summaryrefslogtreecommitdiffstats
path: root/src/shared/dissect-image.c
blob: a9e211f246a1f77bd3bd2006958391e2aa596429 (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
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786
2787
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
2807
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
2824
2825
2826
2827
2828
2829
2830
2831
2832
2833
2834
2835
2836
2837
2838
2839
2840
2841
2842
2843
2844
2845
2846
2847
2848
2849
2850
2851
2852
2853
2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
2867
2868
2869
2870
2871
2872
2873
2874
2875
2876
2877
2878
2879
2880
2881
2882
2883
2884
2885
2886
2887
2888
2889
2890
2891
2892
2893
2894
2895
2896
2897
2898
2899
2900
2901
2902
2903
2904
2905
2906
2907
2908
2909
2910
2911
2912
2913
2914
2915
2916
2917
2918
2919
2920
2921
2922
2923
2924
2925
2926
2927
2928
2929
2930
2931
2932
2933
2934
2935
2936
2937
2938
2939
2940
2941
2942
2943
2944
2945
2946
2947
2948
2949
2950
2951
2952
2953
2954
2955
2956
2957
2958
2959
2960
2961
2962
2963
2964
2965
2966
2967
2968
2969
2970
2971
2972
2973
2974
2975
2976
2977
2978
2979
2980
2981
2982
2983
2984
2985
2986
2987
2988
2989
2990
2991
2992
2993
2994
2995
2996
2997
2998
2999
3000
3001
3002
3003
3004
3005
3006
3007
3008
3009
3010
3011
3012
3013
3014
3015
3016
3017
3018
3019
3020
3021
3022
3023
3024
3025
3026
3027
3028
3029
3030
3031
3032
3033
3034
3035
3036
3037
3038
3039
3040
3041
3042
3043
3044
3045
3046
3047
3048
3049
3050
3051
3052
3053
3054
3055
3056
3057
3058
3059
3060
3061
3062
3063
3064
3065
3066
3067
3068
3069
3070
3071
3072
3073
3074
3075
3076
3077
3078
3079
3080
3081
3082
3083
3084
3085
3086
3087
3088
3089
3090
3091
3092
3093
3094
3095
3096
3097
3098
3099
3100
3101
3102
3103
3104
3105
3106
3107
3108
3109
3110
3111
3112
3113
3114
3115
3116
3117
3118
3119
3120
3121
3122
3123
3124
3125
3126
3127
3128
3129
3130
3131
3132
3133
3134
3135
3136
3137
3138
3139
3140
3141
3142
3143
3144
3145
3146
3147
3148
3149
3150
3151
3152
3153
3154
3155
3156
3157
3158
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
3180
3181
3182
3183
3184
3185
3186
3187
3188
3189
3190
3191
3192
3193
3194
3195
3196
3197
3198
3199
3200
3201
3202
3203
3204
3205
3206
3207
3208
3209
3210
3211
3212
3213
3214
3215
3216
3217
3218
3219
3220
3221
3222
3223
3224
3225
3226
3227
3228
3229
3230
3231
3232
3233
3234
3235
3236
3237
3238
3239
3240
3241
3242
3243
3244
3245
3246
3247
3248
3249
3250
3251
3252
3253
3254
3255
3256
3257
3258
3259
3260
3261
3262
3263
3264
3265
3266
3267
3268
3269
3270
3271
3272
3273
3274
3275
3276
3277
3278
3279
3280
3281
3282
3283
3284
3285
3286
3287
3288
3289
3290
3291
3292
3293
3294
3295
3296
3297
3298
3299
3300
3301
3302
3303
3304
3305
3306
3307
3308
3309
3310
3311
3312
3313
3314
3315
3316
3317
3318
3319
3320
3321
3322
3323
3324
3325
3326
3327
3328
3329
3330
3331
3332
3333
3334
3335
3336
3337
3338
3339
3340
3341
3342
3343
3344
3345
3346
3347
3348
3349
3350
3351
3352
3353
3354
3355
3356
3357
3358
3359
3360
3361
3362
3363
3364
3365
3366
3367
3368
3369
3370
3371
3372
3373
3374
3375
3376
3377
3378
3379
3380
3381
3382
3383
3384
3385
3386
3387
3388
3389
3390
3391
3392
3393
3394
3395
3396
3397
3398
3399
3400
3401
3402
3403
3404
3405
3406
3407
3408
3409
3410
3411
3412
3413
3414
3415
3416
3417
3418
3419
3420
3421
3422
3423
3424
3425
3426
3427
3428
3429
3430
3431
3432
3433
3434
3435
3436
3437
3438
3439
3440
3441
3442
3443
3444
3445
3446
3447
3448
3449
3450
3451
3452
3453
3454
3455
3456
3457
3458
3459
3460
3461
3462
3463
3464
3465
3466
3467
3468
3469
3470
3471
3472
3473
3474
3475
3476
3477
3478
3479
3480
3481
3482
3483
3484
3485
3486
3487
3488
3489
3490
3491
3492
3493
3494
3495
3496
3497
3498
3499
3500
3501
3502
3503
3504
3505
3506
3507
3508
3509
3510
3511
3512
3513
3514
3515
3516
3517
3518
3519
3520
3521
3522
3523
3524
3525
3526
3527
3528
3529
3530
3531
3532
3533
3534
3535
3536
3537
3538
3539
3540
3541
3542
3543
3544
3545
3546
3547
3548
3549
3550
3551
3552
3553
3554
3555
3556
3557
3558
3559
3560
3561
3562
3563
3564
3565
3566
3567
3568
3569
3570
3571
3572
3573
3574
3575
3576
3577
3578
3579
3580
3581
3582
3583
3584
3585
3586
3587
3588
3589
3590
3591
3592
3593
3594
3595
3596
3597
3598
3599
3600
3601
3602
3603
3604
3605
3606
3607
3608
3609
3610
3611
3612
3613
3614
3615
3616
3617
3618
3619
3620
3621
3622
3623
3624
3625
3626
3627
3628
3629
3630
3631
3632
3633
3634
3635
3636
3637
3638
3639
3640
3641
3642
3643
3644
3645
3646
3647
3648
3649
3650
3651
3652
3653
3654
3655
3656
3657
3658
3659
3660
3661
3662
3663
3664
3665
3666
3667
3668
3669
3670
3671
3672
3673
3674
3675
3676
3677
3678
3679
3680
3681
3682
3683
3684
3685
3686
3687
3688
3689
3690
3691
3692
3693
3694
3695
3696
3697
3698
3699
3700
3701
3702
3703
3704
3705
3706
3707
3708
3709
3710
3711
3712
3713
3714
3715
3716
3717
3718
3719
3720
3721
3722
3723
3724
3725
3726
3727
3728
3729
3730
3731
3732
3733
3734
3735
3736
3737
3738
3739
3740
3741
3742
3743
3744
3745
3746
3747
3748
3749
3750
3751
3752
3753
3754
3755
3756
3757
3758
3759
3760
3761
3762
3763
3764
3765
3766
3767
3768
3769
3770
3771
3772
3773
3774
3775
3776
3777
3778
3779
3780
3781
3782
3783
3784
3785
3786
3787
3788
3789
3790
3791
3792
3793
3794
3795
3796
3797
3798
3799
3800
3801
3802
3803
3804
3805
3806
3807
3808
3809
3810
3811
3812
3813
3814
3815
3816
3817
3818
3819
3820
3821
3822
3823
3824
3825
3826
3827
3828
3829
3830
3831
3832
3833
3834
3835
3836
3837
3838
3839
3840
3841
3842
3843
3844
3845
3846
3847
3848
3849
3850
3851
3852
3853
3854
3855
3856
3857
3858
3859
3860
3861
3862
3863
3864
3865
3866
3867
3868
3869
3870
3871
3872
3873
3874
3875
3876
3877
3878
3879
3880
3881
3882
3883
3884
3885
3886
3887
3888
3889
3890
3891
3892
3893
3894
3895
3896
3897
3898
3899
3900
3901
3902
3903
3904
3905
3906
3907
3908
3909
3910
3911
3912
3913
3914
3915
3916
3917
3918
3919
3920
3921
3922
3923
3924
3925
3926
3927
3928
3929
3930
3931
3932
3933
3934
3935
3936
3937
3938
3939
3940
3941
3942
3943
3944
3945
3946
3947
3948
3949
3950
3951
3952
3953
3954
3955
3956
3957
3958
3959
3960
3961
3962
3963
3964
3965
3966
3967
3968
3969
3970
3971
3972
3973
3974
3975
3976
3977
3978
3979
3980
3981
3982
3983
3984
3985
3986
3987
3988
3989
3990
3991
3992
3993
3994
3995
3996
3997
3998
3999
4000
4001
4002
4003
4004
4005
4006
4007
4008
4009
4010
4011
4012
4013
4014
4015
4016
4017
4018
4019
4020
4021
4022
4023
4024
4025
4026
4027
4028
4029
4030
4031
4032
4033
4034
4035
4036
4037
4038
4039
4040
4041
4042
4043
4044
4045
4046
4047
4048
4049
4050
4051
4052
4053
4054
4055
4056
4057
4058
4059
4060
4061
4062
4063
4064
4065
4066
4067
4068
4069
4070
4071
4072
4073
4074
4075
4076
4077
4078
4079
4080
4081
4082
4083
4084
4085
4086
4087
4088
4089
4090
4091
4092
4093
4094
4095
4096
4097
4098
4099
4100
4101
4102
4103
4104
4105
4106
4107
4108
4109
4110
4111
4112
4113
4114
4115
4116
4117
4118
4119
4120
4121
4122
4123
4124
4125
4126
4127
4128
4129
4130
4131
4132
4133
4134
4135
4136
4137
4138
4139
4140
4141
4142
4143
4144
4145
4146
4147
4148
4149
4150
4151
4152
4153
4154
4155
4156
4157
4158
4159
4160
4161
4162
4163
4164
4165
4166
4167
4168
4169
4170
4171
4172
4173
4174
4175
4176
4177
4178
4179
4180
4181
4182
4183
4184
4185
4186
4187
4188
4189
4190
4191
4192
4193
4194
4195
4196
4197
4198
4199
4200
4201
4202
4203
4204
4205
4206
4207
4208
4209
4210
4211
4212
4213
4214
4215
4216
4217
4218
4219
4220
4221
4222
4223
4224
4225
4226
4227
4228
4229
4230
4231
4232
4233
4234
4235
4236
4237
4238
4239
4240
4241
4242
4243
4244
4245
4246
4247
4248
4249
4250
4251
4252
4253
4254
4255
4256
4257
4258
4259
4260
4261
4262
4263
4264
4265
4266
4267
4268
4269
4270
4271
4272
4273
4274
4275
4276
4277
4278
4279
4280
4281
4282
4283
4284
4285
4286
4287
4288
4289
4290
4291
4292
4293
4294
4295
4296
4297
4298
4299
4300
4301
4302
4303
4304
4305
4306
4307
4308
4309
4310
4311
4312
4313
4314
4315
4316
4317
4318
4319
4320
4321
4322
4323
4324
4325
4326
4327
4328
4329
4330
4331
4332
4333
4334
4335
4336
4337
4338
4339
4340
4341
4342
4343
4344
4345
4346
4347
4348
4349
4350
4351
4352
4353
4354
4355
4356
4357
4358
4359
4360
4361
4362
4363
4364
4365
4366
4367
4368
4369
4370
4371
4372
4373
4374
4375
4376
4377
4378
4379
4380
4381
4382
4383
4384
4385
4386
4387
4388
4389
4390
4391
4392
4393
4394
4395
4396
4397
4398
4399
4400
4401
4402
4403
4404
/* SPDX-License-Identifier: LGPL-2.1-or-later */

#if HAVE_VALGRIND_MEMCHECK_H
#include <valgrind/memcheck.h>
#endif

#include <linux/dm-ioctl.h>
#include <linux/loop.h>
#include <sys/file.h>
#include <sys/mount.h>
#include <sys/prctl.h>
#include <sys/wait.h>
#include <sysexits.h>

#if HAVE_OPENSSL
#include <openssl/err.h>
#include <openssl/pem.h>
#include <openssl/x509.h>
#endif

#include "sd-device.h"
#include "sd-id128.h"

#include "architecture.h"
#include "ask-password-api.h"
#include "blkid-util.h"
#include "blockdev-util.h"
#include "btrfs-util.h"
#include "chase.h"
#include "conf-files.h"
#include "constants.h"
#include "copy.h"
#include "cryptsetup-util.h"
#include "device-nodes.h"
#include "device-private.h"
#include "device-util.h"
#include "devnum-util.h"
#include "discover-image.h"
#include "dissect-image.h"
#include "dm-util.h"
#include "env-file.h"
#include "env-util.h"
#include "extension-util.h"
#include "fd-util.h"
#include "fileio.h"
#include "fs-util.h"
#include "fsck-util.h"
#include "gpt.h"
#include "hexdecoct.h"
#include "hostname-setup.h"
#include "id128-util.h"
#include "import-util.h"
#include "io-util.h"
#include "missing_mount.h"
#include "missing_syscall.h"
#include "mkdir-label.h"
#include "mount-util.h"
#include "mountpoint-util.h"
#include "namespace-util.h"
#include "nulstr-util.h"
#include "openssl-util.h"
#include "os-util.h"
#include "path-util.h"
#include "proc-cmdline.h"
#include "process-util.h"
#include "raw-clone.h"
#include "resize-fs.h"
#include "signal-util.h"
#include "sparse-endian.h"
#include "stat-util.h"
#include "stdio-util.h"
#include "string-table.h"
#include "string-util.h"
#include "strv.h"
#include "tmpfile-util.h"
#include "udev-util.h"
#include "user-util.h"
#include "varlink.h"
#include "xattr-util.h"

/* how many times to wait for the device nodes to appear */
#define N_DEVICE_NODE_LIST_ATTEMPTS 10

int dissect_fstype_ok(const char *fstype) {
        const char *e;
        bool b;

        /* When we automatically mount file systems, be a bit conservative by default what we are willing to
         * mount, just as an extra safety net to not mount with badly maintained legacy file system
         * drivers. */

        e = secure_getenv("SYSTEMD_DISSECT_FILE_SYSTEMS");
        if (e) {
                _cleanup_strv_free_ char **l = NULL;

                l = strv_split(e, ":");
                if (!l)
                        return -ENOMEM;

                b = strv_contains(l, fstype);
        } else
                b = STR_IN_SET(fstype,
                               "btrfs",
                               "erofs",
                               "ext4",
                               "f2fs",
                               "squashfs",
                               "vfat",
                               "xfs");
        if (b)
                return true;

        log_debug("File system type '%s' is not allowed to be mounted as result of automatic dissection.", fstype);
        return false;
}

int probe_sector_size(int fd, uint32_t *ret) {

        /* Disk images might be for 512B or for 4096 sector sizes, let's try to auto-detect that by searching
         * for the GPT headers at the relevant byte offsets */

        assert_cc(sizeof(GptHeader) == 92);

        /* We expect a sector size in the range 512…4096. The GPT header is located in the second
         * sector. Hence it could be at byte 512 at the earliest, and at byte 4096 at the latest. And we must
         * read with granularity of the largest sector size we care about. Which means 8K. */
        uint8_t sectors[2 * 4096];
        uint32_t found = 0;
        ssize_t n;

        assert(fd >= 0);
        assert(ret);

        n = pread(fd, sectors, sizeof(sectors), 0);
        if (n < 0)
                return -errno;
        if (n != sizeof(sectors)) /* too short? */
                goto not_found;

        /* Let's see if we find the GPT partition header with various expected sector sizes */
        for (uint32_t sz = 512; sz <= 4096; sz <<= 1) {
                const GptHeader *p;

                assert(sizeof(sectors) >= sz * 2);
                p = (const GptHeader*) (sectors + sz);

                if (!gpt_header_has_signature(p))
                        continue;

                if (found != 0)
                        return log_debug_errno(SYNTHETIC_ERRNO(ENOTUNIQ),
                                               "Detected valid partition table at offsets matching multiple sector sizes, refusing.");

                found = sz;
        }

        if (found != 0) {
                log_debug("Determined sector size %" PRIu32 " based on discovered partition table.", found);
                *ret = found;
                return 1; /* indicate we *did* find it */
        }

not_found:
        log_debug("Couldn't find any partition table to derive sector size of.");
        *ret = 512; /* pick the traditional default */
        return 0;   /* indicate we didn't find it */
}

int probe_sector_size_prefer_ioctl(int fd, uint32_t *ret) {
        struct stat st;

        assert(fd >= 0);
        assert(ret);

        /* Just like probe_sector_size(), but if we are looking at a block device, will use the already
         * configured sector size rather than probing by contents */

        if (fstat(fd, &st) < 0)
                return -errno;

        if (S_ISBLK(st.st_mode))
                return blockdev_get_sector_size(fd, ret);

        return probe_sector_size(fd, ret);
}

int probe_filesystem_full(
                int fd,
                const char *path,
                uint64_t offset,
                uint64_t size,
                char **ret_fstype) {

        /* Try to find device content type and return it in *ret_fstype. If nothing is found,
         * 0/NULL will be returned. -EUCLEAN will be returned for ambiguous results, and a
         * different error otherwise. */

#if HAVE_BLKID
        _cleanup_(blkid_free_probep) blkid_probe b = NULL;
        _cleanup_free_ char *path_by_fd = NULL;
        _cleanup_close_ int fd_close = -EBADF;
        const char *fstype;
        int r;

        assert(fd >= 0 || path);
        assert(ret_fstype);

        if (fd < 0) {
                fd_close = open(path, O_RDONLY|O_NONBLOCK|O_CLOEXEC|O_NOCTTY);
                if (fd_close < 0)
                        return -errno;

                fd = fd_close;
        }

        if (!path) {
                r = fd_get_path(fd, &path_by_fd);
                if (r < 0)
                        return r;

                path = path_by_fd;
        }

        if (size == 0) /* empty size? nothing found! */
                goto not_found;

        b = blkid_new_probe();
        if (!b)
                return -ENOMEM;

        /* The Linux kernel maintains separate block device caches for main ("whole") and partition block
         * devices, which means making a change to one might not be reflected immediately when reading via
         * the other. That's massively confusing when mixing accesses to such devices. Let's address this in
         * a limited way: when probing a file system that is not at the beginning of the block device we
         * apparently probe a partition via the main block device, and in that case let's first flush the
         * main block device cache, so that we get the data that the per-partition block device last
         * sync'ed on.
         *
         * This only works under the assumption that any tools that write to the partition block devices
         * issue an syncfs()/fsync() on the device after making changes. Typically file system formatting
         * tools that write a superblock onto a partition block device do that, however. */
        if (offset != 0)
                if (ioctl(fd, BLKFLSBUF, 0) < 0)
                        log_debug_errno(errno, "Failed to flush block device cache, ignoring: %m");

        errno = 0;
        r = blkid_probe_set_device(
                        b,
                        fd,
                        offset,
                        size == UINT64_MAX ? 0 : size); /* when blkid sees size=0 it understands "everything". We prefer using UINT64_MAX for that */
        if (r != 0)
                return errno_or_else(ENOMEM);

        blkid_probe_enable_superblocks(b, 1);
        blkid_probe_set_superblocks_flags(b, BLKID_SUBLKS_TYPE);

        errno = 0;
        r = blkid_do_safeprobe(b);
        if (r == _BLKID_SAFEPROBE_NOT_FOUND)
                goto not_found;
        if (r == _BLKID_SAFEPROBE_AMBIGUOUS)
                return log_debug_errno(SYNTHETIC_ERRNO(EUCLEAN),
                                       "Results ambiguous for partition %s", path);
        if (r == _BLKID_SAFEPROBE_ERROR)
                return log_debug_errno(errno_or_else(EIO), "Failed to probe partition %s: %m", path);

        assert(r == _BLKID_SAFEPROBE_FOUND);

        (void) blkid_probe_lookup_value(b, "TYPE", &fstype, NULL);

        if (fstype) {
                log_debug("Probed fstype '%s' on partition %s.", fstype, path);
                return strdup_to_full(ret_fstype, fstype);
        }

not_found:
        log_debug("No type detected on partition %s", path);
        *ret_fstype = NULL;
        return 0;
#else
        return -EOPNOTSUPP;
#endif
}

#if HAVE_BLKID
static int image_policy_may_use(
                const ImagePolicy *policy,
                PartitionDesignator designator) {

        PartitionPolicyFlags f;

        /* For each partition we find in the partition table do a first check if it may exist at all given
         * the policy, or if it shall be ignored. */

        f = image_policy_get_exhaustively(policy, designator);
        if (f < 0)
                return f;

        if ((f & _PARTITION_POLICY_USE_MASK) == PARTITION_POLICY_ABSENT)
                /* only flag set in policy is "absent"? then this partition may not exist at all */
                return log_debug_errno(
                                SYNTHETIC_ERRNO(ERFKILL),
                                "Partition of designator '%s' exists, but not allowed by policy, refusing.",
                                partition_designator_to_string(designator));
        if ((f & _PARTITION_POLICY_USE_MASK & ~PARTITION_POLICY_ABSENT) == PARTITION_POLICY_UNUSED) {
                /* only "unused" or "unused" + "absent" are set? then don't use it */
                log_debug("Partition of designator '%s' exists, and policy dictates to ignore it, doing so.",
                          partition_designator_to_string(designator));
                return false; /* ignore! */
        }

        return true; /* use! */
}

static int image_policy_check_protection(
                const ImagePolicy *policy,
                PartitionDesignator designator,
                PartitionPolicyFlags found_flags) {

        PartitionPolicyFlags policy_flags;

        /* Checks if the flags in the policy for the designated partition overlap the flags of what we found */

        if (found_flags < 0)
                return found_flags;

        policy_flags = image_policy_get_exhaustively(policy, designator);
        if (policy_flags < 0)
                return policy_flags;

        if ((found_flags & policy_flags) == 0) {
                _cleanup_free_ char *found_flags_string = NULL, *policy_flags_string = NULL;

                (void) partition_policy_flags_to_string(found_flags, /* simplify= */ true, &found_flags_string);
                (void) partition_policy_flags_to_string(policy_flags, /* simplify= */ true, &policy_flags_string);

                return log_debug_errno(SYNTHETIC_ERRNO(ERFKILL), "Partition %s discovered with policy '%s' but '%s' was required, refusing.",
                                       partition_designator_to_string(designator),
                                       strnull(found_flags_string), strnull(policy_flags_string));
        }

        return 0;
}

static int image_policy_check_partition_flags(
                const ImagePolicy *policy,
                PartitionDesignator designator,
                uint64_t gpt_flags) {

        PartitionPolicyFlags policy_flags;
        bool b;

        /* Checks if the partition flags in the policy match reality */

        policy_flags = image_policy_get_exhaustively(policy, designator);
        if (policy_flags < 0)
                return policy_flags;

        b = FLAGS_SET(gpt_flags, SD_GPT_FLAG_READ_ONLY);
        if ((policy_flags & _PARTITION_POLICY_READ_ONLY_MASK) == (b ? PARTITION_POLICY_READ_ONLY_OFF : PARTITION_POLICY_READ_ONLY_ON))
                return log_debug_errno(SYNTHETIC_ERRNO(ERFKILL), "Partition %s has 'read-only' flag incorrectly set (must be %s, is %s), refusing.",
                                       partition_designator_to_string(designator),
                                       one_zero(!b), one_zero(b));

        b = FLAGS_SET(gpt_flags, SD_GPT_FLAG_GROWFS);
        if ((policy_flags & _PARTITION_POLICY_GROWFS_MASK) == (b ? PARTITION_POLICY_GROWFS_OFF : PARTITION_POLICY_GROWFS_ON))
                return log_debug_errno(SYNTHETIC_ERRNO(ERFKILL), "Partition %s has 'growfs' flag incorrectly set (must be %s, is %s), refusing.",
                                       partition_designator_to_string(designator),
                                       one_zero(!b), one_zero(b));

        return 0;
}

static int dissected_image_probe_filesystems(
                DissectedImage *m,
                int fd,
                const ImagePolicy *policy) {

        int r;

        assert(m);

        /* Fill in file system types if we don't know them yet. */

        for (PartitionDesignator i = 0; i < _PARTITION_DESIGNATOR_MAX; i++) {
                DissectedPartition *p = m->partitions + i;
                PartitionPolicyFlags found_flags;

                if (!p->found)
                        continue;

                if (!p->fstype) {
                        /* If we have an fd referring to the partition block device, use that. Otherwise go
                         * via the whole block device or backing regular file, and read via offset. */
                        if (p->mount_node_fd >= 0)
                                r = probe_filesystem_full(p->mount_node_fd, p->node, 0, UINT64_MAX, &p->fstype);
                        else
                                r = probe_filesystem_full(fd, p->node, p->offset, p->size, &p->fstype);
                        if (r < 0)
                                return r;
                }

                if (streq_ptr(p->fstype, "crypto_LUKS")) {
                        m->encrypted = true;
                        found_flags = PARTITION_POLICY_ENCRYPTED; /* found this one, and its definitely encrypted */
                } else
                        /* found it, but it's definitely not encrypted, hence mask the encrypted flag, but
                         * set all other ways that indicate "present". */
                        found_flags = PARTITION_POLICY_UNPROTECTED|PARTITION_POLICY_VERITY|PARTITION_POLICY_SIGNED;

                if (p->fstype && fstype_is_ro(p->fstype))
                        p->rw = false;

                if (!p->rw)
                        p->growfs = false;

                /* We might have learnt more about the file system now (i.e. whether it is encrypted or not),
                 * hence we need to validate this against policy again, to see if the policy still matches
                 * with this new information. Note that image_policy_check_protection() will check for
                 * overlap between what's allowed in the policy and what we pass as 'found_policy' here. In
                 * the unencrypted case we thus might pass an overly unspecific mask here (i.e. unprotected
                 * OR verity OR signed), but that's fine since the earlier policy check already checked more
                 * specific which of those three cases where OK. Keep in mind that this function here only
                 * looks at specific partitions (and thus can only deduce encryption or not) but not the
                 * overall partition table (and thus cannot deduce verity or not). The earlier dissection
                 * checks already did the relevant checks that look at the whole partition table, and
                 * enforced policy there as needed. */
                r = image_policy_check_protection(policy, i, found_flags);
                if (r < 0)
                        return r;
        }

        return 0;
}

static void check_partition_flags(
                const char *node,
                unsigned long long pflags,
                unsigned long long supported) {

        assert(node);

        /* Mask away all flags supported by this partition's type and the three flags the UEFI spec defines generically */
        pflags &= ~(supported |
                    SD_GPT_FLAG_REQUIRED_PARTITION |
                    SD_GPT_FLAG_NO_BLOCK_IO_PROTOCOL |
                    SD_GPT_FLAG_LEGACY_BIOS_BOOTABLE);

        if (pflags == 0)
                return;

        /* If there are other bits set, then log about it, to make things discoverable */
        for (unsigned i = 0; i < sizeof(pflags) * 8; i++) {
                unsigned long long bit = 1ULL << i;
                if (!FLAGS_SET(pflags, bit))
                        continue;

                log_debug("Unexpected partition flag %llu set on %s!", bit, node);
        }
}

static int dissected_image_new(const char *path, DissectedImage **ret) {
        _cleanup_(dissected_image_unrefp) DissectedImage *m = NULL;
        _cleanup_free_ char *name = NULL;
        int r;

        assert(ret);

        if (path) {
                _cleanup_free_ char *filename = NULL;

                r = path_extract_filename(path, &filename);
                if (r < 0)
                        return r;

                r = raw_strip_suffixes(filename, &name);
                if (r < 0)
                        return r;

                if (!image_name_is_valid(name)) {
                        log_debug("Image name %s is not valid, ignoring.", strna(name));
                        name = mfree(name);
                }
        }

        m = new(DissectedImage, 1);
        if (!m)
                return -ENOMEM;

        *m = (DissectedImage) {
                .has_init_system = -1,
                .image_name = TAKE_PTR(name),
        };

        for (PartitionDesignator i = 0; i < _PARTITION_DESIGNATOR_MAX; i++)
                m->partitions[i] = DISSECTED_PARTITION_NULL;

        *ret = TAKE_PTR(m);
        return 0;
}
#endif

static void dissected_partition_done(DissectedPartition *p) {
        assert(p);

        free(p->fstype);
        free(p->node);
        free(p->label);
        free(p->decrypted_fstype);
        free(p->decrypted_node);
        free(p->mount_options);
        safe_close(p->mount_node_fd);
        safe_close(p->fsmount_fd);

        *p = DISSECTED_PARTITION_NULL;
}

#if HAVE_BLKID
static int diskseq_should_be_used(
                const char *whole_devname,
                uint64_t diskseq,
                DissectImageFlags flags) {

        int r;

        assert(whole_devname);

        /* No diskseq. We cannot use by-diskseq symlink. */
        if (diskseq == 0)
                return false;

        /* Do not use by-diskseq link unless DISSECT_IMAGE_DISKSEQ_DEVNODE flag is explicitly set. */
        if (!FLAGS_SET(flags, DISSECT_IMAGE_DISKSEQ_DEVNODE))
                return false;

        _cleanup_(sd_device_unrefp) sd_device *dev = NULL;
        r = sd_device_new_from_devname(&dev, whole_devname);
        if (r < 0)
                return r;

        /* When ID_IGNORE_DISKSEQ udev property is set, the by-diskseq symlink will not be created. */
        r = device_get_property_bool(dev, "ID_IGNORE_DISKSEQ");
        if (r >= 0)
                return !r; /* If explicitly specified, use it. */
        if (r != -ENOENT)
                return r;

        return true;
}

static int make_partition_devname(
                const char *whole_devname,
                uint64_t diskseq,
                int nr,
                DissectImageFlags flags,
                char **ret) {

        _cleanup_free_ char *s = NULL;
        int r;

        assert(whole_devname);
        assert(nr != 0); /* zero is not a valid partition nr */
        assert(ret);

        r = diskseq_should_be_used(whole_devname, diskseq, flags);
        if (r < 0)
                log_debug_errno(r, "Failed to determine if diskseq should be used for %s, assuming no, ignoring: %m", whole_devname);
        if (r <= 0) {
                /* Given a whole block device node name (e.g. /dev/sda or /dev/loop7) generate a partition
                 * device name (e.g. /dev/sda7 or /dev/loop7p5). The rule the kernel uses is simple: if whole
                 * block device node name ends in a digit, then suffix a 'p', followed by the partition
                 * number. Otherwise, just suffix the partition number without any 'p'. */

                if (nr < 0) { /* whole disk? */
                        s = strdup(whole_devname);
                        if (!s)
                                return -ENOMEM;
                } else {
                        size_t l = strlen(whole_devname);
                        if (l < 1) /* underflow check for the subtraction below */
                                return -EINVAL;

                        bool need_p = ascii_isdigit(whole_devname[l-1]); /* Last char a digit? */

                        if (asprintf(&s, "%s%s%i", whole_devname, need_p ? "p" : "", nr) < 0)
                                return -ENOMEM;
                }
        } else {
                if (nr < 0) /* whole disk? */
                        r = asprintf(&s, "/dev/disk/by-diskseq/%" PRIu64, diskseq);
                else
                        r = asprintf(&s, "/dev/disk/by-diskseq/%" PRIu64 "-part%i", diskseq, nr);
                if (r < 0)
                        return -ENOMEM;
        }

        *ret = TAKE_PTR(s);
        return 0;
}

static int open_partition(
                const char *node,
                bool is_partition,
                const LoopDevice *loop) {

        _cleanup_(sd_device_unrefp) sd_device *dev = NULL;
        _cleanup_close_ int fd = -EBADF;
        dev_t devnum;
        int r;

        assert(node);
        assert(loop);

        fd = open(node, O_RDONLY|O_NONBLOCK|O_CLOEXEC|O_NOCTTY);
        if (fd < 0)
                return -errno;

        /* Check if the block device is a child of (or equivalent to) the originally provided one. */
        r = block_device_new_from_fd(fd, is_partition ? BLOCK_DEVICE_LOOKUP_WHOLE_DISK : 0, &dev);
        if (r < 0)
                return r;

        r = sd_device_get_devnum(dev, &devnum);
        if (r < 0)
                return r;

        if (loop->devno != devnum)
                return -ENXIO;

        /* Also check diskseq. */
        if (loop->diskseq != 0) {
                uint64_t diskseq;

                r = fd_get_diskseq(fd, &diskseq);
                if (r < 0)
                        return r;

                if (loop->diskseq != diskseq)
                        return -ENXIO;
        }

        log_debug("Opened %s (fd=%i, whole_block_devnum=" DEVNUM_FORMAT_STR ", diskseq=%" PRIu64 ").",
                  node, fd, DEVNUM_FORMAT_VAL(loop->devno), loop->diskseq);
        return TAKE_FD(fd);
}

static int compare_arch(Architecture a, Architecture b) {
        if (a == b)
                return 0;

        if (a == native_architecture())
                return 1;

        if (b == native_architecture())
                return -1;

#ifdef ARCHITECTURE_SECONDARY
        if (a == ARCHITECTURE_SECONDARY)
                return 1;

        if (b == ARCHITECTURE_SECONDARY)
                return -1;
#endif

        return 0;
}

static int dissect_image(
                DissectedImage *m,
                int fd,
                const char *devname,
                const VeritySettings *verity,
                const MountOptions *mount_options,
                const ImagePolicy *policy,
                DissectImageFlags flags) {

        sd_id128_t root_uuid = SD_ID128_NULL, root_verity_uuid = SD_ID128_NULL;
        sd_id128_t usr_uuid = SD_ID128_NULL, usr_verity_uuid = SD_ID128_NULL;
        bool is_gpt, is_mbr, multiple_generic = false,
                generic_rw = false,  /* initialize to appease gcc */
                generic_growfs = false;
        _cleanup_(blkid_free_probep) blkid_probe b = NULL;
        _cleanup_free_ char *generic_node = NULL;
        sd_id128_t generic_uuid = SD_ID128_NULL;
        const char *pttype = NULL, *sptuuid = NULL;
        blkid_partlist pl;
        int r, generic_nr = -1, n_partitions;

        assert(m);
        assert(fd >= 0);
        assert(devname);
        assert(!verity || verity->designator < 0 || IN_SET(verity->designator, PARTITION_ROOT, PARTITION_USR));
        assert(!verity || verity->root_hash || verity->root_hash_size == 0);
        assert(!verity || verity->root_hash_sig || verity->root_hash_sig_size == 0);
        assert(!verity || (verity->root_hash || !verity->root_hash_sig));
        assert(!((flags & DISSECT_IMAGE_GPT_ONLY) && (flags & DISSECT_IMAGE_NO_PARTITION_TABLE)));
        assert(m->sector_size > 0);

        /* Probes a disk image, and returns information about what it found in *ret.
         *
         * Returns -ENOPKG if no suitable partition table or file system could be found.
         * Returns -EADDRNOTAVAIL if a root hash was specified but no matching root/verity partitions found.
         * Returns -ENXIO if we couldn't find any partition suitable as root or /usr partition
         * Returns -ENOTUNIQ if we only found multiple generic partitions and thus don't know what to do with that
         * Returns -ERFKILL if image doesn't match image policy
         * Returns -EBADR if verity data was provided externally for an image that has a GPT partition table (i.e. is not just a naked fs)
         * Returns -EPROTONOSUPPORT if DISSECT_IMAGE_ADD_PARTITION_DEVICES is set but the block device does not have partition logic enabled
         * Returns -ENOMSG if we didn't find a single usable partition (and DISSECT_IMAGE_REFUSE_EMPTY is set) */

        uint64_t diskseq = m->loop ? m->loop->diskseq : 0;

        if (verity && verity->root_hash) {
                sd_id128_t fsuuid, vuuid;

                /* If a root hash is supplied, then we use the root partition that has a UUID that match the
                 * first 128-bit of the root hash. And we use the verity partition that has a UUID that match
                 * the final 128-bit. */

                if (verity->root_hash_size < sizeof(sd_id128_t))
                        return -EINVAL;

                memcpy(&fsuuid, verity->root_hash, sizeof(sd_id128_t));
                memcpy(&vuuid, (const uint8_t*) verity->root_hash + verity->root_hash_size - sizeof(sd_id128_t), sizeof(sd_id128_t));

                if (sd_id128_is_null(fsuuid))
                        return -EINVAL;
                if (sd_id128_is_null(vuuid))
                        return -EINVAL;

                /* If the verity data declares it's for the /usr partition, then search for that, in all
                 * other cases assume it's for the root partition. */
                if (verity->designator == PARTITION_USR) {
                        usr_uuid = fsuuid;
                        usr_verity_uuid = vuuid;
                } else {
                        root_uuid = fsuuid;
                        root_verity_uuid = vuuid;
                }
        }

        b = blkid_new_probe();
        if (!b)
                return -ENOMEM;

        errno = 0;
        r = blkid_probe_set_device(b, fd, 0, 0);
        if (r != 0)
                return errno_or_else(ENOMEM);

        errno = 0;
        r = blkid_probe_set_sectorsize(b, m->sector_size);
        if (r != 0)
                return errno_or_else(EIO);

        if ((flags & DISSECT_IMAGE_GPT_ONLY) == 0) {
                /* Look for file system superblocks, unless we only shall look for GPT partition tables */
                blkid_probe_enable_superblocks(b, 1);
                blkid_probe_set_superblocks_flags(b, BLKID_SUBLKS_TYPE|BLKID_SUBLKS_USAGE|BLKID_SUBLKS_UUID);
        }

        blkid_probe_enable_partitions(b, 1);
        blkid_probe_set_partitions_flags(b, BLKID_PARTS_ENTRY_DETAILS);

        errno = 0;
        r = blkid_do_safeprobe(b);
        if (r == _BLKID_SAFEPROBE_ERROR)
                return errno_or_else(EIO);
        if (IN_SET(r, _BLKID_SAFEPROBE_AMBIGUOUS, _BLKID_SAFEPROBE_NOT_FOUND))
                return log_debug_errno(SYNTHETIC_ERRNO(ENOPKG), "Failed to identify any partition table.");

        assert(r == _BLKID_SAFEPROBE_FOUND);

        if ((!(flags & DISSECT_IMAGE_GPT_ONLY) &&
            (flags & DISSECT_IMAGE_GENERIC_ROOT)) ||
            (flags & DISSECT_IMAGE_NO_PARTITION_TABLE)) {
                const char *usage = NULL;

                /* If flags permit this, also allow using non-partitioned single-filesystem images */

                (void) blkid_probe_lookup_value(b, "USAGE", &usage, NULL);
                if (STRPTR_IN_SET(usage, "filesystem", "crypto")) {
                        _cleanup_free_ char *t = NULL, *n = NULL, *o = NULL;
                        const char *fstype = NULL, *options = NULL, *suuid = NULL;
                        _cleanup_close_ int mount_node_fd = -EBADF;
                        sd_id128_t uuid = SD_ID128_NULL;
                        PartitionPolicyFlags found_flags;
                        bool encrypted;

                        /* OK, we have found a file system, that's our root partition then. */

                        r = image_policy_may_use(policy, PARTITION_ROOT);
                        if (r < 0)
                                return r;
                        if (r == 0) /* policy says ignore this, so we ignore it */
                                return -ENOPKG;

                        (void) blkid_probe_lookup_value(b, "TYPE", &fstype, NULL);
                        (void) blkid_probe_lookup_value(b, "UUID", &suuid, NULL);

                        encrypted = streq_ptr(fstype, "crypto_LUKS");

                        if (verity_settings_data_covers(verity, PARTITION_ROOT))
                                found_flags = verity->root_hash_sig ? PARTITION_POLICY_SIGNED : PARTITION_POLICY_VERITY;
                        else
                                found_flags = encrypted ? PARTITION_POLICY_ENCRYPTED : PARTITION_POLICY_UNPROTECTED;

                        r = image_policy_check_protection(policy, PARTITION_ROOT, found_flags);
                        if (r < 0)
                                return r;

                        r = image_policy_check_partition_flags(policy, PARTITION_ROOT, 0); /* we have no gpt partition flags, hence check against all bits off */
                        if (r < 0)
                                return r;

                        if (FLAGS_SET(flags, DISSECT_IMAGE_PIN_PARTITION_DEVICES)) {
                                mount_node_fd = open_partition(devname, /* is_partition = */ false, m->loop);
                                if (mount_node_fd < 0)
                                        return mount_node_fd;
                        }

                        if (fstype) {
                                t = strdup(fstype);
                                if (!t)
                                        return -ENOMEM;
                        }

                        if (suuid) {
                                /* blkid will return FAT's serial number as UUID, hence it is quite possible
                                 * that parsing this will fail. We'll ignore the ID, since it's just too
                                 * short to be useful as true identifier. */
                                r = sd_id128_from_string(suuid, &uuid);
                                if (r < 0)
                                        log_debug_errno(r, "Failed to parse file system UUID '%s', ignoring: %m", suuid);
                        }

                        r = make_partition_devname(devname, diskseq, -1, flags, &n);
                        if (r < 0)
                                return r;

                        m->single_file_system = true;
                        m->encrypted = encrypted;

                        m->has_verity = verity && verity->data_path;
                        m->verity_ready = verity_settings_data_covers(verity, PARTITION_ROOT);

                        m->has_verity_sig = false; /* signature not embedded, must be specified */
                        m->verity_sig_ready = m->verity_ready && verity->root_hash_sig;

                        m->image_uuid = uuid;

                        options = mount_options_from_designator(mount_options, PARTITION_ROOT);
                        if (options) {
                                o = strdup(options);
                                if (!o)
                                        return -ENOMEM;
                        }

                        m->partitions[PARTITION_ROOT] = (DissectedPartition) {
                                .found = true,
                                .rw = !m->verity_ready && !fstype_is_ro(fstype),
                                .partno = -1,
                                .architecture = _ARCHITECTURE_INVALID,
                                .fstype = TAKE_PTR(t),
                                .node = TAKE_PTR(n),
                                .mount_options = TAKE_PTR(o),
                                .mount_node_fd = TAKE_FD(mount_node_fd),
                                .offset = 0,
                                .size = UINT64_MAX,
                                .fsmount_fd = -EBADF,
                        };

                        return 0;
                }
        }

        (void) blkid_probe_lookup_value(b, "PTTYPE", &pttype, NULL);
        if (!pttype)
                return -ENOPKG;

        is_gpt = streq_ptr(pttype, "gpt");
        is_mbr = streq_ptr(pttype, "dos");

        if (!is_gpt && ((flags & DISSECT_IMAGE_GPT_ONLY) || !is_mbr))
                return -ENOPKG;

        /* We support external verity data partitions only if the image has no partition table */
        if (verity && verity->data_path)
                return -EBADR;

        if (FLAGS_SET(flags, DISSECT_IMAGE_ADD_PARTITION_DEVICES)) {
                /* Safety check: refuse block devices that carry a partition table but for which the kernel doesn't
                 * do partition scanning. */
                r = blockdev_partscan_enabled(fd);
                if (r < 0)
                        return r;
                if (r == 0)
                        return -EPROTONOSUPPORT;
        }

        (void) blkid_probe_lookup_value(b, "PTUUID", &sptuuid, NULL);
        if (sptuuid) {
                r = sd_id128_from_string(sptuuid, &m->image_uuid);
                if (r < 0)
                        log_debug_errno(r, "Failed to parse partition table UUID '%s', ignoring: %m", sptuuid);
        }

        errno = 0;
        pl = blkid_probe_get_partitions(b);
        if (!pl)
                return errno_or_else(ENOMEM);

        errno = 0;
        n_partitions = blkid_partlist_numof_partitions(pl);
        if (n_partitions < 0)
                return errno_or_else(EIO);

        for (int i = 0; i < n_partitions; i++) {
                _cleanup_free_ char *node = NULL;
                unsigned long long pflags;
                blkid_loff_t start, size;
                blkid_partition pp;
                int nr;

                errno = 0;
                pp = blkid_partlist_get_partition(pl, i);
                if (!pp)
                        return errno_or_else(EIO);

                pflags = blkid_partition_get_flags(pp);

                errno = 0;
                nr = blkid_partition_get_partno(pp);
                if (nr < 0)
                        return errno_or_else(EIO);

                errno = 0;
                start = blkid_partition_get_start(pp);
                if (start < 0)
                        return errno_or_else(EIO);

                assert((uint64_t) start < UINT64_MAX/512);

                errno = 0;
                size = blkid_partition_get_size(pp);
                if (size < 0)
                        return errno_or_else(EIO);

                assert((uint64_t) size < UINT64_MAX/512);

                /* While probing we need the non-diskseq device node name to access the thing, hence mask off
                 * DISSECT_IMAGE_DISKSEQ_DEVNODE. */
                r = make_partition_devname(devname, diskseq, nr, flags & ~DISSECT_IMAGE_DISKSEQ_DEVNODE, &node);
                if (r < 0)
                        return r;

                /* So here's the thing: after the main ("whole") block device popped up it might take a while
                 * before the kernel fully probed the partition table. Waiting for that to finish is icky in
                 * userspace. So here's what we do instead. We issue the BLKPG_ADD_PARTITION ioctl to add the
                 * partition ourselves, racing against the kernel. Good thing is: if this call fails with
                 * EBUSY then the kernel was quicker than us, and that's totally OK, the outcome is good for
                 * us: the device node will exist. If OTOH our call was successful we won the race. Which is
                 * also good as the outcome is the same: the partition block device exists, and we can use
                 * it.
                 *
                 * Kernel returns EBUSY if there's already a partition by that number or an overlapping
                 * partition already existent. */

                if (FLAGS_SET(flags, DISSECT_IMAGE_ADD_PARTITION_DEVICES)) {
                        r = block_device_add_partition(fd, node, nr, (uint64_t) start * 512, (uint64_t) size * 512);
                        if (r < 0) {
                                if (r != -EBUSY)
                                        return log_debug_errno(r, "BLKPG_ADD_PARTITION failed: %m");

                                log_debug_errno(r, "Kernel was quicker than us in adding partition %i.", nr);
                        } else
                                log_debug("We were quicker than kernel in adding partition %i.", nr);
                }

                if (is_gpt) {
                        const char *fstype = NULL, *label;
                        sd_id128_t type_id, id;
                        GptPartitionType type;
                        bool rw = true, growfs = false;

                        r = blkid_partition_get_uuid_id128(pp, &id);
                        if (r < 0) {
                                log_debug_errno(r, "Failed to read partition UUID, ignoring: %m");
                                continue;
                        }

                        r = blkid_partition_get_type_id128(pp, &type_id);
                        if (r < 0) {
                                log_debug_errno(r, "Failed to read partition type UUID, ignoring: %m");
                                continue;
                        }

                        type = gpt_partition_type_from_uuid(type_id);

                        label = blkid_partition_get_name(pp); /* libblkid returns NULL here if empty */

                        if (IN_SET(type.designator,
                                   PARTITION_HOME,
                                   PARTITION_SRV,
                                   PARTITION_XBOOTLDR,
                                   PARTITION_TMP)) {

                                check_partition_flags(node, pflags,
                                                      SD_GPT_FLAG_NO_AUTO | SD_GPT_FLAG_READ_ONLY | SD_GPT_FLAG_GROWFS);

                                if (pflags & SD_GPT_FLAG_NO_AUTO)
                                        continue;

                                rw = !(pflags & SD_GPT_FLAG_READ_ONLY);
                                growfs = FLAGS_SET(pflags, SD_GPT_FLAG_GROWFS);

                        } else if (type.designator == PARTITION_ESP) {

                                /* Note that we don't check the SD_GPT_FLAG_NO_AUTO flag for the ESP, as it is
                                 * not defined there. We instead check the SD_GPT_FLAG_NO_BLOCK_IO_PROTOCOL, as
                                 * recommended by the UEFI spec (See "12.3.3 Number and Location of System
                                 * Partitions"). */

                                if (pflags & SD_GPT_FLAG_NO_BLOCK_IO_PROTOCOL)
                                        continue;

                                fstype = "vfat";

                        } else if (type.designator == PARTITION_ROOT) {

                                check_partition_flags(node, pflags,
                                                      SD_GPT_FLAG_NO_AUTO | SD_GPT_FLAG_READ_ONLY | SD_GPT_FLAG_GROWFS);

                                if (pflags & SD_GPT_FLAG_NO_AUTO)
                                        continue;

                                /* If a root ID is specified, ignore everything but the root id */
                                if (!sd_id128_is_null(root_uuid) && !sd_id128_equal(root_uuid, id))
                                        continue;

                                rw = !(pflags & SD_GPT_FLAG_READ_ONLY);
                                growfs = FLAGS_SET(pflags, SD_GPT_FLAG_GROWFS);

                        } else if (type.designator == PARTITION_ROOT_VERITY) {

                                check_partition_flags(node, pflags,
                                                      SD_GPT_FLAG_NO_AUTO | SD_GPT_FLAG_READ_ONLY);

                                if (pflags & SD_GPT_FLAG_NO_AUTO)
                                        continue;

                                m->has_verity = true;

                                /* If no verity configuration is specified, then don't do verity */
                                if (!verity)
                                        continue;
                                if (verity->designator >= 0 && verity->designator != PARTITION_ROOT)
                                        continue;

                                /* If root hash is specified, then ignore everything but the root id */
                                if (!sd_id128_is_null(root_verity_uuid) && !sd_id128_equal(root_verity_uuid, id))
                                        continue;

                                fstype = "DM_verity_hash";
                                rw = false;

                        } else if (type.designator == PARTITION_ROOT_VERITY_SIG) {

                                check_partition_flags(node, pflags,
                                                      SD_GPT_FLAG_NO_AUTO | SD_GPT_FLAG_READ_ONLY);

                                if (pflags & SD_GPT_FLAG_NO_AUTO)
                                        continue;

                                m->has_verity_sig = true;

                                if (!verity)
                                        continue;
                                if (verity->designator >= 0 && verity->designator != PARTITION_ROOT)
                                        continue;

                                fstype = "verity_hash_signature";
                                rw = false;

                        } else if (type.designator == PARTITION_USR) {

                                check_partition_flags(node, pflags,
                                                      SD_GPT_FLAG_NO_AUTO | SD_GPT_FLAG_READ_ONLY | SD_GPT_FLAG_GROWFS);

                                if (pflags & SD_GPT_FLAG_NO_AUTO)
                                        continue;

                                /* If a usr ID is specified, ignore everything but the usr id */
                                if (!sd_id128_is_null(usr_uuid) && !sd_id128_equal(usr_uuid, id))
                                        continue;

                                rw = !(pflags & SD_GPT_FLAG_READ_ONLY);
                                growfs = FLAGS_SET(pflags, SD_GPT_FLAG_GROWFS);

                        } else if (type.designator == PARTITION_USR_VERITY) {

                                check_partition_flags(node, pflags,
                                                      SD_GPT_FLAG_NO_AUTO | SD_GPT_FLAG_READ_ONLY);

                                if (pflags & SD_GPT_FLAG_NO_AUTO)
                                        continue;

                                m->has_verity = true;

                                if (!verity)
                                        continue;
                                if (verity->designator >= 0 && verity->designator != PARTITION_USR)
                                        continue;

                                /* If usr hash is specified, then ignore everything but the usr id */
                                if (!sd_id128_is_null(usr_verity_uuid) && !sd_id128_equal(usr_verity_uuid, id))
                                        continue;

                                fstype = "DM_verity_hash";
                                rw = false;

                        } else if (type.designator == PARTITION_USR_VERITY_SIG) {

                                check_partition_flags(node, pflags,
                                                      SD_GPT_FLAG_NO_AUTO | SD_GPT_FLAG_READ_ONLY);

                                if (pflags & SD_GPT_FLAG_NO_AUTO)
                                        continue;

                                m->has_verity_sig = true;

                                if (!verity)
                                        continue;
                                if (verity->designator >= 0 && verity->designator != PARTITION_USR)
                                        continue;

                                fstype = "verity_hash_signature";
                                rw = false;

                        } else if (type.designator == PARTITION_SWAP) {

                                check_partition_flags(node, pflags, SD_GPT_FLAG_NO_AUTO);

                                if (pflags & SD_GPT_FLAG_NO_AUTO)
                                        continue;

                                /* Note: we don't set fstype = "swap" here, because we still need to probe if
                                 * it might be encrypted (i.e. fstype "crypt_LUKS") or unencrypted
                                 * (i.e. fstype "swap"), and the only way to figure that out is via fstype
                                 * probing. */

                        /* We don't have a designator for SD_GPT_LINUX_GENERIC so check the UUID instead. */
                        } else if (sd_id128_equal(type.uuid, SD_GPT_LINUX_GENERIC)) {

                                check_partition_flags(node, pflags,
                                                      SD_GPT_FLAG_NO_AUTO | SD_GPT_FLAG_READ_ONLY | SD_GPT_FLAG_GROWFS);

                                if (pflags & SD_GPT_FLAG_NO_AUTO)
                                        continue;

                                if (generic_node)
                                        multiple_generic = true;
                                else {
                                        generic_nr = nr;
                                        generic_rw = !(pflags & SD_GPT_FLAG_READ_ONLY);
                                        generic_growfs = FLAGS_SET(pflags, SD_GPT_FLAG_GROWFS);
                                        generic_uuid = id;
                                        generic_node = TAKE_PTR(node);
                                }

                        } else if (type.designator == PARTITION_VAR) {

                                check_partition_flags(node, pflags,
                                                      SD_GPT_FLAG_NO_AUTO | SD_GPT_FLAG_READ_ONLY | SD_GPT_FLAG_GROWFS);

                                if (pflags & SD_GPT_FLAG_NO_AUTO)
                                        continue;

                                if (!FLAGS_SET(flags, DISSECT_IMAGE_RELAX_VAR_CHECK)) {
                                        sd_id128_t var_uuid;

                                        /* For /var we insist that the uuid of the partition matches the
                                         * HMAC-SHA256 of the /var GPT partition type uuid, keyed by machine
                                         * ID. Why? Unlike the other partitions /var is inherently
                                         * installation specific, hence we need to be careful not to mount it
                                         * in the wrong installation. By hashing the partition UUID from
                                         * /etc/machine-id we can securely bind the partition to the
                                         * installation. */

                                        r = sd_id128_get_machine_app_specific(SD_GPT_VAR, &var_uuid);
                                        if (r < 0)
                                                return r;

                                        if (!sd_id128_equal(var_uuid, id)) {
                                                log_debug("Found a /var/ partition, but its UUID didn't match our expectations "
                                                          "(found: " SD_ID128_UUID_FORMAT_STR ", expected: " SD_ID128_UUID_FORMAT_STR "), ignoring.",
                                                          SD_ID128_FORMAT_VAL(id), SD_ID128_FORMAT_VAL(var_uuid));
                                                continue;
                                        }
                                }

                                rw = !(pflags & SD_GPT_FLAG_READ_ONLY);
                                growfs = FLAGS_SET(pflags, SD_GPT_FLAG_GROWFS);
                        }

                        if (type.designator != _PARTITION_DESIGNATOR_INVALID) {
                                _cleanup_free_ char *t = NULL, *o = NULL, *l = NULL, *n = NULL;
                                _cleanup_close_ int mount_node_fd = -EBADF;
                                const char *options = NULL;

                                r = image_policy_may_use(policy, type.designator);
                                if (r < 0)
                                        return r;
                                if (r == 0) {
                                        /* Policy says: ignore; Remember this fact, so that we later can distinguish between "found but ignored" and "not found at all" */

                                        if (!m->partitions[type.designator].found)
                                                m->partitions[type.designator].ignored = true;

                                        continue;
                                }

                                if (m->partitions[type.designator].found) {
                                        int c;

                                        /* For most partition types the first one we see wins. Except for the
                                         * rootfs and /usr, where we do a version compare of the label, and
                                         * let the newest version win. This permits a simple A/B versioning
                                         * scheme in OS images. */

                                        c = compare_arch(type.arch, m->partitions[type.designator].architecture);
                                        if (c < 0) /* the arch we already found is better than the one we found now */
                                                continue;
                                        if (c == 0 && /* same arch? then go by version in label */
                                            (!partition_designator_is_versioned(type.designator) ||
                                             strverscmp_improved(label, m->partitions[type.designator].label) <= 0))
                                                continue;

                                        dissected_partition_done(m->partitions + type.designator);
                                }

                                if (FLAGS_SET(flags, DISSECT_IMAGE_PIN_PARTITION_DEVICES) &&
                                    type.designator != PARTITION_SWAP) {
                                        mount_node_fd = open_partition(node, /* is_partition = */ true, m->loop);
                                        if (mount_node_fd < 0)
                                                return mount_node_fd;
                                }

                                r = make_partition_devname(devname, diskseq, nr, flags, &n);
                                if (r < 0)
                                        return r;

                                if (fstype) {
                                        t = strdup(fstype);
                                        if (!t)
                                                return -ENOMEM;
                                }

                                if (label) {
                                        l = strdup(label);
                                        if (!l)
                                                return -ENOMEM;
                                }

                                options = mount_options_from_designator(mount_options, type.designator);
                                if (options) {
                                        o = strdup(options);
                                        if (!o)
                                                return -ENOMEM;
                                }

                                m->partitions[type.designator] = (DissectedPartition) {
                                        .found = true,
                                        .partno = nr,
                                        .rw = rw,
                                        .growfs = growfs,
                                        .architecture = type.arch,
                                        .node = TAKE_PTR(n),
                                        .fstype = TAKE_PTR(t),
                                        .label = TAKE_PTR(l),
                                        .uuid = id,
                                        .mount_options = TAKE_PTR(o),
                                        .mount_node_fd = TAKE_FD(mount_node_fd),
                                        .offset = (uint64_t) start * 512,
                                        .size = (uint64_t) size * 512,
                                        .gpt_flags = pflags,
                                        .fsmount_fd = -EBADF,
                                };
                        }

                } else if (is_mbr) {

                        switch (blkid_partition_get_type(pp)) {

                        case 0x83: /* Linux partition */

                                if (pflags != 0x80) /* Bootable flag */
                                        continue;

                                if (generic_node)
                                        multiple_generic = true;
                                else {
                                        generic_nr = nr;
                                        generic_rw = true;
                                        generic_growfs = false;
                                        generic_node = TAKE_PTR(node);
                                }

                                break;

                        case 0xEA: { /* Boot Loader Spec extended $BOOT partition */
                                _cleanup_close_ int mount_node_fd = -EBADF;
                                _cleanup_free_ char *o = NULL, *n = NULL;
                                sd_id128_t id = SD_ID128_NULL;
                                const char *options = NULL;

                                r = image_policy_may_use(policy, PARTITION_XBOOTLDR);
                                if (r < 0)
                                        return r;
                                if (r == 0) { /* policy says: ignore */
                                        if (!m->partitions[PARTITION_XBOOTLDR].found)
                                                m->partitions[PARTITION_XBOOTLDR].ignored = true;

                                        continue;
                                }

                                /* First one wins */
                                if (m->partitions[PARTITION_XBOOTLDR].found)
                                        continue;

                                if (FLAGS_SET(flags, DISSECT_IMAGE_PIN_PARTITION_DEVICES)) {
                                        mount_node_fd = open_partition(node, /* is_partition = */ true, m->loop);
                                        if (mount_node_fd < 0)
                                                return mount_node_fd;
                                }

                                (void) blkid_partition_get_uuid_id128(pp, &id);

                                r = make_partition_devname(devname, diskseq, nr, flags, &n);
                                if (r < 0)
                                        return r;

                                options = mount_options_from_designator(mount_options, PARTITION_XBOOTLDR);
                                if (options) {
                                        o = strdup(options);
                                        if (!o)
                                                return -ENOMEM;
                                }

                                m->partitions[PARTITION_XBOOTLDR] = (DissectedPartition) {
                                        .found = true,
                                        .partno = nr,
                                        .rw = true,
                                        .growfs = false,
                                        .architecture = _ARCHITECTURE_INVALID,
                                        .node = TAKE_PTR(n),
                                        .uuid = id,
                                        .mount_options = TAKE_PTR(o),
                                        .mount_node_fd = TAKE_FD(mount_node_fd),
                                        .offset = (uint64_t) start * 512,
                                        .size = (uint64_t) size * 512,
                                        .fsmount_fd = -EBADF,
                                };

                                break;
                        }}
                }
        }

        if (!m->partitions[PARTITION_ROOT].found &&
                (m->partitions[PARTITION_ROOT_VERITY].found ||
                 m->partitions[PARTITION_ROOT_VERITY_SIG].found))
                        return -EADDRNOTAVAIL; /* Verity found but no matching rootfs? Something is off, refuse. */

        /* Hmm, we found a signature partition but no Verity data? Something is off. */
        if (m->partitions[PARTITION_ROOT_VERITY_SIG].found && !m->partitions[PARTITION_ROOT_VERITY].found)
                return -EADDRNOTAVAIL;

        if (!m->partitions[PARTITION_USR].found &&
                (m->partitions[PARTITION_USR_VERITY].found ||
                 m->partitions[PARTITION_USR_VERITY_SIG].found))
                        return -EADDRNOTAVAIL; /* as above */

        /* as above */
        if (m->partitions[PARTITION_USR_VERITY_SIG].found && !m->partitions[PARTITION_USR_VERITY].found)
                return -EADDRNOTAVAIL;

        /* If root and /usr are combined then insist that the architecture matches */
        if (m->partitions[PARTITION_ROOT].found &&
            m->partitions[PARTITION_USR].found &&
            (m->partitions[PARTITION_ROOT].architecture >= 0 &&
             m->partitions[PARTITION_USR].architecture >= 0 &&
             m->partitions[PARTITION_ROOT].architecture != m->partitions[PARTITION_USR].architecture))
                return -EADDRNOTAVAIL;

        if (!m->partitions[PARTITION_ROOT].found &&
            !m->partitions[PARTITION_USR].found &&
            (flags & DISSECT_IMAGE_GENERIC_ROOT) &&
            (!verity || !verity->root_hash || verity->designator != PARTITION_USR)) {

                /* OK, we found nothing usable, then check if there's a single generic partition, and use
                 * that. If the root hash was set however, then we won't fall back to a generic node, because
                 * the root hash decides. */

                /* If we didn't find a properly marked root partition, but we did find a single suitable
                 * generic Linux partition, then use this as root partition, if the caller asked for it. */
                if (multiple_generic)
                        return -ENOTUNIQ;

                /* If we didn't find a generic node, then we can't fix this up either */
                if (generic_node) {
                        r = image_policy_may_use(policy, PARTITION_ROOT);
                        if (r < 0)
                                return r;
                        if (r == 0)
                                /* Policy says: ignore; remember that we did */
                                m->partitions[PARTITION_ROOT].ignored = true;
                        else {
                                _cleanup_close_ int mount_node_fd = -EBADF;
                                _cleanup_free_ char *o = NULL, *n = NULL;
                                const char *options;

                                if (FLAGS_SET(flags, DISSECT_IMAGE_PIN_PARTITION_DEVICES)) {
                                        mount_node_fd = open_partition(generic_node, /* is_partition = */ true, m->loop);
                                        if (mount_node_fd < 0)
                                                return mount_node_fd;
                                }

                                r = make_partition_devname(devname, diskseq, generic_nr, flags, &n);
                                if (r < 0)
                                        return r;

                                options = mount_options_from_designator(mount_options, PARTITION_ROOT);
                                if (options) {
                                        o = strdup(options);
                                        if (!o)
                                                return -ENOMEM;
                                }

                                assert(generic_nr >= 0);
                                m->partitions[PARTITION_ROOT] = (DissectedPartition) {
                                        .found = true,
                                        .rw = generic_rw,
                                        .growfs = generic_growfs,
                                        .partno = generic_nr,
                                        .architecture = _ARCHITECTURE_INVALID,
                                        .node = TAKE_PTR(n),
                                        .uuid = generic_uuid,
                                        .mount_options = TAKE_PTR(o),
                                        .mount_node_fd = TAKE_FD(mount_node_fd),
                                        .offset = UINT64_MAX,
                                        .size = UINT64_MAX,
                                        .fsmount_fd = -EBADF,
                                };
                        }
                }
        }

        /* Check if we have a root fs if we are told to do check. /usr alone is fine too, but only if appropriate flag for that is set too */
        if (FLAGS_SET(flags, DISSECT_IMAGE_REQUIRE_ROOT) &&
            !(m->partitions[PARTITION_ROOT].found || (m->partitions[PARTITION_USR].found && FLAGS_SET(flags, DISSECT_IMAGE_USR_NO_ROOT))))
                return -ENXIO;

        if (m->partitions[PARTITION_ROOT_VERITY].found) {
                /* We only support one verity partition per image, i.e. can't do for both /usr and root fs */
                if (m->partitions[PARTITION_USR_VERITY].found)
                        return -ENOTUNIQ;

                /* We don't support verity enabled root with a split out /usr. Neither with nor without
                 * verity there. (Note that we do support verity-less root with verity-full /usr, though.) */
                if (m->partitions[PARTITION_USR].found)
                        return -EADDRNOTAVAIL;
        }

        if (verity) {
                /* If a verity designator is specified, then insist that the matching partition exists */
                if (verity->designator >= 0 && !m->partitions[verity->designator].found)
                        return -EADDRNOTAVAIL;

                bool have_verity_sig_partition;
                if (verity->designator >= 0)
                        have_verity_sig_partition = m->partitions[verity->designator == PARTITION_USR ? PARTITION_USR_VERITY_SIG : PARTITION_ROOT_VERITY_SIG].found;
                else
                        have_verity_sig_partition = m->partitions[PARTITION_USR_VERITY_SIG].found || m->partitions[PARTITION_ROOT_VERITY_SIG].found;

                if (verity->root_hash) {
                        /* If we have an explicit root hash and found the partitions for it, then we are ready to use
                         * Verity, set things up for it */

                        if (verity->designator < 0 || verity->designator == PARTITION_ROOT) {
                                if (!m->partitions[PARTITION_ROOT_VERITY].found || !m->partitions[PARTITION_ROOT].found)
                                        return -EADDRNOTAVAIL;

                                /* If we found a verity setup, then the root partition is necessarily read-only. */
                                m->partitions[PARTITION_ROOT].rw = false;
                                m->verity_ready = true;

                        } else {
                                assert(verity->designator == PARTITION_USR);

                                if (!m->partitions[PARTITION_USR_VERITY].found || !m->partitions[PARTITION_USR].found)
                                        return -EADDRNOTAVAIL;

                                m->partitions[PARTITION_USR].rw = false;
                                m->verity_ready = true;
                        }

                        if (m->verity_ready)
                                m->verity_sig_ready = verity->root_hash_sig || have_verity_sig_partition;

                } else if (have_verity_sig_partition) {

                        /* If we found an embedded signature partition, we are ready, too. */

                        m->verity_ready = m->verity_sig_ready = true;
                        if (verity->designator >= 0)
                                m->partitions[verity->designator == PARTITION_USR ? PARTITION_USR : PARTITION_ROOT].rw = false;
                        else if (m->partitions[PARTITION_USR_VERITY_SIG].found)
                                m->partitions[PARTITION_USR].rw = false;
                        else if (m->partitions[PARTITION_ROOT_VERITY_SIG].found)
                                m->partitions[PARTITION_ROOT].rw = false;
                }
        }

        bool any = false;

        /* After we discovered all partitions let's see if the verity requirements match the policy. (Note:
         * we don't check encryption requirements here, because we haven't probed the file system yet, hence
         * don't know if this is encrypted or not) */
        for (PartitionDesignator di = 0; di < _PARTITION_DESIGNATOR_MAX; di++) {
                PartitionDesignator vi, si;
                PartitionPolicyFlags found_flags;

                any = any || m->partitions[di].found;

                vi = partition_verity_of(di);
                si = partition_verity_sig_of(di);

                /* Determine the verity protection level for this partition. */
                found_flags = m->partitions[di].found ?
                        (vi >= 0 && m->partitions[vi].found ?
                         (si >= 0 && m->partitions[si].found ? PARTITION_POLICY_SIGNED : PARTITION_POLICY_VERITY) :
                         PARTITION_POLICY_ENCRYPTED|PARTITION_POLICY_UNPROTECTED) :
                        (m->partitions[di].ignored ? PARTITION_POLICY_UNUSED : PARTITION_POLICY_ABSENT);

                r = image_policy_check_protection(policy, di, found_flags);
                if (r < 0)
                        return r;

                if (m->partitions[di].found) {
                        r = image_policy_check_partition_flags(policy, di, m->partitions[di].gpt_flags);
                        if (r < 0)
                                return r;
                }
        }

        if (!any && !FLAGS_SET(flags, DISSECT_IMAGE_ALLOW_EMPTY))
                return -ENOMSG;

        r = dissected_image_probe_filesystems(m, fd, policy);
        if (r < 0)
                return r;

        return 0;
}
#endif

int dissect_image_file(
                const char *path,
                const VeritySettings *verity,
                const MountOptions *mount_options,
                const ImagePolicy *image_policy,
                DissectImageFlags flags,
                DissectedImage **ret) {

#if HAVE_BLKID
        _cleanup_(dissected_image_unrefp) DissectedImage *m = NULL;
        _cleanup_close_ int fd = -EBADF;
        struct stat st;
        int r;

        assert(path);

        fd = open(path, O_RDONLY|O_CLOEXEC|O_NONBLOCK|O_NOCTTY);
        if (fd < 0)
                return -errno;

        if (fstat(fd, &st) < 0)
                return -errno;

        r = stat_verify_regular(&st);
        if (r < 0)
                return r;

        r = dissected_image_new(path, &m);
        if (r < 0)
                return r;

        m->image_size = st.st_size;

        r = probe_sector_size(fd, &m->sector_size);
        if (r < 0)
                return r;

        r = dissect_image(m, fd, path, verity, mount_options, image_policy, flags);
        if (r < 0)
                return r;

        if (ret)
                *ret = TAKE_PTR(m);
        return 0;
#else
        return -EOPNOTSUPP;
#endif
}

int dissect_log_error(int log_level, int r, const char *name, const VeritySettings *verity) {
        assert(log_level >= 0 && log_level <= LOG_DEBUG);
        assert(name);

        switch (r) {

        case 0 ... INT_MAX: /* success! */
                return r;

        case -EOPNOTSUPP:
                return log_full_errno(log_level, r, "Dissecting images is not supported, compiled without blkid support.");

        case -ENOPKG:
                return log_full_errno(log_level, r, "%s: Couldn't identify a suitable partition table or file system.", name);

        case -ENOMEDIUM:
                return log_full_errno(log_level, r, "%s: The image does not pass os-release/extension-release validation.", name);

        case -EADDRNOTAVAIL:
                return log_full_errno(log_level, r, "%s: No root partition for specified root hash found.", name);

        case -ENOTUNIQ:
                return log_full_errno(log_level, r, "%s: Multiple suitable root partitions found in image.", name);

        case -ENXIO:
                return log_full_errno(log_level, r, "%s: No suitable root partition found in image.", name);

        case -EPROTONOSUPPORT:
                return log_full_errno(log_level, r, "Device '%s' is a loopback block device with partition scanning turned off, please turn it on.", name);

        case -ENOTBLK:
                return log_full_errno(log_level, r, "%s: Image is not a block device.", name);

        case -EBADR:
                return log_full_errno(log_level, r,
                                      "Combining partitioned images (such as '%s') with external Verity data (such as '%s') not supported. "
                                      "(Consider setting $SYSTEMD_DISSECT_VERITY_SIDECAR=0 to disable automatic discovery of external Verity data.)",
                                      name, strna(verity ? verity->data_path : NULL));

        case -ERFKILL:
                return log_full_errno(log_level, r, "%s: image does not match image policy.", name);

        case -ENOMSG:
                return log_full_errno(log_level, r, "%s: no suitable partitions found.", name);

        default:
                return log_full_errno(log_level, r, "%s: cannot dissect image: %m", name);
        }
}

int dissect_image_file_and_warn(
                const char *path,
                const VeritySettings *verity,
                const MountOptions *mount_options,
                const ImagePolicy *image_policy,
                DissectImageFlags flags,
                DissectedImage **ret) {

        return dissect_log_error(
                        LOG_ERR,
                        dissect_image_file(path, verity, mount_options, image_policy, flags, ret),
                        path,
                        verity);
}

void dissected_image_close(DissectedImage *m) {
        if (!m)
                return;

        /* Closes all fds we keep open associated with this, but nothing else */

        FOREACH_ARRAY(p, m->partitions, _PARTITION_DESIGNATOR_MAX) {
                p->mount_node_fd = safe_close(p->mount_node_fd);
                p->fsmount_fd = safe_close(p->fsmount_fd);
        }

        m->loop = loop_device_unref(m->loop);
}

DissectedImage* dissected_image_unref(DissectedImage *m) {
        if (!m)
                return NULL;

        /* First, clear dissected partitions. */
        for (PartitionDesignator i = 0; i < _PARTITION_DESIGNATOR_MAX; i++)
                dissected_partition_done(m->partitions + i);

        /* Second, free decrypted images. This must be after dissected_partition_done(), as freeing
         * DecryptedImage may try to deactivate partitions. */
        decrypted_image_unref(m->decrypted_image);

        /* Third, unref LoopDevice. This must be called after the above two, as freeing LoopDevice may try to
         * remove existing partitions on the loopback block device. */
        loop_device_unref(m->loop);

        free(m->image_name);
        free(m->hostname);
        strv_free(m->machine_info);
        strv_free(m->os_release);
        strv_free(m->initrd_release);
        strv_free(m->confext_release);
        strv_free(m->sysext_release);

        return mfree(m);
}

static int is_loop_device(const char *path) {
        char s[SYS_BLOCK_PATH_MAX("/../loop/")];
        struct stat st;

        assert(path);

        if (stat(path, &st) < 0)
                return -errno;

        if (!S_ISBLK(st.st_mode))
                return -ENOTBLK;

        xsprintf_sys_block_path(s, "/loop/", st.st_dev);
        if (access(s, F_OK) < 0) {
                if (errno != ENOENT)
                        return -errno;

                /* The device itself isn't a loop device, but maybe it's a partition and its parent is? */
                xsprintf_sys_block_path(s, "/../loop/", st.st_dev);
                if (access(s, F_OK) < 0)
                        return errno == ENOENT ? false : -errno;
        }

        return true;
}

static int run_fsck(int node_fd, const char *fstype) {
        int r, exit_status;
        pid_t pid;

        assert(node_fd >= 0);
        assert(fstype);

        r = fsck_exists_for_fstype(fstype);
        if (r < 0) {
                log_debug_errno(r, "Couldn't determine whether fsck for %s exists, proceeding anyway.", fstype);
                return 0;
        }
        if (r == 0) {
                log_debug("Not checking partition %s, as fsck for %s does not exist.", FORMAT_PROC_FD_PATH(node_fd), fstype);
                return 0;
        }

        r = safe_fork_full(
                        "(fsck)",
                        NULL,
                        &node_fd, 1, /* Leave the node fd open */
                        FORK_RESET_SIGNALS|FORK_CLOSE_ALL_FDS|FORK_RLIMIT_NOFILE_SAFE|FORK_DEATHSIG_SIGTERM|FORK_REARRANGE_STDIO|FORK_CLOEXEC_OFF,
                        &pid);
        if (r < 0)
                return log_debug_errno(r, "Failed to fork off fsck: %m");
        if (r == 0) {
                /* Child */
                execlp("fsck", "fsck", "-aT", FORMAT_PROC_FD_PATH(node_fd), NULL);
                log_open();
                log_debug_errno(errno, "Failed to execl() fsck: %m");
                _exit(FSCK_OPERATIONAL_ERROR);
        }

        exit_status = wait_for_terminate_and_check("fsck", pid, 0);
        if (exit_status < 0)
                return log_debug_errno(exit_status, "Failed to fork off fsck: %m");

        if ((exit_status & ~FSCK_ERROR_CORRECTED) != FSCK_SUCCESS) {
                log_debug("fsck failed with exit status %i.", exit_status);

                if ((exit_status & (FSCK_SYSTEM_SHOULD_REBOOT|FSCK_ERRORS_LEFT_UNCORRECTED)) != 0)
                        return log_debug_errno(SYNTHETIC_ERRNO(EUCLEAN), "File system is corrupted, refusing.");

                log_debug("Ignoring fsck error.");
        }

        return 0;
}

static int fs_grow(const char *node_path, int mount_fd, const char *mount_path) {
        _cleanup_close_ int _mount_fd = -EBADF, node_fd = -EBADF;
        uint64_t size, newsize;
        const char *id;
        int r;

        assert(node_path);
        assert(mount_fd >= 0 || mount_path);

        node_fd = open(node_path, O_RDONLY|O_CLOEXEC|O_NONBLOCK|O_NOCTTY);
        if (node_fd < 0)
                return log_debug_errno(errno, "Failed to open node device %s: %m", node_path);

        r = blockdev_get_device_size(node_fd, &size);
        if (r < 0)
                return log_debug_errno(r, "Failed to get block device size of %s: %m", node_path);

        if (mount_fd < 0) {
                assert(mount_path);

                _mount_fd = open(mount_path, O_RDONLY|O_DIRECTORY|O_CLOEXEC);
                if (_mount_fd < 0)
                        return log_debug_errno(errno, "Failed to open mounted file system %s: %m", mount_path);

                mount_fd = _mount_fd;
        } else {
                mount_fd = fd_reopen_condition(mount_fd, O_RDONLY|O_DIRECTORY|O_CLOEXEC, O_RDONLY|O_DIRECTORY|O_CLOEXEC, &_mount_fd);
                if (mount_fd < 0)
                        return log_debug_errno(errno, "Failed to reopen mount node: %m");
        }

        id = mount_path ?: node_path;

        log_debug("Resizing \"%s\" to %"PRIu64" bytes...", id, size);
        r = resize_fs(mount_fd, size, &newsize);
        if (r < 0)
                return log_debug_errno(r, "Failed to resize \"%s\" to %"PRIu64" bytes: %m", id, size);

        if (newsize == size)
                log_debug("Successfully resized \"%s\" to %s bytes.",
                          id, FORMAT_BYTES(newsize));
        else {
                assert(newsize < size);
                log_debug("Successfully resized \"%s\" to %s bytes (%"PRIu64" bytes lost due to blocksize).",
                          id, FORMAT_BYTES(newsize), size - newsize);
        }

        return 0;
}

int partition_pick_mount_options(
                PartitionDesignator d,
                const char *fstype,
                bool rw,
                bool discard,
                char **ret_options,
                unsigned long *ret_ms_flags) {

        _cleanup_free_ char *options = NULL;

        assert(ret_options);

        /* Selects a baseline of bind mount flags, that should always apply.
         *
         * Firstly, we set MS_NODEV universally on all mounts, since we don't want to allow device nodes outside of /dev/.
         *
         * On /var/tmp/ we'll also set MS_NOSUID, same as we set for /tmp/ on the host.
         *
         * On the ESP and XBOOTLDR partitions we'll also disable symlinks, and execution. These file systems
         * are generally untrusted (i.e. not encrypted or authenticated), and typically VFAT hence we should
         * be as restrictive as possible, and this shouldn't hurt, since the functionality is not available
         * there anyway. */

        unsigned long flags = MS_NODEV;

        if (!rw)
                flags |= MS_RDONLY;

        switch (d) {

        case PARTITION_ESP:
        case PARTITION_XBOOTLDR:
                flags |= MS_NOSUID|MS_NOEXEC|ms_nosymfollow_supported();

                /* The ESP might contain a pre-boot random seed. Let's make this unaccessible to regular
                 * userspace. ESP/XBOOTLDR is almost certainly VFAT, hence if we don't know assume it is. */
                if (!fstype || fstype_can_umask(fstype))
                        if (!strextend_with_separator(&options, ",", "umask=0077"))
                                return -ENOMEM;
                break;

        case PARTITION_TMP:
                flags |= MS_NOSUID;
                break;

        default:
                break;
        }

        /* So, when you request MS_RDONLY from ext4, then this means nothing. It happily still writes to the
         * backing storage. What's worse, the BLKRO[GS]ET flag and (in case of loopback devices)
         * LO_FLAGS_READ_ONLY don't mean anything, they affect userspace accesses only, and write accesses
         * from the upper file system still get propagated through to the underlying file system,
         * unrestricted. To actually get ext4/xfs/btrfs to stop writing to the device we need to specify
         * "norecovery" as mount option, in addition to MS_RDONLY. Yes, this sucks, since it means we need to
         * carry a per file system table here.
         *
         * Note that this means that we might not be able to mount corrupted file systems as read-only
         * anymore (since in some cases the kernel implementations will refuse mounting when corrupted,
         * read-only and "norecovery" is specified). But I think for the case of automatically determined
         * mount options for loopback devices this is the right choice, since otherwise using the same
         * loopback file twice even in read-only mode, is going to fail badly sooner or later. The use case of
         * making reuse of the immutable images "just work" is more relevant to us than having read-only
         * access that actually modifies stuff work on such image files. Or to say this differently: if
         * people want their file systems to be fixed up they should just open them in writable mode, where
         * all these problems don't exist. */
        if (!rw && fstype) {
                const char *option = fstype_norecovery_option(fstype);

                if (option && !strextend_with_separator(&options, ",", option))
                        return -ENOMEM;
        }

        if (discard && fstype && fstype_can_discard(fstype))
                if (!strextend_with_separator(&options, ",", "discard"))
                        return -ENOMEM;

        if (!ret_ms_flags) /* Fold flags into option string if ret_flags specified as NULL */
                if (!strextend_with_separator(&options, ",",
                                              FLAGS_SET(flags, MS_RDONLY) ? "ro" : "rw",
                                              FLAGS_SET(flags, MS_NODEV) ? "nodev" : "dev",
                                              FLAGS_SET(flags, MS_NOSUID) ? "nosuid" : "suid",
                                              FLAGS_SET(flags, MS_NOEXEC) ? "noexec" : "exec",
                                              FLAGS_SET(flags, MS_NOSYMFOLLOW) ? "nosymfollow" : NULL))
                        /* NB: we suppress 'symfollow' here, since it's the default, and old /bin/mount might not know it */
                        return -ENOMEM;

        if (ret_ms_flags)
                *ret_ms_flags = flags;

        *ret_options = TAKE_PTR(options);
        return 0;
}

static bool need_user_mapping(uid_t uid_shift, uid_t uid_range) {

        if (!uid_is_valid(uid_shift))
                return false;

        return uid_shift != 0 || uid_range != UINT32_MAX;
}

static int mount_partition(
                PartitionDesignator d,
                DissectedPartition *m,
                const char *where,
                const char *directory,
                uid_t uid_shift,
                uid_t uid_range,
                int userns_fd,
                DissectImageFlags flags) {

        _cleanup_free_ char *chased = NULL, *options = NULL;
        const char *p = NULL, *node, *fstype = NULL;
        bool rw, discard, grow;
        unsigned long ms_flags;
        int r;

        assert(m);

        if (!m->found)
                return 0;

        /* Check the various combinations when we can't do anything anymore */
        if (m->fsmount_fd < 0 && m->mount_node_fd < 0)
                return 0;
        if (m->fsmount_fd >= 0 && !where)
                return 0;
        if (!where && m->mount_node_fd < 0)
                return 0;

        if (m->fsmount_fd < 0) {
                fstype = dissected_partition_fstype(m);
                if (!fstype)
                        return -EAFNOSUPPORT;

                /* We are looking at an encrypted partition? This either means stacked encryption, or the
                 * caller didn't call dissected_image_decrypt() beforehand. Let's return a recognizable error
                 * for this case. */
                if (streq(fstype, "crypto_LUKS"))
                        return -EUNATCH;

                r = dissect_fstype_ok(fstype);
                if (r < 0)
                        return r;
                if (!r)
                        return -EIDRM; /* Recognizable error */
        }

        node = m->mount_node_fd < 0 ? NULL : FORMAT_PROC_FD_PATH(m->mount_node_fd);
        rw = m->rw && !(flags & DISSECT_IMAGE_MOUNT_READ_ONLY);

        discard = ((flags & DISSECT_IMAGE_DISCARD) ||
                   ((flags & DISSECT_IMAGE_DISCARD_ON_LOOP) && (m->node && is_loop_device(m->node) > 0)));

        grow = rw && m->growfs && FLAGS_SET(flags, DISSECT_IMAGE_GROWFS);

        if (FLAGS_SET(flags, DISSECT_IMAGE_FSCK) && rw && m->mount_node_fd >= 0 && m->fsmount_fd < 0) {
                r = run_fsck(m->mount_node_fd, fstype);
                if (r < 0)
                        return r;
        }

        if (where) {
                if (directory) {
                        /* Automatically create missing mount points inside the image, if necessary. */
                        r = mkdir_p_root(where, directory, uid_shift, (gid_t) uid_shift, 0755);
                        if (r < 0 && r != -EROFS)
                                return r;

                        r = chase(directory, where, CHASE_PREFIX_ROOT, &chased, NULL);
                        if (r < 0)
                                return r;

                        p = chased;
                } else {
                        /* Create top-level mount if missing – but only if this is asked for. This won't modify the
                         * image (as the branch above does) but the host hierarchy, and the created directory might
                         * survive our mount in the host hierarchy hence. */
                        if (FLAGS_SET(flags, DISSECT_IMAGE_MKDIR)) {
                                r = mkdir_p(where, 0755);
                                if (r < 0)
                                        return r;
                        }

                        p = where;
                }
        }

        if (m->fsmount_fd < 0) {
                r = partition_pick_mount_options(d, fstype, rw, discard, &options, &ms_flags);
                if (r < 0)
                        return r;

                if (need_user_mapping(uid_shift, uid_range) && fstype_can_uid_gid(fstype)) {
                        _cleanup_free_ char *uid_option = NULL;

                        if (asprintf(&uid_option, "uid=" UID_FMT ",gid=" GID_FMT, uid_shift, (gid_t) uid_shift) < 0)
                                return -ENOMEM;

                        if (!strextend_with_separator(&options, ",", uid_option))
                                return -ENOMEM;

                        userns_fd = -EBADF; /* Not needed */
                }

                if (!isempty(m->mount_options))
                        if (!strextend_with_separator(&options, ",", m->mount_options))
                                return -ENOMEM;
        }

        if (p) {
                if (m->fsmount_fd >= 0) {
                        /* Case #1: Attach existing fsmount fd to the file system */

                        r = mount_exchange_graceful(
                                        m->fsmount_fd,
                                        p,
                                        FLAGS_SET(flags, DISSECT_IMAGE_TRY_ATOMIC_MOUNT_EXCHANGE));
                        if (r < 0)
                                return log_debug_errno(r, "Failed to mount image on '%s': %m", p);

                } else {
                        assert(node);

                        /* Case #2: Mount directly into place */
                        r = mount_nofollow_verbose(LOG_DEBUG, node, p, fstype, ms_flags, options);
                        if (r < 0)
                                return r;

                        if (grow)
                                (void) fs_grow(node, -EBADF, p);

                        if (userns_fd >= 0) {
                                r = remount_idmap_fd(STRV_MAKE(p), userns_fd);
                                if (r < 0)
                                        return r;
                        }
                }
        } else {
                assert(node);

                /* Case #3: Create fsmount fd */

                m->fsmount_fd = make_fsmount(LOG_DEBUG, node, fstype, ms_flags, options, userns_fd);
                if (m->fsmount_fd < 0)
                        return m->fsmount_fd;

                if (grow)
                        (void) fs_grow(node, m->fsmount_fd, NULL);
        }

        return 1;
}

static int mount_root_tmpfs(const char *where, uid_t uid_shift, uid_t uid_range, DissectImageFlags flags) {
        _cleanup_free_ char *options = NULL;
        int r;

        assert(where);

        /* For images that contain /usr/ but no rootfs, let's mount rootfs as tmpfs */

        if (FLAGS_SET(flags, DISSECT_IMAGE_MKDIR)) {
                r = mkdir_p(where, 0755);
                if (r < 0)
                        return r;
        }

        if (need_user_mapping(uid_shift, uid_range)) {
                if (asprintf(&options, "uid=" UID_FMT ",gid=" GID_FMT, uid_shift, (gid_t) uid_shift) < 0)
                        return -ENOMEM;
        }

        r = mount_nofollow_verbose(LOG_DEBUG, "rootfs", where, "tmpfs", MS_NODEV, options);
        if (r < 0)
                return r;

        return 1;
}

static int mount_point_is_available(const char *where, const char *path, bool missing_ok) {
        _cleanup_free_ char *p = NULL;
        int r;

        /* Check whether <path> is suitable as a mountpoint, i.e. is an empty directory
         * or does not exist at all (when missing_ok). */

        r = chase(path, where, CHASE_PREFIX_ROOT, &p, NULL);
        if (r == -ENOENT)
                return missing_ok;
        if (r < 0)
                return log_debug_errno(r, "Failed to chase \"%s\": %m", path);

        r = dir_is_empty(p, /* ignore_hidden_or_backup= */ false);
        if (r == -ENOTDIR)
                return false;
        if (r < 0)
                return log_debug_errno(r, "Failed to check directory \"%s\": %m", p);
        return r > 0;
}

int dissected_image_mount(
                DissectedImage *m,
                const char *where,
                uid_t uid_shift,
                uid_t uid_range,
                int userns_fd,
                DissectImageFlags flags) {

        _cleanup_close_ int my_userns_fd = -EBADF;
        int r;

        assert(m);

        /* If 'where' is NULL then we'll use the new mount API to create fsmount() fds for the mounts and
         * store them in DissectedPartition.fsmount_fd.
         *
         * If 'where' is not NULL then we'll either mount the partitions to the right places ourselves,
         * or use DissectedPartition.fsmount_fd and bind it to the right places.
         *
         * This allows splitting the setting up the superblocks and the binding to file systems paths into
         * two distinct and differently privileged components: one that gets the fsmount fds, and the other
         * that then applies them.
         *
         * Returns:
         *
         *  -ENXIO        → No root partition found
         *  -EMEDIUMTYPE  → DISSECT_IMAGE_VALIDATE_OS set but no os-release/extension-release file found
         *  -EUNATCH      → Encrypted partition found for which no dm-crypt was set up yet
         *  -EUCLEAN      → fsck for file system failed
         *  -EBUSY        → File system already mounted/used elsewhere (kernel)
         *  -EAFNOSUPPORT → File system type not supported or not known
         *  -EIDRM        → File system is not among allowlisted "common" file systems
         */

        if (!where && (flags & (DISSECT_IMAGE_VALIDATE_OS|DISSECT_IMAGE_VALIDATE_OS_EXT)) != 0)
                return -EOPNOTSUPP; /* for now, not supported */

        if (!(m->partitions[PARTITION_ROOT].found ||
              (m->partitions[PARTITION_USR].found && FLAGS_SET(flags, DISSECT_IMAGE_USR_NO_ROOT))))
                return -ENXIO; /* Require a root fs or at least a /usr/ fs (the latter is subject to a flag of its own) */

        if (userns_fd < 0 && need_user_mapping(uid_shift, uid_range) && FLAGS_SET(flags, DISSECT_IMAGE_MOUNT_IDMAPPED)) {

                my_userns_fd = make_userns(uid_shift, uid_range, UID_INVALID, UID_INVALID, REMOUNT_IDMAPPING_HOST_ROOT);
                if (my_userns_fd < 0)
                        return my_userns_fd;

                userns_fd = my_userns_fd;
        }

        if ((flags & DISSECT_IMAGE_MOUNT_NON_ROOT_ONLY) == 0) {

                /* First mount the root fs. If there's none we use a tmpfs. */
                if (m->partitions[PARTITION_ROOT].found) {
                        r = mount_partition(PARTITION_ROOT, m->partitions + PARTITION_ROOT, where, NULL, uid_shift, uid_range, userns_fd, flags);
                        if (r < 0)
                                return r;

                } else if (where) {
                        r = mount_root_tmpfs(where, uid_shift, uid_range, flags);
                        if (r < 0)
                                return r;
                }

                /* For us mounting root always means mounting /usr as well */
                r = mount_partition(PARTITION_USR, m->partitions + PARTITION_USR, where, "/usr", uid_shift, uid_range, userns_fd, flags);
                if (r < 0)
                        return r;
        }

        if ((flags & DISSECT_IMAGE_MOUNT_NON_ROOT_ONLY) == 0 &&
            (flags & (DISSECT_IMAGE_VALIDATE_OS|DISSECT_IMAGE_VALIDATE_OS_EXT)) != 0) {
                /* If either one of the validation flags are set, ensure that the image qualifies as
                 * one or the other (or both). */
                bool ok = false;

                assert(where);

                if (FLAGS_SET(flags, DISSECT_IMAGE_VALIDATE_OS)) {
                        r = path_is_os_tree(where);
                        if (r < 0)
                                return r;
                        if (r > 0)
                                ok = true;
                }
                if (!ok && FLAGS_SET(flags, DISSECT_IMAGE_VALIDATE_OS_EXT) && m->image_name) {
                        r = extension_has_forbidden_content(where);
                        if (r < 0)
                                return r;
                        if (r == 0) {
                                r = path_is_extension_tree(IMAGE_SYSEXT, where, m->image_name, FLAGS_SET(flags, DISSECT_IMAGE_RELAX_EXTENSION_CHECK));
                                if (r == 0)
                                        r = path_is_extension_tree(IMAGE_CONFEXT, where, m->image_name, FLAGS_SET(flags, DISSECT_IMAGE_RELAX_EXTENSION_CHECK));
                                if (r < 0)
                                        return r;
                                if (r > 0)
                                        ok = true;
                        }
                }

                if (!ok)
                        return -ENOMEDIUM;
        }

        if (flags & DISSECT_IMAGE_MOUNT_ROOT_ONLY)
                return 0;

        r = mount_partition(PARTITION_HOME, m->partitions + PARTITION_HOME, where, "/home", uid_shift, uid_range, userns_fd, flags);
        if (r < 0)
                return r;

        r = mount_partition(PARTITION_SRV, m->partitions + PARTITION_SRV, where, "/srv", uid_shift, uid_range, userns_fd, flags);
        if (r < 0)
                return r;

        r = mount_partition(PARTITION_VAR, m->partitions + PARTITION_VAR, where, "/var", uid_shift, uid_range, userns_fd, flags);
        if (r < 0)
                return r;

        r = mount_partition(PARTITION_TMP, m->partitions + PARTITION_TMP, where, "/var/tmp", uid_shift, uid_range, userns_fd, flags);
        if (r < 0)
                return r;

        int slash_boot_is_available = 0;
        if (where) {
                r = slash_boot_is_available = mount_point_is_available(where, "/boot", /* missing_ok = */ true);
                if (r < 0)
                        return r;
        }
        if (!where || slash_boot_is_available) {
                r = mount_partition(PARTITION_XBOOTLDR, m->partitions + PARTITION_XBOOTLDR, where, "/boot", uid_shift, uid_range, userns_fd, flags);
                if (r < 0)
                        return r;
                slash_boot_is_available = !r;
        }

        if (m->partitions[PARTITION_ESP].found) {
                const char *esp_path = NULL;

                if (where) {
                        /* Mount the ESP to /boot/ if it exists and is empty and we didn't already mount the
                         * XBOOTLDR partition into it. Otherwise, use /efi instead, but only if it exists
                         * and is empty. */

                        if (slash_boot_is_available) {
                                r = mount_point_is_available(where, "/boot", /* missing_ok = */ false);
                                if (r < 0)
                                        return r;
                                if (r > 0)
                                        esp_path = "/boot";
                        }

                        if (!esp_path) {
                                r = mount_point_is_available(where, "/efi", /* missing_ok = */ true);
                                if (r < 0)
                                        return r;
                                if (r > 0)
                                        esp_path = "/efi";
                        }
                }

                /* OK, let's mount the ESP now (possibly creating the dir if missing) */
                r = mount_partition(PARTITION_ESP, m->partitions + PARTITION_ESP, where, esp_path, uid_shift, uid_range, userns_fd, flags);
                if (r < 0)
                        return r;
        }

        return 0;
}

int dissected_image_mount_and_warn(
                DissectedImage *m,
                const char *where,
                uid_t uid_shift,
                uid_t uid_range,
                int userns_fd,
                DissectImageFlags flags) {

        int r;

        assert(m);

        r = dissected_image_mount(m, where, uid_shift, uid_range, userns_fd, flags);
        if (r == -ENXIO)
                return log_error_errno(r, "Failed to mount image: No root file system found in image.");
        if (r == -EMEDIUMTYPE)
                return log_error_errno(r, "Failed to mount image: No suitable os-release/extension-release file in image found.");
        if (r == -EUNATCH)
                return log_error_errno(r, "Failed to mount image: Encrypted file system discovered, but decryption not requested.");
        if (r == -EUCLEAN)
                return log_error_errno(r, "Failed to mount image: File system check on image failed.");
        if (r == -EBUSY)
                return log_error_errno(r, "Failed to mount image: File system already mounted elsewhere.");
        if (r == -EAFNOSUPPORT)
                return log_error_errno(r, "Failed to mount image: File system type not supported or not known.");
        if (r == -EIDRM)
                return log_error_errno(r, "Failed to mount image: File system is too uncommon, refused.");
        if (r < 0)
                return log_error_errno(r, "Failed to mount image: %m");

        return r;
}

#if HAVE_LIBCRYPTSETUP
struct DecryptedPartition {
        struct crypt_device *device;
        char *name;
        bool relinquished;
};
#endif

typedef struct DecryptedPartition DecryptedPartition;

struct DecryptedImage {
        unsigned n_ref;
        DecryptedPartition *decrypted;
        size_t n_decrypted;
};

static DecryptedImage* decrypted_image_free(DecryptedImage *d) {
#if HAVE_LIBCRYPTSETUP
        int r;

        if (!d)
                return NULL;

        for (size_t i = 0; i < d->n_decrypted; i++) {
                DecryptedPartition *p = d->decrypted + i;

                if (p->device && p->name && !p->relinquished) {
                        _cleanup_free_ char *node = NULL;

                        node = path_join("/dev/mapper", p->name);
                        if (node) {
                                r = btrfs_forget_device(node);
                                if (r < 0 && r != -ENOENT)
                                        log_debug_errno(r, "Failed to forget btrfs device %s, ignoring: %m", node);
                        } else
                                log_oom_debug();

                        /* Let's deactivate lazily, as the dm volume may be already/still used by other processes. */
                        r = sym_crypt_deactivate_by_name(p->device, p->name, CRYPT_DEACTIVATE_DEFERRED);
                        if (r < 0)
                                log_debug_errno(r, "Failed to deactivate encrypted partition %s", p->name);
                }

                if (p->device)
                        sym_crypt_free(p->device);
                free(p->name);
        }

        free(d->decrypted);
        free(d);
#endif
        return NULL;
}

DEFINE_TRIVIAL_REF_UNREF_FUNC(DecryptedImage, decrypted_image, decrypted_image_free);

#if HAVE_LIBCRYPTSETUP
static int decrypted_image_new(DecryptedImage **ret) {
        _cleanup_(decrypted_image_unrefp) DecryptedImage *d = NULL;

        assert(ret);

        d = new(DecryptedImage, 1);
        if (!d)
                return -ENOMEM;

        *d = (DecryptedImage) {
                .n_ref = 1,
        };

        *ret = TAKE_PTR(d);
        return 0;
}

static int make_dm_name_and_node(const void *original_node, const char *suffix, char **ret_name, char **ret_node) {
        _cleanup_free_ char *name = NULL, *node = NULL;
        const char *base;

        assert(original_node);
        assert(suffix);
        assert(ret_name);
        assert(ret_node);

        base = strrchr(original_node, '/');
        if (!base)
                base = original_node;
        else
                base++;
        if (isempty(base))
                return -EINVAL;

        name = strjoin(base, suffix);
        if (!name)
                return -ENOMEM;
        if (!filename_is_valid(name))
                return -EINVAL;

        node = path_join(sym_crypt_get_dir(), name);
        if (!node)
                return -ENOMEM;

        *ret_name = TAKE_PTR(name);
        *ret_node = TAKE_PTR(node);

        return 0;
}

static int decrypt_partition(
                DissectedPartition *m,
                const char *passphrase,
                DissectImageFlags flags,
                DecryptedImage *d) {

        _cleanup_free_ char *node = NULL, *name = NULL;
        _cleanup_(sym_crypt_freep) struct crypt_device *cd = NULL;
        _cleanup_close_ int fd = -EBADF;
        int r;

        assert(m);
        assert(d);

        if (!m->found || !m->node || !m->fstype)
                return 0;

        if (!streq(m->fstype, "crypto_LUKS"))
                return 0;

        if (!passphrase)
                return -ENOKEY;

        r = dlopen_cryptsetup();
        if (r < 0)
                return r;

        r = make_dm_name_and_node(m->node, "-decrypted", &name, &node);
        if (r < 0)
                return r;

        if (!GREEDY_REALLOC0(d->decrypted, d->n_decrypted + 1))
                return -ENOMEM;

        r = sym_crypt_init(&cd, m->node);
        if (r < 0)
                return log_debug_errno(r, "Failed to initialize dm-crypt: %m");

        cryptsetup_enable_logging(cd);

        r = sym_crypt_load(cd, CRYPT_LUKS, NULL);
        if (r < 0)
                return log_debug_errno(r, "Failed to load LUKS metadata: %m");

        r = sym_crypt_activate_by_passphrase(cd, name, CRYPT_ANY_SLOT, passphrase, strlen(passphrase),
                                             ((flags & DISSECT_IMAGE_DEVICE_READ_ONLY) ? CRYPT_ACTIVATE_READONLY : 0) |
                                             ((flags & DISSECT_IMAGE_DISCARD_ON_CRYPTO) ? CRYPT_ACTIVATE_ALLOW_DISCARDS : 0));
        if (r < 0) {
                log_debug_errno(r, "Failed to activate LUKS device: %m");
                return r == -EPERM ? -EKEYREJECTED : r;
        }

        fd = open(node, O_RDONLY|O_NONBLOCK|O_CLOEXEC|O_NOCTTY);
        if (fd < 0)
                return log_debug_errno(errno, "Failed to open %s: %m", node);

        d->decrypted[d->n_decrypted++] = (DecryptedPartition) {
                .name = TAKE_PTR(name),
                .device = TAKE_PTR(cd),
        };

        m->decrypted_node = TAKE_PTR(node);
        close_and_replace(m->mount_node_fd, fd);

        return 0;
}

static int verity_can_reuse(
                const VeritySettings *verity,
                const char *name,
                struct crypt_device **ret_cd) {

        /* If the same volume was already open, check that the root hashes match, and reuse it if they do */
        _cleanup_free_ char *root_hash_existing = NULL;
        _cleanup_(sym_crypt_freep) struct crypt_device *cd = NULL;
        struct crypt_params_verity crypt_params = {};
        size_t root_hash_existing_size;
        int r;

        assert(verity);
        assert(name);
        assert(ret_cd);

        r = sym_crypt_init_by_name(&cd, name);
        if (r < 0)
                return log_debug_errno(r, "Error opening verity device, crypt_init_by_name failed: %m");

        cryptsetup_enable_logging(cd);

        r = sym_crypt_get_verity_info(cd, &crypt_params);
        if (r < 0)
                return log_debug_errno(r, "Error opening verity device, crypt_get_verity_info failed: %m");

        root_hash_existing_size = verity->root_hash_size;
        root_hash_existing = malloc0(root_hash_existing_size);
        if (!root_hash_existing)
                return -ENOMEM;

        r = sym_crypt_volume_key_get(cd, CRYPT_ANY_SLOT, root_hash_existing, &root_hash_existing_size, NULL, 0);
        if (r < 0)
                return log_debug_errno(r, "Error opening verity device, crypt_volume_key_get failed: %m");
        if (verity->root_hash_size != root_hash_existing_size ||
            memcmp(root_hash_existing, verity->root_hash, verity->root_hash_size) != 0)
                return log_debug_errno(SYNTHETIC_ERRNO(EINVAL), "Error opening verity device, it already exists but root hashes are different.");

#if HAVE_CRYPT_ACTIVATE_BY_SIGNED_KEY
        /* Ensure that, if signatures are supported, we only reuse the device if the previous mount used the
         * same settings, so that a previous unsigned mount will not be reused if the user asks to use
         * signing for the new one, and vice versa. */
        if (!!verity->root_hash_sig != !!(crypt_params.flags & CRYPT_VERITY_ROOT_HASH_SIGNATURE))
                return log_debug_errno(SYNTHETIC_ERRNO(EINVAL), "Error opening verity device, it already exists but signature settings are not the same.");
#endif

        *ret_cd = TAKE_PTR(cd);
        return 0;
}

static char* dm_deferred_remove_clean(char *name) {
        if (!name)
                return NULL;

        (void) sym_crypt_deactivate_by_name(NULL, name, CRYPT_DEACTIVATE_DEFERRED);
        return mfree(name);
}
DEFINE_TRIVIAL_CLEANUP_FUNC(char *, dm_deferred_remove_clean);

static int validate_signature_userspace(const VeritySettings *verity, DissectImageFlags flags) {
        int r;

        if (!FLAGS_SET(flags, DISSECT_IMAGE_ALLOW_USERSPACE_VERITY)) {
                log_debug("Userspace dm-verity signature authentication disabled via flag.");
                return 0;
        }

        r = secure_getenv_bool("SYSTEMD_ALLOW_USERSPACE_VERITY");
        if (r < 0 && r != -ENXIO) {
                log_debug_errno(r, "Failed to parse $SYSTEMD_ALLOW_USERSPACE_VERITY environment variable, refusing userspace dm-verity signature authentication.");
                return 0;
        }
        if (!r) {
                log_debug("Userspace dm-verity signature authentication disabled via $SYSTEMD_ALLOW_USERSPACE_VERITY environment variable.");
                return 0;
        }

        bool b;
        r = proc_cmdline_get_bool("systemd.allow_userspace_verity", PROC_CMDLINE_TRUE_WHEN_MISSING, &b);
        if (r < 0) {
                log_debug_errno(r, "Failed to parse systemd.allow_userspace_verity= kernel command line option, refusing userspace dm-verity signature authentication.");
                return 0;
        }
        if (!b) {
                log_debug("Userspace dm-verity signature authentication disabled via systemd.allow_userspace_verity= kernel command line variable.");
                return 0;
        }

#if HAVE_OPENSSL
        _cleanup_(sk_X509_free_allp) STACK_OF(X509) *sk = NULL;
        _cleanup_strv_free_ char **certs = NULL;
        _cleanup_(PKCS7_freep) PKCS7 *p7 = NULL;
        _cleanup_free_ char *s = NULL;
        _cleanup_(BIO_freep) BIO *bio = NULL; /* 'bio' must be freed first, 's' second, hence keep this order
                                               * of declaration in place, please */
        const unsigned char *d;

        assert(verity);
        assert(verity->root_hash);
        assert(verity->root_hash_sig);

        /* Because installing a signature certificate into the kernel chain is so messy, let's optionally do
         * userspace validation. */

        r = conf_files_list_nulstr(&certs, ".crt", NULL, CONF_FILES_REGULAR|CONF_FILES_FILTER_MASKED, CONF_PATHS_NULSTR("verity.d"));
        if (r < 0)
                return log_debug_errno(r, "Failed to enumerate certificates: %m");
        if (strv_isempty(certs)) {
                log_debug("No userspace dm-verity certificates found.");
                return 0;
        }

        d = verity->root_hash_sig;
        p7 = d2i_PKCS7(NULL, &d, (long) verity->root_hash_sig_size);
        if (!p7)
                return log_debug_errno(SYNTHETIC_ERRNO(EINVAL), "Failed to parse PKCS7 DER signature data.");

        s = hexmem(verity->root_hash, verity->root_hash_size);
        if (!s)
                return log_oom_debug();

        bio = BIO_new_mem_buf(s, strlen(s));
        if (!bio)
                return log_oom_debug();

        sk = sk_X509_new_null();
        if (!sk)
                return log_oom_debug();

        STRV_FOREACH(i, certs) {
                _cleanup_(X509_freep) X509 *c = NULL;
                _cleanup_fclose_ FILE *f = NULL;

                f = fopen(*i, "re");
                if (!f) {
                        log_debug_errno(errno, "Failed to open '%s', ignoring: %m", *i);
                        continue;
                }

                c = PEM_read_X509(f, NULL, NULL, NULL);
                if (!c) {
                        log_debug("Failed to load X509 certificate '%s', ignoring.", *i);
                        continue;
                }

                if (sk_X509_push(sk, c) == 0)
                        return log_oom_debug();

                TAKE_PTR(c);
        }

        r = PKCS7_verify(p7, sk, NULL, bio, NULL, PKCS7_NOINTERN|PKCS7_NOVERIFY);
        if (r)
                log_debug("Userspace PKCS#7 validation succeeded.");
        else
                log_debug("Userspace PKCS#7 validation failed: %s", ERR_error_string(ERR_get_error(), NULL));

        return r;
#else
        log_debug("Not doing client-side validation of dm-verity root hash signatures, OpenSSL support disabled.");
        return 0;
#endif
}

static int do_crypt_activate_verity(
                struct crypt_device *cd,
                const char *name,
                const VeritySettings *verity,
                DissectImageFlags flags) {

        bool check_signature;
        int r, k;

        assert(cd);
        assert(name);
        assert(verity);

        if (verity->root_hash_sig) {
                r = secure_getenv_bool("SYSTEMD_DISSECT_VERITY_SIGNATURE");
                if (r < 0 && r != -ENXIO)
                        log_debug_errno(r, "Failed to parse $SYSTEMD_DISSECT_VERITY_SIGNATURE");

                check_signature = r != 0;
        } else
                check_signature = false;

        if (check_signature) {

#if HAVE_CRYPT_ACTIVATE_BY_SIGNED_KEY
                /* First, if we have support for signed keys in the kernel, then try that first. */
                r = sym_crypt_activate_by_signed_key(
                                cd,
                                name,
                                verity->root_hash,
                                verity->root_hash_size,
                                verity->root_hash_sig,
                                verity->root_hash_sig_size,
                                CRYPT_ACTIVATE_READONLY);
                if (r >= 0)
                        return r;

                log_debug_errno(r, "Validation of dm-verity signature failed via the kernel, trying userspace validation instead: %m");
#else
                log_debug("Activation of verity device with signature requested, but not supported via the kernel by %s due to missing crypt_activate_by_signed_key(), trying userspace validation instead.",
                          program_invocation_short_name);
                r = 0; /* Set for the propagation below */
#endif

                /* So this didn't work via the kernel, then let's try userspace validation instead. If that
                 * works we'll try to activate without telling the kernel the signature. */

                /* Preferably propagate the original kernel error, so that the fallback logic can work,
                 * as the device-mapper is finicky around concurrent activations of the same volume */
                k = validate_signature_userspace(verity, flags);
                if (k < 0)
                        return r < 0 ? r : k;
                if (k == 0)
                        return log_debug_errno(r < 0 ? r : SYNTHETIC_ERRNO(ENOKEY),
                                               "Activation of signed Verity volume worked neither via the kernel nor in userspace, can't activate.");
        }

        return sym_crypt_activate_by_volume_key(
                        cd,
                        name,
                        verity->root_hash,
                        verity->root_hash_size,
                        CRYPT_ACTIVATE_READONLY);
}

static usec_t verity_timeout(void) {
        usec_t t = 100 * USEC_PER_MSEC;
        const char *e;
        int r;

        /* On slower machines, like non-KVM vm, setting up device may take a long time.
         * Let's make the timeout configurable. */

        e = getenv("SYSTEMD_DISSECT_VERITY_TIMEOUT_SEC");
        if (!e)
                return t;

        r = parse_sec(e, &t);
        if (r < 0)
                log_debug_errno(r,
                                "Failed to parse timeout specified in $SYSTEMD_DISSECT_VERITY_TIMEOUT_SEC, "
                                "using the default timeout (%s).",
                                FORMAT_TIMESPAN(t, USEC_PER_MSEC));

        return t;
}

static int verity_partition(
                PartitionDesignator designator,
                DissectedPartition *m,
                DissectedPartition *v,
                const VeritySettings *verity,
                DissectImageFlags flags,
                DecryptedImage *d) {

        _cleanup_(sym_crypt_freep) struct crypt_device *cd = NULL;
        _cleanup_free_ char *node = NULL, *name = NULL;
        _cleanup_close_ int mount_node_fd = -EBADF;
        int r;

        assert(m);
        assert(v || (verity && verity->data_path));

        if (!verity || !verity->root_hash)
                return 0;
        if (!((verity->designator < 0 && designator == PARTITION_ROOT) ||
              (verity->designator == designator)))
                return 0;

        if (!m->found || !m->node || !m->fstype)
                return 0;
        if (!verity->data_path) {
                if (!v->found || !v->node || !v->fstype)
                        return 0;

                if (!streq(v->fstype, "DM_verity_hash"))
                        return 0;
        }

        r = dlopen_cryptsetup();
        if (r < 0)
                return r;

        if (FLAGS_SET(flags, DISSECT_IMAGE_VERITY_SHARE)) {
                /* Use the roothash, which is unique per volume, as the device node name, so that it can be reused */
                _cleanup_free_ char *root_hash_encoded = NULL;

                root_hash_encoded = hexmem(verity->root_hash, verity->root_hash_size);
                if (!root_hash_encoded)
                        return -ENOMEM;

                r = make_dm_name_and_node(root_hash_encoded, "-verity", &name, &node);
        } else
                r = make_dm_name_and_node(m->node, "-verity", &name, &node);
        if (r < 0)
                return r;

        r = sym_crypt_init(&cd, verity->data_path ?: v->node);
        if (r < 0)
                return r;

        cryptsetup_enable_logging(cd);

        r = sym_crypt_load(cd, CRYPT_VERITY, NULL);
        if (r < 0)
                return r;

        r = sym_crypt_set_data_device(cd, m->node);
        if (r < 0)
                return r;

        if (!GREEDY_REALLOC0(d->decrypted, d->n_decrypted + 1))
                return -ENOMEM;

        /* If activating fails because the device already exists, check the metadata and reuse it if it matches.
         * In case of ENODEV/ENOENT, which can happen if another process is activating at the exact same time,
         * retry a few times before giving up. */
        for (unsigned i = 0; i < N_DEVICE_NODE_LIST_ATTEMPTS; i++) {
                _cleanup_(dm_deferred_remove_cleanp) char *restore_deferred_remove = NULL;
                _cleanup_(sym_crypt_freep) struct crypt_device *existing_cd = NULL;
                _cleanup_close_ int fd = -EBADF;

                /* First, check if the device already exists. */
                fd = open(node, O_RDONLY|O_NONBLOCK|O_CLOEXEC|O_NOCTTY);
                if (fd < 0 && !ERRNO_IS_DEVICE_ABSENT(errno))
                        return log_debug_errno(errno, "Failed to open verity device %s: %m", node);
                if (fd >= 0)
                        goto check; /* The device already exists. Let's check it. */

                /* The symlink to the device node does not exist yet. Assume not activated, and let's activate it. */
                r = do_crypt_activate_verity(cd, name, verity, flags);
                if (r >= 0)
                        goto try_open; /* The device is activated. Let's open it. */
                /* libdevmapper can return EINVAL when the device is already in the activation stage.
                 * There's no way to distinguish this situation from a genuine error due to invalid
                 * parameters, so immediately fall back to activating the device with a unique name.
                 * Improvements in libcrypsetup can ensure this never happens:
                 * https://gitlab.com/cryptsetup/cryptsetup/-/merge_requests/96 */
                if (r == -EINVAL && FLAGS_SET(flags, DISSECT_IMAGE_VERITY_SHARE))
                        break;
                /* Volume is being opened but not ready, crypt_init_by_name would fail, try to open again if
                 * sharing is enabled. */
                if (r == -ENODEV && FLAGS_SET(flags, DISSECT_IMAGE_VERITY_SHARE))
                        goto try_again;
                if (!IN_SET(r,
                            -EEXIST, /* Volume has already been opened and ready to be used. */
                            -EBUSY   /* Volume is being opened but not ready, crypt_init_by_name() can fetch details. */))
                        return log_debug_errno(r, "Failed to activate verity device %s: %m", node);

        check:
                /* To avoid races, disable automatic removal on umount while setting up the new device. Restore it on failure. */
                r = dm_deferred_remove_cancel(name);
                /* -EBUSY and -ENXIO: the device has already been removed or being removed. We cannot
                 * use the device, try to open again. See target_message() in drivers/md/dm-ioctl.c
                 * and dm_cancel_deferred_remove() in drivers/md/dm.c */
                if (IN_SET(r, -EBUSY, -ENXIO))
                        goto try_again;
                if (r < 0)
                        return log_debug_errno(r, "Failed to disable automated deferred removal for verity device %s: %m", node);

                restore_deferred_remove = strdup(name);
                if (!restore_deferred_remove)
                        return log_oom_debug();

                r = verity_can_reuse(verity, name, &existing_cd);
                /* Same as above, -EINVAL can randomly happen when it actually means -EEXIST */
                if (r == -EINVAL && FLAGS_SET(flags, DISSECT_IMAGE_VERITY_SHARE))
                        break;
                if (IN_SET(r,
                           -ENOENT, /* Removed?? */
                           -EBUSY,  /* Volume is being opened but not ready, crypt_init_by_name() can fetch details. */
                           -ENODEV  /* Volume is being opened but not ready, crypt_init_by_name() would fail, try to open again. */ ))
                        goto try_again;
                if (r < 0)
                        return log_debug_errno(r, "Failed to check if existing verity device %s can be reused: %m", node);

                if (fd < 0) {
                        /* devmapper might say that the device exists, but the devlink might not yet have been
                         * created. Check and wait for the udev event in that case. */
                        r = device_wait_for_devlink(node, "block", verity_timeout(), NULL);
                        /* Fallback to activation with a unique device if it's taking too long */
                        if (r == -ETIMEDOUT && FLAGS_SET(flags, DISSECT_IMAGE_VERITY_SHARE))
                                break;
                        if (r < 0)
                                return log_debug_errno(r, "Failed to wait device node symlink %s: %m", node);
                }

        try_open:
                if (fd < 0) {
                        /* Now, the device is activated and devlink is created. Let's open it. */
                        fd = open(node, O_RDONLY|O_NONBLOCK|O_CLOEXEC|O_NOCTTY);
                        if (fd < 0) {
                                if (!ERRNO_IS_DEVICE_ABSENT(errno))
                                        return log_debug_errno(errno, "Failed to open verity device %s: %m", node);

                                /* The device has already been removed?? */
                                goto try_again;
                        }
                }

                /* Everything looks good and we'll be able to mount the device, so deferred remove will be re-enabled at that point. */
                restore_deferred_remove = mfree(restore_deferred_remove);

                mount_node_fd = TAKE_FD(fd);
                if (existing_cd)
                        crypt_free_and_replace(cd, existing_cd);

                goto success;

        try_again:
                /* Device is being removed by another process. Let's wait for a while. */
                (void) usleep_safe(2 * USEC_PER_MSEC);
        }

        /* All trials failed or a conflicting verity device exists. Let's try to activate with a unique name. */
        if (FLAGS_SET(flags, DISSECT_IMAGE_VERITY_SHARE)) {
                /* Before trying to activate with unique name, we need to free crypt_device object.
                 * Otherwise, we get error from libcryptsetup like the following:
                 * ------
                 * systemd[1234]: Cannot use device /dev/loop5 which is in use (already mapped or mounted).
                 * ------
                 */
                sym_crypt_free(cd);
                cd = NULL;
                return verity_partition(designator, m, v, verity, flags & ~DISSECT_IMAGE_VERITY_SHARE, d);
        }

        return log_debug_errno(SYNTHETIC_ERRNO(EBUSY), "All attempts to activate verity device %s failed.", name);

success:
        d->decrypted[d->n_decrypted++] = (DecryptedPartition) {
                .name = TAKE_PTR(name),
                .device = TAKE_PTR(cd),
        };

        m->decrypted_node = TAKE_PTR(node);
        close_and_replace(m->mount_node_fd, mount_node_fd);

        return 0;
}
#endif

int dissected_image_decrypt(
                DissectedImage *m,
                const char *passphrase,
                const VeritySettings *verity,
                DissectImageFlags flags) {

#if HAVE_LIBCRYPTSETUP
        _cleanup_(decrypted_image_unrefp) DecryptedImage *d = NULL;
        int r;
#endif

        assert(m);
        assert(!verity || verity->root_hash || verity->root_hash_size == 0);

        /* Returns:
         *
         *      = 0           → There was nothing to decrypt
         *      > 0           → Decrypted successfully
         *      -ENOKEY       → There's something to decrypt but no key was supplied
         *      -EKEYREJECTED → Passed key was not correct
         *      -EBUSY        → Generic Verity error (kernel is not very explanatory)
         */

        if (verity && verity->root_hash && verity->root_hash_size < sizeof(sd_id128_t))
                return -EINVAL;

        if (!m->encrypted && !m->verity_ready)
                return 0;

#if HAVE_LIBCRYPTSETUP
        r = decrypted_image_new(&d);
        if (r < 0)
                return r;

        for (PartitionDesignator i = 0; i < _PARTITION_DESIGNATOR_MAX; i++) {
                DissectedPartition *p = m->partitions + i;
                PartitionDesignator k;

                if (!p->found)
                        continue;

                r = decrypt_partition(p, passphrase, flags, d);
                if (r < 0)
                        return r;

                k = partition_verity_of(i);
                if (k >= 0) {
                        flags |= getenv_bool("SYSTEMD_VERITY_SHARING") != 0 ? DISSECT_IMAGE_VERITY_SHARE : 0;

                        r = verity_partition(i, p, m->partitions + k, verity, flags, d);
                        if (r < 0)
                                return r;
                }

                if (!p->decrypted_fstype && p->mount_node_fd >= 0 && p->decrypted_node) {
                        r = probe_filesystem_full(p->mount_node_fd, p->decrypted_node, 0, UINT64_MAX, &p->decrypted_fstype);
                        if (r < 0 && r != -EUCLEAN)
                                return r;
                }
        }

        m->decrypted_image = TAKE_PTR(d);

        return 1;
#else
        return -EOPNOTSUPP;
#endif
}

int dissected_image_decrypt_interactively(
                DissectedImage *m,
                const char *passphrase,
                const VeritySettings *verity,
                DissectImageFlags flags) {

        _cleanup_strv_free_erase_ char **z = NULL;
        int n = 3, r;

        if (passphrase)
                n--;

        for (;;) {
                r = dissected_image_decrypt(m, passphrase, verity, flags);
                if (r >= 0)
                        return r;
                if (r == -EKEYREJECTED)
                        log_error_errno(r, "Incorrect passphrase, try again!");
                else if (r != -ENOKEY)
                        return log_error_errno(r, "Failed to decrypt image: %m");

                if (--n < 0)
                        return log_error_errno(SYNTHETIC_ERRNO(EKEYREJECTED),
                                               "Too many retries.");

                z = strv_free_erase(z);

                static const AskPasswordRequest req = {
                        .message = "Please enter image passphrase:",
                        .id = "dissect",
                        .keyring = "dissect",
                        .credential = "dissect.passphrase",
                };

                r = ask_password_auto(&req, USEC_INFINITY, /* flags= */ 0, &z);
                if (r < 0)
                        return log_error_errno(r, "Failed to query for passphrase: %m");

                passphrase = z[0];
        }
}

static int decrypted_image_relinquish(DecryptedImage *d) {
        assert(d);

        /* Turns on automatic removal after the last use ended for all DM devices of this image, and sets a
         * boolean so that we don't clean it up ourselves either anymore */

#if HAVE_LIBCRYPTSETUP
        int r;

        for (size_t i = 0; i < d->n_decrypted; i++) {
                DecryptedPartition *p = d->decrypted + i;

                if (p->relinquished)
                        continue;

                r = sym_crypt_deactivate_by_name(NULL, p->name, CRYPT_DEACTIVATE_DEFERRED);
                if (r < 0)
                        return log_debug_errno(r, "Failed to mark %s for auto-removal: %m", p->name);

                p->relinquished = true;
        }
#endif

        return 0;
}

int dissected_image_relinquish(DissectedImage *m) {
        int r;

        assert(m);

        if (m->decrypted_image) {
                r = decrypted_image_relinquish(m->decrypted_image);
                if (r < 0)
                        return r;
        }

        if (m->loop)
                loop_device_relinquish(m->loop);

        return 0;
}

static char *build_auxiliary_path(const char *image, const char *suffix) {
        const char *e;
        char *n;

        assert(image);
        assert(suffix);

        e = endswith(image, ".raw");
        if (!e)
                return strjoin(e, suffix);

        n = new(char, e - image + strlen(suffix) + 1);
        if (!n)
                return NULL;

        strcpy(mempcpy(n, image, e - image), suffix);
        return n;
}

void verity_settings_done(VeritySettings *v) {
        assert(v);

        v->root_hash = mfree(v->root_hash);
        v->root_hash_size = 0;

        v->root_hash_sig = mfree(v->root_hash_sig);
        v->root_hash_sig_size = 0;

        v->data_path = mfree(v->data_path);
}

int verity_settings_load(
                VeritySettings *verity,
                const char *image,
                const char *root_hash_path,
                const char *root_hash_sig_path) {

        _cleanup_free_ void *root_hash = NULL, *root_hash_sig = NULL;
        size_t root_hash_size = 0, root_hash_sig_size = 0;
        _cleanup_free_ char *verity_data_path = NULL;
        PartitionDesignator designator;
        int r;

        assert(verity);
        assert(image);
        assert(verity->designator < 0 || IN_SET(verity->designator, PARTITION_ROOT, PARTITION_USR));

        /* If we are asked to load the root hash for a device node, exit early */
        if (is_device_path(image))
                return 0;

        r = secure_getenv_bool("SYSTEMD_DISSECT_VERITY_SIDECAR");
        if (r < 0 && r != -ENXIO)
                log_debug_errno(r, "Failed to parse $SYSTEMD_DISSECT_VERITY_SIDECAR, ignoring: %m");
        if (r == 0)
                return 0;

        designator = verity->designator;

        /* We only fill in what isn't already filled in */

        if (!verity->root_hash) {
                _cleanup_free_ char *text = NULL;

                if (root_hash_path) {
                        /* If explicitly specified it takes precedence */
                        r = read_one_line_file(root_hash_path, &text);
                        if (r < 0)
                                return r;

                        if (designator < 0)
                                designator = PARTITION_ROOT;
                } else {
                        /* Otherwise look for xattr and separate file, and first for the data for root and if
                         * that doesn't exist for /usr */

                        if (designator < 0 || designator == PARTITION_ROOT) {
                                r = getxattr_malloc(image, "user.verity.roothash", &text);
                                if (r < 0) {
                                        _cleanup_free_ char *p = NULL;

                                        if (r != -ENOENT && !ERRNO_IS_XATTR_ABSENT(r))
                                                return r;

                                        p = build_auxiliary_path(image, ".roothash");
                                        if (!p)
                                                return -ENOMEM;

                                        r = read_one_line_file(p, &text);
                                        if (r < 0 && r != -ENOENT)
                                                return r;
                                }

                                if (text)
                                        designator = PARTITION_ROOT;
                        }

                        if (!text && (designator < 0 || designator == PARTITION_USR)) {
                                /* So in the "roothash" xattr/file name above the "root" of course primarily
                                 * refers to the root of the Verity Merkle tree. But coincidentally it also
                                 * is the hash for the *root* file system, i.e. the "root" neatly refers to
                                 * two distinct concepts called "root". Taking benefit of this happy
                                 * coincidence we call the file with the root hash for the /usr/ file system
                                 * `usrhash`, because `usrroothash` or `rootusrhash` would just be too
                                 * confusing. We thus drop the reference to the root of the Merkle tree, and
                                 * just indicate which file system it's about. */
                                r = getxattr_malloc(image, "user.verity.usrhash", &text);
                                if (r < 0) {
                                        _cleanup_free_ char *p = NULL;

                                        if (r != -ENOENT && !ERRNO_IS_XATTR_ABSENT(r))
                                                return r;

                                        p = build_auxiliary_path(image, ".usrhash");
                                        if (!p)
                                                return -ENOMEM;

                                        r = read_one_line_file(p, &text);
                                        if (r < 0 && r != -ENOENT)
                                                return r;
                                }

                                if (text)
                                        designator = PARTITION_USR;
                        }
                }

                if (text) {
                        r = unhexmem(text, &root_hash, &root_hash_size);
                        if (r < 0)
                                return r;
                        if (root_hash_size < sizeof(sd_id128_t))
                                return -EINVAL;
                }
        }

        if ((root_hash || verity->root_hash) && !verity->root_hash_sig) {
                if (root_hash_sig_path) {
                        r = read_full_file(root_hash_sig_path, (char**) &root_hash_sig, &root_hash_sig_size);
                        if (r < 0 && r != -ENOENT)
                                return r;

                        if (designator < 0)
                                designator = PARTITION_ROOT;
                } else {
                        if (designator < 0 || designator == PARTITION_ROOT) {
                                _cleanup_free_ char *p = NULL;

                                /* Follow naming convention recommended by the relevant RFC:
                                 * https://tools.ietf.org/html/rfc5751#section-3.2.1 */
                                p = build_auxiliary_path(image, ".roothash.p7s");
                                if (!p)
                                        return -ENOMEM;

                                r = read_full_file(p, (char**) &root_hash_sig, &root_hash_sig_size);
                                if (r < 0 && r != -ENOENT)
                                        return r;
                                if (r >= 0)
                                        designator = PARTITION_ROOT;
                        }

                        if (!root_hash_sig && (designator < 0 || designator == PARTITION_USR)) {
                                _cleanup_free_ char *p = NULL;

                                p = build_auxiliary_path(image, ".usrhash.p7s");
                                if (!p)
                                        return -ENOMEM;

                                r = read_full_file(p, (char**) &root_hash_sig, &root_hash_sig_size);
                                if (r < 0 && r != -ENOENT)
                                        return r;
                                if (r >= 0)
                                        designator = PARTITION_USR;
                        }
                }

                if (root_hash_sig && root_hash_sig_size == 0) /* refuse empty size signatures */
                        return -EINVAL;
        }

        if (!verity->data_path) {
                _cleanup_free_ char *p = NULL;

                p = build_auxiliary_path(image, ".verity");
                if (!p)
                        return -ENOMEM;

                if (access(p, F_OK) < 0) {
                        if (errno != ENOENT)
                                return -errno;
                } else
                        verity_data_path = TAKE_PTR(p);
        }

        if (root_hash) {
                verity->root_hash = TAKE_PTR(root_hash);
                verity->root_hash_size = root_hash_size;
        }

        if (root_hash_sig) {
                verity->root_hash_sig = TAKE_PTR(root_hash_sig);
                verity->root_hash_sig_size = root_hash_sig_size;
        }

        if (verity_data_path)
                verity->data_path = TAKE_PTR(verity_data_path);

        if (verity->designator < 0)
                verity->designator = designator;

        return 1;
}

int dissected_image_load_verity_sig_partition(
                DissectedImage *m,
                int fd,
                VeritySettings *verity) {

        _cleanup_free_ void *root_hash = NULL, *root_hash_sig = NULL;
        _cleanup_(json_variant_unrefp) JsonVariant *v = NULL;
        size_t root_hash_size, root_hash_sig_size;
        _cleanup_free_ char *buf = NULL;
        PartitionDesignator d;
        DissectedPartition *p;
        JsonVariant *rh, *sig;
        ssize_t n;
        char *e;
        int r;

        assert(m);
        assert(fd >= 0);
        assert(verity);

        if (verity->root_hash && verity->root_hash_sig) /* Already loaded? */
                return 0;

        r = secure_getenv_bool("SYSTEMD_DISSECT_VERITY_EMBEDDED");
        if (r < 0 && r != -ENXIO)
                log_debug_errno(r, "Failed to parse $SYSTEMD_DISSECT_VERITY_EMBEDDED, ignoring: %m");
        if (r == 0)
                return 0;

        d = partition_verity_sig_of(verity->designator < 0 ? PARTITION_ROOT : verity->designator);
        assert(d >= 0);

        p = m->partitions + d;
        if (!p->found)
                return 0;
        if (p->offset == UINT64_MAX || p->size == UINT64_MAX)
                return -EINVAL;

        if (p->size > 4*1024*1024) /* Signature data cannot possible be larger than 4M, refuse that */
                return log_debug_errno(SYNTHETIC_ERRNO(EFBIG), "Verity signature partition is larger than 4M, refusing.");

        buf = new(char, p->size+1);
        if (!buf)
                return -ENOMEM;

        n = pread(fd, buf, p->size, p->offset);
        if (n < 0)
                return -ENOMEM;
        if ((uint64_t) n != p->size)
                return -EIO;

        e = memchr(buf, 0, p->size);
        if (e) {
                /* If we found a NUL byte then the rest of the data must be NUL too */
                if (!memeqzero(e, p->size - (e - buf)))
                        return log_debug_errno(SYNTHETIC_ERRNO(EINVAL), "Signature data contains embedded NUL byte.");
        } else
                buf[p->size] = 0;

        r = json_parse(buf, 0, &v, NULL, NULL);
        if (r < 0)
                return log_debug_errno(r, "Failed to parse signature JSON data: %m");

        rh = json_variant_by_key(v, "rootHash");
        if (!rh)
                return log_debug_errno(SYNTHETIC_ERRNO(EINVAL), "Signature JSON object lacks 'rootHash' field.");
        if (!json_variant_is_string(rh))
                return log_debug_errno(SYNTHETIC_ERRNO(EINVAL), "'rootHash' field of signature JSON object is not a string.");

        r = unhexmem(json_variant_string(rh), &root_hash, &root_hash_size);
        if (r < 0)
                return log_debug_errno(r, "Failed to parse root hash field: %m");

        /* Check if specified root hash matches if it is specified */
        if (verity->root_hash &&
            memcmp_nn(verity->root_hash, verity->root_hash_size, root_hash, root_hash_size) != 0) {
                _cleanup_free_ char *a = NULL, *b = NULL;

                a = hexmem(root_hash, root_hash_size);
                b = hexmem(verity->root_hash, verity->root_hash_size);

                return log_debug_errno(r, "Root hash in signature JSON data (%s) doesn't match configured hash (%s).", strna(a), strna(b));
        }

        sig = json_variant_by_key(v, "signature");
        if (!sig)
                return log_debug_errno(SYNTHETIC_ERRNO(EINVAL), "Signature JSON object lacks 'signature' field.");
        if (!json_variant_is_string(sig))
                return log_debug_errno(SYNTHETIC_ERRNO(EINVAL), "'signature' field of signature JSON object is not a string.");

        r = unbase64mem(json_variant_string(sig), &root_hash_sig, &root_hash_sig_size);
        if (r < 0)
                return log_debug_errno(r, "Failed to parse signature field: %m");

        free_and_replace(verity->root_hash, root_hash);
        verity->root_hash_size = root_hash_size;

        free_and_replace(verity->root_hash_sig, root_hash_sig);
        verity->root_hash_sig_size = root_hash_sig_size;

        return 1;
}

int dissected_image_acquire_metadata(
                DissectedImage *m,
                int userns_fd,
                DissectImageFlags extra_flags) {

        enum {
                META_HOSTNAME,
                META_MACHINE_ID,
                META_MACHINE_INFO,
                META_OS_RELEASE,
                META_INITRD_RELEASE,
                META_SYSEXT_RELEASE,
                META_CONFEXT_RELEASE,
                META_HAS_INIT_SYSTEM,
                _META_MAX,
        };

        static const char *const paths[_META_MAX] = {
                [META_HOSTNAME]          = "/etc/hostname\0",
                [META_MACHINE_ID]        = "/etc/machine-id\0",
                [META_MACHINE_INFO]      = "/etc/machine-info\0",
                [META_OS_RELEASE]        = "/etc/os-release\0"
                                           "/usr/lib/os-release\0",
                [META_INITRD_RELEASE]    = "/etc/initrd-release\0"
                                           "/usr/lib/initrd-release\0",
                [META_SYSEXT_RELEASE]    = "sysext-release\0",       /* String used only for logging. */
                [META_CONFEXT_RELEASE]   = "confext-release\0",      /* ditto */
                [META_HAS_INIT_SYSTEM]   = "has-init-system\0",      /* ditto */
        };

        _cleanup_strv_free_ char **machine_info = NULL, **os_release = NULL, **initrd_release = NULL, **sysext_release = NULL, **confext_release = NULL;
        _cleanup_free_ char *hostname = NULL, *t = NULL;
        _cleanup_close_pair_ int error_pipe[2] = EBADF_PAIR;
        _cleanup_(sigkill_waitp) pid_t child = 0;
        sd_id128_t machine_id = SD_ID128_NULL;
        unsigned n_meta_initialized = 0;
        int fds[2 * _META_MAX], r, v;
        int has_init_system = -1;
        ssize_t n;

        BLOCK_SIGNALS(SIGCHLD);

        assert(m);

        for (; n_meta_initialized < _META_MAX; n_meta_initialized++) {
                assert(paths[n_meta_initialized]);

                if (pipe2(fds + 2*n_meta_initialized, O_CLOEXEC) < 0) {
                        r = -errno;
                        goto finish;
                }
        }

        r = get_common_dissect_directory(&t);
        if (r < 0)
                goto finish;

        if (pipe2(error_pipe, O_CLOEXEC) < 0) {
                r = -errno;
                goto finish;
        }

        r = safe_fork("(sd-dissect)", FORK_RESET_SIGNALS|FORK_DEATHSIG_SIGTERM, &child);
        if (r < 0)
                goto finish;
        if (r == 0) {
                /* Child */
                error_pipe[0] = safe_close(error_pipe[0]);

                if (userns_fd < 0)
                        r = detach_mount_namespace_harder(0, 0);
                else
                        r = detach_mount_namespace_userns(userns_fd);
                if (r < 0) {
                        log_debug_errno(r, "Failed to detach mount namespace: %m");
                        goto inner_fail;
                }

                r = dissected_image_mount(
                                m,
                                t,
                                /* uid_shift= */ UID_INVALID,
                                /* uid_range= */ UID_INVALID,
                                /* userns_fd= */ -EBADF,
                                extra_flags |
                                DISSECT_IMAGE_READ_ONLY |
                                DISSECT_IMAGE_MOUNT_ROOT_ONLY |
                                DISSECT_IMAGE_USR_NO_ROOT);
                if (r < 0) {
                        log_debug_errno(r, "Failed to mount dissected image: %m");
                        goto inner_fail;
                }

                for (unsigned k = 0; k < _META_MAX; k++) {
                        _cleanup_close_ int fd = -ENOENT;

                        assert(paths[k]);

                        fds[2*k] = safe_close(fds[2*k]);

                        switch (k) {

                        case META_SYSEXT_RELEASE:
                                if (!m->image_name)
                                        goto next;

                                /* As per the os-release spec, if the image is an extension it will have a
                                 * file named after the image name in extension-release.d/ - we use the image
                                 * name and try to resolve it with the extension-release helpers, as
                                 * sometimes the image names are mangled on deployment and do not match
                                 * anymore.  Unlike other paths this is not fixed, and the image name can be
                                 * mangled on deployment, so by calling into the helper we allow a fallback
                                 * that matches on the first extension-release file found in the directory,
                                 * if one named after the image cannot be found first. */
                                r = open_extension_release(
                                                t,
                                                IMAGE_SYSEXT,
                                                m->image_name,
                                                /* relax_extension_release_check= */ false,
                                                /* ret_path= */ NULL,
                                                &fd);
                                if (r < 0)
                                        fd = r;
                                break;

                        case META_CONFEXT_RELEASE:
                                if (!m->image_name)
                                        goto next;

                                /* As above */
                                r = open_extension_release(
                                                t,
                                                IMAGE_CONFEXT,
                                                m->image_name,
                                                /* relax_extension_release_check= */ false,
                                                /* ret_path= */ NULL,
                                                &fd);
                                if (r < 0)
                                        fd = r;

                                break;

                        case META_HAS_INIT_SYSTEM: {
                                bool found = false;

                                FOREACH_STRING(init,
                                               "/usr/lib/systemd/systemd",  /* systemd on /usr/ merged system */
                                               "/lib/systemd/systemd",      /* systemd on /usr/ non-merged systems */
                                               "/sbin/init") {              /* traditional path the Linux kernel invokes */

                                        r = chase(init, t, CHASE_PREFIX_ROOT, NULL, NULL);
                                        if (r < 0) {
                                                if (r != -ENOENT)
                                                        log_debug_errno(r, "Failed to resolve %s, ignoring: %m", init);
                                        } else {
                                                found = true;
                                                break;
                                        }
                                }

                                r = loop_write(fds[2*k+1], &found, sizeof(found));
                                if (r < 0)
                                        goto inner_fail;

                                goto next;
                        }

                        default:
                                NULSTR_FOREACH(p, paths[k]) {
                                        fd = chase_and_open(p, t, CHASE_PREFIX_ROOT, O_RDONLY|O_CLOEXEC|O_NOCTTY, NULL);
                                        if (fd >= 0)
                                                break;
                                }
                        }

                        if (fd < 0) {
                                log_debug_errno(fd, "Failed to read %s file of image, ignoring: %m", paths[k]);
                                goto next;
                        }

                        r = copy_bytes(fd, fds[2*k+1], UINT64_MAX, 0);
                        if (r < 0)
                                goto inner_fail;

                next:
                        fds[2*k+1] = safe_close(fds[2*k+1]);
                }

                _exit(EXIT_SUCCESS);

        inner_fail:
                /* Let parent know the error */
                (void) write(error_pipe[1], &r, sizeof(r));
                _exit(EXIT_FAILURE);
        }

        error_pipe[1] = safe_close(error_pipe[1]);

        for (unsigned k = 0; k < _META_MAX; k++) {
                _cleanup_fclose_ FILE *f = NULL;

                assert(paths[k]);

                fds[2*k+1] = safe_close(fds[2*k+1]);

                f = take_fdopen(&fds[2*k], "r");
                if (!f) {
                        r = -errno;
                        goto finish;
                }

                switch (k) {

                case META_HOSTNAME:
                        r = read_etc_hostname_stream(f, &hostname);
                        if (r < 0)
                                log_debug_errno(r, "Failed to read /etc/hostname of image: %m");

                        break;

                case META_MACHINE_ID: {
                        _cleanup_free_ char *line = NULL;

                        r = read_line(f, LONG_LINE_MAX, &line);
                        if (r < 0)
                                log_debug_errno(r, "Failed to read /etc/machine-id of image: %m");
                        else if (r == 33) {
                                r = sd_id128_from_string(line, &machine_id);
                                if (r < 0)
                                        log_debug_errno(r, "Image contains invalid /etc/machine-id: %s", line);
                        } else if (r == 0)
                                log_debug("/etc/machine-id file of image is empty.");
                        else if (streq(line, "uninitialized"))
                                log_debug("/etc/machine-id file of image is uninitialized (likely aborted first boot).");
                        else
                                log_debug("/etc/machine-id file of image has unexpected length %i.", r);

                        break;
                }

                case META_MACHINE_INFO:
                        r = load_env_file_pairs(f, "machine-info", &machine_info);
                        if (r < 0)
                                log_debug_errno(r, "Failed to read /etc/machine-info of image: %m");

                        break;

                case META_OS_RELEASE:
                        r = load_env_file_pairs(f, "os-release", &os_release);
                        if (r < 0)
                                log_debug_errno(r, "Failed to read OS release file of image: %m");

                        break;

                case META_INITRD_RELEASE:
                        r = load_env_file_pairs(f, "initrd-release", &initrd_release);
                        if (r < 0)
                                log_debug_errno(r, "Failed to read initrd release file of image: %m");

                        break;

                case META_SYSEXT_RELEASE:
                        r = load_env_file_pairs(f, "sysext-release", &sysext_release);
                        if (r < 0)
                                log_debug_errno(r, "Failed to read sysext release file of image: %m");

                        break;

                case META_CONFEXT_RELEASE:
                        r = load_env_file_pairs(f, "confext-release", &confext_release);
                        if (r < 0)
                                log_debug_errno(r, "Failed to read confext release file of image: %m");

                        break;

                case META_HAS_INIT_SYSTEM: {
                        bool b = false;
                        size_t nr;

                        errno = 0;
                        nr = fread(&b, 1, sizeof(b), f);
                        if (nr != sizeof(b))
                                log_debug_errno(errno_or_else(EIO), "Failed to read has-init-system boolean: %m");
                        else
                                has_init_system = b;

                        break;
                }}
        }

        r = wait_for_terminate_and_check("(sd-dissect)", child, 0);
        child = 0;
        if (r < 0)
                goto finish;

        n = read(error_pipe[0], &v, sizeof(v));
        if (n < 0) {
                r = -errno;
                goto finish;
        }
        if (n == sizeof(v)) {
                r = v; /* propagate error sent to us from child */
                goto finish;
        }
        if (n != 0) {
                r = -EIO;
                goto finish;
        }
        if (r != EXIT_SUCCESS) {
                r = -EPROTO;
                goto finish;
        }

        free_and_replace(m->hostname, hostname);
        m->machine_id = machine_id;
        strv_free_and_replace(m->machine_info, machine_info);
        strv_free_and_replace(m->os_release, os_release);
        strv_free_and_replace(m->initrd_release, initrd_release);
        strv_free_and_replace(m->sysext_release, sysext_release);
        strv_free_and_replace(m->confext_release, confext_release);
        m->has_init_system = has_init_system;

finish:
        for (unsigned k = 0; k < n_meta_initialized; k++)
                safe_close_pair(fds + 2*k);

        return r;
}

Architecture dissected_image_architecture(DissectedImage *img) {
        assert(img);

        if (img->partitions[PARTITION_ROOT].found &&
            img->partitions[PARTITION_ROOT].architecture >= 0)
                return img->partitions[PARTITION_ROOT].architecture;

        if (img->partitions[PARTITION_USR].found &&
            img->partitions[PARTITION_USR].architecture >= 0)
                return img->partitions[PARTITION_USR].architecture;

        return _ARCHITECTURE_INVALID;
}

int dissect_loop_device(
                LoopDevice *loop,
                const VeritySettings *verity,
                const MountOptions *mount_options,
                const ImagePolicy *image_policy,
                DissectImageFlags flags,
                DissectedImage **ret) {

#if HAVE_BLKID
        _cleanup_(dissected_image_unrefp) DissectedImage *m = NULL;
        int r;

        assert(loop);

        r = dissected_image_new(loop->backing_file ?: loop->node, &m);
        if (r < 0)
                return r;

        m->loop = loop_device_ref(loop);
        m->image_size = m->loop->device_size;
        m->sector_size = m->loop->sector_size;

        r = dissect_image(m, loop->fd, loop->node, verity, mount_options, image_policy, flags);
        if (r < 0)
                return r;

        if (ret)
                *ret = TAKE_PTR(m);

        return 0;
#else
        return -EOPNOTSUPP;
#endif
}

int dissect_loop_device_and_warn(
                LoopDevice *loop,
                const VeritySettings *verity,
                const MountOptions *mount_options,
                const ImagePolicy *image_policy,
                DissectImageFlags flags,
                DissectedImage **ret) {

        assert(loop);

        return dissect_log_error(
                        LOG_ERR,
                        dissect_loop_device(loop, verity, mount_options, image_policy, flags, ret),
                        loop->backing_file ?: loop->node,
                        verity);

}

bool dissected_image_verity_candidate(const DissectedImage *image, PartitionDesignator partition_designator) {
        assert(image);

        /* Checks if this partition could theoretically do Verity. For non-partitioned images this only works
         * if there's an external verity file supplied, for which we can consult .has_verity. For partitioned
         * images we only check the partition type.
         *
         * This call is used to decide whether to suppress or show a verity column in tabular output of the
         * image. */

        if (image->single_file_system)
                return partition_designator == PARTITION_ROOT && image->has_verity;

        return partition_verity_of(partition_designator) >= 0;
}

bool dissected_image_verity_ready(const DissectedImage *image, PartitionDesignator partition_designator) {
        PartitionDesignator k;

        assert(image);

        /* Checks if this partition has verity data available that we can activate. For non-partitioned this
         * works for the root partition, for others only if the associated verity partition was found. */

        if (!image->verity_ready)
                return false;

        if (image->single_file_system)
                return partition_designator == PARTITION_ROOT;

        k = partition_verity_of(partition_designator);
        return k >= 0 && image->partitions[k].found;
}

bool dissected_image_verity_sig_ready(const DissectedImage *image, PartitionDesignator partition_designator) {
        PartitionDesignator k;

        assert(image);

        /* Checks if this partition has verity signature data available that we can use. */

        if (!image->verity_sig_ready)
                return false;

        if (image->single_file_system)
                return partition_designator == PARTITION_ROOT;

        k = partition_verity_sig_of(partition_designator);
        return k >= 0 && image->partitions[k].found;
}

MountOptions* mount_options_free_all(MountOptions *options) {
        MountOptions *m;

        while ((m = LIST_POP(mount_options, options))) {
                free(m->options);
                free(m);
        }

        return NULL;
}

const char* mount_options_from_designator(const MountOptions *options, PartitionDesignator designator) {
        LIST_FOREACH(mount_options, m, options)
                if (designator == m->partition_designator && !isempty(m->options))
                        return m->options;

        return NULL;
}

int mount_image_privately_interactively(
                const char *image,
                const ImagePolicy *image_policy,
                DissectImageFlags flags,
                char **ret_directory,
                int *ret_dir_fd,
                LoopDevice **ret_loop_device) {

        _cleanup_(verity_settings_done) VeritySettings verity = VERITY_SETTINGS_DEFAULT;
        _cleanup_(loop_device_unrefp) LoopDevice *d = NULL;
        _cleanup_(dissected_image_unrefp) DissectedImage *dissected_image = NULL;
        _cleanup_free_ char *dir = NULL;
        int r;

        /* Mounts an OS image at a temporary place, inside a newly created mount namespace of our own. This
         * is used by tools such as systemd-tmpfiles or systemd-firstboot to operate on some disk image
         * easily. */

        assert(image);
        assert(ret_loop_device);

        /* We intend to mount this right-away, hence add the partitions if needed and pin them. */
        flags |= DISSECT_IMAGE_ADD_PARTITION_DEVICES |
                DISSECT_IMAGE_PIN_PARTITION_DEVICES;

        r = verity_settings_load(&verity, image, NULL, NULL);
        if (r < 0)
                return log_error_errno(r, "Failed to load root hash data: %m");

        r = loop_device_make_by_path(
                        image,
                        FLAGS_SET(flags, DISSECT_IMAGE_DEVICE_READ_ONLY) ? O_RDONLY : O_RDWR,
                        /* sector_size= */ UINT32_MAX,
                        FLAGS_SET(flags, DISSECT_IMAGE_NO_PARTITION_TABLE) ? 0 : LO_FLAGS_PARTSCAN,
                        LOCK_SH,
                        &d);
        if (r < 0)
                return log_error_errno(r, "Failed to set up loopback device for %s: %m", image);

        r = dissect_loop_device_and_warn(
                        d,
                        &verity,
                        /* mount_options= */ NULL,
                        image_policy,
                        flags,
                        &dissected_image);
        if (r < 0)
                return r;

        r = dissected_image_load_verity_sig_partition(dissected_image, d->fd, &verity);
        if (r < 0)
                return r;

        r = dissected_image_decrypt_interactively(dissected_image, NULL, &verity, flags);
        if (r < 0)
                return r;

        r = detach_mount_namespace();
        if (r < 0)
                return log_error_errno(r, "Failed to detach mount namespace: %m");

        r = mkdir_p("/run/systemd/mount-rootfs", 0555);
        if (r < 0)
                return log_error_errno(r, "Failed to create mount point: %m");

        r = dissected_image_mount_and_warn(
                        dissected_image,
                        "/run/systemd/mount-rootfs",
                        /* uid_shift= */ UID_INVALID,
                        /* uid_range= */ UID_INVALID,
                        /* userns_fd= */ -EBADF,
                        flags);
        if (r < 0)
                return r;

        r = loop_device_flock(d, LOCK_UN);
        if (r < 0)
                return r;

        r = dissected_image_relinquish(dissected_image);
        if (r < 0)
                return log_error_errno(r, "Failed to relinquish DM and loopback block devices: %m");

        if (ret_directory) {
                dir = strdup("/run/systemd/mount-rootfs");
                if (!dir)
                        return log_oom();
        }

        if (ret_dir_fd) {
                _cleanup_close_ int dir_fd = -EBADF;

                dir_fd = open("/run/systemd/mount-rootfs", O_CLOEXEC|O_DIRECTORY);
                if (dir_fd < 0)
                        return log_error_errno(errno, "Failed to open mount point directory: %m");

                *ret_dir_fd = TAKE_FD(dir_fd);
        }

        if (ret_directory)
                *ret_directory = TAKE_PTR(dir);

        *ret_loop_device = TAKE_PTR(d);
        return 0;
}

static bool mount_options_relax_extension_release_checks(const MountOptions *options) {
        if (!options)
                return false;

        return string_contains_word(mount_options_from_designator(options, PARTITION_ROOT), ",", "x-systemd.relax-extension-release-check") ||
                        string_contains_word(mount_options_from_designator(options, PARTITION_USR), ",", "x-systemd.relax-extension-release-check") ||
                        string_contains_word(options->options, ",", "x-systemd.relax-extension-release-check");
}

int verity_dissect_and_mount(
                int src_fd,
                const char *src,
                const char *dest,
                const MountOptions *options,
                const ImagePolicy *image_policy,
                const char *required_host_os_release_id,
                const char *required_host_os_release_version_id,
                const char *required_host_os_release_sysext_level,
                const char *required_host_os_release_confext_level,
                const char *required_sysext_scope,
                DissectedImage **ret_image) {

        _cleanup_(loop_device_unrefp) LoopDevice *loop_device = NULL;
        _cleanup_(dissected_image_unrefp) DissectedImage *dissected_image = NULL;
        _cleanup_(verity_settings_done) VeritySettings verity = VERITY_SETTINGS_DEFAULT;
        DissectImageFlags dissect_image_flags;
        bool relax_extension_release_check;
        int r;

        assert(src);
        /* Verifying release metadata requires mounted image for now, so ensure the check is skipped when
         * opening an image without mounting it immediately (i.e.: 'dest' is NULL). */
        assert(!required_host_os_release_id || dest);

        relax_extension_release_check = mount_options_relax_extension_release_checks(options);

        /* We might get an FD for the image, but we use the original path to look for the dm-verity files */
        r = verity_settings_load(&verity, src, NULL, NULL);
        if (r < 0)
                return log_debug_errno(r, "Failed to load root hash: %m");

        dissect_image_flags =
                (verity.data_path ? DISSECT_IMAGE_NO_PARTITION_TABLE : 0) |
                (relax_extension_release_check ? DISSECT_IMAGE_RELAX_EXTENSION_CHECK : 0) |
                DISSECT_IMAGE_ADD_PARTITION_DEVICES |
                DISSECT_IMAGE_PIN_PARTITION_DEVICES |
                DISSECT_IMAGE_ALLOW_USERSPACE_VERITY;

        /* Note that we don't use loop_device_make here, as the FD is most likely O_PATH which would not be
         * accepted by LOOP_CONFIGURE, so just let loop_device_make_by_path reopen it as a regular FD. */
        r = loop_device_make_by_path(
                        src_fd >= 0 ? FORMAT_PROC_FD_PATH(src_fd) : src,
                        /* open_flags= */ -1,
                        /* sector_size= */ UINT32_MAX,
                        verity.data_path ? 0 : LO_FLAGS_PARTSCAN,
                        LOCK_SH,
                        &loop_device);
        if (r < 0)
                return log_debug_errno(r, "Failed to create loop device for image: %m");

        r = dissect_loop_device(
                        loop_device,
                        &verity,
                        options,
                        image_policy,
                        dissect_image_flags,
                        &dissected_image);
        /* No partition table? Might be a single-filesystem image, try again */
        if (!verity.data_path && r == -ENOPKG)
                 r = dissect_loop_device(
                                loop_device,
                                &verity,
                                options,
                                image_policy,
                                dissect_image_flags | DISSECT_IMAGE_NO_PARTITION_TABLE,
                                &dissected_image);
        if (r < 0)
                return log_debug_errno(r, "Failed to dissect image: %m");

        r = dissected_image_load_verity_sig_partition(dissected_image, loop_device->fd, &verity);
        if (r < 0)
                return r;

        r = dissected_image_decrypt(
                        dissected_image,
                        NULL,
                        &verity,
                        dissect_image_flags);
        if (r < 0)
                return log_debug_errno(r, "Failed to decrypt dissected image: %m");

        if (dest) {
                r = mkdir_p_label(dest, 0755);
                if (r < 0)
                        return log_debug_errno(r, "Failed to create destination directory %s: %m", dest);
                r = umount_recursive(dest, 0);
                if (r < 0)
                        return log_debug_errno(r, "Failed to umount under destination directory %s: %m", dest);
        }

        r = dissected_image_mount(
                        dissected_image,
                        dest,
                        /* uid_shift= */ UID_INVALID,
                        /* uid_range= */ UID_INVALID,
                        /* userns_fd= */ -EBADF,
                        dissect_image_flags);
        if (r < 0)
                return log_debug_errno(r, "Failed to mount image: %m");

        r = loop_device_flock(loop_device, LOCK_UN);
        if (r < 0)
                return log_debug_errno(r, "Failed to unlock loopback device: %m");

        /* If we got os-release values from the caller, then we need to match them with the image's
         * extension-release.d/ content. Return -EINVAL if there's any mismatch.
         * First, check the distro ID. If that matches, then check the new SYSEXT_LEVEL value if
         * available, or else fallback to VERSION_ID. If neither is present (eg: rolling release),
         * then a simple match on the ID will be performed. */
        if (required_host_os_release_id) {
                _cleanup_strv_free_ char **extension_release = NULL;
                ImageClass class = IMAGE_SYSEXT;

                assert(!isempty(required_host_os_release_id));

                r = load_extension_release_pairs(dest, IMAGE_SYSEXT, dissected_image->image_name, relax_extension_release_check, &extension_release);
                if (r == -ENOENT) {
                        r = load_extension_release_pairs(dest, IMAGE_CONFEXT, dissected_image->image_name, relax_extension_release_check, &extension_release);
                        if (r >= 0)
                                class = IMAGE_CONFEXT;
                }
                if (r < 0)
                        return log_debug_errno(r, "Failed to parse image %s extension-release metadata: %m", dissected_image->image_name);

                r = extension_release_validate(
                                dissected_image->image_name,
                                required_host_os_release_id,
                                required_host_os_release_version_id,
                                class == IMAGE_SYSEXT ? required_host_os_release_sysext_level : required_host_os_release_confext_level,
                                required_sysext_scope,
                                extension_release,
                                class);
                if (r == 0)
                        return log_debug_errno(SYNTHETIC_ERRNO(ESTALE), "Image %s extension-release metadata does not match the root's", dissected_image->image_name);
                if (r < 0)
                        return log_debug_errno(r, "Failed to compare image %s extension-release metadata with the root's os-release: %m", dissected_image->image_name);
        }

        r = dissected_image_relinquish(dissected_image);
        if (r < 0)
                return log_debug_errno(r, "Failed to relinquish dissected image: %m");

        if (ret_image)
                *ret_image = TAKE_PTR(dissected_image);

        return 0;
}

int get_common_dissect_directory(char **ret) {
        _cleanup_free_ char *t = NULL;
        int r;

        /* A common location we mount dissected images to. The assumption is that everyone who uses this
         * function runs in their own private mount namespace (with mount propagation off on /run/systemd/,
         * and thus can mount something here without affecting anyone else). */

        t = strdup("/run/systemd/dissect-root");
        if (!t)
                return log_oom_debug();

        r = mkdir_parents(t, 0755);
        if (r < 0)
                return log_debug_errno(r, "Failed to create parent dirs of mount point '%s': %m", t);

        r = RET_NERRNO(mkdir(t, 0000)); /* It's supposed to be overmounted, hence let's make this inaccessible */
        if (r < 0 && r != -EEXIST)
                return log_debug_errno(r, "Failed to create mount point '%s': %m", t);

        if (ret)
                *ret = TAKE_PTR(t);

        return 0;
}

#if HAVE_BLKID

static JSON_DISPATCH_ENUM_DEFINE(dispatch_architecture, Architecture, architecture_from_string);
static JSON_DISPATCH_ENUM_DEFINE(dispatch_partition_designator, PartitionDesignator, partition_designator_from_string);

typedef struct PartitionFields {
        PartitionDesignator designator;
        bool rw;
        bool growfs;
        unsigned partno;
        Architecture architecture;
        sd_id128_t uuid;
        char *fstype;
        char *label;
        uint64_t size;
        uint64_t offset;
        unsigned fsmount_fd_idx;
} PartitionFields;

static void partition_fields_done(PartitionFields *f) {
        assert(f);

        f->fstype = mfree(f->fstype);
        f->label = mfree(f->label);
}

typedef struct ReplyParameters {
        JsonVariant *partitions;
        char *image_policy;
        uint64_t image_size;
        uint32_t sector_size;
        sd_id128_t image_uuid;
} ReplyParameters;

static void reply_parameters_done(ReplyParameters *p) {
        assert(p);

        p->image_policy = mfree(p->image_policy);
        p->partitions = json_variant_unref(p->partitions);
}

#endif

int mountfsd_mount_image(
                const char *path,
                int userns_fd,
                const ImagePolicy *image_policy,
                DissectImageFlags flags,
                DissectedImage **ret) {

#if HAVE_BLKID
        _cleanup_(reply_parameters_done) ReplyParameters p = {};

        static const JsonDispatch dispatch_table[] = {
                { "partitions",  JSON_VARIANT_ARRAY,         json_dispatch_variant,  offsetof(struct ReplyParameters, partitions),   JSON_MANDATORY },
                { "imagePolicy", JSON_VARIANT_STRING,        json_dispatch_string,   offsetof(struct ReplyParameters, image_policy), 0              },
                { "imageSize",   _JSON_VARIANT_TYPE_INVALID, json_dispatch_uint64,   offsetof(struct ReplyParameters, image_size),   JSON_MANDATORY },
                { "sectorSize",  _JSON_VARIANT_TYPE_INVALID, json_dispatch_uint32,   offsetof(struct ReplyParameters, sector_size),  JSON_MANDATORY },
                { "imageUuid",   JSON_VARIANT_STRING,        json_dispatch_id128,    offsetof(struct ReplyParameters, image_uuid),   0              },
                {}
        };

        _cleanup_(dissected_image_unrefp) DissectedImage *di = NULL;
        _cleanup_close_ int image_fd = -EBADF;
        _cleanup_(varlink_unrefp) Varlink *vl = NULL;
        _cleanup_free_ char *ps = NULL;
        unsigned max_fd = UINT_MAX;
        const char *error_id;
        int r;

        assert(path);
        assert(ret);

        r = varlink_connect_address(&vl, "/run/systemd/io.systemd.MountFileSystem");
        if (r < 0)
                return log_error_errno(r, "Failed to connect to mountfsd: %m");

        r = varlink_set_allow_fd_passing_input(vl, true);
        if (r < 0)
                return log_error_errno(r, "Failed to enable varlink fd passing for read: %m");

        r = varlink_set_allow_fd_passing_output(vl, true);
        if (r < 0)
                return log_error_errno(r, "Failed to enable varlink fd passing for write: %m");

        image_fd = open(path, O_RDONLY|O_CLOEXEC);
        if (image_fd < 0)
                return log_error_errno(errno, "Failed to open '%s': %m", path);

        r = varlink_push_dup_fd(vl, image_fd);
        if (r < 0)
                return log_error_errno(r, "Failed to push image fd into varlink connection: %m");

        if (userns_fd >= 0) {
                r = varlink_push_dup_fd(vl, userns_fd);
                if (r < 0)
                        return log_error_errno(r, "Failed to push image fd into varlink connection: %m");
        }

        if (image_policy) {
                r = image_policy_to_string(image_policy, /* simplify= */ false, &ps);
                if (r < 0)
                        return log_error_errno(r, "Failed format image policy to string: %m");
        }

        JsonVariant *reply = NULL;
        r = varlink_callb(
                        vl,
                        "io.systemd.MountFileSystem.MountImage",
                        &reply,
                        &error_id,
                        JSON_BUILD_OBJECT(
                                        JSON_BUILD_PAIR("imageFileDescriptor", JSON_BUILD_UNSIGNED(0)),
                                        JSON_BUILD_PAIR_CONDITION(userns_fd >= 0, "userNamespaceFileDescriptor", JSON_BUILD_UNSIGNED(1)),
                                        JSON_BUILD_PAIR("readOnly", JSON_BUILD_BOOLEAN(FLAGS_SET(flags, DISSECT_IMAGE_MOUNT_READ_ONLY))),
                                        JSON_BUILD_PAIR("growFileSystems", JSON_BUILD_BOOLEAN(FLAGS_SET(flags, DISSECT_IMAGE_GROWFS))),
                                        JSON_BUILD_PAIR_CONDITION(ps, "imagePolicy", JSON_BUILD_STRING(ps)),
                                        JSON_BUILD_PAIR("allowInteractiveAuthentication", JSON_BUILD_BOOLEAN(FLAGS_SET(flags, DISSECT_IMAGE_ALLOW_INTERACTIVE_AUTH)))));
        if (r < 0)
                return log_error_errno(r, "Failed to call MountImage() varlink call: %m");
        if (!isempty(error_id))
                return log_error_errno(varlink_error_to_errno(error_id, reply), "Failed to call MountImage() varlink call: %s", error_id);

        r = json_dispatch(reply, dispatch_table, JSON_ALLOW_EXTENSIONS, &p);
        if (r < 0)
                return log_error_errno(r, "Failed to parse MountImage() reply: %m");

        log_debug("Effective image policy: %s", p.image_policy);

        JsonVariant *i;
        JSON_VARIANT_ARRAY_FOREACH(i, p.partitions) {
                _cleanup_close_ int fsmount_fd = -EBADF;

                _cleanup_(partition_fields_done) PartitionFields pp = {
                        .designator = _PARTITION_DESIGNATOR_INVALID,
                        .architecture = _ARCHITECTURE_INVALID,
                        .size = UINT64_MAX,
                        .offset = UINT64_MAX,
                        .fsmount_fd_idx = UINT_MAX,
                };

                static const JsonDispatch partition_dispatch_table[] = {
                        { "designator",          JSON_VARIANT_STRING,        dispatch_partition_designator, offsetof(struct PartitionFields, designator),       JSON_MANDATORY },
                        { "writable",            JSON_VARIANT_BOOLEAN,       json_dispatch_boolean,         offsetof(struct PartitionFields, rw),               JSON_MANDATORY },
                        { "growFileSystem",      JSON_VARIANT_BOOLEAN,       json_dispatch_boolean,         offsetof(struct PartitionFields, growfs),           JSON_MANDATORY },
                        { "partitionNumber",     _JSON_VARIANT_TYPE_INVALID, json_dispatch_uint,            offsetof(struct PartitionFields, partno),           0              },
                        { "architecture",        JSON_VARIANT_STRING,        dispatch_architecture,         offsetof(struct PartitionFields, architecture),     0              },
                        { "partitionUuid",       JSON_VARIANT_STRING,        json_dispatch_id128,           offsetof(struct PartitionFields, uuid),             0              },
                        { "fileSystemType",      JSON_VARIANT_STRING,        json_dispatch_string,          offsetof(struct PartitionFields, fstype),           JSON_MANDATORY },
                        { "partitionLabel",      JSON_VARIANT_STRING,        json_dispatch_string,          offsetof(struct PartitionFields, label),            0              },
                        { "size",                _JSON_VARIANT_TYPE_INVALID, json_dispatch_uint64,          offsetof(struct PartitionFields, size),             JSON_MANDATORY },
                        { "offset",              _JSON_VARIANT_TYPE_INVALID, json_dispatch_uint64,          offsetof(struct PartitionFields, offset),           JSON_MANDATORY },
                        { "mountFileDescriptor", _JSON_VARIANT_TYPE_INVALID, json_dispatch_uint,            offsetof(struct PartitionFields, fsmount_fd_idx),   JSON_MANDATORY },
                        {}
                };

                r = json_dispatch(i, partition_dispatch_table, JSON_ALLOW_EXTENSIONS, &pp);
                if (r < 0)
                        return log_error_errno(r, "Failed to parse partition data: %m");

                if (pp.fsmount_fd_idx != UINT_MAX) {
                        if (max_fd == UINT_MAX || pp.fsmount_fd_idx > max_fd)
                                max_fd = pp.fsmount_fd_idx;

                        fsmount_fd = varlink_take_fd(vl, pp.fsmount_fd_idx);
                        if (fsmount_fd < 0)
                                return fsmount_fd;
                }

                assert(pp.designator >= 0);

                if (!di) {
                        r = dissected_image_new(path, &di);
                        if (r < 0)
                                return log_error_errno(r, "Failed to allocated new dissected image structure: %m");
                }

                if (di->partitions[pp.designator].found)
                        return log_error_errno(SYNTHETIC_ERRNO(EBADMSG), "Duplicate partition data for '%s'.", partition_designator_to_string(pp.designator));

                di->partitions[pp.designator] = (DissectedPartition) {
                        .found = true,
                        .rw = pp.rw,
                        .growfs = pp.growfs,
                        .partno = pp.partno,
                        .architecture = pp.architecture,
                        .uuid = pp.uuid,
                        .fstype = TAKE_PTR(pp.fstype),
                        .label = TAKE_PTR(pp.label),
                        .mount_node_fd = -EBADF,
                        .size = pp.size,
                        .offset = pp.offset,
                        .fsmount_fd = TAKE_FD(fsmount_fd),
                };
        }

        di->image_size = p.image_size;
        di->sector_size = p.sector_size;
        di->image_uuid = p.image_uuid;

        *ret = TAKE_PTR(di);
        return 0;
#else
        return -EOPNOTSUPP;
#endif
}