xref: /qemu/block/rbd.c (revision 12b35405)
1 /*
2  * QEMU Block driver for RADOS (Ceph)
3  *
4  * Copyright (C) 2010-2011 Christian Brunner <chb@muc.de>,
5  *                         Josh Durgin <josh.durgin@dreamhost.com>
6  *
7  * This work is licensed under the terms of the GNU GPL, version 2.  See
8  * the COPYING file in the top-level directory.
9  *
10  * Contributions after 2012-01-13 are licensed under the terms of the
11  * GNU GPL, version 2 or (at your option) any later version.
12  */
13 
14 #include "qemu/osdep.h"
15 
16 #include <rbd/librbd.h>
17 #include "qapi/error.h"
18 #include "qemu/error-report.h"
19 #include "qemu/module.h"
20 #include "qemu/option.h"
21 #include "block/block_int.h"
22 #include "block/qdict.h"
23 #include "crypto/secret.h"
24 #include "qemu/cutils.h"
25 #include "sysemu/replay.h"
26 #include "qapi/qmp/qstring.h"
27 #include "qapi/qmp/qdict.h"
28 #include "qapi/qmp/qjson.h"
29 #include "qapi/qmp/qlist.h"
30 #include "qapi/qobject-input-visitor.h"
31 #include "qapi/qapi-visit-block-core.h"
32 
33 /*
34  * When specifying the image filename use:
35  *
36  * rbd:poolname/devicename[@snapshotname][:option1=value1[:option2=value2...]]
37  *
38  * poolname must be the name of an existing rados pool.
39  *
40  * devicename is the name of the rbd image.
41  *
42  * Each option given is used to configure rados, and may be any valid
43  * Ceph option, "id", or "conf".
44  *
45  * The "id" option indicates what user we should authenticate as to
46  * the Ceph cluster.  If it is excluded we will use the Ceph default
47  * (normally 'admin').
48  *
49  * The "conf" option specifies a Ceph configuration file to read.  If
50  * it is not specified, we will read from the default Ceph locations
51  * (e.g., /etc/ceph/ceph.conf).  To avoid reading _any_ configuration
52  * file, specify conf=/dev/null.
53  *
54  * Configuration values containing :, @, or = can be escaped with a
55  * leading "\".
56  */
57 
58 /* rbd_aio_discard added in 0.1.2 */
59 #if LIBRBD_VERSION_CODE >= LIBRBD_VERSION(0, 1, 2)
60 #define LIBRBD_SUPPORTS_DISCARD
61 #else
62 #undef LIBRBD_SUPPORTS_DISCARD
63 #endif
64 
65 #define OBJ_MAX_SIZE (1UL << OBJ_DEFAULT_OBJ_ORDER)
66 
67 #define RBD_MAX_SNAPS 100
68 
69 /* The LIBRBD_SUPPORTS_IOVEC is defined in librbd.h */
70 #ifdef LIBRBD_SUPPORTS_IOVEC
71 #define LIBRBD_USE_IOVEC 1
72 #else
73 #define LIBRBD_USE_IOVEC 0
74 #endif
75 
76 typedef enum {
77     RBD_AIO_READ,
78     RBD_AIO_WRITE,
79     RBD_AIO_DISCARD,
80     RBD_AIO_FLUSH
81 } RBDAIOCmd;
82 
83 typedef struct RBDAIOCB {
84     BlockAIOCB common;
85     int64_t ret;
86     QEMUIOVector *qiov;
87     char *bounce;
88     RBDAIOCmd cmd;
89     int error;
90     struct BDRVRBDState *s;
91 } RBDAIOCB;
92 
93 typedef struct RADOSCB {
94     RBDAIOCB *acb;
95     struct BDRVRBDState *s;
96     int64_t size;
97     char *buf;
98     int64_t ret;
99 } RADOSCB;
100 
101 typedef struct BDRVRBDState {
102     rados_t cluster;
103     rados_ioctx_t io_ctx;
104     rbd_image_t image;
105     char *image_name;
106     char *snap;
107     char *namespace;
108     uint64_t image_size;
109 } BDRVRBDState;
110 
111 static int qemu_rbd_connect(rados_t *cluster, rados_ioctx_t *io_ctx,
112                             BlockdevOptionsRbd *opts, bool cache,
113                             const char *keypairs, const char *secretid,
114                             Error **errp);
115 
116 static char *qemu_rbd_next_tok(char *src, char delim, char **p)
117 {
118     char *end;
119 
120     *p = NULL;
121 
122     for (end = src; *end; ++end) {
123         if (*end == delim) {
124             break;
125         }
126         if (*end == '\\' && end[1] != '\0') {
127             end++;
128         }
129     }
130     if (*end == delim) {
131         *p = end + 1;
132         *end = '\0';
133     }
134     return src;
135 }
136 
137 static void qemu_rbd_unescape(char *src)
138 {
139     char *p;
140 
141     for (p = src; *src; ++src, ++p) {
142         if (*src == '\\' && src[1] != '\0') {
143             src++;
144         }
145         *p = *src;
146     }
147     *p = '\0';
148 }
149 
150 static void qemu_rbd_parse_filename(const char *filename, QDict *options,
151                                     Error **errp)
152 {
153     const char *start;
154     char *p, *buf;
155     QList *keypairs = NULL;
156     char *found_str, *image_name;
157 
158     if (!strstart(filename, "rbd:", &start)) {
159         error_setg(errp, "File name must start with 'rbd:'");
160         return;
161     }
162 
163     buf = g_strdup(start);
164     p = buf;
165 
166     found_str = qemu_rbd_next_tok(p, '/', &p);
167     if (!p) {
168         error_setg(errp, "Pool name is required");
169         goto done;
170     }
171     qemu_rbd_unescape(found_str);
172     qdict_put_str(options, "pool", found_str);
173 
174     if (strchr(p, '@')) {
175         image_name = qemu_rbd_next_tok(p, '@', &p);
176 
177         found_str = qemu_rbd_next_tok(p, ':', &p);
178         qemu_rbd_unescape(found_str);
179         qdict_put_str(options, "snapshot", found_str);
180     } else {
181         image_name = qemu_rbd_next_tok(p, ':', &p);
182     }
183     /* Check for namespace in the image_name */
184     if (strchr(image_name, '/')) {
185         found_str = qemu_rbd_next_tok(image_name, '/', &image_name);
186         qemu_rbd_unescape(found_str);
187         qdict_put_str(options, "namespace", found_str);
188     } else {
189         qdict_put_str(options, "namespace", "");
190     }
191     qemu_rbd_unescape(image_name);
192     qdict_put_str(options, "image", image_name);
193     if (!p) {
194         goto done;
195     }
196 
197     /* The following are essentially all key/value pairs, and we treat
198      * 'id' and 'conf' a bit special.  Key/value pairs may be in any order. */
199     while (p) {
200         char *name, *value;
201         name = qemu_rbd_next_tok(p, '=', &p);
202         if (!p) {
203             error_setg(errp, "conf option %s has no value", name);
204             break;
205         }
206 
207         qemu_rbd_unescape(name);
208 
209         value = qemu_rbd_next_tok(p, ':', &p);
210         qemu_rbd_unescape(value);
211 
212         if (!strcmp(name, "conf")) {
213             qdict_put_str(options, "conf", value);
214         } else if (!strcmp(name, "id")) {
215             qdict_put_str(options, "user", value);
216         } else {
217             /*
218              * We pass these internally to qemu_rbd_set_keypairs(), so
219              * we can get away with the simpler list of [ "key1",
220              * "value1", "key2", "value2" ] rather than a raw dict
221              * { "key1": "value1", "key2": "value2" } where we can't
222              * guarantee order, or even a more correct but complex
223              * [ { "key1": "value1" }, { "key2": "value2" } ]
224              */
225             if (!keypairs) {
226                 keypairs = qlist_new();
227             }
228             qlist_append_str(keypairs, name);
229             qlist_append_str(keypairs, value);
230         }
231     }
232 
233     if (keypairs) {
234         qdict_put(options, "=keyvalue-pairs",
235                   qobject_to_json(QOBJECT(keypairs)));
236     }
237 
238 done:
239     g_free(buf);
240     qobject_unref(keypairs);
241     return;
242 }
243 
244 
245 static void qemu_rbd_refresh_limits(BlockDriverState *bs, Error **errp)
246 {
247     /* XXX Does RBD support AIO on less than 512-byte alignment? */
248     bs->bl.request_alignment = 512;
249 }
250 
251 
252 static int qemu_rbd_set_auth(rados_t cluster, BlockdevOptionsRbd *opts,
253                              Error **errp)
254 {
255     char *key, *acr;
256     int r;
257     GString *accu;
258     RbdAuthModeList *auth;
259 
260     if (opts->key_secret) {
261         key = qcrypto_secret_lookup_as_base64(opts->key_secret, errp);
262         if (!key) {
263             return -EIO;
264         }
265         r = rados_conf_set(cluster, "key", key);
266         g_free(key);
267         if (r < 0) {
268             error_setg_errno(errp, -r, "Could not set 'key'");
269             return r;
270         }
271     }
272 
273     if (opts->has_auth_client_required) {
274         accu = g_string_new("");
275         for (auth = opts->auth_client_required; auth; auth = auth->next) {
276             if (accu->str[0]) {
277                 g_string_append_c(accu, ';');
278             }
279             g_string_append(accu, RbdAuthMode_str(auth->value));
280         }
281         acr = g_string_free(accu, FALSE);
282         r = rados_conf_set(cluster, "auth_client_required", acr);
283         g_free(acr);
284         if (r < 0) {
285             error_setg_errno(errp, -r,
286                              "Could not set 'auth_client_required'");
287             return r;
288         }
289     }
290 
291     return 0;
292 }
293 
294 static int qemu_rbd_set_keypairs(rados_t cluster, const char *keypairs_json,
295                                  Error **errp)
296 {
297     QList *keypairs;
298     QString *name;
299     QString *value;
300     const char *key;
301     size_t remaining;
302     int ret = 0;
303 
304     if (!keypairs_json) {
305         return ret;
306     }
307     keypairs = qobject_to(QList,
308                           qobject_from_json(keypairs_json, &error_abort));
309     remaining = qlist_size(keypairs) / 2;
310     assert(remaining);
311 
312     while (remaining--) {
313         name = qobject_to(QString, qlist_pop(keypairs));
314         value = qobject_to(QString, qlist_pop(keypairs));
315         assert(name && value);
316         key = qstring_get_str(name);
317 
318         ret = rados_conf_set(cluster, key, qstring_get_str(value));
319         qobject_unref(value);
320         if (ret < 0) {
321             error_setg_errno(errp, -ret, "invalid conf option %s", key);
322             qobject_unref(name);
323             ret = -EINVAL;
324             break;
325         }
326         qobject_unref(name);
327     }
328 
329     qobject_unref(keypairs);
330     return ret;
331 }
332 
333 static void qemu_rbd_memset(RADOSCB *rcb, int64_t offs)
334 {
335     if (LIBRBD_USE_IOVEC) {
336         RBDAIOCB *acb = rcb->acb;
337         iov_memset(acb->qiov->iov, acb->qiov->niov, offs, 0,
338                    acb->qiov->size - offs);
339     } else {
340         memset(rcb->buf + offs, 0, rcb->size - offs);
341     }
342 }
343 
344 static QemuOptsList runtime_opts = {
345     .name = "rbd",
346     .head = QTAILQ_HEAD_INITIALIZER(runtime_opts.head),
347     .desc = {
348         {
349             .name = "pool",
350             .type = QEMU_OPT_STRING,
351             .help = "Rados pool name",
352         },
353         {
354             .name = "namespace",
355             .type = QEMU_OPT_STRING,
356             .help = "Rados namespace name in the pool",
357         },
358         {
359             .name = "image",
360             .type = QEMU_OPT_STRING,
361             .help = "Image name in the pool",
362         },
363         {
364             .name = "conf",
365             .type = QEMU_OPT_STRING,
366             .help = "Rados config file location",
367         },
368         {
369             .name = "snapshot",
370             .type = QEMU_OPT_STRING,
371             .help = "Ceph snapshot name",
372         },
373         {
374             /* maps to 'id' in rados_create() */
375             .name = "user",
376             .type = QEMU_OPT_STRING,
377             .help = "Rados id name",
378         },
379         /*
380          * server.* extracted manually, see qemu_rbd_mon_host()
381          */
382         { /* end of list */ }
383     },
384 };
385 
386 /* FIXME Deprecate and remove keypairs or make it available in QMP. */
387 static int qemu_rbd_do_create(BlockdevCreateOptions *options,
388                               const char *keypairs, const char *password_secret,
389                               Error **errp)
390 {
391     BlockdevCreateOptionsRbd *opts = &options->u.rbd;
392     rados_t cluster;
393     rados_ioctx_t io_ctx;
394     int obj_order = 0;
395     int ret;
396 
397     assert(options->driver == BLOCKDEV_DRIVER_RBD);
398     if (opts->location->has_snapshot) {
399         error_setg(errp, "Can't use snapshot name for image creation");
400         return -EINVAL;
401     }
402 
403     if (opts->has_cluster_size) {
404         int64_t objsize = opts->cluster_size;
405         if ((objsize - 1) & objsize) {    /* not a power of 2? */
406             error_setg(errp, "obj size needs to be power of 2");
407             return -EINVAL;
408         }
409         if (objsize < 4096) {
410             error_setg(errp, "obj size too small");
411             return -EINVAL;
412         }
413         obj_order = ctz32(objsize);
414     }
415 
416     ret = qemu_rbd_connect(&cluster, &io_ctx, opts->location, false, keypairs,
417                            password_secret, errp);
418     if (ret < 0) {
419         return ret;
420     }
421 
422     ret = rbd_create(io_ctx, opts->location->image, opts->size, &obj_order);
423     if (ret < 0) {
424         error_setg_errno(errp, -ret, "error rbd create");
425         goto out;
426     }
427 
428     ret = 0;
429 out:
430     rados_ioctx_destroy(io_ctx);
431     rados_shutdown(cluster);
432     return ret;
433 }
434 
435 static int qemu_rbd_co_create(BlockdevCreateOptions *options, Error **errp)
436 {
437     return qemu_rbd_do_create(options, NULL, NULL, errp);
438 }
439 
440 static int coroutine_fn qemu_rbd_co_create_opts(BlockDriver *drv,
441                                                 const char *filename,
442                                                 QemuOpts *opts,
443                                                 Error **errp)
444 {
445     BlockdevCreateOptions *create_options;
446     BlockdevCreateOptionsRbd *rbd_opts;
447     BlockdevOptionsRbd *loc;
448     Error *local_err = NULL;
449     const char *keypairs, *password_secret;
450     QDict *options = NULL;
451     int ret = 0;
452 
453     create_options = g_new0(BlockdevCreateOptions, 1);
454     create_options->driver = BLOCKDEV_DRIVER_RBD;
455     rbd_opts = &create_options->u.rbd;
456 
457     rbd_opts->location = g_new0(BlockdevOptionsRbd, 1);
458 
459     password_secret = qemu_opt_get(opts, "password-secret");
460 
461     /* Read out options */
462     rbd_opts->size = ROUND_UP(qemu_opt_get_size_del(opts, BLOCK_OPT_SIZE, 0),
463                               BDRV_SECTOR_SIZE);
464     rbd_opts->cluster_size = qemu_opt_get_size_del(opts,
465                                                    BLOCK_OPT_CLUSTER_SIZE, 0);
466     rbd_opts->has_cluster_size = (rbd_opts->cluster_size != 0);
467 
468     options = qdict_new();
469     qemu_rbd_parse_filename(filename, options, &local_err);
470     if (local_err) {
471         ret = -EINVAL;
472         error_propagate(errp, local_err);
473         goto exit;
474     }
475 
476     /*
477      * Caution: while qdict_get_try_str() is fine, getting non-string
478      * types would require more care.  When @options come from -blockdev
479      * or blockdev_add, its members are typed according to the QAPI
480      * schema, but when they come from -drive, they're all QString.
481      */
482     loc = rbd_opts->location;
483     loc->pool        = g_strdup(qdict_get_try_str(options, "pool"));
484     loc->conf        = g_strdup(qdict_get_try_str(options, "conf"));
485     loc->has_conf    = !!loc->conf;
486     loc->user        = g_strdup(qdict_get_try_str(options, "user"));
487     loc->has_user    = !!loc->user;
488     loc->q_namespace = g_strdup(qdict_get_try_str(options, "namespace"));
489     loc->image       = g_strdup(qdict_get_try_str(options, "image"));
490     keypairs         = qdict_get_try_str(options, "=keyvalue-pairs");
491 
492     ret = qemu_rbd_do_create(create_options, keypairs, password_secret, errp);
493     if (ret < 0) {
494         goto exit;
495     }
496 
497 exit:
498     qobject_unref(options);
499     qapi_free_BlockdevCreateOptions(create_options);
500     return ret;
501 }
502 
503 /*
504  * This aio completion is being called from rbd_finish_bh() and runs in qemu
505  * BH context.
506  */
507 static void qemu_rbd_complete_aio(RADOSCB *rcb)
508 {
509     RBDAIOCB *acb = rcb->acb;
510     int64_t r;
511 
512     r = rcb->ret;
513 
514     if (acb->cmd != RBD_AIO_READ) {
515         if (r < 0) {
516             acb->ret = r;
517             acb->error = 1;
518         } else if (!acb->error) {
519             acb->ret = rcb->size;
520         }
521     } else {
522         if (r < 0) {
523             qemu_rbd_memset(rcb, 0);
524             acb->ret = r;
525             acb->error = 1;
526         } else if (r < rcb->size) {
527             qemu_rbd_memset(rcb, r);
528             if (!acb->error) {
529                 acb->ret = rcb->size;
530             }
531         } else if (!acb->error) {
532             acb->ret = r;
533         }
534     }
535 
536     g_free(rcb);
537 
538     if (!LIBRBD_USE_IOVEC) {
539         if (acb->cmd == RBD_AIO_READ) {
540             qemu_iovec_from_buf(acb->qiov, 0, acb->bounce, acb->qiov->size);
541         }
542         qemu_vfree(acb->bounce);
543     }
544 
545     acb->common.cb(acb->common.opaque, (acb->ret > 0 ? 0 : acb->ret));
546 
547     qemu_aio_unref(acb);
548 }
549 
550 static char *qemu_rbd_mon_host(BlockdevOptionsRbd *opts, Error **errp)
551 {
552     const char **vals;
553     const char *host, *port;
554     char *rados_str;
555     InetSocketAddressBaseList *p;
556     int i, cnt;
557 
558     if (!opts->has_server) {
559         return NULL;
560     }
561 
562     for (cnt = 0, p = opts->server; p; p = p->next) {
563         cnt++;
564     }
565 
566     vals = g_new(const char *, cnt + 1);
567 
568     for (i = 0, p = opts->server; p; p = p->next, i++) {
569         host = p->value->host;
570         port = p->value->port;
571 
572         if (strchr(host, ':')) {
573             vals[i] = g_strdup_printf("[%s]:%s", host, port);
574         } else {
575             vals[i] = g_strdup_printf("%s:%s", host, port);
576         }
577     }
578     vals[i] = NULL;
579 
580     rados_str = i ? g_strjoinv(";", (char **)vals) : NULL;
581     g_strfreev((char **)vals);
582     return rados_str;
583 }
584 
585 static int qemu_rbd_connect(rados_t *cluster, rados_ioctx_t *io_ctx,
586                             BlockdevOptionsRbd *opts, bool cache,
587                             const char *keypairs, const char *secretid,
588                             Error **errp)
589 {
590     char *mon_host = NULL;
591     Error *local_err = NULL;
592     int r;
593 
594     if (secretid) {
595         if (opts->key_secret) {
596             error_setg(errp,
597                        "Legacy 'password-secret' clashes with 'key-secret'");
598             return -EINVAL;
599         }
600         opts->key_secret = g_strdup(secretid);
601         opts->has_key_secret = true;
602     }
603 
604     mon_host = qemu_rbd_mon_host(opts, &local_err);
605     if (local_err) {
606         error_propagate(errp, local_err);
607         r = -EINVAL;
608         goto failed_opts;
609     }
610 
611     r = rados_create(cluster, opts->user);
612     if (r < 0) {
613         error_setg_errno(errp, -r, "error initializing");
614         goto failed_opts;
615     }
616 
617     /* try default location when conf=NULL, but ignore failure */
618     r = rados_conf_read_file(*cluster, opts->conf);
619     if (opts->has_conf && r < 0) {
620         error_setg_errno(errp, -r, "error reading conf file %s", opts->conf);
621         goto failed_shutdown;
622     }
623 
624     r = qemu_rbd_set_keypairs(*cluster, keypairs, errp);
625     if (r < 0) {
626         goto failed_shutdown;
627     }
628 
629     if (mon_host) {
630         r = rados_conf_set(*cluster, "mon_host", mon_host);
631         if (r < 0) {
632             goto failed_shutdown;
633         }
634     }
635 
636     r = qemu_rbd_set_auth(*cluster, opts, errp);
637     if (r < 0) {
638         goto failed_shutdown;
639     }
640 
641     /*
642      * Fallback to more conservative semantics if setting cache
643      * options fails. Ignore errors from setting rbd_cache because the
644      * only possible error is that the option does not exist, and
645      * librbd defaults to no caching. If write through caching cannot
646      * be set up, fall back to no caching.
647      */
648     if (cache) {
649         rados_conf_set(*cluster, "rbd_cache", "true");
650     } else {
651         rados_conf_set(*cluster, "rbd_cache", "false");
652     }
653 
654     r = rados_connect(*cluster);
655     if (r < 0) {
656         error_setg_errno(errp, -r, "error connecting");
657         goto failed_shutdown;
658     }
659 
660     r = rados_ioctx_create(*cluster, opts->pool, io_ctx);
661     if (r < 0) {
662         error_setg_errno(errp, -r, "error opening pool %s", opts->pool);
663         goto failed_shutdown;
664     }
665     /*
666      * Set the namespace after opening the io context on the pool,
667      * if nspace == NULL or if nspace == "", it is just as we did nothing
668      */
669     rados_ioctx_set_namespace(*io_ctx, opts->q_namespace);
670 
671     return 0;
672 
673 failed_shutdown:
674     rados_shutdown(*cluster);
675 failed_opts:
676     g_free(mon_host);
677     return r;
678 }
679 
680 static int qemu_rbd_convert_options(QDict *options, BlockdevOptionsRbd **opts,
681                                     Error **errp)
682 {
683     Visitor *v;
684 
685     /* Convert the remaining options into a QAPI object */
686     v = qobject_input_visitor_new_flat_confused(options, errp);
687     if (!v) {
688         return -EINVAL;
689     }
690 
691     visit_type_BlockdevOptionsRbd(v, NULL, opts, errp);
692     visit_free(v);
693     if (!opts) {
694         return -EINVAL;
695     }
696 
697     return 0;
698 }
699 
700 static int qemu_rbd_attempt_legacy_options(QDict *options,
701                                            BlockdevOptionsRbd **opts,
702                                            char **keypairs)
703 {
704     char *filename;
705     int r;
706 
707     filename = g_strdup(qdict_get_try_str(options, "filename"));
708     if (!filename) {
709         return -EINVAL;
710     }
711     qdict_del(options, "filename");
712 
713     qemu_rbd_parse_filename(filename, options, NULL);
714 
715     /* keypairs freed by caller */
716     *keypairs = g_strdup(qdict_get_try_str(options, "=keyvalue-pairs"));
717     if (*keypairs) {
718         qdict_del(options, "=keyvalue-pairs");
719     }
720 
721     r = qemu_rbd_convert_options(options, opts, NULL);
722 
723     g_free(filename);
724     return r;
725 }
726 
727 static int qemu_rbd_open(BlockDriverState *bs, QDict *options, int flags,
728                          Error **errp)
729 {
730     BDRVRBDState *s = bs->opaque;
731     BlockdevOptionsRbd *opts = NULL;
732     const QDictEntry *e;
733     Error *local_err = NULL;
734     char *keypairs, *secretid;
735     int r;
736 
737     keypairs = g_strdup(qdict_get_try_str(options, "=keyvalue-pairs"));
738     if (keypairs) {
739         qdict_del(options, "=keyvalue-pairs");
740     }
741 
742     secretid = g_strdup(qdict_get_try_str(options, "password-secret"));
743     if (secretid) {
744         qdict_del(options, "password-secret");
745     }
746 
747     r = qemu_rbd_convert_options(options, &opts, &local_err);
748     if (local_err) {
749         /* If keypairs are present, that means some options are present in
750          * the modern option format.  Don't attempt to parse legacy option
751          * formats, as we won't support mixed usage. */
752         if (keypairs) {
753             error_propagate(errp, local_err);
754             goto out;
755         }
756 
757         /* If the initial attempt to convert and process the options failed,
758          * we may be attempting to open an image file that has the rbd options
759          * specified in the older format consisting of all key/value pairs
760          * encoded in the filename.  Go ahead and attempt to parse the
761          * filename, and see if we can pull out the required options. */
762         r = qemu_rbd_attempt_legacy_options(options, &opts, &keypairs);
763         if (r < 0) {
764             /* Propagate the original error, not the legacy parsing fallback
765              * error, as the latter was just a best-effort attempt. */
766             error_propagate(errp, local_err);
767             goto out;
768         }
769         /* Take care whenever deciding to actually deprecate; once this ability
770          * is removed, we will not be able to open any images with legacy-styled
771          * backing image strings. */
772         warn_report("RBD options encoded in the filename as keyvalue pairs "
773                     "is deprecated");
774     }
775 
776     /* Remove the processed options from the QDict (the visitor processes
777      * _all_ options in the QDict) */
778     while ((e = qdict_first(options))) {
779         qdict_del(options, e->key);
780     }
781 
782     r = qemu_rbd_connect(&s->cluster, &s->io_ctx, opts,
783                          !(flags & BDRV_O_NOCACHE), keypairs, secretid, errp);
784     if (r < 0) {
785         goto out;
786     }
787 
788     s->snap = g_strdup(opts->snapshot);
789     s->image_name = g_strdup(opts->image);
790 
791     /* rbd_open is always r/w */
792     r = rbd_open(s->io_ctx, s->image_name, &s->image, s->snap);
793     if (r < 0) {
794         error_setg_errno(errp, -r, "error reading header from %s",
795                          s->image_name);
796         goto failed_open;
797     }
798 
799     r = rbd_get_size(s->image, &s->image_size);
800     if (r < 0) {
801         error_setg_errno(errp, -r, "error getting image size from %s",
802                          s->image_name);
803         rbd_close(s->image);
804         goto failed_open;
805     }
806 
807     /* If we are using an rbd snapshot, we must be r/o, otherwise
808      * leave as-is */
809     if (s->snap != NULL) {
810         r = bdrv_apply_auto_read_only(bs, "rbd snapshots are read-only", errp);
811         if (r < 0) {
812             rbd_close(s->image);
813             goto failed_open;
814         }
815     }
816 
817     /* When extending regular files, we get zeros from the OS */
818     bs->supported_truncate_flags = BDRV_REQ_ZERO_WRITE;
819 
820     r = 0;
821     goto out;
822 
823 failed_open:
824     rados_ioctx_destroy(s->io_ctx);
825     g_free(s->snap);
826     g_free(s->image_name);
827     rados_shutdown(s->cluster);
828 out:
829     qapi_free_BlockdevOptionsRbd(opts);
830     g_free(keypairs);
831     g_free(secretid);
832     return r;
833 }
834 
835 
836 /* Since RBD is currently always opened R/W via the API,
837  * we just need to check if we are using a snapshot or not, in
838  * order to determine if we will allow it to be R/W */
839 static int qemu_rbd_reopen_prepare(BDRVReopenState *state,
840                                    BlockReopenQueue *queue, Error **errp)
841 {
842     BDRVRBDState *s = state->bs->opaque;
843     int ret = 0;
844 
845     if (s->snap && state->flags & BDRV_O_RDWR) {
846         error_setg(errp,
847                    "Cannot change node '%s' to r/w when using RBD snapshot",
848                    bdrv_get_device_or_node_name(state->bs));
849         ret = -EINVAL;
850     }
851 
852     return ret;
853 }
854 
855 static void qemu_rbd_close(BlockDriverState *bs)
856 {
857     BDRVRBDState *s = bs->opaque;
858 
859     rbd_close(s->image);
860     rados_ioctx_destroy(s->io_ctx);
861     g_free(s->snap);
862     g_free(s->image_name);
863     rados_shutdown(s->cluster);
864 }
865 
866 /* Resize the RBD image and update the 'image_size' with the current size */
867 static int qemu_rbd_resize(BlockDriverState *bs, uint64_t size)
868 {
869     BDRVRBDState *s = bs->opaque;
870     int r;
871 
872     r = rbd_resize(s->image, size);
873     if (r < 0) {
874         return r;
875     }
876 
877     s->image_size = size;
878 
879     return 0;
880 }
881 
882 static const AIOCBInfo rbd_aiocb_info = {
883     .aiocb_size = sizeof(RBDAIOCB),
884 };
885 
886 static void rbd_finish_bh(void *opaque)
887 {
888     RADOSCB *rcb = opaque;
889     qemu_rbd_complete_aio(rcb);
890 }
891 
892 /*
893  * This is the callback function for rbd_aio_read and _write
894  *
895  * Note: this function is being called from a non qemu thread so
896  * we need to be careful about what we do here. Generally we only
897  * schedule a BH, and do the rest of the io completion handling
898  * from rbd_finish_bh() which runs in a qemu context.
899  */
900 static void rbd_finish_aiocb(rbd_completion_t c, RADOSCB *rcb)
901 {
902     RBDAIOCB *acb = rcb->acb;
903 
904     rcb->ret = rbd_aio_get_return_value(c);
905     rbd_aio_release(c);
906 
907     replay_bh_schedule_oneshot_event(bdrv_get_aio_context(acb->common.bs),
908                                      rbd_finish_bh, rcb);
909 }
910 
911 static int rbd_aio_discard_wrapper(rbd_image_t image,
912                                    uint64_t off,
913                                    uint64_t len,
914                                    rbd_completion_t comp)
915 {
916 #ifdef LIBRBD_SUPPORTS_DISCARD
917     return rbd_aio_discard(image, off, len, comp);
918 #else
919     return -ENOTSUP;
920 #endif
921 }
922 
923 static int rbd_aio_flush_wrapper(rbd_image_t image,
924                                  rbd_completion_t comp)
925 {
926 #ifdef LIBRBD_SUPPORTS_AIO_FLUSH
927     return rbd_aio_flush(image, comp);
928 #else
929     return -ENOTSUP;
930 #endif
931 }
932 
933 static BlockAIOCB *rbd_start_aio(BlockDriverState *bs,
934                                  int64_t off,
935                                  QEMUIOVector *qiov,
936                                  int64_t size,
937                                  BlockCompletionFunc *cb,
938                                  void *opaque,
939                                  RBDAIOCmd cmd)
940 {
941     RBDAIOCB *acb;
942     RADOSCB *rcb = NULL;
943     rbd_completion_t c;
944     int r;
945 
946     BDRVRBDState *s = bs->opaque;
947 
948     acb = qemu_aio_get(&rbd_aiocb_info, bs, cb, opaque);
949     acb->cmd = cmd;
950     acb->qiov = qiov;
951     assert(!qiov || qiov->size == size);
952 
953     rcb = g_new(RADOSCB, 1);
954 
955     if (!LIBRBD_USE_IOVEC) {
956         if (cmd == RBD_AIO_DISCARD || cmd == RBD_AIO_FLUSH) {
957             acb->bounce = NULL;
958         } else {
959             acb->bounce = qemu_try_blockalign(bs, qiov->size);
960             if (acb->bounce == NULL) {
961                 goto failed;
962             }
963         }
964         if (cmd == RBD_AIO_WRITE) {
965             qemu_iovec_to_buf(acb->qiov, 0, acb->bounce, qiov->size);
966         }
967         rcb->buf = acb->bounce;
968     }
969 
970     acb->ret = 0;
971     acb->error = 0;
972     acb->s = s;
973 
974     rcb->acb = acb;
975     rcb->s = acb->s;
976     rcb->size = size;
977     r = rbd_aio_create_completion(rcb, (rbd_callback_t) rbd_finish_aiocb, &c);
978     if (r < 0) {
979         goto failed;
980     }
981 
982     switch (cmd) {
983     case RBD_AIO_WRITE: {
984         /*
985          * RBD APIs don't allow us to write more than actual size, so in order
986          * to support growing images, we resize the image before write
987          * operations that exceed the current size.
988          */
989         if (off + size > s->image_size) {
990             r = qemu_rbd_resize(bs, off + size);
991             if (r < 0) {
992                 goto failed_completion;
993             }
994         }
995 #ifdef LIBRBD_SUPPORTS_IOVEC
996             r = rbd_aio_writev(s->image, qiov->iov, qiov->niov, off, c);
997 #else
998             r = rbd_aio_write(s->image, off, size, rcb->buf, c);
999 #endif
1000         break;
1001     }
1002     case RBD_AIO_READ:
1003 #ifdef LIBRBD_SUPPORTS_IOVEC
1004             r = rbd_aio_readv(s->image, qiov->iov, qiov->niov, off, c);
1005 #else
1006             r = rbd_aio_read(s->image, off, size, rcb->buf, c);
1007 #endif
1008         break;
1009     case RBD_AIO_DISCARD:
1010         r = rbd_aio_discard_wrapper(s->image, off, size, c);
1011         break;
1012     case RBD_AIO_FLUSH:
1013         r = rbd_aio_flush_wrapper(s->image, c);
1014         break;
1015     default:
1016         r = -EINVAL;
1017     }
1018 
1019     if (r < 0) {
1020         goto failed_completion;
1021     }
1022     return &acb->common;
1023 
1024 failed_completion:
1025     rbd_aio_release(c);
1026 failed:
1027     g_free(rcb);
1028     if (!LIBRBD_USE_IOVEC) {
1029         qemu_vfree(acb->bounce);
1030     }
1031 
1032     qemu_aio_unref(acb);
1033     return NULL;
1034 }
1035 
1036 static BlockAIOCB *qemu_rbd_aio_preadv(BlockDriverState *bs,
1037                                        uint64_t offset, uint64_t bytes,
1038                                        QEMUIOVector *qiov, int flags,
1039                                        BlockCompletionFunc *cb,
1040                                        void *opaque)
1041 {
1042     return rbd_start_aio(bs, offset, qiov, bytes, cb, opaque,
1043                          RBD_AIO_READ);
1044 }
1045 
1046 static BlockAIOCB *qemu_rbd_aio_pwritev(BlockDriverState *bs,
1047                                         uint64_t offset, uint64_t bytes,
1048                                         QEMUIOVector *qiov, int flags,
1049                                         BlockCompletionFunc *cb,
1050                                         void *opaque)
1051 {
1052     return rbd_start_aio(bs, offset, qiov, bytes, cb, opaque,
1053                          RBD_AIO_WRITE);
1054 }
1055 
1056 #ifdef LIBRBD_SUPPORTS_AIO_FLUSH
1057 static BlockAIOCB *qemu_rbd_aio_flush(BlockDriverState *bs,
1058                                       BlockCompletionFunc *cb,
1059                                       void *opaque)
1060 {
1061     return rbd_start_aio(bs, 0, NULL, 0, cb, opaque, RBD_AIO_FLUSH);
1062 }
1063 
1064 #else
1065 
1066 static int qemu_rbd_co_flush(BlockDriverState *bs)
1067 {
1068 #if LIBRBD_VERSION_CODE >= LIBRBD_VERSION(0, 1, 1)
1069     /* rbd_flush added in 0.1.1 */
1070     BDRVRBDState *s = bs->opaque;
1071     return rbd_flush(s->image);
1072 #else
1073     return 0;
1074 #endif
1075 }
1076 #endif
1077 
1078 static int qemu_rbd_getinfo(BlockDriverState *bs, BlockDriverInfo *bdi)
1079 {
1080     BDRVRBDState *s = bs->opaque;
1081     rbd_image_info_t info;
1082     int r;
1083 
1084     r = rbd_stat(s->image, &info, sizeof(info));
1085     if (r < 0) {
1086         return r;
1087     }
1088 
1089     bdi->cluster_size = info.obj_size;
1090     return 0;
1091 }
1092 
1093 static int64_t qemu_rbd_getlength(BlockDriverState *bs)
1094 {
1095     BDRVRBDState *s = bs->opaque;
1096     rbd_image_info_t info;
1097     int r;
1098 
1099     r = rbd_stat(s->image, &info, sizeof(info));
1100     if (r < 0) {
1101         return r;
1102     }
1103 
1104     return info.size;
1105 }
1106 
1107 static int coroutine_fn qemu_rbd_co_truncate(BlockDriverState *bs,
1108                                              int64_t offset,
1109                                              bool exact,
1110                                              PreallocMode prealloc,
1111                                              BdrvRequestFlags flags,
1112                                              Error **errp)
1113 {
1114     int r;
1115 
1116     if (prealloc != PREALLOC_MODE_OFF) {
1117         error_setg(errp, "Unsupported preallocation mode '%s'",
1118                    PreallocMode_str(prealloc));
1119         return -ENOTSUP;
1120     }
1121 
1122     r = qemu_rbd_resize(bs, offset);
1123     if (r < 0) {
1124         error_setg_errno(errp, -r, "Failed to resize file");
1125         return r;
1126     }
1127 
1128     return 0;
1129 }
1130 
1131 static int qemu_rbd_snap_create(BlockDriverState *bs,
1132                                 QEMUSnapshotInfo *sn_info)
1133 {
1134     BDRVRBDState *s = bs->opaque;
1135     int r;
1136 
1137     if (sn_info->name[0] == '\0') {
1138         return -EINVAL; /* we need a name for rbd snapshots */
1139     }
1140 
1141     /*
1142      * rbd snapshots are using the name as the user controlled unique identifier
1143      * we can't use the rbd snapid for that purpose, as it can't be set
1144      */
1145     if (sn_info->id_str[0] != '\0' &&
1146         strcmp(sn_info->id_str, sn_info->name) != 0) {
1147         return -EINVAL;
1148     }
1149 
1150     if (strlen(sn_info->name) >= sizeof(sn_info->id_str)) {
1151         return -ERANGE;
1152     }
1153 
1154     r = rbd_snap_create(s->image, sn_info->name);
1155     if (r < 0) {
1156         error_report("failed to create snap: %s", strerror(-r));
1157         return r;
1158     }
1159 
1160     return 0;
1161 }
1162 
1163 static int qemu_rbd_snap_remove(BlockDriverState *bs,
1164                                 const char *snapshot_id,
1165                                 const char *snapshot_name,
1166                                 Error **errp)
1167 {
1168     BDRVRBDState *s = bs->opaque;
1169     int r;
1170 
1171     if (!snapshot_name) {
1172         error_setg(errp, "rbd need a valid snapshot name");
1173         return -EINVAL;
1174     }
1175 
1176     /* If snapshot_id is specified, it must be equal to name, see
1177        qemu_rbd_snap_list() */
1178     if (snapshot_id && strcmp(snapshot_id, snapshot_name)) {
1179         error_setg(errp,
1180                    "rbd do not support snapshot id, it should be NULL or "
1181                    "equal to snapshot name");
1182         return -EINVAL;
1183     }
1184 
1185     r = rbd_snap_remove(s->image, snapshot_name);
1186     if (r < 0) {
1187         error_setg_errno(errp, -r, "Failed to remove the snapshot");
1188     }
1189     return r;
1190 }
1191 
1192 static int qemu_rbd_snap_rollback(BlockDriverState *bs,
1193                                   const char *snapshot_name)
1194 {
1195     BDRVRBDState *s = bs->opaque;
1196 
1197     return rbd_snap_rollback(s->image, snapshot_name);
1198 }
1199 
1200 static int qemu_rbd_snap_list(BlockDriverState *bs,
1201                               QEMUSnapshotInfo **psn_tab)
1202 {
1203     BDRVRBDState *s = bs->opaque;
1204     QEMUSnapshotInfo *sn_info, *sn_tab = NULL;
1205     int i, snap_count;
1206     rbd_snap_info_t *snaps;
1207     int max_snaps = RBD_MAX_SNAPS;
1208 
1209     do {
1210         snaps = g_new(rbd_snap_info_t, max_snaps);
1211         snap_count = rbd_snap_list(s->image, snaps, &max_snaps);
1212         if (snap_count <= 0) {
1213             g_free(snaps);
1214         }
1215     } while (snap_count == -ERANGE);
1216 
1217     if (snap_count <= 0) {
1218         goto done;
1219     }
1220 
1221     sn_tab = g_new0(QEMUSnapshotInfo, snap_count);
1222 
1223     for (i = 0; i < snap_count; i++) {
1224         const char *snap_name = snaps[i].name;
1225 
1226         sn_info = sn_tab + i;
1227         pstrcpy(sn_info->id_str, sizeof(sn_info->id_str), snap_name);
1228         pstrcpy(sn_info->name, sizeof(sn_info->name), snap_name);
1229 
1230         sn_info->vm_state_size = snaps[i].size;
1231         sn_info->date_sec = 0;
1232         sn_info->date_nsec = 0;
1233         sn_info->vm_clock_nsec = 0;
1234     }
1235     rbd_snap_list_end(snaps);
1236     g_free(snaps);
1237 
1238  done:
1239     *psn_tab = sn_tab;
1240     return snap_count;
1241 }
1242 
1243 #ifdef LIBRBD_SUPPORTS_DISCARD
1244 static BlockAIOCB *qemu_rbd_aio_pdiscard(BlockDriverState *bs,
1245                                          int64_t offset,
1246                                          int bytes,
1247                                          BlockCompletionFunc *cb,
1248                                          void *opaque)
1249 {
1250     return rbd_start_aio(bs, offset, NULL, bytes, cb, opaque,
1251                          RBD_AIO_DISCARD);
1252 }
1253 #endif
1254 
1255 #ifdef LIBRBD_SUPPORTS_INVALIDATE
1256 static void coroutine_fn qemu_rbd_co_invalidate_cache(BlockDriverState *bs,
1257                                                       Error **errp)
1258 {
1259     BDRVRBDState *s = bs->opaque;
1260     int r = rbd_invalidate_cache(s->image);
1261     if (r < 0) {
1262         error_setg_errno(errp, -r, "Failed to invalidate the cache");
1263     }
1264 }
1265 #endif
1266 
1267 static QemuOptsList qemu_rbd_create_opts = {
1268     .name = "rbd-create-opts",
1269     .head = QTAILQ_HEAD_INITIALIZER(qemu_rbd_create_opts.head),
1270     .desc = {
1271         {
1272             .name = BLOCK_OPT_SIZE,
1273             .type = QEMU_OPT_SIZE,
1274             .help = "Virtual disk size"
1275         },
1276         {
1277             .name = BLOCK_OPT_CLUSTER_SIZE,
1278             .type = QEMU_OPT_SIZE,
1279             .help = "RBD object size"
1280         },
1281         {
1282             .name = "password-secret",
1283             .type = QEMU_OPT_STRING,
1284             .help = "ID of secret providing the password",
1285         },
1286         { /* end of list */ }
1287     }
1288 };
1289 
1290 static const char *const qemu_rbd_strong_runtime_opts[] = {
1291     "pool",
1292     "image",
1293     "conf",
1294     "snapshot",
1295     "user",
1296     "server.",
1297     "password-secret",
1298 
1299     NULL
1300 };
1301 
1302 static BlockDriver bdrv_rbd = {
1303     .format_name            = "rbd",
1304     .instance_size          = sizeof(BDRVRBDState),
1305     .bdrv_parse_filename    = qemu_rbd_parse_filename,
1306     .bdrv_refresh_limits    = qemu_rbd_refresh_limits,
1307     .bdrv_file_open         = qemu_rbd_open,
1308     .bdrv_close             = qemu_rbd_close,
1309     .bdrv_reopen_prepare    = qemu_rbd_reopen_prepare,
1310     .bdrv_co_create         = qemu_rbd_co_create,
1311     .bdrv_co_create_opts    = qemu_rbd_co_create_opts,
1312     .bdrv_has_zero_init     = bdrv_has_zero_init_1,
1313     .bdrv_get_info          = qemu_rbd_getinfo,
1314     .create_opts            = &qemu_rbd_create_opts,
1315     .bdrv_getlength         = qemu_rbd_getlength,
1316     .bdrv_co_truncate       = qemu_rbd_co_truncate,
1317     .protocol_name          = "rbd",
1318 
1319     .bdrv_aio_preadv        = qemu_rbd_aio_preadv,
1320     .bdrv_aio_pwritev       = qemu_rbd_aio_pwritev,
1321 
1322 #ifdef LIBRBD_SUPPORTS_AIO_FLUSH
1323     .bdrv_aio_flush         = qemu_rbd_aio_flush,
1324 #else
1325     .bdrv_co_flush_to_disk  = qemu_rbd_co_flush,
1326 #endif
1327 
1328 #ifdef LIBRBD_SUPPORTS_DISCARD
1329     .bdrv_aio_pdiscard      = qemu_rbd_aio_pdiscard,
1330 #endif
1331 
1332     .bdrv_snapshot_create   = qemu_rbd_snap_create,
1333     .bdrv_snapshot_delete   = qemu_rbd_snap_remove,
1334     .bdrv_snapshot_list     = qemu_rbd_snap_list,
1335     .bdrv_snapshot_goto     = qemu_rbd_snap_rollback,
1336 #ifdef LIBRBD_SUPPORTS_INVALIDATE
1337     .bdrv_co_invalidate_cache = qemu_rbd_co_invalidate_cache,
1338 #endif
1339 
1340     .strong_runtime_opts    = qemu_rbd_strong_runtime_opts,
1341 };
1342 
1343 static void bdrv_rbd_init(void)
1344 {
1345     bdrv_register(&bdrv_rbd);
1346 }
1347 
1348 block_init(bdrv_rbd_init);
1349