aboutsummaryrefslogtreecommitdiff
path: root/sys/kern/imgact_elf.c
blob: 245894926ee16fa348dd7ee2c846f76b1459cf50 (plain) (blame)
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
/*-
 * SPDX-License-Identifier: BSD-3-Clause
 *
 * Copyright (c) 2017 Dell EMC
 * Copyright (c) 2000-2001, 2003 David O'Brien
 * Copyright (c) 1995-1996 Søren Schmidt
 * Copyright (c) 1996 Peter Wemm
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer
 *    in this position and unchanged.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 * 3. The name of the author may not be used to endorse or promote products
 *    derived from this software without specific prior written permission
 *
 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
 * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */

#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");

#include "opt_capsicum.h"

#include <sys/param.h>
#include <sys/capsicum.h>
#include <sys/compressor.h>
#include <sys/exec.h>
#include <sys/fcntl.h>
#include <sys/imgact.h>
#include <sys/imgact_elf.h>
#include <sys/jail.h>
#include <sys/kernel.h>
#include <sys/lock.h>
#include <sys/malloc.h>
#include <sys/mount.h>
#include <sys/mman.h>
#include <sys/namei.h>
#include <sys/proc.h>
#include <sys/procfs.h>
#include <sys/ptrace.h>
#include <sys/racct.h>
#include <sys/resourcevar.h>
#include <sys/rwlock.h>
#include <sys/sbuf.h>
#include <sys/sf_buf.h>
#include <sys/smp.h>
#include <sys/systm.h>
#include <sys/signalvar.h>
#include <sys/stat.h>
#include <sys/sx.h>
#include <sys/syscall.h>
#include <sys/sysctl.h>
#include <sys/sysent.h>
#include <sys/vnode.h>
#include <sys/syslog.h>
#include <sys/eventhandler.h>
#include <sys/user.h>

#include <vm/vm.h>
#include <vm/vm_kern.h>
#include <vm/vm_param.h>
#include <vm/pmap.h>
#include <vm/vm_map.h>
#include <vm/vm_object.h>
#include <vm/vm_extern.h>

#include <machine/elf.h>
#include <machine/md_var.h>

#define ELF_NOTE_ROUNDSIZE	4
#define OLD_EI_BRAND	8

static int __elfN(check_header)(const Elf_Ehdr *hdr);
static Elf_Brandinfo *__elfN(get_brandinfo)(struct image_params *imgp,
    const char *interp, int32_t *osrel, uint32_t *fctl0);
static int __elfN(load_file)(struct proc *p, const char *file, u_long *addr,
    u_long *entry);
static int __elfN(load_section)(struct image_params *imgp, vm_ooffset_t offset,
    caddr_t vmaddr, size_t memsz, size_t filsz, vm_prot_t prot);
static int __CONCAT(exec_, __elfN(imgact))(struct image_params *imgp);
static bool __elfN(freebsd_trans_osrel)(const Elf_Note *note,
    int32_t *osrel);
static bool kfreebsd_trans_osrel(const Elf_Note *note, int32_t *osrel);
static boolean_t __elfN(check_note)(struct image_params *imgp,
    Elf_Brandnote *checknote, int32_t *osrel, boolean_t *has_fctl0,
    uint32_t *fctl0);
static vm_prot_t __elfN(trans_prot)(Elf_Word);
static Elf_Word __elfN(untrans_prot)(vm_prot_t);

SYSCTL_NODE(_kern, OID_AUTO, __CONCAT(elf, __ELF_WORD_SIZE),
    CTLFLAG_RW | CTLFLAG_MPSAFE, 0,
    "");

#define	CORE_BUF_SIZE	(16 * 1024)

int __elfN(fallback_brand) = -1;
SYSCTL_INT(__CONCAT(_kern_elf, __ELF_WORD_SIZE), OID_AUTO,
    fallback_brand, CTLFLAG_RWTUN, &__elfN(fallback_brand), 0,
    __XSTRING(__CONCAT(ELF, __ELF_WORD_SIZE)) " brand of last resort");

static int elf_legacy_coredump = 0;
SYSCTL_INT(_debug, OID_AUTO, __elfN(legacy_coredump), CTLFLAG_RW, 
    &elf_legacy_coredump, 0,
    "include all and only RW pages in core dumps");

int __elfN(nxstack) =
#if defined(__amd64__) || defined(__powerpc64__) /* both 64 and 32 bit */ || \
    (defined(__arm__) && __ARM_ARCH >= 7) || defined(__aarch64__) || \
    defined(__riscv)
	1;
#else
	0;
#endif
SYSCTL_INT(__CONCAT(_kern_elf, __ELF_WORD_SIZE), OID_AUTO,
    nxstack, CTLFLAG_RW, &__elfN(nxstack), 0,
    __XSTRING(__CONCAT(ELF, __ELF_WORD_SIZE)) ": enable non-executable stack");

#if __ELF_WORD_SIZE == 32 && (defined(__amd64__) || defined(__i386__))
int i386_read_exec = 0;
SYSCTL_INT(_kern_elf32, OID_AUTO, read_exec, CTLFLAG_RW, &i386_read_exec, 0,
    "enable execution from readable segments");
#endif

static u_long __elfN(pie_base) = ET_DYN_LOAD_ADDR;
static int
sysctl_pie_base(SYSCTL_HANDLER_ARGS)
{
	u_long val;
	int error;

	val = __elfN(pie_base);
	error = sysctl_handle_long(oidp, &val, 0, req);
	if (error != 0 || req->newptr == NULL)
		return (error);
	if ((val & PAGE_MASK) != 0)
		return (EINVAL);
	__elfN(pie_base) = val;
	return (0);
}
SYSCTL_PROC(__CONCAT(_kern_elf, __ELF_WORD_SIZE), OID_AUTO, pie_base,
    CTLTYPE_ULONG | CTLFLAG_MPSAFE | CTLFLAG_RW, NULL, 0,
    sysctl_pie_base, "LU",
    "PIE load base without randomization");

SYSCTL_NODE(__CONCAT(_kern_elf, __ELF_WORD_SIZE), OID_AUTO, aslr,
    CTLFLAG_RW | CTLFLAG_MPSAFE, 0,
    "");
#define	ASLR_NODE_OID	__CONCAT(__CONCAT(_kern_elf, __ELF_WORD_SIZE), _aslr)

static int __elfN(aslr_enabled) = 0;
SYSCTL_INT(ASLR_NODE_OID, OID_AUTO, enable, CTLFLAG_RWTUN,
    &__elfN(aslr_enabled), 0,
    __XSTRING(__CONCAT(ELF, __ELF_WORD_SIZE))
    ": enable address map randomization");

static int __elfN(pie_aslr_enabled) = 0;
SYSCTL_INT(ASLR_NODE_OID, OID_AUTO, pie_enable, CTLFLAG_RWTUN,
    &__elfN(pie_aslr_enabled), 0,
    __XSTRING(__CONCAT(ELF, __ELF_WORD_SIZE))
    ": enable address map randomization for PIE binaries");

static int __elfN(aslr_honor_sbrk) = 1;
SYSCTL_INT(ASLR_NODE_OID, OID_AUTO, honor_sbrk, CTLFLAG_RW,
    &__elfN(aslr_honor_sbrk), 0,
    __XSTRING(__CONCAT(ELF, __ELF_WORD_SIZE)) ": assume sbrk is used");

static int __elfN(aslr_stack_gap) = 3;
SYSCTL_INT(ASLR_NODE_OID, OID_AUTO, stack_gap, CTLFLAG_RW,
    &__elfN(aslr_stack_gap), 0,
    __XSTRING(__CONCAT(ELF, __ELF_WORD_SIZE))
    ": maximum percentage of main stack to waste on a random gap");

static int __elfN(sigfastblock) = 1;
SYSCTL_INT(__CONCAT(_kern_elf, __ELF_WORD_SIZE), OID_AUTO, sigfastblock,
    CTLFLAG_RWTUN, &__elfN(sigfastblock), 0,
    "enable sigfastblock for new processes");

static bool __elfN(allow_wx) = true;
SYSCTL_BOOL(__CONCAT(_kern_elf, __ELF_WORD_SIZE), OID_AUTO, allow_wx,
    CTLFLAG_RWTUN, &__elfN(allow_wx), 0,
    "Allow pages to be mapped simultaneously writable and executable");

static Elf_Brandinfo *elf_brand_list[MAX_BRANDS];

#define	aligned(a, t)	(rounddown2((u_long)(a), sizeof(t)) == (u_long)(a))

static const char FREEBSD_ABI_VENDOR[] = "FreeBSD";

Elf_Brandnote __elfN(freebsd_brandnote) = {
	.hdr.n_namesz	= sizeof(FREEBSD_ABI_VENDOR),
	.hdr.n_descsz	= sizeof(int32_t),
	.hdr.n_type	= NT_FREEBSD_ABI_TAG,
	.vendor		= FREEBSD_ABI_VENDOR,
	.flags		= BN_TRANSLATE_OSREL,
	.trans_osrel	= __elfN(freebsd_trans_osrel)
};

static bool
__elfN(freebsd_trans_osrel)(const Elf_Note *note, int32_t *osrel)
{
	uintptr_t p;

	p = (uintptr_t)(note + 1);
	p += roundup2(note->n_namesz, ELF_NOTE_ROUNDSIZE);
	*osrel = *(const int32_t *)(p);

	return (true);
}

static const char GNU_ABI_VENDOR[] = "GNU";
static int GNU_KFREEBSD_ABI_DESC = 3;

Elf_Brandnote __elfN(kfreebsd_brandnote) = {
	.hdr.n_namesz	= sizeof(GNU_ABI_VENDOR),
	.hdr.n_descsz	= 16,	/* XXX at least 16 */
	.hdr.n_type	= 1,
	.vendor		= GNU_ABI_VENDOR,
	.flags		= BN_TRANSLATE_OSREL,
	.trans_osrel	= kfreebsd_trans_osrel
};

static bool
kfreebsd_trans_osrel(const Elf_Note *note, int32_t *osrel)
{
	const Elf32_Word *desc;
	uintptr_t p;

	p = (uintptr_t)(note + 1);
	p += roundup2(note->n_namesz, ELF_NOTE_ROUNDSIZE);

	desc = (const Elf32_Word *)p;
	if (desc[0] != GNU_KFREEBSD_ABI_DESC)
		return (false);

	/*
	 * Debian GNU/kFreeBSD embed the earliest compatible kernel version
	 * (__FreeBSD_version: <major><two digit minor>Rxx) in the LSB way.
	 */
	*osrel = desc[1] * 100000 + desc[2] * 1000 + desc[3];

	return (true);
}

int
__elfN(insert_brand_entry)(Elf_Brandinfo *entry)
{
	int i;

	for (i = 0; i < MAX_BRANDS; i++) {
		if (elf_brand_list[i] == NULL) {
			elf_brand_list[i] = entry;
			break;
		}
	}
	if (i == MAX_BRANDS) {
		printf("WARNING: %s: could not insert brandinfo entry: %p\n",
			__func__, entry);
		return (-1);
	}
	return (0);
}

int
__elfN(remove_brand_entry)(Elf_Brandinfo *entry)
{
	int i;

	for (i = 0; i < MAX_BRANDS; i++) {
		if (elf_brand_list[i] == entry) {
			elf_brand_list[i] = NULL;
			break;
		}
	}
	if (i == MAX_BRANDS)
		return (-1);
	return (0);
}

int
__elfN(brand_inuse)(Elf_Brandinfo *entry)
{
	struct proc *p;
	int rval = FALSE;

	sx_slock(&allproc_lock);
	FOREACH_PROC_IN_SYSTEM(p) {
		if (p->p_sysent == entry->sysvec) {
			rval = TRUE;
			break;
		}
	}
	sx_sunlock(&allproc_lock);

	return (rval);
}

static Elf_Brandinfo *
__elfN(get_brandinfo)(struct image_params *imgp, const char *interp,
    int32_t *osrel, uint32_t *fctl0)
{
	const Elf_Ehdr *hdr = (const Elf_Ehdr *)imgp->image_header;
	Elf_Brandinfo *bi, *bi_m;
	boolean_t ret, has_fctl0;
	int i, interp_name_len;

	interp_name_len = interp != NULL ? strlen(interp) + 1 : 0;

	/*
	 * We support four types of branding -- (1) the ELF EI_OSABI field
	 * that SCO added to the ELF spec, (2) FreeBSD 3.x's traditional string
	 * branding w/in the ELF header, (3) path of the `interp_path'
	 * field, and (4) the ".note.ABI-tag" ELF section.
	 */

	/* Look for an ".note.ABI-tag" ELF section */
	bi_m = NULL;
	for (i = 0; i < MAX_BRANDS; i++) {
		bi = elf_brand_list[i];
		if (bi == NULL)
			continue;
		if (interp != NULL && (bi->flags & BI_BRAND_ONLY_STATIC) != 0)
			continue;
		if (hdr->e_machine == bi->machine && (bi->flags &
		    (BI_BRAND_NOTE|BI_BRAND_NOTE_MANDATORY)) != 0) {
			has_fctl0 = false;
			*fctl0 = 0;
			*osrel = 0;
			ret = __elfN(check_note)(imgp, bi->brand_note, osrel,
			    &has_fctl0, fctl0);
			/* Give brand a chance to veto check_note's guess */
			if (ret && bi->header_supported) {
				ret = bi->header_supported(imgp, osrel,
				    has_fctl0 ? fctl0 : NULL);
			}
			/*
			 * If note checker claimed the binary, but the
			 * interpreter path in the image does not
			 * match default one for the brand, try to
			 * search for other brands with the same
			 * interpreter.  Either there is better brand
			 * with the right interpreter, or, failing
			 * this, we return first brand which accepted
			 * our note and, optionally, header.
			 */
			if (ret && bi_m == NULL && interp != NULL &&
			    (bi->interp_path == NULL ||
			    (strlen(bi->interp_path) + 1 != interp_name_len ||
			    strncmp(interp, bi->interp_path, interp_name_len)
			    != 0))) {
				bi_m = bi;
				ret = 0;
			}
			if (ret)
				return (bi);
		}
	}
	if (bi_m != NULL)
		return (bi_m);

	/* If the executable has a brand, search for it in the brand list. */
	for (i = 0; i < MAX_BRANDS; i++) {
		bi = elf_brand_list[i];
		if (bi == NULL || (bi->flags & BI_BRAND_NOTE_MANDATORY) != 0 ||
		    (interp != NULL && (bi->flags & BI_BRAND_ONLY_STATIC) != 0))
			continue;
		if (hdr->e_machine == bi->machine &&
		    (hdr->e_ident[EI_OSABI] == bi->brand ||
		    (bi->compat_3_brand != NULL &&
		    strcmp((const char *)&hdr->e_ident[OLD_EI_BRAND],
		    bi->compat_3_brand) == 0))) {
			/* Looks good, but give brand a chance to veto */
			if (bi->header_supported == NULL ||
			    bi->header_supported(imgp, NULL, NULL)) {
				/*
				 * Again, prefer strictly matching
				 * interpreter path.
				 */
				if (interp_name_len == 0 &&
				    bi->interp_path == NULL)
					return (bi);
				if (bi->interp_path != NULL &&
				    strlen(bi->interp_path) + 1 ==
				    interp_name_len && strncmp(interp,
				    bi->interp_path, interp_name_len) == 0)
					return (bi);
				if (bi_m == NULL)
					bi_m = bi;
			}
		}
	}
	if (bi_m != NULL)
		return (bi_m);

	/* No known brand, see if the header is recognized by any brand */
	for (i = 0; i < MAX_BRANDS; i++) {
		bi = elf_brand_list[i];
		if (bi == NULL || bi->flags & BI_BRAND_NOTE_MANDATORY ||
		    bi->header_supported == NULL)
			continue;
		if (hdr->e_machine == bi->machine) {
			ret = bi->header_supported(imgp, NULL, NULL);
			if (ret)
				return (bi);
		}
	}

	/* Lacking a known brand, search for a recognized interpreter. */
	if (interp != NULL) {
		for (i = 0; i < MAX_BRANDS; i++) {
			bi = elf_brand_list[i];
			if (bi == NULL || (bi->flags &
			    (BI_BRAND_NOTE_MANDATORY | BI_BRAND_ONLY_STATIC))
			    != 0)
				continue;
			if (hdr->e_machine == bi->machine &&
			    bi->interp_path != NULL &&
			    /* ELF image p_filesz includes terminating zero */
			    strlen(bi->interp_path) + 1 == interp_name_len &&
			    strncmp(interp, bi->interp_path, interp_name_len)
			    == 0 && (bi->header_supported == NULL ||
			    bi->header_supported(imgp, NULL, NULL)))
				return (bi);
		}
	}

	/* Lacking a recognized interpreter, try the default brand */
	for (i = 0; i < MAX_BRANDS; i++) {
		bi = elf_brand_list[i];
		if (bi == NULL || (bi->flags & BI_BRAND_NOTE_MANDATORY) != 0 ||
		    (interp != NULL && (bi->flags & BI_BRAND_ONLY_STATIC) != 0))
			continue;
		if (hdr->e_machine == bi->machine &&
		    __elfN(fallback_brand) == bi->brand &&
		    (bi->header_supported == NULL ||
		    bi->header_supported(imgp, NULL, NULL)))
			return (bi);
	}
	return (NULL);
}

static bool
__elfN(phdr_in_zero_page)(const Elf_Ehdr *hdr)
{
	return (hdr->e_phoff <= PAGE_SIZE &&
	    (u_int)hdr->e_phentsize * hdr->e_phnum <= PAGE_SIZE - hdr->e_phoff);
}

static int
__elfN(check_header)(const Elf_Ehdr *hdr)
{
	Elf_Brandinfo *bi;
	int i;

	if (!IS_ELF(*hdr) ||
	    hdr->e_ident[EI_CLASS] != ELF_TARG_CLASS ||
	    hdr->e_ident[EI_DATA] != ELF_TARG_DATA ||
	    hdr->e_ident[EI_VERSION] != EV_CURRENT ||
	    hdr->e_phentsize != sizeof(Elf_Phdr) ||
	    hdr->e_version != ELF_TARG_VER)
		return (ENOEXEC);

	/*
	 * Make sure we have at least one brand for this machine.
	 */

	for (i = 0; i < MAX_BRANDS; i++) {
		bi = elf_brand_list[i];
		if (bi != NULL && bi->machine == hdr->e_machine)
			break;
	}
	if (i == MAX_BRANDS)
		return (ENOEXEC);

	return (0);
}

static int
__elfN(map_partial)(vm_map_t map, vm_object_t object, vm_ooffset_t offset,
    vm_offset_t start, vm_offset_t end, vm_prot_t prot)
{
	struct sf_buf *sf;
	int error;
	vm_offset_t off;

	/*
	 * Create the page if it doesn't exist yet. Ignore errors.
	 */
	vm_map_fixed(map, NULL, 0, trunc_page(start), round_page(end) -
	    trunc_page(start), VM_PROT_ALL, VM_PROT_ALL, MAP_CHECK_EXCL);

	/*
	 * Find the page from the underlying object.
	 */
	if (object != NULL) {
		sf = vm_imgact_map_page(object, offset);
		if (sf == NULL)
			return (KERN_FAILURE);
		off = offset - trunc_page(offset);
		error = copyout((caddr_t)sf_buf_kva(sf) + off, (caddr_t)start,
		    end - start);
		vm_imgact_unmap_page(sf);
		if (error != 0)
			return (KERN_FAILURE);
	}

	return (KERN_SUCCESS);
}

static int
__elfN(map_insert)(struct image_params *imgp, vm_map_t map, vm_object_t object,
    vm_ooffset_t offset, vm_offset_t start, vm_offset_t end, vm_prot_t prot,
    int cow)
{
	struct sf_buf *sf;
	vm_offset_t off;
	vm_size_t sz;
	int error, locked, rv;

	if (start != trunc_page(start)) {
		rv = __elfN(map_partial)(map, object, offset, start,
		    round_page(start), prot);
		if (rv != KERN_SUCCESS)
			return (rv);
		offset += round_page(start) - start;
		start = round_page(start);
	}
	if (end != round_page(end)) {
		rv = __elfN(map_partial)(map, object, offset +
		    trunc_page(end) - start, trunc_page(end), end, prot);
		if (rv != KERN_SUCCESS)
			return (rv);
		end = trunc_page(end);
	}
	if (start >= end)
		return (KERN_SUCCESS);
	if ((offset & PAGE_MASK) != 0) {
		/*
		 * The mapping is not page aligned.  This means that we have
		 * to copy the data.
		 */
		rv = vm_map_fixed(map, NULL, 0, start, end - start,
		    prot | VM_PROT_WRITE, VM_PROT_ALL, MAP_CHECK_EXCL);
		if (rv != KERN_SUCCESS)
			return (rv);
		if (object == NULL)
			return (KERN_SUCCESS);
		for (; start < end; start += sz) {
			sf = vm_imgact_map_page(object, offset);
			if (sf == NULL)
				return (KERN_FAILURE);
			off = offset - trunc_page(offset);
			sz = end - start;
			if (sz > PAGE_SIZE - off)
				sz = PAGE_SIZE - off;
			error = copyout((caddr_t)sf_buf_kva(sf) + off,
			    (caddr_t)start, sz);
			vm_imgact_unmap_page(sf);
			if (error != 0)
				return (KERN_FAILURE);
			offset += sz;
		}
	} else {
		vm_object_reference(object);
		rv = vm_map_fixed(map, object, offset, start, end - start,
		    prot, VM_PROT_ALL, cow | MAP_CHECK_EXCL |
		    (object != NULL ? MAP_VN_EXEC : 0));
		if (rv != KERN_SUCCESS) {
			locked = VOP_ISLOCKED(imgp->vp);
			VOP_UNLOCK(imgp->vp);
			vm_object_deallocate(object);
			vn_lock(imgp->vp, locked | LK_RETRY);
			return (rv);
		} else if (object != NULL) {
			MPASS(imgp->vp->v_object == object);
			VOP_SET_TEXT_CHECKED(imgp->vp);
		}
	}
	return (KERN_SUCCESS);
}

static int
__elfN(load_section)(struct image_params *imgp, vm_ooffset_t offset,
    caddr_t vmaddr, size_t memsz, size_t filsz, vm_prot_t prot)
{
	struct sf_buf *sf;
	size_t map_len;
	vm_map_t map;
	vm_object_t object;
	vm_offset_t map_addr;
	int error, rv, cow;
	size_t copy_len;
	vm_ooffset_t file_addr;

	/*
	 * It's necessary to fail if the filsz + offset taken from the
	 * header is greater than the actual file pager object's size.
	 * If we were to allow this, then the vm_map_find() below would
	 * walk right off the end of the file object and into the ether.
	 *
	 * While I'm here, might as well check for something else that
	 * is invalid: filsz cannot be greater than memsz.
	 */
	if ((filsz != 0 && (off_t)filsz + offset > imgp->attr->va_size) ||
	    filsz > memsz) {
		uprintf("elf_load_section: truncated ELF file\n");
		return (ENOEXEC);
	}

	object = imgp->object;
	map = &imgp->proc->p_vmspace->vm_map;
	map_addr = trunc_page((vm_offset_t)vmaddr);
	file_addr = trunc_page(offset);

	/*
	 * We have two choices.  We can either clear the data in the last page
	 * of an oversized mapping, or we can start the anon mapping a page
	 * early and copy the initialized data into that first page.  We
	 * choose the second.
	 */
	if (filsz == 0)
		map_len = 0;
	else if (memsz > filsz)
		map_len = trunc_page(offset + filsz) - file_addr;
	else
		map_len = round_page(offset + filsz) - file_addr;

	if (map_len != 0) {
		/* cow flags: don't dump readonly sections in core */
		cow = MAP_COPY_ON_WRITE | MAP_PREFAULT |
		    (prot & VM_PROT_WRITE ? 0 : MAP_DISABLE_COREDUMP);

		rv = __elfN(map_insert)(imgp, map, object, file_addr,
		    map_addr, map_addr + map_len, prot, cow);
		if (rv != KERN_SUCCESS)
			return (EINVAL);

		/* we can stop now if we've covered it all */
		if (memsz == filsz)
			return (0);
	}

	/*
	 * We have to get the remaining bit of the file into the first part
	 * of the oversized map segment.  This is normally because the .data
	 * segment in the file is extended to provide bss.  It's a neat idea
	 * to try and save a page, but it's a pain in the behind to implement.
	 */
	copy_len = filsz == 0 ? 0 : (offset + filsz) - trunc_page(offset +
	    filsz);
	map_addr = trunc_page((vm_offset_t)vmaddr + filsz);
	map_len = round_page((vm_offset_t)vmaddr + memsz) - map_addr;

	/* This had damn well better be true! */
	if (map_len != 0) {
		rv = __elfN(map_insert)(imgp, map, NULL, 0, map_addr,
		    map_addr + map_len, prot, 0);
		if (rv != KERN_SUCCESS)
			return (EINVAL);
	}

	if (copy_len != 0) {
		sf = vm_imgact_map_page(object, offset + filsz);
		if (sf == NULL)
			return (EIO);

		/* send the page fragment to user space */
		error = copyout((caddr_t)sf_buf_kva(sf), (caddr_t)map_addr,
		    copy_len);
		vm_imgact_unmap_page(sf);
		if (error != 0)
			return (error);
	}

	/*
	 * Remove write access to the page if it was only granted by map_insert
	 * to allow copyout.
	 */
	if ((prot & VM_PROT_WRITE) == 0)
		vm_map_protect(map, trunc_page(map_addr), round_page(map_addr +
		    map_len), prot, 0, VM_MAP_PROTECT_SET_PROT);

	return (0);
}

static int
__elfN(load_sections)(struct image_params *imgp, const Elf_Ehdr *hdr,
    const Elf_Phdr *phdr, u_long rbase, u_long *base_addrp)
{
	vm_prot_t prot;
	u_long base_addr;
	bool first;
	int error, i;

	ASSERT_VOP_LOCKED(imgp->vp, __func__);

	base_addr = 0;
	first = true;

	for (i = 0; i < hdr->e_phnum; i++) {
		if (phdr[i].p_type != PT_LOAD || phdr[i].p_memsz == 0)
			continue;

		/* Loadable segment */
		prot = __elfN(trans_prot)(phdr[i].p_flags);
		error = __elfN(load_section)(imgp, phdr[i].p_offset,
		    (caddr_t)(uintptr_t)phdr[i].p_vaddr + rbase,
		    phdr[i].p_memsz, phdr[i].p_filesz, prot);
		if (error != 0)
			return (error);

		/*
		 * Establish the base address if this is the first segment.
		 */
		if (first) {
  			base_addr = trunc_page(phdr[i].p_vaddr + rbase);
			first = false;
		}
	}

	if (base_addrp != NULL)
		*base_addrp = base_addr;

	return (0);
}

/*
 * Load the file "file" into memory.  It may be either a shared object
 * or an executable.
 *
 * The "addr" reference parameter is in/out.  On entry, it specifies
 * the address where a shared object should be loaded.  If the file is
 * an executable, this value is ignored.  On exit, "addr" specifies
 * where the file was actually loaded.
 *
 * The "entry" reference parameter is out only.  On exit, it specifies
 * the entry point for the loaded file.
 */
static int
__elfN(load_file)(struct proc *p, const char *file, u_long *addr,
	u_long *entry)
{
	struct {
		struct nameidata nd;
		struct vattr attr;
		struct image_params image_params;
	} *tempdata;
	const Elf_Ehdr *hdr = NULL;
	const Elf_Phdr *phdr = NULL;
	struct nameidata *nd;
	struct vattr *attr;
	struct image_params *imgp;
	u_long rbase;
	u_long base_addr = 0;
	int error;

#ifdef CAPABILITY_MODE
	/*
	 * XXXJA: This check can go away once we are sufficiently confident
	 * that the checks in namei() are correct.
	 */
	if (IN_CAPABILITY_MODE(curthread))
		return (ECAPMODE);
#endif

	tempdata = malloc(sizeof(*tempdata), M_TEMP, M_WAITOK | M_ZERO);
	nd = &tempdata->nd;
	attr = &tempdata->attr;
	imgp = &tempdata->image_params;

	/*
	 * Initialize part of the common data
	 */
	imgp->proc = p;
	imgp->attr = attr;

	NDINIT(nd, LOOKUP, ISOPEN | FOLLOW | LOCKSHARED | LOCKLEAF,
	    UIO_SYSSPACE, file, curthread);
	if ((error = namei(nd)) != 0) {
		nd->ni_vp = NULL;
		goto fail;
	}
	NDFREE(nd, NDF_ONLY_PNBUF);
	imgp->vp = nd->ni_vp;

	/*
	 * Check permissions, modes, uid, etc on the file, and "open" it.
	 */
	error = exec_check_permissions(imgp);
	if (error)
		goto fail;

	error = exec_map_first_page(imgp);
	if (error)
		goto fail;

	imgp->object = nd->ni_vp->v_object;

	hdr = (const Elf_Ehdr *)imgp->image_header;
	if ((error = __elfN(check_header)(hdr)) != 0)
		goto fail;
	if (hdr->e_type == ET_DYN)
		rbase = *addr;
	else if (hdr->e_type == ET_EXEC)
		rbase = 0;
	else {
		error = ENOEXEC;
		goto fail;
	}

	/* Only support headers that fit within first page for now      */
	if (!__elfN(phdr_in_zero_page)(hdr)) {
		error = ENOEXEC;
		goto fail;
	}

	phdr = (const Elf_Phdr *)(imgp->image_header + hdr->e_phoff);
	if (!aligned(phdr, Elf_Addr)) {
		error = ENOEXEC;
		goto fail;
	}

	error = __elfN(load_sections)(imgp, hdr, phdr, rbase, &base_addr);
	if (error != 0)
		goto fail;

	*addr = base_addr;
	*entry = (unsigned long)hdr->e_entry + rbase;

fail:
	if (imgp->firstpage)
		exec_unmap_first_page(imgp);

	if (nd->ni_vp) {
		if (imgp->textset)
			VOP_UNSET_TEXT_CHECKED(nd->ni_vp);
		vput(nd->ni_vp);
	}
	free(tempdata, M_TEMP);

	return (error);
}

static u_long
__CONCAT(rnd_, __elfN(base))(vm_map_t map __unused, u_long minv, u_long maxv,
    u_int align)
{
	u_long rbase, res;

	MPASS(vm_map_min(map) <= minv);
	MPASS(maxv <= vm_map_max(map));
	MPASS(minv < maxv);
	MPASS(minv + align < maxv);
	arc4rand(&rbase, sizeof(rbase), 0);
	res = roundup(minv, (u_long)align) + rbase % (maxv - minv);
	res &= ~((u_long)align - 1);
	if (res >= maxv)
		res -= align;
	KASSERT(res >= minv,
	    ("res %#lx < minv %#lx, maxv %#lx rbase %#lx",
	    res, minv, maxv, rbase));
	KASSERT(res < maxv,
	    ("res %#lx > maxv %#lx, minv %#lx rbase %#lx",
	    res, maxv, minv, rbase));
	return (res);
}

static int
__elfN(enforce_limits)(struct image_params *imgp, const Elf_Ehdr *hdr,
    const Elf_Phdr *phdr, u_long et_dyn_addr)
{
	struct vmspace *vmspace;
	const char *err_str;
	u_long text_size, data_size, total_size, text_addr, data_addr;
	u_long seg_size, seg_addr;
	int i;

	err_str = NULL;
	text_size = data_size = total_size = text_addr = data_addr = 0;

	for (i = 0; i < hdr->e_phnum; i++) {
		if (phdr[i].p_type != PT_LOAD || phdr[i].p_memsz == 0)
			continue;

		seg_addr = trunc_page(phdr[i].p_vaddr + et_dyn_addr);
		seg_size = round_page(phdr[i].p_memsz +
		    phdr[i].p_vaddr + et_dyn_addr - seg_addr);

		/*
		 * Make the largest executable segment the official
		 * text segment and all others data.
		 *
		 * Note that obreak() assumes that data_addr + data_size == end
		 * of data load area, and the ELF file format expects segments
		 * to be sorted by address.  If multiple data segments exist,
		 * the last one will be used.
		 */

		if ((phdr[i].p_flags & PF_X) != 0 && text_size < seg_size) {
			text_size = seg_size;
			text_addr = seg_addr;
		} else {
			data_size = seg_size;
			data_addr = seg_addr;
		}
		total_size += seg_size;
	}

	if (data_addr == 0 && data_size == 0) {
		data_addr = text_addr;
		data_size = text_size;
	}

	/*
	 * Check limits.  It should be safe to check the
	 * limits after loading the segments since we do
	 * not actually fault in all the segments pages.
	 */
	PROC_LOCK(imgp->proc);
	if (data_size > lim_cur_proc(imgp->proc, RLIMIT_DATA))
		err_str = "Data segment size exceeds process limit";
	else if (text_size > maxtsiz)
		err_str = "Text segment size exceeds system limit";
	else if (total_size > lim_cur_proc(imgp->proc, RLIMIT_VMEM))
		err_str = "Total segment size exceeds process limit";
	else if (racct_set(imgp->proc, RACCT_DATA, data_size) != 0)
		err_str = "Data segment size exceeds resource limit";
	else if (racct_set(imgp->proc, RACCT_VMEM, total_size) != 0)
		err_str = "Total segment size exceeds resource limit";
	PROC_UNLOCK(imgp->proc);
	if (err_str != NULL) {
		uprintf("%s\n", err_str);
		return (ENOMEM);
	}

	vmspace = imgp->proc->p_vmspace;
	vmspace->vm_tsize = text_size >> PAGE_SHIFT;
	vmspace->vm_taddr = (caddr_t)(uintptr_t)text_addr;
	vmspace->vm_dsize = data_size >> PAGE_SHIFT;
	vmspace->vm_daddr = (caddr_t)(uintptr_t)data_addr;

	return (0);
}

static int
__elfN(get_interp)(struct image_params *imgp, const Elf_Phdr *phdr,
    char **interpp, bool *free_interpp)
{
	struct thread *td;
	char *interp;
	int error, interp_name_len;

	KASSERT(phdr->p_type == PT_INTERP,
	    ("%s: p_type %u != PT_INTERP", __func__, phdr->p_type));
	ASSERT_VOP_LOCKED(imgp->vp, __func__);

	td = curthread;

	/* Path to interpreter */
	if (phdr->p_filesz < 2 || phdr->p_filesz > MAXPATHLEN) {
		uprintf("Invalid PT_INTERP\n");
		return (ENOEXEC);
	}

	interp_name_len = phdr->p_filesz;
	if (phdr->p_offset > PAGE_SIZE ||
	    interp_name_len > PAGE_SIZE - phdr->p_offset) {
		/*
		 * The vnode lock might be needed by the pagedaemon to
		 * clean pages owned by the vnode.  Do not allow sleep
		 * waiting for memory with the vnode locked, instead
		 * try non-sleepable allocation first, and if it
		 * fails, go to the slow path were we drop the lock
		 * and do M_WAITOK.  A text reference prevents
		 * modifications to the vnode content.
		 */
		interp = malloc(interp_name_len + 1, M_TEMP, M_NOWAIT);
		if (interp == NULL) {
			VOP_UNLOCK(imgp->vp);
			interp = malloc(interp_name_len + 1, M_TEMP, M_WAITOK);
			vn_lock(imgp->vp, LK_SHARED | LK_RETRY);
		}

		error = vn_rdwr(UIO_READ, imgp->vp, interp,
		    interp_name_len, phdr->p_offset,
		    UIO_SYSSPACE, IO_NODELOCKED, td->td_ucred,
		    NOCRED, NULL, td);
		if (error != 0) {
			free(interp, M_TEMP);
			uprintf("i/o error PT_INTERP %d\n", error);
			return (error);
		}
		interp[interp_name_len] = '\0';

		*interpp = interp;
		*free_interpp = true;
		return (0);
	}

	interp = __DECONST(char *, imgp->image_header) + phdr->p_offset;
	if (interp[interp_name_len - 1] != '\0') {
		uprintf("Invalid PT_INTERP\n");
		return (ENOEXEC);
	}

	*interpp = interp;
	*free_interpp = false;
	return (0);
}

static int
__elfN(load_interp)(struct image_params *imgp, const Elf_Brandinfo *brand_info,
    const char *interp, u_long *addr, u_long *entry)
{
	char *path;
	int error;

	if (brand_info->emul_path != NULL &&
	    brand_info->emul_path[0] != '\0') {
		path = malloc(MAXPATHLEN, M_TEMP, M_WAITOK);
		snprintf(path, MAXPATHLEN, "%s%s",
		    brand_info->emul_path, interp);
		error = __elfN(load_file)(imgp->proc, path, addr, entry);
		free(path, M_TEMP);
		if (error == 0)
			return (0);
	}

	if (brand_info->interp_newpath != NULL &&
	    (brand_info->interp_path == NULL ||
	    strcmp(interp, brand_info->interp_path) == 0)) {
		error = __elfN(load_file)(imgp->proc,
		    brand_info->interp_newpath, addr, entry);
		if (error == 0)
			return (0);
	}

	error = __elfN(load_file)(imgp->proc, interp, addr, entry);
	if (error == 0)
		return (0);

	uprintf("ELF interpreter %s not found, error %d\n", interp, error);
	return (error);
}

/*
 * Impossible et_dyn_addr initial value indicating that the real base
 * must be calculated later with some randomization applied.
 */
#define	ET_DYN_ADDR_RAND	1

static int
__CONCAT(exec_, __elfN(imgact))(struct image_params *imgp)
{
	struct thread *td;
	const Elf_Ehdr *hdr;
	const Elf_Phdr *phdr;
	Elf_Auxargs *elf_auxargs;
	struct vmspace *vmspace;
	vm_map_t map;
	char *interp;
	Elf_Brandinfo *brand_info;
	struct sysentvec *sv;
	u_long addr, baddr, et_dyn_addr, entry, proghdr;
	u_long maxalign, mapsz, maxv, maxv1;
	uint32_t fctl0;
	int32_t osrel;
	bool free_interp;
	int error, i, n;

	hdr = (const Elf_Ehdr *)imgp->image_header;

	/*
	 * Do we have a valid ELF header ?
	 *
	 * Only allow ET_EXEC & ET_DYN here, reject ET_DYN later
	 * if particular brand doesn't support it.
	 */
	if (__elfN(check_header)(hdr) != 0 ||
	    (hdr->e_type != ET_EXEC && hdr->e_type != ET_DYN))
		return (-1);

	/*
	 * From here on down, we return an errno, not -1, as we've
	 * detected an ELF file.
	 */

	if (!__elfN(phdr_in_zero_page)(hdr)) {
		uprintf("Program headers not in the first page\n");
		return (ENOEXEC);
	}
	phdr = (const Elf_Phdr *)(imgp->image_header + hdr->e_phoff); 
	if (!aligned(phdr, Elf_Addr)) {
		uprintf("Unaligned program headers\n");
		return (ENOEXEC);
	}

	n = error = 0;
	baddr = 0;
	osrel = 0;
	fctl0 = 0;
	entry = proghdr = 0;
	interp = NULL;
	free_interp = false;
	td = curthread;
	maxalign = PAGE_SIZE;
	mapsz = 0;

	for (i = 0; i < hdr->e_phnum; i++) {
		switch (phdr[i].p_type) {
		case PT_LOAD:
			if (n == 0)
				baddr = phdr[i].p_vaddr;
			if (phdr[i].p_align > maxalign)
				maxalign = phdr[i].p_align;
			mapsz += phdr[i].p_memsz;
			n++;

			/*
			 * If this segment contains the program headers,
			 * remember their virtual address for the AT_PHDR
			 * aux entry. Static binaries don't usually include
			 * a PT_PHDR entry.
			 */
			if (phdr[i].p_offset == 0 &&
			    hdr->e_phoff + hdr->e_phnum * hdr->e_phentsize
				<= phdr[i].p_filesz)
				proghdr = phdr[i].p_vaddr + hdr->e_phoff;
			break;
		case PT_INTERP:
			/* Path to interpreter */
			if (interp != NULL) {
				uprintf("Multiple PT_INTERP headers\n");
				error = ENOEXEC;
				goto ret;
			}
			error = __elfN(get_interp)(imgp, &phdr[i], &interp,
			    &free_interp);
			if (error != 0)
				goto ret;
			break;
		case PT_GNU_STACK:
			if (__elfN(nxstack))
				imgp->stack_prot =
				    __elfN(trans_prot)(phdr[i].p_flags);
			imgp->stack_sz = phdr[i].p_memsz;
			break;
		case PT_PHDR: 	/* Program header table info */
			proghdr = phdr[i].p_vaddr;
			break;
		}
	}

	brand_info = __elfN(get_brandinfo)(imgp, interp, &osrel, &fctl0);
	if (brand_info == NULL) {
		uprintf("ELF binary type \"%u\" not known.\n",
		    hdr->e_ident[EI_OSABI]);
		error = ENOEXEC;
		goto ret;
	}
	sv = brand_info->sysvec;
	et_dyn_addr = 0;
	if (hdr->e_type == ET_DYN) {
		if ((brand_info->flags & BI_CAN_EXEC_DYN) == 0) {
			uprintf("Cannot execute shared object\n");
			error = ENOEXEC;
			goto ret;
		}
		/*
		 * Honour the base load address from the dso if it is
		 * non-zero for some reason.
		 */
		if (baddr == 0) {
			if ((sv->sv_flags & SV_ASLR) == 0 ||
			    (fctl0 & NT_FREEBSD_FCTL_ASLR_DISABLE) != 0)
				et_dyn_addr = __elfN(pie_base);
			else if ((__elfN(pie_aslr_enabled) &&
			    (imgp->proc->p_flag2 & P2_ASLR_DISABLE) == 0) ||
			    (imgp->proc->p_flag2 & P2_ASLR_ENABLE) != 0)
				et_dyn_addr = ET_DYN_ADDR_RAND;
			else
				et_dyn_addr = __elfN(pie_base);
		}
	}

	/*
	 * Avoid a possible deadlock if the current address space is destroyed
	 * and that address space maps the locked vnode.  In the common case,
	 * the locked vnode's v_usecount is decremented but remains greater
	 * than zero.  Consequently, the vnode lock is not needed by vrele().
	 * However, in cases where the vnode lock is external, such as nullfs,
	 * v_usecount may become zero.
	 *
	 * The VV_TEXT flag prevents modifications to the executable while
	 * the vnode is unlocked.
	 */
	VOP_UNLOCK(imgp->vp);

	/*
	 * Decide whether to enable randomization of user mappings.
	 * First, reset user preferences for the setid binaries.
	 * Then, account for the support of the randomization by the
	 * ABI, by user preferences, and make special treatment for
	 * PIE binaries.
	 */
	if (imgp->credential_setid) {
		PROC_LOCK(imgp->proc);
		imgp->proc->p_flag2 &= ~(P2_ASLR_ENABLE | P2_ASLR_DISABLE);
		PROC_UNLOCK(imgp->proc);
	}
	if ((sv->sv_flags & SV_ASLR) == 0 ||
	    (imgp->proc->p_flag2 & P2_ASLR_DISABLE) != 0 ||
	    (fctl0 & NT_FREEBSD_FCTL_ASLR_DISABLE) != 0) {
		KASSERT(et_dyn_addr != ET_DYN_ADDR_RAND,
		    ("et_dyn_addr == RAND and !ASLR"));
	} else if ((imgp->proc->p_flag2 & P2_ASLR_ENABLE) != 0 ||
	    (__elfN(aslr_enabled) && hdr->e_type == ET_EXEC) ||
	    et_dyn_addr == ET_DYN_ADDR_RAND) {
		imgp->map_flags |= MAP_ASLR;
		/*
		 * If user does not care about sbrk, utilize the bss
		 * grow region for mappings as well.  We can select
		 * the base for the image anywere and still not suffer
		 * from the fragmentation.
		 */
		if (!__elfN(aslr_honor_sbrk) ||
		    (imgp->proc->p_flag2 & P2_ASLR_IGNSTART) != 0)
			imgp->map_flags |= MAP_ASLR_IGNSTART;
	}

	if (!__elfN(allow_wx) && (fctl0 & NT_FREEBSD_FCTL_WXNEEDED) == 0)
		imgp->map_flags |= MAP_WXORX;

	error = exec_new_vmspace(imgp, sv);
	vmspace = imgp->proc->p_vmspace;
	map = &vmspace->vm_map;

	imgp->proc->p_sysent = sv;

	maxv = vm_map_max(map) - lim_max(td, RLIMIT_STACK);
	if (et_dyn_addr == ET_DYN_ADDR_RAND) {
		KASSERT((map->flags & MAP_ASLR) != 0,
		    ("ET_DYN_ADDR_RAND but !MAP_ASLR"));
		et_dyn_addr = __CONCAT(rnd_, __elfN(base))(map,
		    vm_map_min(map) + mapsz + lim_max(td, RLIMIT_DATA),
		    /* reserve half of the address space to interpreter */
		    maxv / 2, 1UL << flsl(maxalign));
	}

	vn_lock(imgp->vp, LK_SHARED | LK_RETRY);
	if (error != 0)
		goto ret;

	error = __elfN(load_sections)(imgp, hdr, phdr, et_dyn_addr, NULL);
	if (error != 0)
		goto ret;

	error = __elfN(enforce_limits)(imgp, hdr, phdr, et_dyn_addr);
	if (error != 0)
		goto ret;

	entry = (u_long)hdr->e_entry + et_dyn_addr;

	/*
	 * We load the dynamic linker where a userland call
	 * to mmap(0, ...) would put it.  The rationale behind this
	 * calculation is that it leaves room for the heap to grow to
	 * its maximum allowed size.
	 */
	addr = round_page((vm_offset_t)vmspace->vm_daddr + lim_max(td,
	    RLIMIT_DATA));
	if ((map->flags & MAP_ASLR) != 0) {
		maxv1 = maxv / 2 + addr / 2;
		MPASS(maxv1 >= addr);	/* No overflow */
		map->anon_loc = __CONCAT(rnd_, __elfN(base))(map, addr, maxv1,
		    (MAXPAGESIZES > 1 && pagesizes[1] != 0) ?
		    pagesizes[1] : pagesizes[0]);
	} else {
		map->anon_loc = addr;
	}

	imgp->entry_addr = entry;

	if (interp != NULL) {
		VOP_UNLOCK(imgp->vp);
		if ((map->flags & MAP_ASLR) != 0) {
			/* Assume that interpreter fits into 1/4 of AS */
			maxv1 = maxv / 2 + addr / 2;
			MPASS(maxv1 >= addr);	/* No overflow */
			addr = __CONCAT(rnd_, __elfN(base))(map, addr,
			    maxv1, PAGE_SIZE);
		}
		error = __elfN(load_interp)(imgp, brand_info, interp, &addr,
		    &imgp->entry_addr);
		vn_lock(imgp->vp, LK_SHARED | LK_RETRY);
		if (error != 0)
			goto ret;
	} else
		addr = et_dyn_addr;

	/*
	 * Construct auxargs table (used by the copyout_auxargs routine)
	 */
	elf_auxargs = malloc(sizeof(Elf_Auxargs), M_TEMP, M_NOWAIT);
	if (elf_auxargs == NULL) {
		VOP_UNLOCK(imgp->vp);
		elf_auxargs = malloc(sizeof(Elf_Auxargs), M_TEMP, M_WAITOK);
		vn_lock(imgp->vp, LK_SHARED | LK_RETRY);
	}
	elf_auxargs->execfd = -1;
	elf_auxargs->phdr = proghdr + et_dyn_addr;
	elf_auxargs->phent = hdr->e_phentsize;
	elf_auxargs->phnum = hdr->e_phnum;
	elf_auxargs->pagesz = PAGE_SIZE;
	elf_auxargs->base = addr;
	elf_auxargs->flags = 0;
	elf_auxargs->entry = entry;
	elf_auxargs->hdr_eflags = hdr->e_flags;

	imgp->auxargs = elf_auxargs;
	imgp->interpreted = 0;
	imgp->reloc_base = addr;
	imgp->proc->p_osrel = osrel;
	imgp->proc->p_fctl0 = fctl0;
	imgp->proc->p_elf_machine = hdr->e_machine;
	imgp->proc->p_elf_flags = hdr->e_flags;

ret:
	if (free_interp)
		free(interp, M_TEMP);
	return (error);
}

#define	suword __CONCAT(suword, __ELF_WORD_SIZE)

int
__elfN(freebsd_copyout_auxargs)(struct image_params *imgp, uintptr_t base)
{
	Elf_Auxargs *args = (Elf_Auxargs *)imgp->auxargs;
	Elf_Auxinfo *argarray, *pos;
	int error;

	argarray = pos = malloc(AT_COUNT * sizeof(*pos), M_TEMP,
	    M_WAITOK | M_ZERO);

	if (args->execfd != -1)
		AUXARGS_ENTRY(pos, AT_EXECFD, args->execfd);
	AUXARGS_ENTRY(pos, AT_PHDR, args->phdr);
	AUXARGS_ENTRY(pos, AT_PHENT, args->phent);
	AUXARGS_ENTRY(pos, AT_PHNUM, args->phnum);
	AUXARGS_ENTRY(pos, AT_PAGESZ, args->pagesz);
	AUXARGS_ENTRY(pos, AT_FLAGS, args->flags);
	AUXARGS_ENTRY(pos, AT_ENTRY, args->entry);
	AUXARGS_ENTRY(pos, AT_BASE, args->base);
	AUXARGS_ENTRY(pos, AT_EHDRFLAGS, args->hdr_eflags);
	if (imgp->execpathp != 0)
		AUXARGS_ENTRY_PTR(pos, AT_EXECPATH, imgp->execpathp);
	AUXARGS_ENTRY(pos, AT_OSRELDATE,
	    imgp->proc->p_ucred->cr_prison->pr_osreldate);
	if (imgp->canary != 0) {
		AUXARGS_ENTRY_PTR(pos, AT_CANARY, imgp->canary);
		AUXARGS_ENTRY(pos, AT_CANARYLEN, imgp->canarylen);
	}
	AUXARGS_ENTRY(pos, AT_NCPUS, mp_ncpus);
	if (imgp->pagesizes != 0) {
		AUXARGS_ENTRY_PTR(pos, AT_PAGESIZES, imgp->pagesizes);
		AUXARGS_ENTRY(pos, AT_PAGESIZESLEN, imgp->pagesizeslen);
	}
	if (imgp->sysent->sv_timekeep_base != 0) {
		AUXARGS_ENTRY(pos, AT_TIMEKEEP,
		    imgp->sysent->sv_timekeep_base);
	}
	AUXARGS_ENTRY(pos, AT_STACKPROT, imgp->sysent->sv_shared_page_obj
	    != NULL && imgp->stack_prot != 0 ? imgp->stack_prot :
	    imgp->sysent->sv_stackprot);
	if (imgp->sysent->sv_hwcap != NULL)
		AUXARGS_ENTRY(pos, AT_HWCAP, *imgp->sysent->sv_hwcap);
	if (imgp->sysent->sv_hwcap2 != NULL)
		AUXARGS_ENTRY(pos, AT_HWCAP2, *imgp->sysent->sv_hwcap2);
	AUXARGS_ENTRY(pos, AT_BSDFLAGS, __elfN(sigfastblock) ?
	    ELF_BSDF_SIGFASTBLK : 0);
	AUXARGS_ENTRY(pos, AT_ARGC, imgp->args->argc);
	AUXARGS_ENTRY_PTR(pos, AT_ARGV, imgp->argv);
	AUXARGS_ENTRY(pos, AT_ENVC, imgp->args->envc);
	AUXARGS_ENTRY_PTR(pos, AT_ENVV, imgp->envv);
	AUXARGS_ENTRY_PTR(pos, AT_PS_STRINGS, imgp->ps_strings);
	if (imgp->sysent->sv_fxrng_gen_base != 0)
		AUXARGS_ENTRY(pos, AT_FXRNG, imgp->sysent->sv_fxrng_gen_base);
	AUXARGS_ENTRY(pos, AT_NULL, 0);

	free(imgp->auxargs, M_TEMP);
	imgp->auxargs = NULL;
	KASSERT(pos - argarray <= AT_COUNT, ("Too many auxargs"));

	error = copyout(argarray, (void *)base, sizeof(*argarray) * AT_COUNT);
	free(argarray, M_TEMP);
	return (error);
}

int
__elfN(freebsd_fixup)(uintptr_t *stack_base, struct image_params *imgp)
{
	Elf_Addr *base;

	base = (Elf_Addr *)*stack_base;
	base--;
	if (suword(base, imgp->args->argc) == -1)
		return (EFAULT);
	*stack_base = (uintptr_t)base;
	return (0);
}

/*
 * Code for generating ELF core dumps.
 */

typedef void (*segment_callback)(vm_map_entry_t, void *);

/* Closure for cb_put_phdr(). */
struct phdr_closure {
	Elf_Phdr *phdr;		/* Program header to fill in */
	Elf_Off offset;		/* Offset of segment in core file */
};

/* Closure for cb_size_segment(). */
struct sseg_closure {
	int count;		/* Count of writable segments. */
	size_t size;		/* Total size of all writable segments. */
};

typedef void (*outfunc_t)(void *, struct sbuf *, size_t *);

struct note_info {
	int		type;		/* Note type. */
	outfunc_t 	outfunc; 	/* Output function. */
	void		*outarg;	/* Argument for the output function. */
	size_t		outsize;	/* Output size. */
	TAILQ_ENTRY(note_info) link;	/* Link to the next note info. */
};

TAILQ_HEAD(note_info_list, note_info);

/* Coredump output parameters. */
struct coredump_params {
	off_t		offset;
	struct ucred	*active_cred;
	struct ucred	*file_cred;
	struct thread	*td;
	struct vnode	*vp;
	struct compressor *comp;
};

extern int compress_user_cores;
extern int compress_user_cores_level;

static void cb_put_phdr(vm_map_entry_t, void *);
static void cb_size_segment(vm_map_entry_t, void *);
static int core_write(struct coredump_params *, const void *, size_t, off_t,
    enum uio_seg, size_t *);
static void each_dumpable_segment(struct thread *, segment_callback, void *);
static int __elfN(corehdr)(struct coredump_params *, int, void *, size_t,
    struct note_info_list *, size_t);
static void __elfN(prepare_notes)(struct thread *, struct note_info_list *,
    size_t *);
static void __elfN(puthdr)(struct thread *, void *, size_t, int, size_t);
static void __elfN(putnote)(struct note_info *, struct sbuf *);
static size_t register_note(struct note_info_list *, int, outfunc_t, void *);
static int sbuf_drain_core_output(void *, const char *, int);

static void __elfN(note_fpregset)(void *, struct sbuf *, size_t *);
static void __elfN(note_prpsinfo)(void *, struct sbuf *, size_t *);
static void __elfN(note_prstatus)(void *, struct sbuf *, size_t *);
static void __elfN(note_threadmd)(void *, struct sbuf *, size_t *);
static void __elfN(note_thrmisc)(void *, struct sbuf *, size_t *);
static void __elfN(note_ptlwpinfo)(void *, struct sbuf *, size_t *);
static void __elfN(note_procstat_auxv)(void *, struct sbuf *, size_t *);
static void __elfN(note_procstat_proc)(void *, struct sbuf *, size_t *);
static void __elfN(note_procstat_psstrings)(void *, struct sbuf *, size_t *);
static void note_procstat_files(void *, struct sbuf *, size_t *);
static void note_procstat_groups(void *, struct sbuf *, size_t *);
static void note_procstat_osrel(void *, struct sbuf *, size_t *);
static void note_procstat_rlimit(void *, struct sbuf *, size_t *);
static void note_procstat_umask(void *, struct sbuf *, size_t *);
static void note_procstat_vmmap(void *, struct sbuf *, size_t *);

/*
 * Write out a core segment to the compression stream.
 */
static int
compress_chunk(struct coredump_params *p, char *base, char *buf, u_int len)
{
	u_int chunk_len;
	int error;

	while (len > 0) {
		chunk_len = MIN(len, CORE_BUF_SIZE);

		/*
		 * We can get EFAULT error here.
		 * In that case zero out the current chunk of the segment.
		 */
		error = copyin(base, buf, chunk_len);
		if (error != 0)
			bzero(buf, chunk_len);
		error = compressor_write(p->comp, buf, chunk_len);
		if (error != 0)
			break;
		base += chunk_len;
		len -= chunk_len;
	}
	return (error);
}

static int
core_compressed_write(void *base, size_t len, off_t offset, void *arg)
{

	return (core_write((struct coredump_params *)arg, base, len, offset,
	    UIO_SYSSPACE, NULL));
}

static int
core_write(struct coredump_params *p, const void *base, size_t len,
    off_t offset, enum uio_seg seg, size_t *resid)
{

	return (vn_rdwr_inchunks(UIO_WRITE, p->vp, __DECONST(void *, base),
	    len, offset, seg, IO_UNIT | IO_DIRECT | IO_RANGELOCKED,
	    p->active_cred, p->file_cred, resid, p->td));
}

static int
core_output(char *base, size_t len, off_t offset, struct coredump_params *p,
    void *tmpbuf)
{
	vm_map_t map;
	struct mount *mp;
	size_t resid, runlen;
	int error;
	bool success;

	KASSERT((uintptr_t)base % PAGE_SIZE == 0,
	    ("%s: user address %p is not page-aligned", __func__, base));

	if (p->comp != NULL)
		return (compress_chunk(p, base, tmpbuf, len));

	map = &p->td->td_proc->p_vmspace->vm_map;
	for (; len > 0; base += runlen, offset += runlen, len -= runlen) {
		/*
		 * Attempt to page in all virtual pages in the range.  If a
		 * virtual page is not backed by the pager, it is represented as
		 * a hole in the file.  This can occur with zero-filled
		 * anonymous memory or truncated files, for example.
		 */
		for (runlen = 0; runlen < len; runlen += PAGE_SIZE) {
			error = vm_fault(map, (uintptr_t)base + runlen,
			    VM_PROT_READ, VM_FAULT_NOFILL, NULL);
			if (runlen == 0)
				success = error == KERN_SUCCESS;
			else if ((error == KERN_SUCCESS) != success)
				break;
		}

		if (success) {
			error = core_write(p, base, runlen, offset,
			    UIO_USERSPACE, &resid);
			if (error != 0) {
				if (error != EFAULT)
					break;

				/*
				 * EFAULT may be returned if the user mapping
				 * could not be accessed, e.g., because a mapped
				 * file has been truncated.  Skip the page if no
				 * progress was made, to protect against a
				 * hypothetical scenario where vm_fault() was
				 * successful but core_write() returns EFAULT
				 * anyway.
				 */
				runlen -= resid;
				if (runlen == 0) {
					success = false;
					runlen = PAGE_SIZE;
				}
			}
		}
		if (!success) {
			error = vn_start_write(p->vp, &mp, V_WAIT);
			if (error != 0)
				break;
			vn_lock(p->vp, LK_EXCLUSIVE | LK_RETRY);
			error = vn_truncate_locked(p->vp, offset + runlen,
			    false, p->td->td_ucred);
			VOP_UNLOCK(p->vp);
			vn_finished_write(mp);
			if (error != 0)
				break;
		}
	}
	return (error);
}

/*
 * Drain into a core file.
 */
static int
sbuf_drain_core_output(void *arg, const char *data, int len)
{
	struct coredump_params *p;
	int error, locked;

	p = (struct coredump_params *)arg;

	/*
	 * Some kern_proc out routines that print to this sbuf may
	 * call us with the process lock held. Draining with the
	 * non-sleepable lock held is unsafe. The lock is needed for
	 * those routines when dumping a live process. In our case we
	 * can safely release the lock before draining and acquire
	 * again after.
	 */
	locked = PROC_LOCKED(p->td->td_proc);
	if (locked)
		PROC_UNLOCK(p->td->td_proc);
	if (p->comp != NULL)
		error = compressor_write(p->comp, __DECONST(char *, data), len);
	else
		error = core_write(p, __DECONST(void *, data), len, p->offset,
		    UIO_SYSSPACE, NULL);
	if (locked)
		PROC_LOCK(p->td->td_proc);
	if (error != 0)
		return (-error);
	p->offset += len;
	return (len);
}

int
__elfN(coredump)(struct thread *td, struct vnode *vp, off_t limit, int flags)
{
	struct ucred *cred = td->td_ucred;
	int error = 0;
	struct sseg_closure seginfo;
	struct note_info_list notelst;
	struct coredump_params params;
	struct note_info *ninfo;
	void *hdr, *tmpbuf;
	size_t hdrsize, notesz, coresize;

	hdr = NULL;
	tmpbuf = NULL;
	TAILQ_INIT(&notelst);

	/* Size the program segments. */
	seginfo.count = 0;
	seginfo.size = 0;
	each_dumpable_segment(td, cb_size_segment, &seginfo);

	/*
	 * Collect info about the core file header area.
	 */
	hdrsize = sizeof(Elf_Ehdr) + sizeof(Elf_Phdr) * (1 + seginfo.count);
	if (seginfo.count + 1 >= PN_XNUM)
		hdrsize += sizeof(Elf_Shdr);
	__elfN(prepare_notes)(td, &notelst, &notesz);
	coresize = round_page(hdrsize + notesz) + seginfo.size;

	/* Set up core dump parameters. */
	params.offset = 0;
	params.active_cred = cred;
	params.file_cred = NOCRED;
	params.td = td;
	params.vp = vp;
	params.comp = NULL;

#ifdef RACCT
	if (racct_enable) {
		PROC_LOCK(td->td_proc);
		error = racct_add(td->td_proc, RACCT_CORE, coresize);
		PROC_UNLOCK(td->td_proc);
		if (error != 0) {
			error = EFAULT;
			goto done;
		}
	}
#endif
	if (coresize >= limit) {
		error = EFAULT;
		goto done;
	}

	/* Create a compression stream if necessary. */
	if (compress_user_cores != 0) {
		params.comp = compressor_init(core_compressed_write,
		    compress_user_cores, CORE_BUF_SIZE,
		    compress_user_cores_level, &params);
		if (params.comp == NULL) {
			error = EFAULT;
			goto done;
		}
		tmpbuf = malloc(CORE_BUF_SIZE, M_TEMP, M_WAITOK | M_ZERO);
        }

	/*
	 * Allocate memory for building the header, fill it up,
	 * and write it out following the notes.
	 */
	hdr = malloc(hdrsize, M_TEMP, M_WAITOK);
	error = __elfN(corehdr)(&params, seginfo.count, hdr, hdrsize, &notelst,
	    notesz);

	/* Write the contents of all of the writable segments. */
	if (error == 0) {
		Elf_Phdr *php;
		off_t offset;
		int i;

		php = (Elf_Phdr *)((char *)hdr + sizeof(Elf_Ehdr)) + 1;
		offset = round_page(hdrsize + notesz);
		for (i = 0; i < seginfo.count; i++) {
			error = core_output((char *)(uintptr_t)php->p_vaddr,
			    php->p_filesz, offset, &params, tmpbuf);
			if (error != 0)
				break;
			offset += php->p_filesz;
			php++;
		}
		if (error == 0 && params.comp != NULL)
			error = compressor_flush(params.comp);
	}
	if (error) {
		log(LOG_WARNING,
		    "Failed to write core file for process %s (error %d)\n",
		    curproc->p_comm, error);
	}

done:
	free(tmpbuf, M_TEMP);
	if (params.comp != NULL)
		compressor_fini(params.comp);
	while ((ninfo = TAILQ_FIRST(&notelst)) != NULL) {
		TAILQ_REMOVE(&notelst, ninfo, link);
		free(ninfo, M_TEMP);
	}
	if (hdr != NULL)
		free(hdr, M_TEMP);

	return (error);
}

/*
 * A callback for each_dumpable_segment() to write out the segment's
 * program header entry.
 */
static void
cb_put_phdr(vm_map_entry_t entry, void *closure)
{
	struct phdr_closure *phc = (struct phdr_closure *)closure;
	Elf_Phdr *phdr = phc->phdr;

	phc->offset = round_page(phc->offset);

	phdr->p_type = PT_LOAD;
	phdr->p_offset = phc->offset;
	phdr->p_vaddr = entry->start;
	phdr->p_paddr = 0;
	phdr->p_filesz = phdr->p_memsz = entry->end - entry->start;
	phdr->p_align = PAGE_SIZE;
	phdr->p_flags = __elfN(untrans_prot)(entry->protection);

	phc->offset += phdr->p_filesz;
	phc->phdr++;
}

/*
 * A callback for each_dumpable_segment() to gather information about
 * the number of segments and their total size.
 */
static void
cb_size_segment(vm_map_entry_t entry, void *closure)
{
	struct sseg_closure *ssc = (struct sseg_closure *)closure;

	ssc->count++;
	ssc->size += entry->end - entry->start;
}

/*
 * For each writable segment in the process's memory map, call the given
 * function with a pointer to the map entry and some arbitrary
 * caller-supplied data.
 */
static void
each_dumpable_segment(struct thread *td, segment_callback func, void *closure)
{
	struct proc *p = td->td_proc;
	vm_map_t map = &p->p_vmspace->vm_map;
	vm_map_entry_t entry;
	vm_object_t backing_object, object;
	bool ignore_entry;

	vm_map_lock_read(map);
	VM_MAP_ENTRY_FOREACH(entry, map) {
		/*
		 * Don't dump inaccessible mappings, deal with legacy
		 * coredump mode.
		 *
		 * Note that read-only segments related to the elf binary
		 * are marked MAP_ENTRY_NOCOREDUMP now so we no longer
		 * need to arbitrarily ignore such segments.
		 */
		if (elf_legacy_coredump) {
			if ((entry->protection & VM_PROT_RW) != VM_PROT_RW)
				continue;
		} else {
			if ((entry->protection & VM_PROT_ALL) == 0)
				continue;
		}

		/*
		 * Dont include memory segment in the coredump if
		 * MAP_NOCORE is set in mmap(2) or MADV_NOCORE in
		 * madvise(2).  Do not dump submaps (i.e. parts of the
		 * kernel map).
		 */
		if (entry->eflags & (MAP_ENTRY_NOCOREDUMP|MAP_ENTRY_IS_SUB_MAP))
			continue;

		if ((object = entry->object.vm_object) == NULL)
			continue;

		/* Ignore memory-mapped devices and such things. */
		VM_OBJECT_RLOCK(object);
		while ((backing_object = object->backing_object) != NULL) {
			VM_OBJECT_RLOCK(backing_object);
			VM_OBJECT_RUNLOCK(object);
			object = backing_object;
		}
		ignore_entry = (object->flags & OBJ_FICTITIOUS) != 0;
		VM_OBJECT_RUNLOCK(object);
		if (ignore_entry)
			continue;

		(*func)(entry, closure);
	}
	vm_map_unlock_read(map);
}

/*
 * Write the core file header to the file, including padding up to
 * the page boundary.
 */
static int
__elfN(corehdr)(struct coredump_params *p, int numsegs, void *hdr,
    size_t hdrsize, struct note_info_list *notelst, size_t notesz)
{
	struct note_info *ninfo;
	struct sbuf *sb;
	int error;

	/* Fill in the header. */
	bzero(hdr, hdrsize);
	__elfN(puthdr)(p->td, hdr, hdrsize, numsegs, notesz);

	sb = sbuf_new(NULL, NULL, CORE_BUF_SIZE, SBUF_FIXEDLEN);
	sbuf_set_drain(sb, sbuf_drain_core_output, p);
	sbuf_start_section(sb, NULL);
	sbuf_bcat(sb, hdr, hdrsize);
	TAILQ_FOREACH(ninfo, notelst, link)
	    __elfN(putnote)(ninfo, sb);
	/* Align up to a page boundary for the program segments. */
	sbuf_end_section(sb, -1, PAGE_SIZE, 0);
	error = sbuf_finish(sb);
	sbuf_delete(sb);

	return (error);
}

static void
__elfN(prepare_notes)(struct thread *td, struct note_info_list *list,
    size_t *sizep)
{
	struct proc *p;
	struct thread *thr;
	size_t size;

	p = td->td_proc;
	size = 0;

	size += register_note(list, NT_PRPSINFO, __elfN(note_prpsinfo), p);

	/*
	 * To have the debugger select the right thread (LWP) as the initial
	 * thread, we dump the state of the thread passed to us in td first.
	 * This is the thread that causes the core dump and thus likely to
	 * be the right thread one wants to have selected in the debugger.
	 */
	thr = td;
	while (thr != NULL) {
		size += register_note(list, NT_PRSTATUS,
		    __elfN(note_prstatus), thr);
		size += register_note(list, NT_FPREGSET,
		    __elfN(note_fpregset), thr);
		size += register_note(list, NT_THRMISC,
		    __elfN(note_thrmisc), thr);
		size += register_note(list, NT_PTLWPINFO,
		    __elfN(note_ptlwpinfo), thr);
		size += register_note(list, -1,
		    __elfN(note_threadmd), thr);

		thr = (thr == td) ? TAILQ_FIRST(&p->p_threads) :
		    TAILQ_NEXT(thr, td_plist);
		if (thr == td)
			thr = TAILQ_NEXT(thr, td_plist);
	}

	size += register_note(list, NT_PROCSTAT_PROC,
	    __elfN(note_procstat_proc), p);
	size += register_note(list, NT_PROCSTAT_FILES,
	    note_procstat_files, p);
	size += register_note(list, NT_PROCSTAT_VMMAP,
	    note_procstat_vmmap, p);
	size += register_note(list, NT_PROCSTAT_GROUPS,
	    note_procstat_groups, p);
	size += register_note(list, NT_PROCSTAT_UMASK,
	    note_procstat_umask, p);
	size += register_note(list, NT_PROCSTAT_RLIMIT,
	    note_procstat_rlimit, p);
	size += register_note(list, NT_PROCSTAT_OSREL,
	    note_procstat_osrel, p);
	size += register_note(list, NT_PROCSTAT_PSSTRINGS,
	    __elfN(note_procstat_psstrings), p);
	size += register_note(list, NT_PROCSTAT_AUXV,
	    __elfN(note_procstat_auxv), p);

	*sizep = size;
}

static void
__elfN(puthdr)(struct thread *td, void *hdr, size_t hdrsize, int numsegs,
    size_t notesz)
{
	Elf_Ehdr *ehdr;
	Elf_Phdr *phdr;
	Elf_Shdr *shdr;
	struct phdr_closure phc;

	ehdr = (Elf_Ehdr *)hdr;

	ehdr->e_ident[EI_MAG0] = ELFMAG0;
	ehdr->e_ident[EI_MAG1] = ELFMAG1;
	ehdr->e_ident[EI_MAG2] = ELFMAG2;
	ehdr->e_ident[EI_MAG3] = ELFMAG3;
	ehdr->e_ident[EI_CLASS] = ELF_CLASS;
	ehdr->e_ident[EI_DATA] = ELF_DATA;
	ehdr->e_ident[EI_VERSION] = EV_CURRENT;
	ehdr->e_ident[EI_OSABI] = ELFOSABI_FREEBSD;
	ehdr->e_ident[EI_ABIVERSION] = 0;
	ehdr->e_ident[EI_PAD] = 0;
	ehdr->e_type = ET_CORE;
	ehdr->e_machine = td->td_proc->p_elf_machine;
	ehdr->e_version = EV_CURRENT;
	ehdr->e_entry = 0;
	ehdr->e_phoff = sizeof(Elf_Ehdr);
	ehdr->e_flags = td->td_proc->p_elf_flags;
	ehdr->e_ehsize = sizeof(Elf_Ehdr);
	ehdr->e_phentsize = sizeof(Elf_Phdr);
	ehdr->e_shentsize = sizeof(Elf_Shdr);
	ehdr->e_shstrndx = SHN_UNDEF;
	if (numsegs + 1 < PN_XNUM) {
		ehdr->e_phnum = numsegs + 1;
		ehdr->e_shnum = 0;
	} else {
		ehdr->e_phnum = PN_XNUM;
		ehdr->e_shnum = 1;

		ehdr->e_shoff = ehdr->e_phoff +
		    (numsegs + 1) * ehdr->e_phentsize;
		KASSERT(ehdr->e_shoff == hdrsize - sizeof(Elf_Shdr),
		    ("e_shoff: %zu, hdrsize - shdr: %zu",
		     (size_t)ehdr->e_shoff, hdrsize - sizeof(Elf_Shdr)));

		shdr = (Elf_Shdr *)((char *)hdr + ehdr->e_shoff);
		memset(shdr, 0, sizeof(*shdr));
		/*
		 * A special first section is used to hold large segment and
		 * section counts.  This was proposed by Sun Microsystems in
		 * Solaris and has been adopted by Linux; the standard ELF
		 * tools are already familiar with the technique.
		 *
		 * See table 7-7 of the Solaris "Linker and Libraries Guide"
		 * (or 12-7 depending on the version of the document) for more
		 * details.
		 */
		shdr->sh_type = SHT_NULL;
		shdr->sh_size = ehdr->e_shnum;
		shdr->sh_link = ehdr->e_shstrndx;
		shdr->sh_info = numsegs + 1;
	}

	/*
	 * Fill in the program header entries.
	 */
	phdr = (Elf_Phdr *)((char *)hdr + ehdr->e_phoff);

	/* The note segement. */
	phdr->p_type = PT_NOTE;
	phdr->p_offset = hdrsize;
	phdr->p_vaddr = 0;
	phdr->p_paddr = 0;
	phdr->p_filesz = notesz;
	phdr->p_memsz = 0;
	phdr->p_flags = PF_R;
	phdr->p_align = ELF_NOTE_ROUNDSIZE;
	phdr++;

	/* All the writable segments from the program. */
	phc.phdr = phdr;
	phc.offset = round_page(hdrsize + notesz);
	each_dumpable_segment(td, cb_put_phdr, &phc);
}

static size_t
register_note(struct note_info_list *list, int type, outfunc_t out, void *arg)
{
	struct note_info *ninfo;
	size_t size, notesize;

	size = 0;
	out(arg, NULL, &size);
	ninfo = malloc(sizeof(*ninfo), M_TEMP, M_ZERO | M_WAITOK);
	ninfo->type = type;
	ninfo->outfunc = out;
	ninfo->outarg = arg;
	ninfo->outsize = size;
	TAILQ_INSERT_TAIL(list, ninfo, link);

	if (type == -1)
		return (size);

	notesize = sizeof(Elf_Note) +		/* note header */
	    roundup2(sizeof(FREEBSD_ABI_VENDOR), ELF_NOTE_ROUNDSIZE) +
						/* note name */
	    roundup2(size, ELF_NOTE_ROUNDSIZE);	/* note description */

	return (notesize);
}

static size_t
append_note_data(const void *src, void *dst, size_t len)
{
	size_t padded_len;

	padded_len = roundup2(len, ELF_NOTE_ROUNDSIZE);
	if (dst != NULL) {
		bcopy(src, dst, len);
		bzero((char *)dst + len, padded_len - len);
	}
	return (padded_len);
}

size_t
__elfN(populate_note)(int type, void *src, void *dst, size_t size, void **descp)
{
	Elf_Note *note;
	char *buf;
	size_t notesize;

	buf = dst;
	if (buf != NULL) {
		note = (Elf_Note *)buf;
		note->n_namesz = sizeof(FREEBSD_ABI_VENDOR);
		note->n_descsz = size;
		note->n_type = type;
		buf += sizeof(*note);
		buf += append_note_data(FREEBSD_ABI_VENDOR, buf,
		    sizeof(FREEBSD_ABI_VENDOR));
		append_note_data(src, buf, size);
		if (descp != NULL)
			*descp = buf;
	}

	notesize = sizeof(Elf_Note) +		/* note header */
	    roundup2(sizeof(FREEBSD_ABI_VENDOR), ELF_NOTE_ROUNDSIZE) +
						/* note name */
	    roundup2(size, ELF_NOTE_ROUNDSIZE);	/* note description */

	return (notesize);
}

static void
__elfN(putnote)(struct note_info *ninfo, struct sbuf *sb)
{
	Elf_Note note;
	ssize_t old_len, sect_len;
	size_t new_len, descsz, i;

	if (ninfo->type == -1) {
		ninfo->outfunc(ninfo->outarg, sb, &ninfo->outsize);
		return;
	}

	note.n_namesz = sizeof(FREEBSD_ABI_VENDOR);
	note.n_descsz = ninfo->outsize;
	note.n_type = ninfo->type;

	sbuf_bcat(sb, &note, sizeof(note));
	sbuf_start_section(sb, &old_len);
	sbuf_bcat(sb, FREEBSD_ABI_VENDOR, sizeof(FREEBSD_ABI_VENDOR));
	sbuf_end_section(sb, old_len, ELF_NOTE_ROUNDSIZE, 0);
	if (note.n_descsz == 0)
		return;
	sbuf_start_section(sb, &old_len);
	ninfo->outfunc(ninfo->outarg, sb, &ninfo->outsize);
	sect_len = sbuf_end_section(sb, old_len, ELF_NOTE_ROUNDSIZE, 0);
	if (sect_len < 0)
		return;

	new_len = (size_t)sect_len;
	descsz = roundup(note.n_descsz, ELF_NOTE_ROUNDSIZE);
	if (new_len < descsz) {
		/*
		 * It is expected that individual note emitters will correctly
		 * predict their expected output size and fill up to that size
		 * themselves, padding in a format-specific way if needed.
		 * However, in case they don't, just do it here with zeros.
		 */
		for (i = 0; i < descsz - new_len; i++)
			sbuf_putc(sb, 0);
	} else if (new_len > descsz) {
		/*
		 * We can't always truncate sb -- we may have drained some
		 * of it already.
		 */
		KASSERT(new_len == descsz, ("%s: Note type %u changed as we "
		    "read it (%zu > %zu).  Since it is longer than "
		    "expected, this coredump's notes are corrupt.  THIS "
		    "IS A BUG in the note_procstat routine for type %u.\n",
		    __func__, (unsigned)note.n_type, new_len, descsz,
		    (unsigned)note.n_type));
	}
}

/*
 * Miscellaneous note out functions.
 */

#if defined(COMPAT_FREEBSD32) && __ELF_WORD_SIZE == 32
#include <compat/freebsd32/freebsd32.h>
#include <compat/freebsd32/freebsd32_signal.h>

typedef struct prstatus32 elf_prstatus_t;
typedef struct prpsinfo32 elf_prpsinfo_t;
typedef struct fpreg32 elf_prfpregset_t;
typedef struct fpreg32 elf_fpregset_t;
typedef struct reg32 elf_gregset_t;
typedef struct thrmisc32 elf_thrmisc_t;
#define ELF_KERN_PROC_MASK	KERN_PROC_MASK32
typedef struct kinfo_proc32 elf_kinfo_proc_t;
typedef uint32_t elf_ps_strings_t;
#else
typedef prstatus_t elf_prstatus_t;
typedef prpsinfo_t elf_prpsinfo_t;
typedef prfpregset_t elf_prfpregset_t;
typedef prfpregset_t elf_fpregset_t;
typedef gregset_t elf_gregset_t;
typedef thrmisc_t elf_thrmisc_t;
#define ELF_KERN_PROC_MASK	0
typedef struct kinfo_proc elf_kinfo_proc_t;
typedef vm_offset_t elf_ps_strings_t;
#endif

static void
__elfN(note_prpsinfo)(void *arg, struct sbuf *sb, size_t *sizep)
{
	struct sbuf sbarg;
	size_t len;
	char *cp, *end;
	struct proc *p;
	elf_prpsinfo_t *psinfo;
	int error;

	p = (struct proc *)arg;
	if (sb != NULL) {
		KASSERT(*sizep == sizeof(*psinfo), ("invalid size"));
		psinfo = malloc(sizeof(*psinfo), M_TEMP, M_ZERO | M_WAITOK);
		psinfo->pr_version = PRPSINFO_VERSION;
		psinfo->pr_psinfosz = sizeof(elf_prpsinfo_t);
		strlcpy(psinfo->pr_fname, p->p_comm, sizeof(psinfo->pr_fname));
		PROC_LOCK(p);
		if (p->p_args != NULL) {
			len = sizeof(psinfo->pr_psargs) - 1;
			if (len > p->p_args->ar_length)
				len = p->p_args->ar_length;
			memcpy(psinfo->pr_psargs, p->p_args->ar_args, len);
			PROC_UNLOCK(p);
			error = 0;
		} else {
			_PHOLD(p);
			PROC_UNLOCK(p);
			sbuf_new(&sbarg, psinfo->pr_psargs,
			    sizeof(psinfo->pr_psargs), SBUF_FIXEDLEN);
			error = proc_getargv(curthread, p, &sbarg);
			PRELE(p);
			if (sbuf_finish(&sbarg) == 0)
				len = sbuf_len(&sbarg) - 1;
			else
				len = sizeof(psinfo->pr_psargs) - 1;
			sbuf_delete(&sbarg);
		}
		if (error || len == 0)
			strlcpy(psinfo->pr_psargs, p->p_comm,
			    sizeof(psinfo->pr_psargs));
		else {
			KASSERT(len < sizeof(psinfo->pr_psargs),
			    ("len is too long: %zu vs %zu", len,
			    sizeof(psinfo->pr_psargs)));
			cp = psinfo->pr_psargs;
			end = cp + len - 1;
			for (;;) {
				cp = memchr(cp, '\0', end - cp);
				if (cp == NULL)
					break;
				*cp = ' ';
			}
		}
		psinfo->pr_pid = p->p_pid;
		sbuf_bcat(sb, psinfo, sizeof(*psinfo));
		free(psinfo, M_TEMP);
	}
	*sizep = sizeof(*psinfo);
}

static void
__elfN(note_prstatus)(void *arg, struct sbuf *sb, size_t *sizep)
{
	struct thread *td;
	elf_prstatus_t *status;

	td = (struct thread *)arg;
	if (sb != NULL) {
		KASSERT(*sizep == sizeof(*status), ("invalid size"));
		status = malloc(sizeof(*status), M_TEMP, M_ZERO | M_WAITOK);
		status->pr_version = PRSTATUS_VERSION;
		status->pr_statussz = sizeof(elf_prstatus_t);
		status->pr_gregsetsz = sizeof(elf_gregset_t);
		status->pr_fpregsetsz = sizeof(elf_fpregset_t);
		status->pr_osreldate = osreldate;
		status->pr_cursig = td->td_proc->p_sig;
		status->pr_pid = td->td_tid;
#if defined(COMPAT_FREEBSD32) && __ELF_WORD_SIZE == 32
		fill_regs32(td, &status->pr_reg);
#else
		fill_regs(td, &status->pr_reg);
#endif
		sbuf_bcat(sb, status, sizeof(*status));
		free(status, M_TEMP);
	}
	*sizep = sizeof(*status);
}

static void
__elfN(note_fpregset)(void *arg, struct sbuf *sb, size_t *sizep)
{
	struct thread *td;
	elf_prfpregset_t *fpregset;

	td = (struct thread *)arg;
	if (sb != NULL) {
		KASSERT(*sizep == sizeof(*fpregset), ("invalid size"));
		fpregset = malloc(sizeof(*fpregset), M_TEMP, M_ZERO | M_WAITOK);
#if defined(COMPAT_FREEBSD32) && __ELF_WORD_SIZE == 32
		fill_fpregs32(td, fpregset);
#else
		fill_fpregs(td, fpregset);
#endif
		sbuf_bcat(sb, fpregset, sizeof(*fpregset));
		free(fpregset, M_TEMP);
	}
	*sizep = sizeof(*fpregset);
}

static void
__elfN(note_thrmisc)(void *arg, struct sbuf *sb, size_t *sizep)
{
	struct thread *td;
	elf_thrmisc_t thrmisc;

	td = (struct thread *)arg;
	if (sb != NULL) {
		KASSERT(*sizep == sizeof(thrmisc), ("invalid size"));
		bzero(&thrmisc, sizeof(thrmisc));
		strcpy(thrmisc.pr_tname, td->td_name);
		sbuf_bcat(sb, &thrmisc, sizeof(thrmisc));
	}
	*sizep = sizeof(thrmisc);
}

static void
__elfN(note_ptlwpinfo)(void *arg, struct sbuf *sb, size_t *sizep)
{
	struct thread *td;
	size_t size;
	int structsize;
#if defined(COMPAT_FREEBSD32) && __ELF_WORD_SIZE == 32
	struct ptrace_lwpinfo32 pl;
#else
	struct ptrace_lwpinfo pl;
#endif

	td = (struct thread *)arg;
	size = sizeof(structsize) + sizeof(pl);
	if (sb != NULL) {
		KASSERT(*sizep == size, ("invalid size"));
		structsize = sizeof(pl);
		sbuf_bcat(sb, &structsize, sizeof(structsize));
		bzero(&pl, sizeof(pl));
		pl.pl_lwpid = td->td_tid;
		pl.pl_event = PL_EVENT_NONE;
		pl.pl_sigmask = td->td_sigmask;
		pl.pl_siglist = td->td_siglist;
		if (td->td_si.si_signo != 0) {
			pl.pl_event = PL_EVENT_SIGNAL;
			pl.pl_flags |= PL_FLAG_SI;
#if defined(COMPAT_FREEBSD32) && __ELF_WORD_SIZE == 32
			siginfo_to_siginfo32(&td->td_si, &pl.pl_siginfo);
#else
			pl.pl_siginfo = td->td_si;
#endif
		}
		strcpy(pl.pl_tdname, td->td_name);
		/* XXX TODO: supply more information in struct ptrace_lwpinfo*/
		sbuf_bcat(sb, &pl, sizeof(pl));
	}
	*sizep = size;
}

/*
 * Allow for MD specific notes, as well as any MD
 * specific preparations for writing MI notes.
 */
static void
__elfN(note_threadmd)(void *arg, struct sbuf *sb, size_t *sizep)
{
	struct thread *td;
	void *buf;
	size_t size;

	td = (struct thread *)arg;
	size = *sizep;
	if (size != 0 && sb != NULL)
		buf = malloc(size, M_TEMP, M_ZERO | M_WAITOK);
	else
		buf = NULL;
	size = 0;
	__elfN(dump_thread)(td, buf, &size);
	KASSERT(sb == NULL || *sizep == size, ("invalid size"));
	if (size != 0 && sb != NULL)
		sbuf_bcat(sb, buf, size);
	free(buf, M_TEMP);
	*sizep = size;
}

#ifdef KINFO_PROC_SIZE
CTASSERT(sizeof(struct kinfo_proc) == KINFO_PROC_SIZE);
#endif

static void
__elfN(note_procstat_proc)(void *arg, struct sbuf *sb, size_t *sizep)
{
	struct proc *p;
	size_t size;
	int structsize;

	p = (struct proc *)arg;
	size = sizeof(structsize) + p->p_numthreads *
	    sizeof(elf_kinfo_proc_t);

	if (sb != NULL) {
		KASSERT(*sizep == size, ("invalid size"));
		structsize = sizeof(elf_kinfo_proc_t);
		sbuf_bcat(sb, &structsize, sizeof(structsize));
		sx_slock(&proctree_lock);
		PROC_LOCK(p);
		kern_proc_out(p, sb, ELF_KERN_PROC_MASK);
		sx_sunlock(&proctree_lock);
	}
	*sizep = size;
}

#ifdef KINFO_FILE_SIZE
CTASSERT(sizeof(struct kinfo_file) == KINFO_FILE_SIZE);
#endif

static void
note_procstat_files(void *arg, struct sbuf *sb, size_t *sizep)
{
	struct proc *p;
	size_t size, sect_sz, i;
	ssize_t start_len, sect_len;
	int structsize, filedesc_flags;

	if (coredump_pack_fileinfo)
		filedesc_flags = KERN_FILEDESC_PACK_KINFO;
	else
		filedesc_flags = 0;

	p = (struct proc *)arg;
	structsize = sizeof(struct kinfo_file);
	if (sb == NULL) {
		size = 0;
		sb = sbuf_new(NULL, NULL, 128, SBUF_FIXEDLEN);
		sbuf_set_drain(sb, sbuf_count_drain, &size);
		sbuf_bcat(sb, &structsize, sizeof(structsize));
		PROC_LOCK(p);
		kern_proc_filedesc_out(p, sb, -1, filedesc_flags);
		sbuf_finish(sb);
		sbuf_delete(sb);
		*sizep = size;
	} else {
		sbuf_start_section(sb, &start_len);

		sbuf_bcat(sb, &structsize, sizeof(structsize));
		PROC_LOCK(p);
		kern_proc_filedesc_out(p, sb, *sizep - sizeof(structsize),
		    filedesc_flags);

		sect_len = sbuf_end_section(sb, start_len, 0, 0);
		if (sect_len < 0)
			return;
		sect_sz = sect_len;

		KASSERT(sect_sz <= *sizep,
		    ("kern_proc_filedesc_out did not respect maxlen; "
		     "requested %zu, got %zu", *sizep - sizeof(structsize),
		     sect_sz - sizeof(structsize)));

		for (i = 0; i < *sizep - sect_sz && sb->s_error == 0; i++)
			sbuf_putc(sb, 0);
	}
}

#ifdef KINFO_VMENTRY_SIZE
CTASSERT(sizeof(struct kinfo_vmentry) == KINFO_VMENTRY_SIZE);
#endif

static void
note_procstat_vmmap(void *arg, struct sbuf *sb, size_t *sizep)
{
	struct proc *p;
	size_t size;
	int structsize, vmmap_flags;

	if (coredump_pack_vmmapinfo)
		vmmap_flags = KERN_VMMAP_PACK_KINFO;
	else
		vmmap_flags = 0;

	p = (struct proc *)arg;
	structsize = sizeof(struct kinfo_vmentry);
	if (sb == NULL) {
		size = 0;
		sb = sbuf_new(NULL, NULL, 128, SBUF_FIXEDLEN);
		sbuf_set_drain(sb, sbuf_count_drain, &size);
		sbuf_bcat(sb, &structsize, sizeof(structsize));
		PROC_LOCK(p);
		kern_proc_vmmap_out(p, sb, -1, vmmap_flags);
		sbuf_finish(sb);
		sbuf_delete(sb);
		*sizep = size;
	} else {
		sbuf_bcat(sb, &structsize, sizeof(structsize));
		PROC_LOCK(p);
		kern_proc_vmmap_out(p, sb, *sizep - sizeof(structsize),
		    vmmap_flags);
	}
}

static void
note_procstat_groups(void *arg, struct sbuf *sb, size_t *sizep)
{
	struct proc *p;
	size_t size;
	int structsize;

	p = (struct proc *)arg;
	size = sizeof(structsize) + p->p_ucred->cr_ngroups * sizeof(gid_t);
	if (sb != NULL) {
		KASSERT(*sizep == size, ("invalid size"));
		structsize = sizeof(gid_t);
		sbuf_bcat(sb, &structsize, sizeof(structsize));
		sbuf_bcat(sb, p->p_ucred->cr_groups, p->p_ucred->cr_ngroups *
		    sizeof(gid_t));
	}
	*sizep = size;
}

static void
note_procstat_umask(void *arg, struct sbuf *sb, size_t *sizep)
{
	struct proc *p;
	size_t size;
	int structsize;

	p = (struct proc *)arg;
	size = sizeof(structsize) + sizeof(p->p_pd->pd_cmask);
	if (sb != NULL) {
		KASSERT(*sizep == size, ("invalid size"));
		structsize = sizeof(p->p_pd->pd_cmask);
		sbuf_bcat(sb, &structsize, sizeof(structsize));
		sbuf_bcat(sb, &p->p_pd->pd_cmask, sizeof(p->p_pd->pd_cmask));
	}
	*sizep = size;
}

static void
note_procstat_rlimit(void *arg, struct sbuf *sb, size_t *sizep)
{
	struct proc *p;
	struct rlimit rlim[RLIM_NLIMITS];
	size_t size;
	int structsize, i;

	p = (struct proc *)arg;
	size = sizeof(structsize) + sizeof(rlim);
	if (sb != NULL) {
		KASSERT(*sizep == size, ("invalid size"));
		structsize = sizeof(rlim);
		sbuf_bcat(sb, &structsize, sizeof(structsize));
		PROC_LOCK(p);
		for (i = 0; i < RLIM_NLIMITS; i++)
			lim_rlimit_proc(p, i, &rlim[i]);
		PROC_UNLOCK(p);
		sbuf_bcat(sb, rlim, sizeof(rlim));
	}
	*sizep = size;
}

static void
note_procstat_osrel(void *arg, struct sbuf *sb, size_t *sizep)
{
	struct proc *p;
	size_t size;
	int structsize;

	p = (struct proc *)arg;
	size = sizeof(structsize) + sizeof(p->p_osrel);
	if (sb != NULL) {
		KASSERT(*sizep == size, ("invalid size"));
		structsize = sizeof(p->p_osrel);
		sbuf_bcat(sb, &structsize, sizeof(structsize));
		sbuf_bcat(sb, &p->p_osrel, sizeof(p->p_osrel));
	}
	*sizep = size;
}

static void
__elfN(note_procstat_psstrings)(void *arg, struct sbuf *sb, size_t *sizep)
{
	struct proc *p;
	elf_ps_strings_t ps_strings;
	size_t size;
	int structsize;

	p = (struct proc *)arg;
	size = sizeof(structsize) + sizeof(ps_strings);
	if (sb != NULL) {
		KASSERT(*sizep == size, ("invalid size"));
		structsize = sizeof(ps_strings);
#if defined(COMPAT_FREEBSD32) && __ELF_WORD_SIZE == 32
		ps_strings = PTROUT(p->p_sysent->sv_psstrings);
#else
		ps_strings = p->p_sysent->sv_psstrings;
#endif
		sbuf_bcat(sb, &structsize, sizeof(structsize));
		sbuf_bcat(sb, &ps_strings, sizeof(ps_strings));
	}
	*sizep = size;
}

static void
__elfN(note_procstat_auxv)(void *arg, struct sbuf *sb, size_t *sizep)
{
	struct proc *p;
	size_t size;
	int structsize;

	p = (struct proc *)arg;
	if (sb == NULL) {
		size = 0;
		sb = sbuf_new(NULL, NULL, 128, SBUF_FIXEDLEN);
		sbuf_set_drain(sb, sbuf_count_drain, &size);
		sbuf_bcat(sb, &structsize, sizeof(structsize));
		PHOLD(p);
		proc_getauxv(curthread, p, sb);
		PRELE(p);
		sbuf_finish(sb);
		sbuf_delete(sb);
		*sizep = size;
	} else {
		structsize = sizeof(Elf_Auxinfo);
		sbuf_bcat(sb, &structsize, sizeof(structsize));
		PHOLD(p);
		proc_getauxv(curthread, p, sb);
		PRELE(p);
	}
}

static boolean_t
__elfN(parse_notes)(struct image_params *imgp, Elf_Note *checknote,
    const char *note_vendor, const Elf_Phdr *pnote,
    boolean_t (*cb)(const Elf_Note *, void *, boolean_t *), void *cb_arg)
{
	const Elf_Note *note, *note0, *note_end;
	const char *note_name;
	char *buf;
	int i, error;
	boolean_t res;

	/* We need some limit, might as well use PAGE_SIZE. */
	if (pnote == NULL || pnote->p_filesz > PAGE_SIZE)
		return (FALSE);
	ASSERT_VOP_LOCKED(imgp->vp, "parse_notes");
	if (pnote->p_offset > PAGE_SIZE ||
	    pnote->p_filesz > PAGE_SIZE - pnote->p_offset) {
		buf = malloc(pnote->p_filesz, M_TEMP, M_NOWAIT);
		if (buf == NULL) {
			VOP_UNLOCK(imgp->vp);
			buf = malloc(pnote->p_filesz, M_TEMP, M_WAITOK);
			vn_lock(imgp->vp, LK_SHARED | LK_RETRY);
		}
		error = vn_rdwr(UIO_READ, imgp->vp, buf, pnote->p_filesz,
		    pnote->p_offset, UIO_SYSSPACE, IO_NODELOCKED,
		    curthread->td_ucred, NOCRED, NULL, curthread);
		if (error != 0) {
			uprintf("i/o error PT_NOTE\n");
			goto retf;
		}
		note = note0 = (const Elf_Note *)buf;
		note_end = (const Elf_Note *)(buf + pnote->p_filesz);
	} else {
		note = note0 = (const Elf_Note *)(imgp->image_header +
		    pnote->p_offset);
		note_end = (const Elf_Note *)(imgp->image_header +
		    pnote->p_offset + pnote->p_filesz);
		buf = NULL;
	}
	for (i = 0; i < 100 && note >= note0 && note < note_end; i++) {
		if (!aligned(note, Elf32_Addr) || (const char *)note_end -
		    (const char *)note < sizeof(Elf_Note)) {
			goto retf;
		}
		if (note->n_namesz != checknote->n_namesz ||
		    note->n_descsz != checknote->n_descsz ||
		    note->n_type != checknote->n_type)
			goto nextnote;
		note_name = (const char *)(note + 1);
		if (note_name + checknote->n_namesz >=
		    (const char *)note_end || strncmp(note_vendor,
		    note_name, checknote->n_namesz) != 0)
			goto nextnote;

		if (cb(note, cb_arg, &res))
			goto ret;
nextnote:
		note = (const Elf_Note *)((const char *)(note + 1) +
		    roundup2(note->n_namesz, ELF_NOTE_ROUNDSIZE) +
		    roundup2(note->n_descsz, ELF_NOTE_ROUNDSIZE));
	}
retf:
	res = FALSE;
ret:
	free(buf, M_TEMP);
	return (res);
}

struct brandnote_cb_arg {
	Elf_Brandnote *brandnote;
	int32_t *osrel;
};

static boolean_t
brandnote_cb(const Elf_Note *note, void *arg0, boolean_t *res)
{
	struct brandnote_cb_arg *arg;

	arg = arg0;

	/*
	 * Fetch the osreldate for binary from the ELF OSABI-note if
	 * necessary.
	 */
	*res = (arg->brandnote->flags & BN_TRANSLATE_OSREL) != 0 &&
	    arg->brandnote->trans_osrel != NULL ?
	    arg->brandnote->trans_osrel(note, arg->osrel) : TRUE;

	return (TRUE);
}

static Elf_Note fctl_note = {
	.n_namesz = sizeof(FREEBSD_ABI_VENDOR),
	.n_descsz = sizeof(uint32_t),
	.n_type = NT_FREEBSD_FEATURE_CTL,
};

struct fctl_cb_arg {
	boolean_t *has_fctl0;
	uint32_t *fctl0;
};

static boolean_t
note_fctl_cb(const Elf_Note *note, void *arg0, boolean_t *res)
{
	struct fctl_cb_arg *arg;
	const Elf32_Word *desc;
	uintptr_t p;

	arg = arg0;
	p = (uintptr_t)(note + 1);
	p += roundup2(note->n_namesz, ELF_NOTE_ROUNDSIZE);
	desc = (const Elf32_Word *)p;
	*arg->has_fctl0 = TRUE;
	*arg->fctl0 = desc[0];
	return (TRUE);
}

/*
 * Try to find the appropriate ABI-note section for checknote, fetch
 * the osreldate and feature control flags for binary from the ELF
 * OSABI-note.  Only the first page of the image is searched, the same
 * as for headers.
 */
static boolean_t
__elfN(check_note)(struct image_params *imgp, Elf_Brandnote *brandnote,
    int32_t *osrel, boolean_t *has_fctl0, uint32_t *fctl0)
{
	const Elf_Phdr *phdr;
	const Elf_Ehdr *hdr;
	struct brandnote_cb_arg b_arg;
	struct fctl_cb_arg f_arg;
	int i, j;

	hdr = (const Elf_Ehdr *)imgp->image_header;
	phdr = (const Elf_Phdr *)(imgp->image_header + hdr->e_phoff);
	b_arg.brandnote = brandnote;
	b_arg.osrel = osrel;
	f_arg.has_fctl0 = has_fctl0;
	f_arg.fctl0 = fctl0;

	for (i = 0; i < hdr->e_phnum; i++) {
		if (phdr[i].p_type == PT_NOTE && __elfN(parse_notes)(imgp,
		    &brandnote->hdr, brandnote->vendor, &phdr[i], brandnote_cb,
		    &b_arg)) {
			for (j = 0; j < hdr->e_phnum; j++) {
				if (phdr[j].p_type == PT_NOTE &&
				    __elfN(parse_notes)(imgp, &fctl_note,
				    FREEBSD_ABI_VENDOR, &phdr[j],
				    note_fctl_cb, &f_arg))
					break;
			}
			return (TRUE);
		}
	}
	return (FALSE);

}

/*
 * Tell kern_execve.c about it, with a little help from the linker.
 */
static struct execsw __elfN(execsw) = {
	.ex_imgact = __CONCAT(exec_, __elfN(imgact)),
	.ex_name = __XSTRING(__CONCAT(ELF, __ELF_WORD_SIZE))
};
EXEC_SET(__CONCAT(elf, __ELF_WORD_SIZE), __elfN(execsw));

static vm_prot_t
__elfN(trans_prot)(Elf_Word flags)
{
	vm_prot_t prot;

	prot = 0;
	if (flags & PF_X)
		prot |= VM_PROT_EXECUTE;
	if (flags & PF_W)
		prot |= VM_PROT_WRITE;
	if (flags & PF_R)
		prot |= VM_PROT_READ;
#if __ELF_WORD_SIZE == 32 && (defined(__amd64__) || defined(__i386__))
	if (i386_read_exec && (flags & PF_R))
		prot |= VM_PROT_EXECUTE;
#endif
	return (prot);
}

static Elf_Word
__elfN(untrans_prot)(vm_prot_t prot)
{
	Elf_Word flags;

	flags = 0;
	if (prot & VM_PROT_EXECUTE)
		flags |= PF_X;
	if (prot & VM_PROT_READ)
		flags |= PF_R;
	if (prot & VM_PROT_WRITE)
		flags |= PF_W;
	return (flags);
}

void
__elfN(stackgap)(struct image_params *imgp, uintptr_t *stack_base)
{
	uintptr_t range, rbase, gap;
	int pct;

	pct = __elfN(aslr_stack_gap);
	if (pct == 0)
		return;
	if (pct > 50)
		pct = 50;
	range = imgp->eff_stack_sz * pct / 100;
	arc4rand(&rbase, sizeof(rbase), 0);
	gap = rbase % range;
	gap &= ~(sizeof(u_long) - 1);
	*stack_base -= gap;
}