blob: 7f8a26fdcc2c06dfb0eaa7739c9593c0b575e7dd [file] [log] [blame]
Sage Weilf24e9982009-10-06 11:31:10 -07001#include "ceph_debug.h"
2
3#include <linux/err.h>
4#include <linux/highmem.h>
5#include <linux/mm.h>
6#include <linux/pagemap.h>
7#include <linux/slab.h>
8#include <linux/uaccess.h>
9
10#include "super.h"
11#include "osd_client.h"
12#include "messenger.h"
13#include "decode.h"
Sage Weil4e7a5dc2009-11-18 16:19:57 -080014#include "auth.h"
Sage Weilf24e9982009-10-06 11:31:10 -070015
Yehuda Sadeh0d59ab82010-01-13 17:03:23 -080016#define OSD_REPLY_RESERVE_FRONT_LEN 512
17
Sage Weilf24e9982009-10-06 11:31:10 -070018const static struct ceph_connection_operations osd_con_ops;
19
20static void kick_requests(struct ceph_osd_client *osdc, struct ceph_osd *osd);
21
22/*
23 * Implement client access to distributed object storage cluster.
24 *
25 * All data objects are stored within a cluster/cloud of OSDs, or
26 * "object storage devices." (Note that Ceph OSDs have _nothing_ to
27 * do with the T10 OSD extensions to SCSI.) Ceph OSDs are simply
28 * remote daemons serving up and coordinating consistent and safe
29 * access to storage.
30 *
31 * Cluster membership and the mapping of data objects onto storage devices
32 * are described by the osd map.
33 *
34 * We keep track of pending OSD requests (read, write), resubmit
35 * requests to different OSDs when the cluster topology/data layout
36 * change, or retry the affected requests when the communications
37 * channel with an OSD is reset.
38 */
39
40/*
41 * calculate the mapping of a file extent onto an object, and fill out the
42 * request accordingly. shorten extent as necessary if it crosses an
43 * object boundary.
44 *
45 * fill osd op in request message.
46 */
47static void calc_layout(struct ceph_osd_client *osdc,
48 struct ceph_vino vino, struct ceph_file_layout *layout,
49 u64 off, u64 *plen,
50 struct ceph_osd_request *req)
51{
52 struct ceph_osd_request_head *reqhead = req->r_request->front.iov_base;
53 struct ceph_osd_op *op = (void *)(reqhead + 1);
54 u64 orig_len = *plen;
55 u64 objoff, objlen; /* extent in object */
56 u64 bno;
57
58 reqhead->snapid = cpu_to_le64(vino.snap);
59
60 /* object extent? */
61 ceph_calc_file_object_mapping(layout, off, plen, &bno,
62 &objoff, &objlen);
63 if (*plen < orig_len)
64 dout(" skipping last %llu, final file extent %llu~%llu\n",
65 orig_len - *plen, off, *plen);
66
67 sprintf(req->r_oid, "%llx.%08llx", vino.ino, bno);
68 req->r_oid_len = strlen(req->r_oid);
69
70 op->extent.offset = cpu_to_le64(objoff);
71 op->extent.length = cpu_to_le64(objlen);
72 req->r_num_pages = calc_pages_for(off, *plen);
73
74 dout("calc_layout %s (%d) %llu~%llu (%d pages)\n",
75 req->r_oid, req->r_oid_len, objoff, objlen, req->r_num_pages);
76}
77
Yehuda Sadeh0d59ab82010-01-13 17:03:23 -080078static void remove_replies(struct ceph_osd_request *req)
79{
80 int i;
81 int max = ARRAY_SIZE(req->replies);
82
83 for (i=0; i<max; i++) {
84 if (req->replies[i])
85 ceph_msg_put(req->replies[i]);
86 }
87}
Sage Weilf24e9982009-10-06 11:31:10 -070088
89/*
90 * requests
91 */
Sage Weil415e49a2009-12-07 13:37:03 -080092void ceph_osdc_release_request(struct kref *kref)
Sage Weilf24e9982009-10-06 11:31:10 -070093{
Sage Weil415e49a2009-12-07 13:37:03 -080094 struct ceph_osd_request *req = container_of(kref,
95 struct ceph_osd_request,
96 r_kref);
97
98 if (req->r_request)
99 ceph_msg_put(req->r_request);
100 if (req->r_reply)
101 ceph_msg_put(req->r_reply);
Yehuda Sadeh0d59ab82010-01-13 17:03:23 -0800102 remove_replies(req);
103 if (req->r_con_filling_msg) {
Sage Weil350b1c32009-12-22 10:45:45 -0800104 dout("release_request revoking pages %p from con %p\n",
Yehuda Sadeh0d59ab82010-01-13 17:03:23 -0800105 req->r_pages, req->r_con_filling_msg);
106 ceph_con_revoke_message(req->r_con_filling_msg,
107 req->r_reply);
108 ceph_con_put(req->r_con_filling_msg);
Sage Weil350b1c32009-12-22 10:45:45 -0800109 }
Sage Weil415e49a2009-12-07 13:37:03 -0800110 if (req->r_own_pages)
111 ceph_release_page_vector(req->r_pages,
112 req->r_num_pages);
113 ceph_put_snap_context(req->r_snapc);
114 if (req->r_mempool)
115 mempool_free(req, req->r_osdc->req_mempool);
116 else
117 kfree(req);
Sage Weilf24e9982009-10-06 11:31:10 -0700118}
119
Yehuda Sadeh0d59ab82010-01-13 17:03:23 -0800120static int alloc_replies(struct ceph_osd_request *req, int num_reply)
121{
122 int i;
123 int max = ARRAY_SIZE(req->replies);
124
125 BUG_ON(num_reply > max);
126
127 for (i=0; i<num_reply; i++) {
128 req->replies[i] = ceph_msg_new(0, OSD_REPLY_RESERVE_FRONT_LEN, 0, 0, NULL);
129 if (IS_ERR(req->replies[i])) {
130 int j;
131 int err = PTR_ERR(req->replies[i]);
132 for (j = 0; j<=i; j++) {
133 ceph_msg_put(req->replies[j]);
134 }
135 return err;
136 }
137 }
138
139 for (; i<max; i++) {
140 req->replies[i] = NULL;
141 }
142
143 req->cur_reply = 0;
144
145 return 0;
146}
147
148static struct ceph_msg *__get_next_reply(struct ceph_connection *con,
149 struct ceph_osd_request *req,
150 int front_len)
151{
152 struct ceph_msg *reply;
153 if (req->r_con_filling_msg) {
154 dout("revoking reply msg %p from old con %p\n", req->r_reply,
155 req->r_con_filling_msg);
156 ceph_con_revoke_message(req->r_con_filling_msg, req->r_reply);
157 ceph_con_put(req->r_con_filling_msg);
158 req->cur_reply = 0;
159 }
160 reply = req->replies[req->cur_reply];
161 if (!reply || front_len > OSD_REPLY_RESERVE_FRONT_LEN) {
162 /* maybe we can allocate it now? */
163 reply = ceph_msg_new(0, front_len, 0, 0, NULL);
164 if (!reply || IS_ERR(reply)) {
165 pr_err(" reply alloc failed, front_len=%d\n", front_len);
166 return ERR_PTR(-ENOMEM);
167 }
168 }
169 req->r_con_filling_msg = ceph_con_get(con);
170 req->r_reply = ceph_msg_get(reply); /* for duration of read over socket */
171 return ceph_msg_get(reply);
172}
173
Sage Weilf24e9982009-10-06 11:31:10 -0700174/*
175 * build new request AND message, calculate layout, and adjust file
176 * extent as needed.
177 *
178 * if the file was recently truncated, we include information about its
179 * old and new size so that the object can be updated appropriately. (we
180 * avoid synchronously deleting truncated objects because it's slow.)
181 *
182 * if @do_sync, include a 'startsync' command so that the osd will flush
183 * data quickly.
184 */
185struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc,
186 struct ceph_file_layout *layout,
187 struct ceph_vino vino,
188 u64 off, u64 *plen,
189 int opcode, int flags,
190 struct ceph_snap_context *snapc,
191 int do_sync,
192 u32 truncate_seq,
193 u64 truncate_size,
194 struct timespec *mtime,
195 bool use_mempool, int num_reply)
196{
197 struct ceph_osd_request *req;
198 struct ceph_msg *msg;
199 struct ceph_osd_request_head *head;
200 struct ceph_osd_op *op;
201 void *p;
Yehuda Sadeh0c948992010-02-01 16:10:45 -0800202 int num_op = 1 + do_sync;
Sage Weilf24e9982009-10-06 11:31:10 -0700203 size_t msg_size = sizeof(*head) + num_op*sizeof(*op);
204 int err, i;
Sage Weilf24e9982009-10-06 11:31:10 -0700205
206 if (use_mempool) {
207 req = mempool_alloc(osdc->req_mempool, GFP_NOFS);
208 memset(req, 0, sizeof(*req));
209 } else {
210 req = kzalloc(sizeof(*req), GFP_NOFS);
211 }
212 if (req == NULL)
213 return ERR_PTR(-ENOMEM);
214
Yehuda Sadeh0d59ab82010-01-13 17:03:23 -0800215 err = alloc_replies(req, num_reply);
Sage Weilf24e9982009-10-06 11:31:10 -0700216 if (err) {
217 ceph_osdc_put_request(req);
218 return ERR_PTR(-ENOMEM);
219 }
Yehuda Sadeh93c20d92009-12-15 09:50:36 -0800220 req->r_num_prealloc_reply = num_reply;
Sage Weilf24e9982009-10-06 11:31:10 -0700221
222 req->r_osdc = osdc;
223 req->r_mempool = use_mempool;
Sage Weil415e49a2009-12-07 13:37:03 -0800224 kref_init(&req->r_kref);
Sage Weilf24e9982009-10-06 11:31:10 -0700225 init_completion(&req->r_completion);
226 init_completion(&req->r_safe_completion);
227 INIT_LIST_HEAD(&req->r_unsafe_item);
228 req->r_flags = flags;
229
230 WARN_ON((flags & (CEPH_OSD_FLAG_READ|CEPH_OSD_FLAG_WRITE)) == 0);
231
232 /* create message; allow space for oid */
233 msg_size += 40;
234 if (snapc)
235 msg_size += sizeof(u64) * snapc->num_snaps;
236 if (use_mempool)
Sage Weil8f3bc052009-10-14 17:36:07 -0700237 msg = ceph_msgpool_get(&osdc->msgpool_op, 0);
Sage Weilf24e9982009-10-06 11:31:10 -0700238 else
239 msg = ceph_msg_new(CEPH_MSG_OSD_OP, msg_size, 0, 0, NULL);
240 if (IS_ERR(msg)) {
Sage Weilf24e9982009-10-06 11:31:10 -0700241 ceph_osdc_put_request(req);
242 return ERR_PTR(PTR_ERR(msg));
243 }
244 msg->hdr.type = cpu_to_le16(CEPH_MSG_OSD_OP);
245 memset(msg->front.iov_base, 0, msg->front.iov_len);
246 head = msg->front.iov_base;
247 op = (void *)(head + 1);
248 p = (void *)(op + num_op);
249
250 req->r_request = msg;
251 req->r_snapc = ceph_get_snap_context(snapc);
252
253 head->client_inc = cpu_to_le32(1); /* always, for now. */
254 head->flags = cpu_to_le32(flags);
255 if (flags & CEPH_OSD_FLAG_WRITE)
256 ceph_encode_timespec(&head->mtime, mtime);
257 head->num_ops = cpu_to_le16(num_op);
258 op->op = cpu_to_le16(opcode);
259
260 /* calculate max write size */
261 calc_layout(osdc, vino, layout, off, plen, req);
262 req->r_file_layout = *layout; /* keep a copy */
263
264 if (flags & CEPH_OSD_FLAG_WRITE) {
265 req->r_request->hdr.data_off = cpu_to_le16(off);
266 req->r_request->hdr.data_len = cpu_to_le32(*plen);
267 op->payload_len = cpu_to_le32(*plen);
268 }
Yehuda Sadeh0c948992010-02-01 16:10:45 -0800269 op->extent.truncate_size = cpu_to_le64(truncate_size);
270 op->extent.truncate_seq = cpu_to_le32(truncate_seq);
Sage Weilf24e9982009-10-06 11:31:10 -0700271
272 /* fill in oid */
273 head->object_len = cpu_to_le32(req->r_oid_len);
274 memcpy(p, req->r_oid, req->r_oid_len);
275 p += req->r_oid_len;
276
Sage Weilf24e9982009-10-06 11:31:10 -0700277 if (do_sync) {
278 op++;
279 op->op = cpu_to_le16(CEPH_OSD_OP_STARTSYNC);
280 }
281 if (snapc) {
282 head->snap_seq = cpu_to_le64(snapc->seq);
283 head->num_snaps = cpu_to_le32(snapc->num_snaps);
284 for (i = 0; i < snapc->num_snaps; i++) {
285 put_unaligned_le64(snapc->snaps[i], p);
286 p += sizeof(u64);
287 }
288 }
289
290 BUG_ON(p > msg->front.iov_base + msg->front.iov_len);
291 return req;
292}
293
294/*
295 * We keep osd requests in an rbtree, sorted by ->r_tid.
296 */
297static void __insert_request(struct ceph_osd_client *osdc,
298 struct ceph_osd_request *new)
299{
300 struct rb_node **p = &osdc->requests.rb_node;
301 struct rb_node *parent = NULL;
302 struct ceph_osd_request *req = NULL;
303
304 while (*p) {
305 parent = *p;
306 req = rb_entry(parent, struct ceph_osd_request, r_node);
307 if (new->r_tid < req->r_tid)
308 p = &(*p)->rb_left;
309 else if (new->r_tid > req->r_tid)
310 p = &(*p)->rb_right;
311 else
312 BUG();
313 }
314
315 rb_link_node(&new->r_node, parent, p);
316 rb_insert_color(&new->r_node, &osdc->requests);
317}
318
319static struct ceph_osd_request *__lookup_request(struct ceph_osd_client *osdc,
320 u64 tid)
321{
322 struct ceph_osd_request *req;
323 struct rb_node *n = osdc->requests.rb_node;
324
325 while (n) {
326 req = rb_entry(n, struct ceph_osd_request, r_node);
327 if (tid < req->r_tid)
328 n = n->rb_left;
329 else if (tid > req->r_tid)
330 n = n->rb_right;
331 else
332 return req;
333 }
334 return NULL;
335}
336
337static struct ceph_osd_request *
338__lookup_request_ge(struct ceph_osd_client *osdc,
339 u64 tid)
340{
341 struct ceph_osd_request *req;
342 struct rb_node *n = osdc->requests.rb_node;
343
344 while (n) {
345 req = rb_entry(n, struct ceph_osd_request, r_node);
346 if (tid < req->r_tid) {
347 if (!n->rb_left)
348 return req;
349 n = n->rb_left;
350 } else if (tid > req->r_tid) {
351 n = n->rb_right;
352 } else {
353 return req;
354 }
355 }
356 return NULL;
357}
358
359
360/*
Sage Weil81b024e2009-10-09 10:29:18 -0700361 * If the osd connection drops, we need to resubmit all requests.
Sage Weilf24e9982009-10-06 11:31:10 -0700362 */
363static void osd_reset(struct ceph_connection *con)
364{
365 struct ceph_osd *osd = con->private;
366 struct ceph_osd_client *osdc;
367
368 if (!osd)
369 return;
370 dout("osd_reset osd%d\n", osd->o_osd);
371 osdc = osd->o_osdc;
372 osd->o_incarnation++;
373 down_read(&osdc->map_sem);
374 kick_requests(osdc, osd);
375 up_read(&osdc->map_sem);
376}
377
378/*
379 * Track open sessions with osds.
380 */
381static struct ceph_osd *create_osd(struct ceph_osd_client *osdc)
382{
383 struct ceph_osd *osd;
384
385 osd = kzalloc(sizeof(*osd), GFP_NOFS);
386 if (!osd)
387 return NULL;
388
389 atomic_set(&osd->o_ref, 1);
390 osd->o_osdc = osdc;
391 INIT_LIST_HEAD(&osd->o_requests);
Yehuda Sadehf5a20412010-02-03 11:00:26 -0800392 INIT_LIST_HEAD(&osd->o_osd_lru);
Sage Weilf24e9982009-10-06 11:31:10 -0700393 osd->o_incarnation = 1;
394
395 ceph_con_init(osdc->client->msgr, &osd->o_con);
396 osd->o_con.private = osd;
397 osd->o_con.ops = &osd_con_ops;
398 osd->o_con.peer_name.type = CEPH_ENTITY_TYPE_OSD;
Sage Weil4e7a5dc2009-11-18 16:19:57 -0800399
Sage Weilf24e9982009-10-06 11:31:10 -0700400 return osd;
401}
402
403static struct ceph_osd *get_osd(struct ceph_osd *osd)
404{
405 if (atomic_inc_not_zero(&osd->o_ref)) {
406 dout("get_osd %p %d -> %d\n", osd, atomic_read(&osd->o_ref)-1,
407 atomic_read(&osd->o_ref));
408 return osd;
409 } else {
410 dout("get_osd %p FAIL\n", osd);
411 return NULL;
412 }
413}
414
415static void put_osd(struct ceph_osd *osd)
416{
417 dout("put_osd %p %d -> %d\n", osd, atomic_read(&osd->o_ref),
418 atomic_read(&osd->o_ref) - 1);
Sage Weil42ce56e2009-11-18 11:22:36 -0800419 if (atomic_dec_and_test(&osd->o_ref))
Sage Weilf24e9982009-10-06 11:31:10 -0700420 kfree(osd);
Sage Weilf24e9982009-10-06 11:31:10 -0700421}
422
423/*
424 * remove an osd from our map
425 */
Yehuda Sadehf5a20412010-02-03 11:00:26 -0800426static void __remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
Sage Weilf24e9982009-10-06 11:31:10 -0700427{
Yehuda Sadehf5a20412010-02-03 11:00:26 -0800428 dout("__remove_osd %p\n", osd);
Sage Weilf24e9982009-10-06 11:31:10 -0700429 BUG_ON(!list_empty(&osd->o_requests));
430 rb_erase(&osd->o_node, &osdc->osds);
Yehuda Sadehf5a20412010-02-03 11:00:26 -0800431 list_del_init(&osd->o_osd_lru);
Sage Weilf24e9982009-10-06 11:31:10 -0700432 ceph_con_close(&osd->o_con);
433 put_osd(osd);
434}
435
Yehuda Sadehf5a20412010-02-03 11:00:26 -0800436static void __move_osd_to_lru(struct ceph_osd_client *osdc,
437 struct ceph_osd *osd)
438{
439 dout("__move_osd_to_lru %p\n", osd);
440 BUG_ON(!list_empty(&osd->o_osd_lru));
441 list_add_tail(&osd->o_osd_lru, &osdc->osd_lru);
442 osd->lru_ttl = jiffies + osdc->client->mount_args->osd_idle_ttl * HZ;
443}
444
445static void __remove_osd_from_lru(struct ceph_osd *osd)
446{
447 dout("__remove_osd_from_lru %p\n", osd);
448 if (!list_empty(&osd->o_osd_lru))
449 list_del_init(&osd->o_osd_lru);
450}
451
452static void remove_old_osds(struct ceph_osd_client *osdc, int remove_all)
453{
454 struct ceph_osd *osd, *nosd;
455
456 dout("__remove_old_osds %p\n", osdc);
457 mutex_lock(&osdc->request_mutex);
458 list_for_each_entry_safe(osd, nosd, &osdc->osd_lru, o_osd_lru) {
459 if (!remove_all && time_before(jiffies, osd->lru_ttl))
460 break;
461 __remove_osd(osdc, osd);
462 }
463 mutex_unlock(&osdc->request_mutex);
464}
465
Sage Weilf24e9982009-10-06 11:31:10 -0700466/*
467 * reset osd connect
468 */
Yehuda Sadehf5a20412010-02-03 11:00:26 -0800469static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
Sage Weilf24e9982009-10-06 11:31:10 -0700470{
471 int ret = 0;
472
Yehuda Sadehf5a20412010-02-03 11:00:26 -0800473 dout("__reset_osd %p osd%d\n", osd, osd->o_osd);
Sage Weilf24e9982009-10-06 11:31:10 -0700474 if (list_empty(&osd->o_requests)) {
Yehuda Sadehf5a20412010-02-03 11:00:26 -0800475 __remove_osd(osdc, osd);
Sage Weilf24e9982009-10-06 11:31:10 -0700476 } else {
477 ceph_con_close(&osd->o_con);
478 ceph_con_open(&osd->o_con, &osdc->osdmap->osd_addr[osd->o_osd]);
479 osd->o_incarnation++;
480 }
481 return ret;
482}
483
484static void __insert_osd(struct ceph_osd_client *osdc, struct ceph_osd *new)
485{
486 struct rb_node **p = &osdc->osds.rb_node;
487 struct rb_node *parent = NULL;
488 struct ceph_osd *osd = NULL;
489
490 while (*p) {
491 parent = *p;
492 osd = rb_entry(parent, struct ceph_osd, o_node);
493 if (new->o_osd < osd->o_osd)
494 p = &(*p)->rb_left;
495 else if (new->o_osd > osd->o_osd)
496 p = &(*p)->rb_right;
497 else
498 BUG();
499 }
500
501 rb_link_node(&new->o_node, parent, p);
502 rb_insert_color(&new->o_node, &osdc->osds);
503}
504
505static struct ceph_osd *__lookup_osd(struct ceph_osd_client *osdc, int o)
506{
507 struct ceph_osd *osd;
508 struct rb_node *n = osdc->osds.rb_node;
509
510 while (n) {
511 osd = rb_entry(n, struct ceph_osd, o_node);
512 if (o < osd->o_osd)
513 n = n->rb_left;
514 else if (o > osd->o_osd)
515 n = n->rb_right;
516 else
517 return osd;
518 }
519 return NULL;
520}
521
522
523/*
524 * Register request, assign tid. If this is the first request, set up
525 * the timeout event.
526 */
527static void register_request(struct ceph_osd_client *osdc,
528 struct ceph_osd_request *req)
529{
Sage Weilf24e9982009-10-06 11:31:10 -0700530 mutex_lock(&osdc->request_mutex);
531 req->r_tid = ++osdc->last_tid;
Sage Weil6df058c2009-12-22 11:24:33 -0800532 req->r_request->hdr.tid = cpu_to_le64(req->r_tid);
Sage Weilf24e9982009-10-06 11:31:10 -0700533
534 dout("register_request %p tid %lld\n", req, req->r_tid);
535 __insert_request(osdc, req);
536 ceph_osdc_get_request(req);
537 osdc->num_requests++;
538
539 req->r_timeout_stamp =
Sage Weil6b805182009-10-27 11:50:50 -0700540 jiffies + osdc->client->mount_args->osd_timeout*HZ;
Sage Weilf24e9982009-10-06 11:31:10 -0700541
542 if (osdc->num_requests == 1) {
543 osdc->timeout_tid = req->r_tid;
544 dout(" timeout on tid %llu at %lu\n", req->r_tid,
545 req->r_timeout_stamp);
546 schedule_delayed_work(&osdc->timeout_work,
547 round_jiffies_relative(req->r_timeout_stamp - jiffies));
548 }
549 mutex_unlock(&osdc->request_mutex);
550}
551
552/*
553 * called under osdc->request_mutex
554 */
555static void __unregister_request(struct ceph_osd_client *osdc,
556 struct ceph_osd_request *req)
557{
558 dout("__unregister_request %p tid %lld\n", req, req->r_tid);
559 rb_erase(&req->r_node, &osdc->requests);
560 osdc->num_requests--;
561
Sage Weil0ba64782009-10-08 16:57:16 -0700562 if (req->r_osd) {
563 /* make sure the original request isn't in flight. */
564 ceph_con_revoke(&req->r_osd->o_con, req->r_request);
565
566 list_del_init(&req->r_osd_item);
567 if (list_empty(&req->r_osd->o_requests))
Yehuda Sadehf5a20412010-02-03 11:00:26 -0800568 __move_osd_to_lru(osdc, req->r_osd);
Sage Weil0ba64782009-10-08 16:57:16 -0700569 req->r_osd = NULL;
570 }
Sage Weilf24e9982009-10-06 11:31:10 -0700571
572 ceph_osdc_put_request(req);
573
574 if (req->r_tid == osdc->timeout_tid) {
575 if (osdc->num_requests == 0) {
576 dout("no requests, canceling timeout\n");
577 osdc->timeout_tid = 0;
578 cancel_delayed_work(&osdc->timeout_work);
579 } else {
580 req = rb_entry(rb_first(&osdc->requests),
581 struct ceph_osd_request, r_node);
582 osdc->timeout_tid = req->r_tid;
583 dout("rescheduled timeout on tid %llu at %lu\n",
584 req->r_tid, req->r_timeout_stamp);
585 schedule_delayed_work(&osdc->timeout_work,
586 round_jiffies_relative(req->r_timeout_stamp -
587 jiffies));
588 }
589 }
590}
591
592/*
593 * Cancel a previously queued request message
594 */
595static void __cancel_request(struct ceph_osd_request *req)
596{
597 if (req->r_sent) {
598 ceph_con_revoke(&req->r_osd->o_con, req->r_request);
599 req->r_sent = 0;
600 }
601}
602
603/*
604 * Pick an osd (the first 'up' osd in the pg), allocate the osd struct
605 * (as needed), and set the request r_osd appropriately. If there is
606 * no up osd, set r_osd to NULL.
607 *
608 * Return 0 if unchanged, 1 if changed, or negative on error.
609 *
610 * Caller should hold map_sem for read and request_mutex.
611 */
612static int __map_osds(struct ceph_osd_client *osdc,
613 struct ceph_osd_request *req)
614{
615 struct ceph_osd_request_head *reqhead = req->r_request->front.iov_base;
Sage Weil51042122009-11-04 11:39:12 -0800616 struct ceph_pg pgid;
Sage Weilf24e9982009-10-06 11:31:10 -0700617 int o = -1;
618 int err;
619 struct ceph_osd *newosd = NULL;
620
621 dout("map_osds %p tid %lld\n", req, req->r_tid);
622 err = ceph_calc_object_layout(&reqhead->layout, req->r_oid,
623 &req->r_file_layout, osdc->osdmap);
624 if (err)
625 return err;
Sage Weil51042122009-11-04 11:39:12 -0800626 pgid = reqhead->layout.ol_pgid;
Sage Weil7740a422010-01-08 15:58:25 -0800627 req->r_pgid = pgid;
628
Sage Weilf24e9982009-10-06 11:31:10 -0700629 o = ceph_calc_pg_primary(osdc->osdmap, pgid);
630
631 if ((req->r_osd && req->r_osd->o_osd == o &&
632 req->r_sent >= req->r_osd->o_incarnation) ||
633 (req->r_osd == NULL && o == -1))
634 return 0; /* no change */
635
Sage Weil51042122009-11-04 11:39:12 -0800636 dout("map_osds tid %llu pgid %d.%x osd%d (was osd%d)\n",
637 req->r_tid, le32_to_cpu(pgid.pool), le16_to_cpu(pgid.ps), o,
Sage Weilf24e9982009-10-06 11:31:10 -0700638 req->r_osd ? req->r_osd->o_osd : -1);
639
640 if (req->r_osd) {
641 __cancel_request(req);
642 list_del_init(&req->r_osd_item);
643 if (list_empty(&req->r_osd->o_requests)) {
644 /* try to re-use r_osd if possible */
645 newosd = get_osd(req->r_osd);
Yehuda Sadehf5a20412010-02-03 11:00:26 -0800646 __remove_osd(osdc, newosd);
Sage Weilf24e9982009-10-06 11:31:10 -0700647 }
648 req->r_osd = NULL;
649 }
650
651 req->r_osd = __lookup_osd(osdc, o);
652 if (!req->r_osd && o >= 0) {
653 if (newosd) {
654 req->r_osd = newosd;
655 newosd = NULL;
656 } else {
657 err = -ENOMEM;
658 req->r_osd = create_osd(osdc);
659 if (!req->r_osd)
660 goto out;
661 }
662
663 dout("map_osds osd %p is osd%d\n", req->r_osd, o);
664 req->r_osd->o_osd = o;
665 req->r_osd->o_con.peer_name.num = cpu_to_le64(o);
666 __insert_osd(osdc, req->r_osd);
667
668 ceph_con_open(&req->r_osd->o_con, &osdc->osdmap->osd_addr[o]);
669 }
670
Yehuda Sadehf5a20412010-02-03 11:00:26 -0800671 if (req->r_osd) {
672 __remove_osd_from_lru(req->r_osd);
Sage Weilf24e9982009-10-06 11:31:10 -0700673 list_add(&req->r_osd_item, &req->r_osd->o_requests);
Yehuda Sadehf5a20412010-02-03 11:00:26 -0800674 }
Sage Weilf24e9982009-10-06 11:31:10 -0700675 err = 1; /* osd changed */
676
677out:
678 if (newosd)
679 put_osd(newosd);
680 return err;
681}
682
683/*
684 * caller should hold map_sem (for read) and request_mutex
685 */
686static int __send_request(struct ceph_osd_client *osdc,
687 struct ceph_osd_request *req)
688{
689 struct ceph_osd_request_head *reqhead;
690 int err;
691
692 err = __map_osds(osdc, req);
693 if (err < 0)
694 return err;
695 if (req->r_osd == NULL) {
696 dout("send_request %p no up osds in pg\n", req);
697 ceph_monc_request_next_osdmap(&osdc->client->monc);
698 return 0;
699 }
700
701 dout("send_request %p tid %llu to osd%d flags %d\n",
702 req, req->r_tid, req->r_osd->o_osd, req->r_flags);
703
704 reqhead = req->r_request->front.iov_base;
705 reqhead->osdmap_epoch = cpu_to_le32(osdc->osdmap->epoch);
706 reqhead->flags |= cpu_to_le32(req->r_flags); /* e.g., RETRY */
707 reqhead->reassert_version = req->r_reassert_version;
708
Sage Weil6b805182009-10-27 11:50:50 -0700709 req->r_timeout_stamp = jiffies+osdc->client->mount_args->osd_timeout*HZ;
Sage Weilf24e9982009-10-06 11:31:10 -0700710
711 ceph_msg_get(req->r_request); /* send consumes a ref */
712 ceph_con_send(&req->r_osd->o_con, req->r_request);
713 req->r_sent = req->r_osd->o_incarnation;
714 return 0;
715}
716
717/*
718 * Timeout callback, called every N seconds when 1 or more osd
719 * requests has been active for more than N seconds. When this
720 * happens, we ping all OSDs with requests who have timed out to
721 * ensure any communications channel reset is detected. Reset the
722 * request timeouts another N seconds in the future as we go.
723 * Reschedule the timeout event another N seconds in future (unless
724 * there are no open requests).
725 */
726static void handle_timeout(struct work_struct *work)
727{
728 struct ceph_osd_client *osdc =
729 container_of(work, struct ceph_osd_client, timeout_work.work);
730 struct ceph_osd_request *req;
731 struct ceph_osd *osd;
Sage Weil6b805182009-10-27 11:50:50 -0700732 unsigned long timeout = osdc->client->mount_args->osd_timeout * HZ;
Sage Weilf24e9982009-10-06 11:31:10 -0700733 unsigned long next_timeout = timeout + jiffies;
734 struct rb_node *p;
735
736 dout("timeout\n");
737 down_read(&osdc->map_sem);
738
739 ceph_monc_request_next_osdmap(&osdc->client->monc);
740
741 mutex_lock(&osdc->request_mutex);
742 for (p = rb_first(&osdc->requests); p; p = rb_next(p)) {
743 req = rb_entry(p, struct ceph_osd_request, r_node);
744
745 if (req->r_resend) {
746 int err;
747
748 dout("osdc resending prev failed %lld\n", req->r_tid);
749 err = __send_request(osdc, req);
750 if (err)
751 dout("osdc failed again on %lld\n", req->r_tid);
752 else
753 req->r_resend = false;
754 continue;
755 }
756 }
757 for (p = rb_first(&osdc->osds); p; p = rb_next(p)) {
758 osd = rb_entry(p, struct ceph_osd, o_node);
759 if (list_empty(&osd->o_requests))
760 continue;
761 req = list_first_entry(&osd->o_requests,
762 struct ceph_osd_request, r_osd_item);
763 if (time_before(jiffies, req->r_timeout_stamp))
764 continue;
765
766 dout(" tid %llu (at least) timed out on osd%d\n",
767 req->r_tid, osd->o_osd);
768 req->r_timeout_stamp = next_timeout;
769 ceph_con_keepalive(&osd->o_con);
770 }
771
772 if (osdc->timeout_tid)
773 schedule_delayed_work(&osdc->timeout_work,
774 round_jiffies_relative(timeout));
775
776 mutex_unlock(&osdc->request_mutex);
777
778 up_read(&osdc->map_sem);
779}
780
Yehuda Sadehf5a20412010-02-03 11:00:26 -0800781static void handle_osds_timeout(struct work_struct *work)
782{
783 struct ceph_osd_client *osdc =
784 container_of(work, struct ceph_osd_client,
785 osds_timeout_work.work);
786 unsigned long delay =
787 osdc->client->mount_args->osd_idle_ttl * HZ >> 2;
788
789 dout("osds timeout\n");
790 down_read(&osdc->map_sem);
791 remove_old_osds(osdc, 0);
792 up_read(&osdc->map_sem);
793
794 schedule_delayed_work(&osdc->osds_timeout_work,
795 round_jiffies_relative(delay));
796}
797
Sage Weilf24e9982009-10-06 11:31:10 -0700798/*
799 * handle osd op reply. either call the callback if it is specified,
800 * or do the completion to wake up the waiting thread.
801 */
Sage Weil350b1c32009-12-22 10:45:45 -0800802static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg,
803 struct ceph_connection *con)
Sage Weilf24e9982009-10-06 11:31:10 -0700804{
805 struct ceph_osd_reply_head *rhead = msg->front.iov_base;
806 struct ceph_osd_request *req;
807 u64 tid;
808 int numops, object_len, flags;
809
Sage Weil6df058c2009-12-22 11:24:33 -0800810 tid = le64_to_cpu(msg->hdr.tid);
Sage Weilf24e9982009-10-06 11:31:10 -0700811 if (msg->front.iov_len < sizeof(*rhead))
812 goto bad;
Sage Weilf24e9982009-10-06 11:31:10 -0700813 numops = le32_to_cpu(rhead->num_ops);
814 object_len = le32_to_cpu(rhead->object_len);
815 if (msg->front.iov_len != sizeof(*rhead) + object_len +
816 numops * sizeof(struct ceph_osd_op))
817 goto bad;
818 dout("handle_reply %p tid %llu\n", msg, tid);
819
820 /* lookup */
821 mutex_lock(&osdc->request_mutex);
822 req = __lookup_request(osdc, tid);
823 if (req == NULL) {
824 dout("handle_reply tid %llu dne\n", tid);
825 mutex_unlock(&osdc->request_mutex);
826 return;
827 }
828 ceph_osdc_get_request(req);
829 flags = le32_to_cpu(rhead->flags);
830
Sage Weil350b1c32009-12-22 10:45:45 -0800831 /*
Yehuda Sadeh0d59ab82010-01-13 17:03:23 -0800832 * if this connection filled our message, drop our reference now, to
Sage Weil350b1c32009-12-22 10:45:45 -0800833 * avoid a (safe but slower) revoke later.
834 */
Yehuda Sadeh0d59ab82010-01-13 17:03:23 -0800835 if (req->r_con_filling_msg == con && req->r_reply == msg) {
836 dout(" got pages, dropping con_filling_msg ref %p\n", con);
837 req->r_con_filling_msg = NULL;
Sage Weil350b1c32009-12-22 10:45:45 -0800838 ceph_con_put(con);
839 }
840
Sage Weilf24e9982009-10-06 11:31:10 -0700841 if (req->r_reply) {
842 /*
843 * once we see the message has been received, we don't
844 * need a ref (which is only needed for revoking
845 * pages)
846 */
847 ceph_msg_put(req->r_reply);
848 req->r_reply = NULL;
849 }
850
851 if (!req->r_got_reply) {
852 unsigned bytes;
853
854 req->r_result = le32_to_cpu(rhead->result);
855 bytes = le32_to_cpu(msg->hdr.data_len);
856 dout("handle_reply result %d bytes %d\n", req->r_result,
857 bytes);
858 if (req->r_result == 0)
859 req->r_result = bytes;
860
861 /* in case this is a write and we need to replay, */
862 req->r_reassert_version = rhead->reassert_version;
863
864 req->r_got_reply = 1;
865 } else if ((flags & CEPH_OSD_FLAG_ONDISK) == 0) {
866 dout("handle_reply tid %llu dup ack\n", tid);
Sage Weil34b43a52009-12-01 12:23:54 -0800867 mutex_unlock(&osdc->request_mutex);
Sage Weilf24e9982009-10-06 11:31:10 -0700868 goto done;
869 }
870
871 dout("handle_reply tid %llu flags %d\n", tid, flags);
872
873 /* either this is a read, or we got the safe response */
874 if ((flags & CEPH_OSD_FLAG_ONDISK) ||
875 ((flags & CEPH_OSD_FLAG_WRITE) == 0))
876 __unregister_request(osdc, req);
877
878 mutex_unlock(&osdc->request_mutex);
879
880 if (req->r_callback)
881 req->r_callback(req, msg);
882 else
883 complete(&req->r_completion);
884
885 if (flags & CEPH_OSD_FLAG_ONDISK) {
886 if (req->r_safe_callback)
887 req->r_safe_callback(req, msg);
888 complete(&req->r_safe_completion); /* fsync waiter */
889 }
890
891done:
892 ceph_osdc_put_request(req);
893 return;
894
895bad:
896 pr_err("corrupt osd_op_reply got %d %d expected %d\n",
897 (int)msg->front.iov_len, le32_to_cpu(msg->hdr.front_len),
898 (int)sizeof(*rhead));
Sage Weil9ec7cab2009-12-14 15:13:47 -0800899 ceph_msg_dump(msg);
Sage Weilf24e9982009-10-06 11:31:10 -0700900}
901
902
903/*
904 * Resubmit osd requests whose osd or osd address has changed. Request
905 * a new osd map if osds are down, or we are otherwise unable to determine
906 * how to direct a request.
907 *
908 * Close connections to down osds.
909 *
910 * If @who is specified, resubmit requests for that specific osd.
911 *
912 * Caller should hold map_sem for read and request_mutex.
913 */
914static void kick_requests(struct ceph_osd_client *osdc,
915 struct ceph_osd *kickosd)
916{
917 struct ceph_osd_request *req;
918 struct rb_node *p, *n;
919 int needmap = 0;
920 int err;
921
922 dout("kick_requests osd%d\n", kickosd ? kickosd->o_osd : -1);
923 mutex_lock(&osdc->request_mutex);
924 if (!kickosd) {
925 for (p = rb_first(&osdc->osds); p; p = n) {
926 struct ceph_osd *osd =
927 rb_entry(p, struct ceph_osd, o_node);
928
929 n = rb_next(p);
930 if (!ceph_osd_is_up(osdc->osdmap, osd->o_osd) ||
Sage Weil103e2d32010-01-07 16:12:36 -0800931 memcmp(&osd->o_con.peer_addr,
932 ceph_osd_addr(osdc->osdmap,
933 osd->o_osd),
934 sizeof(struct ceph_entity_addr)) != 0)
Yehuda Sadehf5a20412010-02-03 11:00:26 -0800935 __reset_osd(osdc, osd);
Sage Weilf24e9982009-10-06 11:31:10 -0700936 }
937 }
938
939 for (p = rb_first(&osdc->requests); p; p = rb_next(p)) {
940 req = rb_entry(p, struct ceph_osd_request, r_node);
941
942 if (req->r_resend) {
943 dout(" r_resend set on tid %llu\n", req->r_tid);
Sage Weil266673d2009-10-09 10:31:32 -0700944 __cancel_request(req);
Sage Weilf24e9982009-10-06 11:31:10 -0700945 goto kick;
946 }
Sage Weil266673d2009-10-09 10:31:32 -0700947 if (req->r_osd && kickosd == req->r_osd) {
948 __cancel_request(req);
Sage Weilf24e9982009-10-06 11:31:10 -0700949 goto kick;
Sage Weil266673d2009-10-09 10:31:32 -0700950 }
Sage Weilf24e9982009-10-06 11:31:10 -0700951
952 err = __map_osds(osdc, req);
953 if (err == 0)
954 continue; /* no change */
955 if (err < 0) {
956 /*
957 * FIXME: really, we should set the request
958 * error and fail if this isn't a 'nofail'
959 * request, but that's a fair bit more
960 * complicated to do. So retry!
961 */
962 dout(" setting r_resend on %llu\n", req->r_tid);
963 req->r_resend = true;
964 continue;
965 }
966 if (req->r_osd == NULL) {
967 dout("tid %llu maps to no valid osd\n", req->r_tid);
968 needmap++; /* request a newer map */
969 continue;
970 }
971
972kick:
Sage Weilc1ea8822009-10-08 16:55:47 -0700973 dout("kicking %p tid %llu osd%d\n", req, req->r_tid,
974 req->r_osd->o_osd);
Sage Weilf24e9982009-10-06 11:31:10 -0700975 req->r_flags |= CEPH_OSD_FLAG_RETRY;
976 err = __send_request(osdc, req);
977 if (err) {
978 dout(" setting r_resend on %llu\n", req->r_tid);
979 req->r_resend = true;
980 }
981 }
982 mutex_unlock(&osdc->request_mutex);
983
984 if (needmap) {
985 dout("%d requests for down osds, need new map\n", needmap);
986 ceph_monc_request_next_osdmap(&osdc->client->monc);
987 }
988}
989
990/*
991 * Process updated osd map.
992 *
993 * The message contains any number of incremental and full maps, normally
994 * indicating some sort of topology change in the cluster. Kick requests
995 * off to different OSDs as needed.
996 */
997void ceph_osdc_handle_map(struct ceph_osd_client *osdc, struct ceph_msg *msg)
998{
999 void *p, *end, *next;
1000 u32 nr_maps, maplen;
1001 u32 epoch;
1002 struct ceph_osdmap *newmap = NULL, *oldmap;
1003 int err;
1004 struct ceph_fsid fsid;
1005
1006 dout("handle_map have %u\n", osdc->osdmap ? osdc->osdmap->epoch : 0);
1007 p = msg->front.iov_base;
1008 end = p + msg->front.iov_len;
1009
1010 /* verify fsid */
1011 ceph_decode_need(&p, end, sizeof(fsid), bad);
1012 ceph_decode_copy(&p, &fsid, sizeof(fsid));
Sage Weil07433042009-11-18 16:50:41 -08001013 if (ceph_check_fsid(osdc->client, &fsid) < 0)
1014 return;
Sage Weilf24e9982009-10-06 11:31:10 -07001015
1016 down_write(&osdc->map_sem);
1017
1018 /* incremental maps */
1019 ceph_decode_32_safe(&p, end, nr_maps, bad);
1020 dout(" %d inc maps\n", nr_maps);
1021 while (nr_maps > 0) {
1022 ceph_decode_need(&p, end, 2*sizeof(u32), bad);
Sage Weilc89136e2009-10-14 09:59:09 -07001023 epoch = ceph_decode_32(&p);
1024 maplen = ceph_decode_32(&p);
Sage Weilf24e9982009-10-06 11:31:10 -07001025 ceph_decode_need(&p, end, maplen, bad);
1026 next = p + maplen;
1027 if (osdc->osdmap && osdc->osdmap->epoch+1 == epoch) {
1028 dout("applying incremental map %u len %d\n",
1029 epoch, maplen);
1030 newmap = osdmap_apply_incremental(&p, next,
1031 osdc->osdmap,
1032 osdc->client->msgr);
1033 if (IS_ERR(newmap)) {
1034 err = PTR_ERR(newmap);
1035 goto bad;
1036 }
Sage Weil30dc6382009-12-21 14:49:37 -08001037 BUG_ON(!newmap);
Sage Weilf24e9982009-10-06 11:31:10 -07001038 if (newmap != osdc->osdmap) {
1039 ceph_osdmap_destroy(osdc->osdmap);
1040 osdc->osdmap = newmap;
1041 }
1042 } else {
1043 dout("ignoring incremental map %u len %d\n",
1044 epoch, maplen);
1045 }
1046 p = next;
1047 nr_maps--;
1048 }
1049 if (newmap)
1050 goto done;
1051
1052 /* full maps */
1053 ceph_decode_32_safe(&p, end, nr_maps, bad);
1054 dout(" %d full maps\n", nr_maps);
1055 while (nr_maps) {
1056 ceph_decode_need(&p, end, 2*sizeof(u32), bad);
Sage Weilc89136e2009-10-14 09:59:09 -07001057 epoch = ceph_decode_32(&p);
1058 maplen = ceph_decode_32(&p);
Sage Weilf24e9982009-10-06 11:31:10 -07001059 ceph_decode_need(&p, end, maplen, bad);
1060 if (nr_maps > 1) {
1061 dout("skipping non-latest full map %u len %d\n",
1062 epoch, maplen);
1063 } else if (osdc->osdmap && osdc->osdmap->epoch >= epoch) {
1064 dout("skipping full map %u len %d, "
1065 "older than our %u\n", epoch, maplen,
1066 osdc->osdmap->epoch);
1067 } else {
1068 dout("taking full map %u len %d\n", epoch, maplen);
1069 newmap = osdmap_decode(&p, p+maplen);
1070 if (IS_ERR(newmap)) {
1071 err = PTR_ERR(newmap);
1072 goto bad;
1073 }
Sage Weil30dc6382009-12-21 14:49:37 -08001074 BUG_ON(!newmap);
Sage Weilf24e9982009-10-06 11:31:10 -07001075 oldmap = osdc->osdmap;
1076 osdc->osdmap = newmap;
1077 if (oldmap)
1078 ceph_osdmap_destroy(oldmap);
1079 }
1080 p += maplen;
1081 nr_maps--;
1082 }
1083
1084done:
1085 downgrade_write(&osdc->map_sem);
1086 ceph_monc_got_osdmap(&osdc->client->monc, osdc->osdmap->epoch);
1087 if (newmap)
1088 kick_requests(osdc, NULL);
1089 up_read(&osdc->map_sem);
1090 return;
1091
1092bad:
1093 pr_err("osdc handle_map corrupt msg\n");
Sage Weil9ec7cab2009-12-14 15:13:47 -08001094 ceph_msg_dump(msg);
Sage Weilf24e9982009-10-06 11:31:10 -07001095 up_write(&osdc->map_sem);
1096 return;
1097}
1098
1099
1100/*
1101 * A read request prepares specific pages that data is to be read into.
1102 * When a message is being read off the wire, we call prepare_pages to
1103 * find those pages.
1104 * 0 = success, -1 failure.
1105 */
Yehuda Sadeh0d59ab82010-01-13 17:03:23 -08001106static int __prepare_pages(struct ceph_connection *con,
Yehuda Sadeh0547a9b2010-01-11 14:47:13 -08001107 struct ceph_msg_header *hdr,
1108 struct ceph_osd_request *req,
1109 u64 tid,
1110 struct ceph_msg *m)
Sage Weilf24e9982009-10-06 11:31:10 -07001111{
1112 struct ceph_osd *osd = con->private;
1113 struct ceph_osd_client *osdc;
Sage Weilf24e9982009-10-06 11:31:10 -07001114 int ret = -1;
Yehuda Sadeh0547a9b2010-01-11 14:47:13 -08001115 int data_len = le32_to_cpu(hdr->data_len);
1116 unsigned data_off = le16_to_cpu(hdr->data_off);
1117
1118 int want = calc_pages_for(data_off & ~PAGE_MASK, data_len);
Sage Weilf24e9982009-10-06 11:31:10 -07001119
1120 if (!osd)
1121 return -1;
Yehuda Sadeh0547a9b2010-01-11 14:47:13 -08001122
Sage Weilf24e9982009-10-06 11:31:10 -07001123 osdc = osd->o_osdc;
1124
Yehuda Sadeh0d59ab82010-01-13 17:03:23 -08001125 dout("__prepare_pages on msg %p tid %llu, has %d pages, want %d\n", m,
Sage Weilf24e9982009-10-06 11:31:10 -07001126 tid, req->r_num_pages, want);
Sage Weil350b1c32009-12-22 10:45:45 -08001127 if (unlikely(req->r_num_pages < want))
1128 goto out;
Sage Weil350b1c32009-12-22 10:45:45 -08001129 m->pages = req->r_pages;
1130 m->nr_pages = req->r_num_pages;
1131 ret = 0; /* success */
Sage Weilf24e9982009-10-06 11:31:10 -07001132out:
Yehuda Sadeh0547a9b2010-01-11 14:47:13 -08001133 BUG_ON(ret < 0 || m->nr_pages < want);
1134
Sage Weilf24e9982009-10-06 11:31:10 -07001135 return ret;
1136}
1137
1138/*
1139 * Register request, send initial attempt.
1140 */
1141int ceph_osdc_start_request(struct ceph_osd_client *osdc,
1142 struct ceph_osd_request *req,
1143 bool nofail)
1144{
Sage Weilc1ea8822009-10-08 16:55:47 -07001145 int rc = 0;
Sage Weilf24e9982009-10-06 11:31:10 -07001146
1147 req->r_request->pages = req->r_pages;
1148 req->r_request->nr_pages = req->r_num_pages;
1149
1150 register_request(osdc, req);
1151
1152 down_read(&osdc->map_sem);
1153 mutex_lock(&osdc->request_mutex);
Sage Weilc1ea8822009-10-08 16:55:47 -07001154 /*
1155 * a racing kick_requests() may have sent the message for us
1156 * while we dropped request_mutex above, so only send now if
1157 * the request still han't been touched yet.
1158 */
1159 if (req->r_sent == 0) {
1160 rc = __send_request(osdc, req);
1161 if (rc) {
1162 if (nofail) {
1163 dout("osdc_start_request failed send, "
1164 " marking %lld\n", req->r_tid);
1165 req->r_resend = true;
1166 rc = 0;
1167 } else {
1168 __unregister_request(osdc, req);
1169 }
Sage Weilf24e9982009-10-06 11:31:10 -07001170 }
1171 }
1172 mutex_unlock(&osdc->request_mutex);
1173 up_read(&osdc->map_sem);
1174 return rc;
1175}
1176
1177/*
1178 * wait for a request to complete
1179 */
1180int ceph_osdc_wait_request(struct ceph_osd_client *osdc,
1181 struct ceph_osd_request *req)
1182{
1183 int rc;
1184
1185 rc = wait_for_completion_interruptible(&req->r_completion);
1186 if (rc < 0) {
1187 mutex_lock(&osdc->request_mutex);
1188 __cancel_request(req);
Sage Weil529cfcc2009-12-22 10:29:39 -08001189 __unregister_request(osdc, req);
Sage Weilf24e9982009-10-06 11:31:10 -07001190 mutex_unlock(&osdc->request_mutex);
Sage Weil529cfcc2009-12-22 10:29:39 -08001191 dout("wait_request tid %llu canceled/timed out\n", req->r_tid);
Sage Weilf24e9982009-10-06 11:31:10 -07001192 return rc;
1193 }
1194
1195 dout("wait_request tid %llu result %d\n", req->r_tid, req->r_result);
1196 return req->r_result;
1197}
1198
1199/*
1200 * sync - wait for all in-flight requests to flush. avoid starvation.
1201 */
1202void ceph_osdc_sync(struct ceph_osd_client *osdc)
1203{
1204 struct ceph_osd_request *req;
1205 u64 last_tid, next_tid = 0;
1206
1207 mutex_lock(&osdc->request_mutex);
1208 last_tid = osdc->last_tid;
1209 while (1) {
1210 req = __lookup_request_ge(osdc, next_tid);
1211 if (!req)
1212 break;
1213 if (req->r_tid > last_tid)
1214 break;
1215
1216 next_tid = req->r_tid + 1;
1217 if ((req->r_flags & CEPH_OSD_FLAG_WRITE) == 0)
1218 continue;
1219
1220 ceph_osdc_get_request(req);
1221 mutex_unlock(&osdc->request_mutex);
1222 dout("sync waiting on tid %llu (last is %llu)\n",
1223 req->r_tid, last_tid);
1224 wait_for_completion(&req->r_safe_completion);
1225 mutex_lock(&osdc->request_mutex);
1226 ceph_osdc_put_request(req);
1227 }
1228 mutex_unlock(&osdc->request_mutex);
1229 dout("sync done (thru tid %llu)\n", last_tid);
1230}
1231
1232/*
1233 * init, shutdown
1234 */
1235int ceph_osdc_init(struct ceph_osd_client *osdc, struct ceph_client *client)
1236{
1237 int err;
1238
1239 dout("init\n");
1240 osdc->client = client;
1241 osdc->osdmap = NULL;
1242 init_rwsem(&osdc->map_sem);
1243 init_completion(&osdc->map_waiters);
1244 osdc->last_requested_map = 0;
1245 mutex_init(&osdc->request_mutex);
1246 osdc->timeout_tid = 0;
1247 osdc->last_tid = 0;
1248 osdc->osds = RB_ROOT;
Yehuda Sadehf5a20412010-02-03 11:00:26 -08001249 INIT_LIST_HEAD(&osdc->osd_lru);
Sage Weilf24e9982009-10-06 11:31:10 -07001250 osdc->requests = RB_ROOT;
1251 osdc->num_requests = 0;
1252 INIT_DELAYED_WORK(&osdc->timeout_work, handle_timeout);
Yehuda Sadehf5a20412010-02-03 11:00:26 -08001253 INIT_DELAYED_WORK(&osdc->osds_timeout_work, handle_osds_timeout);
1254
1255 schedule_delayed_work(&osdc->osds_timeout_work,
1256 round_jiffies_relative(osdc->client->mount_args->osd_idle_ttl * HZ));
Sage Weilf24e9982009-10-06 11:31:10 -07001257
Sage Weil5f44f142009-11-18 14:52:18 -08001258 err = -ENOMEM;
Sage Weilf24e9982009-10-06 11:31:10 -07001259 osdc->req_mempool = mempool_create_kmalloc_pool(10,
1260 sizeof(struct ceph_osd_request));
1261 if (!osdc->req_mempool)
Sage Weil5f44f142009-11-18 14:52:18 -08001262 goto out;
Sage Weilf24e9982009-10-06 11:31:10 -07001263
1264 err = ceph_msgpool_init(&osdc->msgpool_op, 4096, 10, true);
1265 if (err < 0)
Sage Weil5f44f142009-11-18 14:52:18 -08001266 goto out_mempool;
Sage Weilf24e9982009-10-06 11:31:10 -07001267 return 0;
Sage Weil5f44f142009-11-18 14:52:18 -08001268
Sage Weil5f44f142009-11-18 14:52:18 -08001269out_mempool:
1270 mempool_destroy(osdc->req_mempool);
1271out:
1272 return err;
Sage Weilf24e9982009-10-06 11:31:10 -07001273}
1274
1275void ceph_osdc_stop(struct ceph_osd_client *osdc)
1276{
1277 cancel_delayed_work_sync(&osdc->timeout_work);
Yehuda Sadehf5a20412010-02-03 11:00:26 -08001278 cancel_delayed_work_sync(&osdc->osds_timeout_work);
Sage Weilf24e9982009-10-06 11:31:10 -07001279 if (osdc->osdmap) {
1280 ceph_osdmap_destroy(osdc->osdmap);
1281 osdc->osdmap = NULL;
1282 }
Yehuda Sadehf5a20412010-02-03 11:00:26 -08001283 remove_old_osds(osdc, 1);
Sage Weilf24e9982009-10-06 11:31:10 -07001284 mempool_destroy(osdc->req_mempool);
1285 ceph_msgpool_destroy(&osdc->msgpool_op);
Sage Weilf24e9982009-10-06 11:31:10 -07001286}
1287
1288/*
1289 * Read some contiguous pages. If we cross a stripe boundary, shorten
1290 * *plen. Return number of bytes read, or error.
1291 */
1292int ceph_osdc_readpages(struct ceph_osd_client *osdc,
1293 struct ceph_vino vino, struct ceph_file_layout *layout,
1294 u64 off, u64 *plen,
1295 u32 truncate_seq, u64 truncate_size,
1296 struct page **pages, int num_pages)
1297{
1298 struct ceph_osd_request *req;
1299 int rc = 0;
1300
1301 dout("readpages on ino %llx.%llx on %llu~%llu\n", vino.ino,
1302 vino.snap, off, *plen);
1303 req = ceph_osdc_new_request(osdc, layout, vino, off, plen,
1304 CEPH_OSD_OP_READ, CEPH_OSD_FLAG_READ,
1305 NULL, 0, truncate_seq, truncate_size, NULL,
1306 false, 1);
1307 if (IS_ERR(req))
1308 return PTR_ERR(req);
1309
1310 /* it may be a short read due to an object boundary */
1311 req->r_pages = pages;
1312 num_pages = calc_pages_for(off, *plen);
1313 req->r_num_pages = num_pages;
1314
1315 dout("readpages final extent is %llu~%llu (%d pages)\n",
1316 off, *plen, req->r_num_pages);
1317
1318 rc = ceph_osdc_start_request(osdc, req, false);
1319 if (!rc)
1320 rc = ceph_osdc_wait_request(osdc, req);
1321
1322 ceph_osdc_put_request(req);
1323 dout("readpages result %d\n", rc);
1324 return rc;
1325}
1326
1327/*
1328 * do a synchronous write on N pages
1329 */
1330int ceph_osdc_writepages(struct ceph_osd_client *osdc, struct ceph_vino vino,
1331 struct ceph_file_layout *layout,
1332 struct ceph_snap_context *snapc,
1333 u64 off, u64 len,
1334 u32 truncate_seq, u64 truncate_size,
1335 struct timespec *mtime,
1336 struct page **pages, int num_pages,
1337 int flags, int do_sync, bool nofail)
1338{
1339 struct ceph_osd_request *req;
1340 int rc = 0;
1341
1342 BUG_ON(vino.snap != CEPH_NOSNAP);
1343 req = ceph_osdc_new_request(osdc, layout, vino, off, &len,
1344 CEPH_OSD_OP_WRITE,
1345 flags | CEPH_OSD_FLAG_ONDISK |
1346 CEPH_OSD_FLAG_WRITE,
1347 snapc, do_sync,
1348 truncate_seq, truncate_size, mtime,
1349 nofail, 1);
1350 if (IS_ERR(req))
1351 return PTR_ERR(req);
1352
1353 /* it may be a short write due to an object boundary */
1354 req->r_pages = pages;
1355 req->r_num_pages = calc_pages_for(off, len);
1356 dout("writepages %llu~%llu (%d pages)\n", off, len,
1357 req->r_num_pages);
1358
1359 rc = ceph_osdc_start_request(osdc, req, nofail);
1360 if (!rc)
1361 rc = ceph_osdc_wait_request(osdc, req);
1362
1363 ceph_osdc_put_request(req);
1364 if (rc == 0)
1365 rc = len;
1366 dout("writepages result %d\n", rc);
1367 return rc;
1368}
1369
1370/*
1371 * handle incoming message
1372 */
1373static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
1374{
1375 struct ceph_osd *osd = con->private;
Julia Lawall32c895e2009-11-21 16:53:16 +01001376 struct ceph_osd_client *osdc;
Sage Weilf24e9982009-10-06 11:31:10 -07001377 int type = le16_to_cpu(msg->hdr.type);
1378
1379 if (!osd)
1380 return;
Julia Lawall32c895e2009-11-21 16:53:16 +01001381 osdc = osd->o_osdc;
Sage Weilf24e9982009-10-06 11:31:10 -07001382
1383 switch (type) {
1384 case CEPH_MSG_OSD_MAP:
1385 ceph_osdc_handle_map(osdc, msg);
1386 break;
1387 case CEPH_MSG_OSD_OPREPLY:
Sage Weil350b1c32009-12-22 10:45:45 -08001388 handle_reply(osdc, msg, con);
Sage Weilf24e9982009-10-06 11:31:10 -07001389 break;
1390
1391 default:
1392 pr_err("received unknown message type %d %s\n", type,
1393 ceph_msg_type_name(type));
1394 }
1395 ceph_msg_put(msg);
1396}
1397
1398static struct ceph_msg *alloc_msg(struct ceph_connection *con,
Yehuda Sadeh24504182010-01-08 13:58:34 -08001399 struct ceph_msg_header *hdr,
1400 int *skip)
Sage Weilf24e9982009-10-06 11:31:10 -07001401{
1402 struct ceph_osd *osd = con->private;
1403 struct ceph_osd_client *osdc = osd->o_osdc;
1404 int type = le16_to_cpu(hdr->type);
Sage Weil8f3bc052009-10-14 17:36:07 -07001405 int front = le32_to_cpu(hdr->front_len);
Yehuda Sadeh0547a9b2010-01-11 14:47:13 -08001406 int data_len = le32_to_cpu(hdr->data_len);
Yehuda Sadeh24504182010-01-08 13:58:34 -08001407 struct ceph_msg *m;
Yehuda Sadeh0547a9b2010-01-11 14:47:13 -08001408 struct ceph_osd_request *req;
1409 u64 tid;
1410 int err;
Sage Weilf24e9982009-10-06 11:31:10 -07001411
Yehuda Sadeh24504182010-01-08 13:58:34 -08001412 *skip = 0;
Yehuda Sadeh0547a9b2010-01-11 14:47:13 -08001413 if (type != CEPH_MSG_OSD_OPREPLY)
Yehuda Sadeh24504182010-01-08 13:58:34 -08001414 return NULL;
Yehuda Sadeh0547a9b2010-01-11 14:47:13 -08001415
1416 tid = le64_to_cpu(hdr->tid);
1417 mutex_lock(&osdc->request_mutex);
1418 req = __lookup_request(osdc, tid);
1419 if (!req) {
1420 *skip = 1;
1421 m = NULL;
Yehuda Sadeh0d59ab82010-01-13 17:03:23 -08001422 dout("alloc_msg unknown tid %llu\n", tid);
Yehuda Sadeh0547a9b2010-01-11 14:47:13 -08001423 goto out;
1424 }
Yehuda Sadeh0d59ab82010-01-13 17:03:23 -08001425 m = __get_next_reply(con, req, front);
1426 if (!m || IS_ERR(m)) {
Yehuda Sadeh0547a9b2010-01-11 14:47:13 -08001427 *skip = 1;
1428 goto out;
Sage Weilf24e9982009-10-06 11:31:10 -07001429 }
Yehuda Sadeh24504182010-01-08 13:58:34 -08001430
Yehuda Sadeh0547a9b2010-01-11 14:47:13 -08001431 if (data_len > 0) {
Yehuda Sadeh0d59ab82010-01-13 17:03:23 -08001432 err = __prepare_pages(con, hdr, req, tid, m);
Yehuda Sadeh0547a9b2010-01-11 14:47:13 -08001433 if (err < 0) {
1434 *skip = 1;
1435 ceph_msg_put(m);
1436 m = ERR_PTR(err);
1437 }
1438 }
1439
1440out:
1441 mutex_unlock(&osdc->request_mutex);
Yehuda Sadeh24504182010-01-08 13:58:34 -08001442
1443 return m;
Sage Weilf24e9982009-10-06 11:31:10 -07001444}
1445
1446/*
1447 * Wrappers to refcount containing ceph_osd struct
1448 */
1449static struct ceph_connection *get_osd_con(struct ceph_connection *con)
1450{
1451 struct ceph_osd *osd = con->private;
1452 if (get_osd(osd))
1453 return con;
1454 return NULL;
1455}
1456
1457static void put_osd_con(struct ceph_connection *con)
1458{
1459 struct ceph_osd *osd = con->private;
1460 put_osd(osd);
1461}
1462
Sage Weil4e7a5dc2009-11-18 16:19:57 -08001463/*
1464 * authentication
1465 */
1466static int get_authorizer(struct ceph_connection *con,
Sage Weil50b885b2009-12-01 14:12:07 -08001467 void **buf, int *len, int *proto,
1468 void **reply_buf, int *reply_len, int force_new)
Sage Weil4e7a5dc2009-11-18 16:19:57 -08001469{
1470 struct ceph_osd *o = con->private;
1471 struct ceph_osd_client *osdc = o->o_osdc;
1472 struct ceph_auth_client *ac = osdc->client->monc.auth;
1473 int ret = 0;
1474
1475 if (force_new && o->o_authorizer) {
1476 ac->ops->destroy_authorizer(ac, o->o_authorizer);
1477 o->o_authorizer = NULL;
1478 }
1479 if (o->o_authorizer == NULL) {
1480 ret = ac->ops->create_authorizer(
1481 ac, CEPH_ENTITY_TYPE_OSD,
1482 &o->o_authorizer,
1483 &o->o_authorizer_buf,
1484 &o->o_authorizer_buf_len,
1485 &o->o_authorizer_reply_buf,
1486 &o->o_authorizer_reply_buf_len);
1487 if (ret)
1488 return ret;
1489 }
1490
1491 *proto = ac->protocol;
1492 *buf = o->o_authorizer_buf;
1493 *len = o->o_authorizer_buf_len;
1494 *reply_buf = o->o_authorizer_reply_buf;
1495 *reply_len = o->o_authorizer_reply_buf_len;
1496 return 0;
1497}
1498
1499
1500static int verify_authorizer_reply(struct ceph_connection *con, int len)
1501{
1502 struct ceph_osd *o = con->private;
1503 struct ceph_osd_client *osdc = o->o_osdc;
1504 struct ceph_auth_client *ac = osdc->client->monc.auth;
1505
1506 return ac->ops->verify_authorizer_reply(ac, o->o_authorizer, len);
1507}
1508
Sage Weil9bd2e6f2010-02-02 16:21:06 -08001509static int invalidate_authorizer(struct ceph_connection *con)
1510{
1511 struct ceph_osd *o = con->private;
1512 struct ceph_osd_client *osdc = o->o_osdc;
1513 struct ceph_auth_client *ac = osdc->client->monc.auth;
1514
1515 if (ac->ops->invalidate_authorizer)
1516 ac->ops->invalidate_authorizer(ac, CEPH_ENTITY_TYPE_OSD);
1517
1518 return ceph_monc_validate_auth(&osdc->client->monc);
1519}
Sage Weil4e7a5dc2009-11-18 16:19:57 -08001520
Sage Weilf24e9982009-10-06 11:31:10 -07001521const static struct ceph_connection_operations osd_con_ops = {
1522 .get = get_osd_con,
1523 .put = put_osd_con,
1524 .dispatch = dispatch,
Sage Weil4e7a5dc2009-11-18 16:19:57 -08001525 .get_authorizer = get_authorizer,
1526 .verify_authorizer_reply = verify_authorizer_reply,
Sage Weil9bd2e6f2010-02-02 16:21:06 -08001527 .invalidate_authorizer = invalidate_authorizer,
Sage Weilf24e9982009-10-06 11:31:10 -07001528 .alloc_msg = alloc_msg,
Sage Weil81b024e2009-10-09 10:29:18 -07001529 .fault = osd_reset,
Sage Weilf24e9982009-10-06 11:31:10 -07001530};