aboutsummaryrefslogtreecommitdiff
path: root/nbd.c
blob: d1a67eeaf4a5a00dcdffc39837fa9d266b24d1ae (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
/*
 *  Copyright (C) 2005  Anthony Liguori <anthony@codemonkey.ws>
 *
 *  Network Block Device
 *
 *  This program is free software; you can redistribute it and/or modify
 *  it under the terms of the GNU General Public License as published by
 *  the Free Software Foundation; under version 2 of the License.
 *
 *  This program is distributed in the hope that it will be useful,
 *  but WITHOUT ANY WARRANTY; without even the implied warranty of
 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 *  GNU General Public License for more details.
 *
 *  You should have received a copy of the GNU General Public License
 *  along with this program; if not, see <http://www.gnu.org/licenses/>.
 */

#include "block/nbd.h"
#include "block/block.h"

#include "block/coroutine.h"

#include <errno.h>
#include <string.h>
#ifndef _WIN32
#include <sys/ioctl.h>
#endif
#if defined(__sun__) || defined(__HAIKU__)
#include <sys/ioccom.h>
#endif
#include <ctype.h>
#include <inttypes.h>

#ifdef __linux__
#include <linux/fs.h>
#endif

#include "qemu/sockets.h"
#include "qemu/queue.h"

//#define DEBUG_NBD

#ifdef DEBUG_NBD
#define TRACE(msg, ...) do { \
    LOG(msg, ## __VA_ARGS__); \
} while(0)
#else
#define TRACE(msg, ...) \
    do { } while (0)
#endif

#define LOG(msg, ...) do { \
    fprintf(stderr, "%s:%s():L%d: " msg "\n", \
            __FILE__, __FUNCTION__, __LINE__, ## __VA_ARGS__); \
} while(0)

/* This is all part of the "official" NBD API */

#define NBD_REQUEST_SIZE        (4 + 4 + 8 + 8 + 4)
#define NBD_REPLY_SIZE          (4 + 4 + 8)
#define NBD_REQUEST_MAGIC       0x25609513
#define NBD_REPLY_MAGIC         0x67446698
#define NBD_OPTS_MAGIC          0x49484156454F5054LL
#define NBD_CLIENT_MAGIC        0x0000420281861253LL

#define NBD_SET_SOCK            _IO(0xab, 0)
#define NBD_SET_BLKSIZE         _IO(0xab, 1)
#define NBD_SET_SIZE            _IO(0xab, 2)
#define NBD_DO_IT               _IO(0xab, 3)
#define NBD_CLEAR_SOCK          _IO(0xab, 4)
#define NBD_CLEAR_QUE           _IO(0xab, 5)
#define NBD_PRINT_DEBUG         _IO(0xab, 6)
#define NBD_SET_SIZE_BLOCKS     _IO(0xab, 7)
#define NBD_DISCONNECT          _IO(0xab, 8)
#define NBD_SET_TIMEOUT         _IO(0xab, 9)
#define NBD_SET_FLAGS           _IO(0xab, 10)

#define NBD_OPT_EXPORT_NAME     (1 << 0)

/* Definitions for opaque data types */

typedef struct NBDRequest NBDRequest;

struct NBDRequest {
    QSIMPLEQ_ENTRY(NBDRequest) entry;
    NBDClient *client;
    uint8_t *data;
};

struct NBDExport {
    int refcount;
    void (*close)(NBDExport *exp);

    BlockDriverState *bs;
    char *name;
    off_t dev_offset;
    off_t size;
    uint32_t nbdflags;
    QTAILQ_HEAD(, NBDClient) clients;
    QSIMPLEQ_HEAD(, NBDRequest) requests;
    QTAILQ_ENTRY(NBDExport) next;
};

static QTAILQ_HEAD(, NBDExport) exports = QTAILQ_HEAD_INITIALIZER(exports);

struct NBDClient {
    int refcount;
    void (*close)(NBDClient *client);

    NBDExport *exp;
    int sock;

    Coroutine *recv_coroutine;

    CoMutex send_lock;
    Coroutine *send_coroutine;

    QTAILQ_ENTRY(NBDClient) next;
    int nb_requests;
    bool closing;
};

/* That's all folks */

ssize_t nbd_wr_sync(int fd, void *buffer, size_t size, bool do_read)
{
    size_t offset = 0;
    int err;

    if (qemu_in_coroutine()) {
        if (do_read) {
            return qemu_co_recv(fd, buffer, size);
        } else {
            return qemu_co_send(fd, buffer, size);
        }
    }

    while (offset < size) {
        ssize_t len;

        if (do_read) {
            len = qemu_recv(fd, buffer + offset, size - offset, 0);
        } else {
            len = send(fd, buffer + offset, size - offset, 0);
        }

        if (len < 0) {
            err = socket_error();

            /* recoverable error */
            if (err == EINTR || (offset > 0 && err == EAGAIN)) {
                continue;
            }

            /* unrecoverable error */
            return -err;
        }

        /* eof */
        if (len == 0) {
            break;
        }

        offset += len;
    }

    return offset;
}

static ssize_t read_sync(int fd, void *buffer, size_t size)
{
    /* Sockets are kept in blocking mode in the negotiation phase.  After
     * that, a non-readable socket simply means that another thread stole
     * our request/reply.  Synchronization is done with recv_coroutine, so
     * that this is coroutine-safe.
     */
    return nbd_wr_sync(fd, buffer, size, true);
}

static ssize_t write_sync(int fd, void *buffer, size_t size)
{
    int ret;
    do {
        /* For writes, we do expect the socket to be writable.  */
        ret = nbd_wr_sync(fd, buffer, size, false);
    } while (ret == -EAGAIN);
    return ret;
}

static void combine_addr(char *buf, size_t len, const char* address,
                         uint16_t port)
{
    /* If the address-part contains a colon, it's an IPv6 IP so needs [] */
    if (strstr(address, ":")) {
        snprintf(buf, len, "[%s]:%u", address, port);
    } else {
        snprintf(buf, len, "%s:%u", address, port);
    }
}

int tcp_socket_outgoing_opts(QemuOpts *opts)
{
    Error *local_err = NULL;
    int fd = inet_connect_opts(opts, &local_err, NULL, NULL);
    if (local_err != NULL) {
        qerror_report_err(local_err);
        error_free(local_err);
    }

    return fd;
}

int tcp_socket_incoming(const char *address, uint16_t port)
{
    char address_and_port[128];
    combine_addr(address_and_port, 128, address, port);
    return tcp_socket_incoming_spec(address_and_port);
}

int tcp_socket_incoming_spec(const char *address_and_port)
{
    Error *local_err = NULL;
    int fd = inet_listen(address_and_port, NULL, 0, SOCK_STREAM, 0, &local_err);

    if (local_err != NULL) {
        qerror_report_err(local_err);
        error_free(local_err);
    }
    return fd;
}

int unix_socket_incoming(const char *path)
{
    Error *local_err = NULL;
    int fd = unix_listen(path, NULL, 0, &local_err);

    if (local_err != NULL) {
        qerror_report_err(local_err);
        error_free(local_err);
    }
    return fd;
}

int unix_socket_outgoing(const char *path)
{
    Error *local_err = NULL;
    int fd = unix_connect(path, &local_err);

    if (local_err != NULL) {
        qerror_report_err(local_err);
        error_free(local_err);
    }
    return fd;
}

/* Basic flow for negotiation

   Server         Client
   Negotiate

   or

   Server         Client
   Negotiate #1
                  Option
   Negotiate #2

   ----

   followed by

   Server         Client
                  Request
   Response
                  Request
   Response
                  ...
   ...
                  Request (type == 2)

*/

static int nbd_receive_options(NBDClient *client)
{
    int csock = client->sock;
    char name[256];
    uint32_t tmp, length;
    uint64_t magic;
    int rc;

    /* Client sends:
        [ 0 ..   3]   reserved (0)
        [ 4 ..  11]   NBD_OPTS_MAGIC
        [12 ..  15]   NBD_OPT_EXPORT_NAME
        [16 ..  19]   length
        [20 ..  xx]   export name (length bytes)
     */

    rc = -EINVAL;
    if (read_sync(csock, &tmp, sizeof(tmp)) != sizeof(tmp)) {
        LOG("read failed");
        goto fail;
    }
    TRACE("Checking reserved");
    if (tmp != 0) {
        LOG("Bad reserved received");
        goto fail;
    }

    if (read_sync(csock, &magic, sizeof(magic)) != sizeof(magic)) {
        LOG("read failed");
        goto fail;
    }
    TRACE("Checking reserved");
    if (magic != be64_to_cpu(NBD_OPTS_MAGIC)) {
        LOG("Bad magic received");
        goto fail;
    }

    if (read_sync(csock, &tmp, sizeof(tmp)) != sizeof(tmp)) {
        LOG("read failed");
        goto fail;
    }
    TRACE("Checking option");
    if (tmp != be32_to_cpu(NBD_OPT_EXPORT_NAME)) {
        LOG("Bad option received");
        goto fail;
    }

    if (read_sync(csock, &length, sizeof(length)) != sizeof(length)) {
        LOG("read failed");
        goto fail;
    }
    TRACE("Checking length");
    length = be32_to_cpu(length);
    if (length > 255) {
        LOG("Bad length received");
        goto fail;
    }
    if (read_sync(csock, name, length) != length) {
        LOG("read failed");
        goto fail;
    }
    name[length] = '\0';

    client->exp = nbd_export_find(name);
    if (!client->exp) {
        LOG("export not found");
        goto fail;
    }

    QTAILQ_INSERT_TAIL(&client->exp->clients, client, next);
    nbd_export_get(client->exp);

    TRACE("Option negotiation succeeded.");
    rc = 0;
fail:
    return rc;
}

static int nbd_send_negotiate(NBDClient *client)
{
    int csock = client->sock;
    char buf[8 + 8 + 8 + 128];
    int rc;
    const int myflags = (NBD_FLAG_HAS_FLAGS | NBD_FLAG_SEND_TRIM |
                         NBD_FLAG_SEND_FLUSH | NBD_FLAG_SEND_FUA);

    /* Negotiation header without options:
        [ 0 ..   7]   passwd       ("NBDMAGIC")
        [ 8 ..  15]   magic        (NBD_CLIENT_MAGIC)
        [16 ..  23]   size
        [24 ..  25]   server flags (0)
        [24 ..  27]   export flags
        [28 .. 151]   reserved     (0)

       Negotiation header with options, part 1:
        [ 0 ..   7]   passwd       ("NBDMAGIC")
        [ 8 ..  15]   magic        (NBD_OPTS_MAGIC)
        [16 ..  17]   server flags (0)

       part 2 (after options are sent):
        [18 ..  25]   size
        [26 ..  27]   export flags
        [28 .. 151]   reserved     (0)
     */

    socket_set_block(csock);
    rc = -EINVAL;

    TRACE("Beginning negotiation.");
    memset(buf, 0, sizeof(buf));
    memcpy(buf, "NBDMAGIC", 8);
    if (client->exp) {
        assert ((client->exp->nbdflags & ~65535) == 0);
        cpu_to_be64w((uint64_t*)(buf + 8), NBD_CLIENT_MAGIC);
        cpu_to_be64w((uint64_t*)(buf + 16), client->exp->size);
        cpu_to_be16w((uint16_t*)(buf + 26), client->exp->nbdflags | myflags);
    } else {
        cpu_to_be64w((uint64_t*)(buf + 8), NBD_OPTS_MAGIC);
    }

    if (client->exp) {
        if (write_sync(csock, buf, sizeof(buf)) != sizeof(buf)) {
            LOG("write failed");
            goto fail;
        }
    } else {
        if (write_sync(csock, buf, 18) != 18) {
            LOG("write failed");
            goto fail;
        }
        rc = nbd_receive_options(client);
        if (rc < 0) {
            LOG("option negotiation failed");
            goto fail;
        }

        assert ((client->exp->nbdflags & ~65535) == 0);
        cpu_to_be64w((uint64_t*)(buf + 18), client->exp->size);
        cpu_to_be16w((uint16_t*)(buf + 26), client->exp->nbdflags | myflags);
        if (write_sync(csock, buf + 18, sizeof(buf) - 18) != sizeof(buf) - 18) {
            LOG("write failed");
            goto fail;
        }
    }

    TRACE("Negotiation succeeded.");
    rc = 0;
fail:
    socket_set_nonblock(csock);
    return rc;
}

int nbd_receive_negotiate(int csock, const char *name, uint32_t *flags,
                          off_t *size, size_t *blocksize)
{
    char buf[256];
    uint64_t magic, s;
    uint16_t tmp;
    int rc;

    TRACE("Receiving negotiation.");

    socket_set_block(csock);
    rc = -EINVAL;

    if (read_sync(csock, buf, 8) != 8) {
        LOG("read failed");
        goto fail;
    }

    buf[8] = '\0';
    if (strlen(buf) == 0) {
        LOG("server connection closed");
        goto fail;
    }

    TRACE("Magic is %c%c%c%c%c%c%c%c",
          qemu_isprint(buf[0]) ? buf[0] : '.',
          qemu_isprint(buf[1]) ? buf[1] : '.',
          qemu_isprint(buf[2]) ? buf[2] : '.',
          qemu_isprint(buf[3]) ? buf[3] : '.',
          qemu_isprint(buf[4]) ? buf[4] : '.',
          qemu_isprint(buf[5]) ? buf[5] : '.',
          qemu_isprint(buf[6]) ? buf[6] : '.',
          qemu_isprint(buf[7]) ? buf[7] : '.');

    if (memcmp(buf, "NBDMAGIC", 8) != 0) {
        LOG("Invalid magic received");
        goto fail;
    }

    if (read_sync(csock, &magic, sizeof(magic)) != sizeof(magic)) {
        LOG("read failed");
        goto fail;
    }
    magic = be64_to_cpu(magic);
    TRACE("Magic is 0x%" PRIx64, magic);

    if (name) {
        uint32_t reserved = 0;
        uint32_t opt;
        uint32_t namesize;

        TRACE("Checking magic (opts_magic)");
        if (magic != NBD_OPTS_MAGIC) {
            LOG("Bad magic received");
            goto fail;
        }
        if (read_sync(csock, &tmp, sizeof(tmp)) != sizeof(tmp)) {
            LOG("flags read failed");
            goto fail;
        }
        *flags = be16_to_cpu(tmp) << 16;
        /* reserved for future use */
        if (write_sync(csock, &reserved, sizeof(reserved)) !=
            sizeof(reserved)) {
            LOG("write failed (reserved)");
            goto fail;
        }
        /* write the export name */
        magic = cpu_to_be64(magic);
        if (write_sync(csock, &magic, sizeof(magic)) != sizeof(magic)) {
            LOG("write failed (magic)");
            goto fail;
        }
        opt = cpu_to_be32(NBD_OPT_EXPORT_NAME);
        if (write_sync(csock, &opt, sizeof(opt)) != sizeof(opt)) {
            LOG("write failed (opt)");
            goto fail;
        }
        namesize = cpu_to_be32(strlen(name));
        if (write_sync(csock, &namesize, sizeof(namesize)) !=
            sizeof(namesize)) {
            LOG("write failed (namesize)");
            goto fail;
        }
        if (write_sync(csock, (char*)name, strlen(name)) != strlen(name)) {
            LOG("write failed (name)");
            goto fail;
        }
    } else {
        TRACE("Checking magic (cli_magic)");

        if (magic != NBD_CLIENT_MAGIC) {
            LOG("Bad magic received");
            goto fail;
        }
    }

    if (read_sync(csock, &s, sizeof(s)) != sizeof(s)) {
        LOG("read failed");
        goto fail;
    }
    *size = be64_to_cpu(s);
    *blocksize = 1024;
    TRACE("Size is %" PRIu64, *size);

    if (!name) {
        if (read_sync(csock, flags, sizeof(*flags)) != sizeof(*flags)) {
            LOG("read failed (flags)");
            goto fail;
        }
        *flags = be32_to_cpup(flags);
    } else {
        if (read_sync(csock, &tmp, sizeof(tmp)) != sizeof(tmp)) {
            LOG("read failed (tmp)");
            goto fail;
        }
        *flags |= be32_to_cpu(tmp);
    }
    if (read_sync(csock, &buf, 124) != 124) {
        LOG("read failed (buf)");
        goto fail;
    }
    rc = 0;

fail:
    socket_set_nonblock(csock);
    return rc;
}

#ifdef __linux__
int nbd_init(int fd, int csock, uint32_t flags, off_t size, size_t blocksize)
{
    TRACE("Setting NBD socket");

    if (ioctl(fd, NBD_SET_SOCK, csock) < 0) {
        int serrno = errno;
        LOG("Failed to set NBD socket");
        return -serrno;
    }

    TRACE("Setting block size to %lu", (unsigned long)blocksize);

    if (ioctl(fd, NBD_SET_BLKSIZE, blocksize) < 0) {
        int serrno = errno;
        LOG("Failed setting NBD block size");
        return -serrno;
    }

        TRACE("Setting size to %zd block(s)", (size_t)(size / blocksize));

    if (ioctl(fd, NBD_SET_SIZE_BLOCKS, size / blocksize) < 0) {
        int serrno = errno;
        LOG("Failed setting size (in blocks)");
        return -serrno;
    }

    if (ioctl(fd, NBD_SET_FLAGS, flags) < 0) {
        if (errno == ENOTTY) {
            int read_only = (flags & NBD_FLAG_READ_ONLY) != 0;
            TRACE("Setting readonly attribute");

            if (ioctl(fd, BLKROSET, (unsigned long) &read_only) < 0) {
                int serrno = errno;
                LOG("Failed setting read-only attribute");
                return -serrno;
            }
        } else {
            int serrno = errno;
            LOG("Failed setting flags");
            return -serrno;
        }
    }

    TRACE("Negotiation ended");

    return 0;
}

int nbd_disconnect(int fd)
{
    ioctl(fd, NBD_CLEAR_QUE);
    ioctl(fd, NBD_DISCONNECT);
    ioctl(fd, NBD_CLEAR_SOCK);
    return 0;
}

int nbd_client(int fd)
{
    int ret;
    int serrno;

    TRACE("Doing NBD loop");

    ret = ioctl(fd, NBD_DO_IT);
    if (ret < 0 && errno == EPIPE) {
        /* NBD_DO_IT normally returns EPIPE when someone has disconnected
         * the socket via NBD_DISCONNECT.  We do not want to return 1 in
         * that case.
         */
        ret = 0;
    }
    serrno = errno;

    TRACE("NBD loop returned %d: %s", ret, strerror(serrno));

    TRACE("Clearing NBD queue");
    ioctl(fd, NBD_CLEAR_QUE);

    TRACE("Clearing NBD socket");
    ioctl(fd, NBD_CLEAR_SOCK);

    errno = serrno;
    return ret;
}
#else
int nbd_init(int fd, int csock, uint32_t flags, off_t size, size_t blocksize)
{
    return -ENOTSUP;
}

int nbd_disconnect(int fd)
{
    return -ENOTSUP;
}

int nbd_client(int fd)
{
    return -ENOTSUP;
}
#endif

ssize_t nbd_send_request(int csock, struct nbd_request *request)
{
    uint8_t buf[NBD_REQUEST_SIZE];
    ssize_t ret;

    cpu_to_be32w((uint32_t*)buf, NBD_REQUEST_MAGIC);
    cpu_to_be32w((uint32_t*)(buf + 4), request->type);
    cpu_to_be64w((uint64_t*)(buf + 8), request->handle);
    cpu_to_be64w((uint64_t*)(buf + 16), request->from);
    cpu_to_be32w((uint32_t*)(buf + 24), request->len);

    TRACE("Sending request to client: "
          "{ .from = %" PRIu64", .len = %u, .handle = %" PRIu64", .type=%i}",
          request->from, request->len, request->handle, request->type);

    ret = write_sync(csock, buf, sizeof(buf));
    if (ret < 0) {
        return ret;
    }

    if (ret != sizeof(buf)) {
        LOG("writing to socket failed");
        return -EINVAL;
    }
    return 0;
}

static ssize_t nbd_receive_request(int csock, struct nbd_request *request)
{
    uint8_t buf[NBD_REQUEST_SIZE];
    uint32_t magic;
    ssize_t ret;

    ret = read_sync(csock, buf, sizeof(buf));
    if (ret < 0) {
        return ret;
    }

    if (ret != sizeof(buf)) {
        LOG("read failed");
        return -EINVAL;
    }

    /* Request
       [ 0 ..  3]   magic   (NBD_REQUEST_MAGIC)
       [ 4 ..  7]   type    (0 == READ, 1 == WRITE)
       [ 8 .. 15]   handle
       [16 .. 23]   from
       [24 .. 27]   len
     */

    magic = be32_to_cpup((uint32_t*)buf);
    request->type  = be32_to_cpup((uint32_t*)(buf + 4));
    request->handle = be64_to_cpup((uint64_t*)(buf + 8));
    request->from  = be64_to_cpup((uint64_t*)(buf + 16));
    request->len   = be32_to_cpup((uint32_t*)(buf + 24));

    TRACE("Got request: "
          "{ magic = 0x%x, .type = %d, from = %" PRIu64" , len = %u }",
          magic, request->type, request->from, request->len);

    if (magic != NBD_REQUEST_MAGIC) {
        LOG("invalid magic (got 0x%x)", magic);
        return -EINVAL;
    }
    return 0;
}

ssize_t nbd_receive_reply(int csock, struct nbd_reply *reply)
{
    uint8_t buf[NBD_REPLY_SIZE];
    uint32_t magic;
    ssize_t ret;

    ret = read_sync(csock, buf, sizeof(buf));
    if (ret < 0) {
        return ret;
    }

    if (ret != sizeof(buf)) {
        LOG("read failed");
        return -EINVAL;
    }

    /* Reply
       [ 0 ..  3]    magic   (NBD_REPLY_MAGIC)
       [ 4 ..  7]    error   (0 == no error)
       [ 7 .. 15]    handle
     */

    magic = be32_to_cpup((uint32_t*)buf);
    reply->error  = be32_to_cpup((uint32_t*)(buf + 4));
    reply->handle = be64_to_cpup((uint64_t*)(buf + 8));

    TRACE("Got reply: "
          "{ magic = 0x%x, .error = %d, handle = %" PRIu64" }",
          magic, reply->error, reply->handle);

    if (magic != NBD_REPLY_MAGIC) {
        LOG("invalid magic (got 0x%x)", magic);
        return -EINVAL;
    }
    return 0;
}

static ssize_t nbd_send_reply(int csock, struct nbd_reply *reply)
{
    uint8_t buf[NBD_REPLY_SIZE];
    ssize_t ret;

    /* Reply
       [ 0 ..  3]    magic   (NBD_REPLY_MAGIC)
       [ 4 ..  7]    error   (0 == no error)
       [ 7 .. 15]    handle
     */
    cpu_to_be32w((uint32_t*)buf, NBD_REPLY_MAGIC);
    cpu_to_be32w((uint32_t*)(buf + 4), reply->error);
    cpu_to_be64w((uint64_t*)(buf + 8), reply->handle);

    TRACE("Sending response to client");

    ret = write_sync(csock, buf, sizeof(buf));
    if (ret < 0) {
        return ret;
    }

    if (ret != sizeof(buf)) {
        LOG("writing to socket failed");
        return -EINVAL;
    }
    return 0;
}

#define MAX_NBD_REQUESTS 16

void nbd_client_get(NBDClient *client)
{
    client->refcount++;
}

void nbd_client_put(NBDClient *client)
{
    if (--client->refcount == 0) {
        /* The last reference should be dropped by client->close,
         * which is called by nbd_client_close.
         */
        assert(client->closing);

        qemu_set_fd_handler2(client->sock, NULL, NULL, NULL, NULL);
        close(client->sock);
        client->sock = -1;
        if (client->exp) {
            QTAILQ_REMOVE(&client->exp->clients, client, next);
            nbd_export_put(client->exp);
        }
        g_free(client);
    }
}

void nbd_client_close(NBDClient *client)
{
    if (client->closing) {
        return;
    }

    client->closing = true;

    /* Force requests to finish.  They will drop their own references,
     * then we'll close the socket and free the NBDClient.
     */
    shutdown(client->sock, 2);

    /* Also tell the client, so that they release their reference.  */
    if (client->close) {
        client->close(client);
    }
}

static NBDRequest *nbd_request_get(NBDClient *client)
{
    NBDRequest *req;
    NBDExport *exp = client->exp;

    assert(client->nb_requests <= MAX_NBD_REQUESTS - 1);
    client->nb_requests++;

    if (QSIMPLEQ_EMPTY(&exp->requests)) {
        req = g_malloc0(sizeof(NBDRequest));
        req->data = qemu_blockalign(exp->bs, NBD_BUFFER_SIZE);
    } else {
        req = QSIMPLEQ_FIRST(&exp->requests);
        QSIMPLEQ_REMOVE_HEAD(&exp->requests, entry);
    }
    nbd_client_get(client);
    req->client = client;
    return req;
}

static void nbd_request_put(NBDRequest *req)
{
    NBDClient *client = req->client;
    QSIMPLEQ_INSERT_HEAD(&client->exp->requests, req, entry);
    if (client->nb_requests-- == MAX_NBD_REQUESTS) {
        qemu_notify_event();
    }
    nbd_client_put(client);
}

NBDExport *nbd_export_new(BlockDriverState *bs, off_t dev_offset,
                          off_t size, uint32_t nbdflags,
                          void (*close)(NBDExport *))
{
    NBDExport *exp = g_malloc0(sizeof(NBDExport));
    QSIMPLEQ_INIT(&exp->requests);
    exp->refcount = 1;
    QTAILQ_INIT(&exp->clients);
    exp->bs = bs;
    exp->dev_offset = dev_offset;
    exp->nbdflags = nbdflags;
    exp->size = size == -1 ? bdrv_getlength(bs) : size;
    exp->close = close;
    return exp;
}

NBDExport *nbd_export_find(const char *name)
{
    NBDExport *exp;
    QTAILQ_FOREACH(exp, &exports, next) {
        if (strcmp(name, exp->name) == 0) {
            return exp;
        }
    }

    return NULL;
}

void nbd_export_set_name(NBDExport *exp, const char *name)
{
    if (exp->name == name) {
        return;
    }

    nbd_export_get(exp);
    if (exp->name != NULL) {
        g_free(exp->name);
        exp->name = NULL;
        QTAILQ_REMOVE(&exports, exp, next);
        nbd_export_put(exp);
    }
    if (name != NULL) {
        nbd_export_get(exp);
        exp->name = g_strdup(name);
        QTAILQ_INSERT_TAIL(&exports, exp, next);
    }
    nbd_export_put(exp);
}

void nbd_export_close(NBDExport *exp)
{
    NBDClient *client, *next;

    nbd_export_get(exp);
    QTAILQ_FOREACH_SAFE(client, &exp->clients, next, next) {
        nbd_client_close(client);
    }
    nbd_export_set_name(exp, NULL);
    nbd_export_put(exp);
}

void nbd_export_get(NBDExport *exp)
{
    assert(exp->refcount > 0);
    exp->refcount++;
}

void nbd_export_put(NBDExport *exp)
{
    assert(exp->refcount > 0);
    if (exp->refcount == 1) {
        nbd_export_close(exp);
    }

    if (--exp->refcount == 0) {
        assert(exp->name == NULL);

        if (exp->close) {
            exp->close(exp);
        }

        while (!QSIMPLEQ_EMPTY(&exp->requests)) {
            NBDRequest *first = QSIMPLEQ_FIRST(&exp->requests);
            QSIMPLEQ_REMOVE_HEAD(&exp->requests, entry);
            qemu_vfree(first->data);
            g_free(first);
        }

        g_free(exp);
    }
}

BlockDriverState *nbd_export_get_blockdev(NBDExport *exp)
{
    return exp->bs;
}

void nbd_export_close_all(void)
{
    NBDExport *exp, *next;

    QTAILQ_FOREACH_SAFE(exp, &exports, next, next) {
        nbd_export_close(exp);
    }
}

static int nbd_can_read(void *opaque);
static void nbd_read(void *opaque);
static void nbd_restart_write(void *opaque);

static ssize_t nbd_co_send_reply(NBDRequest *req, struct nbd_reply *reply,
                                 int len)
{
    NBDClient *client = req->client;
    int csock = client->sock;
    ssize_t rc, ret;

    qemu_co_mutex_lock(&client->send_lock);
    qemu_set_fd_handler2(csock, nbd_can_read, nbd_read,
                         nbd_restart_write, client);
    client->send_coroutine = qemu_coroutine_self();

    if (!len) {
        rc = nbd_send_reply(csock, reply);
    } else {
        socket_set_cork(csock, 1);
        rc = nbd_send_reply(csock, reply);
        if (rc >= 0) {
            ret = qemu_co_send(csock, req->data, len);
            if (ret != len) {
                rc = -EIO;
            }
        }
        socket_set_cork(csock, 0);
    }

    client->send_coroutine = NULL;
    qemu_set_fd_handler2(csock, nbd_can_read, nbd_read, NULL, client);
    qemu_co_mutex_unlock(&client->send_lock);
    return rc;
}

static ssize_t nbd_co_receive_request(NBDRequest *req, struct nbd_request *request)
{
    NBDClient *client = req->client;
    int csock = client->sock;
    ssize_t rc;

    client->recv_coroutine = qemu_coroutine_self();
    rc = nbd_receive_request(csock, request);
    if (rc < 0) {
        if (rc != -EAGAIN) {
            rc = -EIO;
        }
        goto out;
    }

    if (request->len > NBD_BUFFER_SIZE) {
        LOG("len (%u) is larger than max len (%u)",
            request->len, NBD_BUFFER_SIZE);
        rc = -EINVAL;
        goto out;
    }

    if ((request->from + request->len) < request->from) {
        LOG("integer overflow detected! "
            "you're probably being attacked");
        rc = -EINVAL;
        goto out;
    }

    TRACE("Decoding type");

    if ((request->type & NBD_CMD_MASK_COMMAND) == NBD_CMD_WRITE) {
        TRACE("Reading %u byte(s)", request->len);

        if (qemu_co_recv(csock, req->data, request->len) != request->len) {
            LOG("reading from socket failed");
            rc = -EIO;
            goto out;
        }
    }
    rc = 0;

out:
    client->recv_coroutine = NULL;
    return rc;
}

static void nbd_trip(void *opaque)
{
    NBDClient *client = opaque;
    NBDExport *exp = client->exp;
    NBDRequest *req;
    struct nbd_request request;
    struct nbd_reply reply;
    ssize_t ret;

    TRACE("Reading request.");
    if (client->closing) {
        return;
    }

    req = nbd_request_get(client);
    ret = nbd_co_receive_request(req, &request);
    if (ret == -EAGAIN) {
        goto done;
    }
    if (ret == -EIO) {
        goto out;
    }

    reply.handle = request.handle;
    reply.error = 0;

    if (ret < 0) {
        reply.error = -ret;
        goto error_reply;
    }

    if ((request.from + request.len) > exp->size) {
            LOG("From: %" PRIu64 ", Len: %u, Size: %" PRIu64
            ", Offset: %" PRIu64 "\n",
                    request.from, request.len,
                    (uint64_t)exp->size, (uint64_t)exp->dev_offset);
        LOG("requested operation past EOF--bad client?");
        goto invalid_request;
    }

    switch (request.type & NBD_CMD_MASK_COMMAND) {
    case NBD_CMD_READ:
        TRACE("Request type is READ");

        if (request.type & NBD_CMD_FLAG_FUA) {
            ret = bdrv_co_flush(exp->bs);
            if (ret < 0) {
                LOG("flush failed");
                reply.error = -ret;
                goto error_reply;
            }
        }

        ret = bdrv_read(exp->bs, (request.from + exp->dev_offset) / 512,
                        req->data, request.len / 512);
        if (ret < 0) {
            LOG("reading from file failed");
            reply.error = -ret;
            goto error_reply;
        }

        TRACE("Read %u byte(s)", request.len);
        if (nbd_co_send_reply(req, &reply, request.len) < 0)
            goto out;
        break;
    case NBD_CMD_WRITE:
        TRACE("Request type is WRITE");

        if (exp->nbdflags & NBD_FLAG_READ_ONLY) {
            TRACE("Server is read-only, return error");
            reply.error = EROFS;
            goto error_reply;
        }

        TRACE("Writing to device");

        ret = bdrv_write(exp->bs, (request.from + exp->dev_offset) / 512,
                         req->data, request.len / 512);
        if (ret < 0) {
            LOG("writing to file failed");
            reply.error = -ret;
            goto error_reply;
        }

        if (request.type & NBD_CMD_FLAG_FUA) {
            ret = bdrv_co_flush(exp->bs);
            if (ret < 0) {
                LOG("flush failed");
                reply.error = -ret;
                goto error_reply;
            }
        }

        if (nbd_co_send_reply(req, &reply, 0) < 0) {
            goto out;
        }
        break;
    case NBD_CMD_DISC:
        TRACE("Request type is DISCONNECT");
        errno = 0;
        goto out;
    case NBD_CMD_FLUSH:
        TRACE("Request type is FLUSH");

        ret = bdrv_co_flush(exp->bs);
        if (ret < 0) {
            LOG("flush failed");
            reply.error = -ret;
        }
        if (nbd_co_send_reply(req, &reply, 0) < 0) {
            goto out;
        }
        break;
    case NBD_CMD_TRIM:
        TRACE("Request type is TRIM");
        ret = bdrv_co_discard(exp->bs, (request.from + exp->dev_offset) / 512,
                              request.len / 512);
        if (ret < 0) {
            LOG("discard failed");
            reply.error = -ret;
        }
        if (nbd_co_send_reply(req, &reply, 0) < 0) {
            goto out;
        }
        break;
    default:
        LOG("invalid request type (%u) received", request.type);
    invalid_request:
        reply.error = -EINVAL;
    error_reply:
        if (nbd_co_send_reply(req, &reply, 0) < 0) {
            goto out;
        }
        break;
    }

    TRACE("Request/Reply complete");

done:
    nbd_request_put(req);
    return;

out:
    nbd_request_put(req);
    nbd_client_close(client);
}

static int nbd_can_read(void *opaque)
{
    NBDClient *client = opaque;

    return client->recv_coroutine || client->nb_requests < MAX_NBD_REQUESTS;
}

static void nbd_read(void *opaque)
{
    NBDClient *client = opaque;

    if (client->recv_coroutine) {
        qemu_coroutine_enter(client->recv_coroutine, NULL);
    } else {
        qemu_coroutine_enter(qemu_coroutine_create(nbd_trip), client);
    }
}

static void nbd_restart_write(void *opaque)
{
    NBDClient *client = opaque;

    qemu_coroutine_enter(client->send_coroutine, NULL);
}

NBDClient *nbd_client_new(NBDExport *exp, int csock,
                          void (*close)(NBDClient *))
{
    NBDClient *client;
    client = g_malloc0(sizeof(NBDClient));
    client->refcount = 1;
    client->exp = exp;
    client->sock = csock;
    if (nbd_send_negotiate(client) < 0) {
        g_free(client);
        return NULL;
    }
    client->close = close;
    qemu_co_mutex_init(&client->send_lock);
    qemu_set_fd_handler2(csock, nbd_can_read, nbd_read, NULL, client);

    if (exp) {
        QTAILQ_INSERT_TAIL(&exp->clients, client, next);
        nbd_export_get(exp);
    }
    return client;
}