projects / linux-2.6-block.git / blame
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (incremental) | history | HEAD
ceph: move max constant definitions
[linux-2.6-block.git] / net / ceph / osd_client.c
CommitLineData
3d14c5d2 1#include <linux/ceph/ceph_debug.h>
f24e9980 2
3d14c5d2 3#include <linux/module.h>
f24e9980
SW		 4#include <linux/err.h>
5#include <linux/highmem.h>
6#include <linux/mm.h>
7#include <linux/pagemap.h>
8#include <linux/slab.h>
9#include <linux/uaccess.h>
68b4476b
YS		 10#ifdef CONFIG_BLOCK
11#include <linux/bio.h>
12#endif
f24e9980 13
3d14c5d2
YS		 14#include <linux/ceph/libceph.h>
15#include <linux/ceph/osd_client.h>
16#include <linux/ceph/messenger.h>
17#include <linux/ceph/decode.h>
18#include <linux/ceph/auth.h>
19#include <linux/ceph/pagelist.h>
f24e9980 20
c16e7869
SW		 21#define OSD_OP_FRONT_LEN 4096
22#define OSD_OPREPLY_FRONT_LEN 512
0d59ab81 23
9e32789f 24static const struct ceph_connection_operations osd_con_ops;
f24e9980 25
f9d25199 26static void __send_queued(struct ceph_osd_client *osdc);
6f6c7006 27static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd);
a40c4f10
YS		 28static void __register_request(struct ceph_osd_client *osdc,
29 struct ceph_osd_request *req);
30static void __unregister_linger_request(struct ceph_osd_client *osdc,
31 struct ceph_osd_request *req);
56e925b6
SW		 32static void __send_request(struct ceph_osd_client *osdc,
33 struct ceph_osd_request *req);
f24e9980 34
68b4476b
YS		 35static int op_has_extent(int op)
36{
37 return (op == CEPH_OSD_OP_READ ||
38 op == CEPH_OSD_OP_WRITE);
39}
40
f24e9980
SW		 41/*
42 * Implement client access to distributed object storage cluster.
43 *
44 * All data objects are stored within a cluster/cloud of OSDs, or
45 * "object storage devices." (Note that Ceph OSDs have _nothing_ to
46 * do with the T10 OSD extensions to SCSI.) Ceph OSDs are simply
47 * remote daemons serving up and coordinating consistent and safe
48 * access to storage.
49 *
50 * Cluster membership and the mapping of data objects onto storage devices
51 * are described by the osd map.
52 *
53 * We keep track of pending OSD requests (read, write), resubmit
54 * requests to different OSDs when the cluster topology/data layout
55 * change, or retry the affected requests when the communications
56 * channel with an OSD is reset.
57 */
58
59/*
60 * calculate the mapping of a file extent onto an object, and fill out the
61 * request accordingly. shorten extent as necessary if it crosses an
62 * object boundary.
63 *
64 * fill osd op in request message.
65 */
dbe0fc41 66static int calc_layout(struct ceph_file_layout *layout, u64 off, u64 *plen,
47a05811 67 struct ceph_osd_req_op *op, u64 *bno)
f24e9980 68{
60e56f13 69 u64 orig_len = *plen;
60e56f13
AE		 70 u64 objoff = 0;
71 u64 objlen = 0;
d63b77f4 72 int r;
f24e9980 73
60e56f13 74 /* object extent? */
47a05811 75 r = ceph_calc_file_object_mapping(layout, off, orig_len, bno,
60e56f13 76 &objoff, &objlen);
d63b77f4
SW		 77 if (r < 0)
78 return r;
60e56f13
AE		 79 if (objlen < orig_len) {
80 *plen = objlen;
81 dout(" skipping last %llu, final file extent %llu~%llu\n",
82 orig_len - *plen, off, *plen);
83 }
84
85 if (op_has_extent(op->op)) {
86 u32 osize = le32_to_cpu(layout->fl_object_size);
87 op->extent.offset = objoff;
88 op->extent.length = objlen;
89 if (op->extent.truncate_size <= off - objoff) {
90 op->extent.truncate_size = 0;
91 } else {
92 op->extent.truncate_size -= off - objoff;
93 if (op->extent.truncate_size > osize)
94 op->extent.truncate_size = osize;
95 }
96 }
60e56f13
AE		 97 if (op->op == CEPH_OSD_OP_WRITE)
98 op->payload_len = *plen;
99
60cf5992 100 dout("calc_layout bno=%llx %llu~%llu\n", *bno, objoff, objlen);
f24e9980 101
3ff5f385 102 return 0;
f24e9980
SW		 103}
104
f24e9980
SW		 105/*
106 * requests
107 */
415e49a9 108void ceph_osdc_release_request(struct kref *kref)
f24e9980 109{
e0c59487 110 int num_pages;
415e49a9
SW		 111 struct ceph_osd_request *req = container_of(kref,
112 struct ceph_osd_request,
113 r_kref);
114
115 if (req->r_request)
116 ceph_msg_put(req->r_request);
0d59ab81 117 if (req->r_con_filling_msg) {
9cbb1d72
AE		 118 dout("%s revoking msg %p from con %p\n", __func__,
119 req->r_reply, req->r_con_filling_msg);
8921d114 120 ceph_msg_revoke_incoming(req->r_reply);
0d47766f 121 req->r_con_filling_msg->ops->put(req->r_con_filling_msg);
9cbb1d72 122 req->r_con_filling_msg = NULL;
350b1c32 123 }
ab8cb34a
AE		 124 if (req->r_reply)
125 ceph_msg_put(req->r_reply);
0fff87ec
AE		 126
127 if (req->r_data_in.type == CEPH_OSD_DATA_TYPE_PAGES &&
e0c59487
AE		 128 req->r_data_in.own_pages) {
129 num_pages = calc_pages_for((u64)req->r_data_in.alignment,
130 (u64)req->r_data_in.length);
131 ceph_release_page_vector(req->r_data_in.pages, num_pages);
132 }
0fff87ec 133 if (req->r_data_out.type == CEPH_OSD_DATA_TYPE_PAGES &&
e0c59487
AE		 134 req->r_data_out.own_pages) {
135 num_pages = calc_pages_for((u64)req->r_data_out.alignment,
136 (u64)req->r_data_out.length);
137 ceph_release_page_vector(req->r_data_out.pages, num_pages);
138 }
0fff87ec 139
415e49a9
SW		 140 ceph_put_snap_context(req->r_snapc);
141 if (req->r_mempool)
142 mempool_free(req, req->r_osdc->req_mempool);
143 else
144 kfree(req);
f24e9980 145}
3d14c5d2 146EXPORT_SYMBOL(ceph_osdc_release_request);
68b4476b 147
3499e8a5 148struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc,
f24e9980 149 struct ceph_snap_context *snapc,
1b83bef2 150 unsigned int num_ops,
3499e8a5 151 bool use_mempool,
54a54007 152 gfp_t gfp_flags)
f24e9980
SW		 153{
154 struct ceph_osd_request *req;
155 struct ceph_msg *msg;
1b83bef2
SW		 156 size_t msg_size;
157
158 msg_size = 4 + 4 + 8 + 8 + 4+8;
159 msg_size += 2 + 4 + 8 + 4 + 4; /* oloc */
160 msg_size += 1 + 8 + 4 + 4; /* pg_t */
161 msg_size += 4 + MAX_OBJ_NAME_SIZE;
162 msg_size += 2 + num_ops*sizeof(struct ceph_osd_op);
163 msg_size += 8; /* snapid */
164 msg_size += 8; /* snap_seq */
165 msg_size += 8 * (snapc ? snapc->num_snaps : 0); /* snaps */
166 msg_size += 4;
f24e9980
SW		 167
168 if (use_mempool) {
3499e8a5 169 req = mempool_alloc(osdc->req_mempool, gfp_flags);
f24e9980
SW		 170 memset(req, 0, sizeof(*req));
171 } else {
3499e8a5 172 req = kzalloc(sizeof(*req), gfp_flags);
f24e9980
SW		 173 }
174 if (req == NULL)
a79832f2 175 return NULL;
f24e9980 176
f24e9980
SW		 177 req->r_osdc = osdc;
178 req->r_mempool = use_mempool;
68b4476b 179
415e49a9 180 kref_init(&req->r_kref);
f24e9980
SW		 181 init_completion(&req->r_completion);
182 init_completion(&req->r_safe_completion);
a978fa20 183 RB_CLEAR_NODE(&req->r_node);
f24e9980 184 INIT_LIST_HEAD(&req->r_unsafe_item);
a40c4f10
YS		 185 INIT_LIST_HEAD(&req->r_linger_item);
186 INIT_LIST_HEAD(&req->r_linger_osd);
935b639a 187 INIT_LIST_HEAD(&req->r_req_lru_item);
cd43045c
SW		 188 INIT_LIST_HEAD(&req->r_osd_item);
189
c16e7869
SW		 190 /* create reply message */
191 if (use_mempool)
192 msg = ceph_msgpool_get(&osdc->msgpool_op_reply, 0);
193 else
194 msg = ceph_msg_new(CEPH_MSG_OSD_OPREPLY,
b61c2763 195 OSD_OPREPLY_FRONT_LEN, gfp_flags, true);
a79832f2 196 if (!msg) {
c16e7869 197 ceph_osdc_put_request(req);
a79832f2 198 return NULL;
c16e7869
SW		 199 }
200 req->r_reply = msg;
201
0fff87ec
AE		 202 req->r_data_in.type = CEPH_OSD_DATA_TYPE_NONE;
203 req->r_data_out.type = CEPH_OSD_DATA_TYPE_NONE;
d50b409f 204
c16e7869 205 /* create request message; allow space for oid */
f24e9980 206 if (use_mempool)
8f3bc053 207 msg = ceph_msgpool_get(&osdc->msgpool_op, 0);
f24e9980 208 else
b61c2763 209 msg = ceph_msg_new(CEPH_MSG_OSD_OP, msg_size, gfp_flags, true);
a79832f2 210 if (!msg) {
f24e9980 211 ceph_osdc_put_request(req);
a79832f2 212 return NULL;
f24e9980 213 }
68b4476b 214
f24e9980 215 memset(msg->front.iov_base, 0, msg->front.iov_len);
3499e8a5
YS		 216
217 req->r_request = msg;
3499e8a5
YS		 218
219 return req;
220}
3d14c5d2 221EXPORT_SYMBOL(ceph_osdc_alloc_request);
3499e8a5 222
175face2 223static u64 osd_req_encode_op(struct ceph_osd_request *req,
68b4476b
YS		 224 struct ceph_osd_op *dst,
225 struct ceph_osd_req_op *src)
226{
175face2 227 u64 out_data_len = 0;
95e072eb 228 struct ceph_pagelist *pagelist;
175face2 229
68b4476b
YS		 230 dst->op = cpu_to_le16(src->op);
231
065a68f9 232 switch (src->op) {
fbfab539
AE		 233 case CEPH_OSD_OP_STAT:
234 break;
68b4476b
YS		 235 case CEPH_OSD_OP_READ:
236 case CEPH_OSD_OP_WRITE:
175face2
AE		 237 if (src->op == CEPH_OSD_OP_WRITE)
238 out_data_len = src->extent.length;
239 dst->extent.offset = cpu_to_le64(src->extent.offset);
240 dst->extent.length = cpu_to_le64(src->extent.length);
68b4476b
YS		 241 dst->extent.truncate_size =
242 cpu_to_le64(src->extent.truncate_size);
243 dst->extent.truncate_seq =
244 cpu_to_le32(src->extent.truncate_seq);
245 break;
ae1533b6 246 case CEPH_OSD_OP_CALL:
95e072eb
AE		 247 pagelist = kmalloc(sizeof (*pagelist), GFP_NOFS);
248 BUG_ON(!pagelist);
249 ceph_pagelist_init(pagelist);
250
ae1533b6
YS		 251 dst->cls.class_len = src->cls.class_len;
252 dst->cls.method_len = src->cls.method_len;
253 dst->cls.indata_len = cpu_to_le32(src->cls.indata_len);
95e072eb 254 ceph_pagelist_append(pagelist, src->cls.class_name,
ae1533b6 255 src->cls.class_len);
95e072eb 256 ceph_pagelist_append(pagelist, src->cls.method_name,
ae1533b6 257 src->cls.method_len);
95e072eb 258 ceph_pagelist_append(pagelist, src->cls.indata,
ae1533b6 259 src->cls.indata_len);
95e072eb
AE		 260
261 req->r_data_out.type = CEPH_OSD_DATA_TYPE_PAGELIST;
262 req->r_data_out.pagelist = pagelist;
263 out_data_len = pagelist->length;
ae1533b6 264 break;
68b4476b
YS		 265 case CEPH_OSD_OP_STARTSYNC:
266 break;
a40c4f10
YS		 267 case CEPH_OSD_OP_NOTIFY_ACK:
268 case CEPH_OSD_OP_WATCH:
269 dst->watch.cookie = cpu_to_le64(src->watch.cookie);
270 dst->watch.ver = cpu_to_le64(src->watch.ver);
271 dst->watch.flag = src->watch.flag;
272 break;
68b4476b 273 default:
8f63ca2d 274 pr_err("unrecognized osd opcode %d\n", src->op);
68b4476b
YS		 275 WARN_ON(1);
276 break;
4c46459c
AE		 277 case CEPH_OSD_OP_MAPEXT:
278 case CEPH_OSD_OP_MASKTRUNC:
279 case CEPH_OSD_OP_SPARSE_READ:
a9f36c3e 280 case CEPH_OSD_OP_NOTIFY:
4c46459c
AE		 281 case CEPH_OSD_OP_ASSERT_VER:
282 case CEPH_OSD_OP_WRITEFULL:
283 case CEPH_OSD_OP_TRUNCATE:
284 case CEPH_OSD_OP_ZERO:
285 case CEPH_OSD_OP_DELETE:
286 case CEPH_OSD_OP_APPEND:
287 case CEPH_OSD_OP_SETTRUNC:
288 case CEPH_OSD_OP_TRIMTRUNC:
289 case CEPH_OSD_OP_TMAPUP:
290 case CEPH_OSD_OP_TMAPPUT:
291 case CEPH_OSD_OP_TMAPGET:
292 case CEPH_OSD_OP_CREATE:
a9f36c3e 293 case CEPH_OSD_OP_ROLLBACK:
4c46459c
AE		 294 case CEPH_OSD_OP_OMAPGETKEYS:
295 case CEPH_OSD_OP_OMAPGETVALS:
296 case CEPH_OSD_OP_OMAPGETHEADER:
297 case CEPH_OSD_OP_OMAPGETVALSBYKEYS:
298 case CEPH_OSD_OP_MODE_RD:
299 case CEPH_OSD_OP_OMAPSETVALS:
300 case CEPH_OSD_OP_OMAPSETHEADER:
301 case CEPH_OSD_OP_OMAPCLEAR:
302 case CEPH_OSD_OP_OMAPRMKEYS:
303 case CEPH_OSD_OP_OMAP_CMP:
304 case CEPH_OSD_OP_CLONERANGE:
305 case CEPH_OSD_OP_ASSERT_SRC_VERSION:
306 case CEPH_OSD_OP_SRC_CMPXATTR:
a9f36c3e 307 case CEPH_OSD_OP_GETXATTR:
4c46459c 308 case CEPH_OSD_OP_GETXATTRS:
a9f36c3e
AE		 309 case CEPH_OSD_OP_CMPXATTR:
310 case CEPH_OSD_OP_SETXATTR:
4c46459c
AE		 311 case CEPH_OSD_OP_SETXATTRS:
312 case CEPH_OSD_OP_RESETXATTRS:
313 case CEPH_OSD_OP_RMXATTR:
314 case CEPH_OSD_OP_PULL:
315 case CEPH_OSD_OP_PUSH:
316 case CEPH_OSD_OP_BALANCEREADS:
317 case CEPH_OSD_OP_UNBALANCEREADS:
318 case CEPH_OSD_OP_SCRUB:
319 case CEPH_OSD_OP_SCRUB_RESERVE:
320 case CEPH_OSD_OP_SCRUB_UNRESERVE:
321 case CEPH_OSD_OP_SCRUB_STOP:
322 case CEPH_OSD_OP_SCRUB_MAP:
323 case CEPH_OSD_OP_WRLOCK:
324 case CEPH_OSD_OP_WRUNLOCK:
325 case CEPH_OSD_OP_RDLOCK:
326 case CEPH_OSD_OP_RDUNLOCK:
327 case CEPH_OSD_OP_UPLOCK:
328 case CEPH_OSD_OP_DNLOCK:
329 case CEPH_OSD_OP_PGLS:
330 case CEPH_OSD_OP_PGLS_FILTER:
331 pr_err("unsupported osd opcode %s\n",
8f63ca2d 332 ceph_osd_op_name(src->op));
4c46459c
AE		 333 WARN_ON(1);
334 break;
68b4476b
YS		 335 }
336 dst->payload_len = cpu_to_le32(src->payload_len);
175face2
AE		 337
338 return out_data_len;
68b4476b
YS		 339}
340
3499e8a5
YS		 341/*
342 * build new request AND message
343 *
344 */
345void ceph_osdc_build_request(struct ceph_osd_request *req,
175face2 346 u64 off, unsigned int num_ops,
68b4476b 347 struct ceph_osd_req_op *src_ops,
4d6b250b 348 struct ceph_snap_context *snapc, u64 snap_id,
af77f26c 349 struct timespec *mtime)
3499e8a5
YS		 350{
351 struct ceph_msg *msg = req->r_request;
68b4476b 352 struct ceph_osd_req_op *src_op;
3499e8a5 353 void *p;
1b83bef2 354 size_t msg_size;
3499e8a5 355 int flags = req->r_flags;
f44246e3 356 u64 data_len;
68b4476b 357 int i;
3499e8a5 358
1b83bef2
SW		 359 req->r_num_ops = num_ops;
360 req->r_snapid = snap_id;
f24e9980
SW		 361 req->r_snapc = ceph_get_snap_context(snapc);
362
1b83bef2
SW		 363 /* encode request */
364 msg->hdr.version = cpu_to_le16(4);
365
366 p = msg->front.iov_base;
367 ceph_encode_32(&p, 1); /* client_inc is always 1 */
368 req->r_request_osdmap_epoch = p;
369 p += 4;
370 req->r_request_flags = p;
371 p += 4;
372 if (req->r_flags & CEPH_OSD_FLAG_WRITE)
373 ceph_encode_timespec(p, mtime);
374 p += sizeof(struct ceph_timespec);
375 req->r_request_reassert_version = p;
376 p += sizeof(struct ceph_eversion); /* will get filled in */
377
378 /* oloc */
379 ceph_encode_8(&p, 4);
380 ceph_encode_8(&p, 4);
381 ceph_encode_32(&p, 8 + 4 + 4);
382 req->r_request_pool = p;
383 p += 8;
384 ceph_encode_32(&p, -1); /* preferred */
385 ceph_encode_32(&p, 0); /* key len */
386
387 ceph_encode_8(&p, 1);
388 req->r_request_pgid = p;
389 p += 8 + 4;
390 ceph_encode_32(&p, -1); /* preferred */
391
392 /* oid */
393 ceph_encode_32(&p, req->r_oid_len);
af77f26c 394 memcpy(p, req->r_oid, req->r_oid_len);
1b83bef2 395 dout("oid '%.*s' len %d\n", req->r_oid_len, req->r_oid, req->r_oid_len);
af77f26c 396 p += req->r_oid_len;
f24e9980 397
175face2 398 /* ops--can imply data */
1b83bef2 399 ceph_encode_16(&p, num_ops);
68b4476b 400 src_op = src_ops;
1b83bef2 401 req->r_request_ops = p;
175face2 402 data_len = 0;
1b83bef2 403 for (i = 0; i < num_ops; i++, src_op++) {
175face2 404 data_len += osd_req_encode_op(req, p, src_op);
1b83bef2
SW		 405 p += sizeof(struct ceph_osd_op);
406 }
68b4476b 407
1b83bef2
SW		 408 /* snaps */
409 ceph_encode_64(&p, req->r_snapid);
410 ceph_encode_64(&p, req->r_snapc ? req->r_snapc->seq : 0);
411 ceph_encode_32(&p, req->r_snapc ? req->r_snapc->num_snaps : 0);
412 if (req->r_snapc) {
f24e9980 413 for (i = 0; i < snapc->num_snaps; i++) {
1b83bef2 414 ceph_encode_64(&p, req->r_snapc->snaps[i]);
f24e9980
SW		 415 }
416 }
417
1b83bef2
SW		 418 req->r_request_attempts = p;
419 p += 4;
420
175face2 421 /* data */
0baa1bd9
AE		 422 if (flags & CEPH_OSD_FLAG_WRITE) {
423 u16 data_off;
424
425 /*
426 * The header "data_off" is a hint to the receiver
427 * allowing it to align received data into its
428 * buffers such that there's no need to re-copy
429 * it before writing it to disk (direct I/O).
430 */
431 data_off = (u16) (off & 0xffff);
432 req->r_request->hdr.data_off = cpu_to_le16(data_off);
433 }
f44246e3 434 req->r_request->hdr.data_len = cpu_to_le32(data_len);
c5c6b19d 435
f24e9980 436 BUG_ON(p > msg->front.iov_base + msg->front.iov_len);
6f863e71
SW		 437 msg_size = p - msg->front.iov_base;
438 msg->front.iov_len = msg_size;
439 msg->hdr.front_len = cpu_to_le32(msg_size);
1b83bef2
SW		 440
441 dout("build_request msg_size was %d num_ops %d\n", (int)msg_size,
442 num_ops);
3499e8a5
YS		 443 return;
444}
3d14c5d2 445EXPORT_SYMBOL(ceph_osdc_build_request);
3499e8a5
YS		 446
447/*
448 * build new request AND message, calculate layout, and adjust file
449 * extent as needed.
450 *
451 * if the file was recently truncated, we include information about its
452 * old and new size so that the object can be updated appropriately. (we
453 * avoid synchronously deleting truncated objects because it's slow.)
454 *
455 * if @do_sync, include a 'startsync' command so that the osd will flush
456 * data quickly.
457 */
458struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc,
459 struct ceph_file_layout *layout,
460 struct ceph_vino vino,
461 u64 off, u64 *plen,
462 int opcode, int flags,
463 struct ceph_snap_context *snapc,
464 int do_sync,
465 u32 truncate_seq,
466 u64 truncate_size,
467 struct timespec *mtime,
153e5167 468 bool use_mempool)
3499e8a5 469{
ae7ca4a3 470 struct ceph_osd_req_op ops[2];
68b4476b 471 struct ceph_osd_request *req;
ae7ca4a3 472 unsigned int num_op = 1;
47a05811 473 u64 bno = 0;
6816282d 474 int r;
68b4476b 475
ae7ca4a3
AE		 476 memset(&ops, 0, sizeof ops);
477
68b4476b
YS		 478 ops[0].op = opcode;
479 ops[0].extent.truncate_seq = truncate_seq;
480 ops[0].extent.truncate_size = truncate_size;
68b4476b
YS		 481
482 if (do_sync) {
483 ops[1].op = CEPH_OSD_OP_STARTSYNC;
ae7ca4a3
AE		 484 num_op++;
485 }
68b4476b 486
ae7ca4a3
AE		 487 req = ceph_osdc_alloc_request(osdc, snapc, num_op, use_mempool,
488 GFP_NOFS);
4ad12621 489 if (!req)
6816282d 490 return ERR_PTR(-ENOMEM);
d178a9e7 491 req->r_flags = flags;
3499e8a5
YS		 492
493 /* calculate max write size */
60cf5992 494 r = calc_layout(layout, off, plen, ops, &bno);
3ff5f385
AE		 495 if (r < 0) {
496 ceph_osdc_put_request(req);
6816282d 497 return ERR_PTR(r);
3ff5f385 498 }
3499e8a5
YS		 499 req->r_file_layout = *layout; /* keep a copy */
500
dbe0fc41
AE		 501 snprintf(req->r_oid, sizeof(req->r_oid), "%llx.%08llx", vino.ino, bno);
502 req->r_oid_len = strlen(req->r_oid);
503
175face2 504 ceph_osdc_build_request(req, off, num_op, ops,
ae7ca4a3 505 snapc, vino.snap, mtime);
3499e8a5 506
f24e9980
SW		 507 return req;
508}
3d14c5d2 509EXPORT_SYMBOL(ceph_osdc_new_request);
f24e9980
SW		 510
511/*
512 * We keep osd requests in an rbtree, sorted by ->r_tid.
513 */
514static void __insert_request(struct ceph_osd_client *osdc,
515 struct ceph_osd_request *new)
516{
517 struct rb_node **p = &osdc->requests.rb_node;
518 struct rb_node *parent = NULL;
519 struct ceph_osd_request *req = NULL;
520
521 while (*p) {
522 parent = *p;
523 req = rb_entry(parent, struct ceph_osd_request, r_node);
524 if (new->r_tid < req->r_tid)
525 p = &(*p)->rb_left;
526 else if (new->r_tid > req->r_tid)
527 p = &(*p)->rb_right;
528 else
529 BUG();
530 }
531
532 rb_link_node(&new->r_node, parent, p);
533 rb_insert_color(&new->r_node, &osdc->requests);
534}
535
536static struct ceph_osd_request *__lookup_request(struct ceph_osd_client *osdc,
537 u64 tid)
538{
539 struct ceph_osd_request *req;
540 struct rb_node *n = osdc->requests.rb_node;
541
542 while (n) {
543 req = rb_entry(n, struct ceph_osd_request, r_node);
544 if (tid < req->r_tid)
545 n = n->rb_left;
546 else if (tid > req->r_tid)
547 n = n->rb_right;
548 else
549 return req;
550 }
551 return NULL;
552}
553
554static struct ceph_osd_request *
555__lookup_request_ge(struct ceph_osd_client *osdc,
556 u64 tid)
557{
558 struct ceph_osd_request *req;
559 struct rb_node *n = osdc->requests.rb_node;
560
561 while (n) {
562 req = rb_entry(n, struct ceph_osd_request, r_node);
563 if (tid < req->r_tid) {
564 if (!n->rb_left)
565 return req;
566 n = n->rb_left;
567 } else if (tid > req->r_tid) {
568 n = n->rb_right;
569 } else {
570 return req;
571 }
572 }
573 return NULL;
574}
575
6f6c7006
SW		 576/*
577 * Resubmit requests pending on the given osd.
578 */
579static void __kick_osd_requests(struct ceph_osd_client *osdc,
580 struct ceph_osd *osd)
581{
a40c4f10 582 struct ceph_osd_request *req, *nreq;
e02493c0 583 LIST_HEAD(resend);
6f6c7006
SW		 584 int err;
585
586 dout("__kick_osd_requests osd%d\n", osd->o_osd);
587 err = __reset_osd(osdc, osd);
685a7555 588 if (err)
6f6c7006 589 return;
e02493c0
AE		 590 /*
591 * Build up a list of requests to resend by traversing the
592 * osd's list of requests. Requests for a given object are
593 * sent in tid order, and that is also the order they're
594 * kept on this list. Therefore all requests that are in
595 * flight will be found first, followed by all requests that
596 * have not yet been sent. And to resend requests while
597 * preserving this order we will want to put any sent
598 * requests back on the front of the osd client's unsent
599 * list.
600 *
601 * So we build a separate ordered list of already-sent
602 * requests for the affected osd and splice it onto the
603 * front of the osd client's unsent list. Once we've seen a
604 * request that has not yet been sent we're done. Those
605 * requests are already sitting right where they belong.
606 */
6f6c7006 607 list_for_each_entry(req, &osd->o_requests, r_osd_item) {
e02493c0
AE		 608 if (!req->r_sent)
609 break;
610 list_move_tail(&req->r_req_lru_item, &resend);
611 dout("requeueing %p tid %llu osd%d\n", req, req->r_tid,
6f6c7006 612 osd->o_osd);
a40c4f10
YS		 613 if (!req->r_linger)
614 req->r_flags |= CEPH_OSD_FLAG_RETRY;
615 }
e02493c0 616 list_splice(&resend, &osdc->req_unsent);
a40c4f10 617
e02493c0
AE		 618 /*
619 * Linger requests are re-registered before sending, which
620 * sets up a new tid for each. We add them to the unsent
621 * list at the end to keep things in tid order.
622 */
a40c4f10
YS		 623 list_for_each_entry_safe(req, nreq, &osd->o_linger_requests,
624 r_linger_osd) {
77f38e0e
SW		 625 /*
626 * reregister request prior to unregistering linger so
627 * that r_osd is preserved.
628 */
629 BUG_ON(!list_empty(&req->r_req_lru_item));
a40c4f10 630 __register_request(osdc, req);
e02493c0 631 list_add_tail(&req->r_req_lru_item, &osdc->req_unsent);
ad885927 632 list_add_tail(&req->r_osd_item, &req->r_osd->o_requests);
77f38e0e 633 __unregister_linger_request(osdc, req);
a40c4f10
YS		 634 dout("requeued lingering %p tid %llu osd%d\n", req, req->r_tid,
635 osd->o_osd);
6f6c7006
SW		 636 }
637}
638
f24e9980 639/*
81b024e7 640 * If the osd connection drops, we need to resubmit all requests.
f24e9980
SW		 641 */
642static void osd_reset(struct ceph_connection *con)
643{
644 struct ceph_osd *osd = con->private;
645 struct ceph_osd_client *osdc;
646
647 if (!osd)
648 return;
649 dout("osd_reset osd%d\n", osd->o_osd);
650 osdc = osd->o_osdc;
f24e9980 651 down_read(&osdc->map_sem);
83aff95e
SW		 652 mutex_lock(&osdc->request_mutex);
653 __kick_osd_requests(osdc, osd);
f9d25199 654 __send_queued(osdc);
83aff95e 655 mutex_unlock(&osdc->request_mutex);
f24e9980
SW		 656 up_read(&osdc->map_sem);
657}
658
659/*
660 * Track open sessions with osds.
661 */
e10006f8 662static struct ceph_osd *create_osd(struct ceph_osd_client *osdc, int onum)
f24e9980
SW		 663{
664 struct ceph_osd *osd;
665
666 osd = kzalloc(sizeof(*osd), GFP_NOFS);
667 if (!osd)
668 return NULL;
669
670 atomic_set(&osd->o_ref, 1);
671 osd->o_osdc = osdc;
e10006f8 672 osd->o_osd = onum;
f407731d 673 RB_CLEAR_NODE(&osd->o_node);
f24e9980 674 INIT_LIST_HEAD(&osd->o_requests);
a40c4f10 675 INIT_LIST_HEAD(&osd->o_linger_requests);
f5a2041b 676 INIT_LIST_HEAD(&osd->o_osd_lru);
f24e9980
SW		 677 osd->o_incarnation = 1;
678
b7a9e5dd 679 ceph_con_init(&osd->o_con, osd, &osd_con_ops, &osdc->client->msgr);
4e7a5dcd 680
422d2cb8 681 INIT_LIST_HEAD(&osd->o_keepalive_item);
f24e9980
SW		 682 return osd;
683}
684
685static struct ceph_osd *get_osd(struct ceph_osd *osd)
686{
687 if (atomic_inc_not_zero(&osd->o_ref)) {
688 dout("get_osd %p %d -> %d\n", osd, atomic_read(&osd->o_ref)-1,
689 atomic_read(&osd->o_ref));
690 return osd;
691 } else {
692 dout("get_osd %p FAIL\n", osd);
693 return NULL;
694 }
695}
696
697static void put_osd(struct ceph_osd *osd)
698{
699 dout("put_osd %p %d -> %d\n", osd, atomic_read(&osd->o_ref),
700 atomic_read(&osd->o_ref) - 1);
a255651d 701 if (atomic_dec_and_test(&osd->o_ref) && osd->o_auth.authorizer) {
79494d1b
SW		 702 struct ceph_auth_client *ac = osd->o_osdc->client->monc.auth;
703
27859f97 704 ceph_auth_destroy_authorizer(ac, osd->o_auth.authorizer);
f24e9980 705 kfree(osd);
79494d1b 706 }
f24e9980
SW		 707}
708
709/*
710 * remove an osd from our map
711 */
f5a2041b 712static void __remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
f24e9980 713{
f5a2041b 714 dout("__remove_osd %p\n", osd);
f24e9980
SW		 715 BUG_ON(!list_empty(&osd->o_requests));
716 rb_erase(&osd->o_node, &osdc->osds);
f5a2041b 717 list_del_init(&osd->o_osd_lru);
f24e9980
SW		 718 ceph_con_close(&osd->o_con);
719 put_osd(osd);
720}
721
aca420bc
SW		 722static void remove_all_osds(struct ceph_osd_client *osdc)
723{
048a9d2d 724 dout("%s %p\n", __func__, osdc);
aca420bc
SW		 725 mutex_lock(&osdc->request_mutex);
726 while (!RB_EMPTY_ROOT(&osdc->osds)) {
727 struct ceph_osd *osd = rb_entry(rb_first(&osdc->osds),
728 struct ceph_osd, o_node);
729 __remove_osd(osdc, osd);
730 }
731 mutex_unlock(&osdc->request_mutex);
732}
733
f5a2041b
YS		 734static void __move_osd_to_lru(struct ceph_osd_client *osdc,
735 struct ceph_osd *osd)
736{
737 dout("__move_osd_to_lru %p\n", osd);
738 BUG_ON(!list_empty(&osd->o_osd_lru));
739 list_add_tail(&osd->o_osd_lru, &osdc->osd_lru);
3d14c5d2 740 osd->lru_ttl = jiffies + osdc->client->options->osd_idle_ttl * HZ;
f5a2041b
YS		 741}
742
743static void __remove_osd_from_lru(struct ceph_osd *osd)
744{
745 dout("__remove_osd_from_lru %p\n", osd);
746 if (!list_empty(&osd->o_osd_lru))
747 list_del_init(&osd->o_osd_lru);
748}
749
aca420bc 750static void remove_old_osds(struct ceph_osd_client *osdc)
f5a2041b
YS		 751{
752 struct ceph_osd *osd, *nosd;
753
754 dout("__remove_old_osds %p\n", osdc);
755 mutex_lock(&osdc->request_mutex);
756 list_for_each_entry_safe(osd, nosd, &osdc->osd_lru, o_osd_lru) {
aca420bc 757 if (time_before(jiffies, osd->lru_ttl))
f5a2041b
YS		 758 break;
759 __remove_osd(osdc, osd);
760 }
761 mutex_unlock(&osdc->request_mutex);
762}
763
f24e9980
SW		 764/*
765 * reset osd connect
766 */
f5a2041b 767static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
f24e9980 768{
c3acb181 769 struct ceph_entity_addr *peer_addr;
f24e9980 770
f5a2041b 771 dout("__reset_osd %p osd%d\n", osd, osd->o_osd);
a40c4f10
YS		 772 if (list_empty(&osd->o_requests) &&
773 list_empty(&osd->o_linger_requests)) {
f5a2041b 774 __remove_osd(osdc, osd);
c3acb181
AE		 775
776 return -ENODEV;
777 }
778
779 peer_addr = &osdc->osdmap->osd_addr[osd->o_osd];
780 if (!memcmp(peer_addr, &osd->o_con.peer_addr, sizeof (*peer_addr)) &&
781 !ceph_con_opened(&osd->o_con)) {
782 struct ceph_osd_request *req;
783
87b315a5
SW		 784 dout(" osd addr hasn't changed and connection never opened,"
785 " letting msgr retry");
786 /* touch each r_stamp for handle_timeout()'s benfit */
787 list_for_each_entry(req, &osd->o_requests, r_osd_item)
788 req->r_stamp = jiffies;
c3acb181
AE		 789
790 return -EAGAIN;
f24e9980 791 }
c3acb181
AE		 792
793 ceph_con_close(&osd->o_con);
794 ceph_con_open(&osd->o_con, CEPH_ENTITY_TYPE_OSD, osd->o_osd, peer_addr);
795 osd->o_incarnation++;
796
797 return 0;
f24e9980
SW		 798}
799
800static void __insert_osd(struct ceph_osd_client *osdc, struct ceph_osd *new)
801{
802 struct rb_node **p = &osdc->osds.rb_node;
803 struct rb_node *parent = NULL;
804 struct ceph_osd *osd = NULL;
805
aca420bc 806 dout("__insert_osd %p osd%d\n", new, new->o_osd);
f24e9980
SW		 807 while (*p) {
808 parent = *p;
809 osd = rb_entry(parent, struct ceph_osd, o_node);
810 if (new->o_osd < osd->o_osd)
811 p = &(*p)->rb_left;
812 else if (new->o_osd > osd->o_osd)
813 p = &(*p)->rb_right;
814 else
815 BUG();
816 }
817
818 rb_link_node(&new->o_node, parent, p);
819 rb_insert_color(&new->o_node, &osdc->osds);
820}
821
822static struct ceph_osd *__lookup_osd(struct ceph_osd_client *osdc, int o)
823{
824 struct ceph_osd *osd;
825 struct rb_node *n = osdc->osds.rb_node;
826
827 while (n) {
828 osd = rb_entry(n, struct ceph_osd, o_node);
829 if (o < osd->o_osd)
830 n = n->rb_left;
831 else if (o > osd->o_osd)
832 n = n->rb_right;
833 else
834 return osd;
835 }
836 return NULL;
837}
838
422d2cb8
YS		 839static void __schedule_osd_timeout(struct ceph_osd_client *osdc)
840{
841 schedule_delayed_work(&osdc->timeout_work,
3d14c5d2 842 osdc->client->options->osd_keepalive_timeout * HZ);
422d2cb8
YS		 843}
844
845static void __cancel_osd_timeout(struct ceph_osd_client *osdc)
846{
847 cancel_delayed_work(&osdc->timeout_work);
848}
f24e9980
SW		 849
850/*
851 * Register request, assign tid. If this is the first request, set up
852 * the timeout event.
853 */
a40c4f10
YS		 854static void __register_request(struct ceph_osd_client *osdc,
855 struct ceph_osd_request *req)
f24e9980 856{
f24e9980 857 req->r_tid = ++osdc->last_tid;
6df058c0 858 req->r_request->hdr.tid = cpu_to_le64(req->r_tid);
77f38e0e 859 dout("__register_request %p tid %lld\n", req, req->r_tid);
f24e9980
SW		 860 __insert_request(osdc, req);
861 ceph_osdc_get_request(req);
862 osdc->num_requests++;
f24e9980 863 if (osdc->num_requests == 1) {
422d2cb8
YS		 864 dout(" first request, scheduling timeout\n");
865 __schedule_osd_timeout(osdc);
f24e9980 866 }
a40c4f10
YS		 867}
868
f24e9980
SW		 869/*
870 * called under osdc->request_mutex
871 */
872static void __unregister_request(struct ceph_osd_client *osdc,
873 struct ceph_osd_request *req)
874{
35f9f8a0
SW		 875 if (RB_EMPTY_NODE(&req->r_node)) {
876 dout("__unregister_request %p tid %lld not registered\n",
877 req, req->r_tid);
878 return;
879 }
880
f24e9980
SW		 881 dout("__unregister_request %p tid %lld\n", req, req->r_tid);
882 rb_erase(&req->r_node, &osdc->requests);
883 osdc->num_requests--;
884
0ba6478d
SW		 885 if (req->r_osd) {
886 /* make sure the original request isn't in flight. */
6740a845 887 ceph_msg_revoke(req->r_request);
0ba6478d
SW		 888
889 list_del_init(&req->r_osd_item);
a40c4f10
YS		 890 if (list_empty(&req->r_osd->o_requests) &&
891 list_empty(&req->r_osd->o_linger_requests)) {
892 dout("moving osd to %p lru\n", req->r_osd);
f5a2041b 893 __move_osd_to_lru(osdc, req->r_osd);
a40c4f10 894 }
fbdb9190 895 if (list_empty(&req->r_linger_item))
a40c4f10 896 req->r_osd = NULL;
0ba6478d 897 }
f24e9980 898
7d5f2481 899 list_del_init(&req->r_req_lru_item);
f24e9980
SW		 900 ceph_osdc_put_request(req);
901
422d2cb8
YS		 902 if (osdc->num_requests == 0) {
903 dout(" no requests, canceling timeout\n");
904 __cancel_osd_timeout(osdc);
f24e9980
SW		 905 }
906}
907
908/*
909 * Cancel a previously queued request message
910 */
911static void __cancel_request(struct ceph_osd_request *req)
912{
6bc18876 913 if (req->r_sent && req->r_osd) {
6740a845 914 ceph_msg_revoke(req->r_request);
f24e9980
SW		 915 req->r_sent = 0;
916 }
917}
918
a40c4f10
YS		 919static void __register_linger_request(struct ceph_osd_client *osdc,
920 struct ceph_osd_request *req)
921{
922 dout("__register_linger_request %p\n", req);
923 list_add_tail(&req->r_linger_item, &osdc->req_linger);
6194ea89
SW		 924 if (req->r_osd)
925 list_add_tail(&req->r_linger_osd,
926 &req->r_osd->o_linger_requests);
a40c4f10
YS		 927}
928
929static void __unregister_linger_request(struct ceph_osd_client *osdc,
930 struct ceph_osd_request *req)
931{
932 dout("__unregister_linger_request %p\n", req);
61c74035 933 list_del_init(&req->r_linger_item);
a40c4f10 934 if (req->r_osd) {
a40c4f10
YS		 935 list_del_init(&req->r_linger_osd);
936
937 if (list_empty(&req->r_osd->o_requests) &&
938 list_empty(&req->r_osd->o_linger_requests)) {
939 dout("moving osd to %p lru\n", req->r_osd);
940 __move_osd_to_lru(osdc, req->r_osd);
941 }
fbdb9190
SW		 942 if (list_empty(&req->r_osd_item))
943 req->r_osd = NULL;
a40c4f10
YS		 944 }
945}
946
947void ceph_osdc_unregister_linger_request(struct ceph_osd_client *osdc,
948 struct ceph_osd_request *req)
949{
950 mutex_lock(&osdc->request_mutex);
951 if (req->r_linger) {
952 __unregister_linger_request(osdc, req);
953 ceph_osdc_put_request(req);
954 }
955 mutex_unlock(&osdc->request_mutex);
956}
957EXPORT_SYMBOL(ceph_osdc_unregister_linger_request);
958
959void ceph_osdc_set_request_linger(struct ceph_osd_client *osdc,
960 struct ceph_osd_request *req)
961{
962 if (!req->r_linger) {
963 dout("set_request_linger %p\n", req);
964 req->r_linger = 1;
965 /*
966 * caller is now responsible for calling
967 * unregister_linger_request
968 */
969 ceph_osdc_get_request(req);
970 }
971}
972EXPORT_SYMBOL(ceph_osdc_set_request_linger);
973
f24e9980
SW		 974/*
975 * Pick an osd (the first 'up' osd in the pg), allocate the osd struct
976 * (as needed), and set the request r_osd appropriately. If there is
25985edc 977 * no up osd, set r_osd to NULL. Move the request to the appropriate list
6f6c7006 978 * (unsent, homeless) or leave on in-flight lru.
f24e9980
SW		 979 *
980 * Return 0 if unchanged, 1 if changed, or negative on error.
981 *
982 * Caller should hold map_sem for read and request_mutex.
983 */
6f6c7006 984static int __map_request(struct ceph_osd_client *osdc,
38d6453c 985 struct ceph_osd_request *req, int force_resend)
f24e9980 986{
5b191d99 987 struct ceph_pg pgid;
d85b7056
SW		 988 int acting[CEPH_PG_MAX_SIZE];
989 int o = -1, num = 0;
f24e9980 990 int err;
f24e9980 991
6f6c7006 992 dout("map_request %p tid %lld\n", req, req->r_tid);
41766f87
AE		 993 err = ceph_calc_ceph_pg(&pgid, req->r_oid, osdc->osdmap,
994 ceph_file_layout_pg_pool(req->r_file_layout));
6f6c7006
SW		 995 if (err) {
996 list_move(&req->r_req_lru_item, &osdc->req_notarget);
f24e9980 997 return err;
6f6c7006 998 }
7740a42f
SW		 999 req->r_pgid = pgid;
1000
d85b7056
SW		 1001 err = ceph_calc_pg_acting(osdc->osdmap, pgid, acting);
1002 if (err > 0) {
1003 o = acting[0];
1004 num = err;
1005 }
f24e9980 1006
38d6453c
SW		 1007 if ((!force_resend &&
1008 req->r_osd && req->r_osd->o_osd == o &&
d85b7056
SW		 1009 req->r_sent >= req->r_osd->o_incarnation &&
1010 req->r_num_pg_osds == num &&
1011 memcmp(req->r_pg_osds, acting, sizeof(acting[0])*num) == 0) ||
f24e9980
SW		 1012 (req->r_osd == NULL && o == -1))
1013 return 0; /* no change */
1014
5b191d99
SW		 1015 dout("map_request tid %llu pgid %lld.%x osd%d (was osd%d)\n",
1016 req->r_tid, pgid.pool, pgid.seed, o,
f24e9980
SW		 1017 req->r_osd ? req->r_osd->o_osd : -1);
1018
d85b7056
SW		 1019 /* record full pg acting set */
1020 memcpy(req->r_pg_osds, acting, sizeof(acting[0]) * num);
1021 req->r_num_pg_osds = num;
1022
f24e9980
SW		 1023 if (req->r_osd) {
1024 __cancel_request(req);
1025 list_del_init(&req->r_osd_item);
f24e9980
SW		 1026 req->r_osd = NULL;
1027 }
1028
1029 req->r_osd = __lookup_osd(osdc, o);
1030 if (!req->r_osd && o >= 0) {
c99eb1c7 1031 err = -ENOMEM;
e10006f8 1032 req->r_osd = create_osd(osdc, o);
6f6c7006
SW		 1033 if (!req->r_osd) {
1034 list_move(&req->r_req_lru_item, &osdc->req_notarget);
c99eb1c7 1035 goto out;
6f6c7006 1036 }
f24e9980 1037
6f6c7006 1038 dout("map_request osd %p is osd%d\n", req->r_osd, o);
f24e9980
SW		 1039 __insert_osd(osdc, req->r_osd);
1040
b7a9e5dd
SW		 1041 ceph_con_open(&req->r_osd->o_con,
1042 CEPH_ENTITY_TYPE_OSD, o,
1043 &osdc->osdmap->osd_addr[o]);
f24e9980
SW		 1044 }
1045
f5a2041b
YS		 1046 if (req->r_osd) {
1047 __remove_osd_from_lru(req->r_osd);
ad885927
AE		 1048 list_add_tail(&req->r_osd_item, &req->r_osd->o_requests);
1049 list_move_tail(&req->r_req_lru_item, &osdc->req_unsent);
6f6c7006 1050 } else {
ad885927 1051 list_move_tail(&req->r_req_lru_item, &osdc->req_notarget);
f5a2041b 1052 }
d85b7056 1053 err = 1; /* osd or pg changed */
f24e9980
SW		 1054
1055out:
f24e9980
SW		 1056 return err;
1057}
1058
1059/*
1060 * caller should hold map_sem (for read) and request_mutex
1061 */
56e925b6
SW		 1062static void __send_request(struct ceph_osd_client *osdc,
1063 struct ceph_osd_request *req)
f24e9980 1064{
1b83bef2 1065 void *p;
f24e9980 1066
1b83bef2
SW		 1067 dout("send_request %p tid %llu to osd%d flags %d pg %lld.%x\n",
1068 req, req->r_tid, req->r_osd->o_osd, req->r_flags,
1069 (unsigned long long)req->r_pgid.pool, req->r_pgid.seed);
1070
1071 /* fill in message content that changes each time we send it */
1072 put_unaligned_le32(osdc->osdmap->epoch, req->r_request_osdmap_epoch);
1073 put_unaligned_le32(req->r_flags, req->r_request_flags);
1074 put_unaligned_le64(req->r_pgid.pool, req->r_request_pool);
1075 p = req->r_request_pgid;
1076 ceph_encode_64(&p, req->r_pgid.pool);
1077 ceph_encode_32(&p, req->r_pgid.seed);
1078 put_unaligned_le64(1, req->r_request_attempts); /* FIXME */
1079 memcpy(req->r_request_reassert_version, &req->r_reassert_version,
1080 sizeof(req->r_reassert_version));
2169aea6 1081
3dd72fc0 1082 req->r_stamp = jiffies;
07a27e22 1083 list_move_tail(&req->r_req_lru_item, &osdc->req_lru);
f24e9980
SW		 1084
1085 ceph_msg_get(req->r_request); /* send consumes a ref */
1086 ceph_con_send(&req->r_osd->o_con, req->r_request);
1087 req->r_sent = req->r_osd->o_incarnation;
f24e9980
SW		 1088}
1089
6f6c7006
SW		 1090/*
1091 * Send any requests in the queue (req_unsent).
1092 */
f9d25199 1093static void __send_queued(struct ceph_osd_client *osdc)
6f6c7006
SW		 1094{
1095 struct ceph_osd_request *req, *tmp;
1096
f9d25199
AE		 1097 dout("__send_queued\n");
1098 list_for_each_entry_safe(req, tmp, &osdc->req_unsent, r_req_lru_item)
6f6c7006 1099 __send_request(osdc, req);
6f6c7006
SW		 1100}
1101
f24e9980
SW		 1102/*
1103 * Timeout callback, called every N seconds when 1 or more osd
1104 * requests has been active for more than N seconds. When this
1105 * happens, we ping all OSDs with requests who have timed out to
1106 * ensure any communications channel reset is detected. Reset the
1107 * request timeouts another N seconds in the future as we go.
1108 * Reschedule the timeout event another N seconds in future (unless
1109 * there are no open requests).
1110 */
1111static void handle_timeout(struct work_struct *work)
1112{
1113 struct ceph_osd_client *osdc =
1114 container_of(work, struct ceph_osd_client, timeout_work.work);
83aff95e 1115 struct ceph_osd_request *req;
f24e9980 1116 struct ceph_osd *osd;
422d2cb8 1117 unsigned long keepalive =
3d14c5d2 1118 osdc->client->options->osd_keepalive_timeout * HZ;
422d2cb8 1119 struct list_head slow_osds;
f24e9980
SW		 1120 dout("timeout\n");
1121 down_read(&osdc->map_sem);
1122
1123 ceph_monc_request_next_osdmap(&osdc->client->monc);
1124
1125 mutex_lock(&osdc->request_mutex);
f24e9980 1126
422d2cb8
YS		 1127 /*
1128 * ping osds that are a bit slow. this ensures that if there
1129 * is a break in the TCP connection we will notice, and reopen
1130 * a connection with that osd (from the fault callback).
1131 */
1132 INIT_LIST_HEAD(&slow_osds);
1133 list_for_each_entry(req, &osdc->req_lru, r_req_lru_item) {
3dd72fc0 1134 if (time_before(jiffies, req->r_stamp + keepalive))
422d2cb8
YS		 1135 break;
1136
1137 osd = req->r_osd;
1138 BUG_ON(!osd);
1139 dout(" tid %llu is slow, will send keepalive on osd%d\n",
f24e9980 1140 req->r_tid, osd->o_osd);
422d2cb8
YS		 1141 list_move_tail(&osd->o_keepalive_item, &slow_osds);
1142 }
1143 while (!list_empty(&slow_osds)) {
1144 osd = list_entry(slow_osds.next, struct ceph_osd,
1145 o_keepalive_item);
1146 list_del_init(&osd->o_keepalive_item);
f24e9980
SW		 1147 ceph_con_keepalive(&osd->o_con);
1148 }
1149
422d2cb8 1150 __schedule_osd_timeout(osdc);
f9d25199 1151 __send_queued(osdc);
f24e9980 1152 mutex_unlock(&osdc->request_mutex);
f24e9980
SW		 1153 up_read(&osdc->map_sem);
1154}
1155
f5a2041b
YS		 1156static void handle_osds_timeout(struct work_struct *work)
1157{
1158 struct ceph_osd_client *osdc =
1159 container_of(work, struct ceph_osd_client,
1160 osds_timeout_work.work);
1161 unsigned long delay =
3d14c5d2 1162 osdc->client->options->osd_idle_ttl * HZ >> 2;
f5a2041b
YS		 1163
1164 dout("osds timeout\n");
1165 down_read(&osdc->map_sem);
aca420bc 1166 remove_old_osds(osdc);
f5a2041b
YS		 1167 up_read(&osdc->map_sem);
1168
1169 schedule_delayed_work(&osdc->osds_timeout_work,
1170 round_jiffies_relative(delay));
1171}
1172
25845472
SW		 1173static void complete_request(struct ceph_osd_request *req)
1174{
1175 if (req->r_safe_callback)
1176 req->r_safe_callback(req, NULL);
1177 complete_all(&req->r_safe_completion); /* fsync waiter */
1178}
1179
1b83bef2
SW		 1180static int __decode_pgid(void **p, void *end, struct ceph_pg *pgid)
1181{
1182 __u8 v;
1183
1184 ceph_decode_need(p, end, 1 + 8 + 4 + 4, bad);
1185 v = ceph_decode_8(p);
1186 if (v > 1) {
1187 pr_warning("do not understand pg encoding %d > 1", v);
1188 return -EINVAL;
1189 }
1190 pgid->pool = ceph_decode_64(p);
1191 pgid->seed = ceph_decode_32(p);
1192 *p += 4;
1193 return 0;
1194
1195bad:
1196 pr_warning("incomplete pg encoding");
1197 return -EINVAL;
1198}
1199
f24e9980
SW		 1200/*
1201 * handle osd op reply. either call the callback if it is specified,
1202 * or do the completion to wake up the waiting thread.
1203 */
350b1c32
SW		 1204static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg,
1205 struct ceph_connection *con)
f24e9980 1206{
1b83bef2 1207 void *p, *end;
f24e9980
SW		 1208 struct ceph_osd_request *req;
1209 u64 tid;
1b83bef2
SW		 1210 int object_len;
1211 int numops, payload_len, flags;
0ceed5db 1212 s32 result;
1b83bef2
SW		 1213 s32 retry_attempt;
1214 struct ceph_pg pg;
1215 int err;
1216 u32 reassert_epoch;
1217 u64 reassert_version;
1218 u32 osdmap_epoch;
0d5af164 1219 int already_completed;
1b83bef2 1220 int i;
f24e9980 1221
6df058c0 1222 tid = le64_to_cpu(msg->hdr.tid);
1b83bef2
SW		 1223 dout("handle_reply %p tid %llu\n", msg, tid);
1224
1225 p = msg->front.iov_base;
1226 end = p + msg->front.iov_len;
1227
1228 ceph_decode_need(&p, end, 4, bad);
1229 object_len = ceph_decode_32(&p);
1230 ceph_decode_need(&p, end, object_len, bad);
1231 p += object_len;
1232
1233 err = __decode_pgid(&p, end, &pg);
1234 if (err)
f24e9980 1235 goto bad;
1b83bef2
SW		 1236
1237 ceph_decode_need(&p, end, 8 + 4 + 4 + 8 + 4, bad);
1238 flags = ceph_decode_64(&p);
1239 result = ceph_decode_32(&p);
1240 reassert_epoch = ceph_decode_32(&p);
1241 reassert_version = ceph_decode_64(&p);
1242 osdmap_epoch = ceph_decode_32(&p);
1243
f24e9980
SW		 1244 /* lookup */
1245 mutex_lock(&osdc->request_mutex);
1246 req = __lookup_request(osdc, tid);
1247 if (req == NULL) {
1248 dout("handle_reply tid %llu dne\n", tid);
1249 mutex_unlock(&osdc->request_mutex);
1250 return;
1251 }
1252 ceph_osdc_get_request(req);
1b83bef2
SW		 1253
1254 dout("handle_reply %p tid %llu req %p result %d\n", msg, tid,
1255 req, result);
1256
1257 ceph_decode_need(&p, end, 4, bad);
1258 numops = ceph_decode_32(&p);
1259 if (numops > CEPH_OSD_MAX_OP)
1260 goto bad_put;
1261 if (numops != req->r_num_ops)
1262 goto bad_put;
1263 payload_len = 0;
1264 ceph_decode_need(&p, end, numops * sizeof(struct ceph_osd_op), bad);
1265 for (i = 0; i < numops; i++) {
1266 struct ceph_osd_op *op = p;
1267 int len;
1268
1269 len = le32_to_cpu(op->payload_len);
1270 req->r_reply_op_len[i] = len;
1271 dout(" op %d has %d bytes\n", i, len);
1272 payload_len += len;
1273 p += sizeof(*op);
1274 }
1275 if (payload_len != le32_to_cpu(msg->hdr.data_len)) {
1276 pr_warning("sum of op payload lens %d != data_len %d",
1277 payload_len, le32_to_cpu(msg->hdr.data_len));
1278 goto bad_put;
1279 }
1280
1281 ceph_decode_need(&p, end, 4 + numops * 4, bad);
1282 retry_attempt = ceph_decode_32(&p);
1283 for (i = 0; i < numops; i++)
1284 req->r_reply_op_result[i] = ceph_decode_32(&p);
f24e9980 1285
350b1c32 1286 /*
0d59ab81 1287 * if this connection filled our message, drop our reference now, to
350b1c32
SW		 1288 * avoid a (safe but slower) revoke later.
1289 */
0d59ab81 1290 if (req->r_con_filling_msg == con && req->r_reply == msg) {
c16e7869 1291 dout(" dropping con_filling_msg ref %p\n", con);
0d59ab81 1292 req->r_con_filling_msg = NULL;
0d47766f 1293 con->ops->put(con);
350b1c32
SW		 1294 }
1295
f24e9980 1296 if (!req->r_got_reply) {
95c96174 1297 unsigned int bytes;
f24e9980 1298
1b83bef2 1299 req->r_result = result;
f24e9980
SW		 1300 bytes = le32_to_cpu(msg->hdr.data_len);
1301 dout("handle_reply result %d bytes %d\n", req->r_result,
1302 bytes);
1303 if (req->r_result == 0)
1304 req->r_result = bytes;
1305
1306 /* in case this is a write and we need to replay, */
1b83bef2
SW		 1307 req->r_reassert_version.epoch = cpu_to_le32(reassert_epoch);
1308 req->r_reassert_version.version = cpu_to_le64(reassert_version);
f24e9980
SW		 1309
1310 req->r_got_reply = 1;
1311 } else if ((flags & CEPH_OSD_FLAG_ONDISK) == 0) {
1312 dout("handle_reply tid %llu dup ack\n", tid);
34b43a56 1313 mutex_unlock(&osdc->request_mutex);
f24e9980
SW		 1314 goto done;
1315 }
1316
1317 dout("handle_reply tid %llu flags %d\n", tid, flags);
1318
a40c4f10
YS		 1319 if (req->r_linger && (flags & CEPH_OSD_FLAG_ONDISK))
1320 __register_linger_request(osdc, req);
1321
f24e9980 1322 /* either this is a read, or we got the safe response */
0ceed5db
SW		 1323 if (result < 0 ||
1324 (flags & CEPH_OSD_FLAG_ONDISK) ||
f24e9980
SW		 1325 ((flags & CEPH_OSD_FLAG_WRITE) == 0))
1326 __unregister_request(osdc, req);
1327
0d5af164
AE		 1328 already_completed = req->r_completed;
1329 req->r_completed = 1;
f24e9980 1330 mutex_unlock(&osdc->request_mutex);
0d5af164
AE		 1331 if (already_completed)
1332 goto done;
f24e9980
SW		 1333
1334 if (req->r_callback)
1335 req->r_callback(req, msg);
1336 else
03066f23 1337 complete_all(&req->r_completion);
f24e9980 1338
25845472
SW		 1339 if (flags & CEPH_OSD_FLAG_ONDISK)
1340 complete_request(req);
f24e9980
SW		 1341
1342done:
a40c4f10 1343 dout("req=%p req->r_linger=%d\n", req, req->r_linger);
f24e9980
SW		 1344 ceph_osdc_put_request(req);
1345 return;
1346
1b83bef2
SW		 1347bad_put:
1348 ceph_osdc_put_request(req);
f24e9980 1349bad:
1b83bef2
SW		 1350 pr_err("corrupt osd_op_reply got %d %d\n",
1351 (int)msg->front.iov_len, le32_to_cpu(msg->hdr.front_len));
9ec7cab1 1352 ceph_msg_dump(msg);
f24e9980
SW		 1353}
1354
6f6c7006 1355static void reset_changed_osds(struct ceph_osd_client *osdc)
f24e9980 1356{
f24e9980 1357 struct rb_node *p, *n;
f24e9980 1358
6f6c7006
SW		 1359 for (p = rb_first(&osdc->osds); p; p = n) {
1360 struct ceph_osd *osd = rb_entry(p, struct ceph_osd, o_node);
f24e9980 1361
6f6c7006
SW		 1362 n = rb_next(p);
1363 if (!ceph_osd_is_up(osdc->osdmap, osd->o_osd) ||
1364 memcmp(&osd->o_con.peer_addr,
1365 ceph_osd_addr(osdc->osdmap,
1366 osd->o_osd),
1367 sizeof(struct ceph_entity_addr)) != 0)
1368 __reset_osd(osdc, osd);
f24e9980 1369 }
422d2cb8
YS		 1370}
1371
1372/*
6f6c7006
SW		 1373 * Requeue requests whose mapping to an OSD has changed. If requests map to
1374 * no osd, request a new map.
422d2cb8 1375 *
e6d50f67 1376 * Caller should hold map_sem for read.
422d2cb8 1377 */
38d6453c 1378static void kick_requests(struct ceph_osd_client *osdc, int force_resend)
422d2cb8 1379{
a40c4f10 1380 struct ceph_osd_request *req, *nreq;
6f6c7006
SW		 1381 struct rb_node *p;
1382 int needmap = 0;
1383 int err;
422d2cb8 1384
38d6453c 1385 dout("kick_requests %s\n", force_resend ? " (force resend)" : "");
422d2cb8 1386 mutex_lock(&osdc->request_mutex);
6194ea89 1387 for (p = rb_first(&osdc->requests); p; ) {
6f6c7006 1388 req = rb_entry(p, struct ceph_osd_request, r_node);
6194ea89 1389 p = rb_next(p);
ab60b16d
AE		 1390
1391 /*
1392 * For linger requests that have not yet been
1393 * registered, move them to the linger list; they'll
1394 * be sent to the osd in the loop below. Unregister
1395 * the request before re-registering it as a linger
1396 * request to ensure the __map_request() below
1397 * will decide it needs to be sent.
1398 */
1399 if (req->r_linger && list_empty(&req->r_linger_item)) {
1400 dout("%p tid %llu restart on osd%d\n",
1401 req, req->r_tid,
1402 req->r_osd ? req->r_osd->o_osd : -1);
1403 __unregister_request(osdc, req);
1404 __register_linger_request(osdc, req);
1405 continue;
1406 }
1407
38d6453c 1408 err = __map_request(osdc, req, force_resend);
6f6c7006
SW		 1409 if (err < 0)
1410 continue; /* error */
1411 if (req->r_osd == NULL) {
1412 dout("%p tid %llu maps to no osd\n", req, req->r_tid);
1413 needmap++; /* request a newer map */
1414 } else if (err > 0) {
6194ea89
SW		 1415 if (!req->r_linger) {
1416 dout("%p tid %llu requeued on osd%d\n", req,
1417 req->r_tid,
1418 req->r_osd ? req->r_osd->o_osd : -1);
a40c4f10 1419 req->r_flags |= CEPH_OSD_FLAG_RETRY;
6194ea89
SW		 1420 }
1421 }
a40c4f10
YS		 1422 }
1423
1424 list_for_each_entry_safe(req, nreq, &osdc->req_linger,
1425 r_linger_item) {
1426 dout("linger req=%p req->r_osd=%p\n", req, req->r_osd);
1427
38d6453c 1428 err = __map_request(osdc, req, force_resend);
ab60b16d 1429 dout("__map_request returned %d\n", err);
a40c4f10
YS		 1430 if (err == 0)
1431 continue; /* no change and no osd was specified */
1432 if (err < 0)
1433 continue; /* hrm! */
1434 if (req->r_osd == NULL) {
1435 dout("tid %llu maps to no valid osd\n", req->r_tid);
1436 needmap++; /* request a newer map */
1437 continue;
6f6c7006 1438 }
a40c4f10
YS		 1439
1440 dout("kicking lingering %p tid %llu osd%d\n", req, req->r_tid,
1441 req->r_osd ? req->r_osd->o_osd : -1);
a40c4f10 1442 __register_request(osdc, req);
c89ce05e 1443 __unregister_linger_request(osdc, req);
6f6c7006 1444 }
f24e9980
SW		 1445 mutex_unlock(&osdc->request_mutex);
1446
1447 if (needmap) {
1448 dout("%d requests for down osds, need new map\n", needmap);
1449 ceph_monc_request_next_osdmap(&osdc->client->monc);
1450 }
e6d50f67 1451 reset_changed_osds(osdc);
422d2cb8 1452}
6f6c7006
SW		 1453
1454
f24e9980
SW		 1455/*
1456 * Process updated osd map.
1457 *
1458 * The message contains any number of incremental and full maps, normally
1459 * indicating some sort of topology change in the cluster. Kick requests
1460 * off to different OSDs as needed.
1461 */
1462void ceph_osdc_handle_map(struct ceph_osd_client *osdc, struct ceph_msg *msg)
1463{
1464 void *p, *end, *next;
1465 u32 nr_maps, maplen;
1466 u32 epoch;
1467 struct ceph_osdmap *newmap = NULL, *oldmap;
1468 int err;
1469 struct ceph_fsid fsid;
1470
1471 dout("handle_map have %u\n", osdc->osdmap ? osdc->osdmap->epoch : 0);
1472 p = msg->front.iov_base;
1473 end = p + msg->front.iov_len;
1474
1475 /* verify fsid */
1476 ceph_decode_need(&p, end, sizeof(fsid), bad);
1477 ceph_decode_copy(&p, &fsid, sizeof(fsid));
0743304d
SW		 1478 if (ceph_check_fsid(osdc->client, &fsid) < 0)
1479 return;
f24e9980
SW		 1480
1481 down_write(&osdc->map_sem);
1482
1483 /* incremental maps */
1484 ceph_decode_32_safe(&p, end, nr_maps, bad);
1485 dout(" %d inc maps\n", nr_maps);
1486 while (nr_maps > 0) {
1487 ceph_decode_need(&p, end, 2*sizeof(u32), bad);
c89136ea
SW		 1488 epoch = ceph_decode_32(&p);
1489 maplen = ceph_decode_32(&p);
f24e9980
SW		 1490 ceph_decode_need(&p, end, maplen, bad);
1491 next = p + maplen;
1492 if (osdc->osdmap && osdc->osdmap->epoch+1 == epoch) {
1493 dout("applying incremental map %u len %d\n",
1494 epoch, maplen);
1495 newmap = osdmap_apply_incremental(&p, next,
1496 osdc->osdmap,
15d9882c 1497 &osdc->client->msgr);
f24e9980
SW		 1498 if (IS_ERR(newmap)) {
1499 err = PTR_ERR(newmap);
1500 goto bad;
1501 }
30dc6381 1502 BUG_ON(!newmap);
f24e9980
SW		 1503 if (newmap != osdc->osdmap) {
1504 ceph_osdmap_destroy(osdc->osdmap);
1505 osdc->osdmap = newmap;
1506 }
38d6453c 1507 kick_requests(osdc, 0);
f24e9980
SW		 1508 } else {
1509 dout("ignoring incremental map %u len %d\n",
1510 epoch, maplen);
1511 }
1512 p = next;
1513 nr_maps--;
1514 }
1515 if (newmap)
1516 goto done;
1517
1518 /* full maps */
1519 ceph_decode_32_safe(&p, end, nr_maps, bad);
1520 dout(" %d full maps\n", nr_maps);
1521 while (nr_maps) {
1522 ceph_decode_need(&p, end, 2*sizeof(u32), bad);
c89136ea
SW		 1523 epoch = ceph_decode_32(&p);
1524 maplen = ceph_decode_32(&p);
f24e9980
SW		 1525 ceph_decode_need(&p, end, maplen, bad);
1526 if (nr_maps > 1) {
1527 dout("skipping non-latest full map %u len %d\n",
1528 epoch, maplen);
1529 } else if (osdc->osdmap && osdc->osdmap->epoch >= epoch) {
1530 dout("skipping full map %u len %d, "
1531 "older than our %u\n", epoch, maplen,
1532 osdc->osdmap->epoch);
1533 } else {
38d6453c
SW		 1534 int skipped_map = 0;
1535
f24e9980
SW		 1536 dout("taking full map %u len %d\n", epoch, maplen);
1537 newmap = osdmap_decode(&p, p+maplen);
1538 if (IS_ERR(newmap)) {
1539 err = PTR_ERR(newmap);
1540 goto bad;
1541 }
30dc6381 1542 BUG_ON(!newmap);
f24e9980
SW		 1543 oldmap = osdc->osdmap;
1544 osdc->osdmap = newmap;
38d6453c
SW		 1545 if (oldmap) {
1546 if (oldmap->epoch + 1 < newmap->epoch)
1547 skipped_map = 1;
f24e9980 1548 ceph_osdmap_destroy(oldmap);
38d6453c
SW		 1549 }
1550 kick_requests(osdc, skipped_map);
f24e9980
SW		 1551 }
1552 p += maplen;
1553 nr_maps--;
1554 }
1555
1556done:
1557 downgrade_write(&osdc->map_sem);
1558 ceph_monc_got_osdmap(&osdc->client->monc, osdc->osdmap->epoch);
cd634fb6
SW		 1559
1560 /*
1561 * subscribe to subsequent osdmap updates if full to ensure
1562 * we find out when we are no longer full and stop returning
1563 * ENOSPC.
1564 */
1565 if (ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL))
1566 ceph_monc_request_next_osdmap(&osdc->client->monc);
1567
f9d25199
AE		 1568 mutex_lock(&osdc->request_mutex);
1569 __send_queued(osdc);
1570 mutex_unlock(&osdc->request_mutex);
f24e9980 1571 up_read(&osdc->map_sem);
03066f23 1572 wake_up_all(&osdc->client->auth_wq);
f24e9980
SW		 1573 return;
1574
1575bad:
1576 pr_err("osdc handle_map corrupt msg\n");
9ec7cab1 1577 ceph_msg_dump(msg);
f24e9980
SW		 1578 up_write(&osdc->map_sem);
1579 return;
1580}
1581
a40c4f10
YS		 1582/*
1583 * watch/notify callback event infrastructure
1584 *
1585 * These callbacks are used both for watch and notify operations.
1586 */
1587static void __release_event(struct kref *kref)
1588{
1589 struct ceph_osd_event *event =
1590 container_of(kref, struct ceph_osd_event, kref);
1591
1592 dout("__release_event %p\n", event);
1593 kfree(event);
1594}
1595
1596static void get_event(struct ceph_osd_event *event)
1597{
1598 kref_get(&event->kref);
1599}
1600
1601void ceph_osdc_put_event(struct ceph_osd_event *event)
1602{
1603 kref_put(&event->kref, __release_event);
1604}
1605EXPORT_SYMBOL(ceph_osdc_put_event);
1606
1607static void __insert_event(struct ceph_osd_client *osdc,
1608 struct ceph_osd_event *new)
1609{
1610 struct rb_node **p = &osdc->event_tree.rb_node;
1611 struct rb_node *parent = NULL;
1612 struct ceph_osd_event *event = NULL;
1613
1614 while (*p) {
1615 parent = *p;
1616 event = rb_entry(parent, struct ceph_osd_event, node);
1617 if (new->cookie < event->cookie)
1618 p = &(*p)->rb_left;
1619 else if (new->cookie > event->cookie)
1620 p = &(*p)->rb_right;
1621 else
1622 BUG();
1623 }
1624
1625 rb_link_node(&new->node, parent, p);
1626 rb_insert_color(&new->node, &osdc->event_tree);
1627}
1628
1629static struct ceph_osd_event *__find_event(struct ceph_osd_client *osdc,
1630 u64 cookie)
1631{
1632 struct rb_node **p = &osdc->event_tree.rb_node;
1633 struct rb_node *parent = NULL;
1634 struct ceph_osd_event *event = NULL;
1635
1636 while (*p) {
1637 parent = *p;
1638 event = rb_entry(parent, struct ceph_osd_event, node);
1639 if (cookie < event->cookie)
1640 p = &(*p)->rb_left;
1641 else if (cookie > event->cookie)
1642 p = &(*p)->rb_right;
1643 else
1644 return event;
1645 }
1646 return NULL;
1647}
1648
1649static void __remove_event(struct ceph_osd_event *event)
1650{
1651 struct ceph_osd_client *osdc = event->osdc;
1652
1653 if (!RB_EMPTY_NODE(&event->node)) {
1654 dout("__remove_event removed %p\n", event);
1655 rb_erase(&event->node, &osdc->event_tree);
1656 ceph_osdc_put_event(event);
1657 } else {
1658 dout("__remove_event didn't remove %p\n", event);
1659 }
1660}
1661
1662int ceph_osdc_create_event(struct ceph_osd_client *osdc,
1663 void (*event_cb)(u64, u64, u8, void *),
3c663bbd 1664 void *data, struct ceph_osd_event **pevent)
a40c4f10
YS		 1665{
1666 struct ceph_osd_event *event;
1667
1668 event = kmalloc(sizeof(*event), GFP_NOIO);
1669 if (!event)
1670 return -ENOMEM;
1671
1672 dout("create_event %p\n", event);
1673 event->cb = event_cb;
3c663bbd 1674 event->one_shot = 0;
a40c4f10
YS		 1675 event->data = data;
1676 event->osdc = osdc;
1677 INIT_LIST_HEAD(&event->osd_node);
3ee5234d 1678 RB_CLEAR_NODE(&event->node);
a40c4f10
YS		 1679 kref_init(&event->kref); /* one ref for us */
1680 kref_get(&event->kref); /* one ref for the caller */
a40c4f10
YS		 1681
1682 spin_lock(&osdc->event_lock);
1683 event->cookie = ++osdc->event_count;
1684 __insert_event(osdc, event);
1685 spin_unlock(&osdc->event_lock);
1686
1687 *pevent = event;
1688 return 0;
1689}
1690EXPORT_SYMBOL(ceph_osdc_create_event);
1691
1692void ceph_osdc_cancel_event(struct ceph_osd_event *event)
1693{
1694 struct ceph_osd_client *osdc = event->osdc;
1695
1696 dout("cancel_event %p\n", event);
1697 spin_lock(&osdc->event_lock);
1698 __remove_event(event);
1699 spin_unlock(&osdc->event_lock);
1700 ceph_osdc_put_event(event); /* caller's */
1701}
1702EXPORT_SYMBOL(ceph_osdc_cancel_event);
1703
1704
1705static void do_event_work(struct work_struct *work)
1706{
1707 struct ceph_osd_event_work *event_work =
1708 container_of(work, struct ceph_osd_event_work, work);
1709 struct ceph_osd_event *event = event_work->event;
1710 u64 ver = event_work->ver;
1711 u64 notify_id = event_work->notify_id;
1712 u8 opcode = event_work->opcode;
1713
1714 dout("do_event_work completing %p\n", event);
1715 event->cb(ver, notify_id, opcode, event->data);
a40c4f10
YS		 1716 dout("do_event_work completed %p\n", event);
1717 ceph_osdc_put_event(event);
1718 kfree(event_work);
1719}
1720
1721
1722/*
1723 * Process osd watch notifications
1724 */
3c663bbd
AE		 1725static void handle_watch_notify(struct ceph_osd_client *osdc,
1726 struct ceph_msg *msg)
a40c4f10
YS		 1727{
1728 void *p, *end;
1729 u8 proto_ver;
1730 u64 cookie, ver, notify_id;
1731 u8 opcode;
1732 struct ceph_osd_event *event;
1733 struct ceph_osd_event_work *event_work;
1734
1735 p = msg->front.iov_base;
1736 end = p + msg->front.iov_len;
1737
1738 ceph_decode_8_safe(&p, end, proto_ver, bad);
1739 ceph_decode_8_safe(&p, end, opcode, bad);
1740 ceph_decode_64_safe(&p, end, cookie, bad);
1741 ceph_decode_64_safe(&p, end, ver, bad);
1742 ceph_decode_64_safe(&p, end, notify_id, bad);
1743
1744 spin_lock(&osdc->event_lock);
1745 event = __find_event(osdc, cookie);
1746 if (event) {
3c663bbd 1747 BUG_ON(event->one_shot);
a40c4f10 1748 get_event(event);
a40c4f10
YS		 1749 }
1750 spin_unlock(&osdc->event_lock);
1751 dout("handle_watch_notify cookie %lld ver %lld event %p\n",
1752 cookie, ver, event);
1753 if (event) {
1754 event_work = kmalloc(sizeof(*event_work), GFP_NOIO);
a40c4f10
YS		 1755 if (!event_work) {
1756 dout("ERROR: could not allocate event_work\n");
1757 goto done_err;
1758 }
6b0ae409 1759 INIT_WORK(&event_work->work, do_event_work);
a40c4f10
YS		 1760 event_work->event = event;
1761 event_work->ver = ver;
1762 event_work->notify_id = notify_id;
1763 event_work->opcode = opcode;
1764 if (!queue_work(osdc->notify_wq, &event_work->work)) {
1765 dout("WARNING: failed to queue notify event work\n");
1766 goto done_err;
1767 }
1768 }
1769
1770 return;
1771
1772done_err:
a40c4f10
YS		 1773 ceph_osdc_put_event(event);
1774 return;
1775
1776bad:
1777 pr_err("osdc handle_watch_notify corrupt msg\n");
1778 return;
1779}
1780
70636773
AE		 1781static void ceph_osdc_msg_data_set(struct ceph_msg *msg,
1782 struct ceph_osd_data *osd_data)
f24e9980 1783{
0fff87ec 1784 if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGES) {
f1baeb2b 1785 BUG_ON(osd_data->length > (u64) SIZE_MAX);
ebf18f47 1786 if (osd_data->length)
70636773
AE		 1787 ceph_msg_data_set_pages(msg, osd_data->pages,
1788 osd_data->length, osd_data->alignment);
9a5e6d09
AE		 1789 } else if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGELIST) {
1790 BUG_ON(!osd_data->pagelist->length);
1791 ceph_msg_data_set_pagelist(msg, osd_data->pagelist);
68b4476b 1792#ifdef CONFIG_BLOCK
0fff87ec 1793 } else if (osd_data->type == CEPH_OSD_DATA_TYPE_BIO) {
70636773 1794 ceph_msg_data_set_bio(msg, osd_data->bio);
68b4476b 1795#endif
2ac2b7a6 1796 } else {
0fff87ec 1797 BUG_ON(osd_data->type != CEPH_OSD_DATA_TYPE_NONE);
2ac2b7a6 1798 }
70636773
AE		 1799}
1800
1801/*
1802 * Register request, send initial attempt.
1803 */
1804int ceph_osdc_start_request(struct ceph_osd_client *osdc,
1805 struct ceph_osd_request *req,
1806 bool nofail)
1807{
1808 int rc = 0;
1809
1810 /* Set up response incoming data and request outgoing data fields */
1811
1812 ceph_osdc_msg_data_set(req->r_reply, &req->r_data_in);
1813 ceph_osdc_msg_data_set(req->r_request, &req->r_data_out);
f24e9980 1814
f24e9980
SW		 1815 down_read(&osdc->map_sem);
1816 mutex_lock(&osdc->request_mutex);
dc4b870c 1817 __register_request(osdc, req);
92451b49
AE		 1818 WARN_ON(req->r_sent);
1819 rc = __map_request(osdc, req, 0);
1820 if (rc < 0) {
1821 if (nofail) {
1822 dout("osdc_start_request failed map, "
1823 " will retry %lld\n", req->r_tid);
1824 rc = 0;
f24e9980 1825 }
92451b49 1826 goto out_unlock;
f24e9980 1827 }
92451b49
AE		 1828 if (req->r_osd == NULL) {
1829 dout("send_request %p no up osds in pg\n", req);
1830 ceph_monc_request_next_osdmap(&osdc->client->monc);
1831 } else {
7e2766a1 1832 __send_queued(osdc);
92451b49
AE		 1833 }
1834 rc = 0;
234af26f 1835out_unlock:
f24e9980
SW		 1836 mutex_unlock(&osdc->request_mutex);
1837 up_read(&osdc->map_sem);
1838 return rc;
1839}
3d14c5d2 1840EXPORT_SYMBOL(ceph_osdc_start_request);
f24e9980
SW		 1841
1842/*
1843 * wait for a request to complete
1844 */
1845int ceph_osdc_wait_request(struct ceph_osd_client *osdc,
1846 struct ceph_osd_request *req)
1847{
1848 int rc;
1849
1850 rc = wait_for_completion_interruptible(&req->r_completion);
1851 if (rc < 0) {
1852 mutex_lock(&osdc->request_mutex);
1853 __cancel_request(req);
529cfcc4 1854 __unregister_request(osdc, req);
f24e9980 1855 mutex_unlock(&osdc->request_mutex);
25845472 1856 complete_request(req);
529cfcc4 1857 dout("wait_request tid %llu canceled/timed out\n", req->r_tid);
f24e9980
SW		 1858 return rc;
1859 }
1860
1861 dout("wait_request tid %llu result %d\n", req->r_tid, req->r_result);
1862 return req->r_result;
1863}
3d14c5d2 1864EXPORT_SYMBOL(ceph_osdc_wait_request);
f24e9980
SW		 1865
1866/*
1867 * sync - wait for all in-flight requests to flush. avoid starvation.
1868 */
1869void ceph_osdc_sync(struct ceph_osd_client *osdc)
1870{
1871 struct ceph_osd_request *req;
1872 u64 last_tid, next_tid = 0;
1873
1874 mutex_lock(&osdc->request_mutex);
1875 last_tid = osdc->last_tid;
1876 while (1) {
1877 req = __lookup_request_ge(osdc, next_tid);
1878 if (!req)
1879 break;
1880 if (req->r_tid > last_tid)
1881 break;
1882
1883 next_tid = req->r_tid + 1;
1884 if ((req->r_flags & CEPH_OSD_FLAG_WRITE) == 0)
1885 continue;
1886
1887 ceph_osdc_get_request(req);
1888 mutex_unlock(&osdc->request_mutex);
1889 dout("sync waiting on tid %llu (last is %llu)\n",
1890 req->r_tid, last_tid);
1891 wait_for_completion(&req->r_safe_completion);
1892 mutex_lock(&osdc->request_mutex);
1893 ceph_osdc_put_request(req);
1894 }
1895 mutex_unlock(&osdc->request_mutex);
1896 dout("sync done (thru tid %llu)\n", last_tid);
1897}
3d14c5d2 1898EXPORT_SYMBOL(ceph_osdc_sync);
f24e9980
SW		 1899
1900/*
1901 * init, shutdown
1902 */
1903int ceph_osdc_init(struct ceph_osd_client *osdc, struct ceph_client *client)
1904{
1905 int err;
1906
1907 dout("init\n");
1908 osdc->client = client;
1909 osdc->osdmap = NULL;
1910 init_rwsem(&osdc->map_sem);
1911 init_completion(&osdc->map_waiters);
1912 osdc->last_requested_map = 0;
1913 mutex_init(&osdc->request_mutex);
f24e9980
SW		 1914 osdc->last_tid = 0;
1915 osdc->osds = RB_ROOT;
f5a2041b 1916 INIT_LIST_HEAD(&osdc->osd_lru);
f24e9980 1917 osdc->requests = RB_ROOT;
422d2cb8 1918 INIT_LIST_HEAD(&osdc->req_lru);
6f6c7006
SW		 1919 INIT_LIST_HEAD(&osdc->req_unsent);
1920 INIT_LIST_HEAD(&osdc->req_notarget);
a40c4f10 1921 INIT_LIST_HEAD(&osdc->req_linger);
f24e9980
SW		 1922 osdc->num_requests = 0;
1923 INIT_DELAYED_WORK(&osdc->timeout_work, handle_timeout);
f5a2041b 1924 INIT_DELAYED_WORK(&osdc->osds_timeout_work, handle_osds_timeout);
a40c4f10
YS		 1925 spin_lock_init(&osdc->event_lock);
1926 osdc->event_tree = RB_ROOT;
1927 osdc->event_count = 0;
f5a2041b
YS		 1928
1929 schedule_delayed_work(&osdc->osds_timeout_work,
3d14c5d2 1930 round_jiffies_relative(osdc->client->options->osd_idle_ttl * HZ));
f24e9980 1931
5f44f142 1932 err = -ENOMEM;
f24e9980
SW		 1933 osdc->req_mempool = mempool_create_kmalloc_pool(10,
1934 sizeof(struct ceph_osd_request));
1935 if (!osdc->req_mempool)
5f44f142 1936 goto out;
f24e9980 1937
d50b409f
SW		 1938 err = ceph_msgpool_init(&osdc->msgpool_op, CEPH_MSG_OSD_OP,
1939 OSD_OP_FRONT_LEN, 10, true,
4f48280e 1940 "osd_op");
f24e9980 1941 if (err < 0)
5f44f142 1942 goto out_mempool;
d50b409f 1943 err = ceph_msgpool_init(&osdc->msgpool_op_reply, CEPH_MSG_OSD_OPREPLY,
4f48280e
SW		 1944 OSD_OPREPLY_FRONT_LEN, 10, true,
1945 "osd_op_reply");
c16e7869
SW		 1946 if (err < 0)
1947 goto out_msgpool;
a40c4f10
YS		 1948
1949 osdc->notify_wq = create_singlethread_workqueue("ceph-watch-notify");
1950 if (IS_ERR(osdc->notify_wq)) {
1951 err = PTR_ERR(osdc->notify_wq);
1952 osdc->notify_wq = NULL;
1953 goto out_msgpool;
1954 }
f24e9980 1955 return 0;
5f44f142 1956
c16e7869
SW		 1957out_msgpool:
1958 ceph_msgpool_destroy(&osdc->msgpool_op);
5f44f142
SW		 1959out_mempool:
1960 mempool_destroy(osdc->req_mempool);
1961out:
1962 return err;
f24e9980
SW		 1963}
1964
1965void ceph_osdc_stop(struct ceph_osd_client *osdc)
1966{
a40c4f10
YS		 1967 flush_workqueue(osdc->notify_wq);
1968 destroy_workqueue(osdc->notify_wq);
f24e9980 1969 cancel_delayed_work_sync(&osdc->timeout_work);
f5a2041b 1970 cancel_delayed_work_sync(&osdc->osds_timeout_work);
f24e9980
SW		 1971 if (osdc->osdmap) {
1972 ceph_osdmap_destroy(osdc->osdmap);
1973 osdc->osdmap = NULL;
1974 }
aca420bc 1975 remove_all_osds(osdc);
f24e9980
SW		 1976 mempool_destroy(osdc->req_mempool);
1977 ceph_msgpool_destroy(&osdc->msgpool_op);
c16e7869 1978 ceph_msgpool_destroy(&osdc->msgpool_op_reply);
f24e9980
SW		 1979}
1980
1981/*
1982 * Read some contiguous pages. If we cross a stripe boundary, shorten
1983 * *plen. Return number of bytes read, or error.
1984 */
1985int ceph_osdc_readpages(struct ceph_osd_client *osdc,
1986 struct ceph_vino vino, struct ceph_file_layout *layout,
1987 u64 off, u64 *plen,
1988 u32 truncate_seq, u64 truncate_size,
b7495fc2 1989 struct page **pages, int num_pages, int page_align)
f24e9980
SW		 1990{
1991 struct ceph_osd_request *req;
0fff87ec 1992 struct ceph_osd_data *osd_data;
f24e9980
SW		 1993 int rc = 0;
1994
1995 dout("readpages on ino %llx.%llx on %llu~%llu\n", vino.ino,
1996 vino.snap, off, *plen);
1997 req = ceph_osdc_new_request(osdc, layout, vino, off, plen,
1998 CEPH_OSD_OP_READ, CEPH_OSD_FLAG_READ,
1999 NULL, 0, truncate_seq, truncate_size, NULL,
153e5167 2000 false);
6816282d
SW		 2001 if (IS_ERR(req))
2002 return PTR_ERR(req);
f24e9980
SW		 2003
2004 /* it may be a short read due to an object boundary */
0fff87ec
AE		 2005
2006 osd_data = &req->r_data_in;
2007 osd_data->type = CEPH_OSD_DATA_TYPE_PAGES;
2008 osd_data->pages = pages;
e0c59487 2009 osd_data->length = *plen;
0fff87ec 2010 osd_data->alignment = page_align;
f24e9980 2011
e0c59487
AE		 2012 dout("readpages final extent is %llu~%llu (%llu bytes align %d)\n",
2013 off, *plen, osd_data->length, page_align);
f24e9980
SW		 2014
2015 rc = ceph_osdc_start_request(osdc, req, false);
2016 if (!rc)
2017 rc = ceph_osdc_wait_request(osdc, req);
2018
2019 ceph_osdc_put_request(req);
2020 dout("readpages result %d\n", rc);
2021 return rc;
2022}
3d14c5d2 2023EXPORT_SYMBOL(ceph_osdc_readpages);
f24e9980
SW		 2024
2025/*
2026 * do a synchronous write on N pages
2027 */
2028int ceph_osdc_writepages(struct ceph_osd_client *osdc, struct ceph_vino vino,
2029 struct ceph_file_layout *layout,
2030 struct ceph_snap_context *snapc,
2031 u64 off, u64 len,
2032 u32 truncate_seq, u64 truncate_size,
2033 struct timespec *mtime,
24808826 2034 struct page **pages, int num_pages)
f24e9980
SW		 2035{
2036 struct ceph_osd_request *req;
0fff87ec 2037 struct ceph_osd_data *osd_data;
f24e9980 2038 int rc = 0;
b7495fc2 2039 int page_align = off & ~PAGE_MASK;
f24e9980
SW		 2040
2041 BUG_ON(vino.snap != CEPH_NOSNAP);
2042 req = ceph_osdc_new_request(osdc, layout, vino, off, &len,
2043 CEPH_OSD_OP_WRITE,
24808826 2044 CEPH_OSD_FLAG_ONDISK | CEPH_OSD_FLAG_WRITE,
fbf8685f 2045 snapc, 0,
f24e9980 2046 truncate_seq, truncate_size, mtime,
153e5167 2047 true);
6816282d
SW		 2048 if (IS_ERR(req))
2049 return PTR_ERR(req);
f24e9980
SW		 2050
2051 /* it may be a short write due to an object boundary */
0fff87ec
AE		 2052 osd_data = &req->r_data_out;
2053 osd_data->type = CEPH_OSD_DATA_TYPE_PAGES;
2054 osd_data->pages = pages;
e0c59487 2055 osd_data->length = len;
0fff87ec 2056 osd_data->alignment = page_align;
e0c59487 2057 dout("writepages %llu~%llu (%llu bytes)\n", off, len, osd_data->length);
f24e9980 2058
87f979d3 2059 rc = ceph_osdc_start_request(osdc, req, true);
f24e9980
SW		 2060 if (!rc)
2061 rc = ceph_osdc_wait_request(osdc, req);
2062
2063 ceph_osdc_put_request(req);
2064 if (rc == 0)
2065 rc = len;
2066 dout("writepages result %d\n", rc);
2067 return rc;
2068}
3d14c5d2 2069EXPORT_SYMBOL(ceph_osdc_writepages);
f24e9980
SW		 2070
2071/*
2072 * handle incoming message
2073 */
2074static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
2075{
2076 struct ceph_osd *osd = con->private;
32c895e7 2077 struct ceph_osd_client *osdc;
f24e9980
SW		 2078 int type = le16_to_cpu(msg->hdr.type);
2079
2080 if (!osd)
4a32f93d 2081 goto out;
32c895e7 2082 osdc = osd->o_osdc;
f24e9980
SW		 2083
2084 switch (type) {
2085 case CEPH_MSG_OSD_MAP:
2086 ceph_osdc_handle_map(osdc, msg);
2087 break;
2088 case CEPH_MSG_OSD_OPREPLY:
350b1c32 2089 handle_reply(osdc, msg, con);
f24e9980 2090 break;
a40c4f10
YS		 2091 case CEPH_MSG_WATCH_NOTIFY:
2092 handle_watch_notify(osdc, msg);
2093 break;
f24e9980
SW		 2094
2095 default:
2096 pr_err("received unknown message type %d %s\n", type,
2097 ceph_msg_type_name(type));
2098 }
4a32f93d 2099out:
f24e9980
SW		 2100 ceph_msg_put(msg);
2101}
2102
5b3a4db3 2103/*
21b667f6
SW		 2104 * lookup and return message for incoming reply. set up reply message
2105 * pages.
5b3a4db3
SW		 2106 */
2107static struct ceph_msg *get_reply(struct ceph_connection *con,
2450418c
YS		 2108 struct ceph_msg_header *hdr,
2109 int *skip)
f24e9980
SW		 2110{
2111 struct ceph_osd *osd = con->private;
2112 struct ceph_osd_client *osdc = osd->o_osdc;
2450418c 2113 struct ceph_msg *m;
0547a9b3 2114 struct ceph_osd_request *req;
5b3a4db3
SW		 2115 int front = le32_to_cpu(hdr->front_len);
2116 int data_len = le32_to_cpu(hdr->data_len);
0547a9b3 2117 u64 tid;
f24e9980 2118
0547a9b3
YS		 2119 tid = le64_to_cpu(hdr->tid);
2120 mutex_lock(&osdc->request_mutex);
2121 req = __lookup_request(osdc, tid);
2122 if (!req) {
2123 *skip = 1;
2124 m = NULL;
756a16a5
SW		 2125 dout("get_reply unknown tid %llu from osd%d\n", tid,
2126 osd->o_osd);
0547a9b3
YS		 2127 goto out;
2128 }
c16e7869
SW		 2129
2130 if (req->r_con_filling_msg) {
8921d114 2131 dout("%s revoking msg %p from old con %p\n", __func__,
c16e7869 2132 req->r_reply, req->r_con_filling_msg);
8921d114 2133 ceph_msg_revoke_incoming(req->r_reply);
0d47766f 2134 req->r_con_filling_msg->ops->put(req->r_con_filling_msg);
6f46cb29 2135 req->r_con_filling_msg = NULL;
0547a9b3
YS		 2136 }
2137
c16e7869
SW		 2138 if (front > req->r_reply->front.iov_len) {
2139 pr_warning("get_reply front %d > preallocated %d\n",
2140 front, (int)req->r_reply->front.iov_len);
b61c2763 2141 m = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, front, GFP_NOFS, false);
a79832f2 2142 if (!m)
c16e7869
SW		 2143 goto out;
2144 ceph_msg_put(req->r_reply);
2145 req->r_reply = m;
2146 }
2147 m = ceph_msg_get(req->r_reply);
2148
0547a9b3 2149 if (data_len > 0) {
0fff87ec
AE		 2150 struct ceph_osd_data *osd_data = &req->r_data_in;
2151
2152 if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGES) {
0fff87ec 2153 if (osd_data->pages &&
e0c59487 2154 unlikely(osd_data->length < data_len)) {
2ac2b7a6 2155
e0c59487
AE		 2156 pr_warning("tid %lld reply has %d bytes "
2157 "we had only %llu bytes ready\n",
2158 tid, data_len, osd_data->length);
2ac2b7a6
AE		 2159 *skip = 1;
2160 ceph_msg_put(m);
2161 m = NULL;
2162 goto out;
2163 }
2ac2b7a6 2164 }
0547a9b3 2165 }
5b3a4db3 2166 *skip = 0;
0d47766f 2167 req->r_con_filling_msg = con->ops->get(con);
c16e7869 2168 dout("get_reply tid %lld %p\n", tid, m);
0547a9b3
YS		 2169
2170out:
2171 mutex_unlock(&osdc->request_mutex);
2450418c 2172 return m;
5b3a4db3
SW		 2173
2174}
2175
2176static struct ceph_msg *alloc_msg(struct ceph_connection *con,
2177 struct ceph_msg_header *hdr,
2178 int *skip)
2179{
2180 struct ceph_osd *osd = con->private;
2181 int type = le16_to_cpu(hdr->type);
2182 int front = le32_to_cpu(hdr->front_len);
2183
1c20f2d2 2184 *skip = 0;
5b3a4db3
SW		 2185 switch (type) {
2186 case CEPH_MSG_OSD_MAP:
a40c4f10 2187 case CEPH_MSG_WATCH_NOTIFY:
b61c2763 2188 return ceph_msg_new(type, front, GFP_NOFS, false);
5b3a4db3
SW		 2189 case CEPH_MSG_OSD_OPREPLY:
2190 return get_reply(con, hdr, skip);
2191 default:
2192 pr_info("alloc_msg unexpected msg type %d from osd%d\n", type,
2193 osd->o_osd);
2194 *skip = 1;
2195 return NULL;
2196 }
f24e9980
SW		 2197}
2198
2199/*
2200 * Wrappers to refcount containing ceph_osd struct
2201 */
2202static struct ceph_connection *get_osd_con(struct ceph_connection *con)
2203{
2204 struct ceph_osd *osd = con->private;
2205 if (get_osd(osd))
2206 return con;
2207 return NULL;
2208}
2209
2210static void put_osd_con(struct ceph_connection *con)
2211{
2212 struct ceph_osd *osd = con->private;
2213 put_osd(osd);
2214}
2215
4e7a5dcd
SW		 2216/*
2217 * authentication
2218 */
a3530df3
AE		 2219/*
2220 * Note: returned pointer is the address of a structure that's
2221 * managed separately. Caller must *not* attempt to free it.
2222 */
2223static struct ceph_auth_handshake *get_authorizer(struct ceph_connection *con,
8f43fb53 2224 int *proto, int force_new)
4e7a5dcd
SW		 2225{
2226 struct ceph_osd *o = con->private;
2227 struct ceph_osd_client *osdc = o->o_osdc;
2228 struct ceph_auth_client *ac = osdc->client->monc.auth;
74f1869f 2229 struct ceph_auth_handshake *auth = &o->o_auth;
4e7a5dcd 2230
74f1869f 2231 if (force_new && auth->authorizer) {
27859f97 2232 ceph_auth_destroy_authorizer(ac, auth->authorizer);
74f1869f
AE		 2233 auth->authorizer = NULL;
2234 }
27859f97
SW		 2235 if (!auth->authorizer) {
2236 int ret = ceph_auth_create_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
2237 auth);
4e7a5dcd 2238 if (ret)
a3530df3 2239 return ERR_PTR(ret);
27859f97
SW		 2240 } else {
2241 int ret = ceph_auth_update_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
0bed9b5c
SW		 2242 auth);
2243 if (ret)
2244 return ERR_PTR(ret);
4e7a5dcd 2245 }
4e7a5dcd 2246 *proto = ac->protocol;
74f1869f 2247
a3530df3 2248 return auth;
4e7a5dcd
SW		 2249}
2250
2251
2252static int verify_authorizer_reply(struct ceph_connection *con, int len)
2253{
2254 struct ceph_osd *o = con->private;
2255 struct ceph_osd_client *osdc = o->o_osdc;
2256 struct ceph_auth_client *ac = osdc->client->monc.auth;
2257
27859f97 2258 return ceph_auth_verify_authorizer_reply(ac, o->o_auth.authorizer, len);
4e7a5dcd
SW		 2259}
2260
9bd2e6f8
SW		 2261static int invalidate_authorizer(struct ceph_connection *con)
2262{
2263 struct ceph_osd *o = con->private;
2264 struct ceph_osd_client *osdc = o->o_osdc;
2265 struct ceph_auth_client *ac = osdc->client->monc.auth;
2266
27859f97 2267 ceph_auth_invalidate_authorizer(ac, CEPH_ENTITY_TYPE_OSD);
9bd2e6f8
SW		 2268 return ceph_monc_validate_auth(&osdc->client->monc);
2269}
4e7a5dcd 2270
9e32789f 2271static const struct ceph_connection_operations osd_con_ops = {
f24e9980
SW		 2272 .get = get_osd_con,
2273 .put = put_osd_con,
2274 .dispatch = dispatch,
4e7a5dcd
SW		 2275 .get_authorizer = get_authorizer,
2276 .verify_authorizer_reply = verify_authorizer_reply,
9bd2e6f8 2277 .invalidate_authorizer = invalidate_authorizer,
f24e9980 2278 .alloc_msg = alloc_msg,
81b024e7 2279 .fault = osd_reset,
f24e9980 2280};
Linux 6.x block layer and io_uring tree(s)