summaryrefslogtreecommitdiffstats
path: root/src/launch-libvirt.c
blob: 8d1cb678af8fc37ae473b10a1c61fac27704a477 (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
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
/* libguestfs
 * Copyright (C) 2009-2012 Red Hat Inc.
 *
 * This library is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2 of the License, or (at your option) any later version.
 *
 * This library 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
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with this library; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

#include <config.h>

#include <stdio.h>
#include <stdlib.h>
#include <stdarg.h>
#include <stdbool.h>
#include <unistd.h>
#include <fcntl.h>
#include <limits.h>
#include <grp.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <sys/wait.h>
#include <assert.h>

#ifdef HAVE_LIBVIRT
#include <libvirt/libvirt.h>
#include <libvirt/virterror.h>
#endif

#ifdef HAVE_LIBXML2
#include <libxml/xmlIO.h>
#include <libxml/xmlwriter.h>
#include <libxml/xpath.h>
#include <libxml/parser.h>
#include <libxml/tree.h>
#include <libxml/xmlsave.h>
#endif

#include "glthread/lock.h"

#include "guestfs.h"
#include "guestfs-internal.h"
#include "guestfs-internal-actions.h"
#include "guestfs_protocol.h"

/* Check minimum required version of libvirt.  The libvirt backend
 * is new and not the default, so we can get away with forcing
 * people who want to try it to have a reasonably new version of
 * libvirt, so we don't have to work around any bugs in libvirt.
 *
 * This is also checked at runtime because you can dynamically link
 * with a different version from what you were compiled with.
 */
#define MIN_LIBVIRT_MAJOR 0
#define MIN_LIBVIRT_MINOR 10
#define MIN_LIBVIRT_MICRO 2 /* XXX patches in > 2 already */
#define MIN_LIBVIRT_VERSION (MIN_LIBVIRT_MAJOR * 1000000 + \
                             MIN_LIBVIRT_MINOR * 1000 + \
                             MIN_LIBVIRT_MICRO)

#if defined(HAVE_LIBVIRT) && \
  LIBVIR_VERSION_NUMBER >= MIN_LIBVIRT_VERSION && \
  defined(HAVE_LIBXML2)

#ifndef HAVE_XMLBUFFERDETACH
/* Added in libxml2 2.8.0.  This is mostly a copy of the function from
 * upstream libxml2, which is under a more permissive license.
 */
static xmlChar *
xmlBufferDetach (xmlBufferPtr buf)
{
  xmlChar *ret;

  if (buf == NULL)
    return NULL;
  if (buf->alloc == XML_BUFFER_ALLOC_IMMUTABLE)
    return NULL;

  ret = buf->content;
  buf->content = NULL;
  buf->size = 0;
  buf->use = 0;

  return ret;
}
#endif

/* Pointed to by 'struct drive *' -> priv field. */
struct drive_libvirt {
  /* This is either the original path, made absolute.  Or for readonly
   * drives, it is an (absolute path to) the overlay file that we
   * create.  This is always non-NULL.
   */
  char *path;
  /* The format of priv->path. */
  char *format;
};

/* Parameters passed to construct_libvirt_xml and subfunctions.  We
 * keep them all in a structure for convenience!
 */
struct libvirt_xml_params {
  char *capabilities_xml;       /* libvirt capabilities XML */
  char *kernel;                 /* paths to kernel and initrd */
  char *initrd;
  char *appliance_overlay;      /* path to qcow2 overlay backed by appliance */
  char appliance_dev[64];       /* appliance device name */
  size_t appliance_index;       /* index of appliance */
  char guestfsd_sock[UNIX_PATH_MAX]; /* paths to sockets */
  char console_sock[UNIX_PATH_MAX];
  bool enable_svirt;            /* false if we decided to disable sVirt */
  bool is_kvm;                  /* false = qemu, true = kvm */
};

static xmlChar *construct_libvirt_xml (guestfs_h *g, const struct libvirt_xml_params *params);
static void libvirt_error (guestfs_h *g, const char *fs, ...);
static int is_custom_qemu (guestfs_h *g);
static int is_blk (const char *path);
static int random_chars (char *ret, size_t len);
static void ignore_errors (void *ignore, virErrorPtr ignore2);
static char *make_qcow2_overlay (guestfs_h *g, const char *path, const char *format);
static int make_qcow2_overlay_for_drive (guestfs_h *g, struct drive *drv);
static void drive_free_priv (void *);

static int
launch_libvirt (guestfs_h *g, const char *libvirt_uri)
{
  unsigned long version;
  int is_root = geteuid () == 0;
  virConnectPtr conn = NULL;
  virDomainPtr dom = NULL;
  struct libvirt_xml_params params = {
    .capabilities_xml = NULL,
    .kernel = NULL,
    .initrd = NULL,
    .appliance_overlay = NULL,
  };
  xmlChar *xml = NULL;
  char *appliance = NULL;
  struct sockaddr_un addr;
  int console = -1, r;
  uint32_t size;
  void *buf = NULL;
  struct drive *drv;
  size_t i;

  /* XXX: It should be possible to make this work. */
  if (g->direct) {
    error (g, _("direct mode flag is not supported yet for libvirt attach method"));
    return -1;
  }

  virGetVersion (&version, NULL, NULL);
  debug (g, "libvirt version = %lu", version);
  if (version < MIN_LIBVIRT_VERSION) {
    error (g, _("you must have libvirt >= %d.%d.%d "
                "to use the 'libvirt' attach-method"),
           MIN_LIBVIRT_MAJOR, MIN_LIBVIRT_MINOR, MIN_LIBVIRT_MICRO);
    return -1;
  }

  guestfs___launch_send_progress (g, 0);
  TRACE0 (launch_libvirt_start);

  if (g->verbose)
    guestfs___print_timestamped_message (g, "connect to libvirt");

  /* Connect to libvirt, get capabilities. */
  conn = guestfs___open_libvirt_connection (g, libvirt_uri, 0);
  if (!conn) {
    libvirt_error (g, _("could not connect to libvirt (URI = %s)"),
           libvirt_uri ? : "NULL");
    goto cleanup;
  }

  /* Suppress default behaviour of printing errors to stderr.  Note
   * you can't set this to NULL to ignore errors; setting it to NULL
   * restores the default error handler ...
   */
  virConnSetErrorFunc (conn, NULL, ignore_errors);

  if (g->verbose)
    guestfs___print_timestamped_message (g, "get libvirt capabilities");

  params.capabilities_xml = virConnectGetCapabilities (conn);
  if (!params.capabilities_xml) {
    libvirt_error (g, _("could not get libvirt capabilities"));
    goto cleanup;
  }

  /* Locate and/or build the appliance. */
  TRACE0 (launch_build_libvirt_appliance_start);

  if (g->verbose)
    guestfs___print_timestamped_message (g, "build appliance");

  if (guestfs___build_appliance (g, &params.kernel, &params.initrd,
                                 &appliance) == -1)
    goto cleanup;

  guestfs___launch_send_progress (g, 3);
  TRACE0 (launch_build_libvirt_appliance_end);

  /* Create overlays for read-only drives and the appliance.  This
   * works around lack of support for <transient/> disks in libvirt.
   * Note that appliance can be NULL if using the old-style appliance.
   */
  if (appliance) {
    params.appliance_overlay = make_qcow2_overlay (g, appliance, "raw");
    if (!params.appliance_overlay)
      goto cleanup;
  }

  ITER_DRIVES (g, i, drv) {
    if (make_qcow2_overlay_for_drive (g, drv) == -1)
      goto cleanup;
  }

  TRACE0 (launch_build_libvirt_qcow2_overlay_end);

  /* Using virtio-serial, we need to create a local Unix domain socket
   * for qemu to connect to.
   */
  snprintf (params.guestfsd_sock, sizeof params.guestfsd_sock,
            "%s/guestfsd.sock", g->tmpdir);
  unlink (params.guestfsd_sock);

  g->sock = socket (AF_UNIX, SOCK_STREAM|SOCK_CLOEXEC, 0);
  if (g->sock == -1) {
    perrorf (g, "socket");
    goto cleanup;
  }

  if (fcntl (g->sock, F_SETFL, O_NONBLOCK) == -1) {
    perrorf (g, "fcntl");
    goto cleanup;
  }

  addr.sun_family = AF_UNIX;
  memcpy (addr.sun_path, params.guestfsd_sock, UNIX_PATH_MAX);

  if (bind (g->sock, &addr, sizeof addr) == -1) {
    perrorf (g, "bind");
    goto cleanup;
  }

  if (listen (g->sock, 1) == -1) {
    perrorf (g, "listen");
    goto cleanup;
  }

  /* For the serial console. */
  snprintf (params.console_sock, sizeof params.console_sock,
            "%s/console.sock", g->tmpdir);
  unlink (params.console_sock);

  console = socket (AF_UNIX, SOCK_STREAM|SOCK_CLOEXEC, 0);
  if (console == -1) {
    perrorf (g, "socket");
    goto cleanup;
  }

  addr.sun_family = AF_UNIX;
  memcpy (addr.sun_path, params.console_sock, UNIX_PATH_MAX);

  if (bind (console, &addr, sizeof addr) == -1) {
    perrorf (g, "bind");
    goto cleanup;
  }

  if (listen (console, 1) == -1) {
    perrorf (g, "listen");
    goto cleanup;
  }

  /* libvirt, if running as root, will run the qemu process as
   * qemu.qemu, which means it won't be able to access the socket.
   * There are roughly three things that get in the way:
   * (1) Permissions of the socket.
   * (2) Permissions of the parent directory(-ies).  Remember this
   *     if $TMPDIR is located in your home directory.
   * (3) SELinux/sVirt will prevent access.  libvirt ought to
   *     label the socket.
   */
  if (is_root) {
    struct group *grp;

    if (chmod (params.guestfsd_sock, 0775) == -1) {
      perrorf (g, "chmod: %s", params.guestfsd_sock);
      goto cleanup;
    }

    if (chmod (params.console_sock, 0775) == -1) {
      perrorf (g, "chmod: %s", params.console_sock);
      goto cleanup;
    }

    grp = getgrnam ("qemu");
    if (grp != NULL) {
      if (chown (params.guestfsd_sock, 0, grp->gr_gid) == -1) {
        perrorf (g, "chown: %s", params.guestfsd_sock);
        goto cleanup;
      }
      if (chown (params.console_sock, 0, grp->gr_gid) == -1) {
        perrorf (g, "chown: %s", params.console_sock);
        goto cleanup;
      }
    } else
      debug (g, "cannot find group 'qemu'");
  }

  /* Construct the libvirt XML. */
  if (g->verbose)
    guestfs___print_timestamped_message (g, "create libvirt XML");

  params.appliance_index = g->nr_drives;
  strcpy (params.appliance_dev, "/dev/sd");
  guestfs___drive_name (params.appliance_index, &params.appliance_dev[7]);
  params.enable_svirt = ! is_custom_qemu (g);
  /* XXX big hack, instead of actually parsing the capabilities XML. */
  params.is_kvm = strstr (params.capabilities_xml, "'kvm'") != NULL;

  xml = construct_libvirt_xml (g, &params);
  if (!xml)
    goto cleanup;

  /* Launch the libvirt guest. */
  if (g->verbose)
    guestfs___print_timestamped_message (g, "launch libvirt guest");

  dom = virDomainCreateXML (conn, (char *) xml, VIR_DOMAIN_START_AUTODESTROY);
  if (!dom) {
    libvirt_error (g, _("could not create appliance through libvirt"));
    goto cleanup;
  }

  g->state = LAUNCHING;

  /* Wait for console socket to open. */
  r = accept4 (console, NULL, NULL, SOCK_NONBLOCK|SOCK_CLOEXEC);
  if (r == -1) {
    perrorf (g, "accept");
    goto cleanup;
  }
  if (close (console) == -1) {
    perrorf (g, "close: console socket");
    close (r);
    goto cleanup;
  }
  g->fd[0] = r; /* This is the accepted console socket. */

  g->fd[1] = dup (g->fd[0]);
  if (g->fd[1] == -1) {
    perrorf (g, "dup");
    goto cleanup;
  }

  /* Wait for libvirt domain to start and to connect back to us via
   * virtio-serial and send the GUESTFS_LAUNCH_FLAG message.
   */
  r = guestfs___accept_from_daemon (g);
  if (r == -1)
    goto cleanup;

  /* NB: We reach here just because qemu has opened the socket.  It
   * does not mean the daemon is up until we read the
   * GUESTFS_LAUNCH_FLAG below.  Failures in qemu startup can still
   * happen even if we reach here, even early failures like not being
   * able to open a drive.
   */

  /* Close the listening socket. */
  if (close (g->sock) == -1) {
    perrorf (g, "close: listening socket");
    close (r);
    g->sock = -1;
    goto cleanup;
  }
  g->sock = r; /* This is the accepted data socket. */

  if (fcntl (g->sock, F_SETFL, O_NONBLOCK) == -1) {
    perrorf (g, "fcntl");
    goto cleanup;
  }

  r = guestfs___recv_from_daemon (g, &size, &buf);
  free (buf);

  if (r == -1) {
    guestfs___launch_failed_error (g);
    goto cleanup;
  }

  if (size != GUESTFS_LAUNCH_FLAG) {
    guestfs___launch_failed_error (g);
    goto cleanup;
  }

  if (g->verbose)
    guestfs___print_timestamped_message (g, "appliance is up");

  /* This is possible in some really strange situations, such as
   * guestfsd starts up OK but then qemu immediately exits.  Check for
   * it because the caller is probably expecting to be able to send
   * commands after this function returns.
   */
  if (g->state != READY) {
    error (g, _("qemu launched and contacted daemon, but state != READY"));
    goto cleanup;
  }

  if (appliance)
    guestfs___add_dummy_appliance_drive (g);

  TRACE0 (launch_libvirt_end);

  guestfs___launch_send_progress (g, 12);

  g->virt.conn = conn;
  g->virt.dom = dom;

  free (params.kernel);
  free (params.initrd);
  free (appliance);
  free (params.appliance_overlay);
  free (params.capabilities_xml);
  free (xml);

  return 0;

 cleanup:
  if (console >= 0)
    close (console);
  if (g->fd[0] >= 0) {
    close (g->fd[0]);
    g->fd[0] = -1;
  }
  if (g->fd[1] >= 0) {
    close (g->fd[1]);
    g->fd[1] = -1;
  }
  if (g->sock >= 0) {
    close (g->sock);
    g->sock = -1;
  }

  if (dom) {
    virDomainDestroy (dom);
    virDomainFree (dom);
  }
  if (conn)
    virConnectClose (conn);

  free (params.kernel);
  free (params.initrd);
  free (appliance);
  free (params.appliance_overlay);
  free (params.capabilities_xml);
  free (xml);

  g->state = CONFIG;

  return -1;
}

static int
is_custom_qemu (guestfs_h *g)
{
  return g->qemu && STRNEQ (g->qemu, QEMU);
}

static int construct_libvirt_xml_name (guestfs_h *g, const struct libvirt_xml_params *params, xmlTextWriterPtr xo);
static int construct_libvirt_xml_cpu (guestfs_h *g, const struct libvirt_xml_params *params, xmlTextWriterPtr xo);
static int construct_libvirt_xml_boot (guestfs_h *g, const struct libvirt_xml_params *params, xmlTextWriterPtr xo);
static int construct_libvirt_xml_seclabel (guestfs_h *g, const struct libvirt_xml_params *params, xmlTextWriterPtr xo);
static int construct_libvirt_xml_lifecycle (guestfs_h *g, const struct libvirt_xml_params *params, xmlTextWriterPtr xo);
static int construct_libvirt_xml_devices (guestfs_h *g, const struct libvirt_xml_params *params, xmlTextWriterPtr xo);
static int construct_libvirt_xml_qemu_cmdline (guestfs_h *g, const struct libvirt_xml_params *params, xmlTextWriterPtr xo);
static int construct_libvirt_xml_disk (guestfs_h *g, xmlTextWriterPtr xo, struct drive *drv, size_t drv_index);
static int construct_libvirt_xml_appliance (guestfs_h *g, const struct libvirt_xml_params *params, xmlTextWriterPtr xo);

#define XMLERROR(code,e) do {                                           \
    if ((e) == (code)) {                                                \
      perrorf (g, _("error constructing libvirt XML at \"%s\""),        \
               #e);                                                     \
      goto err;                                                         \
    }                                                                   \
  } while (0)

static xmlChar *
construct_libvirt_xml (guestfs_h *g, const struct libvirt_xml_params *params)
{
  xmlChar *ret = NULL;
  xmlBufferPtr xb = NULL;
  xmlOutputBufferPtr ob;
  xmlTextWriterPtr xo = NULL;

  XMLERROR (NULL, xb = xmlBufferCreate ());
  XMLERROR (NULL, ob = xmlOutputBufferCreateBuffer (xb, NULL));
  XMLERROR (NULL, xo = xmlNewTextWriter (ob));

  XMLERROR (-1, xmlTextWriterSetIndent (xo, 1));
  XMLERROR (-1, xmlTextWriterSetIndentString (xo, BAD_CAST "  "));
  XMLERROR (-1, xmlTextWriterStartDocument (xo, NULL, NULL, NULL));

  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "domain"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "type",
                         params->is_kvm ? BAD_CAST "kvm" : BAD_CAST "qemu"));
  XMLERROR (-1,
            xmlTextWriterWriteAttributeNS (xo,
                                           BAD_CAST "xmlns",
                                           BAD_CAST "qemu",
                                           NULL,
                                           BAD_CAST "http://libvirt.org/schemas/domain/qemu/1.0"));

  if (construct_libvirt_xml_name (g, params, xo) == -1)
    goto err;
  if (construct_libvirt_xml_cpu (g, params, xo) == -1)
    goto err;
  if (construct_libvirt_xml_boot (g, params, xo) == -1)
    goto err;
  if (!params->enable_svirt)
    if (construct_libvirt_xml_seclabel (g, params, xo) == -1)
      goto err;
  if (construct_libvirt_xml_lifecycle (g, params, xo) == -1)
    goto err;
  if (construct_libvirt_xml_devices (g, params, xo) == -1)
    goto err;
  if (construct_libvirt_xml_qemu_cmdline (g, params, xo) == -1)
    goto err;

  XMLERROR (-1, xmlTextWriterEndElement (xo));

  XMLERROR (-1, xmlTextWriterEndDocument (xo));
  XMLERROR (NULL, ret = xmlBufferDetach (xb)); /* caller frees ret */

  debug (g, "libvirt XML:\n%s", ret);

 err:
  if (xo)
    xmlFreeTextWriter (xo); /* frees 'ob' too */
  if (xb)
    xmlBufferFree (xb);

  return ret;
}

/* Construct a securely random name.  We don't need to save the name
 * because if we ever needed it, it's available from libvirt.
 */
#define DOMAIN_NAME_LEN 16

static int
construct_libvirt_xml_name (guestfs_h *g,
                            const struct libvirt_xml_params *params,
                            xmlTextWriterPtr xo)
{
  char name[DOMAIN_NAME_LEN+1];

  if (random_chars (name, DOMAIN_NAME_LEN) == -1) {
    perrorf (g, "/dev/urandom");
    goto err;
  }

  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "name"));
  XMLERROR (-1, xmlTextWriterWriteFormatString (xo, "guestfs-%s", name));
  XMLERROR (-1, xmlTextWriterEndElement (xo));

  return 0;

 err:
  return -1;
}

/* CPU and memory features. */
static int
construct_libvirt_xml_cpu (guestfs_h *g,
                           const struct libvirt_xml_params *params,
                           xmlTextWriterPtr xo)
{
  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "memory"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "unit", BAD_CAST "MiB"));
  XMLERROR (-1, xmlTextWriterWriteFormatString (xo, "%d", g->memsize));
  XMLERROR (-1, xmlTextWriterEndElement (xo));

  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "currentMemory"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "unit", BAD_CAST "MiB"));
  XMLERROR (-1, xmlTextWriterWriteFormatString (xo, "%d", g->memsize));
  XMLERROR (-1, xmlTextWriterEndElement (xo));

  /* It would be faster to pass the CPU host model to the appliance,
   * allowing maximum speed for things like checksums, encryption.
   * However this doesn't work because KVM doesn't emulate all of the
   * required guest insns (RHBZ#870071).  This is why the following
   * section is commented out.
   */
#if 0
  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "cpu"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "mode",
                                         BAD_CAST "host-model"));
  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "model"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "fallback",
                                         BAD_CAST "allow"));
  XMLERROR (-1, xmlTextWriterEndElement (xo));
  XMLERROR (-1, xmlTextWriterEndElement (xo));
#endif

  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "vcpu"));
  XMLERROR (-1, xmlTextWriterWriteFormatString (xo, "%d", g->smp));
  XMLERROR (-1, xmlTextWriterEndElement (xo));

  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "clock"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "offset",
                                         BAD_CAST "utc"));
  XMLERROR (-1, xmlTextWriterEndElement (xo));

  return 0;

 err:
  return -1;
}

/* Boot parameters. */
static int
construct_libvirt_xml_boot (guestfs_h *g,
                            const struct libvirt_xml_params *params,
                            xmlTextWriterPtr xo)
{
  char *cmdline;
  int flags;

  /* Linux kernel command line. */
  flags = 0;
  if (!params->is_kvm)
    flags |= APPLIANCE_COMMAND_LINE_IS_TCG;
  cmdline = guestfs___appliance_command_line (g, params->appliance_dev, flags);

  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "os"));

  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "type"));
  XMLERROR (-1, xmlTextWriterWriteString (xo, BAD_CAST "hvm"));
  XMLERROR (-1, xmlTextWriterEndElement (xo));

  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "kernel"));
  XMLERROR (-1, xmlTextWriterWriteString (xo, BAD_CAST params->kernel));
  XMLERROR (-1, xmlTextWriterEndElement (xo));

  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "initrd"));
  XMLERROR (-1, xmlTextWriterWriteString (xo, BAD_CAST params->initrd));
  XMLERROR (-1, xmlTextWriterEndElement (xo));

  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "cmdline"));
  XMLERROR (-1, xmlTextWriterWriteString (xo, BAD_CAST cmdline));
  XMLERROR (-1, xmlTextWriterEndElement (xo));

  XMLERROR (-1, xmlTextWriterEndElement (xo));

  free (cmdline);
  return 0;

 err:
  free (cmdline);
  return -1;
}

static int
construct_libvirt_xml_seclabel (guestfs_h *g,
                                const struct libvirt_xml_params *params,
                                xmlTextWriterPtr xo)
{
  /* This disables SELinux/sVirt confinement. */
  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "seclabel"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "type",
                                         BAD_CAST "none"));
  XMLERROR (-1, xmlTextWriterEndElement (xo));

  return 0;

 err:
  return -1;
}

/* qemu -no-reboot */
static int
construct_libvirt_xml_lifecycle (guestfs_h *g,
                                 const struct libvirt_xml_params *params,
                                 xmlTextWriterPtr xo)
{
  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "on_reboot"));
  XMLERROR (-1, xmlTextWriterWriteString (xo, BAD_CAST "destroy"));
  XMLERROR (-1, xmlTextWriterEndElement (xo));

  return 0;

 err:
  return -1;
}

/* Devices. */
static int
construct_libvirt_xml_devices (guestfs_h *g,
                               const struct libvirt_xml_params *params,
                               xmlTextWriterPtr xo)
{
  struct drive *drv;
  size_t i;

  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "devices"));

  /* Path to qemu.  Only write this if the user has changed the
   * default, otherwise allow libvirt to choose the best one.
   */
  if (is_custom_qemu (g)) {
    XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "emulator"));
    XMLERROR (-1, xmlTextWriterWriteString (xo, BAD_CAST g->qemu));
    XMLERROR (-1, xmlTextWriterEndElement (xo));
  }

  /* virtio-scsi controller. */
  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "controller"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "type",
                                         BAD_CAST "scsi"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "index",
                                         BAD_CAST "0"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "model",
                                         BAD_CAST "virtio-scsi"));
  XMLERROR (-1, xmlTextWriterEndElement (xo));

  /* Disks. */
  ITER_DRIVES (g, i, drv) {
    if (construct_libvirt_xml_disk (g, xo, drv, i) == -1)
      goto err;
  }

  if (params->appliance_overlay) {
    /* Appliance disk. */
    if (construct_libvirt_xml_appliance (g, params, xo) == -1)
      goto err;
  }

  /* Console. */
  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "serial"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "type",
                                         BAD_CAST "unix"));
  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "source"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "mode",
                                         BAD_CAST "connect"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "path",
                                         BAD_CAST params->console_sock));
  XMLERROR (-1, xmlTextWriterEndElement (xo));
  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "target"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "port",
                                         BAD_CAST "0"));
  XMLERROR (-1, xmlTextWriterEndElement (xo));
  XMLERROR (-1, xmlTextWriterEndElement (xo));

  /* Virtio-serial for guestfsd communication. */
  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "channel"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "type",
                                         BAD_CAST "unix"));
  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "source"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "mode",
                                         BAD_CAST "connect"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "path",
                                         BAD_CAST params->guestfsd_sock));
  XMLERROR (-1, xmlTextWriterEndElement (xo));
  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "target"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "type",
                                         BAD_CAST "virtio"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "name",
                                         BAD_CAST "org.libguestfs.channel.0"));
  XMLERROR (-1, xmlTextWriterEndElement (xo));
  XMLERROR (-1, xmlTextWriterEndElement (xo));

  XMLERROR (-1, xmlTextWriterEndElement (xo));

  return 0;

 err:
  return -1;
}

static int
construct_libvirt_xml_disk (guestfs_h *g,
                            xmlTextWriterPtr xo,
                            struct drive *drv, size_t drv_index)
{
  char drive_name[64] = "sd";
  char scsi_target[64];
  struct drive_libvirt *drv_priv;
  char *format = NULL;
  int is_host_device;

  /* XXX We probably could support this if we thought about it some more. */
  if (drv->iface) {
    error (g, _("'iface' parameter is not supported by the libvirt attach-method"));
    goto err;
  }

  guestfs___drive_name (drv_index, &drive_name[2]);
  snprintf (scsi_target, sizeof scsi_target, "%zu", drv_index);

  drv_priv = (struct drive_libvirt *) drv->priv;

  /* Change the libvirt XML according to whether the host path is
   * a device or a file.  For devices, use:
   *   <disk type=block device=disk>
   *     <source dev=[path]>
   * For files, use:
   *   <disk type=file device=disk>
   *     <source file=[path]>
   */
  is_host_device = is_blk (drv_priv->path);

  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "disk"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "device",
                                         BAD_CAST "disk"));
  if (!is_host_device) {
    XMLERROR (-1,
              xmlTextWriterWriteAttribute (xo, BAD_CAST "type",
                                           BAD_CAST "file"));

    XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "source"));
    XMLERROR (-1,
              xmlTextWriterWriteAttribute (xo, BAD_CAST "file",
                                           BAD_CAST drv_priv->path));
    XMLERROR (-1, xmlTextWriterEndElement (xo));
  }
  else {
    XMLERROR (-1,
              xmlTextWriterWriteAttribute (xo, BAD_CAST "type",
                                           BAD_CAST "block"));

    XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "source"));
    XMLERROR (-1,
              xmlTextWriterWriteAttribute (xo, BAD_CAST "dev",
                                           BAD_CAST drv_priv->path));
    XMLERROR (-1, xmlTextWriterEndElement (xo));
  }

  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "target"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "dev",
                                         BAD_CAST drive_name));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "bus",
                                         BAD_CAST "scsi"));
  XMLERROR (-1, xmlTextWriterEndElement (xo));

  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "driver"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "name",
                                         BAD_CAST "qemu"));
  if (drv_priv->format) {
    XMLERROR (-1,
              xmlTextWriterWriteAttribute (xo, BAD_CAST "type",
                                           BAD_CAST drv_priv->format));
  }
  else {
    /* libvirt has disabled the feature of detecting the disk format,
     * unless the administrator sets allow_disk_format_probing=1 in
     * qemu.conf.  There is no way to detect if this option is set, so we
     * have to do format detection here using qemu-img and pass that to
     * libvirt.
     *
     * This is still a security issue, so in most cases it is recommended
     * the users pass the format to libguestfs which will faithfully pass
     * that to libvirt and this function won't be used.
     */
    format = guestfs_disk_format (g, drv_priv->path);
    if (!format)
      goto err;

    if (STREQ (format, "unknown")) {
      error (g, _("could not auto-detect the format of '%s'\n"
                  "If the format is known, pass the format to libguestfs, eg. using the\n"
                  "'--format' option, or via the optional 'format' argument to 'add-drive'."),
             drv->path);
      goto err;
    }

    XMLERROR (-1,
              xmlTextWriterWriteAttribute (xo, BAD_CAST "type",
                                           BAD_CAST format));
  }
  if (drv->use_cache_none) {
    XMLERROR (-1,
              xmlTextWriterWriteAttribute (xo, BAD_CAST "cache",
                                           BAD_CAST "none"));
  }
  XMLERROR (-1, xmlTextWriterEndElement (xo));

  if (drv->disk_label) {
    XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "serial"));
    XMLERROR (-1, xmlTextWriterWriteString (xo, BAD_CAST drv->disk_label));
    XMLERROR (-1, xmlTextWriterEndElement (xo));
  }

  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "address"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "type",
                                         BAD_CAST "drive"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "controller",
                                         BAD_CAST "0"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "bus",
                                         BAD_CAST "0"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "target",
                                         BAD_CAST scsi_target));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "unit",
                                         BAD_CAST "0"));
  XMLERROR (-1, xmlTextWriterEndElement (xo));

  XMLERROR (-1, xmlTextWriterEndElement (xo));

  free (format);
  return 0;

 err:
  free (format);
  return -1;
}

static int
construct_libvirt_xml_appliance (guestfs_h *g,
                                 const struct libvirt_xml_params *params,
                                 xmlTextWriterPtr xo)
{
  char scsi_target[64];

  snprintf (scsi_target, sizeof scsi_target, "%zu", params->appliance_index);

  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "disk"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "type",
                                         BAD_CAST "file"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "device",
                                         BAD_CAST "disk"));

  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "source"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "file",
                                         BAD_CAST params->appliance_overlay));
  XMLERROR (-1, xmlTextWriterEndElement (xo));

  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "target"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "dev",
                                         BAD_CAST &params->appliance_dev[5]));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "bus",
                                         BAD_CAST "scsi"));
  XMLERROR (-1, xmlTextWriterEndElement (xo));

  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "driver"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "name",
                                         BAD_CAST "qemu"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "type",
                                         BAD_CAST "qcow2"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "cache",
                                         BAD_CAST "unsafe"));
  XMLERROR (-1, xmlTextWriterEndElement (xo));

  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "address"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "type",
                                         BAD_CAST "drive"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "controller",
                                         BAD_CAST "0"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "bus",
                                         BAD_CAST "0"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "target",
                                         BAD_CAST scsi_target));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "unit",
                                         BAD_CAST "0"));
  XMLERROR (-1, xmlTextWriterEndElement (xo));

  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "shareable"));
  XMLERROR (-1, xmlTextWriterEndElement (xo));

  /* We'd like to do this, but it's not supported by libvirt.
   * See construct_libvirt_xml_qemu_cmdline for the workaround.
   *
   * XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "transient"));
   * XMLERROR (-1, xmlTextWriterEndElement (xo));
   */

  XMLERROR (-1, xmlTextWriterEndElement (xo));

  return 0;

 err:
  return -1;
}

static int
construct_libvirt_xml_qemu_cmdline (guestfs_h *g,
                                    const struct libvirt_xml_params *params,
                                    xmlTextWriterPtr xo)
{
  struct qemu_param *qp;
  char *tmpdir;

  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "qemu:commandline"));

  /* We need to ensure the snapshots are created in the persistent
   * temporary directory (RHBZ#856619).  We must set one, because
   * otherwise libvirt will use a random TMPDIR (RHBZ#865464).
   */
  tmpdir = guestfs_get_cachedir (g);

  XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "qemu:env"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "name",
                                         BAD_CAST "TMPDIR"));
  XMLERROR (-1,
            xmlTextWriterWriteAttribute (xo, BAD_CAST "value",
                                         BAD_CAST tmpdir));
  XMLERROR (-1, xmlTextWriterEndElement (xo));

  free (tmpdir);

  /* Workaround because libvirt user networking cannot specify "net="
   * parameter.
   */
  if (g->enable_network) {
    XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "qemu:arg"));
    XMLERROR (-1,
              xmlTextWriterWriteAttribute (xo, BAD_CAST "value",
                                           BAD_CAST "-netdev"));
    XMLERROR (-1, xmlTextWriterEndElement (xo));

    XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "qemu:arg"));
    XMLERROR (-1,
              xmlTextWriterWriteAttribute (xo, BAD_CAST "value",
                                           BAD_CAST "user,id=usernet,net=169.254.0.0/16"));
    XMLERROR (-1, xmlTextWriterEndElement (xo));

    XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "qemu:arg"));
    XMLERROR (-1,
              xmlTextWriterWriteAttribute (xo, BAD_CAST "value",
                                           BAD_CAST "-device"));
    XMLERROR (-1, xmlTextWriterEndElement (xo));

    XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "qemu:arg"));
    XMLERROR (-1,
              xmlTextWriterWriteAttribute (xo, BAD_CAST "value",
                                           BAD_CAST "virtio-net-pci,netdev=usernet"));
    XMLERROR (-1, xmlTextWriterEndElement (xo));
  }

  /* The qemu command line arguments requested by the caller. */
  for (qp = g->qemu_params; qp; qp = qp->next) {
    XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "qemu:arg"));
    XMLERROR (-1,
              xmlTextWriterWriteAttribute (xo, BAD_CAST "value",
                                           BAD_CAST qp->qemu_param));
    XMLERROR (-1, xmlTextWriterEndElement (xo));

    if (qp->qemu_value) {
      XMLERROR (-1, xmlTextWriterStartElement (xo, BAD_CAST "qemu:arg"));
      XMLERROR (-1,
                xmlTextWriterWriteAttribute (xo, BAD_CAST "value",
                                             BAD_CAST qp->qemu_value));
      XMLERROR (-1, xmlTextWriterEndElement (xo));
    }
  }

  XMLERROR (-1, xmlTextWriterEndElement (xo));

  return 0;

 err:
  return -1;
}

static int
is_blk (const char *path)
{
  struct stat statbuf;

  if (stat (path, &statbuf) == -1)
    return 0;
  return S_ISBLK (statbuf.st_mode);
}

static int
random_chars (char *ret, size_t len)
{
  int fd;
  size_t i;
  unsigned char c;
  int saved_errno;

  fd = open ("/dev/urandom", O_RDONLY|O_CLOEXEC);
  if (fd == -1)
    return -1;

  for (i = 0; i < len; ++i) {
    if (read (fd, &c, 1) != 1) {
      saved_errno = errno;
      close (fd);
      errno = saved_errno;
      return -1;
    }
    ret[i] = "0123456789abcdefghijklmnopqrstuvwxyz"[c % 36];
  }
  ret[len] = '\0';

  if (close (fd) == -1)
    return -1;

  return 0;
}

static void
ignore_errors (void *ignore, virErrorPtr ignore2)
{
  /* empty */
}

/* Create a temporary qcow2 overlay on top of 'path'. */
static char *
make_qcow2_overlay (guestfs_h *g, const char *path, const char *format)
{
  char *tmpfile;
  struct command *cmd;
  int r;

  /* Path must be absolute. */
  assert (path);
  assert (path[0] == '/');

  tmpfile = safe_asprintf (g, "%s/snapshot%d", g->tmpdir, ++g->unique);

  cmd = guestfs___new_command (g);
  guestfs___cmd_add_arg (cmd, "qemu-img");
  guestfs___cmd_add_arg (cmd, "create");
  guestfs___cmd_add_arg (cmd, "-f");
  guestfs___cmd_add_arg (cmd, "qcow2");
  guestfs___cmd_add_arg (cmd, "-b");
  guestfs___cmd_add_arg (cmd, path);
  if (format) {
    guestfs___cmd_add_arg (cmd, "-o");
    guestfs___cmd_add_arg_format (cmd, "backing_fmt=%s", format);
  }
  guestfs___cmd_add_arg (cmd, tmpfile);
  r = guestfs___cmd_run (cmd);
  guestfs___cmd_close (cmd);
  if (r == -1)
    goto error;
  if (!WIFEXITED (r) || WEXITSTATUS (r) != 0) {
    error (g, _("qemu-img create: could not create snapshot over %s"), path);
    goto error;
  }

  return tmpfile;               /* caller frees */

 error:
  free (tmpfile);

  return NULL;
}

static int
make_qcow2_overlay_for_drive (guestfs_h *g, struct drive *drv)
{
  char *path;
  struct drive_libvirt *drv_priv;

  if (drv->priv && drv->free_priv)
    drv->free_priv (drv->priv);

  drv->priv = drv_priv = safe_calloc (g, 1, sizeof (struct drive_libvirt));
  drv->free_priv = drive_free_priv;

  /* Even for non-readonly paths, we need to make the paths absolute here. */
  path = realpath (drv->path, NULL);
  if (path == NULL) {
    perrorf (g, _("realpath: could not convert '%s' to absolute path"),
             drv->path);
    return -1;
  }

  if (!drv->readonly) {
    drv_priv->path = path;
    drv_priv->format = drv->format ? safe_strdup (g, drv->format) : NULL;
  }
  else {
    drv_priv->path = make_qcow2_overlay (g, path, drv->format);
    free (path);
    if (!drv_priv->path)
      return -1;
    drv_priv->format = safe_strdup (g, "qcow2");
  }

  return 0;
}

static void
drive_free_priv (void *priv)
{
  struct drive_libvirt *drv_priv = priv;

  free (drv_priv->path);
  free (drv_priv->format);
  free (drv_priv);
}

static int
shutdown_libvirt (guestfs_h *g, int check_for_errors)
{
  virConnectPtr conn = g->virt.conn;
  virDomainPtr dom = g->virt.dom;
  int ret = 0;
  int flags;

  /* Note that we can be called back very early in launch (specifically
   * from launch_libvirt itself), when conn and dom might be NULL.
   */

  if (dom != NULL) {
    flags = check_for_errors ? VIR_DOMAIN_DESTROY_GRACEFUL : 0;
    if (virDomainDestroyFlags (dom, flags) == -1) {
      libvirt_error (g, _("could not destroy libvirt domain"));
      ret = -1;
    }
    virDomainFree (dom);
  }

  if (conn != NULL)
    virConnectClose (conn);

  g->virt.conn = NULL;
  g->virt.dom = NULL;

  return ret;
}

/* Wrapper around error() which produces better errors for
 * libvirt functions.
 */
static void
libvirt_error (guestfs_h *g, const char *fs, ...)
{
  va_list args;
  char *msg;
  int len;
  virErrorPtr err;

  va_start (args, fs);
  len = vasprintf (&msg, fs, args);
  va_end (args);

  if (len < 0)
    msg = safe_asprintf (g,
                         _("%s: internal error forming error message"),
                         __func__);

  /* In all recent libvirt, this retrieves the thread-local error. */
  err = virGetLastError ();
  if (err)
    error (g, "%s: %s [code=%d domain=%d]",
           msg, err->message, err->code, err->domain);
  else
    error (g, "%s", msg);

  /* NB. 'err' must not be freed! */
  free (msg);
}

/* This backend assumes virtio-scsi is available. */
static int
max_disks_libvirt (guestfs_h *g)
{
  return 255;
}

static xmlChar *construct_libvirt_xml_hot_add_disk (guestfs_h *g, struct drive *drv, size_t drv_index);

/* Hot-add a drive.  Note the appliance is up when this is called. */
static int
hot_add_drive_libvirt (guestfs_h *g, struct drive *drv, size_t drv_index)
{
  virConnectPtr conn = g->virt.conn;
  virDomainPtr dom = g->virt.dom;
  xmlChar *xml = NULL;

  if (!conn || !dom) {
    /* This is essentially an internal error if it happens. */
    error (g, "%s: conn == NULL or dom == NULL", __func__);
    return -1;
  }

  /* Create overlay for read-only drive.  This works around lack of
   * support for <transient/> disks in libvirt.
   */
  if (make_qcow2_overlay_for_drive (g, drv) == -1)
    return -1;

  /* Create the XML for the new disk. */
  xml = construct_libvirt_xml_hot_add_disk (g, drv, drv_index);
  if (xml == NULL)
    return -1;

  /* Attach it. */
  if (virDomainAttachDeviceFlags (dom, (char *) xml,
                                  VIR_DOMAIN_DEVICE_MODIFY_LIVE) == -1) {
    libvirt_error (g, _("could not attach disk to libvirt domain"));
    goto error;
  }

  free (xml);
  return 0;

 error:
  free (xml);
  return -1;
}

/* Hot-remove a drive.  Note the appliance is up when this is called. */
static int
hot_remove_drive_libvirt (guestfs_h *g, struct drive *drv, size_t drv_index)
{
  virConnectPtr conn = g->virt.conn;
  virDomainPtr dom = g->virt.dom;
  xmlChar *xml = NULL;

  if (!conn || !dom) {
    /* This is essentially an internal error if it happens. */
    error (g, "%s: conn == NULL or dom == NULL", __func__);
    return -1;
  }

  /* Re-create the XML for the disk. */
  xml = construct_libvirt_xml_hot_add_disk (g, drv, drv_index);
  if (xml == NULL)
    return -1;

  /* Detach it. */
  if (virDomainDetachDeviceFlags (dom, (char *) xml,
                                  VIR_DOMAIN_DEVICE_MODIFY_LIVE) == -1) {
    libvirt_error (g, _("could not detach disk from libvirt domain"));
    goto error;
  }

  free (xml);
  return 0;

 error:
  free (xml);
  return -1;
}

static xmlChar *
construct_libvirt_xml_hot_add_disk (guestfs_h *g, struct drive *drv,
                                    size_t drv_index)
{
  xmlChar *ret = NULL;
  xmlBufferPtr xb = NULL;
  xmlOutputBufferPtr ob;
  xmlTextWriterPtr xo = NULL;

  XMLERROR (NULL, xb = xmlBufferCreate ());
  XMLERROR (NULL, ob = xmlOutputBufferCreateBuffer (xb, NULL));
  XMLERROR (NULL, xo = xmlNewTextWriter (ob));

  XMLERROR (-1, xmlTextWriterSetIndent (xo, 1));
  XMLERROR (-1, xmlTextWriterSetIndentString (xo, BAD_CAST "  "));
  XMLERROR (-1, xmlTextWriterStartDocument (xo, NULL, NULL, NULL));

  if (construct_libvirt_xml_disk (g, xo, drv, drv_index) == -1)
    goto err;

  XMLERROR (-1, xmlTextWriterEndDocument (xo));
  XMLERROR (NULL, ret = xmlBufferDetach (xb)); /* caller frees ret */

  debug (g, "hot-add disk XML:\n%s", ret);

 err:
  if (xo)
    xmlFreeTextWriter (xo); /* frees 'ob' too */
  if (xb)
    xmlBufferFree (xb);

  return ret;
}

struct attach_ops attach_ops_libvirt = {
  .launch = launch_libvirt,
  .shutdown = shutdown_libvirt,
  .max_disks = max_disks_libvirt,
  .hot_add_drive = hot_add_drive_libvirt,
  .hot_remove_drive = hot_remove_drive_libvirt,
};

#else /* no libvirt or libxml2 at compile time */

#define NOT_IMPL(r)                                                     \
  error (g, _("libvirt attach-method is not available because "         \
              "this version of libguestfs was compiled "                \
              "without libvirt or libvirt < %d.%d.%d or without libxml2"), \
         MIN_LIBVIRT_MAJOR, MIN_LIBVIRT_MINOR, MIN_LIBVIRT_MICRO);      \
  return r

static int
launch_libvirt (guestfs_h *g, const char *arg)
{
  NOT_IMPL (-1);
}

static int
shutdown_libvirt (guestfs_h *g, int check_for_errors)
{
  NOT_IMPL (-1);
}

static int
max_disks_libvirt (guestfs_h *g)
{
  NOT_IMPL (-1);
}

struct attach_ops attach_ops_libvirt = {
  .launch = launch_libvirt,
  .shutdown = shutdown_libvirt,
  .max_disks = max_disks_libvirt,
};

#endif /* no libvirt or libxml2 at compile time */