aboutsummaryrefslogtreecommitdiffstats
path: root/fs/udf/super.c
blob: 0f8b3cb355852acb362b7fbc6f8f7f23fdfdc7cd (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
/*
 * super.c
 *
 * PURPOSE
 *  Super block routines for the OSTA-UDF(tm) filesystem.
 *
 * DESCRIPTION
 *  OSTA-UDF(tm) = Optical Storage Technology Association
 *  Universal Disk Format.
 *
 *  This code is based on version 2.00 of the UDF specification,
 *  and revision 3 of the ECMA 167 standard [equivalent to ISO 13346].
 *    http://www.osta.org/
 *    http://www.ecma.ch/
 *    http://www.iso.org/
 *
 * COPYRIGHT
 *  This file is distributed under the terms of the GNU General Public
 *  License (GPL). Copies of the GPL can be obtained from:
 *    ftp://prep.ai.mit.edu/pub/gnu/GPL
 *  Each contributing author retains all rights to their own work.
 *
 *  (C) 1998 Dave Boynton
 *  (C) 1998-2004 Ben Fennema
 *  (C) 2000 Stelias Computing Inc
 *
 * HISTORY
 *
 *  09/24/98 dgb  changed to allow compiling outside of kernel, and
 *                added some debugging.
 *  10/01/98 dgb  updated to allow (some) possibility of compiling w/2.0.34
 *  10/16/98      attempting some multi-session support
 *  10/17/98      added freespace count for "df"
 *  11/11/98 gr   added novrs option
 *  11/26/98 dgb  added fileset,anchor mount options
 *  12/06/98 blf  really hosed things royally. vat/sparing support. sequenced
 *                vol descs. rewrote option handling based on isofs
 *  12/20/98      find the free space bitmap (if it exists)
 */

#include "udfdecl.h"

#include <linux/blkdev.h>
#include <linux/slab.h>
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/parser.h>
#include <linux/stat.h>
#include <linux/cdrom.h>
#include <linux/nls.h>
#include <linux/vfs.h>
#include <linux/vmalloc.h>
#include <linux/errno.h>
#include <linux/mount.h>
#include <linux/seq_file.h>
#include <linux/bitmap.h>
#include <linux/crc-itu-t.h>
#include <linux/log2.h>
#include <asm/byteorder.h>
#include <linux/iversion.h>

#include "udf_sb.h"
#include "udf_i.h"

#include <linux/init.h>
#include <linux/uaccess.h>

enum {
	VDS_POS_PRIMARY_VOL_DESC,
	VDS_POS_UNALLOC_SPACE_DESC,
	VDS_POS_LOGICAL_VOL_DESC,
	VDS_POS_IMP_USE_VOL_DESC,
	VDS_POS_LENGTH
};

#define VSD_FIRST_SECTOR_OFFSET		32768
#define VSD_MAX_SECTOR_OFFSET		0x800000

/*
 * Maximum number of Terminating Descriptor / Logical Volume Integrity
 * Descriptor redirections. The chosen numbers are arbitrary - just that we
 * hopefully don't limit any real use of rewritten inode on write-once media
 * but avoid looping for too long on corrupted media.
 */
#define UDF_MAX_TD_NESTING 64
#define UDF_MAX_LVID_NESTING 1000

enum { UDF_MAX_LINKS = 0xffff };

/* These are the "meat" - everything else is stuffing */
static int udf_fill_super(struct super_block *, void *, int);
static void udf_put_super(struct super_block *);
static int udf_sync_fs(struct super_block *, int);
static int udf_remount_fs(struct super_block *, int *, char *);
static void udf_load_logicalvolint(struct super_block *, struct kernel_extent_ad);
static void udf_open_lvid(struct super_block *);
static void udf_close_lvid(struct super_block *);
static unsigned int udf_count_free(struct super_block *);
static int udf_statfs(struct dentry *, struct kstatfs *);
static int udf_show_options(struct seq_file *, struct dentry *);

struct logicalVolIntegrityDescImpUse *udf_sb_lvidiu(struct super_block *sb)
{
	struct logicalVolIntegrityDesc *lvid;
	unsigned int partnum;
	unsigned int offset;

	if (!UDF_SB(sb)->s_lvid_bh)
		return NULL;
	lvid = (struct logicalVolIntegrityDesc *)UDF_SB(sb)->s_lvid_bh->b_data;
	partnum = le32_to_cpu(lvid->numOfPartitions);
	/* The offset is to skip freeSpaceTable and sizeTable arrays */
	offset = partnum * 2 * sizeof(uint32_t);
	return (struct logicalVolIntegrityDescImpUse *)
					(((uint8_t *)(lvid + 1)) + offset);
}

/* UDF filesystem type */
static struct dentry *udf_mount(struct file_system_type *fs_type,
		      int flags, const char *dev_name, void *data)
{
	return mount_bdev(fs_type, flags, dev_name, data, udf_fill_super);
}

static struct file_system_type udf_fstype = {
	.owner		= THIS_MODULE,
	.name		= "udf",
	.mount		= udf_mount,
	.kill_sb	= kill_block_super,
	.fs_flags	= FS_REQUIRES_DEV,
};
MODULE_ALIAS_FS("udf");

static struct kmem_cache *udf_inode_cachep;

static struct inode *udf_alloc_inode(struct super_block *sb)
{
	struct udf_inode_info *ei;
	ei = kmem_cache_alloc(udf_inode_cachep, GFP_KERNEL);
	if (!ei)
		return NULL;

	ei->i_unique = 0;
	ei->i_lenExtents = 0;
	ei->i_lenStreams = 0;
	ei->i_next_alloc_block = 0;
	ei->i_next_alloc_goal = 0;
	ei->i_strat4096 = 0;
	ei->i_streamdir = 0;
	ei->i_hidden = 0;
	init_rwsem(&ei->i_data_sem);
	ei->cached_extent.lstart = -1;
	spin_lock_init(&ei->i_extent_cache_lock);
	inode_set_iversion(&ei->vfs_inode, 1);

	return &ei->vfs_inode;
}

static void udf_free_in_core_inode(struct inode *inode)
{
	kmem_cache_free(udf_inode_cachep, UDF_I(inode));
}

static void init_once(void *foo)
{
	struct udf_inode_info *ei = (struct udf_inode_info *)foo;

	ei->i_ext.i_data = NULL;
	inode_init_once(&ei->vfs_inode);
}

static int __init init_inodecache(void)
{
	udf_inode_cachep = kmem_cache_create("udf_inode_cache",
					     sizeof(struct udf_inode_info),
					     0, (SLAB_RECLAIM_ACCOUNT |
						 SLAB_MEM_SPREAD |
						 SLAB_ACCOUNT),
					     init_once);
	if (!udf_inode_cachep)
		return -ENOMEM;
	return 0;
}

static void destroy_inodecache(void)
{
	/*
	 * Make sure all delayed rcu free inodes are flushed before we
	 * destroy cache.
	 */
	rcu_barrier();
	kmem_cache_destroy(udf_inode_cachep);
}

/* Superblock operations */
static const struct super_operations udf_sb_ops = {
	.alloc_inode	= udf_alloc_inode,
	.free_inode	= udf_free_in_core_inode,
	.write_inode	= udf_write_inode,
	.evict_inode	= udf_evict_inode,
	.put_super	= udf_put_super,
	.sync_fs	= udf_sync_fs,
	.statfs		= udf_statfs,
	.remount_fs	= udf_remount_fs,
	.show_options	= udf_show_options,
};

struct udf_options {
	unsigned char novrs;
	unsigned int blocksize;
	unsigned int session;
	unsigned int lastblock;
	unsigned int anchor;
	unsigned int flags;
	umode_t umask;
	kgid_t gid;
	kuid_t uid;
	umode_t fmode;
	umode_t dmode;
	struct nls_table *nls_map;
};

static int __init init_udf_fs(void)
{
	int err;

	err = init_inodecache();
	if (err)
		goto out1;
	err = register_filesystem(&udf_fstype);
	if (err)
		goto out;

	return 0;

out:
	destroy_inodecache();

out1:
	return err;
}

static void __exit exit_udf_fs(void)
{
	unregister_filesystem(&udf_fstype);
	destroy_inodecache();
}

static int udf_sb_alloc_partition_maps(struct super_block *sb, u32 count)
{
	struct udf_sb_info *sbi = UDF_SB(sb);

	sbi->s_partmaps = kcalloc(count, sizeof(*sbi->s_partmaps), GFP_KERNEL);
	if (!sbi->s_partmaps) {
		sbi->s_partitions = 0;
		return -ENOMEM;
	}

	sbi->s_partitions = count;
	return 0;
}

static void udf_sb_free_bitmap(struct udf_bitmap *bitmap)
{
	int i;
	int nr_groups = bitmap->s_nr_groups;

	for (i = 0; i < nr_groups; i++)
		brelse(bitmap->s_block_bitmap[i]);

	kvfree(bitmap);
}

static void udf_free_partition(struct udf_part_map *map)
{
	int i;
	struct udf_meta_data *mdata;

	if (map->s_partition_flags & UDF_PART_FLAG_UNALLOC_TABLE)
		iput(map->s_uspace.s_table);
	if (map->s_partition_flags & UDF_PART_FLAG_UNALLOC_BITMAP)
		udf_sb_free_bitmap(map->s_uspace.s_bitmap);
	if (map->s_partition_type == UDF_SPARABLE_MAP15)
		for (i = 0; i < 4; i++)
			brelse(map->s_type_specific.s_sparing.s_spar_map[i]);
	else if (map->s_partition_type == UDF_METADATA_MAP25) {
		mdata = &map->s_type_specific.s_metadata;
		iput(mdata->s_metadata_fe);
		mdata->s_metadata_fe = NULL;

		iput(mdata->s_mirror_fe);
		mdata->s_mirror_fe = NULL;

		iput(mdata->s_bitmap_fe);
		mdata->s_bitmap_fe = NULL;
	}
}

static void udf_sb_free_partitions(struct super_block *sb)
{
	struct udf_sb_info *sbi = UDF_SB(sb);
	int i;

	if (!sbi->s_partmaps)
		return;
	for (i = 0; i < sbi->s_partitions; i++)
		udf_free_partition(&sbi->s_partmaps[i]);
	kfree(sbi->s_partmaps);
	sbi->s_partmaps = NULL;
}

static int udf_show_options(struct seq_file *seq, struct dentry *root)
{
	struct super_block *sb = root->d_sb;
	struct udf_sb_info *sbi = UDF_SB(sb);

	if (!UDF_QUERY_FLAG(sb, UDF_FLAG_STRICT))
		seq_puts(seq, ",nostrict");
	if (UDF_QUERY_FLAG(sb, UDF_FLAG_BLOCKSIZE_SET))
		seq_printf(seq, ",bs=%lu", sb->s_blocksize);
	if (UDF_QUERY_FLAG(sb, UDF_FLAG_UNHIDE))
		seq_puts(seq, ",unhide");
	if (UDF_QUERY_FLAG(sb, UDF_FLAG_UNDELETE))
		seq_puts(seq, ",undelete");
	if (!UDF_QUERY_FLAG(sb, UDF_FLAG_USE_AD_IN_ICB))
		seq_puts(seq, ",noadinicb");
	if (UDF_QUERY_FLAG(sb, UDF_FLAG_USE_SHORT_AD))
		seq_puts(seq, ",shortad");
	if (UDF_QUERY_FLAG(sb, UDF_FLAG_UID_FORGET))
		seq_puts(seq, ",uid=forget");
	if (UDF_QUERY_FLAG(sb, UDF_FLAG_GID_FORGET))
		seq_puts(seq, ",gid=forget");
	if (UDF_QUERY_FLAG(sb, UDF_FLAG_UID_SET))
		seq_printf(seq, ",uid=%u", from_kuid(&init_user_ns, sbi->s_uid));
	if (UDF_QUERY_FLAG(sb, UDF_FLAG_GID_SET))
		seq_printf(seq, ",gid=%u", from_kgid(&init_user_ns, sbi->s_gid));
	if (sbi->s_umask != 0)
		seq_printf(seq, ",umask=%ho", sbi->s_umask);
	if (sbi->s_fmode != UDF_INVALID_MODE)
		seq_printf(seq, ",mode=%ho", sbi->s_fmode);
	if (sbi->s_dmode != UDF_INVALID_MODE)
		seq_printf(seq, ",dmode=%ho", sbi->s_dmode);
	if (UDF_QUERY_FLAG(sb, UDF_FLAG_SESSION_SET))
		seq_printf(seq, ",session=%d", sbi->s_session);
	if (UDF_QUERY_FLAG(sb, UDF_FLAG_LASTBLOCK_SET))
		seq_printf(seq, ",lastblock=%u", sbi->s_last_block);
	if (sbi->s_anchor != 0)
		seq_printf(seq, ",anchor=%u", sbi->s_anchor);
	if (sbi->s_nls_map)
		seq_printf(seq, ",iocharset=%s", sbi->s_nls_map->charset);
	else
		seq_puts(seq, ",iocharset=utf8");

	return 0;
}

/*
 * udf_parse_options
 *
 * PURPOSE
 *	Parse mount options.
 *
 * DESCRIPTION
 *	The following mount options are supported:
 *
 *	gid=		Set the default group.
 *	umask=		Set the default umask.
 *	mode=		Set the default file permissions.
 *	dmode=		Set the default directory permissions.
 *	uid=		Set the default user.
 *	bs=		Set the block size.
 *	unhide		Show otherwise hidden files.
 *	undelete	Show deleted files in lists.
 *	adinicb		Embed data in the inode (default)
 *	noadinicb	Don't embed data in the inode
 *	shortad		Use short ad's
 *	longad		Use long ad's (default)
 *	nostrict	Unset strict conformance
 *	iocharset=	Set the NLS character set
 *
 *	The remaining are for debugging and disaster recovery:
 *
 *	novrs		Skip volume sequence recognition
 *
 *	The following expect a offset from 0.
 *
 *	session=	Set the CDROM session (default= last session)
 *	anchor=		Override standard anchor location. (default= 256)
 *	volume=		Override the VolumeDesc location. (unused)
 *	partition=	Override the PartitionDesc location. (unused)
 *	lastblock=	Set the last block of the filesystem/
 *
 *	The following expect a offset from the partition root.
 *
 *	fileset=	Override the fileset block location. (unused)
 *	rootdir=	Override the root directory location. (unused)
 *		WARNING: overriding the rootdir to a non-directory may
 *		yield highly unpredictable results.
 *
 * PRE-CONDITIONS
 *	options		Pointer to mount options string.
 *	uopts		Pointer to mount options variable.
 *
 * POST-CONDITIONS
 *	<return>	1	Mount options parsed okay.
 *	<return>	0	Error parsing mount options.
 *
 * HISTORY
 *	July 1, 1997 - Andrew E. Mileski
 *	Written, tested, and released.
 */

enum {
	Opt_novrs, Opt_nostrict, Opt_bs, Opt_unhide, Opt_undelete,
	Opt_noadinicb, Opt_adinicb, Opt_shortad, Opt_longad,
	Opt_gid, Opt_uid, Opt_umask, Opt_session, Opt_lastblock,
	Opt_anchor, Opt_volume, Opt_partition, Opt_fileset,
	Opt_rootdir, Opt_utf8, Opt_iocharset,
	Opt_err, Opt_uforget, Opt_uignore, Opt_gforget, Opt_gignore,
	Opt_fmode, Opt_dmode
};

static const match_table_t tokens = {
	{Opt_novrs,	"novrs"},
	{Opt_nostrict,	"nostrict"},
	{Opt_bs,	"bs=%u"},
	{Opt_unhide,	"unhide"},
	{Opt_undelete,	"undelete"},
	{Opt_noadinicb,	"noadinicb"},
	{Opt_adinicb,	"adinicb"},
	{Opt_shortad,	"shortad"},
	{Opt_longad,	"longad"},
	{Opt_uforget,	"uid=forget"},
	{Opt_uignore,	"uid=ignore"},
	{Opt_gforget,	"gid=forget"},
	{Opt_gignore,	"gid=ignore"},
	{Opt_gid,	"gid=%u"},
	{Opt_uid,	"uid=%u"},
	{Opt_umask,	"umask=%o"},
	{Opt_session,	"session=%u"},
	{Opt_lastblock,	"lastblock=%u"},
	{Opt_anchor,	"anchor=%u"},
	{Opt_volume,	"volume=%u"},
	{Opt_partition,	"partition=%u"},
	{Opt_fileset,	"fileset=%u"},
	{Opt_rootdir,	"rootdir=%u"},
	{Opt_utf8,	"utf8"},
	{Opt_iocharset,	"iocharset=%s"},
	{Opt_fmode,     "mode=%o"},
	{Opt_dmode,     "dmode=%o"},
	{Opt_err,	NULL}
};

static int udf_parse_options(char *options, struct udf_options *uopt,
			     bool remount)
{
	char *p;
	int option;

	uopt->novrs = 0;
	uopt->session = 0xFFFFFFFF;
	uopt->lastblock = 0;
	uopt->anchor = 0;

	if (!options)
		return 1;

	while ((p = strsep(&options, ",")) != NULL) {
		substring_t args[MAX_OPT_ARGS];
		int token;
		unsigned n;
		if (!*p)
			continue;

		token = match_token(p, tokens, args);
		switch (token) {
		case Opt_novrs:
			uopt->novrs = 1;
			break;
		case Opt_bs:
			if (match_int(&args[0], &option))
				return 0;
			n = option;
			if (n != 512 && n != 1024 && n != 2048 && n != 4096)
				return 0;
			uopt->blocksize = n;
			uopt->flags |= (1 << UDF_FLAG_BLOCKSIZE_SET);
			break;
		case Opt_unhide:
			uopt->flags |= (1 << UDF_FLAG_UNHIDE);
			break;
		case Opt_undelete:
			uopt->flags |= (1 << UDF_FLAG_UNDELETE);
			break;
		case Opt_noadinicb:
			uopt->flags &= ~(1 << UDF_FLAG_USE_AD_IN_ICB);
			break;
		case Opt_adinicb:
			uopt->flags |= (1 << UDF_FLAG_USE_AD_IN_ICB);
			break;
		case Opt_shortad:
			uopt->flags |= (1 << UDF_FLAG_USE_SHORT_AD);
			break;
		case Opt_longad:
			uopt->flags &= ~(1 << UDF_FLAG_USE_SHORT_AD);
			break;
		case Opt_gid:
			if (match_int(args, &option))
				return 0;
			uopt->gid = make_kgid(current_user_ns(), option);
			if (!gid_valid(uopt->gid))
				return 0;
			uopt->flags |= (1 << UDF_FLAG_GID_SET);
			break;
		case Opt_uid:
			if (match_int(args, &option))
				return 0;
			uopt->uid = make_kuid(current_user_ns(), option);
			if (!uid_valid(uopt->uid))
				return 0;
			uopt->flags |= (1 << UDF_FLAG_UID_SET);
			break;
		case Opt_umask:
			if (match_octal(args, &option))
				return 0;
			uopt->umask = option;
			break;
		case Opt_nostrict:
			uopt->flags &= ~(1 << UDF_FLAG_STRICT);
			break;
		case Opt_session:
			if (match_int(args, &option))
				return 0;
			uopt->session = option;
			if (!remount)
				uopt->flags |= (1 << UDF_FLAG_SESSION_SET);
			break;
		case Opt_lastblock:
			if (match_int(args, &option))
				return 0;
			uopt->lastblock = option;
			if (!remount)
				uopt->flags |= (1 << UDF_FLAG_LASTBLOCK_SET);
			break;
		case Opt_anchor:
			if (match_int(args, &option))
				return 0;
			uopt->anchor = option;
			break;
		case Opt_volume:
		case Opt_partition:
		case Opt_fileset:
		case Opt_rootdir:
			/* Ignored (never implemented properly) */
			break;
		case Opt_utf8:
			if (!remount) {
				unload_nls(uopt->nls_map);
				uopt->nls_map = NULL;
			}
			break;
		case Opt_iocharset:
			if (!remount) {
				unload_nls(uopt->nls_map);
				uopt->nls_map = NULL;
			}
			/* When nls_map is not loaded then UTF-8 is used */
			if (!remount && strcmp(args[0].from, "utf8") != 0) {
				uopt->nls_map = load_nls(args[0].from);
				if (!uopt->nls_map) {
					pr_err("iocharset %s not found\n",
						args[0].from);
					return 0;
				}
			}
			break;
		case Opt_uforget:
			uopt->flags |= (1 << UDF_FLAG_UID_FORGET);
			break;
		case Opt_uignore:
		case Opt_gignore:
			/* These options are superseeded by uid=<number> */
			break;
		case Opt_gforget:
			uopt->flags |= (1 << UDF_FLAG_GID_FORGET);
			break;
		case Opt_fmode:
			if (match_octal(args, &option))
				return 0;
			uopt->fmode = option & 0777;
			break;
		case Opt_dmode:
			if (match_octal(args, &option))
				return 0;
			uopt->dmode = option & 0777;
			break;
		default:
			pr_err("bad mount option \"%s\" or missing value\n", p);
			return 0;
		}
	}
	return 1;
}

static int udf_remount_fs(struct super_block *sb, int *flags, char *options)
{
	struct udf_options uopt;
	struct udf_sb_info *sbi = UDF_SB(sb);
	int error = 0;

	if (!(*flags & SB_RDONLY) && UDF_QUERY_FLAG(sb, UDF_FLAG_RW_INCOMPAT))
		return -EACCES;

	sync_filesystem(sb);

	uopt.flags = sbi->s_flags;
	uopt.uid   = sbi->s_uid;
	uopt.gid   = sbi->s_gid;
	uopt.umask = sbi->s_umask;
	uopt.fmode = sbi->s_fmode;
	uopt.dmode = sbi->s_dmode;
	uopt.nls_map = NULL;

	if (!udf_parse_options(options, &uopt, true))
		return -EINVAL;

	write_lock(&sbi->s_cred_lock);
	sbi->s_flags = uopt.flags;
	sbi->s_uid   = uopt.uid;
	sbi->s_gid   = uopt.gid;
	sbi->s_umask = uopt.umask;
	sbi->s_fmode = uopt.fmode;
	sbi->s_dmode = uopt.dmode;
	write_unlock(&sbi->s_cred_lock);

	if ((bool)(*flags & SB_RDONLY) == sb_rdonly(sb))
		goto out_unlock;

	if (*flags & SB_RDONLY)
		udf_close_lvid(sb);
	else
		udf_open_lvid(sb);

out_unlock:
	return error;
}

/*
 * Check VSD descriptor. Returns -1 in case we are at the end of volume
 * recognition area, 0 if the descriptor is valid but non-interesting, 1 if
 * we found one of NSR descriptors we are looking for.
 */
static int identify_vsd(const struct volStructDesc *vsd)
{
	int ret = 0;

	if (!memcmp(vsd->stdIdent, VSD_STD_ID_CD001, VSD_STD_ID_LEN)) {
		switch (vsd->structType) {
		case 0:
			udf_debug("ISO9660 Boot Record found\n");
			break;
		case 1:
			udf_debug("ISO9660 Primary Volume Descriptor found\n");
			break;
		case 2:
			udf_debug("ISO9660 Supplementary Volume Descriptor found\n");
			break;
		case 3:
			udf_debug("ISO9660 Volume Partition Descriptor found\n");
			break;
		case 255:
			udf_debug("ISO9660 Volume Descriptor Set Terminator found\n");
			break;
		default:
			udf_debug("ISO9660 VRS (%u) found\n", vsd->structType);
			break;
		}
	} else if (!memcmp(vsd->stdIdent, VSD_STD_ID_BEA01, VSD_STD_ID_LEN))
		; /* ret = 0 */
	else if (!memcmp(vsd->stdIdent, VSD_STD_ID_NSR02, VSD_STD_ID_LEN))
		ret = 1;
	else if (!memcmp(vsd->stdIdent, VSD_STD_ID_NSR03, VSD_STD_ID_LEN))
		ret = 1;
	else if (!memcmp(vsd->stdIdent, VSD_STD_ID_BOOT2, VSD_STD_ID_LEN))
		; /* ret = 0 */
	else if (!memcmp(vsd->stdIdent, VSD_STD_ID_CDW02, VSD_STD_ID_LEN))
		; /* ret = 0 */
	else {
		/* TEA01 or invalid id : end of volume recognition area */
		ret = -1;
	}

	return ret;
}

/*
 * Check Volume Structure Descriptors (ECMA 167 2/9.1)
 * We also check any "CD-ROM Volume Descriptor Set" (ECMA 167 2/8.3.1)
 * @return   1 if NSR02 or NSR03 found,
 *	    -1 if first sector read error, 0 otherwise
 */
static int udf_check_vsd(struct super_block *sb)
{
	struct volStructDesc *vsd = NULL;
	loff_t sector = VSD_FIRST_SECTOR_OFFSET;
	int sectorsize;
	struct buffer_head *bh = NULL;
	int nsr = 0;
	struct udf_sb_info *sbi;
	loff_t session_offset;

	sbi = UDF_SB(sb);
	if (sb->s_blocksize < sizeof(struct volStructDesc))
		sectorsize = sizeof(struct volStructDesc);
	else
		sectorsize = sb->s_blocksize;

	session_offset = (loff_t)sbi->s_session << sb->s_blocksize_bits;
	sector += session_offset;

	udf_debug("Starting at sector %u (%lu byte sectors)\n",
		  (unsigned int)(sector >> sb->s_blocksize_bits),
		  sb->s_blocksize);
	/* Process the sequence (if applicable). The hard limit on the sector
	 * offset is arbitrary, hopefully large enough so that all valid UDF
	 * filesystems will be recognised. There is no mention of an upper
	 * bound to the size of the volume recognition area in the standard.
	 *  The limit will prevent the code to read all the sectors of a
	 * specially crafted image (like a bluray disc full of CD001 sectors),
	 * potentially causing minutes or even hours of uninterruptible I/O
	 * activity. This actually happened with uninitialised SSD partitions
	 * (all 0xFF) before the check for the limit and all valid IDs were
	 * added */
	for (; !nsr && sector < VSD_MAX_SECTOR_OFFSET; sector += sectorsize) {
		/* Read a block */
		bh = udf_tread(sb, sector >> sb->s_blocksize_bits);
		if (!bh)
			break;

		vsd = (struct volStructDesc *)(bh->b_data +
					      (sector & (sb->s_blocksize - 1)));
		nsr = identify_vsd(vsd);
		/* Found NSR or end? */
		if (nsr) {
			brelse(bh);
			break;
		}
		/*
		 * Special handling for improperly formatted VRS (e.g., Win10)
		 * where components are separated by 2048 bytes even though
		 * sectors are 4K
		 */
		if (sb->s_blocksize == 4096) {
			nsr = identify_vsd(vsd + 1);
			/* Ignore unknown IDs... */
			if (nsr < 0)
				nsr = 0;
		}
		brelse(bh);
	}

	if (nsr > 0)
		return 1;
	else if (!bh && sector - session_offset == VSD_FIRST_SECTOR_OFFSET)
		return -1;
	else
		return 0;
}

static int udf_verify_domain_identifier(struct super_block *sb,
					struct regid *ident, char *dname)
{
	struct domainEntityIDSuffix *suffix;

	if (memcmp(ident->ident, UDF_ID_COMPLIANT, strlen(UDF_ID_COMPLIANT))) {
		udf_warn(sb, "Not OSTA UDF compliant %s descriptor.\n", dname);
		goto force_ro;
	}
	if (ident->flags & (1 << ENTITYID_FLAGS_DIRTY)) {
		udf_warn(sb, "Possibly not OSTA UDF compliant %s descriptor.\n",
			 dname);
		goto force_ro;
	}
	suffix = (struct domainEntityIDSuffix *)ident->identSuffix;
	if (suffix->flags & (1 << ENTITYIDSUFFIX_FLAGS_HARDWRITEPROTECT) ||
	    suffix->flags & (1 << ENTITYIDSUFFIX_FLAGS_SOFTWRITEPROTECT)) {
		if (!sb_rdonly(sb)) {
			udf_warn(sb, "Descriptor for %s marked write protected."
				 " Forcing read only mount.\n", dname);
		}
		goto force_ro;
	}
	return 0;

force_ro:
	if (!sb_rdonly(sb))
		return -EACCES;
	UDF_SET_FLAG(sb, UDF_FLAG_RW_INCOMPAT);
	return 0;
}

static int udf_load_fileset(struct super_block *sb, struct fileSetDesc *fset,
			    struct kernel_lb_addr *root)
{
	int ret;

	ret = udf_verify_domain_identifier(sb, &fset->domainIdent, "file set");
	if (ret < 0)
		return ret;

	*root = lelb_to_cpu(fset->rootDirectoryICB.extLocation);
	UDF_SB(sb)->s_serial_number = le16_to_cpu(fset->descTag.tagSerialNum);

	udf_debug("Rootdir at block=%u, partition=%u\n",
		  root->logicalBlockNum, root->partitionReferenceNum);
	return 0;
}

static int udf_find_fileset(struct super_block *sb,
			    struct kernel_lb_addr *fileset,
			    struct kernel_lb_addr *root)
{
	struct buffer_head *bh = NULL;
	uint16_t ident;
	int ret;

	if (fileset->logicalBlockNum == 0xFFFFFFFF &&
	    fileset->partitionReferenceNum == 0xFFFF)
		return -EINVAL;

	bh = udf_read_ptagged(sb, fileset, 0, &ident);
	if (!bh)
		return -EIO;
	if (ident != TAG_IDENT_FSD) {
		brelse(bh);
		return -EINVAL;
	}

	udf_debug("Fileset at block=%u, partition=%u\n",
		  fileset->logicalBlockNum, fileset->partitionReferenceNum);

	UDF_SB(sb)->s_partition = fileset->partitionReferenceNum;
	ret = udf_load_fileset(sb, (struct fileSetDesc *)bh->b_data, root);
	brelse(bh);
	return ret;
}

/*
 * Load primary Volume Descriptor Sequence
 *
 * Return <0 on error, 0 on success. -EAGAIN is special meaning next sequence
 * should be tried.
 */
static int udf_load_pvoldesc(struct super_block *sb, sector_t block)
{
	struct primaryVolDesc *pvoldesc;
	uint8_t *outstr;
	struct buffer_head *bh;
	uint16_t ident;
	int ret = -ENOMEM;
	struct timestamp *ts;

	outstr = kmalloc(128, GFP_NOFS);
	if (!outstr)
		return -ENOMEM;

	bh = udf_read_tagged(sb, block, block, &ident);
	if (!bh) {
		ret = -EAGAIN;
		goto out2;
	}

	if (ident != TAG_IDENT_PVD) {
		ret = -EIO;
		goto out_bh;
	}

	pvoldesc = (struct primaryVolDesc *)bh->b_data;

	udf_disk_stamp_to_time(&UDF_SB(sb)->s_record_time,
			      pvoldesc->recordingDateAndTime);
	ts = &pvoldesc->recordingDateAndTime;
	udf_debug("recording time %04u/%02u/%02u %02u:%02u (%x)\n",
		  le16_to_cpu(ts->year), ts->month, ts->day, ts->hour,
		  ts->minute, le16_to_cpu(ts->typeAndTimezone));

	ret = udf_dstrCS0toChar(sb, outstr, 31, pvoldesc->volIdent, 32);
	if (ret < 0) {
		strcpy(UDF_SB(sb)->s_volume_ident, "InvalidName");
		pr_warn("incorrect volume identification, setting to "
			"'InvalidName'\n");
	} else {
		strncpy(UDF_SB(sb)->s_volume_ident, outstr, ret);
	}
	udf_debug("volIdent[] = '%s'\n", UDF_SB(sb)->s_volume_ident);

	ret = udf_dstrCS0toChar(sb, outstr, 127, pvoldesc->volSetIdent, 128);
	if (ret < 0) {
		ret = 0;
		goto out_bh;
	}
	outstr[ret] = 0;
	udf_debug("volSetIdent[] = '%s'\n", outstr);

	ret = 0;
out_bh:
	brelse(bh);
out2:
	kfree(outstr);
	return ret;
}

struct inode *udf_find_metadata_inode_efe(struct super_block *sb,
					u32 meta_file_loc, u32 partition_ref)
{
	struct kernel_lb_addr addr;
	struct inode *metadata_fe;

	addr.logicalBlockNum = meta_file_loc;
	addr.partitionReferenceNum = partition_ref;

	metadata_fe = udf_iget_special(sb, &addr);

	if (IS_ERR(metadata_fe)) {
		udf_warn(sb, "metadata inode efe not found\n");
		return metadata_fe;
	}
	if (UDF_I(metadata_fe)->i_alloc_type != ICBTAG_FLAG_AD_SHORT) {
		udf_warn(sb, "metadata inode efe does not have short allocation descriptors!\n");
		iput(metadata_fe);
		return ERR_PTR(-EIO);
	}

	return metadata_fe;
}

static int udf_load_metadata_files(struct super_block *sb, int partition,
				   int type1_index)
{
	struct udf_sb_info *sbi = UDF_SB(sb);
	struct udf_part_map *map;
	struct udf_meta_data *mdata;
	struct kernel_lb_addr addr;
	struct inode *fe;

	map = &sbi->s_partmaps[partition];
	mdata = &map->s_type_specific.s_metadata;
	mdata->s_phys_partition_ref = type1_index;

	/* metadata address */
	udf_debug("Metadata file location: block = %u part = %u\n",
		  mdata->s_meta_file_loc, mdata->s_phys_partition_ref);

	fe = udf_find_metadata_inode_efe(sb, mdata->s_meta_file_loc,
					 mdata->s_phys_partition_ref);
	if (IS_ERR(fe)) {
		/* mirror file entry */
		udf_debug("Mirror metadata file location: block = %u part = %u\n",
			  mdata->s_mirror_file_loc, mdata->s_phys_partition_ref);

		fe = udf_find_metadata_inode_efe(sb, mdata->s_mirror_file_loc,
						 mdata->s_phys_partition_ref);

		if (IS_ERR(fe)) {
			udf_err(sb, "Both metadata and mirror metadata inode efe can not found\n");
			return PTR_ERR(fe);
		}
		mdata->s_mirror_fe = fe;
	} else
		mdata->s_metadata_fe = fe;


	/*
	 * bitmap file entry
	 * Note:
	 * Load only if bitmap file location differs from 0xFFFFFFFF (DCN-5102)
	*/
	if (mdata->s_bitmap_file_loc != 0xFFFFFFFF) {
		addr.logicalBlockNum = mdata->s_bitmap_file_loc;
		addr.partitionReferenceNum = mdata->s_phys_partition_ref;

		udf_debug("Bitmap file location: block = %u part = %u\n",
			  addr.logicalBlockNum, addr.partitionReferenceNum);

		fe = udf_iget_special(sb, &addr);
		if (IS_ERR(fe)) {
			if (sb_rdonly(sb))
				udf_warn(sb, "bitmap inode efe not found but it's ok since the disc is mounted read-only\n");
			else {
				udf_err(sb, "bitmap inode efe not found and attempted read-write mount\n");
				return PTR_ERR(fe);
			}
		} else
			mdata->s_bitmap_fe = fe;
	}

	udf_debug("udf_load_metadata_files Ok\n");
	return 0;
}

int udf_compute_nr_groups(struct super_block *sb, u32 partition)
{
	struct udf_part_map *map = &UDF_SB(sb)->s_partmaps[partition];
	return DIV_ROUND_UP(map->s_partition_len +
			    (sizeof(struct spaceBitmapDesc) << 3),
			    sb->s_blocksize * 8);
}

static struct udf_bitmap *udf_sb_alloc_bitmap(struct super_block *sb, u32 index)
{
	struct udf_bitmap *bitmap;
	int nr_groups;
	int size;

	nr_groups = udf_compute_nr_groups(sb, index);
	size = sizeof(struct udf_bitmap) +
		(sizeof(struct buffer_head *) * nr_groups);

	if (size <= PAGE_SIZE)
		bitmap = kzalloc(size, GFP_KERNEL);
	else
		bitmap = vzalloc(size); /* TODO: get rid of vzalloc */

	if (!bitmap)
		return NULL;

	bitmap->s_nr_groups = nr_groups;
	return bitmap;
}

static int check_partition_desc(struct super_block *sb,
				struct partitionDesc *p,
				struct udf_part_map *map)
{
	bool umap, utable, fmap, ftable;
	struct partitionHeaderDesc *phd;

	switch (le32_to_cpu(p->accessType)) {
	case PD_ACCESS_TYPE_READ_ONLY:
	case PD_ACCESS_TYPE_WRITE_ONCE:
	case PD_ACCESS_TYPE_NONE:
		goto force_ro;
	}

	/* No Partition Header Descriptor? */
	if (strcmp(p->partitionContents.ident, PD_PARTITION_CONTENTS_NSR02) &&
	    strcmp(p->partitionContents.ident, PD_PARTITION_CONTENTS_NSR03))
		goto force_ro;

	phd = (struct partitionHeaderDesc *)p->partitionContentsUse;
	utable = phd->unallocSpaceTable.extLength;
	umap = phd->unallocSpaceBitmap.extLength;
	ftable = phd->freedSpaceTable.extLength;
	fmap = phd->freedSpaceBitmap.extLength;

	/* No allocation info? */
	if (!utable && !umap && !ftable && !fmap)
		goto force_ro;

	/* We don't support blocks that require erasing before overwrite */
	if (ftable || fmap)
		goto force_ro;
	/* UDF 2.60: 2.3.3 - no mixing of tables & bitmaps, no VAT. */
	if (utable && umap)
		goto force_ro;

	if (map->s_partition_type == UDF_VIRTUAL_MAP15 ||
	    map->s_partition_type == UDF_VIRTUAL_MAP20)
		goto force_ro;

	return 0;
force_ro:
	if (!sb_rdonly(sb))
		return -EACCES;
	UDF_SET_FLAG(sb, UDF_FLAG_RW_INCOMPAT);
	return 0;
}

static int udf_fill_partdesc_info(struct super_block *sb,
		struct partitionDesc *p, int p_index)
{
	struct udf_part_map *map;
	struct udf_sb_info *sbi = UDF_SB(sb);
	struct partitionHeaderDesc *phd;
	int err;

	map = &sbi->s_partmaps[p_index];

	map->s_partition_len = le32_to_cpu(p->partitionLength); /* blocks */
	map->s_partition_root = le32_to_cpu(p->partitionStartingLocation);

	if (p->accessType == cpu_to_le32(PD_ACCESS_TYPE_READ_ONLY))
		map->s_partition_flags |= UDF_PART_FLAG_READ_ONLY;
	if (p->accessType == cpu_to_le32(PD_ACCESS_TYPE_WRITE_ONCE))
		map->s_partition_flags |= UDF_PART_FLAG_WRITE_ONCE;
	if (p->accessType == cpu_to_le32(PD_ACCESS_TYPE_REWRITABLE))
		map->s_partition_flags |= UDF_PART_FLAG_REWRITABLE;
	if (p->accessType == cpu_to_le32(PD_ACCESS_TYPE_OVERWRITABLE))
		map->s_partition_flags |= UDF_PART_FLAG_OVERWRITABLE;

	udf_debug("Partition (%d type %x) starts at physical %u, block length %u\n",
		  p_index, map->s_partition_type,
		  map->s_partition_root, map->s_partition_len);

	err = check_partition_desc(sb, p, map);
	if (err)
		return err;

	/*
	 * Skip loading allocation info it we cannot ever write to the fs.
	 * This is a correctness thing as we may have decided to force ro mount
	 * to avoid allocation info we don't support.
	 */
	if (UDF_QUERY_FLAG(sb, UDF_FLAG_RW_INCOMPAT))
		return 0;

	phd = (struct partitionHeaderDesc *)p->partitionContentsUse;
	if (phd->unallocSpaceTable.extLength) {
		struct kernel_lb_addr loc = {
			.logicalBlockNum = le32_to_cpu(
				phd->unallocSpaceTable.extPosition),
			.partitionReferenceNum = p_index,
		};
		struct inode *inode;

		inode = udf_iget_special(sb, &loc);
		if (IS_ERR(inode)) {
			udf_debug("cannot load unallocSpaceTable (part %d)\n",
				  p_index);
			return PTR_ERR(inode);
		}
		map->s_uspace.s_table = inode;
		map->s_partition_flags |= UDF_PART_FLAG_UNALLOC_TABLE;
		udf_debug("unallocSpaceTable (part %d) @ %lu\n",
			  p_index, map->s_uspace.s_table->i_ino);
	}

	if (phd->unallocSpaceBitmap.extLength) {
		struct udf_bitmap *bitmap = udf_sb_alloc_bitmap(sb, p_index);
		if (!bitmap)
			return -ENOMEM;
		map->s_uspace.s_bitmap = bitmap;
		bitmap->s_extPosition = le32_to_cpu(
				phd->unallocSpaceBitmap.extPosition);
		map->s_partition_flags |= UDF_PART_FLAG_UNALLOC_BITMAP;
		udf_debug("unallocSpaceBitmap (part %d) @ %u\n",
			  p_index, bitmap->s_extPosition);
	}

	return 0;
}

static void udf_find_vat_block(struct super_block *sb, int p_index,
			       int type1_index, sector_t start_block)
{
	struct udf_sb_info *sbi = UDF_SB(sb);
	struct udf_part_map *map = &sbi->s_partmaps[p_index];
	sector_t vat_block;
	struct kernel_lb_addr ino;
	struct inode *inode;

	/*
	 * VAT file entry is in the last recorded block. Some broken disks have
	 * it a few blocks before so try a bit harder...
	 */
	ino.partitionReferenceNum = type1_index;
	for (vat_block = start_block;
	     vat_block >= map->s_partition_root &&
	     vat_block >= start_block - 3; vat_block--) {
		ino.logicalBlockNum = vat_block - map->s_partition_root;
		inode = udf_iget_special(sb, &ino);
		if (!IS_ERR(inode)) {
			sbi->s_vat_inode = inode;
			break;
		}
	}
}

static int udf_load_vat(struct super_block *sb, int p_index, int type1_index)
{
	struct udf_sb_info *sbi = UDF_SB(sb);
	struct udf_part_map *map = &sbi->s_partmaps[p_index];
	struct buffer_head *bh = NULL;
	struct udf_inode_info *vati;
	uint32_t pos;
	struct virtualAllocationTable20 *vat20;
	sector_t blocks = i_size_read(sb->s_bdev->bd_inode) >>
			  sb->s_blocksize_bits;

	udf_find_vat_block(sb, p_index, type1_index, sbi->s_last_block);
	if (!sbi->s_vat_inode &&
	    sbi->s_last_block != blocks - 1) {
		pr_notice("Failed to read VAT inode from the last recorded block (%lu), retrying with the last block of the device (%lu).\n",
			  (unsigned long)sbi->s_last_block,
			  (unsigned long)blocks - 1);
		udf_find_vat_block(sb, p_index, type1_index, blocks - 1);
	}
	if (!sbi->s_vat_inode)
		return -EIO;

	if (map->s_partition_type == UDF_VIRTUAL_MAP15) {
		map->s_type_specific.s_virtual.s_start_offset = 0;
		map->s_type_specific.s_virtual.s_num_entries =
			(sbi->s_vat_inode->i_size - 36) >> 2;
	} else if (map->s_partition_type == UDF_VIRTUAL_MAP20) {
		vati = UDF_I(sbi->s_vat_inode);
		if (vati->i_alloc_type != ICBTAG_FLAG_AD_IN_ICB) {
			pos = udf_block_map(sbi->s_vat_inode, 0);
			bh = sb_bread(sb, pos);
			if (!bh)
				return -EIO;
			vat20 = (struct virtualAllocationTable20 *)bh->b_data;
		} else {
			vat20 = (struct virtualAllocationTable20 *)
							vati->i_ext.i_data;
		}

		map->s_type_specific.s_virtual.s_start_offset =
			le16_to_cpu(vat20->lengthHeader);
		map->s_type_specific.s_virtual.s_num_entries =
			(sbi->s_vat_inode->i_size -
				map->s_type_specific.s_virtual.
					s_start_offset) >> 2;
		brelse(bh);
	}
	return 0;
}

/*
 * Load partition descriptor block
 *
 * Returns <0 on error, 0 on success, -EAGAIN is special - try next descriptor
 * sequence.
 */
static int udf_load_partdesc(struct super_block *sb, sector_t block)
{
	struct buffer_head *bh;
	struct partitionDesc *p;
	struct udf_part_map *map;
	struct udf_sb_info *sbi = UDF_SB(sb);
	int i, type1_idx;
	uint16_t partitionNumber;
	uint16_t ident;
	int ret;

	bh = udf_read_tagged(sb, block, block, &ident);
	if (!bh)
		return -EAGAIN;
	if (ident != TAG_IDENT_PD) {
		ret = 0;
		goto out_bh;
	}

	p = (struct partitionDesc *)bh->b_data;
	partitionNumber = le16_to_cpu(p->partitionNumber);

	/* First scan for TYPE1 and SPARABLE partitions */
	for (i = 0; i < sbi->s_partitions; i++) {
		map = &sbi->s_partmaps[i];
		udf_debug("Searching map: (%u == %u)\n",
			  map->s_partition_num, partitionNumber);
		if (map->s_partition_num == partitionNumber &&
		    (map->s_partition_type == UDF_TYPE1_MAP15 ||
		     map->s_partition_type == UDF_SPARABLE_MAP15))
			break;
	}

	if (i >= sbi->s_partitions) {
		udf_debug("Partition (%u) not found in partition map\n",
			  partitionNumber);
		ret = 0;
		goto out_bh;
	}

	ret = udf_fill_partdesc_info(sb, p, i);
	if (ret < 0)
		goto out_bh;

	/*
	 * Now rescan for VIRTUAL or METADATA partitions when SPARABLE and
	 * PHYSICAL partitions are already set up
	 */
	type1_idx = i;
	map = NULL; /* supress 'maybe used uninitialized' warning */
	for (i = 0; i < sbi->s_partitions; i++) {
		map = &sbi->s_partmaps[i];

		if (map->s_partition_num == partitionNumber &&
		    (map->s_partition_type == UDF_VIRTUAL_MAP15 ||
		     map->s_partition_type == UDF_VIRTUAL_MAP20 ||
		     map->s_partition_type == UDF_METADATA_MAP25))
			break;
	}

	if (i >= sbi->s_partitions) {
		ret = 0;
		goto out_bh;
	}

	ret = udf_fill_partdesc_info(sb, p, i);
	if (ret < 0)
		goto out_bh;

	if (map->s_partition_type == UDF_METADATA_MAP25) {
		ret = udf_load_metadata_files(sb, i, type1_idx);
		if (ret < 0) {
			udf_err(sb, "error loading MetaData partition map %d\n",
				i);
			goto out_bh;
		}
	} else {
		/*
		 * If we have a partition with virtual map, we don't handle
		 * writing to it (we overwrite blocks instead of relocating
		 * them).
		 */
		if (!sb_rdonly(sb)) {
			ret = -EACCES;
			goto out_bh;
		}
		UDF_SET_FLAG(sb, UDF_FLAG_RW_INCOMPAT);
		ret = udf_load_vat(sb, i, type1_idx);
		if (ret < 0)
			goto out_bh;
	}
	ret = 0;
out_bh:
	/* In case loading failed, we handle cleanup in udf_fill_super */
	brelse(bh);
	return ret;
}

static int udf_load_sparable_map(struct super_block *sb,
				 struct udf_part_map *map,
				 struct sparablePartitionMap *spm)
{
	uint32_t loc;
	uint16_t ident;
	struct sparingTable *st;
	struct udf_sparing_data *sdata = &map->s_type_specific.s_sparing;
	int i;
	struct buffer_head *bh;

	map->s_partition_type = UDF_SPARABLE_MAP15;
	sdata->s_packet_len = le16_to_cpu(spm->packetLength);
	if (!is_power_of_2(sdata->s_packet_len)) {
		udf_err(sb, "error loading logical volume descriptor: "
			"Invalid packet length %u\n",
			(unsigned)sdata->s_packet_len);
		return -EIO;
	}
	if (spm->numSparingTables > 4) {
		udf_err(sb, "error loading logical volume descriptor: "
			"Too many sparing tables (%d)\n",
			(int)spm->numSparingTables);
		return -EIO;
	}
	if (le32_to_cpu(spm->sizeSparingTable) > sb->s_blocksize) {
		udf_err(sb, "error loading logical volume descriptor: "
			"Too big sparing table size (%u)\n",
			le32_to_cpu(spm->sizeSparingTable));
		return -EIO;
	}

	for (i = 0; i < spm->numSparingTables; i++) {
		loc = le32_to_cpu(spm->locSparingTable[i]);
		bh = udf_read_tagged(sb, loc, loc, &ident);
		if (!bh)
			continue;

		st = (struct sparingTable *)bh->b_data;
		if (ident != 0 ||
		    strncmp(st->sparingIdent.ident, UDF_ID_SPARING,
			    strlen(UDF_ID_SPARING)) ||
		    sizeof(*st) + le16_to_cpu(st->reallocationTableLen) >
							sb->s_blocksize) {
			brelse(bh);
			continue;
		}

		sdata->s_spar_map[i] = bh;
	}
	map->s_partition_func = udf_get_pblock_spar15;
	return 0;
}

static int udf_load_logicalvol(struct super_block *sb, sector_t block,
			       struct kernel_lb_addr *fileset)
{
	struct logicalVolDesc *lvd;
	int i, offset;
	uint8_t type;
	struct udf_sb_info *sbi = UDF_SB(sb);
	struct genericPartitionMap *gpm;
	uint16_t ident;
	struct buffer_head *bh;
	unsigned int table_len;
	int ret;

	bh = udf_read_tagged(sb, block, block, &ident);
	if (!bh)
		return -EAGAIN;
	BUG_ON(ident != TAG_IDENT_LVD);
	lvd = (struct logicalVolDesc *)bh->b_data;
	table_len = le32_to_cpu(lvd->mapTableLength);
	if (table_len > sb->s_blocksize - sizeof(*lvd)) {
		udf_err(sb, "error loading logical volume descriptor: "
			"Partition table too long (%u > %lu)\n", table_len,
			sb->s_blocksize - sizeof(*lvd));
		ret = -EIO;
		goto out_bh;
	}

	ret = udf_verify_domain_identifier(sb, &lvd->domainIdent,
					   "logical volume");
	if (ret)
		goto out_bh;
	ret = udf_sb_alloc_partition_maps(sb, le32_to_cpu(lvd->numPartitionMaps));
	if (ret)
		goto out_bh;

	for (i = 0, offset = 0;
	     i < sbi->s_partitions && offset < table_len;
	     i++, offset += gpm->partitionMapLength) {
		struct udf_part_map *map = &sbi->s_partmaps[i];
		gpm = (struct genericPartitionMap *)
				&(lvd->partitionMaps[offset]);
		type = gpm->partitionMapType;
		if (type == 1) {
			struct genericPartitionMap1 *gpm1 =
				(struct genericPartitionMap1 *)gpm;
			map->s_partition_type = UDF_TYPE1_MAP15;
			map->s_volumeseqnum = le16_to_cpu(gpm1->volSeqNum);
			map->s_partition_num = le16_to_cpu(gpm1->partitionNum);
			map->s_partition_func = NULL;
		} else if (type == 2) {
			struct udfPartitionMap2 *upm2 =
						(struct udfPartitionMap2 *)gpm;
			if (!strncmp(upm2->partIdent.ident, UDF_ID_VIRTUAL,
						strlen(UDF_ID_VIRTUAL))) {
				u16 suf =
					le16_to_cpu(((__le16 *)upm2->partIdent.
							identSuffix)[0]);
				if (suf < 0x0200) {
					map->s_partition_type =
							UDF_VIRTUAL_MAP15;
					map->s_partition_func =
							udf_get_pblock_virt15;
				} else {
					map->s_partition_type =
							UDF_VIRTUAL_MAP20;
					map->s_partition_func =
							udf_get_pblock_virt20;
				}
			} else if (!strncmp(upm2->partIdent.ident,
						UDF_ID_SPARABLE,
						strlen(UDF_ID_SPARABLE))) {
				ret = udf_load_sparable_map(sb, map,
					(struct sparablePartitionMap *)gpm);
				if (ret < 0)
					goto out_bh;
			} else if (!strncmp(upm2->partIdent.ident,
						UDF_ID_METADATA,
						strlen(UDF_ID_METADATA))) {
				struct udf_meta_data *mdata =
					&map->s_type_specific.s_metadata;
				struct metadataPartitionMap *mdm =
						(struct metadataPartitionMap *)
						&(lvd->partitionMaps[offset]);
				udf_debug("Parsing Logical vol part %d type %u  id=%s\n",
					  i, type, UDF_ID_METADATA);

				map->s_partition_type = UDF_METADATA_MAP25;
				map->s_partition_func = udf_get_pblock_meta25;

				mdata->s_meta_file_loc   =
					le32_to_cpu(mdm->metadataFileLoc);
				mdata->s_mirror_file_loc =
					le32_to_cpu(mdm->metadataMirrorFileLoc);
				mdata->s_bitmap_file_loc =
					le32_to_cpu(mdm->metadataBitmapFileLoc);
				mdata->s_alloc_unit_size =
					le32_to_cpu(mdm->allocUnitSize);
				mdata->s_align_unit_size =
					le16_to_cpu(mdm->alignUnitSize);
				if (mdm->flags & 0x01)
					mdata->s_flags |= MF_DUPLICATE_MD;

				udf_debug("Metadata Ident suffix=0x%x\n",
					  le16_to_cpu(*(__le16 *)
						      mdm->partIdent.identSuffix));
				udf_debug("Metadata part num=%u\n",
					  le16_to_cpu(mdm->partitionNum));
				udf_debug("Metadata part alloc unit size=%u\n",
					  le32_to_cpu(mdm->allocUnitSize));
				udf_debug("Metadata file loc=%u\n",
					  le32_to_cpu(mdm->metadataFileLoc));
				udf_debug("Mirror file loc=%u\n",
					  le32_to_cpu(mdm->metadataMirrorFileLoc));
				udf_debug("Bitmap file loc=%u\n",
					  le32_to_cpu(mdm->metadataBitmapFileLoc));
				udf_debug("Flags: %d %u\n",
					  mdata->s_flags, mdm->flags);
			} else {
				udf_debug("Unknown ident: %s\n",
					  upm2->partIdent.ident);
				continue;
			}
			map->s_volumeseqnum = le16_to_cpu(upm2->volSeqNum);
			map->s_partition_num = le16_to_cpu(upm2->partitionNum);
		}
		udf_debug("Partition (%d:%u) type %u on volume %u\n",
			  i, map->s_partition_num, type, map->s_volumeseqnum);
	}

	if (fileset) {
		struct long_ad *la = (struct long_ad *)&(lvd->logicalVolContentsUse[0]);

		*fileset = lelb_to_cpu(la->extLocation);
		udf_debug("FileSet found in LogicalVolDesc at block=%u, partition=%u\n",
			  fileset->logicalBlockNum,
			  fileset->partitionReferenceNum);
	}
	if (lvd->integritySeqExt.extLength)
		udf_load_logicalvolint(sb, leea_to_cpu(lvd->integritySeqExt));
	ret = 0;

	if (!sbi->s_lvid_bh) {
		/* We can't generate unique IDs without a valid LVID */
		if (sb_rdonly(sb)) {
			UDF_SET_FLAG(sb, UDF_FLAG_RW_INCOMPAT);
		} else {
			udf_warn(sb, "Damaged or missing LVID, forcing "
				     "readonly mount\n");
			ret = -EACCES;
		}
	}
out_bh:
	brelse(bh);
	return ret;
}

/*
 * Find the prevailing Logical Volume Integrity Descriptor.
 */
static void udf_load_logicalvolint(struct super_block *sb, struct kernel_extent_ad loc)
{
	struct buffer_head *bh, *final_bh;
	uint16_t ident;
	struct udf_sb_info *sbi = UDF_SB(sb);
	struct logicalVolIntegrityDesc *lvid;
	int indirections = 0;
	u32 parts, impuselen;

	while (++indirections <= UDF_MAX_LVID_NESTING) {
		final_bh = NULL;
		while (loc.extLength > 0 &&
			(bh = udf_read_tagged(sb, loc.extLocation,
					loc.extLocation, &ident))) {
			if (ident != TAG_IDENT_LVID) {
				brelse(bh);
				break;
			}

			brelse(final_bh);
			final_bh = bh;

			loc.extLength -= sb->s_blocksize;
			loc.extLocation++;
		}

		if (!final_bh)
			return;

		brelse(sbi->s_lvid_bh);
		sbi->s_lvid_bh = final_bh;

		lvid = (struct logicalVolIntegrityDesc *)final_bh->b_data;
		if (lvid->nextIntegrityExt.extLength == 0)
			goto check;

		loc = leea_to_cpu(lvid->nextIntegrityExt);
	}

	udf_warn(sb, "Too many LVID indirections (max %u), ignoring.\n",
		UDF_MAX_LVID_NESTING);
out_err:
	brelse(sbi->s_lvid_bh);
	sbi->s_lvid_bh = NULL;
	return;
check:
	parts = le32_to_cpu(lvid->numOfPartitions);
	impuselen = le32_to_cpu(lvid->lengthOfImpUse);
	if (parts >= sb->s_blocksize || impuselen >= sb->s_blocksize ||
	    sizeof(struct logicalVolIntegrityDesc) + impuselen +
	    2 * parts * sizeof(u32) > sb->s_blocksize) {
		udf_warn(sb, "Corrupted LVID (parts=%u, impuselen=%u), "
			 "ignoring.\n", parts, impuselen);
		goto out_err;
	}
}

/*
 * Step for reallocation of table of partition descriptor sequence numbers.
 * Must be power of 2.
 */
#define PART_DESC_ALLOC_STEP 32

struct part_desc_seq_scan_data {
	struct udf_vds_record rec;
	u32 partnum;
};

struct desc_seq_scan_data {
	struct udf_vds_record vds[VDS_POS_LENGTH];
	unsigned int size_part_descs;
	unsigned int num_part_descs;
	struct part_desc_seq_scan_data *part_descs_loc;
};

static struct udf_vds_record *handle_partition_descriptor(
				struct buffer_head *bh,
				struct desc_seq_scan_data *data)
{
	struct partitionDesc *desc = (struct partitionDesc *)bh->b_data;
	int partnum;
	int i;

	partnum = le16_to_cpu(desc->partitionNumber);
	for (i = 0; i < data->num_part_descs; i++)
		if (partnum == data->part_descs_loc[i].partnum)
			return &(data->part_descs_loc[i].rec);
	if (data->num_part_descs >= data->size_part_descs) {
		struct part_desc_seq_scan_data *new_loc;
		unsigned int new_size = ALIGN(partnum, PART_DESC_ALLOC_STEP);

		new_loc = kcalloc(new_size, sizeof(*new_loc), GFP_KERNEL);
		if (!new_loc)
			return ERR_PTR(-ENOMEM);
		memcpy(new_loc, data->part_descs_loc,
		       data->size_part_descs * sizeof(*new_loc));
		kfree(data->part_descs_loc);
		data->part_descs_loc = new_loc;
		data->size_part_descs = new_size;
	}
	return &(data->part_descs_loc[data->num_part_descs++].rec);
}


static struct udf_vds_record *get_volume_descriptor_record(uint16_t ident,
		struct buffer_head *bh, struct desc_seq_scan_data *data)
{
	switch (ident) {
	case TAG_IDENT_PVD: /* ISO 13346 3/10.1 */
		return &(data->vds[VDS_POS_PRIMARY_VOL_DESC]);
	case TAG_IDENT_IUVD: /* ISO 13346 3/10.4 */
		return &(data->vds[VDS_POS_IMP_USE_VOL_DESC]);
	case TAG_IDENT_LVD: /* ISO 13346 3/10.6 */
		return &(data->vds[VDS_POS_LOGICAL_VOL_DESC]);
	case TAG_IDENT_USD: /* ISO 13346 3/10.8 */
		return &(data->vds[VDS_POS_UNALLOC_SPACE_DESC]);
	case TAG_IDENT_PD: /* ISO 13346 3/10.5 */
		return handle_partition_descriptor(bh, data);
	}
	return NULL;
}

/*
 * Process a main/reserve volume descriptor sequence.
 *   @block		First block of first extent of the sequence.
 *   @lastblock		Lastblock of first extent of the sequence.
 *   @fileset		There we store extent containing root fileset
 *
 * Returns <0 on error, 0 on success. -EAGAIN is special - try next descriptor
 * sequence
 */
static noinline int udf_process_sequence(
		struct super_block *sb,
		sector_t block, sector_t lastblock,
		struct kernel_lb_addr *fileset)
{
	struct buffer_head *bh = NULL;
	struct udf_vds_record *curr;
	struct generic_desc *gd;
	struct volDescPtr *vdp;
	bool done = false;
	uint32_t vdsn;
	uint16_t ident;
	int ret;
	unsigned int indirections = 0;
	struct desc_seq_scan_data data;
	unsigned int i;

	memset(data.vds, 0, sizeof(struct udf_vds_record) * VDS_POS_LENGTH);
	data.size_part_descs = PART_DESC_ALLOC_STEP;
	data.num_part_descs = 0;
	data.part_descs_loc = kcalloc(data.size_part_descs,
				      sizeof(*data.part_descs_loc),
				      GFP_KERNEL);
	if (!data.part_descs_loc)
		return -ENOMEM;

	/*
	 * Read the main descriptor sequence and find which descriptors
	 * are in it.
	 */
	for (; (!done && block <= lastblock); block++) {
		bh = udf_read_tagged(sb, block, block, &ident);
		if (!bh)
			break;

		/* Process each descriptor (ISO 13346 3/8.3-8.4) */
		gd = (struct generic_desc *)bh->b_data;
		vdsn = le32_to_cpu(gd->volDescSeqNum);
		switch (ident) {
		case TAG_IDENT_VDP: /* ISO 13346 3/10.3 */
			if (++indirections > UDF_MAX_TD_NESTING) {
				udf_err(sb, "too many Volume Descriptor "
					"Pointers (max %u supported)\n",
					UDF_MAX_TD_NESTING);
				brelse(bh);
				ret = -EIO;
				goto out;
			}

			vdp = (struct volDescPtr *)bh->b_data;
			block = le32_to_cpu(vdp->nextVolDescSeqExt.extLocation);
			lastblock = le32_to_cpu(
				vdp->nextVolDescSeqExt.extLength) >>
				sb->s_blocksize_bits;
			lastblock += block - 1;
			/* For loop is going to increment 'block' again */
			block--;
			break;
		case TAG_IDENT_PVD: /* ISO 13346 3/10.1 */
		case TAG_IDENT_IUVD: /* ISO 13346 3/10.4 */
		case TAG_IDENT_LVD: /* ISO 13346 3/10.6 */
		case TAG_IDENT_USD: /* ISO 13346 3/10.8 */
		case TAG_IDENT_PD: /* ISO 13346 3/10.5 */
			curr = get_volume_descriptor_record(ident, bh, &data);
			if (IS_ERR(curr)) {
				brelse(bh);
				ret = PTR_ERR(curr);
				goto out;
			}
			/* Descriptor we don't care about? */
			if (!curr)
				break;
			if (vdsn >= curr->volDescSeqNum) {
				curr->volDescSeqNum = vdsn;
				curr->block = block;
			}
			break;
		case TAG_IDENT_TD: /* ISO 13346 3/10.9 */
			done = true;
			break;
		}
		brelse(bh);
	}
	/*
	 * Now read interesting descriptors again and process them
	 * in a suitable order
	 */
	if (!data.vds[VDS_POS_PRIMARY_VOL_DESC].block) {
		udf_err(sb, "Primary Volume Descriptor not found!\n");
		ret = -EAGAIN;
		goto out;
	}
	ret = udf_load_pvoldesc(sb, data.vds[VDS_POS_PRIMARY_VOL_DESC].block);
	if (ret < 0)
		goto out;

	if (data.vds[VDS_POS_LOGICAL_VOL_DESC].block) {
		ret = udf_load_logicalvol(sb,
				data.vds[VDS_POS_LOGICAL_VOL_DESC].block,
				fileset);
		if (ret < 0)
			goto out;
	}

	/* Now handle prevailing Partition Descriptors */
	for (i = 0; i < data.num_part_descs; i++) {
		ret = udf_load_partdesc(sb, data.part_descs_loc[i].rec.block);
		if (ret < 0)
			goto out;
	}
	ret = 0;
out:
	kfree(data.part_descs_loc);
	return ret;
}

/*
 * Load Volume Descriptor Sequence described by anchor in bh
 *
 * Returns <0 on error, 0 on success
 */
static int udf_load_sequence(struct super_block *sb, struct buffer_head *bh,
			     struct kernel_lb_addr *fileset)
{
	struct anchorVolDescPtr *anchor;
	sector_t main_s, main_e, reserve_s, reserve_e;
	int ret;

	anchor = (struct anchorVolDescPtr *)bh->b_data;

	/* Locate the main sequence */
	main_s = le32_to_cpu(anchor->mainVolDescSeqExt.extLocation);
	main_e = le32_to_cpu(anchor->mainVolDescSeqExt.extLength);
	main_e = main_e >> sb->s_blocksize_bits;
	main_e += main_s - 1;

	/* Locate the reserve sequence */
	reserve_s = le32_to_cpu(anchor->reserveVolDescSeqExt.extLocation);
	reserve_e = le32_to_cpu(anchor->reserveVolDescSeqExt.extLength);
	reserve_e = reserve_e >> sb->s_blocksize_bits;
	reserve_e += reserve_s - 1;

	/* Process the main & reserve sequences */
	/* responsible for finding the PartitionDesc(s) */
	ret = udf_process_sequence(sb, main_s, main_e, fileset);
	if (ret != -EAGAIN)
		return ret;
	udf_sb_free_partitions(sb);
	ret = udf_process_sequence(sb, reserve_s, reserve_e, fileset);
	if (ret < 0) {
		udf_sb_free_partitions(sb);
		/* No sequence was OK, return -EIO */
		if (ret == -EAGAIN)
			ret = -EIO;
	}
	return ret;
}

/*
 * Check whether there is an anchor block in the given block and
 * load Volume Descriptor Sequence if so.
 *
 * Returns <0 on error, 0 on success, -EAGAIN is special - try next anchor
 * block
 */
static int udf_check_anchor_block(struct super_block *sb, sector_t block,
				  struct kernel_lb_addr *fileset)
{
	struct buffer_head *bh;
	uint16_t ident;
	int ret;

	if (UDF_QUERY_FLAG(sb, UDF_FLAG_VARCONV) &&
	    udf_fixed_to_variable(block) >=
	    i_size_read(sb->s_bdev->bd_inode) >> sb->s_blocksize_bits)
		return -EAGAIN;

	bh = udf_read_tagged(sb, block, block, &ident);
	if (!bh)
		return -EAGAIN;
	if (ident != TAG_IDENT_AVDP) {
		brelse(bh);
		return -EAGAIN;
	}
	ret = udf_load_sequence(sb, bh, fileset);
	brelse(bh);
	return ret;
}

/*
 * Search for an anchor volume descriptor pointer.
 *
 * Returns < 0 on error, 0 on success. -EAGAIN is special - try next set
 * of anchors.
 */
static int udf_scan_anchors(struct super_block *sb, sector_t *lastblock,
			    struct kernel_lb_addr *fileset)
{
	sector_t last[6];
	int i;
	struct udf_sb_info *sbi = UDF_SB(sb);
	int last_count = 0;
	int ret;

	/* First try user provided anchor */
	if (sbi->s_anchor) {
		ret = udf_check_anchor_block(sb, sbi->s_anchor, fileset);
		if (ret != -EAGAIN)
			return ret;
	}
	/*
	 * according to spec, anchor is in either:
	 *     block 256
	 *     lastblock-256
	 *     lastblock
	 *  however, if the disc isn't closed, it could be 512.
	 */
	ret = udf_check_anchor_block(sb, sbi->s_session + 256, fileset);
	if (ret != -EAGAIN)
		return ret;
	/*
	 * The trouble is which block is the last one. Drives often misreport
	 * this so we try various possibilities.
	 */
	last[last_count++] = *lastblock;
	if (*lastblock >= 1)
		last[last_count++] = *lastblock - 1;
	last[last_count++] = *lastblock + 1;
	if (*lastblock >= 2)
		last[last_count++] = *lastblock - 2;
	if (*lastblock >= 150)
		last[last_count++] = *lastblock - 150;
	if (*lastblock >= 152)
		last[last_count++] = *lastblock - 152;

	for (i = 0; i < last_count; i++) {
		if (last[i] >= i_size_read(sb->s_bdev->bd_inode) >>
				sb->s_blocksize_bits)
			continue;
		ret = udf_check_anchor_block(sb, last[i], fileset);
		if (ret != -EAGAIN) {
			if (!ret)
				*lastblock = last[i];
			return ret;
		}
		if (last[i] < 256)
			continue;
		ret = udf_check_anchor_block(sb, last[i] - 256, fileset);
		if (ret != -EAGAIN) {
			if (!ret)
				*lastblock = last[i];
			return ret;
		}
	}

	/* Finally try block 512 in case media is open */
	return udf_check_anchor_block(sb, sbi->s_session + 512, fileset);
}

/*
 * Find an anchor volume descriptor and load Volume Descriptor Sequence from
 * area specified by it. The function expects sbi->s_lastblock to be the last
 * block on the media.
 *
 * Return <0 on error, 0 if anchor found. -EAGAIN is special meaning anchor
 * was not found.
 */
static int udf_find_anchor(struct super_block *sb,
			   struct kernel_lb_addr *fileset)
{
	struct udf_sb_info *sbi = UDF_SB(sb);
	sector_t lastblock = sbi->s_last_block;
	int ret;

	ret = udf_scan_anchors(sb, &lastblock, fileset);
	if (ret != -EAGAIN)
		goto out;

	/* No anchor found? Try VARCONV conversion of block numbers */
	UDF_SET_FLAG(sb, UDF_FLAG_VARCONV);
	lastblock = udf_variable_to_fixed(sbi->s_last_block);
	/* Firstly, we try to not convert number of the last block */
	ret = udf_scan_anchors(sb, &lastblock, fileset);
	if (ret != -EAGAIN)
		goto out;

	lastblock = sbi->s_last_block;
	/* Secondly, we try with converted number of the last block */
	ret = udf_scan_anchors(sb, &lastblock, fileset);
	if (ret < 0) {
		/* VARCONV didn't help. Clear it. */
		UDF_CLEAR_FLAG(sb, UDF_FLAG_VARCONV);
	}
out:
	if (ret == 0)
		sbi->s_last_block = lastblock;
	return ret;
}

/*
 * Check Volume Structure Descriptor, find Anchor block and load Volume
 * Descriptor Sequence.
 *
 * Returns < 0 on error, 0 on success. -EAGAIN is special meaning anchor
 * block was not found.
 */
static int udf_load_vrs(struct super_block *sb, struct udf_options *uopt,
			int silent, struct kernel_lb_addr *fileset)
{
	struct udf_sb_info *sbi = UDF_SB(sb);
	int nsr = 0;
	int ret;

	if (!sb_set_blocksize(sb, uopt->blocksize)) {
		if (!silent)
			udf_warn(sb, "Bad block size\n");
		return -EINVAL;
	}
	sbi->s_last_block = uopt->lastblock;
	if (!uopt->novrs) {
		/* Check that it is NSR02 compliant */
		nsr = udf_check_vsd(sb);
		if (!nsr) {
			if (!silent)
				udf_warn(sb, "No VRS found\n");
			return -EINVAL;
		}
		if (nsr == -1)
			udf_debug("Failed to read sector at offset %d. "
				  "Assuming open disc. Skipping validity "
				  "check\n", VSD_FIRST_SECTOR_OFFSET);
		if (!sbi->s_last_block)
			sbi->s_last_block = udf_get_last_block(sb);
	} else {
		udf_debug("Validity check skipped because of novrs option\n");
	}

	/* Look for anchor block and load Volume Descriptor Sequence */
	sbi->s_anchor = uopt->anchor;
	ret = udf_find_anchor(sb, fileset);
	if (ret < 0) {
		if (!silent && ret == -EAGAIN)
			udf_warn(sb, "No anchor found\n");
		return ret;
	}
	return 0;
}

static void udf_finalize_lvid(struct logicalVolIntegrityDesc *lvid)
{
	struct timespec64 ts;

	ktime_get_real_ts64(&ts);
	udf_time_to_disk_stamp(&lvid->recordingDateAndTime, ts);
	lvid->descTag.descCRC = cpu_to_le16(
		crc_itu_t(0, (char *)lvid + sizeof(struct tag),
			le16_to_cpu(lvid->descTag.descCRCLength)));
	lvid->descTag.tagChecksum = udf_tag_checksum(&lvid->descTag);
}

static void udf_open_lvid(struct super_block *sb)
{
	struct udf_sb_info *sbi = UDF_SB(sb);
	struct buffer_head *bh = sbi->s_lvid_bh;
	struct logicalVolIntegrityDesc *lvid;
	struct logicalVolIntegrityDescImpUse *lvidiu;

	if (!bh)
		return;
	lvid = (struct logicalVolIntegrityDesc *)bh->b_data;
	lvidiu = udf_sb_lvidiu(sb);
	if (!lvidiu)
		return;

	mutex_lock(&sbi->s_alloc_mutex);
	lvidiu->impIdent.identSuffix[0] = UDF_OS_CLASS_UNIX;
	lvidiu->impIdent.identSuffix[1] = UDF_OS_ID_LINUX;
	if (le32_to_cpu(lvid->integrityType) == LVID_INTEGRITY_TYPE_CLOSE)
		lvid->integrityType = cpu_to_le32(LVID_INTEGRITY_TYPE_OPEN);
	else
		UDF_SET_FLAG(sb, UDF_FLAG_INCONSISTENT);

	udf_finalize_lvid(lvid);
	mark_buffer_dirty(bh);
	sbi->s_lvid_dirty = 0;
	mutex_unlock(&sbi->s_alloc_mutex);
	/* Make opening of filesystem visible on the media immediately */
	sync_dirty_buffer(bh);
}

static void udf_close_lvid(struct super_block *sb)
{
	struct udf_sb_info *sbi = UDF_SB(sb);
	struct buffer_head *bh = sbi->s_lvid_bh;
	struct logicalVolIntegrityDesc *lvid;
	struct logicalVolIntegrityDescImpUse *lvidiu;

	if (!bh)
		return;
	lvid = (struct logicalVolIntegrityDesc *)bh->b_data;
	lvidiu = udf_sb_lvidiu(sb);
	if (!lvidiu)
		return;

	mutex_lock(&sbi->s_alloc_mutex);
	lvidiu->impIdent.identSuffix[0] = UDF_OS_CLASS_UNIX;
	lvidiu->impIdent.identSuffix[1] = UDF_OS_ID_LINUX;
	if (UDF_MAX_WRITE_VERSION > le16_to_cpu(lvidiu->maxUDFWriteRev))
		lvidiu->maxUDFWriteRev = cpu_to_le16(UDF_MAX_WRITE_VERSION);
	if (sbi->s_udfrev > le16_to_cpu(lvidiu->minUDFReadRev))
		lvidiu->minUDFReadRev = cpu_to_le16(sbi->s_udfrev);
	if (sbi->s_udfrev > le16_to_cpu(lvidiu->minUDFWriteRev))
		lvidiu->minUDFWriteRev = cpu_to_le16(sbi->s_udfrev);
	if (!UDF_QUERY_FLAG(sb, UDF_FLAG_INCONSISTENT))
		lvid->integrityType = cpu_to_le32(LVID_INTEGRITY_TYPE_CLOSE);

	/*
	 * We set buffer uptodate unconditionally here to avoid spurious
	 * warnings from mark_buffer_dirty() when previous EIO has marked
	 * the buffer as !uptodate
	 */
	set_buffer_uptodate(bh);
	udf_finalize_lvid(lvid);
	mark_buffer_dirty(bh);
	sbi->s_lvid_dirty = 0;
	mutex_unlock(&sbi->s_alloc_mutex);
	/* Make closing of filesystem visible on the media immediately */
	sync_dirty_buffer(bh);
}

u64 lvid_get_unique_id(struct super_block *sb)
{
	struct buffer_head *bh;
	struct udf_sb_info *sbi = UDF_SB(sb);
	struct logicalVolIntegrityDesc *lvid;
	struct logicalVolHeaderDesc *lvhd;
	u64 uniqueID;
	u64 ret;

	bh = sbi->s_lvid_bh;
	if (!bh)
		return 0;

	lvid = (struct logicalVolIntegrityDesc *)bh->b_data;
	lvhd = (struct logicalVolHeaderDesc *)lvid->logicalVolContentsUse;

	mutex_lock(&sbi->s_alloc_mutex);
	ret = uniqueID = le64_to_cpu(lvhd->uniqueID);
	if (!(++uniqueID & 0xFFFFFFFF))
		uniqueID += 16;
	lvhd->uniqueID = cpu_to_le64(uniqueID);
	udf_updated_lvid(sb);
	mutex_unlock(&sbi->s_alloc_mutex);

	return ret;
}

static int udf_fill_super(struct super_block *sb, void *options, int silent)
{
	int ret = -EINVAL;
	struct inode *inode = NULL;
	struct udf_options uopt;
	struct kernel_lb_addr rootdir, fileset;
	struct udf_sb_info *sbi;
	bool lvid_open = false;

	uopt.flags = (1 << UDF_FLAG_USE_AD_IN_ICB) | (1 << UDF_FLAG_STRICT);
	/* By default we'll use overflow[ug]id when UDF inode [ug]id == -1 */
	uopt.uid = make_kuid(current_user_ns(), overflowuid);
	uopt.gid = make_kgid(current_user_ns(), overflowgid);
	uopt.umask = 0;
	uopt.fmode = UDF_INVALID_MODE;
	uopt.dmode = UDF_INVALID_MODE;
	uopt.nls_map = NULL;

	sbi = kzalloc(sizeof(*sbi), GFP_KERNEL);
	if (!sbi)
		return -ENOMEM;

	sb->s_fs_info = sbi;

	mutex_init(&sbi->s_alloc_mutex);

	if (!udf_parse_options((char *)options, &uopt, false))
		goto parse_options_failure;

	fileset.logicalBlockNum = 0xFFFFFFFF;
	fileset.partitionReferenceNum = 0xFFFF;

	sbi->s_flags = uopt.flags;
	sbi->s_uid = uopt.uid;
	sbi->s_gid = uopt.gid;
	sbi->s_umask = uopt.umask;
	sbi->s_fmode = uopt.fmode;
	sbi->s_dmode = uopt.dmode;
	sbi->s_nls_map = uopt.nls_map;
	rwlock_init(&sbi->s_cred_lock);

	if (uopt.session == 0xFFFFFFFF)
		sbi->s_session = udf_get_last_session(sb);
	else
		sbi->s_session = uopt.session;

	udf_debug("Multi-session=%d\n", sbi->s_session);

	/* Fill in the rest of the superblock */
	sb->s_op = &udf_sb_ops;
	sb->s_export_op = &udf_export_ops;

	sb->s_magic = UDF_SUPER_MAGIC;
	sb->s_time_gran = 1000;

	if (uopt.flags & (1 << UDF_FLAG_BLOCKSIZE_SET)) {
		ret = udf_load_vrs(sb, &uopt, silent, &fileset);
	} else {
		uopt.blocksize = bdev_logical_block_size(sb->s_bdev);
		while (uopt.blocksize <= 4096) {
			ret = udf_load_vrs(sb, &uopt, silent, &fileset);
			if (ret < 0) {
				if (!silent && ret != -EACCES) {
					pr_notice("Scanning with blocksize %u failed\n",
						  uopt.blocksize);
				}
				brelse(sbi->s_lvid_bh);
				sbi->s_lvid_bh = NULL;
				/*
				 * EACCES is special - we want to propagate to
				 * upper layers that we cannot handle RW mount.
				 */
				if (ret == -EACCES)
					break;
			} else
				break;

			uopt.blocksize <<= 1;
		}
	}
	if (ret < 0) {
		if (ret == -EAGAIN) {
			udf_warn(sb, "No partition found (1)\n");
			ret = -EINVAL;
		}
		goto error_out;
	}

	udf_debug("Lastblock=%u\n", sbi->s_last_block);

	if (sbi->s_lvid_bh) {
		struct logicalVolIntegrityDescImpUse *lvidiu =
							udf_sb_lvidiu(sb);
		uint16_t minUDFReadRev;
		uint16_t minUDFWriteRev;

		if (!lvidiu) {
			ret = -EINVAL;
			goto error_out;
		}
		minUDFReadRev = le16_to_cpu(lvidiu->minUDFReadRev);
		minUDFWriteRev = le16_to_cpu(lvidiu->minUDFWriteRev);
		if (minUDFReadRev > UDF_MAX_READ_VERSION) {
			udf_err(sb, "minUDFReadRev=%x (max is %x)\n",
				minUDFReadRev,
				UDF_MAX_READ_VERSION);
			ret = -EINVAL;
			goto error_out;
		} else if (minUDFWriteRev > UDF_MAX_WRITE_VERSION) {
			if (!sb_rdonly(sb)) {
				ret = -EACCES;
				goto error_out;
			}
			UDF_SET_FLAG(sb, UDF_FLAG_RW_INCOMPAT);
		}

		sbi->s_udfrev = minUDFWriteRev;

		if (minUDFReadRev >= UDF_VERS_USE_EXTENDED_FE)
			UDF_SET_FLAG(sb, UDF_FLAG_USE_EXTENDED_FE);
		if (minUDFReadRev >= UDF_VERS_USE_STREAMS)
			UDF_SET_FLAG(sb, UDF_FLAG_USE_STREAMS);
	}

	if (!sbi->s_partitions) {
		udf_warn(sb, "No partition found (2)\n");
		ret = -EINVAL;
		goto error_out;
	}

	if (sbi->s_partmaps[sbi->s_partition].s_partition_flags &
			UDF_PART_FLAG_READ_ONLY) {
		if (!sb_rdonly(sb)) {
			ret = -EACCES;
			goto error_out;
		}
		UDF_SET_FLAG(sb, UDF_FLAG_RW_INCOMPAT);
	}

	ret = udf_find_fileset(sb, &fileset, &rootdir);
	if (ret < 0) {
		udf_warn(sb, "No fileset found\n");
		goto error_out;
	}

	if (!silent) {
		struct timestamp ts;
		udf_time_to_disk_stamp(&ts, sbi->s_record_time);
		udf_info("Mounting volume '%s', timestamp %04u/%02u/%02u %02u:%02u (%x)\n",
			 sbi->s_volume_ident,
			 le16_to_cpu(ts.year), ts.month, ts.day,
			 ts.hour, ts.minute, le16_to_cpu(ts.typeAndTimezone));
	}
	if (!sb_rdonly(sb)) {
		udf_open_lvid(sb);
		lvid_open = true;
	}

	/* Assign the root inode */
	/* assign inodes by physical block number */
	/* perhaps it's not extensible enough, but for now ... */
	inode = udf_iget(sb, &rootdir);
	if (IS_ERR(inode)) {
		udf_err(sb, "Error in udf_iget, block=%u, partition=%u\n",
		       rootdir.logicalBlockNum, rootdir.partitionReferenceNum);
		ret = PTR_ERR(inode);
		goto error_out;
	}

	/* Allocate a dentry for the root inode */
	sb->s_root = d_make_root(inode);
	if (!sb->s_root) {
		udf_err(sb, "Couldn't allocate root dentry\n");
		ret = -ENOMEM;
		goto error_out;
	}
	sb->s_maxbytes = MAX_LFS_FILESIZE;
	sb->s_max_links = UDF_MAX_LINKS;
	return 0;

error_out:
	iput(sbi->s_vat_inode);
parse_options_failure:
	unload_nls(uopt.nls_map);
	if (lvid_open)
		udf_close_lvid(sb);
	brelse(sbi->s_lvid_bh);
	udf_sb_free_partitions(sb);
	kfree(sbi);
	sb->s_fs_info = NULL;

	return ret;
}

void _udf_err(struct super_block *sb, const char *function,
	      const char *fmt, ...)
{
	struct va_format vaf;
	va_list args;

	va_start(args, fmt);

	vaf.fmt = fmt;
	vaf.va = &args;

	pr_err("error (device %s): %s: %pV", sb->s_id, function, &vaf);

	va_end(args);
}

void _udf_warn(struct super_block *sb, const char *function,
	       const char *fmt, ...)
{
	struct va_format vaf;
	va_list args;

	va_start(args, fmt);

	vaf.fmt = fmt;
	vaf.va = &args;

	pr_warn("warning (device %s): %s: %pV", sb->s_id, function, &vaf);

	va_end(args);
}

static void udf_put_super(struct super_block *sb)
{
	struct udf_sb_info *sbi;

	sbi = UDF_SB(sb);

	iput(sbi->s_vat_inode);
	unload_nls(sbi->s_nls_map);
	if (!sb_rdonly(sb))
		udf_close_lvid(sb);
	brelse(sbi->s_lvid_bh);
	udf_sb_free_partitions(sb);
	mutex_destroy(&sbi->s_alloc_mutex);
	kfree(sb->s_fs_info);
	sb->s_fs_info = NULL;
}

static int udf_sync_fs(struct super_block *sb, int wait)
{
	struct udf_sb_info *sbi = UDF_SB(sb);

	mutex_lock(&sbi->s_alloc_mutex);
	if (sbi->s_lvid_dirty) {
		struct buffer_head *bh = sbi->s_lvid_bh;
		struct logicalVolIntegrityDesc *lvid;

		lvid = (struct logicalVolIntegrityDesc *)bh->b_data;
		udf_finalize_lvid(lvid);

		/*
		 * Blockdevice will be synced later so we don't have to submit
		 * the buffer for IO
		 */
		mark_buffer_dirty(bh);
		sbi->s_lvid_dirty = 0;
	}
	mutex_unlock(&sbi->s_alloc_mutex);

	return 0;
}

static int udf_statfs(struct dentry *dentry, struct kstatfs *buf)
{
	struct super_block *sb = dentry->d_sb;
	struct udf_sb_info *sbi = UDF_SB(sb);
	struct logicalVolIntegrityDescImpUse *lvidiu;
	u64 id = huge_encode_dev(sb->s_bdev->bd_dev);

	lvidiu = udf_sb_lvidiu(sb);
	buf->f_type = UDF_SUPER_MAGIC;
	buf->f_bsize = sb->s_blocksize;
	buf->f_blocks = sbi->s_partmaps[sbi->s_partition].s_partition_len;
	buf->f_bfree = udf_count_free(sb);
	buf->f_bavail = buf->f_bfree;
	buf->f_files = (lvidiu != NULL ? (le32_to_cpu(lvidiu->numFiles) +
					  le32_to_cpu(lvidiu->numDirs)) : 0)
			+ buf->f_bfree;
	buf->f_ffree = buf->f_bfree;
	buf->f_namelen = UDF_NAME_LEN;
	buf->f_fsid.val[0] = (u32)id;
	buf->f_fsid.val[1] = (u32)(id >> 32);

	return 0;
}

static unsigned int udf_count_free_bitmap(struct super_block *sb,
					  struct udf_bitmap *bitmap)
{
	struct buffer_head *bh = NULL;
	unsigned int accum = 0;
	int index;
	udf_pblk_t block = 0, newblock;
	struct kernel_lb_addr loc;
	uint32_t bytes;
	uint8_t *ptr;
	uint16_t ident;
	struct spaceBitmapDesc *bm;

	loc.logicalBlockNum = bitmap->s_extPosition;
	loc.partitionReferenceNum = UDF_SB(sb)->s_partition;
	bh = udf_read_ptagged(sb, &loc, 0, &ident);

	if (!bh) {
		udf_err(sb, "udf_count_free failed\n");
		goto out;
	} else if (ident != TAG_IDENT_SBD) {
		brelse(bh);
		udf_err(sb, "udf_count_free failed\n");
		goto out;
	}

	bm = (struct spaceBitmapDesc *)bh->b_data;
	bytes = le32_to_cpu(bm->numOfBytes);
	index = sizeof(struct spaceBitmapDesc); /* offset in first block only */
	ptr = (uint8_t *)bh->b_data;

	while (bytes > 0) {
		u32 cur_bytes = min_t(u32, bytes, sb->s_blocksize - index);
		accum += bitmap_weight((const unsigned long *)(ptr + index),
					cur_bytes * 8);
		bytes -= cur_bytes;
		if (bytes) {
			brelse(bh);
			newblock = udf_get_lb_pblock(sb, &loc, ++block);
			bh = udf_tread(sb, newblock);
			if (!bh) {
				udf_debug("read failed\n");
				goto out;
			}
			index = 0;
			ptr = (uint8_t *)bh->b_data;
		}
	}
	brelse(bh);
out:
	return accum;
}

static unsigned int udf_count_free_table(struct super_block *sb,
					 struct inode *table)
{
	unsigned int accum = 0;
	uint32_t elen;
	struct kernel_lb_addr eloc;
	int8_t etype;
	struct extent_position epos;

	mutex_lock(&UDF_SB(sb)->s_alloc_mutex);
	epos.block = UDF_I(table)->i_location;
	epos.offset = sizeof(struct unallocSpaceEntry);
	epos.bh = NULL;

	while ((etype = udf_next_aext(table, &epos, &eloc, &elen, 1)) != -1)
		accum += (elen >> table->i_sb->s_blocksize_bits);

	brelse(epos.bh);
	mutex_unlock(&UDF_SB(sb)->s_alloc_mutex);

	return accum;
}

static unsigned int udf_count_free(struct super_block *sb)
{
	unsigned int accum = 0;
	struct udf_sb_info *sbi = UDF_SB(sb);
	struct udf_part_map *map;
	unsigned int part = sbi->s_partition;
	int ptype = sbi->s_partmaps[part].s_partition_type;

	if (ptype == UDF_METADATA_MAP25) {
		part = sbi->s_partmaps[part].s_type_specific.s_metadata.
							s_phys_partition_ref;
	} else if (ptype == UDF_VIRTUAL_MAP15 || ptype == UDF_VIRTUAL_MAP20) {
		/*
		 * Filesystems with VAT are append-only and we cannot write to
 		 * them. Let's just report 0 here.
		 */
		return 0;
	}

	if (sbi->s_lvid_bh) {
		struct logicalVolIntegrityDesc *lvid =
			(struct logicalVolIntegrityDesc *)
			sbi->s_lvid_bh->b_data;
		if (le32_to_cpu(lvid->numOfPartitions) > part) {
			accum = le32_to_cpu(
					lvid->freeSpaceTable[part]);
			if (accum == 0xFFFFFFFF)
				accum = 0;
		}
	}

	if (accum)
		return accum;

	map = &sbi->s_partmaps[part];
	if (map->s_partition_flags & UDF_PART_FLAG_UNALLOC_BITMAP) {
		accum += udf_count_free_bitmap(sb,
					       map->s_uspace.s_bitmap);
	}
	if (accum)
		return accum;

	if (map->s_partition_flags & UDF_PART_FLAG_UNALLOC_TABLE) {
		accum += udf_count_free_table(sb,
					      map->s_uspace.s_table);
	}
	return accum;
}

MODULE_AUTHOR("Ben Fennema");
MODULE_DESCRIPTION("Universal Disk Format Filesystem");
MODULE_LICENSE("GPL");
module_init(init_udf_fs)
module_exit(exit_udf_fs)
"kt">int entry, const u64 *words, struct sk_buff **skbref) { int dlen, hlen, len, i, alloclen; int off, swivel = RX_SWIVEL_OFF_VAL; struct cas_page *page; struct sk_buff *skb; void *addr, *crcaddr; __sum16 csum; char *p; hlen = CAS_VAL(RX_COMP2_HDR_SIZE, words[1]); dlen = CAS_VAL(RX_COMP1_DATA_SIZE, words[0]); len = hlen + dlen; if (RX_COPY_ALWAYS || (words[2] & RX_COMP3_SMALL_PKT)) alloclen = len; else alloclen = max(hlen, RX_COPY_MIN); skb = netdev_alloc_skb(cp->dev, alloclen + swivel + cp->crc_size); if (skb == NULL) return -1; *skbref = skb; skb_reserve(skb, swivel); p = skb->data; addr = crcaddr = NULL; if (hlen) { /* always copy header pages */ i = CAS_VAL(RX_COMP2_HDR_INDEX, words[1]); page = cp->rx_pages[CAS_VAL(RX_INDEX_RING, i)][CAS_VAL(RX_INDEX_NUM, i)]; off = CAS_VAL(RX_COMP2_HDR_OFF, words[1]) * 0x100 + swivel; i = hlen; if (!dlen) /* attach FCS */ i += cp->crc_size; pci_dma_sync_single_for_cpu(cp->pdev, page->dma_addr + off, i, PCI_DMA_FROMDEVICE); addr = cas_page_map(page->buffer); memcpy(p, addr + off, i); pci_dma_sync_single_for_device(cp->pdev, page->dma_addr + off, i, PCI_DMA_FROMDEVICE); cas_page_unmap(addr); RX_USED_ADD(page, 0x100); p += hlen; swivel = 0; } if (alloclen < (hlen + dlen)) { skb_frag_t *frag = skb_shinfo(skb)->frags; /* normal or jumbo packets. we use frags */ i = CAS_VAL(RX_COMP1_DATA_INDEX, words[0]); page = cp->rx_pages[CAS_VAL(RX_INDEX_RING, i)][CAS_VAL(RX_INDEX_NUM, i)]; off = CAS_VAL(RX_COMP1_DATA_OFF, words[0]) + swivel; hlen = min(cp->page_size - off, dlen); if (hlen < 0) { netif_printk(cp, rx_err, KERN_DEBUG, cp->dev, "rx page overflow: %d\n", hlen); dev_kfree_skb_irq(skb); return -1; } i = hlen; if (i == dlen) /* attach FCS */ i += cp->crc_size; pci_dma_sync_single_for_cpu(cp->pdev, page->dma_addr + off, i, PCI_DMA_FROMDEVICE); /* make sure we always copy a header */ swivel = 0; if (p == (char *) skb->data) { /* not split */ addr = cas_page_map(page->buffer); memcpy(p, addr + off, RX_COPY_MIN); pci_dma_sync_single_for_device(cp->pdev, page->dma_addr + off, i, PCI_DMA_FROMDEVICE); cas_page_unmap(addr); off += RX_COPY_MIN; swivel = RX_COPY_MIN; RX_USED_ADD(page, cp->mtu_stride); } else { RX_USED_ADD(page, hlen); } skb_put(skb, alloclen); skb_shinfo(skb)->nr_frags++; skb->data_len += hlen - swivel; skb->truesize += hlen - swivel; skb->len += hlen - swivel; __skb_frag_set_page(frag, page->buffer); __skb_frag_ref(frag); skb_frag_off_set(frag, off); skb_frag_size_set(frag, hlen - swivel); /* any more data? */ if ((words[0] & RX_COMP1_SPLIT_PKT) && ((dlen -= hlen) > 0)) { hlen = dlen; off = 0; i = CAS_VAL(RX_COMP2_NEXT_INDEX, words[1]); page = cp->rx_pages[CAS_VAL(RX_INDEX_RING, i)][CAS_VAL(RX_INDEX_NUM, i)]; pci_dma_sync_single_for_cpu(cp->pdev, page->dma_addr, hlen + cp->crc_size, PCI_DMA_FROMDEVICE); pci_dma_sync_single_for_device(cp->pdev, page->dma_addr, hlen + cp->crc_size, PCI_DMA_FROMDEVICE); skb_shinfo(skb)->nr_frags++; skb->data_len += hlen; skb->len += hlen; frag++; __skb_frag_set_page(frag, page->buffer); __skb_frag_ref(frag); skb_frag_off_set(frag, 0); skb_frag_size_set(frag, hlen); RX_USED_ADD(page, hlen + cp->crc_size); } if (cp->crc_size) { addr = cas_page_map(page->buffer); crcaddr = addr + off + hlen; } } else { /* copying packet */ if (!dlen) goto end_copy_pkt; i = CAS_VAL(RX_COMP1_DATA_INDEX, words[0]); page = cp->rx_pages[CAS_VAL(RX_INDEX_RING, i)][CAS_VAL(RX_INDEX_NUM, i)]; off = CAS_VAL(RX_COMP1_DATA_OFF, words[0]) + swivel; hlen = min(cp->page_size - off, dlen); if (hlen < 0) { netif_printk(cp, rx_err, KERN_DEBUG, cp->dev, "rx page overflow: %d\n", hlen); dev_kfree_skb_irq(skb); return -1; } i = hlen; if (i == dlen) /* attach FCS */ i += cp->crc_size; pci_dma_sync_single_for_cpu(cp->pdev, page->dma_addr + off, i, PCI_DMA_FROMDEVICE); addr = cas_page_map(page->buffer); memcpy(p, addr + off, i); pci_dma_sync_single_for_device(cp->pdev, page->dma_addr + off, i, PCI_DMA_FROMDEVICE); cas_page_unmap(addr); if (p == (char *) skb->data) /* not split */ RX_USED_ADD(page, cp->mtu_stride); else RX_USED_ADD(page, i); /* any more data? */ if ((words[0] & RX_COMP1_SPLIT_PKT) && ((dlen -= hlen) > 0)) { p += hlen; i = CAS_VAL(RX_COMP2_NEXT_INDEX, words[1]); page = cp->rx_pages[CAS_VAL(RX_INDEX_RING, i)][CAS_VAL(RX_INDEX_NUM, i)]; pci_dma_sync_single_for_cpu(cp->pdev, page->dma_addr, dlen + cp->crc_size, PCI_DMA_FROMDEVICE); addr = cas_page_map(page->buffer); memcpy(p, addr, dlen + cp->crc_size); pci_dma_sync_single_for_device(cp->pdev, page->dma_addr, dlen + cp->crc_size, PCI_DMA_FROMDEVICE); cas_page_unmap(addr); RX_USED_ADD(page, dlen + cp->crc_size); } end_copy_pkt: if (cp->crc_size) { addr = NULL; crcaddr = skb->data + alloclen; } skb_put(skb, alloclen); } csum = (__force __sum16)htons(CAS_VAL(RX_COMP4_TCP_CSUM, words[3])); if (cp->crc_size) { /* checksum includes FCS. strip it out. */ csum = csum_fold(csum_partial(crcaddr, cp->crc_size, csum_unfold(csum))); if (addr) cas_page_unmap(addr); } skb->protocol = eth_type_trans(skb, cp->dev); if (skb->protocol == htons(ETH_P_IP)) { skb->csum = csum_unfold(~csum); skb->ip_summed = CHECKSUM_COMPLETE; } else skb_checksum_none_assert(skb); return len; } /* we can handle up to 64 rx flows at a time. we do the same thing * as nonreassm except that we batch up the buffers. * NOTE: we currently just treat each flow as a bunch of packets that * we pass up. a better way would be to coalesce the packets * into a jumbo packet. to do that, we need to do the following: * 1) the first packet will have a clean split between header and * data. save both. * 2) each time the next flow packet comes in, extend the * data length and merge the checksums. * 3) on flow release, fix up the header. * 4) make sure the higher layer doesn't care. * because packets get coalesced, we shouldn't run into fragment count * issues. */ static inline void cas_rx_flow_pkt(struct cas *cp, const u64 *words, struct sk_buff *skb) { int flowid = CAS_VAL(RX_COMP3_FLOWID, words[2]) & (N_RX_FLOWS - 1); struct sk_buff_head *flow = &cp->rx_flows[flowid]; /* this is protected at a higher layer, so no need to * do any additional locking here. stick the buffer * at the end. */ __skb_queue_tail(flow, skb); if (words[0] & RX_COMP1_RELEASE_FLOW) { while ((skb = __skb_dequeue(flow))) { cas_skb_release(skb); } } } /* put rx descriptor back on ring. if a buffer is in use by a higher * layer, this will need to put in a replacement. */ static void cas_post_page(struct cas *cp, const int ring, const int index) { cas_page_t *new; int entry; entry = cp->rx_old[ring]; new = cas_page_swap(cp, ring, index); cp->init_rxds[ring][entry].buffer = cpu_to_le64(new->dma_addr); cp->init_rxds[ring][entry].index = cpu_to_le64(CAS_BASE(RX_INDEX_NUM, index) | CAS_BASE(RX_INDEX_RING, ring)); entry = RX_DESC_ENTRY(ring, entry + 1); cp->rx_old[ring] = entry; if (entry % 4) return; if (ring == 0) writel(entry, cp->regs + REG_RX_KICK); else if ((N_RX_DESC_RINGS > 1) && (cp->cas_flags & CAS_FLAG_REG_PLUS)) writel(entry, cp->regs + REG_PLUS_RX_KICK1); } /* only when things are bad */ static int cas_post_rxds_ringN(struct cas *cp, int ring, int num) { unsigned int entry, last, count, released; int cluster; cas_page_t **page = cp->rx_pages[ring]; entry = cp->rx_old[ring]; netif_printk(cp, intr, KERN_DEBUG, cp->dev, "rxd[%d] interrupt, done: %d\n", ring, entry); cluster = -1; count = entry & 0x3; last = RX_DESC_ENTRY(ring, num ? entry + num - 4: entry - 4); released = 0; while (entry != last) { /* make a new buffer if it's still in use */ if (page_count(page[entry]->buffer) > 1) { cas_page_t *new = cas_page_dequeue(cp); if (!new) { /* let the timer know that we need to * do this again */ cp->cas_flags |= CAS_FLAG_RXD_POST(ring); if (!timer_pending(&cp->link_timer)) mod_timer(&cp->link_timer, jiffies + CAS_LINK_FAST_TIMEOUT); cp->rx_old[ring] = entry; cp->rx_last[ring] = num ? num - released : 0; return -ENOMEM; } spin_lock(&cp->rx_inuse_lock); list_add(&page[entry]->list, &cp->rx_inuse_list); spin_unlock(&cp->rx_inuse_lock); cp->init_rxds[ring][entry].buffer = cpu_to_le64(new->dma_addr); page[entry] = new; } if (++count == 4) { cluster = entry; count = 0; } released++; entry = RX_DESC_ENTRY(ring, entry + 1); } cp->rx_old[ring] = entry; if (cluster < 0) return 0; if (ring == 0) writel(cluster, cp->regs + REG_RX_KICK); else if ((N_RX_DESC_RINGS > 1) && (cp->cas_flags & CAS_FLAG_REG_PLUS)) writel(cluster, cp->regs + REG_PLUS_RX_KICK1); return 0; } /* process a completion ring. packets are set up in three basic ways: * small packets: should be copied header + data in single buffer. * large packets: header and data in a single buffer. * split packets: header in a separate buffer from data. * data may be in multiple pages. data may be > 256 * bytes but in a single page. * * NOTE: RX page posting is done in this routine as well. while there's * the capability of using multiple RX completion rings, it isn't * really worthwhile due to the fact that the page posting will * force serialization on the single descriptor ring. */ static int cas_rx_ringN(struct cas *cp, int ring, int budget) { struct cas_rx_comp *rxcs = cp->init_rxcs[ring]; int entry, drops; int npackets = 0; netif_printk(cp, intr, KERN_DEBUG, cp->dev, "rx[%d] interrupt, done: %d/%d\n", ring, readl(cp->regs + REG_RX_COMP_HEAD), cp->rx_new[ring]); entry = cp->rx_new[ring]; drops = 0; while (1) { struct cas_rx_comp *rxc = rxcs + entry; struct sk_buff *uninitialized_var(skb); int type, len; u64 words[4]; int i, dring; words[0] = le64_to_cpu(rxc->word1); words[1] = le64_to_cpu(rxc->word2); words[2] = le64_to_cpu(rxc->word3); words[3] = le64_to_cpu(rxc->word4); /* don't touch if still owned by hw */ type = CAS_VAL(RX_COMP1_TYPE, words[0]); if (type == 0) break; /* hw hasn't cleared the zero bit yet */ if (words[3] & RX_COMP4_ZERO) { break; } /* get info on the packet */ if (words[3] & (RX_COMP4_LEN_MISMATCH | RX_COMP4_BAD)) { spin_lock(&cp->stat_lock[ring]); cp->net_stats[ring].rx_errors++; if (words[3] & RX_COMP4_LEN_MISMATCH) cp->net_stats[ring].rx_length_errors++; if (words[3] & RX_COMP4_BAD) cp->net_stats[ring].rx_crc_errors++; spin_unlock(&cp->stat_lock[ring]); /* We'll just return it to Cassini. */ drop_it: spin_lock(&cp->stat_lock[ring]); ++cp->net_stats[ring].rx_dropped; spin_unlock(&cp->stat_lock[ring]); goto next; } len = cas_rx_process_pkt(cp, rxc, entry, words, &skb); if (len < 0) { ++drops; goto drop_it; } /* see if it's a flow re-assembly or not. the driver * itself handles release back up. */ if (RX_DONT_BATCH || (type == 0x2)) { /* non-reassm: these always get released */ cas_skb_release(skb); } else { cas_rx_flow_pkt(cp, words, skb); } spin_lock(&cp->stat_lock[ring]); cp->net_stats[ring].rx_packets++; cp->net_stats[ring].rx_bytes += len; spin_unlock(&cp->stat_lock[ring]); next: npackets++; /* should it be released? */ if (words[0] & RX_COMP1_RELEASE_HDR) { i = CAS_VAL(RX_COMP2_HDR_INDEX, words[1]); dring = CAS_VAL(RX_INDEX_RING, i); i = CAS_VAL(RX_INDEX_NUM, i); cas_post_page(cp, dring, i); } if (words[0] & RX_COMP1_RELEASE_DATA) { i = CAS_VAL(RX_COMP1_DATA_INDEX, words[0]); dring = CAS_VAL(RX_INDEX_RING, i); i = CAS_VAL(RX_INDEX_NUM, i); cas_post_page(cp, dring, i); } if (words[0] & RX_COMP1_RELEASE_NEXT) { i = CAS_VAL(RX_COMP2_NEXT_INDEX, words[1]); dring = CAS_VAL(RX_INDEX_RING, i); i = CAS_VAL(RX_INDEX_NUM, i); cas_post_page(cp, dring, i); } /* skip to the next entry */ entry = RX_COMP_ENTRY(ring, entry + 1 + CAS_VAL(RX_COMP1_SKIP, words[0])); #ifdef USE_NAPI if (budget && (npackets >= budget)) break; #endif } cp->rx_new[ring] = entry; if (drops) netdev_info(cp->dev, "Memory squeeze, deferring packet\n"); return npackets; } /* put completion entries back on the ring */ static void cas_post_rxcs_ringN(struct net_device *dev, struct cas *cp, int ring) { struct cas_rx_comp *rxc = cp->init_rxcs[ring]; int last, entry; last = cp->rx_cur[ring]; entry = cp->rx_new[ring]; netif_printk(cp, intr, KERN_DEBUG, dev, "rxc[%d] interrupt, done: %d/%d\n", ring, readl(cp->regs + REG_RX_COMP_HEAD), entry); /* zero and re-mark descriptors */ while (last != entry) { cas_rxc_init(rxc + last); last = RX_COMP_ENTRY(ring, last + 1); } cp->rx_cur[ring] = last; if (ring == 0) writel(last, cp->regs + REG_RX_COMP_TAIL); else if (cp->cas_flags & CAS_FLAG_REG_PLUS) writel(last, cp->regs + REG_PLUS_RX_COMPN_TAIL(ring)); } /* cassini can use all four PCI interrupts for the completion ring. * rings 3 and 4 are identical */ #if defined(USE_PCI_INTC) || defined(USE_PCI_INTD) static inline void cas_handle_irqN(struct net_device *dev, struct cas *cp, const u32 status, const int ring) { if (status & (INTR_RX_COMP_FULL_ALT | INTR_RX_COMP_AF_ALT)) cas_post_rxcs_ringN(dev, cp, ring); } static irqreturn_t cas_interruptN(int irq, void *dev_id) { struct net_device *dev = dev_id; struct cas *cp = netdev_priv(dev); unsigned long flags; int ring = (irq == cp->pci_irq_INTC) ? 2 : 3; u32 status = readl(cp->regs + REG_PLUS_INTRN_STATUS(ring)); /* check for shared irq */ if (status == 0) return IRQ_NONE; spin_lock_irqsave(&cp->lock, flags); if (status & INTR_RX_DONE_ALT) { /* handle rx separately */ #ifdef USE_NAPI cas_mask_intr(cp); napi_schedule(&cp->napi); #else cas_rx_ringN(cp, ring, 0); #endif status &= ~INTR_RX_DONE_ALT; } if (status) cas_handle_irqN(dev, cp, status, ring); spin_unlock_irqrestore(&cp->lock, flags); return IRQ_HANDLED; } #endif #ifdef USE_PCI_INTB /* everything but rx packets */ static inline void cas_handle_irq1(struct cas *cp, const u32 status) { if (status & INTR_RX_BUF_UNAVAIL_1) { /* Frame arrived, no free RX buffers available. * NOTE: we can get this on a link transition. */ cas_post_rxds_ringN(cp, 1, 0); spin_lock(&cp->stat_lock[1]); cp->net_stats[1].rx_dropped++; spin_unlock(&cp->stat_lock[1]); } if (status & INTR_RX_BUF_AE_1) cas_post_rxds_ringN(cp, 1, RX_DESC_RINGN_SIZE(1) - RX_AE_FREEN_VAL(1)); if (status & (INTR_RX_COMP_AF | INTR_RX_COMP_FULL)) cas_post_rxcs_ringN(cp, 1); } /* ring 2 handles a few more events than 3 and 4 */ static irqreturn_t cas_interrupt1(int irq, void *dev_id) { struct net_device *dev = dev_id; struct cas *cp = netdev_priv(dev); unsigned long flags; u32 status = readl(cp->regs + REG_PLUS_INTRN_STATUS(1)); /* check for shared interrupt */ if (status == 0) return IRQ_NONE; spin_lock_irqsave(&cp->lock, flags); if (status & INTR_RX_DONE_ALT) { /* handle rx separately */ #ifdef USE_NAPI cas_mask_intr(cp); napi_schedule(&cp->napi); #else cas_rx_ringN(cp, 1, 0); #endif status &= ~INTR_RX_DONE_ALT; } if (status) cas_handle_irq1(cp, status); spin_unlock_irqrestore(&cp->lock, flags); return IRQ_HANDLED; } #endif static inline void cas_handle_irq(struct net_device *dev, struct cas *cp, const u32 status) { /* housekeeping interrupts */ if (status & INTR_ERROR_MASK) cas_abnormal_irq(dev, cp, status); if (status & INTR_RX_BUF_UNAVAIL) { /* Frame arrived, no free RX buffers available. * NOTE: we can get this on a link transition. */ cas_post_rxds_ringN(cp, 0, 0); spin_lock(&cp->stat_lock[0]); cp->net_stats[0].rx_dropped++; spin_unlock(&cp->stat_lock[0]); } else if (status & INTR_RX_BUF_AE) { cas_post_rxds_ringN(cp, 0, RX_DESC_RINGN_SIZE(0) - RX_AE_FREEN_VAL(0)); } if (status & (INTR_RX_COMP_AF | INTR_RX_COMP_FULL)) cas_post_rxcs_ringN(dev, cp, 0); } static irqreturn_t cas_interrupt(int irq, void *dev_id) { struct net_device *dev = dev_id; struct cas *cp = netdev_priv(dev); unsigned long flags; u32 status = readl(cp->regs + REG_INTR_STATUS); if (status == 0) return IRQ_NONE; spin_lock_irqsave(&cp->lock, flags); if (status & (INTR_TX_ALL | INTR_TX_INTME)) { cas_tx(dev, cp, status); status &= ~(INTR_TX_ALL | INTR_TX_INTME); } if (status & INTR_RX_DONE) { #ifdef USE_NAPI cas_mask_intr(cp); napi_schedule(&cp->napi); #else cas_rx_ringN(cp, 0, 0); #endif status &= ~INTR_RX_DONE; } if (status) cas_handle_irq(dev, cp, status); spin_unlock_irqrestore(&cp->lock, flags); return IRQ_HANDLED; } #ifdef USE_NAPI static int cas_poll(struct napi_struct *napi, int budget) { struct cas *cp = container_of(napi, struct cas, napi); struct net_device *dev = cp->dev; int i, enable_intr, credits; u32 status = readl(cp->regs + REG_INTR_STATUS); unsigned long flags; spin_lock_irqsave(&cp->lock, flags); cas_tx(dev, cp, status); spin_unlock_irqrestore(&cp->lock, flags); /* NAPI rx packets. we spread the credits across all of the * rxc rings * * to make sure we're fair with the work we loop through each * ring N_RX_COMP_RING times with a request of * budget / N_RX_COMP_RINGS */ enable_intr = 1; credits = 0; for (i = 0; i < N_RX_COMP_RINGS; i++) { int j; for (j = 0; j < N_RX_COMP_RINGS; j++) { credits += cas_rx_ringN(cp, j, budget / N_RX_COMP_RINGS); if (credits >= budget) { enable_intr = 0; goto rx_comp; } } } rx_comp: /* final rx completion */ spin_lock_irqsave(&cp->lock, flags); if (status) cas_handle_irq(dev, cp, status); #ifdef USE_PCI_INTB if (N_RX_COMP_RINGS > 1) { status = readl(cp->regs + REG_PLUS_INTRN_STATUS(1)); if (status) cas_handle_irq1(dev, cp, status); } #endif #ifdef USE_PCI_INTC if (N_RX_COMP_RINGS > 2) { status = readl(cp->regs + REG_PLUS_INTRN_STATUS(2)); if (status) cas_handle_irqN(dev, cp, status, 2); } #endif #ifdef USE_PCI_INTD if (N_RX_COMP_RINGS > 3) { status = readl(cp->regs + REG_PLUS_INTRN_STATUS(3)); if (status) cas_handle_irqN(dev, cp, status, 3); } #endif spin_unlock_irqrestore(&cp->lock, flags); if (enable_intr) { napi_complete(napi); cas_unmask_intr(cp); } return credits; } #endif #ifdef CONFIG_NET_POLL_CONTROLLER static void cas_netpoll(struct net_device *dev) { struct cas *cp = netdev_priv(dev); cas_disable_irq(cp, 0); cas_interrupt(cp->pdev->irq, dev); cas_enable_irq(cp, 0); #ifdef USE_PCI_INTB if (N_RX_COMP_RINGS > 1) { /* cas_interrupt1(); */ } #endif #ifdef USE_PCI_INTC if (N_RX_COMP_RINGS > 2) { /* cas_interruptN(); */ } #endif #ifdef USE_PCI_INTD if (N_RX_COMP_RINGS > 3) { /* cas_interruptN(); */ } #endif } #endif static void cas_tx_timeout(struct net_device *dev) { struct cas *cp = netdev_priv(dev); netdev_err(dev, "transmit timed out, resetting\n"); if (!cp->hw_running) { netdev_err(dev, "hrm.. hw not running!\n"); return; } netdev_err(dev, "MIF_STATE[%08x]\n", readl(cp->regs + REG_MIF_STATE_MACHINE)); netdev_err(dev, "MAC_STATE[%08x]\n", readl(cp->regs + REG_MAC_STATE_MACHINE)); netdev_err(dev, "TX_STATE[%08x:%08x:%08x] FIFO[%08x:%08x:%08x] SM1[%08x] SM2[%08x]\n", readl(cp->regs + REG_TX_CFG), readl(cp->regs + REG_MAC_TX_STATUS), readl(cp->regs + REG_MAC_TX_CFG), readl(cp->regs + REG_TX_FIFO_PKT_CNT), readl(cp->regs + REG_TX_FIFO_WRITE_PTR), readl(cp->regs + REG_TX_FIFO_READ_PTR), readl(cp->regs + REG_TX_SM_1), readl(cp->regs + REG_TX_SM_2)); netdev_err(dev, "RX_STATE[%08x:%08x:%08x]\n", readl(cp->regs + REG_RX_CFG), readl(cp->regs + REG_MAC_RX_STATUS), readl(cp->regs + REG_MAC_RX_CFG)); netdev_err(dev, "HP_STATE[%08x:%08x:%08x:%08x]\n", readl(cp->regs + REG_HP_STATE_MACHINE), readl(cp->regs + REG_HP_STATUS0), readl(cp->regs + REG_HP_STATUS1), readl(cp->regs + REG_HP_STATUS2)); #if 1 atomic_inc(&cp->reset_task_pending); atomic_inc(&cp->reset_task_pending_all); schedule_work(&cp->reset_task); #else atomic_set(&cp->reset_task_pending, CAS_RESET_ALL); schedule_work(&cp->reset_task); #endif } static inline int cas_intme(int ring, int entry) { /* Algorithm: IRQ every 1/2 of descriptors. */ if (!(entry & ((TX_DESC_RINGN_SIZE(ring) >> 1) - 1))) return 1; return 0; } static void cas_write_txd(struct cas *cp, int ring, int entry, dma_addr_t mapping, int len, u64 ctrl, int last) { struct cas_tx_desc *txd = cp->init_txds[ring] + entry; ctrl |= CAS_BASE(TX_DESC_BUFLEN, len); if (cas_intme(ring, entry)) ctrl |= TX_DESC_INTME; if (last) ctrl |= TX_DESC_EOF; txd->control = cpu_to_le64(ctrl); txd->buffer = cpu_to_le64(mapping); } static inline void *tx_tiny_buf(struct cas *cp, const int ring, const int entry) { return cp->tx_tiny_bufs[ring] + TX_TINY_BUF_LEN*entry; } static inline dma_addr_t tx_tiny_map(struct cas *cp, const int ring, const int entry, const int tentry) { cp->tx_tiny_use[ring][tentry].nbufs++; cp->tx_tiny_use[ring][entry].used = 1; return cp->tx_tiny_dvma[ring] + TX_TINY_BUF_LEN*entry; } static inline int cas_xmit_tx_ringN(struct cas *cp, int ring, struct sk_buff *skb) { struct net_device *dev = cp->dev; int entry, nr_frags, frag, tabort, tentry; dma_addr_t mapping; unsigned long flags; u64 ctrl; u32 len; spin_lock_irqsave(&cp->tx_lock[ring], flags); /* This is a hard error, log it. */ if (TX_BUFFS_AVAIL(cp, ring) <= CAS_TABORT(cp)*(skb_shinfo(skb)->nr_frags + 1)) { netif_stop_queue(dev); spin_unlock_irqrestore(&cp->tx_lock[ring], flags); netdev_err(dev, "BUG! Tx Ring full when queue awake!\n"); return 1; } ctrl = 0; if (skb->ip_summed == CHECKSUM_PARTIAL) { const u64 csum_start_off = skb_checksum_start_offset(skb); const u64 csum_stuff_off = csum_start_off + skb->csum_offset; ctrl = TX_DESC_CSUM_EN | CAS_BASE(TX_DESC_CSUM_START, csum_start_off) | CAS_BASE(TX_DESC_CSUM_STUFF, csum_stuff_off); } entry = cp->tx_new[ring]; cp->tx_skbs[ring][entry] = skb; nr_frags = skb_shinfo(skb)->nr_frags; len = skb_headlen(skb); mapping = pci_map_page(cp->pdev, virt_to_page(skb->data), offset_in_page(skb->data), len, PCI_DMA_TODEVICE); tentry = entry; tabort = cas_calc_tabort(cp, (unsigned long) skb->data, len); if (unlikely(tabort)) { /* NOTE: len is always > tabort */ cas_write_txd(cp, ring, entry, mapping, len - tabort, ctrl | TX_DESC_SOF, 0); entry = TX_DESC_NEXT(ring, entry); skb_copy_from_linear_data_offset(skb, len - tabort, tx_tiny_buf(cp, ring, entry), tabort); mapping = tx_tiny_map(cp, ring, entry, tentry); cas_write_txd(cp, ring, entry, mapping, tabort, ctrl, (nr_frags == 0)); } else { cas_write_txd(cp, ring, entry, mapping, len, ctrl | TX_DESC_SOF, (nr_frags == 0)); } entry = TX_DESC_NEXT(ring, entry); for (frag = 0; frag < nr_frags; frag++) { const skb_frag_t *fragp = &skb_shinfo(skb)->frags[frag]; len = skb_frag_size(fragp); mapping = skb_frag_dma_map(&cp->pdev->dev, fragp, 0, len, DMA_TO_DEVICE); tabort = cas_calc_tabort(cp, skb_frag_off(fragp), len); if (unlikely(tabort)) { void *addr; /* NOTE: len is always > tabort */ cas_write_txd(cp, ring, entry, mapping, len - tabort, ctrl, 0); entry = TX_DESC_NEXT(ring, entry); addr = cas_page_map(skb_frag_page(fragp)); memcpy(tx_tiny_buf(cp, ring, entry), addr + skb_frag_off(fragp) + len - tabort, tabort); cas_page_unmap(addr); mapping = tx_tiny_map(cp, ring, entry, tentry); len = tabort; } cas_write_txd(cp, ring, entry, mapping, len, ctrl, (frag + 1 == nr_frags)); entry = TX_DESC_NEXT(ring, entry); } cp->tx_new[ring] = entry; if (TX_BUFFS_AVAIL(cp, ring) <= CAS_TABORT(cp)*(MAX_SKB_FRAGS + 1)) netif_stop_queue(dev); netif_printk(cp, tx_queued, KERN_DEBUG, dev, "tx[%d] queued, slot %d, skblen %d, avail %d\n", ring, entry, skb->len, TX_BUFFS_AVAIL(cp, ring)); writel(entry, cp->regs + REG_TX_KICKN(ring)); spin_unlock_irqrestore(&cp->tx_lock[ring], flags); return 0; } static netdev_tx_t cas_start_xmit(struct sk_buff *skb, struct net_device *dev) { struct cas *cp = netdev_priv(dev); /* this is only used as a load-balancing hint, so it doesn't * need to be SMP safe */ static int ring; if (skb_padto(skb, cp->min_frame_size)) return NETDEV_TX_OK; /* XXX: we need some higher-level QoS hooks to steer packets to * individual queues. */ if (cas_xmit_tx_ringN(cp, ring++ & N_TX_RINGS_MASK, skb)) return NETDEV_TX_BUSY; return NETDEV_TX_OK; } static void cas_init_tx_dma(struct cas *cp) { u64 desc_dma = cp->block_dvma; unsigned long off; u32 val; int i; /* set up tx completion writeback registers. must be 8-byte aligned */ #ifdef USE_TX_COMPWB off = offsetof(struct cas_init_block, tx_compwb); writel((desc_dma + off) >> 32, cp->regs + REG_TX_COMPWB_DB_HI); writel((desc_dma + off) & 0xffffffff, cp->regs + REG_TX_COMPWB_DB_LOW); #endif /* enable completion writebacks, enable paced mode, * disable read pipe, and disable pre-interrupt compwbs */ val = TX_CFG_COMPWB_Q1 | TX_CFG_COMPWB_Q2 | TX_CFG_COMPWB_Q3 | TX_CFG_COMPWB_Q4 | TX_CFG_DMA_RDPIPE_DIS | TX_CFG_PACED_MODE | TX_CFG_INTR_COMPWB_DIS; /* write out tx ring info and tx desc bases */ for (i = 0; i < MAX_TX_RINGS; i++) { off = (unsigned long) cp->init_txds[i] - (unsigned long) cp->init_block; val |= CAS_TX_RINGN_BASE(i); writel((desc_dma + off) >> 32, cp->regs + REG_TX_DBN_HI(i)); writel((desc_dma + off) & 0xffffffff, cp->regs + REG_TX_DBN_LOW(i)); /* don't zero out the kick register here as the system * will wedge */ } writel(val, cp->regs + REG_TX_CFG); /* program max burst sizes. these numbers should be different * if doing QoS. */ #ifdef USE_QOS writel(0x800, cp->regs + REG_TX_MAXBURST_0); writel(0x1600, cp->regs + REG_TX_MAXBURST_1); writel(0x2400, cp->regs + REG_TX_MAXBURST_2); writel(0x4800, cp->regs + REG_TX_MAXBURST_3); #else writel(0x800, cp->regs + REG_TX_MAXBURST_0); writel(0x800, cp->regs + REG_TX_MAXBURST_1); writel(0x800, cp->regs + REG_TX_MAXBURST_2); writel(0x800, cp->regs + REG_TX_MAXBURST_3); #endif } /* Must be invoked under cp->lock. */ static inline void cas_init_dma(struct cas *cp) { cas_init_tx_dma(cp); cas_init_rx_dma(cp); } static void cas_process_mc_list(struct cas *cp) { u16 hash_table[16]; u32 crc; struct netdev_hw_addr *ha; int i = 1; memset(hash_table, 0, sizeof(hash_table)); netdev_for_each_mc_addr(ha, cp->dev) { if (i <= CAS_MC_EXACT_MATCH_SIZE) { /* use the alternate mac address registers for the * first 15 multicast addresses */ writel((ha->addr[4] << 8) | ha->addr[5], cp->regs + REG_MAC_ADDRN(i*3 + 0)); writel((ha->addr[2] << 8) | ha->addr[3], cp->regs + REG_MAC_ADDRN(i*3 + 1)); writel((ha->addr[0] << 8) | ha->addr[1], cp->regs + REG_MAC_ADDRN(i*3 + 2)); i++; } else { /* use hw hash table for the next series of * multicast addresses */ crc = ether_crc_le(ETH_ALEN, ha->addr); crc >>= 24; hash_table[crc >> 4] |= 1 << (15 - (crc & 0xf)); } } for (i = 0; i < 16; i++) writel(hash_table[i], cp->regs + REG_MAC_HASH_TABLEN(i)); } /* Must be invoked under cp->lock. */ static u32 cas_setup_multicast(struct cas *cp) { u32 rxcfg = 0; int i; if (cp->dev->flags & IFF_PROMISC) { rxcfg |= MAC_RX_CFG_PROMISC_EN; } else if (cp->dev->flags & IFF_ALLMULTI) { for (i=0; i < 16; i++) writel(0xFFFF, cp->regs + REG_MAC_HASH_TABLEN(i)); rxcfg |= MAC_RX_CFG_HASH_FILTER_EN; } else { cas_process_mc_list(cp); rxcfg |= MAC_RX_CFG_HASH_FILTER_EN; } return rxcfg; } /* must be invoked under cp->stat_lock[N_TX_RINGS] */ static void cas_clear_mac_err(struct cas *cp) { writel(0, cp->regs + REG_MAC_COLL_NORMAL); writel(0, cp->regs + REG_MAC_COLL_FIRST); writel(0, cp->regs + REG_MAC_COLL_EXCESS); writel(0, cp->regs + REG_MAC_COLL_LATE); writel(0, cp->regs + REG_MAC_TIMER_DEFER); writel(0, cp->regs + REG_MAC_ATTEMPTS_PEAK); writel(0, cp->regs + REG_MAC_RECV_FRAME); writel(0, cp->regs + REG_MAC_LEN_ERR); writel(0, cp->regs + REG_MAC_ALIGN_ERR); writel(0, cp->regs + REG_MAC_FCS_ERR); writel(0, cp->regs + REG_MAC_RX_CODE_ERR); } static void cas_mac_reset(struct cas *cp) { int i; /* do both TX and RX reset */ writel(0x1, cp->regs + REG_MAC_TX_RESET); writel(0x1, cp->regs + REG_MAC_RX_RESET); /* wait for TX */ i = STOP_TRIES; while (i-- > 0) { if (readl(cp->regs + REG_MAC_TX_RESET) == 0) break; udelay(10); } /* wait for RX */ i = STOP_TRIES; while (i-- > 0) { if (readl(cp->regs + REG_MAC_RX_RESET) == 0) break; udelay(10); } if (readl(cp->regs + REG_MAC_TX_RESET) | readl(cp->regs + REG_MAC_RX_RESET)) netdev_err(cp->dev, "mac tx[%d]/rx[%d] reset failed [%08x]\n", readl(cp->regs + REG_MAC_TX_RESET), readl(cp->regs + REG_MAC_RX_RESET), readl(cp->regs + REG_MAC_STATE_MACHINE)); } /* Must be invoked under cp->lock. */ static void cas_init_mac(struct cas *cp) { unsigned char *e = &cp->dev->dev_addr[0]; int i; cas_mac_reset(cp); /* setup core arbitration weight register */ writel(CAWR_RR_DIS, cp->regs + REG_CAWR); #if !defined(CONFIG_SPARC64) && !defined(CONFIG_ALPHA) /* set the infinite burst register for chips that don't have * pci issues. */ if ((cp->cas_flags & CAS_FLAG_TARGET_ABORT) == 0) writel(INF_BURST_EN, cp->regs + REG_INF_BURST); #endif writel(0x1BF0, cp->regs + REG_MAC_SEND_PAUSE); writel(0x00, cp->regs + REG_MAC_IPG0); writel(0x08, cp->regs + REG_MAC_IPG1); writel(0x04, cp->regs + REG_MAC_IPG2); /* change later for 802.3z */ writel(0x40, cp->regs + REG_MAC_SLOT_TIME); /* min frame + FCS */ writel(ETH_ZLEN + 4, cp->regs + REG_MAC_FRAMESIZE_MIN); /* Ethernet payload + header + FCS + optional VLAN tag. NOTE: we * specify the maximum frame size to prevent RX tag errors on * oversized frames. */ writel(CAS_BASE(MAC_FRAMESIZE_MAX_BURST, 0x2000) | CAS_BASE(MAC_FRAMESIZE_MAX_FRAME, (CAS_MAX_MTU + ETH_HLEN + 4 + 4)), cp->regs + REG_MAC_FRAMESIZE_MAX); /* NOTE: crc_size is used as a surrogate for half-duplex. * workaround saturn half-duplex issue by increasing preamble * size to 65 bytes. */ if ((cp->cas_flags & CAS_FLAG_SATURN) && cp->crc_size) writel(0x41, cp->regs + REG_MAC_PA_SIZE); else writel(0x07, cp->regs + REG_MAC_PA_SIZE); writel(0x04, cp->regs + REG_MAC_JAM_SIZE); writel(0x10, cp->regs + REG_MAC_ATTEMPT_LIMIT); writel(0x8808, cp->regs + REG_MAC_CTRL_TYPE); writel((e[5] | (e[4] << 8)) & 0x3ff, cp->regs + REG_MAC_RANDOM_SEED); writel(0, cp->regs + REG_MAC_ADDR_FILTER0); writel(0, cp->regs + REG_MAC_ADDR_FILTER1); writel(0, cp->regs + REG_MAC_ADDR_FILTER2); writel(0, cp->regs + REG_MAC_ADDR_FILTER2_1_MASK); writel(0, cp->regs + REG_MAC_ADDR_FILTER0_MASK); /* setup mac address in perfect filter array */ for (i = 0; i < 45; i++) writel(0x0, cp->regs + REG_MAC_ADDRN(i)); writel((e[4] << 8) | e[5], cp->regs + REG_MAC_ADDRN(0)); writel((e[2] << 8) | e[3], cp->regs + REG_MAC_ADDRN(1)); writel((e[0] << 8) | e[1], cp->regs + REG_MAC_ADDRN(2)); writel(0x0001, cp->regs + REG_MAC_ADDRN(42)); writel(0xc200, cp->regs + REG_MAC_ADDRN(43)); writel(0x0180, cp->regs + REG_MAC_ADDRN(44)); cp->mac_rx_cfg = cas_setup_multicast(cp); spin_lock(&cp->stat_lock[N_TX_RINGS]); cas_clear_mac_err(cp); spin_unlock(&cp->stat_lock[N_TX_RINGS]); /* Setup MAC interrupts. We want to get all of the interesting * counter expiration events, but we do not want to hear about * normal rx/tx as the DMA engine tells us that. */ writel(MAC_TX_FRAME_XMIT, cp->regs + REG_MAC_TX_MASK); writel(MAC_RX_FRAME_RECV, cp->regs + REG_MAC_RX_MASK); /* Don't enable even the PAUSE interrupts for now, we * make no use of those events other than to record them. */ writel(0xffffffff, cp->regs + REG_MAC_CTRL_MASK); } /* Must be invoked under cp->lock. */ static void cas_init_pause_thresholds(struct cas *cp) { /* Calculate pause thresholds. Setting the OFF threshold to the * full RX fifo size effectively disables PAUSE generation */ if (cp->rx_fifo_size <= (2 * 1024)) { cp->rx_pause_off = cp->rx_pause_on = cp->rx_fifo_size; } else { int max_frame = (cp->dev->mtu + ETH_HLEN + 4 + 4 + 64) & ~63; if (max_frame * 3 > cp->rx_fifo_size) { cp->rx_pause_off = 7104; cp->rx_pause_on = 960; } else { int off = (cp->rx_fifo_size - (max_frame * 2)); int on = off - max_frame; cp->rx_pause_off = off; cp->rx_pause_on = on; } } } static int cas_vpd_match(const void __iomem *p, const char *str) { int len = strlen(str) + 1; int i; for (i = 0; i < len; i++) { if (readb(p + i) != str[i]) return 0; } return 1; } /* get the mac address by reading the vpd information in the rom. * also get the phy type and determine if there's an entropy generator. * NOTE: this is a bit convoluted for the following reasons: * 1) vpd info has order-dependent mac addresses for multinic cards * 2) the only way to determine the nic order is to use the slot * number. * 3) fiber cards don't have bridges, so their slot numbers don't * mean anything. * 4) we don't actually know we have a fiber card until after * the mac addresses are parsed. */ static int cas_get_vpd_info(struct cas *cp, unsigned char *dev_addr, const int offset) { void __iomem *p = cp->regs + REG_EXPANSION_ROM_RUN_START; void __iomem *base, *kstart; int i, len; int found = 0; #define VPD_FOUND_MAC 0x01 #define VPD_FOUND_PHY 0x02 int phy_type = CAS_PHY_MII_MDIO0; /* default phy type */ int mac_off = 0; #if defined(CONFIG_SPARC) const unsigned char *addr; #endif /* give us access to the PROM */ writel(BIM_LOCAL_DEV_PROM | BIM_LOCAL_DEV_PAD, cp->regs + REG_BIM_LOCAL_DEV_EN); /* check for an expansion rom */ if (readb(p) != 0x55 || readb(p + 1) != 0xaa) goto use_random_mac_addr; /* search for beginning of vpd */ base = NULL; for (i = 2; i < EXPANSION_ROM_SIZE; i++) { /* check for PCIR */ if ((readb(p + i + 0) == 0x50) && (readb(p + i + 1) == 0x43) && (readb(p + i + 2) == 0x49) && (readb(p + i + 3) == 0x52)) { base = p + (readb(p + i + 8) | (readb(p + i + 9) << 8)); break; } } if (!base || (readb(base) != 0x82)) goto use_random_mac_addr; i = (readb(base + 1) | (readb(base + 2) << 8)) + 3; while (i < EXPANSION_ROM_SIZE) { if (readb(base + i) != 0x90) /* no vpd found */ goto use_random_mac_addr; /* found a vpd field */ len = readb(base + i + 1) | (readb(base + i + 2) << 8); /* extract keywords */ kstart = base + i + 3; p = kstart; while ((p - kstart) < len) { int klen = readb(p + 2); int j; char type; p += 3; /* look for the following things: * -- correct length == 29 * 3 (type) + 2 (size) + * 18 (strlen("local-mac-address") + 1) + * 6 (mac addr) * -- VPD Instance 'I' * -- VPD Type Bytes 'B' * -- VPD data length == 6 * -- property string == local-mac-address * * -- correct length == 24 * 3 (type) + 2 (size) + * 12 (strlen("entropy-dev") + 1) + * 7 (strlen("vms110") + 1) * -- VPD Instance 'I' * -- VPD Type String 'B' * -- VPD data length == 7 * -- property string == entropy-dev * * -- correct length == 18 * 3 (type) + 2 (size) + * 9 (strlen("phy-type") + 1) + * 4 (strlen("pcs") + 1) * -- VPD Instance 'I' * -- VPD Type String 'S' * -- VPD data length == 4 * -- property string == phy-type * * -- correct length == 23 * 3 (type) + 2 (size) + * 14 (strlen("phy-interface") + 1) + * 4 (strlen("pcs") + 1) * -- VPD Instance 'I' * -- VPD Type String 'S' * -- VPD data length == 4 * -- property string == phy-interface */ if (readb(p) != 'I') goto next; /* finally, check string and length */ type = readb(p + 3); if (type == 'B') { if ((klen == 29) && readb(p + 4) == 6 && cas_vpd_match(p + 5, "local-mac-address")) { if (mac_off++ > offset) goto next; /* set mac address */ for (j = 0; j < 6; j++) dev_addr[j] = readb(p + 23 + j); goto found_mac; } } if (type != 'S') goto next; #ifdef USE_ENTROPY_DEV if ((klen == 24) && cas_vpd_match(p + 5, "entropy-dev") && cas_vpd_match(p + 17, "vms110")) { cp->cas_flags |= CAS_FLAG_ENTROPY_DEV; goto next; } #endif if (found & VPD_FOUND_PHY) goto next; if ((klen == 18) && readb(p + 4) == 4 && cas_vpd_match(p + 5, "phy-type")) { if (cas_vpd_match(p + 14, "pcs")) { phy_type = CAS_PHY_SERDES; goto found_phy; } } if ((klen == 23) && readb(p + 4) == 4 && cas_vpd_match(p + 5, "phy-interface")) { if (cas_vpd_match(p + 19, "pcs")) { phy_type = CAS_PHY_SERDES; goto found_phy; } } found_mac: found |= VPD_FOUND_MAC; goto next; found_phy: found |= VPD_FOUND_PHY; next: p += klen; } i += len + 3; } use_random_mac_addr: if (found & VPD_FOUND_MAC) goto done; #if defined(CONFIG_SPARC) addr = of_get_property(cp->of_node, "local-mac-address", NULL); if (addr != NULL) { memcpy(dev_addr, addr, ETH_ALEN); goto done; } #endif /* Sun MAC prefix then 3 random bytes. */ pr_info("MAC address not found in ROM VPD\n"); dev_addr[0] = 0x08; dev_addr[1] = 0x00; dev_addr[2] = 0x20; get_random_bytes(dev_addr + 3, 3); done: writel(0, cp->regs + REG_BIM_LOCAL_DEV_EN); return phy_type; } /* check pci invariants */ static void cas_check_pci_invariants(struct cas *cp) { struct pci_dev *pdev = cp->pdev; cp->cas_flags = 0; if ((pdev->vendor == PCI_VENDOR_ID_SUN) && (pdev->device == PCI_DEVICE_ID_SUN_CASSINI)) { if (pdev->revision >= CAS_ID_REVPLUS) cp->cas_flags |= CAS_FLAG_REG_PLUS; if (pdev->revision < CAS_ID_REVPLUS02u) cp->cas_flags |= CAS_FLAG_TARGET_ABORT; /* Original Cassini supports HW CSUM, but it's not * enabled by default as it can trigger TX hangs. */ if (pdev->revision < CAS_ID_REV2) cp->cas_flags |= CAS_FLAG_NO_HW_CSUM; } else { /* Only sun has original cassini chips. */ cp->cas_flags |= CAS_FLAG_REG_PLUS; /* We use a flag because the same phy might be externally * connected. */ if ((pdev->vendor == PCI_VENDOR_ID_NS) && (pdev->device == PCI_DEVICE_ID_NS_SATURN)) cp->cas_flags |= CAS_FLAG_SATURN; } } static int cas_check_invariants(struct cas *cp) { struct pci_dev *pdev = cp->pdev; u32 cfg; int i; /* get page size for rx buffers. */ cp->page_order = 0; #ifdef USE_PAGE_ORDER if (PAGE_SHIFT < CAS_JUMBO_PAGE_SHIFT) { /* see if we can allocate larger pages */ struct page *page = alloc_pages(GFP_ATOMIC, CAS_JUMBO_PAGE_SHIFT - PAGE_SHIFT); if (page) { __free_pages(page, CAS_JUMBO_PAGE_SHIFT - PAGE_SHIFT); cp->page_order = CAS_JUMBO_PAGE_SHIFT - PAGE_SHIFT; } else { printk("MTU limited to %d bytes\n", CAS_MAX_MTU); } } #endif cp->page_size = (PAGE_SIZE << cp->page_order); /* Fetch the FIFO configurations. */ cp->tx_fifo_size = readl(cp->regs + REG_TX_FIFO_SIZE) * 64; cp->rx_fifo_size = RX_FIFO_SIZE; /* finish phy determination. MDIO1 takes precedence over MDIO0 if * they're both connected. */ cp->phy_type = cas_get_vpd_info(cp, cp->dev->dev_addr, PCI_SLOT(pdev->devfn)); if (cp->phy_type & CAS_PHY_SERDES) { cp->cas_flags |= CAS_FLAG_1000MB_CAP; return 0; /* no more checking needed */ } /* MII */ cfg = readl(cp->regs + REG_MIF_CFG); if (cfg & MIF_CFG_MDIO_1) { cp->phy_type = CAS_PHY_MII_MDIO1; } else if (cfg & MIF_CFG_MDIO_0) { cp->phy_type = CAS_PHY_MII_MDIO0; } cas_mif_poll(cp, 0); writel(PCS_DATAPATH_MODE_MII, cp->regs + REG_PCS_DATAPATH_MODE); for (i = 0; i < 32; i++) { u32 phy_id; int j; for (j = 0; j < 3; j++) { cp->phy_addr = i; phy_id = cas_phy_read(cp, MII_PHYSID1) << 16; phy_id |= cas_phy_read(cp, MII_PHYSID2); if (phy_id && (phy_id != 0xFFFFFFFF)) { cp->phy_id = phy_id; goto done; } } } pr_err("MII phy did not respond [%08x]\n", readl(cp->regs + REG_MIF_STATE_MACHINE)); return -1; done: /* see if we can do gigabit */ cfg = cas_phy_read(cp, MII_BMSR); if ((cfg & CAS_BMSR_1000_EXTEND) && cas_phy_read(cp, CAS_MII_1000_EXTEND)) cp->cas_flags |= CAS_FLAG_1000MB_CAP; return 0; } /* Must be invoked under cp->lock. */ static inline void cas_start_dma(struct cas *cp) { int i; u32 val; int txfailed = 0; /* enable dma */ val = readl(cp->regs + REG_TX_CFG) | TX_CFG_DMA_EN; writel(val, cp->regs + REG_TX_CFG); val = readl(cp->regs + REG_RX_CFG) | RX_CFG_DMA_EN; writel(val, cp->regs + REG_RX_CFG); /* enable the mac */ val = readl(cp->regs + REG_MAC_TX_CFG) | MAC_TX_CFG_EN; writel(val, cp->regs + REG_MAC_TX_CFG); val = readl(cp->regs + REG_MAC_RX_CFG) | MAC_RX_CFG_EN; writel(val, cp->regs + REG_MAC_RX_CFG); i = STOP_TRIES; while (i-- > 0) { val = readl(cp->regs + REG_MAC_TX_CFG); if ((val & MAC_TX_CFG_EN)) break; udelay(10); } if (i < 0) txfailed = 1; i = STOP_TRIES; while (i-- > 0) { val = readl(cp->regs + REG_MAC_RX_CFG); if ((val & MAC_RX_CFG_EN)) { if (txfailed) { netdev_err(cp->dev, "enabling mac failed [tx:%08x:%08x]\n", readl(cp->regs + REG_MIF_STATE_MACHINE), readl(cp->regs + REG_MAC_STATE_MACHINE)); } goto enable_rx_done; } udelay(10); } netdev_err(cp->dev, "enabling mac failed [%s:%08x:%08x]\n", (txfailed ? "tx,rx" : "rx"), readl(cp->regs + REG_MIF_STATE_MACHINE), readl(cp->regs + REG_MAC_STATE_MACHINE)); enable_rx_done: cas_unmask_intr(cp); /* enable interrupts */ writel(RX_DESC_RINGN_SIZE(0) - 4, cp->regs + REG_RX_KICK); writel(0, cp->regs + REG_RX_COMP_TAIL); if (cp->cas_flags & CAS_FLAG_REG_PLUS) { if (N_RX_DESC_RINGS > 1) writel(RX_DESC_RINGN_SIZE(1) - 4, cp->regs + REG_PLUS_RX_KICK1); for (i = 1; i < N_RX_COMP_RINGS; i++) writel(0, cp->regs + REG_PLUS_RX_COMPN_TAIL(i)); } } /* Must be invoked under cp->lock. */ static void cas_read_pcs_link_mode(struct cas *cp, int *fd, int *spd, int *pause) { u32 val = readl(cp->regs + REG_PCS_MII_LPA); *fd = (val & PCS_MII_LPA_FD) ? 1 : 0; *pause = (val & PCS_MII_LPA_SYM_PAUSE) ? 0x01 : 0x00; if (val & PCS_MII_LPA_ASYM_PAUSE) *pause |= 0x10; *spd = 1000; } /* Must be invoked under cp->lock. */ static void cas_read_mii_link_mode(struct cas *cp, int *fd, int *spd, int *pause) { u32 val; *fd = 0; *spd = 10; *pause = 0; /* use GMII registers */ val = cas_phy_read(cp, MII_LPA); if (val & CAS_LPA_PAUSE) *pause = 0x01; if (val & CAS_LPA_ASYM_PAUSE) *pause |= 0x10; if (val & LPA_DUPLEX) *fd = 1; if (val & LPA_100) *spd = 100; if (cp->cas_flags & CAS_FLAG_1000MB_CAP) { val = cas_phy_read(cp, CAS_MII_1000_STATUS); if (val & (CAS_LPA_1000FULL | CAS_LPA_1000HALF)) *spd = 1000; if (val & CAS_LPA_1000FULL) *fd = 1; } } /* A link-up condition has occurred, initialize and enable the * rest of the chip. * * Must be invoked under cp->lock. */ static void cas_set_link_modes(struct cas *cp) { u32 val; int full_duplex, speed, pause; full_duplex = 0; speed = 10; pause = 0; if (CAS_PHY_MII(cp->phy_type)) { cas_mif_poll(cp, 0); val = cas_phy_read(cp, MII_BMCR); if (val & BMCR_ANENABLE) { cas_read_mii_link_mode(cp, &full_duplex, &speed, &pause); } else { if (val & BMCR_FULLDPLX) full_duplex = 1; if (val & BMCR_SPEED100) speed = 100; else if (val & CAS_BMCR_SPEED1000) speed = (cp->cas_flags & CAS_FLAG_1000MB_CAP) ? 1000 : 100; } cas_mif_poll(cp, 1); } else { val = readl(cp->regs + REG_PCS_MII_CTRL); cas_read_pcs_link_mode(cp, &full_duplex, &speed, &pause); if ((val & PCS_MII_AUTONEG_EN) == 0) { if (val & PCS_MII_CTRL_DUPLEX) full_duplex = 1; } } netif_info(cp, link, cp->dev, "Link up at %d Mbps, %s-duplex\n", speed, full_duplex ? "full" : "half"); val = MAC_XIF_TX_MII_OUTPUT_EN | MAC_XIF_LINK_LED; if (CAS_PHY_MII(cp->phy_type)) { val |= MAC_XIF_MII_BUFFER_OUTPUT_EN; if (!full_duplex) val |= MAC_XIF_DISABLE_ECHO; } if (full_duplex) val |= MAC_XIF_FDPLX_LED; if (speed == 1000) val |= MAC_XIF_GMII_MODE; writel(val, cp->regs + REG_MAC_XIF_CFG); /* deal with carrier and collision detect. */ val = MAC_TX_CFG_IPG_EN; if (full_duplex) { val |= MAC_TX_CFG_IGNORE_CARRIER; val |= MAC_TX_CFG_IGNORE_COLL; } else { #ifndef USE_CSMA_CD_PROTO val |= MAC_TX_CFG_NEVER_GIVE_UP_EN; val |= MAC_TX_CFG_NEVER_GIVE_UP_LIM; #endif } /* val now set up for REG_MAC_TX_CFG */ /* If gigabit and half-duplex, enable carrier extension * mode. increase slot time to 512 bytes as well. * else, disable it and make sure slot time is 64 bytes. * also activate checksum bug workaround */ if ((speed == 1000) && !full_duplex) { writel(val | MAC_TX_CFG_CARRIER_EXTEND, cp->regs + REG_MAC_TX_CFG); val = readl(cp->regs + REG_MAC_RX_CFG); val &= ~MAC_RX_CFG_STRIP_FCS; /* checksum workaround */ writel(val | MAC_RX_CFG_CARRIER_EXTEND, cp->regs + REG_MAC_RX_CFG); writel(0x200, cp->regs + REG_MAC_SLOT_TIME); cp->crc_size = 4; /* minimum size gigabit frame at half duplex */ cp->min_frame_size = CAS_1000MB_MIN_FRAME; } else { writel(val, cp->regs + REG_MAC_TX_CFG); /* checksum bug workaround. don't strip FCS when in * half-duplex mode */ val = readl(cp->regs + REG_MAC_RX_CFG); if (full_duplex) { val |= MAC_RX_CFG_STRIP_FCS; cp->crc_size = 0; cp->min_frame_size = CAS_MIN_MTU; } else { val &= ~MAC_RX_CFG_STRIP_FCS; cp->crc_size = 4; cp->min_frame_size = CAS_MIN_FRAME; } writel(val & ~MAC_RX_CFG_CARRIER_EXTEND, cp->regs + REG_MAC_RX_CFG); writel(0x40, cp->regs + REG_MAC_SLOT_TIME); } if (netif_msg_link(cp)) { if (pause & 0x01) { netdev_info(cp->dev, "Pause is enabled (rxfifo: %d off: %d on: %d)\n", cp->rx_fifo_size, cp->rx_pause_off, cp->rx_pause_on); } else if (pause & 0x10) { netdev_info(cp->dev, "TX pause enabled\n"); } else { netdev_info(cp->dev, "Pause is disabled\n"); } } val = readl(cp->regs + REG_MAC_CTRL_CFG); val &= ~(MAC_CTRL_CFG_SEND_PAUSE_EN | MAC_CTRL_CFG_RECV_PAUSE_EN); if (pause) { /* symmetric or asymmetric pause */ val |= MAC_CTRL_CFG_SEND_PAUSE_EN; if (pause & 0x01) { /* symmetric pause */ val |= MAC_CTRL_CFG_RECV_PAUSE_EN; } } writel(val, cp->regs + REG_MAC_CTRL_CFG); cas_start_dma(cp); } /* Must be invoked under cp->lock. */ static void cas_init_hw(struct cas *cp, int restart_link) { if (restart_link) cas_phy_init(cp); cas_init_pause_thresholds(cp); cas_init_mac(cp); cas_init_dma(cp); if (restart_link) { /* Default aneg parameters */ cp->timer_ticks = 0; cas_begin_auto_negotiation(cp, NULL); } else if (cp->lstate == link_up) { cas_set_link_modes(cp); netif_carrier_on(cp->dev); } } /* Must be invoked under cp->lock. on earlier cassini boards, * SOFT_0 is tied to PCI reset. we use this to force a pci reset, * let it settle out, and then restore pci state. */ static void cas_hard_reset(struct cas *cp) { writel(BIM_LOCAL_DEV_SOFT_0, cp->regs + REG_BIM_LOCAL_DEV_EN); udelay(20); pci_restore_state(cp->pdev); } static void cas_global_reset(struct cas *cp, int blkflag) { int limit; /* issue a global reset. don't use RSTOUT. */ if (blkflag && !CAS_PHY_MII(cp->phy_type)) { /* For PCS, when the blkflag is set, we should set the * SW_REST_BLOCK_PCS_SLINK bit to prevent the results of * the last autonegotiation from being cleared. We'll * need some special handling if the chip is set into a * loopback mode. */ writel((SW_RESET_TX | SW_RESET_RX | SW_RESET_BLOCK_PCS_SLINK), cp->regs + REG_SW_RESET); } else { writel(SW_RESET_TX | SW_RESET_RX, cp->regs + REG_SW_RESET); } /* need to wait at least 3ms before polling register */ mdelay(3); limit = STOP_TRIES; while (limit-- > 0) { u32 val = readl(cp->regs + REG_SW_RESET); if ((val & (SW_RESET_TX | SW_RESET_RX)) == 0) goto done; udelay(10); } netdev_err(cp->dev, "sw reset failed\n"); done: /* enable various BIM interrupts */ writel(BIM_CFG_DPAR_INTR_ENABLE | BIM_CFG_RMA_INTR_ENABLE | BIM_CFG_RTA_INTR_ENABLE, cp->regs + REG_BIM_CFG); /* clear out pci error status mask for handled errors. * we don't deal with DMA counter overflows as they happen * all the time. */ writel(0xFFFFFFFFU & ~(PCI_ERR_BADACK | PCI_ERR_DTRTO | PCI_ERR_OTHER | PCI_ERR_BIM_DMA_WRITE | PCI_ERR_BIM_DMA_READ), cp->regs + REG_PCI_ERR_STATUS_MASK); /* set up for MII by default to address mac rx reset timeout * issue */ writel(PCS_DATAPATH_MODE_MII, cp->regs + REG_PCS_DATAPATH_MODE); } static void cas_reset(struct cas *cp, int blkflag) { u32 val; cas_mask_intr(cp); cas_global_reset(cp, blkflag); cas_mac_reset(cp); cas_entropy_reset(cp); /* disable dma engines. */ val = readl(cp->regs + REG_TX_CFG); val &= ~TX_CFG_DMA_EN; writel(val, cp->regs + REG_TX_CFG); val = readl(cp->regs + REG_RX_CFG); val &= ~RX_CFG_DMA_EN; writel(val, cp->regs + REG_RX_CFG); /* program header parser */ if ((cp->cas_flags & CAS_FLAG_TARGET_ABORT) || (CAS_HP_ALT_FIRMWARE == cas_prog_null)) { cas_load_firmware(cp, CAS_HP_FIRMWARE); } else { cas_load_firmware(cp, CAS_HP_ALT_FIRMWARE); } /* clear out error registers */ spin_lock(&cp->stat_lock[N_TX_RINGS]); cas_clear_mac_err(cp); spin_unlock(&cp->stat_lock[N_TX_RINGS]); } /* Shut down the chip, must be called with pm_mutex held. */ static void cas_shutdown(struct cas *cp) { unsigned long flags; /* Make us not-running to avoid timers respawning */ cp->hw_running = 0; del_timer_sync(&cp->link_timer); /* Stop the reset task */ #if 0 while (atomic_read(&cp->reset_task_pending_mtu) || atomic_read(&cp->reset_task_pending_spare) || atomic_read(&cp->reset_task_pending_all)) schedule(); #else while (atomic_read(&cp->reset_task_pending)) schedule(); #endif /* Actually stop the chip */ cas_lock_all_save(cp, flags); cas_reset(cp, 0); if (cp->cas_flags & CAS_FLAG_SATURN) cas_phy_powerdown(cp); cas_unlock_all_restore(cp, flags); } static int cas_change_mtu(struct net_device *dev, int new_mtu) { struct cas *cp = netdev_priv(dev); dev->mtu = new_mtu; if (!netif_running(dev) || !netif_device_present(dev)) return 0; /* let the reset task handle it */ #if 1 atomic_inc(&cp->reset_task_pending); if ((cp->phy_type & CAS_PHY_SERDES)) { atomic_inc(&cp->reset_task_pending_all); } else { atomic_inc(&cp->reset_task_pending_mtu); } schedule_work(&cp->reset_task); #else atomic_set(&cp->reset_task_pending, (cp->phy_type & CAS_PHY_SERDES) ? CAS_RESET_ALL : CAS_RESET_MTU); pr_err("reset called in cas_change_mtu\n"); schedule_work(&cp->reset_task); #endif flush_work(&cp->reset_task); return 0; } static void cas_clean_txd(struct cas *cp, int ring) { struct cas_tx_desc *txd = cp->init_txds[ring]; struct sk_buff *skb, **skbs = cp->tx_skbs[ring]; u64 daddr, dlen; int i, size; size = TX_DESC_RINGN_SIZE(ring); for (i = 0; i < size; i++) { int frag; if (skbs[i] == NULL) continue; skb = skbs[i]; skbs[i] = NULL; for (frag = 0; frag <= skb_shinfo(skb)->nr_frags; frag++) { int ent = i & (size - 1); /* first buffer is never a tiny buffer and so * needs to be unmapped. */ daddr = le64_to_cpu(txd[ent].buffer); dlen = CAS_VAL(TX_DESC_BUFLEN, le64_to_cpu(txd[ent].control)); pci_unmap_page(cp->pdev, daddr, dlen, PCI_DMA_TODEVICE); if (frag != skb_shinfo(skb)->nr_frags) { i++; /* next buffer might by a tiny buffer. * skip past it. */ ent = i & (size - 1); if (cp->tx_tiny_use[ring][ent].used) i++; } } dev_kfree_skb_any(skb); } /* zero out tiny buf usage */ memset(cp->tx_tiny_use[ring], 0, size*sizeof(*cp->tx_tiny_use[ring])); } /* freed on close */ static inline void cas_free_rx_desc(struct cas *cp, int ring) { cas_page_t **page = cp->rx_pages[ring]; int i, size; size = RX_DESC_RINGN_SIZE(ring); for (i = 0; i < size; i++) { if (page[i]) { cas_page_free(cp, page[i]); page[i] = NULL; } } } static void cas_free_rxds(struct cas *cp) { int i; for (i = 0; i < N_RX_DESC_RINGS; i++) cas_free_rx_desc(cp, i); } /* Must be invoked under cp->lock. */ static void cas_clean_rings(struct cas *cp) { int i; /* need to clean all tx rings */ memset(cp->tx_old, 0, sizeof(*cp->tx_old)*N_TX_RINGS); memset(cp->tx_new, 0, sizeof(*cp->tx_new)*N_TX_RINGS); for (i = 0; i < N_TX_RINGS; i++) cas_clean_txd(cp, i); /* zero out init block */ memset(cp->init_block, 0, sizeof(struct cas_init_block)); cas_clean_rxds(cp); cas_clean_rxcs(cp); } /* allocated on open */ static inline int cas_alloc_rx_desc(struct cas *cp, int ring) { cas_page_t **page = cp->rx_pages[ring]; int size, i = 0; size = RX_DESC_RINGN_SIZE(ring); for (i = 0; i < size; i++) { if ((page[i] = cas_page_alloc(cp, GFP_KERNEL)) == NULL) return -1; } return 0; } static int cas_alloc_rxds(struct cas *cp) { int i; for (i = 0; i < N_RX_DESC_RINGS; i++) { if (cas_alloc_rx_desc(cp, i) < 0) { cas_free_rxds(cp); return -1; } } return 0; } static void cas_reset_task(struct work_struct *work) { struct cas *cp = container_of(work, struct cas, reset_task); #if 0 int pending = atomic_read(&cp->reset_task_pending); #else int pending_all = atomic_read(&cp->reset_task_pending_all); int pending_spare = atomic_read(&cp->reset_task_pending_spare); int pending_mtu = atomic_read(&cp->reset_task_pending_mtu); if (pending_all == 0 && pending_spare == 0 && pending_mtu == 0) { /* We can have more tasks scheduled than actually * needed. */ atomic_dec(&cp->reset_task_pending); return; } #endif /* The link went down, we reset the ring, but keep * DMA stopped. Use this function for reset * on error as well. */ if (cp->hw_running) { unsigned long flags; /* Make sure we don't get interrupts or tx packets */ netif_device_detach(cp->dev); cas_lock_all_save(cp, flags); if (cp->opened) { /* We call cas_spare_recover when we call cas_open. * but we do not initialize the lists cas_spare_recover * uses until cas_open is called. */ cas_spare_recover(cp, GFP_ATOMIC); } #if 1 /* test => only pending_spare set */ if (!pending_all && !pending_mtu) goto done; #else if (pending == CAS_RESET_SPARE) goto done; #endif /* when pending == CAS_RESET_ALL, the following * call to cas_init_hw will restart auto negotiation. * Setting the second argument of cas_reset to * !(pending == CAS_RESET_ALL) will set this argument * to 1 (avoiding reinitializing the PHY for the normal * PCS case) when auto negotiation is not restarted. */ #if 1 cas_reset(cp, !(pending_all > 0)); if (cp->opened) cas_clean_rings(cp); cas_init_hw(cp, (pending_all > 0)); #else cas_reset(cp, !(pending == CAS_RESET_ALL)); if (cp->opened) cas_clean_rings(cp); cas_init_hw(cp, pending == CAS_RESET_ALL); #endif done: cas_unlock_all_restore(cp, flags); netif_device_attach(cp->dev); } #if 1 atomic_sub(pending_all, &cp->reset_task_pending_all); atomic_sub(pending_spare, &cp->reset_task_pending_spare); atomic_sub(pending_mtu, &cp->reset_task_pending_mtu); atomic_dec(&cp->reset_task_pending); #else atomic_set(&cp->reset_task_pending, 0); #endif } static void cas_link_timer(struct timer_list *t) { struct cas *cp = from_timer(cp, t, link_timer); int mask, pending = 0, reset = 0; unsigned long flags; if (link_transition_timeout != 0 && cp->link_transition_jiffies_valid && ((jiffies - cp->link_transition_jiffies) > (link_transition_timeout))) { /* One-second counter so link-down workaround doesn't * cause resets to occur so fast as to fool the switch * into thinking the link is down. */ cp->link_transition_jiffies_valid = 0; } if (!cp->hw_running) return; spin_lock_irqsave(&cp->lock, flags); cas_lock_tx(cp); cas_entropy_gather(cp); /* If the link task is still pending, we just * reschedule the link timer */ #if 1 if (atomic_read(&cp->reset_task_pending_all) || atomic_read(&cp->reset_task_pending_spare) || atomic_read(&cp->reset_task_pending_mtu)) goto done; #else if (atomic_read(&cp->reset_task_pending)) goto done; #endif /* check for rx cleaning */ if ((mask = (cp->cas_flags & CAS_FLAG_RXD_POST_MASK))) { int i, rmask; for (i = 0; i < MAX_RX_DESC_RINGS; i++) { rmask = CAS_FLAG_RXD_POST(i); if ((mask & rmask) == 0) continue; /* post_rxds will do a mod_timer */ if (cas_post_rxds_ringN(cp, i, cp->rx_last[i]) < 0) { pending = 1; continue; } cp->cas_flags &= ~rmask; } } if (CAS_PHY_MII(cp->phy_type)) { u16 bmsr; cas_mif_poll(cp, 0); bmsr = cas_phy_read(cp, MII_BMSR); /* WTZ: Solaris driver reads this twice, but that * may be due to the PCS case and the use of a * common implementation. Read it twice here to be * safe. */ bmsr = cas_phy_read(cp, MII_BMSR); cas_mif_poll(cp, 1); readl(cp->regs + REG_MIF_STATUS); /* avoid dups */ reset = cas_mii_link_check(cp, bmsr); } else { reset = cas_pcs_link_check(cp); } if (reset) goto done; /* check for tx state machine confusion */ if ((readl(cp->regs + REG_MAC_TX_STATUS) & MAC_TX_FRAME_XMIT) == 0) { u32 val = readl(cp->regs + REG_MAC_STATE_MACHINE); u32 wptr, rptr; int tlm = CAS_VAL(MAC_SM_TLM, val); if (((tlm == 0x5) || (tlm == 0x3)) && (CAS_VAL(MAC_SM_ENCAP_SM, val) == 0)) { netif_printk(cp, tx_err, KERN_DEBUG, cp->dev, "tx err: MAC_STATE[%08x]\n", val); reset = 1; goto done; } val = readl(cp->regs + REG_TX_FIFO_PKT_CNT); wptr = readl(cp->regs + REG_TX_FIFO_WRITE_PTR); rptr = readl(cp->regs + REG_TX_FIFO_READ_PTR); if ((val == 0) && (wptr != rptr)) { netif_printk(cp, tx_err, KERN_DEBUG, cp->dev, "tx err: TX_FIFO[%08x:%08x:%08x]\n", val, wptr, rptr); reset = 1; } if (reset) cas_hard_reset(cp); } done: if (reset) { #if 1 atomic_inc(&cp->reset_task_pending); atomic_inc(&cp->reset_task_pending_all); schedule_work(&cp->reset_task); #else atomic_set(&cp->reset_task_pending, CAS_RESET_ALL); pr_err("reset called in cas_link_timer\n"); schedule_work(&cp->reset_task); #endif } if (!pending) mod_timer(&cp->link_timer, jiffies + CAS_LINK_TIMEOUT); cas_unlock_tx(cp); spin_unlock_irqrestore(&cp->lock, flags); } /* tiny buffers are used to avoid target abort issues with * older cassini's */ static void cas_tx_tiny_free(struct cas *cp) { struct pci_dev *pdev = cp->pdev; int i; for (i = 0; i < N_TX_RINGS; i++) { if (!cp->tx_tiny_bufs[i]) continue; pci_free_consistent(pdev, TX_TINY_BUF_BLOCK, cp->tx_tiny_bufs[i], cp->tx_tiny_dvma[i]); cp->tx_tiny_bufs[i] = NULL; } } static int cas_tx_tiny_alloc(struct cas *cp) { struct pci_dev *pdev = cp->pdev; int i; for (i = 0; i < N_TX_RINGS; i++) { cp->tx_tiny_bufs[i] = pci_alloc_consistent(pdev, TX_TINY_BUF_BLOCK, &cp->tx_tiny_dvma[i]); if (!cp->tx_tiny_bufs[i]) { cas_tx_tiny_free(cp); return -1; } } return 0; } static int cas_open(struct net_device *dev) { struct cas *cp = netdev_priv(dev); int hw_was_up, err; unsigned long flags; mutex_lock(&cp->pm_mutex); hw_was_up = cp->hw_running; /* The power-management mutex protects the hw_running * etc. state so it is safe to do this bit without cp->lock */ if (!cp->hw_running) { /* Reset the chip */ cas_lock_all_save(cp, flags); /* We set the second arg to cas_reset to zero * because cas_init_hw below will have its second * argument set to non-zero, which will force * autonegotiation to start. */ cas_reset(cp, 0); cp->hw_running = 1; cas_unlock_all_restore(cp, flags); } err = -ENOMEM; if (cas_tx_tiny_alloc(cp) < 0) goto err_unlock; /* alloc rx descriptors */ if (cas_alloc_rxds(cp) < 0) goto err_tx_tiny; /* allocate spares */ cas_spare_init(cp); cas_spare_recover(cp, GFP_KERNEL); /* We can now request the interrupt as we know it's masked * on the controller. cassini+ has up to 4 interrupts * that can be used, but you need to do explicit pci interrupt * mapping to expose them */ if (request_irq(cp->pdev->irq, cas_interrupt, IRQF_SHARED, dev->name, (void *) dev)) { netdev_err(cp->dev, "failed to request irq !\n"); err = -EAGAIN; goto err_spare; } #ifdef USE_NAPI napi_enable(&cp->napi); #endif /* init hw */ cas_lock_all_save(cp, flags); cas_clean_rings(cp); cas_init_hw(cp, !hw_was_up); cp->opened = 1; cas_unlock_all_restore(cp, flags); netif_start_queue(dev); mutex_unlock(&cp->pm_mutex); return 0; err_spare: cas_spare_free(cp); cas_free_rxds(cp); err_tx_tiny: cas_tx_tiny_free(cp); err_unlock: mutex_unlock(&cp->pm_mutex); return err; } static int cas_close(struct net_device *dev) { unsigned long flags; struct cas *cp = netdev_priv(dev); #ifdef USE_NAPI napi_disable(&cp->napi); #endif /* Make sure we don't get distracted by suspend/resume */ mutex_lock(&cp->pm_mutex); netif_stop_queue(dev); /* Stop traffic, mark us closed */ cas_lock_all_save(cp, flags); cp->opened = 0; cas_reset(cp, 0); cas_phy_init(cp); cas_begin_auto_negotiation(cp, NULL); cas_clean_rings(cp); cas_unlock_all_restore(cp, flags); free_irq(cp->pdev->irq, (void *) dev); cas_spare_free(cp); cas_free_rxds(cp); cas_tx_tiny_free(cp); mutex_unlock(&cp->pm_mutex); return 0; } static struct { const char name[ETH_GSTRING_LEN]; } ethtool_cassini_statnames[] = { {"collisions"}, {"rx_bytes"}, {"rx_crc_errors"}, {"rx_dropped"}, {"rx_errors"}, {"rx_fifo_errors"}, {"rx_frame_errors"}, {"rx_length_errors"}, {"rx_over_errors"}, {"rx_packets"}, {"tx_aborted_errors"}, {"tx_bytes"}, {"tx_dropped"}, {"tx_errors"}, {"tx_fifo_errors"}, {"tx_packets"} }; #define CAS_NUM_STAT_KEYS ARRAY_SIZE(ethtool_cassini_statnames) static struct { const int offsets; /* neg. values for 2nd arg to cas_read_phy */ } ethtool_register_table[] = { {-MII_BMSR}, {-MII_BMCR}, {REG_CAWR}, {REG_INF_BURST}, {REG_BIM_CFG}, {REG_RX_CFG}, {REG_HP_CFG}, {REG_MAC_TX_CFG}, {REG_MAC_RX_CFG}, {REG_MAC_CTRL_CFG}, {REG_MAC_XIF_CFG}, {REG_MIF_CFG}, {REG_PCS_CFG}, {REG_SATURN_PCFG}, {REG_PCS_MII_STATUS}, {REG_PCS_STATE_MACHINE}, {REG_MAC_COLL_EXCESS}, {REG_MAC_COLL_LATE} }; #define CAS_REG_LEN ARRAY_SIZE(ethtool_register_table) #define CAS_MAX_REGS (sizeof (u32)*CAS_REG_LEN) static void cas_read_regs(struct cas *cp, u8 *ptr, int len) { u8 *p; int i; unsigned long flags; spin_lock_irqsave(&cp->lock, flags); for (i = 0, p = ptr; i < len ; i ++, p += sizeof(u32)) { u16 hval; u32 val; if (ethtool_register_table[i].offsets < 0) { hval = cas_phy_read(cp, -ethtool_register_table[i].offsets); val = hval; } else { val= readl(cp->regs+ethtool_register_table[i].offsets); } memcpy(p, (u8 *)&val, sizeof(u32)); } spin_unlock_irqrestore(&cp->lock, flags); } static struct net_device_stats *cas_get_stats(struct net_device *dev) { struct cas *cp = netdev_priv(dev); struct net_device_stats *stats = cp->net_stats; unsigned long flags; int i; unsigned long tmp; /* we collate all of the stats into net_stats[N_TX_RING] */ if (!cp->hw_running) return stats + N_TX_RINGS; /* collect outstanding stats */ /* WTZ: the Cassini spec gives these as 16 bit counters but * stored in 32-bit words. Added a mask of 0xffff to be safe, * in case the chip somehow puts any garbage in the other bits. * Also, counter usage didn't seem to mach what Adrian did * in the parts of the code that set these quantities. Made * that consistent. */ spin_lock_irqsave(&cp->stat_lock[N_TX_RINGS], flags); stats[N_TX_RINGS].rx_crc_errors += readl(cp->regs + REG_MAC_FCS_ERR) & 0xffff; stats[N_TX_RINGS].rx_frame_errors += readl(cp->regs + REG_MAC_ALIGN_ERR) &0xffff; stats[N_TX_RINGS].rx_length_errors += readl(cp->regs + REG_MAC_LEN_ERR) & 0xffff; #if 1 tmp = (readl(cp->regs + REG_MAC_COLL_EXCESS) & 0xffff) + (readl(cp->regs + REG_MAC_COLL_LATE) & 0xffff); stats[N_TX_RINGS].tx_aborted_errors += tmp; stats[N_TX_RINGS].collisions += tmp + (readl(cp->regs + REG_MAC_COLL_NORMAL) & 0xffff); #else stats[N_TX_RINGS].tx_aborted_errors += readl(cp->regs + REG_MAC_COLL_EXCESS); stats[N_TX_RINGS].collisions += readl(cp->regs + REG_MAC_COLL_EXCESS) + readl(cp->regs + REG_MAC_COLL_LATE); #endif cas_clear_mac_err(cp); /* saved bits that are unique to ring 0 */ spin_lock(&cp->stat_lock[0]); stats[N_TX_RINGS].collisions += stats[0].collisions; stats[N_TX_RINGS].rx_over_errors += stats[0].rx_over_errors; stats[N_TX_RINGS].rx_frame_errors += stats[0].rx_frame_errors; stats[N_TX_RINGS].rx_fifo_errors += stats[0].rx_fifo_errors; stats[N_TX_RINGS].tx_aborted_errors += stats[0].tx_aborted_errors; stats[N_TX_RINGS].tx_fifo_errors += stats[0].tx_fifo_errors; spin_unlock(&cp->stat_lock[0]); for (i = 0; i < N_TX_RINGS; i++) { spin_lock(&cp->stat_lock[i]); stats[N_TX_RINGS].rx_length_errors += stats[i].rx_length_errors; stats[N_TX_RINGS].rx_crc_errors += stats[i].rx_crc_errors; stats[N_TX_RINGS].rx_packets += stats[i].rx_packets; stats[N_TX_RINGS].tx_packets += stats[i].tx_packets; stats[N_TX_RINGS].rx_bytes += stats[i].rx_bytes; stats[N_TX_RINGS].tx_bytes += stats[i].tx_bytes; stats[N_TX_RINGS].rx_errors += stats[i].rx_errors; stats[N_TX_RINGS].tx_errors += stats[i].tx_errors; stats[N_TX_RINGS].rx_dropped += stats[i].rx_dropped; stats[N_TX_RINGS].tx_dropped += stats[i].tx_dropped; memset(stats + i, 0, sizeof(struct net_device_stats)); spin_unlock(&cp->stat_lock[i]); } spin_unlock_irqrestore(&cp->stat_lock[N_TX_RINGS], flags); return stats + N_TX_RINGS; } static void cas_set_multicast(struct net_device *dev) { struct cas *cp = netdev_priv(dev); u32 rxcfg, rxcfg_new; unsigned long flags; int limit = STOP_TRIES; if (!cp->hw_running) return; spin_lock_irqsave(&cp->lock, flags); rxcfg = readl(cp->regs + REG_MAC_RX_CFG); /* disable RX MAC and wait for completion */ writel(rxcfg & ~MAC_RX_CFG_EN, cp->regs + REG_MAC_RX_CFG); while (readl(cp->regs + REG_MAC_RX_CFG) & MAC_RX_CFG_EN) { if (!limit--) break; udelay(10); } /* disable hash filter and wait for completion */ limit = STOP_TRIES; rxcfg &= ~(MAC_RX_CFG_PROMISC_EN | MAC_RX_CFG_HASH_FILTER_EN); writel(rxcfg & ~MAC_RX_CFG_EN, cp->regs + REG_MAC_RX_CFG); while (readl(cp->regs + REG_MAC_RX_CFG) & MAC_RX_CFG_HASH_FILTER_EN) { if (!limit--) break; udelay(10); } /* program hash filters */ cp->mac_rx_cfg = rxcfg_new = cas_setup_multicast(cp); rxcfg |= rxcfg_new; writel(rxcfg, cp->regs + REG_MAC_RX_CFG); spin_unlock_irqrestore(&cp->lock, flags); } static void cas_get_drvinfo(struct net_device *dev, struct ethtool_drvinfo *info) { struct cas *cp = netdev_priv(dev); strlcpy(info->driver, DRV_MODULE_NAME, sizeof(info->driver)); strlcpy(info->version, DRV_MODULE_VERSION, sizeof(info->version)); strlcpy(info->bus_info, pci_name(cp->pdev), sizeof(info->bus_info)); } static int cas_get_link_ksettings(struct net_device *dev, struct ethtool_link_ksettings *cmd) { struct cas *cp = netdev_priv(dev); u16 bmcr; int full_duplex, speed, pause; unsigned long flags; enum link_state linkstate = link_up; u32 supported, advertising; advertising = 0; supported = SUPPORTED_Autoneg; if (cp->cas_flags & CAS_FLAG_1000MB_CAP) { supported |= SUPPORTED_1000baseT_Full; advertising |= ADVERTISED_1000baseT_Full; } /* Record PHY settings if HW is on. */ spin_lock_irqsave(&cp->lock, flags); bmcr = 0; linkstate = cp->lstate; if (CAS_PHY_MII(cp->phy_type)) { cmd->base.port = PORT_MII; cmd->base.phy_address = cp->phy_addr; advertising |= ADVERTISED_TP | ADVERTISED_MII | ADVERTISED_10baseT_Half | ADVERTISED_10baseT_Full | ADVERTISED_100baseT_Half | ADVERTISED_100baseT_Full; supported |= (SUPPORTED_10baseT_Half | SUPPORTED_10baseT_Full | SUPPORTED_100baseT_Half | SUPPORTED_100baseT_Full | SUPPORTED_TP | SUPPORTED_MII); if (cp->hw_running) { cas_mif_poll(cp, 0); bmcr = cas_phy_read(cp, MII_BMCR); cas_read_mii_link_mode(cp, &full_duplex, &speed, &pause); cas_mif_poll(cp, 1); } } else { cmd->base.port = PORT_FIBRE; cmd->base.phy_address = 0; supported |= SUPPORTED_FIBRE; advertising |= ADVERTISED_FIBRE; if (cp->hw_running) { /* pcs uses the same bits as mii */ bmcr = readl(cp->regs + REG_PCS_MII_CTRL); cas_read_pcs_link_mode(cp, &full_duplex, &speed, &pause); } } spin_unlock_irqrestore(&cp->lock, flags); if (bmcr & BMCR_ANENABLE) { advertising |= ADVERTISED_Autoneg; cmd->base.autoneg = AUTONEG_ENABLE; cmd->base.speed = ((speed == 10) ? SPEED_10 : ((speed == 1000) ? SPEED_1000 : SPEED_100)); cmd->base.duplex = full_duplex ? DUPLEX_FULL : DUPLEX_HALF; } else { cmd->base.autoneg = AUTONEG_DISABLE; cmd->base.speed = ((bmcr & CAS_BMCR_SPEED1000) ? SPEED_1000 : ((bmcr & BMCR_SPEED100) ? SPEED_100 : SPEED_10)); cmd->base.duplex = (bmcr & BMCR_FULLDPLX) ? DUPLEX_FULL : DUPLEX_HALF; } if (linkstate != link_up) { /* Force these to "unknown" if the link is not up and * autonogotiation in enabled. We can set the link * speed to 0, but not cmd->duplex, * because its legal values are 0 and 1. Ethtool will * print the value reported in parentheses after the * word "Unknown" for unrecognized values. * * If in forced mode, we report the speed and duplex * settings that we configured. */ if (cp->link_cntl & BMCR_ANENABLE) { cmd->base.speed = 0; cmd->base.duplex = 0xff; } else { cmd->base.speed = SPEED_10; if (cp->link_cntl & BMCR_SPEED100) { cmd->base.speed = SPEED_100; } else if (cp->link_cntl & CAS_BMCR_SPEED1000) { cmd->base.speed = SPEED_1000; } cmd->base.duplex = (cp->link_cntl & BMCR_FULLDPLX) ? DUPLEX_FULL : DUPLEX_HALF; } } ethtool_convert_legacy_u32_to_link_mode(cmd->link_modes.supported, supported); ethtool_convert_legacy_u32_to_link_mode(cmd->link_modes.advertising, advertising); return 0; } static int cas_set_link_ksettings(struct net_device *dev, const struct ethtool_link_ksettings *cmd) { struct cas *cp = netdev_priv(dev); unsigned long flags; u32 speed = cmd->base.speed; /* Verify the settings we care about. */ if (cmd->base.autoneg != AUTONEG_ENABLE && cmd->base.autoneg != AUTONEG_DISABLE) return -EINVAL; if (cmd->base.autoneg == AUTONEG_DISABLE && ((speed != SPEED_1000 && speed != SPEED_100 && speed != SPEED_10) || (cmd->base.duplex != DUPLEX_HALF && cmd->base.duplex != DUPLEX_FULL))) return -EINVAL; /* Apply settings and restart link process. */ spin_lock_irqsave(&cp->lock, flags); cas_begin_auto_negotiation(cp, cmd); spin_unlock_irqrestore(&cp->lock, flags); return 0; } static int cas_nway_reset(struct net_device *dev) { struct cas *cp = netdev_priv(dev); unsigned long flags; if ((cp->link_cntl & BMCR_ANENABLE) == 0) return -EINVAL; /* Restart link process. */ spin_lock_irqsave(&cp->lock, flags); cas_begin_auto_negotiation(cp, NULL); spin_unlock_irqrestore(&cp->lock, flags); return 0; } static u32 cas_get_link(struct net_device *dev) { struct cas *cp = netdev_priv(dev); return cp->lstate == link_up; } static u32 cas_get_msglevel(struct net_device *dev) { struct cas *cp = netdev_priv(dev); return cp->msg_enable; } static void cas_set_msglevel(struct net_device *dev, u32 value) { struct cas *cp = netdev_priv(dev); cp->msg_enable = value; } static int cas_get_regs_len(struct net_device *dev) { struct cas *cp = netdev_priv(dev); return cp->casreg_len < CAS_MAX_REGS ? cp->casreg_len: CAS_MAX_REGS; } static void cas_get_regs(struct net_device *dev, struct ethtool_regs *regs, void *p) { struct cas *cp = netdev_priv(dev); regs->version = 0; /* cas_read_regs handles locks (cp->lock). */ cas_read_regs(cp, p, regs->len / sizeof(u32)); } static int cas_get_sset_count(struct net_device *dev, int sset) { switch (sset) { case ETH_SS_STATS: return CAS_NUM_STAT_KEYS; default: return -EOPNOTSUPP; } } static void cas_get_strings(struct net_device *dev, u32 stringset, u8 *data) { memcpy(data, &ethtool_cassini_statnames, CAS_NUM_STAT_KEYS * ETH_GSTRING_LEN); } static void cas_get_ethtool_stats(struct net_device *dev, struct ethtool_stats *estats, u64 *data) { struct cas *cp = netdev_priv(dev); struct net_device_stats *stats = cas_get_stats(cp->dev); int i = 0; data[i++] = stats->collisions; data[i++] = stats->rx_bytes; data[i++] = stats->rx_crc_errors; data[i++] = stats->rx_dropped; data[i++] = stats->rx_errors; data[i++] = stats->rx_fifo_errors; data[i++] = stats->rx_frame_errors; data[i++] = stats->rx_length_errors; data[i++] = stats->rx_over_errors; data[i++] = stats->rx_packets; data[i++] = stats->tx_aborted_errors; data[i++] = stats->tx_bytes; data[i++] = stats->tx_dropped; data[i++] = stats->tx_errors; data[i++] = stats->tx_fifo_errors; data[i++] = stats->tx_packets; BUG_ON(i != CAS_NUM_STAT_KEYS); } static const struct ethtool_ops cas_ethtool_ops = { .get_drvinfo = cas_get_drvinfo, .nway_reset = cas_nway_reset, .get_link = cas_get_link, .get_msglevel = cas_get_msglevel, .set_msglevel = cas_set_msglevel, .get_regs_len = cas_get_regs_len, .get_regs = cas_get_regs, .get_sset_count = cas_get_sset_count, .get_strings = cas_get_strings, .get_ethtool_stats = cas_get_ethtool_stats, .get_link_ksettings = cas_get_link_ksettings, .set_link_ksettings = cas_set_link_ksettings, }; static int cas_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd) { struct cas *cp = netdev_priv(dev); struct mii_ioctl_data *data = if_mii(ifr); unsigned long flags; int rc = -EOPNOTSUPP; /* Hold the PM mutex while doing ioctl's or we may collide * with open/close and power management and oops. */ mutex_lock(&cp->pm_mutex); switch (cmd) { case SIOCGMIIPHY: /* Get address of MII PHY in use. */ data->phy_id = cp->phy_addr; /* Fallthrough... */ case SIOCGMIIREG: /* Read MII PHY register. */ spin_lock_irqsave(&cp->lock, flags); cas_mif_poll(cp, 0); data->val_out = cas_phy_read(cp, data->reg_num & 0x1f); cas_mif_poll(cp, 1); spin_unlock_irqrestore(&cp->lock, flags); rc = 0; break; case SIOCSMIIREG: /* Write MII PHY register. */ spin_lock_irqsave(&cp->lock, flags); cas_mif_poll(cp, 0); rc = cas_phy_write(cp, data->reg_num & 0x1f, data->val_in); cas_mif_poll(cp, 1); spin_unlock_irqrestore(&cp->lock, flags); break; default: break; } mutex_unlock(&cp->pm_mutex); return rc; } /* When this chip sits underneath an Intel 31154 bridge, it is the * only subordinate device and we can tweak the bridge settings to * reflect that fact. */ static void cas_program_bridge(struct pci_dev *cas_pdev) { struct pci_dev *pdev = cas_pdev->bus->self; u32 val; if (!pdev) return; if (pdev->vendor != 0x8086 || pdev->device != 0x537c) return; /* Clear bit 10 (Bus Parking Control) in the Secondary * Arbiter Control/Status Register which lives at offset * 0x41. Using a 32-bit word read/modify/write at 0x40 * is much simpler so that's how we do this. */ pci_read_config_dword(pdev, 0x40, &val); val &= ~0x00040000; pci_write_config_dword(pdev, 0x40, val); /* Max out the Multi-Transaction Timer settings since * Cassini is the only device present. * * The register is 16-bit and lives at 0x50. When the * settings are enabled, it extends the GRANT# signal * for a requestor after a transaction is complete. This * allows the next request to run without first needing * to negotiate the GRANT# signal back. * * Bits 12:10 define the grant duration: * * 1 -- 16 clocks * 2 -- 32 clocks * 3 -- 64 clocks * 4 -- 128 clocks * 5 -- 256 clocks * * All other values are illegal. * * Bits 09:00 define which REQ/GNT signal pairs get the * GRANT# signal treatment. We set them all. */ pci_write_config_word(pdev, 0x50, (5 << 10) | 0x3ff); /* The Read Prefecth Policy register is 16-bit and sits at * offset 0x52. It enables a "smart" pre-fetch policy. We * enable it and max out all of the settings since only one * device is sitting underneath and thus bandwidth sharing is * not an issue. * * The register has several 3 bit fields, which indicates a * multiplier applied to the base amount of prefetching the * chip would do. These fields are at: * * 15:13 --- ReRead Primary Bus * 12:10 --- FirstRead Primary Bus * 09:07 --- ReRead Secondary Bus * 06:04 --- FirstRead Secondary Bus * * Bits 03:00 control which REQ/GNT pairs the prefetch settings * get enabled on. Bit 3 is a grouped enabler which controls * all of the REQ/GNT pairs from [8:3]. Bits 2 to 0 control * the individual REQ/GNT pairs [2:0]. */ pci_write_config_word(pdev, 0x52, (0x7 << 13) | (0x7 << 10) | (0x7 << 7) | (0x7 << 4) | (0xf << 0)); /* Force cacheline size to 0x8 */ pci_write_config_byte(pdev, PCI_CACHE_LINE_SIZE, 0x08); /* Force latency timer to maximum setting so Cassini can * sit on the bus as long as it likes. */ pci_write_config_byte(pdev, PCI_LATENCY_TIMER, 0xff); } static const struct net_device_ops cas_netdev_ops = { .ndo_open = cas_open, .ndo_stop = cas_close, .ndo_start_xmit = cas_start_xmit, .ndo_get_stats = cas_get_stats, .ndo_set_rx_mode = cas_set_multicast, .ndo_do_ioctl = cas_ioctl, .ndo_tx_timeout = cas_tx_timeout, .ndo_change_mtu = cas_change_mtu, .ndo_set_mac_address = eth_mac_addr, .ndo_validate_addr = eth_validate_addr, #ifdef CONFIG_NET_POLL_CONTROLLER .ndo_poll_controller = cas_netpoll, #endif }; static int cas_init_one(struct pci_dev *pdev, const struct pci_device_id *ent) { static int cas_version_printed = 0; unsigned long casreg_len; struct net_device *dev; struct cas *cp; int i, err, pci_using_dac; u16 pci_cmd; u8 orig_cacheline_size = 0, cas_cacheline_size = 0; if (cas_version_printed++ == 0) pr_info("%s", version); err = pci_enable_device(pdev); if (err) { dev_err(&pdev->dev, "Cannot enable PCI device, aborting\n"); return err; } if (!(pci_resource_flags(pdev, 0) & IORESOURCE_MEM)) { dev_err(&pdev->dev, "Cannot find proper PCI device " "base address, aborting\n"); err = -ENODEV; goto err_out_disable_pdev; } dev = alloc_etherdev(sizeof(*cp)); if (!dev) { err = -ENOMEM; goto err_out_disable_pdev; } SET_NETDEV_DEV(dev, &pdev->dev); err = pci_request_regions(pdev, dev->name); if (err) { dev_err(&pdev->dev, "Cannot obtain PCI resources, aborting\n"); goto err_out_free_netdev; } pci_set_master(pdev); /* we must always turn on parity response or else parity * doesn't get generated properly. disable SERR/PERR as well. * in addition, we want to turn MWI on. */ pci_read_config_word(pdev, PCI_COMMAND, &pci_cmd); pci_cmd &= ~PCI_COMMAND_SERR; pci_cmd |= PCI_COMMAND_PARITY; pci_write_config_word(pdev, PCI_COMMAND, pci_cmd); if (pci_try_set_mwi(pdev)) pr_warn("Could not enable MWI for %s\n", pci_name(pdev)); cas_program_bridge(pdev); /* * On some architectures, the default cache line size set * by pci_try_set_mwi reduces perforamnce. We have to increase * it for this case. To start, we'll print some configuration * data. */ #if 1 pci_read_config_byte(pdev, PCI_CACHE_LINE_SIZE, &orig_cacheline_size); if (orig_cacheline_size < CAS_PREF_CACHELINE_SIZE) { cas_cacheline_size = (CAS_PREF_CACHELINE_SIZE < SMP_CACHE_BYTES) ? CAS_PREF_CACHELINE_SIZE : SMP_CACHE_BYTES; if (pci_write_config_byte(pdev, PCI_CACHE_LINE_SIZE, cas_cacheline_size)) { dev_err(&pdev->dev, "Could not set PCI cache " "line size\n"); goto err_out_free_res; } } #endif /* Configure DMA attributes. */ if (!pci_set_dma_mask(pdev, DMA_BIT_MASK(64))) { pci_using_dac = 1; err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(64)); if (err < 0) { dev_err(&pdev->dev, "Unable to obtain 64-bit DMA " "for consistent allocations\n"); goto err_out_free_res; } } else { err = pci_set_dma_mask(pdev, DMA_BIT_MASK(32)); if (err) { dev_err(&pdev->dev, "No usable DMA configuration, " "aborting\n"); goto err_out_free_res; } pci_using_dac = 0; } casreg_len = pci_resource_len(pdev, 0); cp = netdev_priv(dev); cp->pdev = pdev; #if 1 /* A value of 0 indicates we never explicitly set it */ cp->orig_cacheline_size = cas_cacheline_size ? orig_cacheline_size: 0; #endif cp->dev = dev; cp->msg_enable = (cassini_debug < 0) ? CAS_DEF_MSG_ENABLE : cassini_debug; #if defined(CONFIG_SPARC) cp->of_node = pci_device_to_OF_node(pdev); #endif cp->link_transition = LINK_TRANSITION_UNKNOWN; cp->link_transition_jiffies_valid = 0; spin_lock_init(&cp->lock); spin_lock_init(&cp->rx_inuse_lock); spin_lock_init(&cp->rx_spare_lock); for (i = 0; i < N_TX_RINGS; i++) { spin_lock_init(&cp->stat_lock[i]); spin_lock_init(&cp->tx_lock[i]); } spin_lock_init(&cp->stat_lock[N_TX_RINGS]); mutex_init(&cp->pm_mutex); timer_setup(&cp->link_timer, cas_link_timer, 0); #if 1 /* Just in case the implementation of atomic operations * change so that an explicit initialization is necessary. */ atomic_set(&cp->reset_task_pending, 0); atomic_set(&cp->reset_task_pending_all, 0); atomic_set(&cp->reset_task_pending_spare, 0); atomic_set(&cp->reset_task_pending_mtu, 0); #endif INIT_WORK(&cp->reset_task, cas_reset_task); /* Default link parameters */ if (link_mode >= 0 && link_mode < 6) cp->link_cntl = link_modes[link_mode]; else cp->link_cntl = BMCR_ANENABLE; cp->lstate = link_down; cp->link_transition = LINK_TRANSITION_LINK_DOWN; netif_carrier_off(cp->dev); cp->timer_ticks = 0; /* give us access to cassini registers */ cp->regs = pci_iomap(pdev, 0, casreg_len); if (!cp->regs) { dev_err(&pdev->dev, "Cannot map device registers, aborting\n"); goto err_out_free_res; } cp->casreg_len = casreg_len; pci_save_state(pdev); cas_check_pci_invariants(cp); cas_hard_reset(cp); cas_reset(cp, 0); if (cas_check_invariants(cp)) goto err_out_iounmap; if (cp->cas_flags & CAS_FLAG_SATURN) cas_saturn_firmware_init(cp); cp->init_block = (struct cas_init_block *) pci_alloc_consistent(pdev, sizeof(struct cas_init_block), &cp->block_dvma); if (!cp->init_block) { dev_err(&pdev->dev, "Cannot allocate init block, aborting\n"); goto err_out_iounmap; } for (i = 0; i < N_TX_RINGS; i++) cp->init_txds[i] = cp->init_block->txds[i]; for (i = 0; i < N_RX_DESC_RINGS; i++) cp->init_rxds[i] = cp->init_block->rxds[i]; for (i = 0; i < N_RX_COMP_RINGS; i++) cp->init_rxcs[i] = cp->init_block->rxcs[i]; for (i = 0; i < N_RX_FLOWS; i++) skb_queue_head_init(&cp->rx_flows[i]); dev->netdev_ops = &cas_netdev_ops; dev->ethtool_ops = &cas_ethtool_ops; dev->watchdog_timeo = CAS_TX_TIMEOUT; #ifdef USE_NAPI netif_napi_add(dev, &cp->napi, cas_poll, 64); #endif dev->irq = pdev->irq; dev->dma = 0; /* Cassini features. */ if ((cp->cas_flags & CAS_FLAG_NO_HW_CSUM) == 0) dev->features |= NETIF_F_HW_CSUM | NETIF_F_SG; if (pci_using_dac) dev->features |= NETIF_F_HIGHDMA; /* MTU range: 60 - varies or 9000 */ dev->min_mtu = CAS_MIN_MTU; dev->max_mtu = CAS_MAX_MTU; if (register_netdev(dev)) { dev_err(&pdev->dev, "Cannot register net device, aborting\n"); goto err_out_free_consistent; } i = readl(cp->regs + REG_BIM_CFG); netdev_info(dev, "Sun Cassini%s (%sbit/%sMHz PCI/%s) Ethernet[%d] %pM\n", (cp->cas_flags & CAS_FLAG_REG_PLUS) ? "+" : "", (i & BIM_CFG_32BIT) ? "32" : "64", (i & BIM_CFG_66MHZ) ? "66" : "33", (cp->phy_type == CAS_PHY_SERDES) ? "Fi" : "Cu", pdev->irq, dev->dev_addr); pci_set_drvdata(pdev, dev); cp->hw_running = 1; cas_entropy_reset(cp); cas_phy_init(cp); cas_begin_auto_negotiation(cp, NULL); return 0; err_out_free_consistent: pci_free_consistent(pdev, sizeof(struct cas_init_block), cp->init_block, cp->block_dvma); err_out_iounmap: mutex_lock(&cp->pm_mutex); if (cp->hw_running) cas_shutdown(cp); mutex_unlock(&cp->pm_mutex); pci_iounmap(pdev, cp->regs); err_out_free_res: pci_release_regions(pdev); /* Try to restore it in case the error occurred after we * set it. */ pci_write_config_byte(pdev, PCI_CACHE_LINE_SIZE, orig_cacheline_size); err_out_free_netdev: free_netdev(dev); err_out_disable_pdev: pci_disable_device(pdev); return -ENODEV; } static void cas_remove_one(struct pci_dev *pdev) { struct net_device *dev = pci_get_drvdata(pdev); struct cas *cp; if (!dev) return; cp = netdev_priv(dev); unregister_netdev(dev); vfree(cp->fw_data); mutex_lock(&cp->pm_mutex); cancel_work_sync(&cp->reset_task); if (cp->hw_running) cas_shutdown(cp); mutex_unlock(&cp->pm_mutex); #if 1 if (cp->orig_cacheline_size) { /* Restore the cache line size if we had modified * it. */ pci_write_config_byte(pdev, PCI_CACHE_LINE_SIZE, cp->orig_cacheline_size); } #endif pci_free_consistent(pdev, sizeof(struct cas_init_block), cp->init_block, cp->block_dvma); pci_iounmap(pdev, cp->regs); free_netdev(dev); pci_release_regions(pdev); pci_disable_device(pdev); } #ifdef CONFIG_PM static int cas_suspend(struct pci_dev *pdev, pm_message_t state) { struct net_device *dev = pci_get_drvdata(pdev); struct cas *cp = netdev_priv(dev); unsigned long flags; mutex_lock(&cp->pm_mutex); /* If the driver is opened, we stop the DMA */ if (cp->opened) { netif_device_detach(dev); cas_lock_all_save(cp, flags); /* We can set the second arg of cas_reset to 0 * because on resume, we'll call cas_init_hw with * its second arg set so that autonegotiation is * restarted. */ cas_reset(cp, 0); cas_clean_rings(cp); cas_unlock_all_restore(cp, flags); } if (cp->hw_running) cas_shutdown(cp); mutex_unlock(&cp->pm_mutex); return 0; } static int cas_resume(struct pci_dev *pdev) { struct net_device *dev = pci_get_drvdata(pdev); struct cas *cp = netdev_priv(dev); netdev_info(dev, "resuming\n"); mutex_lock(&cp->pm_mutex); cas_hard_reset(cp); if (cp->opened) { unsigned long flags; cas_lock_all_save(cp, flags); cas_reset(cp, 0); cp->hw_running = 1; cas_clean_rings(cp); cas_init_hw(cp, 1); cas_unlock_all_restore(cp, flags); netif_device_attach(dev); } mutex_unlock(&cp->pm_mutex); return 0; } #endif /* CONFIG_PM */ static struct pci_driver cas_driver = { .name = DRV_MODULE_NAME, .id_table = cas_pci_tbl, .probe = cas_init_one, .remove = cas_remove_one, #ifdef CONFIG_PM .suspend = cas_suspend, .resume = cas_resume #endif }; static int __init cas_init(void) { if (linkdown_timeout > 0) link_transition_timeout = linkdown_timeout * HZ; else link_transition_timeout = 0; return pci_register_driver(&cas_driver); } static void __exit cas_cleanup(void) { pci_unregister_driver(&cas_driver); } module_init(cas_init); module_exit(cas_cleanup);