]>
Commit | Line | Data |
---|---|---|
f5db4af4 JB |
1 | /* |
2 | * Copyright (C) 2006-2009 Red Hat, Inc. | |
3 | * | |
4 | * This file is released under the LGPL. | |
5 | */ | |
6 | ||
7 | #include <linux/bio.h> | |
5a0e3ad6 | 8 | #include <linux/slab.h> |
0f30af98 | 9 | #include <linux/jiffies.h> |
f5db4af4 JB |
10 | #include <linux/dm-dirty-log.h> |
11 | #include <linux/device-mapper.h> | |
12 | #include <linux/dm-log-userspace.h> | |
056075c7 | 13 | #include <linux/module.h> |
5066a4df | 14 | #include <linux/workqueue.h> |
f5db4af4 JB |
15 | |
16 | #include "dm-log-userspace-transfer.h" | |
17 | ||
5066a4df | 18 | #define DM_LOG_USERSPACE_VSN "1.3.0" |
86a54a48 | 19 | |
f5db4af4 JB |
20 | struct flush_entry { |
21 | int type; | |
22 | region_t region; | |
23 | struct list_head list; | |
24 | }; | |
25 | ||
085ae065 JB |
26 | /* |
27 | * This limit on the number of mark and clear request is, to a degree, | |
28 | * arbitrary. However, there is some basis for the choice in the limits | |
29 | * imposed on the size of data payload by dm-log-userspace-transfer.c: | |
30 | * dm_consult_userspace(). | |
31 | */ | |
32 | #define MAX_FLUSH_GROUP_COUNT 32 | |
33 | ||
f5db4af4 JB |
34 | struct log_c { |
35 | struct dm_target *ti; | |
5a25f0eb | 36 | struct dm_dev *log_dev; |
f5db4af4 JB |
37 | uint32_t region_size; |
38 | region_t region_count; | |
7ec23d50 | 39 | uint64_t luid; |
f5db4af4 JB |
40 | char uuid[DM_UUID_LEN]; |
41 | ||
42 | char *usr_argv_str; | |
43 | uint32_t usr_argc; | |
44 | ||
45 | /* | |
46 | * in_sync_hint gets set when doing is_remote_recovering. It | |
47 | * represents the first region that needs recovery. IOW, the | |
48 | * first zero bit of sync_bits. This can be useful for to limit | |
49 | * traffic for calls like is_remote_recovering and get_resync_work, | |
50 | * but be take care in its use for anything else. | |
51 | */ | |
52 | uint64_t in_sync_hint; | |
53 | ||
909cc4fb JB |
54 | /* |
55 | * Mark and clear requests are held until a flush is issued | |
56 | * so that we can group, and thereby limit, the amount of | |
57 | * network traffic between kernel and userspace. The 'flush_lock' | |
58 | * is used to protect these lists. | |
59 | */ | |
f5db4af4 | 60 | spinlock_t flush_lock; |
909cc4fb JB |
61 | struct list_head mark_list; |
62 | struct list_head clear_list; | |
5066a4df DZ |
63 | |
64 | /* | |
65 | * Workqueue for flush of clear region requests. | |
66 | */ | |
67 | struct workqueue_struct *dmlog_wq; | |
68 | struct delayed_work flush_log_work; | |
69 | atomic_t sched_flush; | |
70 | ||
71 | /* | |
72 | * Combine userspace flush and mark requests for efficiency. | |
73 | */ | |
74 | uint32_t integrated_flush; | |
f5db4af4 JB |
75 | }; |
76 | ||
77 | static mempool_t *flush_entry_pool; | |
78 | ||
79 | static void *flush_entry_alloc(gfp_t gfp_mask, void *pool_data) | |
80 | { | |
81 | return kmalloc(sizeof(struct flush_entry), gfp_mask); | |
82 | } | |
83 | ||
84 | static void flush_entry_free(void *element, void *pool_data) | |
85 | { | |
86 | kfree(element); | |
87 | } | |
88 | ||
89 | static int userspace_do_request(struct log_c *lc, const char *uuid, | |
90 | int request_type, char *data, size_t data_size, | |
91 | char *rdata, size_t *rdata_size) | |
92 | { | |
93 | int r; | |
94 | ||
95 | /* | |
96 | * If the server isn't there, -ESRCH is returned, | |
97 | * and we must keep trying until the server is | |
98 | * restored. | |
99 | */ | |
100 | retry: | |
7ec23d50 | 101 | r = dm_consult_userspace(uuid, lc->luid, request_type, data, |
f5db4af4 JB |
102 | data_size, rdata, rdata_size); |
103 | ||
104 | if (r != -ESRCH) | |
105 | return r; | |
106 | ||
107 | DMERR(" Userspace log server not found."); | |
108 | while (1) { | |
109 | set_current_state(TASK_INTERRUPTIBLE); | |
110 | schedule_timeout(2*HZ); | |
111 | DMWARN("Attempting to contact userspace log server..."); | |
7ec23d50 JB |
112 | r = dm_consult_userspace(uuid, lc->luid, DM_ULOG_CTR, |
113 | lc->usr_argv_str, | |
f5db4af4 JB |
114 | strlen(lc->usr_argv_str) + 1, |
115 | NULL, NULL); | |
116 | if (!r) | |
117 | break; | |
118 | } | |
119 | DMINFO("Reconnected to userspace log server... DM_ULOG_CTR complete"); | |
7ec23d50 | 120 | r = dm_consult_userspace(uuid, lc->luid, DM_ULOG_RESUME, NULL, |
f5db4af4 JB |
121 | 0, NULL, NULL); |
122 | if (!r) | |
123 | goto retry; | |
124 | ||
125 | DMERR("Error trying to resume userspace log: %d", r); | |
126 | ||
127 | return -ESRCH; | |
128 | } | |
129 | ||
130 | static int build_constructor_string(struct dm_target *ti, | |
131 | unsigned argc, char **argv, | |
132 | char **ctr_str) | |
133 | { | |
134 | int i, str_size; | |
135 | char *str = NULL; | |
136 | ||
137 | *ctr_str = NULL; | |
138 | ||
5066a4df DZ |
139 | /* |
140 | * Determine overall size of the string. | |
141 | */ | |
f5db4af4 JB |
142 | for (i = 0, str_size = 0; i < argc; i++) |
143 | str_size += strlen(argv[i]) + 1; /* +1 for space between args */ | |
144 | ||
145 | str_size += 20; /* Max number of chars in a printed u64 number */ | |
146 | ||
147 | str = kzalloc(str_size, GFP_KERNEL); | |
148 | if (!str) { | |
149 | DMWARN("Unable to allocate memory for constructor string"); | |
150 | return -ENOMEM; | |
151 | } | |
152 | ||
b8313b6d JB |
153 | str_size = sprintf(str, "%llu", (unsigned long long)ti->len); |
154 | for (i = 0; i < argc; i++) | |
155 | str_size += sprintf(str + str_size, " %s", argv[i]); | |
f5db4af4 JB |
156 | |
157 | *ctr_str = str; | |
158 | return str_size; | |
159 | } | |
160 | ||
5066a4df DZ |
161 | static void do_flush(struct work_struct *work) |
162 | { | |
163 | int r; | |
164 | struct log_c *lc = container_of(work, struct log_c, flush_log_work.work); | |
165 | ||
166 | atomic_set(&lc->sched_flush, 0); | |
167 | ||
168 | r = userspace_do_request(lc, lc->uuid, DM_ULOG_FLUSH, NULL, 0, NULL, NULL); | |
169 | ||
170 | if (r) | |
171 | dm_table_event(lc->ti->table); | |
172 | } | |
173 | ||
f5db4af4 JB |
174 | /* |
175 | * userspace_ctr | |
176 | * | |
177 | * argv contains: | |
5066a4df DZ |
178 | * <UUID> [integrated_flush] <other args> |
179 | * Where 'other args' are the userspace implementation-specific log | |
180 | * arguments. | |
181 | * | |
182 | * Example: | |
183 | * <UUID> [integrated_flush] clustered-disk <arg count> <log dev> | |
184 | * <region_size> [[no]sync] | |
185 | * | |
186 | * This module strips off the <UUID> and uses it for identification | |
187 | * purposes when communicating with userspace about a log. | |
f5db4af4 | 188 | * |
5066a4df DZ |
189 | * If integrated_flush is defined, the kernel combines flush |
190 | * and mark requests. | |
191 | * | |
192 | * The rest of the line, beginning with 'clustered-disk', is passed | |
193 | * to the userspace ctr function. | |
f5db4af4 JB |
194 | */ |
195 | static int userspace_ctr(struct dm_dirty_log *log, struct dm_target *ti, | |
196 | unsigned argc, char **argv) | |
197 | { | |
198 | int r = 0; | |
199 | int str_size; | |
200 | char *ctr_str = NULL; | |
201 | struct log_c *lc = NULL; | |
202 | uint64_t rdata; | |
203 | size_t rdata_size = sizeof(rdata); | |
5a25f0eb JB |
204 | char *devices_rdata = NULL; |
205 | size_t devices_rdata_size = DM_NAME_LEN; | |
f5db4af4 JB |
206 | |
207 | if (argc < 3) { | |
208 | DMWARN("Too few arguments to userspace dirty log"); | |
209 | return -EINVAL; | |
210 | } | |
211 | ||
5a25f0eb | 212 | lc = kzalloc(sizeof(*lc), GFP_KERNEL); |
f5db4af4 JB |
213 | if (!lc) { |
214 | DMWARN("Unable to allocate userspace log context."); | |
215 | return -ENOMEM; | |
216 | } | |
217 | ||
7ec23d50 | 218 | /* The ptr value is sufficient for local unique id */ |
bca915aa | 219 | lc->luid = (unsigned long)lc; |
7ec23d50 | 220 | |
f5db4af4 JB |
221 | lc->ti = ti; |
222 | ||
223 | if (strlen(argv[0]) > (DM_UUID_LEN - 1)) { | |
224 | DMWARN("UUID argument too long."); | |
225 | kfree(lc); | |
226 | return -EINVAL; | |
227 | } | |
228 | ||
5066a4df DZ |
229 | lc->usr_argc = argc; |
230 | ||
f5db4af4 | 231 | strncpy(lc->uuid, argv[0], DM_UUID_LEN); |
5066a4df DZ |
232 | argc--; |
233 | argv++; | |
f5db4af4 | 234 | spin_lock_init(&lc->flush_lock); |
909cc4fb JB |
235 | INIT_LIST_HEAD(&lc->mark_list); |
236 | INIT_LIST_HEAD(&lc->clear_list); | |
f5db4af4 | 237 | |
5066a4df DZ |
238 | if (!strcasecmp(argv[0], "integrated_flush")) { |
239 | lc->integrated_flush = 1; | |
240 | argc--; | |
241 | argv++; | |
242 | } | |
243 | ||
244 | str_size = build_constructor_string(ti, argc, argv, &ctr_str); | |
f5db4af4 JB |
245 | if (str_size < 0) { |
246 | kfree(lc); | |
247 | return str_size; | |
248 | } | |
249 | ||
5a25f0eb JB |
250 | devices_rdata = kzalloc(devices_rdata_size, GFP_KERNEL); |
251 | if (!devices_rdata) { | |
252 | DMERR("Failed to allocate memory for device information"); | |
253 | r = -ENOMEM; | |
254 | goto out; | |
255 | } | |
256 | ||
257 | /* | |
258 | * Send table string and get back any opened device. | |
259 | */ | |
7ec23d50 | 260 | r = dm_consult_userspace(lc->uuid, lc->luid, DM_ULOG_CTR, |
5a25f0eb JB |
261 | ctr_str, str_size, |
262 | devices_rdata, &devices_rdata_size); | |
f5db4af4 | 263 | |
4a038677 JB |
264 | if (r < 0) { |
265 | if (r == -ESRCH) | |
266 | DMERR("Userspace log server not found"); | |
267 | else | |
268 | DMERR("Userspace log server failed to create log"); | |
f5db4af4 JB |
269 | goto out; |
270 | } | |
271 | ||
272 | /* Since the region size does not change, get it now */ | |
273 | rdata_size = sizeof(rdata); | |
7ec23d50 | 274 | r = dm_consult_userspace(lc->uuid, lc->luid, DM_ULOG_GET_REGION_SIZE, |
f5db4af4 JB |
275 | NULL, 0, (char *)&rdata, &rdata_size); |
276 | ||
277 | if (r) { | |
278 | DMERR("Failed to get region size of dirty log"); | |
279 | goto out; | |
280 | } | |
281 | ||
282 | lc->region_size = (uint32_t)rdata; | |
283 | lc->region_count = dm_sector_div_up(ti->len, lc->region_size); | |
284 | ||
5a25f0eb JB |
285 | if (devices_rdata_size) { |
286 | if (devices_rdata[devices_rdata_size - 1] != '\0') { | |
287 | DMERR("DM_ULOG_CTR device return string not properly terminated"); | |
288 | r = -EINVAL; | |
289 | goto out; | |
290 | } | |
291 | r = dm_get_device(ti, devices_rdata, | |
292 | dm_table_get_mode(ti->table), &lc->log_dev); | |
293 | if (r) | |
294 | DMERR("Failed to register %s with device-mapper", | |
295 | devices_rdata); | |
296 | } | |
5066a4df DZ |
297 | |
298 | if (lc->integrated_flush) { | |
299 | lc->dmlog_wq = alloc_workqueue("dmlogd", WQ_MEM_RECLAIM, 0); | |
300 | if (!lc->dmlog_wq) { | |
301 | DMERR("couldn't start dmlogd"); | |
302 | r = -ENOMEM; | |
303 | goto out; | |
304 | } | |
305 | ||
306 | INIT_DELAYED_WORK(&lc->flush_log_work, do_flush); | |
307 | atomic_set(&lc->sched_flush, 0); | |
308 | } | |
309 | ||
f5db4af4 | 310 | out: |
5a25f0eb | 311 | kfree(devices_rdata); |
f5db4af4 JB |
312 | if (r) { |
313 | kfree(lc); | |
314 | kfree(ctr_str); | |
315 | } else { | |
316 | lc->usr_argv_str = ctr_str; | |
f5db4af4 JB |
317 | log->context = lc; |
318 | } | |
319 | ||
320 | return r; | |
321 | } | |
322 | ||
323 | static void userspace_dtr(struct dm_dirty_log *log) | |
324 | { | |
f5db4af4 JB |
325 | struct log_c *lc = log->context; |
326 | ||
5066a4df DZ |
327 | if (lc->integrated_flush) { |
328 | /* flush workqueue */ | |
329 | if (atomic_read(&lc->sched_flush)) | |
330 | flush_delayed_work(&lc->flush_log_work); | |
331 | ||
332 | destroy_workqueue(lc->dmlog_wq); | |
333 | } | |
334 | ||
4a038677 | 335 | (void) dm_consult_userspace(lc->uuid, lc->luid, DM_ULOG_DTR, |
5066a4df | 336 | NULL, 0, NULL, NULL); |
f5db4af4 | 337 | |
5a25f0eb JB |
338 | if (lc->log_dev) |
339 | dm_put_device(lc->ti, lc->log_dev); | |
340 | ||
f5db4af4 JB |
341 | kfree(lc->usr_argv_str); |
342 | kfree(lc); | |
343 | ||
344 | return; | |
345 | } | |
346 | ||
347 | static int userspace_presuspend(struct dm_dirty_log *log) | |
348 | { | |
349 | int r; | |
350 | struct log_c *lc = log->context; | |
351 | ||
7ec23d50 | 352 | r = dm_consult_userspace(lc->uuid, lc->luid, DM_ULOG_PRESUSPEND, |
5066a4df | 353 | NULL, 0, NULL, NULL); |
f5db4af4 JB |
354 | |
355 | return r; | |
356 | } | |
357 | ||
358 | static int userspace_postsuspend(struct dm_dirty_log *log) | |
359 | { | |
360 | int r; | |
361 | struct log_c *lc = log->context; | |
362 | ||
5066a4df DZ |
363 | /* |
364 | * Run planned flush earlier. | |
365 | */ | |
366 | if (lc->integrated_flush && atomic_read(&lc->sched_flush)) | |
367 | flush_delayed_work(&lc->flush_log_work); | |
368 | ||
7ec23d50 | 369 | r = dm_consult_userspace(lc->uuid, lc->luid, DM_ULOG_POSTSUSPEND, |
5066a4df | 370 | NULL, 0, NULL, NULL); |
f5db4af4 JB |
371 | |
372 | return r; | |
373 | } | |
374 | ||
375 | static int userspace_resume(struct dm_dirty_log *log) | |
376 | { | |
377 | int r; | |
378 | struct log_c *lc = log->context; | |
379 | ||
380 | lc->in_sync_hint = 0; | |
7ec23d50 | 381 | r = dm_consult_userspace(lc->uuid, lc->luid, DM_ULOG_RESUME, |
5066a4df | 382 | NULL, 0, NULL, NULL); |
f5db4af4 JB |
383 | |
384 | return r; | |
385 | } | |
386 | ||
387 | static uint32_t userspace_get_region_size(struct dm_dirty_log *log) | |
388 | { | |
389 | struct log_c *lc = log->context; | |
390 | ||
391 | return lc->region_size; | |
392 | } | |
393 | ||
394 | /* | |
395 | * userspace_is_clean | |
396 | * | |
397 | * Check whether a region is clean. If there is any sort of | |
398 | * failure when consulting the server, we return not clean. | |
399 | * | |
400 | * Returns: 1 if clean, 0 otherwise | |
401 | */ | |
402 | static int userspace_is_clean(struct dm_dirty_log *log, region_t region) | |
403 | { | |
404 | int r; | |
405 | uint64_t region64 = (uint64_t)region; | |
406 | int64_t is_clean; | |
407 | size_t rdata_size; | |
408 | struct log_c *lc = log->context; | |
409 | ||
410 | rdata_size = sizeof(is_clean); | |
411 | r = userspace_do_request(lc, lc->uuid, DM_ULOG_IS_CLEAN, | |
412 | (char *)®ion64, sizeof(region64), | |
413 | (char *)&is_clean, &rdata_size); | |
414 | ||
415 | return (r) ? 0 : (int)is_clean; | |
416 | } | |
417 | ||
418 | /* | |
419 | * userspace_in_sync | |
420 | * | |
421 | * Check if the region is in-sync. If there is any sort | |
422 | * of failure when consulting the server, we assume that | |
423 | * the region is not in sync. | |
424 | * | |
425 | * If 'can_block' is set, return immediately | |
426 | * | |
427 | * Returns: 1 if in-sync, 0 if not-in-sync, -EWOULDBLOCK | |
428 | */ | |
429 | static int userspace_in_sync(struct dm_dirty_log *log, region_t region, | |
430 | int can_block) | |
431 | { | |
432 | int r; | |
433 | uint64_t region64 = region; | |
434 | int64_t in_sync; | |
435 | size_t rdata_size; | |
436 | struct log_c *lc = log->context; | |
437 | ||
438 | /* | |
439 | * We can never respond directly - even if in_sync_hint is | |
440 | * set. This is because another machine could see a device | |
441 | * failure and mark the region out-of-sync. If we don't go | |
442 | * to userspace to ask, we might think the region is in-sync | |
443 | * and allow a read to pick up data that is stale. (This is | |
444 | * very unlikely if a device actually fails; but it is very | |
445 | * likely if a connection to one device from one machine fails.) | |
446 | * | |
447 | * There still might be a problem if the mirror caches the region | |
448 | * state as in-sync... but then this call would not be made. So, | |
449 | * that is a mirror problem. | |
450 | */ | |
451 | if (!can_block) | |
452 | return -EWOULDBLOCK; | |
453 | ||
454 | rdata_size = sizeof(in_sync); | |
455 | r = userspace_do_request(lc, lc->uuid, DM_ULOG_IN_SYNC, | |
456 | (char *)®ion64, sizeof(region64), | |
457 | (char *)&in_sync, &rdata_size); | |
458 | return (r) ? 0 : (int)in_sync; | |
459 | } | |
460 | ||
085ae065 JB |
461 | static int flush_one_by_one(struct log_c *lc, struct list_head *flush_list) |
462 | { | |
463 | int r = 0; | |
464 | struct flush_entry *fe; | |
465 | ||
466 | list_for_each_entry(fe, flush_list, list) { | |
467 | r = userspace_do_request(lc, lc->uuid, fe->type, | |
468 | (char *)&fe->region, | |
469 | sizeof(fe->region), | |
470 | NULL, NULL); | |
471 | if (r) | |
472 | break; | |
473 | } | |
474 | ||
475 | return r; | |
476 | } | |
477 | ||
5066a4df DZ |
478 | static int flush_by_group(struct log_c *lc, struct list_head *flush_list, |
479 | int flush_with_payload) | |
085ae065 JB |
480 | { |
481 | int r = 0; | |
482 | int count; | |
483 | uint32_t type = 0; | |
484 | struct flush_entry *fe, *tmp_fe; | |
485 | LIST_HEAD(tmp_list); | |
486 | uint64_t group[MAX_FLUSH_GROUP_COUNT]; | |
487 | ||
488 | /* | |
489 | * Group process the requests | |
490 | */ | |
491 | while (!list_empty(flush_list)) { | |
492 | count = 0; | |
493 | ||
494 | list_for_each_entry_safe(fe, tmp_fe, flush_list, list) { | |
495 | group[count] = fe->region; | |
496 | count++; | |
497 | ||
6c9b27ab | 498 | list_move(&fe->list, &tmp_list); |
085ae065 JB |
499 | |
500 | type = fe->type; | |
501 | if (count >= MAX_FLUSH_GROUP_COUNT) | |
502 | break; | |
503 | } | |
504 | ||
5066a4df DZ |
505 | if (flush_with_payload) { |
506 | r = userspace_do_request(lc, lc->uuid, DM_ULOG_FLUSH, | |
507 | (char *)(group), | |
508 | count * sizeof(uint64_t), | |
509 | NULL, NULL); | |
510 | /* | |
511 | * Integrated flush failed. | |
512 | */ | |
513 | if (r) | |
514 | break; | |
515 | } else { | |
516 | r = userspace_do_request(lc, lc->uuid, type, | |
517 | (char *)(group), | |
518 | count * sizeof(uint64_t), | |
519 | NULL, NULL); | |
520 | if (r) { | |
521 | /* | |
522 | * Group send failed. Attempt one-by-one. | |
523 | */ | |
524 | list_splice_init(&tmp_list, flush_list); | |
525 | r = flush_one_by_one(lc, flush_list); | |
526 | break; | |
527 | } | |
085ae065 JB |
528 | } |
529 | } | |
530 | ||
531 | /* | |
532 | * Must collect flush_entrys that were successfully processed | |
533 | * as a group so that they will be free'd by the caller. | |
534 | */ | |
535 | list_splice_init(&tmp_list, flush_list); | |
536 | ||
537 | return r; | |
538 | } | |
539 | ||
f5db4af4 JB |
540 | /* |
541 | * userspace_flush | |
542 | * | |
543 | * This function is ok to block. | |
544 | * The flush happens in two stages. First, it sends all | |
545 | * clear/mark requests that are on the list. Then it | |
546 | * tells the server to commit them. This gives the | |
547 | * server a chance to optimise the commit, instead of | |
548 | * doing it for every request. | |
549 | * | |
550 | * Additionally, we could implement another thread that | |
551 | * sends the requests up to the server - reducing the | |
552 | * load on flush. Then the flush would have less in | |
553 | * the list and be responsible for the finishing commit. | |
554 | * | |
555 | * Returns: 0 on success, < 0 on failure | |
556 | */ | |
557 | static int userspace_flush(struct dm_dirty_log *log) | |
558 | { | |
559 | int r = 0; | |
560 | unsigned long flags; | |
561 | struct log_c *lc = log->context; | |
909cc4fb JB |
562 | LIST_HEAD(mark_list); |
563 | LIST_HEAD(clear_list); | |
5066a4df DZ |
564 | int mark_list_is_empty; |
565 | int clear_list_is_empty; | |
f5db4af4 JB |
566 | struct flush_entry *fe, *tmp_fe; |
567 | ||
568 | spin_lock_irqsave(&lc->flush_lock, flags); | |
909cc4fb JB |
569 | list_splice_init(&lc->mark_list, &mark_list); |
570 | list_splice_init(&lc->clear_list, &clear_list); | |
f5db4af4 JB |
571 | spin_unlock_irqrestore(&lc->flush_lock, flags); |
572 | ||
5066a4df DZ |
573 | mark_list_is_empty = list_empty(&mark_list); |
574 | clear_list_is_empty = list_empty(&clear_list); | |
575 | ||
576 | if (mark_list_is_empty && clear_list_is_empty) | |
f5db4af4 JB |
577 | return 0; |
578 | ||
5066a4df | 579 | r = flush_by_group(lc, &clear_list, 0); |
085ae065 | 580 | if (r) |
5066a4df DZ |
581 | goto out; |
582 | ||
583 | if (!lc->integrated_flush) { | |
584 | r = flush_by_group(lc, &mark_list, 0); | |
585 | if (r) | |
586 | goto out; | |
587 | r = userspace_do_request(lc, lc->uuid, DM_ULOG_FLUSH, | |
588 | NULL, 0, NULL, NULL); | |
589 | goto out; | |
590 | } | |
909cc4fb | 591 | |
5066a4df DZ |
592 | /* |
593 | * Send integrated flush request with mark_list as payload. | |
594 | */ | |
595 | r = flush_by_group(lc, &mark_list, 1); | |
085ae065 | 596 | if (r) |
5066a4df | 597 | goto out; |
f5db4af4 | 598 | |
5066a4df DZ |
599 | if (mark_list_is_empty && !atomic_read(&lc->sched_flush)) { |
600 | /* | |
601 | * When there are only clear region requests, | |
602 | * we schedule a flush in the future. | |
603 | */ | |
604 | queue_delayed_work(lc->dmlog_wq, &lc->flush_log_work, 3 * HZ); | |
605 | atomic_set(&lc->sched_flush, 1); | |
606 | } else { | |
607 | /* | |
608 | * Cancel pending flush because we | |
609 | * have already flushed in mark_region. | |
610 | */ | |
611 | cancel_delayed_work(&lc->flush_log_work); | |
612 | atomic_set(&lc->sched_flush, 0); | |
613 | } | |
f5db4af4 | 614 | |
5066a4df | 615 | out: |
f5db4af4 | 616 | /* |
5066a4df | 617 | * We can safely remove these entries, even after failure. |
f5db4af4 JB |
618 | * Calling code will receive an error and will know that |
619 | * the log facility has failed. | |
620 | */ | |
909cc4fb JB |
621 | list_for_each_entry_safe(fe, tmp_fe, &mark_list, list) { |
622 | list_del(&fe->list); | |
623 | mempool_free(fe, flush_entry_pool); | |
624 | } | |
625 | list_for_each_entry_safe(fe, tmp_fe, &clear_list, list) { | |
f5db4af4 JB |
626 | list_del(&fe->list); |
627 | mempool_free(fe, flush_entry_pool); | |
628 | } | |
629 | ||
630 | if (r) | |
631 | dm_table_event(lc->ti->table); | |
632 | ||
633 | return r; | |
634 | } | |
635 | ||
636 | /* | |
637 | * userspace_mark_region | |
638 | * | |
639 | * This function should avoid blocking unless absolutely required. | |
640 | * (Memory allocation is valid for blocking.) | |
641 | */ | |
642 | static void userspace_mark_region(struct dm_dirty_log *log, region_t region) | |
643 | { | |
644 | unsigned long flags; | |
645 | struct log_c *lc = log->context; | |
646 | struct flush_entry *fe; | |
647 | ||
648 | /* Wait for an allocation, but _never_ fail */ | |
649 | fe = mempool_alloc(flush_entry_pool, GFP_NOIO); | |
650 | BUG_ON(!fe); | |
651 | ||
652 | spin_lock_irqsave(&lc->flush_lock, flags); | |
653 | fe->type = DM_ULOG_MARK_REGION; | |
654 | fe->region = region; | |
909cc4fb | 655 | list_add(&fe->list, &lc->mark_list); |
f5db4af4 JB |
656 | spin_unlock_irqrestore(&lc->flush_lock, flags); |
657 | ||
658 | return; | |
659 | } | |
660 | ||
661 | /* | |
662 | * userspace_clear_region | |
663 | * | |
664 | * This function must not block. | |
665 | * So, the alloc can't block. In the worst case, it is ok to | |
666 | * fail. It would simply mean we can't clear the region. | |
667 | * Does nothing to current sync context, but does mean | |
668 | * the region will be re-sync'ed on a reload of the mirror | |
669 | * even though it is in-sync. | |
670 | */ | |
671 | static void userspace_clear_region(struct dm_dirty_log *log, region_t region) | |
672 | { | |
673 | unsigned long flags; | |
674 | struct log_c *lc = log->context; | |
675 | struct flush_entry *fe; | |
676 | ||
677 | /* | |
678 | * If we fail to allocate, we skip the clearing of | |
679 | * the region. This doesn't hurt us in any way, except | |
680 | * to cause the region to be resync'ed when the | |
681 | * device is activated next time. | |
682 | */ | |
683 | fe = mempool_alloc(flush_entry_pool, GFP_ATOMIC); | |
684 | if (!fe) { | |
685 | DMERR("Failed to allocate memory to clear region."); | |
686 | return; | |
687 | } | |
688 | ||
689 | spin_lock_irqsave(&lc->flush_lock, flags); | |
690 | fe->type = DM_ULOG_CLEAR_REGION; | |
691 | fe->region = region; | |
909cc4fb | 692 | list_add(&fe->list, &lc->clear_list); |
f5db4af4 JB |
693 | spin_unlock_irqrestore(&lc->flush_lock, flags); |
694 | ||
695 | return; | |
696 | } | |
697 | ||
698 | /* | |
699 | * userspace_get_resync_work | |
700 | * | |
701 | * Get a region that needs recovery. It is valid to return | |
702 | * an error for this function. | |
703 | * | |
704 | * Returns: 1 if region filled, 0 if no work, <0 on error | |
705 | */ | |
706 | static int userspace_get_resync_work(struct dm_dirty_log *log, region_t *region) | |
707 | { | |
708 | int r; | |
709 | size_t rdata_size; | |
710 | struct log_c *lc = log->context; | |
711 | struct { | |
712 | int64_t i; /* 64-bit for mix arch compatibility */ | |
713 | region_t r; | |
714 | } pkg; | |
715 | ||
716 | if (lc->in_sync_hint >= lc->region_count) | |
717 | return 0; | |
718 | ||
719 | rdata_size = sizeof(pkg); | |
720 | r = userspace_do_request(lc, lc->uuid, DM_ULOG_GET_RESYNC_WORK, | |
5066a4df | 721 | NULL, 0, (char *)&pkg, &rdata_size); |
f5db4af4 JB |
722 | |
723 | *region = pkg.r; | |
724 | return (r) ? r : (int)pkg.i; | |
725 | } | |
726 | ||
727 | /* | |
728 | * userspace_set_region_sync | |
729 | * | |
730 | * Set the sync status of a given region. This function | |
731 | * must not fail. | |
732 | */ | |
733 | static void userspace_set_region_sync(struct dm_dirty_log *log, | |
734 | region_t region, int in_sync) | |
735 | { | |
736 | int r; | |
737 | struct log_c *lc = log->context; | |
738 | struct { | |
739 | region_t r; | |
740 | int64_t i; | |
741 | } pkg; | |
742 | ||
743 | pkg.r = region; | |
744 | pkg.i = (int64_t)in_sync; | |
745 | ||
746 | r = userspace_do_request(lc, lc->uuid, DM_ULOG_SET_REGION_SYNC, | |
5066a4df | 747 | (char *)&pkg, sizeof(pkg), NULL, NULL); |
f5db4af4 JB |
748 | |
749 | /* | |
750 | * It would be nice to be able to report failures. | |
751 | * However, it is easy emough to detect and resolve. | |
752 | */ | |
753 | return; | |
754 | } | |
755 | ||
756 | /* | |
757 | * userspace_get_sync_count | |
758 | * | |
759 | * If there is any sort of failure when consulting the server, | |
760 | * we assume that the sync count is zero. | |
761 | * | |
762 | * Returns: sync count on success, 0 on failure | |
763 | */ | |
764 | static region_t userspace_get_sync_count(struct dm_dirty_log *log) | |
765 | { | |
766 | int r; | |
767 | size_t rdata_size; | |
768 | uint64_t sync_count; | |
769 | struct log_c *lc = log->context; | |
770 | ||
771 | rdata_size = sizeof(sync_count); | |
772 | r = userspace_do_request(lc, lc->uuid, DM_ULOG_GET_SYNC_COUNT, | |
5066a4df | 773 | NULL, 0, (char *)&sync_count, &rdata_size); |
f5db4af4 JB |
774 | |
775 | if (r) | |
776 | return 0; | |
777 | ||
778 | if (sync_count >= lc->region_count) | |
779 | lc->in_sync_hint = lc->region_count; | |
780 | ||
781 | return (region_t)sync_count; | |
782 | } | |
783 | ||
784 | /* | |
785 | * userspace_status | |
786 | * | |
787 | * Returns: amount of space consumed | |
788 | */ | |
789 | static int userspace_status(struct dm_dirty_log *log, status_type_t status_type, | |
790 | char *result, unsigned maxlen) | |
791 | { | |
792 | int r = 0; | |
b8313b6d | 793 | char *table_args; |
f5db4af4 JB |
794 | size_t sz = (size_t)maxlen; |
795 | struct log_c *lc = log->context; | |
796 | ||
797 | switch (status_type) { | |
798 | case STATUSTYPE_INFO: | |
799 | r = userspace_do_request(lc, lc->uuid, DM_ULOG_STATUS_INFO, | |
5066a4df | 800 | NULL, 0, result, &sz); |
f5db4af4 JB |
801 | |
802 | if (r) { | |
803 | sz = 0; | |
804 | DMEMIT("%s 1 COM_FAILURE", log->type->name); | |
805 | } | |
806 | break; | |
807 | case STATUSTYPE_TABLE: | |
808 | sz = 0; | |
0d03d59d | 809 | table_args = strchr(lc->usr_argv_str, ' '); |
b8313b6d JB |
810 | BUG_ON(!table_args); /* There will always be a ' ' */ |
811 | table_args++; | |
812 | ||
5066a4df DZ |
813 | DMEMIT("%s %u %s ", log->type->name, lc->usr_argc, lc->uuid); |
814 | if (lc->integrated_flush) | |
815 | DMEMIT("integrated_flush "); | |
816 | DMEMIT("%s ", table_args); | |
f5db4af4 JB |
817 | break; |
818 | } | |
819 | return (r) ? 0 : (int)sz; | |
820 | } | |
821 | ||
822 | /* | |
823 | * userspace_is_remote_recovering | |
824 | * | |
825 | * Returns: 1 if region recovering, 0 otherwise | |
826 | */ | |
827 | static int userspace_is_remote_recovering(struct dm_dirty_log *log, | |
828 | region_t region) | |
829 | { | |
830 | int r; | |
831 | uint64_t region64 = region; | |
832 | struct log_c *lc = log->context; | |
0f30af98 | 833 | static unsigned long limit; |
f5db4af4 JB |
834 | struct { |
835 | int64_t is_recovering; | |
836 | uint64_t in_sync_hint; | |
837 | } pkg; | |
838 | size_t rdata_size = sizeof(pkg); | |
839 | ||
840 | /* | |
841 | * Once the mirror has been reported to be in-sync, | |
842 | * it will never again ask for recovery work. So, | |
843 | * we can safely say there is not a remote machine | |
844 | * recovering if the device is in-sync. (in_sync_hint | |
845 | * must be reset at resume time.) | |
846 | */ | |
847 | if (region < lc->in_sync_hint) | |
848 | return 0; | |
0f30af98 | 849 | else if (time_after(limit, jiffies)) |
f5db4af4 JB |
850 | return 1; |
851 | ||
852 | limit = jiffies + (HZ / 4); | |
853 | r = userspace_do_request(lc, lc->uuid, DM_ULOG_IS_REMOTE_RECOVERING, | |
854 | (char *)®ion64, sizeof(region64), | |
855 | (char *)&pkg, &rdata_size); | |
856 | if (r) | |
857 | return 1; | |
858 | ||
859 | lc->in_sync_hint = pkg.in_sync_hint; | |
860 | ||
861 | return (int)pkg.is_recovering; | |
862 | } | |
863 | ||
864 | static struct dm_dirty_log_type _userspace_type = { | |
865 | .name = "userspace", | |
866 | .module = THIS_MODULE, | |
867 | .ctr = userspace_ctr, | |
868 | .dtr = userspace_dtr, | |
869 | .presuspend = userspace_presuspend, | |
870 | .postsuspend = userspace_postsuspend, | |
871 | .resume = userspace_resume, | |
872 | .get_region_size = userspace_get_region_size, | |
873 | .is_clean = userspace_is_clean, | |
874 | .in_sync = userspace_in_sync, | |
875 | .flush = userspace_flush, | |
876 | .mark_region = userspace_mark_region, | |
877 | .clear_region = userspace_clear_region, | |
878 | .get_resync_work = userspace_get_resync_work, | |
879 | .set_region_sync = userspace_set_region_sync, | |
880 | .get_sync_count = userspace_get_sync_count, | |
881 | .status = userspace_status, | |
882 | .is_remote_recovering = userspace_is_remote_recovering, | |
883 | }; | |
884 | ||
885 | static int __init userspace_dirty_log_init(void) | |
886 | { | |
887 | int r = 0; | |
888 | ||
889 | flush_entry_pool = mempool_create(100, flush_entry_alloc, | |
890 | flush_entry_free, NULL); | |
891 | ||
892 | if (!flush_entry_pool) { | |
893 | DMWARN("Unable to create flush_entry_pool: No memory."); | |
894 | return -ENOMEM; | |
895 | } | |
896 | ||
897 | r = dm_ulog_tfr_init(); | |
898 | if (r) { | |
899 | DMWARN("Unable to initialize userspace log communications"); | |
900 | mempool_destroy(flush_entry_pool); | |
901 | return r; | |
902 | } | |
903 | ||
904 | r = dm_dirty_log_type_register(&_userspace_type); | |
905 | if (r) { | |
906 | DMWARN("Couldn't register userspace dirty log type"); | |
907 | dm_ulog_tfr_exit(); | |
908 | mempool_destroy(flush_entry_pool); | |
909 | return r; | |
910 | } | |
911 | ||
86a54a48 | 912 | DMINFO("version " DM_LOG_USERSPACE_VSN " loaded"); |
f5db4af4 JB |
913 | return 0; |
914 | } | |
915 | ||
916 | static void __exit userspace_dirty_log_exit(void) | |
917 | { | |
918 | dm_dirty_log_type_unregister(&_userspace_type); | |
919 | dm_ulog_tfr_exit(); | |
920 | mempool_destroy(flush_entry_pool); | |
921 | ||
86a54a48 | 922 | DMINFO("version " DM_LOG_USERSPACE_VSN " unloaded"); |
f5db4af4 JB |
923 | return; |
924 | } | |
925 | ||
926 | module_init(userspace_dirty_log_init); | |
927 | module_exit(userspace_dirty_log_exit); | |
928 | ||
929 | MODULE_DESCRIPTION(DM_NAME " userspace dirty log link"); | |
930 | MODULE_AUTHOR("Jonathan Brassow <dm-devel@redhat.com>"); | |
931 | MODULE_LICENSE("GPL"); |