver), | |
417 | * retry may or may not help. | |
418 | * If it does not, you may need to force disconnect. */ | |
0500813f | 419 | spin_lock_irq(&device->resource->req_lock); |
a8cd15ba | 420 | list_del(&peer_req->w.list); |
0500813f | 421 | spin_unlock_irq(&device->resource->req_lock); |
22cc37a9 | 422 | |
b30ab791 | 423 | drbd_free_peer_req(device, peer_req); |
80a40e43 | 424 | defer: |
b30ab791 | 425 | put_ldev(device); |
80a40e43 | 426 | return -EAGAIN; |
b411b363 PR |
427 | } |
428 | ||
99920dc5 | 429 | int w_resync_timer(struct drbd_work *w, int cancel) |
b411b363 | 430 | { |
84b8c06b AG |
431 | struct drbd_device *device = |
432 | container_of(w, struct drbd_device, resync_work); | |
433 | ||
b30ab791 | 434 | switch (device->state.conn) { |
63106d3c | 435 | case C_VERIFY_S: |
d448a2e1 | 436 | make_ov_request(device, cancel); |
63106d3c PR |
437 | break; |
438 | case C_SYNC_TARGET: | |
d448a2e1 | 439 | make_resync_request(device, cancel); |
63106d3c | 440 | break; |
b411b363 PR |
441 | } |
442 | ||
99920dc5 | 443 | return 0; |
794abb75 PR |
444 | } |
445 | ||
446 | void resync_timer_fn(unsigned long data) | |
447 | { | |
b30ab791 | 448 | struct drbd_device *device = (struct drbd_device *) data; |
794abb75 | 449 | |
15e26f6a LE |
450 | drbd_queue_work_if_unqueued( |
451 | &first_peer_device(device)->connection->sender_work, | |
452 | &device->resync_work); | |
b411b363 PR |
453 | } |
454 | ||
778f271d PR |
455 | static void fifo_set(struct fifo_buffer *fb, int value) |
456 | { | |
457 | int i; | |
458 | ||
459 | for (i = 0; i < fb->size; i++) | |
f10f2623 | 460 | fb->values[i] = value; |
778f271d PR |
461 | } |
462 | ||
463 | static int fifo_push(struct fifo_buffer *fb, int value) | |
464 | { | |
465 | int ov; | |
466 | ||
467 | ov = fb->values[fb->head_index]; | |
468 | fb->values[fb->head_index++] = value; | |
469 | ||
470 | if (fb->head_index >= fb->size) | |
471 | fb->head_index = 0; | |
472 | ||
473 | return ov; | |
474 | } | |
475 | ||
476 | static void fifo_add_val(struct fifo_buffer *fb, int value) | |
477 | { | |
478 | int i; | |
479 | ||
480 | for (i = 0; i < fb->size; i++) | |
481 | fb->values[i] += value; | |
482 | } | |
483 | ||
9958c857 PR |
484 | struct fifo_buffer *fifo_alloc(int fifo_size) |
485 | { | |
486 | struct fifo_buffer *fb; | |
487 | ||
8747d30a | 488 | fb = kzalloc(sizeof(struct fifo_buffer) + sizeof(int) * fifo_size, GFP_NOIO); |
9958c857 PR |
489 | if (!fb) |
490 | return NULL; | |
491 | ||
492 | fb->head_index = 0; | |
493 | fb->size = fifo_size; | |
494 | fb->total = 0; | |
495 | ||
496 | return fb; | |
497 | } | |
498 | ||
0e49d7b0 | 499 | static int drbd_rs_controller(struct drbd_device *device, unsigned int sect_in) |
778f271d | 500 | { |
daeda1cc | 501 | struct disk_conf *dc; |
7f34f614 | 502 | unsigned int want; /* The number of sectors we want in-flight */ |
778f271d | 503 | int req_sect; /* Number of sectors to request in this turn */ |
7f34f614 | 504 | int correction; /* Number of sectors more we need in-flight */ |
778f271d PR |
505 | int cps; /* correction per invocation of drbd_rs_controller() */ |
506 | int steps; /* Number of time steps to plan ahead */ | |
507 | int curr_corr; | |
508 | int max_sect; | |
813472ce | 509 | struct fifo_buffer *plan; |
778f271d | 510 | |
b30ab791 AG |
511 | dc = rcu_dereference(device->ldev->disk_conf); |
512 | plan = rcu_dereference(device->rs_plan_s); | |
778f271d | 513 | |
813472ce | 514 | steps = plan->size; /* (dc->c_plan_ahead * 10 * SLEEP_TIME) / HZ; */ |
778f271d | 515 | |
b30ab791 | 516 | if (device->rs_in_flight + sect_in == 0) { /* At start of resync */ |
daeda1cc | 517 | want = ((dc->resync_rate * 2 * SLEEP_TIME) / HZ) * steps; |
778f271d | 518 | } else { /* normal path */ |
daeda1cc PR |
519 | want = dc->c_fill_target ? dc->c_fill_target : |
520 | sect_in * dc->c_delay_target * HZ / (SLEEP_TIME * 10); | |
778f271d PR |
521 | } |
522 | ||
b30ab791 | 523 | correction = want - device->rs_in_flight - plan->total; |
778f271d PR |
524 | |
525 | /* Plan ahead */ | |
526 | cps = correction / steps; | |
813472ce PR |
527 | fifo_add_val(plan, cps); |
528 | plan->total += cps * steps; | |
778f271d PR |
529 | |
530 | /* What we do in this step */ | |
813472ce PR |
531 | curr_corr = fifo_push(plan, 0); |
532 | plan->total -= curr_corr; | |
778f271d PR |
533 | |
534 | req_sect = sect_in + curr_corr; | |
535 | if (req_sect < 0) | |
536 | req_sect = 0; | |
537 | ||
daeda1cc | 538 | max_sect = (dc->c_max_rate * 2 * SLEEP_TIME) / HZ; |
778f271d PR |
539 | if (req_sect > max_sect) |
540 | req_sect = max_sect; | |
541 | ||
542 | /* | |
d0180171 | 543 | drbd_warn(device, "si=%u if=%d wa=%u co=%d st=%d cps=%d pl=%d cc=%d rs=%d\n", |
b30ab791 AG |
544 | sect_in, device->rs_in_flight, want, correction, |
545 | steps, cps, device->rs_planed, curr_corr, req_sect); | |
778f271d PR |
546 | */ |
547 | ||
548 | return req_sect; | |
549 | } | |
550 | ||
b30ab791 | 551 | static int drbd_rs_number_requests(struct drbd_device *device) |
e65f440d | 552 | { |
0e49d7b0 LE |
553 | unsigned int sect_in; /* Number of sectors that came in since the last turn */ |
554 | int number, mxb; | |
555 | ||
556 | sect_in = atomic_xchg(&device->rs_sect_in, 0); | |
557 | device->rs_in_flight -= sect_in; | |
813472ce PR |
558 | |
559 | rcu_read_lock(); | |
0e49d7b0 | 560 | mxb = drbd_get_max_buffers(device) / 2; |
b30ab791 | 561 | if (rcu_dereference(device->rs_plan_s)->size) { |
0e49d7b0 | 562 | number = drbd_rs_controller(device, sect_in) >> (BM_BLOCK_SHIFT - 9); |
b30ab791 | 563 | device->c_sync_rate = number * HZ * (BM_BLOCK_SIZE / 1024) / SLEEP_TIME; |
e65f440d | 564 | } else { |
b30ab791 AG |
565 | device->c_sync_rate = rcu_dereference(device->ldev->disk_conf)->resync_rate; |
566 | number = SLEEP_TIME * device->c_sync_rate / ((BM_BLOCK_SIZE / 1024) * HZ); | |
e65f440d | 567 | } |
813472ce | 568 | rcu_read_unlock(); |
e65f440d | 569 | |
0e49d7b0 LE |
570 | /* Don't have more than "max-buffers"/2 in-flight. |
571 | * Otherwise we may cause the remote site to stall on drbd_alloc_pages(), | |
572 | * potentially causing a distributed deadlock on congestion during | |
573 | * online-verify or (checksum-based) resync, if max-buffers, | |
574 | * socket buffer sizes and resync rate settings are mis-configured. */ | |
7f34f614 LE |
575 | |
576 | /* note that "number" is in units of "BM_BLOCK_SIZE" (which is 4k), | |
577 | * mxb (as used here, and in drbd_alloc_pages on the peer) is | |
578 | * "number of pages" (typically also 4k), | |
579 | * but "rs_in_flight" is in "sectors" (512 Byte). */ | |
580 | if (mxb - device->rs_in_flight/8 < number) | |
581 | number = mxb - device->rs_in_flight/8; | |
0e49d7b0 | 582 | |
e65f440d LE |
583 | return number; |
584 | } | |
585 | ||
44a4d551 | 586 | static int make_resync_request(struct drbd_device *const device, int cancel) |
b411b363 | 587 | { |
44a4d551 LE |
588 | struct drbd_peer_device *const peer_device = first_peer_device(device); |
589 | struct drbd_connection *const connection = peer_device ? peer_device->connection : NULL; | |
b411b363 PR |
590 | unsigned long bit; |
591 | sector_t sector; | |
b30ab791 | 592 | const sector_t capacity = drbd_get_capacity(device->this_bdev); |
1816a2b4 | 593 | int max_bio_size; |
e65f440d | 594 | int number, rollback_i, size; |
506afb62 | 595 | int align, requeue = 0; |
0f0601f4 | 596 | int i = 0; |
b411b363 PR |
597 | |
598 | if (unlikely(cancel)) | |
99920dc5 | 599 | return 0; |
b411b363 | 600 | |
b30ab791 | 601 | if (device->rs_total == 0) { |
af85e8e8 | 602 | /* empty resync? */ |
b30ab791 | 603 | drbd_resync_finished(device); |
99920dc5 | 604 | return 0; |
af85e8e8 LE |
605 | } |
606 | ||
b30ab791 AG |
607 | if (!get_ldev(device)) { |
608 | /* Since we only need to access device->rsync a | |
609 | get_ldev_if_state(device,D_FAILED) would be sufficient, but | |
b411b363 PR |
610 | to continue resync with a broken disk makes no sense at |
611 | all */ | |
d0180171 | 612 | drbd_err(device, "Disk broke down during resync!\n"); |
99920dc5 | 613 | return 0; |
b411b363 PR |
614 | } |
615 | ||
b30ab791 AG |
616 | max_bio_size = queue_max_hw_sectors(device->rq_queue) << 9; |
617 | number = drbd_rs_number_requests(device); | |
0e49d7b0 | 618 | if (number <= 0) |
0f0601f4 | 619 | goto requeue; |
b411b363 | 620 | |
b411b363 | 621 | for (i = 0; i < number; i++) { |
506afb62 LE |
622 | /* Stop generating RS requests when half of the send buffer is filled, |
623 | * but notify TCP that we'd like to have more space. */ | |
44a4d551 LE |
624 | mutex_lock(&connection->data.mutex); |
625 | if (connection->data.socket) { | |
506afb62 LE |
626 | struct sock *sk = connection->data.socket->sk; |
627 | int queued = sk->sk_wmem_queued; | |
628 | int sndbuf = sk->sk_sndbuf; | |
629 | if (queued > sndbuf / 2) { | |
630 | requeue = 1; | |
631 | if (sk->sk_socket) | |
632 | set_bit(SOCK_NOSPACE, &sk->sk_socket->flags); | |
633 | } | |
634 | } else | |
635 | requeue = 1; | |
44a4d551 | 636 | mutex_unlock(&connection->data.mutex); |
506afb62 | 637 | if (requeue) |
b411b363 PR |
638 | goto requeue; |
639 | ||
640 | next_sector: | |
641 | size = BM_BLOCK_SIZE; | |
b30ab791 | 642 | bit = drbd_bm_find_next(device, device->bm_resync_fo); |
b411b363 | 643 | |
4b0715f0 | 644 | if (bit == DRBD_END_OF_BITMAP) { |
b30ab791 AG |
645 | device->bm_resync_fo = drbd_bm_bits(device); |
646 | put_ldev(device); | |
99920dc5 | 647 | return 0; |
b411b363 PR |
648 | } |
649 | ||
650 | sector = BM_BIT_TO_SECT(bit); | |
651 | ||
ad3fee79 | 652 | if (drbd_try_rs_begin_io(device, sector)) { |
b30ab791 | 653 | device->bm_resync_fo = bit; |
b411b363 PR |
654 | goto requeue; |
655 | } | |
b30ab791 | 656 | device->bm_resync_fo = bit + 1; |
b411b363 | 657 | |
b30ab791 AG |
658 | if (unlikely(drbd_bm_test_bit(device, bit) == 0)) { |
659 | drbd_rs_complete_io(device, sector); | |
b411b363 PR |
660 | goto next_sector; |
661 | } | |
662 | ||
1816a2b4 | 663 | #if DRBD_MAX_BIO_SIZE > BM_BLOCK_SIZE |
b411b363 PR |
664 | /* try to find some adjacent bits. |
665 | * we stop if we have already the maximum req size. | |
666 | * | |
667 | * Additionally always align bigger requests, in order to | |
668 | * be prepared for all stripe sizes of software RAIDs. | |
b411b363 PR |
669 | */ |
670 | align = 1; | |
d207450c | 671 | rollback_i = i; |
6377b923 | 672 | while (i < number) { |
1816a2b4 | 673 | if (size + BM_BLOCK_SIZE > max_bio_size) |
b411b363 PR |
674 | break; |
675 | ||
676 | /* Be always aligned */ | |
677 | if (sector & ((1<<(align+3))-1)) | |
678 | break; | |
679 | ||
680 | /* do not cross extent boundaries */ | |
681 | if (((bit+1) & BM_BLOCKS_PER_BM_EXT_MASK) == 0) | |
682 | break; | |
683 | /* now, is it actually dirty, after all? | |
684 | * caution, drbd_bm_test_bit is tri-state for some | |
685 | * obscure reason; ( b == 0 ) would get the out-of-band | |
686 | * only accidentally right because of the "oddly sized" | |
687 | * adjustment below */ | |
b30ab791 | 688 | if (drbd_bm_test_bit(device, bit+1) != 1) |
b411b363 PR |
689 | break; |
690 | bit++; | |
691 | size += BM_BLOCK_SIZE; | |
692 | if ((BM_BLOCK_SIZE << align) <= size) | |
693 | align++; | |
694 | i++; | |
695 | } | |
696 | /* if we merged some, | |
697 | * reset the offset to start the next drbd_bm_find_next from */ | |
698 | if (size > BM_BLOCK_SIZE) | |
b30ab791 | 699 | device->bm_resync_fo = bit + 1; |
b411b363 PR |
700 | #endif |
701 | ||
702 | /* adjust very last sectors, in case we are oddly sized */ | |
703 | if (sector + (size>>9) > capacity) | |
704 | size = (capacity-sector)<<9; | |
aaaba345 LE |
705 | |
706 | if (device->use_csums) { | |
44a4d551 | 707 | switch (read_for_csum(peer_device, sector, size)) { |
80a40e43 | 708 | case -EIO: /* Disk failure */ |
b30ab791 | 709 | put_ldev(device); |
99920dc5 | 710 | return -EIO; |
80a40e43 | 711 | case -EAGAIN: /* allocation failed, or ldev busy */ |
b30ab791 AG |
712 | drbd_rs_complete_io(device, sector); |
713 | device->bm_resync_fo = BM_SECT_TO_BIT(sector); | |
d207450c | 714 | i = rollback_i; |
b411b363 | 715 | goto requeue; |
80a40e43 LE |
716 | case 0: |
717 | /* everything ok */ | |
718 | break; | |
719 | default: | |
720 | BUG(); | |
b411b363 PR |
721 | } |
722 | } else { | |
99920dc5 AG |
723 | int err; |
724 | ||
b30ab791 | 725 | inc_rs_pending(device); |
44a4d551 | 726 | err = drbd_send_drequest(peer_device, P_RS_DATA_REQUEST, |
99920dc5 AG |
727 | sector, size, ID_SYNCER); |
728 | if (err) { | |
d0180171 | 729 | drbd_err(device, "drbd_send_drequest() failed, aborting...\n"); |
b30ab791 AG |
730 | dec_rs_pending(device); |
731 | put_ldev(device); | |
99920dc5 | 732 | return err; |
b411b363 PR |
733 | } |
734 | } | |
735 | } | |
736 | ||
b30ab791 | 737 | if (device->bm_resync_fo >= drbd_bm_bits(device)) { |
b411b363 PR |
738 | /* last syncer _request_ was sent, |
739 | * but the P_RS_DATA_REPLY not yet received. sync will end (and | |
740 | * next sync group will resume), as soon as we receive the last | |
741 | * resync data block, and the last bit is cleared. | |
742 | * until then resync "work" is "inactive" ... | |
743 | */ | |
b30ab791 | 744 | put_ldev(device); |
99920dc5 | 745 | return 0; |
b411b363 PR |
746 | } |
747 | ||
748 | requeue: | |
b30ab791 AG |
749 | device->rs_in_flight += (i << (BM_BLOCK_SHIFT - 9)); |
750 | mod_timer(&device->resync_timer, jiffies + SLEEP_TIME); | |
751 | put_ldev(device); | |
99920dc5 | 752 | return 0; |
b411b363 PR |
753 | } |
754 | ||
d448a2e1 | 755 | static int make_ov_request(struct drbd_device *device, int cancel) |
b411b363 PR |
756 | { |
757 | int number, i, size; | |
758 | sector_t sector; | |
b30ab791 | 759 | const sector_t capacity = drbd_get_capacity(device->this_bdev); |
58ffa580 | 760 | bool stop_sector_reached = false; |
b411b363 PR |
761 | |
762 | if (unlikely(cancel)) | |
763 | return 1; | |
764 | ||
b30ab791 | 765 | number = drbd_rs_number_requests(device); |
b411b363 | 766 | |
b30ab791 | 767 | sector = device->ov_position; |
b411b363 | 768 | for (i = 0; i < number; i++) { |
58ffa580 | 769 | if (sector >= capacity) |
b411b363 | 770 | return 1; |
58ffa580 LE |
771 | |
772 | /* We check for "finished" only in the reply path: | |
773 | * w_e_end_ov_reply(). | |
774 | * We need to send at least one request out. */ | |
775 | stop_sector_reached = i > 0 | |
b30ab791 AG |
776 | && verify_can_do_stop_sector(device) |
777 | && sector >= device->ov_stop_sector; | |
58ffa580 LE |
778 | if (stop_sector_reached) |
779 | break; | |
b411b363 PR |
780 | |
781 | size = BM_BLOCK_SIZE; | |
782 | ||
ad3fee79 | 783 | if (drbd_try_rs_begin_io(device, sector)) { |
b30ab791 | 784 | device->ov_position = sector; |
b411b363 PR |
785 | goto requeue; |
786 | } | |
787 | ||
788 | if (sector + (size>>9) > capacity) | |
789 | size = (capacity-sector)<<9; | |
790 | ||
b30ab791 | 791 | inc_rs_pending(device); |
69a22773 | 792 | if (drbd_send_ov_request(first_peer_device(device), sector, size)) { |
b30ab791 | 793 | dec_rs_pending(device); |
b411b363 PR |
794 | return 0; |
795 | } | |
796 | sector += BM_SECT_PER_BIT; | |
797 | } | |
b30ab791 | 798 | device->ov_position = sector; |
b411b363 PR |
799 | |
800 | requeue: | |
b30ab791 | 801 | device->rs_in_flight += (i << (BM_BLOCK_SHIFT - 9)); |
58ffa580 | 802 | if (i == 0 || !stop_sector_reached) |
b30ab791 | 803 | mod_timer(&device->resync_timer, jiffies + SLEEP_TIME); |
b411b363 PR |
804 | return 1; |
805 | } | |
806 | ||
99920dc5 | 807 | int w_ov_finished(struct drbd_work *w, int cancel) |
b411b363 | 808 | { |
84b8c06b AG |
809 | struct drbd_device_work *dw = |
810 | container_of(w, struct drbd_device_work, w); | |
811 | struct drbd_device *device = dw->device; | |
812 | kfree(dw); | |
b30ab791 AG |
813 | ov_out_of_sync_print(device); |
814 | drbd_resync_finished(device); | |
b411b363 | 815 | |
99920dc5 | 816 | return 0; |
b411b363 PR |
817 | } |
818 | ||
99920dc5 | 819 | static int w_resync_finished(struct drbd_work *w, int cancel) |
b411b363 | 820 | { |
84b8c06b AG |
821 | struct drbd_device_work *dw = |
822 | container_of(w, struct drbd_device_work, w); | |
823 | struct drbd_device *device = dw->device; | |
824 | kfree(dw); | |
b411b363 | 825 | |
b30ab791 | 826 | drbd_resync_finished(device); |
b411b363 | 827 | |
99920dc5 | 828 | return 0; |
b411b363 PR |
829 | } |
830 | ||
b30ab791 | 831 | static void ping_peer(struct drbd_device *device) |
af85e8e8 | 832 | { |
a6b32bc3 | 833 | struct drbd_connection *connection = first_peer_device(device)->connection; |
2a67d8b9 | 834 | |
bde89a9e AG |
835 | clear_bit(GOT_PING_ACK, &connection->flags); |
836 | request_ping(connection); | |
837 | wait_event(connection->ping_wait, | |
838 | test_bit(GOT_PING_ACK, &connection->flags) || device->state.conn < C_CONNECTED); | |
af85e8e8 LE |
839 | } |
840 | ||
b30ab791 | 841 | int drbd_resync_finished(struct drbd_device *device) |
b411b363 PR |
842 | { |
843 | unsigned long db, dt, dbdt; | |
844 | unsigned long n_oos; | |
845 | union drbd_state os, ns; | |
84b8c06b | 846 | struct drbd_device_work *dw; |
b411b363 | 847 | char *khelper_cmd = NULL; |
26525618 | 848 | int verify_done = 0; |
b411b363 PR |
849 | |
850 | /* Remove all elements from the resync LRU. Since future actions | |
851 | * might set bits in the (main) bitmap, then the entries in the | |
852 | * resync LRU would be wrong. */ | |
b30ab791 | 853 | if (drbd_rs_del_all(device)) { |
b411b363 PR |
854 | /* In case this is not possible now, most probably because |
855 | * there are P_RS_DATA_REPLY Packets lingering on the worker's | |
856 | * queue (or even the read operations for those packets | |
857 | * is not finished by now). Retry in 100ms. */ | |
858 | ||
20ee6390 | 859 | schedule_timeout_interruptible(HZ / 10); |
84b8c06b AG |
860 | dw = kmalloc(sizeof(struct drbd_device_work), GFP_ATOMIC); |
861 | if (dw) { | |
862 | dw->w.cb = w_resync_finished; | |
863 | dw->device = device; | |
864 | drbd_queue_work(&first_peer_device(device)->connection->sender_work, | |
865 | &dw->w); | |
b411b363 PR |
866 | return 1; |
867 | } | |
84b8c06b | 868 | drbd_err(device, "Warn failed to drbd_rs_del_all() and to kmalloc(dw).\n"); |
b411b363 PR |
869 | } |
870 | ||
b30ab791 | 871 | dt = (jiffies - device->rs_start - device->rs_paused) / HZ; |
b411b363 PR |
872 | if (dt <= 0) |
873 | dt = 1; | |
84b8c06b | 874 | |
b30ab791 | 875 | db = device->rs_total; |
58ffa580 | 876 | /* adjust for verify start and stop sectors, respective reached position */ |
b30ab791 AG |
877 | if (device->state.conn == C_VERIFY_S || device->state.conn == C_VERIFY_T) |
878 | db -= device->ov_left; | |
58ffa580 | 879 | |
b411b363 | 880 | dbdt = Bit2KB(db/dt); |
b30ab791 | 881 | device->rs_paused /= HZ; |
b411b363 | 882 | |
b30ab791 | 883 | if (!get_ldev(device)) |
b411b363 PR |
884 | goto out; |
885 | ||
b30ab791 | 886 | ping_peer(device); |
af85e8e8 | 887 | |
0500813f | 888 | spin_lock_irq(&device->resource->req_lock); |
b30ab791 | 889 | os = drbd_read_state(device); |
b411b363 | 890 | |
26525618 LE |
891 | verify_done = (os.conn == C_VERIFY_S || os.conn == C_VERIFY_T); |
892 | ||
b411b363 PR |
893 | /* This protects us against multiple calls (that can happen in the presence |
894 | of application IO), and against connectivity loss just before we arrive here. */ | |
895 | if (os.conn <= C_CONNECTED) | |
896 | goto out_unlock; | |
897 | ||
898 | ns = os; | |
899 | ns.conn = C_CONNECTED; | |
900 | ||
d0180171 | 901 | drbd_info(device, "%s done (total %lu sec; paused %lu sec; %lu K/sec)\n", |
58ffa580 | 902 | verify_done ? "Online verify" : "Resync", |
b30ab791 | 903 | dt + device->rs_paused, device->rs_paused, dbdt); |
b411b363 | 904 | |
b30ab791 | 905 | n_oos = drbd_bm_total_weight(device); |
b411b363 PR |
906 | |
907 | if (os.conn == C_VERIFY_S || os.conn == C_VERIFY_T) { | |
908 | if (n_oos) { | |
d0180171 | 909 | drbd_alert(device, "Online verify found %lu %dk block out of sync!\n", |
b411b363 PR |
910 | n_oos, Bit2KB(1)); |
911 | khelper_cmd = "out-of-sync"; | |
912 | } | |
913 | } else { | |
0b0ba1ef | 914 | D_ASSERT(device, (n_oos - device->rs_failed) == 0); |
b411b363 PR |
915 | |
916 | if (os.conn == C_SYNC_TARGET || os.conn == C_PAUSED_SYNC_T) | |
917 | khelper_cmd = "after-resync-target"; | |
918 | ||
aaaba345 | 919 | if (device->use_csums && device->rs_total) { |
b30ab791 AG |
920 | const unsigned long s = device->rs_same_csum; |
921 | const unsigned long t = device->rs_total; | |
b411b363 PR |
922 | const int ratio = |
923 | (t == 0) ? 0 : | |
924 | (t < 100000) ? ((s*100)/t) : (s/(t/100)); | |
d0180171 | 925 | drbd_info(device, "%u %% had equal checksums, eliminated: %luK; " |
b411b363 PR |
926 | "transferred %luK total %luK\n", |
927 | ratio, | |
b30ab791 AG |
928 | Bit2KB(device->rs_same_csum), |
929 | Bit2KB(device->rs_total - device->rs_same_csum), | |
930 | Bit2KB(device->rs_total)); | |
b411b363 PR |
931 | } |
932 | } | |
933 | ||
b30ab791 | 934 | if (device->rs_failed) { |
d0180171 | 935 | drbd_info(device, " %lu failed blocks\n", device->rs_failed); |
b411b363 PR |
936 | |
937 | if (os.conn == C_SYNC_TARGET || os.conn == C_PAUSED_SYNC_T) { | |
938 | ns.disk = D_INCONSISTENT; | |
939 | ns.pdsk = D_UP_TO_DATE; | |
940 | } else { | |
941 | ns.disk = D_UP_TO_DATE; | |
942 | ns.pdsk = D_INCONSISTENT; | |
943 | } | |
944 | } else { | |
945 | ns.disk = D_UP_TO_DATE; | |
946 | ns.pdsk = D_UP_TO_DATE; | |
947 | ||
948 | if (os.conn == C_SYNC_TARGET || os.conn == C_PAUSED_SYNC_T) { | |
b30ab791 | 949 | if (device->p_uuid) { |
b411b363 PR |
950 | int i; |
951 | for (i = UI_BITMAP ; i <= UI_HISTORY_END ; i++) | |
b30ab791 AG |
952 | _drbd_uuid_set(device, i, device->p_uuid[i]); |
953 | drbd_uuid_set(device, UI_BITMAP, device->ldev->md.uuid[UI_CURRENT]); | |
954 | _drbd_uuid_set(device, UI_CURRENT, device->p_uuid[UI_CURRENT]); | |
b411b363 | 955 | } else { |
d0180171 | 956 | drbd_err(device, "device->p_uuid is NULL! BUG\n"); |
b411b363 PR |
957 | } |
958 | } | |
959 | ||
62b0da3a LE |
960 | if (!(os.conn == C_VERIFY_S || os.conn == C_VERIFY_T)) { |
961 | /* for verify runs, we don't update uuids here, | |
962 | * so there would be nothing to report. */ | |
b30ab791 AG |
963 | drbd_uuid_set_bm(device, 0UL); |
964 | drbd_print_uuids(device, "updated UUIDs"); | |
965 | if (device->p_uuid) { | |
62b0da3a LE |
966 | /* Now the two UUID sets are equal, update what we |
967 | * know of the peer. */ | |
968 | int i; | |
969 | for (i = UI_CURRENT ; i <= UI_HISTORY_END ; i++) | |
b30ab791 | 970 | device->p_uuid[i] = device->ldev->md.uuid[i]; |
62b0da3a | 971 | } |
b411b363 PR |
972 | } |
973 | } | |
974 | ||
b30ab791 | 975 | _drbd_set_state(device, ns, CS_VERBOSE, NULL); |
b411b363 | 976 | out_unlock: |
0500813f | 977 | spin_unlock_irq(&device->resource->req_lock); |
b30ab791 | 978 | put_ldev(device); |
b411b363 | 979 | out: |
b30ab791 AG |
980 | device->rs_total = 0; |
981 | device->rs_failed = 0; | |
982 | device->rs_paused = 0; | |
58ffa580 LE |
983 | |
984 | /* reset start sector, if we reached end of device */ | |
b30ab791 AG |
985 | if (verify_done && device->ov_left == 0) |
986 | device->ov_start_sector = 0; | |
b411b363 | 987 | |
b30ab791 | 988 | drbd_md_sync(device); |
13d42685 | 989 | |
b411b363 | 990 | if (khelper_cmd) |
b30ab791 | 991 | drbd_khelper(device, khelper_cmd); |
b411b363 PR |
992 | |
993 | return 1; | |
994 | } | |
995 | ||
996 | /* helper */ | |
b30ab791 | 997 | static void move_to_net_ee_or_free(struct drbd_device *device, struct drbd_peer_request *peer_req) |
b411b363 | 998 | { |
045417f7 | 999 | if (drbd_peer_req_has_active_page(peer_req)) { |
b411b363 | 1000 | /* This might happen if sendpage() has not finished */ |
db830c46 | 1001 | int i = (peer_req->i.size + PAGE_SIZE -1) >> PAGE_SHIFT; |
b30ab791 AG |
1002 | atomic_add(i, &device->pp_in_use_by_net); |
1003 | atomic_sub(i, &device->pp_in_use); | |
0500813f | 1004 | spin_lock_irq(&device->resource->req_lock); |
a8cd15ba | 1005 | list_add_tail(&peer_req->w.list, &device->net_ee); |
0500813f | 1006 | spin_unlock_irq(&device->resource->req_lock); |
435f0740 | 1007 | wake_up(&drbd_pp_wait); |
b411b363 | 1008 | } else |
b30ab791 | 1009 | drbd_free_peer_req(device, peer_req); |
b411b363 PR |
1010 | } |
1011 | ||
1012 | /** | |
1013 | * w_e_end_data_req() - Worker callback, to send a P_DATA_REPLY packet in response to a P_DATA_REQUEST | |
b30ab791 | 1014 | * @device: DRBD device. |
b411b363 PR |
1015 | * @w: work object. |
1016 | * @cancel: The connection will be closed anyways | |
1017 | */ | |
99920dc5 | 1018 | int w_e_end_data_req(struct drbd_work *w, int cancel) |
b411b363 | 1019 | { |
a8cd15ba | 1020 | struct drbd_peer_request *peer_req = container_of(w, struct drbd_peer_request, w); |
6780139c AG |
1021 | struct drbd_peer_device *peer_device = peer_req->peer_device; |
1022 | struct drbd_device *device = peer_device->device; | |
99920dc5 | 1023 | int err; |
b411b363 PR |
1024 | |
1025 | if (unlikely(cancel)) { | |
b30ab791 AG |
1026 | drbd_free_peer_req(device, peer_req); |
1027 | dec_unacked(device); | |
99920dc5 | 1028 | return 0; |
b411b363 PR |
1029 | } |
1030 | ||
db830c46 | 1031 | if (likely((peer_req->flags & EE_WAS_ERROR) == 0)) { |
6780139c | 1032 | err = drbd_send_block(peer_device, P_DATA_REPLY, peer_req); |
b411b363 PR |
1033 | } else { |
1034 | if (__ratelimit(&drbd_ratelimit_state)) | |
d0180171 | 1035 | drbd_err(device, "Sending NegDReply. sector=%llus.\n", |
db830c46 | 1036 | (unsigned long long)peer_req->i.sector); |
b411b363 | 1037 | |
6780139c | 1038 | err = drbd_send_ack(peer_device, P_NEG_DREPLY, peer_req); |
b411b363 PR |
1039 | } |
1040 | ||
b30ab791 | 1041 | dec_unacked(device); |
b411b363 | 1042 | |
b30ab791 | 1043 | move_to_net_ee_or_free(device, peer_req); |
b411b363 | 1044 | |
99920dc5 | 1045 | if (unlikely(err)) |
d0180171 | 1046 | drbd_err(device, "drbd_send_block() failed\n"); |
99920dc5 | 1047 | return err; |
b411b363 PR |
1048 | } |
1049 | ||
1050 | /** | |
a209b4ae | 1051 | * w_e_end_rsdata_req() - Worker callback to send a P_RS_DATA_REPLY packet in response to a P_RS_DATA_REQUEST |
b411b363 PR |
1052 | * @w: work object. |
1053 | * @cancel: The connection will be closed anyways | |
1054 | */ | |
99920dc5 | 1055 | int w_e_end_rsdata_req(struct drbd_work *w, int cancel) |
b411b363 | 1056 | { |
a8cd15ba | 1057 | struct drbd_peer_request *peer_req = container_of(w, struct drbd_peer_request, w); |
6780139c AG |
1058 | struct drbd_peer_device *peer_device = peer_req->peer_device; |
1059 | struct drbd_device *device = peer_device->device; | |
99920dc5 | 1060 | int err; |
b411b363 PR |
1061 | |
1062 | if (unlikely(cancel)) { | |
b30ab791 AG |
1063 | drbd_free_peer_req(device, peer_req); |
1064 | dec_unacked(device); | |
99920dc5 | 1065 | return 0; |
b411b363 PR |
1066 | } |
1067 | ||
b30ab791 AG |
1068 | if (get_ldev_if_state(device, D_FAILED)) { |
1069 | drbd_rs_complete_io(device, peer_req->i.sector); | |
1070 | put_ldev(device); | |
b411b363 PR |
1071 | } |
1072 | ||
b30ab791 | 1073 | if (device->state.conn == C_AHEAD) { |
6780139c | 1074 | err = drbd_send_ack(peer_device, P_RS_CANCEL, peer_req); |
db830c46 | 1075 | } else if (likely((peer_req->flags & EE_WAS_ERROR) == 0)) { |
b30ab791 AG |
1076 | if (likely(device->state.pdsk >= D_INCONSISTENT)) { |
1077 | inc_rs_pending(device); | |
6780139c | 1078 | err = drbd_send_block(peer_device, P_RS_DATA_REPLY, peer_req); |
b411b363 PR |
1079 | } else { |
1080 | if (__ratelimit(&drbd_ratelimit_state)) | |
d0180171 | 1081 | drbd_err(device, "Not sending RSDataReply, " |
b411b363 | 1082 | "partner DISKLESS!\n"); |
99920dc5 | 1083 | err = 0; |
b411b363 PR |
1084 | } |
1085 | } else { | |
1086 | if (__ratelimit(&drbd_ratelimit_state)) | |
d0180171 | 1087 | drbd_err(device, "Sending NegRSDReply. sector %llus.\n", |
db830c46 | 1088 | (unsigned long long)peer_req->i.sector); |
b411b363 | 1089 | |
6780139c | 1090 | err = drbd_send_ack(peer_device, P_NEG_RS_DREPLY, peer_req); |
b411b363 PR |
1091 | |
1092 | /* update resync data with failure */ | |
b30ab791 | 1093 | drbd_rs_failed_io(device, peer_req->i.sector, peer_req->i.size); |
b411b363 PR |
1094 | } |
1095 | ||
b30ab791 | 1096 | dec_unacked(device); |
b411b363 | 1097 | |
b30ab791 | 1098 | move_to_net_ee_or_free(device, peer_req); |
b411b363 | 1099 | |
99920dc5 | 1100 | if (unlikely(err)) |
d0180171 | 1101 | drbd_err(device, "drbd_send_block() failed\n"); |
99920dc5 | 1102 | return err; |
b411b363 PR |
1103 | } |
1104 | ||
99920dc5 | 1105 | int w_e_end_csum_rs_req(struct drbd_work *w, int cancel) |
b411b363 | 1106 | { |
a8cd15ba | 1107 | struct drbd_peer_request *peer_req = container_of(w, struct drbd_peer_request, w); |
6780139c AG |
1108 | struct drbd_peer_device *peer_device = peer_req->peer_device; |
1109 | struct drbd_device *device = peer_device->device; | |
b411b363 PR |
1110 | struct digest_info *di; |
1111 | int digest_size; | |
1112 | void *digest = NULL; | |
99920dc5 | 1113 | int err, eq = 0; |
b411b363 PR |
1114 | |
1115 | if (unlikely(cancel)) { | |
b30ab791 AG |
1116 | drbd_free_peer_req(device, peer_req); |
1117 | dec_unacked(device); | |
99920dc5 | 1118 | return 0; |
b411b363 PR |
1119 | } |
1120 | ||
b30ab791 AG |
1121 | if (get_ldev(device)) { |
1122 | drbd_rs_complete_io(device, peer_req->i.sector); | |
1123 | put_ldev(device); | |
1d53f09e | 1124 | } |
b411b363 | 1125 | |
db830c46 | 1126 | di = peer_req->digest; |
b411b363 | 1127 | |
db830c46 | 1128 | if (likely((peer_req->flags & EE_WAS_ERROR) == 0)) { |
b411b363 PR |
1129 | /* quick hack to try to avoid a race against reconfiguration. |
1130 | * a real fix would be much more involved, | |
1131 | * introducing more locking mechanisms */ | |
6780139c AG |
1132 | if (peer_device->connection->csums_tfm) { |
1133 | digest_size = crypto_hash_digestsize(peer_device->connection->csums_tfm); | |
0b0ba1ef | 1134 | D_ASSERT(device, digest_size == di->digest_size); |
b411b363 PR |
1135 | digest = kmalloc(digest_size, GFP_NOIO); |
1136 | } | |
1137 | if (digest) { | |
6780139c | 1138 | drbd_csum_ee(peer_device->connection->csums_tfm, peer_req, digest); |
b411b363 PR |
1139 | eq = !memcmp(digest, di->digest, digest_size); |
1140 | kfree(digest); | |
1141 | } | |
1142 | ||
1143 | if (eq) { | |
b30ab791 | 1144 | drbd_set_in_sync(device, peer_req->i.sector, peer_req->i.size); |
676396d5 | 1145 | /* rs_same_csums unit is BM_BLOCK_SIZE */ |
b30ab791 | 1146 | device->rs_same_csum += peer_req->i.size >> BM_BLOCK_SHIFT; |
6780139c | 1147 | err = drbd_send_ack(peer_device, P_RS_IS_IN_SYNC, peer_req); |
b411b363 | 1148 | } else { |
b30ab791 | 1149 | inc_rs_pending(device); |
db830c46 AG |
1150 | peer_req->block_id = ID_SYNCER; /* By setting block_id, digest pointer becomes invalid! */ |
1151 | peer_req->flags &= ~EE_HAS_DIGEST; /* This peer request no longer has a digest pointer */ | |
204bba99 | 1152 | kfree(di); |
6780139c | 1153 | err = drbd_send_block(peer_device, P_RS_DATA_REPLY, peer_req); |
b411b363 PR |
1154 | } |
1155 | } else { | |
6780139c | 1156 | err = drbd_send_ack(peer_device, P_NEG_RS_DREPLY, peer_req); |
b411b363 | 1157 | if (__ratelimit(&drbd_ratelimit_state)) |
d0180171 | 1158 | drbd_err(device, "Sending NegDReply. I guess it gets messy.\n"); |
b411b363 PR |
1159 | } |
1160 | ||
b30ab791 AG |
1161 | dec_unacked(device); |
1162 | move_to_net_ee_or_free(device, peer_req); | |
b411b363 | 1163 | |
99920dc5 | 1164 | if (unlikely(err)) |
d0180171 | 1165 | drbd_err(device, "drbd_send_block/ack() failed\n"); |
99920dc5 | 1166 | return err; |
b411b363 PR |
1167 | } |
1168 | ||
99920dc5 | 1169 | int w_e_end_ov_req(struct drbd_work *w, int cancel) |
b411b363 | 1170 | { |
a8cd15ba | 1171 | struct drbd_peer_request *peer_req = container_of(w, struct drbd_peer_request, w); |
6780139c AG |
1172 | struct drbd_peer_device *peer_device = peer_req->peer_device; |
1173 | struct drbd_device *device = peer_device->device; | |
db830c46 AG |
1174 | sector_t sector = peer_req->i.sector; |
1175 | unsigned int size = peer_req->i.size; | |
b411b363 PR |
1176 | int digest_size; |
1177 | void *digest; | |
99920dc5 | 1178 | int err = 0; |
b411b363 PR |
1179 | |
1180 | if (unlikely(cancel)) | |
1181 | goto out; | |
1182 | ||
6780139c | 1183 | digest_size = crypto_hash_digestsize(peer_device->connection->verify_tfm); |
b411b363 | 1184 | digest = kmalloc(digest_size, GFP_NOIO); |
8f21420e | 1185 | if (!digest) { |
99920dc5 | 1186 | err = 1; /* terminate the connection in case the allocation failed */ |
8f21420e | 1187 | goto out; |
b411b363 PR |
1188 | } |
1189 | ||
db830c46 | 1190 | if (likely(!(peer_req->flags & EE_WAS_ERROR))) |
6780139c | 1191 | drbd_csum_ee(peer_device->connection->verify_tfm, peer_req, digest); |
8f21420e PR |
1192 | else |
1193 | memset(digest, 0, digest_size); | |
1194 | ||
53ea4331 LE |
1195 | /* Free e and pages before send. |
1196 | * In case we block on congestion, we could otherwise run into | |
1197 | * some distributed deadlock, if the other side blocks on | |
1198 | * congestion as well, because our receiver blocks in | |
c37c8ecf | 1199 | * drbd_alloc_pages due to pp_in_use > max_buffers. */ |
b30ab791 | 1200 | drbd_free_peer_req(device, peer_req); |
db830c46 | 1201 | peer_req = NULL; |
b30ab791 | 1202 | inc_rs_pending(device); |
6780139c | 1203 | err = drbd_send_drequest_csum(peer_device, sector, size, digest, digest_size, P_OV_REPLY); |
99920dc5 | 1204 | if (err) |
b30ab791 | 1205 | dec_rs_pending(device); |
8f21420e PR |
1206 | kfree(digest); |
1207 | ||
b411b363 | 1208 | out: |
db830c46 | 1209 | if (peer_req) |
b30ab791 AG |
1210 | drbd_free_peer_req(device, peer_req); |
1211 | dec_unacked(device); | |
99920dc5 | 1212 | return err; |
b411b363 PR |
1213 | } |
1214 | ||
b30ab791 | 1215 | void drbd_ov_out_of_sync_found(struct drbd_device *device, sector_t sector, int size) |
b411b363 | 1216 | { |
b30ab791 AG |
1217 | if (device->ov_last_oos_start + device->ov_last_oos_size == sector) { |
1218 | device->ov_last_oos_size += size>>9; | |
b411b363 | 1219 | } else { |
b30ab791 AG |
1220 | device->ov_last_oos_start = sector; |
1221 | device->ov_last_oos_size = size>>9; | |
b411b363 | 1222 | } |
b30ab791 | 1223 | drbd_set_out_of_sync(device, sector, size); |
b411b363 PR |
1224 | } |
1225 | ||
99920dc5 | 1226 | int w_e_end_ov_reply(struct drbd_work *w, int cancel) |
b411b363 | 1227 | { |
a8cd15ba | 1228 | struct drbd_peer_request *peer_req = container_of(w, struct drbd_peer_request, w); |
6780139c AG |
1229 | struct drbd_peer_device *peer_device = peer_req->peer_device; |
1230 | struct drbd_device *device = peer_device->device; | |
b411b363 | 1231 | struct digest_info *di; |
b411b363 | 1232 | void *digest; |
db830c46 AG |
1233 | sector_t sector = peer_req->i.sector; |
1234 | unsigned int size = peer_req->i.size; | |
53ea4331 | 1235 | int digest_size; |
99920dc5 | 1236 | int err, eq = 0; |
58ffa580 | 1237 | bool stop_sector_reached = false; |
b411b363 PR |
1238 | |
1239 | if (unlikely(cancel)) { | |
b30ab791 AG |
1240 | drbd_free_peer_req(device, peer_req); |
1241 | dec_unacked(device); | |
99920dc5 | 1242 | return 0; |
b411b363 PR |
1243 | } |
1244 | ||
1245 | /* after "cancel", because after drbd_disconnect/drbd_rs_cancel_all | |
1246 | * the resync lru has been cleaned up already */ | |
b30ab791 AG |
1247 | if (get_ldev(device)) { |
1248 | drbd_rs_complete_io(device, peer_req->i.sector); | |
1249 | put_ldev(device); | |
1d53f09e | 1250 | } |
b411b363 | 1251 | |
db830c46 | 1252 | di = peer_req->digest; |
b411b363 | 1253 | |
db830c46 | 1254 | if (likely((peer_req->flags & EE_WAS_ERROR) == 0)) { |
6780139c | 1255 | digest_size = crypto_hash_digestsize(peer_device->connection->verify_tfm); |
b411b363 PR |
1256 | digest = kmalloc(digest_size, GFP_NOIO); |
1257 | if (digest) { | |
6780139c | 1258 | drbd_csum_ee(peer_device->connection->verify_tfm, peer_req, digest); |
b411b363 | 1259 | |
0b0ba1ef | 1260 | D_ASSERT(device, digest_size == di->digest_size); |
b411b363 PR |
1261 | eq = !memcmp(digest, di->digest, digest_size); |
1262 | kfree(digest); | |
1263 | } | |
b411b363 PR |
1264 | } |
1265 | ||
9676c760 LE |
1266 | /* Free peer_req and pages before send. |
1267 | * In case we block on congestion, we could otherwise run into | |
1268 | * some distributed deadlock, if the other side blocks on | |
1269 | * congestion as well, because our receiver blocks in | |
c37c8ecf | 1270 | * drbd_alloc_pages due to pp_in_use > max_buffers. */ |
b30ab791 | 1271 | drbd_free_peer_req(device, peer_req); |
b411b363 | 1272 | if (!eq) |
b30ab791 | 1273 | drbd_ov_out_of_sync_found(device, sector, size); |
b411b363 | 1274 | else |
b30ab791 | 1275 | ov_out_of_sync_print(device); |
b411b363 | 1276 | |
6780139c | 1277 | err = drbd_send_ack_ex(peer_device, P_OV_RESULT, sector, size, |
fa79abd8 | 1278 | eq ? ID_IN_SYNC : ID_OUT_OF_SYNC); |
b411b363 | 1279 | |
b30ab791 | 1280 | dec_unacked(device); |
b411b363 | 1281 | |
b30ab791 | 1282 | --device->ov_left; |
ea5442af LE |
1283 | |
1284 | /* let's advance progress step marks only for every other megabyte */ | |
b30ab791 AG |
1285 | if ((device->ov_left & 0x200) == 0x200) |
1286 | drbd_advance_rs_marks(device, device->ov_left); | |
ea5442af | 1287 | |
b30ab791 AG |
1288 | stop_sector_reached = verify_can_do_stop_sector(device) && |
1289 | (sector + (size>>9)) >= device->ov_stop_sector; | |
58ffa580 | 1290 | |
b30ab791 AG |
1291 | if (device->ov_left == 0 || stop_sector_reached) { |
1292 | ov_out_of_sync_print(device); | |
1293 | drbd_resync_finished(device); | |
b411b363 PR |
1294 | } |
1295 | ||
99920dc5 | 1296 | return err; |
b411b363 PR |
1297 | } |
1298 | ||
b6dd1a89 LE |
1299 | /* FIXME |
1300 | * We need to track the number of pending barrier acks, | |
1301 | * and to be able to wait for them. | |
1302 | * See also comment in drbd_adm_attach before drbd_suspend_io. | |
1303 | */ | |
bde89a9e | 1304 | static int drbd_send_barrier(struct drbd_connection *connection) |
b411b363 | 1305 | { |
9f5bdc33 | 1306 | struct p_barrier *p; |
b6dd1a89 | 1307 | struct drbd_socket *sock; |
b411b363 | 1308 | |
bde89a9e AG |
1309 | sock = &connection->data; |
1310 | p = conn_prepare_command(connection, sock); | |
9f5bdc33 AG |
1311 | if (!p) |
1312 | return -EIO; | |
bde89a9e | 1313 | p->barrier = connection->send.current_epoch_nr; |
b6dd1a89 | 1314 | p->pad = 0; |
bde89a9e | 1315 | connection->send.current_epoch_writes = 0; |
b6dd1a89 | 1316 | |
bde89a9e | 1317 | return conn_send_command(connection, sock, P_BARRIER, sizeof(*p), NULL, 0); |
b411b363 PR |
1318 | } |
1319 | ||
99920dc5 | 1320 | int w_send_write_hint(struct drbd_work *w, int cancel) |
b411b363 | 1321 | { |
84b8c06b AG |
1322 | struct drbd_device *device = |
1323 | container_of(w, struct drbd_device, unplug_work); | |
9f5bdc33 AG |
1324 | struct drbd_socket *sock; |
1325 | ||
b411b363 | 1326 | if (cancel) |
99920dc5 | 1327 | return 0; |
a6b32bc3 | 1328 | sock = &first_peer_device(device)->connection->data; |
69a22773 | 1329 | if (!drbd_prepare_command(first_peer_device(device), sock)) |
9f5bdc33 | 1330 | return -EIO; |
69a22773 | 1331 | return drbd_send_command(first_peer_device(device), sock, P_UNPLUG_REMOTE, 0, NULL, 0); |
b411b363 PR |
1332 | } |
1333 | ||
bde89a9e | 1334 | static void re_init_if_first_write(struct drbd_connection *connection, unsigned int epoch) |
4eb9b3cb | 1335 | { |
bde89a9e AG |
1336 | if (!connection->send.seen_any_write_yet) { |
1337 | connection->send.seen_any_write_yet = true; | |
1338 | connection->send.current_epoch_nr = epoch; | |
1339 | connection->send.current_epoch_writes = 0; | |
4eb9b3cb LE |
1340 | } |
1341 | } | |
1342 | ||
bde89a9e | 1343 | static void maybe_send_barrier(struct drbd_connection *connection, unsigned int epoch) |
4eb9b3cb LE |
1344 | { |
1345 | /* re-init if first write on this connection */ | |
bde89a9e | 1346 | if (!connection->send.seen_any_write_yet) |
4eb9b3cb | 1347 | return; |
bde89a9e AG |
1348 | if (connection->send.current_epoch_nr != epoch) { |
1349 | if (connection->send.current_epoch_writes) | |
1350 | drbd_send_barrier(connection); | |
1351 | connection->send.current_epoch_nr = epoch; | |
4eb9b3cb LE |
1352 | } |
1353 | } | |
1354 | ||
8f7bed77 | 1355 | int w_send_out_of_sync(struct drbd_work *w, int cancel) |
73a01a18 PR |
1356 | { |
1357 | struct drbd_request *req = container_of(w, struct drbd_request, w); | |
84b8c06b | 1358 | struct drbd_device *device = req->device; |
44a4d551 LE |
1359 | struct drbd_peer_device *const peer_device = first_peer_device(device); |
1360 | struct drbd_connection *const connection = peer_device->connection; | |
99920dc5 | 1361 | int err; |
73a01a18 PR |
1362 | |
1363 | if (unlikely(cancel)) { | |
8554df1c | 1364 | req_mod(req, SEND_CANCELED); |
99920dc5 | 1365 | return 0; |
73a01a18 | 1366 | } |
e5f891b2 | 1367 | req->pre_send_jif = jiffies; |
73a01a18 | 1368 | |
bde89a9e | 1369 | /* this time, no connection->send.current_epoch_writes++; |
b6dd1a89 LE |
1370 | * If it was sent, it was the closing barrier for the last |
1371 | * replicated epoch, before we went into AHEAD mode. | |
1372 | * No more barriers will be sent, until we leave AHEAD mode again. */ | |
bde89a9e | 1373 | maybe_send_barrier(connection, req->epoch); |
b6dd1a89 | 1374 | |
44a4d551 | 1375 | err = drbd_send_out_of_sync(peer_device, req); |
8554df1c | 1376 | req_mod(req, OOS_HANDED_TO_NETWORK); |
73a01a18 | 1377 | |
99920dc5 | 1378 | return err; |
73a01a18 PR |
1379 | } |
1380 | ||
b411b363 PR |
1381 | /** |
1382 | * w_send_dblock() - Worker callback to send a P_DATA packet in order to mirror a write request | |
b411b363 PR |
1383 | * @w: work object. |
1384 | * @cancel: The connection will be closed anyways | |
1385 | */ | |
99920dc5 | 1386 | int w_send_dblock(struct drbd_work *w, int cancel) |
b411b363 PR |
1387 | { |
1388 | struct drbd_request *req = container_of(w, struct drbd_request, w); | |
84b8c06b | 1389 | struct drbd_device *device = req->device; |
44a4d551 LE |
1390 | struct drbd_peer_device *const peer_device = first_peer_device(device); |
1391 | struct drbd_connection *connection = peer_device->connection; | |
99920dc5 | 1392 | int err; |
b411b363 PR |
1393 | |
1394 | if (unlikely(cancel)) { | |
8554df1c | 1395 | req_mod(req, SEND_CANCELED); |
99920dc5 | 1396 | return 0; |
b411b363 | 1397 | } |
e5f891b2 | 1398 | req->pre_send_jif = jiffies; |
b411b363 | 1399 | |
bde89a9e AG |
1400 | re_init_if_first_write(connection, req->epoch); |
1401 | maybe_send_barrier(connection, req->epoch); | |
1402 | connection->send.current_epoch_writes++; | |
b6dd1a89 | 1403 | |
44a4d551 | 1404 | err = drbd_send_dblock(peer_device, req); |
99920dc5 | 1405 | req_mod(req, err ? SEND_FAILED : HANDED_OVER_TO_NETWORK); |
b411b363 | 1406 | |
99920dc5 | 1407 | return err; |
b411b363 PR |
1408 | } |
1409 | ||
1410 | /** | |
1411 | * w_send_read_req() - Worker callback to send a read request (P_DATA_REQUEST) packet | |
b411b363 PR |
1412 | * @w: work object. |
1413 | * @cancel: The connection will be closed anyways | |
1414 | */ | |
99920dc5 | 1415 | int w_send_read_req(struct drbd_work *w, int cancel) |
b411b363 PR |
1416 | { |
1417 | struct drbd_request *req = container_of(w, struct drbd_request, w); | |
84b8c06b | 1418 | struct drbd_device *device = req->device; |
44a4d551 LE |
1419 | struct drbd_peer_device *const peer_device = first_peer_device(device); |
1420 | struct drbd_connection *connection = peer_device->connection; | |
99920dc5 | 1421 | int err; |
b411b363 PR |
1422 | |
1423 | if (unlikely(cancel)) { | |
8554df1c | 1424 | req_mod(req, SEND_CANCELED); |
99920dc5 | 1425 | return 0; |
b411b363 | 1426 | } |
e5f891b2 | 1427 | req->pre_send_jif = jiffies; |
b411b363 | 1428 | |
b6dd1a89 LE |
1429 | /* Even read requests may close a write epoch, |
1430 | * if there was any yet. */ | |
bde89a9e | 1431 | maybe_send_barrier(connection, req->epoch); |
b6dd1a89 | 1432 | |
44a4d551 | 1433 | err = drbd_send_drequest(peer_device, P_DATA_REQUEST, req->i.sector, req->i.size, |
6c1005e7 | 1434 | (unsigned long)req); |
b411b363 | 1435 | |
99920dc5 | 1436 | req_mod(req, err ? SEND_FAILED : HANDED_OVER_TO_NETWORK); |
b411b363 | 1437 | |
99920dc5 | 1438 | return err; |
b411b363 PR |
1439 | } |
1440 | ||
99920dc5 | 1441 | int w_restart_disk_io(struct drbd_work *w, int cancel) |
265be2d0 PR |
1442 | { |
1443 | struct drbd_request *req = container_of(w, struct drbd_request, w); | |
84b8c06b | 1444 | struct drbd_device *device = req->device; |
265be2d0 | 1445 | |
0778286a | 1446 | if (bio_data_dir(req->master_bio) == WRITE && req->rq_state & RQ_IN_ACT_LOG) |
4dd726f0 | 1447 | drbd_al_begin_io(device, &req->i); |
265be2d0 PR |
1448 | |
1449 | drbd_req_make_private_bio(req, req->master_bio); | |
b30ab791 | 1450 | req->private_bio->bi_bdev = device->ldev->backing_bdev; |
265be2d0 PR |
1451 | generic_make_request(req->private_bio); |
1452 | ||
99920dc5 | 1453 | return 0; |
265be2d0 PR |
1454 | } |
1455 | ||
b30ab791 | 1456 | static int _drbd_may_sync_now(struct drbd_device *device) |
b411b363 | 1457 | { |
b30ab791 | 1458 | struct drbd_device *odev = device; |
95f8efd0 | 1459 | int resync_after; |
b411b363 PR |
1460 | |
1461 | while (1) { | |
a3f8f7dc | 1462 | if (!odev->ldev || odev->state.disk == D_DISKLESS) |
438c8374 | 1463 | return 1; |
daeda1cc | 1464 | rcu_read_lock(); |
95f8efd0 | 1465 | resync_after = rcu_dereference(odev->ldev->disk_conf)->resync_after; |
daeda1cc | 1466 | rcu_read_unlock(); |
95f8efd0 | 1467 | if (resync_after == -1) |
b411b363 | 1468 | return 1; |
b30ab791 | 1469 | odev = minor_to_device(resync_after); |
a3f8f7dc | 1470 | if (!odev) |
841ce241 | 1471 | return 1; |
b411b363 PR |
1472 | if ((odev->state.conn >= C_SYNC_SOURCE && |
1473 | odev->state.conn <= C_PAUSED_SYNC_T) || | |
1474 | odev->state.aftr_isp || odev->state.peer_isp || | |
1475 | odev->state.user_isp) | |
1476 | return 0; | |
1477 | } | |
1478 | } | |
1479 | ||
1480 | /** | |
1481 | * _drbd_pause_after() - Pause resync on all devices that may not resync now | |
b30ab791 | 1482 | * @device: DRBD device. |
b411b363 PR |
1483 | * |
1484 | * Called from process context only (admin command and after_state_ch). | |
1485 | */ | |
b30ab791 | 1486 | static int _drbd_pause_after(struct drbd_device *device) |
b411b363 | 1487 | { |
54761697 | 1488 | struct drbd_device *odev; |
b411b363 PR |
1489 | int i, rv = 0; |
1490 | ||
695d08fa | 1491 | rcu_read_lock(); |
05a10ec7 | 1492 | idr_for_each_entry(&drbd_devices, odev, i) { |
b411b363 PR |
1493 | if (odev->state.conn == C_STANDALONE && odev->state.disk == D_DISKLESS) |
1494 | continue; | |
1495 | if (!_drbd_may_sync_now(odev)) | |
1496 | rv |= (__drbd_set_state(_NS(odev, aftr_isp, 1), CS_HARD, NULL) | |
1497 | != SS_NOTHING_TO_DO); | |
1498 | } | |
695d08fa | 1499 | rcu_read_unlock(); |
b411b363 PR |
1500 | |
1501 | return rv; | |
1502 | } | |
1503 | ||
1504 | /** | |
1505 | * _drbd_resume_next() - Resume resync on all devices that may resync now | |
b30ab791 | 1506 | * @device: DRBD device. |
b411b363 PR |
1507 | * |
1508 | * Called from process context only (admin command and worker). | |
1509 | */ | |
b30ab791 | 1510 | static int _drbd_resume_next(struct drbd_device *device) |
b411b363 | 1511 | { |
54761697 | 1512 | struct drbd_device *odev; |
b411b363 PR |
1513 | int i, rv = 0; |
1514 | ||
695d08fa | 1515 | rcu_read_lock(); |
05a10ec7 | 1516 | idr_for_each_entry(&drbd_devices, odev, i) { |
b411b363 PR |
1517 | if (odev->state.conn == C_STANDALONE && odev->state.disk == D_DISKLESS) |
1518 | continue; | |
1519 | if (odev->state.aftr_isp) { | |
1520 | if (_drbd_may_sync_now(odev)) | |
1521 | rv |= (__drbd_set_state(_NS(odev, aftr_isp, 0), | |
1522 | CS_HARD, NULL) | |
1523 | != SS_NOTHING_TO_DO) ; | |
1524 | } | |
1525 | } | |
695d08fa | 1526 | rcu_read_unlock(); |
b411b363 PR |
1527 | return rv; |
1528 | } | |
1529 | ||
b30ab791 | 1530 | void resume_next_sg(struct drbd_device *device) |
b411b363 PR |
1531 | { |
1532 | write_lock_irq(&global_state_lock); | |
b30ab791 | 1533 | _drbd_resume_next(device); |
b411b363 PR |
1534 | write_unlock_irq(&global_state_lock); |
1535 | } | |
1536 | ||
b30ab791 | 1537 | void suspend_other_sg(struct drbd_device *device) |
b411b363 PR |
1538 | { |
1539 | write_lock_irq(&global_state_lock); | |
b30ab791 | 1540 | _drbd_pause_after(device); |
b411b363 PR |
1541 | write_unlock_irq(&global_state_lock); |
1542 | } | |
1543 | ||
dc97b708 | 1544 | /* caller must hold global_state_lock */ |
b30ab791 | 1545 | enum drbd_ret_code drbd_resync_after_valid(struct drbd_device *device, int o_minor) |
b411b363 | 1546 | { |
54761697 | 1547 | struct drbd_device *odev; |
95f8efd0 | 1548 | int resync_after; |
b411b363 PR |
1549 | |
1550 | if (o_minor == -1) | |
1551 | return NO_ERROR; | |
a3f8f7dc | 1552 | if (o_minor < -1 || o_minor > MINORMASK) |
95f8efd0 | 1553 | return ERR_RESYNC_AFTER; |
b411b363 PR |
1554 | |
1555 | /* check for loops */ | |
b30ab791 | 1556 | odev = minor_to_device(o_minor); |
b411b363 | 1557 | while (1) { |
b30ab791 | 1558 | if (odev == device) |
95f8efd0 | 1559 | return ERR_RESYNC_AFTER_CYCLE; |
b411b363 | 1560 | |
a3f8f7dc LE |
1561 | /* You are free to depend on diskless, non-existing, |
1562 | * or not yet/no longer existing minors. | |
1563 | * We only reject dependency loops. | |
1564 | * We cannot follow the dependency chain beyond a detached or | |
1565 | * missing minor. | |
1566 | */ | |
1567 | if (!odev || !odev->ldev || odev->state.disk == D_DISKLESS) | |
1568 | return NO_ERROR; | |
1569 | ||
daeda1cc | 1570 | rcu_read_lock(); |
95f8efd0 | 1571 | resync_after = rcu_dereference(odev->ldev->disk_conf)->resync_after; |
daeda1cc | 1572 | rcu_read_unlock(); |
b411b363 | 1573 | /* dependency chain ends here, no cycles. */ |
95f8efd0 | 1574 | if (resync_after == -1) |
b411b363 PR |
1575 | return NO_ERROR; |
1576 | ||
1577 | /* follow the dependency chain */ | |
b30ab791 | 1578 | odev = minor_to_device(resync_after); |
b411b363 PR |
1579 | } |
1580 | } | |
1581 | ||
dc97b708 | 1582 | /* caller must hold global_state_lock */ |
b30ab791 | 1583 | void drbd_resync_after_changed(struct drbd_device *device) |
b411b363 PR |
1584 | { |
1585 | int changes; | |
b411b363 | 1586 | |
dc97b708 | 1587 | do { |
b30ab791 AG |
1588 | changes = _drbd_pause_after(device); |
1589 | changes |= _drbd_resume_next(device); | |
dc97b708 | 1590 | } while (changes); |
b411b363 PR |
1591 | } |
1592 | ||
b30ab791 | 1593 | void drbd_rs_controller_reset(struct drbd_device *device) |
9bd28d3c | 1594 | { |
813472ce PR |
1595 | struct fifo_buffer *plan; |
1596 | ||
b30ab791 AG |
1597 | atomic_set(&device->rs_sect_in, 0); |
1598 | atomic_set(&device->rs_sect_ev, 0); | |
1599 | device->rs_in_flight = 0; | |
813472ce PR |
1600 | |
1601 | /* Updating the RCU protected object in place is necessary since | |
1602 | this function gets called from atomic context. | |
1603 | It is valid since all other updates also lead to an completely | |
1604 | empty fifo */ | |
1605 | rcu_read_lock(); | |
b30ab791 | 1606 | plan = rcu_dereference(device->rs_plan_s); |
813472ce PR |
1607 | plan->total = 0; |
1608 | fifo_set(plan, 0); | |
1609 | rcu_read_unlock(); | |
9bd28d3c LE |
1610 | } |
1611 | ||
1f04af33 PR |
1612 | void start_resync_timer_fn(unsigned long data) |
1613 | { | |
b30ab791 | 1614 | struct drbd_device *device = (struct drbd_device *) data; |
ac0acb9e | 1615 | drbd_device_post_work(device, RS_START); |
1f04af33 PR |
1616 | } |
1617 | ||
ac0acb9e | 1618 | static void do_start_resync(struct drbd_device *device) |
1f04af33 | 1619 | { |
b30ab791 | 1620 | if (atomic_read(&device->unacked_cnt) || atomic_read(&device->rs_pending_cnt)) { |
ac0acb9e | 1621 | drbd_warn(device, "postponing start_resync ...\n"); |
b30ab791 AG |
1622 | device->start_resync_timer.expires = jiffies + HZ/10; |
1623 | add_timer(&device->start_resync_timer); | |
ac0acb9e | 1624 | return; |
1f04af33 PR |
1625 | } |
1626 | ||
b30ab791 AG |
1627 | drbd_start_resync(device, C_SYNC_SOURCE); |
1628 | clear_bit(AHEAD_TO_SYNC_SOURCE, &device->flags); | |
1f04af33 PR |
1629 | } |
1630 | ||
aaaba345 LE |
1631 | static bool use_checksum_based_resync(struct drbd_connection *connection, struct drbd_device *device) |
1632 | { | |
1633 | bool csums_after_crash_only; | |
1634 | rcu_read_lock(); | |
1635 | csums_after_crash_only = rcu_dereference(connection->net_conf)->csums_after_crash_only; | |
1636 | rcu_read_unlock(); | |
1637 | return connection->agreed_pro_version >= 89 && /* supported? */ | |
1638 | connection->csums_tfm && /* configured? */ | |
1639 | (csums_after_crash_only == 0 /* use for each resync? */ | |
1640 | || test_bit(CRASHED_PRIMARY, &device->flags)); /* or only after Primary crash? */ | |
1641 | } | |
1642 | ||
b411b363 PR |
1643 | /** |
1644 | * drbd_start_resync() - Start the resync process | |
b30ab791 | 1645 | * @device: DRBD device. |
b411b363 PR |
1646 | * @side: Either C_SYNC_SOURCE or C_SYNC_TARGET |
1647 | * | |
1648 | * This function might bring you directly into one of the | |
1649 | * C_PAUSED_SYNC_* states. | |
1650 | */ | |
b30ab791 | 1651 | void drbd_start_resync(struct drbd_device *device, enum drbd_conns side) |
b411b363 | 1652 | { |
44a4d551 LE |
1653 | struct drbd_peer_device *peer_device = first_peer_device(device); |
1654 | struct drbd_connection *connection = peer_device ? peer_device->connection : NULL; | |
b411b363 PR |
1655 | union drbd_state ns; |
1656 | int r; | |
1657 | ||
b30ab791 | 1658 | if (device->state.conn >= C_SYNC_SOURCE && device->state.conn < C_AHEAD) { |
d0180171 | 1659 | drbd_err(device, "Resync already running!\n"); |
b411b363 PR |
1660 | return; |
1661 | } | |
1662 | ||
b30ab791 | 1663 | if (!test_bit(B_RS_H_DONE, &device->flags)) { |
e64a3294 PR |
1664 | if (side == C_SYNC_TARGET) { |
1665 | /* Since application IO was locked out during C_WF_BITMAP_T and | |
1666 | C_WF_SYNC_UUID we are still unmodified. Before going to C_SYNC_TARGET | |
1667 | we check that we might make the data inconsistent. */ | |
b30ab791 | 1668 | r = drbd_khelper(device, "before-resync-target"); |
e64a3294 PR |
1669 | r = (r >> 8) & 0xff; |
1670 | if (r > 0) { | |
d0180171 | 1671 | drbd_info(device, "before-resync-target handler returned %d, " |
09b9e797 | 1672 | "dropping connection.\n", r); |
44a4d551 | 1673 | conn_request_state(connection, NS(conn, C_DISCONNECTING), CS_HARD); |
09b9e797 PR |
1674 | return; |
1675 | } | |
e64a3294 | 1676 | } else /* C_SYNC_SOURCE */ { |
b30ab791 | 1677 | r = drbd_khelper(device, "before-resync-source"); |
e64a3294 PR |
1678 | r = (r >> 8) & 0xff; |
1679 | if (r > 0) { | |
1680 | if (r == 3) { | |
d0180171 | 1681 | drbd_info(device, "before-resync-source handler returned %d, " |
e64a3294 PR |
1682 | "ignoring. Old userland tools?", r); |
1683 | } else { | |
d0180171 | 1684 | drbd_info(device, "before-resync-source handler returned %d, " |
e64a3294 | 1685 | "dropping connection.\n", r); |
44a4d551 | 1686 | conn_request_state(connection, |
a6b32bc3 | 1687 | NS(conn, C_DISCONNECTING), CS_HARD); |
e64a3294 PR |
1688 | return; |
1689 | } | |
1690 | } | |
09b9e797 | 1691 | } |
b411b363 PR |
1692 | } |
1693 | ||
44a4d551 | 1694 | if (current == connection->worker.task) { |
dad20554 | 1695 | /* The worker should not sleep waiting for state_mutex, |
e64a3294 | 1696 | that can take long */ |
b30ab791 AG |
1697 | if (!mutex_trylock(device->state_mutex)) { |
1698 | set_bit(B_RS_H_DONE, &device->flags); | |
1699 | device->start_resync_timer.expires = jiffies + HZ/5; | |
1700 | add_timer(&device->start_resync_timer); | |
e64a3294 PR |
1701 | return; |
1702 | } | |
1703 | } else { | |
b30ab791 | 1704 | mutex_lock(device->state_mutex); |
e64a3294 | 1705 | } |
b30ab791 | 1706 | clear_bit(B_RS_H_DONE, &device->flags); |
b411b363 | 1707 | |
074f4afe LE |
1708 | /* req_lock: serialize with drbd_send_and_submit() and others |
1709 | * global_state_lock: for stable sync-after dependencies */ | |
1710 | spin_lock_irq(&device->resource->req_lock); | |
1711 | write_lock(&global_state_lock); | |
a700471b | 1712 | /* Did some connection breakage or IO error race with us? */ |
b30ab791 AG |
1713 | if (device->state.conn < C_CONNECTED |
1714 | || !get_ldev_if_state(device, D_NEGOTIATING)) { | |
074f4afe LE |
1715 | write_unlock(&global_state_lock); |
1716 | spin_unlock_irq(&device->resource->req_lock); | |
b30ab791 | 1717 | mutex_unlock(device->state_mutex); |
b411b363 PR |
1718 | return; |
1719 | } | |
1720 | ||
b30ab791 | 1721 | ns = drbd_read_state(device); |
b411b363 | 1722 | |
b30ab791 | 1723 | ns.aftr_isp = !_drbd_may_sync_now(device); |
b411b363 PR |
1724 | |
1725 | ns.conn = side; | |
1726 | ||
1727 | if (side == C_SYNC_TARGET) | |
1728 | ns.disk = D_INCONSISTENT; | |
1729 | else /* side == C_SYNC_SOURCE */ | |
1730 | ns.pdsk = D_INCONSISTENT; | |
1731 | ||
b30ab791 AG |
1732 | r = __drbd_set_state(device, ns, CS_VERBOSE, NULL); |
1733 | ns = drbd_read_state(device); | |
b411b363 PR |
1734 | |
1735 | if (ns.conn < C_CONNECTED) | |
1736 | r = SS_UNKNOWN_ERROR; | |
1737 | ||
1738 | if (r == SS_SUCCESS) { | |
b30ab791 | 1739 | unsigned long tw = drbd_bm_total_weight(device); |
1d7734a0 LE |
1740 | unsigned long now = jiffies; |
1741 | int i; | |
1742 | ||
b30ab791 AG |
1743 | device->rs_failed = 0; |
1744 | device->rs_paused = 0; | |
1745 | device->rs_same_csum = 0; | |
1746 | device->rs_last_events = 0; | |
1747 | device->rs_last_sect_ev = 0; | |
1748 | device->rs_total = tw; | |
1749 | device->rs_start = now; | |
1d7734a0 | 1750 | for (i = 0; i < DRBD_SYNC_MARKS; i++) { |
b30ab791 AG |
1751 | device->rs_mark_left[i] = tw; |
1752 | device->rs_mark_time[i] = now; | |
1d7734a0 | 1753 | } |
b30ab791 | 1754 | _drbd_pause_after(device); |
5ab7d2c0 LE |
1755 | /* Forget potentially stale cached per resync extent bit-counts. |
1756 | * Open coded drbd_rs_cancel_all(device), we already have IRQs | |
1757 | * disabled, and know the disk state is ok. */ | |
1758 | spin_lock(&device->al_lock); | |
1759 | lc_reset(device->resync); | |
1760 | device->resync_locked = 0; | |
1761 | device->resync_wenr = LC_FREE; | |
1762 | spin_unlock(&device->al_lock); | |
b411b363 | 1763 | } |
074f4afe LE |
1764 | write_unlock(&global_state_lock); |
1765 | spin_unlock_irq(&device->resource->req_lock); | |
5a22db89 | 1766 | |
b411b363 | 1767 | if (r == SS_SUCCESS) { |
5ab7d2c0 | 1768 | wake_up(&device->al_wait); /* for lc_reset() above */ |
328e0f12 PR |
1769 | /* reset rs_last_bcast when a resync or verify is started, |
1770 | * to deal with potential jiffies wrap. */ | |
b30ab791 | 1771 | device->rs_last_bcast = jiffies - HZ; |
328e0f12 | 1772 | |
d0180171 | 1773 | drbd_info(device, "Began resync as %s (will sync %lu KB [%lu bits set]).\n", |
b411b363 | 1774 | drbd_conn_str(ns.conn), |
b30ab791 AG |
1775 | (unsigned long) device->rs_total << (BM_BLOCK_SHIFT-10), |
1776 | (unsigned long) device->rs_total); | |
aaaba345 | 1777 | if (side == C_SYNC_TARGET) { |
b30ab791 | 1778 | device->bm_resync_fo = 0; |
aaaba345 LE |
1779 | device->use_csums = use_checksum_based_resync(connection, device); |
1780 | } else { | |
1781 | device->use_csums = 0; | |
1782 | } | |
6c922ed5 LE |
1783 | |
1784 | /* Since protocol 96, we must serialize drbd_gen_and_send_sync_uuid | |
1785 | * with w_send_oos, or the sync target will get confused as to | |
1786 | * how much bits to resync. We cannot do that always, because for an | |
1787 | * empty resync and protocol < 95, we need to do it here, as we call | |
1788 | * drbd_resync_finished from here in that case. | |
1789 | * We drbd_gen_and_send_sync_uuid here for protocol < 96, | |
1790 | * and from after_state_ch otherwise. */ | |
44a4d551 LE |
1791 | if (side == C_SYNC_SOURCE && connection->agreed_pro_version < 96) |
1792 | drbd_gen_and_send_sync_uuid(peer_device); | |
b411b363 | 1793 | |
44a4d551 | 1794 | if (connection->agreed_pro_version < 95 && device->rs_total == 0) { |
af85e8e8 LE |
1795 | /* This still has a race (about when exactly the peers |
1796 | * detect connection loss) that can lead to a full sync | |
1797 | * on next handshake. In 8.3.9 we fixed this with explicit | |
1798 | * resync-finished notifications, but the fix | |
1799 | * introduces a protocol change. Sleeping for some | |
1800 | * time longer than the ping interval + timeout on the | |
1801 | * SyncSource, to give the SyncTarget the chance to | |
1802 | * detect connection loss, then waiting for a ping | |
1803 | * response (implicit in drbd_resync_finished) reduces | |
1804 | * the race considerably, but does not solve it. */ | |
44ed167d PR |
1805 | if (side == C_SYNC_SOURCE) { |
1806 | struct net_conf *nc; | |
1807 | int timeo; | |
1808 | ||
1809 | rcu_read_lock(); | |
44a4d551 | 1810 | nc = rcu_dereference(connection->net_conf); |
44ed167d PR |
1811 | timeo = nc->ping_int * HZ + nc->ping_timeo * HZ / 9; |
1812 | rcu_read_unlock(); | |
1813 | schedule_timeout_interruptible(timeo); | |
1814 | } | |
b30ab791 | 1815 | drbd_resync_finished(device); |
b411b363 PR |
1816 | } |
1817 | ||
b30ab791 AG |
1818 | drbd_rs_controller_reset(device); |
1819 | /* ns.conn may already be != device->state.conn, | |
b411b363 PR |
1820 | * we may have been paused in between, or become paused until |
1821 | * the timer triggers. | |
1822 | * No matter, that is handled in resync_timer_fn() */ | |
1823 | if (ns.conn == C_SYNC_TARGET) | |
b30ab791 | 1824 | mod_timer(&device->resync_timer, jiffies); |
b411b363 | 1825 | |
b30ab791 | 1826 | drbd_md_sync(device); |
b411b363 | 1827 | } |
b30ab791 AG |
1828 | put_ldev(device); |
1829 | mutex_unlock(device->state_mutex); | |
b411b363 PR |
1830 | } |
1831 | ||
e334f550 | 1832 | static void update_on_disk_bitmap(struct drbd_device *device, bool resync_done) |
c7a58db4 LE |
1833 | { |
1834 | struct sib_info sib = { .sib_reason = SIB_SYNC_PROGRESS, }; | |
1835 | device->rs_last_bcast = jiffies; | |
1836 | ||
1837 | if (!get_ldev(device)) | |
1838 | return; | |
1839 | ||
1840 | drbd_bm_write_lazy(device, 0); | |
5ab7d2c0 | 1841 | if (resync_done && is_sync_state(device->state.conn)) |
c7a58db4 | 1842 | drbd_resync_finished(device); |
5ab7d2c0 | 1843 | |
c7a58db4 LE |
1844 | drbd_bcast_event(device, &sib); |
1845 | /* update timestamp, in case it took a while to write out stuff */ | |
1846 | device->rs_last_bcast = jiffies; | |
1847 | put_ldev(device); | |
1848 | } | |
1849 | ||
e334f550 LE |
1850 | static void drbd_ldev_destroy(struct drbd_device *device) |
1851 | { | |
1852 | lc_destroy(device->resync); | |
1853 | device->resync = NULL; | |
1854 | lc_destroy(device->act_log); | |
1855 | device->act_log = NULL; | |
d1b80853 AG |
1856 | |
1857 | __acquire(local); | |
1858 | drbd_free_ldev(device->ldev); | |
1859 | device->ldev = NULL; | |
1860 | __release(local); | |
1861 | ||
e334f550 LE |
1862 | clear_bit(GOING_DISKLESS, &device->flags); |
1863 | wake_up(&device->misc_wait); | |
1864 | } | |
1865 | ||
1866 | static void go_diskless(struct drbd_device *device) | |
1867 | { | |
1868 | D_ASSERT(device, device->state.disk == D_FAILED); | |
1869 | /* we cannot assert local_cnt == 0 here, as get_ldev_if_state will | |
1870 | * inc/dec it frequently. Once we are D_DISKLESS, no one will touch | |
1871 | * the protected members anymore, though, so once put_ldev reaches zero | |
1872 | * again, it will be safe to free them. */ | |
1873 | ||
1874 | /* Try to write changed bitmap pages, read errors may have just | |
1875 | * set some bits outside the area covered by the activity log. | |
1876 | * | |
1877 | * If we have an IO error during the bitmap writeout, | |
1878 | * we will want a full sync next time, just in case. | |
1879 | * (Do we want a specific meta data flag for this?) | |
1880 | * | |
1881 | * If that does not make it to stable storage either, | |
1882 | * we cannot do anything about that anymore. | |
1883 | * | |
1884 | * We still need to check if both bitmap and ldev are present, we may | |
1885 | * end up here after a failed attach, before ldev was even assigned. | |
1886 | */ | |
1887 | if (device->bitmap && device->ldev) { | |
1888 | /* An interrupted resync or similar is allowed to recounts bits | |
1889 | * while we detach. | |
1890 | * Any modifications would not be expected anymore, though. | |
1891 | */ | |
1892 | if (drbd_bitmap_io_from_worker(device, drbd_bm_write, | |
1893 | "detach", BM_LOCKED_TEST_ALLOWED)) { | |
1894 | if (test_bit(WAS_READ_ERROR, &device->flags)) { | |
1895 | drbd_md_set_flag(device, MDF_FULL_SYNC); | |
1896 | drbd_md_sync(device); | |
1897 | } | |
1898 | } | |
1899 | } | |
1900 | ||
1901 | drbd_force_state(device, NS(disk, D_DISKLESS)); | |
1902 | } | |
1903 | ||
ac0acb9e LE |
1904 | static int do_md_sync(struct drbd_device *device) |
1905 | { | |
1906 | drbd_warn(device, "md_sync_timer expired! Worker calls drbd_md_sync().\n"); | |
1907 | drbd_md_sync(device); | |
1908 | return 0; | |
1909 | } | |
1910 | ||
944410e9 LE |
1911 | /* only called from drbd_worker thread, no locking */ |
1912 | void __update_timing_details( | |
1913 | struct drbd_thread_timing_details *tdp, | |
1914 | unsigned int *cb_nr, | |
1915 | void *cb, | |
1916 | const char *fn, const unsigned int line) | |
1917 | { | |
1918 | unsigned int i = *cb_nr % DRBD_THREAD_DETAILS_HIST; | |
1919 | struct drbd_thread_timing_details *td = tdp + i; | |
1920 | ||
1921 | td->start_jif = jiffies; | |
1922 | td->cb_addr = cb; | |
1923 | td->caller_fn = fn; | |
1924 | td->line = line; | |
1925 | td->cb_nr = *cb_nr; | |
1926 | ||
1927 | i = (i+1) % DRBD_THREAD_DETAILS_HIST; | |
1928 | td = tdp + i; | |
1929 | memset(td, 0, sizeof(*td)); | |
1930 | ||
1931 | ++(*cb_nr); | |
1932 | } | |
1933 | ||
e334f550 LE |
1934 | static void do_device_work(struct drbd_device *device, const unsigned long todo) |
1935 | { | |
b47a06d1 | 1936 | if (test_bit(MD_SYNC, &todo)) |
ac0acb9e | 1937 | do_md_sync(device); |
b47a06d1 AG |
1938 | if (test_bit(RS_DONE, &todo) || |
1939 | test_bit(RS_PROGRESS, &todo)) | |
1940 | update_on_disk_bitmap(device, test_bit(RS_DONE, &todo)); | |
1941 | if (test_bit(GO_DISKLESS, &todo)) | |
e334f550 | 1942 | go_diskless(device); |
b47a06d1 | 1943 | if (test_bit(DESTROY_DISK, &todo)) |
e334f550 | 1944 | drbd_ldev_destroy(device); |
b47a06d1 | 1945 | if (test_bit(RS_START, &todo)) |
ac0acb9e | 1946 | do_start_resync(device); |
e334f550 LE |
1947 | } |
1948 | ||
1949 | #define DRBD_DEVICE_WORK_MASK \ | |
1950 | ((1UL << GO_DISKLESS) \ | |
1951 | |(1UL << DESTROY_DISK) \ | |
ac0acb9e LE |
1952 | |(1UL << MD_SYNC) \ |
1953 | |(1UL << RS_START) \ | |
e334f550 LE |
1954 | |(1UL << RS_PROGRESS) \ |
1955 | |(1UL << RS_DONE) \ | |
1956 | ) | |
1957 | ||
1958 | static unsigned long get_work_bits(unsigned long *flags) | |
1959 | { | |
1960 | unsigned long old, new; | |
1961 | do { | |
1962 | old = *flags; | |
1963 | new = old & ~DRBD_DEVICE_WORK_MASK; | |
1964 | } while (cmpxchg(flags, old, new) != old); | |
1965 | return old & DRBD_DEVICE_WORK_MASK; | |
1966 | } | |
1967 | ||
1968 | static void do_unqueued_work(struct drbd_connection *connection) | |
c7a58db4 LE |
1969 | { |
1970 | struct drbd_peer_device *peer_device; | |
1971 | int vnr; | |
1972 | ||
1973 | rcu_read_lock(); | |
1974 | idr_for_each_entry(&connection->peer_devices, peer_device, vnr) { | |
1975 | struct drbd_device *device = peer_device->device; | |
e334f550 LE |
1976 | unsigned long todo = get_work_bits(&device->flags); |
1977 | if (!todo) | |
c7a58db4 | 1978 | continue; |
5ab7d2c0 | 1979 | |
c7a58db4 LE |
1980 | kref_get(&device->kref); |
1981 | rcu_read_unlock(); | |
e334f550 | 1982 | do_device_work(device, todo); |
c7a58db4 LE |
1983 | kref_put(&device->kref, drbd_destroy_device); |
1984 | rcu_read_lock(); | |
1985 | } | |
1986 | rcu_read_unlock(); | |
1987 | } | |
1988 | ||
a186e478 | 1989 | static bool dequeue_work_batch(struct drbd_work_queue *queue, struct list_head *work_list) |
8c0785a5 LE |
1990 | { |
1991 | spin_lock_irq(&queue->q_lock); | |
15e26f6a | 1992 | list_splice_tail_init(&queue->q, work_list); |
8c0785a5 LE |
1993 | spin_unlock_irq(&queue->q_lock); |
1994 | return !list_empty(work_list); | |
1995 | } | |
1996 | ||
bde89a9e | 1997 | static void wait_for_work(struct drbd_connection *connection, struct list_head *work_list) |
b6dd1a89 LE |
1998 | { |
1999 | DEFINE_WAIT(wait); | |
2000 | struct net_conf *nc; | |
2001 | int uncork, cork; | |
2002 | ||
abde9cc6 | 2003 | dequeue_work_batch(&connection->sender_work, work_list); |
b6dd1a89 LE |
2004 | if (!list_empty(work_list)) |
2005 | return; | |
2006 | ||
2007 | /* Still nothing to do? | |
2008 | * Maybe we still need to close the current epoch, | |
2009 | * even if no new requests are queued yet. | |
2010 | * | |
2011 | * Also, poke TCP, just in case. | |
2012 | * Then wait for new work (or signal). */ | |
2013 | rcu_read_lock(); | |
2014 | nc = rcu_dereference(connection->net_conf); | |
2015 | uncork = nc ? nc->tcp_cork : 0; | |
2016 | rcu_read_unlock(); | |
2017 | if (uncork) { | |
2018 | mutex_lock(&connection->data.mutex); | |
2019 | if (connection->data.socket) | |
2020 | drbd_tcp_uncork(connection->data.socket); | |
2021 | mutex_unlock(&connection->data.mutex); | |
2022 | } | |
2023 | ||
2024 | for (;;) { | |
2025 | int send_barrier; | |
2026 | prepare_to_wait(&connection->sender_work.q_wait, &wait, TASK_INTERRUPTIBLE); | |
0500813f | 2027 | spin_lock_irq(&connection->resource->req_lock); |
b6dd1a89 | 2028 | spin_lock(&connection->sender_work.q_lock); /* FIXME get rid of this one? */ |
bc317a9e LE |
2029 | /* dequeue single item only, |
2030 | * we still use drbd_queue_work_front() in some places */ | |
2031 | if (!list_empty(&connection->sender_work.q)) | |
4dd726f0 | 2032 | list_splice_tail_init(&connection->sender_work.q, work_list); |
b6dd1a89 LE |
2033 | spin_unlock(&connection->sender_work.q_lock); /* FIXME get rid of this one? */ |
2034 | if (!list_empty(work_list) || signal_pending(current)) { | |
0500813f | 2035 | spin_unlock_irq(&connection->resource->req_lock); |
b6dd1a89 LE |
2036 | break; |
2037 | } | |
f9c78128 LE |
2038 | |
2039 | /* We found nothing new to do, no to-be-communicated request, | |
2040 | * no other work item. We may still need to close the last | |
2041 | * epoch. Next incoming request epoch will be connection -> | |
2042 | * current transfer log epoch number. If that is different | |
2043 | * from the epoch of the last request we communicated, it is | |
2044 | * safe to send the epoch separating barrier now. | |
2045 | */ | |
2046 | send_barrier = | |
2047 | atomic_read(&connection->current_tle_nr) != | |
2048 | connection->send.current_epoch_nr; | |
0500813f | 2049 | spin_unlock_irq(&connection->resource->req_lock); |
f9c78128 LE |
2050 | |
2051 | if (send_barrier) | |
2052 | maybe_send_barrier(connection, | |
2053 | connection->send.current_epoch_nr + 1); | |
5ab7d2c0 | 2054 | |
e334f550 | 2055 | if (test_bit(DEVICE_WORK_PENDING, &connection->flags)) |
5ab7d2c0 LE |
2056 | break; |
2057 | ||
a80ca1ae LE |
2058 | /* drbd_send() may have called flush_signals() */ |
2059 | if (get_t_state(&connection->worker) != RUNNING) | |
2060 | break; | |
5ab7d2c0 | 2061 | |
b6dd1a89 LE |
2062 | schedule(); |
2063 | /* may be woken up for other things but new work, too, | |
2064 | * e.g. if the current epoch got closed. | |
2065 | * In which case we send the barrier above. */ | |
2066 | } | |
2067 | finish_wait(&connection->sender_work.q_wait, &wait); | |
2068 | ||
2069 | /* someone may have changed the config while we have been waiting above. */ | |
2070 | rcu_read_lock(); | |
2071 | nc = rcu_dereference(connection->net_conf); | |
2072 | cork = nc ? nc->tcp_cork : 0; | |
2073 | rcu_read_unlock(); | |
2074 | mutex_lock(&connection->data.mutex); | |
2075 | if (connection->data.socket) { | |
2076 | if (cork) | |
2077 | drbd_tcp_cork(connection->data.socket); | |
2078 | else if (!uncork) | |
2079 | drbd_tcp_uncork(connection->data.socket); | |
2080 | } | |
2081 | mutex_unlock(&connection->data.mutex); | |
2082 | } | |
2083 | ||
b411b363 PR |
2084 | int drbd_worker(struct drbd_thread *thi) |
2085 | { | |
bde89a9e | 2086 | struct drbd_connection *connection = thi->connection; |
6db7e50a | 2087 | struct drbd_work *w = NULL; |
c06ece6b | 2088 | struct drbd_peer_device *peer_device; |
b411b363 | 2089 | LIST_HEAD(work_list); |
8c0785a5 | 2090 | int vnr; |
b411b363 | 2091 | |
e77a0a5c | 2092 | while (get_t_state(thi) == RUNNING) { |
80822284 | 2093 | drbd_thread_current_set_cpu(thi); |
b411b363 | 2094 | |
944410e9 LE |
2095 | if (list_empty(&work_list)) { |
2096 | update_worker_timing_details(connection, wait_for_work); | |
bde89a9e | 2097 | wait_for_work(connection, &work_list); |
944410e9 | 2098 | } |
b411b363 | 2099 | |
944410e9 LE |
2100 | if (test_and_clear_bit(DEVICE_WORK_PENDING, &connection->flags)) { |
2101 | update_worker_timing_details(connection, do_unqueued_work); | |
e334f550 | 2102 | do_unqueued_work(connection); |
944410e9 | 2103 | } |
5ab7d2c0 | 2104 | |
8c0785a5 | 2105 | if (signal_pending(current)) { |
b411b363 | 2106 | flush_signals(current); |
19393e10 | 2107 | if (get_t_state(thi) == RUNNING) { |
1ec861eb | 2108 | drbd_warn(connection, "Worker got an unexpected signal\n"); |
b411b363 | 2109 | continue; |
19393e10 | 2110 | } |
b411b363 PR |
2111 | break; |
2112 | } | |
2113 | ||
e77a0a5c | 2114 | if (get_t_state(thi) != RUNNING) |
b411b363 | 2115 | break; |
b411b363 | 2116 | |
8c0785a5 | 2117 | while (!list_empty(&work_list)) { |
6db7e50a AG |
2118 | w = list_first_entry(&work_list, struct drbd_work, list); |
2119 | list_del_init(&w->list); | |
944410e9 | 2120 | update_worker_timing_details(connection, w->cb); |
6db7e50a | 2121 | if (w->cb(w, connection->cstate < C_WF_REPORT_PARAMS) == 0) |
8c0785a5 | 2122 | continue; |
bde89a9e AG |
2123 | if (connection->cstate >= C_WF_REPORT_PARAMS) |
2124 | conn_request_state(connection, NS(conn, C_NETWORK_FAILURE), CS_HARD); | |
b411b363 PR |
2125 | } |
2126 | } | |
b411b363 | 2127 | |
8c0785a5 | 2128 | do { |
944410e9 LE |
2129 | if (test_and_clear_bit(DEVICE_WORK_PENDING, &connection->flags)) { |
2130 | update_worker_timing_details(connection, do_unqueued_work); | |
e334f550 | 2131 | do_unqueued_work(connection); |
944410e9 | 2132 | } |
b411b363 | 2133 | while (!list_empty(&work_list)) { |
6db7e50a AG |
2134 | w = list_first_entry(&work_list, struct drbd_work, list); |
2135 | list_del_init(&w->list); | |
944410e9 | 2136 | update_worker_timing_details(connection, w->cb); |
6db7e50a | 2137 | w->cb(w, 1); |
b411b363 | 2138 | } |
bde89a9e | 2139 | dequeue_work_batch(&connection->sender_work, &work_list); |
e334f550 | 2140 | } while (!list_empty(&work_list) || test_bit(DEVICE_WORK_PENDING, &connection->flags)); |
b411b363 | 2141 | |
c141ebda | 2142 | rcu_read_lock(); |
c06ece6b AG |
2143 | idr_for_each_entry(&connection->peer_devices, peer_device, vnr) { |
2144 | struct drbd_device *device = peer_device->device; | |
0b0ba1ef | 2145 | D_ASSERT(device, device->state.disk == D_DISKLESS && device->state.conn == C_STANDALONE); |
b30ab791 | 2146 | kref_get(&device->kref); |
c141ebda | 2147 | rcu_read_unlock(); |
b30ab791 | 2148 | drbd_device_cleanup(device); |
05a10ec7 | 2149 | kref_put(&device->kref, drbd_destroy_device); |
c141ebda | 2150 | rcu_read_lock(); |
0e29d163 | 2151 | } |
c141ebda | 2152 | rcu_read_unlock(); |
b411b363 PR |
2153 | |
2154 | return 0; | |
2155 | } |