Merge tag 'for-upstream' of https://gitlab.com/bonzini/qemu into staging
[qemu.git] / nbd / server.c
1 /*
2 * Copyright (C) 2016-2021 Red Hat, Inc.
3 * Copyright (C) 2005 Anthony Liguori <anthony@codemonkey.ws>
4 *
5 * Network Block Device Server Side
6 *
7 * This program is free software; you can redistribute it and/or modify
8 * it under the terms of the GNU General Public License as published by
9 * the Free Software Foundation; under version 2 of the License.
10 *
11 * This program is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 * GNU General Public License for more details.
15 *
16 * You should have received a copy of the GNU General Public License
17 * along with this program; if not, see <http://www.gnu.org/licenses/>.
18 */
19
20 #include "qemu/osdep.h"
21
22 #include "block/export.h"
23 #include "qapi/error.h"
24 #include "qemu/queue.h"
25 #include "trace.h"
26 #include "nbd-internal.h"
27 #include "qemu/units.h"
28
29 #define NBD_META_ID_BASE_ALLOCATION 0
30 #define NBD_META_ID_ALLOCATION_DEPTH 1
31 /* Dirty bitmaps use 'NBD_META_ID_DIRTY_BITMAP + i', so keep this id last. */
32 #define NBD_META_ID_DIRTY_BITMAP 2
33
34 /*
35 * NBD_MAX_BLOCK_STATUS_EXTENTS: 1 MiB of extents data. An empirical
36 * constant. If an increase is needed, note that the NBD protocol
37 * recommends no larger than 32 mb, so that the client won't consider
38 * the reply as a denial of service attack.
39 */
40 #define NBD_MAX_BLOCK_STATUS_EXTENTS (1 * MiB / 8)
41
42 static int system_errno_to_nbd_errno(int err)
43 {
44 switch (err) {
45 case 0:
46 return NBD_SUCCESS;
47 case EPERM:
48 case EROFS:
49 return NBD_EPERM;
50 case EIO:
51 return NBD_EIO;
52 case ENOMEM:
53 return NBD_ENOMEM;
54 #ifdef EDQUOT
55 case EDQUOT:
56 #endif
57 case EFBIG:
58 case ENOSPC:
59 return NBD_ENOSPC;
60 case EOVERFLOW:
61 return NBD_EOVERFLOW;
62 case ENOTSUP:
63 #if ENOTSUP != EOPNOTSUPP
64 case EOPNOTSUPP:
65 #endif
66 return NBD_ENOTSUP;
67 case ESHUTDOWN:
68 return NBD_ESHUTDOWN;
69 case EINVAL:
70 default:
71 return NBD_EINVAL;
72 }
73 }
74
75 /* Definitions for opaque data types */
76
77 typedef struct NBDRequestData NBDRequestData;
78
79 struct NBDRequestData {
80 QSIMPLEQ_ENTRY(NBDRequestData) entry;
81 NBDClient *client;
82 uint8_t *data;
83 bool complete;
84 };
85
86 struct NBDExport {
87 BlockExport common;
88
89 char *name;
90 char *description;
91 uint64_t size;
92 uint16_t nbdflags;
93 QTAILQ_HEAD(, NBDClient) clients;
94 QTAILQ_ENTRY(NBDExport) next;
95
96 BlockBackend *eject_notifier_blk;
97 Notifier eject_notifier;
98
99 bool allocation_depth;
100 BdrvDirtyBitmap **export_bitmaps;
101 size_t nr_export_bitmaps;
102 };
103
104 static QTAILQ_HEAD(, NBDExport) exports = QTAILQ_HEAD_INITIALIZER(exports);
105
106 /* NBDExportMetaContexts represents a list of contexts to be exported,
107 * as selected by NBD_OPT_SET_META_CONTEXT. Also used for
108 * NBD_OPT_LIST_META_CONTEXT. */
109 typedef struct NBDExportMetaContexts {
110 NBDExport *exp;
111 size_t count; /* number of negotiated contexts */
112 bool base_allocation; /* export base:allocation context (block status) */
113 bool allocation_depth; /* export qemu:allocation-depth */
114 bool *bitmaps; /*
115 * export qemu:dirty-bitmap:<export bitmap name>,
116 * sized by exp->nr_export_bitmaps
117 */
118 } NBDExportMetaContexts;
119
120 struct NBDClient {
121 int refcount;
122 void (*close_fn)(NBDClient *client, bool negotiated);
123
124 NBDExport *exp;
125 QCryptoTLSCreds *tlscreds;
126 char *tlsauthz;
127 QIOChannelSocket *sioc; /* The underlying data channel */
128 QIOChannel *ioc; /* The current I/O channel which may differ (eg TLS) */
129
130 Coroutine *recv_coroutine;
131
132 CoMutex send_lock;
133 Coroutine *send_coroutine;
134
135 bool read_yielding;
136 bool quiescing;
137
138 QTAILQ_ENTRY(NBDClient) next;
139 int nb_requests;
140 bool closing;
141
142 uint32_t check_align; /* If non-zero, check for aligned client requests */
143
144 bool structured_reply;
145 NBDExportMetaContexts export_meta;
146
147 uint32_t opt; /* Current option being negotiated */
148 uint32_t optlen; /* remaining length of data in ioc for the option being
149 negotiated now */
150 };
151
152 static void nbd_client_receive_next_request(NBDClient *client);
153
154 /* Basic flow for negotiation
155
156 Server Client
157 Negotiate
158
159 or
160
161 Server Client
162 Negotiate #1
163 Option
164 Negotiate #2
165
166 ----
167
168 followed by
169
170 Server Client
171 Request
172 Response
173 Request
174 Response
175 ...
176 ...
177 Request (type == 2)
178
179 */
180
181 static inline void set_be_option_rep(NBDOptionReply *rep, uint32_t option,
182 uint32_t type, uint32_t length)
183 {
184 stq_be_p(&rep->magic, NBD_REP_MAGIC);
185 stl_be_p(&rep->option, option);
186 stl_be_p(&rep->type, type);
187 stl_be_p(&rep->length, length);
188 }
189
190 /* Send a reply header, including length, but no payload.
191 * Return -errno on error, 0 on success. */
192 static int nbd_negotiate_send_rep_len(NBDClient *client, uint32_t type,
193 uint32_t len, Error **errp)
194 {
195 NBDOptionReply rep;
196
197 trace_nbd_negotiate_send_rep_len(client->opt, nbd_opt_lookup(client->opt),
198 type, nbd_rep_lookup(type), len);
199
200 assert(len < NBD_MAX_BUFFER_SIZE);
201
202 set_be_option_rep(&rep, client->opt, type, len);
203 return nbd_write(client->ioc, &rep, sizeof(rep), errp);
204 }
205
206 /* Send a reply header with default 0 length.
207 * Return -errno on error, 0 on success. */
208 static int nbd_negotiate_send_rep(NBDClient *client, uint32_t type,
209 Error **errp)
210 {
211 return nbd_negotiate_send_rep_len(client, type, 0, errp);
212 }
213
214 /* Send an error reply.
215 * Return -errno on error, 0 on success. */
216 static int GCC_FMT_ATTR(4, 0)
217 nbd_negotiate_send_rep_verr(NBDClient *client, uint32_t type,
218 Error **errp, const char *fmt, va_list va)
219 {
220 ERRP_GUARD();
221 g_autofree char *msg = NULL;
222 int ret;
223 size_t len;
224
225 msg = g_strdup_vprintf(fmt, va);
226 len = strlen(msg);
227 assert(len < NBD_MAX_STRING_SIZE);
228 trace_nbd_negotiate_send_rep_err(msg);
229 ret = nbd_negotiate_send_rep_len(client, type, len, errp);
230 if (ret < 0) {
231 return ret;
232 }
233 if (nbd_write(client->ioc, msg, len, errp) < 0) {
234 error_prepend(errp, "write failed (error message): ");
235 return -EIO;
236 }
237
238 return 0;
239 }
240
241 /*
242 * Return a malloc'd copy of @name suitable for use in an error reply.
243 */
244 static char *
245 nbd_sanitize_name(const char *name)
246 {
247 if (strnlen(name, 80) < 80) {
248 return g_strdup(name);
249 }
250 /* XXX Should we also try to sanitize any control characters? */
251 return g_strdup_printf("%.80s...", name);
252 }
253
254 /* Send an error reply.
255 * Return -errno on error, 0 on success. */
256 static int GCC_FMT_ATTR(4, 5)
257 nbd_negotiate_send_rep_err(NBDClient *client, uint32_t type,
258 Error **errp, const char *fmt, ...)
259 {
260 va_list va;
261 int ret;
262
263 va_start(va, fmt);
264 ret = nbd_negotiate_send_rep_verr(client, type, errp, fmt, va);
265 va_end(va);
266 return ret;
267 }
268
269 /* Drop remainder of the current option, and send a reply with the
270 * given error type and message. Return -errno on read or write
271 * failure; or 0 if connection is still live. */
272 static int GCC_FMT_ATTR(4, 0)
273 nbd_opt_vdrop(NBDClient *client, uint32_t type, Error **errp,
274 const char *fmt, va_list va)
275 {
276 int ret = nbd_drop(client->ioc, client->optlen, errp);
277
278 client->optlen = 0;
279 if (!ret) {
280 ret = nbd_negotiate_send_rep_verr(client, type, errp, fmt, va);
281 }
282 return ret;
283 }
284
285 static int GCC_FMT_ATTR(4, 5)
286 nbd_opt_drop(NBDClient *client, uint32_t type, Error **errp,
287 const char *fmt, ...)
288 {
289 int ret;
290 va_list va;
291
292 va_start(va, fmt);
293 ret = nbd_opt_vdrop(client, type, errp, fmt, va);
294 va_end(va);
295
296 return ret;
297 }
298
299 static int GCC_FMT_ATTR(3, 4)
300 nbd_opt_invalid(NBDClient *client, Error **errp, const char *fmt, ...)
301 {
302 int ret;
303 va_list va;
304
305 va_start(va, fmt);
306 ret = nbd_opt_vdrop(client, NBD_REP_ERR_INVALID, errp, fmt, va);
307 va_end(va);
308
309 return ret;
310 }
311
312 /* Read size bytes from the unparsed payload of the current option.
313 * If @check_nul, require that no NUL bytes appear in buffer.
314 * Return -errno on I/O error, 0 if option was completely handled by
315 * sending a reply about inconsistent lengths, or 1 on success. */
316 static int nbd_opt_read(NBDClient *client, void *buffer, size_t size,
317 bool check_nul, Error **errp)
318 {
319 if (size > client->optlen) {
320 return nbd_opt_invalid(client, errp,
321 "Inconsistent lengths in option %s",
322 nbd_opt_lookup(client->opt));
323 }
324 client->optlen -= size;
325 if (qio_channel_read_all(client->ioc, buffer, size, errp) < 0) {
326 return -EIO;
327 }
328
329 if (check_nul && strnlen(buffer, size) != size) {
330 return nbd_opt_invalid(client, errp,
331 "Unexpected embedded NUL in option %s",
332 nbd_opt_lookup(client->opt));
333 }
334 return 1;
335 }
336
337 /* Drop size bytes from the unparsed payload of the current option.
338 * Return -errno on I/O error, 0 if option was completely handled by
339 * sending a reply about inconsistent lengths, or 1 on success. */
340 static int nbd_opt_skip(NBDClient *client, size_t size, Error **errp)
341 {
342 if (size > client->optlen) {
343 return nbd_opt_invalid(client, errp,
344 "Inconsistent lengths in option %s",
345 nbd_opt_lookup(client->opt));
346 }
347 client->optlen -= size;
348 return nbd_drop(client->ioc, size, errp) < 0 ? -EIO : 1;
349 }
350
351 /* nbd_opt_read_name
352 *
353 * Read a string with the format:
354 * uint32_t len (<= NBD_MAX_STRING_SIZE)
355 * len bytes string (not 0-terminated)
356 *
357 * On success, @name will be allocated.
358 * If @length is non-null, it will be set to the actual string length.
359 *
360 * Return -errno on I/O error, 0 if option was completely handled by
361 * sending a reply about inconsistent lengths, or 1 on success.
362 */
363 static int nbd_opt_read_name(NBDClient *client, char **name, uint32_t *length,
364 Error **errp)
365 {
366 int ret;
367 uint32_t len;
368 g_autofree char *local_name = NULL;
369
370 *name = NULL;
371 ret = nbd_opt_read(client, &len, sizeof(len), false, errp);
372 if (ret <= 0) {
373 return ret;
374 }
375 len = cpu_to_be32(len);
376
377 if (len > NBD_MAX_STRING_SIZE) {
378 return nbd_opt_invalid(client, errp,
379 "Invalid name length: %" PRIu32, len);
380 }
381
382 local_name = g_malloc(len + 1);
383 ret = nbd_opt_read(client, local_name, len, true, errp);
384 if (ret <= 0) {
385 return ret;
386 }
387 local_name[len] = '\0';
388
389 if (length) {
390 *length = len;
391 }
392 *name = g_steal_pointer(&local_name);
393
394 return 1;
395 }
396
397 /* Send a single NBD_REP_SERVER reply to NBD_OPT_LIST, including payload.
398 * Return -errno on error, 0 on success. */
399 static int nbd_negotiate_send_rep_list(NBDClient *client, NBDExport *exp,
400 Error **errp)
401 {
402 ERRP_GUARD();
403 size_t name_len, desc_len;
404 uint32_t len;
405 const char *name = exp->name ? exp->name : "";
406 const char *desc = exp->description ? exp->description : "";
407 QIOChannel *ioc = client->ioc;
408 int ret;
409
410 trace_nbd_negotiate_send_rep_list(name, desc);
411 name_len = strlen(name);
412 desc_len = strlen(desc);
413 assert(name_len <= NBD_MAX_STRING_SIZE && desc_len <= NBD_MAX_STRING_SIZE);
414 len = name_len + desc_len + sizeof(len);
415 ret = nbd_negotiate_send_rep_len(client, NBD_REP_SERVER, len, errp);
416 if (ret < 0) {
417 return ret;
418 }
419
420 len = cpu_to_be32(name_len);
421 if (nbd_write(ioc, &len, sizeof(len), errp) < 0) {
422 error_prepend(errp, "write failed (name length): ");
423 return -EINVAL;
424 }
425
426 if (nbd_write(ioc, name, name_len, errp) < 0) {
427 error_prepend(errp, "write failed (name buffer): ");
428 return -EINVAL;
429 }
430
431 if (nbd_write(ioc, desc, desc_len, errp) < 0) {
432 error_prepend(errp, "write failed (description buffer): ");
433 return -EINVAL;
434 }
435
436 return 0;
437 }
438
439 /* Process the NBD_OPT_LIST command, with a potential series of replies.
440 * Return -errno on error, 0 on success. */
441 static int nbd_negotiate_handle_list(NBDClient *client, Error **errp)
442 {
443 NBDExport *exp;
444 assert(client->opt == NBD_OPT_LIST);
445
446 /* For each export, send a NBD_REP_SERVER reply. */
447 QTAILQ_FOREACH(exp, &exports, next) {
448 if (nbd_negotiate_send_rep_list(client, exp, errp)) {
449 return -EINVAL;
450 }
451 }
452 /* Finish with a NBD_REP_ACK. */
453 return nbd_negotiate_send_rep(client, NBD_REP_ACK, errp);
454 }
455
456 static void nbd_check_meta_export(NBDClient *client)
457 {
458 if (client->exp != client->export_meta.exp) {
459 client->export_meta.count = 0;
460 }
461 }
462
463 /* Send a reply to NBD_OPT_EXPORT_NAME.
464 * Return -errno on error, 0 on success. */
465 static int nbd_negotiate_handle_export_name(NBDClient *client, bool no_zeroes,
466 Error **errp)
467 {
468 ERRP_GUARD();
469 g_autofree char *name = NULL;
470 char buf[NBD_REPLY_EXPORT_NAME_SIZE] = "";
471 size_t len;
472 int ret;
473 uint16_t myflags;
474
475 /* Client sends:
476 [20 .. xx] export name (length bytes)
477 Server replies:
478 [ 0 .. 7] size
479 [ 8 .. 9] export flags
480 [10 .. 133] reserved (0) [unless no_zeroes]
481 */
482 trace_nbd_negotiate_handle_export_name();
483 if (client->optlen > NBD_MAX_STRING_SIZE) {
484 error_setg(errp, "Bad length received");
485 return -EINVAL;
486 }
487 name = g_malloc(client->optlen + 1);
488 if (nbd_read(client->ioc, name, client->optlen, "export name", errp) < 0) {
489 return -EIO;
490 }
491 name[client->optlen] = '\0';
492 client->optlen = 0;
493
494 trace_nbd_negotiate_handle_export_name_request(name);
495
496 client->exp = nbd_export_find(name);
497 if (!client->exp) {
498 error_setg(errp, "export not found");
499 return -EINVAL;
500 }
501
502 myflags = client->exp->nbdflags;
503 if (client->structured_reply) {
504 myflags |= NBD_FLAG_SEND_DF;
505 }
506 trace_nbd_negotiate_new_style_size_flags(client->exp->size, myflags);
507 stq_be_p(buf, client->exp->size);
508 stw_be_p(buf + 8, myflags);
509 len = no_zeroes ? 10 : sizeof(buf);
510 ret = nbd_write(client->ioc, buf, len, errp);
511 if (ret < 0) {
512 error_prepend(errp, "write failed: ");
513 return ret;
514 }
515
516 QTAILQ_INSERT_TAIL(&client->exp->clients, client, next);
517 blk_exp_ref(&client->exp->common);
518 nbd_check_meta_export(client);
519
520 return 0;
521 }
522
523 /* Send a single NBD_REP_INFO, with a buffer @buf of @length bytes.
524 * The buffer does NOT include the info type prefix.
525 * Return -errno on error, 0 if ready to send more. */
526 static int nbd_negotiate_send_info(NBDClient *client,
527 uint16_t info, uint32_t length, void *buf,
528 Error **errp)
529 {
530 int rc;
531
532 trace_nbd_negotiate_send_info(info, nbd_info_lookup(info), length);
533 rc = nbd_negotiate_send_rep_len(client, NBD_REP_INFO,
534 sizeof(info) + length, errp);
535 if (rc < 0) {
536 return rc;
537 }
538 info = cpu_to_be16(info);
539 if (nbd_write(client->ioc, &info, sizeof(info), errp) < 0) {
540 return -EIO;
541 }
542 if (nbd_write(client->ioc, buf, length, errp) < 0) {
543 return -EIO;
544 }
545 return 0;
546 }
547
548 /* nbd_reject_length: Handle any unexpected payload.
549 * @fatal requests that we quit talking to the client, even if we are able
550 * to successfully send an error reply.
551 * Return:
552 * -errno transmission error occurred or @fatal was requested, errp is set
553 * 0 error message successfully sent to client, errp is not set
554 */
555 static int nbd_reject_length(NBDClient *client, bool fatal, Error **errp)
556 {
557 int ret;
558
559 assert(client->optlen);
560 ret = nbd_opt_invalid(client, errp, "option '%s' has unexpected length",
561 nbd_opt_lookup(client->opt));
562 if (fatal && !ret) {
563 error_setg(errp, "option '%s' has unexpected length",
564 nbd_opt_lookup(client->opt));
565 return -EINVAL;
566 }
567 return ret;
568 }
569
570 /* Handle NBD_OPT_INFO and NBD_OPT_GO.
571 * Return -errno on error, 0 if ready for next option, and 1 to move
572 * into transmission phase. */
573 static int nbd_negotiate_handle_info(NBDClient *client, Error **errp)
574 {
575 int rc;
576 g_autofree char *name = NULL;
577 NBDExport *exp;
578 uint16_t requests;
579 uint16_t request;
580 uint32_t namelen = 0;
581 bool sendname = false;
582 bool blocksize = false;
583 uint32_t sizes[3];
584 char buf[sizeof(uint64_t) + sizeof(uint16_t)];
585 uint32_t check_align = 0;
586 uint16_t myflags;
587
588 /* Client sends:
589 4 bytes: L, name length (can be 0)
590 L bytes: export name
591 2 bytes: N, number of requests (can be 0)
592 N * 2 bytes: N requests
593 */
594 rc = nbd_opt_read_name(client, &name, &namelen, errp);
595 if (rc <= 0) {
596 return rc;
597 }
598 trace_nbd_negotiate_handle_export_name_request(name);
599
600 rc = nbd_opt_read(client, &requests, sizeof(requests), false, errp);
601 if (rc <= 0) {
602 return rc;
603 }
604 requests = be16_to_cpu(requests);
605 trace_nbd_negotiate_handle_info_requests(requests);
606 while (requests--) {
607 rc = nbd_opt_read(client, &request, sizeof(request), false, errp);
608 if (rc <= 0) {
609 return rc;
610 }
611 request = be16_to_cpu(request);
612 trace_nbd_negotiate_handle_info_request(request,
613 nbd_info_lookup(request));
614 /* We care about NBD_INFO_NAME and NBD_INFO_BLOCK_SIZE;
615 * everything else is either a request we don't know or
616 * something we send regardless of request */
617 switch (request) {
618 case NBD_INFO_NAME:
619 sendname = true;
620 break;
621 case NBD_INFO_BLOCK_SIZE:
622 blocksize = true;
623 break;
624 }
625 }
626 if (client->optlen) {
627 return nbd_reject_length(client, false, errp);
628 }
629
630 exp = nbd_export_find(name);
631 if (!exp) {
632 g_autofree char *sane_name = nbd_sanitize_name(name);
633
634 return nbd_negotiate_send_rep_err(client, NBD_REP_ERR_UNKNOWN,
635 errp, "export '%s' not present",
636 sane_name);
637 }
638
639 /* Don't bother sending NBD_INFO_NAME unless client requested it */
640 if (sendname) {
641 rc = nbd_negotiate_send_info(client, NBD_INFO_NAME, namelen, name,
642 errp);
643 if (rc < 0) {
644 return rc;
645 }
646 }
647
648 /* Send NBD_INFO_DESCRIPTION only if available, regardless of
649 * client request */
650 if (exp->description) {
651 size_t len = strlen(exp->description);
652
653 assert(len <= NBD_MAX_STRING_SIZE);
654 rc = nbd_negotiate_send_info(client, NBD_INFO_DESCRIPTION,
655 len, exp->description, errp);
656 if (rc < 0) {
657 return rc;
658 }
659 }
660
661 /* Send NBD_INFO_BLOCK_SIZE always, but tweak the minimum size
662 * according to whether the client requested it, and according to
663 * whether this is OPT_INFO or OPT_GO. */
664 /* minimum - 1 for back-compat, or actual if client will obey it. */
665 if (client->opt == NBD_OPT_INFO || blocksize) {
666 check_align = sizes[0] = blk_get_request_alignment(exp->common.blk);
667 } else {
668 sizes[0] = 1;
669 }
670 assert(sizes[0] <= NBD_MAX_BUFFER_SIZE);
671 /* preferred - Hard-code to 4096 for now.
672 * TODO: is blk_bs(blk)->bl.opt_transfer appropriate? */
673 sizes[1] = MAX(4096, sizes[0]);
674 /* maximum - At most 32M, but smaller as appropriate. */
675 sizes[2] = MIN(blk_get_max_transfer(exp->common.blk), NBD_MAX_BUFFER_SIZE);
676 trace_nbd_negotiate_handle_info_block_size(sizes[0], sizes[1], sizes[2]);
677 sizes[0] = cpu_to_be32(sizes[0]);
678 sizes[1] = cpu_to_be32(sizes[1]);
679 sizes[2] = cpu_to_be32(sizes[2]);
680 rc = nbd_negotiate_send_info(client, NBD_INFO_BLOCK_SIZE,
681 sizeof(sizes), sizes, errp);
682 if (rc < 0) {
683 return rc;
684 }
685
686 /* Send NBD_INFO_EXPORT always */
687 myflags = exp->nbdflags;
688 if (client->structured_reply) {
689 myflags |= NBD_FLAG_SEND_DF;
690 }
691 trace_nbd_negotiate_new_style_size_flags(exp->size, myflags);
692 stq_be_p(buf, exp->size);
693 stw_be_p(buf + 8, myflags);
694 rc = nbd_negotiate_send_info(client, NBD_INFO_EXPORT,
695 sizeof(buf), buf, errp);
696 if (rc < 0) {
697 return rc;
698 }
699
700 /*
701 * If the client is just asking for NBD_OPT_INFO, but forgot to
702 * request block sizes in a situation that would impact
703 * performance, then return an error. But for NBD_OPT_GO, we
704 * tolerate all clients, regardless of alignments.
705 */
706 if (client->opt == NBD_OPT_INFO && !blocksize &&
707 blk_get_request_alignment(exp->common.blk) > 1) {
708 return nbd_negotiate_send_rep_err(client,
709 NBD_REP_ERR_BLOCK_SIZE_REQD,
710 errp,
711 "request NBD_INFO_BLOCK_SIZE to "
712 "use this export");
713 }
714
715 /* Final reply */
716 rc = nbd_negotiate_send_rep(client, NBD_REP_ACK, errp);
717 if (rc < 0) {
718 return rc;
719 }
720
721 if (client->opt == NBD_OPT_GO) {
722 client->exp = exp;
723 client->check_align = check_align;
724 QTAILQ_INSERT_TAIL(&client->exp->clients, client, next);
725 blk_exp_ref(&client->exp->common);
726 nbd_check_meta_export(client);
727 rc = 1;
728 }
729 return rc;
730 }
731
732
733 /* Handle NBD_OPT_STARTTLS. Return NULL to drop connection, or else the
734 * new channel for all further (now-encrypted) communication. */
735 static QIOChannel *nbd_negotiate_handle_starttls(NBDClient *client,
736 Error **errp)
737 {
738 QIOChannel *ioc;
739 QIOChannelTLS *tioc;
740 struct NBDTLSHandshakeData data = { 0 };
741
742 assert(client->opt == NBD_OPT_STARTTLS);
743
744 trace_nbd_negotiate_handle_starttls();
745 ioc = client->ioc;
746
747 if (nbd_negotiate_send_rep(client, NBD_REP_ACK, errp) < 0) {
748 return NULL;
749 }
750
751 tioc = qio_channel_tls_new_server(ioc,
752 client->tlscreds,
753 client->tlsauthz,
754 errp);
755 if (!tioc) {
756 return NULL;
757 }
758
759 qio_channel_set_name(QIO_CHANNEL(tioc), "nbd-server-tls");
760 trace_nbd_negotiate_handle_starttls_handshake();
761 data.loop = g_main_loop_new(g_main_context_default(), FALSE);
762 qio_channel_tls_handshake(tioc,
763 nbd_tls_handshake,
764 &data,
765 NULL,
766 NULL);
767
768 if (!data.complete) {
769 g_main_loop_run(data.loop);
770 }
771 g_main_loop_unref(data.loop);
772 if (data.error) {
773 object_unref(OBJECT(tioc));
774 error_propagate(errp, data.error);
775 return NULL;
776 }
777
778 return QIO_CHANNEL(tioc);
779 }
780
781 /* nbd_negotiate_send_meta_context
782 *
783 * Send one chunk of reply to NBD_OPT_{LIST,SET}_META_CONTEXT
784 *
785 * For NBD_OPT_LIST_META_CONTEXT @context_id is ignored, 0 is used instead.
786 */
787 static int nbd_negotiate_send_meta_context(NBDClient *client,
788 const char *context,
789 uint32_t context_id,
790 Error **errp)
791 {
792 NBDOptionReplyMetaContext opt;
793 struct iovec iov[] = {
794 {.iov_base = &opt, .iov_len = sizeof(opt)},
795 {.iov_base = (void *)context, .iov_len = strlen(context)}
796 };
797
798 assert(iov[1].iov_len <= NBD_MAX_STRING_SIZE);
799 if (client->opt == NBD_OPT_LIST_META_CONTEXT) {
800 context_id = 0;
801 }
802
803 trace_nbd_negotiate_meta_query_reply(context, context_id);
804 set_be_option_rep(&opt.h, client->opt, NBD_REP_META_CONTEXT,
805 sizeof(opt) - sizeof(opt.h) + iov[1].iov_len);
806 stl_be_p(&opt.context_id, context_id);
807
808 return qio_channel_writev_all(client->ioc, iov, 2, errp) < 0 ? -EIO : 0;
809 }
810
811 /*
812 * Return true if @query matches @pattern, or if @query is empty when
813 * the @client is performing _LIST_.
814 */
815 static bool nbd_meta_empty_or_pattern(NBDClient *client, const char *pattern,
816 const char *query)
817 {
818 if (!*query) {
819 trace_nbd_negotiate_meta_query_parse("empty");
820 return client->opt == NBD_OPT_LIST_META_CONTEXT;
821 }
822 if (strcmp(query, pattern) == 0) {
823 trace_nbd_negotiate_meta_query_parse(pattern);
824 return true;
825 }
826 trace_nbd_negotiate_meta_query_skip("pattern not matched");
827 return false;
828 }
829
830 /*
831 * Return true and adjust @str in place if it begins with @prefix.
832 */
833 static bool nbd_strshift(const char **str, const char *prefix)
834 {
835 size_t len = strlen(prefix);
836
837 if (strncmp(*str, prefix, len) == 0) {
838 *str += len;
839 return true;
840 }
841 return false;
842 }
843
844 /* nbd_meta_base_query
845 *
846 * Handle queries to 'base' namespace. For now, only the base:allocation
847 * context is available. Return true if @query has been handled.
848 */
849 static bool nbd_meta_base_query(NBDClient *client, NBDExportMetaContexts *meta,
850 const char *query)
851 {
852 if (!nbd_strshift(&query, "base:")) {
853 return false;
854 }
855 trace_nbd_negotiate_meta_query_parse("base:");
856
857 if (nbd_meta_empty_or_pattern(client, "allocation", query)) {
858 meta->base_allocation = true;
859 }
860 return true;
861 }
862
863 /* nbd_meta_qemu_query
864 *
865 * Handle queries to 'qemu' namespace. For now, only the qemu:dirty-bitmap:
866 * and qemu:allocation-depth contexts are available. Return true if @query
867 * has been handled.
868 */
869 static bool nbd_meta_qemu_query(NBDClient *client, NBDExportMetaContexts *meta,
870 const char *query)
871 {
872 size_t i;
873
874 if (!nbd_strshift(&query, "qemu:")) {
875 return false;
876 }
877 trace_nbd_negotiate_meta_query_parse("qemu:");
878
879 if (!*query) {
880 if (client->opt == NBD_OPT_LIST_META_CONTEXT) {
881 meta->allocation_depth = meta->exp->allocation_depth;
882 if (meta->exp->nr_export_bitmaps) {
883 memset(meta->bitmaps, 1, meta->exp->nr_export_bitmaps);
884 }
885 }
886 trace_nbd_negotiate_meta_query_parse("empty");
887 return true;
888 }
889
890 if (strcmp(query, "allocation-depth") == 0) {
891 trace_nbd_negotiate_meta_query_parse("allocation-depth");
892 meta->allocation_depth = meta->exp->allocation_depth;
893 return true;
894 }
895
896 if (nbd_strshift(&query, "dirty-bitmap:")) {
897 trace_nbd_negotiate_meta_query_parse("dirty-bitmap:");
898 if (!*query) {
899 if (client->opt == NBD_OPT_LIST_META_CONTEXT &&
900 meta->exp->nr_export_bitmaps) {
901 memset(meta->bitmaps, 1, meta->exp->nr_export_bitmaps);
902 }
903 trace_nbd_negotiate_meta_query_parse("empty");
904 return true;
905 }
906
907 for (i = 0; i < meta->exp->nr_export_bitmaps; i++) {
908 const char *bm_name;
909
910 bm_name = bdrv_dirty_bitmap_name(meta->exp->export_bitmaps[i]);
911 if (strcmp(bm_name, query) == 0) {
912 meta->bitmaps[i] = true;
913 trace_nbd_negotiate_meta_query_parse(query);
914 return true;
915 }
916 }
917 trace_nbd_negotiate_meta_query_skip("no dirty-bitmap match");
918 return true;
919 }
920
921 trace_nbd_negotiate_meta_query_skip("unknown qemu context");
922 return true;
923 }
924
925 /* nbd_negotiate_meta_query
926 *
927 * Parse namespace name and call corresponding function to parse body of the
928 * query.
929 *
930 * The only supported namespaces are 'base' and 'qemu'.
931 *
932 * Return -errno on I/O error, 0 if option was completely handled by
933 * sending a reply about inconsistent lengths, or 1 on success. */
934 static int nbd_negotiate_meta_query(NBDClient *client,
935 NBDExportMetaContexts *meta, Error **errp)
936 {
937 int ret;
938 g_autofree char *query = NULL;
939 uint32_t len;
940
941 ret = nbd_opt_read(client, &len, sizeof(len), false, errp);
942 if (ret <= 0) {
943 return ret;
944 }
945 len = cpu_to_be32(len);
946
947 if (len > NBD_MAX_STRING_SIZE) {
948 trace_nbd_negotiate_meta_query_skip("length too long");
949 return nbd_opt_skip(client, len, errp);
950 }
951
952 query = g_malloc(len + 1);
953 ret = nbd_opt_read(client, query, len, true, errp);
954 if (ret <= 0) {
955 return ret;
956 }
957 query[len] = '\0';
958
959 if (nbd_meta_base_query(client, meta, query)) {
960 return 1;
961 }
962 if (nbd_meta_qemu_query(client, meta, query)) {
963 return 1;
964 }
965
966 trace_nbd_negotiate_meta_query_skip("unknown namespace");
967 return 1;
968 }
969
970 /* nbd_negotiate_meta_queries
971 * Handle NBD_OPT_LIST_META_CONTEXT and NBD_OPT_SET_META_CONTEXT
972 *
973 * Return -errno on I/O error, or 0 if option was completely handled. */
974 static int nbd_negotiate_meta_queries(NBDClient *client,
975 NBDExportMetaContexts *meta, Error **errp)
976 {
977 int ret;
978 g_autofree char *export_name = NULL;
979 /* Mark unused to work around https://bugs.llvm.org/show_bug.cgi?id=3888 */
980 g_autofree G_GNUC_UNUSED bool *bitmaps = NULL;
981 NBDExportMetaContexts local_meta = {0};
982 uint32_t nb_queries;
983 size_t i;
984 size_t count = 0;
985
986 if (client->opt == NBD_OPT_SET_META_CONTEXT && !client->structured_reply) {
987 return nbd_opt_invalid(client, errp,
988 "request option '%s' when structured reply "
989 "is not negotiated",
990 nbd_opt_lookup(client->opt));
991 }
992
993 if (client->opt == NBD_OPT_LIST_META_CONTEXT) {
994 /* Only change the caller's meta on SET. */
995 meta = &local_meta;
996 }
997
998 g_free(meta->bitmaps);
999 memset(meta, 0, sizeof(*meta));
1000
1001 ret = nbd_opt_read_name(client, &export_name, NULL, errp);
1002 if (ret <= 0) {
1003 return ret;
1004 }
1005
1006 meta->exp = nbd_export_find(export_name);
1007 if (meta->exp == NULL) {
1008 g_autofree char *sane_name = nbd_sanitize_name(export_name);
1009
1010 return nbd_opt_drop(client, NBD_REP_ERR_UNKNOWN, errp,
1011 "export '%s' not present", sane_name);
1012 }
1013 meta->bitmaps = g_new0(bool, meta->exp->nr_export_bitmaps);
1014 if (client->opt == NBD_OPT_LIST_META_CONTEXT) {
1015 bitmaps = meta->bitmaps;
1016 }
1017
1018 ret = nbd_opt_read(client, &nb_queries, sizeof(nb_queries), false, errp);
1019 if (ret <= 0) {
1020 return ret;
1021 }
1022 nb_queries = cpu_to_be32(nb_queries);
1023 trace_nbd_negotiate_meta_context(nbd_opt_lookup(client->opt),
1024 export_name, nb_queries);
1025
1026 if (client->opt == NBD_OPT_LIST_META_CONTEXT && !nb_queries) {
1027 /* enable all known contexts */
1028 meta->base_allocation = true;
1029 meta->allocation_depth = meta->exp->allocation_depth;
1030 if (meta->exp->nr_export_bitmaps) {
1031 memset(meta->bitmaps, 1, meta->exp->nr_export_bitmaps);
1032 }
1033 } else {
1034 for (i = 0; i < nb_queries; ++i) {
1035 ret = nbd_negotiate_meta_query(client, meta, errp);
1036 if (ret <= 0) {
1037 return ret;
1038 }
1039 }
1040 }
1041
1042 if (meta->base_allocation) {
1043 ret = nbd_negotiate_send_meta_context(client, "base:allocation",
1044 NBD_META_ID_BASE_ALLOCATION,
1045 errp);
1046 if (ret < 0) {
1047 return ret;
1048 }
1049 count++;
1050 }
1051
1052 if (meta->allocation_depth) {
1053 ret = nbd_negotiate_send_meta_context(client, "qemu:allocation-depth",
1054 NBD_META_ID_ALLOCATION_DEPTH,
1055 errp);
1056 if (ret < 0) {
1057 return ret;
1058 }
1059 count++;
1060 }
1061
1062 for (i = 0; i < meta->exp->nr_export_bitmaps; i++) {
1063 const char *bm_name;
1064 g_autofree char *context = NULL;
1065
1066 if (!meta->bitmaps[i]) {
1067 continue;
1068 }
1069
1070 bm_name = bdrv_dirty_bitmap_name(meta->exp->export_bitmaps[i]);
1071 context = g_strdup_printf("qemu:dirty-bitmap:%s", bm_name);
1072
1073 ret = nbd_negotiate_send_meta_context(client, context,
1074 NBD_META_ID_DIRTY_BITMAP + i,
1075 errp);
1076 if (ret < 0) {
1077 return ret;
1078 }
1079 count++;
1080 }
1081
1082 ret = nbd_negotiate_send_rep(client, NBD_REP_ACK, errp);
1083 if (ret == 0) {
1084 meta->count = count;
1085 }
1086
1087 return ret;
1088 }
1089
1090 /* nbd_negotiate_options
1091 * Process all NBD_OPT_* client option commands, during fixed newstyle
1092 * negotiation.
1093 * Return:
1094 * -errno on error, errp is set
1095 * 0 on successful negotiation, errp is not set
1096 * 1 if client sent NBD_OPT_ABORT, i.e. on valid disconnect,
1097 * errp is not set
1098 */
1099 static int nbd_negotiate_options(NBDClient *client, Error **errp)
1100 {
1101 uint32_t flags;
1102 bool fixedNewstyle = false;
1103 bool no_zeroes = false;
1104
1105 /* Client sends:
1106 [ 0 .. 3] client flags
1107
1108 Then we loop until NBD_OPT_EXPORT_NAME or NBD_OPT_GO:
1109 [ 0 .. 7] NBD_OPTS_MAGIC
1110 [ 8 .. 11] NBD option
1111 [12 .. 15] Data length
1112 ... Rest of request
1113
1114 [ 0 .. 7] NBD_OPTS_MAGIC
1115 [ 8 .. 11] Second NBD option
1116 [12 .. 15] Data length
1117 ... Rest of request
1118 */
1119
1120 if (nbd_read32(client->ioc, &flags, "flags", errp) < 0) {
1121 return -EIO;
1122 }
1123 trace_nbd_negotiate_options_flags(flags);
1124 if (flags & NBD_FLAG_C_FIXED_NEWSTYLE) {
1125 fixedNewstyle = true;
1126 flags &= ~NBD_FLAG_C_FIXED_NEWSTYLE;
1127 }
1128 if (flags & NBD_FLAG_C_NO_ZEROES) {
1129 no_zeroes = true;
1130 flags &= ~NBD_FLAG_C_NO_ZEROES;
1131 }
1132 if (flags != 0) {
1133 error_setg(errp, "Unknown client flags 0x%" PRIx32 " received", flags);
1134 return -EINVAL;
1135 }
1136
1137 while (1) {
1138 int ret;
1139 uint32_t option, length;
1140 uint64_t magic;
1141
1142 if (nbd_read64(client->ioc, &magic, "opts magic", errp) < 0) {
1143 return -EINVAL;
1144 }
1145 trace_nbd_negotiate_options_check_magic(magic);
1146 if (magic != NBD_OPTS_MAGIC) {
1147 error_setg(errp, "Bad magic received");
1148 return -EINVAL;
1149 }
1150
1151 if (nbd_read32(client->ioc, &option, "option", errp) < 0) {
1152 return -EINVAL;
1153 }
1154 client->opt = option;
1155
1156 if (nbd_read32(client->ioc, &length, "option length", errp) < 0) {
1157 return -EINVAL;
1158 }
1159 assert(!client->optlen);
1160 client->optlen = length;
1161
1162 if (length > NBD_MAX_BUFFER_SIZE) {
1163 error_setg(errp, "len (%" PRIu32" ) is larger than max len (%u)",
1164 length, NBD_MAX_BUFFER_SIZE);
1165 return -EINVAL;
1166 }
1167
1168 trace_nbd_negotiate_options_check_option(option,
1169 nbd_opt_lookup(option));
1170 if (client->tlscreds &&
1171 client->ioc == (QIOChannel *)client->sioc) {
1172 QIOChannel *tioc;
1173 if (!fixedNewstyle) {
1174 error_setg(errp, "Unsupported option 0x%" PRIx32, option);
1175 return -EINVAL;
1176 }
1177 switch (option) {
1178 case NBD_OPT_STARTTLS:
1179 if (length) {
1180 /* Unconditionally drop the connection if the client
1181 * can't start a TLS negotiation correctly */
1182 return nbd_reject_length(client, true, errp);
1183 }
1184 tioc = nbd_negotiate_handle_starttls(client, errp);
1185 if (!tioc) {
1186 return -EIO;
1187 }
1188 ret = 0;
1189 object_unref(OBJECT(client->ioc));
1190 client->ioc = QIO_CHANNEL(tioc);
1191 break;
1192
1193 case NBD_OPT_EXPORT_NAME:
1194 /* No way to return an error to client, so drop connection */
1195 error_setg(errp, "Option 0x%x not permitted before TLS",
1196 option);
1197 return -EINVAL;
1198
1199 default:
1200 /* Let the client keep trying, unless they asked to
1201 * quit. Always try to give an error back to the
1202 * client; but when replying to OPT_ABORT, be aware
1203 * that the client may hang up before receiving the
1204 * error, in which case we are fine ignoring the
1205 * resulting EPIPE. */
1206 ret = nbd_opt_drop(client, NBD_REP_ERR_TLS_REQD,
1207 option == NBD_OPT_ABORT ? NULL : errp,
1208 "Option 0x%" PRIx32
1209 " not permitted before TLS", option);
1210 if (option == NBD_OPT_ABORT) {
1211 return 1;
1212 }
1213 break;
1214 }
1215 } else if (fixedNewstyle) {
1216 switch (option) {
1217 case NBD_OPT_LIST:
1218 if (length) {
1219 ret = nbd_reject_length(client, false, errp);
1220 } else {
1221 ret = nbd_negotiate_handle_list(client, errp);
1222 }
1223 break;
1224
1225 case NBD_OPT_ABORT:
1226 /* NBD spec says we must try to reply before
1227 * disconnecting, but that we must also tolerate
1228 * guests that don't wait for our reply. */
1229 nbd_negotiate_send_rep(client, NBD_REP_ACK, NULL);
1230 return 1;
1231
1232 case NBD_OPT_EXPORT_NAME:
1233 return nbd_negotiate_handle_export_name(client, no_zeroes,
1234 errp);
1235
1236 case NBD_OPT_INFO:
1237 case NBD_OPT_GO:
1238 ret = nbd_negotiate_handle_info(client, errp);
1239 if (ret == 1) {
1240 assert(option == NBD_OPT_GO);
1241 return 0;
1242 }
1243 break;
1244
1245 case NBD_OPT_STARTTLS:
1246 if (length) {
1247 ret = nbd_reject_length(client, false, errp);
1248 } else if (client->tlscreds) {
1249 ret = nbd_negotiate_send_rep_err(client,
1250 NBD_REP_ERR_INVALID, errp,
1251 "TLS already enabled");
1252 } else {
1253 ret = nbd_negotiate_send_rep_err(client,
1254 NBD_REP_ERR_POLICY, errp,
1255 "TLS not configured");
1256 }
1257 break;
1258
1259 case NBD_OPT_STRUCTURED_REPLY:
1260 if (length) {
1261 ret = nbd_reject_length(client, false, errp);
1262 } else if (client->structured_reply) {
1263 ret = nbd_negotiate_send_rep_err(
1264 client, NBD_REP_ERR_INVALID, errp,
1265 "structured reply already negotiated");
1266 } else {
1267 ret = nbd_negotiate_send_rep(client, NBD_REP_ACK, errp);
1268 client->structured_reply = true;
1269 }
1270 break;
1271
1272 case NBD_OPT_LIST_META_CONTEXT:
1273 case NBD_OPT_SET_META_CONTEXT:
1274 ret = nbd_negotiate_meta_queries(client, &client->export_meta,
1275 errp);
1276 break;
1277
1278 default:
1279 ret = nbd_opt_drop(client, NBD_REP_ERR_UNSUP, errp,
1280 "Unsupported option %" PRIu32 " (%s)",
1281 option, nbd_opt_lookup(option));
1282 break;
1283 }
1284 } else {
1285 /*
1286 * If broken new-style we should drop the connection
1287 * for anything except NBD_OPT_EXPORT_NAME
1288 */
1289 switch (option) {
1290 case NBD_OPT_EXPORT_NAME:
1291 return nbd_negotiate_handle_export_name(client, no_zeroes,
1292 errp);
1293
1294 default:
1295 error_setg(errp, "Unsupported option %" PRIu32 " (%s)",
1296 option, nbd_opt_lookup(option));
1297 return -EINVAL;
1298 }
1299 }
1300 if (ret < 0) {
1301 return ret;
1302 }
1303 }
1304 }
1305
1306 /* nbd_negotiate
1307 * Return:
1308 * -errno on error, errp is set
1309 * 0 on successful negotiation, errp is not set
1310 * 1 if client sent NBD_OPT_ABORT, i.e. on valid disconnect,
1311 * errp is not set
1312 */
1313 static coroutine_fn int nbd_negotiate(NBDClient *client, Error **errp)
1314 {
1315 ERRP_GUARD();
1316 char buf[NBD_OLDSTYLE_NEGOTIATE_SIZE] = "";
1317 int ret;
1318
1319 /* Old style negotiation header, no room for options
1320 [ 0 .. 7] passwd ("NBDMAGIC")
1321 [ 8 .. 15] magic (NBD_CLIENT_MAGIC)
1322 [16 .. 23] size
1323 [24 .. 27] export flags (zero-extended)
1324 [28 .. 151] reserved (0)
1325
1326 New style negotiation header, client can send options
1327 [ 0 .. 7] passwd ("NBDMAGIC")
1328 [ 8 .. 15] magic (NBD_OPTS_MAGIC)
1329 [16 .. 17] server flags (0)
1330 ....options sent, ending in NBD_OPT_EXPORT_NAME or NBD_OPT_GO....
1331 */
1332
1333 qio_channel_set_blocking(client->ioc, false, NULL);
1334
1335 trace_nbd_negotiate_begin();
1336 memcpy(buf, "NBDMAGIC", 8);
1337
1338 stq_be_p(buf + 8, NBD_OPTS_MAGIC);
1339 stw_be_p(buf + 16, NBD_FLAG_FIXED_NEWSTYLE | NBD_FLAG_NO_ZEROES);
1340
1341 if (nbd_write(client->ioc, buf, 18, errp) < 0) {
1342 error_prepend(errp, "write failed: ");
1343 return -EINVAL;
1344 }
1345 ret = nbd_negotiate_options(client, errp);
1346 if (ret != 0) {
1347 if (ret < 0) {
1348 error_prepend(errp, "option negotiation failed: ");
1349 }
1350 return ret;
1351 }
1352
1353 /* Attach the channel to the same AioContext as the export */
1354 if (client->exp && client->exp->common.ctx) {
1355 qio_channel_attach_aio_context(client->ioc, client->exp->common.ctx);
1356 }
1357
1358 assert(!client->optlen);
1359 trace_nbd_negotiate_success();
1360
1361 return 0;
1362 }
1363
1364 /* nbd_read_eof
1365 * Tries to read @size bytes from @ioc. This is a local implementation of
1366 * qio_channel_readv_all_eof. We have it here because we need it to be
1367 * interruptible and to know when the coroutine is yielding.
1368 * Returns 1 on success
1369 * 0 on eof, when no data was read (errp is not set)
1370 * negative errno on failure (errp is set)
1371 */
1372 static inline int coroutine_fn
1373 nbd_read_eof(NBDClient *client, void *buffer, size_t size, Error **errp)
1374 {
1375 bool partial = false;
1376
1377 assert(size);
1378 while (size > 0) {
1379 struct iovec iov = { .iov_base = buffer, .iov_len = size };
1380 ssize_t len;
1381
1382 len = qio_channel_readv(client->ioc, &iov, 1, errp);
1383 if (len == QIO_CHANNEL_ERR_BLOCK) {
1384 client->read_yielding = true;
1385 qio_channel_yield(client->ioc, G_IO_IN);
1386 client->read_yielding = false;
1387 if (client->quiescing) {
1388 return -EAGAIN;
1389 }
1390 continue;
1391 } else if (len < 0) {
1392 return -EIO;
1393 } else if (len == 0) {
1394 if (partial) {
1395 error_setg(errp,
1396 "Unexpected end-of-file before all bytes were read");
1397 return -EIO;
1398 } else {
1399 return 0;
1400 }
1401 }
1402
1403 partial = true;
1404 size -= len;
1405 buffer = (uint8_t *) buffer + len;
1406 }
1407 return 1;
1408 }
1409
1410 static int nbd_receive_request(NBDClient *client, NBDRequest *request,
1411 Error **errp)
1412 {
1413 uint8_t buf[NBD_REQUEST_SIZE];
1414 uint32_t magic;
1415 int ret;
1416
1417 ret = nbd_read_eof(client, buf, sizeof(buf), errp);
1418 if (ret < 0) {
1419 return ret;
1420 }
1421 if (ret == 0) {
1422 return -EIO;
1423 }
1424
1425 /* Request
1426 [ 0 .. 3] magic (NBD_REQUEST_MAGIC)
1427 [ 4 .. 5] flags (NBD_CMD_FLAG_FUA, ...)
1428 [ 6 .. 7] type (NBD_CMD_READ, ...)
1429 [ 8 .. 15] handle
1430 [16 .. 23] from
1431 [24 .. 27] len
1432 */
1433
1434 magic = ldl_be_p(buf);
1435 request->flags = lduw_be_p(buf + 4);
1436 request->type = lduw_be_p(buf + 6);
1437 request->handle = ldq_be_p(buf + 8);
1438 request->from = ldq_be_p(buf + 16);
1439 request->len = ldl_be_p(buf + 24);
1440
1441 trace_nbd_receive_request(magic, request->flags, request->type,
1442 request->from, request->len);
1443
1444 if (magic != NBD_REQUEST_MAGIC) {
1445 error_setg(errp, "invalid magic (got 0x%" PRIx32 ")", magic);
1446 return -EINVAL;
1447 }
1448 return 0;
1449 }
1450
1451 #define MAX_NBD_REQUESTS 16
1452
1453 void nbd_client_get(NBDClient *client)
1454 {
1455 client->refcount++;
1456 }
1457
1458 void nbd_client_put(NBDClient *client)
1459 {
1460 if (--client->refcount == 0) {
1461 /* The last reference should be dropped by client->close,
1462 * which is called by client_close.
1463 */
1464 assert(client->closing);
1465
1466 qio_channel_detach_aio_context(client->ioc);
1467 object_unref(OBJECT(client->sioc));
1468 object_unref(OBJECT(client->ioc));
1469 if (client->tlscreds) {
1470 object_unref(OBJECT(client->tlscreds));
1471 }
1472 g_free(client->tlsauthz);
1473 if (client->exp) {
1474 QTAILQ_REMOVE(&client->exp->clients, client, next);
1475 blk_exp_unref(&client->exp->common);
1476 }
1477 g_free(client->export_meta.bitmaps);
1478 g_free(client);
1479 }
1480 }
1481
1482 static void client_close(NBDClient *client, bool negotiated)
1483 {
1484 if (client->closing) {
1485 return;
1486 }
1487
1488 client->closing = true;
1489
1490 /* Force requests to finish. They will drop their own references,
1491 * then we'll close the socket and free the NBDClient.
1492 */
1493 qio_channel_shutdown(client->ioc, QIO_CHANNEL_SHUTDOWN_BOTH,
1494 NULL);
1495
1496 /* Also tell the client, so that they release their reference. */
1497 if (client->close_fn) {
1498 client->close_fn(client, negotiated);
1499 }
1500 }
1501
1502 static NBDRequestData *nbd_request_get(NBDClient *client)
1503 {
1504 NBDRequestData *req;
1505
1506 assert(client->nb_requests <= MAX_NBD_REQUESTS - 1);
1507 client->nb_requests++;
1508
1509 req = g_new0(NBDRequestData, 1);
1510 nbd_client_get(client);
1511 req->client = client;
1512 return req;
1513 }
1514
1515 static void nbd_request_put(NBDRequestData *req)
1516 {
1517 NBDClient *client = req->client;
1518
1519 if (req->data) {
1520 qemu_vfree(req->data);
1521 }
1522 g_free(req);
1523
1524 client->nb_requests--;
1525
1526 if (client->quiescing && client->nb_requests == 0) {
1527 aio_wait_kick();
1528 }
1529
1530 nbd_client_receive_next_request(client);
1531
1532 nbd_client_put(client);
1533 }
1534
1535 static void blk_aio_attached(AioContext *ctx, void *opaque)
1536 {
1537 NBDExport *exp = opaque;
1538 NBDClient *client;
1539
1540 trace_nbd_blk_aio_attached(exp->name, ctx);
1541
1542 exp->common.ctx = ctx;
1543
1544 QTAILQ_FOREACH(client, &exp->clients, next) {
1545 qio_channel_attach_aio_context(client->ioc, ctx);
1546
1547 assert(client->nb_requests == 0);
1548 assert(client->recv_coroutine == NULL);
1549 assert(client->send_coroutine == NULL);
1550 }
1551 }
1552
1553 static void blk_aio_detach(void *opaque)
1554 {
1555 NBDExport *exp = opaque;
1556 NBDClient *client;
1557
1558 trace_nbd_blk_aio_detach(exp->name, exp->common.ctx);
1559
1560 QTAILQ_FOREACH(client, &exp->clients, next) {
1561 qio_channel_detach_aio_context(client->ioc);
1562 }
1563
1564 exp->common.ctx = NULL;
1565 }
1566
1567 static void nbd_drained_begin(void *opaque)
1568 {
1569 NBDExport *exp = opaque;
1570 NBDClient *client;
1571
1572 QTAILQ_FOREACH(client, &exp->clients, next) {
1573 client->quiescing = true;
1574 }
1575 }
1576
1577 static void nbd_drained_end(void *opaque)
1578 {
1579 NBDExport *exp = opaque;
1580 NBDClient *client;
1581
1582 QTAILQ_FOREACH(client, &exp->clients, next) {
1583 client->quiescing = false;
1584 nbd_client_receive_next_request(client);
1585 }
1586 }
1587
1588 static bool nbd_drained_poll(void *opaque)
1589 {
1590 NBDExport *exp = opaque;
1591 NBDClient *client;
1592
1593 QTAILQ_FOREACH(client, &exp->clients, next) {
1594 if (client->nb_requests != 0) {
1595 /*
1596 * If there's a coroutine waiting for a request on nbd_read_eof()
1597 * enter it here so we don't depend on the client to wake it up.
1598 */
1599 if (client->recv_coroutine != NULL && client->read_yielding) {
1600 qemu_aio_coroutine_enter(exp->common.ctx,
1601 client->recv_coroutine);
1602 }
1603
1604 return true;
1605 }
1606 }
1607
1608 return false;
1609 }
1610
1611 static void nbd_eject_notifier(Notifier *n, void *data)
1612 {
1613 NBDExport *exp = container_of(n, NBDExport, eject_notifier);
1614
1615 blk_exp_request_shutdown(&exp->common);
1616 }
1617
1618 void nbd_export_set_on_eject_blk(BlockExport *exp, BlockBackend *blk)
1619 {
1620 NBDExport *nbd_exp = container_of(exp, NBDExport, common);
1621 assert(exp->drv == &blk_exp_nbd);
1622 assert(nbd_exp->eject_notifier_blk == NULL);
1623
1624 blk_ref(blk);
1625 nbd_exp->eject_notifier_blk = blk;
1626 nbd_exp->eject_notifier.notify = nbd_eject_notifier;
1627 blk_add_remove_bs_notifier(blk, &nbd_exp->eject_notifier);
1628 }
1629
1630 static const BlockDevOps nbd_block_ops = {
1631 .drained_begin = nbd_drained_begin,
1632 .drained_end = nbd_drained_end,
1633 .drained_poll = nbd_drained_poll,
1634 };
1635
1636 static int nbd_export_create(BlockExport *blk_exp, BlockExportOptions *exp_args,
1637 Error **errp)
1638 {
1639 NBDExport *exp = container_of(blk_exp, NBDExport, common);
1640 BlockExportOptionsNbd *arg = &exp_args->u.nbd;
1641 BlockBackend *blk = blk_exp->blk;
1642 int64_t size;
1643 uint64_t perm, shared_perm;
1644 bool readonly = !exp_args->writable;
1645 bool shared = !exp_args->writable;
1646 strList *bitmaps;
1647 size_t i;
1648 int ret;
1649
1650 assert(exp_args->type == BLOCK_EXPORT_TYPE_NBD);
1651
1652 if (!nbd_server_is_running()) {
1653 error_setg(errp, "NBD server not running");
1654 return -EINVAL;
1655 }
1656
1657 if (!arg->has_name) {
1658 arg->name = exp_args->node_name;
1659 }
1660
1661 if (strlen(arg->name) > NBD_MAX_STRING_SIZE) {
1662 error_setg(errp, "export name '%s' too long", arg->name);
1663 return -EINVAL;
1664 }
1665
1666 if (arg->description && strlen(arg->description) > NBD_MAX_STRING_SIZE) {
1667 error_setg(errp, "description '%s' too long", arg->description);
1668 return -EINVAL;
1669 }
1670
1671 if (nbd_export_find(arg->name)) {
1672 error_setg(errp, "NBD server already has export named '%s'", arg->name);
1673 return -EEXIST;
1674 }
1675
1676 size = blk_getlength(blk);
1677 if (size < 0) {
1678 error_setg_errno(errp, -size,
1679 "Failed to determine the NBD export's length");
1680 return size;
1681 }
1682
1683 /* Don't allow resize while the NBD server is running, otherwise we don't
1684 * care what happens with the node. */
1685 blk_get_perm(blk, &perm, &shared_perm);
1686 ret = blk_set_perm(blk, perm, shared_perm & ~BLK_PERM_RESIZE, errp);
1687 if (ret < 0) {
1688 return ret;
1689 }
1690
1691 QTAILQ_INIT(&exp->clients);
1692 exp->name = g_strdup(arg->name);
1693 exp->description = g_strdup(arg->description);
1694 exp->nbdflags = (NBD_FLAG_HAS_FLAGS | NBD_FLAG_SEND_FLUSH |
1695 NBD_FLAG_SEND_FUA | NBD_FLAG_SEND_CACHE);
1696 if (readonly) {
1697 exp->nbdflags |= NBD_FLAG_READ_ONLY;
1698 if (shared) {
1699 exp->nbdflags |= NBD_FLAG_CAN_MULTI_CONN;
1700 }
1701 } else {
1702 exp->nbdflags |= (NBD_FLAG_SEND_TRIM | NBD_FLAG_SEND_WRITE_ZEROES |
1703 NBD_FLAG_SEND_FAST_ZERO);
1704 }
1705 exp->size = QEMU_ALIGN_DOWN(size, BDRV_SECTOR_SIZE);
1706
1707 for (bitmaps = arg->bitmaps; bitmaps; bitmaps = bitmaps->next) {
1708 exp->nr_export_bitmaps++;
1709 }
1710 exp->export_bitmaps = g_new0(BdrvDirtyBitmap *, exp->nr_export_bitmaps);
1711 for (i = 0, bitmaps = arg->bitmaps; bitmaps;
1712 i++, bitmaps = bitmaps->next) {
1713 const char *bitmap = bitmaps->value;
1714 BlockDriverState *bs = blk_bs(blk);
1715 BdrvDirtyBitmap *bm = NULL;
1716
1717 while (bs) {
1718 bm = bdrv_find_dirty_bitmap(bs, bitmap);
1719 if (bm != NULL) {
1720 break;
1721 }
1722
1723 bs = bdrv_filter_or_cow_bs(bs);
1724 }
1725
1726 if (bm == NULL) {
1727 ret = -ENOENT;
1728 error_setg(errp, "Bitmap '%s' is not found", bitmap);
1729 goto fail;
1730 }
1731
1732 if (bdrv_dirty_bitmap_check(bm, BDRV_BITMAP_ALLOW_RO, errp)) {
1733 ret = -EINVAL;
1734 goto fail;
1735 }
1736
1737 if (readonly && bdrv_is_writable(bs) &&
1738 bdrv_dirty_bitmap_enabled(bm)) {
1739 ret = -EINVAL;
1740 error_setg(errp,
1741 "Enabled bitmap '%s' incompatible with readonly export",
1742 bitmap);
1743 goto fail;
1744 }
1745
1746 exp->export_bitmaps[i] = bm;
1747 assert(strlen(bitmap) <= BDRV_BITMAP_MAX_NAME_SIZE);
1748 }
1749
1750 /* Mark bitmaps busy in a separate loop, to simplify roll-back concerns. */
1751 for (i = 0; i < exp->nr_export_bitmaps; i++) {
1752 bdrv_dirty_bitmap_set_busy(exp->export_bitmaps[i], true);
1753 }
1754
1755 exp->allocation_depth = arg->allocation_depth;
1756
1757 /*
1758 * We need to inhibit request queuing in the block layer to ensure we can
1759 * be properly quiesced when entering a drained section, as our coroutines
1760 * servicing pending requests might enter blk_pread().
1761 */
1762 blk_set_disable_request_queuing(blk, true);
1763
1764 blk_add_aio_context_notifier(blk, blk_aio_attached, blk_aio_detach, exp);
1765
1766 blk_set_dev_ops(blk, &nbd_block_ops, exp);
1767
1768 QTAILQ_INSERT_TAIL(&exports, exp, next);
1769
1770 return 0;
1771
1772 fail:
1773 g_free(exp->export_bitmaps);
1774 g_free(exp->name);
1775 g_free(exp->description);
1776 return ret;
1777 }
1778
1779 NBDExport *nbd_export_find(const char *name)
1780 {
1781 NBDExport *exp;
1782 QTAILQ_FOREACH(exp, &exports, next) {
1783 if (strcmp(name, exp->name) == 0) {
1784 return exp;
1785 }
1786 }
1787
1788 return NULL;
1789 }
1790
1791 AioContext *
1792 nbd_export_aio_context(NBDExport *exp)
1793 {
1794 return exp->common.ctx;
1795 }
1796
1797 static void nbd_export_request_shutdown(BlockExport *blk_exp)
1798 {
1799 NBDExport *exp = container_of(blk_exp, NBDExport, common);
1800 NBDClient *client, *next;
1801
1802 blk_exp_ref(&exp->common);
1803 /*
1804 * TODO: Should we expand QMP NbdServerRemoveNode enum to allow a
1805 * close mode that stops advertising the export to new clients but
1806 * still permits existing clients to run to completion? Because of
1807 * that possibility, nbd_export_close() can be called more than
1808 * once on an export.
1809 */
1810 QTAILQ_FOREACH_SAFE(client, &exp->clients, next, next) {
1811 client_close(client, true);
1812 }
1813 if (exp->name) {
1814 g_free(exp->name);
1815 exp->name = NULL;
1816 QTAILQ_REMOVE(&exports, exp, next);
1817 }
1818 blk_exp_unref(&exp->common);
1819 }
1820
1821 static void nbd_export_delete(BlockExport *blk_exp)
1822 {
1823 size_t i;
1824 NBDExport *exp = container_of(blk_exp, NBDExport, common);
1825
1826 assert(exp->name == NULL);
1827 assert(QTAILQ_EMPTY(&exp->clients));
1828
1829 g_free(exp->description);
1830 exp->description = NULL;
1831
1832 if (exp->common.blk) {
1833 if (exp->eject_notifier_blk) {
1834 notifier_remove(&exp->eject_notifier);
1835 blk_unref(exp->eject_notifier_blk);
1836 }
1837 blk_remove_aio_context_notifier(exp->common.blk, blk_aio_attached,
1838 blk_aio_detach, exp);
1839 blk_set_disable_request_queuing(exp->common.blk, false);
1840 }
1841
1842 for (i = 0; i < exp->nr_export_bitmaps; i++) {
1843 bdrv_dirty_bitmap_set_busy(exp->export_bitmaps[i], false);
1844 }
1845 }
1846
1847 const BlockExportDriver blk_exp_nbd = {
1848 .type = BLOCK_EXPORT_TYPE_NBD,
1849 .instance_size = sizeof(NBDExport),
1850 .create = nbd_export_create,
1851 .delete = nbd_export_delete,
1852 .request_shutdown = nbd_export_request_shutdown,
1853 };
1854
1855 static int coroutine_fn nbd_co_send_iov(NBDClient *client, struct iovec *iov,
1856 unsigned niov, Error **errp)
1857 {
1858 int ret;
1859
1860 g_assert(qemu_in_coroutine());
1861 qemu_co_mutex_lock(&client->send_lock);
1862 client->send_coroutine = qemu_coroutine_self();
1863
1864 ret = qio_channel_writev_all(client->ioc, iov, niov, errp) < 0 ? -EIO : 0;
1865
1866 client->send_coroutine = NULL;
1867 qemu_co_mutex_unlock(&client->send_lock);
1868
1869 return ret;
1870 }
1871
1872 static inline void set_be_simple_reply(NBDSimpleReply *reply, uint64_t error,
1873 uint64_t handle)
1874 {
1875 stl_be_p(&reply->magic, NBD_SIMPLE_REPLY_MAGIC);
1876 stl_be_p(&reply->error, error);
1877 stq_be_p(&reply->handle, handle);
1878 }
1879
1880 static int nbd_co_send_simple_reply(NBDClient *client,
1881 uint64_t handle,
1882 uint32_t error,
1883 void *data,
1884 size_t len,
1885 Error **errp)
1886 {
1887 NBDSimpleReply reply;
1888 int nbd_err = system_errno_to_nbd_errno(error);
1889 struct iovec iov[] = {
1890 {.iov_base = &reply, .iov_len = sizeof(reply)},
1891 {.iov_base = data, .iov_len = len}
1892 };
1893
1894 trace_nbd_co_send_simple_reply(handle, nbd_err, nbd_err_lookup(nbd_err),
1895 len);
1896 set_be_simple_reply(&reply, nbd_err, handle);
1897
1898 return nbd_co_send_iov(client, iov, len ? 2 : 1, errp);
1899 }
1900
1901 static inline void set_be_chunk(NBDStructuredReplyChunk *chunk, uint16_t flags,
1902 uint16_t type, uint64_t handle, uint32_t length)
1903 {
1904 stl_be_p(&chunk->magic, NBD_STRUCTURED_REPLY_MAGIC);
1905 stw_be_p(&chunk->flags, flags);
1906 stw_be_p(&chunk->type, type);
1907 stq_be_p(&chunk->handle, handle);
1908 stl_be_p(&chunk->length, length);
1909 }
1910
1911 static int coroutine_fn nbd_co_send_structured_done(NBDClient *client,
1912 uint64_t handle,
1913 Error **errp)
1914 {
1915 NBDStructuredReplyChunk chunk;
1916 struct iovec iov[] = {
1917 {.iov_base = &chunk, .iov_len = sizeof(chunk)},
1918 };
1919
1920 trace_nbd_co_send_structured_done(handle);
1921 set_be_chunk(&chunk, NBD_REPLY_FLAG_DONE, NBD_REPLY_TYPE_NONE, handle, 0);
1922
1923 return nbd_co_send_iov(client, iov, 1, errp);
1924 }
1925
1926 static int coroutine_fn nbd_co_send_structured_read(NBDClient *client,
1927 uint64_t handle,
1928 uint64_t offset,
1929 void *data,
1930 size_t size,
1931 bool final,
1932 Error **errp)
1933 {
1934 NBDStructuredReadData chunk;
1935 struct iovec iov[] = {
1936 {.iov_base = &chunk, .iov_len = sizeof(chunk)},
1937 {.iov_base = data, .iov_len = size}
1938 };
1939
1940 assert(size);
1941 trace_nbd_co_send_structured_read(handle, offset, data, size);
1942 set_be_chunk(&chunk.h, final ? NBD_REPLY_FLAG_DONE : 0,
1943 NBD_REPLY_TYPE_OFFSET_DATA, handle,
1944 sizeof(chunk) - sizeof(chunk.h) + size);
1945 stq_be_p(&chunk.offset, offset);
1946
1947 return nbd_co_send_iov(client, iov, 2, errp);
1948 }
1949
1950 static int coroutine_fn nbd_co_send_structured_error(NBDClient *client,
1951 uint64_t handle,
1952 uint32_t error,
1953 const char *msg,
1954 Error **errp)
1955 {
1956 NBDStructuredError chunk;
1957 int nbd_err = system_errno_to_nbd_errno(error);
1958 struct iovec iov[] = {
1959 {.iov_base = &chunk, .iov_len = sizeof(chunk)},
1960 {.iov_base = (char *)msg, .iov_len = msg ? strlen(msg) : 0},
1961 };
1962
1963 assert(nbd_err);
1964 trace_nbd_co_send_structured_error(handle, nbd_err,
1965 nbd_err_lookup(nbd_err), msg ? msg : "");
1966 set_be_chunk(&chunk.h, NBD_REPLY_FLAG_DONE, NBD_REPLY_TYPE_ERROR, handle,
1967 sizeof(chunk) - sizeof(chunk.h) + iov[1].iov_len);
1968 stl_be_p(&chunk.error, nbd_err);
1969 stw_be_p(&chunk.message_length, iov[1].iov_len);
1970
1971 return nbd_co_send_iov(client, iov, 1 + !!iov[1].iov_len, errp);
1972 }
1973
1974 /* Do a sparse read and send the structured reply to the client.
1975 * Returns -errno if sending fails. bdrv_block_status_above() failure is
1976 * reported to the client, at which point this function succeeds.
1977 */
1978 static int coroutine_fn nbd_co_send_sparse_read(NBDClient *client,
1979 uint64_t handle,
1980 uint64_t offset,
1981 uint8_t *data,
1982 size_t size,
1983 Error **errp)
1984 {
1985 int ret = 0;
1986 NBDExport *exp = client->exp;
1987 size_t progress = 0;
1988
1989 while (progress < size) {
1990 int64_t pnum;
1991 int status = bdrv_block_status_above(blk_bs(exp->common.blk), NULL,
1992 offset + progress,
1993 size - progress, &pnum, NULL,
1994 NULL);
1995 bool final;
1996
1997 if (status < 0) {
1998 char *msg = g_strdup_printf("unable to check for holes: %s",
1999 strerror(-status));
2000
2001 ret = nbd_co_send_structured_error(client, handle, -status, msg,
2002 errp);
2003 g_free(msg);
2004 return ret;
2005 }
2006 assert(pnum && pnum <= size - progress);
2007 final = progress + pnum == size;
2008 if (status & BDRV_BLOCK_ZERO) {
2009 NBDStructuredReadHole chunk;
2010 struct iovec iov[] = {
2011 {.iov_base = &chunk, .iov_len = sizeof(chunk)},
2012 };
2013
2014 trace_nbd_co_send_structured_read_hole(handle, offset + progress,
2015 pnum);
2016 set_be_chunk(&chunk.h, final ? NBD_REPLY_FLAG_DONE : 0,
2017 NBD_REPLY_TYPE_OFFSET_HOLE,
2018 handle, sizeof(chunk) - sizeof(chunk.h));
2019 stq_be_p(&chunk.offset, offset + progress);
2020 stl_be_p(&chunk.length, pnum);
2021 ret = nbd_co_send_iov(client, iov, 1, errp);
2022 } else {
2023 ret = blk_pread(exp->common.blk, offset + progress,
2024 data + progress, pnum);
2025 if (ret < 0) {
2026 error_setg_errno(errp, -ret, "reading from file failed");
2027 break;
2028 }
2029 ret = nbd_co_send_structured_read(client, handle, offset + progress,
2030 data + progress, pnum, final,
2031 errp);
2032 }
2033
2034 if (ret < 0) {
2035 break;
2036 }
2037 progress += pnum;
2038 }
2039 return ret;
2040 }
2041
2042 typedef struct NBDExtentArray {
2043 NBDExtent *extents;
2044 unsigned int nb_alloc;
2045 unsigned int count;
2046 uint64_t total_length;
2047 bool can_add;
2048 bool converted_to_be;
2049 } NBDExtentArray;
2050
2051 static NBDExtentArray *nbd_extent_array_new(unsigned int nb_alloc)
2052 {
2053 NBDExtentArray *ea = g_new0(NBDExtentArray, 1);
2054
2055 ea->nb_alloc = nb_alloc;
2056 ea->extents = g_new(NBDExtent, nb_alloc);
2057 ea->can_add = true;
2058
2059 return ea;
2060 }
2061
2062 static void nbd_extent_array_free(NBDExtentArray *ea)
2063 {
2064 g_free(ea->extents);
2065 g_free(ea);
2066 }
2067 G_DEFINE_AUTOPTR_CLEANUP_FUNC(NBDExtentArray, nbd_extent_array_free);
2068
2069 /* Further modifications of the array after conversion are abandoned */
2070 static void nbd_extent_array_convert_to_be(NBDExtentArray *ea)
2071 {
2072 int i;
2073
2074 assert(!ea->converted_to_be);
2075 ea->can_add = false;
2076 ea->converted_to_be = true;
2077
2078 for (i = 0; i < ea->count; i++) {
2079 ea->extents[i].flags = cpu_to_be32(ea->extents[i].flags);
2080 ea->extents[i].length = cpu_to_be32(ea->extents[i].length);
2081 }
2082 }
2083
2084 /*
2085 * Add extent to NBDExtentArray. If extent can't be added (no available space),
2086 * return -1.
2087 * For safety, when returning -1 for the first time, .can_add is set to false,
2088 * further call to nbd_extent_array_add() will crash.
2089 * (to avoid the situation, when after failing to add an extent (returned -1),
2090 * user miss this failure and add another extent, which is successfully added
2091 * (array is full, but new extent may be squashed into the last one), then we
2092 * have invalid array with skipped extent)
2093 */
2094 static int nbd_extent_array_add(NBDExtentArray *ea,
2095 uint32_t length, uint32_t flags)
2096 {
2097 assert(ea->can_add);
2098
2099 if (!length) {
2100 return 0;
2101 }
2102
2103 /* Extend previous extent if flags are the same */
2104 if (ea->count > 0 && flags == ea->extents[ea->count - 1].flags) {
2105 uint64_t sum = (uint64_t)length + ea->extents[ea->count - 1].length;
2106
2107 if (sum <= UINT32_MAX) {
2108 ea->extents[ea->count - 1].length = sum;
2109 ea->total_length += length;
2110 return 0;
2111 }
2112 }
2113
2114 if (ea->count >= ea->nb_alloc) {
2115 ea->can_add = false;
2116 return -1;
2117 }
2118
2119 ea->total_length += length;
2120 ea->extents[ea->count] = (NBDExtent) {.length = length, .flags = flags};
2121 ea->count++;
2122
2123 return 0;
2124 }
2125
2126 static int blockstatus_to_extents(BlockDriverState *bs, uint64_t offset,
2127 uint64_t bytes, NBDExtentArray *ea)
2128 {
2129 while (bytes) {
2130 uint32_t flags;
2131 int64_t num;
2132 int ret = bdrv_block_status_above(bs, NULL, offset, bytes, &num,
2133 NULL, NULL);
2134
2135 if (ret < 0) {
2136 return ret;
2137 }
2138
2139 flags = (ret & BDRV_BLOCK_DATA ? 0 : NBD_STATE_HOLE) |
2140 (ret & BDRV_BLOCK_ZERO ? NBD_STATE_ZERO : 0);
2141
2142 if (nbd_extent_array_add(ea, num, flags) < 0) {
2143 return 0;
2144 }
2145
2146 offset += num;
2147 bytes -= num;
2148 }
2149
2150 return 0;
2151 }
2152
2153 static int blockalloc_to_extents(BlockDriverState *bs, uint64_t offset,
2154 uint64_t bytes, NBDExtentArray *ea)
2155 {
2156 while (bytes) {
2157 int64_t num;
2158 int ret = bdrv_is_allocated_above(bs, NULL, false, offset, bytes,
2159 &num);
2160
2161 if (ret < 0) {
2162 return ret;
2163 }
2164
2165 if (nbd_extent_array_add(ea, num, ret) < 0) {
2166 return 0;
2167 }
2168
2169 offset += num;
2170 bytes -= num;
2171 }
2172
2173 return 0;
2174 }
2175
2176 /*
2177 * nbd_co_send_extents
2178 *
2179 * @ea is converted to BE by the function
2180 * @last controls whether NBD_REPLY_FLAG_DONE is sent.
2181 */
2182 static int nbd_co_send_extents(NBDClient *client, uint64_t handle,
2183 NBDExtentArray *ea,
2184 bool last, uint32_t context_id, Error **errp)
2185 {
2186 NBDStructuredMeta chunk;
2187 struct iovec iov[] = {
2188 {.iov_base = &chunk, .iov_len = sizeof(chunk)},
2189 {.iov_base = ea->extents, .iov_len = ea->count * sizeof(ea->extents[0])}
2190 };
2191
2192 nbd_extent_array_convert_to_be(ea);
2193
2194 trace_nbd_co_send_extents(handle, ea->count, context_id, ea->total_length,
2195 last);
2196 set_be_chunk(&chunk.h, last ? NBD_REPLY_FLAG_DONE : 0,
2197 NBD_REPLY_TYPE_BLOCK_STATUS,
2198 handle, sizeof(chunk) - sizeof(chunk.h) + iov[1].iov_len);
2199 stl_be_p(&chunk.context_id, context_id);
2200
2201 return nbd_co_send_iov(client, iov, 2, errp);
2202 }
2203
2204 /* Get block status from the exported device and send it to the client */
2205 static int nbd_co_send_block_status(NBDClient *client, uint64_t handle,
2206 BlockDriverState *bs, uint64_t offset,
2207 uint32_t length, bool dont_fragment,
2208 bool last, uint32_t context_id,
2209 Error **errp)
2210 {
2211 int ret;
2212 unsigned int nb_extents = dont_fragment ? 1 : NBD_MAX_BLOCK_STATUS_EXTENTS;
2213 g_autoptr(NBDExtentArray) ea = nbd_extent_array_new(nb_extents);
2214
2215 if (context_id == NBD_META_ID_BASE_ALLOCATION) {
2216 ret = blockstatus_to_extents(bs, offset, length, ea);
2217 } else {
2218 ret = blockalloc_to_extents(bs, offset, length, ea);
2219 }
2220 if (ret < 0) {
2221 return nbd_co_send_structured_error(
2222 client, handle, -ret, "can't get block status", errp);
2223 }
2224
2225 return nbd_co_send_extents(client, handle, ea, last, context_id, errp);
2226 }
2227
2228 /* Populate @ea from a dirty bitmap. */
2229 static void bitmap_to_extents(BdrvDirtyBitmap *bitmap,
2230 uint64_t offset, uint64_t length,
2231 NBDExtentArray *es)
2232 {
2233 int64_t start, dirty_start, dirty_count;
2234 int64_t end = offset + length;
2235 bool full = false;
2236
2237 bdrv_dirty_bitmap_lock(bitmap);
2238
2239 for (start = offset;
2240 bdrv_dirty_bitmap_next_dirty_area(bitmap, start, end, INT32_MAX,
2241 &dirty_start, &dirty_count);
2242 start = dirty_start + dirty_count)
2243 {
2244 if ((nbd_extent_array_add(es, dirty_start - start, 0) < 0) ||
2245 (nbd_extent_array_add(es, dirty_count, NBD_STATE_DIRTY) < 0))
2246 {
2247 full = true;
2248 break;
2249 }
2250 }
2251
2252 if (!full) {
2253 /* last non dirty extent, nothing to do if array is now full */
2254 (void) nbd_extent_array_add(es, end - start, 0);
2255 }
2256
2257 bdrv_dirty_bitmap_unlock(bitmap);
2258 }
2259
2260 static int nbd_co_send_bitmap(NBDClient *client, uint64_t handle,
2261 BdrvDirtyBitmap *bitmap, uint64_t offset,
2262 uint32_t length, bool dont_fragment, bool last,
2263 uint32_t context_id, Error **errp)
2264 {
2265 unsigned int nb_extents = dont_fragment ? 1 : NBD_MAX_BLOCK_STATUS_EXTENTS;
2266 g_autoptr(NBDExtentArray) ea = nbd_extent_array_new(nb_extents);
2267
2268 bitmap_to_extents(bitmap, offset, length, ea);
2269
2270 return nbd_co_send_extents(client, handle, ea, last, context_id, errp);
2271 }
2272
2273 /* nbd_co_receive_request
2274 * Collect a client request. Return 0 if request looks valid, -EIO to drop
2275 * connection right away, -EAGAIN to indicate we were interrupted and the
2276 * channel should be quiesced, and any other negative value to report an error
2277 * to the client (although the caller may still need to disconnect after
2278 * reporting the error).
2279 */
2280 static int nbd_co_receive_request(NBDRequestData *req, NBDRequest *request,
2281 Error **errp)
2282 {
2283 NBDClient *client = req->client;
2284 int valid_flags;
2285 int ret;
2286
2287 g_assert(qemu_in_coroutine());
2288 assert(client->recv_coroutine == qemu_coroutine_self());
2289 ret = nbd_receive_request(client, request, errp);
2290 if (ret < 0) {
2291 return ret;
2292 }
2293
2294 trace_nbd_co_receive_request_decode_type(request->handle, request->type,
2295 nbd_cmd_lookup(request->type));
2296
2297 if (request->type != NBD_CMD_WRITE) {
2298 /* No payload, we are ready to read the next request. */
2299 req->complete = true;
2300 }
2301
2302 if (request->type == NBD_CMD_DISC) {
2303 /* Special case: we're going to disconnect without a reply,
2304 * whether or not flags, from, or len are bogus */
2305 return -EIO;
2306 }
2307
2308 if (request->type == NBD_CMD_READ || request->type == NBD_CMD_WRITE ||
2309 request->type == NBD_CMD_CACHE)
2310 {
2311 if (request->len > NBD_MAX_BUFFER_SIZE) {
2312 error_setg(errp, "len (%" PRIu32" ) is larger than max len (%u)",
2313 request->len, NBD_MAX_BUFFER_SIZE);
2314 return -EINVAL;
2315 }
2316
2317 if (request->type != NBD_CMD_CACHE) {
2318 req->data = blk_try_blockalign(client->exp->common.blk,
2319 request->len);
2320 if (req->data == NULL) {
2321 error_setg(errp, "No memory");
2322 return -ENOMEM;
2323 }
2324 }
2325 }
2326
2327 if (request->type == NBD_CMD_WRITE) {
2328 if (nbd_read(client->ioc, req->data, request->len, "CMD_WRITE data",
2329 errp) < 0)
2330 {
2331 return -EIO;
2332 }
2333 req->complete = true;
2334
2335 trace_nbd_co_receive_request_payload_received(request->handle,
2336 request->len);
2337 }
2338
2339 /* Sanity checks. */
2340 if (client->exp->nbdflags & NBD_FLAG_READ_ONLY &&
2341 (request->type == NBD_CMD_WRITE ||
2342 request->type == NBD_CMD_WRITE_ZEROES ||
2343 request->type == NBD_CMD_TRIM)) {
2344 error_setg(errp, "Export is read-only");
2345 return -EROFS;
2346 }
2347 if (request->from > client->exp->size ||
2348 request->len > client->exp->size - request->from) {
2349 error_setg(errp, "operation past EOF; From: %" PRIu64 ", Len: %" PRIu32
2350 ", Size: %" PRIu64, request->from, request->len,
2351 client->exp->size);
2352 return (request->type == NBD_CMD_WRITE ||
2353 request->type == NBD_CMD_WRITE_ZEROES) ? -ENOSPC : -EINVAL;
2354 }
2355 if (client->check_align && !QEMU_IS_ALIGNED(request->from | request->len,
2356 client->check_align)) {
2357 /*
2358 * The block layer gracefully handles unaligned requests, but
2359 * it's still worth tracing client non-compliance
2360 */
2361 trace_nbd_co_receive_align_compliance(nbd_cmd_lookup(request->type),
2362 request->from,
2363 request->len,
2364 client->check_align);
2365 }
2366 valid_flags = NBD_CMD_FLAG_FUA;
2367 if (request->type == NBD_CMD_READ && client->structured_reply) {
2368 valid_flags |= NBD_CMD_FLAG_DF;
2369 } else if (request->type == NBD_CMD_WRITE_ZEROES) {
2370 valid_flags |= NBD_CMD_FLAG_NO_HOLE | NBD_CMD_FLAG_FAST_ZERO;
2371 } else if (request->type == NBD_CMD_BLOCK_STATUS) {
2372 valid_flags |= NBD_CMD_FLAG_REQ_ONE;
2373 }
2374 if (request->flags & ~valid_flags) {
2375 error_setg(errp, "unsupported flags for command %s (got 0x%x)",
2376 nbd_cmd_lookup(request->type), request->flags);
2377 return -EINVAL;
2378 }
2379
2380 return 0;
2381 }
2382
2383 /* Send simple reply without a payload, or a structured error
2384 * @error_msg is ignored if @ret >= 0
2385 * Returns 0 if connection is still live, -errno on failure to talk to client
2386 */
2387 static coroutine_fn int nbd_send_generic_reply(NBDClient *client,
2388 uint64_t handle,
2389 int ret,
2390 const char *error_msg,
2391 Error **errp)
2392 {
2393 if (client->structured_reply && ret < 0) {
2394 return nbd_co_send_structured_error(client, handle, -ret, error_msg,
2395 errp);
2396 } else {
2397 return nbd_co_send_simple_reply(client, handle, ret < 0 ? -ret : 0,
2398 NULL, 0, errp);
2399 }
2400 }
2401
2402 /* Handle NBD_CMD_READ request.
2403 * Return -errno if sending fails. Other errors are reported directly to the
2404 * client as an error reply. */
2405 static coroutine_fn int nbd_do_cmd_read(NBDClient *client, NBDRequest *request,
2406 uint8_t *data, Error **errp)
2407 {
2408 int ret;
2409 NBDExport *exp = client->exp;
2410
2411 assert(request->type == NBD_CMD_READ);
2412
2413 /* XXX: NBD Protocol only documents use of FUA with WRITE */
2414 if (request->flags & NBD_CMD_FLAG_FUA) {
2415 ret = blk_co_flush(exp->common.blk);
2416 if (ret < 0) {
2417 return nbd_send_generic_reply(client, request->handle, ret,
2418 "flush failed", errp);
2419 }
2420 }
2421
2422 if (client->structured_reply && !(request->flags & NBD_CMD_FLAG_DF) &&
2423 request->len)
2424 {
2425 return nbd_co_send_sparse_read(client, request->handle, request->from,
2426 data, request->len, errp);
2427 }
2428
2429 ret = blk_pread(exp->common.blk, request->from, data, request->len);
2430 if (ret < 0) {
2431 return nbd_send_generic_reply(client, request->handle, ret,
2432 "reading from file failed", errp);
2433 }
2434
2435 if (client->structured_reply) {
2436 if (request->len) {
2437 return nbd_co_send_structured_read(client, request->handle,
2438 request->from, data,
2439 request->len, true, errp);
2440 } else {
2441 return nbd_co_send_structured_done(client, request->handle, errp);
2442 }
2443 } else {
2444 return nbd_co_send_simple_reply(client, request->handle, 0,
2445 data, request->len, errp);
2446 }
2447 }
2448
2449 /*
2450 * nbd_do_cmd_cache
2451 *
2452 * Handle NBD_CMD_CACHE request.
2453 * Return -errno if sending fails. Other errors are reported directly to the
2454 * client as an error reply.
2455 */
2456 static coroutine_fn int nbd_do_cmd_cache(NBDClient *client, NBDRequest *request,
2457 Error **errp)
2458 {
2459 int ret;
2460 NBDExport *exp = client->exp;
2461
2462 assert(request->type == NBD_CMD_CACHE);
2463
2464 ret = blk_co_preadv(exp->common.blk, request->from, request->len,
2465 NULL, BDRV_REQ_COPY_ON_READ | BDRV_REQ_PREFETCH);
2466
2467 return nbd_send_generic_reply(client, request->handle, ret,
2468 "caching data failed", errp);
2469 }
2470
2471 /* Handle NBD request.
2472 * Return -errno if sending fails. Other errors are reported directly to the
2473 * client as an error reply. */
2474 static coroutine_fn int nbd_handle_request(NBDClient *client,
2475 NBDRequest *request,
2476 uint8_t *data, Error **errp)
2477 {
2478 int ret;
2479 int flags;
2480 NBDExport *exp = client->exp;
2481 char *msg;
2482 size_t i;
2483
2484 switch (request->type) {
2485 case NBD_CMD_CACHE:
2486 return nbd_do_cmd_cache(client, request, errp);
2487
2488 case NBD_CMD_READ:
2489 return nbd_do_cmd_read(client, request, data, errp);
2490
2491 case NBD_CMD_WRITE:
2492 flags = 0;
2493 if (request->flags & NBD_CMD_FLAG_FUA) {
2494 flags |= BDRV_REQ_FUA;
2495 }
2496 ret = blk_pwrite(exp->common.blk, request->from, data, request->len,
2497 flags);
2498 return nbd_send_generic_reply(client, request->handle, ret,
2499 "writing to file failed", errp);
2500
2501 case NBD_CMD_WRITE_ZEROES:
2502 flags = 0;
2503 if (request->flags & NBD_CMD_FLAG_FUA) {
2504 flags |= BDRV_REQ_FUA;
2505 }
2506 if (!(request->flags & NBD_CMD_FLAG_NO_HOLE)) {
2507 flags |= BDRV_REQ_MAY_UNMAP;
2508 }
2509 if (request->flags & NBD_CMD_FLAG_FAST_ZERO) {
2510 flags |= BDRV_REQ_NO_FALLBACK;
2511 }
2512 ret = blk_pwrite_zeroes(exp->common.blk, request->from, request->len,
2513 flags);
2514 return nbd_send_generic_reply(client, request->handle, ret,
2515 "writing to file failed", errp);
2516
2517 case NBD_CMD_DISC:
2518 /* unreachable, thanks to special case in nbd_co_receive_request() */
2519 abort();
2520
2521 case NBD_CMD_FLUSH:
2522 ret = blk_co_flush(exp->common.blk);
2523 return nbd_send_generic_reply(client, request->handle, ret,
2524 "flush failed", errp);
2525
2526 case NBD_CMD_TRIM:
2527 ret = blk_co_pdiscard(exp->common.blk, request->from, request->len);
2528 if (ret >= 0 && request->flags & NBD_CMD_FLAG_FUA) {
2529 ret = blk_co_flush(exp->common.blk);
2530 }
2531 return nbd_send_generic_reply(client, request->handle, ret,
2532 "discard failed", errp);
2533
2534 case NBD_CMD_BLOCK_STATUS:
2535 if (!request->len) {
2536 return nbd_send_generic_reply(client, request->handle, -EINVAL,
2537 "need non-zero length", errp);
2538 }
2539 if (client->export_meta.count) {
2540 bool dont_fragment = request->flags & NBD_CMD_FLAG_REQ_ONE;
2541 int contexts_remaining = client->export_meta.count;
2542
2543 if (client->export_meta.base_allocation) {
2544 ret = nbd_co_send_block_status(client, request->handle,
2545 blk_bs(exp->common.blk),
2546 request->from,
2547 request->len, dont_fragment,
2548 !--contexts_remaining,
2549 NBD_META_ID_BASE_ALLOCATION,
2550 errp);
2551 if (ret < 0) {
2552 return ret;
2553 }
2554 }
2555
2556 if (client->export_meta.allocation_depth) {
2557 ret = nbd_co_send_block_status(client, request->handle,
2558 blk_bs(exp->common.blk),
2559 request->from, request->len,
2560 dont_fragment,
2561 !--contexts_remaining,
2562 NBD_META_ID_ALLOCATION_DEPTH,
2563 errp);
2564 if (ret < 0) {
2565 return ret;
2566 }
2567 }
2568
2569 for (i = 0; i < client->exp->nr_export_bitmaps; i++) {
2570 if (!client->export_meta.bitmaps[i]) {
2571 continue;
2572 }
2573 ret = nbd_co_send_bitmap(client, request->handle,
2574 client->exp->export_bitmaps[i],
2575 request->from, request->len,
2576 dont_fragment, !--contexts_remaining,
2577 NBD_META_ID_DIRTY_BITMAP + i, errp);
2578 if (ret < 0) {
2579 return ret;
2580 }
2581 }
2582
2583 assert(!contexts_remaining);
2584
2585 return 0;
2586 } else {
2587 return nbd_send_generic_reply(client, request->handle, -EINVAL,
2588 "CMD_BLOCK_STATUS not negotiated",
2589 errp);
2590 }
2591
2592 default:
2593 msg = g_strdup_printf("invalid request type (%" PRIu32 ") received",
2594 request->type);
2595 ret = nbd_send_generic_reply(client, request->handle, -EINVAL, msg,
2596 errp);
2597 g_free(msg);
2598 return ret;
2599 }
2600 }
2601
2602 /* Owns a reference to the NBDClient passed as opaque. */
2603 static coroutine_fn void nbd_trip(void *opaque)
2604 {
2605 NBDClient *client = opaque;
2606 NBDRequestData *req;
2607 NBDRequest request = { 0 }; /* GCC thinks it can be used uninitialized */
2608 int ret;
2609 Error *local_err = NULL;
2610
2611 trace_nbd_trip();
2612 if (client->closing) {
2613 nbd_client_put(client);
2614 return;
2615 }
2616
2617 if (client->quiescing) {
2618 /*
2619 * We're switching between AIO contexts. Don't attempt to receive a new
2620 * request and kick the main context which may be waiting for us.
2621 */
2622 nbd_client_put(client);
2623 client->recv_coroutine = NULL;
2624 aio_wait_kick();
2625 return;
2626 }
2627
2628 req = nbd_request_get(client);
2629 ret = nbd_co_receive_request(req, &request, &local_err);
2630 client->recv_coroutine = NULL;
2631
2632 if (client->closing) {
2633 /*
2634 * The client may be closed when we are blocked in
2635 * nbd_co_receive_request()
2636 */
2637 goto done;
2638 }
2639
2640 if (ret == -EAGAIN) {
2641 assert(client->quiescing);
2642 goto done;
2643 }
2644
2645 nbd_client_receive_next_request(client);
2646 if (ret == -EIO) {
2647 goto disconnect;
2648 }
2649
2650 if (ret < 0) {
2651 /* It wans't -EIO, so, according to nbd_co_receive_request()
2652 * semantics, we should return the error to the client. */
2653 Error *export_err = local_err;
2654
2655 local_err = NULL;
2656 ret = nbd_send_generic_reply(client, request.handle, -EINVAL,
2657 error_get_pretty(export_err), &local_err);
2658 error_free(export_err);
2659 } else {
2660 ret = nbd_handle_request(client, &request, req->data, &local_err);
2661 }
2662 if (ret < 0) {
2663 error_prepend(&local_err, "Failed to send reply: ");
2664 goto disconnect;
2665 }
2666
2667 /* We must disconnect after NBD_CMD_WRITE if we did not
2668 * read the payload.
2669 */
2670 if (!req->complete) {
2671 error_setg(&local_err, "Request handling failed in intermediate state");
2672 goto disconnect;
2673 }
2674
2675 done:
2676 nbd_request_put(req);
2677 nbd_client_put(client);
2678 return;
2679
2680 disconnect:
2681 if (local_err) {
2682 error_reportf_err(local_err, "Disconnect client, due to: ");
2683 }
2684 nbd_request_put(req);
2685 client_close(client, true);
2686 nbd_client_put(client);
2687 }
2688
2689 static void nbd_client_receive_next_request(NBDClient *client)
2690 {
2691 if (!client->recv_coroutine && client->nb_requests < MAX_NBD_REQUESTS &&
2692 !client->quiescing) {
2693 nbd_client_get(client);
2694 client->recv_coroutine = qemu_coroutine_create(nbd_trip, client);
2695 aio_co_schedule(client->exp->common.ctx, client->recv_coroutine);
2696 }
2697 }
2698
2699 static coroutine_fn void nbd_co_client_start(void *opaque)
2700 {
2701 NBDClient *client = opaque;
2702 Error *local_err = NULL;
2703
2704 qemu_co_mutex_init(&client->send_lock);
2705
2706 if (nbd_negotiate(client, &local_err)) {
2707 if (local_err) {
2708 error_report_err(local_err);
2709 }
2710 client_close(client, false);
2711 return;
2712 }
2713
2714 nbd_client_receive_next_request(client);
2715 }
2716
2717 /*
2718 * Create a new client listener using the given channel @sioc.
2719 * Begin servicing it in a coroutine. When the connection closes, call
2720 * @close_fn with an indication of whether the client completed negotiation.
2721 */
2722 void nbd_client_new(QIOChannelSocket *sioc,
2723 QCryptoTLSCreds *tlscreds,
2724 const char *tlsauthz,
2725 void (*close_fn)(NBDClient *, bool))
2726 {
2727 NBDClient *client;
2728 Coroutine *co;
2729
2730 client = g_new0(NBDClient, 1);
2731 client->refcount = 1;
2732 client->tlscreds = tlscreds;
2733 if (tlscreds) {
2734 object_ref(OBJECT(client->tlscreds));
2735 }
2736 client->tlsauthz = g_strdup(tlsauthz);
2737 client->sioc = sioc;
2738 object_ref(OBJECT(client->sioc));
2739 client->ioc = QIO_CHANNEL(sioc);
2740 object_ref(OBJECT(client->ioc));
2741 client->close_fn = close_fn;
2742
2743 co = qemu_coroutine_create(nbd_co_client_start, client);
2744 qemu_coroutine_enter(co);
2745 }